AMD Ryzen 7 5800U support

ivan-volnov commented 2 years ago

Describe the bug Hi there!

The new Ryzen 7 5800U doesn't work on 13.0-RELEASE with drm-fbsd13-kmod. Hope I will be fixed on 13.1-RELEASE.

Can help with testing and patching but don't want to upgrade the whole system.

FreeBSD version

FreeBSD hostname 13.0-RELEASE-p8 FreeBSD 13.0-RELEASE-p8 #0: Tue Mar 15 09:36:28 UTC 2022     root@amd64-builder.daemonology.net:/usr/obj/usr/src/amd64.amd64/sys/GENERIC  amd64

PCI Info

pciconf -lv

``` hostb0@pci0:0:0:0: class=0x060000 rev=0x00 hdr=0x00 vendor=0x1022 device=0x1630 subvendor=0x1022 subdevice=0x1630 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Renoir/Cezanne Root Complex' class = bridge subclass = HOST-PCI none0@pci0:0:0:2: class=0x080600 rev=0x00 hdr=0x00 vendor=0x1022 device=0x1631 subvendor=0x1022 subdevice=0x1631 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Renoir/Cezanne IOMMU' class = base peripheral subclass = IOMMU hostb1@pci0:0:1:0: class=0x060000 rev=0x00 hdr=0x00 vendor=0x1022 device=0x1632 subvendor=0x0000 subdevice=0x0000 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Renoir PCIe Dummy Host Bridge' class = bridge subclass = HOST-PCI pcib1@pci0:0:1:1: class=0x060400 rev=0x00 hdr=0x01 vendor=0x1022 device=0x1633 subvendor=0x1022 subdevice=0x1453 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Renoir PCIe GPP Bridge' class = bridge subclass = PCI-PCI pcib2@pci0:0:1:3: class=0x060400 rev=0x00 hdr=0x01 vendor=0x1022 device=0x1634 subvendor=0x1022 subdevice=0x1453 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Renoir/Cezanne PCIe GPP Bridge' class = bridge subclass = PCI-PCI hostb2@pci0:0:2:0: class=0x060000 rev=0x00 hdr=0x00 vendor=0x1022 device=0x1632 subvendor=0x0000 subdevice=0x0000 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Renoir PCIe Dummy Host Bridge' class = bridge subclass = HOST-PCI pcib3@pci0:0:2:1: class=0x060400 rev=0x00 hdr=0x01 vendor=0x1022 device=0x1634 subvendor=0x1022 subdevice=0x1453 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Renoir/Cezanne PCIe GPP Bridge' class = bridge subclass = PCI-PCI pcib4@pci0:0:2:2: class=0x060400 rev=0x00 hdr=0x01 vendor=0x1022 device=0x1634 subvendor=0x1022 subdevice=0x1453 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Renoir/Cezanne PCIe GPP Bridge' class = bridge subclass = PCI-PCI pcib5@pci0:0:2:3: class=0x060400 rev=0x00 hdr=0x01 vendor=0x1022 device=0x1634 subvendor=0x1022 subdevice=0x1453 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Renoir/Cezanne PCIe GPP Bridge' class = bridge subclass = PCI-PCI hostb3@pci0:0:8:0: class=0x060000 rev=0x00 hdr=0x00 vendor=0x1022 device=0x1632 subvendor=0x0000 subdevice=0x0000 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Renoir PCIe Dummy Host Bridge' class = bridge subclass = HOST-PCI pcib6@pci0:0:8:1: class=0x060400 rev=0x00 hdr=0x01 vendor=0x1022 device=0x1635 subvendor=0x1022 subdevice=0x1635 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Renoir Internal PCIe GPP Bridge to Bus' class = bridge subclass = PCI-PCI pcib7@pci0:0:8:2: class=0x060400 rev=0x00 hdr=0x01 vendor=0x1022 device=0x1635 subvendor=0x1022 subdevice=0x1635 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Renoir Internal PCIe GPP Bridge to Bus' class = bridge subclass = PCI-PCI intsmb0@pci0:0:20:0: class=0x0c0500 rev=0x51 hdr=0x00 vendor=0x1022 device=0x790b subvendor=0x1022 subdevice=0x790b vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'FCH SMBus Controller' class = serial bus subclass = SMBus isab0@pci0:0:20:3: class=0x060100 rev=0x51 hdr=0x00 vendor=0x1022 device=0x790e subvendor=0x1022 subdevice=0x790e vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'FCH LPC Bridge' class = bridge subclass = PCI-ISA hostb4@pci0:0:24:0: class=0x060000 rev=0x00 hdr=0x00 vendor=0x1022 device=0x166a subvendor=0x0000 subdevice=0x0000 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Cezanne Data Fabric; Function 0' class = bridge subclass = HOST-PCI hostb5@pci0:0:24:1: class=0x060000 rev=0x00 hdr=0x00 vendor=0x1022 device=0x166b subvendor=0x0000 subdevice=0x0000 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Cezanne Data Fabric; Function 1' class = bridge subclass = HOST-PCI hostb6@pci0:0:24:2: class=0x060000 rev=0x00 hdr=0x00 vendor=0x1022 device=0x166c subvendor=0x0000 subdevice=0x0000 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Cezanne Data Fabric; Function 2' class = bridge subclass = HOST-PCI hostb7@pci0:0:24:3: class=0x060000 rev=0x00 hdr=0x00 vendor=0x1022 device=0x166d subvendor=0x0000 subdevice=0x0000 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Cezanne Data Fabric; Function 3' class = bridge subclass = HOST-PCI hostb8@pci0:0:24:4: class=0x060000 rev=0x00 hdr=0x00 vendor=0x1022 device=0x166e subvendor=0x0000 subdevice=0x0000 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Cezanne Data Fabric; Function 4' class = bridge subclass = HOST-PCI hostb9@pci0:0:24:5: class=0x060000 rev=0x00 hdr=0x00 vendor=0x1022 device=0x166f subvendor=0x0000 subdevice=0x0000 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Cezanne Data Fabric; Function 5' class = bridge subclass = HOST-PCI hostb10@pci0:0:24:6: class=0x060000 rev=0x00 hdr=0x00 vendor=0x1022 device=0x1670 subvendor=0x0000 subdevice=0x0000 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Cezanne Data Fabric; Function 6' class = bridge subclass = HOST-PCI hostb11@pci0:0:24:7: class=0x060000 rev=0x00 hdr=0x00 vendor=0x1022 device=0x1671 subvendor=0x0000 subdevice=0x0000 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Cezanne Data Fabric; Function 7' class = bridge subclass = HOST-PCI nvme0@pci0:1:0:0: class=0x010802 rev=0x01 hdr=0x00 vendor=0x1e4b device=0x1202 subvendor=0x1e4b subdevice=0x1202 vendor = 'MAXIO Technology (Hangzhou) Ltd.' device = 'NVMe SSD Controller MAP1202' class = mass storage subclass = NVM none1@pci0:2:0:0: class=0x020000 rev=0x05 hdr=0x00 vendor=0x10ec device=0x8125 subvendor=0x10ec subdevice=0x0123 vendor = 'Realtek Semiconductor Co., Ltd.' device = 'RTL8125 2.5GbE Controller' class = network subclass = ethernet re0@pci0:3:0:0: class=0x020000 rev=0x15 hdr=0x00 vendor=0x10ec device=0x8168 subvendor=0x10ec subdevice=0x0123 vendor = 'Realtek Semiconductor Co., Ltd.' device = 'RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller' class = network subclass = ethernet nvme1@pci0:5:0:0: class=0x010802 rev=0x03 hdr=0x00 vendor=0x126f device=0x2263 subvendor=0x126f subdevice=0x2263 vendor = 'Silicon Motion, Inc.' device = 'SM2263EN/SM2263XT SSD Controller' class = mass storage subclass = NVM vgapci0@pci0:6:0:0: class=0x030000 rev=0xc1 hdr=0x00 vendor=0x1002 device=0x1638 subvendor=0x1002 subdevice=0x0123 vendor = 'Advanced Micro Devices, Inc. [AMD/ATI]' device = 'Cezanne' class = display subclass = VGA hdac0@pci0:6:0:1: class=0x040300 rev=0x00 hdr=0x00 vendor=0x1002 device=0x1637 subvendor=0x1002 subdevice=0x1637 vendor = 'Advanced Micro Devices, Inc. [AMD/ATI]' device = 'Renoir Radeon High Definition Audio Controller' class = multimedia subclass = HDA none2@pci0:6:0:2: class=0x108000 rev=0x00 hdr=0x00 vendor=0x1022 device=0x15df subvendor=0x1022 subdevice=0x15df vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Family 17h (Models 10h-1fh) Platform Security Processor' class = encrypt/decrypt xhci0@pci0:6:0:3: class=0x0c0330 rev=0x00 hdr=0x00 vendor=0x1022 device=0x1639 subvendor=0x1022 subdevice=0x1639 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Renoir/Cezanne USB 3.1' class = serial bus subclass = USB xhci1@pci0:6:0:4: class=0x0c0330 rev=0x00 hdr=0x00 vendor=0x1022 device=0x1639 subvendor=0x1022 subdevice=0x1639 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Renoir/Cezanne USB 3.1' class = serial bus subclass = USB none3@pci0:6:0:5: class=0x048000 rev=0x01 hdr=0x00 vendor=0x1022 device=0x15e2 subvendor=0x1022 subdevice=0x15e2 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Raven/Raven2/FireFlight/Renoir Audio Processor' class = multimedia hdac1@pci0:6:0:6: class=0x040300 rev=0x00 hdr=0x00 vendor=0x1022 device=0x15e3 subvendor=0x10ec subdevice=0x0000 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'Family 17h (Models 10h-1fh) HD Audio Controller' class = multimedia subclass = HDA ahci0@pci0:7:0:0: class=0x010601 rev=0x81 hdr=0x00 vendor=0x1022 device=0x7901 subvendor=0x1022 subdevice=0x7901 vendor = 'Advanced Micro Devices, Inc. [AMD]' device = 'FCH SATA Controller [AHCI mode]' class = mass storage subclass = SATA ```

DRM KMOD version

drm-fbsd13-kmod 5.4.144.g20220128
drm-kmod g20190710_1

To Reproduce Try to launch a wayland DM dwl:

❯ dwl
00:00:10.505 [backend/backend.c:217] Found 0 GPUs, cannot create backend
00:00:10.505 [backend/backend.c:386] Failed to open any DRM device
couldn't create backend

Additional context

Found here a tread about the Cezanne core cpu. They say that drm-devel-kmod can be built on 13.1-BETA1, but Xorg 1.20.14 (also 1.20.13) crashes while accessing amdgpu.

neelchauhan commented 2 years ago

You need a newer version of drm-kmod. 13.0 won't work, but 13.1 can if you use 5.7 or newer.

ivan-volnov commented 2 years ago

Ok. Thanks. Switched to 13.1-RC1 (RC2 now) and installed drm-devel-kmod from ports. Works pretty well. Although there is no GPU accelerated video.

The only thing i have changed in /boot/loader.conf is hw.vga.textmode=1 to set booting fonts.

I also get this line in the logs:

amdgpu: os_same_file_description couldn't determine if two DRM fds reference the same file description.
If they do, bad things may happen!

Can I fix it somehow?

pointcheck commented 2 years ago

Xorg 1.20.14 still does not work on Cezenne CPUs, it crashes while, presumably, accessing AMDGPU. Also GPU offloading to NVIDIA does not work, I fail to make it detect more than one providers.

Using FreeBSD 13.1-RC1, drm-devel-kmod 5.7.19.g20220223, NVIDIA driver 510.60.02.

[drm] initializing kernel modesetting (RENOIR 0x1002:0x1638 0x17AA:0x3A5D 0xC6). nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms 510.60.02 Wed Mar 16 11:03:12 UTC 2022

ivan-volnov commented 2 years ago

Updated to FreeBSD 13.1-RC5. And added a second monitor via DP.

I experience reboots on the DP monitor start.

On driver load during reboot
On wayland session create
On wake up from sleep

Some of dmesg hope it helps:

VT: Replacing driver "efifb" with new "fb".
start FB_INFO:
type=11 height=1440 width=2560 depth=32
pbase=0x1010c8d000 vbase=0xfffff81010c8d000
name=drmn0 flags=0x0 stride=10240 bpp=32
end FB_INFO
drmn0: ring gfx uses VM inv eng 0 on hub 0
drmn0: ring comp_1.0.0 uses VM inv eng 1 on hub 0
drmn0: ring comp_1.1.0 uses VM inv eng 4 on hub 0
drmn0: ring comp_1.2.0 uses VM inv eng 5 on hub 0
drmn0: ring comp_1.3.0 uses VM inv eng 6 on hub 0
drmn0: ring comp_1.0.1 uses VM inv eng 7 on hub 0
drmn0: ring comp_1.1.1 uses VM inv eng 8 on hub 0
drmn0: ring comp_1.2.1 uses VM inv eng 9 on hub 0
drmn0: ring comp_1.3.1 uses VM inv eng 10 on hub 0
drmn0: ring kiq_2.1.0 uses VM inv eng 11 on hub 0
drmn0: ring sdma0 uses VM inv eng 0 on hub 1
drmn0: ring vcn_dec uses VM inv eng 1 on hub 1
drmn0: ring vcn_enc0 uses VM inv eng 4 on hub 1
drmn0: ring vcn_enc1 uses VM inv eng 5 on hub 1
drmn0: ring jpeg_dec uses VM inv eng 6 on hub 1
sysctl_warn_reuse: can't re-use a leaf (hw.dri.debug)!
[drm] Initialized amdgpu 3.37.0 20150101 for drmn0 on minor 0
acpi_wmi0: <ACPI-WMI mapping> on acpi0
acpi_wmi0: cannot find EC device
acpi_wmi0: Embedded MOF found
ACPI: \134AOD.WQBA: 1 arguments were passed to a non-method ACPI object (Buffer) (20201113/nsarguments-361)
driver bug: Unable to set devclass (class: ppc devname: (unknown))
intsmb0: <AMD FCH SMBus Controller> at device 20.0 on pci0
smbus0: <System Management Bus> on intsmb0
lo0: link state changed to UP
re0: link state changed to DOWN
uhid0 on uhub1
uhid0: <vendor 0x0c45 FC660M, class 0/0, rev 2.00/0.01, addr 1> on usbus1
ums0 on uhub1
ums0: <ROCCAT ROCCAT Kone Pure Military, class 0/0, rev 2.00/2.00, addr 2> on usbus1
ums0: 5 buttons and [XYZT] coordinates ID=1
pflog0: promiscuous mode enabled
re0: link state changed to UP
Security policy loaded: MAC/ntpd (mac_ntpd)
[drm] Fence fallback timer expired on ring gfx
[drm] Fence fallback timer expired on ring sdma0
[drm ERROR :drm_atomic_helper_wait_for_flip_done] [CRTC:62:crtc-3] flip_done timed out
[drm ERROR :drm_atomic_helper_wait_for_flip_done] [CRTC:62:crtc-3] flip_done timed out
WARNING atomic_read(&vblank->refcount) == 0 failed at /usr/ports/graphics/drm-devel-kmod/work/drm-kmod-drm_v5.7.19_3/drivers/gpu/drm/drm_vblank.c:1143
[drm] Fence fallback timer expired on ring gfx
[drm] Fence fallback timer expired on ring gfx

xmhd commented 2 years ago

I have a 5850u (which is the same cpu, just the 'pro' version) and it is working nicely (with gpu acceleration) on 14.0-CURRENT from April 7th 2022 along with drm-devel-kmod.

This morning I tried 13.1-RC5 and drm-510-kmod, which unfortunately do not work. I am able to 'kldload amdgpu' which compared to the same command on the aforementioned working setup, but doesn't seem to load the module properly -- when I load the module on 14.0-CURRENT, I see a decent amount of information output. On 13.1-RC5 I see the following info output:

isab1: <PCI-ISA bridge> at device 20.3 on pci0
device_attach: isab1 attach returned 6
<6>[drm] amdgpu kernel modesetting enabled

And when running 'startx' afterwards, the xorg server does not start.

edit: I have also observed the xf86-video-amdgpu 2d driver providing a significantly smoother experience than the modesetting driver that is part of xorg-server.

edit2: I was able to build drm-devel-kmod port on 13.1-RC5 and the 'amdgpu' module loads correctly. However I have now found that I am unable to move the mouse cursor (key input still works). I'm assuming this is i2c related and specific to my thinkpad x13 gen 2 (it works in 14.0-CURRENT).

edit3: drm-54-kmod does not work on 13.1-RC5 either.

tldr; the moral of this is:

if you need amd cezanne support, then drm-devel-kmod (5.7.x) will work
drm-510-kmod and drm-54-kmod likely do not work at this point in time

xmhd commented 2 years ago

so I done some digging and found the following missing pieces:

firmware for 5000 series/cezanne is currently missing, see: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=263793
however, as cezanne is a modified renoir, they work (mostly) with renoir firmware ...
drm-devel-kmod (v5.7) which works for my 5850u has the following commit adding the relevant device ID's https://github.com/freebsd/drm-kmod/commit/44a2eaf653479067ddee2c2db79e68993ec5b4b8
v5.4 and v5.4 do not have these device ID's added

I will test this out and report back.

wulf7 commented 2 years ago

Try to build recently created 5.10-lts branch from sources. It includes upstream support for all these Renoir chips. But I am not sure if we have all required firmwares in gpu-firmware-kmod port now or not.

xmhd commented 2 years ago

I will try that @wulf7 thanks for the advice

pointcheck commented 2 years ago

Hi all.

I have just tried 5.10-lts, it does not load on Cezanne CPU at all, reports 'amdgpu/green_sardine_sdma.bin' is missing, see my dmesg output below. So, 5.7-stable branch is the only that loads but still does not work - Xorg crashes at AMDGPU access.

I have been investigating this issue for quite some time and came to conclusion that the problem is not in DRM driver, neither in Xorg, but in /lib/libthr.so.3 or even in /lib/libc.so.7 library which comes with FreeBSD/LLVM. In other words it is a bug in FreeBSD.

drmn1: on vgapci1 vgapci1: child drmn1 requested pci_enable_io vgapci1: child drmn1 requested pci_enable_io [drm] initializing kernel modesetting (RENOIR 0x1002:0x1638 0x17AA:0x3A5D 0xC6). drmn1: Trusted Memory Zone (TMZ) feature disabled as experimental (default) [drm] register mmio base: 0xD1400000 [drm] register mmio size: 524288 [drm] add ip block number 0 [drm] add ip block number 1 [drm] add ip block number 2 [drm] add ip block number 3 [drm] add ip block number 4 [drm] add ip block number 5 [drm] add ip block number 6 [drm] add ip block number 7 [drm] add ip block number 8 [drm] add ip block number 9 drmn1: Fetched VBIOS from VFCT drmn1: could not load firmware image 'amdgpu/green_sardine_sdma.bin' [drm ERROR :sdma_v4_0_init_microcode] sdma_v4_0: Failed to load firmware "amdgpu/green_sardine_sdma.bin" [drm ERROR :sdma_v4_0_early_init] Failed to load sdma firmware! [drm] VCN decode is enabled in VM mode [drm] VCN encode is enabled in VM mode [drm] JPEG decode is enabled in VM mode [drm] vm size is 262144 GB, 4 levels, block size is 9-bit, fragment size is 9-bit drmn1: VRAM: 4096M 0x000000F400000000 - 0x000000F4FFFFFFFF (4096M used) drmn1: GART: 1024M 0x0000000000000000 - 0x000000003FFFFFFF drmn1: AGP: 267419648M 0x000000F800000000 - 0x0000FFFFFFFFFFFF [drm] Detected VRAM RAM=4096M, BAR=4096M [drm] RAM width 128bits DDR4 [drm] amdgpu: 4096M of VRAM memory ready [drm] amdgpu: 4096M of GTT memory ready. [drm] GART: num cpu pages 262144, num gpu pages 262144 [drm] PCIE GART of 1024M enabled (table at 0x000000F400900000). drmn1: could not load firmware image 'amdgpu/green_sardine_asd.bin' drmn1: fail to initialize asd microcode [drm ERROR :psp_sw_init] Failed to load psp firmware! [drm ERROR :amdgpu_device_ip_init] sw_init of IP block failed -2 drmn1: amdgpu_device_ip_init failed drmn1: Fatal error during GPU init drmn1: amdgpu: finishing device. [drm] amdgpu: ttm finalized device_attach: drmn1 attach returned 2

xmhd commented 2 years ago

Hi all.

I have just tried 5.10-lts, it does not load on Cezanne CPU at all, reports 'amdgpu/green_sardine_sdma.bin' is missing, see my dmesg output below. So, 5.7-stable branch is the only that loads but still does not work - Xorg crashes at AMDGPU access.

I have been investigating this issue for quite some time and came to conclusion that the problem is not in DRM driver, neither in Xorg, but in /lib/libthr.so.3 or even in /lib/libc.so.7 library which comes with FreeBSD/LLVM. In other words it is a bug in FreeBSD.

drmn1: on vgapci1 vgapci1: child drmn1 requested pci_enable_io vgapci1: child drmn1 requested pci_enable_io [drm] initializing kernel modesetting (RENOIR 0x1002:0x1638 0x17AA:0x3A5D 0xC6). drmn1: Trusted Memory Zone (TMZ) feature disabled as experimental (default) [drm] register mmio base: 0xD1400000 [drm] register mmio size: 524288 [drm] add ip block number 0 [drm] add ip block number 1 [drm] add ip block number 2 [drm] add ip block number 3 [drm] add ip block number 4 [drm] add ip block number 5 [drm] add ip block number 6 [drm] add ip block number 7 [drm] add ip block number 8 [drm] add ip block number 9 drmn1: Fetched VBIOS from VFCT drmn1: could not load firmware image 'amdgpu/green_sardine_sdma.bin' [drm ERROR :sdma_v4_0_init_microcode] sdma_v4_0: Failed to load firmware "amdgpu/green_sardine_sdma.bin" [drm ERROR :sdma_v4_0_early_init] Failed to load sdma firmware! [drm] VCN decode is enabled in VM mode [drm] VCN encode is enabled in VM mode [drm] JPEG decode is enabled in VM mode [drm] vm size is 262144 GB, 4 levels, block size is 9-bit, fragment size is 9-bit drmn1: VRAM: 4096M 0x000000F400000000 - 0x000000F4FFFFFFFF (4096M used) drmn1: GART: 1024M 0x0000000000000000 - 0x000000003FFFFFFF drmn1: AGP: 267419648M 0x000000F800000000 - 0x0000FFFFFFFFFFFF [drm] Detected VRAM RAM=4096M, BAR=4096M [drm] RAM width 128bits DDR4 [drm] amdgpu: 4096M of VRAM memory ready [drm] amdgpu: 4096M of GTT memory ready. [drm] GART: num cpu pages 262144, num gpu pages 262144 [drm] PCIE GART of 1024M enabled (table at 0x000000F400900000). drmn1: could not load firmware image 'amdgpu/green_sardine_asd.bin' drmn1: fail to initialize asd microcode [drm ERROR :psp_sw_init] Failed to load psp firmware! [drm ERROR :amdgpu_device_ip_init] sw_init of IP block failed -2 drmn1: amdgpu_device_ip_init failed drmn1: Fatal error during GPU init drmn1: amdgpu: finishing device. [drm] amdgpu: ttm finalized device_attach: drmn1 attach returned 2

Yes, even with the green sardine firmware missing the driver should still work (to a certain extent) with the renoir firmware as it had been doing so with the 5.7 branch.

I had last tested 5.7 branch with 14.0-CURRENT with the April 6th release of install media. So if it is broken in FreeBSD as you say, the change has happened somewhere in that time.

I did have the same outcome as you re: 5.10-lts branch though.

pointcheck commented 2 years ago

Can you please show me your Xorg.0.log as well as /etc/X11/xorg.conf ? Thanks.

I have installed missing green_sardine firmware manually from /usr/ports/graphics/gpu-firmware-kmod/ (strangely it's just missing in Makefile), now 5.10-lts driver loads ok, modesetting works, yet same issue with Xorg using amdgpu.

xmhd commented 2 years ago

Can you please show me your Xorg.0.log as well as /etc/X11/xorg.conf ? Thanks.

I have installed missing green_sardine firmware manually from /usr/ports/graphics/gpu-firmware-kmod/ (strangely it's just missing in Makefile), now 5.10-lts driver loads ok, modesetting works, yet same issue with Xorg using amdgpu.

I can try that tonight. If you don't mind, can you please provide the changes that you made?

I was looking at adding the firmware files to gpu-firmware-amd-kmod which doesn't look too difficult. But if we're debugging this and trying to come to the same outcome, I would prefer we follow the same steps.

pointcheck commented 2 years ago

I did not do any changes. I just installed missing firmware from their resting directories:

# cd /usr/ports/graphics/gpu-firmware-kmod && make install

# ls -d /usr/ports/graphics/gpu-firmware-kmod/work/drm-kmod-firmware-*/amdgpukmsfw/green_sardine_* | xargs -I % /bin/sh -c '(cd %; make install)'

xmhd commented 2 years ago

I did not do any changes. I just installed missing firmware from their resting directories:
# cd /usr/ports/graphics/gpu-firmware-kmod && make install

# ls -d /usr/ports/graphics/gpu-firmware-kmod/work/drm-kmod-firmware-*/amdgpukmsfw/green_sardine_* | xargs -I % /bin/sh -c '(cd %; make install)'

I wasn't able to replicate an install of greensardine* firmware files with this method, sorry. While I have some experience with the amdgpu driver, my experience with drm-kmod + gpu-firmware-kmod is limited :-)

I would imagine that xf86-video-amdgpu being stuck at version 19.x is the reason behind the crashing that you are experiencing.

ivan-volnov commented 2 years ago

Hi there!

Can anybody add a pull request to https://github.com/freebsd/drm-kmod-firmware to restore green_sardine support?

I opened an issue https://github.com/freebsd/drm-kmod-firmware/issues/21 for that.

xmhd commented 2 years ago

Hi there!

Can anybody add a pull request to https://github.com/freebsd/drm-kmod-firmware to restore green_sardine support?

I opened an issue freebsd/drm-kmod-firmware#21 for that.

I have done so. As mentioned in my PR, the next piece to be added is the relevant bits in the Makefile for the gpu-firmware-kmod port.

pointcheck commented 2 years ago

I would imagine that xf86-video-amdgpu being stuck at version 19.x is the reason behind the crashing that you are experiencing.

As I said I begin thinking that the problem is not in Xorg/xf86-video-amdgpu per se, but in libc library used in FreeBSD 13. I would like to see if this same issue persists in FreeBSD 14-CURRENT. Can you please send me your Xorg.0.log for investigation ?

xmhd commented 2 years ago

I would imagine that xf86-video-amdgpu being stuck at version 19.x is the reason behind the crashing that you are experiencing.

As I said I begin thinking that the problem is not in Xorg/xf86-video-amdgpu per se, but in libc library used in FreeBSD 13. I would like to see if this same issue persists in FreeBSD 14-CURRENT. Can you please send me your Xorg.0.log for investigation ?

Ok now I see what you're talking about. I tried out the 14.0-CURRENT image from April 14th 2022 and now I experience the same problem, even with v5.7 in use. So the breakage occured somewhere between April 7th and April 14th.

With regards to xorg.conf - I don't change any defaults. The only 'tweaks' I make is to /usr/local/share/X11/xorg.conf.d/10-amdgpu.conf where I enable TearFree for no screen tearing.

ivan-volnov commented 2 years ago

The firmware is available now.

Updated to 13.1-RELEASE, installed gpu-firmware-amd-kmod-green-sardine and drm-510-kmod. No luck. Described the issue here #167

Can anybody test it also?

evadot commented 2 years ago

Closing this issue, #167 will track the problem now that we (should) have everything to support this GPU generation.

freebsd / drm-kmod

AMD Ryzen 7 5800U support #155