On Thu, Aug 15, 2024 at 8:48 PM Jakub Kicinski [off-list ref] wrote:

On Tue, 13 Aug 2024 21:13:05 +0000 Mina Almasry wrote:

quoted

+int dev_get_max_mp_channel(const struct net_device *dev)
+{
+     int i, max = -1;

I presume the bug from yesterday is self evident once reported? :)

Yes, my apologies. The int return value from this function was being
implicitly cast to unsigned in the check I think, and that failed the
check.

My test didn't catch that due to a test environment issue. I had 2
ethtools installed by accident on the machine, and the test was
invoking the wrong one. I think I have that ironed out now.

quoted

+     ASSERT_RTNL();
+
+     for (i = 0; i < dev->real_num_rx_queues; i++)
+             if (dev->_rx[i].mp_params.mp_priv)
+                     /* The number of queues is the idx plus 1. */
+                     max = i + 1;

The +1 is odd. The function as it stands reports min channel count.
Not max_mp_channel, if you ask me. And if you renamed it, you don't
have to use -1 as "not installed".

Will do.

--
Thanks,
Mina