Re: [PATCH] virtio-mmio : fix the crash in the vm shutdown

[Jason Wang <jasowang@redhat.com>]

On Tue, Nov 9, 2021 at 10:37 AM Cindy Lu <lulu@redhat.com> wrote:
>
> The root cause for this crash is the ioeventfd not stopped while the VM stop.
> The callback for vmstate_change was not implement in virtio-mmio bus
>
> Reproduce step
> load the vm with
>  -M microvm \
>   -netdev tap,id=net0,vhostforce,script=no,downscript=no  \
>   -device virtio-net-device,netdev=net0\
>
> After the VM boot, login the vm and then shutdown the vm
>
> System will crash
> [Current thread is 1 (Thread 0x7ffff6edde00 (LWP 374378))]
> (gdb) bt
> 0  0x00005555558f18b4 in qemu_flush_or_purge_queued_packets (purge=false, nc=0x55500252e850) at ../net/net.c:636
> 1  qemu_flush_queued_packets (nc=0x55500252e850) at ../net/net.c:656
> 2  0x0000555555b6c363 in virtio_queue_notify_vq (vq=0x7fffe7e2b010) at ../hw/virtio/virtio.c:2339
> 3  virtio_queue_host_notifier_read (n=0x7fffe7e2b08c) at ../hw/virtio/virtio.c:3583
> 4  0x0000555555de7b5a in aio_dispatch_handler (ctx=ctx@entry=0x5555567c5780, node=0x555556b83fd0) at ../util/aio-posix.c:329
> 5  0x0000555555de8454 in aio_dispatch_ready_handlers (ready_list=<optimized out>, ctx=<optimized out>) at ../util/aio-posix.c:359
> 6  aio_poll (ctx=0x5555567c5780, blocking=blocking@entry=false) at ../util/aio-posix.c:662
> 7  0x0000555555cce0cc in monitor_cleanup () at ../monitor/monitor.c:645
> 8  0x0000555555b06bd2 in qemu_cleanup () at ../softmmu/runstate.c:822
> 9  0x000055555586e693 in main (argc=<optimized out>, argv=<optimized out>, envp=<optimized out>) at ../softmmu/main.c:51
>
> Signed-off-by: Cindy Lu <lulu@redhat.com>
> ---

Acked-by: Jason Wang <jasowang@redhat.com

Thanks

>  hw/virtio/virtio-mmio.c | 12 ++++++++++++
>  1 file changed, 12 insertions(+)
>
> diff --git a/hw/virtio/virtio-mmio.c b/hw/virtio/virtio-mmio.c
> index 7b3ebca178..72da12fea5 100644
> --- a/hw/virtio/virtio-mmio.c
> +++ b/hw/virtio/virtio-mmio.c
> @@ -817,6 +817,17 @@ static char *virtio_mmio_bus_get_dev_path(DeviceState *dev)
>      return path;
>  }
>
> +static void virtio_mmio_vmstate_change(DeviceState *d, bool running)
> +{
> +    VirtIOMMIOProxy *proxy = VIRTIO_MMIO(d);
> +
> +    if (running) {
> +        virtio_mmio_start_ioeventfd(proxy);
> +    } else {
> +        virtio_mmio_stop_ioeventfd(proxy);
> +    }
> +}
> +
>  static void virtio_mmio_bus_class_init(ObjectClass *klass, void *data)
>  {
>      BusClass *bus_class = BUS_CLASS(klass);
> @@ -832,6 +843,7 @@ static void virtio_mmio_bus_class_init(ObjectClass *klass, void *data)
>      k->ioeventfd_enabled = virtio_mmio_ioeventfd_enabled;
>      k->ioeventfd_assign = virtio_mmio_ioeventfd_assign;
>      k->pre_plugged = virtio_mmio_pre_plugged;
> +    k->vmstate_change = virtio_mmio_vmstate_change;
>      k->has_variable_vring_alignment = true;
>      bus_class->max_dev = 1;
>      bus_class->get_dev_path = virtio_mmio_bus_get_dev_path;
> --
> 2.21.3
>