--- Log for 17.09.119 Server: livingstone.freenode.net Channel: #rockbox --- Nick: logbot Version: Dancer V4.16 Started: 13 days and 23 hours ago 00.07.55 Join krabador [0] (~krabador@unaffiliated/krabador) 00.09.41 Quit krabador (Remote host closed the connection) 00.14.37 Quit ZincAlloy (Quit: Leaving.) 00.15.25 *** Saving seen data "./dancer.seen" 01.34.14 Quit PimpiN8 (Quit: My MacBook has gone to sleep. ZZZzzz…) 02.15.28 *** Saving seen data "./dancer.seen" 02.45.38 Quit Natch (Ping timeout: 246 seconds) 02.57.57 Join Jinx [0] (~Jinx@unaffiliated/jinx) 03.31.39 Quit michaelni (Ping timeout: 246 seconds) 03.44.25 Join michaelni [0] (~michael@213-47-68-29.cable.dynamic.surfer.at) 04.15.30 *** Saving seen data "./dancer.seen" 04.19.09 Quit olavx200 (Read error: Connection reset by peer) 04.19.12 Quit danielp3344 (Read error: Connection reset by peer) 04.19.16 Quit Halamix2[m] (Write error: Connection reset by peer) 04.29.57 Join danielp3344 [0] (danielp334@gateway/shell/matrix.org/x-doufflilxmcgrhmn) 04.50.53 Quit igitoor (*.net *.split) 04.56.20 Join igitoor [0] (igitur@unaffiliated/contempt) 05.10.42 Quit TheSeven (Ping timeout: 245 seconds) 05.11.06 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven) 05.18.57 Quit igitoor (*.net *.split) 05.20.48 Join Halamix2[m] [0] (halamix2ma@gateway/shell/matrix.org/x-jyigsxzrptpffooo) 05.20.48 Join olavx200 [0] (olavx200ma@gateway/shell/matrix.org/x-aautnocfenbwjdhk) 05.24.22 Join igitoor [0] (igitur@unaffiliated/contempt) 05.36.20 Join Natch [0] (~Natch@h-112-130.A444.priv.bahnhof.se) 06.15.32 *** Saving seen data "./dancer.seen" 07.15.02 Join ZincAlloy [0] (~Adium@2a02:8108:9440:dfc:8121:a6d:117d:bda0) 07.20.10 Quit ZincAlloy (Ping timeout: 276 seconds) 07.52.17 Join advcomp2019_ [0] (~advcomp20@65-131-186-179.sxct.qwest.net) 07.52.17 Quit advcomp2019_ (Changing host) 07.52.17 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019) 07.55.35 Quit advcomp2019 (Ping timeout: 246 seconds) 07.57.10 Join ZincAlloy [0] (~Adium@2a02:8108:9440:dfc:8121:a6d:117d:bda0) 08.02.25 Quit ZincAlloy (Ping timeout: 276 seconds) 08.13.11 Join advcomp2019__ [0] (~advcomp20@65-131-186-179.sxct.qwest.net) 08.13.11 Quit advcomp2019__ (Changing host) 08.13.11 Join advcomp2019__ [0] (~advcomp20@unaffiliated/advcomp2019) 08.15.33 *** Saving seen data "./dancer.seen" 08.16.38 Quit advcomp2019_ (Ping timeout: 265 seconds) 08.21.32 Quit TheSeven (Ping timeout: 245 seconds) 08.25.30 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven) 08.59.48 Join petur [0] (~petur@mx1.dekimo.be) 08.59.49 Quit petur (Changing host) 08.59.49 Join petur [0] (~petur@rockbox/developer/petur) 09.13.11 Quit JanC (Remote host closed the connection) 09.13.31 Join JanC [0] (~janc@lugwv/member/JanC) 10.15.35 *** Saving seen data "./dancer.seen" 10.47.29 Join PimpiN8 [0] (~textual@77.60.122.82) 10.51.44 Join pamaury [0] (~pamaury@rockbox/developer/pamaury) 11.00.17 Join dys [0] (~dys@2003:5b:203b:100:a64c:c8ff:fef4:13a6) 12.06.03 Join vmx [0] (~vmx@ip5f5ac62a.dynamic.kabel-deutschland.de) 12.06.07 Quit dys (Ping timeout: 245 seconds) 12.15.37 *** Saving seen data "./dancer.seen" 12.25.55 Quit PimpiN8 (Quit: My MacBook has gone to sleep. ZZZzzz…) 12.31.52 Join dys [0] (~dys@2003:5b:203b:100:a64c:c8ff:fef4:13a6) 12.36.07 Quit dys (Ping timeout: 245 seconds) 12.39.00 Join evilwombat [0] (~evilwomba@c-73-241-77-99.hsd1.ca.comcast.net) 12.40.52 # Hello. I seem to be seeing memory corruption on top-of-tree (and v3.13 and v3.14) when dealing with bookmarks and large playlists (3900+ entries). Loading a recent bookmark sometimes shows a playlist of the correct length, and sometimes shows a playlist whose length is like, 2x what I expect (with some variability in the actual number of tracks). In the latter instances, pulling up the playlist menu results in an error saying "E 12.40.53 # rror loading playlist file". Target is iHP-120. Has anyone run into this? 12.57.04 # evilwombat, There are a lot of possible memory corruption issues fixed in the latest DEV 12.57.58 # Bilgus, thanks. Do you know when this was last cut? I tried using RockboxUtility to install the "latest" version as of today. Would that include this? 12.58.15 # you say you have seen the same behavior in all three? 12.58.45 # its easily been 5 weeks ago so it should be in there.. 12.59.07 # Yeah. I've got a playlist of 3830 songs or so. Loading this playlist multiple times in a row sometimes shows 3830 tracks, and sometimes 77xx tracks, where xx is usually random. 13.00.23 # I've got an H320 that doesn't have this problem, and that's running 3.13. When loading the same playlist multiple times in a row, the H320 sometimes shows slightly fewer tracks than the actual playlist length, but never more. (race between UI and loader?) 13.01.07 # hmm or perhaps something isn't getting freed 13.01.09 # Both units have CF card mods, which probably complicates analysis further. Trying to swap them around to see if the problem follows the card 13.01.38 # Ill see if I can repro on one of my sansas 13.02.11 # ... although for what it's worth, the speed of the cards might be bringing out a race condition that would be less likely in an HDD-based unit, due to timing of things 13.02.15 # thank you 13.04.28 # Sometimes the "oversized playlist" issue can be reproduced by loading the same playlist repeatedly. Other times I'm seeing it by starting a playlist, creating a bookmark, and then resuming from that bookmark via the "recent bookmarks" menu 13.05.29 # Will try to put a different card into the H120 and see if the problem still happens, before I swap the card with the H320. Those things are a beast to take apart 13.09.21 # you might try deleting the playlistcontrol file after it happens 13.12.36 # It basically saves the changes to get to the end result rather than saving the whole playlist it might be getting corrupted and if that works I'd like to see it the next round 13.12.49 Join dys [0] (~dys@2003:5b:203b:100:a64c:c8ff:fef4:13a6) 13.17.44 # playlistcontrol file? 13.18.25 # ah, .rockbox/.playlist_control. Thanks, I'll give this a try 13.19.37 # I am still copying everything around, in preparation for an upcoming card swap. Will try this after all the cards are synced. I'll be back in about 30 minutes 13.21.34 Quit dys (Ping timeout: 250 seconds) 13.22.54 # I'm seeing some junk appear in the playlist control file when the playlist shows less then the actual number of tracks so this is probably the right 'track' so to speak 13.23.46 # and slow storage is probably exacerbating it.. 13.37.06 Join krabador [0] (~krabador@unaffiliated/krabador) 13.56.03 # Hmmm... interesting! What purpose does this file serve? Would you like me to try anything else on my end? 14.03.03 # Crap... just saw this happen on the H320 also, with a different card 14.04.15 # .playlist_control looks fine in this case - no extra characters or junk. Nice, cleanly-terminated ASCII text 14.05.46 # Oh, wait, sorry, false alarm - I mis-read the screen. I saw the "incomplete playlist load" problem, but not the oversized playlist problem. 14.06.52 Quit Halamix2[m] (Changing host) 14.06.52 Join Halamix2[m] [0] (halamix2ma@unaffiliated/halamix2) 14.06.52 Quit Halamix2[m] (Changing host) 14.06.52 Join Halamix2[m] [0] (halamix2ma@gateway/shell/matrix.org/x-jyigsxzrptpffooo) 14.07.33 Join ZincAlloy [0] (~Adium@ip5f5acea4.dynamic.kabel-deutschland.de) 14.08.52 # ... but even in the case where I saw an incomplete playlist, my .playlist_control looked fine 14.11.37 Quit ZincAlloy (Ping timeout: 240 seconds) 14.12.41 # Not to speak too soon, but it almost seems as if disabling dircache is making this less frequent (or working around it). It could be just due to a shift in timing of things, or just not enough testing on my part. I shall fsck my storage, rebuild my playlists, and exercise the playlist/bookmark features tomorrow during the day, and come back with my findings tomorrow. Thank you for the help! I must go crash now 14.15.10 Join dys [0] (~dys@tmo-115-103.customers.d1-online.com) 14.15.36 # ok 14.15.39 *** Saving seen data "./dancer.seen" 14.17.34 Join massiveH [0] (~massiveH@ool-18e4e9f2.dyn.optonline.net) 14.17.58 # FYI the playliost control file stores all the changes made to the playlist since it was loaded its how rockbox stores the changes rather than saving the whole playlist and possibly corrupting it on unexpected power down 14.19.21 # try saving it to a new file name and deleting it next time you have the oversized playlist and see what happens 14.28.01 # Sounds good. Last time I tried to view the oversized playlist, I saw an error message, but I'll see if it's possible to get to the "Save current playlist" option without using the playlist context menu 14.54.41 Quit krabador (Remote host closed the connection) 15.01.45 Join PimpiN8 [0] (~textual@77.60.122.82) 15.08.49 Quit Jinx (Ping timeout: 276 seconds) 15.17.47 Join Jinx [0] (~Jinx@unaffiliated/jinx) 15.24.49 Quit dys (Ping timeout: 240 seconds) 15.43.22 Quit evilwombat (Quit: Leaving) 16.03.12 Quit massiveH (Quit: Leaving) 16.03.24 Join evilwombat [0] (~evilwomba@c-73-241-77-99.hsd1.ca.comcast.net) 16.06.19 # Okay, stupid silly out-of-the-blue question - inside playlist_flush_callback(), why aren't we take the control_mutex *before* evaluating the two 'if' conditions? Or am I just too used to preemption and Rockbox has a non-preemptive kernel? 16.08.39 Join ZincAlloy [0] (~Adium@ip5f5acea4.dynamic.kabel-deutschland.de) 16.12.30 Quit Natch (Ping timeout: 245 seconds) 16.13.11 Quit ZincAlloy (Ping timeout: 265 seconds) 16.15.40 *** Saving seen data "./dancer.seen" 16.27.27 # https://github.com/Rockbox/rockbox/blob/master/apps/playlist.c#L1257 16.29.46 # https://www.rockbox.org/wiki/RockboxKernel 16.31.29 # there is nothing there that will cede control of the thread... 16.59.02 Quit PimpiN8 (Quit: My MacBook has gone to sleep. ZZZzzz…) 17.13.30 Quit petur (Quit: Connection reset by beer) 17.50.28 Quit pamaury (Quit: Konversation terminated!) 17.51.28 Join petur [0] (~petur@rockbox/developer/petur) 17.55.02 Join krabador [0] (~krabador@unaffiliated/krabador) 18.08.40 Join ZincAlloy [0] (~Adium@2a02:8108:9440:dfc:b8b1:8259:4305:9117) 18.15.41 *** Saving seen data "./dancer.seen" 18.38.20 Quit krabador (Remote host closed the connection) 18.45.29 Join Natch [0] (~Natch@h-112-130.A444.priv.bahnhof.se) 19.11.14 Join lebellium [0] (~lebellium@89-92-69-72.hfc.dyn.abo.bbox.fr) 19.41.40 Quit vmx (Remote host closed the connection) 20.15.45 *** Saving seen data "./dancer.seen" 21.25.53 Join dys [0] (~dys@2a01:598:a80b:ca23:9eb6:d0ff:feee:f529) 21.31.00 Quit dys (Ping timeout: 276 seconds) 22.12.44 Join PimpiN8 [0] (~textual@2001:1c04:3305:b700:2971:69ee:7806:d8f5) 22.13.12 Join quadjfet [0] (quadjfet@gateway/vpn/privateinternetaccess/quadjfet) 22.15.46 *** Saving seen data "./dancer.seen" 22.23.22 Join dys [0] (~dys@84.46.53.122) 22.43.53 Quit petur (Remote host closed the connection) 23.06.26 Quit PimpiN8 (Quit: My MacBook has gone to sleep. ZZZzzz…) 23.06.39 Quit ZincAlloy (Quit: Leaving.) 23.35.42 Join MrZeus [0] (~MrZeus@88-145-85-240.host.pobb.as13285.net) 23.46.02 Quit lebellium (Quit: Leaving)