afalout wrote: > Hi Philippe, > > I'll let you pay if you can convince me that LMS sending a volume > control change because it felt like it is your fault :) > > On more serious topic: > > I recompiled everything (inc. libupnp-1.6.19) from source, as it kept > crashing on my FC20 32 bit box; used -g and -01 so valgring will track > symbols. Few isses to report > > A) - Looks like something does not like the increase to 200 jobs: > > [09:54:03.630860] TimerLoop:616 discovered service > > http://192.168.1.87:1763/ConnectionManager/66b388dd-cef8-d1b6-d18e-ec5024bf7f36/control.xml > > http://192.168.1.87:1763/ConnectionManager/66b388dd-cef8-d1b6-d18e-ec5024bf7f36/event.xml > [09:54:04.182824] CallbackActionHandler:522 Error in action callback -- > -204 (cookie 2667) > total jobs = 200, too many jobstotal jobs = 200, too many jobstotal jobs > = 200, too many jobstotal jobs = 200, too many jobstotal jobs = 200, too > many jobstotal jobs = 200, too many jobstotal jobs = 200, too many > jobs[09:54:06.358837] CallbackActionHandler:522 Error in action callback > -- -204 (cookie 2777) > [09:54:07.126946] CallbackActionHandler:522 Error in action callback -- > -204 (cookie 2778) > total jobs = 200, too many jobstotal jobs = 200, too many > jobs[09:54:07.896619] CallbackActionHandler:522 Error in action callback > -- -204 (cookie 2779) > [09:54:07.940042] CallbackEventHandler:661 uPNP search timeout >
How many players do you have currently ? libupnp consumes a lot of jobs and when a player becomes unresponsive, a few pending jkobs will be generated before the player is removed from the list. That can create the "too many pending job" but increasing the #of jobs is not making the problem worse, quite the opposite. What you might have to do in fact is increasing even more. I have 7 players activated on my system > > b) Running via valgrind, it seems there is quite a bit of memory leaking > : > > ==24927== LEAK SUMMARY: > ==24927== definitely lost: 420,685 bytes in 9,673 blocks <<<<<<<< a > lot! > ==24927== indirectly lost: 3,131,371 bytes in 137,994 blocks > ==24927== possibly lost: 7,113 bytes in 104 blocks > ==24927== still reachable: 13,165,191 bytes in 5,724 blocks > ==24927== suppressed: 0 bytes in 0 blocks > ==24927== Reachable blocks (those to which a pointer was found) are not > shown. > ==24927== To see them, rerun with: --leak-check=full > --show-leak-kinds=all > ==24927== > ==24927== For counts of detected and suppressed errors, rerun with: -v > ==24927== Use --track-origins=yes to see where uninitialised values come > from > ==24927== ERROR SUMMARY: 134 errors from 42 contexts (suppressed: 0 from > 0) > > ... with stack overruns, invalid reads, and the such. > I've not run it through valgrind. The difficulty is that is also creates a lot of "false alarm" and it takes _a_lot_ of time to go through them. But okay, I'll do a pass. There is also libupnp that is not (reall) under my control > > c) When running via valgrind, it seems to run without crashing; when > running without it, it keeps crashing after an hour or two (No coredump > tho) This behaviour is quite tipycal for code with memory handling > isses. > > I would suggest checking out valgring output carefully, quite a bit of > "interesting" stuff reported, and GCC compile warnings might be of > interest too. > WRT gcc, I've done a passes on warning and so far, I think that none of them is impacting, but I'll re-check today. There a re few ones that are due to a cast that I should make on a call that is made 30 times, , but it has not consequence. WRT to crash, it is pretty strange. I'm not saying that it is crash-free, but a few other users are using it 24/7 and crash are very very rare - there has to be something specific. Do you have another machine where you could let it run for a while ? Cheers, Andrej LMS 7.7.2 - 5 radio, 3 Boom, 4 Duet, 1 Touch, 1 SB2. Sonos 2xPLAY:1, PLAY:3, PLAY:5, Marantz NR1603, JBL OnBeat, XBMC, Foobar2000, XBoxOne (sort of) ------------------------------------------------------------------------ philippe_44's Profile: http://forums.slimdevices.com/member.php?userid=17261 View this thread: http://forums.slimdevices.com/showthread.php?t=102496 _______________________________________________ plugins mailing list [email protected] http://lists.slimdevices.com/mailman/listinfo/plugins
