amdgpu fails to run VF for s7150x2 device

Tue Sep 11 14:48:53 UTC 2018

All of the necessary changes are working their way upstream.  Try the
latest amd-staging-drm-next branch:
https://cgit.freedesktop.org/~agd5f/linux/log/?h=amd-staging-drm-next

Alex
On Tue, Sep 11, 2018 at 8:49 AM Alexander Frolov <frolov at nicevt.ru> wrote:
>
> Hi all!
>
> I am running S7150x2 with single VF enabled. VF is created with gim.ko
> and passed through to the Ubuntu 18.04 LTE guest. The default driver
> (amdgpu) fails to run VF.
> The dmesg output is following:
>
> [    5.029131] [drm] amdgpu kernel modesetting enabled.
> [    6.084808] [drm] Device Version 0.0
> [    6.085153] [drm] Compression level 0 log level 0
> [    6.085472] [drm] 12286 io pages at offset 0x1000000
> [    6.085797] [drm] 16777216 byte draw area at offset 0x0
> [    6.086141] [drm] RAM header offset: 0x3ffe000
> [    6.088019] [drm] qxl: 16M of VRAM memory size
> [    6.088497] [drm] qxl: 63M of IO pages memory ready (VRAM domain)
> [    6.088999] [drm] qxl: 64M of Surface memory size
> [    6.146983] [drm] main mem slot 1 [f4000000,3ffe000]
> [    6.147415] [drm] surface mem slot 2 [f8000000,4000000]
> [    6.148248] [drm] fb mappable at 0xF4000000, size 3145728
> [    6.148672] [drm] fb: depth 24, pitch 4096, width 1024, height 768
> [    6.149354] fbcon: qxldrmfb (fb0) is primary device
> [    6.161713] qxl 0000:00:02.0: fb0: qxldrmfb frame buffer device
> [    6.180939] [drm] Initialized qxl 0.1.0 20120117 for 0000:00:02.0 on
> minor 0
> [    6.188972] [drm] initializing kernel modesetting (TONGA
> 0x1002:0x692F 0x0000:0x0000 0x00).
> [    6.189153] [drm] register mmio base: 0xFC000000
> [    6.189173] [drm] register mmio size: 4194304
> [    6.189391] [drm] PCI I/O BAR is not found.
> [    6.236438] [drm] PCI I/O BAR is not found. Using MMIO to access ATOM
> BIOS
> [    6.236745] [drm] GPU post is not needed
> [    6.236764] [drm] Changing default dispclk from 600Mhz to 625Mhz
> [    6.237744] [drm] vm size is 64 GB, block size is 13-bit, fragment
> size is 9-bit
> [    6.237965] amdgpu 0000:00:08.0: VRAM: 7888M 0x0000000000000000 -
> 0x00000001ECFFFFFF (7888M used)
> [    6.237999] amdgpu 0000:00:08.0: GTT: 1024M 0x00000001ED000000 -
> 0x000000022CFFFFFF
> [    6.238266] [drm] Detected VRAM RAM=7888M, BAR=256M
> [    6.238287] [drm] RAM width 256bits GDDR5
> [    6.238453] [drm] amdgpu: 7888M of VRAM memory ready
> [    6.238566] [drm] amdgpu: 7888M of GTT memory ready.
> [    6.240119] [drm] GART: num cpu pages 262144, num gpu pages 262144
> [    6.241330] [drm] PCIE GART of 1024M enabled (table at
> 0x0000000000040000).
> [    6.242955] amdgpu 0000:00:08.0: amdgpu: using MSI.
> [    6.243734] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
> [    6.244473] [drm] Driver supports precise vblank timestamp query.
> [    6.245645] [drm] amdgpu: irq initialized.
> [    6.246424] amdgpu: [powerplay] amdgpu: powerplay sw initialized
> [    6.247940] amdgpu 0000:00:08.0: fence driver on ring 0 use gpu addr
> 0x00000001ed400040, cpu addr 0x        (ptrval)
> [    6.248921] amdgpu 0000:00:08.0: fence driver on ring 1 use gpu addr
> 0x00000001ed4000c0, cpu addr 0x        (ptrval)
> [    6.249797] amdgpu 0000:00:08.0: fence driver on ring 2 use gpu addr
> 0x00000001ed400140, cpu addr 0x        (ptrval)
> [    6.250741] amdgpu 0000:00:08.0: fence driver on ring 3 use gpu addr
> 0x00000001ed4001c0, cpu addr 0x        (ptrval)
> [    6.251577] amdgpu 0000:00:08.0: fence driver on ring 4 use gpu addr
> 0x00000001ed400240, cpu addr 0x        (ptrval)
> [    6.252479] amdgpu 0000:00:08.0: fence driver on ring 5 use gpu addr
> 0x00000001ed4002c0, cpu addr 0x        (ptrval)
> [    6.253379] amdgpu 0000:00:08.0: fence driver on ring 6 use gpu addr
> 0x00000001ed400340, cpu addr 0x        (ptrval)
> [    6.254243] amdgpu 0000:00:08.0: fence driver on ring 7 use gpu addr
> 0x00000001ed4003c0, cpu addr 0x        (ptrval)
> [    6.255062] amdgpu 0000:00:08.0: fence driver on ring 8 use gpu addr
> 0x00000001ed400440, cpu addr 0x        (ptrval)
> [    6.255767] amdgpu 0000:00:08.0: fence driver on ring 9 use gpu addr
> 0x00000001ed4004e0, cpu addr 0x        (ptrval)
> [    6.257432] amdgpu 0000:00:08.0: fence driver on ring 10 use gpu addr
> 0x00000001ed400560, cpu addr 0x        (ptrval)
> [    6.258289] amdgpu 0000:00:08.0: fence driver on ring 11 use gpu addr
> 0x00000001ed4005e0, cpu addr 0x        (ptrval)
> [    6.267196] [drm] ring test on 0 succeeded in 15 usecs
> [    6.460505] [drm:gfx_v8_0_kiq_resume [amdgpu]] *ERROR* KCQ enable
> failed (scratch(0xC040)=0xCAFEDEAD)
> [    6.461246] [drm:amdgpu_device_init [amdgpu]] *ERROR* hw_init of IP
> block <gfx_v8_0> failed -22
> [    6.461933] amdgpu 0000:00:08.0: amdgpu_init failed
> [    6.465365] amdgpu 0000:00:08.0:         (ptrval) unpin not necessary
> [    6.466753] [drm] amdgpu: ttm finalized
> [    6.472417] amdgpu 0000:00:08.0: Fatal error during GPU init
> [   18.400713] [drm] amdgpu: finishing device.
> [   18.411047] amdgpu: probe of 0000:00:08.0 failed with error -22
>
> Does amdgpu support VF mode? Any help will be appreciated!
> Thank you!
>
> Best,
>    Alexander Frolov
>
> _______________________________________________
> amd-gfx mailing list
> amd-gfx at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx