linux/drivers/md/dm-ioctl.c

1674 lines
35 KiB
C
Raw Normal View History

/*
* Copyright (C) 2001, 2002 Sistina Software (UK) Limited.
* Copyright (C) 2004 - 2006 Red Hat, Inc. All rights reserved.
*
* This file is released under the GPL.
*/
#include "dm.h"
#include <linux/module.h>
#include <linux/vmalloc.h>
#include <linux/miscdevice.h>
#include <linux/init.h>
#include <linux/wait.h>
#include <linux/slab.h>
#include <linux/dm-ioctl.h>
#include <linux/hdreg.h>
#include <linux/compat.h>
#include <asm/uaccess.h>
#define DM_MSG_PREFIX "ioctl"
#define DM_DRIVER_EMAIL "dm-devel@redhat.com"
/*-----------------------------------------------------------------
* The ioctl interface needs to be able to look up devices by
* name or uuid.
*---------------------------------------------------------------*/
struct hash_cell {
struct list_head name_list;
struct list_head uuid_list;
char *name;
char *uuid;
struct mapped_device *md;
struct dm_table *new_map;
};
struct vers_iter {
size_t param_size;
struct dm_target_versions *vers, *old_vers;
char *end;
uint32_t flags;
};
#define NUM_BUCKETS 64
#define MASK_BUCKETS (NUM_BUCKETS - 1)
static struct list_head _name_buckets[NUM_BUCKETS];
static struct list_head _uuid_buckets[NUM_BUCKETS];
static void dm_hash_remove_all(int keep_open_devices);
/*
* Guards access to both hash tables.
*/
static DECLARE_RWSEM(_hash_lock);
dm: avoid _hash_lock deadlock Fix a reported deadlock if there are still unprocessed multipath events on a device that is being removed. _hash_lock is held during dev_remove while trying to send the outstanding events. Sending the events requests the _hash_lock again in dm_copy_name_and_uuid. This patch introduces a separate lock around regions that modify the link to the hash table (dm_set_mdptr) or the name or uuid so that dm_copy_name_and_uuid no longer needs _hash_lock. Additionally, dm_copy_name_and_uuid can only be called if md exists so we can drop the dm_get() and dm_put() which can lead to a BUG() while md is being freed. The deadlock: #0 [ffff8106298dfb48] schedule at ffffffff80063035 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad #7 [ffff8106298dfd30] dev_remove at ffffffff88250865 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4 #10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9 #11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf #12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call) Cc: stable@kernel.org Reported-by: guy keren <choo@actcom.co.il> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-12-11 07:51:52 +08:00
/*
* Protects use of mdptr to obtain hash cell name and uuid from mapped device.
*/
static DEFINE_MUTEX(dm_hash_cells_mutex);
static void init_buckets(struct list_head *buckets)
{
unsigned int i;
for (i = 0; i < NUM_BUCKETS; i++)
INIT_LIST_HEAD(buckets + i);
}
static int dm_hash_init(void)
{
init_buckets(_name_buckets);
init_buckets(_uuid_buckets);
return 0;
}
static void dm_hash_exit(void)
{
dm_hash_remove_all(0);
}
/*-----------------------------------------------------------------
* Hash function:
* We're not really concerned with the str hash function being
* fast since it's only used by the ioctl interface.
*---------------------------------------------------------------*/
static unsigned int hash_str(const char *str)
{
const unsigned int hash_mult = 2654435387U;
unsigned int h = 0;
while (*str)
h = (h + (unsigned int) *str++) * hash_mult;
return h & MASK_BUCKETS;
}
/*-----------------------------------------------------------------
* Code for looking up a device by name
*---------------------------------------------------------------*/
static struct hash_cell *__get_name_cell(const char *str)
{
struct hash_cell *hc;
unsigned int h = hash_str(str);
list_for_each_entry (hc, _name_buckets + h, name_list)
if (!strcmp(hc->name, str)) {
dm_get(hc->md);
return hc;
}
return NULL;
}
static struct hash_cell *__get_uuid_cell(const char *str)
{
struct hash_cell *hc;
unsigned int h = hash_str(str);
list_for_each_entry (hc, _uuid_buckets + h, uuid_list)
if (!strcmp(hc->uuid, str)) {
dm_get(hc->md);
return hc;
}
return NULL;
}
/*-----------------------------------------------------------------
* Inserting, removing and renaming a device.
*---------------------------------------------------------------*/
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;
hc->name = kstrdup(name, GFP_KERNEL);
if (!hc->name) {
kfree(hc);
return NULL;
}
if (!uuid)
hc->uuid = NULL;
else {
hc->uuid = kstrdup(uuid, GFP_KERNEL);
if (!hc->uuid) {
kfree(hc->name);
kfree(hc);
return NULL;
}
}
INIT_LIST_HEAD(&hc->name_list);
INIT_LIST_HEAD(&hc->uuid_list);
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)
{
struct hash_cell *cell, *hc;
/*
* 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);
hc = __get_name_cell(name);
if (hc) {
dm_put(hc->md);
goto bad;
}
list_add(&cell->name_list, _name_buckets + hash_str(name));
if (uuid) {
hc = __get_uuid_cell(uuid);
if (hc) {
list_del(&cell->name_list);
dm_put(hc->md);
goto bad;
}
list_add(&cell->uuid_list, _uuid_buckets + hash_str(uuid));
}
dm_get(md);
dm: avoid _hash_lock deadlock Fix a reported deadlock if there are still unprocessed multipath events on a device that is being removed. _hash_lock is held during dev_remove while trying to send the outstanding events. Sending the events requests the _hash_lock again in dm_copy_name_and_uuid. This patch introduces a separate lock around regions that modify the link to the hash table (dm_set_mdptr) or the name or uuid so that dm_copy_name_and_uuid no longer needs _hash_lock. Additionally, dm_copy_name_and_uuid can only be called if md exists so we can drop the dm_get() and dm_put() which can lead to a BUG() while md is being freed. The deadlock: #0 [ffff8106298dfb48] schedule at ffffffff80063035 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad #7 [ffff8106298dfd30] dev_remove at ffffffff88250865 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4 #10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9 #11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf #12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call) Cc: stable@kernel.org Reported-by: guy keren <choo@actcom.co.il> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-12-11 07:51:52 +08:00
mutex_lock(&dm_hash_cells_mutex);
dm_set_mdptr(md, cell);
dm: avoid _hash_lock deadlock Fix a reported deadlock if there are still unprocessed multipath events on a device that is being removed. _hash_lock is held during dev_remove while trying to send the outstanding events. Sending the events requests the _hash_lock again in dm_copy_name_and_uuid. This patch introduces a separate lock around regions that modify the link to the hash table (dm_set_mdptr) or the name or uuid so that dm_copy_name_and_uuid no longer needs _hash_lock. Additionally, dm_copy_name_and_uuid can only be called if md exists so we can drop the dm_get() and dm_put() which can lead to a BUG() while md is being freed. The deadlock: #0 [ffff8106298dfb48] schedule at ffffffff80063035 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad #7 [ffff8106298dfd30] dev_remove at ffffffff88250865 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4 #10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9 #11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf #12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call) Cc: stable@kernel.org Reported-by: guy keren <choo@actcom.co.il> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-12-11 07:51:52 +08:00
mutex_unlock(&dm_hash_cells_mutex);
up_write(&_hash_lock);
return 0;
bad:
up_write(&_hash_lock);
free_cell(cell);
return -EBUSY;
}
static void __hash_remove(struct hash_cell *hc)
{
struct dm_table *table;
/* remove from the dev hash */
list_del(&hc->uuid_list);
list_del(&hc->name_list);
dm: avoid _hash_lock deadlock Fix a reported deadlock if there are still unprocessed multipath events on a device that is being removed. _hash_lock is held during dev_remove while trying to send the outstanding events. Sending the events requests the _hash_lock again in dm_copy_name_and_uuid. This patch introduces a separate lock around regions that modify the link to the hash table (dm_set_mdptr) or the name or uuid so that dm_copy_name_and_uuid no longer needs _hash_lock. Additionally, dm_copy_name_and_uuid can only be called if md exists so we can drop the dm_get() and dm_put() which can lead to a BUG() while md is being freed. The deadlock: #0 [ffff8106298dfb48] schedule at ffffffff80063035 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad #7 [ffff8106298dfd30] dev_remove at ffffffff88250865 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4 #10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9 #11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf #12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call) Cc: stable@kernel.org Reported-by: guy keren <choo@actcom.co.il> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-12-11 07:51:52 +08:00
mutex_lock(&dm_hash_cells_mutex);
dm_set_mdptr(hc->md, NULL);
dm: avoid _hash_lock deadlock Fix a reported deadlock if there are still unprocessed multipath events on a device that is being removed. _hash_lock is held during dev_remove while trying to send the outstanding events. Sending the events requests the _hash_lock again in dm_copy_name_and_uuid. This patch introduces a separate lock around regions that modify the link to the hash table (dm_set_mdptr) or the name or uuid so that dm_copy_name_and_uuid no longer needs _hash_lock. Additionally, dm_copy_name_and_uuid can only be called if md exists so we can drop the dm_get() and dm_put() which can lead to a BUG() while md is being freed. The deadlock: #0 [ffff8106298dfb48] schedule at ffffffff80063035 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad #7 [ffff8106298dfd30] dev_remove at ffffffff88250865 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4 #10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9 #11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf #12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call) Cc: stable@kernel.org Reported-by: guy keren <choo@actcom.co.il> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-12-11 07:51:52 +08:00
mutex_unlock(&dm_hash_cells_mutex);
table = dm_get_live_table(hc->md);
if (table) {
dm_table_event(table);
dm_table_put(table);
}
if (hc->new_map)
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 11:05:10 +08:00
dm_table_destroy(hc->new_map);
dm_put(hc->md);
free_cell(hc);
}
static void dm_hash_remove_all(int keep_open_devices)
{
int i, dev_skipped;
struct hash_cell *hc;
struct mapped_device *md;
retry:
dev_skipped = 0;
down_write(&_hash_lock);
for (i = 0; i < NUM_BUCKETS; i++) {
list_for_each_entry(hc, _name_buckets + i, name_list) {
md = hc->md;
dm_get(md);
if (keep_open_devices && dm_lock_for_deletion(md)) {
dm_put(md);
dev_skipped++;
continue;
}
__hash_remove(hc);
up_write(&_hash_lock);
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 11:13:56 +08:00
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;
}
}
up_write(&_hash_lock);
if (dev_skipped)
DMWARN("remove_all left %d open device(s)", dev_skipped);
}
static struct mapped_device *dm_hash_rename(struct dm_ioctl *param,
const char *new)
{
char *new_name, *old_name;
struct hash_cell *hc;
struct dm_table *table;
struct mapped_device *md;
/*
* duplicate new.
*/
new_name = kstrdup(new, GFP_KERNEL);
if (!new_name)
return ERR_PTR(-ENOMEM);
down_write(&_hash_lock);
/*
* Is new free ?
*/
hc = __get_name_cell(new);
if (hc) {
DMWARN("asked to rename to an already-existing name %s -> %s",
param->name, new);
dm_put(hc->md);
up_write(&_hash_lock);
kfree(new_name);
return ERR_PTR(-EBUSY);
}
/*
* Is there such a device as 'old' ?
*/
hc = __get_name_cell(param->name);
if (!hc) {
DMWARN("asked to rename a non-existent device %s -> %s",
param->name, new);
up_write(&_hash_lock);
kfree(new_name);
return ERR_PTR(-ENXIO);
}
/*
* rename and move the name cell.
*/
list_del(&hc->name_list);
old_name = hc->name;
dm: avoid _hash_lock deadlock Fix a reported deadlock if there are still unprocessed multipath events on a device that is being removed. _hash_lock is held during dev_remove while trying to send the outstanding events. Sending the events requests the _hash_lock again in dm_copy_name_and_uuid. This patch introduces a separate lock around regions that modify the link to the hash table (dm_set_mdptr) or the name or uuid so that dm_copy_name_and_uuid no longer needs _hash_lock. Additionally, dm_copy_name_and_uuid can only be called if md exists so we can drop the dm_get() and dm_put() which can lead to a BUG() while md is being freed. The deadlock: #0 [ffff8106298dfb48] schedule at ffffffff80063035 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad #7 [ffff8106298dfd30] dev_remove at ffffffff88250865 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4 #10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9 #11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf #12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call) Cc: stable@kernel.org Reported-by: guy keren <choo@actcom.co.il> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-12-11 07:51:52 +08:00
mutex_lock(&dm_hash_cells_mutex);
hc->name = new_name;
dm: avoid _hash_lock deadlock Fix a reported deadlock if there are still unprocessed multipath events on a device that is being removed. _hash_lock is held during dev_remove while trying to send the outstanding events. Sending the events requests the _hash_lock again in dm_copy_name_and_uuid. This patch introduces a separate lock around regions that modify the link to the hash table (dm_set_mdptr) or the name or uuid so that dm_copy_name_and_uuid no longer needs _hash_lock. Additionally, dm_copy_name_and_uuid can only be called if md exists so we can drop the dm_get() and dm_put() which can lead to a BUG() while md is being freed. The deadlock: #0 [ffff8106298dfb48] schedule at ffffffff80063035 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad #7 [ffff8106298dfd30] dev_remove at ffffffff88250865 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4 #10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9 #11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf #12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call) Cc: stable@kernel.org Reported-by: guy keren <choo@actcom.co.il> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-12-11 07:51:52 +08:00
mutex_unlock(&dm_hash_cells_mutex);
list_add(&hc->name_list, _name_buckets + hash_str(new_name));
/*
* Wake up any dm event waiters.
*/
table = dm_get_live_table(hc->md);
if (table) {
dm_table_event(table);
dm_table_put(table);
}
if (!dm_kobject_uevent(hc->md, KOBJ_CHANGE, param->event_nr))
param->flags |= DM_UEVENT_GENERATED_FLAG;
md = hc->md;
up_write(&_hash_lock);
kfree(old_name);
return md;
}
/*-----------------------------------------------------------------
* Implementation of the ioctl commands
*---------------------------------------------------------------*/
/*
* All the ioctl commands get dispatched to functions with this
* prototype.
*/
typedef int (*ioctl_fn)(struct dm_ioctl *param, size_t param_size);
static int remove_all(struct dm_ioctl *param, size_t param_size)
{
dm_hash_remove_all(1);
param->data_size = 0;
return 0;
}
/*
* Round up the ptr to an 8-byte boundary.
*/
#define ALIGN_MASK 7
static inline void *align_ptr(void *ptr)
{
return (void *) (((size_t) (ptr + ALIGN_MASK)) & ~ALIGN_MASK);
}
/*
* 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;
}
static int list_devices(struct dm_ioctl *param, size_t param_size)
{
unsigned int i;
struct hash_cell *hc;
size_t len, needed = 0;
struct gendisk *disk;
struct dm_name_list *nl, *old_nl = NULL;
down_write(&_hash_lock);
/*
* Loop through all the devices working out how much
* space we need.
*/
for (i = 0; i < NUM_BUCKETS; i++) {
list_for_each_entry (hc, _name_buckets + i, name_list) {
needed += sizeof(struct dm_name_list);
needed += strlen(hc->name) + 1;
needed += ALIGN_MASK;
}
}
/*
* Grab our output buffer.
*/
nl = 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;
nl->dev = 0; /* Flags no data */
/*
* Now loop through filling out the names.
*/
for (i = 0; i < NUM_BUCKETS; i++) {
list_for_each_entry (hc, _name_buckets + i, name_list) {
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;
nl = align_ptr(((void *) ++nl) + strlen(hc->name) + 1);
}
}
out:
up_write(&_hash_lock);
return 0;
}
static void list_version_get_needed(struct target_type *tt, void *needed_param)
{
size_t *needed = needed_param;
*needed += sizeof(struct dm_target_versions);
*needed += strlen(tt->name);
*needed += ALIGN_MASK;
}
static void list_version_get_info(struct target_type *tt, void *param)
{
struct vers_iter *info = param;
/* 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;
}
if (info->old_vers)
info->old_vers->next = (uint32_t) ((void *)info->vers -
(void *)info->old_vers);
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);
info->old_vers = info->vers;
info->vers = align_ptr(((void *) ++info->vers) + strlen(tt->name) + 1);
}
static int list_versions(struct dm_ioctl *param, size_t param_size)
{
size_t len, needed = 0;
struct dm_target_versions *vers;
struct vers_iter iter_info;
/*
* Loop through all the devices working out how much
* space we need.
*/
dm_target_iterate(list_version_get_needed, &needed);
/*
* 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;
iter_info.end = (char *)vers+len;
/*
* Now loop through filling out the names & versions.
*/
dm_target_iterate(list_version_get_info, &iter_info);
param->flags |= iter_info.flags;
out:
return 0;
}
static int check_name(const char *name)
{
if (strchr(name, '/')) {
DMWARN("invalid device name");
return -EINVAL;
}
return 0;
}
/*
* On successful return, the caller must not attempt to acquire
* _hash_lock without first calling dm_table_put, because dm_table_destroy
* waits for this dm_table_put and could be called under this lock.
*/
static struct dm_table *dm_get_inactive_table(struct mapped_device *md)
{
struct hash_cell *hc;
struct dm_table *table = NULL;
down_read(&_hash_lock);
hc = dm_get_mdptr(md);
if (!hc || hc->md != md) {
DMWARN("device has been removed from the dev hash table.");
goto out;
}
table = hc->new_map;
if (table)
dm_table_get(table);
out:
up_read(&_hash_lock);
return table;
}
static struct dm_table *dm_get_live_or_inactive_table(struct mapped_device *md,
struct dm_ioctl *param)
{
return (param->flags & DM_QUERY_INACTIVE_TABLE_FLAG) ?
dm_get_inactive_table(md) : dm_get_live_table(md);
}
/*
* Fills in a dm_ioctl structure, ready for sending back to
* userland.
*/
static void __dev_status(struct mapped_device *md, struct dm_ioctl *param)
{
struct gendisk *disk = dm_disk(md);
struct dm_table *table;
param->flags &= ~(DM_SUSPEND_FLAG | DM_READONLY_FLAG |
DM_ACTIVE_PRESENT_FLAG);
if (dm_suspended_md(md))
param->flags |= DM_SUSPEND_FLAG;
param->dev = huge_encode_dev(disk_devt(disk));
/*
* 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);
param->event_nr = dm_get_event_nr(md);
param->target_count = 0;
table = dm_get_live_table(md);
if (table) {
if (!(param->flags & DM_QUERY_INACTIVE_TABLE_FLAG)) {
if (get_disk_ro(disk))
param->flags |= DM_READONLY_FLAG;
param->target_count = dm_table_get_num_targets(table);
}
dm_table_put(table);
param->flags |= DM_ACTIVE_PRESENT_FLAG;
}
if (param->flags & DM_QUERY_INACTIVE_TABLE_FLAG) {
table = dm_get_inactive_table(md);
if (table) {
if (!(dm_table_get_mode(table) & FMODE_WRITE))
param->flags |= DM_READONLY_FLAG;
param->target_count = dm_table_get_num_targets(table);
dm_table_put(table);
}
}
}
static int dev_create(struct dm_ioctl *param, size_t param_size)
{
int r, m = DM_ANY_MINOR;
struct mapped_device *md;
r = check_name(param->name);
if (r)
return r;
if (param->flags & DM_PERSISTENT_DEV_FLAG)
m = MINOR(huge_decode_dev(param->dev));
r = dm_create(m, &md);
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 11:13:56 +08:00
if (r) {
dm_put(md);
dm_destroy(md);
return r;
}
param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
__dev_status(md, param);
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 11:13:56 +08:00
return 0;
}
/*
* Always use UUID for lookups if it's present, otherwise use name or dev.
*/
static struct hash_cell *__find_device_hash_cell(struct dm_ioctl *param)
{
struct mapped_device *md;
void *mdptr = NULL;
if (*param->uuid)
return __get_uuid_cell(param->uuid);
if (*param->name)
return __get_name_cell(param->name);
md = dm_get_md(huge_decode_dev(param->dev));
if (!md)
goto out;
mdptr = dm_get_mdptr(md);
if (!mdptr)
dm_put(md);
out:
return mdptr;
}
static struct mapped_device *find_device(struct dm_ioctl *param)
{
struct hash_cell *hc;
struct mapped_device *md = NULL;
down_read(&_hash_lock);
hc = __find_device_hash_cell(param);
if (hc) {
md = hc->md;
/*
* Sneakily write in both the name and the uuid
* while we have the cell.
*/
strlcpy(param->name, hc->name, sizeof(param->name));
if (hc->uuid)
strlcpy(param->uuid, hc->uuid, sizeof(param->uuid));
else
param->uuid[0] = '\0';
if (hc->new_map)
param->flags |= DM_INACTIVE_PRESENT_FLAG;
else
param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
}
up_read(&_hash_lock);
return md;
}
static int dev_remove(struct dm_ioctl *param, size_t param_size)
{
struct hash_cell *hc;
struct mapped_device *md;
int r;
down_write(&_hash_lock);
hc = __find_device_hash_cell(param);
if (!hc) {
DMWARN("device doesn't appear to be in the dev hash table.");
up_write(&_hash_lock);
return -ENXIO;
}
md = hc->md;
/*
* Ensure the device is not open and nothing further can open it.
*/
r = dm_lock_for_deletion(md);
if (r) {
DMWARN("unable to remove open device %s", hc->name);
up_write(&_hash_lock);
dm_put(md);
return r;
}
__hash_remove(hc);
up_write(&_hash_lock);
if (!dm_kobject_uevent(md, KOBJ_REMOVE, param->event_nr))
param->flags |= DM_UEVENT_GENERATED_FLAG;
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 11:13:56 +08:00
dm_destroy(md);
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;
}
static int dev_rename(struct dm_ioctl *param, size_t param_size)
{
int r;
char *new_name = (char *) param + param->data_start;
struct mapped_device *md;
if (new_name < param->data ||
invalid_str(new_name, (void *) param + param_size) ||
strlen(new_name) > DM_NAME_LEN - 1) {
DMWARN("Invalid new logical volume name supplied.");
return -EINVAL;
}
r = check_name(new_name);
if (r)
return r;
md = dm_hash_rename(param, new_name);
if (IS_ERR(md))
return PTR_ERR(md);
__dev_status(md, param);
dm_put(md);
return 0;
}
static int dev_set_geometry(struct dm_ioctl *param, size_t param_size)
{
int r = -EINVAL, x;
struct mapped_device *md;
struct hd_geometry geometry;
unsigned long indata[4];
char *geostr = (char *) param + param->data_start;
md = find_device(param);
if (!md)
return -ENXIO;
if (geostr < param->data ||
invalid_str(geostr, (void *) param + param_size)) {
DMWARN("Invalid geometry supplied.");
goto out;
}
x = sscanf(geostr, "%lu %lu %lu %lu", indata,
indata + 1, indata + 2, indata + 3);
if (x != 4) {
DMWARN("Unable to interpret geometry settings.");
goto out;
}
if (indata[0] > 65535 || indata[1] > 255 ||
indata[2] > 255 || indata[3] > ULONG_MAX) {
DMWARN("Geometry exceeds range limits.");
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;
}
static int do_suspend(struct dm_ioctl *param)
{
int r = 0;
unsigned suspend_flags = DM_SUSPEND_LOCKFS_FLAG;
struct mapped_device *md;
md = find_device(param);
if (!md)
return -ENXIO;
if (param->flags & DM_SKIP_LOCKFS_FLAG)
suspend_flags &= ~DM_SUSPEND_LOCKFS_FLAG;
if (param->flags & DM_NOFLUSH_FLAG)
suspend_flags |= DM_SUSPEND_NOFLUSH_FLAG;
if (!dm_suspended_md(md)) {
r = dm_suspend(md, suspend_flags);
if (r)
goto out;
}
__dev_status(md, param);
out:
dm_put(md);
return r;
}
static int do_resume(struct dm_ioctl *param)
{
int r = 0;
unsigned suspend_flags = DM_SUSPEND_LOCKFS_FLAG;
struct hash_cell *hc;
struct mapped_device *md;
struct dm_table *new_map, *old_map = NULL;
down_write(&_hash_lock);
hc = __find_device_hash_cell(param);
if (!hc) {
DMWARN("device doesn't appear to be in the dev hash table.");
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) {
/* Suspend if it isn't already suspended */
if (param->flags & DM_SKIP_LOCKFS_FLAG)
suspend_flags &= ~DM_SUSPEND_LOCKFS_FLAG;
if (param->flags & DM_NOFLUSH_FLAG)
suspend_flags |= DM_SUSPEND_NOFLUSH_FLAG;
if (!dm_suspended_md(md))
dm_suspend(md, suspend_flags);
old_map = dm_swap_table(md, new_map);
if (IS_ERR(old_map)) {
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 11:05:10 +08:00
dm_table_destroy(new_map);
dm_put(md);
return PTR_ERR(old_map);
}
if (dm_table_get_mode(new_map) & FMODE_WRITE)
set_disk_ro(dm_disk(md), 0);
else
set_disk_ro(dm_disk(md), 1);
}
if (dm_suspended_md(md)) {
r = dm_resume(md);
if (!r && !dm_kobject_uevent(md, KOBJ_CHANGE, param->event_nr))
param->flags |= DM_UEVENT_GENERATED_FLAG;
}
if (old_map)
dm_table_destroy(old_map);
if (!r)
__dev_status(md, param);
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.
*/
static int dev_suspend(struct dm_ioctl *param, size_t param_size)
{
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.
*/
static int dev_status(struct dm_ioctl *param, size_t param_size)
{
struct mapped_device *md;
md = find_device(param);
if (!md)
return -ENXIO;
__dev_status(md, param);
dm_put(md);
return 0;
}
/*
* 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;
outptr = outbuf = get_result_buffer(param, param_size, &len);
if (param->flags & DM_STATUS_TABLE_FLAG)
type = STATUSTYPE_TABLE;
else
type = STATUSTYPE_INFO;
/* Get all the target info */
num_targets = dm_table_get_num_targets(table);
for (i = 0; i < num_targets; i++) {
struct dm_target *ti = dm_table_get_target(table, i);
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;
strncpy(spec->target_type, ti->type->name,
sizeof(spec->target_type));
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) {
if (ti->type->status(ti, type, outptr, remaining)) {
param->flags |= DM_BUFFER_FULL_FLAG;
break;
}
} else
outptr[0] = '\0';
outptr += strlen(outptr) + 1;
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
*/
static int dev_wait(struct dm_ioctl *param, size_t param_size)
{
int r = 0;
struct mapped_device *md;
struct dm_table *table;
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.
*/
__dev_status(md, param);
table = dm_get_live_or_inactive_table(md, param);
if (table) {
retrieve_status(table, param, param_size);
dm_table_put(table);
}
out:
dm_put(md);
return r;
}
static inline fmode_t get_mode(struct dm_ioctl *param)
{
fmode_t mode = FMODE_READ | FMODE_WRITE;
if (param->flags & DM_READONLY_FLAG)
mode = FMODE_READ;
return mode;
}
static int next_target(struct dm_target_spec *last, uint32_t next, void *end,
struct dm_target_spec **spec, char **target_params)
{
*spec = (struct dm_target_spec *) ((unsigned char *) last + next);
*target_params = (char *) (*spec + 1);
if (*spec < (last + 1))
return -EINVAL;
return invalid_str(*target_params, end);
}
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;
void *end = (void *) param + param_size;
char *target_params;
if (!param->target_count) {
DMWARN("populate_table: no targets specified");
return -EINVAL;
}
for (i = 0; i < param->target_count; i++) {
r = next_target(spec, next, end, &spec, &target_params);
if (r) {
DMWARN("unable to find target");
return r;
}
r = dm_table_add_target(table, spec->target_type,
(sector_t) spec->sector_start,
(sector_t) spec->length,
target_params);
if (r) {
DMWARN("error adding target to table");
return r;
}
next = spec->next;
}
return dm_table_complete(table);
}
static int table_load(struct dm_ioctl *param, size_t param_size)
{
int r;
struct hash_cell *hc;
struct dm_table *t;
struct mapped_device *md;
md = find_device(param);
if (!md)
return -ENXIO;
r = dm_table_create(&t, get_mode(param), param->target_count, md);
if (r)
goto out;
r = populate_table(t, param, param_size);
if (r) {
dm_table_destroy(t);
goto out;
}
dm: do not initialise full request queue when bio based Change bio-based mapped devices no longer to have a fully initialized request_queue (request_fn, elevator, etc). This means bio-based DM devices no longer register elevator sysfs attributes ('iosched/' tree or 'scheduler' other than "none"). In contrast, a request-based DM device will continue to have a full request_queue and will register elevator sysfs attributes. Therefore a user can determine a DM device's type by checking if elevator sysfs attributes exist. First allocate a minimalist request_queue structure for a DM device (needed for both bio and request-based DM). Initialization of a full request_queue is deferred until it is known that the DM device is request-based, at the end of the table load sequence. Factor DM device's request_queue initialization: - common to both request-based and bio-based into dm_init_md_queue(). - specific to request-based into dm_init_request_based_queue(). The md->type_lock mutex is used to protect md->queue, in addition to md->type, during table_load(). A DM device's first table_load will establish the immutable md->type. But md->queue initialization, based on md->type, may fail at that time (because blk_init_allocated_queue cannot allocate memory). Therefore any subsequent table_load must (re)try dm_setup_md_queue independently of establishing md->type. Signed-off-by: Mike Snitzer <snitzer@redhat.com> Acked-by: Kiyoshi Ueda <k-ueda@ct.jp.nec.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2010-08-12 11:14:02 +08:00
/* Protect md->type and md->queue against concurrent table loads. */
dm_lock_md_type(md);
if (dm_get_md_type(md) == DM_TYPE_NONE)
/* Initial table load: acquire type of table. */
dm_set_md_type(md, dm_table_get_type(t));
else if (dm_get_md_type(md) != dm_table_get_type(t)) {
DMWARN("can't change device type after initial table load.");
dm_table_destroy(t);
dm_unlock_md_type(md);
r = -EINVAL;
goto out;
}
dm: do not initialise full request queue when bio based Change bio-based mapped devices no longer to have a fully initialized request_queue (request_fn, elevator, etc). This means bio-based DM devices no longer register elevator sysfs attributes ('iosched/' tree or 'scheduler' other than "none"). In contrast, a request-based DM device will continue to have a full request_queue and will register elevator sysfs attributes. Therefore a user can determine a DM device's type by checking if elevator sysfs attributes exist. First allocate a minimalist request_queue structure for a DM device (needed for both bio and request-based DM). Initialization of a full request_queue is deferred until it is known that the DM device is request-based, at the end of the table load sequence. Factor DM device's request_queue initialization: - common to both request-based and bio-based into dm_init_md_queue(). - specific to request-based into dm_init_request_based_queue(). The md->type_lock mutex is used to protect md->queue, in addition to md->type, during table_load(). A DM device's first table_load will establish the immutable md->type. But md->queue initialization, based on md->type, may fail at that time (because blk_init_allocated_queue cannot allocate memory). Therefore any subsequent table_load must (re)try dm_setup_md_queue independently of establishing md->type. Signed-off-by: Mike Snitzer <snitzer@redhat.com> Acked-by: Kiyoshi Ueda <k-ueda@ct.jp.nec.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2010-08-12 11:14:02 +08:00
/* setup md->queue to reflect md's type (may block) */
r = dm_setup_md_queue(md);
if (r) {
DMWARN("unable to set up device queue for new table.");
dm_table_destroy(t);
dm_unlock_md_type(md);
goto out;
}
dm_unlock_md_type(md);
/* stage inactive table */
down_write(&_hash_lock);
hc = dm_get_mdptr(md);
if (!hc || hc->md != md) {
DMWARN("device has been removed from the dev hash table.");
dm_table_destroy(t);
up_write(&_hash_lock);
r = -ENXIO;
goto out;
}
if (hc->new_map)
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 11:05:10 +08:00
dm_table_destroy(hc->new_map);
hc->new_map = t;
up_write(&_hash_lock);
param->flags |= DM_INACTIVE_PRESENT_FLAG;
__dev_status(md, param);
out:
dm_put(md);
return r;
}
static int table_clear(struct dm_ioctl *param, size_t param_size)
{
struct hash_cell *hc;
struct mapped_device *md;
down_write(&_hash_lock);
hc = __find_device_hash_cell(param);
if (!hc) {
DMWARN("device doesn't appear to be in the dev hash table.");
up_write(&_hash_lock);
return -ENXIO;
}
if (hc->new_map) {
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 11:05:10 +08:00
dm_table_destroy(hc->new_map);
hc->new_map = NULL;
}
param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
__dev_status(hc->md, param);
md = hc->md;
up_write(&_hash_lock);
dm_put(md);
return 0;
}
/*
* 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;
struct dm_dev_internal *dd;
struct dm_target_deps *deps;
deps = get_result_buffer(param, param_size, &len);
/*
* Count the devices.
*/
list_for_each (tmp, dm_table_get_devices(table))
count++;
/*
* Check we have enough space.
*/
needed = sizeof(*deps) + (sizeof(*deps->dev) * count);
if (len < needed) {
param->flags |= DM_BUFFER_FULL_FLAG;
return;
}
/*
* Fill in the devices.
*/
deps->count = count;
count = 0;
list_for_each_entry (dd, dm_table_get_devices(table), list)
deps->dev[count++] = huge_encode_dev(dd->dm_dev.bdev->bd_dev);
param->data_size = param->data_start + needed;
}
static int table_deps(struct dm_ioctl *param, size_t param_size)
{
struct mapped_device *md;
struct dm_table *table;
md = find_device(param);
if (!md)
return -ENXIO;
__dev_status(md, param);
table = dm_get_live_or_inactive_table(md, param);
if (table) {
retrieve_deps(table, param, param_size);
dm_table_put(table);
}
dm_put(md);
return 0;
}
/*
* Return the status of a device as a text string for each
* target.
*/
static int table_status(struct dm_ioctl *param, size_t param_size)
{
struct mapped_device *md;
struct dm_table *table;
md = find_device(param);
if (!md)
return -ENXIO;
__dev_status(md, param);
table = dm_get_live_or_inactive_table(md, param);
if (table) {
retrieve_status(table, param, param_size);
dm_table_put(table);
}
dm_put(md);
return 0;
}
/*
* Pass a message to the target that's at the supplied device offset.
*/
static int target_message(struct dm_ioctl *param, size_t param_size)
{
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;
md = find_device(param);
if (!md)
return -ENXIO;
if (tmsg < (struct dm_target_msg *) param->data ||
invalid_str(tmsg->message, (void *) param + param_size)) {
DMWARN("Invalid target message parameters.");
r = -EINVAL;
goto out;
}
r = dm_split_args(&argc, &argv, tmsg->message);
if (r) {
DMWARN("Failed to split target message parameters");
goto out;
}
table = dm_get_live_table(md);
if (!table)
goto out_argv;
if (dm_deleting_md(md)) {
r = -ENXIO;
goto out_table;
}
ti = dm_table_find_target(table, tmsg->sector);
if (!dm_target_is_valid(ti)) {
DMWARN("Target message sector outside device.");
r = -EINVAL;
} else if (ti->type->message)
r = ti->type->message(ti, argc, argv);
else {
DMWARN("Target type does not support messages");
r = -EINVAL;
}
out_table:
dm_table_put(table);
out_argv:
kfree(argv);
out:
param->data_size = 0;
dm_put(md);
return r;
}
/*-----------------------------------------------------------------
* Implementation of open/close/ioctl on the special char
* device.
*---------------------------------------------------------------*/
static ioctl_fn lookup_ioctl(unsigned int cmd)
{
static struct {
int cmd;
ioctl_fn fn;
} _ioctls[] = {
{DM_VERSION_CMD, NULL}, /* version is dealt with elsewhere */
{DM_REMOVE_ALL_CMD, remove_all},
{DM_LIST_DEVICES_CMD, list_devices},
{DM_DEV_CREATE_CMD, dev_create},
{DM_DEV_REMOVE_CMD, dev_remove},
{DM_DEV_RENAME_CMD, dev_rename},
{DM_DEV_SUSPEND_CMD, dev_suspend},
{DM_DEV_STATUS_CMD, dev_status},
{DM_DEV_WAIT_CMD, dev_wait},
{DM_TABLE_LOAD_CMD, table_load},
{DM_TABLE_CLEAR_CMD, table_clear},
{DM_TABLE_DEPS_CMD, table_deps},
{DM_TABLE_STATUS_CMD, table_status},
{DM_LIST_VERSIONS_CMD, list_versions},
{DM_TARGET_MSG_CMD, target_message},
{DM_DEV_SET_GEOMETRY_CMD, dev_set_geometry}
};
return (cmd >= ARRAY_SIZE(_ioctls)) ? NULL : _ioctls[cmd].fn;
}
/*
* As well as checking the version compatibility this always
* copies the kernel interface version out.
*/
static int check_version(unsigned int cmd, struct dm_ioctl __user *user)
{
uint32_t version[3];
int r = 0;
if (copy_from_user(version, user->version, sizeof(version)))
return -EFAULT;
if ((DM_VERSION_MAJOR != version[0]) ||
(DM_VERSION_MINOR < version[1])) {
DMWARN("ioctl interface mismatch: "
"kernel(%u.%u.%u), user(%u.%u.%u), cmd(%d)",
DM_VERSION_MAJOR, DM_VERSION_MINOR,
DM_VERSION_PATCHLEVEL,
version[0], version[1], version[2], cmd);
r = -EINVAL;
}
/*
* Fill in the kernel version.
*/
version[0] = DM_VERSION_MAJOR;
version[1] = DM_VERSION_MINOR;
version[2] = DM_VERSION_PATCHLEVEL;
if (copy_to_user(user->version, version, sizeof(version)))
return -EFAULT;
return r;
}
static void free_params(struct dm_ioctl *param)
{
vfree(param);
}
static int copy_params(struct dm_ioctl __user *user, struct dm_ioctl **param)
{
struct dm_ioctl tmp, *dmi;
if (copy_from_user(&tmp, user, sizeof(tmp) - sizeof(tmp.data)))
return -EFAULT;
if (tmp.data_size < (sizeof(tmp) - sizeof(tmp.data)))
return -EINVAL;
dmi = vmalloc(tmp.data_size);
if (!dmi)
return -ENOMEM;
if (copy_from_user(dmi, user, tmp.data_size)) {
vfree(dmi);
return -EFAULT;
}
*param = dmi;
return 0;
}
static int validate_params(uint cmd, struct dm_ioctl *param)
{
/* Always clear this flag */
param->flags &= ~DM_BUFFER_FULL_FLAG;
param->flags &= ~DM_UEVENT_GENERATED_FLAG;
/* Ignores parameters */
if (cmd == DM_REMOVE_ALL_CMD ||
cmd == DM_LIST_DEVICES_CMD ||
cmd == DM_LIST_VERSIONS_CMD)
return 0;
if ((cmd == DM_DEV_CREATE_CMD)) {
if (!*param->name) {
DMWARN("name not supplied when creating device");
return -EINVAL;
}
} else if ((*param->uuid && *param->name)) {
DMWARN("only supply one of name or uuid, cmd(%u)", cmd);
return -EINVAL;
}
/* Ensure strings are terminated */
param->name[DM_NAME_LEN - 1] = '\0';
param->uuid[DM_UUID_LEN - 1] = '\0';
return 0;
}
static int ctl_ioctl(uint command, struct dm_ioctl __user *user)
{
int r = 0;
unsigned int cmd;
struct dm_ioctl *uninitialized_var(param);
ioctl_fn fn = NULL;
size_t param_size;
/* 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.
*/
r = check_version(cmd, user);
if (r)
return r;
/*
* Nothing more to do for the version command.
*/
if (cmd == DM_VERSION_CMD)
return 0;
fn = lookup_ioctl(cmd);
if (!fn) {
DMWARN("dm_ctl_ioctl: unknown command 0x%x", command);
return -ENOTTY;
}
/*
* Trying to avoid low memory issues when a device is
* suspended.
*/
current->flags |= PF_MEMALLOC;
/*
* Copy the parameters into kernel space.
*/
r = copy_params(user, &param);
current->flags &= ~PF_MEMALLOC;
if (r)
return r;
r = validate_params(cmd, param);
if (r)
goto out;
param_size = param->data_size;
param->data_size = sizeof(*param);
r = fn(param, param_size);
/*
* Copy the results back to userland.
*/
if (!r && copy_to_user(user, param, param->data_size))
r = -EFAULT;
out:
free_params(param);
return r;
}
static long dm_ctl_ioctl(struct file *file, uint command, ulong u)
{
return (long)ctl_ioctl(command, (struct dm_ioctl __user *)u);
}
#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
static const struct file_operations _ctl_fops = {
.open = nonseekable_open,
.unlocked_ioctl = dm_ctl_ioctl,
.compat_ioctl = dm_compat_ctl_ioctl,
.owner = THIS_MODULE,
};
static struct miscdevice _dm_misc = {
.minor = MAPPER_CTRL_MINOR,
.name = DM_NAME,
.nodename = DM_DIR "/" DM_CONTROL_NODE,
.fops = &_ctl_fops
};
MODULE_ALIAS_MISCDEV(MAPPER_CTRL_MINOR);
MODULE_ALIAS("devname:" DM_DIR "/" DM_CONTROL_NODE);
/*
* Create misc character device and link to DM_DIR/control.
*/
int __init dm_interface_init(void)
{
int r;
r = dm_hash_init();
if (r)
return r;
r = misc_register(&_dm_misc);
if (r) {
DMERR("misc_register failed for control device");
dm_hash_exit();
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)
{
if (misc_deregister(&_dm_misc) < 0)
DMERR("misc_deregister failed for control device");
dm_hash_exit();
}
/**
* 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;
dm: avoid _hash_lock deadlock Fix a reported deadlock if there are still unprocessed multipath events on a device that is being removed. _hash_lock is held during dev_remove while trying to send the outstanding events. Sending the events requests the _hash_lock again in dm_copy_name_and_uuid. This patch introduces a separate lock around regions that modify the link to the hash table (dm_set_mdptr) or the name or uuid so that dm_copy_name_and_uuid no longer needs _hash_lock. Additionally, dm_copy_name_and_uuid can only be called if md exists so we can drop the dm_get() and dm_put() which can lead to a BUG() while md is being freed. The deadlock: #0 [ffff8106298dfb48] schedule at ffffffff80063035 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad #7 [ffff8106298dfd30] dev_remove at ffffffff88250865 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4 #10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9 #11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf #12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call) Cc: stable@kernel.org Reported-by: guy keren <choo@actcom.co.il> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-12-11 07:51:52 +08:00
mutex_lock(&dm_hash_cells_mutex);
hc = dm_get_mdptr(md);
if (!hc || hc->md != md) {
r = -ENXIO;
goto out;
}
if (name)
strcpy(name, hc->name);
if (uuid)
strcpy(uuid, hc->uuid ? : "");
out:
dm: avoid _hash_lock deadlock Fix a reported deadlock if there are still unprocessed multipath events on a device that is being removed. _hash_lock is held during dev_remove while trying to send the outstanding events. Sending the events requests the _hash_lock again in dm_copy_name_and_uuid. This patch introduces a separate lock around regions that modify the link to the hash table (dm_set_mdptr) or the name or uuid so that dm_copy_name_and_uuid no longer needs _hash_lock. Additionally, dm_copy_name_and_uuid can only be called if md exists so we can drop the dm_get() and dm_put() which can lead to a BUG() while md is being freed. The deadlock: #0 [ffff8106298dfb48] schedule at ffffffff80063035 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad #7 [ffff8106298dfd30] dev_remove at ffffffff88250865 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4 #10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9 #11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf #12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call) Cc: stable@kernel.org Reported-by: guy keren <choo@actcom.co.il> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
2009-12-11 07:51:52 +08:00
mutex_unlock(&dm_hash_cells_mutex);
return r;
}