News: 0001565042

  ARM Give a man a fire and he's warm for a day, but set fire to him and he's warm for the rest of his life (Terry Pratchett, Jingo)

xf86-video-amdgpu 25.0.0 Released With Two Years Worth Of Fixes

([Radeon] 4 Hours Ago xf86-video-amdgpu 25.0.0)


It had been two years since the last update to the AMDGPU X.Org DDX driver but now xf86-video-amdgpu 25.0.0 is now available for those relying on this driver/hardware-specific driver for X.Org enabled Linux systems rather than the xf86-video-modesetting generic driver or a Wayland-based desktop.

Coming as a surprise is xf86-video-amdgpu 25.0 as the successor to xf86-video-amdgpu 23.0 that released back in February 2023... While two years have passed, there still aren't too many changes in this xf86-video-amdgpu 25.0 release.

The updated xf86-video-amdgpu 25.0 release drops compatibility with "ancient" X.Org Server versions, brings various low-level fixes, GitLab / Continuous Integration churn, and improved man page formatting. For those building xf86-video-amdgpu 25.0, it does drop the Autotools build system support and now depends upon Meson.

[1]

Arguably the two most interesting patches of xf86-video-amdgpu 25.0 are the handling of color management properties and correctly checking the GFX12 (RDNA4) swizzle mode.

For those making use of the xf86-video-amdgpu 25.0 DDX, downloads and more details via the [2]mailing list announcement .

While two years between DDX releases may seem like a long time, it's now been ten years since [3]the last xf86-video-intel release that itself was just a development snapshot, due to now just focusing on the generic xf86-video-modesetting DDX for X.Org Server Intel graphics customers.



[1] https://www.phoronix.com/image-viewer.php?id=2025&image=xf86_amdgpu_25_lrg

[2] https://lists.freedesktop.org/archives/dri-devel/2025-July/517470.html

[3] https://gitlab.freedesktop.org/xorg/driver/xf86-video-intel/-/tags



ihatemichael

aviallon

davidbepo

zaps166

Svyatko

Risch's decision procedure for integration, not surprisingly,
uses a recursion on the number and type of the extensions from the
rational functions needed to represent the integrand. Although the
algorithm follows and critically depends upon the appropriate structure
of the input, as in the case of multivariate factorization, we cannot
claim that the algorithm is a natural one. In fact, the creator of
differential algebra, Ritt, committed suicide in the early 1950's,
largely, it is claimed, because few paid attention to his work. Probably
he would have received more attention had he obtained the algorithm as well.
-- Joel Moses, "Algorithms and Complexity", ed. J. F. Traub