linux-kernel - Re: [PATCH v2] drm/i915: Setting/clearing the memory access bit when en/disabling i915

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>] [day] [month] [year] [list]

Message-ID: <20251009011010.37763-1-safinaskar@gmail.com>
Date: Thu,  9 Oct 2025 04:10:00 +0300
From: Askar Safin <safinaskar@...il.com>
To: jia.yao@...el.com
Cc: alex.zuo@...el.com,
	chris.p.wilson@...ux.intel.com,
	intel-gfx@...ts.freedesktop.org,
	piliu@...hat.com,
	linux-kernel@...r.kernel.org,
	shuicheng.lin@...el.com
Subject: Re: [PATCH v2] drm/i915: Setting/clearing the memory access bit when en/disabling i915

[It seems this letter got lost, so I'm sending it again, this time using git-send-email.
Please, CC lkml additionally next time, hopefully it will work better than intel-gfx@...edesktop ]

On Tue, Oct 7, 2025 at 11:25 PM Jia Yao <jia.yao@...el.com> wrote:
> Make i915's PCI device management more robust by always setting/clearing

Thank you! I tested this patch, and it indeed fixes the problem. You may add:

Tested-By: Askar Safin <safinaskar@...il.com>

Please, CC me with further patches, I will test them.

Also, I will repeat that the bug doesn't happen if we have
CONFIG_INTEL_IOMMU_DEFAULT_ON=y.
So, probably we just need to patch all distro's configs to include
CONFIG_INTEL_IOMMU_DEFAULT_ON=y
(Ubuntu already does that.)
On the other hand, if there is some deeper issue to fix, then, of
course, I'm glad it is being fixed!

Also, please, add some "Fixes" tag and CC stable to make sure this fix
will be backported to stable kernels.

For example:

  Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
  Cc: <stable@...r.kernel.org>

Here I chose the initial git commit, because i915 always was in git history.

-- 
Askar Safin