Can't issue Smart Reboot due to PCI Passthrough

[cloudrootab]

Are you using XOA or XO from the sources?

XO from the sources

Which release channel?

None

Provide your commit number

fd9c9

Describe the bug

Raising bug report, as per Oliviers request in thread https://xcp-ng.org/forum/topic/10232/feature-request-disable-smart-reboot-if-guests-have-pci-passthru

I have two GPUs. However, this leads to Smart Reboot not being able to complete. Which leads to it aborting the operation, and reverts the suspending of VMs. Instead I'm met with the error message, as described below.

GPU0: AMD RX6400 (attached through GUI)
GPU1: Nvidia GTX1060 (attached through CLI as per compute-docs)

Error message

host.restart
{
  "force": false,
  "suspendResidentVms": true,
  "bypassBlockedSuspend": false,
  "bypassCurrentVmCheck": false,
  "id": "3c08401c-446b-445d-9ea2-24393beae8e4"
}
{
  "errors": [
    {
      "code": "VM_HAS_PCI_ATTACHED",
      "params": [
        "OpaqueRef:ad56d720-83ff-4a57-dc26-51841905fcb9"
      ],
      "task": {
        "uuid": "fbb8b196-88e9-bc18-0d1f-06143897380e",
        "name_label": "Async.VM.suspend",
        "name_description": "",
        "allowed_operations": [],
        "current_operations": {},
        "created": "20250108T16:55:30Z",
        "finished": "20250108T16:55:30Z",
        "status": "failure",
        "resident_on": "OpaqueRef:8b1095b3-beee-1d7f-93f2-bf5a88896299",
        "progress": 1,
        "type": "<none/>",
        "result": "",
        "error_info": [
          "VM_HAS_PCI_ATTACHED",
          "OpaqueRef:ad56d720-83ff-4a57-dc26-51841905fcb9"
        ],
        "other_config": {},
        "subtask_of": "OpaqueRef:NULL",
        "subtasks": [],
        "backtrace": "(((process xapi)(filename ocaml/xapi/xapi_vm_lifecycle.ml)(line 742))((process xapi)(filename ocaml/xapi/xapi_vm_helpers.ml)(line 1653))((process xapi)(filename ocaml/libs/xapi-stdext/lib/xapi-stdext-pervasives/pervasiveext.ml)(line 24))((process xapi)(filename ocaml/libs/xapi-stdext/lib/xapi-stdext-pervasives/pervasiveext.ml)(line 39))((process xapi)(filename ocaml/xapi/helpers.ml)(line 1662))((process xapi)(filename ocaml/xapi/xapi_vm_helpers.ml)(line 1652))((process xapi)(filename ocaml/xapi/message_forwarding.ml)(line 2241))((process xapi)(filename ocaml/xapi/rbac.ml)(line 191))((process xapi)(filename ocaml/xapi/rbac.ml)(line 200))((process xapi)(filename ocaml/xapi/server_helpers.ml)(line 75)))"
      }
    },
    {
      "code": "VM_HAS_PCI_ATTACHED",
      "params": [
        "OpaqueRef:ef20a143-4ca5-b293-c111-a6905c49394c"
      ],
      "task": {
        "uuid": "d75716e1-ec5f-40af-0e05-d496c0379e96",
        "name_label": "Async.VM.suspend",
        "name_description": "",
        "allowed_operations": [],
        "current_operations": {},
        "created": "20250108T16:55:30Z",
        "finished": "20250108T16:55:30Z",
        "status": "failure",
        "resident_on": "OpaqueRef:8b1095b3-beee-1d7f-93f2-bf5a88896299",
        "progress": 1,
        "type": "<none/>",
        "result": "",
        "error_info": [
          "VM_HAS_PCI_ATTACHED",
          "OpaqueRef:ef20a143-4ca5-b293-c111-a6905c49394c"
        ],
        "other_config": {},
        "subtask_of": "OpaqueRef:NULL",
        "subtasks": [],
        "backtrace": "(((process xapi)(filename ocaml/xapi/xapi_vm_lifecycle.ml)(line 742))((process xapi)(filename ocaml/xapi/xapi_vm_helpers.ml)(line 1653))((process xapi)(filename ocaml/libs/xapi-stdext/lib/xapi-stdext-pervasives/pervasiveext.ml)(line 24))((process xapi)(filename ocaml/libs/xapi-stdext/lib/xapi-stdext-pervasives/pervasiveext.ml)(line 39))((process xapi)(filename ocaml/xapi/helpers.ml)(line 1662))((process xapi)(filename ocaml/xapi/xapi_vm_helpers.ml)(line 1652))((process xapi)(filename ocaml/xapi/message_forwarding.ml)(line 2241))((process xapi)(filename ocaml/xapi/rbac.ml)(line 191))((process xapi)(filename ocaml/xapi/rbac.ml)(line 200))((process xapi)(filename ocaml/xapi/server_helpers.ml)(line 75)))"
      }
    }
  ],
  "message": "",
  "name": "Error",
  "stack": "Error: 
    at next (/opt/xen-orchestra/@vates/async-each/index.js:83:24)
    at onFulfilled (/opt/xen-orchestra/@vates/async-each/index.js:56:7)
    at onFulfilledWrapper (/opt/xen-orchestra/@vates/async-each/index.js:58:41)"
}

To reproduce

Have GPUs attached through PCI-passthrough (I have two)
Issue a smart reboot from Host/Advanced.

Expected behavior

If it isn't possible to issue and do a Smart Reboot due to PCI devices being attached. Then this should disable the Smart Reboot button altogether. Preferably with a note that makes it clear why the button is disabled.

Screenshots

No response

Node

20.18.1

Hypervisor

XCP-NG 8.3

Additional context

No response

[olivierlambert]

Thanks for the report!