Emacs ELPA policy fights and DRM atomic infrastructure

Emacs packaging policy around LLM-assisted code and several large DRM patch series dominated Desktop and Graphics traffic. NonGNU ELPA submissions drew maintainer disagreement while kernel graphics work advanced atomic state handling and driver support.

NonGNU ELPA clutch submission draws RMS opposition

A submission inquiry for clutch, an interactive database client, reached NonGNU ELPA discussion on emacs-devel and produced a multi-message policy exchange. RMS opposed inclusion while Stefan and Eli pointed to NonGNU exemptions. The thread underscores unresolved rules for packages that rely on LLM assistance.

DRM atomic state readout reaches v7

A 34-patch v7 series on dri-devel implements core DRM atomic state readout infrastructure together with fastboot-style initialization for TIDSS and bridges. The work supplies missing readout paths that drivers need for reliable atomic modesetting bring-up. Readers tracking display stack correctness will care about the new core helpers and their first consumers.

Explicit kobj_type release and __must_init

A nine-patch v2 series adds a __must_init attribute and requires every kobj_type initializer to supply an explicit release method. The changes target latent use-after-free bugs that span several subsystems. Hardening of object lifetime rules matters for long-term driver stability.

Fork threats over Emacs LLM policy

Users on emacs-devel threatened an Emacs fork in response to LLM code policy, citing community harm that they said extends beyond copyright. A maintainer pressed for concrete descriptions of those harms. The heated exchange shows how packaging rules are colliding with contributor expectations.

ekp package and Knuth-Plass policy debate

A NonGNU ELPA proposal for ekp, a Knuth-Plass line-breaking package, reopened LLM-generated code policy and naming questions with RMS participating. The discussion repeats the same inclusion tensions seen in the clutch thread. Package authors and ELPA maintainers both need clearer boundaries.

DRM panels move into atomic commits

Dmitry Baryshkov posted a 35-patch series that lets DRM panels join atomic commits by embedding bridges rather than extending panel function tables. The design keeps the panel API narrow while giving panels full atomic state participation. Panel and bridge maintainers gain a cleaner path for atomic conversion.

TIDSS BeagleY-AI and AM62P display support

A 14-patch v4 series adds tidss support for AM62P and J722S SoCs plus BeagleY-AI HDMI, OLDI, and DPI outputs along with related cleanups. The changes expand mainline coverage for Texas Instruments display hardware used on popular boards. Board bring-up and embedded graphics users benefit directly.

DMA-BUF CPU sync flag for videobuf2

A four-patch series proposes an opt-in DMA-BUF CPU cache synchronization flag for videobuf2. Multiple reviewers questioned the design and offered alternative approaches. Media and graphics buffer-sharing paths remain under active scrutiny for correctness.