This is merely a historical archive of years 2008-2021, before the migration to mailman3.
A maintained and still updated list archive can be found at https://lists.osmocom.org/hyperkitty/list/gerrit-log@lists.osmocom.org/.
Holger Freyther gerrit-no-reply at lists.osmocom.orgPatch Set 1: Code-Review-2 > There's nothing to be raised with upstream, I think. You want to > remove access to a private API, and I agree we should try that. > However, then remove the select() based RTCP code completely and > rely on polling, and verify that it all still works. If it does, > we can marge such a patch [but not the current one]. Only if it > doensn't we indeed should talk to upstream. I was about to write the same. Max, if you want to move things forward do the full work and not just half of it. A patch like this with all good intentions is making it a lot worse than moving things forward. So what you changed is: * select the RTCP socket * If it is readable ignore it * Hope/Assume that reading from the RTP socket will try to read RTCP too (yes, it seems to do that even in old oRTP versions). BUT: Let's assume this case: * oRTP is send/recv * RTP is blocked by a firewall/bad masquerading * RTCP arrives.. Now RTCP will never be read and the select will always trigger.. making any user of this code spend 99% CPU time.. and now you, Harald and me have spent time with it. Try to understand how the code around is using this method, it is static which makes it a lot easier. -- To view, visit https://gerrit.osmocom.org/908 To unsubscribe, visit https://gerrit.osmocom.org/settings Gerrit-MessageType: comment Gerrit-Change-Id: Iff1b15c68efca3e02267e0308142c6a7a0c2a974 Gerrit-PatchSet: 1 Gerrit-Project: libosmo-abis Gerrit-Branch: master Gerrit-Owner: Max <msuraev at sysmocom.de> Gerrit-Reviewer: Harald Welte <laforge at gnumonks.org> Gerrit-Reviewer: Holger Freyther <holger at freyther.de> Gerrit-Reviewer: Jenkins Builder Gerrit-Reviewer: Max <msuraev at sysmocom.de> Gerrit-HasComments: No