[PATCH] btrfs: Allow read-only mount with corrupted extent tree

[PATCH] btrfs: Allow read-only mount with corrupted extent tree

[Dāvis Mosāns]

Currently if there's any corruption at all in extent tree
(eg. even single bit) then mounting will fail with:
"failed to read block groups: -5" (-EIO)
It happens because we immediately abort on first error when
searching in extent tree for block groups.

Now with this patch if `ignorebadroots` option is specified
then we handle such case and continue by removing already
created block groups and creating dummy block groups.

Signed-off-by: Dāvis Mosāns <davispuh@gmail.com>
---
 fs/btrfs/block-group.c | 20 ++++++++++++++++++++
 fs/btrfs/disk-io.c     |  4 ++--
 fs/btrfs/disk-io.h     |  2 ++
 3 files changed, 24 insertions(+), 2 deletions(-)

diff --git a/fs/btrfs/block-group.c b/fs/btrfs/block-group.c
index 9e7d9d0c763d..80b9bb9afb8c 100644
--- a/fs/btrfs/block-group.c
+++ b/fs/btrfs/block-group.c
@@ -2212,6 +2212,26 @@ int btrfs_read_block_groups(struct btrfs_fs_info *info)
 	ret = check_chunk_block_group_mappings(info);
 error:
 	btrfs_free_path(path);
+
+	if (ret == -EIO && btrfs_test_opt(info, IGNOREBADROOTS)) {
+
+		if (btrfs_super_log_root(info->super_copy) != 0) {
+			btrfs_warn(info, "Ignoring tree-log replay due to extent tree corruption!");
+			btrfs_set_super_log_root(info->super_copy, 0);
+		}
+
+		btrfs_put_block_group_cache(info);
+		btrfs_stop_all_workers(info);
+		btrfs_free_block_groups(info);
+		ret = btrfs_init_workqueues(info, NULL);
+		if (ret)
+			return ret;
+		ret = btrfs_init_space_info(info);
+		if (ret)
+			return ret;
+		return fill_dummy_bgs(info);
+	}
+
 	return ret;
 }
 
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index a59ab7b9aea0..b1ad9c85d578 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -2115,7 +2115,7 @@ static int read_backup_root(struct btrfs_fs_info *fs_info, u8 priority)
 }
 
 /* helper to cleanup workers */
-static void btrfs_stop_all_workers(struct btrfs_fs_info *fs_info)
+void btrfs_stop_all_workers(struct btrfs_fs_info *fs_info)
 {
 	btrfs_destroy_workqueue(fs_info->fixup_workers);
 	btrfs_destroy_workqueue(fs_info->delalloc_workers);
@@ -2283,7 +2283,7 @@ static void btrfs_init_qgroup(struct btrfs_fs_info *fs_info)
 	mutex_init(&fs_info->qgroup_rescan_lock);
 }
 
-static int btrfs_init_workqueues(struct btrfs_fs_info *fs_info,
+int btrfs_init_workqueues(struct btrfs_fs_info *fs_info,
 		struct btrfs_fs_devices *fs_devices)
 {
 	u32 max_active = fs_info->thread_pool_size;
diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h
index 0e7e9526b6a8..41348c8d3f9a 100644
--- a/fs/btrfs/disk-io.h
+++ b/fs/btrfs/disk-io.h
@@ -139,6 +139,8 @@ int btrfs_get_free_objectid(struct btrfs_root *root, u64 *objectid);
 int btrfs_init_root_free_objectid(struct btrfs_root *root);
 int __init btrfs_end_io_wq_init(void);
 void __cold btrfs_end_io_wq_exit(void);
+void btrfs_stop_all_workers(struct btrfs_fs_info *fs_info);
+int btrfs_init_workqueues(struct btrfs_fs_info *fs_info, struct btrfs_fs_devices *fs_devices);
 
 #ifdef CONFIG_DEBUG_LOCK_ALLOC
 void btrfs_set_buffer_lockdep_class(u64 objectid,
-- 
2.32.0

Re: [PATCH] btrfs: Allow read-only mount with corrupted extent tree

[Josef Bacik]

On 8/11/21 4:07 PM, Dāvis Mosāns wrote:
> Currently if there's any corruption at all in extent tree
> (eg. even single bit) then mounting will fail with:
> "failed to read block groups: -5" (-EIO)
> It happens because we immediately abort on first error when
> searching in extent tree for block groups.
> 
> Now with this patch if `ignorebadroots` option is specified
> then we handle such case and continue by removing already
> created block groups and creating dummy block groups.
> 

Already done and queue'ed up for the next release

btrfs: rescue: allow ibadroots to skip bad extent tree when reading block group 
items

thanks,

Josef

Re: [PATCH] btrfs: Allow read-only mount with corrupted extent tree

[Dāvis Mosāns]

trešd., 2021. g. 11. aug., plkst. 23:41 — lietotājs Josef Bacik
(<josef@toxicpanda.com>) rakstīja:
>
> On 8/11/21 4:07 PM, Dāvis Mosāns wrote:
> > Currently if there's any corruption at all in extent tree
> > (eg. even single bit) then mounting will fail with:
> > "failed to read block groups: -5" (-EIO)
> > It happens because we immediately abort on first error when
> > searching in extent tree for block groups.
> >
> > Now with this patch if `ignorebadroots` option is specified
> > then we handle such case and continue by removing already
> > created block groups and creating dummy block groups.
> >
>
> Already done and queue'ed up for the next release
>
> btrfs: rescue: allow ibadroots to skip bad extent tree when reading block group
> items

Nice!
Originally I submitted this patch in March, but it went nowhere so resent.
Anyway I tested that other patch and it has limitation that it won't
handle log replay error so need to use both
rescue=ignorebadroots,nologreplay
But other than it seems to work fine.

Here's output when not using nologreplay

[ 1886.281376] BTRFS info (device sde): ignoring bad roots
[ 1886.281381] BTRFS info (device sde): disk space caching is enabled
[ 1886.281383] BTRFS info (device sde): has skinny extents
[ 1886.365608] BTRFS info (device sde): bdev /dev/sdq errs: wr 0, rd
0, flush 0, corrupt 473, gen 0
[ 1886.365618] BTRFS info (device sde): bdev /dev/sdi errs: wr 31626,
rd 18765, flush 178, corrupt 5841, gen 0
[ 1886.365623] BTRFS info (device sde): bdev /dev/sdo errs: wr 6867,
rd 2640, flush 178, corrupt 1066, gen 0
[ 1900.249267] BTRFS warning (device sde): checksum verify failed on
21057125941248 wanted 0x5a4526a7 found 0x25949991 level 0
[ 1902.323362] BTRFS error (device sde): parent transid verify failed
on 21057111523328 wanted 2262739 found 2262698
[ 1902.519338] BTRFS error (device sde): bad tree block start, want
21057108836352 have 524288
[ 1902.940022] BTRFS warning (device sde): checksum verify failed on
21057097302016 wanted 0x8b2501e9 found 0x5b8ab9a3 level 0
[ 1902.946715] BTRFS error (device sde): parent transid verify failed
on 21057097302016 wanted 2262739 found 2262696
[ 1902.950727] BTRFS info (device sde): start tree-log replay
[ 1911.289738] BTRFS warning (device sde): checksum verify failed on
21057127661568 wanted 0xd1506ed9 found 0x22ab750a level 0
[ 1911.293266] BTRFS warning (device sde): checksum verify failed on
21057107017728 wanted 0x9120eaee found 0x35b2df28 level 0
[ 1911.326093] BTRFS: error (device sde) in __btrfs_free_extent:3069:
errno=-5 IO failure
[ 1911.326101] BTRFS: error (device sde) in
btrfs_run_delayed_refs:2150: errno=-5 IO failure
[ 1911.326106] BTRFS warning (device sde): Skipping commit of aborted
transaction.
[ 1911.326108] BTRFS: error (device sde) in cleanup_transaction:1945:
errno=-5 IO failure
[ 1911.326190] BTRFS warning (device sde): checksum verify failed on
21057107017728 wanted 0x9120eaee found 0xd00e16cf level 0
[ 1911.326317] BTRFS warning (device sde): checksum verify failed on
21057107017728 wanted 0x9120eaee found 0xd00e16cf level 0
[ 1911.326922] BTRFS: error (device sde) in btrfs_replay_log:2417:
errno=-5 IO failure (Failed to recover log tree)
[ 1912.160913] BTRFS error (device sde): parent transid verify failed
on 21057111736320 wanted 2262739 found 2262698
[ 1914.666177] BTRFS error (device sde): open_ctree failed

Best regards,
Dāvis

Re: [PATCH] btrfs: Allow read-only mount with corrupted extent tree

[Dāvis Mosāns]

svētd., 2021. g. 21. marts, plkst. 23:46 — lietotājs Dāvis Mosāns
(<davispuh@gmail.com>) rakstīja:
>
> Currently if there's any corruption at all in extent tree
> (eg. even single bit) then mounting will fail with:
> "failed to read block groups: -5" (-EIO)
> It happens because we immediately abort on first error when
> searching in extent tree for block groups.
>
> Now with this patch if `ignorebadroots` option is specified
> then we handle such case and continue by removing already
> created block groups and creating dummy block groups.
>
> Signed-off-by: Dāvis Mosāns <davispuh@gmail.com>
> ---
>  fs/btrfs/block-group.c | 20 ++++++++++++++++++++
>  fs/btrfs/disk-io.c     |  4 ++--
>  fs/btrfs/disk-io.h     |  2 ++
>  3 files changed, 24 insertions(+), 2 deletions(-)
>
> diff --git a/fs/btrfs/block-group.c b/fs/btrfs/block-group.c
> index 48ebc106a606..f485cf14c2f8 100644
> --- a/fs/btrfs/block-group.c
> +++ b/fs/btrfs/block-group.c
> @@ -2048,6 +2048,26 @@ int btrfs_read_block_groups(struct btrfs_fs_info *info)
>         ret = check_chunk_block_group_mappings(info);
>  error:
>         btrfs_free_path(path);
> +
> +       if (ret == -EIO && btrfs_test_opt(info, IGNOREBADROOTS)) {
> +
> +               if (btrfs_super_log_root(info->super_copy) != 0) {
> +                       btrfs_warn(info, "Ignoring tree-log replay due to extent tree corruption!");
> +                       btrfs_set_super_log_root(info->super_copy, 0);
> +               }
> +
> +               btrfs_put_block_group_cache(info);
> +               btrfs_stop_all_workers(info);
> +               btrfs_free_block_groups(info);
> +               ret = btrfs_init_workqueues(info, NULL);
> +               if (ret)
> +                       return ret;
> +               ret = btrfs_init_space_info(info);
> +               if (ret)
> +                       return ret;
> +               return fill_dummy_bgs(info);
> +       }
> +
>         return ret;
>  }
>
> diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
> index 07a2b4f69b10..dc744f76d075 100644
> --- a/fs/btrfs/disk-io.c
> +++ b/fs/btrfs/disk-io.c
> @@ -1955,7 +1955,7 @@ static int read_backup_root(struct btrfs_fs_info *fs_info, u8 priority)
>  }
>
>  /* helper to cleanup workers */
> -static void btrfs_stop_all_workers(struct btrfs_fs_info *fs_info)
> +void btrfs_stop_all_workers(struct btrfs_fs_info *fs_info)
>  {
>         btrfs_destroy_workqueue(fs_info->fixup_workers);
>         btrfs_destroy_workqueue(fs_info->delalloc_workers);
> @@ -2122,7 +2122,7 @@ static void btrfs_init_qgroup(struct btrfs_fs_info *fs_info)
>         mutex_init(&fs_info->qgroup_rescan_lock);
>  }
>
> -static int btrfs_init_workqueues(struct btrfs_fs_info *fs_info,
> +int btrfs_init_workqueues(struct btrfs_fs_info *fs_info,
>                 struct btrfs_fs_devices *fs_devices)
>  {
>         u32 max_active = fs_info->thread_pool_size;
> diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h
> index e45057c0c016..f9bfcba86a04 100644
> --- a/fs/btrfs/disk-io.h
> +++ b/fs/btrfs/disk-io.h
> @@ -137,6 +137,8 @@ int btrfs_find_free_objectid(struct btrfs_root *root, u64 *objectid);
>  int btrfs_find_highest_objectid(struct btrfs_root *root, u64 *objectid);
>  int __init btrfs_end_io_wq_init(void);
>  void __cold btrfs_end_io_wq_exit(void);
> +void btrfs_stop_all_workers(struct btrfs_fs_info *fs_info);
> +int btrfs_init_workqueues(struct btrfs_fs_info *fs_info, struct btrfs_fs_devices *fs_devices);
>
>  #ifdef CONFIG_DEBUG_LOCK_ALLOC
>  void btrfs_set_buffer_lockdep_class(u64 objectid,
> --
> 2.30.2
>

Ping? Could anyone take a look?

[PATCH] btrfs: Allow read-only mount with corrupted extent tree

[Dāvis Mosāns]

Currently if there's any corruption at all in extent tree
(eg. even single bit) then mounting will fail with:
"failed to read block groups: -5" (-EIO)
It happens because we immediately abort on first error when
searching in extent tree for block groups.

Now with this patch if `ignorebadroots` option is specified
then we handle such case and continue by removing already
created block groups and creating dummy block groups.

Signed-off-by: Dāvis Mosāns <davispuh@gmail.com>
---
 fs/btrfs/block-group.c | 20 ++++++++++++++++++++
 fs/btrfs/disk-io.c     |  4 ++--
 fs/btrfs/disk-io.h     |  2 ++
 3 files changed, 24 insertions(+), 2 deletions(-)

diff --git a/fs/btrfs/block-group.c b/fs/btrfs/block-group.c
index 48ebc106a606..f485cf14c2f8 100644
--- a/fs/btrfs/block-group.c
+++ b/fs/btrfs/block-group.c
@@ -2048,6 +2048,26 @@ int btrfs_read_block_groups(struct btrfs_fs_info *info)
 	ret = check_chunk_block_group_mappings(info);
 error:
 	btrfs_free_path(path);
+
+	if (ret == -EIO && btrfs_test_opt(info, IGNOREBADROOTS)) {
+
+		if (btrfs_super_log_root(info->super_copy) != 0) {
+			btrfs_warn(info, "Ignoring tree-log replay due to extent tree corruption!");
+			btrfs_set_super_log_root(info->super_copy, 0);
+		}
+
+		btrfs_put_block_group_cache(info);
+		btrfs_stop_all_workers(info);
+		btrfs_free_block_groups(info);
+		ret = btrfs_init_workqueues(info, NULL);
+		if (ret)
+			return ret;
+		ret = btrfs_init_space_info(info);
+		if (ret)
+			return ret;
+		return fill_dummy_bgs(info);
+	}
+
 	return ret;
 }
 
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 07a2b4f69b10..dc744f76d075 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -1955,7 +1955,7 @@ static int read_backup_root(struct btrfs_fs_info *fs_info, u8 priority)
 }
 
 /* helper to cleanup workers */
-static void btrfs_stop_all_workers(struct btrfs_fs_info *fs_info)
+void btrfs_stop_all_workers(struct btrfs_fs_info *fs_info)
 {
 	btrfs_destroy_workqueue(fs_info->fixup_workers);
 	btrfs_destroy_workqueue(fs_info->delalloc_workers);
@@ -2122,7 +2122,7 @@ static void btrfs_init_qgroup(struct btrfs_fs_info *fs_info)
 	mutex_init(&fs_info->qgroup_rescan_lock);
 }
 
-static int btrfs_init_workqueues(struct btrfs_fs_info *fs_info,
+int btrfs_init_workqueues(struct btrfs_fs_info *fs_info,
 		struct btrfs_fs_devices *fs_devices)
 {
 	u32 max_active = fs_info->thread_pool_size;
diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h
index e45057c0c016..f9bfcba86a04 100644
--- a/fs/btrfs/disk-io.h
+++ b/fs/btrfs/disk-io.h
@@ -137,6 +137,8 @@ int btrfs_find_free_objectid(struct btrfs_root *root, u64 *objectid);
 int btrfs_find_highest_objectid(struct btrfs_root *root, u64 *objectid);
 int __init btrfs_end_io_wq_init(void);
 void __cold btrfs_end_io_wq_exit(void);
+void btrfs_stop_all_workers(struct btrfs_fs_info *fs_info);
+int btrfs_init_workqueues(struct btrfs_fs_info *fs_info, struct btrfs_fs_devices *fs_devices);
 
 #ifdef CONFIG_DEBUG_LOCK_ALLOC
 void btrfs_set_buffer_lockdep_class(u64 objectid,
-- 
2.30.2