benpicco
benpicco
Yes, I try not to do so many things at once in a PR. This one already - makes `NETDEV_EVENT_LINK_UP` mandatory - makes router advertisements depend on `NETDEV_EVENT_LINK_UP`/`NETDEV_EVENT_LINK_DOWN` events Dropping...
Thank you, missed that one.
I had to rebase to make static checks happy - spell checks have been fixed in a different PR.
Hm this seems to be one of those problems that go away when setting `ENABLE_DEBUG` to 1 :weary:
Wait what :fearful: Is that with the current `master`? When I try % coap-server-notls -A fe80::90a7:a6ff:fe4b:2e32%tapbr0 -d 1 and > client put fe80::90a7:a6ff:fe4b:2e32 5683 /somepath somedata I get ``` client...
I don't see this on `native`, is this on stm32?
Could also be an out of bounds write on the Ethernet driver, corrupting the pktbuf free list. can you try with https://github.com/RIOT-OS/RIOT/pull/18227
What 'works' surprisingly well is ```patch --- a/sys/net/application_layer/nanocoap/sock.c +++ b/sys/net/application_layer/nanocoap/sock.c @@ -174,7 +174,7 @@ ssize_t nanocoap_sock_request_cb(nanocoap_sock_t *sock, coap_pkt_t *pkt, case STATE_RESPONSE_RCVD: case STATE_RESPONSE_OK: if (ctx == NULL) { - DEBUG("nanocoap:...
So my issue was caused by a different thread being scheduled every 10 ms (and calling `ztimer_set` each time), which in turn caused the UART RX interrupt to be delayed...
Can you add ``` FEATURES_OPTIONAL += periph_uart_rs485 ``` to `tests/periph_uart` so this is compile-tested? And is the DE pin really all that's needed for RS485 operation?