gnrc_netif_lorawan: add missing NULL check [backport 2021.01]

[jia200x]

Backport of #16110

Contribution description

This PR adds a missing NULL check when allocating a gnrc_pktsnip_t on reception. If the pktbuf is full, GNRC netif doesn't return immediately.

This PR fixes that.

Testing procedure

Configure the Network Server to send a downlink to the node and try this patch:

diff --git a/sys/net/gnrc/netif/lorawan/gnrc_netif_lorawan.c b/sys/net/gnrc/netif/lorawan/gnrc_netif_lorawan.c
index c30519220f..a6fb79cb72 100644
--- a/sys/net/gnrc/netif/lorawan/gnrc_netif_lorawan.c
+++ b/sys/net/gnrc/netif/lorawan/gnrc_netif_lorawan.c
@@ -26,7 +26,7 @@
 #include "net/loramac.h"
 #include "net/gnrc/netreg.h"
 
-#define ENABLE_DEBUG 0
+#define ENABLE_DEBUG 1
 #include "debug.h"
 
 static uint8_t _nwkskey[LORAMAC_NWKSKEY_LEN];
@@ -91,7 +91,7 @@ void gnrc_lorawan_mcps_indication(gnrc_lorawan_t *mac, mcps_indication_t *ind)
 {
     (void)mac;
     gnrc_pktsnip_t *pkt = gnrc_pktbuf_add(NULL, ind->data.pkt->iol_base,
-                                          ind->data.pkt->iol_len,
+                                          1024,
                                           GNRC_NETTYPE_LORAWAN);
     if (!pkt) {
         DEBUG("gnrc_lorawan: mcps_indication: couldn't allocate pktbuf\n");

It should print "gnrc_lorawan: mcps_indication: couldn't allocate pktbuf" when a packet is received

Issues/PRs references

None so far

[leandrolanzieri]

ACK

[leandrolanzieri]

@miri64 any idea of why the python-tests action is failing?

[miri64]

@miri64 any idea of why the python-tests action is failing?

dd88ead, which was not part of the release, fixed some linting errors, so it seems like #16058 needs to be backported as well. :-/

[miri64]

(it happens due to a newer version of pylint which spots some previously false negatives).

[leandrolanzieri]

@miri64 any idea of why the python-tests action is failing?

dd88ead, which was not part of the release, fixed some linting errors, so it seems like #16058 needs to be backported as well. :-/

Thanks!

[leandrolanzieri]

This needs rebasing

[jia200x]

done!

[leandrolanzieri]

The CI is showing a lot of unrelated errors :/

[miri64]

Was there a toolchain update in the docker container?

[fjmolinas]

Was there a toolchain update in the docker container?

Yes MIPS toolchain was updated after the release, which could explain the issue RIOT-OS/riotdocker#107

[miri64]

Is there a PR that fixed the issues?

[miri64]

Versioned docker containers would be great ;-).

[fjmolinas]

Is there a PR that fixed the issues?

Can't find one...

[aabadie]

I think #16042 should be backported.

[fjmolinas]

@leandrolanzieri can you rebase? CI should pass now.

[leandrolanzieri]

@leandrolanzieri can you rebase? CI should pass now.

@jia200x 's PR :)

[jia200x]

done!

[leandrolanzieri]

Re-ACK

[jia200x]

thanks everybody for the review and related backports!!

[miri64]

🎉 hardest backport ever ^^