btrfs-progs: enhance error handling of __free_extent() #1065
+120
−92
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
There is a known bug in older kernels that orphan items can be missing
for dropped subvolumes.
This makes those subvolumes unable to be removed on the next mount, and
recent kernel commit 4289b494ac55 ("btrfs: do not allow relocation of
partially dropped subvolumes") introduced one extra safe net to catch
such problem.
But unfortunately there is no way to repair it.
Add the repair ability to both the original and lowmem modes.
Reviewed-by: Boris Burkov <boris@bur.io>
Signed-off-by: Qu Wenruo <wqu@suse.com>
The test case fsck/066 is only to verify we can detect the missing root orphan item, no repair for it yet. Now the repair ability is added, change the test case to verify the repair is also properly done. Reviewed-by: Boris Burkov <boris@bur.io> Signed-off-by: Qu Wenruo <wqu@suse.com>
…xtent() [BUG] There is a bug report that btrfs-convert crashed during converting an ext4 image which is almost full. [CAUSE] Just before the crash, btrfs-convert is committing the current transaction but failed to locate the backref inside __free_extent(). Then it went through the error handling path, which prints an error message and try to dump the leaf. But in this particular case, the error code is not -ENOENT, thus the path is already released, resuling path->nodes[0] to be NULL, and btrfs_print_leaf() triggers a NULL pointer dereference. [FIX] The kernel version of btrfs_free_extent() is only dumping the tree for -ENOENT error code. And patch "btrfs: refactor the error handling of __btrfs_free_extent()" is submitted to make abort_and_dump() to only dump the leaf if the path is not released. So follow the same kernel patch, by only dumping the leaf if the path is not released. Issue: kdave#1064 Signed-off-by: Qu Wenruo <wqu@suse.com>
Just follow the kernel patch "btrfs: refactor the error handling of __btrfs_free_extent()", to handle the error first for lookup_extent_backref(), so we can reduce one indent level. Furthermore remove the unnessary forced type casting of the error message, and replace the old printk() with proper the error() helper. Signed-off-by: Qu Wenruo <wqu@suse.com>
kdave
pushed a commit
that referenced
this pull request
Nov 25, 2025
…xtent() [BUG] There is a bug report that btrfs-convert crashed during converting an ext4 image which is almost full. [CAUSE] Just before the crash, btrfs-convert is committing the current transaction but failed to locate the backref inside __free_extent(). Then it went through the error handling path, which printed an error message and tried to dump the leaf. But in this particular case, the error code is not -ENOENT, thus the path is already released, resulting path->nodes[0] to be NULL, and btrfs_print_leaf() triggers a NULL pointer dereference. [FIX] The kernel version of btrfs_free_extent() is only dumping the tree for -ENOENT error code. And patch "btrfs: refactor the error handling of __btrfs_free_extent()" was submitted to make abort_and_dump() to only dump the leaf if the path is not released. So follow the same kernel patch, by only dumping the leaf if the path is not released. Issue: #1064 Pull-request: #1065 Signed-off-by: Qu Wenruo <wqu@suse.com> Signed-off-by: David Sterba <dsterba@suse.com>
kdave
pushed a commit
that referenced
this pull request
Nov 25, 2025
Just follow the kernel patch "btrfs: refactor the error handling of __btrfs_free_extent()", to handle the error first for lookup_extent_backref(), so we can reduce one indent level. Furthermore remove the unnecessary forced type casting of the error message, and replace the old printk() with proper the error() helper. Pull-request: #1065 Signed-off-by: Qu Wenruo <wqu@suse.com> Signed-off-by: David Sterba <dsterba@suse.com>
Owner
|
Merged to devel, thanks. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There is a github bug report about btrfs-convert crash, where
btrfs_print_leaf() is called on NULL path->nodes[0].
The first patch fix the bug by cross-port a fix from kernel part.
The second patch refactor the error handling of __free_extent(), mostly
follow the kernel patch "btrfs: refactor the error handling of __btrfs_free_extent()".