Re: [Qemu-devel] [PATCH] vhost-user-test: support VHOST_USER_SET_VRING_ENABLE

[Marc-André Lureau]

Hi

----- Original Message -----
> Hi
> 
> That patch doesn't work, vhost-user-test fails:
>  
> ERROR:tests/vhost-user-test.c:491:test_migrate_source_check: assertion
> failed: (!overlap)

Actually, with "vhost: let SET_VRING_ENABLE message depends on protocol 
feature" it works

ack after that patch.

> 
> ----- Original Message -----
> > vhost-user-test is broken now: it assumes
> > QEMU sends RESET_OWNER, and we stopped doing that.
> > Wait for ENABLE_RING with 0 instead.
> > 
> > Signed-off-by: Michael S. Tsirkin <address@hidden>
> > ---
> >  tests/vhost-user-test.c | 7 +++++--
> >  1 file changed, 5 insertions(+), 2 deletions(-)
> > 
> > diff --git a/tests/vhost-user-test.c b/tests/vhost-user-test.c
> > index 01cfc7e..022223b 100644
> > --- a/tests/vhost-user-test.c
> > +++ b/tests/vhost-user-test.c
> > @@ -70,6 +70,7 @@ typedef enum VhostUserRequest {
> >      VHOST_USER_SET_VRING_ERR = 14,
> >      VHOST_USER_GET_PROTOCOL_FEATURES = 15,
> >      VHOST_USER_SET_PROTOCOL_FEATURES = 16,
> > +    VHOST_USER_SET_VRING_ENABLE = 18,
> >      VHOST_USER_MAX
> >  } VhostUserRequest;
> >  
> > @@ -315,8 +316,10 @@ static void chr_read(void *opaque, const uint8_t *buf,
> > int size)
> >          g_cond_signal(&s->data_cond);
> >          break;
> >  
> > -    case VHOST_USER_RESET_OWNER:
> > -        s->fds_num = 0;
> > +    case VHOST_USER_SET_VRING_ENABLE:
> > +        if (!msg.payload.state.num) {
> > +            s->fds_num = 0;
> > +        }
> >          break;
> >  
> >      default:
> > --
> > MST
> > 
>