On 2023/7/4 0:52, Michael S. Tsirkin wrote: > On Wed, Jun 14, 2023 at 09:01:47PM +0800, Hawkins Jiawei wrote: >> According to VirtIO standard, "The class, command and >> command-specific-data are set by the driver, >> and the device sets the ack byte. >> There is little it can do except issue a diagnostic >> if ack is not VIRTIO_NET_OK." >> >> Therefore, QEMU should stop sending the queued SVQ commands and >> cancel the device startup if the device's ack is not VIRTIO_NET_OK. >> >> Yet the problem is that, vhost_vdpa_net_load_x() returns 1 based on >> `*s->status != VIRTIO_NET_OK` when the device's ack is VIRTIO_NET_ERR. >> As a result, net->nc->info->load() also returns 1, this makes >> vhost_net_start_one() incorrectly assume the device state is >> successfully loaded by vhost_vdpa_net_load() and return 0, instead of >> goto `fail` label to cancel the device startup, as vhost_net_start_one() >> only cancels the device startup when net->nc->info->load() returns a >> negative value. >> >> This patch fixes this problem by returning -EINVAL when the device's >> ack is not VIRTIO_NET_OK. >> >> Fixes: f73c0c43ac ("vdpa: extract vhost_vdpa_net_load_mac from >> vhost_vdpa_net_load") >> Fixes: f64c7cda69 ("vdpa: Add vhost_vdpa_net_load_mq") >> Signed-off-by: Hawkins Jiawei <yin31...@gmail.com> >> --- >> net/vhost-vdpa.c | 10 +++++++--- >> 1 file changed, 7 insertions(+), 3 deletions(-) >> >> diff --git a/net/vhost-vdpa.c b/net/vhost-vdpa.c >> index 37cdc84562..630c9bf71e 100644 >> --- a/net/vhost-vdpa.c >> +++ b/net/vhost-vdpa.c >> @@ -651,8 +651,9 @@ static int vhost_vdpa_net_load_mac(VhostVDPAState *s, >> const VirtIONet *n) >> if (unlikely(dev_written < 0)) { >> return dev_written; >> } >> - >> - return *s->status != VIRTIO_NET_OK; >> + if (*s->status != VIRTIO_NET_OK) { >> + return -EINVAL; >> + } >> } >> >> return 0; >> @@ -676,8 +677,11 @@ static int vhost_vdpa_net_load_mq(VhostVDPAState *s, >> if (unlikely(dev_written < 0)) { >> return dev_written; >> } >> + if (*s->status != VIRTIO_NET_OK) { >> + return -EINVAL; >> + } >> >> - return *s->status != VIRTIO_NET_OK; >> + return 0; >> } > > I think EIO would be better, we have too many EINVAL cases, > making things hard to debug.
I will refactor this patch to return -EIO according to your suggestion. Thanks! > > >> >> static int vhost_vdpa_net_load(NetClientState *nc) >> -- >> 2.25.1 >