[AMD Official Use Only - AMD Internal Distribution Only] Hi Mike,
Thanks for the logs. I've reverted the patch in amd-staging-drm-next and also sent the patch to the stable list. If its possible, please also collect the dmcub trace log (cat /sys/kernel/debug/dri/0/amdgpu_dm_dmub_tracebuffer) after the hang. Its usually dri/0 or dri/1 folder -- Regards, Jay ________________________________ From: Mikhail Gavrilov <[email protected]> Sent: Friday, May 30, 2025 3:59 PM To: Pillai, Aurabindo <[email protected]> Cc: Chung, ChiaHsuan (Tom) <[email protected]>; Wu, Ray <[email protected]>; Wheeler, Daniel <[email protected]>; Deucher, Alexander <[email protected]>; amd-gfx list <[email protected]>; dri-devel <[email protected]>; Linux List Kernel Mailing <[email protected]>; Linux regressions mailing list <[email protected]> Subject: Re: 6.15-rc6/regression/bisected - after commit f1c6be3999d2 error appeared: *ERROR* dc_dmub_srv_log_diagnostic_data: DMCUB error On Fri, May 30, 2025 at 6:48 PM Pillai, Aurabindo <[email protected]> wrote: > > [AMD Official Use Only - AMD Internal Distribution Only] > > > Hi Mike, > > We were trying to see if we can repro the issue on newer cards as well, but > it seems only 6000 series can repro at our end. > If you can repro more easily on other cards, please add "drm.debug=0x116 > log_buf_len=20M" to your kernel cmdline and grab the dmesg please. Hi Aurabindo, With drm.debug=0x116, I was able to capture the DMCUB error on the 7900XTX even during system boot. Here’s a snippet from the log: [ 140.307960] amdgpu 0000:03:00.0: [drm:drm_atomic_state_init] Allocated atomic state 000000003bcb4982 [ 140.307978] amdgpu 0000:03:00.0: [drm:drm_atomic_get_plane_state] Added [PLANE:77:plane-6] 00000000d20ccca3 state to 000000003bcb4982 [ 140.307985] amdgpu 0000:03:00.0: [drm:drm_atomic_get_crtc_state] Added [CRTC:80:crtc-0] 00000000267a47e8 state to 000000003bcb4982 [ 140.307992] amdgpu 0000:03:00.0: [drm:drm_atomic_set_fb_for_plane] Set [FB:132] for [PLANE:77:plane-6] state 00000000d20ccca3 [ 140.308214] amdgpu 0000:03:00.0: [drm:drm_mode_addfb2] [FB:134] [ 140.506110] amdgpu 0000:03:00.0: [drm:dc_dmub_srv_wait_for_idle [amdgpu]] No reply for DMUB command: status=3 [ 140.506572] amdgpu 0000:03:00.0: [drm] *ERROR* dc_dmub_srv_log_diagnostic_data: DMCUB error - collecting diagnostic data [ 140.506605] amdgpu 0000:03:00.0: [drm:dc_dmub_srv_log_diagnostic_data [amdgpu]] DMCUB STATE: [ 140.507065] amdgpu 0000:03:00.0: [drm:dc_dmub_srv_log_diagnostic_data [amdgpu]] dmcub_version : 07002d00 [ 140.507500] amdgpu 0000:03:00.0: [drm:dc_dmub_srv_log_diagnostic_data [amdgpu]] scratch [0] : 00000003 [ 140.507924] amdgpu 0000:03:00.0: [drm:dc_dmub_srv_log_diagnostic_data [amdgpu]] scratch [1] : 07002d00 [ 140.508341] amdgpu 0000:03:00.0: [drm:dc_dmub_srv_log_diagnostic_data [amdgpu]] scratch [2] : 00000000 [ 140.508591] amdgpu 0000:03:00.0: [drm:dc_dmub_srv_wait_for_idle [amdgpu]] No reply for DMUB command: status=3 [ 140.508944] amdgpu 0000:03:00.0: [drm] *ERROR* dc_dmub_srv_log_diagnostic_data: DMCUB error - collecting diagnostic data > > I'd also like to know if your issue is fully resolved if "drm/amd/display: > more liberal vmin/vmax update for freesync" is reverted. Yes, the issue was fully resolved by reverting commit f1c6be3999d2. I’ve attached the full kernel log below. -- Best Regards, Mike Gavrilov.
