Claude review: drm: bridge: ti-sn65dsi83: Disable video burst mode for LVDS stability

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

Patch Review

**Verdict: Acceptable but with a reservation about regression risk.**

The change removes `MIPI_DSI_MODE_VIDEO_BURST`:

```c
-	dsi->mode_flags = MIPI_DSI_MODE_VIDEO | MIPI_DSI_MODE_VIDEO_BURST |
-			  MIPI_DSI_MODE_VIDEO_NO_HSA | MIPI_DSI_MODE_NO_EOT_PACKET;
+	dsi->mode_flags = MIPI_DSI_MODE_VIDEO | MIPI_DSI_MODE_VIDEO_NO_HSA |
+			  MIPI_DSI_MODE_NO_EOT_PACKET;
```

Without `MIPI_DSI_MODE_VIDEO_BURST`, the DSI link will operate in non-burst sync-event mode. The justification — that burst mode depends on continuous clocking behavior from the DSI host and may be unreliable — is plausible but somewhat generic.

**Points to consider:**

1. **Regression risk:** Unlike patch 1 (where the logic was clearly contradictory), burst mode is a valid DSI mode and this driver has shipped with it since its introduction. Removing it could cause issues on platforms where burst mode is currently working and non-burst mode is not — different DSI host controllers may behave differently. The Tested-by tags are from only two platforms.

2. **The commit message is honest** about the nature of the problem ("depending on the host controller implementation"), which is good. However, it would benefit from naming the specific DSI host controller or SoC where the instability was observed to help future readers understand the scope.

3. **No `Fixes:` tag** is present on either patch. If this is considered a bug fix (which the title "for LVDS stability" implies), a Fixes tag pointing to the original commit that added burst mode would be appropriate for backport tracking.

4. **`MIPI_DSI_MODE_VIDEO_NO_HSA` and `MIPI_DSI_MODE_NO_EOT_PACKET` are retained.** This is fine — `NO_HSA` is a reasonable optimization when the bridge programs hsync width via register, and `NO_EOT_PACKET` is commonly used with TI bridges.

The code change itself is trivially correct — it's just flag removal and reformatting the line continuation. No functional issues with the implementation.

---
Generated by Claude Code Patch Reviewer