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: perf/core: out-of-line and export perf_allow_cpu/tracepoint()
Date: Mon, 25 May 2026 18:04:51 +1000
Message-ID: <review-patch1-20260523013326.129491-2-jhubbard@nvidia.com>
In-Reply-To: <20260523013326.129491-2-jhubbard@nvidia.com>
References: <20260523013326.129491-1-jhubbard@nvidia.com> <20260523013326.129491-2-jhubbard@nvidia.com>
X-Mailer: Claude Code Patch Reviewer
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit
MIME-Version: 1.0

Patch Review

**Good:**

The out-of-line implementations are exact copies of the previous static inline bodies, preserving the distinct threshold levels and error codes:

```c
int perf_allow_cpu(void)
{
	if (sysctl_perf_event_paranoid > 0 && !perfmon_capable())
		return -EACCES;
	return security_perf_event_open(PERF_SECURITY_CPU);
}

int perf_allow_tracepoint(void)
{
	if (sysctl_perf_event_paranoid > -1 && !perfmon_capable())
		return -EPERM;
	return security_perf_event_open(PERF_SECURITY_TRACEPOINT);
}
```

The graduated thresholds (`> 0` for CPU, `> -1` for tracepoint, `> 1` for kernel in the existing code) and distinct error codes (`-EACCES` vs `-EPERM`) are all faithfully preserved.

**Good: !CONFIG_PERF_EVENTS stubs are complete.** The patch adds stubs for all three helpers:

```c
static inline int perf_allow_kernel(void)
{
	return perfmon_capable() ? 0 : -EACCES;
}
static inline int perf_allow_cpu(void)
{
	return perfmon_capable() ? 0 : -EACCES;
}
static inline int perf_allow_tracepoint(void)
{
	return perfmon_capable() ? 0 : -EPERM;
}
```

Note that `perf_allow_kernel()` didn't have a `!CONFIG_PERF_EVENTS` stub before this patch either -- the existing code declared it `extern` in the `CONFIG_PERF_EVENTS` block with no fallback. Adding it here is a bonus fix that makes the trio complete. The cover letter v2 changelog correctly attributes this to AI review feedback catching the build failure with `PERF_EVENTS=n`.

**Minor observation:** The stubs use `perfmon_capable()` as the fallback, which is a reasonable conservative default -- it requires `CAP_PERFMON` or `CAP_SYS_ADMIN`. The LSM hook is naturally unavailable without `CONFIG_PERF_EVENTS`, so this is the best approximation.

**Placement:** The new functions are placed in `kernel/events/core.c` right after the existing `perf_allow_kernel()` / `EXPORT_SYMBOL_GPL`, which is the natural location. No issues.

No problems with this patch.

---

---
Generated by Claude Code Patch Reviewer