ddcutil
ddcutil copied to clipboard
Second monitor in MST chain not supported
I've got 2 of my monitors setup using MST of which only the first in the chain is detected and works with ddcutil. Both monitors are Dell U2515H monitors that work when not using MST. If you need any more information, feel free to ask. I will try not to use MST, but I'll have to find myself a longer cable. The Acer monitor mentioned in the information is working correctly btw and is not having any problems atm.
Andreas,
Interesting report. I know there are problems with MST support in the i915 driver. This is the first I've seen for AMD. Some observations:
-
Both the radeon and amdgpu drivers are loaded. Is there a reason for this?
-
Looking at the filtered log files, it appears there's program failure at line 88 of amdgpu driver file amdgpu_dm_mst_types.c, when doing an i2c transfer. Can you send the full dmesg output for a failure, along with file Xorg.0.log?
Regards, Sanford
On 09/03/2018 08:05 AM, Andreas Backx wrote:
I've got 2 of my monitors setup using MST of which only the first in the chain is detected and works with ddcutil. Both monitors are Dell U2515H monitors that work when not using MST. If you need any more information, feel free to ask. I will try not to use MST, but I'll have to find myself a longer cable. The Acer monitor mentioned in the information is working correctly btw and is not having any problems atm.
Here's a gist of all the information I could get, (I am part of the group for i2c devices, but I thought I'd run the environment command as root so it might contain more information). https://gist.github.com/AndreasBackx/d7ab65f74c494aebf0edb4f32a7c24f3
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/rockowitz/ddcutil/issues/61, or mute the thread https://github.com/notifications/unsubscribe-auth/ANhsbjC1xdVbdjGIdYIevbeL_jSMMJtxks5uXRsegaJpZM4WXgBv.
1) Both the radeon and amdgpu drivers are loaded. Is there a reason for this?
Ah my bad, the wiki page on AMD is a bit confusing. I'll remove the radeon one.
2) Looking at the filtered log files, it appears there's program failure at line 88 of amdgpu driver file amdgpu_dm_mst_types.c, when doing an i2c transfer. Can you send the full dmesg output for a failure, along with file Xorg.0.log?
I'll report back to you when I get home tonight.
Also, please re-run "ddcutil env -v" without the radeon driver, on the off-chance that's triggering the problem. Thanks.
On 09/04/2018 06:03 AM, Andreas Backx wrote:
1. Both the radeon and amdgpu drivers are loaded. Is there a reason for this? Ah my bad, the wiki page on AMD is a bit confusing. I'll remove the radeon one. 2. Looking at the filtered log files, it appears there's program failure at line 88 of amdgpu driver file amdgpu_dm_mst_types.c, when doing an i2c transfer. Can you send the full dmesg output for a failure, along with file Xorg.0.log? I'll report back to you when I get home tonight.— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/rockowitz/ddcutil/issues/61#issuecomment-418311837, or mute the thread https://github.com/notifications/unsubscribe-auth/ANhsbr4IiS9rzjxKbKJNm8rhuuq2cXKHks5uXk_wgaJpZM4WXgBv.
Alright, I've removed radeon from the modules, and updated/added the files you asked for in the gist. It doesn't seem to have made a difference on the surface as I assume by placing amdgpu first in the modules list, it used that (which is what the wiki says).
The Xorg.0.log file was first removed after having already booted with only the amdgpu module, then I rebooted, and finally logged into i3. I hope that gives you some concise logs.
Any update on this?
As noted previously, there's a failure in the amdgpu driver near function dm_dp_aux_transfer() at line 88 of file amdgpu_dm_mst_types.c. This is a failure in the amdgpu driver. Not being able to replicate your hardware configuration I can't say more, beyond the general comment that I2C MST support in Linux video drivers is still problematic. I suggest you report this to the amdgpu driver folks.
Sanford
On 11/11/2018 08:17 PM, Andreas Backx wrote:
Any update on this?
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/rockowitz/ddcutil/issues/61#issuecomment-437724577, or mute the thread https://github.com/notifications/unsubscribe-auth/ANhsbr1fRHTz34UQvO9K6iIG4nu6C2Zwks5uuMw2gaJpZM4WXgBv.
I've been changing my setup a bit and I moved to Wayland. I realised that I completely forgot about this issue. I was ready to create a bug report on the issue tracker for amdgpu, though when I recreated the logs, I no longer saw the warning you were referring to. Here are the new logs. I'm now running Sway. Has anything else changed? Do you need some more information perhaps?
Andreas,
I apologize for taking so long to respond.
Even though the amdgpu warnings are no longer there, the issue is the lack of MST I2C support in the amdgpu driver, and should be directed to the amdgpu folks. There is no /dev/i2c device associated with the second Dell monitor associated with DP-5.
I suggest you install program i2cdetect (normally found in package i2c-tools). "ddcutil environment" uses this program as an additional check if it's available. What I expect it will confirm is that there's no /dev/i2c device for the second Dell monitor.
Sanford
On 4/14/19 3:58 PM, Andreas Backx wrote:
I've been changing my setup a bit and I moved to Wayland. I realised that I completely forgot about this issue. I was ready to create a bug report on the issue tracker for amdgpu, though when I recreated the logs, I no longer saw the warning you were referring to. Here are the new logs. https://gist.github.com/AndreasBackx/c8e295da821e4df9f3daa7a04768f84d I'm now running Sway https://github.com/swaywm/sway. Has anything else changed? Do you need some more information perhaps?
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/rockowitz/ddcutil/issues/61#issuecomment-483051613, or mute the thread https://github.com/notifications/unsubscribe-auth/ANhsbvfTdxPIbBndr8LG1JOWF4oFURpCks5vg4hcgaJpZM4WXgBv.