linux-kernel - Re: [PATCH v4 0/3] vfio: handle DMA map/unmap up to the addressable limit

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <aPe0E6Jj9BJA2Bd5@devgpu012.nha5.facebook.com>
Date: Tue, 21 Oct 2025 09:25:55 -0700
From: Alex Mastro <amastro@...com>
To: Alex Williamson <alex@...zbot.org>
CC: Alejandro Jimenez <alejandro.j.jimenez@...cle.com>,
        Jason Gunthorpe
	<jgg@...pe.ca>, <kvm@...r.kernel.org>,
        <linux-kernel@...r.kernel.org>, David
 Matlack <dmatlack@...gle.com>
Subject: Re: [PATCH v4 0/3] vfio: handle DMA map/unmap up to the addressable
 limit

On Mon, Oct 20, 2025 at 03:36:33PM -0600, Alex Williamson wrote:
> I do note that we're missing a Fixes: tag.  I think we've had hints of
> this issue all the way back to the original implementation, so perhaps
> the last commit should include:
> 
> Fixes: 73fa0d10d077 ("vfio: Type1 IOMMU implementation")

SGTM

> Unless you've identified a more specific target.

I have not.

> Along with the tag, it would probably be useful in that same commit to
> expand on the scope of the issue in the commit log.  I believe we allow
> mappings to be created at the top of the address space that cannot be
> removed via ioctl, but such inconsistency should result in an
> application error due to the failed ioctl and does not affect cleanup
> on release.

Makes sense. I will clarify the commit msg in v5 to be more specific about what
this series changes relative to existing functionality.

> Should we also therefore expand the DMA mapping tests in
> tools/testing/selftests/vfio to include an end of address space test?

Yes. I will append such a commit to the end of the series in v5. Our VFIO tests
are built on top of a hermetic rust wrapper library over VFIO ioctls, but they
aren't quite ready to be open sourced yet.