Bug#1079394: linux-image-6.10.6-amd64: causes cifs regression, flatpak & ostree signature corruption

[Salvatore Bonaccorso <carnil@debian.org>]

Hi

On Sat, Aug 24, 2024 at 01:08:01AM -0700, Forest wrote:
> After further testing today, that commit also seems to be (intermittently)
> causing ffmpeg and mkvmerge to write corrupt files to cifs mounts.
> 
> 
> On Fri, 23 Aug 2024 22:13:07 +0200, Salvatore Bonaccorso wrote:
> 
> >Since you identified this as upstream regression including the
> >breaking commit, can you please report this upstream (and please do
> >keep us in the loop).
> 
> That is my plan. Last time I reported a bug to the cifs team, it was
> ignored, but perhaps I'll have more luck through the additional email
> addresses you provided. Thanks for those.
> 
> 
> In the meantime, I'm in an awkward position on Debian: The Stable kernel is
> too old for my GPU, and both Testing and Unstable now have kernels with this
> data-corrupting bug.
> 
> Looks like I'm stuck with custom kernel builds until this is sorted out.
> 6.9.12 doesn't have the bug, but it's labeled EOL on kernel.org, which I
> think means no more security updates. I guess I'll have to try the 6.6
> branch.

AFIU, this should be fixed with c26096ee0278 ("mm: Fix
filemap_invalidate_inode() to use invalidate_inode_pages2_range()")
upstream, correct?

Regards,
Salvatore