KMS does not work on MSI B350M MORTAR with AMD Ryzen 3 2200g

[Paul Menzel]

Dear Alex,


On 07/17/18 00:00, Paul Menzel wrote:

> Am 16.07.2018 um 18:30 schrieb Alex Deucher:
>> On Mon, Jul 16, 2018 at 12:14 PM, Paul Menzel
>> <pmenzel+amd-gfx at molgen.mpg.de> wrote:
>>> Dear Linux folks,
>>>
>>>
>>> Trying to boot Debian Buster/testing with Linux 4.16.16 on a MSI B350M MORTAR [1]
>>> with a Ryzen 3 2200g and a monitor connected over HDMI, when kernel mode setting
>>> (KMS) is enabled the monitor does not receive any updates, that means I do not
>>> see the login prompt and only the Linux and systemd service messages from before.
>>> The system seems to running though.
>>>
>>> Debian’s package *firmware-amd-graphics* [2] is from August 2017 and therefore
>>> outdated. I manually installed the firmware files from the Linux firmware
>>> repository [3].
>>>
>>> Here is the log.
>>>
>>> ```
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652026] [drm] BIOS signature incorrect 0 0
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652058] ATOM BIOS: 113-RAVEN-106
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652096] [drm] vm size is 262144 GB, 4 levels, block size is 9-bit, fragment size is 9-bit
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652107] amdgpu 0000:38:00.0: VRAM: 1024M 0x000000F400000000 - 0x000000F43FFFFFFF (1024M used)
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652109] amdgpu 0000:38:00.0: GTT: 1024M 0x000000F500000000 - 0x000000F53FFFFFFF
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652113] [drm] Detected VRAM RAM=1024M, BAR=256M
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652114] [drm] RAM width 128bits UNKNOWN
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652170] [TTM] Zone  kernel: Available graphics memory: 7703370 kiB
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652171] [TTM] Zone   dma32: Available graphics memory: 2097152 kiB
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652171] [TTM] Initializing pool allocator
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652174] [TTM] Initializing DMA pool allocator
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652202] [drm] amdgpu: 1024M of VRAM memory ready
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652203] [drm] amdgpu: 3072M of GTT memory ready.
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652220] [drm] GART: num cpu pages 262144, num gpu pages 262144
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652433] [drm] PCIE GART of 1024M enabled (table at 0x000000F400800000).
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652541] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652541] [drm] Driver supports precise vblank timestamp query.
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.652983] amdgpu 0000:38:00.0: firmware: direct-loading firmware amdgpu/raven_asd.bin
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.653276] amdgpu 0000:38:00.0: firmware: direct-loading firmware amdgpu/raven_pfp.bin
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.653498] amdgpu 0000:38:00.0: firmware: direct-loading firmware amdgpu/raven_me.bin
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.653668] amdgpu 0000:38:00.0: firmware: direct-loading firmware amdgpu/raven_ce.bin
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.653960] amdgpu 0000:38:00.0: firmware: direct-loading firmware amdgpu/raven_rlc.bin
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.654817] amdgpu 0000:38:00.0: firmware: direct-loading firmware amdgpu/raven_mec.bin
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.656137] amdgpu 0000:38:00.0: firmware: direct-loading firmware amdgpu/raven_mec2.bin
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.658217] amdgpu 0000:38:00.0: firmware: direct-loading firmware amdgpu/raven_sdma.bin
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.658220] [drm] use_doorbell being set to: [true]
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.659663] amdgpu 0000:38:00.0: firmware: direct-loading firmware amdgpu/raven_vcn.bin
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.659667] [drm] Found VCN firmware Version: 1.73 Family ID: 18
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.847424] [drm] VCN decode and encode initialized successfully.
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.852189] amdgpu 0000:38:00.0: ring 0(gfx) uses VM inv eng 4 on hub 0
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.852191] amdgpu 0000:38:00.0: ring 1(comp_1.0.0) uses VM inv eng 5 on hub 0
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.852193] amdgpu 0000:38:00.0: ring 2(comp_1.1.0) uses VM inv eng 6 on hub 0
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.852194] amdgpu 0000:38:00.0: ring 3(comp_1.2.0) uses VM inv eng 7 on hub 0
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.852195] amdgpu 0000:38:00.0: ring 4(comp_1.3.0) uses VM inv eng 8 on hub 0
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.852197] amdgpu 0000:38:00.0: ring 5(comp_1.0.1) uses VM inv eng 9 on hub 0
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.852198] amdgpu 0000:38:00.0: ring 6(comp_1.1.1) uses VM inv eng 10 on hub 0
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.852200] amdgpu 0000:38:00.0: ring 7(comp_1.2.1) uses VM inv eng 11 on hub 0
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.852201] amdgpu 0000:38:00.0: ring 8(comp_1.3.1) uses VM inv eng 12 on hub 0
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.852202] amdgpu 0000:38:00.0: ring 9(kiq_2.1.0) uses VM inv eng 13 on hub 0
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.852204] amdgpu 0000:38:00.0: ring 10(sdma0) uses VM inv eng 4 on hub 1
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.852205] amdgpu 0000:38:00.0: ring 11(vcn_dec) uses VM inv eng 5 on hub 1
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.852207] amdgpu 0000:38:00.0: ring 12(vcn_enc0) uses VM inv eng 6 on hub 1
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.852208] amdgpu 0000:38:00.0: ring 13(vcn_enc1) uses VM inv eng 7 on hub 1
>>> Jul 16 17:54:39 tokeiihto kernel: [   25.856580] [drm] Initialized amdgpu 3.23.0 20150101 for 0000:38:00.0 on minor 0
>>> ```
>>>
>>> The messages `[drm] BIOS signature incorrect 0 0` stands out. The latest
>>> mainboard firmware version 7A37v1F from May 2018 is installed [4].
>>>
>>> Booting with `nomodeset` works.
>>>
>>> Please find attached the full log with the KMS boot messages in the
>>> beginning, and the nomodeset messages in the second part.
>>>
>>> Do you have any hints how to get modesetting to work?
>>
>> Are
>> CONFIG_DRM_AMD_DC and CONFIG_DRM_AMD_DC_DCN1_0 set in your kernel config?
>> Those are required for modesetting on raven.
> 
> No, the second option is not selected in linux-image-4.16.0-2-amd64 [5].
> 
> ```
> #
> # Display Engine Configuration
> #
> CONFIG_DRM_AMD_DC=y
> # CONFIG_DRM_AMD_DC_PRE_VEGA is not set
> # CONFIG_DRM_AMD_DC_FBC is not set
> # CONFIG_DRM_AMD_DC_DCN1_0 is not set
> # CONFIG_DEBUG_KERNEL_DC is not set
> ```
> 
> It’s selected in linux-image-4.17.0-1 [6]. That version is not yet in Debian
> Buster/testing.

I upgraded to Debian Sid/unstable, and KMS works with linux-image-4.17.0-1,
which has both of options selected.

> I wonder, if in the future the Linux kernel could give feedback to the user> if options are unselected.


Kind regards,

Paul


> [5]: https://packages.debian.org/sid/amd64/linux-image-4.16.0-2-amd64
> [6]: https://packages.debian.org/sid/amd64/linux-image-4.17.0-1-amd64

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5174 bytes
Desc: S/MIME Cryptographic Signature
URL: <https://lists.freedesktop.org/archives/amd-gfx/attachments/20180717/04ddc98c/attachment.bin>