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 2008-10-29

00:00:54rasherHrm, seems a bit unfit to build a crosscompiled sdl
00:01:03 Join dany_21a_ [0] (
00:01:53 Quit bertrik ("Leaving")
00:02:02funmandoesn't sdl distribute built dev packages?
00:03:18rasherIt would seem so
00:05:05rasherAlthough I'm not sure it'd work
00:06:42***Saving seen data "./dancer.seen"
00:09:53rasherI don't know if the mingw32 package is identical to what we'd want. But maybe
00:09:58 Join neddy [0] (n=john@nat/sun/x-b6528d6e5c509da3)
00:11:27 Quit n1s ()
00:12:27 Quit DerDome ("Leaving.")
00:12:38rasherWell, I'm getting a compile error
00:12:58 Part dany_21a_
00:16:48rasherWhich I don't get if I compile SDL myself, so it would seem that *something* isn't right
00:18:05funmansome difference in the headers?
00:18:14funmanor a missing symbol?
00:18:28rasherYes :)
00:18:45 Quit dany_21a (Read error: 110 (Connection timed out))
00:20:15 Quit courtc (Read error: 60 (Operation timed out))
00:20:38funmanI played a bit with the cpu and peripheral clocks on the Clip but no result for SD :(
00:24:10 Quit culture (Read error: 60 (Operation timed out))
00:24:10rasherBagder: do you use tools/release/ for building the dailies?
00:24:10Bagderno, only releases
00:24:10 Quit blkhawk (Remote closed the connection)
00:24:10 Join blkhawk [0] (
00:24:10Bagderit was derived from my daily-build script
00:24:10funmanby the way the NAND/SD isr branches to code loaded from the flash into SDRAM
00:24:10funmanso the interrupts are not required
00:24:10 Join courtc [0] (
00:24:10rasherBagder: wouldn't it make sense to merge them? Aren't they practically the same, except for the version in the filename?
00:24:10 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
00:24:10Bagderbasically, yes
00:24:23Bagderthe dailies are in different subdirs too
00:24:33rasherSo if I made a tools/releases/, we'd be nearly there?
00:24:45Bagderit'd be a fine start yes
00:25:11 Join mattgs_ [0] (
00:25:18 Quit mattgs_ (Remote closed the connection)
00:25:54 Join miepchen^schlaf [0] (
00:26:23funmanwould you have documentation resources on fastbus/synchronous/asynchronous mode of ARM cpu ?
00:26:23 Join mattgs [0] (
00:29:32 Quit miepchen^schlaf (Client Quit)
00:31:14denes_funman: that I was wondering about myself. isn't the meaning simply fastbus:cpu_clock=preipheral_clock asynchronous=cpu_clock>peripheral_clock? some documentation would be good though.
00:32:07funmanthat makes sense: the cpu is in fastbus at reset, so you have time to setup the clocks
00:32:15 Quit ender` (" cd /pub && get beer")
00:32:19 Join miepchen^schlaf [0] (
00:32:24funmangoogle mostly leads to website
00:35:34denes_funman: I gathered this from the s5l8700 docs. when it talks about the DSP (and not the ARM core), this is how it describes fastbus and asynchronous mode.
00:40:07funman"In FastBus mode GCLK is sourced from BCLK. The FCLK input is ignored. This means that BCLK is used to control the AMBA ASB interface and the internal ARM922T processor core."
00:40:14mattgsHi. I have a 5.5 gen 80 gig windows ipod, and I am running linux. I am trying to use the rockbox utility to install rockbox on my ipod, but I keep getting the error message "could not open ipod" I don't know what I am doing wrong
00:40:20funmanfclk is the cpu clock, not sure about gclk and bclk
00:40:32funmanmattgs: are you running as root ?
00:41:00mattgsshould I?
00:41:08funmanyes, you must :)
00:41:15mattgsthank you
00:41:53funmandenes_: how far are you with your work?
00:42:34denes_funman: trying to reverse engineer the FTL. That will take a long time. Otherwise I think the port is in good shape.
00:42:56 Quit DaaT ()
00:42:58funmancan you somehow dump a full image of the flash to analyze it on your host ?
00:43:31denes_funman: no yet, because gevaerts has not hacked in the usb i think.
00:44:00funmandenes_: dany_21_a managed to dump the e200v2 ROM by connecting a pin he controlled (the button led) to its soundcard
00:44:01denes_funman: but i am afraid it has to be reverse engineered from the OF anyway.
00:44:14denes_funman: hehe :) nice
00:44:16funmanI see
00:44:35funmanhe has put some code he used in the sansav2 git repository if that can help you
00:44:42denes_funman: you are lucky, that the clip uses sd. there is no ftl.
00:44:56funmanmaybe a sdtl :P
00:45:08denes_funman: that's probably done by the hardware.
00:45:21funmanyes that's what I understand
00:45:25denes_funman: I mean I guess
00:45:34funmanI hope ^^
00:45:51 Join CaptainKewl [0] (
00:46:49denes_funman: the way I "dumped" the bootrom of another mp3 player, was to display it on the screen, and a colleague typed it in the computer :)
00:47:00denes_funman: only 4KB :)
00:47:18funmanouch you were motivated :)
00:47:59denes_funman: yes, but then I didn't find anything useable in the bootrom :)
00:48:20funmanI have another 'dead' mp3 player but I need to find a usable OF
00:48:37funmanthe one flashed on it only allow to use it as a USB mass storage - no more DAP features
00:48:53denes_funman: what player is this?
00:49:18funman bmp6607
00:49:33funmanplays 'amv' videos : a kind of mjpeg in avi
00:50:18denes_funman: ah this s1mp3, with the z80 core+dsp?
00:50:22funmanI have the Clip running at 240MHz now: display errors very fast !
00:50:35funmandenes_: no idea
00:50:51funmanwell, there's an Ali chip inside
00:51:18funmanI kind of suppose it was sold under different brands
00:51:30funmanbluesky is a 'discount' brand of carrefour
00:51:35denes_funman: you could take a look at
00:52:14 Quit bluebrother ("leaving")
00:52:52funmanI have browsed it already: the CPU is an ATJ
00:53:39funmanif I want to work on it ever again I need at least the OF ..
00:54:12denes_funman: you probably can't do much with it - if it really has a z80 with an undocumented dsp
00:54:52funmanat least I could do with the z80
00:55:52denes_funman: yes, play wav :) anyway, I gotta go to bed. so see you later.
00:55:57funmansee you ;)
00:58:04 Join goffa [0] (n=goffa@
00:59:13funmanit seems obvious that SanDisk uses the IDE and MemoryStick clocks for the pl180
00:59:36funmanas3525 datasheet mentions max clock frequencies of 90 and 40MHz respectively for IDE and MS
00:59:37 Join mc2739 [0] (
00:59:54funmanand the OF uses 128MHz and 80MHz
01:04:03 Quit Horscht (Read error: 110 (Connection timed out))
01:04:28funmannow I notice the OF enables the DMA controller - I wonder if that would help
01:06:21mc2739funman: just tested e200 build from svn - get this error -
01:07:11 Quit kugel ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
01:08:13funmanmc2739: did you select bootloader build?
01:08:54mc2739I had a good build with your earlier commits (r18900)
01:09:12funmanI suspect it is rev 18902
01:09:34mc2739maybe this -
01:09:36*linuxstb investigates
01:09:48funmanmc2739: yes exact
01:10:00 Quit goffa_ (Read error: 110 (Connection timed out))
01:10:16 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
01:14:00linuxstbmc2739: Comment out the HAVE_USBSTACK define in that file
01:14:27mc2739linuxstb: OK - brb
01:16:44mc2739taht worked - it builds now
01:21:28mc2739still no backlight or display activity
01:21:29 Quit mattgs (Read error: 104 (Connection reset by peer))
01:21:40 Join Avalon_Scythe [0] (n=18630b99@gateway/web/cgi-irc/
01:21:57Avalon_ScytheEllo Govna BRB
01:22:05linuxstbmc2739: Did you make the required change in the bootloader to attempt to enable the backlight?
01:23:26mc2739linuxstb: which change?
01:25:39 Quit perrikwp (" ajax IRC Client")
01:25:48JdGordonlinuxstb: from the apps/ level there should be no difference at all between the e200 v1 and v2 model?
01:25:56linuxstbmc2739: Something like this -
01:25:58funmanmc2739: some code with i2c
01:26:07JdGordonso all those lang changes may not have been needed
01:26:57linuxstbThe lang files use the model name, which is different.
01:27:31JdGordonoh, hmm
01:28:07Avalon_Scythehey can u guys help me, my Sansa e200 is in recovery mode and it has a curropted txt file with funny greek looking letters
01:28:24Avalon_ScytheCan You tell me How I can fix that txt File
01:28:24linuxstbJdGordon: But the lang file changes were trivial - just replacing "e200" with "e200*"
01:29:02 Part pixelma
01:29:23 Join pixelma2 [0] (n=marianne@rockbox/staff/pixelma)
01:30:20funmanis there any interest disabling temporarily interrupts if it's not to access some peripheral device ?
01:32:44mc2739still no go on backlight - never returns from lcd_init_device();
01:34:19linuxstbmc2739: Ah... Have you tried moving that backlight code to before the lcd code? You'll need to move as3525_codec_init(); as well.
01:34:39linuxstbi.e. right after system_init.
01:34:57 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
01:36:01linuxstbfunman: I'm not sure what you mean.
01:36:18funmanI'm not sure what I read in the disassembler either :)
01:37:40funmanI try to read step-by-step what the OF does from reset vector to nand initialization
01:37:47funmanthere is still some unanalyzed bits
01:38:22funmanand all these unanalyzed bits enable and disable interrupts a lot of times
01:38:39linuxstbfunman: BTW, the bootloader should call lcd_init(), which in turn calls lcd_init_device(), and lcd_clear_display, rather than calling lcd_init_device() directly. But it's not important yet.
01:38:53mc2739linuxstb, funman: we now have a backlight on e200v2
01:39:18 Join webguest06 [0] (n=5c03818d@gateway/web/cgi-irc/
01:39:35funmanare the lcd_write_*() functions written yet ?
01:39:50linuxstbmc2739: Nice.
01:40:17linuxstbfunman: Yes, but something seems wrong - mc2739 said that lcd_init_device never returns.
01:41:02funmanthe culprit seems to be _display_on()
01:41:50funmanis fb_data typedef always short ?
01:41:55Avalon_Scythecan someone help me
01:42:04 Join perrikwp [0] (i=d1a8d351@gateway/web/ajax/
01:42:08krazykitAvalon_Scythe, only if you ask a question
01:42:10funmanAvalon_Scythe: you could try google for "how to unbrick e200"
01:42:11linuxstbfunman: On 16bpp displays, yes.
01:42:19Avalon_Scytheno no no
01:42:40Avalon_ScytheYou See My Sansa e200 is in recovery mode and it ahs a currpoted txt file
01:42:49Avalon_Scythecan you show me how to fix it
01:43:07funmanlinuxstb: mc2739 the only problem I see would be waiting for a bit set in DBOP_STAT
01:44:19linuxstbEither that, or I do something so bad that the e200 dies...
01:44:24funmanlinuxstb: you don't enable the dbop clock
01:44:44 Quit webguest06 ("CGI:IRC (EOF)")
01:44:47linuxstbGuess that will help...
01:45:00funmanI don't know what is the exact setting for CGU_DBOP in e200
01:45:50 Join kkurbjun [0] (
01:46:57*Avalon_Scythe waits patiently for an answer
01:47:08funmanmc2739: what if you add a CGU_DBOP = (1<<3)|3; at the top of ams3525_dbop_init() ?
01:47:15mc2739Avalon_Scythe: have you searched the forum for "corrupt version.txt"?
01:47:45Avalon_Scytheum no
01:47:49Avalon_Scythei tried google leme check
01:48:10mc2739funman: where is that?
01:51:42funmanor just CGU_DBOP = (1<<3);
01:52:56 Join Darksair [0] (n=user@
01:53:50mc2739ok - CGU_DBOP = (1<<3)|3; gives noise on display, then mirrored sideways rockbox logo which then gets partially overwritten
01:54:15funmanwoah cool !
01:54:19funmanhow does it get overritten ?
01:54:47 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
01:54:47*rasher cheers linuxstb, funman and mc2739 on
01:54:50linuxstbfunman: Are you looking at the OF disassembly? I'm seeing CGU_DBOP = (1<<3) at 0x1184
01:55:12funmanwell |=
01:55:14linuxstb(but some code I'm not following before it)
01:55:19linuxstbfunman: Yes, that's what I meant.
01:55:20funmanbut the divider might have been set elsewhere
01:55:21mc2739funman: it is drawn rotated 90 degrees - the bottom of the logo overwrites the top
01:55:48funmanhorizontal versus vertical packing maybe
01:56:05funmanyou can change that, but I guess the functions need to be changed as well
01:56:46linuxstbmc2739: Have you re-run ../tools/configure today?
01:56:50kkurbjundoes anyone know how to check out a specific revision from SVN? Either my measurements are really off, or the gigabeat F's idle current draw has gone up 20 mA since I started measuring the current.
01:56:51funmanwhat about the colours ?
01:57:03funmankkurbjun: svn co -r
01:57:27mc2739yes, after r18906
01:57:32Avalon_ScytheOH I C mc
01:57:39mc2739colors look good
01:57:41Avalon_Scythedo u know were i can get a linux live CD
01:57:59linuxstbmc2739: Can you take a photo?
01:58:09funmanpic or it didn't happen
01:58:16kkurbjunfunman: great, thanks
01:58:20Avalon_Scythehey MC
01:58:32Avalon_Scythedo u know were i can get a linux live CD
01:58:45linuxstbThis should be relatively easy to fix - I expect Sandisk use the LCD in a different way to Rockbox, so we'll need to init it differently.
01:59:12mc2739yes I can get pic, but may take a bit - need to get camera from my brother
01:59:17Avalon_Scythehey ur name is linuxstb so u should know were i can get a linux live CD
01:59:21funmanlinuxstb: to be honest I don't remember where I saw the clock divider for dbop module
01:59:33funmanmaybe I just tried the lowest speed to be safe ;)
01:59:51linuxstbAvalon_Scythe: I've no idea, but I wouldd try entering "linux live cd" in google
02:00:58mc2739can the rotation and mirroring be fixed easily?
02:01:10 Join webguest404 [0] (n=d9e1eb8a@gateway/web/cgi-irc/
02:01:22Avalon_ScytheOH can u change windows TEMPORARILY tolinux
02:01:35Avalon_Scytheto linux* so u can get a linux live CD
02:01:40 Quit ameyer (Read error: 104 (Connection reset by peer))
02:01:43funmanAvalon_Scythe: I think it's a question for other channels
02:01:48webguest404hi I just downloaded a current build and it didn't have sudoku in the games app
02:01:57funmanmc2739: not that easily because the hardware is unknown
02:02:05linuxstbmc2739: I think it's "just" a matter of changing the inits for the LCD - and we have datasheets, so it shouldn't be a problem.
02:02:14funmanoh do you have ?
02:02:29 Join lupin_sansei [0] (
02:02:29linuxstbfunman: It's a known LCD controller - lots of Rockbox targets use it (iriver, iaudio, ipods...)
02:02:38Avalon_Scytheawwww damn,
02:02:46Avalon_ScytheI gota find someone with a linux
02:03:01funmanon #linux ?
02:03:21 Nick Darksair is now known as Darksair{away} (n=user@
02:03:23Avalon_ScytheOh hold on can u put a linux live CD in windows and it still works?
02:03:46lupin_sanseiHello, just installed RB on my 4th gen iPod and I love it, but when I do a db import it seems to get stuck at a certain point even if I leave it overnight. I suspect I have a dodgy hard drive. Would it be worth reformatting the drive in disk mode and trying again?
02:03:48krazykitAvalon_Scythe, this is not a linux channel. take it to ##linux or elsewhere
02:03:58webguest404about the sudoku game, is it me or is it gone ?
02:04:34kkurbjunwebguest404: sudoku is broken in SVN right now
02:04:37linuxstblupin_sansei: Have you tried checking it for errors - e.g. with "chkdsk" or "fsck.vfat" ?
02:04:55kkurbjunI'm not sure if it's missing from the build zips because of it
02:05:00 Quit Avalon_Scythe ("CGI:IRC (EOF)")
02:05:03webguest404ah, thanks. I'm not going mad :-)
02:05:17lupin_sanseilinuxstb: I tried the Windows gui around chkdsk (I assume that's what it is) and it did fix errors but BR still gets stuck
02:05:33 Quit webguest404 (Client Quit)
02:05:41 Quit bughunter2 ("bye")
02:06:43***Saving seen data "./dancer.seen"
02:07:59funmanlinuxstb: can you add the backlight code at the right location ?
02:08:29linuxstblupin_sansei: If I was you, I think I would reformat and reinstall - just to be sure that nothing is corrupted.
02:09:11linuxstbfunman: If I knew what that location was. I'm not sure where that as3525_codec_init should go - I'll have a look...
02:09:19lupin_sanseilinuxstb: okay I'm happy doing that. I basically just mount the ipod in disk mode and use windows to reformat don't I?
02:09:56linuxstblupin_sansei: How big is the disk?
02:10:05lupin_sanseilinuxstb: 20GB
02:10:09funmanwhy not in the bootloader ? I was thinking of something like backlight_on/off() to enable/disable
02:10:55linuxstbfunman: It's too specific - I think it should be hidden inside another init. Remember that all this code we're writing will eventually be used to build the main Rockbox.
02:11:37funmanwhat's the problem ?
02:12:14lupin_sanseirandom question - are there any commercial mp3 players which use Rockbox as their firmware?
02:13:36funmanlinuxstb: I see _backlight_on/off() for e200v1
02:15:29linuxstbfunman: Yes, that's where the actual backlight code will go - I'm just thinking about where that as3525_codec_init() call should go.
02:16:27cool_walking_lupin_sansei: As their stock firmware you mean? Nope.
02:16:38 Quit lupin_sansei (Read error: 104 (Connection reset by peer))
02:17:17funmanlinuxstb: i2c_init() is called in the bootloaders (main-pp for example)
02:17:28 Join lupin_sansei [0] (
02:17:37funman#if defined(SANSA_E200) || defined(PHILIPS_SA9200) i2c_init(); _backlight_on();
02:18:45 Quit homielowe ()
02:21:01 Nick Darksair{away} is now known as Darksair (n=user@
02:21:57linuxstbfunman: OK, I'll leave it there for now, and just implement the backlight functions - a copy/paste from the v1s for now, but maybe we'll need to try and unify the code a little. Other things are likely to be the same as well.
02:25:28funmanpower_off() is identical also
02:26:44funmanoh .. I wonder if some tuning of the voltages by i2c is needed for SD
02:27:54 Quit krazykit ("dinnertime")
02:29:05lupin_sanseithanks for the tips
02:29:06 Part lupin_sansei
02:30:48 Join Nibbl [0] (
02:34:57 Quit Darksair ("-_-")
02:35:51 Quit XavierGr ()
02:36:07 Join Darksair [0] (n=user@
02:43:09linuxstbfunman: backlight stuff committed (plus a little lcd cleanup)
02:43:16*linuxstb wonders if he'll see mc2739's photo tonight, or if he should go to sleep
02:44:05*funman can't go to sleep before the picture
02:44:18*linuxstb has a job
02:44:43funmanwith nice pictures as reward ? ;)
02:45:25linuxstbIf you want some nighttime reading, the LCD datasheet is linked from here -
02:45:36funmanno thanks ;)
02:45:44funmandoes lcd_init() call lcd_enable() ?
02:45:47 Join orly_owl_ [0] (
02:46:11funmanoh it was in the next chunk ;)
02:47:10linuxstbIIUC, the default state should be enabled when initialised. Also, HAVE_LCD_ENABLE doesn't seem to be defined for bootloader builds normally (for e200v1 at least)
02:48:19linuxstbMy guess (for the LCD) is that we should copy the e200v1 init more closely - that also appears to be using the same controller (but the update is done differently)
02:49:58funmanrequires some testing from e200 owners
02:50:05funmanthanks to you they have something to start from
02:51:02linuxstbOr possibly someone that knows these controllers (e.g. amiconn) will immediately tell us what's wrong (if we ever get a photo...)
02:51:35funmanmirroring should be easy to fix, rotation looks weird
02:51:53linuxstbI think that's just part of the init - you can configure the orientation.
02:52:57linuxstbBut I gave up commenting all the inits about half way through...
02:54:36 Quit midkay ("Leaving")
02:54:48funmanlook page 61 of the datasheet
02:56:26linuxstbThat's it ;)
02:57:26 Join Soap-phone [0] (n=4bdacff2@gateway/web/cgi-irc/
02:57:26funmanshould try 0x38 for a start
02:57:30funmanmc2739: still around ?
02:58:24*funman wonders why all e200v2 owners are only available for testing when there is nothing to test
02:58:40linuxstbMaybe those writes in lcd_update() and lcd_update_reg() aren't needed. I think they're just copied from the iaudio x5 or h300 driver.
02:59:20funmanI think a write modifies the start address
02:59:36 Quit orly_owl (Read error: 110 (Connection timed out))
03:00:42linuxstbI think I'll leave you to it - hopefully a tester with a camera will appear soon, and I'll wake up to a nice photo.
03:00:58 Quit Soap-phone (Client Quit)
03:03:07funmanoh .. the OF definitely messes with i2c
03:04:44 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
03:05:26 Nick orly_owl_ is now known as orly_owl (
03:09:25funmanoh it's probably due to the code using R_ENTRY_MODE_*VERT
03:09:41*funman understands what linuxstb meant
03:10:01 Quit funman ("leaving")
03:11:06 Quit tarbo (Remote closed the connection)
03:11:25 Join tarbo [0] (n=me@unaffiliated/tarbo)
03:12:54 Quit synergist ("changing servers")
03:14:28 Join BiptoN [0] (n=62948a24@gateway/web/cgi-irc/
03:14:34JdGordonoh bloody hell... missed them by 5 min :(
03:16:00linuxstbNot quite...
03:16:18linuxstbJdGordon: Still here?
03:16:51linuxstbLet me commit one change to SVN, then you can test.
03:18:10linuxstbOK, hopefully you should now see something on the LCD.
03:18:22linuxstbHave you built for the e200v2 before?
03:18:51JdGordonyeah, once, but never got it installed yet though because my sansa and comp dont like eachother
03:18:57JdGordonhave to use my laptop to install...
03:19:10linuxstbOops, my commit failed - update again now (if you already have)
03:19:47JdGordonline 97?
03:20:00linuxstbYes, the addition of CGU_DBOP
03:20:17JdGordontarget/arm/as3525/sansa-e200v2/lcd-e200v2.c:97: error: stray ‘\194’ in program
03:20:20linuxstbOops, seems funman already comitted....
03:20:50 Join synergist [0] (
03:21:06*linuxstb is having a bad committing day
03:21:36JdGordon2am? bad yesterday or bad tomorow? :D
03:22:57JdGordonwierd.. bad whitespace or something
03:22:59linuxstbSeems to be both.
03:24:26linuxstbYes, it's a dodgy character added by funman - between the "|" and the "(4-1)"
03:24:53JdGordonok, got it built, 1 min
03:26:07*JdGordon curses technology
03:27:33mc2739linuxstb: picture posted (not a very good one though) -
03:28:50linuxstbmc2739: Hopefully we can improve that...
03:29:50JdGordon... nothing is happening... ?
03:31:08JdGordoni can still get into the of though so not completly dead
03:31:17JdGordonbut yeah, black lcd, no backlight
03:32:03JdGordon"mkamsboot e200pA.bin.of bootloader-e200v2.sansa e200pA.bin" is correct yeah?
03:32:29linuxstbYes. If dual-booting works, then the mkamsboot part was fine.
03:32:30mc2739JdGordon: yes, that looks correct
03:33:52linuxstbmc2739: What was the line you added to lcd-e200v2.c ?
03:34:10linuxstbWas it "CGU_DBOP |=" or "CGU_DBOP ="
03:34:13mc2739linuxstb: CGU_DBOP = (1<<3)|3;
03:34:42linuxstbOK, so same as SVN...
03:36:05linuxstbJdGordon: Could you try commenting out the calls to lcd_init() and show_logo() in bootloader/sansa_as3525.c and see if the backlight comes on?
03:37:30linuxstbmc2739: Do you still have the working code? (i.e. you haven't updated to current SVN) ?
03:38:46mc2739yes, still at 18906 - haven't changed anything since it worked
03:38:50JdGordonstill nothing
03:38:59linuxstbmc2739: Good. In the lcd_update() and lcd_update_rect() functions, can you replace "R_ENTRY_MODE_SOLID_VERT" with "R_ENTRY_MODE_HORZ" ?
03:40:27linuxstbJdGordon: OK, then the final thing (to change back to the code that works for mc2739) is to add #include "as3514.h" to the bootloader, and then replace the call to _backlight_on() with "as3525_codec_write(AS3514_DCDC15, 6);" (keep the lcd stuff commented out for now)
03:41:47JdGordonhmm odd... put it back on and yes i have backlight
03:42:24linuxstbSo calling as3525_codec_write() works, but _backlight_on() doesn't?
03:42:51JdGordonno, i tihnk it was the reconfigure i just did... uncommenting out the l\cd_init line now
03:43:12linuxstbAH yes, there were changes earlier today to the configure script.
03:43:24JdGordon:( that would explain it then.. sorry about that
03:44:16JdGordonrotated 90 deg, and backwards!
03:44:21linuxstbPerfect ;)
03:44:22mc2739linuxstb: logo is not mirrored or rotated now
03:44:44linuxstbIs there a "but", or does that mean it's perfect?
03:45:10 Join goffa_ [0] (n=goffa@
03:45:24mc2739but it looks like it draws 176 pixels and then starts at the top again and overwrites
03:46:12mc2739the logo looks exactly square with noise at the bottom
03:47:10mc2739like the picture, but with the correct orientation
03:47:33linuxstbOK, I definitely need to go to bed now. I'll leave it up to you two to continue if you want - the datasheet is here...
03:48:02mc2739g'nite - good work so far
03:48:24JdGordonwee,, gnight
03:52:19JdGordonshould I commit that change to get it rotated correctly?
03:53:34 Nick fxb is now known as fxb__ (
03:54:51mc2739it will be one less thing to do lates
03:56:36 Quit goffa (Read error: 110 (Connection timed out))
04:06:02 Quit Acksaw (Read error: 104 (Connection reset by peer))
04:06:27*JdGordon thinks he should go over his lecture notes before that datasheet :'(
04:06:48***Saving seen data "./dancer.seen"
04:09:05JdGordon8-colour mode could be an interesting option to save power and have the display on
04:12:09 Quit Slack (Client Quit)
04:17:09 Join blkhawk- [0] (
04:17:48 Join Acksaw [0] (
04:24:32 Quit gromit` ("Coyote finally caught me")
04:24:51 Join gromit` [0] (
04:27:50 Quit reacocard (Read error: 110 (Connection timed out))
04:29:23 Nick Darksair is now known as Darksair{away} (n=user@
04:30:31 Quit mc2739 (""Zzzzz"")
04:33:05 Join miepchen^schlaf_ [0] (
04:34:23 Quit blkhawk (Read error: 110 (Connection timed out))
04:35:00 Join reacocard [0] (n=reacocar@WL-112.CINE.HMC.Edu)
04:35:07 Nick blkhawk- is now known as blkhawk (
04:42:45 Quit Nico_P (Remote closed the connection)
04:50:54 Quit miepchen^schlaf (Connection timed out)
04:56:24 Quit BiptoN ("CGI:IRC (EOF)")
05:08:25 Nick Darksair{away} is now known as Darksair (n=user@
05:32:26 Quit Zarggg ()
05:32:52 Join gregorovius [0] (
05:34:31gregoroviussomeone told me the other day they were selling gigabeat F players, but I can't remember who has logs if it was in the channel
05:36:03 Join nuonguy [0] (
05:36:33gregoroviusoh, thanks jd
05:40:53 Quit saratoga ("CGI:IRC (EOF)")
06:06:51***Saving seen data "./dancer.seen"
06:25:16 Join goffa [0] (n=goffa@
06:27:26 Join homielowe [0] (
06:28:26 Quit gregorovius ()
06:30:03 Join aarcane [0] (
06:31:55 Quit aarcane (Client Quit)
06:32:32 Join mc2739 [0] (
06:32:41mc2739JdGordon: ping
06:34:14mc2739here is a fix for the e200v2 display -
06:34:33JdGordoncan you dcc me the diff? my net is fubar
06:34:41JdGordonor direct link the diff?
06:35:04mc2739how about pastebin?
06:35:13JdGordonpastebin sometimes works
06:35:26JdGordonim capped atm, and flyspray almost always times out
06:35:48JdGordonah dw, worked
06:37:07JdGordonhehe, silly change :p
06:37:38mc2739lcd_update_rect() probably needs the same change
06:38:26 Quit goffa_ (Read error: 110 (Connection timed out))
06:38:50mc2739but it looked a bit too complicated for me to figure out tonight
06:40:47JdGordonwhy is the button light code in the lcd code?
06:40:52JdGordonor just hasnt been moved yet?
06:41:14JdGordoncool, its using the whole screen now, but the imge is still fubar?
06:42:04JdGordonbassically a yellow rectangle at the top?
06:42:15JdGordonits correct for you?
06:42:35mc2739button light code was there - just #if 0'd
06:43:02JdGordonoh yeah..
06:43:10JdGordonis the lcd correct for you?
06:43:19mc2739its correct on mine - rockbox logo - version number on bottom - ID: on top line above logo
06:43:45JdGordonok, ive done a make clean and trying again
06:45:48JdGordonweee :) all working
06:46:01JdGordonill fix that button light code and commit
06:46:17mc2739ok - thanks
06:46:41JdGordonare both lines needed each time its toggled?
06:48:13mc2739probably yes
06:48:35JdGordonone way to find out :p
06:49:14mc2739you can just toggle bit 7, but if the direction gets changed elsewhere, it probably won't work
06:49:43JdGordonwhy would the direction change?
06:50:13 Quit nuonguy ("This computer has gone to sleep")
06:50:27mc2739I don't know that it would
06:50:50JdGordonisnt the calls backwards? shouldnt the direction be set before setting the value?
06:51:38mc2739thats what I would think, but it works both ways
06:52:05JdGordonok, its all working
06:57:58JdGordonmc2739: seems to work
06:58:22JdGordonreplcaing lcd_update() with lcd_update_rect() doing the full screen... probably not the best way to test though?
07:04:42mc2739I'm get Hunk failed - on lcd-e200v2.c - at 364 and at 408
07:06:20JdGordondw, its working
07:06:35JdGordoni think...
07:10:03JdGordonyeah, it works :) in it goes
07:12:00mc2739 has an #if 0 around the some of the code
07:12:26JdGordonyeah, cleaned that up.. that was just for testing _rect
07:12:45JdGordonthe gram doesnt get cleared on bootup...
07:12:50 Quit reacocard (Remote closed the connection)
07:13:54mc2739maybe it should be cleared before the backlight gets turned on
07:14:29 Quit BHSPitMonkey (Remote closed the connection)
07:15:24JdGordonwierd.. ok, apparently it should be.. probably ignore everything i just said about it not clearing and put it down to how i tested update_rect
07:16:04mc2739ok - ignoring
07:17:42mc2739looks good - I'm off to bed again
07:18:36 Quit mc2739 ()
07:28:59 Join JdGordon|uni [0] (i=d31c9189@gateway/web/ajax/
07:29:00 Quit CaptainKewl (Read error: 110 (Connection timed out))
07:29:34 Quit JdGordon (Read error: 113 (No route to host))
07:29:40 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
07:29:47 Quit JdGordon|uni (Client Quit)
07:30:26JdGordonwhat does target/arm/ffs-arm.S:46: Error: selected processor does not support `clz r0,r1' mean?
07:30:35JdGordonother than what it says....
07:32:24 Join reacocard [0] (n=reacocar@WL-112.CINE.HMC.Edu)
07:33:09 Join AndyIL [0] (i=AndyI@
07:36:48 Quit reacocard (Remote closed the connection)
07:38:44 Quit advcomp2019 (Read error: 104 (Connection reset by peer))
07:38:58 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
07:40:19 Join reacocard [0] (n=reacocar@WL-112.CINE.HMC.Edu)
07:44:14 Quit AndyI (Read error: 110 (Connection timed out))
07:44:48 Quit parafin ("So long and thanks for all the fish")
07:45:00 Join parafin [0] (i=parafin@
07:45:24 Quit parafin (Client Quit)
07:45:37 Join parafin [0] (i=parafin@
07:46:11 Quit parafin (Client Quit)
07:46:17 Join parafin [0] (
07:52:07amiconnJdGordon: It means exactly what it says... the selected arm processor doesn't support the 'clz' instruction
07:52:22amiconnFor which target do you get this error?
07:52:46JdGordone200v2, thought id try a regular build../. its needed for priority scheduling apparently
07:52:56JdGordonso i gave up and back to playing in the bootloader
07:53:10amiconne200v2 is ams?
07:55:56amiconnIt has an ARM922T core, which is armv4
07:56:15amiconnSo the ARM_ARCH macro for the e200v2 is wrong
07:56:44*JdGordon back in 10
07:58:35 Quit AndyIL ()
08:02:15 Quit miepchen^schlaf_ ()
08:06:52***Saving seen data "./dancer.seen"
08:07:48 Join AndyI [0] (i=AndyI@
08:09:35*JdGordon back
08:11:44JdGordonif a gpio direction bit is set its output yeah? (on the ams)
08:12:26amiconnJdGordon: I've just committed the ams arm version fix
08:14:29*JdGordon thinks he is reading the comments in the m200v2 button driver wrong
08:17:09*amiconn glares at the build table
08:19:37JdGordonis there something more I need to do to read a gpio pin than setting the DIR bit?
08:20:13 Quit parafin (Remote closed the connection)
08:21:08 Join gromit`` [0] (
08:21:18 Join parafin [0] (
08:22:11*amiconn summons domonoky
08:26:47 Join bertrik [0] (
08:26:48JdGordonwow! thats alot of red :D
08:27:50 Quit cool_walking_ (" ajax IRC Client")
08:28:27amiconnYeah, domonoky goofed
08:28:36*amiconn ponders reverting for now
08:28:54amiconnSeems he didn't commit all necessary changes
08:29:09 Quit gromit` (Read error: 110 (Connection timed out))
08:29:37bertrikwow, e200v2 displays works! :)
08:29:42amiconnHmm, build is hanging *again*, grr
08:30:04JdGordonwhy did the red start on rashers commit though?
08:30:16JdGordonbertrik: yeah :D im trying to get buttons reading now and not having much luck
08:30:51 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
08:31:54amiconnJdGordon: Because some commits before that didn't trigger a build
08:31:57JdGordonbertrik: is what i was hoping would work (with calls to button_init and button_read_device in the bootloader, but nothing
08:32:17amiconnSomething is borked on the build master it seems :/
08:35:43 Quit BigBambi (Read error: 113 (No route to host))
08:36:04bertrikJdGordon, that should do it. I don't know how the hardware mappings table was filled for the e200v2, maybe a mistake was made when analysing it from the OF.
08:37:07JdGordonhow do I read the value of each gpio?
08:37:24JdGordonin one hit.. instead of each pin
08:38:03bertrikread (gpio base + 0xFF << 2) IIRC
08:42:08JdGordongcc wants brackets...
08:42:31JdGordon(gpiobase+ff)<<2 of gpio+(ff<<2)?
08:42:46bertrikthere's also an AFSEL register for each GPIO which can select an alternative function for the pins of a GPIO port. You should disable the alternative function, but I don't know if that's going to interfere with the dbop/display
08:43:06bertrikgpio_base + (0xFF << 2)
08:45:03*amiconn would just write that as gpio_base + 0x3fc then...
08:45:44bertrikJdGordon, do you have the datasheet?
08:46:21*JdGordon shouldnt be fiddling untill after exams :'(
08:46:42 Join Rob2222 [0] (
08:46:44JdGordondinner time :)
08:48:15 Quit bertrik ("Leaving")
08:48:49 Join ender` [0] (
08:48:49 Join avis [0] (n=ident@pdpc/supporter/student/avis)
08:49:17 Join Bagderr [241] (n=daniel@rockbox/developer/bagder)
08:49:42 Nick Bagderr is now known as B4gder (n=daniel@rockbox/developer/bagder)
08:50:16amiconnB4gder: Something seems to be wrong with the build master. Not picking up commits, hanging builds, ...
08:50:35*amiconn is about to fix domonoky's red
08:51:20B4gderit's building right now
08:51:28 Quit Rob2223 (Read error: 60 (Operation timed out))
08:52:31amiconnYes, but that's a restarted build which is already overdue again
08:52:38 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
08:52:49B4gderprobably one of the servers that are borked
08:53:38amiconnBut yesterday it missed at least one commit
08:53:51B4gderwell that's hardly a new bug...
08:54:06B4gderand nothing has been changed in that area for a very long time
08:54:10 Quit avis (Remote closed the connection)
08:55:04 Join culture [0] (
08:56:09amiconnB4gder: Maybe, but it seems to happen more and more often recently
08:59:05 Quit AndyI ()
09:01:09 Quit Nibbl (Read error: 110 (Connection timed out))
09:01:42 Join Nibbl [0] (
09:02:41amiconnNow there are gaps in the current row of builds... and seems to have diskspace problems
09:04:03cgseems like has disk full
09:04:24LinusNamiconn: oops
09:05:18 Join BigBambi [0] (i=86ceaf40@gateway/web/ajax/
09:06:20*amiconn confused domonoky and bluiebrother earlier
09:06:54 Join AndyI [0] (i=AndyI@
09:06:55 Join dany_21a_ [0] (
09:18:57 Quit GodEater (" ajax IRC Client")
09:19:02 Join GodEater [0] (i=c2cbc962@gateway/web/ajax/
09:23:29 Quit culture (Connection timed out)
09:25:07GodEaterDid anyone get around to adding that CPPFLAGS=-U_FORTIFY_SOURCE fix to yet ?
09:27:05B4gderah latest-ubuntu-fix wasn't it?
09:27:29GodEaterwe've just had a post in the forums about it as well
09:27:39GodEaterso I'm guessing we're going to get more and more until we commit a fix
09:28:04*JdGordon sends a virtual 6-pack to LinusN
09:29:24BigBambiGodEater: wasn't there another one for 64 bit?
09:29:51 Quit jhulst (Read error: 113 (No route to host))
09:29:55B4gder-U_FORTIFY_SOURCE is for the gcc build, right?
09:30:16LinusN"Gammar correction" - priceless :-)
09:30:27B4gderhaha, I liked that one too
09:30:40BigBambiI didn't even notice until someone pointed it out :)
09:30:47GodEaterBigBambi: you tell me - I've not come across one since I'm not running a 64-bit build at home
09:31:08BigBambiGodEater: I thought I remembered you discussing this when it came up first
09:32:40*B4gder committed an attempt
09:33:55BigBambiI can't really tell from that if -fno-stack-protector ended up being required or not :)
09:34:37 Join homielowe_ [0] (n=homielow@
09:34:52 Quit kachna (Read error: 113 (No route to host))
09:35:15GodEaterand nor could I, since it wasn't me using it
09:37:16BigBambiI know, I just thought you might have done, sorry
09:37:52GodEaterit's ok - I just have no idea whether it's required or not - so couldn't possibly tell whether it should be commited :D
09:41:34 Quit homielowe (Read error: 110 (Connection timed out))
09:42:06JdGordonis there a lofi version of flyspray? im getting timeouts maybe 50% of the time (because my internet is capped to rediculous speeds)
09:43:08 Join BigBambi_L [0] (i=86ceaf34@gateway/web/ajax/
09:43:41B4gderI don't think there is
09:45:48JdGordonoh well
09:48:06B4gder21 fine build servers again
09:48:49*B4gder spots a 2-second build in there...
09:49:04B4gderthat's a pretty fine build time
09:53:03GodEaterbest I managed is 8 seconds it would seem
09:54:17 Join TheArchitect [0] (n=447b9fdd@gateway/web/cgi-irc/
09:55:39 Quit Darksair ("To Arch or Gentoo? That is the question...")
09:56:01 Quit TheArchitect (Client Quit)
09:56:14 Join Darksair [0] (n=user@
09:58:58 Quit advcomp2019 (Read error: 104 (Connection reset by peer))
09:58:59 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
09:59:42 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
10:03:30 Join petur [50] (n=petur@rockbox/developer/petur)
10:05:40 Join lasser [0] (
10:06:55***Saving seen data "./dancer.seen"
10:07:00 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
10:09:24JdGordonis crossfeed something that once you have it configured you either have it on or off? you wouldnt change the options often?
10:13:34 Quit BigBambi (" ajax IRC Client")
10:14:51LloreanMaybe if you bought new headphones (swapping from IEMs to over-the-ear, or whatnot)
10:14:58LloreanBut generally only on/off.
10:15:14JdGordonok, thanks
10:15:34 Quit perrikwp (" ajax IRC Client")
10:15:39JdGordonthe EQ shuold be used the same yeah? but probably isnt?
10:15:57LloreanYeah, the EQ is used in all kinds of crazy ways
10:17:49LloreanPeople even request that it automatically load eq files based on genre ID3 tags...
10:19:27*JdGordon wonders why list acceleration settings are under scrolling options?!
10:20:25LloreanIf I had to guess, because people constantly look there for list scrolling, rather than line scrolling.
10:20:58JdGordoneven if its under display?
10:21:16*JdGordon is being a masochist and going to have a go at reordering the menu to make sense
10:22:24LloreanWell, one could consider that a "display" option if one hammered a square peg into a round hole long enough.
10:22:59JdGordonit almakes more sense going under scrollbar options
10:23:07JdGordonalmost makes even
10:28:39 Quit homielowe_ (Read error: 110 (Connection timed out))
10:29:07 Quit dany_21a_ (Read error: 54 (Connection reset by peer))
10:29:28 Join dany_21a_ [0] (
10:39:16 Join grndslm [0] (
10:43:28 Join kachna [0] (
10:46:06JdGordonchanel configuration and balance... again not usually changed unless the headphones/speakers change?
10:46:55JdGordonand whats dithering?
10:47:15GodEaterI wondered that the other day too
10:47:20GodEaterI'm still none the wiser
10:47:38GodEaterMind you, I understand less than 1% of all our sound based options. I leave them well alone.
10:48:14GodEaterI had a play with stereo width the other day. Made no difference that I could hear at all
10:49:04 Join maha [0] (i=juma@gateway/gpg-tor/key-0xB9002659)
10:49:11*JdGordon doesnt actually care what it does.. just if its a set and forget setting
10:49:42*GodEater is more of the opinion if it doesn't do anything, why have it at all
10:50:01crwlwas it so that stereo width doesn't do anything if you don't set channel mode (or whatever it was) to custom ?
10:51:38LloreanJdGordon: Dithering can, apparently, improve the apparent audio quality of low sample rate files.
10:52:04JdGordonso you may want to change it frequently?
10:52:10LloreanI don't know.
10:52:39LloreanI *think* it's set-and-forget
10:52:43Lloreanbut I'm not sure.
10:59:08 Quit kachna ("Konversation terminated!")
10:59:17 Join DerDome [0] (n=DerDome@
11:01:19 Join n1s [0] (n=nils@rockbox/developer/n1s)
11:02:14n1slinuxstb: thanks for the grammar fix :)
11:02:43 Join tvelocity [0] (
11:02:46n1shaha, indeed :)
11:03:17linuxstbrasher suggested that I should say that I did it on purpose - I'm not arguing with him.
11:05:37JdGordonlets say the replaygain (as an example) options were split into all its settings and a single "enable replaygain" setting... which is better.. having the enable option twice, or only have the enable option in with the enable option for other not-set-and-forget-sound settings?
11:05:49*JdGordon thinks the second
11:06:19LloreanIsn't replaygain one of the *most* set-and-forget options?
11:06:46LloreanIn fact, I'm kinda surprised we offer the option to turn it off/on or anything beyond the "track gain" "album gain" "track gain if shuffled" options.
11:07:05LloreanI'd personally assume the tags presence means "use them", and that the songs would be replaygained to the level the user wanted anyway.
11:07:06 Quit jhMikeS (Nick collision from services.)
11:07:07JdGordonthat was a bad example then
11:07:12 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
11:07:21JdGordoncrossfade then...
11:07:27LloreanI see what you're getting at though.
11:07:41LloreanIf you're going to split "enable" away from "details", I wouldn't duplicate "enable"
11:09:39 Join Vodi [0] (
11:09:52n1sbtw, the various enable * settings are a mess of yes/no and on/off
11:10:10JdGordonthats easy to fix, just needs someone to do it
11:10:25n1syes of course
11:10:54linuxstbWhat about a new kind of settings screen, which has "enable/disable" at the top, and then the various settings underneath?
11:11:24JdGordonhow would that work with our limited screen space and buttons?
11:12:01JdGordonunless you mean that it should say if its enabled or not in the setting screen?
11:12:01linuxstbThe same way as currently. It would just be a list
11:13:05LloreanAnother option would be to get rid of all the explicit enable/disable settings, and make it implicit such as enabling replaygain by having "Replaygain Mode" include a "none" option.
11:13:36LloreanFor any that have multiple sub-options that is. As soon as you've picked non-null values for it, it's on.
11:13:41n1sLlorean: that would be bad for EQ for example
11:14:03Lloreann1s: There's a "flat" preset, isn't there?
11:14:13 Quit dany_21a_ (Remote closed the connection)
11:14:40n1syes, but then you lose your own eq preset unless you save it explicitly, seems inconvenient to me at least
11:15:18JdGordonLlorean: I dont like that.. what happens when your playing with options and want to disable it, but leave it as is so you can play later? saving the config and then manuaally resetting each option is too much hassle
11:15:57 Join dany_21a_ [0] (
11:16:02LloreanJdGordon: Well, having "Enable" not immediately visible when you're playing with the settings can cause added confusion when you change them, and don't hear the effect.
11:16:34JdGordonbut thats a seperate issue
11:17:01 Quit Darksair ("Having dinner...")
11:17:38LloreanI think implicit-enable allows you to both cut down the number of settings, and have users know that when they set a value it's just going to "take effect", rather than them having to also change some other setting to make that value actually be active.
11:17:56LloreanIt's not perfect for *every* setting (EQ might be a good example)
11:18:07LloreanBut I'm sure it's good for at least one, and there's probably others where it could work out.
11:19:12JdGordon"crossfade settings (currently disabled)" as the submenu title should be enough
11:23:47*JdGordon doesnt know if "play selected first" should be under file browser options or playlist r playback options :(
11:24:12JdGordonI'm thinking both it and auto-change directory should be in browser options
11:24:31rasherJdGordon: "crossfade settings (currently disabled)" won't fit on many screen/font combinations
11:24:48JdGordonunfrotunatly true
11:25:56Vodihey guys
11:26:07Vodiany developer of the "Cowon D2" here?
11:26:59BigBambi_LVodi: You are better just to ask. If people know, they will answer. IF not, people read the logs
11:27:31Vodiwhat's the status now of the flash driver?
11:27:36Vodiis something going on?
11:27:42Vodican i help testing?
11:28:00B4gderI don't think testing is what's needed atm
11:29:17Vodiso i have to wait and look at the forums?
11:29:34B4gderyes, or dive in and help it happen
11:30:48Vodisorry, i don't have expirience in hardware programming, only silly school-java-programming :(
11:32:37linuxstbI was imagining a first line in the settings menu would be an inline setting, something like "Status: Enabled" or even "Status: On"
11:34:39 Part Vodi
11:35:26JdGordon is the start of my reworking... it shuold be editable by anyone so if you want to contribute....
11:35:37JdGordonill move it to the wiki when its more finished
11:36:38 Join AJCantos [0] (n=d9141192@gateway/web/cgi-irc/
11:37:16JdGordonlinuxstb: I thought we didnt really like inline settings? also mixing inline and regular is bound to cause headaches
11:37:47linuxstbWhat kind of headaches?
11:38:30JdGordoni guess if its only toggle settings it would be ok
11:40:02linuxstbAnd who do you mean by "we"? Inline settings everywhere have lots of problems (but some people like them a lot), and I'm just suggesting a specific inline setting, which will contain the same text everywhere it's used.
11:40:33JdGordonhavnt inline settings been rejected before?
11:40:53JdGordonbut yeah, i see what your saying and it could work
11:41:32linuxstbI think they've been rejected, because they won't work well in general (for all languages, screen sizes, fonts etc).
11:41:43amiconnLlorean: I think the option to disable rg is necessary. What if a user who never wants rg (e.g. me) gets a track containing rg tags?
11:42:20 Join Darksair [0] (n=user@
11:43:47jhMikeSamiconn: How did AS3525 compile with ARM_ARCH set to 5? The clz instruction should have cause a problem.
11:44:06amiconnIt did, that's why I fixed it
11:44:31amiconnBootloaders didn't run into the clz issue, but JdGordon reported that a full build did
11:44:32linuxstbjhMikeS: Not much of Rockbox is being compiled for the AS3525 at the moment, just some target-specific drivers.
11:45:36Lloreanamiconn: If you don't want rg, do you think there'll be any noticeable effect if a track you get has rg tags? Generally speaking, in a mixed collection it'd be exactly the same as the varied levels which normal CDs might be mastered at.
11:45:41LloreanEspecially if the default is to "album gain"
11:47:45amiconnI don't care whether the effect is small or large. I want to be sure it has *no* effect
11:47:57LloreanWhy, exactly?
11:48:17LloreanEvery CD you own is mastered at a different level. Since you want RG off, you obviously don't care what that level is.
11:48:55n1sJdGordon: i'm not sure i understand your spreadsheet correctly but if you suggest putting shuffle/repeat/party-mode in sound settings I don't like that, they have nothing to do with sound also DSP options would be more confusing to users imho
11:48:59LloreanIf you're getting a random MP3 from the internet where you don't even control what tags are on it, surely you don't know that volume normalization wasn't applied before encoding either.
11:49:03amiconnI do care. I want it at the level it was mastered at.
11:49:43Lloreanamiconn: For what reason do you want it at that level?
11:49:50JdGordonn1s: no, they would either be seperate at the same level as "sound settings", or in a submenu at the same level
11:50:02JdGordonthey are attached to the bold text, not the sound settings menu
11:50:08amiconnUmm, because it was meant to be at that level?
11:50:33Lloreanamiconn: That's a pointless statement. The instant you change the volume from 0db, it's not at that level.
11:50:46LloreanI'm assuming you adjust the volume from time to time.
11:51:12LloreanIt was meant to be *stored* at that level.
11:51:21JdGordonupdated again... has every settings i think now
11:51:35LloreanThe level you listen to it at, on the other hand, is affected by your audio hardware anyway so amplification or reduction is expected.
11:52:08amiconnYes, but then I still get the differences between albums at their original levels
11:52:18 Quit DerDome ("Leaving.")
11:52:27rasherThis would make sense if you were listening to a collection of connected albums
11:52:48B4gder'Björn' seems to have been mangled a bit there in a few sources in jhMikeS' latest commit...
11:53:05Lloreanrasher: So, it's problematic if you're listening to connected albums, that have not been replaygained properly but have been replaygained.
11:53:32n1sJdGordon: why move time and date to the info menu? it is a setting.
11:53:33rasherLlorean: Don't ask me - I don't really understand the reasoning either
11:54:26GodEaterLlorean: what do mean "not been replaygained properly" ?
11:54:26amiconnBasically I want to have the choice to tell rockbox to not fiddle with stuff I don't want it to fiddle with.
11:54:35JdGordonn1s: have a look at the patch... bassically the point was to move all time date stuff together
11:55:04JdGordonn1s: the time/date stuff would come out of the rockbox info screen
11:55:04LloreanGodEater: If you have songs whose volume relative to other songs needs to be maintained, you make sure that all their "album" replaygain tags are the same by scanning them at once.
11:55:07n1sJdGordon: I'm just saying that i would be looking for settings in the settings menu...
11:55:11amiconnRockbox *is* about choices after all, and that's a choice affecting core functionality
11:55:16Lloreanamiconn: And you'd have that choice by not including replaygain tags...
11:55:44LloreanWe assume the presence of tags means "use them" for every other tag we support. There's not an option "skip Artist tags when indexing the database, and use folders instead"
11:55:50amiconnLlorean: Without that setting I can never be sure unless I check all my tags - rather cumbersome, I'd say
11:55:55JdGordonn1s: to be honest, I dont like the info menu anyway
11:56:04JdGordonnot sure where its options should go thoughg
11:56:14amiconnLlorean: All other tags don't affect the sound
11:56:19n1sJdGordon: maybe have a "Set alarm" thing that isn't in the settings menu?
11:56:46Lloreanamiconn: But they do affect other functionality. I suppose being about choices only means "choices about sound"?
11:57:11JdGordonn1s: move "Time & date" up into the main menu? (I'm trying to get support for that patch :p )
11:57:42 Quit larstobi (Remote closed the connection)
11:57:57 Join larstobi [0] (n=larstobi@
11:58:04n1sJdGordon: I like that better
11:58:05JdGordonthe idea of the new layout is settings which are not set/forget ones are no more than 3 away from almost anywhere in the UI
11:58:08Lloreanamiconn: There are tools that can pretty easily remove replaygain tags from a whole collection of audio. And if a song comes from a copyright holder with replaygain tags, then that's how it's *supposed* to be heard anyway.
11:58:13 Join moos [0] (
12:00:30JdGordona big downside to this is the backlight options are maybe too far down
12:02:04n1sI like your FS #9173 and would like that in either System or the main menu, does it change the "Set time and date" screen to look similar or is it still its ugly old self? :)
12:02:36JdGordoni dont remember, I think its still ugly, but changing the setting screen would become a priority
12:03:34 Join einhirn [0] (
12:04:23n1sJdGordon: I think you could flatten the system settings menu a bit, dropping the "hardware" menu and moving its submenus up
12:04:45 Part dany_21a_
12:05:45JdGordonI'd like to get brightness and contrast up higher... any ideas?
12:05:55 Join robin0800 [0] (
12:05:57n1salso the theme settings->colours menu seems to only have one submenu and nothing else...
12:06:23JdGordonthats a paste error i thnk
12:06:55JdGordonoh no its not...
12:06:56***Saving seen data "./dancer.seen"
12:06:56n1sah, yes there should be fg and bg too,
12:07:16JdGordonthey wernt in the wiki layout :p
12:09:25n1sno good ideas about moving contrast and brightness up...
12:10:01JdGordonive put then as the first options under settings/display
12:10:18JdGordonnot pretty, but better
12:11:59 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
12:13:30JdGordonupdated again
12:17:39JdGordonand again to put file view under display
12:17:58 Join {phoenix} [0] (
12:21:42 Quit linuxstb (Read error: 110 (Connection timed out))
12:22:38 Quit denes_ (Read error: 110 (Connection timed out))
12:22:53 Join denes_ [0] (
12:22:56n1sJdGordon: I think it's starting to look good but i have to run
12:28:24 Join kushal_12_27_200 [0] (n=kushal@
12:32:01 Join homielowe [0] (
12:53:16LinusNLlorean: still, rockbox should definitely have the option to ignore replaygain tags, as all other software that supports replaygain. i have the same files on several systems, and i might want to use replaygain on my xbox but not on rockbox. i don't want to be forced to remove all replaygain tags to be able to disable it. what if i later decide that i want it?
12:53:39 Join stooo [0] (
12:55:05 Nick fxb__ is now known as fxb (
13:13:55 Join pvbcharon [0] (n=charon@
13:22:22 Quit pvbcharon ()
13:25:27 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
13:25:52 Join BigBambi [0] (i=86ceaf40@rockbox/staff/BigBambi)
13:28:39 Quit amiconn (Nick collision from services.)
13:28:45 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
13:28:54 Quit pixelma2 (Nick collision from services.)
13:29:02 Join pixelma2_ [0] (n=marianne@rockbox/staff/pixelma)
13:29:04 Nick pixelma2_ is now known as pixelma2 (n=marianne@rockbox/staff/pixelma)
13:29:29 Join CaptainKewl [0] (
13:29:29 Quit CaptainKewl (Read error: 54 (Connection reset by peer))
13:36:33 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
13:41:27 Quit kushal_12_27_200 ("Leaving")
13:44:27B4gderthose v2 hackers have too crappy cameras! ;-)
13:45:42BigBambi"Dear Rockbox Fund. I need a DSLR to take photos of my v2" :)
13:47:58B4gder"dear Mr Somebody. Here's a few spare ones sent your way."
13:48:50n1sB4gder: what do you think of FS #9492? From what i can tell he is giving our buildsystem a much needed overhault but since i am pretty clueless i don't exactly know what cmake brings in terms of maintainability / issues and i remember you being sceptical
13:49:02*linuxstb wonders how we can know those screenshots aren't just from a v1
13:49:55B4gdern1s: I haven't checked it out really
13:49:58GodEatern1s: unless he's improved the patch in the last couple of days, it needs a LOT of work
13:50:04GodEaterit didn't build at all when I tried it
13:50:43n1sgodeater: he posted an updated one today, that is supposed to build a sim afaict
13:51:00Zagorit's also very annoying as a patch. he renames and moves around things, so it's hard to see what really changed
13:51:04B4gderI my eyes cmake is slightly problematic since we are exactly zero people with experience on it
13:51:27GodEaterthat's very true as well
13:51:30B4gderit moves much stuff from regular make magic to cmake magic
13:51:59Zagorwhat does cmake do better than make?
13:52:02B4gderI realize one argument is that we have long-standing make problems in our build system now
13:52:16n1sI think that sort of comments should go on the task so he knows how you feel and if it is unlikely to be accepted please tell him
13:52:26B4gderZagor: cmake is rather like configure+automake in a single package kind of
13:52:41Zagorn1s: you're right
13:53:00B4gderyes, but I still would like to try out a working approach to see what it's like
13:53:20B4gderI'm not shooting down the idea completely
13:56:00B4gderhe does change a lot of things in configure with no good reason
13:57:37*GodEater suggests asking him why he's done this :)
13:58:05B4gderyes, but I'd rather test it too and bundle all my comments at once
13:58:22B4gderI'll have a go later
13:58:25ZagorI'm primarily worried about adding a new layer of complexity on top of make. a layer nobody in the project is familiar with.
13:58:41B4gderinstead of make rather, this removes make I think...
13:58:47 Join midkay [0] (n=midkay@rockbox/developer/midkay)
13:58:51Zagorno, it creates Makefiles
13:59:30B4gdereven more like automake than I thought then
13:59:36linuxstbSo it creates "genuine" GNU Makefiles, rather than its own intermediate format?
13:59:56Zagorlinuxstb: as far as I understand, yes
14:00:20Zagor says "Simple configuration files placed in each source directory (called CMakeLists.txt files) are used to generate standard build files (e.g., makefiles on Unix and projects/workspaces in Windows MSVC) which are used in the usual way."
14:01:32 Join funman [0] (
14:01:47B4gderin this case I don't think they can be called "simple" :-)
14:02:22GodEaterit's make output is pretty colours!
14:04:02Zagordo we have a clear problem description? what is the real problem with the current build system?
14:04:20Zagoradding -pipe is fine, but that doesn't really address any core problem
14:05:00B4gderI think the main problems we have are some dependency ones
14:05:18B4gderie some file changes don't trigger rebuilds of other parts properly
14:06:21 Join PaulJam [0] (
14:06:29 Quit Nibbl (Read error: 110 (Connection timed out))
14:06:44funmanJdGordon: sorry for the weird space character - it happens very often with my french keyboard layout (altgr+space)
14:06:57***Saving seen data "./dancer.seen"
14:07:00ZagorB4gder: do you have an example I can peek at?
14:07:08funmanI'll look at preventing that with git precommit hooks
14:07:32B4gderZagor: not right now, although amiconn may have a few of them in memory
14:07:39n1sZagor: FS #6847 is an example
14:07:58Zagorn1s: thanks!
14:08:04n1salso i think we build bitmaps when making manuals
14:08:55B4gderI think some of the problems are due to the large amount of generated c files
14:09:11B4gderI mean, how we treat them
14:09:33n1snope, no bitmaps built with manual
14:10:36amiconnZagor: Building bitmaps and some bin stuff when building manuals or voices, not relinking codecs when a codec library changed, not recreating the self-extracting binary if only something in firmware/decompressor/ changes, ...
14:11:01amiconns/manuals or//
14:13:13 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
14:14:15*GodEater sees that the new patch still doesn't work
14:19:50Zagoruhm, how are voices built?
14:21:07B4gderselected voice then 'make voice'
14:21:14B4gderselect even
14:21:26B4gderI mean in configure
14:21:36Zagorah, gotcha
14:22:04robin0800GodEater: what patch is that ?
14:23:00B4gderFS #9492
14:23:39 Quit rvvs89_ (Remote closed the connection)
14:25:58JdGordonfunman: dw about it...
14:27:19funmanI have put "git diff-index -p $against −−|grep ' ' && exit 1" in .git/hooks/pre-commit
14:27:50BigBambi_Lfunman: don't worry
14:27:55 Join mofux [0] (
14:28:16mofuxis it possible to install rockbox on my ipod nano video 4gb?
14:28:26robin0800r18918 still will not compile on ubuntu 8.10 same errors as before!
14:28:28mofux(dunno exactly what generation this is, but could be 5?)
14:28:59B4gderrobin0800: what errors would that be?
14:29:05funmanmofux: if it's a quite large nano, it's 3rd gen and you can not use rockbox on it
14:29:38mofuxfunman: it could be this one:
14:29:44robin0800B4gder: it dosn't make correctly
14:29:58BigBambi_Lmofux: If it is any Nano other than 1st generation it won't work
14:30:00n1sB4gder: robin0800: IIRC you have to add -fno-stack-protector to CPPFLAGS on 64 bit ubuntu 8.10 too
14:30:17B4gderfor rockbox builds?
14:30:25BigBambi_Ln1s: Isn't that to build the cross compilers though, not to build?
14:30:33n1sno for gcc builds, maybe i'm confused
14:30:48funmanmofux: I don't know of nano 5th gen, google for "what is my ipod model" : there is a page on Apple website which list them all
14:30:54n1srobin0800: _what_ doesn't make correctly?
14:30:58robin0800n1s: not on 64 bit only 32 here
14:31:00B4gderI don't think we're getting details here so I don't know...
14:31:02BigBambi_Lfunman: The one I just linked? :)
14:31:23 Nick Darksair is now known as Darksair{away} (n=user@
14:31:26funmanyes - I can't stand the high speed of this channel !
14:31:34BigBambi_Lrobin0800: If you put the log on pastebin it would help
14:31:52mofuxfunman: accordingly to the apple page it must be a iPod nano 3rd gen
14:32:00mofuxhmmm, what a shame :)
14:32:02BigBambi_Lmofux: Then it won't work
14:32:02 Join rvvs89 [0] (
14:39:46robin0800BigBambi_L: running script now willpaste bin when finished
14:40:02GodEaterrobin0800: The cmake patch
14:40:15Zagorhmm, what is the benefit of using the $(call PRINTS) construct instead of simply having one echo line and one executing line?
14:40:26B4gderrobin0800: so you _are_ talking about ?
14:40:36B4gderZagor: cygwin craziness
14:40:48robin0800B4gder: yes
14:40:49B4gderweird stuff to make it run faster
14:41:07BigBambi_Lrobin0800: So why mention r18918?
14:41:20B4gderI suppose that was after my commit?
14:41:34*BigBambi_L slowly sees the light - he was trying the after Bagder's commit
14:42:05GodEaterB4gder: 18919 builds fine here, and I'm on Ubuntu 8.10
14:42:07BigBambi_LThat was confusing
14:42:50amiconnZagor: Significant speedup on cygwin (if using gnu make 3.81). 'echo' is a command call, 'info' is a built-in function of gnu make
14:42:56*BigBambi_L assumes robin0800 has build-essential installed? :P
14:43:39Zagoramiconn: and why don't we call info directly?
14:43:44JdGordonfunman: linuxstb: any suggestions on how to figure out the buttons? the gpio pins in the wiki dont seen to be correct
14:43:53amiconnBecause gnu make <3.81 doesn't know 'info'
14:44:32robin0800BigBambi_L: yes
14:44:48ZagorI'm not seeing any conditional use of info.
14:44:48 Part B4gder
14:45:43Zagori.e. this won't work on gmake < 3.81 anyway, will it?
14:45:44amiconnZagor: The trick is in the toplevel makefile generated by configure
14:45:44JdGordon is what i tried earlier (with button_init and read_device() calls in the bootloader of course)
14:45:44funmanJdGordon: put them all as input, and display their content in a infinite loop
14:45:59amiconnIt *does* work on gnu make 3.80, tried and tested
14:46:08Zagoramiconn: ah, found it. thanks.
14:46:47funmanJdGordon: if the wiki doesn't match, it's likely a keyscan like in m200 and clip
14:47:14funmani.e. if you don't write to the 'good' pin, the results will vary across different devices (like for my clip and linuxstb's one)
14:47:43JdGordonso how do we figure out the proper way?
14:47:53funmandisassembly :o
14:47:59*linuxstb thinks JdGordon has it easy now that there's an LCD to help debug things
14:48:02 Part stooo
14:48:24n1scongrats on the e200v2 lcd guys!
14:48:31*linuxstb starts to read the Fuze OF's LCD code
14:48:32funmanat least the power button should be on it's standalone gpio pin
14:49:07*funman wonders if JdGordon will post a picture of e200v2 LCD correctly rotated
14:49:16JdGordonis there really any point?
14:49:23JdGordonits excatly the same as a v1 :D
14:49:32JdGordonexcept it has the ams id at the top
14:49:36JdGordonok.. ok...
14:49:49funmanshort negociation ;)
14:50:33BigBambi_Lfunman: Those are some fine persuasive skills :)
14:50:43JdGordonmy camera isnt really good enough :p
14:51:11*JdGordon looks around for his card reader
14:51:36funmanJdGordon: also you could try the SD code on e200v2's SD slot
14:52:41funmanI'm still not sure how to handle dynamic base address
14:53:19JdGordonno photo tonight.. cant find my reader
14:54:54 Quit faemir (Read error: 110 (Connection timed out))
14:55:41robin0800GodEater: build failure on lunux 8.10
14:56:00JdGordonhmm... maybe not... my camera actually has a microSD chip in the adaptor! and i have a sansa cable!
14:56:45funmanrobin0800: can you paste line 51 and around of /usr/include/bits/fcntl2.h ?
14:58:33 Join J-23 [0] (
14:59:01GodEaterrobin0800: well that's definitely the error you get when CPPFLAGS hasn't been set properly
14:59:15J-23Hi! Was LCD code for e200v2 commited to SVN?
14:59:20robin0800funman: will have to run it again to do that
14:59:27JdGordonJ-23: twas
15:00:12funmanrobin0800: why ?
15:01:10*funman powers up the ubuntu 8.10 computer
15:03:54funman __open_missing_mode ();
15:04:22*n1s takes the plunge
15:05:22funmanif you use O_CREAT in open() you need 3 arguments
15:05:43Zagorapps/Makefile:97 explains the unwanted bitmap creation
15:05:44funmanso check what is at gcc-4.0.3/gcc/collect2.c:1580
15:06:41funmanI suppose it's a protection to not create files with unknown modes
15:07:08funmanand if ubuntu ships gcc 4.0.3 the patch is in the debdiff
15:07:10*J-23 tests
15:07:21 Join evilnick [0] (i=0c140464@gateway/web/ajax/
15:09:43J-23hmm, doesn't work
15:10:25 Quit PaulJam (".")
15:11:17funmando you mean no backlight at all ?
15:11:27linuxstbJ-23: Have you re-run tools/configure?
15:11:28 Join kachna [0] (
15:11:34 Quit robin0800 (Remote closed the connection)
15:11:35J-23linuxstb: yes.
15:11:39J-23funman: yes
15:11:39JdGordonZagor: if you're playing in the Makefiles, can you see if you can fix the problem where if a file is missing it doesnt complain about missing sysfont.h?
15:11:40linuxstbOr better, start from a clean build directory.
15:11:55J-23yes, I ran rm -rf * in build dir
15:12:18ZagorJdGordon: hmm? if any other file is missing, a missing sysfont.h is ignored?
15:12:44funmanthere is an entry about that in the bug tracker
15:12:55JdGordonif a file is missing the error is that sysfont.h cant be found (going from memory now), instead of saying that file is missing
15:13:49Zagorok. I'll take a look.
15:14:01*linuxstb notes that a way to get bugs fixed is to threaten to rewrite that part of Rockbox in a foreign language
15:14:43Zagorfrankly I wasn't aware that people felt the build system was so bad it needed a complete rewrite
15:15:04funmanthat's what they all say
15:15:57funmanI think it's a problem when the build system grows in complexity
15:16:20linuxstbAnother often-requested "feature" (there's even a current thread on the mailing list) is a way to quickly compile a single plugin, without doing a full build (I know about "make rocks", but that's still not really what people want)
15:17:43JdGordonwouldnt that be fixed if it didnt need to relink every .rock ever time?
15:17:54J-23s/@/!/. :)
15:19:05funmanwas robin0800 building the arm cross compiler?
15:19:19funmanJ-23: great ;) what did you need to change ?
15:19:33 Join fiXXXerMet [0] (
15:20:45J-23I don't know what helped, but I again removed everything from build dir and I recompiled (make clean and then make) mkamsboot
15:21:44 Nick Darksair{away} is now known as Darksair (n=user@
15:22:16linuxstbZagor: I don't think people have been saying the build system needs a complete rewrite - just that it is need of some love, as lots of bugs have crept in over the years
15:22:30linuxstb(apart from the person rewriting it of course...)
15:23:37Zagorwas there ever a bug filed for "voice builds bitmaps"?
15:24:15ZagorI can't seem to find one, anyway
15:24:47 Quit AJCantos ("CGI:IRC (EOF)")
15:25:35JdGordonthe build system doesnt really have bugs, just annoyances which untill now no one could be bothered trying to work out?
15:26:23J-23 - high resolution photos of LCD on e280v2
15:27:13BigBambi_LJ-23: And you know all about Zagor
15:29:58funmanwhy not putting the gcc patches in the svn repository ?
15:31:05Zagorbecause they are "files", not "code"
15:31:20funmanlike bitmaps?
15:31:38n1sfunman: they are not part of rockbox
15:31:38Zagorsorry, poor explanation. they are not our code.
15:32:33Zagorhaving them in svn would be using svn as a file serving tool. apache is a lot better at that. :-)
15:32:37fiXXXerMetI added a 4gb microsd card to my sansa e260r. I can mount the internal storage as /dev/sdb1, and dmesg says "[sdc] Attached SCSI removable disk" but I can not mount sdc1 or fdisk sdc... Any tips for accessing the extra storage?
15:33:01JdGordonthe e200 OF cant read microSDHC
15:33:01funman should fix the build of gcc on ubuntu (the build hasn't finished yet)
15:33:16BigBambi_LfiXXXerMet: Are you using the OF, or have you enabled Rockbox USB then compiled?
15:33:36Zagorfunman: iirc that error occurs in more than one place
15:33:38BigBambi_Loriginal firmware
15:33:47fiXXXerMetI'm using rockbox firmware
15:34:04BigBambi_LYou have enabled Rockbox USB in the code, then compiled?
15:34:09fiXXXerMetProbably not
15:34:11 Quit kachna (Read error: 110 (Connection timed out))
15:34:19BigBambi_LThen you are using the Sansa firmware for USB
15:34:33BigBambi_LWhich as JdGordon says, cannot handle SDHC
15:35:15 Quit rvvs89 (Remote closed the connection)
15:35:16BigBambi_LRockbox does not yet have a USB mode for the Sansas enabled as there are some problems. When you insert USB, it should reboot to the Sansa firmware
15:36:14fiXXXerMetI'm confused. The sandisk firmware can't handle the microsd cards? There is a slot specifically for them, and the documentation mentions using them numerous times
15:36:33BigBambi_LIt can handle SD only, not SDHC
15:36:43JdGordonfiXXXerMet: sdHC... 4+ gb cards
15:36:54fiXXXerMetoh, so it supports 2gb, not 4gb
15:37:05BigBambi_LLower than 2 GB fine, more than 2 GB not fine. Rockbox can handle both, but for now there is no USB mode
15:37:31BigBambi_LSo to use it, you need an external reader to add files, then Rockbox will use it just fine
15:37:53fiXXXerMetLuckily I have an external reader
15:38:01fiXXXerMetrockbox 3 doesn't support it yet either?
15:38:04BigBambi_LWhen the Rockbox USB is enabled, you will be able to do it all via Rockbox without an external reader
15:38:20fiXXXerMetAnd to do that I need to compile with the usb option?
15:38:23BigBambi_LNo, othewise I wouldn't keep saying Rockbox doesn't do it
15:38:34BigBambi_LfiXXXerMet: Yes, but it isn't recommended
15:38:55fiXXXerMetwhy not?
15:39:00BigBambi_LA bug in the SD card driver that is much more likely with USB can lead to corruption either of data or the filesystem
15:39:24fiXXXerMetso my best bet is to use the external reader?
15:40:21BigBambi_LI didn't of course mean SD card driver I meant flash
15:40:35BigBambi_LEither way −−> corruption
15:41:42funman*** buffer overflow detected ***: arm-elf-ar terminated
15:41:51fiXXXerMetThanks for the tips BigBambi
15:42:02funmanlooks like 256MB of RAM isn't enough to build a cross compiler !
15:42:25 Join rvvs89 [0] (
15:42:30BigBambi_LfiXXXerMet: No probs
15:42:43funmanand thinks everything succeeded
15:43:35funmanthere is no status checking
15:46:38 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
15:47:23Zagorsome build servers are causing build errors that I can't replicate...
15:47:53 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
15:47:59JdGordonis mine one of them?
15:48:30Zagorthere's an awful lot of them. must be my fault.
15:48:35funmanshouldn't use set -e ?
15:48:48funmanto leave the possibility to fix the build
15:49:55 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
15:53:41Zagorfunman: that's probably a good idea
15:54:16funmanor maybe an error variable to produce an error message
15:54:25Zagorotoh doesn't require bash
15:54:42ZagorI'm not sure how widespread -e is
15:55:03funmanI'll get the posix spec
15:56:08 Quit rvvs89 (Remote closed the connection)
15:56:28J-23so what's most important now on e200v2?
15:57:25funman : it's standard
15:57:28funmanJ-23: buttons
15:57:35 Join mf0102 [0] (
15:58:40J-23maybe we should create GPIO monitor like Debug->View I/O ports?
15:59:32funmanit's easy to do
16:00:39funmanproblem is if there is a keyscan routine like in other models it will be useless
16:01:36 Join culture [0] (
16:02:37Zagorlooks like a synchronisation issue. are the build servers running -j2 or more?
16:05:19n1sZagor: that's server specific
16:05:55*n1s uses make -j when building and would like it to continue to work :)
16:06:59***Saving seen data "./dancer.seen"
16:07:04rasherBagder: did you see the script? Especially the top bit, which should allow to use it for both daily and releases
16:10:11Zagorah, -j without a number causes it to fail
16:10:11J-23can SDRAM test use LCD to show results?
16:12:18 Join goffa_ [0] (n=goffa@
16:13:42 Join rvvs89 [0] (
16:14:19 Join meven [0] (
16:16:23 Quit BigBambi_L (" ajax IRC Client")
16:17:06 Join BigBambi_L [0] (i=86ceaf34@gateway/web/ajax/
16:22:56 Join meven_ [0] (
16:22:58 Quit goffa (Read error: 110 (Connection timed out))
16:27:11 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
16:31:33funmanBagder: why using FORTIFY_SOURCE ? I think it just enables unwanted extra sanity checks
16:31:47funmanthis is why the arm-elf-ar I just built doesn't run
16:32:04funmanoh .. it's undefined
16:36:14 Part LinusN
16:40:15 Join faemir [0] (
16:42:36 Quit meven (Read error: 110 (Connection timed out))
16:43:27 Join meven__ [0] (
16:44:15Zagorwhat a nasty problem
16:44:58saratogaif anyone gets a chance, could they profile AAC using the 192k test track from the wiki on Coldfire?
16:45:09 Quit n1s ()
16:45:21saratogai have the ARM results, and when I eventually get down to working on the codec, having the Coldfire ones would be nice too
16:45:28saratogano hurry though
16:50:18 Quit evilnick (" ajax IRC Client")
16:51:00funman < diff needed to build an arm-elf-ar which doesn't crash on ubuntu 8.10
16:51:43 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
16:52:11 Join perrikwp [0] (i=d1a8d351@gateway/web/ajax/
16:52:15amiconnZagor: Btw, a probably related problem is that the build system always tries to build the bitmaps twice
16:52:15Zagorfunman: excellent
16:52:45funmangcc seems to have built correctly with only the change in gcc/collect2.c
16:52:47Zagoramiconn: no the problem is that -j tries to compile the .c files that need the bitmap in parallell with creating the bitmap
16:53:14amiconnI once tried to fix that, but failed. When I removed what I thought would cause the double build, it didn't build the bitmaps at all (or at least not soon enough for the dependent steps)
16:53:25amiconnZagor: The effect happens without -j as well
16:53:46ZagorI don't see that it's doing it twice
16:54:08Zagorit's doing both "mono" and "native", is that it?
16:55:11Zagoralso I'm puzzled by why dep-apps has some dependencies without paths. and why make accepts that...
16:55:39funmanI'll send the patches on the ML
16:56:01Zagorsilly me, the pathless ones are the local files...
16:56:54 Join AJCantos [0] (n=d9141192@gateway/web/cgi-irc/
16:57:04 Quit AJCantos (Client Quit)
16:57:13Zagoryesss, adding "recorder/icons.h: $(BITMAPLIBS)" fixes the -j problem
16:57:39 Join AJCantos [0] (n=d9141192@gateway/web/cgi-irc/
16:57:49amiconnZagor: mono-native-mono-native
16:58:02Zagoramiconn: even with my latest commit?
16:58:03 Join ameyer [0] (
16:58:22Zagorthat's why I don't see it :)
16:59:42 Quit meven_ (Read error: 110 (Connection timed out))
16:59:43Zagortry this latest
17:00:38*pondlife picked the wrong time to build a D2 sim... ;)
17:03:42Zagorpondlife: did I break anything for you?
17:03:49Zagorlooks fine here
17:04:02pondlifeI was one revision behind, just building
17:04:05pondlifeSlooow Cygwin
17:04:35pondlifeI guess I might need to make veryclean?
17:05:54AJCantosHi guys, I'm having some problems executing RockboxDev.h, I get this error:
17:06:07pondlifeZagor: I'm still getting... make[1]: *** No rule to make target `rockboxlogo.h', needed by `/home/Steve/rockbox/buildsim/apps/bookmark.o'. Stop.
17:06:17AJCantosmake[3]: *** [crtbegin.o] Error 1, any idea?
17:06:26pondlifeThis is with r18922
17:06:27funmanAJCantos: ubuntu 8.10 ?
17:06:44AJCantosubunto 8.04.1-eee
17:06:48Zagorpondlife: and how did you run make?
17:07:07funmanAJCantos: paste the full error, including message before this last line
17:07:08pondlifemake veryclean, then configure, then make
17:07:55funmangevaerts: can you advice me how to write the SD/flash driver for sansav2 ?
17:07:56Zagorpondlife: the build servers just suceeded, so I don't think I'm at fault here
17:08:18gevaertsfunman: no. I know nothing about those
17:08:19funmanthe same driver should be used both for flash and SD; with the only difference being the register base address
17:08:21AJCantosIt goes /tmp/cc60t44W.s: Assembler messages:
17:08:22pondlifeProbably not, but I guess I have some cached bad build from a few revs ago
17:08:39AJCantosand then shows loads of messages of assembler commands
17:08:41Zagorgotta run
17:08:43funmangevaerts: I wanted to know if I could use multivolume API for that (IF_MV(blabla)
17:08:55funmanAJCantos: paste the whole thing on a pastebin
17:08:56 Quit Zagor ("Client exiting")
17:09:01gevaertsAh, ok. That's different :)
17:09:17AJCantosand then it shows:make[3]: *** [crtbegin.o] Error 1 make[3]: Leaving directory `/media/MUSIC/tmp/rbdev-build/build-gcc/ gcc' make[2]: *** [extra] Error 2 make[2]: Leaving directory `/media/MUSIC/tmp/rbdev-build/build-gcc/ gcc' make[1]: *** [stmp-multilib] Error 2 make[1]: Leaving directory `/media/MUSIC/tmp/rbdev-build/build-gcc/ gcc' make: *** [install-gcc] Error 2
17:09:41gevaertsfunman: do you have specific questions?
17:09:43AJCantosups, how can I do that pastebin thing?
17:09:47funmanthat way I could use in the code, if drive is 0, flash base address, if drive is 1, sd base address - does it seem sensible ?
17:09:52amiconnmeh, Zagor's gone :(
17:10:02funmanAJCantos: go to
17:10:24amiconnZagor (for the logs): Now the bitmaps aren't built twice, but instead they aren't rebuilt at all when _only_ a bitmap changed :(
17:10:24gevaertsThat makes sense, yes.
17:10:25pondlifeHmm, I have a rockboxlogo.h, but it's not being found
17:10:32funmangevaerts: thanks
17:10:38 Join Hillshum [0] (
17:10:47 Join blahrus [0] (n=blahrus@
17:11:14Hillshumhow does mkamsboot work now?
17:11:25amiconnI.e. same problem as with codeclibs, pluginlib, ...
17:12:30AJCantosI've pasted it, and now I guess I should give you the link:
17:12:34gevaertsfunman: it's probably a good idea to use the IF_MV* macros properly, so you can easily support both targets with and without SD slot
17:13:21pixelma2speaking of building the bitmaps - will the recent commits fix the problem with when changing an already used bitmap to a different size and you build without a "make clean" or manually deleting the old ones the buildsystem will use the old data of the bitmap and stuff it into the new size (IIUC) - so it'll look all garbled
17:13:44pixelma2? I hope someone understands what I meant...
17:15:23 Join IudeX [0] (n=52a0f8f7@gateway/web/cgi-irc/
17:15:27 Join kachna|lappy [0] (
17:15:59 Quit IudeX (Client Quit)
17:16:13 Join IudeX [0] (n=52a0f8f7@gateway/web/cgi-irc/
17:16:18 Quit meven__ (Read error: 110 (Connection timed out))
17:16:29IudeXAny progress in clip rockboxing?
17:16:34funmangevaerts: if the e200 only has 2 volumes (SD slot and flash), IF_MV() would be enough ? no need to use IF_MV2() ?
17:16:52funmanIudeX: not yet .. but I have found a new way to follow for the SD
17:17:02IudeXiudex: nice
17:17:11IudeXfunman: nice *ouch
17:17:12Hillshumfunman: some have 2 NAND chips?
17:17:27funmanHillshum: what do you mean ?
17:17:31IudeXoh i have to go out :/
17:17:33IudeXgood luck!
17:17:44Hillshumi think the e280 may have two NAND chips
17:17:49 Quit IudeX (Client Quit)
17:18:04amiconnpixelma2: No it won't - that's a different problem
17:18:14funmanHillshum: I believe they can be accessed by software as a unique volume
17:18:32gevaertsfunman: look at the macros. IF_MV() and IF_MV2() mean something else than that :)
17:18:52funmanah right ..
17:19:01Hillshumalso, how do i use mkamsboot?
17:19:10funman#define NUM_VOLUMES 2 : so there is no target with more than 2 volumes
17:19:17funmanHillshum: ./mkamsboot :)
17:19:51gevaertsNo. That NUM_VOLUMES handling is a bit (or a lot) unclean though, but don't bother about that too much
17:20:07funmanpixelma2: yes I think that's the goal
17:20:10HillshumI get this
17:20:20 Nick cg is now known as cg|aw (
17:20:25funmanAJCantos: strange, can you paste what's before ?
17:20:56funmanfirst unrecognized character '@' : @ means comment in gas syntax so that shouldn't be unrecognized
17:22:33amiconnfunman: You need IF_MV2 in most cases due to how cpp macros work
17:22:38 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
17:24:54 Join n1s [0] (n=nils@rockbox/developer/n1s)
17:25:53 Quit rvvs89 (Remote closed the connection)
17:26:41AJCantosI can't actually, because I'm not in that computer right now and that's the only thing I have
17:27:19funmandoes eeepc have an arm cpu ?
17:27:41HillshumXP runs on it so no?
17:27:59Tuplanollano, it has celeron or atom, depending on the model
17:28:10funmanthen why building gcc does assemble arm code ? o:
17:28:35linuxstbI assume it's building arm-elf-gcc
17:28:55funmanit's not a reason to build arm code
17:28:57linuxstbWhich requires building some arm libs
17:29:02AJCantosyes, that's the one
17:29:14funmanand crt*
17:29:55linuxstbIn fact, our script builds about 500 of the things, in different combinations of arm, thumb, big-endian, little-endian, different arm architectures, ....
17:31:06funmanI wonder if the intel assembler is called
17:31:37linuxstbLooks that way - which is normally caused by the arm binutils (which should be compiled and installed prior to building gcc) not being compiled and installed...
17:31:45linuxstb(or not being in the path)
17:31:54 Join rvvs89 [0] (
17:32:33 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
17:32:53 Quit BigBambi_L (" ajax IRC Client")
17:34:22 Quit BigBambi (" ajax IRC Client")
17:37:40AJCantosNow that you say so I haven't added anything to the path like /usr/local/arm-elf/bin or whatever, so maybe that's the problem?
17:38:03 Join AhtiK [0] (n=ahti@
17:39:02funmanAJCantos: were the binutils installed in that directory ?
17:39:09 Quit parafin (
17:39:09 Quit freqmod_qu (
17:39:09 Quit faemir (
17:39:09 Quit culture (
17:39:09 Quit Darksair (
17:39:09 Quit reacocard (
17:39:09 Quit kkurbjun (
17:39:09 Quit preglow (
17:39:09 Quit lastebil_ (
17:39:09 Quit Zambezi (
17:39:09 Quit BlakeJohnson861 (
17:39:09 Quit balou_ (
17:39:09 Quit linuxstb (
17:39:09 Quit jhMikeS (
17:39:09 Quit liiwi (
17:39:09 Quit maraz (
17:39:09 Quit lostlogic (
17:39:09 Quit Kopfgeldjaeger (
17:39:09 Quit FOAD (
17:39:09 Quit cg|aw (
17:39:09 Quit DataGhost (
17:40:15AJCantosI think so
17:41:00NJoinfaemir [0] (
17:41:00NJoinculture [0] (
17:41:00NJoinlinuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
17:41:00NJoinDarksair [0] (n=user@
17:41:00NJoinjhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
17:41:00NJoinparafin [0] (
17:41:00NJoinreacocard [0] (n=reacocar@WL-112.CINE.HMC.Edu)
17:41:00NJoinkkurbjun [0] (
17:41:00NJoinpreglow [0] (
17:41:00NJoinlastebil_ [0] (
17:41:00NJoinZambezi [0] (
17:41:00NJoinfreqmod_qu [0] (i=quassel@2001:700:300:1430:213:d3ff:fee9:5ed0)
17:41:00NJoinBlakeJohnson861 [0] (
17:41:00NJoincg|aw [0] (
17:41:00NJoinFOAD [0] (
17:41:00NJoinDataGhost [0] (n=dataghos@unaffiliated/dataghost)
17:41:00NJoinKopfgeldjaeger [0] (
17:41:00NJoinbalou_ [0] (
17:41:00NJoinliiwi [0] (
17:41:00NJoinmaraz [0] (
17:41:00NJoinlostlogic [50] (n=lostlogi@rockbox/developer/lostlogic)
17:41:05 Part pondlife
17:42:58 Quit mofux (Remote closed the connection)
17:47:12n1smarkun: what do you think of FS #9206 ?
17:49:22markunn1s: I can't look into it right now, but if it just fixes the missing glyphs I guess it's ok to commit, right?
17:49:52n1sit seems so to me but i know next to nothing about fonts :)
17:50:30markunI think rasher should take a look then
17:51:47AJCantosWell, thanks guys, I think you've made to the bottom of it, I'll try and add that info to the path and hopefully that would be it, thanks!
17:54:35 Join christophe [0] (
17:55:13GodEaterseems Bagder's commit this morning didn't help the gcc issue after all - I still had to specify CPPFLAGS= on the line calling for gcc to get built
17:56:04funmanGodEater: did you notice the patches I have sent ?
17:56:43 Nick J-23 is now known as Moarc (
17:56:49 Nick Moarc is now known as J-23 (
17:56:56GodEaterfunman: sorry no - been trying to get the bottom of my ftp issue :D
17:58:46domonokywow, lcd is working on e200v2.. congrats to all involved :-)
18:01:39 Quit AhtiK (Read error: 110 (Connection timed out))
18:06:26 Join jgarvey [0] (
18:06:26 Quit AJCantos ("CGI:IRC (EOF)")
18:06:47linuxstbdomonoky: What v2 device do you have?
18:06:56 Quit petur ("work->home")
18:07:02***Saving seen data "./dancer.seen"
18:07:37linuxstbAh yes. And there are a couple of Fuze owners around, so that just leaves the c200v2
18:07:42domonokyand i am awaiting a e200v2 from froobie
18:07:51 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
18:08:02scorchedomonoky: erm...about that...
18:08:30domonokyscorche: any problems with this ?
18:08:48scorchejoin -community
18:09:28 Quit mf0102 ("Ex-Chat")
18:11:03 Join arohtar [0] (
18:11:12 Join nuonguy [0] (
18:11:30 Join Strife89 [0] (n=michael@
18:12:29 Quit faemir (Read error: 104 (Connection reset by peer))
18:13:03funmangevaerts: in fact I'm not using IF_MV*() but something like that :
18:13:11 Quit {phoenix} (Read error: 60 (Operation timed out))
18:13:11 Quit yhuang ("Leaving.")
18:13:16funmanand a mandatory 'volume' argument to all the functions
18:13:58gevaertsSo how does fat.c know that?
18:14:25 Join {phoenix} [0] (
18:14:33funmanhm ok
18:15:21funmanmaybe a #ifndef HAVE_MULTIVOLUME \n int volume = 0 \n #endif
18:16:04gevaertsThat's probably the easiest way
18:16:34gevaertsMultivolume needs to be cleaned up though, so this can be talked about I think
18:16:35 Quit saratoga ("CGI:IRC (EOF)")
18:17:30*gevaerts wonders how much the bin+ram cost would be if we just dropped IF_MV* and made that argument mandatory
18:17:45 Join |J-23| [0] (
18:19:01 Quit J-23 (Nick collision from services.)
18:19:05 Nick |J-23| is now known as J-23 (
18:19:07n1sgevaerts: going over 4 args for a function is costly on some archs at least
18:19:19funmanalso if several targets use the sd protocol, a SD layer like in Linux could be worth it
18:19:20gevaertshm, ok
18:19:32 Join miepchen^schlaf [0] (
18:19:44 Join bughunter2 [0] (n=Jelle@
18:20:09*gevaerts managed to work on the ata reorg a bit today (train ride)
18:21:34gevaertsfunman: I think there are four sd targets in the tree (in various states of completion): sansa, sansav2, d2 and elio
18:22:38funmanshould I use 'drive' instead of 'volume' ? (like in the header)
18:23:35gevaertsI think so, yes. "volume" can also mean a separate partition (and HAVE_MULTIVOLUME has both meanings right now, depending on where you look in the code)
18:23:44J-23What's D2?
18:23:58gevaertsCowon D2
18:31:09 Quit linuxstb (Read error: 110 (Connection timed out))
18:32:07 Quit rvvs89 (Remote closed the connection)
18:32:21 Join rvvs89 [0] (
18:32:54 Join Xerion_ [0] (
18:33:43 Join {-phoenix-} [0] (
18:33:47 Quit arohtar (Read error: 110 (Connection timed out))
18:34:14 Join Lear [0] (i=chatzill@rockbox/developer/lear)
18:34:24 Quit {-phoenix-} (Client Quit)
18:34:40 Join {-phoenix-} [0] (
18:35:56 Quit {phoenix} (Read error: 104 (Connection reset by peer))
18:38:18 Quit pixelma2 ("-")
18:38:37 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
18:40:37 Quit bmbl ("Woah!")
18:41:51 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
18:43:46 Join Acky [0] (
18:44:23 Join m0f0x [0] (
18:48:23 Join faemir [0] (
18:48:35 Join Schmogel [0] (
18:48:47 Join kugel [0] (n=chatzill@unaffiliated/kugel)
18:49:18 Join AhtiK [0] (
18:53:52amiconnfunman: The IF_MV2 macro is a better readable method than ifdefing the function argument literally. That's why it was introduced - it does essentially the same
18:54:21amiconnAdding a mandatory 'volume' parameter for all targets is out of question, since it'd waste binsize
18:55:11funmanamiconn: what's your other solution ?
18:55:27 Quit rvvs89 (Remote closed the connection)
18:55:36amiconnWhat other solution? The current macros are working...
18:55:53 Join rvvs89 [0] (
18:56:03funmanwell .. for 2 different drivers maybe
18:56:14amiconngevaerts: Making the 'volume' argument mandatory could easily increase binsize by several KB for non-multivolume target, especially on arm and sh1 if a function is pushed over the magic "4 arguments" limit
18:56:34funmanit's only for sansav2 though, not all targets
18:57:11amiconnWe do need an abstraction layer if the volumes are handled by different drivers. That's what gevaerts is working on afaik
18:57:42amiconnBut the sansav2 doesn't need that afaik
18:57:49gevaertsfunman: ondio and sansav1 basically do the same thing as sansav2 as far as I can see
18:59:02amiconnAnd the IF_MV2 should be used for all multivolume targets in the respective drivers, in order to be able to compile a single-volume build for testing purposes
18:59:03 Quit GodEater_ (Remote closed the connection)
18:59:10funmanok .. I must take HAVE_HOTSWAP as HAVE_MULTIVOLUME ?
18:59:15amiconnThis has proven to be useful several times already
18:59:47amiconnNo, they're 2 different things. You can have a multivolume target without hotswap
18:59:54amiconn(though not the other way round)
19:00:33funmansee ata_write_sectors() in ata-sd-pp.c for example: is 'drive' declaration a bug ?
19:01:05funmandrive as an argument conditional on HAVE_MULTIVOLUME, and drive as a local variable conditional on HAVE_HOTSWAP
19:01:17 Quit Acksaw (Connection timed out)
19:01:21gevaertsfunman: that looks like a bug to me, yes
19:01:57amiconnThat's a bug, yes
19:02:58amiconnSee firmware/drivers/ata_mmc.c: ata_*_sectors() for an example how it's meant to be
19:03:55 Quit Xerion_ (Connection reset by peer)
19:03:57amiconn(yes I know that driver should be moved to target tree, and yes that driver is a bit special because it can (statically, at boot) select either internal storage or MMC when compiled without hotswap)
19:04:39J-23anything new to test on e200v2?
19:04:51amiconnWithout multivolume and hotswap I mean
19:05:08 Join GodEater_ [0] (n=ge@rockbox/staff/GodEater)
19:05:18funmanamiconn: for example TDR1 register is fixed (in sh7034.h) - in sansav2 case it's not : different for each volume
19:06:26 Join Thundercloud [0] (
19:06:45*amiconn wonders what this has to do with multivolume and/or hotswap
19:07:48funmanI use an addition to get the exact register address and this addition is useless without multivolume
19:08:44amiconnThat's why the Ondio driver uses 'const int drive' if multivolume is undefined
19:09:02 Join Badger [0] (i=badger@tremulous/player/badger)
19:09:17amiconnBy declaring it const, gcc is able to do the addition at build time and optimize the runtime code away
19:10:01amiconnThis trick is used in a number of places, not only related to multivolume (iirc it's used in the fat driver as well when building w/o fat16 support)
19:10:44 Quit christophe (Read error: 110 (Connection timed out))
19:11:07amiconnThis trick saves quite some #ifdefs without increasing binsize
19:11:38 Join christophe [0] (
19:12:51 Quit culture (Read error: 110 (Connection timed out))
19:13:35 Join bertrik [0] (
19:14:11funmanhello bertrik
19:14:39bertrikhi funman, read the good news this morning about the e200v2 display
19:16:12bertrikoh and nice to read that you accepted the commit offer
19:16:37funmanI wanted to know a bit more the project's insides
19:17:42funmanI had an idea yesterday: I was looking at the disassembly and I noted a lot of references to i2c before the code that inits the SD
19:18:04funmanmaybe some power need to be activated through the i2c bus
19:18:26 Part fiXXXerMet
19:19:11bertrikthe codec audio/PMU (or whatever it is called) is indeed responsible for the generation of some internal voltages
19:20:09funmanit used some registers different from what is in your code, so I thought I would call you for help :)
19:20:32funmanwhat stroke me is the POWER management unit name
19:21:16bertrikI think i2c itself is not related to power management at all, but possibly the codec that sits on the i2c bus is
19:22:21 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:25:21funmanright, it's just a transport bus
19:25:43funmanI want to clean a bit the SD code I have to commit it
19:25:53funmanbut right now I'm facing some problems with the clock settings
19:26:27funmanI think I don't set the pclk correctly since I don't have any display anymore when I use the PLLA as a source for it
19:26:28 Quit nuonguy ("This computer has gone to sleep")
19:26:51funmanbut earlier in my tests I saw a data abort in the scheduler code - a part that shouldn't have been called
19:27:10funmanmaybe I need to 'invalidate the instruction and data cache' when I change the cp15 control register for asynchronous mode
19:27:27 Join massiveH [0] (
19:27:56funmanoh .. or maybe I should wait for the PLLA to be locked like the datasheet mentions
19:28:24 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
19:28:37 Quit linuxstb (Read error: 104 (Connection reset by peer))
19:28:44 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
19:28:54 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
19:28:54bertriksounds like a very good idea
19:29:19 Join stoffel_ [0] (
19:29:25*bertrik spots something in the as3525 datasheet about voltage scaling, for Fcpu < 200 MHz we can use 1.1V instead of 1.2V
19:30:55 Quit m0f0x ()
19:31:22 Quit grndslm (Read error: 104 (Connection reset by peer))
19:31:39linuxstbbertrik: What do you think about trying to share the "as3514" (I'm guessing the as3525 parts are similar to the as3514) code with the v1 sansas? The backlight code I committed last night was identical to the c200v1/e200v1 code, apart from replacing the i2c write for PP with your as3525_codec_write function.
19:33:21 Join robin0800 [0] (
19:34:42 Join grndslm [0] (
19:35:31gevaertsfunman: what HAVE_HOTSWAP usage in ata_write_sectors()? ;)
19:38:07amiconnlinuxstb: I would rather live with a bit of duplication than going through hoops in order to unify them
19:38:09 Quit Darksair ("Zzz...")
19:39:13amiconnI'm not saying it shouldn't be done here, but rather that it depends on how much complexity it would add to these builds
19:42:05linuxstbI agree that it depends on the complexity. But it just seems that all I've done recently is copy/paste code from elsewhere in Rockbox...
19:43:44amiconnYeah, but probably you didn't copypaste a whole file without subsequent adjustments
19:43:51bertriklinuxstb, I thought about that too. It seems the as3525 only adds one more register to the as3514 (but I haven't checked all of the register bits)
19:44:23linuxstbBut I think in this case, it's just similar to the wmcodec "API" - the wmcodec_write() function. We can maybe just implement an ascodec_write() function (for PP as well), and use that.
19:44:51*bertrik agrees so far
19:45:21*gevaerts wonders what the build servers are doing
19:45:32linuxstbDinner time?
19:45:42*amiconn experiments with putting the asm ucl_nrv2e_decompress_8() into the flash loader
19:45:48 Join funman_ [0] (
19:46:08amiconnI hope I can get away w/o serial unbricking session(s)
19:47:56amiconnDoes SNV support something like symlinks?
19:49:53*gevaerts doesn't really understand what that means in practice for rockbox
19:50:46bertrikno compiling rockbox on a fat partition
19:50:49amiconnHmm, I wonder whether I should use that, or just cross-reference a file from a different directory
19:51:09amiconnbertrik: cygwin does symlinks even on fat partitions
19:51:12 Quit bughunter2 (Read error: 54 (Connection reset by peer))
19:51:31linuxstbNot everyone uses the cygwin client though
19:52:04amiconnThing is that I need firmware/decompressor/sh_nrv2e_d8.S when building flash/bootloader/ as well
19:52:31amiconnRight now I just use ../../firmware/decompressor/sh_nrv2e_d8.S in the Makefile...
19:53:09gevaertsBagder: the build seems to be locked up again
19:53:34 Quit lasser ("ChatZilla 0.9.83 [Iceweasel 3.0.3/2008092816]")
19:53:37linuxstbDoes decompressor/ need to be in firmware?
19:54:22amiconnI think so
19:54:44amiconnBut even if it doesn't, it would definitely not belong to flash/
19:56:00linuxstbAm I right in thinking that the current code just has a copy of the C function?
19:57:00amiconnIn fact flash/bootloader/bootloader.c had it first, and I nicked it for the decompressor
19:58:17bertriklinuxstb, if the codec register bits are also similar between as3514 and as3525, then calling it ascodec_write sounds like an excellent idea to me
19:59:44linuxstbamiconn: I can't think of anything better than your current solution.
20:02:35 Quit stoffel_ (Read error: 113 (No route to host))
20:02:36 Quit funman (Read error: 110 (Connection timed out))
20:02:38 Join evilnick [0] (i=0c140464@gateway/web/ajax/
20:02:45 Join IudeX [0] (n=52a0f8f7@gateway/web/cgi-irc/
20:03:29IudeXfunman: Why are u have a _ at the end of nick?
20:04:06 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
20:04:06Strife89IudeX: It's a backup for when one gets bumped offline.
20:04:32IudeXStife89: Ok. thanks
20:07:05***Saving seen data "./dancer.seen"
20:13:51n1shm, building the lng files take 1/4 of the total build time on my machine
20:14:42n1s'someone' should look at speeding up genlang :)
20:15:32Strife89n1s: It bothers me, too. My machine is infinitely slower than yours. :)
20:15:44 Join bughunter2 [0] (n=Jelle@
20:15:45*rasher runs away in a hurry
20:15:47Strife89n1s: 800MHz, before you ask.
20:16:10 Quit robin0800 (Connection timed out)
20:16:21Learn1s: I noticed a lot of pattern matching going on, but that shouldn't take very much time...
20:16:46rasherIt's not like my idea of moving to XML is helping any... (it gets a bit slower to build lngs from the xml language format)
20:17:59Learn1s: Looked at perldoc -q profile?
20:18:01 Nick funman_ is now known as funman (
20:20:24*amiconn didn't have the impression that genlang is slow so far
20:20:30n1sLear: nope, as with so many other things I'm clueless when it comes to perl :)
20:20:55n1samiconn: maybe because cygwin makes everything else much slower
20:21:24amiconnI also build on linux occasionally
20:22:08amiconnI wouldn't mind some speedup of the build system of course
20:22:26n1samiconn: building the lng files "time make -j2 langs" takes 11 seconds for me, and a whole build for h300 takes 43
20:23:28*domonoky gets a "no rule to make target 'rockboxlogo.h'" error, but the file was generated.. :-/
20:23:45funmann1s: did you compare to linux running in a virtual machine ?
20:24:18rashern1s: well, there are 6.4MB of text, and a good deal of comparing and what have you
20:24:38n1sfunman: no, i have no virtual machine, it was just a quick comparison between a couple of runs of make langs and regular make
20:25:07funmanI think it's worth a try, seeing cygwin's slowness
20:25:08n1srasher: yes, i am not saying that it is overly slow but it stuck out
20:25:09amiconnDid you disable ccache for that test?
20:25:17n1samiconn: no
20:25:33*n1s tries
20:25:36Learn1s: Single lang takes half a second here...
20:26:45 Quit IudeX ("CGI:IRC (Ping timeout)")
20:27:03n1sLear: there's 46 of them and i have 2 cores so that matches my results
20:27:52Learn1s: Removing the pattern matching cuts some 20%, so that's not a huge issue at least...
20:28:36n1samiconn: yeah, disabling ccache makes the whole build take twice as long so lng generation is only 1/8 of that
20:28:40rasherI guess you could speed it up by allowing to compile multiple langs in the same instance (thereby only reading english.lang once)
20:28:52n1sLear: maybe I'm overreacting :)
20:29:19 Join mc2739 [0] (
20:29:50LearAnd btw, the pattern matching I looked at was only the bit where determines if a string should be included for the current target or not...
20:31:24n1sLear: so if it does that based on english.lang the results should be possible to cache
20:31:42n1sthis seems like it would be too much work for little gain
20:32:00rashern1s: only if the translation properly matches english.lang
20:32:01*funman fought one hour to find that an 'and' in asm code should have been an 'orr'
20:32:24rasherAnd yeah, I'm pretty sure this isn't exactly low hanging fruit
20:36:27 Join jhulst_ [0] (n=jhulst@unaffiliated/jhulst)
20:37:12Learn1s: Hm, some more testing suggests that optimizing parsetarget could help. E.g., it splits and iterates over $target a lot. Doing it once should save some time, as could pre-compiling some patterns.
20:37:50 Join spaceboy [0] (n=50ee5801@gateway/web/cgi-irc/
20:37:53rasherIt hardly seems worth it
20:38:26spaceboyHello - where can I get "eraser" - as mentioned here: ?
20:38:41spaceboyI am having same issue as thread creator ;s
20:39:04rasherI think you need to talk to gevaerts
20:39:49amiconnIf I had done for 3.0.1 what I'm doing now, moving the rombox link address for the Player would have been unnecessary
20:39:50mc2739spaceboy: try here -
20:40:21spaceboyOh, thanks :)
20:40:25amiconnNow I have the problem that I need to pad either bootbox or the bootloader a little
20:40:41mc2739you're welcome
20:41:54spaceboyGoinig to unbrick, have a nice evening!
20:42:06 Quit spaceboy (Client Quit)
20:45:32Learn1s: He, according to a profile run, 40% of the time is spent in parsetarget, and another 40% in match... :)
20:46:26funmanwhere are the other 40% spent ?
20:47:07*gevaerts tells funman about the decimal system :)
20:47:59 Join Nic0_P [0] (
20:49:00 Quit Nico_P (Read error: 104 (Connection reset by peer))
20:49:14mc2739funman: did anyone tell you that your sd test works on the e200v2?
20:49:43 Quit jhulst (No route to host)
20:49:56funmanmc2739: no, nobody did
20:49:57*amiconn did brick is Player
20:50:15*funman wonders what mc2739 will tell ..
20:50:26mc2739it does!
20:50:45kugellinuxstb: I think I can have the patch done in 1h or 2
20:51:14funmanmc2739: did you use the NAND base address or SD ?
20:51:47linuxstbfunman: This isn't fair, the e200v2 port is working better than our own devices... ;)
20:52:08*funman hopes bertrik will make linuxstb change his mind shortly
20:52:32bertrikhmm, change his mind how?
20:52:44funmanfind the missing i2c bit for SD card power-up !
20:52:47linuxstbkugel: OK, thanks.
20:52:53mc2739funman: I used the code that was in the git version of bootloader/sansa_as3525.c
20:53:33funmanmc2739: can you look at the top of ata_inand.c and change the condition to see if it also works the same for the SD slot ?
20:54:16mc2739funman: I can try it later - I'm not at home now
20:54:27*amiconn wonders what he did wrong...
20:55:49 Join _guest [0] (n=4d63e631@gateway/web/cgi-irc/
20:56:01 Nick _guest is now known as obo (n=4d63e631@gateway/web/cgi-irc/
20:56:09 Quit evilnick (" ajax IRC Client")
20:57:25 Join stu8ball [0] (
20:57:28obois there any reason to keep the gigabeat-s branch around?
20:57:34funmanamiconn: gevaerts , can search for 'IF_MV' in and tell me if you think the usage is correct ?
20:57:55stu8ballWhat all languages can plugins be written in ?
20:58:03stu8ballThere's C and, on the wiki example, C++.
20:58:17funmanIn fact I think overloading function calls with a IF_MV2() macro is a bit overdoing, even if it saves one argument when calling the function (the functions have maximum 2 arguments)
20:58:21obostu8ball: c only
20:58:28stu8ballok, thanks
20:59:03 Join petur [50] (n=petur@rockbox/developer/petur)
20:59:28bertrikfunman, after looking at the datasheet it seems most generated voltages have pretty sane defaults
21:00:07funmancouldn't a on/off switch exist on the bus ?
21:00:08 Join culture [0] (
21:00:20funmanmaybe undocumented in the datasheet since sandisk added its own stuff in the SoC
21:00:23gevaertsfunman: looks ok to me. For static functions with only one argument, you can probably just add the drive argument. It's mainly the public official ata_* ones that you need to be careful about
21:01:31linuxstbmarkun: ping
21:01:52linuxstb(or anyone else that knows the audiohw API)
21:01:54gevaertsfunman: actually, you now have a global and several local drive variables. That doesn't look right
21:02:25funmangevaerts: well the global variable should have been removed obviously
21:02:34 Join DerDome [0] (
21:02:48funmangevaerts: why 'only one argument', is there a problem with function having 2 ? (like send_cmd() ) . I thought the barrier was 4
21:03:13gevaertsfunman: one is certainly safe, four is certainly not. I don't know where the limit is :)
21:03:48bertrikfunman, there is a USB voltage of 3.26V which is disabled by default
21:04:03gevaertsUSB 3.26V?
21:04:34bertrikto supply the USB 2.0 OTG interface block
21:05:18*amiconn probably shouldn't have lame-padded the bootloader...
21:05:19bertrikofcourse I don't know exactly what the internal bootloader does, it may put other values in the voltage generation registers than the codec defaults
21:05:40*amiconn goes to do some soldering :\
21:06:43 Join Schmo [0] (
21:07:40linuxstbCan anyone familiar with the audiohw API look at this patch, and say if it looks like the Right Thing ?
21:09:17gevaertsStrife89: PluginSpacerocks is indeed not very fancy, but at least in my view most of those plugin wiki pages don't make much sense, as the information should be in the manual
21:10:38Strife89gevaerts: I've edited some plugin pages already but didn't remove the "lack of documentation" symbol if you'd like to look.
21:11:10Strife89gevaerts: I'd actually like to see the Wiki pages and the manual text merged somehow, but it's wishful thinking on my part.
21:11:50 Join robin0800 [0] (
21:11:57gevaertsI think they should be merged with the manual as the result, not to have both with the same content, as then they will diverge again
21:12:02 Quit Schmogel (Read error: 110 (Connection timed out))
21:12:04 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
21:12:44funmanNow that I know my code will not break any existing setup: should I commit it or post it on flyspray?
21:12:48bertrikfunman, in the OF I something weird, apparently something is written to the sd_mci_base address and gpio d7 at the same time
21:13:13Strife89gevaerts: Four I've edited:
21:13:27funmanbertrik: I noticed that : but xpd should be linked to the sdmci controller, and not to the gpio at this time
21:13:40Strife89Make that three:
21:13:53funmanI'm not sure if writing into gpio d7 will write into xpd[7] (which is the resistor open drain pin IIRC)
21:14:46funman < sd patch
21:15:37Strife89Oh, is this work on the SD driver?
21:15:51*Strife89 is clueless here.
21:15:59J-23can I test it?
21:18:07 Quit Strife89 ("Signing off for now.")
21:18:12 Join evilnick [0] (i=0c140464@gateway/web/ajax/
21:18:26bertrikI thought D7 was the button light
21:18:27 Join Acksaw [0] (
21:18:32bertrik(on the clip at least)
21:19:11gevaertsfunman: I think that if it doesn't brick players, and if it has no chance of introducing bugs (i.e. breaking someone's nice rockbox experience somehow), it should just go in.
21:20:07funmanbertrik: look at the BGA mapping : some pins have different meaning (selectable in CCU_IO for xpd[])
21:20:39funmangevaerts: ok, if I don't follow the 'good style' (tm) I hope someone will tell me ;)
21:21:10gevaertsNew ports have bit looser rules than e.g. playback.c :)
21:21:44*funman looks for something to modify in playback.c
21:21:51domonokyit seem r18920 broke building on my system (cygwin) :-/ (it seems it cant find rockboxlogo.h and other bmp headers)
21:22:29domonokyr18919 builds fine... *searching for the bug*
21:22:49n1sdomonoky: you did 'make clean' right?
21:22:56bertrikThat confuses me, does it mean we can't do anything with the button light is the sd it busy?
21:23:03pixelmamaybe reconfigured?
21:23:16domonokynls: yes i did, clean and reconfigured source code
21:23:50funmanbertrik: I'm not sure ..
21:26:02gevaertsShould I even try to change ata_flash.c in my storage overhaul? iFP7xx compilation fails before it reaches that file, so I have no way at all to see if I did it properly
21:26:17 Quit mc2739 ("bbl")
21:27:33funmanbertrik: or you could modify ccu_io, modify the button light, and modify ccu_io again but you need to be sure the SD is not busy and your code will not be interrupted
21:29:55pixelmadomonoky: is this for every build you tried or only target or sim or? Currently trying myself on cygwin too (Ondio target build as a start)
21:30:33pixelmalooks good sofar (it's building the plugin bitmaps currently)
21:30:36domonokypixelma: i was trying h100 sim.
21:31:03domonokythe bitmaps are built correctly, but it cant find the headers later.. :-/
21:31:33 Quit Acky (Connection timed out)
21:32:21domonokylooks like target builds works... it got past the point it errors for sims..
21:33:09 Quit J-23 ("Flying cow pressed ^D on my keyboard.")
21:33:26pixelmaOndio build went fine here
21:33:46domonokycould you try a sim build ?
21:34:02pixelmayes, configuring right now
21:34:44pixelmathat sandisk target list got too long and confused me ;)
21:35:06 Join arohtar [0] (
21:35:31domonoky and we will extend it even more...:-)
21:36:28 Quit arohtar (Read error: 104 (Connection reset by peer))
21:36:44 Join arohtar [0] (
21:37:17pixelmaright, stopped at *** No rule to make target `rockboxlogo.h' ...
21:38:05domonokyso sim build on cygwin are broken..lets blame zagor :-)
21:38:16funmanonly on cygwin?
21:38:36domonokyhm.. dont know. can somebody try on linux ?
21:39:02pixelmanothing to see in the build table
21:39:18n1ssim builds fine here
21:40:02gevaertsr18922 should have fixed that if I read the log correctly
21:40:07 Join Nibbler [0] (
21:40:23pixelmaI just updated to r18926
21:40:29 Quit faemir (Read error: 60 (Operation timed out))
21:41:21pixelmaI mean before trying
21:41:28 Join Kingbuzzo [0] (
21:43:45 Join meven [0] (
21:46:18pixelmacould be a hint for "someone" though - the error that occured on some buildservers and which was fixed by r18922 looks similar but not completely the same. So maybe it's related...
21:46:35domonokyyes, from r18920 to now its broken..
21:46:51funmanbertrik: the OF enables the i2c audio before using the SD, not the i2c master
21:47:09n1sfunman: there is one problem with importing code from linux, we are GPL "v2 or later" while linux is strictly v2 so if anyone want's to ever release rockbox as GPL v3 that file could not be included...
21:47:35funmann1s: it's an header so the only problem will be someone has to type again the list of registers
21:48:10 Join meven_ [0] (
21:48:26funmanand I thought invidivual files could be licensed at the discretion of their author, and the 'gplv2 only' apply only to files with an unspecified licence
21:48:28bluebrotherfunman: shouldn't the file itself have a note about it being imported? Your mail about the tabs will get forgotten rather soon
21:48:32 Quit massiveH ("Leaving")
21:48:43funmanbluebrother: I thought the path in the file was enough
21:49:45n1sfunman: i know, just saying as we had a _lengthy_ debate and finally decide to go "or later" and went through all headers and updated them, got permission to use that "or later" for IPL code etc, so i don't think we should add more restricted code, i don't know how others see it though...
21:50:39funmanoh I see
21:50:52*linuxstb agrees with n1s - we don't want to commit any new "v2 only" files
21:51:10funmanthis file explicitely says 'gpl version 2'
21:51:36linuxstbYes, that's the problem.
21:51:51funmanI think mailing Deep Blue is shorter than rewriting the whole file
21:51:58*bluebrother is with n1s here −− making the license situation unclear for the whole thing isn't a good thing.
21:52:06 Quit obo ("CGI:IRC")
21:54:17 Quit meven (Connection reset by peer)
21:55:59funmanI sent them a short mail asking the permission to redistribute under GPLv2 or later
21:56:15funmanif they don't answer this week I'll retype the registers list
21:56:35funmanwell there is even no registers, only bitfields
21:57:42 Quit petur ("*plop*")
21:58:22funmanbertrik: a problem is I see the OF accessing undocumented registers in the i2c_audio memory space
21:58:41funmanoffsets 0x14 and 0x18 for example
21:59:08 Quit Thundercloud (Remote closed the connection)
22:01:34 Join Thundercloud [0] (
22:02:12*bluebrother spots that he overlooked a red build yesterday
22:03:49bluebrotherhmm, but if that memcpy is wrong due to the value being a short then it's wrong at other locations too
22:05:18amiconnThe struct member you tried to copy is an array of shorts
22:05:27amiconnThe other one is an array of chars
22:05:47 Quit amiconn (" reboot... :(")
22:05:51 Quit Lear ("ChatZilla 0.9.83 [Firefox 3.1b2pre/20081029033642]")
22:05:54 Quit meven_ (Read error: 110 (Connection timed out))
22:07:10***Saving seen data "./dancer.seen"
22:07:35bluebrotherameyer: no, I mean other places where the array is memcpy'ed
22:07:48bluebrotherlike in restore_state()
22:08:56bluebrotherah, that wasn't meant at ameyer ...
22:10:31bertrikfunman, those are some other kind of slave id, I think they are used for i2c in slave mode, see I2C1_SEAD0/1
22:11:14funman bertrik but I don't see them listed for I2C2
22:11:17bertrikit looks like there is a single i2c driver that can handle both the audio master and i2cmsi bus
22:11:19linuxstbbluebrother: savedboard/currentboard are char arrays. It's only possiblevals that's short.
22:11:45funmanah I just found the 'i2c_busy' function in the OF
22:11:48kugellinuxstb: so, I have a running sim now (without background picture though, since I can't find 64bit driver for my old scanner). What about the bootloader?
22:12:15bluebrotherlinuxstb: yep, but sudoku.c:870 memcpy-ies possiblevals
22:12:21kugelI could compile a bootloader in the git repo (which didn't more but add the 5s delay). Now it #errors because of the unknown memory
22:12:52bluebrother(plus a few other places)
22:13:23linuxstbbluebrother: No it doesn't... Are you looking at current SVN?
22:13:42*bluebrother goes getting a fresh copy
22:14:12 Join stoffel_ [0] (
22:14:19linuxstbkugel: Fixing the real (i.e. not sim build) is what's needed. i.e. adding the Fuze correctly to the build system, with either dummy functions, or real code.
22:15:20kugellinuxstb: Yes I know, but making the sim work was rather trivial. And thus, I already added a keymap.
22:15:39funmankugel: for the fuze it should be the exact same settings than for clip & m200
22:15:44kugelBut that's why I'm asking, should the bootloader build work or is it not too serious now?
22:15:52*bluebrother notices his local copy is messed up and for some reason svn didn't notice?
22:15:59 Quit arohtar (Connection timed out)
22:16:09kugelI added it to the build system, that's not the question
22:16:18kugelto the rockbox tree, rather
22:16:57kugelfunman: i.e. 2MiB ram?
22:17:23domonokywith fuze added, mkamsboot could get dangerous for some m200v2 firmwares (it detects them as fuze, which errors for now)
22:17:46linuxstbkugel: The aim is to get the bootloader compiling.
22:17:51funmandomonoky: I wanted to ask you if you can find something to use A3 for dual boot
22:17:53*bluebrother realizes how he managed to fool himself and goes standing in the corner
22:18:34funmanthe problem is that it could be problematic : rockbox running when we don't want, or the reverse, because we HAVE to use the power button at startup
22:18:54funmanmaybe check power button; delay 0.5s and check again
22:19:20domonokyfunman: yes, thats the problem.. i will try to find something..
22:20:40linuxstbdomonoky: Maybe the suggestion of md5 checksums should be added?
22:20:41 Quit maha (Remote closed the connection)
22:21:06funmanthe only problem I can think of is bricking a device due to bad button code
22:21:24 Quit moos ("Rockbox rules the DAP world")
22:22:10 Join maha [0] (i=juma@gateway/gpg-tor/key-0xB9002659)
22:27:06bertrikfunman, on line 301 of ata_sd_as3525.c it seems comment are code are inconsistent, is the code correct or the comment?
22:28:06 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
22:28:35funmanread, 3rd bit
22:28:41 Quit {-phoenix-} (Remote closed the connection)
22:28:56 Join {-phoenix-} [0] (
22:31:12*linuxstb now needs an md5 function for mkamsboot, and sees one in tools/fwpatcher/
22:31:49domonokyit seems adding fuze is not as dangerios as i though, it will boot into the OF, if the A§
22:31:55domonokyA3 check is enabled..
22:33:11 Join dabujo [0] (
22:33:53domonokybut md5 sums like for the iriver hx00 bootloaders would be good..
22:34:22funmanextra safety ^^
22:34:30funmancan iriver hx00 be unbricked?
22:34:49domonokyonly with a bdm.. (extra hardware and skills)
22:34:54bluebrotheronly by using a bdm
22:36:09domonokyafaik rbutil does this md5 checking for the irivers, so we dont really need the md5sums in mkamsboot..
22:36:31 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
22:36:37*amiconn really wonders what he did wrong
22:36:40funmanthere is a big warning anyway
22:36:47domonokyand trust that people who use mkamsboot directly to know what they do.. :-)
22:36:53amiconnuart_boot now refuses to work, at least the important part :(
22:37:16funmanamiconn: something due to the decompressor ?
22:37:25amiconnA few weeks ago I tested it before I tried the new flash image, and then it worked
22:38:08amiconnfunman: uart_boot is an old program to load a serial monitor into an archos and run stuff using that monitor
22:38:29 Quit jgarvey ("Leaving")
22:38:50amiconnIt has nothing to do with the decompressor. What I did was flashing an obviously wrong image into the flash rom, and so my Player no longer boots
22:39:00 Quit Thundercloud (Remote closed the connection)
22:39:45amiconnWith uart_boot it should be possible to transfer minimon and unbrick the box, but somehow minimon refuses to work, even though the download *appears* to work
22:39:57 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
22:41:26Slasheridomonoky: iriver bootloaders (from version 7) have internal crc32 checksum
22:41:32 Join Thundercloud [0] (
22:42:26 Join tyfoo [0] (
22:42:28tyfoohi, how to backup my existing rockbox setup including all configs? (ipod mini 2nd gen)
22:42:53domonokySlasheri: but that doesnt save you if use flash a broken bootloader with the OF ?
22:43:20domonokytyfoo: just backup the .rockbox dir..
22:43:25Slasheridomonoky: indeed
22:43:42Slasheridomonoky: but normally you shouldn't patch the bootloader from the .bin file yourself
22:43:59Slasherieven the OF does have a checksum in its .hex-format
22:44:00linuxstbdomonoky: I don't want the command-line utilities to become second-class citizens though - I still prefer to use them to rbutil.
22:44:40domonokylinuxstb: the iriver commandline tools also dont do this :-)
22:45:02linuxstbI know, but I've long since removed the iriver firmware, so they're no use to me.
22:45:05domonokySlasheri: extra safety is always nice..
22:45:07tyfoodomonoky: i heard, rockbox saves its settings in an unpartitionated diskpart. is that wrong?
22:45:10 Join Zarggg [0] (
22:45:33linuxstbtyfoo: It used to, but not for a few years.
22:45:42tyfoodomonoky: because i want to install the original apple fiormware for testing purpose
22:47:48Slasheridomonoky: that's true of course. and in fact version 7 bootloader flash sequence has a double checksum check (a simple checksum and hard-coded crc32)
22:48:11linuxstbshotofadds: Hi, any news?
22:50:13 Quit stoffel_ ("leaving")
22:50:30 Join n17ikh|Lappy [0] (
22:51:06*gevaerts thinks that he should actually test his 109 files FS #9500 diff some time :)
22:51:15 Join midgey [0] (
22:52:03gevaertsIt looks like that one will give an interesting delta though.
22:52:55funman32MB of ipodlinux source - still downloading ..
22:53:06*bluebrother wonders what color that delta will have
22:53:17funmanI hope to find descriptions of the i2c registers
22:53:23shotofaddslinuxstb: I haven't had much spare time lately i'm afraid. I'm trying to get usb going on the m200, but I think windows might be thwarting that right now. Don't suppose anyone has a tcctool that might run in the Rockbox debian VM (Sarge?) since I can't built one myself..
22:53:34gevaertsCurrent state is +600 for ipod, and -1400 for D2
22:53:42 Quit Kingbuzzo ("Ex-Chat")
22:53:54 Quit reacocard (Read error: 104 (Connection reset by peer))
22:54:02shotofaddsgevaerts: nice! I vote it goes it :p
22:54:38gevaertsNot yet. I still need to get rid of some wrapper functions (they need to be macros), so it will look even nicer :)
22:54:59gevaertsAnd maybe test a bit of course
22:56:29*gevaerts wouldn't mind if some people reviewed this to check if he's not too far from the right track
22:56:37shotofaddstesting, schmesting ;-) I'll try to take a look later tonight, but for now at least I have an excuse for not committing NAND driver updates... waiting for gevaerts' re-work ;)
22:56:39kugelfunman: I get "unsupported model: fuze" when building the bootloader
22:56:53funmanin mkamsboot?
22:56:59kugelin make
22:57:11gevaertsshotofadds: I don't think the patch will touch parts where you are likely to change things :)
22:57:13kugelat the very end of the compilation
22:57:28funmanplease paste the thing, I have no idea where this message comes from
22:58:04 Quit Kopfgeldjaeger (Read error: 113 (No route to host))
22:58:11domonokyits probably the sdram mode "read"
22:58:33linuxstbkugel: That's probably in tools/scramble
22:58:40funmankugel: if it's in system-as3525.c , just add the fuze case next to m200v2 and clip
22:59:13kugelscramble sounds promising
22:59:35linuxstbCheck the svn history of scramble.c - you'll see where I added the other v2 targets
22:59:40 Join Kopfgeldjaeger [0] (
22:59:47*shotofadds looks at scorche and wonders if an updated VMware image is still on the todo list
22:59:49linuxstbThe number used there is MODEL_NUMBER from config-fuze.h
23:00:00*scorche|sh hides
23:00:10shotofaddsthe lack of libusb-dev (and inability to apt-get it) is rather frustrating :(
23:00:30gevaertsshotofadds: why can't you apt-get it?
23:02:21funmanshotofadds: you can download the .deb and dpkg -i it
23:03:19linuxstbOr just do it the old fashioned way and get the source .tgz and ./configure && make && make install
23:03:54*gevaerts recommends downloading the deb
23:04:25scorche|shDebain Policy and all that jazz
23:04:34 Join fml [0] (n=4fd3c894@gateway/web/cgi-irc/
23:04:37*scorche|sh transposes a couple of letters
23:04:44 Part evilnick
23:06:12gevaertsOf course it's entirely possible that tcctool doesn't work with that version
23:06:32fmlHas any consensus been reached about including/not including new features? Low mem targets are not an argument for me because, in order to enable RB on them, many other features have to be cut off as well. Hence it will be a raid :-)
23:06:47shotofaddsgevaerts: we'll soon find out...
23:07:34fmlAnother question: in one of the recent commits, memcpy is used with a parameter 81. Would sizeof work there (2-d array)? Or would sizeof give wrong size in that case, i.e. not 81?
23:07:34Bagderfml: what consensus is there to reach?
23:07:39kugellinuxstb: FS #9512
23:07:52gevaertsfml: the consensus is that we won't do that RAID :)
23:08:01 Join AJCantos [0] (
23:08:21funmanrockbox is not emacs (tm)
23:08:36fmlBagder: about what features are accepted into the main stream and what need a custom build to be used
23:08:57 Join meven [0] (
23:08:59Bagderfml: we haven't discussed individual patches afaik
23:09:03*linuxstb thinks the concensus is to treat everything on a case-by-case basis
23:09:09funmankugel: a Clip background could be used for the Fuze
23:09:23kugelfunman: Well, but both are good oses lacking a decent text editor ;)
23:09:30*Badger looks rather surprised.
23:09:31Bagderthe discussion has been more like waving hands in the air
23:09:32kugel(regarding the emacs)
23:09:37fmlre. 81: I mean r18906 (sudoku.c)
23:09:47gevaertsfml: only well-coded patches need apply. I'm not aware of any more consensus than that
23:10:09funmankugel: the memory on the fuze is not unknown, it is 2MB ;) (tools/configure)
23:10:24 Quit bughunter2 ("bye")
23:10:26kugelWell, I don't know it. You told me to use 2MiB
23:10:34*gevaerts also doesn't see the need for much more consensus
23:11:05fmlAnd what about sizeof(2-d array) =? 81?
23:11:58gevaertsfml: I personally think sizeof would be better in this case.
23:12:04fmlI ask because I was surprised to see a number there.
23:12:23funmankugel: in the config-fuze.h there is mention of button wheel, also do you know the screen resolution ?
23:12:50kugelit's 220x176
23:12:55shotofaddsgevaerts: i don't think this is going to work in a VM. VMware shows the device is connected (140e:b021) but lsusb inside the VM shows nothing
23:13:20funmankugel: why didn't you test the bootloader ? (to see if backlight works for example)
23:13:22gevaertsCan't you tell vmware to pass it on to the guest?
23:13:29fmlgevaerts: I'm I right in the assumption that if I have char x[9][9] then sizeof(x) is 81? Haven't used such thing for a while and am too lazy to try it out.
23:13:58gevaertsI don't see why not
23:13:59kugelfunman: a) let you/ linuxstb review my work b) no time :)
23:14:07shotofaddsgevaerts: I did, it already shows as connected..#
23:14:39gevaertsshotofadds: did you run lsusb as root?
23:14:41 Quit domonoky (Read error: 104 (Connection reset by peer))
23:15:25 Quit tvelocity ("Αποχώρησε")
23:15:48shotofaddsyep. nothing, which is weird as I'm sure I've seen this working before
23:17:26*gevaerts can't help further then
23:17:39gevaertsI've never actually used vmware
23:18:06shotofaddsgevaerts: when I tried a usb m200 build from windows the usb stack fails with a panic "usb_drv_send_nonblocking(129, 0): Not supported". I guess that's another one to ask vitja about...
23:18:17shotofaddshence why I wanted to try from linux
23:18:28*shotofadds will attempt to track down a real linux install
23:18:57rasherBagder: did you see
23:19:08funmanshotofadds: I know virtualbox can do usb passthrough
23:19:11BagderI noticed it but I haven't checked it out closer
23:19:25 Quit fml ("CGI:IRC 0.5.9 (2006/06/06)")
23:19:29rasherIt's basically a copy of, but slightly smarter
23:19:44 Part pondlife
23:19:44shotofaddsfunman: I'll give it a try later, but right now the "better half" needs some attention... back later.
23:20:03gevaertsshotofadds: that means there is an issue with endpoint allocation I think. Are you trying serial or storage?
23:22:05rasherBagder: I've modified the other scripts in the same way, but hesitate to commit
23:22:14AJCantosHi guys, I'm having problems executing, here's what I get
23:23:13funmanAJCantos: you really need to know what happens before these lines
23:23:44funmandid you verify arm-elf-as etc were installed when the script reaches this point ?
23:24:49AJCantosfunman: No, how should I do that?
23:25:10funmancheck in /usr/local/arm-elf (the default)
23:25:26funmanthe script will modify it's $PATH to look there itself
23:25:56amiconnHah, the SVN minimon is linked to a wrong address
23:26:11n1sAJCantos: looks like it tries to use the wrong assembelr
23:26:17amiconnuart_boot loads it to 0x0ffff000 (IRAM), but it's linked for DRAM (0x09000000)
23:26:22Bagderit's /usr/local/arm-elf/bin
23:26:23funmanamiconn: can't those addresses be determined at buildtime ?
23:26:36*amiconn wonders what [IDC]Dragon did....
23:26:53 Quit Kopfgeldjaeger (Read error: 104 (Connection reset by peer))
23:26:57amiconnAt least I have a working minimon for unbricking my Player
23:27:09robin0800funman: don't know if it helps but now CPPFLAGS=-O_FORTIFY_SOURCE is inside I don't see it in the print out should I?
23:28:19funmanrobin0800: Yes you should see it, did you try removing /tmp/rbdev-build ?
23:28:21BagderAJCantos: seeing you're doing that on a /media/ device, I would rather recommend to use a "native" hard drive for speed (and non-fat32 filesystem)
23:28:36AJCantosnls: and what am I supposed to do?
23:29:21AJCantosBadger: I can't, my eee doesn't have much storage!
23:29:42robin0800funman: yes you can't run it again if you don't remove them
23:29:57BagderAJCantos: ok, that's a good motivation but are you using fat32 on that?
23:30:06n1sAJCantos: that sort of depends on _why_ this happens, did the binutils build and install work as expected, is arm-elf-as and friends in the path where the script needs them, do you have the right permissions etc
23:30:20funmanrobin0800: I built the arm compiler today on ubuntu 8.10, but I needed to apply 2 patches to gcc and binutils (binutils would crash and gcc fail to build)
23:30:49funmanAJCantos: try /usr/local/arm-elf/bin/arm-elf-as
23:30:59funmanmaybe you can't execute them if they are on a separate partition
23:31:31AJCantosfunman: Fair enoug
23:32:12AJCantosBagder: My in-board storage is ext3, but the externl is fat32
23:32:47BagderI wouldn't count on that to work 100%
23:33:04funmanyou could create a big file on the fat32, and use it as a ext3 partition
23:33:34BagderI mean, it depends on how you mount it etc but fat32 isn't case sensitive and more
23:33:57BagderI suspect that can lead to "funny" side-effects
23:34:04robin0800funman: yes saw those on the mailing list, does Badger know its broken again?
23:34:05AJCantosn1s: Is difficult to tell, as I don't know what should be in there...
23:34:36AJCantosAlright, I'll try to build them in my ext3 disk and see what happens
23:34:50AJCantosThanks guys
23:35:32*Bagder can see how Badger will offload him some of his traffic...
23:35:46robin0800funman: How do you apply your patches ?
23:36:37funmanrobin0800: you can add a pause in the script after it has unpacked the source ("echo press enter;read i" will do)
23:36:48funmanand apply them in /tmp/rbdev-build before compilation begins
23:37:11 Join krazykit [0] (
23:38:28*Bagder tries to understand what the YH920 guy means
23:40:18*linuxstb looks at the Sansa V2 firmwares, and sees them all described as "e200 plus", "m200 plus" etc in the included header.txt file
23:40:43funmanlinuxstb: you mean in the SanDisk archive ?
23:41:11linuxstbI'm not sure where the header.txt comes from - it's in some of the downloads on Bagder's site
23:41:21 Quit Schmo (Read error: 104 (Connection reset by peer))
23:41:36BagderI think they are more or less directly off SanDisk's site
23:41:46Bagderthe packages I mean
23:42:52funmanI download them on the Sansa forums
23:43:15funmanif e200v2 is "e200 plus", why isn't m200 "m200 plus plus plus" ? :)
23:43:32gevaertsAnd is the c200v2 a c plus plus?
23:44:01linuxstbBecause there is no such thing as "e200v2", "c200v2", it's just something Bagder invented ;)
23:44:16*Bagder goes to the corner
23:44:29funmandoes it mean Bagder is now our messiah ?
23:44:40Bagderyou may kneel now
23:44:43funman"thou shall not crash"
23:46:24*amiconn got his Player back up :D
23:46:54amiconnDarn minimon - gotta commit that fix
23:47:43 Quit bertrik ("Leaving")
23:50:18 Quit jhulst_ (Remote closed the connection)
23:50:34 Quit n1s ()
23:55:34amiconnAnd a flash image built without lame manual padding works as expected...
23:55:37*amiconn slaps self

Previous day | Next day