Claude review: video/fbdev/via: check ioremap return value in viafb_lcd_get_mobile_state

[Claude Code Review Bot <claude-review@example.com>]

Patch Review

**The bug is real.** At `lcd.c:956`, `ioremap()` is called and the result is used immediately at line 957 via `readw(biosptr)` without any NULL check. If `ioremap()` fails, this would cause a NULL pointer dereference.

**Issues:**

1. **Commit message says `-ENOMEM` but the code returns `false`.** The commit message states:
   > This patch adds a proper NULL check after ioremap() and returns -ENOMEM in case of failure.
   
   But the actual code returns `false`, which is consistent with the function's `bool` return type. The commit message is wrong and should be corrected to say it returns `false`.

2. **DEBUG_MSG format string has a leading space:**
   ```c
   DEBUG_MSG(KERN_ERR " Failed to remap BIOS memory\n");
   ```
   The space after `KERN_ERR` before `"Failed"` looks like a typo. It should be:
   ```c
   DEBUG_MSG(KERN_ERR "Failed to remap BIOS memory\n");
   ```

3. **Use of `DEBUG_MSG` with `KERN_ERR`:** Looking at existing error paths in this function, neither of the existing `return false` paths (line 988) logs anything. Using `DEBUG_MSG` (which is typically compiled out) with `KERN_ERR` is a bit contradictory — if this is truly an error worth reporting, `pr_err()` would be more appropriate and wouldn't be compiled away. That said, using the existing codebase conventions is also reasonable; the author should clarify intent.

4. **Minor: the `*mobile` output parameter is left uninitialized on the error path.** This matches the existing behavior when the BIOS signature doesn't match (the `else` branch at line 986-988 also returns `false` without setting `*mobile`), so this is a pre-existing issue, not something this patch introduces.

**The fix itself is correct in substance** — the NULL check is placed in the right location, the return value is appropriate for the function signature, and the existing `iounmap()` calls in later paths don't need adjustment since we never successfully mapped in this error case.

---
Generated by Claude Code Patch Reviewer