2023-01-25 20:00:44 +00:00
|
|
|
// SPDX-License-Identifier: GPL-2.0-only
|
2005-04-16 22:20:36 +00:00
|
|
|
/*
|
|
|
|
* Copyright (C) 2001, 2002 Sistina Software (UK) Limited.
|
2006-06-26 07:27:32 +00:00
|
|
|
* Copyright (C) 2004 - 2006 Red Hat, Inc. All rights reserved.
|
2005-04-16 22:20:36 +00:00
|
|
|
*
|
|
|
|
* This file is released under the GPL.
|
|
|
|
*/
|
|
|
|
|
2016-05-12 20:28:10 +00:00
|
|
|
#include "dm-core.h"
|
dm ima: measure data on table load
DM configures a block device with various target specific attributes
passed to it as a table. DM loads the table, and calls each target’s
respective constructors with the attributes as input parameters.
Some of these attributes are critical to ensure the device meets
certain security bar. Thus, IMA should measure these attributes, to
ensure they are not tampered with, during the lifetime of the device.
So that the external services can have high confidence in the
configuration of the block-devices on a given system.
Some devices may have large tables. And a given device may change its
state (table-load, suspend, resume, rename, remove, table-clear etc.)
many times. Measuring these attributes each time when the device
changes its state will significantly increase the size of the IMA logs.
Further, once configured, these attributes are not expected to change
unless a new table is loaded, or a device is removed and recreated.
Therefore the clear-text of the attributes should only be measured
during table load, and the hash of the active/inactive table should be
measured for the remaining device state changes.
Export IMA function ima_measure_critical_data() to allow measurement
of DM device parameters, as well as target specific attributes, during
table load. Compute the hash of the inactive table and store it for
measurements during future state change. If a load is called multiple
times, update the inactive table hash with the hash of the latest
populated table. So that the correct inactive table hash is measured
when the device transitions to different states like resume, remove,
rename, etc.
Signed-off-by: Tushar Sugandhi <tusharsu@linux.microsoft.com>
Signed-off-by: Colin Ian King <colin.king@canonical.com> # leak fix
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
2021-07-13 00:48:58 +00:00
|
|
|
#include "dm-ima.h"
|
2005-04-16 22:20:36 +00:00
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/vmalloc.h>
|
|
|
|
#include <linux/miscdevice.h>
|
2017-02-02 19:43:54 +00:00
|
|
|
#include <linux/sched/mm.h>
|
2005-04-16 22:20:36 +00:00
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/wait.h>
|
|
|
|
#include <linux/slab.h>
|
2021-03-10 10:18:03 +00:00
|
|
|
#include <linux/rbtree.h>
|
2005-04-16 22:20:36 +00:00
|
|
|
#include <linux/dm-ioctl.h>
|
2006-03-27 09:17:54 +00:00
|
|
|
#include <linux/hdreg.h>
|
2008-02-08 02:09:56 +00:00
|
|
|
#include <linux/compat.h>
|
2022-01-29 14:58:39 +00:00
|
|
|
#include <linux/nospec.h>
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2016-12-24 19:46:01 +00:00
|
|
|
#include <linux/uaccess.h>
|
dm ima: measure data on table load
DM configures a block device with various target specific attributes
passed to it as a table. DM loads the table, and calls each target’s
respective constructors with the attributes as input parameters.
Some of these attributes are critical to ensure the device meets
certain security bar. Thus, IMA should measure these attributes, to
ensure they are not tampered with, during the lifetime of the device.
So that the external services can have high confidence in the
configuration of the block-devices on a given system.
Some devices may have large tables. And a given device may change its
state (table-load, suspend, resume, rename, remove, table-clear etc.)
many times. Measuring these attributes each time when the device
changes its state will significantly increase the size of the IMA logs.
Further, once configured, these attributes are not expected to change
unless a new table is loaded, or a device is removed and recreated.
Therefore the clear-text of the attributes should only be measured
during table load, and the hash of the active/inactive table should be
measured for the remaining device state changes.
Export IMA function ima_measure_critical_data() to allow measurement
of DM device parameters, as well as target specific attributes, during
table load. Compute the hash of the inactive table and store it for
measurements during future state change. If a load is called multiple
times, update the inactive table hash with the hash of the latest
populated table. So that the correct inactive table hash is measured
when the device transitions to different states like resume, remove,
rename, etc.
Signed-off-by: Tushar Sugandhi <tusharsu@linux.microsoft.com>
Signed-off-by: Colin Ian King <colin.king@canonical.com> # leak fix
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
2021-07-13 00:48:58 +00:00
|
|
|
#include <linux/ima.h>
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2006-06-26 07:27:35 +00:00
|
|
|
#define DM_MSG_PREFIX "ioctl"
|
2024-02-07 16:42:53 +00:00
|
|
|
#define DM_DRIVER_EMAIL "dm-devel@lists.linux.dev"
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2017-01-16 21:05:59 +00:00
|
|
|
struct dm_file {
|
|
|
|
/*
|
|
|
|
* poll will wait until the global event number is greater than
|
|
|
|
* this value.
|
|
|
|
*/
|
2023-01-25 20:14:58 +00:00
|
|
|
volatile unsigned int global_event_nr;
|
2017-01-16 21:05:59 +00:00
|
|
|
};
|
|
|
|
|
2023-01-26 14:48:30 +00:00
|
|
|
/*
|
|
|
|
*---------------------------------------------------------------
|
2005-04-16 22:20:36 +00:00
|
|
|
* The ioctl interface needs to be able to look up devices by
|
|
|
|
* name or uuid.
|
2023-01-26 14:48:30 +00:00
|
|
|
*---------------------------------------------------------------
|
|
|
|
*/
|
2005-04-16 22:20:36 +00:00
|
|
|
struct hash_cell {
|
2021-03-10 10:18:03 +00:00
|
|
|
struct rb_node name_node;
|
|
|
|
struct rb_node uuid_node;
|
|
|
|
bool name_set;
|
|
|
|
bool uuid_set;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
char *name;
|
|
|
|
char *uuid;
|
|
|
|
struct mapped_device *md;
|
|
|
|
struct dm_table *new_map;
|
|
|
|
};
|
|
|
|
|
|
|
|
struct vers_iter {
|
2023-02-01 21:31:43 +00:00
|
|
|
size_t param_size;
|
|
|
|
struct dm_target_versions *vers, *old_vers;
|
|
|
|
char *end;
|
|
|
|
uint32_t flags;
|
2005-04-16 22:20:36 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
|
2021-03-10 10:18:03 +00:00
|
|
|
static struct rb_root name_rb_tree = RB_ROOT;
|
|
|
|
static struct rb_root uuid_rb_tree = RB_ROOT;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2013-11-01 22:27:41 +00:00
|
|
|
static void dm_hash_remove_all(bool keep_open_devices, bool mark_deferred, bool only_deferred);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Guards access to both hash tables.
|
|
|
|
*/
|
|
|
|
static DECLARE_RWSEM(_hash_lock);
|
|
|
|
|
2009-12-10 23:51:52 +00:00
|
|
|
/*
|
|
|
|
* Protects use of mdptr to obtain hash cell name and uuid from mapped device.
|
|
|
|
*/
|
|
|
|
static DEFINE_MUTEX(dm_hash_cells_mutex);
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
static void dm_hash_exit(void)
|
|
|
|
{
|
2013-11-01 22:27:41 +00:00
|
|
|
dm_hash_remove_all(false, false, false);
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
2023-01-26 14:48:30 +00:00
|
|
|
/*
|
|
|
|
*---------------------------------------------------------------
|
2005-04-16 22:20:36 +00:00
|
|
|
* Code for looking up a device by name
|
2023-01-26 14:48:30 +00:00
|
|
|
*---------------------------------------------------------------
|
|
|
|
*/
|
2005-04-16 22:20:36 +00:00
|
|
|
static struct hash_cell *__get_name_cell(const char *str)
|
|
|
|
{
|
2021-03-10 10:18:03 +00:00
|
|
|
struct rb_node *n = name_rb_tree.rb_node;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2021-03-10 10:18:03 +00:00
|
|
|
while (n) {
|
|
|
|
struct hash_cell *hc = container_of(n, struct hash_cell, name_node);
|
2023-02-01 22:42:29 +00:00
|
|
|
int c;
|
|
|
|
|
|
|
|
c = strcmp(hc->name, str);
|
2021-03-10 10:18:03 +00:00
|
|
|
if (!c) {
|
2006-06-26 07:27:24 +00:00
|
|
|
dm_get(hc->md);
|
2005-04-16 22:20:36 +00:00
|
|
|
return hc;
|
2006-06-26 07:27:24 +00:00
|
|
|
}
|
2021-03-10 10:18:03 +00:00
|
|
|
n = c >= 0 ? n->rb_left : n->rb_right;
|
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct hash_cell *__get_uuid_cell(const char *str)
|
|
|
|
{
|
2021-03-10 10:18:03 +00:00
|
|
|
struct rb_node *n = uuid_rb_tree.rb_node;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2021-03-10 10:18:03 +00:00
|
|
|
while (n) {
|
|
|
|
struct hash_cell *hc = container_of(n, struct hash_cell, uuid_node);
|
2023-02-01 22:42:29 +00:00
|
|
|
int c;
|
|
|
|
|
|
|
|
c = strcmp(hc->uuid, str);
|
2021-03-10 10:18:03 +00:00
|
|
|
if (!c) {
|
2006-06-26 07:27:24 +00:00
|
|
|
dm_get(hc->md);
|
2005-04-16 22:20:36 +00:00
|
|
|
return hc;
|
2006-06-26 07:27:24 +00:00
|
|
|
}
|
2021-03-10 10:18:03 +00:00
|
|
|
n = c >= 0 ? n->rb_left : n->rb_right;
|
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2021-03-10 10:18:03 +00:00
|
|
|
static void __unlink_name(struct hash_cell *hc)
|
|
|
|
{
|
|
|
|
if (hc->name_set) {
|
|
|
|
hc->name_set = false;
|
|
|
|
rb_erase(&hc->name_node, &name_rb_tree);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void __unlink_uuid(struct hash_cell *hc)
|
|
|
|
{
|
|
|
|
if (hc->uuid_set) {
|
|
|
|
hc->uuid_set = false;
|
|
|
|
rb_erase(&hc->uuid_node, &uuid_rb_tree);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void __link_name(struct hash_cell *new_hc)
|
|
|
|
{
|
|
|
|
struct rb_node **n, *parent;
|
|
|
|
|
|
|
|
__unlink_name(new_hc);
|
|
|
|
|
|
|
|
new_hc->name_set = true;
|
|
|
|
|
|
|
|
n = &name_rb_tree.rb_node;
|
|
|
|
parent = NULL;
|
|
|
|
|
|
|
|
while (*n) {
|
|
|
|
struct hash_cell *hc = container_of(*n, struct hash_cell, name_node);
|
2023-02-01 22:42:29 +00:00
|
|
|
int c;
|
|
|
|
|
|
|
|
c = strcmp(hc->name, new_hc->name);
|
2021-03-10 10:18:03 +00:00
|
|
|
BUG_ON(!c);
|
|
|
|
parent = *n;
|
|
|
|
n = c >= 0 ? &hc->name_node.rb_left : &hc->name_node.rb_right;
|
|
|
|
}
|
|
|
|
|
|
|
|
rb_link_node(&new_hc->name_node, parent, n);
|
|
|
|
rb_insert_color(&new_hc->name_node, &name_rb_tree);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void __link_uuid(struct hash_cell *new_hc)
|
|
|
|
{
|
|
|
|
struct rb_node **n, *parent;
|
|
|
|
|
|
|
|
__unlink_uuid(new_hc);
|
|
|
|
|
|
|
|
new_hc->uuid_set = true;
|
|
|
|
|
|
|
|
n = &uuid_rb_tree.rb_node;
|
|
|
|
parent = NULL;
|
|
|
|
|
|
|
|
while (*n) {
|
|
|
|
struct hash_cell *hc = container_of(*n, struct hash_cell, uuid_node);
|
2023-02-01 22:42:29 +00:00
|
|
|
int c;
|
|
|
|
|
|
|
|
c = strcmp(hc->uuid, new_hc->uuid);
|
2021-03-10 10:18:03 +00:00
|
|
|
BUG_ON(!c);
|
|
|
|
parent = *n;
|
|
|
|
n = c > 0 ? &hc->uuid_node.rb_left : &hc->uuid_node.rb_right;
|
|
|
|
}
|
|
|
|
|
|
|
|
rb_link_node(&new_hc->uuid_node, parent, n);
|
|
|
|
rb_insert_color(&new_hc->uuid_node, &uuid_rb_tree);
|
|
|
|
}
|
|
|
|
|
2011-08-02 11:32:06 +00:00
|
|
|
static struct hash_cell *__get_dev_cell(uint64_t dev)
|
|
|
|
{
|
|
|
|
struct mapped_device *md;
|
|
|
|
struct hash_cell *hc;
|
|
|
|
|
|
|
|
md = dm_get_md(huge_decode_dev(dev));
|
|
|
|
if (!md)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
hc = dm_get_mdptr(md);
|
|
|
|
if (!hc) {
|
|
|
|
dm_put(md);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
return hc;
|
|
|
|
}
|
|
|
|
|
2023-01-26 14:48:30 +00:00
|
|
|
/*
|
|
|
|
*---------------------------------------------------------------
|
2005-04-16 22:20:36 +00:00
|
|
|
* Inserting, removing and renaming a device.
|
2023-01-26 14:48:30 +00:00
|
|
|
*---------------------------------------------------------------
|
|
|
|
*/
|
2005-04-16 22:20:36 +00:00
|
|
|
static struct hash_cell *alloc_cell(const char *name, const char *uuid,
|
|
|
|
struct mapped_device *md)
|
|
|
|
{
|
|
|
|
struct hash_cell *hc;
|
|
|
|
|
|
|
|
hc = kmalloc(sizeof(*hc), GFP_KERNEL);
|
|
|
|
if (!hc)
|
|
|
|
return NULL;
|
|
|
|
|
2005-06-23 07:09:02 +00:00
|
|
|
hc->name = kstrdup(name, GFP_KERNEL);
|
2005-04-16 22:20:36 +00:00
|
|
|
if (!hc->name) {
|
|
|
|
kfree(hc);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!uuid)
|
|
|
|
hc->uuid = NULL;
|
|
|
|
|
|
|
|
else {
|
2005-06-23 07:09:02 +00:00
|
|
|
hc->uuid = kstrdup(uuid, GFP_KERNEL);
|
2005-04-16 22:20:36 +00:00
|
|
|
if (!hc->uuid) {
|
|
|
|
kfree(hc->name);
|
|
|
|
kfree(hc);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2021-03-10 10:18:03 +00:00
|
|
|
hc->name_set = hc->uuid_set = false;
|
2005-04-16 22:20:36 +00:00
|
|
|
hc->md = md;
|
|
|
|
hc->new_map = NULL;
|
|
|
|
return hc;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void free_cell(struct hash_cell *hc)
|
|
|
|
{
|
|
|
|
if (hc) {
|
|
|
|
kfree(hc->name);
|
|
|
|
kfree(hc->uuid);
|
|
|
|
kfree(hc);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The kdev_t and uuid of a device can never change once it is
|
|
|
|
* initially inserted.
|
|
|
|
*/
|
|
|
|
static int dm_hash_insert(const char *name, const char *uuid, struct mapped_device *md)
|
|
|
|
{
|
2006-06-26 07:27:24 +00:00
|
|
|
struct hash_cell *cell, *hc;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Allocate the new cells.
|
|
|
|
*/
|
|
|
|
cell = alloc_cell(name, uuid, md);
|
|
|
|
if (!cell)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Insert the cell into both hash tables.
|
|
|
|
*/
|
|
|
|
down_write(&_hash_lock);
|
2006-06-26 07:27:24 +00:00
|
|
|
hc = __get_name_cell(name);
|
|
|
|
if (hc) {
|
|
|
|
dm_put(hc->md);
|
2005-04-16 22:20:36 +00:00
|
|
|
goto bad;
|
2006-06-26 07:27:24 +00:00
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2021-03-10 10:18:03 +00:00
|
|
|
__link_name(cell);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
if (uuid) {
|
2006-06-26 07:27:24 +00:00
|
|
|
hc = __get_uuid_cell(uuid);
|
|
|
|
if (hc) {
|
2021-03-10 10:18:03 +00:00
|
|
|
__unlink_name(cell);
|
2006-06-26 07:27:24 +00:00
|
|
|
dm_put(hc->md);
|
2005-04-16 22:20:36 +00:00
|
|
|
goto bad;
|
|
|
|
}
|
2021-03-10 10:18:03 +00:00
|
|
|
__link_uuid(cell);
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
dm_get(md);
|
2009-12-10 23:51:52 +00:00
|
|
|
mutex_lock(&dm_hash_cells_mutex);
|
2005-04-16 22:20:36 +00:00
|
|
|
dm_set_mdptr(md, cell);
|
2009-12-10 23:51:52 +00:00
|
|
|
mutex_unlock(&dm_hash_cells_mutex);
|
2005-04-16 22:20:36 +00:00
|
|
|
up_write(&_hash_lock);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
bad:
|
|
|
|
up_write(&_hash_lock);
|
|
|
|
free_cell(cell);
|
|
|
|
return -EBUSY;
|
|
|
|
}
|
|
|
|
|
2013-07-10 22:41:18 +00:00
|
|
|
static struct dm_table *__hash_remove(struct hash_cell *hc)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
2005-09-28 04:45:44 +00:00
|
|
|
struct dm_table *table;
|
2013-07-10 22:41:18 +00:00
|
|
|
int srcu_idx;
|
2005-09-28 04:45:44 +00:00
|
|
|
|
2023-02-17 18:08:17 +00:00
|
|
|
lockdep_assert_held(&_hash_lock);
|
|
|
|
|
2021-03-10 10:18:03 +00:00
|
|
|
/* remove from the dev trees */
|
|
|
|
__unlink_name(hc);
|
|
|
|
__unlink_uuid(hc);
|
2009-12-10 23:51:52 +00:00
|
|
|
mutex_lock(&dm_hash_cells_mutex);
|
2005-04-16 22:20:36 +00:00
|
|
|
dm_set_mdptr(hc->md, NULL);
|
2009-12-10 23:51:52 +00:00
|
|
|
mutex_unlock(&dm_hash_cells_mutex);
|
2005-09-28 04:45:44 +00:00
|
|
|
|
2013-07-10 22:41:18 +00:00
|
|
|
table = dm_get_live_table(hc->md, &srcu_idx);
|
|
|
|
if (table)
|
2005-09-28 04:45:44 +00:00
|
|
|
dm_table_event(table);
|
2013-07-10 22:41:18 +00:00
|
|
|
dm_put_live_table(hc->md, srcu_idx);
|
2005-09-28 04:45:44 +00:00
|
|
|
|
2013-07-10 22:41:18 +00:00
|
|
|
table = NULL;
|
2005-04-16 22:20:36 +00:00
|
|
|
if (hc->new_map)
|
2013-07-10 22:41:18 +00:00
|
|
|
table = hc->new_map;
|
2006-03-27 09:17:54 +00:00
|
|
|
dm_put(hc->md);
|
2005-04-16 22:20:36 +00:00
|
|
|
free_cell(hc);
|
2013-07-10 22:41:18 +00:00
|
|
|
|
|
|
|
return table;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
2013-11-01 22:27:41 +00:00
|
|
|
static void dm_hash_remove_all(bool keep_open_devices, bool mark_deferred, bool only_deferred)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
2021-03-10 10:18:03 +00:00
|
|
|
int dev_skipped;
|
|
|
|
struct rb_node *n;
|
2005-04-16 22:20:36 +00:00
|
|
|
struct hash_cell *hc;
|
2010-08-12 03:13:55 +00:00
|
|
|
struct mapped_device *md;
|
2013-07-10 22:41:18 +00:00
|
|
|
struct dm_table *t;
|
2010-08-12 03:13:55 +00:00
|
|
|
|
|
|
|
retry:
|
|
|
|
dev_skipped = 0;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
down_write(&_hash_lock);
|
2006-06-26 07:27:34 +00:00
|
|
|
|
2021-03-10 10:18:03 +00:00
|
|
|
for (n = rb_first(&name_rb_tree); n; n = rb_next(n)) {
|
|
|
|
hc = container_of(n, struct hash_cell, name_node);
|
|
|
|
md = hc->md;
|
|
|
|
dm_get(md);
|
2006-06-26 07:27:34 +00:00
|
|
|
|
2021-03-10 10:18:03 +00:00
|
|
|
if (keep_open_devices &&
|
|
|
|
dm_lock_for_deletion(md, mark_deferred, only_deferred)) {
|
|
|
|
dm_put(md);
|
|
|
|
dev_skipped++;
|
|
|
|
continue;
|
|
|
|
}
|
2010-08-12 03:13:55 +00:00
|
|
|
|
2021-03-10 10:18:03 +00:00
|
|
|
t = __hash_remove(hc);
|
2006-06-26 07:27:34 +00:00
|
|
|
|
2021-03-10 10:18:03 +00:00
|
|
|
up_write(&_hash_lock);
|
2006-06-26 07:27:34 +00:00
|
|
|
|
2021-03-10 10:18:03 +00:00
|
|
|
if (t) {
|
|
|
|
dm_sync_table(md);
|
|
|
|
dm_table_destroy(t);
|
2010-08-12 03:13:55 +00:00
|
|
|
}
|
2021-07-13 00:49:00 +00:00
|
|
|
dm_ima_measure_on_device_remove(md, true);
|
2021-03-10 10:18:03 +00:00
|
|
|
dm_put(md);
|
|
|
|
if (likely(keep_open_devices))
|
|
|
|
dm_destroy(md);
|
|
|
|
else
|
|
|
|
dm_destroy_immediate(md);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Some mapped devices may be using other mapped
|
|
|
|
* devices, so repeat until we make no further
|
|
|
|
* progress. If a new mapped device is created
|
|
|
|
* here it will also get removed.
|
|
|
|
*/
|
|
|
|
goto retry;
|
2006-06-26 07:27:34 +00:00
|
|
|
}
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
up_write(&_hash_lock);
|
2010-08-12 03:13:55 +00:00
|
|
|
|
|
|
|
if (dev_skipped)
|
|
|
|
DMWARN("remove_all left %d open device(s)", dev_skipped);
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
2011-01-13 19:59:47 +00:00
|
|
|
/*
|
|
|
|
* Set the uuid of a hash_cell that isn't already set.
|
|
|
|
*/
|
|
|
|
static void __set_cell_uuid(struct hash_cell *hc, char *new_uuid)
|
|
|
|
{
|
|
|
|
mutex_lock(&dm_hash_cells_mutex);
|
|
|
|
hc->uuid = new_uuid;
|
|
|
|
mutex_unlock(&dm_hash_cells_mutex);
|
|
|
|
|
2021-03-10 10:18:03 +00:00
|
|
|
__link_uuid(hc);
|
2011-01-13 19:59:47 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Changes the name of a hash_cell and returns the old name for
|
|
|
|
* the caller to free.
|
|
|
|
*/
|
|
|
|
static char *__change_cell_name(struct hash_cell *hc, char *new_name)
|
|
|
|
{
|
|
|
|
char *old_name;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Rename and move the name cell.
|
|
|
|
*/
|
2021-03-10 10:18:03 +00:00
|
|
|
__unlink_name(hc);
|
2011-01-13 19:59:47 +00:00
|
|
|
old_name = hc->name;
|
|
|
|
|
|
|
|
mutex_lock(&dm_hash_cells_mutex);
|
|
|
|
hc->name = new_name;
|
|
|
|
mutex_unlock(&dm_hash_cells_mutex);
|
|
|
|
|
2021-03-10 10:18:03 +00:00
|
|
|
__link_name(hc);
|
2011-01-13 19:59:47 +00:00
|
|
|
|
|
|
|
return old_name;
|
|
|
|
}
|
|
|
|
|
2010-08-12 03:13:53 +00:00
|
|
|
static struct mapped_device *dm_hash_rename(struct dm_ioctl *param,
|
|
|
|
const char *new)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
2011-01-13 19:59:47 +00:00
|
|
|
char *new_data, *old_name = NULL;
|
2005-04-16 22:20:36 +00:00
|
|
|
struct hash_cell *hc;
|
2006-01-06 08:20:01 +00:00
|
|
|
struct dm_table *table;
|
2010-08-12 03:13:53 +00:00
|
|
|
struct mapped_device *md;
|
2023-01-25 20:14:58 +00:00
|
|
|
unsigned int change_uuid = (param->flags & DM_UUID_FLAG) ? 1 : 0;
|
2013-07-10 22:41:18 +00:00
|
|
|
int srcu_idx;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* duplicate new.
|
|
|
|
*/
|
2011-01-13 19:59:47 +00:00
|
|
|
new_data = kstrdup(new, GFP_KERNEL);
|
|
|
|
if (!new_data)
|
2010-08-12 03:13:53 +00:00
|
|
|
return ERR_PTR(-ENOMEM);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
down_write(&_hash_lock);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Is new free ?
|
|
|
|
*/
|
2011-01-13 19:59:47 +00:00
|
|
|
if (change_uuid)
|
|
|
|
hc = __get_uuid_cell(new);
|
|
|
|
else
|
|
|
|
hc = __get_name_cell(new);
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
if (hc) {
|
2023-02-03 17:55:47 +00:00
|
|
|
DMERR("Unable to change %s on mapped device %s to one that already exists: %s",
|
2022-08-24 11:25:57 +00:00
|
|
|
change_uuid ? "uuid" : "name",
|
|
|
|
param->name, new);
|
2006-06-26 07:27:24 +00:00
|
|
|
dm_put(hc->md);
|
2005-04-16 22:20:36 +00:00
|
|
|
up_write(&_hash_lock);
|
2011-01-13 19:59:47 +00:00
|
|
|
kfree(new_data);
|
2010-08-12 03:13:53 +00:00
|
|
|
return ERR_PTR(-EBUSY);
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Is there such a device as 'old' ?
|
|
|
|
*/
|
2010-08-12 03:13:53 +00:00
|
|
|
hc = __get_name_cell(param->name);
|
2005-04-16 22:20:36 +00:00
|
|
|
if (!hc) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("Unable to rename non-existent device, %s to %s%s",
|
|
|
|
param->name, change_uuid ? "uuid " : "", new);
|
2005-04-16 22:20:36 +00:00
|
|
|
up_write(&_hash_lock);
|
2011-01-13 19:59:47 +00:00
|
|
|
kfree(new_data);
|
2010-08-12 03:13:53 +00:00
|
|
|
return ERR_PTR(-ENXIO);
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2011-01-13 19:59:47 +00:00
|
|
|
* Does this device already have a uuid?
|
2005-04-16 22:20:36 +00:00
|
|
|
*/
|
2011-01-13 19:59:47 +00:00
|
|
|
if (change_uuid && hc->uuid) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("Unable to change uuid of mapped device %s to %s "
|
|
|
|
"because uuid is already set to %s",
|
|
|
|
param->name, new, hc->uuid);
|
2011-01-13 19:59:47 +00:00
|
|
|
dm_put(hc->md);
|
|
|
|
up_write(&_hash_lock);
|
|
|
|
kfree(new_data);
|
|
|
|
return ERR_PTR(-EINVAL);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (change_uuid)
|
|
|
|
__set_cell_uuid(hc, new_data);
|
|
|
|
else
|
|
|
|
old_name = __change_cell_name(hc, new_data);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2006-01-06 08:20:01 +00:00
|
|
|
/*
|
|
|
|
* Wake up any dm event waiters.
|
|
|
|
*/
|
2013-07-10 22:41:18 +00:00
|
|
|
table = dm_get_live_table(hc->md, &srcu_idx);
|
|
|
|
if (table)
|
2006-01-06 08:20:01 +00:00
|
|
|
dm_table_event(table);
|
2013-07-10 22:41:18 +00:00
|
|
|
dm_put_live_table(hc->md, srcu_idx);
|
2006-01-06 08:20:01 +00:00
|
|
|
|
dm: send just one event on resize, not two
Device mapper sends an uevent when the device is suspended, using the
function set_capacity_and_notify. However, this causes a race condition
with udev.
Udev skips scanning dm devices that are suspended. If we send an uevent
while we are suspended, udev will be racing with device mapper resume
code. If the device mapper resume code wins the race, udev will process
the uevent after the device is resumed and it will properly scan the
device.
However, if udev wins the race, it will receive the uevent, find out that
the dm device is suspended and skip scanning the device. This causes bugs
such as systemd unmounting the device - see
https://bugzilla.redhat.com/show_bug.cgi?id=2158628
This commit fixes this race.
We replace the function set_capacity_and_notify with set_capacity, so that
the uevent is not sent at this point. In do_resume, we detect if the
capacity has changed and we pass a boolean variable need_resize_uevent to
dm_kobject_uevent. dm_kobject_uevent adds "RESIZE=1" to the uevent if
need_resize_uevent is set.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Tested-by: Peter Rajnoha <prajnoha@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Mike Snitzer <snitzer@kernel.org>
2023-02-07 13:33:06 +00:00
|
|
|
if (!dm_kobject_uevent(hc->md, KOBJ_CHANGE, param->event_nr, false))
|
2010-08-12 03:13:53 +00:00
|
|
|
param->flags |= DM_UEVENT_GENERATED_FLAG;
|
2007-12-13 14:15:57 +00:00
|
|
|
|
2010-08-12 03:13:53 +00:00
|
|
|
md = hc->md;
|
2021-07-13 00:49:02 +00:00
|
|
|
|
|
|
|
dm_ima_measure_on_device_rename(md);
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
up_write(&_hash_lock);
|
|
|
|
kfree(old_name);
|
2010-08-12 03:13:53 +00:00
|
|
|
|
|
|
|
return md;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
2013-11-01 22:27:41 +00:00
|
|
|
void dm_deferred_remove(void)
|
|
|
|
{
|
|
|
|
dm_hash_remove_all(true, false, true);
|
|
|
|
}
|
|
|
|
|
2023-01-26 14:48:30 +00:00
|
|
|
/*
|
|
|
|
*---------------------------------------------------------------
|
2005-04-16 22:20:36 +00:00
|
|
|
* Implementation of the ioctl commands
|
2023-01-26 14:48:30 +00:00
|
|
|
*---------------------------------------------------------------
|
|
|
|
*/
|
2005-04-16 22:20:36 +00:00
|
|
|
/*
|
|
|
|
* All the ioctl commands get dispatched to functions with this
|
|
|
|
* prototype.
|
|
|
|
*/
|
2017-05-05 18:12:52 +00:00
|
|
|
typedef int (*ioctl_fn)(struct file *filp, struct dm_ioctl *param, size_t param_size);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2017-05-05 18:12:52 +00:00
|
|
|
static int remove_all(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
2013-11-01 22:27:41 +00:00
|
|
|
dm_hash_remove_all(true, !!(param->flags & DM_DEFERRED_REMOVE), false);
|
2005-04-16 22:20:36 +00:00
|
|
|
param->data_size = 0;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Round up the ptr to an 8-byte boundary.
|
|
|
|
*/
|
|
|
|
#define ALIGN_MASK 7
|
2017-09-20 11:29:49 +00:00
|
|
|
static inline size_t align_val(size_t val)
|
|
|
|
{
|
|
|
|
return (val + ALIGN_MASK) & ~ALIGN_MASK;
|
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
static inline void *align_ptr(void *ptr)
|
|
|
|
{
|
2017-09-20 11:29:49 +00:00
|
|
|
return (void *)align_val((size_t)ptr);
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Retrieves the data payload buffer from an already allocated
|
|
|
|
* struct dm_ioctl.
|
|
|
|
*/
|
|
|
|
static void *get_result_buffer(struct dm_ioctl *param, size_t param_size,
|
|
|
|
size_t *len)
|
|
|
|
{
|
|
|
|
param->data_start = align_ptr(param + 1) - (void *) param;
|
|
|
|
|
|
|
|
if (param->data_start < param_size)
|
|
|
|
*len = param_size - param->data_start;
|
|
|
|
else
|
|
|
|
*len = 0;
|
|
|
|
|
|
|
|
return ((void *) param) + param->data_start;
|
|
|
|
}
|
|
|
|
|
2021-03-11 18:27:29 +00:00
|
|
|
static bool filter_device(struct hash_cell *hc, const char *pfx_name, const char *pfx_uuid)
|
|
|
|
{
|
|
|
|
const char *val;
|
|
|
|
size_t val_len, pfx_len;
|
|
|
|
|
|
|
|
val = hc->name;
|
|
|
|
val_len = strlen(val);
|
|
|
|
pfx_len = strnlen(pfx_name, DM_NAME_LEN);
|
|
|
|
if (pfx_len > val_len)
|
|
|
|
return false;
|
|
|
|
if (memcmp(val, pfx_name, pfx_len))
|
|
|
|
return false;
|
|
|
|
|
|
|
|
val = hc->uuid ? hc->uuid : "";
|
|
|
|
val_len = strlen(val);
|
|
|
|
pfx_len = strnlen(pfx_uuid, DM_UUID_LEN);
|
|
|
|
if (pfx_len > val_len)
|
|
|
|
return false;
|
|
|
|
if (memcmp(val, pfx_uuid, pfx_len))
|
|
|
|
return false;
|
|
|
|
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
2017-05-05 18:12:52 +00:00
|
|
|
static int list_devices(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
2021-03-10 10:18:03 +00:00
|
|
|
struct rb_node *n;
|
2005-04-16 22:20:36 +00:00
|
|
|
struct hash_cell *hc;
|
|
|
|
size_t len, needed = 0;
|
|
|
|
struct gendisk *disk;
|
2017-09-20 11:29:49 +00:00
|
|
|
struct dm_name_list *orig_nl, *nl, *old_nl = NULL;
|
2017-01-16 21:07:01 +00:00
|
|
|
uint32_t *event_nr;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
down_write(&_hash_lock);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Loop through all the devices working out how much
|
|
|
|
* space we need.
|
|
|
|
*/
|
2021-03-10 10:18:03 +00:00
|
|
|
for (n = rb_first(&name_rb_tree); n; n = rb_next(n)) {
|
|
|
|
hc = container_of(n, struct hash_cell, name_node);
|
2021-03-11 18:27:29 +00:00
|
|
|
if (!filter_device(hc, param->name, param->uuid))
|
|
|
|
continue;
|
2021-03-10 10:18:03 +00:00
|
|
|
needed += align_val(offsetof(struct dm_name_list, name) + strlen(hc->name) + 1);
|
2021-03-12 14:07:30 +00:00
|
|
|
needed += align_val(sizeof(uint32_t) * 2);
|
|
|
|
if (param->flags & DM_UUID_FLAG && hc->uuid)
|
|
|
|
needed += align_val(strlen(hc->uuid) + 1);
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Grab our output buffer.
|
|
|
|
*/
|
2017-09-20 11:29:49 +00:00
|
|
|
nl = orig_nl = get_result_buffer(param, param_size, &len);
|
2021-03-26 18:32:32 +00:00
|
|
|
if (len < needed || len < sizeof(nl->dev)) {
|
2005-04-16 22:20:36 +00:00
|
|
|
param->flags |= DM_BUFFER_FULL_FLAG;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
param->data_size = param->data_start + needed;
|
|
|
|
|
|
|
|
nl->dev = 0; /* Flags no data */
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Now loop through filling out the names.
|
|
|
|
*/
|
2021-03-10 10:18:03 +00:00
|
|
|
for (n = rb_first(&name_rb_tree); n; n = rb_next(n)) {
|
2021-03-12 14:07:30 +00:00
|
|
|
void *uuid_ptr;
|
2023-02-01 22:42:29 +00:00
|
|
|
|
2021-03-10 10:18:03 +00:00
|
|
|
hc = container_of(n, struct hash_cell, name_node);
|
2021-03-11 18:27:29 +00:00
|
|
|
if (!filter_device(hc, param->name, param->uuid))
|
|
|
|
continue;
|
2021-03-10 10:18:03 +00:00
|
|
|
if (old_nl)
|
|
|
|
old_nl->next = (uint32_t) ((void *) nl -
|
|
|
|
(void *) old_nl);
|
|
|
|
disk = dm_disk(hc->md);
|
|
|
|
nl->dev = huge_encode_dev(disk_devt(disk));
|
|
|
|
nl->next = 0;
|
|
|
|
strcpy(nl->name, hc->name);
|
|
|
|
|
|
|
|
old_nl = nl;
|
|
|
|
event_nr = align_ptr(nl->name + strlen(hc->name) + 1);
|
2021-03-12 14:07:30 +00:00
|
|
|
event_nr[0] = dm_get_event_nr(hc->md);
|
|
|
|
event_nr[1] = 0;
|
|
|
|
uuid_ptr = align_ptr(event_nr + 2);
|
|
|
|
if (param->flags & DM_UUID_FLAG) {
|
|
|
|
if (hc->uuid) {
|
|
|
|
event_nr[1] |= DM_NAME_LIST_FLAG_HAS_UUID;
|
|
|
|
strcpy(uuid_ptr, hc->uuid);
|
|
|
|
uuid_ptr = align_ptr(uuid_ptr + strlen(hc->uuid) + 1);
|
|
|
|
} else {
|
|
|
|
event_nr[1] |= DM_NAME_LIST_FLAG_DOESNT_HAVE_UUID;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
nl = uuid_ptr;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
2017-09-20 11:29:49 +00:00
|
|
|
/*
|
|
|
|
* If mismatch happens, security may be compromised due to buffer
|
|
|
|
* overflow, so it's better to crash.
|
|
|
|
*/
|
|
|
|
BUG_ON((char *)nl - (char *)orig_nl != needed);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
out:
|
|
|
|
up_write(&_hash_lock);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void list_version_get_needed(struct target_type *tt, void *needed_param)
|
|
|
|
{
|
2023-02-01 21:31:43 +00:00
|
|
|
size_t *needed = needed_param;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2023-02-01 21:31:43 +00:00
|
|
|
*needed += sizeof(struct dm_target_versions);
|
|
|
|
*needed += strlen(tt->name) + 1;
|
|
|
|
*needed += ALIGN_MASK;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static void list_version_get_info(struct target_type *tt, void *param)
|
|
|
|
{
|
2023-02-01 21:31:43 +00:00
|
|
|
struct vers_iter *info = param;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2023-02-01 21:31:43 +00:00
|
|
|
/* Check space - it might have changed since the first iteration */
|
|
|
|
if ((char *)info->vers + sizeof(tt->version) + strlen(tt->name) + 1 > info->end) {
|
|
|
|
info->flags = DM_BUFFER_FULL_FLAG;
|
|
|
|
return;
|
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2023-02-01 21:31:43 +00:00
|
|
|
if (info->old_vers)
|
|
|
|
info->old_vers->next = (uint32_t) ((void *)info->vers - (void *)info->old_vers);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2023-02-01 21:31:43 +00:00
|
|
|
info->vers->version[0] = tt->version[0];
|
|
|
|
info->vers->version[1] = tt->version[1];
|
|
|
|
info->vers->version[2] = tt->version[2];
|
|
|
|
info->vers->next = 0;
|
|
|
|
strcpy(info->vers->name, tt->name);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2023-02-01 21:31:43 +00:00
|
|
|
info->old_vers = info->vers;
|
|
|
|
info->vers = align_ptr((void *)(info->vers + 1) + strlen(tt->name) + 1);
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
2019-09-16 09:55:42 +00:00
|
|
|
static int __list_versions(struct dm_ioctl *param, size_t param_size, const char *name)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
size_t len, needed = 0;
|
|
|
|
struct dm_target_versions *vers;
|
|
|
|
struct vers_iter iter_info;
|
2019-09-16 09:55:42 +00:00
|
|
|
struct target_type *tt = NULL;
|
|
|
|
|
|
|
|
if (name) {
|
|
|
|
tt = dm_get_target_type(name);
|
|
|
|
if (!tt)
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Loop through all the devices working out how much
|
|
|
|
* space we need.
|
|
|
|
*/
|
2019-09-16 09:55:42 +00:00
|
|
|
if (!tt)
|
|
|
|
dm_target_iterate(list_version_get_needed, &needed);
|
|
|
|
else
|
|
|
|
list_version_get_needed(tt, &needed);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Grab our output buffer.
|
|
|
|
*/
|
|
|
|
vers = get_result_buffer(param, param_size, &len);
|
|
|
|
if (len < needed) {
|
|
|
|
param->flags |= DM_BUFFER_FULL_FLAG;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
param->data_size = param->data_start + needed;
|
|
|
|
|
|
|
|
iter_info.param_size = param_size;
|
|
|
|
iter_info.old_vers = NULL;
|
|
|
|
iter_info.vers = vers;
|
|
|
|
iter_info.flags = 0;
|
dm ioctl: fix misbehavior if list_versions races with module loading
__list_versions will first estimate the required space using the
"dm_target_iterate(list_version_get_needed, &needed)" call and then will
fill the space using the "dm_target_iterate(list_version_get_info,
&iter_info)" call. Each of these calls locks the targets using the
"down_read(&_lock)" and "up_read(&_lock)" calls, however between the first
and second "dm_target_iterate" there is no lock held and the target
modules can be loaded at this point, so the second "dm_target_iterate"
call may need more space than what was the first "dm_target_iterate"
returned.
The code tries to handle this overflow (see the beginning of
list_version_get_info), however this handling is incorrect.
The code sets "param->data_size = param->data_start + needed" and
"iter_info.end = (char *)vers+len" - "needed" is the size returned by the
first dm_target_iterate call; "len" is the size of the buffer allocated by
userspace.
"len" may be greater than "needed"; in this case, the code will write up
to "len" bytes into the buffer, however param->data_size is set to
"needed", so it may write data past the param->data_size value. The ioctl
interface copies only up to param->data_size into userspace, thus part of
the result will be truncated.
Fix this bug by setting "iter_info.end = (char *)vers + needed;" - this
guarantees that the second "dm_target_iterate" call will write only up to
the "needed" buffer and it will exit with "DM_BUFFER_FULL_FLAG" if it
overflows the "needed" space - in this case, userspace will allocate a
larger buffer and retry.
Note that there is also a bug in list_version_get_needed - we need to add
"strlen(tt->name) + 1" to the needed size, not "strlen(tt->name)".
Cc: stable@vger.kernel.org
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@kernel.org>
2022-11-01 20:53:35 +00:00
|
|
|
iter_info.end = (char *)vers + needed;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Now loop through filling out the names & versions.
|
|
|
|
*/
|
2019-09-16 09:55:42 +00:00
|
|
|
if (!tt)
|
|
|
|
dm_target_iterate(list_version_get_info, &iter_info);
|
|
|
|
else
|
|
|
|
list_version_get_info(tt, &iter_info);
|
2005-04-16 22:20:36 +00:00
|
|
|
param->flags |= iter_info.flags;
|
|
|
|
|
|
|
|
out:
|
2019-09-16 09:55:42 +00:00
|
|
|
if (tt)
|
|
|
|
dm_put_target_type(tt);
|
2005-04-16 22:20:36 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2019-09-16 09:55:42 +00:00
|
|
|
static int list_versions(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
|
|
|
{
|
|
|
|
return __list_versions(param, param_size, NULL);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int get_target_version(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
|
|
|
{
|
|
|
|
return __list_versions(param, param_size, param->name);
|
|
|
|
}
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
static int check_name(const char *name)
|
|
|
|
{
|
|
|
|
if (strchr(name, '/')) {
|
2023-06-03 14:52:43 +00:00
|
|
|
DMERR("device name cannot contain '/'");
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
2023-06-03 14:52:44 +00:00
|
|
|
if (strcmp(name, DM_CONTROL_NODE) == 0 ||
|
|
|
|
strcmp(name, ".") == 0 ||
|
|
|
|
strcmp(name, "..") == 0) {
|
|
|
|
DMERR("device name cannot be \"%s\", \".\", or \"..\"", DM_CONTROL_NODE);
|
2005-04-16 22:20:36 +00:00
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-12-10 23:52:22 +00:00
|
|
|
/*
|
|
|
|
* On successful return, the caller must not attempt to acquire
|
2014-12-24 06:48:12 +00:00
|
|
|
* _hash_lock without first calling dm_put_live_table, because dm_table_destroy
|
|
|
|
* waits for this dm_put_live_table and could be called under this lock.
|
2009-12-10 23:52:22 +00:00
|
|
|
*/
|
2013-07-10 22:41:18 +00:00
|
|
|
static struct dm_table *dm_get_inactive_table(struct mapped_device *md, int *srcu_idx)
|
2009-12-10 23:52:22 +00:00
|
|
|
{
|
|
|
|
struct hash_cell *hc;
|
|
|
|
struct dm_table *table = NULL;
|
|
|
|
|
2013-07-10 22:41:18 +00:00
|
|
|
/* increment rcu count, we don't care about the table pointer */
|
|
|
|
dm_get_live_table(md, srcu_idx);
|
|
|
|
|
2009-12-10 23:52:22 +00:00
|
|
|
down_read(&_hash_lock);
|
|
|
|
hc = dm_get_mdptr(md);
|
2022-12-16 04:23:53 +00:00
|
|
|
if (!hc) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("device has been removed from the dev hash table.");
|
2009-12-10 23:52:22 +00:00
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
table = hc->new_map;
|
|
|
|
|
|
|
|
out:
|
|
|
|
up_read(&_hash_lock);
|
|
|
|
|
|
|
|
return table;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct dm_table *dm_get_live_or_inactive_table(struct mapped_device *md,
|
2013-07-10 22:41:18 +00:00
|
|
|
struct dm_ioctl *param,
|
|
|
|
int *srcu_idx)
|
2009-12-10 23:52:22 +00:00
|
|
|
{
|
|
|
|
return (param->flags & DM_QUERY_INACTIVE_TABLE_FLAG) ?
|
2013-07-10 22:41:18 +00:00
|
|
|
dm_get_inactive_table(md, srcu_idx) : dm_get_live_table(md, srcu_idx);
|
2009-12-10 23:52:22 +00:00
|
|
|
}
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
/*
|
|
|
|
* Fills in a dm_ioctl structure, ready for sending back to
|
|
|
|
* userland.
|
|
|
|
*/
|
2010-08-12 03:13:52 +00:00
|
|
|
static void __dev_status(struct mapped_device *md, struct dm_ioctl *param)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
struct gendisk *disk = dm_disk(md);
|
|
|
|
struct dm_table *table;
|
2013-07-10 22:41:18 +00:00
|
|
|
int srcu_idx;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
param->flags &= ~(DM_SUSPEND_FLAG | DM_READONLY_FLAG |
|
2014-10-28 22:34:52 +00:00
|
|
|
DM_ACTIVE_PRESENT_FLAG | DM_INTERNAL_SUSPEND_FLAG);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2009-12-10 23:52:26 +00:00
|
|
|
if (dm_suspended_md(md))
|
2005-04-16 22:20:36 +00:00
|
|
|
param->flags |= DM_SUSPEND_FLAG;
|
|
|
|
|
2014-10-28 22:34:52 +00:00
|
|
|
if (dm_suspended_internally_md(md))
|
|
|
|
param->flags |= DM_INTERNAL_SUSPEND_FLAG;
|
|
|
|
|
2013-11-01 22:27:41 +00:00
|
|
|
if (dm_test_deferred_remove_flag(md))
|
|
|
|
param->flags |= DM_DEFERRED_REMOVE;
|
|
|
|
|
2008-09-03 07:01:48 +00:00
|
|
|
param->dev = huge_encode_dev(disk_devt(disk));
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2006-06-26 07:27:34 +00:00
|
|
|
/*
|
|
|
|
* Yes, this will be out of date by the time it gets back
|
|
|
|
* to userland, but it is still very useful for
|
|
|
|
* debugging.
|
|
|
|
*/
|
|
|
|
param->open_count = dm_open_count(md);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
param->event_nr = dm_get_event_nr(md);
|
2009-12-10 23:52:22 +00:00
|
|
|
param->target_count = 0;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2013-07-10 22:41:18 +00:00
|
|
|
table = dm_get_live_table(md, &srcu_idx);
|
2005-04-16 22:20:36 +00:00
|
|
|
if (table) {
|
2009-12-10 23:52:22 +00:00
|
|
|
if (!(param->flags & DM_QUERY_INACTIVE_TABLE_FLAG)) {
|
|
|
|
if (get_disk_ro(disk))
|
|
|
|
param->flags |= DM_READONLY_FLAG;
|
2022-07-05 18:00:36 +00:00
|
|
|
param->target_count = table->num_targets;
|
2009-12-10 23:52:22 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
param->flags |= DM_ACTIVE_PRESENT_FLAG;
|
|
|
|
}
|
2013-07-10 22:41:18 +00:00
|
|
|
dm_put_live_table(md, srcu_idx);
|
2009-12-10 23:52:22 +00:00
|
|
|
|
|
|
|
if (param->flags & DM_QUERY_INACTIVE_TABLE_FLAG) {
|
2013-07-10 22:41:18 +00:00
|
|
|
int srcu_idx;
|
2023-02-01 22:42:29 +00:00
|
|
|
|
2013-07-10 22:41:18 +00:00
|
|
|
table = dm_get_inactive_table(md, &srcu_idx);
|
2009-12-10 23:52:22 +00:00
|
|
|
if (table) {
|
2023-06-08 11:02:55 +00:00
|
|
|
if (!(dm_table_get_mode(table) & BLK_OPEN_WRITE))
|
2009-12-10 23:52:22 +00:00
|
|
|
param->flags |= DM_READONLY_FLAG;
|
2022-07-05 18:00:36 +00:00
|
|
|
param->target_count = table->num_targets;
|
2009-12-10 23:52:22 +00:00
|
|
|
}
|
2013-07-10 22:41:18 +00:00
|
|
|
dm_put_live_table(md, srcu_idx);
|
2009-12-10 23:52:22 +00:00
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
2017-05-05 18:12:52 +00:00
|
|
|
static int dev_create(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
2006-06-26 07:27:32 +00:00
|
|
|
int r, m = DM_ANY_MINOR;
|
2005-04-16 22:20:36 +00:00
|
|
|
struct mapped_device *md;
|
|
|
|
|
|
|
|
r = check_name(param->name);
|
|
|
|
if (r)
|
|
|
|
return r;
|
|
|
|
|
|
|
|
if (param->flags & DM_PERSISTENT_DEV_FLAG)
|
2006-06-26 07:27:32 +00:00
|
|
|
m = MINOR(huge_decode_dev(param->dev));
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2006-06-26 07:27:32 +00:00
|
|
|
r = dm_create(m, &md);
|
2005-04-16 22:20:36 +00:00
|
|
|
if (r)
|
|
|
|
return r;
|
|
|
|
|
|
|
|
r = dm_hash_insert(param->name, *param->uuid ? param->uuid : NULL, md);
|
dm: separate device deletion from dm_put
This patch separates the device deletion code from dm_put()
to make sure the deletion happens in the process context.
By this patch, device deletion always occurs in an ioctl (process)
context and dm_put() can be called in interrupt context.
As a result, the request-based dm's bad dm_put() usage pointed out
by Mikulas below disappears.
http://marc.info/?l=dm-devel&m=126699981019735&w=2
Without this patch, I confirmed there is a case to crash the system:
dm_put() => dm_table_destroy() => vfree() => BUG_ON(in_interrupt())
Some more backgrounds and details:
In request-based dm, a device opener can remove a mapped_device
while the last request is still completing, because bios in the last
request complete first and then the device opener can close and remove
the mapped_device before the last request completes:
CPU0 CPU1
=================================================================
<<INTERRUPT>>
blk_end_request_all(clone_rq)
blk_update_request(clone_rq)
bio_endio(clone_bio) == end_clone_bio
blk_update_request(orig_rq)
bio_endio(orig_bio)
<<I/O completed>>
dm_blk_close()
dev_remove()
dm_put(md)
<<Free md>>
blk_finish_request(clone_rq)
....
dm_end_request(clone_rq)
free_rq_clone(clone_rq)
blk_end_request_all(orig_rq)
rq_completed(md)
So request-based dm used dm_get()/dm_put() to hold md for each I/O
until its request completion handling is fully done.
However, the final dm_put() can call the device deletion code which
must not be run in interrupt context and may cause kernel panic.
To solve the problem, this patch moves the device deletion code,
dm_destroy(), to predetermined places that is actually deleting
the mapped_device in ioctl (process) context, and changes dm_put()
just to decrement the reference count of the mapped_device.
By this change, dm_put() can be used in any context and the symmetric
model below is introduced:
dm_create(): create a mapped_device
dm_destroy(): destroy a mapped_device
dm_get(): increment the reference count of a mapped_device
dm_put(): decrement the reference count of a mapped_device
dm_destroy() waits for all references of the mapped_device to disappear,
then deletes the mapped_device.
dm_destroy() uses active waiting with msleep(1), since deleting
the mapped_device isn't performance-critical task.
And since at this point, nobody opens the mapped_device and no new
reference will be taken, the pending counts are just for racing
completing activity and will eventually decrease to zero.
For the unlikely case of the forced module unload, dm_destroy_immediate(),
which doesn't wait and forcibly deletes the mapped_device, is also
introduced and used in dm_hash_remove_all(). Otherwise, "rmmod -f"
may be stuck and never return.
And now, because the mapped_device is deleted at this point, subsequent
accesses to the mapped_device may cause NULL pointer references.
Cc: stable@kernel.org
Signed-off-by: Kiyoshi Ueda <k-ueda@ct.jp.nec.com>
Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2010-08-12 03:13:56 +00:00
|
|
|
if (r) {
|
|
|
|
dm_put(md);
|
|
|
|
dm_destroy(md);
|
|
|
|
return r;
|
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
|
|
|
|
|
2010-08-12 03:13:52 +00:00
|
|
|
__dev_status(md, param);
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
dm_put(md);
|
|
|
|
|
dm: separate device deletion from dm_put
This patch separates the device deletion code from dm_put()
to make sure the deletion happens in the process context.
By this patch, device deletion always occurs in an ioctl (process)
context and dm_put() can be called in interrupt context.
As a result, the request-based dm's bad dm_put() usage pointed out
by Mikulas below disappears.
http://marc.info/?l=dm-devel&m=126699981019735&w=2
Without this patch, I confirmed there is a case to crash the system:
dm_put() => dm_table_destroy() => vfree() => BUG_ON(in_interrupt())
Some more backgrounds and details:
In request-based dm, a device opener can remove a mapped_device
while the last request is still completing, because bios in the last
request complete first and then the device opener can close and remove
the mapped_device before the last request completes:
CPU0 CPU1
=================================================================
<<INTERRUPT>>
blk_end_request_all(clone_rq)
blk_update_request(clone_rq)
bio_endio(clone_bio) == end_clone_bio
blk_update_request(orig_rq)
bio_endio(orig_bio)
<<I/O completed>>
dm_blk_close()
dev_remove()
dm_put(md)
<<Free md>>
blk_finish_request(clone_rq)
....
dm_end_request(clone_rq)
free_rq_clone(clone_rq)
blk_end_request_all(orig_rq)
rq_completed(md)
So request-based dm used dm_get()/dm_put() to hold md for each I/O
until its request completion handling is fully done.
However, the final dm_put() can call the device deletion code which
must not be run in interrupt context and may cause kernel panic.
To solve the problem, this patch moves the device deletion code,
dm_destroy(), to predetermined places that is actually deleting
the mapped_device in ioctl (process) context, and changes dm_put()
just to decrement the reference count of the mapped_device.
By this change, dm_put() can be used in any context and the symmetric
model below is introduced:
dm_create(): create a mapped_device
dm_destroy(): destroy a mapped_device
dm_get(): increment the reference count of a mapped_device
dm_put(): decrement the reference count of a mapped_device
dm_destroy() waits for all references of the mapped_device to disappear,
then deletes the mapped_device.
dm_destroy() uses active waiting with msleep(1), since deleting
the mapped_device isn't performance-critical task.
And since at this point, nobody opens the mapped_device and no new
reference will be taken, the pending counts are just for racing
completing activity and will eventually decrease to zero.
For the unlikely case of the forced module unload, dm_destroy_immediate(),
which doesn't wait and forcibly deletes the mapped_device, is also
introduced and used in dm_hash_remove_all(). Otherwise, "rmmod -f"
may be stuck and never return.
And now, because the mapped_device is deleted at this point, subsequent
accesses to the mapped_device may cause NULL pointer references.
Cc: stable@kernel.org
Signed-off-by: Kiyoshi Ueda <k-ueda@ct.jp.nec.com>
Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2010-08-12 03:13:56 +00:00
|
|
|
return 0;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Always use UUID for lookups if it's present, otherwise use name or dev.
|
|
|
|
*/
|
2006-01-14 21:20:43 +00:00
|
|
|
static struct hash_cell *__find_device_hash_cell(struct dm_ioctl *param)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
2011-08-02 11:32:06 +00:00
|
|
|
struct hash_cell *hc = NULL;
|
2006-03-27 09:17:53 +00:00
|
|
|
|
2011-08-02 11:32:06 +00:00
|
|
|
if (*param->uuid) {
|
2022-03-20 21:12:46 +00:00
|
|
|
if (*param->name || param->dev) {
|
|
|
|
DMERR("Invalid ioctl structure: uuid %s, name %s, dev %llx",
|
|
|
|
param->uuid, param->name, (unsigned long long)param->dev);
|
2011-08-02 11:32:06 +00:00
|
|
|
return NULL;
|
2022-03-20 21:12:46 +00:00
|
|
|
}
|
2011-08-02 11:32:06 +00:00
|
|
|
|
2011-08-02 11:32:06 +00:00
|
|
|
hc = __get_uuid_cell(param->uuid);
|
|
|
|
if (!hc)
|
|
|
|
return NULL;
|
2011-08-02 11:32:06 +00:00
|
|
|
} else if (*param->name) {
|
2022-03-20 21:12:46 +00:00
|
|
|
if (param->dev) {
|
|
|
|
DMERR("Invalid ioctl structure: name %s, dev %llx",
|
|
|
|
param->name, (unsigned long long)param->dev);
|
2011-08-02 11:32:06 +00:00
|
|
|
return NULL;
|
2022-03-20 21:12:46 +00:00
|
|
|
}
|
2011-08-02 11:32:06 +00:00
|
|
|
|
2011-08-02 11:32:06 +00:00
|
|
|
hc = __get_name_cell(param->name);
|
|
|
|
if (!hc)
|
|
|
|
return NULL;
|
2011-08-02 11:32:06 +00:00
|
|
|
} else if (param->dev) {
|
|
|
|
hc = __get_dev_cell(param->dev);
|
|
|
|
if (!hc)
|
|
|
|
return NULL;
|
|
|
|
} else
|
2011-08-02 11:32:06 +00:00
|
|
|
return NULL;
|
2006-03-27 09:17:53 +00:00
|
|
|
|
2011-08-02 11:32:06 +00:00
|
|
|
/*
|
|
|
|
* Sneakily write in both the name and the uuid
|
|
|
|
* while we have the cell.
|
|
|
|
*/
|
2023-02-07 22:04:01 +00:00
|
|
|
strscpy(param->name, hc->name, sizeof(param->name));
|
2011-08-02 11:32:06 +00:00
|
|
|
if (hc->uuid)
|
2023-02-07 22:04:01 +00:00
|
|
|
strscpy(param->uuid, hc->uuid, sizeof(param->uuid));
|
2011-08-02 11:32:06 +00:00
|
|
|
else
|
|
|
|
param->uuid[0] = '\0';
|
|
|
|
|
|
|
|
if (hc->new_map)
|
|
|
|
param->flags |= DM_INACTIVE_PRESENT_FLAG;
|
|
|
|
else
|
|
|
|
param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
|
|
|
|
|
|
|
|
return hc;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
2006-01-14 21:20:43 +00:00
|
|
|
static struct mapped_device *find_device(struct dm_ioctl *param)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
struct hash_cell *hc;
|
|
|
|
struct mapped_device *md = NULL;
|
|
|
|
|
|
|
|
down_read(&_hash_lock);
|
|
|
|
hc = __find_device_hash_cell(param);
|
2011-08-02 11:32:06 +00:00
|
|
|
if (hc)
|
2005-04-16 22:20:36 +00:00
|
|
|
md = hc->md;
|
|
|
|
up_read(&_hash_lock);
|
|
|
|
|
|
|
|
return md;
|
|
|
|
}
|
|
|
|
|
2017-05-05 18:12:52 +00:00
|
|
|
static int dev_remove(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
struct hash_cell *hc;
|
2006-06-26 07:27:24 +00:00
|
|
|
struct mapped_device *md;
|
2006-06-26 07:27:34 +00:00
|
|
|
int r;
|
2013-07-10 22:41:18 +00:00
|
|
|
struct dm_table *t;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
down_write(&_hash_lock);
|
|
|
|
hc = __find_device_hash_cell(param);
|
|
|
|
|
|
|
|
if (!hc) {
|
2011-01-13 19:59:55 +00:00
|
|
|
DMDEBUG_LIMIT("device doesn't appear to be in the dev hash table.");
|
2005-04-16 22:20:36 +00:00
|
|
|
up_write(&_hash_lock);
|
|
|
|
return -ENXIO;
|
|
|
|
}
|
|
|
|
|
2006-06-26 07:27:24 +00:00
|
|
|
md = hc->md;
|
|
|
|
|
2006-06-26 07:27:34 +00:00
|
|
|
/*
|
|
|
|
* Ensure the device is not open and nothing further can open it.
|
|
|
|
*/
|
2013-11-01 22:27:41 +00:00
|
|
|
r = dm_lock_for_deletion(md, !!(param->flags & DM_DEFERRED_REMOVE), false);
|
2006-06-26 07:27:34 +00:00
|
|
|
if (r) {
|
2013-11-01 22:27:41 +00:00
|
|
|
if (r == -EBUSY && param->flags & DM_DEFERRED_REMOVE) {
|
|
|
|
up_write(&_hash_lock);
|
|
|
|
dm_put(md);
|
|
|
|
return 0;
|
|
|
|
}
|
2011-01-13 19:59:55 +00:00
|
|
|
DMDEBUG_LIMIT("unable to remove open device %s", hc->name);
|
2006-06-26 07:27:34 +00:00
|
|
|
up_write(&_hash_lock);
|
|
|
|
dm_put(md);
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2013-07-10 22:41:18 +00:00
|
|
|
t = __hash_remove(hc);
|
2005-04-16 22:20:36 +00:00
|
|
|
up_write(&_hash_lock);
|
2009-06-22 09:12:30 +00:00
|
|
|
|
2013-07-10 22:41:18 +00:00
|
|
|
if (t) {
|
|
|
|
dm_sync_table(md);
|
|
|
|
dm_table_destroy(t);
|
|
|
|
}
|
|
|
|
|
2013-11-01 22:27:41 +00:00
|
|
|
param->flags &= ~DM_DEFERRED_REMOVE;
|
|
|
|
|
2021-07-13 00:49:00 +00:00
|
|
|
dm_ima_measure_on_device_remove(md, false);
|
|
|
|
|
dm: send just one event on resize, not two
Device mapper sends an uevent when the device is suspended, using the
function set_capacity_and_notify. However, this causes a race condition
with udev.
Udev skips scanning dm devices that are suspended. If we send an uevent
while we are suspended, udev will be racing with device mapper resume
code. If the device mapper resume code wins the race, udev will process
the uevent after the device is resumed and it will properly scan the
device.
However, if udev wins the race, it will receive the uevent, find out that
the dm device is suspended and skip scanning the device. This causes bugs
such as systemd unmounting the device - see
https://bugzilla.redhat.com/show_bug.cgi?id=2158628
This commit fixes this race.
We replace the function set_capacity_and_notify with set_capacity, so that
the uevent is not sent at this point. In do_resume, we detect if the
capacity has changed and we pass a boolean variable need_resize_uevent to
dm_kobject_uevent. dm_kobject_uevent adds "RESIZE=1" to the uevent if
need_resize_uevent is set.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Tested-by: Peter Rajnoha <prajnoha@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Mike Snitzer <snitzer@kernel.org>
2023-02-07 13:33:06 +00:00
|
|
|
if (!dm_kobject_uevent(md, KOBJ_REMOVE, param->event_nr, false))
|
2010-03-06 02:32:31 +00:00
|
|
|
param->flags |= DM_UEVENT_GENERATED_FLAG;
|
2009-06-22 09:12:30 +00:00
|
|
|
|
2006-06-26 07:27:24 +00:00
|
|
|
dm_put(md);
|
dm: separate device deletion from dm_put
This patch separates the device deletion code from dm_put()
to make sure the deletion happens in the process context.
By this patch, device deletion always occurs in an ioctl (process)
context and dm_put() can be called in interrupt context.
As a result, the request-based dm's bad dm_put() usage pointed out
by Mikulas below disappears.
http://marc.info/?l=dm-devel&m=126699981019735&w=2
Without this patch, I confirmed there is a case to crash the system:
dm_put() => dm_table_destroy() => vfree() => BUG_ON(in_interrupt())
Some more backgrounds and details:
In request-based dm, a device opener can remove a mapped_device
while the last request is still completing, because bios in the last
request complete first and then the device opener can close and remove
the mapped_device before the last request completes:
CPU0 CPU1
=================================================================
<<INTERRUPT>>
blk_end_request_all(clone_rq)
blk_update_request(clone_rq)
bio_endio(clone_bio) == end_clone_bio
blk_update_request(orig_rq)
bio_endio(orig_bio)
<<I/O completed>>
dm_blk_close()
dev_remove()
dm_put(md)
<<Free md>>
blk_finish_request(clone_rq)
....
dm_end_request(clone_rq)
free_rq_clone(clone_rq)
blk_end_request_all(orig_rq)
rq_completed(md)
So request-based dm used dm_get()/dm_put() to hold md for each I/O
until its request completion handling is fully done.
However, the final dm_put() can call the device deletion code which
must not be run in interrupt context and may cause kernel panic.
To solve the problem, this patch moves the device deletion code,
dm_destroy(), to predetermined places that is actually deleting
the mapped_device in ioctl (process) context, and changes dm_put()
just to decrement the reference count of the mapped_device.
By this change, dm_put() can be used in any context and the symmetric
model below is introduced:
dm_create(): create a mapped_device
dm_destroy(): destroy a mapped_device
dm_get(): increment the reference count of a mapped_device
dm_put(): decrement the reference count of a mapped_device
dm_destroy() waits for all references of the mapped_device to disappear,
then deletes the mapped_device.
dm_destroy() uses active waiting with msleep(1), since deleting
the mapped_device isn't performance-critical task.
And since at this point, nobody opens the mapped_device and no new
reference will be taken, the pending counts are just for racing
completing activity and will eventually decrease to zero.
For the unlikely case of the forced module unload, dm_destroy_immediate(),
which doesn't wait and forcibly deletes the mapped_device, is also
introduced and used in dm_hash_remove_all(). Otherwise, "rmmod -f"
may be stuck and never return.
And now, because the mapped_device is deleted at this point, subsequent
accesses to the mapped_device may cause NULL pointer references.
Cc: stable@kernel.org
Signed-off-by: Kiyoshi Ueda <k-ueda@ct.jp.nec.com>
Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2010-08-12 03:13:56 +00:00
|
|
|
dm_destroy(md);
|
2005-04-16 22:20:36 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Check a string doesn't overrun the chunk of
|
|
|
|
* memory we copied from userland.
|
|
|
|
*/
|
|
|
|
static int invalid_str(char *str, void *end)
|
|
|
|
{
|
|
|
|
while ((void *) str < end)
|
|
|
|
if (!*str++)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
2017-05-05 18:12:52 +00:00
|
|
|
static int dev_rename(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
int r;
|
2011-01-13 19:59:47 +00:00
|
|
|
char *new_data = (char *) param + param->data_start;
|
2010-08-12 03:13:53 +00:00
|
|
|
struct mapped_device *md;
|
2023-01-25 20:14:58 +00:00
|
|
|
unsigned int change_uuid = (param->flags & DM_UUID_FLAG) ? 1 : 0;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2011-01-13 19:59:47 +00:00
|
|
|
if (new_data < param->data ||
|
2013-08-29 15:37:45 +00:00
|
|
|
invalid_str(new_data, (void *) param + param_size) || !*new_data ||
|
2011-01-13 19:59:47 +00:00
|
|
|
strlen(new_data) > (change_uuid ? DM_UUID_LEN - 1 : DM_NAME_LEN - 1)) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("Invalid new mapped device name or uuid string supplied.");
|
2005-04-16 22:20:36 +00:00
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
2011-01-13 19:59:47 +00:00
|
|
|
if (!change_uuid) {
|
|
|
|
r = check_name(new_data);
|
|
|
|
if (r)
|
|
|
|
return r;
|
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2011-01-13 19:59:47 +00:00
|
|
|
md = dm_hash_rename(param, new_data);
|
2010-08-12 03:13:53 +00:00
|
|
|
if (IS_ERR(md))
|
|
|
|
return PTR_ERR(md);
|
2010-03-06 02:32:31 +00:00
|
|
|
|
2010-08-12 03:13:53 +00:00
|
|
|
__dev_status(md, param);
|
|
|
|
dm_put(md);
|
|
|
|
|
|
|
|
return 0;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
2017-05-05 18:12:52 +00:00
|
|
|
static int dev_set_geometry(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
2006-03-27 09:17:54 +00:00
|
|
|
{
|
|
|
|
int r = -EINVAL, x;
|
|
|
|
struct mapped_device *md;
|
|
|
|
struct hd_geometry geometry;
|
|
|
|
unsigned long indata[4];
|
|
|
|
char *geostr = (char *) param + param->data_start;
|
dm: reject trailing characters in sccanf input
Device mapper uses sscanf to convert arguments to numbers. The problem is that
the way we use it ignores additional unmatched characters in the scanned string.
For example, this `if (sscanf(string, "%d", &number) == 1)' will match a number,
but also it will match number with some garbage appended, like "123abc".
As a result, device mapper accepts garbage after some numbers. For example
the command `dmsetup create vg1-new --table "0 16384 linear 254:1bla 34816bla"'
will pass without an error.
This patch fixes all sscanf uses in device mapper. It appends "%c" with
a pointer to a dummy character variable to every sscanf statement.
The construct `if (sscanf(string, "%d%c", &number, &dummy) == 1)' succeeds
only if string is a null-terminated number (optionally preceded by some
whitespace characters). If there is some character appended after the number,
sscanf matches "%c", writes the character to the dummy variable and returns 2.
We check the return value for 1 and consequently reject numbers with some
garbage appended.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Acked-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2012-03-28 17:41:26 +00:00
|
|
|
char dummy;
|
2006-03-27 09:17:54 +00:00
|
|
|
|
|
|
|
md = find_device(param);
|
|
|
|
if (!md)
|
|
|
|
return -ENXIO;
|
|
|
|
|
2008-02-08 02:09:53 +00:00
|
|
|
if (geostr < param->data ||
|
2006-03-27 09:17:54 +00:00
|
|
|
invalid_str(geostr, (void *) param + param_size)) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("Invalid geometry supplied.");
|
2006-03-27 09:17:54 +00:00
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
dm: reject trailing characters in sccanf input
Device mapper uses sscanf to convert arguments to numbers. The problem is that
the way we use it ignores additional unmatched characters in the scanned string.
For example, this `if (sscanf(string, "%d", &number) == 1)' will match a number,
but also it will match number with some garbage appended, like "123abc".
As a result, device mapper accepts garbage after some numbers. For example
the command `dmsetup create vg1-new --table "0 16384 linear 254:1bla 34816bla"'
will pass without an error.
This patch fixes all sscanf uses in device mapper. It appends "%c" with
a pointer to a dummy character variable to every sscanf statement.
The construct `if (sscanf(string, "%d%c", &number, &dummy) == 1)' succeeds
only if string is a null-terminated number (optionally preceded by some
whitespace characters). If there is some character appended after the number,
sscanf matches "%c", writes the character to the dummy variable and returns 2.
We check the return value for 1 and consequently reject numbers with some
garbage appended.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Acked-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2012-03-28 17:41:26 +00:00
|
|
|
x = sscanf(geostr, "%lu %lu %lu %lu%c", indata,
|
|
|
|
indata + 1, indata + 2, indata + 3, &dummy);
|
2006-03-27 09:17:54 +00:00
|
|
|
|
|
|
|
if (x != 4) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("Unable to interpret geometry settings.");
|
2006-03-27 09:17:54 +00:00
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2023-01-17 20:59:55 +00:00
|
|
|
if (indata[0] > 65535 || indata[1] > 255 || indata[2] > 255) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("Geometry exceeds range limits.");
|
2006-03-27 09:17:54 +00:00
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
geometry.cylinders = indata[0];
|
|
|
|
geometry.heads = indata[1];
|
|
|
|
geometry.sectors = indata[2];
|
|
|
|
geometry.start = indata[3];
|
|
|
|
|
|
|
|
r = dm_set_geometry(md, &geometry);
|
|
|
|
|
|
|
|
param->data_size = 0;
|
|
|
|
|
|
|
|
out:
|
|
|
|
dm_put(md);
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
static int do_suspend(struct dm_ioctl *param)
|
|
|
|
{
|
|
|
|
int r = 0;
|
2023-01-25 20:14:58 +00:00
|
|
|
unsigned int suspend_flags = DM_SUSPEND_LOCKFS_FLAG;
|
2005-04-16 22:20:36 +00:00
|
|
|
struct mapped_device *md;
|
|
|
|
|
|
|
|
md = find_device(param);
|
|
|
|
if (!md)
|
|
|
|
return -ENXIO;
|
|
|
|
|
2006-01-06 08:20:07 +00:00
|
|
|
if (param->flags & DM_SKIP_LOCKFS_FLAG)
|
2006-12-08 10:41:04 +00:00
|
|
|
suspend_flags &= ~DM_SUSPEND_LOCKFS_FLAG;
|
2006-12-08 10:41:07 +00:00
|
|
|
if (param->flags & DM_NOFLUSH_FLAG)
|
|
|
|
suspend_flags |= DM_SUSPEND_NOFLUSH_FLAG;
|
2006-01-06 08:20:07 +00:00
|
|
|
|
2010-08-12 03:13:52 +00:00
|
|
|
if (!dm_suspended_md(md)) {
|
2006-12-08 10:41:04 +00:00
|
|
|
r = dm_suspend(md, suspend_flags);
|
2010-08-12 03:13:52 +00:00
|
|
|
if (r)
|
|
|
|
goto out;
|
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2010-08-12 03:13:52 +00:00
|
|
|
__dev_status(md, param);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2010-08-12 03:13:52 +00:00
|
|
|
out:
|
2005-04-16 22:20:36 +00:00
|
|
|
dm_put(md);
|
2010-08-12 03:13:52 +00:00
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int do_resume(struct dm_ioctl *param)
|
|
|
|
{
|
|
|
|
int r = 0;
|
2023-01-25 20:14:58 +00:00
|
|
|
unsigned int suspend_flags = DM_SUSPEND_LOCKFS_FLAG;
|
2005-04-16 22:20:36 +00:00
|
|
|
struct hash_cell *hc;
|
|
|
|
struct mapped_device *md;
|
2009-12-10 23:52:24 +00:00
|
|
|
struct dm_table *new_map, *old_map = NULL;
|
dm: send just one event on resize, not two
Device mapper sends an uevent when the device is suspended, using the
function set_capacity_and_notify. However, this causes a race condition
with udev.
Udev skips scanning dm devices that are suspended. If we send an uevent
while we are suspended, udev will be racing with device mapper resume
code. If the device mapper resume code wins the race, udev will process
the uevent after the device is resumed and it will properly scan the
device.
However, if udev wins the race, it will receive the uevent, find out that
the dm device is suspended and skip scanning the device. This causes bugs
such as systemd unmounting the device - see
https://bugzilla.redhat.com/show_bug.cgi?id=2158628
This commit fixes this race.
We replace the function set_capacity_and_notify with set_capacity, so that
the uevent is not sent at this point. In do_resume, we detect if the
capacity has changed and we pass a boolean variable need_resize_uevent to
dm_kobject_uevent. dm_kobject_uevent adds "RESIZE=1" to the uevent if
need_resize_uevent is set.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Tested-by: Peter Rajnoha <prajnoha@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Mike Snitzer <snitzer@kernel.org>
2023-02-07 13:33:06 +00:00
|
|
|
bool need_resize_uevent = false;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
down_write(&_hash_lock);
|
|
|
|
|
|
|
|
hc = __find_device_hash_cell(param);
|
|
|
|
if (!hc) {
|
2011-01-13 19:59:55 +00:00
|
|
|
DMDEBUG_LIMIT("device doesn't appear to be in the dev hash table.");
|
2005-04-16 22:20:36 +00:00
|
|
|
up_write(&_hash_lock);
|
|
|
|
return -ENXIO;
|
|
|
|
}
|
|
|
|
|
|
|
|
md = hc->md;
|
|
|
|
|
|
|
|
new_map = hc->new_map;
|
|
|
|
hc->new_map = NULL;
|
|
|
|
param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
|
|
|
|
|
|
|
|
up_write(&_hash_lock);
|
|
|
|
|
|
|
|
/* Do we need to load a new map ? */
|
|
|
|
if (new_map) {
|
dm: send just one event on resize, not two
Device mapper sends an uevent when the device is suspended, using the
function set_capacity_and_notify. However, this causes a race condition
with udev.
Udev skips scanning dm devices that are suspended. If we send an uevent
while we are suspended, udev will be racing with device mapper resume
code. If the device mapper resume code wins the race, udev will process
the uevent after the device is resumed and it will properly scan the
device.
However, if udev wins the race, it will receive the uevent, find out that
the dm device is suspended and skip scanning the device. This causes bugs
such as systemd unmounting the device - see
https://bugzilla.redhat.com/show_bug.cgi?id=2158628
This commit fixes this race.
We replace the function set_capacity_and_notify with set_capacity, so that
the uevent is not sent at this point. In do_resume, we detect if the
capacity has changed and we pass a boolean variable need_resize_uevent to
dm_kobject_uevent. dm_kobject_uevent adds "RESIZE=1" to the uevent if
need_resize_uevent is set.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Tested-by: Peter Rajnoha <prajnoha@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Mike Snitzer <snitzer@kernel.org>
2023-02-07 13:33:06 +00:00
|
|
|
sector_t old_size, new_size;
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
/* Suspend if it isn't already suspended */
|
2023-06-01 06:14:23 +00:00
|
|
|
if (param->flags & DM_SKIP_LOCKFS_FLAG)
|
2006-12-08 10:41:04 +00:00
|
|
|
suspend_flags &= ~DM_SUSPEND_LOCKFS_FLAG;
|
2006-12-08 10:41:07 +00:00
|
|
|
if (param->flags & DM_NOFLUSH_FLAG)
|
|
|
|
suspend_flags |= DM_SUSPEND_NOFLUSH_FLAG;
|
2024-08-13 10:39:52 +00:00
|
|
|
if (!dm_suspended_md(md)) {
|
|
|
|
r = dm_suspend(md, suspend_flags);
|
|
|
|
if (r) {
|
|
|
|
down_write(&_hash_lock);
|
|
|
|
hc = dm_get_mdptr(md);
|
|
|
|
if (hc && !hc->new_map) {
|
|
|
|
hc->new_map = new_map;
|
|
|
|
new_map = NULL;
|
|
|
|
} else {
|
|
|
|
r = -ENXIO;
|
|
|
|
}
|
|
|
|
up_write(&_hash_lock);
|
|
|
|
if (new_map) {
|
|
|
|
dm_sync_table(md);
|
|
|
|
dm_table_destroy(new_map);
|
|
|
|
}
|
|
|
|
dm_put(md);
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
dm: send just one event on resize, not two
Device mapper sends an uevent when the device is suspended, using the
function set_capacity_and_notify. However, this causes a race condition
with udev.
Udev skips scanning dm devices that are suspended. If we send an uevent
while we are suspended, udev will be racing with device mapper resume
code. If the device mapper resume code wins the race, udev will process
the uevent after the device is resumed and it will properly scan the
device.
However, if udev wins the race, it will receive the uevent, find out that
the dm device is suspended and skip scanning the device. This causes bugs
such as systemd unmounting the device - see
https://bugzilla.redhat.com/show_bug.cgi?id=2158628
This commit fixes this race.
We replace the function set_capacity_and_notify with set_capacity, so that
the uevent is not sent at this point. In do_resume, we detect if the
capacity has changed and we pass a boolean variable need_resize_uevent to
dm_kobject_uevent. dm_kobject_uevent adds "RESIZE=1" to the uevent if
need_resize_uevent is set.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Tested-by: Peter Rajnoha <prajnoha@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Mike Snitzer <snitzer@kernel.org>
2023-02-07 13:33:06 +00:00
|
|
|
old_size = dm_get_size(md);
|
2009-12-10 23:52:24 +00:00
|
|
|
old_map = dm_swap_table(md, new_map);
|
|
|
|
if (IS_ERR(old_map)) {
|
2013-07-10 22:41:18 +00:00
|
|
|
dm_sync_table(md);
|
dm table: rework reference counting
Rework table reference counting.
The existing code uses a reference counter. When the last reference is
dropped and the counter reaches zero, the table destructor is called.
Table reference counters are acquired/released from upcalls from other
kernel code (dm_any_congested, dm_merge_bvec, dm_unplug_all).
If the reference counter reaches zero in one of the upcalls, the table
destructor is called from almost random kernel code.
This leads to various problems:
* dm_any_congested being called under a spinlock, which calls the
destructor, which calls some sleeping function.
* the destructor attempting to take a lock that is already taken by the
same process.
* stale reference from some other kernel code keeps the table
constructed, which keeps some devices open, even after successful
return from "dmsetup remove". This can confuse lvm and prevent closing
of underlying devices or reusing device minor numbers.
The patch changes reference counting so that the table destructor can be
called only at predetermined places.
The table has always exactly one reference from either mapped_device->map
or hash_cell->new_map. After this patch, this reference is not counted
in table->holders. A pair of dm_create_table/dm_destroy_table functions
is used for table creation/destruction.
Temporary references from the other code increase table->holders. A pair
of dm_table_get/dm_table_put functions is used to manipulate it.
When the table is about to be destroyed, we wait for table->holders to
reach 0. Then, we call the table destructor. We use active waiting with
msleep(1), because the situation happens rarely (to one user in 5 years)
and removing the device isn't performance-critical task: the user doesn't
care if it takes one tick more or not.
This way, the destructor is called only at specific points
(dm_table_destroy function) and the above problems associated with lazy
destruction can't happen.
Finally remove the temporary protection added to dm_any_congested().
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-01-06 03:05:10 +00:00
|
|
|
dm_table_destroy(new_map);
|
2005-04-16 22:20:36 +00:00
|
|
|
dm_put(md);
|
2009-12-10 23:52:24 +00:00
|
|
|
return PTR_ERR(old_map);
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
dm: send just one event on resize, not two
Device mapper sends an uevent when the device is suspended, using the
function set_capacity_and_notify. However, this causes a race condition
with udev.
Udev skips scanning dm devices that are suspended. If we send an uevent
while we are suspended, udev will be racing with device mapper resume
code. If the device mapper resume code wins the race, udev will process
the uevent after the device is resumed and it will properly scan the
device.
However, if udev wins the race, it will receive the uevent, find out that
the dm device is suspended and skip scanning the device. This causes bugs
such as systemd unmounting the device - see
https://bugzilla.redhat.com/show_bug.cgi?id=2158628
This commit fixes this race.
We replace the function set_capacity_and_notify with set_capacity, so that
the uevent is not sent at this point. In do_resume, we detect if the
capacity has changed and we pass a boolean variable need_resize_uevent to
dm_kobject_uevent. dm_kobject_uevent adds "RESIZE=1" to the uevent if
need_resize_uevent is set.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Tested-by: Peter Rajnoha <prajnoha@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Mike Snitzer <snitzer@kernel.org>
2023-02-07 13:33:06 +00:00
|
|
|
new_size = dm_get_size(md);
|
|
|
|
if (old_size && new_size && old_size != new_size)
|
|
|
|
need_resize_uevent = true;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2023-06-08 11:02:55 +00:00
|
|
|
if (dm_table_get_mode(new_map) & BLK_OPEN_WRITE)
|
2005-04-16 22:20:36 +00:00
|
|
|
set_disk_ro(dm_disk(md), 0);
|
|
|
|
else
|
|
|
|
set_disk_ro(dm_disk(md), 1);
|
|
|
|
}
|
|
|
|
|
2010-03-06 02:32:24 +00:00
|
|
|
if (dm_suspended_md(md)) {
|
2005-04-16 22:20:36 +00:00
|
|
|
r = dm_resume(md);
|
dm ima: measure data on device resume
A given block device can load a table multiple times, with different
input parameters, before eventually resuming it. Further, a device may
be suspended and then resumed. The device may never resume after a
table-load. Because of the above valid scenarios for a given device,
it is important to measure and log the device resume event using IMA.
Also, if the table is large, measuring it in clear-text each time the
device changes state, will unnecessarily increase the size of IMA log.
Since the table clear-text is already measured during table-load event,
measuring the hash during resume should be sufficient to validate the
table contents.
Measure the device parameters, and hash of the active table, when the
device is resumed.
Signed-off-by: Tushar Sugandhi <tusharsu@linux.microsoft.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
2021-07-13 00:48:59 +00:00
|
|
|
if (!r) {
|
|
|
|
dm_ima_measure_on_device_resume(md, new_map ? true : false);
|
|
|
|
|
dm: send just one event on resize, not two
Device mapper sends an uevent when the device is suspended, using the
function set_capacity_and_notify. However, this causes a race condition
with udev.
Udev skips scanning dm devices that are suspended. If we send an uevent
while we are suspended, udev will be racing with device mapper resume
code. If the device mapper resume code wins the race, udev will process
the uevent after the device is resumed and it will properly scan the
device.
However, if udev wins the race, it will receive the uevent, find out that
the dm device is suspended and skip scanning the device. This causes bugs
such as systemd unmounting the device - see
https://bugzilla.redhat.com/show_bug.cgi?id=2158628
This commit fixes this race.
We replace the function set_capacity_and_notify with set_capacity, so that
the uevent is not sent at this point. In do_resume, we detect if the
capacity has changed and we pass a boolean variable need_resize_uevent to
dm_kobject_uevent. dm_kobject_uevent adds "RESIZE=1" to the uevent if
need_resize_uevent is set.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Tested-by: Peter Rajnoha <prajnoha@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Mike Snitzer <snitzer@kernel.org>
2023-02-07 13:33:06 +00:00
|
|
|
if (!dm_kobject_uevent(md, KOBJ_CHANGE, param->event_nr, need_resize_uevent))
|
dm ima: measure data on device resume
A given block device can load a table multiple times, with different
input parameters, before eventually resuming it. Further, a device may
be suspended and then resumed. The device may never resume after a
table-load. Because of the above valid scenarios for a given device,
it is important to measure and log the device resume event using IMA.
Also, if the table is large, measuring it in clear-text each time the
device changes state, will unnecessarily increase the size of IMA log.
Since the table clear-text is already measured during table-load event,
measuring the hash during resume should be sufficient to validate the
table contents.
Measure the device parameters, and hash of the active table, when the
device is resumed.
Signed-off-by: Tushar Sugandhi <tusharsu@linux.microsoft.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
2021-07-13 00:48:59 +00:00
|
|
|
param->flags |= DM_UEVENT_GENERATED_FLAG;
|
|
|
|
}
|
2010-03-06 02:32:24 +00:00
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2013-07-10 22:41:18 +00:00
|
|
|
/*
|
|
|
|
* Since dm_swap_table synchronizes RCU, nobody should be in
|
|
|
|
* read-side critical section already.
|
|
|
|
*/
|
2009-12-10 23:52:24 +00:00
|
|
|
if (old_map)
|
|
|
|
dm_table_destroy(old_map);
|
2009-06-22 09:12:30 +00:00
|
|
|
|
2010-03-06 02:32:24 +00:00
|
|
|
if (!r)
|
2010-08-12 03:13:52 +00:00
|
|
|
__dev_status(md, param);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
dm_put(md);
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Set or unset the suspension state of a device.
|
|
|
|
* If the device already is in the requested state we just return its status.
|
|
|
|
*/
|
2017-05-05 18:12:52 +00:00
|
|
|
static int dev_suspend(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
if (param->flags & DM_SUSPEND_FLAG)
|
|
|
|
return do_suspend(param);
|
|
|
|
|
|
|
|
return do_resume(param);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Copies device info back to user space, used by
|
|
|
|
* the create and info ioctls.
|
|
|
|
*/
|
2017-05-05 18:12:52 +00:00
|
|
|
static int dev_status(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
struct mapped_device *md;
|
|
|
|
|
|
|
|
md = find_device(param);
|
|
|
|
if (!md)
|
|
|
|
return -ENXIO;
|
|
|
|
|
2010-08-12 03:13:52 +00:00
|
|
|
__dev_status(md, param);
|
2005-04-16 22:20:36 +00:00
|
|
|
dm_put(md);
|
2010-08-12 03:13:52 +00:00
|
|
|
|
|
|
|
return 0;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Build up the status struct for each target
|
|
|
|
*/
|
|
|
|
static void retrieve_status(struct dm_table *table,
|
|
|
|
struct dm_ioctl *param, size_t param_size)
|
|
|
|
{
|
|
|
|
unsigned int i, num_targets;
|
|
|
|
struct dm_target_spec *spec;
|
|
|
|
char *outbuf, *outptr;
|
|
|
|
status_type_t type;
|
|
|
|
size_t remaining, len, used = 0;
|
2023-01-25 20:14:58 +00:00
|
|
|
unsigned int status_flags = 0;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
outptr = outbuf = get_result_buffer(param, param_size, &len);
|
|
|
|
|
|
|
|
if (param->flags & DM_STATUS_TABLE_FLAG)
|
|
|
|
type = STATUSTYPE_TABLE;
|
dm ima: measure data on table load
DM configures a block device with various target specific attributes
passed to it as a table. DM loads the table, and calls each target’s
respective constructors with the attributes as input parameters.
Some of these attributes are critical to ensure the device meets
certain security bar. Thus, IMA should measure these attributes, to
ensure they are not tampered with, during the lifetime of the device.
So that the external services can have high confidence in the
configuration of the block-devices on a given system.
Some devices may have large tables. And a given device may change its
state (table-load, suspend, resume, rename, remove, table-clear etc.)
many times. Measuring these attributes each time when the device
changes its state will significantly increase the size of the IMA logs.
Further, once configured, these attributes are not expected to change
unless a new table is loaded, or a device is removed and recreated.
Therefore the clear-text of the attributes should only be measured
during table load, and the hash of the active/inactive table should be
measured for the remaining device state changes.
Export IMA function ima_measure_critical_data() to allow measurement
of DM device parameters, as well as target specific attributes, during
table load. Compute the hash of the inactive table and store it for
measurements during future state change. If a load is called multiple
times, update the inactive table hash with the hash of the latest
populated table. So that the correct inactive table hash is measured
when the device transitions to different states like resume, remove,
rename, etc.
Signed-off-by: Tushar Sugandhi <tusharsu@linux.microsoft.com>
Signed-off-by: Colin Ian King <colin.king@canonical.com> # leak fix
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
2021-07-13 00:48:58 +00:00
|
|
|
else if (param->flags & DM_IMA_MEASUREMENT_FLAG)
|
|
|
|
type = STATUSTYPE_IMA;
|
2005-04-16 22:20:36 +00:00
|
|
|
else
|
|
|
|
type = STATUSTYPE_INFO;
|
|
|
|
|
|
|
|
/* Get all the target info */
|
2022-07-05 18:00:36 +00:00
|
|
|
num_targets = table->num_targets;
|
2005-04-16 22:20:36 +00:00
|
|
|
for (i = 0; i < num_targets; i++) {
|
|
|
|
struct dm_target *ti = dm_table_get_target(table, i);
|
2013-03-01 22:45:44 +00:00
|
|
|
size_t l;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
remaining = len - (outptr - outbuf);
|
|
|
|
if (remaining <= sizeof(struct dm_target_spec)) {
|
|
|
|
param->flags |= DM_BUFFER_FULL_FLAG;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
spec = (struct dm_target_spec *) outptr;
|
|
|
|
|
|
|
|
spec->status = 0;
|
|
|
|
spec->sector_start = ti->begin;
|
|
|
|
spec->length = ti->len;
|
2023-09-25 06:54:51 +00:00
|
|
|
strscpy_pad(spec->target_type, ti->type->name,
|
|
|
|
sizeof(spec->target_type));
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
outptr += sizeof(struct dm_target_spec);
|
|
|
|
remaining = len - (outptr - outbuf);
|
|
|
|
if (remaining <= 0) {
|
|
|
|
param->flags |= DM_BUFFER_FULL_FLAG;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Get the status/table string from the target driver */
|
|
|
|
if (ti->type->status) {
|
2012-07-27 14:08:16 +00:00
|
|
|
if (param->flags & DM_NOFLUSH_FLAG)
|
|
|
|
status_flags |= DM_STATUS_NOFLUSH_FLAG;
|
2013-03-01 22:45:44 +00:00
|
|
|
ti->type->status(ti, type, status_flags, outptr, remaining);
|
2005-04-16 22:20:36 +00:00
|
|
|
} else
|
|
|
|
outptr[0] = '\0';
|
|
|
|
|
2013-03-01 22:45:44 +00:00
|
|
|
l = strlen(outptr) + 1;
|
|
|
|
if (l == remaining) {
|
|
|
|
param->flags |= DM_BUFFER_FULL_FLAG;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
outptr += l;
|
2005-04-16 22:20:36 +00:00
|
|
|
used = param->data_start + (outptr - outbuf);
|
|
|
|
|
|
|
|
outptr = align_ptr(outptr);
|
|
|
|
spec->next = outptr - outbuf;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (used)
|
|
|
|
param->data_size = used;
|
|
|
|
|
|
|
|
param->target_count = num_targets;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Wait for a device to report an event
|
|
|
|
*/
|
2017-05-05 18:12:52 +00:00
|
|
|
static int dev_wait(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
2010-08-12 03:13:52 +00:00
|
|
|
int r = 0;
|
2005-04-16 22:20:36 +00:00
|
|
|
struct mapped_device *md;
|
|
|
|
struct dm_table *table;
|
2013-07-10 22:41:18 +00:00
|
|
|
int srcu_idx;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
md = find_device(param);
|
|
|
|
if (!md)
|
|
|
|
return -ENXIO;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Wait for a notification event
|
|
|
|
*/
|
|
|
|
if (dm_wait_event(md, param->event_nr)) {
|
|
|
|
r = -ERESTARTSYS;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The userland program is going to want to know what
|
|
|
|
* changed to trigger the event, so we may as well tell
|
|
|
|
* him and save an ioctl.
|
|
|
|
*/
|
2010-08-12 03:13:52 +00:00
|
|
|
__dev_status(md, param);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2013-07-10 22:41:18 +00:00
|
|
|
table = dm_get_live_or_inactive_table(md, param, &srcu_idx);
|
|
|
|
if (table)
|
2005-04-16 22:20:36 +00:00
|
|
|
retrieve_status(table, param, param_size);
|
2013-07-10 22:41:18 +00:00
|
|
|
dm_put_live_table(md, srcu_idx);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2010-08-12 03:13:52 +00:00
|
|
|
out:
|
2005-04-16 22:20:36 +00:00
|
|
|
dm_put(md);
|
2010-08-12 03:13:52 +00:00
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2017-05-05 18:12:52 +00:00
|
|
|
/*
|
|
|
|
* Remember the global event number and make it possible to poll
|
|
|
|
* for further events.
|
|
|
|
*/
|
|
|
|
static int dev_arm_poll(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
|
|
|
{
|
|
|
|
struct dm_file *priv = filp->private_data;
|
|
|
|
|
|
|
|
priv->global_event_nr = atomic_read(&dm_global_event_nr);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2023-06-08 11:02:55 +00:00
|
|
|
static inline blk_mode_t get_mode(struct dm_ioctl *param)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
2023-06-08 11:02:55 +00:00
|
|
|
blk_mode_t mode = BLK_OPEN_READ | BLK_OPEN_WRITE;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
if (param->flags & DM_READONLY_FLAG)
|
2023-06-08 11:02:55 +00:00
|
|
|
mode = BLK_OPEN_READ;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
return mode;
|
|
|
|
}
|
|
|
|
|
2023-06-03 14:52:41 +00:00
|
|
|
static int next_target(struct dm_target_spec *last, uint32_t next, const char *end,
|
2005-04-16 22:20:36 +00:00
|
|
|
struct dm_target_spec **spec, char **target_params)
|
|
|
|
{
|
2023-06-03 14:52:39 +00:00
|
|
|
static_assert(__alignof__(struct dm_target_spec) <= 8,
|
|
|
|
"struct dm_target_spec must not require more than 8-byte alignment");
|
|
|
|
|
2023-06-03 14:52:40 +00:00
|
|
|
/*
|
|
|
|
* Number of bytes remaining, starting with last. This is always
|
|
|
|
* sizeof(struct dm_target_spec) or more, as otherwise *last was
|
|
|
|
* out of bounds already.
|
|
|
|
*/
|
2023-06-03 14:52:41 +00:00
|
|
|
size_t remaining = end - (char *)last;
|
2023-06-03 14:52:40 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* There must be room for both the next target spec and the
|
|
|
|
* NUL-terminator of the target itself.
|
|
|
|
*/
|
|
|
|
if (remaining - sizeof(struct dm_target_spec) <= next) {
|
|
|
|
DMERR("Target spec extends beyond end of parameters");
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2023-06-03 14:52:39 +00:00
|
|
|
if (next % __alignof__(struct dm_target_spec)) {
|
|
|
|
DMERR("Next dm_target_spec (offset %u) is not %zu-byte aligned",
|
|
|
|
next, __alignof__(struct dm_target_spec));
|
2005-04-16 22:20:36 +00:00
|
|
|
return -EINVAL;
|
2023-06-03 14:52:39 +00:00
|
|
|
}
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
*spec = (struct dm_target_spec *) ((unsigned char *) last + next);
|
|
|
|
*target_params = (char *) (*spec + 1);
|
|
|
|
|
2023-06-03 14:52:41 +00:00
|
|
|
return 0;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static int populate_table(struct dm_table *table,
|
|
|
|
struct dm_ioctl *param, size_t param_size)
|
|
|
|
{
|
|
|
|
int r;
|
|
|
|
unsigned int i = 0;
|
|
|
|
struct dm_target_spec *spec = (struct dm_target_spec *) param;
|
|
|
|
uint32_t next = param->data_start;
|
2023-06-03 14:52:41 +00:00
|
|
|
const char *const end = (const char *) param + param_size;
|
2005-04-16 22:20:36 +00:00
|
|
|
char *target_params;
|
2023-06-03 14:52:41 +00:00
|
|
|
size_t min_size = sizeof(struct dm_ioctl);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
if (!param->target_count) {
|
2023-02-06 22:42:32 +00:00
|
|
|
DMERR("%s: no targets specified", __func__);
|
2005-04-16 22:20:36 +00:00
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (i = 0; i < param->target_count; i++) {
|
2023-06-03 14:52:41 +00:00
|
|
|
const char *nul_terminator;
|
|
|
|
|
|
|
|
if (next < min_size) {
|
|
|
|
DMERR("%s: next target spec (offset %u) overlaps %s",
|
|
|
|
__func__, next, i ? "previous target" : "'struct dm_ioctl'");
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
r = next_target(spec, next, end, &spec, &target_params);
|
|
|
|
if (r) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("unable to find target");
|
2005-04-16 22:20:36 +00:00
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2023-06-03 14:52:41 +00:00
|
|
|
nul_terminator = memchr(target_params, 0, (size_t)(end - target_params));
|
|
|
|
if (nul_terminator == NULL) {
|
|
|
|
DMERR("%s: target parameters not NUL-terminated", __func__);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Add 1 for NUL terminator */
|
|
|
|
min_size = (size_t)(nul_terminator - (const char *)spec) + 1;
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
r = dm_table_add_target(table, spec->target_type,
|
|
|
|
(sector_t) spec->sector_start,
|
|
|
|
(sector_t) spec->length,
|
|
|
|
target_params);
|
|
|
|
if (r) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("error adding target to table");
|
2005-04-16 22:20:36 +00:00
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
next = spec->next;
|
|
|
|
}
|
|
|
|
|
|
|
|
return dm_table_complete(table);
|
|
|
|
}
|
|
|
|
|
2017-04-27 17:11:23 +00:00
|
|
|
static bool is_valid_type(enum dm_queue_mode cur, enum dm_queue_mode new)
|
2016-06-28 19:37:15 +00:00
|
|
|
{
|
|
|
|
if (cur == new ||
|
|
|
|
(cur == DM_TYPE_BIO_BASED && new == DM_TYPE_DAX_BIO_BASED))
|
|
|
|
return true;
|
|
|
|
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2017-05-05 18:12:52 +00:00
|
|
|
static int table_load(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
int r;
|
|
|
|
struct hash_cell *hc;
|
2013-07-10 22:41:18 +00:00
|
|
|
struct dm_table *t, *old_map = NULL;
|
2006-03-27 09:17:54 +00:00
|
|
|
struct mapped_device *md;
|
2011-10-31 20:19:04 +00:00
|
|
|
struct target_type *immutable_target_type;
|
2006-03-27 09:17:54 +00:00
|
|
|
|
|
|
|
md = find_device(param);
|
|
|
|
if (!md)
|
|
|
|
return -ENXIO;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2006-03-27 09:17:54 +00:00
|
|
|
r = dm_table_create(&t, get_mode(param), param->target_count, md);
|
2005-04-16 22:20:36 +00:00
|
|
|
if (r)
|
2013-08-28 00:03:00 +00:00
|
|
|
goto err;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2013-08-27 22:57:03 +00:00
|
|
|
/* Protect md->type and md->queue against concurrent table loads. */
|
|
|
|
dm_lock_md_type(md);
|
2005-04-16 22:20:36 +00:00
|
|
|
r = populate_table(t, param, param_size);
|
2013-08-28 00:03:00 +00:00
|
|
|
if (r)
|
|
|
|
goto err_unlock_md_type;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
dm ima: measure data on table load
DM configures a block device with various target specific attributes
passed to it as a table. DM loads the table, and calls each target’s
respective constructors with the attributes as input parameters.
Some of these attributes are critical to ensure the device meets
certain security bar. Thus, IMA should measure these attributes, to
ensure they are not tampered with, during the lifetime of the device.
So that the external services can have high confidence in the
configuration of the block-devices on a given system.
Some devices may have large tables. And a given device may change its
state (table-load, suspend, resume, rename, remove, table-clear etc.)
many times. Measuring these attributes each time when the device
changes its state will significantly increase the size of the IMA logs.
Further, once configured, these attributes are not expected to change
unless a new table is loaded, or a device is removed and recreated.
Therefore the clear-text of the attributes should only be measured
during table load, and the hash of the active/inactive table should be
measured for the remaining device state changes.
Export IMA function ima_measure_critical_data() to allow measurement
of DM device parameters, as well as target specific attributes, during
table load. Compute the hash of the inactive table and store it for
measurements during future state change. If a load is called multiple
times, update the inactive table hash with the hash of the latest
populated table. So that the correct inactive table hash is measured
when the device transitions to different states like resume, remove,
rename, etc.
Signed-off-by: Tushar Sugandhi <tusharsu@linux.microsoft.com>
Signed-off-by: Colin Ian King <colin.king@canonical.com> # leak fix
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
2021-07-13 00:48:58 +00:00
|
|
|
dm_ima_measure_on_table_load(t, STATUSTYPE_IMA);
|
|
|
|
|
2011-10-31 20:19:04 +00:00
|
|
|
immutable_target_type = dm_get_immutable_target_type(md);
|
|
|
|
if (immutable_target_type &&
|
2016-02-06 23:38:46 +00:00
|
|
|
(immutable_target_type != dm_table_get_immutable_target_type(t)) &&
|
|
|
|
!dm_table_get_wildcard_target(t)) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("can't replace immutable target type %s",
|
|
|
|
immutable_target_type->name);
|
2011-10-31 20:19:04 +00:00
|
|
|
r = -EINVAL;
|
2013-08-28 00:03:00 +00:00
|
|
|
goto err_unlock_md_type;
|
2011-10-31 20:19:04 +00:00
|
|
|
}
|
|
|
|
|
2015-04-30 14:10:36 +00:00
|
|
|
if (dm_get_md_type(md) == DM_TYPE_NONE) {
|
|
|
|
/* setup md->queue to reflect md's type (may block) */
|
2016-01-31 17:05:42 +00:00
|
|
|
r = dm_setup_md_queue(md, t);
|
2015-04-30 14:10:36 +00:00
|
|
|
if (r) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("unable to set up device queue for new table.");
|
2015-04-30 14:10:36 +00:00
|
|
|
goto err_unlock_md_type;
|
|
|
|
}
|
2016-06-28 19:37:15 +00:00
|
|
|
} else if (!is_valid_type(dm_get_md_type(md), dm_table_get_type(t))) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("can't change device type (old=%u vs new=%u) after initial table load.",
|
|
|
|
dm_get_md_type(md), dm_table_get_type(t));
|
2010-08-12 03:14:01 +00:00
|
|
|
r = -EINVAL;
|
2013-08-28 00:03:00 +00:00
|
|
|
goto err_unlock_md_type;
|
2010-08-12 03:14:01 +00:00
|
|
|
}
|
2010-08-12 03:14:02 +00:00
|
|
|
|
2010-08-12 03:14:01 +00:00
|
|
|
dm_unlock_md_type(md);
|
|
|
|
|
|
|
|
/* stage inactive table */
|
2005-04-16 22:20:36 +00:00
|
|
|
down_write(&_hash_lock);
|
2006-03-27 09:17:54 +00:00
|
|
|
hc = dm_get_mdptr(md);
|
2022-12-16 04:23:53 +00:00
|
|
|
if (!hc) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("device has been removed from the dev hash table.");
|
2006-03-27 09:17:54 +00:00
|
|
|
up_write(&_hash_lock);
|
|
|
|
r = -ENXIO;
|
2013-08-28 00:03:00 +00:00
|
|
|
goto err_destroy_table;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
if (hc->new_map)
|
2013-07-10 22:41:18 +00:00
|
|
|
old_map = hc->new_map;
|
2005-04-16 22:20:36 +00:00
|
|
|
hc->new_map = t;
|
2006-03-27 09:17:54 +00:00
|
|
|
up_write(&_hash_lock);
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
param->flags |= DM_INACTIVE_PRESENT_FLAG;
|
2010-08-12 03:13:52 +00:00
|
|
|
__dev_status(md, param);
|
2006-03-27 09:17:54 +00:00
|
|
|
|
2013-07-10 22:41:18 +00:00
|
|
|
if (old_map) {
|
|
|
|
dm_sync_table(md);
|
|
|
|
dm_table_destroy(old_map);
|
|
|
|
}
|
|
|
|
|
2006-03-27 09:17:54 +00:00
|
|
|
dm_put(md);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2013-08-28 00:03:00 +00:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
err_unlock_md_type:
|
|
|
|
dm_unlock_md_type(md);
|
|
|
|
err_destroy_table:
|
|
|
|
dm_table_destroy(t);
|
|
|
|
err:
|
|
|
|
dm_put(md);
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2017-05-05 18:12:52 +00:00
|
|
|
static int table_clear(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
struct hash_cell *hc;
|
2006-06-26 07:27:24 +00:00
|
|
|
struct mapped_device *md;
|
2013-07-10 22:41:18 +00:00
|
|
|
struct dm_table *old_map = NULL;
|
2021-07-13 00:49:01 +00:00
|
|
|
bool has_new_map = false;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
down_write(&_hash_lock);
|
|
|
|
|
|
|
|
hc = __find_device_hash_cell(param);
|
|
|
|
if (!hc) {
|
2011-01-13 19:59:55 +00:00
|
|
|
DMDEBUG_LIMIT("device doesn't appear to be in the dev hash table.");
|
2005-04-16 22:20:36 +00:00
|
|
|
up_write(&_hash_lock);
|
|
|
|
return -ENXIO;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (hc->new_map) {
|
2013-07-10 22:41:18 +00:00
|
|
|
old_map = hc->new_map;
|
2005-04-16 22:20:36 +00:00
|
|
|
hc->new_map = NULL;
|
2021-07-13 00:49:01 +00:00
|
|
|
has_new_map = true;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
2006-06-26 07:27:24 +00:00
|
|
|
md = hc->md;
|
2005-04-16 22:20:36 +00:00
|
|
|
up_write(&_hash_lock);
|
2023-04-17 15:59:56 +00:00
|
|
|
|
|
|
|
param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
|
|
|
|
__dev_status(md, param);
|
|
|
|
|
2013-07-10 22:41:18 +00:00
|
|
|
if (old_map) {
|
|
|
|
dm_sync_table(md);
|
|
|
|
dm_table_destroy(old_map);
|
|
|
|
}
|
2021-07-13 00:49:01 +00:00
|
|
|
dm_ima_measure_on_table_clear(md, has_new_map);
|
2006-06-26 07:27:24 +00:00
|
|
|
dm_put(md);
|
2010-08-12 03:13:52 +00:00
|
|
|
|
|
|
|
return 0;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Retrieves a list of devices used by a particular dm device.
|
|
|
|
*/
|
|
|
|
static void retrieve_deps(struct dm_table *table,
|
|
|
|
struct dm_ioctl *param, size_t param_size)
|
|
|
|
{
|
|
|
|
unsigned int count = 0;
|
|
|
|
struct list_head *tmp;
|
|
|
|
size_t len, needed;
|
2008-10-10 12:37:09 +00:00
|
|
|
struct dm_dev_internal *dd;
|
2005-04-16 22:20:36 +00:00
|
|
|
struct dm_target_deps *deps;
|
|
|
|
|
2023-08-09 10:44:20 +00:00
|
|
|
down_read(&table->devices_lock);
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
deps = get_result_buffer(param, param_size, &len);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Count the devices.
|
|
|
|
*/
|
2023-01-30 20:43:57 +00:00
|
|
|
list_for_each(tmp, dm_table_get_devices(table))
|
2005-04-16 22:20:36 +00:00
|
|
|
count++;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Check we have enough space.
|
|
|
|
*/
|
2020-06-17 16:31:45 +00:00
|
|
|
needed = struct_size(deps, dev, count);
|
2005-04-16 22:20:36 +00:00
|
|
|
if (len < needed) {
|
|
|
|
param->flags |= DM_BUFFER_FULL_FLAG;
|
2023-08-09 10:44:20 +00:00
|
|
|
goto out;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Fill in the devices.
|
|
|
|
*/
|
|
|
|
deps->count = count;
|
|
|
|
count = 0;
|
2023-01-30 20:43:57 +00:00
|
|
|
list_for_each_entry(dd, dm_table_get_devices(table), list)
|
2014-08-13 18:53:43 +00:00
|
|
|
deps->dev[count++] = huge_encode_dev(dd->dm_dev->bdev->bd_dev);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
param->data_size = param->data_start + needed;
|
2023-08-09 10:44:20 +00:00
|
|
|
|
|
|
|
out:
|
|
|
|
up_read(&table->devices_lock);
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
2017-05-05 18:12:52 +00:00
|
|
|
static int table_deps(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
struct mapped_device *md;
|
|
|
|
struct dm_table *table;
|
2013-07-10 22:41:18 +00:00
|
|
|
int srcu_idx;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
md = find_device(param);
|
|
|
|
if (!md)
|
|
|
|
return -ENXIO;
|
|
|
|
|
2010-08-12 03:13:52 +00:00
|
|
|
__dev_status(md, param);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2013-07-10 22:41:18 +00:00
|
|
|
table = dm_get_live_or_inactive_table(md, param, &srcu_idx);
|
|
|
|
if (table)
|
2005-04-16 22:20:36 +00:00
|
|
|
retrieve_deps(table, param, param_size);
|
2013-07-10 22:41:18 +00:00
|
|
|
dm_put_live_table(md, srcu_idx);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
dm_put(md);
|
2010-08-12 03:13:52 +00:00
|
|
|
|
|
|
|
return 0;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Return the status of a device as a text string for each
|
|
|
|
* target.
|
|
|
|
*/
|
2017-05-05 18:12:52 +00:00
|
|
|
static int table_status(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
struct mapped_device *md;
|
|
|
|
struct dm_table *table;
|
2013-07-10 22:41:18 +00:00
|
|
|
int srcu_idx;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
md = find_device(param);
|
|
|
|
if (!md)
|
|
|
|
return -ENXIO;
|
|
|
|
|
2010-08-12 03:13:52 +00:00
|
|
|
__dev_status(md, param);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2013-07-10 22:41:18 +00:00
|
|
|
table = dm_get_live_or_inactive_table(md, param, &srcu_idx);
|
|
|
|
if (table)
|
2005-04-16 22:20:36 +00:00
|
|
|
retrieve_status(table, param, param_size);
|
2013-07-10 22:41:18 +00:00
|
|
|
dm_put_live_table(md, srcu_idx);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
dm_put(md);
|
2010-08-12 03:13:52 +00:00
|
|
|
|
|
|
|
return 0;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
2013-03-01 22:45:49 +00:00
|
|
|
/*
|
2013-08-16 14:54:23 +00:00
|
|
|
* Process device-mapper dependent messages. Messages prefixed with '@'
|
|
|
|
* are processed by the DM core. All others are delivered to the target.
|
2013-03-01 22:45:49 +00:00
|
|
|
* Returns a number <= 1 if message was processed by device mapper.
|
|
|
|
* Returns 2 if message should be delivered to the target.
|
|
|
|
*/
|
2023-01-25 20:14:58 +00:00
|
|
|
static int message_for_md(struct mapped_device *md, unsigned int argc, char **argv,
|
|
|
|
char *result, unsigned int maxlen)
|
2013-03-01 22:45:49 +00:00
|
|
|
{
|
2013-08-16 14:54:23 +00:00
|
|
|
int r;
|
|
|
|
|
|
|
|
if (**argv != '@')
|
|
|
|
return 2; /* no '@' prefix, deliver to target */
|
|
|
|
|
2013-11-01 22:27:41 +00:00
|
|
|
if (!strcasecmp(argv[0], "@cancel_deferred_remove")) {
|
|
|
|
if (argc != 1) {
|
|
|
|
DMERR("Invalid arguments for @cancel_deferred_remove");
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
return dm_cancel_deferred_remove(md);
|
|
|
|
}
|
|
|
|
|
2013-08-16 14:54:23 +00:00
|
|
|
r = dm_stats_message(md, argc, argv, result, maxlen);
|
|
|
|
if (r < 2)
|
|
|
|
return r;
|
|
|
|
|
|
|
|
DMERR("Unsupported message sent to DM core: %s", argv[0]);
|
|
|
|
return -EINVAL;
|
2013-03-01 22:45:49 +00:00
|
|
|
}
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
/*
|
|
|
|
* Pass a message to the target that's at the supplied device offset.
|
|
|
|
*/
|
2017-05-05 18:12:52 +00:00
|
|
|
static int target_message(struct file *filp, struct dm_ioctl *param, size_t param_size)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
int r, argc;
|
|
|
|
char **argv;
|
|
|
|
struct mapped_device *md;
|
|
|
|
struct dm_table *table;
|
|
|
|
struct dm_target *ti;
|
|
|
|
struct dm_target_msg *tmsg = (void *) param + param->data_start;
|
2013-03-01 22:45:49 +00:00
|
|
|
size_t maxlen;
|
|
|
|
char *result = get_result_buffer(param, param_size, &maxlen);
|
2013-07-10 22:41:18 +00:00
|
|
|
int srcu_idx;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
md = find_device(param);
|
|
|
|
if (!md)
|
|
|
|
return -ENXIO;
|
|
|
|
|
2007-10-19 21:38:36 +00:00
|
|
|
if (tmsg < (struct dm_target_msg *) param->data ||
|
2005-04-16 22:20:36 +00:00
|
|
|
invalid_str(tmsg->message, (void *) param + param_size)) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("Invalid target message parameters.");
|
2005-04-16 22:20:36 +00:00
|
|
|
r = -EINVAL;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
r = dm_split_args(&argc, &argv, tmsg->message);
|
|
|
|
if (r) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("Failed to split target message parameters");
|
2005-04-16 22:20:36 +00:00
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2011-08-02 11:32:03 +00:00
|
|
|
if (!argc) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("Empty message received.");
|
2020-11-28 10:19:59 +00:00
|
|
|
r = -EINVAL;
|
2012-03-07 19:09:34 +00:00
|
|
|
goto out_argv;
|
2011-08-02 11:32:03 +00:00
|
|
|
}
|
|
|
|
|
2013-03-01 22:45:49 +00:00
|
|
|
r = message_for_md(md, argc, argv, result, maxlen);
|
|
|
|
if (r <= 1)
|
|
|
|
goto out_argv;
|
|
|
|
|
2013-07-10 22:41:18 +00:00
|
|
|
table = dm_get_live_table(md, &srcu_idx);
|
2005-04-16 22:20:36 +00:00
|
|
|
if (!table)
|
2013-07-10 22:41:18 +00:00
|
|
|
goto out_table;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2009-12-10 23:52:20 +00:00
|
|
|
if (dm_deleting_md(md)) {
|
|
|
|
r = -ENXIO;
|
|
|
|
goto out_table;
|
|
|
|
}
|
|
|
|
|
2007-12-13 14:15:25 +00:00
|
|
|
ti = dm_table_find_target(table, tmsg->sector);
|
2019-08-23 13:55:26 +00:00
|
|
|
if (!ti) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("Target message sector outside device.");
|
2005-04-16 22:20:36 +00:00
|
|
|
r = -EINVAL;
|
2007-12-13 14:15:25 +00:00
|
|
|
} else if (ti->type->message)
|
2018-02-28 20:59:59 +00:00
|
|
|
r = ti->type->message(ti, argc, argv, result, maxlen);
|
2005-04-16 22:20:36 +00:00
|
|
|
else {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("Target type does not support messages");
|
2005-04-16 22:20:36 +00:00
|
|
|
r = -EINVAL;
|
|
|
|
}
|
|
|
|
|
2009-12-10 23:52:20 +00:00
|
|
|
out_table:
|
2013-07-10 22:41:18 +00:00
|
|
|
dm_put_live_table(md, srcu_idx);
|
2005-04-16 22:20:36 +00:00
|
|
|
out_argv:
|
|
|
|
kfree(argv);
|
|
|
|
out:
|
2013-03-01 22:45:49 +00:00
|
|
|
if (r >= 0)
|
|
|
|
__dev_status(md, param);
|
|
|
|
|
|
|
|
if (r == 1) {
|
|
|
|
param->flags |= DM_DATA_OUT_FLAG;
|
2013-08-16 14:54:23 +00:00
|
|
|
if (dm_message_test_buffer_overflow(result, maxlen))
|
2013-03-01 22:45:49 +00:00
|
|
|
param->flags |= DM_BUFFER_FULL_FLAG;
|
|
|
|
else
|
|
|
|
param->data_size = param->data_start + strlen(result) + 1;
|
|
|
|
r = 0;
|
|
|
|
}
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
dm_put(md);
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2013-03-01 22:45:48 +00:00
|
|
|
/*
|
|
|
|
* The ioctl parameter block consists of two parts, a dm_ioctl struct
|
|
|
|
* followed by a data buffer. This flag is set if the second part,
|
|
|
|
* which has a variable size, is not used by the function processing
|
|
|
|
* the ioctl.
|
|
|
|
*/
|
2017-09-20 11:29:49 +00:00
|
|
|
#define IOCTL_FLAGS_NO_PARAMS 1
|
|
|
|
#define IOCTL_FLAGS_ISSUE_GLOBAL_EVENT 2
|
2013-03-01 22:45:48 +00:00
|
|
|
|
2023-01-26 14:48:30 +00:00
|
|
|
/*
|
|
|
|
*---------------------------------------------------------------
|
|
|
|
* Implementation of open/close/ioctl on the special char device.
|
|
|
|
*---------------------------------------------------------------
|
|
|
|
*/
|
2013-03-01 22:45:48 +00:00
|
|
|
static ioctl_fn lookup_ioctl(unsigned int cmd, int *ioctl_flags)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
2017-06-22 18:33:47 +00:00
|
|
|
static const struct {
|
2005-04-16 22:20:36 +00:00
|
|
|
int cmd;
|
2013-03-01 22:45:48 +00:00
|
|
|
int flags;
|
2005-04-16 22:20:36 +00:00
|
|
|
ioctl_fn fn;
|
|
|
|
} _ioctls[] = {
|
2013-03-01 22:45:48 +00:00
|
|
|
{DM_VERSION_CMD, 0, NULL}, /* version is dealt with elsewhere */
|
2017-09-20 11:29:49 +00:00
|
|
|
{DM_REMOVE_ALL_CMD, IOCTL_FLAGS_NO_PARAMS | IOCTL_FLAGS_ISSUE_GLOBAL_EVENT, remove_all},
|
2013-03-01 22:45:48 +00:00
|
|
|
{DM_LIST_DEVICES_CMD, 0, list_devices},
|
|
|
|
|
2017-09-20 11:29:49 +00:00
|
|
|
{DM_DEV_CREATE_CMD, IOCTL_FLAGS_NO_PARAMS | IOCTL_FLAGS_ISSUE_GLOBAL_EVENT, dev_create},
|
|
|
|
{DM_DEV_REMOVE_CMD, IOCTL_FLAGS_NO_PARAMS | IOCTL_FLAGS_ISSUE_GLOBAL_EVENT, dev_remove},
|
|
|
|
{DM_DEV_RENAME_CMD, IOCTL_FLAGS_ISSUE_GLOBAL_EVENT, dev_rename},
|
2013-03-01 22:45:48 +00:00
|
|
|
{DM_DEV_SUSPEND_CMD, IOCTL_FLAGS_NO_PARAMS, dev_suspend},
|
|
|
|
{DM_DEV_STATUS_CMD, IOCTL_FLAGS_NO_PARAMS, dev_status},
|
|
|
|
{DM_DEV_WAIT_CMD, 0, dev_wait},
|
|
|
|
|
|
|
|
{DM_TABLE_LOAD_CMD, 0, table_load},
|
|
|
|
{DM_TABLE_CLEAR_CMD, IOCTL_FLAGS_NO_PARAMS, table_clear},
|
|
|
|
{DM_TABLE_DEPS_CMD, 0, table_deps},
|
|
|
|
{DM_TABLE_STATUS_CMD, 0, table_status},
|
|
|
|
|
|
|
|
{DM_LIST_VERSIONS_CMD, 0, list_versions},
|
|
|
|
|
|
|
|
{DM_TARGET_MSG_CMD, 0, target_message},
|
2017-05-05 18:12:52 +00:00
|
|
|
{DM_DEV_SET_GEOMETRY_CMD, 0, dev_set_geometry},
|
2022-11-01 20:54:27 +00:00
|
|
|
{DM_DEV_ARM_POLL_CMD, IOCTL_FLAGS_NO_PARAMS, dev_arm_poll},
|
|
|
|
{DM_GET_TARGET_VERSION_CMD, 0, get_target_version},
|
2005-04-16 22:20:36 +00:00
|
|
|
};
|
|
|
|
|
2013-03-01 22:45:48 +00:00
|
|
|
if (unlikely(cmd >= ARRAY_SIZE(_ioctls)))
|
|
|
|
return NULL;
|
|
|
|
|
2022-01-29 14:58:39 +00:00
|
|
|
cmd = array_index_nospec(cmd, ARRAY_SIZE(_ioctls));
|
2013-03-01 22:45:48 +00:00
|
|
|
*ioctl_flags = _ioctls[cmd].flags;
|
|
|
|
return _ioctls[cmd].fn;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* As well as checking the version compatibility this always
|
|
|
|
* copies the kernel interface version out.
|
|
|
|
*/
|
2023-06-03 14:52:42 +00:00
|
|
|
static int check_version(unsigned int cmd, struct dm_ioctl __user *user,
|
|
|
|
struct dm_ioctl *kernel_params)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
int r = 0;
|
|
|
|
|
2023-06-03 14:52:42 +00:00
|
|
|
/* Make certain version is first member of dm_ioctl struct */
|
|
|
|
BUILD_BUG_ON(offsetof(struct dm_ioctl, version) != 0);
|
|
|
|
|
|
|
|
if (copy_from_user(kernel_params->version, user->version, sizeof(kernel_params->version)))
|
2005-04-16 22:20:36 +00:00
|
|
|
return -EFAULT;
|
|
|
|
|
2023-06-03 14:52:42 +00:00
|
|
|
if ((kernel_params->version[0] != DM_VERSION_MAJOR) ||
|
|
|
|
(kernel_params->version[1] > DM_VERSION_MINOR)) {
|
2023-02-03 17:55:47 +00:00
|
|
|
DMERR("ioctl interface mismatch: kernel(%u.%u.%u), user(%u.%u.%u), cmd(%d)",
|
2022-08-24 11:25:57 +00:00
|
|
|
DM_VERSION_MAJOR, DM_VERSION_MINOR,
|
|
|
|
DM_VERSION_PATCHLEVEL,
|
2023-06-03 14:52:42 +00:00
|
|
|
kernel_params->version[0],
|
|
|
|
kernel_params->version[1],
|
|
|
|
kernel_params->version[2],
|
|
|
|
cmd);
|
2005-04-16 22:20:36 +00:00
|
|
|
r = -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Fill in the kernel version.
|
|
|
|
*/
|
2023-06-03 14:52:42 +00:00
|
|
|
kernel_params->version[0] = DM_VERSION_MAJOR;
|
|
|
|
kernel_params->version[1] = DM_VERSION_MINOR;
|
|
|
|
kernel_params->version[2] = DM_VERSION_PATCHLEVEL;
|
|
|
|
if (copy_to_user(user->version, kernel_params->version, sizeof(kernel_params->version)))
|
2005-04-16 22:20:36 +00:00
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2016-06-28 14:36:46 +00:00
|
|
|
#define DM_PARAMS_MALLOC 0x0001 /* Params allocated with kvmalloc() */
|
2012-12-21 20:23:36 +00:00
|
|
|
#define DM_WIPE_BUFFER 0x0010 /* Wipe input buffer before returning from ioctl */
|
|
|
|
|
|
|
|
static void free_params(struct dm_ioctl *param, size_t param_size, int param_flags)
|
|
|
|
{
|
|
|
|
if (param_flags & DM_WIPE_BUFFER)
|
|
|
|
memset(param, 0, param_size);
|
|
|
|
|
2016-06-28 14:36:46 +00:00
|
|
|
if (param_flags & DM_PARAMS_MALLOC)
|
|
|
|
kvfree(param);
|
2012-12-21 20:23:36 +00:00
|
|
|
}
|
|
|
|
|
2013-03-01 22:45:49 +00:00
|
|
|
static int copy_params(struct dm_ioctl __user *user, struct dm_ioctl *param_kernel,
|
2018-10-03 16:43:59 +00:00
|
|
|
int ioctl_flags, struct dm_ioctl **param, int *param_flags)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
2013-03-01 22:45:49 +00:00
|
|
|
struct dm_ioctl *dmi;
|
2011-03-24 13:54:30 +00:00
|
|
|
int secure_data;
|
2016-11-18 22:28:00 +00:00
|
|
|
const size_t minimum_data_size = offsetof(struct dm_ioctl, data);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2023-06-03 14:52:42 +00:00
|
|
|
/* check_version() already copied version from userspace, avoid TOCTOU */
|
|
|
|
if (copy_from_user((char *)param_kernel + sizeof(param_kernel->version),
|
|
|
|
(char __user *)user + sizeof(param_kernel->version),
|
|
|
|
minimum_data_size - sizeof(param_kernel->version)))
|
2005-04-16 22:20:36 +00:00
|
|
|
return -EFAULT;
|
|
|
|
|
2024-01-09 14:57:56 +00:00
|
|
|
if (unlikely(param_kernel->data_size < minimum_data_size) ||
|
|
|
|
unlikely(param_kernel->data_size > DM_MAX_TARGETS * DM_MAX_TARGET_PARAMS)) {
|
2022-03-20 21:12:46 +00:00
|
|
|
DMERR("Invalid data size in the ioctl structure: %u",
|
|
|
|
param_kernel->data_size);
|
2005-04-16 22:20:36 +00:00
|
|
|
return -EINVAL;
|
2022-03-20 21:12:46 +00:00
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2013-03-01 22:45:49 +00:00
|
|
|
secure_data = param_kernel->flags & DM_SECURE_DATA_FLAG;
|
2011-03-24 13:54:30 +00:00
|
|
|
|
2012-12-21 20:23:36 +00:00
|
|
|
*param_flags = secure_data ? DM_WIPE_BUFFER : 0;
|
|
|
|
|
2013-03-01 22:45:49 +00:00
|
|
|
if (ioctl_flags & IOCTL_FLAGS_NO_PARAMS) {
|
|
|
|
dmi = param_kernel;
|
|
|
|
dmi->data_size = minimum_data_size;
|
|
|
|
goto data_copied;
|
|
|
|
}
|
|
|
|
|
2012-12-21 20:23:36 +00:00
|
|
|
/*
|
2017-05-18 19:00:51 +00:00
|
|
|
* Use __GFP_HIGH to avoid low memory issues when a device is
|
|
|
|
* suspended and the ioctl is needed to resume it.
|
2012-12-21 20:23:36 +00:00
|
|
|
* Use kmalloc() rather than vmalloc() when we can.
|
2012-12-21 20:23:36 +00:00
|
|
|
*/
|
2012-12-21 20:23:36 +00:00
|
|
|
dmi = NULL;
|
2023-06-26 14:48:40 +00:00
|
|
|
dmi = kvmalloc(param_kernel->data_size, GFP_NOIO | __GFP_HIGH);
|
2012-12-21 20:23:36 +00:00
|
|
|
|
2011-03-24 13:54:30 +00:00
|
|
|
if (!dmi) {
|
2013-03-01 22:45:49 +00:00
|
|
|
if (secure_data && clear_user(user, param_kernel->data_size))
|
2011-03-24 13:54:30 +00:00
|
|
|
return -EFAULT;
|
2005-04-16 22:20:36 +00:00
|
|
|
return -ENOMEM;
|
2011-03-24 13:54:30 +00:00
|
|
|
}
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2016-06-28 14:36:46 +00:00
|
|
|
*param_flags |= DM_PARAMS_MALLOC;
|
|
|
|
|
2018-10-03 16:43:59 +00:00
|
|
|
/* Copy from param_kernel (which was already copied from user) */
|
|
|
|
memcpy(dmi, param_kernel, minimum_data_size);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2018-10-03 16:43:59 +00:00
|
|
|
if (copy_from_user(&dmi->data, (char __user *)user + minimum_data_size,
|
|
|
|
param_kernel->data_size - minimum_data_size))
|
2012-12-21 20:23:30 +00:00
|
|
|
goto bad;
|
2018-10-03 16:43:59 +00:00
|
|
|
data_copied:
|
2011-03-24 13:54:30 +00:00
|
|
|
/* Wipe the user buffer so we do not return it to userspace */
|
2013-03-01 22:45:49 +00:00
|
|
|
if (secure_data && clear_user(user, param_kernel->data_size))
|
2011-03-24 13:54:30 +00:00
|
|
|
goto bad;
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
*param = dmi;
|
|
|
|
return 0;
|
2011-03-24 13:54:28 +00:00
|
|
|
|
|
|
|
bad:
|
2013-03-01 22:45:49 +00:00
|
|
|
free_params(dmi, param_kernel->data_size, *param_flags);
|
2012-12-21 20:23:36 +00:00
|
|
|
|
2011-03-24 13:54:28 +00:00
|
|
|
return -EFAULT;
|
2005-04-16 22:20:36 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static int validate_params(uint cmd, struct dm_ioctl *param)
|
|
|
|
{
|
|
|
|
/* Always clear this flag */
|
|
|
|
param->flags &= ~DM_BUFFER_FULL_FLAG;
|
2010-03-06 02:32:31 +00:00
|
|
|
param->flags &= ~DM_UEVENT_GENERATED_FLAG;
|
2011-03-24 13:54:30 +00:00
|
|
|
param->flags &= ~DM_SECURE_DATA_FLAG;
|
2013-03-01 22:45:49 +00:00
|
|
|
param->flags &= ~DM_DATA_OUT_FLAG;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
/* Ignores parameters */
|
|
|
|
if (cmd == DM_REMOVE_ALL_CMD ||
|
|
|
|
cmd == DM_LIST_DEVICES_CMD ||
|
|
|
|
cmd == DM_LIST_VERSIONS_CMD)
|
|
|
|
return 0;
|
|
|
|
|
2017-04-17 18:05:03 +00:00
|
|
|
if (cmd == DM_DEV_CREATE_CMD) {
|
2005-04-16 22:20:36 +00:00
|
|
|
if (!*param->name) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("name not supplied when creating device");
|
2005-04-16 22:20:36 +00:00
|
|
|
return -EINVAL;
|
|
|
|
}
|
2017-04-17 18:05:03 +00:00
|
|
|
} else if (*param->uuid && *param->name) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("only supply one of name or uuid, cmd(%u)", cmd);
|
2005-04-16 22:20:36 +00:00
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Ensure strings are terminated */
|
|
|
|
param->name[DM_NAME_LEN - 1] = '\0';
|
|
|
|
param->uuid[DM_UUID_LEN - 1] = '\0';
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2017-05-05 18:12:52 +00:00
|
|
|
static int ctl_ioctl(struct file *file, uint command, struct dm_ioctl __user *user)
|
2005-04-16 22:20:36 +00:00
|
|
|
{
|
|
|
|
int r = 0;
|
2013-03-01 22:45:48 +00:00
|
|
|
int ioctl_flags;
|
2012-12-21 20:23:36 +00:00
|
|
|
int param_flags;
|
2005-04-16 22:20:36 +00:00
|
|
|
unsigned int cmd;
|
treewide: Remove uninitialized_var() usage
Using uninitialized_var() is dangerous as it papers over real bugs[1]
(or can in the future), and suppresses unrelated compiler warnings
(e.g. "unused variable"). If the compiler thinks it is uninitialized,
either simply initialize the variable or make compiler changes.
In preparation for removing[2] the[3] macro[4], remove all remaining
needless uses with the following script:
git grep '\buninitialized_var\b' | cut -d: -f1 | sort -u | \
xargs perl -pi -e \
's/\buninitialized_var\(([^\)]+)\)/\1/g;
s:\s*/\* (GCC be quiet|to make compiler happy) \*/$::g;'
drivers/video/fbdev/riva/riva_hw.c was manually tweaked to avoid
pathological white-space.
No outstanding warnings were found building allmodconfig with GCC 9.3.0
for x86_64, i386, arm64, arm, powerpc, powerpc64le, s390x, mips, sparc64,
alpha, and m68k.
[1] https://lore.kernel.org/lkml/20200603174714.192027-1-glider@google.com/
[2] https://lore.kernel.org/lkml/CA+55aFw+Vbj0i=1TGqCR5vQkCzWJ0QxK6CernOU6eedsudAixw@mail.gmail.com/
[3] https://lore.kernel.org/lkml/CA+55aFwgbgqhbp1fkxvRKEpzyR5J8n1vKT1VZdz9knmPuXhOeg@mail.gmail.com/
[4] https://lore.kernel.org/lkml/CA+55aFz2500WfbKXAx8s67wrm9=yVJu65TpLgN_ybYNv0VEOKA@mail.gmail.com/
Reviewed-by: Leon Romanovsky <leonro@mellanox.com> # drivers/infiniband and mlx4/mlx5
Acked-by: Jason Gunthorpe <jgg@mellanox.com> # IB
Acked-by: Kalle Valo <kvalo@codeaurora.org> # wireless drivers
Reviewed-by: Chao Yu <yuchao0@huawei.com> # erofs
Signed-off-by: Kees Cook <keescook@chromium.org>
2020-06-03 20:09:38 +00:00
|
|
|
struct dm_ioctl *param;
|
2005-04-16 22:20:36 +00:00
|
|
|
ioctl_fn fn = NULL;
|
2011-03-24 13:54:28 +00:00
|
|
|
size_t input_param_size;
|
2013-03-01 22:45:49 +00:00
|
|
|
struct dm_ioctl param_kernel;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
|
|
|
/* only root can play with this */
|
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
return -EACCES;
|
|
|
|
|
|
|
|
if (_IOC_TYPE(command) != DM_IOCTL)
|
|
|
|
return -ENOTTY;
|
|
|
|
|
|
|
|
cmd = _IOC_NR(command);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Check the interface version passed in. This also
|
|
|
|
* writes out the kernel's interface version.
|
|
|
|
*/
|
2023-06-03 14:52:42 +00:00
|
|
|
r = check_version(cmd, user, ¶m_kernel);
|
2005-04-16 22:20:36 +00:00
|
|
|
if (r)
|
|
|
|
return r;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Nothing more to do for the version command.
|
|
|
|
*/
|
|
|
|
if (cmd == DM_VERSION_CMD)
|
|
|
|
return 0;
|
|
|
|
|
2013-03-01 22:45:48 +00:00
|
|
|
fn = lookup_ioctl(cmd, &ioctl_flags);
|
2005-04-16 22:20:36 +00:00
|
|
|
if (!fn) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("dm_ctl_ioctl: unknown command 0x%x", command);
|
2005-04-16 22:20:36 +00:00
|
|
|
return -ENOTTY;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Copy the parameters into kernel space.
|
|
|
|
*/
|
2013-03-01 22:45:49 +00:00
|
|
|
r = copy_params(user, ¶m_kernel, ioctl_flags, ¶m, ¶m_flags);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2006-02-01 11:04:52 +00:00
|
|
|
if (r)
|
|
|
|
return r;
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2011-03-24 13:54:30 +00:00
|
|
|
input_param_size = param->data_size;
|
2005-04-16 22:20:36 +00:00
|
|
|
r = validate_params(cmd, param);
|
|
|
|
if (r)
|
|
|
|
goto out;
|
|
|
|
|
2017-04-27 17:32:55 +00:00
|
|
|
param->data_size = offsetof(struct dm_ioctl, data);
|
2017-05-05 18:12:52 +00:00
|
|
|
r = fn(file, param, input_param_size);
|
2005-04-16 22:20:36 +00:00
|
|
|
|
2013-03-01 22:45:48 +00:00
|
|
|
if (unlikely(param->flags & DM_BUFFER_FULL_FLAG) &&
|
|
|
|
unlikely(ioctl_flags & IOCTL_FLAGS_NO_PARAMS))
|
|
|
|
DMERR("ioctl %d tried to output some data but has IOCTL_FLAGS_NO_PARAMS set", cmd);
|
|
|
|
|
2017-09-20 11:29:49 +00:00
|
|
|
if (!r && ioctl_flags & IOCTL_FLAGS_ISSUE_GLOBAL_EVENT)
|
|
|
|
dm_issue_global_event();
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
/*
|
|
|
|
* Copy the results back to userland.
|
|
|
|
*/
|
|
|
|
if (!r && copy_to_user(user, param, param->data_size))
|
|
|
|
r = -EFAULT;
|
|
|
|
|
2011-03-24 13:54:28 +00:00
|
|
|
out:
|
2012-12-21 20:23:36 +00:00
|
|
|
free_params(param, input_param_size, param_flags);
|
2005-04-16 22:20:36 +00:00
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
2008-02-08 02:09:53 +00:00
|
|
|
static long dm_ctl_ioctl(struct file *file, uint command, ulong u)
|
|
|
|
{
|
2017-05-05 18:12:52 +00:00
|
|
|
return (long)ctl_ioctl(file, command, (struct dm_ioctl __user *)u);
|
2008-02-08 02:09:53 +00:00
|
|
|
}
|
|
|
|
|
2008-02-08 02:09:56 +00:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
static long dm_compat_ctl_ioctl(struct file *file, uint command, ulong u)
|
|
|
|
{
|
|
|
|
return (long)dm_ctl_ioctl(file, command, (ulong) compat_ptr(u));
|
|
|
|
}
|
|
|
|
#else
|
|
|
|
#define dm_compat_ctl_ioctl NULL
|
|
|
|
#endif
|
|
|
|
|
2017-01-16 21:05:59 +00:00
|
|
|
static int dm_open(struct inode *inode, struct file *filp)
|
|
|
|
{
|
|
|
|
int r;
|
|
|
|
struct dm_file *priv;
|
|
|
|
|
|
|
|
r = nonseekable_open(inode, filp);
|
|
|
|
if (unlikely(r))
|
|
|
|
return r;
|
|
|
|
|
|
|
|
priv = filp->private_data = kmalloc(sizeof(struct dm_file), GFP_KERNEL);
|
|
|
|
if (!priv)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
priv->global_event_nr = atomic_read(&dm_global_event_nr);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int dm_release(struct inode *inode, struct file *filp)
|
|
|
|
{
|
|
|
|
kfree(filp->private_data);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2017-07-03 10:39:46 +00:00
|
|
|
static __poll_t dm_poll(struct file *filp, poll_table *wait)
|
2017-01-16 21:05:59 +00:00
|
|
|
{
|
|
|
|
struct dm_file *priv = filp->private_data;
|
2017-07-03 10:39:46 +00:00
|
|
|
__poll_t mask = 0;
|
2017-01-16 21:05:59 +00:00
|
|
|
|
|
|
|
poll_wait(filp, &dm_global_eventq, wait);
|
|
|
|
|
|
|
|
if ((int)(atomic_read(&dm_global_event_nr) - priv->global_event_nr) > 0)
|
2018-02-11 22:34:03 +00:00
|
|
|
mask |= EPOLLIN;
|
2017-01-16 21:05:59 +00:00
|
|
|
|
|
|
|
return mask;
|
|
|
|
}
|
|
|
|
|
2007-02-12 08:55:33 +00:00
|
|
|
static const struct file_operations _ctl_fops = {
|
2017-01-16 21:05:59 +00:00
|
|
|
.open = dm_open,
|
|
|
|
.release = dm_release,
|
|
|
|
.poll = dm_poll,
|
2008-02-08 02:09:53 +00:00
|
|
|
.unlocked_ioctl = dm_ctl_ioctl,
|
2008-02-08 02:09:56 +00:00
|
|
|
.compat_ioctl = dm_compat_ctl_ioctl,
|
2005-04-16 22:20:36 +00:00
|
|
|
.owner = THIS_MODULE,
|
llseek: automatically add .llseek fop
All file_operations should get a .llseek operation so we can make
nonseekable_open the default for future file operations without a
.llseek pointer.
The three cases that we can automatically detect are no_llseek, seq_lseek
and default_llseek. For cases where we can we can automatically prove that
the file offset is always ignored, we use noop_llseek, which maintains
the current behavior of not returning an error from a seek.
New drivers should normally not use noop_llseek but instead use no_llseek
and call nonseekable_open at open time. Existing drivers can be converted
to do the same when the maintainer knows for certain that no user code
relies on calling seek on the device file.
The generated code is often incorrectly indented and right now contains
comments that clarify for each added line why a specific variant was
chosen. In the version that gets submitted upstream, the comments will
be gone and I will manually fix the indentation, because there does not
seem to be a way to do that using coccinelle.
Some amount of new code is currently sitting in linux-next that should get
the same modifications, which I will do at the end of the merge window.
Many thanks to Julia Lawall for helping me learn to write a semantic
patch that does all this.
===== begin semantic patch =====
// This adds an llseek= method to all file operations,
// as a preparation for making no_llseek the default.
//
// The rules are
// - use no_llseek explicitly if we do nonseekable_open
// - use seq_lseek for sequential files
// - use default_llseek if we know we access f_pos
// - use noop_llseek if we know we don't access f_pos,
// but we still want to allow users to call lseek
//
@ open1 exists @
identifier nested_open;
@@
nested_open(...)
{
<+...
nonseekable_open(...)
...+>
}
@ open exists@
identifier open_f;
identifier i, f;
identifier open1.nested_open;
@@
int open_f(struct inode *i, struct file *f)
{
<+...
(
nonseekable_open(...)
|
nested_open(...)
)
...+>
}
@ read disable optional_qualifier exists @
identifier read_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
expression E;
identifier func;
@@
ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off)
{
<+...
(
*off = E
|
*off += E
|
func(..., off, ...)
|
E = *off
)
...+>
}
@ read_no_fpos disable optional_qualifier exists @
identifier read_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
@@
ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off)
{
... when != off
}
@ write @
identifier write_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
expression E;
identifier func;
@@
ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off)
{
<+...
(
*off = E
|
*off += E
|
func(..., off, ...)
|
E = *off
)
...+>
}
@ write_no_fpos @
identifier write_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
@@
ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off)
{
... when != off
}
@ fops0 @
identifier fops;
@@
struct file_operations fops = {
...
};
@ has_llseek depends on fops0 @
identifier fops0.fops;
identifier llseek_f;
@@
struct file_operations fops = {
...
.llseek = llseek_f,
...
};
@ has_read depends on fops0 @
identifier fops0.fops;
identifier read_f;
@@
struct file_operations fops = {
...
.read = read_f,
...
};
@ has_write depends on fops0 @
identifier fops0.fops;
identifier write_f;
@@
struct file_operations fops = {
...
.write = write_f,
...
};
@ has_open depends on fops0 @
identifier fops0.fops;
identifier open_f;
@@
struct file_operations fops = {
...
.open = open_f,
...
};
// use no_llseek if we call nonseekable_open
////////////////////////////////////////////
@ nonseekable1 depends on !has_llseek && has_open @
identifier fops0.fops;
identifier nso ~= "nonseekable_open";
@@
struct file_operations fops = {
... .open = nso, ...
+.llseek = no_llseek, /* nonseekable */
};
@ nonseekable2 depends on !has_llseek @
identifier fops0.fops;
identifier open.open_f;
@@
struct file_operations fops = {
... .open = open_f, ...
+.llseek = no_llseek, /* open uses nonseekable */
};
// use seq_lseek for sequential files
/////////////////////////////////////
@ seq depends on !has_llseek @
identifier fops0.fops;
identifier sr ~= "seq_read";
@@
struct file_operations fops = {
... .read = sr, ...
+.llseek = seq_lseek, /* we have seq_read */
};
// use default_llseek if there is a readdir
///////////////////////////////////////////
@ fops1 depends on !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier readdir_e;
@@
// any other fop is used that changes pos
struct file_operations fops = {
... .readdir = readdir_e, ...
+.llseek = default_llseek, /* readdir is present */
};
// use default_llseek if at least one of read/write touches f_pos
/////////////////////////////////////////////////////////////////
@ fops2 depends on !fops1 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read.read_f;
@@
// read fops use offset
struct file_operations fops = {
... .read = read_f, ...
+.llseek = default_llseek, /* read accesses f_pos */
};
@ fops3 depends on !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier write.write_f;
@@
// write fops use offset
struct file_operations fops = {
... .write = write_f, ...
+ .llseek = default_llseek, /* write accesses f_pos */
};
// Use noop_llseek if neither read nor write accesses f_pos
///////////////////////////////////////////////////////////
@ fops4 depends on !fops1 && !fops2 && !fops3 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read_no_fpos.read_f;
identifier write_no_fpos.write_f;
@@
// write fops use offset
struct file_operations fops = {
...
.write = write_f,
.read = read_f,
...
+.llseek = noop_llseek, /* read and write both use no f_pos */
};
@ depends on has_write && !has_read && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier write_no_fpos.write_f;
@@
struct file_operations fops = {
... .write = write_f, ...
+.llseek = noop_llseek, /* write uses no f_pos */
};
@ depends on has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read_no_fpos.read_f;
@@
struct file_operations fops = {
... .read = read_f, ...
+.llseek = noop_llseek, /* read uses no f_pos */
};
@ depends on !has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
@@
struct file_operations fops = {
...
+.llseek = noop_llseek, /* no read or write fn */
};
===== End semantic patch =====
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Julia Lawall <julia@diku.dk>
Cc: Christoph Hellwig <hch@infradead.org>
2010-08-15 16:52:59 +00:00
|
|
|
.llseek = noop_llseek,
|
2005-04-16 22:20:36 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
static struct miscdevice _dm_misc = {
|
2010-08-12 03:14:05 +00:00
|
|
|
.minor = MAPPER_CTRL_MINOR,
|
2023-02-01 21:31:43 +00:00
|
|
|
.name = DM_NAME,
|
2010-08-12 03:14:05 +00:00
|
|
|
.nodename = DM_DIR "/" DM_CONTROL_NODE,
|
2023-02-01 21:31:43 +00:00
|
|
|
.fops = &_ctl_fops
|
2005-04-16 22:20:36 +00:00
|
|
|
};
|
|
|
|
|
2010-08-12 03:14:05 +00:00
|
|
|
MODULE_ALIAS_MISCDEV(MAPPER_CTRL_MINOR);
|
|
|
|
MODULE_ALIAS("devname:" DM_DIR "/" DM_CONTROL_NODE);
|
|
|
|
|
2005-04-16 22:20:36 +00:00
|
|
|
/*
|
|
|
|
* Create misc character device and link to DM_DIR/control.
|
|
|
|
*/
|
|
|
|
int __init dm_interface_init(void)
|
|
|
|
{
|
|
|
|
int r;
|
|
|
|
|
|
|
|
r = misc_register(&_dm_misc);
|
|
|
|
if (r) {
|
|
|
|
DMERR("misc_register failed for control device");
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
DMINFO("%d.%d.%d%s initialised: %s", DM_VERSION_MAJOR,
|
|
|
|
DM_VERSION_MINOR, DM_VERSION_PATCHLEVEL, DM_VERSION_EXTRA,
|
|
|
|
DM_DRIVER_EMAIL);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
void dm_interface_exit(void)
|
|
|
|
{
|
2015-07-30 22:59:57 +00:00
|
|
|
misc_deregister(&_dm_misc);
|
2005-04-16 22:20:36 +00:00
|
|
|
dm_hash_exit();
|
|
|
|
}
|
2007-10-19 21:47:59 +00:00
|
|
|
|
|
|
|
/**
|
|
|
|
* dm_copy_name_and_uuid - Copy mapped device name & uuid into supplied buffers
|
|
|
|
* @md: Pointer to mapped_device
|
|
|
|
* @name: Buffer (size DM_NAME_LEN) for name
|
|
|
|
* @uuid: Buffer (size DM_UUID_LEN) for uuid or empty string if uuid not defined
|
|
|
|
*/
|
|
|
|
int dm_copy_name_and_uuid(struct mapped_device *md, char *name, char *uuid)
|
|
|
|
{
|
|
|
|
int r = 0;
|
|
|
|
struct hash_cell *hc;
|
|
|
|
|
|
|
|
if (!md)
|
|
|
|
return -ENXIO;
|
|
|
|
|
2009-12-10 23:51:52 +00:00
|
|
|
mutex_lock(&dm_hash_cells_mutex);
|
2007-10-19 21:47:59 +00:00
|
|
|
hc = dm_get_mdptr(md);
|
2022-12-16 04:23:53 +00:00
|
|
|
if (!hc) {
|
2007-10-19 21:47:59 +00:00
|
|
|
r = -ENXIO;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2009-01-06 03:05:04 +00:00
|
|
|
if (name)
|
|
|
|
strcpy(name, hc->name);
|
|
|
|
if (uuid)
|
|
|
|
strcpy(uuid, hc->uuid ? : "");
|
2007-10-19 21:47:59 +00:00
|
|
|
|
|
|
|
out:
|
2009-12-10 23:51:52 +00:00
|
|
|
mutex_unlock(&dm_hash_cells_mutex);
|
2007-10-19 21:47:59 +00:00
|
|
|
|
|
|
|
return r;
|
|
|
|
}
|
2020-10-01 19:00:56 +00:00
|
|
|
EXPORT_SYMBOL_GPL(dm_copy_name_and_uuid);
|
dm: add support to directly boot to a mapped device
Add a "create" module parameter, which allows device-mapper targets to
be configured at boot time. This enables early use of DM targets in the
boot process (as the root device or otherwise) without the need of an
initramfs.
The syntax used in the boot param is based on the concise format from
the dmsetup tool to follow the rule of least surprise:
dmsetup table --concise /dev/mapper/lroot
Which is:
dm-mod.create=<name>,<uuid>,<minor>,<flags>,<table>[,<table>+][;<name>,<uuid>,<minor>,<flags>,<table>[,<table>+]+]
Where,
<name> ::= The device name.
<uuid> ::= xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx | ""
<minor> ::= The device minor number | ""
<flags> ::= "ro" | "rw"
<table> ::= <start_sector> <num_sectors> <target_type> <target_args>
<target_type> ::= "verity" | "linear" | ...
For example, the following could be added in the boot parameters:
dm-mod.create="lroot,,,rw, 0 4096 linear 98:16 0, 4096 4096 linear 98:32 0" root=/dev/dm-0
Only the targets that were tested are allowed and the ones that don't
change any block device when the device is create as read-only. For
example, mirror and cache targets are not allowed. The rationale behind
this is that if the user makes a mistake, choosing the wrong device to
be the mirror or the cache can corrupt data.
The only targets initially allowed are:
* crypt
* delay
* linear
* snapshot-origin
* striped
* verity
Co-developed-by: Will Drewry <wad@chromium.org>
Co-developed-by: Kees Cook <keescook@chromium.org>
Co-developed-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Signed-off-by: Helen Koike <helen.koike@collabora.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
2019-02-21 20:33:34 +00:00
|
|
|
|
|
|
|
/**
|
|
|
|
* dm_early_create - create a mapped device in early boot.
|
|
|
|
*
|
|
|
|
* @dmi: Contains main information of the device mapping to be created.
|
|
|
|
* @spec_array: array of pointers to struct dm_target_spec. Describes the
|
|
|
|
* mapping table of the device.
|
|
|
|
* @target_params_array: array of strings with the parameters to a specific
|
|
|
|
* target.
|
|
|
|
*
|
|
|
|
* Instead of having the struct dm_target_spec and the parameters for every
|
|
|
|
* target embedded at the end of struct dm_ioctl (as performed in a normal
|
|
|
|
* ioctl), pass them as arguments, so the caller doesn't need to serialize them.
|
|
|
|
* The size of the spec_array and target_params_array is given by
|
|
|
|
* @dmi->target_count.
|
|
|
|
* This function is supposed to be called in early boot, so locking mechanisms
|
|
|
|
* to protect against concurrent loads are not required.
|
|
|
|
*/
|
|
|
|
int __init dm_early_create(struct dm_ioctl *dmi,
|
|
|
|
struct dm_target_spec **spec_array,
|
|
|
|
char **target_params_array)
|
|
|
|
{
|
|
|
|
int r, m = DM_ANY_MINOR;
|
|
|
|
struct dm_table *t, *old_map;
|
|
|
|
struct mapped_device *md;
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
if (!dmi->target_count)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
r = check_name(dmi->name);
|
|
|
|
if (r)
|
|
|
|
return r;
|
|
|
|
|
|
|
|
if (dmi->flags & DM_PERSISTENT_DEV_FLAG)
|
|
|
|
m = MINOR(huge_decode_dev(dmi->dev));
|
|
|
|
|
|
|
|
/* alloc dm device */
|
|
|
|
r = dm_create(m, &md);
|
|
|
|
if (r)
|
|
|
|
return r;
|
|
|
|
|
|
|
|
/* hash insert */
|
|
|
|
r = dm_hash_insert(dmi->name, *dmi->uuid ? dmi->uuid : NULL, md);
|
|
|
|
if (r)
|
|
|
|
goto err_destroy_dm;
|
|
|
|
|
|
|
|
/* alloc table */
|
|
|
|
r = dm_table_create(&t, get_mode(dmi), dmi->target_count, md);
|
|
|
|
if (r)
|
2019-05-15 16:50:54 +00:00
|
|
|
goto err_hash_remove;
|
dm: add support to directly boot to a mapped device
Add a "create" module parameter, which allows device-mapper targets to
be configured at boot time. This enables early use of DM targets in the
boot process (as the root device or otherwise) without the need of an
initramfs.
The syntax used in the boot param is based on the concise format from
the dmsetup tool to follow the rule of least surprise:
dmsetup table --concise /dev/mapper/lroot
Which is:
dm-mod.create=<name>,<uuid>,<minor>,<flags>,<table>[,<table>+][;<name>,<uuid>,<minor>,<flags>,<table>[,<table>+]+]
Where,
<name> ::= The device name.
<uuid> ::= xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx | ""
<minor> ::= The device minor number | ""
<flags> ::= "ro" | "rw"
<table> ::= <start_sector> <num_sectors> <target_type> <target_args>
<target_type> ::= "verity" | "linear" | ...
For example, the following could be added in the boot parameters:
dm-mod.create="lroot,,,rw, 0 4096 linear 98:16 0, 4096 4096 linear 98:32 0" root=/dev/dm-0
Only the targets that were tested are allowed and the ones that don't
change any block device when the device is create as read-only. For
example, mirror and cache targets are not allowed. The rationale behind
this is that if the user makes a mistake, choosing the wrong device to
be the mirror or the cache can corrupt data.
The only targets initially allowed are:
* crypt
* delay
* linear
* snapshot-origin
* striped
* verity
Co-developed-by: Will Drewry <wad@chromium.org>
Co-developed-by: Kees Cook <keescook@chromium.org>
Co-developed-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Signed-off-by: Helen Koike <helen.koike@collabora.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
2019-02-21 20:33:34 +00:00
|
|
|
|
|
|
|
/* add targets */
|
|
|
|
for (i = 0; i < dmi->target_count; i++) {
|
|
|
|
r = dm_table_add_target(t, spec_array[i]->target_type,
|
|
|
|
(sector_t) spec_array[i]->sector_start,
|
|
|
|
(sector_t) spec_array[i]->length,
|
|
|
|
target_params_array[i]);
|
|
|
|
if (r) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("error adding target to table");
|
dm: add support to directly boot to a mapped device
Add a "create" module parameter, which allows device-mapper targets to
be configured at boot time. This enables early use of DM targets in the
boot process (as the root device or otherwise) without the need of an
initramfs.
The syntax used in the boot param is based on the concise format from
the dmsetup tool to follow the rule of least surprise:
dmsetup table --concise /dev/mapper/lroot
Which is:
dm-mod.create=<name>,<uuid>,<minor>,<flags>,<table>[,<table>+][;<name>,<uuid>,<minor>,<flags>,<table>[,<table>+]+]
Where,
<name> ::= The device name.
<uuid> ::= xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx | ""
<minor> ::= The device minor number | ""
<flags> ::= "ro" | "rw"
<table> ::= <start_sector> <num_sectors> <target_type> <target_args>
<target_type> ::= "verity" | "linear" | ...
For example, the following could be added in the boot parameters:
dm-mod.create="lroot,,,rw, 0 4096 linear 98:16 0, 4096 4096 linear 98:32 0" root=/dev/dm-0
Only the targets that were tested are allowed and the ones that don't
change any block device when the device is create as read-only. For
example, mirror and cache targets are not allowed. The rationale behind
this is that if the user makes a mistake, choosing the wrong device to
be the mirror or the cache can corrupt data.
The only targets initially allowed are:
* crypt
* delay
* linear
* snapshot-origin
* striped
* verity
Co-developed-by: Will Drewry <wad@chromium.org>
Co-developed-by: Kees Cook <keescook@chromium.org>
Co-developed-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Signed-off-by: Helen Koike <helen.koike@collabora.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
2019-02-21 20:33:34 +00:00
|
|
|
goto err_destroy_table;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* finish table */
|
|
|
|
r = dm_table_complete(t);
|
|
|
|
if (r)
|
|
|
|
goto err_destroy_table;
|
|
|
|
|
|
|
|
/* setup md->queue to reflect md's type (may block) */
|
|
|
|
r = dm_setup_md_queue(md, t);
|
|
|
|
if (r) {
|
2022-08-24 11:25:57 +00:00
|
|
|
DMERR("unable to set up device queue for new table.");
|
dm: add support to directly boot to a mapped device
Add a "create" module parameter, which allows device-mapper targets to
be configured at boot time. This enables early use of DM targets in the
boot process (as the root device or otherwise) without the need of an
initramfs.
The syntax used in the boot param is based on the concise format from
the dmsetup tool to follow the rule of least surprise:
dmsetup table --concise /dev/mapper/lroot
Which is:
dm-mod.create=<name>,<uuid>,<minor>,<flags>,<table>[,<table>+][;<name>,<uuid>,<minor>,<flags>,<table>[,<table>+]+]
Where,
<name> ::= The device name.
<uuid> ::= xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx | ""
<minor> ::= The device minor number | ""
<flags> ::= "ro" | "rw"
<table> ::= <start_sector> <num_sectors> <target_type> <target_args>
<target_type> ::= "verity" | "linear" | ...
For example, the following could be added in the boot parameters:
dm-mod.create="lroot,,,rw, 0 4096 linear 98:16 0, 4096 4096 linear 98:32 0" root=/dev/dm-0
Only the targets that were tested are allowed and the ones that don't
change any block device when the device is create as read-only. For
example, mirror and cache targets are not allowed. The rationale behind
this is that if the user makes a mistake, choosing the wrong device to
be the mirror or the cache can corrupt data.
The only targets initially allowed are:
* crypt
* delay
* linear
* snapshot-origin
* striped
* verity
Co-developed-by: Will Drewry <wad@chromium.org>
Co-developed-by: Kees Cook <keescook@chromium.org>
Co-developed-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Signed-off-by: Helen Koike <helen.koike@collabora.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
2019-02-21 20:33:34 +00:00
|
|
|
goto err_destroy_table;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Set new map */
|
|
|
|
dm_suspend(md, 0);
|
|
|
|
old_map = dm_swap_table(md, t);
|
|
|
|
if (IS_ERR(old_map)) {
|
|
|
|
r = PTR_ERR(old_map);
|
|
|
|
goto err_destroy_table;
|
|
|
|
}
|
|
|
|
set_disk_ro(dm_disk(md), !!(dmi->flags & DM_READONLY_FLAG));
|
|
|
|
|
|
|
|
/* resume device */
|
|
|
|
r = dm_resume(md);
|
|
|
|
if (r)
|
|
|
|
goto err_destroy_table;
|
|
|
|
|
|
|
|
DMINFO("%s (%s) is ready", md->disk->disk_name, dmi->name);
|
|
|
|
dm_put(md);
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
err_destroy_table:
|
|
|
|
dm_table_destroy(t);
|
2019-05-15 16:50:54 +00:00
|
|
|
err_hash_remove:
|
2023-02-17 18:08:17 +00:00
|
|
|
down_write(&_hash_lock);
|
2019-05-15 16:50:54 +00:00
|
|
|
(void) __hash_remove(__get_name_cell(dmi->name));
|
2023-02-17 18:08:17 +00:00
|
|
|
up_write(&_hash_lock);
|
2019-05-15 16:50:54 +00:00
|
|
|
/* release reference from __get_name_cell */
|
|
|
|
dm_put(md);
|
dm: add support to directly boot to a mapped device
Add a "create" module parameter, which allows device-mapper targets to
be configured at boot time. This enables early use of DM targets in the
boot process (as the root device or otherwise) without the need of an
initramfs.
The syntax used in the boot param is based on the concise format from
the dmsetup tool to follow the rule of least surprise:
dmsetup table --concise /dev/mapper/lroot
Which is:
dm-mod.create=<name>,<uuid>,<minor>,<flags>,<table>[,<table>+][;<name>,<uuid>,<minor>,<flags>,<table>[,<table>+]+]
Where,
<name> ::= The device name.
<uuid> ::= xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx | ""
<minor> ::= The device minor number | ""
<flags> ::= "ro" | "rw"
<table> ::= <start_sector> <num_sectors> <target_type> <target_args>
<target_type> ::= "verity" | "linear" | ...
For example, the following could be added in the boot parameters:
dm-mod.create="lroot,,,rw, 0 4096 linear 98:16 0, 4096 4096 linear 98:32 0" root=/dev/dm-0
Only the targets that were tested are allowed and the ones that don't
change any block device when the device is create as read-only. For
example, mirror and cache targets are not allowed. The rationale behind
this is that if the user makes a mistake, choosing the wrong device to
be the mirror or the cache can corrupt data.
The only targets initially allowed are:
* crypt
* delay
* linear
* snapshot-origin
* striped
* verity
Co-developed-by: Will Drewry <wad@chromium.org>
Co-developed-by: Kees Cook <keescook@chromium.org>
Co-developed-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Signed-off-by: Helen Koike <helen.koike@collabora.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
2019-02-21 20:33:34 +00:00
|
|
|
err_destroy_dm:
|
|
|
|
dm_put(md);
|
|
|
|
dm_destroy(md);
|
|
|
|
return r;
|
|
|
|
}
|