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/panthor: Store IRQ register base iomem pointer in panthor_irq Date: Mon, 13 Apr 2026 19:05:52 +1000 Message-ID: In-Reply-To: <20260412142951.2309135-5-karunika.choo@arm.com> References: <20260412142951.2309135-1-karunika.choo@arm.com> <20260412142951.2309135-5-karunika.choo@arm.com> X-Mailer: Claude Code Patch Reviewer Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Patch Review **Status: Good, with one minor observation.** Adds `void __iomem *iomem` to `struct panthor_irq` and introduces generic I= RQ register offset defines: ```c +#define INT_RAWSTAT 0x0 +#define INT_CLEAR 0x4 +#define INT_MASK 0x8 +#define INT_STAT 0xc ``` The `PANTHOR_IRQ_HANDLER` macro signature simplifies from `(__name, __reg_p= refix, __handler)` to `(__name, __handler)` =E2=80=94 the per-block `__reg_= prefix` (e.g., `GPU_INT_`, `JOB_INT_`, `MMU_INT_`) is no longer needed beca= use each IRQ instance now carries its own iomem base pointing to the right = register block. The IRQ request functions gain an `iomem` parameter: ```c -int panthor_request_gpu_irq(struct panthor_device *ptdev, ...); +int panthor_request_gpu_irq(struct panthor_device *ptdev, + void __iomem *iomem, ...); ``` **Observation:** The names `INT_RAWSTAT`, `INT_CLEAR`, `INT_MASK`, `INT_STA= T` are quite generic. They're defined in `panthor_device.h` which is only i= ncluded within the panthor driver, so namespace collision risk is low in pr= actice. But if this header ever gets transitively included from a broader c= ontext, these names could clash. A `PANTHOR_` prefix might be safer, though= this is a minor style point and not a blocking issue. --- --- Generated by Claude Code Patch Reviewer