Created attachment 145903 [details] Kernel logs with resume errors In the last two weeks the gpu has failed to resume 4 out of 26 suspend/resume cycles. Prior to that, the 5.0 kernel (Ubuntu 9.04) did not have this problem. Here are excerpts from the resume failures. I am attaching the full kernel logs from the four crashes. The first group (gfx_v8_0) has happened three times, the second group (sdma_v3_0) has happened one time. -- amdgpu 0000:07:00.0: [drm:amdgpu_ring_test_helper [amdgpu]] *ERROR* ring gfx test failed (-110) [drm:amdgpu_device_ip_resume_phase2 [amdgpu]] *ERROR* resume of IP block <gfx_v8_0> failed -110 [drm:amdgpu_device_resume [amdgpu]] *ERROR* amdgpu_device_ip_resume failed (-110). PM: dpm_run_callback(): pci_pm_resume+0x0/0xa0 returns -110 PM: Device 0000:07:00.0 failed to resume async: error -110 -- amdgpu 0000:07:00.0: [drm:amdgpu_ring_test_helper [amdgpu]] *ERROR* ring sdma0 test failed (-110) [drm:amdgpu_device_ip_resume_phase2 [amdgpu]] *ERROR* resume of IP block <sdma_v3_0> failed -110 [drm:amdgpu_device_resume [amdgpu]] *ERROR* amdgpu_device_ip_resume failed (-110). PM: dpm_run_callback(): pci_pm_resume+0x0/0xa0 returns -110 PM: Device 0000:07:00.0 failed to resume async: error -110
Can you try kernel 5.4? There have a number of suspend and resume fixes. They should filter back to stable kernels as well.
With the 5.4-rc6 kernel I was able to run a suspend/resume cycle 50 times without anything breaking.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.