diff options
Diffstat (limited to 'gitlab/issues/target_missing/host_missing/accel_missing/1717.toml')
| -rw-r--r-- | gitlab/issues/target_missing/host_missing/accel_missing/1717.toml | 37 |
1 files changed, 37 insertions, 0 deletions
diff --git a/gitlab/issues/target_missing/host_missing/accel_missing/1717.toml b/gitlab/issues/target_missing/host_missing/accel_missing/1717.toml new file mode 100644 index 00000000..cb02010c --- /dev/null +++ b/gitlab/issues/target_missing/host_missing/accel_missing/1717.toml @@ -0,0 +1,37 @@ +id = 1717 +title = "GPU passthrough (NV h100)case vfio Error" +state = "closed" +created_at = "2023-06-20T07:33:34.268Z" +closed_at = "2025-01-13T17:36:45.580Z" +labels = ["VFIO"] +url = "https://gitlab.com/qemu-project/qemu/-/issues/1717" +host-os = "ubuntu 22.04,redhat linux 9,centos9" +host-arch = "x86," +qemu-version = "6.2.0 8.0.2" +guest-os = "ubuntu22.04" +guest-arch = "x86" +description = """GPU passthrough (NV h100) will case a error + + +qemu-system-x86_64: vfio_err_notifier_handler(0000:17:00.0) Unrecoverable error detected. Please collect any data possible and then kill the guest + + +this error happen in centos, redhat linux,ubuntu with some kernel i have try( 5.19.0,6.0,6.2) +The same server insert L4,L40 GPU, will not happen. Only happen on H100 GPU +The same server install esxios. everything is normal. GPU work fine + +With vfio error. there is some idrac log error on my dell server + +``` +A bus fatal error was detected on a component at slot 2.\tTue Jun 20 2023 05:51:51 +A fatal error was detected on a component at bus 23 device 0 function 0.\tTue Jun 20 2023 05:51:51 +A fatal error was detected on a component at bus 22 device 2 function 0.\tTue Jun 20 2023 05:51:51 +``` + +Otherwise, I have try to passthrough gpu on dell amd and intel server both. +With AMD CPU , gpu not working in vm. but will not case vfio error +With INTEL CPU, will case vfio error.""" +reproduce = """1. Set GPU passthrought +2. Start VM +3. Do something in vm""" +additional = """""" |