amdgpu 0000:84:00.0: gpu post error! \\ Fatal error during GPU init

[Deucher, Alexander]

> -----Original Message-----
> From: amd-gfx [mailto:amd-gfx-bounces at lists.freedesktop.org] On Behalf
> Of Dennis Schridde
> Sent: Thursday, April 13, 2017 10:18 AM
> To: amd-gfx at lists.freedesktop.org
> Subject: amdgpu 0000:84:00.0: gpu post error! \\ Fatal error during GPU init
> 
> Hello!
> 
> I am trying to use an AMD FirePro S7150X2 with the AMDGPU driver of a
> Linux
> 4.10.9 kernel (CoreOS Container Linux) and linux-firmware
> e39f0e3e6897ad865b3704f61218ae83f98a85da, but I run into the following
> error
> after the amdgpu module is being loaded:
> 
> [   17.692746] amdgpu 0000:84:00.0: enabling device (0000 -> 0003)
> [   17.692940] [drm] initializing kernel modesetting (TONGA 0x1002:0x6929
> 0x1002:0x0334 0x00).
> [   17.692963] [drm] register mmio base: 0xD0100000
> [   17.692964] [drm] register mmio size: 262144
> [   17.692970] [drm] doorbell mmio base: 0xF0000000
> [   17.692971] [drm] doorbell mmio size: 2097152
> [   17.692980] [drm] probing gen 2 caps for device 10b5:8747 = 8796103/10e
> [   17.692981] [drm] probing mlw for device 10b5:8747 = 8796103
> [   17.692992] [drm] VCE enabled in physical mode
> [   18.648132] ATOM BIOS: C76301
> [   18.651758] [drm] GPU posting now...
> [   23.661513] [drm:amdgpu_connector_add [amdgpu]] *ERROR* atombios
> stuck in
> loop for more than 5secs aborting
> [   23.673155] [drm:amdgpu_connector_add [amdgpu]] *ERROR* atombios
> stuck
> executing F250 (len 334, WS 4, PS 0) @ 0xF365
> [   23.685453] [drm:amdgpu_connector_add [amdgpu]] *ERROR* atombios
> stuck
> executing DB34 (len 324, WS 4, PS 0) @ 0xDC2C
> [   23.697816] [drm:amdgpu_connector_add [amdgpu]] *ERROR* atombios
> stuck
> executing BCDE (len 254, WS 0, PS 4) @ 0xBDB4
> [   23.710137] [drm:amdgpu_connector_add [amdgpu]] *ERROR* atombios
> stuck
> executing B832 (len 143, WS 0, PS 8) @ 0xB8A9
> [   23.722451] amdgpu 0000:84:00.0: gpu post error!
> [   23.727950] amdgpu 0000:84:00.0: Fatal error during GPU init

Posting the GPU is failing.  The is the initial basic asic setup that is required before anything else can happen.  There seem to be timeouts waiting for some register states.  Is there anything special about your setup?  Can you try a vanilla kernel?

Alex

> [   23.734594] [drm] amdgpu: finishing device.
> [   23.739592] ------------[ cut here ]------------
> ...
> [   24.096608] ---[ end trace 88c8cb35b32e3b88 ]---
> [   24.102086] BUG: unable to handle kernel NULL pointer dereference at
> 0000000000000018
> [   24.111438] IP: __ww_mutex_lock+0x24/0xa0
> [   24.116222] PGD 0
> [   24.116223]
> [   24.120737] Oops: 0002 [#1] SMP
> ...
> 
> Please find a full log attached.
> 
> My kernel configuration is available at:
>  https://github.com/urzds/coreos-overlay/blob/hpc_support/sys-
> kernel/coreos-modules/files/{commonconfig-4.10,amd64_defconfig-4.10}
> Please refer to the the commit log of the "hpc_support" branch for my
> changes
> compared to the CoreOS CL stock config.
> 
> I would be very glad if you could help me in debugging the issue and getting
> the GPU running.
> 
> Thanks,
> Dennis