fabian18

https://github.com/RIOT-OS/RIOT/blob/7ee1b585737b87c5a7df355f81ff113515289dfe/core/msg.c#L274-L276 This assertion is triggered, after the call to ```C netif_get_opt(&netif->netif, NETOPT_LINK, 0, &link, sizeof(netopt_enable_t)); ``` ,for me at least.

I tested a bit with the following patch: Patch ```diff diff --git a/drivers/netdev/ieee802154.c b/drivers/netdev/ieee802154.c index 95fb902225..4465420995 100644 --- a/drivers/netdev/ieee802154.c +++ b/drivers/netdev/ieee802154.c @@ -33,7 +33,6 @@ void netdev_ieee802154_reset(netdev_ieee802154_t *dev) { /*...

Btw if someone does not want to use the procedures given in "Testing procedure": I am testing with `stm32eth`. And I can trigger `LINK_UP` / `LINK_DOWN` events when I enable...

> LGTM, but @benpicco offline told me that sometimes the prefix is not received, regardless, after the interface is brought up. If that could be debugged (or at least some...

> What is still needed here? I would [wish](https://github.com/RIOT-OS/RIOT/pull/17893#issuecomment-1220800404) that addresses are auto (re-)configured in `gnrc_ipv6_nib_iface_up()`.

They are lost. The application must take care of that, I think. If I add an address to an interface in Linux and put the interface down and up again,...

So the status in this PR is, as far as I understand. You would rather keep the addresses of an interface during a link restart, until there is a way...

> I searched the code base for netdev_driver_t , and i think there are some more netdevs to be adjusted. Oh sorry those are the netdevs which support `NETDEV_EVENT_LINK_*`. Ignore...

Ok but `cdc_ecm_netdev` does not support `NETDEV_EVENT_LINK_*`. What about that?

With `ENABLE_DEBUG=1` in `nanocoap/sock.c`, I get a failed assertion: ``` > client put fe80::3478:9157:9fd5:81c3%5 5683 /somepath somedata 2022-06-22 11:22:45,749 # client put fe80::3478:9157:9fd5:81c3%5 5683 /somepath somedata 2022-06-22 11:22:45,752 # nanocli:...