Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider: - [Medium] drm/ch1115: out-of-range display offset 64 sent in ch1115_oled_init() - [Medium] drm/ch1115: missing DC-DC disable in ch1115_encoder_atomic_disable() --
commit e5268b4f2a7c534421925bdee0145caf63c50632 Author: Nicolás Antinori <[email protected]> drm/ch1115: add support for Chipwealth CH1115 OLED controller This commit adds a DRM driver for Chipwealth CH1115 monochrome dot matrix OLED panels connected over I2C. > diff --git a/drivers/gpu/drm/tiny/ch1115-i2c.c > b/drivers/gpu/drm/tiny/ch1115-i2c.c > new file mode 100644 > index 0000000000000..08fee5a09f4f9 > --- /dev/null > +++ b/drivers/gpu/drm/tiny/ch1115-i2c.c > @@ -0,0 +1,859 @@ [ ... ] > +static int ch1115_oled_init(struct ch1115_device *ch1115) > +{ > + u8 commands[] = { > + CH1115_CMD_OFF, > + > + CH1115_CMD_CLOCK_DIVIDE_RATIO, > + CH1115_VAL_CLOCK_DIVIDE_RATIO(CH1115_VAL_CLOCK_OSC_FREQ, > + CH1115_VAL_CLOCK_DIV_RATIO), > + > + CH1115_CMD_MUX_RATIO, > + ch1115->height - 1, > + > + CH1115_CMD_OFFSET, > + ch1115->flip_vertical > + ? ch1115->y_start_offset > + : CH1115_MAX_HEIGHT - ch1115->y_start_offset, [Severity: Medium] Does ch1115_oled_init() send an out-of-range value for CH1115_CMD_OFFSET when flip_vertical is false and y_start_offset is 0? In ch1115_parse_dt(), y_start_offset defaults to 0 when the chipwealth,offset property is omitted, and is otherwise restricted to [0, CH1115_MAX_HEIGHT - 1] (0 to 63): ret = of_property_read_u32(np, "chipwealth,offset", &val); if (ret) { ch1115->y_start_offset = 0; } else { if (val > CH1115_MAX_HEIGHT - 1) { dev_err(dev, "chipwealth,offset=%d must be [0-%d]\n", val, CH1115_MAX_HEIGHT - 1); return -EINVAL; } ch1115->y_start_offset = (u8)val; } When ch1115_encoder_atomic_enable() calls ch1115_oled_init() with flip_vertical set to false and y_start_offset set to 0, the expression CH1115_MAX_HEIGHT - ch1115->y_start_offset evaluates to 64 (0x40), which overflows the 6-bit [0, 63] vertical offset field instead of wrapping modulo CH1115_MAX_HEIGHT to 0. [ ... ] > +static void ch1115_encoder_atomic_enable(struct drm_encoder *encoder, > + struct drm_atomic_commit *state) > +{ > + struct drm_device *drm = encoder->dev; > + struct ch1115_device *ch1115 = drm_to_ch1115(drm); > + int ret; > + > + ret = ch1115_oled_init(ch1115); > + if (ret < 0) > + return; > + > + u8 command = CH1115_CMD_ON; > + > + ret = ch1115_send_command_list(ch1115, &command, 1); > + if (ret < 0) > + return; > + > + backlight_enable(ch1115->bl_dev); > +} > + > +static void ch1115_encoder_atomic_disable(struct drm_encoder *encoder, > + struct drm_atomic_commit *state) > +{ > + struct drm_device *drm = encoder->dev; > + struct ch1115_device *ch1115 = drm_to_ch1115(drm); > + > + backlight_disable(ch1115->bl_dev); > + > + u8 command = CH1115_CMD_OFF; > + > + ch1115_send_command_list(ch1115, &command, 1); > +} [Severity: Medium] Should ch1115_encoder_atomic_disable() also disable the internal DC-DC converter? When the encoder is enabled, ch1115_encoder_atomic_enable() calls ch1115_oled_init(), which turns on the DC-DC converter with CH1115_CMD_DC_DC and CH1115_VAL_DC_DC_ON. However, when the display is disabled via KMS or during system shutdown via ch1115_shutdown() -> drm_atomic_helper_shutdown(), ch1115_encoder_atomic_disable() only sends CH1115_CMD_OFF, leaving the DC-DC converter enabled. By contrast, ch1115_remove() explicitly sends CH1115_CMD_OFF, CH1115_CMD_DC_DC, and CH1115_VAL_DC_DC_OFF after calling drm_atomic_helper_shutdown(): u8 cmd_list[] = { CH1115_CMD_OFF, CH1115_CMD_DC_DC, CH1115_VAL_DC_DC_OFF, }; drm_dev_unplug(&ch1115->dev); drm_atomic_helper_shutdown(&ch1115->dev); ch1115_send_command_list(ch1115, cmd_list, ARRAY_SIZE(cmd_list)); If ch1115_encoder_atomic_disable() sent CH1115_CMD_OFF, CH1115_CMD_DC_DC, and CH1115_VAL_DC_DC_OFF, wouldn't that properly power down the DC-DC converter on both KMS disable and ch1115_shutdown(), while avoiding the duplicate command sequence in ch1115_remove()? -- Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=2
