"Invalid display" talking to Philips 436M6VBP over HDMI

[l0b0]

Not sure what to do about this error:

❯ ddcutil detect
Invalid display
   I2C bus:  /dev/i2c-7
   DRM connector:           card0-HDMI-A-1
   EDID synopsis:
      Mfg id:               PHL - Philips Consumer Electronics Company
      Model:                PHL 436M6VBP
      Product code:         49529  (0xc179)
      Serial number:        
      Binary serial number: 1335 (0x00000537)
      Manufacture year:     2019,  Week: 9
   DDC communication failed. (getvcp of feature x10 returned Error_Info[EIO in ddc_write_read_with_retry, causes: EIO])

After looking around at similar issues I've tried both --disable-dynamic-sleep --sleep-multiplier 3 and --use-file-io, but to no effect.

interrogate.log

[rnhmjoj]

I'm not sure, given it has identified the display I'd say i2c is set up correctly.
There is a number of similar issues in the upstream tracker, so it's probably unrelated.

[rockowitz]

I²C communication is working on /dev/i2c-7. The EDID (slave address x50) can be read. Error EIO is an indication that slave address x37 (DDC) is unresponsive.

If you install i2cdetect what I expect it will show is that, for i2c bus 7, slave address x50 is responsive, and slave address x37 is not. So no amount of fiddling with ddcutil settings will make things work.

You have 2 video devices (card0 and card1) enabled. Apparently one is on the Ryzen CPU. and the other is a discrete card. Do you need both? That may be adding complexity that confuses the amdgpu driver.

Have you tested whether I²C communication works on X11 instead of Wayland?

Beyond that, I would guess that the problem is in some way related to how NixOS is configured, but I have no experience with NixOS.

[l0b0]

There is a number of similar issues in the upstream tracker, so it's probably unrelated.

Which upstream tracker do you mean? i2c, the kernel, the display manufacturer, or something else?

[l0b0]

I2C communication is working on /dev/i2c-7. The EDID (slave address x50) can be read. Error EIO is an indication that slave address x37 (DDC) is unresponsive.

If you install i2cdetect what I expect it will show is that, for i2c bus 7, slave address x50 is responsive, and slave address x37 is not. So no amount of fiddling with ddcutil settings will make things work.

Oh, OK, that clarifies things.

You have 2 video devices (card0 and card1) enabled. Apparently one is on the Ryzen CPU. and the other is a discrete card. Do you need both? That may be adding complexity that confuses the amdgpu driver.

Disabling the onboard GPU didn't help.

Have you tested whether I2C communication works on X11 instead of Wayland?

Just did, and the result was the same on X11 as on Wayland.

Beyond that, I would guess that the problem is in some way related to how NixOS is configured, but I have no experience with NixOS.

I'd expect it to be something wrong with the display itself, but I don't know enough about this to debug any of it.