Vfio/gpu guide for linux guest?

Software & Operating Systems Linux
,
1

tl;dr? Share your favorite up to date guide for GPU passthrough to a Linux guest under a Linux KVM host.

The long story:

I’m trying to minimize what the host OS does, to make it easier to “replace” at a whim. I’ve built VMs for various functions, one for containers, one for plex, several for Windows functions.

I’ve got no trouble getting GPUs to fully work for Windows guests, but can’t say the same for Linux. By fully work, I mean, X11 / Wayland / whatever they use these days. My Plex server is ubuntu bionic and I have a quadro p400 assigned, that works “good enough” for hw acceleration, but not enough that the kernel or X consider it a viable console (it will not output video to it at all).

I’ve tried masking the hypervisor like you do for a Windows guest and I’ve made sure to enable multifunction on the pass-through device entry in the XML, so the video + audio components are on the same bus/slot

X can see the GPU when it’s doing its configuration pass, but doesn’t seem to like it. Originally I thought this was a quirk of the platform, as I was experimenting with an i7-8700 at work on a Dell 5060 SFF machine, using a Firepro as the target VM. Windows loves it, Linux hates it seemingly. So I tried the same on my Threadripper at home, and similar result with my Quadro P400, the guest knows its there, loads kernel modules for it, but won’t display to it.

What should I post to troubleshoot this? Guest XML? Xorg log from the guest? lspci from the guest? something else?

2

show us your config files relating to how vfio-pci binds to your GPU.

can you access the linux VM over SSH to diagnose a potential driver issue on the guest?

3

The guest won’t boot without Spice installed, it turns on and sits at 50% cpu usage indefinitely.

Nothing fancy, the old fashioned way of blacklisting the radeon driver in modprobe and using vfio options:

Intel UHD 630 graphics is the first ID, FirePro GPU and AUDIO the second two

# cat vfio.conf
options vfio-pci ids=8086:3e92,1002:674a,1002:aa90

# cat /proc/cmdline
BOOT_IMAGE=(hd3,gpt2)/vmlinuz-5.4.15-200.fc31.x86_64 root=/dev/mapper/fedora-root ro resume=/dev/mapper/fedora-swap rd.lvm.lv=fedora/root rd.lvm.lv=fedora/swap rhgb quiet intel_iommu=on,igfx_off iommu=pt rd.driver.pre=vfio-pci video=vesafb:off i915.modeset=0

# lspci -vvs 01:00.0
01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Turks GL [FirePro V3900] (prog-if 00 [VGA controller])
        Subsystem: Advanced Micro Devices, Inc. [AMD/ATI] Device 0106
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR+ FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0, Cache Line Size: 64 bytes
        Interrupt: pin A routed to IRQ 16
        Region 0: Memory at d0000000 (64-bit, prefetchable) [size=256M]
        Region 2: Memory at e1220000 (64-bit, non-prefetchable) [size=128K]
        Region 4: I/O ports at 3000 [size=256]
        Expansion ROM at e1200000 [disabled] [size=128K]
        Capabilities: [50] Power Management version 3
                Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
                Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=0 PME-
        Capabilities: [58] Express (v2) Legacy Endpoint, MSI 00
                DevCap: MaxPayload 256 bytes, PhantFunc 0, Latency L0s <4us, L1 unlimited
                        ExtTag+ AttnBtn- AttnInd- PwrInd- RBE+ FLReset-
                DevCtl: CorrErr- NonFatalErr- FatalErr- UnsupReq-
                        RlxdOrd+ ExtTag+ PhantFunc- AuxPwr- NoSnoop+
                        MaxPayload 256 bytes, MaxReadReq 512 bytes
                DevSta: CorrErr- NonFatalErr- FatalErr- UnsupReq- AuxPwr- TransPend-
                LnkCap: Port #0, Speed 2.5GT/s, Width x16, ASPM L0s L1, Exit Latency L0s <64ns, L1 <1us
                        ClockPM- Surprise- LLActRep- BwNot- ASPMOptComp-
                LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes Disabled- CommClk+
                        ExtSynch- ClockPM- AutWidDis- BWInt- AutBWInt-
                LnkSta: Speed 2.5GT/s (ok), Width x16 (ok)
                        TrErr- Train- SlotClk+ DLActive- BWMgmt- ABWMgmt-
                DevCap2: Completion Timeout: Range ABCD, TimeoutDis+, LTR-, OBFF Not Supported
                         AtomicOpsCap: 32bit- 64bit- 128bitCAS-
                DevCtl2: Completion Timeout: 50us to 50ms, TimeoutDis-, LTR-, OBFF Disabled
                         AtomicOpsCtl: ReqEn-
                LnkCtl2: Target Link Speed: 2.5GT/s, EnterCompliance- SpeedDis-
                         Transmit Margin: Normal Operating Range, EnterModifiedCompliance- ComplianceSOS-
                         Compliance De-emphasis: -6dB
                LnkSta2: Current De-emphasis Level: -6dB, EqualizationComplete-, EqualizationPhase1-
                         EqualizationPhase2-, EqualizationPhase3-, LinkEqualizationRequest-
        Capabilities: [a0] MSI: Enable- Count=1/1 Maskable- 64bit+
                Address: 0000000000000000  Data: 0000
        Capabilities: [100 v1] Vendor Specific Information: ID=0001 Rev=1 Len=010 <?>
        Capabilities: [150 v1] Advanced Error Reporting
                UESta:  DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt- UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSViol-
                UEMsk:  DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt- UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSViol-
                UESvrt: DLP+ SDES+ TLP- FCP+ CmpltTO- CmpltAbrt- UnxCmplt- RxOF+ MalfTLP+ ECRC- UnsupReq- ACSViol-
                CESta:  RxErr- BadTLP- BadDLLP- Rollover- Timeout- AdvNonFatalErr-
                CEMsk:  RxErr- BadTLP- BadDLLP- Rollover- Timeout- AdvNonFatalErr-
                AERCap: First Error Pointer: 00, ECRCGenCap+ ECRCGenEn- ECRCChkCap+ ECRCChkEn-
                        MultHdrRecCap- MultHdrRecEn- TLPPfxPres- HdrLogCap-
                HeaderLog: 00000000 00000000 00000000 00000000
        Kernel driver in use: vfio-pci
        Kernel modules: radeon

# lspci -vvs 01:00.1
01:00.1 Audio device: Advanced Micro Devices, Inc. [AMD/ATI] Turks HDMI Audio [Radeon HD 6500/6600 / 6700M Series]
        Subsystem: Advanced Micro Devices, Inc. [AMD/ATI] Turks HDMI Audio [Radeon HD 6500/6600 / 6700M Series]
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR+ FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0, Cache Line Size: 64 bytes
        Interrupt: pin B routed to IRQ 17
        Region 0: Memory at e1240000 (64-bit, non-prefetchable) [size=16K]
        Capabilities: [50] Power Management version 3
                Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
                Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=0 PME-
        Capabilities: [58] Express (v2) Legacy Endpoint, MSI 00
                DevCap: MaxPayload 256 bytes, PhantFunc 0, Latency L0s <4us, L1 unlimited
                        ExtTag+ AttnBtn- AttnInd- PwrInd- RBE+ FLReset-
                DevCtl: CorrErr+ NonFatalErr+ FatalErr+ UnsupReq-
                        RlxdOrd+ ExtTag+ PhantFunc- AuxPwr- NoSnoop-
                        MaxPayload 256 bytes, MaxReadReq 512 bytes
                DevSta: CorrErr- NonFatalErr- FatalErr- UnsupReq- AuxPwr- TransPend-
                LnkCap: Port #0, Speed 2.5GT/s, Width x16, ASPM L0s L1, Exit Latency L0s <64ns, L1 <1us
                        ClockPM- Surprise- LLActRep- BwNot- ASPMOptComp-
                LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes Disabled- CommClk+
                        ExtSynch- ClockPM- AutWidDis- BWInt- AutBWInt-
                LnkSta: Speed 2.5GT/s (ok), Width x16 (ok)
                        TrErr- Train- SlotClk+ DLActive- BWMgmt- ABWMgmt-
                DevCap2: Completion Timeout: Range ABCD, TimeoutDis+, LTR-, OBFF Not Supported
                         AtomicOpsCap: 32bit- 64bit- 128bitCAS-
                DevCtl2: Completion Timeout: 50us to 50ms, TimeoutDis-, LTR-, OBFF Disabled
                         AtomicOpsCtl: ReqEn-
                LnkSta2: Current De-emphasis Level: -6dB, EqualizationComplete-, EqualizationPhase1-
                         EqualizationPhase2-, EqualizationPhase3-, LinkEqualizationRequest-
        Capabilities: [a0] MSI: Enable- Count=1/1 Maskable- 64bit+
                Address: 0000000000000000  Data: 0000
        Capabilities: [100 v1] Vendor Specific Information: ID=0001 Rev=1 Len=010 <?>
        Capabilities: [150 v1] Advanced Error Reporting
                UESta:  DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt- UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSViol-
                UEMsk:  DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt- UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSViol-
                UESvrt: DLP+ SDES+ TLP- FCP+ CmpltTO- CmpltAbrt- UnxCmplt- RxOF+ MalfTLP+ ECRC- UnsupReq- ACSViol-
                CESta:  RxErr- BadTLP- BadDLLP- Rollover- Timeout- AdvNonFatalErr-
                CEMsk:  RxErr- BadTLP- BadDLLP- Rollover- Timeout- AdvNonFatalErr-
                AERCap: First Error Pointer: 00, ECRCGenCap+ ECRCGenEn- ECRCChkCap+ ECRCChkEn-
                        MultHdrRecCap- MultHdrRecEn- TLPPfxPres- HdrLogCap-
                HeaderLog: 00000000 00000000 00000000 00000000
        Kernel driver in use: vfio-pci
        Kernel modules: snd_hda_intel
4

something is wrong with your libvirt/qemu setup. the VM should boot without spice. can you provide the XML for your VM?