Skip to content
View in the app

A better way to browse. Learn more.
Unraid

A full-screen app on your home screen with push notifications, badges and more.

To install this app on iOS and iPadOS
  1. Tap the Share icon in Safari
  2. Scroll the menu and tap Add to Home Screen.
  3. Tap Add in the top-right corner.
To install this app on Android
  1. Tap the 3-dot menu (⋮) in the top-right corner of the browser.
  2. Tap Add to Home screen or Install app.
  3. Confirm by tapping Install.
Unraid Summer Sale in live: 20% off Starter and Unleashed + buy one, get one 50% off

Problem with AMD Radeon VEGA 64 GPU when VM is shutting down

chenci
in VM Engine (KVM)

Featured Replies

Hi, I started doing some testing with an AMD VEGA 64 GPU that I have it lying around, and I am checking if there is a way to passthrough correctly to a Windows 10 VM in Unraid. What I noticed is that it correctly detects it on VM boot, and I was able to install the AMD driver through Windows update (I heard that Adrenaline driver may have issues). Everything was working perfectly fine, until I shut down the VM. It will give me "amdgpu: Fatal error during GPU init", in which the VM just get stuck and Unraid is also unresponsive, that I have to hard-reset the server. I was reading in the forum that a lot of people were having issues with VEGA 64, I am wondering is this is a kernel issue or something else. I will appreciate anyone that can help.

I noticed another user have similar problem with his/her AMD GPU recently, and also with Unraid V.6.11.1.

The error that I get is this 

Oct 19 19:26:38 UnR kernel: amdgpu 0000:84:00.0: amdgpu: PSP runtime database doesn't exist
Oct 19 19:26:38 UnR kernel: amdgpu 0000:84:00.0: amdgpu: PSP runtime database doesn't exist
Oct 19 19:26:38 UnR kernel: amdgpu: hwmgr_sw_init smu backed is vega10_smu
Oct 19 19:26:38 UnR kernel: [drm] Found UVD firmware Version: 66.43 Family ID: 17
Oct 19 19:26:38 UnR kernel: [drm] PSP loading UVD firmware
Oct 19 19:26:38 UnR kernel: [drm] Found VCE firmware Version: 57.6 Binary ID: 4
Oct 19 19:26:38 UnR kernel: [drm] PSP loading VCE firmware
Oct 19 19:26:38 UnR kernel: [drm:psp_hw_start [amdgpu]] *ERROR* PSP create ring failed!
Oct 19 19:26:38 UnR kernel: [drm:psp_hw_init [amdgpu]] *ERROR* PSP firmware loading failed
Oct 19 19:26:38 UnR kernel: [drm:amdgpu_device_fw_loading [amdgpu]] *ERROR* hw_init of IP block <psp> failed -22
Oct 19 19:26:38 UnR kernel: amdgpu 0000:84:00.0: amdgpu: amdgpu_device_ip_init failed
Oct 19 19:26:38 UnR kernel: amdgpu 0000:84:00.0: amdgpu: Fatal error during GPU init
Oct 19 19:26:38 UnR kernel: amdgpu 0000:84:00.0: amdgpu: amdgpu: finishing device.
Oct 19 19:26:38 UnR kernel: amdgpu: probe of 0000:84:00.0 failed with error -22
Oct 19 19:26:38 UnR kernel: BUG: kernel NULL pointer dereference, address: 0000000000000090
Oct 19 19:26:38 UnR kernel: #PF: supervisor write access in kernel mode
Oct 19 19:26:38 UnR kernel: #PF: error_code(0x0002) - not-present page
Oct 19 19:26:38 UnR kernel: PGD 0 P4D 0 
Oct 19 19:26:38 UnR kernel: Oops: 0002 [#1] PREEMPT SMP PTI
Oct 19 19:26:38 UnR kernel: CPU: 14 PID: 23675 Comm: qemu-event Not tainted 5.19.14-Unraid #1
Oct 19 19:26:38 UnR kernel: Hardware name: Dell Inc. PowerEdge R730xd/072T6D, BIOS 2.11.0 11/02/2019
Oct 19 19:26:38 UnR kernel: RIP: 0010:drm_sched_fini+0x42/0x90 [gpu_sched]
Oct 19 19:26:38 UnR kernel: Code: e8 6c f4 c0 e0 48 8d ab 98 00 00 00 4c 8d 63 f8 48 85 ed 74 29 48 89 ef e8 b5 73 37 e1 48 8b 45 10 48 8d 55 10 48 39 d0 74 0c <c6> 80 90 00 00 00 01 48 8b 00 eb ef 48 89 ef e8 6d 74 37 e1 48 83
Oct 19 19:26:38 UnR kernel: RSP: 0018:ffffc9000b79fba0 EFLAGS: 00010217
Oct 19 19:26:38 UnR kernel: RAX: 0000000000000000 RBX: ffff888bacd09630 RCX: ffff88876d54cbc0
Oct 19 19:26:38 UnR kernel: RDX: ffff888bacd096d8 RSI: ffff88876d54cbe8 RDI: ffff888bacd096c8
Oct 19 19:26:38 UnR kernel: RBP: ffff888bacd096c8 R08: ffff88876d54c1c0 R09: 0000000080400037
Oct 19 19:26:38 UnR kernel: R10: ffff88876d54c1c0 R11: ffff8885874900b8 R12: ffff888bacd09628
Oct 19 19:26:38 UnR kernel: R13: ffff888bacd00000 R14: ffff8881055e5370 R15: ffff88820281d9a0
Oct 19 19:26:38 UnR kernel: FS:  00001528a88986c0(0000) GS:ffff8886679c0000(0000) knlGS:0000000000000000
Oct 19 19:26:38 UnR kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
Oct 19 19:26:38 UnR kernel: CR2: 0000000000000090 CR3: 0000000686992003 CR4: 00000000001706e0
Oct 19 19:26:38 UnR kernel: Call Trace:
Oct 19 19:26:38 UnR kernel: <TASK>
Oct 19 19:26:38 UnR kernel: amdgpu_fence_driver_sw_fini+0x35/0x7c [amdgpu]
Oct 19 19:26:38 UnR kernel: amdgpu_device_fini_sw+0x27/0x2bf [amdgpu]
Oct 19 19:26:38 UnR kernel: amdgpu_driver_release_kms+0x12/0x25 [amdgpu]
Oct 19 19:26:38 UnR kernel: drm_dev_put+0x34/0x62 [drm]
Oct 19 19:26:38 UnR kernel: release_nodes+0x40/0x5c
Oct 19 19:26:38 UnR kernel: devres_release_all+0x91/0xb9
Oct 19 19:26:38 UnR kernel: device_unbind_cleanup+0xe/0x61
Oct 19 19:26:38 UnR kernel: really_probe+0x268/0x273
Oct 19 19:26:38 UnR kernel: __driver_probe_device+0x8d/0xbd
Oct 19 19:26:38 UnR kernel: driver_probe_device+0x1f/0x77
Oct 19 19:26:38 UnR kernel: __device_attach_driver+0x83/0x97
Oct 19 19:26:38 UnR kernel: ? driver_allows_async_probing+0x58/0x58
Oct 19 19:26:38 UnR kernel: bus_for_each_drv+0x85/0xad
Oct 19 19:26:38 UnR kernel: __device_attach+0xb9/0x154
Oct 19 19:26:38 UnR kernel: bus_rescan_devices_helper+0x3a/0x69
Oct 19 19:26:38 UnR kernel: drivers_probe_store+0x34/0x50
Oct 19 19:26:38 UnR kernel: kernfs_fop_write_iter+0x137/0x17f
Oct 19 19:26:38 UnR kernel: new_sync_write+0x7f/0xbb
Oct 19 19:26:38 UnR kernel: vfs_write+0xda/0x129
Oct 19 19:26:38 UnR kernel: ksys_write+0x76/0xc2
Oct 19 19:26:38 UnR kernel: do_syscall_64+0x6b/0x81
Oct 19 19:26:38 UnR kernel: entry_SYSCALL_64_after_hwframe+0x63/0xcd
Oct 19 19:26:38 UnR kernel: RIP: 0033:0x1528acad842f
Oct 19 19:26:38 UnR kernel: Code: 89 54 24 18 48 89 74 24 10 89 7c 24 08 e8 19 2d f8 ff 48 8b 54 24 18 48 8b 74 24 10 41 89 c0 8b 7c 24 08 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 31 44 89 c7 48 89 44 24 08 e8 6c 2d f8 ff 48
Oct 19 19:26:38 UnR kernel: RSP: 002b:00001528a8897780 EFLAGS: 00000293 ORIG_RAX: 0000000000000001
Oct 19 19:26:38 UnR kernel: RAX: ffffffffffffffda RBX: 000000000000001d RCX: 00001528acad842f
Oct 19 19:26:38 UnR kernel: RDX: 000000000000000c RSI: 0000152894045e90 RDI: 000000000000001d
Oct 19 19:26:38 UnR kernel: RBP: 000000000000000c R08: 0000000000000000 R09: 0000152850003060
Oct 19 19:26:38 UnR kernel: R10: 0000000000000000 R11: 0000000000000293 R12: 0000152894045e90
Oct 19 19:26:38 UnR kernel: R13: 000000000000001d R14: 0000000000000000 R15: 00001528ad27a5c0
Oct 19 19:26:38 UnR kernel: </TASK>
Oct 19 19:26:38 UnR kernel: Modules linked in: xt_nat veth xt_CHECKSUM ipt_REJECT nf_reject_ipv4 xt_tcpudp ip6table_mangle ip6table_nat iptable_mangle vhost_net tun vhost vhost_iotlb tap xt_conntrack xt_MASQUERADE nf_conntrack_netlink nfnetlink xfrm_user xfrm_algo xt_addrtype iptable_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 br_netfilter xfs md_mod wireguard curve25519_x86_64 libcurve25519_generic libchacha20poly1305 chacha_x86_64 poly1305_x86_64 ip6_udp_tunnel udp_tunnel libchacha ip6table_filter ip6_tables iptable_filter ip_tables x_tables af_packet 8021q garp mrp bridge stp llc bonding tls ipv6 igb i40e amdgpu mxm_wmi gpu_sched drm_ttm_helper ttm drm_display_helper mgag200 drm_shmem_helper drm_kms_helper drm agpgart backlight syscopyarea ipmi_ssif sysfillrect x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm crct10dif_pclmul crc32_pclmul crc32c_intel ghash_clmulni_intel aesni_intel crypto_simd cryptd rapl intel_cstate intel_uncore sysimgblt input_leds fb_sys_fops i2c_algo_bit
Oct 19 19:26:38 UnR kernel: led_class joydev i2c_core mpt3sas ahci nvme libahci raid_class nvme_core scsi_transport_sas ipmi_si wmi acpi_power_meter button unix [last unloaded: igb]
Oct 19 19:26:38 UnR kernel: CR2: 0000000000000090
Oct 19 19:26:38 UnR kernel: ---[ end trace 0000000000000000 ]---
Oct 19 19:26:38 UnR kernel: RIP: 0010:drm_sched_fini+0x42/0x90 [gpu_sched]
Oct 19 19:26:38 UnR kernel: Code: e8 6c f4 c0 e0 48 8d ab 98 00 00 00 4c 8d 63 f8 48 85 ed 74 29 48 89 ef e8 b5 73 37 e1 48 8b 45 10 48 8d 55 10 48 39 d0 74 0c <c6> 80 90 00 00 00 01 48 8b 00 eb ef 48 89 ef e8 6d 74 37 e1 48 83
Oct 19 19:26:38 UnR kernel: RSP: 0018:ffffc9000b79fba0 EFLAGS: 00010217
Oct 19 19:26:38 UnR kernel: RAX: 0000000000000000 RBX: ffff888bacd09630 RCX: ffff88876d54cbc0
Oct 19 19:26:38 UnR kernel: RDX: ffff888bacd096d8 RSI: ffff88876d54cbe8 RDI: ffff888bacd096c8
Oct 19 19:26:38 UnR kernel: RBP: ffff888bacd096c8 R08: ffff88876d54c1c0 R09: 0000000080400037
Oct 19 19:26:38 UnR kernel: R10: ffff88876d54c1c0 R11: ffff8885874900b8 R12: ffff888bacd09628
Oct 19 19:26:38 UnR kernel: R13: ffff888bacd00000 R14: ffff8881055e5370 R15: ffff88820281d9a0
Oct 19 19:26:38 UnR kernel: FS:  00001528a88986c0(0000) GS:ffff8886679c0000(0000) knlGS:0000000000000000
Oct 19 19:26:38 UnR kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
Oct 19 19:26:38 UnR kernel: CR2: 0000000000000090 CR3: 0000000686992003 CR4: 00000000001706e0

 

Hardware

  • Dell PowerEdge R730xd with 2 x Intel Xeon 2650L V3 48GB RAM
  • Gigabyte AMD Radeon Vega 64, with VGA Bios loaded

Software

  • Unraid 6.11.1
  • VM: 8 cores passthrough, 12288MB RAM, Q35-7.1, Bios OVMF, Hyper-V yes, USB 3.0 qemu XHCI
  • Windows 10 pro

 

  • Quote
    • Author

    Found it, it's still the AMD reset bug problem. Strangely the script for the bug didn't work for me. So I used AMD Vendor Reset plugin from ICH777 and the problem was fixed. Now it works perfectly.

    Just be mindful the when you install the driver, I was not able to go above AMD Adrenalin 22.10.1 (I don't know why). You need to install the driver, otherwise, you won't have audio coming from HDMI.

    I also increased the fan speed a little bit, so when playing games, thermals won't be an issue. Very happy with the result inside the server.

  • Quote

    • Join the conversation

    You can post now and register later. If you have an account, sign in now to post with your account.
    Note: Your post will require moderator approval before it will be visible.

    Guest
    Reply to this topic...
    Go to topic listing

    Account

    Navigation

    Search

    Search

    Configure browser push notifications
    Chrome (Android)
    1. Tap the lock icon next to the address bar.
    2. Tap Permissions → Notifications.
    3. Adjust your preference.
    Chrome (Desktop)
    1. Click the padlock icon in the address bar.
    2. Select Site settings.
    3. Find Notifications and adjust your preference.