Re: [PATCH 2/2] i2c: virtio: fix completion handling

[Vincent Whitchurch]

On Wed, Oct 20, 2021 at 11:17:21AM +0200, Viresh Kumar wrote:
> On 20-10-21, 16:54, Jie Deng wrote:
> > 
> > On 2021/10/19 16:22, Viresh Kumar wrote:
> > > On 19-10-21, 09:46, Vincent Whitchurch wrote:
> > > >   static void virtio_i2c_msg_done(struct virtqueue *vq)
> > > >   {
> > > > -	struct virtio_i2c *vi = vq->vdev->priv;
> > > > +	struct virtio_i2c_req *req;
> > > > +	unsigned int len;
> > > > -	complete(&vi->completion);
> > > > +	while ((req = virtqueue_get_buf(vq, &len)))
> > > > +		complete(&req->completion);
> > > Instead of adding a completion for each request and using only the
> > > last one, maybe we can do this instead here:
> > > 
> > > 	while ((req = virtqueue_get_buf(vq, &len))) {
> > >                  if (req->out_hdr.flags == cpu_to_le32(VIRTIO_I2C_FLAGS_FAIL_NEXT))
> > 
> > 
> > Is this for the last one check ? For the last one, this bit should be
> > cleared, right ?
> 
> Oops, you are right. This should be `!=` instead. Thanks.

I don't quite understand how that would be safe since
virtqueue_add_sgs() can fail after a few iterations and all queued
request buffers can have FAIL_NEXT set.  In such a case, we would end up
waiting forever with your proposed change, wouldn't we?