The latest update for the open source graphics drivers is here with Mesa 24.2.7 out now. Plus we might see Mesa 24.3 a bit sooner than expected.
For Mesa 24.2.7 the release notes mention these fixes:
- brw_fs_opt_copy_propagation incorrectly handles size changes of uniforms.
- nouveau paraview msaa corruption 23.1 bisected regression.
- Missing xshmfence dependency for X11 WSI on macOS.
- [AMDGPU RDNA3] Black square artifacts in viewport renders in Blender.
- radv: test_sm67_sample_cmp_level test in vkd3d-proton fails.
- [ANV] LNL tiled corruption in background traci Blackops3-trace-dx11-1080p-high.
- [anv] Possible regression in e3814dee1ac0f90771b921a4f6f5aed10f06e8d4.
- Confidential issue #12092.
No new features, since it's a point-release for the last stable release.
However, we also have news that Mesa 24.3 development has gone seemingly better than expected, which will have new features. In the release announcement email for the second Release Candidate, developer Dylan Baker mentioned:
We have a fairly small RC2 today, with a bit of work across the board. We currently only have 2 open blocking issues, and given that we delayed the branchpoint by 3 weeks, I wouldn't object to pulling in the RC period by 1 week and have rc3 rather than rc4 be the final release, if we can get all of the issues closed in time.
Based on the release schedule, that might then put the Mesa 24.3 release around November 20th.
QuoteConfidential issue #12092
What? I switched to AMD specifically to use drivers that offer a first-class FOSS experience, and to avoid this kind of vague chicanery.
Part of the point of FOSS is, you know, being open about issues! I get paranoid about this crap ever since I first started reading the likes of Brian Krebs and Bruce Schneier years ago.
Quoting: redneckdrowIt's a common practice in almost any opensource when dealing with vulnerabilities.QuoteConfidential issue #12092
What? I switched to AMD specifically to use drivers that offer a first-class FOSS experience, and to avoid this kind of vague chicanery.
Part of the point of FOSS is, you know, being open about issues! I get paranoid about this crap ever since I first started reading the likes of Brian Krebs and Bruce Schneier years ago.
Quoting: redneckdrowThis seems to be the fix for the confidential bug:QuoteConfidential issue #12092
What? I switched to AMD specifically to use drivers that offer a first-class FOSS experience, and to avoid this kind of vague chicanery.
Part of the point of FOSS is, you know, being open about issues! I get paranoid about this crap ever since I first started reading the likes of Brian Krebs and Bruce Schneier years ago.
https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31914
It's related to WebGL so I'm guessing a security issue in a browser, that's about to be disclosed?
Last edited by whizse on 14 November 2024 at 4:39 pm UTC
Quoting: whizseQuoting: redneckdrowThis seems to be the fix for the confidential bug:QuoteConfidential issue #12092
What? I switched to AMD specifically to use drivers that offer a first-class FOSS experience, and to avoid this kind of vague chicanery.
Part of the point of FOSS is, you know, being open about issues! I get paranoid about this crap ever since I first started reading the likes of Brian Krebs and Bruce Schneier years ago.
https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31914
It's related to WebGL so I'm guessing a security issue in a browser, that's about to be disclosed?
Kind of, based on their description it's WebGL reachable, implying that it's a wider attack demonstrated to them in the form of webGL code, which is an understandable angle to take, because webGL is basically the least trusted rendering tool in the entire system.
Edit: based on the patch I would guess it's a buffer overflow attack.
Last edited by LoudTechie on 14 November 2024 at 6:53 pm UTC
See more from me