| Message ID | 20250415063342.3033071-1-davechen@synology.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | [v2] btrfs: fix COW handling in run_delalloc_nocow function | expand |
On Tue, Apr 15, 2025 at 7:34 AM davechen <davechen@synology.com> wrote: > > In run_delalloc_nocow(), when the found btrfs_key's offset > cur_offset, > it indicates a gap between the current processing region and > the next file extent. The original code would directly jump to > the "must_cow" label, which increments the slot and forces a fallback > to COW. This behavior might skip an extent item and result in an > overestimated COW fallback range. > > This patch modifies the logic so that when a gap is detected: > - If no COW range is already being recorded (cow_start is unset), > cow_start is set to cur_offset. > - cur_offset is then advanced to the beginning of the next extent. > - Instead of jumping to "must_cow", control flows directly to > "next_slot" so that the same extent item can be reexamined properly. > > The change ensures that we accurately account for the extent gap and > avoid accidentally extending the range that needs to fallback to COW. > > Signed-off-by: Dave Chen <davechen@synology.com> > --- In the future, don't forget to add here, after the "---", a description of what changed in this version of the patch compared to the previous one (look for examples in the mailing list). > fs/btrfs/inode.c | 9 +++++---- > 1 file changed, 5 insertions(+), 4 deletions(-) > > diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c > index 5b842276573e..394d5113b6cb 100644 > --- a/fs/btrfs/inode.c > +++ b/fs/btrfs/inode.c > @@ -2155,12 +2155,13 @@ static noinline int run_delalloc_nocow(struct btrfs_inode *inode, > > /* > * If the found extent starts after requested offset, then > - * adjust extent_end to be right before this extent begins > + * adjust cur_offset to be right before this extent begins I'll add punctuation to the end of the sentence to make it comply with the preferred coding style. Reviewed-by: Filipe Manana <fdmanana@suse.com> Patch looks good to me now, I'll add it to for-next, thanks. > */ > if (found_key.offset > cur_offset) { > - extent_end = found_key.offset; > - extent_type = 0; > - goto must_cow; > + if (cow_start == (u64)-1) > + cow_start = cur_offset; > + cur_offset = found_key.offset; > + goto next_slot; > } > > /* > -- > 2.43.0 > >
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 5b842276573e..394d5113b6cb 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -2155,12 +2155,13 @@ static noinline int run_delalloc_nocow(struct btrfs_inode *inode, /* * If the found extent starts after requested offset, then - * adjust extent_end to be right before this extent begins + * adjust cur_offset to be right before this extent begins */ if (found_key.offset > cur_offset) { - extent_end = found_key.offset; - extent_type = 0; - goto must_cow; + if (cow_start == (u64)-1) + cow_start = cur_offset; + cur_offset = found_key.offset; + goto next_slot; } /*
In run_delalloc_nocow(), when the found btrfs_key's offset > cur_offset, it indicates a gap between the current processing region and the next file extent. The original code would directly jump to the "must_cow" label, which increments the slot and forces a fallback to COW. This behavior might skip an extent item and result in an overestimated COW fallback range. This patch modifies the logic so that when a gap is detected: - If no COW range is already being recorded (cow_start is unset), cow_start is set to cur_offset. - cur_offset is then advanced to the beginning of the next extent. - Instead of jumping to "must_cow", control flows directly to "next_slot" so that the same extent item can be reexamined properly. The change ensures that we accurately account for the extent gap and avoid accidentally extending the range that needs to fallback to COW. Signed-off-by: Dave Chen <davechen@synology.com> --- fs/btrfs/inode.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-)