Re: [PATCH 5.15.y] xfs: fix super block buf log item UAF during force shutdown

[Greg KH <gregkh@linuxfoundation.org>]

On Fri, Jul 04, 2025 at 02:15:46PM +0530, Pranav Tyagi wrote:
> On Fri, Jul 4, 2025 at 1:58 PM Greg KH <gregkh@linuxfoundation.org> wrote:
> >
> > On Fri, Jul 04, 2025 at 01:16:01PM +0530, Pranav Tyagi wrote:
> > > On Wed, Jun 25, 2025 at 7:39 PM Sasha Levin <sashal@kernel.org> wrote:
> > > >
> > > > [ Sasha's backport helper bot ]
> > > >
> > > > Hi,
> > > >
> > > > ✅ All tests passed successfully. No issues detected.
> > > > No action required from the submitter.
> > > >
> > > > The upstream commit SHA1 provided is correct: 575689fc0ffa6c4bb4e72fd18e31a6525a6124e0
> > > >
> > > > WARNING: Author mismatch between patch and upstream commit:
> > > > Backport author: Pranav Tyagi<pranav.tyagi03@gmail.com>
> > > > Commit author: Guo Xuenan<guoxuenan@huawei.com>
> > > >
> > > > Status in newer kernel trees:
> > > > 6.15.y | Present (exact SHA1)
> > > > 6.12.y | Present (exact SHA1)
> > > > 6.6.y | Present (exact SHA1)
> > > > 6.1.y | Present (different SHA1: 0d889ae85fcf)
> > > >
> > > > Note: The patch differs from the upstream commit:
> > > > ---
> > > > 1:  575689fc0ffa6 ! 1:  9876b048d8f68 xfs: fix super block buf log item UAF during force shutdown
> > > >     @@ Metadata
> > > >       ## Commit message ##
> > > >          xfs: fix super block buf log item UAF during force shutdown
> > > >
> > > >     +    [ Upstream commit 575689fc0ffa6c4bb4e72fd18e31a6525a6124e0 ]
> > > >     +
> > > >          xfs log io error will trigger xlog shut down, and end_io worker call
> > > >          xlog_state_shutdown_callbacks to unpin and release the buf log item.
> > > >          The race condition is that when there are some thread doing transaction
> > > >     @@ Commit message
> > > >          ==================================================================
> > > >          Disabling lock debugging due to kernel taint
> > > >
> > > >     +    [ Backport to 5.15: context cleanly applied with no semantic changes.
> > > >     +    Build-tested. ]
> > > >     +
> > > >          Signed-off-by: Guo Xuenan <guoxuenan@huawei.com>
> > > >          Reviewed-by: Darrick J. Wong <djwong@kernel.org>
> > > >          Signed-off-by: Darrick J. Wong <djwong@kernel.org>
> > > >     +    Signed-off-by: Pranav Tyagi <pranav.tyagi03@gmail.com>
> > > >
> > > >       ## fs/xfs/xfs_buf_item.c ##
> > > >      @@ fs/xfs/xfs_buf_item.c: xfs_buf_item_relse(
> > > > ---
> > > >
> > > > Results of testing on various branches:
> > > >
> > > > | Branch                    | Patch Apply | Build Test |
> > > > |---------------------------|-------------|------------|
> > > > | stable/linux-5.15.y       |  Success    |  Success   |
> > >
> > > Hi,
> > >
> > > Just following up on this 5.15.y backport.
> > > Please let me know if anything else is needed from my side.
> >
> > xfs patches need to go through the xfs maintainers for their approval.
> > "build tested" just doesn't cut it at all, you MUST actually test this
> > at runtime.  Otherwise, why would you even want this patch applied if
> > you don't have xfs systems that you care about?
> >
> > thanks,
> >
> > greg k-h
> 
> Hi Greg,
> 
> Apologies for the oversight. I had assumed that a build test would
> suffice for a backport since the patch is already merged upstream. I now
> understand the importance of runtime testing, even for backports. I’ll
> test it on the XFS setup and resend the patch.

Why do you want/need this backported if it wasn't even tested?  Why do
the backport at all?

confused,

greg k-h