Re: [Qemu-devel] [PATCH v2] net: Flush queued packets when guest resumes

[Fam Zheng]

On Tue, 07/07 12:10, Michael S. Tsirkin wrote:
> On Tue, Jul 07, 2015 at 04:58:29PM +0800, Jason Wang wrote:
> > 
> > 
> > On 07/07/2015 04:13 PM, Michael S. Tsirkin wrote:
> > > On Tue, Jul 07, 2015 at 09:21:07AM +0800, Fam Zheng wrote:
> > >> Since commit 6e99c63 "net/socket: Drop net_socket_can_send" and friends,
> > >> net queues need to be explicitly flushed after qemu_can_send_packet()
> > >> returns false, because the netdev side will disable the polling of fd.
> > >>
> > >> This fixes the case of "cont" after "stop" (or migration).
> > >>
> > >> Signed-off-by: Fam Zheng <address@hidden>
> > > Note virtio has its own handler which must be used to
> > > flush packets - this one might run too early or too late.
> > 
> > If runs too realy (DRIVER_OK is not set), then: packet will be dropped
> > (if this patch is used with "drop virtio_net_can_receive()", or the
> > queue will be purged since qemu_can_send_packet() returns false. If too
> > late, at least tap read poll will be enabled. So still looks ok?
> 
> It's still not helpful for virtio. So which cards are fixed?
> 
> I just find the comment
>       This fixes the case of "cont" after "stop" (or migration)
> too vague.
> Can you please include the info about what's broken in the commit log?

It does fix virtio-net, as well as most other cards.  HMP "stop" then "cont"
breaks net, tested with host pinging guest while doing it.

Fam

> 
> > >
> > >> ---
> > >>
> > >> v2: Unify with VM stop handler. (Stefan)
> > >> ---
> > >>  net/net.c | 19 ++++++++++++-------
> > >>  1 file changed, 12 insertions(+), 7 deletions(-)
> > >>
> > >> diff --git a/net/net.c b/net/net.c
> > >> index 6ff7fec..28a5597 100644
> > >> --- a/net/net.c
> > >> +++ b/net/net.c
> > >> @@ -1257,14 +1257,19 @@ void qmp_set_link(const char *name, bool up, 
> > >> Error **errp)
> > >>  static void net_vm_change_state_handler(void *opaque, int running,
> > >>                                          RunState state)
> > >>  {
> > >> -    /* Complete all queued packets, to guarantee we don't modify
> > >> -     * state later when VM is not running.
> > >> -     */
> > >> -    if (!running) {
> > >> -        NetClientState *nc;
> > >> -        NetClientState *tmp;
> > >> +    NetClientState *nc;
> > >> +    NetClientState *tmp;
> > >>  
> > >> -        QTAILQ_FOREACH_SAFE(nc, &net_clients, next, tmp) {
> > >> +    QTAILQ_FOREACH_SAFE(nc, &net_clients, next, tmp) {
> > >> +        if (running) {
> > >> +            /* Flush queued packets and wake up backends. */
> > >> +            if (nc->peer && qemu_can_send_packet(nc)) {
> > >> +                qemu_flush_queued_packets(nc->peer);
> > >> +            }
> > >> +        } else {
> > >> +            /* Complete all queued packets, to guarantee we don't modify
> > >> +             * state later when VM is not running.
> > >> +             */
> > >>              qemu_flush_or_purge_queued_packets(nc, true);
> > >>          }
> > >>      }
> > >> -- 
> > >> 2.4.3