On Thu, Jul 25, 2019 at 11:05:24AM +0000, Oleksandr Suvorov wrote:

DDC and VGA channels are independent, and therefore we cannot decide whether the monitor is connected or not, depending on the information from the DDC.

So the monitor should always be considered connected. Thus there is no reason to use connector detect callback for this driver.

Fixes DRM error of dumb monitor detection like: ... DRM: head 'VGA-1' found, connector 32 is disconnected. ...

Cc: stable@vger.kernel.org Fixes: 56fe8b6f4991 ("drm/bridge: Add RGB to VGA bridge support") Signed-off-by: Oleksandr Suvorov oleksandr.suvorov@toradex.com

Uh nope :-)

Yes VGA monitors are broken, but the way to fix that is to either override that on the kernel cmdline, or in your userspace somewhere. Not hardcode this in the kernel for everyone. Because not everyone does have a broken VGA monitor, but if you do this _every_ desktop will try to light up that broken monitor. Which leads to lots of bug reports and regressions.

This case is exactly what connector_status_unknown is meant for: The kernel couldn't authoritatively figure out whether there is a monitor or not. Userspace should/can try this into account for autoconfiguration.

Note a more proper fix would be to somehow wire up load detection. That will work even for dumb VGA monitors, and e.g. i915 then gives you an authoritative connector_status_disconnected if it could execute a load detect cycle (not always possible on some hw) and there's no screen detected with that.

Maybe we should document this better somewhere in docs? Would be great if you can type a patch for this ... -Daniel

---

drivers/gpu/drm/bridge/dumb-vga-dac.c | 18 ------------------ 1 file changed, 18 deletions(-)

diff --git a/drivers/gpu/drm/bridge/dumb-vga-dac.c b/drivers/gpu/drm/bridge/dumb-vga-dac.c index d32885b906ae..e37c19356d12 100644 --- a/drivers/gpu/drm/bridge/dumb-vga-dac.c +++ b/drivers/gpu/drm/bridge/dumb-vga-dac.c @@ -73,25 +73,7 @@ static const struct drm_connector_helper_funcs dumb_vga_con_helper_funcs = { .get_modes = dumb_vga_get_modes, }; -static enum drm_connector_status -dumb_vga_connector_detect(struct drm_connector *connector, bool force) -{

• struct dumb_vga *vga = drm_connector_to_dumb_vga(connector);
• /*

• * Even if we have an I2C bus, we can't assume that the cable
  

• * is disconnected if drm_probe_ddc fails. Some cables don't
  

• * wire the DDC pins, or the I2C bus might not be working at
  

• * all.
  

• */
  

• if (!IS_ERR(vga->ddc) && drm_probe_ddc(vga->ddc))

• return connector_status_connected;
  

• return connector_status_unknown;

-}

static const struct drm_connector_funcs dumb_vga_con_funcs = {

• .detect = dumb_vga_connector_detect, .fill_modes = drm_helper_probe_single_connector_modes, .destroy = drm_connector_cleanup, .reset = drm_atomic_helper_connector_reset,

-- 2.20.1

On Thu, Jul 25, 2019 at 5:41 PM maxime.ripard@free-electrons.com maxime.ripard@free-electrons.com wrote:

On Thu, Jul 25, 2019 at 11:05:23AM +0000, Oleksandr Suvorov wrote:

...

Even in source code of this driver there is an author's description: /* * Even if we have an I2C bus, we can't assume that the cable * is disconnected if drm_probe_ddc fails. Some cables don't * wire the DDC pins, or the I2C bus might not be working at * all. */

That's true. DDC and VGA channels are independent, and therefore we cannot decide whether the monitor is connected or not, depending on the information from the DDC.

So the monitor should always be considered connected.

Well, no. Like you said, we cannot decided whether is connected or not.

Maxime, thanks, I agree that's a bad solution. But I still think we should be able to define the DT node of a device for this driver to claim the connector is always connected. Please see my following thoughts.

...

Thus there is no reason to use connector detect callback for this driver: DRM sub-system considers monitor always connected if there is no detect() callback registered with drm_connector_init().

How to reproduce the bug:

• setup: i.MX8QXP, LCDIF video module + gpu/drm/mxsfb driver, adv712x VGA DAC + dumb-vga-dac driver, VGA-connector w/o DDC;
• try to use drivers chain mxsfb-drm + dumb-vga-dac;

• any DRM applications consider the monitor is not connected:

  $ weston-start $ cat /var/log/weston.log ... DRM: head 'VGA-1' found, connector 32 is disconnected. ... $ cat /sys/devices/platform/5a180000.lcdif/drm/card0/card0-VGA-1/status unknown

And that's exactly what's being reported here: we cannot decide if it is connected or not, so it's unknown.

If weston chooses to ignore connectors that are in an unknown state, I'd say it's weston's problem, since it's much broader than this particular device.

If we look at the code of drm_probe_helper.c, we can see, the drm_helper_probe_detect_ctx() assume the cable is connected if there is no detect() callback registered. ... if (funcs->detect_ctx) ret = funcs->detect_ctx(connector, &ctx, force); else if (connector->funcs->detect) ret = connector->funcs->detect(connector, force); else ret = connector_status_connected; ...

The driver dumb-vga-dac supports both DT configurations: - with DDC channel, that allows us to detect if the cable is connected; - without DDC channel. In this case, IMHO, the driver should behave the same way as a connector driver without registered detect() callback.

So what about the patch like?

@@ -81,6 +81,13 @@ dumb_vga_connector_detect(struct drm_connector *connector, bool force) { struct dumb_vga *vga = drm_connector_to_dumb_vga(connector);

+ /* + * If I2C bus for DDC is not defined, asume that the cable + * is always connected. + */ + if (PTR_ERR(vga->ddc) == -ENODEV) + return connector_status_connected; + /* * Even if we have an I2C bus, we can't assume that the cable * is disconnected if drm_probe_ddc fails. Some cables don't