PCI Passthorugh INTERNAL_ERROR

TITUS MAXIMUS

XCPNG: 8.3
node: 20.18.3
npm: 10.8.3
xen-orchestra-upload-ova: 0.1.6
xen-orchestra-web: 0.17.1
xo-server: 5.177.4
xo-server-telemetry: 0.7.0
xo-server-xoa: 0.30.1
xo-web-free: 5.174.3
xoa-cli: 0.40.3
xoa-updater: 0.50.10

When I go to pass through a device I get the following error on the screen.

olivierlambert

Hi,

As stated in https://docs.xen-orchestra.com/community, you should provide the commit number and double check you are on the latest version available on master

TITUS MAXIMUS

@olivierlambert I was actually using the official XOA not from source, but I do have source instance installed with the commit of c8f9d and I have having the same issue.

olivierlambert

Oh sorry, I thought you were on the sources In that case, you just have to provide the XOA channel you are on, that's enough

Regarding your issue, it's hard to tell, it seems you have a problem at the host level. Have you enabled IOMMU in the bios?

TITUS MAXIMUS

@olivierlambert I have SR-IOV enabled in the BIOS.

olivierlambert

What's the result of lspci on your machine?

Also pinging @Team-Hypervisor-Kernel

TeddyAstie

Not a Xen issue.
This seems to be either a configuration issue (knowing /opt/xensource/libexec/xen-cmdline --get-dom0 may help) causing a issue in XAPI (@Team-XAPI-Network).

Maybe crashing in xapi/pciops.ml#L71-L80 or xapi/xapi_pci_helpers.ml#L179-L207.

TITUS MAXIMUS

@TeddyAstie I ran /opt/xensource/libexec/xen-cmdline --get-dom0 on the host and it didn't return anything.

EDIT: I also want to mention that I upgraded from 8.2 to 8.3 is that makes any difference. But I only started to testing passing through devices after I upgraded.

andriy.sultanov

@TITUS-MAXIMUS Could you please attach /var/log/xensource.log from the time of the error? Would be very useful to have a backtrace from where the error occurs

andriy.sultanov

@TITUS-MAXIMUS --get-dom0 being empty means no PCI devices were hidden from dom0 either, did you follow this step of the guide? https://docs.xcp-ng.org/compute/#2-tell-xcp-ng-not-to-use-this-device-id-for-dom0

TITUS MAXIMUS

@andriy.sultanov Here are the logs. To your second question. I thought I was able to pass through devices from XO and then the host reboots. Maybe I misunderstood how this worked. Are you saying I need to do this in the console and then XO?

Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] PCI.get_dom0_access_status R:1f40280dd69a failed with exception Server_error(INTERNAL_ERROR, [ 126 ])
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] Raised Server_error(INTERNAL_ERROR, [ 126 ])
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] 1/15 xapi Raised at file ocaml/xapi-aux/helper_process.ml, line 17
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] 2/15 xapi Called from file ocaml/xapi/pciops.ml, line 73
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] 3/15 xapi Called from file ocaml/xapi/pciops.ml, line 95
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] 4/15 xapi Called from file ocaml/xapi/pciops.ml, line 99
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] 5/15 xapi Called from file ocaml/libs/xapi-stdext/lib/xapi-stdext-pervasives/pervasiveext.ml, line 24
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] 6/15 xapi Called from file ocaml/libs/xapi-stdext/lib/xapi-stdext-pervasives/pervasiveext.ml, line 39
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] 7/15 xapi Called from file ocaml/xapi/xapi_pci_helpers.ml, line 213
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] 8/15 xapi Called from file ocaml/xapi/message_forwarding.ml, line 141
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] 9/15 xapi Called from file ocaml/xapi/rbac.ml, line 188
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] 10/15 xapi Called from file ocaml/xapi/rbac.ml, line 197
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] 11/15 xapi Called from file ocaml/xapi/server_helpers.ml, line 77
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] 12/15 xapi Called from file ocaml/xapi/server_helpers.ml, line 96
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] 13/15 xapi Called from file ocaml/libs/xapi-stdext/lib/xapi-stdext-pervasives/pervasiveext.ml, line 24
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] 14/15 xapi Called from file ocaml/libs/xapi-stdext/lib/xapi-stdext-pervasives/pervasiveext.ml, line 39
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace] 15/15 xapi Called from file ocaml/libs/log/debug.ml, line 250
Jul  7 16:42:08 xcp-ng-01 xapi: [error||62070 :::80|dispatch:PCI.get_dom0_access_status D:7ee7c7e624f9|backtrace]

andriy.sultanov

@TITUS-MAXIMUS You are correct. Sorry, the command to run is /opt/xensource/libexec/xen-cmdline --get-dom0 xen-pciback.hide - does this return anything? what's the return code of the command?