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/amd/pm: Use guard(mutex) instead of manual lock+unlock
Date: Sat, 16 May 2026 12:18:26 +1000
Message-ID: <review-patch1-20260508024744.6523-1-andrejhirata@usp.br>
In-Reply-To: <20260508024744.6523-1-andrejhirata@usp.br>
References: <20260508024744.6523-1-andrejhirata@usp.br> <20260508024744.6523-1-andrejhirata@usp.br>
X-Mailer: Claude Code Patch Reviewer
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit
MIME-Version: 1.0

Patch Review

**Bug 1 (critical): Logic inversion in `amdgpu_dpm_is_link_reset_supported`**

The original code calls `smu_link_reset_is_support()` only when `is_support_sw_smu(adev)` is **true**:
```c
if (is_support_sw_smu(adev)) {
    mutex_lock(&adev->pm.mutex);
    support_link_reset = smu_link_reset_is_support(smu);
    mutex_unlock(&adev->pm.mutex);
}
```

The v3 patch inverts the condition:
```c
if (is_support_sw_smu(adev))
    return false;

guard(mutex)(&adev->pm.mutex);
return smu_link_reset_is_support(smu);
```

This returns `false` when sw_smu IS supported (should query the function), and calls `smu_link_reset_is_support` when it's NOT supported (should return `false`). The condition should be `if (!is_support_sw_smu(adev))`. Compare with the correctly negated `amdgpu_dpm_mode1_reset` and `amdgpu_dpm_link_reset` conversions in the same patch.

**Bug 2 (build-breaking): Missing semicolon in `amdgpu_dpm_is_mode1_reset_supported`**

```c
guard(mutex)(&adev->pm.mutex);
return smu_mode1_reset_is_support(smu)
}
```

Missing semicolon after the `return` statement.

**Bug 3 (build-breaking): `return =` syntax errors**

Four instances of `return =` instead of `return`:

1. `amdgpu_dpm_smu_i2c_bus_access`:
   ```c
   return = pp_funcs->smu_i2c_bus_access(pp_handle, acquire);
   ```

2. `amdgpu_dpm_get_pp_table`:
   ```c
   return = pp_funcs->get_pp_table(adev->powerplay.pp_handle, table);
   ```

3. `amdgpu_dpm_set_fine_grain_clk_vol`:
   ```c
   return = pp_funcs->set_fine_grain_clk_vol(adev->powerplay.pp_handle, ...);
   ```

4. `amdgpu_dpm_get_display_mode_validation_clocks`:
   ```c
   return = pp_funcs->get_display_mode_validation_clocks(adev->powerplay.pp_handle, clocks);
   ```

All four should be just `return` (no `=`).

**Nit: Superfluous `return;` at end of `amdgpu_dpm_get_current_power_state`**

```c
*state = adev->pm.dpm.user_state;

-out:
-	mutex_unlock(&adev->pm.mutex);
+	return;
 }
```

The trailing `return;` at the end of a void function is unnecessary. The old `goto out` -> `return;` conversion is correct, but the one at the very end of the function body should be dropped.

**Observation: The RESEND/v1 patch did not have any of these bugs** -- it was a simpler mechanical conversion that kept `ret` variables and only swapped lock/unlock for `guard()`. The bugs were introduced in v2/v3 during the more aggressive refactoring to use direct `return f()` and early-return patterns. It would be worth running a `make` build test before submitting v4.

---
Generated by Claude Code Patch Reviewer