Skip to content

Fix fsync bug in ext3 and ext4

Jan Kara linux-ext4 1 version In review

Get series with b4
b4 am [email protected]

Versions newest first

  1. v1 Latest4 patches 4 repliesIn review

Patches in v1 No applies recorded yet

#Subject and filesSizeRepliesApplied
1/4ext3: Wait for proper transaction commit on fsyncfs/ext3/fsync.c fs/ext3/inode.c include/linux/ext3_fs_i.h+58 -210 replies
2/4ext4: Avoid issuing barriers on error recovery pathfs/ext4/fsync.c+3 -20 replies
3/4ext4: Fix error handling ext4_ind_get_blocksfs/ext4/inode.c+1 -10 replies
4/4ext4: Wait for proper transaction commit on fsyncfs/ext4/ext4.h fs/ext4/extents.c fs/ext4/fsync.c +1 more+63 -234 replies

Reviews trailers per patch

No Reviewed-by, Acked-by or Tested-by trailers for v1 yet.

Read the v1 thread

Cover letter [PATCH 0/4] Fix fsync bug in ext3 and ext4

Hi,

the patch series below fixes the problem ext3 and ext4 has that we
rely on inode dirty bits for correct fsync. This has two problems:
a) ext3 and ext4 often does not set them (as they call ext?_mark_inode_dirty) and thus only quota code sets the dirty bit as a sideeffect.
b) pdflush can come in and clear the dirty bit any time (thanks to Chris for pointing out this).

The second and third patch in the series are just minor bug fixes to ext4.
I did some basic testing to verify that the code does what I'd expect it to do but an extra pair of eyes checking the new code would be helpful...

Honza