From mboxrd@z Thu Jan  1 00:00:00 1970
From: Claude Code Review Bot <claude-review@example.com>
To: dri-devel-reviews@example.com
Subject: Claude review: drm/msm: Add basic perfcntr infrastructure
Date: Tue, 05 May 2026 08:06:03 +1000
Message-ID: <review-patch10-20260504190751.61052-11-robin.clark@oss.qualcomm.com>
In-Reply-To: <20260504190751.61052-11-robin.clark@oss.qualcomm.com>
References: <20260504190751.61052-1-robin.clark@oss.qualcomm.com> <20260504190751.61052-11-robin.clark@oss.qualcomm.com>
X-Mailer: Claude Code Patch Reviewer
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit
MIME-Version: 1.0

Patch Review

This adds `msm_perfcntr.c` with init/cleanup/resume/suspend stubs and group_idx/counter_base helpers.

**Issue: `msm_gpu_sysprof_no_perfcntr_zap()` NULL dereference.**

The inline function is changed to:

```c
return (refcount_read(&gpu->sysprof_active) > 1) ||
	READ_ONCE(gpu->perfcntrs->stream);
```

But `gpu->perfcntrs` is only set when `gpu->num_perfcntr_groups > 0` (which is a6xx+ only). For older GPUs (a2xx-a5xx), `gpu->perfcntrs` will be NULL, causing a NULL pointer dereference if this function is ever called on those GPUs. This should be guarded:

```c
return (refcount_read(&gpu->sysprof_active) > 1) ||
	(gpu->perfcntrs && READ_ONCE(gpu->perfcntrs->stream));
```

**Minor: `devm_kzalloc`/`devm_kfree` usage.** Using `devm_*` allocations then manually freeing in `msm_perfcntr_cleanup()` is a bit unusual. The `devm_kfree` calls in cleanup are necessary since the device outlives the cleanup, but mixing devm with manual lifetime management can be confusing. Not a bug, just worth noting.

**Minor: `msm_perfcntr_resume()` doesn't check `gpu->perfcntrs` for NULL** before accessing `gpu->perfcntrs->stream`, same concern as above.

---
Generated by Claude Code Patch Reviewer