On Wed, Apr 4, 2018 at 6:57 PM, Yann Ylavic <ylavic....@gmail.com> wrote: > On Wed, Apr 4, 2018 at 6:38 PM, William A Rowe Jr <wr...@rowe-clan.net> wrote: >> One question here; >> >> isn't it more efficient to offer push_resource_fifo() and >> push_resource_lifo() >> API's to the user, as opposed to extra conditionals? In theory, the consumer >> developer knows which ring they want. > > So we'd have two more _fifo versions of apr_reslist_release() and > apr_reslist_maintain() if we wanted to avoid the test somewhere. > Given that we already acquire the "listlock" mutex and signal the > "avail" condvar, this new test is pretty much in the noise already, > it's not worth duplicating the above functions IMHO.
Not to talk about apr_time_now() called on every push_resource()..
