On 1/27/2021 4:52 PM, Andrew Lunn wrote:
On Thu, Jan 28, 2021 at 02:50:14AM +0200, Vladimir Oltean wrote:
quoted
On Thu, Jan 28, 2021 at 01:46:30AM +0100, Andrew Lunn wrote:
quoted
quoted
+ case NETDEV_GOING_DOWN: {
+ struct dsa_port *dp, *cpu_dp;
+ struct dsa_switch_tree *dst;
+ int err = 0;
+
+ if (!netdev_uses_dsa(dev))
+ return NOTIFY_DONE;
+
+ cpu_dp = dev->dsa_ptr;
+ dst = cpu_dp->ds->dst;
+
+ list_for_each_entry(dp, &dst->ports, list) {
+ if (!dsa_is_user_port(dp->ds, dp->index)) {
!dsa_is_user_port() ??
That ! seems odd.
Oops, that's something that I refactored at the last minute after I
prototyped the idea from:
if (!dsa_is_user_port(dp->ds, dp->index))
continue;
because it looked uglier that way.
I was guessing it would be something like that. With that fixed:
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
When you fix it:
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
--
Florian