2009-12-08 22:09:11 +00:00
|
|
|
/*
|
|
|
|
|
* Disk Array driver for HP Smart Array SAS controllers
|
2016-02-23 21:16:46 +00:00
|
|
|
* Copyright 2016 Microsemi Corporation
|
2015-07-18 16:12:38 +00:00
|
|
|
* Copyright 2014-2015 PMC-Sierra, Inc.
|
|
|
|
|
* Copyright 2000,2009-2015 Hewlett-Packard Development Company, L.P.
|
2009-12-08 22:09:11 +00:00
|
|
|
*
|
|
|
|
|
* This program is free software; you can redistribute it and/or modify
|
|
|
|
|
* it under the terms of the GNU General Public License as published by
|
|
|
|
|
* the Free Software Foundation; version 2 of the License.
|
|
|
|
|
*
|
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
|
* MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
|
|
|
|
|
* NON INFRINGEMENT. See the GNU General Public License for more details.
|
|
|
|
|
*
|
2016-02-23 21:16:46 +00:00
|
|
|
* Questions/Comments/Bugfixes to esc.storagedev@microsemi.com
|
2009-12-08 22:09:11 +00:00
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#include <linux/module.h>
|
|
|
|
|
#include <linux/interrupt.h>
|
|
|
|
|
#include <linux/types.h>
|
|
|
|
|
#include <linux/pci.h>
|
2011-11-11 16:14:23 +00:00
|
|
|
#include <linux/pci-aspm.h>
|
2009-12-08 22:09:11 +00:00
|
|
|
#include <linux/kernel.h>
|
|
|
|
|
#include <linux/slab.h>
|
|
|
|
|
#include <linux/delay.h>
|
|
|
|
|
#include <linux/fs.h>
|
|
|
|
|
#include <linux/timer.h>
|
|
|
|
|
#include <linux/init.h>
|
|
|
|
|
#include <linux/spinlock.h>
|
|
|
|
|
#include <linux/compat.h>
|
|
|
|
|
#include <linux/blktrace_api.h>
|
|
|
|
|
#include <linux/uaccess.h>
|
|
|
|
|
#include <linux/io.h>
|
|
|
|
|
#include <linux/dma-mapping.h>
|
|
|
|
|
#include <linux/completion.h>
|
|
|
|
|
#include <linux/moduleparam.h>
|
|
|
|
|
#include <scsi/scsi.h>
|
|
|
|
|
#include <scsi/scsi_cmnd.h>
|
|
|
|
|
#include <scsi/scsi_device.h>
|
|
|
|
|
#include <scsi/scsi_host.h>
|
2010-02-25 20:02:51 +00:00
|
|
|
#include <scsi/scsi_tcq.h>
|
2015-04-23 14:32:16 +00:00
|
|
|
#include <scsi/scsi_eh.h>
|
2015-11-04 21:52:34 +00:00
|
|
|
#include <scsi/scsi_transport_sas.h>
|
2015-04-23 14:35:04 +00:00
|
|
|
#include <scsi/scsi_dbg.h>
|
2009-12-08 22:09:11 +00:00
|
|
|
#include <linux/cciss_ioctl.h>
|
|
|
|
|
#include <linux/string.h>
|
|
|
|
|
#include <linux/bitmap.h>
|
2011-07-26 23:09:06 +00:00
|
|
|
#include <linux/atomic.h>
|
2011-10-26 21:22:04 +00:00
|
|
|
#include <linux/jiffies.h>
|
2014-11-14 23:26:27 +00:00
|
|
|
#include <linux/percpu-defs.h>
|
2014-05-29 15:53:18 +00:00
|
|
|
#include <linux/percpu.h>
|
2015-01-23 22:41:09 +00:00
|
|
|
#include <asm/unaligned.h>
|
2014-02-18 19:55:33 +00:00
|
|
|
#include <asm/div64.h>
|
2009-12-08 22:09:11 +00:00
|
|
|
#include "hpsa_cmd.h"
|
|
|
|
|
#include "hpsa.h"
|
|
|
|
|
|
2015-11-04 21:52:40 +00:00
|
|
|
/*
|
|
|
|
|
* HPSA_DRIVER_VERSION must be 3 byte values (0-255) separated by '.'
|
|
|
|
|
* with an optional trailing '-' followed by a byte value (0-255).
|
|
|
|
|
*/
|
|
|
|
|
#define HPSA_DRIVER_VERSION "3.4.14-0"
|
2009-12-08 22:09:11 +00:00
|
|
|
#define DRIVER_NAME "HP HPSA Driver (v " HPSA_DRIVER_VERSION ")"
|
2012-01-19 20:00:59 +00:00
|
|
|
#define HPSA "hpsa"
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-01-23 22:44:56 +00:00
|
|
|
/* How long to wait for CISS doorbell communication */
|
|
|
|
|
#define CLEAR_EVENT_WAIT_INTERVAL 20 /* ms for each msleep() call */
|
|
|
|
|
#define MODE_CHANGE_WAIT_INTERVAL 10 /* ms for each msleep() call */
|
|
|
|
|
#define MAX_CLEAR_EVENT_WAIT 30000 /* times 20 ms = 600 s */
|
|
|
|
|
#define MAX_MODE_CHANGE_WAIT 2000 /* times 10 ms = 20 s */
|
2009-12-08 22:09:11 +00:00
|
|
|
#define MAX_IOCTL_CONFIG_WAIT 1000
|
|
|
|
|
|
|
|
|
|
/*define how many times we will try a command because of bus resets */
|
|
|
|
|
#define MAX_CMD_RETRIES 3
|
|
|
|
|
|
|
|
|
|
/* Embedded module documentation macros - see modules.h */
|
|
|
|
|
MODULE_AUTHOR("Hewlett-Packard Company");
|
|
|
|
|
MODULE_DESCRIPTION("Driver for HP Smart Array Controller version " \
|
|
|
|
|
HPSA_DRIVER_VERSION);
|
|
|
|
|
MODULE_SUPPORTED_DEVICE("HP Smart Array Controllers");
|
|
|
|
|
MODULE_VERSION(HPSA_DRIVER_VERSION);
|
|
|
|
|
MODULE_LICENSE("GPL");
|
|
|
|
|
|
|
|
|
|
static int hpsa_allow_any;
|
|
|
|
|
module_param(hpsa_allow_any, int, S_IRUGO|S_IWUSR);
|
|
|
|
|
MODULE_PARM_DESC(hpsa_allow_any,
|
|
|
|
|
"Allow hpsa driver to access unknown HP Smart Array hardware");
|
2011-01-06 20:48:29 +00:00
|
|
|
static int hpsa_simple_mode;
|
|
|
|
|
module_param(hpsa_simple_mode, int, S_IRUGO|S_IWUSR);
|
|
|
|
|
MODULE_PARM_DESC(hpsa_simple_mode,
|
|
|
|
|
"Use 'simple mode' rather than 'performant mode'");
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
/* define the PCI info for the cards we can control */
|
|
|
|
|
static const struct pci_device_id hpsa_pci_device_id[] = {
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSE, 0x103C, 0x3241},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSE, 0x103C, 0x3243},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSE, 0x103C, 0x3245},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSE, 0x103C, 0x3247},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSE, 0x103C, 0x3249},
|
2013-09-04 20:11:10 +00:00
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSE, 0x103C, 0x324A},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSE, 0x103C, 0x324B},
|
2010-02-04 14:42:45 +00:00
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSE, 0x103C, 0x3233},
|
2011-03-07 16:44:16 +00:00
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSF, 0x103C, 0x3350},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSF, 0x103C, 0x3351},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSF, 0x103C, 0x3352},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSF, 0x103C, 0x3353},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSF, 0x103C, 0x3354},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSF, 0x103C, 0x3355},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSF, 0x103C, 0x3356},
|
2012-09-20 21:05:18 +00:00
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSH, 0x103C, 0x1921},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSH, 0x103C, 0x1922},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSH, 0x103C, 0x1923},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSH, 0x103C, 0x1924},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSH, 0x103C, 0x1926},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSH, 0x103C, 0x1928},
|
2013-09-04 20:05:55 +00:00
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSH, 0x103C, 0x1929},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21BD},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21BE},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21BF},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21C0},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21C1},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21C2},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21C3},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21C4},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21C5},
|
2014-05-08 19:27:24 +00:00
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21C6},
|
2013-09-04 20:05:55 +00:00
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21C7},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21C8},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21C9},
|
2014-05-08 19:27:24 +00:00
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21CA},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21CB},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21CC},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21CD},
|
|
|
|
|
{PCI_VENDOR_ID_HP, PCI_DEVICE_ID_HP_CISSI, 0x103C, 0x21CE},
|
2015-04-23 14:35:27 +00:00
|
|
|
{PCI_VENDOR_ID_ADAPTEC2, 0x0290, 0x9005, 0x0580},
|
2015-07-18 16:12:54 +00:00
|
|
|
{PCI_VENDOR_ID_ADAPTEC2, 0x0290, 0x9005, 0x0581},
|
|
|
|
|
{PCI_VENDOR_ID_ADAPTEC2, 0x0290, 0x9005, 0x0582},
|
|
|
|
|
{PCI_VENDOR_ID_ADAPTEC2, 0x0290, 0x9005, 0x0583},
|
|
|
|
|
{PCI_VENDOR_ID_ADAPTEC2, 0x0290, 0x9005, 0x0584},
|
|
|
|
|
{PCI_VENDOR_ID_ADAPTEC2, 0x0290, 0x9005, 0x0585},
|
2014-02-18 19:58:02 +00:00
|
|
|
{PCI_VENDOR_ID_HP_3PAR, 0x0075, 0x1590, 0x0076},
|
|
|
|
|
{PCI_VENDOR_ID_HP_3PAR, 0x0075, 0x1590, 0x0087},
|
|
|
|
|
{PCI_VENDOR_ID_HP_3PAR, 0x0075, 0x1590, 0x007D},
|
|
|
|
|
{PCI_VENDOR_ID_HP_3PAR, 0x0075, 0x1590, 0x0088},
|
|
|
|
|
{PCI_VENDOR_ID_HP, 0x333f, 0x103c, 0x333f},
|
2010-12-01 17:16:07 +00:00
|
|
|
{PCI_VENDOR_ID_HP, PCI_ANY_ID, PCI_ANY_ID, PCI_ANY_ID,
|
2010-06-16 18:51:20 +00:00
|
|
|
PCI_CLASS_STORAGE_RAID << 8, 0xffff << 8, 0},
|
2009-12-08 22:09:11 +00:00
|
|
|
{0,}
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
MODULE_DEVICE_TABLE(pci, hpsa_pci_device_id);
|
|
|
|
|
|
|
|
|
|
/* board_id = Subsystem Device ID & Vendor ID
|
|
|
|
|
* product = Marketing Name for the board
|
|
|
|
|
* access = Address of the struct of function pointers
|
|
|
|
|
*/
|
|
|
|
|
static struct board_type products[] = {
|
|
|
|
|
{0x3241103C, "Smart Array P212", &SA5_access},
|
|
|
|
|
{0x3243103C, "Smart Array P410", &SA5_access},
|
|
|
|
|
{0x3245103C, "Smart Array P410i", &SA5_access},
|
|
|
|
|
{0x3247103C, "Smart Array P411", &SA5_access},
|
|
|
|
|
{0x3249103C, "Smart Array P812", &SA5_access},
|
2013-09-04 20:11:10 +00:00
|
|
|
{0x324A103C, "Smart Array P712m", &SA5_access},
|
|
|
|
|
{0x324B103C, "Smart Array P711m", &SA5_access},
|
2014-11-14 23:26:38 +00:00
|
|
|
{0x3233103C, "HP StorageWorks 1210m", &SA5_access}, /* alias of 333f */
|
2012-09-20 21:05:18 +00:00
|
|
|
{0x3350103C, "Smart Array P222", &SA5_access},
|
|
|
|
|
{0x3351103C, "Smart Array P420", &SA5_access},
|
|
|
|
|
{0x3352103C, "Smart Array P421", &SA5_access},
|
|
|
|
|
{0x3353103C, "Smart Array P822", &SA5_access},
|
|
|
|
|
{0x3354103C, "Smart Array P420i", &SA5_access},
|
|
|
|
|
{0x3355103C, "Smart Array P220i", &SA5_access},
|
|
|
|
|
{0x3356103C, "Smart Array P721m", &SA5_access},
|
2013-09-04 20:08:29 +00:00
|
|
|
{0x1921103C, "Smart Array P830i", &SA5_access},
|
|
|
|
|
{0x1922103C, "Smart Array P430", &SA5_access},
|
|
|
|
|
{0x1923103C, "Smart Array P431", &SA5_access},
|
|
|
|
|
{0x1924103C, "Smart Array P830", &SA5_access},
|
|
|
|
|
{0x1926103C, "Smart Array P731m", &SA5_access},
|
|
|
|
|
{0x1928103C, "Smart Array P230i", &SA5_access},
|
|
|
|
|
{0x1929103C, "Smart Array P530", &SA5_access},
|
2015-01-23 22:45:07 +00:00
|
|
|
{0x21BD103C, "Smart Array P244br", &SA5_access},
|
|
|
|
|
{0x21BE103C, "Smart Array P741m", &SA5_access},
|
|
|
|
|
{0x21BF103C, "Smart HBA H240ar", &SA5_access},
|
|
|
|
|
{0x21C0103C, "Smart Array P440ar", &SA5_access},
|
2015-01-23 22:45:12 +00:00
|
|
|
{0x21C1103C, "Smart Array P840ar", &SA5_access},
|
2015-01-23 22:45:07 +00:00
|
|
|
{0x21C2103C, "Smart Array P440", &SA5_access},
|
|
|
|
|
{0x21C3103C, "Smart Array P441", &SA5_access},
|
2013-09-04 20:05:55 +00:00
|
|
|
{0x21C4103C, "Smart Array", &SA5_access},
|
2015-01-23 22:45:07 +00:00
|
|
|
{0x21C5103C, "Smart Array P841", &SA5_access},
|
|
|
|
|
{0x21C6103C, "Smart HBA H244br", &SA5_access},
|
|
|
|
|
{0x21C7103C, "Smart HBA H240", &SA5_access},
|
|
|
|
|
{0x21C8103C, "Smart HBA H241", &SA5_access},
|
2013-09-04 20:05:55 +00:00
|
|
|
{0x21C9103C, "Smart Array", &SA5_access},
|
2015-01-23 22:45:07 +00:00
|
|
|
{0x21CA103C, "Smart Array P246br", &SA5_access},
|
|
|
|
|
{0x21CB103C, "Smart Array P840", &SA5_access},
|
2014-05-08 19:27:24 +00:00
|
|
|
{0x21CC103C, "Smart Array", &SA5_access},
|
|
|
|
|
{0x21CD103C, "Smart Array", &SA5_access},
|
2015-01-23 22:45:07 +00:00
|
|
|
{0x21CE103C, "Smart HBA", &SA5_access},
|
2015-04-23 14:35:27 +00:00
|
|
|
{0x05809005, "SmartHBA-SA", &SA5_access},
|
2015-07-18 16:12:54 +00:00
|
|
|
{0x05819005, "SmartHBA-SA 8i", &SA5_access},
|
|
|
|
|
{0x05829005, "SmartHBA-SA 8i8e", &SA5_access},
|
|
|
|
|
{0x05839005, "SmartHBA-SA 8e", &SA5_access},
|
|
|
|
|
{0x05849005, "SmartHBA-SA 16i", &SA5_access},
|
|
|
|
|
{0x05859005, "SmartHBA-SA 4i4e", &SA5_access},
|
2014-02-18 19:58:02 +00:00
|
|
|
{0x00761590, "HP Storage P1224 Array Controller", &SA5_access},
|
|
|
|
|
{0x00871590, "HP Storage P1224e Array Controller", &SA5_access},
|
|
|
|
|
{0x007D1590, "HP Storage P1228 Array Controller", &SA5_access},
|
|
|
|
|
{0x00881590, "HP Storage P1228e Array Controller", &SA5_access},
|
|
|
|
|
{0x333f103c, "HP StorageWorks 1210m Array Controller", &SA5_access},
|
2009-12-08 22:09:11 +00:00
|
|
|
{0xFFFF103C, "Unknown Smart Array", &SA5_access},
|
|
|
|
|
};
|
|
|
|
|
|
2015-11-04 21:52:34 +00:00
|
|
|
static struct scsi_transport_template *hpsa_sas_transport_template;
|
|
|
|
|
static int hpsa_add_sas_host(struct ctlr_info *h);
|
|
|
|
|
static void hpsa_delete_sas_host(struct ctlr_info *h);
|
|
|
|
|
static int hpsa_add_sas_device(struct hpsa_sas_node *hpsa_sas_node,
|
|
|
|
|
struct hpsa_scsi_dev_t *device);
|
|
|
|
|
static void hpsa_remove_sas_device(struct hpsa_scsi_dev_t *device);
|
|
|
|
|
static struct hpsa_scsi_dev_t
|
|
|
|
|
*hpsa_find_device_by_sas_rphy(struct ctlr_info *h,
|
|
|
|
|
struct sas_rphy *rphy);
|
|
|
|
|
|
2015-04-23 14:34:16 +00:00
|
|
|
#define SCSI_CMD_BUSY ((struct scsi_cmnd *)&hpsa_cmd_busy)
|
|
|
|
|
static const struct scsi_cmnd hpsa_cmd_busy;
|
|
|
|
|
#define SCSI_CMD_IDLE ((struct scsi_cmnd *)&hpsa_cmd_idle)
|
|
|
|
|
static const struct scsi_cmnd hpsa_cmd_idle;
|
2009-12-08 22:09:11 +00:00
|
|
|
static int number_of_controllers;
|
|
|
|
|
|
2010-06-16 18:51:50 +00:00
|
|
|
static irqreturn_t do_hpsa_intr_intx(int irq, void *dev_id);
|
|
|
|
|
static irqreturn_t do_hpsa_intr_msi(int irq, void *dev_id);
|
2014-11-14 23:26:27 +00:00
|
|
|
static int hpsa_ioctl(struct scsi_device *dev, int cmd, void __user *arg);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
#ifdef CONFIG_COMPAT
|
2014-11-14 23:26:27 +00:00
|
|
|
static int hpsa_compat_ioctl(struct scsi_device *dev, int cmd,
|
|
|
|
|
void __user *arg);
|
2009-12-08 22:09:11 +00:00
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
static void cmd_free(struct ctlr_info *h, struct CommandList *c);
|
|
|
|
|
static struct CommandList *cmd_alloc(struct ctlr_info *h);
|
2015-04-23 14:35:04 +00:00
|
|
|
static void cmd_tagged_free(struct ctlr_info *h, struct CommandList *c);
|
|
|
|
|
static struct CommandList *cmd_tagged_alloc(struct ctlr_info *h,
|
|
|
|
|
struct scsi_cmnd *scmd);
|
2013-02-20 17:24:41 +00:00
|
|
|
static int fill_cmd(struct CommandList *c, u8 cmd, struct ctlr_info *h,
|
2014-02-18 19:57:11 +00:00
|
|
|
void *buff, size_t size, u16 page_code, unsigned char *scsi3addr,
|
2009-12-08 22:09:11 +00:00
|
|
|
int cmd_type);
|
2015-01-23 22:42:48 +00:00
|
|
|
static void hpsa_free_cmd_pool(struct ctlr_info *h);
|
2014-02-18 19:57:11 +00:00
|
|
|
#define VPD_PAGE (1 << 8)
|
2015-11-04 21:50:13 +00:00
|
|
|
#define HPSA_SIMPLE_ERROR_BITS 0x03
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2010-11-16 07:10:29 +00:00
|
|
|
static int hpsa_scsi_queue_command(struct Scsi_Host *h, struct scsi_cmnd *cmd);
|
2010-02-04 14:43:16 +00:00
|
|
|
static void hpsa_scan_start(struct Scsi_Host *);
|
|
|
|
|
static int hpsa_scan_finished(struct Scsi_Host *sh,
|
|
|
|
|
unsigned long elapsed_time);
|
2015-01-23 22:41:30 +00:00
|
|
|
static int hpsa_change_queue_depth(struct scsi_device *sdev, int qdepth);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
static int hpsa_eh_device_reset_handler(struct scsi_cmnd *scsicmd);
|
2012-05-01 16:42:51 +00:00
|
|
|
static int hpsa_eh_abort_handler(struct scsi_cmnd *scsicmd);
|
2009-12-08 22:09:11 +00:00
|
|
|
static int hpsa_slave_alloc(struct scsi_device *sdev);
|
2015-04-23 14:31:47 +00:00
|
|
|
static int hpsa_slave_configure(struct scsi_device *sdev);
|
2009-12-08 22:09:11 +00:00
|
|
|
static void hpsa_slave_destroy(struct scsi_device *sdev);
|
|
|
|
|
|
2015-11-04 21:50:01 +00:00
|
|
|
static void hpsa_update_scsi_devices(struct ctlr_info *h);
|
2009-12-08 22:09:11 +00:00
|
|
|
static int check_for_unit_attention(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c);
|
|
|
|
|
static void check_ioctl_unit_attention(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c);
|
2010-02-04 14:42:40 +00:00
|
|
|
/* performant mode helper functions */
|
|
|
|
|
static void calc_bucket_map(int *bucket, int num_buckets,
|
2015-01-23 22:41:09 +00:00
|
|
|
int nsgs, int min_blocks, u32 *bucket_map);
|
2015-04-23 14:33:48 +00:00
|
|
|
static void hpsa_free_performant_mode(struct ctlr_info *h);
|
|
|
|
|
static int hpsa_put_ctlr_into_performant_mode(struct ctlr_info *h);
|
2012-05-01 16:43:06 +00:00
|
|
|
static inline u32 next_command(struct ctlr_info *h, u8 q);
|
2012-12-21 21:08:55 +00:00
|
|
|
static int hpsa_find_cfg_addrs(struct pci_dev *pdev, void __iomem *vaddr,
|
|
|
|
|
u32 *cfg_base_addr, u64 *cfg_base_addr_index,
|
|
|
|
|
u64 *cfg_offset);
|
|
|
|
|
static int hpsa_pci_find_memory_BAR(struct pci_dev *pdev,
|
|
|
|
|
unsigned long *memory_bar);
|
|
|
|
|
static int hpsa_lookup_board_id(struct pci_dev *pdev, u32 *board_id);
|
|
|
|
|
static int hpsa_wait_for_board_state(struct pci_dev *pdev, void __iomem *vaddr,
|
|
|
|
|
int wait_for_ready);
|
2012-05-01 16:42:51 +00:00
|
|
|
static inline void finish_cmd(struct CommandList *c);
|
2015-01-23 22:45:01 +00:00
|
|
|
static int hpsa_wait_for_mode_change_ack(struct ctlr_info *h);
|
2011-01-06 20:48:03 +00:00
|
|
|
#define BOARD_NOT_READY 0
|
|
|
|
|
#define BOARD_READY 1
|
2014-02-18 19:57:37 +00:00
|
|
|
static void hpsa_drain_accel_commands(struct ctlr_info *h);
|
2014-02-18 19:55:43 +00:00
|
|
|
static void hpsa_flush_cache(struct ctlr_info *h);
|
2014-02-18 19:56:34 +00:00
|
|
|
static int hpsa_scsi_ioaccel_queue_command(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c, u32 ioaccel_handle, u8 *cdb, int cdb_len,
|
2015-01-23 22:43:30 +00:00
|
|
|
u8 *scsi3addr, struct hpsa_scsi_dev_t *phys_disk);
|
2015-01-23 22:43:25 +00:00
|
|
|
static void hpsa_command_resubmit_worker(struct work_struct *work);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
static u32 lockup_detected(struct ctlr_info *h);
|
|
|
|
|
static int detect_controller_lockup(struct ctlr_info *h);
|
2015-11-04 21:52:16 +00:00
|
|
|
static void hpsa_disable_rld_caching(struct ctlr_info *h);
|
2015-11-04 21:52:34 +00:00
|
|
|
static inline int hpsa_scsi_do_report_phys_luns(struct ctlr_info *h,
|
|
|
|
|
struct ReportExtendedLUNdata *buf, int bufsize);
|
2015-11-04 21:52:09 +00:00
|
|
|
static int hpsa_luns_changed(struct ctlr_info *h);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
static inline struct ctlr_info *sdev_to_hba(struct scsi_device *sdev)
|
|
|
|
|
{
|
|
|
|
|
unsigned long *priv = shost_priv(sdev->host);
|
|
|
|
|
return (struct ctlr_info *) *priv;
|
|
|
|
|
}
|
|
|
|
|
|
2010-02-04 14:43:11 +00:00
|
|
|
static inline struct ctlr_info *shost_to_hba(struct Scsi_Host *sh)
|
|
|
|
|
{
|
|
|
|
|
unsigned long *priv = shost_priv(sh);
|
|
|
|
|
return (struct ctlr_info *) *priv;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:34:16 +00:00
|
|
|
static inline bool hpsa_is_cmd_idle(struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
return c->scsi_cmd == SCSI_CMD_IDLE;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:35:22 +00:00
|
|
|
static inline bool hpsa_is_pending_event(struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
return c->abort_pending || c->reset_pending;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:32:16 +00:00
|
|
|
/* extract sense key, asc, and ascq from sense data. -1 means invalid. */
|
|
|
|
|
static void decode_sense_data(const u8 *sense_data, int sense_data_len,
|
|
|
|
|
u8 *sense_key, u8 *asc, u8 *ascq)
|
|
|
|
|
{
|
|
|
|
|
struct scsi_sense_hdr sshdr;
|
|
|
|
|
bool rc;
|
|
|
|
|
|
|
|
|
|
*sense_key = -1;
|
|
|
|
|
*asc = -1;
|
|
|
|
|
*ascq = -1;
|
|
|
|
|
|
|
|
|
|
if (sense_data_len < 1)
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
rc = scsi_normalize_sense(sense_data, sense_data_len, &sshdr);
|
|
|
|
|
if (rc) {
|
|
|
|
|
*sense_key = sshdr.sense_key;
|
|
|
|
|
*asc = sshdr.asc;
|
|
|
|
|
*ascq = sshdr.ascq;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
static int check_for_unit_attention(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c)
|
|
|
|
|
{
|
2015-04-23 14:32:16 +00:00
|
|
|
u8 sense_key, asc, ascq;
|
|
|
|
|
int sense_len;
|
|
|
|
|
|
|
|
|
|
if (c->err_info->SenseLen > sizeof(c->err_info->SenseInfo))
|
|
|
|
|
sense_len = sizeof(c->err_info->SenseInfo);
|
|
|
|
|
else
|
|
|
|
|
sense_len = c->err_info->SenseLen;
|
|
|
|
|
|
|
|
|
|
decode_sense_data(c->err_info->SenseInfo, sense_len,
|
|
|
|
|
&sense_key, &asc, &ascq);
|
2015-07-18 16:12:28 +00:00
|
|
|
if (sense_key != UNIT_ATTENTION || asc == 0xff)
|
2009-12-08 22:09:11 +00:00
|
|
|
return 0;
|
|
|
|
|
|
2015-04-23 14:32:16 +00:00
|
|
|
switch (asc) {
|
2009-12-08 22:09:11 +00:00
|
|
|
case STATE_CHANGED:
|
2015-04-23 14:32:16 +00:00
|
|
|
dev_warn(&h->pdev->dev,
|
2015-04-23 14:35:09 +00:00
|
|
|
"%s: a state change detected, command retried\n",
|
|
|
|
|
h->devname);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case LUN_FAILED:
|
2014-11-14 23:26:48 +00:00
|
|
|
dev_warn(&h->pdev->dev,
|
2015-04-23 14:35:09 +00:00
|
|
|
"%s: LUN failure detected\n", h->devname);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case REPORT_LUNS_CHANGED:
|
2014-11-14 23:26:48 +00:00
|
|
|
dev_warn(&h->pdev->dev,
|
2015-04-23 14:35:09 +00:00
|
|
|
"%s: report LUN data changed\n", h->devname);
|
2009-12-08 22:09:11 +00:00
|
|
|
/*
|
2012-01-19 20:01:25 +00:00
|
|
|
* Note: this REPORT_LUNS_CHANGED condition only occurs on the external
|
|
|
|
|
* target (array) devices.
|
2009-12-08 22:09:11 +00:00
|
|
|
*/
|
|
|
|
|
break;
|
|
|
|
|
case POWER_OR_RESET:
|
2015-04-23 14:35:09 +00:00
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"%s: a power on or device reset detected\n",
|
|
|
|
|
h->devname);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case UNIT_ATTENTION_CLEARED:
|
2015-04-23 14:35:09 +00:00
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"%s: unit attention cleared by another initiator\n",
|
|
|
|
|
h->devname);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
default:
|
2015-04-23 14:35:09 +00:00
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"%s: unknown unit attention detected\n",
|
|
|
|
|
h->devname);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
return 1;
|
|
|
|
|
}
|
|
|
|
|
|
2012-05-01 16:42:35 +00:00
|
|
|
static int check_for_busy(struct ctlr_info *h, struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
if (c->err_info->CommandStatus != CMD_TARGET_STATUS ||
|
|
|
|
|
(c->err_info->ScsiStatus != SAM_STAT_BUSY &&
|
|
|
|
|
c->err_info->ScsiStatus != SAM_STAT_TASK_SET_FULL))
|
|
|
|
|
return 0;
|
|
|
|
|
dev_warn(&h->pdev->dev, HPSA "device busy");
|
|
|
|
|
return 1;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:32:22 +00:00
|
|
|
static u32 lockup_detected(struct ctlr_info *h);
|
|
|
|
|
static ssize_t host_show_lockup_detected(struct device *dev,
|
|
|
|
|
struct device_attribute *attr, char *buf)
|
|
|
|
|
{
|
|
|
|
|
int ld;
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct Scsi_Host *shost = class_to_shost(dev);
|
|
|
|
|
|
|
|
|
|
h = shost_to_hba(shost);
|
|
|
|
|
ld = lockup_detected(h);
|
|
|
|
|
|
|
|
|
|
return sprintf(buf, "ld=%d\n", ld);
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:57:00 +00:00
|
|
|
static ssize_t host_store_hp_ssd_smart_path_status(struct device *dev,
|
|
|
|
|
struct device_attribute *attr,
|
|
|
|
|
const char *buf, size_t count)
|
|
|
|
|
{
|
|
|
|
|
int status, len;
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct Scsi_Host *shost = class_to_shost(dev);
|
|
|
|
|
char tmpbuf[10];
|
|
|
|
|
|
|
|
|
|
if (!capable(CAP_SYS_ADMIN) || !capable(CAP_SYS_RAWIO))
|
|
|
|
|
return -EACCES;
|
|
|
|
|
len = count > sizeof(tmpbuf) - 1 ? sizeof(tmpbuf) - 1 : count;
|
|
|
|
|
strncpy(tmpbuf, buf, len);
|
|
|
|
|
tmpbuf[len] = '\0';
|
|
|
|
|
if (sscanf(tmpbuf, "%d", &status) != 1)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
h = shost_to_hba(shost);
|
|
|
|
|
h->acciopath_status = !!status;
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"hpsa: HP SSD Smart Path %s via sysfs update.\n",
|
|
|
|
|
h->acciopath_status ? "enabled" : "disabled");
|
|
|
|
|
return count;
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:57:52 +00:00
|
|
|
static ssize_t host_store_raid_offload_debug(struct device *dev,
|
|
|
|
|
struct device_attribute *attr,
|
|
|
|
|
const char *buf, size_t count)
|
|
|
|
|
{
|
|
|
|
|
int debug_level, len;
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct Scsi_Host *shost = class_to_shost(dev);
|
|
|
|
|
char tmpbuf[10];
|
|
|
|
|
|
|
|
|
|
if (!capable(CAP_SYS_ADMIN) || !capable(CAP_SYS_RAWIO))
|
|
|
|
|
return -EACCES;
|
|
|
|
|
len = count > sizeof(tmpbuf) - 1 ? sizeof(tmpbuf) - 1 : count;
|
|
|
|
|
strncpy(tmpbuf, buf, len);
|
|
|
|
|
tmpbuf[len] = '\0';
|
|
|
|
|
if (sscanf(tmpbuf, "%d", &debug_level) != 1)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
if (debug_level < 0)
|
|
|
|
|
debug_level = 0;
|
|
|
|
|
h = shost_to_hba(shost);
|
|
|
|
|
h->raid_offload_debug = debug_level;
|
|
|
|
|
dev_warn(&h->pdev->dev, "hpsa: Set raid_offload_debug level = %d\n",
|
|
|
|
|
h->raid_offload_debug);
|
|
|
|
|
return count;
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
static ssize_t host_store_rescan(struct device *dev,
|
|
|
|
|
struct device_attribute *attr,
|
|
|
|
|
const char *buf, size_t count)
|
|
|
|
|
{
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct Scsi_Host *shost = class_to_shost(dev);
|
2010-02-04 14:43:11 +00:00
|
|
|
h = shost_to_hba(shost);
|
2010-02-25 20:03:12 +00:00
|
|
|
hpsa_scan_start(h->scsi_host);
|
2009-12-08 22:09:11 +00:00
|
|
|
return count;
|
|
|
|
|
}
|
|
|
|
|
|
2010-05-27 20:14:34 +00:00
|
|
|
static ssize_t host_show_firmware_revision(struct device *dev,
|
|
|
|
|
struct device_attribute *attr, char *buf)
|
|
|
|
|
{
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct Scsi_Host *shost = class_to_shost(dev);
|
|
|
|
|
unsigned char *fwrev;
|
|
|
|
|
|
|
|
|
|
h = shost_to_hba(shost);
|
|
|
|
|
if (!h->hba_inquiry_data)
|
|
|
|
|
return 0;
|
|
|
|
|
fwrev = &h->hba_inquiry_data[32];
|
|
|
|
|
return snprintf(buf, 20, "%c%c%c%c\n",
|
|
|
|
|
fwrev[0], fwrev[1], fwrev[2], fwrev[3]);
|
|
|
|
|
}
|
|
|
|
|
|
2011-01-06 20:48:39 +00:00
|
|
|
static ssize_t host_show_commands_outstanding(struct device *dev,
|
|
|
|
|
struct device_attribute *attr, char *buf)
|
|
|
|
|
{
|
|
|
|
|
struct Scsi_Host *shost = class_to_shost(dev);
|
|
|
|
|
struct ctlr_info *h = shost_to_hba(shost);
|
|
|
|
|
|
2014-11-14 23:27:09 +00:00
|
|
|
return snprintf(buf, 20, "%d\n",
|
|
|
|
|
atomic_read(&h->commands_outstanding));
|
2011-01-06 20:48:39 +00:00
|
|
|
}
|
|
|
|
|
|
2011-02-15 21:32:58 +00:00
|
|
|
static ssize_t host_show_transport_mode(struct device *dev,
|
|
|
|
|
struct device_attribute *attr, char *buf)
|
|
|
|
|
{
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct Scsi_Host *shost = class_to_shost(dev);
|
|
|
|
|
|
|
|
|
|
h = shost_to_hba(shost);
|
|
|
|
|
return snprintf(buf, 20, "%s\n",
|
2011-02-15 21:33:03 +00:00
|
|
|
h->transMethod & CFGTBL_Trans_Performant ?
|
2011-02-15 21:32:58 +00:00
|
|
|
"performant" : "simple");
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:57:00 +00:00
|
|
|
static ssize_t host_show_hp_ssd_smart_path_status(struct device *dev,
|
|
|
|
|
struct device_attribute *attr, char *buf)
|
|
|
|
|
{
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct Scsi_Host *shost = class_to_shost(dev);
|
|
|
|
|
|
|
|
|
|
h = shost_to_hba(shost);
|
|
|
|
|
return snprintf(buf, 30, "HP SSD Smart Path %s\n",
|
|
|
|
|
(h->acciopath_status == 1) ? "enabled" : "disabled");
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-03 20:00:01 +00:00
|
|
|
/* List of controllers which cannot be hard reset on kexec with reset_devices */
|
2011-03-09 23:00:06 +00:00
|
|
|
static u32 unresettable_controller[] = {
|
|
|
|
|
0x324a103C, /* Smart Array P712m */
|
2015-04-23 14:32:06 +00:00
|
|
|
0x324b103C, /* Smart Array P711m */
|
2011-03-09 23:00:06 +00:00
|
|
|
0x3223103C, /* Smart Array P800 */
|
|
|
|
|
0x3234103C, /* Smart Array P400 */
|
|
|
|
|
0x3235103C, /* Smart Array P400i */
|
|
|
|
|
0x3211103C, /* Smart Array E200i */
|
|
|
|
|
0x3212103C, /* Smart Array E200 */
|
|
|
|
|
0x3213103C, /* Smart Array E200i */
|
|
|
|
|
0x3214103C, /* Smart Array E200i */
|
|
|
|
|
0x3215103C, /* Smart Array E200i */
|
|
|
|
|
0x3237103C, /* Smart Array E500 */
|
|
|
|
|
0x323D103C, /* Smart Array P700m */
|
2011-11-28 14:39:55 +00:00
|
|
|
0x40800E11, /* Smart Array 5i */
|
2011-03-09 23:00:06 +00:00
|
|
|
0x409C0E11, /* Smart Array 6400 */
|
|
|
|
|
0x409D0E11, /* Smart Array 6400 EM */
|
2012-02-14 17:07:59 +00:00
|
|
|
0x40700E11, /* Smart Array 5300 */
|
|
|
|
|
0x40820E11, /* Smart Array 532 */
|
|
|
|
|
0x40830E11, /* Smart Array 5312 */
|
|
|
|
|
0x409A0E11, /* Smart Array 641 */
|
|
|
|
|
0x409B0E11, /* Smart Array 642 */
|
|
|
|
|
0x40910E11, /* Smart Array 6i */
|
2011-03-09 23:00:06 +00:00
|
|
|
};
|
|
|
|
|
|
2011-05-03 20:00:01 +00:00
|
|
|
/* List of controllers which cannot even be soft reset */
|
|
|
|
|
static u32 soft_unresettable_controller[] = {
|
2011-11-28 14:39:55 +00:00
|
|
|
0x40800E11, /* Smart Array 5i */
|
2012-02-14 17:07:59 +00:00
|
|
|
0x40700E11, /* Smart Array 5300 */
|
|
|
|
|
0x40820E11, /* Smart Array 532 */
|
|
|
|
|
0x40830E11, /* Smart Array 5312 */
|
|
|
|
|
0x409A0E11, /* Smart Array 641 */
|
|
|
|
|
0x409B0E11, /* Smart Array 642 */
|
|
|
|
|
0x40910E11, /* Smart Array 6i */
|
2011-05-03 20:00:01 +00:00
|
|
|
/* Exclude 640x boards. These are two pci devices in one slot
|
|
|
|
|
* which share a battery backed cache module. One controls the
|
|
|
|
|
* cache, the other accesses the cache through the one that controls
|
|
|
|
|
* it. If we reset the one controlling the cache, the other will
|
|
|
|
|
* likely not be happy. Just forbid resetting this conjoined mess.
|
|
|
|
|
* The 640x isn't really supported by hpsa anyway.
|
|
|
|
|
*/
|
|
|
|
|
0x409C0E11, /* Smart Array 6400 */
|
|
|
|
|
0x409D0E11, /* Smart Array 6400 EM */
|
|
|
|
|
};
|
|
|
|
|
|
2015-04-23 14:32:06 +00:00
|
|
|
static u32 needs_abort_tags_swizzled[] = {
|
|
|
|
|
0x323D103C, /* Smart Array P700m */
|
|
|
|
|
0x324a103C, /* Smart Array P712m */
|
|
|
|
|
0x324b103C, /* SmartArray P711m */
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static int board_id_in_array(u32 a[], int nelems, u32 board_id)
|
2011-03-09 23:00:06 +00:00
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
2015-04-23 14:32:06 +00:00
|
|
|
for (i = 0; i < nelems; i++)
|
|
|
|
|
if (a[i] == board_id)
|
|
|
|
|
return 1;
|
|
|
|
|
return 0;
|
2011-05-03 20:00:01 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:32:06 +00:00
|
|
|
static int ctlr_is_hard_resettable(u32 board_id)
|
2011-05-03 20:00:01 +00:00
|
|
|
{
|
2015-04-23 14:32:06 +00:00
|
|
|
return !board_id_in_array(unresettable_controller,
|
|
|
|
|
ARRAY_SIZE(unresettable_controller), board_id);
|
|
|
|
|
}
|
2011-05-03 20:00:01 +00:00
|
|
|
|
2015-04-23 14:32:06 +00:00
|
|
|
static int ctlr_is_soft_resettable(u32 board_id)
|
|
|
|
|
{
|
|
|
|
|
return !board_id_in_array(soft_unresettable_controller,
|
|
|
|
|
ARRAY_SIZE(soft_unresettable_controller), board_id);
|
2011-03-09 23:00:06 +00:00
|
|
|
}
|
|
|
|
|
|
2011-05-03 20:00:01 +00:00
|
|
|
static int ctlr_is_resettable(u32 board_id)
|
|
|
|
|
{
|
|
|
|
|
return ctlr_is_hard_resettable(board_id) ||
|
|
|
|
|
ctlr_is_soft_resettable(board_id);
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:32:06 +00:00
|
|
|
static int ctlr_needs_abort_tags_swizzled(u32 board_id)
|
|
|
|
|
{
|
|
|
|
|
return board_id_in_array(needs_abort_tags_swizzled,
|
|
|
|
|
ARRAY_SIZE(needs_abort_tags_swizzled), board_id);
|
|
|
|
|
}
|
|
|
|
|
|
2011-03-09 23:00:06 +00:00
|
|
|
static ssize_t host_show_resettable(struct device *dev,
|
|
|
|
|
struct device_attribute *attr, char *buf)
|
|
|
|
|
{
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct Scsi_Host *shost = class_to_shost(dev);
|
|
|
|
|
|
|
|
|
|
h = shost_to_hba(shost);
|
2011-05-03 20:00:01 +00:00
|
|
|
return snprintf(buf, 20, "%d\n", ctlr_is_resettable(h->board_id));
|
2011-03-09 23:00:06 +00:00
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
static inline int is_logical_dev_addr_mode(unsigned char scsi3addr[])
|
|
|
|
|
{
|
|
|
|
|
return (scsi3addr[3] & 0xC0) == 0x40;
|
|
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:41:35 +00:00
|
|
|
static const char * const raid_label[] = { "0", "4", "1(+0)", "5", "5+1", "6",
|
2015-11-04 21:52:22 +00:00
|
|
|
"1(+0)ADM", "UNKNOWN", "PHYS DRV"
|
2009-12-08 22:09:11 +00:00
|
|
|
};
|
2014-02-18 19:56:55 +00:00
|
|
|
#define HPSA_RAID_0 0
|
|
|
|
|
#define HPSA_RAID_4 1
|
|
|
|
|
#define HPSA_RAID_1 2 /* also used for RAID 10 */
|
|
|
|
|
#define HPSA_RAID_5 3 /* also used for RAID 50 */
|
|
|
|
|
#define HPSA_RAID_51 4
|
|
|
|
|
#define HPSA_RAID_6 5 /* also used for RAID 60 */
|
|
|
|
|
#define HPSA_RAID_ADM 6 /* also used for RAID 1+0 ADM */
|
2015-11-04 21:52:22 +00:00
|
|
|
#define RAID_UNKNOWN (ARRAY_SIZE(raid_label) - 2)
|
|
|
|
|
#define PHYSICAL_DRIVE (ARRAY_SIZE(raid_label) - 1)
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-11-04 21:51:33 +00:00
|
|
|
static inline bool is_logical_device(struct hpsa_scsi_dev_t *device)
|
|
|
|
|
{
|
|
|
|
|
return !device->physical_device;
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
static ssize_t raid_level_show(struct device *dev,
|
|
|
|
|
struct device_attribute *attr, char *buf)
|
|
|
|
|
{
|
|
|
|
|
ssize_t l = 0;
|
2010-02-04 14:41:38 +00:00
|
|
|
unsigned char rlevel;
|
2009-12-08 22:09:11 +00:00
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct scsi_device *sdev;
|
|
|
|
|
struct hpsa_scsi_dev_t *hdev;
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
|
|
sdev = to_scsi_device(dev);
|
|
|
|
|
h = sdev_to_hba(sdev);
|
|
|
|
|
spin_lock_irqsave(&h->lock, flags);
|
|
|
|
|
hdev = sdev->hostdata;
|
|
|
|
|
if (!hdev) {
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
|
|
|
|
return -ENODEV;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Is this even a logical drive? */
|
2015-11-04 21:51:33 +00:00
|
|
|
if (!is_logical_device(hdev)) {
|
2009-12-08 22:09:11 +00:00
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
|
|
|
|
l = snprintf(buf, PAGE_SIZE, "N/A\n");
|
|
|
|
|
return l;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
rlevel = hdev->raid_level;
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
2010-02-04 14:41:38 +00:00
|
|
|
if (rlevel > RAID_UNKNOWN)
|
2009-12-08 22:09:11 +00:00
|
|
|
rlevel = RAID_UNKNOWN;
|
|
|
|
|
l = snprintf(buf, PAGE_SIZE, "RAID %s\n", raid_label[rlevel]);
|
|
|
|
|
return l;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static ssize_t lunid_show(struct device *dev,
|
|
|
|
|
struct device_attribute *attr, char *buf)
|
|
|
|
|
{
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct scsi_device *sdev;
|
|
|
|
|
struct hpsa_scsi_dev_t *hdev;
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
unsigned char lunid[8];
|
|
|
|
|
|
|
|
|
|
sdev = to_scsi_device(dev);
|
|
|
|
|
h = sdev_to_hba(sdev);
|
|
|
|
|
spin_lock_irqsave(&h->lock, flags);
|
|
|
|
|
hdev = sdev->hostdata;
|
|
|
|
|
if (!hdev) {
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
|
|
|
|
return -ENODEV;
|
|
|
|
|
}
|
|
|
|
|
memcpy(lunid, hdev->scsi3addr, sizeof(lunid));
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
|
|
|
|
return snprintf(buf, 20, "0x%02x%02x%02x%02x%02x%02x%02x%02x\n",
|
|
|
|
|
lunid[0], lunid[1], lunid[2], lunid[3],
|
|
|
|
|
lunid[4], lunid[5], lunid[6], lunid[7]);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static ssize_t unique_id_show(struct device *dev,
|
|
|
|
|
struct device_attribute *attr, char *buf)
|
|
|
|
|
{
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct scsi_device *sdev;
|
|
|
|
|
struct hpsa_scsi_dev_t *hdev;
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
unsigned char sn[16];
|
|
|
|
|
|
|
|
|
|
sdev = to_scsi_device(dev);
|
|
|
|
|
h = sdev_to_hba(sdev);
|
|
|
|
|
spin_lock_irqsave(&h->lock, flags);
|
|
|
|
|
hdev = sdev->hostdata;
|
|
|
|
|
if (!hdev) {
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
|
|
|
|
return -ENODEV;
|
|
|
|
|
}
|
|
|
|
|
memcpy(sn, hdev->device_id, sizeof(sn));
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
|
|
|
|
return snprintf(buf, 16 * 2 + 2,
|
|
|
|
|
"%02X%02X%02X%02X%02X%02X%02X%02X"
|
|
|
|
|
"%02X%02X%02X%02X%02X%02X%02X%02X\n",
|
|
|
|
|
sn[0], sn[1], sn[2], sn[3],
|
|
|
|
|
sn[4], sn[5], sn[6], sn[7],
|
|
|
|
|
sn[8], sn[9], sn[10], sn[11],
|
|
|
|
|
sn[12], sn[13], sn[14], sn[15]);
|
|
|
|
|
}
|
|
|
|
|
|
2016-04-27 22:13:33 +00:00
|
|
|
static ssize_t sas_address_show(struct device *dev,
|
|
|
|
|
struct device_attribute *attr, char *buf)
|
|
|
|
|
{
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct scsi_device *sdev;
|
|
|
|
|
struct hpsa_scsi_dev_t *hdev;
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
u64 sas_address;
|
|
|
|
|
|
|
|
|
|
sdev = to_scsi_device(dev);
|
|
|
|
|
h = sdev_to_hba(sdev);
|
|
|
|
|
spin_lock_irqsave(&h->lock, flags);
|
|
|
|
|
hdev = sdev->hostdata;
|
|
|
|
|
if (!hdev || is_logical_device(hdev) || !hdev->expose_device) {
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
|
|
|
|
return -ENODEV;
|
|
|
|
|
}
|
|
|
|
|
sas_address = hdev->sas_address;
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
|
|
|
|
|
|
|
|
|
return snprintf(buf, PAGE_SIZE, "0x%016llx\n", sas_address);
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:55:54 +00:00
|
|
|
static ssize_t host_show_hp_ssd_smart_path_enabled(struct device *dev,
|
|
|
|
|
struct device_attribute *attr, char *buf)
|
|
|
|
|
{
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct scsi_device *sdev;
|
|
|
|
|
struct hpsa_scsi_dev_t *hdev;
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
int offload_enabled;
|
|
|
|
|
|
|
|
|
|
sdev = to_scsi_device(dev);
|
|
|
|
|
h = sdev_to_hba(sdev);
|
|
|
|
|
spin_lock_irqsave(&h->lock, flags);
|
|
|
|
|
hdev = sdev->hostdata;
|
|
|
|
|
if (!hdev) {
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
|
|
|
|
return -ENODEV;
|
|
|
|
|
}
|
|
|
|
|
offload_enabled = hdev->offload_enabled;
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
|
|
|
|
return snprintf(buf, 20, "%d\n", offload_enabled);
|
|
|
|
|
}
|
|
|
|
|
|
2015-07-18 16:12:43 +00:00
|
|
|
#define MAX_PATHS 8
|
|
|
|
|
static ssize_t path_info_show(struct device *dev,
|
|
|
|
|
struct device_attribute *attr, char *buf)
|
|
|
|
|
{
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct scsi_device *sdev;
|
|
|
|
|
struct hpsa_scsi_dev_t *hdev;
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
int i;
|
|
|
|
|
int output_len = 0;
|
|
|
|
|
u8 box;
|
|
|
|
|
u8 bay;
|
|
|
|
|
u8 path_map_index = 0;
|
|
|
|
|
char *active;
|
|
|
|
|
unsigned char phys_connector[2];
|
|
|
|
|
|
|
|
|
|
sdev = to_scsi_device(dev);
|
|
|
|
|
h = sdev_to_hba(sdev);
|
|
|
|
|
spin_lock_irqsave(&h->devlock, flags);
|
|
|
|
|
hdev = sdev->hostdata;
|
|
|
|
|
if (!hdev) {
|
|
|
|
|
spin_unlock_irqrestore(&h->devlock, flags);
|
|
|
|
|
return -ENODEV;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
bay = hdev->bay;
|
|
|
|
|
for (i = 0; i < MAX_PATHS; i++) {
|
|
|
|
|
path_map_index = 1<<i;
|
|
|
|
|
if (i == hdev->active_path_index)
|
|
|
|
|
active = "Active";
|
|
|
|
|
else if (hdev->path_map & path_map_index)
|
|
|
|
|
active = "Inactive";
|
|
|
|
|
else
|
|
|
|
|
continue;
|
|
|
|
|
|
hpsa: fix multiple issues in path_info_show
path_info_show() seems to be broken in multiple ways.
First, there's
817 return snprintf(buf, output_len+1, "%s%s%s%s%s%s%s%s",
818 path[0], path[1], path[2], path[3],
819 path[4], path[5], path[6], path[7]);
so hopefully output_len contains the combined length of the eight
strings. Otherwise, snprintf will stop copying to the output
buffer, but still end up reporting that combined length - which
in turn would result in user-space getting a bunch of useless nul
bytes (thankfully the upper sysfs layer seems to clear the output
buffer before passing it to the various ->show routines). But we have
767 output_len = snprintf(path[i],
768 PATH_STRING_LEN, "[%d:%d:%d:%d] %20.20s ",
769 h->scsi_host->host_no,
770 hdev->bus, hdev->target, hdev->lun,
771 scsi_device_type(hdev->devtype));
so output_len at best contains the length of the last string printed.
Inside the loop, we then otherwise add to output_len. By magic,
we still have PATH_STRING_LEN available every time... This
wouldn't really be a problem if the bean-counting has been done
properly and each line actually does fit in 50 bytes, and maybe
it does, but I don't immediately see why. Suppose we end up
taking this branch:
802 output_len += snprintf(path[i] + output_len,
803 PATH_STRING_LEN,
804 "BOX: %hhu BAY: %hhu %s\n",
805 box, bay, active);
An optimistic estimate says this uses strlen("BOX: 1 BAY: 2
Active\n") which is 21. Now add the 20 bytes guaranteed by the
%20.20s and then some for the rest of that format string, and
we're easily over 50 bytes. I don't think we can get over 100
bytes even being pessimistic, so this just means we'll scribble
into the next path[i+1] and maybe get that overwritten later,
leading to some garbled output (in fact, since we'd overwrite the
previous string's 0-terminator, we could end up with one very
long string and then print various suffixes of that, leading to
much more than 400 bytes of output). Except of course when we're
filling path[7], where overrunning it means writing random stuff
to the kernel stack, which is usually a lot of fun.
We can fix all of that and get rid of the 400 byte stack buffer by
simply writing directly to the given output buffer, which the upper
layer guarantees is at least PAGE_SIZE. s[c]nprintf doesn't care where
it is writing to, so this doesn't make the spin lock hold time any
longer. Using scnprintf ensures that output_len always represents the
number of bytes actually written to the buffer, so we'll report the
proper amount to the upper layer.
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2015-11-04 21:52:28 +00:00
|
|
|
output_len += scnprintf(buf + output_len,
|
|
|
|
|
PAGE_SIZE - output_len,
|
|
|
|
|
"[%d:%d:%d:%d] %20.20s ",
|
2015-07-18 16:12:43 +00:00
|
|
|
h->scsi_host->host_no,
|
|
|
|
|
hdev->bus, hdev->target, hdev->lun,
|
|
|
|
|
scsi_device_type(hdev->devtype));
|
|
|
|
|
|
2015-12-22 16:36:48 +00:00
|
|
|
if (hdev->devtype == TYPE_RAID || is_logical_device(hdev)) {
|
2015-12-22 16:36:36 +00:00
|
|
|
output_len += scnprintf(buf + output_len,
|
hpsa: fix multiple issues in path_info_show
path_info_show() seems to be broken in multiple ways.
First, there's
817 return snprintf(buf, output_len+1, "%s%s%s%s%s%s%s%s",
818 path[0], path[1], path[2], path[3],
819 path[4], path[5], path[6], path[7]);
so hopefully output_len contains the combined length of the eight
strings. Otherwise, snprintf will stop copying to the output
buffer, but still end up reporting that combined length - which
in turn would result in user-space getting a bunch of useless nul
bytes (thankfully the upper sysfs layer seems to clear the output
buffer before passing it to the various ->show routines). But we have
767 output_len = snprintf(path[i],
768 PATH_STRING_LEN, "[%d:%d:%d:%d] %20.20s ",
769 h->scsi_host->host_no,
770 hdev->bus, hdev->target, hdev->lun,
771 scsi_device_type(hdev->devtype));
so output_len at best contains the length of the last string printed.
Inside the loop, we then otherwise add to output_len. By magic,
we still have PATH_STRING_LEN available every time... This
wouldn't really be a problem if the bean-counting has been done
properly and each line actually does fit in 50 bytes, and maybe
it does, but I don't immediately see why. Suppose we end up
taking this branch:
802 output_len += snprintf(path[i] + output_len,
803 PATH_STRING_LEN,
804 "BOX: %hhu BAY: %hhu %s\n",
805 box, bay, active);
An optimistic estimate says this uses strlen("BOX: 1 BAY: 2
Active\n") which is 21. Now add the 20 bytes guaranteed by the
%20.20s and then some for the rest of that format string, and
we're easily over 50 bytes. I don't think we can get over 100
bytes even being pessimistic, so this just means we'll scribble
into the next path[i+1] and maybe get that overwritten later,
leading to some garbled output (in fact, since we'd overwrite the
previous string's 0-terminator, we could end up with one very
long string and then print various suffixes of that, leading to
much more than 400 bytes of output). Except of course when we're
filling path[7], where overrunning it means writing random stuff
to the kernel stack, which is usually a lot of fun.
We can fix all of that and get rid of the 400 byte stack buffer by
simply writing directly to the given output buffer, which the upper
layer guarantees is at least PAGE_SIZE. s[c]nprintf doesn't care where
it is writing to, so this doesn't make the spin lock hold time any
longer. Using scnprintf ensures that output_len always represents the
number of bytes actually written to the buffer, so we'll report the
proper amount to the upper layer.
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2015-11-04 21:52:28 +00:00
|
|
|
PAGE_SIZE - output_len,
|
|
|
|
|
"%s\n", active);
|
2015-07-18 16:12:43 +00:00
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
box = hdev->box[i];
|
|
|
|
|
memcpy(&phys_connector, &hdev->phys_connector[i],
|
|
|
|
|
sizeof(phys_connector));
|
|
|
|
|
if (phys_connector[0] < '0')
|
|
|
|
|
phys_connector[0] = '0';
|
|
|
|
|
if (phys_connector[1] < '0')
|
|
|
|
|
phys_connector[1] = '0';
|
2015-12-22 16:36:48 +00:00
|
|
|
output_len += scnprintf(buf + output_len,
|
hpsa: fix multiple issues in path_info_show
path_info_show() seems to be broken in multiple ways.
First, there's
817 return snprintf(buf, output_len+1, "%s%s%s%s%s%s%s%s",
818 path[0], path[1], path[2], path[3],
819 path[4], path[5], path[6], path[7]);
so hopefully output_len contains the combined length of the eight
strings. Otherwise, snprintf will stop copying to the output
buffer, but still end up reporting that combined length - which
in turn would result in user-space getting a bunch of useless nul
bytes (thankfully the upper sysfs layer seems to clear the output
buffer before passing it to the various ->show routines). But we have
767 output_len = snprintf(path[i],
768 PATH_STRING_LEN, "[%d:%d:%d:%d] %20.20s ",
769 h->scsi_host->host_no,
770 hdev->bus, hdev->target, hdev->lun,
771 scsi_device_type(hdev->devtype));
so output_len at best contains the length of the last string printed.
Inside the loop, we then otherwise add to output_len. By magic,
we still have PATH_STRING_LEN available every time... This
wouldn't really be a problem if the bean-counting has been done
properly and each line actually does fit in 50 bytes, and maybe
it does, but I don't immediately see why. Suppose we end up
taking this branch:
802 output_len += snprintf(path[i] + output_len,
803 PATH_STRING_LEN,
804 "BOX: %hhu BAY: %hhu %s\n",
805 box, bay, active);
An optimistic estimate says this uses strlen("BOX: 1 BAY: 2
Active\n") which is 21. Now add the 20 bytes guaranteed by the
%20.20s and then some for the rest of that format string, and
we're easily over 50 bytes. I don't think we can get over 100
bytes even being pessimistic, so this just means we'll scribble
into the next path[i+1] and maybe get that overwritten later,
leading to some garbled output (in fact, since we'd overwrite the
previous string's 0-terminator, we could end up with one very
long string and then print various suffixes of that, leading to
much more than 400 bytes of output). Except of course when we're
filling path[7], where overrunning it means writing random stuff
to the kernel stack, which is usually a lot of fun.
We can fix all of that and get rid of the 400 byte stack buffer by
simply writing directly to the given output buffer, which the upper
layer guarantees is at least PAGE_SIZE. s[c]nprintf doesn't care where
it is writing to, so this doesn't make the spin lock hold time any
longer. Using scnprintf ensures that output_len always represents the
number of bytes actually written to the buffer, so we'll report the
proper amount to the upper layer.
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2015-11-04 21:52:28 +00:00
|
|
|
PAGE_SIZE - output_len,
|
2015-07-18 16:12:43 +00:00
|
|
|
"PORT: %.2s ",
|
|
|
|
|
phys_connector);
|
2016-02-23 21:16:15 +00:00
|
|
|
if ((hdev->devtype == TYPE_DISK || hdev->devtype == TYPE_ZBC) &&
|
|
|
|
|
hdev->expose_device) {
|
2015-07-18 16:12:43 +00:00
|
|
|
if (box == 0 || box == 0xFF) {
|
2015-12-22 16:36:36 +00:00
|
|
|
output_len += scnprintf(buf + output_len,
|
hpsa: fix multiple issues in path_info_show
path_info_show() seems to be broken in multiple ways.
First, there's
817 return snprintf(buf, output_len+1, "%s%s%s%s%s%s%s%s",
818 path[0], path[1], path[2], path[3],
819 path[4], path[5], path[6], path[7]);
so hopefully output_len contains the combined length of the eight
strings. Otherwise, snprintf will stop copying to the output
buffer, but still end up reporting that combined length - which
in turn would result in user-space getting a bunch of useless nul
bytes (thankfully the upper sysfs layer seems to clear the output
buffer before passing it to the various ->show routines). But we have
767 output_len = snprintf(path[i],
768 PATH_STRING_LEN, "[%d:%d:%d:%d] %20.20s ",
769 h->scsi_host->host_no,
770 hdev->bus, hdev->target, hdev->lun,
771 scsi_device_type(hdev->devtype));
so output_len at best contains the length of the last string printed.
Inside the loop, we then otherwise add to output_len. By magic,
we still have PATH_STRING_LEN available every time... This
wouldn't really be a problem if the bean-counting has been done
properly and each line actually does fit in 50 bytes, and maybe
it does, but I don't immediately see why. Suppose we end up
taking this branch:
802 output_len += snprintf(path[i] + output_len,
803 PATH_STRING_LEN,
804 "BOX: %hhu BAY: %hhu %s\n",
805 box, bay, active);
An optimistic estimate says this uses strlen("BOX: 1 BAY: 2
Active\n") which is 21. Now add the 20 bytes guaranteed by the
%20.20s and then some for the rest of that format string, and
we're easily over 50 bytes. I don't think we can get over 100
bytes even being pessimistic, so this just means we'll scribble
into the next path[i+1] and maybe get that overwritten later,
leading to some garbled output (in fact, since we'd overwrite the
previous string's 0-terminator, we could end up with one very
long string and then print various suffixes of that, leading to
much more than 400 bytes of output). Except of course when we're
filling path[7], where overrunning it means writing random stuff
to the kernel stack, which is usually a lot of fun.
We can fix all of that and get rid of the 400 byte stack buffer by
simply writing directly to the given output buffer, which the upper
layer guarantees is at least PAGE_SIZE. s[c]nprintf doesn't care where
it is writing to, so this doesn't make the spin lock hold time any
longer. Using scnprintf ensures that output_len always represents the
number of bytes actually written to the buffer, so we'll report the
proper amount to the upper layer.
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2015-11-04 21:52:28 +00:00
|
|
|
PAGE_SIZE - output_len,
|
2015-07-18 16:12:43 +00:00
|
|
|
"BAY: %hhu %s\n",
|
|
|
|
|
bay, active);
|
|
|
|
|
} else {
|
2015-12-22 16:36:36 +00:00
|
|
|
output_len += scnprintf(buf + output_len,
|
hpsa: fix multiple issues in path_info_show
path_info_show() seems to be broken in multiple ways.
First, there's
817 return snprintf(buf, output_len+1, "%s%s%s%s%s%s%s%s",
818 path[0], path[1], path[2], path[3],
819 path[4], path[5], path[6], path[7]);
so hopefully output_len contains the combined length of the eight
strings. Otherwise, snprintf will stop copying to the output
buffer, but still end up reporting that combined length - which
in turn would result in user-space getting a bunch of useless nul
bytes (thankfully the upper sysfs layer seems to clear the output
buffer before passing it to the various ->show routines). But we have
767 output_len = snprintf(path[i],
768 PATH_STRING_LEN, "[%d:%d:%d:%d] %20.20s ",
769 h->scsi_host->host_no,
770 hdev->bus, hdev->target, hdev->lun,
771 scsi_device_type(hdev->devtype));
so output_len at best contains the length of the last string printed.
Inside the loop, we then otherwise add to output_len. By magic,
we still have PATH_STRING_LEN available every time... This
wouldn't really be a problem if the bean-counting has been done
properly and each line actually does fit in 50 bytes, and maybe
it does, but I don't immediately see why. Suppose we end up
taking this branch:
802 output_len += snprintf(path[i] + output_len,
803 PATH_STRING_LEN,
804 "BOX: %hhu BAY: %hhu %s\n",
805 box, bay, active);
An optimistic estimate says this uses strlen("BOX: 1 BAY: 2
Active\n") which is 21. Now add the 20 bytes guaranteed by the
%20.20s and then some for the rest of that format string, and
we're easily over 50 bytes. I don't think we can get over 100
bytes even being pessimistic, so this just means we'll scribble
into the next path[i+1] and maybe get that overwritten later,
leading to some garbled output (in fact, since we'd overwrite the
previous string's 0-terminator, we could end up with one very
long string and then print various suffixes of that, leading to
much more than 400 bytes of output). Except of course when we're
filling path[7], where overrunning it means writing random stuff
to the kernel stack, which is usually a lot of fun.
We can fix all of that and get rid of the 400 byte stack buffer by
simply writing directly to the given output buffer, which the upper
layer guarantees is at least PAGE_SIZE. s[c]nprintf doesn't care where
it is writing to, so this doesn't make the spin lock hold time any
longer. Using scnprintf ensures that output_len always represents the
number of bytes actually written to the buffer, so we'll report the
proper amount to the upper layer.
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2015-11-04 21:52:28 +00:00
|
|
|
PAGE_SIZE - output_len,
|
2015-07-18 16:12:43 +00:00
|
|
|
"BOX: %hhu BAY: %hhu %s\n",
|
|
|
|
|
box, bay, active);
|
|
|
|
|
}
|
|
|
|
|
} else if (box != 0 && box != 0xFF) {
|
2015-12-22 16:36:36 +00:00
|
|
|
output_len += scnprintf(buf + output_len,
|
hpsa: fix multiple issues in path_info_show
path_info_show() seems to be broken in multiple ways.
First, there's
817 return snprintf(buf, output_len+1, "%s%s%s%s%s%s%s%s",
818 path[0], path[1], path[2], path[3],
819 path[4], path[5], path[6], path[7]);
so hopefully output_len contains the combined length of the eight
strings. Otherwise, snprintf will stop copying to the output
buffer, but still end up reporting that combined length - which
in turn would result in user-space getting a bunch of useless nul
bytes (thankfully the upper sysfs layer seems to clear the output
buffer before passing it to the various ->show routines). But we have
767 output_len = snprintf(path[i],
768 PATH_STRING_LEN, "[%d:%d:%d:%d] %20.20s ",
769 h->scsi_host->host_no,
770 hdev->bus, hdev->target, hdev->lun,
771 scsi_device_type(hdev->devtype));
so output_len at best contains the length of the last string printed.
Inside the loop, we then otherwise add to output_len. By magic,
we still have PATH_STRING_LEN available every time... This
wouldn't really be a problem if the bean-counting has been done
properly and each line actually does fit in 50 bytes, and maybe
it does, but I don't immediately see why. Suppose we end up
taking this branch:
802 output_len += snprintf(path[i] + output_len,
803 PATH_STRING_LEN,
804 "BOX: %hhu BAY: %hhu %s\n",
805 box, bay, active);
An optimistic estimate says this uses strlen("BOX: 1 BAY: 2
Active\n") which is 21. Now add the 20 bytes guaranteed by the
%20.20s and then some for the rest of that format string, and
we're easily over 50 bytes. I don't think we can get over 100
bytes even being pessimistic, so this just means we'll scribble
into the next path[i+1] and maybe get that overwritten later,
leading to some garbled output (in fact, since we'd overwrite the
previous string's 0-terminator, we could end up with one very
long string and then print various suffixes of that, leading to
much more than 400 bytes of output). Except of course when we're
filling path[7], where overrunning it means writing random stuff
to the kernel stack, which is usually a lot of fun.
We can fix all of that and get rid of the 400 byte stack buffer by
simply writing directly to the given output buffer, which the upper
layer guarantees is at least PAGE_SIZE. s[c]nprintf doesn't care where
it is writing to, so this doesn't make the spin lock hold time any
longer. Using scnprintf ensures that output_len always represents the
number of bytes actually written to the buffer, so we'll report the
proper amount to the upper layer.
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2015-11-04 21:52:28 +00:00
|
|
|
PAGE_SIZE - output_len, "BOX: %hhu %s\n",
|
2015-07-18 16:12:43 +00:00
|
|
|
box, active);
|
|
|
|
|
} else
|
2015-12-22 16:36:36 +00:00
|
|
|
output_len += scnprintf(buf + output_len,
|
hpsa: fix multiple issues in path_info_show
path_info_show() seems to be broken in multiple ways.
First, there's
817 return snprintf(buf, output_len+1, "%s%s%s%s%s%s%s%s",
818 path[0], path[1], path[2], path[3],
819 path[4], path[5], path[6], path[7]);
so hopefully output_len contains the combined length of the eight
strings. Otherwise, snprintf will stop copying to the output
buffer, but still end up reporting that combined length - which
in turn would result in user-space getting a bunch of useless nul
bytes (thankfully the upper sysfs layer seems to clear the output
buffer before passing it to the various ->show routines). But we have
767 output_len = snprintf(path[i],
768 PATH_STRING_LEN, "[%d:%d:%d:%d] %20.20s ",
769 h->scsi_host->host_no,
770 hdev->bus, hdev->target, hdev->lun,
771 scsi_device_type(hdev->devtype));
so output_len at best contains the length of the last string printed.
Inside the loop, we then otherwise add to output_len. By magic,
we still have PATH_STRING_LEN available every time... This
wouldn't really be a problem if the bean-counting has been done
properly and each line actually does fit in 50 bytes, and maybe
it does, but I don't immediately see why. Suppose we end up
taking this branch:
802 output_len += snprintf(path[i] + output_len,
803 PATH_STRING_LEN,
804 "BOX: %hhu BAY: %hhu %s\n",
805 box, bay, active);
An optimistic estimate says this uses strlen("BOX: 1 BAY: 2
Active\n") which is 21. Now add the 20 bytes guaranteed by the
%20.20s and then some for the rest of that format string, and
we're easily over 50 bytes. I don't think we can get over 100
bytes even being pessimistic, so this just means we'll scribble
into the next path[i+1] and maybe get that overwritten later,
leading to some garbled output (in fact, since we'd overwrite the
previous string's 0-terminator, we could end up with one very
long string and then print various suffixes of that, leading to
much more than 400 bytes of output). Except of course when we're
filling path[7], where overrunning it means writing random stuff
to the kernel stack, which is usually a lot of fun.
We can fix all of that and get rid of the 400 byte stack buffer by
simply writing directly to the given output buffer, which the upper
layer guarantees is at least PAGE_SIZE. s[c]nprintf doesn't care where
it is writing to, so this doesn't make the spin lock hold time any
longer. Using scnprintf ensures that output_len always represents the
number of bytes actually written to the buffer, so we'll report the
proper amount to the upper layer.
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2015-11-04 21:52:28 +00:00
|
|
|
PAGE_SIZE - output_len, "%s\n", active);
|
2015-07-18 16:12:43 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
spin_unlock_irqrestore(&h->devlock, flags);
|
hpsa: fix multiple issues in path_info_show
path_info_show() seems to be broken in multiple ways.
First, there's
817 return snprintf(buf, output_len+1, "%s%s%s%s%s%s%s%s",
818 path[0], path[1], path[2], path[3],
819 path[4], path[5], path[6], path[7]);
so hopefully output_len contains the combined length of the eight
strings. Otherwise, snprintf will stop copying to the output
buffer, but still end up reporting that combined length - which
in turn would result in user-space getting a bunch of useless nul
bytes (thankfully the upper sysfs layer seems to clear the output
buffer before passing it to the various ->show routines). But we have
767 output_len = snprintf(path[i],
768 PATH_STRING_LEN, "[%d:%d:%d:%d] %20.20s ",
769 h->scsi_host->host_no,
770 hdev->bus, hdev->target, hdev->lun,
771 scsi_device_type(hdev->devtype));
so output_len at best contains the length of the last string printed.
Inside the loop, we then otherwise add to output_len. By magic,
we still have PATH_STRING_LEN available every time... This
wouldn't really be a problem if the bean-counting has been done
properly and each line actually does fit in 50 bytes, and maybe
it does, but I don't immediately see why. Suppose we end up
taking this branch:
802 output_len += snprintf(path[i] + output_len,
803 PATH_STRING_LEN,
804 "BOX: %hhu BAY: %hhu %s\n",
805 box, bay, active);
An optimistic estimate says this uses strlen("BOX: 1 BAY: 2
Active\n") which is 21. Now add the 20 bytes guaranteed by the
%20.20s and then some for the rest of that format string, and
we're easily over 50 bytes. I don't think we can get over 100
bytes even being pessimistic, so this just means we'll scribble
into the next path[i+1] and maybe get that overwritten later,
leading to some garbled output (in fact, since we'd overwrite the
previous string's 0-terminator, we could end up with one very
long string and then print various suffixes of that, leading to
much more than 400 bytes of output). Except of course when we're
filling path[7], where overrunning it means writing random stuff
to the kernel stack, which is usually a lot of fun.
We can fix all of that and get rid of the 400 byte stack buffer by
simply writing directly to the given output buffer, which the upper
layer guarantees is at least PAGE_SIZE. s[c]nprintf doesn't care where
it is writing to, so this doesn't make the spin lock hold time any
longer. Using scnprintf ensures that output_len always represents the
number of bytes actually written to the buffer, so we'll report the
proper amount to the upper layer.
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2015-11-04 21:52:28 +00:00
|
|
|
return output_len;
|
2015-07-18 16:12:43 +00:00
|
|
|
}
|
|
|
|
|
|
2011-03-09 23:00:01 +00:00
|
|
|
static DEVICE_ATTR(raid_level, S_IRUGO, raid_level_show, NULL);
|
|
|
|
|
static DEVICE_ATTR(lunid, S_IRUGO, lunid_show, NULL);
|
|
|
|
|
static DEVICE_ATTR(unique_id, S_IRUGO, unique_id_show, NULL);
|
|
|
|
|
static DEVICE_ATTR(rescan, S_IWUSR, NULL, host_store_rescan);
|
2016-04-27 22:13:33 +00:00
|
|
|
static DEVICE_ATTR(sas_address, S_IRUGO, sas_address_show, NULL);
|
2014-02-18 19:55:54 +00:00
|
|
|
static DEVICE_ATTR(hp_ssd_smart_path_enabled, S_IRUGO,
|
|
|
|
|
host_show_hp_ssd_smart_path_enabled, NULL);
|
2015-07-18 16:12:43 +00:00
|
|
|
static DEVICE_ATTR(path_info, S_IRUGO, path_info_show, NULL);
|
2014-02-18 19:57:00 +00:00
|
|
|
static DEVICE_ATTR(hp_ssd_smart_path_status, S_IWUSR|S_IRUGO|S_IROTH,
|
|
|
|
|
host_show_hp_ssd_smart_path_status,
|
|
|
|
|
host_store_hp_ssd_smart_path_status);
|
2014-02-18 19:57:52 +00:00
|
|
|
static DEVICE_ATTR(raid_offload_debug, S_IWUSR, NULL,
|
|
|
|
|
host_store_raid_offload_debug);
|
2011-03-09 23:00:01 +00:00
|
|
|
static DEVICE_ATTR(firmware_revision, S_IRUGO,
|
|
|
|
|
host_show_firmware_revision, NULL);
|
|
|
|
|
static DEVICE_ATTR(commands_outstanding, S_IRUGO,
|
|
|
|
|
host_show_commands_outstanding, NULL);
|
|
|
|
|
static DEVICE_ATTR(transport_mode, S_IRUGO,
|
|
|
|
|
host_show_transport_mode, NULL);
|
2011-03-09 23:00:06 +00:00
|
|
|
static DEVICE_ATTR(resettable, S_IRUGO,
|
|
|
|
|
host_show_resettable, NULL);
|
2015-04-23 14:32:22 +00:00
|
|
|
static DEVICE_ATTR(lockup_detected, S_IRUGO,
|
|
|
|
|
host_show_lockup_detected, NULL);
|
2011-03-09 23:00:01 +00:00
|
|
|
|
|
|
|
|
static struct device_attribute *hpsa_sdev_attrs[] = {
|
|
|
|
|
&dev_attr_raid_level,
|
|
|
|
|
&dev_attr_lunid,
|
|
|
|
|
&dev_attr_unique_id,
|
2014-02-18 19:55:54 +00:00
|
|
|
&dev_attr_hp_ssd_smart_path_enabled,
|
2015-07-18 16:12:43 +00:00
|
|
|
&dev_attr_path_info,
|
2016-04-27 22:13:33 +00:00
|
|
|
&dev_attr_sas_address,
|
2011-03-09 23:00:01 +00:00
|
|
|
NULL,
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
static struct device_attribute *hpsa_shost_attrs[] = {
|
|
|
|
|
&dev_attr_rescan,
|
|
|
|
|
&dev_attr_firmware_revision,
|
|
|
|
|
&dev_attr_commands_outstanding,
|
|
|
|
|
&dev_attr_transport_mode,
|
2011-03-09 23:00:06 +00:00
|
|
|
&dev_attr_resettable,
|
2014-02-18 19:57:00 +00:00
|
|
|
&dev_attr_hp_ssd_smart_path_status,
|
2014-02-18 19:57:52 +00:00
|
|
|
&dev_attr_raid_offload_debug,
|
2015-11-06 15:24:09 +00:00
|
|
|
&dev_attr_lockup_detected,
|
2011-03-09 23:00:01 +00:00
|
|
|
NULL,
|
|
|
|
|
};
|
|
|
|
|
|
2015-04-23 14:31:47 +00:00
|
|
|
#define HPSA_NRESERVED_CMDS (HPSA_CMDS_RESERVED_FOR_ABORTS + \
|
|
|
|
|
HPSA_CMDS_RESERVED_FOR_DRIVER + HPSA_MAX_CONCURRENT_PASSTHRUS)
|
|
|
|
|
|
2011-03-09 23:00:01 +00:00
|
|
|
static struct scsi_host_template hpsa_driver_template = {
|
|
|
|
|
.module = THIS_MODULE,
|
2012-01-19 20:00:59 +00:00
|
|
|
.name = HPSA,
|
|
|
|
|
.proc_name = HPSA,
|
2011-03-09 23:00:01 +00:00
|
|
|
.queuecommand = hpsa_scsi_queue_command,
|
|
|
|
|
.scan_start = hpsa_scan_start,
|
|
|
|
|
.scan_finished = hpsa_scan_finished,
|
2015-01-23 22:41:30 +00:00
|
|
|
.change_queue_depth = hpsa_change_queue_depth,
|
2011-03-09 23:00:01 +00:00
|
|
|
.this_id = -1,
|
|
|
|
|
.use_clustering = ENABLE_CLUSTERING,
|
2012-05-01 16:42:51 +00:00
|
|
|
.eh_abort_handler = hpsa_eh_abort_handler,
|
2011-03-09 23:00:01 +00:00
|
|
|
.eh_device_reset_handler = hpsa_eh_device_reset_handler,
|
|
|
|
|
.ioctl = hpsa_ioctl,
|
|
|
|
|
.slave_alloc = hpsa_slave_alloc,
|
2015-04-23 14:31:47 +00:00
|
|
|
.slave_configure = hpsa_slave_configure,
|
2011-03-09 23:00:01 +00:00
|
|
|
.slave_destroy = hpsa_slave_destroy,
|
|
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
|
.compat_ioctl = hpsa_compat_ioctl,
|
|
|
|
|
#endif
|
|
|
|
|
.sdev_attrs = hpsa_sdev_attrs,
|
|
|
|
|
.shost_attrs = hpsa_shost_attrs,
|
2011-10-26 21:20:53 +00:00
|
|
|
.max_sectors = 8192,
|
2013-10-23 10:25:40 +00:00
|
|
|
.no_write_same = 1,
|
2011-03-09 23:00:01 +00:00
|
|
|
};
|
|
|
|
|
|
2012-05-01 16:43:06 +00:00
|
|
|
static inline u32 next_command(struct ctlr_info *h, u8 q)
|
2011-03-09 23:00:01 +00:00
|
|
|
{
|
|
|
|
|
u32 a;
|
2014-05-29 15:53:07 +00:00
|
|
|
struct reply_queue_buffer *rq = &h->reply_queue[q];
|
2011-03-09 23:00:01 +00:00
|
|
|
|
2014-02-18 19:55:17 +00:00
|
|
|
if (h->transMethod & CFGTBL_Trans_io_accel1)
|
|
|
|
|
return h->access.command_completed(h, q);
|
|
|
|
|
|
2011-03-09 23:00:01 +00:00
|
|
|
if (unlikely(!(h->transMethod & CFGTBL_Trans_Performant)))
|
2012-05-01 16:43:06 +00:00
|
|
|
return h->access.command_completed(h, q);
|
2011-03-09 23:00:01 +00:00
|
|
|
|
2012-05-01 16:43:06 +00:00
|
|
|
if ((rq->head[rq->current_entry] & 1) == rq->wraparound) {
|
|
|
|
|
a = rq->head[rq->current_entry];
|
|
|
|
|
rq->current_entry++;
|
2014-11-14 23:27:09 +00:00
|
|
|
atomic_dec(&h->commands_outstanding);
|
2011-03-09 23:00:01 +00:00
|
|
|
} else {
|
|
|
|
|
a = FIFO_EMPTY;
|
|
|
|
|
}
|
|
|
|
|
/* Check for wraparound */
|
2012-05-01 16:43:06 +00:00
|
|
|
if (rq->current_entry == h->max_commands) {
|
|
|
|
|
rq->current_entry = 0;
|
|
|
|
|
rq->wraparound ^= 1;
|
2011-03-09 23:00:01 +00:00
|
|
|
}
|
|
|
|
|
return a;
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:56:34 +00:00
|
|
|
/*
|
|
|
|
|
* There are some special bits in the bus address of the
|
|
|
|
|
* command that we have to set for the controller to know
|
|
|
|
|
* how to process the command:
|
|
|
|
|
*
|
|
|
|
|
* Normal performant mode:
|
|
|
|
|
* bit 0: 1 means performant mode, 0 means simple mode.
|
|
|
|
|
* bits 1-3 = block fetch table entry
|
|
|
|
|
* bits 4-6 = command type (== 0)
|
|
|
|
|
*
|
|
|
|
|
* ioaccel1 mode:
|
|
|
|
|
* bit 0 = "performant mode" bit.
|
|
|
|
|
* bits 1-3 = block fetch table entry
|
|
|
|
|
* bits 4-6 = command type (== 110)
|
|
|
|
|
* (command type is needed because ioaccel1 mode
|
|
|
|
|
* commands are submitted through the same register as normal
|
|
|
|
|
* mode commands, so this is how the controller knows whether
|
|
|
|
|
* the command is normal mode or ioaccel1 mode.)
|
|
|
|
|
*
|
|
|
|
|
* ioaccel2 mode:
|
|
|
|
|
* bit 0 = "performant mode" bit.
|
|
|
|
|
* bits 1-4 = block fetch table entry (note extra bit)
|
|
|
|
|
* bits 4-6 = not needed, because ioaccel2 mode has
|
|
|
|
|
* a separate special register for submitting commands.
|
|
|
|
|
*/
|
|
|
|
|
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
/*
|
|
|
|
|
* set_performant_mode: Modify the tag for cciss performant
|
2011-03-09 23:00:01 +00:00
|
|
|
* set bit 0 for pull model, bits 3-1 for block fetch
|
|
|
|
|
* register number
|
|
|
|
|
*/
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
#define DEFAULT_REPLY_QUEUE (-1)
|
|
|
|
|
static void set_performant_mode(struct ctlr_info *h, struct CommandList *c,
|
|
|
|
|
int reply_queue)
|
2011-03-09 23:00:01 +00:00
|
|
|
{
|
2012-05-01 16:43:06 +00:00
|
|
|
if (likely(h->transMethod & CFGTBL_Trans_Performant)) {
|
2011-03-09 23:00:01 +00:00
|
|
|
c->busaddr |= 1 | (h->blockFetchTable[c->Header.SGList] << 1);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
if (unlikely(!h->msix_vector))
|
|
|
|
|
return;
|
|
|
|
|
if (likely(reply_queue == DEFAULT_REPLY_QUEUE))
|
2012-05-01 16:43:06 +00:00
|
|
|
c->Header.ReplyQueue =
|
2013-07-26 14:06:18 +00:00
|
|
|
raw_smp_processor_id() % h->nreply_queues;
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
else
|
|
|
|
|
c->Header.ReplyQueue = reply_queue % h->nreply_queues;
|
2012-05-01 16:43:06 +00:00
|
|
|
}
|
2011-03-09 23:00:01 +00:00
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:56:34 +00:00
|
|
|
static void set_ioaccel1_performant_mode(struct ctlr_info *h,
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
struct CommandList *c,
|
|
|
|
|
int reply_queue)
|
2014-02-18 19:56:34 +00:00
|
|
|
{
|
|
|
|
|
struct io_accel1_cmd *cp = &h->ioaccel_cmd_pool[c->cmdindex];
|
|
|
|
|
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
/*
|
|
|
|
|
* Tell the controller to post the reply to the queue for this
|
2014-02-18 19:56:34 +00:00
|
|
|
* processor. This seems to give the best I/O throughput.
|
|
|
|
|
*/
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
if (likely(reply_queue == DEFAULT_REPLY_QUEUE))
|
|
|
|
|
cp->ReplyQueue = smp_processor_id() % h->nreply_queues;
|
|
|
|
|
else
|
|
|
|
|
cp->ReplyQueue = reply_queue % h->nreply_queues;
|
|
|
|
|
/*
|
|
|
|
|
* Set the bits in the address sent down to include:
|
2014-02-18 19:56:34 +00:00
|
|
|
* - performant mode bit (bit 0)
|
|
|
|
|
* - pull count (bits 1-3)
|
|
|
|
|
* - command type (bits 4-6)
|
|
|
|
|
*/
|
|
|
|
|
c->busaddr |= 1 | (h->ioaccel1_blockFetchTable[c->Header.SGList] << 1) |
|
|
|
|
|
IOACCEL1_BUSADDR_CMDTYPE;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:34:06 +00:00
|
|
|
static void set_ioaccel2_tmf_performant_mode(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c,
|
|
|
|
|
int reply_queue)
|
|
|
|
|
{
|
|
|
|
|
struct hpsa_tmf_struct *cp = (struct hpsa_tmf_struct *)
|
|
|
|
|
&h->ioaccel2_cmd_pool[c->cmdindex];
|
|
|
|
|
|
|
|
|
|
/* Tell the controller to post the reply to the queue for this
|
|
|
|
|
* processor. This seems to give the best I/O throughput.
|
|
|
|
|
*/
|
|
|
|
|
if (likely(reply_queue == DEFAULT_REPLY_QUEUE))
|
|
|
|
|
cp->reply_queue = smp_processor_id() % h->nreply_queues;
|
|
|
|
|
else
|
|
|
|
|
cp->reply_queue = reply_queue % h->nreply_queues;
|
|
|
|
|
/* Set the bits in the address sent down to include:
|
|
|
|
|
* - performant mode bit not used in ioaccel mode 2
|
|
|
|
|
* - pull count (bits 0-3)
|
|
|
|
|
* - command type isn't needed for ioaccel2
|
|
|
|
|
*/
|
|
|
|
|
c->busaddr |= h->ioaccel2_blockFetchTable[0];
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:56:34 +00:00
|
|
|
static void set_ioaccel2_performant_mode(struct ctlr_info *h,
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
struct CommandList *c,
|
|
|
|
|
int reply_queue)
|
2014-02-18 19:56:34 +00:00
|
|
|
{
|
|
|
|
|
struct io_accel2_cmd *cp = &h->ioaccel2_cmd_pool[c->cmdindex];
|
|
|
|
|
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
/*
|
|
|
|
|
* Tell the controller to post the reply to the queue for this
|
2014-02-18 19:56:34 +00:00
|
|
|
* processor. This seems to give the best I/O throughput.
|
|
|
|
|
*/
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
if (likely(reply_queue == DEFAULT_REPLY_QUEUE))
|
|
|
|
|
cp->reply_queue = smp_processor_id() % h->nreply_queues;
|
|
|
|
|
else
|
|
|
|
|
cp->reply_queue = reply_queue % h->nreply_queues;
|
|
|
|
|
/*
|
|
|
|
|
* Set the bits in the address sent down to include:
|
2014-02-18 19:56:34 +00:00
|
|
|
* - performant mode bit not used in ioaccel mode 2
|
|
|
|
|
* - pull count (bits 0-3)
|
|
|
|
|
* - command type isn't needed for ioaccel2
|
|
|
|
|
*/
|
|
|
|
|
c->busaddr |= (h->ioaccel2_blockFetchTable[cp->sg_count]);
|
|
|
|
|
}
|
|
|
|
|
|
2012-05-01 16:43:42 +00:00
|
|
|
static int is_firmware_flash_cmd(u8 *cdb)
|
|
|
|
|
{
|
|
|
|
|
return cdb[0] == BMIC_WRITE && cdb[6] == BMIC_FLASH_FIRMWARE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* During firmware flash, the heartbeat register may not update as frequently
|
|
|
|
|
* as it should. So we dial down lockup detection during firmware flash. and
|
|
|
|
|
* dial it back up when firmware flash completes.
|
|
|
|
|
*/
|
|
|
|
|
#define HEARTBEAT_SAMPLE_INTERVAL_DURING_FLASH (240 * HZ)
|
|
|
|
|
#define HEARTBEAT_SAMPLE_INTERVAL (30 * HZ)
|
|
|
|
|
static void dial_down_lockup_detection_during_fw_flash(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
if (!is_firmware_flash_cmd(c->Request.CDB))
|
|
|
|
|
return;
|
|
|
|
|
atomic_inc(&h->firmware_flash_in_progress);
|
|
|
|
|
h->heartbeat_sample_interval = HEARTBEAT_SAMPLE_INTERVAL_DURING_FLASH;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void dial_up_lockup_detection_on_fw_flash_complete(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
if (is_firmware_flash_cmd(c->Request.CDB) &&
|
|
|
|
|
atomic_dec_and_test(&h->firmware_flash_in_progress))
|
|
|
|
|
h->heartbeat_sample_interval = HEARTBEAT_SAMPLE_INTERVAL;
|
|
|
|
|
}
|
|
|
|
|
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
static void __enqueue_cmd_and_start_io(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c, int reply_queue)
|
2011-03-09 23:00:01 +00:00
|
|
|
{
|
2015-01-23 22:44:40 +00:00
|
|
|
dial_down_lockup_detection_during_fw_flash(h, c);
|
|
|
|
|
atomic_inc(&h->commands_outstanding);
|
2014-02-18 19:56:34 +00:00
|
|
|
switch (c->cmd_type) {
|
|
|
|
|
case CMD_IOACCEL1:
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
set_ioaccel1_performant_mode(h, c, reply_queue);
|
2015-01-23 22:44:40 +00:00
|
|
|
writel(c->busaddr, h->vaddr + SA5_REQUEST_PORT_OFFSET);
|
2014-02-18 19:56:34 +00:00
|
|
|
break;
|
|
|
|
|
case CMD_IOACCEL2:
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
set_ioaccel2_performant_mode(h, c, reply_queue);
|
2015-01-23 22:44:40 +00:00
|
|
|
writel(c->busaddr, h->vaddr + IOACCEL2_INBOUND_POSTQ_32);
|
2014-02-18 19:56:34 +00:00
|
|
|
break;
|
2015-04-23 14:34:06 +00:00
|
|
|
case IOACCEL2_TMF:
|
|
|
|
|
set_ioaccel2_tmf_performant_mode(h, c, reply_queue);
|
|
|
|
|
writel(c->busaddr, h->vaddr + IOACCEL2_INBOUND_POSTQ_32);
|
|
|
|
|
break;
|
2014-02-18 19:56:34 +00:00
|
|
|
default:
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
set_performant_mode(h, c, reply_queue);
|
2015-01-23 22:44:40 +00:00
|
|
|
h->access.submit_command(h, c);
|
2014-02-18 19:56:34 +00:00
|
|
|
}
|
2011-03-09 23:00:01 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:34:16 +00:00
|
|
|
static void enqueue_cmd_and_start_io(struct ctlr_info *h, struct CommandList *c)
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
{
|
2015-04-23 14:35:22 +00:00
|
|
|
if (unlikely(hpsa_is_pending_event(c)))
|
2015-04-23 14:34:16 +00:00
|
|
|
return finish_cmd(c);
|
|
|
|
|
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
__enqueue_cmd_and_start_io(h, c, DEFAULT_REPLY_QUEUE);
|
|
|
|
|
}
|
|
|
|
|
|
2011-03-09 23:00:01 +00:00
|
|
|
static inline int is_hba_lunid(unsigned char scsi3addr[])
|
|
|
|
|
{
|
|
|
|
|
return memcmp(scsi3addr, RAID_CTLR_LUNID, 8) == 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline int is_scsi_rev_5(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
if (!h->hba_inquiry_data)
|
|
|
|
|
return 0;
|
|
|
|
|
if ((h->hba_inquiry_data[2] & 0x07) == 5)
|
|
|
|
|
return 1;
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
static int hpsa_find_target_lun(struct ctlr_info *h,
|
|
|
|
|
unsigned char scsi3addr[], int bus, int *target, int *lun)
|
|
|
|
|
{
|
|
|
|
|
/* finds an unused bus, target, lun for a new physical device
|
|
|
|
|
* assumes h->devlock is held
|
|
|
|
|
*/
|
|
|
|
|
int i, found = 0;
|
2011-10-26 21:21:07 +00:00
|
|
|
DECLARE_BITMAP(lun_taken, HPSA_MAX_DEVICES);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2012-01-20 15:15:27 +00:00
|
|
|
bitmap_zero(lun_taken, HPSA_MAX_DEVICES);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
for (i = 0; i < h->ndevices; i++) {
|
|
|
|
|
if (h->dev[i]->bus == bus && h->dev[i]->target != -1)
|
2012-01-20 15:15:27 +00:00
|
|
|
__set_bit(h->dev[i]->target, lun_taken);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
2012-01-20 15:15:27 +00:00
|
|
|
i = find_first_zero_bit(lun_taken, HPSA_MAX_DEVICES);
|
|
|
|
|
if (i < HPSA_MAX_DEVICES) {
|
|
|
|
|
/* *bus = 1; */
|
|
|
|
|
*target = i;
|
|
|
|
|
*lun = 0;
|
|
|
|
|
found = 1;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
return !found;
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-04 21:50:31 +00:00
|
|
|
static void hpsa_show_dev_msg(const char *level, struct ctlr_info *h,
|
2015-04-23 14:31:55 +00:00
|
|
|
struct hpsa_scsi_dev_t *dev, char *description)
|
|
|
|
|
{
|
2015-11-04 21:52:22 +00:00
|
|
|
#define LABEL_SIZE 25
|
|
|
|
|
char label[LABEL_SIZE];
|
|
|
|
|
|
2015-11-04 21:50:25 +00:00
|
|
|
if (h == NULL || h->pdev == NULL || h->scsi_host == NULL)
|
|
|
|
|
return;
|
|
|
|
|
|
2015-11-04 21:52:22 +00:00
|
|
|
switch (dev->devtype) {
|
|
|
|
|
case TYPE_RAID:
|
|
|
|
|
snprintf(label, LABEL_SIZE, "controller");
|
|
|
|
|
break;
|
|
|
|
|
case TYPE_ENCLOSURE:
|
|
|
|
|
snprintf(label, LABEL_SIZE, "enclosure");
|
|
|
|
|
break;
|
|
|
|
|
case TYPE_DISK:
|
2016-02-23 21:16:15 +00:00
|
|
|
case TYPE_ZBC:
|
2015-11-04 21:52:22 +00:00
|
|
|
if (dev->external)
|
|
|
|
|
snprintf(label, LABEL_SIZE, "external");
|
|
|
|
|
else if (!is_logical_dev_addr_mode(dev->scsi3addr))
|
|
|
|
|
snprintf(label, LABEL_SIZE, "%s",
|
|
|
|
|
raid_label[PHYSICAL_DRIVE]);
|
|
|
|
|
else
|
|
|
|
|
snprintf(label, LABEL_SIZE, "RAID-%s",
|
|
|
|
|
dev->raid_level > RAID_UNKNOWN ? "?" :
|
|
|
|
|
raid_label[dev->raid_level]);
|
|
|
|
|
break;
|
|
|
|
|
case TYPE_ROM:
|
|
|
|
|
snprintf(label, LABEL_SIZE, "rom");
|
|
|
|
|
break;
|
|
|
|
|
case TYPE_TAPE:
|
|
|
|
|
snprintf(label, LABEL_SIZE, "tape");
|
|
|
|
|
break;
|
|
|
|
|
case TYPE_MEDIUM_CHANGER:
|
|
|
|
|
snprintf(label, LABEL_SIZE, "changer");
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
snprintf(label, LABEL_SIZE, "UNKNOWN");
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:31:55 +00:00
|
|
|
dev_printk(level, &h->pdev->dev,
|
2015-11-04 21:52:22 +00:00
|
|
|
"scsi %d:%d:%d:%d: %s %s %.8s %.16s %s SSDSmartPathCap%c En%c Exp=%d\n",
|
2015-04-23 14:31:55 +00:00
|
|
|
h->scsi_host->host_no, dev->bus, dev->target, dev->lun,
|
|
|
|
|
description,
|
|
|
|
|
scsi_device_type(dev->devtype),
|
|
|
|
|
dev->vendor,
|
|
|
|
|
dev->model,
|
2015-11-04 21:52:22 +00:00
|
|
|
label,
|
2015-04-23 14:31:55 +00:00
|
|
|
dev->offload_config ? '+' : '-',
|
|
|
|
|
dev->offload_enabled ? '+' : '-',
|
2015-11-04 21:51:21 +00:00
|
|
|
dev->expose_device);
|
2015-04-23 14:31:55 +00:00
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/* Add an entry into h->dev[] array. */
|
2015-11-04 21:50:01 +00:00
|
|
|
static int hpsa_scsi_add_entry(struct ctlr_info *h,
|
2009-12-08 22:09:11 +00:00
|
|
|
struct hpsa_scsi_dev_t *device,
|
|
|
|
|
struct hpsa_scsi_dev_t *added[], int *nadded)
|
|
|
|
|
{
|
|
|
|
|
/* assumes h->devlock is held */
|
|
|
|
|
int n = h->ndevices;
|
|
|
|
|
int i;
|
|
|
|
|
unsigned char addr1[8], addr2[8];
|
|
|
|
|
struct hpsa_scsi_dev_t *sd;
|
|
|
|
|
|
2011-10-26 21:21:07 +00:00
|
|
|
if (n >= HPSA_MAX_DEVICES) {
|
2009-12-08 22:09:11 +00:00
|
|
|
dev_err(&h->pdev->dev, "too many devices, some will be "
|
|
|
|
|
"inaccessible.\n");
|
|
|
|
|
return -1;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* physical devices do not have lun or target assigned until now. */
|
|
|
|
|
if (device->lun != -1)
|
|
|
|
|
/* Logical device, lun is already assigned. */
|
|
|
|
|
goto lun_assigned;
|
|
|
|
|
|
|
|
|
|
/* If this device a non-zero lun of a multi-lun device
|
|
|
|
|
* byte 4 of the 8-byte LUN addr will contain the logical
|
2015-01-23 22:41:09 +00:00
|
|
|
* unit no, zero otherwise.
|
2009-12-08 22:09:11 +00:00
|
|
|
*/
|
|
|
|
|
if (device->scsi3addr[4] == 0) {
|
|
|
|
|
/* This is not a non-zero lun of a multi-lun device */
|
|
|
|
|
if (hpsa_find_target_lun(h, device->scsi3addr,
|
|
|
|
|
device->bus, &device->target, &device->lun) != 0)
|
|
|
|
|
return -1;
|
|
|
|
|
goto lun_assigned;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* This is a non-zero lun of a multi-lun device.
|
|
|
|
|
* Search through our list and find the device which
|
2015-07-18 16:13:09 +00:00
|
|
|
* has the same 8 byte LUN address, excepting byte 4 and 5.
|
2009-12-08 22:09:11 +00:00
|
|
|
* Assign the same bus and target for this new LUN.
|
|
|
|
|
* Use the logical unit number from the firmware.
|
|
|
|
|
*/
|
|
|
|
|
memcpy(addr1, device->scsi3addr, 8);
|
|
|
|
|
addr1[4] = 0;
|
2015-07-18 16:13:09 +00:00
|
|
|
addr1[5] = 0;
|
2009-12-08 22:09:11 +00:00
|
|
|
for (i = 0; i < n; i++) {
|
|
|
|
|
sd = h->dev[i];
|
|
|
|
|
memcpy(addr2, sd->scsi3addr, 8);
|
|
|
|
|
addr2[4] = 0;
|
2015-07-18 16:13:09 +00:00
|
|
|
addr2[5] = 0;
|
|
|
|
|
/* differ only in byte 4 and 5? */
|
2009-12-08 22:09:11 +00:00
|
|
|
if (memcmp(addr1, addr2, 8) == 0) {
|
|
|
|
|
device->bus = sd->bus;
|
|
|
|
|
device->target = sd->target;
|
|
|
|
|
device->lun = device->scsi3addr[4];
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
if (device->lun == -1) {
|
|
|
|
|
dev_warn(&h->pdev->dev, "physical device with no LUN=0,"
|
|
|
|
|
" suspect firmware bug or unsupported hardware "
|
|
|
|
|
"configuration.\n");
|
|
|
|
|
return -1;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
lun_assigned:
|
|
|
|
|
|
|
|
|
|
h->dev[n] = device;
|
|
|
|
|
h->ndevices++;
|
|
|
|
|
added[*nadded] = device;
|
|
|
|
|
(*nadded)++;
|
2015-04-23 14:31:55 +00:00
|
|
|
hpsa_show_dev_msg(KERN_INFO, h, device,
|
2015-11-04 21:51:21 +00:00
|
|
|
device->expose_device ? "added" : "masked");
|
2015-04-23 14:32:54 +00:00
|
|
|
device->offload_to_be_enabled = device->offload_enabled;
|
|
|
|
|
device->offload_enabled = 0;
|
2009-12-08 22:09:11 +00:00
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2012-01-19 20:01:30 +00:00
|
|
|
/* Update an entry in h->dev[] array. */
|
2015-11-04 21:50:01 +00:00
|
|
|
static void hpsa_scsi_update_entry(struct ctlr_info *h,
|
2012-01-19 20:01:30 +00:00
|
|
|
int entry, struct hpsa_scsi_dev_t *new_entry)
|
|
|
|
|
{
|
2015-04-23 14:32:54 +00:00
|
|
|
int offload_enabled;
|
2012-01-19 20:01:30 +00:00
|
|
|
/* assumes h->devlock is held */
|
|
|
|
|
BUG_ON(entry < 0 || entry >= HPSA_MAX_DEVICES);
|
|
|
|
|
|
|
|
|
|
/* Raid level changed. */
|
|
|
|
|
h->dev[entry]->raid_level = new_entry->raid_level;
|
2014-02-18 19:55:38 +00:00
|
|
|
|
2015-01-23 22:43:30 +00:00
|
|
|
/* Raid offload parameters changed. Careful about the ordering. */
|
|
|
|
|
if (new_entry->offload_config && new_entry->offload_enabled) {
|
|
|
|
|
/*
|
|
|
|
|
* if drive is newly offload_enabled, we want to copy the
|
|
|
|
|
* raid map data first. If previously offload_enabled and
|
|
|
|
|
* offload_config were set, raid map data had better be
|
|
|
|
|
* the same as it was before. if raid map data is changed
|
|
|
|
|
* then it had better be the case that
|
|
|
|
|
* h->dev[entry]->offload_enabled is currently 0.
|
|
|
|
|
*/
|
|
|
|
|
h->dev[entry]->raid_map = new_entry->raid_map;
|
|
|
|
|
h->dev[entry]->ioaccel_handle = new_entry->ioaccel_handle;
|
|
|
|
|
}
|
2015-04-23 14:32:59 +00:00
|
|
|
if (new_entry->hba_ioaccel_enabled) {
|
|
|
|
|
h->dev[entry]->ioaccel_handle = new_entry->ioaccel_handle;
|
|
|
|
|
wmb(); /* set ioaccel_handle *before* hba_ioaccel_enabled */
|
|
|
|
|
}
|
|
|
|
|
h->dev[entry]->hba_ioaccel_enabled = new_entry->hba_ioaccel_enabled;
|
2014-02-18 19:55:38 +00:00
|
|
|
h->dev[entry]->offload_config = new_entry->offload_config;
|
2014-02-18 19:56:50 +00:00
|
|
|
h->dev[entry]->offload_to_mirror = new_entry->offload_to_mirror;
|
2015-01-23 22:43:30 +00:00
|
|
|
h->dev[entry]->queue_depth = new_entry->queue_depth;
|
2014-02-18 19:55:38 +00:00
|
|
|
|
2015-04-23 14:31:47 +00:00
|
|
|
/*
|
|
|
|
|
* We can turn off ioaccel offload now, but need to delay turning
|
|
|
|
|
* it on until we can update h->dev[entry]->phys_disk[], but we
|
|
|
|
|
* can't do that until all the devices are updated.
|
|
|
|
|
*/
|
|
|
|
|
h->dev[entry]->offload_to_be_enabled = new_entry->offload_enabled;
|
|
|
|
|
if (!new_entry->offload_enabled)
|
|
|
|
|
h->dev[entry]->offload_enabled = 0;
|
|
|
|
|
|
2015-04-23 14:32:54 +00:00
|
|
|
offload_enabled = h->dev[entry]->offload_enabled;
|
|
|
|
|
h->dev[entry]->offload_enabled = h->dev[entry]->offload_to_be_enabled;
|
2015-04-23 14:31:55 +00:00
|
|
|
hpsa_show_dev_msg(KERN_INFO, h, h->dev[entry], "updated");
|
2015-04-23 14:32:54 +00:00
|
|
|
h->dev[entry]->offload_enabled = offload_enabled;
|
2012-01-19 20:01:30 +00:00
|
|
|
}
|
|
|
|
|
|
[SCSI] hpsa: fix bug in adjust_hpsa_scsi_table
fix bug in adjust_hpsa_scsi_table which caused devices which have
changed size, etc. to do the wrong thing.
The problem was as follows:
The driver maintains its current idea of what devices are present
in the h->dev[] array. When it updates this array, it scans the
hardware, and produces a new list of devices, call it sd[], for
scsi devices.
Then, it compares each item in h->dev[] vs. sd[], and any items which
are not present sd it removes from h->dev[], and any items present
in sd[], but different, it modifies in h->dev[].
Then, it looks for items in sd[] which are not present in h->dev[],
and adds those items into h->dev[]. All the while, it keeps track
of what items were added and removed to/from h->dev[].
Finally, it updates the SCSI mid-layer by removing and adding
the same devices it removed and added to/from h->dev[]. (modified
devices count as a remove then add.)
originally, when a "changed" device was discovered, it was
removed then added to h->dev[]. The item was added to the *end*
of h->dev[]. And, the item was removed from sd[] as well
(nulled out). As it processed h->dev[], these newly added items
at the end of the list were encountered, and sd[] was searched,
but those items were nulled out. So they ended up getting removed
immediately after they were added.
The solution is to have a way to replace items in the h->dev[]
array instead of doing a remove + add. Then the "changed" items.
are not encountered a second time, and removed.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-02-04 14:43:41 +00:00
|
|
|
/* Replace an entry from h->dev[] array. */
|
2015-11-04 21:50:01 +00:00
|
|
|
static void hpsa_scsi_replace_entry(struct ctlr_info *h,
|
[SCSI] hpsa: fix bug in adjust_hpsa_scsi_table
fix bug in adjust_hpsa_scsi_table which caused devices which have
changed size, etc. to do the wrong thing.
The problem was as follows:
The driver maintains its current idea of what devices are present
in the h->dev[] array. When it updates this array, it scans the
hardware, and produces a new list of devices, call it sd[], for
scsi devices.
Then, it compares each item in h->dev[] vs. sd[], and any items which
are not present sd it removes from h->dev[], and any items present
in sd[], but different, it modifies in h->dev[].
Then, it looks for items in sd[] which are not present in h->dev[],
and adds those items into h->dev[]. All the while, it keeps track
of what items were added and removed to/from h->dev[].
Finally, it updates the SCSI mid-layer by removing and adding
the same devices it removed and added to/from h->dev[]. (modified
devices count as a remove then add.)
originally, when a "changed" device was discovered, it was
removed then added to h->dev[]. The item was added to the *end*
of h->dev[]. And, the item was removed from sd[] as well
(nulled out). As it processed h->dev[], these newly added items
at the end of the list were encountered, and sd[] was searched,
but those items were nulled out. So they ended up getting removed
immediately after they were added.
The solution is to have a way to replace items in the h->dev[]
array instead of doing a remove + add. Then the "changed" items.
are not encountered a second time, and removed.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-02-04 14:43:41 +00:00
|
|
|
int entry, struct hpsa_scsi_dev_t *new_entry,
|
|
|
|
|
struct hpsa_scsi_dev_t *added[], int *nadded,
|
|
|
|
|
struct hpsa_scsi_dev_t *removed[], int *nremoved)
|
|
|
|
|
{
|
|
|
|
|
/* assumes h->devlock is held */
|
2011-10-26 21:21:07 +00:00
|
|
|
BUG_ON(entry < 0 || entry >= HPSA_MAX_DEVICES);
|
[SCSI] hpsa: fix bug in adjust_hpsa_scsi_table
fix bug in adjust_hpsa_scsi_table which caused devices which have
changed size, etc. to do the wrong thing.
The problem was as follows:
The driver maintains its current idea of what devices are present
in the h->dev[] array. When it updates this array, it scans the
hardware, and produces a new list of devices, call it sd[], for
scsi devices.
Then, it compares each item in h->dev[] vs. sd[], and any items which
are not present sd it removes from h->dev[], and any items present
in sd[], but different, it modifies in h->dev[].
Then, it looks for items in sd[] which are not present in h->dev[],
and adds those items into h->dev[]. All the while, it keeps track
of what items were added and removed to/from h->dev[].
Finally, it updates the SCSI mid-layer by removing and adding
the same devices it removed and added to/from h->dev[]. (modified
devices count as a remove then add.)
originally, when a "changed" device was discovered, it was
removed then added to h->dev[]. The item was added to the *end*
of h->dev[]. And, the item was removed from sd[] as well
(nulled out). As it processed h->dev[], these newly added items
at the end of the list were encountered, and sd[] was searched,
but those items were nulled out. So they ended up getting removed
immediately after they were added.
The solution is to have a way to replace items in the h->dev[]
array instead of doing a remove + add. Then the "changed" items.
are not encountered a second time, and removed.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-02-04 14:43:41 +00:00
|
|
|
removed[*nremoved] = h->dev[entry];
|
|
|
|
|
(*nremoved)++;
|
2011-08-09 13:18:01 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* New physical devices won't have target/lun assigned yet
|
|
|
|
|
* so we need to preserve the values in the slot we are replacing.
|
|
|
|
|
*/
|
|
|
|
|
if (new_entry->target == -1) {
|
|
|
|
|
new_entry->target = h->dev[entry]->target;
|
|
|
|
|
new_entry->lun = h->dev[entry]->lun;
|
|
|
|
|
}
|
|
|
|
|
|
[SCSI] hpsa: fix bug in adjust_hpsa_scsi_table
fix bug in adjust_hpsa_scsi_table which caused devices which have
changed size, etc. to do the wrong thing.
The problem was as follows:
The driver maintains its current idea of what devices are present
in the h->dev[] array. When it updates this array, it scans the
hardware, and produces a new list of devices, call it sd[], for
scsi devices.
Then, it compares each item in h->dev[] vs. sd[], and any items which
are not present sd it removes from h->dev[], and any items present
in sd[], but different, it modifies in h->dev[].
Then, it looks for items in sd[] which are not present in h->dev[],
and adds those items into h->dev[]. All the while, it keeps track
of what items were added and removed to/from h->dev[].
Finally, it updates the SCSI mid-layer by removing and adding
the same devices it removed and added to/from h->dev[]. (modified
devices count as a remove then add.)
originally, when a "changed" device was discovered, it was
removed then added to h->dev[]. The item was added to the *end*
of h->dev[]. And, the item was removed from sd[] as well
(nulled out). As it processed h->dev[], these newly added items
at the end of the list were encountered, and sd[] was searched,
but those items were nulled out. So they ended up getting removed
immediately after they were added.
The solution is to have a way to replace items in the h->dev[]
array instead of doing a remove + add. Then the "changed" items.
are not encountered a second time, and removed.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-02-04 14:43:41 +00:00
|
|
|
h->dev[entry] = new_entry;
|
|
|
|
|
added[*nadded] = new_entry;
|
|
|
|
|
(*nadded)++;
|
2015-04-23 14:31:55 +00:00
|
|
|
hpsa_show_dev_msg(KERN_INFO, h, new_entry, "replaced");
|
2015-04-23 14:32:54 +00:00
|
|
|
new_entry->offload_to_be_enabled = new_entry->offload_enabled;
|
|
|
|
|
new_entry->offload_enabled = 0;
|
[SCSI] hpsa: fix bug in adjust_hpsa_scsi_table
fix bug in adjust_hpsa_scsi_table which caused devices which have
changed size, etc. to do the wrong thing.
The problem was as follows:
The driver maintains its current idea of what devices are present
in the h->dev[] array. When it updates this array, it scans the
hardware, and produces a new list of devices, call it sd[], for
scsi devices.
Then, it compares each item in h->dev[] vs. sd[], and any items which
are not present sd it removes from h->dev[], and any items present
in sd[], but different, it modifies in h->dev[].
Then, it looks for items in sd[] which are not present in h->dev[],
and adds those items into h->dev[]. All the while, it keeps track
of what items were added and removed to/from h->dev[].
Finally, it updates the SCSI mid-layer by removing and adding
the same devices it removed and added to/from h->dev[]. (modified
devices count as a remove then add.)
originally, when a "changed" device was discovered, it was
removed then added to h->dev[]. The item was added to the *end*
of h->dev[]. And, the item was removed from sd[] as well
(nulled out). As it processed h->dev[], these newly added items
at the end of the list were encountered, and sd[] was searched,
but those items were nulled out. So they ended up getting removed
immediately after they were added.
The solution is to have a way to replace items in the h->dev[]
array instead of doing a remove + add. Then the "changed" items.
are not encountered a second time, and removed.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-02-04 14:43:41 +00:00
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/* Remove an entry from h->dev[] array. */
|
2015-11-04 21:50:01 +00:00
|
|
|
static void hpsa_scsi_remove_entry(struct ctlr_info *h, int entry,
|
2009-12-08 22:09:11 +00:00
|
|
|
struct hpsa_scsi_dev_t *removed[], int *nremoved)
|
|
|
|
|
{
|
|
|
|
|
/* assumes h->devlock is held */
|
|
|
|
|
int i;
|
|
|
|
|
struct hpsa_scsi_dev_t *sd;
|
|
|
|
|
|
2011-10-26 21:21:07 +00:00
|
|
|
BUG_ON(entry < 0 || entry >= HPSA_MAX_DEVICES);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
sd = h->dev[entry];
|
|
|
|
|
removed[*nremoved] = h->dev[entry];
|
|
|
|
|
(*nremoved)++;
|
|
|
|
|
|
|
|
|
|
for (i = entry; i < h->ndevices-1; i++)
|
|
|
|
|
h->dev[i] = h->dev[i+1];
|
|
|
|
|
h->ndevices--;
|
2015-04-23 14:31:55 +00:00
|
|
|
hpsa_show_dev_msg(KERN_INFO, h, sd, "removed");
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#define SCSI3ADDR_EQ(a, b) ( \
|
|
|
|
|
(a)[7] == (b)[7] && \
|
|
|
|
|
(a)[6] == (b)[6] && \
|
|
|
|
|
(a)[5] == (b)[5] && \
|
|
|
|
|
(a)[4] == (b)[4] && \
|
|
|
|
|
(a)[3] == (b)[3] && \
|
|
|
|
|
(a)[2] == (b)[2] && \
|
|
|
|
|
(a)[1] == (b)[1] && \
|
|
|
|
|
(a)[0] == (b)[0])
|
|
|
|
|
|
|
|
|
|
static void fixup_botched_add(struct ctlr_info *h,
|
|
|
|
|
struct hpsa_scsi_dev_t *added)
|
|
|
|
|
{
|
|
|
|
|
/* called when scsi_add_device fails in order to re-adjust
|
|
|
|
|
* h->dev[] to match the mid layer's view.
|
|
|
|
|
*/
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
int i, j;
|
|
|
|
|
|
|
|
|
|
spin_lock_irqsave(&h->lock, flags);
|
|
|
|
|
for (i = 0; i < h->ndevices; i++) {
|
|
|
|
|
if (h->dev[i] == added) {
|
|
|
|
|
for (j = i; j < h->ndevices-1; j++)
|
|
|
|
|
h->dev[j] = h->dev[j+1];
|
|
|
|
|
h->ndevices--;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
|
|
|
|
kfree(added);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline int device_is_the_same(struct hpsa_scsi_dev_t *dev1,
|
|
|
|
|
struct hpsa_scsi_dev_t *dev2)
|
|
|
|
|
{
|
|
|
|
|
/* we compare everything except lun and target as these
|
|
|
|
|
* are not yet assigned. Compare parts likely
|
|
|
|
|
* to differ first
|
|
|
|
|
*/
|
|
|
|
|
if (memcmp(dev1->scsi3addr, dev2->scsi3addr,
|
|
|
|
|
sizeof(dev1->scsi3addr)) != 0)
|
|
|
|
|
return 0;
|
|
|
|
|
if (memcmp(dev1->device_id, dev2->device_id,
|
|
|
|
|
sizeof(dev1->device_id)) != 0)
|
|
|
|
|
return 0;
|
|
|
|
|
if (memcmp(dev1->model, dev2->model, sizeof(dev1->model)) != 0)
|
|
|
|
|
return 0;
|
|
|
|
|
if (memcmp(dev1->vendor, dev2->vendor, sizeof(dev1->vendor)) != 0)
|
|
|
|
|
return 0;
|
|
|
|
|
if (dev1->devtype != dev2->devtype)
|
|
|
|
|
return 0;
|
|
|
|
|
if (dev1->bus != dev2->bus)
|
|
|
|
|
return 0;
|
|
|
|
|
return 1;
|
|
|
|
|
}
|
|
|
|
|
|
2012-01-19 20:01:30 +00:00
|
|
|
static inline int device_updated(struct hpsa_scsi_dev_t *dev1,
|
|
|
|
|
struct hpsa_scsi_dev_t *dev2)
|
|
|
|
|
{
|
|
|
|
|
/* Device attributes that can change, but don't mean
|
|
|
|
|
* that the device is a different device, nor that the OS
|
|
|
|
|
* needs to be told anything about the change.
|
|
|
|
|
*/
|
|
|
|
|
if (dev1->raid_level != dev2->raid_level)
|
|
|
|
|
return 1;
|
2014-02-18 19:55:38 +00:00
|
|
|
if (dev1->offload_config != dev2->offload_config)
|
|
|
|
|
return 1;
|
|
|
|
|
if (dev1->offload_enabled != dev2->offload_enabled)
|
|
|
|
|
return 1;
|
2015-07-18 16:12:49 +00:00
|
|
|
if (!is_logical_dev_addr_mode(dev1->scsi3addr))
|
|
|
|
|
if (dev1->queue_depth != dev2->queue_depth)
|
|
|
|
|
return 1;
|
2012-01-19 20:01:30 +00:00
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/* Find needle in haystack. If exact match found, return DEVICE_SAME,
|
|
|
|
|
* and return needle location in *index. If scsi3addr matches, but not
|
|
|
|
|
* vendor, model, serial num, etc. return DEVICE_CHANGED, and return needle
|
2012-01-19 20:01:30 +00:00
|
|
|
* location in *index.
|
|
|
|
|
* In the case of a minor device attribute change, such as RAID level, just
|
|
|
|
|
* return DEVICE_UPDATED, along with the updated device's location in index.
|
|
|
|
|
* If needle not found, return DEVICE_NOT_FOUND.
|
2009-12-08 22:09:11 +00:00
|
|
|
*/
|
|
|
|
|
static int hpsa_scsi_find_entry(struct hpsa_scsi_dev_t *needle,
|
|
|
|
|
struct hpsa_scsi_dev_t *haystack[], int haystack_size,
|
|
|
|
|
int *index)
|
|
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
#define DEVICE_NOT_FOUND 0
|
|
|
|
|
#define DEVICE_CHANGED 1
|
|
|
|
|
#define DEVICE_SAME 2
|
2012-01-19 20:01:30 +00:00
|
|
|
#define DEVICE_UPDATED 3
|
2015-11-04 21:50:31 +00:00
|
|
|
if (needle == NULL)
|
|
|
|
|
return DEVICE_NOT_FOUND;
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
for (i = 0; i < haystack_size; i++) {
|
2010-02-04 14:43:36 +00:00
|
|
|
if (haystack[i] == NULL) /* previously removed. */
|
|
|
|
|
continue;
|
2009-12-08 22:09:11 +00:00
|
|
|
if (SCSI3ADDR_EQ(needle->scsi3addr, haystack[i]->scsi3addr)) {
|
|
|
|
|
*index = i;
|
2012-01-19 20:01:30 +00:00
|
|
|
if (device_is_the_same(needle, haystack[i])) {
|
|
|
|
|
if (device_updated(needle, haystack[i]))
|
|
|
|
|
return DEVICE_UPDATED;
|
2009-12-08 22:09:11 +00:00
|
|
|
return DEVICE_SAME;
|
2012-01-19 20:01:30 +00:00
|
|
|
} else {
|
2014-02-21 22:25:00 +00:00
|
|
|
/* Keep offline devices offline */
|
|
|
|
|
if (needle->volume_offline)
|
|
|
|
|
return DEVICE_NOT_FOUND;
|
2009-12-08 22:09:11 +00:00
|
|
|
return DEVICE_CHANGED;
|
2012-01-19 20:01:30 +00:00
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
*index = -1;
|
|
|
|
|
return DEVICE_NOT_FOUND;
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-21 22:25:00 +00:00
|
|
|
static void hpsa_monitor_offline_device(struct ctlr_info *h,
|
|
|
|
|
unsigned char scsi3addr[])
|
|
|
|
|
{
|
|
|
|
|
struct offline_device_entry *device;
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
|
|
/* Check to see if device is already on the list */
|
|
|
|
|
spin_lock_irqsave(&h->offline_device_lock, flags);
|
|
|
|
|
list_for_each_entry(device, &h->offline_device_list, offline_list) {
|
|
|
|
|
if (memcmp(device->scsi3addr, scsi3addr,
|
|
|
|
|
sizeof(device->scsi3addr)) == 0) {
|
|
|
|
|
spin_unlock_irqrestore(&h->offline_device_lock, flags);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
spin_unlock_irqrestore(&h->offline_device_lock, flags);
|
|
|
|
|
|
|
|
|
|
/* Device is not on the list, add it. */
|
|
|
|
|
device = kmalloc(sizeof(*device), GFP_KERNEL);
|
|
|
|
|
if (!device) {
|
|
|
|
|
dev_warn(&h->pdev->dev, "out of memory in %s\n", __func__);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
memcpy(device->scsi3addr, scsi3addr, sizeof(device->scsi3addr));
|
|
|
|
|
spin_lock_irqsave(&h->offline_device_lock, flags);
|
|
|
|
|
list_add_tail(&device->offline_list, &h->offline_device_list);
|
|
|
|
|
spin_unlock_irqrestore(&h->offline_device_lock, flags);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Print a message explaining various offline volume states */
|
|
|
|
|
static void hpsa_show_volume_status(struct ctlr_info *h,
|
|
|
|
|
struct hpsa_scsi_dev_t *sd)
|
|
|
|
|
{
|
|
|
|
|
if (sd->volume_offline == HPSA_VPD_LV_STATUS_UNSUPPORTED)
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"C%d:B%d:T%d:L%d Volume status is not available through vital product data pages.\n",
|
|
|
|
|
h->scsi_host->host_no,
|
|
|
|
|
sd->bus, sd->target, sd->lun);
|
|
|
|
|
switch (sd->volume_offline) {
|
|
|
|
|
case HPSA_LV_OK:
|
|
|
|
|
break;
|
|
|
|
|
case HPSA_LV_UNDERGOING_ERASE:
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"C%d:B%d:T%d:L%d Volume is undergoing background erase process.\n",
|
|
|
|
|
h->scsi_host->host_no,
|
|
|
|
|
sd->bus, sd->target, sd->lun);
|
|
|
|
|
break;
|
2015-07-18 16:13:04 +00:00
|
|
|
case HPSA_LV_NOT_AVAILABLE:
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"C%d:B%d:T%d:L%d Volume is waiting for transforming volume.\n",
|
|
|
|
|
h->scsi_host->host_no,
|
|
|
|
|
sd->bus, sd->target, sd->lun);
|
|
|
|
|
break;
|
2014-02-21 22:25:00 +00:00
|
|
|
case HPSA_LV_UNDERGOING_RPI:
|
|
|
|
|
dev_info(&h->pdev->dev,
|
2015-07-18 16:13:04 +00:00
|
|
|
"C%d:B%d:T%d:L%d Volume is undergoing rapid parity init.\n",
|
2014-02-21 22:25:00 +00:00
|
|
|
h->scsi_host->host_no,
|
|
|
|
|
sd->bus, sd->target, sd->lun);
|
|
|
|
|
break;
|
|
|
|
|
case HPSA_LV_PENDING_RPI:
|
|
|
|
|
dev_info(&h->pdev->dev,
|
2015-07-18 16:13:04 +00:00
|
|
|
"C%d:B%d:T%d:L%d Volume is queued for rapid parity initialization process.\n",
|
|
|
|
|
h->scsi_host->host_no,
|
|
|
|
|
sd->bus, sd->target, sd->lun);
|
2014-02-21 22:25:00 +00:00
|
|
|
break;
|
|
|
|
|
case HPSA_LV_ENCRYPTED_NO_KEY:
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"C%d:B%d:T%d:L%d Volume is encrypted and cannot be accessed because key is not present.\n",
|
|
|
|
|
h->scsi_host->host_no,
|
|
|
|
|
sd->bus, sd->target, sd->lun);
|
|
|
|
|
break;
|
|
|
|
|
case HPSA_LV_PLAINTEXT_IN_ENCRYPT_ONLY_CONTROLLER:
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"C%d:B%d:T%d:L%d Volume is not encrypted and cannot be accessed because controller is in encryption-only mode.\n",
|
|
|
|
|
h->scsi_host->host_no,
|
|
|
|
|
sd->bus, sd->target, sd->lun);
|
|
|
|
|
break;
|
|
|
|
|
case HPSA_LV_UNDERGOING_ENCRYPTION:
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"C%d:B%d:T%d:L%d Volume is undergoing encryption process.\n",
|
|
|
|
|
h->scsi_host->host_no,
|
|
|
|
|
sd->bus, sd->target, sd->lun);
|
|
|
|
|
break;
|
|
|
|
|
case HPSA_LV_UNDERGOING_ENCRYPTION_REKEYING:
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"C%d:B%d:T%d:L%d Volume is undergoing encryption re-keying process.\n",
|
|
|
|
|
h->scsi_host->host_no,
|
|
|
|
|
sd->bus, sd->target, sd->lun);
|
|
|
|
|
break;
|
|
|
|
|
case HPSA_LV_ENCRYPTED_IN_NON_ENCRYPTED_CONTROLLER:
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"C%d:B%d:T%d:L%d Volume is encrypted and cannot be accessed because controller does not have encryption enabled.\n",
|
|
|
|
|
h->scsi_host->host_no,
|
|
|
|
|
sd->bus, sd->target, sd->lun);
|
|
|
|
|
break;
|
|
|
|
|
case HPSA_LV_PENDING_ENCRYPTION:
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"C%d:B%d:T%d:L%d Volume is pending migration to encrypted state, but process has not started.\n",
|
|
|
|
|
h->scsi_host->host_no,
|
|
|
|
|
sd->bus, sd->target, sd->lun);
|
|
|
|
|
break;
|
|
|
|
|
case HPSA_LV_PENDING_ENCRYPTION_REKEYING:
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"C%d:B%d:T%d:L%d Volume is encrypted and is pending encryption rekeying.\n",
|
|
|
|
|
h->scsi_host->host_no,
|
|
|
|
|
sd->bus, sd->target, sd->lun);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:43:30 +00:00
|
|
|
/*
|
|
|
|
|
* Figure the list of physical drive pointers for a logical drive with
|
|
|
|
|
* raid offload configured.
|
|
|
|
|
*/
|
|
|
|
|
static void hpsa_figure_phys_disk_ptrs(struct ctlr_info *h,
|
|
|
|
|
struct hpsa_scsi_dev_t *dev[], int ndevices,
|
|
|
|
|
struct hpsa_scsi_dev_t *logical_drive)
|
|
|
|
|
{
|
|
|
|
|
struct raid_map_data *map = &logical_drive->raid_map;
|
|
|
|
|
struct raid_map_disk_data *dd = &map->data[0];
|
|
|
|
|
int i, j;
|
|
|
|
|
int total_disks_per_row = le16_to_cpu(map->data_disks_per_row) +
|
|
|
|
|
le16_to_cpu(map->metadata_disks_per_row);
|
|
|
|
|
int nraid_map_entries = le16_to_cpu(map->row_cnt) *
|
|
|
|
|
le16_to_cpu(map->layout_map_count) *
|
|
|
|
|
total_disks_per_row;
|
|
|
|
|
int nphys_disk = le16_to_cpu(map->layout_map_count) *
|
|
|
|
|
total_disks_per_row;
|
|
|
|
|
int qdepth;
|
|
|
|
|
|
|
|
|
|
if (nraid_map_entries > RAID_MAP_MAX_ENTRIES)
|
|
|
|
|
nraid_map_entries = RAID_MAP_MAX_ENTRIES;
|
|
|
|
|
|
2015-04-23 14:35:22 +00:00
|
|
|
logical_drive->nphysical_disks = nraid_map_entries;
|
|
|
|
|
|
2015-01-23 22:43:30 +00:00
|
|
|
qdepth = 0;
|
|
|
|
|
for (i = 0; i < nraid_map_entries; i++) {
|
|
|
|
|
logical_drive->phys_disk[i] = NULL;
|
|
|
|
|
if (!logical_drive->offload_config)
|
|
|
|
|
continue;
|
|
|
|
|
for (j = 0; j < ndevices; j++) {
|
2015-11-04 21:50:31 +00:00
|
|
|
if (dev[j] == NULL)
|
|
|
|
|
continue;
|
2015-01-23 22:43:30 +00:00
|
|
|
if (dev[j]->devtype != TYPE_DISK)
|
|
|
|
|
continue;
|
2016-02-23 21:16:15 +00:00
|
|
|
if (dev[j]->devtype != TYPE_ZBC)
|
|
|
|
|
continue;
|
2015-11-04 21:51:33 +00:00
|
|
|
if (is_logical_device(dev[j]))
|
2015-01-23 22:43:30 +00:00
|
|
|
continue;
|
|
|
|
|
if (dev[j]->ioaccel_handle != dd[i].ioaccel_handle)
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
logical_drive->phys_disk[i] = dev[j];
|
|
|
|
|
if (i < nphys_disk)
|
|
|
|
|
qdepth = min(h->nr_cmds, qdepth +
|
|
|
|
|
logical_drive->phys_disk[i]->queue_depth);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* This can happen if a physical drive is removed and
|
|
|
|
|
* the logical drive is degraded. In that case, the RAID
|
|
|
|
|
* map data will refer to a physical disk which isn't actually
|
|
|
|
|
* present. And in that case offload_enabled should already
|
|
|
|
|
* be 0, but we'll turn it off here just in case
|
|
|
|
|
*/
|
|
|
|
|
if (!logical_drive->phys_disk[i]) {
|
|
|
|
|
logical_drive->offload_enabled = 0;
|
2015-04-23 14:31:47 +00:00
|
|
|
logical_drive->offload_to_be_enabled = 0;
|
|
|
|
|
logical_drive->queue_depth = 8;
|
2015-01-23 22:43:30 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
if (nraid_map_entries)
|
|
|
|
|
/*
|
|
|
|
|
* This is correct for reads, too high for full stripe writes,
|
|
|
|
|
* way too high for partial stripe writes
|
|
|
|
|
*/
|
|
|
|
|
logical_drive->queue_depth = qdepth;
|
|
|
|
|
else
|
|
|
|
|
logical_drive->queue_depth = h->nr_cmds;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_update_log_drive_phys_drive_ptrs(struct ctlr_info *h,
|
|
|
|
|
struct hpsa_scsi_dev_t *dev[], int ndevices)
|
|
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < ndevices; i++) {
|
2015-11-04 21:50:31 +00:00
|
|
|
if (dev[i] == NULL)
|
|
|
|
|
continue;
|
2015-01-23 22:43:30 +00:00
|
|
|
if (dev[i]->devtype != TYPE_DISK)
|
|
|
|
|
continue;
|
2016-02-23 21:16:15 +00:00
|
|
|
if (dev[i]->devtype != TYPE_ZBC)
|
|
|
|
|
continue;
|
2015-11-04 21:51:33 +00:00
|
|
|
if (!is_logical_device(dev[i]))
|
2015-01-23 22:43:30 +00:00
|
|
|
continue;
|
2015-04-23 14:31:47 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* If offload is currently enabled, the RAID map and
|
|
|
|
|
* phys_disk[] assignment *better* not be changing
|
|
|
|
|
* and since it isn't changing, we do not need to
|
|
|
|
|
* update it.
|
|
|
|
|
*/
|
|
|
|
|
if (dev[i]->offload_enabled)
|
|
|
|
|
continue;
|
|
|
|
|
|
2015-01-23 22:43:30 +00:00
|
|
|
hpsa_figure_phys_disk_ptrs(h, dev, ndevices, dev[i]);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-04 21:51:51 +00:00
|
|
|
static int hpsa_add_device(struct ctlr_info *h, struct hpsa_scsi_dev_t *device)
|
|
|
|
|
{
|
|
|
|
|
int rc = 0;
|
|
|
|
|
|
|
|
|
|
if (!h->scsi_host)
|
|
|
|
|
return 1;
|
|
|
|
|
|
2015-11-04 21:52:34 +00:00
|
|
|
if (is_logical_device(device)) /* RAID */
|
|
|
|
|
rc = scsi_add_device(h->scsi_host, device->bus,
|
2015-11-04 21:51:51 +00:00
|
|
|
device->target, device->lun);
|
2015-11-04 21:52:34 +00:00
|
|
|
else /* HBA */
|
|
|
|
|
rc = hpsa_add_sas_device(h->sas_host, device);
|
|
|
|
|
|
2015-11-04 21:51:51 +00:00
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_remove_device(struct ctlr_info *h,
|
|
|
|
|
struct hpsa_scsi_dev_t *device)
|
|
|
|
|
{
|
|
|
|
|
struct scsi_device *sdev = NULL;
|
|
|
|
|
|
|
|
|
|
if (!h->scsi_host)
|
|
|
|
|
return;
|
|
|
|
|
|
2015-11-04 21:52:34 +00:00
|
|
|
if (is_logical_device(device)) { /* RAID */
|
|
|
|
|
sdev = scsi_device_lookup(h->scsi_host, device->bus,
|
2015-11-04 21:51:51 +00:00
|
|
|
device->target, device->lun);
|
2015-11-04 21:52:34 +00:00
|
|
|
if (sdev) {
|
|
|
|
|
scsi_remove_device(sdev);
|
|
|
|
|
scsi_device_put(sdev);
|
|
|
|
|
} else {
|
|
|
|
|
/*
|
|
|
|
|
* We don't expect to get here. Future commands
|
|
|
|
|
* to this device will get a selection timeout as
|
|
|
|
|
* if the device were gone.
|
|
|
|
|
*/
|
|
|
|
|
hpsa_show_dev_msg(KERN_WARNING, h, device,
|
2015-11-04 21:51:51 +00:00
|
|
|
"didn't find device for removal.");
|
2015-11-04 21:52:34 +00:00
|
|
|
}
|
|
|
|
|
} else /* HBA */
|
|
|
|
|
hpsa_remove_sas_device(device);
|
2015-11-04 21:51:51 +00:00
|
|
|
}
|
|
|
|
|
|
2015-11-04 21:50:01 +00:00
|
|
|
static void adjust_hpsa_scsi_table(struct ctlr_info *h,
|
2009-12-08 22:09:11 +00:00
|
|
|
struct hpsa_scsi_dev_t *sd[], int nsds)
|
|
|
|
|
{
|
|
|
|
|
/* sd contains scsi3 addresses and devtypes, and inquiry
|
|
|
|
|
* data. This function takes what's in sd to be the current
|
|
|
|
|
* reality and updates h->dev[] to reflect that reality.
|
|
|
|
|
*/
|
|
|
|
|
int i, entry, device_change, changes = 0;
|
|
|
|
|
struct hpsa_scsi_dev_t *csd;
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
struct hpsa_scsi_dev_t **added, **removed;
|
|
|
|
|
int nadded, nremoved;
|
|
|
|
|
|
2015-11-04 21:50:56 +00:00
|
|
|
/*
|
|
|
|
|
* A reset can cause a device status to change
|
|
|
|
|
* re-schedule the scan to see what happened.
|
|
|
|
|
*/
|
|
|
|
|
if (h->reset_in_progress) {
|
|
|
|
|
h->drv_req_rescan = 1;
|
|
|
|
|
return;
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2011-10-26 21:21:07 +00:00
|
|
|
added = kzalloc(sizeof(*added) * HPSA_MAX_DEVICES, GFP_KERNEL);
|
|
|
|
|
removed = kzalloc(sizeof(*removed) * HPSA_MAX_DEVICES, GFP_KERNEL);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
if (!added || !removed) {
|
|
|
|
|
dev_warn(&h->pdev->dev, "out of memory in "
|
|
|
|
|
"adjust_hpsa_scsi_table\n");
|
|
|
|
|
goto free_and_out;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
spin_lock_irqsave(&h->devlock, flags);
|
|
|
|
|
|
|
|
|
|
/* find any devices in h->dev[] that are not in
|
|
|
|
|
* sd[] and remove them from h->dev[], and for any
|
|
|
|
|
* devices which have changed, remove the old device
|
|
|
|
|
* info and add the new device info.
|
2012-01-19 20:01:30 +00:00
|
|
|
* If minor device attributes change, just update
|
|
|
|
|
* the existing device structure.
|
2009-12-08 22:09:11 +00:00
|
|
|
*/
|
|
|
|
|
i = 0;
|
|
|
|
|
nremoved = 0;
|
|
|
|
|
nadded = 0;
|
|
|
|
|
while (i < h->ndevices) {
|
|
|
|
|
csd = h->dev[i];
|
|
|
|
|
device_change = hpsa_scsi_find_entry(csd, sd, nsds, &entry);
|
|
|
|
|
if (device_change == DEVICE_NOT_FOUND) {
|
|
|
|
|
changes++;
|
2015-11-04 21:50:01 +00:00
|
|
|
hpsa_scsi_remove_entry(h, i, removed, &nremoved);
|
2009-12-08 22:09:11 +00:00
|
|
|
continue; /* remove ^^^, hence i not incremented */
|
|
|
|
|
} else if (device_change == DEVICE_CHANGED) {
|
|
|
|
|
changes++;
|
2015-11-04 21:50:01 +00:00
|
|
|
hpsa_scsi_replace_entry(h, i, sd[entry],
|
[SCSI] hpsa: fix bug in adjust_hpsa_scsi_table
fix bug in adjust_hpsa_scsi_table which caused devices which have
changed size, etc. to do the wrong thing.
The problem was as follows:
The driver maintains its current idea of what devices are present
in the h->dev[] array. When it updates this array, it scans the
hardware, and produces a new list of devices, call it sd[], for
scsi devices.
Then, it compares each item in h->dev[] vs. sd[], and any items which
are not present sd it removes from h->dev[], and any items present
in sd[], but different, it modifies in h->dev[].
Then, it looks for items in sd[] which are not present in h->dev[],
and adds those items into h->dev[]. All the while, it keeps track
of what items were added and removed to/from h->dev[].
Finally, it updates the SCSI mid-layer by removing and adding
the same devices it removed and added to/from h->dev[]. (modified
devices count as a remove then add.)
originally, when a "changed" device was discovered, it was
removed then added to h->dev[]. The item was added to the *end*
of h->dev[]. And, the item was removed from sd[] as well
(nulled out). As it processed h->dev[], these newly added items
at the end of the list were encountered, and sd[] was searched,
but those items were nulled out. So they ended up getting removed
immediately after they were added.
The solution is to have a way to replace items in the h->dev[]
array instead of doing a remove + add. Then the "changed" items.
are not encountered a second time, and removed.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-02-04 14:43:41 +00:00
|
|
|
added, &nadded, removed, &nremoved);
|
2010-02-04 14:43:31 +00:00
|
|
|
/* Set it to NULL to prevent it from being freed
|
|
|
|
|
* at the bottom of hpsa_update_scsi_devices()
|
|
|
|
|
*/
|
|
|
|
|
sd[entry] = NULL;
|
2012-01-19 20:01:30 +00:00
|
|
|
} else if (device_change == DEVICE_UPDATED) {
|
2015-11-04 21:50:01 +00:00
|
|
|
hpsa_scsi_update_entry(h, i, sd[entry]);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
i++;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Now, make sure every device listed in sd[] is also
|
|
|
|
|
* listed in h->dev[], adding them if they aren't found
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < nsds; i++) {
|
|
|
|
|
if (!sd[i]) /* if already added above. */
|
|
|
|
|
continue;
|
2014-02-21 22:25:00 +00:00
|
|
|
|
|
|
|
|
/* Don't add devices which are NOT READY, FORMAT IN PROGRESS
|
|
|
|
|
* as the SCSI mid-layer does not handle such devices well.
|
|
|
|
|
* It relentlessly loops sending TUR at 3Hz, then READ(10)
|
|
|
|
|
* at 160Hz, and prevents the system from coming up.
|
|
|
|
|
*/
|
|
|
|
|
if (sd[i]->volume_offline) {
|
|
|
|
|
hpsa_show_volume_status(h, sd[i]);
|
2015-04-23 14:31:55 +00:00
|
|
|
hpsa_show_dev_msg(KERN_INFO, h, sd[i], "offline");
|
2014-02-21 22:25:00 +00:00
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
device_change = hpsa_scsi_find_entry(sd[i], h->dev,
|
|
|
|
|
h->ndevices, &entry);
|
|
|
|
|
if (device_change == DEVICE_NOT_FOUND) {
|
|
|
|
|
changes++;
|
2015-11-04 21:50:01 +00:00
|
|
|
if (hpsa_scsi_add_entry(h, sd[i], added, &nadded) != 0)
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
sd[i] = NULL; /* prevent from being freed later. */
|
|
|
|
|
} else if (device_change == DEVICE_CHANGED) {
|
|
|
|
|
/* should never happen... */
|
|
|
|
|
changes++;
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"device unexpectedly changed.\n");
|
|
|
|
|
/* but if it does happen, we just ignore that device */
|
|
|
|
|
}
|
|
|
|
|
}
|
2015-04-23 14:31:47 +00:00
|
|
|
hpsa_update_log_drive_phys_drive_ptrs(h, h->dev, h->ndevices);
|
|
|
|
|
|
|
|
|
|
/* Now that h->dev[]->phys_disk[] is coherent, we can enable
|
|
|
|
|
* any logical drives that need it enabled.
|
|
|
|
|
*/
|
2015-11-04 21:50:31 +00:00
|
|
|
for (i = 0; i < h->ndevices; i++) {
|
|
|
|
|
if (h->dev[i] == NULL)
|
|
|
|
|
continue;
|
2015-04-23 14:31:47 +00:00
|
|
|
h->dev[i]->offload_enabled = h->dev[i]->offload_to_be_enabled;
|
2015-11-04 21:50:31 +00:00
|
|
|
}
|
2015-04-23 14:31:47 +00:00
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
spin_unlock_irqrestore(&h->devlock, flags);
|
|
|
|
|
|
2014-02-21 22:25:00 +00:00
|
|
|
/* Monitor devices which are in one of several NOT READY states to be
|
|
|
|
|
* brought online later. This must be done without holding h->devlock,
|
|
|
|
|
* so don't touch h->dev[]
|
|
|
|
|
*/
|
|
|
|
|
for (i = 0; i < nsds; i++) {
|
|
|
|
|
if (!sd[i]) /* if already added above. */
|
|
|
|
|
continue;
|
|
|
|
|
if (sd[i]->volume_offline)
|
|
|
|
|
hpsa_monitor_offline_device(h, sd[i]->scsi3addr);
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/* Don't notify scsi mid layer of any changes the first time through
|
|
|
|
|
* (or if there are no changes) scsi_scan_host will do it later the
|
|
|
|
|
* first time through.
|
|
|
|
|
*/
|
2015-11-04 21:50:01 +00:00
|
|
|
if (!changes)
|
2009-12-08 22:09:11 +00:00
|
|
|
goto free_and_out;
|
|
|
|
|
|
|
|
|
|
/* Notify scsi mid layer of any removed devices */
|
|
|
|
|
for (i = 0; i < nremoved; i++) {
|
2015-11-04 21:50:31 +00:00
|
|
|
if (removed[i] == NULL)
|
|
|
|
|
continue;
|
2015-11-04 21:51:51 +00:00
|
|
|
if (removed[i]->expose_device)
|
|
|
|
|
hpsa_remove_device(h, removed[i]);
|
2009-12-08 22:09:11 +00:00
|
|
|
kfree(removed[i]);
|
|
|
|
|
removed[i] = NULL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Notify scsi mid layer of any added devices */
|
|
|
|
|
for (i = 0; i < nadded; i++) {
|
2015-11-04 21:51:51 +00:00
|
|
|
int rc = 0;
|
|
|
|
|
|
2015-11-04 21:50:31 +00:00
|
|
|
if (added[i] == NULL)
|
|
|
|
|
continue;
|
2015-11-04 21:51:21 +00:00
|
|
|
if (!(added[i]->expose_device))
|
2015-04-23 14:31:47 +00:00
|
|
|
continue;
|
2015-11-04 21:51:51 +00:00
|
|
|
rc = hpsa_add_device(h, added[i]);
|
|
|
|
|
if (!rc)
|
2009-12-08 22:09:11 +00:00
|
|
|
continue;
|
2015-11-04 21:51:51 +00:00
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"addition failed %d, device not added.", rc);
|
2009-12-08 22:09:11 +00:00
|
|
|
/* now we have to remove it from h->dev,
|
|
|
|
|
* since it didn't get added to scsi mid layer
|
|
|
|
|
*/
|
|
|
|
|
fixup_botched_add(h, added[i]);
|
2015-11-04 21:50:37 +00:00
|
|
|
h->drv_req_rescan = 1;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
free_and_out:
|
|
|
|
|
kfree(added);
|
|
|
|
|
kfree(removed);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
2013-09-03 20:45:58 +00:00
|
|
|
* Lookup bus/target/lun and return corresponding struct hpsa_scsi_dev_t *
|
2009-12-08 22:09:11 +00:00
|
|
|
* Assume's h->devlock is held.
|
|
|
|
|
*/
|
|
|
|
|
static struct hpsa_scsi_dev_t *lookup_hpsa_scsi_dev(struct ctlr_info *h,
|
|
|
|
|
int bus, int target, int lun)
|
|
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
struct hpsa_scsi_dev_t *sd;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < h->ndevices; i++) {
|
|
|
|
|
sd = h->dev[i];
|
|
|
|
|
if (sd->bus == bus && sd->target == target && sd->lun == lun)
|
|
|
|
|
return sd;
|
|
|
|
|
}
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int hpsa_slave_alloc(struct scsi_device *sdev)
|
|
|
|
|
{
|
|
|
|
|
struct hpsa_scsi_dev_t *sd;
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
|
|
|
|
|
h = sdev_to_hba(sdev);
|
|
|
|
|
spin_lock_irqsave(&h->devlock, flags);
|
2015-11-04 21:52:34 +00:00
|
|
|
if (sdev_channel(sdev) == HPSA_PHYSICAL_DEVICE_BUS) {
|
|
|
|
|
struct scsi_target *starget;
|
|
|
|
|
struct sas_rphy *rphy;
|
|
|
|
|
|
|
|
|
|
starget = scsi_target(sdev);
|
|
|
|
|
rphy = target_to_rphy(starget);
|
|
|
|
|
sd = hpsa_find_device_by_sas_rphy(h, rphy);
|
|
|
|
|
if (sd) {
|
|
|
|
|
sd->target = sdev_id(sdev);
|
|
|
|
|
sd->lun = sdev->lun;
|
|
|
|
|
}
|
|
|
|
|
} else
|
|
|
|
|
sd = lookup_hpsa_scsi_dev(h, sdev_channel(sdev),
|
|
|
|
|
sdev_id(sdev), sdev->lun);
|
|
|
|
|
|
|
|
|
|
if (sd && sd->expose_device) {
|
2015-01-23 22:43:30 +00:00
|
|
|
atomic_set(&sd->ioaccel_cmds_out, 0);
|
2015-11-04 21:52:34 +00:00
|
|
|
sdev->hostdata = sd;
|
2015-04-23 14:31:47 +00:00
|
|
|
} else
|
|
|
|
|
sdev->hostdata = NULL;
|
2009-12-08 22:09:11 +00:00
|
|
|
spin_unlock_irqrestore(&h->devlock, flags);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:31:47 +00:00
|
|
|
/* configure scsi device based on internal per-device structure */
|
|
|
|
|
static int hpsa_slave_configure(struct scsi_device *sdev)
|
|
|
|
|
{
|
|
|
|
|
struct hpsa_scsi_dev_t *sd;
|
|
|
|
|
int queue_depth;
|
|
|
|
|
|
|
|
|
|
sd = sdev->hostdata;
|
2015-11-04 21:51:21 +00:00
|
|
|
sdev->no_uld_attach = !sd || !sd->expose_device;
|
2015-04-23 14:31:47 +00:00
|
|
|
|
|
|
|
|
if (sd)
|
|
|
|
|
queue_depth = sd->queue_depth != 0 ?
|
|
|
|
|
sd->queue_depth : sdev->host->can_queue;
|
|
|
|
|
else
|
|
|
|
|
queue_depth = sdev->host->can_queue;
|
|
|
|
|
|
|
|
|
|
scsi_change_queue_depth(sdev, queue_depth);
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
static void hpsa_slave_destroy(struct scsi_device *sdev)
|
|
|
|
|
{
|
2010-02-04 14:41:54 +00:00
|
|
|
/* nothing to do. */
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:33:27 +00:00
|
|
|
static void hpsa_free_ioaccel2_sg_chain_blocks(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
if (!h->ioaccel2_cmd_sg_list)
|
|
|
|
|
return;
|
|
|
|
|
for (i = 0; i < h->nr_cmds; i++) {
|
|
|
|
|
kfree(h->ioaccel2_cmd_sg_list[i]);
|
|
|
|
|
h->ioaccel2_cmd_sg_list[i] = NULL;
|
|
|
|
|
}
|
|
|
|
|
kfree(h->ioaccel2_cmd_sg_list);
|
|
|
|
|
h->ioaccel2_cmd_sg_list = NULL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int hpsa_allocate_ioaccel2_sg_chain_blocks(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
if (h->chainsize <= 0)
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
h->ioaccel2_cmd_sg_list =
|
|
|
|
|
kzalloc(sizeof(*h->ioaccel2_cmd_sg_list) * h->nr_cmds,
|
|
|
|
|
GFP_KERNEL);
|
|
|
|
|
if (!h->ioaccel2_cmd_sg_list)
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
for (i = 0; i < h->nr_cmds; i++) {
|
|
|
|
|
h->ioaccel2_cmd_sg_list[i] =
|
|
|
|
|
kmalloc(sizeof(*h->ioaccel2_cmd_sg_list[i]) *
|
|
|
|
|
h->maxsgentries, GFP_KERNEL);
|
|
|
|
|
if (!h->ioaccel2_cmd_sg_list[i])
|
|
|
|
|
goto clean;
|
|
|
|
|
}
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
clean:
|
|
|
|
|
hpsa_free_ioaccel2_sg_chain_blocks(h);
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
}
|
|
|
|
|
|
2010-02-25 20:03:27 +00:00
|
|
|
static void hpsa_free_sg_chain_blocks(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
if (!h->cmd_sg_list)
|
|
|
|
|
return;
|
|
|
|
|
for (i = 0; i < h->nr_cmds; i++) {
|
|
|
|
|
kfree(h->cmd_sg_list[i]);
|
|
|
|
|
h->cmd_sg_list[i] = NULL;
|
|
|
|
|
}
|
|
|
|
|
kfree(h->cmd_sg_list);
|
|
|
|
|
h->cmd_sg_list = NULL;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:33:48 +00:00
|
|
|
static int hpsa_alloc_sg_chain_blocks(struct ctlr_info *h)
|
2010-02-25 20:03:27 +00:00
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
if (h->chainsize <= 0)
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
h->cmd_sg_list = kzalloc(sizeof(*h->cmd_sg_list) * h->nr_cmds,
|
|
|
|
|
GFP_KERNEL);
|
2015-01-23 22:42:42 +00:00
|
|
|
if (!h->cmd_sg_list) {
|
|
|
|
|
dev_err(&h->pdev->dev, "Failed to allocate SG list\n");
|
2010-02-25 20:03:27 +00:00
|
|
|
return -ENOMEM;
|
2015-01-23 22:42:42 +00:00
|
|
|
}
|
2010-02-25 20:03:27 +00:00
|
|
|
for (i = 0; i < h->nr_cmds; i++) {
|
|
|
|
|
h->cmd_sg_list[i] = kmalloc(sizeof(*h->cmd_sg_list[i]) *
|
|
|
|
|
h->chainsize, GFP_KERNEL);
|
2015-01-23 22:42:42 +00:00
|
|
|
if (!h->cmd_sg_list[i]) {
|
|
|
|
|
dev_err(&h->pdev->dev, "Failed to allocate cmd SG\n");
|
2010-02-25 20:03:27 +00:00
|
|
|
goto clean;
|
2015-01-23 22:42:42 +00:00
|
|
|
}
|
2010-02-25 20:03:27 +00:00
|
|
|
}
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
clean:
|
|
|
|
|
hpsa_free_sg_chain_blocks(h);
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:33:27 +00:00
|
|
|
static int hpsa_map_ioaccel2_sg_chain_block(struct ctlr_info *h,
|
|
|
|
|
struct io_accel2_cmd *cp, struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
struct ioaccel2_sg_element *chain_block;
|
|
|
|
|
u64 temp64;
|
|
|
|
|
u32 chain_size;
|
|
|
|
|
|
|
|
|
|
chain_block = h->ioaccel2_cmd_sg_list[c->cmdindex];
|
2015-11-04 21:51:14 +00:00
|
|
|
chain_size = le32_to_cpu(cp->sg[0].length);
|
2015-04-23 14:33:27 +00:00
|
|
|
temp64 = pci_map_single(h->pdev, chain_block, chain_size,
|
|
|
|
|
PCI_DMA_TODEVICE);
|
|
|
|
|
if (dma_mapping_error(&h->pdev->dev, temp64)) {
|
|
|
|
|
/* prevent subsequent unmapping */
|
|
|
|
|
cp->sg->address = 0;
|
|
|
|
|
return -1;
|
|
|
|
|
}
|
|
|
|
|
cp->sg->address = cpu_to_le64(temp64);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_unmap_ioaccel2_sg_chain_block(struct ctlr_info *h,
|
|
|
|
|
struct io_accel2_cmd *cp)
|
|
|
|
|
{
|
|
|
|
|
struct ioaccel2_sg_element *chain_sg;
|
|
|
|
|
u64 temp64;
|
|
|
|
|
u32 chain_size;
|
|
|
|
|
|
|
|
|
|
chain_sg = cp->sg;
|
|
|
|
|
temp64 = le64_to_cpu(chain_sg->address);
|
2015-11-04 21:51:14 +00:00
|
|
|
chain_size = le32_to_cpu(cp->sg[0].length);
|
2015-04-23 14:33:27 +00:00
|
|
|
pci_unmap_single(h->pdev, temp64, chain_size, PCI_DMA_TODEVICE);
|
|
|
|
|
}
|
|
|
|
|
|
2013-02-20 17:24:46 +00:00
|
|
|
static int hpsa_map_sg_chain_block(struct ctlr_info *h,
|
2010-02-25 20:03:27 +00:00
|
|
|
struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
struct SGDescriptor *chain_sg, *chain_block;
|
|
|
|
|
u64 temp64;
|
2014-11-14 23:26:59 +00:00
|
|
|
u32 chain_len;
|
2010-02-25 20:03:27 +00:00
|
|
|
|
|
|
|
|
chain_sg = &c->SG[h->max_cmd_sg_entries - 1];
|
|
|
|
|
chain_block = h->cmd_sg_list[c->cmdindex];
|
2014-11-14 23:26:59 +00:00
|
|
|
chain_sg->Ext = cpu_to_le32(HPSA_SG_CHAIN);
|
|
|
|
|
chain_len = sizeof(*chain_sg) *
|
2015-01-23 22:41:09 +00:00
|
|
|
(le16_to_cpu(c->Header.SGTotal) - h->max_cmd_sg_entries);
|
2014-11-14 23:26:59 +00:00
|
|
|
chain_sg->Len = cpu_to_le32(chain_len);
|
|
|
|
|
temp64 = pci_map_single(h->pdev, chain_block, chain_len,
|
2010-02-25 20:03:27 +00:00
|
|
|
PCI_DMA_TODEVICE);
|
2013-02-20 17:24:46 +00:00
|
|
|
if (dma_mapping_error(&h->pdev->dev, temp64)) {
|
|
|
|
|
/* prevent subsequent unmapping */
|
2014-11-14 23:26:59 +00:00
|
|
|
chain_sg->Addr = cpu_to_le64(0);
|
2013-02-20 17:24:46 +00:00
|
|
|
return -1;
|
|
|
|
|
}
|
2014-11-14 23:26:59 +00:00
|
|
|
chain_sg->Addr = cpu_to_le64(temp64);
|
2013-02-20 17:24:46 +00:00
|
|
|
return 0;
|
2010-02-25 20:03:27 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_unmap_sg_chain_block(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
struct SGDescriptor *chain_sg;
|
|
|
|
|
|
2014-11-14 23:26:59 +00:00
|
|
|
if (le16_to_cpu(c->Header.SGTotal) <= h->max_cmd_sg_entries)
|
2010-02-25 20:03:27 +00:00
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
chain_sg = &c->SG[h->max_cmd_sg_entries - 1];
|
2014-11-14 23:26:59 +00:00
|
|
|
pci_unmap_single(h->pdev, le64_to_cpu(chain_sg->Addr),
|
|
|
|
|
le32_to_cpu(chain_sg->Len), PCI_DMA_TODEVICE);
|
2010-02-25 20:03:27 +00:00
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:57:21 +00:00
|
|
|
|
|
|
|
|
/* Decode the various types of errors on ioaccel2 path.
|
|
|
|
|
* Return 1 for any error that should generate a RAID path retry.
|
|
|
|
|
* Return 0 for errors that don't require a RAID path retry.
|
|
|
|
|
*/
|
|
|
|
|
static int handle_ioaccel_mode2_error(struct ctlr_info *h,
|
2014-02-18 19:56:34 +00:00
|
|
|
struct CommandList *c,
|
|
|
|
|
struct scsi_cmnd *cmd,
|
|
|
|
|
struct io_accel2_cmd *c2)
|
|
|
|
|
{
|
|
|
|
|
int data_len;
|
2014-02-18 19:57:21 +00:00
|
|
|
int retry = 0;
|
2015-04-23 14:33:32 +00:00
|
|
|
u32 ioaccel2_resid = 0;
|
2014-02-18 19:56:34 +00:00
|
|
|
|
|
|
|
|
switch (c2->error_data.serv_response) {
|
|
|
|
|
case IOACCEL2_SERV_RESPONSE_COMPLETE:
|
|
|
|
|
switch (c2->error_data.status) {
|
|
|
|
|
case IOACCEL2_STATUS_SR_TASK_COMP_GOOD:
|
|
|
|
|
break;
|
|
|
|
|
case IOACCEL2_STATUS_SR_TASK_COMP_CHK_COND:
|
2014-05-29 15:53:54 +00:00
|
|
|
cmd->result |= SAM_STAT_CHECK_CONDITION;
|
2014-02-18 19:56:34 +00:00
|
|
|
if (c2->error_data.data_present !=
|
2014-05-29 15:53:54 +00:00
|
|
|
IOACCEL2_SENSE_DATA_PRESENT) {
|
|
|
|
|
memset(cmd->sense_buffer, 0,
|
|
|
|
|
SCSI_SENSE_BUFFERSIZE);
|
2014-02-18 19:56:34 +00:00
|
|
|
break;
|
2014-05-29 15:53:54 +00:00
|
|
|
}
|
2014-02-18 19:56:34 +00:00
|
|
|
/* copy the sense data */
|
|
|
|
|
data_len = c2->error_data.sense_data_len;
|
|
|
|
|
if (data_len > SCSI_SENSE_BUFFERSIZE)
|
|
|
|
|
data_len = SCSI_SENSE_BUFFERSIZE;
|
|
|
|
|
if (data_len > sizeof(c2->error_data.sense_data_buff))
|
|
|
|
|
data_len =
|
|
|
|
|
sizeof(c2->error_data.sense_data_buff);
|
|
|
|
|
memcpy(cmd->sense_buffer,
|
|
|
|
|
c2->error_data.sense_data_buff, data_len);
|
2014-02-18 19:57:21 +00:00
|
|
|
retry = 1;
|
2014-02-18 19:56:34 +00:00
|
|
|
break;
|
|
|
|
|
case IOACCEL2_STATUS_SR_TASK_COMP_BUSY:
|
2014-02-18 19:57:21 +00:00
|
|
|
retry = 1;
|
2014-02-18 19:56:34 +00:00
|
|
|
break;
|
|
|
|
|
case IOACCEL2_STATUS_SR_TASK_COMP_RES_CON:
|
2014-02-18 19:57:21 +00:00
|
|
|
retry = 1;
|
2014-02-18 19:56:34 +00:00
|
|
|
break;
|
|
|
|
|
case IOACCEL2_STATUS_SR_TASK_COMP_SET_FULL:
|
2015-04-23 14:32:43 +00:00
|
|
|
retry = 1;
|
2014-02-18 19:56:34 +00:00
|
|
|
break;
|
|
|
|
|
case IOACCEL2_STATUS_SR_TASK_COMP_ABORTED:
|
2014-02-18 19:57:21 +00:00
|
|
|
retry = 1;
|
2014-02-18 19:56:34 +00:00
|
|
|
break;
|
|
|
|
|
default:
|
2014-02-18 19:57:21 +00:00
|
|
|
retry = 1;
|
2014-02-18 19:56:34 +00:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
case IOACCEL2_SERV_RESPONSE_FAILURE:
|
2015-04-23 14:33:32 +00:00
|
|
|
switch (c2->error_data.status) {
|
|
|
|
|
case IOACCEL2_STATUS_SR_IO_ERROR:
|
|
|
|
|
case IOACCEL2_STATUS_SR_IO_ABORTED:
|
|
|
|
|
case IOACCEL2_STATUS_SR_OVERRUN:
|
|
|
|
|
retry = 1;
|
|
|
|
|
break;
|
|
|
|
|
case IOACCEL2_STATUS_SR_UNDERRUN:
|
|
|
|
|
cmd->result = (DID_OK << 16); /* host byte */
|
|
|
|
|
cmd->result |= (COMMAND_COMPLETE << 8); /* msg byte */
|
|
|
|
|
ioaccel2_resid = get_unaligned_le32(
|
|
|
|
|
&c2->error_data.resid_cnt[0]);
|
|
|
|
|
scsi_set_resid(cmd, ioaccel2_resid);
|
|
|
|
|
break;
|
|
|
|
|
case IOACCEL2_STATUS_SR_NO_PATH_TO_DEVICE:
|
|
|
|
|
case IOACCEL2_STATUS_SR_INVALID_DEVICE:
|
|
|
|
|
case IOACCEL2_STATUS_SR_IOACCEL_DISABLED:
|
|
|
|
|
/* We will get an event from ctlr to trigger rescan */
|
|
|
|
|
retry = 1;
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
retry = 1;
|
|
|
|
|
}
|
2014-02-18 19:56:34 +00:00
|
|
|
break;
|
|
|
|
|
case IOACCEL2_SERV_RESPONSE_TMF_COMPLETE:
|
|
|
|
|
break;
|
|
|
|
|
case IOACCEL2_SERV_RESPONSE_TMF_SUCCESS:
|
|
|
|
|
break;
|
|
|
|
|
case IOACCEL2_SERV_RESPONSE_TMF_REJECTED:
|
2014-02-18 19:57:21 +00:00
|
|
|
retry = 1;
|
2014-02-18 19:56:34 +00:00
|
|
|
break;
|
|
|
|
|
case IOACCEL2_SERV_RESPONSE_TMF_WRONG_LUN:
|
|
|
|
|
break;
|
|
|
|
|
default:
|
2014-02-18 19:57:21 +00:00
|
|
|
retry = 1;
|
2014-02-18 19:56:34 +00:00
|
|
|
break;
|
|
|
|
|
}
|
2014-02-18 19:57:21 +00:00
|
|
|
|
|
|
|
|
return retry; /* retry on raid path? */
|
2014-02-18 19:56:34 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:34:16 +00:00
|
|
|
static void hpsa_cmd_resolve_events(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c)
|
|
|
|
|
{
|
2015-04-23 14:35:22 +00:00
|
|
|
bool do_wake = false;
|
|
|
|
|
|
2015-04-23 14:34:16 +00:00
|
|
|
/*
|
|
|
|
|
* Prevent the following race in the abort handler:
|
|
|
|
|
*
|
|
|
|
|
* 1. LLD is requested to abort a SCSI command
|
|
|
|
|
* 2. The SCSI command completes
|
|
|
|
|
* 3. The struct CommandList associated with step 2 is made available
|
|
|
|
|
* 4. New I/O request to LLD to another LUN re-uses struct CommandList
|
|
|
|
|
* 5. Abort handler follows scsi_cmnd->host_scribble and
|
|
|
|
|
* finds struct CommandList and tries to aborts it
|
|
|
|
|
* Now we have aborted the wrong command.
|
|
|
|
|
*
|
2015-04-23 14:35:22 +00:00
|
|
|
* Reset c->scsi_cmd here so that the abort or reset handler will know
|
|
|
|
|
* this command has completed. Then, check to see if the handler is
|
2015-04-23 14:34:16 +00:00
|
|
|
* waiting for this command, and, if so, wake it.
|
|
|
|
|
*/
|
|
|
|
|
c->scsi_cmd = SCSI_CMD_IDLE;
|
2015-04-23 14:35:22 +00:00
|
|
|
mb(); /* Declare command idle before checking for pending events. */
|
2015-04-23 14:34:16 +00:00
|
|
|
if (c->abort_pending) {
|
2015-04-23 14:35:22 +00:00
|
|
|
do_wake = true;
|
2015-04-23 14:34:16 +00:00
|
|
|
c->abort_pending = false;
|
|
|
|
|
}
|
2015-04-23 14:35:22 +00:00
|
|
|
if (c->reset_pending) {
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
struct hpsa_scsi_dev_t *dev;
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* There appears to be a reset pending; lock the lock and
|
|
|
|
|
* reconfirm. If so, then decrement the count of outstanding
|
|
|
|
|
* commands and wake the reset command if this is the last one.
|
|
|
|
|
*/
|
|
|
|
|
spin_lock_irqsave(&h->lock, flags);
|
|
|
|
|
dev = c->reset_pending; /* Re-fetch under the lock. */
|
|
|
|
|
if (dev && atomic_dec_and_test(&dev->reset_cmds_out))
|
|
|
|
|
do_wake = true;
|
|
|
|
|
c->reset_pending = NULL;
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (do_wake)
|
|
|
|
|
wake_up_all(&h->event_sync_wait_queue);
|
2015-04-23 14:34:16 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:35:04 +00:00
|
|
|
static void hpsa_cmd_resolve_and_free(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
hpsa_cmd_resolve_events(h, c);
|
|
|
|
|
cmd_tagged_free(h, c);
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:34:11 +00:00
|
|
|
static void hpsa_cmd_free_and_done(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c, struct scsi_cmnd *cmd)
|
|
|
|
|
{
|
2015-04-23 14:35:04 +00:00
|
|
|
hpsa_cmd_resolve_and_free(h, c);
|
2015-04-23 14:34:11 +00:00
|
|
|
cmd->scsi_done(cmd);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_retry_cmd(struct ctlr_info *h, struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
INIT_WORK(&c->work, hpsa_command_resubmit_worker);
|
|
|
|
|
queue_work_on(raw_smp_processor_id(), h->resubmit_wq, &c->work);
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:34:16 +00:00
|
|
|
static void hpsa_set_scsi_cmd_aborted(struct scsi_cmnd *cmd)
|
|
|
|
|
{
|
|
|
|
|
cmd->result = DID_ABORT << 16;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_cmd_abort_and_free(struct ctlr_info *h, struct CommandList *c,
|
|
|
|
|
struct scsi_cmnd *cmd)
|
|
|
|
|
{
|
|
|
|
|
hpsa_set_scsi_cmd_aborted(cmd);
|
|
|
|
|
dev_warn(&h->pdev->dev, "CDB %16phN was aborted with status 0x%x\n",
|
|
|
|
|
c->Request.CDB, c->err_info->ScsiStatus);
|
2015-04-23 14:35:04 +00:00
|
|
|
hpsa_cmd_resolve_and_free(h, c);
|
2015-04-23 14:34:16 +00:00
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:56:34 +00:00
|
|
|
static void process_ioaccel2_completion(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c, struct scsi_cmnd *cmd,
|
|
|
|
|
struct hpsa_scsi_dev_t *dev)
|
|
|
|
|
{
|
|
|
|
|
struct io_accel2_cmd *c2 = &h->ioaccel2_cmd_pool[c->cmdindex];
|
|
|
|
|
|
|
|
|
|
/* check for good status */
|
|
|
|
|
if (likely(c2->error_data.serv_response == 0 &&
|
2015-04-23 14:34:11 +00:00
|
|
|
c2->error_data.status == 0))
|
|
|
|
|
return hpsa_cmd_free_and_done(h, c, cmd);
|
2014-02-18 19:56:34 +00:00
|
|
|
|
2015-04-23 14:34:11 +00:00
|
|
|
/*
|
|
|
|
|
* Any RAID offload error results in retry which will use
|
2014-02-18 19:56:34 +00:00
|
|
|
* the normal I/O path so the controller can handle whatever's
|
|
|
|
|
* wrong.
|
|
|
|
|
*/
|
2015-11-04 21:51:33 +00:00
|
|
|
if (is_logical_device(dev) &&
|
2014-02-18 19:56:34 +00:00
|
|
|
c2->error_data.serv_response ==
|
|
|
|
|
IOACCEL2_SERV_RESPONSE_FAILURE) {
|
2015-01-23 22:43:25 +00:00
|
|
|
if (c2->error_data.status ==
|
|
|
|
|
IOACCEL2_STATUS_SR_IOACCEL_DISABLED)
|
|
|
|
|
dev->offload_enabled = 0;
|
2015-04-23 14:34:11 +00:00
|
|
|
|
|
|
|
|
return hpsa_retry_cmd(h, c);
|
2014-02-18 19:57:21 +00:00
|
|
|
}
|
2015-01-23 22:43:25 +00:00
|
|
|
|
|
|
|
|
if (handle_ioaccel_mode2_error(h, c, cmd, c2))
|
2015-04-23 14:34:11 +00:00
|
|
|
return hpsa_retry_cmd(h, c);
|
2015-01-23 22:43:25 +00:00
|
|
|
|
2015-04-23 14:34:11 +00:00
|
|
|
return hpsa_cmd_free_and_done(h, c, cmd);
|
2014-02-18 19:56:34 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:32:16 +00:00
|
|
|
/* Returns 0 on success, < 0 otherwise. */
|
|
|
|
|
static int hpsa_evaluate_tmf_status(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *cp)
|
|
|
|
|
{
|
|
|
|
|
u8 tmf_status = cp->err_info->ScsiStatus;
|
|
|
|
|
|
|
|
|
|
switch (tmf_status) {
|
|
|
|
|
case CISS_TMF_COMPLETE:
|
|
|
|
|
/*
|
|
|
|
|
* CISS_TMF_COMPLETE never happens, instead,
|
|
|
|
|
* ei->CommandStatus == 0 for this case.
|
|
|
|
|
*/
|
|
|
|
|
case CISS_TMF_SUCCESS:
|
|
|
|
|
return 0;
|
|
|
|
|
case CISS_TMF_INVALID_FRAME:
|
|
|
|
|
case CISS_TMF_NOT_SUPPORTED:
|
|
|
|
|
case CISS_TMF_FAILED:
|
|
|
|
|
case CISS_TMF_WRONG_LUN:
|
|
|
|
|
case CISS_TMF_OVERLAPPED_TAG:
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
dev_warn(&h->pdev->dev, "Unknown TMF status: 0x%02x\n",
|
|
|
|
|
tmf_status);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
return -tmf_status;
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-03 19:59:00 +00:00
|
|
|
static void complete_scsi_command(struct CommandList *cp)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
|
|
|
|
struct scsi_cmnd *cmd;
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct ErrorInfo *ei;
|
2014-02-18 19:55:33 +00:00
|
|
|
struct hpsa_scsi_dev_t *dev;
|
2015-04-23 14:33:27 +00:00
|
|
|
struct io_accel2_cmd *c2;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-04-23 14:32:16 +00:00
|
|
|
u8 sense_key;
|
|
|
|
|
u8 asc; /* additional sense code */
|
|
|
|
|
u8 ascq; /* additional sense code qualifier */
|
[SCSI] hpsa: fix potential overrun while memcpy'ing sense data
This memcpy:
memcpy(cmd->sense_buffer, ei->SenseInfo,
ei->SenseLen > SCSI_SENSE_BUFFERSIZE ?
SCSI_SENSE_BUFFERSIZE :
ei->SenseLen);
The ei->SenseLen field is filled in by the Smart Array. For requests to
logical drives, it will not exceed 32 bytes, so should be ok, but for physical
requests it depends on the target device, not the Smart Array. It's conceivable
that this could exceed the 32 byte size of ei->SenseInfo. In that case, the memcpy
would read past the end of ei->SenseInfo, copying data from the next command,
as if it were sense data, or, if it happened to be the very last command in the
block of allocated commands, could fall off the end of the allocated area and
crash. I'm not aware of anyone ever encountering this behavior, but it could
conceivably happen. This bug was found by Coverity.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2011-06-03 14:57:34 +00:00
|
|
|
unsigned long sense_data_size;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
ei = cp->err_info;
|
2015-01-23 22:44:30 +00:00
|
|
|
cmd = cp->scsi_cmd;
|
2009-12-08 22:09:11 +00:00
|
|
|
h = cp->h;
|
2014-02-18 19:55:33 +00:00
|
|
|
dev = cmd->device->hostdata;
|
2015-04-23 14:33:27 +00:00
|
|
|
c2 = &h->ioaccel2_cmd_pool[cp->cmdindex];
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
scsi_dma_unmap(cmd); /* undo the DMA mappings */
|
2014-02-18 19:55:17 +00:00
|
|
|
if ((cp->cmd_type == CMD_SCSI) &&
|
2015-01-23 22:41:09 +00:00
|
|
|
(le16_to_cpu(cp->Header.SGTotal) > h->max_cmd_sg_entries))
|
2010-02-25 20:03:27 +00:00
|
|
|
hpsa_unmap_sg_chain_block(h, cp);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-04-23 14:33:27 +00:00
|
|
|
if ((cp->cmd_type == CMD_IOACCEL2) &&
|
|
|
|
|
(c2->sg[0].chain_indicator == IOACCEL2_CHAIN))
|
|
|
|
|
hpsa_unmap_ioaccel2_sg_chain_block(h, c2);
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
cmd->result = (DID_OK << 16); /* host byte */
|
|
|
|
|
cmd->result |= (COMMAND_COMPLETE << 8); /* msg byte */
|
2014-02-18 19:56:34 +00:00
|
|
|
|
2015-01-23 22:43:30 +00:00
|
|
|
if (cp->cmd_type == CMD_IOACCEL2 || cp->cmd_type == CMD_IOACCEL1)
|
|
|
|
|
atomic_dec(&cp->phys_disk->ioaccel_cmds_out);
|
|
|
|
|
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
/*
|
|
|
|
|
* We check for lockup status here as it may be set for
|
|
|
|
|
* CMD_SCSI, CMD_IOACCEL1 and CMD_IOACCEL2 commands by
|
|
|
|
|
* fail_all_oustanding_cmds()
|
|
|
|
|
*/
|
|
|
|
|
if (unlikely(ei->CommandStatus == CMD_CTLR_LOCKUP)) {
|
|
|
|
|
/* DID_NO_CONNECT will prevent a retry */
|
|
|
|
|
cmd->result = DID_NO_CONNECT << 16;
|
2015-04-23 14:34:11 +00:00
|
|
|
return hpsa_cmd_free_and_done(h, cp, cmd);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:35:22 +00:00
|
|
|
if ((unlikely(hpsa_is_pending_event(cp)))) {
|
|
|
|
|
if (cp->reset_pending)
|
|
|
|
|
return hpsa_cmd_resolve_and_free(h, cp);
|
|
|
|
|
if (cp->abort_pending)
|
|
|
|
|
return hpsa_cmd_abort_and_free(h, cp, cmd);
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:56:34 +00:00
|
|
|
if (cp->cmd_type == CMD_IOACCEL2)
|
|
|
|
|
return process_ioaccel2_completion(h, cp, cmd, dev);
|
|
|
|
|
|
2014-07-03 15:18:19 +00:00
|
|
|
scsi_set_resid(cmd, ei->ResidualCnt);
|
2015-04-23 14:34:11 +00:00
|
|
|
if (ei->CommandStatus == 0)
|
|
|
|
|
return hpsa_cmd_free_and_done(h, cp, cmd);
|
2014-07-03 15:18:19 +00:00
|
|
|
|
2014-02-18 19:55:17 +00:00
|
|
|
/* For I/O accelerator commands, copy over some fields to the normal
|
|
|
|
|
* CISS header used below for error handling.
|
|
|
|
|
*/
|
|
|
|
|
if (cp->cmd_type == CMD_IOACCEL1) {
|
|
|
|
|
struct io_accel1_cmd *c = &h->ioaccel_cmd_pool[cp->cmdindex];
|
2015-01-23 22:41:09 +00:00
|
|
|
cp->Header.SGList = scsi_sg_count(cmd);
|
|
|
|
|
cp->Header.SGTotal = cpu_to_le16(cp->Header.SGList);
|
|
|
|
|
cp->Request.CDBLen = le16_to_cpu(c->io_flags) &
|
|
|
|
|
IOACCEL1_IOFLAGS_CDBLEN_MASK;
|
2014-11-14 23:26:59 +00:00
|
|
|
cp->Header.tag = c->tag;
|
2014-02-18 19:55:17 +00:00
|
|
|
memcpy(cp->Header.LUN.LunAddrBytes, c->CISS_LUN, 8);
|
|
|
|
|
memcpy(cp->Request.CDB, c->CDB, cp->Request.CDBLen);
|
2014-02-18 19:55:33 +00:00
|
|
|
|
|
|
|
|
/* Any RAID offload error results in retry which will use
|
|
|
|
|
* the normal I/O path so the controller can handle whatever's
|
|
|
|
|
* wrong.
|
|
|
|
|
*/
|
2015-11-04 21:51:33 +00:00
|
|
|
if (is_logical_device(dev)) {
|
2014-02-18 19:55:33 +00:00
|
|
|
if (ei->CommandStatus == CMD_IOACCEL_DISABLED)
|
|
|
|
|
dev->offload_enabled = 0;
|
2015-04-23 14:35:22 +00:00
|
|
|
return hpsa_retry_cmd(h, cp);
|
2014-02-18 19:55:33 +00:00
|
|
|
}
|
2014-02-18 19:55:17 +00:00
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/* an error has occurred */
|
|
|
|
|
switch (ei->CommandStatus) {
|
|
|
|
|
|
|
|
|
|
case CMD_TARGET_STATUS:
|
2015-04-23 14:32:16 +00:00
|
|
|
cmd->result |= ei->ScsiStatus;
|
|
|
|
|
/* copy the sense data */
|
|
|
|
|
if (SCSI_SENSE_BUFFERSIZE < sizeof(ei->SenseInfo))
|
|
|
|
|
sense_data_size = SCSI_SENSE_BUFFERSIZE;
|
|
|
|
|
else
|
|
|
|
|
sense_data_size = sizeof(ei->SenseInfo);
|
|
|
|
|
if (ei->SenseLen < sense_data_size)
|
|
|
|
|
sense_data_size = ei->SenseLen;
|
|
|
|
|
memcpy(cmd->sense_buffer, ei->SenseInfo, sense_data_size);
|
|
|
|
|
if (ei->ScsiStatus)
|
|
|
|
|
decode_sense_data(ei->SenseInfo, sense_data_size,
|
|
|
|
|
&sense_key, &asc, &ascq);
|
2009-12-08 22:09:11 +00:00
|
|
|
if (ei->ScsiStatus == SAM_STAT_CHECK_CONDITION) {
|
2010-02-04 14:43:00 +00:00
|
|
|
if (sense_key == ABORTED_COMMAND) {
|
2013-09-23 18:33:41 +00:00
|
|
|
cmd->result |= DID_SOFT_ERROR << 16;
|
2010-02-04 14:43:00 +00:00
|
|
|
break;
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
/* Problem was not a check condition
|
|
|
|
|
* Pass it up to the upper layers...
|
|
|
|
|
*/
|
|
|
|
|
if (ei->ScsiStatus) {
|
|
|
|
|
dev_warn(&h->pdev->dev, "cp %p has status 0x%x "
|
|
|
|
|
"Sense: 0x%x, ASC: 0x%x, ASCQ: 0x%x, "
|
|
|
|
|
"Returning result: 0x%x\n",
|
|
|
|
|
cp, ei->ScsiStatus,
|
|
|
|
|
sense_key, asc, ascq,
|
|
|
|
|
cmd->result);
|
|
|
|
|
} else { /* scsi status is zero??? How??? */
|
|
|
|
|
dev_warn(&h->pdev->dev, "cp %p SCSI status was 0. "
|
|
|
|
|
"Returning no connection.\n", cp),
|
|
|
|
|
|
|
|
|
|
/* Ordinarily, this case should never happen,
|
|
|
|
|
* but there is a bug in some released firmware
|
|
|
|
|
* revisions that allows it to happen if, for
|
|
|
|
|
* example, a 4100 backplane loses power and
|
|
|
|
|
* the tape drive is in it. We assume that
|
|
|
|
|
* it's a fatal error of some kind because we
|
|
|
|
|
* can't show that it wasn't. We will make it
|
|
|
|
|
* look like selection timeout since that is
|
|
|
|
|
* the most common reason for this to occur,
|
|
|
|
|
* and it's severe enough.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
cmd->result = DID_NO_CONNECT << 16;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case CMD_DATA_UNDERRUN: /* let mid layer handle it. */
|
|
|
|
|
break;
|
|
|
|
|
case CMD_DATA_OVERRUN:
|
2015-01-23 22:44:35 +00:00
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"CDB %16phN data overrun\n", cp->Request.CDB);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case CMD_INVALID: {
|
|
|
|
|
/* print_bytes(cp, sizeof(*cp), 1, 0);
|
|
|
|
|
print_cmd(cp); */
|
|
|
|
|
/* We get CMD_INVALID if you address a non-existent device
|
|
|
|
|
* instead of a selection timeout (no response). You will
|
|
|
|
|
* see this if you yank out a drive, then try to access it.
|
|
|
|
|
* This is kind of a shame because it means that any other
|
|
|
|
|
* CMD_INVALID (e.g. driver bug) will get interpreted as a
|
|
|
|
|
* missing target. */
|
|
|
|
|
cmd->result = DID_NO_CONNECT << 16;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
case CMD_PROTOCOL_ERR:
|
2012-09-14 21:34:25 +00:00
|
|
|
cmd->result = DID_ERROR << 16;
|
2015-01-23 22:44:35 +00:00
|
|
|
dev_warn(&h->pdev->dev, "CDB %16phN : protocol error\n",
|
|
|
|
|
cp->Request.CDB);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case CMD_HARDWARE_ERR:
|
|
|
|
|
cmd->result = DID_ERROR << 16;
|
2015-01-23 22:44:35 +00:00
|
|
|
dev_warn(&h->pdev->dev, "CDB %16phN : hardware error\n",
|
|
|
|
|
cp->Request.CDB);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case CMD_CONNECTION_LOST:
|
|
|
|
|
cmd->result = DID_ERROR << 16;
|
2015-01-23 22:44:35 +00:00
|
|
|
dev_warn(&h->pdev->dev, "CDB %16phN : connection lost\n",
|
|
|
|
|
cp->Request.CDB);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case CMD_ABORTED:
|
2015-04-23 14:34:16 +00:00
|
|
|
/* Return now to avoid calling scsi_done(). */
|
|
|
|
|
return hpsa_cmd_abort_and_free(h, cp, cmd);
|
2009-12-08 22:09:11 +00:00
|
|
|
case CMD_ABORT_FAILED:
|
|
|
|
|
cmd->result = DID_ERROR << 16;
|
2015-01-23 22:44:35 +00:00
|
|
|
dev_warn(&h->pdev->dev, "CDB %16phN : abort failed\n",
|
|
|
|
|
cp->Request.CDB);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case CMD_UNSOLICITED_ABORT:
|
2011-07-26 16:08:52 +00:00
|
|
|
cmd->result = DID_SOFT_ERROR << 16; /* retry the command */
|
2015-01-23 22:44:35 +00:00
|
|
|
dev_warn(&h->pdev->dev, "CDB %16phN : unsolicited abort\n",
|
|
|
|
|
cp->Request.CDB);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case CMD_TIMEOUT:
|
|
|
|
|
cmd->result = DID_TIME_OUT << 16;
|
2015-01-23 22:44:35 +00:00
|
|
|
dev_warn(&h->pdev->dev, "CDB %16phN timed out\n",
|
|
|
|
|
cp->Request.CDB);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
2011-01-07 16:55:48 +00:00
|
|
|
case CMD_UNABORTABLE:
|
|
|
|
|
cmd->result = DID_ERROR << 16;
|
|
|
|
|
dev_warn(&h->pdev->dev, "Command unabortable\n");
|
|
|
|
|
break;
|
2015-04-23 14:32:16 +00:00
|
|
|
case CMD_TMF_STATUS:
|
|
|
|
|
if (hpsa_evaluate_tmf_status(h, cp)) /* TMF failed? */
|
|
|
|
|
cmd->result = DID_ERROR << 16;
|
|
|
|
|
break;
|
2014-02-18 19:55:33 +00:00
|
|
|
case CMD_IOACCEL_DISABLED:
|
|
|
|
|
/* This only handles the direct pass-through case since RAID
|
|
|
|
|
* offload is handled above. Just attempt a retry.
|
|
|
|
|
*/
|
|
|
|
|
cmd->result = DID_SOFT_ERROR << 16;
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"cp %p had HP SSD Smart Path error\n", cp);
|
|
|
|
|
break;
|
2009-12-08 22:09:11 +00:00
|
|
|
default:
|
|
|
|
|
cmd->result = DID_ERROR << 16;
|
|
|
|
|
dev_warn(&h->pdev->dev, "cp %p returned unknown status %x\n",
|
|
|
|
|
cp, ei->CommandStatus);
|
|
|
|
|
}
|
2015-04-23 14:34:11 +00:00
|
|
|
|
|
|
|
|
return hpsa_cmd_free_and_done(h, cp, cmd);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_pci_unmap(struct pci_dev *pdev,
|
|
|
|
|
struct CommandList *c, int sg_used, int data_direction)
|
|
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
2014-11-14 23:26:59 +00:00
|
|
|
for (i = 0; i < sg_used; i++)
|
|
|
|
|
pci_unmap_single(pdev, (dma_addr_t) le64_to_cpu(c->SG[i].Addr),
|
|
|
|
|
le32_to_cpu(c->SG[i].Len),
|
|
|
|
|
data_direction);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
2013-02-20 17:24:41 +00:00
|
|
|
static int hpsa_map_one(struct pci_dev *pdev,
|
2009-12-08 22:09:11 +00:00
|
|
|
struct CommandList *cp,
|
|
|
|
|
unsigned char *buf,
|
|
|
|
|
size_t buflen,
|
|
|
|
|
int data_direction)
|
|
|
|
|
{
|
2010-02-04 14:41:33 +00:00
|
|
|
u64 addr64;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
if (buflen == 0 || data_direction == PCI_DMA_NONE) {
|
|
|
|
|
cp->Header.SGList = 0;
|
2014-11-14 23:26:59 +00:00
|
|
|
cp->Header.SGTotal = cpu_to_le16(0);
|
2013-02-20 17:24:41 +00:00
|
|
|
return 0;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
2014-11-14 23:26:59 +00:00
|
|
|
addr64 = pci_map_single(pdev, buf, buflen, data_direction);
|
2013-02-20 17:24:34 +00:00
|
|
|
if (dma_mapping_error(&pdev->dev, addr64)) {
|
2013-02-20 17:24:41 +00:00
|
|
|
/* Prevent subsequent unmap of something never mapped */
|
2013-02-20 17:24:34 +00:00
|
|
|
cp->Header.SGList = 0;
|
2014-11-14 23:26:59 +00:00
|
|
|
cp->Header.SGTotal = cpu_to_le16(0);
|
2013-02-20 17:24:41 +00:00
|
|
|
return -1;
|
2013-02-20 17:24:34 +00:00
|
|
|
}
|
2014-11-14 23:26:59 +00:00
|
|
|
cp->SG[0].Addr = cpu_to_le64(addr64);
|
|
|
|
|
cp->SG[0].Len = cpu_to_le32(buflen);
|
|
|
|
|
cp->SG[0].Ext = cpu_to_le32(HPSA_SG_LAST); /* we are not chaining */
|
|
|
|
|
cp->Header.SGList = 1; /* no. SGs contig in this cmd */
|
|
|
|
|
cp->Header.SGTotal = cpu_to_le16(1); /* total sgs in cmd list */
|
2013-02-20 17:24:41 +00:00
|
|
|
return 0;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
#define NO_TIMEOUT ((unsigned long) -1)
|
|
|
|
|
#define DEFAULT_TIMEOUT 30000 /* milliseconds */
|
|
|
|
|
static int hpsa_scsi_do_simple_cmd_core(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c, int reply_queue, unsigned long timeout_msecs)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
|
|
|
|
DECLARE_COMPLETION_ONSTACK(wait);
|
|
|
|
|
|
|
|
|
|
c->waiting = &wait;
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
__enqueue_cmd_and_start_io(h, c, reply_queue);
|
|
|
|
|
if (timeout_msecs == NO_TIMEOUT) {
|
|
|
|
|
/* TODO: get rid of this no-timeout thing */
|
|
|
|
|
wait_for_completion_io(&wait);
|
|
|
|
|
return IO_OK;
|
|
|
|
|
}
|
|
|
|
|
if (!wait_for_completion_io_timeout(&wait,
|
|
|
|
|
msecs_to_jiffies(timeout_msecs))) {
|
|
|
|
|
dev_warn(&h->pdev->dev, "Command timed out.\n");
|
|
|
|
|
return -ETIMEDOUT;
|
|
|
|
|
}
|
|
|
|
|
return IO_OK;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int hpsa_scsi_do_simple_cmd(struct ctlr_info *h, struct CommandList *c,
|
|
|
|
|
int reply_queue, unsigned long timeout_msecs)
|
|
|
|
|
{
|
|
|
|
|
if (unlikely(lockup_detected(h))) {
|
|
|
|
|
c->err_info->CommandStatus = CMD_CTLR_LOCKUP;
|
|
|
|
|
return IO_OK;
|
|
|
|
|
}
|
|
|
|
|
return hpsa_scsi_do_simple_cmd_core(h, c, reply_queue, timeout_msecs);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
2014-05-29 15:53:18 +00:00
|
|
|
static u32 lockup_detected(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
int cpu;
|
|
|
|
|
u32 rc, *lockup_detected;
|
|
|
|
|
|
|
|
|
|
cpu = get_cpu();
|
|
|
|
|
lockup_detected = per_cpu_ptr(h->lockup_detected, cpu);
|
|
|
|
|
rc = *lockup_detected;
|
|
|
|
|
put_cpu();
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2012-05-01 16:42:40 +00:00
|
|
|
#define MAX_DRIVER_CMD_RETRIES 25
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
static int hpsa_scsi_do_simple_cmd_with_retry(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c, int data_direction, unsigned long timeout_msecs)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
2012-05-01 16:42:40 +00:00
|
|
|
int backoff_time = 10, retry_count = 0;
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
int rc;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
do {
|
2011-05-09 06:32:40 +00:00
|
|
|
memset(c->err_info, 0, sizeof(*c->err_info));
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
rc = hpsa_scsi_do_simple_cmd(h, c, DEFAULT_REPLY_QUEUE,
|
|
|
|
|
timeout_msecs);
|
|
|
|
|
if (rc)
|
|
|
|
|
break;
|
2009-12-08 22:09:11 +00:00
|
|
|
retry_count++;
|
2012-05-01 16:42:40 +00:00
|
|
|
if (retry_count > 3) {
|
|
|
|
|
msleep(backoff_time);
|
|
|
|
|
if (backoff_time < 1000)
|
|
|
|
|
backoff_time *= 2;
|
|
|
|
|
}
|
2012-05-01 16:42:35 +00:00
|
|
|
} while ((check_for_unit_attention(h, c) ||
|
2012-05-01 16:42:40 +00:00
|
|
|
check_for_busy(h, c)) &&
|
|
|
|
|
retry_count <= MAX_DRIVER_CMD_RETRIES);
|
2009-12-08 22:09:11 +00:00
|
|
|
hpsa_pci_unmap(h->pdev, c, 1, data_direction);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
if (retry_count > MAX_DRIVER_CMD_RETRIES)
|
|
|
|
|
rc = -EIO;
|
|
|
|
|
return rc;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:57:47 +00:00
|
|
|
static void hpsa_print_cmd(struct ctlr_info *h, char *txt,
|
|
|
|
|
struct CommandList *c)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
2014-02-18 19:57:47 +00:00
|
|
|
const u8 *cdb = c->Request.CDB;
|
|
|
|
|
const u8 *lun = c->Header.LUN.LunAddrBytes;
|
|
|
|
|
|
|
|
|
|
dev_warn(&h->pdev->dev, "%s: LUN:%02x%02x%02x%02x%02x%02x%02x%02x"
|
|
|
|
|
" CDB:%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x%02x\n",
|
|
|
|
|
txt, lun[0], lun[1], lun[2], lun[3],
|
|
|
|
|
lun[4], lun[5], lun[6], lun[7],
|
|
|
|
|
cdb[0], cdb[1], cdb[2], cdb[3],
|
|
|
|
|
cdb[4], cdb[5], cdb[6], cdb[7],
|
|
|
|
|
cdb[8], cdb[9], cdb[10], cdb[11],
|
|
|
|
|
cdb[12], cdb[13], cdb[14], cdb[15]);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_scsi_interpret_error(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *cp)
|
|
|
|
|
{
|
|
|
|
|
const struct ErrorInfo *ei = cp->err_info;
|
2009-12-08 22:09:11 +00:00
|
|
|
struct device *d = &cp->h->pdev->dev;
|
2015-04-23 14:32:16 +00:00
|
|
|
u8 sense_key, asc, ascq;
|
|
|
|
|
int sense_len;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
switch (ei->CommandStatus) {
|
|
|
|
|
case CMD_TARGET_STATUS:
|
2015-04-23 14:32:16 +00:00
|
|
|
if (ei->SenseLen > sizeof(ei->SenseInfo))
|
|
|
|
|
sense_len = sizeof(ei->SenseInfo);
|
|
|
|
|
else
|
|
|
|
|
sense_len = ei->SenseLen;
|
|
|
|
|
decode_sense_data(ei->SenseInfo, sense_len,
|
|
|
|
|
&sense_key, &asc, &ascq);
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_print_cmd(h, "SCSI status", cp);
|
|
|
|
|
if (ei->ScsiStatus == SAM_STAT_CHECK_CONDITION)
|
2015-04-23 14:32:16 +00:00
|
|
|
dev_warn(d, "SCSI Status = 02, Sense key = 0x%02x, ASC = 0x%02x, ASCQ = 0x%02x\n",
|
|
|
|
|
sense_key, asc, ascq);
|
2014-02-18 19:57:47 +00:00
|
|
|
else
|
2015-04-23 14:32:16 +00:00
|
|
|
dev_warn(d, "SCSI Status = 0x%02x\n", ei->ScsiStatus);
|
2009-12-08 22:09:11 +00:00
|
|
|
if (ei->ScsiStatus == 0)
|
|
|
|
|
dev_warn(d, "SCSI status is abnormally zero. "
|
|
|
|
|
"(probably indicates selection timeout "
|
|
|
|
|
"reported incorrectly due to a known "
|
|
|
|
|
"firmware bug, circa July, 2001.)\n");
|
|
|
|
|
break;
|
|
|
|
|
case CMD_DATA_UNDERRUN: /* let mid layer handle it. */
|
|
|
|
|
break;
|
|
|
|
|
case CMD_DATA_OVERRUN:
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_print_cmd(h, "overrun condition", cp);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case CMD_INVALID: {
|
|
|
|
|
/* controller unfortunately reports SCSI passthru's
|
|
|
|
|
* to non-existent targets as invalid commands.
|
|
|
|
|
*/
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_print_cmd(h, "invalid command", cp);
|
|
|
|
|
dev_warn(d, "probably means device no longer present\n");
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
case CMD_PROTOCOL_ERR:
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_print_cmd(h, "protocol error", cp);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case CMD_HARDWARE_ERR:
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_print_cmd(h, "hardware error", cp);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case CMD_CONNECTION_LOST:
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_print_cmd(h, "connection lost", cp);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case CMD_ABORTED:
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_print_cmd(h, "aborted", cp);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case CMD_ABORT_FAILED:
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_print_cmd(h, "abort failed", cp);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case CMD_UNSOLICITED_ABORT:
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_print_cmd(h, "unsolicited abort", cp);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case CMD_TIMEOUT:
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_print_cmd(h, "timed out", cp);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
2011-01-07 16:55:48 +00:00
|
|
|
case CMD_UNABORTABLE:
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_print_cmd(h, "unabortable", cp);
|
2011-01-07 16:55:48 +00:00
|
|
|
break;
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
case CMD_CTLR_LOCKUP:
|
|
|
|
|
hpsa_print_cmd(h, "controller lockup detected", cp);
|
|
|
|
|
break;
|
2009-12-08 22:09:11 +00:00
|
|
|
default:
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_print_cmd(h, "unknown status", cp);
|
|
|
|
|
dev_warn(d, "Unknown command status %x\n",
|
2009-12-08 22:09:11 +00:00
|
|
|
ei->CommandStatus);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int hpsa_scsi_do_inquiry(struct ctlr_info *h, unsigned char *scsi3addr,
|
2014-02-18 19:57:11 +00:00
|
|
|
u16 page, unsigned char *buf,
|
2009-12-08 22:09:11 +00:00
|
|
|
unsigned char bufsize)
|
|
|
|
|
{
|
|
|
|
|
int rc = IO_OK;
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
struct ErrorInfo *ei;
|
|
|
|
|
|
2015-01-23 22:43:04 +00:00
|
|
|
c = cmd_alloc(h);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2013-02-20 17:24:41 +00:00
|
|
|
if (fill_cmd(c, HPSA_INQUIRY, h, buf, bufsize,
|
|
|
|
|
page, scsi3addr, TYPE_CMD)) {
|
|
|
|
|
rc = -1;
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
rc = hpsa_scsi_do_simple_cmd_with_retry(h, c,
|
2016-04-27 22:13:51 +00:00
|
|
|
PCI_DMA_FROMDEVICE, DEFAULT_TIMEOUT);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
if (rc)
|
|
|
|
|
goto out;
|
2009-12-08 22:09:11 +00:00
|
|
|
ei = c->err_info;
|
|
|
|
|
if (ei->CommandStatus != 0 && ei->CommandStatus != CMD_DATA_UNDERRUN) {
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_scsi_interpret_error(h, c);
|
2009-12-08 22:09:11 +00:00
|
|
|
rc = -1;
|
|
|
|
|
}
|
2013-02-20 17:24:41 +00:00
|
|
|
out:
|
2015-01-23 22:43:04 +00:00
|
|
|
cmd_free(h, c);
|
2009-12-08 22:09:11 +00:00
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:56:39 +00:00
|
|
|
static int hpsa_send_reset(struct ctlr_info *h, unsigned char *scsi3addr,
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
u8 reset_type, int reply_queue)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
|
|
|
|
int rc = IO_OK;
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
struct ErrorInfo *ei;
|
|
|
|
|
|
2015-01-23 22:43:04 +00:00
|
|
|
c = cmd_alloc(h);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
|
2013-02-20 17:24:41 +00:00
|
|
|
/* fill_cmd can't fail here, no data buffer to map. */
|
2015-11-04 21:51:02 +00:00
|
|
|
(void) fill_cmd(c, reset_type, h, NULL, 0, 0,
|
2014-02-18 19:56:39 +00:00
|
|
|
scsi3addr, TYPE_MSG);
|
2016-04-27 22:13:51 +00:00
|
|
|
rc = hpsa_scsi_do_simple_cmd(h, c, reply_queue, DEFAULT_TIMEOUT);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
if (rc) {
|
|
|
|
|
dev_warn(&h->pdev->dev, "Failed to send reset command\n");
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
/* no unmap needed here because no data xfer. */
|
|
|
|
|
|
|
|
|
|
ei = c->err_info;
|
|
|
|
|
if (ei->CommandStatus != 0) {
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_scsi_interpret_error(h, c);
|
2009-12-08 22:09:11 +00:00
|
|
|
rc = -1;
|
|
|
|
|
}
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
out:
|
2015-01-23 22:43:04 +00:00
|
|
|
cmd_free(h, c);
|
2009-12-08 22:09:11 +00:00
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:35:22 +00:00
|
|
|
static bool hpsa_cmd_dev_match(struct ctlr_info *h, struct CommandList *c,
|
|
|
|
|
struct hpsa_scsi_dev_t *dev,
|
|
|
|
|
unsigned char *scsi3addr)
|
|
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
bool match = false;
|
|
|
|
|
struct io_accel2_cmd *c2 = &h->ioaccel2_cmd_pool[c->cmdindex];
|
|
|
|
|
struct hpsa_tmf_struct *ac = (struct hpsa_tmf_struct *) c2;
|
|
|
|
|
|
|
|
|
|
if (hpsa_is_cmd_idle(c))
|
|
|
|
|
return false;
|
|
|
|
|
|
|
|
|
|
switch (c->cmd_type) {
|
|
|
|
|
case CMD_SCSI:
|
|
|
|
|
case CMD_IOCTL_PEND:
|
|
|
|
|
match = !memcmp(scsi3addr, &c->Header.LUN.LunAddrBytes,
|
|
|
|
|
sizeof(c->Header.LUN.LunAddrBytes));
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case CMD_IOACCEL1:
|
|
|
|
|
case CMD_IOACCEL2:
|
|
|
|
|
if (c->phys_disk == dev) {
|
|
|
|
|
/* HBA mode match */
|
|
|
|
|
match = true;
|
|
|
|
|
} else {
|
|
|
|
|
/* Possible RAID mode -- check each phys dev. */
|
|
|
|
|
/* FIXME: Do we need to take out a lock here? If
|
|
|
|
|
* so, we could just call hpsa_get_pdisk_of_ioaccel2()
|
|
|
|
|
* instead. */
|
|
|
|
|
for (i = 0; i < dev->nphysical_disks && !match; i++) {
|
|
|
|
|
/* FIXME: an alternate test might be
|
|
|
|
|
*
|
|
|
|
|
* match = dev->phys_disk[i]->ioaccel_handle
|
|
|
|
|
* == c2->scsi_nexus; */
|
|
|
|
|
match = dev->phys_disk[i] == c->phys_disk;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case IOACCEL2_TMF:
|
|
|
|
|
for (i = 0; i < dev->nphysical_disks && !match; i++) {
|
|
|
|
|
match = dev->phys_disk[i]->ioaccel_handle ==
|
|
|
|
|
le32_to_cpu(ac->it_nexus);
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case 0: /* The command is in the middle of being initialized. */
|
|
|
|
|
match = false;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
dev_err(&h->pdev->dev, "unexpected cmd_type: %d\n",
|
|
|
|
|
c->cmd_type);
|
|
|
|
|
BUG();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return match;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int hpsa_do_reset(struct ctlr_info *h, struct hpsa_scsi_dev_t *dev,
|
|
|
|
|
unsigned char *scsi3addr, u8 reset_type, int reply_queue)
|
|
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
int rc = 0;
|
|
|
|
|
|
|
|
|
|
/* We can really only handle one reset at a time */
|
|
|
|
|
if (mutex_lock_interruptible(&h->reset_mutex) == -EINTR) {
|
|
|
|
|
dev_warn(&h->pdev->dev, "concurrent reset wait interrupted.\n");
|
|
|
|
|
return -EINTR;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
BUG_ON(atomic_read(&dev->reset_cmds_out) != 0);
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < h->nr_cmds; i++) {
|
|
|
|
|
struct CommandList *c = h->cmd_pool + i;
|
|
|
|
|
int refcount = atomic_inc_return(&c->refcount);
|
|
|
|
|
|
|
|
|
|
if (refcount > 1 && hpsa_cmd_dev_match(h, c, dev, scsi3addr)) {
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Mark the target command as having a reset pending,
|
|
|
|
|
* then lock a lock so that the command cannot complete
|
|
|
|
|
* while we're considering it. If the command is not
|
|
|
|
|
* idle then count it; otherwise revoke the event.
|
|
|
|
|
*/
|
|
|
|
|
c->reset_pending = dev;
|
|
|
|
|
spin_lock_irqsave(&h->lock, flags); /* Implied MB */
|
|
|
|
|
if (!hpsa_is_cmd_idle(c))
|
|
|
|
|
atomic_inc(&dev->reset_cmds_out);
|
|
|
|
|
else
|
|
|
|
|
c->reset_pending = NULL;
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
cmd_free(h, c);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
rc = hpsa_send_reset(h, scsi3addr, reset_type, reply_queue);
|
|
|
|
|
if (!rc)
|
|
|
|
|
wait_event(h->event_sync_wait_queue,
|
|
|
|
|
atomic_read(&dev->reset_cmds_out) == 0 ||
|
|
|
|
|
lockup_detected(h));
|
|
|
|
|
|
|
|
|
|
if (unlikely(lockup_detected(h))) {
|
2015-07-18 16:12:22 +00:00
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"Controller lockup detected during reset wait\n");
|
|
|
|
|
rc = -ENODEV;
|
|
|
|
|
}
|
2015-04-23 14:35:22 +00:00
|
|
|
|
|
|
|
|
if (unlikely(rc))
|
|
|
|
|
atomic_set(&dev->reset_cmds_out, 0);
|
|
|
|
|
|
|
|
|
|
mutex_unlock(&h->reset_mutex);
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
static void hpsa_get_raid_level(struct ctlr_info *h,
|
|
|
|
|
unsigned char *scsi3addr, unsigned char *raid_level)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
unsigned char *buf;
|
|
|
|
|
|
|
|
|
|
*raid_level = RAID_UNKNOWN;
|
|
|
|
|
buf = kzalloc(64, GFP_KERNEL);
|
|
|
|
|
if (!buf)
|
|
|
|
|
return;
|
2014-02-18 19:57:11 +00:00
|
|
|
rc = hpsa_scsi_do_inquiry(h, scsi3addr, VPD_PAGE | 0xC1, buf, 64);
|
2009-12-08 22:09:11 +00:00
|
|
|
if (rc == 0)
|
|
|
|
|
*raid_level = buf[8];
|
|
|
|
|
if (*raid_level > RAID_UNKNOWN)
|
|
|
|
|
*raid_level = RAID_UNKNOWN;
|
|
|
|
|
kfree(buf);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:55:33 +00:00
|
|
|
#define HPSA_MAP_DEBUG
|
|
|
|
|
#ifdef HPSA_MAP_DEBUG
|
|
|
|
|
static void hpsa_debug_map_buff(struct ctlr_info *h, int rc,
|
|
|
|
|
struct raid_map_data *map_buff)
|
|
|
|
|
{
|
|
|
|
|
struct raid_map_disk_data *dd = &map_buff->data[0];
|
|
|
|
|
int map, row, col;
|
|
|
|
|
u16 map_cnt, row_cnt, disks_per_row;
|
|
|
|
|
|
|
|
|
|
if (rc != 0)
|
|
|
|
|
return;
|
|
|
|
|
|
2014-02-18 19:57:52 +00:00
|
|
|
/* Show details only if debugging has been activated. */
|
|
|
|
|
if (h->raid_offload_debug < 2)
|
|
|
|
|
return;
|
|
|
|
|
|
2014-02-18 19:55:33 +00:00
|
|
|
dev_info(&h->pdev->dev, "structure_size = %u\n",
|
|
|
|
|
le32_to_cpu(map_buff->structure_size));
|
|
|
|
|
dev_info(&h->pdev->dev, "volume_blk_size = %u\n",
|
|
|
|
|
le32_to_cpu(map_buff->volume_blk_size));
|
|
|
|
|
dev_info(&h->pdev->dev, "volume_blk_cnt = 0x%llx\n",
|
|
|
|
|
le64_to_cpu(map_buff->volume_blk_cnt));
|
|
|
|
|
dev_info(&h->pdev->dev, "physicalBlockShift = %u\n",
|
|
|
|
|
map_buff->phys_blk_shift);
|
|
|
|
|
dev_info(&h->pdev->dev, "parity_rotation_shift = %u\n",
|
|
|
|
|
map_buff->parity_rotation_shift);
|
|
|
|
|
dev_info(&h->pdev->dev, "strip_size = %u\n",
|
|
|
|
|
le16_to_cpu(map_buff->strip_size));
|
|
|
|
|
dev_info(&h->pdev->dev, "disk_starting_blk = 0x%llx\n",
|
|
|
|
|
le64_to_cpu(map_buff->disk_starting_blk));
|
|
|
|
|
dev_info(&h->pdev->dev, "disk_blk_cnt = 0x%llx\n",
|
|
|
|
|
le64_to_cpu(map_buff->disk_blk_cnt));
|
|
|
|
|
dev_info(&h->pdev->dev, "data_disks_per_row = %u\n",
|
|
|
|
|
le16_to_cpu(map_buff->data_disks_per_row));
|
|
|
|
|
dev_info(&h->pdev->dev, "metadata_disks_per_row = %u\n",
|
|
|
|
|
le16_to_cpu(map_buff->metadata_disks_per_row));
|
|
|
|
|
dev_info(&h->pdev->dev, "row_cnt = %u\n",
|
|
|
|
|
le16_to_cpu(map_buff->row_cnt));
|
|
|
|
|
dev_info(&h->pdev->dev, "layout_map_count = %u\n",
|
|
|
|
|
le16_to_cpu(map_buff->layout_map_count));
|
2015-01-23 22:41:09 +00:00
|
|
|
dev_info(&h->pdev->dev, "flags = 0x%x\n",
|
2014-02-18 19:57:31 +00:00
|
|
|
le16_to_cpu(map_buff->flags));
|
2015-01-23 22:41:09 +00:00
|
|
|
dev_info(&h->pdev->dev, "encrypytion = %s\n",
|
|
|
|
|
le16_to_cpu(map_buff->flags) &
|
|
|
|
|
RAID_MAP_FLAG_ENCRYPT_ON ? "ON" : "OFF");
|
2014-02-18 19:57:31 +00:00
|
|
|
dev_info(&h->pdev->dev, "dekindex = %u\n",
|
|
|
|
|
le16_to_cpu(map_buff->dekindex));
|
2014-02-18 19:55:33 +00:00
|
|
|
map_cnt = le16_to_cpu(map_buff->layout_map_count);
|
|
|
|
|
for (map = 0; map < map_cnt; map++) {
|
|
|
|
|
dev_info(&h->pdev->dev, "Map%u:\n", map);
|
|
|
|
|
row_cnt = le16_to_cpu(map_buff->row_cnt);
|
|
|
|
|
for (row = 0; row < row_cnt; row++) {
|
|
|
|
|
dev_info(&h->pdev->dev, " Row%u:\n", row);
|
|
|
|
|
disks_per_row =
|
|
|
|
|
le16_to_cpu(map_buff->data_disks_per_row);
|
|
|
|
|
for (col = 0; col < disks_per_row; col++, dd++)
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
" D%02u: h=0x%04x xor=%u,%u\n",
|
|
|
|
|
col, dd->ioaccel_handle,
|
|
|
|
|
dd->xor_mult[0], dd->xor_mult[1]);
|
|
|
|
|
disks_per_row =
|
|
|
|
|
le16_to_cpu(map_buff->metadata_disks_per_row);
|
|
|
|
|
for (col = 0; col < disks_per_row; col++, dd++)
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
" M%02u: h=0x%04x xor=%u,%u\n",
|
|
|
|
|
col, dd->ioaccel_handle,
|
|
|
|
|
dd->xor_mult[0], dd->xor_mult[1]);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
#else
|
|
|
|
|
static void hpsa_debug_map_buff(__attribute__((unused)) struct ctlr_info *h,
|
|
|
|
|
__attribute__((unused)) int rc,
|
|
|
|
|
__attribute__((unused)) struct raid_map_data *map_buff)
|
|
|
|
|
{
|
|
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
static int hpsa_get_raid_map(struct ctlr_info *h,
|
|
|
|
|
unsigned char *scsi3addr, struct hpsa_scsi_dev_t *this_device)
|
|
|
|
|
{
|
|
|
|
|
int rc = 0;
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
struct ErrorInfo *ei;
|
|
|
|
|
|
2015-01-23 22:43:04 +00:00
|
|
|
c = cmd_alloc(h);
|
2015-04-23 14:33:38 +00:00
|
|
|
|
2014-02-18 19:55:33 +00:00
|
|
|
if (fill_cmd(c, HPSA_GET_RAID_MAP, h, &this_device->raid_map,
|
|
|
|
|
sizeof(this_device->raid_map), 0,
|
|
|
|
|
scsi3addr, TYPE_CMD)) {
|
2015-04-23 14:33:43 +00:00
|
|
|
dev_warn(&h->pdev->dev, "hpsa_get_raid_map fill_cmd failed\n");
|
|
|
|
|
cmd_free(h, c);
|
|
|
|
|
return -1;
|
2014-02-18 19:55:33 +00:00
|
|
|
}
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
rc = hpsa_scsi_do_simple_cmd_with_retry(h, c,
|
2016-04-27 22:13:51 +00:00
|
|
|
PCI_DMA_FROMDEVICE, DEFAULT_TIMEOUT);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
if (rc)
|
|
|
|
|
goto out;
|
2014-02-18 19:55:33 +00:00
|
|
|
ei = c->err_info;
|
|
|
|
|
if (ei->CommandStatus != 0 && ei->CommandStatus != CMD_DATA_UNDERRUN) {
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_scsi_interpret_error(h, c);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
rc = -1;
|
|
|
|
|
goto out;
|
2014-02-18 19:55:33 +00:00
|
|
|
}
|
2015-01-23 22:43:04 +00:00
|
|
|
cmd_free(h, c);
|
2014-02-18 19:55:33 +00:00
|
|
|
|
|
|
|
|
/* @todo in the future, dynamically allocate RAID map memory */
|
|
|
|
|
if (le32_to_cpu(this_device->raid_map.structure_size) >
|
|
|
|
|
sizeof(this_device->raid_map)) {
|
|
|
|
|
dev_warn(&h->pdev->dev, "RAID map size is too large!\n");
|
|
|
|
|
rc = -1;
|
|
|
|
|
}
|
|
|
|
|
hpsa_debug_map_buff(h, rc, &this_device->raid_map);
|
|
|
|
|
return rc;
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
out:
|
|
|
|
|
cmd_free(h, c);
|
|
|
|
|
return rc;
|
2014-02-18 19:55:33 +00:00
|
|
|
}
|
|
|
|
|
|
2015-11-04 21:52:34 +00:00
|
|
|
static int hpsa_bmic_sense_subsystem_information(struct ctlr_info *h,
|
|
|
|
|
unsigned char scsi3addr[], u16 bmic_device_index,
|
|
|
|
|
struct bmic_sense_subsystem_info *buf, size_t bufsize)
|
|
|
|
|
{
|
|
|
|
|
int rc = IO_OK;
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
struct ErrorInfo *ei;
|
|
|
|
|
|
|
|
|
|
c = cmd_alloc(h);
|
|
|
|
|
|
|
|
|
|
rc = fill_cmd(c, BMIC_SENSE_SUBSYSTEM_INFORMATION, h, buf, bufsize,
|
|
|
|
|
0, RAID_CTLR_LUNID, TYPE_CMD);
|
|
|
|
|
if (rc)
|
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
|
|
c->Request.CDB[2] = bmic_device_index & 0xff;
|
|
|
|
|
c->Request.CDB[9] = (bmic_device_index >> 8) & 0xff;
|
|
|
|
|
|
|
|
|
|
rc = hpsa_scsi_do_simple_cmd_with_retry(h, c,
|
2016-04-27 22:13:51 +00:00
|
|
|
PCI_DMA_FROMDEVICE, DEFAULT_TIMEOUT);
|
2015-11-04 21:52:34 +00:00
|
|
|
if (rc)
|
|
|
|
|
goto out;
|
|
|
|
|
ei = c->err_info;
|
|
|
|
|
if (ei->CommandStatus != 0 && ei->CommandStatus != CMD_DATA_UNDERRUN) {
|
|
|
|
|
hpsa_scsi_interpret_error(h, c);
|
|
|
|
|
rc = -1;
|
|
|
|
|
}
|
|
|
|
|
out:
|
|
|
|
|
cmd_free(h, c);
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-04 21:51:57 +00:00
|
|
|
static int hpsa_bmic_id_controller(struct ctlr_info *h,
|
|
|
|
|
struct bmic_identify_controller *buf, size_t bufsize)
|
|
|
|
|
{
|
|
|
|
|
int rc = IO_OK;
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
struct ErrorInfo *ei;
|
|
|
|
|
|
|
|
|
|
c = cmd_alloc(h);
|
|
|
|
|
|
|
|
|
|
rc = fill_cmd(c, BMIC_IDENTIFY_CONTROLLER, h, buf, bufsize,
|
|
|
|
|
0, RAID_CTLR_LUNID, TYPE_CMD);
|
|
|
|
|
if (rc)
|
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
|
|
rc = hpsa_scsi_do_simple_cmd_with_retry(h, c,
|
2016-04-27 22:13:51 +00:00
|
|
|
PCI_DMA_FROMDEVICE, DEFAULT_TIMEOUT);
|
2015-11-04 21:51:57 +00:00
|
|
|
if (rc)
|
|
|
|
|
goto out;
|
|
|
|
|
ei = c->err_info;
|
|
|
|
|
if (ei->CommandStatus != 0 && ei->CommandStatus != CMD_DATA_UNDERRUN) {
|
|
|
|
|
hpsa_scsi_interpret_error(h, c);
|
|
|
|
|
rc = -1;
|
|
|
|
|
}
|
|
|
|
|
out:
|
|
|
|
|
cmd_free(h, c);
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:43:30 +00:00
|
|
|
static int hpsa_bmic_id_physical_device(struct ctlr_info *h,
|
|
|
|
|
unsigned char scsi3addr[], u16 bmic_device_index,
|
|
|
|
|
struct bmic_identify_physical_device *buf, size_t bufsize)
|
|
|
|
|
{
|
|
|
|
|
int rc = IO_OK;
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
struct ErrorInfo *ei;
|
|
|
|
|
|
|
|
|
|
c = cmd_alloc(h);
|
|
|
|
|
rc = fill_cmd(c, BMIC_IDENTIFY_PHYSICAL_DEVICE, h, buf, bufsize,
|
|
|
|
|
0, RAID_CTLR_LUNID, TYPE_CMD);
|
|
|
|
|
if (rc)
|
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
|
|
c->Request.CDB[2] = bmic_device_index & 0xff;
|
|
|
|
|
c->Request.CDB[9] = (bmic_device_index >> 8) & 0xff;
|
|
|
|
|
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
hpsa_scsi_do_simple_cmd_with_retry(h, c, PCI_DMA_FROMDEVICE,
|
2016-04-27 22:13:51 +00:00
|
|
|
DEFAULT_TIMEOUT);
|
2015-01-23 22:43:30 +00:00
|
|
|
ei = c->err_info;
|
|
|
|
|
if (ei->CommandStatus != 0 && ei->CommandStatus != CMD_DATA_UNDERRUN) {
|
|
|
|
|
hpsa_scsi_interpret_error(h, c);
|
|
|
|
|
rc = -1;
|
|
|
|
|
}
|
|
|
|
|
out:
|
|
|
|
|
cmd_free(h, c);
|
2015-11-04 21:52:34 +00:00
|
|
|
|
2015-01-23 22:43:30 +00:00
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2015-12-22 16:36:48 +00:00
|
|
|
/*
|
|
|
|
|
* get enclosure information
|
|
|
|
|
* struct ReportExtendedLUNdata *rlep - Used for BMIC drive number
|
|
|
|
|
* struct hpsa_scsi_dev_t *encl_dev - device entry for enclosure
|
|
|
|
|
* Uses id_physical_device to determine the box_index.
|
|
|
|
|
*/
|
|
|
|
|
static void hpsa_get_enclosure_info(struct ctlr_info *h,
|
|
|
|
|
unsigned char *scsi3addr,
|
|
|
|
|
struct ReportExtendedLUNdata *rlep, int rle_index,
|
|
|
|
|
struct hpsa_scsi_dev_t *encl_dev)
|
|
|
|
|
{
|
|
|
|
|
int rc = -1;
|
|
|
|
|
struct CommandList *c = NULL;
|
|
|
|
|
struct ErrorInfo *ei = NULL;
|
|
|
|
|
struct bmic_sense_storage_box_params *bssbp = NULL;
|
|
|
|
|
struct bmic_identify_physical_device *id_phys = NULL;
|
|
|
|
|
struct ext_report_lun_entry *rle = &rlep->LUN[rle_index];
|
|
|
|
|
u16 bmic_device_index = 0;
|
|
|
|
|
|
|
|
|
|
bmic_device_index = GET_BMIC_DRIVE_NUMBER(&rle->lunid[0]);
|
|
|
|
|
|
2016-02-23 21:16:09 +00:00
|
|
|
if (bmic_device_index == 0xFF00 || MASKED_DEVICE(&rle->lunid[0])) {
|
|
|
|
|
rc = IO_OK;
|
2015-12-22 16:36:48 +00:00
|
|
|
goto out;
|
2016-02-23 21:16:09 +00:00
|
|
|
}
|
2015-12-22 16:36:48 +00:00
|
|
|
|
|
|
|
|
bssbp = kzalloc(sizeof(*bssbp), GFP_KERNEL);
|
|
|
|
|
if (!bssbp)
|
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
|
|
id_phys = kzalloc(sizeof(*id_phys), GFP_KERNEL);
|
|
|
|
|
if (!id_phys)
|
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
|
|
rc = hpsa_bmic_id_physical_device(h, scsi3addr, bmic_device_index,
|
|
|
|
|
id_phys, sizeof(*id_phys));
|
|
|
|
|
if (rc) {
|
|
|
|
|
dev_warn(&h->pdev->dev, "%s: id_phys failed %d bdi[0x%x]\n",
|
|
|
|
|
__func__, encl_dev->external, bmic_device_index);
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
c = cmd_alloc(h);
|
|
|
|
|
|
|
|
|
|
rc = fill_cmd(c, BMIC_SENSE_STORAGE_BOX_PARAMS, h, bssbp,
|
|
|
|
|
sizeof(*bssbp), 0, RAID_CTLR_LUNID, TYPE_CMD);
|
|
|
|
|
|
|
|
|
|
if (rc)
|
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
|
|
if (id_phys->phys_connector[1] == 'E')
|
|
|
|
|
c->Request.CDB[5] = id_phys->box_index;
|
|
|
|
|
else
|
|
|
|
|
c->Request.CDB[5] = 0;
|
|
|
|
|
|
|
|
|
|
rc = hpsa_scsi_do_simple_cmd_with_retry(h, c, PCI_DMA_FROMDEVICE,
|
2016-04-27 22:13:51 +00:00
|
|
|
DEFAULT_TIMEOUT);
|
2015-12-22 16:36:48 +00:00
|
|
|
if (rc)
|
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
|
|
ei = c->err_info;
|
|
|
|
|
if (ei->CommandStatus != 0 && ei->CommandStatus != CMD_DATA_UNDERRUN) {
|
|
|
|
|
rc = -1;
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
encl_dev->box[id_phys->active_path_number] = bssbp->phys_box_on_port;
|
|
|
|
|
memcpy(&encl_dev->phys_connector[id_phys->active_path_number],
|
|
|
|
|
bssbp->phys_connector, sizeof(bssbp->phys_connector));
|
|
|
|
|
|
|
|
|
|
rc = IO_OK;
|
|
|
|
|
out:
|
|
|
|
|
kfree(bssbp);
|
|
|
|
|
kfree(id_phys);
|
|
|
|
|
|
|
|
|
|
if (c)
|
|
|
|
|
cmd_free(h, c);
|
|
|
|
|
|
|
|
|
|
if (rc != IO_OK)
|
|
|
|
|
hpsa_show_dev_msg(KERN_INFO, h, encl_dev,
|
|
|
|
|
"Error, could not get enclosure information\n");
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-04 21:52:34 +00:00
|
|
|
static u64 hpsa_get_sas_address_from_report_physical(struct ctlr_info *h,
|
|
|
|
|
unsigned char *scsi3addr)
|
|
|
|
|
{
|
|
|
|
|
struct ReportExtendedLUNdata *physdev;
|
|
|
|
|
u32 nphysicals;
|
|
|
|
|
u64 sa = 0;
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
physdev = kzalloc(sizeof(*physdev), GFP_KERNEL);
|
|
|
|
|
if (!physdev)
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
if (hpsa_scsi_do_report_phys_luns(h, physdev, sizeof(*physdev))) {
|
|
|
|
|
dev_err(&h->pdev->dev, "report physical LUNs failed.\n");
|
|
|
|
|
kfree(physdev);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
nphysicals = get_unaligned_be32(physdev->LUNListLength) / 24;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < nphysicals; i++)
|
|
|
|
|
if (!memcmp(&physdev->LUN[i].lunid[0], scsi3addr, 8)) {
|
|
|
|
|
sa = get_unaligned_be64(&physdev->LUN[i].wwid[0]);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
kfree(physdev);
|
|
|
|
|
|
|
|
|
|
return sa;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_get_sas_address(struct ctlr_info *h, unsigned char *scsi3addr,
|
|
|
|
|
struct hpsa_scsi_dev_t *dev)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
u64 sa = 0;
|
|
|
|
|
|
|
|
|
|
if (is_hba_lunid(scsi3addr)) {
|
|
|
|
|
struct bmic_sense_subsystem_info *ssi;
|
|
|
|
|
|
|
|
|
|
ssi = kzalloc(sizeof(*ssi), GFP_KERNEL);
|
|
|
|
|
if (ssi == NULL) {
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"%s: out of memory\n", __func__);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
rc = hpsa_bmic_sense_subsystem_information(h,
|
|
|
|
|
scsi3addr, 0, ssi, sizeof(*ssi));
|
|
|
|
|
if (rc == 0) {
|
|
|
|
|
sa = get_unaligned_be64(ssi->primary_world_wide_id);
|
|
|
|
|
h->sas_address = sa;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
kfree(ssi);
|
|
|
|
|
} else
|
|
|
|
|
sa = hpsa_get_sas_address_from_report_physical(h, scsi3addr);
|
|
|
|
|
|
|
|
|
|
dev->sas_address = sa;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Get a device id from inquiry page 0x83 */
|
2014-02-18 19:57:16 +00:00
|
|
|
static int hpsa_vpd_page_supported(struct ctlr_info *h,
|
|
|
|
|
unsigned char scsi3addr[], u8 page)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
int i;
|
|
|
|
|
int pages;
|
|
|
|
|
unsigned char *buf, bufsize;
|
|
|
|
|
|
|
|
|
|
buf = kzalloc(256, GFP_KERNEL);
|
|
|
|
|
if (!buf)
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
/* Get the size of the page list first */
|
|
|
|
|
rc = hpsa_scsi_do_inquiry(h, scsi3addr,
|
|
|
|
|
VPD_PAGE | HPSA_VPD_SUPPORTED_PAGES,
|
|
|
|
|
buf, HPSA_VPD_HEADER_SZ);
|
|
|
|
|
if (rc != 0)
|
|
|
|
|
goto exit_unsupported;
|
|
|
|
|
pages = buf[3];
|
|
|
|
|
if ((pages + HPSA_VPD_HEADER_SZ) <= 255)
|
|
|
|
|
bufsize = pages + HPSA_VPD_HEADER_SZ;
|
|
|
|
|
else
|
|
|
|
|
bufsize = 255;
|
|
|
|
|
|
|
|
|
|
/* Get the whole VPD page list */
|
|
|
|
|
rc = hpsa_scsi_do_inquiry(h, scsi3addr,
|
|
|
|
|
VPD_PAGE | HPSA_VPD_SUPPORTED_PAGES,
|
|
|
|
|
buf, bufsize);
|
|
|
|
|
if (rc != 0)
|
|
|
|
|
goto exit_unsupported;
|
|
|
|
|
|
|
|
|
|
pages = buf[3];
|
|
|
|
|
for (i = 1; i <= pages; i++)
|
|
|
|
|
if (buf[3 + i] == page)
|
|
|
|
|
goto exit_supported;
|
|
|
|
|
exit_unsupported:
|
|
|
|
|
kfree(buf);
|
|
|
|
|
return 0;
|
|
|
|
|
exit_supported:
|
|
|
|
|
kfree(buf);
|
|
|
|
|
return 1;
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:55:33 +00:00
|
|
|
static void hpsa_get_ioaccel_status(struct ctlr_info *h,
|
|
|
|
|
unsigned char *scsi3addr, struct hpsa_scsi_dev_t *this_device)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
unsigned char *buf;
|
|
|
|
|
u8 ioaccel_status;
|
|
|
|
|
|
|
|
|
|
this_device->offload_config = 0;
|
|
|
|
|
this_device->offload_enabled = 0;
|
2015-04-23 14:31:47 +00:00
|
|
|
this_device->offload_to_be_enabled = 0;
|
2014-02-18 19:55:33 +00:00
|
|
|
|
|
|
|
|
buf = kzalloc(64, GFP_KERNEL);
|
|
|
|
|
if (!buf)
|
|
|
|
|
return;
|
2014-02-18 19:57:16 +00:00
|
|
|
if (!hpsa_vpd_page_supported(h, scsi3addr, HPSA_VPD_LV_IOACCEL_STATUS))
|
|
|
|
|
goto out;
|
2014-02-18 19:55:33 +00:00
|
|
|
rc = hpsa_scsi_do_inquiry(h, scsi3addr,
|
2014-02-18 19:57:11 +00:00
|
|
|
VPD_PAGE | HPSA_VPD_LV_IOACCEL_STATUS, buf, 64);
|
2014-02-18 19:55:33 +00:00
|
|
|
if (rc != 0)
|
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
|
|
#define IOACCEL_STATUS_BYTE 4
|
|
|
|
|
#define OFFLOAD_CONFIGURED_BIT 0x01
|
|
|
|
|
#define OFFLOAD_ENABLED_BIT 0x02
|
|
|
|
|
ioaccel_status = buf[IOACCEL_STATUS_BYTE];
|
|
|
|
|
this_device->offload_config =
|
|
|
|
|
!!(ioaccel_status & OFFLOAD_CONFIGURED_BIT);
|
|
|
|
|
if (this_device->offload_config) {
|
|
|
|
|
this_device->offload_enabled =
|
|
|
|
|
!!(ioaccel_status & OFFLOAD_ENABLED_BIT);
|
|
|
|
|
if (hpsa_get_raid_map(h, scsi3addr, this_device))
|
|
|
|
|
this_device->offload_enabled = 0;
|
|
|
|
|
}
|
2015-04-23 14:31:47 +00:00
|
|
|
this_device->offload_to_be_enabled = this_device->offload_enabled;
|
2014-02-18 19:55:33 +00:00
|
|
|
out:
|
|
|
|
|
kfree(buf);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/* Get the device id from inquiry page 0x83 */
|
|
|
|
|
static int hpsa_get_device_id(struct ctlr_info *h, unsigned char *scsi3addr,
|
2015-11-04 21:51:39 +00:00
|
|
|
unsigned char *device_id, int index, int buflen)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
unsigned char *buf;
|
|
|
|
|
|
|
|
|
|
if (buflen > 16)
|
|
|
|
|
buflen = 16;
|
|
|
|
|
buf = kzalloc(64, GFP_KERNEL);
|
|
|
|
|
if (!buf)
|
2014-05-29 15:54:20 +00:00
|
|
|
return -ENOMEM;
|
2014-02-18 19:57:11 +00:00
|
|
|
rc = hpsa_scsi_do_inquiry(h, scsi3addr, VPD_PAGE | 0x83, buf, 64);
|
2009-12-08 22:09:11 +00:00
|
|
|
if (rc == 0)
|
2015-11-04 21:51:39 +00:00
|
|
|
memcpy(device_id, &buf[index], buflen);
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
kfree(buf);
|
2015-11-04 21:51:39 +00:00
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
return rc != 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int hpsa_scsi_do_report_luns(struct ctlr_info *h, int logical,
|
2015-01-23 22:43:30 +00:00
|
|
|
void *buf, int bufsize,
|
2009-12-08 22:09:11 +00:00
|
|
|
int extended_response)
|
|
|
|
|
{
|
|
|
|
|
int rc = IO_OK;
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
unsigned char scsi3addr[8];
|
|
|
|
|
struct ErrorInfo *ei;
|
|
|
|
|
|
2015-01-23 22:43:04 +00:00
|
|
|
c = cmd_alloc(h);
|
2015-04-23 14:33:38 +00:00
|
|
|
|
2010-02-04 14:42:04 +00:00
|
|
|
/* address the controller */
|
|
|
|
|
memset(scsi3addr, 0, sizeof(scsi3addr));
|
2013-02-20 17:24:41 +00:00
|
|
|
if (fill_cmd(c, logical ? HPSA_REPORT_LOG : HPSA_REPORT_PHYS, h,
|
|
|
|
|
buf, bufsize, 0, scsi3addr, TYPE_CMD)) {
|
|
|
|
|
rc = -1;
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
if (extended_response)
|
|
|
|
|
c->Request.CDB[1] = extended_response;
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
rc = hpsa_scsi_do_simple_cmd_with_retry(h, c,
|
2016-04-27 22:13:51 +00:00
|
|
|
PCI_DMA_FROMDEVICE, DEFAULT_TIMEOUT);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
if (rc)
|
|
|
|
|
goto out;
|
2009-12-08 22:09:11 +00:00
|
|
|
ei = c->err_info;
|
|
|
|
|
if (ei->CommandStatus != 0 &&
|
|
|
|
|
ei->CommandStatus != CMD_DATA_UNDERRUN) {
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_scsi_interpret_error(h, c);
|
2009-12-08 22:09:11 +00:00
|
|
|
rc = -1;
|
2014-02-18 19:55:33 +00:00
|
|
|
} else {
|
2015-01-23 22:43:30 +00:00
|
|
|
struct ReportLUNdata *rld = buf;
|
|
|
|
|
|
|
|
|
|
if (rld->extended_response_flag != extended_response) {
|
2014-02-18 19:55:33 +00:00
|
|
|
dev_err(&h->pdev->dev,
|
|
|
|
|
"report luns requested format %u, got %u\n",
|
|
|
|
|
extended_response,
|
2015-01-23 22:43:30 +00:00
|
|
|
rld->extended_response_flag);
|
2014-02-18 19:55:33 +00:00
|
|
|
rc = -1;
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
2013-02-20 17:24:41 +00:00
|
|
|
out:
|
2015-01-23 22:43:04 +00:00
|
|
|
cmd_free(h, c);
|
2009-12-08 22:09:11 +00:00
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline int hpsa_scsi_do_report_phys_luns(struct ctlr_info *h,
|
2015-01-23 22:43:30 +00:00
|
|
|
struct ReportExtendedLUNdata *buf, int bufsize)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
2015-01-23 22:43:30 +00:00
|
|
|
return hpsa_scsi_do_report_luns(h, 0, buf, bufsize,
|
|
|
|
|
HPSA_REPORT_PHYS_EXTENDED);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline int hpsa_scsi_do_report_log_luns(struct ctlr_info *h,
|
|
|
|
|
struct ReportLUNdata *buf, int bufsize)
|
|
|
|
|
{
|
|
|
|
|
return hpsa_scsi_do_report_luns(h, 1, buf, bufsize, 0);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline void hpsa_set_bus_target_lun(struct hpsa_scsi_dev_t *device,
|
|
|
|
|
int bus, int target, int lun)
|
|
|
|
|
{
|
|
|
|
|
device->bus = bus;
|
|
|
|
|
device->target = target;
|
|
|
|
|
device->lun = lun;
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-21 22:25:00 +00:00
|
|
|
/* Use VPD inquiry to get details of volume status */
|
|
|
|
|
static int hpsa_get_volume_status(struct ctlr_info *h,
|
|
|
|
|
unsigned char scsi3addr[])
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
int status;
|
|
|
|
|
int size;
|
|
|
|
|
unsigned char *buf;
|
|
|
|
|
|
|
|
|
|
buf = kzalloc(64, GFP_KERNEL);
|
|
|
|
|
if (!buf)
|
|
|
|
|
return HPSA_VPD_LV_STATUS_UNSUPPORTED;
|
|
|
|
|
|
|
|
|
|
/* Does controller have VPD for logical volume status? */
|
2014-05-29 15:54:10 +00:00
|
|
|
if (!hpsa_vpd_page_supported(h, scsi3addr, HPSA_VPD_LV_STATUS))
|
2014-02-21 22:25:00 +00:00
|
|
|
goto exit_failed;
|
|
|
|
|
|
|
|
|
|
/* Get the size of the VPD return buffer */
|
|
|
|
|
rc = hpsa_scsi_do_inquiry(h, scsi3addr, VPD_PAGE | HPSA_VPD_LV_STATUS,
|
|
|
|
|
buf, HPSA_VPD_HEADER_SZ);
|
2014-05-29 15:54:10 +00:00
|
|
|
if (rc != 0)
|
2014-02-21 22:25:00 +00:00
|
|
|
goto exit_failed;
|
|
|
|
|
size = buf[3];
|
|
|
|
|
|
|
|
|
|
/* Now get the whole VPD buffer */
|
|
|
|
|
rc = hpsa_scsi_do_inquiry(h, scsi3addr, VPD_PAGE | HPSA_VPD_LV_STATUS,
|
|
|
|
|
buf, size + HPSA_VPD_HEADER_SZ);
|
2014-05-29 15:54:10 +00:00
|
|
|
if (rc != 0)
|
2014-02-21 22:25:00 +00:00
|
|
|
goto exit_failed;
|
|
|
|
|
status = buf[4]; /* status byte */
|
|
|
|
|
|
|
|
|
|
kfree(buf);
|
|
|
|
|
return status;
|
|
|
|
|
exit_failed:
|
|
|
|
|
kfree(buf);
|
|
|
|
|
return HPSA_VPD_LV_STATUS_UNSUPPORTED;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Determine offline status of a volume.
|
|
|
|
|
* Return either:
|
|
|
|
|
* 0 (not offline)
|
2014-05-29 15:54:25 +00:00
|
|
|
* 0xff (offline for unknown reasons)
|
2014-02-21 22:25:00 +00:00
|
|
|
* # (integer code indicating one of several NOT READY states
|
|
|
|
|
* describing why a volume is to be kept offline)
|
|
|
|
|
*/
|
2014-05-29 15:54:25 +00:00
|
|
|
static int hpsa_volume_offline(struct ctlr_info *h,
|
2014-02-21 22:25:00 +00:00
|
|
|
unsigned char scsi3addr[])
|
|
|
|
|
{
|
|
|
|
|
struct CommandList *c;
|
2015-04-23 14:32:16 +00:00
|
|
|
unsigned char *sense;
|
|
|
|
|
u8 sense_key, asc, ascq;
|
|
|
|
|
int sense_len;
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
int rc, ldstat = 0;
|
2014-02-21 22:25:00 +00:00
|
|
|
u16 cmd_status;
|
|
|
|
|
u8 scsi_status;
|
|
|
|
|
#define ASC_LUN_NOT_READY 0x04
|
|
|
|
|
#define ASCQ_LUN_NOT_READY_FORMAT_IN_PROGRESS 0x04
|
|
|
|
|
#define ASCQ_LUN_NOT_READY_INITIALIZING_CMD_REQ 0x02
|
|
|
|
|
|
|
|
|
|
c = cmd_alloc(h);
|
2015-04-23 14:33:38 +00:00
|
|
|
|
2014-02-21 22:25:00 +00:00
|
|
|
(void) fill_cmd(c, TEST_UNIT_READY, h, NULL, 0, 0, scsi3addr, TYPE_CMD);
|
2016-04-27 22:13:51 +00:00
|
|
|
rc = hpsa_scsi_do_simple_cmd(h, c, DEFAULT_REPLY_QUEUE,
|
|
|
|
|
DEFAULT_TIMEOUT);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
if (rc) {
|
|
|
|
|
cmd_free(h, c);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
2014-02-21 22:25:00 +00:00
|
|
|
sense = c->err_info->SenseInfo;
|
2015-04-23 14:32:16 +00:00
|
|
|
if (c->err_info->SenseLen > sizeof(c->err_info->SenseInfo))
|
|
|
|
|
sense_len = sizeof(c->err_info->SenseInfo);
|
|
|
|
|
else
|
|
|
|
|
sense_len = c->err_info->SenseLen;
|
|
|
|
|
decode_sense_data(sense, sense_len, &sense_key, &asc, &ascq);
|
2014-02-21 22:25:00 +00:00
|
|
|
cmd_status = c->err_info->CommandStatus;
|
|
|
|
|
scsi_status = c->err_info->ScsiStatus;
|
|
|
|
|
cmd_free(h, c);
|
|
|
|
|
/* Is the volume 'not ready'? */
|
|
|
|
|
if (cmd_status != CMD_TARGET_STATUS ||
|
|
|
|
|
scsi_status != SAM_STAT_CHECK_CONDITION ||
|
|
|
|
|
sense_key != NOT_READY ||
|
|
|
|
|
asc != ASC_LUN_NOT_READY) {
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Determine the reason for not ready state */
|
|
|
|
|
ldstat = hpsa_get_volume_status(h, scsi3addr);
|
|
|
|
|
|
|
|
|
|
/* Keep volume offline in certain cases: */
|
|
|
|
|
switch (ldstat) {
|
|
|
|
|
case HPSA_LV_UNDERGOING_ERASE:
|
2015-07-18 16:13:04 +00:00
|
|
|
case HPSA_LV_NOT_AVAILABLE:
|
2014-02-21 22:25:00 +00:00
|
|
|
case HPSA_LV_UNDERGOING_RPI:
|
|
|
|
|
case HPSA_LV_PENDING_RPI:
|
|
|
|
|
case HPSA_LV_ENCRYPTED_NO_KEY:
|
|
|
|
|
case HPSA_LV_PLAINTEXT_IN_ENCRYPT_ONLY_CONTROLLER:
|
|
|
|
|
case HPSA_LV_UNDERGOING_ENCRYPTION:
|
|
|
|
|
case HPSA_LV_UNDERGOING_ENCRYPTION_REKEYING:
|
|
|
|
|
case HPSA_LV_ENCRYPTED_IN_NON_ENCRYPTED_CONTROLLER:
|
|
|
|
|
return ldstat;
|
|
|
|
|
case HPSA_VPD_LV_STATUS_UNSUPPORTED:
|
|
|
|
|
/* If VPD status page isn't available,
|
|
|
|
|
* use ASC/ASCQ to determine state
|
|
|
|
|
*/
|
|
|
|
|
if ((ascq == ASCQ_LUN_NOT_READY_FORMAT_IN_PROGRESS) ||
|
|
|
|
|
(ascq == ASCQ_LUN_NOT_READY_INITIALIZING_CMD_REQ))
|
|
|
|
|
return ldstat;
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:32:06 +00:00
|
|
|
/*
|
|
|
|
|
* Find out if a logical device supports aborts by simply trying one.
|
|
|
|
|
* Smart Array may claim not to support aborts on logical drives, but
|
|
|
|
|
* if a MSA2000 * is connected, the drives on that will be presented
|
|
|
|
|
* by the Smart Array as logical drives, and aborts may be sent to
|
|
|
|
|
* those devices successfully. So the simplest way to find out is
|
|
|
|
|
* to simply try an abort and see how the device responds.
|
|
|
|
|
*/
|
|
|
|
|
static int hpsa_device_supports_aborts(struct ctlr_info *h,
|
|
|
|
|
unsigned char *scsi3addr)
|
|
|
|
|
{
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
struct ErrorInfo *ei;
|
|
|
|
|
int rc = 0;
|
|
|
|
|
|
|
|
|
|
u64 tag = (u64) -1; /* bogus tag */
|
|
|
|
|
|
|
|
|
|
/* Assume that physical devices support aborts */
|
|
|
|
|
if (!is_logical_dev_addr_mode(scsi3addr))
|
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
|
|
c = cmd_alloc(h);
|
2015-04-23 14:33:38 +00:00
|
|
|
|
2015-04-23 14:32:06 +00:00
|
|
|
(void) fill_cmd(c, HPSA_ABORT_MSG, h, &tag, 0, 0, scsi3addr, TYPE_MSG);
|
2016-04-27 22:13:51 +00:00
|
|
|
(void) hpsa_scsi_do_simple_cmd(h, c, DEFAULT_REPLY_QUEUE,
|
|
|
|
|
DEFAULT_TIMEOUT);
|
2015-04-23 14:32:06 +00:00
|
|
|
/* no unmap needed here because no data xfer. */
|
|
|
|
|
ei = c->err_info;
|
|
|
|
|
switch (ei->CommandStatus) {
|
|
|
|
|
case CMD_INVALID:
|
|
|
|
|
rc = 0;
|
|
|
|
|
break;
|
|
|
|
|
case CMD_UNABORTABLE:
|
|
|
|
|
case CMD_ABORT_FAILED:
|
|
|
|
|
rc = 1;
|
|
|
|
|
break;
|
2015-04-23 14:32:16 +00:00
|
|
|
case CMD_TMF_STATUS:
|
|
|
|
|
rc = hpsa_evaluate_tmf_status(h, c);
|
|
|
|
|
break;
|
2015-04-23 14:32:06 +00:00
|
|
|
default:
|
|
|
|
|
rc = 0;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
cmd_free(h, c);
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
static int hpsa_update_device_info(struct ctlr_info *h,
|
2011-08-09 13:17:30 +00:00
|
|
|
unsigned char scsi3addr[], struct hpsa_scsi_dev_t *this_device,
|
|
|
|
|
unsigned char *is_OBDR_device)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
2011-08-09 13:17:30 +00:00
|
|
|
|
|
|
|
|
#define OBDR_SIG_OFFSET 43
|
|
|
|
|
#define OBDR_TAPE_SIG "$DR-10"
|
|
|
|
|
#define OBDR_SIG_LEN (sizeof(OBDR_TAPE_SIG) - 1)
|
|
|
|
|
#define OBDR_TAPE_INQ_SIZE (OBDR_SIG_OFFSET + OBDR_SIG_LEN)
|
|
|
|
|
|
2010-02-04 14:42:09 +00:00
|
|
|
unsigned char *inq_buff;
|
2011-08-09 13:17:30 +00:00
|
|
|
unsigned char *obdr_sig;
|
2015-11-04 21:50:44 +00:00
|
|
|
int rc = 0;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2010-02-04 14:42:09 +00:00
|
|
|
inq_buff = kzalloc(OBDR_TAPE_INQ_SIZE, GFP_KERNEL);
|
2015-11-04 21:50:44 +00:00
|
|
|
if (!inq_buff) {
|
|
|
|
|
rc = -ENOMEM;
|
2009-12-08 22:09:11 +00:00
|
|
|
goto bail_out;
|
2015-11-04 21:50:44 +00:00
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
/* Do an inquiry to the device to see what it is. */
|
|
|
|
|
if (hpsa_scsi_do_inquiry(h, scsi3addr, 0, inq_buff,
|
|
|
|
|
(unsigned char) OBDR_TAPE_INQ_SIZE) != 0) {
|
|
|
|
|
/* Inquiry failed (msg printed already) */
|
|
|
|
|
dev_err(&h->pdev->dev,
|
|
|
|
|
"hpsa_update_device_info: inquiry failed\n");
|
2015-11-04 21:50:44 +00:00
|
|
|
rc = -EIO;
|
2009-12-08 22:09:11 +00:00
|
|
|
goto bail_out;
|
|
|
|
|
}
|
|
|
|
|
|
2016-02-23 21:16:40 +00:00
|
|
|
scsi_sanitize_inquiry_string(&inq_buff[8], 8);
|
|
|
|
|
scsi_sanitize_inquiry_string(&inq_buff[16], 16);
|
2015-11-04 21:51:39 +00:00
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
this_device->devtype = (inq_buff[0] & 0x1f);
|
|
|
|
|
memcpy(this_device->scsi3addr, scsi3addr, 8);
|
|
|
|
|
memcpy(this_device->vendor, &inq_buff[8],
|
|
|
|
|
sizeof(this_device->vendor));
|
|
|
|
|
memcpy(this_device->model, &inq_buff[16],
|
|
|
|
|
sizeof(this_device->model));
|
|
|
|
|
memset(this_device->device_id, 0,
|
|
|
|
|
sizeof(this_device->device_id));
|
2015-11-04 21:51:39 +00:00
|
|
|
hpsa_get_device_id(h, scsi3addr, this_device->device_id, 8,
|
2009-12-08 22:09:11 +00:00
|
|
|
sizeof(this_device->device_id));
|
|
|
|
|
|
2016-02-23 21:16:15 +00:00
|
|
|
if ((this_device->devtype == TYPE_DISK ||
|
|
|
|
|
this_device->devtype == TYPE_ZBC) &&
|
2014-02-18 19:55:33 +00:00
|
|
|
is_logical_dev_addr_mode(scsi3addr)) {
|
2014-05-29 15:54:25 +00:00
|
|
|
int volume_offline;
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
hpsa_get_raid_level(h, scsi3addr, &this_device->raid_level);
|
2014-02-18 19:55:33 +00:00
|
|
|
if (h->fw_support & MISC_FW_RAID_OFFLOAD_BASIC)
|
|
|
|
|
hpsa_get_ioaccel_status(h, scsi3addr, this_device);
|
2014-05-29 15:54:25 +00:00
|
|
|
volume_offline = hpsa_volume_offline(h, scsi3addr);
|
|
|
|
|
if (volume_offline < 0 || volume_offline > 0xff)
|
|
|
|
|
volume_offline = HPSA_VPD_LV_STATUS_UNSUPPORTED;
|
|
|
|
|
this_device->volume_offline = volume_offline & 0xff;
|
2014-02-18 19:55:33 +00:00
|
|
|
} else {
|
2009-12-08 22:09:11 +00:00
|
|
|
this_device->raid_level = RAID_UNKNOWN;
|
2014-02-18 19:55:33 +00:00
|
|
|
this_device->offload_config = 0;
|
|
|
|
|
this_device->offload_enabled = 0;
|
2015-04-23 14:31:47 +00:00
|
|
|
this_device->offload_to_be_enabled = 0;
|
2015-04-23 14:32:59 +00:00
|
|
|
this_device->hba_ioaccel_enabled = 0;
|
2014-02-21 22:25:00 +00:00
|
|
|
this_device->volume_offline = 0;
|
2015-01-23 22:43:30 +00:00
|
|
|
this_device->queue_depth = h->nr_cmds;
|
2014-02-18 19:55:33 +00:00
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2011-08-09 13:17:30 +00:00
|
|
|
if (is_OBDR_device) {
|
|
|
|
|
/* See if this is a One-Button-Disaster-Recovery device
|
|
|
|
|
* by looking for "$DR-10" at offset 43 in inquiry data.
|
|
|
|
|
*/
|
|
|
|
|
obdr_sig = &inq_buff[OBDR_SIG_OFFSET];
|
|
|
|
|
*is_OBDR_device = (this_device->devtype == TYPE_ROM &&
|
|
|
|
|
strncmp(obdr_sig, OBDR_TAPE_SIG,
|
|
|
|
|
OBDR_SIG_LEN) == 0);
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
kfree(inq_buff);
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
bail_out:
|
|
|
|
|
kfree(inq_buff);
|
2015-11-04 21:50:44 +00:00
|
|
|
return rc;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:32:06 +00:00
|
|
|
static void hpsa_update_device_supports_aborts(struct ctlr_info *h,
|
|
|
|
|
struct hpsa_scsi_dev_t *dev, u8 *scsi3addr)
|
|
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
int rc, entry;
|
|
|
|
|
/*
|
|
|
|
|
* See if this device supports aborts. If we already know
|
|
|
|
|
* the device, we already know if it supports aborts, otherwise
|
|
|
|
|
* we have to find out if it supports aborts by trying one.
|
|
|
|
|
*/
|
|
|
|
|
spin_lock_irqsave(&h->devlock, flags);
|
|
|
|
|
rc = hpsa_scsi_find_entry(dev, h->dev, h->ndevices, &entry);
|
|
|
|
|
if ((rc == DEVICE_SAME || rc == DEVICE_UPDATED) &&
|
|
|
|
|
entry >= 0 && entry < h->ndevices) {
|
|
|
|
|
dev->supports_aborts = h->dev[entry]->supports_aborts;
|
|
|
|
|
spin_unlock_irqrestore(&h->devlock, flags);
|
|
|
|
|
} else {
|
|
|
|
|
spin_unlock_irqrestore(&h->devlock, flags);
|
|
|
|
|
dev->supports_aborts =
|
|
|
|
|
hpsa_device_supports_aborts(h, scsi3addr);
|
|
|
|
|
if (dev->supports_aborts < 0)
|
|
|
|
|
dev->supports_aborts = 0;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-04 21:51:45 +00:00
|
|
|
/*
|
|
|
|
|
* Helper function to assign bus, target, lun mapping of devices.
|
2009-12-08 22:09:11 +00:00
|
|
|
* Logical drive target and lun are assigned at this time, but
|
|
|
|
|
* physical device lun and target assignment are deferred (assigned
|
|
|
|
|
* in hpsa_find_target_lun, called by hpsa_scsi_add_entry.)
|
2015-11-04 21:51:45 +00:00
|
|
|
*/
|
2009-12-08 22:09:11 +00:00
|
|
|
static void figure_bus_target_lun(struct ctlr_info *h,
|
2012-01-19 20:01:14 +00:00
|
|
|
u8 *lunaddrbytes, struct hpsa_scsi_dev_t *device)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
2015-11-04 21:51:45 +00:00
|
|
|
u32 lunid = get_unaligned_le32(lunaddrbytes);
|
2012-01-19 20:01:14 +00:00
|
|
|
|
|
|
|
|
if (!is_logical_dev_addr_mode(lunaddrbytes)) {
|
|
|
|
|
/* physical device, target and lun filled in later */
|
2009-12-08 22:09:11 +00:00
|
|
|
if (is_hba_lunid(lunaddrbytes))
|
2015-11-04 21:51:45 +00:00
|
|
|
hpsa_set_bus_target_lun(device,
|
|
|
|
|
HPSA_HBA_BUS, 0, lunid & 0x3fff);
|
2009-12-08 22:09:11 +00:00
|
|
|
else
|
2012-01-19 20:01:14 +00:00
|
|
|
/* defer target, lun assignment for physical devices */
|
2015-11-04 21:51:45 +00:00
|
|
|
hpsa_set_bus_target_lun(device,
|
|
|
|
|
HPSA_PHYSICAL_DEVICE_BUS, -1, -1);
|
2012-01-19 20:01:14 +00:00
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
/* It's a logical device */
|
2015-11-04 21:51:57 +00:00
|
|
|
if (device->external) {
|
2012-01-19 20:01:14 +00:00
|
|
|
hpsa_set_bus_target_lun(device,
|
2015-11-04 21:51:45 +00:00
|
|
|
HPSA_EXTERNAL_RAID_VOLUME_BUS, (lunid >> 16) & 0x3fff,
|
|
|
|
|
lunid & 0x00ff);
|
2012-01-19 20:01:14 +00:00
|
|
|
return;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
2015-11-04 21:51:45 +00:00
|
|
|
hpsa_set_bus_target_lun(device, HPSA_RAID_VOLUME_BUS,
|
|
|
|
|
0, lunid & 0x3fff);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2014-02-18 19:56:45 +00:00
|
|
|
/*
|
|
|
|
|
* Get address of physical disk used for an ioaccel2 mode command:
|
|
|
|
|
* 1. Extract ioaccel2 handle from the command.
|
|
|
|
|
* 2. Find a matching ioaccel2 handle from list of physical disks.
|
|
|
|
|
* 3. Return:
|
|
|
|
|
* 1 and set scsi3addr to address of matching physical
|
|
|
|
|
* 0 if no matching physical disk was found.
|
|
|
|
|
*/
|
|
|
|
|
static int hpsa_get_pdisk_of_ioaccel2(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *ioaccel2_cmd_to_abort, unsigned char *scsi3addr)
|
|
|
|
|
{
|
2015-04-23 14:31:47 +00:00
|
|
|
struct io_accel2_cmd *c2 =
|
|
|
|
|
&h->ioaccel2_cmd_pool[ioaccel2_cmd_to_abort->cmdindex];
|
|
|
|
|
unsigned long flags;
|
2014-02-18 19:56:45 +00:00
|
|
|
int i;
|
|
|
|
|
|
2015-04-23 14:31:47 +00:00
|
|
|
spin_lock_irqsave(&h->devlock, flags);
|
|
|
|
|
for (i = 0; i < h->ndevices; i++)
|
|
|
|
|
if (h->dev[i]->ioaccel_handle == le32_to_cpu(c2->scsi_nexus)) {
|
|
|
|
|
memcpy(scsi3addr, h->dev[i]->scsi3addr,
|
|
|
|
|
sizeof(h->dev[i]->scsi3addr));
|
|
|
|
|
spin_unlock_irqrestore(&h->devlock, flags);
|
|
|
|
|
return 1;
|
|
|
|
|
}
|
|
|
|
|
spin_unlock_irqrestore(&h->devlock, flags);
|
|
|
|
|
return 0;
|
2014-02-18 19:56:45 +00:00
|
|
|
}
|
2015-04-23 14:31:47 +00:00
|
|
|
|
2015-11-04 21:51:57 +00:00
|
|
|
static int figure_external_status(struct ctlr_info *h, int raid_ctlr_position,
|
|
|
|
|
int i, int nphysicals, int nlocal_logicals)
|
|
|
|
|
{
|
|
|
|
|
/* In report logicals, local logicals are listed first,
|
|
|
|
|
* then any externals.
|
|
|
|
|
*/
|
|
|
|
|
int logicals_start = nphysicals + (raid_ctlr_position == 0);
|
|
|
|
|
|
|
|
|
|
if (i == raid_ctlr_position)
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
if (i < logicals_start)
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
/* i is in logicals range, but still within local logicals */
|
|
|
|
|
if ((i - nphysicals - (raid_ctlr_position == 0)) < nlocal_logicals)
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
return 1; /* it's an external lun */
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/*
|
|
|
|
|
* Do CISS_REPORT_PHYS and CISS_REPORT_LOG. Data is returned in physdev,
|
|
|
|
|
* logdev. The number of luns in physdev and logdev are returned in
|
|
|
|
|
* *nphysicals and *nlogicals, respectively.
|
|
|
|
|
* Returns 0 on success, -1 otherwise.
|
|
|
|
|
*/
|
|
|
|
|
static int hpsa_gather_lun_info(struct ctlr_info *h,
|
2015-01-23 22:43:30 +00:00
|
|
|
struct ReportExtendedLUNdata *physdev, u32 *nphysicals,
|
2010-02-04 14:41:33 +00:00
|
|
|
struct ReportLUNdata *logdev, u32 *nlogicals)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
2015-01-23 22:43:30 +00:00
|
|
|
if (hpsa_scsi_do_report_phys_luns(h, physdev, sizeof(*physdev))) {
|
2009-12-08 22:09:11 +00:00
|
|
|
dev_err(&h->pdev->dev, "report physical LUNs failed.\n");
|
|
|
|
|
return -1;
|
|
|
|
|
}
|
2015-01-23 22:43:30 +00:00
|
|
|
*nphysicals = be32_to_cpu(*((__be32 *)physdev->LUNListLength)) / 24;
|
2009-12-08 22:09:11 +00:00
|
|
|
if (*nphysicals > HPSA_MAX_PHYS_LUN) {
|
2015-01-23 22:43:30 +00:00
|
|
|
dev_warn(&h->pdev->dev, "maximum physical LUNs (%d) exceeded. %d LUNs ignored.\n",
|
|
|
|
|
HPSA_MAX_PHYS_LUN, *nphysicals - HPSA_MAX_PHYS_LUN);
|
2009-12-08 22:09:11 +00:00
|
|
|
*nphysicals = HPSA_MAX_PHYS_LUN;
|
|
|
|
|
}
|
2015-01-23 22:43:30 +00:00
|
|
|
if (hpsa_scsi_do_report_log_luns(h, logdev, sizeof(*logdev))) {
|
2009-12-08 22:09:11 +00:00
|
|
|
dev_err(&h->pdev->dev, "report logical LUNs failed.\n");
|
|
|
|
|
return -1;
|
|
|
|
|
}
|
2010-02-04 14:42:19 +00:00
|
|
|
*nlogicals = be32_to_cpu(*((__be32 *) logdev->LUNListLength)) / 8;
|
2009-12-08 22:09:11 +00:00
|
|
|
/* Reject Logicals in excess of our max capability. */
|
|
|
|
|
if (*nlogicals > HPSA_MAX_LUN) {
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"maximum logical LUNs (%d) exceeded. "
|
|
|
|
|
"%d LUNs ignored.\n", HPSA_MAX_LUN,
|
|
|
|
|
*nlogicals - HPSA_MAX_LUN);
|
|
|
|
|
*nlogicals = HPSA_MAX_LUN;
|
|
|
|
|
}
|
|
|
|
|
if (*nlogicals + *nphysicals > HPSA_MAX_PHYS_LUN) {
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"maximum logical + physical LUNs (%d) exceeded. "
|
|
|
|
|
"%d LUNs ignored.\n", HPSA_MAX_PHYS_LUN,
|
|
|
|
|
*nphysicals + *nlogicals - HPSA_MAX_PHYS_LUN);
|
|
|
|
|
*nlogicals = HPSA_MAX_PHYS_LUN - *nphysicals;
|
|
|
|
|
}
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2014-11-14 23:26:27 +00:00
|
|
|
static u8 *figure_lunaddrbytes(struct ctlr_info *h, int raid_ctlr_position,
|
|
|
|
|
int i, int nphysicals, int nlogicals,
|
2014-02-18 19:55:07 +00:00
|
|
|
struct ReportExtendedLUNdata *physdev_list,
|
2010-02-04 14:42:50 +00:00
|
|
|
struct ReportLUNdata *logdev_list)
|
|
|
|
|
{
|
|
|
|
|
/* Helper function, figure out where the LUN ID info is coming from
|
|
|
|
|
* given index i, lists of physical and logical devices, where in
|
|
|
|
|
* the list the raid controller is supposed to appear (first or last)
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
int logicals_start = nphysicals + (raid_ctlr_position == 0);
|
|
|
|
|
int last_device = nphysicals + nlogicals + (raid_ctlr_position == 0);
|
|
|
|
|
|
|
|
|
|
if (i == raid_ctlr_position)
|
|
|
|
|
return RAID_CTLR_LUNID;
|
|
|
|
|
|
|
|
|
|
if (i < logicals_start)
|
2014-05-29 15:53:34 +00:00
|
|
|
return &physdev_list->LUN[i -
|
|
|
|
|
(raid_ctlr_position == 0)].lunid[0];
|
2010-02-04 14:42:50 +00:00
|
|
|
|
|
|
|
|
if (i < last_device)
|
|
|
|
|
return &logdev_list->LUN[i - nphysicals -
|
|
|
|
|
(raid_ctlr_position == 0)][0];
|
|
|
|
|
BUG();
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:43:30 +00:00
|
|
|
/* get physical drive ioaccel handle and queue depth */
|
|
|
|
|
static void hpsa_get_ioaccel_drive_info(struct ctlr_info *h,
|
|
|
|
|
struct hpsa_scsi_dev_t *dev,
|
2015-11-04 21:51:08 +00:00
|
|
|
struct ReportExtendedLUNdata *rlep, int rle_index,
|
2015-01-23 22:43:30 +00:00
|
|
|
struct bmic_identify_physical_device *id_phys)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
2015-11-04 21:51:08 +00:00
|
|
|
struct ext_report_lun_entry *rle = &rlep->LUN[rle_index];
|
2015-01-23 22:43:30 +00:00
|
|
|
|
|
|
|
|
dev->ioaccel_handle = rle->ioaccel_handle;
|
2015-11-04 21:51:08 +00:00
|
|
|
if ((rle->device_flags & 0x08) && dev->ioaccel_handle)
|
2015-04-23 14:32:59 +00:00
|
|
|
dev->hba_ioaccel_enabled = 1;
|
2015-01-23 22:43:30 +00:00
|
|
|
memset(id_phys, 0, sizeof(*id_phys));
|
2015-11-04 21:51:08 +00:00
|
|
|
rc = hpsa_bmic_id_physical_device(h, &rle->lunid[0],
|
|
|
|
|
GET_BMIC_DRIVE_NUMBER(&rle->lunid[0]), id_phys,
|
2015-01-23 22:43:30 +00:00
|
|
|
sizeof(*id_phys));
|
|
|
|
|
if (!rc)
|
|
|
|
|
/* Reserve space for FW operations */
|
|
|
|
|
#define DRIVE_CMDS_RESERVED_FOR_FW 2
|
|
|
|
|
#define DRIVE_QUEUE_DEPTH 7
|
|
|
|
|
dev->queue_depth =
|
|
|
|
|
le16_to_cpu(id_phys->current_queue_depth_limit) -
|
|
|
|
|
DRIVE_CMDS_RESERVED_FOR_FW;
|
|
|
|
|
else
|
|
|
|
|
dev->queue_depth = DRIVE_QUEUE_DEPTH; /* conservative */
|
|
|
|
|
}
|
|
|
|
|
|
2015-07-18 16:12:43 +00:00
|
|
|
static void hpsa_get_path_info(struct hpsa_scsi_dev_t *this_device,
|
2015-11-04 21:51:08 +00:00
|
|
|
struct ReportExtendedLUNdata *rlep, int rle_index,
|
2015-07-18 16:12:43 +00:00
|
|
|
struct bmic_identify_physical_device *id_phys)
|
|
|
|
|
{
|
2015-11-04 21:51:08 +00:00
|
|
|
struct ext_report_lun_entry *rle = &rlep->LUN[rle_index];
|
|
|
|
|
|
|
|
|
|
if ((rle->device_flags & 0x08) && this_device->ioaccel_handle)
|
2015-07-18 16:12:43 +00:00
|
|
|
this_device->hba_ioaccel_enabled = 1;
|
|
|
|
|
|
|
|
|
|
memcpy(&this_device->active_path_index,
|
|
|
|
|
&id_phys->active_path_number,
|
|
|
|
|
sizeof(this_device->active_path_index));
|
|
|
|
|
memcpy(&this_device->path_map,
|
|
|
|
|
&id_phys->redundant_path_present_map,
|
|
|
|
|
sizeof(this_device->path_map));
|
|
|
|
|
memcpy(&this_device->box,
|
|
|
|
|
&id_phys->alternate_paths_phys_box_on_port,
|
|
|
|
|
sizeof(this_device->box));
|
|
|
|
|
memcpy(&this_device->phys_connector,
|
|
|
|
|
&id_phys->alternate_paths_phys_connector,
|
|
|
|
|
sizeof(this_device->phys_connector));
|
|
|
|
|
memcpy(&this_device->bay,
|
|
|
|
|
&id_phys->phys_bay_in_box,
|
|
|
|
|
sizeof(this_device->bay));
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-04 21:51:57 +00:00
|
|
|
/* get number of local logical disks. */
|
|
|
|
|
static int hpsa_set_local_logical_count(struct ctlr_info *h,
|
|
|
|
|
struct bmic_identify_controller *id_ctlr,
|
|
|
|
|
u32 *nlocals)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
|
|
if (!id_ctlr) {
|
|
|
|
|
dev_warn(&h->pdev->dev, "%s: id_ctlr buffer is NULL.\n",
|
|
|
|
|
__func__);
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
}
|
|
|
|
|
memset(id_ctlr, 0, sizeof(*id_ctlr));
|
|
|
|
|
rc = hpsa_bmic_id_controller(h, id_ctlr, sizeof(*id_ctlr));
|
|
|
|
|
if (!rc)
|
|
|
|
|
if (id_ctlr->configured_logical_drive_count < 256)
|
|
|
|
|
*nlocals = id_ctlr->configured_logical_drive_count;
|
|
|
|
|
else
|
|
|
|
|
*nlocals = le16_to_cpu(
|
|
|
|
|
id_ctlr->extended_logical_unit_count);
|
|
|
|
|
else
|
|
|
|
|
*nlocals = -1;
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2015-11-04 21:50:01 +00:00
|
|
|
static void hpsa_update_scsi_devices(struct ctlr_info *h)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
|
|
|
|
/* the idea here is we could get notified
|
|
|
|
|
* that some devices have changed, so we do a report
|
|
|
|
|
* physical luns and report logical luns cmd, and adjust
|
|
|
|
|
* our list of devices accordingly.
|
|
|
|
|
*
|
|
|
|
|
* The scsi3addr's of devices won't change so long as the
|
|
|
|
|
* adapter is not reset. That means we can rescan and
|
|
|
|
|
* tell which devices we already know about, vs. new
|
|
|
|
|
* devices, vs. disappearing devices.
|
|
|
|
|
*/
|
2014-02-18 19:55:07 +00:00
|
|
|
struct ReportExtendedLUNdata *physdev_list = NULL;
|
2009-12-08 22:09:11 +00:00
|
|
|
struct ReportLUNdata *logdev_list = NULL;
|
2015-01-23 22:43:30 +00:00
|
|
|
struct bmic_identify_physical_device *id_phys = NULL;
|
2015-11-04 21:51:57 +00:00
|
|
|
struct bmic_identify_controller *id_ctlr = NULL;
|
2010-02-04 14:41:33 +00:00
|
|
|
u32 nphysicals = 0;
|
|
|
|
|
u32 nlogicals = 0;
|
2015-11-04 21:51:57 +00:00
|
|
|
u32 nlocal_logicals = 0;
|
2010-02-04 14:41:33 +00:00
|
|
|
u32 ndev_allocated = 0;
|
2009-12-08 22:09:11 +00:00
|
|
|
struct hpsa_scsi_dev_t **currentsd, *this_device, *tmpdevice;
|
|
|
|
|
int ncurrent = 0;
|
2012-01-19 20:01:25 +00:00
|
|
|
int i, n_ext_target_devs, ndevs_to_allocate;
|
2010-02-04 14:42:50 +00:00
|
|
|
int raid_ctlr_position;
|
2015-11-04 21:51:27 +00:00
|
|
|
bool physical_device;
|
2012-01-19 20:01:19 +00:00
|
|
|
DECLARE_BITMAP(lunzerobits, MAX_EXT_TARGETS);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2011-10-26 21:21:07 +00:00
|
|
|
currentsd = kzalloc(sizeof(*currentsd) * HPSA_MAX_DEVICES, GFP_KERNEL);
|
2014-11-14 23:26:54 +00:00
|
|
|
physdev_list = kzalloc(sizeof(*physdev_list), GFP_KERNEL);
|
|
|
|
|
logdev_list = kzalloc(sizeof(*logdev_list), GFP_KERNEL);
|
2009-12-08 22:09:11 +00:00
|
|
|
tmpdevice = kzalloc(sizeof(*tmpdevice), GFP_KERNEL);
|
2015-01-23 22:43:30 +00:00
|
|
|
id_phys = kzalloc(sizeof(*id_phys), GFP_KERNEL);
|
2015-11-04 21:51:57 +00:00
|
|
|
id_ctlr = kzalloc(sizeof(*id_ctlr), GFP_KERNEL);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-01-23 22:43:30 +00:00
|
|
|
if (!currentsd || !physdev_list || !logdev_list ||
|
2015-11-04 21:51:57 +00:00
|
|
|
!tmpdevice || !id_phys || !id_ctlr) {
|
2009-12-08 22:09:11 +00:00
|
|
|
dev_err(&h->pdev->dev, "out of memory\n");
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
|
|
|
|
memset(lunzerobits, 0, sizeof(lunzerobits));
|
|
|
|
|
|
2015-11-04 21:50:37 +00:00
|
|
|
h->drv_req_rescan = 0; /* cancel scheduled rescan - we're doing it. */
|
|
|
|
|
|
2015-01-23 22:43:30 +00:00
|
|
|
if (hpsa_gather_lun_info(h, physdev_list, &nphysicals,
|
2015-11-04 21:50:37 +00:00
|
|
|
logdev_list, &nlogicals)) {
|
|
|
|
|
h->drv_req_rescan = 1;
|
2009-12-08 22:09:11 +00:00
|
|
|
goto out;
|
2015-11-04 21:50:37 +00:00
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-11-04 21:51:57 +00:00
|
|
|
/* Set number of local logicals (non PTRAID) */
|
|
|
|
|
if (hpsa_set_local_logical_count(h, id_ctlr, &nlocal_logicals)) {
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"%s: Can't determine number of local logical devices.\n",
|
|
|
|
|
__func__);
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2012-01-19 20:01:19 +00:00
|
|
|
/* We might see up to the maximum number of logical and physical disks
|
|
|
|
|
* plus external target devices, and a device for the local RAID
|
|
|
|
|
* controller.
|
2009-12-08 22:09:11 +00:00
|
|
|
*/
|
2012-01-19 20:01:19 +00:00
|
|
|
ndevs_to_allocate = nphysicals + nlogicals + MAX_EXT_TARGETS + 1;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
/* Allocate the per device structures */
|
|
|
|
|
for (i = 0; i < ndevs_to_allocate; i++) {
|
2011-10-26 21:21:12 +00:00
|
|
|
if (i >= HPSA_MAX_DEVICES) {
|
|
|
|
|
dev_warn(&h->pdev->dev, "maximum devices (%d) exceeded."
|
|
|
|
|
" %d devices ignored.\n", HPSA_MAX_DEVICES,
|
|
|
|
|
ndevs_to_allocate - HPSA_MAX_DEVICES);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
currentsd[i] = kzalloc(sizeof(*currentsd[i]), GFP_KERNEL);
|
|
|
|
|
if (!currentsd[i]) {
|
|
|
|
|
dev_warn(&h->pdev->dev, "out of memory at %s:%d\n",
|
|
|
|
|
__FILE__, __LINE__);
|
2015-11-04 21:50:37 +00:00
|
|
|
h->drv_req_rescan = 1;
|
2009-12-08 22:09:11 +00:00
|
|
|
goto out;
|
|
|
|
|
}
|
|
|
|
|
ndev_allocated++;
|
|
|
|
|
}
|
|
|
|
|
|
2014-05-29 15:53:49 +00:00
|
|
|
if (is_scsi_rev_5(h))
|
2010-02-04 14:42:50 +00:00
|
|
|
raid_ctlr_position = 0;
|
|
|
|
|
else
|
|
|
|
|
raid_ctlr_position = nphysicals + nlogicals;
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/* adjust our table of devices */
|
2012-01-19 20:01:25 +00:00
|
|
|
n_ext_target_devs = 0;
|
2009-12-08 22:09:11 +00:00
|
|
|
for (i = 0; i < nphysicals + nlogicals + 1; i++) {
|
2011-08-09 13:17:30 +00:00
|
|
|
u8 *lunaddrbytes, is_OBDR = 0;
|
2015-11-04 21:50:44 +00:00
|
|
|
int rc = 0;
|
2015-11-04 21:51:08 +00:00
|
|
|
int phys_dev_index = i - (raid_ctlr_position == 0);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-11-04 21:51:27 +00:00
|
|
|
physical_device = i < nphysicals + (raid_ctlr_position == 0);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
/* Figure out where the LUN ID info is coming from */
|
2010-02-04 14:42:50 +00:00
|
|
|
lunaddrbytes = figure_lunaddrbytes(h, raid_ctlr_position,
|
|
|
|
|
i, nphysicals, nlogicals, physdev_list, logdev_list);
|
2015-04-23 14:31:47 +00:00
|
|
|
|
|
|
|
|
/* skip masked non-disk devices */
|
2015-11-04 21:51:27 +00:00
|
|
|
if (MASKED_DEVICE(lunaddrbytes) && physical_device &&
|
2015-12-22 16:36:48 +00:00
|
|
|
(physdev_list->LUN[phys_dev_index].device_type != 0x06) &&
|
|
|
|
|
(physdev_list->LUN[phys_dev_index].device_flags & 0x01))
|
2015-11-04 21:51:27 +00:00
|
|
|
continue;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
/* Get device type, vendor, model, device id */
|
2015-11-04 21:50:44 +00:00
|
|
|
rc = hpsa_update_device_info(h, lunaddrbytes, tmpdevice,
|
|
|
|
|
&is_OBDR);
|
|
|
|
|
if (rc == -ENOMEM) {
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"Out of memory, rescan deferred.\n");
|
2015-11-04 21:50:37 +00:00
|
|
|
h->drv_req_rescan = 1;
|
2015-11-04 21:50:44 +00:00
|
|
|
goto out;
|
2015-11-04 21:50:37 +00:00
|
|
|
}
|
2015-11-04 21:50:44 +00:00
|
|
|
if (rc) {
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"Inquiry failed, skipping device.\n");
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-04 21:51:57 +00:00
|
|
|
/* Determine if this is a lun from an external target array */
|
|
|
|
|
tmpdevice->external =
|
|
|
|
|
figure_external_status(h, raid_ctlr_position, i,
|
|
|
|
|
nphysicals, nlocal_logicals);
|
|
|
|
|
|
2012-01-19 20:01:14 +00:00
|
|
|
figure_bus_target_lun(h, lunaddrbytes, tmpdevice);
|
2015-04-23 14:32:06 +00:00
|
|
|
hpsa_update_device_supports_aborts(h, tmpdevice, lunaddrbytes);
|
2009-12-08 22:09:11 +00:00
|
|
|
this_device = currentsd[ncurrent];
|
|
|
|
|
|
2015-11-04 21:52:09 +00:00
|
|
|
/* Turn on discovery_polling if there are ext target devices.
|
|
|
|
|
* Event-based change notification is unreliable for those.
|
2009-12-08 22:09:11 +00:00
|
|
|
*/
|
2015-11-04 21:52:09 +00:00
|
|
|
if (!h->discovery_polling) {
|
|
|
|
|
if (tmpdevice->external) {
|
|
|
|
|
h->discovery_polling = 1;
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"External target, activate discovery polling.\n");
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
2015-11-04 21:52:09 +00:00
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
*this_device = *tmpdevice;
|
2015-11-04 21:51:27 +00:00
|
|
|
this_device->physical_device = physical_device;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-11-04 21:51:27 +00:00
|
|
|
/*
|
|
|
|
|
* Expose all devices except for physical devices that
|
|
|
|
|
* are masked.
|
|
|
|
|
*/
|
|
|
|
|
if (MASKED_DEVICE(lunaddrbytes) && this_device->physical_device)
|
2015-11-04 21:51:21 +00:00
|
|
|
this_device->expose_device = 0;
|
|
|
|
|
else
|
|
|
|
|
this_device->expose_device = 1;
|
2015-04-23 14:31:47 +00:00
|
|
|
|
2015-11-04 21:52:34 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Get the SAS address for physical devices that are exposed.
|
|
|
|
|
*/
|
|
|
|
|
if (this_device->physical_device && this_device->expose_device)
|
|
|
|
|
hpsa_get_sas_address(h, lunaddrbytes, this_device);
|
2015-04-23 14:31:47 +00:00
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
switch (this_device->devtype) {
|
2011-08-09 13:17:30 +00:00
|
|
|
case TYPE_ROM:
|
2009-12-08 22:09:11 +00:00
|
|
|
/* We don't *really* support actual CD-ROM devices,
|
|
|
|
|
* just "One Button Disaster Recovery" tape drive
|
|
|
|
|
* which temporarily pretends to be a CD-ROM drive.
|
|
|
|
|
* So we check that the device is really an OBDR tape
|
|
|
|
|
* device by checking for "$DR-10" in bytes 43-48 of
|
|
|
|
|
* the inquiry data.
|
|
|
|
|
*/
|
2011-08-09 13:17:30 +00:00
|
|
|
if (is_OBDR)
|
|
|
|
|
ncurrent++;
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case TYPE_DISK:
|
2016-02-23 21:16:15 +00:00
|
|
|
case TYPE_ZBC:
|
2015-11-04 21:51:27 +00:00
|
|
|
if (this_device->physical_device) {
|
2015-07-18 16:12:59 +00:00
|
|
|
/* The disk is in HBA mode. */
|
|
|
|
|
/* Never use RAID mapper in HBA mode. */
|
2015-04-23 14:33:04 +00:00
|
|
|
this_device->offload_enabled = 0;
|
2015-07-18 16:12:59 +00:00
|
|
|
hpsa_get_ioaccel_drive_info(h, this_device,
|
2015-11-04 21:51:08 +00:00
|
|
|
physdev_list, phys_dev_index, id_phys);
|
|
|
|
|
hpsa_get_path_info(this_device,
|
|
|
|
|
physdev_list, phys_dev_index, id_phys);
|
2015-07-18 16:12:59 +00:00
|
|
|
}
|
2015-04-23 14:33:04 +00:00
|
|
|
ncurrent++;
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case TYPE_TAPE:
|
|
|
|
|
case TYPE_MEDIUM_CHANGER:
|
2015-12-22 16:36:48 +00:00
|
|
|
ncurrent++;
|
|
|
|
|
break;
|
2015-04-23 14:31:47 +00:00
|
|
|
case TYPE_ENCLOSURE:
|
2016-02-23 21:16:09 +00:00
|
|
|
if (!this_device->external)
|
|
|
|
|
hpsa_get_enclosure_info(h, lunaddrbytes,
|
2015-12-22 16:36:48 +00:00
|
|
|
physdev_list, phys_dev_index,
|
|
|
|
|
this_device);
|
2015-07-18 16:12:59 +00:00
|
|
|
ncurrent++;
|
2015-04-23 14:31:47 +00:00
|
|
|
break;
|
2009-12-08 22:09:11 +00:00
|
|
|
case TYPE_RAID:
|
|
|
|
|
/* Only present the Smartarray HBA as a RAID controller.
|
|
|
|
|
* If it's a RAID controller other than the HBA itself
|
|
|
|
|
* (an external RAID controller, MSA500 or similar)
|
|
|
|
|
* don't present it.
|
|
|
|
|
*/
|
|
|
|
|
if (!is_hba_lunid(lunaddrbytes))
|
|
|
|
|
break;
|
|
|
|
|
ncurrent++;
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
2011-10-26 21:21:07 +00:00
|
|
|
if (ncurrent >= HPSA_MAX_DEVICES)
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
}
|
2015-11-04 21:52:34 +00:00
|
|
|
|
|
|
|
|
if (h->sas_host == NULL) {
|
|
|
|
|
int rc = 0;
|
|
|
|
|
|
|
|
|
|
rc = hpsa_add_sas_host(h);
|
|
|
|
|
if (rc) {
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"Could not add sas host %d\n", rc);
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-04 21:50:01 +00:00
|
|
|
adjust_hpsa_scsi_table(h, currentsd, ncurrent);
|
2009-12-08 22:09:11 +00:00
|
|
|
out:
|
|
|
|
|
kfree(tmpdevice);
|
|
|
|
|
for (i = 0; i < ndev_allocated; i++)
|
|
|
|
|
kfree(currentsd[i]);
|
|
|
|
|
kfree(currentsd);
|
|
|
|
|
kfree(physdev_list);
|
|
|
|
|
kfree(logdev_list);
|
2015-11-04 21:51:57 +00:00
|
|
|
kfree(id_ctlr);
|
2015-01-23 22:43:30 +00:00
|
|
|
kfree(id_phys);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:44:45 +00:00
|
|
|
static void hpsa_set_sg_descriptor(struct SGDescriptor *desc,
|
|
|
|
|
struct scatterlist *sg)
|
|
|
|
|
{
|
|
|
|
|
u64 addr64 = (u64) sg_dma_address(sg);
|
|
|
|
|
unsigned int len = sg_dma_len(sg);
|
|
|
|
|
|
|
|
|
|
desc->Addr = cpu_to_le64(addr64);
|
|
|
|
|
desc->Len = cpu_to_le32(len);
|
|
|
|
|
desc->Ext = 0;
|
|
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:42:17 +00:00
|
|
|
/*
|
|
|
|
|
* hpsa_scatter_gather takes a struct scsi_cmnd, (cmd), and does the pci
|
2009-12-08 22:09:11 +00:00
|
|
|
* dma mapping and fills in the scatter gather entries of the
|
|
|
|
|
* hpsa command, cp.
|
|
|
|
|
*/
|
2010-02-25 20:03:27 +00:00
|
|
|
static int hpsa_scatter_gather(struct ctlr_info *h,
|
2009-12-08 22:09:11 +00:00
|
|
|
struct CommandList *cp,
|
|
|
|
|
struct scsi_cmnd *cmd)
|
|
|
|
|
{
|
|
|
|
|
struct scatterlist *sg;
|
2015-04-23 14:34:27 +00:00
|
|
|
int use_sg, i, sg_limit, chained, last_sg;
|
2010-02-25 20:03:27 +00:00
|
|
|
struct SGDescriptor *curr_sg;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2010-02-25 20:03:27 +00:00
|
|
|
BUG_ON(scsi_sg_count(cmd) > h->maxsgentries);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
use_sg = scsi_dma_map(cmd);
|
|
|
|
|
if (use_sg < 0)
|
|
|
|
|
return use_sg;
|
|
|
|
|
|
|
|
|
|
if (!use_sg)
|
|
|
|
|
goto sglist_finished;
|
|
|
|
|
|
2015-04-23 14:34:27 +00:00
|
|
|
/*
|
|
|
|
|
* If the number of entries is greater than the max for a single list,
|
|
|
|
|
* then we have a chained list; we will set up all but one entry in the
|
|
|
|
|
* first list (the last entry is saved for link information);
|
|
|
|
|
* otherwise, we don't have a chained list and we'll set up at each of
|
|
|
|
|
* the entries in the one list.
|
|
|
|
|
*/
|
2010-02-25 20:03:27 +00:00
|
|
|
curr_sg = cp->SG;
|
2015-04-23 14:34:27 +00:00
|
|
|
chained = use_sg > h->max_cmd_sg_entries;
|
|
|
|
|
sg_limit = chained ? h->max_cmd_sg_entries - 1 : use_sg;
|
|
|
|
|
last_sg = scsi_sg_count(cmd) - 1;
|
|
|
|
|
scsi_for_each_sg(cmd, sg, sg_limit, i) {
|
2015-01-23 22:44:45 +00:00
|
|
|
hpsa_set_sg_descriptor(curr_sg, sg);
|
2010-02-25 20:03:27 +00:00
|
|
|
curr_sg++;
|
|
|
|
|
}
|
2015-01-23 22:44:45 +00:00
|
|
|
|
2015-04-23 14:34:27 +00:00
|
|
|
if (chained) {
|
|
|
|
|
/*
|
|
|
|
|
* Continue with the chained list. Set curr_sg to the chained
|
|
|
|
|
* list. Modify the limit to the total count less the entries
|
|
|
|
|
* we've already set up. Resume the scan at the list entry
|
|
|
|
|
* where the previous loop left off.
|
|
|
|
|
*/
|
|
|
|
|
curr_sg = h->cmd_sg_list[cp->cmdindex];
|
|
|
|
|
sg_limit = use_sg - sg_limit;
|
|
|
|
|
for_each_sg(sg, sg, sg_limit, i) {
|
|
|
|
|
hpsa_set_sg_descriptor(curr_sg, sg);
|
|
|
|
|
curr_sg++;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:44:45 +00:00
|
|
|
/* Back the pointer up to the last entry and mark it as "last". */
|
2015-04-23 14:34:27 +00:00
|
|
|
(curr_sg - 1)->Ext = cpu_to_le32(HPSA_SG_LAST);
|
2010-02-25 20:03:27 +00:00
|
|
|
|
|
|
|
|
if (use_sg + chained > h->maxSG)
|
|
|
|
|
h->maxSG = use_sg + chained;
|
|
|
|
|
|
|
|
|
|
if (chained) {
|
|
|
|
|
cp->Header.SGList = h->max_cmd_sg_entries;
|
2014-11-14 23:26:59 +00:00
|
|
|
cp->Header.SGTotal = cpu_to_le16(use_sg + 1);
|
2013-02-20 17:24:46 +00:00
|
|
|
if (hpsa_map_sg_chain_block(h, cp)) {
|
|
|
|
|
scsi_dma_unmap(cmd);
|
|
|
|
|
return -1;
|
|
|
|
|
}
|
2010-02-25 20:03:27 +00:00
|
|
|
return 0;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
sglist_finished:
|
|
|
|
|
|
2010-02-04 14:41:33 +00:00
|
|
|
cp->Header.SGList = (u8) use_sg; /* no. SGs contig in this cmd */
|
2015-01-23 22:42:17 +00:00
|
|
|
cp->Header.SGTotal = cpu_to_le16(use_sg); /* total sgs in cmd list */
|
2009-12-08 22:09:11 +00:00
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:55:33 +00:00
|
|
|
#define IO_ACCEL_INELIGIBLE (1)
|
|
|
|
|
static int fixup_ioaccel_cdb(u8 *cdb, int *cdb_len)
|
|
|
|
|
{
|
|
|
|
|
int is_write = 0;
|
|
|
|
|
u32 block;
|
|
|
|
|
u32 block_cnt;
|
|
|
|
|
|
|
|
|
|
/* Perform some CDB fixups if needed using 10 byte reads/writes only */
|
|
|
|
|
switch (cdb[0]) {
|
|
|
|
|
case WRITE_6:
|
|
|
|
|
case WRITE_12:
|
|
|
|
|
is_write = 1;
|
|
|
|
|
case READ_6:
|
|
|
|
|
case READ_12:
|
|
|
|
|
if (*cdb_len == 6) {
|
2015-11-04 21:50:50 +00:00
|
|
|
block = get_unaligned_be16(&cdb[2]);
|
2014-02-18 19:55:33 +00:00
|
|
|
block_cnt = cdb[4];
|
2015-11-04 21:50:50 +00:00
|
|
|
if (block_cnt == 0)
|
|
|
|
|
block_cnt = 256;
|
2014-02-18 19:55:33 +00:00
|
|
|
} else {
|
|
|
|
|
BUG_ON(*cdb_len != 12);
|
2015-11-04 21:50:50 +00:00
|
|
|
block = get_unaligned_be32(&cdb[2]);
|
|
|
|
|
block_cnt = get_unaligned_be32(&cdb[6]);
|
2014-02-18 19:55:33 +00:00
|
|
|
}
|
|
|
|
|
if (block_cnt > 0xffff)
|
|
|
|
|
return IO_ACCEL_INELIGIBLE;
|
|
|
|
|
|
|
|
|
|
cdb[0] = is_write ? WRITE_10 : READ_10;
|
|
|
|
|
cdb[1] = 0;
|
|
|
|
|
cdb[2] = (u8) (block >> 24);
|
|
|
|
|
cdb[3] = (u8) (block >> 16);
|
|
|
|
|
cdb[4] = (u8) (block >> 8);
|
|
|
|
|
cdb[5] = (u8) (block);
|
|
|
|
|
cdb[6] = 0;
|
|
|
|
|
cdb[7] = (u8) (block_cnt >> 8);
|
|
|
|
|
cdb[8] = (u8) (block_cnt);
|
|
|
|
|
cdb[9] = 0;
|
|
|
|
|
*cdb_len = 10;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:56:34 +00:00
|
|
|
static int hpsa_scsi_ioaccel1_queue_command(struct ctlr_info *h,
|
2014-02-18 19:55:33 +00:00
|
|
|
struct CommandList *c, u32 ioaccel_handle, u8 *cdb, int cdb_len,
|
2015-01-23 22:43:30 +00:00
|
|
|
u8 *scsi3addr, struct hpsa_scsi_dev_t *phys_disk)
|
2014-02-18 19:55:17 +00:00
|
|
|
{
|
|
|
|
|
struct scsi_cmnd *cmd = c->scsi_cmd;
|
|
|
|
|
struct io_accel1_cmd *cp = &h->ioaccel_cmd_pool[c->cmdindex];
|
|
|
|
|
unsigned int len;
|
|
|
|
|
unsigned int total_len = 0;
|
|
|
|
|
struct scatterlist *sg;
|
|
|
|
|
u64 addr64;
|
|
|
|
|
int use_sg, i;
|
|
|
|
|
struct SGDescriptor *curr_sg;
|
|
|
|
|
u32 control = IOACCEL1_CONTROL_SIMPLEQUEUE;
|
|
|
|
|
|
2014-02-18 19:55:33 +00:00
|
|
|
/* TODO: implement chaining support */
|
2015-01-23 22:43:30 +00:00
|
|
|
if (scsi_sg_count(cmd) > h->ioaccel_maxsg) {
|
|
|
|
|
atomic_dec(&phys_disk->ioaccel_cmds_out);
|
2014-02-18 19:55:33 +00:00
|
|
|
return IO_ACCEL_INELIGIBLE;
|
2015-01-23 22:43:30 +00:00
|
|
|
}
|
2014-02-18 19:55:33 +00:00
|
|
|
|
2014-02-18 19:55:17 +00:00
|
|
|
BUG_ON(cmd->cmd_len > IOACCEL1_IOFLAGS_CDBLEN_MAX);
|
|
|
|
|
|
2015-01-23 22:43:30 +00:00
|
|
|
if (fixup_ioaccel_cdb(cdb, &cdb_len)) {
|
|
|
|
|
atomic_dec(&phys_disk->ioaccel_cmds_out);
|
2014-02-18 19:55:33 +00:00
|
|
|
return IO_ACCEL_INELIGIBLE;
|
2015-01-23 22:43:30 +00:00
|
|
|
}
|
2014-02-18 19:55:33 +00:00
|
|
|
|
2014-02-18 19:55:17 +00:00
|
|
|
c->cmd_type = CMD_IOACCEL1;
|
|
|
|
|
|
|
|
|
|
/* Adjust the DMA address to point to the accelerated command buffer */
|
|
|
|
|
c->busaddr = (u32) h->ioaccel_cmd_pool_dhandle +
|
|
|
|
|
(c->cmdindex * sizeof(*cp));
|
|
|
|
|
BUG_ON(c->busaddr & 0x0000007F);
|
|
|
|
|
|
|
|
|
|
use_sg = scsi_dma_map(cmd);
|
2015-01-23 22:43:30 +00:00
|
|
|
if (use_sg < 0) {
|
|
|
|
|
atomic_dec(&phys_disk->ioaccel_cmds_out);
|
2014-02-18 19:55:17 +00:00
|
|
|
return use_sg;
|
2015-01-23 22:43:30 +00:00
|
|
|
}
|
2014-02-18 19:55:17 +00:00
|
|
|
|
|
|
|
|
if (use_sg) {
|
|
|
|
|
curr_sg = cp->SG;
|
|
|
|
|
scsi_for_each_sg(cmd, sg, use_sg, i) {
|
|
|
|
|
addr64 = (u64) sg_dma_address(sg);
|
|
|
|
|
len = sg_dma_len(sg);
|
|
|
|
|
total_len += len;
|
2014-11-14 23:26:59 +00:00
|
|
|
curr_sg->Addr = cpu_to_le64(addr64);
|
|
|
|
|
curr_sg->Len = cpu_to_le32(len);
|
|
|
|
|
curr_sg->Ext = cpu_to_le32(0);
|
2014-02-18 19:55:17 +00:00
|
|
|
curr_sg++;
|
|
|
|
|
}
|
2014-11-14 23:26:59 +00:00
|
|
|
(--curr_sg)->Ext = cpu_to_le32(HPSA_SG_LAST);
|
2014-02-18 19:55:17 +00:00
|
|
|
|
|
|
|
|
switch (cmd->sc_data_direction) {
|
|
|
|
|
case DMA_TO_DEVICE:
|
|
|
|
|
control |= IOACCEL1_CONTROL_DATA_OUT;
|
|
|
|
|
break;
|
|
|
|
|
case DMA_FROM_DEVICE:
|
|
|
|
|
control |= IOACCEL1_CONTROL_DATA_IN;
|
|
|
|
|
break;
|
|
|
|
|
case DMA_NONE:
|
|
|
|
|
control |= IOACCEL1_CONTROL_NODATAXFER;
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
dev_err(&h->pdev->dev, "unknown data direction: %d\n",
|
|
|
|
|
cmd->sc_data_direction);
|
|
|
|
|
BUG();
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
control |= IOACCEL1_CONTROL_NODATAXFER;
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:56:34 +00:00
|
|
|
c->Header.SGList = use_sg;
|
2014-02-18 19:55:17 +00:00
|
|
|
/* Fill out the command structure to submit */
|
2015-01-23 22:41:09 +00:00
|
|
|
cp->dev_handle = cpu_to_le16(ioaccel_handle & 0xFFFF);
|
|
|
|
|
cp->transfer_len = cpu_to_le32(total_len);
|
|
|
|
|
cp->io_flags = cpu_to_le16(IOACCEL1_IOFLAGS_IO_REQ |
|
|
|
|
|
(cdb_len & IOACCEL1_IOFLAGS_CDBLEN_MASK));
|
|
|
|
|
cp->control = cpu_to_le32(control);
|
2014-02-18 19:55:33 +00:00
|
|
|
memcpy(cp->CDB, cdb, cdb_len);
|
|
|
|
|
memcpy(cp->CISS_LUN, scsi3addr, 8);
|
2014-02-18 19:56:34 +00:00
|
|
|
/* Tag was already set at init time. */
|
2014-02-18 19:55:33 +00:00
|
|
|
enqueue_cmd_and_start_io(h, c);
|
2014-02-18 19:55:17 +00:00
|
|
|
return 0;
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2014-02-18 19:55:33 +00:00
|
|
|
/*
|
|
|
|
|
* Queue a command directly to a device behind the controller using the
|
|
|
|
|
* I/O accelerator path.
|
|
|
|
|
*/
|
|
|
|
|
static int hpsa_scsi_ioaccel_direct_map(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
struct scsi_cmnd *cmd = c->scsi_cmd;
|
|
|
|
|
struct hpsa_scsi_dev_t *dev = cmd->device->hostdata;
|
|
|
|
|
|
2015-01-23 22:43:30 +00:00
|
|
|
c->phys_disk = dev;
|
|
|
|
|
|
2014-02-18 19:55:33 +00:00
|
|
|
return hpsa_scsi_ioaccel_queue_command(h, c, dev->ioaccel_handle,
|
2015-01-23 22:43:30 +00:00
|
|
|
cmd->cmnd, cmd->cmd_len, dev->scsi3addr, dev);
|
2014-02-18 19:55:33 +00:00
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:57:31 +00:00
|
|
|
/*
|
|
|
|
|
* Set encryption parameters for the ioaccel2 request
|
|
|
|
|
*/
|
|
|
|
|
static void set_encrypt_ioaccel2(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c, struct io_accel2_cmd *cp)
|
|
|
|
|
{
|
|
|
|
|
struct scsi_cmnd *cmd = c->scsi_cmd;
|
|
|
|
|
struct hpsa_scsi_dev_t *dev = cmd->device->hostdata;
|
|
|
|
|
struct raid_map_data *map = &dev->raid_map;
|
|
|
|
|
u64 first_block;
|
|
|
|
|
|
|
|
|
|
/* Are we doing encryption on this device */
|
2015-01-23 22:41:09 +00:00
|
|
|
if (!(le16_to_cpu(map->flags) & RAID_MAP_FLAG_ENCRYPT_ON))
|
2014-02-18 19:57:31 +00:00
|
|
|
return;
|
|
|
|
|
/* Set the data encryption key index. */
|
|
|
|
|
cp->dekindex = map->dekindex;
|
|
|
|
|
|
|
|
|
|
/* Set the encryption enable flag, encoded into direction field. */
|
|
|
|
|
cp->direction |= IOACCEL2_DIRECTION_ENCRYPT_MASK;
|
|
|
|
|
|
|
|
|
|
/* Set encryption tweak values based on logical block address
|
|
|
|
|
* If block size is 512, tweak value is LBA.
|
|
|
|
|
* For other block sizes, tweak is (LBA * block size)/ 512)
|
|
|
|
|
*/
|
|
|
|
|
switch (cmd->cmnd[0]) {
|
|
|
|
|
/* Required? 6-byte cdbs eliminated by fixup_ioaccel_cdb */
|
|
|
|
|
case WRITE_6:
|
|
|
|
|
case READ_6:
|
2015-01-23 22:41:09 +00:00
|
|
|
first_block = get_unaligned_be16(&cmd->cmnd[2]);
|
2014-02-18 19:57:31 +00:00
|
|
|
break;
|
|
|
|
|
case WRITE_10:
|
|
|
|
|
case READ_10:
|
|
|
|
|
/* Required? 12-byte cdbs eliminated by fixup_ioaccel_cdb */
|
|
|
|
|
case WRITE_12:
|
|
|
|
|
case READ_12:
|
2015-01-23 22:41:09 +00:00
|
|
|
first_block = get_unaligned_be32(&cmd->cmnd[2]);
|
2014-02-18 19:57:31 +00:00
|
|
|
break;
|
|
|
|
|
case WRITE_16:
|
|
|
|
|
case READ_16:
|
2015-01-23 22:41:09 +00:00
|
|
|
first_block = get_unaligned_be64(&cmd->cmnd[2]);
|
2014-02-18 19:57:31 +00:00
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
dev_err(&h->pdev->dev,
|
2015-01-23 22:41:09 +00:00
|
|
|
"ERROR: %s: size (0x%x) not supported for encryption\n",
|
|
|
|
|
__func__, cmd->cmnd[0]);
|
2014-02-18 19:57:31 +00:00
|
|
|
BUG();
|
|
|
|
|
break;
|
|
|
|
|
}
|
2015-01-23 22:41:09 +00:00
|
|
|
|
|
|
|
|
if (le32_to_cpu(map->volume_blk_size) != 512)
|
|
|
|
|
first_block = first_block *
|
|
|
|
|
le32_to_cpu(map->volume_blk_size)/512;
|
|
|
|
|
|
|
|
|
|
cp->tweak_lower = cpu_to_le32(first_block);
|
|
|
|
|
cp->tweak_upper = cpu_to_le32(first_block >> 32);
|
2014-02-18 19:57:31 +00:00
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:56:34 +00:00
|
|
|
static int hpsa_scsi_ioaccel2_queue_command(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c, u32 ioaccel_handle, u8 *cdb, int cdb_len,
|
2015-01-23 22:43:30 +00:00
|
|
|
u8 *scsi3addr, struct hpsa_scsi_dev_t *phys_disk)
|
2014-02-18 19:56:34 +00:00
|
|
|
{
|
|
|
|
|
struct scsi_cmnd *cmd = c->scsi_cmd;
|
|
|
|
|
struct io_accel2_cmd *cp = &h->ioaccel2_cmd_pool[c->cmdindex];
|
|
|
|
|
struct ioaccel2_sg_element *curr_sg;
|
|
|
|
|
int use_sg, i;
|
|
|
|
|
struct scatterlist *sg;
|
|
|
|
|
u64 addr64;
|
|
|
|
|
u32 len;
|
|
|
|
|
u32 total_len = 0;
|
|
|
|
|
|
2015-04-23 14:33:27 +00:00
|
|
|
BUG_ON(scsi_sg_count(cmd) > h->maxsgentries);
|
2014-02-18 19:56:34 +00:00
|
|
|
|
2015-01-23 22:43:30 +00:00
|
|
|
if (fixup_ioaccel_cdb(cdb, &cdb_len)) {
|
|
|
|
|
atomic_dec(&phys_disk->ioaccel_cmds_out);
|
2014-02-18 19:56:34 +00:00
|
|
|
return IO_ACCEL_INELIGIBLE;
|
2015-01-23 22:43:30 +00:00
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:56:34 +00:00
|
|
|
c->cmd_type = CMD_IOACCEL2;
|
|
|
|
|
/* Adjust the DMA address to point to the accelerated command buffer */
|
|
|
|
|
c->busaddr = (u32) h->ioaccel2_cmd_pool_dhandle +
|
|
|
|
|
(c->cmdindex * sizeof(*cp));
|
|
|
|
|
BUG_ON(c->busaddr & 0x0000007F);
|
|
|
|
|
|
|
|
|
|
memset(cp, 0, sizeof(*cp));
|
|
|
|
|
cp->IU_type = IOACCEL2_IU_TYPE;
|
|
|
|
|
|
|
|
|
|
use_sg = scsi_dma_map(cmd);
|
2015-01-23 22:43:30 +00:00
|
|
|
if (use_sg < 0) {
|
|
|
|
|
atomic_dec(&phys_disk->ioaccel_cmds_out);
|
2014-02-18 19:56:34 +00:00
|
|
|
return use_sg;
|
2015-01-23 22:43:30 +00:00
|
|
|
}
|
2014-02-18 19:56:34 +00:00
|
|
|
|
|
|
|
|
if (use_sg) {
|
|
|
|
|
curr_sg = cp->sg;
|
2015-04-23 14:33:27 +00:00
|
|
|
if (use_sg > h->ioaccel_maxsg) {
|
|
|
|
|
addr64 = le64_to_cpu(
|
|
|
|
|
h->ioaccel2_cmd_sg_list[c->cmdindex]->address);
|
|
|
|
|
curr_sg->address = cpu_to_le64(addr64);
|
|
|
|
|
curr_sg->length = 0;
|
|
|
|
|
curr_sg->reserved[0] = 0;
|
|
|
|
|
curr_sg->reserved[1] = 0;
|
|
|
|
|
curr_sg->reserved[2] = 0;
|
|
|
|
|
curr_sg->chain_indicator = 0x80;
|
|
|
|
|
|
|
|
|
|
curr_sg = h->ioaccel2_cmd_sg_list[c->cmdindex];
|
|
|
|
|
}
|
2014-02-18 19:56:34 +00:00
|
|
|
scsi_for_each_sg(cmd, sg, use_sg, i) {
|
|
|
|
|
addr64 = (u64) sg_dma_address(sg);
|
|
|
|
|
len = sg_dma_len(sg);
|
|
|
|
|
total_len += len;
|
|
|
|
|
curr_sg->address = cpu_to_le64(addr64);
|
|
|
|
|
curr_sg->length = cpu_to_le32(len);
|
|
|
|
|
curr_sg->reserved[0] = 0;
|
|
|
|
|
curr_sg->reserved[1] = 0;
|
|
|
|
|
curr_sg->reserved[2] = 0;
|
|
|
|
|
curr_sg->chain_indicator = 0;
|
|
|
|
|
curr_sg++;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
switch (cmd->sc_data_direction) {
|
|
|
|
|
case DMA_TO_DEVICE:
|
2014-02-18 19:57:31 +00:00
|
|
|
cp->direction &= ~IOACCEL2_DIRECTION_MASK;
|
|
|
|
|
cp->direction |= IOACCEL2_DIR_DATA_OUT;
|
2014-02-18 19:56:34 +00:00
|
|
|
break;
|
|
|
|
|
case DMA_FROM_DEVICE:
|
2014-02-18 19:57:31 +00:00
|
|
|
cp->direction &= ~IOACCEL2_DIRECTION_MASK;
|
|
|
|
|
cp->direction |= IOACCEL2_DIR_DATA_IN;
|
2014-02-18 19:56:34 +00:00
|
|
|
break;
|
|
|
|
|
case DMA_NONE:
|
2014-02-18 19:57:31 +00:00
|
|
|
cp->direction &= ~IOACCEL2_DIRECTION_MASK;
|
|
|
|
|
cp->direction |= IOACCEL2_DIR_NO_DATA;
|
2014-02-18 19:56:34 +00:00
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
dev_err(&h->pdev->dev, "unknown data direction: %d\n",
|
|
|
|
|
cmd->sc_data_direction);
|
|
|
|
|
BUG();
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
} else {
|
2014-02-18 19:57:31 +00:00
|
|
|
cp->direction &= ~IOACCEL2_DIRECTION_MASK;
|
|
|
|
|
cp->direction |= IOACCEL2_DIR_NO_DATA;
|
2014-02-18 19:56:34 +00:00
|
|
|
}
|
2014-02-18 19:57:31 +00:00
|
|
|
|
|
|
|
|
/* Set encryption parameters, if necessary */
|
|
|
|
|
set_encrypt_ioaccel2(h, c, cp);
|
|
|
|
|
|
2015-01-23 22:41:09 +00:00
|
|
|
cp->scsi_nexus = cpu_to_le32(ioaccel_handle);
|
2015-01-23 22:43:09 +00:00
|
|
|
cp->Tag = cpu_to_le32(c->cmdindex << DIRECT_LOOKUP_SHIFT);
|
2014-02-18 19:56:34 +00:00
|
|
|
memcpy(cp->cdb, cdb, sizeof(cp->cdb));
|
|
|
|
|
|
|
|
|
|
cp->data_len = cpu_to_le32(total_len);
|
|
|
|
|
cp->err_ptr = cpu_to_le64(c->busaddr +
|
|
|
|
|
offsetof(struct io_accel2_cmd, error_data));
|
2014-11-14 23:26:59 +00:00
|
|
|
cp->err_len = cpu_to_le32(sizeof(cp->error_data));
|
2014-02-18 19:56:34 +00:00
|
|
|
|
2015-04-23 14:33:27 +00:00
|
|
|
/* fill in sg elements */
|
|
|
|
|
if (use_sg > h->ioaccel_maxsg) {
|
|
|
|
|
cp->sg_count = 1;
|
2015-11-04 21:51:14 +00:00
|
|
|
cp->sg[0].length = cpu_to_le32(use_sg * sizeof(cp->sg[0]));
|
2015-04-23 14:33:27 +00:00
|
|
|
if (hpsa_map_ioaccel2_sg_chain_block(h, cp, c)) {
|
|
|
|
|
atomic_dec(&phys_disk->ioaccel_cmds_out);
|
|
|
|
|
scsi_dma_unmap(cmd);
|
|
|
|
|
return -1;
|
|
|
|
|
}
|
|
|
|
|
} else
|
|
|
|
|
cp->sg_count = (u8) use_sg;
|
|
|
|
|
|
2014-02-18 19:56:34 +00:00
|
|
|
enqueue_cmd_and_start_io(h, c);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Queue a command to the correct I/O accelerator path.
|
|
|
|
|
*/
|
|
|
|
|
static int hpsa_scsi_ioaccel_queue_command(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c, u32 ioaccel_handle, u8 *cdb, int cdb_len,
|
2015-01-23 22:43:30 +00:00
|
|
|
u8 *scsi3addr, struct hpsa_scsi_dev_t *phys_disk)
|
2014-02-18 19:56:34 +00:00
|
|
|
{
|
2015-01-23 22:43:30 +00:00
|
|
|
/* Try to honor the device's queue depth */
|
|
|
|
|
if (atomic_inc_return(&phys_disk->ioaccel_cmds_out) >
|
|
|
|
|
phys_disk->queue_depth) {
|
|
|
|
|
atomic_dec(&phys_disk->ioaccel_cmds_out);
|
|
|
|
|
return IO_ACCEL_INELIGIBLE;
|
|
|
|
|
}
|
2014-02-18 19:56:34 +00:00
|
|
|
if (h->transMethod & CFGTBL_Trans_io_accel1)
|
|
|
|
|
return hpsa_scsi_ioaccel1_queue_command(h, c, ioaccel_handle,
|
2015-01-23 22:43:30 +00:00
|
|
|
cdb, cdb_len, scsi3addr,
|
|
|
|
|
phys_disk);
|
2014-02-18 19:56:34 +00:00
|
|
|
else
|
|
|
|
|
return hpsa_scsi_ioaccel2_queue_command(h, c, ioaccel_handle,
|
2015-01-23 22:43:30 +00:00
|
|
|
cdb, cdb_len, scsi3addr,
|
|
|
|
|
phys_disk);
|
2014-02-18 19:56:34 +00:00
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:56:55 +00:00
|
|
|
static void raid_map_helper(struct raid_map_data *map,
|
|
|
|
|
int offload_to_mirror, u32 *map_index, u32 *current_group)
|
|
|
|
|
{
|
|
|
|
|
if (offload_to_mirror == 0) {
|
|
|
|
|
/* use physical disk in the first mirrored group. */
|
2015-01-23 22:41:09 +00:00
|
|
|
*map_index %= le16_to_cpu(map->data_disks_per_row);
|
2014-02-18 19:56:55 +00:00
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
do {
|
|
|
|
|
/* determine mirror group that *map_index indicates */
|
2015-01-23 22:41:09 +00:00
|
|
|
*current_group = *map_index /
|
|
|
|
|
le16_to_cpu(map->data_disks_per_row);
|
2014-02-18 19:56:55 +00:00
|
|
|
if (offload_to_mirror == *current_group)
|
|
|
|
|
continue;
|
2015-01-23 22:41:09 +00:00
|
|
|
if (*current_group < le16_to_cpu(map->layout_map_count) - 1) {
|
2014-02-18 19:56:55 +00:00
|
|
|
/* select map index from next group */
|
2015-01-23 22:41:09 +00:00
|
|
|
*map_index += le16_to_cpu(map->data_disks_per_row);
|
2014-02-18 19:56:55 +00:00
|
|
|
(*current_group)++;
|
|
|
|
|
} else {
|
|
|
|
|
/* select map index from first group */
|
2015-01-23 22:41:09 +00:00
|
|
|
*map_index %= le16_to_cpu(map->data_disks_per_row);
|
2014-02-18 19:56:55 +00:00
|
|
|
*current_group = 0;
|
|
|
|
|
}
|
|
|
|
|
} while (offload_to_mirror != *current_group);
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:55:33 +00:00
|
|
|
/*
|
|
|
|
|
* Attempt to perform offload RAID mapping for a logical volume I/O.
|
|
|
|
|
*/
|
|
|
|
|
static int hpsa_scsi_ioaccel_raid_map(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
struct scsi_cmnd *cmd = c->scsi_cmd;
|
|
|
|
|
struct hpsa_scsi_dev_t *dev = cmd->device->hostdata;
|
|
|
|
|
struct raid_map_data *map = &dev->raid_map;
|
|
|
|
|
struct raid_map_disk_data *dd = &map->data[0];
|
|
|
|
|
int is_write = 0;
|
|
|
|
|
u32 map_index;
|
|
|
|
|
u64 first_block, last_block;
|
|
|
|
|
u32 block_cnt;
|
|
|
|
|
u32 blocks_per_row;
|
|
|
|
|
u64 first_row, last_row;
|
|
|
|
|
u32 first_row_offset, last_row_offset;
|
|
|
|
|
u32 first_column, last_column;
|
2014-02-18 19:56:55 +00:00
|
|
|
u64 r0_first_row, r0_last_row;
|
|
|
|
|
u32 r5or6_blocks_per_row;
|
|
|
|
|
u64 r5or6_first_row, r5or6_last_row;
|
|
|
|
|
u32 r5or6_first_row_offset, r5or6_last_row_offset;
|
|
|
|
|
u32 r5or6_first_column, r5or6_last_column;
|
|
|
|
|
u32 total_disks_per_row;
|
|
|
|
|
u32 stripesize;
|
|
|
|
|
u32 first_group, last_group, current_group;
|
2014-02-18 19:55:33 +00:00
|
|
|
u32 map_row;
|
|
|
|
|
u32 disk_handle;
|
|
|
|
|
u64 disk_block;
|
|
|
|
|
u32 disk_block_cnt;
|
|
|
|
|
u8 cdb[16];
|
|
|
|
|
u8 cdb_len;
|
2015-01-23 22:41:09 +00:00
|
|
|
u16 strip_size;
|
2014-02-18 19:55:33 +00:00
|
|
|
#if BITS_PER_LONG == 32
|
|
|
|
|
u64 tmpdiv;
|
|
|
|
|
#endif
|
2014-02-18 19:56:55 +00:00
|
|
|
int offload_to_mirror;
|
2014-02-18 19:55:33 +00:00
|
|
|
|
|
|
|
|
/* check for valid opcode, get LBA and block count */
|
|
|
|
|
switch (cmd->cmnd[0]) {
|
|
|
|
|
case WRITE_6:
|
|
|
|
|
is_write = 1;
|
|
|
|
|
case READ_6:
|
2015-11-04 21:50:50 +00:00
|
|
|
first_block = get_unaligned_be16(&cmd->cmnd[2]);
|
2014-02-18 19:55:33 +00:00
|
|
|
block_cnt = cmd->cmnd[4];
|
2014-07-03 15:18:14 +00:00
|
|
|
if (block_cnt == 0)
|
|
|
|
|
block_cnt = 256;
|
2014-02-18 19:55:33 +00:00
|
|
|
break;
|
|
|
|
|
case WRITE_10:
|
|
|
|
|
is_write = 1;
|
|
|
|
|
case READ_10:
|
|
|
|
|
first_block =
|
|
|
|
|
(((u64) cmd->cmnd[2]) << 24) |
|
|
|
|
|
(((u64) cmd->cmnd[3]) << 16) |
|
|
|
|
|
(((u64) cmd->cmnd[4]) << 8) |
|
|
|
|
|
cmd->cmnd[5];
|
|
|
|
|
block_cnt =
|
|
|
|
|
(((u32) cmd->cmnd[7]) << 8) |
|
|
|
|
|
cmd->cmnd[8];
|
|
|
|
|
break;
|
|
|
|
|
case WRITE_12:
|
|
|
|
|
is_write = 1;
|
|
|
|
|
case READ_12:
|
|
|
|
|
first_block =
|
|
|
|
|
(((u64) cmd->cmnd[2]) << 24) |
|
|
|
|
|
(((u64) cmd->cmnd[3]) << 16) |
|
|
|
|
|
(((u64) cmd->cmnd[4]) << 8) |
|
|
|
|
|
cmd->cmnd[5];
|
|
|
|
|
block_cnt =
|
|
|
|
|
(((u32) cmd->cmnd[6]) << 24) |
|
|
|
|
|
(((u32) cmd->cmnd[7]) << 16) |
|
|
|
|
|
(((u32) cmd->cmnd[8]) << 8) |
|
|
|
|
|
cmd->cmnd[9];
|
|
|
|
|
break;
|
|
|
|
|
case WRITE_16:
|
|
|
|
|
is_write = 1;
|
|
|
|
|
case READ_16:
|
|
|
|
|
first_block =
|
|
|
|
|
(((u64) cmd->cmnd[2]) << 56) |
|
|
|
|
|
(((u64) cmd->cmnd[3]) << 48) |
|
|
|
|
|
(((u64) cmd->cmnd[4]) << 40) |
|
|
|
|
|
(((u64) cmd->cmnd[5]) << 32) |
|
|
|
|
|
(((u64) cmd->cmnd[6]) << 24) |
|
|
|
|
|
(((u64) cmd->cmnd[7]) << 16) |
|
|
|
|
|
(((u64) cmd->cmnd[8]) << 8) |
|
|
|
|
|
cmd->cmnd[9];
|
|
|
|
|
block_cnt =
|
|
|
|
|
(((u32) cmd->cmnd[10]) << 24) |
|
|
|
|
|
(((u32) cmd->cmnd[11]) << 16) |
|
|
|
|
|
(((u32) cmd->cmnd[12]) << 8) |
|
|
|
|
|
cmd->cmnd[13];
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
return IO_ACCEL_INELIGIBLE; /* process via normal I/O path */
|
|
|
|
|
}
|
|
|
|
|
last_block = first_block + block_cnt - 1;
|
|
|
|
|
|
|
|
|
|
/* check for write to non-RAID-0 */
|
|
|
|
|
if (is_write && dev->raid_level != 0)
|
|
|
|
|
return IO_ACCEL_INELIGIBLE;
|
|
|
|
|
|
|
|
|
|
/* check for invalid block or wraparound */
|
2015-01-23 22:41:09 +00:00
|
|
|
if (last_block >= le64_to_cpu(map->volume_blk_cnt) ||
|
|
|
|
|
last_block < first_block)
|
2014-02-18 19:55:33 +00:00
|
|
|
return IO_ACCEL_INELIGIBLE;
|
|
|
|
|
|
|
|
|
|
/* calculate stripe information for the request */
|
2015-01-23 22:41:09 +00:00
|
|
|
blocks_per_row = le16_to_cpu(map->data_disks_per_row) *
|
|
|
|
|
le16_to_cpu(map->strip_size);
|
|
|
|
|
strip_size = le16_to_cpu(map->strip_size);
|
2014-02-18 19:55:33 +00:00
|
|
|
#if BITS_PER_LONG == 32
|
|
|
|
|
tmpdiv = first_block;
|
|
|
|
|
(void) do_div(tmpdiv, blocks_per_row);
|
|
|
|
|
first_row = tmpdiv;
|
|
|
|
|
tmpdiv = last_block;
|
|
|
|
|
(void) do_div(tmpdiv, blocks_per_row);
|
|
|
|
|
last_row = tmpdiv;
|
|
|
|
|
first_row_offset = (u32) (first_block - (first_row * blocks_per_row));
|
|
|
|
|
last_row_offset = (u32) (last_block - (last_row * blocks_per_row));
|
|
|
|
|
tmpdiv = first_row_offset;
|
2015-01-23 22:41:09 +00:00
|
|
|
(void) do_div(tmpdiv, strip_size);
|
2014-02-18 19:55:33 +00:00
|
|
|
first_column = tmpdiv;
|
|
|
|
|
tmpdiv = last_row_offset;
|
2015-01-23 22:41:09 +00:00
|
|
|
(void) do_div(tmpdiv, strip_size);
|
2014-02-18 19:55:33 +00:00
|
|
|
last_column = tmpdiv;
|
|
|
|
|
#else
|
|
|
|
|
first_row = first_block / blocks_per_row;
|
|
|
|
|
last_row = last_block / blocks_per_row;
|
|
|
|
|
first_row_offset = (u32) (first_block - (first_row * blocks_per_row));
|
|
|
|
|
last_row_offset = (u32) (last_block - (last_row * blocks_per_row));
|
2015-01-23 22:41:09 +00:00
|
|
|
first_column = first_row_offset / strip_size;
|
|
|
|
|
last_column = last_row_offset / strip_size;
|
2014-02-18 19:55:33 +00:00
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
/* if this isn't a single row/column then give to the controller */
|
|
|
|
|
if ((first_row != last_row) || (first_column != last_column))
|
|
|
|
|
return IO_ACCEL_INELIGIBLE;
|
|
|
|
|
|
|
|
|
|
/* proceeding with driver mapping */
|
2015-01-23 22:41:09 +00:00
|
|
|
total_disks_per_row = le16_to_cpu(map->data_disks_per_row) +
|
|
|
|
|
le16_to_cpu(map->metadata_disks_per_row);
|
2014-02-18 19:55:33 +00:00
|
|
|
map_row = ((u32)(first_row >> map->parity_rotation_shift)) %
|
2015-01-23 22:41:09 +00:00
|
|
|
le16_to_cpu(map->row_cnt);
|
2014-02-18 19:56:55 +00:00
|
|
|
map_index = (map_row * total_disks_per_row) + first_column;
|
|
|
|
|
|
|
|
|
|
switch (dev->raid_level) {
|
|
|
|
|
case HPSA_RAID_0:
|
|
|
|
|
break; /* nothing special to do */
|
|
|
|
|
case HPSA_RAID_1:
|
|
|
|
|
/* Handles load balance across RAID 1 members.
|
|
|
|
|
* (2-drive R1 and R10 with even # of drives.)
|
|
|
|
|
* Appropriate for SSDs, not optimal for HDDs
|
2014-02-18 19:55:33 +00:00
|
|
|
*/
|
2015-01-23 22:41:09 +00:00
|
|
|
BUG_ON(le16_to_cpu(map->layout_map_count) != 2);
|
2014-02-18 19:55:33 +00:00
|
|
|
if (dev->offload_to_mirror)
|
2015-01-23 22:41:09 +00:00
|
|
|
map_index += le16_to_cpu(map->data_disks_per_row);
|
2014-02-18 19:55:33 +00:00
|
|
|
dev->offload_to_mirror = !dev->offload_to_mirror;
|
2014-02-18 19:56:55 +00:00
|
|
|
break;
|
|
|
|
|
case HPSA_RAID_ADM:
|
|
|
|
|
/* Handles N-way mirrors (R1-ADM)
|
|
|
|
|
* and R10 with # of drives divisible by 3.)
|
|
|
|
|
*/
|
2015-01-23 22:41:09 +00:00
|
|
|
BUG_ON(le16_to_cpu(map->layout_map_count) != 3);
|
2014-02-18 19:56:55 +00:00
|
|
|
|
|
|
|
|
offload_to_mirror = dev->offload_to_mirror;
|
|
|
|
|
raid_map_helper(map, offload_to_mirror,
|
|
|
|
|
&map_index, ¤t_group);
|
|
|
|
|
/* set mirror group to use next time */
|
|
|
|
|
offload_to_mirror =
|
2015-01-23 22:41:09 +00:00
|
|
|
(offload_to_mirror >=
|
|
|
|
|
le16_to_cpu(map->layout_map_count) - 1)
|
2014-02-18 19:56:55 +00:00
|
|
|
? 0 : offload_to_mirror + 1;
|
|
|
|
|
dev->offload_to_mirror = offload_to_mirror;
|
|
|
|
|
/* Avoid direct use of dev->offload_to_mirror within this
|
|
|
|
|
* function since multiple threads might simultaneously
|
|
|
|
|
* increment it beyond the range of dev->layout_map_count -1.
|
|
|
|
|
*/
|
|
|
|
|
break;
|
|
|
|
|
case HPSA_RAID_5:
|
|
|
|
|
case HPSA_RAID_6:
|
2015-01-23 22:41:09 +00:00
|
|
|
if (le16_to_cpu(map->layout_map_count) <= 1)
|
2014-02-18 19:56:55 +00:00
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
/* Verify first and last block are in same RAID group */
|
|
|
|
|
r5or6_blocks_per_row =
|
2015-01-23 22:41:09 +00:00
|
|
|
le16_to_cpu(map->strip_size) *
|
|
|
|
|
le16_to_cpu(map->data_disks_per_row);
|
2014-02-18 19:56:55 +00:00
|
|
|
BUG_ON(r5or6_blocks_per_row == 0);
|
2015-01-23 22:41:09 +00:00
|
|
|
stripesize = r5or6_blocks_per_row *
|
|
|
|
|
le16_to_cpu(map->layout_map_count);
|
2014-02-18 19:56:55 +00:00
|
|
|
#if BITS_PER_LONG == 32
|
|
|
|
|
tmpdiv = first_block;
|
|
|
|
|
first_group = do_div(tmpdiv, stripesize);
|
|
|
|
|
tmpdiv = first_group;
|
|
|
|
|
(void) do_div(tmpdiv, r5or6_blocks_per_row);
|
|
|
|
|
first_group = tmpdiv;
|
|
|
|
|
tmpdiv = last_block;
|
|
|
|
|
last_group = do_div(tmpdiv, stripesize);
|
|
|
|
|
tmpdiv = last_group;
|
|
|
|
|
(void) do_div(tmpdiv, r5or6_blocks_per_row);
|
|
|
|
|
last_group = tmpdiv;
|
|
|
|
|
#else
|
|
|
|
|
first_group = (first_block % stripesize) / r5or6_blocks_per_row;
|
|
|
|
|
last_group = (last_block % stripesize) / r5or6_blocks_per_row;
|
|
|
|
|
#endif
|
2014-03-13 22:12:50 +00:00
|
|
|
if (first_group != last_group)
|
2014-02-18 19:56:55 +00:00
|
|
|
return IO_ACCEL_INELIGIBLE;
|
|
|
|
|
|
|
|
|
|
/* Verify request is in a single row of RAID 5/6 */
|
|
|
|
|
#if BITS_PER_LONG == 32
|
|
|
|
|
tmpdiv = first_block;
|
|
|
|
|
(void) do_div(tmpdiv, stripesize);
|
|
|
|
|
first_row = r5or6_first_row = r0_first_row = tmpdiv;
|
|
|
|
|
tmpdiv = last_block;
|
|
|
|
|
(void) do_div(tmpdiv, stripesize);
|
|
|
|
|
r5or6_last_row = r0_last_row = tmpdiv;
|
|
|
|
|
#else
|
|
|
|
|
first_row = r5or6_first_row = r0_first_row =
|
|
|
|
|
first_block / stripesize;
|
|
|
|
|
r5or6_last_row = r0_last_row = last_block / stripesize;
|
|
|
|
|
#endif
|
|
|
|
|
if (r5or6_first_row != r5or6_last_row)
|
|
|
|
|
return IO_ACCEL_INELIGIBLE;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/* Verify request is in a single column */
|
|
|
|
|
#if BITS_PER_LONG == 32
|
|
|
|
|
tmpdiv = first_block;
|
|
|
|
|
first_row_offset = do_div(tmpdiv, stripesize);
|
|
|
|
|
tmpdiv = first_row_offset;
|
|
|
|
|
first_row_offset = (u32) do_div(tmpdiv, r5or6_blocks_per_row);
|
|
|
|
|
r5or6_first_row_offset = first_row_offset;
|
|
|
|
|
tmpdiv = last_block;
|
|
|
|
|
r5or6_last_row_offset = do_div(tmpdiv, stripesize);
|
|
|
|
|
tmpdiv = r5or6_last_row_offset;
|
|
|
|
|
r5or6_last_row_offset = do_div(tmpdiv, r5or6_blocks_per_row);
|
|
|
|
|
tmpdiv = r5or6_first_row_offset;
|
|
|
|
|
(void) do_div(tmpdiv, map->strip_size);
|
|
|
|
|
first_column = r5or6_first_column = tmpdiv;
|
|
|
|
|
tmpdiv = r5or6_last_row_offset;
|
|
|
|
|
(void) do_div(tmpdiv, map->strip_size);
|
|
|
|
|
r5or6_last_column = tmpdiv;
|
|
|
|
|
#else
|
|
|
|
|
first_row_offset = r5or6_first_row_offset =
|
|
|
|
|
(u32)((first_block % stripesize) %
|
|
|
|
|
r5or6_blocks_per_row);
|
|
|
|
|
|
|
|
|
|
r5or6_last_row_offset =
|
|
|
|
|
(u32)((last_block % stripesize) %
|
|
|
|
|
r5or6_blocks_per_row);
|
|
|
|
|
|
|
|
|
|
first_column = r5or6_first_column =
|
2015-01-23 22:41:09 +00:00
|
|
|
r5or6_first_row_offset / le16_to_cpu(map->strip_size);
|
2014-02-18 19:56:55 +00:00
|
|
|
r5or6_last_column =
|
2015-01-23 22:41:09 +00:00
|
|
|
r5or6_last_row_offset / le16_to_cpu(map->strip_size);
|
2014-02-18 19:56:55 +00:00
|
|
|
#endif
|
|
|
|
|
if (r5or6_first_column != r5or6_last_column)
|
|
|
|
|
return IO_ACCEL_INELIGIBLE;
|
|
|
|
|
|
|
|
|
|
/* Request is eligible */
|
|
|
|
|
map_row = ((u32)(first_row >> map->parity_rotation_shift)) %
|
2015-01-23 22:41:09 +00:00
|
|
|
le16_to_cpu(map->row_cnt);
|
2014-02-18 19:56:55 +00:00
|
|
|
|
|
|
|
|
map_index = (first_group *
|
2015-01-23 22:41:09 +00:00
|
|
|
(le16_to_cpu(map->row_cnt) * total_disks_per_row)) +
|
2014-02-18 19:56:55 +00:00
|
|
|
(map_row * total_disks_per_row) + first_column;
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
return IO_ACCEL_INELIGIBLE;
|
2014-02-18 19:55:33 +00:00
|
|
|
}
|
2014-02-18 19:56:55 +00:00
|
|
|
|
2015-01-23 22:44:14 +00:00
|
|
|
if (unlikely(map_index >= RAID_MAP_MAX_ENTRIES))
|
|
|
|
|
return IO_ACCEL_INELIGIBLE;
|
|
|
|
|
|
2015-01-23 22:43:30 +00:00
|
|
|
c->phys_disk = dev->phys_disk[map_index];
|
2016-02-23 21:16:34 +00:00
|
|
|
if (!c->phys_disk)
|
|
|
|
|
return IO_ACCEL_INELIGIBLE;
|
2015-01-23 22:43:30 +00:00
|
|
|
|
2014-02-18 19:55:33 +00:00
|
|
|
disk_handle = dd[map_index].ioaccel_handle;
|
2015-01-23 22:41:09 +00:00
|
|
|
disk_block = le64_to_cpu(map->disk_starting_blk) +
|
|
|
|
|
first_row * le16_to_cpu(map->strip_size) +
|
|
|
|
|
(first_row_offset - first_column *
|
|
|
|
|
le16_to_cpu(map->strip_size));
|
2014-02-18 19:55:33 +00:00
|
|
|
disk_block_cnt = block_cnt;
|
|
|
|
|
|
|
|
|
|
/* handle differing logical/physical block sizes */
|
|
|
|
|
if (map->phys_blk_shift) {
|
|
|
|
|
disk_block <<= map->phys_blk_shift;
|
|
|
|
|
disk_block_cnt <<= map->phys_blk_shift;
|
|
|
|
|
}
|
|
|
|
|
BUG_ON(disk_block_cnt > 0xffff);
|
|
|
|
|
|
|
|
|
|
/* build the new CDB for the physical disk I/O */
|
|
|
|
|
if (disk_block > 0xffffffff) {
|
|
|
|
|
cdb[0] = is_write ? WRITE_16 : READ_16;
|
|
|
|
|
cdb[1] = 0;
|
|
|
|
|
cdb[2] = (u8) (disk_block >> 56);
|
|
|
|
|
cdb[3] = (u8) (disk_block >> 48);
|
|
|
|
|
cdb[4] = (u8) (disk_block >> 40);
|
|
|
|
|
cdb[5] = (u8) (disk_block >> 32);
|
|
|
|
|
cdb[6] = (u8) (disk_block >> 24);
|
|
|
|
|
cdb[7] = (u8) (disk_block >> 16);
|
|
|
|
|
cdb[8] = (u8) (disk_block >> 8);
|
|
|
|
|
cdb[9] = (u8) (disk_block);
|
|
|
|
|
cdb[10] = (u8) (disk_block_cnt >> 24);
|
|
|
|
|
cdb[11] = (u8) (disk_block_cnt >> 16);
|
|
|
|
|
cdb[12] = (u8) (disk_block_cnt >> 8);
|
|
|
|
|
cdb[13] = (u8) (disk_block_cnt);
|
|
|
|
|
cdb[14] = 0;
|
|
|
|
|
cdb[15] = 0;
|
|
|
|
|
cdb_len = 16;
|
|
|
|
|
} else {
|
|
|
|
|
cdb[0] = is_write ? WRITE_10 : READ_10;
|
|
|
|
|
cdb[1] = 0;
|
|
|
|
|
cdb[2] = (u8) (disk_block >> 24);
|
|
|
|
|
cdb[3] = (u8) (disk_block >> 16);
|
|
|
|
|
cdb[4] = (u8) (disk_block >> 8);
|
|
|
|
|
cdb[5] = (u8) (disk_block);
|
|
|
|
|
cdb[6] = 0;
|
|
|
|
|
cdb[7] = (u8) (disk_block_cnt >> 8);
|
|
|
|
|
cdb[8] = (u8) (disk_block_cnt);
|
|
|
|
|
cdb[9] = 0;
|
|
|
|
|
cdb_len = 10;
|
|
|
|
|
}
|
|
|
|
|
return hpsa_scsi_ioaccel_queue_command(h, c, disk_handle, cdb, cdb_len,
|
2015-01-23 22:43:30 +00:00
|
|
|
dev->scsi3addr,
|
|
|
|
|
dev->phys_disk[map_index]);
|
2014-02-18 19:55:33 +00:00
|
|
|
}
|
|
|
|
|
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
/*
|
|
|
|
|
* Submit commands down the "normal" RAID stack path
|
|
|
|
|
* All callers to hpsa_ciss_submit must check lockup_detected
|
|
|
|
|
* beforehand, before (opt.) and after calling cmd_alloc
|
|
|
|
|
*/
|
2015-01-23 22:43:20 +00:00
|
|
|
static int hpsa_ciss_submit(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c, struct scsi_cmnd *cmd,
|
|
|
|
|
unsigned char scsi3addr[])
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
|
|
|
|
cmd->host_scribble = (unsigned char *) c;
|
|
|
|
|
c->cmd_type = CMD_SCSI;
|
|
|
|
|
c->scsi_cmd = cmd;
|
|
|
|
|
c->Header.ReplyQueue = 0; /* unused in simple mode */
|
|
|
|
|
memcpy(&c->Header.LUN.LunAddrBytes[0], &scsi3addr[0], 8);
|
2015-01-23 22:43:09 +00:00
|
|
|
c->Header.tag = cpu_to_le64((c->cmdindex << DIRECT_LOOKUP_SHIFT));
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
/* Fill in the request block... */
|
|
|
|
|
|
|
|
|
|
c->Request.Timeout = 0;
|
|
|
|
|
BUG_ON(cmd->cmd_len > sizeof(c->Request.CDB));
|
|
|
|
|
c->Request.CDBLen = cmd->cmd_len;
|
|
|
|
|
memcpy(c->Request.CDB, cmd->cmnd, cmd->cmd_len);
|
|
|
|
|
switch (cmd->sc_data_direction) {
|
|
|
|
|
case DMA_TO_DEVICE:
|
2014-11-14 23:27:04 +00:00
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(TYPE_CMD, ATTR_SIMPLE, XFER_WRITE);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case DMA_FROM_DEVICE:
|
2014-11-14 23:27:04 +00:00
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(TYPE_CMD, ATTR_SIMPLE, XFER_READ);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case DMA_NONE:
|
2014-11-14 23:27:04 +00:00
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(TYPE_CMD, ATTR_SIMPLE, XFER_NONE);
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case DMA_BIDIRECTIONAL:
|
|
|
|
|
/* This can happen if a buggy application does a scsi passthru
|
|
|
|
|
* and sets both inlen and outlen to non-zero. ( see
|
|
|
|
|
* ../scsi/scsi_ioctl.c:scsi_ioctl_send_command() )
|
|
|
|
|
*/
|
|
|
|
|
|
2014-11-14 23:27:04 +00:00
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(TYPE_CMD, ATTR_SIMPLE, XFER_RSVD);
|
2009-12-08 22:09:11 +00:00
|
|
|
/* This is technically wrong, and hpsa controllers should
|
|
|
|
|
* reject it with CMD_INVALID, which is the most correct
|
|
|
|
|
* response, but non-fibre backends appear to let it
|
|
|
|
|
* slide by, and give the same results as if this field
|
|
|
|
|
* were set correctly. Either way is acceptable for
|
|
|
|
|
* our purposes here.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
dev_err(&h->pdev->dev, "unknown data direction: %d\n",
|
|
|
|
|
cmd->sc_data_direction);
|
|
|
|
|
BUG();
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
2010-02-25 20:03:27 +00:00
|
|
|
if (hpsa_scatter_gather(h, c, cmd) < 0) { /* Fill SG list */
|
2015-04-23 14:35:04 +00:00
|
|
|
hpsa_cmd_resolve_and_free(h, c);
|
2009-12-08 22:09:11 +00:00
|
|
|
return SCSI_MLQUEUE_HOST_BUSY;
|
|
|
|
|
}
|
|
|
|
|
enqueue_cmd_and_start_io(h, c);
|
|
|
|
|
/* the cmd'll come back via intr handler in complete_scsi_command() */
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:32:32 +00:00
|
|
|
static void hpsa_cmd_init(struct ctlr_info *h, int index,
|
|
|
|
|
struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
dma_addr_t cmd_dma_handle, err_dma_handle;
|
|
|
|
|
|
|
|
|
|
/* Zero out all of commandlist except the last field, refcount */
|
|
|
|
|
memset(c, 0, offsetof(struct CommandList, refcount));
|
|
|
|
|
c->Header.tag = cpu_to_le64((u64) (index << DIRECT_LOOKUP_SHIFT));
|
|
|
|
|
cmd_dma_handle = h->cmd_pool_dhandle + index * sizeof(*c);
|
|
|
|
|
c->err_info = h->errinfo_pool + index;
|
|
|
|
|
memset(c->err_info, 0, sizeof(*c->err_info));
|
|
|
|
|
err_dma_handle = h->errinfo_pool_dhandle
|
|
|
|
|
+ index * sizeof(*c->err_info);
|
|
|
|
|
c->cmdindex = index;
|
|
|
|
|
c->busaddr = (u32) cmd_dma_handle;
|
|
|
|
|
c->ErrDesc.Addr = cpu_to_le64((u64) err_dma_handle);
|
|
|
|
|
c->ErrDesc.Len = cpu_to_le32((u32) sizeof(*c->err_info));
|
|
|
|
|
c->h = h;
|
2015-04-23 14:34:16 +00:00
|
|
|
c->scsi_cmd = SCSI_CMD_IDLE;
|
2015-04-23 14:32:32 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_preinitialize_commands(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < h->nr_cmds; i++) {
|
|
|
|
|
struct CommandList *c = h->cmd_pool + i;
|
|
|
|
|
|
|
|
|
|
hpsa_cmd_init(h, i, c);
|
|
|
|
|
atomic_set(&c->refcount, 0);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline void hpsa_cmd_partial_init(struct ctlr_info *h, int index,
|
|
|
|
|
struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
dma_addr_t cmd_dma_handle = h->cmd_pool_dhandle + index * sizeof(*c);
|
|
|
|
|
|
2015-04-23 14:35:04 +00:00
|
|
|
BUG_ON(c->cmdindex != index);
|
|
|
|
|
|
2015-04-23 14:32:32 +00:00
|
|
|
memset(c->Request.CDB, 0, sizeof(c->Request.CDB));
|
|
|
|
|
memset(c->err_info, 0, sizeof(*c->err_info));
|
|
|
|
|
c->busaddr = (u32) cmd_dma_handle;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:32:48 +00:00
|
|
|
static int hpsa_ioaccel_submit(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c, struct scsi_cmnd *cmd,
|
|
|
|
|
unsigned char *scsi3addr)
|
|
|
|
|
{
|
|
|
|
|
struct hpsa_scsi_dev_t *dev = cmd->device->hostdata;
|
|
|
|
|
int rc = IO_ACCEL_INELIGIBLE;
|
|
|
|
|
|
|
|
|
|
cmd->host_scribble = (unsigned char *) c;
|
|
|
|
|
|
|
|
|
|
if (dev->offload_enabled) {
|
|
|
|
|
hpsa_cmd_init(h, c->cmdindex, c);
|
|
|
|
|
c->cmd_type = CMD_SCSI;
|
|
|
|
|
c->scsi_cmd = cmd;
|
|
|
|
|
rc = hpsa_scsi_ioaccel_raid_map(h, c);
|
|
|
|
|
if (rc < 0) /* scsi_dma_map failed. */
|
|
|
|
|
rc = SCSI_MLQUEUE_HOST_BUSY;
|
2015-04-23 14:32:59 +00:00
|
|
|
} else if (dev->hba_ioaccel_enabled) {
|
2015-04-23 14:32:48 +00:00
|
|
|
hpsa_cmd_init(h, c->cmdindex, c);
|
|
|
|
|
c->cmd_type = CMD_SCSI;
|
|
|
|
|
c->scsi_cmd = cmd;
|
|
|
|
|
rc = hpsa_scsi_ioaccel_direct_map(h, c);
|
|
|
|
|
if (rc < 0) /* scsi_dma_map failed. */
|
|
|
|
|
rc = SCSI_MLQUEUE_HOST_BUSY;
|
|
|
|
|
}
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:43:25 +00:00
|
|
|
static void hpsa_command_resubmit_worker(struct work_struct *work)
|
|
|
|
|
{
|
|
|
|
|
struct scsi_cmnd *cmd;
|
|
|
|
|
struct hpsa_scsi_dev_t *dev;
|
2015-04-23 14:34:11 +00:00
|
|
|
struct CommandList *c = container_of(work, struct CommandList, work);
|
2015-01-23 22:43:25 +00:00
|
|
|
|
|
|
|
|
cmd = c->scsi_cmd;
|
|
|
|
|
dev = cmd->device->hostdata;
|
|
|
|
|
if (!dev) {
|
|
|
|
|
cmd->result = DID_NO_CONNECT << 16;
|
2015-04-23 14:34:11 +00:00
|
|
|
return hpsa_cmd_free_and_done(c->h, c, cmd);
|
2015-01-23 22:43:25 +00:00
|
|
|
}
|
2015-04-23 14:35:22 +00:00
|
|
|
if (c->reset_pending)
|
|
|
|
|
return hpsa_cmd_resolve_and_free(c->h, c);
|
2015-04-23 14:34:16 +00:00
|
|
|
if (c->abort_pending)
|
|
|
|
|
return hpsa_cmd_abort_and_free(c->h, c, cmd);
|
2015-04-23 14:32:48 +00:00
|
|
|
if (c->cmd_type == CMD_IOACCEL2) {
|
|
|
|
|
struct ctlr_info *h = c->h;
|
|
|
|
|
struct io_accel2_cmd *c2 = &h->ioaccel2_cmd_pool[c->cmdindex];
|
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
|
|
if (c2->error_data.serv_response ==
|
|
|
|
|
IOACCEL2_STATUS_SR_TASK_COMP_SET_FULL) {
|
|
|
|
|
rc = hpsa_ioaccel_submit(h, c, cmd, dev->scsi3addr);
|
|
|
|
|
if (rc == 0)
|
|
|
|
|
return;
|
|
|
|
|
if (rc == SCSI_MLQUEUE_HOST_BUSY) {
|
|
|
|
|
/*
|
|
|
|
|
* If we get here, it means dma mapping failed.
|
|
|
|
|
* Try again via scsi mid layer, which will
|
|
|
|
|
* then get SCSI_MLQUEUE_HOST_BUSY.
|
|
|
|
|
*/
|
|
|
|
|
cmd->result = DID_IMM_RETRY << 16;
|
2015-04-23 14:34:11 +00:00
|
|
|
return hpsa_cmd_free_and_done(h, c, cmd);
|
2015-04-23 14:32:48 +00:00
|
|
|
}
|
|
|
|
|
/* else, fall thru and resubmit down CISS path */
|
|
|
|
|
}
|
|
|
|
|
}
|
2015-04-23 14:32:32 +00:00
|
|
|
hpsa_cmd_partial_init(c->h, c->cmdindex, c);
|
2015-01-23 22:43:25 +00:00
|
|
|
if (hpsa_ciss_submit(c->h, c, cmd, dev->scsi3addr)) {
|
|
|
|
|
/*
|
|
|
|
|
* If we get here, it means dma mapping failed. Try
|
|
|
|
|
* again via scsi mid layer, which will then get
|
|
|
|
|
* SCSI_MLQUEUE_HOST_BUSY.
|
2015-04-23 14:32:48 +00:00
|
|
|
*
|
|
|
|
|
* hpsa_ciss_submit will have already freed c
|
|
|
|
|
* if it encountered a dma mapping failure.
|
2015-01-23 22:43:25 +00:00
|
|
|
*/
|
|
|
|
|
cmd->result = DID_IMM_RETRY << 16;
|
|
|
|
|
cmd->scsi_done(cmd);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:43:20 +00:00
|
|
|
/* Running in struct Scsi_Host->host_lock less mode */
|
|
|
|
|
static int hpsa_scsi_queue_command(struct Scsi_Host *sh, struct scsi_cmnd *cmd)
|
|
|
|
|
{
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct hpsa_scsi_dev_t *dev;
|
|
|
|
|
unsigned char scsi3addr[8];
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
int rc = 0;
|
|
|
|
|
|
|
|
|
|
/* Get the ptr to our adapter structure out of cmd->host. */
|
|
|
|
|
h = sdev_to_hba(cmd->device);
|
2015-04-23 14:35:04 +00:00
|
|
|
|
|
|
|
|
BUG_ON(cmd->request->tag < 0);
|
|
|
|
|
|
2015-01-23 22:43:20 +00:00
|
|
|
dev = cmd->device->hostdata;
|
|
|
|
|
if (!dev) {
|
|
|
|
|
cmd->result = DID_NO_CONNECT << 16;
|
|
|
|
|
cmd->scsi_done(cmd);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:35:04 +00:00
|
|
|
memcpy(scsi3addr, dev->scsi3addr, sizeof(scsi3addr));
|
2015-04-23 14:33:38 +00:00
|
|
|
|
2015-01-23 22:44:19 +00:00
|
|
|
if (unlikely(lockup_detected(h))) {
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
cmd->result = DID_NO_CONNECT << 16;
|
2015-01-23 22:44:19 +00:00
|
|
|
cmd->scsi_done(cmd);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
2015-04-23 14:35:04 +00:00
|
|
|
c = cmd_tagged_alloc(h, cmd);
|
2015-01-23 22:43:20 +00:00
|
|
|
|
2015-01-23 22:44:19 +00:00
|
|
|
/*
|
|
|
|
|
* Call alternate submit routine for I/O accelerated commands.
|
2015-01-23 22:43:20 +00:00
|
|
|
* Retries always go down the normal I/O path.
|
|
|
|
|
*/
|
|
|
|
|
if (likely(cmd->retries == 0 &&
|
|
|
|
|
cmd->request->cmd_type == REQ_TYPE_FS &&
|
|
|
|
|
h->acciopath_status)) {
|
2015-04-23 14:32:48 +00:00
|
|
|
rc = hpsa_ioaccel_submit(h, c, cmd, scsi3addr);
|
|
|
|
|
if (rc == 0)
|
|
|
|
|
return 0;
|
|
|
|
|
if (rc == SCSI_MLQUEUE_HOST_BUSY) {
|
2015-04-23 14:35:04 +00:00
|
|
|
hpsa_cmd_resolve_and_free(h, c);
|
2015-04-23 14:32:48 +00:00
|
|
|
return SCSI_MLQUEUE_HOST_BUSY;
|
2015-01-23 22:43:20 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
return hpsa_ciss_submit(h, c, cmd, scsi3addr);
|
|
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:44:50 +00:00
|
|
|
static void hpsa_scan_complete(struct ctlr_info *h)
|
2014-02-18 19:55:48 +00:00
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
2015-01-23 22:44:50 +00:00
|
|
|
spin_lock_irqsave(&h->scan_lock, flags);
|
|
|
|
|
h->scan_finished = 1;
|
|
|
|
|
wake_up_all(&h->scan_wait_queue);
|
|
|
|
|
spin_unlock_irqrestore(&h->scan_lock, flags);
|
2014-02-18 19:55:48 +00:00
|
|
|
}
|
|
|
|
|
|
2010-02-04 14:43:16 +00:00
|
|
|
static void hpsa_scan_start(struct Scsi_Host *sh)
|
|
|
|
|
{
|
|
|
|
|
struct ctlr_info *h = shost_to_hba(sh);
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
2015-01-23 22:44:50 +00:00
|
|
|
/*
|
|
|
|
|
* Don't let rescans be initiated on a controller known to be locked
|
|
|
|
|
* up. If the controller locks up *during* a rescan, that thread is
|
|
|
|
|
* probably hosed, but at least we can prevent new rescan threads from
|
|
|
|
|
* piling up on a locked up controller.
|
|
|
|
|
*/
|
|
|
|
|
if (unlikely(lockup_detected(h)))
|
|
|
|
|
return hpsa_scan_complete(h);
|
2014-02-18 19:55:48 +00:00
|
|
|
|
2010-02-04 14:43:16 +00:00
|
|
|
/* wait until any scan already in progress is finished. */
|
|
|
|
|
while (1) {
|
|
|
|
|
spin_lock_irqsave(&h->scan_lock, flags);
|
|
|
|
|
if (h->scan_finished)
|
|
|
|
|
break;
|
|
|
|
|
spin_unlock_irqrestore(&h->scan_lock, flags);
|
|
|
|
|
wait_event(h->scan_wait_queue, h->scan_finished);
|
|
|
|
|
/* Note: We don't need to worry about a race between this
|
|
|
|
|
* thread and driver unload because the midlayer will
|
|
|
|
|
* have incremented the reference count, so unload won't
|
|
|
|
|
* happen if we're in here.
|
|
|
|
|
*/
|
|
|
|
|
}
|
|
|
|
|
h->scan_finished = 0; /* mark scan as in progress */
|
|
|
|
|
spin_unlock_irqrestore(&h->scan_lock, flags);
|
|
|
|
|
|
2015-01-23 22:44:50 +00:00
|
|
|
if (unlikely(lockup_detected(h)))
|
|
|
|
|
return hpsa_scan_complete(h);
|
2014-02-18 19:55:48 +00:00
|
|
|
|
2015-11-04 21:50:01 +00:00
|
|
|
hpsa_update_scsi_devices(h);
|
2010-02-04 14:43:16 +00:00
|
|
|
|
2015-01-23 22:44:50 +00:00
|
|
|
hpsa_scan_complete(h);
|
2010-02-04 14:43:16 +00:00
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:41:30 +00:00
|
|
|
static int hpsa_change_queue_depth(struct scsi_device *sdev, int qdepth)
|
|
|
|
|
{
|
2015-01-23 22:43:30 +00:00
|
|
|
struct hpsa_scsi_dev_t *logical_drive = sdev->hostdata;
|
|
|
|
|
|
|
|
|
|
if (!logical_drive)
|
|
|
|
|
return -ENODEV;
|
2015-01-23 22:41:30 +00:00
|
|
|
|
|
|
|
|
if (qdepth < 1)
|
|
|
|
|
qdepth = 1;
|
2015-01-23 22:43:30 +00:00
|
|
|
else if (qdepth > logical_drive->queue_depth)
|
|
|
|
|
qdepth = logical_drive->queue_depth;
|
|
|
|
|
|
|
|
|
|
return scsi_change_queue_depth(sdev, qdepth);
|
2015-01-23 22:41:30 +00:00
|
|
|
}
|
|
|
|
|
|
2010-02-04 14:43:16 +00:00
|
|
|
static int hpsa_scan_finished(struct Scsi_Host *sh,
|
|
|
|
|
unsigned long elapsed_time)
|
|
|
|
|
{
|
|
|
|
|
struct ctlr_info *h = shost_to_hba(sh);
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
int finished;
|
|
|
|
|
|
|
|
|
|
spin_lock_irqsave(&h->scan_lock, flags);
|
|
|
|
|
finished = h->scan_finished;
|
|
|
|
|
spin_unlock_irqrestore(&h->scan_lock, flags);
|
|
|
|
|
return finished;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:35:09 +00:00
|
|
|
static int hpsa_scsi_host_alloc(struct ctlr_info *h)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
2012-01-19 20:00:53 +00:00
|
|
|
struct Scsi_Host *sh;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2012-01-19 20:00:53 +00:00
|
|
|
sh = scsi_host_alloc(&hpsa_driver_template, sizeof(h));
|
2015-04-23 14:35:09 +00:00
|
|
|
if (sh == NULL) {
|
|
|
|
|
dev_err(&h->pdev->dev, "scsi_host_alloc failed\n");
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
}
|
2012-01-19 20:00:53 +00:00
|
|
|
|
|
|
|
|
sh->io_port = 0;
|
|
|
|
|
sh->n_io_port = 0;
|
|
|
|
|
sh->this_id = -1;
|
|
|
|
|
sh->max_channel = 3;
|
|
|
|
|
sh->max_cmd_len = MAX_COMMAND_SIZE;
|
|
|
|
|
sh->max_lun = HPSA_MAX_LUN;
|
|
|
|
|
sh->max_id = HPSA_MAX_LUN;
|
2015-04-23 14:31:47 +00:00
|
|
|
sh->can_queue = h->nr_cmds - HPSA_NRESERVED_CMDS;
|
2015-01-23 22:43:30 +00:00
|
|
|
sh->cmd_per_lun = sh->can_queue;
|
2012-01-19 20:00:53 +00:00
|
|
|
sh->sg_tablesize = h->maxsgentries;
|
2015-11-04 21:52:34 +00:00
|
|
|
sh->transportt = hpsa_sas_transport_template;
|
2012-01-19 20:00:53 +00:00
|
|
|
sh->hostdata[0] = (unsigned long) h;
|
|
|
|
|
sh->irq = h->intr[h->intr_mode];
|
|
|
|
|
sh->unique_id = sh->irq;
|
2015-10-08 08:28:04 +00:00
|
|
|
|
2015-04-23 14:35:09 +00:00
|
|
|
h->scsi_host = sh;
|
2012-01-19 20:00:53 +00:00
|
|
|
return 0;
|
2015-04-23 14:35:09 +00:00
|
|
|
}
|
2012-01-19 20:00:53 +00:00
|
|
|
|
2015-04-23 14:35:09 +00:00
|
|
|
static int hpsa_scsi_add_host(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
int rv;
|
|
|
|
|
|
|
|
|
|
rv = scsi_add_host(h->scsi_host, &h->pdev->dev);
|
|
|
|
|
if (rv) {
|
|
|
|
|
dev_err(&h->pdev->dev, "scsi_add_host failed\n");
|
|
|
|
|
return rv;
|
|
|
|
|
}
|
|
|
|
|
scsi_scan_host(h->scsi_host);
|
|
|
|
|
return 0;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:35:04 +00:00
|
|
|
/*
|
|
|
|
|
* The block layer has already gone to the trouble of picking out a unique,
|
|
|
|
|
* small-integer tag for this request. We use an offset from that value as
|
|
|
|
|
* an index to select our command block. (The offset allows us to reserve the
|
|
|
|
|
* low-numbered entries for our own uses.)
|
|
|
|
|
*/
|
|
|
|
|
static int hpsa_get_cmd_index(struct scsi_cmnd *scmd)
|
|
|
|
|
{
|
|
|
|
|
int idx = scmd->request->tag;
|
|
|
|
|
|
|
|
|
|
if (idx < 0)
|
|
|
|
|
return idx;
|
|
|
|
|
|
|
|
|
|
/* Offset to leave space for internal cmds. */
|
|
|
|
|
return idx += HPSA_NRESERVED_CMDS;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:34:22 +00:00
|
|
|
/*
|
|
|
|
|
* Send a TEST_UNIT_READY command to the specified LUN using the specified
|
|
|
|
|
* reply queue; returns zero if the unit is ready, and non-zero otherwise.
|
|
|
|
|
*/
|
|
|
|
|
static int hpsa_send_test_unit_ready(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c, unsigned char lunaddr[],
|
|
|
|
|
int reply_queue)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
|
|
/* Send the Test Unit Ready, fill_cmd can't fail, no mapping */
|
|
|
|
|
(void) fill_cmd(c, TEST_UNIT_READY, h,
|
|
|
|
|
NULL, 0, 0, lunaddr, TYPE_CMD);
|
2016-04-27 22:13:51 +00:00
|
|
|
rc = hpsa_scsi_do_simple_cmd(h, c, reply_queue, DEFAULT_TIMEOUT);
|
2015-04-23 14:34:22 +00:00
|
|
|
if (rc)
|
|
|
|
|
return rc;
|
|
|
|
|
/* no unmap needed here because no data xfer. */
|
|
|
|
|
|
|
|
|
|
/* Check if the unit is already ready. */
|
|
|
|
|
if (c->err_info->CommandStatus == CMD_SUCCESS)
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* The first command sent after reset will receive "unit attention" to
|
|
|
|
|
* indicate that the LUN has been reset...this is actually what we're
|
|
|
|
|
* looking for (but, success is good too).
|
|
|
|
|
*/
|
|
|
|
|
if (c->err_info->CommandStatus == CMD_TARGET_STATUS &&
|
|
|
|
|
c->err_info->ScsiStatus == SAM_STAT_CHECK_CONDITION &&
|
|
|
|
|
(c->err_info->SenseInfo[2] == NO_SENSE ||
|
|
|
|
|
c->err_info->SenseInfo[2] == UNIT_ATTENTION))
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
return 1;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Wait for a TEST_UNIT_READY command to complete, retrying as necessary;
|
|
|
|
|
* returns zero when the unit is ready, and non-zero when giving up.
|
|
|
|
|
*/
|
|
|
|
|
static int hpsa_wait_for_test_unit_ready(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c,
|
|
|
|
|
unsigned char lunaddr[], int reply_queue)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
2014-02-21 22:25:05 +00:00
|
|
|
int rc;
|
2009-12-08 22:09:11 +00:00
|
|
|
int count = 0;
|
|
|
|
|
int waittime = 1; /* seconds */
|
|
|
|
|
|
|
|
|
|
/* Send test unit ready until device ready, or give up. */
|
2015-04-23 14:34:22 +00:00
|
|
|
for (count = 0; count < HPSA_TUR_RETRY_LIMIT; count++) {
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-04-23 14:34:22 +00:00
|
|
|
/*
|
|
|
|
|
* Wait for a bit. do this first, because if we send
|
2009-12-08 22:09:11 +00:00
|
|
|
* the TUR right away, the reset will just abort it.
|
|
|
|
|
*/
|
|
|
|
|
msleep(1000 * waittime);
|
2015-04-23 14:34:22 +00:00
|
|
|
|
|
|
|
|
rc = hpsa_send_test_unit_ready(h, c, lunaddr, reply_queue);
|
|
|
|
|
if (!rc)
|
|
|
|
|
break;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
/* Increase wait time with each try, up to a point. */
|
|
|
|
|
if (waittime < HPSA_MAX_WAIT_INTERVAL_SECS)
|
2015-04-23 14:34:22 +00:00
|
|
|
waittime *= 2;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-04-23 14:34:22 +00:00
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"waiting %d secs for device to become ready.\n",
|
|
|
|
|
waittime);
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-04-23 14:34:22 +00:00
|
|
|
return rc;
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-04-23 14:34:22 +00:00
|
|
|
static int wait_for_device_to_become_ready(struct ctlr_info *h,
|
|
|
|
|
unsigned char lunaddr[],
|
|
|
|
|
int reply_queue)
|
|
|
|
|
{
|
|
|
|
|
int first_queue;
|
|
|
|
|
int last_queue;
|
|
|
|
|
int rq;
|
|
|
|
|
int rc = 0;
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
|
|
|
|
|
c = cmd_alloc(h);
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* If no specific reply queue was requested, then send the TUR
|
|
|
|
|
* repeatedly, requesting a reply on each reply queue; otherwise execute
|
|
|
|
|
* the loop exactly once using only the specified queue.
|
|
|
|
|
*/
|
|
|
|
|
if (reply_queue == DEFAULT_REPLY_QUEUE) {
|
|
|
|
|
first_queue = 0;
|
|
|
|
|
last_queue = h->nreply_queues - 1;
|
|
|
|
|
} else {
|
|
|
|
|
first_queue = reply_queue;
|
|
|
|
|
last_queue = reply_queue;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
for (rq = first_queue; rq <= last_queue; rq++) {
|
|
|
|
|
rc = hpsa_wait_for_test_unit_ready(h, c, lunaddr, rq);
|
|
|
|
|
if (rc)
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (rc)
|
|
|
|
|
dev_warn(&h->pdev->dev, "giving up on device.\n");
|
|
|
|
|
else
|
|
|
|
|
dev_warn(&h->pdev->dev, "device is ready.\n");
|
|
|
|
|
|
2015-01-23 22:43:04 +00:00
|
|
|
cmd_free(h, c);
|
2009-12-08 22:09:11 +00:00
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Need at least one of these error handlers to keep ../scsi/hosts.c from
|
|
|
|
|
* complaining. Doing a host- or bus-reset can't do anything good here.
|
|
|
|
|
*/
|
|
|
|
|
static int hpsa_eh_device_reset_handler(struct scsi_cmnd *scsicmd)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct hpsa_scsi_dev_t *dev;
|
2015-11-04 21:51:02 +00:00
|
|
|
u8 reset_type;
|
2015-06-04 14:47:56 +00:00
|
|
|
char msg[48];
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
/* find the controller to which the command to be aborted was sent */
|
|
|
|
|
h = sdev_to_hba(scsicmd->device);
|
|
|
|
|
if (h == NULL) /* paranoia */
|
|
|
|
|
return FAILED;
|
2015-01-23 22:44:24 +00:00
|
|
|
|
|
|
|
|
if (lockup_detected(h))
|
|
|
|
|
return FAILED;
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
dev = scsicmd->device->hostdata;
|
|
|
|
|
if (!dev) {
|
2015-04-23 14:35:22 +00:00
|
|
|
dev_err(&h->pdev->dev, "%s: device lookup failed\n", __func__);
|
2009-12-08 22:09:11 +00:00
|
|
|
return FAILED;
|
|
|
|
|
}
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
|
|
|
|
|
/* if controller locked up, we can guarantee command won't complete */
|
|
|
|
|
if (lockup_detected(h)) {
|
2015-06-04 14:47:56 +00:00
|
|
|
snprintf(msg, sizeof(msg),
|
|
|
|
|
"cmd %d RESET FAILED, lockup detected",
|
|
|
|
|
hpsa_get_cmd_index(scsicmd));
|
2015-04-23 14:35:04 +00:00
|
|
|
hpsa_show_dev_msg(KERN_WARNING, h, dev, msg);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
return FAILED;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* this reset request might be the result of a lockup; check */
|
|
|
|
|
if (detect_controller_lockup(h)) {
|
2015-06-04 14:47:56 +00:00
|
|
|
snprintf(msg, sizeof(msg),
|
|
|
|
|
"cmd %d RESET FAILED, new lockup detected",
|
|
|
|
|
hpsa_get_cmd_index(scsicmd));
|
2015-04-23 14:35:04 +00:00
|
|
|
hpsa_show_dev_msg(KERN_WARNING, h, dev, msg);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
return FAILED;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:35:22 +00:00
|
|
|
/* Do not attempt on controller */
|
|
|
|
|
if (is_hba_lunid(dev->scsi3addr))
|
|
|
|
|
return SUCCESS;
|
|
|
|
|
|
2015-11-04 21:51:02 +00:00
|
|
|
if (is_logical_dev_addr_mode(dev->scsi3addr))
|
|
|
|
|
reset_type = HPSA_DEVICE_RESET_MSG;
|
|
|
|
|
else
|
|
|
|
|
reset_type = HPSA_PHYS_TARGET_RESET;
|
|
|
|
|
|
|
|
|
|
sprintf(msg, "resetting %s",
|
|
|
|
|
reset_type == HPSA_DEVICE_RESET_MSG ? "logical " : "physical ");
|
|
|
|
|
hpsa_show_dev_msg(KERN_WARNING, h, dev, msg);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
|
2015-11-04 21:50:56 +00:00
|
|
|
h->reset_in_progress = 1;
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/* send a reset to the SCSI LUN which the command was sent to */
|
2015-11-04 21:51:02 +00:00
|
|
|
rc = hpsa_do_reset(h, dev, dev->scsi3addr, reset_type,
|
2015-04-23 14:35:22 +00:00
|
|
|
DEFAULT_REPLY_QUEUE);
|
2015-11-04 21:51:02 +00:00
|
|
|
sprintf(msg, "reset %s %s",
|
|
|
|
|
reset_type == HPSA_DEVICE_RESET_MSG ? "logical " : "physical ",
|
|
|
|
|
rc == 0 ? "completed successfully" : "failed");
|
2015-04-23 14:35:22 +00:00
|
|
|
hpsa_show_dev_msg(KERN_WARNING, h, dev, msg);
|
2015-11-04 21:50:56 +00:00
|
|
|
h->reset_in_progress = 0;
|
2015-04-23 14:35:22 +00:00
|
|
|
return rc == 0 ? SUCCESS : FAILED;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
2012-05-01 16:42:56 +00:00
|
|
|
static void swizzle_abort_tag(u8 *tag)
|
|
|
|
|
{
|
|
|
|
|
u8 original_tag[8];
|
|
|
|
|
|
|
|
|
|
memcpy(original_tag, tag, 8);
|
|
|
|
|
tag[0] = original_tag[3];
|
|
|
|
|
tag[1] = original_tag[2];
|
|
|
|
|
tag[2] = original_tag[1];
|
|
|
|
|
tag[3] = original_tag[0];
|
|
|
|
|
tag[4] = original_tag[7];
|
|
|
|
|
tag[5] = original_tag[6];
|
|
|
|
|
tag[6] = original_tag[5];
|
|
|
|
|
tag[7] = original_tag[4];
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:55:28 +00:00
|
|
|
static void hpsa_get_tag(struct ctlr_info *h,
|
2015-01-23 22:41:09 +00:00
|
|
|
struct CommandList *c, __le32 *taglower, __le32 *tagupper)
|
2014-02-18 19:55:28 +00:00
|
|
|
{
|
2015-01-23 22:41:09 +00:00
|
|
|
u64 tag;
|
2014-02-18 19:55:28 +00:00
|
|
|
if (c->cmd_type == CMD_IOACCEL1) {
|
|
|
|
|
struct io_accel1_cmd *cm1 = (struct io_accel1_cmd *)
|
|
|
|
|
&h->ioaccel_cmd_pool[c->cmdindex];
|
2015-01-23 22:41:09 +00:00
|
|
|
tag = le64_to_cpu(cm1->tag);
|
|
|
|
|
*tagupper = cpu_to_le32(tag >> 32);
|
|
|
|
|
*taglower = cpu_to_le32(tag);
|
2014-02-18 19:56:45 +00:00
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
if (c->cmd_type == CMD_IOACCEL2) {
|
|
|
|
|
struct io_accel2_cmd *cm2 = (struct io_accel2_cmd *)
|
|
|
|
|
&h->ioaccel2_cmd_pool[c->cmdindex];
|
2014-02-18 19:57:31 +00:00
|
|
|
/* upper tag not used in ioaccel2 mode */
|
|
|
|
|
memset(tagupper, 0, sizeof(*tagupper));
|
|
|
|
|
*taglower = cm2->Tag;
|
2014-02-18 19:56:45 +00:00
|
|
|
return;
|
2014-02-18 19:55:28 +00:00
|
|
|
}
|
2015-01-23 22:41:09 +00:00
|
|
|
tag = le64_to_cpu(c->Header.tag);
|
|
|
|
|
*tagupper = cpu_to_le32(tag >> 32);
|
|
|
|
|
*taglower = cpu_to_le32(tag);
|
2014-02-18 19:55:28 +00:00
|
|
|
}
|
|
|
|
|
|
2012-05-01 16:42:51 +00:00
|
|
|
static int hpsa_send_abort(struct ctlr_info *h, unsigned char *scsi3addr,
|
2015-04-23 14:32:06 +00:00
|
|
|
struct CommandList *abort, int reply_queue)
|
2012-05-01 16:42:51 +00:00
|
|
|
{
|
|
|
|
|
int rc = IO_OK;
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
struct ErrorInfo *ei;
|
2015-01-23 22:41:09 +00:00
|
|
|
__le32 tagupper, taglower;
|
2012-05-01 16:42:51 +00:00
|
|
|
|
2015-01-23 22:43:04 +00:00
|
|
|
c = cmd_alloc(h);
|
2012-05-01 16:42:51 +00:00
|
|
|
|
2013-02-20 17:24:41 +00:00
|
|
|
/* fill_cmd can't fail here, no buffer to map */
|
2015-04-23 14:32:06 +00:00
|
|
|
(void) fill_cmd(c, HPSA_ABORT_MSG, h, &abort->Header.tag,
|
2013-02-20 17:24:41 +00:00
|
|
|
0, 0, scsi3addr, TYPE_MSG);
|
2015-04-23 14:32:06 +00:00
|
|
|
if (h->needs_abort_tags_swizzled)
|
2012-05-01 16:42:56 +00:00
|
|
|
swizzle_abort_tag(&c->Request.CDB[4]);
|
2016-04-27 22:13:51 +00:00
|
|
|
(void) hpsa_scsi_do_simple_cmd(h, c, reply_queue, DEFAULT_TIMEOUT);
|
2014-02-18 19:55:28 +00:00
|
|
|
hpsa_get_tag(h, abort, &taglower, &tagupper);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
dev_dbg(&h->pdev->dev, "%s: Tag:0x%08x:%08x: do_simple_cmd(abort) completed.\n",
|
2014-02-18 19:55:28 +00:00
|
|
|
__func__, tagupper, taglower);
|
2012-05-01 16:42:51 +00:00
|
|
|
/* no unmap needed here because no data xfer. */
|
|
|
|
|
|
|
|
|
|
ei = c->err_info;
|
|
|
|
|
switch (ei->CommandStatus) {
|
|
|
|
|
case CMD_SUCCESS:
|
|
|
|
|
break;
|
2015-04-23 14:32:16 +00:00
|
|
|
case CMD_TMF_STATUS:
|
|
|
|
|
rc = hpsa_evaluate_tmf_status(h, c);
|
|
|
|
|
break;
|
2012-05-01 16:42:51 +00:00
|
|
|
case CMD_UNABORTABLE: /* Very common, don't make noise. */
|
|
|
|
|
rc = -1;
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
dev_dbg(&h->pdev->dev, "%s: Tag:0x%08x:%08x: interpreting error.\n",
|
2014-02-18 19:55:28 +00:00
|
|
|
__func__, tagupper, taglower);
|
2014-02-18 19:57:47 +00:00
|
|
|
hpsa_scsi_interpret_error(h, c);
|
2012-05-01 16:42:51 +00:00
|
|
|
rc = -1;
|
|
|
|
|
break;
|
|
|
|
|
}
|
2015-01-23 22:43:04 +00:00
|
|
|
cmd_free(h, c);
|
2014-02-18 19:57:31 +00:00
|
|
|
dev_dbg(&h->pdev->dev, "%s: Tag:0x%08x:%08x: Finished.\n",
|
|
|
|
|
__func__, tagupper, taglower);
|
2012-05-01 16:42:51 +00:00
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:34:06 +00:00
|
|
|
static void setup_ioaccel2_abort_cmd(struct CommandList *c, struct ctlr_info *h,
|
|
|
|
|
struct CommandList *command_to_abort, int reply_queue)
|
|
|
|
|
{
|
|
|
|
|
struct io_accel2_cmd *c2 = &h->ioaccel2_cmd_pool[c->cmdindex];
|
|
|
|
|
struct hpsa_tmf_struct *ac = (struct hpsa_tmf_struct *) c2;
|
|
|
|
|
struct io_accel2_cmd *c2a =
|
|
|
|
|
&h->ioaccel2_cmd_pool[command_to_abort->cmdindex];
|
2015-04-23 14:34:16 +00:00
|
|
|
struct scsi_cmnd *scmd = command_to_abort->scsi_cmd;
|
2015-04-23 14:34:06 +00:00
|
|
|
struct hpsa_scsi_dev_t *dev = scmd->device->hostdata;
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* We're overlaying struct hpsa_tmf_struct on top of something which
|
|
|
|
|
* was allocated as a struct io_accel2_cmd, so we better be sure it
|
|
|
|
|
* actually fits, and doesn't overrun the error info space.
|
|
|
|
|
*/
|
|
|
|
|
BUILD_BUG_ON(sizeof(struct hpsa_tmf_struct) >
|
|
|
|
|
sizeof(struct io_accel2_cmd));
|
|
|
|
|
BUG_ON(offsetof(struct io_accel2_cmd, error_data) <
|
|
|
|
|
offsetof(struct hpsa_tmf_struct, error_len) +
|
|
|
|
|
sizeof(ac->error_len));
|
|
|
|
|
|
|
|
|
|
c->cmd_type = IOACCEL2_TMF;
|
2015-04-23 14:34:16 +00:00
|
|
|
c->scsi_cmd = SCSI_CMD_BUSY;
|
|
|
|
|
|
2015-04-23 14:34:06 +00:00
|
|
|
/* Adjust the DMA address to point to the accelerated command buffer */
|
|
|
|
|
c->busaddr = (u32) h->ioaccel2_cmd_pool_dhandle +
|
|
|
|
|
(c->cmdindex * sizeof(struct io_accel2_cmd));
|
|
|
|
|
BUG_ON(c->busaddr & 0x0000007F);
|
|
|
|
|
|
|
|
|
|
memset(ac, 0, sizeof(*c2)); /* yes this is correct */
|
|
|
|
|
ac->iu_type = IOACCEL2_IU_TMF_TYPE;
|
|
|
|
|
ac->reply_queue = reply_queue;
|
|
|
|
|
ac->tmf = IOACCEL2_TMF_ABORT;
|
|
|
|
|
ac->it_nexus = cpu_to_le32(dev->ioaccel_handle);
|
|
|
|
|
memset(ac->lun_id, 0, sizeof(ac->lun_id));
|
|
|
|
|
ac->tag = cpu_to_le64(c->cmdindex << DIRECT_LOOKUP_SHIFT);
|
|
|
|
|
ac->abort_tag = cpu_to_le64(le32_to_cpu(c2a->Tag));
|
|
|
|
|
ac->error_ptr = cpu_to_le64(c->busaddr +
|
|
|
|
|
offsetof(struct io_accel2_cmd, error_data));
|
|
|
|
|
ac->error_len = cpu_to_le32(sizeof(c2->error_data));
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:56:45 +00:00
|
|
|
/* ioaccel2 path firmware cannot handle abort task requests.
|
|
|
|
|
* Change abort requests to physical target reset, and send to the
|
|
|
|
|
* address of the physical disk used for the ioaccel 2 command.
|
|
|
|
|
* Return 0 on success (IO_OK)
|
|
|
|
|
* -1 on failure
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
static int hpsa_send_reset_as_abort_ioaccel2(struct ctlr_info *h,
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
unsigned char *scsi3addr, struct CommandList *abort, int reply_queue)
|
2014-02-18 19:56:45 +00:00
|
|
|
{
|
|
|
|
|
int rc = IO_OK;
|
|
|
|
|
struct scsi_cmnd *scmd; /* scsi command within request being aborted */
|
|
|
|
|
struct hpsa_scsi_dev_t *dev; /* device to which scsi cmd was sent */
|
|
|
|
|
unsigned char phys_scsi3addr[8]; /* addr of phys disk with volume */
|
|
|
|
|
unsigned char *psa = &phys_scsi3addr[0];
|
|
|
|
|
|
|
|
|
|
/* Get a pointer to the hpsa logical device. */
|
2015-01-23 22:44:30 +00:00
|
|
|
scmd = abort->scsi_cmd;
|
2014-02-18 19:56:45 +00:00
|
|
|
dev = (struct hpsa_scsi_dev_t *)(scmd->device->hostdata);
|
|
|
|
|
if (dev == NULL) {
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"Cannot abort: no device pointer for command.\n");
|
|
|
|
|
return -1; /* not abortable */
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:57:52 +00:00
|
|
|
if (h->raid_offload_debug > 0)
|
|
|
|
|
dev_info(&h->pdev->dev,
|
2015-04-23 14:31:55 +00:00
|
|
|
"scsi %d:%d:%d:%d %s scsi3addr 0x%02x%02x%02x%02x%02x%02x%02x%02x\n",
|
2014-02-18 19:57:52 +00:00
|
|
|
h->scsi_host->host_no, dev->bus, dev->target, dev->lun,
|
2015-04-23 14:31:55 +00:00
|
|
|
"Reset as abort",
|
2014-02-18 19:57:52 +00:00
|
|
|
scsi3addr[0], scsi3addr[1], scsi3addr[2], scsi3addr[3],
|
|
|
|
|
scsi3addr[4], scsi3addr[5], scsi3addr[6], scsi3addr[7]);
|
|
|
|
|
|
2014-02-18 19:56:45 +00:00
|
|
|
if (!dev->offload_enabled) {
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"Can't abort: device is not operating in HP SSD Smart Path mode.\n");
|
|
|
|
|
return -1; /* not abortable */
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Incoming scsi3addr is logical addr. We need physical disk addr. */
|
|
|
|
|
if (!hpsa_get_pdisk_of_ioaccel2(h, abort, psa)) {
|
|
|
|
|
dev_warn(&h->pdev->dev, "Can't abort: Failed lookup of physical address.\n");
|
|
|
|
|
return -1; /* not abortable */
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* send the reset */
|
2014-02-18 19:57:52 +00:00
|
|
|
if (h->raid_offload_debug > 0)
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"Reset as abort: Resetting physical device at scsi3addr 0x%02x%02x%02x%02x%02x%02x%02x%02x\n",
|
|
|
|
|
psa[0], psa[1], psa[2], psa[3],
|
|
|
|
|
psa[4], psa[5], psa[6], psa[7]);
|
2015-04-23 14:35:22 +00:00
|
|
|
rc = hpsa_do_reset(h, dev, psa, HPSA_RESET_TYPE_TARGET, reply_queue);
|
2014-02-18 19:56:45 +00:00
|
|
|
if (rc != 0) {
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"Reset as abort: Failed on physical device at scsi3addr 0x%02x%02x%02x%02x%02x%02x%02x%02x\n",
|
|
|
|
|
psa[0], psa[1], psa[2], psa[3],
|
|
|
|
|
psa[4], psa[5], psa[6], psa[7]);
|
|
|
|
|
return rc; /* failed to reset */
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* wait for device to recover */
|
2015-04-23 14:34:22 +00:00
|
|
|
if (wait_for_device_to_become_ready(h, psa, reply_queue) != 0) {
|
2014-02-18 19:56:45 +00:00
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"Reset as abort: Failed: Device never recovered from reset: 0x%02x%02x%02x%02x%02x%02x%02x%02x\n",
|
|
|
|
|
psa[0], psa[1], psa[2], psa[3],
|
|
|
|
|
psa[4], psa[5], psa[6], psa[7]);
|
|
|
|
|
return -1; /* failed to recover */
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* device recovered */
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"Reset as abort: Device recovered from reset: scsi3addr 0x%02x%02x%02x%02x%02x%02x%02x%02x\n",
|
|
|
|
|
psa[0], psa[1], psa[2], psa[3],
|
|
|
|
|
psa[4], psa[5], psa[6], psa[7]);
|
|
|
|
|
|
|
|
|
|
return rc; /* success */
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:34:06 +00:00
|
|
|
static int hpsa_send_abort_ioaccel2(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *abort, int reply_queue)
|
|
|
|
|
{
|
|
|
|
|
int rc = IO_OK;
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
__le32 taglower, tagupper;
|
|
|
|
|
struct hpsa_scsi_dev_t *dev;
|
|
|
|
|
struct io_accel2_cmd *c2;
|
|
|
|
|
|
|
|
|
|
dev = abort->scsi_cmd->device->hostdata;
|
|
|
|
|
if (!dev->offload_enabled && !dev->hba_ioaccel_enabled)
|
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
|
|
c = cmd_alloc(h);
|
|
|
|
|
setup_ioaccel2_abort_cmd(c, h, abort, reply_queue);
|
|
|
|
|
c2 = &h->ioaccel2_cmd_pool[c->cmdindex];
|
2016-04-27 22:13:51 +00:00
|
|
|
(void) hpsa_scsi_do_simple_cmd(h, c, reply_queue, DEFAULT_TIMEOUT);
|
2015-04-23 14:34:06 +00:00
|
|
|
hpsa_get_tag(h, abort, &taglower, &tagupper);
|
|
|
|
|
dev_dbg(&h->pdev->dev,
|
|
|
|
|
"%s: Tag:0x%08x:%08x: do_simple_cmd(ioaccel2 abort) completed.\n",
|
|
|
|
|
__func__, tagupper, taglower);
|
|
|
|
|
/* no unmap needed here because no data xfer. */
|
|
|
|
|
|
|
|
|
|
dev_dbg(&h->pdev->dev,
|
|
|
|
|
"%s: Tag:0x%08x:%08x: abort service response = 0x%02x.\n",
|
|
|
|
|
__func__, tagupper, taglower, c2->error_data.serv_response);
|
|
|
|
|
switch (c2->error_data.serv_response) {
|
|
|
|
|
case IOACCEL2_SERV_RESPONSE_TMF_COMPLETE:
|
|
|
|
|
case IOACCEL2_SERV_RESPONSE_TMF_SUCCESS:
|
|
|
|
|
rc = 0;
|
|
|
|
|
break;
|
|
|
|
|
case IOACCEL2_SERV_RESPONSE_TMF_REJECTED:
|
|
|
|
|
case IOACCEL2_SERV_RESPONSE_FAILURE:
|
|
|
|
|
case IOACCEL2_SERV_RESPONSE_TMF_WRONG_LUN:
|
|
|
|
|
rc = -1;
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"%s: Tag:0x%08x:%08x: unknown abort service response 0x%02x\n",
|
|
|
|
|
__func__, tagupper, taglower,
|
|
|
|
|
c2->error_data.serv_response);
|
|
|
|
|
rc = -1;
|
|
|
|
|
}
|
|
|
|
|
cmd_free(h, c);
|
|
|
|
|
dev_dbg(&h->pdev->dev, "%s: Tag:0x%08x:%08x: Finished.\n", __func__,
|
|
|
|
|
tagupper, taglower);
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2012-05-01 16:42:56 +00:00
|
|
|
static int hpsa_send_abort_both_ways(struct ctlr_info *h,
|
2016-02-23 21:16:28 +00:00
|
|
|
struct hpsa_scsi_dev_t *dev, struct CommandList *abort, int reply_queue)
|
2012-05-01 16:42:56 +00:00
|
|
|
{
|
2015-04-23 14:34:06 +00:00
|
|
|
/*
|
|
|
|
|
* ioccelerator mode 2 commands should be aborted via the
|
2014-02-18 19:56:45 +00:00
|
|
|
* accelerated path, since RAID path is unaware of these commands,
|
2015-04-23 14:34:06 +00:00
|
|
|
* but not all underlying firmware can handle abort TMF.
|
|
|
|
|
* Change abort to physical device reset when abort TMF is unsupported.
|
2014-02-18 19:56:45 +00:00
|
|
|
*/
|
2015-04-23 14:34:06 +00:00
|
|
|
if (abort->cmd_type == CMD_IOACCEL2) {
|
2016-02-23 21:16:28 +00:00
|
|
|
if ((HPSATMF_IOACCEL_ENABLED & h->TMFSupportFlags) ||
|
|
|
|
|
dev->physical_device)
|
2015-04-23 14:34:06 +00:00
|
|
|
return hpsa_send_abort_ioaccel2(h, abort,
|
|
|
|
|
reply_queue);
|
|
|
|
|
else
|
2016-02-23 21:16:28 +00:00
|
|
|
return hpsa_send_reset_as_abort_ioaccel2(h,
|
|
|
|
|
dev->scsi3addr,
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
abort, reply_queue);
|
2015-04-23 14:34:06 +00:00
|
|
|
}
|
2016-02-23 21:16:28 +00:00
|
|
|
return hpsa_send_abort(h, dev->scsi3addr, abort, reply_queue);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
}
|
2014-02-18 19:56:45 +00:00
|
|
|
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
/* Find out which reply queue a command was meant to return on */
|
|
|
|
|
static int hpsa_extract_reply_queue(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
if (c->cmd_type == CMD_IOACCEL2)
|
|
|
|
|
return h->ioaccel2_cmd_pool[c->cmdindex].reply_queue;
|
|
|
|
|
return c->Header.ReplyQueue;
|
2012-05-01 16:42:56 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:32:06 +00:00
|
|
|
/*
|
|
|
|
|
* Limit concurrency of abort commands to prevent
|
|
|
|
|
* over-subscription of commands
|
|
|
|
|
*/
|
|
|
|
|
static inline int wait_for_available_abort_cmd(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
#define ABORT_CMD_WAIT_MSECS 5000
|
|
|
|
|
return !wait_event_timeout(h->abort_cmd_wait_queue,
|
|
|
|
|
atomic_dec_if_positive(&h->abort_cmds_available) >= 0,
|
|
|
|
|
msecs_to_jiffies(ABORT_CMD_WAIT_MSECS));
|
|
|
|
|
}
|
|
|
|
|
|
2012-05-01 16:42:51 +00:00
|
|
|
/* Send an abort for the specified command.
|
|
|
|
|
* If the device and controller support it,
|
|
|
|
|
* send a task abort request.
|
|
|
|
|
*/
|
|
|
|
|
static int hpsa_eh_abort_handler(struct scsi_cmnd *sc)
|
|
|
|
|
{
|
|
|
|
|
|
2015-04-23 14:34:16 +00:00
|
|
|
int rc;
|
2012-05-01 16:42:51 +00:00
|
|
|
struct ctlr_info *h;
|
|
|
|
|
struct hpsa_scsi_dev_t *dev;
|
|
|
|
|
struct CommandList *abort; /* pointer to command to be aborted */
|
|
|
|
|
struct scsi_cmnd *as; /* ptr to scsi cmd inside aborted command. */
|
|
|
|
|
char msg[256]; /* For debug messaging. */
|
|
|
|
|
int ml = 0;
|
2015-01-23 22:41:09 +00:00
|
|
|
__le32 tagupper, taglower;
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
int refcount, reply_queue;
|
|
|
|
|
|
|
|
|
|
if (sc == NULL)
|
|
|
|
|
return FAILED;
|
2012-05-01 16:42:51 +00:00
|
|
|
|
2015-04-23 14:32:06 +00:00
|
|
|
if (sc->device == NULL)
|
|
|
|
|
return FAILED;
|
|
|
|
|
|
2012-05-01 16:42:51 +00:00
|
|
|
/* Find the controller of the command to be aborted */
|
|
|
|
|
h = sdev_to_hba(sc->device);
|
2015-04-23 14:32:06 +00:00
|
|
|
if (h == NULL)
|
2012-05-01 16:42:51 +00:00
|
|
|
return FAILED;
|
|
|
|
|
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
/* Find the device of the command to be aborted */
|
|
|
|
|
dev = sc->device->hostdata;
|
|
|
|
|
if (!dev) {
|
|
|
|
|
dev_err(&h->pdev->dev, "%s FAILED, Device lookup failed.\n",
|
|
|
|
|
msg);
|
2015-01-23 22:44:24 +00:00
|
|
|
return FAILED;
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* If controller locked up, we can guarantee command won't complete */
|
|
|
|
|
if (lockup_detected(h)) {
|
|
|
|
|
hpsa_show_dev_msg(KERN_WARNING, h, dev,
|
|
|
|
|
"ABORT FAILED, lockup detected");
|
|
|
|
|
return FAILED;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* This is a good time to check if controller lockup has occurred */
|
|
|
|
|
if (detect_controller_lockup(h)) {
|
|
|
|
|
hpsa_show_dev_msg(KERN_WARNING, h, dev,
|
|
|
|
|
"ABORT FAILED, new lockup detected");
|
|
|
|
|
return FAILED;
|
|
|
|
|
}
|
2015-01-23 22:44:24 +00:00
|
|
|
|
2012-05-01 16:42:51 +00:00
|
|
|
/* Check that controller supports some kind of task abort */
|
|
|
|
|
if (!(HPSATMF_PHYS_TASK_ABORT & h->TMFSupportFlags) &&
|
|
|
|
|
!(HPSATMF_LOG_TASK_ABORT & h->TMFSupportFlags))
|
|
|
|
|
return FAILED;
|
|
|
|
|
|
|
|
|
|
memset(msg, 0, sizeof(msg));
|
2015-04-23 14:33:54 +00:00
|
|
|
ml += sprintf(msg+ml, "scsi %d:%d:%d:%llu %s %p",
|
2012-05-01 16:42:51 +00:00
|
|
|
h->scsi_host->host_no, sc->device->channel,
|
2015-04-23 14:31:55 +00:00
|
|
|
sc->device->id, sc->device->lun,
|
2015-04-23 14:33:54 +00:00
|
|
|
"Aborting command", sc);
|
2012-05-01 16:42:51 +00:00
|
|
|
|
|
|
|
|
/* Get SCSI command to be aborted */
|
|
|
|
|
abort = (struct CommandList *) sc->host_scribble;
|
|
|
|
|
if (abort == NULL) {
|
2015-01-23 22:43:35 +00:00
|
|
|
/* This can happen if the command already completed. */
|
|
|
|
|
return SUCCESS;
|
|
|
|
|
}
|
|
|
|
|
refcount = atomic_inc_return(&abort->refcount);
|
|
|
|
|
if (refcount == 1) { /* Command is done already. */
|
|
|
|
|
cmd_free(h, abort);
|
|
|
|
|
return SUCCESS;
|
2012-05-01 16:42:51 +00:00
|
|
|
}
|
2015-04-23 14:32:06 +00:00
|
|
|
|
|
|
|
|
/* Don't bother trying the abort if we know it won't work. */
|
|
|
|
|
if (abort->cmd_type != CMD_IOACCEL2 &&
|
|
|
|
|
abort->cmd_type != CMD_IOACCEL1 && !dev->supports_aborts) {
|
|
|
|
|
cmd_free(h, abort);
|
|
|
|
|
return FAILED;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:34:16 +00:00
|
|
|
/*
|
|
|
|
|
* Check that we're aborting the right command.
|
|
|
|
|
* It's possible the CommandList already completed and got re-used.
|
|
|
|
|
*/
|
|
|
|
|
if (abort->scsi_cmd != sc) {
|
|
|
|
|
cmd_free(h, abort);
|
|
|
|
|
return SUCCESS;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
abort->abort_pending = true;
|
2014-02-18 19:55:28 +00:00
|
|
|
hpsa_get_tag(h, abort, &taglower, &tagupper);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
reply_queue = hpsa_extract_reply_queue(h, abort);
|
2014-02-18 19:55:28 +00:00
|
|
|
ml += sprintf(msg+ml, "Tag:0x%08x:%08x ", tagupper, taglower);
|
2015-01-23 22:44:30 +00:00
|
|
|
as = abort->scsi_cmd;
|
2012-05-01 16:42:51 +00:00
|
|
|
if (as != NULL)
|
2015-04-23 14:33:54 +00:00
|
|
|
ml += sprintf(msg+ml,
|
|
|
|
|
"CDBLen: %d CDB: 0x%02x%02x... SN: 0x%lx ",
|
|
|
|
|
as->cmd_len, as->cmnd[0], as->cmnd[1],
|
|
|
|
|
as->serial_number);
|
|
|
|
|
dev_warn(&h->pdev->dev, "%s BEING SENT\n", msg);
|
2015-04-23 14:31:55 +00:00
|
|
|
hpsa_show_dev_msg(KERN_WARNING, h, dev, "Aborting command");
|
2015-04-23 14:33:54 +00:00
|
|
|
|
2012-05-01 16:42:51 +00:00
|
|
|
/*
|
|
|
|
|
* Command is in flight, or possibly already completed
|
|
|
|
|
* by the firmware (but not to the scsi mid layer) but we can't
|
|
|
|
|
* distinguish which. Send the abort down.
|
|
|
|
|
*/
|
2015-04-23 14:32:06 +00:00
|
|
|
if (wait_for_available_abort_cmd(h)) {
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
2015-04-23 14:33:54 +00:00
|
|
|
"%s FAILED, timeout waiting for an abort command to become available.\n",
|
|
|
|
|
msg);
|
2015-04-23 14:32:06 +00:00
|
|
|
cmd_free(h, abort);
|
|
|
|
|
return FAILED;
|
|
|
|
|
}
|
2016-02-23 21:16:28 +00:00
|
|
|
rc = hpsa_send_abort_both_ways(h, dev, abort, reply_queue);
|
2015-04-23 14:32:06 +00:00
|
|
|
atomic_inc(&h->abort_cmds_available);
|
|
|
|
|
wake_up_all(&h->abort_cmd_wait_queue);
|
2012-05-01 16:42:51 +00:00
|
|
|
if (rc != 0) {
|
2015-04-23 14:33:54 +00:00
|
|
|
dev_warn(&h->pdev->dev, "%s SENT, FAILED\n", msg);
|
2015-04-23 14:31:55 +00:00
|
|
|
hpsa_show_dev_msg(KERN_WARNING, h, dev,
|
2015-04-23 14:33:54 +00:00
|
|
|
"FAILED to abort command");
|
2015-01-23 22:43:35 +00:00
|
|
|
cmd_free(h, abort);
|
2012-05-01 16:42:51 +00:00
|
|
|
return FAILED;
|
|
|
|
|
}
|
2015-04-23 14:33:54 +00:00
|
|
|
dev_info(&h->pdev->dev, "%s SENT, SUCCESS\n", msg);
|
2015-04-23 14:35:22 +00:00
|
|
|
wait_event(h->event_sync_wait_queue,
|
2015-04-23 14:34:16 +00:00
|
|
|
abort->scsi_cmd != sc || lockup_detected(h));
|
2015-01-23 22:43:35 +00:00
|
|
|
cmd_free(h, abort);
|
2015-04-23 14:34:16 +00:00
|
|
|
return !lockup_detected(h) ? SUCCESS : FAILED;
|
2012-05-01 16:42:51 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:35:04 +00:00
|
|
|
/*
|
|
|
|
|
* For operations with an associated SCSI command, a command block is allocated
|
|
|
|
|
* at init, and managed by cmd_tagged_alloc() and cmd_tagged_free() using the
|
|
|
|
|
* block request tag as an index into a table of entries. cmd_tagged_free() is
|
|
|
|
|
* the complement, although cmd_free() may be called instead.
|
|
|
|
|
*/
|
|
|
|
|
static struct CommandList *cmd_tagged_alloc(struct ctlr_info *h,
|
|
|
|
|
struct scsi_cmnd *scmd)
|
|
|
|
|
{
|
|
|
|
|
int idx = hpsa_get_cmd_index(scmd);
|
|
|
|
|
struct CommandList *c = h->cmd_pool + idx;
|
|
|
|
|
|
|
|
|
|
if (idx < HPSA_NRESERVED_CMDS || idx >= h->nr_cmds) {
|
|
|
|
|
dev_err(&h->pdev->dev, "Bad block tag: %d not in [%d..%d]\n",
|
|
|
|
|
idx, HPSA_NRESERVED_CMDS, h->nr_cmds - 1);
|
|
|
|
|
/* The index value comes from the block layer, so if it's out of
|
|
|
|
|
* bounds, it's probably not our bug.
|
|
|
|
|
*/
|
|
|
|
|
BUG();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
atomic_inc(&c->refcount);
|
|
|
|
|
if (unlikely(!hpsa_is_cmd_idle(c))) {
|
|
|
|
|
/*
|
|
|
|
|
* We expect that the SCSI layer will hand us a unique tag
|
|
|
|
|
* value. Thus, there should never be a collision here between
|
|
|
|
|
* two requests...because if the selected command isn't idle
|
|
|
|
|
* then someone is going to be very disappointed.
|
|
|
|
|
*/
|
|
|
|
|
dev_err(&h->pdev->dev,
|
|
|
|
|
"tag collision (tag=%d) in cmd_tagged_alloc().\n",
|
|
|
|
|
idx);
|
|
|
|
|
if (c->scsi_cmd != NULL)
|
|
|
|
|
scsi_print_command(c->scsi_cmd);
|
|
|
|
|
scsi_print_command(scmd);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
hpsa_cmd_partial_init(h, idx, c);
|
|
|
|
|
return c;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void cmd_tagged_free(struct ctlr_info *h, struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
/*
|
|
|
|
|
* Release our reference to the block. We don't need to do anything
|
|
|
|
|
* else to free it, because it is accessed by index. (There's no point
|
|
|
|
|
* in checking the result of the decrement, since we cannot guarantee
|
|
|
|
|
* that there isn't a concurrent abort which is also accessing it.)
|
|
|
|
|
*/
|
|
|
|
|
(void)atomic_dec(&c->refcount);
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/*
|
|
|
|
|
* For operations that cannot sleep, a command block is allocated at init,
|
|
|
|
|
* and managed by cmd_alloc() and cmd_free() using a simple bitmap to track
|
|
|
|
|
* which ones are free or in use. Lock must be held when calling this.
|
|
|
|
|
* cmd_free() is the complement.
|
2015-04-23 14:33:38 +00:00
|
|
|
* This function never gives up and returns NULL. If it hangs,
|
|
|
|
|
* another thread must call cmd_free() to free some tags.
|
2009-12-08 22:09:11 +00:00
|
|
|
*/
|
2015-01-23 22:43:35 +00:00
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
static struct CommandList *cmd_alloc(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
struct CommandList *c;
|
2015-04-23 14:32:32 +00:00
|
|
|
int refcount, i;
|
2015-04-23 14:35:04 +00:00
|
|
|
int offset = 0;
|
2014-11-14 23:27:29 +00:00
|
|
|
|
2015-01-23 22:43:41 +00:00
|
|
|
/*
|
|
|
|
|
* There is some *extremely* small but non-zero chance that that
|
2014-11-14 23:27:29 +00:00
|
|
|
* multiple threads could get in here, and one thread could
|
|
|
|
|
* be scanning through the list of bits looking for a free
|
|
|
|
|
* one, but the free ones are always behind him, and other
|
|
|
|
|
* threads sneak in behind him and eat them before he can
|
|
|
|
|
* get to them, so that while there is always a free one, a
|
|
|
|
|
* very unlucky thread might be starved anyway, never able to
|
|
|
|
|
* beat the other threads. In reality, this happens so
|
|
|
|
|
* infrequently as to be indistinguishable from never.
|
2015-04-23 14:35:04 +00:00
|
|
|
*
|
|
|
|
|
* Note that we start allocating commands before the SCSI host structure
|
|
|
|
|
* is initialized. Since the search starts at bit zero, this
|
|
|
|
|
* all works, since we have at least one command structure available;
|
|
|
|
|
* however, it means that the structures with the low indexes have to be
|
|
|
|
|
* reserved for driver-initiated requests, while requests from the block
|
|
|
|
|
* layer will use the higher indexes.
|
2014-11-14 23:27:29 +00:00
|
|
|
*/
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-01-23 22:43:35 +00:00
|
|
|
for (;;) {
|
2015-04-23 14:35:04 +00:00
|
|
|
i = find_next_zero_bit(h->cmd_pool_bits,
|
|
|
|
|
HPSA_NRESERVED_CMDS,
|
|
|
|
|
offset);
|
|
|
|
|
if (unlikely(i >= HPSA_NRESERVED_CMDS)) {
|
2015-01-23 22:43:35 +00:00
|
|
|
offset = 0;
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
c = h->cmd_pool + i;
|
|
|
|
|
refcount = atomic_inc_return(&c->refcount);
|
|
|
|
|
if (unlikely(refcount > 1)) {
|
|
|
|
|
cmd_free(h, c); /* already in use */
|
2015-04-23 14:35:04 +00:00
|
|
|
offset = (i + 1) % HPSA_NRESERVED_CMDS;
|
2015-01-23 22:43:35 +00:00
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
set_bit(i & (BITS_PER_LONG - 1),
|
|
|
|
|
h->cmd_pool_bits + (i / BITS_PER_LONG));
|
|
|
|
|
break; /* it's ours now. */
|
|
|
|
|
}
|
2015-04-23 14:32:32 +00:00
|
|
|
hpsa_cmd_partial_init(h, i, c);
|
2009-12-08 22:09:11 +00:00
|
|
|
return c;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:35:04 +00:00
|
|
|
/*
|
|
|
|
|
* This is the complementary operation to cmd_alloc(). Note, however, in some
|
|
|
|
|
* corner cases it may also be used to free blocks allocated by
|
|
|
|
|
* cmd_tagged_alloc() in which case the ref-count decrement does the trick and
|
|
|
|
|
* the clear-bit is harmless.
|
|
|
|
|
*/
|
2009-12-08 22:09:11 +00:00
|
|
|
static void cmd_free(struct ctlr_info *h, struct CommandList *c)
|
|
|
|
|
{
|
2015-01-23 22:43:35 +00:00
|
|
|
if (atomic_dec_and_test(&c->refcount)) {
|
|
|
|
|
int i;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-01-23 22:43:35 +00:00
|
|
|
i = c - h->cmd_pool;
|
|
|
|
|
clear_bit(i & (BITS_PER_LONG - 1),
|
|
|
|
|
h->cmd_pool_bits + (i / BITS_PER_LONG));
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
|
|
2014-11-14 23:26:27 +00:00
|
|
|
static int hpsa_ioctl32_passthru(struct scsi_device *dev, int cmd,
|
|
|
|
|
void __user *arg)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
|
|
|
|
IOCTL32_Command_struct __user *arg32 =
|
|
|
|
|
(IOCTL32_Command_struct __user *) arg;
|
|
|
|
|
IOCTL_Command_struct arg64;
|
|
|
|
|
IOCTL_Command_struct __user *p = compat_alloc_user_space(sizeof(arg64));
|
|
|
|
|
int err;
|
|
|
|
|
u32 cp;
|
|
|
|
|
|
2011-01-07 16:55:53 +00:00
|
|
|
memset(&arg64, 0, sizeof(arg64));
|
2009-12-08 22:09:11 +00:00
|
|
|
err = 0;
|
|
|
|
|
err |= copy_from_user(&arg64.LUN_info, &arg32->LUN_info,
|
|
|
|
|
sizeof(arg64.LUN_info));
|
|
|
|
|
err |= copy_from_user(&arg64.Request, &arg32->Request,
|
|
|
|
|
sizeof(arg64.Request));
|
|
|
|
|
err |= copy_from_user(&arg64.error_info, &arg32->error_info,
|
|
|
|
|
sizeof(arg64.error_info));
|
|
|
|
|
err |= get_user(arg64.buf_size, &arg32->buf_size);
|
|
|
|
|
err |= get_user(cp, &arg32->buf);
|
|
|
|
|
arg64.buf = compat_ptr(cp);
|
|
|
|
|
err |= copy_to_user(p, &arg64, sizeof(arg64));
|
|
|
|
|
|
|
|
|
|
if (err)
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
2014-11-14 23:26:27 +00:00
|
|
|
err = hpsa_ioctl(dev, CCISS_PASSTHRU, p);
|
2009-12-08 22:09:11 +00:00
|
|
|
if (err)
|
|
|
|
|
return err;
|
|
|
|
|
err |= copy_in_user(&arg32->error_info, &p->error_info,
|
|
|
|
|
sizeof(arg32->error_info));
|
|
|
|
|
if (err)
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int hpsa_ioctl32_big_passthru(struct scsi_device *dev,
|
2014-11-14 23:26:27 +00:00
|
|
|
int cmd, void __user *arg)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
|
|
|
|
BIG_IOCTL32_Command_struct __user *arg32 =
|
|
|
|
|
(BIG_IOCTL32_Command_struct __user *) arg;
|
|
|
|
|
BIG_IOCTL_Command_struct arg64;
|
|
|
|
|
BIG_IOCTL_Command_struct __user *p =
|
|
|
|
|
compat_alloc_user_space(sizeof(arg64));
|
|
|
|
|
int err;
|
|
|
|
|
u32 cp;
|
|
|
|
|
|
2011-01-07 16:55:53 +00:00
|
|
|
memset(&arg64, 0, sizeof(arg64));
|
2009-12-08 22:09:11 +00:00
|
|
|
err = 0;
|
|
|
|
|
err |= copy_from_user(&arg64.LUN_info, &arg32->LUN_info,
|
|
|
|
|
sizeof(arg64.LUN_info));
|
|
|
|
|
err |= copy_from_user(&arg64.Request, &arg32->Request,
|
|
|
|
|
sizeof(arg64.Request));
|
|
|
|
|
err |= copy_from_user(&arg64.error_info, &arg32->error_info,
|
|
|
|
|
sizeof(arg64.error_info));
|
|
|
|
|
err |= get_user(arg64.buf_size, &arg32->buf_size);
|
|
|
|
|
err |= get_user(arg64.malloc_size, &arg32->malloc_size);
|
|
|
|
|
err |= get_user(cp, &arg32->buf);
|
|
|
|
|
arg64.buf = compat_ptr(cp);
|
|
|
|
|
err |= copy_to_user(p, &arg64, sizeof(arg64));
|
|
|
|
|
|
|
|
|
|
if (err)
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
2014-11-14 23:26:27 +00:00
|
|
|
err = hpsa_ioctl(dev, CCISS_BIG_PASSTHRU, p);
|
2009-12-08 22:09:11 +00:00
|
|
|
if (err)
|
|
|
|
|
return err;
|
|
|
|
|
err |= copy_in_user(&arg32->error_info, &p->error_info,
|
|
|
|
|
sizeof(arg32->error_info));
|
|
|
|
|
if (err)
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
return err;
|
|
|
|
|
}
|
2010-02-04 14:43:51 +00:00
|
|
|
|
2014-11-14 23:26:27 +00:00
|
|
|
static int hpsa_compat_ioctl(struct scsi_device *dev, int cmd, void __user *arg)
|
2010-02-04 14:43:51 +00:00
|
|
|
{
|
|
|
|
|
switch (cmd) {
|
|
|
|
|
case CCISS_GETPCIINFO:
|
|
|
|
|
case CCISS_GETINTINFO:
|
|
|
|
|
case CCISS_SETINTINFO:
|
|
|
|
|
case CCISS_GETNODENAME:
|
|
|
|
|
case CCISS_SETNODENAME:
|
|
|
|
|
case CCISS_GETHEARTBEAT:
|
|
|
|
|
case CCISS_GETBUSTYPES:
|
|
|
|
|
case CCISS_GETFIRMVER:
|
|
|
|
|
case CCISS_GETDRIVVER:
|
|
|
|
|
case CCISS_REVALIDVOLS:
|
|
|
|
|
case CCISS_DEREGDISK:
|
|
|
|
|
case CCISS_REGNEWDISK:
|
|
|
|
|
case CCISS_REGNEWD:
|
|
|
|
|
case CCISS_RESCANDISK:
|
|
|
|
|
case CCISS_GETLUNINFO:
|
|
|
|
|
return hpsa_ioctl(dev, cmd, arg);
|
|
|
|
|
|
|
|
|
|
case CCISS_PASSTHRU32:
|
|
|
|
|
return hpsa_ioctl32_passthru(dev, cmd, arg);
|
|
|
|
|
case CCISS_BIG_PASSTHRU32:
|
|
|
|
|
return hpsa_ioctl32_big_passthru(dev, cmd, arg);
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
return -ENOIOCTLCMD;
|
|
|
|
|
}
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
static int hpsa_getpciinfo_ioctl(struct ctlr_info *h, void __user *argp)
|
|
|
|
|
{
|
|
|
|
|
struct hpsa_pci_info pciinfo;
|
|
|
|
|
|
|
|
|
|
if (!argp)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
pciinfo.domain = pci_domain_nr(h->pdev->bus);
|
|
|
|
|
pciinfo.bus = h->pdev->bus->number;
|
|
|
|
|
pciinfo.dev_fn = h->pdev->devfn;
|
|
|
|
|
pciinfo.board_id = h->board_id;
|
|
|
|
|
if (copy_to_user(argp, &pciinfo, sizeof(pciinfo)))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int hpsa_getdrivver_ioctl(struct ctlr_info *h, void __user *argp)
|
|
|
|
|
{
|
|
|
|
|
DriverVer_type DriverVer;
|
|
|
|
|
unsigned char vmaj, vmin, vsubmin;
|
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
|
|
rc = sscanf(HPSA_DRIVER_VERSION, "%hhu.%hhu.%hhu",
|
|
|
|
|
&vmaj, &vmin, &vsubmin);
|
|
|
|
|
if (rc != 3) {
|
|
|
|
|
dev_info(&h->pdev->dev, "driver version string '%s' "
|
|
|
|
|
"unrecognized.", HPSA_DRIVER_VERSION);
|
|
|
|
|
vmaj = 0;
|
|
|
|
|
vmin = 0;
|
|
|
|
|
vsubmin = 0;
|
|
|
|
|
}
|
|
|
|
|
DriverVer = (vmaj << 16) | (vmin << 8) | vsubmin;
|
|
|
|
|
if (!argp)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
if (copy_to_user(argp, &DriverVer, sizeof(DriverVer_type)))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int hpsa_passthru_ioctl(struct ctlr_info *h, void __user *argp)
|
|
|
|
|
{
|
|
|
|
|
IOCTL_Command_struct iocommand;
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
char *buff = NULL;
|
2014-11-14 23:26:59 +00:00
|
|
|
u64 temp64;
|
2013-02-20 17:24:52 +00:00
|
|
|
int rc = 0;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
if (!argp)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
if (!capable(CAP_SYS_RAWIO))
|
|
|
|
|
return -EPERM;
|
|
|
|
|
if (copy_from_user(&iocommand, argp, sizeof(iocommand)))
|
|
|
|
|
return -EFAULT;
|
|
|
|
|
if ((iocommand.buf_size < 1) &&
|
|
|
|
|
(iocommand.Request.Type.Direction != XFER_NONE)) {
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
|
|
|
|
if (iocommand.buf_size > 0) {
|
|
|
|
|
buff = kmalloc(iocommand.buf_size, GFP_KERNEL);
|
|
|
|
|
if (buff == NULL)
|
2015-04-23 14:33:43 +00:00
|
|
|
return -ENOMEM;
|
2014-05-29 15:52:41 +00:00
|
|
|
if (iocommand.Request.Type.Direction & XFER_WRITE) {
|
2011-01-06 20:47:48 +00:00
|
|
|
/* Copy the data into the buffer we created */
|
|
|
|
|
if (copy_from_user(buff, iocommand.buf,
|
|
|
|
|
iocommand.buf_size)) {
|
2013-02-20 17:24:52 +00:00
|
|
|
rc = -EFAULT;
|
|
|
|
|
goto out_kfree;
|
2011-01-06 20:47:48 +00:00
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
memset(buff, 0, iocommand.buf_size);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
2011-01-06 20:47:48 +00:00
|
|
|
}
|
2015-01-23 22:43:04 +00:00
|
|
|
c = cmd_alloc(h);
|
2015-04-23 14:33:38 +00:00
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/* Fill in the command type */
|
|
|
|
|
c->cmd_type = CMD_IOCTL_PEND;
|
2015-04-23 14:34:16 +00:00
|
|
|
c->scsi_cmd = SCSI_CMD_BUSY;
|
2009-12-08 22:09:11 +00:00
|
|
|
/* Fill in Command Header */
|
|
|
|
|
c->Header.ReplyQueue = 0; /* unused in simple mode */
|
|
|
|
|
if (iocommand.buf_size > 0) { /* buffer to fill */
|
|
|
|
|
c->Header.SGList = 1;
|
2014-11-14 23:26:59 +00:00
|
|
|
c->Header.SGTotal = cpu_to_le16(1);
|
2009-12-08 22:09:11 +00:00
|
|
|
} else { /* no buffers to fill */
|
|
|
|
|
c->Header.SGList = 0;
|
2014-11-14 23:26:59 +00:00
|
|
|
c->Header.SGTotal = cpu_to_le16(0);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
memcpy(&c->Header.LUN, &iocommand.LUN_info, sizeof(c->Header.LUN));
|
|
|
|
|
|
|
|
|
|
/* Fill in Request block */
|
|
|
|
|
memcpy(&c->Request, &iocommand.Request,
|
|
|
|
|
sizeof(c->Request));
|
|
|
|
|
|
|
|
|
|
/* Fill in the scatter gather information */
|
|
|
|
|
if (iocommand.buf_size > 0) {
|
2014-11-14 23:26:59 +00:00
|
|
|
temp64 = pci_map_single(h->pdev, buff,
|
2009-12-08 22:09:11 +00:00
|
|
|
iocommand.buf_size, PCI_DMA_BIDIRECTIONAL);
|
2014-11-14 23:26:59 +00:00
|
|
|
if (dma_mapping_error(&h->pdev->dev, (dma_addr_t) temp64)) {
|
|
|
|
|
c->SG[0].Addr = cpu_to_le64(0);
|
|
|
|
|
c->SG[0].Len = cpu_to_le32(0);
|
2013-02-20 17:24:57 +00:00
|
|
|
rc = -ENOMEM;
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
2014-11-14 23:26:59 +00:00
|
|
|
c->SG[0].Addr = cpu_to_le64(temp64);
|
|
|
|
|
c->SG[0].Len = cpu_to_le32(iocommand.buf_size);
|
|
|
|
|
c->SG[0].Ext = cpu_to_le32(HPSA_SG_LAST); /* not chaining */
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
2016-04-27 22:13:51 +00:00
|
|
|
rc = hpsa_scsi_do_simple_cmd(h, c, DEFAULT_REPLY_QUEUE,
|
|
|
|
|
DEFAULT_TIMEOUT);
|
2011-06-03 14:57:29 +00:00
|
|
|
if (iocommand.buf_size > 0)
|
|
|
|
|
hpsa_pci_unmap(h->pdev, c, 1, PCI_DMA_BIDIRECTIONAL);
|
2009-12-08 22:09:11 +00:00
|
|
|
check_ioctl_unit_attention(h, c);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
if (rc) {
|
|
|
|
|
rc = -EIO;
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
/* Copy the error information out */
|
|
|
|
|
memcpy(&iocommand.error_info, c->err_info,
|
|
|
|
|
sizeof(iocommand.error_info));
|
|
|
|
|
if (copy_to_user(argp, &iocommand, sizeof(iocommand))) {
|
2013-02-20 17:24:52 +00:00
|
|
|
rc = -EFAULT;
|
|
|
|
|
goto out;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
2014-05-29 15:52:41 +00:00
|
|
|
if ((iocommand.Request.Type.Direction & XFER_READ) &&
|
2011-01-06 20:47:48 +00:00
|
|
|
iocommand.buf_size > 0) {
|
2009-12-08 22:09:11 +00:00
|
|
|
/* Copy the data out of the buffer we created */
|
|
|
|
|
if (copy_to_user(iocommand.buf, buff, iocommand.buf_size)) {
|
2013-02-20 17:24:52 +00:00
|
|
|
rc = -EFAULT;
|
|
|
|
|
goto out;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
}
|
2013-02-20 17:24:52 +00:00
|
|
|
out:
|
2015-01-23 22:43:04 +00:00
|
|
|
cmd_free(h, c);
|
2013-02-20 17:24:52 +00:00
|
|
|
out_kfree:
|
|
|
|
|
kfree(buff);
|
|
|
|
|
return rc;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int hpsa_big_passthru_ioctl(struct ctlr_info *h, void __user *argp)
|
|
|
|
|
{
|
|
|
|
|
BIG_IOCTL_Command_struct *ioc;
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
unsigned char **buff = NULL;
|
|
|
|
|
int *buff_size = NULL;
|
2014-11-14 23:26:59 +00:00
|
|
|
u64 temp64;
|
2009-12-08 22:09:11 +00:00
|
|
|
BYTE sg_used = 0;
|
|
|
|
|
int status = 0;
|
2010-02-04 14:41:33 +00:00
|
|
|
u32 left;
|
|
|
|
|
u32 sz;
|
2009-12-08 22:09:11 +00:00
|
|
|
BYTE __user *data_ptr;
|
|
|
|
|
|
|
|
|
|
if (!argp)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
if (!capable(CAP_SYS_RAWIO))
|
|
|
|
|
return -EPERM;
|
|
|
|
|
ioc = (BIG_IOCTL_Command_struct *)
|
|
|
|
|
kmalloc(sizeof(*ioc), GFP_KERNEL);
|
|
|
|
|
if (!ioc) {
|
|
|
|
|
status = -ENOMEM;
|
|
|
|
|
goto cleanup1;
|
|
|
|
|
}
|
|
|
|
|
if (copy_from_user(ioc, argp, sizeof(*ioc))) {
|
|
|
|
|
status = -EFAULT;
|
|
|
|
|
goto cleanup1;
|
|
|
|
|
}
|
|
|
|
|
if ((ioc->buf_size < 1) &&
|
|
|
|
|
(ioc->Request.Type.Direction != XFER_NONE)) {
|
|
|
|
|
status = -EINVAL;
|
|
|
|
|
goto cleanup1;
|
|
|
|
|
}
|
|
|
|
|
/* Check kmalloc limits using all SGs */
|
|
|
|
|
if (ioc->malloc_size > MAX_KMALLOC_SIZE) {
|
|
|
|
|
status = -EINVAL;
|
|
|
|
|
goto cleanup1;
|
|
|
|
|
}
|
2012-01-19 20:00:48 +00:00
|
|
|
if (ioc->buf_size > ioc->malloc_size * SG_ENTRIES_IN_CMD) {
|
2009-12-08 22:09:11 +00:00
|
|
|
status = -EINVAL;
|
|
|
|
|
goto cleanup1;
|
|
|
|
|
}
|
2012-01-19 20:00:48 +00:00
|
|
|
buff = kzalloc(SG_ENTRIES_IN_CMD * sizeof(char *), GFP_KERNEL);
|
2009-12-08 22:09:11 +00:00
|
|
|
if (!buff) {
|
|
|
|
|
status = -ENOMEM;
|
|
|
|
|
goto cleanup1;
|
|
|
|
|
}
|
2012-01-19 20:00:48 +00:00
|
|
|
buff_size = kmalloc(SG_ENTRIES_IN_CMD * sizeof(int), GFP_KERNEL);
|
2009-12-08 22:09:11 +00:00
|
|
|
if (!buff_size) {
|
|
|
|
|
status = -ENOMEM;
|
|
|
|
|
goto cleanup1;
|
|
|
|
|
}
|
|
|
|
|
left = ioc->buf_size;
|
|
|
|
|
data_ptr = ioc->buf;
|
|
|
|
|
while (left) {
|
|
|
|
|
sz = (left > ioc->malloc_size) ? ioc->malloc_size : left;
|
|
|
|
|
buff_size[sg_used] = sz;
|
|
|
|
|
buff[sg_used] = kmalloc(sz, GFP_KERNEL);
|
|
|
|
|
if (buff[sg_used] == NULL) {
|
|
|
|
|
status = -ENOMEM;
|
|
|
|
|
goto cleanup1;
|
|
|
|
|
}
|
2014-05-29 15:52:41 +00:00
|
|
|
if (ioc->Request.Type.Direction & XFER_WRITE) {
|
2009-12-08 22:09:11 +00:00
|
|
|
if (copy_from_user(buff[sg_used], data_ptr, sz)) {
|
2014-07-03 15:18:03 +00:00
|
|
|
status = -EFAULT;
|
2009-12-08 22:09:11 +00:00
|
|
|
goto cleanup1;
|
|
|
|
|
}
|
|
|
|
|
} else
|
|
|
|
|
memset(buff[sg_used], 0, sz);
|
|
|
|
|
left -= sz;
|
|
|
|
|
data_ptr += sz;
|
|
|
|
|
sg_used++;
|
|
|
|
|
}
|
2015-01-23 22:43:04 +00:00
|
|
|
c = cmd_alloc(h);
|
2015-04-23 14:33:38 +00:00
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
c->cmd_type = CMD_IOCTL_PEND;
|
2015-04-23 14:34:16 +00:00
|
|
|
c->scsi_cmd = SCSI_CMD_BUSY;
|
2009-12-08 22:09:11 +00:00
|
|
|
c->Header.ReplyQueue = 0;
|
2014-11-14 23:26:59 +00:00
|
|
|
c->Header.SGList = (u8) sg_used;
|
|
|
|
|
c->Header.SGTotal = cpu_to_le16(sg_used);
|
2009-12-08 22:09:11 +00:00
|
|
|
memcpy(&c->Header.LUN, &ioc->LUN_info, sizeof(c->Header.LUN));
|
|
|
|
|
memcpy(&c->Request, &ioc->Request, sizeof(c->Request));
|
|
|
|
|
if (ioc->buf_size > 0) {
|
|
|
|
|
int i;
|
|
|
|
|
for (i = 0; i < sg_used; i++) {
|
2014-11-14 23:26:59 +00:00
|
|
|
temp64 = pci_map_single(h->pdev, buff[i],
|
2009-12-08 22:09:11 +00:00
|
|
|
buff_size[i], PCI_DMA_BIDIRECTIONAL);
|
2014-11-14 23:26:59 +00:00
|
|
|
if (dma_mapping_error(&h->pdev->dev,
|
|
|
|
|
(dma_addr_t) temp64)) {
|
|
|
|
|
c->SG[i].Addr = cpu_to_le64(0);
|
|
|
|
|
c->SG[i].Len = cpu_to_le32(0);
|
2013-02-20 17:24:57 +00:00
|
|
|
hpsa_pci_unmap(h->pdev, c, i,
|
|
|
|
|
PCI_DMA_BIDIRECTIONAL);
|
|
|
|
|
status = -ENOMEM;
|
2013-09-23 18:33:51 +00:00
|
|
|
goto cleanup0;
|
2013-02-20 17:24:57 +00:00
|
|
|
}
|
2014-11-14 23:26:59 +00:00
|
|
|
c->SG[i].Addr = cpu_to_le64(temp64);
|
|
|
|
|
c->SG[i].Len = cpu_to_le32(buff_size[i]);
|
|
|
|
|
c->SG[i].Ext = cpu_to_le32(0);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
2014-11-14 23:26:59 +00:00
|
|
|
c->SG[--i].Ext = cpu_to_le32(HPSA_SG_LAST);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
2016-04-27 22:13:51 +00:00
|
|
|
status = hpsa_scsi_do_simple_cmd(h, c, DEFAULT_REPLY_QUEUE,
|
|
|
|
|
DEFAULT_TIMEOUT);
|
2011-01-06 20:47:48 +00:00
|
|
|
if (sg_used)
|
|
|
|
|
hpsa_pci_unmap(h->pdev, c, sg_used, PCI_DMA_BIDIRECTIONAL);
|
2009-12-08 22:09:11 +00:00
|
|
|
check_ioctl_unit_attention(h, c);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
if (status) {
|
|
|
|
|
status = -EIO;
|
|
|
|
|
goto cleanup0;
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/* Copy the error information out */
|
|
|
|
|
memcpy(&ioc->error_info, c->err_info, sizeof(ioc->error_info));
|
|
|
|
|
if (copy_to_user(argp, ioc, sizeof(*ioc))) {
|
|
|
|
|
status = -EFAULT;
|
2013-09-23 18:33:51 +00:00
|
|
|
goto cleanup0;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
2014-05-29 15:52:41 +00:00
|
|
|
if ((ioc->Request.Type.Direction & XFER_READ) && ioc->buf_size > 0) {
|
2015-01-23 22:41:09 +00:00
|
|
|
int i;
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/* Copy the data out of the buffer we created */
|
|
|
|
|
BYTE __user *ptr = ioc->buf;
|
|
|
|
|
for (i = 0; i < sg_used; i++) {
|
|
|
|
|
if (copy_to_user(ptr, buff[i], buff_size[i])) {
|
|
|
|
|
status = -EFAULT;
|
2013-09-23 18:33:51 +00:00
|
|
|
goto cleanup0;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
ptr += buff_size[i];
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
status = 0;
|
2013-09-23 18:33:51 +00:00
|
|
|
cleanup0:
|
2015-01-23 22:43:04 +00:00
|
|
|
cmd_free(h, c);
|
2009-12-08 22:09:11 +00:00
|
|
|
cleanup1:
|
|
|
|
|
if (buff) {
|
2015-01-23 22:41:09 +00:00
|
|
|
int i;
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
for (i = 0; i < sg_used; i++)
|
|
|
|
|
kfree(buff[i]);
|
|
|
|
|
kfree(buff);
|
|
|
|
|
}
|
|
|
|
|
kfree(buff_size);
|
|
|
|
|
kfree(ioc);
|
|
|
|
|
return status;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void check_ioctl_unit_attention(struct ctlr_info *h,
|
|
|
|
|
struct CommandList *c)
|
|
|
|
|
{
|
|
|
|
|
if (c->err_info->CommandStatus == CMD_TARGET_STATUS &&
|
|
|
|
|
c->err_info->ScsiStatus != SAM_STAT_CHECK_CONDITION)
|
|
|
|
|
(void) check_for_unit_attention(h, c);
|
|
|
|
|
}
|
2013-09-23 18:34:12 +00:00
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/*
|
|
|
|
|
* ioctl
|
|
|
|
|
*/
|
2014-11-14 23:26:27 +00:00
|
|
|
static int hpsa_ioctl(struct scsi_device *dev, int cmd, void __user *arg)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
void __user *argp = (void __user *)arg;
|
2013-09-23 18:34:12 +00:00
|
|
|
int rc;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
h = sdev_to_hba(dev);
|
|
|
|
|
|
|
|
|
|
switch (cmd) {
|
|
|
|
|
case CCISS_DEREGDISK:
|
|
|
|
|
case CCISS_REGNEWDISK:
|
|
|
|
|
case CCISS_REGNEWD:
|
2010-02-04 14:43:16 +00:00
|
|
|
hpsa_scan_start(h->scsi_host);
|
2009-12-08 22:09:11 +00:00
|
|
|
return 0;
|
|
|
|
|
case CCISS_GETPCIINFO:
|
|
|
|
|
return hpsa_getpciinfo_ioctl(h, argp);
|
|
|
|
|
case CCISS_GETDRIVVER:
|
|
|
|
|
return hpsa_getdrivver_ioctl(h, argp);
|
|
|
|
|
case CCISS_PASSTHRU:
|
2015-01-23 22:43:46 +00:00
|
|
|
if (atomic_dec_if_positive(&h->passthru_cmds_avail) < 0)
|
2013-09-23 18:34:12 +00:00
|
|
|
return -EAGAIN;
|
|
|
|
|
rc = hpsa_passthru_ioctl(h, argp);
|
2015-01-23 22:43:46 +00:00
|
|
|
atomic_inc(&h->passthru_cmds_avail);
|
2013-09-23 18:34:12 +00:00
|
|
|
return rc;
|
2009-12-08 22:09:11 +00:00
|
|
|
case CCISS_BIG_PASSTHRU:
|
2015-01-23 22:43:46 +00:00
|
|
|
if (atomic_dec_if_positive(&h->passthru_cmds_avail) < 0)
|
2013-09-23 18:34:12 +00:00
|
|
|
return -EAGAIN;
|
|
|
|
|
rc = hpsa_big_passthru_ioctl(h, argp);
|
2015-01-23 22:43:46 +00:00
|
|
|
atomic_inc(&h->passthru_cmds_avail);
|
2013-09-23 18:34:12 +00:00
|
|
|
return rc;
|
2009-12-08 22:09:11 +00:00
|
|
|
default:
|
|
|
|
|
return -ENOTTY;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:33:38 +00:00
|
|
|
static void hpsa_send_host_reset(struct ctlr_info *h, unsigned char *scsi3addr,
|
2012-12-21 21:08:55 +00:00
|
|
|
u8 reset_type)
|
2011-05-03 19:59:51 +00:00
|
|
|
{
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
|
|
|
|
|
c = cmd_alloc(h);
|
2015-04-23 14:33:38 +00:00
|
|
|
|
2013-02-20 17:24:41 +00:00
|
|
|
/* fill_cmd can't fail here, no data buffer to map */
|
|
|
|
|
(void) fill_cmd(c, HPSA_DEVICE_RESET_MSG, h, NULL, 0, 0,
|
2011-05-03 19:59:51 +00:00
|
|
|
RAID_CTLR_LUNID, TYPE_MSG);
|
|
|
|
|
c->Request.CDB[1] = reset_type; /* fill_cmd defaults to target reset */
|
|
|
|
|
c->waiting = NULL;
|
|
|
|
|
enqueue_cmd_and_start_io(h, c);
|
|
|
|
|
/* Don't wait for completion, the reset won't complete. Don't free
|
|
|
|
|
* the command either. This is the last command we will send before
|
|
|
|
|
* re-initializing everything, so it doesn't matter and won't leak.
|
|
|
|
|
*/
|
2015-04-23 14:33:38 +00:00
|
|
|
return;
|
2011-05-03 19:59:51 +00:00
|
|
|
}
|
|
|
|
|
|
2013-02-20 17:24:41 +00:00
|
|
|
static int fill_cmd(struct CommandList *c, u8 cmd, struct ctlr_info *h,
|
2014-02-18 19:57:11 +00:00
|
|
|
void *buff, size_t size, u16 page_code, unsigned char *scsi3addr,
|
2009-12-08 22:09:11 +00:00
|
|
|
int cmd_type)
|
|
|
|
|
{
|
|
|
|
|
int pci_dir = XFER_NONE;
|
2015-04-23 14:32:06 +00:00
|
|
|
u64 tag; /* for commands to be aborted */
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
c->cmd_type = CMD_IOCTL_PEND;
|
2015-04-23 14:34:16 +00:00
|
|
|
c->scsi_cmd = SCSI_CMD_BUSY;
|
2009-12-08 22:09:11 +00:00
|
|
|
c->Header.ReplyQueue = 0;
|
|
|
|
|
if (buff != NULL && size > 0) {
|
|
|
|
|
c->Header.SGList = 1;
|
2014-11-14 23:26:59 +00:00
|
|
|
c->Header.SGTotal = cpu_to_le16(1);
|
2009-12-08 22:09:11 +00:00
|
|
|
} else {
|
|
|
|
|
c->Header.SGList = 0;
|
2014-11-14 23:26:59 +00:00
|
|
|
c->Header.SGTotal = cpu_to_le16(0);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
memcpy(c->Header.LUN.LunAddrBytes, scsi3addr, 8);
|
|
|
|
|
|
|
|
|
|
if (cmd_type == TYPE_CMD) {
|
|
|
|
|
switch (cmd) {
|
|
|
|
|
case HPSA_INQUIRY:
|
|
|
|
|
/* are we trying to read a vital product page */
|
2014-02-18 19:57:11 +00:00
|
|
|
if (page_code & VPD_PAGE) {
|
2009-12-08 22:09:11 +00:00
|
|
|
c->Request.CDB[1] = 0x01;
|
2014-02-18 19:57:11 +00:00
|
|
|
c->Request.CDB[2] = (page_code & 0xff);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
c->Request.CDBLen = 6;
|
2014-11-14 23:27:04 +00:00
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(cmd_type, ATTR_SIMPLE, XFER_READ);
|
2009-12-08 22:09:11 +00:00
|
|
|
c->Request.Timeout = 0;
|
|
|
|
|
c->Request.CDB[0] = HPSA_INQUIRY;
|
|
|
|
|
c->Request.CDB[4] = size & 0xFF;
|
|
|
|
|
break;
|
|
|
|
|
case HPSA_REPORT_LOG:
|
|
|
|
|
case HPSA_REPORT_PHYS:
|
|
|
|
|
/* Talking to controller so It's a physical command
|
|
|
|
|
mode = 00 target = 0. Nothing to write.
|
|
|
|
|
*/
|
|
|
|
|
c->Request.CDBLen = 12;
|
2014-11-14 23:27:04 +00:00
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(cmd_type, ATTR_SIMPLE, XFER_READ);
|
2009-12-08 22:09:11 +00:00
|
|
|
c->Request.Timeout = 0;
|
|
|
|
|
c->Request.CDB[0] = cmd;
|
|
|
|
|
c->Request.CDB[6] = (size >> 24) & 0xFF; /* MSB */
|
|
|
|
|
c->Request.CDB[7] = (size >> 16) & 0xFF;
|
|
|
|
|
c->Request.CDB[8] = (size >> 8) & 0xFF;
|
|
|
|
|
c->Request.CDB[9] = size & 0xFF;
|
|
|
|
|
break;
|
2015-11-04 21:52:16 +00:00
|
|
|
case BMIC_SENSE_DIAG_OPTIONS:
|
|
|
|
|
c->Request.CDBLen = 16;
|
|
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(cmd_type, ATTR_SIMPLE, XFER_READ);
|
|
|
|
|
c->Request.Timeout = 0;
|
|
|
|
|
/* Spec says this should be BMIC_WRITE */
|
|
|
|
|
c->Request.CDB[0] = BMIC_READ;
|
|
|
|
|
c->Request.CDB[6] = BMIC_SENSE_DIAG_OPTIONS;
|
|
|
|
|
break;
|
|
|
|
|
case BMIC_SET_DIAG_OPTIONS:
|
|
|
|
|
c->Request.CDBLen = 16;
|
|
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(cmd_type,
|
|
|
|
|
ATTR_SIMPLE, XFER_WRITE);
|
|
|
|
|
c->Request.Timeout = 0;
|
|
|
|
|
c->Request.CDB[0] = BMIC_WRITE;
|
|
|
|
|
c->Request.CDB[6] = BMIC_SET_DIAG_OPTIONS;
|
|
|
|
|
break;
|
2009-12-08 22:09:11 +00:00
|
|
|
case HPSA_CACHE_FLUSH:
|
|
|
|
|
c->Request.CDBLen = 12;
|
2014-11-14 23:27:04 +00:00
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(cmd_type,
|
|
|
|
|
ATTR_SIMPLE, XFER_WRITE);
|
2009-12-08 22:09:11 +00:00
|
|
|
c->Request.Timeout = 0;
|
|
|
|
|
c->Request.CDB[0] = BMIC_WRITE;
|
|
|
|
|
c->Request.CDB[6] = BMIC_CACHE_FLUSH;
|
2011-10-26 21:21:17 +00:00
|
|
|
c->Request.CDB[7] = (size >> 8) & 0xFF;
|
|
|
|
|
c->Request.CDB[8] = size & 0xFF;
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
case TEST_UNIT_READY:
|
|
|
|
|
c->Request.CDBLen = 6;
|
2014-11-14 23:27:04 +00:00
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(cmd_type, ATTR_SIMPLE, XFER_NONE);
|
2009-12-08 22:09:11 +00:00
|
|
|
c->Request.Timeout = 0;
|
|
|
|
|
break;
|
2014-02-18 19:55:33 +00:00
|
|
|
case HPSA_GET_RAID_MAP:
|
|
|
|
|
c->Request.CDBLen = 12;
|
2014-11-14 23:27:04 +00:00
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(cmd_type, ATTR_SIMPLE, XFER_READ);
|
2014-02-18 19:55:33 +00:00
|
|
|
c->Request.Timeout = 0;
|
|
|
|
|
c->Request.CDB[0] = HPSA_CISS_READ;
|
|
|
|
|
c->Request.CDB[1] = cmd;
|
|
|
|
|
c->Request.CDB[6] = (size >> 24) & 0xFF; /* MSB */
|
|
|
|
|
c->Request.CDB[7] = (size >> 16) & 0xFF;
|
|
|
|
|
c->Request.CDB[8] = (size >> 8) & 0xFF;
|
|
|
|
|
c->Request.CDB[9] = size & 0xFF;
|
|
|
|
|
break;
|
2014-02-21 22:25:15 +00:00
|
|
|
case BMIC_SENSE_CONTROLLER_PARAMETERS:
|
|
|
|
|
c->Request.CDBLen = 10;
|
2014-11-14 23:27:04 +00:00
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(cmd_type, ATTR_SIMPLE, XFER_READ);
|
2014-02-21 22:25:15 +00:00
|
|
|
c->Request.Timeout = 0;
|
|
|
|
|
c->Request.CDB[0] = BMIC_READ;
|
|
|
|
|
c->Request.CDB[6] = BMIC_SENSE_CONTROLLER_PARAMETERS;
|
|
|
|
|
c->Request.CDB[7] = (size >> 16) & 0xFF;
|
|
|
|
|
c->Request.CDB[8] = (size >> 8) & 0xFF;
|
|
|
|
|
break;
|
2015-01-23 22:43:30 +00:00
|
|
|
case BMIC_IDENTIFY_PHYSICAL_DEVICE:
|
|
|
|
|
c->Request.CDBLen = 10;
|
|
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(cmd_type, ATTR_SIMPLE, XFER_READ);
|
|
|
|
|
c->Request.Timeout = 0;
|
|
|
|
|
c->Request.CDB[0] = BMIC_READ;
|
|
|
|
|
c->Request.CDB[6] = BMIC_IDENTIFY_PHYSICAL_DEVICE;
|
|
|
|
|
c->Request.CDB[7] = (size >> 16) & 0xFF;
|
|
|
|
|
c->Request.CDB[8] = (size >> 8) & 0XFF;
|
|
|
|
|
break;
|
2015-11-04 21:52:34 +00:00
|
|
|
case BMIC_SENSE_SUBSYSTEM_INFORMATION:
|
|
|
|
|
c->Request.CDBLen = 10;
|
|
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(cmd_type, ATTR_SIMPLE, XFER_READ);
|
|
|
|
|
c->Request.Timeout = 0;
|
|
|
|
|
c->Request.CDB[0] = BMIC_READ;
|
|
|
|
|
c->Request.CDB[6] = BMIC_SENSE_SUBSYSTEM_INFORMATION;
|
|
|
|
|
c->Request.CDB[7] = (size >> 16) & 0xFF;
|
|
|
|
|
c->Request.CDB[8] = (size >> 8) & 0XFF;
|
|
|
|
|
break;
|
2015-12-22 16:36:48 +00:00
|
|
|
case BMIC_SENSE_STORAGE_BOX_PARAMS:
|
|
|
|
|
c->Request.CDBLen = 10;
|
|
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(cmd_type, ATTR_SIMPLE, XFER_READ);
|
|
|
|
|
c->Request.Timeout = 0;
|
|
|
|
|
c->Request.CDB[0] = BMIC_READ;
|
|
|
|
|
c->Request.CDB[6] = BMIC_SENSE_STORAGE_BOX_PARAMS;
|
|
|
|
|
c->Request.CDB[7] = (size >> 16) & 0xFF;
|
|
|
|
|
c->Request.CDB[8] = (size >> 8) & 0XFF;
|
|
|
|
|
break;
|
2015-11-04 21:51:57 +00:00
|
|
|
case BMIC_IDENTIFY_CONTROLLER:
|
|
|
|
|
c->Request.CDBLen = 10;
|
|
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(cmd_type, ATTR_SIMPLE, XFER_READ);
|
|
|
|
|
c->Request.Timeout = 0;
|
|
|
|
|
c->Request.CDB[0] = BMIC_READ;
|
|
|
|
|
c->Request.CDB[1] = 0;
|
|
|
|
|
c->Request.CDB[2] = 0;
|
|
|
|
|
c->Request.CDB[3] = 0;
|
|
|
|
|
c->Request.CDB[4] = 0;
|
|
|
|
|
c->Request.CDB[5] = 0;
|
|
|
|
|
c->Request.CDB[6] = BMIC_IDENTIFY_CONTROLLER;
|
|
|
|
|
c->Request.CDB[7] = (size >> 16) & 0xFF;
|
|
|
|
|
c->Request.CDB[8] = (size >> 8) & 0XFF;
|
|
|
|
|
c->Request.CDB[9] = 0;
|
|
|
|
|
break;
|
2009-12-08 22:09:11 +00:00
|
|
|
default:
|
|
|
|
|
dev_warn(&h->pdev->dev, "unknown command 0x%c\n", cmd);
|
|
|
|
|
BUG();
|
2013-02-20 17:24:41 +00:00
|
|
|
return -1;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
} else if (cmd_type == TYPE_MSG) {
|
|
|
|
|
switch (cmd) {
|
|
|
|
|
|
2015-11-04 21:51:02 +00:00
|
|
|
case HPSA_PHYS_TARGET_RESET:
|
|
|
|
|
c->Request.CDBLen = 16;
|
|
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(cmd_type, ATTR_SIMPLE, XFER_NONE);
|
|
|
|
|
c->Request.Timeout = 0; /* Don't time out */
|
|
|
|
|
memset(&c->Request.CDB[0], 0, sizeof(c->Request.CDB));
|
|
|
|
|
c->Request.CDB[0] = HPSA_RESET;
|
|
|
|
|
c->Request.CDB[1] = HPSA_TARGET_RESET_TYPE;
|
|
|
|
|
/* Physical target reset needs no control bytes 4-7*/
|
|
|
|
|
c->Request.CDB[4] = 0x00;
|
|
|
|
|
c->Request.CDB[5] = 0x00;
|
|
|
|
|
c->Request.CDB[6] = 0x00;
|
|
|
|
|
c->Request.CDB[7] = 0x00;
|
|
|
|
|
break;
|
2009-12-08 22:09:11 +00:00
|
|
|
case HPSA_DEVICE_RESET_MSG:
|
|
|
|
|
c->Request.CDBLen = 16;
|
2014-11-14 23:27:04 +00:00
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(cmd_type, ATTR_SIMPLE, XFER_NONE);
|
2009-12-08 22:09:11 +00:00
|
|
|
c->Request.Timeout = 0; /* Don't time out */
|
2011-05-03 19:59:51 +00:00
|
|
|
memset(&c->Request.CDB[0], 0, sizeof(c->Request.CDB));
|
|
|
|
|
c->Request.CDB[0] = cmd;
|
2012-07-26 16:34:10 +00:00
|
|
|
c->Request.CDB[1] = HPSA_RESET_TYPE_LUN;
|
2009-12-08 22:09:11 +00:00
|
|
|
/* If bytes 4-7 are zero, it means reset the */
|
|
|
|
|
/* LunID device */
|
|
|
|
|
c->Request.CDB[4] = 0x00;
|
|
|
|
|
c->Request.CDB[5] = 0x00;
|
|
|
|
|
c->Request.CDB[6] = 0x00;
|
|
|
|
|
c->Request.CDB[7] = 0x00;
|
2012-05-01 16:42:51 +00:00
|
|
|
break;
|
|
|
|
|
case HPSA_ABORT_MSG:
|
2015-04-23 14:32:06 +00:00
|
|
|
memcpy(&tag, buff, sizeof(tag));
|
2015-01-23 22:41:09 +00:00
|
|
|
dev_dbg(&h->pdev->dev,
|
2015-04-23 14:32:06 +00:00
|
|
|
"Abort Tag:0x%016llx using rqst Tag:0x%016llx",
|
|
|
|
|
tag, c->Header.tag);
|
2012-05-01 16:42:51 +00:00
|
|
|
c->Request.CDBLen = 16;
|
2014-11-14 23:27:04 +00:00
|
|
|
c->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(cmd_type,
|
|
|
|
|
ATTR_SIMPLE, XFER_WRITE);
|
2012-05-01 16:42:51 +00:00
|
|
|
c->Request.Timeout = 0; /* Don't time out */
|
|
|
|
|
c->Request.CDB[0] = HPSA_TASK_MANAGEMENT;
|
|
|
|
|
c->Request.CDB[1] = HPSA_TMF_ABORT_TASK;
|
|
|
|
|
c->Request.CDB[2] = 0x00; /* reserved */
|
|
|
|
|
c->Request.CDB[3] = 0x00; /* reserved */
|
|
|
|
|
/* Tag to abort goes in CDB[4]-CDB[11] */
|
2015-04-23 14:32:06 +00:00
|
|
|
memcpy(&c->Request.CDB[4], &tag, sizeof(tag));
|
2012-05-01 16:42:51 +00:00
|
|
|
c->Request.CDB[12] = 0x00; /* reserved */
|
|
|
|
|
c->Request.CDB[13] = 0x00; /* reserved */
|
|
|
|
|
c->Request.CDB[14] = 0x00; /* reserved */
|
|
|
|
|
c->Request.CDB[15] = 0x00; /* reserved */
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
dev_warn(&h->pdev->dev, "unknown message type %d\n",
|
|
|
|
|
cmd);
|
|
|
|
|
BUG();
|
|
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
dev_warn(&h->pdev->dev, "unknown command type %d\n", cmd_type);
|
|
|
|
|
BUG();
|
|
|
|
|
}
|
|
|
|
|
|
2014-11-14 23:27:04 +00:00
|
|
|
switch (GET_DIR(c->Request.type_attr_dir)) {
|
2009-12-08 22:09:11 +00:00
|
|
|
case XFER_READ:
|
|
|
|
|
pci_dir = PCI_DMA_FROMDEVICE;
|
|
|
|
|
break;
|
|
|
|
|
case XFER_WRITE:
|
|
|
|
|
pci_dir = PCI_DMA_TODEVICE;
|
|
|
|
|
break;
|
|
|
|
|
case XFER_NONE:
|
|
|
|
|
pci_dir = PCI_DMA_NONE;
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
pci_dir = PCI_DMA_BIDIRECTIONAL;
|
|
|
|
|
}
|
2013-02-20 17:24:41 +00:00
|
|
|
if (hpsa_map_one(h->pdev, c, buff, size, pci_dir))
|
|
|
|
|
return -1;
|
|
|
|
|
return 0;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Map (physical) PCI mem into (virtual) kernel space
|
|
|
|
|
*/
|
|
|
|
|
static void __iomem *remap_pci_mem(ulong base, ulong size)
|
|
|
|
|
{
|
|
|
|
|
ulong page_base = ((ulong) base) & PAGE_MASK;
|
|
|
|
|
ulong page_offs = ((ulong) base) - page_base;
|
2012-07-26 16:34:23 +00:00
|
|
|
void __iomem *page_remapped = ioremap_nocache(page_base,
|
|
|
|
|
page_offs + size);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
return page_remapped ? (page_remapped + page_offs) : NULL;
|
|
|
|
|
}
|
|
|
|
|
|
2012-05-01 16:43:06 +00:00
|
|
|
static inline unsigned long get_next_completion(struct ctlr_info *h, u8 q)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
2012-05-01 16:43:06 +00:00
|
|
|
return h->access.command_completed(h, q);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
2010-02-04 14:42:35 +00:00
|
|
|
static inline bool interrupt_pending(struct ctlr_info *h)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
|
|
|
|
return h->access.intr_pending(h);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline long interrupt_not_for_us(struct ctlr_info *h)
|
|
|
|
|
{
|
2010-06-16 18:51:50 +00:00
|
|
|
return (h->access.intr_pending(h) == 0) ||
|
|
|
|
|
(h->interrupts_enabled == 0);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
2010-02-04 14:41:33 +00:00
|
|
|
static inline int bad_tag(struct ctlr_info *h, u32 tag_index,
|
|
|
|
|
u32 raw_tag)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
|
|
|
|
if (unlikely(tag_index >= h->nr_cmds)) {
|
|
|
|
|
dev_warn(&h->pdev->dev, "bad tag 0x%08x ignored.\n", raw_tag);
|
|
|
|
|
return 1;
|
|
|
|
|
}
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2012-05-01 16:42:46 +00:00
|
|
|
static inline void finish_cmd(struct CommandList *c)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
2012-05-01 16:43:42 +00:00
|
|
|
dial_up_lockup_detection_on_fw_flash_complete(c->h, c);
|
2014-02-18 19:56:34 +00:00
|
|
|
if (likely(c->cmd_type == CMD_IOACCEL1 || c->cmd_type == CMD_SCSI
|
|
|
|
|
|| c->cmd_type == CMD_IOACCEL2))
|
2011-05-03 19:59:00 +00:00
|
|
|
complete_scsi_command(c);
|
2015-04-23 14:34:06 +00:00
|
|
|
else if (c->cmd_type == CMD_IOCTL_PEND || c->cmd_type == IOACCEL2_TMF)
|
2009-12-08 22:09:11 +00:00
|
|
|
complete(c->waiting);
|
2010-02-04 14:42:24 +00:00
|
|
|
}
|
|
|
|
|
|
2010-02-04 14:42:40 +00:00
|
|
|
/* process completion of an indexed ("direct lookup") command */
|
2012-05-01 16:43:01 +00:00
|
|
|
static inline void process_indexed_cmd(struct ctlr_info *h,
|
2010-02-04 14:42:40 +00:00
|
|
|
u32 raw_tag)
|
|
|
|
|
{
|
|
|
|
|
u32 tag_index;
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
|
2015-01-23 22:43:09 +00:00
|
|
|
tag_index = raw_tag >> DIRECT_LOOKUP_SHIFT;
|
2012-05-01 16:43:01 +00:00
|
|
|
if (!bad_tag(h, tag_index, raw_tag)) {
|
|
|
|
|
c = h->cmd_pool + tag_index;
|
|
|
|
|
finish_cmd(c);
|
|
|
|
|
}
|
2010-02-04 14:42:40 +00:00
|
|
|
}
|
|
|
|
|
|
2011-05-03 19:59:51 +00:00
|
|
|
/* Some controllers, like p400, will give us one interrupt
|
|
|
|
|
* after a soft reset, even if we turned interrupts off.
|
|
|
|
|
* Only need to check for this in the hpsa_xxx_discard_completions
|
|
|
|
|
* functions.
|
|
|
|
|
*/
|
|
|
|
|
static int ignore_bogus_interrupt(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
if (likely(!reset_devices))
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
if (likely(h->interrupts_enabled))
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
dev_info(&h->pdev->dev, "Received interrupt while interrupts disabled "
|
|
|
|
|
"(known firmware bug.) Ignoring.\n");
|
|
|
|
|
|
|
|
|
|
return 1;
|
|
|
|
|
}
|
|
|
|
|
|
2012-05-01 16:43:06 +00:00
|
|
|
/*
|
|
|
|
|
* Convert &h->q[x] (passed to interrupt handlers) back to h.
|
|
|
|
|
* Relies on (h-q[x] == x) being true for x such that
|
|
|
|
|
* 0 <= x < MAX_REPLY_QUEUES.
|
|
|
|
|
*/
|
|
|
|
|
static struct ctlr_info *queue_to_hba(u8 *queue)
|
2011-05-03 19:59:51 +00:00
|
|
|
{
|
2012-05-01 16:43:06 +00:00
|
|
|
return container_of((queue - *queue), struct ctlr_info, q[0]);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static irqreturn_t hpsa_intx_discard_completions(int irq, void *queue)
|
|
|
|
|
{
|
|
|
|
|
struct ctlr_info *h = queue_to_hba(queue);
|
|
|
|
|
u8 q = *(u8 *) queue;
|
2011-05-03 19:59:51 +00:00
|
|
|
u32 raw_tag;
|
|
|
|
|
|
|
|
|
|
if (ignore_bogus_interrupt(h))
|
|
|
|
|
return IRQ_NONE;
|
|
|
|
|
|
|
|
|
|
if (interrupt_not_for_us(h))
|
|
|
|
|
return IRQ_NONE;
|
2011-10-26 21:22:04 +00:00
|
|
|
h->last_intr_timestamp = get_jiffies_64();
|
2011-05-03 19:59:51 +00:00
|
|
|
while (interrupt_pending(h)) {
|
2012-05-01 16:43:06 +00:00
|
|
|
raw_tag = get_next_completion(h, q);
|
2011-05-03 19:59:51 +00:00
|
|
|
while (raw_tag != FIFO_EMPTY)
|
2012-05-01 16:43:06 +00:00
|
|
|
raw_tag = next_command(h, q);
|
2011-05-03 19:59:51 +00:00
|
|
|
}
|
|
|
|
|
return IRQ_HANDLED;
|
|
|
|
|
}
|
|
|
|
|
|
2012-05-01 16:43:06 +00:00
|
|
|
static irqreturn_t hpsa_msix_discard_completions(int irq, void *queue)
|
2011-05-03 19:59:51 +00:00
|
|
|
{
|
2012-05-01 16:43:06 +00:00
|
|
|
struct ctlr_info *h = queue_to_hba(queue);
|
2011-05-03 19:59:51 +00:00
|
|
|
u32 raw_tag;
|
2012-05-01 16:43:06 +00:00
|
|
|
u8 q = *(u8 *) queue;
|
2011-05-03 19:59:51 +00:00
|
|
|
|
|
|
|
|
if (ignore_bogus_interrupt(h))
|
|
|
|
|
return IRQ_NONE;
|
|
|
|
|
|
2011-10-26 21:22:04 +00:00
|
|
|
h->last_intr_timestamp = get_jiffies_64();
|
2012-05-01 16:43:06 +00:00
|
|
|
raw_tag = get_next_completion(h, q);
|
2011-05-03 19:59:51 +00:00
|
|
|
while (raw_tag != FIFO_EMPTY)
|
2012-05-01 16:43:06 +00:00
|
|
|
raw_tag = next_command(h, q);
|
2011-05-03 19:59:51 +00:00
|
|
|
return IRQ_HANDLED;
|
|
|
|
|
}
|
|
|
|
|
|
2012-05-01 16:43:06 +00:00
|
|
|
static irqreturn_t do_hpsa_intr_intx(int irq, void *queue)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
2012-05-01 16:43:06 +00:00
|
|
|
struct ctlr_info *h = queue_to_hba((u8 *) queue);
|
2010-02-04 14:42:40 +00:00
|
|
|
u32 raw_tag;
|
2012-05-01 16:43:06 +00:00
|
|
|
u8 q = *(u8 *) queue;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
if (interrupt_not_for_us(h))
|
|
|
|
|
return IRQ_NONE;
|
2011-10-26 21:22:04 +00:00
|
|
|
h->last_intr_timestamp = get_jiffies_64();
|
2010-06-16 18:51:50 +00:00
|
|
|
while (interrupt_pending(h)) {
|
2012-05-01 16:43:06 +00:00
|
|
|
raw_tag = get_next_completion(h, q);
|
2010-06-16 18:51:50 +00:00
|
|
|
while (raw_tag != FIFO_EMPTY) {
|
2015-01-23 22:43:09 +00:00
|
|
|
process_indexed_cmd(h, raw_tag);
|
2012-05-01 16:43:06 +00:00
|
|
|
raw_tag = next_command(h, q);
|
2010-06-16 18:51:50 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
return IRQ_HANDLED;
|
|
|
|
|
}
|
|
|
|
|
|
2012-05-01 16:43:06 +00:00
|
|
|
static irqreturn_t do_hpsa_intr_msi(int irq, void *queue)
|
2010-06-16 18:51:50 +00:00
|
|
|
{
|
2012-05-01 16:43:06 +00:00
|
|
|
struct ctlr_info *h = queue_to_hba(queue);
|
2010-06-16 18:51:50 +00:00
|
|
|
u32 raw_tag;
|
2012-05-01 16:43:06 +00:00
|
|
|
u8 q = *(u8 *) queue;
|
2010-06-16 18:51:50 +00:00
|
|
|
|
2011-10-26 21:22:04 +00:00
|
|
|
h->last_intr_timestamp = get_jiffies_64();
|
2012-05-01 16:43:06 +00:00
|
|
|
raw_tag = get_next_completion(h, q);
|
2010-02-04 14:42:40 +00:00
|
|
|
while (raw_tag != FIFO_EMPTY) {
|
2015-01-23 22:43:09 +00:00
|
|
|
process_indexed_cmd(h, raw_tag);
|
2012-05-01 16:43:06 +00:00
|
|
|
raw_tag = next_command(h, q);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
return IRQ_HANDLED;
|
|
|
|
|
}
|
|
|
|
|
|
2011-02-15 21:32:53 +00:00
|
|
|
/* Send a message CDB to the firmware. Careful, this only works
|
|
|
|
|
* in simple mode, not performant mode due to the tag lookup.
|
|
|
|
|
* We only ever use this immediately after a controller reset.
|
|
|
|
|
*/
|
2012-12-21 21:08:55 +00:00
|
|
|
static int hpsa_message(struct pci_dev *pdev, unsigned char opcode,
|
|
|
|
|
unsigned char type)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
|
|
|
|
struct Command {
|
|
|
|
|
struct CommandListHeader CommandHeader;
|
|
|
|
|
struct RequestBlock Request;
|
|
|
|
|
struct ErrDescriptor ErrorDescriptor;
|
|
|
|
|
};
|
|
|
|
|
struct Command *cmd;
|
|
|
|
|
static const size_t cmd_sz = sizeof(*cmd) +
|
|
|
|
|
sizeof(cmd->ErrorDescriptor);
|
|
|
|
|
dma_addr_t paddr64;
|
2015-01-23 22:41:09 +00:00
|
|
|
__le32 paddr32;
|
|
|
|
|
u32 tag;
|
2009-12-08 22:09:11 +00:00
|
|
|
void __iomem *vaddr;
|
|
|
|
|
int i, err;
|
|
|
|
|
|
|
|
|
|
vaddr = pci_ioremap_bar(pdev, 0);
|
|
|
|
|
if (vaddr == NULL)
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
|
|
/* The Inbound Post Queue only accepts 32-bit physical addresses for the
|
|
|
|
|
* CCISS commands, so they must be allocated from the lower 4GiB of
|
|
|
|
|
* memory.
|
|
|
|
|
*/
|
|
|
|
|
err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(32));
|
|
|
|
|
if (err) {
|
|
|
|
|
iounmap(vaddr);
|
2015-01-23 22:42:37 +00:00
|
|
|
return err;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
cmd = pci_alloc_consistent(pdev, cmd_sz, &paddr64);
|
|
|
|
|
if (cmd == NULL) {
|
|
|
|
|
iounmap(vaddr);
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* This must fit, because of the 32-bit consistent DMA mask. Also,
|
|
|
|
|
* although there's no guarantee, we assume that the address is at
|
|
|
|
|
* least 4-byte aligned (most likely, it's page-aligned).
|
|
|
|
|
*/
|
2015-01-23 22:41:09 +00:00
|
|
|
paddr32 = cpu_to_le32(paddr64);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
cmd->CommandHeader.ReplyQueue = 0;
|
|
|
|
|
cmd->CommandHeader.SGList = 0;
|
2014-11-14 23:26:59 +00:00
|
|
|
cmd->CommandHeader.SGTotal = cpu_to_le16(0);
|
2015-01-23 22:41:09 +00:00
|
|
|
cmd->CommandHeader.tag = cpu_to_le64(paddr64);
|
2009-12-08 22:09:11 +00:00
|
|
|
memset(&cmd->CommandHeader.LUN.LunAddrBytes, 0, 8);
|
|
|
|
|
|
|
|
|
|
cmd->Request.CDBLen = 16;
|
2014-11-14 23:27:04 +00:00
|
|
|
cmd->Request.type_attr_dir =
|
|
|
|
|
TYPE_ATTR_DIR(TYPE_MSG, ATTR_HEADOFQUEUE, XFER_NONE);
|
2009-12-08 22:09:11 +00:00
|
|
|
cmd->Request.Timeout = 0; /* Don't time out */
|
|
|
|
|
cmd->Request.CDB[0] = opcode;
|
|
|
|
|
cmd->Request.CDB[1] = type;
|
|
|
|
|
memset(&cmd->Request.CDB[2], 0, 14); /* rest of the CDB is reserved */
|
2014-11-14 23:26:59 +00:00
|
|
|
cmd->ErrorDescriptor.Addr =
|
2015-01-23 22:41:09 +00:00
|
|
|
cpu_to_le64((le32_to_cpu(paddr32) + sizeof(*cmd)));
|
2014-11-14 23:26:59 +00:00
|
|
|
cmd->ErrorDescriptor.Len = cpu_to_le32(sizeof(struct ErrorInfo));
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-01-23 22:41:09 +00:00
|
|
|
writel(le32_to_cpu(paddr32), vaddr + SA5_REQUEST_PORT_OFFSET);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
for (i = 0; i < HPSA_MSG_SEND_RETRY_LIMIT; i++) {
|
|
|
|
|
tag = readl(vaddr + SA5_REPLY_PORT_OFFSET);
|
2015-01-23 22:41:09 +00:00
|
|
|
if ((tag & ~HPSA_SIMPLE_ERROR_BITS) == paddr64)
|
2009-12-08 22:09:11 +00:00
|
|
|
break;
|
|
|
|
|
msleep(HPSA_MSG_SEND_RETRY_INTERVAL_MSECS);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
iounmap(vaddr);
|
|
|
|
|
|
|
|
|
|
/* we leak the DMA buffer here ... no choice since the controller could
|
|
|
|
|
* still complete the command.
|
|
|
|
|
*/
|
|
|
|
|
if (i == HPSA_MSG_SEND_RETRY_LIMIT) {
|
|
|
|
|
dev_err(&pdev->dev, "controller message %02x:%02x timed out\n",
|
|
|
|
|
opcode, type);
|
|
|
|
|
return -ETIMEDOUT;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
pci_free_consistent(pdev, cmd_sz, cmd, paddr64);
|
|
|
|
|
|
|
|
|
|
if (tag & HPSA_ERROR_BIT) {
|
|
|
|
|
dev_err(&pdev->dev, "controller message %02x:%02x failed\n",
|
|
|
|
|
opcode, type);
|
|
|
|
|
return -EIO;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
dev_info(&pdev->dev, "controller message %02x:%02x succeeded\n",
|
|
|
|
|
opcode, type);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#define hpsa_noop(p) hpsa_message(p, 3, 0)
|
|
|
|
|
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
static int hpsa_controller_hard_reset(struct pci_dev *pdev,
|
2014-11-14 23:26:27 +00:00
|
|
|
void __iomem *vaddr, u32 use_doorbell)
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
{
|
|
|
|
|
|
|
|
|
|
if (use_doorbell) {
|
|
|
|
|
/* For everything after the P600, the PCI power state method
|
|
|
|
|
* of resetting the controller doesn't work, so we have this
|
|
|
|
|
* other way using the doorbell register.
|
|
|
|
|
*/
|
|
|
|
|
dev_info(&pdev->dev, "using doorbell to reset controller\n");
|
2011-05-03 19:59:46 +00:00
|
|
|
writel(use_doorbell, vaddr + SA5_DOORBELL);
|
2013-09-23 18:33:36 +00:00
|
|
|
|
2014-05-29 15:52:47 +00:00
|
|
|
/* PMC hardware guys tell us we need a 10 second delay after
|
2013-09-23 18:33:36 +00:00
|
|
|
* doorbell reset and before any attempt to talk to the board
|
|
|
|
|
* at all to ensure that this actually works and doesn't fall
|
|
|
|
|
* over in some weird corner cases.
|
|
|
|
|
*/
|
2014-05-29 15:52:47 +00:00
|
|
|
msleep(10000);
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
} else { /* Try to do it the PCI power state way */
|
|
|
|
|
|
|
|
|
|
/* Quoting from the Open CISS Specification: "The Power
|
|
|
|
|
* Management Control/Status Register (CSR) controls the power
|
|
|
|
|
* state of the device. The normal operating state is D0,
|
|
|
|
|
* CSR=00h. The software off state is D3, CSR=03h. To reset
|
|
|
|
|
* the controller, place the interface device in D3 then to D0,
|
|
|
|
|
* this causes a secondary PCI reset which will reset the
|
|
|
|
|
* controller." */
|
2015-01-23 22:41:25 +00:00
|
|
|
|
|
|
|
|
int rc = 0;
|
|
|
|
|
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
dev_info(&pdev->dev, "using PCI PM to reset controller\n");
|
2015-01-23 22:41:25 +00:00
|
|
|
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
/* enter the D3hot power management state */
|
2015-01-23 22:41:25 +00:00
|
|
|
rc = pci_set_power_state(pdev, PCI_D3hot);
|
|
|
|
|
if (rc)
|
|
|
|
|
return rc;
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
|
|
|
|
|
msleep(500);
|
|
|
|
|
|
|
|
|
|
/* enter the D0 power management state */
|
2015-01-23 22:41:25 +00:00
|
|
|
rc = pci_set_power_state(pdev, PCI_D0);
|
|
|
|
|
if (rc)
|
|
|
|
|
return rc;
|
2011-10-21 06:19:43 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* The P600 requires a small delay when changing states.
|
|
|
|
|
* Otherwise we may think the board did not reset and we bail.
|
|
|
|
|
* This for kdump only and is particular to the P600.
|
|
|
|
|
*/
|
|
|
|
|
msleep(500);
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
}
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2012-12-21 21:08:55 +00:00
|
|
|
static void init_driver_version(char *driver_version, int len)
|
2011-05-03 19:59:10 +00:00
|
|
|
{
|
|
|
|
|
memset(driver_version, 0, len);
|
2012-01-19 20:00:59 +00:00
|
|
|
strncpy(driver_version, HPSA " " HPSA_DRIVER_VERSION, len - 1);
|
2011-05-03 19:59:10 +00:00
|
|
|
}
|
|
|
|
|
|
2012-12-21 21:08:55 +00:00
|
|
|
static int write_driver_ver_to_cfgtable(struct CfgTable __iomem *cfgtable)
|
2011-05-03 19:59:10 +00:00
|
|
|
{
|
|
|
|
|
char *driver_version;
|
|
|
|
|
int i, size = sizeof(cfgtable->driver_version);
|
|
|
|
|
|
|
|
|
|
driver_version = kmalloc(size, GFP_KERNEL);
|
|
|
|
|
if (!driver_version)
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
|
|
init_driver_version(driver_version, size);
|
|
|
|
|
for (i = 0; i < size; i++)
|
|
|
|
|
writeb(driver_version[i], &cfgtable->driver_version[i]);
|
|
|
|
|
kfree(driver_version);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2012-12-21 21:08:55 +00:00
|
|
|
static void read_driver_ver_from_cfgtable(struct CfgTable __iomem *cfgtable,
|
|
|
|
|
unsigned char *driver_ver)
|
2011-05-03 19:59:10 +00:00
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < sizeof(cfgtable->driver_version); i++)
|
|
|
|
|
driver_ver[i] = readb(&cfgtable->driver_version[i]);
|
|
|
|
|
}
|
|
|
|
|
|
2012-12-21 21:08:55 +00:00
|
|
|
static int controller_reset_failed(struct CfgTable __iomem *cfgtable)
|
2011-05-03 19:59:10 +00:00
|
|
|
{
|
|
|
|
|
|
|
|
|
|
char *driver_ver, *old_driver_ver;
|
|
|
|
|
int rc, size = sizeof(cfgtable->driver_version);
|
|
|
|
|
|
|
|
|
|
old_driver_ver = kmalloc(2 * size, GFP_KERNEL);
|
|
|
|
|
if (!old_driver_ver)
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
driver_ver = old_driver_ver + size;
|
|
|
|
|
|
|
|
|
|
/* After a reset, the 32 bytes of "driver version" in the cfgtable
|
|
|
|
|
* should have been changed, otherwise we know the reset failed.
|
|
|
|
|
*/
|
|
|
|
|
init_driver_version(old_driver_ver, size);
|
|
|
|
|
read_driver_ver_from_cfgtable(cfgtable, driver_ver);
|
|
|
|
|
rc = !memcmp(driver_ver, old_driver_ver, size);
|
|
|
|
|
kfree(old_driver_ver);
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
/* This does a hard reset of the controller using PCI power management
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
* states or the using the doorbell register.
|
2009-12-08 22:09:11 +00:00
|
|
|
*/
|
2015-04-02 13:25:54 +00:00
|
|
|
static int hpsa_kdump_hard_reset_controller(struct pci_dev *pdev, u32 board_id)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
u64 cfg_offset;
|
|
|
|
|
u32 cfg_base_addr;
|
|
|
|
|
u64 cfg_base_addr_index;
|
|
|
|
|
void __iomem *vaddr;
|
|
|
|
|
unsigned long paddr;
|
2011-05-03 19:59:10 +00:00
|
|
|
u32 misc_fw_support;
|
2011-01-06 20:48:08 +00:00
|
|
|
int rc;
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
struct CfgTable __iomem *cfgtable;
|
2011-05-03 19:59:46 +00:00
|
|
|
u32 use_doorbell;
|
2011-01-06 20:48:08 +00:00
|
|
|
u16 command_register;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
/* For controllers as old as the P600, this is very nearly
|
|
|
|
|
* the same thing as
|
2009-12-08 22:09:11 +00:00
|
|
|
*
|
|
|
|
|
* pci_save_state(pci_dev);
|
|
|
|
|
* pci_set_power_state(pci_dev, PCI_D3hot);
|
|
|
|
|
* pci_set_power_state(pci_dev, PCI_D0);
|
|
|
|
|
* pci_restore_state(pci_dev);
|
|
|
|
|
*
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
* For controllers newer than the P600, the pci power state
|
|
|
|
|
* method of resetting doesn't work so we have another way
|
|
|
|
|
* using the doorbell register.
|
2009-12-08 22:09:11 +00:00
|
|
|
*/
|
2010-06-16 18:51:45 +00:00
|
|
|
|
2015-01-23 22:42:06 +00:00
|
|
|
if (!ctlr_is_resettable(board_id)) {
|
|
|
|
|
dev_warn(&pdev->dev, "Controller not resettable\n");
|
2011-01-06 20:48:18 +00:00
|
|
|
return -ENODEV;
|
|
|
|
|
}
|
2011-05-03 20:00:01 +00:00
|
|
|
|
|
|
|
|
/* if controller is soft- but not hard resettable... */
|
|
|
|
|
if (!ctlr_is_hard_resettable(board_id))
|
|
|
|
|
return -ENOTSUPP; /* try soft reset later. */
|
2010-06-16 18:51:45 +00:00
|
|
|
|
2011-01-06 20:48:08 +00:00
|
|
|
/* Save the PCI command register */
|
|
|
|
|
pci_read_config_word(pdev, 4, &command_register);
|
|
|
|
|
pci_save_state(pdev);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
/* find the first memory BAR, so we can find the cfg table */
|
|
|
|
|
rc = hpsa_pci_find_memory_BAR(pdev, &paddr);
|
|
|
|
|
if (rc)
|
|
|
|
|
return rc;
|
|
|
|
|
vaddr = remap_pci_mem(paddr, 0x250);
|
|
|
|
|
if (!vaddr)
|
|
|
|
|
return -ENOMEM;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
/* find cfgtable in order to check if reset via doorbell is supported */
|
|
|
|
|
rc = hpsa_find_cfg_addrs(pdev, vaddr, &cfg_base_addr,
|
|
|
|
|
&cfg_base_addr_index, &cfg_offset);
|
|
|
|
|
if (rc)
|
|
|
|
|
goto unmap_vaddr;
|
|
|
|
|
cfgtable = remap_pci_mem(pci_resource_start(pdev,
|
|
|
|
|
cfg_base_addr_index) + cfg_offset, sizeof(*cfgtable));
|
|
|
|
|
if (!cfgtable) {
|
|
|
|
|
rc = -ENOMEM;
|
|
|
|
|
goto unmap_vaddr;
|
|
|
|
|
}
|
2011-05-03 19:59:10 +00:00
|
|
|
rc = write_driver_ver_to_cfgtable(cfgtable);
|
|
|
|
|
if (rc)
|
2015-01-23 22:41:14 +00:00
|
|
|
goto unmap_cfgtable;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2011-05-03 19:59:46 +00:00
|
|
|
/* If reset via doorbell register is supported, use that.
|
|
|
|
|
* There are two such methods. Favor the newest method.
|
|
|
|
|
*/
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
misc_fw_support = readl(&cfgtable->misc_fw_support);
|
2011-05-03 19:59:46 +00:00
|
|
|
use_doorbell = misc_fw_support & MISC_FW_DOORBELL_RESET2;
|
|
|
|
|
if (use_doorbell) {
|
|
|
|
|
use_doorbell = DOORBELL_CTLR_RESET2;
|
|
|
|
|
} else {
|
|
|
|
|
use_doorbell = misc_fw_support & MISC_FW_DOORBELL_RESET;
|
|
|
|
|
if (use_doorbell) {
|
2015-01-23 22:42:22 +00:00
|
|
|
dev_warn(&pdev->dev,
|
|
|
|
|
"Soft reset not supported. Firmware update is required.\n");
|
2011-05-03 19:59:51 +00:00
|
|
|
rc = -ENOTSUPP; /* try soft reset */
|
2011-05-03 19:59:46 +00:00
|
|
|
goto unmap_cfgtable;
|
|
|
|
|
}
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
rc = hpsa_controller_hard_reset(pdev, vaddr, use_doorbell);
|
|
|
|
|
if (rc)
|
|
|
|
|
goto unmap_cfgtable;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2011-01-06 20:48:08 +00:00
|
|
|
pci_restore_state(pdev);
|
|
|
|
|
pci_write_config_word(pdev, 4, command_register);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
/* Some devices (notably the HP Smart Array 5i Controller)
|
|
|
|
|
need a little pause here */
|
|
|
|
|
msleep(HPSA_POST_RESET_PAUSE_MSECS);
|
|
|
|
|
|
2011-01-06 20:48:03 +00:00
|
|
|
rc = hpsa_wait_for_board_state(pdev, vaddr, BOARD_READY);
|
|
|
|
|
if (rc) {
|
|
|
|
|
dev_warn(&pdev->dev,
|
2015-01-23 22:42:22 +00:00
|
|
|
"Failed waiting for board to become ready after hard reset\n");
|
2011-01-06 20:48:03 +00:00
|
|
|
goto unmap_cfgtable;
|
|
|
|
|
}
|
|
|
|
|
|
2011-05-03 19:59:10 +00:00
|
|
|
rc = controller_reset_failed(vaddr);
|
|
|
|
|
if (rc < 0)
|
|
|
|
|
goto unmap_cfgtable;
|
|
|
|
|
if (rc) {
|
2011-05-03 19:59:51 +00:00
|
|
|
dev_warn(&pdev->dev, "Unable to successfully reset "
|
|
|
|
|
"controller. Will try soft reset.\n");
|
|
|
|
|
rc = -ENOTSUPP;
|
2011-05-03 19:59:10 +00:00
|
|
|
} else {
|
2011-05-03 19:59:51 +00:00
|
|
|
dev_info(&pdev->dev, "board ready after hard reset.\n");
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
unmap_cfgtable:
|
|
|
|
|
iounmap(cfgtable);
|
|
|
|
|
|
|
|
|
|
unmap_vaddr:
|
|
|
|
|
iounmap(vaddr);
|
|
|
|
|
return rc;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* We cannot read the structure directly, for portability we must use
|
|
|
|
|
* the io functions.
|
|
|
|
|
* This is for debug only.
|
|
|
|
|
*/
|
2014-11-14 23:26:27 +00:00
|
|
|
static void print_cfg_table(struct device *dev, struct CfgTable __iomem *tb)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
2010-05-27 20:13:58 +00:00
|
|
|
#ifdef HPSA_DEBUG
|
2009-12-08 22:09:11 +00:00
|
|
|
int i;
|
|
|
|
|
char temp_name[17];
|
|
|
|
|
|
|
|
|
|
dev_info(dev, "Controller Configuration information\n");
|
|
|
|
|
dev_info(dev, "------------------------------------\n");
|
|
|
|
|
for (i = 0; i < 4; i++)
|
|
|
|
|
temp_name[i] = readb(&(tb->Signature[i]));
|
|
|
|
|
temp_name[4] = '\0';
|
|
|
|
|
dev_info(dev, " Signature = %s\n", temp_name);
|
|
|
|
|
dev_info(dev, " Spec Number = %d\n", readl(&(tb->SpecValence)));
|
|
|
|
|
dev_info(dev, " Transport methods supported = 0x%x\n",
|
|
|
|
|
readl(&(tb->TransportSupport)));
|
|
|
|
|
dev_info(dev, " Transport methods active = 0x%x\n",
|
|
|
|
|
readl(&(tb->TransportActive)));
|
|
|
|
|
dev_info(dev, " Requested transport Method = 0x%x\n",
|
|
|
|
|
readl(&(tb->HostWrite.TransportRequest)));
|
|
|
|
|
dev_info(dev, " Coalesce Interrupt Delay = 0x%x\n",
|
|
|
|
|
readl(&(tb->HostWrite.CoalIntDelay)));
|
|
|
|
|
dev_info(dev, " Coalesce Interrupt Count = 0x%x\n",
|
|
|
|
|
readl(&(tb->HostWrite.CoalIntCount)));
|
2015-01-23 22:41:56 +00:00
|
|
|
dev_info(dev, " Max outstanding commands = %d\n",
|
2009-12-08 22:09:11 +00:00
|
|
|
readl(&(tb->CmdsOutMax)));
|
|
|
|
|
dev_info(dev, " Bus Types = 0x%x\n", readl(&(tb->BusTypes)));
|
|
|
|
|
for (i = 0; i < 16; i++)
|
|
|
|
|
temp_name[i] = readb(&(tb->ServerName[i]));
|
|
|
|
|
temp_name[16] = '\0';
|
|
|
|
|
dev_info(dev, " Server Name = %s\n", temp_name);
|
|
|
|
|
dev_info(dev, " Heartbeat Counter = 0x%x\n\n\n",
|
|
|
|
|
readl(&(tb->HeartBeat)));
|
|
|
|
|
#endif /* HPSA_DEBUG */
|
2010-05-27 20:13:58 +00:00
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
static int find_PCI_BAR_index(struct pci_dev *pdev, unsigned long pci_bar_addr)
|
|
|
|
|
{
|
|
|
|
|
int i, offset, mem_type, bar_type;
|
|
|
|
|
|
|
|
|
|
if (pci_bar_addr == PCI_BASE_ADDRESS_0) /* looking for BAR zero? */
|
|
|
|
|
return 0;
|
|
|
|
|
offset = 0;
|
|
|
|
|
for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) {
|
|
|
|
|
bar_type = pci_resource_flags(pdev, i) & PCI_BASE_ADDRESS_SPACE;
|
|
|
|
|
if (bar_type == PCI_BASE_ADDRESS_SPACE_IO)
|
|
|
|
|
offset += 4;
|
|
|
|
|
else {
|
|
|
|
|
mem_type = pci_resource_flags(pdev, i) &
|
|
|
|
|
PCI_BASE_ADDRESS_MEM_TYPE_MASK;
|
|
|
|
|
switch (mem_type) {
|
|
|
|
|
case PCI_BASE_ADDRESS_MEM_TYPE_32:
|
|
|
|
|
case PCI_BASE_ADDRESS_MEM_TYPE_1M:
|
|
|
|
|
offset += 4; /* 32 bit */
|
|
|
|
|
break;
|
|
|
|
|
case PCI_BASE_ADDRESS_MEM_TYPE_64:
|
|
|
|
|
offset += 8;
|
|
|
|
|
break;
|
|
|
|
|
default: /* reserved in PCI 2.2 */
|
|
|
|
|
dev_warn(&pdev->dev,
|
|
|
|
|
"base address is invalid\n");
|
|
|
|
|
return -1;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
if (offset == pci_bar_addr - PCI_BASE_ADDRESS_0)
|
|
|
|
|
return i + 1;
|
|
|
|
|
}
|
|
|
|
|
return -1;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:33:12 +00:00
|
|
|
static void hpsa_disable_interrupt_mode(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
if (h->msix_vector) {
|
|
|
|
|
if (h->pdev->msix_enabled)
|
|
|
|
|
pci_disable_msix(h->pdev);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->msix_vector = 0;
|
2015-04-23 14:33:12 +00:00
|
|
|
} else if (h->msi_vector) {
|
|
|
|
|
if (h->pdev->msi_enabled)
|
|
|
|
|
pci_disable_msi(h->pdev);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->msi_vector = 0;
|
2015-04-23 14:33:12 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/* If MSI/MSI-X is supported by the kernel we will try to enable it on
|
2015-01-23 22:42:22 +00:00
|
|
|
* controllers that are capable. If not, we use legacy INTx mode.
|
2009-12-08 22:09:11 +00:00
|
|
|
*/
|
2012-12-21 21:08:55 +00:00
|
|
|
static void hpsa_interrupt_mode(struct ctlr_info *h)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
|
|
|
|
#ifdef CONFIG_PCI_MSI
|
2012-05-01 16:43:06 +00:00
|
|
|
int err, i;
|
|
|
|
|
struct msix_entry hpsa_msix_entries[MAX_REPLY_QUEUES];
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < MAX_REPLY_QUEUES; i++) {
|
|
|
|
|
hpsa_msix_entries[i].vector = 0;
|
|
|
|
|
hpsa_msix_entries[i].entry = i;
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
/* Some boards advertise MSI but don't really support it */
|
2010-05-27 20:13:02 +00:00
|
|
|
if ((h->board_id == 0x40700E11) || (h->board_id == 0x40800E11) ||
|
|
|
|
|
(h->board_id == 0x40820E11) || (h->board_id == 0x40830E11))
|
2009-12-08 22:09:11 +00:00
|
|
|
goto default_int_mode;
|
2010-05-27 20:12:46 +00:00
|
|
|
if (pci_find_capability(h->pdev, PCI_CAP_ID_MSIX)) {
|
2015-01-23 22:42:22 +00:00
|
|
|
dev_info(&h->pdev->dev, "MSI-X capable controller\n");
|
2014-01-15 12:30:53 +00:00
|
|
|
h->msix_vector = MAX_REPLY_QUEUES;
|
2014-05-29 15:53:02 +00:00
|
|
|
if (h->msix_vector > num_online_cpus())
|
|
|
|
|
h->msix_vector = num_online_cpus();
|
2014-08-18 06:01:42 +00:00
|
|
|
err = pci_enable_msix_range(h->pdev, hpsa_msix_entries,
|
|
|
|
|
1, h->msix_vector);
|
|
|
|
|
if (err < 0) {
|
|
|
|
|
dev_warn(&h->pdev->dev, "MSI-X init failed %d\n", err);
|
|
|
|
|
h->msix_vector = 0;
|
|
|
|
|
goto single_msi_mode;
|
|
|
|
|
} else if (err < h->msix_vector) {
|
2010-05-27 20:12:46 +00:00
|
|
|
dev_warn(&h->pdev->dev, "only %d MSI-X vectors "
|
2009-12-08 22:09:11 +00:00
|
|
|
"available\n", err);
|
|
|
|
|
}
|
2014-08-18 06:01:42 +00:00
|
|
|
h->msix_vector = err;
|
|
|
|
|
for (i = 0; i < h->msix_vector; i++)
|
|
|
|
|
h->intr[i] = hpsa_msix_entries[i].vector;
|
|
|
|
|
return;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
2014-08-18 06:01:42 +00:00
|
|
|
single_msi_mode:
|
2010-05-27 20:12:46 +00:00
|
|
|
if (pci_find_capability(h->pdev, PCI_CAP_ID_MSI)) {
|
2015-01-23 22:42:22 +00:00
|
|
|
dev_info(&h->pdev->dev, "MSI capable controller\n");
|
2010-05-27 20:12:46 +00:00
|
|
|
if (!pci_enable_msi(h->pdev))
|
2009-12-08 22:09:11 +00:00
|
|
|
h->msi_vector = 1;
|
|
|
|
|
else
|
2010-05-27 20:12:46 +00:00
|
|
|
dev_warn(&h->pdev->dev, "MSI init failed\n");
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
default_int_mode:
|
|
|
|
|
#endif /* CONFIG_PCI_MSI */
|
|
|
|
|
/* if we get here we're going to use the default interrupt mode */
|
2011-02-15 21:32:53 +00:00
|
|
|
h->intr[h->intr_mode] = h->pdev->irq;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
2012-12-21 21:08:55 +00:00
|
|
|
static int hpsa_lookup_board_id(struct pci_dev *pdev, u32 *board_id)
|
2010-05-27 20:12:52 +00:00
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
u32 subsystem_vendor_id, subsystem_device_id;
|
|
|
|
|
|
|
|
|
|
subsystem_vendor_id = pdev->subsystem_vendor;
|
|
|
|
|
subsystem_device_id = pdev->subsystem_device;
|
|
|
|
|
*board_id = ((subsystem_device_id << 16) & 0xffff0000) |
|
|
|
|
|
subsystem_vendor_id;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < ARRAY_SIZE(products); i++)
|
|
|
|
|
if (*board_id == products[i].board_id)
|
|
|
|
|
return i;
|
|
|
|
|
|
2010-06-16 18:51:20 +00:00
|
|
|
if ((subsystem_vendor_id != PCI_VENDOR_ID_HP &&
|
|
|
|
|
subsystem_vendor_id != PCI_VENDOR_ID_COMPAQ) ||
|
|
|
|
|
!hpsa_allow_any) {
|
2010-05-27 20:12:52 +00:00
|
|
|
dev_warn(&pdev->dev, "unrecognized board ID: "
|
|
|
|
|
"0x%08x, ignoring.\n", *board_id);
|
|
|
|
|
return -ENODEV;
|
|
|
|
|
}
|
|
|
|
|
return ARRAY_SIZE(products) - 1; /* generic unknown smart array */
|
|
|
|
|
}
|
|
|
|
|
|
2012-12-21 21:08:55 +00:00
|
|
|
static int hpsa_pci_find_memory_BAR(struct pci_dev *pdev,
|
|
|
|
|
unsigned long *memory_bar)
|
2010-05-27 20:13:07 +00:00
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < DEVICE_COUNT_RESOURCE; i++)
|
2010-06-16 18:51:25 +00:00
|
|
|
if (pci_resource_flags(pdev, i) & IORESOURCE_MEM) {
|
2010-05-27 20:13:07 +00:00
|
|
|
/* addressing mode bits already removed */
|
2010-06-16 18:51:25 +00:00
|
|
|
*memory_bar = pci_resource_start(pdev, i);
|
|
|
|
|
dev_dbg(&pdev->dev, "memory BAR = %lx\n",
|
2010-05-27 20:13:07 +00:00
|
|
|
*memory_bar);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
2010-06-16 18:51:25 +00:00
|
|
|
dev_warn(&pdev->dev, "no memory BAR found\n");
|
2010-05-27 20:13:07 +00:00
|
|
|
return -ENODEV;
|
|
|
|
|
}
|
|
|
|
|
|
2012-12-21 21:08:55 +00:00
|
|
|
static int hpsa_wait_for_board_state(struct pci_dev *pdev, void __iomem *vaddr,
|
|
|
|
|
int wait_for_ready)
|
2010-05-27 20:13:12 +00:00
|
|
|
{
|
2011-01-06 20:48:03 +00:00
|
|
|
int i, iterations;
|
2010-05-27 20:13:12 +00:00
|
|
|
u32 scratchpad;
|
2011-01-06 20:48:03 +00:00
|
|
|
if (wait_for_ready)
|
|
|
|
|
iterations = HPSA_BOARD_READY_ITERATIONS;
|
|
|
|
|
else
|
|
|
|
|
iterations = HPSA_BOARD_NOT_READY_ITERATIONS;
|
2010-05-27 20:13:12 +00:00
|
|
|
|
2011-01-06 20:48:03 +00:00
|
|
|
for (i = 0; i < iterations; i++) {
|
|
|
|
|
scratchpad = readl(vaddr + SA5_SCRATCHPAD_OFFSET);
|
|
|
|
|
if (wait_for_ready) {
|
|
|
|
|
if (scratchpad == HPSA_FIRMWARE_READY)
|
|
|
|
|
return 0;
|
|
|
|
|
} else {
|
|
|
|
|
if (scratchpad != HPSA_FIRMWARE_READY)
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
2010-05-27 20:13:12 +00:00
|
|
|
msleep(HPSA_BOARD_READY_POLL_INTERVAL_MSECS);
|
|
|
|
|
}
|
2011-01-06 20:48:03 +00:00
|
|
|
dev_warn(&pdev->dev, "board not ready, timed out.\n");
|
2010-05-27 20:13:12 +00:00
|
|
|
return -ENODEV;
|
|
|
|
|
}
|
|
|
|
|
|
2012-12-21 21:08:55 +00:00
|
|
|
static int hpsa_find_cfg_addrs(struct pci_dev *pdev, void __iomem *vaddr,
|
|
|
|
|
u32 *cfg_base_addr, u64 *cfg_base_addr_index,
|
|
|
|
|
u64 *cfg_offset)
|
2010-06-16 18:51:30 +00:00
|
|
|
{
|
|
|
|
|
*cfg_base_addr = readl(vaddr + SA5_CTCFG_OFFSET);
|
|
|
|
|
*cfg_offset = readl(vaddr + SA5_CTMEM_OFFSET);
|
|
|
|
|
*cfg_base_addr &= (u32) 0x0000ffff;
|
|
|
|
|
*cfg_base_addr_index = find_PCI_BAR_index(pdev, *cfg_base_addr);
|
|
|
|
|
if (*cfg_base_addr_index == -1) {
|
|
|
|
|
dev_warn(&pdev->dev, "cannot find cfg_base_addr_index\n");
|
|
|
|
|
return -ENODEV;
|
|
|
|
|
}
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:33:17 +00:00
|
|
|
static void hpsa_free_cfgtables(struct ctlr_info *h)
|
|
|
|
|
{
|
2015-04-23 14:33:48 +00:00
|
|
|
if (h->transtable) {
|
2015-04-23 14:33:17 +00:00
|
|
|
iounmap(h->transtable);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->transtable = NULL;
|
|
|
|
|
}
|
|
|
|
|
if (h->cfgtable) {
|
2015-04-23 14:33:17 +00:00
|
|
|
iounmap(h->cfgtable);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->cfgtable = NULL;
|
|
|
|
|
}
|
2015-04-23 14:33:17 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Find and map CISS config table and transfer table
|
|
|
|
|
+ * several items must be unmapped (freed) later
|
|
|
|
|
+ * */
|
2012-12-21 21:08:55 +00:00
|
|
|
static int hpsa_find_cfgtables(struct ctlr_info *h)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
2010-02-04 14:41:33 +00:00
|
|
|
u64 cfg_offset;
|
|
|
|
|
u32 cfg_base_addr;
|
|
|
|
|
u64 cfg_base_addr_index;
|
2010-02-04 14:42:40 +00:00
|
|
|
u32 trans_offset;
|
2010-06-16 18:51:30 +00:00
|
|
|
int rc;
|
2010-05-27 20:13:17 +00:00
|
|
|
|
2010-06-16 18:51:30 +00:00
|
|
|
rc = hpsa_find_cfg_addrs(h->pdev, h->vaddr, &cfg_base_addr,
|
|
|
|
|
&cfg_base_addr_index, &cfg_offset);
|
|
|
|
|
if (rc)
|
|
|
|
|
return rc;
|
2010-05-27 20:13:17 +00:00
|
|
|
h->cfgtable = remap_pci_mem(pci_resource_start(h->pdev,
|
2010-06-16 18:51:30 +00:00
|
|
|
cfg_base_addr_index) + cfg_offset, sizeof(*h->cfgtable));
|
2015-01-23 22:42:27 +00:00
|
|
|
if (!h->cfgtable) {
|
|
|
|
|
dev_err(&h->pdev->dev, "Failed mapping cfgtable\n");
|
2010-05-27 20:13:17 +00:00
|
|
|
return -ENOMEM;
|
2015-01-23 22:42:27 +00:00
|
|
|
}
|
2011-05-03 19:59:10 +00:00
|
|
|
rc = write_driver_ver_to_cfgtable(h->cfgtable);
|
|
|
|
|
if (rc)
|
|
|
|
|
return rc;
|
2010-05-27 20:13:17 +00:00
|
|
|
/* Find performant mode table. */
|
2010-06-16 18:51:30 +00:00
|
|
|
trans_offset = readl(&h->cfgtable->TransMethodOffset);
|
2010-05-27 20:13:17 +00:00
|
|
|
h->transtable = remap_pci_mem(pci_resource_start(h->pdev,
|
|
|
|
|
cfg_base_addr_index)+cfg_offset+trans_offset,
|
|
|
|
|
sizeof(*h->transtable));
|
2015-04-23 14:33:17 +00:00
|
|
|
if (!h->transtable) {
|
|
|
|
|
dev_err(&h->pdev->dev, "Failed mapping transfer table\n");
|
|
|
|
|
hpsa_free_cfgtables(h);
|
2010-05-27 20:13:17 +00:00
|
|
|
return -ENOMEM;
|
2015-04-23 14:33:17 +00:00
|
|
|
}
|
2010-05-27 20:13:17 +00:00
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2012-12-21 21:08:55 +00:00
|
|
|
static void hpsa_get_max_perf_mode_cmds(struct ctlr_info *h)
|
2010-06-16 18:51:56 +00:00
|
|
|
{
|
2015-04-23 14:31:47 +00:00
|
|
|
#define MIN_MAX_COMMANDS 16
|
|
|
|
|
BUILD_BUG_ON(MIN_MAX_COMMANDS <= HPSA_NRESERVED_CMDS);
|
|
|
|
|
|
|
|
|
|
h->max_commands = readl(&h->cfgtable->MaxPerformantModeCommands);
|
2011-01-06 20:48:13 +00:00
|
|
|
|
|
|
|
|
/* Limit commands in memory limited kdump scenario. */
|
|
|
|
|
if (reset_devices && h->max_commands > 32)
|
|
|
|
|
h->max_commands = 32;
|
|
|
|
|
|
2015-04-23 14:31:47 +00:00
|
|
|
if (h->max_commands < MIN_MAX_COMMANDS) {
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"Controller reports max supported commands of %d Using %d instead. Ensure that firmware is up to date.\n",
|
|
|
|
|
h->max_commands,
|
|
|
|
|
MIN_MAX_COMMANDS);
|
|
|
|
|
h->max_commands = MIN_MAX_COMMANDS;
|
2010-06-16 18:51:56 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:42:17 +00:00
|
|
|
/* If the controller reports that the total max sg entries is greater than 512,
|
|
|
|
|
* then we know that chained SG blocks work. (Original smart arrays did not
|
|
|
|
|
* support chained SG blocks and would return zero for max sg entries.)
|
|
|
|
|
*/
|
|
|
|
|
static int hpsa_supports_chained_sg_blocks(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
return h->maxsgentries > 512;
|
|
|
|
|
}
|
|
|
|
|
|
2010-05-27 20:13:27 +00:00
|
|
|
/* Interrogate the hardware for some limits:
|
|
|
|
|
* max commands, max SG elements without chaining, and with chaining,
|
|
|
|
|
* SG chain block size, etc.
|
|
|
|
|
*/
|
2012-12-21 21:08:55 +00:00
|
|
|
static void hpsa_find_board_params(struct ctlr_info *h)
|
2010-05-27 20:13:27 +00:00
|
|
|
{
|
2010-06-16 18:51:56 +00:00
|
|
|
hpsa_get_max_perf_mode_cmds(h);
|
2015-01-23 22:43:04 +00:00
|
|
|
h->nr_cmds = h->max_commands;
|
2010-05-27 20:13:27 +00:00
|
|
|
h->maxsgentries = readl(&(h->cfgtable->MaxScatterGatherElements));
|
2014-02-18 19:55:33 +00:00
|
|
|
h->fw_support = readl(&(h->cfgtable->misc_fw_support));
|
2015-01-23 22:42:17 +00:00
|
|
|
if (hpsa_supports_chained_sg_blocks(h)) {
|
|
|
|
|
/* Limit in-command s/g elements to 32 save dma'able memory. */
|
2010-05-27 20:13:27 +00:00
|
|
|
h->max_cmd_sg_entries = 32;
|
2014-11-14 23:26:43 +00:00
|
|
|
h->chainsize = h->maxsgentries - h->max_cmd_sg_entries;
|
2010-05-27 20:13:27 +00:00
|
|
|
h->maxsgentries--; /* save one for chain pointer */
|
|
|
|
|
} else {
|
2015-01-23 22:42:17 +00:00
|
|
|
/*
|
|
|
|
|
* Original smart arrays supported at most 31 s/g entries
|
|
|
|
|
* embedded inline in the command (trying to use more
|
|
|
|
|
* would lock up the controller)
|
|
|
|
|
*/
|
|
|
|
|
h->max_cmd_sg_entries = 31;
|
2014-11-14 23:26:43 +00:00
|
|
|
h->maxsgentries = 31; /* default to traditional values */
|
2015-01-23 22:42:17 +00:00
|
|
|
h->chainsize = 0;
|
2010-05-27 20:13:27 +00:00
|
|
|
}
|
2012-05-01 16:42:51 +00:00
|
|
|
|
|
|
|
|
/* Find out what task management functions are supported and cache */
|
|
|
|
|
h->TMFSupportFlags = readl(&(h->cfgtable->TMFSupportFlags));
|
2014-02-18 19:55:59 +00:00
|
|
|
if (!(HPSATMF_PHYS_TASK_ABORT & h->TMFSupportFlags))
|
|
|
|
|
dev_warn(&h->pdev->dev, "Physical aborts not supported\n");
|
|
|
|
|
if (!(HPSATMF_LOG_TASK_ABORT & h->TMFSupportFlags))
|
|
|
|
|
dev_warn(&h->pdev->dev, "Logical aborts not supported\n");
|
2015-04-23 14:34:06 +00:00
|
|
|
if (!(HPSATMF_IOACCEL_ENABLED & h->TMFSupportFlags))
|
|
|
|
|
dev_warn(&h->pdev->dev, "HP SSD Smart Path aborts not supported\n");
|
2010-05-27 20:13:27 +00:00
|
|
|
}
|
|
|
|
|
|
2010-05-27 20:13:32 +00:00
|
|
|
static inline bool hpsa_CISS_signature_present(struct ctlr_info *h)
|
|
|
|
|
{
|
2012-04-04 13:14:59 +00:00
|
|
|
if (!check_signature(h->cfgtable->Signature, "CISS", 4)) {
|
2015-01-23 22:42:22 +00:00
|
|
|
dev_err(&h->pdev->dev, "not a valid CISS config table\n");
|
2010-05-27 20:13:32 +00:00
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
|
2013-12-04 23:10:16 +00:00
|
|
|
static inline void hpsa_set_driver_support_bits(struct ctlr_info *h)
|
2010-05-27 20:13:38 +00:00
|
|
|
{
|
2013-12-04 23:10:16 +00:00
|
|
|
u32 driver_support;
|
2010-05-27 20:13:38 +00:00
|
|
|
|
2013-12-04 23:10:16 +00:00
|
|
|
driver_support = readl(&(h->cfgtable->driver_support));
|
2014-06-26 13:44:52 +00:00
|
|
|
/* Need to enable prefetch in the SCSI core for 6400 in x86 */
|
|
|
|
|
#ifdef CONFIG_X86
|
2013-12-04 23:10:16 +00:00
|
|
|
driver_support |= ENABLE_SCSI_PREFETCH;
|
2010-05-27 20:13:38 +00:00
|
|
|
#endif
|
2013-12-04 23:10:21 +00:00
|
|
|
driver_support |= ENABLE_UNIT_ATTN;
|
|
|
|
|
writel(driver_support, &(h->cfgtable->driver_support));
|
2010-05-27 20:13:38 +00:00
|
|
|
}
|
|
|
|
|
|
2010-05-27 20:13:43 +00:00
|
|
|
/* Disable DMA prefetch for the P600. Otherwise an ASIC bug may result
|
|
|
|
|
* in a prefetch beyond physical memory.
|
|
|
|
|
*/
|
|
|
|
|
static inline void hpsa_p600_dma_prefetch_quirk(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
u32 dma_prefetch;
|
|
|
|
|
|
|
|
|
|
if (h->board_id != 0x3225103C)
|
|
|
|
|
return;
|
|
|
|
|
dma_prefetch = readl(h->vaddr + I2O_DMA1_CFG);
|
|
|
|
|
dma_prefetch |= 0x8000;
|
|
|
|
|
writel(dma_prefetch, h->vaddr + I2O_DMA1_CFG);
|
|
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:45:01 +00:00
|
|
|
static int hpsa_wait_for_clear_event_notify_ack(struct ctlr_info *h)
|
2014-02-18 19:55:43 +00:00
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
u32 doorbell_value;
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
/* wait until the clear_event_notify bit 6 is cleared by controller. */
|
2015-01-23 22:44:56 +00:00
|
|
|
for (i = 0; i < MAX_CLEAR_EVENT_WAIT; i++) {
|
2014-02-18 19:55:43 +00:00
|
|
|
spin_lock_irqsave(&h->lock, flags);
|
|
|
|
|
doorbell_value = readl(h->vaddr + SA5_DOORBELL);
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
|
|
|
|
if (!(doorbell_value & DOORBELL_CLEAR_EVENTS))
|
2015-01-23 22:45:01 +00:00
|
|
|
goto done;
|
2014-02-18 19:55:43 +00:00
|
|
|
/* delay and try again */
|
2015-01-23 22:44:56 +00:00
|
|
|
msleep(CLEAR_EVENT_WAIT_INTERVAL);
|
2014-02-18 19:55:43 +00:00
|
|
|
}
|
2015-01-23 22:45:01 +00:00
|
|
|
return -ENODEV;
|
|
|
|
|
done:
|
|
|
|
|
return 0;
|
2014-02-18 19:55:43 +00:00
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:45:01 +00:00
|
|
|
static int hpsa_wait_for_mode_change_ack(struct ctlr_info *h)
|
2010-05-27 20:13:48 +00:00
|
|
|
{
|
|
|
|
|
int i;
|
2011-01-06 20:48:24 +00:00
|
|
|
u32 doorbell_value;
|
|
|
|
|
unsigned long flags;
|
2010-05-27 20:13:48 +00:00
|
|
|
|
|
|
|
|
/* under certain very rare conditions, this can take awhile.
|
|
|
|
|
* (e.g.: hot replace a failed 144GB drive in a RAID 5 set right
|
|
|
|
|
* as we enter this code.)
|
|
|
|
|
*/
|
2015-01-23 22:44:56 +00:00
|
|
|
for (i = 0; i < MAX_MODE_CHANGE_WAIT; i++) {
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
if (h->remove_in_progress)
|
|
|
|
|
goto done;
|
2011-01-06 20:48:24 +00:00
|
|
|
spin_lock_irqsave(&h->lock, flags);
|
|
|
|
|
doorbell_value = readl(h->vaddr + SA5_DOORBELL);
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
2011-02-15 21:33:13 +00:00
|
|
|
if (!(doorbell_value & CFGTBL_ChangeReq))
|
2015-01-23 22:45:01 +00:00
|
|
|
goto done;
|
2010-05-27 20:13:48 +00:00
|
|
|
/* delay and try again */
|
2015-01-23 22:44:56 +00:00
|
|
|
msleep(MODE_CHANGE_WAIT_INTERVAL);
|
2010-05-27 20:13:48 +00:00
|
|
|
}
|
2015-01-23 22:45:01 +00:00
|
|
|
return -ENODEV;
|
|
|
|
|
done:
|
|
|
|
|
return 0;
|
2010-05-27 20:14:08 +00:00
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:45:01 +00:00
|
|
|
/* return -ENODEV or other reason on error, 0 on success */
|
2012-12-21 21:08:55 +00:00
|
|
|
static int hpsa_enter_simple_mode(struct ctlr_info *h)
|
2010-05-27 20:14:08 +00:00
|
|
|
{
|
|
|
|
|
u32 trans_support;
|
|
|
|
|
|
|
|
|
|
trans_support = readl(&(h->cfgtable->TransportSupport));
|
|
|
|
|
if (!(trans_support & SIMPLE_MODE))
|
|
|
|
|
return -ENOTSUPP;
|
|
|
|
|
|
|
|
|
|
h->max_commands = readl(&(h->cfgtable->CmdsOutMax));
|
2014-02-18 19:55:33 +00:00
|
|
|
|
2010-05-27 20:14:08 +00:00
|
|
|
/* Update the field, and then ring the doorbell */
|
|
|
|
|
writel(CFGTBL_Trans_Simple, &(h->cfgtable->HostWrite.TransportRequest));
|
2014-02-18 19:56:29 +00:00
|
|
|
writel(0, &h->cfgtable->HostWrite.command_pool_addr_hi);
|
2010-05-27 20:14:08 +00:00
|
|
|
writel(CFGTBL_ChangeReq, h->vaddr + SA5_DOORBELL);
|
2015-01-23 22:45:01 +00:00
|
|
|
if (hpsa_wait_for_mode_change_ack(h))
|
|
|
|
|
goto error;
|
2010-05-27 20:13:48 +00:00
|
|
|
print_cfg_table(&h->pdev->dev, h->cfgtable);
|
2014-02-18 19:55:33 +00:00
|
|
|
if (!(readl(&(h->cfgtable->TransportActive)) & CFGTBL_Trans_Simple))
|
|
|
|
|
goto error;
|
2011-02-15 21:33:03 +00:00
|
|
|
h->transMethod = CFGTBL_Trans_Simple;
|
2010-05-27 20:13:48 +00:00
|
|
|
return 0;
|
2014-02-18 19:55:33 +00:00
|
|
|
error:
|
2015-01-23 22:42:22 +00:00
|
|
|
dev_err(&h->pdev->dev, "failed to enter simple mode\n");
|
2014-02-18 19:55:33 +00:00
|
|
|
return -ENODEV;
|
2010-05-27 20:13:48 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:33:17 +00:00
|
|
|
/* free items allocated or mapped by hpsa_pci_init */
|
|
|
|
|
static void hpsa_free_pci_init(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
hpsa_free_cfgtables(h); /* pci_init 4 */
|
|
|
|
|
iounmap(h->vaddr); /* pci_init 3 */
|
2015-04-23 14:33:48 +00:00
|
|
|
h->vaddr = NULL;
|
2015-04-23 14:33:17 +00:00
|
|
|
hpsa_disable_interrupt_mode(h); /* pci_init 2 */
|
2015-04-23 14:34:32 +00:00
|
|
|
/*
|
|
|
|
|
* call pci_disable_device before pci_release_regions per
|
|
|
|
|
* Documentation/PCI/pci.txt
|
|
|
|
|
*/
|
2015-04-23 14:33:17 +00:00
|
|
|
pci_disable_device(h->pdev); /* pci_init 1 */
|
2015-04-23 14:34:32 +00:00
|
|
|
pci_release_regions(h->pdev); /* pci_init 2 */
|
2015-04-23 14:33:17 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* several items must be freed later */
|
2012-12-21 21:08:55 +00:00
|
|
|
static int hpsa_pci_init(struct ctlr_info *h)
|
2010-05-27 20:13:17 +00:00
|
|
|
{
|
2010-05-27 20:13:48 +00:00
|
|
|
int prod_index, err;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2010-05-27 20:12:52 +00:00
|
|
|
prod_index = hpsa_lookup_board_id(h->pdev, &h->board_id);
|
|
|
|
|
if (prod_index < 0)
|
2015-01-23 22:42:06 +00:00
|
|
|
return prod_index;
|
2010-05-27 20:12:52 +00:00
|
|
|
h->product_name = products[prod_index].product_name;
|
|
|
|
|
h->access = *(products[prod_index].access);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-04-23 14:32:06 +00:00
|
|
|
h->needs_abort_tags_swizzled =
|
|
|
|
|
ctlr_needs_abort_tags_swizzled(h->board_id);
|
|
|
|
|
|
2011-11-11 16:14:23 +00:00
|
|
|
pci_disable_link_state(h->pdev, PCIE_LINK_STATE_L0S |
|
|
|
|
|
PCIE_LINK_STATE_L1 | PCIE_LINK_STATE_CLKPM);
|
|
|
|
|
|
2010-05-27 20:12:46 +00:00
|
|
|
err = pci_enable_device(h->pdev);
|
2009-12-08 22:09:11 +00:00
|
|
|
if (err) {
|
2015-04-23 14:33:17 +00:00
|
|
|
dev_err(&h->pdev->dev, "failed to enable PCI device\n");
|
2015-04-23 14:34:32 +00:00
|
|
|
pci_disable_device(h->pdev);
|
2009-12-08 22:09:11 +00:00
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2012-01-19 20:00:59 +00:00
|
|
|
err = pci_request_regions(h->pdev, HPSA);
|
2009-12-08 22:09:11 +00:00
|
|
|
if (err) {
|
2010-05-27 20:12:46 +00:00
|
|
|
dev_err(&h->pdev->dev,
|
2015-04-23 14:33:17 +00:00
|
|
|
"failed to obtain PCI resources\n");
|
2015-04-23 14:34:32 +00:00
|
|
|
pci_disable_device(h->pdev);
|
|
|
|
|
return err;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
2014-11-14 23:27:24 +00:00
|
|
|
|
|
|
|
|
pci_set_master(h->pdev);
|
|
|
|
|
|
2010-05-27 20:13:02 +00:00
|
|
|
hpsa_interrupt_mode(h);
|
2010-06-16 18:51:25 +00:00
|
|
|
err = hpsa_pci_find_memory_BAR(h->pdev, &h->paddr);
|
2010-05-27 20:13:07 +00:00
|
|
|
if (err)
|
2015-04-23 14:33:17 +00:00
|
|
|
goto clean2; /* intmode+region, pci */
|
2009-12-08 22:09:11 +00:00
|
|
|
h->vaddr = remap_pci_mem(h->paddr, 0x250);
|
2010-05-27 20:13:22 +00:00
|
|
|
if (!h->vaddr) {
|
2015-04-23 14:33:17 +00:00
|
|
|
dev_err(&h->pdev->dev, "failed to remap PCI mem\n");
|
2010-05-27 20:13:22 +00:00
|
|
|
err = -ENOMEM;
|
2015-04-23 14:33:17 +00:00
|
|
|
goto clean2; /* intmode+region, pci */
|
2010-05-27 20:13:22 +00:00
|
|
|
}
|
2011-01-06 20:48:03 +00:00
|
|
|
err = hpsa_wait_for_board_state(h->pdev, h->vaddr, BOARD_READY);
|
2010-05-27 20:13:12 +00:00
|
|
|
if (err)
|
2015-04-23 14:33:17 +00:00
|
|
|
goto clean3; /* vaddr, intmode+region, pci */
|
2010-05-27 20:13:17 +00:00
|
|
|
err = hpsa_find_cfgtables(h);
|
|
|
|
|
if (err)
|
2015-04-23 14:33:17 +00:00
|
|
|
goto clean3; /* vaddr, intmode+region, pci */
|
2010-05-27 20:13:27 +00:00
|
|
|
hpsa_find_board_params(h);
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2010-05-27 20:13:32 +00:00
|
|
|
if (!hpsa_CISS_signature_present(h)) {
|
2009-12-08 22:09:11 +00:00
|
|
|
err = -ENODEV;
|
2015-04-23 14:33:17 +00:00
|
|
|
goto clean4; /* cfgtables, vaddr, intmode+region, pci */
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
2013-12-04 23:10:16 +00:00
|
|
|
hpsa_set_driver_support_bits(h);
|
2010-05-27 20:13:43 +00:00
|
|
|
hpsa_p600_dma_prefetch_quirk(h);
|
2010-05-27 20:13:48 +00:00
|
|
|
err = hpsa_enter_simple_mode(h);
|
|
|
|
|
if (err)
|
2015-04-23 14:33:17 +00:00
|
|
|
goto clean4; /* cfgtables, vaddr, intmode+region, pci */
|
2009-12-08 22:09:11 +00:00
|
|
|
return 0;
|
|
|
|
|
|
2015-04-23 14:33:17 +00:00
|
|
|
clean4: /* cfgtables, vaddr, intmode+region, pci */
|
|
|
|
|
hpsa_free_cfgtables(h);
|
|
|
|
|
clean3: /* vaddr, intmode+region, pci */
|
|
|
|
|
iounmap(h->vaddr);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->vaddr = NULL;
|
2015-04-23 14:33:17 +00:00
|
|
|
clean2: /* intmode+region, pci */
|
|
|
|
|
hpsa_disable_interrupt_mode(h);
|
2015-04-23 14:34:32 +00:00
|
|
|
/*
|
|
|
|
|
* call pci_disable_device before pci_release_regions per
|
|
|
|
|
* Documentation/PCI/pci.txt
|
|
|
|
|
*/
|
2015-04-23 14:33:17 +00:00
|
|
|
pci_disable_device(h->pdev);
|
2015-04-23 14:34:32 +00:00
|
|
|
pci_release_regions(h->pdev);
|
2009-12-08 22:09:11 +00:00
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
2012-12-21 21:08:55 +00:00
|
|
|
static void hpsa_hba_inquiry(struct ctlr_info *h)
|
2010-02-04 14:42:50 +00:00
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
|
|
#define HBA_INQUIRY_BYTE_COUNT 64
|
|
|
|
|
h->hba_inquiry_data = kmalloc(HBA_INQUIRY_BYTE_COUNT, GFP_KERNEL);
|
|
|
|
|
if (!h->hba_inquiry_data)
|
|
|
|
|
return;
|
|
|
|
|
rc = hpsa_scsi_do_inquiry(h, RAID_CTLR_LUNID, 0,
|
|
|
|
|
h->hba_inquiry_data, HBA_INQUIRY_BYTE_COUNT);
|
|
|
|
|
if (rc != 0) {
|
|
|
|
|
kfree(h->hba_inquiry_data);
|
|
|
|
|
h->hba_inquiry_data = NULL;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-02 13:25:54 +00:00
|
|
|
static int hpsa_init_reset_devices(struct pci_dev *pdev, u32 board_id)
|
2010-06-16 18:51:35 +00:00
|
|
|
{
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
int rc, i;
|
2015-01-23 22:41:20 +00:00
|
|
|
void __iomem *vaddr;
|
2010-06-16 18:51:35 +00:00
|
|
|
|
|
|
|
|
if (!reset_devices)
|
|
|
|
|
return 0;
|
|
|
|
|
|
2014-08-14 14:12:39 +00:00
|
|
|
/* kdump kernel is loading, we don't know in which state is
|
|
|
|
|
* the pci interface. The dev->enable_cnt is equal zero
|
|
|
|
|
* so we call enable+disable, wait a while and switch it on.
|
|
|
|
|
*/
|
|
|
|
|
rc = pci_enable_device(pdev);
|
|
|
|
|
if (rc) {
|
|
|
|
|
dev_warn(&pdev->dev, "Failed to enable PCI device\n");
|
|
|
|
|
return -ENODEV;
|
|
|
|
|
}
|
|
|
|
|
pci_disable_device(pdev);
|
|
|
|
|
msleep(260); /* a randomly chosen number */
|
|
|
|
|
rc = pci_enable_device(pdev);
|
|
|
|
|
if (rc) {
|
|
|
|
|
dev_warn(&pdev->dev, "failed to enable device.\n");
|
|
|
|
|
return -ENODEV;
|
|
|
|
|
}
|
2014-11-14 23:27:24 +00:00
|
|
|
|
2014-09-12 12:44:15 +00:00
|
|
|
pci_set_master(pdev);
|
2014-11-14 23:27:24 +00:00
|
|
|
|
2015-01-23 22:41:20 +00:00
|
|
|
vaddr = pci_ioremap_bar(pdev, 0);
|
|
|
|
|
if (vaddr == NULL) {
|
|
|
|
|
rc = -ENOMEM;
|
|
|
|
|
goto out_disable;
|
|
|
|
|
}
|
|
|
|
|
writel(SA5_INTR_OFF, vaddr + SA5_REPLY_INTR_MASK_OFFSET);
|
|
|
|
|
iounmap(vaddr);
|
|
|
|
|
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
/* Reset the controller with a PCI power-cycle or via doorbell */
|
2015-04-02 13:25:54 +00:00
|
|
|
rc = hpsa_kdump_hard_reset_controller(pdev, board_id);
|
2010-06-16 18:51:35 +00:00
|
|
|
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
/* -ENOTSUPP here means we cannot reset the controller
|
|
|
|
|
* but it's already (and still) up and running in
|
2010-06-16 18:51:45 +00:00
|
|
|
* "performant mode". Or, it might be 640x, which can't reset
|
|
|
|
|
* due to concerns about shared bbwc between 6402/6404 pair.
|
[SCSI] hpsa: Fix hard reset code.
Smart Array controllers newer than the P600 do not honor the
PCI power state method of resetting the controllers. Instead,
in these cases we can get them to reset via the "doorbell" register.
This escaped notice until we began using "performant" mode because
the fact that the controllers did not reset did not normally
impede subsequent operation, and so things generally appeared to
"work". Once the performant mode code was added, if the controller
does not reset, it remains in performant mode. The code immediately
after the reset presumes the controller is in "simple" mode
(which previously, it had remained in simple mode the whole time).
If the controller remains in performant mode any code which presumes
it is in simple mode will not work. So the reset needs to be fixed.
Unfortunately there are some controllers which cannot be reset by
either method. (eg. p800). We detect these cases by noticing that
the controller seems to remain in performant mode even after a
reset has been attempted. In those case, we proceed anyway,
as if the reset has happened (and skip the step of waiting for
the controller to become ready -- which is expecting it to be in
"simple" mode.) To sum up, we try to do a better job of resetting
the controller if "reset_devices" is set, and if it doesn't work,
we print a message and try to continue anyway.
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-06-16 18:51:40 +00:00
|
|
|
*/
|
2015-01-23 22:42:01 +00:00
|
|
|
if (rc)
|
2014-08-14 14:12:39 +00:00
|
|
|
goto out_disable;
|
2010-06-16 18:51:35 +00:00
|
|
|
|
|
|
|
|
/* Now try to get the controller to respond to a no-op */
|
2015-01-23 22:42:11 +00:00
|
|
|
dev_info(&pdev->dev, "Waiting for controller to respond to no-op\n");
|
2010-06-16 18:51:35 +00:00
|
|
|
for (i = 0; i < HPSA_POST_RESET_NOOP_RETRIES; i++) {
|
|
|
|
|
if (hpsa_noop(pdev) == 0)
|
|
|
|
|
break;
|
|
|
|
|
else
|
|
|
|
|
dev_warn(&pdev->dev, "no-op failed%s\n",
|
|
|
|
|
(i < 11 ? "; re-trying" : ""));
|
|
|
|
|
}
|
2014-08-14 14:12:39 +00:00
|
|
|
|
|
|
|
|
out_disable:
|
|
|
|
|
|
|
|
|
|
pci_disable_device(pdev);
|
|
|
|
|
return rc;
|
2010-06-16 18:51:35 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:33:22 +00:00
|
|
|
static void hpsa_free_cmd_pool(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
kfree(h->cmd_pool_bits);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->cmd_pool_bits = NULL;
|
|
|
|
|
if (h->cmd_pool) {
|
2015-04-23 14:33:22 +00:00
|
|
|
pci_free_consistent(h->pdev,
|
|
|
|
|
h->nr_cmds * sizeof(struct CommandList),
|
|
|
|
|
h->cmd_pool,
|
|
|
|
|
h->cmd_pool_dhandle);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->cmd_pool = NULL;
|
|
|
|
|
h->cmd_pool_dhandle = 0;
|
|
|
|
|
}
|
|
|
|
|
if (h->errinfo_pool) {
|
2015-04-23 14:33:22 +00:00
|
|
|
pci_free_consistent(h->pdev,
|
|
|
|
|
h->nr_cmds * sizeof(struct ErrorInfo),
|
|
|
|
|
h->errinfo_pool,
|
|
|
|
|
h->errinfo_pool_dhandle);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->errinfo_pool = NULL;
|
|
|
|
|
h->errinfo_pool_dhandle = 0;
|
|
|
|
|
}
|
2015-04-23 14:33:22 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:32:27 +00:00
|
|
|
static int hpsa_alloc_cmd_pool(struct ctlr_info *h)
|
2011-05-03 19:59:20 +00:00
|
|
|
{
|
|
|
|
|
h->cmd_pool_bits = kzalloc(
|
|
|
|
|
DIV_ROUND_UP(h->nr_cmds, BITS_PER_LONG) *
|
|
|
|
|
sizeof(unsigned long), GFP_KERNEL);
|
|
|
|
|
h->cmd_pool = pci_alloc_consistent(h->pdev,
|
|
|
|
|
h->nr_cmds * sizeof(*h->cmd_pool),
|
|
|
|
|
&(h->cmd_pool_dhandle));
|
|
|
|
|
h->errinfo_pool = pci_alloc_consistent(h->pdev,
|
|
|
|
|
h->nr_cmds * sizeof(*h->errinfo_pool),
|
|
|
|
|
&(h->errinfo_pool_dhandle));
|
|
|
|
|
if ((h->cmd_pool_bits == NULL)
|
|
|
|
|
|| (h->cmd_pool == NULL)
|
|
|
|
|
|| (h->errinfo_pool == NULL)) {
|
|
|
|
|
dev_err(&h->pdev->dev, "out of memory in %s", __func__);
|
2015-01-23 22:42:48 +00:00
|
|
|
goto clean_up;
|
2011-05-03 19:59:20 +00:00
|
|
|
}
|
2015-04-23 14:32:32 +00:00
|
|
|
hpsa_preinitialize_commands(h);
|
2011-05-03 19:59:20 +00:00
|
|
|
return 0;
|
2015-01-23 22:42:48 +00:00
|
|
|
clean_up:
|
|
|
|
|
hpsa_free_cmd_pool(h);
|
|
|
|
|
return -ENOMEM;
|
2011-05-03 19:59:20 +00:00
|
|
|
}
|
|
|
|
|
|
2014-05-29 15:53:13 +00:00
|
|
|
static void hpsa_irq_affinity_hints(struct ctlr_info *h)
|
|
|
|
|
{
|
2015-01-23 22:41:46 +00:00
|
|
|
int i, cpu;
|
2014-05-29 15:53:13 +00:00
|
|
|
|
|
|
|
|
cpu = cpumask_first(cpu_online_mask);
|
|
|
|
|
for (i = 0; i < h->msix_vector; i++) {
|
2015-01-23 22:41:46 +00:00
|
|
|
irq_set_affinity_hint(h->intr[i], get_cpu_mask(cpu));
|
2014-05-29 15:53:13 +00:00
|
|
|
cpu = cpumask_next(cpu, cpu_online_mask);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:41:40 +00:00
|
|
|
/* clear affinity hints and free MSI-X, MSI, or legacy INTx vectors */
|
|
|
|
|
static void hpsa_free_irqs(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
if (!h->msix_vector || h->intr_mode != PERF_MODE_INT) {
|
|
|
|
|
/* Single reply queue, only one irq to free */
|
|
|
|
|
i = h->intr_mode;
|
|
|
|
|
irq_set_affinity_hint(h->intr[i], NULL);
|
|
|
|
|
free_irq(h->intr[i], &h->q[i]);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->q[i] = 0;
|
2015-01-23 22:41:40 +00:00
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < h->msix_vector; i++) {
|
|
|
|
|
irq_set_affinity_hint(h->intr[i], NULL);
|
|
|
|
|
free_irq(h->intr[i], &h->q[i]);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->q[i] = 0;
|
2015-01-23 22:41:40 +00:00
|
|
|
}
|
2015-01-23 22:41:51 +00:00
|
|
|
for (; i < MAX_REPLY_QUEUES; i++)
|
|
|
|
|
h->q[i] = 0;
|
2015-01-23 22:41:40 +00:00
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:42:32 +00:00
|
|
|
/* returns 0 on success; cleans up and returns -Enn on error */
|
|
|
|
|
static int hpsa_request_irqs(struct ctlr_info *h,
|
2011-05-03 19:59:25 +00:00
|
|
|
irqreturn_t (*msixhandler)(int, void *),
|
|
|
|
|
irqreturn_t (*intxhandler)(int, void *))
|
|
|
|
|
{
|
2012-05-01 16:43:06 +00:00
|
|
|
int rc, i;
|
2011-05-03 19:59:25 +00:00
|
|
|
|
2012-05-01 16:43:06 +00:00
|
|
|
/*
|
|
|
|
|
* initialize h->q[x] = x so that interrupt handlers know which
|
|
|
|
|
* queue to process.
|
|
|
|
|
*/
|
|
|
|
|
for (i = 0; i < MAX_REPLY_QUEUES; i++)
|
|
|
|
|
h->q[i] = (u8) i;
|
|
|
|
|
|
2014-01-15 12:30:53 +00:00
|
|
|
if (h->intr_mode == PERF_MODE_INT && h->msix_vector > 0) {
|
2012-05-01 16:43:06 +00:00
|
|
|
/* If performant mode and MSI-X, use multiple reply queues */
|
2015-01-23 22:41:51 +00:00
|
|
|
for (i = 0; i < h->msix_vector; i++) {
|
2015-04-23 14:34:58 +00:00
|
|
|
sprintf(h->intrname[i], "%s-msix%d", h->devname, i);
|
2012-05-01 16:43:06 +00:00
|
|
|
rc = request_irq(h->intr[i], msixhandler,
|
2015-04-23 14:34:58 +00:00
|
|
|
0, h->intrname[i],
|
2012-05-01 16:43:06 +00:00
|
|
|
&h->q[i]);
|
2015-01-23 22:41:51 +00:00
|
|
|
if (rc) {
|
|
|
|
|
int j;
|
|
|
|
|
|
|
|
|
|
dev_err(&h->pdev->dev,
|
|
|
|
|
"failed to get irq %d for %s\n",
|
|
|
|
|
h->intr[i], h->devname);
|
|
|
|
|
for (j = 0; j < i; j++) {
|
|
|
|
|
free_irq(h->intr[j], &h->q[j]);
|
|
|
|
|
h->q[j] = 0;
|
|
|
|
|
}
|
|
|
|
|
for (; j < MAX_REPLY_QUEUES; j++)
|
|
|
|
|
h->q[j] = 0;
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
}
|
2014-05-29 15:53:13 +00:00
|
|
|
hpsa_irq_affinity_hints(h);
|
2012-05-01 16:43:06 +00:00
|
|
|
} else {
|
|
|
|
|
/* Use single reply pool */
|
2014-01-15 12:30:53 +00:00
|
|
|
if (h->msix_vector > 0 || h->msi_vector) {
|
2015-04-23 14:34:58 +00:00
|
|
|
if (h->msix_vector)
|
|
|
|
|
sprintf(h->intrname[h->intr_mode],
|
|
|
|
|
"%s-msix", h->devname);
|
|
|
|
|
else
|
|
|
|
|
sprintf(h->intrname[h->intr_mode],
|
|
|
|
|
"%s-msi", h->devname);
|
2012-05-01 16:43:06 +00:00
|
|
|
rc = request_irq(h->intr[h->intr_mode],
|
2015-04-23 14:34:58 +00:00
|
|
|
msixhandler, 0,
|
|
|
|
|
h->intrname[h->intr_mode],
|
2012-05-01 16:43:06 +00:00
|
|
|
&h->q[h->intr_mode]);
|
|
|
|
|
} else {
|
2015-04-23 14:34:58 +00:00
|
|
|
sprintf(h->intrname[h->intr_mode],
|
|
|
|
|
"%s-intx", h->devname);
|
2012-05-01 16:43:06 +00:00
|
|
|
rc = request_irq(h->intr[h->intr_mode],
|
2015-04-23 14:34:58 +00:00
|
|
|
intxhandler, IRQF_SHARED,
|
|
|
|
|
h->intrname[h->intr_mode],
|
2012-05-01 16:43:06 +00:00
|
|
|
&h->q[h->intr_mode]);
|
|
|
|
|
}
|
2015-04-23 14:33:48 +00:00
|
|
|
irq_set_affinity_hint(h->intr[h->intr_mode], NULL);
|
2012-05-01 16:43:06 +00:00
|
|
|
}
|
2011-05-03 19:59:25 +00:00
|
|
|
if (rc) {
|
2015-04-23 14:33:17 +00:00
|
|
|
dev_err(&h->pdev->dev, "failed to get irq %d for %s\n",
|
2011-05-03 19:59:25 +00:00
|
|
|
h->intr[h->intr_mode], h->devname);
|
2015-04-23 14:33:17 +00:00
|
|
|
hpsa_free_irqs(h);
|
2011-05-03 19:59:25 +00:00
|
|
|
return -ENODEV;
|
|
|
|
|
}
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2012-12-21 21:08:55 +00:00
|
|
|
static int hpsa_kdump_soft_reset(struct ctlr_info *h)
|
2011-05-03 19:59:51 +00:00
|
|
|
{
|
2015-04-23 14:35:14 +00:00
|
|
|
int rc;
|
2015-04-23 14:33:38 +00:00
|
|
|
hpsa_send_host_reset(h, RAID_CTLR_LUNID, HPSA_RESET_TYPE_CONTROLLER);
|
2011-05-03 19:59:51 +00:00
|
|
|
|
|
|
|
|
dev_info(&h->pdev->dev, "Waiting for board to soft reset.\n");
|
2015-04-23 14:35:14 +00:00
|
|
|
rc = hpsa_wait_for_board_state(h->pdev, h->vaddr, BOARD_NOT_READY);
|
|
|
|
|
if (rc) {
|
2011-05-03 19:59:51 +00:00
|
|
|
dev_warn(&h->pdev->dev, "Soft reset had no effect.\n");
|
2015-04-23 14:35:14 +00:00
|
|
|
return rc;
|
2011-05-03 19:59:51 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
dev_info(&h->pdev->dev, "Board reset, awaiting READY status.\n");
|
2015-04-23 14:35:14 +00:00
|
|
|
rc = hpsa_wait_for_board_state(h->pdev, h->vaddr, BOARD_READY);
|
|
|
|
|
if (rc) {
|
2011-05-03 19:59:51 +00:00
|
|
|
dev_warn(&h->pdev->dev, "Board failed to become ready "
|
|
|
|
|
"after soft reset.\n");
|
2015-04-23 14:35:14 +00:00
|
|
|
return rc;
|
2011-05-03 19:59:51 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2014-05-29 15:53:07 +00:00
|
|
|
static void hpsa_free_reply_queues(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < h->nreply_queues; i++) {
|
|
|
|
|
if (!h->reply_queue[i].head)
|
|
|
|
|
continue;
|
2015-04-23 14:33:22 +00:00
|
|
|
pci_free_consistent(h->pdev,
|
|
|
|
|
h->reply_queue_size,
|
|
|
|
|
h->reply_queue[i].head,
|
|
|
|
|
h->reply_queue[i].busaddr);
|
2014-05-29 15:53:07 +00:00
|
|
|
h->reply_queue[i].head = NULL;
|
|
|
|
|
h->reply_queue[i].busaddr = 0;
|
|
|
|
|
}
|
2015-04-23 14:33:48 +00:00
|
|
|
h->reply_queue_size = 0;
|
2014-05-29 15:53:07 +00:00
|
|
|
}
|
|
|
|
|
|
2012-05-01 16:43:21 +00:00
|
|
|
static void hpsa_undo_allocations_after_kdump_soft_reset(struct ctlr_info *h)
|
|
|
|
|
{
|
2015-04-23 14:33:48 +00:00
|
|
|
hpsa_free_performant_mode(h); /* init_one 7 */
|
|
|
|
|
hpsa_free_sg_chain_blocks(h); /* init_one 6 */
|
|
|
|
|
hpsa_free_cmd_pool(h); /* init_one 5 */
|
|
|
|
|
hpsa_free_irqs(h); /* init_one 4 */
|
2015-04-23 14:35:09 +00:00
|
|
|
scsi_host_put(h->scsi_host); /* init_one 3 */
|
|
|
|
|
h->scsi_host = NULL; /* init_one 3 */
|
|
|
|
|
hpsa_free_pci_init(h); /* init_one 2_5 */
|
2015-04-23 14:34:43 +00:00
|
|
|
free_percpu(h->lockup_detected); /* init_one 2 */
|
|
|
|
|
h->lockup_detected = NULL; /* init_one 2 */
|
|
|
|
|
if (h->resubmit_wq) {
|
|
|
|
|
destroy_workqueue(h->resubmit_wq); /* init_one 1 */
|
|
|
|
|
h->resubmit_wq = NULL;
|
|
|
|
|
}
|
|
|
|
|
if (h->rescan_ctlr_wq) {
|
|
|
|
|
destroy_workqueue(h->rescan_ctlr_wq);
|
|
|
|
|
h->rescan_ctlr_wq = NULL;
|
|
|
|
|
}
|
2015-04-23 14:33:48 +00:00
|
|
|
kfree(h); /* init_one 1 */
|
2011-05-03 19:59:51 +00:00
|
|
|
}
|
|
|
|
|
|
2011-10-26 21:22:04 +00:00
|
|
|
/* Called when controller lockup detected. */
|
2015-01-23 22:43:09 +00:00
|
|
|
static void fail_all_outstanding_cmds(struct ctlr_info *h)
|
2011-10-26 21:22:04 +00:00
|
|
|
{
|
2015-01-23 22:43:35 +00:00
|
|
|
int i, refcount;
|
|
|
|
|
struct CommandList *c;
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
int failcount = 0;
|
2011-10-26 21:22:04 +00:00
|
|
|
|
2015-01-23 22:43:25 +00:00
|
|
|
flush_workqueue(h->resubmit_wq); /* ensure all cmds are fully built */
|
2015-01-23 22:43:09 +00:00
|
|
|
for (i = 0; i < h->nr_cmds; i++) {
|
|
|
|
|
c = h->cmd_pool + i;
|
2015-01-23 22:43:35 +00:00
|
|
|
refcount = atomic_inc_return(&c->refcount);
|
|
|
|
|
if (refcount > 1) {
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
c->err_info->CommandStatus = CMD_CTLR_LOCKUP;
|
2015-01-23 22:43:35 +00:00
|
|
|
finish_cmd(c);
|
2015-04-23 14:32:11 +00:00
|
|
|
atomic_dec(&h->commands_outstanding);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
failcount++;
|
2015-01-23 22:43:35 +00:00
|
|
|
}
|
|
|
|
|
cmd_free(h, c);
|
2011-10-26 21:22:04 +00:00
|
|
|
}
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"failed %d commands in fail_all\n", failcount);
|
2011-10-26 21:22:04 +00:00
|
|
|
}
|
|
|
|
|
|
2014-05-29 15:53:18 +00:00
|
|
|
static void set_lockup_detected_for_all_cpus(struct ctlr_info *h, u32 value)
|
|
|
|
|
{
|
2015-03-05 00:19:19 +00:00
|
|
|
int cpu;
|
2014-05-29 15:53:18 +00:00
|
|
|
|
2015-03-05 00:19:19 +00:00
|
|
|
for_each_online_cpu(cpu) {
|
2014-05-29 15:53:18 +00:00
|
|
|
u32 *lockup_detected;
|
|
|
|
|
lockup_detected = per_cpu_ptr(h->lockup_detected, cpu);
|
|
|
|
|
*lockup_detected = value;
|
|
|
|
|
}
|
|
|
|
|
wmb(); /* be sure the per-cpu variables are out to memory */
|
|
|
|
|
}
|
|
|
|
|
|
2011-10-26 21:22:04 +00:00
|
|
|
static void controller_lockup_detected(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
2014-05-29 15:53:18 +00:00
|
|
|
u32 lockup_detected;
|
2011-10-26 21:22:04 +00:00
|
|
|
|
|
|
|
|
h->access.set_intr_mask(h, HPSA_INTR_OFF);
|
|
|
|
|
spin_lock_irqsave(&h->lock, flags);
|
2014-05-29 15:53:18 +00:00
|
|
|
lockup_detected = readl(h->vaddr + SA5_SCRATCHPAD_OFFSET);
|
|
|
|
|
if (!lockup_detected) {
|
|
|
|
|
/* no heartbeat, but controller gave us a zero. */
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
"lockup detected after %d but scratchpad register is zero\n",
|
|
|
|
|
h->heartbeat_sample_interval / HZ);
|
2014-05-29 15:53:18 +00:00
|
|
|
lockup_detected = 0xffffffff;
|
|
|
|
|
}
|
|
|
|
|
set_lockup_detected_for_all_cpus(h, lockup_detected);
|
2011-10-26 21:22:04 +00:00
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
dev_warn(&h->pdev->dev, "Controller lockup detected: 0x%08x after %d\n",
|
|
|
|
|
lockup_detected, h->heartbeat_sample_interval / HZ);
|
2011-10-26 21:22:04 +00:00
|
|
|
pci_disable_device(h->pdev);
|
2015-01-23 22:43:09 +00:00
|
|
|
fail_all_outstanding_cmds(h);
|
2011-10-26 21:22:04 +00:00
|
|
|
}
|
|
|
|
|
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
static int detect_controller_lockup(struct ctlr_info *h)
|
2011-10-26 21:22:04 +00:00
|
|
|
{
|
|
|
|
|
u64 now;
|
|
|
|
|
u32 heartbeat;
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
|
|
now = get_jiffies_64();
|
|
|
|
|
/* If we've received an interrupt recently, we're ok. */
|
|
|
|
|
if (time_after64(h->last_intr_timestamp +
|
2012-05-01 16:43:42 +00:00
|
|
|
(h->heartbeat_sample_interval), now))
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
return false;
|
2011-10-26 21:22:04 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* If we've already checked the heartbeat recently, we're ok.
|
|
|
|
|
* This could happen if someone sends us a signal. We
|
|
|
|
|
* otherwise don't care about signals in this thread.
|
|
|
|
|
*/
|
|
|
|
|
if (time_after64(h->last_heartbeat_timestamp +
|
2012-05-01 16:43:42 +00:00
|
|
|
(h->heartbeat_sample_interval), now))
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
return false;
|
2011-10-26 21:22:04 +00:00
|
|
|
|
|
|
|
|
/* If heartbeat has not changed since we last looked, we're not ok. */
|
|
|
|
|
spin_lock_irqsave(&h->lock, flags);
|
|
|
|
|
heartbeat = readl(&h->cfgtable->HeartBeat);
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
|
|
|
|
if (h->last_heartbeat == heartbeat) {
|
|
|
|
|
controller_lockup_detected(h);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
return true;
|
2011-10-26 21:22:04 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* We're ok. */
|
|
|
|
|
h->last_heartbeat = heartbeat;
|
|
|
|
|
h->last_heartbeat_timestamp = now;
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
return false;
|
2011-10-26 21:22:04 +00:00
|
|
|
}
|
|
|
|
|
|
2014-02-21 22:25:00 +00:00
|
|
|
static void hpsa_ack_ctlr_events(struct ctlr_info *h)
|
2014-02-18 19:55:43 +00:00
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
char *event_type;
|
|
|
|
|
|
2015-01-23 22:44:07 +00:00
|
|
|
if (!(h->fw_support & MISC_FW_EVENT_NOTIFY))
|
|
|
|
|
return;
|
|
|
|
|
|
2014-02-18 19:55:43 +00:00
|
|
|
/* Ask the controller to clear the events we're handling. */
|
2014-02-18 19:56:09 +00:00
|
|
|
if ((h->transMethod & (CFGTBL_Trans_io_accel1
|
|
|
|
|
| CFGTBL_Trans_io_accel2)) &&
|
2014-02-18 19:55:43 +00:00
|
|
|
(h->events & HPSA_EVENT_NOTIFY_ACCEL_IO_PATH_STATE_CHANGE ||
|
|
|
|
|
h->events & HPSA_EVENT_NOTIFY_ACCEL_IO_PATH_CONFIG_CHANGE)) {
|
|
|
|
|
|
|
|
|
|
if (h->events & HPSA_EVENT_NOTIFY_ACCEL_IO_PATH_STATE_CHANGE)
|
|
|
|
|
event_type = "state change";
|
|
|
|
|
if (h->events & HPSA_EVENT_NOTIFY_ACCEL_IO_PATH_CONFIG_CHANGE)
|
|
|
|
|
event_type = "configuration change";
|
|
|
|
|
/* Stop sending new RAID offload reqs via the IO accelerator */
|
|
|
|
|
scsi_block_requests(h->scsi_host);
|
|
|
|
|
for (i = 0; i < h->ndevices; i++)
|
|
|
|
|
h->dev[i]->offload_enabled = 0;
|
2014-02-18 19:57:37 +00:00
|
|
|
hpsa_drain_accel_commands(h);
|
2014-02-18 19:55:43 +00:00
|
|
|
/* Set 'accelerator path config change' bit */
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"Acknowledging event: 0x%08x (HP SSD Smart Path %s)\n",
|
|
|
|
|
h->events, event_type);
|
|
|
|
|
writel(h->events, &(h->cfgtable->clear_event_notify));
|
|
|
|
|
/* Set the "clear event notify field update" bit 6 */
|
|
|
|
|
writel(DOORBELL_CLEAR_EVENTS, h->vaddr + SA5_DOORBELL);
|
|
|
|
|
/* Wait until ctlr clears 'clear event notify field', bit 6 */
|
|
|
|
|
hpsa_wait_for_clear_event_notify_ack(h);
|
|
|
|
|
scsi_unblock_requests(h->scsi_host);
|
|
|
|
|
} else {
|
|
|
|
|
/* Acknowledge controller notification events. */
|
|
|
|
|
writel(h->events, &(h->cfgtable->clear_event_notify));
|
|
|
|
|
writel(DOORBELL_CLEAR_EVENTS, h->vaddr + SA5_DOORBELL);
|
|
|
|
|
hpsa_wait_for_clear_event_notify_ack(h);
|
|
|
|
|
#if 0
|
|
|
|
|
writel(CFGTBL_ChangeReq, h->vaddr + SA5_DOORBELL);
|
|
|
|
|
hpsa_wait_for_mode_change_ack(h);
|
|
|
|
|
#endif
|
|
|
|
|
}
|
2014-02-21 22:25:00 +00:00
|
|
|
return;
|
2014-02-18 19:55:43 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Check a register on the controller to see if there are configuration
|
|
|
|
|
* changes (added/changed/removed logical drives, etc.) which mean that
|
2014-02-18 19:57:05 +00:00
|
|
|
* we should rescan the controller for devices.
|
|
|
|
|
* Also check flag for driver-initiated rescan.
|
2014-02-18 19:55:43 +00:00
|
|
|
*/
|
2014-02-21 22:25:00 +00:00
|
|
|
static int hpsa_ctlr_needs_rescan(struct ctlr_info *h)
|
2014-02-18 19:55:43 +00:00
|
|
|
{
|
2015-11-04 21:50:37 +00:00
|
|
|
if (h->drv_req_rescan) {
|
|
|
|
|
h->drv_req_rescan = 0;
|
|
|
|
|
return 1;
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:55:43 +00:00
|
|
|
if (!(h->fw_support & MISC_FW_EVENT_NOTIFY))
|
2014-02-21 22:25:00 +00:00
|
|
|
return 0;
|
2014-02-18 19:55:43 +00:00
|
|
|
|
|
|
|
|
h->events = readl(&(h->cfgtable->event_notify));
|
2014-02-21 22:25:00 +00:00
|
|
|
return h->events & RESCAN_REQUIRED_EVENT_BITS;
|
|
|
|
|
}
|
2014-02-18 19:55:43 +00:00
|
|
|
|
2014-02-21 22:25:00 +00:00
|
|
|
/*
|
|
|
|
|
* Check if any of the offline devices have become ready
|
|
|
|
|
*/
|
|
|
|
|
static int hpsa_offline_devices_ready(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
struct offline_device_entry *d;
|
|
|
|
|
struct list_head *this, *tmp;
|
|
|
|
|
|
|
|
|
|
spin_lock_irqsave(&h->offline_device_lock, flags);
|
|
|
|
|
list_for_each_safe(this, tmp, &h->offline_device_list) {
|
|
|
|
|
d = list_entry(this, struct offline_device_entry,
|
|
|
|
|
offline_list);
|
|
|
|
|
spin_unlock_irqrestore(&h->offline_device_lock, flags);
|
2014-07-03 15:17:58 +00:00
|
|
|
if (!hpsa_volume_offline(h, d->scsi3addr)) {
|
|
|
|
|
spin_lock_irqsave(&h->offline_device_lock, flags);
|
|
|
|
|
list_del(&d->offline_list);
|
|
|
|
|
spin_unlock_irqrestore(&h->offline_device_lock, flags);
|
2014-02-21 22:25:00 +00:00
|
|
|
return 1;
|
2014-07-03 15:17:58 +00:00
|
|
|
}
|
2014-02-21 22:25:00 +00:00
|
|
|
spin_lock_irqsave(&h->offline_device_lock, flags);
|
|
|
|
|
}
|
|
|
|
|
spin_unlock_irqrestore(&h->offline_device_lock, flags);
|
|
|
|
|
return 0;
|
2014-02-18 19:55:43 +00:00
|
|
|
}
|
|
|
|
|
|
2015-11-04 21:52:09 +00:00
|
|
|
static int hpsa_luns_changed(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
int rc = 1; /* assume there are changes */
|
|
|
|
|
struct ReportLUNdata *logdev = NULL;
|
|
|
|
|
|
|
|
|
|
/* if we can't find out if lun data has changed,
|
|
|
|
|
* assume that it has.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
if (!h->lastlogicals)
|
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
|
|
logdev = kzalloc(sizeof(*logdev), GFP_KERNEL);
|
|
|
|
|
if (!logdev) {
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"Out of memory, can't track lun changes.\n");
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
|
|
|
|
if (hpsa_scsi_do_report_luns(h, 1, logdev, sizeof(*logdev), 0)) {
|
|
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"report luns failed, can't track lun changes.\n");
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
|
|
|
|
if (memcmp(logdev, h->lastlogicals, sizeof(*logdev))) {
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"Lun changes detected.\n");
|
|
|
|
|
memcpy(h->lastlogicals, logdev, sizeof(*logdev));
|
|
|
|
|
goto out;
|
|
|
|
|
} else
|
|
|
|
|
rc = 0; /* no changes detected. */
|
|
|
|
|
out:
|
|
|
|
|
kfree(logdev);
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:45:17 +00:00
|
|
|
static void hpsa_rescan_ctlr_worker(struct work_struct *work)
|
2011-10-26 21:22:04 +00:00
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
2013-12-04 23:10:07 +00:00
|
|
|
struct ctlr_info *h = container_of(to_delayed_work(work),
|
2015-01-23 22:45:17 +00:00
|
|
|
struct ctlr_info, rescan_ctlr_work);
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
if (h->remove_in_progress)
|
2013-12-04 23:10:07 +00:00
|
|
|
return;
|
2014-02-21 22:25:00 +00:00
|
|
|
|
|
|
|
|
if (hpsa_ctlr_needs_rescan(h) || hpsa_offline_devices_ready(h)) {
|
|
|
|
|
scsi_host_get(h->scsi_host);
|
|
|
|
|
hpsa_ack_ctlr_events(h);
|
|
|
|
|
hpsa_scan_start(h->scsi_host);
|
|
|
|
|
scsi_host_put(h->scsi_host);
|
2015-11-04 21:52:09 +00:00
|
|
|
} else if (h->discovery_polling) {
|
2015-11-04 21:52:16 +00:00
|
|
|
hpsa_disable_rld_caching(h);
|
2015-11-04 21:52:09 +00:00
|
|
|
if (hpsa_luns_changed(h)) {
|
|
|
|
|
struct Scsi_Host *sh = NULL;
|
|
|
|
|
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"driver discovery polling rescan.\n");
|
|
|
|
|
sh = scsi_host_get(h->scsi_host);
|
|
|
|
|
if (sh != NULL) {
|
|
|
|
|
hpsa_scan_start(sh);
|
|
|
|
|
scsi_host_put(sh);
|
|
|
|
|
}
|
|
|
|
|
}
|
2014-02-21 22:25:00 +00:00
|
|
|
}
|
2013-12-04 23:10:07 +00:00
|
|
|
spin_lock_irqsave(&h->lock, flags);
|
2015-01-23 22:45:17 +00:00
|
|
|
if (!h->remove_in_progress)
|
|
|
|
|
queue_delayed_work(h->rescan_ctlr_wq, &h->rescan_ctlr_work,
|
|
|
|
|
h->heartbeat_sample_interval);
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_monitor_ctlr_worker(struct work_struct *work)
|
|
|
|
|
{
|
|
|
|
|
unsigned long flags;
|
|
|
|
|
struct ctlr_info *h = container_of(to_delayed_work(work),
|
|
|
|
|
struct ctlr_info, monitor_ctlr_work);
|
|
|
|
|
|
|
|
|
|
detect_controller_lockup(h);
|
|
|
|
|
if (lockup_detected(h))
|
2011-10-26 21:22:04 +00:00
|
|
|
return;
|
2015-01-23 22:45:17 +00:00
|
|
|
|
|
|
|
|
spin_lock_irqsave(&h->lock, flags);
|
|
|
|
|
if (!h->remove_in_progress)
|
|
|
|
|
schedule_delayed_work(&h->monitor_ctlr_work,
|
2013-12-04 23:10:07 +00:00
|
|
|
h->heartbeat_sample_interval);
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
2011-10-26 21:22:04 +00:00
|
|
|
}
|
|
|
|
|
|
2015-01-23 22:45:17 +00:00
|
|
|
static struct workqueue_struct *hpsa_create_controller_wq(struct ctlr_info *h,
|
|
|
|
|
char *name)
|
|
|
|
|
{
|
|
|
|
|
struct workqueue_struct *wq = NULL;
|
|
|
|
|
|
2015-02-06 23:44:15 +00:00
|
|
|
wq = alloc_ordered_workqueue("%s_%d_hpsa", 0, name, h->ctlr);
|
2015-01-23 22:45:17 +00:00
|
|
|
if (!wq)
|
|
|
|
|
dev_err(&h->pdev->dev, "failed to create %s workqueue\n", name);
|
|
|
|
|
|
|
|
|
|
return wq;
|
|
|
|
|
}
|
|
|
|
|
|
2012-12-21 21:08:55 +00:00
|
|
|
static int hpsa_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
2010-06-16 18:51:35 +00:00
|
|
|
int dac, rc;
|
2009-12-08 22:09:11 +00:00
|
|
|
struct ctlr_info *h;
|
2011-05-03 19:59:51 +00:00
|
|
|
int try_soft_reset = 0;
|
|
|
|
|
unsigned long flags;
|
2015-04-02 13:25:54 +00:00
|
|
|
u32 board_id;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
if (number_of_controllers == 0)
|
|
|
|
|
printk(KERN_INFO DRIVER_NAME "\n");
|
|
|
|
|
|
2015-04-02 13:25:54 +00:00
|
|
|
rc = hpsa_lookup_board_id(pdev, &board_id);
|
|
|
|
|
if (rc < 0) {
|
|
|
|
|
dev_warn(&pdev->dev, "Board ID not found\n");
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
rc = hpsa_init_reset_devices(pdev, board_id);
|
2011-05-03 19:59:51 +00:00
|
|
|
if (rc) {
|
|
|
|
|
if (rc != -ENOTSUPP)
|
|
|
|
|
return rc;
|
|
|
|
|
/* If the reset fails in a particular way (it has no way to do
|
|
|
|
|
* a proper hard reset, so returns -ENOTSUPP) we can try to do
|
|
|
|
|
* a soft reset once we get the controller configured up to the
|
|
|
|
|
* point that it can accept a command.
|
|
|
|
|
*/
|
|
|
|
|
try_soft_reset = 1;
|
|
|
|
|
rc = 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
reinit_after_soft_reset:
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2010-02-04 14:42:40 +00:00
|
|
|
/* Command structures must be aligned on a 32-byte boundary because
|
|
|
|
|
* the 5 lower bits of the address are used by the hardware. and by
|
|
|
|
|
* the driver. See comments in hpsa.h for more info.
|
|
|
|
|
*/
|
|
|
|
|
BUILD_BUG_ON(sizeof(struct CommandList) % COMMANDLIST_ALIGNMENT);
|
2009-12-08 22:09:11 +00:00
|
|
|
h = kzalloc(sizeof(*h), GFP_KERNEL);
|
2015-04-23 14:33:48 +00:00
|
|
|
if (!h) {
|
|
|
|
|
dev_err(&pdev->dev, "Failed to allocate controller head\n");
|
2010-02-04 14:41:59 +00:00
|
|
|
return -ENOMEM;
|
2015-04-23 14:33:48 +00:00
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2010-05-27 20:12:46 +00:00
|
|
|
h->pdev = pdev;
|
2015-04-23 14:33:48 +00:00
|
|
|
|
2011-02-15 21:32:53 +00:00
|
|
|
h->intr_mode = hpsa_simple_mode ? SIMPLE_MODE_INT : PERF_MODE_INT;
|
2014-02-21 22:25:00 +00:00
|
|
|
INIT_LIST_HEAD(&h->offline_device_list);
|
2011-01-06 20:48:24 +00:00
|
|
|
spin_lock_init(&h->lock);
|
2014-02-21 22:25:00 +00:00
|
|
|
spin_lock_init(&h->offline_device_lock);
|
2011-01-06 20:48:24 +00:00
|
|
|
spin_lock_init(&h->scan_lock);
|
2015-01-23 22:43:46 +00:00
|
|
|
atomic_set(&h->passthru_cmds_avail, HPSA_MAX_CONCURRENT_PASSTHRUS);
|
2015-04-23 14:32:06 +00:00
|
|
|
atomic_set(&h->abort_cmds_available, HPSA_CMDS_RESERVED_FOR_ABORTS);
|
2014-05-29 15:53:18 +00:00
|
|
|
|
|
|
|
|
/* Allocate and clear per-cpu variable lockup_detected */
|
|
|
|
|
h->lockup_detected = alloc_percpu(u32);
|
2014-07-03 15:18:08 +00:00
|
|
|
if (!h->lockup_detected) {
|
2015-04-23 14:33:48 +00:00
|
|
|
dev_err(&h->pdev->dev, "Failed to allocate lockup detector\n");
|
2014-07-03 15:18:08 +00:00
|
|
|
rc = -ENOMEM;
|
2015-04-23 14:34:53 +00:00
|
|
|
goto clean1; /* aer/h */
|
2014-07-03 15:18:08 +00:00
|
|
|
}
|
2014-05-29 15:53:18 +00:00
|
|
|
set_lockup_detected_for_all_cpus(h, 0);
|
|
|
|
|
|
2010-05-27 20:12:46 +00:00
|
|
|
rc = hpsa_pci_init(h);
|
2015-04-23 14:33:48 +00:00
|
|
|
if (rc)
|
2015-04-23 14:35:09 +00:00
|
|
|
goto clean2; /* lu, aer/h */
|
|
|
|
|
|
|
|
|
|
/* relies on h-> settings made by hpsa_pci_init, including
|
|
|
|
|
* interrupt_mode h->intr */
|
|
|
|
|
rc = hpsa_scsi_host_alloc(h);
|
|
|
|
|
if (rc)
|
|
|
|
|
goto clean2_5; /* pci, lu, aer/h */
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-04-23 14:35:09 +00:00
|
|
|
sprintf(h->devname, HPSA "%d", h->scsi_host->host_no);
|
2009-12-08 22:09:11 +00:00
|
|
|
h->ctlr = number_of_controllers;
|
|
|
|
|
number_of_controllers++;
|
|
|
|
|
|
|
|
|
|
/* configure PCI DMA stuff */
|
2010-02-04 14:41:59 +00:00
|
|
|
rc = pci_set_dma_mask(pdev, DMA_BIT_MASK(64));
|
|
|
|
|
if (rc == 0) {
|
2009-12-08 22:09:11 +00:00
|
|
|
dac = 1;
|
2010-02-04 14:41:59 +00:00
|
|
|
} else {
|
|
|
|
|
rc = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
|
|
|
|
|
if (rc == 0) {
|
|
|
|
|
dac = 0;
|
|
|
|
|
} else {
|
|
|
|
|
dev_err(&pdev->dev, "no suitable DMA available\n");
|
2015-04-23 14:35:09 +00:00
|
|
|
goto clean3; /* shost, pci, lu, aer/h */
|
2010-02-04 14:41:59 +00:00
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* make sure the board interrupts are off */
|
|
|
|
|
h->access.set_intr_mask(h, HPSA_INTR_OFF);
|
2010-06-16 18:51:50 +00:00
|
|
|
|
2015-04-23 14:33:48 +00:00
|
|
|
rc = hpsa_request_irqs(h, do_hpsa_intr_msi, do_hpsa_intr_intx);
|
|
|
|
|
if (rc)
|
2015-04-23 14:35:09 +00:00
|
|
|
goto clean3; /* shost, pci, lu, aer/h */
|
2015-04-23 14:32:27 +00:00
|
|
|
rc = hpsa_alloc_cmd_pool(h);
|
2015-01-23 22:42:54 +00:00
|
|
|
if (rc)
|
2015-04-23 14:35:09 +00:00
|
|
|
goto clean4; /* irq, shost, pci, lu, aer/h */
|
2015-04-23 14:33:48 +00:00
|
|
|
rc = hpsa_alloc_sg_chain_blocks(h);
|
|
|
|
|
if (rc)
|
2015-04-23 14:35:09 +00:00
|
|
|
goto clean5; /* cmd, irq, shost, pci, lu, aer/h */
|
2010-02-04 14:43:16 +00:00
|
|
|
init_waitqueue_head(&h->scan_wait_queue);
|
2015-04-23 14:32:06 +00:00
|
|
|
init_waitqueue_head(&h->abort_cmd_wait_queue);
|
2015-04-23 14:35:22 +00:00
|
|
|
init_waitqueue_head(&h->event_sync_wait_queue);
|
|
|
|
|
mutex_init(&h->reset_mutex);
|
2010-02-04 14:43:16 +00:00
|
|
|
h->scan_finished = 1; /* no scan currently in progress */
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
pci_set_drvdata(pdev, h);
|
2011-05-03 19:59:41 +00:00
|
|
|
h->ndevices = 0;
|
2015-04-23 14:35:09 +00:00
|
|
|
|
2011-05-03 19:59:41 +00:00
|
|
|
spin_lock_init(&h->devlock);
|
2015-04-23 14:33:48 +00:00
|
|
|
rc = hpsa_put_ctlr_into_performant_mode(h);
|
|
|
|
|
if (rc)
|
2015-04-23 14:35:09 +00:00
|
|
|
goto clean6; /* sg, cmd, irq, shost, pci, lu, aer/h */
|
|
|
|
|
|
2015-04-23 14:34:53 +00:00
|
|
|
/* create the resubmit workqueue */
|
|
|
|
|
h->rescan_ctlr_wq = hpsa_create_controller_wq(h, "rescan");
|
|
|
|
|
if (!h->rescan_ctlr_wq) {
|
|
|
|
|
rc = -ENOMEM;
|
|
|
|
|
goto clean7;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
h->resubmit_wq = hpsa_create_controller_wq(h, "resubmit");
|
|
|
|
|
if (!h->resubmit_wq) {
|
|
|
|
|
rc = -ENOMEM;
|
|
|
|
|
goto clean7; /* aer/h */
|
|
|
|
|
}
|
2011-05-03 19:59:51 +00:00
|
|
|
|
2015-04-23 14:33:48 +00:00
|
|
|
/*
|
|
|
|
|
* At this point, the controller is ready to take commands.
|
2011-05-03 19:59:51 +00:00
|
|
|
* Now, if reset_devices and the hard reset didn't work, try
|
|
|
|
|
* the soft reset and see if that works.
|
|
|
|
|
*/
|
|
|
|
|
if (try_soft_reset) {
|
|
|
|
|
|
|
|
|
|
/* This is kind of gross. We may or may not get a completion
|
|
|
|
|
* from the soft reset command, and if we do, then the value
|
|
|
|
|
* from the fifo may or may not be valid. So, we wait 10 secs
|
|
|
|
|
* after the reset throwing away any completions we get during
|
|
|
|
|
* that time. Unregister the interrupt handler and register
|
|
|
|
|
* fake ones to scoop up any residual completions.
|
|
|
|
|
*/
|
|
|
|
|
spin_lock_irqsave(&h->lock, flags);
|
|
|
|
|
h->access.set_intr_mask(h, HPSA_INTR_OFF);
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
2015-01-23 22:41:40 +00:00
|
|
|
hpsa_free_irqs(h);
|
2015-01-23 22:42:32 +00:00
|
|
|
rc = hpsa_request_irqs(h, hpsa_msix_discard_completions,
|
2011-05-03 19:59:51 +00:00
|
|
|
hpsa_intx_discard_completions);
|
|
|
|
|
if (rc) {
|
2015-01-23 22:42:32 +00:00
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"Failed to request_irq after soft reset.\n");
|
2015-04-23 14:34:37 +00:00
|
|
|
/*
|
2015-04-23 14:34:48 +00:00
|
|
|
* cannot goto clean7 or free_irqs will be called
|
|
|
|
|
* again. Instead, do its work
|
|
|
|
|
*/
|
|
|
|
|
hpsa_free_performant_mode(h); /* clean7 */
|
|
|
|
|
hpsa_free_sg_chain_blocks(h); /* clean6 */
|
|
|
|
|
hpsa_free_cmd_pool(h); /* clean5 */
|
|
|
|
|
/*
|
|
|
|
|
* skip hpsa_free_irqs(h) clean4 since that
|
|
|
|
|
* was just called before request_irqs failed
|
2015-04-23 14:34:37 +00:00
|
|
|
*/
|
|
|
|
|
goto clean3;
|
2011-05-03 19:59:51 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
rc = hpsa_kdump_soft_reset(h);
|
|
|
|
|
if (rc)
|
|
|
|
|
/* Neither hard nor soft reset worked, we're hosed. */
|
2015-07-18 16:12:33 +00:00
|
|
|
goto clean7;
|
2011-05-03 19:59:51 +00:00
|
|
|
|
|
|
|
|
dev_info(&h->pdev->dev, "Board READY.\n");
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"Waiting for stale completions to drain.\n");
|
|
|
|
|
h->access.set_intr_mask(h, HPSA_INTR_ON);
|
|
|
|
|
msleep(10000);
|
|
|
|
|
h->access.set_intr_mask(h, HPSA_INTR_OFF);
|
|
|
|
|
|
|
|
|
|
rc = controller_reset_failed(h->cfgtable);
|
|
|
|
|
if (rc)
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"Soft reset appears to have failed.\n");
|
|
|
|
|
|
|
|
|
|
/* since the controller's reset, we have to go back and re-init
|
|
|
|
|
* everything. Easiest to just forget what we've done and do it
|
|
|
|
|
* all over again.
|
|
|
|
|
*/
|
|
|
|
|
hpsa_undo_allocations_after_kdump_soft_reset(h);
|
|
|
|
|
try_soft_reset = 0;
|
|
|
|
|
if (rc)
|
2015-04-23 14:34:48 +00:00
|
|
|
/* don't goto clean, we already unallocated */
|
2011-05-03 19:59:51 +00:00
|
|
|
return -ENODEV;
|
|
|
|
|
|
|
|
|
|
goto reinit_after_soft_reset;
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-04-23 14:33:48 +00:00
|
|
|
/* Enable Accelerated IO path at driver layer */
|
|
|
|
|
h->acciopath_status = 1;
|
2015-11-04 21:52:09 +00:00
|
|
|
/* Disable discovery polling.*/
|
|
|
|
|
h->discovery_polling = 0;
|
2014-02-18 19:57:00 +00:00
|
|
|
|
2014-02-18 19:57:05 +00:00
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/* Turn the interrupts on so we can service requests */
|
|
|
|
|
h->access.set_intr_mask(h, HPSA_INTR_ON);
|
|
|
|
|
|
2010-02-04 14:42:50 +00:00
|
|
|
hpsa_hba_inquiry(h);
|
2013-12-04 23:10:07 +00:00
|
|
|
|
2015-11-04 21:52:09 +00:00
|
|
|
h->lastlogicals = kzalloc(sizeof(*(h->lastlogicals)), GFP_KERNEL);
|
|
|
|
|
if (!h->lastlogicals)
|
|
|
|
|
dev_info(&h->pdev->dev,
|
|
|
|
|
"Can't track change to report lun data\n");
|
|
|
|
|
|
2016-04-27 22:13:26 +00:00
|
|
|
/* hook into SCSI subsystem */
|
|
|
|
|
rc = hpsa_scsi_add_host(h);
|
|
|
|
|
if (rc)
|
|
|
|
|
goto clean7; /* perf, sg, cmd, irq, shost, pci, lu, aer/h */
|
|
|
|
|
|
2013-12-04 23:10:07 +00:00
|
|
|
/* Monitor the controller for firmware lockups */
|
|
|
|
|
h->heartbeat_sample_interval = HEARTBEAT_SAMPLE_INTERVAL;
|
|
|
|
|
INIT_DELAYED_WORK(&h->monitor_ctlr_work, hpsa_monitor_ctlr_worker);
|
|
|
|
|
schedule_delayed_work(&h->monitor_ctlr_work,
|
|
|
|
|
h->heartbeat_sample_interval);
|
2015-01-23 22:45:17 +00:00
|
|
|
INIT_DELAYED_WORK(&h->rescan_ctlr_work, hpsa_rescan_ctlr_worker);
|
|
|
|
|
queue_delayed_work(h->rescan_ctlr_wq, &h->rescan_ctlr_work,
|
|
|
|
|
h->heartbeat_sample_interval);
|
[SCSI] hpsa: return 0 from driver probe function on success, not 1
A return value of 1 is interpreted as an error. See pci_driver.
in local_pci_probe(). If you're wondering how this ever could
have worked, it's because it used to be the case that only return
values less than zero were interpreted as failure. But even in
the current kernel if the driver registers its various entry
points with the kernel, and then returns a value which is
interpreted as failure, those registrations aren't undone, so
the driver still mostly works. However, the driver's remove
function wouldn't be called on rmmod, and pci power management
functions wouldn't work. In the case of Smart Array, since it
has a battery backed cache (or else no cache) even if the driver
is not shut down properly as long as there is no outstanding
i/o, nothing too bad happens, which is why it took so long to
notice.
Requesting backport to stable because the change to pci-driver.c
which requires driver probe functions to return 0 occurred between
2.6.35 and 2.6.36 (the pci power management breakage) and again
between 3.7 and 3.8 (pci_dev->driver getting set to NULL in
local_pci_probe() preventing driver remove function from being
called on rmmod.)
Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com>
Cc: stable@vger.kernel.org
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-11-01 16:02:25 +00:00
|
|
|
return 0;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
2015-04-23 14:35:09 +00:00
|
|
|
clean7: /* perf, sg, cmd, irq, shost, pci, lu, aer/h */
|
2015-04-23 14:33:48 +00:00
|
|
|
hpsa_free_performant_mode(h);
|
|
|
|
|
h->access.set_intr_mask(h, HPSA_INTR_OFF);
|
|
|
|
|
clean6: /* sg, cmd, irq, pci, lockup, wq/aer/h */
|
2010-02-25 20:03:27 +00:00
|
|
|
hpsa_free_sg_chain_blocks(h);
|
2015-04-23 14:35:09 +00:00
|
|
|
clean5: /* cmd, irq, shost, pci, lu, aer/h */
|
2011-05-03 19:59:20 +00:00
|
|
|
hpsa_free_cmd_pool(h);
|
2015-04-23 14:35:09 +00:00
|
|
|
clean4: /* irq, shost, pci, lu, aer/h */
|
2015-01-23 22:41:40 +00:00
|
|
|
hpsa_free_irqs(h);
|
2015-04-23 14:35:09 +00:00
|
|
|
clean3: /* shost, pci, lu, aer/h */
|
|
|
|
|
scsi_host_put(h->scsi_host);
|
|
|
|
|
h->scsi_host = NULL;
|
|
|
|
|
clean2_5: /* pci, lu, aer/h */
|
2015-04-23 14:33:17 +00:00
|
|
|
hpsa_free_pci_init(h);
|
2015-04-23 14:35:09 +00:00
|
|
|
clean2: /* lu, aer/h */
|
2015-04-23 14:33:48 +00:00
|
|
|
if (h->lockup_detected) {
|
|
|
|
|
free_percpu(h->lockup_detected);
|
|
|
|
|
h->lockup_detected = NULL;
|
|
|
|
|
}
|
|
|
|
|
clean1: /* wq/aer/h */
|
|
|
|
|
if (h->resubmit_wq) {
|
2015-01-23 22:43:25 +00:00
|
|
|
destroy_workqueue(h->resubmit_wq);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->resubmit_wq = NULL;
|
|
|
|
|
}
|
|
|
|
|
if (h->rescan_ctlr_wq) {
|
2015-01-23 22:45:17 +00:00
|
|
|
destroy_workqueue(h->rescan_ctlr_wq);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->rescan_ctlr_wq = NULL;
|
|
|
|
|
}
|
2009-12-08 22:09:11 +00:00
|
|
|
kfree(h);
|
2010-02-04 14:41:59 +00:00
|
|
|
return rc;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_flush_cache(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
char *flush_buf;
|
|
|
|
|
struct CommandList *c;
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
int rc;
|
2013-09-23 18:33:30 +00:00
|
|
|
|
2014-05-29 15:53:18 +00:00
|
|
|
if (unlikely(lockup_detected(h)))
|
2013-09-23 18:33:30 +00:00
|
|
|
return;
|
2009-12-08 22:09:11 +00:00
|
|
|
flush_buf = kzalloc(4, GFP_KERNEL);
|
|
|
|
|
if (!flush_buf)
|
|
|
|
|
return;
|
|
|
|
|
|
2015-01-23 22:43:04 +00:00
|
|
|
c = cmd_alloc(h);
|
2015-04-23 14:33:38 +00:00
|
|
|
|
2013-02-20 17:24:41 +00:00
|
|
|
if (fill_cmd(c, HPSA_CACHE_FLUSH, h, flush_buf, 4, 0,
|
|
|
|
|
RAID_CTLR_LUNID, TYPE_CMD)) {
|
|
|
|
|
goto out;
|
|
|
|
|
}
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
rc = hpsa_scsi_do_simple_cmd_with_retry(h, c,
|
2016-04-27 22:13:51 +00:00
|
|
|
PCI_DMA_TODEVICE, DEFAULT_TIMEOUT);
|
hpsa: rework controller command submission
Allow driver initiated commands to have a timeout. It does not
yet try to do anything with timeouts on such commands.
We are sending a reset in order to get rid of a command we want to abort.
If we make it return on the same reply queue as the command we want to abort,
the completion of the aborted command will not race with the completion of
the reset command.
Rename hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd(), since
this function is the interface for issuing commands to the controller and
not the "core" of that implementation. Add a parameter to it which allows
the caller to specify the reply queue to be used. Modify existing callers
to specify the default reply queue.
Rename __hpsa_scsi_do_simple_cmd_core() to hpsa_scsi_do_simple_cmd_core(),
since this routine is the "core" implementation of the "do simple command"
function and there is no longer any other function with a similar name.
Modify the existing callers of this routine (other than
hpsa_scsi_do_simple_cmd()) to instead call hpsa_scsi_do_simple_cmd(), since
it will now accept the reply_queue paramenter, and it provides a controller
lock-up check. (Also, tweak two related message strings to make them
distinct from each other.)
Submitting a command to a locked up controller always results in a timeout,
so check for controller lock-up before submitting.
This is to enable fixing a race between command completions and
abort completions on different reply queues in a subsequent patch.
We want to be able to specify which reply queue an abort completion
should occur on so that it cannot race the completion of the command
it is trying to abort.
The following race was possible in theory:
1. Abort command is sent to hardware.
2. Command to be aborted simultaneously completes on another
reply queue.
3. Hardware receives abort command, decides command has already
completed and indicates this to the driver via another different
reply queue.
4. driver processes abort completion finds that the hardware does not know
about the command, concludes that therefore the command cannot complete,
returns SUCCESS indicating to the mid-layer that the scsi_cmnd may be
re-used.
5. Command from step 2 is processed and completed back to scsi mid
layer (after we already promised that would never happen.)
Fix by forcing aborts to complete on the same reply queue as the command
they are aborting.
Piggybacking device rescanning functionality onto the lockup
detection thread is not a good idea because if the controller
locks up during device rescanning, then the thread could get
stuck, then the lockup isn't detected. Use separate work
queues for device rescanning and lockup detection.
Detect controller lockup in abort handler.
After a lockup is detected, return DO_NO_CONNECT which results in immediate
termination of commands rather than DID_ERR which results in retries.
Modify detect_controller_lockup() to return the result, to remove the need for
a separate check.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
2015-04-23 14:32:00 +00:00
|
|
|
if (rc)
|
|
|
|
|
goto out;
|
2009-12-08 22:09:11 +00:00
|
|
|
if (c->err_info->CommandStatus != 0)
|
2013-02-20 17:24:41 +00:00
|
|
|
out:
|
2009-12-08 22:09:11 +00:00
|
|
|
dev_warn(&h->pdev->dev,
|
|
|
|
|
"error flushing cache on controller\n");
|
2015-01-23 22:43:04 +00:00
|
|
|
cmd_free(h, c);
|
2009-12-08 22:09:11 +00:00
|
|
|
kfree(flush_buf);
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-04 21:52:16 +00:00
|
|
|
/* Make controller gather fresh report lun data each time we
|
|
|
|
|
* send down a report luns request
|
|
|
|
|
*/
|
|
|
|
|
static void hpsa_disable_rld_caching(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
u32 *options;
|
|
|
|
|
struct CommandList *c;
|
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
|
|
/* Don't bother trying to set diag options if locked up */
|
|
|
|
|
if (unlikely(h->lockup_detected))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
options = kzalloc(sizeof(*options), GFP_KERNEL);
|
|
|
|
|
if (!options) {
|
|
|
|
|
dev_err(&h->pdev->dev,
|
|
|
|
|
"Error: failed to disable rld caching, during alloc.\n");
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
c = cmd_alloc(h);
|
|
|
|
|
|
|
|
|
|
/* first, get the current diag options settings */
|
|
|
|
|
if (fill_cmd(c, BMIC_SENSE_DIAG_OPTIONS, h, options, 4, 0,
|
|
|
|
|
RAID_CTLR_LUNID, TYPE_CMD))
|
|
|
|
|
goto errout;
|
|
|
|
|
|
|
|
|
|
rc = hpsa_scsi_do_simple_cmd_with_retry(h, c,
|
2016-04-27 22:13:51 +00:00
|
|
|
PCI_DMA_FROMDEVICE, DEFAULT_TIMEOUT);
|
2015-11-04 21:52:16 +00:00
|
|
|
if ((rc != 0) || (c->err_info->CommandStatus != 0))
|
|
|
|
|
goto errout;
|
|
|
|
|
|
|
|
|
|
/* Now, set the bit for disabling the RLD caching */
|
|
|
|
|
*options |= HPSA_DIAG_OPTS_DISABLE_RLD_CACHING;
|
|
|
|
|
|
|
|
|
|
if (fill_cmd(c, BMIC_SET_DIAG_OPTIONS, h, options, 4, 0,
|
|
|
|
|
RAID_CTLR_LUNID, TYPE_CMD))
|
|
|
|
|
goto errout;
|
|
|
|
|
|
|
|
|
|
rc = hpsa_scsi_do_simple_cmd_with_retry(h, c,
|
2016-04-27 22:13:51 +00:00
|
|
|
PCI_DMA_TODEVICE, DEFAULT_TIMEOUT);
|
2015-11-04 21:52:16 +00:00
|
|
|
if ((rc != 0) || (c->err_info->CommandStatus != 0))
|
|
|
|
|
goto errout;
|
|
|
|
|
|
|
|
|
|
/* Now verify that it got set: */
|
|
|
|
|
if (fill_cmd(c, BMIC_SENSE_DIAG_OPTIONS, h, options, 4, 0,
|
|
|
|
|
RAID_CTLR_LUNID, TYPE_CMD))
|
|
|
|
|
goto errout;
|
|
|
|
|
|
|
|
|
|
rc = hpsa_scsi_do_simple_cmd_with_retry(h, c,
|
2016-04-27 22:13:51 +00:00
|
|
|
PCI_DMA_FROMDEVICE, DEFAULT_TIMEOUT);
|
2015-11-04 21:52:16 +00:00
|
|
|
if ((rc != 0) || (c->err_info->CommandStatus != 0))
|
|
|
|
|
goto errout;
|
|
|
|
|
|
2015-11-12 09:43:38 +00:00
|
|
|
if (*options & HPSA_DIAG_OPTS_DISABLE_RLD_CACHING)
|
2015-11-04 21:52:16 +00:00
|
|
|
goto out;
|
|
|
|
|
|
|
|
|
|
errout:
|
|
|
|
|
dev_err(&h->pdev->dev,
|
|
|
|
|
"Error: failed to disable report lun data caching.\n");
|
|
|
|
|
out:
|
|
|
|
|
cmd_free(h, c);
|
|
|
|
|
kfree(options);
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
static void hpsa_shutdown(struct pci_dev *pdev)
|
|
|
|
|
{
|
|
|
|
|
struct ctlr_info *h;
|
|
|
|
|
|
|
|
|
|
h = pci_get_drvdata(pdev);
|
|
|
|
|
/* Turn board interrupts off and send the flush cache command
|
|
|
|
|
* sendcmd will turn off interrupt, and send the flush...
|
|
|
|
|
* To write all data in the battery backed cache to disks
|
|
|
|
|
*/
|
|
|
|
|
hpsa_flush_cache(h);
|
|
|
|
|
h->access.set_intr_mask(h, HPSA_INTR_OFF);
|
2015-04-23 14:33:48 +00:00
|
|
|
hpsa_free_irqs(h); /* init_one 4 */
|
2015-04-23 14:33:12 +00:00
|
|
|
hpsa_disable_interrupt_mode(h); /* pci_init 2 */
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
2012-12-21 21:08:55 +00:00
|
|
|
static void hpsa_free_device_info(struct ctlr_info *h)
|
2012-01-19 20:00:42 +00:00
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
|
2015-04-23 14:33:48 +00:00
|
|
|
for (i = 0; i < h->ndevices; i++) {
|
2012-01-19 20:00:42 +00:00
|
|
|
kfree(h->dev[i]);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->dev[i] = NULL;
|
|
|
|
|
}
|
2012-01-19 20:00:42 +00:00
|
|
|
}
|
|
|
|
|
|
2012-12-21 21:08:55 +00:00
|
|
|
static void hpsa_remove_one(struct pci_dev *pdev)
|
2009-12-08 22:09:11 +00:00
|
|
|
{
|
|
|
|
|
struct ctlr_info *h;
|
2013-12-04 23:10:07 +00:00
|
|
|
unsigned long flags;
|
2009-12-08 22:09:11 +00:00
|
|
|
|
|
|
|
|
if (pci_get_drvdata(pdev) == NULL) {
|
2011-10-26 21:22:04 +00:00
|
|
|
dev_err(&pdev->dev, "unable to remove device\n");
|
2009-12-08 22:09:11 +00:00
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
h = pci_get_drvdata(pdev);
|
2013-12-04 23:10:07 +00:00
|
|
|
|
|
|
|
|
/* Get rid of any controller monitoring work items */
|
|
|
|
|
spin_lock_irqsave(&h->lock, flags);
|
|
|
|
|
h->remove_in_progress = 1;
|
|
|
|
|
spin_unlock_irqrestore(&h->lock, flags);
|
2015-01-23 22:45:17 +00:00
|
|
|
cancel_delayed_work_sync(&h->monitor_ctlr_work);
|
|
|
|
|
cancel_delayed_work_sync(&h->rescan_ctlr_work);
|
|
|
|
|
destroy_workqueue(h->rescan_ctlr_wq);
|
|
|
|
|
destroy_workqueue(h->resubmit_wq);
|
2015-04-23 14:33:12 +00:00
|
|
|
|
2015-07-18 16:13:15 +00:00
|
|
|
/*
|
|
|
|
|
* Call before disabling interrupts.
|
|
|
|
|
* scsi_remove_host can trigger I/O operations especially
|
|
|
|
|
* when multipath is enabled. There can be SYNCHRONIZE CACHE
|
|
|
|
|
* operations which cannot complete and will hang the system.
|
|
|
|
|
*/
|
|
|
|
|
if (h->scsi_host)
|
|
|
|
|
scsi_remove_host(h->scsi_host); /* init_one 8 */
|
2015-04-23 14:33:48 +00:00
|
|
|
/* includes hpsa_free_irqs - init_one 4 */
|
2015-04-23 14:33:17 +00:00
|
|
|
/* includes hpsa_disable_interrupt_mode - pci_init 2 */
|
2009-12-08 22:09:11 +00:00
|
|
|
hpsa_shutdown(pdev);
|
2015-04-23 14:33:12 +00:00
|
|
|
|
2015-04-23 14:33:48 +00:00
|
|
|
hpsa_free_device_info(h); /* scan */
|
|
|
|
|
|
2015-04-23 14:35:09 +00:00
|
|
|
kfree(h->hba_inquiry_data); /* init_one 10 */
|
|
|
|
|
h->hba_inquiry_data = NULL; /* init_one 10 */
|
|
|
|
|
hpsa_free_ioaccel2_sg_chain_blocks(h);
|
2015-04-23 14:33:48 +00:00
|
|
|
hpsa_free_performant_mode(h); /* init_one 7 */
|
|
|
|
|
hpsa_free_sg_chain_blocks(h); /* init_one 6 */
|
|
|
|
|
hpsa_free_cmd_pool(h); /* init_one 5 */
|
2015-11-04 21:52:09 +00:00
|
|
|
kfree(h->lastlogicals);
|
2015-04-23 14:33:48 +00:00
|
|
|
|
|
|
|
|
/* hpsa_free_irqs already called via hpsa_shutdown init_one 4 */
|
2015-04-23 14:33:17 +00:00
|
|
|
|
2015-04-23 14:35:09 +00:00
|
|
|
scsi_host_put(h->scsi_host); /* init_one 3 */
|
|
|
|
|
h->scsi_host = NULL; /* init_one 3 */
|
|
|
|
|
|
2015-04-23 14:33:17 +00:00
|
|
|
/* includes hpsa_disable_interrupt_mode - pci_init 2 */
|
2015-04-23 14:35:09 +00:00
|
|
|
hpsa_free_pci_init(h); /* init_one 2.5 */
|
2015-04-23 14:33:17 +00:00
|
|
|
|
2015-04-23 14:33:48 +00:00
|
|
|
free_percpu(h->lockup_detected); /* init_one 2 */
|
|
|
|
|
h->lockup_detected = NULL; /* init_one 2 */
|
|
|
|
|
/* (void) pci_disable_pcie_error_reporting(pdev); */ /* init_one 1 */
|
2015-11-04 21:52:34 +00:00
|
|
|
|
|
|
|
|
hpsa_delete_sas_host(h);
|
|
|
|
|
|
2015-04-23 14:33:48 +00:00
|
|
|
kfree(h); /* init_one 1 */
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int hpsa_suspend(__attribute__((unused)) struct pci_dev *pdev,
|
|
|
|
|
__attribute__((unused)) pm_message_t state)
|
|
|
|
|
{
|
|
|
|
|
return -ENOSYS;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int hpsa_resume(__attribute__((unused)) struct pci_dev *pdev)
|
|
|
|
|
{
|
|
|
|
|
return -ENOSYS;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static struct pci_driver hpsa_pci_driver = {
|
2012-01-19 20:00:59 +00:00
|
|
|
.name = HPSA,
|
2009-12-08 22:09:11 +00:00
|
|
|
.probe = hpsa_init_one,
|
2012-12-21 21:08:55 +00:00
|
|
|
.remove = hpsa_remove_one,
|
2009-12-08 22:09:11 +00:00
|
|
|
.id_table = hpsa_pci_device_id, /* id_table */
|
|
|
|
|
.shutdown = hpsa_shutdown,
|
|
|
|
|
.suspend = hpsa_suspend,
|
|
|
|
|
.resume = hpsa_resume,
|
|
|
|
|
};
|
|
|
|
|
|
2010-02-04 14:42:40 +00:00
|
|
|
/* Fill in bucket_map[], given nsgs (the max number of
|
|
|
|
|
* scatter gather elements supported) and bucket[],
|
|
|
|
|
* which is an array of 8 integers. The bucket[] array
|
|
|
|
|
* contains 8 different DMA transfer sizes (in 16
|
|
|
|
|
* byte increments) which the controller uses to fetch
|
|
|
|
|
* commands. This function fills in bucket_map[], which
|
|
|
|
|
* maps a given number of scatter gather elements to one of
|
|
|
|
|
* the 8 DMA transfer sizes. The point of it is to allow the
|
|
|
|
|
* controller to only do as much DMA as needed to fetch the
|
|
|
|
|
* command, with the DMA transfer size encoded in the lower
|
|
|
|
|
* bits of the command address.
|
|
|
|
|
*/
|
|
|
|
|
static void calc_bucket_map(int bucket[], int num_buckets,
|
2015-01-23 22:41:09 +00:00
|
|
|
int nsgs, int min_blocks, u32 *bucket_map)
|
2010-02-04 14:42:40 +00:00
|
|
|
{
|
|
|
|
|
int i, j, b, size;
|
|
|
|
|
|
|
|
|
|
/* Note, bucket_map must have nsgs+1 entries. */
|
|
|
|
|
for (i = 0; i <= nsgs; i++) {
|
|
|
|
|
/* Compute size of a command with i SG entries */
|
2014-02-18 19:55:17 +00:00
|
|
|
size = i + min_blocks;
|
2010-02-04 14:42:40 +00:00
|
|
|
b = num_buckets; /* Assume the biggest bucket */
|
|
|
|
|
/* Find the bucket that is just big enough */
|
2014-02-18 19:55:17 +00:00
|
|
|
for (j = 0; j < num_buckets; j++) {
|
2010-02-04 14:42:40 +00:00
|
|
|
if (bucket[j] >= size) {
|
|
|
|
|
b = j;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
/* for a command with i SG entries, use bucket b. */
|
|
|
|
|
bucket_map[i] = b;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:33:48 +00:00
|
|
|
/*
|
|
|
|
|
* return -ENODEV on err, 0 on success (or no action)
|
|
|
|
|
* allocates numerous items that must be freed later
|
|
|
|
|
*/
|
2015-01-23 22:45:01 +00:00
|
|
|
static int hpsa_enter_performant_mode(struct ctlr_info *h, u32 trans_support)
|
2010-02-04 14:42:40 +00:00
|
|
|
{
|
2010-05-27 20:14:19 +00:00
|
|
|
int i;
|
|
|
|
|
unsigned long register_value;
|
2014-02-18 19:55:17 +00:00
|
|
|
unsigned long transMethod = CFGTBL_Trans_Performant |
|
|
|
|
|
(trans_support & CFGTBL_Trans_use_short_tags) |
|
2014-02-18 19:56:29 +00:00
|
|
|
CFGTBL_Trans_enable_directed_msix |
|
|
|
|
|
(trans_support & (CFGTBL_Trans_io_accel1 |
|
|
|
|
|
CFGTBL_Trans_io_accel2));
|
2014-02-18 19:55:17 +00:00
|
|
|
struct access_method access = SA5_performant_access;
|
2010-05-27 20:14:39 +00:00
|
|
|
|
|
|
|
|
/* This is a bit complicated. There are 8 registers on
|
|
|
|
|
* the controller which we write to to tell it 8 different
|
|
|
|
|
* sizes of commands which there may be. It's a way of
|
|
|
|
|
* reducing the DMA done to fetch each command. Encoded into
|
|
|
|
|
* each command's tag are 3 bits which communicate to the controller
|
|
|
|
|
* which of the eight sizes that command fits within. The size of
|
|
|
|
|
* each command depends on how many scatter gather entries there are.
|
|
|
|
|
* Each SG entry requires 16 bytes. The eight registers are programmed
|
|
|
|
|
* with the number of 16-byte blocks a command of that size requires.
|
|
|
|
|
* The smallest command possible requires 5 such 16 byte blocks.
|
2012-01-19 20:00:48 +00:00
|
|
|
* the largest command possible requires SG_ENTRIES_IN_CMD + 4 16-byte
|
2010-05-27 20:14:39 +00:00
|
|
|
* blocks. Note, this only extends to the SG entries contained
|
|
|
|
|
* within the command block, and does not extend to chained blocks
|
|
|
|
|
* of SG elements. bft[] contains the eight values we write to
|
|
|
|
|
* the registers. They are not evenly distributed, but have more
|
|
|
|
|
* sizes for small commands, and fewer sizes for larger commands.
|
|
|
|
|
*/
|
2012-01-19 20:00:48 +00:00
|
|
|
int bft[8] = {5, 6, 8, 10, 12, 20, 28, SG_ENTRIES_IN_CMD + 4};
|
2014-02-18 19:56:29 +00:00
|
|
|
#define MIN_IOACCEL2_BFT_ENTRY 5
|
|
|
|
|
#define HPSA_IOACCEL2_HEADER_SZ 4
|
|
|
|
|
int bft2[16] = {MIN_IOACCEL2_BFT_ENTRY, 6, 7, 8, 9, 10, 11, 12,
|
|
|
|
|
13, 14, 15, 16, 17, 18, 19,
|
|
|
|
|
HPSA_IOACCEL2_HEADER_SZ + IOACCEL2_MAXSGENTRIES};
|
|
|
|
|
BUILD_BUG_ON(ARRAY_SIZE(bft2) != 16);
|
|
|
|
|
BUILD_BUG_ON(ARRAY_SIZE(bft) != 8);
|
|
|
|
|
BUILD_BUG_ON(offsetof(struct io_accel2_cmd, sg) >
|
|
|
|
|
16 * MIN_IOACCEL2_BFT_ENTRY);
|
|
|
|
|
BUILD_BUG_ON(sizeof(struct ioaccel2_sg_element) != 16);
|
2012-01-19 20:00:48 +00:00
|
|
|
BUILD_BUG_ON(28 > SG_ENTRIES_IN_CMD + 4);
|
2010-02-04 14:42:40 +00:00
|
|
|
/* 5 = 1 s/g entry or 4k
|
|
|
|
|
* 6 = 2 s/g entry or 8k
|
|
|
|
|
* 8 = 4 s/g entry or 16k
|
|
|
|
|
* 10 = 6 s/g entry or 24k
|
|
|
|
|
*/
|
|
|
|
|
|
2014-05-29 15:53:23 +00:00
|
|
|
/* If the controller supports either ioaccel method then
|
|
|
|
|
* we can also use the RAID stack submit path that does not
|
|
|
|
|
* perform the superfluous readl() after each command submission.
|
|
|
|
|
*/
|
|
|
|
|
if (trans_support & (CFGTBL_Trans_io_accel1 | CFGTBL_Trans_io_accel2))
|
|
|
|
|
access = SA5_performant_access_no_read;
|
|
|
|
|
|
2010-02-04 14:42:40 +00:00
|
|
|
/* Controller spec: zero out this buffer. */
|
2014-05-29 15:53:07 +00:00
|
|
|
for (i = 0; i < h->nreply_queues; i++)
|
|
|
|
|
memset(h->reply_queue[i].head, 0, h->reply_queue_size);
|
2010-02-04 14:42:40 +00:00
|
|
|
|
2012-01-19 20:00:48 +00:00
|
|
|
bft[7] = SG_ENTRIES_IN_CMD + 4;
|
|
|
|
|
calc_bucket_map(bft, ARRAY_SIZE(bft),
|
2014-02-18 19:55:17 +00:00
|
|
|
SG_ENTRIES_IN_CMD, 4, h->blockFetchTable);
|
2010-02-04 14:42:40 +00:00
|
|
|
for (i = 0; i < 8; i++)
|
|
|
|
|
writel(bft[i], &h->transtable->BlockFetch[i]);
|
|
|
|
|
|
|
|
|
|
/* size of controller ring buffer */
|
|
|
|
|
writel(h->max_commands, &h->transtable->RepQSize);
|
2012-05-01 16:43:06 +00:00
|
|
|
writel(h->nreply_queues, &h->transtable->RepQCount);
|
2010-02-04 14:42:40 +00:00
|
|
|
writel(0, &h->transtable->RepQCtrAddrLow32);
|
|
|
|
|
writel(0, &h->transtable->RepQCtrAddrHigh32);
|
2012-05-01 16:43:06 +00:00
|
|
|
|
|
|
|
|
for (i = 0; i < h->nreply_queues; i++) {
|
|
|
|
|
writel(0, &h->transtable->RepQAddr[i].upper);
|
2014-05-29 15:53:07 +00:00
|
|
|
writel(h->reply_queue[i].busaddr,
|
2012-05-01 16:43:06 +00:00
|
|
|
&h->transtable->RepQAddr[i].lower);
|
|
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:56:29 +00:00
|
|
|
writel(0, &h->cfgtable->HostWrite.command_pool_addr_hi);
|
2014-02-18 19:55:17 +00:00
|
|
|
writel(transMethod, &(h->cfgtable->HostWrite.TransportRequest));
|
|
|
|
|
/*
|
|
|
|
|
* enable outbound interrupt coalescing in accelerator mode;
|
|
|
|
|
*/
|
|
|
|
|
if (trans_support & CFGTBL_Trans_io_accel1) {
|
|
|
|
|
access = SA5_ioaccel_mode1_access;
|
|
|
|
|
writel(10, &h->cfgtable->HostWrite.CoalIntDelay);
|
|
|
|
|
writel(4, &h->cfgtable->HostWrite.CoalIntCount);
|
2014-02-18 19:56:34 +00:00
|
|
|
} else {
|
|
|
|
|
if (trans_support & CFGTBL_Trans_io_accel2) {
|
|
|
|
|
access = SA5_ioaccel_mode2_access;
|
|
|
|
|
writel(10, &h->cfgtable->HostWrite.CoalIntDelay);
|
|
|
|
|
writel(4, &h->cfgtable->HostWrite.CoalIntCount);
|
|
|
|
|
}
|
2014-02-18 19:55:17 +00:00
|
|
|
}
|
2010-02-04 14:42:40 +00:00
|
|
|
writel(CFGTBL_ChangeReq, h->vaddr + SA5_DOORBELL);
|
2015-01-23 22:45:01 +00:00
|
|
|
if (hpsa_wait_for_mode_change_ack(h)) {
|
|
|
|
|
dev_err(&h->pdev->dev,
|
|
|
|
|
"performant mode problem - doorbell timeout\n");
|
|
|
|
|
return -ENODEV;
|
|
|
|
|
}
|
2010-02-04 14:42:40 +00:00
|
|
|
register_value = readl(&(h->cfgtable->TransportActive));
|
|
|
|
|
if (!(register_value & CFGTBL_Trans_Performant)) {
|
2015-01-23 22:42:22 +00:00
|
|
|
dev_err(&h->pdev->dev,
|
|
|
|
|
"performant mode problem - transport not active\n");
|
2015-01-23 22:45:01 +00:00
|
|
|
return -ENODEV;
|
2010-02-04 14:42:40 +00:00
|
|
|
}
|
2011-02-15 21:33:03 +00:00
|
|
|
/* Change the access methods to the performant access methods */
|
2014-02-18 19:55:17 +00:00
|
|
|
h->access = access;
|
|
|
|
|
h->transMethod = transMethod;
|
|
|
|
|
|
2014-02-18 19:56:29 +00:00
|
|
|
if (!((trans_support & CFGTBL_Trans_io_accel1) ||
|
|
|
|
|
(trans_support & CFGTBL_Trans_io_accel2)))
|
2015-01-23 22:45:01 +00:00
|
|
|
return 0;
|
2014-02-18 19:55:17 +00:00
|
|
|
|
2014-02-18 19:56:29 +00:00
|
|
|
if (trans_support & CFGTBL_Trans_io_accel1) {
|
|
|
|
|
/* Set up I/O accelerator mode */
|
|
|
|
|
for (i = 0; i < h->nreply_queues; i++) {
|
|
|
|
|
writel(i, h->vaddr + IOACCEL_MODE1_REPLY_QUEUE_INDEX);
|
|
|
|
|
h->reply_queue[i].current_entry =
|
|
|
|
|
readl(h->vaddr + IOACCEL_MODE1_PRODUCER_INDEX);
|
|
|
|
|
}
|
|
|
|
|
bft[7] = h->ioaccel_maxsg + 8;
|
|
|
|
|
calc_bucket_map(bft, ARRAY_SIZE(bft), h->ioaccel_maxsg, 8,
|
|
|
|
|
h->ioaccel1_blockFetchTable);
|
2014-02-18 19:55:17 +00:00
|
|
|
|
2014-02-18 19:56:29 +00:00
|
|
|
/* initialize all reply queue entries to unused */
|
2014-05-29 15:53:07 +00:00
|
|
|
for (i = 0; i < h->nreply_queues; i++)
|
|
|
|
|
memset(h->reply_queue[i].head,
|
|
|
|
|
(u8) IOACCEL_MODE1_REPLY_UNUSED,
|
|
|
|
|
h->reply_queue_size);
|
2014-02-18 19:55:17 +00:00
|
|
|
|
2014-02-18 19:56:29 +00:00
|
|
|
/* set all the constant fields in the accelerator command
|
|
|
|
|
* frames once at init time to save CPU cycles later.
|
|
|
|
|
*/
|
|
|
|
|
for (i = 0; i < h->nr_cmds; i++) {
|
|
|
|
|
struct io_accel1_cmd *cp = &h->ioaccel_cmd_pool[i];
|
|
|
|
|
|
|
|
|
|
cp->function = IOACCEL1_FUNCTION_SCSIIO;
|
|
|
|
|
cp->err_info = (u32) (h->errinfo_pool_dhandle +
|
|
|
|
|
(i * sizeof(struct ErrorInfo)));
|
|
|
|
|
cp->err_info_len = sizeof(struct ErrorInfo);
|
|
|
|
|
cp->sgl_offset = IOACCEL1_SGLOFFSET;
|
2015-01-23 22:41:09 +00:00
|
|
|
cp->host_context_flags =
|
|
|
|
|
cpu_to_le16(IOACCEL1_HCFLAGS_CISS_FORMAT);
|
2014-02-18 19:56:29 +00:00
|
|
|
cp->timeout_sec = 0;
|
|
|
|
|
cp->ReplyQueue = 0;
|
2014-11-14 23:26:59 +00:00
|
|
|
cp->tag =
|
2015-01-23 22:43:09 +00:00
|
|
|
cpu_to_le64((i << DIRECT_LOOKUP_SHIFT));
|
2014-11-14 23:26:59 +00:00
|
|
|
cp->host_addr =
|
|
|
|
|
cpu_to_le64(h->ioaccel_cmd_pool_dhandle +
|
2014-02-18 19:56:29 +00:00
|
|
|
(i * sizeof(struct io_accel1_cmd)));
|
|
|
|
|
}
|
|
|
|
|
} else if (trans_support & CFGTBL_Trans_io_accel2) {
|
|
|
|
|
u64 cfg_offset, cfg_base_addr_index;
|
|
|
|
|
u32 bft2_offset, cfg_base_addr;
|
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
|
|
rc = hpsa_find_cfg_addrs(h->pdev, h->vaddr, &cfg_base_addr,
|
|
|
|
|
&cfg_base_addr_index, &cfg_offset);
|
|
|
|
|
BUILD_BUG_ON(offsetof(struct io_accel2_cmd, sg) != 64);
|
|
|
|
|
bft2[15] = h->ioaccel_maxsg + HPSA_IOACCEL2_HEADER_SZ;
|
|
|
|
|
calc_bucket_map(bft2, ARRAY_SIZE(bft2), h->ioaccel_maxsg,
|
|
|
|
|
4, h->ioaccel2_blockFetchTable);
|
|
|
|
|
bft2_offset = readl(&h->cfgtable->io_accel_request_size_offset);
|
|
|
|
|
BUILD_BUG_ON(offsetof(struct CfgTable,
|
|
|
|
|
io_accel_request_size_offset) != 0xb8);
|
|
|
|
|
h->ioaccel2_bft2_regs =
|
|
|
|
|
remap_pci_mem(pci_resource_start(h->pdev,
|
|
|
|
|
cfg_base_addr_index) +
|
|
|
|
|
cfg_offset + bft2_offset,
|
|
|
|
|
ARRAY_SIZE(bft2) *
|
|
|
|
|
sizeof(*h->ioaccel2_bft2_regs));
|
|
|
|
|
for (i = 0; i < ARRAY_SIZE(bft2); i++)
|
|
|
|
|
writel(bft2[i], &h->ioaccel2_bft2_regs[i]);
|
2014-02-18 19:55:17 +00:00
|
|
|
}
|
2014-02-18 19:56:29 +00:00
|
|
|
writel(CFGTBL_ChangeReq, h->vaddr + SA5_DOORBELL);
|
2015-01-23 22:45:01 +00:00
|
|
|
if (hpsa_wait_for_mode_change_ack(h)) {
|
|
|
|
|
dev_err(&h->pdev->dev,
|
|
|
|
|
"performant mode problem - enabling ioaccel mode\n");
|
|
|
|
|
return -ENODEV;
|
|
|
|
|
}
|
|
|
|
|
return 0;
|
2014-02-18 19:55:17 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:33:22 +00:00
|
|
|
/* Free ioaccel1 mode command blocks and block fetch table */
|
|
|
|
|
static void hpsa_free_ioaccel1_cmd_and_bft(struct ctlr_info *h)
|
|
|
|
|
{
|
2015-04-23 14:33:48 +00:00
|
|
|
if (h->ioaccel_cmd_pool) {
|
2015-04-23 14:33:22 +00:00
|
|
|
pci_free_consistent(h->pdev,
|
|
|
|
|
h->nr_cmds * sizeof(*h->ioaccel_cmd_pool),
|
|
|
|
|
h->ioaccel_cmd_pool,
|
|
|
|
|
h->ioaccel_cmd_pool_dhandle);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->ioaccel_cmd_pool = NULL;
|
|
|
|
|
h->ioaccel_cmd_pool_dhandle = 0;
|
|
|
|
|
}
|
2015-04-23 14:33:22 +00:00
|
|
|
kfree(h->ioaccel1_blockFetchTable);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->ioaccel1_blockFetchTable = NULL;
|
2015-04-23 14:33:22 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:32:27 +00:00
|
|
|
/* Allocate ioaccel1 mode command blocks and block fetch table */
|
|
|
|
|
static int hpsa_alloc_ioaccel1_cmd_and_bft(struct ctlr_info *h)
|
2014-02-18 19:55:17 +00:00
|
|
|
{
|
2014-02-18 19:55:33 +00:00
|
|
|
h->ioaccel_maxsg =
|
|
|
|
|
readl(&(h->cfgtable->io_accel_max_embedded_sg_count));
|
|
|
|
|
if (h->ioaccel_maxsg > IOACCEL1_MAXSGENTRIES)
|
|
|
|
|
h->ioaccel_maxsg = IOACCEL1_MAXSGENTRIES;
|
|
|
|
|
|
2014-02-18 19:55:17 +00:00
|
|
|
/* Command structures must be aligned on a 128-byte boundary
|
|
|
|
|
* because the 7 lower bits of the address are used by the
|
|
|
|
|
* hardware.
|
|
|
|
|
*/
|
|
|
|
|
BUILD_BUG_ON(sizeof(struct io_accel1_cmd) %
|
|
|
|
|
IOACCEL1_COMMANDLIST_ALIGNMENT);
|
|
|
|
|
h->ioaccel_cmd_pool =
|
|
|
|
|
pci_alloc_consistent(h->pdev,
|
|
|
|
|
h->nr_cmds * sizeof(*h->ioaccel_cmd_pool),
|
|
|
|
|
&(h->ioaccel_cmd_pool_dhandle));
|
|
|
|
|
|
|
|
|
|
h->ioaccel1_blockFetchTable =
|
2014-02-18 19:55:33 +00:00
|
|
|
kmalloc(((h->ioaccel_maxsg + 1) *
|
2014-02-18 19:55:17 +00:00
|
|
|
sizeof(u32)), GFP_KERNEL);
|
|
|
|
|
|
|
|
|
|
if ((h->ioaccel_cmd_pool == NULL) ||
|
|
|
|
|
(h->ioaccel1_blockFetchTable == NULL))
|
|
|
|
|
goto clean_up;
|
|
|
|
|
|
|
|
|
|
memset(h->ioaccel_cmd_pool, 0,
|
|
|
|
|
h->nr_cmds * sizeof(*h->ioaccel_cmd_pool));
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
clean_up:
|
2015-04-23 14:33:22 +00:00
|
|
|
hpsa_free_ioaccel1_cmd_and_bft(h);
|
2015-04-23 14:33:43 +00:00
|
|
|
return -ENOMEM;
|
2010-05-27 20:14:19 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:33:22 +00:00
|
|
|
/* Free ioaccel2 mode command blocks and block fetch table */
|
|
|
|
|
static void hpsa_free_ioaccel2_cmd_and_bft(struct ctlr_info *h)
|
|
|
|
|
{
|
2015-04-23 14:33:27 +00:00
|
|
|
hpsa_free_ioaccel2_sg_chain_blocks(h);
|
|
|
|
|
|
2015-04-23 14:33:48 +00:00
|
|
|
if (h->ioaccel2_cmd_pool) {
|
2015-04-23 14:33:22 +00:00
|
|
|
pci_free_consistent(h->pdev,
|
|
|
|
|
h->nr_cmds * sizeof(*h->ioaccel2_cmd_pool),
|
|
|
|
|
h->ioaccel2_cmd_pool,
|
|
|
|
|
h->ioaccel2_cmd_pool_dhandle);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->ioaccel2_cmd_pool = NULL;
|
|
|
|
|
h->ioaccel2_cmd_pool_dhandle = 0;
|
|
|
|
|
}
|
2015-04-23 14:33:22 +00:00
|
|
|
kfree(h->ioaccel2_blockFetchTable);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->ioaccel2_blockFetchTable = NULL;
|
2015-04-23 14:33:22 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:32:27 +00:00
|
|
|
/* Allocate ioaccel2 mode command blocks and block fetch table */
|
|
|
|
|
static int hpsa_alloc_ioaccel2_cmd_and_bft(struct ctlr_info *h)
|
2014-02-18 19:56:14 +00:00
|
|
|
{
|
2015-04-23 14:33:27 +00:00
|
|
|
int rc;
|
|
|
|
|
|
2014-02-18 19:56:14 +00:00
|
|
|
/* Allocate ioaccel2 mode command blocks and block fetch table */
|
|
|
|
|
|
|
|
|
|
h->ioaccel_maxsg =
|
|
|
|
|
readl(&(h->cfgtable->io_accel_max_embedded_sg_count));
|
|
|
|
|
if (h->ioaccel_maxsg > IOACCEL2_MAXSGENTRIES)
|
|
|
|
|
h->ioaccel_maxsg = IOACCEL2_MAXSGENTRIES;
|
|
|
|
|
|
|
|
|
|
BUILD_BUG_ON(sizeof(struct io_accel2_cmd) %
|
|
|
|
|
IOACCEL2_COMMANDLIST_ALIGNMENT);
|
|
|
|
|
h->ioaccel2_cmd_pool =
|
|
|
|
|
pci_alloc_consistent(h->pdev,
|
|
|
|
|
h->nr_cmds * sizeof(*h->ioaccel2_cmd_pool),
|
|
|
|
|
&(h->ioaccel2_cmd_pool_dhandle));
|
|
|
|
|
|
|
|
|
|
h->ioaccel2_blockFetchTable =
|
|
|
|
|
kmalloc(((h->ioaccel_maxsg + 1) *
|
|
|
|
|
sizeof(u32)), GFP_KERNEL);
|
|
|
|
|
|
|
|
|
|
if ((h->ioaccel2_cmd_pool == NULL) ||
|
2015-04-23 14:33:27 +00:00
|
|
|
(h->ioaccel2_blockFetchTable == NULL)) {
|
|
|
|
|
rc = -ENOMEM;
|
|
|
|
|
goto clean_up;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
rc = hpsa_allocate_ioaccel2_sg_chain_blocks(h);
|
|
|
|
|
if (rc)
|
2014-02-18 19:56:14 +00:00
|
|
|
goto clean_up;
|
|
|
|
|
|
|
|
|
|
memset(h->ioaccel2_cmd_pool, 0,
|
|
|
|
|
h->nr_cmds * sizeof(*h->ioaccel2_cmd_pool));
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
clean_up:
|
2015-04-23 14:33:22 +00:00
|
|
|
hpsa_free_ioaccel2_cmd_and_bft(h);
|
2015-04-23 14:33:27 +00:00
|
|
|
return rc;
|
2014-02-18 19:56:14 +00:00
|
|
|
}
|
|
|
|
|
|
2015-04-23 14:33:48 +00:00
|
|
|
/* Free items allocated by hpsa_put_ctlr_into_performant_mode */
|
|
|
|
|
static void hpsa_free_performant_mode(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
kfree(h->blockFetchTable);
|
|
|
|
|
h->blockFetchTable = NULL;
|
|
|
|
|
hpsa_free_reply_queues(h);
|
|
|
|
|
hpsa_free_ioaccel1_cmd_and_bft(h);
|
|
|
|
|
hpsa_free_ioaccel2_cmd_and_bft(h);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* return -ENODEV on error, 0 on success (or no action)
|
|
|
|
|
* allocates numerous items that must be freed later
|
|
|
|
|
*/
|
|
|
|
|
static int hpsa_put_ctlr_into_performant_mode(struct ctlr_info *h)
|
2010-05-27 20:14:19 +00:00
|
|
|
{
|
|
|
|
|
u32 trans_support;
|
2014-02-18 19:55:17 +00:00
|
|
|
unsigned long transMethod = CFGTBL_Trans_Performant |
|
|
|
|
|
CFGTBL_Trans_use_short_tags;
|
2015-04-23 14:33:48 +00:00
|
|
|
int i, rc;
|
2010-05-27 20:14:19 +00:00
|
|
|
|
2011-01-06 20:48:29 +00:00
|
|
|
if (hpsa_simple_mode)
|
2015-04-23 14:33:48 +00:00
|
|
|
return 0;
|
2011-01-06 20:48:29 +00:00
|
|
|
|
2014-04-14 19:01:09 +00:00
|
|
|
trans_support = readl(&(h->cfgtable->TransportSupport));
|
|
|
|
|
if (!(trans_support & PERFORMANT_MODE))
|
2015-04-23 14:33:48 +00:00
|
|
|
return 0;
|
2014-04-14 19:01:09 +00:00
|
|
|
|
2014-02-18 19:55:17 +00:00
|
|
|
/* Check for I/O accelerator mode support */
|
|
|
|
|
if (trans_support & CFGTBL_Trans_io_accel1) {
|
|
|
|
|
transMethod |= CFGTBL_Trans_io_accel1 |
|
|
|
|
|
CFGTBL_Trans_enable_directed_msix;
|
2015-04-23 14:33:48 +00:00
|
|
|
rc = hpsa_alloc_ioaccel1_cmd_and_bft(h);
|
|
|
|
|
if (rc)
|
|
|
|
|
return rc;
|
|
|
|
|
} else if (trans_support & CFGTBL_Trans_io_accel2) {
|
|
|
|
|
transMethod |= CFGTBL_Trans_io_accel2 |
|
2014-02-18 19:56:14 +00:00
|
|
|
CFGTBL_Trans_enable_directed_msix;
|
2015-04-23 14:33:48 +00:00
|
|
|
rc = hpsa_alloc_ioaccel2_cmd_and_bft(h);
|
|
|
|
|
if (rc)
|
|
|
|
|
return rc;
|
2014-02-18 19:55:17 +00:00
|
|
|
}
|
|
|
|
|
|
2014-01-15 12:30:53 +00:00
|
|
|
h->nreply_queues = h->msix_vector > 0 ? h->msix_vector : 1;
|
2010-06-16 18:51:56 +00:00
|
|
|
hpsa_get_max_perf_mode_cmds(h);
|
2010-05-27 20:14:19 +00:00
|
|
|
/* Performant mode ring buffer and supporting data structures */
|
2014-05-29 15:53:07 +00:00
|
|
|
h->reply_queue_size = h->max_commands * sizeof(u64);
|
2010-05-27 20:14:19 +00:00
|
|
|
|
2012-05-01 16:43:06 +00:00
|
|
|
for (i = 0; i < h->nreply_queues; i++) {
|
2014-05-29 15:53:07 +00:00
|
|
|
h->reply_queue[i].head = pci_alloc_consistent(h->pdev,
|
|
|
|
|
h->reply_queue_size,
|
|
|
|
|
&(h->reply_queue[i].busaddr));
|
2015-04-23 14:33:48 +00:00
|
|
|
if (!h->reply_queue[i].head) {
|
|
|
|
|
rc = -ENOMEM;
|
|
|
|
|
goto clean1; /* rq, ioaccel */
|
|
|
|
|
}
|
2012-05-01 16:43:06 +00:00
|
|
|
h->reply_queue[i].size = h->max_commands;
|
|
|
|
|
h->reply_queue[i].wraparound = 1; /* spec: init to 1 */
|
|
|
|
|
h->reply_queue[i].current_entry = 0;
|
|
|
|
|
}
|
|
|
|
|
|
2010-05-27 20:14:19 +00:00
|
|
|
/* Need a block fetch table for performant mode */
|
2012-01-19 20:00:48 +00:00
|
|
|
h->blockFetchTable = kmalloc(((SG_ENTRIES_IN_CMD + 1) *
|
2010-05-27 20:14:19 +00:00
|
|
|
sizeof(u32)), GFP_KERNEL);
|
2015-04-23 14:33:48 +00:00
|
|
|
if (!h->blockFetchTable) {
|
|
|
|
|
rc = -ENOMEM;
|
|
|
|
|
goto clean1; /* rq, ioaccel */
|
|
|
|
|
}
|
2010-05-27 20:14:19 +00:00
|
|
|
|
2015-04-23 14:33:48 +00:00
|
|
|
rc = hpsa_enter_performant_mode(h, trans_support);
|
|
|
|
|
if (rc)
|
|
|
|
|
goto clean2; /* bft, rq, ioaccel */
|
|
|
|
|
return 0;
|
2010-02-04 14:42:40 +00:00
|
|
|
|
2015-04-23 14:33:48 +00:00
|
|
|
clean2: /* bft, rq, ioaccel */
|
2010-02-04 14:42:40 +00:00
|
|
|
kfree(h->blockFetchTable);
|
2015-04-23 14:33:48 +00:00
|
|
|
h->blockFetchTable = NULL;
|
|
|
|
|
clean1: /* rq, ioaccel */
|
|
|
|
|
hpsa_free_reply_queues(h);
|
|
|
|
|
hpsa_free_ioaccel1_cmd_and_bft(h);
|
|
|
|
|
hpsa_free_ioaccel2_cmd_and_bft(h);
|
|
|
|
|
return rc;
|
2010-02-04 14:42:40 +00:00
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:57:37 +00:00
|
|
|
static int is_accelerated_cmd(struct CommandList *c)
|
2014-02-18 19:55:43 +00:00
|
|
|
{
|
2014-02-18 19:57:37 +00:00
|
|
|
return c->cmd_type == CMD_IOACCEL1 || c->cmd_type == CMD_IOACCEL2;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_drain_accel_commands(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
struct CommandList *c = NULL;
|
2015-01-23 22:43:09 +00:00
|
|
|
int i, accel_cmds_out;
|
2015-01-23 22:43:35 +00:00
|
|
|
int refcount;
|
2014-02-18 19:55:43 +00:00
|
|
|
|
2015-01-23 22:43:09 +00:00
|
|
|
do { /* wait for all outstanding ioaccel commands to drain out */
|
2014-02-18 19:57:37 +00:00
|
|
|
accel_cmds_out = 0;
|
2015-01-23 22:43:09 +00:00
|
|
|
for (i = 0; i < h->nr_cmds; i++) {
|
|
|
|
|
c = h->cmd_pool + i;
|
2015-01-23 22:43:35 +00:00
|
|
|
refcount = atomic_inc_return(&c->refcount);
|
|
|
|
|
if (refcount > 1) /* Command is allocated */
|
|
|
|
|
accel_cmds_out += is_accelerated_cmd(c);
|
|
|
|
|
cmd_free(h, c);
|
2015-01-23 22:43:09 +00:00
|
|
|
}
|
2014-02-18 19:57:37 +00:00
|
|
|
if (accel_cmds_out <= 0)
|
2015-01-23 22:43:35 +00:00
|
|
|
break;
|
2014-02-18 19:55:43 +00:00
|
|
|
msleep(100);
|
|
|
|
|
} while (1);
|
|
|
|
|
}
|
|
|
|
|
|
2015-11-04 21:52:34 +00:00
|
|
|
static struct hpsa_sas_phy *hpsa_alloc_sas_phy(
|
|
|
|
|
struct hpsa_sas_port *hpsa_sas_port)
|
|
|
|
|
{
|
|
|
|
|
struct hpsa_sas_phy *hpsa_sas_phy;
|
|
|
|
|
struct sas_phy *phy;
|
|
|
|
|
|
|
|
|
|
hpsa_sas_phy = kzalloc(sizeof(*hpsa_sas_phy), GFP_KERNEL);
|
|
|
|
|
if (!hpsa_sas_phy)
|
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
|
|
phy = sas_phy_alloc(hpsa_sas_port->parent_node->parent_dev,
|
|
|
|
|
hpsa_sas_port->next_phy_index);
|
|
|
|
|
if (!phy) {
|
|
|
|
|
kfree(hpsa_sas_phy);
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
hpsa_sas_port->next_phy_index++;
|
|
|
|
|
hpsa_sas_phy->phy = phy;
|
|
|
|
|
hpsa_sas_phy->parent_port = hpsa_sas_port;
|
|
|
|
|
|
|
|
|
|
return hpsa_sas_phy;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_free_sas_phy(struct hpsa_sas_phy *hpsa_sas_phy)
|
|
|
|
|
{
|
|
|
|
|
struct sas_phy *phy = hpsa_sas_phy->phy;
|
|
|
|
|
|
|
|
|
|
sas_port_delete_phy(hpsa_sas_phy->parent_port->port, phy);
|
|
|
|
|
sas_phy_free(phy);
|
|
|
|
|
if (hpsa_sas_phy->added_to_port)
|
|
|
|
|
list_del(&hpsa_sas_phy->phy_list_entry);
|
|
|
|
|
kfree(hpsa_sas_phy);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int hpsa_sas_port_add_phy(struct hpsa_sas_phy *hpsa_sas_phy)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
struct hpsa_sas_port *hpsa_sas_port;
|
|
|
|
|
struct sas_phy *phy;
|
|
|
|
|
struct sas_identify *identify;
|
|
|
|
|
|
|
|
|
|
hpsa_sas_port = hpsa_sas_phy->parent_port;
|
|
|
|
|
phy = hpsa_sas_phy->phy;
|
|
|
|
|
|
|
|
|
|
identify = &phy->identify;
|
|
|
|
|
memset(identify, 0, sizeof(*identify));
|
|
|
|
|
identify->sas_address = hpsa_sas_port->sas_address;
|
|
|
|
|
identify->device_type = SAS_END_DEVICE;
|
|
|
|
|
identify->initiator_port_protocols = SAS_PROTOCOL_STP;
|
|
|
|
|
identify->target_port_protocols = SAS_PROTOCOL_STP;
|
|
|
|
|
phy->minimum_linkrate_hw = SAS_LINK_RATE_UNKNOWN;
|
|
|
|
|
phy->maximum_linkrate_hw = SAS_LINK_RATE_UNKNOWN;
|
|
|
|
|
phy->minimum_linkrate = SAS_LINK_RATE_UNKNOWN;
|
|
|
|
|
phy->maximum_linkrate = SAS_LINK_RATE_UNKNOWN;
|
|
|
|
|
phy->negotiated_linkrate = SAS_LINK_RATE_UNKNOWN;
|
|
|
|
|
|
|
|
|
|
rc = sas_phy_add(hpsa_sas_phy->phy);
|
|
|
|
|
if (rc)
|
|
|
|
|
return rc;
|
|
|
|
|
|
|
|
|
|
sas_port_add_phy(hpsa_sas_port->port, hpsa_sas_phy->phy);
|
|
|
|
|
list_add_tail(&hpsa_sas_phy->phy_list_entry,
|
|
|
|
|
&hpsa_sas_port->phy_list_head);
|
|
|
|
|
hpsa_sas_phy->added_to_port = true;
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int
|
|
|
|
|
hpsa_sas_port_add_rphy(struct hpsa_sas_port *hpsa_sas_port,
|
|
|
|
|
struct sas_rphy *rphy)
|
|
|
|
|
{
|
|
|
|
|
struct sas_identify *identify;
|
|
|
|
|
|
|
|
|
|
identify = &rphy->identify;
|
|
|
|
|
identify->sas_address = hpsa_sas_port->sas_address;
|
|
|
|
|
identify->initiator_port_protocols = SAS_PROTOCOL_STP;
|
|
|
|
|
identify->target_port_protocols = SAS_PROTOCOL_STP;
|
|
|
|
|
|
|
|
|
|
return sas_rphy_add(rphy);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static struct hpsa_sas_port
|
|
|
|
|
*hpsa_alloc_sas_port(struct hpsa_sas_node *hpsa_sas_node,
|
|
|
|
|
u64 sas_address)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
struct hpsa_sas_port *hpsa_sas_port;
|
|
|
|
|
struct sas_port *port;
|
|
|
|
|
|
|
|
|
|
hpsa_sas_port = kzalloc(sizeof(*hpsa_sas_port), GFP_KERNEL);
|
|
|
|
|
if (!hpsa_sas_port)
|
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
|
|
INIT_LIST_HEAD(&hpsa_sas_port->phy_list_head);
|
|
|
|
|
hpsa_sas_port->parent_node = hpsa_sas_node;
|
|
|
|
|
|
|
|
|
|
port = sas_port_alloc_num(hpsa_sas_node->parent_dev);
|
|
|
|
|
if (!port)
|
|
|
|
|
goto free_hpsa_port;
|
|
|
|
|
|
|
|
|
|
rc = sas_port_add(port);
|
|
|
|
|
if (rc)
|
|
|
|
|
goto free_sas_port;
|
|
|
|
|
|
|
|
|
|
hpsa_sas_port->port = port;
|
|
|
|
|
hpsa_sas_port->sas_address = sas_address;
|
|
|
|
|
list_add_tail(&hpsa_sas_port->port_list_entry,
|
|
|
|
|
&hpsa_sas_node->port_list_head);
|
|
|
|
|
|
|
|
|
|
return hpsa_sas_port;
|
|
|
|
|
|
|
|
|
|
free_sas_port:
|
|
|
|
|
sas_port_free(port);
|
|
|
|
|
free_hpsa_port:
|
|
|
|
|
kfree(hpsa_sas_port);
|
|
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_free_sas_port(struct hpsa_sas_port *hpsa_sas_port)
|
|
|
|
|
{
|
|
|
|
|
struct hpsa_sas_phy *hpsa_sas_phy;
|
|
|
|
|
struct hpsa_sas_phy *next;
|
|
|
|
|
|
|
|
|
|
list_for_each_entry_safe(hpsa_sas_phy, next,
|
|
|
|
|
&hpsa_sas_port->phy_list_head, phy_list_entry)
|
|
|
|
|
hpsa_free_sas_phy(hpsa_sas_phy);
|
|
|
|
|
|
|
|
|
|
sas_port_delete(hpsa_sas_port->port);
|
|
|
|
|
list_del(&hpsa_sas_port->port_list_entry);
|
|
|
|
|
kfree(hpsa_sas_port);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static struct hpsa_sas_node *hpsa_alloc_sas_node(struct device *parent_dev)
|
|
|
|
|
{
|
|
|
|
|
struct hpsa_sas_node *hpsa_sas_node;
|
|
|
|
|
|
|
|
|
|
hpsa_sas_node = kzalloc(sizeof(*hpsa_sas_node), GFP_KERNEL);
|
|
|
|
|
if (hpsa_sas_node) {
|
|
|
|
|
hpsa_sas_node->parent_dev = parent_dev;
|
|
|
|
|
INIT_LIST_HEAD(&hpsa_sas_node->port_list_head);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return hpsa_sas_node;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_free_sas_node(struct hpsa_sas_node *hpsa_sas_node)
|
|
|
|
|
{
|
|
|
|
|
struct hpsa_sas_port *hpsa_sas_port;
|
|
|
|
|
struct hpsa_sas_port *next;
|
|
|
|
|
|
|
|
|
|
if (!hpsa_sas_node)
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
list_for_each_entry_safe(hpsa_sas_port, next,
|
|
|
|
|
&hpsa_sas_node->port_list_head, port_list_entry)
|
|
|
|
|
hpsa_free_sas_port(hpsa_sas_port);
|
|
|
|
|
|
|
|
|
|
kfree(hpsa_sas_node);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static struct hpsa_scsi_dev_t
|
|
|
|
|
*hpsa_find_device_by_sas_rphy(struct ctlr_info *h,
|
|
|
|
|
struct sas_rphy *rphy)
|
|
|
|
|
{
|
|
|
|
|
int i;
|
|
|
|
|
struct hpsa_scsi_dev_t *device;
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < h->ndevices; i++) {
|
|
|
|
|
device = h->dev[i];
|
|
|
|
|
if (!device->sas_port)
|
|
|
|
|
continue;
|
|
|
|
|
if (device->sas_port->rphy == rphy)
|
|
|
|
|
return device;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int hpsa_add_sas_host(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
struct device *parent_dev;
|
|
|
|
|
struct hpsa_sas_node *hpsa_sas_node;
|
|
|
|
|
struct hpsa_sas_port *hpsa_sas_port;
|
|
|
|
|
struct hpsa_sas_phy *hpsa_sas_phy;
|
|
|
|
|
|
|
|
|
|
parent_dev = &h->scsi_host->shost_gendev;
|
|
|
|
|
|
|
|
|
|
hpsa_sas_node = hpsa_alloc_sas_node(parent_dev);
|
|
|
|
|
if (!hpsa_sas_node)
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
|
|
hpsa_sas_port = hpsa_alloc_sas_port(hpsa_sas_node, h->sas_address);
|
|
|
|
|
if (!hpsa_sas_port) {
|
|
|
|
|
rc = -ENODEV;
|
|
|
|
|
goto free_sas_node;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
hpsa_sas_phy = hpsa_alloc_sas_phy(hpsa_sas_port);
|
|
|
|
|
if (!hpsa_sas_phy) {
|
|
|
|
|
rc = -ENODEV;
|
|
|
|
|
goto free_sas_port;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
rc = hpsa_sas_port_add_phy(hpsa_sas_phy);
|
|
|
|
|
if (rc)
|
|
|
|
|
goto free_sas_phy;
|
|
|
|
|
|
|
|
|
|
h->sas_host = hpsa_sas_node;
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
free_sas_phy:
|
|
|
|
|
hpsa_free_sas_phy(hpsa_sas_phy);
|
|
|
|
|
free_sas_port:
|
|
|
|
|
hpsa_free_sas_port(hpsa_sas_port);
|
|
|
|
|
free_sas_node:
|
|
|
|
|
hpsa_free_sas_node(hpsa_sas_node);
|
|
|
|
|
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_delete_sas_host(struct ctlr_info *h)
|
|
|
|
|
{
|
|
|
|
|
hpsa_free_sas_node(h->sas_host);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int hpsa_add_sas_device(struct hpsa_sas_node *hpsa_sas_node,
|
|
|
|
|
struct hpsa_scsi_dev_t *device)
|
|
|
|
|
{
|
|
|
|
|
int rc;
|
|
|
|
|
struct hpsa_sas_port *hpsa_sas_port;
|
|
|
|
|
struct sas_rphy *rphy;
|
|
|
|
|
|
|
|
|
|
hpsa_sas_port = hpsa_alloc_sas_port(hpsa_sas_node, device->sas_address);
|
|
|
|
|
if (!hpsa_sas_port)
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
|
|
rphy = sas_end_device_alloc(hpsa_sas_port->port);
|
|
|
|
|
if (!rphy) {
|
|
|
|
|
rc = -ENODEV;
|
|
|
|
|
goto free_sas_port;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
hpsa_sas_port->rphy = rphy;
|
|
|
|
|
device->sas_port = hpsa_sas_port;
|
|
|
|
|
|
|
|
|
|
rc = hpsa_sas_port_add_rphy(hpsa_sas_port, rphy);
|
|
|
|
|
if (rc)
|
|
|
|
|
goto free_sas_port;
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
free_sas_port:
|
|
|
|
|
hpsa_free_sas_port(hpsa_sas_port);
|
|
|
|
|
device->sas_port = NULL;
|
|
|
|
|
|
|
|
|
|
return rc;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void hpsa_remove_sas_device(struct hpsa_scsi_dev_t *device)
|
|
|
|
|
{
|
|
|
|
|
if (device->sas_port) {
|
|
|
|
|
hpsa_free_sas_port(device->sas_port);
|
|
|
|
|
device->sas_port = NULL;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int
|
|
|
|
|
hpsa_sas_get_linkerrors(struct sas_phy *phy)
|
|
|
|
|
{
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int
|
|
|
|
|
hpsa_sas_get_enclosure_identifier(struct sas_rphy *rphy, u64 *identifier)
|
|
|
|
|
{
|
2016-04-14 09:37:44 +00:00
|
|
|
*identifier = 0;
|
2015-11-04 21:52:34 +00:00
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int
|
|
|
|
|
hpsa_sas_get_bay_identifier(struct sas_rphy *rphy)
|
|
|
|
|
{
|
|
|
|
|
return -ENXIO;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int
|
|
|
|
|
hpsa_sas_phy_reset(struct sas_phy *phy, int hard_reset)
|
|
|
|
|
{
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int
|
|
|
|
|
hpsa_sas_phy_enable(struct sas_phy *phy, int enable)
|
|
|
|
|
{
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int
|
|
|
|
|
hpsa_sas_phy_setup(struct sas_phy *phy)
|
|
|
|
|
{
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
hpsa_sas_phy_release(struct sas_phy *phy)
|
|
|
|
|
{
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int
|
|
|
|
|
hpsa_sas_phy_speed(struct sas_phy *phy, struct sas_phy_linkrates *rates)
|
|
|
|
|
{
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* SMP = Serial Management Protocol */
|
|
|
|
|
static int
|
|
|
|
|
hpsa_sas_smp_handler(struct Scsi_Host *shost, struct sas_rphy *rphy,
|
|
|
|
|
struct request *req)
|
|
|
|
|
{
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static struct sas_function_template hpsa_sas_transport_functions = {
|
|
|
|
|
.get_linkerrors = hpsa_sas_get_linkerrors,
|
|
|
|
|
.get_enclosure_identifier = hpsa_sas_get_enclosure_identifier,
|
|
|
|
|
.get_bay_identifier = hpsa_sas_get_bay_identifier,
|
|
|
|
|
.phy_reset = hpsa_sas_phy_reset,
|
|
|
|
|
.phy_enable = hpsa_sas_phy_enable,
|
|
|
|
|
.phy_setup = hpsa_sas_phy_setup,
|
|
|
|
|
.phy_release = hpsa_sas_phy_release,
|
|
|
|
|
.set_phy_speed = hpsa_sas_phy_speed,
|
|
|
|
|
.smp_handler = hpsa_sas_smp_handler,
|
|
|
|
|
};
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
/*
|
|
|
|
|
* This is it. Register the PCI driver information for the cards we control
|
|
|
|
|
* the OS will call our registered routines when it finds one of our cards.
|
|
|
|
|
*/
|
|
|
|
|
static int __init hpsa_init(void)
|
|
|
|
|
{
|
2015-11-04 21:52:34 +00:00
|
|
|
int rc;
|
|
|
|
|
|
|
|
|
|
hpsa_sas_transport_template =
|
|
|
|
|
sas_attach_transport(&hpsa_sas_transport_functions);
|
|
|
|
|
if (!hpsa_sas_transport_template)
|
|
|
|
|
return -ENODEV;
|
|
|
|
|
|
|
|
|
|
rc = pci_register_driver(&hpsa_pci_driver);
|
|
|
|
|
|
|
|
|
|
if (rc)
|
|
|
|
|
sas_release_transport(hpsa_sas_transport_template);
|
|
|
|
|
|
|
|
|
|
return rc;
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void __exit hpsa_cleanup(void)
|
|
|
|
|
{
|
|
|
|
|
pci_unregister_driver(&hpsa_pci_driver);
|
2015-11-04 21:52:34 +00:00
|
|
|
sas_release_transport(hpsa_sas_transport_template);
|
2009-12-08 22:09:11 +00:00
|
|
|
}
|
|
|
|
|
|
2014-02-18 19:55:17 +00:00
|
|
|
static void __attribute__((unused)) verify_offsets(void)
|
|
|
|
|
{
|
2014-02-18 19:57:31 +00:00
|
|
|
#define VERIFY_OFFSET(member, offset) \
|
|
|
|
|
BUILD_BUG_ON(offsetof(struct raid_map_data, member) != offset)
|
|
|
|
|
|
|
|
|
|
VERIFY_OFFSET(structure_size, 0);
|
|
|
|
|
VERIFY_OFFSET(volume_blk_size, 4);
|
|
|
|
|
VERIFY_OFFSET(volume_blk_cnt, 8);
|
|
|
|
|
VERIFY_OFFSET(phys_blk_shift, 16);
|
|
|
|
|
VERIFY_OFFSET(parity_rotation_shift, 17);
|
|
|
|
|
VERIFY_OFFSET(strip_size, 18);
|
|
|
|
|
VERIFY_OFFSET(disk_starting_blk, 20);
|
|
|
|
|
VERIFY_OFFSET(disk_blk_cnt, 28);
|
|
|
|
|
VERIFY_OFFSET(data_disks_per_row, 36);
|
|
|
|
|
VERIFY_OFFSET(metadata_disks_per_row, 38);
|
|
|
|
|
VERIFY_OFFSET(row_cnt, 40);
|
|
|
|
|
VERIFY_OFFSET(layout_map_count, 42);
|
|
|
|
|
VERIFY_OFFSET(flags, 44);
|
|
|
|
|
VERIFY_OFFSET(dekindex, 46);
|
|
|
|
|
/* VERIFY_OFFSET(reserved, 48 */
|
|
|
|
|
VERIFY_OFFSET(data, 64);
|
|
|
|
|
|
|
|
|
|
#undef VERIFY_OFFSET
|
|
|
|
|
|
2014-02-18 19:56:04 +00:00
|
|
|
#define VERIFY_OFFSET(member, offset) \
|
|
|
|
|
BUILD_BUG_ON(offsetof(struct io_accel2_cmd, member) != offset)
|
|
|
|
|
|
|
|
|
|
VERIFY_OFFSET(IU_type, 0);
|
|
|
|
|
VERIFY_OFFSET(direction, 1);
|
|
|
|
|
VERIFY_OFFSET(reply_queue, 2);
|
|
|
|
|
/* VERIFY_OFFSET(reserved1, 3); */
|
|
|
|
|
VERIFY_OFFSET(scsi_nexus, 4);
|
|
|
|
|
VERIFY_OFFSET(Tag, 8);
|
|
|
|
|
VERIFY_OFFSET(cdb, 16);
|
|
|
|
|
VERIFY_OFFSET(cciss_lun, 32);
|
|
|
|
|
VERIFY_OFFSET(data_len, 40);
|
|
|
|
|
VERIFY_OFFSET(cmd_priority_task_attr, 44);
|
|
|
|
|
VERIFY_OFFSET(sg_count, 45);
|
|
|
|
|
/* VERIFY_OFFSET(reserved3 */
|
|
|
|
|
VERIFY_OFFSET(err_ptr, 48);
|
|
|
|
|
VERIFY_OFFSET(err_len, 56);
|
|
|
|
|
/* VERIFY_OFFSET(reserved4 */
|
|
|
|
|
VERIFY_OFFSET(sg, 64);
|
|
|
|
|
|
|
|
|
|
#undef VERIFY_OFFSET
|
|
|
|
|
|
2014-02-18 19:55:17 +00:00
|
|
|
#define VERIFY_OFFSET(member, offset) \
|
|
|
|
|
BUILD_BUG_ON(offsetof(struct io_accel1_cmd, member) != offset)
|
|
|
|
|
|
|
|
|
|
VERIFY_OFFSET(dev_handle, 0x00);
|
|
|
|
|
VERIFY_OFFSET(reserved1, 0x02);
|
|
|
|
|
VERIFY_OFFSET(function, 0x03);
|
|
|
|
|
VERIFY_OFFSET(reserved2, 0x04);
|
|
|
|
|
VERIFY_OFFSET(err_info, 0x0C);
|
|
|
|
|
VERIFY_OFFSET(reserved3, 0x10);
|
|
|
|
|
VERIFY_OFFSET(err_info_len, 0x12);
|
|
|
|
|
VERIFY_OFFSET(reserved4, 0x13);
|
|
|
|
|
VERIFY_OFFSET(sgl_offset, 0x14);
|
|
|
|
|
VERIFY_OFFSET(reserved5, 0x15);
|
|
|
|
|
VERIFY_OFFSET(transfer_len, 0x1C);
|
|
|
|
|
VERIFY_OFFSET(reserved6, 0x20);
|
|
|
|
|
VERIFY_OFFSET(io_flags, 0x24);
|
|
|
|
|
VERIFY_OFFSET(reserved7, 0x26);
|
|
|
|
|
VERIFY_OFFSET(LUN, 0x34);
|
|
|
|
|
VERIFY_OFFSET(control, 0x3C);
|
|
|
|
|
VERIFY_OFFSET(CDB, 0x40);
|
|
|
|
|
VERIFY_OFFSET(reserved8, 0x50);
|
|
|
|
|
VERIFY_OFFSET(host_context_flags, 0x60);
|
|
|
|
|
VERIFY_OFFSET(timeout_sec, 0x62);
|
|
|
|
|
VERIFY_OFFSET(ReplyQueue, 0x64);
|
|
|
|
|
VERIFY_OFFSET(reserved9, 0x65);
|
2014-11-14 23:26:59 +00:00
|
|
|
VERIFY_OFFSET(tag, 0x68);
|
2014-02-18 19:55:17 +00:00
|
|
|
VERIFY_OFFSET(host_addr, 0x70);
|
|
|
|
|
VERIFY_OFFSET(CISS_LUN, 0x78);
|
|
|
|
|
VERIFY_OFFSET(SG, 0x78 + 8);
|
|
|
|
|
#undef VERIFY_OFFSET
|
|
|
|
|
}
|
|
|
|
|
|
2009-12-08 22:09:11 +00:00
|
|
|
module_init(hpsa_init);
|
|
|
|
|
module_exit(hpsa_cleanup);
|
