On Wed, Nov 19, 2025 at 12:04:12PM +0100, Paolo Abeni wrote:
> On 11/19/25 7:55 AM, Michael S. Tsirkin wrote:
> > @@ -1720,6 +1720,7 @@ static long vhost_net_set_owner(struct vhost_net *n)
> > static long vhost_net_ioctl(struct file *f, unsigned int ioctl,
> > unsigned long arg)
> > {
> > + const DEFINE_VHOST_FEATURES_ARRAY(features_array, vhost_net_features);
>
> I'm sorry for the late feedback, I was drowning in other stuff.
>
> I have just a couple of non blocking suggestions, feel free to ignore.
>
> I think that if you rename `vhost_net_features` as
> `vhost_net_features_bits` and `features_array` as `vhost_net_features`
> the diffstat could be smaller and possibly clearer.
>
> > u64 all_features[VIRTIO_FEATURES_U64S];
> > struct vhost_net *n = f->private_data;
> > void __user *argp = (void __user *)arg;
> > @@ -1734,14 +1735,14 @@ static long vhost_net_ioctl(struct file *f,
> > unsigned int ioctl,
> > return -EFAULT;
> > return vhost_net_set_backend(n, backend.index, backend.fd);
> > case VHOST_GET_FEATURES:
> > - features = vhost_net_features[0];
> > + features = VHOST_FEATURES_U64(vhost_net_features, 0);
>
> Here and below you could use directly:
>
> features = features_array[0];
Ah. Good point.
> if you apply the rename mentioned above, this chunk and the following 3
> should not be needed.
>
> [...]> diff --git a/drivers/vhost/test.c b/drivers/vhost/test.c
> > index 42c955a5b211..af727fccfe40 100644
> > --- a/drivers/vhost/test.c
> > +++ b/drivers/vhost/test.c
> > @@ -308,6 +308,12 @@ static long vhost_test_set_backend(struct vhost_test
> > *n, unsigned index, int fd)
> > return r;
> > }
> >
> > +static const int vhost_test_features[] = {
> > + VHOST_FEATURES
> > +};
> > +
> > +#define VHOST_TEST_FEATURES VHOST_FEATURES_U64(vhost_test_features, 0)
>
> If you rename `VHOST_FEATURES` to `VHOST_FEATURES_BITS` and
> `VHOST_TEST_FEATURES` to `VHOST_FEATURES`,
This one I don't want to do, people tend to copy/paste code
and this is not what devices should be doing.
> the following two chunks
> should not be needed.
>
> Thanks,
>
> Paolo
