linux-kernel - Re: [PATCH 2/3] drm/nouveau/disp: Always accept linear modifier

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: <DC99MMSCWQSM.5KF4ISD63U4X@kernel.org>
Date: Fri, 22 Aug 2025 23:14:14 +0200
From: "Danilo Krummrich" <dakr@...nel.org>
To: "James Jones" <jajones@...dia.com>
Cc: "Lyude Paul" <lyude@...hat.com>, "Faith Ekstrand"
 <faith.ekstrand@...labora.com>, <nouveau@...ts.freedesktop.org>,
 <dri-devel@...ts.freedesktop.org>, <linux-kernel@...r.kernel.org>, "David
 Airlie" <airlied@...il.com>, "Simona Vetter" <simona@...ll.ch>, "Joel
 Fernandes" <joelagnelf@...dia.com>
Subject: Re: [PATCH 2/3] drm/nouveau/disp: Always accept linear modifier

On Fri Aug 22, 2025 at 11:11 PM CEST, James Jones wrote:
> On 8/22/25 13:55, Danilo Krummrich wrote:
>> On Tue Aug 12, 2025 at 12:00 AM CEST, James Jones wrote:
>>> On some chipsets, which block-linear modifiers are
>>> supported is format-specific. However, linear
>>> modifiers are always be supported. The prior
>>> modifier filtering logic was not accounting for
>>> the linear case.
>>>
>>> Fixes: c586f30bf74c ("drm/nouveau/kms: Add format mod prop to base/ovly/nvdisp")
>>> Signed-off-by: James Jones <jajones@...dia.com>
>> 
>> This issue seems to be present since v5.10, what's the implication of this? I
>> assume this has to be backported into stable releases?
>> 
>> Does the subsequent patch break strictly depend on this fix, or can it go
>> separately?
>
> Without this fix, the next patch breaks linear modifier use on 
> Blackwell2+. In my testing, that meant fbcon was severely corrupted (In 
> a manner that suggests it ends up with a block-linear surface rendered 
> to as if it was linear).
>
> Yes, it has to go back to a fair number of stable branches to fix 
> similar issues on pre-fermi GPUs, though oddly in my testing 
> before/after this patch, fbcon came up fine on my NV50, so the effects 
> might not be as severe there for some reason.

Ok, thanks! This sounds like we should apply the fix, backmerge the -rc it lands
in and then merge the rest of this series.

- Danilo