Claude review: drm/msm/dp: add missing drm_edid_connector_update() before add_modes on cached EDID

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

Patch Review

**Correctness: Good.**

The change adds the required `drm_edid_connector_update()` call in the cached-EDID fast path of `msm_dp_panel_get_modes()`:

```c
if (msm_dp_panel->drm_edid) {
    drm_edid_connector_update(connector, msm_dp_panel->drm_edid);
    return drm_edid_connector_add_modes(connector);
}
```

This mirrors the pattern already used in `msm_dp_panel_read_sink_caps()` (line 281) where `drm_edid_connector_update()` is called right after `drm_edid_read_ddc()`. The documentation in `drm_edid.c` is explicit that `drm_edid_connector_update()` "must be called before calling `drm_edid_connector_add_modes()`", so this is the correct fix.

**Calling `drm_edid_connector_update()` redundantly is safe.** It calls `update_display_info()`, `_drm_update_tile_info()`, and `_drm_edid_connector_property_update()` — all of which are idempotent for the same EDID data. If `get_modes` is called more than once during a single connection lifecycle, the second call to `drm_edid_connector_update()` simply re-sets the same values. No concern here.

**Return value of `drm_edid_connector_update()` is intentionally ignored.** The function returns an error code from property updates, but the caller is `.get_modes` which returns mode count. Other drivers (e.g., `drm_bridge_connector.c`) also ignore this return value in the `get_modes` path, so this is consistent with existing patterns.

**Fixes tag is correct.** Commit 5bea90ad9743 is the refactor that introduced the `struct drm_edid` API usage and dropped the previously-implicit association.

**Scope is appropriate.** V1 had retry/re-read logic; v2 correctly identified the root cause and dropped the workaround. The fix is minimal — one added line plus the braces to wrap the block.

**No issues found.**

---
Generated by Claude Code Patch Reviewer