Panic under load when ran under QEMU

One of my NetBSD nodes started having panics under load - when running a fu=
ll bob build - recently. It is a VM on a ProxMox server, all updated and se=
tup somewhat non-standard - stuff accumulated over the years and perhaps no=
t following the best in practice advice; the virtual disks live on two ZFS =
pools backed by fast NVME disks. Anyway, I passed the panic traces to Claud=
e Code, which found a very plausible issue as follows:

----
Confirmed fixed. Summary for the NetBSD PR:

**Bug:** `sys/dev/pci/ld_virtio.c` doesn't negotiate `VIRTIO_F_RING_EVENT_I=
DX`. Under heavy I/O with `iothread=3D1` on QEMU, completion interrupts flo=
od the handler faster than it drains the used ring =E2=86=92 livelock =
=E2=86=92 watchdog panic.

**Evidence:** `if_vioif.c:508` negotiates it; `ld_virtio.c` has zero mentio=
ns. Block device features at boot show no `EVENT_IDX`; net device shows it.

**Workaround:** switch virtio-blk to virtio-scsi (`vioscsi` driver unaffect=
ed).

**Fix:** add `VIRTIO_F_RING_EVENT_IDX` to `req_features` in `ld_virtio.c` +=
 event-idx update after completion drain, mirroring `if_vioif.c` lines 508/=
2212/2381.

----

I won't be filing a PR as I understand LLM contributions are not accepted, =
but perhaps somebody might be interested to confirm it. I switched to virti=
o-scsi and I no longer have the issue.=20

Chavdar=20
=20


Sent with Proton Mail secure email.