linux/fs/btrfs/dev-replace.c

953 lines
29 KiB
C
Raw Normal View History

/*
* Copyright (C) STRATO AG 2012. All rights reserved.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public
* License v2 as published by the Free Software Foundation.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* You should have received a copy of the GNU General Public
* License along with this program; if not, write to the
* Free Software Foundation, Inc., 59 Temple Place - Suite 330,
* Boston, MA 021110-1307, USA.
*/
#include <linux/sched.h>
#include <linux/bio.h>
#include <linux/slab.h>
#include <linux/buffer_head.h>
#include <linux/blkdev.h>
#include <linux/random.h>
#include <linux/iocontext.h>
#include <linux/capability.h>
#include <linux/kthread.h>
#include <linux/math64.h>
#include <asm/div64.h>
#include "ctree.h"
#include "extent_map.h"
#include "disk-io.h"
#include "transaction.h"
#include "print-tree.h"
#include "volumes.h"
#include "async-thread.h"
#include "check-integrity.h"
#include "rcu-string.h"
#include "dev-replace.h"
#include "sysfs.h"
static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
int scrub_ret);
static void btrfs_dev_replace_update_device_in_mapping_tree(
struct btrfs_fs_info *fs_info,
struct btrfs_device *srcdev,
struct btrfs_device *tgtdev);
static u64 __btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info);
static int btrfs_dev_replace_kthread(void *data);
static int btrfs_dev_replace_continue_on_mount(struct btrfs_fs_info *fs_info);
int btrfs_init_dev_replace(struct btrfs_fs_info *fs_info)
{
struct btrfs_key key;
struct btrfs_root *dev_root = fs_info->dev_root;
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
struct extent_buffer *eb;
int slot;
int ret = 0;
struct btrfs_path *path = NULL;
int item_size;
struct btrfs_dev_replace_item *ptr;
u64 src_devid;
path = btrfs_alloc_path();
if (!path) {
ret = -ENOMEM;
goto out;
}
key.objectid = 0;
key.type = BTRFS_DEV_REPLACE_KEY;
key.offset = 0;
ret = btrfs_search_slot(NULL, dev_root, &key, path, 0, 0);
if (ret) {
no_valid_dev_replace_entry_found:
ret = 0;
dev_replace->replace_state =
BTRFS_DEV_REPLACE_ITEM_STATE_NEVER_STARTED;
dev_replace->cont_reading_from_srcdev_mode =
BTRFS_DEV_REPLACE_ITEM_CONT_READING_FROM_SRCDEV_MODE_ALWAYS;
dev_replace->replace_state = 0;
dev_replace->time_started = 0;
dev_replace->time_stopped = 0;
atomic64_set(&dev_replace->num_write_errors, 0);
atomic64_set(&dev_replace->num_uncorrectable_read_errors, 0);
dev_replace->cursor_left = 0;
dev_replace->committed_cursor_left = 0;
dev_replace->cursor_left_last_write_of_item = 0;
dev_replace->cursor_right = 0;
dev_replace->srcdev = NULL;
dev_replace->tgtdev = NULL;
dev_replace->is_valid = 0;
dev_replace->item_needs_writeback = 0;
goto out;
}
slot = path->slots[0];
eb = path->nodes[0];
item_size = btrfs_item_size_nr(eb, slot);
ptr = btrfs_item_ptr(eb, slot, struct btrfs_dev_replace_item);
if (item_size != sizeof(struct btrfs_dev_replace_item)) {
btrfs_warn(fs_info,
"dev_replace entry found has unexpected size, ignore entry");
goto no_valid_dev_replace_entry_found;
}
src_devid = btrfs_dev_replace_src_devid(eb, ptr);
dev_replace->cont_reading_from_srcdev_mode =
btrfs_dev_replace_cont_reading_from_srcdev_mode(eb, ptr);
dev_replace->replace_state = btrfs_dev_replace_replace_state(eb, ptr);
dev_replace->time_started = btrfs_dev_replace_time_started(eb, ptr);
dev_replace->time_stopped =
btrfs_dev_replace_time_stopped(eb, ptr);
atomic64_set(&dev_replace->num_write_errors,
btrfs_dev_replace_num_write_errors(eb, ptr));
atomic64_set(&dev_replace->num_uncorrectable_read_errors,
btrfs_dev_replace_num_uncorrectable_read_errors(eb, ptr));
dev_replace->cursor_left = btrfs_dev_replace_cursor_left(eb, ptr);
dev_replace->committed_cursor_left = dev_replace->cursor_left;
dev_replace->cursor_left_last_write_of_item = dev_replace->cursor_left;
dev_replace->cursor_right = btrfs_dev_replace_cursor_right(eb, ptr);
dev_replace->is_valid = 1;
dev_replace->item_needs_writeback = 0;
switch (dev_replace->replace_state) {
case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
dev_replace->srcdev = NULL;
dev_replace->tgtdev = NULL;
break;
case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
dev_replace->srcdev = btrfs_find_device(fs_info, src_devid,
NULL, NULL);
dev_replace->tgtdev = btrfs_find_device(fs_info,
BTRFS_DEV_REPLACE_DEVID,
NULL, NULL);
/*
* allow 'btrfs dev replace_cancel' if src/tgt device is
* missing
*/
if (!dev_replace->srcdev &&
!btrfs_test_opt(fs_info, DEGRADED)) {
ret = -EIO;
btrfs_warn(fs_info,
"cannot mount because device replace operation is ongoing and");
btrfs_warn(fs_info,
"srcdev (devid %llu) is missing, need to run 'btrfs dev scan'?",
src_devid);
}
if (!dev_replace->tgtdev &&
!btrfs_test_opt(fs_info, DEGRADED)) {
ret = -EIO;
btrfs_warn(fs_info,
"cannot mount because device replace operation is ongoing and");
btrfs_warn(fs_info,
"tgtdev (devid %llu) is missing, need to run 'btrfs dev scan'?",
BTRFS_DEV_REPLACE_DEVID);
}
if (dev_replace->tgtdev) {
if (dev_replace->srcdev) {
dev_replace->tgtdev->total_bytes =
dev_replace->srcdev->total_bytes;
dev_replace->tgtdev->disk_total_bytes =
dev_replace->srcdev->disk_total_bytes;
dev_replace->tgtdev->commit_total_bytes =
dev_replace->srcdev->commit_total_bytes;
dev_replace->tgtdev->bytes_used =
dev_replace->srcdev->bytes_used;
dev_replace->tgtdev->commit_bytes_used =
dev_replace->srcdev->commit_bytes_used;
}
dev_replace->tgtdev->is_tgtdev_for_dev_replace = 1;
btrfs_init_dev_replace_tgtdev_for_resume(fs_info,
dev_replace->tgtdev);
}
break;
}
out:
btrfs_free_path(path);
return ret;
}
/*
* called from commit_transaction. Writes changed device replace state to
* disk.
*/
int btrfs_run_dev_replace(struct btrfs_trans_handle *trans,
struct btrfs_fs_info *fs_info)
{
int ret;
struct btrfs_root *dev_root = fs_info->dev_root;
struct btrfs_path *path;
struct btrfs_key key;
struct extent_buffer *eb;
struct btrfs_dev_replace_item *ptr;
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_lock(dev_replace, 0);
if (!dev_replace->is_valid ||
!dev_replace->item_needs_writeback) {
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_unlock(dev_replace, 0);
return 0;
}
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_unlock(dev_replace, 0);
key.objectid = 0;
key.type = BTRFS_DEV_REPLACE_KEY;
key.offset = 0;
path = btrfs_alloc_path();
if (!path) {
ret = -ENOMEM;
goto out;
}
ret = btrfs_search_slot(trans, dev_root, &key, path, -1, 1);
if (ret < 0) {
btrfs_warn(fs_info,
"error %d while searching for dev_replace item!",
ret);
goto out;
}
if (ret == 0 &&
btrfs_item_size_nr(path->nodes[0], path->slots[0]) < sizeof(*ptr)) {
/*
* need to delete old one and insert a new one.
* Since no attempt is made to recover any old state, if the
* dev_replace state is 'running', the data on the target
* drive is lost.
* It would be possible to recover the state: just make sure
* that the beginning of the item is never changed and always
* contains all the essential information. Then read this
* minimal set of information and use it as a base for the
* new state.
*/
ret = btrfs_del_item(trans, dev_root, path);
if (ret != 0) {
btrfs_warn(fs_info,
"delete too small dev_replace item failed %d!",
ret);
goto out;
}
ret = 1;
}
if (ret == 1) {
/* need to insert a new item */
btrfs_release_path(path);
ret = btrfs_insert_empty_item(trans, dev_root, path,
&key, sizeof(*ptr));
if (ret < 0) {
btrfs_warn(fs_info,
"insert dev_replace item failed %d!", ret);
goto out;
}
}
eb = path->nodes[0];
ptr = btrfs_item_ptr(eb, path->slots[0],
struct btrfs_dev_replace_item);
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_lock(dev_replace, 1);
if (dev_replace->srcdev)
btrfs_set_dev_replace_src_devid(eb, ptr,
dev_replace->srcdev->devid);
else
btrfs_set_dev_replace_src_devid(eb, ptr, (u64)-1);
btrfs_set_dev_replace_cont_reading_from_srcdev_mode(eb, ptr,
dev_replace->cont_reading_from_srcdev_mode);
btrfs_set_dev_replace_replace_state(eb, ptr,
dev_replace->replace_state);
btrfs_set_dev_replace_time_started(eb, ptr, dev_replace->time_started);
btrfs_set_dev_replace_time_stopped(eb, ptr, dev_replace->time_stopped);
btrfs_set_dev_replace_num_write_errors(eb, ptr,
atomic64_read(&dev_replace->num_write_errors));
btrfs_set_dev_replace_num_uncorrectable_read_errors(eb, ptr,
atomic64_read(&dev_replace->num_uncorrectable_read_errors));
dev_replace->cursor_left_last_write_of_item =
dev_replace->cursor_left;
btrfs_set_dev_replace_cursor_left(eb, ptr,
dev_replace->cursor_left_last_write_of_item);
btrfs_set_dev_replace_cursor_right(eb, ptr,
dev_replace->cursor_right);
dev_replace->item_needs_writeback = 0;
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_unlock(dev_replace, 1);
btrfs_mark_buffer_dirty(eb);
out:
btrfs_free_path(path);
return ret;
}
void btrfs_after_dev_replace_commit(struct btrfs_fs_info *fs_info)
{
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
dev_replace->committed_cursor_left =
dev_replace->cursor_left_last_write_of_item;
}
int btrfs_dev_replace_start(struct btrfs_fs_info *fs_info,
const char *tgtdev_name, u64 srcdevid, const char *srcdev_name,
int read_src)
{
struct btrfs_root *root = fs_info->dev_root;
struct btrfs_trans_handle *trans;
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
int ret;
struct btrfs_device *tgt_device = NULL;
struct btrfs_device *src_device = NULL;
/* the disk copy procedure reuses the scrub code */
mutex_lock(&fs_info->volume_mutex);
ret = btrfs_find_device_by_devspec(fs_info, srcdevid,
srcdev_name, &src_device);
if (ret) {
mutex_unlock(&fs_info->volume_mutex);
return ret;
}
ret = btrfs_init_dev_replace_tgtdev(fs_info, tgtdev_name,
src_device, &tgt_device);
mutex_unlock(&fs_info->volume_mutex);
if (ret)
return ret;
/*
* Here we commit the transaction to make sure commit_total_bytes
* of all the devices are updated.
*/
trans = btrfs_attach_transaction(root);
if (!IS_ERR(trans)) {
ret = btrfs_commit_transaction(trans);
if (ret)
return ret;
} else if (PTR_ERR(trans) != -ENOENT) {
return PTR_ERR(trans);
}
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_lock(dev_replace, 1);
switch (dev_replace->replace_state) {
case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
break;
case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
ret = BTRFS_IOCTL_DEV_REPLACE_RESULT_ALREADY_STARTED;
goto leave;
}
dev_replace->cont_reading_from_srcdev_mode = read_src;
WARN_ON(!src_device);
dev_replace->srcdev = src_device;
WARN_ON(!tgt_device);
dev_replace->tgtdev = tgt_device;
btrfs_info_in_rcu(fs_info,
"dev_replace from %s (devid %llu) to %s started",
src_device->missing ? "<missing disk>" :
rcu_str_deref(src_device->name),
src_device->devid,
rcu_str_deref(tgt_device->name));
/*
* from now on, the writes to the srcdev are all duplicated to
* go to the tgtdev as well (refer to btrfs_map_block()).
*/
dev_replace->replace_state = BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED;
dev_replace->time_started = get_seconds();
dev_replace->cursor_left = 0;
dev_replace->committed_cursor_left = 0;
dev_replace->cursor_left_last_write_of_item = 0;
dev_replace->cursor_right = 0;
dev_replace->is_valid = 1;
dev_replace->item_needs_writeback = 1;
atomic64_set(&dev_replace->num_write_errors, 0);
atomic64_set(&dev_replace->num_uncorrectable_read_errors, 0);
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_unlock(dev_replace, 1);
ret = btrfs_sysfs_add_device_link(tgt_device->fs_devices, tgt_device);
if (ret)
btrfs_err(fs_info, "kobj add dev failed %d", ret);
btrfs_wait_ordered_roots(fs_info, U64_MAX, 0, (u64)-1);
/* force writing the updated state information to disk */
trans = btrfs_start_transaction(root, 0);
if (IS_ERR(trans)) {
ret = PTR_ERR(trans);
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_lock(dev_replace, 1);
goto leave;
}
ret = btrfs_commit_transaction(trans);
WARN_ON(ret);
/* the disk copy procedure reuses the scrub code */
ret = btrfs_scrub_dev(fs_info, src_device->devid, 0,
btrfs_device_get_total_bytes(src_device),
&dev_replace->scrub_progress, 0, 1);
ret = btrfs_dev_replace_finishing(fs_info, ret);
if (ret == -EINPROGRESS) {
ret = BTRFS_IOCTL_DEV_REPLACE_RESULT_SCRUB_INPROGRESS;
} else {
WARN_ON(ret);
}
return ret;
leave:
dev_replace->srcdev = NULL;
dev_replace->tgtdev = NULL;
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_unlock(dev_replace, 1);
btrfs_destroy_dev_replace_tgtdev(fs_info, tgt_device);
return ret;
}
int btrfs_dev_replace_by_ioctl(struct btrfs_fs_info *fs_info,
struct btrfs_ioctl_dev_replace_args *args)
{
int ret;
switch (args->start.cont_reading_from_srcdev_mode) {
case BTRFS_IOCTL_DEV_REPLACE_CONT_READING_FROM_SRCDEV_MODE_ALWAYS:
case BTRFS_IOCTL_DEV_REPLACE_CONT_READING_FROM_SRCDEV_MODE_AVOID:
break;
default:
return -EINVAL;
}
if ((args->start.srcdevid == 0 && args->start.srcdev_name[0] == '\0') ||
args->start.tgtdev_name[0] == '\0')
return -EINVAL;
ret = btrfs_dev_replace_start(fs_info, args->start.tgtdev_name,
args->start.srcdevid,
args->start.srcdev_name,
args->start.cont_reading_from_srcdev_mode);
args->result = ret;
/* don't warn if EINPROGRESS, someone else might be running scrub */
if (ret == BTRFS_IOCTL_DEV_REPLACE_RESULT_SCRUB_INPROGRESS)
ret = 0;
return ret;
}
Btrfs: fix use-after-free in the finishing procedure of the device replace During device replace test, we hit a null pointer deference (It was very easy to reproduce it by running xfstests' btrfs/011 on the devices with the virtio scsi driver). There were two bugs that caused this problem: - We might allocate new chunks on the replaced device after we updated the mapping tree. And we forgot to replace the source device in those mapping of the new chunks. - We might get the mapping information which including the source device before the mapping information update. And then submit the bio which was based on that mapping information after we freed the source device. For the first bug, we can fix it by doing mapping tree update and source device remove in the same context of the chunk mutex. The chunk mutex is used to protect the allocable device list, the above method can avoid the new chunk allocation, and after we remove the source device, all the new chunks will be allocated on the new device. So it can fix the first bug. For the second bug, we need make sure all flighting bios are finished and no new bios are produced during we are removing the source device. To fix this problem, we introduced a global @bio_counter, we not only inc/dec @bio_counter outsize of map_blocks, but also inc it before submitting bio and dec @bio_counter when ending bios. Since Raid56 is a little different and device replace dosen't support raid56 yet, it is not addressed in the patch and I add comments to make sure we will fix it in the future. Reported-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> Signed-off-by: Miao Xie <miaox@cn.fujitsu.com> Signed-off-by: Josef Bacik <jbacik@fb.com>
2014-01-30 08:46:55 +00:00
/*
* blocked until all in-flight bios operations are finished.
Btrfs: fix use-after-free in the finishing procedure of the device replace During device replace test, we hit a null pointer deference (It was very easy to reproduce it by running xfstests' btrfs/011 on the devices with the virtio scsi driver). There were two bugs that caused this problem: - We might allocate new chunks on the replaced device after we updated the mapping tree. And we forgot to replace the source device in those mapping of the new chunks. - We might get the mapping information which including the source device before the mapping information update. And then submit the bio which was based on that mapping information after we freed the source device. For the first bug, we can fix it by doing mapping tree update and source device remove in the same context of the chunk mutex. The chunk mutex is used to protect the allocable device list, the above method can avoid the new chunk allocation, and after we remove the source device, all the new chunks will be allocated on the new device. So it can fix the first bug. For the second bug, we need make sure all flighting bios are finished and no new bios are produced during we are removing the source device. To fix this problem, we introduced a global @bio_counter, we not only inc/dec @bio_counter outsize of map_blocks, but also inc it before submitting bio and dec @bio_counter when ending bios. Since Raid56 is a little different and device replace dosen't support raid56 yet, it is not addressed in the patch and I add comments to make sure we will fix it in the future. Reported-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> Signed-off-by: Miao Xie <miaox@cn.fujitsu.com> Signed-off-by: Josef Bacik <jbacik@fb.com>
2014-01-30 08:46:55 +00:00
*/
static void btrfs_rm_dev_replace_blocked(struct btrfs_fs_info *fs_info)
{
set_bit(BTRFS_FS_STATE_DEV_REPLACING, &fs_info->fs_state);
wait_event(fs_info->replace_wait, !percpu_counter_sum(
&fs_info->bio_counter));
Btrfs: fix use-after-free in the finishing procedure of the device replace During device replace test, we hit a null pointer deference (It was very easy to reproduce it by running xfstests' btrfs/011 on the devices with the virtio scsi driver). There were two bugs that caused this problem: - We might allocate new chunks on the replaced device after we updated the mapping tree. And we forgot to replace the source device in those mapping of the new chunks. - We might get the mapping information which including the source device before the mapping information update. And then submit the bio which was based on that mapping information after we freed the source device. For the first bug, we can fix it by doing mapping tree update and source device remove in the same context of the chunk mutex. The chunk mutex is used to protect the allocable device list, the above method can avoid the new chunk allocation, and after we remove the source device, all the new chunks will be allocated on the new device. So it can fix the first bug. For the second bug, we need make sure all flighting bios are finished and no new bios are produced during we are removing the source device. To fix this problem, we introduced a global @bio_counter, we not only inc/dec @bio_counter outsize of map_blocks, but also inc it before submitting bio and dec @bio_counter when ending bios. Since Raid56 is a little different and device replace dosen't support raid56 yet, it is not addressed in the patch and I add comments to make sure we will fix it in the future. Reported-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> Signed-off-by: Miao Xie <miaox@cn.fujitsu.com> Signed-off-by: Josef Bacik <jbacik@fb.com>
2014-01-30 08:46:55 +00:00
}
/*
* we have removed target device, it is safe to allow new bios request.
*/
static void btrfs_rm_dev_replace_unblocked(struct btrfs_fs_info *fs_info)
{
clear_bit(BTRFS_FS_STATE_DEV_REPLACING, &fs_info->fs_state);
wake_up(&fs_info->replace_wait);
Btrfs: fix use-after-free in the finishing procedure of the device replace During device replace test, we hit a null pointer deference (It was very easy to reproduce it by running xfstests' btrfs/011 on the devices with the virtio scsi driver). There were two bugs that caused this problem: - We might allocate new chunks on the replaced device after we updated the mapping tree. And we forgot to replace the source device in those mapping of the new chunks. - We might get the mapping information which including the source device before the mapping information update. And then submit the bio which was based on that mapping information after we freed the source device. For the first bug, we can fix it by doing mapping tree update and source device remove in the same context of the chunk mutex. The chunk mutex is used to protect the allocable device list, the above method can avoid the new chunk allocation, and after we remove the source device, all the new chunks will be allocated on the new device. So it can fix the first bug. For the second bug, we need make sure all flighting bios are finished and no new bios are produced during we are removing the source device. To fix this problem, we introduced a global @bio_counter, we not only inc/dec @bio_counter outsize of map_blocks, but also inc it before submitting bio and dec @bio_counter when ending bios. Since Raid56 is a little different and device replace dosen't support raid56 yet, it is not addressed in the patch and I add comments to make sure we will fix it in the future. Reported-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> Signed-off-by: Miao Xie <miaox@cn.fujitsu.com> Signed-off-by: Josef Bacik <jbacik@fb.com>
2014-01-30 08:46:55 +00:00
}
static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
int scrub_ret)
{
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
struct btrfs_device *tgt_device;
struct btrfs_device *src_device;
struct btrfs_root *root = fs_info->tree_root;
u8 uuid_tmp[BTRFS_UUID_SIZE];
struct btrfs_trans_handle *trans;
int ret = 0;
/* don't allow cancel or unmount to disturb the finishing procedure */
mutex_lock(&dev_replace->lock_finishing_cancel_unmount);
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_lock(dev_replace, 0);
/* was the operation canceled, or is it finished? */
if (dev_replace->replace_state !=
BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED) {
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_unlock(dev_replace, 0);
mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
return 0;
}
tgt_device = dev_replace->tgtdev;
src_device = dev_replace->srcdev;
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_unlock(dev_replace, 0);
/*
* flush all outstanding I/O and inode extent mappings before the
* copy operation is declared as being finished
*/
ret = btrfs_start_delalloc_roots(fs_info, 0, -1);
if (ret) {
mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
return ret;
}
btrfs_wait_ordered_roots(fs_info, U64_MAX, 0, (u64)-1);
trans = btrfs_start_transaction(root, 0);
if (IS_ERR(trans)) {
mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
return PTR_ERR(trans);
}
ret = btrfs_commit_transaction(trans);
WARN_ON(ret);
mutex_lock(&uuid_mutex);
/* keep away write_all_supers() during the finishing procedure */
mutex_lock(&fs_info->fs_devices->device_list_mutex);
mutex_lock(&fs_info->chunk_mutex);
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_lock(dev_replace, 1);
dev_replace->replace_state =
scrub_ret ? BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED
: BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED;
dev_replace->tgtdev = NULL;
dev_replace->srcdev = NULL;
dev_replace->time_stopped = get_seconds();
dev_replace->item_needs_writeback = 1;
Btrfs: fix use-after-free in the finishing procedure of the device replace During device replace test, we hit a null pointer deference (It was very easy to reproduce it by running xfstests' btrfs/011 on the devices with the virtio scsi driver). There were two bugs that caused this problem: - We might allocate new chunks on the replaced device after we updated the mapping tree. And we forgot to replace the source device in those mapping of the new chunks. - We might get the mapping information which including the source device before the mapping information update. And then submit the bio which was based on that mapping information after we freed the source device. For the first bug, we can fix it by doing mapping tree update and source device remove in the same context of the chunk mutex. The chunk mutex is used to protect the allocable device list, the above method can avoid the new chunk allocation, and after we remove the source device, all the new chunks will be allocated on the new device. So it can fix the first bug. For the second bug, we need make sure all flighting bios are finished and no new bios are produced during we are removing the source device. To fix this problem, we introduced a global @bio_counter, we not only inc/dec @bio_counter outsize of map_blocks, but also inc it before submitting bio and dec @bio_counter when ending bios. Since Raid56 is a little different and device replace dosen't support raid56 yet, it is not addressed in the patch and I add comments to make sure we will fix it in the future. Reported-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> Signed-off-by: Miao Xie <miaox@cn.fujitsu.com> Signed-off-by: Josef Bacik <jbacik@fb.com>
2014-01-30 08:46:55 +00:00
/* replace old device with new one in mapping tree */
if (!scrub_ret) {
btrfs_dev_replace_update_device_in_mapping_tree(fs_info,
src_device,
tgt_device);
} else {
btrfs_err_in_rcu(fs_info,
"btrfs_scrub_dev(%s, %llu, %s) failed %d",
src_device->missing ? "<missing disk>" :
rcu_str_deref(src_device->name),
src_device->devid,
rcu_str_deref(tgt_device->name), scrub_ret);
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_unlock(dev_replace, 1);
mutex_unlock(&fs_info->chunk_mutex);
mutex_unlock(&fs_info->fs_devices->device_list_mutex);
mutex_unlock(&uuid_mutex);
btrfs: Wait for in-flight bios before freeing target device for raid56 When raid56 dev-replace is cancelled by running scrub, we will free target device without waiting for in-flight bios, causing the following NULL pointer deference or general protection failure. BUG: unable to handle kernel NULL pointer dereference at 00000000000005e0 IP: generic_make_request_checks+0x4d/0x610 CPU: 1 PID: 11676 Comm: kworker/u4:14 Tainted: G O 4.11.0-rc2 #72 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.10.2-20170228_101828-anatol 04/01/2014 Workqueue: btrfs-endio-raid56 btrfs_endio_raid56_helper [btrfs] task: ffff88002875b4c0 task.stack: ffffc90001334000 RIP: 0010:generic_make_request_checks+0x4d/0x610 Call Trace: ? generic_make_request+0xc7/0x360 generic_make_request+0x24/0x360 ? generic_make_request+0xc7/0x360 submit_bio+0x64/0x120 ? page_in_rbio+0x4d/0x80 [btrfs] ? rbio_orig_end_io+0x80/0x80 [btrfs] finish_rmw+0x3f4/0x540 [btrfs] validate_rbio_for_rmw+0x36/0x40 [btrfs] raid_rmw_end_io+0x7a/0x90 [btrfs] bio_endio+0x56/0x60 end_workqueue_fn+0x3c/0x40 [btrfs] btrfs_scrubparity_helper+0xef/0x620 [btrfs] btrfs_endio_raid56_helper+0xe/0x10 [btrfs] process_one_work+0x2af/0x720 ? process_one_work+0x22b/0x720 worker_thread+0x4b/0x4f0 kthread+0x10f/0x150 ? process_one_work+0x720/0x720 ? kthread_create_on_node+0x40/0x40 ret_from_fork+0x2e/0x40 RIP: generic_make_request_checks+0x4d/0x610 RSP: ffffc90001337bb8 In btrfs_dev_replace_finishing(), we will call btrfs_rm_dev_replace_blocked() to wait bios before destroying the target device when scrub is finished normally. However when dev-replace is aborted, either due to error or cancelled by scrub, we didn't wait for bios, this can lead to use-after-free if there are bios holding the target device. Furthermore, for raid56 scrub, at least 2 places are calling btrfs_map_sblock() without protection of bio_counter, leading to the problem. This patch fixes the problem: 1) Wait for bio_counter before freeing target device when canceling replace 2) When calling btrfs_map_sblock() for raid56, use bio_counter to protect the call. Cc: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Reviewed-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2017-03-29 01:33:21 +00:00
btrfs_rm_dev_replace_blocked(fs_info);
if (tgt_device)
btrfs_destroy_dev_replace_tgtdev(fs_info, tgt_device);
btrfs: Wait for in-flight bios before freeing target device for raid56 When raid56 dev-replace is cancelled by running scrub, we will free target device without waiting for in-flight bios, causing the following NULL pointer deference or general protection failure. BUG: unable to handle kernel NULL pointer dereference at 00000000000005e0 IP: generic_make_request_checks+0x4d/0x610 CPU: 1 PID: 11676 Comm: kworker/u4:14 Tainted: G O 4.11.0-rc2 #72 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.10.2-20170228_101828-anatol 04/01/2014 Workqueue: btrfs-endio-raid56 btrfs_endio_raid56_helper [btrfs] task: ffff88002875b4c0 task.stack: ffffc90001334000 RIP: 0010:generic_make_request_checks+0x4d/0x610 Call Trace: ? generic_make_request+0xc7/0x360 generic_make_request+0x24/0x360 ? generic_make_request+0xc7/0x360 submit_bio+0x64/0x120 ? page_in_rbio+0x4d/0x80 [btrfs] ? rbio_orig_end_io+0x80/0x80 [btrfs] finish_rmw+0x3f4/0x540 [btrfs] validate_rbio_for_rmw+0x36/0x40 [btrfs] raid_rmw_end_io+0x7a/0x90 [btrfs] bio_endio+0x56/0x60 end_workqueue_fn+0x3c/0x40 [btrfs] btrfs_scrubparity_helper+0xef/0x620 [btrfs] btrfs_endio_raid56_helper+0xe/0x10 [btrfs] process_one_work+0x2af/0x720 ? process_one_work+0x22b/0x720 worker_thread+0x4b/0x4f0 kthread+0x10f/0x150 ? process_one_work+0x720/0x720 ? kthread_create_on_node+0x40/0x40 ret_from_fork+0x2e/0x40 RIP: generic_make_request_checks+0x4d/0x610 RSP: ffffc90001337bb8 In btrfs_dev_replace_finishing(), we will call btrfs_rm_dev_replace_blocked() to wait bios before destroying the target device when scrub is finished normally. However when dev-replace is aborted, either due to error or cancelled by scrub, we didn't wait for bios, this can lead to use-after-free if there are bios holding the target device. Furthermore, for raid56 scrub, at least 2 places are calling btrfs_map_sblock() without protection of bio_counter, leading to the problem. This patch fixes the problem: 1) Wait for bio_counter before freeing target device when canceling replace 2) When calling btrfs_map_sblock() for raid56, use bio_counter to protect the call. Cc: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Reviewed-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2017-03-29 01:33:21 +00:00
btrfs_rm_dev_replace_unblocked(fs_info);
mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
return scrub_ret;
}
btrfs_info_in_rcu(fs_info,
"dev_replace from %s (devid %llu) to %s finished",
src_device->missing ? "<missing disk>" :
rcu_str_deref(src_device->name),
src_device->devid,
rcu_str_deref(tgt_device->name));
tgt_device->is_tgtdev_for_dev_replace = 0;
tgt_device->devid = src_device->devid;
src_device->devid = BTRFS_DEV_REPLACE_DEVID;
memcpy(uuid_tmp, tgt_device->uuid, sizeof(uuid_tmp));
memcpy(tgt_device->uuid, src_device->uuid, sizeof(tgt_device->uuid));
memcpy(src_device->uuid, uuid_tmp, sizeof(src_device->uuid));
btrfs_device_set_total_bytes(tgt_device, src_device->total_bytes);
btrfs_device_set_disk_total_bytes(tgt_device,
src_device->disk_total_bytes);
btrfs_device_set_bytes_used(tgt_device, src_device->bytes_used);
ASSERT(list_empty(&src_device->resized_list));
tgt_device->commit_total_bytes = src_device->commit_total_bytes;
tgt_device->commit_bytes_used = src_device->bytes_used;
btrfs_assign_next_active_device(fs_info, src_device, tgt_device);
list_add(&tgt_device->dev_alloc_list, &fs_info->fs_devices->alloc_list);
fs_info->fs_devices->rw_devices++;
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_unlock(dev_replace, 1);
Btrfs: fix use-after-free in the finishing procedure of the device replace During device replace test, we hit a null pointer deference (It was very easy to reproduce it by running xfstests' btrfs/011 on the devices with the virtio scsi driver). There were two bugs that caused this problem: - We might allocate new chunks on the replaced device after we updated the mapping tree. And we forgot to replace the source device in those mapping of the new chunks. - We might get the mapping information which including the source device before the mapping information update. And then submit the bio which was based on that mapping information after we freed the source device. For the first bug, we can fix it by doing mapping tree update and source device remove in the same context of the chunk mutex. The chunk mutex is used to protect the allocable device list, the above method can avoid the new chunk allocation, and after we remove the source device, all the new chunks will be allocated on the new device. So it can fix the first bug. For the second bug, we need make sure all flighting bios are finished and no new bios are produced during we are removing the source device. To fix this problem, we introduced a global @bio_counter, we not only inc/dec @bio_counter outsize of map_blocks, but also inc it before submitting bio and dec @bio_counter when ending bios. Since Raid56 is a little different and device replace dosen't support raid56 yet, it is not addressed in the patch and I add comments to make sure we will fix it in the future. Reported-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> Signed-off-by: Miao Xie <miaox@cn.fujitsu.com> Signed-off-by: Josef Bacik <jbacik@fb.com>
2014-01-30 08:46:55 +00:00
btrfs_rm_dev_replace_blocked(fs_info);
btrfs: Fix a lockdep warning when running xfstest. The following lockdep warning is triggered during xfstests: [ 1702.980872] ========================================================= [ 1702.981181] [ INFO: possible irq lock inversion dependency detected ] [ 1702.981482] 3.18.0-rc1 #27 Not tainted [ 1702.981781] --------------------------------------------------------- [ 1702.982095] kswapd0/77 just changed the state of lock: [ 1702.982415] (&delayed_node->mutex){+.+.-.}, at: [<ffffffffa03b0b51>] __btrfs_release_delayed_node+0x41/0x1f0 [btrfs] [ 1702.982794] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1702.983160] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1702.984675] other info that might help us debug this: [ 1702.985524] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1702.986799] Possible interrupt unsafe locking scenario: [ 1702.987681] CPU0 CPU1 [ 1702.988137] ---- ---- [ 1702.988598] lock(&fs_info->dev_replace.lock); [ 1702.989069] local_irq_disable(); [ 1702.989534] lock(&delayed_node->mutex); [ 1702.990038] lock(&found->groups_sem); [ 1702.990494] <Interrupt> [ 1702.990938] lock(&delayed_node->mutex); [ 1702.991407] *** DEADLOCK *** It is because the btrfs_kobj_{add/rm}_device() will call memory allocation with GFP_KERNEL, which may flush fs page cache to free space, waiting for it self to do the commit, causing the deadlock. To solve the problem, move btrfs_kobj_{add/rm}_device() out of the dev_replace lock range, also involing split the btrfs_rm_dev_replace_srcdev() function into remove and free parts. Now only btrfs_rm_dev_replace_remove_srcdev() is called in dev_replace lock range, and kobj_{add/rm} and btrfs_rm_dev_replace_free_srcdev() are called out of the lock range. Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
2014-10-30 08:52:31 +00:00
btrfs_rm_dev_replace_remove_srcdev(fs_info, src_device);
Btrfs: fix use-after-free in the finishing procedure of the device replace During device replace test, we hit a null pointer deference (It was very easy to reproduce it by running xfstests' btrfs/011 on the devices with the virtio scsi driver). There were two bugs that caused this problem: - We might allocate new chunks on the replaced device after we updated the mapping tree. And we forgot to replace the source device in those mapping of the new chunks. - We might get the mapping information which including the source device before the mapping information update. And then submit the bio which was based on that mapping information after we freed the source device. For the first bug, we can fix it by doing mapping tree update and source device remove in the same context of the chunk mutex. The chunk mutex is used to protect the allocable device list, the above method can avoid the new chunk allocation, and after we remove the source device, all the new chunks will be allocated on the new device. So it can fix the first bug. For the second bug, we need make sure all flighting bios are finished and no new bios are produced during we are removing the source device. To fix this problem, we introduced a global @bio_counter, we not only inc/dec @bio_counter outsize of map_blocks, but also inc it before submitting bio and dec @bio_counter when ending bios. Since Raid56 is a little different and device replace dosen't support raid56 yet, it is not addressed in the patch and I add comments to make sure we will fix it in the future. Reported-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> Signed-off-by: Miao Xie <miaox@cn.fujitsu.com> Signed-off-by: Josef Bacik <jbacik@fb.com>
2014-01-30 08:46:55 +00:00
btrfs_rm_dev_replace_unblocked(fs_info);
/*
* this is again a consistent state where no dev_replace procedure
* is running, the target device is part of the filesystem, the
* source device is not part of the filesystem anymore and its 1st
* superblock is scratched out so that it is no longer marked to
* belong to this filesystem.
*/
mutex_unlock(&fs_info->chunk_mutex);
mutex_unlock(&fs_info->fs_devices->device_list_mutex);
mutex_unlock(&uuid_mutex);
btrfs: Fix a lockdep warning when running xfstest. The following lockdep warning is triggered during xfstests: [ 1702.980872] ========================================================= [ 1702.981181] [ INFO: possible irq lock inversion dependency detected ] [ 1702.981482] 3.18.0-rc1 #27 Not tainted [ 1702.981781] --------------------------------------------------------- [ 1702.982095] kswapd0/77 just changed the state of lock: [ 1702.982415] (&delayed_node->mutex){+.+.-.}, at: [<ffffffffa03b0b51>] __btrfs_release_delayed_node+0x41/0x1f0 [btrfs] [ 1702.982794] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1702.983160] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1702.984675] other info that might help us debug this: [ 1702.985524] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1702.986799] Possible interrupt unsafe locking scenario: [ 1702.987681] CPU0 CPU1 [ 1702.988137] ---- ---- [ 1702.988598] lock(&fs_info->dev_replace.lock); [ 1702.989069] local_irq_disable(); [ 1702.989534] lock(&delayed_node->mutex); [ 1702.990038] lock(&found->groups_sem); [ 1702.990494] <Interrupt> [ 1702.990938] lock(&delayed_node->mutex); [ 1702.991407] *** DEADLOCK *** It is because the btrfs_kobj_{add/rm}_device() will call memory allocation with GFP_KERNEL, which may flush fs page cache to free space, waiting for it self to do the commit, causing the deadlock. To solve the problem, move btrfs_kobj_{add/rm}_device() out of the dev_replace lock range, also involing split the btrfs_rm_dev_replace_srcdev() function into remove and free parts. Now only btrfs_rm_dev_replace_remove_srcdev() is called in dev_replace lock range, and kobj_{add/rm} and btrfs_rm_dev_replace_free_srcdev() are called out of the lock range. Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
2014-10-30 08:52:31 +00:00
/* replace the sysfs entry */
btrfs_sysfs_rm_device_link(fs_info->fs_devices, src_device);
btrfs: Fix a lockdep warning when running xfstest. The following lockdep warning is triggered during xfstests: [ 1702.980872] ========================================================= [ 1702.981181] [ INFO: possible irq lock inversion dependency detected ] [ 1702.981482] 3.18.0-rc1 #27 Not tainted [ 1702.981781] --------------------------------------------------------- [ 1702.982095] kswapd0/77 just changed the state of lock: [ 1702.982415] (&delayed_node->mutex){+.+.-.}, at: [<ffffffffa03b0b51>] __btrfs_release_delayed_node+0x41/0x1f0 [btrfs] [ 1702.982794] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1702.983160] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1702.984675] other info that might help us debug this: [ 1702.985524] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1702.986799] Possible interrupt unsafe locking scenario: [ 1702.987681] CPU0 CPU1 [ 1702.988137] ---- ---- [ 1702.988598] lock(&fs_info->dev_replace.lock); [ 1702.989069] local_irq_disable(); [ 1702.989534] lock(&delayed_node->mutex); [ 1702.990038] lock(&found->groups_sem); [ 1702.990494] <Interrupt> [ 1702.990938] lock(&delayed_node->mutex); [ 1702.991407] *** DEADLOCK *** It is because the btrfs_kobj_{add/rm}_device() will call memory allocation with GFP_KERNEL, which may flush fs page cache to free space, waiting for it self to do the commit, causing the deadlock. To solve the problem, move btrfs_kobj_{add/rm}_device() out of the dev_replace lock range, also involing split the btrfs_rm_dev_replace_srcdev() function into remove and free parts. Now only btrfs_rm_dev_replace_remove_srcdev() is called in dev_replace lock range, and kobj_{add/rm} and btrfs_rm_dev_replace_free_srcdev() are called out of the lock range. Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
2014-10-30 08:52:31 +00:00
btrfs_rm_dev_replace_free_srcdev(fs_info, src_device);
/* write back the superblocks */
trans = btrfs_start_transaction(root, 0);
if (!IS_ERR(trans))
btrfs_commit_transaction(trans);
mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
return 0;
}
static void btrfs_dev_replace_update_device_in_mapping_tree(
struct btrfs_fs_info *fs_info,
struct btrfs_device *srcdev,
struct btrfs_device *tgtdev)
{
struct extent_map_tree *em_tree = &fs_info->mapping_tree.map_tree;
struct extent_map *em;
struct map_lookup *map;
u64 start = 0;
int i;
write_lock(&em_tree->lock);
do {
em = lookup_extent_mapping(em_tree, start, (u64)-1);
if (!em)
break;
map = em->map_lookup;
for (i = 0; i < map->num_stripes; i++)
if (srcdev == map->stripes[i].dev)
map->stripes[i].dev = tgtdev;
start = em->start + em->len;
free_extent_map(em);
} while (start);
write_unlock(&em_tree->lock);
}
/*
* Read progress of device replace status according to the state and last
* stored position. The value format is the same as for
* btrfs_dev_replace::progress_1000
*/
static u64 btrfs_dev_replace_progress(struct btrfs_fs_info *fs_info)
{
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
u64 ret = 0;
switch (dev_replace->replace_state) {
case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
ret = 0;
break;
case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
ret = 1000;
break;
case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
ret = div64_u64(dev_replace->cursor_left,
div_u64(btrfs_device_get_total_bytes(
dev_replace->srcdev), 1000));
break;
}
return ret;
}
void btrfs_dev_replace_status(struct btrfs_fs_info *fs_info,
struct btrfs_ioctl_dev_replace_args *args)
{
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_lock(dev_replace, 0);
/* even if !dev_replace_is_valid, the values are good enough for
* the replace_status ioctl */
args->result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR;
args->status.replace_state = dev_replace->replace_state;
args->status.time_started = dev_replace->time_started;
args->status.time_stopped = dev_replace->time_stopped;
args->status.num_write_errors =
atomic64_read(&dev_replace->num_write_errors);
args->status.num_uncorrectable_read_errors =
atomic64_read(&dev_replace->num_uncorrectable_read_errors);
args->status.progress_1000 = btrfs_dev_replace_progress(fs_info);
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_unlock(dev_replace, 0);
}
int btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info,
struct btrfs_ioctl_dev_replace_args *args)
{
args->result = __btrfs_dev_replace_cancel(fs_info);
return 0;
}
static u64 __btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info)
{
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
struct btrfs_device *tgt_device = NULL;
struct btrfs_trans_handle *trans;
struct btrfs_root *root = fs_info->tree_root;
u64 result;
int ret;
if (sb_rdonly(fs_info->sb))
return -EROFS;
mutex_lock(&dev_replace->lock_finishing_cancel_unmount);
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_lock(dev_replace, 1);
switch (dev_replace->replace_state) {
case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NOT_STARTED;
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_unlock(dev_replace, 1);
goto leave;
case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR;
tgt_device = dev_replace->tgtdev;
dev_replace->tgtdev = NULL;
dev_replace->srcdev = NULL;
break;
}
dev_replace->replace_state = BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED;
dev_replace->time_stopped = get_seconds();
dev_replace->item_needs_writeback = 1;
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_unlock(dev_replace, 1);
btrfs_scrub_cancel(fs_info);
trans = btrfs_start_transaction(root, 0);
if (IS_ERR(trans)) {
mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
return PTR_ERR(trans);
}
ret = btrfs_commit_transaction(trans);
WARN_ON(ret);
if (tgt_device)
btrfs_destroy_dev_replace_tgtdev(fs_info, tgt_device);
leave:
mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
return result;
}
void btrfs_dev_replace_suspend_for_unmount(struct btrfs_fs_info *fs_info)
{
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
mutex_lock(&dev_replace->lock_finishing_cancel_unmount);
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_lock(dev_replace, 1);
switch (dev_replace->replace_state) {
case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
break;
case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
dev_replace->replace_state =
BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED;
dev_replace->time_stopped = get_seconds();
dev_replace->item_needs_writeback = 1;
btrfs_info(fs_info, "suspending dev_replace for unmount");
break;
}
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_unlock(dev_replace, 1);
mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
}
/* resume dev_replace procedure that was interrupted by unmount */
int btrfs_resume_dev_replace_async(struct btrfs_fs_info *fs_info)
{
struct task_struct *task;
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_lock(dev_replace, 1);
switch (dev_replace->replace_state) {
case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_unlock(dev_replace, 1);
return 0;
case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
break;
case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
dev_replace->replace_state =
BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED;
break;
}
if (!dev_replace->tgtdev || !dev_replace->tgtdev->bdev) {
btrfs_info(fs_info,
"cannot continue dev_replace, tgtdev is missing");
btrfs_info(fs_info,
"you may cancel the operation after 'mount -o degraded'");
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_unlock(dev_replace, 1);
return 0;
}
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
btrfs_dev_replace_unlock(dev_replace, 1);
WARN_ON(test_and_set_bit(BTRFS_FS_EXCL_OP, &fs_info->flags));
task = kthread_run(btrfs_dev_replace_kthread, fs_info, "btrfs-devrepl");
return PTR_ERR_OR_ZERO(task);
}
static int btrfs_dev_replace_kthread(void *data)
{
struct btrfs_fs_info *fs_info = data;
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
u64 progress;
progress = btrfs_dev_replace_progress(fs_info);
progress = div_u64(progress, 10);
btrfs_info_in_rcu(fs_info,
"continuing dev_replace from %s (devid %llu) to %s @%u%%",
dev_replace->srcdev->missing ? "<missing disk>"
: rcu_str_deref(dev_replace->srcdev->name),
dev_replace->srcdev->devid,
dev_replace->tgtdev ? rcu_str_deref(dev_replace->tgtdev->name)
: "<missing target disk>",
(unsigned int)progress);
btrfs_dev_replace_continue_on_mount(fs_info);
clear_bit(BTRFS_FS_EXCL_OP, &fs_info->flags);
return 0;
}
static int btrfs_dev_replace_continue_on_mount(struct btrfs_fs_info *fs_info)
{
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
int ret;
ret = btrfs_scrub_dev(fs_info, dev_replace->srcdev->devid,
dev_replace->committed_cursor_left,
btrfs_device_get_total_bytes(dev_replace->srcdev),
&dev_replace->scrub_progress, 0, 1);
ret = btrfs_dev_replace_finishing(fs_info, ret);
WARN_ON(ret);
return 0;
}
int btrfs_dev_replace_is_ongoing(struct btrfs_dev_replace *dev_replace)
{
if (!dev_replace->is_valid)
return 0;
switch (dev_replace->replace_state) {
case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
return 0;
case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
/*
* return true even if tgtdev is missing (this is
* something that can happen if the dev_replace
* procedure is suspended by an umount and then
* the tgtdev is missing (or "btrfs dev scan") was
* not called and the the filesystem is remounted
* in degraded state. This does not stop the
* dev_replace procedure. It needs to be canceled
* manually if the cancellation is wanted.
*/
break;
}
return 1;
}
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
void btrfs_dev_replace_lock(struct btrfs_dev_replace *dev_replace, int rw)
{
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
if (rw == 1) {
/* write */
again:
wait_event(dev_replace->read_lock_wq,
atomic_read(&dev_replace->blocking_readers) == 0);
write_lock(&dev_replace->lock);
if (atomic_read(&dev_replace->blocking_readers)) {
write_unlock(&dev_replace->lock);
goto again;
}
} else {
read_lock(&dev_replace->lock);
atomic_inc(&dev_replace->read_locks);
}
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
}
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
void btrfs_dev_replace_unlock(struct btrfs_dev_replace *dev_replace, int rw)
{
if (rw == 1) {
/* write */
ASSERT(atomic_read(&dev_replace->blocking_readers) == 0);
write_unlock(&dev_replace->lock);
} else {
ASSERT(atomic_read(&dev_replace->read_locks) > 0);
atomic_dec(&dev_replace->read_locks);
read_unlock(&dev_replace->lock);
}
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
}
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
/* inc blocking cnt and release read lock */
void btrfs_dev_replace_set_lock_blocking(
struct btrfs_dev_replace *dev_replace)
{
/* only set blocking for read lock */
ASSERT(atomic_read(&dev_replace->read_locks) > 0);
atomic_inc(&dev_replace->blocking_readers);
read_unlock(&dev_replace->lock);
}
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
/* acquire read lock and dec blocking cnt */
void btrfs_dev_replace_clear_lock_blocking(
struct btrfs_dev_replace *dev_replace)
{
Btrfs: fix lockdep deadlock warning due to dev_replace Xfstests btrfs/011 complains about a deadlock warning, [ 1226.649039] ========================================================= [ 1226.649039] [ INFO: possible irq lock inversion dependency detected ] [ 1226.649039] 4.1.0+ #270 Not tainted [ 1226.649039] --------------------------------------------------------- [ 1226.652955] kswapd0/46 just changed the state of lock: [ 1226.652955] (&delayed_node->mutex){+.+.-.}, at: [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] but this lock took another, RECLAIM_FS-unsafe lock in the past: [ 1226.652955] (&fs_info->dev_replace.lock){+.+.+.} and interrupts could create inverse lock ordering between them. [ 1226.652955] other info that might help us debug this: [ 1226.652955] Chain exists of: &delayed_node->mutex --> &found->groups_sem --> &fs_info->dev_replace.lock [ 1226.652955] Possible interrupt unsafe locking scenario: [ 1226.652955] CPU0 CPU1 [ 1226.652955] ---- ---- [ 1226.652955] lock(&fs_info->dev_replace.lock); [ 1226.652955] local_irq_disable(); [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] lock(&found->groups_sem); [ 1226.652955] <Interrupt> [ 1226.652955] lock(&delayed_node->mutex); [ 1226.652955] *** DEADLOCK *** Commit 084b6e7c7607 ("btrfs: Fix a lockdep warning when running xfstest.") tried to fix a similar one that has the exactly same warning, but with that, we still run to this. The above lock chain comes from btrfs_commit_transaction ->btrfs_run_delayed_items ... ->__btrfs_update_delayed_inode ... ->__btrfs_cow_block ... ->find_free_extent ->cache_block_group ->load_free_space_cache ->btrfs_readpages ->submit_one_bio ... ->__btrfs_map_block ->btrfs_dev_replace_lock However, with high memory pressure, tasks which hold dev_replace.lock can be interrupted by kswapd and then kswapd is intended to release memory occupied by superblock, inodes and dentries, where we may call evict_inode, and it comes to [ 1226.652955] [<ffffffff81458735>] __btrfs_release_delayed_node+0x45/0x1d0 [ 1226.652955] [<ffffffff81459e74>] btrfs_remove_delayed_node+0x24/0x30 [ 1226.652955] [<ffffffff8140c5fe>] btrfs_evict_inode+0x34e/0x700 delayed_node->mutex may be acquired in __btrfs_release_delayed_node(), and it leads to a ABBA deadlock. To fix this, we can use "blocking rwlock" used in the case of extent_buffer, but things are simpler here since we only needs read's spinlock to blocking lock. With this, btrfs/011 no more produces warnings in dmesg. Signed-off-by: Liu Bo <bo.li.liu@oracle.com> Signed-off-by: David Sterba <dsterba@suse.com>
2015-07-17 08:49:19 +00:00
/* only set blocking for read lock */
ASSERT(atomic_read(&dev_replace->read_locks) > 0);
ASSERT(atomic_read(&dev_replace->blocking_readers) > 0);
read_lock(&dev_replace->lock);
if (atomic_dec_and_test(&dev_replace->blocking_readers) &&
waitqueue_active(&dev_replace->read_lock_wq))
wake_up(&dev_replace->read_lock_wq);
}
Btrfs: fix use-after-free in the finishing procedure of the device replace During device replace test, we hit a null pointer deference (It was very easy to reproduce it by running xfstests' btrfs/011 on the devices with the virtio scsi driver). There were two bugs that caused this problem: - We might allocate new chunks on the replaced device after we updated the mapping tree. And we forgot to replace the source device in those mapping of the new chunks. - We might get the mapping information which including the source device before the mapping information update. And then submit the bio which was based on that mapping information after we freed the source device. For the first bug, we can fix it by doing mapping tree update and source device remove in the same context of the chunk mutex. The chunk mutex is used to protect the allocable device list, the above method can avoid the new chunk allocation, and after we remove the source device, all the new chunks will be allocated on the new device. So it can fix the first bug. For the second bug, we need make sure all flighting bios are finished and no new bios are produced during we are removing the source device. To fix this problem, we introduced a global @bio_counter, we not only inc/dec @bio_counter outsize of map_blocks, but also inc it before submitting bio and dec @bio_counter when ending bios. Since Raid56 is a little different and device replace dosen't support raid56 yet, it is not addressed in the patch and I add comments to make sure we will fix it in the future. Reported-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> Signed-off-by: Miao Xie <miaox@cn.fujitsu.com> Signed-off-by: Josef Bacik <jbacik@fb.com>
2014-01-30 08:46:55 +00:00
void btrfs_bio_counter_inc_noblocked(struct btrfs_fs_info *fs_info)
{
percpu_counter_inc(&fs_info->bio_counter);
}
void btrfs_bio_counter_sub(struct btrfs_fs_info *fs_info, s64 amount)
Btrfs: fix use-after-free in the finishing procedure of the device replace During device replace test, we hit a null pointer deference (It was very easy to reproduce it by running xfstests' btrfs/011 on the devices with the virtio scsi driver). There were two bugs that caused this problem: - We might allocate new chunks on the replaced device after we updated the mapping tree. And we forgot to replace the source device in those mapping of the new chunks. - We might get the mapping information which including the source device before the mapping information update. And then submit the bio which was based on that mapping information after we freed the source device. For the first bug, we can fix it by doing mapping tree update and source device remove in the same context of the chunk mutex. The chunk mutex is used to protect the allocable device list, the above method can avoid the new chunk allocation, and after we remove the source device, all the new chunks will be allocated on the new device. So it can fix the first bug. For the second bug, we need make sure all flighting bios are finished and no new bios are produced during we are removing the source device. To fix this problem, we introduced a global @bio_counter, we not only inc/dec @bio_counter outsize of map_blocks, but also inc it before submitting bio and dec @bio_counter when ending bios. Since Raid56 is a little different and device replace dosen't support raid56 yet, it is not addressed in the patch and I add comments to make sure we will fix it in the future. Reported-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> Signed-off-by: Miao Xie <miaox@cn.fujitsu.com> Signed-off-by: Josef Bacik <jbacik@fb.com>
2014-01-30 08:46:55 +00:00
{
percpu_counter_sub(&fs_info->bio_counter, amount);
Btrfs: fix use-after-free in the finishing procedure of the device replace During device replace test, we hit a null pointer deference (It was very easy to reproduce it by running xfstests' btrfs/011 on the devices with the virtio scsi driver). There were two bugs that caused this problem: - We might allocate new chunks on the replaced device after we updated the mapping tree. And we forgot to replace the source device in those mapping of the new chunks. - We might get the mapping information which including the source device before the mapping information update. And then submit the bio which was based on that mapping information after we freed the source device. For the first bug, we can fix it by doing mapping tree update and source device remove in the same context of the chunk mutex. The chunk mutex is used to protect the allocable device list, the above method can avoid the new chunk allocation, and after we remove the source device, all the new chunks will be allocated on the new device. So it can fix the first bug. For the second bug, we need make sure all flighting bios are finished and no new bios are produced during we are removing the source device. To fix this problem, we introduced a global @bio_counter, we not only inc/dec @bio_counter outsize of map_blocks, but also inc it before submitting bio and dec @bio_counter when ending bios. Since Raid56 is a little different and device replace dosen't support raid56 yet, it is not addressed in the patch and I add comments to make sure we will fix it in the future. Reported-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> Signed-off-by: Miao Xie <miaox@cn.fujitsu.com> Signed-off-by: Josef Bacik <jbacik@fb.com>
2014-01-30 08:46:55 +00:00
if (waitqueue_active(&fs_info->replace_wait))
wake_up(&fs_info->replace_wait);
}
void btrfs_bio_counter_inc_blocked(struct btrfs_fs_info *fs_info)
{
while (1) {
percpu_counter_inc(&fs_info->bio_counter);
if (likely(!test_bit(BTRFS_FS_STATE_DEV_REPLACING,
&fs_info->fs_state)))
break;
Btrfs: fix use-after-free in the finishing procedure of the device replace During device replace test, we hit a null pointer deference (It was very easy to reproduce it by running xfstests' btrfs/011 on the devices with the virtio scsi driver). There were two bugs that caused this problem: - We might allocate new chunks on the replaced device after we updated the mapping tree. And we forgot to replace the source device in those mapping of the new chunks. - We might get the mapping information which including the source device before the mapping information update. And then submit the bio which was based on that mapping information after we freed the source device. For the first bug, we can fix it by doing mapping tree update and source device remove in the same context of the chunk mutex. The chunk mutex is used to protect the allocable device list, the above method can avoid the new chunk allocation, and after we remove the source device, all the new chunks will be allocated on the new device. So it can fix the first bug. For the second bug, we need make sure all flighting bios are finished and no new bios are produced during we are removing the source device. To fix this problem, we introduced a global @bio_counter, we not only inc/dec @bio_counter outsize of map_blocks, but also inc it before submitting bio and dec @bio_counter when ending bios. Since Raid56 is a little different and device replace dosen't support raid56 yet, it is not addressed in the patch and I add comments to make sure we will fix it in the future. Reported-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> Signed-off-by: Miao Xie <miaox@cn.fujitsu.com> Signed-off-by: Josef Bacik <jbacik@fb.com>
2014-01-30 08:46:55 +00:00
btrfs_bio_counter_dec(fs_info);
wait_event(fs_info->replace_wait,
!test_bit(BTRFS_FS_STATE_DEV_REPLACING,
&fs_info->fs_state));
}
}