Claude review: drm: bridge: ti-sn65dsi83: Remove NO_HFP and NO_HBP mode flags for LVDS stability

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

Patch Review

**Verdict: Good patch, should be accepted.**

The change removes `MIPI_DSI_MODE_VIDEO_NO_HFP` and `MIPI_DSI_MODE_VIDEO_NO_HBP` from the DSI mode flags:

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

This is well-justified. The `NO_HFP`/`NO_HBP` flags tell the DSI host it can skip sending horizontal front/back porch periods, but the driver programs these exact values into the bridge chip's registers at lines 672–677:

```c
regmap_write(ctx->regmap, REG_VID_CHA_HORIZONTAL_BACK_PORCH,
             (mode->htotal - mode->hsync_end) / dual_factor);
regmap_write(ctx->regmap, REG_VID_CHA_HORIZONTAL_FRONT_PORCH,
             (mode->hsync_start - mode->hdisplay) / dual_factor);
```

The bridge chip clearly needs these blanking periods to reconstruct valid LVDS timings. Suppressing them on the DSI link while expecting the bridge to use them is inconsistent. The datasheet reference strengthens the argument. Note that `MIPI_DSI_MODE_VIDEO_NO_HSA` is intentionally retained — this is reasonable since the bridge has its own hsync-width register (`REG_VID_CHA_HSYNC_PULSE_WIDTH`) and can reconstruct HSA from the programmed value.

No issues with the code change itself. Two independent Tested-by tags add confidence.

---
Generated by Claude Code Patch Reviewer