minor memory leak in threadpool_queue_update()
Running Tor 0.2.9.4-alpha ish on moria1 for a day under valgrind, I found
==37985== 128 bytes in 8 blocks are definitely lost in loss record 33 of 92
==37985== at 0x4A06A2E: malloc (vg_replace_malloc.c:270)
==37985== by 0x269C27: tor_malloc_ (util.c:150)
==37985== by 0x26B5A5: tor_malloc_zero_ (util.c:178)
==37985== by 0x21E39E: worker_state_new (cpuworker.c:44)
==37985== by 0x26FD1C: threadpool_queue_update (workqueue.c:363)
==37985== by 0x21E25C: cpuworkers_rotate_keyinfo (cpuworker.c:187)
==37985== by 0x14D3E9: signal_callback (main.c:2173)
==37985== by 0x2B8FFA: event_signal_closure (in /home/tord/git/src/or/tor)
==37985== by 0x2B9CE9: event_process_active_single_queue (in /home/tord/git/src/or/tor)
==37985== by 0x2BA3BF: event_process_active (in /home/tord/git/src/or/tor)
==37985== by 0x2BAB23: event_base_loop (in /home/tord/git/src/or/tor)
==37985== by 0x14A05E: do_main_loop (main.c:2370)
Check out git show a52e5491
, which went into Tor 0.2.6.3-alpha.