Attention is currently required from: laforge, pespin, fixeria. msuraev has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-msc/+/28849 )
Change subject: libosmonetify SMPP ......................................................................
Patch Set 57:
(4 comments)
File src/libsmpputil/smpp_msc.c:
https://gerrit.osmocom.org/c/osmo-msc/+/28849/comment/3f238ae1_e34abfe4 PS57, Line 849:
The bad side of using auto-formatting :)
Done
File src/libsmpputil/smpp_smsc.c:
https://gerrit.osmocom.org/c/osmo-msc/+/28849/comment/5d896a19_0ffd4a94 PS49, Line 985: osmo_stream_srv_link_close(smsc->link);
Then this needs to be investigated further and fixed, it probably means there's a bug.
Indeed, would be nice to dig deeper to understand why changing refcounting breaks tests. I don't see at as a reason to not merge a patch which doesn't change it though - this is clearly outside of the scope for this ticket.
File src/libsmpputil/smpp_smsc.c:
https://gerrit.osmocom.org/c/osmo-msc/+/28849/comment/26da2d31_76f68a19 PS57, Line 767: return -ENOMEM;
I know it's unlikely that msgb_alloc() would return NULL, but still: shouldn't we close the connecti […]
That depends whether we treat memory allocation as recoverable error or not. If it is than one of subsequent calls might succeed if memory pressure was reduced in a meantime. There's nothing wrong with the socket so why would we close it for unrelated error?
https://gerrit.osmocom.org/c/osmo-msc/+/28849/comment/52f8fc48_d103e0b3 PS57, Line 855: lost
Ack
Done
-
msuraev