mirror of
https://github.com/torvalds/linux.git
synced 2024-11-11 22:51:42 +00:00
5af5eee7ca
Adapter Shutdown and Unregistration cleanup - Correct the logic around hba shutdown. Prior to final reset, the driver must wait for all XRIs to return from the adapter. Added logic to poll, progressively slowing the poll rate as delay gets longer. - Correct behavior around the rsvd1 field in UNREG_RPI_ALL mailbox completion and final rpi cleanup. - Updated logic to move pending VPI registrations to their completion in cases where a CVL may be received while registration in progress. - Added unreg all rpi mailbox command before unreg vpi. Signed-off-by: Alex Iannicelli <alex.iannicelli@emulex.com> Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
9095 lines
269 KiB
C
9095 lines
269 KiB
C
/*******************************************************************
|
|
* This file is part of the Emulex Linux Device Driver for *
|
|
* Fibre Channel Host Bus Adapters. *
|
|
* Copyright (C) 2004-2010 Emulex. All rights reserved. *
|
|
* EMULEX and SLI are trademarks of Emulex. *
|
|
* www.emulex.com *
|
|
* Portions Copyright (C) 2004-2005 Christoph Hellwig *
|
|
* *
|
|
* This program is free software; you can redistribute it and/or *
|
|
* modify it under the terms of version 2 of the GNU General *
|
|
* Public License as published by the Free Software Foundation. *
|
|
* This program is distributed in the hope that it will be useful. *
|
|
* ALL EXPRESS OR IMPLIED CONDITIONS, REPRESENTATIONS AND *
|
|
* WARRANTIES, INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY, *
|
|
* FITNESS FOR A PARTICULAR PURPOSE, OR NON-INFRINGEMENT, ARE *
|
|
* DISCLAIMED, EXCEPT TO THE EXTENT THAT SUCH DISCLAIMERS ARE HELD *
|
|
* TO BE LEGALLY INVALID. See the GNU General Public License for *
|
|
* more details, a copy of which can be found in the file COPYING *
|
|
* included with this package. *
|
|
*******************************************************************/
|
|
|
|
#include <linux/blkdev.h>
|
|
#include <linux/delay.h>
|
|
#include <linux/dma-mapping.h>
|
|
#include <linux/idr.h>
|
|
#include <linux/interrupt.h>
|
|
#include <linux/kthread.h>
|
|
#include <linux/pci.h>
|
|
#include <linux/spinlock.h>
|
|
#include <linux/ctype.h>
|
|
#include <linux/aer.h>
|
|
#include <linux/slab.h>
|
|
|
|
#include <scsi/scsi.h>
|
|
#include <scsi/scsi_device.h>
|
|
#include <scsi/scsi_host.h>
|
|
#include <scsi/scsi_transport_fc.h>
|
|
|
|
#include "lpfc_hw4.h"
|
|
#include "lpfc_hw.h"
|
|
#include "lpfc_sli.h"
|
|
#include "lpfc_sli4.h"
|
|
#include "lpfc_nl.h"
|
|
#include "lpfc_disc.h"
|
|
#include "lpfc_scsi.h"
|
|
#include "lpfc.h"
|
|
#include "lpfc_logmsg.h"
|
|
#include "lpfc_crtn.h"
|
|
#include "lpfc_vport.h"
|
|
#include "lpfc_version.h"
|
|
|
|
char *_dump_buf_data;
|
|
unsigned long _dump_buf_data_order;
|
|
char *_dump_buf_dif;
|
|
unsigned long _dump_buf_dif_order;
|
|
spinlock_t _dump_buf_lock;
|
|
|
|
static void lpfc_get_hba_model_desc(struct lpfc_hba *, uint8_t *, uint8_t *);
|
|
static int lpfc_post_rcv_buf(struct lpfc_hba *);
|
|
static int lpfc_sli4_queue_create(struct lpfc_hba *);
|
|
static void lpfc_sli4_queue_destroy(struct lpfc_hba *);
|
|
static int lpfc_create_bootstrap_mbox(struct lpfc_hba *);
|
|
static int lpfc_setup_endian_order(struct lpfc_hba *);
|
|
static int lpfc_sli4_read_config(struct lpfc_hba *);
|
|
static void lpfc_destroy_bootstrap_mbox(struct lpfc_hba *);
|
|
static void lpfc_free_sgl_list(struct lpfc_hba *);
|
|
static int lpfc_init_sgl_list(struct lpfc_hba *);
|
|
static int lpfc_init_active_sgl_array(struct lpfc_hba *);
|
|
static void lpfc_free_active_sgl(struct lpfc_hba *);
|
|
static int lpfc_hba_down_post_s3(struct lpfc_hba *phba);
|
|
static int lpfc_hba_down_post_s4(struct lpfc_hba *phba);
|
|
static int lpfc_sli4_cq_event_pool_create(struct lpfc_hba *);
|
|
static void lpfc_sli4_cq_event_pool_destroy(struct lpfc_hba *);
|
|
static void lpfc_sli4_cq_event_release_all(struct lpfc_hba *);
|
|
|
|
static struct scsi_transport_template *lpfc_transport_template = NULL;
|
|
static struct scsi_transport_template *lpfc_vport_transport_template = NULL;
|
|
static DEFINE_IDR(lpfc_hba_index);
|
|
|
|
/**
|
|
* lpfc_config_port_prep - Perform lpfc initialization prior to config port
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine will do LPFC initialization prior to issuing the CONFIG_PORT
|
|
* mailbox command. It retrieves the revision information from the HBA and
|
|
* collects the Vital Product Data (VPD) about the HBA for preparing the
|
|
* configuration of the HBA.
|
|
*
|
|
* Return codes:
|
|
* 0 - success.
|
|
* -ERESTART - requests the SLI layer to reset the HBA and try again.
|
|
* Any other value - indicates an error.
|
|
**/
|
|
int
|
|
lpfc_config_port_prep(struct lpfc_hba *phba)
|
|
{
|
|
lpfc_vpd_t *vp = &phba->vpd;
|
|
int i = 0, rc;
|
|
LPFC_MBOXQ_t *pmb;
|
|
MAILBOX_t *mb;
|
|
char *lpfc_vpd_data = NULL;
|
|
uint16_t offset = 0;
|
|
static char licensed[56] =
|
|
"key unlock for use with gnu public licensed code only\0";
|
|
static int init_key = 1;
|
|
|
|
pmb = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
|
|
if (!pmb) {
|
|
phba->link_state = LPFC_HBA_ERROR;
|
|
return -ENOMEM;
|
|
}
|
|
|
|
mb = &pmb->u.mb;
|
|
phba->link_state = LPFC_INIT_MBX_CMDS;
|
|
|
|
if (lpfc_is_LC_HBA(phba->pcidev->device)) {
|
|
if (init_key) {
|
|
uint32_t *ptext = (uint32_t *) licensed;
|
|
|
|
for (i = 0; i < 56; i += sizeof (uint32_t), ptext++)
|
|
*ptext = cpu_to_be32(*ptext);
|
|
init_key = 0;
|
|
}
|
|
|
|
lpfc_read_nv(phba, pmb);
|
|
memset((char*)mb->un.varRDnvp.rsvd3, 0,
|
|
sizeof (mb->un.varRDnvp.rsvd3));
|
|
memcpy((char*)mb->un.varRDnvp.rsvd3, licensed,
|
|
sizeof (licensed));
|
|
|
|
rc = lpfc_sli_issue_mbox(phba, pmb, MBX_POLL);
|
|
|
|
if (rc != MBX_SUCCESS) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_MBOX,
|
|
"0324 Config Port initialization "
|
|
"error, mbxCmd x%x READ_NVPARM, "
|
|
"mbxStatus x%x\n",
|
|
mb->mbxCommand, mb->mbxStatus);
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
return -ERESTART;
|
|
}
|
|
memcpy(phba->wwnn, (char *)mb->un.varRDnvp.nodename,
|
|
sizeof(phba->wwnn));
|
|
memcpy(phba->wwpn, (char *)mb->un.varRDnvp.portname,
|
|
sizeof(phba->wwpn));
|
|
}
|
|
|
|
phba->sli3_options = 0x0;
|
|
|
|
/* Setup and issue mailbox READ REV command */
|
|
lpfc_read_rev(phba, pmb);
|
|
rc = lpfc_sli_issue_mbox(phba, pmb, MBX_POLL);
|
|
if (rc != MBX_SUCCESS) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0439 Adapter failed to init, mbxCmd x%x "
|
|
"READ_REV, mbxStatus x%x\n",
|
|
mb->mbxCommand, mb->mbxStatus);
|
|
mempool_free( pmb, phba->mbox_mem_pool);
|
|
return -ERESTART;
|
|
}
|
|
|
|
|
|
/*
|
|
* The value of rr must be 1 since the driver set the cv field to 1.
|
|
* This setting requires the FW to set all revision fields.
|
|
*/
|
|
if (mb->un.varRdRev.rr == 0) {
|
|
vp->rev.rBit = 0;
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0440 Adapter failed to init, READ_REV has "
|
|
"missing revision information.\n");
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
return -ERESTART;
|
|
}
|
|
|
|
if (phba->sli_rev == 3 && !mb->un.varRdRev.v3rsp) {
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
return -EINVAL;
|
|
}
|
|
|
|
/* Save information as VPD data */
|
|
vp->rev.rBit = 1;
|
|
memcpy(&vp->sli3Feat, &mb->un.varRdRev.sli3Feat, sizeof(uint32_t));
|
|
vp->rev.sli1FwRev = mb->un.varRdRev.sli1FwRev;
|
|
memcpy(vp->rev.sli1FwName, (char*) mb->un.varRdRev.sli1FwName, 16);
|
|
vp->rev.sli2FwRev = mb->un.varRdRev.sli2FwRev;
|
|
memcpy(vp->rev.sli2FwName, (char *) mb->un.varRdRev.sli2FwName, 16);
|
|
vp->rev.biuRev = mb->un.varRdRev.biuRev;
|
|
vp->rev.smRev = mb->un.varRdRev.smRev;
|
|
vp->rev.smFwRev = mb->un.varRdRev.un.smFwRev;
|
|
vp->rev.endecRev = mb->un.varRdRev.endecRev;
|
|
vp->rev.fcphHigh = mb->un.varRdRev.fcphHigh;
|
|
vp->rev.fcphLow = mb->un.varRdRev.fcphLow;
|
|
vp->rev.feaLevelHigh = mb->un.varRdRev.feaLevelHigh;
|
|
vp->rev.feaLevelLow = mb->un.varRdRev.feaLevelLow;
|
|
vp->rev.postKernRev = mb->un.varRdRev.postKernRev;
|
|
vp->rev.opFwRev = mb->un.varRdRev.opFwRev;
|
|
|
|
/* If the sli feature level is less then 9, we must
|
|
* tear down all RPIs and VPIs on link down if NPIV
|
|
* is enabled.
|
|
*/
|
|
if (vp->rev.feaLevelHigh < 9)
|
|
phba->sli3_options |= LPFC_SLI3_VPORT_TEARDOWN;
|
|
|
|
if (lpfc_is_LC_HBA(phba->pcidev->device))
|
|
memcpy(phba->RandomData, (char *)&mb->un.varWords[24],
|
|
sizeof (phba->RandomData));
|
|
|
|
/* Get adapter VPD information */
|
|
lpfc_vpd_data = kmalloc(DMP_VPD_SIZE, GFP_KERNEL);
|
|
if (!lpfc_vpd_data)
|
|
goto out_free_mbox;
|
|
|
|
do {
|
|
lpfc_dump_mem(phba, pmb, offset, DMP_REGION_VPD);
|
|
rc = lpfc_sli_issue_mbox(phba, pmb, MBX_POLL);
|
|
|
|
if (rc != MBX_SUCCESS) {
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0441 VPD not present on adapter, "
|
|
"mbxCmd x%x DUMP VPD, mbxStatus x%x\n",
|
|
mb->mbxCommand, mb->mbxStatus);
|
|
mb->un.varDmp.word_cnt = 0;
|
|
}
|
|
/* dump mem may return a zero when finished or we got a
|
|
* mailbox error, either way we are done.
|
|
*/
|
|
if (mb->un.varDmp.word_cnt == 0)
|
|
break;
|
|
if (mb->un.varDmp.word_cnt > DMP_VPD_SIZE - offset)
|
|
mb->un.varDmp.word_cnt = DMP_VPD_SIZE - offset;
|
|
lpfc_sli_pcimem_bcopy(((uint8_t *)mb) + DMP_RSP_OFFSET,
|
|
lpfc_vpd_data + offset,
|
|
mb->un.varDmp.word_cnt);
|
|
offset += mb->un.varDmp.word_cnt;
|
|
} while (mb->un.varDmp.word_cnt && offset < DMP_VPD_SIZE);
|
|
lpfc_parse_vpd(phba, lpfc_vpd_data, offset);
|
|
|
|
kfree(lpfc_vpd_data);
|
|
out_free_mbox:
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_config_async_cmpl - Completion handler for config async event mbox cmd
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @pmboxq: pointer to the driver internal queue element for mailbox command.
|
|
*
|
|
* This is the completion handler for driver's configuring asynchronous event
|
|
* mailbox command to the device. If the mailbox command returns successfully,
|
|
* it will set internal async event support flag to 1; otherwise, it will
|
|
* set internal async event support flag to 0.
|
|
**/
|
|
static void
|
|
lpfc_config_async_cmpl(struct lpfc_hba * phba, LPFC_MBOXQ_t * pmboxq)
|
|
{
|
|
if (pmboxq->u.mb.mbxStatus == MBX_SUCCESS)
|
|
phba->temp_sensor_support = 1;
|
|
else
|
|
phba->temp_sensor_support = 0;
|
|
mempool_free(pmboxq, phba->mbox_mem_pool);
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_dump_wakeup_param_cmpl - dump memory mailbox command completion handler
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @pmboxq: pointer to the driver internal queue element for mailbox command.
|
|
*
|
|
* This is the completion handler for dump mailbox command for getting
|
|
* wake up parameters. When this command complete, the response contain
|
|
* Option rom version of the HBA. This function translate the version number
|
|
* into a human readable string and store it in OptionROMVersion.
|
|
**/
|
|
static void
|
|
lpfc_dump_wakeup_param_cmpl(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmboxq)
|
|
{
|
|
struct prog_id *prg;
|
|
uint32_t prog_id_word;
|
|
char dist = ' ';
|
|
/* character array used for decoding dist type. */
|
|
char dist_char[] = "nabx";
|
|
|
|
if (pmboxq->u.mb.mbxStatus != MBX_SUCCESS) {
|
|
mempool_free(pmboxq, phba->mbox_mem_pool);
|
|
return;
|
|
}
|
|
|
|
prg = (struct prog_id *) &prog_id_word;
|
|
|
|
/* word 7 contain option rom version */
|
|
prog_id_word = pmboxq->u.mb.un.varWords[7];
|
|
|
|
/* Decode the Option rom version word to a readable string */
|
|
if (prg->dist < 4)
|
|
dist = dist_char[prg->dist];
|
|
|
|
if ((prg->dist == 3) && (prg->num == 0))
|
|
sprintf(phba->OptionROMVersion, "%d.%d%d",
|
|
prg->ver, prg->rev, prg->lev);
|
|
else
|
|
sprintf(phba->OptionROMVersion, "%d.%d%d%c%d",
|
|
prg->ver, prg->rev, prg->lev,
|
|
dist, prg->num);
|
|
mempool_free(pmboxq, phba->mbox_mem_pool);
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_config_port_post - Perform lpfc initialization after config port
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine will do LPFC initialization after the CONFIG_PORT mailbox
|
|
* command call. It performs all internal resource and state setups on the
|
|
* port: post IOCB buffers, enable appropriate host interrupt attentions,
|
|
* ELS ring timers, etc.
|
|
*
|
|
* Return codes
|
|
* 0 - success.
|
|
* Any other value - error.
|
|
**/
|
|
int
|
|
lpfc_config_port_post(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_vport *vport = phba->pport;
|
|
struct Scsi_Host *shost = lpfc_shost_from_vport(vport);
|
|
LPFC_MBOXQ_t *pmb;
|
|
MAILBOX_t *mb;
|
|
struct lpfc_dmabuf *mp;
|
|
struct lpfc_sli *psli = &phba->sli;
|
|
uint32_t status, timeout;
|
|
int i, j;
|
|
int rc;
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
/*
|
|
* If the Config port completed correctly the HBA is not
|
|
* over heated any more.
|
|
*/
|
|
if (phba->over_temp_state == HBA_OVER_TEMP)
|
|
phba->over_temp_state = HBA_NORMAL_TEMP;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
pmb = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
|
|
if (!pmb) {
|
|
phba->link_state = LPFC_HBA_ERROR;
|
|
return -ENOMEM;
|
|
}
|
|
mb = &pmb->u.mb;
|
|
|
|
/* Get login parameters for NID. */
|
|
rc = lpfc_read_sparam(phba, pmb, 0);
|
|
if (rc) {
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
return -ENOMEM;
|
|
}
|
|
|
|
pmb->vport = vport;
|
|
if (lpfc_sli_issue_mbox(phba, pmb, MBX_POLL) != MBX_SUCCESS) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0448 Adapter failed init, mbxCmd x%x "
|
|
"READ_SPARM mbxStatus x%x\n",
|
|
mb->mbxCommand, mb->mbxStatus);
|
|
phba->link_state = LPFC_HBA_ERROR;
|
|
mp = (struct lpfc_dmabuf *) pmb->context1;
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
lpfc_mbuf_free(phba, mp->virt, mp->phys);
|
|
kfree(mp);
|
|
return -EIO;
|
|
}
|
|
|
|
mp = (struct lpfc_dmabuf *) pmb->context1;
|
|
|
|
memcpy(&vport->fc_sparam, mp->virt, sizeof (struct serv_parm));
|
|
lpfc_mbuf_free(phba, mp->virt, mp->phys);
|
|
kfree(mp);
|
|
pmb->context1 = NULL;
|
|
|
|
if (phba->cfg_soft_wwnn)
|
|
u64_to_wwn(phba->cfg_soft_wwnn,
|
|
vport->fc_sparam.nodeName.u.wwn);
|
|
if (phba->cfg_soft_wwpn)
|
|
u64_to_wwn(phba->cfg_soft_wwpn,
|
|
vport->fc_sparam.portName.u.wwn);
|
|
memcpy(&vport->fc_nodename, &vport->fc_sparam.nodeName,
|
|
sizeof (struct lpfc_name));
|
|
memcpy(&vport->fc_portname, &vport->fc_sparam.portName,
|
|
sizeof (struct lpfc_name));
|
|
|
|
/* Update the fc_host data structures with new wwn. */
|
|
fc_host_node_name(shost) = wwn_to_u64(vport->fc_nodename.u.wwn);
|
|
fc_host_port_name(shost) = wwn_to_u64(vport->fc_portname.u.wwn);
|
|
fc_host_max_npiv_vports(shost) = phba->max_vpi;
|
|
|
|
/* If no serial number in VPD data, use low 6 bytes of WWNN */
|
|
/* This should be consolidated into parse_vpd ? - mr */
|
|
if (phba->SerialNumber[0] == 0) {
|
|
uint8_t *outptr;
|
|
|
|
outptr = &vport->fc_nodename.u.s.IEEE[0];
|
|
for (i = 0; i < 12; i++) {
|
|
status = *outptr++;
|
|
j = ((status & 0xf0) >> 4);
|
|
if (j <= 9)
|
|
phba->SerialNumber[i] =
|
|
(char)((uint8_t) 0x30 + (uint8_t) j);
|
|
else
|
|
phba->SerialNumber[i] =
|
|
(char)((uint8_t) 0x61 + (uint8_t) (j - 10));
|
|
i++;
|
|
j = (status & 0xf);
|
|
if (j <= 9)
|
|
phba->SerialNumber[i] =
|
|
(char)((uint8_t) 0x30 + (uint8_t) j);
|
|
else
|
|
phba->SerialNumber[i] =
|
|
(char)((uint8_t) 0x61 + (uint8_t) (j - 10));
|
|
}
|
|
}
|
|
|
|
lpfc_read_config(phba, pmb);
|
|
pmb->vport = vport;
|
|
if (lpfc_sli_issue_mbox(phba, pmb, MBX_POLL) != MBX_SUCCESS) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0453 Adapter failed to init, mbxCmd x%x "
|
|
"READ_CONFIG, mbxStatus x%x\n",
|
|
mb->mbxCommand, mb->mbxStatus);
|
|
phba->link_state = LPFC_HBA_ERROR;
|
|
mempool_free( pmb, phba->mbox_mem_pool);
|
|
return -EIO;
|
|
}
|
|
|
|
/* Check if the port is disabled */
|
|
lpfc_sli_read_link_ste(phba);
|
|
|
|
/* Reset the DFT_HBA_Q_DEPTH to the max xri */
|
|
if (phba->cfg_hba_queue_depth > (mb->un.varRdConfig.max_xri+1))
|
|
phba->cfg_hba_queue_depth =
|
|
(mb->un.varRdConfig.max_xri + 1) -
|
|
lpfc_sli4_get_els_iocb_cnt(phba);
|
|
|
|
phba->lmt = mb->un.varRdConfig.lmt;
|
|
|
|
/* Get the default values for Model Name and Description */
|
|
lpfc_get_hba_model_desc(phba, phba->ModelName, phba->ModelDesc);
|
|
|
|
if ((phba->cfg_link_speed > LINK_SPEED_10G)
|
|
|| ((phba->cfg_link_speed == LINK_SPEED_1G)
|
|
&& !(phba->lmt & LMT_1Gb))
|
|
|| ((phba->cfg_link_speed == LINK_SPEED_2G)
|
|
&& !(phba->lmt & LMT_2Gb))
|
|
|| ((phba->cfg_link_speed == LINK_SPEED_4G)
|
|
&& !(phba->lmt & LMT_4Gb))
|
|
|| ((phba->cfg_link_speed == LINK_SPEED_8G)
|
|
&& !(phba->lmt & LMT_8Gb))
|
|
|| ((phba->cfg_link_speed == LINK_SPEED_10G)
|
|
&& !(phba->lmt & LMT_10Gb))) {
|
|
/* Reset link speed to auto */
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_LINK_EVENT,
|
|
"1302 Invalid speed for this board: "
|
|
"Reset link speed to auto: x%x\n",
|
|
phba->cfg_link_speed);
|
|
phba->cfg_link_speed = LINK_SPEED_AUTO;
|
|
}
|
|
|
|
phba->link_state = LPFC_LINK_DOWN;
|
|
|
|
/* Only process IOCBs on ELS ring till hba_state is READY */
|
|
if (psli->ring[psli->extra_ring].cmdringaddr)
|
|
psli->ring[psli->extra_ring].flag |= LPFC_STOP_IOCB_EVENT;
|
|
if (psli->ring[psli->fcp_ring].cmdringaddr)
|
|
psli->ring[psli->fcp_ring].flag |= LPFC_STOP_IOCB_EVENT;
|
|
if (psli->ring[psli->next_ring].cmdringaddr)
|
|
psli->ring[psli->next_ring].flag |= LPFC_STOP_IOCB_EVENT;
|
|
|
|
/* Post receive buffers for desired rings */
|
|
if (phba->sli_rev != 3)
|
|
lpfc_post_rcv_buf(phba);
|
|
|
|
/*
|
|
* Configure HBA MSI-X attention conditions to messages if MSI-X mode
|
|
*/
|
|
if (phba->intr_type == MSIX) {
|
|
rc = lpfc_config_msi(phba, pmb);
|
|
if (rc) {
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
return -EIO;
|
|
}
|
|
rc = lpfc_sli_issue_mbox(phba, pmb, MBX_POLL);
|
|
if (rc != MBX_SUCCESS) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_MBOX,
|
|
"0352 Config MSI mailbox command "
|
|
"failed, mbxCmd x%x, mbxStatus x%x\n",
|
|
pmb->u.mb.mbxCommand,
|
|
pmb->u.mb.mbxStatus);
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
return -EIO;
|
|
}
|
|
}
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
/* Initialize ERATT handling flag */
|
|
phba->hba_flag &= ~HBA_ERATT_HANDLED;
|
|
|
|
/* Enable appropriate host interrupts */
|
|
status = readl(phba->HCregaddr);
|
|
status |= HC_MBINT_ENA | HC_ERINT_ENA | HC_LAINT_ENA;
|
|
if (psli->num_rings > 0)
|
|
status |= HC_R0INT_ENA;
|
|
if (psli->num_rings > 1)
|
|
status |= HC_R1INT_ENA;
|
|
if (psli->num_rings > 2)
|
|
status |= HC_R2INT_ENA;
|
|
if (psli->num_rings > 3)
|
|
status |= HC_R3INT_ENA;
|
|
|
|
if ((phba->cfg_poll & ENABLE_FCP_RING_POLLING) &&
|
|
(phba->cfg_poll & DISABLE_FCP_RING_INT))
|
|
status &= ~(HC_R0INT_ENA);
|
|
|
|
writel(status, phba->HCregaddr);
|
|
readl(phba->HCregaddr); /* flush */
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
/* Set up ring-0 (ELS) timer */
|
|
timeout = phba->fc_ratov * 2;
|
|
mod_timer(&vport->els_tmofunc, jiffies + HZ * timeout);
|
|
/* Set up heart beat (HB) timer */
|
|
mod_timer(&phba->hb_tmofunc, jiffies + HZ * LPFC_HB_MBOX_INTERVAL);
|
|
phba->hb_outstanding = 0;
|
|
phba->last_completion_time = jiffies;
|
|
/* Set up error attention (ERATT) polling timer */
|
|
mod_timer(&phba->eratt_poll, jiffies + HZ * LPFC_ERATT_POLL_INTERVAL);
|
|
|
|
if (phba->hba_flag & LINK_DISABLED) {
|
|
lpfc_printf_log(phba,
|
|
KERN_ERR, LOG_INIT,
|
|
"2598 Adapter Link is disabled.\n");
|
|
lpfc_down_link(phba, pmb);
|
|
pmb->mbox_cmpl = lpfc_sli_def_mbox_cmpl;
|
|
rc = lpfc_sli_issue_mbox(phba, pmb, MBX_NOWAIT);
|
|
if ((rc != MBX_SUCCESS) && (rc != MBX_BUSY)) {
|
|
lpfc_printf_log(phba,
|
|
KERN_ERR, LOG_INIT,
|
|
"2599 Adapter failed to issue DOWN_LINK"
|
|
" mbox command rc 0x%x\n", rc);
|
|
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
return -EIO;
|
|
}
|
|
} else if (phba->cfg_suppress_link_up == LPFC_INITIALIZE_LINK) {
|
|
lpfc_init_link(phba, pmb, phba->cfg_topology,
|
|
phba->cfg_link_speed);
|
|
pmb->mbox_cmpl = lpfc_sli_def_mbox_cmpl;
|
|
lpfc_set_loopback_flag(phba);
|
|
rc = lpfc_sli_issue_mbox(phba, pmb, MBX_NOWAIT);
|
|
if (rc != MBX_SUCCESS) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0454 Adapter failed to init, mbxCmd x%x "
|
|
"INIT_LINK, mbxStatus x%x\n",
|
|
mb->mbxCommand, mb->mbxStatus);
|
|
|
|
/* Clear all interrupt enable conditions */
|
|
writel(0, phba->HCregaddr);
|
|
readl(phba->HCregaddr); /* flush */
|
|
/* Clear all pending interrupts */
|
|
writel(0xffffffff, phba->HAregaddr);
|
|
readl(phba->HAregaddr); /* flush */
|
|
|
|
phba->link_state = LPFC_HBA_ERROR;
|
|
if (rc != MBX_BUSY)
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
return -EIO;
|
|
}
|
|
}
|
|
/* MBOX buffer will be freed in mbox compl */
|
|
pmb = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
|
|
if (!pmb) {
|
|
phba->link_state = LPFC_HBA_ERROR;
|
|
return -ENOMEM;
|
|
}
|
|
|
|
lpfc_config_async(phba, pmb, LPFC_ELS_RING);
|
|
pmb->mbox_cmpl = lpfc_config_async_cmpl;
|
|
pmb->vport = phba->pport;
|
|
rc = lpfc_sli_issue_mbox(phba, pmb, MBX_NOWAIT);
|
|
|
|
if ((rc != MBX_BUSY) && (rc != MBX_SUCCESS)) {
|
|
lpfc_printf_log(phba,
|
|
KERN_ERR,
|
|
LOG_INIT,
|
|
"0456 Adapter failed to issue "
|
|
"ASYNCEVT_ENABLE mbox status x%x\n",
|
|
rc);
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
}
|
|
|
|
/* Get Option rom version */
|
|
pmb = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
|
|
if (!pmb) {
|
|
phba->link_state = LPFC_HBA_ERROR;
|
|
return -ENOMEM;
|
|
}
|
|
|
|
lpfc_dump_wakeup_param(phba, pmb);
|
|
pmb->mbox_cmpl = lpfc_dump_wakeup_param_cmpl;
|
|
pmb->vport = phba->pport;
|
|
rc = lpfc_sli_issue_mbox(phba, pmb, MBX_NOWAIT);
|
|
|
|
if ((rc != MBX_BUSY) && (rc != MBX_SUCCESS)) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT, "0435 Adapter failed "
|
|
"to get Option ROM version status x%x\n", rc);
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_hba_init_link - Initialize the FC link
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @flag: mailbox command issue mode - either MBX_POLL or MBX_NOWAIT
|
|
*
|
|
* This routine will issue the INIT_LINK mailbox command call.
|
|
* It is available to other drivers through the lpfc_hba data
|
|
* structure for use as a delayed link up mechanism with the
|
|
* module parameter lpfc_suppress_link_up.
|
|
*
|
|
* Return code
|
|
* 0 - success
|
|
* Any other value - error
|
|
**/
|
|
int
|
|
lpfc_hba_init_link(struct lpfc_hba *phba, uint32_t flag)
|
|
{
|
|
struct lpfc_vport *vport = phba->pport;
|
|
LPFC_MBOXQ_t *pmb;
|
|
MAILBOX_t *mb;
|
|
int rc;
|
|
|
|
pmb = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
|
|
if (!pmb) {
|
|
phba->link_state = LPFC_HBA_ERROR;
|
|
return -ENOMEM;
|
|
}
|
|
mb = &pmb->u.mb;
|
|
pmb->vport = vport;
|
|
|
|
lpfc_init_link(phba, pmb, phba->cfg_topology,
|
|
phba->cfg_link_speed);
|
|
pmb->mbox_cmpl = lpfc_sli_def_mbox_cmpl;
|
|
lpfc_set_loopback_flag(phba);
|
|
rc = lpfc_sli_issue_mbox(phba, pmb, flag);
|
|
if (rc != MBX_SUCCESS) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0498 Adapter failed to init, mbxCmd x%x "
|
|
"INIT_LINK, mbxStatus x%x\n",
|
|
mb->mbxCommand, mb->mbxStatus);
|
|
/* Clear all interrupt enable conditions */
|
|
writel(0, phba->HCregaddr);
|
|
readl(phba->HCregaddr); /* flush */
|
|
/* Clear all pending interrupts */
|
|
writel(0xffffffff, phba->HAregaddr);
|
|
readl(phba->HAregaddr); /* flush */
|
|
phba->link_state = LPFC_HBA_ERROR;
|
|
if (rc != MBX_BUSY || flag == MBX_POLL)
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
return -EIO;
|
|
}
|
|
phba->cfg_suppress_link_up = LPFC_INITIALIZE_LINK;
|
|
if (flag == MBX_POLL)
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_hba_down_link - this routine downs the FC link
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @flag: mailbox command issue mode - either MBX_POLL or MBX_NOWAIT
|
|
*
|
|
* This routine will issue the DOWN_LINK mailbox command call.
|
|
* It is available to other drivers through the lpfc_hba data
|
|
* structure for use to stop the link.
|
|
*
|
|
* Return code
|
|
* 0 - success
|
|
* Any other value - error
|
|
**/
|
|
int
|
|
lpfc_hba_down_link(struct lpfc_hba *phba, uint32_t flag)
|
|
{
|
|
LPFC_MBOXQ_t *pmb;
|
|
int rc;
|
|
|
|
pmb = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
|
|
if (!pmb) {
|
|
phba->link_state = LPFC_HBA_ERROR;
|
|
return -ENOMEM;
|
|
}
|
|
|
|
lpfc_printf_log(phba,
|
|
KERN_ERR, LOG_INIT,
|
|
"0491 Adapter Link is disabled.\n");
|
|
lpfc_down_link(phba, pmb);
|
|
pmb->mbox_cmpl = lpfc_sli_def_mbox_cmpl;
|
|
rc = lpfc_sli_issue_mbox(phba, pmb, flag);
|
|
if ((rc != MBX_SUCCESS) && (rc != MBX_BUSY)) {
|
|
lpfc_printf_log(phba,
|
|
KERN_ERR, LOG_INIT,
|
|
"2522 Adapter failed to issue DOWN_LINK"
|
|
" mbox command rc 0x%x\n", rc);
|
|
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
return -EIO;
|
|
}
|
|
if (flag == MBX_POLL)
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_hba_down_prep - Perform lpfc uninitialization prior to HBA reset
|
|
* @phba: pointer to lpfc HBA data structure.
|
|
*
|
|
* This routine will do LPFC uninitialization before the HBA is reset when
|
|
* bringing down the SLI Layer.
|
|
*
|
|
* Return codes
|
|
* 0 - success.
|
|
* Any other value - error.
|
|
**/
|
|
int
|
|
lpfc_hba_down_prep(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_vport **vports;
|
|
int i;
|
|
|
|
if (phba->sli_rev <= LPFC_SLI_REV3) {
|
|
/* Disable interrupts */
|
|
writel(0, phba->HCregaddr);
|
|
readl(phba->HCregaddr); /* flush */
|
|
}
|
|
|
|
if (phba->pport->load_flag & FC_UNLOADING)
|
|
lpfc_cleanup_discovery_resources(phba->pport);
|
|
else {
|
|
vports = lpfc_create_vport_work_array(phba);
|
|
if (vports != NULL)
|
|
for (i = 0; i <= phba->max_vports &&
|
|
vports[i] != NULL; i++)
|
|
lpfc_cleanup_discovery_resources(vports[i]);
|
|
lpfc_destroy_vport_work_array(phba, vports);
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_hba_down_post_s3 - Perform lpfc uninitialization after HBA reset
|
|
* @phba: pointer to lpfc HBA data structure.
|
|
*
|
|
* This routine will do uninitialization after the HBA is reset when bring
|
|
* down the SLI Layer.
|
|
*
|
|
* Return codes
|
|
* 0 - success.
|
|
* Any other value - error.
|
|
**/
|
|
static int
|
|
lpfc_hba_down_post_s3(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_sli *psli = &phba->sli;
|
|
struct lpfc_sli_ring *pring;
|
|
struct lpfc_dmabuf *mp, *next_mp;
|
|
LIST_HEAD(completions);
|
|
int i;
|
|
|
|
if (phba->sli3_options & LPFC_SLI3_HBQ_ENABLED)
|
|
lpfc_sli_hbqbuf_free_all(phba);
|
|
else {
|
|
/* Cleanup preposted buffers on the ELS ring */
|
|
pring = &psli->ring[LPFC_ELS_RING];
|
|
list_for_each_entry_safe(mp, next_mp, &pring->postbufq, list) {
|
|
list_del(&mp->list);
|
|
pring->postbufq_cnt--;
|
|
lpfc_mbuf_free(phba, mp->virt, mp->phys);
|
|
kfree(mp);
|
|
}
|
|
}
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
for (i = 0; i < psli->num_rings; i++) {
|
|
pring = &psli->ring[i];
|
|
|
|
/* At this point in time the HBA is either reset or DOA. Either
|
|
* way, nothing should be on txcmplq as it will NEVER complete.
|
|
*/
|
|
list_splice_init(&pring->txcmplq, &completions);
|
|
pring->txcmplq_cnt = 0;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
/* Cancel all the IOCBs from the completions list */
|
|
lpfc_sli_cancel_iocbs(phba, &completions, IOSTAT_LOCAL_REJECT,
|
|
IOERR_SLI_ABORTED);
|
|
|
|
lpfc_sli_abort_iocb_ring(phba, pring);
|
|
spin_lock_irq(&phba->hbalock);
|
|
}
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_hba_down_post_s4 - Perform lpfc uninitialization after HBA reset
|
|
* @phba: pointer to lpfc HBA data structure.
|
|
*
|
|
* This routine will do uninitialization after the HBA is reset when bring
|
|
* down the SLI Layer.
|
|
*
|
|
* Return codes
|
|
* 0 - success.
|
|
* Any other value - error.
|
|
**/
|
|
static int
|
|
lpfc_hba_down_post_s4(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_scsi_buf *psb, *psb_next;
|
|
LIST_HEAD(aborts);
|
|
int ret;
|
|
unsigned long iflag = 0;
|
|
struct lpfc_sglq *sglq_entry = NULL;
|
|
|
|
ret = lpfc_hba_down_post_s3(phba);
|
|
if (ret)
|
|
return ret;
|
|
/* At this point in time the HBA is either reset or DOA. Either
|
|
* way, nothing should be on lpfc_abts_els_sgl_list, it needs to be
|
|
* on the lpfc_sgl_list so that it can either be freed if the
|
|
* driver is unloading or reposted if the driver is restarting
|
|
* the port.
|
|
*/
|
|
spin_lock_irq(&phba->hbalock); /* required for lpfc_sgl_list and */
|
|
/* scsl_buf_list */
|
|
/* abts_sgl_list_lock required because worker thread uses this
|
|
* list.
|
|
*/
|
|
spin_lock(&phba->sli4_hba.abts_sgl_list_lock);
|
|
list_for_each_entry(sglq_entry,
|
|
&phba->sli4_hba.lpfc_abts_els_sgl_list, list)
|
|
sglq_entry->state = SGL_FREED;
|
|
|
|
list_splice_init(&phba->sli4_hba.lpfc_abts_els_sgl_list,
|
|
&phba->sli4_hba.lpfc_sgl_list);
|
|
spin_unlock(&phba->sli4_hba.abts_sgl_list_lock);
|
|
/* abts_scsi_buf_list_lock required because worker thread uses this
|
|
* list.
|
|
*/
|
|
spin_lock(&phba->sli4_hba.abts_scsi_buf_list_lock);
|
|
list_splice_init(&phba->sli4_hba.lpfc_abts_scsi_buf_list,
|
|
&aborts);
|
|
spin_unlock(&phba->sli4_hba.abts_scsi_buf_list_lock);
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
list_for_each_entry_safe(psb, psb_next, &aborts, list) {
|
|
psb->pCmd = NULL;
|
|
psb->status = IOSTAT_SUCCESS;
|
|
}
|
|
spin_lock_irqsave(&phba->scsi_buf_list_lock, iflag);
|
|
list_splice(&aborts, &phba->lpfc_scsi_buf_list);
|
|
spin_unlock_irqrestore(&phba->scsi_buf_list_lock, iflag);
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_hba_down_post - Wrapper func for hba down post routine
|
|
* @phba: pointer to lpfc HBA data structure.
|
|
*
|
|
* This routine wraps the actual SLI3 or SLI4 routine for performing
|
|
* uninitialization after the HBA is reset when bring down the SLI Layer.
|
|
*
|
|
* Return codes
|
|
* 0 - success.
|
|
* Any other value - error.
|
|
**/
|
|
int
|
|
lpfc_hba_down_post(struct lpfc_hba *phba)
|
|
{
|
|
return (*phba->lpfc_hba_down_post)(phba);
|
|
}
|
|
|
|
/**
|
|
* lpfc_hb_timeout - The HBA-timer timeout handler
|
|
* @ptr: unsigned long holds the pointer to lpfc hba data structure.
|
|
*
|
|
* This is the HBA-timer timeout handler registered to the lpfc driver. When
|
|
* this timer fires, a HBA timeout event shall be posted to the lpfc driver
|
|
* work-port-events bitmap and the worker thread is notified. This timeout
|
|
* event will be used by the worker thread to invoke the actual timeout
|
|
* handler routine, lpfc_hb_timeout_handler. Any periodical operations will
|
|
* be performed in the timeout handler and the HBA timeout event bit shall
|
|
* be cleared by the worker thread after it has taken the event bitmap out.
|
|
**/
|
|
static void
|
|
lpfc_hb_timeout(unsigned long ptr)
|
|
{
|
|
struct lpfc_hba *phba;
|
|
uint32_t tmo_posted;
|
|
unsigned long iflag;
|
|
|
|
phba = (struct lpfc_hba *)ptr;
|
|
|
|
/* Check for heart beat timeout conditions */
|
|
spin_lock_irqsave(&phba->pport->work_port_lock, iflag);
|
|
tmo_posted = phba->pport->work_port_events & WORKER_HB_TMO;
|
|
if (!tmo_posted)
|
|
phba->pport->work_port_events |= WORKER_HB_TMO;
|
|
spin_unlock_irqrestore(&phba->pport->work_port_lock, iflag);
|
|
|
|
/* Tell the worker thread there is work to do */
|
|
if (!tmo_posted)
|
|
lpfc_worker_wake_up(phba);
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_hb_mbox_cmpl - The lpfc heart-beat mailbox command callback function
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @pmboxq: pointer to the driver internal queue element for mailbox command.
|
|
*
|
|
* This is the callback function to the lpfc heart-beat mailbox command.
|
|
* If configured, the lpfc driver issues the heart-beat mailbox command to
|
|
* the HBA every LPFC_HB_MBOX_INTERVAL (current 5) seconds. At the time the
|
|
* heart-beat mailbox command is issued, the driver shall set up heart-beat
|
|
* timeout timer to LPFC_HB_MBOX_TIMEOUT (current 30) seconds and marks
|
|
* heart-beat outstanding state. Once the mailbox command comes back and
|
|
* no error conditions detected, the heart-beat mailbox command timer is
|
|
* reset to LPFC_HB_MBOX_INTERVAL seconds and the heart-beat outstanding
|
|
* state is cleared for the next heart-beat. If the timer expired with the
|
|
* heart-beat outstanding state set, the driver will put the HBA offline.
|
|
**/
|
|
static void
|
|
lpfc_hb_mbox_cmpl(struct lpfc_hba * phba, LPFC_MBOXQ_t * pmboxq)
|
|
{
|
|
unsigned long drvr_flag;
|
|
|
|
spin_lock_irqsave(&phba->hbalock, drvr_flag);
|
|
phba->hb_outstanding = 0;
|
|
spin_unlock_irqrestore(&phba->hbalock, drvr_flag);
|
|
|
|
/* Check and reset heart-beat timer is necessary */
|
|
mempool_free(pmboxq, phba->mbox_mem_pool);
|
|
if (!(phba->pport->fc_flag & FC_OFFLINE_MODE) &&
|
|
!(phba->link_state == LPFC_HBA_ERROR) &&
|
|
!(phba->pport->load_flag & FC_UNLOADING))
|
|
mod_timer(&phba->hb_tmofunc,
|
|
jiffies + HZ * LPFC_HB_MBOX_INTERVAL);
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_hb_timeout_handler - The HBA-timer timeout handler
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This is the actual HBA-timer timeout handler to be invoked by the worker
|
|
* thread whenever the HBA timer fired and HBA-timeout event posted. This
|
|
* handler performs any periodic operations needed for the device. If such
|
|
* periodic event has already been attended to either in the interrupt handler
|
|
* or by processing slow-ring or fast-ring events within the HBA-timer
|
|
* timeout window (LPFC_HB_MBOX_INTERVAL), this handler just simply resets
|
|
* the timer for the next timeout period. If lpfc heart-beat mailbox command
|
|
* is configured and there is no heart-beat mailbox command outstanding, a
|
|
* heart-beat mailbox is issued and timer set properly. Otherwise, if there
|
|
* has been a heart-beat mailbox command outstanding, the HBA shall be put
|
|
* to offline.
|
|
**/
|
|
void
|
|
lpfc_hb_timeout_handler(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_vport **vports;
|
|
LPFC_MBOXQ_t *pmboxq;
|
|
struct lpfc_dmabuf *buf_ptr;
|
|
int retval, i;
|
|
struct lpfc_sli *psli = &phba->sli;
|
|
LIST_HEAD(completions);
|
|
|
|
vports = lpfc_create_vport_work_array(phba);
|
|
if (vports != NULL)
|
|
for (i = 0; i <= phba->max_vports && vports[i] != NULL; i++)
|
|
lpfc_rcv_seq_check_edtov(vports[i]);
|
|
lpfc_destroy_vport_work_array(phba, vports);
|
|
|
|
if ((phba->link_state == LPFC_HBA_ERROR) ||
|
|
(phba->pport->load_flag & FC_UNLOADING) ||
|
|
(phba->pport->fc_flag & FC_OFFLINE_MODE))
|
|
return;
|
|
|
|
spin_lock_irq(&phba->pport->work_port_lock);
|
|
|
|
if (time_after(phba->last_completion_time + LPFC_HB_MBOX_INTERVAL * HZ,
|
|
jiffies)) {
|
|
spin_unlock_irq(&phba->pport->work_port_lock);
|
|
if (!phba->hb_outstanding)
|
|
mod_timer(&phba->hb_tmofunc,
|
|
jiffies + HZ * LPFC_HB_MBOX_INTERVAL);
|
|
else
|
|
mod_timer(&phba->hb_tmofunc,
|
|
jiffies + HZ * LPFC_HB_MBOX_TIMEOUT);
|
|
return;
|
|
}
|
|
spin_unlock_irq(&phba->pport->work_port_lock);
|
|
|
|
if (phba->elsbuf_cnt &&
|
|
(phba->elsbuf_cnt == phba->elsbuf_prev_cnt)) {
|
|
spin_lock_irq(&phba->hbalock);
|
|
list_splice_init(&phba->elsbuf, &completions);
|
|
phba->elsbuf_cnt = 0;
|
|
phba->elsbuf_prev_cnt = 0;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
while (!list_empty(&completions)) {
|
|
list_remove_head(&completions, buf_ptr,
|
|
struct lpfc_dmabuf, list);
|
|
lpfc_mbuf_free(phba, buf_ptr->virt, buf_ptr->phys);
|
|
kfree(buf_ptr);
|
|
}
|
|
}
|
|
phba->elsbuf_prev_cnt = phba->elsbuf_cnt;
|
|
|
|
/* If there is no heart beat outstanding, issue a heartbeat command */
|
|
if (phba->cfg_enable_hba_heartbeat) {
|
|
if (!phba->hb_outstanding) {
|
|
if ((!(psli->sli_flag & LPFC_SLI_MBOX_ACTIVE)) &&
|
|
(list_empty(&psli->mboxq))) {
|
|
pmboxq = mempool_alloc(phba->mbox_mem_pool,
|
|
GFP_KERNEL);
|
|
if (!pmboxq) {
|
|
mod_timer(&phba->hb_tmofunc,
|
|
jiffies +
|
|
HZ * LPFC_HB_MBOX_INTERVAL);
|
|
return;
|
|
}
|
|
|
|
lpfc_heart_beat(phba, pmboxq);
|
|
pmboxq->mbox_cmpl = lpfc_hb_mbox_cmpl;
|
|
pmboxq->vport = phba->pport;
|
|
retval = lpfc_sli_issue_mbox(phba, pmboxq,
|
|
MBX_NOWAIT);
|
|
|
|
if (retval != MBX_BUSY &&
|
|
retval != MBX_SUCCESS) {
|
|
mempool_free(pmboxq,
|
|
phba->mbox_mem_pool);
|
|
mod_timer(&phba->hb_tmofunc,
|
|
jiffies +
|
|
HZ * LPFC_HB_MBOX_INTERVAL);
|
|
return;
|
|
}
|
|
phba->skipped_hb = 0;
|
|
phba->hb_outstanding = 1;
|
|
} else if (time_before_eq(phba->last_completion_time,
|
|
phba->skipped_hb)) {
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"2857 Last completion time not "
|
|
" updated in %d ms\n",
|
|
jiffies_to_msecs(jiffies
|
|
- phba->last_completion_time));
|
|
} else
|
|
phba->skipped_hb = jiffies;
|
|
|
|
mod_timer(&phba->hb_tmofunc,
|
|
jiffies + HZ * LPFC_HB_MBOX_TIMEOUT);
|
|
return;
|
|
} else {
|
|
/*
|
|
* If heart beat timeout called with hb_outstanding set
|
|
* we need to give the hb mailbox cmd a chance to
|
|
* complete or TMO.
|
|
*/
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_INIT,
|
|
"0459 Adapter heartbeat still out"
|
|
"standing:last compl time was %d ms.\n",
|
|
jiffies_to_msecs(jiffies
|
|
- phba->last_completion_time));
|
|
mod_timer(&phba->hb_tmofunc,
|
|
jiffies + HZ * LPFC_HB_MBOX_TIMEOUT);
|
|
}
|
|
}
|
|
}
|
|
|
|
/**
|
|
* lpfc_offline_eratt - Bring lpfc offline on hardware error attention
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is called to bring the HBA offline when HBA hardware error
|
|
* other than Port Error 6 has been detected.
|
|
**/
|
|
static void
|
|
lpfc_offline_eratt(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_sli *psli = &phba->sli;
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
psli->sli_flag &= ~LPFC_SLI_ACTIVE;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
lpfc_offline_prep(phba);
|
|
|
|
lpfc_offline(phba);
|
|
lpfc_reset_barrier(phba);
|
|
spin_lock_irq(&phba->hbalock);
|
|
lpfc_sli_brdreset(phba);
|
|
spin_unlock_irq(&phba->hbalock);
|
|
lpfc_hba_down_post(phba);
|
|
lpfc_sli_brdready(phba, HS_MBRDY);
|
|
lpfc_unblock_mgmt_io(phba);
|
|
phba->link_state = LPFC_HBA_ERROR;
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_offline_eratt - Bring lpfc offline on SLI4 hardware error attention
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is called to bring a SLI4 HBA offline when HBA hardware error
|
|
* other than Port Error 6 has been detected.
|
|
**/
|
|
static void
|
|
lpfc_sli4_offline_eratt(struct lpfc_hba *phba)
|
|
{
|
|
lpfc_offline_prep(phba);
|
|
lpfc_offline(phba);
|
|
lpfc_sli4_brdreset(phba);
|
|
lpfc_hba_down_post(phba);
|
|
lpfc_sli4_post_status_check(phba);
|
|
lpfc_unblock_mgmt_io(phba);
|
|
phba->link_state = LPFC_HBA_ERROR;
|
|
}
|
|
|
|
/**
|
|
* lpfc_handle_deferred_eratt - The HBA hardware deferred error handler
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to handle the deferred HBA hardware error
|
|
* conditions. This type of error is indicated by HBA by setting ER1
|
|
* and another ER bit in the host status register. The driver will
|
|
* wait until the ER1 bit clears before handling the error condition.
|
|
**/
|
|
static void
|
|
lpfc_handle_deferred_eratt(struct lpfc_hba *phba)
|
|
{
|
|
uint32_t old_host_status = phba->work_hs;
|
|
struct lpfc_sli_ring *pring;
|
|
struct lpfc_sli *psli = &phba->sli;
|
|
|
|
/* If the pci channel is offline, ignore possible errors,
|
|
* since we cannot communicate with the pci card anyway.
|
|
*/
|
|
if (pci_channel_offline(phba->pcidev)) {
|
|
spin_lock_irq(&phba->hbalock);
|
|
phba->hba_flag &= ~DEFER_ERATT;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
return;
|
|
}
|
|
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0479 Deferred Adapter Hardware Error "
|
|
"Data: x%x x%x x%x\n",
|
|
phba->work_hs,
|
|
phba->work_status[0], phba->work_status[1]);
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
psli->sli_flag &= ~LPFC_SLI_ACTIVE;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
|
|
/*
|
|
* Firmware stops when it triggred erratt. That could cause the I/Os
|
|
* dropped by the firmware. Error iocb (I/O) on txcmplq and let the
|
|
* SCSI layer retry it after re-establishing link.
|
|
*/
|
|
pring = &psli->ring[psli->fcp_ring];
|
|
lpfc_sli_abort_iocb_ring(phba, pring);
|
|
|
|
/*
|
|
* There was a firmware error. Take the hba offline and then
|
|
* attempt to restart it.
|
|
*/
|
|
lpfc_offline_prep(phba);
|
|
lpfc_offline(phba);
|
|
|
|
/* Wait for the ER1 bit to clear.*/
|
|
while (phba->work_hs & HS_FFER1) {
|
|
msleep(100);
|
|
phba->work_hs = readl(phba->HSregaddr);
|
|
/* If driver is unloading let the worker thread continue */
|
|
if (phba->pport->load_flag & FC_UNLOADING) {
|
|
phba->work_hs = 0;
|
|
break;
|
|
}
|
|
}
|
|
|
|
/*
|
|
* This is to ptrotect against a race condition in which
|
|
* first write to the host attention register clear the
|
|
* host status register.
|
|
*/
|
|
if ((!phba->work_hs) && (!(phba->pport->load_flag & FC_UNLOADING)))
|
|
phba->work_hs = old_host_status & ~HS_FFER1;
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
phba->hba_flag &= ~DEFER_ERATT;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
phba->work_status[0] = readl(phba->MBslimaddr + 0xa8);
|
|
phba->work_status[1] = readl(phba->MBslimaddr + 0xac);
|
|
}
|
|
|
|
static void
|
|
lpfc_board_errevt_to_mgmt(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_board_event_header board_event;
|
|
struct Scsi_Host *shost;
|
|
|
|
board_event.event_type = FC_REG_BOARD_EVENT;
|
|
board_event.subcategory = LPFC_EVENT_PORTINTERR;
|
|
shost = lpfc_shost_from_vport(phba->pport);
|
|
fc_host_post_vendor_event(shost, fc_get_event_number(),
|
|
sizeof(board_event),
|
|
(char *) &board_event,
|
|
LPFC_NL_VENDOR_ID);
|
|
}
|
|
|
|
/**
|
|
* lpfc_handle_eratt_s3 - The SLI3 HBA hardware error handler
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to handle the following HBA hardware error
|
|
* conditions:
|
|
* 1 - HBA error attention interrupt
|
|
* 2 - DMA ring index out of range
|
|
* 3 - Mailbox command came back as unknown
|
|
**/
|
|
static void
|
|
lpfc_handle_eratt_s3(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_vport *vport = phba->pport;
|
|
struct lpfc_sli *psli = &phba->sli;
|
|
struct lpfc_sli_ring *pring;
|
|
uint32_t event_data;
|
|
unsigned long temperature;
|
|
struct temp_event temp_event_data;
|
|
struct Scsi_Host *shost;
|
|
|
|
/* If the pci channel is offline, ignore possible errors,
|
|
* since we cannot communicate with the pci card anyway.
|
|
*/
|
|
if (pci_channel_offline(phba->pcidev)) {
|
|
spin_lock_irq(&phba->hbalock);
|
|
phba->hba_flag &= ~DEFER_ERATT;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
return;
|
|
}
|
|
|
|
/* If resets are disabled then leave the HBA alone and return */
|
|
if (!phba->cfg_enable_hba_reset)
|
|
return;
|
|
|
|
/* Send an internal error event to mgmt application */
|
|
lpfc_board_errevt_to_mgmt(phba);
|
|
|
|
if (phba->hba_flag & DEFER_ERATT)
|
|
lpfc_handle_deferred_eratt(phba);
|
|
|
|
if ((phba->work_hs & HS_FFER6) || (phba->work_hs & HS_FFER8)) {
|
|
if (phba->work_hs & HS_FFER6)
|
|
/* Re-establishing Link */
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_LINK_EVENT,
|
|
"1301 Re-establishing Link "
|
|
"Data: x%x x%x x%x\n",
|
|
phba->work_hs, phba->work_status[0],
|
|
phba->work_status[1]);
|
|
if (phba->work_hs & HS_FFER8)
|
|
/* Device Zeroization */
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_LINK_EVENT,
|
|
"2861 Host Authentication device "
|
|
"zeroization Data:x%x x%x x%x\n",
|
|
phba->work_hs, phba->work_status[0],
|
|
phba->work_status[1]);
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
psli->sli_flag &= ~LPFC_SLI_ACTIVE;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
/*
|
|
* Firmware stops when it triggled erratt with HS_FFER6.
|
|
* That could cause the I/Os dropped by the firmware.
|
|
* Error iocb (I/O) on txcmplq and let the SCSI layer
|
|
* retry it after re-establishing link.
|
|
*/
|
|
pring = &psli->ring[psli->fcp_ring];
|
|
lpfc_sli_abort_iocb_ring(phba, pring);
|
|
|
|
/*
|
|
* There was a firmware error. Take the hba offline and then
|
|
* attempt to restart it.
|
|
*/
|
|
lpfc_offline_prep(phba);
|
|
lpfc_offline(phba);
|
|
lpfc_sli_brdrestart(phba);
|
|
if (lpfc_online(phba) == 0) { /* Initialize the HBA */
|
|
lpfc_unblock_mgmt_io(phba);
|
|
return;
|
|
}
|
|
lpfc_unblock_mgmt_io(phba);
|
|
} else if (phba->work_hs & HS_CRIT_TEMP) {
|
|
temperature = readl(phba->MBslimaddr + TEMPERATURE_OFFSET);
|
|
temp_event_data.event_type = FC_REG_TEMPERATURE_EVENT;
|
|
temp_event_data.event_code = LPFC_CRIT_TEMP;
|
|
temp_event_data.data = (uint32_t)temperature;
|
|
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0406 Adapter maximum temperature exceeded "
|
|
"(%ld), taking this port offline "
|
|
"Data: x%x x%x x%x\n",
|
|
temperature, phba->work_hs,
|
|
phba->work_status[0], phba->work_status[1]);
|
|
|
|
shost = lpfc_shost_from_vport(phba->pport);
|
|
fc_host_post_vendor_event(shost, fc_get_event_number(),
|
|
sizeof(temp_event_data),
|
|
(char *) &temp_event_data,
|
|
SCSI_NL_VID_TYPE_PCI
|
|
| PCI_VENDOR_ID_EMULEX);
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
phba->over_temp_state = HBA_OVER_TEMP;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
lpfc_offline_eratt(phba);
|
|
|
|
} else {
|
|
/* The if clause above forces this code path when the status
|
|
* failure is a value other than FFER6. Do not call the offline
|
|
* twice. This is the adapter hardware error path.
|
|
*/
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0457 Adapter Hardware Error "
|
|
"Data: x%x x%x x%x\n",
|
|
phba->work_hs,
|
|
phba->work_status[0], phba->work_status[1]);
|
|
|
|
event_data = FC_REG_DUMP_EVENT;
|
|
shost = lpfc_shost_from_vport(vport);
|
|
fc_host_post_vendor_event(shost, fc_get_event_number(),
|
|
sizeof(event_data), (char *) &event_data,
|
|
SCSI_NL_VID_TYPE_PCI | PCI_VENDOR_ID_EMULEX);
|
|
|
|
lpfc_offline_eratt(phba);
|
|
}
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_handle_eratt_s4 - The SLI4 HBA hardware error handler
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to handle the SLI4 HBA hardware error attention
|
|
* conditions.
|
|
**/
|
|
static void
|
|
lpfc_handle_eratt_s4(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_vport *vport = phba->pport;
|
|
uint32_t event_data;
|
|
struct Scsi_Host *shost;
|
|
|
|
/* If the pci channel is offline, ignore possible errors, since
|
|
* we cannot communicate with the pci card anyway.
|
|
*/
|
|
if (pci_channel_offline(phba->pcidev))
|
|
return;
|
|
/* If resets are disabled then leave the HBA alone and return */
|
|
if (!phba->cfg_enable_hba_reset)
|
|
return;
|
|
|
|
/* Send an internal error event to mgmt application */
|
|
lpfc_board_errevt_to_mgmt(phba);
|
|
|
|
/* For now, the actual action for SLI4 device handling is not
|
|
* specified yet, just treated it as adaptor hardware failure
|
|
*/
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0143 SLI4 Adapter Hardware Error Data: x%x x%x\n",
|
|
phba->work_status[0], phba->work_status[1]);
|
|
|
|
event_data = FC_REG_DUMP_EVENT;
|
|
shost = lpfc_shost_from_vport(vport);
|
|
fc_host_post_vendor_event(shost, fc_get_event_number(),
|
|
sizeof(event_data), (char *) &event_data,
|
|
SCSI_NL_VID_TYPE_PCI | PCI_VENDOR_ID_EMULEX);
|
|
|
|
lpfc_sli4_offline_eratt(phba);
|
|
}
|
|
|
|
/**
|
|
* lpfc_handle_eratt - Wrapper func for handling hba error attention
|
|
* @phba: pointer to lpfc HBA data structure.
|
|
*
|
|
* This routine wraps the actual SLI3 or SLI4 hba error attention handling
|
|
* routine from the API jump table function pointer from the lpfc_hba struct.
|
|
*
|
|
* Return codes
|
|
* 0 - success.
|
|
* Any other value - error.
|
|
**/
|
|
void
|
|
lpfc_handle_eratt(struct lpfc_hba *phba)
|
|
{
|
|
(*phba->lpfc_handle_eratt)(phba);
|
|
}
|
|
|
|
/**
|
|
* lpfc_handle_latt - The HBA link event handler
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked from the worker thread to handle a HBA host
|
|
* attention link event.
|
|
**/
|
|
void
|
|
lpfc_handle_latt(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_vport *vport = phba->pport;
|
|
struct lpfc_sli *psli = &phba->sli;
|
|
LPFC_MBOXQ_t *pmb;
|
|
volatile uint32_t control;
|
|
struct lpfc_dmabuf *mp;
|
|
int rc = 0;
|
|
|
|
pmb = (LPFC_MBOXQ_t *)mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
|
|
if (!pmb) {
|
|
rc = 1;
|
|
goto lpfc_handle_latt_err_exit;
|
|
}
|
|
|
|
mp = kmalloc(sizeof(struct lpfc_dmabuf), GFP_KERNEL);
|
|
if (!mp) {
|
|
rc = 2;
|
|
goto lpfc_handle_latt_free_pmb;
|
|
}
|
|
|
|
mp->virt = lpfc_mbuf_alloc(phba, 0, &mp->phys);
|
|
if (!mp->virt) {
|
|
rc = 3;
|
|
goto lpfc_handle_latt_free_mp;
|
|
}
|
|
|
|
/* Cleanup any outstanding ELS commands */
|
|
lpfc_els_flush_all_cmd(phba);
|
|
|
|
psli->slistat.link_event++;
|
|
lpfc_read_la(phba, pmb, mp);
|
|
pmb->mbox_cmpl = lpfc_mbx_cmpl_read_la;
|
|
pmb->vport = vport;
|
|
/* Block ELS IOCBs until we have processed this mbox command */
|
|
phba->sli.ring[LPFC_ELS_RING].flag |= LPFC_STOP_IOCB_EVENT;
|
|
rc = lpfc_sli_issue_mbox (phba, pmb, MBX_NOWAIT);
|
|
if (rc == MBX_NOT_FINISHED) {
|
|
rc = 4;
|
|
goto lpfc_handle_latt_free_mbuf;
|
|
}
|
|
|
|
/* Clear Link Attention in HA REG */
|
|
spin_lock_irq(&phba->hbalock);
|
|
writel(HA_LATT, phba->HAregaddr);
|
|
readl(phba->HAregaddr); /* flush */
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
return;
|
|
|
|
lpfc_handle_latt_free_mbuf:
|
|
phba->sli.ring[LPFC_ELS_RING].flag &= ~LPFC_STOP_IOCB_EVENT;
|
|
lpfc_mbuf_free(phba, mp->virt, mp->phys);
|
|
lpfc_handle_latt_free_mp:
|
|
kfree(mp);
|
|
lpfc_handle_latt_free_pmb:
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
lpfc_handle_latt_err_exit:
|
|
/* Enable Link attention interrupts */
|
|
spin_lock_irq(&phba->hbalock);
|
|
psli->sli_flag |= LPFC_PROCESS_LA;
|
|
control = readl(phba->HCregaddr);
|
|
control |= HC_LAINT_ENA;
|
|
writel(control, phba->HCregaddr);
|
|
readl(phba->HCregaddr); /* flush */
|
|
|
|
/* Clear Link Attention in HA REG */
|
|
writel(HA_LATT, phba->HAregaddr);
|
|
readl(phba->HAregaddr); /* flush */
|
|
spin_unlock_irq(&phba->hbalock);
|
|
lpfc_linkdown(phba);
|
|
phba->link_state = LPFC_HBA_ERROR;
|
|
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_MBOX,
|
|
"0300 LATT: Cannot issue READ_LA: Data:%d\n", rc);
|
|
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_parse_vpd - Parse VPD (Vital Product Data)
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @vpd: pointer to the vital product data.
|
|
* @len: length of the vital product data in bytes.
|
|
*
|
|
* This routine parses the Vital Product Data (VPD). The VPD is treated as
|
|
* an array of characters. In this routine, the ModelName, ProgramType, and
|
|
* ModelDesc, etc. fields of the phba data structure will be populated.
|
|
*
|
|
* Return codes
|
|
* 0 - pointer to the VPD passed in is NULL
|
|
* 1 - success
|
|
**/
|
|
int
|
|
lpfc_parse_vpd(struct lpfc_hba *phba, uint8_t *vpd, int len)
|
|
{
|
|
uint8_t lenlo, lenhi;
|
|
int Length;
|
|
int i, j;
|
|
int finished = 0;
|
|
int index = 0;
|
|
|
|
if (!vpd)
|
|
return 0;
|
|
|
|
/* Vital Product */
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0455 Vital Product Data: x%x x%x x%x x%x\n",
|
|
(uint32_t) vpd[0], (uint32_t) vpd[1], (uint32_t) vpd[2],
|
|
(uint32_t) vpd[3]);
|
|
while (!finished && (index < (len - 4))) {
|
|
switch (vpd[index]) {
|
|
case 0x82:
|
|
case 0x91:
|
|
index += 1;
|
|
lenlo = vpd[index];
|
|
index += 1;
|
|
lenhi = vpd[index];
|
|
index += 1;
|
|
i = ((((unsigned short)lenhi) << 8) + lenlo);
|
|
index += i;
|
|
break;
|
|
case 0x90:
|
|
index += 1;
|
|
lenlo = vpd[index];
|
|
index += 1;
|
|
lenhi = vpd[index];
|
|
index += 1;
|
|
Length = ((((unsigned short)lenhi) << 8) + lenlo);
|
|
if (Length > len - index)
|
|
Length = len - index;
|
|
while (Length > 0) {
|
|
/* Look for Serial Number */
|
|
if ((vpd[index] == 'S') && (vpd[index+1] == 'N')) {
|
|
index += 2;
|
|
i = vpd[index];
|
|
index += 1;
|
|
j = 0;
|
|
Length -= (3+i);
|
|
while(i--) {
|
|
phba->SerialNumber[j++] = vpd[index++];
|
|
if (j == 31)
|
|
break;
|
|
}
|
|
phba->SerialNumber[j] = 0;
|
|
continue;
|
|
}
|
|
else if ((vpd[index] == 'V') && (vpd[index+1] == '1')) {
|
|
phba->vpd_flag |= VPD_MODEL_DESC;
|
|
index += 2;
|
|
i = vpd[index];
|
|
index += 1;
|
|
j = 0;
|
|
Length -= (3+i);
|
|
while(i--) {
|
|
phba->ModelDesc[j++] = vpd[index++];
|
|
if (j == 255)
|
|
break;
|
|
}
|
|
phba->ModelDesc[j] = 0;
|
|
continue;
|
|
}
|
|
else if ((vpd[index] == 'V') && (vpd[index+1] == '2')) {
|
|
phba->vpd_flag |= VPD_MODEL_NAME;
|
|
index += 2;
|
|
i = vpd[index];
|
|
index += 1;
|
|
j = 0;
|
|
Length -= (3+i);
|
|
while(i--) {
|
|
phba->ModelName[j++] = vpd[index++];
|
|
if (j == 79)
|
|
break;
|
|
}
|
|
phba->ModelName[j] = 0;
|
|
continue;
|
|
}
|
|
else if ((vpd[index] == 'V') && (vpd[index+1] == '3')) {
|
|
phba->vpd_flag |= VPD_PROGRAM_TYPE;
|
|
index += 2;
|
|
i = vpd[index];
|
|
index += 1;
|
|
j = 0;
|
|
Length -= (3+i);
|
|
while(i--) {
|
|
phba->ProgramType[j++] = vpd[index++];
|
|
if (j == 255)
|
|
break;
|
|
}
|
|
phba->ProgramType[j] = 0;
|
|
continue;
|
|
}
|
|
else if ((vpd[index] == 'V') && (vpd[index+1] == '4')) {
|
|
phba->vpd_flag |= VPD_PORT;
|
|
index += 2;
|
|
i = vpd[index];
|
|
index += 1;
|
|
j = 0;
|
|
Length -= (3+i);
|
|
while(i--) {
|
|
phba->Port[j++] = vpd[index++];
|
|
if (j == 19)
|
|
break;
|
|
}
|
|
phba->Port[j] = 0;
|
|
continue;
|
|
}
|
|
else {
|
|
index += 2;
|
|
i = vpd[index];
|
|
index += 1;
|
|
index += i;
|
|
Length -= (3 + i);
|
|
}
|
|
}
|
|
finished = 0;
|
|
break;
|
|
case 0x78:
|
|
finished = 1;
|
|
break;
|
|
default:
|
|
index ++;
|
|
break;
|
|
}
|
|
}
|
|
|
|
return(1);
|
|
}
|
|
|
|
/**
|
|
* lpfc_get_hba_model_desc - Retrieve HBA device model name and description
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @mdp: pointer to the data structure to hold the derived model name.
|
|
* @descp: pointer to the data structure to hold the derived description.
|
|
*
|
|
* This routine retrieves HBA's description based on its registered PCI device
|
|
* ID. The @descp passed into this function points to an array of 256 chars. It
|
|
* shall be returned with the model name, maximum speed, and the host bus type.
|
|
* The @mdp passed into this function points to an array of 80 chars. When the
|
|
* function returns, the @mdp will be filled with the model name.
|
|
**/
|
|
static void
|
|
lpfc_get_hba_model_desc(struct lpfc_hba *phba, uint8_t *mdp, uint8_t *descp)
|
|
{
|
|
lpfc_vpd_t *vp;
|
|
uint16_t dev_id = phba->pcidev->device;
|
|
int max_speed;
|
|
int GE = 0;
|
|
int oneConnect = 0; /* default is not a oneConnect */
|
|
struct {
|
|
char *name;
|
|
char *bus;
|
|
char *function;
|
|
} m = {"<Unknown>", "", ""};
|
|
|
|
if (mdp && mdp[0] != '\0'
|
|
&& descp && descp[0] != '\0')
|
|
return;
|
|
|
|
if (phba->lmt & LMT_10Gb)
|
|
max_speed = 10;
|
|
else if (phba->lmt & LMT_8Gb)
|
|
max_speed = 8;
|
|
else if (phba->lmt & LMT_4Gb)
|
|
max_speed = 4;
|
|
else if (phba->lmt & LMT_2Gb)
|
|
max_speed = 2;
|
|
else
|
|
max_speed = 1;
|
|
|
|
vp = &phba->vpd;
|
|
|
|
switch (dev_id) {
|
|
case PCI_DEVICE_ID_FIREFLY:
|
|
m = (typeof(m)){"LP6000", "PCI", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_SUPERFLY:
|
|
if (vp->rev.biuRev >= 1 && vp->rev.biuRev <= 3)
|
|
m = (typeof(m)){"LP7000", "PCI",
|
|
"Fibre Channel Adapter"};
|
|
else
|
|
m = (typeof(m)){"LP7000E", "PCI",
|
|
"Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_DRAGONFLY:
|
|
m = (typeof(m)){"LP8000", "PCI",
|
|
"Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_CENTAUR:
|
|
if (FC_JEDEC_ID(vp->rev.biuRev) == CENTAUR_2G_JEDEC_ID)
|
|
m = (typeof(m)){"LP9002", "PCI",
|
|
"Fibre Channel Adapter"};
|
|
else
|
|
m = (typeof(m)){"LP9000", "PCI",
|
|
"Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_RFLY:
|
|
m = (typeof(m)){"LP952", "PCI",
|
|
"Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_PEGASUS:
|
|
m = (typeof(m)){"LP9802", "PCI-X",
|
|
"Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_THOR:
|
|
m = (typeof(m)){"LP10000", "PCI-X",
|
|
"Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_VIPER:
|
|
m = (typeof(m)){"LPX1000", "PCI-X",
|
|
"Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_PFLY:
|
|
m = (typeof(m)){"LP982", "PCI-X",
|
|
"Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_TFLY:
|
|
m = (typeof(m)){"LP1050", "PCI-X",
|
|
"Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_HELIOS:
|
|
m = (typeof(m)){"LP11000", "PCI-X2",
|
|
"Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_HELIOS_SCSP:
|
|
m = (typeof(m)){"LP11000-SP", "PCI-X2",
|
|
"Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_HELIOS_DCSP:
|
|
m = (typeof(m)){"LP11002-SP", "PCI-X2",
|
|
"Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_NEPTUNE:
|
|
m = (typeof(m)){"LPe1000", "PCIe", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_NEPTUNE_SCSP:
|
|
m = (typeof(m)){"LPe1000-SP", "PCIe", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_NEPTUNE_DCSP:
|
|
m = (typeof(m)){"LPe1002-SP", "PCIe", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_BMID:
|
|
m = (typeof(m)){"LP1150", "PCI-X2", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_BSMB:
|
|
m = (typeof(m)){"LP111", "PCI-X2", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_ZEPHYR:
|
|
m = (typeof(m)){"LPe11000", "PCIe", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_ZEPHYR_SCSP:
|
|
m = (typeof(m)){"LPe11000", "PCIe", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_ZEPHYR_DCSP:
|
|
m = (typeof(m)){"LP2105", "PCIe", "FCoE Adapter"};
|
|
GE = 1;
|
|
break;
|
|
case PCI_DEVICE_ID_ZMID:
|
|
m = (typeof(m)){"LPe1150", "PCIe", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_ZSMB:
|
|
m = (typeof(m)){"LPe111", "PCIe", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_LP101:
|
|
m = (typeof(m)){"LP101", "PCI-X", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_LP10000S:
|
|
m = (typeof(m)){"LP10000-S", "PCI", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_LP11000S:
|
|
m = (typeof(m)){"LP11000-S", "PCI-X2", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_LPE11000S:
|
|
m = (typeof(m)){"LPe11000-S", "PCIe", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_SAT:
|
|
m = (typeof(m)){"LPe12000", "PCIe", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_SAT_MID:
|
|
m = (typeof(m)){"LPe1250", "PCIe", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_SAT_SMB:
|
|
m = (typeof(m)){"LPe121", "PCIe", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_SAT_DCSP:
|
|
m = (typeof(m)){"LPe12002-SP", "PCIe", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_SAT_SCSP:
|
|
m = (typeof(m)){"LPe12000-SP", "PCIe", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_SAT_S:
|
|
m = (typeof(m)){"LPe12000-S", "PCIe", "Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_HORNET:
|
|
m = (typeof(m)){"LP21000", "PCIe", "FCoE Adapter"};
|
|
GE = 1;
|
|
break;
|
|
case PCI_DEVICE_ID_PROTEUS_VF:
|
|
m = (typeof(m)){"LPev12000", "PCIe IOV",
|
|
"Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_PROTEUS_PF:
|
|
m = (typeof(m)){"LPev12000", "PCIe IOV",
|
|
"Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_PROTEUS_S:
|
|
m = (typeof(m)){"LPemv12002-S", "PCIe IOV",
|
|
"Fibre Channel Adapter"};
|
|
break;
|
|
case PCI_DEVICE_ID_TIGERSHARK:
|
|
oneConnect = 1;
|
|
m = (typeof(m)){"OCe10100", "PCIe", "FCoE"};
|
|
break;
|
|
case PCI_DEVICE_ID_TOMCAT:
|
|
oneConnect = 1;
|
|
m = (typeof(m)){"OCe11100", "PCIe", "FCoE"};
|
|
break;
|
|
case PCI_DEVICE_ID_FALCON:
|
|
m = (typeof(m)){"LPSe12002-ML1-E", "PCIe",
|
|
"EmulexSecure Fibre"};
|
|
break;
|
|
case PCI_DEVICE_ID_BALIUS:
|
|
m = (typeof(m)){"LPVe12002", "PCIe Shared I/O",
|
|
"Fibre Channel Adapter"};
|
|
break;
|
|
default:
|
|
m = (typeof(m)){"Unknown", "", ""};
|
|
break;
|
|
}
|
|
|
|
if (mdp && mdp[0] == '\0')
|
|
snprintf(mdp, 79,"%s", m.name);
|
|
/* oneConnect hba requires special processing, they are all initiators
|
|
* and we put the port number on the end
|
|
*/
|
|
if (descp && descp[0] == '\0') {
|
|
if (oneConnect)
|
|
snprintf(descp, 255,
|
|
"Emulex OneConnect %s, %s Initiator, Port %s",
|
|
m.name, m.function,
|
|
phba->Port);
|
|
else
|
|
snprintf(descp, 255,
|
|
"Emulex %s %d%s %s %s",
|
|
m.name, max_speed, (GE) ? "GE" : "Gb",
|
|
m.bus, m.function);
|
|
}
|
|
}
|
|
|
|
/**
|
|
* lpfc_post_buffer - Post IOCB(s) with DMA buffer descriptor(s) to a IOCB ring
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @pring: pointer to a IOCB ring.
|
|
* @cnt: the number of IOCBs to be posted to the IOCB ring.
|
|
*
|
|
* This routine posts a given number of IOCBs with the associated DMA buffer
|
|
* descriptors specified by the cnt argument to the given IOCB ring.
|
|
*
|
|
* Return codes
|
|
* The number of IOCBs NOT able to be posted to the IOCB ring.
|
|
**/
|
|
int
|
|
lpfc_post_buffer(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, int cnt)
|
|
{
|
|
IOCB_t *icmd;
|
|
struct lpfc_iocbq *iocb;
|
|
struct lpfc_dmabuf *mp1, *mp2;
|
|
|
|
cnt += pring->missbufcnt;
|
|
|
|
/* While there are buffers to post */
|
|
while (cnt > 0) {
|
|
/* Allocate buffer for command iocb */
|
|
iocb = lpfc_sli_get_iocbq(phba);
|
|
if (iocb == NULL) {
|
|
pring->missbufcnt = cnt;
|
|
return cnt;
|
|
}
|
|
icmd = &iocb->iocb;
|
|
|
|
/* 2 buffers can be posted per command */
|
|
/* Allocate buffer to post */
|
|
mp1 = kmalloc(sizeof (struct lpfc_dmabuf), GFP_KERNEL);
|
|
if (mp1)
|
|
mp1->virt = lpfc_mbuf_alloc(phba, MEM_PRI, &mp1->phys);
|
|
if (!mp1 || !mp1->virt) {
|
|
kfree(mp1);
|
|
lpfc_sli_release_iocbq(phba, iocb);
|
|
pring->missbufcnt = cnt;
|
|
return cnt;
|
|
}
|
|
|
|
INIT_LIST_HEAD(&mp1->list);
|
|
/* Allocate buffer to post */
|
|
if (cnt > 1) {
|
|
mp2 = kmalloc(sizeof (struct lpfc_dmabuf), GFP_KERNEL);
|
|
if (mp2)
|
|
mp2->virt = lpfc_mbuf_alloc(phba, MEM_PRI,
|
|
&mp2->phys);
|
|
if (!mp2 || !mp2->virt) {
|
|
kfree(mp2);
|
|
lpfc_mbuf_free(phba, mp1->virt, mp1->phys);
|
|
kfree(mp1);
|
|
lpfc_sli_release_iocbq(phba, iocb);
|
|
pring->missbufcnt = cnt;
|
|
return cnt;
|
|
}
|
|
|
|
INIT_LIST_HEAD(&mp2->list);
|
|
} else {
|
|
mp2 = NULL;
|
|
}
|
|
|
|
icmd->un.cont64[0].addrHigh = putPaddrHigh(mp1->phys);
|
|
icmd->un.cont64[0].addrLow = putPaddrLow(mp1->phys);
|
|
icmd->un.cont64[0].tus.f.bdeSize = FCELSSIZE;
|
|
icmd->ulpBdeCount = 1;
|
|
cnt--;
|
|
if (mp2) {
|
|
icmd->un.cont64[1].addrHigh = putPaddrHigh(mp2->phys);
|
|
icmd->un.cont64[1].addrLow = putPaddrLow(mp2->phys);
|
|
icmd->un.cont64[1].tus.f.bdeSize = FCELSSIZE;
|
|
cnt--;
|
|
icmd->ulpBdeCount = 2;
|
|
}
|
|
|
|
icmd->ulpCommand = CMD_QUE_RING_BUF64_CN;
|
|
icmd->ulpLe = 1;
|
|
|
|
if (lpfc_sli_issue_iocb(phba, pring->ringno, iocb, 0) ==
|
|
IOCB_ERROR) {
|
|
lpfc_mbuf_free(phba, mp1->virt, mp1->phys);
|
|
kfree(mp1);
|
|
cnt++;
|
|
if (mp2) {
|
|
lpfc_mbuf_free(phba, mp2->virt, mp2->phys);
|
|
kfree(mp2);
|
|
cnt++;
|
|
}
|
|
lpfc_sli_release_iocbq(phba, iocb);
|
|
pring->missbufcnt = cnt;
|
|
return cnt;
|
|
}
|
|
lpfc_sli_ringpostbuf_put(phba, pring, mp1);
|
|
if (mp2)
|
|
lpfc_sli_ringpostbuf_put(phba, pring, mp2);
|
|
}
|
|
pring->missbufcnt = 0;
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_post_rcv_buf - Post the initial receive IOCB buffers to ELS ring
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine posts initial receive IOCB buffers to the ELS ring. The
|
|
* current number of initial IOCB buffers specified by LPFC_BUF_RING0 is
|
|
* set to 64 IOCBs.
|
|
*
|
|
* Return codes
|
|
* 0 - success (currently always success)
|
|
**/
|
|
static int
|
|
lpfc_post_rcv_buf(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_sli *psli = &phba->sli;
|
|
|
|
/* Ring 0, ELS / CT buffers */
|
|
lpfc_post_buffer(phba, &psli->ring[LPFC_ELS_RING], LPFC_BUF_RING0);
|
|
/* Ring 2 - FCP no buffers needed */
|
|
|
|
return 0;
|
|
}
|
|
|
|
#define S(N,V) (((V)<<(N))|((V)>>(32-(N))))
|
|
|
|
/**
|
|
* lpfc_sha_init - Set up initial array of hash table entries
|
|
* @HashResultPointer: pointer to an array as hash table.
|
|
*
|
|
* This routine sets up the initial values to the array of hash table entries
|
|
* for the LC HBAs.
|
|
**/
|
|
static void
|
|
lpfc_sha_init(uint32_t * HashResultPointer)
|
|
{
|
|
HashResultPointer[0] = 0x67452301;
|
|
HashResultPointer[1] = 0xEFCDAB89;
|
|
HashResultPointer[2] = 0x98BADCFE;
|
|
HashResultPointer[3] = 0x10325476;
|
|
HashResultPointer[4] = 0xC3D2E1F0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sha_iterate - Iterate initial hash table with the working hash table
|
|
* @HashResultPointer: pointer to an initial/result hash table.
|
|
* @HashWorkingPointer: pointer to an working hash table.
|
|
*
|
|
* This routine iterates an initial hash table pointed by @HashResultPointer
|
|
* with the values from the working hash table pointeed by @HashWorkingPointer.
|
|
* The results are putting back to the initial hash table, returned through
|
|
* the @HashResultPointer as the result hash table.
|
|
**/
|
|
static void
|
|
lpfc_sha_iterate(uint32_t * HashResultPointer, uint32_t * HashWorkingPointer)
|
|
{
|
|
int t;
|
|
uint32_t TEMP;
|
|
uint32_t A, B, C, D, E;
|
|
t = 16;
|
|
do {
|
|
HashWorkingPointer[t] =
|
|
S(1,
|
|
HashWorkingPointer[t - 3] ^ HashWorkingPointer[t -
|
|
8] ^
|
|
HashWorkingPointer[t - 14] ^ HashWorkingPointer[t - 16]);
|
|
} while (++t <= 79);
|
|
t = 0;
|
|
A = HashResultPointer[0];
|
|
B = HashResultPointer[1];
|
|
C = HashResultPointer[2];
|
|
D = HashResultPointer[3];
|
|
E = HashResultPointer[4];
|
|
|
|
do {
|
|
if (t < 20) {
|
|
TEMP = ((B & C) | ((~B) & D)) + 0x5A827999;
|
|
} else if (t < 40) {
|
|
TEMP = (B ^ C ^ D) + 0x6ED9EBA1;
|
|
} else if (t < 60) {
|
|
TEMP = ((B & C) | (B & D) | (C & D)) + 0x8F1BBCDC;
|
|
} else {
|
|
TEMP = (B ^ C ^ D) + 0xCA62C1D6;
|
|
}
|
|
TEMP += S(5, A) + E + HashWorkingPointer[t];
|
|
E = D;
|
|
D = C;
|
|
C = S(30, B);
|
|
B = A;
|
|
A = TEMP;
|
|
} while (++t <= 79);
|
|
|
|
HashResultPointer[0] += A;
|
|
HashResultPointer[1] += B;
|
|
HashResultPointer[2] += C;
|
|
HashResultPointer[3] += D;
|
|
HashResultPointer[4] += E;
|
|
|
|
}
|
|
|
|
/**
|
|
* lpfc_challenge_key - Create challenge key based on WWPN of the HBA
|
|
* @RandomChallenge: pointer to the entry of host challenge random number array.
|
|
* @HashWorking: pointer to the entry of the working hash array.
|
|
*
|
|
* This routine calculates the working hash array referred by @HashWorking
|
|
* from the challenge random numbers associated with the host, referred by
|
|
* @RandomChallenge. The result is put into the entry of the working hash
|
|
* array and returned by reference through @HashWorking.
|
|
**/
|
|
static void
|
|
lpfc_challenge_key(uint32_t * RandomChallenge, uint32_t * HashWorking)
|
|
{
|
|
*HashWorking = (*RandomChallenge ^ *HashWorking);
|
|
}
|
|
|
|
/**
|
|
* lpfc_hba_init - Perform special handling for LC HBA initialization
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @hbainit: pointer to an array of unsigned 32-bit integers.
|
|
*
|
|
* This routine performs the special handling for LC HBA initialization.
|
|
**/
|
|
void
|
|
lpfc_hba_init(struct lpfc_hba *phba, uint32_t *hbainit)
|
|
{
|
|
int t;
|
|
uint32_t *HashWorking;
|
|
uint32_t *pwwnn = (uint32_t *) phba->wwnn;
|
|
|
|
HashWorking = kcalloc(80, sizeof(uint32_t), GFP_KERNEL);
|
|
if (!HashWorking)
|
|
return;
|
|
|
|
HashWorking[0] = HashWorking[78] = *pwwnn++;
|
|
HashWorking[1] = HashWorking[79] = *pwwnn;
|
|
|
|
for (t = 0; t < 7; t++)
|
|
lpfc_challenge_key(phba->RandomData + t, HashWorking + t);
|
|
|
|
lpfc_sha_init(hbainit);
|
|
lpfc_sha_iterate(hbainit, HashWorking);
|
|
kfree(HashWorking);
|
|
}
|
|
|
|
/**
|
|
* lpfc_cleanup - Performs vport cleanups before deleting a vport
|
|
* @vport: pointer to a virtual N_Port data structure.
|
|
*
|
|
* This routine performs the necessary cleanups before deleting the @vport.
|
|
* It invokes the discovery state machine to perform necessary state
|
|
* transitions and to release the ndlps associated with the @vport. Note,
|
|
* the physical port is treated as @vport 0.
|
|
**/
|
|
void
|
|
lpfc_cleanup(struct lpfc_vport *vport)
|
|
{
|
|
struct lpfc_hba *phba = vport->phba;
|
|
struct lpfc_nodelist *ndlp, *next_ndlp;
|
|
int i = 0;
|
|
|
|
if (phba->link_state > LPFC_LINK_DOWN)
|
|
lpfc_port_link_failure(vport);
|
|
|
|
list_for_each_entry_safe(ndlp, next_ndlp, &vport->fc_nodes, nlp_listp) {
|
|
if (!NLP_CHK_NODE_ACT(ndlp)) {
|
|
ndlp = lpfc_enable_node(vport, ndlp,
|
|
NLP_STE_UNUSED_NODE);
|
|
if (!ndlp)
|
|
continue;
|
|
spin_lock_irq(&phba->ndlp_lock);
|
|
NLP_SET_FREE_REQ(ndlp);
|
|
spin_unlock_irq(&phba->ndlp_lock);
|
|
/* Trigger the release of the ndlp memory */
|
|
lpfc_nlp_put(ndlp);
|
|
continue;
|
|
}
|
|
spin_lock_irq(&phba->ndlp_lock);
|
|
if (NLP_CHK_FREE_REQ(ndlp)) {
|
|
/* The ndlp should not be in memory free mode already */
|
|
spin_unlock_irq(&phba->ndlp_lock);
|
|
continue;
|
|
} else
|
|
/* Indicate request for freeing ndlp memory */
|
|
NLP_SET_FREE_REQ(ndlp);
|
|
spin_unlock_irq(&phba->ndlp_lock);
|
|
|
|
if (vport->port_type != LPFC_PHYSICAL_PORT &&
|
|
ndlp->nlp_DID == Fabric_DID) {
|
|
/* Just free up ndlp with Fabric_DID for vports */
|
|
lpfc_nlp_put(ndlp);
|
|
continue;
|
|
}
|
|
|
|
if (ndlp->nlp_type & NLP_FABRIC)
|
|
lpfc_disc_state_machine(vport, ndlp, NULL,
|
|
NLP_EVT_DEVICE_RECOVERY);
|
|
|
|
lpfc_disc_state_machine(vport, ndlp, NULL,
|
|
NLP_EVT_DEVICE_RM);
|
|
|
|
}
|
|
|
|
/* At this point, ALL ndlp's should be gone
|
|
* because of the previous NLP_EVT_DEVICE_RM.
|
|
* Lets wait for this to happen, if needed.
|
|
*/
|
|
while (!list_empty(&vport->fc_nodes)) {
|
|
if (i++ > 3000) {
|
|
lpfc_printf_vlog(vport, KERN_ERR, LOG_DISCOVERY,
|
|
"0233 Nodelist not empty\n");
|
|
list_for_each_entry_safe(ndlp, next_ndlp,
|
|
&vport->fc_nodes, nlp_listp) {
|
|
lpfc_printf_vlog(ndlp->vport, KERN_ERR,
|
|
LOG_NODE,
|
|
"0282 did:x%x ndlp:x%p "
|
|
"usgmap:x%x refcnt:%d\n",
|
|
ndlp->nlp_DID, (void *)ndlp,
|
|
ndlp->nlp_usg_map,
|
|
atomic_read(
|
|
&ndlp->kref.refcount));
|
|
}
|
|
break;
|
|
}
|
|
|
|
/* Wait for any activity on ndlps to settle */
|
|
msleep(10);
|
|
}
|
|
}
|
|
|
|
/**
|
|
* lpfc_stop_vport_timers - Stop all the timers associated with a vport
|
|
* @vport: pointer to a virtual N_Port data structure.
|
|
*
|
|
* This routine stops all the timers associated with a @vport. This function
|
|
* is invoked before disabling or deleting a @vport. Note that the physical
|
|
* port is treated as @vport 0.
|
|
**/
|
|
void
|
|
lpfc_stop_vport_timers(struct lpfc_vport *vport)
|
|
{
|
|
del_timer_sync(&vport->els_tmofunc);
|
|
del_timer_sync(&vport->fc_fdmitmo);
|
|
lpfc_can_disctmo(vport);
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* __lpfc_sli4_stop_fcf_redisc_wait_timer - Stop FCF rediscovery wait timer
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine stops the SLI4 FCF rediscover wait timer if it's on. The
|
|
* caller of this routine should already hold the host lock.
|
|
**/
|
|
void
|
|
__lpfc_sli4_stop_fcf_redisc_wait_timer(struct lpfc_hba *phba)
|
|
{
|
|
/* Clear pending FCF rediscovery wait flag */
|
|
phba->fcf.fcf_flag &= ~FCF_REDISC_PEND;
|
|
|
|
/* Now, try to stop the timer */
|
|
del_timer(&phba->fcf.redisc_wait);
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_stop_fcf_redisc_wait_timer - Stop FCF rediscovery wait timer
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine stops the SLI4 FCF rediscover wait timer if it's on. It
|
|
* checks whether the FCF rediscovery wait timer is pending with the host
|
|
* lock held before proceeding with disabling the timer and clearing the
|
|
* wait timer pendig flag.
|
|
**/
|
|
void
|
|
lpfc_sli4_stop_fcf_redisc_wait_timer(struct lpfc_hba *phba)
|
|
{
|
|
spin_lock_irq(&phba->hbalock);
|
|
if (!(phba->fcf.fcf_flag & FCF_REDISC_PEND)) {
|
|
/* FCF rediscovery timer already fired or stopped */
|
|
spin_unlock_irq(&phba->hbalock);
|
|
return;
|
|
}
|
|
__lpfc_sli4_stop_fcf_redisc_wait_timer(phba);
|
|
/* Clear failover in progress flags */
|
|
phba->fcf.fcf_flag &= ~(FCF_DEAD_DISC | FCF_ACVL_DISC);
|
|
spin_unlock_irq(&phba->hbalock);
|
|
}
|
|
|
|
/**
|
|
* lpfc_stop_hba_timers - Stop all the timers associated with an HBA
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine stops all the timers associated with a HBA. This function is
|
|
* invoked before either putting a HBA offline or unloading the driver.
|
|
**/
|
|
void
|
|
lpfc_stop_hba_timers(struct lpfc_hba *phba)
|
|
{
|
|
lpfc_stop_vport_timers(phba->pport);
|
|
del_timer_sync(&phba->sli.mbox_tmo);
|
|
del_timer_sync(&phba->fabric_block_timer);
|
|
del_timer_sync(&phba->eratt_poll);
|
|
del_timer_sync(&phba->hb_tmofunc);
|
|
phba->hb_outstanding = 0;
|
|
|
|
switch (phba->pci_dev_grp) {
|
|
case LPFC_PCI_DEV_LP:
|
|
/* Stop any LightPulse device specific driver timers */
|
|
del_timer_sync(&phba->fcp_poll_timer);
|
|
break;
|
|
case LPFC_PCI_DEV_OC:
|
|
/* Stop any OneConnect device sepcific driver timers */
|
|
lpfc_sli4_stop_fcf_redisc_wait_timer(phba);
|
|
break;
|
|
default:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0297 Invalid device group (x%x)\n",
|
|
phba->pci_dev_grp);
|
|
break;
|
|
}
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_block_mgmt_io - Mark a HBA's management interface as blocked
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine marks a HBA's management interface as blocked. Once the HBA's
|
|
* management interface is marked as blocked, all the user space access to
|
|
* the HBA, whether they are from sysfs interface or libdfc interface will
|
|
* all be blocked. The HBA is set to block the management interface when the
|
|
* driver prepares the HBA interface for online or offline.
|
|
**/
|
|
static void
|
|
lpfc_block_mgmt_io(struct lpfc_hba * phba)
|
|
{
|
|
unsigned long iflag;
|
|
uint8_t actcmd = MBX_HEARTBEAT;
|
|
unsigned long timeout;
|
|
|
|
|
|
spin_lock_irqsave(&phba->hbalock, iflag);
|
|
phba->sli.sli_flag |= LPFC_BLOCK_MGMT_IO;
|
|
if (phba->sli.mbox_active)
|
|
actcmd = phba->sli.mbox_active->u.mb.mbxCommand;
|
|
spin_unlock_irqrestore(&phba->hbalock, iflag);
|
|
/* Determine how long we might wait for the active mailbox
|
|
* command to be gracefully completed by firmware.
|
|
*/
|
|
timeout = msecs_to_jiffies(lpfc_mbox_tmo_val(phba, actcmd) * 1000) +
|
|
jiffies;
|
|
/* Wait for the outstnading mailbox command to complete */
|
|
while (phba->sli.mbox_active) {
|
|
/* Check active mailbox complete status every 2ms */
|
|
msleep(2);
|
|
if (time_after(jiffies, timeout)) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
|
|
"2813 Mgmt IO is Blocked %x "
|
|
"- mbox cmd %x still active\n",
|
|
phba->sli.sli_flag, actcmd);
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
|
|
/**
|
|
* lpfc_online - Initialize and bring a HBA online
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine initializes the HBA and brings a HBA online. During this
|
|
* process, the management interface is blocked to prevent user space access
|
|
* to the HBA interfering with the driver initialization.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* 1 - failed
|
|
**/
|
|
int
|
|
lpfc_online(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_vport *vport;
|
|
struct lpfc_vport **vports;
|
|
int i;
|
|
|
|
if (!phba)
|
|
return 0;
|
|
vport = phba->pport;
|
|
|
|
if (!(vport->fc_flag & FC_OFFLINE_MODE))
|
|
return 0;
|
|
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_INIT,
|
|
"0458 Bring Adapter online\n");
|
|
|
|
lpfc_block_mgmt_io(phba);
|
|
|
|
if (!lpfc_sli_queue_setup(phba)) {
|
|
lpfc_unblock_mgmt_io(phba);
|
|
return 1;
|
|
}
|
|
|
|
if (phba->sli_rev == LPFC_SLI_REV4) {
|
|
if (lpfc_sli4_hba_setup(phba)) { /* Initialize SLI4 HBA */
|
|
lpfc_unblock_mgmt_io(phba);
|
|
return 1;
|
|
}
|
|
} else {
|
|
if (lpfc_sli_hba_setup(phba)) { /* Initialize SLI2/SLI3 HBA */
|
|
lpfc_unblock_mgmt_io(phba);
|
|
return 1;
|
|
}
|
|
}
|
|
|
|
vports = lpfc_create_vport_work_array(phba);
|
|
if (vports != NULL)
|
|
for (i = 0; i <= phba->max_vports && vports[i] != NULL; i++) {
|
|
struct Scsi_Host *shost;
|
|
shost = lpfc_shost_from_vport(vports[i]);
|
|
spin_lock_irq(shost->host_lock);
|
|
vports[i]->fc_flag &= ~FC_OFFLINE_MODE;
|
|
if (phba->sli3_options & LPFC_SLI3_NPIV_ENABLED)
|
|
vports[i]->fc_flag |= FC_VPORT_NEEDS_REG_VPI;
|
|
if (phba->sli_rev == LPFC_SLI_REV4)
|
|
vports[i]->fc_flag |= FC_VPORT_NEEDS_INIT_VPI;
|
|
spin_unlock_irq(shost->host_lock);
|
|
}
|
|
lpfc_destroy_vport_work_array(phba, vports);
|
|
|
|
lpfc_unblock_mgmt_io(phba);
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_unblock_mgmt_io - Mark a HBA's management interface to be not blocked
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine marks a HBA's management interface as not blocked. Once the
|
|
* HBA's management interface is marked as not blocked, all the user space
|
|
* access to the HBA, whether they are from sysfs interface or libdfc
|
|
* interface will be allowed. The HBA is set to block the management interface
|
|
* when the driver prepares the HBA interface for online or offline and then
|
|
* set to unblock the management interface afterwards.
|
|
**/
|
|
void
|
|
lpfc_unblock_mgmt_io(struct lpfc_hba * phba)
|
|
{
|
|
unsigned long iflag;
|
|
|
|
spin_lock_irqsave(&phba->hbalock, iflag);
|
|
phba->sli.sli_flag &= ~LPFC_BLOCK_MGMT_IO;
|
|
spin_unlock_irqrestore(&phba->hbalock, iflag);
|
|
}
|
|
|
|
/**
|
|
* lpfc_offline_prep - Prepare a HBA to be brought offline
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to prepare a HBA to be brought offline. It performs
|
|
* unregistration login to all the nodes on all vports and flushes the mailbox
|
|
* queue to make it ready to be brought offline.
|
|
**/
|
|
void
|
|
lpfc_offline_prep(struct lpfc_hba * phba)
|
|
{
|
|
struct lpfc_vport *vport = phba->pport;
|
|
struct lpfc_nodelist *ndlp, *next_ndlp;
|
|
struct lpfc_vport **vports;
|
|
struct Scsi_Host *shost;
|
|
int i;
|
|
|
|
if (vport->fc_flag & FC_OFFLINE_MODE)
|
|
return;
|
|
|
|
lpfc_block_mgmt_io(phba);
|
|
|
|
lpfc_linkdown(phba);
|
|
|
|
/* Issue an unreg_login to all nodes on all vports */
|
|
vports = lpfc_create_vport_work_array(phba);
|
|
if (vports != NULL) {
|
|
for (i = 0; i <= phba->max_vports && vports[i] != NULL; i++) {
|
|
if (vports[i]->load_flag & FC_UNLOADING)
|
|
continue;
|
|
shost = lpfc_shost_from_vport(vports[i]);
|
|
spin_lock_irq(shost->host_lock);
|
|
vports[i]->vpi_state &= ~LPFC_VPI_REGISTERED;
|
|
vports[i]->fc_flag |= FC_VPORT_NEEDS_REG_VPI;
|
|
vports[i]->fc_flag &= ~FC_VFI_REGISTERED;
|
|
spin_unlock_irq(shost->host_lock);
|
|
|
|
shost = lpfc_shost_from_vport(vports[i]);
|
|
list_for_each_entry_safe(ndlp, next_ndlp,
|
|
&vports[i]->fc_nodes,
|
|
nlp_listp) {
|
|
if (!NLP_CHK_NODE_ACT(ndlp))
|
|
continue;
|
|
if (ndlp->nlp_state == NLP_STE_UNUSED_NODE)
|
|
continue;
|
|
if (ndlp->nlp_type & NLP_FABRIC) {
|
|
lpfc_disc_state_machine(vports[i], ndlp,
|
|
NULL, NLP_EVT_DEVICE_RECOVERY);
|
|
lpfc_disc_state_machine(vports[i], ndlp,
|
|
NULL, NLP_EVT_DEVICE_RM);
|
|
}
|
|
spin_lock_irq(shost->host_lock);
|
|
ndlp->nlp_flag &= ~NLP_NPR_ADISC;
|
|
spin_unlock_irq(shost->host_lock);
|
|
lpfc_unreg_rpi(vports[i], ndlp);
|
|
}
|
|
}
|
|
}
|
|
lpfc_destroy_vport_work_array(phba, vports);
|
|
|
|
lpfc_sli_mbox_sys_shutdown(phba);
|
|
}
|
|
|
|
/**
|
|
* lpfc_offline - Bring a HBA offline
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine actually brings a HBA offline. It stops all the timers
|
|
* associated with the HBA, brings down the SLI layer, and eventually
|
|
* marks the HBA as in offline state for the upper layer protocol.
|
|
**/
|
|
void
|
|
lpfc_offline(struct lpfc_hba *phba)
|
|
{
|
|
struct Scsi_Host *shost;
|
|
struct lpfc_vport **vports;
|
|
int i;
|
|
|
|
if (phba->pport->fc_flag & FC_OFFLINE_MODE)
|
|
return;
|
|
|
|
/* stop port and all timers associated with this hba */
|
|
lpfc_stop_port(phba);
|
|
vports = lpfc_create_vport_work_array(phba);
|
|
if (vports != NULL)
|
|
for (i = 0; i <= phba->max_vports && vports[i] != NULL; i++)
|
|
lpfc_stop_vport_timers(vports[i]);
|
|
lpfc_destroy_vport_work_array(phba, vports);
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_INIT,
|
|
"0460 Bring Adapter offline\n");
|
|
/* Bring down the SLI Layer and cleanup. The HBA is offline
|
|
now. */
|
|
lpfc_sli_hba_down(phba);
|
|
spin_lock_irq(&phba->hbalock);
|
|
phba->work_ha = 0;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
vports = lpfc_create_vport_work_array(phba);
|
|
if (vports != NULL)
|
|
for (i = 0; i <= phba->max_vports && vports[i] != NULL; i++) {
|
|
shost = lpfc_shost_from_vport(vports[i]);
|
|
spin_lock_irq(shost->host_lock);
|
|
vports[i]->work_port_events = 0;
|
|
vports[i]->fc_flag |= FC_OFFLINE_MODE;
|
|
spin_unlock_irq(shost->host_lock);
|
|
}
|
|
lpfc_destroy_vport_work_array(phba, vports);
|
|
}
|
|
|
|
/**
|
|
* lpfc_scsi_free - Free all the SCSI buffers and IOCBs from driver lists
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is to free all the SCSI buffers and IOCBs from the driver
|
|
* list back to kernel. It is called from lpfc_pci_remove_one to free
|
|
* the internal resources before the device is removed from the system.
|
|
*
|
|
* Return codes
|
|
* 0 - successful (for now, it always returns 0)
|
|
**/
|
|
static int
|
|
lpfc_scsi_free(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_scsi_buf *sb, *sb_next;
|
|
struct lpfc_iocbq *io, *io_next;
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
/* Release all the lpfc_scsi_bufs maintained by this host. */
|
|
spin_lock(&phba->scsi_buf_list_lock);
|
|
list_for_each_entry_safe(sb, sb_next, &phba->lpfc_scsi_buf_list, list) {
|
|
list_del(&sb->list);
|
|
pci_pool_free(phba->lpfc_scsi_dma_buf_pool, sb->data,
|
|
sb->dma_handle);
|
|
kfree(sb);
|
|
phba->total_scsi_bufs--;
|
|
}
|
|
spin_unlock(&phba->scsi_buf_list_lock);
|
|
|
|
/* Release all the lpfc_iocbq entries maintained by this host. */
|
|
list_for_each_entry_safe(io, io_next, &phba->lpfc_iocb_list, list) {
|
|
list_del(&io->list);
|
|
kfree(io);
|
|
phba->total_iocbq_bufs--;
|
|
}
|
|
spin_unlock_irq(&phba->hbalock);
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_create_port - Create an FC port
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @instance: a unique integer ID to this FC port.
|
|
* @dev: pointer to the device data structure.
|
|
*
|
|
* This routine creates a FC port for the upper layer protocol. The FC port
|
|
* can be created on top of either a physical port or a virtual port provided
|
|
* by the HBA. This routine also allocates a SCSI host data structure (shost)
|
|
* and associates the FC port created before adding the shost into the SCSI
|
|
* layer.
|
|
*
|
|
* Return codes
|
|
* @vport - pointer to the virtual N_Port data structure.
|
|
* NULL - port create failed.
|
|
**/
|
|
struct lpfc_vport *
|
|
lpfc_create_port(struct lpfc_hba *phba, int instance, struct device *dev)
|
|
{
|
|
struct lpfc_vport *vport;
|
|
struct Scsi_Host *shost;
|
|
int error = 0;
|
|
|
|
if (dev != &phba->pcidev->dev)
|
|
shost = scsi_host_alloc(&lpfc_vport_template,
|
|
sizeof(struct lpfc_vport));
|
|
else
|
|
shost = scsi_host_alloc(&lpfc_template,
|
|
sizeof(struct lpfc_vport));
|
|
if (!shost)
|
|
goto out;
|
|
|
|
vport = (struct lpfc_vport *) shost->hostdata;
|
|
vport->phba = phba;
|
|
vport->load_flag |= FC_LOADING;
|
|
vport->fc_flag |= FC_VPORT_NEEDS_REG_VPI;
|
|
vport->fc_rscn_flush = 0;
|
|
|
|
lpfc_get_vport_cfgparam(vport);
|
|
shost->unique_id = instance;
|
|
shost->max_id = LPFC_MAX_TARGET;
|
|
shost->max_lun = vport->cfg_max_luns;
|
|
shost->this_id = -1;
|
|
shost->max_cmd_len = 16;
|
|
if (phba->sli_rev == LPFC_SLI_REV4) {
|
|
shost->dma_boundary =
|
|
phba->sli4_hba.pc_sli4_params.sge_supp_len-1;
|
|
shost->sg_tablesize = phba->cfg_sg_seg_cnt;
|
|
}
|
|
|
|
/*
|
|
* Set initial can_queue value since 0 is no longer supported and
|
|
* scsi_add_host will fail. This will be adjusted later based on the
|
|
* max xri value determined in hba setup.
|
|
*/
|
|
shost->can_queue = phba->cfg_hba_queue_depth - 10;
|
|
if (dev != &phba->pcidev->dev) {
|
|
shost->transportt = lpfc_vport_transport_template;
|
|
vport->port_type = LPFC_NPIV_PORT;
|
|
} else {
|
|
shost->transportt = lpfc_transport_template;
|
|
vport->port_type = LPFC_PHYSICAL_PORT;
|
|
}
|
|
|
|
/* Initialize all internally managed lists. */
|
|
INIT_LIST_HEAD(&vport->fc_nodes);
|
|
INIT_LIST_HEAD(&vport->rcv_buffer_list);
|
|
spin_lock_init(&vport->work_port_lock);
|
|
|
|
init_timer(&vport->fc_disctmo);
|
|
vport->fc_disctmo.function = lpfc_disc_timeout;
|
|
vport->fc_disctmo.data = (unsigned long)vport;
|
|
|
|
init_timer(&vport->fc_fdmitmo);
|
|
vport->fc_fdmitmo.function = lpfc_fdmi_tmo;
|
|
vport->fc_fdmitmo.data = (unsigned long)vport;
|
|
|
|
init_timer(&vport->els_tmofunc);
|
|
vport->els_tmofunc.function = lpfc_els_timeout;
|
|
vport->els_tmofunc.data = (unsigned long)vport;
|
|
error = scsi_add_host_with_dma(shost, dev, &phba->pcidev->dev);
|
|
if (error)
|
|
goto out_put_shost;
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
list_add_tail(&vport->listentry, &phba->port_list);
|
|
spin_unlock_irq(&phba->hbalock);
|
|
return vport;
|
|
|
|
out_put_shost:
|
|
scsi_host_put(shost);
|
|
out:
|
|
return NULL;
|
|
}
|
|
|
|
/**
|
|
* destroy_port - destroy an FC port
|
|
* @vport: pointer to an lpfc virtual N_Port data structure.
|
|
*
|
|
* This routine destroys a FC port from the upper layer protocol. All the
|
|
* resources associated with the port are released.
|
|
**/
|
|
void
|
|
destroy_port(struct lpfc_vport *vport)
|
|
{
|
|
struct Scsi_Host *shost = lpfc_shost_from_vport(vport);
|
|
struct lpfc_hba *phba = vport->phba;
|
|
|
|
lpfc_debugfs_terminate(vport);
|
|
fc_remove_host(shost);
|
|
scsi_remove_host(shost);
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
list_del_init(&vport->listentry);
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
lpfc_cleanup(vport);
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_get_instance - Get a unique integer ID
|
|
*
|
|
* This routine allocates a unique integer ID from lpfc_hba_index pool. It
|
|
* uses the kernel idr facility to perform the task.
|
|
*
|
|
* Return codes:
|
|
* instance - a unique integer ID allocated as the new instance.
|
|
* -1 - lpfc get instance failed.
|
|
**/
|
|
int
|
|
lpfc_get_instance(void)
|
|
{
|
|
int instance = 0;
|
|
|
|
/* Assign an unused number */
|
|
if (!idr_pre_get(&lpfc_hba_index, GFP_KERNEL))
|
|
return -1;
|
|
if (idr_get_new(&lpfc_hba_index, NULL, &instance))
|
|
return -1;
|
|
return instance;
|
|
}
|
|
|
|
/**
|
|
* lpfc_scan_finished - method for SCSI layer to detect whether scan is done
|
|
* @shost: pointer to SCSI host data structure.
|
|
* @time: elapsed time of the scan in jiffies.
|
|
*
|
|
* This routine is called by the SCSI layer with a SCSI host to determine
|
|
* whether the scan host is finished.
|
|
*
|
|
* Note: there is no scan_start function as adapter initialization will have
|
|
* asynchronously kicked off the link initialization.
|
|
*
|
|
* Return codes
|
|
* 0 - SCSI host scan is not over yet.
|
|
* 1 - SCSI host scan is over.
|
|
**/
|
|
int lpfc_scan_finished(struct Scsi_Host *shost, unsigned long time)
|
|
{
|
|
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
|
|
struct lpfc_hba *phba = vport->phba;
|
|
int stat = 0;
|
|
|
|
spin_lock_irq(shost->host_lock);
|
|
|
|
if (vport->load_flag & FC_UNLOADING) {
|
|
stat = 1;
|
|
goto finished;
|
|
}
|
|
if (time >= 30 * HZ) {
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0461 Scanning longer than 30 "
|
|
"seconds. Continuing initialization\n");
|
|
stat = 1;
|
|
goto finished;
|
|
}
|
|
if (time >= 15 * HZ && phba->link_state <= LPFC_LINK_DOWN) {
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0465 Link down longer than 15 "
|
|
"seconds. Continuing initialization\n");
|
|
stat = 1;
|
|
goto finished;
|
|
}
|
|
|
|
if (vport->port_state != LPFC_VPORT_READY)
|
|
goto finished;
|
|
if (vport->num_disc_nodes || vport->fc_prli_sent)
|
|
goto finished;
|
|
if (vport->fc_map_cnt == 0 && time < 2 * HZ)
|
|
goto finished;
|
|
if ((phba->sli.sli_flag & LPFC_SLI_MBOX_ACTIVE) != 0)
|
|
goto finished;
|
|
|
|
stat = 1;
|
|
|
|
finished:
|
|
spin_unlock_irq(shost->host_lock);
|
|
return stat;
|
|
}
|
|
|
|
/**
|
|
* lpfc_host_attrib_init - Initialize SCSI host attributes on a FC port
|
|
* @shost: pointer to SCSI host data structure.
|
|
*
|
|
* This routine initializes a given SCSI host attributes on a FC port. The
|
|
* SCSI host can be either on top of a physical port or a virtual port.
|
|
**/
|
|
void lpfc_host_attrib_init(struct Scsi_Host *shost)
|
|
{
|
|
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
|
|
struct lpfc_hba *phba = vport->phba;
|
|
/*
|
|
* Set fixed host attributes. Must done after lpfc_sli_hba_setup().
|
|
*/
|
|
|
|
fc_host_node_name(shost) = wwn_to_u64(vport->fc_nodename.u.wwn);
|
|
fc_host_port_name(shost) = wwn_to_u64(vport->fc_portname.u.wwn);
|
|
fc_host_supported_classes(shost) = FC_COS_CLASS3;
|
|
|
|
memset(fc_host_supported_fc4s(shost), 0,
|
|
sizeof(fc_host_supported_fc4s(shost)));
|
|
fc_host_supported_fc4s(shost)[2] = 1;
|
|
fc_host_supported_fc4s(shost)[7] = 1;
|
|
|
|
lpfc_vport_symbolic_node_name(vport, fc_host_symbolic_name(shost),
|
|
sizeof fc_host_symbolic_name(shost));
|
|
|
|
fc_host_supported_speeds(shost) = 0;
|
|
if (phba->lmt & LMT_10Gb)
|
|
fc_host_supported_speeds(shost) |= FC_PORTSPEED_10GBIT;
|
|
if (phba->lmt & LMT_8Gb)
|
|
fc_host_supported_speeds(shost) |= FC_PORTSPEED_8GBIT;
|
|
if (phba->lmt & LMT_4Gb)
|
|
fc_host_supported_speeds(shost) |= FC_PORTSPEED_4GBIT;
|
|
if (phba->lmt & LMT_2Gb)
|
|
fc_host_supported_speeds(shost) |= FC_PORTSPEED_2GBIT;
|
|
if (phba->lmt & LMT_1Gb)
|
|
fc_host_supported_speeds(shost) |= FC_PORTSPEED_1GBIT;
|
|
|
|
fc_host_maxframe_size(shost) =
|
|
(((uint32_t) vport->fc_sparam.cmn.bbRcvSizeMsb & 0x0F) << 8) |
|
|
(uint32_t) vport->fc_sparam.cmn.bbRcvSizeLsb;
|
|
|
|
fc_host_dev_loss_tmo(shost) = vport->cfg_devloss_tmo;
|
|
|
|
/* This value is also unchanging */
|
|
memset(fc_host_active_fc4s(shost), 0,
|
|
sizeof(fc_host_active_fc4s(shost)));
|
|
fc_host_active_fc4s(shost)[2] = 1;
|
|
fc_host_active_fc4s(shost)[7] = 1;
|
|
|
|
fc_host_max_npiv_vports(shost) = phba->max_vpi;
|
|
spin_lock_irq(shost->host_lock);
|
|
vport->load_flag &= ~FC_LOADING;
|
|
spin_unlock_irq(shost->host_lock);
|
|
}
|
|
|
|
/**
|
|
* lpfc_stop_port_s3 - Stop SLI3 device port
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to stop an SLI3 device port, it stops the device
|
|
* from generating interrupts and stops the device driver's timers for the
|
|
* device.
|
|
**/
|
|
static void
|
|
lpfc_stop_port_s3(struct lpfc_hba *phba)
|
|
{
|
|
/* Clear all interrupt enable conditions */
|
|
writel(0, phba->HCregaddr);
|
|
readl(phba->HCregaddr); /* flush */
|
|
/* Clear all pending interrupts */
|
|
writel(0xffffffff, phba->HAregaddr);
|
|
readl(phba->HAregaddr); /* flush */
|
|
|
|
/* Reset some HBA SLI setup states */
|
|
lpfc_stop_hba_timers(phba);
|
|
phba->pport->work_port_events = 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_stop_port_s4 - Stop SLI4 device port
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to stop an SLI4 device port, it stops the device
|
|
* from generating interrupts and stops the device driver's timers for the
|
|
* device.
|
|
**/
|
|
static void
|
|
lpfc_stop_port_s4(struct lpfc_hba *phba)
|
|
{
|
|
/* Reset some HBA SLI4 setup states */
|
|
lpfc_stop_hba_timers(phba);
|
|
phba->pport->work_port_events = 0;
|
|
phba->sli4_hba.intr_enable = 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_stop_port - Wrapper function for stopping hba port
|
|
* @phba: Pointer to HBA context object.
|
|
*
|
|
* This routine wraps the actual SLI3 or SLI4 hba stop port routine from
|
|
* the API jump table function pointer from the lpfc_hba struct.
|
|
**/
|
|
void
|
|
lpfc_stop_port(struct lpfc_hba *phba)
|
|
{
|
|
phba->lpfc_stop_port(phba);
|
|
}
|
|
|
|
/**
|
|
* lpfc_fcf_redisc_wait_start_timer - Start fcf rediscover wait timer
|
|
* @phba: Pointer to hba for which this call is being executed.
|
|
*
|
|
* This routine starts the timer waiting for the FCF rediscovery to complete.
|
|
**/
|
|
void
|
|
lpfc_fcf_redisc_wait_start_timer(struct lpfc_hba *phba)
|
|
{
|
|
unsigned long fcf_redisc_wait_tmo =
|
|
(jiffies + msecs_to_jiffies(LPFC_FCF_REDISCOVER_WAIT_TMO));
|
|
/* Start fcf rediscovery wait period timer */
|
|
mod_timer(&phba->fcf.redisc_wait, fcf_redisc_wait_tmo);
|
|
spin_lock_irq(&phba->hbalock);
|
|
/* Allow action to new fcf asynchronous event */
|
|
phba->fcf.fcf_flag &= ~(FCF_AVAILABLE | FCF_SCAN_DONE);
|
|
/* Mark the FCF rediscovery pending state */
|
|
phba->fcf.fcf_flag |= FCF_REDISC_PEND;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_fcf_redisc_wait_tmo - FCF table rediscover wait timeout
|
|
* @ptr: Map to lpfc_hba data structure pointer.
|
|
*
|
|
* This routine is invoked when waiting for FCF table rediscover has been
|
|
* timed out. If new FCF record(s) has (have) been discovered during the
|
|
* wait period, a new FCF event shall be added to the FCOE async event
|
|
* list, and then worker thread shall be waked up for processing from the
|
|
* worker thread context.
|
|
**/
|
|
void
|
|
lpfc_sli4_fcf_redisc_wait_tmo(unsigned long ptr)
|
|
{
|
|
struct lpfc_hba *phba = (struct lpfc_hba *)ptr;
|
|
|
|
/* Don't send FCF rediscovery event if timer cancelled */
|
|
spin_lock_irq(&phba->hbalock);
|
|
if (!(phba->fcf.fcf_flag & FCF_REDISC_PEND)) {
|
|
spin_unlock_irq(&phba->hbalock);
|
|
return;
|
|
}
|
|
/* Clear FCF rediscovery timer pending flag */
|
|
phba->fcf.fcf_flag &= ~FCF_REDISC_PEND;
|
|
/* FCF rediscovery event to worker thread */
|
|
phba->fcf.fcf_flag |= FCF_REDISC_EVT;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_FIP,
|
|
"2776 FCF rediscover quiescent timer expired\n");
|
|
/* wake up worker thread */
|
|
lpfc_worker_wake_up(phba);
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_fw_cfg_check - Read the firmware config and verify FCoE support
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This function uses the QUERY_FW_CFG mailbox command to determine if the
|
|
* firmware loaded supports FCoE. A return of zero indicates that the mailbox
|
|
* was successful and the firmware supports FCoE. Any other return indicates
|
|
* a error. It is assumed that this function will be called before interrupts
|
|
* are enabled.
|
|
**/
|
|
static int
|
|
lpfc_sli4_fw_cfg_check(struct lpfc_hba *phba)
|
|
{
|
|
int rc = 0;
|
|
LPFC_MBOXQ_t *mboxq;
|
|
struct lpfc_mbx_query_fw_cfg *query_fw_cfg;
|
|
uint32_t length;
|
|
uint32_t shdr_status, shdr_add_status;
|
|
|
|
mboxq = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
|
|
if (!mboxq) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2621 Failed to allocate mbox for "
|
|
"query firmware config cmd\n");
|
|
return -ENOMEM;
|
|
}
|
|
query_fw_cfg = &mboxq->u.mqe.un.query_fw_cfg;
|
|
length = (sizeof(struct lpfc_mbx_query_fw_cfg) -
|
|
sizeof(struct lpfc_sli4_cfg_mhdr));
|
|
lpfc_sli4_config(phba, mboxq, LPFC_MBOX_SUBSYSTEM_COMMON,
|
|
LPFC_MBOX_OPCODE_QUERY_FW_CFG,
|
|
length, LPFC_SLI4_MBX_EMBED);
|
|
rc = lpfc_sli_issue_mbox(phba, mboxq, MBX_POLL);
|
|
/* The IOCTL status is embedded in the mailbox subheader. */
|
|
shdr_status = bf_get(lpfc_mbox_hdr_status,
|
|
&query_fw_cfg->header.cfg_shdr.response);
|
|
shdr_add_status = bf_get(lpfc_mbox_hdr_add_status,
|
|
&query_fw_cfg->header.cfg_shdr.response);
|
|
if (shdr_status || shdr_add_status || rc != MBX_SUCCESS) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
|
|
"2622 Query Firmware Config failed "
|
|
"mbx status x%x, status x%x add_status x%x\n",
|
|
rc, shdr_status, shdr_add_status);
|
|
return -EINVAL;
|
|
}
|
|
if (!bf_get(lpfc_function_mode_fcoe_i, query_fw_cfg)) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
|
|
"2623 FCoE Function not supported by firmware. "
|
|
"Function mode = %08x\n",
|
|
query_fw_cfg->function_mode);
|
|
return -EINVAL;
|
|
}
|
|
if (rc != MBX_TIMEOUT)
|
|
mempool_free(mboxq, phba->mbox_mem_pool);
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_parse_latt_fault - Parse sli4 link-attention link fault code
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @acqe_link: pointer to the async link completion queue entry.
|
|
*
|
|
* This routine is to parse the SLI4 link-attention link fault code and
|
|
* translate it into the base driver's read link attention mailbox command
|
|
* status.
|
|
*
|
|
* Return: Link-attention status in terms of base driver's coding.
|
|
**/
|
|
static uint16_t
|
|
lpfc_sli4_parse_latt_fault(struct lpfc_hba *phba,
|
|
struct lpfc_acqe_link *acqe_link)
|
|
{
|
|
uint16_t latt_fault;
|
|
|
|
switch (bf_get(lpfc_acqe_link_fault, acqe_link)) {
|
|
case LPFC_ASYNC_LINK_FAULT_NONE:
|
|
case LPFC_ASYNC_LINK_FAULT_LOCAL:
|
|
case LPFC_ASYNC_LINK_FAULT_REMOTE:
|
|
latt_fault = 0;
|
|
break;
|
|
default:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0398 Invalid link fault code: x%x\n",
|
|
bf_get(lpfc_acqe_link_fault, acqe_link));
|
|
latt_fault = MBXERR_ERROR;
|
|
break;
|
|
}
|
|
return latt_fault;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_parse_latt_type - Parse sli4 link attention type
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @acqe_link: pointer to the async link completion queue entry.
|
|
*
|
|
* This routine is to parse the SLI4 link attention type and translate it
|
|
* into the base driver's link attention type coding.
|
|
*
|
|
* Return: Link attention type in terms of base driver's coding.
|
|
**/
|
|
static uint8_t
|
|
lpfc_sli4_parse_latt_type(struct lpfc_hba *phba,
|
|
struct lpfc_acqe_link *acqe_link)
|
|
{
|
|
uint8_t att_type;
|
|
|
|
switch (bf_get(lpfc_acqe_link_status, acqe_link)) {
|
|
case LPFC_ASYNC_LINK_STATUS_DOWN:
|
|
case LPFC_ASYNC_LINK_STATUS_LOGICAL_DOWN:
|
|
att_type = AT_LINK_DOWN;
|
|
break;
|
|
case LPFC_ASYNC_LINK_STATUS_UP:
|
|
/* Ignore physical link up events - wait for logical link up */
|
|
att_type = AT_RESERVED;
|
|
break;
|
|
case LPFC_ASYNC_LINK_STATUS_LOGICAL_UP:
|
|
att_type = AT_LINK_UP;
|
|
break;
|
|
default:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0399 Invalid link attention type: x%x\n",
|
|
bf_get(lpfc_acqe_link_status, acqe_link));
|
|
att_type = AT_RESERVED;
|
|
break;
|
|
}
|
|
return att_type;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_parse_latt_link_speed - Parse sli4 link-attention link speed
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @acqe_link: pointer to the async link completion queue entry.
|
|
*
|
|
* This routine is to parse the SLI4 link-attention link speed and translate
|
|
* it into the base driver's link-attention link speed coding.
|
|
*
|
|
* Return: Link-attention link speed in terms of base driver's coding.
|
|
**/
|
|
static uint8_t
|
|
lpfc_sli4_parse_latt_link_speed(struct lpfc_hba *phba,
|
|
struct lpfc_acqe_link *acqe_link)
|
|
{
|
|
uint8_t link_speed;
|
|
|
|
switch (bf_get(lpfc_acqe_link_speed, acqe_link)) {
|
|
case LPFC_ASYNC_LINK_SPEED_ZERO:
|
|
link_speed = LA_UNKNW_LINK;
|
|
break;
|
|
case LPFC_ASYNC_LINK_SPEED_10MBPS:
|
|
link_speed = LA_UNKNW_LINK;
|
|
break;
|
|
case LPFC_ASYNC_LINK_SPEED_100MBPS:
|
|
link_speed = LA_UNKNW_LINK;
|
|
break;
|
|
case LPFC_ASYNC_LINK_SPEED_1GBPS:
|
|
link_speed = LA_1GHZ_LINK;
|
|
break;
|
|
case LPFC_ASYNC_LINK_SPEED_10GBPS:
|
|
link_speed = LA_10GHZ_LINK;
|
|
break;
|
|
default:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0483 Invalid link-attention link speed: x%x\n",
|
|
bf_get(lpfc_acqe_link_speed, acqe_link));
|
|
link_speed = LA_UNKNW_LINK;
|
|
break;
|
|
}
|
|
return link_speed;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_async_link_evt - Process the asynchronous link event
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @acqe_link: pointer to the async link completion queue entry.
|
|
*
|
|
* This routine is to handle the SLI4 asynchronous link event.
|
|
**/
|
|
static void
|
|
lpfc_sli4_async_link_evt(struct lpfc_hba *phba,
|
|
struct lpfc_acqe_link *acqe_link)
|
|
{
|
|
struct lpfc_dmabuf *mp;
|
|
LPFC_MBOXQ_t *pmb;
|
|
MAILBOX_t *mb;
|
|
READ_LA_VAR *la;
|
|
uint8_t att_type;
|
|
|
|
att_type = lpfc_sli4_parse_latt_type(phba, acqe_link);
|
|
if (att_type != AT_LINK_DOWN && att_type != AT_LINK_UP)
|
|
return;
|
|
phba->fcoe_eventtag = acqe_link->event_tag;
|
|
pmb = (LPFC_MBOXQ_t *)mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
|
|
if (!pmb) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
|
|
"0395 The mboxq allocation failed\n");
|
|
return;
|
|
}
|
|
mp = kmalloc(sizeof(struct lpfc_dmabuf), GFP_KERNEL);
|
|
if (!mp) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
|
|
"0396 The lpfc_dmabuf allocation failed\n");
|
|
goto out_free_pmb;
|
|
}
|
|
mp->virt = lpfc_mbuf_alloc(phba, 0, &mp->phys);
|
|
if (!mp->virt) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
|
|
"0397 The mbuf allocation failed\n");
|
|
goto out_free_dmabuf;
|
|
}
|
|
|
|
/* Cleanup any outstanding ELS commands */
|
|
lpfc_els_flush_all_cmd(phba);
|
|
|
|
/* Block ELS IOCBs until we have done process link event */
|
|
phba->sli.ring[LPFC_ELS_RING].flag |= LPFC_STOP_IOCB_EVENT;
|
|
|
|
/* Update link event statistics */
|
|
phba->sli.slistat.link_event++;
|
|
|
|
/* Create pseudo lpfc_handle_latt mailbox command from link ACQE */
|
|
lpfc_read_la(phba, pmb, mp);
|
|
pmb->vport = phba->pport;
|
|
|
|
/* Parse and translate status field */
|
|
mb = &pmb->u.mb;
|
|
mb->mbxStatus = lpfc_sli4_parse_latt_fault(phba, acqe_link);
|
|
|
|
/* Parse and translate link attention fields */
|
|
la = (READ_LA_VAR *) &pmb->u.mb.un.varReadLA;
|
|
la->eventTag = acqe_link->event_tag;
|
|
la->attType = att_type;
|
|
la->UlnkSpeed = lpfc_sli4_parse_latt_link_speed(phba, acqe_link);
|
|
|
|
/* Fake the the following irrelvant fields */
|
|
la->topology = TOPOLOGY_PT_PT;
|
|
la->granted_AL_PA = 0;
|
|
la->il = 0;
|
|
la->pb = 0;
|
|
la->fa = 0;
|
|
la->mm = 0;
|
|
|
|
/* Keep the link status for extra SLI4 state machine reference */
|
|
phba->sli4_hba.link_state.speed =
|
|
bf_get(lpfc_acqe_link_speed, acqe_link);
|
|
phba->sli4_hba.link_state.duplex =
|
|
bf_get(lpfc_acqe_link_duplex, acqe_link);
|
|
phba->sli4_hba.link_state.status =
|
|
bf_get(lpfc_acqe_link_status, acqe_link);
|
|
phba->sli4_hba.link_state.physical =
|
|
bf_get(lpfc_acqe_link_physical, acqe_link);
|
|
phba->sli4_hba.link_state.fault =
|
|
bf_get(lpfc_acqe_link_fault, acqe_link);
|
|
phba->sli4_hba.link_state.logical_speed =
|
|
bf_get(lpfc_acqe_qos_link_speed, acqe_link);
|
|
|
|
/* Invoke the lpfc_handle_latt mailbox command callback function */
|
|
lpfc_mbx_cmpl_read_la(phba, pmb);
|
|
|
|
return;
|
|
|
|
out_free_dmabuf:
|
|
kfree(mp);
|
|
out_free_pmb:
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_perform_vport_cvl - Perform clear virtual link on a vport
|
|
* @vport: pointer to vport data structure.
|
|
*
|
|
* This routine is to perform Clear Virtual Link (CVL) on a vport in
|
|
* response to a CVL event.
|
|
*
|
|
* Return the pointer to the ndlp with the vport if successful, otherwise
|
|
* return NULL.
|
|
**/
|
|
static struct lpfc_nodelist *
|
|
lpfc_sli4_perform_vport_cvl(struct lpfc_vport *vport)
|
|
{
|
|
struct lpfc_nodelist *ndlp;
|
|
struct Scsi_Host *shost;
|
|
struct lpfc_hba *phba;
|
|
|
|
if (!vport)
|
|
return NULL;
|
|
phba = vport->phba;
|
|
if (!phba)
|
|
return NULL;
|
|
ndlp = lpfc_findnode_did(vport, Fabric_DID);
|
|
if (!ndlp) {
|
|
/* Cannot find existing Fabric ndlp, so allocate a new one */
|
|
ndlp = mempool_alloc(phba->nlp_mem_pool, GFP_KERNEL);
|
|
if (!ndlp)
|
|
return 0;
|
|
lpfc_nlp_init(vport, ndlp, Fabric_DID);
|
|
/* Set the node type */
|
|
ndlp->nlp_type |= NLP_FABRIC;
|
|
/* Put ndlp onto node list */
|
|
lpfc_enqueue_node(vport, ndlp);
|
|
} else if (!NLP_CHK_NODE_ACT(ndlp)) {
|
|
/* re-setup ndlp without removing from node list */
|
|
ndlp = lpfc_enable_node(vport, ndlp, NLP_STE_UNUSED_NODE);
|
|
if (!ndlp)
|
|
return 0;
|
|
}
|
|
if (phba->pport->port_state < LPFC_FLOGI)
|
|
return NULL;
|
|
/* If virtual link is not yet instantiated ignore CVL */
|
|
if ((vport != phba->pport) && (vport->port_state < LPFC_FDISC))
|
|
return NULL;
|
|
shost = lpfc_shost_from_vport(vport);
|
|
if (!shost)
|
|
return NULL;
|
|
lpfc_linkdown_port(vport);
|
|
lpfc_cleanup_pending_mbox(vport);
|
|
spin_lock_irq(shost->host_lock);
|
|
vport->fc_flag |= FC_VPORT_CVL_RCVD;
|
|
spin_unlock_irq(shost->host_lock);
|
|
|
|
return ndlp;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_perform_all_vport_cvl - Perform clear virtual link on all vports
|
|
* @vport: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is to perform Clear Virtual Link (CVL) on all vports in
|
|
* response to a FCF dead event.
|
|
**/
|
|
static void
|
|
lpfc_sli4_perform_all_vport_cvl(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_vport **vports;
|
|
int i;
|
|
|
|
vports = lpfc_create_vport_work_array(phba);
|
|
if (vports)
|
|
for (i = 0; i <= phba->max_vports && vports[i] != NULL; i++)
|
|
lpfc_sli4_perform_vport_cvl(vports[i]);
|
|
lpfc_destroy_vport_work_array(phba, vports);
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_async_fcoe_evt - Process the asynchronous fcoe event
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @acqe_link: pointer to the async fcoe completion queue entry.
|
|
*
|
|
* This routine is to handle the SLI4 asynchronous fcoe event.
|
|
**/
|
|
static void
|
|
lpfc_sli4_async_fcoe_evt(struct lpfc_hba *phba,
|
|
struct lpfc_acqe_fcoe *acqe_fcoe)
|
|
{
|
|
uint8_t event_type = bf_get(lpfc_acqe_fcoe_event_type, acqe_fcoe);
|
|
int rc;
|
|
struct lpfc_vport *vport;
|
|
struct lpfc_nodelist *ndlp;
|
|
struct Scsi_Host *shost;
|
|
int active_vlink_present;
|
|
struct lpfc_vport **vports;
|
|
int i;
|
|
|
|
phba->fc_eventTag = acqe_fcoe->event_tag;
|
|
phba->fcoe_eventtag = acqe_fcoe->event_tag;
|
|
switch (event_type) {
|
|
case LPFC_FCOE_EVENT_TYPE_NEW_FCF:
|
|
case LPFC_FCOE_EVENT_TYPE_FCF_PARAM_MOD:
|
|
if (event_type == LPFC_FCOE_EVENT_TYPE_NEW_FCF)
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_FIP |
|
|
LOG_DISCOVERY,
|
|
"2546 New FCF event, evt_tag:x%x, "
|
|
"index:x%x\n",
|
|
acqe_fcoe->event_tag,
|
|
acqe_fcoe->index);
|
|
else
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_FIP |
|
|
LOG_DISCOVERY,
|
|
"2788 FCF param modified event, "
|
|
"evt_tag:x%x, index:x%x\n",
|
|
acqe_fcoe->event_tag,
|
|
acqe_fcoe->index);
|
|
if (phba->fcf.fcf_flag & FCF_DISCOVERY) {
|
|
/*
|
|
* During period of FCF discovery, read the FCF
|
|
* table record indexed by the event to update
|
|
* FCF roundrobin failover eligible FCF bmask.
|
|
*/
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_FIP |
|
|
LOG_DISCOVERY,
|
|
"2779 Read FCF (x%x) for updating "
|
|
"roundrobin FCF failover bmask\n",
|
|
acqe_fcoe->index);
|
|
rc = lpfc_sli4_read_fcf_rec(phba, acqe_fcoe->index);
|
|
}
|
|
|
|
/* If the FCF discovery is in progress, do nothing. */
|
|
spin_lock_irq(&phba->hbalock);
|
|
if (phba->hba_flag & FCF_TS_INPROG) {
|
|
spin_unlock_irq(&phba->hbalock);
|
|
break;
|
|
}
|
|
/* If fast FCF failover rescan event is pending, do nothing */
|
|
if (phba->fcf.fcf_flag & FCF_REDISC_EVT) {
|
|
spin_unlock_irq(&phba->hbalock);
|
|
break;
|
|
}
|
|
|
|
/* If the FCF has been in discovered state, do nothing. */
|
|
if (phba->fcf.fcf_flag & FCF_SCAN_DONE) {
|
|
spin_unlock_irq(&phba->hbalock);
|
|
break;
|
|
}
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
/* Otherwise, scan the entire FCF table and re-discover SAN */
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_FIP | LOG_DISCOVERY,
|
|
"2770 Start FCF table scan per async FCF "
|
|
"event, evt_tag:x%x, index:x%x\n",
|
|
acqe_fcoe->event_tag, acqe_fcoe->index);
|
|
rc = lpfc_sli4_fcf_scan_read_fcf_rec(phba,
|
|
LPFC_FCOE_FCF_GET_FIRST);
|
|
if (rc)
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_FIP | LOG_DISCOVERY,
|
|
"2547 Issue FCF scan read FCF mailbox "
|
|
"command failed (x%x)\n", rc);
|
|
break;
|
|
|
|
case LPFC_FCOE_EVENT_TYPE_FCF_TABLE_FULL:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
|
|
"2548 FCF Table full count 0x%x tag 0x%x\n",
|
|
bf_get(lpfc_acqe_fcoe_fcf_count, acqe_fcoe),
|
|
acqe_fcoe->event_tag);
|
|
break;
|
|
|
|
case LPFC_FCOE_EVENT_TYPE_FCF_DEAD:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_FIP | LOG_DISCOVERY,
|
|
"2549 FCF (x%x) disconnected from network, "
|
|
"tag:x%x\n", acqe_fcoe->index, acqe_fcoe->event_tag);
|
|
/*
|
|
* If we are in the middle of FCF failover process, clear
|
|
* the corresponding FCF bit in the roundrobin bitmap.
|
|
*/
|
|
spin_lock_irq(&phba->hbalock);
|
|
if (phba->fcf.fcf_flag & FCF_DISCOVERY) {
|
|
spin_unlock_irq(&phba->hbalock);
|
|
/* Update FLOGI FCF failover eligible FCF bmask */
|
|
lpfc_sli4_fcf_rr_index_clear(phba, acqe_fcoe->index);
|
|
break;
|
|
}
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
/* If the event is not for currently used fcf do nothing */
|
|
if (phba->fcf.current_rec.fcf_indx != acqe_fcoe->index)
|
|
break;
|
|
|
|
/*
|
|
* Otherwise, request the port to rediscover the entire FCF
|
|
* table for a fast recovery from case that the current FCF
|
|
* is no longer valid as we are not in the middle of FCF
|
|
* failover process already.
|
|
*/
|
|
spin_lock_irq(&phba->hbalock);
|
|
/* Mark the fast failover process in progress */
|
|
phba->fcf.fcf_flag |= FCF_DEAD_DISC;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_FIP | LOG_DISCOVERY,
|
|
"2771 Start FCF fast failover process due to "
|
|
"FCF DEAD event: evt_tag:x%x, fcf_index:x%x "
|
|
"\n", acqe_fcoe->event_tag, acqe_fcoe->index);
|
|
rc = lpfc_sli4_redisc_fcf_table(phba);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_FIP |
|
|
LOG_DISCOVERY,
|
|
"2772 Issue FCF rediscover mabilbox "
|
|
"command failed, fail through to FCF "
|
|
"dead event\n");
|
|
spin_lock_irq(&phba->hbalock);
|
|
phba->fcf.fcf_flag &= ~FCF_DEAD_DISC;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
/*
|
|
* Last resort will fail over by treating this
|
|
* as a link down to FCF registration.
|
|
*/
|
|
lpfc_sli4_fcf_dead_failthrough(phba);
|
|
} else {
|
|
/* Reset FCF roundrobin bmask for new discovery */
|
|
memset(phba->fcf.fcf_rr_bmask, 0,
|
|
sizeof(*phba->fcf.fcf_rr_bmask));
|
|
/*
|
|
* Handling fast FCF failover to a DEAD FCF event is
|
|
* considered equalivant to receiving CVL to all vports.
|
|
*/
|
|
lpfc_sli4_perform_all_vport_cvl(phba);
|
|
}
|
|
break;
|
|
case LPFC_FCOE_EVENT_TYPE_CVL:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_FIP | LOG_DISCOVERY,
|
|
"2718 Clear Virtual Link Received for VPI 0x%x"
|
|
" tag 0x%x\n", acqe_fcoe->index, acqe_fcoe->event_tag);
|
|
vport = lpfc_find_vport_by_vpid(phba,
|
|
acqe_fcoe->index - phba->vpi_base);
|
|
ndlp = lpfc_sli4_perform_vport_cvl(vport);
|
|
if (!ndlp)
|
|
break;
|
|
active_vlink_present = 0;
|
|
|
|
vports = lpfc_create_vport_work_array(phba);
|
|
if (vports) {
|
|
for (i = 0; i <= phba->max_vports && vports[i] != NULL;
|
|
i++) {
|
|
if ((!(vports[i]->fc_flag &
|
|
FC_VPORT_CVL_RCVD)) &&
|
|
(vports[i]->port_state > LPFC_FDISC)) {
|
|
active_vlink_present = 1;
|
|
break;
|
|
}
|
|
}
|
|
lpfc_destroy_vport_work_array(phba, vports);
|
|
}
|
|
|
|
if (active_vlink_present) {
|
|
/*
|
|
* If there are other active VLinks present,
|
|
* re-instantiate the Vlink using FDISC.
|
|
*/
|
|
mod_timer(&ndlp->nlp_delayfunc, jiffies + HZ);
|
|
shost = lpfc_shost_from_vport(vport);
|
|
spin_lock_irq(shost->host_lock);
|
|
ndlp->nlp_flag |= NLP_DELAY_TMO;
|
|
spin_unlock_irq(shost->host_lock);
|
|
ndlp->nlp_last_elscmd = ELS_CMD_FDISC;
|
|
vport->port_state = LPFC_FDISC;
|
|
} else {
|
|
/*
|
|
* Otherwise, we request port to rediscover
|
|
* the entire FCF table for a fast recovery
|
|
* from possible case that the current FCF
|
|
* is no longer valid if we are not already
|
|
* in the FCF failover process.
|
|
*/
|
|
spin_lock_irq(&phba->hbalock);
|
|
if (phba->fcf.fcf_flag & FCF_DISCOVERY) {
|
|
spin_unlock_irq(&phba->hbalock);
|
|
break;
|
|
}
|
|
/* Mark the fast failover process in progress */
|
|
phba->fcf.fcf_flag |= FCF_ACVL_DISC;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_FIP |
|
|
LOG_DISCOVERY,
|
|
"2773 Start FCF failover per CVL, "
|
|
"evt_tag:x%x\n", acqe_fcoe->event_tag);
|
|
rc = lpfc_sli4_redisc_fcf_table(phba);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_FIP |
|
|
LOG_DISCOVERY,
|
|
"2774 Issue FCF rediscover "
|
|
"mabilbox command failed, "
|
|
"through to CVL event\n");
|
|
spin_lock_irq(&phba->hbalock);
|
|
phba->fcf.fcf_flag &= ~FCF_ACVL_DISC;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
/*
|
|
* Last resort will be re-try on the
|
|
* the current registered FCF entry.
|
|
*/
|
|
lpfc_retry_pport_discovery(phba);
|
|
} else
|
|
/*
|
|
* Reset FCF roundrobin bmask for new
|
|
* discovery.
|
|
*/
|
|
memset(phba->fcf.fcf_rr_bmask, 0,
|
|
sizeof(*phba->fcf.fcf_rr_bmask));
|
|
}
|
|
break;
|
|
default:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
|
|
"0288 Unknown FCoE event type 0x%x event tag "
|
|
"0x%x\n", event_type, acqe_fcoe->event_tag);
|
|
break;
|
|
}
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_async_dcbx_evt - Process the asynchronous dcbx event
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @acqe_link: pointer to the async dcbx completion queue entry.
|
|
*
|
|
* This routine is to handle the SLI4 asynchronous dcbx event.
|
|
**/
|
|
static void
|
|
lpfc_sli4_async_dcbx_evt(struct lpfc_hba *phba,
|
|
struct lpfc_acqe_dcbx *acqe_dcbx)
|
|
{
|
|
phba->fc_eventTag = acqe_dcbx->event_tag;
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
|
|
"0290 The SLI4 DCBX asynchronous event is not "
|
|
"handled yet\n");
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_async_grp5_evt - Process the asynchronous group5 event
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @acqe_link: pointer to the async grp5 completion queue entry.
|
|
*
|
|
* This routine is to handle the SLI4 asynchronous grp5 event. A grp5 event
|
|
* is an asynchronous notified of a logical link speed change. The Port
|
|
* reports the logical link speed in units of 10Mbps.
|
|
**/
|
|
static void
|
|
lpfc_sli4_async_grp5_evt(struct lpfc_hba *phba,
|
|
struct lpfc_acqe_grp5 *acqe_grp5)
|
|
{
|
|
uint16_t prev_ll_spd;
|
|
|
|
phba->fc_eventTag = acqe_grp5->event_tag;
|
|
phba->fcoe_eventtag = acqe_grp5->event_tag;
|
|
prev_ll_spd = phba->sli4_hba.link_state.logical_speed;
|
|
phba->sli4_hba.link_state.logical_speed =
|
|
(bf_get(lpfc_acqe_grp5_llink_spd, acqe_grp5));
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_SLI,
|
|
"2789 GRP5 Async Event: Updating logical link speed "
|
|
"from %dMbps to %dMbps\n", (prev_ll_spd * 10),
|
|
(phba->sli4_hba.link_state.logical_speed*10));
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_async_event_proc - Process all the pending asynchronous event
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked by the worker thread to process all the pending
|
|
* SLI4 asynchronous events.
|
|
**/
|
|
void lpfc_sli4_async_event_proc(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_cq_event *cq_event;
|
|
|
|
/* First, declare the async event has been handled */
|
|
spin_lock_irq(&phba->hbalock);
|
|
phba->hba_flag &= ~ASYNC_EVENT;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
/* Now, handle all the async events */
|
|
while (!list_empty(&phba->sli4_hba.sp_asynce_work_queue)) {
|
|
/* Get the first event from the head of the event queue */
|
|
spin_lock_irq(&phba->hbalock);
|
|
list_remove_head(&phba->sli4_hba.sp_asynce_work_queue,
|
|
cq_event, struct lpfc_cq_event, list);
|
|
spin_unlock_irq(&phba->hbalock);
|
|
/* Process the asynchronous event */
|
|
switch (bf_get(lpfc_trailer_code, &cq_event->cqe.mcqe_cmpl)) {
|
|
case LPFC_TRAILER_CODE_LINK:
|
|
lpfc_sli4_async_link_evt(phba,
|
|
&cq_event->cqe.acqe_link);
|
|
break;
|
|
case LPFC_TRAILER_CODE_FCOE:
|
|
lpfc_sli4_async_fcoe_evt(phba,
|
|
&cq_event->cqe.acqe_fcoe);
|
|
break;
|
|
case LPFC_TRAILER_CODE_DCBX:
|
|
lpfc_sli4_async_dcbx_evt(phba,
|
|
&cq_event->cqe.acqe_dcbx);
|
|
break;
|
|
case LPFC_TRAILER_CODE_GRP5:
|
|
lpfc_sli4_async_grp5_evt(phba,
|
|
&cq_event->cqe.acqe_grp5);
|
|
break;
|
|
default:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
|
|
"1804 Invalid asynchrous event code: "
|
|
"x%x\n", bf_get(lpfc_trailer_code,
|
|
&cq_event->cqe.mcqe_cmpl));
|
|
break;
|
|
}
|
|
/* Free the completion event processed to the free pool */
|
|
lpfc_sli4_cq_event_release(phba, cq_event);
|
|
}
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_fcf_redisc_event_proc - Process fcf table rediscovery event
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked by the worker thread to process FCF table
|
|
* rediscovery pending completion event.
|
|
**/
|
|
void lpfc_sli4_fcf_redisc_event_proc(struct lpfc_hba *phba)
|
|
{
|
|
int rc;
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
/* Clear FCF rediscovery timeout event */
|
|
phba->fcf.fcf_flag &= ~FCF_REDISC_EVT;
|
|
/* Clear driver fast failover FCF record flag */
|
|
phba->fcf.failover_rec.flag = 0;
|
|
/* Set state for FCF fast failover */
|
|
phba->fcf.fcf_flag |= FCF_REDISC_FOV;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
/* Scan FCF table from the first entry to re-discover SAN */
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_FIP | LOG_DISCOVERY,
|
|
"2777 Start post-quiescent FCF table scan\n");
|
|
rc = lpfc_sli4_fcf_scan_read_fcf_rec(phba, LPFC_FCOE_FCF_GET_FIRST);
|
|
if (rc)
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_FIP | LOG_DISCOVERY,
|
|
"2747 Issue FCF scan read FCF mailbox "
|
|
"command failed 0x%x\n", rc);
|
|
}
|
|
|
|
/**
|
|
* lpfc_api_table_setup - Set up per hba pci-device group func api jump table
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @dev_grp: The HBA PCI-Device group number.
|
|
*
|
|
* This routine is invoked to set up the per HBA PCI-Device group function
|
|
* API jump table entries.
|
|
*
|
|
* Return: 0 if success, otherwise -ENODEV
|
|
**/
|
|
int
|
|
lpfc_api_table_setup(struct lpfc_hba *phba, uint8_t dev_grp)
|
|
{
|
|
int rc;
|
|
|
|
/* Set up lpfc PCI-device group */
|
|
phba->pci_dev_grp = dev_grp;
|
|
|
|
/* The LPFC_PCI_DEV_OC uses SLI4 */
|
|
if (dev_grp == LPFC_PCI_DEV_OC)
|
|
phba->sli_rev = LPFC_SLI_REV4;
|
|
|
|
/* Set up device INIT API function jump table */
|
|
rc = lpfc_init_api_table_setup(phba, dev_grp);
|
|
if (rc)
|
|
return -ENODEV;
|
|
/* Set up SCSI API function jump table */
|
|
rc = lpfc_scsi_api_table_setup(phba, dev_grp);
|
|
if (rc)
|
|
return -ENODEV;
|
|
/* Set up SLI API function jump table */
|
|
rc = lpfc_sli_api_table_setup(phba, dev_grp);
|
|
if (rc)
|
|
return -ENODEV;
|
|
/* Set up MBOX API function jump table */
|
|
rc = lpfc_mbox_api_table_setup(phba, dev_grp);
|
|
if (rc)
|
|
return -ENODEV;
|
|
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_log_intr_mode - Log the active interrupt mode
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @intr_mode: active interrupt mode adopted.
|
|
*
|
|
* This routine it invoked to log the currently used active interrupt mode
|
|
* to the device.
|
|
**/
|
|
static void lpfc_log_intr_mode(struct lpfc_hba *phba, uint32_t intr_mode)
|
|
{
|
|
switch (intr_mode) {
|
|
case 0:
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0470 Enable INTx interrupt mode.\n");
|
|
break;
|
|
case 1:
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0481 Enabled MSI interrupt mode.\n");
|
|
break;
|
|
case 2:
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0480 Enabled MSI-X interrupt mode.\n");
|
|
break;
|
|
default:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0482 Illegal interrupt mode.\n");
|
|
break;
|
|
}
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_enable_pci_dev - Enable a generic PCI device.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to enable the PCI device that is common to all
|
|
* PCI devices.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* other values - error
|
|
**/
|
|
static int
|
|
lpfc_enable_pci_dev(struct lpfc_hba *phba)
|
|
{
|
|
struct pci_dev *pdev;
|
|
int bars;
|
|
|
|
/* Obtain PCI device reference */
|
|
if (!phba->pcidev)
|
|
goto out_error;
|
|
else
|
|
pdev = phba->pcidev;
|
|
/* Select PCI BARs */
|
|
bars = pci_select_bars(pdev, IORESOURCE_MEM);
|
|
/* Enable PCI device */
|
|
if (pci_enable_device_mem(pdev))
|
|
goto out_error;
|
|
/* Request PCI resource for the device */
|
|
if (pci_request_selected_regions(pdev, bars, LPFC_DRIVER_NAME))
|
|
goto out_disable_device;
|
|
/* Set up device as PCI master and save state for EEH */
|
|
pci_set_master(pdev);
|
|
pci_try_set_mwi(pdev);
|
|
pci_save_state(pdev);
|
|
|
|
return 0;
|
|
|
|
out_disable_device:
|
|
pci_disable_device(pdev);
|
|
out_error:
|
|
return -ENODEV;
|
|
}
|
|
|
|
/**
|
|
* lpfc_disable_pci_dev - Disable a generic PCI device.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to disable the PCI device that is common to all
|
|
* PCI devices.
|
|
**/
|
|
static void
|
|
lpfc_disable_pci_dev(struct lpfc_hba *phba)
|
|
{
|
|
struct pci_dev *pdev;
|
|
int bars;
|
|
|
|
/* Obtain PCI device reference */
|
|
if (!phba->pcidev)
|
|
return;
|
|
else
|
|
pdev = phba->pcidev;
|
|
/* Select PCI BARs */
|
|
bars = pci_select_bars(pdev, IORESOURCE_MEM);
|
|
/* Release PCI resource and disable PCI device */
|
|
pci_release_selected_regions(pdev, bars);
|
|
pci_disable_device(pdev);
|
|
/* Null out PCI private reference to driver */
|
|
pci_set_drvdata(pdev, NULL);
|
|
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_reset_hba - Reset a hba
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to reset a hba device. It brings the HBA
|
|
* offline, performs a board restart, and then brings the board back
|
|
* online. The lpfc_offline calls lpfc_sli_hba_down which will clean up
|
|
* on outstanding mailbox commands.
|
|
**/
|
|
void
|
|
lpfc_reset_hba(struct lpfc_hba *phba)
|
|
{
|
|
/* If resets are disabled then set error state and return. */
|
|
if (!phba->cfg_enable_hba_reset) {
|
|
phba->link_state = LPFC_HBA_ERROR;
|
|
return;
|
|
}
|
|
lpfc_offline_prep(phba);
|
|
lpfc_offline(phba);
|
|
lpfc_sli_brdrestart(phba);
|
|
lpfc_online(phba);
|
|
lpfc_unblock_mgmt_io(phba);
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli_driver_resource_setup - Setup driver internal resources for SLI3 dev.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to set up the driver internal resources specific to
|
|
* support the SLI-3 HBA device it attached to.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* other values - error
|
|
**/
|
|
static int
|
|
lpfc_sli_driver_resource_setup(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_sli *psli;
|
|
|
|
/*
|
|
* Initialize timers used by driver
|
|
*/
|
|
|
|
/* Heartbeat timer */
|
|
init_timer(&phba->hb_tmofunc);
|
|
phba->hb_tmofunc.function = lpfc_hb_timeout;
|
|
phba->hb_tmofunc.data = (unsigned long)phba;
|
|
|
|
psli = &phba->sli;
|
|
/* MBOX heartbeat timer */
|
|
init_timer(&psli->mbox_tmo);
|
|
psli->mbox_tmo.function = lpfc_mbox_timeout;
|
|
psli->mbox_tmo.data = (unsigned long) phba;
|
|
/* FCP polling mode timer */
|
|
init_timer(&phba->fcp_poll_timer);
|
|
phba->fcp_poll_timer.function = lpfc_poll_timeout;
|
|
phba->fcp_poll_timer.data = (unsigned long) phba;
|
|
/* Fabric block timer */
|
|
init_timer(&phba->fabric_block_timer);
|
|
phba->fabric_block_timer.function = lpfc_fabric_block_timeout;
|
|
phba->fabric_block_timer.data = (unsigned long) phba;
|
|
/* EA polling mode timer */
|
|
init_timer(&phba->eratt_poll);
|
|
phba->eratt_poll.function = lpfc_poll_eratt;
|
|
phba->eratt_poll.data = (unsigned long) phba;
|
|
|
|
/* Host attention work mask setup */
|
|
phba->work_ha_mask = (HA_ERATT | HA_MBATT | HA_LATT);
|
|
phba->work_ha_mask |= (HA_RXMASK << (LPFC_ELS_RING * 4));
|
|
|
|
/* Get all the module params for configuring this host */
|
|
lpfc_get_cfgparam(phba);
|
|
if (phba->pcidev->device == PCI_DEVICE_ID_HORNET) {
|
|
phba->menlo_flag |= HBA_MENLO_SUPPORT;
|
|
/* check for menlo minimum sg count */
|
|
if (phba->cfg_sg_seg_cnt < LPFC_DEFAULT_MENLO_SG_SEG_CNT)
|
|
phba->cfg_sg_seg_cnt = LPFC_DEFAULT_MENLO_SG_SEG_CNT;
|
|
}
|
|
|
|
/*
|
|
* Since the sg_tablesize is module parameter, the sg_dma_buf_size
|
|
* used to create the sg_dma_buf_pool must be dynamically calculated.
|
|
* 2 segments are added since the IOCB needs a command and response bde.
|
|
*/
|
|
phba->cfg_sg_dma_buf_size = sizeof(struct fcp_cmnd) +
|
|
sizeof(struct fcp_rsp) +
|
|
((phba->cfg_sg_seg_cnt + 2) * sizeof(struct ulp_bde64));
|
|
|
|
if (phba->cfg_enable_bg) {
|
|
phba->cfg_sg_seg_cnt = LPFC_MAX_SG_SEG_CNT;
|
|
phba->cfg_sg_dma_buf_size +=
|
|
phba->cfg_prot_sg_seg_cnt * sizeof(struct ulp_bde64);
|
|
}
|
|
|
|
/* Also reinitialize the host templates with new values. */
|
|
lpfc_vport_template.sg_tablesize = phba->cfg_sg_seg_cnt;
|
|
lpfc_template.sg_tablesize = phba->cfg_sg_seg_cnt;
|
|
|
|
phba->max_vpi = LPFC_MAX_VPI;
|
|
/* This will be set to correct value after config_port mbox */
|
|
phba->max_vports = 0;
|
|
|
|
/*
|
|
* Initialize the SLI Layer to run with lpfc HBAs.
|
|
*/
|
|
lpfc_sli_setup(phba);
|
|
lpfc_sli_queue_setup(phba);
|
|
|
|
/* Allocate device driver memory */
|
|
if (lpfc_mem_alloc(phba, BPL_ALIGN_SZ))
|
|
return -ENOMEM;
|
|
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli_driver_resource_unset - Unset drvr internal resources for SLI3 dev
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to unset the driver internal resources set up
|
|
* specific for supporting the SLI-3 HBA device it attached to.
|
|
**/
|
|
static void
|
|
lpfc_sli_driver_resource_unset(struct lpfc_hba *phba)
|
|
{
|
|
/* Free device driver memory allocated */
|
|
lpfc_mem_free_all(phba);
|
|
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_driver_resource_setup - Setup drvr internal resources for SLI4 dev
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to set up the driver internal resources specific to
|
|
* support the SLI-4 HBA device it attached to.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* other values - error
|
|
**/
|
|
static int
|
|
lpfc_sli4_driver_resource_setup(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_sli *psli;
|
|
LPFC_MBOXQ_t *mboxq;
|
|
int rc, i, hbq_count, buf_size, dma_buf_size, max_buf_size;
|
|
uint8_t pn_page[LPFC_MAX_SUPPORTED_PAGES] = {0};
|
|
struct lpfc_mqe *mqe;
|
|
int longs;
|
|
|
|
/* Before proceed, wait for POST done and device ready */
|
|
rc = lpfc_sli4_post_status_check(phba);
|
|
if (rc)
|
|
return -ENODEV;
|
|
|
|
/*
|
|
* Initialize timers used by driver
|
|
*/
|
|
|
|
/* Heartbeat timer */
|
|
init_timer(&phba->hb_tmofunc);
|
|
phba->hb_tmofunc.function = lpfc_hb_timeout;
|
|
phba->hb_tmofunc.data = (unsigned long)phba;
|
|
|
|
psli = &phba->sli;
|
|
/* MBOX heartbeat timer */
|
|
init_timer(&psli->mbox_tmo);
|
|
psli->mbox_tmo.function = lpfc_mbox_timeout;
|
|
psli->mbox_tmo.data = (unsigned long) phba;
|
|
/* Fabric block timer */
|
|
init_timer(&phba->fabric_block_timer);
|
|
phba->fabric_block_timer.function = lpfc_fabric_block_timeout;
|
|
phba->fabric_block_timer.data = (unsigned long) phba;
|
|
/* EA polling mode timer */
|
|
init_timer(&phba->eratt_poll);
|
|
phba->eratt_poll.function = lpfc_poll_eratt;
|
|
phba->eratt_poll.data = (unsigned long) phba;
|
|
/* FCF rediscover timer */
|
|
init_timer(&phba->fcf.redisc_wait);
|
|
phba->fcf.redisc_wait.function = lpfc_sli4_fcf_redisc_wait_tmo;
|
|
phba->fcf.redisc_wait.data = (unsigned long)phba;
|
|
|
|
/*
|
|
* We need to do a READ_CONFIG mailbox command here before
|
|
* calling lpfc_get_cfgparam. For VFs this will report the
|
|
* MAX_XRI, MAX_VPI, MAX_RPI, MAX_IOCB, and MAX_VFI settings.
|
|
* All of the resources allocated
|
|
* for this Port are tied to these values.
|
|
*/
|
|
/* Get all the module params for configuring this host */
|
|
lpfc_get_cfgparam(phba);
|
|
phba->max_vpi = LPFC_MAX_VPI;
|
|
/* This will be set to correct value after the read_config mbox */
|
|
phba->max_vports = 0;
|
|
|
|
/* Program the default value of vlan_id and fc_map */
|
|
phba->valid_vlan = 0;
|
|
phba->fc_map[0] = LPFC_FCOE_FCF_MAP0;
|
|
phba->fc_map[1] = LPFC_FCOE_FCF_MAP1;
|
|
phba->fc_map[2] = LPFC_FCOE_FCF_MAP2;
|
|
|
|
/*
|
|
* Since the sg_tablesize is module parameter, the sg_dma_buf_size
|
|
* used to create the sg_dma_buf_pool must be dynamically calculated.
|
|
* 2 segments are added since the IOCB needs a command and response bde.
|
|
* To insure that the scsi sgl does not cross a 4k page boundary only
|
|
* sgl sizes of must be a power of 2.
|
|
*/
|
|
buf_size = (sizeof(struct fcp_cmnd) + sizeof(struct fcp_rsp) +
|
|
((phba->cfg_sg_seg_cnt + 2) * sizeof(struct sli4_sge)));
|
|
/* Feature Level 1 hardware is limited to 2 pages */
|
|
if ((bf_get(lpfc_sli_intf_featurelevel1, &phba->sli4_hba.sli_intf) ==
|
|
LPFC_SLI_INTF_FEATURELEVEL1_1))
|
|
max_buf_size = LPFC_SLI4_FL1_MAX_BUF_SIZE;
|
|
else
|
|
max_buf_size = LPFC_SLI4_MAX_BUF_SIZE;
|
|
for (dma_buf_size = LPFC_SLI4_MIN_BUF_SIZE;
|
|
dma_buf_size < max_buf_size && buf_size > dma_buf_size;
|
|
dma_buf_size = dma_buf_size << 1)
|
|
;
|
|
if (dma_buf_size == max_buf_size)
|
|
phba->cfg_sg_seg_cnt = (dma_buf_size -
|
|
sizeof(struct fcp_cmnd) - sizeof(struct fcp_rsp) -
|
|
(2 * sizeof(struct sli4_sge))) /
|
|
sizeof(struct sli4_sge);
|
|
phba->cfg_sg_dma_buf_size = dma_buf_size;
|
|
|
|
/* Initialize buffer queue management fields */
|
|
hbq_count = lpfc_sli_hbq_count();
|
|
for (i = 0; i < hbq_count; ++i)
|
|
INIT_LIST_HEAD(&phba->hbqs[i].hbq_buffer_list);
|
|
INIT_LIST_HEAD(&phba->rb_pend_list);
|
|
phba->hbqs[LPFC_ELS_HBQ].hbq_alloc_buffer = lpfc_sli4_rb_alloc;
|
|
phba->hbqs[LPFC_ELS_HBQ].hbq_free_buffer = lpfc_sli4_rb_free;
|
|
|
|
/*
|
|
* Initialize the SLI Layer to run with lpfc SLI4 HBAs.
|
|
*/
|
|
/* Initialize the Abort scsi buffer list used by driver */
|
|
spin_lock_init(&phba->sli4_hba.abts_scsi_buf_list_lock);
|
|
INIT_LIST_HEAD(&phba->sli4_hba.lpfc_abts_scsi_buf_list);
|
|
/* This abort list used by worker thread */
|
|
spin_lock_init(&phba->sli4_hba.abts_sgl_list_lock);
|
|
|
|
/*
|
|
* Initialize dirver internal slow-path work queues
|
|
*/
|
|
|
|
/* Driver internel slow-path CQ Event pool */
|
|
INIT_LIST_HEAD(&phba->sli4_hba.sp_cqe_event_pool);
|
|
/* Response IOCB work queue list */
|
|
INIT_LIST_HEAD(&phba->sli4_hba.sp_queue_event);
|
|
/* Asynchronous event CQ Event work queue list */
|
|
INIT_LIST_HEAD(&phba->sli4_hba.sp_asynce_work_queue);
|
|
/* Fast-path XRI aborted CQ Event work queue list */
|
|
INIT_LIST_HEAD(&phba->sli4_hba.sp_fcp_xri_aborted_work_queue);
|
|
/* Slow-path XRI aborted CQ Event work queue list */
|
|
INIT_LIST_HEAD(&phba->sli4_hba.sp_els_xri_aborted_work_queue);
|
|
/* Receive queue CQ Event work queue list */
|
|
INIT_LIST_HEAD(&phba->sli4_hba.sp_unsol_work_queue);
|
|
|
|
/* Initialize the driver internal SLI layer lists. */
|
|
lpfc_sli_setup(phba);
|
|
lpfc_sli_queue_setup(phba);
|
|
|
|
/* Allocate device driver memory */
|
|
rc = lpfc_mem_alloc(phba, SGL_ALIGN_SZ);
|
|
if (rc)
|
|
return -ENOMEM;
|
|
|
|
/* Create the bootstrap mailbox command */
|
|
rc = lpfc_create_bootstrap_mbox(phba);
|
|
if (unlikely(rc))
|
|
goto out_free_mem;
|
|
|
|
/* Set up the host's endian order with the device. */
|
|
rc = lpfc_setup_endian_order(phba);
|
|
if (unlikely(rc))
|
|
goto out_free_bsmbx;
|
|
|
|
rc = lpfc_sli4_fw_cfg_check(phba);
|
|
if (unlikely(rc))
|
|
goto out_free_bsmbx;
|
|
|
|
/* Set up the hba's configuration parameters. */
|
|
rc = lpfc_sli4_read_config(phba);
|
|
if (unlikely(rc))
|
|
goto out_free_bsmbx;
|
|
|
|
/* Perform a function reset */
|
|
rc = lpfc_pci_function_reset(phba);
|
|
if (unlikely(rc))
|
|
goto out_free_bsmbx;
|
|
|
|
mboxq = (LPFC_MBOXQ_t *) mempool_alloc(phba->mbox_mem_pool,
|
|
GFP_KERNEL);
|
|
if (!mboxq) {
|
|
rc = -ENOMEM;
|
|
goto out_free_bsmbx;
|
|
}
|
|
|
|
/* Get the Supported Pages. It is always available. */
|
|
lpfc_supported_pages(mboxq);
|
|
rc = lpfc_sli_issue_mbox(phba, mboxq, MBX_POLL);
|
|
if (unlikely(rc)) {
|
|
rc = -EIO;
|
|
mempool_free(mboxq, phba->mbox_mem_pool);
|
|
goto out_free_bsmbx;
|
|
}
|
|
|
|
mqe = &mboxq->u.mqe;
|
|
memcpy(&pn_page[0], ((uint8_t *)&mqe->un.supp_pages.word3),
|
|
LPFC_MAX_SUPPORTED_PAGES);
|
|
for (i = 0; i < LPFC_MAX_SUPPORTED_PAGES; i++) {
|
|
switch (pn_page[i]) {
|
|
case LPFC_SLI4_PARAMETERS:
|
|
phba->sli4_hba.pc_sli4_params.supported = 1;
|
|
break;
|
|
default:
|
|
break;
|
|
}
|
|
}
|
|
|
|
/* Read the port's SLI4 Parameters capabilities if supported. */
|
|
if (phba->sli4_hba.pc_sli4_params.supported)
|
|
rc = lpfc_pc_sli4_params_get(phba, mboxq);
|
|
mempool_free(mboxq, phba->mbox_mem_pool);
|
|
if (rc) {
|
|
rc = -EIO;
|
|
goto out_free_bsmbx;
|
|
}
|
|
/* Create all the SLI4 queues */
|
|
rc = lpfc_sli4_queue_create(phba);
|
|
if (rc)
|
|
goto out_free_bsmbx;
|
|
|
|
/* Create driver internal CQE event pool */
|
|
rc = lpfc_sli4_cq_event_pool_create(phba);
|
|
if (rc)
|
|
goto out_destroy_queue;
|
|
|
|
/* Initialize and populate the iocb list per host */
|
|
rc = lpfc_init_sgl_list(phba);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1400 Failed to initialize sgl list.\n");
|
|
goto out_destroy_cq_event_pool;
|
|
}
|
|
rc = lpfc_init_active_sgl_array(phba);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1430 Failed to initialize sgl list.\n");
|
|
goto out_free_sgl_list;
|
|
}
|
|
|
|
rc = lpfc_sli4_init_rpi_hdrs(phba);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1432 Failed to initialize rpi headers.\n");
|
|
goto out_free_active_sgl;
|
|
}
|
|
|
|
/* Allocate eligible FCF bmask memory for FCF roundrobin failover */
|
|
longs = (LPFC_SLI4_FCF_TBL_INDX_MAX + BITS_PER_LONG - 1)/BITS_PER_LONG;
|
|
phba->fcf.fcf_rr_bmask = kzalloc(longs * sizeof(unsigned long),
|
|
GFP_KERNEL);
|
|
if (!phba->fcf.fcf_rr_bmask) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2759 Failed allocate memory for FCF round "
|
|
"robin failover bmask\n");
|
|
goto out_remove_rpi_hdrs;
|
|
}
|
|
|
|
phba->sli4_hba.fcp_eq_hdl = kzalloc((sizeof(struct lpfc_fcp_eq_hdl) *
|
|
phba->cfg_fcp_eq_count), GFP_KERNEL);
|
|
if (!phba->sli4_hba.fcp_eq_hdl) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2572 Failed allocate memory for fast-path "
|
|
"per-EQ handle array\n");
|
|
goto out_free_fcf_rr_bmask;
|
|
}
|
|
|
|
phba->sli4_hba.msix_entries = kzalloc((sizeof(struct msix_entry) *
|
|
phba->sli4_hba.cfg_eqn), GFP_KERNEL);
|
|
if (!phba->sli4_hba.msix_entries) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2573 Failed allocate memory for msi-x "
|
|
"interrupt vector entries\n");
|
|
goto out_free_fcp_eq_hdl;
|
|
}
|
|
|
|
return rc;
|
|
|
|
out_free_fcp_eq_hdl:
|
|
kfree(phba->sli4_hba.fcp_eq_hdl);
|
|
out_free_fcf_rr_bmask:
|
|
kfree(phba->fcf.fcf_rr_bmask);
|
|
out_remove_rpi_hdrs:
|
|
lpfc_sli4_remove_rpi_hdrs(phba);
|
|
out_free_active_sgl:
|
|
lpfc_free_active_sgl(phba);
|
|
out_free_sgl_list:
|
|
lpfc_free_sgl_list(phba);
|
|
out_destroy_cq_event_pool:
|
|
lpfc_sli4_cq_event_pool_destroy(phba);
|
|
out_destroy_queue:
|
|
lpfc_sli4_queue_destroy(phba);
|
|
out_free_bsmbx:
|
|
lpfc_destroy_bootstrap_mbox(phba);
|
|
out_free_mem:
|
|
lpfc_mem_free(phba);
|
|
return rc;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_driver_resource_unset - Unset drvr internal resources for SLI4 dev
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to unset the driver internal resources set up
|
|
* specific for supporting the SLI-4 HBA device it attached to.
|
|
**/
|
|
static void
|
|
lpfc_sli4_driver_resource_unset(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_fcf_conn_entry *conn_entry, *next_conn_entry;
|
|
|
|
/* Free memory allocated for msi-x interrupt vector entries */
|
|
kfree(phba->sli4_hba.msix_entries);
|
|
|
|
/* Free memory allocated for fast-path work queue handles */
|
|
kfree(phba->sli4_hba.fcp_eq_hdl);
|
|
|
|
/* Free the allocated rpi headers. */
|
|
lpfc_sli4_remove_rpi_hdrs(phba);
|
|
lpfc_sli4_remove_rpis(phba);
|
|
|
|
/* Free eligible FCF index bmask */
|
|
kfree(phba->fcf.fcf_rr_bmask);
|
|
|
|
/* Free the ELS sgl list */
|
|
lpfc_free_active_sgl(phba);
|
|
lpfc_free_sgl_list(phba);
|
|
|
|
/* Free the SCSI sgl management array */
|
|
kfree(phba->sli4_hba.lpfc_scsi_psb_array);
|
|
|
|
/* Free the SLI4 queues */
|
|
lpfc_sli4_queue_destroy(phba);
|
|
|
|
/* Free the completion queue EQ event pool */
|
|
lpfc_sli4_cq_event_release_all(phba);
|
|
lpfc_sli4_cq_event_pool_destroy(phba);
|
|
|
|
/* Free the bsmbx region. */
|
|
lpfc_destroy_bootstrap_mbox(phba);
|
|
|
|
/* Free the SLI Layer memory with SLI4 HBAs */
|
|
lpfc_mem_free_all(phba);
|
|
|
|
/* Free the current connect table */
|
|
list_for_each_entry_safe(conn_entry, next_conn_entry,
|
|
&phba->fcf_conn_rec_list, list) {
|
|
list_del_init(&conn_entry->list);
|
|
kfree(conn_entry);
|
|
}
|
|
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_init_api_table_setup - Set up init api fucntion jump table
|
|
* @phba: The hba struct for which this call is being executed.
|
|
* @dev_grp: The HBA PCI-Device group number.
|
|
*
|
|
* This routine sets up the device INIT interface API function jump table
|
|
* in @phba struct.
|
|
*
|
|
* Returns: 0 - success, -ENODEV - failure.
|
|
**/
|
|
int
|
|
lpfc_init_api_table_setup(struct lpfc_hba *phba, uint8_t dev_grp)
|
|
{
|
|
phba->lpfc_hba_init_link = lpfc_hba_init_link;
|
|
phba->lpfc_hba_down_link = lpfc_hba_down_link;
|
|
switch (dev_grp) {
|
|
case LPFC_PCI_DEV_LP:
|
|
phba->lpfc_hba_down_post = lpfc_hba_down_post_s3;
|
|
phba->lpfc_handle_eratt = lpfc_handle_eratt_s3;
|
|
phba->lpfc_stop_port = lpfc_stop_port_s3;
|
|
break;
|
|
case LPFC_PCI_DEV_OC:
|
|
phba->lpfc_hba_down_post = lpfc_hba_down_post_s4;
|
|
phba->lpfc_handle_eratt = lpfc_handle_eratt_s4;
|
|
phba->lpfc_stop_port = lpfc_stop_port_s4;
|
|
break;
|
|
default:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1431 Invalid HBA PCI-device group: 0x%x\n",
|
|
dev_grp);
|
|
return -ENODEV;
|
|
break;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_setup_driver_resource_phase1 - Phase1 etup driver internal resources.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to set up the driver internal resources before the
|
|
* device specific resource setup to support the HBA device it attached to.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* other values - error
|
|
**/
|
|
static int
|
|
lpfc_setup_driver_resource_phase1(struct lpfc_hba *phba)
|
|
{
|
|
/*
|
|
* Driver resources common to all SLI revisions
|
|
*/
|
|
atomic_set(&phba->fast_event_count, 0);
|
|
spin_lock_init(&phba->hbalock);
|
|
|
|
/* Initialize ndlp management spinlock */
|
|
spin_lock_init(&phba->ndlp_lock);
|
|
|
|
INIT_LIST_HEAD(&phba->port_list);
|
|
INIT_LIST_HEAD(&phba->work_list);
|
|
init_waitqueue_head(&phba->wait_4_mlo_m_q);
|
|
|
|
/* Initialize the wait queue head for the kernel thread */
|
|
init_waitqueue_head(&phba->work_waitq);
|
|
|
|
/* Initialize the scsi buffer list used by driver for scsi IO */
|
|
spin_lock_init(&phba->scsi_buf_list_lock);
|
|
INIT_LIST_HEAD(&phba->lpfc_scsi_buf_list);
|
|
|
|
/* Initialize the fabric iocb list */
|
|
INIT_LIST_HEAD(&phba->fabric_iocb_list);
|
|
|
|
/* Initialize list to save ELS buffers */
|
|
INIT_LIST_HEAD(&phba->elsbuf);
|
|
|
|
/* Initialize FCF connection rec list */
|
|
INIT_LIST_HEAD(&phba->fcf_conn_rec_list);
|
|
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_setup_driver_resource_phase2 - Phase2 setup driver internal resources.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to set up the driver internal resources after the
|
|
* device specific resource setup to support the HBA device it attached to.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* other values - error
|
|
**/
|
|
static int
|
|
lpfc_setup_driver_resource_phase2(struct lpfc_hba *phba)
|
|
{
|
|
int error;
|
|
|
|
/* Startup the kernel thread for this host adapter. */
|
|
phba->worker_thread = kthread_run(lpfc_do_work, phba,
|
|
"lpfc_worker_%d", phba->brd_no);
|
|
if (IS_ERR(phba->worker_thread)) {
|
|
error = PTR_ERR(phba->worker_thread);
|
|
return error;
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_unset_driver_resource_phase2 - Phase2 unset driver internal resources.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to unset the driver internal resources set up after
|
|
* the device specific resource setup for supporting the HBA device it
|
|
* attached to.
|
|
**/
|
|
static void
|
|
lpfc_unset_driver_resource_phase2(struct lpfc_hba *phba)
|
|
{
|
|
/* Stop kernel worker thread */
|
|
kthread_stop(phba->worker_thread);
|
|
}
|
|
|
|
/**
|
|
* lpfc_free_iocb_list - Free iocb list.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to free the driver's IOCB list and memory.
|
|
**/
|
|
static void
|
|
lpfc_free_iocb_list(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_iocbq *iocbq_entry = NULL, *iocbq_next = NULL;
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
list_for_each_entry_safe(iocbq_entry, iocbq_next,
|
|
&phba->lpfc_iocb_list, list) {
|
|
list_del(&iocbq_entry->list);
|
|
kfree(iocbq_entry);
|
|
phba->total_iocbq_bufs--;
|
|
}
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_init_iocb_list - Allocate and initialize iocb list.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to allocate and initizlize the driver's IOCB
|
|
* list and set up the IOCB tag array accordingly.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* other values - error
|
|
**/
|
|
static int
|
|
lpfc_init_iocb_list(struct lpfc_hba *phba, int iocb_count)
|
|
{
|
|
struct lpfc_iocbq *iocbq_entry = NULL;
|
|
uint16_t iotag;
|
|
int i;
|
|
|
|
/* Initialize and populate the iocb list per host. */
|
|
INIT_LIST_HEAD(&phba->lpfc_iocb_list);
|
|
for (i = 0; i < iocb_count; i++) {
|
|
iocbq_entry = kzalloc(sizeof(struct lpfc_iocbq), GFP_KERNEL);
|
|
if (iocbq_entry == NULL) {
|
|
printk(KERN_ERR "%s: only allocated %d iocbs of "
|
|
"expected %d count. Unloading driver.\n",
|
|
__func__, i, LPFC_IOCB_LIST_CNT);
|
|
goto out_free_iocbq;
|
|
}
|
|
|
|
iotag = lpfc_sli_next_iotag(phba, iocbq_entry);
|
|
if (iotag == 0) {
|
|
kfree(iocbq_entry);
|
|
printk(KERN_ERR "%s: failed to allocate IOTAG. "
|
|
"Unloading driver.\n", __func__);
|
|
goto out_free_iocbq;
|
|
}
|
|
iocbq_entry->sli4_xritag = NO_XRI;
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
list_add(&iocbq_entry->list, &phba->lpfc_iocb_list);
|
|
phba->total_iocbq_bufs++;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
}
|
|
|
|
return 0;
|
|
|
|
out_free_iocbq:
|
|
lpfc_free_iocb_list(phba);
|
|
|
|
return -ENOMEM;
|
|
}
|
|
|
|
/**
|
|
* lpfc_free_sgl_list - Free sgl list.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to free the driver's sgl list and memory.
|
|
**/
|
|
static void
|
|
lpfc_free_sgl_list(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_sglq *sglq_entry = NULL, *sglq_next = NULL;
|
|
LIST_HEAD(sglq_list);
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
list_splice_init(&phba->sli4_hba.lpfc_sgl_list, &sglq_list);
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
list_for_each_entry_safe(sglq_entry, sglq_next,
|
|
&sglq_list, list) {
|
|
list_del(&sglq_entry->list);
|
|
lpfc_mbuf_free(phba, sglq_entry->virt, sglq_entry->phys);
|
|
kfree(sglq_entry);
|
|
phba->sli4_hba.total_sglq_bufs--;
|
|
}
|
|
kfree(phba->sli4_hba.lpfc_els_sgl_array);
|
|
}
|
|
|
|
/**
|
|
* lpfc_init_active_sgl_array - Allocate the buf to track active ELS XRIs.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to allocate the driver's active sgl memory.
|
|
* This array will hold the sglq_entry's for active IOs.
|
|
**/
|
|
static int
|
|
lpfc_init_active_sgl_array(struct lpfc_hba *phba)
|
|
{
|
|
int size;
|
|
size = sizeof(struct lpfc_sglq *);
|
|
size *= phba->sli4_hba.max_cfg_param.max_xri;
|
|
|
|
phba->sli4_hba.lpfc_sglq_active_list =
|
|
kzalloc(size, GFP_KERNEL);
|
|
if (!phba->sli4_hba.lpfc_sglq_active_list)
|
|
return -ENOMEM;
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_free_active_sgl - Free the buf that tracks active ELS XRIs.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to walk through the array of active sglq entries
|
|
* and free all of the resources.
|
|
* This is just a place holder for now.
|
|
**/
|
|
static void
|
|
lpfc_free_active_sgl(struct lpfc_hba *phba)
|
|
{
|
|
kfree(phba->sli4_hba.lpfc_sglq_active_list);
|
|
}
|
|
|
|
/**
|
|
* lpfc_init_sgl_list - Allocate and initialize sgl list.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to allocate and initizlize the driver's sgl
|
|
* list and set up the sgl xritag tag array accordingly.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* other values - error
|
|
**/
|
|
static int
|
|
lpfc_init_sgl_list(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_sglq *sglq_entry = NULL;
|
|
int i;
|
|
int els_xri_cnt;
|
|
|
|
els_xri_cnt = lpfc_sli4_get_els_iocb_cnt(phba);
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_SLI,
|
|
"2400 lpfc_init_sgl_list els %d.\n",
|
|
els_xri_cnt);
|
|
/* Initialize and populate the sglq list per host/VF. */
|
|
INIT_LIST_HEAD(&phba->sli4_hba.lpfc_sgl_list);
|
|
INIT_LIST_HEAD(&phba->sli4_hba.lpfc_abts_els_sgl_list);
|
|
|
|
/* Sanity check on XRI management */
|
|
if (phba->sli4_hba.max_cfg_param.max_xri <= els_xri_cnt) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
|
|
"2562 No room left for SCSI XRI allocation: "
|
|
"max_xri=%d, els_xri=%d\n",
|
|
phba->sli4_hba.max_cfg_param.max_xri,
|
|
els_xri_cnt);
|
|
return -ENOMEM;
|
|
}
|
|
|
|
/* Allocate memory for the ELS XRI management array */
|
|
phba->sli4_hba.lpfc_els_sgl_array =
|
|
kzalloc((sizeof(struct lpfc_sglq *) * els_xri_cnt),
|
|
GFP_KERNEL);
|
|
|
|
if (!phba->sli4_hba.lpfc_els_sgl_array) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
|
|
"2401 Failed to allocate memory for ELS "
|
|
"XRI management array of size %d.\n",
|
|
els_xri_cnt);
|
|
return -ENOMEM;
|
|
}
|
|
|
|
/* Keep the SCSI XRI into the XRI management array */
|
|
phba->sli4_hba.scsi_xri_max =
|
|
phba->sli4_hba.max_cfg_param.max_xri - els_xri_cnt;
|
|
phba->sli4_hba.scsi_xri_cnt = 0;
|
|
|
|
phba->sli4_hba.lpfc_scsi_psb_array =
|
|
kzalloc((sizeof(struct lpfc_scsi_buf *) *
|
|
phba->sli4_hba.scsi_xri_max), GFP_KERNEL);
|
|
|
|
if (!phba->sli4_hba.lpfc_scsi_psb_array) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
|
|
"2563 Failed to allocate memory for SCSI "
|
|
"XRI management array of size %d.\n",
|
|
phba->sli4_hba.scsi_xri_max);
|
|
kfree(phba->sli4_hba.lpfc_els_sgl_array);
|
|
return -ENOMEM;
|
|
}
|
|
|
|
for (i = 0; i < els_xri_cnt; i++) {
|
|
sglq_entry = kzalloc(sizeof(struct lpfc_sglq), GFP_KERNEL);
|
|
if (sglq_entry == NULL) {
|
|
printk(KERN_ERR "%s: only allocated %d sgls of "
|
|
"expected %d count. Unloading driver.\n",
|
|
__func__, i, els_xri_cnt);
|
|
goto out_free_mem;
|
|
}
|
|
|
|
sglq_entry->sli4_xritag = lpfc_sli4_next_xritag(phba);
|
|
if (sglq_entry->sli4_xritag == NO_XRI) {
|
|
kfree(sglq_entry);
|
|
printk(KERN_ERR "%s: failed to allocate XRI.\n"
|
|
"Unloading driver.\n", __func__);
|
|
goto out_free_mem;
|
|
}
|
|
sglq_entry->buff_type = GEN_BUFF_TYPE;
|
|
sglq_entry->virt = lpfc_mbuf_alloc(phba, 0, &sglq_entry->phys);
|
|
if (sglq_entry->virt == NULL) {
|
|
kfree(sglq_entry);
|
|
printk(KERN_ERR "%s: failed to allocate mbuf.\n"
|
|
"Unloading driver.\n", __func__);
|
|
goto out_free_mem;
|
|
}
|
|
sglq_entry->sgl = sglq_entry->virt;
|
|
memset(sglq_entry->sgl, 0, LPFC_BPL_SIZE);
|
|
|
|
/* The list order is used by later block SGL registraton */
|
|
spin_lock_irq(&phba->hbalock);
|
|
sglq_entry->state = SGL_FREED;
|
|
list_add_tail(&sglq_entry->list, &phba->sli4_hba.lpfc_sgl_list);
|
|
phba->sli4_hba.lpfc_els_sgl_array[i] = sglq_entry;
|
|
phba->sli4_hba.total_sglq_bufs++;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
}
|
|
return 0;
|
|
|
|
out_free_mem:
|
|
kfree(phba->sli4_hba.lpfc_scsi_psb_array);
|
|
lpfc_free_sgl_list(phba);
|
|
return -ENOMEM;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_init_rpi_hdrs - Post the rpi header memory region to the port
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to post rpi header templates to the
|
|
* HBA consistent with the SLI-4 interface spec. This routine
|
|
* posts a PAGE_SIZE memory region to the port to hold up to
|
|
* PAGE_SIZE modulo 64 rpi context headers.
|
|
* No locks are held here because this is an initialization routine
|
|
* called only from probe or lpfc_online when interrupts are not
|
|
* enabled and the driver is reinitializing the device.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* -ENOMEM - No availble memory
|
|
* -EIO - The mailbox failed to complete successfully.
|
|
**/
|
|
int
|
|
lpfc_sli4_init_rpi_hdrs(struct lpfc_hba *phba)
|
|
{
|
|
int rc = 0;
|
|
int longs;
|
|
uint16_t rpi_count;
|
|
struct lpfc_rpi_hdr *rpi_hdr;
|
|
|
|
INIT_LIST_HEAD(&phba->sli4_hba.lpfc_rpi_hdr_list);
|
|
|
|
/*
|
|
* Provision an rpi bitmask range for discovery. The total count
|
|
* is the difference between max and base + 1.
|
|
*/
|
|
rpi_count = phba->sli4_hba.max_cfg_param.rpi_base +
|
|
phba->sli4_hba.max_cfg_param.max_rpi - 1;
|
|
|
|
longs = ((rpi_count) + BITS_PER_LONG - 1) / BITS_PER_LONG;
|
|
phba->sli4_hba.rpi_bmask = kzalloc(longs * sizeof(unsigned long),
|
|
GFP_KERNEL);
|
|
if (!phba->sli4_hba.rpi_bmask)
|
|
return -ENOMEM;
|
|
|
|
rpi_hdr = lpfc_sli4_create_rpi_hdr(phba);
|
|
if (!rpi_hdr) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_MBOX | LOG_SLI,
|
|
"0391 Error during rpi post operation\n");
|
|
lpfc_sli4_remove_rpis(phba);
|
|
rc = -ENODEV;
|
|
}
|
|
|
|
return rc;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_create_rpi_hdr - Allocate an rpi header memory region
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to allocate a single 4KB memory region to
|
|
* support rpis and stores them in the phba. This single region
|
|
* provides support for up to 64 rpis. The region is used globally
|
|
* by the device.
|
|
*
|
|
* Returns:
|
|
* A valid rpi hdr on success.
|
|
* A NULL pointer on any failure.
|
|
**/
|
|
struct lpfc_rpi_hdr *
|
|
lpfc_sli4_create_rpi_hdr(struct lpfc_hba *phba)
|
|
{
|
|
uint16_t rpi_limit, curr_rpi_range;
|
|
struct lpfc_dmabuf *dmabuf;
|
|
struct lpfc_rpi_hdr *rpi_hdr;
|
|
|
|
rpi_limit = phba->sli4_hba.max_cfg_param.rpi_base +
|
|
phba->sli4_hba.max_cfg_param.max_rpi - 1;
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
curr_rpi_range = phba->sli4_hba.next_rpi;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
/*
|
|
* The port has a limited number of rpis. The increment here
|
|
* is LPFC_RPI_HDR_COUNT - 1 to account for the starting value
|
|
* and to allow the full max_rpi range per port.
|
|
*/
|
|
if ((curr_rpi_range + (LPFC_RPI_HDR_COUNT - 1)) > rpi_limit)
|
|
return NULL;
|
|
|
|
/*
|
|
* First allocate the protocol header region for the port. The
|
|
* port expects a 4KB DMA-mapped memory region that is 4K aligned.
|
|
*/
|
|
dmabuf = kzalloc(sizeof(struct lpfc_dmabuf), GFP_KERNEL);
|
|
if (!dmabuf)
|
|
return NULL;
|
|
|
|
dmabuf->virt = dma_alloc_coherent(&phba->pcidev->dev,
|
|
LPFC_HDR_TEMPLATE_SIZE,
|
|
&dmabuf->phys,
|
|
GFP_KERNEL);
|
|
if (!dmabuf->virt) {
|
|
rpi_hdr = NULL;
|
|
goto err_free_dmabuf;
|
|
}
|
|
|
|
memset(dmabuf->virt, 0, LPFC_HDR_TEMPLATE_SIZE);
|
|
if (!IS_ALIGNED(dmabuf->phys, LPFC_HDR_TEMPLATE_SIZE)) {
|
|
rpi_hdr = NULL;
|
|
goto err_free_coherent;
|
|
}
|
|
|
|
/* Save the rpi header data for cleanup later. */
|
|
rpi_hdr = kzalloc(sizeof(struct lpfc_rpi_hdr), GFP_KERNEL);
|
|
if (!rpi_hdr)
|
|
goto err_free_coherent;
|
|
|
|
rpi_hdr->dmabuf = dmabuf;
|
|
rpi_hdr->len = LPFC_HDR_TEMPLATE_SIZE;
|
|
rpi_hdr->page_count = 1;
|
|
spin_lock_irq(&phba->hbalock);
|
|
rpi_hdr->start_rpi = phba->sli4_hba.next_rpi;
|
|
list_add_tail(&rpi_hdr->list, &phba->sli4_hba.lpfc_rpi_hdr_list);
|
|
|
|
/*
|
|
* The next_rpi stores the next module-64 rpi value to post
|
|
* in any subsequent rpi memory region postings.
|
|
*/
|
|
phba->sli4_hba.next_rpi += LPFC_RPI_HDR_COUNT;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
return rpi_hdr;
|
|
|
|
err_free_coherent:
|
|
dma_free_coherent(&phba->pcidev->dev, LPFC_HDR_TEMPLATE_SIZE,
|
|
dmabuf->virt, dmabuf->phys);
|
|
err_free_dmabuf:
|
|
kfree(dmabuf);
|
|
return NULL;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_remove_rpi_hdrs - Remove all rpi header memory regions
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to remove all memory resources allocated
|
|
* to support rpis. This routine presumes the caller has released all
|
|
* rpis consumed by fabric or port logins and is prepared to have
|
|
* the header pages removed.
|
|
**/
|
|
void
|
|
lpfc_sli4_remove_rpi_hdrs(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_rpi_hdr *rpi_hdr, *next_rpi_hdr;
|
|
|
|
list_for_each_entry_safe(rpi_hdr, next_rpi_hdr,
|
|
&phba->sli4_hba.lpfc_rpi_hdr_list, list) {
|
|
list_del(&rpi_hdr->list);
|
|
dma_free_coherent(&phba->pcidev->dev, rpi_hdr->len,
|
|
rpi_hdr->dmabuf->virt, rpi_hdr->dmabuf->phys);
|
|
kfree(rpi_hdr->dmabuf);
|
|
kfree(rpi_hdr);
|
|
}
|
|
|
|
phba->sli4_hba.next_rpi = phba->sli4_hba.max_cfg_param.rpi_base;
|
|
memset(phba->sli4_hba.rpi_bmask, 0, sizeof(*phba->sli4_hba.rpi_bmask));
|
|
}
|
|
|
|
/**
|
|
* lpfc_hba_alloc - Allocate driver hba data structure for a device.
|
|
* @pdev: pointer to pci device data structure.
|
|
*
|
|
* This routine is invoked to allocate the driver hba data structure for an
|
|
* HBA device. If the allocation is successful, the phba reference to the
|
|
* PCI device data structure is set.
|
|
*
|
|
* Return codes
|
|
* pointer to @phba - successful
|
|
* NULL - error
|
|
**/
|
|
static struct lpfc_hba *
|
|
lpfc_hba_alloc(struct pci_dev *pdev)
|
|
{
|
|
struct lpfc_hba *phba;
|
|
|
|
/* Allocate memory for HBA structure */
|
|
phba = kzalloc(sizeof(struct lpfc_hba), GFP_KERNEL);
|
|
if (!phba) {
|
|
dev_err(&pdev->dev, "failed to allocate hba struct\n");
|
|
return NULL;
|
|
}
|
|
|
|
/* Set reference to PCI device in HBA structure */
|
|
phba->pcidev = pdev;
|
|
|
|
/* Assign an unused board number */
|
|
phba->brd_no = lpfc_get_instance();
|
|
if (phba->brd_no < 0) {
|
|
kfree(phba);
|
|
return NULL;
|
|
}
|
|
|
|
spin_lock_init(&phba->ct_ev_lock);
|
|
INIT_LIST_HEAD(&phba->ct_ev_waiters);
|
|
|
|
return phba;
|
|
}
|
|
|
|
/**
|
|
* lpfc_hba_free - Free driver hba data structure with a device.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to free the driver hba data structure with an
|
|
* HBA device.
|
|
**/
|
|
static void
|
|
lpfc_hba_free(struct lpfc_hba *phba)
|
|
{
|
|
/* Release the driver assigned board number */
|
|
idr_remove(&lpfc_hba_index, phba->brd_no);
|
|
|
|
kfree(phba);
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_create_shost - Create hba physical port with associated scsi host.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to create HBA physical port and associate a SCSI
|
|
* host with it.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* other values - error
|
|
**/
|
|
static int
|
|
lpfc_create_shost(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_vport *vport;
|
|
struct Scsi_Host *shost;
|
|
|
|
/* Initialize HBA FC structure */
|
|
phba->fc_edtov = FF_DEF_EDTOV;
|
|
phba->fc_ratov = FF_DEF_RATOV;
|
|
phba->fc_altov = FF_DEF_ALTOV;
|
|
phba->fc_arbtov = FF_DEF_ARBTOV;
|
|
|
|
atomic_set(&phba->sdev_cnt, 0);
|
|
vport = lpfc_create_port(phba, phba->brd_no, &phba->pcidev->dev);
|
|
if (!vport)
|
|
return -ENODEV;
|
|
|
|
shost = lpfc_shost_from_vport(vport);
|
|
phba->pport = vport;
|
|
lpfc_debugfs_initialize(vport);
|
|
/* Put reference to SCSI host to driver's device private data */
|
|
pci_set_drvdata(phba->pcidev, shost);
|
|
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_destroy_shost - Destroy hba physical port with associated scsi host.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to destroy HBA physical port and the associated
|
|
* SCSI host.
|
|
**/
|
|
static void
|
|
lpfc_destroy_shost(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_vport *vport = phba->pport;
|
|
|
|
/* Destroy physical port that associated with the SCSI host */
|
|
destroy_port(vport);
|
|
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_setup_bg - Setup Block guard structures and debug areas.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @shost: the shost to be used to detect Block guard settings.
|
|
*
|
|
* This routine sets up the local Block guard protocol settings for @shost.
|
|
* This routine also allocates memory for debugging bg buffers.
|
|
**/
|
|
static void
|
|
lpfc_setup_bg(struct lpfc_hba *phba, struct Scsi_Host *shost)
|
|
{
|
|
int pagecnt = 10;
|
|
if (lpfc_prot_mask && lpfc_prot_guard) {
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"1478 Registering BlockGuard with the "
|
|
"SCSI layer\n");
|
|
scsi_host_set_prot(shost, lpfc_prot_mask);
|
|
scsi_host_set_guard(shost, lpfc_prot_guard);
|
|
}
|
|
if (!_dump_buf_data) {
|
|
while (pagecnt) {
|
|
spin_lock_init(&_dump_buf_lock);
|
|
_dump_buf_data =
|
|
(char *) __get_free_pages(GFP_KERNEL, pagecnt);
|
|
if (_dump_buf_data) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_BG,
|
|
"9043 BLKGRD: allocated %d pages for "
|
|
"_dump_buf_data at 0x%p\n",
|
|
(1 << pagecnt), _dump_buf_data);
|
|
_dump_buf_data_order = pagecnt;
|
|
memset(_dump_buf_data, 0,
|
|
((1 << PAGE_SHIFT) << pagecnt));
|
|
break;
|
|
} else
|
|
--pagecnt;
|
|
}
|
|
if (!_dump_buf_data_order)
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_BG,
|
|
"9044 BLKGRD: ERROR unable to allocate "
|
|
"memory for hexdump\n");
|
|
} else
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_BG,
|
|
"9045 BLKGRD: already allocated _dump_buf_data=0x%p"
|
|
"\n", _dump_buf_data);
|
|
if (!_dump_buf_dif) {
|
|
while (pagecnt) {
|
|
_dump_buf_dif =
|
|
(char *) __get_free_pages(GFP_KERNEL, pagecnt);
|
|
if (_dump_buf_dif) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_BG,
|
|
"9046 BLKGRD: allocated %d pages for "
|
|
"_dump_buf_dif at 0x%p\n",
|
|
(1 << pagecnt), _dump_buf_dif);
|
|
_dump_buf_dif_order = pagecnt;
|
|
memset(_dump_buf_dif, 0,
|
|
((1 << PAGE_SHIFT) << pagecnt));
|
|
break;
|
|
} else
|
|
--pagecnt;
|
|
}
|
|
if (!_dump_buf_dif_order)
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_BG,
|
|
"9047 BLKGRD: ERROR unable to allocate "
|
|
"memory for hexdump\n");
|
|
} else
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_BG,
|
|
"9048 BLKGRD: already allocated _dump_buf_dif=0x%p\n",
|
|
_dump_buf_dif);
|
|
}
|
|
|
|
/**
|
|
* lpfc_post_init_setup - Perform necessary device post initialization setup.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to perform all the necessary post initialization
|
|
* setup for the device.
|
|
**/
|
|
static void
|
|
lpfc_post_init_setup(struct lpfc_hba *phba)
|
|
{
|
|
struct Scsi_Host *shost;
|
|
struct lpfc_adapter_event_header adapter_event;
|
|
|
|
/* Get the default values for Model Name and Description */
|
|
lpfc_get_hba_model_desc(phba, phba->ModelName, phba->ModelDesc);
|
|
|
|
/*
|
|
* hba setup may have changed the hba_queue_depth so we need to
|
|
* adjust the value of can_queue.
|
|
*/
|
|
shost = pci_get_drvdata(phba->pcidev);
|
|
shost->can_queue = phba->cfg_hba_queue_depth - 10;
|
|
if (phba->sli3_options & LPFC_SLI3_BG_ENABLED)
|
|
lpfc_setup_bg(phba, shost);
|
|
|
|
lpfc_host_attrib_init(shost);
|
|
|
|
if (phba->cfg_poll & DISABLE_FCP_RING_INT) {
|
|
spin_lock_irq(shost->host_lock);
|
|
lpfc_poll_start_timer(phba);
|
|
spin_unlock_irq(shost->host_lock);
|
|
}
|
|
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0428 Perform SCSI scan\n");
|
|
/* Send board arrival event to upper layer */
|
|
adapter_event.event_type = FC_REG_ADAPTER_EVENT;
|
|
adapter_event.subcategory = LPFC_EVENT_ARRIVAL;
|
|
fc_host_post_vendor_event(shost, fc_get_event_number(),
|
|
sizeof(adapter_event),
|
|
(char *) &adapter_event,
|
|
LPFC_NL_VENDOR_ID);
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli_pci_mem_setup - Setup SLI3 HBA PCI memory space.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to set up the PCI device memory space for device
|
|
* with SLI-3 interface spec.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* other values - error
|
|
**/
|
|
static int
|
|
lpfc_sli_pci_mem_setup(struct lpfc_hba *phba)
|
|
{
|
|
struct pci_dev *pdev;
|
|
unsigned long bar0map_len, bar2map_len;
|
|
int i, hbq_count;
|
|
void *ptr;
|
|
int error = -ENODEV;
|
|
|
|
/* Obtain PCI device reference */
|
|
if (!phba->pcidev)
|
|
return error;
|
|
else
|
|
pdev = phba->pcidev;
|
|
|
|
/* Set the device DMA mask size */
|
|
if (pci_set_dma_mask(pdev, DMA_BIT_MASK(64)) != 0
|
|
|| pci_set_consistent_dma_mask(pdev,DMA_BIT_MASK(64)) != 0) {
|
|
if (pci_set_dma_mask(pdev, DMA_BIT_MASK(32)) != 0
|
|
|| pci_set_consistent_dma_mask(pdev,DMA_BIT_MASK(32)) != 0) {
|
|
return error;
|
|
}
|
|
}
|
|
|
|
/* Get the bus address of Bar0 and Bar2 and the number of bytes
|
|
* required by each mapping.
|
|
*/
|
|
phba->pci_bar0_map = pci_resource_start(pdev, 0);
|
|
bar0map_len = pci_resource_len(pdev, 0);
|
|
|
|
phba->pci_bar2_map = pci_resource_start(pdev, 2);
|
|
bar2map_len = pci_resource_len(pdev, 2);
|
|
|
|
/* Map HBA SLIM to a kernel virtual address. */
|
|
phba->slim_memmap_p = ioremap(phba->pci_bar0_map, bar0map_len);
|
|
if (!phba->slim_memmap_p) {
|
|
dev_printk(KERN_ERR, &pdev->dev,
|
|
"ioremap failed for SLIM memory.\n");
|
|
goto out;
|
|
}
|
|
|
|
/* Map HBA Control Registers to a kernel virtual address. */
|
|
phba->ctrl_regs_memmap_p = ioremap(phba->pci_bar2_map, bar2map_len);
|
|
if (!phba->ctrl_regs_memmap_p) {
|
|
dev_printk(KERN_ERR, &pdev->dev,
|
|
"ioremap failed for HBA control registers.\n");
|
|
goto out_iounmap_slim;
|
|
}
|
|
|
|
/* Allocate memory for SLI-2 structures */
|
|
phba->slim2p.virt = dma_alloc_coherent(&pdev->dev,
|
|
SLI2_SLIM_SIZE,
|
|
&phba->slim2p.phys,
|
|
GFP_KERNEL);
|
|
if (!phba->slim2p.virt)
|
|
goto out_iounmap;
|
|
|
|
memset(phba->slim2p.virt, 0, SLI2_SLIM_SIZE);
|
|
phba->mbox = phba->slim2p.virt + offsetof(struct lpfc_sli2_slim, mbx);
|
|
phba->mbox_ext = (phba->slim2p.virt +
|
|
offsetof(struct lpfc_sli2_slim, mbx_ext_words));
|
|
phba->pcb = (phba->slim2p.virt + offsetof(struct lpfc_sli2_slim, pcb));
|
|
phba->IOCBs = (phba->slim2p.virt +
|
|
offsetof(struct lpfc_sli2_slim, IOCBs));
|
|
|
|
phba->hbqslimp.virt = dma_alloc_coherent(&pdev->dev,
|
|
lpfc_sli_hbq_size(),
|
|
&phba->hbqslimp.phys,
|
|
GFP_KERNEL);
|
|
if (!phba->hbqslimp.virt)
|
|
goto out_free_slim;
|
|
|
|
hbq_count = lpfc_sli_hbq_count();
|
|
ptr = phba->hbqslimp.virt;
|
|
for (i = 0; i < hbq_count; ++i) {
|
|
phba->hbqs[i].hbq_virt = ptr;
|
|
INIT_LIST_HEAD(&phba->hbqs[i].hbq_buffer_list);
|
|
ptr += (lpfc_hbq_defs[i]->entry_count *
|
|
sizeof(struct lpfc_hbq_entry));
|
|
}
|
|
phba->hbqs[LPFC_ELS_HBQ].hbq_alloc_buffer = lpfc_els_hbq_alloc;
|
|
phba->hbqs[LPFC_ELS_HBQ].hbq_free_buffer = lpfc_els_hbq_free;
|
|
|
|
memset(phba->hbqslimp.virt, 0, lpfc_sli_hbq_size());
|
|
|
|
INIT_LIST_HEAD(&phba->rb_pend_list);
|
|
|
|
phba->MBslimaddr = phba->slim_memmap_p;
|
|
phba->HAregaddr = phba->ctrl_regs_memmap_p + HA_REG_OFFSET;
|
|
phba->CAregaddr = phba->ctrl_regs_memmap_p + CA_REG_OFFSET;
|
|
phba->HSregaddr = phba->ctrl_regs_memmap_p + HS_REG_OFFSET;
|
|
phba->HCregaddr = phba->ctrl_regs_memmap_p + HC_REG_OFFSET;
|
|
|
|
return 0;
|
|
|
|
out_free_slim:
|
|
dma_free_coherent(&pdev->dev, SLI2_SLIM_SIZE,
|
|
phba->slim2p.virt, phba->slim2p.phys);
|
|
out_iounmap:
|
|
iounmap(phba->ctrl_regs_memmap_p);
|
|
out_iounmap_slim:
|
|
iounmap(phba->slim_memmap_p);
|
|
out:
|
|
return error;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli_pci_mem_unset - Unset SLI3 HBA PCI memory space.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to unset the PCI device memory space for device
|
|
* with SLI-3 interface spec.
|
|
**/
|
|
static void
|
|
lpfc_sli_pci_mem_unset(struct lpfc_hba *phba)
|
|
{
|
|
struct pci_dev *pdev;
|
|
|
|
/* Obtain PCI device reference */
|
|
if (!phba->pcidev)
|
|
return;
|
|
else
|
|
pdev = phba->pcidev;
|
|
|
|
/* Free coherent DMA memory allocated */
|
|
dma_free_coherent(&pdev->dev, lpfc_sli_hbq_size(),
|
|
phba->hbqslimp.virt, phba->hbqslimp.phys);
|
|
dma_free_coherent(&pdev->dev, SLI2_SLIM_SIZE,
|
|
phba->slim2p.virt, phba->slim2p.phys);
|
|
|
|
/* I/O memory unmap */
|
|
iounmap(phba->ctrl_regs_memmap_p);
|
|
iounmap(phba->slim_memmap_p);
|
|
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_post_status_check - Wait for SLI4 POST done and check status
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to wait for SLI4 device Power On Self Test (POST)
|
|
* done and check status.
|
|
*
|
|
* Return 0 if successful, otherwise -ENODEV.
|
|
**/
|
|
int
|
|
lpfc_sli4_post_status_check(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_register sta_reg, uerrlo_reg, uerrhi_reg;
|
|
int i, port_error = -ENODEV;
|
|
|
|
if (!phba->sli4_hba.STAregaddr)
|
|
return -ENODEV;
|
|
|
|
/* Wait up to 30 seconds for the SLI Port POST done and ready */
|
|
for (i = 0; i < 3000; i++) {
|
|
sta_reg.word0 = readl(phba->sli4_hba.STAregaddr);
|
|
/* Encounter fatal POST error, break out */
|
|
if (bf_get(lpfc_hst_state_perr, &sta_reg)) {
|
|
port_error = -ENODEV;
|
|
break;
|
|
}
|
|
if (LPFC_POST_STAGE_ARMFW_READY ==
|
|
bf_get(lpfc_hst_state_port_status, &sta_reg)) {
|
|
port_error = 0;
|
|
break;
|
|
}
|
|
msleep(10);
|
|
}
|
|
|
|
if (port_error)
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1408 Failure HBA POST Status: sta_reg=0x%x, "
|
|
"perr=x%x, sfi=x%x, nip=x%x, ipc=x%x, xrom=x%x, "
|
|
"dl=x%x, pstatus=x%x\n", sta_reg.word0,
|
|
bf_get(lpfc_hst_state_perr, &sta_reg),
|
|
bf_get(lpfc_hst_state_sfi, &sta_reg),
|
|
bf_get(lpfc_hst_state_nip, &sta_reg),
|
|
bf_get(lpfc_hst_state_ipc, &sta_reg),
|
|
bf_get(lpfc_hst_state_xrom, &sta_reg),
|
|
bf_get(lpfc_hst_state_dl, &sta_reg),
|
|
bf_get(lpfc_hst_state_port_status, &sta_reg));
|
|
|
|
/* Log device information */
|
|
phba->sli4_hba.sli_intf.word0 = readl(phba->sli4_hba.SLIINTFregaddr);
|
|
if (bf_get(lpfc_sli_intf_valid,
|
|
&phba->sli4_hba.sli_intf) == LPFC_SLI_INTF_VALID) {
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"2534 Device Info: ChipType=0x%x, SliRev=0x%x, "
|
|
"FeatureL1=0x%x, FeatureL2=0x%x\n",
|
|
bf_get(lpfc_sli_intf_sli_family,
|
|
&phba->sli4_hba.sli_intf),
|
|
bf_get(lpfc_sli_intf_slirev,
|
|
&phba->sli4_hba.sli_intf),
|
|
bf_get(lpfc_sli_intf_featurelevel1,
|
|
&phba->sli4_hba.sli_intf),
|
|
bf_get(lpfc_sli_intf_featurelevel2,
|
|
&phba->sli4_hba.sli_intf));
|
|
}
|
|
phba->sli4_hba.ue_mask_lo = readl(phba->sli4_hba.UEMASKLOregaddr);
|
|
phba->sli4_hba.ue_mask_hi = readl(phba->sli4_hba.UEMASKHIregaddr);
|
|
/* With uncoverable error, log the error message and return error */
|
|
uerrlo_reg.word0 = readl(phba->sli4_hba.UERRLOregaddr);
|
|
uerrhi_reg.word0 = readl(phba->sli4_hba.UERRHIregaddr);
|
|
if ((~phba->sli4_hba.ue_mask_lo & uerrlo_reg.word0) ||
|
|
(~phba->sli4_hba.ue_mask_hi & uerrhi_reg.word0)) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1422 HBA Unrecoverable error: "
|
|
"uerr_lo_reg=0x%x, uerr_hi_reg=0x%x, "
|
|
"ue_mask_lo_reg=0x%x, ue_mask_hi_reg=0x%x\n",
|
|
uerrlo_reg.word0, uerrhi_reg.word0,
|
|
phba->sli4_hba.ue_mask_lo,
|
|
phba->sli4_hba.ue_mask_hi);
|
|
return -ENODEV;
|
|
}
|
|
|
|
return port_error;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_bar0_register_memmap - Set up SLI4 BAR0 register memory map.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to set up SLI4 BAR0 PCI config space register
|
|
* memory map.
|
|
**/
|
|
static void
|
|
lpfc_sli4_bar0_register_memmap(struct lpfc_hba *phba)
|
|
{
|
|
phba->sli4_hba.UERRLOregaddr = phba->sli4_hba.conf_regs_memmap_p +
|
|
LPFC_UERR_STATUS_LO;
|
|
phba->sli4_hba.UERRHIregaddr = phba->sli4_hba.conf_regs_memmap_p +
|
|
LPFC_UERR_STATUS_HI;
|
|
phba->sli4_hba.UEMASKLOregaddr = phba->sli4_hba.conf_regs_memmap_p +
|
|
LPFC_UE_MASK_LO;
|
|
phba->sli4_hba.UEMASKHIregaddr = phba->sli4_hba.conf_regs_memmap_p +
|
|
LPFC_UE_MASK_HI;
|
|
phba->sli4_hba.SLIINTFregaddr = phba->sli4_hba.conf_regs_memmap_p +
|
|
LPFC_SLI_INTF;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_bar1_register_memmap - Set up SLI4 BAR1 register memory map.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to set up SLI4 BAR1 control status register (CSR)
|
|
* memory map.
|
|
**/
|
|
static void
|
|
lpfc_sli4_bar1_register_memmap(struct lpfc_hba *phba)
|
|
{
|
|
|
|
phba->sli4_hba.STAregaddr = phba->sli4_hba.ctrl_regs_memmap_p +
|
|
LPFC_HST_STATE;
|
|
phba->sli4_hba.ISRregaddr = phba->sli4_hba.ctrl_regs_memmap_p +
|
|
LPFC_HST_ISR0;
|
|
phba->sli4_hba.IMRregaddr = phba->sli4_hba.ctrl_regs_memmap_p +
|
|
LPFC_HST_IMR0;
|
|
phba->sli4_hba.ISCRregaddr = phba->sli4_hba.ctrl_regs_memmap_p +
|
|
LPFC_HST_ISCR0;
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_bar2_register_memmap - Set up SLI4 BAR2 register memory map.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @vf: virtual function number
|
|
*
|
|
* This routine is invoked to set up SLI4 BAR2 doorbell register memory map
|
|
* based on the given viftual function number, @vf.
|
|
*
|
|
* Return 0 if successful, otherwise -ENODEV.
|
|
**/
|
|
static int
|
|
lpfc_sli4_bar2_register_memmap(struct lpfc_hba *phba, uint32_t vf)
|
|
{
|
|
if (vf > LPFC_VIR_FUNC_MAX)
|
|
return -ENODEV;
|
|
|
|
phba->sli4_hba.RQDBregaddr = (phba->sli4_hba.drbl_regs_memmap_p +
|
|
vf * LPFC_VFR_PAGE_SIZE + LPFC_RQ_DOORBELL);
|
|
phba->sli4_hba.WQDBregaddr = (phba->sli4_hba.drbl_regs_memmap_p +
|
|
vf * LPFC_VFR_PAGE_SIZE + LPFC_WQ_DOORBELL);
|
|
phba->sli4_hba.EQCQDBregaddr = (phba->sli4_hba.drbl_regs_memmap_p +
|
|
vf * LPFC_VFR_PAGE_SIZE + LPFC_EQCQ_DOORBELL);
|
|
phba->sli4_hba.MQDBregaddr = (phba->sli4_hba.drbl_regs_memmap_p +
|
|
vf * LPFC_VFR_PAGE_SIZE + LPFC_MQ_DOORBELL);
|
|
phba->sli4_hba.BMBXregaddr = (phba->sli4_hba.drbl_regs_memmap_p +
|
|
vf * LPFC_VFR_PAGE_SIZE + LPFC_BMBX);
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_create_bootstrap_mbox - Create the bootstrap mailbox
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to create the bootstrap mailbox
|
|
* region consistent with the SLI-4 interface spec. This
|
|
* routine allocates all memory necessary to communicate
|
|
* mailbox commands to the port and sets up all alignment
|
|
* needs. No locks are expected to be held when calling
|
|
* this routine.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* -ENOMEM - could not allocated memory.
|
|
**/
|
|
static int
|
|
lpfc_create_bootstrap_mbox(struct lpfc_hba *phba)
|
|
{
|
|
uint32_t bmbx_size;
|
|
struct lpfc_dmabuf *dmabuf;
|
|
struct dma_address *dma_address;
|
|
uint32_t pa_addr;
|
|
uint64_t phys_addr;
|
|
|
|
dmabuf = kzalloc(sizeof(struct lpfc_dmabuf), GFP_KERNEL);
|
|
if (!dmabuf)
|
|
return -ENOMEM;
|
|
|
|
/*
|
|
* The bootstrap mailbox region is comprised of 2 parts
|
|
* plus an alignment restriction of 16 bytes.
|
|
*/
|
|
bmbx_size = sizeof(struct lpfc_bmbx_create) + (LPFC_ALIGN_16_BYTE - 1);
|
|
dmabuf->virt = dma_alloc_coherent(&phba->pcidev->dev,
|
|
bmbx_size,
|
|
&dmabuf->phys,
|
|
GFP_KERNEL);
|
|
if (!dmabuf->virt) {
|
|
kfree(dmabuf);
|
|
return -ENOMEM;
|
|
}
|
|
memset(dmabuf->virt, 0, bmbx_size);
|
|
|
|
/*
|
|
* Initialize the bootstrap mailbox pointers now so that the register
|
|
* operations are simple later. The mailbox dma address is required
|
|
* to be 16-byte aligned. Also align the virtual memory as each
|
|
* maibox is copied into the bmbx mailbox region before issuing the
|
|
* command to the port.
|
|
*/
|
|
phba->sli4_hba.bmbx.dmabuf = dmabuf;
|
|
phba->sli4_hba.bmbx.bmbx_size = bmbx_size;
|
|
|
|
phba->sli4_hba.bmbx.avirt = PTR_ALIGN(dmabuf->virt,
|
|
LPFC_ALIGN_16_BYTE);
|
|
phba->sli4_hba.bmbx.aphys = ALIGN(dmabuf->phys,
|
|
LPFC_ALIGN_16_BYTE);
|
|
|
|
/*
|
|
* Set the high and low physical addresses now. The SLI4 alignment
|
|
* requirement is 16 bytes and the mailbox is posted to the port
|
|
* as two 30-bit addresses. The other data is a bit marking whether
|
|
* the 30-bit address is the high or low address.
|
|
* Upcast bmbx aphys to 64bits so shift instruction compiles
|
|
* clean on 32 bit machines.
|
|
*/
|
|
dma_address = &phba->sli4_hba.bmbx.dma_address;
|
|
phys_addr = (uint64_t)phba->sli4_hba.bmbx.aphys;
|
|
pa_addr = (uint32_t) ((phys_addr >> 34) & 0x3fffffff);
|
|
dma_address->addr_hi = (uint32_t) ((pa_addr << 2) |
|
|
LPFC_BMBX_BIT1_ADDR_HI);
|
|
|
|
pa_addr = (uint32_t) ((phba->sli4_hba.bmbx.aphys >> 4) & 0x3fffffff);
|
|
dma_address->addr_lo = (uint32_t) ((pa_addr << 2) |
|
|
LPFC_BMBX_BIT1_ADDR_LO);
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_destroy_bootstrap_mbox - Destroy all bootstrap mailbox resources
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to teardown the bootstrap mailbox
|
|
* region and release all host resources. This routine requires
|
|
* the caller to ensure all mailbox commands recovered, no
|
|
* additional mailbox comands are sent, and interrupts are disabled
|
|
* before calling this routine.
|
|
*
|
|
**/
|
|
static void
|
|
lpfc_destroy_bootstrap_mbox(struct lpfc_hba *phba)
|
|
{
|
|
dma_free_coherent(&phba->pcidev->dev,
|
|
phba->sli4_hba.bmbx.bmbx_size,
|
|
phba->sli4_hba.bmbx.dmabuf->virt,
|
|
phba->sli4_hba.bmbx.dmabuf->phys);
|
|
|
|
kfree(phba->sli4_hba.bmbx.dmabuf);
|
|
memset(&phba->sli4_hba.bmbx, 0, sizeof(struct lpfc_bmbx));
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_read_config - Get the config parameters.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to read the configuration parameters from the HBA.
|
|
* The configuration parameters are used to set the base and maximum values
|
|
* for RPI's XRI's VPI's VFI's and FCFIs. These values also affect the resource
|
|
* allocation for the port.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* -ENOMEM - No availble memory
|
|
* -EIO - The mailbox failed to complete successfully.
|
|
**/
|
|
static int
|
|
lpfc_sli4_read_config(struct lpfc_hba *phba)
|
|
{
|
|
LPFC_MBOXQ_t *pmb;
|
|
struct lpfc_mbx_read_config *rd_config;
|
|
uint32_t rc = 0;
|
|
|
|
pmb = (LPFC_MBOXQ_t *) mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
|
|
if (!pmb) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
|
|
"2011 Unable to allocate memory for issuing "
|
|
"SLI_CONFIG_SPECIAL mailbox command\n");
|
|
return -ENOMEM;
|
|
}
|
|
|
|
lpfc_read_config(phba, pmb);
|
|
|
|
rc = lpfc_sli_issue_mbox(phba, pmb, MBX_POLL);
|
|
if (rc != MBX_SUCCESS) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
|
|
"2012 Mailbox failed , mbxCmd x%x "
|
|
"READ_CONFIG, mbxStatus x%x\n",
|
|
bf_get(lpfc_mqe_command, &pmb->u.mqe),
|
|
bf_get(lpfc_mqe_status, &pmb->u.mqe));
|
|
rc = -EIO;
|
|
} else {
|
|
rd_config = &pmb->u.mqe.un.rd_config;
|
|
phba->sli4_hba.max_cfg_param.max_xri =
|
|
bf_get(lpfc_mbx_rd_conf_xri_count, rd_config);
|
|
phba->sli4_hba.max_cfg_param.xri_base =
|
|
bf_get(lpfc_mbx_rd_conf_xri_base, rd_config);
|
|
phba->sli4_hba.max_cfg_param.max_vpi =
|
|
bf_get(lpfc_mbx_rd_conf_vpi_count, rd_config);
|
|
phba->sli4_hba.max_cfg_param.vpi_base =
|
|
bf_get(lpfc_mbx_rd_conf_vpi_base, rd_config);
|
|
phba->sli4_hba.max_cfg_param.max_rpi =
|
|
bf_get(lpfc_mbx_rd_conf_rpi_count, rd_config);
|
|
phba->sli4_hba.max_cfg_param.rpi_base =
|
|
bf_get(lpfc_mbx_rd_conf_rpi_base, rd_config);
|
|
phba->sli4_hba.max_cfg_param.max_vfi =
|
|
bf_get(lpfc_mbx_rd_conf_vfi_count, rd_config);
|
|
phba->sli4_hba.max_cfg_param.vfi_base =
|
|
bf_get(lpfc_mbx_rd_conf_vfi_base, rd_config);
|
|
phba->sli4_hba.max_cfg_param.max_fcfi =
|
|
bf_get(lpfc_mbx_rd_conf_fcfi_count, rd_config);
|
|
phba->sli4_hba.max_cfg_param.fcfi_base =
|
|
bf_get(lpfc_mbx_rd_conf_fcfi_base, rd_config);
|
|
phba->sli4_hba.max_cfg_param.max_eq =
|
|
bf_get(lpfc_mbx_rd_conf_eq_count, rd_config);
|
|
phba->sli4_hba.max_cfg_param.max_rq =
|
|
bf_get(lpfc_mbx_rd_conf_rq_count, rd_config);
|
|
phba->sli4_hba.max_cfg_param.max_wq =
|
|
bf_get(lpfc_mbx_rd_conf_wq_count, rd_config);
|
|
phba->sli4_hba.max_cfg_param.max_cq =
|
|
bf_get(lpfc_mbx_rd_conf_cq_count, rd_config);
|
|
phba->lmt = bf_get(lpfc_mbx_rd_conf_lmt, rd_config);
|
|
phba->sli4_hba.next_xri = phba->sli4_hba.max_cfg_param.xri_base;
|
|
phba->vpi_base = phba->sli4_hba.max_cfg_param.vpi_base;
|
|
phba->vfi_base = phba->sli4_hba.max_cfg_param.vfi_base;
|
|
phba->sli4_hba.next_rpi = phba->sli4_hba.max_cfg_param.rpi_base;
|
|
phba->max_vpi = (phba->sli4_hba.max_cfg_param.max_vpi > 0) ?
|
|
(phba->sli4_hba.max_cfg_param.max_vpi - 1) : 0;
|
|
phba->max_vports = phba->max_vpi;
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_SLI,
|
|
"2003 cfg params XRI(B:%d M:%d), "
|
|
"VPI(B:%d M:%d) "
|
|
"VFI(B:%d M:%d) "
|
|
"RPI(B:%d M:%d) "
|
|
"FCFI(B:%d M:%d)\n",
|
|
phba->sli4_hba.max_cfg_param.xri_base,
|
|
phba->sli4_hba.max_cfg_param.max_xri,
|
|
phba->sli4_hba.max_cfg_param.vpi_base,
|
|
phba->sli4_hba.max_cfg_param.max_vpi,
|
|
phba->sli4_hba.max_cfg_param.vfi_base,
|
|
phba->sli4_hba.max_cfg_param.max_vfi,
|
|
phba->sli4_hba.max_cfg_param.rpi_base,
|
|
phba->sli4_hba.max_cfg_param.max_rpi,
|
|
phba->sli4_hba.max_cfg_param.fcfi_base,
|
|
phba->sli4_hba.max_cfg_param.max_fcfi);
|
|
}
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
|
|
/* Reset the DFT_HBA_Q_DEPTH to the max xri */
|
|
if (phba->cfg_hba_queue_depth >
|
|
(phba->sli4_hba.max_cfg_param.max_xri -
|
|
lpfc_sli4_get_els_iocb_cnt(phba)))
|
|
phba->cfg_hba_queue_depth =
|
|
phba->sli4_hba.max_cfg_param.max_xri -
|
|
lpfc_sli4_get_els_iocb_cnt(phba);
|
|
return rc;
|
|
}
|
|
|
|
/**
|
|
* lpfc_dev_endian_order_setup - Notify the port of the host's endian order.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to setup the host-side endian order to the
|
|
* HBA consistent with the SLI-4 interface spec.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* -ENOMEM - No availble memory
|
|
* -EIO - The mailbox failed to complete successfully.
|
|
**/
|
|
static int
|
|
lpfc_setup_endian_order(struct lpfc_hba *phba)
|
|
{
|
|
LPFC_MBOXQ_t *mboxq;
|
|
uint32_t rc = 0;
|
|
uint32_t endian_mb_data[2] = {HOST_ENDIAN_LOW_WORD0,
|
|
HOST_ENDIAN_HIGH_WORD1};
|
|
|
|
mboxq = (LPFC_MBOXQ_t *) mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
|
|
if (!mboxq) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0492 Unable to allocate memory for issuing "
|
|
"SLI_CONFIG_SPECIAL mailbox command\n");
|
|
return -ENOMEM;
|
|
}
|
|
|
|
/*
|
|
* The SLI4_CONFIG_SPECIAL mailbox command requires the first two
|
|
* words to contain special data values and no other data.
|
|
*/
|
|
memset(mboxq, 0, sizeof(LPFC_MBOXQ_t));
|
|
memcpy(&mboxq->u.mqe, &endian_mb_data, sizeof(endian_mb_data));
|
|
rc = lpfc_sli_issue_mbox(phba, mboxq, MBX_POLL);
|
|
if (rc != MBX_SUCCESS) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0493 SLI_CONFIG_SPECIAL mailbox failed with "
|
|
"status x%x\n",
|
|
rc);
|
|
rc = -EIO;
|
|
}
|
|
|
|
mempool_free(mboxq, phba->mbox_mem_pool);
|
|
return rc;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_queue_create - Create all the SLI4 queues
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to allocate all the SLI4 queues for the FCoE HBA
|
|
* operation. For each SLI4 queue type, the parameters such as queue entry
|
|
* count (queue depth) shall be taken from the module parameter. For now,
|
|
* we just use some constant number as place holder.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* -ENOMEM - No availble memory
|
|
* -EIO - The mailbox failed to complete successfully.
|
|
**/
|
|
static int
|
|
lpfc_sli4_queue_create(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_queue *qdesc;
|
|
int fcp_eqidx, fcp_cqidx, fcp_wqidx;
|
|
int cfg_fcp_wq_count;
|
|
int cfg_fcp_eq_count;
|
|
|
|
/*
|
|
* Sanity check for confiugred queue parameters against the run-time
|
|
* device parameters
|
|
*/
|
|
|
|
/* Sanity check on FCP fast-path WQ parameters */
|
|
cfg_fcp_wq_count = phba->cfg_fcp_wq_count;
|
|
if (cfg_fcp_wq_count >
|
|
(phba->sli4_hba.max_cfg_param.max_wq - LPFC_SP_WQN_DEF)) {
|
|
cfg_fcp_wq_count = phba->sli4_hba.max_cfg_param.max_wq -
|
|
LPFC_SP_WQN_DEF;
|
|
if (cfg_fcp_wq_count < LPFC_FP_WQN_MIN) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2581 Not enough WQs (%d) from "
|
|
"the pci function for supporting "
|
|
"FCP WQs (%d)\n",
|
|
phba->sli4_hba.max_cfg_param.max_wq,
|
|
phba->cfg_fcp_wq_count);
|
|
goto out_error;
|
|
}
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_INIT,
|
|
"2582 Not enough WQs (%d) from the pci "
|
|
"function for supporting the requested "
|
|
"FCP WQs (%d), the actual FCP WQs can "
|
|
"be supported: %d\n",
|
|
phba->sli4_hba.max_cfg_param.max_wq,
|
|
phba->cfg_fcp_wq_count, cfg_fcp_wq_count);
|
|
}
|
|
/* The actual number of FCP work queues adopted */
|
|
phba->cfg_fcp_wq_count = cfg_fcp_wq_count;
|
|
|
|
/* Sanity check on FCP fast-path EQ parameters */
|
|
cfg_fcp_eq_count = phba->cfg_fcp_eq_count;
|
|
if (cfg_fcp_eq_count >
|
|
(phba->sli4_hba.max_cfg_param.max_eq - LPFC_SP_EQN_DEF)) {
|
|
cfg_fcp_eq_count = phba->sli4_hba.max_cfg_param.max_eq -
|
|
LPFC_SP_EQN_DEF;
|
|
if (cfg_fcp_eq_count < LPFC_FP_EQN_MIN) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2574 Not enough EQs (%d) from the "
|
|
"pci function for supporting FCP "
|
|
"EQs (%d)\n",
|
|
phba->sli4_hba.max_cfg_param.max_eq,
|
|
phba->cfg_fcp_eq_count);
|
|
goto out_error;
|
|
}
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_INIT,
|
|
"2575 Not enough EQs (%d) from the pci "
|
|
"function for supporting the requested "
|
|
"FCP EQs (%d), the actual FCP EQs can "
|
|
"be supported: %d\n",
|
|
phba->sli4_hba.max_cfg_param.max_eq,
|
|
phba->cfg_fcp_eq_count, cfg_fcp_eq_count);
|
|
}
|
|
/* It does not make sense to have more EQs than WQs */
|
|
if (cfg_fcp_eq_count > phba->cfg_fcp_wq_count) {
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_INIT,
|
|
"2593 The FCP EQ count(%d) cannot be greater "
|
|
"than the FCP WQ count(%d), limiting the "
|
|
"FCP EQ count to %d\n", cfg_fcp_eq_count,
|
|
phba->cfg_fcp_wq_count,
|
|
phba->cfg_fcp_wq_count);
|
|
cfg_fcp_eq_count = phba->cfg_fcp_wq_count;
|
|
}
|
|
/* The actual number of FCP event queues adopted */
|
|
phba->cfg_fcp_eq_count = cfg_fcp_eq_count;
|
|
/* The overall number of event queues used */
|
|
phba->sli4_hba.cfg_eqn = phba->cfg_fcp_eq_count + LPFC_SP_EQN_DEF;
|
|
|
|
/*
|
|
* Create Event Queues (EQs)
|
|
*/
|
|
|
|
/* Get EQ depth from module parameter, fake the default for now */
|
|
phba->sli4_hba.eq_esize = LPFC_EQE_SIZE_4B;
|
|
phba->sli4_hba.eq_ecount = LPFC_EQE_DEF_COUNT;
|
|
|
|
/* Create slow path event queue */
|
|
qdesc = lpfc_sli4_queue_alloc(phba, phba->sli4_hba.eq_esize,
|
|
phba->sli4_hba.eq_ecount);
|
|
if (!qdesc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0496 Failed allocate slow-path EQ\n");
|
|
goto out_error;
|
|
}
|
|
phba->sli4_hba.sp_eq = qdesc;
|
|
|
|
/* Create fast-path FCP Event Queue(s) */
|
|
phba->sli4_hba.fp_eq = kzalloc((sizeof(struct lpfc_queue *) *
|
|
phba->cfg_fcp_eq_count), GFP_KERNEL);
|
|
if (!phba->sli4_hba.fp_eq) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2576 Failed allocate memory for fast-path "
|
|
"EQ record array\n");
|
|
goto out_free_sp_eq;
|
|
}
|
|
for (fcp_eqidx = 0; fcp_eqidx < phba->cfg_fcp_eq_count; fcp_eqidx++) {
|
|
qdesc = lpfc_sli4_queue_alloc(phba, phba->sli4_hba.eq_esize,
|
|
phba->sli4_hba.eq_ecount);
|
|
if (!qdesc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0497 Failed allocate fast-path EQ\n");
|
|
goto out_free_fp_eq;
|
|
}
|
|
phba->sli4_hba.fp_eq[fcp_eqidx] = qdesc;
|
|
}
|
|
|
|
/*
|
|
* Create Complete Queues (CQs)
|
|
*/
|
|
|
|
/* Get CQ depth from module parameter, fake the default for now */
|
|
phba->sli4_hba.cq_esize = LPFC_CQE_SIZE;
|
|
phba->sli4_hba.cq_ecount = LPFC_CQE_DEF_COUNT;
|
|
|
|
/* Create slow-path Mailbox Command Complete Queue */
|
|
qdesc = lpfc_sli4_queue_alloc(phba, phba->sli4_hba.cq_esize,
|
|
phba->sli4_hba.cq_ecount);
|
|
if (!qdesc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0500 Failed allocate slow-path mailbox CQ\n");
|
|
goto out_free_fp_eq;
|
|
}
|
|
phba->sli4_hba.mbx_cq = qdesc;
|
|
|
|
/* Create slow-path ELS Complete Queue */
|
|
qdesc = lpfc_sli4_queue_alloc(phba, phba->sli4_hba.cq_esize,
|
|
phba->sli4_hba.cq_ecount);
|
|
if (!qdesc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0501 Failed allocate slow-path ELS CQ\n");
|
|
goto out_free_mbx_cq;
|
|
}
|
|
phba->sli4_hba.els_cq = qdesc;
|
|
|
|
|
|
/* Create fast-path FCP Completion Queue(s), one-to-one with EQs */
|
|
phba->sli4_hba.fcp_cq = kzalloc((sizeof(struct lpfc_queue *) *
|
|
phba->cfg_fcp_eq_count), GFP_KERNEL);
|
|
if (!phba->sli4_hba.fcp_cq) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2577 Failed allocate memory for fast-path "
|
|
"CQ record array\n");
|
|
goto out_free_els_cq;
|
|
}
|
|
for (fcp_cqidx = 0; fcp_cqidx < phba->cfg_fcp_eq_count; fcp_cqidx++) {
|
|
qdesc = lpfc_sli4_queue_alloc(phba, phba->sli4_hba.cq_esize,
|
|
phba->sli4_hba.cq_ecount);
|
|
if (!qdesc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0499 Failed allocate fast-path FCP "
|
|
"CQ (%d)\n", fcp_cqidx);
|
|
goto out_free_fcp_cq;
|
|
}
|
|
phba->sli4_hba.fcp_cq[fcp_cqidx] = qdesc;
|
|
}
|
|
|
|
/* Create Mailbox Command Queue */
|
|
phba->sli4_hba.mq_esize = LPFC_MQE_SIZE;
|
|
phba->sli4_hba.mq_ecount = LPFC_MQE_DEF_COUNT;
|
|
|
|
qdesc = lpfc_sli4_queue_alloc(phba, phba->sli4_hba.mq_esize,
|
|
phba->sli4_hba.mq_ecount);
|
|
if (!qdesc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0505 Failed allocate slow-path MQ\n");
|
|
goto out_free_fcp_cq;
|
|
}
|
|
phba->sli4_hba.mbx_wq = qdesc;
|
|
|
|
/*
|
|
* Create all the Work Queues (WQs)
|
|
*/
|
|
phba->sli4_hba.wq_esize = LPFC_WQE_SIZE;
|
|
phba->sli4_hba.wq_ecount = LPFC_WQE_DEF_COUNT;
|
|
|
|
/* Create slow-path ELS Work Queue */
|
|
qdesc = lpfc_sli4_queue_alloc(phba, phba->sli4_hba.wq_esize,
|
|
phba->sli4_hba.wq_ecount);
|
|
if (!qdesc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0504 Failed allocate slow-path ELS WQ\n");
|
|
goto out_free_mbx_wq;
|
|
}
|
|
phba->sli4_hba.els_wq = qdesc;
|
|
|
|
/* Create fast-path FCP Work Queue(s) */
|
|
phba->sli4_hba.fcp_wq = kzalloc((sizeof(struct lpfc_queue *) *
|
|
phba->cfg_fcp_wq_count), GFP_KERNEL);
|
|
if (!phba->sli4_hba.fcp_wq) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2578 Failed allocate memory for fast-path "
|
|
"WQ record array\n");
|
|
goto out_free_els_wq;
|
|
}
|
|
for (fcp_wqidx = 0; fcp_wqidx < phba->cfg_fcp_wq_count; fcp_wqidx++) {
|
|
qdesc = lpfc_sli4_queue_alloc(phba, phba->sli4_hba.wq_esize,
|
|
phba->sli4_hba.wq_ecount);
|
|
if (!qdesc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0503 Failed allocate fast-path FCP "
|
|
"WQ (%d)\n", fcp_wqidx);
|
|
goto out_free_fcp_wq;
|
|
}
|
|
phba->sli4_hba.fcp_wq[fcp_wqidx] = qdesc;
|
|
}
|
|
|
|
/*
|
|
* Create Receive Queue (RQ)
|
|
*/
|
|
phba->sli4_hba.rq_esize = LPFC_RQE_SIZE;
|
|
phba->sli4_hba.rq_ecount = LPFC_RQE_DEF_COUNT;
|
|
|
|
/* Create Receive Queue for header */
|
|
qdesc = lpfc_sli4_queue_alloc(phba, phba->sli4_hba.rq_esize,
|
|
phba->sli4_hba.rq_ecount);
|
|
if (!qdesc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0506 Failed allocate receive HRQ\n");
|
|
goto out_free_fcp_wq;
|
|
}
|
|
phba->sli4_hba.hdr_rq = qdesc;
|
|
|
|
/* Create Receive Queue for data */
|
|
qdesc = lpfc_sli4_queue_alloc(phba, phba->sli4_hba.rq_esize,
|
|
phba->sli4_hba.rq_ecount);
|
|
if (!qdesc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0507 Failed allocate receive DRQ\n");
|
|
goto out_free_hdr_rq;
|
|
}
|
|
phba->sli4_hba.dat_rq = qdesc;
|
|
|
|
return 0;
|
|
|
|
out_free_hdr_rq:
|
|
lpfc_sli4_queue_free(phba->sli4_hba.hdr_rq);
|
|
phba->sli4_hba.hdr_rq = NULL;
|
|
out_free_fcp_wq:
|
|
for (--fcp_wqidx; fcp_wqidx >= 0; fcp_wqidx--) {
|
|
lpfc_sli4_queue_free(phba->sli4_hba.fcp_wq[fcp_wqidx]);
|
|
phba->sli4_hba.fcp_wq[fcp_wqidx] = NULL;
|
|
}
|
|
kfree(phba->sli4_hba.fcp_wq);
|
|
out_free_els_wq:
|
|
lpfc_sli4_queue_free(phba->sli4_hba.els_wq);
|
|
phba->sli4_hba.els_wq = NULL;
|
|
out_free_mbx_wq:
|
|
lpfc_sli4_queue_free(phba->sli4_hba.mbx_wq);
|
|
phba->sli4_hba.mbx_wq = NULL;
|
|
out_free_fcp_cq:
|
|
for (--fcp_cqidx; fcp_cqidx >= 0; fcp_cqidx--) {
|
|
lpfc_sli4_queue_free(phba->sli4_hba.fcp_cq[fcp_cqidx]);
|
|
phba->sli4_hba.fcp_cq[fcp_cqidx] = NULL;
|
|
}
|
|
kfree(phba->sli4_hba.fcp_cq);
|
|
out_free_els_cq:
|
|
lpfc_sli4_queue_free(phba->sli4_hba.els_cq);
|
|
phba->sli4_hba.els_cq = NULL;
|
|
out_free_mbx_cq:
|
|
lpfc_sli4_queue_free(phba->sli4_hba.mbx_cq);
|
|
phba->sli4_hba.mbx_cq = NULL;
|
|
out_free_fp_eq:
|
|
for (--fcp_eqidx; fcp_eqidx >= 0; fcp_eqidx--) {
|
|
lpfc_sli4_queue_free(phba->sli4_hba.fp_eq[fcp_eqidx]);
|
|
phba->sli4_hba.fp_eq[fcp_eqidx] = NULL;
|
|
}
|
|
kfree(phba->sli4_hba.fp_eq);
|
|
out_free_sp_eq:
|
|
lpfc_sli4_queue_free(phba->sli4_hba.sp_eq);
|
|
phba->sli4_hba.sp_eq = NULL;
|
|
out_error:
|
|
return -ENOMEM;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_queue_destroy - Destroy all the SLI4 queues
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to release all the SLI4 queues with the FCoE HBA
|
|
* operation.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* -ENOMEM - No availble memory
|
|
* -EIO - The mailbox failed to complete successfully.
|
|
**/
|
|
static void
|
|
lpfc_sli4_queue_destroy(struct lpfc_hba *phba)
|
|
{
|
|
int fcp_qidx;
|
|
|
|
/* Release mailbox command work queue */
|
|
lpfc_sli4_queue_free(phba->sli4_hba.mbx_wq);
|
|
phba->sli4_hba.mbx_wq = NULL;
|
|
|
|
/* Release ELS work queue */
|
|
lpfc_sli4_queue_free(phba->sli4_hba.els_wq);
|
|
phba->sli4_hba.els_wq = NULL;
|
|
|
|
/* Release FCP work queue */
|
|
for (fcp_qidx = 0; fcp_qidx < phba->cfg_fcp_wq_count; fcp_qidx++)
|
|
lpfc_sli4_queue_free(phba->sli4_hba.fcp_wq[fcp_qidx]);
|
|
kfree(phba->sli4_hba.fcp_wq);
|
|
phba->sli4_hba.fcp_wq = NULL;
|
|
|
|
/* Release unsolicited receive queue */
|
|
lpfc_sli4_queue_free(phba->sli4_hba.hdr_rq);
|
|
phba->sli4_hba.hdr_rq = NULL;
|
|
lpfc_sli4_queue_free(phba->sli4_hba.dat_rq);
|
|
phba->sli4_hba.dat_rq = NULL;
|
|
|
|
/* Release ELS complete queue */
|
|
lpfc_sli4_queue_free(phba->sli4_hba.els_cq);
|
|
phba->sli4_hba.els_cq = NULL;
|
|
|
|
/* Release mailbox command complete queue */
|
|
lpfc_sli4_queue_free(phba->sli4_hba.mbx_cq);
|
|
phba->sli4_hba.mbx_cq = NULL;
|
|
|
|
/* Release FCP response complete queue */
|
|
for (fcp_qidx = 0; fcp_qidx < phba->cfg_fcp_eq_count; fcp_qidx++)
|
|
lpfc_sli4_queue_free(phba->sli4_hba.fcp_cq[fcp_qidx]);
|
|
kfree(phba->sli4_hba.fcp_cq);
|
|
phba->sli4_hba.fcp_cq = NULL;
|
|
|
|
/* Release fast-path event queue */
|
|
for (fcp_qidx = 0; fcp_qidx < phba->cfg_fcp_eq_count; fcp_qidx++)
|
|
lpfc_sli4_queue_free(phba->sli4_hba.fp_eq[fcp_qidx]);
|
|
kfree(phba->sli4_hba.fp_eq);
|
|
phba->sli4_hba.fp_eq = NULL;
|
|
|
|
/* Release slow-path event queue */
|
|
lpfc_sli4_queue_free(phba->sli4_hba.sp_eq);
|
|
phba->sli4_hba.sp_eq = NULL;
|
|
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_queue_setup - Set up all the SLI4 queues
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to set up all the SLI4 queues for the FCoE HBA
|
|
* operation.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* -ENOMEM - No availble memory
|
|
* -EIO - The mailbox failed to complete successfully.
|
|
**/
|
|
int
|
|
lpfc_sli4_queue_setup(struct lpfc_hba *phba)
|
|
{
|
|
int rc = -ENOMEM;
|
|
int fcp_eqidx, fcp_cqidx, fcp_wqidx;
|
|
int fcp_cq_index = 0;
|
|
|
|
/*
|
|
* Set up Event Queues (EQs)
|
|
*/
|
|
|
|
/* Set up slow-path event queue */
|
|
if (!phba->sli4_hba.sp_eq) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0520 Slow-path EQ not allocated\n");
|
|
goto out_error;
|
|
}
|
|
rc = lpfc_eq_create(phba, phba->sli4_hba.sp_eq,
|
|
LPFC_SP_DEF_IMAX);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0521 Failed setup of slow-path EQ: "
|
|
"rc = 0x%x\n", rc);
|
|
goto out_error;
|
|
}
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"2583 Slow-path EQ setup: queue-id=%d\n",
|
|
phba->sli4_hba.sp_eq->queue_id);
|
|
|
|
/* Set up fast-path event queue */
|
|
for (fcp_eqidx = 0; fcp_eqidx < phba->cfg_fcp_eq_count; fcp_eqidx++) {
|
|
if (!phba->sli4_hba.fp_eq[fcp_eqidx]) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0522 Fast-path EQ (%d) not "
|
|
"allocated\n", fcp_eqidx);
|
|
goto out_destroy_fp_eq;
|
|
}
|
|
rc = lpfc_eq_create(phba, phba->sli4_hba.fp_eq[fcp_eqidx],
|
|
phba->cfg_fcp_imax);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0523 Failed setup of fast-path EQ "
|
|
"(%d), rc = 0x%x\n", fcp_eqidx, rc);
|
|
goto out_destroy_fp_eq;
|
|
}
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"2584 Fast-path EQ setup: "
|
|
"queue[%d]-id=%d\n", fcp_eqidx,
|
|
phba->sli4_hba.fp_eq[fcp_eqidx]->queue_id);
|
|
}
|
|
|
|
/*
|
|
* Set up Complete Queues (CQs)
|
|
*/
|
|
|
|
/* Set up slow-path MBOX Complete Queue as the first CQ */
|
|
if (!phba->sli4_hba.mbx_cq) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0528 Mailbox CQ not allocated\n");
|
|
goto out_destroy_fp_eq;
|
|
}
|
|
rc = lpfc_cq_create(phba, phba->sli4_hba.mbx_cq, phba->sli4_hba.sp_eq,
|
|
LPFC_MCQ, LPFC_MBOX);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0529 Failed setup of slow-path mailbox CQ: "
|
|
"rc = 0x%x\n", rc);
|
|
goto out_destroy_fp_eq;
|
|
}
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"2585 MBX CQ setup: cq-id=%d, parent eq-id=%d\n",
|
|
phba->sli4_hba.mbx_cq->queue_id,
|
|
phba->sli4_hba.sp_eq->queue_id);
|
|
|
|
/* Set up slow-path ELS Complete Queue */
|
|
if (!phba->sli4_hba.els_cq) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0530 ELS CQ not allocated\n");
|
|
goto out_destroy_mbx_cq;
|
|
}
|
|
rc = lpfc_cq_create(phba, phba->sli4_hba.els_cq, phba->sli4_hba.sp_eq,
|
|
LPFC_WCQ, LPFC_ELS);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0531 Failed setup of slow-path ELS CQ: "
|
|
"rc = 0x%x\n", rc);
|
|
goto out_destroy_mbx_cq;
|
|
}
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"2586 ELS CQ setup: cq-id=%d, parent eq-id=%d\n",
|
|
phba->sli4_hba.els_cq->queue_id,
|
|
phba->sli4_hba.sp_eq->queue_id);
|
|
|
|
/* Set up fast-path FCP Response Complete Queue */
|
|
for (fcp_cqidx = 0; fcp_cqidx < phba->cfg_fcp_eq_count; fcp_cqidx++) {
|
|
if (!phba->sli4_hba.fcp_cq[fcp_cqidx]) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0526 Fast-path FCP CQ (%d) not "
|
|
"allocated\n", fcp_cqidx);
|
|
goto out_destroy_fcp_cq;
|
|
}
|
|
rc = lpfc_cq_create(phba, phba->sli4_hba.fcp_cq[fcp_cqidx],
|
|
phba->sli4_hba.fp_eq[fcp_cqidx],
|
|
LPFC_WCQ, LPFC_FCP);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0527 Failed setup of fast-path FCP "
|
|
"CQ (%d), rc = 0x%x\n", fcp_cqidx, rc);
|
|
goto out_destroy_fcp_cq;
|
|
}
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"2588 FCP CQ setup: cq[%d]-id=%d, "
|
|
"parent eq[%d]-id=%d\n",
|
|
fcp_cqidx,
|
|
phba->sli4_hba.fcp_cq[fcp_cqidx]->queue_id,
|
|
fcp_cqidx,
|
|
phba->sli4_hba.fp_eq[fcp_cqidx]->queue_id);
|
|
}
|
|
|
|
/*
|
|
* Set up all the Work Queues (WQs)
|
|
*/
|
|
|
|
/* Set up Mailbox Command Queue */
|
|
if (!phba->sli4_hba.mbx_wq) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0538 Slow-path MQ not allocated\n");
|
|
goto out_destroy_fcp_cq;
|
|
}
|
|
rc = lpfc_mq_create(phba, phba->sli4_hba.mbx_wq,
|
|
phba->sli4_hba.mbx_cq, LPFC_MBOX);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0539 Failed setup of slow-path MQ: "
|
|
"rc = 0x%x\n", rc);
|
|
goto out_destroy_fcp_cq;
|
|
}
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"2589 MBX MQ setup: wq-id=%d, parent cq-id=%d\n",
|
|
phba->sli4_hba.mbx_wq->queue_id,
|
|
phba->sli4_hba.mbx_cq->queue_id);
|
|
|
|
/* Set up slow-path ELS Work Queue */
|
|
if (!phba->sli4_hba.els_wq) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0536 Slow-path ELS WQ not allocated\n");
|
|
goto out_destroy_mbx_wq;
|
|
}
|
|
rc = lpfc_wq_create(phba, phba->sli4_hba.els_wq,
|
|
phba->sli4_hba.els_cq, LPFC_ELS);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0537 Failed setup of slow-path ELS WQ: "
|
|
"rc = 0x%x\n", rc);
|
|
goto out_destroy_mbx_wq;
|
|
}
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"2590 ELS WQ setup: wq-id=%d, parent cq-id=%d\n",
|
|
phba->sli4_hba.els_wq->queue_id,
|
|
phba->sli4_hba.els_cq->queue_id);
|
|
|
|
/* Set up fast-path FCP Work Queue */
|
|
for (fcp_wqidx = 0; fcp_wqidx < phba->cfg_fcp_wq_count; fcp_wqidx++) {
|
|
if (!phba->sli4_hba.fcp_wq[fcp_wqidx]) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0534 Fast-path FCP WQ (%d) not "
|
|
"allocated\n", fcp_wqidx);
|
|
goto out_destroy_fcp_wq;
|
|
}
|
|
rc = lpfc_wq_create(phba, phba->sli4_hba.fcp_wq[fcp_wqidx],
|
|
phba->sli4_hba.fcp_cq[fcp_cq_index],
|
|
LPFC_FCP);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0535 Failed setup of fast-path FCP "
|
|
"WQ (%d), rc = 0x%x\n", fcp_wqidx, rc);
|
|
goto out_destroy_fcp_wq;
|
|
}
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"2591 FCP WQ setup: wq[%d]-id=%d, "
|
|
"parent cq[%d]-id=%d\n",
|
|
fcp_wqidx,
|
|
phba->sli4_hba.fcp_wq[fcp_wqidx]->queue_id,
|
|
fcp_cq_index,
|
|
phba->sli4_hba.fcp_cq[fcp_cq_index]->queue_id);
|
|
/* Round robin FCP Work Queue's Completion Queue assignment */
|
|
fcp_cq_index = ((fcp_cq_index + 1) % phba->cfg_fcp_eq_count);
|
|
}
|
|
|
|
/*
|
|
* Create Receive Queue (RQ)
|
|
*/
|
|
if (!phba->sli4_hba.hdr_rq || !phba->sli4_hba.dat_rq) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0540 Receive Queue not allocated\n");
|
|
goto out_destroy_fcp_wq;
|
|
}
|
|
rc = lpfc_rq_create(phba, phba->sli4_hba.hdr_rq, phba->sli4_hba.dat_rq,
|
|
phba->sli4_hba.els_cq, LPFC_USOL);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0541 Failed setup of Receive Queue: "
|
|
"rc = 0x%x\n", rc);
|
|
goto out_destroy_fcp_wq;
|
|
}
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"2592 USL RQ setup: hdr-rq-id=%d, dat-rq-id=%d "
|
|
"parent cq-id=%d\n",
|
|
phba->sli4_hba.hdr_rq->queue_id,
|
|
phba->sli4_hba.dat_rq->queue_id,
|
|
phba->sli4_hba.els_cq->queue_id);
|
|
return 0;
|
|
|
|
out_destroy_fcp_wq:
|
|
for (--fcp_wqidx; fcp_wqidx >= 0; fcp_wqidx--)
|
|
lpfc_wq_destroy(phba, phba->sli4_hba.fcp_wq[fcp_wqidx]);
|
|
lpfc_wq_destroy(phba, phba->sli4_hba.els_wq);
|
|
out_destroy_mbx_wq:
|
|
lpfc_mq_destroy(phba, phba->sli4_hba.mbx_wq);
|
|
out_destroy_fcp_cq:
|
|
for (--fcp_cqidx; fcp_cqidx >= 0; fcp_cqidx--)
|
|
lpfc_cq_destroy(phba, phba->sli4_hba.fcp_cq[fcp_cqidx]);
|
|
lpfc_cq_destroy(phba, phba->sli4_hba.els_cq);
|
|
out_destroy_mbx_cq:
|
|
lpfc_cq_destroy(phba, phba->sli4_hba.mbx_cq);
|
|
out_destroy_fp_eq:
|
|
for (--fcp_eqidx; fcp_eqidx >= 0; fcp_eqidx--)
|
|
lpfc_eq_destroy(phba, phba->sli4_hba.fp_eq[fcp_eqidx]);
|
|
lpfc_eq_destroy(phba, phba->sli4_hba.sp_eq);
|
|
out_error:
|
|
return rc;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_queue_unset - Unset all the SLI4 queues
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to unset all the SLI4 queues with the FCoE HBA
|
|
* operation.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* -ENOMEM - No availble memory
|
|
* -EIO - The mailbox failed to complete successfully.
|
|
**/
|
|
void
|
|
lpfc_sli4_queue_unset(struct lpfc_hba *phba)
|
|
{
|
|
int fcp_qidx;
|
|
|
|
/* Unset mailbox command work queue */
|
|
lpfc_mq_destroy(phba, phba->sli4_hba.mbx_wq);
|
|
/* Unset ELS work queue */
|
|
lpfc_wq_destroy(phba, phba->sli4_hba.els_wq);
|
|
/* Unset unsolicited receive queue */
|
|
lpfc_rq_destroy(phba, phba->sli4_hba.hdr_rq, phba->sli4_hba.dat_rq);
|
|
/* Unset FCP work queue */
|
|
for (fcp_qidx = 0; fcp_qidx < phba->cfg_fcp_wq_count; fcp_qidx++)
|
|
lpfc_wq_destroy(phba, phba->sli4_hba.fcp_wq[fcp_qidx]);
|
|
/* Unset mailbox command complete queue */
|
|
lpfc_cq_destroy(phba, phba->sli4_hba.mbx_cq);
|
|
/* Unset ELS complete queue */
|
|
lpfc_cq_destroy(phba, phba->sli4_hba.els_cq);
|
|
/* Unset FCP response complete queue */
|
|
for (fcp_qidx = 0; fcp_qidx < phba->cfg_fcp_eq_count; fcp_qidx++)
|
|
lpfc_cq_destroy(phba, phba->sli4_hba.fcp_cq[fcp_qidx]);
|
|
/* Unset fast-path event queue */
|
|
for (fcp_qidx = 0; fcp_qidx < phba->cfg_fcp_eq_count; fcp_qidx++)
|
|
lpfc_eq_destroy(phba, phba->sli4_hba.fp_eq[fcp_qidx]);
|
|
/* Unset slow-path event queue */
|
|
lpfc_eq_destroy(phba, phba->sli4_hba.sp_eq);
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_cq_event_pool_create - Create completion-queue event free pool
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to allocate and set up a pool of completion queue
|
|
* events. The body of the completion queue event is a completion queue entry
|
|
* CQE. For now, this pool is used for the interrupt service routine to queue
|
|
* the following HBA completion queue events for the worker thread to process:
|
|
* - Mailbox asynchronous events
|
|
* - Receive queue completion unsolicited events
|
|
* Later, this can be used for all the slow-path events.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* -ENOMEM - No availble memory
|
|
**/
|
|
static int
|
|
lpfc_sli4_cq_event_pool_create(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_cq_event *cq_event;
|
|
int i;
|
|
|
|
for (i = 0; i < (4 * phba->sli4_hba.cq_ecount); i++) {
|
|
cq_event = kmalloc(sizeof(struct lpfc_cq_event), GFP_KERNEL);
|
|
if (!cq_event)
|
|
goto out_pool_create_fail;
|
|
list_add_tail(&cq_event->list,
|
|
&phba->sli4_hba.sp_cqe_event_pool);
|
|
}
|
|
return 0;
|
|
|
|
out_pool_create_fail:
|
|
lpfc_sli4_cq_event_pool_destroy(phba);
|
|
return -ENOMEM;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_cq_event_pool_destroy - Free completion-queue event free pool
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to free the pool of completion queue events at
|
|
* driver unload time. Note that, it is the responsibility of the driver
|
|
* cleanup routine to free all the outstanding completion-queue events
|
|
* allocated from this pool back into the pool before invoking this routine
|
|
* to destroy the pool.
|
|
**/
|
|
static void
|
|
lpfc_sli4_cq_event_pool_destroy(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_cq_event *cq_event, *next_cq_event;
|
|
|
|
list_for_each_entry_safe(cq_event, next_cq_event,
|
|
&phba->sli4_hba.sp_cqe_event_pool, list) {
|
|
list_del(&cq_event->list);
|
|
kfree(cq_event);
|
|
}
|
|
}
|
|
|
|
/**
|
|
* __lpfc_sli4_cq_event_alloc - Allocate a completion-queue event from free pool
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is the lock free version of the API invoked to allocate a
|
|
* completion-queue event from the free pool.
|
|
*
|
|
* Return: Pointer to the newly allocated completion-queue event if successful
|
|
* NULL otherwise.
|
|
**/
|
|
struct lpfc_cq_event *
|
|
__lpfc_sli4_cq_event_alloc(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_cq_event *cq_event = NULL;
|
|
|
|
list_remove_head(&phba->sli4_hba.sp_cqe_event_pool, cq_event,
|
|
struct lpfc_cq_event, list);
|
|
return cq_event;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_cq_event_alloc - Allocate a completion-queue event from free pool
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is the lock version of the API invoked to allocate a
|
|
* completion-queue event from the free pool.
|
|
*
|
|
* Return: Pointer to the newly allocated completion-queue event if successful
|
|
* NULL otherwise.
|
|
**/
|
|
struct lpfc_cq_event *
|
|
lpfc_sli4_cq_event_alloc(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_cq_event *cq_event;
|
|
unsigned long iflags;
|
|
|
|
spin_lock_irqsave(&phba->hbalock, iflags);
|
|
cq_event = __lpfc_sli4_cq_event_alloc(phba);
|
|
spin_unlock_irqrestore(&phba->hbalock, iflags);
|
|
return cq_event;
|
|
}
|
|
|
|
/**
|
|
* __lpfc_sli4_cq_event_release - Release a completion-queue event to free pool
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @cq_event: pointer to the completion queue event to be freed.
|
|
*
|
|
* This routine is the lock free version of the API invoked to release a
|
|
* completion-queue event back into the free pool.
|
|
**/
|
|
void
|
|
__lpfc_sli4_cq_event_release(struct lpfc_hba *phba,
|
|
struct lpfc_cq_event *cq_event)
|
|
{
|
|
list_add_tail(&cq_event->list, &phba->sli4_hba.sp_cqe_event_pool);
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_cq_event_release - Release a completion-queue event to free pool
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @cq_event: pointer to the completion queue event to be freed.
|
|
*
|
|
* This routine is the lock version of the API invoked to release a
|
|
* completion-queue event back into the free pool.
|
|
**/
|
|
void
|
|
lpfc_sli4_cq_event_release(struct lpfc_hba *phba,
|
|
struct lpfc_cq_event *cq_event)
|
|
{
|
|
unsigned long iflags;
|
|
spin_lock_irqsave(&phba->hbalock, iflags);
|
|
__lpfc_sli4_cq_event_release(phba, cq_event);
|
|
spin_unlock_irqrestore(&phba->hbalock, iflags);
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_cq_event_release_all - Release all cq events to the free pool
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is to free all the pending completion-queue events to the
|
|
* back into the free pool for device reset.
|
|
**/
|
|
static void
|
|
lpfc_sli4_cq_event_release_all(struct lpfc_hba *phba)
|
|
{
|
|
LIST_HEAD(cqelist);
|
|
struct lpfc_cq_event *cqe;
|
|
unsigned long iflags;
|
|
|
|
/* Retrieve all the pending WCQEs from pending WCQE lists */
|
|
spin_lock_irqsave(&phba->hbalock, iflags);
|
|
/* Pending FCP XRI abort events */
|
|
list_splice_init(&phba->sli4_hba.sp_fcp_xri_aborted_work_queue,
|
|
&cqelist);
|
|
/* Pending ELS XRI abort events */
|
|
list_splice_init(&phba->sli4_hba.sp_els_xri_aborted_work_queue,
|
|
&cqelist);
|
|
/* Pending asynnc events */
|
|
list_splice_init(&phba->sli4_hba.sp_asynce_work_queue,
|
|
&cqelist);
|
|
spin_unlock_irqrestore(&phba->hbalock, iflags);
|
|
|
|
while (!list_empty(&cqelist)) {
|
|
list_remove_head(&cqelist, cqe, struct lpfc_cq_event, list);
|
|
lpfc_sli4_cq_event_release(phba, cqe);
|
|
}
|
|
}
|
|
|
|
/**
|
|
* lpfc_pci_function_reset - Reset pci function.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to request a PCI function reset. It will destroys
|
|
* all resources assigned to the PCI function which originates this request.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* -ENOMEM - No availble memory
|
|
* -EIO - The mailbox failed to complete successfully.
|
|
**/
|
|
int
|
|
lpfc_pci_function_reset(struct lpfc_hba *phba)
|
|
{
|
|
LPFC_MBOXQ_t *mboxq;
|
|
uint32_t rc = 0;
|
|
uint32_t shdr_status, shdr_add_status;
|
|
union lpfc_sli4_cfg_shdr *shdr;
|
|
|
|
mboxq = (LPFC_MBOXQ_t *) mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
|
|
if (!mboxq) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0494 Unable to allocate memory for issuing "
|
|
"SLI_FUNCTION_RESET mailbox command\n");
|
|
return -ENOMEM;
|
|
}
|
|
|
|
/* Set up PCI function reset SLI4_CONFIG mailbox-ioctl command */
|
|
lpfc_sli4_config(phba, mboxq, LPFC_MBOX_SUBSYSTEM_COMMON,
|
|
LPFC_MBOX_OPCODE_FUNCTION_RESET, 0,
|
|
LPFC_SLI4_MBX_EMBED);
|
|
rc = lpfc_sli_issue_mbox(phba, mboxq, MBX_POLL);
|
|
shdr = (union lpfc_sli4_cfg_shdr *)
|
|
&mboxq->u.mqe.un.sli4_config.header.cfg_shdr;
|
|
shdr_status = bf_get(lpfc_mbox_hdr_status, &shdr->response);
|
|
shdr_add_status = bf_get(lpfc_mbox_hdr_add_status, &shdr->response);
|
|
if (rc != MBX_TIMEOUT)
|
|
mempool_free(mboxq, phba->mbox_mem_pool);
|
|
if (shdr_status || shdr_add_status || rc) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0495 SLI_FUNCTION_RESET mailbox failed with "
|
|
"status x%x add_status x%x, mbx status x%x\n",
|
|
shdr_status, shdr_add_status, rc);
|
|
rc = -ENXIO;
|
|
}
|
|
return rc;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_send_nop_mbox_cmds - Send sli-4 nop mailbox commands
|
|
* @phba: pointer to lpfc hba data structure.
|
|
* @cnt: number of nop mailbox commands to send.
|
|
*
|
|
* This routine is invoked to send a number @cnt of NOP mailbox command and
|
|
* wait for each command to complete.
|
|
*
|
|
* Return: the number of NOP mailbox command completed.
|
|
**/
|
|
static int
|
|
lpfc_sli4_send_nop_mbox_cmds(struct lpfc_hba *phba, uint32_t cnt)
|
|
{
|
|
LPFC_MBOXQ_t *mboxq;
|
|
int length, cmdsent;
|
|
uint32_t mbox_tmo;
|
|
uint32_t rc = 0;
|
|
uint32_t shdr_status, shdr_add_status;
|
|
union lpfc_sli4_cfg_shdr *shdr;
|
|
|
|
if (cnt == 0) {
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_INIT,
|
|
"2518 Requested to send 0 NOP mailbox cmd\n");
|
|
return cnt;
|
|
}
|
|
|
|
mboxq = (LPFC_MBOXQ_t *)mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
|
|
if (!mboxq) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2519 Unable to allocate memory for issuing "
|
|
"NOP mailbox command\n");
|
|
return 0;
|
|
}
|
|
|
|
/* Set up NOP SLI4_CONFIG mailbox-ioctl command */
|
|
length = (sizeof(struct lpfc_mbx_nop) -
|
|
sizeof(struct lpfc_sli4_cfg_mhdr));
|
|
lpfc_sli4_config(phba, mboxq, LPFC_MBOX_SUBSYSTEM_COMMON,
|
|
LPFC_MBOX_OPCODE_NOP, length, LPFC_SLI4_MBX_EMBED);
|
|
|
|
mbox_tmo = lpfc_mbox_tmo_val(phba, MBX_SLI4_CONFIG);
|
|
for (cmdsent = 0; cmdsent < cnt; cmdsent++) {
|
|
if (!phba->sli4_hba.intr_enable)
|
|
rc = lpfc_sli_issue_mbox(phba, mboxq, MBX_POLL);
|
|
else
|
|
rc = lpfc_sli_issue_mbox_wait(phba, mboxq, mbox_tmo);
|
|
if (rc == MBX_TIMEOUT)
|
|
break;
|
|
/* Check return status */
|
|
shdr = (union lpfc_sli4_cfg_shdr *)
|
|
&mboxq->u.mqe.un.sli4_config.header.cfg_shdr;
|
|
shdr_status = bf_get(lpfc_mbox_hdr_status, &shdr->response);
|
|
shdr_add_status = bf_get(lpfc_mbox_hdr_add_status,
|
|
&shdr->response);
|
|
if (shdr_status || shdr_add_status || rc) {
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_INIT,
|
|
"2520 NOP mailbox command failed "
|
|
"status x%x add_status x%x mbx "
|
|
"status x%x\n", shdr_status,
|
|
shdr_add_status, rc);
|
|
break;
|
|
}
|
|
}
|
|
|
|
if (rc != MBX_TIMEOUT)
|
|
mempool_free(mboxq, phba->mbox_mem_pool);
|
|
|
|
return cmdsent;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_pci_mem_setup - Setup SLI4 HBA PCI memory space.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to set up the PCI device memory space for device
|
|
* with SLI-4 interface spec.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* other values - error
|
|
**/
|
|
static int
|
|
lpfc_sli4_pci_mem_setup(struct lpfc_hba *phba)
|
|
{
|
|
struct pci_dev *pdev;
|
|
unsigned long bar0map_len, bar1map_len, bar2map_len;
|
|
int error = -ENODEV;
|
|
|
|
/* Obtain PCI device reference */
|
|
if (!phba->pcidev)
|
|
return error;
|
|
else
|
|
pdev = phba->pcidev;
|
|
|
|
/* Set the device DMA mask size */
|
|
if (pci_set_dma_mask(pdev, DMA_BIT_MASK(64)) != 0
|
|
|| pci_set_consistent_dma_mask(pdev,DMA_BIT_MASK(64)) != 0) {
|
|
if (pci_set_dma_mask(pdev, DMA_BIT_MASK(32)) != 0
|
|
|| pci_set_consistent_dma_mask(pdev,DMA_BIT_MASK(32)) != 0) {
|
|
return error;
|
|
}
|
|
}
|
|
|
|
/* Get the bus address of SLI4 device Bar0, Bar1, and Bar2 and the
|
|
* number of bytes required by each mapping. They are actually
|
|
* mapping to the PCI BAR regions 0 or 1, 2, and 4 by the SLI4 device.
|
|
*/
|
|
if (pci_resource_start(pdev, 0)) {
|
|
phba->pci_bar0_map = pci_resource_start(pdev, 0);
|
|
bar0map_len = pci_resource_len(pdev, 0);
|
|
} else {
|
|
phba->pci_bar0_map = pci_resource_start(pdev, 1);
|
|
bar0map_len = pci_resource_len(pdev, 1);
|
|
}
|
|
phba->pci_bar1_map = pci_resource_start(pdev, 2);
|
|
bar1map_len = pci_resource_len(pdev, 2);
|
|
|
|
phba->pci_bar2_map = pci_resource_start(pdev, 4);
|
|
bar2map_len = pci_resource_len(pdev, 4);
|
|
|
|
/* Map SLI4 PCI Config Space Register base to a kernel virtual addr */
|
|
phba->sli4_hba.conf_regs_memmap_p =
|
|
ioremap(phba->pci_bar0_map, bar0map_len);
|
|
if (!phba->sli4_hba.conf_regs_memmap_p) {
|
|
dev_printk(KERN_ERR, &pdev->dev,
|
|
"ioremap failed for SLI4 PCI config registers.\n");
|
|
goto out;
|
|
}
|
|
|
|
/* Map SLI4 HBA Control Register base to a kernel virtual address. */
|
|
phba->sli4_hba.ctrl_regs_memmap_p =
|
|
ioremap(phba->pci_bar1_map, bar1map_len);
|
|
if (!phba->sli4_hba.ctrl_regs_memmap_p) {
|
|
dev_printk(KERN_ERR, &pdev->dev,
|
|
"ioremap failed for SLI4 HBA control registers.\n");
|
|
goto out_iounmap_conf;
|
|
}
|
|
|
|
/* Map SLI4 HBA Doorbell Register base to a kernel virtual address. */
|
|
phba->sli4_hba.drbl_regs_memmap_p =
|
|
ioremap(phba->pci_bar2_map, bar2map_len);
|
|
if (!phba->sli4_hba.drbl_regs_memmap_p) {
|
|
dev_printk(KERN_ERR, &pdev->dev,
|
|
"ioremap failed for SLI4 HBA doorbell registers.\n");
|
|
goto out_iounmap_ctrl;
|
|
}
|
|
|
|
/* Set up BAR0 PCI config space register memory map */
|
|
lpfc_sli4_bar0_register_memmap(phba);
|
|
|
|
/* Set up BAR1 register memory map */
|
|
lpfc_sli4_bar1_register_memmap(phba);
|
|
|
|
/* Set up BAR2 register memory map */
|
|
error = lpfc_sli4_bar2_register_memmap(phba, LPFC_VF0);
|
|
if (error)
|
|
goto out_iounmap_all;
|
|
|
|
return 0;
|
|
|
|
out_iounmap_all:
|
|
iounmap(phba->sli4_hba.drbl_regs_memmap_p);
|
|
out_iounmap_ctrl:
|
|
iounmap(phba->sli4_hba.ctrl_regs_memmap_p);
|
|
out_iounmap_conf:
|
|
iounmap(phba->sli4_hba.conf_regs_memmap_p);
|
|
out:
|
|
return error;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_pci_mem_unset - Unset SLI4 HBA PCI memory space.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to unset the PCI device memory space for device
|
|
* with SLI-4 interface spec.
|
|
**/
|
|
static void
|
|
lpfc_sli4_pci_mem_unset(struct lpfc_hba *phba)
|
|
{
|
|
struct pci_dev *pdev;
|
|
|
|
/* Obtain PCI device reference */
|
|
if (!phba->pcidev)
|
|
return;
|
|
else
|
|
pdev = phba->pcidev;
|
|
|
|
/* Free coherent DMA memory allocated */
|
|
|
|
/* Unmap I/O memory space */
|
|
iounmap(phba->sli4_hba.drbl_regs_memmap_p);
|
|
iounmap(phba->sli4_hba.ctrl_regs_memmap_p);
|
|
iounmap(phba->sli4_hba.conf_regs_memmap_p);
|
|
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli_enable_msix - Enable MSI-X interrupt mode on SLI-3 device
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to enable the MSI-X interrupt vectors to device
|
|
* with SLI-3 interface specs. The kernel function pci_enable_msix() is
|
|
* called to enable the MSI-X vectors. Note that pci_enable_msix(), once
|
|
* invoked, enables either all or nothing, depending on the current
|
|
* availability of PCI vector resources. The device driver is responsible
|
|
* for calling the individual request_irq() to register each MSI-X vector
|
|
* with a interrupt handler, which is done in this function. Note that
|
|
* later when device is unloading, the driver should always call free_irq()
|
|
* on all MSI-X vectors it has done request_irq() on before calling
|
|
* pci_disable_msix(). Failure to do so results in a BUG_ON() and a device
|
|
* will be left with MSI-X enabled and leaks its vectors.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* other values - error
|
|
**/
|
|
static int
|
|
lpfc_sli_enable_msix(struct lpfc_hba *phba)
|
|
{
|
|
int rc, i;
|
|
LPFC_MBOXQ_t *pmb;
|
|
|
|
/* Set up MSI-X multi-message vectors */
|
|
for (i = 0; i < LPFC_MSIX_VECTORS; i++)
|
|
phba->msix_entries[i].entry = i;
|
|
|
|
/* Configure MSI-X capability structure */
|
|
rc = pci_enable_msix(phba->pcidev, phba->msix_entries,
|
|
ARRAY_SIZE(phba->msix_entries));
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0420 PCI enable MSI-X failed (%d)\n", rc);
|
|
goto msi_fail_out;
|
|
}
|
|
for (i = 0; i < LPFC_MSIX_VECTORS; i++)
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0477 MSI-X entry[%d]: vector=x%x "
|
|
"message=%d\n", i,
|
|
phba->msix_entries[i].vector,
|
|
phba->msix_entries[i].entry);
|
|
/*
|
|
* Assign MSI-X vectors to interrupt handlers
|
|
*/
|
|
|
|
/* vector-0 is associated to slow-path handler */
|
|
rc = request_irq(phba->msix_entries[0].vector,
|
|
&lpfc_sli_sp_intr_handler, IRQF_SHARED,
|
|
LPFC_SP_DRIVER_HANDLER_NAME, phba);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_INIT,
|
|
"0421 MSI-X slow-path request_irq failed "
|
|
"(%d)\n", rc);
|
|
goto msi_fail_out;
|
|
}
|
|
|
|
/* vector-1 is associated to fast-path handler */
|
|
rc = request_irq(phba->msix_entries[1].vector,
|
|
&lpfc_sli_fp_intr_handler, IRQF_SHARED,
|
|
LPFC_FP_DRIVER_HANDLER_NAME, phba);
|
|
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_INIT,
|
|
"0429 MSI-X fast-path request_irq failed "
|
|
"(%d)\n", rc);
|
|
goto irq_fail_out;
|
|
}
|
|
|
|
/*
|
|
* Configure HBA MSI-X attention conditions to messages
|
|
*/
|
|
pmb = (LPFC_MBOXQ_t *) mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
|
|
|
|
if (!pmb) {
|
|
rc = -ENOMEM;
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0474 Unable to allocate memory for issuing "
|
|
"MBOX_CONFIG_MSI command\n");
|
|
goto mem_fail_out;
|
|
}
|
|
rc = lpfc_config_msi(phba, pmb);
|
|
if (rc)
|
|
goto mbx_fail_out;
|
|
rc = lpfc_sli_issue_mbox(phba, pmb, MBX_POLL);
|
|
if (rc != MBX_SUCCESS) {
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_MBOX,
|
|
"0351 Config MSI mailbox command failed, "
|
|
"mbxCmd x%x, mbxStatus x%x\n",
|
|
pmb->u.mb.mbxCommand, pmb->u.mb.mbxStatus);
|
|
goto mbx_fail_out;
|
|
}
|
|
|
|
/* Free memory allocated for mailbox command */
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
return rc;
|
|
|
|
mbx_fail_out:
|
|
/* Free memory allocated for mailbox command */
|
|
mempool_free(pmb, phba->mbox_mem_pool);
|
|
|
|
mem_fail_out:
|
|
/* free the irq already requested */
|
|
free_irq(phba->msix_entries[1].vector, phba);
|
|
|
|
irq_fail_out:
|
|
/* free the irq already requested */
|
|
free_irq(phba->msix_entries[0].vector, phba);
|
|
|
|
msi_fail_out:
|
|
/* Unconfigure MSI-X capability structure */
|
|
pci_disable_msix(phba->pcidev);
|
|
return rc;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli_disable_msix - Disable MSI-X interrupt mode on SLI-3 device.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to release the MSI-X vectors and then disable the
|
|
* MSI-X interrupt mode to device with SLI-3 interface spec.
|
|
**/
|
|
static void
|
|
lpfc_sli_disable_msix(struct lpfc_hba *phba)
|
|
{
|
|
int i;
|
|
|
|
/* Free up MSI-X multi-message vectors */
|
|
for (i = 0; i < LPFC_MSIX_VECTORS; i++)
|
|
free_irq(phba->msix_entries[i].vector, phba);
|
|
/* Disable MSI-X */
|
|
pci_disable_msix(phba->pcidev);
|
|
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli_enable_msi - Enable MSI interrupt mode on SLI-3 device.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to enable the MSI interrupt mode to device with
|
|
* SLI-3 interface spec. The kernel function pci_enable_msi() is called to
|
|
* enable the MSI vector. The device driver is responsible for calling the
|
|
* request_irq() to register MSI vector with a interrupt the handler, which
|
|
* is done in this function.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* other values - error
|
|
*/
|
|
static int
|
|
lpfc_sli_enable_msi(struct lpfc_hba *phba)
|
|
{
|
|
int rc;
|
|
|
|
rc = pci_enable_msi(phba->pcidev);
|
|
if (!rc)
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0462 PCI enable MSI mode success.\n");
|
|
else {
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0471 PCI enable MSI mode failed (%d)\n", rc);
|
|
return rc;
|
|
}
|
|
|
|
rc = request_irq(phba->pcidev->irq, lpfc_sli_intr_handler,
|
|
IRQF_SHARED, LPFC_DRIVER_NAME, phba);
|
|
if (rc) {
|
|
pci_disable_msi(phba->pcidev);
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_INIT,
|
|
"0478 MSI request_irq failed (%d)\n", rc);
|
|
}
|
|
return rc;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli_disable_msi - Disable MSI interrupt mode to SLI-3 device.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to disable the MSI interrupt mode to device with
|
|
* SLI-3 interface spec. The driver calls free_irq() on MSI vector it has
|
|
* done request_irq() on before calling pci_disable_msi(). Failure to do so
|
|
* results in a BUG_ON() and a device will be left with MSI enabled and leaks
|
|
* its vector.
|
|
*/
|
|
static void
|
|
lpfc_sli_disable_msi(struct lpfc_hba *phba)
|
|
{
|
|
free_irq(phba->pcidev->irq, phba);
|
|
pci_disable_msi(phba->pcidev);
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli_enable_intr - Enable device interrupt to SLI-3 device.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to enable device interrupt and associate driver's
|
|
* interrupt handler(s) to interrupt vector(s) to device with SLI-3 interface
|
|
* spec. Depends on the interrupt mode configured to the driver, the driver
|
|
* will try to fallback from the configured interrupt mode to an interrupt
|
|
* mode which is supported by the platform, kernel, and device in the order
|
|
* of:
|
|
* MSI-X -> MSI -> IRQ.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* other values - error
|
|
**/
|
|
static uint32_t
|
|
lpfc_sli_enable_intr(struct lpfc_hba *phba, uint32_t cfg_mode)
|
|
{
|
|
uint32_t intr_mode = LPFC_INTR_ERROR;
|
|
int retval;
|
|
|
|
if (cfg_mode == 2) {
|
|
/* Need to issue conf_port mbox cmd before conf_msi mbox cmd */
|
|
retval = lpfc_sli_config_port(phba, LPFC_SLI_REV3);
|
|
if (!retval) {
|
|
/* Now, try to enable MSI-X interrupt mode */
|
|
retval = lpfc_sli_enable_msix(phba);
|
|
if (!retval) {
|
|
/* Indicate initialization to MSI-X mode */
|
|
phba->intr_type = MSIX;
|
|
intr_mode = 2;
|
|
}
|
|
}
|
|
}
|
|
|
|
/* Fallback to MSI if MSI-X initialization failed */
|
|
if (cfg_mode >= 1 && phba->intr_type == NONE) {
|
|
retval = lpfc_sli_enable_msi(phba);
|
|
if (!retval) {
|
|
/* Indicate initialization to MSI mode */
|
|
phba->intr_type = MSI;
|
|
intr_mode = 1;
|
|
}
|
|
}
|
|
|
|
/* Fallback to INTx if both MSI-X/MSI initalization failed */
|
|
if (phba->intr_type == NONE) {
|
|
retval = request_irq(phba->pcidev->irq, lpfc_sli_intr_handler,
|
|
IRQF_SHARED, LPFC_DRIVER_NAME, phba);
|
|
if (!retval) {
|
|
/* Indicate initialization to INTx mode */
|
|
phba->intr_type = INTx;
|
|
intr_mode = 0;
|
|
}
|
|
}
|
|
return intr_mode;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli_disable_intr - Disable device interrupt to SLI-3 device.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to disable device interrupt and disassociate the
|
|
* driver's interrupt handler(s) from interrupt vector(s) to device with
|
|
* SLI-3 interface spec. Depending on the interrupt mode, the driver will
|
|
* release the interrupt vector(s) for the message signaled interrupt.
|
|
**/
|
|
static void
|
|
lpfc_sli_disable_intr(struct lpfc_hba *phba)
|
|
{
|
|
/* Disable the currently initialized interrupt mode */
|
|
if (phba->intr_type == MSIX)
|
|
lpfc_sli_disable_msix(phba);
|
|
else if (phba->intr_type == MSI)
|
|
lpfc_sli_disable_msi(phba);
|
|
else if (phba->intr_type == INTx)
|
|
free_irq(phba->pcidev->irq, phba);
|
|
|
|
/* Reset interrupt management states */
|
|
phba->intr_type = NONE;
|
|
phba->sli.slistat.sli_intr = 0;
|
|
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_enable_msix - Enable MSI-X interrupt mode to SLI-4 device
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to enable the MSI-X interrupt vectors to device
|
|
* with SLI-4 interface spec. The kernel function pci_enable_msix() is called
|
|
* to enable the MSI-X vectors. Note that pci_enable_msix(), once invoked,
|
|
* enables either all or nothing, depending on the current availability of
|
|
* PCI vector resources. The device driver is responsible for calling the
|
|
* individual request_irq() to register each MSI-X vector with a interrupt
|
|
* handler, which is done in this function. Note that later when device is
|
|
* unloading, the driver should always call free_irq() on all MSI-X vectors
|
|
* it has done request_irq() on before calling pci_disable_msix(). Failure
|
|
* to do so results in a BUG_ON() and a device will be left with MSI-X
|
|
* enabled and leaks its vectors.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* other values - error
|
|
**/
|
|
static int
|
|
lpfc_sli4_enable_msix(struct lpfc_hba *phba)
|
|
{
|
|
int vectors, rc, index;
|
|
|
|
/* Set up MSI-X multi-message vectors */
|
|
for (index = 0; index < phba->sli4_hba.cfg_eqn; index++)
|
|
phba->sli4_hba.msix_entries[index].entry = index;
|
|
|
|
/* Configure MSI-X capability structure */
|
|
vectors = phba->sli4_hba.cfg_eqn;
|
|
enable_msix_vectors:
|
|
rc = pci_enable_msix(phba->pcidev, phba->sli4_hba.msix_entries,
|
|
vectors);
|
|
if (rc > 1) {
|
|
vectors = rc;
|
|
goto enable_msix_vectors;
|
|
} else if (rc) {
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0484 PCI enable MSI-X failed (%d)\n", rc);
|
|
goto msi_fail_out;
|
|
}
|
|
|
|
/* Log MSI-X vector assignment */
|
|
for (index = 0; index < vectors; index++)
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0489 MSI-X entry[%d]: vector=x%x "
|
|
"message=%d\n", index,
|
|
phba->sli4_hba.msix_entries[index].vector,
|
|
phba->sli4_hba.msix_entries[index].entry);
|
|
/*
|
|
* Assign MSI-X vectors to interrupt handlers
|
|
*/
|
|
|
|
/* The first vector must associated to slow-path handler for MQ */
|
|
rc = request_irq(phba->sli4_hba.msix_entries[0].vector,
|
|
&lpfc_sli4_sp_intr_handler, IRQF_SHARED,
|
|
LPFC_SP_DRIVER_HANDLER_NAME, phba);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_INIT,
|
|
"0485 MSI-X slow-path request_irq failed "
|
|
"(%d)\n", rc);
|
|
goto msi_fail_out;
|
|
}
|
|
|
|
/* The rest of the vector(s) are associated to fast-path handler(s) */
|
|
for (index = 1; index < vectors; index++) {
|
|
phba->sli4_hba.fcp_eq_hdl[index - 1].idx = index - 1;
|
|
phba->sli4_hba.fcp_eq_hdl[index - 1].phba = phba;
|
|
rc = request_irq(phba->sli4_hba.msix_entries[index].vector,
|
|
&lpfc_sli4_fp_intr_handler, IRQF_SHARED,
|
|
LPFC_FP_DRIVER_HANDLER_NAME,
|
|
&phba->sli4_hba.fcp_eq_hdl[index - 1]);
|
|
if (rc) {
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_INIT,
|
|
"0486 MSI-X fast-path (%d) "
|
|
"request_irq failed (%d)\n", index, rc);
|
|
goto cfg_fail_out;
|
|
}
|
|
}
|
|
phba->sli4_hba.msix_vec_nr = vectors;
|
|
|
|
return rc;
|
|
|
|
cfg_fail_out:
|
|
/* free the irq already requested */
|
|
for (--index; index >= 1; index--)
|
|
free_irq(phba->sli4_hba.msix_entries[index - 1].vector,
|
|
&phba->sli4_hba.fcp_eq_hdl[index - 1]);
|
|
|
|
/* free the irq already requested */
|
|
free_irq(phba->sli4_hba.msix_entries[0].vector, phba);
|
|
|
|
msi_fail_out:
|
|
/* Unconfigure MSI-X capability structure */
|
|
pci_disable_msix(phba->pcidev);
|
|
return rc;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_disable_msix - Disable MSI-X interrupt mode to SLI-4 device
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to release the MSI-X vectors and then disable the
|
|
* MSI-X interrupt mode to device with SLI-4 interface spec.
|
|
**/
|
|
static void
|
|
lpfc_sli4_disable_msix(struct lpfc_hba *phba)
|
|
{
|
|
int index;
|
|
|
|
/* Free up MSI-X multi-message vectors */
|
|
free_irq(phba->sli4_hba.msix_entries[0].vector, phba);
|
|
|
|
for (index = 1; index < phba->sli4_hba.msix_vec_nr; index++)
|
|
free_irq(phba->sli4_hba.msix_entries[index].vector,
|
|
&phba->sli4_hba.fcp_eq_hdl[index - 1]);
|
|
|
|
/* Disable MSI-X */
|
|
pci_disable_msix(phba->pcidev);
|
|
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_enable_msi - Enable MSI interrupt mode to SLI-4 device
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to enable the MSI interrupt mode to device with
|
|
* SLI-4 interface spec. The kernel function pci_enable_msi() is called
|
|
* to enable the MSI vector. The device driver is responsible for calling
|
|
* the request_irq() to register MSI vector with a interrupt the handler,
|
|
* which is done in this function.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* other values - error
|
|
**/
|
|
static int
|
|
lpfc_sli4_enable_msi(struct lpfc_hba *phba)
|
|
{
|
|
int rc, index;
|
|
|
|
rc = pci_enable_msi(phba->pcidev);
|
|
if (!rc)
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0487 PCI enable MSI mode success.\n");
|
|
else {
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0488 PCI enable MSI mode failed (%d)\n", rc);
|
|
return rc;
|
|
}
|
|
|
|
rc = request_irq(phba->pcidev->irq, lpfc_sli4_intr_handler,
|
|
IRQF_SHARED, LPFC_DRIVER_NAME, phba);
|
|
if (rc) {
|
|
pci_disable_msi(phba->pcidev);
|
|
lpfc_printf_log(phba, KERN_WARNING, LOG_INIT,
|
|
"0490 MSI request_irq failed (%d)\n", rc);
|
|
return rc;
|
|
}
|
|
|
|
for (index = 0; index < phba->cfg_fcp_eq_count; index++) {
|
|
phba->sli4_hba.fcp_eq_hdl[index].idx = index;
|
|
phba->sli4_hba.fcp_eq_hdl[index].phba = phba;
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_disable_msi - Disable MSI interrupt mode to SLI-4 device
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to disable the MSI interrupt mode to device with
|
|
* SLI-4 interface spec. The driver calls free_irq() on MSI vector it has
|
|
* done request_irq() on before calling pci_disable_msi(). Failure to do so
|
|
* results in a BUG_ON() and a device will be left with MSI enabled and leaks
|
|
* its vector.
|
|
**/
|
|
static void
|
|
lpfc_sli4_disable_msi(struct lpfc_hba *phba)
|
|
{
|
|
free_irq(phba->pcidev->irq, phba);
|
|
pci_disable_msi(phba->pcidev);
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_enable_intr - Enable device interrupt to SLI-4 device
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to enable device interrupt and associate driver's
|
|
* interrupt handler(s) to interrupt vector(s) to device with SLI-4
|
|
* interface spec. Depends on the interrupt mode configured to the driver,
|
|
* the driver will try to fallback from the configured interrupt mode to an
|
|
* interrupt mode which is supported by the platform, kernel, and device in
|
|
* the order of:
|
|
* MSI-X -> MSI -> IRQ.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* other values - error
|
|
**/
|
|
static uint32_t
|
|
lpfc_sli4_enable_intr(struct lpfc_hba *phba, uint32_t cfg_mode)
|
|
{
|
|
uint32_t intr_mode = LPFC_INTR_ERROR;
|
|
int retval, index;
|
|
|
|
if (cfg_mode == 2) {
|
|
/* Preparation before conf_msi mbox cmd */
|
|
retval = 0;
|
|
if (!retval) {
|
|
/* Now, try to enable MSI-X interrupt mode */
|
|
retval = lpfc_sli4_enable_msix(phba);
|
|
if (!retval) {
|
|
/* Indicate initialization to MSI-X mode */
|
|
phba->intr_type = MSIX;
|
|
intr_mode = 2;
|
|
}
|
|
}
|
|
}
|
|
|
|
/* Fallback to MSI if MSI-X initialization failed */
|
|
if (cfg_mode >= 1 && phba->intr_type == NONE) {
|
|
retval = lpfc_sli4_enable_msi(phba);
|
|
if (!retval) {
|
|
/* Indicate initialization to MSI mode */
|
|
phba->intr_type = MSI;
|
|
intr_mode = 1;
|
|
}
|
|
}
|
|
|
|
/* Fallback to INTx if both MSI-X/MSI initalization failed */
|
|
if (phba->intr_type == NONE) {
|
|
retval = request_irq(phba->pcidev->irq, lpfc_sli4_intr_handler,
|
|
IRQF_SHARED, LPFC_DRIVER_NAME, phba);
|
|
if (!retval) {
|
|
/* Indicate initialization to INTx mode */
|
|
phba->intr_type = INTx;
|
|
intr_mode = 0;
|
|
for (index = 0; index < phba->cfg_fcp_eq_count;
|
|
index++) {
|
|
phba->sli4_hba.fcp_eq_hdl[index].idx = index;
|
|
phba->sli4_hba.fcp_eq_hdl[index].phba = phba;
|
|
}
|
|
}
|
|
}
|
|
return intr_mode;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_disable_intr - Disable device interrupt to SLI-4 device
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to disable device interrupt and disassociate
|
|
* the driver's interrupt handler(s) from interrupt vector(s) to device
|
|
* with SLI-4 interface spec. Depending on the interrupt mode, the driver
|
|
* will release the interrupt vector(s) for the message signaled interrupt.
|
|
**/
|
|
static void
|
|
lpfc_sli4_disable_intr(struct lpfc_hba *phba)
|
|
{
|
|
/* Disable the currently initialized interrupt mode */
|
|
if (phba->intr_type == MSIX)
|
|
lpfc_sli4_disable_msix(phba);
|
|
else if (phba->intr_type == MSI)
|
|
lpfc_sli4_disable_msi(phba);
|
|
else if (phba->intr_type == INTx)
|
|
free_irq(phba->pcidev->irq, phba);
|
|
|
|
/* Reset interrupt management states */
|
|
phba->intr_type = NONE;
|
|
phba->sli.slistat.sli_intr = 0;
|
|
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_unset_hba - Unset SLI3 hba device initialization
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to unset the HBA device initialization steps to
|
|
* a device with SLI-3 interface spec.
|
|
**/
|
|
static void
|
|
lpfc_unset_hba(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_vport *vport = phba->pport;
|
|
struct Scsi_Host *shost = lpfc_shost_from_vport(vport);
|
|
|
|
spin_lock_irq(shost->host_lock);
|
|
vport->load_flag |= FC_UNLOADING;
|
|
spin_unlock_irq(shost->host_lock);
|
|
|
|
lpfc_stop_hba_timers(phba);
|
|
|
|
phba->pport->work_port_events = 0;
|
|
|
|
lpfc_sli_hba_down(phba);
|
|
|
|
lpfc_sli_brdrestart(phba);
|
|
|
|
lpfc_sli_disable_intr(phba);
|
|
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_unset_hba - Unset SLI4 hba device initialization.
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is invoked to unset the HBA device initialization steps to
|
|
* a device with SLI-4 interface spec.
|
|
**/
|
|
static void
|
|
lpfc_sli4_unset_hba(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_vport *vport = phba->pport;
|
|
struct Scsi_Host *shost = lpfc_shost_from_vport(vport);
|
|
|
|
spin_lock_irq(shost->host_lock);
|
|
vport->load_flag |= FC_UNLOADING;
|
|
spin_unlock_irq(shost->host_lock);
|
|
|
|
phba->pport->work_port_events = 0;
|
|
|
|
/* Stop the SLI4 device port */
|
|
lpfc_stop_port(phba);
|
|
|
|
lpfc_sli4_disable_intr(phba);
|
|
|
|
/* Reset SLI4 HBA FCoE function */
|
|
lpfc_pci_function_reset(phba);
|
|
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_xri_exchange_busy_wait - Wait for device XRI exchange busy
|
|
* @phba: Pointer to HBA context object.
|
|
*
|
|
* This function is called in the SLI4 code path to wait for completion
|
|
* of device's XRIs exchange busy. It will check the XRI exchange busy
|
|
* on outstanding FCP and ELS I/Os every 10ms for up to 10 seconds; after
|
|
* that, it will check the XRI exchange busy on outstanding FCP and ELS
|
|
* I/Os every 30 seconds, log error message, and wait forever. Only when
|
|
* all XRI exchange busy complete, the driver unload shall proceed with
|
|
* invoking the function reset ioctl mailbox command to the CNA and the
|
|
* the rest of the driver unload resource release.
|
|
**/
|
|
static void
|
|
lpfc_sli4_xri_exchange_busy_wait(struct lpfc_hba *phba)
|
|
{
|
|
int wait_time = 0;
|
|
int fcp_xri_cmpl = list_empty(&phba->sli4_hba.lpfc_abts_scsi_buf_list);
|
|
int els_xri_cmpl = list_empty(&phba->sli4_hba.lpfc_abts_els_sgl_list);
|
|
|
|
while (!fcp_xri_cmpl || !els_xri_cmpl) {
|
|
if (wait_time > LPFC_XRI_EXCH_BUSY_WAIT_TMO) {
|
|
if (!fcp_xri_cmpl)
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2877 FCP XRI exchange busy "
|
|
"wait time: %d seconds.\n",
|
|
wait_time/1000);
|
|
if (!els_xri_cmpl)
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2878 ELS XRI exchange busy "
|
|
"wait time: %d seconds.\n",
|
|
wait_time/1000);
|
|
msleep(LPFC_XRI_EXCH_BUSY_WAIT_T2);
|
|
wait_time += LPFC_XRI_EXCH_BUSY_WAIT_T2;
|
|
} else {
|
|
msleep(LPFC_XRI_EXCH_BUSY_WAIT_T1);
|
|
wait_time += LPFC_XRI_EXCH_BUSY_WAIT_T1;
|
|
}
|
|
fcp_xri_cmpl =
|
|
list_empty(&phba->sli4_hba.lpfc_abts_scsi_buf_list);
|
|
els_xri_cmpl =
|
|
list_empty(&phba->sli4_hba.lpfc_abts_els_sgl_list);
|
|
}
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_hba_unset - Unset the fcoe hba
|
|
* @phba: Pointer to HBA context object.
|
|
*
|
|
* This function is called in the SLI4 code path to reset the HBA's FCoE
|
|
* function. The caller is not required to hold any lock. This routine
|
|
* issues PCI function reset mailbox command to reset the FCoE function.
|
|
* At the end of the function, it calls lpfc_hba_down_post function to
|
|
* free any pending commands.
|
|
**/
|
|
static void
|
|
lpfc_sli4_hba_unset(struct lpfc_hba *phba)
|
|
{
|
|
int wait_cnt = 0;
|
|
LPFC_MBOXQ_t *mboxq;
|
|
|
|
lpfc_stop_hba_timers(phba);
|
|
phba->sli4_hba.intr_enable = 0;
|
|
|
|
/*
|
|
* Gracefully wait out the potential current outstanding asynchronous
|
|
* mailbox command.
|
|
*/
|
|
|
|
/* First, block any pending async mailbox command from posted */
|
|
spin_lock_irq(&phba->hbalock);
|
|
phba->sli.sli_flag |= LPFC_SLI_ASYNC_MBX_BLK;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
/* Now, trying to wait it out if we can */
|
|
while (phba->sli.sli_flag & LPFC_SLI_MBOX_ACTIVE) {
|
|
msleep(10);
|
|
if (++wait_cnt > LPFC_ACTIVE_MBOX_WAIT_CNT)
|
|
break;
|
|
}
|
|
/* Forcefully release the outstanding mailbox command if timed out */
|
|
if (phba->sli.sli_flag & LPFC_SLI_MBOX_ACTIVE) {
|
|
spin_lock_irq(&phba->hbalock);
|
|
mboxq = phba->sli.mbox_active;
|
|
mboxq->u.mb.mbxStatus = MBX_NOT_FINISHED;
|
|
__lpfc_mbox_cmpl_put(phba, mboxq);
|
|
phba->sli.sli_flag &= ~LPFC_SLI_MBOX_ACTIVE;
|
|
phba->sli.mbox_active = NULL;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
}
|
|
|
|
/* Abort all iocbs associated with the hba */
|
|
lpfc_sli_hba_iocb_abort(phba);
|
|
|
|
/* Wait for completion of device XRI exchange busy */
|
|
lpfc_sli4_xri_exchange_busy_wait(phba);
|
|
|
|
/* Disable PCI subsystem interrupt */
|
|
lpfc_sli4_disable_intr(phba);
|
|
|
|
/* Stop kthread signal shall trigger work_done one more time */
|
|
kthread_stop(phba->worker_thread);
|
|
|
|
/* Reset SLI4 HBA FCoE function */
|
|
lpfc_pci_function_reset(phba);
|
|
|
|
/* Stop the SLI4 device port */
|
|
phba->pport->work_port_events = 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_pc_sli4_params_get - Get the SLI4_PARAMS port capabilities.
|
|
* @phba: Pointer to HBA context object.
|
|
* @mboxq: Pointer to the mailboxq memory for the mailbox command response.
|
|
*
|
|
* This function is called in the SLI4 code path to read the port's
|
|
* sli4 capabilities.
|
|
*
|
|
* This function may be be called from any context that can block-wait
|
|
* for the completion. The expectation is that this routine is called
|
|
* typically from probe_one or from the online routine.
|
|
**/
|
|
int
|
|
lpfc_pc_sli4_params_get(struct lpfc_hba *phba, LPFC_MBOXQ_t *mboxq)
|
|
{
|
|
int rc;
|
|
struct lpfc_mqe *mqe;
|
|
struct lpfc_pc_sli4_params *sli4_params;
|
|
uint32_t mbox_tmo;
|
|
|
|
rc = 0;
|
|
mqe = &mboxq->u.mqe;
|
|
|
|
/* Read the port's SLI4 Parameters port capabilities */
|
|
lpfc_sli4_params(mboxq);
|
|
if (!phba->sli4_hba.intr_enable)
|
|
rc = lpfc_sli_issue_mbox(phba, mboxq, MBX_POLL);
|
|
else {
|
|
mbox_tmo = lpfc_mbox_tmo_val(phba, MBX_PORT_CAPABILITIES);
|
|
rc = lpfc_sli_issue_mbox_wait(phba, mboxq, mbox_tmo);
|
|
}
|
|
|
|
if (unlikely(rc))
|
|
return 1;
|
|
|
|
sli4_params = &phba->sli4_hba.pc_sli4_params;
|
|
sli4_params->if_type = bf_get(if_type, &mqe->un.sli4_params);
|
|
sli4_params->sli_rev = bf_get(sli_rev, &mqe->un.sli4_params);
|
|
sli4_params->sli_family = bf_get(sli_family, &mqe->un.sli4_params);
|
|
sli4_params->featurelevel_1 = bf_get(featurelevel_1,
|
|
&mqe->un.sli4_params);
|
|
sli4_params->featurelevel_2 = bf_get(featurelevel_2,
|
|
&mqe->un.sli4_params);
|
|
sli4_params->proto_types = mqe->un.sli4_params.word3;
|
|
sli4_params->sge_supp_len = mqe->un.sli4_params.sge_supp_len;
|
|
sli4_params->if_page_sz = bf_get(if_page_sz, &mqe->un.sli4_params);
|
|
sli4_params->rq_db_window = bf_get(rq_db_window, &mqe->un.sli4_params);
|
|
sli4_params->loopbk_scope = bf_get(loopbk_scope, &mqe->un.sli4_params);
|
|
sli4_params->eq_pages_max = bf_get(eq_pages, &mqe->un.sli4_params);
|
|
sli4_params->eqe_size = bf_get(eqe_size, &mqe->un.sli4_params);
|
|
sli4_params->cq_pages_max = bf_get(cq_pages, &mqe->un.sli4_params);
|
|
sli4_params->cqe_size = bf_get(cqe_size, &mqe->un.sli4_params);
|
|
sli4_params->mq_pages_max = bf_get(mq_pages, &mqe->un.sli4_params);
|
|
sli4_params->mqe_size = bf_get(mqe_size, &mqe->un.sli4_params);
|
|
sli4_params->mq_elem_cnt = bf_get(mq_elem_cnt, &mqe->un.sli4_params);
|
|
sli4_params->wq_pages_max = bf_get(wq_pages, &mqe->un.sli4_params);
|
|
sli4_params->wqe_size = bf_get(wqe_size, &mqe->un.sli4_params);
|
|
sli4_params->rq_pages_max = bf_get(rq_pages, &mqe->un.sli4_params);
|
|
sli4_params->rqe_size = bf_get(rqe_size, &mqe->un.sli4_params);
|
|
sli4_params->hdr_pages_max = bf_get(hdr_pages, &mqe->un.sli4_params);
|
|
sli4_params->hdr_size = bf_get(hdr_size, &mqe->un.sli4_params);
|
|
sli4_params->hdr_pp_align = bf_get(hdr_pp_align, &mqe->un.sli4_params);
|
|
sli4_params->sgl_pages_max = bf_get(sgl_pages, &mqe->un.sli4_params);
|
|
sli4_params->sgl_pp_align = bf_get(sgl_pp_align, &mqe->un.sli4_params);
|
|
return rc;
|
|
}
|
|
|
|
/**
|
|
* lpfc_pci_probe_one_s3 - PCI probe func to reg SLI-3 device to PCI subsystem.
|
|
* @pdev: pointer to PCI device
|
|
* @pid: pointer to PCI device identifier
|
|
*
|
|
* This routine is to be called to attach a device with SLI-3 interface spec
|
|
* to the PCI subsystem. When an Emulex HBA with SLI-3 interface spec is
|
|
* presented on PCI bus, the kernel PCI subsystem looks at PCI device-specific
|
|
* information of the device and driver to see if the driver state that it can
|
|
* support this kind of device. If the match is successful, the driver core
|
|
* invokes this routine. If this routine determines it can claim the HBA, it
|
|
* does all the initialization that it needs to do to handle the HBA properly.
|
|
*
|
|
* Return code
|
|
* 0 - driver can claim the device
|
|
* negative value - driver can not claim the device
|
|
**/
|
|
static int __devinit
|
|
lpfc_pci_probe_one_s3(struct pci_dev *pdev, const struct pci_device_id *pid)
|
|
{
|
|
struct lpfc_hba *phba;
|
|
struct lpfc_vport *vport = NULL;
|
|
struct Scsi_Host *shost = NULL;
|
|
int error;
|
|
uint32_t cfg_mode, intr_mode;
|
|
|
|
/* Allocate memory for HBA structure */
|
|
phba = lpfc_hba_alloc(pdev);
|
|
if (!phba)
|
|
return -ENOMEM;
|
|
|
|
/* Perform generic PCI device enabling operation */
|
|
error = lpfc_enable_pci_dev(phba);
|
|
if (error) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1401 Failed to enable pci device.\n");
|
|
goto out_free_phba;
|
|
}
|
|
|
|
/* Set up SLI API function jump table for PCI-device group-0 HBAs */
|
|
error = lpfc_api_table_setup(phba, LPFC_PCI_DEV_LP);
|
|
if (error)
|
|
goto out_disable_pci_dev;
|
|
|
|
/* Set up SLI-3 specific device PCI memory space */
|
|
error = lpfc_sli_pci_mem_setup(phba);
|
|
if (error) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1402 Failed to set up pci memory space.\n");
|
|
goto out_disable_pci_dev;
|
|
}
|
|
|
|
/* Set up phase-1 common device driver resources */
|
|
error = lpfc_setup_driver_resource_phase1(phba);
|
|
if (error) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1403 Failed to set up driver resource.\n");
|
|
goto out_unset_pci_mem_s3;
|
|
}
|
|
|
|
/* Set up SLI-3 specific device driver resources */
|
|
error = lpfc_sli_driver_resource_setup(phba);
|
|
if (error) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1404 Failed to set up driver resource.\n");
|
|
goto out_unset_pci_mem_s3;
|
|
}
|
|
|
|
/* Initialize and populate the iocb list per host */
|
|
error = lpfc_init_iocb_list(phba, LPFC_IOCB_LIST_CNT);
|
|
if (error) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1405 Failed to initialize iocb list.\n");
|
|
goto out_unset_driver_resource_s3;
|
|
}
|
|
|
|
/* Set up common device driver resources */
|
|
error = lpfc_setup_driver_resource_phase2(phba);
|
|
if (error) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1406 Failed to set up driver resource.\n");
|
|
goto out_free_iocb_list;
|
|
}
|
|
|
|
/* Create SCSI host to the physical port */
|
|
error = lpfc_create_shost(phba);
|
|
if (error) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1407 Failed to create scsi host.\n");
|
|
goto out_unset_driver_resource;
|
|
}
|
|
|
|
/* Configure sysfs attributes */
|
|
vport = phba->pport;
|
|
error = lpfc_alloc_sysfs_attr(vport);
|
|
if (error) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1476 Failed to allocate sysfs attr\n");
|
|
goto out_destroy_shost;
|
|
}
|
|
|
|
shost = lpfc_shost_from_vport(vport); /* save shost for error cleanup */
|
|
/* Now, trying to enable interrupt and bring up the device */
|
|
cfg_mode = phba->cfg_use_msi;
|
|
while (true) {
|
|
/* Put device to a known state before enabling interrupt */
|
|
lpfc_stop_port(phba);
|
|
/* Configure and enable interrupt */
|
|
intr_mode = lpfc_sli_enable_intr(phba, cfg_mode);
|
|
if (intr_mode == LPFC_INTR_ERROR) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0431 Failed to enable interrupt.\n");
|
|
error = -ENODEV;
|
|
goto out_free_sysfs_attr;
|
|
}
|
|
/* SLI-3 HBA setup */
|
|
if (lpfc_sli_hba_setup(phba)) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1477 Failed to set up hba\n");
|
|
error = -ENODEV;
|
|
goto out_remove_device;
|
|
}
|
|
|
|
/* Wait 50ms for the interrupts of previous mailbox commands */
|
|
msleep(50);
|
|
/* Check active interrupts on message signaled interrupts */
|
|
if (intr_mode == 0 ||
|
|
phba->sli.slistat.sli_intr > LPFC_MSIX_VECTORS) {
|
|
/* Log the current active interrupt mode */
|
|
phba->intr_mode = intr_mode;
|
|
lpfc_log_intr_mode(phba, intr_mode);
|
|
break;
|
|
} else {
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0447 Configure interrupt mode (%d) "
|
|
"failed active interrupt test.\n",
|
|
intr_mode);
|
|
/* Disable the current interrupt mode */
|
|
lpfc_sli_disable_intr(phba);
|
|
/* Try next level of interrupt mode */
|
|
cfg_mode = --intr_mode;
|
|
}
|
|
}
|
|
|
|
/* Perform post initialization setup */
|
|
lpfc_post_init_setup(phba);
|
|
|
|
/* Check if there are static vports to be created. */
|
|
lpfc_create_static_vport(phba);
|
|
|
|
return 0;
|
|
|
|
out_remove_device:
|
|
lpfc_unset_hba(phba);
|
|
out_free_sysfs_attr:
|
|
lpfc_free_sysfs_attr(vport);
|
|
out_destroy_shost:
|
|
lpfc_destroy_shost(phba);
|
|
out_unset_driver_resource:
|
|
lpfc_unset_driver_resource_phase2(phba);
|
|
out_free_iocb_list:
|
|
lpfc_free_iocb_list(phba);
|
|
out_unset_driver_resource_s3:
|
|
lpfc_sli_driver_resource_unset(phba);
|
|
out_unset_pci_mem_s3:
|
|
lpfc_sli_pci_mem_unset(phba);
|
|
out_disable_pci_dev:
|
|
lpfc_disable_pci_dev(phba);
|
|
if (shost)
|
|
scsi_host_put(shost);
|
|
out_free_phba:
|
|
lpfc_hba_free(phba);
|
|
return error;
|
|
}
|
|
|
|
/**
|
|
* lpfc_pci_remove_one_s3 - PCI func to unreg SLI-3 device from PCI subsystem.
|
|
* @pdev: pointer to PCI device
|
|
*
|
|
* This routine is to be called to disattach a device with SLI-3 interface
|
|
* spec from PCI subsystem. When an Emulex HBA with SLI-3 interface spec is
|
|
* removed from PCI bus, it performs all the necessary cleanup for the HBA
|
|
* device to be removed from the PCI subsystem properly.
|
|
**/
|
|
static void __devexit
|
|
lpfc_pci_remove_one_s3(struct pci_dev *pdev)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
|
|
struct lpfc_vport **vports;
|
|
struct lpfc_hba *phba = vport->phba;
|
|
int i;
|
|
int bars = pci_select_bars(pdev, IORESOURCE_MEM);
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
vport->load_flag |= FC_UNLOADING;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
lpfc_free_sysfs_attr(vport);
|
|
|
|
/* Release all the vports against this physical port */
|
|
vports = lpfc_create_vport_work_array(phba);
|
|
if (vports != NULL)
|
|
for (i = 1; i <= phba->max_vports && vports[i] != NULL; i++)
|
|
fc_vport_terminate(vports[i]->fc_vport);
|
|
lpfc_destroy_vport_work_array(phba, vports);
|
|
|
|
/* Remove FC host and then SCSI host with the physical port */
|
|
fc_remove_host(shost);
|
|
scsi_remove_host(shost);
|
|
lpfc_cleanup(vport);
|
|
|
|
/*
|
|
* Bring down the SLI Layer. This step disable all interrupts,
|
|
* clears the rings, discards all mailbox commands, and resets
|
|
* the HBA.
|
|
*/
|
|
|
|
/* HBA interrupt will be diabled after this call */
|
|
lpfc_sli_hba_down(phba);
|
|
/* Stop kthread signal shall trigger work_done one more time */
|
|
kthread_stop(phba->worker_thread);
|
|
/* Final cleanup of txcmplq and reset the HBA */
|
|
lpfc_sli_brdrestart(phba);
|
|
|
|
lpfc_stop_hba_timers(phba);
|
|
spin_lock_irq(&phba->hbalock);
|
|
list_del_init(&vport->listentry);
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
lpfc_debugfs_terminate(vport);
|
|
|
|
/* Disable interrupt */
|
|
lpfc_sli_disable_intr(phba);
|
|
|
|
pci_set_drvdata(pdev, NULL);
|
|
scsi_host_put(shost);
|
|
|
|
/*
|
|
* Call scsi_free before mem_free since scsi bufs are released to their
|
|
* corresponding pools here.
|
|
*/
|
|
lpfc_scsi_free(phba);
|
|
lpfc_mem_free_all(phba);
|
|
|
|
dma_free_coherent(&pdev->dev, lpfc_sli_hbq_size(),
|
|
phba->hbqslimp.virt, phba->hbqslimp.phys);
|
|
|
|
/* Free resources associated with SLI2 interface */
|
|
dma_free_coherent(&pdev->dev, SLI2_SLIM_SIZE,
|
|
phba->slim2p.virt, phba->slim2p.phys);
|
|
|
|
/* unmap adapter SLIM and Control Registers */
|
|
iounmap(phba->ctrl_regs_memmap_p);
|
|
iounmap(phba->slim_memmap_p);
|
|
|
|
lpfc_hba_free(phba);
|
|
|
|
pci_release_selected_regions(pdev, bars);
|
|
pci_disable_device(pdev);
|
|
}
|
|
|
|
/**
|
|
* lpfc_pci_suspend_one_s3 - PCI func to suspend SLI-3 device for power mgmnt
|
|
* @pdev: pointer to PCI device
|
|
* @msg: power management message
|
|
*
|
|
* This routine is to be called from the kernel's PCI subsystem to support
|
|
* system Power Management (PM) to device with SLI-3 interface spec. When
|
|
* PM invokes this method, it quiesces the device by stopping the driver's
|
|
* worker thread for the device, turning off device's interrupt and DMA,
|
|
* and bring the device offline. Note that as the driver implements the
|
|
* minimum PM requirements to a power-aware driver's PM support for the
|
|
* suspend/resume -- all the possible PM messages (SUSPEND, HIBERNATE, FREEZE)
|
|
* to the suspend() method call will be treated as SUSPEND and the driver will
|
|
* fully reinitialize its device during resume() method call, the driver will
|
|
* set device to PCI_D3hot state in PCI config space instead of setting it
|
|
* according to the @msg provided by the PM.
|
|
*
|
|
* Return code
|
|
* 0 - driver suspended the device
|
|
* Error otherwise
|
|
**/
|
|
static int
|
|
lpfc_pci_suspend_one_s3(struct pci_dev *pdev, pm_message_t msg)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
|
|
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0473 PCI device Power Management suspend.\n");
|
|
|
|
/* Bring down the device */
|
|
lpfc_offline_prep(phba);
|
|
lpfc_offline(phba);
|
|
kthread_stop(phba->worker_thread);
|
|
|
|
/* Disable interrupt from device */
|
|
lpfc_sli_disable_intr(phba);
|
|
|
|
/* Save device state to PCI config space */
|
|
pci_save_state(pdev);
|
|
pci_set_power_state(pdev, PCI_D3hot);
|
|
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_pci_resume_one_s3 - PCI func to resume SLI-3 device for power mgmnt
|
|
* @pdev: pointer to PCI device
|
|
*
|
|
* This routine is to be called from the kernel's PCI subsystem to support
|
|
* system Power Management (PM) to device with SLI-3 interface spec. When PM
|
|
* invokes this method, it restores the device's PCI config space state and
|
|
* fully reinitializes the device and brings it online. Note that as the
|
|
* driver implements the minimum PM requirements to a power-aware driver's
|
|
* PM for suspend/resume -- all the possible PM messages (SUSPEND, HIBERNATE,
|
|
* FREEZE) to the suspend() method call will be treated as SUSPEND and the
|
|
* driver will fully reinitialize its device during resume() method call,
|
|
* the device will be set to PCI_D0 directly in PCI config space before
|
|
* restoring the state.
|
|
*
|
|
* Return code
|
|
* 0 - driver suspended the device
|
|
* Error otherwise
|
|
**/
|
|
static int
|
|
lpfc_pci_resume_one_s3(struct pci_dev *pdev)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
|
|
uint32_t intr_mode;
|
|
int error;
|
|
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0452 PCI device Power Management resume.\n");
|
|
|
|
/* Restore device state from PCI config space */
|
|
pci_set_power_state(pdev, PCI_D0);
|
|
pci_restore_state(pdev);
|
|
|
|
/*
|
|
* As the new kernel behavior of pci_restore_state() API call clears
|
|
* device saved_state flag, need to save the restored state again.
|
|
*/
|
|
pci_save_state(pdev);
|
|
|
|
if (pdev->is_busmaster)
|
|
pci_set_master(pdev);
|
|
|
|
/* Startup the kernel thread for this host adapter. */
|
|
phba->worker_thread = kthread_run(lpfc_do_work, phba,
|
|
"lpfc_worker_%d", phba->brd_no);
|
|
if (IS_ERR(phba->worker_thread)) {
|
|
error = PTR_ERR(phba->worker_thread);
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0434 PM resume failed to start worker "
|
|
"thread: error=x%x.\n", error);
|
|
return error;
|
|
}
|
|
|
|
/* Configure and enable interrupt */
|
|
intr_mode = lpfc_sli_enable_intr(phba, phba->intr_mode);
|
|
if (intr_mode == LPFC_INTR_ERROR) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0430 PM resume Failed to enable interrupt\n");
|
|
return -EIO;
|
|
} else
|
|
phba->intr_mode = intr_mode;
|
|
|
|
/* Restart HBA and bring it online */
|
|
lpfc_sli_brdrestart(phba);
|
|
lpfc_online(phba);
|
|
|
|
/* Log the current active interrupt mode */
|
|
lpfc_log_intr_mode(phba, phba->intr_mode);
|
|
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli_prep_dev_for_recover - Prepare SLI3 device for pci slot recover
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is called to prepare the SLI3 device for PCI slot recover. It
|
|
* aborts all the outstanding SCSI I/Os to the pci device.
|
|
**/
|
|
static void
|
|
lpfc_sli_prep_dev_for_recover(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_sli *psli = &phba->sli;
|
|
struct lpfc_sli_ring *pring;
|
|
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2723 PCI channel I/O abort preparing for recovery\n");
|
|
|
|
/*
|
|
* There may be errored I/Os through HBA, abort all I/Os on txcmplq
|
|
* and let the SCSI mid-layer to retry them to recover.
|
|
*/
|
|
pring = &psli->ring[psli->fcp_ring];
|
|
lpfc_sli_abort_iocb_ring(phba, pring);
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli_prep_dev_for_reset - Prepare SLI3 device for pci slot reset
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is called to prepare the SLI3 device for PCI slot reset. It
|
|
* disables the device interrupt and pci device, and aborts the internal FCP
|
|
* pending I/Os.
|
|
**/
|
|
static void
|
|
lpfc_sli_prep_dev_for_reset(struct lpfc_hba *phba)
|
|
{
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2710 PCI channel disable preparing for reset\n");
|
|
|
|
/* Block any management I/Os to the device */
|
|
lpfc_block_mgmt_io(phba);
|
|
|
|
/* Block all SCSI devices' I/Os on the host */
|
|
lpfc_scsi_dev_block(phba);
|
|
|
|
/* stop all timers */
|
|
lpfc_stop_hba_timers(phba);
|
|
|
|
/* Disable interrupt and pci device */
|
|
lpfc_sli_disable_intr(phba);
|
|
pci_disable_device(phba->pcidev);
|
|
|
|
/* Flush all driver's outstanding SCSI I/Os as we are to reset */
|
|
lpfc_sli_flush_fcp_rings(phba);
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli_prep_dev_for_perm_failure - Prepare SLI3 dev for pci slot disable
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is called to prepare the SLI3 device for PCI slot permanently
|
|
* disabling. It blocks the SCSI transport layer traffic and flushes the FCP
|
|
* pending I/Os.
|
|
**/
|
|
static void
|
|
lpfc_sli_prep_dev_for_perm_failure(struct lpfc_hba *phba)
|
|
{
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2711 PCI channel permanent disable for failure\n");
|
|
/* Block all SCSI devices' I/Os on the host */
|
|
lpfc_scsi_dev_block(phba);
|
|
|
|
/* stop all timers */
|
|
lpfc_stop_hba_timers(phba);
|
|
|
|
/* Clean up all driver's outstanding SCSI I/Os */
|
|
lpfc_sli_flush_fcp_rings(phba);
|
|
}
|
|
|
|
/**
|
|
* lpfc_io_error_detected_s3 - Method for handling SLI-3 device PCI I/O error
|
|
* @pdev: pointer to PCI device.
|
|
* @state: the current PCI connection state.
|
|
*
|
|
* This routine is called from the PCI subsystem for I/O error handling to
|
|
* device with SLI-3 interface spec. This function is called by the PCI
|
|
* subsystem after a PCI bus error affecting this device has been detected.
|
|
* When this function is invoked, it will need to stop all the I/Os and
|
|
* interrupt(s) to the device. Once that is done, it will return
|
|
* PCI_ERS_RESULT_NEED_RESET for the PCI subsystem to perform proper recovery
|
|
* as desired.
|
|
*
|
|
* Return codes
|
|
* PCI_ERS_RESULT_CAN_RECOVER - can be recovered with reset_link
|
|
* PCI_ERS_RESULT_NEED_RESET - need to reset before recovery
|
|
* PCI_ERS_RESULT_DISCONNECT - device could not be recovered
|
|
**/
|
|
static pci_ers_result_t
|
|
lpfc_io_error_detected_s3(struct pci_dev *pdev, pci_channel_state_t state)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
|
|
|
|
switch (state) {
|
|
case pci_channel_io_normal:
|
|
/* Non-fatal error, prepare for recovery */
|
|
lpfc_sli_prep_dev_for_recover(phba);
|
|
return PCI_ERS_RESULT_CAN_RECOVER;
|
|
case pci_channel_io_frozen:
|
|
/* Fatal error, prepare for slot reset */
|
|
lpfc_sli_prep_dev_for_reset(phba);
|
|
return PCI_ERS_RESULT_NEED_RESET;
|
|
case pci_channel_io_perm_failure:
|
|
/* Permanent failure, prepare for device down */
|
|
lpfc_sli_prep_dev_for_perm_failure(phba);
|
|
return PCI_ERS_RESULT_DISCONNECT;
|
|
default:
|
|
/* Unknown state, prepare and request slot reset */
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0472 Unknown PCI error state: x%x\n", state);
|
|
lpfc_sli_prep_dev_for_reset(phba);
|
|
return PCI_ERS_RESULT_NEED_RESET;
|
|
}
|
|
}
|
|
|
|
/**
|
|
* lpfc_io_slot_reset_s3 - Method for restarting PCI SLI-3 device from scratch.
|
|
* @pdev: pointer to PCI device.
|
|
*
|
|
* This routine is called from the PCI subsystem for error handling to
|
|
* device with SLI-3 interface spec. This is called after PCI bus has been
|
|
* reset to restart the PCI card from scratch, as if from a cold-boot.
|
|
* During the PCI subsystem error recovery, after driver returns
|
|
* PCI_ERS_RESULT_NEED_RESET, the PCI subsystem will perform proper error
|
|
* recovery and then call this routine before calling the .resume method
|
|
* to recover the device. This function will initialize the HBA device,
|
|
* enable the interrupt, but it will just put the HBA to offline state
|
|
* without passing any I/O traffic.
|
|
*
|
|
* Return codes
|
|
* PCI_ERS_RESULT_RECOVERED - the device has been recovered
|
|
* PCI_ERS_RESULT_DISCONNECT - device could not be recovered
|
|
*/
|
|
static pci_ers_result_t
|
|
lpfc_io_slot_reset_s3(struct pci_dev *pdev)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
|
|
struct lpfc_sli *psli = &phba->sli;
|
|
uint32_t intr_mode;
|
|
|
|
dev_printk(KERN_INFO, &pdev->dev, "recovering from a slot reset.\n");
|
|
if (pci_enable_device_mem(pdev)) {
|
|
printk(KERN_ERR "lpfc: Cannot re-enable "
|
|
"PCI device after reset.\n");
|
|
return PCI_ERS_RESULT_DISCONNECT;
|
|
}
|
|
|
|
pci_restore_state(pdev);
|
|
|
|
/*
|
|
* As the new kernel behavior of pci_restore_state() API call clears
|
|
* device saved_state flag, need to save the restored state again.
|
|
*/
|
|
pci_save_state(pdev);
|
|
|
|
if (pdev->is_busmaster)
|
|
pci_set_master(pdev);
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
psli->sli_flag &= ~LPFC_SLI_ACTIVE;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
/* Configure and enable interrupt */
|
|
intr_mode = lpfc_sli_enable_intr(phba, phba->intr_mode);
|
|
if (intr_mode == LPFC_INTR_ERROR) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0427 Cannot re-enable interrupt after "
|
|
"slot reset.\n");
|
|
return PCI_ERS_RESULT_DISCONNECT;
|
|
} else
|
|
phba->intr_mode = intr_mode;
|
|
|
|
/* Take device offline, it will perform cleanup */
|
|
lpfc_offline_prep(phba);
|
|
lpfc_offline(phba);
|
|
lpfc_sli_brdrestart(phba);
|
|
|
|
/* Log the current active interrupt mode */
|
|
lpfc_log_intr_mode(phba, phba->intr_mode);
|
|
|
|
return PCI_ERS_RESULT_RECOVERED;
|
|
}
|
|
|
|
/**
|
|
* lpfc_io_resume_s3 - Method for resuming PCI I/O operation on SLI-3 device.
|
|
* @pdev: pointer to PCI device
|
|
*
|
|
* This routine is called from the PCI subsystem for error handling to device
|
|
* with SLI-3 interface spec. It is called when kernel error recovery tells
|
|
* the lpfc driver that it is ok to resume normal PCI operation after PCI bus
|
|
* error recovery. After this call, traffic can start to flow from this device
|
|
* again.
|
|
*/
|
|
static void
|
|
lpfc_io_resume_s3(struct pci_dev *pdev)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
|
|
|
|
/* Bring device online, it will be no-op for non-fatal error resume */
|
|
lpfc_online(phba);
|
|
|
|
/* Clean up Advanced Error Reporting (AER) if needed */
|
|
if (phba->hba_flag & HBA_AER_ENABLED)
|
|
pci_cleanup_aer_uncorrect_error_status(pdev);
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_get_els_iocb_cnt - Calculate the # of ELS IOCBs to reserve
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* returns the number of ELS/CT IOCBs to reserve
|
|
**/
|
|
int
|
|
lpfc_sli4_get_els_iocb_cnt(struct lpfc_hba *phba)
|
|
{
|
|
int max_xri = phba->sli4_hba.max_cfg_param.max_xri;
|
|
|
|
if (phba->sli_rev == LPFC_SLI_REV4) {
|
|
if (max_xri <= 100)
|
|
return 10;
|
|
else if (max_xri <= 256)
|
|
return 25;
|
|
else if (max_xri <= 512)
|
|
return 50;
|
|
else if (max_xri <= 1024)
|
|
return 100;
|
|
else
|
|
return 150;
|
|
} else
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_pci_probe_one_s4 - PCI probe func to reg SLI-4 device to PCI subsys
|
|
* @pdev: pointer to PCI device
|
|
* @pid: pointer to PCI device identifier
|
|
*
|
|
* This routine is called from the kernel's PCI subsystem to device with
|
|
* SLI-4 interface spec. When an Emulex HBA with SLI-4 interface spec is
|
|
* presented on PCI bus, the kernel PCI subsystem looks at PCI device-specific
|
|
* information of the device and driver to see if the driver state that it
|
|
* can support this kind of device. If the match is successful, the driver
|
|
* core invokes this routine. If this routine determines it can claim the HBA,
|
|
* it does all the initialization that it needs to do to handle the HBA
|
|
* properly.
|
|
*
|
|
* Return code
|
|
* 0 - driver can claim the device
|
|
* negative value - driver can not claim the device
|
|
**/
|
|
static int __devinit
|
|
lpfc_pci_probe_one_s4(struct pci_dev *pdev, const struct pci_device_id *pid)
|
|
{
|
|
struct lpfc_hba *phba;
|
|
struct lpfc_vport *vport = NULL;
|
|
struct Scsi_Host *shost = NULL;
|
|
int error;
|
|
uint32_t cfg_mode, intr_mode;
|
|
int mcnt;
|
|
|
|
/* Allocate memory for HBA structure */
|
|
phba = lpfc_hba_alloc(pdev);
|
|
if (!phba)
|
|
return -ENOMEM;
|
|
|
|
/* Perform generic PCI device enabling operation */
|
|
error = lpfc_enable_pci_dev(phba);
|
|
if (error) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1409 Failed to enable pci device.\n");
|
|
goto out_free_phba;
|
|
}
|
|
|
|
/* Set up SLI API function jump table for PCI-device group-1 HBAs */
|
|
error = lpfc_api_table_setup(phba, LPFC_PCI_DEV_OC);
|
|
if (error)
|
|
goto out_disable_pci_dev;
|
|
|
|
/* Set up SLI-4 specific device PCI memory space */
|
|
error = lpfc_sli4_pci_mem_setup(phba);
|
|
if (error) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1410 Failed to set up pci memory space.\n");
|
|
goto out_disable_pci_dev;
|
|
}
|
|
|
|
/* Set up phase-1 common device driver resources */
|
|
error = lpfc_setup_driver_resource_phase1(phba);
|
|
if (error) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1411 Failed to set up driver resource.\n");
|
|
goto out_unset_pci_mem_s4;
|
|
}
|
|
|
|
/* Set up SLI-4 Specific device driver resources */
|
|
error = lpfc_sli4_driver_resource_setup(phba);
|
|
if (error) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1412 Failed to set up driver resource.\n");
|
|
goto out_unset_pci_mem_s4;
|
|
}
|
|
|
|
/* Initialize and populate the iocb list per host */
|
|
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"2821 initialize iocb list %d.\n",
|
|
phba->cfg_iocb_cnt*1024);
|
|
error = lpfc_init_iocb_list(phba, phba->cfg_iocb_cnt*1024);
|
|
|
|
if (error) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1413 Failed to initialize iocb list.\n");
|
|
goto out_unset_driver_resource_s4;
|
|
}
|
|
|
|
/* Set up common device driver resources */
|
|
error = lpfc_setup_driver_resource_phase2(phba);
|
|
if (error) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1414 Failed to set up driver resource.\n");
|
|
goto out_free_iocb_list;
|
|
}
|
|
|
|
/* Create SCSI host to the physical port */
|
|
error = lpfc_create_shost(phba);
|
|
if (error) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1415 Failed to create scsi host.\n");
|
|
goto out_unset_driver_resource;
|
|
}
|
|
|
|
/* Configure sysfs attributes */
|
|
vport = phba->pport;
|
|
error = lpfc_alloc_sysfs_attr(vport);
|
|
if (error) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1416 Failed to allocate sysfs attr\n");
|
|
goto out_destroy_shost;
|
|
}
|
|
|
|
shost = lpfc_shost_from_vport(vport); /* save shost for error cleanup */
|
|
/* Now, trying to enable interrupt and bring up the device */
|
|
cfg_mode = phba->cfg_use_msi;
|
|
while (true) {
|
|
/* Put device to a known state before enabling interrupt */
|
|
lpfc_stop_port(phba);
|
|
/* Configure and enable interrupt */
|
|
intr_mode = lpfc_sli4_enable_intr(phba, cfg_mode);
|
|
if (intr_mode == LPFC_INTR_ERROR) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0426 Failed to enable interrupt.\n");
|
|
error = -ENODEV;
|
|
goto out_free_sysfs_attr;
|
|
}
|
|
/* Default to single FCP EQ for non-MSI-X */
|
|
if (phba->intr_type != MSIX)
|
|
phba->cfg_fcp_eq_count = 1;
|
|
else if (phba->sli4_hba.msix_vec_nr < phba->cfg_fcp_eq_count)
|
|
phba->cfg_fcp_eq_count = phba->sli4_hba.msix_vec_nr - 1;
|
|
/* Set up SLI-4 HBA */
|
|
if (lpfc_sli4_hba_setup(phba)) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1421 Failed to set up hba\n");
|
|
error = -ENODEV;
|
|
goto out_disable_intr;
|
|
}
|
|
|
|
/* Send NOP mbx cmds for non-INTx mode active interrupt test */
|
|
if (intr_mode != 0)
|
|
mcnt = lpfc_sli4_send_nop_mbox_cmds(phba,
|
|
LPFC_ACT_INTR_CNT);
|
|
|
|
/* Check active interrupts received only for MSI/MSI-X */
|
|
if (intr_mode == 0 ||
|
|
phba->sli.slistat.sli_intr >= LPFC_ACT_INTR_CNT) {
|
|
/* Log the current active interrupt mode */
|
|
phba->intr_mode = intr_mode;
|
|
lpfc_log_intr_mode(phba, intr_mode);
|
|
break;
|
|
}
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0451 Configure interrupt mode (%d) "
|
|
"failed active interrupt test.\n",
|
|
intr_mode);
|
|
/* Unset the preivous SLI-4 HBA setup */
|
|
lpfc_sli4_unset_hba(phba);
|
|
/* Try next level of interrupt mode */
|
|
cfg_mode = --intr_mode;
|
|
}
|
|
|
|
/* Perform post initialization setup */
|
|
lpfc_post_init_setup(phba);
|
|
|
|
/* Check if there are static vports to be created. */
|
|
lpfc_create_static_vport(phba);
|
|
|
|
return 0;
|
|
|
|
out_disable_intr:
|
|
lpfc_sli4_disable_intr(phba);
|
|
out_free_sysfs_attr:
|
|
lpfc_free_sysfs_attr(vport);
|
|
out_destroy_shost:
|
|
lpfc_destroy_shost(phba);
|
|
out_unset_driver_resource:
|
|
lpfc_unset_driver_resource_phase2(phba);
|
|
out_free_iocb_list:
|
|
lpfc_free_iocb_list(phba);
|
|
out_unset_driver_resource_s4:
|
|
lpfc_sli4_driver_resource_unset(phba);
|
|
out_unset_pci_mem_s4:
|
|
lpfc_sli4_pci_mem_unset(phba);
|
|
out_disable_pci_dev:
|
|
lpfc_disable_pci_dev(phba);
|
|
if (shost)
|
|
scsi_host_put(shost);
|
|
out_free_phba:
|
|
lpfc_hba_free(phba);
|
|
return error;
|
|
}
|
|
|
|
/**
|
|
* lpfc_pci_remove_one_s4 - PCI func to unreg SLI-4 device from PCI subsystem
|
|
* @pdev: pointer to PCI device
|
|
*
|
|
* This routine is called from the kernel's PCI subsystem to device with
|
|
* SLI-4 interface spec. When an Emulex HBA with SLI-4 interface spec is
|
|
* removed from PCI bus, it performs all the necessary cleanup for the HBA
|
|
* device to be removed from the PCI subsystem properly.
|
|
**/
|
|
static void __devexit
|
|
lpfc_pci_remove_one_s4(struct pci_dev *pdev)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
|
|
struct lpfc_vport **vports;
|
|
struct lpfc_hba *phba = vport->phba;
|
|
int i;
|
|
|
|
/* Mark the device unloading flag */
|
|
spin_lock_irq(&phba->hbalock);
|
|
vport->load_flag |= FC_UNLOADING;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
/* Free the HBA sysfs attributes */
|
|
lpfc_free_sysfs_attr(vport);
|
|
|
|
/* Release all the vports against this physical port */
|
|
vports = lpfc_create_vport_work_array(phba);
|
|
if (vports != NULL)
|
|
for (i = 1; i <= phba->max_vports && vports[i] != NULL; i++)
|
|
fc_vport_terminate(vports[i]->fc_vport);
|
|
lpfc_destroy_vport_work_array(phba, vports);
|
|
|
|
/* Remove FC host and then SCSI host with the physical port */
|
|
fc_remove_host(shost);
|
|
scsi_remove_host(shost);
|
|
|
|
/* Perform cleanup on the physical port */
|
|
lpfc_cleanup(vport);
|
|
|
|
/*
|
|
* Bring down the SLI Layer. This step disables all interrupts,
|
|
* clears the rings, discards all mailbox commands, and resets
|
|
* the HBA FCoE function.
|
|
*/
|
|
lpfc_debugfs_terminate(vport);
|
|
lpfc_sli4_hba_unset(phba);
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
list_del_init(&vport->listentry);
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
/* Perform scsi free before driver resource_unset since scsi
|
|
* buffers are released to their corresponding pools here.
|
|
*/
|
|
lpfc_scsi_free(phba);
|
|
lpfc_sli4_driver_resource_unset(phba);
|
|
|
|
/* Unmap adapter Control and Doorbell registers */
|
|
lpfc_sli4_pci_mem_unset(phba);
|
|
|
|
/* Release PCI resources and disable device's PCI function */
|
|
scsi_host_put(shost);
|
|
lpfc_disable_pci_dev(phba);
|
|
|
|
/* Finally, free the driver's device data structure */
|
|
lpfc_hba_free(phba);
|
|
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_pci_suspend_one_s4 - PCI func to suspend SLI-4 device for power mgmnt
|
|
* @pdev: pointer to PCI device
|
|
* @msg: power management message
|
|
*
|
|
* This routine is called from the kernel's PCI subsystem to support system
|
|
* Power Management (PM) to device with SLI-4 interface spec. When PM invokes
|
|
* this method, it quiesces the device by stopping the driver's worker
|
|
* thread for the device, turning off device's interrupt and DMA, and bring
|
|
* the device offline. Note that as the driver implements the minimum PM
|
|
* requirements to a power-aware driver's PM support for suspend/resume -- all
|
|
* the possible PM messages (SUSPEND, HIBERNATE, FREEZE) to the suspend()
|
|
* method call will be treated as SUSPEND and the driver will fully
|
|
* reinitialize its device during resume() method call, the driver will set
|
|
* device to PCI_D3hot state in PCI config space instead of setting it
|
|
* according to the @msg provided by the PM.
|
|
*
|
|
* Return code
|
|
* 0 - driver suspended the device
|
|
* Error otherwise
|
|
**/
|
|
static int
|
|
lpfc_pci_suspend_one_s4(struct pci_dev *pdev, pm_message_t msg)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
|
|
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"2843 PCI device Power Management suspend.\n");
|
|
|
|
/* Bring down the device */
|
|
lpfc_offline_prep(phba);
|
|
lpfc_offline(phba);
|
|
kthread_stop(phba->worker_thread);
|
|
|
|
/* Disable interrupt from device */
|
|
lpfc_sli4_disable_intr(phba);
|
|
|
|
/* Save device state to PCI config space */
|
|
pci_save_state(pdev);
|
|
pci_set_power_state(pdev, PCI_D3hot);
|
|
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_pci_resume_one_s4 - PCI func to resume SLI-4 device for power mgmnt
|
|
* @pdev: pointer to PCI device
|
|
*
|
|
* This routine is called from the kernel's PCI subsystem to support system
|
|
* Power Management (PM) to device with SLI-4 interface spac. When PM invokes
|
|
* this method, it restores the device's PCI config space state and fully
|
|
* reinitializes the device and brings it online. Note that as the driver
|
|
* implements the minimum PM requirements to a power-aware driver's PM for
|
|
* suspend/resume -- all the possible PM messages (SUSPEND, HIBERNATE, FREEZE)
|
|
* to the suspend() method call will be treated as SUSPEND and the driver
|
|
* will fully reinitialize its device during resume() method call, the device
|
|
* will be set to PCI_D0 directly in PCI config space before restoring the
|
|
* state.
|
|
*
|
|
* Return code
|
|
* 0 - driver suspended the device
|
|
* Error otherwise
|
|
**/
|
|
static int
|
|
lpfc_pci_resume_one_s4(struct pci_dev *pdev)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
|
|
uint32_t intr_mode;
|
|
int error;
|
|
|
|
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
|
|
"0292 PCI device Power Management resume.\n");
|
|
|
|
/* Restore device state from PCI config space */
|
|
pci_set_power_state(pdev, PCI_D0);
|
|
pci_restore_state(pdev);
|
|
|
|
/*
|
|
* As the new kernel behavior of pci_restore_state() API call clears
|
|
* device saved_state flag, need to save the restored state again.
|
|
*/
|
|
pci_save_state(pdev);
|
|
|
|
if (pdev->is_busmaster)
|
|
pci_set_master(pdev);
|
|
|
|
/* Startup the kernel thread for this host adapter. */
|
|
phba->worker_thread = kthread_run(lpfc_do_work, phba,
|
|
"lpfc_worker_%d", phba->brd_no);
|
|
if (IS_ERR(phba->worker_thread)) {
|
|
error = PTR_ERR(phba->worker_thread);
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0293 PM resume failed to start worker "
|
|
"thread: error=x%x.\n", error);
|
|
return error;
|
|
}
|
|
|
|
/* Configure and enable interrupt */
|
|
intr_mode = lpfc_sli4_enable_intr(phba, phba->intr_mode);
|
|
if (intr_mode == LPFC_INTR_ERROR) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"0294 PM resume Failed to enable interrupt\n");
|
|
return -EIO;
|
|
} else
|
|
phba->intr_mode = intr_mode;
|
|
|
|
/* Restart HBA and bring it online */
|
|
lpfc_sli_brdrestart(phba);
|
|
lpfc_online(phba);
|
|
|
|
/* Log the current active interrupt mode */
|
|
lpfc_log_intr_mode(phba, phba->intr_mode);
|
|
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_prep_dev_for_recover - Prepare SLI4 device for pci slot recover
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is called to prepare the SLI4 device for PCI slot recover. It
|
|
* aborts all the outstanding SCSI I/Os to the pci device.
|
|
**/
|
|
static void
|
|
lpfc_sli4_prep_dev_for_recover(struct lpfc_hba *phba)
|
|
{
|
|
struct lpfc_sli *psli = &phba->sli;
|
|
struct lpfc_sli_ring *pring;
|
|
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2828 PCI channel I/O abort preparing for recovery\n");
|
|
/*
|
|
* There may be errored I/Os through HBA, abort all I/Os on txcmplq
|
|
* and let the SCSI mid-layer to retry them to recover.
|
|
*/
|
|
pring = &psli->ring[psli->fcp_ring];
|
|
lpfc_sli_abort_iocb_ring(phba, pring);
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_prep_dev_for_reset - Prepare SLI4 device for pci slot reset
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is called to prepare the SLI4 device for PCI slot reset. It
|
|
* disables the device interrupt and pci device, and aborts the internal FCP
|
|
* pending I/Os.
|
|
**/
|
|
static void
|
|
lpfc_sli4_prep_dev_for_reset(struct lpfc_hba *phba)
|
|
{
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2826 PCI channel disable preparing for reset\n");
|
|
|
|
/* Block any management I/Os to the device */
|
|
lpfc_block_mgmt_io(phba);
|
|
|
|
/* Block all SCSI devices' I/Os on the host */
|
|
lpfc_scsi_dev_block(phba);
|
|
|
|
/* stop all timers */
|
|
lpfc_stop_hba_timers(phba);
|
|
|
|
/* Disable interrupt and pci device */
|
|
lpfc_sli4_disable_intr(phba);
|
|
pci_disable_device(phba->pcidev);
|
|
|
|
/* Flush all driver's outstanding SCSI I/Os as we are to reset */
|
|
lpfc_sli_flush_fcp_rings(phba);
|
|
}
|
|
|
|
/**
|
|
* lpfc_sli4_prep_dev_for_perm_failure - Prepare SLI4 dev for pci slot disable
|
|
* @phba: pointer to lpfc hba data structure.
|
|
*
|
|
* This routine is called to prepare the SLI4 device for PCI slot permanently
|
|
* disabling. It blocks the SCSI transport layer traffic and flushes the FCP
|
|
* pending I/Os.
|
|
**/
|
|
static void
|
|
lpfc_sli4_prep_dev_for_perm_failure(struct lpfc_hba *phba)
|
|
{
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2827 PCI channel permanent disable for failure\n");
|
|
|
|
/* Block all SCSI devices' I/Os on the host */
|
|
lpfc_scsi_dev_block(phba);
|
|
|
|
/* stop all timers */
|
|
lpfc_stop_hba_timers(phba);
|
|
|
|
/* Clean up all driver's outstanding SCSI I/Os */
|
|
lpfc_sli_flush_fcp_rings(phba);
|
|
}
|
|
|
|
/**
|
|
* lpfc_io_error_detected_s4 - Method for handling PCI I/O error to SLI-4 device
|
|
* @pdev: pointer to PCI device.
|
|
* @state: the current PCI connection state.
|
|
*
|
|
* This routine is called from the PCI subsystem for error handling to device
|
|
* with SLI-4 interface spec. This function is called by the PCI subsystem
|
|
* after a PCI bus error affecting this device has been detected. When this
|
|
* function is invoked, it will need to stop all the I/Os and interrupt(s)
|
|
* to the device. Once that is done, it will return PCI_ERS_RESULT_NEED_RESET
|
|
* for the PCI subsystem to perform proper recovery as desired.
|
|
*
|
|
* Return codes
|
|
* PCI_ERS_RESULT_NEED_RESET - need to reset before recovery
|
|
* PCI_ERS_RESULT_DISCONNECT - device could not be recovered
|
|
**/
|
|
static pci_ers_result_t
|
|
lpfc_io_error_detected_s4(struct pci_dev *pdev, pci_channel_state_t state)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
|
|
|
|
switch (state) {
|
|
case pci_channel_io_normal:
|
|
/* Non-fatal error, prepare for recovery */
|
|
lpfc_sli4_prep_dev_for_recover(phba);
|
|
return PCI_ERS_RESULT_CAN_RECOVER;
|
|
case pci_channel_io_frozen:
|
|
/* Fatal error, prepare for slot reset */
|
|
lpfc_sli4_prep_dev_for_reset(phba);
|
|
return PCI_ERS_RESULT_NEED_RESET;
|
|
case pci_channel_io_perm_failure:
|
|
/* Permanent failure, prepare for device down */
|
|
lpfc_sli4_prep_dev_for_perm_failure(phba);
|
|
return PCI_ERS_RESULT_DISCONNECT;
|
|
default:
|
|
/* Unknown state, prepare and request slot reset */
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2825 Unknown PCI error state: x%x\n", state);
|
|
lpfc_sli4_prep_dev_for_reset(phba);
|
|
return PCI_ERS_RESULT_NEED_RESET;
|
|
}
|
|
}
|
|
|
|
/**
|
|
* lpfc_io_slot_reset_s4 - Method for restart PCI SLI-4 device from scratch
|
|
* @pdev: pointer to PCI device.
|
|
*
|
|
* This routine is called from the PCI subsystem for error handling to device
|
|
* with SLI-4 interface spec. It is called after PCI bus has been reset to
|
|
* restart the PCI card from scratch, as if from a cold-boot. During the
|
|
* PCI subsystem error recovery, after the driver returns
|
|
* PCI_ERS_RESULT_NEED_RESET, the PCI subsystem will perform proper error
|
|
* recovery and then call this routine before calling the .resume method to
|
|
* recover the device. This function will initialize the HBA device, enable
|
|
* the interrupt, but it will just put the HBA to offline state without
|
|
* passing any I/O traffic.
|
|
*
|
|
* Return codes
|
|
* PCI_ERS_RESULT_RECOVERED - the device has been recovered
|
|
* PCI_ERS_RESULT_DISCONNECT - device could not be recovered
|
|
*/
|
|
static pci_ers_result_t
|
|
lpfc_io_slot_reset_s4(struct pci_dev *pdev)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
|
|
struct lpfc_sli *psli = &phba->sli;
|
|
uint32_t intr_mode;
|
|
|
|
dev_printk(KERN_INFO, &pdev->dev, "recovering from a slot reset.\n");
|
|
if (pci_enable_device_mem(pdev)) {
|
|
printk(KERN_ERR "lpfc: Cannot re-enable "
|
|
"PCI device after reset.\n");
|
|
return PCI_ERS_RESULT_DISCONNECT;
|
|
}
|
|
|
|
pci_restore_state(pdev);
|
|
if (pdev->is_busmaster)
|
|
pci_set_master(pdev);
|
|
|
|
spin_lock_irq(&phba->hbalock);
|
|
psli->sli_flag &= ~LPFC_SLI_ACTIVE;
|
|
spin_unlock_irq(&phba->hbalock);
|
|
|
|
/* Configure and enable interrupt */
|
|
intr_mode = lpfc_sli4_enable_intr(phba, phba->intr_mode);
|
|
if (intr_mode == LPFC_INTR_ERROR) {
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"2824 Cannot re-enable interrupt after "
|
|
"slot reset.\n");
|
|
return PCI_ERS_RESULT_DISCONNECT;
|
|
} else
|
|
phba->intr_mode = intr_mode;
|
|
|
|
/* Log the current active interrupt mode */
|
|
lpfc_log_intr_mode(phba, phba->intr_mode);
|
|
|
|
return PCI_ERS_RESULT_RECOVERED;
|
|
}
|
|
|
|
/**
|
|
* lpfc_io_resume_s4 - Method for resuming PCI I/O operation to SLI-4 device
|
|
* @pdev: pointer to PCI device
|
|
*
|
|
* This routine is called from the PCI subsystem for error handling to device
|
|
* with SLI-4 interface spec. It is called when kernel error recovery tells
|
|
* the lpfc driver that it is ok to resume normal PCI operation after PCI bus
|
|
* error recovery. After this call, traffic can start to flow from this device
|
|
* again.
|
|
**/
|
|
static void
|
|
lpfc_io_resume_s4(struct pci_dev *pdev)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
|
|
|
|
/*
|
|
* In case of slot reset, as function reset is performed through
|
|
* mailbox command which needs DMA to be enabled, this operation
|
|
* has to be moved to the io resume phase. Taking device offline
|
|
* will perform the necessary cleanup.
|
|
*/
|
|
if (!(phba->sli.sli_flag & LPFC_SLI_ACTIVE)) {
|
|
/* Perform device reset */
|
|
lpfc_offline_prep(phba);
|
|
lpfc_offline(phba);
|
|
lpfc_sli_brdrestart(phba);
|
|
/* Bring the device back online */
|
|
lpfc_online(phba);
|
|
}
|
|
|
|
/* Clean up Advanced Error Reporting (AER) if needed */
|
|
if (phba->hba_flag & HBA_AER_ENABLED)
|
|
pci_cleanup_aer_uncorrect_error_status(pdev);
|
|
}
|
|
|
|
/**
|
|
* lpfc_pci_probe_one - lpfc PCI probe func to reg dev to PCI subsystem
|
|
* @pdev: pointer to PCI device
|
|
* @pid: pointer to PCI device identifier
|
|
*
|
|
* This routine is to be registered to the kernel's PCI subsystem. When an
|
|
* Emulex HBA device is presented on PCI bus, the kernel PCI subsystem looks
|
|
* at PCI device-specific information of the device and driver to see if the
|
|
* driver state that it can support this kind of device. If the match is
|
|
* successful, the driver core invokes this routine. This routine dispatches
|
|
* the action to the proper SLI-3 or SLI-4 device probing routine, which will
|
|
* do all the initialization that it needs to do to handle the HBA device
|
|
* properly.
|
|
*
|
|
* Return code
|
|
* 0 - driver can claim the device
|
|
* negative value - driver can not claim the device
|
|
**/
|
|
static int __devinit
|
|
lpfc_pci_probe_one(struct pci_dev *pdev, const struct pci_device_id *pid)
|
|
{
|
|
int rc;
|
|
struct lpfc_sli_intf intf;
|
|
|
|
if (pci_read_config_dword(pdev, LPFC_SLI_INTF, &intf.word0))
|
|
return -ENODEV;
|
|
|
|
if ((bf_get(lpfc_sli_intf_valid, &intf) == LPFC_SLI_INTF_VALID) &&
|
|
(bf_get(lpfc_sli_intf_slirev, &intf) == LPFC_SLI_INTF_REV_SLI4))
|
|
rc = lpfc_pci_probe_one_s4(pdev, pid);
|
|
else
|
|
rc = lpfc_pci_probe_one_s3(pdev, pid);
|
|
|
|
return rc;
|
|
}
|
|
|
|
/**
|
|
* lpfc_pci_remove_one - lpfc PCI func to unreg dev from PCI subsystem
|
|
* @pdev: pointer to PCI device
|
|
*
|
|
* This routine is to be registered to the kernel's PCI subsystem. When an
|
|
* Emulex HBA is removed from PCI bus, the driver core invokes this routine.
|
|
* This routine dispatches the action to the proper SLI-3 or SLI-4 device
|
|
* remove routine, which will perform all the necessary cleanup for the
|
|
* device to be removed from the PCI subsystem properly.
|
|
**/
|
|
static void __devexit
|
|
lpfc_pci_remove_one(struct pci_dev *pdev)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
|
|
|
|
switch (phba->pci_dev_grp) {
|
|
case LPFC_PCI_DEV_LP:
|
|
lpfc_pci_remove_one_s3(pdev);
|
|
break;
|
|
case LPFC_PCI_DEV_OC:
|
|
lpfc_pci_remove_one_s4(pdev);
|
|
break;
|
|
default:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1424 Invalid PCI device group: 0x%x\n",
|
|
phba->pci_dev_grp);
|
|
break;
|
|
}
|
|
return;
|
|
}
|
|
|
|
/**
|
|
* lpfc_pci_suspend_one - lpfc PCI func to suspend dev for power management
|
|
* @pdev: pointer to PCI device
|
|
* @msg: power management message
|
|
*
|
|
* This routine is to be registered to the kernel's PCI subsystem to support
|
|
* system Power Management (PM). When PM invokes this method, it dispatches
|
|
* the action to the proper SLI-3 or SLI-4 device suspend routine, which will
|
|
* suspend the device.
|
|
*
|
|
* Return code
|
|
* 0 - driver suspended the device
|
|
* Error otherwise
|
|
**/
|
|
static int
|
|
lpfc_pci_suspend_one(struct pci_dev *pdev, pm_message_t msg)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
|
|
int rc = -ENODEV;
|
|
|
|
switch (phba->pci_dev_grp) {
|
|
case LPFC_PCI_DEV_LP:
|
|
rc = lpfc_pci_suspend_one_s3(pdev, msg);
|
|
break;
|
|
case LPFC_PCI_DEV_OC:
|
|
rc = lpfc_pci_suspend_one_s4(pdev, msg);
|
|
break;
|
|
default:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1425 Invalid PCI device group: 0x%x\n",
|
|
phba->pci_dev_grp);
|
|
break;
|
|
}
|
|
return rc;
|
|
}
|
|
|
|
/**
|
|
* lpfc_pci_resume_one - lpfc PCI func to resume dev for power management
|
|
* @pdev: pointer to PCI device
|
|
*
|
|
* This routine is to be registered to the kernel's PCI subsystem to support
|
|
* system Power Management (PM). When PM invokes this method, it dispatches
|
|
* the action to the proper SLI-3 or SLI-4 device resume routine, which will
|
|
* resume the device.
|
|
*
|
|
* Return code
|
|
* 0 - driver suspended the device
|
|
* Error otherwise
|
|
**/
|
|
static int
|
|
lpfc_pci_resume_one(struct pci_dev *pdev)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
|
|
int rc = -ENODEV;
|
|
|
|
switch (phba->pci_dev_grp) {
|
|
case LPFC_PCI_DEV_LP:
|
|
rc = lpfc_pci_resume_one_s3(pdev);
|
|
break;
|
|
case LPFC_PCI_DEV_OC:
|
|
rc = lpfc_pci_resume_one_s4(pdev);
|
|
break;
|
|
default:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1426 Invalid PCI device group: 0x%x\n",
|
|
phba->pci_dev_grp);
|
|
break;
|
|
}
|
|
return rc;
|
|
}
|
|
|
|
/**
|
|
* lpfc_io_error_detected - lpfc method for handling PCI I/O error
|
|
* @pdev: pointer to PCI device.
|
|
* @state: the current PCI connection state.
|
|
*
|
|
* This routine is registered to the PCI subsystem for error handling. This
|
|
* function is called by the PCI subsystem after a PCI bus error affecting
|
|
* this device has been detected. When this routine is invoked, it dispatches
|
|
* the action to the proper SLI-3 or SLI-4 device error detected handling
|
|
* routine, which will perform the proper error detected operation.
|
|
*
|
|
* Return codes
|
|
* PCI_ERS_RESULT_NEED_RESET - need to reset before recovery
|
|
* PCI_ERS_RESULT_DISCONNECT - device could not be recovered
|
|
**/
|
|
static pci_ers_result_t
|
|
lpfc_io_error_detected(struct pci_dev *pdev, pci_channel_state_t state)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
|
|
pci_ers_result_t rc = PCI_ERS_RESULT_DISCONNECT;
|
|
|
|
switch (phba->pci_dev_grp) {
|
|
case LPFC_PCI_DEV_LP:
|
|
rc = lpfc_io_error_detected_s3(pdev, state);
|
|
break;
|
|
case LPFC_PCI_DEV_OC:
|
|
rc = lpfc_io_error_detected_s4(pdev, state);
|
|
break;
|
|
default:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1427 Invalid PCI device group: 0x%x\n",
|
|
phba->pci_dev_grp);
|
|
break;
|
|
}
|
|
return rc;
|
|
}
|
|
|
|
/**
|
|
* lpfc_io_slot_reset - lpfc method for restart PCI dev from scratch
|
|
* @pdev: pointer to PCI device.
|
|
*
|
|
* This routine is registered to the PCI subsystem for error handling. This
|
|
* function is called after PCI bus has been reset to restart the PCI card
|
|
* from scratch, as if from a cold-boot. When this routine is invoked, it
|
|
* dispatches the action to the proper SLI-3 or SLI-4 device reset handling
|
|
* routine, which will perform the proper device reset.
|
|
*
|
|
* Return codes
|
|
* PCI_ERS_RESULT_RECOVERED - the device has been recovered
|
|
* PCI_ERS_RESULT_DISCONNECT - device could not be recovered
|
|
**/
|
|
static pci_ers_result_t
|
|
lpfc_io_slot_reset(struct pci_dev *pdev)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
|
|
pci_ers_result_t rc = PCI_ERS_RESULT_DISCONNECT;
|
|
|
|
switch (phba->pci_dev_grp) {
|
|
case LPFC_PCI_DEV_LP:
|
|
rc = lpfc_io_slot_reset_s3(pdev);
|
|
break;
|
|
case LPFC_PCI_DEV_OC:
|
|
rc = lpfc_io_slot_reset_s4(pdev);
|
|
break;
|
|
default:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1428 Invalid PCI device group: 0x%x\n",
|
|
phba->pci_dev_grp);
|
|
break;
|
|
}
|
|
return rc;
|
|
}
|
|
|
|
/**
|
|
* lpfc_io_resume - lpfc method for resuming PCI I/O operation
|
|
* @pdev: pointer to PCI device
|
|
*
|
|
* This routine is registered to the PCI subsystem for error handling. It
|
|
* is called when kernel error recovery tells the lpfc driver that it is
|
|
* OK to resume normal PCI operation after PCI bus error recovery. When
|
|
* this routine is invoked, it dispatches the action to the proper SLI-3
|
|
* or SLI-4 device io_resume routine, which will resume the device operation.
|
|
**/
|
|
static void
|
|
lpfc_io_resume(struct pci_dev *pdev)
|
|
{
|
|
struct Scsi_Host *shost = pci_get_drvdata(pdev);
|
|
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
|
|
|
|
switch (phba->pci_dev_grp) {
|
|
case LPFC_PCI_DEV_LP:
|
|
lpfc_io_resume_s3(pdev);
|
|
break;
|
|
case LPFC_PCI_DEV_OC:
|
|
lpfc_io_resume_s4(pdev);
|
|
break;
|
|
default:
|
|
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
|
|
"1429 Invalid PCI device group: 0x%x\n",
|
|
phba->pci_dev_grp);
|
|
break;
|
|
}
|
|
return;
|
|
}
|
|
|
|
static struct pci_device_id lpfc_id_table[] = {
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_VIPER,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_FIREFLY,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_THOR,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_PEGASUS,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_CENTAUR,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_DRAGONFLY,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_SUPERFLY,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_RFLY,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_PFLY,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_NEPTUNE,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_NEPTUNE_SCSP,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_NEPTUNE_DCSP,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_HELIOS,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_HELIOS_SCSP,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_HELIOS_DCSP,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_BMID,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_BSMB,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_ZEPHYR,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_HORNET,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_ZEPHYR_SCSP,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_ZEPHYR_DCSP,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_ZMID,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_ZSMB,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_TFLY,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_LP101,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_LP10000S,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_LP11000S,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_LPE11000S,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_SAT,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_SAT_MID,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_SAT_SMB,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_SAT_DCSP,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_SAT_SCSP,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_SAT_S,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_PROTEUS_VF,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_PROTEUS_PF,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_PROTEUS_S,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_SERVERENGINE, PCI_DEVICE_ID_TIGERSHARK,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_SERVERENGINE, PCI_DEVICE_ID_TOMCAT,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_FALCON,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{PCI_VENDOR_ID_EMULEX, PCI_DEVICE_ID_BALIUS,
|
|
PCI_ANY_ID, PCI_ANY_ID, },
|
|
{ 0 }
|
|
};
|
|
|
|
MODULE_DEVICE_TABLE(pci, lpfc_id_table);
|
|
|
|
static struct pci_error_handlers lpfc_err_handler = {
|
|
.error_detected = lpfc_io_error_detected,
|
|
.slot_reset = lpfc_io_slot_reset,
|
|
.resume = lpfc_io_resume,
|
|
};
|
|
|
|
static struct pci_driver lpfc_driver = {
|
|
.name = LPFC_DRIVER_NAME,
|
|
.id_table = lpfc_id_table,
|
|
.probe = lpfc_pci_probe_one,
|
|
.remove = __devexit_p(lpfc_pci_remove_one),
|
|
.suspend = lpfc_pci_suspend_one,
|
|
.resume = lpfc_pci_resume_one,
|
|
.err_handler = &lpfc_err_handler,
|
|
};
|
|
|
|
/**
|
|
* lpfc_init - lpfc module initialization routine
|
|
*
|
|
* This routine is to be invoked when the lpfc module is loaded into the
|
|
* kernel. The special kernel macro module_init() is used to indicate the
|
|
* role of this routine to the kernel as lpfc module entry point.
|
|
*
|
|
* Return codes
|
|
* 0 - successful
|
|
* -ENOMEM - FC attach transport failed
|
|
* all others - failed
|
|
*/
|
|
static int __init
|
|
lpfc_init(void)
|
|
{
|
|
int error = 0;
|
|
|
|
printk(LPFC_MODULE_DESC "\n");
|
|
printk(LPFC_COPYRIGHT "\n");
|
|
|
|
if (lpfc_enable_npiv) {
|
|
lpfc_transport_functions.vport_create = lpfc_vport_create;
|
|
lpfc_transport_functions.vport_delete = lpfc_vport_delete;
|
|
}
|
|
lpfc_transport_template =
|
|
fc_attach_transport(&lpfc_transport_functions);
|
|
if (lpfc_transport_template == NULL)
|
|
return -ENOMEM;
|
|
if (lpfc_enable_npiv) {
|
|
lpfc_vport_transport_template =
|
|
fc_attach_transport(&lpfc_vport_transport_functions);
|
|
if (lpfc_vport_transport_template == NULL) {
|
|
fc_release_transport(lpfc_transport_template);
|
|
return -ENOMEM;
|
|
}
|
|
}
|
|
error = pci_register_driver(&lpfc_driver);
|
|
if (error) {
|
|
fc_release_transport(lpfc_transport_template);
|
|
if (lpfc_enable_npiv)
|
|
fc_release_transport(lpfc_vport_transport_template);
|
|
}
|
|
|
|
return error;
|
|
}
|
|
|
|
/**
|
|
* lpfc_exit - lpfc module removal routine
|
|
*
|
|
* This routine is invoked when the lpfc module is removed from the kernel.
|
|
* The special kernel macro module_exit() is used to indicate the role of
|
|
* this routine to the kernel as lpfc module exit point.
|
|
*/
|
|
static void __exit
|
|
lpfc_exit(void)
|
|
{
|
|
pci_unregister_driver(&lpfc_driver);
|
|
fc_release_transport(lpfc_transport_template);
|
|
if (lpfc_enable_npiv)
|
|
fc_release_transport(lpfc_vport_transport_template);
|
|
if (_dump_buf_data) {
|
|
printk(KERN_ERR "9062 BLKGRD: freeing %lu pages for "
|
|
"_dump_buf_data at 0x%p\n",
|
|
(1L << _dump_buf_data_order), _dump_buf_data);
|
|
free_pages((unsigned long)_dump_buf_data, _dump_buf_data_order);
|
|
}
|
|
|
|
if (_dump_buf_dif) {
|
|
printk(KERN_ERR "9049 BLKGRD: freeing %lu pages for "
|
|
"_dump_buf_dif at 0x%p\n",
|
|
(1L << _dump_buf_dif_order), _dump_buf_dif);
|
|
free_pages((unsigned long)_dump_buf_dif, _dump_buf_dif_order);
|
|
}
|
|
}
|
|
|
|
module_init(lpfc_init);
|
|
module_exit(lpfc_exit);
|
|
MODULE_LICENSE("GPL");
|
|
MODULE_DESCRIPTION(LPFC_MODULE_DESC);
|
|
MODULE_AUTHOR("Emulex Corporation - tech.support@emulex.com");
|
|
MODULE_VERSION("0:" LPFC_DRIVER_VERSION);
|