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 2011-09-04

00:00:04akaWolfplayers with 7 buttons are?
00:00:29akaWolf7 hard buttons
00:00:36 Quit linuxguy3 (Ping timeout: 276 seconds)
00:00:39LloreanAre you asking which players have 7 buttons?
00:00:51 Join linuxguy3 [0] (
00:01:01LloreanYou'll have to do a little research for that. I don't think anyone will know off the top of their head.
00:01:06LloreanRockbox runs on a decent amount of hardware.
00:01:28saratogai think most have more then that, aside from a few touch players
00:01:50n1sipods if you count spinning the wheel
00:02:14LloreanThe 7 buttons are often in very different configurations.
00:03:11saratogasomeone reported that the new VM i setup worked fine for them so i'm putting it in the dev guide
00:03:30saratogaits smaller then the sourceforge one, but also includes the arm and sdl compiler bits
00:03:35n1ssaratoga: sounds good
00:03:48saratogayou still need to run rockboxdev if you want coldfire or sh
00:03:55saratogasorry i forgot about this, just so busy this month
00:04:31pamaury_weird, it looks like sometimes button_get_w_tmo won't pass control to other threads :-/
00:04:54 Quit ender` (Read error: Connection reset by peer)
00:05:05n1spamaury_: is that recent?
00:05:29 Join ender` [0] (
00:05:30n1si've noticed my h300 hanging while navigating the filebrowser quite often recently
00:06:49 Join JasonT [0] (
00:07:01pamaury_no idea, I'm in the fuze+ bootloader so perhaps I'm missing some init bit
00:07:20pamaury_or perhaps there is a timing problem and the usb thread is not ready at this point
00:07:31pamaury_or perhaps someone broke it :)
00:07:50 Quit JasonT (Client Quit)
00:09:13pamaury_I'm betting on the second option for my problem
00:10:25 Nick pamaury_ is now known as pamaury (
00:10:34n1sthe gigabeast bootloader does sw usb, maybe you can see what it does as iirc there's some stuff that is needed that bootloaders dion't normally do
00:10:36 Quit pamaury (Changing host)
00:10:37 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
00:11:02pamauryI had a very close look at it and I'm doing pretty much the same
00:11:46pamaurythe thing is that the problem happens randomly, effects ranges from no usb detection, late usb thread creation to data abort :-/
00:16:02 Quit Keripo (Quit: Leaving.)
00:21:17pamauryhum, looks to me that usb_init has an inherent race condition
00:24:48 Quit petur (Quit: Leaving)
00:25:22pamauryjhMikeS: ping
00:29:01 Join Jerom1 [0] (~jerome@
00:29:08 Join Keripo [0] (
00:31:44 Quit n1s (Remote host closed the connection)
00:32:13pamauryhum, perhaps that's because there is no main thread in the bootloader...
00:37:00***Saving seen data "./dancer.seen"
00:47:25 Quit pamaury (Remote host closed the connection)
00:49:31 Quit Jerom1 (Quit: Leaving.)
01:03:04 Quit v4hn (Quit: leaving)
01:07:09 Quit ender` (Quit: The reason people use a crucifix against vampires is that vampires are allergic to bullshit. -- Richard Pryor)
01:09:40 Join stripwax [0] (
01:10:02 Quit stripwax (Client Quit)
01:13:22 Quit guymann (Ping timeout: 245 seconds)
01:17:43CIA-14New commit by buschel (r30424): FS #12259: Remove '-w' compiler option for MikMod. Resolve all yet unreported compiler warnings and fix a bug in load_gt2.
01:20:09CIA-14r30424 build result: 1 errors, 3892 warnings (buschel committed)
01:21:35*Buschel loves that his build environment does not really match our build farms...
01:34:39Buschelanyone awake who can reproduce the build issues?
01:39:13 Join ingoldey [0] (~quassel@2002:51dd:9133:0:21a:a0ff:fe50:5977)
01:42:20ingoldeyhave a ipod nano 2g and a linux with "RockboxUtility"
01:42:51CIA-14New commit by buschel (r30425): Fix yellow and red.
01:45:16ingoldeybefore i use the ipod as usbstick whit an ext2 partition
01:46:09CIA-14r30425 build result: 0 errors, 5 warnings (buschel committed)
01:47:18ingoldeynow can not write the bootloader to it.
01:49:06ingoldeydo i have to "reinitialisation" with an iTunes before i can insatll RockBox? (I have no iTunes on Linux)
01:51:15mc2739ingoldey: you might be able to restore it using the directions here:
01:54:44CIA-14New commit by buschel (r30426): Remove several 'set but not used' GCC 4.6.1 warnings from MikMod.
01:54:53ingoldeymc2739: Thank you, i think i can do this. (but)
01:55:36 Join ReimuHakurei [0] (
01:56:27ingoldeymc2739: Do you think i can over to step 7? an reformating to fat32 directly?
01:57:57CIA-14r30426 build result: All green
02:00:58mc2739ingoldey: I think you need to follow all steps - the apple firmware needs to work before you can install Rockbox
02:06:05 Quit Buschel (Quit: ChatZilla 0.9.87 [Firefox 6.0.1/20110830092941])
02:08:19ingoldeyarg, I fail on step "one" there is no MBR file for my ipod nano 8GB
02:10:00ingoldeyi think it is easyer for me to go the "itunes" way.
02:10:32ingoldeymc2739: Tank you, and gn
02:17:11 Quit bertrik_ (Ping timeout: 260 seconds)
02:19:15 Join funman [0] (~fun@rockbox/developer/funman)
02:20:04 Nick madskiny is now known as dre (dre@gateway/shell/
02:24:40 Quit funman (Quit: leaving)
02:37:02***Saving seen data "./dancer.seen"
02:57:20 Quit bertrik (Ping timeout: 264 seconds)
03:06:30 Join ingoldey_ [0] (~quassel@2002:51dd:ce91:0:21a:a0ff:fe50:5977)
03:07:56 Quit ingoldey (Ping timeout: 260 seconds)
03:14:43 Quit Keripo (Quit: Leaving.)
03:14:48 Quit liar (Ping timeout: 258 seconds)
03:15:38 Join liar [0] (
03:37:25 Quit ReimuHakurei (Ping timeout: 258 seconds)
03:47:22 Join ReimuHakurei [0] (
03:58:42 Join soap_ [0] (~soap@rockbox/staff/soap)
04:01:51 Part nick-p ("Leaving")
04:07:17 Quit pixelma (Disconnected by services)
04:07:17 Quit amiconn (Disconnected by services)
04:07:18 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:07:19 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:07:21 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:07:32 Quit ReimuHakurei (Ping timeout: 260 seconds)
04:07:37 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:28:03 Quit akaWolf (Ping timeout: 268 seconds)
04:37:05***Saving seen data "./dancer.seen"
04:40:57 Quit TheSeven (Disconnected by services)
04:41:19 Join [7] [0] (~TheSeven@rockbox/developer/TheSeven)
04:45:31 Join ReimuHakurei [0] (
04:49:20 Quit robin0800 (Ping timeout: 258 seconds)
04:50:05 Join soap__ [0] (
04:56:28 Join white_magic [0] (white_magi@
04:58:23white_magichow can i create playlists manually (playlists that Rockbox will recognize)? That is - which directory do I place them, and does it matter if they're .pls or .m3u?
05:00:57 Join ReimuHakurei_ [0] (
05:01:42 Quit ReimuHakurei (Read error: Connection reset by peer)
05:10:13JdGordonwhite_magic: anywhere you want
05:10:31 Quit CaptainKewl (Quit: Ex-Chat)
05:16:48 Join mystica555_ [0] (
05:28:15 Quit ReimuHakurei_ (Ping timeout: 264 seconds)
05:37:38 Join Rob2223 [0] (
05:39:08 Join JdGord [0] (~AndChat@
05:41:10 Quit Rob2222 (Ping timeout: 240 seconds)
05:51:02 Quit JdGord (Ping timeout: 258 seconds)
05:54:34 Join BHSPitMonkey_ [0] (
06:01:39 Join ReimuHakurei_ [0] (
06:10:35 Join powell14ski [0] (
06:33:28 Quit soap__ (Quit: Leaving)
06:35:03 Join JdGord [0] (~AndChat@
06:37:08***Saving seen data "./dancer.seen"
06:45:24 Quit logbot (Ping timeout: 260 seconds)
06:45:24***ERROR: (Closing Link: (Ping timeout: 260 seconds)) from
06:45:24***Saving seen data "./dancer.seen"
06:45:27***Started Dancer V4.16
06:45:27***Connected to on port 6667
06:45:27***Logfile for #rockbox started
06:45:30Mode"logbot :+i" by logbot
06:45:33***Server message 501: 'logbot :Unknown MODE flag'
06:45:33 Join logbot [0] (
06:45:33 Join JdGord [0] (~AndChat@
06:45:33 Join powell14ski [0] (
06:45:33 Join ReimuHakurei_ [0] (
06:45:33 Join BHSPitMonkey_ [0] (
06:45:33 Join Rob2223 [0] (
06:45:33 Join mystica555_ [0] (
06:45:33 Join white_magic [0] (white_magi@
06:45:33 Join [7] [0] (~TheSeven@rockbox/developer/TheSeven)
06:45:33 Join pixelma [0] (quassel@rockbox/staff/pixelma)
06:45:33 Join amiconn [0] (quassel@rockbox/developer/amiconn)
06:45:33 Join soap_ [0] (~soap@rockbox/staff/soap)
06:45:33 Join liar [0] (
06:45:33 Join ingoldey_ [0] (~quassel@2002:51dd:ce91:0:21a:a0ff:fe50:5977)
06:45:33 Join linuxguy3 [0] (
06:45:33 Join saratoga [0] (9803ec71@gateway/web/freenode/ip.
06:45:33 Join froggyman [0] (~seth@unaffiliated/froggyman)
06:45:33 Join bluefoxx_ [0] (
06:45:33 Join antil33t [0] (
06:45:33 Join fs-bluebot [0] (
06:45:33 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
06:45:33 Join mystica555 [0] (
06:45:33 Join Horschti [0] (~Horscht@xbmc/user/horscht)
06:45:33 Join user890104 [0] (
06:45:33 Join GeekShadow [0] (
06:45:33 Join user829385 [0] (~aoeu@
06:45:33 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
06:45:33 Join ukleinek [0] (
06:45:33 Join simonlnu [0] (ozA16RVHaB@unaffiliated/simonrvn)
06:45:33 Join balintx [0] (
06:45:33 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
06:45:33 Join dunkaist [0] (~dunkaist@
06:45:33 Join neferty [0] (~andor@
06:45:33 Join scorche|sh [0] (~scorche@rockbox/administrator/scorche)
06:45:33 Join God_Eater [0] (93722cd1@rockbox/staff/GodEater)
06:45:33 Join [Saint] [0] (~st.lasciv@
06:45:33 Join tjb0607 [0] (
06:45:34 Join factor [0] (~factor@
06:45:34 Join Whopper [0] (colin@about/cooking/sandwich/whopper)
06:45:34 Join beta2k [0] (
06:45:34 Join Barahir [0] (
06:45:34 Join otih [0] (~otih@CAcert/Assurer/otih)
06:45:34 Join feisar-_ [0] (
06:45:34 Join preglow [0] (thomj@rockbox/developer/preglow)
06:45:34 Join crwl [0] (
06:45:34 Join rudi_s [0] (
06:45:34 Join rasher [0] (~rasher@rockbox/developer/rasher)
06:45:34 Join ruskie [0] (ruskie@sourcemage/mage/ruskie)
06:45:34 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
06:45:34 Join ft [0] (
06:45:34 Join mikroflops [0] (
06:45:34 Join avacore [0] (
06:45:34 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
06:45:34 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
06:45:34 Join TBCOOL [0] (
06:45:34 Join Zambezi [0] (Zulu@unaffiliated/zambezi)
06:45:34 Join dionoea [0] (~dionoea@videolan/developer/dionoea)
06:45:34 Join bieber [0] (
06:45:34 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
06:45:34 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
06:45:34 Join kugel [0] (~kugel@rockbox/developer/kugel)
06:45:34 Join dre [0] (dre@gateway/shell/
06:45:34 Join sinthetek [0] (~sinthetek@unaffiliated/sinthetek)
06:45:34 Join soap [0] (~soap@rockbox/staff/soap)
06:45:34 Join markun [0] (
06:45:34 Join [fred] [0] (
06:45:34 Join efyx [0] (
06:45:34 Join Slasheri [0] (miipekk@rockbox/developer/Slasheri)
06:45:34 Join aevin [0] (eivindsy@unaffiliated/aevin)
06:45:34 Join desowin [0] (
06:45:34 Join Unhelpful [0] (~quassel@rockbox/developer/Unhelpful)
06:45:34 Join ved [0] (
06:45:34 Join eintopf [0] (
06:45:34 Join zu [0] (
06:45:34 Join jordan` [0] (~gromit@2001:660:3302:2826:225:90ff:fe20:d9a8)
06:45:34 Join pjm0616 [0] (~user@
06:45:34 Join ender| [0] (
06:45:34 Join GodEater_ [0] (~bibble@rockbox/staff/GodEater)
06:45:34 Join Staphylo [0] (
06:45:34 Join Elfish [0] (amba@2a01:4f8:100:90a1:abc:abc:abc:abc)
06:45:34 Join alexbobp [0] (~alex@
06:45:34 Join evildaemon [0] (
06:45:34 Join bzed [0] (
06:45:34 Join niekie [0] (~niek@CAcert/Assurer/niekie)
06:45:34 Join Galois [0] (
06:45:34 Join Guest91631 [0] (
06:45:34 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
06:45:34 Join jae [0] (
06:45:34 Join Zarggg [0] (
06:45:34 Join Torne [0] (~torne@rockbox/developer/Torne)
06:45:34 Join bug2000 [0] (~bug@unaffiliated/bug2000)
06:45:34 Join Guinness [0] (
06:45:34 Join Utchybann [0] (
06:45:34 Join Hadaka [0] (
06:45:34 Join parafin [0] (
06:45:34 Join maraz [0] (
06:45:34 Join krazykit [0] (
06:45:34 Join tmzt_ [0] (~tmzt@
06:45:34 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
06:45:34 Join jacekowski [0] (
06:45:34 Join knittl [0] (~knittl@unaffiliated/knittl)
06:45:34 Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
06:45:34 Join FoolOnHill [0] (
06:45:34 Join FOAD [0] (~foad@unaffiliated/foad)
06:45:34 Join simabeis [0] (
06:45:34 Join scorche [0] (~scorche@rockbox/administrator/scorche)
06:45:34 Join yosafbridge [0] (
06:45:34 Join B4gder [241] (~daniel@rockbox/developer/bagder)
06:45:34 Join nomada [0] (
06:45:34 Join ranmachan [0] (ranma@2a01:4f8:130:9321::2)
06:45:34 Join CIA-14 [0] (
06:45:34 Join @ChanServ [0] (ChanServ@services.)
06:58:57 Quit jae (Read error: Operation timed out)
06:59:02 Join jae [0] (
07:00:18 Quit Utchybann (Read error: Operation timed out)
07:00:54 Join Utchybann [0] (
07:16:10 Join nick-p [0] (
07:21:39 Quit JdGord (Ping timeout: 240 seconds)
07:56:43 Join JdGord [0] (~AndChat@
08:09:59 Quit powell14ski (Quit: powell14ski)
08:12:52 Quit simonlnu (Quit: brb)
08:17:23 Join simonrvn [0] (rs0jUjs2RC@unaffiliated/simonrvn)
08:18:50 Nick simonrvn is now known as simonlnu (rs0jUjs2RC@unaffiliated/simonrvn)
08:19:55 Quit simonlnu (Client Quit)
08:20:53 Quit factor (Read error: Connection reset by peer)
08:25:48 Quit nick-p (Quit: Leaving)
08:30:46 Quit JdGord (Quit: Bye)
08:38:19 Join factor [0] (~factor@
08:41:22 Quit Horschti (Quit: Verlassend)
08:42:34 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
08:45:30***Saving seen data "./dancer.seen"
09:01:36 Join stripwax [0] (
09:01:54 Join whiteh [0] (white_magi@
09:02:13 Join Guinness` [0] (
09:03:34 Join Unhelpful_ [0] (~quassel@rockbox/developer/Unhelpful)
09:04:16 Quit white_magic (Read error: Connection reset by peer)
09:04:16 Quit Unhelpful (Quit: No Ping reply in 180 seconds.)
09:04:18 Quit GeekShadow (Ping timeout: 240 seconds)
09:04:18 Quit Guinness (Read error: Connection reset by peer)
09:04:35 Quit Slasheri (Ping timeout: 240 seconds)
09:04:41 Join Slasheri [0] (
09:04:53 Quit Slasheri (Changing host)
09:04:53 Join Slasheri [0] (miipekk@rockbox/developer/Slasheri)
09:05:23 Join GeekShadow [0] (
09:14:27 Quit stripwax (Read error: Connection reset by peer)
09:20:01 Join stripwax [0] (
09:31:29JdGordonwhile everyone is afk... would there be any arguments with a straight replacement of the curent shortcuts plugin with a more functional one inbuilt?
09:31:49JdGordonwith minimal ram waste for those who dont want to use it
09:32:08[Saint]"while everyone is AFK...would there be any argument"
09:32:13[Saint] ;)
09:32:18JdGordonok good :)
09:32:40[Saint]is the shortcuts menu optional now?
09:32:45[Saint]ie. not present if unused?
09:32:54JdGordonI have no intention of commiting such a menu
09:33:02JdGordonwell, not yet
09:33:03 Join n1s [0] (~quassel@rockbox/developer/n1s)
09:33:43[Saint]Oh, so its just a viewer...kinda?
09:34:58JdGordonI'll make a nice editor plugin for it but otherwise its all in core
09:35:27JdGordonI may add the menu #if 0'd though as I know for sure there is people that want that
09:35:38[Saint]As long as there's no gegression, I really couldn't think of an argument for it not to go in personally.
09:35:47[Saint]Given time, I'm sure the usual suspects could.
09:36:04[Saint]*regression, aslso.
09:36:15JdGordonjust give up on typos already :)
09:45:33 Quit stripwax (Read error: Connection reset by peer)
09:48:36JdGordongevaerts: kugel: in the fast_readline() callback (so reasonable disk access happening) do I *need* to reget my handle every time I want to access it?
09:50:24 Join stripwax [0] (
09:51:30 Join funman [0] (~fun@rockbox/developer/funman)
09:52:17JdGordonwhat the heck is the error value of buflib_alloc()?
09:53:00JdGordonmy initial guess was <0, code suggests 0? would be nice if the function doc actually said so
09:53:17JdGordon"* Returns: An integer handle identifying this allocation" :/
10:15:01 Join Horscht [0] (
10:15:01 Quit Horscht (Changing host)
10:15:01 Join Horscht [0] (~Horscht@xbmc/user/horscht)
10:19:20n1shmm, the mikmod plugin could probably get quite a speedup on cf
10:22:57 Quit Horscht (Quit: Verlassend)
10:25:50 Join y4n [0] (y4n@unaffiliated/y4ndexx)
10:32:04CIA-14New commit by nls (r30427): mikmod: enable compiler optimizations for coldfire, they were disabled to work around a bug in the old toolchain.
10:34:56CIA-14r30427 build result: All green
10:39:44 Join mudd1 [0] (
10:39:54 Quit BHSPitMonkey_ (Quit: Ex-Chat)
10:45:21 Join simonlnu [0] (4uZR5e9TYd@unaffiliated/simonrvn)
10:45:33***Saving seen data "./dancer.seen"
10:45:35 Join Horscht [0] (
10:45:35 Quit Horscht (Changing host)
10:45:35 Join Horscht [0] (~Horscht@xbmc/user/horscht)
10:47:26JdGordonwould anyone like to have a play with fs#12251?
10:47:27fs-bluebot user shortcuts in the main menu (patches, new)
10:47:56JdGordon[Saint]: :)
10:48:19JdGordonI'm cleaning it up now so any requests would be nice to hear
10:48:52[Saint]"make for to it workings goodly plz"
10:50:00JdGordonwell duh!
10:50:03JdGordonit alreayd does
10:50:11JdGordonI'd like to know what is missing
10:50:33*JdGordon guesses if he adds a "shutdown" shortcut [Saint] would put this in his build instantly
10:52:42[Saint]I was just lost in my own head for a second there wondering about how cool it would be if, when it does have an (optional) menu, if said menu coul dbe called from the quickscreen.
10:53:22JdGordonway ahead of you :)
10:53:26[Saint]quick list of custom shortcuts, anywhere, "yes please".
10:53:36JdGordonfirst step is replace the current shortcuts plugin
10:53:47*[Saint] nods
10:55:36 Join TheLemonMan [0] (
11:03:54 Quit mystica555 (Ping timeout: 252 seconds)
11:04:33 Quit mystica555_ (Ping timeout: 268 seconds)
11:08:17 Join Buschel [0] (
11:10:44bluebrotherok. Any thoughts on making Rockbox extract its resources to SD card when /sdcard/rockbox/rockbox-info.txt exists? That would make new installations behave as svn does right now, and if a user wants to put resources on SD card he only needs to create a file.
11:10:59bluebrotherplus, on updates it will use the locating used before
11:11:25JdGordonwhy not always?
11:11:37[Saint]what if a user installs to internal, then creates /sdcard/rockbox/rockbox-info.txt?
11:11:41bluebrotherthe SD card might not be always present.
11:11:48[Saint]where would it install during an update then?
11:12:13bluebrother[Saint]: then the next extraction would store resources on SD card. But why should a user create that file by himself if he doesn't want to change the behaviour?
11:12:42JdGordonhow many phones actually have a uSD slot which is accessable without taking the battery out?
11:12:44[Saint]Oh, it was just your wording suggeted to me that it would be ignored in that case.
11:13:06[Saint]JdGordon: Its not that.
11:13:14[Saint]It might literally *not be present*
11:13:20[Saint]as in, not possible to insert.
11:13:22bluebrotherah, no. The idea is to be able easily move from internal to SD
11:14:18[Saint]bluebrother: Right. Sounds perfectly reasonable.
11:14:22bluebrotherbut keeping the old behaviour. That way it could go into svn without needing to consider cases when the SD card is not available (people with large internal memory might want to prefer installing to internal memory for exactly that reason)
11:14:29[Saint]Users with low internal space will appreciate this.
11:14:51bluebrotherand we could get rid of FS #12063 (aka close it :)
11:14:51fs-bluebot Android: put resources on SD card (patches, new)
11:14:54*Llorean owns a phone with a MicroSD slot that doesn't require removing the battery.
11:15:04LloreanNot to mention tablets and stuff.
11:15:38bluebrotherwell, if people prefer we can also check for some file /sdcard/rockbox/put-resources-on-sdcard.txt or anything else :)
11:15:51bluebrotherit's not like that file will get lost during updates
11:16:14LloreanIs there some reason it'd be impossible to say, prompt on first install?
11:16:26 Join mystica555 [0] (
11:16:36bluebrotherthat would require more changes ;-)
11:16:41[Saint]More work ;)
11:17:14LloreanBut it seems like a more "proper" solution. No weird instructions people need to know about before the install.
11:17:40bluebrotherI tried that some time ago but gave up. It's definitely doable, but I guess my Android knowledge is a bit too limited for that right now
11:18:01bluebrothersure, but we can always add some startup dialog later that creates said file
11:18:19[Saint]took the words right outta my err...keyboard.
11:19:29LloreanWould creating the file, then updating with a newer build, remove those resources from the internal while creating them on the external?
11:20:14[Saint]I just thought about that.
11:20:20[Saint]That would require additional magic.
11:20:33[Saint]and, probably create some really weird side effects.
11:21:11LloreanHow about an option within Rockbox itself, a plugin or menu option, "move resources to SD"?
11:23:53bluebrotherfor removing the resources from internal memory you can just use the "delete data" option in Androids application management.
11:24:03bluebrotherbut yeah, that could be done automatically later.
11:24:24 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
11:25:06[Saint]bluebrother: Would you not *have* to delete the data, beofre the next run with the data on uSD?
11:25:19[Saint]Otherwise I can imagine "very weird shit" happening.
11:26:51bluebrother[Saint]: not really. Plugins are still on internal memory (I'd like to move them to libs as done with codecs some time), codecs are libs, and everthing else is "just" data like themes and fonts
11:27:22bluebrothernot sure which order Rockbox tries internal and SD card.
11:27:32[Saint]what about duplicate config.cfg?
11:27:44JdGordonLlorean: and how is rockbox supposed to find that setting? a setting to tell it to look for its settings?
11:27:44[Saint]I'd assume it tried internal first.
11:27:54bluebrotherbtw, your latest cabbiev2 has a major drawback: it gets always overwritten when resources are on SD card and Rockbox updates ;-)
11:28:03bluebrotherconfig.cfg is always on SD card
11:28:07bluebrotherat least AFAIK
11:28:12 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
11:28:16[Saint]Ah, right.
11:29:55bluebrotherjust checked, config.cfg is always saved on SD. Might make sense to have that always on internal memory though.
11:30:25[Saint]wait...its *always* on the sdcard? I assume that's "only if the sdcard is present"?
11:30:31[Saint]Otherwise, we fail ;)
11:30:36bluebrotherwe do :)
11:30:52[Saint]...right, heh ;)
11:31:11bluebrothersee firmware/common/rbpaths.c:121
11:31:42bluebrotherat least I don't see any "try other location if it fails" path in the code
11:33:03[Saint]pretty hillarious that one of the reasons for this patch hitting the back-burner is "not all Android devices have a sdcard" when the code that decides where user settings are stored assumes they do :D
11:33:27bluebrotherbtw, there's an interesting report in the forums that claims that vorbis files with higher bitrate don't play. That thing happens to me as well since buflib, but I never tried low bitrate files.
11:33:57[Saint].ogg files have been warped for me since long before bufflid.
11:34:07[Saint]and, they're definitely not high bitrate.
11:34:19[Saint]*bufflib, too.
11:34:47[Saint]for me, it seems to be all the packaged Android rintones/notification tones.
11:34:51bluebrotherI had problems with vorbis files before as well, but at some time those were gone and after buflib got in they came back for me.
11:35:54[Saint]for me, most of the notification tones (you can't keep the bastards out of the database in a sane way) play as a white noise, where the played time, current time, and time remaining all report 00:00
11:36:09[Saint]and the progressbar never moves, it just keeps screaming at me.
11:36:13[Saint]most unpleasant.
11:36:42bluebrotherstock tones work fine for me.
11:36:49[Saint]Hmmm, weird.
11:37:01n1s[Saint]: care to post one of those somewhere?
11:37:05 Quit whiteh ()
11:37:26[Saint]n1s: I'll rip one of the ones that definitely doesn't play. One sec.
11:38:20 Join mystica555_ [0] (
11:41:37LloreanJdGordon: I wasn't talking about a setting. I was talking about an option to install them to the SD and remove them from internal, after which the usual detection they were already talking about would come into play.
11:41:46 Quit mystica555 (Ping timeout: 245 seconds)
11:42:44[Saint]Ok, now *alllll* .ogg files play at like ~1% volume, even at 0dB
11:43:01[Saint]and they're *all* doing this "screeching white-noise-of-death.
11:43:16n1sbluebrother: the only thread i've seen about oggs not playing in the forum is the usual huge album art
11:43:19[Saint](I just have to hold the phone right next to my head to hear it.
11:44:56[Saint]RaaA is reporting 54~86kbps
11:45:50 Join MikeH__ [0] (
11:46:11n1sunfortunately getting libtremor to skip albumart is not easy
11:46:17[Saint]but yeah, the new hilarious .ogg oddity is all of the tracks being *really* quiet.
11:46:38[Saint]even with volume at 0db, and IEMs, I can *just* hear it.
11:46:46 Quit mystica555_ (Read error: Operation timed out)
11:47:32[Saint]n1s: uploading now.
11:49:10 Join mystica555_ [0] (
11:49:47[Saint]none of these play, or, they do...but its all a high-pitched *scream*
11:51:55Buscheldoes anybody have any midi file that actually plays in rockbox? none of the ones I have just downloaded is played in the simulator
11:52:10n1sBuschel: i think i have some
11:52:34 Join Jerom1 [0] (~jerome@
11:52:38 Quit MikeH__ (Ping timeout: 240 seconds)
11:53:04Buscheln1s: could you upload one or two of them somewhere?
11:53:21n1sBuschel: if i find then, sure
11:53:30*Buschel hopes nqs will find them
11:56:24bluebrothern1s: I was referring to a report in the RaaA thread, not a new one
11:56:35n1sbluebrother: ah
11:56:44n1s[Saint]: ok, got them, thanks
11:56:48bluebrother[Saint]: that's exactly the issue I have. You can barely hear the files playing
11:56:51 Join MikeH__ [0] (
11:57:16[Saint]bluebrother: for me, its new...but, I'm actually thankful.
11:57:30[Saint]barely hearing an intolerable screaming is appreciated ;)
11:57:37bluebrotherhehe :)
11:57:41 Join nick-p [0] (
11:59:39Buscheln1s: forget about my request... I forgot to install the patchset
11:59:43Buschelworks now
12:00:21 Quit mystica555_ (Ping timeout: 246 seconds)
12:00:26Buschelthanks anyway, I will put them in my test folder as well :)
12:01:32[Saint]Are they not redistributable? Could they not go in some sub-folder of test_files?
12:02:02n1si have no idea
12:02:26n1shmm, fileroller doesn't like your zipfile
12:03:03[Saint]I zipped it right off the phone with RootExplorer.
12:03:12[Saint]7zip's treating it fine here.
12:03:23n1syeah, it's some 7zip thing
12:04:11pamaurygevaerts: ping
12:04:14n1sah, no it's trying to use 7z and it doesn't like it
12:05:18n1sunzip did it but complained a lot
12:06:21[Saint]that's odd. I've never had a problem with the .zip files RootExplorer creates, but I've only ever used 7zip and RootExplorer itself to extract them.
12:07:50n1sthey do play fine in the sim though
12:08:49[Saint]Must be some Android thing.
12:10:15pamauryhaha, I found my bootloader bug, that was really nasty
12:11:15n1s[Saint]: so all vorbis files play back as very faint noise on android?
12:11:26[Saint]For me, yeah.
12:11:53[Saint]It just so happens that the only .ogg files I have ready access to unless I'm specifically testing are my ringtones, though.
12:16:29n1snothing seems weird with them, one file seems to have broken rg tags but nothing else suspicious
12:30:04 Quit MikeH__ (Ping timeout: 252 seconds)
12:34:18n1sdoes that ogg problem on android affect everyone?
12:34:44[Saint]bluebrother and I, at least.
12:34:59[Saint]different HW, different revisions.
12:35:07n1sdid it work at some point?
12:35:46[Saint]never completely. It was either random .oggs would scream at me, and some would play fine. Or none played at all.
12:35:57[Saint]this "incredibly low volume" thing is new to me.
12:36:45bluebrotherit works fine for me for some revisions. Haven't tried to find a pattern though.
12:37:11n1svorbis uses a lot of memory, maybe something is messing with its allocations
12:37:11bluebrotherright now r30379 works, r30380 and r30381 start to break things
12:37:16 Join v4hn [0] (
12:37:23n1sdoes aac work?
12:37:38bluebrothernever tried.
12:37:42*bluebrother looks for an aac file
12:37:52n1sshould be a few in thest_files
12:41:51pamauryfuze+ usb doesn't work, I don't if I ran into the errata situation or if something else is wrong :(
12:42:46bluebrotherhmm, I guess I should install a build that exposes the problem first :)
12:43:03[Saint]there's builds that don't?
12:43:08[Saint] one on me.
12:43:35n1sbluebrother: i'm asking because aac also uses a lot of memory
12:43:39bluebrotheryes, my build including AA in notification area :)
12:43:47bluebrotherat least for me −− vorbis plays fine.
12:43:52*bluebrother builds trunk
12:44:34gevaertspamaury: pong
12:45:36***Saving seen data "./dancer.seen"
12:48:04gevaertsJdGordon: I'm not sure. I'd say you want to lock the buffer for fast_readline() Nuse
12:48:40 Quit [Saint] (Quit: Imagination is for turbo-nerds who can't handle how kick-butt reality is. I'm a kick-butt reality master! I would rather die, than be imaginative. I mean that.)
12:49:01pamaurygevaerts: I don't know if you remember well the code but I noticed that in usb_init(), we first call usb_init_device and then init the event queue. Although that's unlikely a controller might generate an interrupt inbetween and the core will push an event on an uninitialized queue
12:49:36bluebrothern1s: nero_192 and nero_he_64 play fine for me with trunk.
12:50:00 Join mystica555 [0] (
12:50:04 Join mystica555_ [0] (
12:50:30n1sbluebrother: but vorbis doesn't?
12:51:04 Join [Saint] [0] (~st.lasciv@
12:51:17bluebrothern1s: no, neither vorbis_192 nor vorbis_96
12:51:22 Quit ingoldey_ (Remote host closed the connection)
12:51:38gevaertspamaury: hm, that sounds wrong indeed
12:51:41bluebrotherwell, they _do_ play but are so soft that you barely hear anything at maximum value
12:52:19bluebrotherurgh, wget was downloading a lot more files than the test_files only :(
12:53:24pamaurygevaerts: ok, if you agree I'll swap the other anytime soon, it can't hurt
12:53:41gevaertspamaury: I suspect usb_init_device() should be at the very end
12:53:47pamauryyes me too
12:55:10gevaertsMost (or all?) stuff seems to check usb_monitor_enabled so we're probably safe though
12:55:27gevaertsBut it's still not good
12:56:23pamaurythere is one exception: transfer completion
12:57:20n1sbluebrother: is the volume problem also only present on some builds?
12:57:39*gevaerts nods
12:58:25gevaertsAs far as I can see there aren't any downsides with moving, and some (albeit unlikely) problems with the current order, so...
12:59:35n1swhat do i need to do to run the sim under valgrind?
13:00:06gevaertsn1s: valgrind and a sim ;)
13:00:36n1sgevaerts: something else too i think since that doesn't work
13:01:35gevaertsn1s: maybe you want −−sdl-threads. Our own thread implementation can break some assumptions tools like valgrind and gdb make
13:01:59*n1s tries
13:03:42bluebrothern1s: I had that problem earlier, but I haven't paid attention when it appeared / disappeared.
13:04:00 Quit y4n (Disconnected by services)
13:04:07 Join y4n [0] (y4n@unaffiliated/y4ndexx)
13:04:28n1sbluebrother: bur does it affect all formats or only some?
13:04:38 Quit factor (Ping timeout: 240 seconds)
13:04:49bluebrotherI'm trying to play all test files to see :)
13:04:59bluebrotherI'm usually only using mp3 and ogg
13:05:20n1sthe volume problem affects mp3 then?
13:05:28[Saint]not for me.
13:05:31bluebrotherno, it never did affect mp3
13:06:01 Quit funman (Quit: leaving)
13:06:11n1sgevaerts: −−sdl-threads works
13:08:36bluebrothern1s: ok, it does _only_ affect vorbis_* out of the files in test_files. Except vorbis_350 (which plays) and vorbis_500 (which gets skipped)
13:09:51bluebrotherthough vorbis_500 plays fine with mplayer, so unless there was some error when transferring to the phone it's not a broken file
13:10:14bluebrotherwhat really confuses me is that vorbis_350 *does* play
13:10:38bluebrotherand that track is really annoying :)
13:11:17 Join ageis [0] (
13:12:57pamauryfuckkkkkkkkk, I just understood my usb problem
13:14:31n1svalgrind on the sim doesn't say anything, possibly i don't understand it well enough, is there a way to use it on android?
13:14:52bluebrotherno idea.
13:15:16bluebrotherdo I need to do anything special to enable debugf with libtremor other than enabling it in configure?
13:16:06n1si don't remember there's so much weird stuff with logf and debugf
13:16:30*bluebrother builds with debugf and logf enabled
13:16:41 Quit mystica555 (Ping timeout: 258 seconds)
13:17:42 Quit mystica555_ (Ping timeout: 260 seconds)
13:17:42n1si *think* logf needs to be enabled both in configure and per-file that uses it but i'm not sure about debugf except that i think debugf works in the sim by default
13:18:42pamauryyes logf needs to be enable in each file (#define LOGF_ENABLE)
13:22:13n1siiuc the codecbuf as such on android id just a 1M static buffer and the actual codec is not loaded into it at all, other than that it seems to work pretty much like native so i don't know where to look
13:23:30bluebrotherhmm, now vorbis_350 doesn't play either.
13:24:12bluebrotherlog looks like this:
13:26:15bluebrotherhmm, I can't see much of DEBUGF in libtremor
13:27:09bluebrotherwhat do I need to define to make DEBUGF work?
13:28:45n1sther's only a single DEBUGF in libtremor it seems (for failed alloc, which might be interesting)
13:28:46*bluebrother found it
13:29:01bluebrotherthere are a few in vorbis.c
13:29:20n1sbluebrother: are those java: set_volume() calls supposed to be there?
13:29:20*bluebrother triggers a new build and leaves for food
13:32:02 Join mystica555 [0] (
13:32:43 Join mystica555_ [0] (
13:33:05bluebrothern1s: yes
13:57:29 Quit Jerom1 (Quit: Leaving.)
14:06:45JdGordongevaerts: comments say get_data() is basically free so as long as i dont yield between that call and usage I should be safe?
14:11:31gevaertsJdGordon: not sure...
14:30:58 Quit nick-p (Quit: Leaving)
14:33:42JdGordonn1s: GodEater_ : do you happen to remember why the lang string is LANG_ADD_TO_FAVES but the english is "add to shortcuts"?
14:35:50Buschelis anything speaking against submitting the midi change in FS #10590 −− especially if I found a nice way to speed optimize the synthesis? even PP5022 can now handle 24 voices without loosing sync
14:35:51fs-bluebot Aliasing when playing MIDI files on iPod 5G (bugs, unconfirmed)
14:37:56n1sBuschel: either my hearing isn't good enough or that only affects some targets as i don't hear any nasty aliasing on cf
14:38:43n1si also thougt midi usually was played at 22.05kHz
14:39:05n1sbut if it's fast enough either way but one sounds better then by all means
14:39:26n1sbuffer misses do really sound like crap though
14:40:34Buscheln1s: does this work for you? ->
14:43:43n1si'll try to test it later nut i'm pretty sure the cf targets will not be able to pull of 44.1kHz
14:43:58n1spull off
14:44:43n1sanyway, there's a subtle bug in the sample packing that i introduced at some point but forgot to fix
14:45:26n1sthe lower sample can bleed into the upper sampleif it's negative or something
14:45:37***Saving seen data "./dancer.seen"
14:46:54JdGordondoes anyone know functionality the shortcuts plugin provides? specifically referring to what it lets you do with the .link file
14:48:18Buscheln1s: where? from what I see there is no saturation (so, it will invert sign if overflowing), but no bleeding between samples
14:49:12n1sBuschel: it's from the addition, the two 16 bits values are packed and then added as a single 32 bit value to a 32 bit value
14:49:21v4hnn1s: thanks for your comment on negative lcds. Just fixed the patch
14:49:25JdGordon... looks like just delete and append...
14:50:15n1sBuschel: to make it correct they should be added as two separate 16 bit values
14:50:17Buscheln1s: understood
14:50:54n1si like that pan== 64 thing :)
14:51:16Buschelthat gave the most speedup :)
14:51:35n1si can imagine
14:56:13n1siirc armv6 has instructions for packed values but as always that's where it's needed the least
14:58:44n1sBuschel: i guess it would be cleanest to just make the output buffer int16_t and adress the samples individually
14:58:52n1spossibly a bit slower though
14:59:27BuschelI am exactly working on this right now :)
14:59:41JdGordonwhich target generally has the most bloated binaries?
15:01:42 Join soap__ [0] (~soap@
15:02:30n1sJdGordon: regular armv4 i think
15:03:06JdGordonarg, 2.5k in code, probably from all the extra strings :/
15:04:34n1sJdGordon: about the string, i think it was called favourites initially but changed to shortcuts later
15:04:47 Quit soap_ (Ping timeout: 246 seconds)
15:06:41n1sstring was added like that...
15:07:10n1sweird, although changing the name of the id is no problem as long as it's done for all .langs
15:12:33n1sBuschel: i'm still somewhat inclined to think the aliasing bug reported in the tracker is something specific for ipods as i've never heard anyone with a different target using 22.05kHz complain so, or it might just be the nature of ipod users :)
15:13:45 Quit rudi_s (Ping timeout: 240 seconds)
15:13:50n1sbtw the mikmod plugin's menu causes the playback to loop the last few samples as long as it's visible
15:13:51 Join rudi_s [0] (
15:14:34Buscheln1s: I assume this caused by the resampler −− which does not do a good job in upsampling...
15:15:36 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
15:17:10CIA-14New commit by jdgordon (r30428): Remove duplicated code to search for a setting by its cfg name
15:20:09CIA-14r30428 build result: 13 errors, 0 warnings (jdgordon committed)
15:20:21JdGordonI wasnt expecting 13 errors! 1 warning maybe...
15:23:35n1sBuschel: but it shouldn't upsample, the hardware should play back at 22.05kHz if that is supported and selected
15:23:41CIA-14New commit by jdgordon (r30429): fix error, none of this code is needed or would work in the checkwps program anyway
15:25:44Buscheln1s: after chaning the buffer to uint16 the performance has gone down −− as expected
15:26:21CIA-14r30429 build result: All green
15:26:36n1sBuschel: how much?
15:27:21Buscheldon't know, at least I get a few buffer misses now
15:31:36n1si did hack up some benchmark code for the midi plugin at one point but i think i've lost it :(
15:34:40n1sBuschel: btw, the samp_buf hack is not needed for non cf targets but is probably not slowing down too much
15:35:12Buschelyes, I played with this, but there was no remarkable difference
15:36:32Buscheln1s: latest version ->
15:39:18n1sBuschel: looks good but i have to run now
15:41:17Buscheln1s: but I am still a bit unsure what causes the aliasing... at least the bug report is valid for iPod Video. maybe I should only set 44100 kHz for this target for now?
15:43:54amiconnn1s, Buschel: You can do packed addition of two 16 bit values in a 32 bit word with a bit of extra processing
15:44:40*amiconn used such trickery in the ape filters, except for arm7tdmi which is faster when using 32 bits throughout
15:46:09amiconnThe latter is not because the packed additons would be slow, but rather because of multiplications needing unpacking
15:48:19 Join Lear [0] (
15:53:59 Quit Horscht (Quit: Verlassend)
15:54:54 Join GeekShad1w [0] (
15:55:05 Quit GeekShad1w (Client Quit)
15:55:24 Quit GeekShadow (Remote host closed the connection)
15:55:43 Join GeekShadow [0] (
16:02:12 Quit antil33t (Read error: Connection reset by peer)
16:02:36 Join antil33t [0] (
16:07:50 Quit ReimuHakurei_ (Ping timeout: 276 seconds)
16:15:02 Join Horscht [0] (
16:15:02 Quit Horscht (Changing host)
16:15:02 Join Horscht [0] (~Horscht@xbmc/user/horscht)
16:20:52 Quit [7] (Remote host closed the connection)
16:20:58 Quit Buschel (Quit: ChatZilla 0.9.87 [Firefox 6.0.1/20110830092941])
16:28:04 Quit Horscht (Quit: Verlassend)
16:35:02 Quit v4hn (Quit: leaving)
16:41:26 Quit antil33t (Read error: Connection reset by peer)
16:41:46 Join antil33t [0] (
16:45:38***Saving seen data "./dancer.seen"
16:50:21 Join Horscht [0] (
16:50:21 Quit Horscht (Changing host)
16:50:21 Join Horscht [0] (~Horscht@xbmc/user/horscht)
16:51:15pamauryit's a shame that usb-drv-arc assume that uncached stuff has virtual address = physical address
16:51:29 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
16:52:21 Quit TheSeven (Read error: Connection reset by peer)
16:59:48 Join powell14ski_ [0] (
17:00:28 Quit Horscht (Quit: Verlassend)
17:01:36 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
17:13:41 Join Horscht [0] (
17:13:41 Quit Horscht (Changing host)
17:13:41 Join Horscht [0] (~Horscht@xbmc/user/horscht)
17:14:01 Join fyrestorm [0] (
17:35:21pamaurygevaerts: why was usb_devbss_attr introduced ? what is it supposed to mean ?
17:38:29 Quit Horscht (Quit: Verlassend)
17:47:01 Join Horscht [0] (~Horscht@xbmc/user/horscht)
17:48:02pamauryhum, I'm really puzzled, I need to define something for USB_DEVBSS_ATTR on fuze+ since I used the mmu to cache the dram. I can create a noncached bss section with has 1:1 mapping (required implictely by the usb driver) but then I'll have to map it using the mmu which will result in a 1Mb noncached section just for a few kilobytes !
17:49:17soapDid I miss where it was announced that FS emails were going to come from a new address?
17:49:45pamaurysoap: nope, I had to change my filters and didn't see any email
17:50:10pamauryalternatively I can setup a noncache mirror of the dram with UNCACHED_ADDR/PHYSICAL_ADDR macros but then I need to fix the usb driver :-/
17:50:16pamauryany thought ?
17:58:59 Quit Lear (Quit: ChatZilla 0.9.87 [Firefox 7.0/20110831124126])
17:59:17 Join Horschti [0] (
17:59:17 Quit Horschti (Changing host)
17:59:17 Join Horschti [0] (~Horscht@xbmc/user/horscht)
18:00:07 Quit Horscht (Ping timeout: 240 seconds)
18:04:30 Nick Unhelpful_ is now known as Unhelpful (~quassel@rockbox/developer/Unhelpful)
18:05:12jhMikeSpamaury: USB_DEVBSS_ATTR is mapped to iram on portalplayer and uncached dram on i.MX31...and you can map sections in many ways at once, whatever suits
18:07:00jhMikeShmmm...maybe that was just QHARRAY
18:08:28jhMikeSnope, it's that too
18:08:56pamauryqharray is different
18:09:29pamaurybut mapping 1Mb of uncached dram just for a few kilobytes seems like a waste
18:09:33jhMikeSboth are still in IRAM on pp
18:09:48pamauryyes but I can't put it in iram on the fuze+
18:10:03pamauryotoh qharray *needs* to be in iram
18:10:12jhMikeSit's not in iram i.MX31 either
18:11:54pamauryI guess the mapping solution is the simplest...
18:13:01jhMikeSthe beast has a 1MB section mapped three different ways, cached virtual, uncached physical, and write buffered virtual, so the whole thing is pretty flexible
18:14:03pamaurythe problem is not flexibility, it's just about granularity
18:15:02jhMikeSwhat for?
18:15:16pamaurywell mapping 1Mb for a few kilobytes...
18:16:15jhMikeSwhy's that an issue?
18:16:26pamaurythat's mean either lots of wasted memory or some cached memory which will endup being uncached which is a shame
18:16:55jhMikeSare you trying to map cached onto physical? if so, why?
18:17:33pamaurythe usb-arc-drv assumes uncached addr = physical addr
18:17:51pamaury= virtual addr
18:18:01jhMikeSso, why not map it that way?
18:18:32jhMikeSI don't know why as3525 for example had to do it all differently
18:18:50pamaurythen back to my problem about mapping 1Mb of uncached for a few kilobytes
18:19:24pamauryexcept if I don't identity map the whole dram
18:19:55jhMikeSwhy do that?
18:20:06pamaurydo what ?
18:20:12 Join CaptainKewl [0] (captainkew@
18:20:15jhMikeSyou mean identity map the dram as cached dram?
18:20:32pamaurythat's what I do currently
18:21:06pamaurybut if I don't identity map it, then I'll probably have to fix the usb driver anyway because it assumes input buffers have virtual = physical addr :)
18:21:18jhMikeSah, then it's like as3525 is done. the gigabeats f/x/s map cached dram to 0 and identity map uncached
18:21:54pamauryEither way I'll have to fix the driver x-(
18:22:11jhMikeSwhy? you'd have a map like gigabeat S, which uses arc
18:22:38 Join ender` [0] (
18:24:09pamauryI can't see how the gigebeat S works since my iram has a physical addr of 0 so if I map the dram at 0, I need to map iram somewhere else and since the driver assumes virtual = physical for qharray which is in iram it will break
18:25:02jhMikeSah, so then you could map cached dram somewhere else, anywhere really
18:25:40pamauryanswer: the imx31 iram is not at 0
18:26:49pamaurythen I think need to fix the driver because the usb driver uses the virtual address as physical address
18:27:13jhMikeStrue, it's not at 0x0, but it could easily have cached dram start at any number of addresses
18:27:38jhMikeSit got it in place for PP at the time it was created
18:28:15pamaurywhat I could do is identity map uncached the dram and the iram and map cached dram somewhere else and fix the driver if needed
18:28:58jhMikeSI think that's the easiest way and it probably wouldn't need fixing other than to provide correct macros
18:30:24jhMikeSone advantage of identity mapped being uncached is there's no need to maintain coherence
18:31:03 Join factor [0] (~factor@
18:32:07pamaurytrue, you only need to take care when giving pointers to the hardware
18:33:33n1samiconn: ah, interesting, care to point out where you did it?
18:35:03jhMikeSportal player's memory controller isn't quite so nice as the other arm ones that let you put stuff anywhere
18:37:04pamauryhow iram do the pp have ?
18:37:07pamaury*how much
18:38:39jhMikeS48K I think
18:39:01jhMikeScrap I forget, it's more than that
18:39:39gevaerts96 or 128 depending on the PP variant IIRC
18:42:36pamauryI will need a uncached but buffered mapping for frame probably
18:44:19jhMikeSjust map that 1MB bit a 3rd way somewhere else if you can tell the HW where the frambuffer should be
18:45:26jhMikeSon the beast I just stuck it right after the remapped cached dram, buffered only but still give the hw the physical address
18:45:40***Saving seen data "./dancer.seen"
18:49:06jhMikeSso, the last MB is mapped three times
18:49:28 Quit antil33t (Read error: Connection reset by peer)
18:49:51 Join antil33t [0] (
18:51:49amiconnn1s: apps/codecs/demac/libdemac/vector_math16_*.h
18:52:10amiconnFor * = armv5te and cf there are ADDHALFREGS and SUBHALFREGS macros
18:52:46amiconnarmv6 and higher and mmx have dedicated instructions for packed addition/ subtraction
18:53:30amiconn(the armv7 version uses neon)
18:55:48*jhMikeS wonders why the VFP was enabled on imx31 (what's using it?)
18:57:19n1snothing, iirc
18:57:25n1samiconn: ah, thanks
18:59:07jhMikeSany FIQ related items should probably be pulled since FIQ is completely borked on imx31 without crazy workaround code
19:02:07n1siirc Unhelpful enabled the vfp but i don't think he ended up using it for anything
19:02:20pamauryimx233 the errata is quite frightening :)
19:03:48Unhelpfuln1s: I tried using it for integer division but with conversions it was slower than even the stock gcc trial-subtraction algorithm.
19:06:40UnhelpfulI really should submit our divider for inclusion in libgcc
19:07:08n1sUnhelpful: aha, is the thouroughput better than fixed point for multiplication if the conversions are disregarded?
19:07:41n1sUnhelpful: you need the contrubutor agreement paperwork, no?
19:08:35UnhelpfulI rather doubt it, if you can avoid tight data deps multiply is what, two cycles? With four or so latency?
19:09:43UnhelpfulI have no idea how to contribute to libgcc, but we have better dividers than it does for arm.
19:13:41jhMikeSamiconn submitted a patch to them regarding a wrong msac opcode on coldfire, so I guess he would know
19:14:41 Join MethoS- [0] (~clemens@
19:15:25n1sjhMikeS: wasn't that for binutils? i think they are more lax
19:15:35UnhelpfulDoes newer arm have idiv? Or vector idiv in neon?
19:15:59n1si think so
19:16:36UnhelpfulI mean, this would still be good... For arm 4-6 or so.
19:16:46jhMikeSn1s: I just recall preglow getting frustrated and then discovering the problem, then something about amiconn mentioning an accepted patch for the issue
19:17:37 Join BHSPitMonkey_ [0] (
19:17:37jhMikeSthe discussion is likely logged
19:17:57n1sjhMikeS: i remember, it was a bug in AS that emitted the wrong opcode or something, gcc doesn't deal with the mac stuff at all
19:19:02jhMikeSwell, the assember does, so that's binutils I guess
19:19:22n1sfor gcc itself, nontrivial things are not accepted without contributor assignment at least
19:19:28UnhelpfulYeah, this is support library stuff.
19:19:45B4gdersurely that's the same for binutils?
19:19:54B4gderas that's a GNU and FSF thing
19:20:41B4gderthe trick is just to call it trivial =)
19:21:02gevaertsTell them it took you ten minutes :)
19:21:03jhMikeSgradeschoolers know how to divide afterall
19:21:25 Quit stripwax (Read error: Connection reset by peer)
19:21:31 Join guymann [0] (
19:22:28UnhelpfulIt took me longer than that to understand what the asdg example was doing, screen. :/
19:23:41UnhelpfuljhMikeS: they do not typically know newton-raphson estimation, though. Well, not in the USA, anyway.
19:25:15jhMikeSwell, not after "no child left behind" in any case
19:26:51 Join stripwax [0] (
19:27:57 Quit TheLemonMan (Quit: WeeChat 0.3.5)
19:28:19 Join TheLemonMan [0] (
19:31:02 Quit TheLemonMan (Client Quit)
19:31:20 Join lorenzo92 [0] (~chatzilla@
19:31:42 Join TheLemonMan [0] (
19:31:47 Quit antil33t (Read error: Connection reset by peer)
19:31:57 Quit lorenzo92 (Client Quit)
19:32:11 Join antil33t [0] (
19:34:00n1sUnhelpful: it seems *some* v7 has integer division but not the "A" versions
19:35:51UnhelpfulThey also keep increasing mul latency, and that will eventually make the n-r divider slow
19:38:22 Join ReimuHakurei [0] (~kudo@
19:38:29UnhelpfulOr at least they did in going from 5e to 6
19:39:15 Quit stripwax (Read error: Connection reset by peer)
19:43:00 Join Jerom1 [0] (~jerome@
19:46:03 Quit bluebrother (Disconnected by services)
19:46:05 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
19:48:24 Quit fs-bluebot (Ping timeout: 245 seconds)
19:49:04 Join funman [0] (~fun@rockbox/developer/funman)
19:49:31 Join fs-bluebot [0] (
19:50:45pamaurywow, the bootloader runs how of the box even after remapping the dram, that's close to being a miracle. Of course I need to fix all my dma code to use physical addresses now...
19:56:50 Join FoH_Phobos [0] (~foh@
19:58:33 Join stripwax [0] (
19:59:35amiconnjhMikeS: My patch was for binutils, not gcc
19:59:48pamauryomg, it works on the first run, I must be on a doing-things-right spree :)
20:04:02 Quit funman (Quit: leaving)
20:04:29 Join robin0800 [0] (
20:05:13 Quit antil33t (Read error: Connection reset by peer)
20:05:35 Join antil33t [0] (
20:06:19 Join soap_ [0] (~soap@rockbox/staff/soap)
20:08:25 Quit soap__ (Ping timeout: 258 seconds)
20:13:38 Quit stripwax (Quit: Miranda IM!- Smaller, Faster, Easier.
20:13:39 Quit TheLemonMan (Quit: WeeChat 0.3.5)
20:16:26 Quit soap_ (Ping timeout: 252 seconds)
20:18:16 Join TheLemonMan [0] (
20:22:23 Quit antil33t (Read error: Connection reset by peer)
20:22:44 Join antil33t [0] (
20:41:19 Quit Jerom1 (Quit: Leaving.)
20:44:28 Quit factor (Read error: Connection reset by peer)
20:45:43***Saving seen data "./dancer.seen"
20:49:37 Join Keripo [0] (
21:01:56 Join factor [0] (~factor@
21:06:23 Join funman [0] (~fun@rockbox/developer/funman)
21:07:22CIA-14New commit by bluebrother (r30430): Android: store resources on SD card. ...
21:07:46 Quit benedikt93 (Quit: Bye ;))
21:09:52CIA-14r30430 build result: All green
21:11:41 Quit y4n (*.net *.split)
21:11:41 Quit otih (*.net *.split)
21:11:41 Quit aevin (*.net *.split)
21:14:14 Quit Keripo (Quit: Leaving.)
21:14:44 Join Keripo [0] (
21:14:46 Join petur [0] (
21:14:51 Quit petur (Changing host)
21:14:51 Join petur [0] (~petur@rockbox/developer/petur)
21:15:28 Quit Keripo (Client Quit)
21:20:26*bluebroth3r wonders if it's possible to link libplugin as dynamic library on Android (and RaaA in general)
21:21:00 Join aevin [0] (
21:21:00 Join y4n [0] (y4n@unaffiliated/y4ndexx)
21:21:00 Join otih [0] (~otih@CAcert/Assurer/otih)
21:21:52 Quit dunkaist (Quit: leaving)
21:24:25 Quit Llorean (Read error: Connection reset by peer)
21:42:33 Quit akaWolf (Ping timeout: 245 seconds)
21:43:09 Join Xerion [0] (
21:56:12 Quit Horschti (Quit: Verlassend)
21:57:46 Join Keripo [0] (
21:59:46 Join Horscht [0] (
21:59:46 Quit Horscht (Changing host)
21:59:46 Join Horscht [0] (~Horscht@xbmc/user/horscht)
22:16:46 Quit mudd1 (Read error: Operation timed out)
22:18:47bertrikhi funman, do you have any comments on FS #12255 ?
22:18:48fs-bluebot sansa clipzip: dualboot code for review (patches, new)
22:22:45 Join dunkaist [0] (~dunkaist@
22:28:27 Quit Horscht (Quit: Verlassend)
22:33:24 Join mudd1 [0] (
22:33:57 Quit petur (Quit: here today, gone tomorrow)
22:38:56funmanbertrik: let me take a quick look
22:43:01bertrikone thing not mentioned there is that the final step would be to also boot the OF on wake-by-USB-plug-in
22:43:17 Quit robin0800 (Read error: Connection timed out)
22:43:21funmandid you find a clip zip already?
22:44:18 Join robin0800 [0] (
22:44:59funmanbertrik: all patches look ok (hard to make a mistake with that few changes)
22:45:45***Saving seen data "./dancer.seen"
22:46:00bertrikfunman, no. I think I could get one from US amazon, but that would cost something like E80,- because of shipping / duties / taxes etc.
22:46:31funmanah right
22:48:35saratogaonce they come down a little i could send you one
22:49:43saratoga$50 is a bit steep for me this month though
22:50:04funmancan't the rockbox fund help you?
22:50:55 Quit ender` (Read error: Connection reset by peer)
22:51:34 Join ender` [0] (
22:51:35bertrikI don't know. I don't mind buying it myself, but I think I'll wait until it is E50,- or lower.
22:52:36saratogagoogle says its launching in europe in September, so i guess you'll get a chance soon enough
22:54:27 Quit ender| (Ping timeout: 260 seconds)
22:54:50 Quit ender` (Read error: Connection reset by peer)
22:55:25 Join ender` [0] (
22:56:18 Quit evildaemon (Quit: Lost terminal)
22:56:58 Quit ageis ()
22:57:07 Join ender| [0] (
22:57:37 Join ageis [0] (
22:58:24 Join nick-p [0] (
23:07:29soapI can front the cash if you want me to ship you one.
23:11:10bertrikthanks for the offer, but I think I'll just wait until it becomes for sale locally
23:13:57 Quit efyx (Remote host closed the connection)
23:34:52 Quit CaptainKewl (Quit: ( :: NoNameScript 4.22 :: ))
23:54:59 Join Horscht [0] (
23:54:59 Quit Horscht (Changing host)
23:54:59 Join Horscht [0] (~Horscht@xbmc/user/horscht)

Previous day | Next day