Attention is currently required from: arehbein.
pespin has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-bsc/+/33892 )
Change subject: bsc (WIP): Make socket queue max. length configurable ......................................................................
Patch Set 2:
(4 comments)
Patchset:
PS2: I think the problem is that you are storing the config value directly in the wqueue, and then when initializing the wqueue you may erase it? Then add an extra field which is updated by VTY and use that when initializing the wqueue.
File src/osmo-bsc/bsc_vty.c:
https://gerrit.osmocom.org/c/osmo-bsc/+/33892/comment/6b5794f4_a314410c PS1, Line 2487: if (fd >= 0 && ((ofd = osmo_fd_get_by_fd(fd)) != NULL)) {
This is to check whether or not the socket and hence has been (properly) opened via `pcu_sock_init`, […]
why do you care about the socket being opened here? if you want to check for pcu-state presence, then simply: if(net->pcu_state) { ...}
https://gerrit.osmocom.org/c/osmo-bsc/+/33892/comment/9b834f0c_34581e20 PS1, Line 2491: }
That's what this does; osmo write queue code (the code in `osmo_wqueue_bfd_cb()`) will use the addre […]
ah I thought there may be an API so I didn't see it sorry.
Actually, there should be an API, since it should internally free the extra messages already enequeued if the new max_len is smaller.
File src/osmo-bsc/pcu_sock.c:
https://gerrit.osmocom.org/c/osmo-bsc/+/33892/comment/0d190f8a_33730028 PS1, Line 956: net->pcu_sock_fd = rc;
it enables the check commented on [above](https://gerrit.osmocom. […]
That check is not needed imho.
-
pespin