From mboxrd@z Thu Jan  1 00:00:00 1970
From: Claude Code Review Bot <claude-review@example.com>
To: dri-devel-reviews@example.com
Subject: Claude review: drm/panel: Add panel driver for ChipWealth CH13726A based panels
Date: Tue, 28 Apr 2026 15:10:46 +1000
Message-ID: <review-patch2-20260426-ch13726a-v7-2-554247c569e5@gmail.com>
In-Reply-To: <20260426-ch13726a-v7-2-554247c569e5@gmail.com>
References: <20260426-ch13726a-v7-0-554247c569e5@gmail.com> <20260426-ch13726a-v7-2-554247c569e5@gmail.com>
X-Mailer: Claude Code Patch Reviewer
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0

Patch Review

**Issue 1 (Kconfig): Unnecessary DP helper selections**

```c
select DRM_DISPLAY_DP_HELPER
select DRM_DISPLAY_HELPER
```

This is a pure MIPI-DSI panel. The only other panel drivers that `select DR=
M_DISPLAY_DP_HELPER` are eDP panels (Samsung ATNA33XC20, simple eDP). Compa=
rable MIPI-DSI panel drivers like `DRM_PANEL_BOE_TD4320` select neither of =
these. Both `select` lines should be removed =E2=80=94 the driver doesn't u=
se any display helper or DP helper APIs.

**Issue 2 (const-correctness): `desc` pointer and `thor_bottom_desc` should=
 be const**

The descriptor struct is declared as:

```c
struct ch13726a_desc *desc;
```

and `thor_bottom_desc` is:

```c
static struct ch13726a_desc thor_bottom_desc =3D {
```

Since `of_device_get_match_data()` returns `const void *`, the current cast=
 discards const:

```c
ctx->desc =3D (struct ch13726a_desc *)of_device_get_match_data(dev);
```

The field should be `const struct ch13726a_desc *desc`, the variable should=
 be `static const struct ch13726a_desc thor_bottom_desc`, and the cast shou=
ld be `(const struct ch13726a_desc *)` (or better, dropped entirely since C=
 allows implicit conversion from `const void *` to `const T *`).

**Issue 3 (style): `uint8_t` loop variable in `get_modes`**

```c
for (uint8_t i =3D 0; i < ctx->desc->num_modes; i++) {
```

Kernel convention is to use `unsigned int` or `int` for loop counters, not =
fixed-width types. `unsigned int` would match the type of `num_modes`.

**Issue 4 (style): `__typeof(*ctx)` in `devm_drm_panel_alloc`**

```c
ctx =3D devm_drm_panel_alloc(dev, __typeof(*ctx), panel,
                           &ch13726a_panel_funcs,
                           DRM_MODE_CONNECTOR_DSI);
```

While functionally correct, all other panel drivers pass the struct name di=
rectly:

```c
ctx =3D devm_drm_panel_alloc(dev, struct ch13726a_panel, panel,
                           &ch13726a_panel_funcs,
                           DRM_MODE_CONNECTOR_DSI);
```

**Observation (not a blocker): `of_drm_get_panel_orientation` error path**

```c
ret =3D of_drm_get_panel_orientation(dev->of_node, &ctx->orientation);
if (ret < 0) {
    dev_err(dev, "%pOF: failed to get orientation %d\n", dev->of_node, ret);
    return ret;
}
```

This could use `dev_err_probe()` for consistency with the other error paths=
 in `probe()`, which would also handle `-EPROBE_DEFER` cleanly:

```c
return dev_err_probe(dev, ret, "Failed to get orientation\n");
```

**Overall for patch 2:** The driver logic itself is solid =E2=80=94 the res=
et sequence, init commands, backlight handling, and multi-rate mode support=
 all look correct. The `mipi_dsi_multi_context` usage is clean, `prepare_pr=
ev_first =3D true` is properly set, and the remove path correctly detaches =
DSI before removing the panel. The issues above are all minor fixups.

---
Generated by Claude Code Patch Reviewer