>
> On 20.10.2016 19:33, Loftus, Ciara wrote:
> >>>> @@ -894,6 +1006,66 @@ dpdk_dev_parse_name(const char
> >> dev_name[], const char prefix[],
> >>>> }
> >>>> }
> >>>>
> >>>> +/* When attaching a vhost device to DPDK, a unique name of the
> format
> >>>> + * 'net_vhostX' is expected, where X is a unique identifier.
> >>>> + * get_vhost_drv_id returns a valid X value to provide to DPDK.
> >>>> + */
> >>>> +static int
> >>>> +get_vhost_drv_id(void)
> >>>> +{
> >>>> + int i = 0;
> >>>> +
> >>>> + for (i = 0; i < RTE_MAX_ETHPORTS; i++) {
> >>>> + if (vhost_drv_ids[i] == 0) {
> >>>> + return i;
> >>>> + }
> >>>> + }
> >>>> +
> >>>> + return -1;
> >>>> +}
> >>>> +
> >>>> +static void
> >>>> +set_vhost_drv_id(int id, int val)
> >>>> +{
> >>>> + vhost_drv_ids[id] = val;
> >>>> +}
> >>>
> >>> I think we can just increment global atomic counter instead of above
> >> solution.
> >>> Numbers just need to be uniqe. Driver never use them for anything.
> >
> > But that way we would be limiting the number of vHost ports to the max
> value of the int we use as the counter.
>
> In that case I recommend to use existing 'id-pool' implementation instead.
I wasn't aware of the id-pool implementation. It's ideal for this case. I've
added it in the v4. Thanks!
>
> Best regards, Ilya Maximets.
>
> P.S.: Could you, please, add me in CC list for the future versions of
> this patch set? I just can't find it in patchwork.
Apologies - I missed you on the next version. You can find it here:
http://openvswitch.org/pipermail/dev/2016-October/080893.html
I'll add your CC from now onwards.
Not sure why they are not showing up in patchwork.
Thanks,
Ciara
_______________________________________________
dev mailing list
[email protected]
http://openvswitch.org/mailman/listinfo/dev
