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/imagination: Synchronize interrupts before suspending the GPU Date: Wed, 27 May 2026 15:07:57 +1000 Message-ID: In-Reply-To: <20260526-sync-irqs-6-12-v2-1-b4306e95194d@imgtec.com> References: <20260526-sync-irqs-6-12-v2-1-b4306e95194d@imgtec.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/imagination: Synchronize interrupts before suspending the GPU Author: Alessio Belle Patches: 2 Reviewed: 2026-05-27T15:07:57.136689 --- This is a single-patch stable backport for kernel 6.12 that addresses a rea= l race condition in the `drm/imagination` (PowerVR) driver. During runtime = PM suspend, the GPU can be powered off while an IRQ handler is still execut= ing on another CPU core, causing a synchronous external abort when the hand= ler tries to access GPU registers. The upstream fix (commit `2d7f05cddf4c`) uses `disable_irq()`/`enable_irq()= ` pairs in `pvr_power_fw_disable`/`pvr_power_fw_enable`, which is the compl= ete solution. This backport correctly identifies that 6.12 lacks the `pvr_p= ower_fw_enable` counterpart infrastructure (and other code added in 6.16) a= nd instead uses the simpler `synchronize_irq()` =E2=80=94 a minimal, approp= riate fix for the stable branch. The patch is well-motivated with a clear crash signature, properly tagged w= ith `Fixes:` and `Cc: stable`, carries the upstream `Reviewed-by` and `Sign= ed-off-by` trailers, and is v2 addressing prior review feedback about missi= ng trailers. **Verdict: This patch looks correct and appropriate for 6.12 stable.** --- Generated by Claude Code Patch Reviewer