Neels Hofmeyr has posted comments on this change. ( https://gerrit.osmocom.org/12266 )
Change subject: vty: show subscriber: put() before printing the use count ...................................................................... Patch Set 1: > looks a bit like a hack and rely on not-really-always-guaranteed-behavior. I > guess the cleaner way would be to tell the printing function to subtract one.. I thought about that, but then also the 'show subscriber cache' iterates subscribers without get()ting them. And, since this is not multithreaded, there actually is no way how a subscriber can exist before this invocation and then be discarded by the put(). By definition we only find subscribers with a use count initially >= 1, and this count cannot decrease externally while the vty command is running. right? Instead, printing a use count that is not actually there is IMHO the hacky solution. Thinking if we might at some point add a used-by bitflag like for ran_conn, then we'd have to hack around in that, too... I guess the least hacky is to always get() and put() and expect a use count of 2 on the vty output. Yet ... the single-threaded-there-is-no-way-to-lose-the-vsub argument makes sense to me. Further opinions welcome... -- To view, visit https://gerrit.osmocom.org/12266 To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings Gerrit-Project: osmo-msc Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: Id02b57b7ed299b010b9f8b9e809548eb1e6aa699 Gerrit-Change-Number: 12266 Gerrit-PatchSet: 1 Gerrit-Owner: Neels Hofmeyr <nhofm...@sysmocom.de> Gerrit-Reviewer: Jenkins Builder (1000002) Gerrit-Reviewer: Neels Hofmeyr <nhofm...@sysmocom.de> Gerrit-CC: Harald Welte <lafo...@gnumonks.org> Gerrit-Comment-Date: Wed, 12 Dec 2018 11:33:01 +0000 Gerrit-HasComments: No Gerrit-HasLabels: No
