Re: [PATCH v2 11/50] drm/bridge: Add bridge driver for display connectors

Hi Boris,

On Thu, Aug 22, 2019 at 06:05:20PM +0200, Boris Brezillon wrote:

On Tue, 20 Aug 2019 04:16:42 +0300 Laurent Pinchart wrote:

...

• /*

• * Get the HPD GPIO for DVI and HDMI connectors. If the GPIO can provide
  

• * interrupts, register an interrupt handler.
  

• */
  

• if (type == DRM_MODE_CONNECTOR_DVII ||

•    type == DRM_MODE_CONNECTOR_HDMIA) {
  

• conn->hpd_gpio = devm_gpiod_get_optional(&pdev->dev, "hpd",
  

• 					 GPIOD_IN);
  

• if (IS_ERR(conn->hpd_gpio)) {
  

• 	if (PTR_ERR(conn->hpd_gpio) != -EPROBE_DEFER)
  

• 		dev_err(&pdev->dev,
  

• 			"Unable to retrieve HPD GPIO\n");
  

• 	return PTR_ERR(conn->hpd_gpio);
  

• }
  

• conn->hpd_irq = gpiod_to_irq(conn->hpd_gpio);
  

• } else {

• conn->hpd_irq = -EINVAL;
  

• }
• if (conn->hpd_irq >= 0) {

• ret = devm_request_threaded_irq(&pdev->dev, conn->hpd_irq,
  

• 				NULL, display_connector_hpd_irq,
  

• 				IRQF_TRIGGER_RISING |
  

• 				IRQF_TRIGGER_FALLING |
  

• 				IRQF_ONESHOT,
  

• 				"HPD", conn);
  

• if (ret) {
  

• 	dev_err(&pdev->dev,
  

• 		"Failed to request HPD interrupt\n");
  

• 	return ret;
  

Is there anything that mandates support of edge events on GPIO chips? I know it's quite common, but maybe we should fallback to polling mode when devm_request_threaded_irq() fails.

That's a good point, I'll change this.

...

• }
  

• }
• /* Retrieve the DDC I2C adapter for DVI, HDMI and VGA connectors. */
• if (type == DRM_MODE_CONNECTOR_DVII ||

•    type == DRM_MODE_CONNECTOR_HDMIA ||
  

•    type == DRM_MODE_CONNECTOR_VGA) {
  

• struct device_node *phandle;
  

• phandle = of_parse_phandle(pdev->dev.of_node, "ddc-i2c-bus", 0);
  

• if (phandle) {
  

• 	conn->bridge.ddc = of_get_i2c_adapter_by_node(phandle);
  

• 	of_node_put(phandle);
  

• 	if (!conn->bridge.ddc)
  

• 		return -EPROBE_DEFER;
  

• } else {
  

• 	dev_dbg(&pdev->dev,
  

• 		"No I2C bus specified, disabling EDID readout\n");
  

• }
  

• }