Attention is currently required from: neels, laforge, pespin, fixeria.
14 comments:
File src/sccp_scoc.c:
Patch Set #14, Line 1110: xua_opt_data_send_cache(conn, SUA_CO_CORE, xua->hdr.msg_class);
maybe i was misreading this code. […]
Ack
File src/sccp_scoc.c:
Patch Set #15, Line 642: /* optional: importance */
like i said before, a patch should ideally do one thing. […]
Done
Simply drop it then.
Done
Patch Set #15, Line 606: } else
not critical, but IMHO there should be curly braces. […]
Ack
Patch Set #15, Line 606: } else
not critical, but IMHO there should be curly braces. […]
Ack
Patch Set #15, Line 635: LOGP(DLSCCP, LOGL_ERROR, "replacing unsent %u bytes of optional data cache with %s optional data\n",
I understand, this case is where we want to cache optional data, but there already is other data in […]
Ack
Patch Set #15, Line 635: LOGP(DLSCCP, LOGL_ERROR, "replacing unsent %u bytes of optional data cache with %s optional data\n",
I understand, this case is where we want to cache optional data, but there already is other data in […]
Ack
Patch Set #15, Line 784: xua_msg_add_sccp_addr(xua, SUA_IEI_DEST_ADDR, &conn->calling_addr);
Then write a new comit "comments are placed to match the fields order in the spec. […]
Done
File src/sccp_scoc.c:
Patch Set #16, Line 592: nt exp_type
could you explain a situation where there might be a mismatch from the expected type? Isn't it alway […]
Both expected types refer to the message which was source of the optional data. This check is simply an additional safeguard to ensure we hadn't screwed up while adjusting FSM. The caching happens in one place, sending in another but we always know what kind of Optional Data we're about to send (i. e. from which message it was cached). So here we compare the type of the message as recorded while caching with the type of the message FSM thinks it's sending.
Patch Set #16, Line 638: msgb_trim(conn->opt_data_cache, 0);
(i'd prefer sanitation: msgb_free() here, and always alloc a new one. […]
I don't: if msgb_trim() is implemented properly there should be no difference in our case (fixed buffer size) besides wasted CPU cycle on unnecessary function calls.
Patch Set #16, Line 667: see Figure C.3 / Q.714 (sheet 2 of 6) */
(osmocom asks to put '*' on every line of comment)
Done
Patch Set #16, Line 673: if (xua_drop_data_check_drop(prim, SCCP_MAX_DATA, "cache overrun"))
the optional data is larger than the upper limit of an SCCP DATA section -- this is not related to c […]
Done
Patch Set #16, Line 675: /* There's no need to cache the optional data since the connection is still active at this point */
/* Send the Optional Data in a DT1 ahead of the RLSD, because it is too large to be sent in one mess […]
Done
(again modifying unrelated comments)
Done
To view, visit change 29084. To unsubscribe, or for help writing mail filters, visit settings.