Attention is currently required from: daniel. pespin has posted comments on this change. ( https://gerrit.osmocom.org/c/libosmocore/+/34144 ) Change subject: osmo_io: Avoid potential double free when sending msgb ...................................................................... Patch Set 1: (3 comments) Commit Message: https://gerrit.osmocom.org/c/libosmocore/+/34144/comment/cd32b5e7_41a2bc70 PS1, Line 16: mean time. To avoid that set the parent of any msgb in the tx queue to the time is indeed mean sometimes. Patchset: PS1: All the talloc parent in here looks already quite complex and I have the feeling it's becoming even more complex here. I think all this deserves a bit more of explanation/documentation since I think I'm not gasping it completely. IIUC we used to enqueue messages to be written to the socket in an internal queue of iofd, but those messages where actually not owned (talloc parent) by the iofd, but by some other random context outside of it. This looks wrong to me. messages enqueued in an iofd should always be owned by iofd. I think you are going into this direction (iofd owning it) but the solution passing lots of extra ctx doesn't look good to me at first sight. I'd welcome further explanation on why all those ctx pointers are needed. File src/core/osmo_io.c: https://gerrit.osmocom.org/c/libosmocore/+/34144/comment/197b573b_3b9482a1 PS1, Line 324: non-related ws -- To view, visit https://gerrit.osmocom.org/c/libosmocore/+/34144 To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings Gerrit-Project: libosmocore Gerrit-Branch: master Gerrit-Change-Id: I3a279b55a3adff96948120683c844e1508d0ba94 Gerrit-Change-Number: 34144 Gerrit-PatchSet: 1 Gerrit-Owner: daniel <dwillmann(a)sysmocom.de> Gerrit-Reviewer: Jenkins Builder Gerrit-Reviewer: laforge <laforge(a)osmocom.org> Gerrit-CC: pespin <pespin(a)sysmocom.de> Gerrit-Attention: daniel <dwillmann(a)sysmocom.de> Gerrit-Comment-Date: Fri, 11 Aug 2023 16:32:41 +0000 Gerrit-HasComments: Yes Gerrit-Has-Labels: No Gerrit-MessageType: comment