i.MX93 + IT6263 LVDS-to-HDMI: DRM cannot get modes (EDID not read) Hello NXP community, I’m bringing up an HDMI display on a custom board based on the i.MX93 (running Debian from the NXP BSP). The monitor works perfectly as an extended screen on Windows (hot-plug, correct resolution), but on the i.MX93 I see no picture. Symptoms: - Kernel log shows (the system still boots): [ 0.587871] /ldb-display-controller/lvds-channel@0: Fixed dependency cycle(s) with /soc@0/bus@44000000/i2c@44340000/lvds-to-hdmi-bridge@4c - The IT6263 appears on I2C-0 at address 0x4c as 'UU' (driver bound). - On I2C-1 (44350000.i2c) I can see a device at 0x50 – this should be the monitor’s EDID EEPROM via the DDC channel. - /sys/class/drm/*/status reports 'connected', but /sys/class/drm/*/modes is empty (no modes at all). - Consequently, the DRM subsystem never picks up a valid mode, and the display stays in power-save. - lsmod | grep ite shows nothing – the driver seems to be statically built into the kernel. What I’ve tried / checked: i2cdetect -y 1 repeatedly shows 0x50, so the DDC path is physically working. I also made sure the monitor’s EDID is fine (tested on a PC). My questions: What is the correct, up-to-date device-tree configuration for the IT6263 on i.MX93 so that the DRM subsystem can read the monitor’s EDID and get the mode list? Are there any specific driver patches or configuration steps required (e.g., for the IT6263 to pass through DDC correctly)? Could the 'Fixed dependency cycle(s)' warning be the root cause, and if so, how do I break the cycle properly? Any guidance, known-working DTS snippets, or references to relevant patches would be enormously appreciated. Thank you in advance for your time! Best regards, Wilbur FRDM-Training Graphics & Display Re: i.MX93 + IT6263 LVDS-to-HDMI: DRM cannot get modes (EDID not read) Hi @absolute-006
"Fixed dependency cycle(s)" is normal; you don't need to worry about it.Add some print statements to `it6263_read_edid` to see if it's working properly.
Best regards, Zhiming
查看全文