Claude review: drm/panel-edp: Change BOE NV140WUM-N64 timings

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

Patch Review

**Commit message:**
The commit message says:

> This screen timing requires a backlight off time of more than 100ms
> from the end of the data stream to avoid screen flickering and red
> light caused by screen material limitations.

This is a bit vague. It would be helpful to clarify:
- Whether this is a `.disable` delay (T9 timing per the eDP spec), which it is based on the code. The commit message says "backlight off time" but the `.disable` field is documented in the struct as *"Time for panel to turn the display off (panel-off to backlight-off, T9-min in eDP timing diagrams)"*. The terminology should be more precise — this is a T9 minimum delay, not a backlight-off time per se.
- How the issue was observed (which platform/board), which would help maintainers evaluate the change.

**Code changes — new delay struct:**

```c
+static const struct panel_delay delay_200_500_e200_d100 = {
+	.hpd_absent = 200,
+	.unprepare = 500,
+	.enable = 200,
+	.disable = 100,
+};
```

This correctly follows the established naming convention (`delay_[hpd_absent]_[unprepare]_e[enable]_d[disable]`), and the field values are correct. There are already `_d10`, `_d50`, and `_d200` variants, so `_d100` fits in naturally.

**Minor issue — placement:** The new struct is inserted at line 1938 (after `delay_200_500_e200`), directly before `delay_200_500_e200_d200`. Looking at the existing code in the drm-next tree, the `_d` variants are ordered as `_d200` (line 1791), `_d10` (line 1798), `_d50` (line 1805). The patch places `_d100` between `delay_200_500_e200` and `delay_200_500_e200_d200`. Since the existing `_d` variants aren't numerically ordered either, this placement is acceptable, though ideally all the `_d` variants would be grouped together.

**Code change — panel entry update:**

```c
-	EDP_PANEL_ENTRY('B', 'O', 'E', 0x0cf6, &delay_200_500_e200, "NV140WUM-N64"),
+	EDP_PANEL_ENTRY('B', 'O', 'E', 0x0cf6, &delay_200_500_e200_d100, "NV140WUM-N64"),
```

This is straightforward and correct — swapping the delay struct pointer for the NV140WUM-N64 panel entry.

**Functional assessment:** The `.disable` field is used in `panel_edp_disable()` where it triggers an `msleep()` call. Previously this panel had `.disable` implicitly zero (no delay). Adding a 100ms delay is a conservative, low-risk change. The 100ms value seems reasonable for a T9-min requirement, and similar delays exist for other panels (10ms, 50ms, 200ms).

**Nits:**
- The patch is missing a `Cc: stable` tag if this is fixing a visible user-facing issue (flickering). If this has always been broken, it may warrant backporting.
- It would be good to have a `Tested-by:` tag from someone who verified the fix on actual hardware.

**Verdict:** The patch looks functionally correct and safe. The commit message could be improved with more precise eDP timing terminology and platform context, but the code change itself is clean and follows existing conventions.

---
Generated by Claude Code Patch Reviewer