[ANNOUNCE] xfs: for-next rebased to d4d12c02bf5f

[ANNOUNCE] xfs: for-next rebased to d4d12c02bf5f

[Dave Chinner]

Hi folks,

I just rebased the for-next tree to correct a bad fixes tag in
the tree that was flags by a linux-next sanity check. The code is
the same, just a commit message needed rewriting, but that means all
the commit change and you'll need to do forced update if you pulled
the branch I pushed a few hours ago.

-Dave.

----------------------------------------------------------------

  git://git.kernel.org/pub/scm/fs/xfs/xfs-linux.git for-next

  Head Commit: d4d12c02bf5f768f1b423c7ae2909c5afdfe0d5f

  xfs: collect errors from inodegc for unlinked inode recovery (2023-06-05 14:48:15 +1000)

----------------------------------------------------------------
Darrick J. Wong (1):
      xfs: fix broken logic when detecting mergeable bmap records

Dave Chinner (9):
      xfs: buffer pins need to hold a buffer reference
      xfs: restore allocation trylock iteration
      xfs: defered work could create precommits
      xfs: fix AGF vs inode cluster buffer deadlock
      xfs: fix double xfs_perag_rele() in xfs_filestream_pick_ag()
      xfs: fix agf/agfl verification on v4 filesystems
      xfs: validity check agbnos on the AGFL
      xfs: validate block number being freed before adding to xefi
      xfs: collect errors from inodegc for unlinked inode recovery

Geert Uytterhoeven (1):
      xfs: Fix undefined behavior of shift into sign bit

 fs/xfs/libxfs/xfs_ag.c          |   5 +++-
 fs/xfs/libxfs/xfs_alloc.c       |  91 ++++++++++++++++++++++++++++++++++++++++++++++++++---------------------
 fs/xfs/libxfs/xfs_alloc.h       |   6 ++---
 fs/xfs/libxfs/xfs_bmap.c        |  10 ++++++--
 fs/xfs/libxfs/xfs_bmap_btree.c  |   7 ++++--
 fs/xfs/libxfs/xfs_ialloc.c      |  24 ++++++++++++-------
 fs/xfs/libxfs/xfs_log_format.h  |   9 ++++++-
 fs/xfs/libxfs/xfs_refcount.c    |  13 ++++++++---
 fs/xfs/libxfs/xfs_trans_inode.c | 113 +++++++---------------------------------------------------------------------------------
 fs/xfs/scrub/bmap.c             |  25 ++++++++++----------
 fs/xfs/scrub/scrub.h            |   8 +++----
 fs/xfs/xfs_buf_item.c           |  88 ++++++++++++++++++++++++++++++++++++++++++++++++++------------------
 fs/xfs/xfs_filestream.c         |   1 -
 fs/xfs/xfs_icache.c             |  46 +++++++++++++++++++++++++++++-------
 fs/xfs/xfs_icache.h             |   4 ++--
 fs/xfs/xfs_inode.c              |  20 +++++-----------
 fs/xfs/xfs_inode.h              |   2 +-
 fs/xfs/xfs_inode_item.c         | 149 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 fs/xfs/xfs_inode_item.h         |   1 +
 fs/xfs/xfs_log_recover.c        |  19 +++++++--------
 fs/xfs/xfs_mount.h              |   1 +
 fs/xfs/xfs_reflink.c            |   4 +++-
 fs/xfs/xfs_super.c              |   1 +
 fs/xfs/xfs_trans.c              |   9 ++++++-
 24 files changed, 427 insertions(+), 229 deletions(-)

-- 
Dave Chinner
david@fromorbit.com

Re: [ANNOUNCE] xfs: for-next rebased to d4d12c02bf5f

[Darrick J. Wong]

On Mon, Jun 05, 2023 at 03:07:20PM +1000, Dave Chinner wrote:
> Hi folks,
> 
> I just rebased the for-next tree to correct a bad fixes tag in
> the tree that was flags by a linux-next sanity check. The code is
> the same, just a commit message needed rewriting, but that means all
> the commit change and you'll need to do forced update if you pulled
> the branch I pushed a few hours ago.
> 
> -Dave.
> 
> ----------------------------------------------------------------
> 
>   git://git.kernel.org/pub/scm/fs/xfs/xfs-linux.git for-next
> 
>   Head Commit: d4d12c02bf5f768f1b423c7ae2909c5afdfe0d5f
> 
>   xfs: collect errors from inodegc for unlinked inode recovery (2023-06-05 14:48:15 +1000)
> 
> ----------------------------------------------------------------
> Darrick J. Wong (1):
>       xfs: fix broken logic when detecting mergeable bmap records
> 
> Dave Chinner (9):
>       xfs: buffer pins need to hold a buffer reference
>       xfs: restore allocation trylock iteration
>       xfs: defered work could create precommits
>       xfs: fix AGF vs inode cluster buffer deadlock
>       xfs: fix double xfs_perag_rele() in xfs_filestream_pick_ag()
>       xfs: fix agf/agfl verification on v4 filesystems
>       xfs: validity check agbnos on the AGFL
>       xfs: validate block number being freed before adding to xefi
>       xfs: collect errors from inodegc for unlinked inode recovery
> 
> Geert Uytterhoeven (1):
>       xfs: Fix undefined behavior of shift into sign bit

Hmm, I don't see "xfs: fix ag count overflow during growfs" in here.
Dave, do you want to do another 6.4 bug release, or throw things back
over the wall so I can merge all the rest of the pending fixes for 6.5?

--D

> 
>  fs/xfs/libxfs/xfs_ag.c          |   5 +++-
>  fs/xfs/libxfs/xfs_alloc.c       |  91 ++++++++++++++++++++++++++++++++++++++++++++++++++---------------------
>  fs/xfs/libxfs/xfs_alloc.h       |   6 ++---
>  fs/xfs/libxfs/xfs_bmap.c        |  10 ++++++--
>  fs/xfs/libxfs/xfs_bmap_btree.c  |   7 ++++--
>  fs/xfs/libxfs/xfs_ialloc.c      |  24 ++++++++++++-------
>  fs/xfs/libxfs/xfs_log_format.h  |   9 ++++++-
>  fs/xfs/libxfs/xfs_refcount.c    |  13 ++++++++---
>  fs/xfs/libxfs/xfs_trans_inode.c | 113 +++++++---------------------------------------------------------------------------------
>  fs/xfs/scrub/bmap.c             |  25 ++++++++++----------
>  fs/xfs/scrub/scrub.h            |   8 +++----
>  fs/xfs/xfs_buf_item.c           |  88 ++++++++++++++++++++++++++++++++++++++++++++++++++------------------
>  fs/xfs/xfs_filestream.c         |   1 -
>  fs/xfs/xfs_icache.c             |  46 +++++++++++++++++++++++++++++-------
>  fs/xfs/xfs_icache.h             |   4 ++--
>  fs/xfs/xfs_inode.c              |  20 +++++-----------
>  fs/xfs/xfs_inode.h              |   2 +-
>  fs/xfs/xfs_inode_item.c         | 149 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>  fs/xfs/xfs_inode_item.h         |   1 +
>  fs/xfs/xfs_log_recover.c        |  19 +++++++--------
>  fs/xfs/xfs_mount.h              |   1 +
>  fs/xfs/xfs_reflink.c            |   4 +++-
>  fs/xfs/xfs_super.c              |   1 +
>  fs/xfs/xfs_trans.c              |   9 ++++++-
>  24 files changed, 427 insertions(+), 229 deletions(-)
> 
> -- 
> Dave Chinner
> david@fromorbit.com

Re: [ANNOUNCE] xfs: for-next rebased to d4d12c02bf5f

[Dave Chinner]

On Mon, Jun 05, 2023 at 05:09:51PM -0700, Darrick J. Wong wrote:
> On Mon, Jun 05, 2023 at 03:07:20PM +1000, Dave Chinner wrote:
> > Hi folks,
> > 
> > I just rebased the for-next tree to correct a bad fixes tag in
> > the tree that was flags by a linux-next sanity check. The code is
> > the same, just a commit message needed rewriting, but that means all
> > the commit change and you'll need to do forced update if you pulled
> > the branch I pushed a few hours ago.
> > 
> > -Dave.
> > 
> > ----------------------------------------------------------------
> > 
> >   git://git.kernel.org/pub/scm/fs/xfs/xfs-linux.git for-next
> > 
> >   Head Commit: d4d12c02bf5f768f1b423c7ae2909c5afdfe0d5f
> > 
> >   xfs: collect errors from inodegc for unlinked inode recovery (2023-06-05 14:48:15 +1000)
> > 
> > ----------------------------------------------------------------
> > Darrick J. Wong (1):
> >       xfs: fix broken logic when detecting mergeable bmap records
> > 
> > Dave Chinner (9):
> >       xfs: buffer pins need to hold a buffer reference
> >       xfs: restore allocation trylock iteration
> >       xfs: defered work could create precommits
> >       xfs: fix AGF vs inode cluster buffer deadlock
> >       xfs: fix double xfs_perag_rele() in xfs_filestream_pick_ag()
> >       xfs: fix agf/agfl verification on v4 filesystems
> >       xfs: validity check agbnos on the AGFL
> >       xfs: validate block number being freed before adding to xefi
> >       xfs: collect errors from inodegc for unlinked inode recovery
> > 
> > Geert Uytterhoeven (1):
> >       xfs: Fix undefined behavior of shift into sign bit
> 
> Hmm, I don't see "xfs: fix ag count overflow during growfs" in here.

No, I didn't pick it up because it conflicted with other bug fix
stuff I am currently working on and I needed to look at it in more
detail before doing anything with it. I hadn't followed the
development of the patch at all, and it was up to v4 so I was going
to need to spend a little bit of time on it to see what the history
of it was first....

> Dave, do you want to do another 6.4 bug release, or throw things back
> over the wall so I can merge all the rest of the pending fixes for 6.5?

If you want, you can pick it up once I've sent a pull request for
the current set of fixes in for-next. That will be later this week;
it needs to spend a couple of days in linux-next before that
happens, though.

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

Re: [ANNOUNCE] xfs: for-next rebased to d4d12c02bf5f

[Darrick J. Wong]

On Tue, Jun 06, 2023 at 10:59:35AM +1000, Dave Chinner wrote:
> On Mon, Jun 05, 2023 at 05:09:51PM -0700, Darrick J. Wong wrote:
> > On Mon, Jun 05, 2023 at 03:07:20PM +1000, Dave Chinner wrote:
> > > Hi folks,
> > > 
> > > I just rebased the for-next tree to correct a bad fixes tag in
> > > the tree that was flags by a linux-next sanity check. The code is
> > > the same, just a commit message needed rewriting, but that means all
> > > the commit change and you'll need to do forced update if you pulled
> > > the branch I pushed a few hours ago.
> > > 
> > > -Dave.
> > > 
> > > ----------------------------------------------------------------
> > > 
> > >   git://git.kernel.org/pub/scm/fs/xfs/xfs-linux.git for-next
> > > 
> > >   Head Commit: d4d12c02bf5f768f1b423c7ae2909c5afdfe0d5f
> > > 
> > >   xfs: collect errors from inodegc for unlinked inode recovery (2023-06-05 14:48:15 +1000)
> > > 
> > > ----------------------------------------------------------------
> > > Darrick J. Wong (1):
> > >       xfs: fix broken logic when detecting mergeable bmap records
> > > 
> > > Dave Chinner (9):
> > >       xfs: buffer pins need to hold a buffer reference
> > >       xfs: restore allocation trylock iteration
> > >       xfs: defered work could create precommits
> > >       xfs: fix AGF vs inode cluster buffer deadlock
> > >       xfs: fix double xfs_perag_rele() in xfs_filestream_pick_ag()
> > >       xfs: fix agf/agfl verification on v4 filesystems
> > >       xfs: validity check agbnos on the AGFL
> > >       xfs: validate block number being freed before adding to xefi
> > >       xfs: collect errors from inodegc for unlinked inode recovery
> > > 
> > > Geert Uytterhoeven (1):
> > >       xfs: Fix undefined behavior of shift into sign bit
> > 
> > Hmm, I don't see "xfs: fix ag count overflow during growfs" in here.
> 
> No, I didn't pick it up because it conflicted with other bug fix
> stuff I am currently working on and I needed to look at it in more
> detail before doing anything with it. I hadn't followed the
> development of the patch at all, and it was up to v4 so I was going
> to need to spend a little bit of time on it to see what the history
> of it was first....

Ah, ok.  Most of the history was the author and I going 'round and
'round about how to validate the incoming fsblocks to prevent agcount
overflow without stomping on any other weird uses.

> > Dave, do you want to do another 6.4 bug release, or throw things back
> > over the wall so I can merge all the rest of the pending fixes for 6.5?
> 
> If you want, you can pick it up once I've sent a pull request for
> the current set of fixes in for-next. That will be later this week;
> it needs to spend a couple of days in linux-next before that
> happens, though.

Yeah, sounds good to me.

--D

> Cheers,
> 
> Dave.
> -- 
> Dave Chinner
> david@fromorbit.com