linux-kernel - Re: [GIT PULL] MM updates for 6.19-rc1

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20251206025434.GA1422789@nvidia.com>
Date: Fri, 5 Dec 2025 22:54:34 -0400
From: Jason Gunthorpe <jgg@...dia.com>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: Andrew Morton <akpm@...ux-foundation.org>,
	Ankit Agrawal <ankita@...dia.com>,
	Alex Williamson <alex@...zbot.org>, linux-mm@...ck.org,
	linux-kernel@...r.kernel.org, mm-commits@...r.kernel.org
Subject: Re: [GIT PULL] MM updates for 6.19-rc1

On Fri, Dec 05, 2025 at 01:56:17PM -0800, Linus Torvalds wrote:
> On Wed, 3 Dec 2025 at 21:29, Andrew Morton <akpm@...ux-foundation.org> wrote:
> >
> > Ankit Agrawal:
> >       vfio/nvgrace-gpu: register device memory for poison handling
> 
> So I had to look at this one because there was a conflict ("int ret"
> had been removed by another commit), and after looking at it I ended
> up deciding to revert this patch entirely.
> 
> I may be wrong, but that patch looks *COMPETELY* broken.

You are right, it is security broken, it worked functionally in
testing because it used qemu which only mmaps once. mmaping twice will
corrupt an interval tree.

Alex and I also noticed it this week also when we were looking at the
same conflcit. We also found another serious bug in how it uses pgoff
that renders it functionally broken after merging all the branches
together as the merge'd tree also changes how pgoff worked. :\

Ankit has been working on a fix this week and we planned to send you a
fix/revert in the next week, travel permitting and so on.

> Again - maybe I'm missing something. If so, that commit can be re-done
> with explanations of why it's all correct.

Reverting it now works well enough too. Let's see if Ankit can prepare
a new revision and perhaps Alex can forward it next week or very early
rc1 after LPC.

Regards,
Jason