Created attachment 285089 [details] Xorg.log that talks about removed devices since about kernel-5.2.8 my box does not thaw properly. first it was once per 10 thaws... but now it became quite bad (like every second thaw)... it seems like X thinks, that all devices are gone shortly after the thaw (see attached Xorg.log). furthermore I observed that the screen becomes black before writing of the suspend image starts (normally i can read text about the suspend proceeding until the power is turned off). i use s2disk for suspend...
Created attachment 285143 [details] another Xorg.log this time the screen looked normal at the first glimpse. but: then i saw it was frozen and the mouse and keyboard were not working...
in the journalctl output i found this suspicious looking messages, that were recorded during todays thaw: kernel: [drm] PCIE GART of 256M enabled (table at 0x000000F400000000). kernel: amdgpu 0000:01:00.0: [drm:amdgpu_ring_test_helper [amdgpu]] *ERROR* ring sdma0 test failed (-110) kernel: [drm:amdgpu_device_ip_resume_phase2 [amdgpu]] *ERROR* resume of IP block <sdma_v3_0> failed -110 kernel: [drm:amdgpu_device_resume [amdgpu]] *ERROR* amdgpu_device_ip_resume failed (-110). kernel: PM: dpm_run_callback(): pci_pm_restore+0x0/0x90 returns -110 kernel: PM: Device 0000:01:00.0 failed to restore async: error -110
yesterday i did that mesa-19.2.0-2 upgrade (without restarting X)... today the screen stayed black again and the kernel says the same as before... so the new linux-firmware (20190923.417a9c6-1) does not help... the kernel is still 5.3.1.arch1-1... the GPU is a RX550 (Device: Radeon 550 Series (POLARIS12, DRM 3.33.0, 5.3.1-arch1-1-ARCH, LLVM 8.0.1) (0x699f)).
can it be that these crashes are caused by certain changes to /boot/initramfs-linux.img ? if yes: rebooting after such changes could be a workaround... -arne
using 5.3.7-arch1-1-ARCH it also happens without any change of initramfs/firmware/kernel... somehow the suspend proceeding does not work so good anymore... -arne
somehow nobody finds this report... the Assignee seems to be wrong... so i filed a new one... -arne *** This bug has been marked as a duplicate of bug 205335 ***