On 01/09/2026 09:12, Thorsten Leemhuis wrote:
On 8/28/26 14:19, Viktor Jägersküpper wrote:
On 8/28/26 09:22, Thorsten Leemhuis wrote:
On 8/27/26 19:14, Melissa Wen wrote:
On 27/08/2026 18:37, Melissa Wen wrote:
On 27/08/2026 15:34, Alex Deucher wrote:
On Wed, Aug 26, 2026 at 7:20 AM Melissa Wen <[email protected]> wrote:
PREMULTI was the default DRM blend mode until 9813e158d13d
("drm/drm_blend: allow blend mode property without PREMULTI")
introduced
the possibility of exposing the blend mode property without PREMULTI
being supported. However, __drm_atomic_helper_plane_state_init() still
resets pixel_blend_mode to PREMULTI (hardcoded), ignoring the supported
mode in this property. In the same series, 860e748bddcc ("drm: ensure
blend mode supported if pixel format with alpha exposed") starts
warning
drivers that supports alpha formats in a given planes but doesn't
support blend mode property, which includes AMD primary and cursor
planes.
[...]
I presume you are planning to apply this directly to drm-misc for 7.3?
Right, I'll apply to drm-misc-fixes, just doing it now.
Or not now. I have to wait for -rc1.
I'll keep it on my radar.
Hmmm, that sounds like these fixes will miss -rc1. But wouldn't it be
good to get them included there, as they afaik (please correct me if I'm
wrong, is easy to mix something up from the outside!) the warnings that
Greg described as ""It's not good to boot and have a bunch of warnings
already happening, resulting in a "tainted" kernel before I have the
chance to break anything manually myself...""[1], as more people then
will likely encounter this (and then waste time on bisecting an issue
known for weeks for which a fix exists).
[1]
https://lore.kernel.org/all/2026082651-spinning-cornbread-33ba@gregkh/
If I understand correctly, the *actual* fix (for all affected drivers)
will make it into 7.3-rc1, see this answer to my question:
https://lore.kernel.org/dri-devel/20260826-brave-nyala-of-elevation-d23aea@houat/
This fix is queued in drm-next:
https://gitlab.freedesktop.org/drm/kernel/-/commit/271e90eb5f9ff34951647e5ed33c1775eebcca50
Ahh, yeah, many thx, I had see this downgrade to drm_warn() (which made
it to -rc1, yeah \o/), but had not made the right connections. :-/
FWIW, even with that done there is now one amdgpu regression report
about this already post-rc1:
https://lore.kernel.org/all/CANkdJ2VvJQSBzyOq=LgSW1=z8-yin_r9py2ij0ujdo55+db...@mail.gmail.com/
"""
[...] After updating from kernel v7.2.2 to v7.3-rc1, several DRM plane
warnings regarding unset blend modes for pixel formats with alpha
channels started appearing in dmesg during initialization.
[drm] [PLANE:44:plane-0] pixel format with alpha exposed but blend
mode not setup. Please fix. [...]
"""
Will reply there now. Wondering if we need to do something about that if
more reports like that show up, but let's wait if that's the case.
I just applied this series to drm-misc-fixes:
- c3080b58d81d ("drm/atomic-state-helper: set pixel_blend_mode to prop
default on reset")
- f0c75da0a6b4 ("drm/amd/display: fix missing blend-mode-prop warning
for DCN")
- 332ad707e38f ("drm/amd/display: advertise PIXEL_NONE and PREMULTI
blend mode for DCE")
With this, the dmesg message will no longer appear on the AMD driver.
Melissa
Ciao, Thorsten
