So with my and Alex's patches you still observe corruption ? What if you
remove my patch and keep Alex's patch ?
On 04/10/2018 06:53 AM, Huang Rui wrote:
On Mon, Apr 09, 2018 at 11:17:58AM -0400, Andrey Grodzovsky wrote:
OK, tested with DC disabled , no issues on resume (no visible
corruption on display or errors in log). Now the display itself
freezes after amdgpu is loaded with DC disabled, this happens only
when BIOS in VGA mode , in console mode no such problem. Happens
before my and Alex patches, looks like a separate issue.
So anyway, if corruption would be there (beginning of VRAM and hence
scanout FB corrupted) , i should have seen it with grub in console
mode where display is fine and not freezing.
1. sudo modprobe amdgpu dc=0 ip_block_mask=0x7f
2. pm-suspend/resume two times.
You will see the start of vram is corrupted after S3 resume.
[ 570.343635] [drm] PCIE GART of 512M enabled (table at 0x000000F400000000).
[ 570.343642] [drm] PSP is resuming...
[ 570.343713] gmc_v9_0_process_interrupt: 12 callbacks suppressed
[ 570.343715] amdgpu 0000:03:00.0: [mmhub] VMC page fault (src_id:0 ring:0
[ 570.343716] amdgpu 0000:03:00.0: at page 0x000000f600700000 from 18
[ 570.343716] amdgpu 0000:03:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000010
[ 570.525510] [drm:psp_resume [amdgpu]] *ERROR* PSP resume failed
[ 570.525523] [drm:amdgpu_device_ip_resume_phase2 [amdgpu]] *ERROR* resume of IP
block <psp> failed -62
[ 570.525536] [drm:amdgpu_device_resume [amdgpu]] *ERROR*
amdgpu_device_ip_resume failed (-62).
[ 570.536704] e1000e: enp0s31f6 NIC Link is Up 100 Mbps Full Duplex, Flow
[ 570.540496] dpm_run_callback(): pci_pm_resume+0x0/0xa0 returns -62
[ 570.547879] e1000e 0000:00:1f.6 enp0s31f6: 10/100 speed: disabling TSO
[ 570.555434] call 0000:03:00.0+ returned -62 after 1973202 usecs
[ 570.689812] PM: Device 0000:03:00.0 failed to resume async: error -62
I attached the whole dmesg.
amd-gfx mailing list