From mboxrd@z Thu Jan 1 00:00:00 1970 From: Claude Code Review Bot To: dri-devel-reviews@example.com Subject: Claude review: drm/xe: Don't emit UTS_RELEASE as part of device coredump Date: Tue, 28 Apr 2026 14:25:21 +1000 Message-ID: In-Reply-To: <20260427160902.1126027-2-u.kleine-koenig@baylibre.com> References: <20260427160902.1126027-2-u.kleine-koenig@baylibre.com> X-Mailer: Claude Code Patch Reviewer Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Overall Series Review Subject: drm/xe: Don't emit UTS_RELEASE as part of device coredump Author: =3D?UTF-8?q?Uwe=3D20Kleine-K=3DC3=3DB6nig=3D20=3D28The=3D20Capable= =3D20Hub=3D29?=3D Patches: 3 Reviewed: 2026-04-28T14:25:21.271181 --- This is a single-patch series that removes the `UTS_RELEASE` (kernel versio= n string) from the Xe device coredump output. The stated motivation is buil= d reproducibility: `UTS_RELEASE` changes with any uncommitted change or new= commit, causing binary differences that are noise when evaluating whether = a source patch changes the compiled output. The patch is technically trivial and correct =E2=80=94 it cleanly removes o= ne line. However, the author themselves express doubt in the cover letter: = *"I'm not really sure if my assertion above that the kernel version isn't h= elpful is actually true."* **This is the key concern.** The kernel version in a device coredump is act= ually quite useful for debugging. When a user files a GPU hang bug report w= ith a coredump attached, knowing the exact kernel version is important for = reproducing the issue and correlating with known bugs. Other GPU drivers (a= mdgpu, msm) emit the same `kernel: UTS_RELEASE` line in their coredumps for= this reason. Removing it from xe while leaving it in amdgpu and msm create= s an inconsistency across DRM drivers. The build reproducibility concern is legitimate, but it's a development-tim= e convenience issue, not a correctness issue. A developer tracking binary c= hanges can filter or exclude `xe_devcoredump.o` from their comparison. The = debugging value of having the kernel version in production coredumps likely= outweighs the minor build-comparison annoyance. **Recommendation:** This patch should probably be **NAK'd** unless there's = a stronger justification. If the reproducibility concern is important enoug= h, a better approach might be to emit the version at runtime (e.g., using `= utsname()->release`) rather than at compile time, which would remove the bu= ild-reproducibility issue while preserving the debugging information. Alter= natively, if the broader community agrees that `UTS_RELEASE` should be remo= ved from coredumps, it should be done consistently across all DRM drivers, = not just xe. --- Generated by Claude Code Patch Reviewer