Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2011-08-17

00:03:13 Quit funman (Quit: Lost terminal)
00:05:18 Quit domonoky (Read error: Connection reset by peer)
00:09:18 Join GigaBrick [0] (
00:11:05 Quit kadoban (Ping timeout: 240 seconds)
00:18:31 Part toffe82
00:24:40 Quit pamaury (Remote host closed the connection)
00:30:03 Part Zagor
00:32:32 Join Scromple [0] (
00:33:05 Quit TheLemonMan (Ping timeout: 240 seconds)
00:34:15 Join advcomp2019 [0] (
00:34:15 Quit advcomp2019 (Changing host)
00:34:15 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
00:37:05 Quit advcomp2019_ (Ping timeout: 240 seconds)
00:37:49 Join FoH [0] (
00:38:52 Quit FoH_Phobos (Quit: And so, the universe ended.)
00:39:54 Quit ender` (Quit: /* Return code=1: generic error condition, Return code=2: all other error conditions */)
00:48:18 Quit lebellium (Quit: ChatZilla 0.9.87 [Firefox 6.0/20110811165603])
00:51:18 Join Poodlemastah [0] (
01:12:07 Join HaimN [0] (~quassel@
01:12:15 Quit wtachi (Quit: &)
01:15:05 Quit FoH (Quit: ¡ooʇ ‘ǝןdoǝd ǝɹɐ sʇɐq)
01:15:09 Quit MethoS- (Remote host closed the connection)
01:18:45 Quit bertrik (Ping timeout: 260 seconds)
01:19:15 Quit glued (Quit: CGI:IRC)
01:20:53 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
01:22:16 Quit HaimN (Ping timeout: 264 seconds)
01:22:37 Quit sideral (Ping timeout: 258 seconds)
01:31:22 Quit drezon (Quit: So long and thanks for all the fish)
01:33:03 Quit Poodlemastah (Quit: ChatZilla 0.9.87 [Firefox 7.0a2/20110729175641])
01:42:25 Quit mudd1 (Ping timeout: 240 seconds)
01:53:00 Quit stripwax (Quit:
01:56:41***Saving seen data "./dancer.seen"
01:58:39 Join kadoban [0] (
02:14:24 Quit Xerion (Read error: Connection reset by peer)
02:46:35 Quit GigaBrick (Ping timeout: 252 seconds)
02:59:20 Join GigaBrick [0] (
03:04:22 Quit kadoban (Read error: Operation timed out)
03:13:03 Join Scr0mple [0] (
03:15:32 Quit Scromple (Ping timeout: 252 seconds)
03:22:07 Join Horschti [0] (~Horscht@xbmc/user/horscht)
03:23:44 Quit mgue (Ping timeout: 240 seconds)
03:25:32 Join mgue [0] (
03:25:57 Quit Horscht (Ping timeout: 276 seconds)
03:30:12 Quit liar (Quit: hallowed are the ori!)
03:42:39 Quit mgue (Ping timeout: 264 seconds)
03:43:17 Join mgue [0] (
03:54:35 Quit efyx (Ping timeout: 240 seconds)
03:54:52 Join efyx [0] (
03:56:43***Saving seen data "./dancer.seen"
04:17:29 Quit amiconn (Disconnected by services)
04:17:30 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:17:50 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:18:47 Quit pixelma (Disconnected by services)
04:18:49 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:18:51 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
05:32:07 Join Rob2223 [0] (
05:35:19 Quit Rob2222 (Ping timeout: 252 seconds)
05:42:25 Quit Horschti (Quit: Verlassend)
05:56:47***Saving seen data "./dancer.seen"
07:08:59 Quit froggyman (Ping timeout: 240 seconds)
07:18:49 Join froggyman [0] (~seth@unaffiliated/froggyman)
07:19:11 Join froggyman_ [0] (~seth@
07:23:29 Quit froggyman_ (Client Quit)
07:42:38 Join kadoban [0] (
07:56:49***Saving seen data "./dancer.seen"
07:56:52 Join sideral [0] (~sideral@rockbox/developer/sideral)
08:07:10 Join mudd1 [0] (
08:09:58 Quit powell14ski (Quit: powell14ski)
08:14:14 Quit shai_ (Read error: Connection reset by peer)
08:14:43 Join shai_ [0] (
08:16:15 Join Rob2222 [0] (
08:17:49 Quit Rob2223 (Read error: Connection reset by peer)
08:18:37 Join Guest_13323 [0] (
08:19:41 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
08:22:11 Nick Guest_13323 is now known as antil33t (
08:26:06 Quit kadoban (Remote host closed the connection)
08:28:07 Quit Keripo (Read error: Connection reset by peer)
08:31:38 Join Bagder [0] (
08:31:38 Quit Bagder (Changing host)
08:31:38 Join Bagder [241] (~daniel@rockbox/developer/bagder)
08:39:37 Join ender` [0] (
08:44:41 Quit GigaBrick (Remote host closed the connection)
08:45:28 Join GigaBrick [0] (
08:58:30GigaBrickkugelp, was the "buffer_alloc(): exclusive buffer owner" panic gevaerts found with USB disconnect fixed already? Just happened for me in the current revision
08:59:45 Quit factor (Ping timeout: 240 seconds)
09:14:39GigaBrick at 15:03 is what I got... Is r30321 supposed to address that too or just on usb insert?
09:22:16 Join HaimN [0] (~quassel@
09:24:22 Join dfkt [0] (dfkt@unaffiliated/dfkt)
09:27:28 Join factor [0] (~factor@
09:28:17 Quit Scr0mple (Quit: Leaving)
09:29:22 Join einhirn [0] (
09:41:02 Quit GeekShadow (Read error: Operation timed out)
09:41:16 Join n1s [0] (~quassel@rockbox/developer/n1s)
09:41:39 Join petur [0] (~petur@rockbox/developer/petur)
09:45:39 Join GeekShadow [0] (
09:50:06 Quit God_Eater (Ping timeout: 252 seconds)
09:52:32 Quit iq_ (Read error: Operation timed out)
09:56:50***Saving seen data "./dancer.seen"
09:59:23 Join bertrik [0] (
09:59:24 Quit bertrik (Changing host)
09:59:24 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
10:05:14 Join TheLemonMan [0] (
10:10:41 Quit HaimN (Ping timeout: 252 seconds)
10:13:48 Quit GeekShadow (Ping timeout: 240 seconds)
10:16:07 Join GeekShadow [0] (
10:16:10 Quit factor (Read error: Connection reset by peer)
10:16:40kugelpGigaBrick: that one should be fixed
10:17:46GigaBrick:/ I just got it again with the latest svn
10:17:54GigaBrickThis was the USB disconnect one though
10:22:11 Join God_Eater [0] (93722cc9@rockbox/staff/GodEater)
10:23:08 Quit GeekShadow (Ping timeout: 252 seconds)
10:25:17 Join GeekShadow [0] (
10:25:40kugelpGigaBrick: my commit should fix a panic on disconnect right
10:25:56 Quit antil33t ()
10:25:58GigaBrickI'll give it another try
10:26:14GigaBrickBut when I tired the SVN tonight and disconnected I got the same panic message gevaerts posted originally
10:26:22GigaBrickThe "exclusive buffer owner" one
10:26:27 Join antil33t [0] (
10:28:21gevaertsGigaBrick: did you connect usb right after booting?
10:28:42GigaBrickYeah, i was just going to mention that I waited until the disk activity icon stopped spinning
10:28:45GigaBrickAnd then plugged it in
10:29:01gevaertshm, might be a different one then
10:29:14GigaBrickWell, I know it was right after I turned it on, but I know I also waited for the disk to settle
10:29:23GigaBrickI'll give it another shot to see if I can reproduce it
10:29:38 Quit antil33t (Client Quit)
10:30:03 Join antil33t [0] (
10:30:55 Quit user890104 (Ping timeout: 240 seconds)
10:30:57GigaBrickOkay, got it again... Rebooted after upgrading it, waited for disk to settle, put USB back in, then took it out with the backlight still on
10:33:16 Join factor [0] (~factor@
10:33:38GigaBrickGot it again...
10:36:00 Quit antil33t ()
10:36:14GigaBrickgevaerts, after I played some music it won't do it again
10:36:21 Join antil33t [0] (
10:37:12 Quit antil33t (Client Quit)
10:37:35 Join antil33t [0] (
10:41:46 Join user890104 [0] (
10:42:47GigaBrickWeirder, now I can't reproduce it at all
10:43:06GigaBrickNot sure if that's good or bad. lol
10:54:07 Join iq [0] (~iq@unaffiliated/iq)
10:54:48 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
11:02:38GigaBrickWell, I got it again, so, whatever is happening doesn't seem very consistent
11:04:55 Join pamaury [0] (
11:04:55 Quit pamaury (Changing host)
11:04:55 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
11:07:54 Join swilde [0] (
11:11:22 Quit shai_ (Read error: Connection reset by peer)
11:11:48 Join shai_ [0] (
11:15:20GigaBrickgevaerts, does it happen to you on usb disconnect at all anymore?
11:15:45gevaertsNo idea. I haven't disconnected USB with current svn
11:16:36gevaertsCan you add the patch at ? That should provide a bit more information in the panic message
11:17:19GigaBrickAll right
11:23:04GigaBrickgevaerts, I got "buffer_alloc() : scrobbler_init:251 exclusive buffer owner"
11:23:11 Join dunkaist [0] (
11:23:37GigaBrickOh, that makes sense why it stopped reproducing
11:23:50GigaBrickI turned the log off anticipating a lot of crashes
11:23:58GigaBrickThat's when the panics stopped too...
11:25:03gevaertsGigaBrick: maybe a silly question, but is the panic text wrapped?
11:25:17GigaBrickMeaning it starts on a new line?
11:25:53gevaertskugel: do you mind if I commit that patch so these panics get easier to trace?
11:26:07gevaertsOr will it mess up your tree too much?
11:26:27GigaBrickgevaerts, it wraps at "...lusive buffer owner" if that's helpful to know
11:26:28kugelgevaerts: yes
11:26:46kugelit messes up my tree, and it will be obsolete
11:26:59kugelbecause with buflib it doesn't panic anymore
11:27:05gevaertsah, ok
11:27:21*gevaerts isn't awake, clearly :)
11:27:33gevaertsI forgot. That's what buflib is *for*...
11:28:15GigaBrickWell, I was right about disabling the log making the panic stop
11:28:47GigaBrickSo I'm guessing there's a function called scrobbler_init that's to blame?
11:29:55gevaertsYes, or that area anyway
11:31:18kugelI'll have a look in a few minutes
11:32:26 Quit mudd1 (Ping timeout: 252 seconds)
11:38:02 Quit dunkaist (Quit: leaving)
11:38:40 Quit GigaBrick (Remote host closed the connection)
11:42:15 Join GigaBrick [0] (
11:43:27GigaBrickI miss anything?
11:52:27kugelgevaerts: scrobbler cache is re-allocated after usb extraction (and scrobbler is shut down on insertion)
11:52:54kugelI assume it's to free memory for usb but it doesn't make a lot of sense
11:53:10 Join FBI_Guy [0] (
11:55:22kugelbecause audiobuf isn't reset back, it still points to whatever audio it was set to. e.g. right after the scrobbler buffer if scrobbler was the last allocation
11:56:31kugelhm, I suspect the shutdown is rather to force flushing of the logs, but that's also handled already by the ata idle callbacks
11:56:53***Saving seen data "./dancer.seen"
11:59:48 Quit ruskie (Ping timeout: 240 seconds)
12:00:19*kugel would propose this patch
12:00:58kugelGigaBrick: that should fix it
12:01:07GigaBrickOkay, I'll give it a try
12:01:44 Part Zagor
12:01:55kugelgevaerts: the ata notify callbacks always work, no? or is there some trick?
12:02:00 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
12:02:03kugellike it doesn't do anything on flash
12:03:46gevaertsAs far as I know they should always work
12:04:36kugelso the patch relies on that callback to flush log to disk on usb
12:04:44kugeland stops the stupid re-allocation
12:06:03kugelmy mail to -dev didn't catch a lot of attention yet :(
12:09:08GigaBrickI don't get that panic
12:09:19GigaBrickBut now after disconnecting the usb I get "Error accessing plyalist control file"
12:09:24GigaBrickThen "Invalid control file"
12:10:27GigaBrickThen when I try to build the playlist again it says it can't update the control file
12:11:05kugelis that when you played music before usb?
12:11:49GigaBrickWell, I rebooted and can load playlists and stuff now
12:11:57GigaBrickThough looks like I ran into that database duplication thing
12:12:01GigaBrickBecause now I have double of everything
12:13:17GigaBrickYeah, I've seen that "Error accessing playlist control file" before
12:13:25GigaBrickBut it always loads after I try it again
12:13:36GigaBrickHaven't seen it say Invalid before
12:13:52kugelis that also reproducable?
12:15:47kugelI'm not sure that's caused by my recent changes
12:15:59GigaBrickI don't think so
12:16:09GigaBrickI think it's related to the issue gevaerts fixed
12:16:26GigaBrickBecause I notice it only does it when it would panic ( well, now instead of panicing it pauses for a second)
12:16:59gevaertsGigaBrick: try increasing that sleep time a bit more, maybe?
12:17:43GigaBrickI'll give that a try
12:17:51GigaBrickI have to let my database re-initalize right now though
12:17:59GigaBrickPIty, I just rated a bunch of songs :(
12:20:37sideralGigaBrick: Try exporting the DB data before regenerating it
12:20:45sideralthere's a good chance the data can be recovered
12:21:02GigaBrickYeah, I considered that only after hitting "Initialize"
12:21:06GigaBrickSo too late now heh
12:21:18sideralAlso, please log the steps that led to the duplication to FS #12129
12:21:19fs-bluebot Duplicate database entries (bugs, unconfirmed)
12:21:39sideralDid you interrupt a DB update with a reboot?
12:22:09GigaBrickNo, it was related to an error acessing the playlist control file
12:22:12FBI_Guyis there a way to copy replaygain info from file to file without replacing the entire file?
12:22:51sideralGigaBrick: interesting...
12:22:59GigaBrickYeah, I'm not really sure what happened to be honest
12:23:14sideralI don't really understand when that error can actually occur
12:23:33sideralPerhaps it's related to a dircache shutdown
12:24:01GigaBrickWe were working through an unrelated panic from scrobbler.c I got when disocnnecting from USB... When it tried to resume I got "Error accessing playlist control file", which I've seen before, but usually can just resume... THis time it said "Invalid control file" and then wiped the playlist, so then when I tried it said "Nothing to resume"
12:24:34sideral(Just a guess because Slasheri and I recently fixed two bugs related to that in the playlist and DB code; there may be more lurking around)
12:24:35GigaBrickSo then once that happened I went to my database to reload the playlist, and when I clicked I got the message "~9000 found" and knew it was dup'ing because my collection is only about 5000 tracks
12:25:49sideralWhen that happens next time, could you please check the database and dircache status pages in the debug menu?
12:26:08GigaBrickThe "Error accessing control file" is fairly reproducable though
12:26:13GigaBrickSo I can probably check that right now
12:26:15 Join metaphys [0] (
12:26:49sideralDo you have DB auto-update enabled, BTW?
12:26:56LloreanFBI_Guy: I'm not even sure what you think you're asking, honestly. Replaygain info is unique to either the track or album anyway, and it's just metadata. I can't even imagine why you'd need to "replace" a file anywhere in the replaygain process.
12:27:01gevaertsGigaBrick: try checking the filesystem. You've had so many panics recently that I'm not entirely convinced that that's still going to be clean
12:27:12GigaBrickYeah, sideral
12:27:38GigaBrickGood idea, gevaerts
12:28:11GigaBricksideral, I've got the dircache debug page open
12:28:19GigaBrickAfter getting the "Error accessing playlist control file" splash
12:28:29sideralis the dircache still enabled?
12:28:53GigaBrickSays "Cache initialized: Yes" if that means yes
12:29:01sideralyeah, that's what I meant
12:30:10sideraland what does the DB debug menu say?
12:30:21sideralfirst three lines?
12:30:38GigaBrick"Intialized: yes, DB Ready: Yes, RAM Cache: Yes"
12:31:01GigaBrickAlso, under progress it says 87% (9145 entries)
12:31:10GigaBrickDoes that mean it duped again?
12:31:43sideralNo idea −− you'll have to check in the DB browser
12:31:44GigaBrickOr maybe that's all still loaded in RAM from before I re-initialized?
12:32:07Slasherifound entries does not correlate to the number of tracks
12:32:07GigaBrickNah, doesn't seem like it duped again
12:32:27Slasheriit includes every file and directory on the player
12:32:40GigaBrickOh, okay
12:32:51sideralLooks like everything is fine and dandy this time. But you didn't add/remove files over USB this time around, right?
12:32:55GigaBrickJust coincidence it happened to be almost twice my collection
12:32:55 Join JacekNiesobski [0] (
12:32:58Slasheri(maybe a good idea to change it to include just tracks)
12:33:23GigaBrickNope, sideral, just connected then disconnected
12:33:36JacekNiesobskihello everyone
12:34:00GigaBrickgevaerts, looks like you were right
12:34:12GigaBrickJust fsck'd and ./.rockbox/.playlist_control got flagged
12:34:22sideralGigaBrick: So the playlist error is probably unrelated to the DB problem −− if there ever was any (you did observe dupes in the DB browser, right? or was it just the 9000 that confused you?)
12:34:37GigaBrickNo, I definitely saw the dupes in the database
12:34:49GigaBrickBut I think the playlist thing is unreleated since I just found the filesystem dirty
12:35:17JacekNiesobskican I ask to be added to the WikiUsersGroup?
12:36:33sideralGigaBrick: Slasheri and I have a theory about the dupe issue (double commit into a dirty database). I now have a filesystem copy from someone who ran into it (and also FS #12228, which may be related), and hope to be able to reproduce the issue now.
12:36:34fs-bluebot Database fails to commit on Fuze v2 fresh SVN build (bugs, unconfirmed)
12:37:36 Join MethoS- [0] (~clemens@
12:37:37GigaBrickHaven't ran into anything like 12228
12:37:50GigaBrickAnything helpful I can do if I get the duplication again?
12:37:54GigaBrickLIke specifically what to note down?
12:38:01FBI_GuyLlorean: sorry for late response :P, but i mean is there a way to copy the replaygain metadata so i dont have to replace the songs i already have on my player which dont have replaygain info with the ones from my computer which now do
12:38:08sideralGigaBrick: But if you have a recipe for reproducing it, I'd still be grateful if you posted it to FS #12129
12:38:09fs-bluebot Duplicate database entries (bugs, unconfirmed)
12:38:13 Quit JacekNiesobski (Quit: CGI:IRC)
12:38:36sideralGigaBrick: Yeah, try to remember the exact steps that lead to it, and attach a copy of your DB to the tracker item
12:38:47GigaBrickOkay, will do if it happens again
12:38:50sideralthat is, all /.rockbox/database*.* files
12:39:02sideralcool, thanks!
12:39:20GigaBrickYeah, this time was so mixed with other things I'm not really sure it'd be any help to recall exactly what happened
12:39:42GigaBrickOr if I really could
12:39:52sideralSlasheri: Have you seen the dircache-related "crazy idea" session I tortured everyone with two nights ago?
12:40:34GigaBrickI think I read through that
12:40:39GigaBrickPlease don't nix the db, <3 the db
12:41:01sideralSlasheri: Might be an entertaining read:
12:41:37Slasherisideral: hehe, i read that =)
12:42:54Slasherisideral: in fact, i thought about block cache implementation before implementing the dircache :)
12:43:13sideralGigaBrick: No one plans to remove it, no worries :)
12:43:45GigaBrickHeh, okay, just had to get that out there since there doesn't seem to be a lot of love for it
12:44:15sideralYeah, it's probably the best solution. I do worry about the complexity of its interface for clients, though, especially when the dircache shuts down
12:45:14sideralPerhaps an alternative idea would be to hide the dircache behind the filesystem interface and just offer an interface for retrieving the filename of an "inode" (aka dircache ID)
12:45:31 Join ruskie [0] (ruskie@sourcemage/mage/ruskie)
12:45:34 Join Rob2223 [0] (
12:47:34GigaBrickJust reproduced the db duplication I had earlier
12:48:39sideralGigaBrick: dang, there goes my after-lunch nap ;)
12:48:48GigaBrickDisconnected USB, had the "Error accessing playlist control file" splash, tried to resume again anyway, got "Invalid control file" splash so then I went to load a new playlist and got "Error updating control playlist" and the player hung up hard
12:48:53GigaBrickWould not retrun to main menu or even power down
12:48:56GigaBrickSo I had to flip the battery
12:49:04 Quit Rob2222 (Ping timeout: 252 seconds)
12:49:06GigaBrickLoaded the device on usb and checked my filesystem to make sure that wasn't it again
12:49:23GigaBrickFilsystem was clean, so I started it again, the playlist contro file loaded the playlist I had just tried and started playing
12:49:26LloreanFBI_Guy: Whatever tool scanned them on your PC should also be able to scan them on your player.
12:49:30GigaBrickBut then when I went to check the database, duplicates
12:49:45GigaBrickOne thing I noticed that when it was hung, the disk was doing an awful lot
12:50:07FBI_GuyLlorean: yeah but that takes forever :P (sorry if this is the wrong place ro be asking)
12:50:50sideralGigaBrick: OK, please add a comment containing the steps and a copy of your DB to the tracker task
12:51:10LloreanFBI_Guy: Typically tagging tools are not designed to identify duplicate songs and match the tags between them. So you'll either need to copy them over, or re-scan
12:51:13sideralah, before you shut down
12:51:35sideralcan you check in the FS browser whether everything looks OK?
12:51:36GigaBrickYeah, I've got the .rockbox dir open
12:52:09FBI_GuyLlorean: alright, thanks, I guess ill just scan again
12:53:11sideralGigaBrick: enable "view all" in the file browser, and look out for strange file or dir names in the root directory
12:53:41 Join wodz [0] (
12:53:44GigaBrickEverything looks right
12:54:04sideralOK, that rules out the dircache corruption issue we've had a few revisions ago
12:54:26sideralwhich could lead to loops in the dir hierarchy, which in turn lead to duplicates
12:54:42GigaBrickHmm, interesting....
12:54:49GigaBrickI copied my db before checking the files
12:54:57GigaBrickSO when I disconnected I got "Error accessing playlist control file" again
12:55:15GigaBrickI thought I'd try to reproduce the whole thing again just for kicks and loaded a playlist, but it worked fine this time
12:55:32GigaBrickOnly difference I can see is that before I inserted the USB while music was playing, this time it was inserted from the main menuy
12:55:34 Quit metaphys (Quit: CGI:IRC (Ping timeout))
12:56:35sideralthat's probably an unrelated issue, but worth mentioning in the FS comment anyway
12:56:56sideralMaybe you should report a separate bug for this behavior as well
12:57:06GigaBrickIn fact I ran into the duplication while trying to figure out why the playlist control file was doing that
12:57:18GigaBrickI know it's related to a panic that gevaerts fixed
12:57:47GigaBrickHad to add a 100 ms delay before the disk mounted again after leaving usb mode, or else it would panic because nothing was mounted
12:58:02GigaBrickWell, the only times I get the "Error accessing control playlist file" are times that it would otherwise panic without that sleep
12:58:08GigaBrickSo there's something else going on there
12:58:35sideralsound like kind of a crude workaround
12:59:02GigaBrickHeh, can't comment, it stops it from panic'ing at least
12:59:27GigaBrickBut I think a similar thing is happening where the playlist is trying to be resumed but there's problems acessing the disk
13:00:01GigaBrickMainly basing that off the fact that I only get the playlist control access error on disconnects that would have caused the panic, so it seems fairly related
13:01:19*sideral need to go offline now
13:01:27sideralI'll check FS later
13:01:39sideralthanks GigaBrick for helping to nail this one down!
13:02:04GigaBrickYou still want me to add those steps I took to the FS right?
13:02:22sideralyes, please
13:02:59GigaBrickOkay... Still have the link in your clipboard, having trouble finding it in the history
13:04:42GigaBrickNvm, found it
13:07:20 Join HaimN_ [0] (~quassel@
13:08:26kugel"Perhaps an alternative idea would be to hide the dircache behind the filesystem interface and just offer an interface for retrieving the filename of an "inode" (aka dircache ID)
13:08:46kugel" <−− how is this different from now?
13:08:51kugelsideral: ^
13:09:24kugelthis is exactly how dircache currently works isn't it?
13:09:35Tornepretty much :)
13:10:56kugeldircache is completely transparent behind opendir and friends, additionally one can exploit the cache by getting dircache IDs and from that the filename
13:11:16*kugel must be missing something
13:16:18GigaBrickUgh, my forgetfulness is making a mockery of this FS page...
13:22:08 Join XavierGr [0] (~xavier@rockbox/staff/XavierGr)
13:22:16GigaBrickkugel, is there anything you can think of in the scrobbler patch you made that would cause the database to rebuild itself like this?
13:23:46GigaBrickI wonder why I keep getting the "Error accessing playlist control file"
13:24:14GigaBrickBecause it seems to all happen after that. Once I click that flash away I get "Invalid control file" and then the splash that pops up when you update the database saying 9000 something found...
13:24:39GigaBrickWeird difference this time though... When I powered down and checked the database for duplicates, instead of finding duplicates it asked to initialize it
13:27:01 Join CoLdAsSauLt [0] (
13:27:15CoLdAsSauLtcan I please get some help on patching Rockbox?
13:27:34CoLdAsSauLtI installed GnuWin32...
13:27:35CIA-14New commit by kugel (r30325): Fix panic after usb extraction if lastfm logging is enabled. ...
13:27:38CoLdAsSauLtbut now what ^^
13:27:58GigaBrickOkay, this is getting really weird now... I just picked up the gigabeat after hitting Initiali database to rebuild my database
13:28:09GigaBrickI have the "Error accessing playlist control file" splash
13:29:10GigaBrickkugel, I'm going to try reverting to my 30295 build and apply the scrobble patch and see if I still get this behavior, does that sound like it will be helpful at all in determining what might be causing it?
13:29:20kugelGigaBrick: does this happen automatically or do you try tor esume playback?
13:29:29GigaBrickThis appeared to happen automatically
13:29:37GigaBrickI was still in the settings menu for database
13:29:46kugelvery strange
13:29:59CIA-14r30325 build result: All green
13:30:55kugelnormally the playlist control file is only accessed when you stop or resume playback, or when you edit the playlist in the playlist viewer
13:30:56 Quit merbanan (Ping timeout: 258 seconds)
13:31:21GigaBrickWould the system automatically try to resume playback?
13:31:37GigaBrickI mean, that's the first thing I thought of but it seems strange it would do that from the settings menu
13:32:34kugelno it shouldn't do that AFAIK
13:33:58 Join merbanan [0] (
13:34:10CoLdAsSauLtsomebody willing to help me out? Sorry for my lack of technical knowledge :s
13:35:22BagderCoLdAsSauLt: you probably want to build rockbox after you've pathed it, so follow the docs to setup a dev environment
13:36:11 Quit HaimN_ (Remote host closed the connection)
13:38:54GigaBrickkugel, I just patched 30295 with your scrobbler patch and I'm not seeing the same behavior
13:39:28GigaBrickNo error on playlist control file access or duplication or anything
13:43:30 Quit merbanan (Ping timeout: 276 seconds)
13:47:06 Join merbanan [0] (
13:50:05kugelGigaBrick: what kind of playlist did you have?
13:50:11kugelplaying an m3u by chance?
13:50:25GigaBrickYeah, this time around
13:50:34GigaBrickBut most of the time I have a dynamic playlist
13:50:41kugelI see a possible code path for your problem
13:51:00GigaBrickDoes it still apply if I get this behavior when using dynamic playlists too?
13:51:05kugelcan you repro this with a dynamic playlist?
13:51:28pamaurygevaerts: I thought we had a define for player with no high speed usb
13:51:42GigaBrickkugel, yeah, happens with either type
13:52:24pamauryI found USB_NO_HIGH_SPEED but it seems to be unused
13:52:45kugelGigaBrick: can you check if the playlist control thread is active (in view OS stacks debug menu)?
13:53:09GigaBrickAfter I repro or just right now?
13:53:17gevaertspamaury: yes, I don't think there's more. That was basically debug code
13:53:38pamauryso all players have high speed usb or hardware usb bridge ?
13:54:08gevaertsAs far as I know, yes
13:54:16gevaertsWell, up to now anyway
13:54:31kugelGigaBrick: after you repro
13:54:35gevaertsDo we care actually, outside of the driver?
13:55:00GigaBrickOkay, I'll have to put the current revision back on
13:55:20pamaurygevaerts: in the core, for the descriptors
13:55:47pamaurywe report as 1.1 usb when we don't have high speed
13:55:58pamaurybut that's not mandatory iirc
13:56:09gevaertsAh, yes, the other speed config
13:56:57***Saving seen data "./dancer.seen"
13:57:28gevaertsI'd say this is something to do when we actually have a target where it matters
13:58:19pamauryI have a FS only device
13:58:29gevaertsAh, you have some work to do then :)
13:59:05pamauryI'm not sure what needs to change,the only use of USB_NO_HIGH_SPEED us the bcdUSB field andI think you can report 2.00 for a FS device
13:59:55gevaertsI'm not sure either. This is a *long* time ago...
14:00:31pamauryfound: If a full-speed only device (with a device descriptor version number equal to 0200H) receives a
14:00:31pamauryGetDescriptor() request for a device_qualifier, it must respond with a request error. The host must not make
14:00:31pamaurya request for an other_speed_configuration descriptor unless it first successfully retrieves the
14:00:31DBUGEnqueued KICK pamaury
14:00:31pamaurydevice_qualifier descriptor.
14:01:00gevaertsright, I think we don't do that yet
14:01:21gevaertsI suspect that's all there is to it
14:03:19pamauryso in fact we could probably keep the current code, assume the host don't request device_qualifier and always use 2.00 version. In fact there is not need to disctinction
14:07:00pamauryExcept for packet size: bulk is 64 at FS
14:07:19pamaurybut that's already handled anyway
14:08:21GigaBrickkugel, I got the "Erorr accessing playlist control file" to reproduce
14:08:33GigaBrickChecked the playlistcachectrl thread and didn't see much of anything
14:08:41GigaBrickChanged from T to something, but went so fast I didn't see what it changed to
14:12:42GigaBrickWeird, now it's showing the splash while the playlist is actually playing on the WPS screen
14:12:51GigaBrickAnd now it's showing it in the debug menu when I went to check on that thread
14:13:19GigaBrickExcept now it's doing something weird where it shows up briefly and disappears... :/
14:14:02gevaertspamaury: the host will ask for the device_qualifier. If it does, though, the worst that happens is that windows tells you that a different port might be faster. Returning an error there (for FS-only devices) is probably a good idea in the long term
14:16:23pamauryokay, I'll do that
14:17:16 Join Rob2222 [0] (
14:17:33kugelGigaBrick: I just wanted to know if it's there
14:17:43kugelcan you disable dircache and see if it persists?
14:18:02GigaBrickYeah, I'll try that
14:20:14GigaBrickWhere is that setting located again?
14:20:46GigaBrickOh, under Disk, nvm
14:21:03 Quit Rob2223 (Ping timeout: 250 seconds)
14:23:33GigaBrickkugel, only got it to reproduce once with dir cache off, just disconnected 5 times in a row now without the playlist control splash
14:23:56kugeldid you reboot?
14:24:20GigaBrickNo, should I?
14:25:02kugelyou can try
14:25:29GigaBrickOkay, I'll reboot and give it a try too
14:26:20GigaBrickYeah, seems to eliminate that splash when dir cache is off
14:27:01kugelthanks for your debugging, I'll have a look later
14:27:37GigaBrickAlso, I thought I should mention
14:27:56GigaBrickI applied the sleep patch and the scrobbler patch to 30307
14:27:58kugelthe playlist control thread tries to fetch dircache ids in the background for the filenames, that seems to fail
14:28:05GigaBrickSTill got the playlist splash
14:28:10GigaBrickBut for some reason in 30295 I don't
14:28:17kugeloh interesting
14:29:22GigaBrickYeah, I haven't seen what's the highest revision I can go to yet without reproducing, but I figured I'd try that and that might answer something
14:29:27kugelnot sure what commit can cause that then
14:30:07kugelyou did try bumping the sleep time higher in the sleep patch, didn't you?
14:30:22GigaBrickYeah, I had it up to a full second and it didn't make a difference
14:30:27GigaBrickDidn't know if I should try any higher though
14:30:39GigaBrickWell, a full HZ (I'm assuming that's a second)
14:31:26kugelthat's a second yes
14:31:56GigaBrickHertz is rolling in his grave at my unsuredness. :P
14:32:52GigaBrickAnyway, I suppose I can go up the revisions incrementally from 30295 until I find which one starts reproducing the behavior
14:33:06kugelthat would be fine :)
14:33:46GigaBrickOh, one quesiton about svn updating
14:33:57GigaBrickHow can I get it to actually replace a file that has custom modifications?
14:34:35kugelnot at all
14:34:40kugelyou need to undo the modifications
14:34:47GigaBrickOh, all right
14:34:49kugele.g. svn revert
14:35:22GigaBrickOh, come to think of it, is there a way to revert back to a specific revision?
14:35:32GigaBrickI should probably just man svn...
14:35:56kugelsvn up -rXXXX
14:37:31GigaBrickOh, cool, it takes care of the changed files too
14:37:37GigaBrickAll right, that will make this go a lot faster
14:38:29kugelit doesn't touch modifications. it tries to merge and abort the update if it doesn't manage
14:38:56GigaBrickHmm... It just replaced my patched ./firmware/usb.c with the one from the svn
14:39:50kugelwell, it would be news to me if it does that. but I haven't used svn in a long time
14:40:05GigaBrickYeah, I have barely ever used it
15:18:36GigaBrickkugel, I must have done something wrong before, I just tried 30307 and I didn't get the playlist splash
15:21:41 Quit CoLdAsSauLt (Quit: CGI:IRC (Ping timeout))
15:46:01 Join FBIGuy [0] (
15:46:01 Quit FBI_Guy (Read error: Connection reset by peer)
15:49:31 Join ntkm [0] (
15:50:44 Quit ntkm (Client Quit)
15:56:21 Join CoLdAsSauLt [0] (
15:56:34CoLdAsSauLtDo I really have to rebuild rockbox, just to apply a patch?
15:56:53Bagderno, but you won't get a rockbox to install unless you build it...
15:56:58***Saving seen data "./dancer.seen"
15:57:50CoLdAsSauLtmm, so the way to go is
15:58:00CoLdAsSauLtinstalling linux
15:58:08CoLdAsSauLtwhich I have already done, virtualbox)
15:58:42CoLdAsSauLtand rebuilding it there
15:58:54CoLdAsSauLtpatching an already installed rockbox isn't possible?
15:59:02CoLdAsSauLtI 'm running it on a Cowon D2)
15:59:19Bagderpatches are little changes to source code
15:59:39Bagderor sometimes big changes...
16:00:23Bagderbut still source code, and once you have applied a patch, you have modified the source code, and then you want that modified source code to get turned into a rockbox install for you
16:01:03CoLdAsSauLtI'm a music lover
16:01:11CoLdAsSauLtbut nowhere a programmer :)
16:01:30CoLdAsSauLtso it's kind of overwhelming
16:01:50CoLdAsSauLtwhen reading through the patch pages
16:02:22 Quit antil33t (Read error: Connection reset by peer)
16:02:42 Join antil33t [0] (
16:03:47 Quit shai_ (Ping timeout: 240 seconds)
16:22:22 Join mshathlonxp [0] (
16:38:16 Quit XavierGr (Read error: Connection reset by peer)
16:38:40 Join XavierGr [0] (~xavier@rockbox/staff/XavierGr)
16:39:26 Quit Rob2222 (Quit: Rob2222)
16:41:37 Quit GigaBrick (Ping timeout: 240 seconds)
16:56:00 Join GigaBrick [0] (
17:01:46 Quit petur (Quit: *plop*)
17:04:21wodzIs buffer passed to sd_read_sectors() alligned?
17:08:56gevaertsNot in general
17:09:35gevaertsPerformance-critical code does align it, but that's not what you mean I guess
17:09:59 Quit Bagder (Quit: Konversation terminated!)
17:11:41wodzeven word aligned on ARM?
17:12:59gevaertshm, I'm not sure actually now. It depends on what the FAT code does
17:18:31wodzHow to build single plugin (or only test plugins)?
17:19:32 Quit antil33t (Read error: Connection reset by peer)
17:19:52 Join antil33t [0] (
17:19:57 Join petur [0] (~petur@rockbox/developer/petur)
17:20:50 Part Zagor
17:49:32 Join Stummi [0] (~Stummi@rockbox/developer/Stummi)
17:56:59***Saving seen data "./dancer.seen"
17:57:29 Join y4n [0] (y4n@unaffiliated/y4ndexx)
18:00:15 Join saratoga [0] (9803ec71@rockbox/developer/saratoga)
18:07:04 Quit wodz (Quit: Leaving)
18:10:56 Quit Stummi (Quit: Bye!)
18:11:09 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
18:11:18 Quit petur (Quit: *plop*)
18:12:40 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
18:13:17 Quit user890104 (Read error: Connection reset by peer)
18:14:31 Join user890104 [0] (
18:18:06 Join mystica555 [0] (
18:23:37 Quit mystica555 (Ping timeout: 276 seconds)
18:28:09 Join mystica555 [0] (
18:32:46 Join T44 [0] (
18:36:28 Quit Topy (Ping timeout: 250 seconds)
18:41:30 Quit mystica555 (Ping timeout: 264 seconds)
18:43:56 Join dunkaist [0] (
18:45:53 Quit dunkaist (Client Quit)
18:51:26 Join Amadiro [0] (
18:51:39 Join lorenzo92 [0] (
18:51:41 Join Rob2222 [0] (
18:51:54 Quit lorenzo92 (Client Quit)
18:52:48 Quit swilde (Remote host closed the connection)
19:01:33 Quit TheLemonMan (Quit: .)
19:06:15 Quit XavierGr (Read error: Connection reset by peer)
19:06:24 Join XavierGr [0] (~xavier@rockbox/staff/XavierGr)
19:10:42 Quit XavierGr (Ping timeout: 250 seconds)
19:10:50 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
19:12:50 Join dunkaist [0] (~dunkaist@
19:37:18 Quit bluebrother (Disconnected by services)
19:37:20 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
19:39:05 Join Horscht [0] (
19:39:05 Quit Horscht (Changing host)
19:39:05 Join Horscht [0] (~Horscht@xbmc/user/horscht)
19:40:36 Quit fs-bluebot (Ping timeout: 250 seconds)
19:41:43 Join fs-bluebot [0] (
19:41:54 Join lixxus [0] (
19:42:14lixxusanyone interested in buying a toshiba gigbeat s 30gb ?
19:42:49 Join stoffel [0] (
19:50:39 Join robin0800 [0] (
19:51:49 Join wodz [0] (
19:57:01***Saving seen data "./dancer.seen"
19:59:40wodzI get Data abort at the address of the plugin__start when loading plugin - what may cause this?
20:02:25 Quit Horscht (Ping timeout: 276 seconds)
20:04:49 Join Horscht [0] (
20:04:49 Quit Horscht (Changing host)
20:04:49 Join Horscht [0] (~Horscht@xbmc/user/horscht)
20:06:49 Join XavierGr [0] (~xavier@rockbox/staff/XavierGr)
20:07:00 Join ChickeNES [0] (
20:09:48 Join ibhan [0] (
20:10:04 Quit ibhan (Remote host closed the connection)
20:10:43 Join ibhan [0] (
20:10:43 Quit antil33t (Read error: No route to host)
20:11:26 Join antil33t [0] (
20:11:32 Quit ibhan (Remote host closed the connection)
20:19:06 Quit user890104 (Read error: Connection reset by peer)
20:20:14 Join user890104 [0] (
20:37:47 Join [fred] [0] (
20:54:13 Quit wodz (Quit: Leaving)
20:54:35 Quit knittl (Ping timeout: 260 seconds)
20:54:35 Quit Staphylo (Ping timeout: 260 seconds)
20:54:36 Join kadoban [0] (
20:55:07 Join knittl [0] (~knittl@unaffiliated/knittl)
20:56:33 Quit stoffel (Remote host closed the connection)
21:03:34 Quit kadoban (Remote host closed the connection)
21:07:17 Join Staphylo [0] (
21:09:01 Quit sideral (Remote host closed the connection)
21:10:22 Join wodz [0] (
21:12:39 Join sideral [0] (~sideral@rockbox/developer/sideral)
21:16:48 Join stoffel [0] (
21:20:37 Join mudd1 [0] (
21:23:11bluebroth3rniiiice :)
21:23:26CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
21:23:26*bluebroth3r stumbled across the LaTeX silence package:
21:29:57bluebroth3rargh, had some other output filter active. Doesn't work the way I thought :(
21:32:59wodz\o/ - I was able to run test_codec on rk27xx :-)
21:36:54wodzso, arm7ej-s gives 51.29MHz for realtime to decode mp3 cbr 128
21:38:36 Join liar [0] (
21:43:34 Quit stoffel (Remote host closed the connection)
21:44:30 Join Buschel [0] (
21:45:57Buschelwhich is quite a lot to decode mp3...
21:46:01Buschelslow RAM?
21:46:34BuschelS5L8701 (ARM940T) shows similar results formp3
21:49:38 Join Strife89 [0] (~Strife89@
21:54:27 Quit factor (Read error: Connection reset by peer)
21:57:04***Saving seen data "./dancer.seen"
22:10:42 Join factor [0] (~factor@
22:12:05 Quit antil33t (Ping timeout: 240 seconds)
22:12:31 Join antil33t [0] (
22:12:45 Quit XavierGr ()
22:13:21*pamaury curses the usb controller of the s3c2440
22:15:44pamaurythe fucking manual sounds like it was written by someone who doesn't know how does it work and which features is has !
22:16:02pamaury(the s3c2440 manual)
22:17:09 Part lixxus
22:22:19wodzBuschel: test_mem says read ~100MB/s, write ~51MB/s, memset ~50MB/s, memcpy 48MB/s
22:23:02wodzI don't have any target to compare this results
22:23:27Buschelis this for IRAM?
22:24:15wodziram is (almost) useless as it is 4kB or so
22:24:20BuschelS5L8701 DRAM = 59 read, 35 write, 35 set, 22 copy
22:24:43Buschelyour numbers are similar to S5L8701 IRAM
22:25:42bertrikBuschel, maybe it's just misconfigured then?
22:25:54Buschelyep, maybe...
22:26:10wodzhmm, so why mp3 decoding takes so much may it be (extremely) slow lcd updates?
22:27:08Buschelon S5L8701 I can sqeeze a bit more performance from my nano 2G. but this might most likely impact other nano's... for the rk27xx it might be interesting to check the RAM timings
22:28:05Buschelwodz, if the memory is that slow the filterbank will be slow as well. I think the codec measurement is fine
22:28:56Buschelwodz, I bet you'll see similar results if you do not update the screen during the test_codec run
22:29:36wodzyou said s5l8701 dram read 59 vs. rk27xx dram read 100 - I don't get
22:30:49Buschelmost codecs use IRAM for the performance critical parts. IRAM S5L8701 = 107 read, 80 write/set, 45 copy
22:31:54Buschelso, a bit faster than rk27xx
22:36:18 Quit FBIGuy (Quit: I'm outta here)
22:39:40Buschelcan you configure to use a higher bus clock for the DRAM?
22:40:44wodzI can't remember the config
22:42:38 Quit benedikt93 (Quit: Bye ;))
22:46:57wodzsd reads aren't fast also
22:47:30Buschelwhat is the CPU clock?
22:47:36wodz200 MHz
22:47:52Buschelbus clock is 0.5 * CPU clock?
22:48:07wodzI think so
22:48:21wodzand APB is Fcpu/4
22:49:09wodzthis particular version of the SoC can run at 240MHz
22:49:44 Join Keripo [0] (
22:50:56Buschelif you do not change the ration fcpu : ahb you should not be able to reach higher codec efficiency (like 53 MHz) but only a higher realtime factor
22:51:29 Quit dunkaist (Quit: leaving)
22:51:52Buschelyou might for example run test_codec with CPU and AHB at 133 MHz
22:52:21Buscheldou you already have clock scaling?
22:54:34Buschelyou will most likely see that test_codec will result in higher efficiency when CPU:AHB = 1:1. this way the CPU could be clocked to lets say 50 MHZ and be able to play mp3 in realtime.
22:57:27Buschelwhat is SCU_DIVCON1 set to?
22:57:49 Join FBI_Guy [0] (
22:58:12wodzthe same as reset value stated in manual
23:05:21wodzeh, I fucked up filesystem on SD playing with sd_write_sectors(). Time to sleep() apparently.
23:05:57 Quit wodz (Quit: Leaving)
23:06:42 Quit FBI_Guy (Quit: I'm outta here)
23:07:01 Join FBI_Guy [0] (
23:37:43 Join TheLemonMan [0] (
23:40:36 Quit n1s (Remote host closed the connection)
23:48:31CIA-14New commit by buschel (r30326): Reduce memory consumption of VGM codec for low memry targets at the costs of some performance for tracks using the 2616 emulator.
23:51:28CIA-14r30326 build result: All green
23:57:05***Saving seen data "./dancer.seen"
23:57:16 Quit ender` (Quit: You don't have to think too hard when you talk to teachers. -- J. D. Salinger)
23:57:54 Part Strife89 ("Vamoose!")

Previous day | Next day