amdgpu crashes badly

[Lukasz Piatkowski]

I'm running current ubuntu 19.10 with kernel "Ubuntu 5.3.0-24.26-generic 5.3.10" on a ThinkPad T495 laptop, which is confirmed to be certified here: https://certification.ubuntu.com/hardware/201905-27049

Unfortunately, the experience is really bad: `amdgpu` driver in the kernel crashes sometimes. I managed to get the `dmesg` output, please see atachemnt. This is confirmed at least on linux-image-5.3.0-18 and linux-image-5.3.0-24 "linux-image-X-generic

ProblemType: Bug
DistroRelease: Ubuntu 19.10
Package: linux-image-generic 5.3.0.24.28
ProcVersionSignature: Ubuntu 5.3.0-24.26-generic 5.3.10
Uname: Linux 5.3.0-24-generic x86_64
ApportVersion: 2.20.11-0ubuntu8.2
Architecture: amd64
Date: Mon Dec 23 14:50:14 2019
InstallationDate: Installed on 2019-12-07 (16 days ago)
InstallationMedia: Ubuntu 19.10 "Eoan Ermine" - Release amd64 (20191017)
MachineType: LENOVO 20NJ0010PB
ProcFB: 0 amdgpudrmfb
ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-5.3.0-24-generic root=/dev/mapper/ubuntu--vg-rootfs ro rootflags=subvol=@ quiet splash vt.handoff=7
PulseList: Error: command ['pacmd', 'list'] failed with exit code 1: No PulseAudio daemon running, or not running as session daemon.
RelatedPackageVersions:
 linux-restricted-modules-5.3.0-24-generic N/A
 linux-backports-modules-5.3.0-24-generic N/A
 linux-firmware 1.183.3
SourcePackage: linux
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 10/28/2019
dmi.bios.vendor: LENOVO
dmi.bios.version: R12ET46W(1.16 )
dmi.board.asset.tag: Not Available
dmi.board.name: 20NJ0010PB
dmi.board.vendor: LENOVO
dmi.board.version: Not Defined
dmi.chassis.asset.tag: No Asset Information
dmi.chassis.type: 10
dmi.chassis.vendor: LENOVO
dmi.chassis.version: None
dmi.modalias: dmi:bvnLENOVO:bvrR12ET46W(1.16):bd10/28/2019:svnLENOVO:pn20NJ0010PB:pvrThinkPadT495:rvnLENOVO:rn20NJ0010PB:rvrNotDefined:cvnLENOVO:ct10:cvrNone:
dmi.product.family: ThinkPad T495
dmi.product.name: 20NJ0010PB
dmi.product.sku: LENOVO_MT_20NJ_BU_Think_FM_ThinkPad T495
dmi.product.version: ThinkPad T495
dmi.sys.vendor: LENOVO

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This change was made by a bot.

[Lukasz Piatkowski (piontec)]

Please let me know if I can help to dubug this issue.

[Lukasz Piatkowski (piontec)]

I updated the kernel to mainline 5.5.0-050500rc4-generic #201912291930 ,but now I get this in dmesg:

[ 1.198203] amdgpu 0000:06:00.0: Direct firmware load for amdgpu/picasso_ta.bin failed with error -2
[ 1.198207] amdgpu 0000:06:00.0: psp v10.0: Failed to load firmware "amdgpu/picasso_ta.bin"

[Kai-Heng Feng (kaihengfeng)]

Firmware can be found here:
https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/

Once the firmwares are copied to /lib/firmwares, initramfs also requires an update:
$ sudo update-initramfs -u -k all

[piontec (piontec-gmail)]

I checked there, but there's no 'picasso_ta.bin' file there: https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/amdgpu

[You-Sheng Yang (vicamo)]

That can currently be found at least in amdgpu proprietary driver 19.50-969691, so you may want to download that from AMD's support site, and extract that firmware blob from extracted amdgpu-dkms deb.

[Lukasz Piatkowski (piontec)]

Can you please share a link? I can't find this download on AMD's page :|

[Lukasz Piatkowski (piontec)]

OK, for the interested: it's available here: https://www.amd.com/en/support/kb/release-notes/rn-amdgpu-unified-linux