On Thu, Feb 17, 2022 at 1:48 PM Ferruh Yigit <ferruh.yi...@intel.com> wrote:
>
> On 2/17/2022 12:44 PM, David Marchand wrote:
> > On Wed, Feb 9, 2022 at 10:48 AM Ciara Loftus <ciara.lof...@intel.com> wrote:
> >> @@ -1836,6 +2008,17 @@ rte_pmd_af_xdp_probe(struct rte_vdev_device *dev)
> >>                  return -1;
> >>          }
> >>
> >> +       /* Register IPC callback which shares xsk fds from primary to 
> >> secondary */
> >> +       if (!afxdp_dev_count) {
> >> +               ret = rte_mp_action_register(ETH_AF_XDP_MP_KEY, 
> >> afxdp_mp_send_fds);
> >> +               if (ret < 0) {
> >
> > This breaks --in-memory mode.
> > It should be instead ret < 0 && rte_errno != ENOTSUP
> >
> > Can you double check?
> > Thanks.
> >
>
> There is already a patch for it [1], if it is reviewed/tested,
> I can merge quickly.

I had tested my suggestion.
The patch [1] does the same and lgtm, I'll let Ciara ack it.

>
> [1]
> https://patches.dpdk.org/project/dpdk/patch/2c422567d3640...@cs.arizona.edu/


-- 
David Marchand

Reply via email to