Hi Linus,
please consider pulling the following gfs2 fixes for -rc2. The one thing that
stands out is this commit:
gfs2: Prevent endless loops in gfs2_file_buffered_write
We have a choice here between the proper fix in iomap_write_iter and
generic_perform_write and this patch, which is just a workaround. I've tried
to get your feedback on this before:
https://lore.kernel.org/all/[email protected]/
To me, it seems to be too late for the proper fix for v5.16 now; please advise
if you think otherwise.
Thanks,
Andreas
The following changes since commit 78805cbe5d72ad27a56962a8072edbcb45ca1180:
Merge tag 'gfs2-v5.15-rc5-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2 (2021-11-02 12:35:04 -0700)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2.git tags/gfs2-v5.16-rc2-fixes
for you to fetch changes up to 554c577cee95bdc1d03d9f457e57dc96eb791845:
gfs2: Prevent endless loops in gfs2_file_buffered_write (2021-11-10 18:22:37 +0100)
----------------------------------------------------------------
Fixes in gfs2:
* The current iomap_file_buffered_write behavior of failing the entire
write when part of the user buffer cannot be faulted in leads to an
endless loop in gfs2. Work around that in gfs2 for now.
* Various other bugs all over the place.
----------------------------------------------------------------
Andreas Gruenbacher (5):
gfs2: Only dereference i->iov when iter_is_iovec(i)
gfs2: Fix atomic bug in gfs2_instantiate
gfs2: Fix length of holes reported at end-of-file
gfs2: Fix "Introduce flag for glock holder auto-demotion"
gfs2: Prevent endless loops in gfs2_file_buffered_write
Bob Peterson (1):
gfs2: release iopen glock early in evict
fs/gfs2/bmap.c | 2 +-
fs/gfs2/file.c | 9 ++++++---
fs/gfs2/glock.c | 12 ++++--------
fs/gfs2/super.c | 14 +++++++-------
4 files changed, 18 insertions(+), 19 deletions(-)
On Wed, Nov 17, 2021 at 3:37 PM Andreas Gruenbacher <[email protected]> wrote:
>
> We have a choice here between the proper fix in iomap_write_iter and
> generic_perform_write and this patch, which is just a workaround. I've tried
> to get your feedback on this before:
Well, you did get Catalin's feedback, and I didn't have any strong
opinions on it, so..
I do think that iomap_write_iter() should be fixed, but I also agree
with your "by now too late", so I've pulled your gfs2 workaround.
Linus
The pull request you sent on Thu, 18 Nov 2021 00:36:56 +0100:
> git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2.git tags/gfs2-v5.16-rc2-fixes
has been merged into torvalds/linux.git:
https://git.kernel.org/torvalds/c/42eb8fdac2fc5d62392dcfcf0253753e821a97b0
Thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/prtracker.html