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 2008-11-02

00:01:11 Join webguest85 [0] (n=3e2fb83a@gateway/web/cgi-irc/
00:02:51 Quit petur ("Zzzzz")
00:05:35 Quit bluebrother ("leaving")
00:05:53 Join saratoga [0] (n=98039edd@gateway/web/cgi-irc/
00:06:01 Quit {phoenix} (Remote closed the connection)
00:07:14 Quit Strife89 ("Going to free up my uncle's PC now.....")
00:10:18saratogafunman: [for the logs] it might be interesting to compare the relative speed of the two internal DRAMs on the AMS chip to see if theres a lot of difference between them
00:10:48 Quit webguest85 ("CGI:IRC")
00:11:21linuxstbkugel: Could you test this patch? It just cleans up the LCD driver, so should hopefully work the same as before (it includes the colour fix) -
00:11:56 Join funman [0] (
00:12:11kugellinuxstb: will do
00:12:26funmansaratoga: right but we need some reference timing first
00:12:27bertriksaratoga, for the i2c driver I used a simple delay loop to see how long an i2c transfer took. This appeared to indicate that the 384k (or so) internal RAM is single-cycle
00:12:53funmankugel linuxstb: cool news that the fuze's lcd works
00:13:11kugelfunman: hello :) Yep, great success today
00:13:26bertrikwell, at 24 MHz at least
00:14:20 Quit bertrik ("Leaving")
00:14:29amiconnbertrik: Is the cache enabled or disabled for that ram area?
00:15:00kugelfunman: can you help with the buttons now? You said they're on ADC not GPIO, is this a guess?
00:15:46funmankugel: yes total guess based on the datasheet
00:15:58kugelfunman: I've not tested if they're in such a table keyscan
00:16:17saratogabertrik: thats good news
00:16:29funmanwell you would need to have an idea of which pins are used, and we don't know at all
00:16:32gevaertsamiconn: the IF_MV2(drive,) usage interferes a bit with proper argument-taking macros. The preprocessor counts the number of arguments wrong then
00:16:52gevaertsWe can still do non-argument macros, but that's not as nice
00:16:59saratogabertrik: do we set the IRAM clock directly or does the OF bootloader handle that?
00:17:14amiconngevaerts: Then we probably need 2 sets of macros
00:17:43gevaertsThe problem is that it complains at the place they get used, not at the declaration, so 2 sets won't help
00:18:03funmankugel: check out the datasheet pages 163 & 164
00:18:07saratogathe datasheet indicates that its max clock is 60MHz, so i guess theres some divider, but if you're getting 1 cycle access times, it must be running synchronous below that speed
00:18:22amiconngevaerts: Then you might need to use IF_MV() instead of IF_MV2()
00:18:25funmanyou must select the adc source with the i2c bus (I_MicSup2) and read the value
00:18:44gevaertsi.e. if code calls storage_read_sectors(IF_MV2(drive,) start, count, buf); gets counted as three argumens, even in the HAVE_MULTIVOLUME case
00:18:45kugeland why should that be fuze only?
00:19:10funmansaratoga: I don't remember seeing any settings for the IRAM in the datasheet
00:19:20funmankugel: ask SanDisk :)
00:19:30gevaertsamiconn: won't work. we need the comma
00:19:45kugelfunman: is they keyscan method unlikely for the fuze?
00:20:03amiconngevaerts: If storage_*() are macros anyway, the IF_MV2() isn't needed in the call
00:20:14saratogafuman: have you tried clocking the CPU above 60MHz?
00:20:20funmankugel: I won't prevent you to dig in the fuze OF to look for buttons
00:20:21amiconnThe macro itself could drop the drive argument as needed
00:20:21gevaertshm, true
00:20:33funmansaratoga: yes but I have some strange behaviours (data abort in code that shouldn't be called)
00:20:49funmanbut sometimes it works (fclk at 240MHz with icache/dcache enabled)
00:20:50gevaertsThat will solve it
00:21:44saratogai wonder if the hardware auto sets the IRAM clock, or if we're just able to run it out of spec without serious issues
00:21:54amiconnHmm, although that means a little waste for targets where storage_* won't be macros
00:22:00kugellinuxstb: so weird, now it only works with _buttonlight_on() again
00:22:08amiconnBut I guess those will actually need the drive number anyway
00:22:58funmankugel: so is the buttonlight and LCD backlight on pin D7 ?
00:24:09kugelfunman: no idea. When the backlight worked for the first time today, it only worked after calling _buttonlight_on(), but shortly after it did without as well
00:25:09funmanstart from a working state and comment out line by line :)
00:25:17kugelfunman: but they aren't the same. backlight is done via i2c isn't it?
00:25:36kugelin my test loop, the buttonlight is on without backlight so...
00:26:12funmanmaybe backlight is in fact a circuit which powers both the backlight and the buttonlight
00:26:22funmanand the buttonlight can be toggled on/off with D7 pin
00:26:39funmanI see the IRAM has a dedicated controller but no mention of how we can control this controller nor what it is
00:26:44kugelsee, now it's working without buttonlight_on again
00:26:55kugelit's rather weird
00:26:56*funman looks but doesn't see
00:27:07kugeli see it though ;)
00:27:35kugellinuxstb: the clean up works
00:27:47funmanthe IRAM clock supports "20MHz to 65 MHz operation speed" so I suppose it's directly using pclk
00:29:26 Join tvelocity [0] (n=tony@
00:29:29linuxstbkugel: OK, thanks.
00:30:45kugelfunman: how far is the sd driver? I've heard there's problems with the internel sd, even though externel works
00:31:51 Quit tchan ("WeeChat 0.2.7-dev")
00:31:52funmanpersonally I can't develop anything since I can't get the driver to work with the embedded SD (in all models). At the moment only the models with a SD slot can be used to develop the driver further
00:32:04 Join tchan [0] (
00:32:16funmanSince the SD protocol is quite well documented, and there are numerous open source drivers it shouldn't be a difficult task anyway
00:33:58kugelfunman: there's existing targets with a microsd slot, so we don't need a whole new driver. The fuze can carry a microsd btw
00:34:28funmankugel: so you can develop the driver and write a bootloader which loads rockbox from a SD card :)
00:34:39domonokykugel: do have a micro sd ? would be good to tes the driver on this :-)
00:34:40kugelI don't think so
00:34:57kugeldomonoky: not yet, but I wouldn't mind buying one
00:35:14funmankugel: the driver for the embedded SD and the SD slot is _exactly_ the same - safe for what we miss to power on the embedded one
00:35:56funmansince it works on e200v2's slot, I don't see why it wouldn't on the Fuze. the SD code is identical on all the sansav2 targets rockbox now 'supports'
00:36:19kugellinuxstb: lcd_clear_display() works, but it's really slow
00:36:39funmantry removing some sleeps maybe
00:37:13WarpTIbut rockbox doesn't support the v2
00:37:20WarpTIi happend to have e280
00:37:31WarpTIunless someone compiled a rockbox for it
00:37:41funmanWarpTI: that's why I put the quotes, I meant all the sansav2 targets which have started being ported
00:38:16WarpTIahh what is there to complete on the port?
00:38:26linuxstbkugel: lcd_clear_display() just clears the Rockbox framebuffer array, it's not a driver function.
00:38:31funmanWarpTI: lots of stuff
00:38:50 Nick fxb__ is now known as fxb (
00:39:26 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
00:39:27kugellinuxstb: I know, but that indicates that the refresh is rather slow
00:41:09funmankugel: svn up to get some speedup
00:42:13funmanamiconn: it's easy to remove binsize if you had added 'nop's :)
00:42:31kugelfunman: the commit is the patch I've already applied
00:42:35amiconnfunman: The SH1 brnach instruction has a delay slot
00:42:54amiconnIf you don't have something useful to do there, the nop is needed
00:43:23amiconnWhat I did now actually costs a single cycle for the replaced 'bra' that didn't have the nop in the delay slot
00:43:30funmankugel: r18969
00:43:38amiconn(bt/bf need 3 cycles while bra only needs 2)
00:43:49saratogawas there ever a patch that does JPEG decoding of album art ?
00:43:51kugelfunman: oh, the page doesn't show it yet
00:44:32amiconnAnd the other 2 instructions were saved by clever use of 'addc' (which does update the T flag, unlike most other arithmetic instructions)
00:46:43*amiconn was hoping to get below 200 bytes, but that seems to be not possible
00:47:58WarpTIhow fast is e200 v2 cpu btw?
00:48:23WarpTIfaster than nintendo ds
00:48:32amiconnNot overly fast for an armv4
00:48:45saratogaalmost everything is faster then the Nintendo DS
00:48:48saratogaits an old CPU
00:48:50funmanthe RTC uses 1 second steps
00:49:11kugelfunman: not really faster
00:49:16funmanI should look at timer1
00:49:19WarpTIhmm sansa c200 faster too?
00:49:34funmankugel: maybe you should try using higher pclk frequency
00:49:51funmantry a CGU_PERI &= ~0x7f; to be sure it's at 24MHz
00:49:54amiconnThe v1 is dual core @80MHz (also armv4)
00:50:37 Quit mofux (Remote closed the connection)
00:51:09mc2739funman: large speed improvement on e200
00:51:35funmanmc2739: it could be even faster, if I understood the problem with the clock settings
00:51:43 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
00:53:03funmanthe CPU is clocked at 42MHz now I believe
00:53:37WarpTINice display's almost done , bootloader works on e series v2 but there's still no sound.
00:54:05funmanbootloader is not finished either
00:54:27funmanaccess to the storage is still missing
00:54:30 Quit ajonat ()
00:54:32liiwiprogress is progress :)
00:54:39funmanright ;)
00:55:39saratogaall targets but the c200 have button and display then?
00:55:43WarpTIit will be easier to debug now however with a kinda running rockbox
00:56:02funmansaratoga: no buttons for the fuze, and nothing for c200 and Clipv2
00:56:22mc2739saratoga: not all buttons on e200 yet - still missing hold, rec, scroll wheel
00:56:35WarpTIclip v2 has display kinda working but that's all :)
00:56:52funmanWarpTI: no. clip v2 is different from clip
00:57:11WarpTIclip v1 is like e200 v2
00:57:19WarpTIthen hows clip v2?
00:57:35funmanI don't know, nobody has opened it and posted pictures yet
00:57:43saratogaWarpTI: if you find yourself about to try and correct funman, probably better to reconsider
00:57:58WarpTIim reconsidering
00:58:22WarpTIfunman's a sansa v2 expert wouldnt mess with him :)
00:58:35funmandon't worry I didn't think you were correcting me :)
00:58:48WarpTIthx funman
00:59:05funmanI'm not an expert I'm just the person interested in sansav2 with the largest amount of free time :)
00:59:58WarpTIhow long till its released u think?
01:00:31funmanwhen it's ready
01:01:06 Join DerDome [0] (
01:01:14WarpTIu know i wouldn't mind test the beta either
01:01:31domonokyWarpTI: no time estimates in the rockbox world, its all done voluntary...
01:01:43*gevaerts tells WarpTI about the channel guidelines and real words
01:01:58funmanWarpTI: not much to test, but developers are very welcome
01:02:11WarpTIman wish i could code
01:02:17WarpTIall i do is making xm's
01:02:40liiwiif you know how to do them well, that's very appreciated.
01:03:03liiwiit's bloody black art to me
01:03:28WarpTIheh been on trackers since a600 :)
01:03:38funman'xm' ?
01:03:47WarpTIeXtended module
01:03:48gevaertsfunman: something off-topic I think
01:03:51saratogaWarpTI: again, guidelines
01:04:02domonokyplease move off-topic talk to -community
01:04:05WarpTIv2 of the .mod tracker format
01:04:11WarpTIoh ok
01:04:22WarpTIGoing to the off topic rockbox channel.
01:05:10WarpTIsorry for Offtopic-ing in Support chan guys.
01:05:18liiwiWarpTI: heh, played with trackers on a500 and few modded c64's, but haven't touched anything related in decade and a half.
01:05:46*liiwi feels moved
01:06:38WarpTIBtw i noticed .mod songs loads , did anyone start on .xm codec for rockbox yet?
01:07:36 Quit parafin ("So long and thanks for all the fish")
01:07:39 Join parafin [0] (
01:08:14domonokywe have a few working tracker formats in rockbox, but i think xm is missing..
01:09:16domonokyFS #8806 is a port of MikMod whch should play .xm but i dont know the state of this patch...
01:12:09 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
01:13:13linuxstbWhat's the status of the mkamsboot dualboot.S code now? I know the Clip dual-boot button check is working fine, what about the e200 and m200?
01:13:52funmanthey all work fine
01:13:53mc2739linuxstb: e200 is fine - down button or usb boot OF
01:14:15 Quit jhMikeS (Nick collision from services.)
01:14:16funmanusb boots OF for the fuze since there is no button known
01:14:21 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
01:14:40mc2739funman: is GPIOD shared also?
01:15:26 Join bimbel [0] (n=Miranda@unaffiliated/bmbl)
01:15:35funmanyes, with SDMCI and memorystick interfaces
01:15:41funmanmc2739: do you have the AS3525 datasheet ?
01:16:42mc2739I just noticed that the buttonlight turns off when storage is initialized
01:16:46funmanyou can ask Bagder for it
01:18:17 Quit bimbel (Client Quit)
01:18:37domonokygevaerts: config-m200v2.h also needs your new CONFIG_STORAGE define.. :-)
01:21:35*gevaerts thinks that there are too many ports in progress to track :)
01:21:59 Quit saratoga ("CGI:IRC (EOF)")
01:22:11 Nick fxb is now known as fxb__ (
01:22:17gevaertsdomonoky: what sort of storage does it have?
01:22:47funmanSD, like all as3525 targets
01:22:53domonokygevaerts: the same as the other v2s (STORAGE_SD i think)
01:23:34gevaertsdone :)
01:25:33*domonoky confirms that dualboot.S in svn work also for m200v2, but the comments are wrong... its button left (not select) and booting is inverted for no (only boot into rb if left is pressed)
01:26:19domonokyand i get a data-abort instead of the logo .. :-/
01:26:49funmanhum same here .. although this data abort is not always present if you modify some code
01:27:18*linuxstb starts looking at the c200v2's LCD driver, and sees that it looks the same as the c200v1 ;)
01:27:46funmandomonoky: if I disable the interrupts (set bit 7 in cpsr) it's fine now, but that might be my 'modification'
01:28:29 Quit jhulst (Remote closed the connection)
01:30:50 Quit ender` (" Light a fire for a man, and he's warm for a day. Light a man on fire, and he's warm for the rest of his life.")
01:32:42 Quit bmbl (Read error: 110 (Connection timed out))
01:33:59JdGordonlinuxstb: can you change the dual boot code for the e200v2 to check for left instead of down to make it the same as the v1's?
01:34:11 Quit Nico_P (Read error: 104 (Connection reset by peer))
01:34:58funmandomonoky: still a data abort with r18972 ?
01:35:19linuxstbJdGordon: Sounds sensible to me, but I'll leave it for someone else. I don't know the button detection code, and I can't test it.
01:36:30gevaertsWhat's the best way to indent preprocessor directives?
01:37:03funman^#<tab>* ?
01:37:08domonokyfunman: will test..
01:38:15***Saving seen data "./dancer.seen"
01:39:14 Quit Xerion (Read error: 104 (Connection reset by peer))
01:39:14 Join blkhawk- [0] (
01:39:14 Quit amiconn (Nick collision from services.)
01:39:18 Join amiconn_ [0] (
01:39:20linuxstbgevaerts: The general style in Rockbox seems to be to not indent at all.... But if I did, I would do ^<tab>#*...
01:39:21domonokyfunman: no data abort anymore, but a scrambled display..
01:39:27 Join bughunter21 [0] (n=Jelle@
01:39:38gevaertslinuxstb: this is 120 lines of only ifdefed defines
01:39:41 Join Xerion [0] (
01:39:51funmandomonoky: it looks like the clocks are too fast for the lcd driver
01:41:09 Quit kachna|lappy (Read error: 104 (Connection reset by peer))
01:41:10 Quit blkhawk- (Success)
01:41:10 Quit domonoky (Success)
01:41:17 Join kachna|lappy [0] (
01:41:22 Join blkhawk- [0] (
01:41:27 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
01:41:41 Join Strife89 [0] (n=michael@
01:41:43 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
01:42:06funmandomonoky: can you try adding some sleeps in lcd_write_command() and/or lcd_write_data() ?
01:42:36 Join Kohlrabi_ [0] (
01:42:38funmankugel: did I break fuze display as well ?
01:42:43domonokywill do. also storage_init() somehow switches my backlight off ?
01:43:05funmanit's certainly gevaerts's fault :)
01:43:06 Quit Kohlrabi (Read error: 104 (Connection reset by peer))
01:43:31 Join Rob2223 [0] (
01:43:36linuxstbgevaerts: Then that sounds like a little indenting is in order...
01:43:37funmanseriously, I don't know why it would do that
01:44:23 Quit Rob2222 (Read error: 104 (Connection reset by peer))
01:44:48funmandomonoky: if the drivers don't work at different clock speeds, we will need to use freq-dependant sleeps I guess
01:44:50mc2739linuxstb: here is a patch for dualboot.S to move e200 to left button -
01:45:17funmanmc2739: you forgot to change the comment :-)
01:45:45 Join goffa_ [0] (n=goffa@
01:46:37mc2739oops, fixed the other two
01:46:47kugelfunman: nope
01:47:02funmanis it faster at least ?
01:47:36funmanmc2739: any side effect on the e200v2 caused by r18972 ?
01:48:05kugelfunman: nope
01:49:02mc2739comment fixed - same link
01:50:03 Quit culture (Connection timed out)
01:52:23funmanmc2739: applied, thanks
01:52:40mc2739you're welcome
01:52:47 Join massiveH [0] (
01:53:34domonokyfunman: some small delay loops make the lcd driver work again...
01:54:09*domonoky is still wondering what switches off his backlight..
01:54:30*linuxstb is still wondering if anyone owns a c200v2
01:54:47funmandomonoky: if they are not too important they can be fixed for now (even for low clock speeds), but conditional on AS3525 probably
01:55:58domonokyfunman: these lcd_write_command and lcd_write_data functions are already conditional on AS3525
01:56:03 Quit Schmogel (Read error: 110 (Connection timed out))
01:56:04 Join sarixe [0] (
01:56:16 Quit bughunter2 (Read error: 110 (Connection timed out))
01:56:33 Quit blkhawk (Read error: 113 (No route to host))
01:56:39mc2739funman: no side effects detected on e200 with r18972
01:57:23funmancool, I have not broken too much and it was repaired anyway ;)
01:57:43funmandomonoky: ah right
01:58:20*funman notices again the data abort, always at the same instruction
01:58:33 Quit Strife89 ("Time to turn in.")
01:58:55 Quit goffa (Read error: 110 (Connection timed out))
01:59:47 Quit amiconn_ (Read error: 110 (Connection timed out))
01:59:53funmanfirmware/thread.c:1214 (when accessing l->l.prev)
02:00:12funmanI don't get why this function is called anyway
02:00:40JdGordonisnt threading disabled in the bootloader?
02:01:08funmanI don't see any call to something which is not a hardware driver in sansav2_as3525.c's main()
02:01:58*domonoky notices that a little delay loop of 10 in the lcd_write_data (after the dbop write is ready) is enough to get the lcd running again..
02:03:14funmanJdGordon: hmm .. I don't know how these functions work at all
02:04:37funmannor how to effectively disable threads (a #define perhaps ?)
02:05:01funmansee you tomorrow
02:05:04 Quit funman ("leaving")
02:05:09linuxstbI don't think you disable threads, you just don't create any...
02:07:39kugeloh well
02:07:45kugelmy fuze is way faster now
02:08:18kugeland lcd_clear_display too
02:09:04JdGordonwhat the heck is the bootloader doing now? lots of looping?
02:09:55JdGordonwhere is the code for the loops?
02:10:05JdGordonsansa_as3525.c is pretty empty
02:10:05kugelmy bootloader loops through the backlight brightness settings
02:10:15kugelthe svn version loops due to storage_init()
02:10:46*domonoky fixed m200v2 lcd in svn...
02:12:11kugelhaha, I already implemented lcd fade :D
02:14:15 Join kronflux [0] (
02:15:02*gevaerts hopes for lots of green
02:15:12 Quit kronflux (Client Quit)
02:15:19JdGordonah bugger... how do i read the whole GPIO in one hit?
02:15:26 Join ajonat [0] (
02:16:28*kugel would like to know that as well
02:16:55domonokyjust dont use the GPIO macros and read from the GPIO base address ?
02:18:02JdGordon*(unsigned char*)GPIOA_BASE+(0xFF << 2) apparentyl
02:20:43linuxstbYes, address lines 2..9 act as a mask.
02:22:32 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
02:23:21 Part WarpTI
02:25:42JdGordonwierd, im trying to get buttons working and only the down button registers
02:26:19 Quit sarixe ("Ex-Chat")
02:27:10JdGordon is the code... the sansaV2 harware mappings page says i have t set the gpio afsel which im hopefuly doing correctly?
02:28:22 Quit tvelocity ("Αποχώρησε")
02:29:24mc2739code looks good to me
02:29:39JdGordonshame it doesnt work though :p
02:30:05mc2739i could not get it to work either, I had to read each bit separately
02:30:09JdGordonits also backwards with the down button... if its unpressed it return BTTON_DOWN, if its pressed it return BUTTON_NONE
02:30:49JdGordonyou mean unset the afsel bit, read, reset it for every button?
02:30:55JdGordoninstead of doing it in 1 hit?
02:31:22mc2739no unset afsel, read each bit, set afsel
02:33:06mc2739although I know reading all bit should work. there was asm code in an early dualboot.S that did it
02:34:16JdGordonthe current dualboot.S reads the down button
02:34:38mc2739changed in r18973
02:35:15JdGordonoh now dualboot is left?
02:35:49mc2739make clean and make mkamsboot
02:38:18 Quit tyfoo2 ("Carpe diem")
02:38:58 Join Darksair [0] (n=user@
02:43:04 Join allele [0] (
02:43:36JdGordonok, with the new dualboot.S it still only reading the down button
02:43:46*gevaerts is confused
02:44:12gevaertsThe h300 sim builds fine here
02:44:20alleleI built the ui simulator for sansa e200 and it compiled fine, but where do I put the .rockbox directory?
02:44:34JdGordonallele: in the archos folder
02:44:38JdGordondo make install
02:46:29 Quit avis (Read error: 104 (Connection reset by peer))
02:46:50alleleah, thanks
02:50:40 Join saratoga [0] (n=98039edd@gateway/web/cgi-irc/
02:52:08gevaertsThe h300 sim fails to build currently, but only for gcc 4.0 or older
02:53:25JdGordonmc2739: where does the 0x20 in dualboot.S come from?
02:53:52JdGordonah, bah my maths sucks.. ok
02:55:35mc2739bit 3 shifted left twice
02:56:28 Quit Bensawsome ("The awsome is gone :(")
02:57:38 Part allele
02:57:47JdGordonhmm, are we reasonably sure the mappings is correct?
02:58:36JdGordonarg, we probably have to reset the direction after the afsel?
02:58:45kugelsaratoga: hey. read the logs as soon as possible :)
02:58:54JdGordonim guessing the lcd doesnt change the down button ins directoin?
02:59:45saratogakugel: i've been reading them, anything specific you're referring to?
03:00:03kugelsaratoga: fuze + lcd = working
03:00:03 Nick blkhawk- is now known as blkhawk (
03:00:20saratogakugel: yeah I saw
03:00:23saratogasounds good
03:01:38kugellooks good* :p
03:02:18kugelsaratoga: if we find the buttons now we can catch up to the clip & e200v2
03:02:27kugel(and m200v2)
03:04:22*JdGordon does his best to sabotage the fuze port :D
03:04:55JdGordonfixing the dir's doesnt change anything :(
03:05:17kugelso you are the one who told sandisk to not use gpio for the buttons, but adc (as assumed by funman)
03:06:24saratogai'm still wondering why profiling codecs doesn't profile the time they spend inside the codeclib functions
03:08:50 Quit domonoky (Read error: 104 (Connection reset by peer))
03:11:04JdGordongevaerts: storage-sim.c in firware and not uisimulator/?
03:14:03gevaertsJdGordon: good point.
03:17:52soapif I were to buy a Sansa v2 simply to test and debug - is there a model which would be useful for me to own?
03:18:31saratogasoap: perhaps the C200 since noone has one yet
03:20:45soapcan do.
03:21:39saratogathough maybe thats getting in a bit deeper then you want
03:21:49saratogadepending how different the hardware is
03:23:33 Quit ajonat (Connection timed out)
03:24:45 Nick Darksair is now known as Darksair{away} (n=user@
03:26:50kugelsoap: haven't you had some philips devices in a good state?
03:27:23kugelI remember the rockbox logo on one of your philips
03:27:30kugelor even the main menu
03:29:38soapkugel, I have no idea of what you speak. There is no money to be made reselling Phillips (not to mention the 3160 scam of the late 90's) therefore I would never buy a Phillips ;)
03:30:16kugelthen my memory lies to me
03:30:36soapsaratoga, I have no problem getting deep - as long as I can be more help than harm. I'm willing to DO anything, but not at a skill level where I can _create_ squat.
03:30:37kugelbut you've shown some devices on the devcon didn't you?
03:30:55kugelor am I totally confusing you with someone else?
03:31:05soapDCW 2007 and DWPG 2007.
03:31:37 Quit faemir (Remote closed the connection)
03:32:55soapI cooked and cleaned at dev con west, and drank beer at dev con pub germany in 2007
03:33:13kugelhmm well, then my bad
03:33:22soapboth instances had photos of my players - no Phillips in sight. ;)
03:33:47kugelmaybe it was scorche then
03:33:53scorcheit was
03:34:00 Quit agaffney (Read error: 104 (Connection reset by peer))
03:34:13soaphe IS my brother.
03:34:14 Join agaffney [0] (n=agaffney@gentoo/developer/
03:37:01 Quit DerDome (Nick collision from services.)
03:37:02 Join DerDome1 [0] (
03:37:14 Nick DerDome1 is now known as DerDome (
03:38:19***Saving seen data "./dancer.seen"
03:39:42 Quit esthar ("KVIrc 3.4.0 Virgo")
03:41:01 Nick Darksair{away} is now known as Darksair (n=user@
04:03:57 Join esthar [0] (
04:13:07 Join hd [0] (
04:13:47 Quit HellDragon (Read error: 104 (Connection reset by peer))
04:13:47 Quit hd (Remote closed the connection)
04:13:53 Join HellDragon [0] (
04:14:00 Join blkhawk- [0] (
04:15:37 Join Falco98 [0] (
04:16:02 Quit massiveH ("Leaving")
04:17:15 Quit kugel ("ChatZilla 0.9.83 [Firefox 3.0.3/2008101315]")
04:20:24 Join massiveH [0] (
04:30:33 Quit blkhawk (Read error: 110 (Connection timed out))
04:30:59 Nick blkhawk- is now known as blkhawk (
04:39:07 Quit esthar (Remote closed the connection)
04:39:41 Join esthar [0] (
05:00:10 Join miepchen^schlaf_ [0] (
05:03:12 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
05:07:17 Quit miepchen^schlaf (Connection timed out)
05:18:13 Nick Darksair is now known as Darksair{away} (n=user@
05:21:27 Quit mc2739 ("Zzzz...")
05:23:30 Join SxSBruce [0] (
05:32:51 Quit saratoga ("CGI:IRC (EOF)")
05:38:23***Saving seen data "./dancer.seen"
05:40:22 Join Llorean1 [0] (
05:40:50 Quit Llorean (Nick collision from services.)
05:40:53 Nick Llorean1 is now known as Llorean (
05:41:59 Join easy_way [0] (n=dave@unaffiliated/easyway/x-973744)
05:44:21 Join pabs_ [0] (
05:49:08 Join Nimdae_ [0] (
05:53:24 Quit Nimdae (Read error: 104 (Connection reset by peer))
05:57:08 Quit pabs (Read error: 110 (Connection timed out))
06:05:38JdGordonare the sliders in the graphical EQ really important?
06:05:57JdGordonI mean, to save space, could only the slider for the selected band be displayed?
06:07:10 Quit Falco98 ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
06:09:04aarcaneJdGordon, sure. but having the other sliders visible provides a frame of reference for adjustment and fine-tuning.
06:10:42JdGordonsure, but if there is not enough room on the screen the slider is the first thing to be removed yeah?
06:10:53JdGordonassuming you want to display as much info as possible
06:11:45 Join courtc_ [0] (
06:11:51aarcaneremoving or shrinking the sliders would make sense to free up space, yes.
06:15:30 Join Bensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
06:15:34 Quit XavierGr ()
06:20:56 Join nuonguy [0] (
06:22:38 Quit SxSBruce ("Leaving.")
06:23:50BensawsomeI dont knwo if anyone is on or not but... hey guys my database wont intilialize properly... it starts adding songs and a little after 8000 songs it freezes sayhing that mean prefetch abort error -_-
06:24:19 Quit courtc (Read error: 110 (Connection timed out))
06:25:21 Quit easy_way (Remote closed the connection)
06:25:32 Nick courtc_ is now known as courtc (
06:38:15 Quit JdGordon ("Konversation terminated!")
06:39:49Bensawsomeno guess not well nite :)
06:39:53 Quit Bensawsome ("The awsome is gone :(")
06:40:38 Join JdGordon [0] (
06:42:07 Quit jhulst (Read error: 113 (No route to host))
06:43:29 Join kkurbjun [0] (
06:55:17 Quit massiveH ("goodnight")
06:58:52 Nick pabs_ is now known as pabs (
07:11:03 Join J-23 [0] (
07:20:56 Quit kkurbjun (Read error: 113 (No route to host))
07:21:25J-23hmm, how does SD test on e200v2's work?
07:34:10 Join AndyIL [0] (i=AndyI@
07:38:28***Saving seen data "./dancer.seen"
07:42:11 Nick Darksair{away} is now known as Darksair (n=user@
07:45:12 Quit AndyI (Read error: 110 (Connection timed out))
07:48:44 Quit perrikwp (" ajax IRC Client")
07:48:57 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
08:10:49 Join Zarggg [0] (
08:16:03 Quit gevaerts (Nick collision from services.)
08:16:15 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
08:30:12 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
08:41:38 Join Rob2222 [0] (
08:59:47 Quit Rob2223 (Read error: 110 (Connection timed out))
09:13:43 Quit BHSPitLappy (Remote closed the connection)
09:38:30***Saving seen data "./dancer.seen"
09:51:35 Quit markun ("leaving")
09:51:59 Join markun [50] (n=markun@rockbox/developer/markun)
09:58:29 Join bertrik [0] (
09:59:22 Quit J-23 (Remote closed the connection)
10:00:59 Join J-23 [0] (
10:01:37J-23hmm, did you consider making Jabber notification system for SVN changes?
10:15:02 Join ender` [0] (
10:16:08amiconnlinuxstb: ping?
10:23:47 Quit amiconn (" reboot")
10:24:19 Join Shiny [0] (
10:27:02 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
10:33:53linuxstbamiconn: pong
10:34:18amiconnlinuxstb: (re logo and ucl)
10:39:25 Join merbanan [0] (n=banan@
10:43:10 Join Wenthof [0] (
10:44:09 Join n1s [0] (n=nils@rockbox/developer/n1s)
10:44:36 Quit Wenthof (Client Quit)
10:50:34linuxstbamiconn: I've just tried compressing a 172x54x16 logo (19008 bytes), and with the default blocksize it's 5839 bytes, -b8192 is 6025 bytes, -b4096 is 6270 bytes, -b2048 is 6645 bytes and -b1024 is 7235 bytes...
10:51:36linuxstbSome of those sizes would hopefully fit on the stack...
10:52:53amiconnEven if not, a 4KB or 8KB buffer might be worth the binsize savings (rockbox logo + usb logo should save more than 8KB on most colour targets)
10:57:57linuxstbAnd if we compressed the bitmap using a blocksize which is a multiple of the width, then the display of the bitmap can simply be a series of calls to uclunpack followed by lcd_bitmap(), so not much code.
10:58:04 Quit pixelma2 ("-")
10:58:36 Join pixelma [0] (n=Marianne@rockbox/staff/pixelma)
10:58:59 Join {phoenix} [0] (
11:01:39amiconnlinuxstb: That would mean extremely small blocksize thoguh, which also aren't a power of two
11:01:56linuxstbDoes a power of 2 mean anything with ucl?
11:02:06amiconnTry -b176 on your test file
11:02:12amiconnProbably not
11:02:14linuxstbThere's a check for < 1024...
11:02:34linuxstb1760 gives 6785 bytes...
11:05:18pixelmahmm... looks like we have a new record?... ;)
11:07:37*amiconn obviously overlooked the word 'multiple'
11:10:39linuxstbAh ;)
11:11:37amiconnThis multiple should then be chosen dynamically, as a best fit for our chosen buffer size (/me would probably go for 4KB)
11:12:32linuxstbWhat about a smaller buffer, but on the stack (and hence in IRAM...) ?
11:13:02amiconnHandling partial bitmap drawing will need some additional calculations in order to avoid unpacking unneeded blocks
11:14:17amiconnOne of my todos still is to make the bitmap drawing api more uniform, so that the caller doesn't need to know whether this is a mono, opaque or transparent bitmap
11:14:25amiconn(i.e. the bitmap header idea)
11:15:50 Join lasser [0] (
11:17:58 Nick Darksair is now known as Darksair{away} (n=user@
11:25:01linuxstbJdGordon: Reading the logs, what about vertical sliders for the EQ (either everywhere, or on height-restricted screens) ?
11:25:45amiconnWe tried vertical sliders for eq first... didn't work nicely
11:26:15amiconnThey need way too much room for the text
11:28:09 Quit denes_ (Read error: 110 (Connection timed out))
11:28:37 Join denes_ [0] (
11:28:49*linuxstb forgot there were three different things to adjust...
11:29:43pixelmawhat does the "HD1" exactly stand for in the lang files (for "multivolume" targets which are not Ondio or E200/C200) - "Hard disk" or should it be used more generically?
11:29:59amiconnhard disk
11:30:16 Nick bughunter21 is now known as bughunter2 (n=Jelle@
11:30:26amiconnIt's for hdd targets with multivolume (i.e. archos with >128GiB disk, needing 2 partitions)
11:31:10pixelmabut hard disk makes it sound as if you have two hard disks, not partitions
11:31:25pixelma(to me)
11:31:33amiconnThis isn't clearly differentiated atm anyway
11:34:09 Nick Darksair{away} is now known as Darksair (n=user@
11:36:29n1spixelma: it itsn't used in any default builds yet so i think gevaerts just put something there to be able to build.
11:36:57 Quit {phoenix} (Remote closed the connection)
11:37:25*n1s wonders if the patch he posted in fs9522 is doing the Right Thing (tm)? I am not so familar with the wole id3 buisness
11:38:33***Saving seen data "./dancer.seen"
11:41:35 Quit amiconn (" bbl")
11:56:00 Join {phoenix} [0] (
12:29:25 Join culture [0] (
12:40:04 Quit JdGordon (Read error: 104 (Connection reset by peer))
12:42:12 Join JdGordon [0] (
12:44:04 Quit reacocard (Read error: 110 (Connection timed out))
12:46:14 Join reacocard [0] (n=reacocar@
12:47:42 Quit reacocard (Client Quit)
12:48:00 Join reacocard [0] (n=reacocar@
12:55:38 Join kugel [0] (n=chatzill@unaffiliated/kugel)
13:03:26 Join domobile [0] (n=c3bd8eb4@gateway/web/cgi-irc/
13:05:45 Quit domobile (Client Quit)
13:05:58 Join domobile [0] (n=c3bd8eb4@gateway/web/cgi-irc/
13:06:08 Quit domobile (Client Quit)
13:07:23 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
13:17:08 Join moos [0] (
13:21:18 Nick fxb__ is now known as fxb (
13:21:20 Quit merbanan (Remote closed the connection)
13:28:32 Quit flux (Read error: 104 (Connection reset by peer))
13:28:36 Join flux [0] (
13:28:58 Quit bertrik (Remote closed the connection)
13:29:12 Quit AndyIL (Read error: 54 (Connection reset by peer))
13:29:42 Join AndyI [0] (i=AndyI@
13:30:41 Join SxSBruce [0] (
13:37:21 Part SxSBruce
13:38:35***Saving seen data "./dancer.seen"
13:43:55 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
13:50:08 Join PaulJam [0] (
13:53:03 Join bertrik [0] (
14:04:32 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
14:04:53 Join denes [0] (
14:19:01 Quit kugel (Read error: 110 (Connection timed out))
14:20:16 Join faemir [0] (
14:20:21 Join mofux_ [0] (
14:20:21 Quit homielowe ()
14:22:01 Quit denes_ (Read error: 110 (Connection timed out))
14:23:42 Join funman [0] (
14:25:26pixelmaa German (speaking) person around who has a nice idea how to translate the "Album Art:" line, stating the album art size of the loaded WPS in the "Rockbox Info" screen - preferrable not too English but still what people would use/understand?
14:27:38 Join fdinel [0] (
14:29:37domonokygoogle translator translates "album art" -> "album cover" :-) But i dont think that should be translated..
14:30:54funmann1s: the id3 tools I use fail on the file in FS #9522 , I'll try to find why (the id3v2 header is quite short)
14:31:10pixelmadomonoky: you would only use "Album art: "?
14:31:39domonokypixelma: yes..
14:32:45*rasher translated as album cover
14:33:01 Quit Horscht ("I got raided by the FBI and all i got is this lousy quit message")
14:33:02rasherCover is an often-used word in Danish. Art isn't.
14:33:06pixelmaI really have a problem with that space in there and it would be nice if it could be a bit more descriptive (that this is about the size)... but maybe that's impossible without making it a sentence :\
14:33:48pixelmaI'd probably prefer "cover" there too in German
14:34:01domonoky"Albumcover" would also work..
14:34:09blkhawkCover would be fine
14:34:21blkhawkAlbumcover yes
14:35:08n1sAFAIR the Swedish translation also uses cover
14:36:15 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
14:36:33pixelmaswedish translation uses the Swedish word for album cover there though (Skivomslag)
14:36:46domonoky"Cover größe" if you want the size in it :-)
14:37:10*pixelma often looks at some other translations for inspiration
14:37:58pixelmadomonoky: I'd use it without the space though, but IMO that's the best version so far
14:38:37pixelmaok, so it will be then. Improvements can always be done later too
14:39:24rasherOh, it's about the size?!
14:40:10n1sah, i thought people meant "cover" translated to that language, that's what i meant at least :)
14:40:39pixelmarasher: yes, states album art size of the loaded WPS in the "Rockbox info" screen
14:40:48rasherHuh, then the original string is pretty odd
14:41:16pixelmaI agree
14:44:12moosI'am working on resorting the english.lang, I renamed one item already (the one I ask about yesterday), +1 to this albumart name change
14:44:20pixelman1s: I could translate "cover" too but it's not very commonly used anymore so I'm afraid people would probably be more confused
14:44:50pixelmathe German expression I mean
14:52:28funmann1s: well the file in FS #9522 appears to heavily violate the id3 spec
14:53:10funmanthe strings are not 0 terminated
14:53:24 Quit nuonguy ("This computer has gone to sleep")
14:53:41funmanand the comment frame's header is missing the language and the comment description
14:55:54funmanhum well in fact the comment frame's header is fine .. ('eng'lish, 'iTunNORM')
15:05:19linuxstbbertrik: How different to your ascodec I2C code would a generic i2c driver be? Do you think the ascodec code can be adapted to be a generic I2C interface, or should they be two independent drivers?
15:05:40 Join tyfoo [0] (
15:07:14linuxstbfunman: Hmm, the "id3v2" tool shows it as "id3v2 v24.0"...
15:07:28funmann1s: hmm no in fact there is an empty string in a tag, due to iTunes authors not understanding an unclear part of the id3v2 spec
15:07:47funmanlinuxstb: here "id3v2" show "No ID3 tag"
15:08:04 Join kugel [0] (n=chatzill@unaffiliated/kugel)
15:08:08linuxstbAh sorry, I meant "file" shows v24.0
15:08:18funmanmost if not all text frames have the p flag set "Data length indicator"
15:08:20linuxstbAnd yes, id3v2 shows "No ID3 tag"
15:08:44funmanI don't even understand why this flag is here because every frame must indicate it's length already
15:09:21funmanso the parser skips 4 bytes (data length), and in the case of a 4 bytes frame (encoding + 3 bytes) it reaches the end of the frame > BOOM
15:10:20linuxstbfunman: What makes you think it was generated by itunes? I see "LAME3.94" near the end of the file.
15:10:25 Quit {phoenix} (Remote closed the connection)
15:10:38funmanlinuxstb: I didn't see what encoded the mp3 data
15:10:42*kugel removed himself from the sudoers
15:10:51funmanthe comment named 'iTunNORM' makes me think that
15:11:10linuxstbProbably a badly-written third-party app trying to add that tag...
15:11:11n1smy "file" outputs: "SOIZ.mp3: Audio file with ID3 version 2.4, MP3 encoding"
15:11:27linuxstb"SOIZ.mp3: Audio file with ID3 version 24.0 tag, MP3 encoding"
15:11:28funmanand also the iTunes-like naming of the file of course
15:12:18funmantry changing the bytes at position 0x33 and 0x5E from 01 to 00
15:12:30kugelhmm, id3 talk, maybe a good point to mention
15:12:39rasherHm, getid3 seems to grok it fine
15:12:49funmanyou'll notice that the tag loads, and the title tag is 4 bytes tooshort
15:13:29funmanrasher: what's the title of the tag ?
15:13:35rasher Title/songname/content description (TIT2): We are the champions
15:13:46kugelfunman: You wanted to help me finding the buttons after lcd is working
15:14:06Bagderanyone knows of a working URL to an existing h120 iriver firmware for install?
15:14:16Bagderit seems all links in the wiki are broken
15:14:19rasherBagder: my site
15:14:29funmanrasher: perhaps it detects broken iTunes tag or ignore the data_length_indicator flag (or I, rockbox, eyeD3, and id3v2 authors misunderstood it)">
15:14:55kugellinuxstb: I find something weird while experimenting. Depending on how I did various printfs, the backlight and buttonlight worked and worked not
15:15:46rasherfunman: no idea..
15:17:44funmann1s: comment added
15:17:55rasherfunman: it lists datalength as [datalength] => 21
15:18:02rasherFor the TIT2 tag
15:19:25funmanno, this is the common frame size
15:19:45funmanfollowed by a 2 bytes flags indicator which indicates data length indicator is present
15:20:26funmanI don't see why there would be 2 data length indicators though, 1 mandatory and 1 optional
15:21:02rasherWhat's the datalength indicator set to? Maybe getid3 catches that it can't be true, and picks the frame size instead?
15:21:12funmanthe data length indicator is not present
15:21:26funmanrasher: very probable
15:21:41rasherI see. That seems reasonable then - the flag is mistakenly set?
15:21:42funmanis it the getid3 php parser ?
15:21:49funmanI believe yes
15:22:46 Quit fdinel (
15:22:46 Quit faemir (
15:22:46 Quit domonoky (
15:22:46 Quit culture (
15:22:46 Quit lasser (
15:22:46 Quit markun (
15:22:46 Quit Rob2222 (
15:22:46 Quit miepchen^schlaf_ (
15:22:46 Quit esthar (
15:22:46 Quit blkhawk (
15:22:46 Quit agaffney (
15:22:46 Quit kachna|lappy (
15:22:46 Quit bughunter2 (
15:22:46 Quit blithe_ (
15:22:46 Quit linuxstb (
15:22:46 Quit gromit`` (
15:22:46 Quit Jabone (
15:22:46 Quit Tuplanolla (
15:22:46 Quit Bawitdaba (
15:22:46 Quit sbhsu (
15:22:46 Quit desowin_ (
15:22:46 Quit Dieterbe (
15:22:46 Quit fxb (
15:23:34NJoinfdinel [0] (
15:23:34NJoinfaemir [0] (
15:23:34NJoindomonoky [0] (n=Domonoky@rockbox/developer/domonoky)
15:23:34NJoinculture [0] (
15:23:34NJoinlasser [0] (
15:23:34NJoinmarkun [50] (n=markun@rockbox/developer/markun)
15:23:34NJoinmiepchen^schlaf_ [0] (
15:23:34NJoinesthar [0] (
15:23:34NJoinblkhawk [0] (
15:23:34NJoinagaffney [0] (n=agaffney@gentoo/developer/
15:23:34NJoinkachna|lappy [0] (
15:23:34NJoinbughunter2 [0] (n=Jelle@
15:23:34NJoinblithe_ [0] (
15:23:34NJoinlinuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
15:23:34NJoingromit`` [0] (
15:23:34NJoinJabone [0] (i=jpylvana@
15:23:34NJoinTuplanolla [0] (n=jani@unaffiliated/tuplanolla)
15:23:34NJoinBawitdaba [0] (
15:23:34NJoinsbhsu [0] (
15:23:34NJoindesowin_ [0] (n=desowin@
15:23:34NJoinDieterbe [0] (
15:23:34NJoinfxb [0] (
15:23:47NJoinRob2222 [0] (
15:24:15funmanin fact it reads the flag but doesn't use it ^^
15:24:18 Join PaulJam_ [0] (
15:25:10rasherSo it lucked out (this time)
15:25:26rasherStill, if the indicator is not present, using the frame size seems reasonable
15:26:01rasherOr just aborting reading completely on account of a broken tag. I've no real opinion one way or the other
15:27:52rasherHm, if the Data length indicator present flag is set, and all other flags are zeroed, using the frame size is safe
15:28:29funmanwell no
15:28:42funmanif the flag is set, the frame size is reduced from 4 bytes
15:28:54funmanwhich is what rockbox does, ignoring the meaning of the data length
15:29:20funmanperhaps a sanity check here would help as well
15:30:33funmanbut the minimal size is frame dependant (for text frames, 1 encoding byte and 1 data byte at east)
15:34:19 Join obo [0] (
15:37:35funmana comment in metadata/mp3.c lets me think that this file was tagged by iTunes 7
15:38:17 Join LambdaCalculus37 [0] (
15:38:38***Saving seen data "./dancer.seen"
15:39:17pixelmaI thought that this "iTunNorm" comment tag was their way of doing gapless? (Not sure it was mentioned yet)
15:39:37rasherPretty sure it's their way of doing replaygain
15:40:57domonokykugel: you want find the buttons on fuze ? want to try the adc ?
15:41:17funmankugel: oops sorry I had forgotten - too busy with this id3 thing :)
15:42:02domonokywith this code: in the bootloader, it should be possible to check the adc channels..
15:42:05kugeldomonoky: well, yes
15:42:08 Quit PaulJam (Read error: 113 (No route to host))
15:42:22kugelsince funman doubts they use gpio
15:42:26domonokyi have successfully measured the USB voiltage on a m200v2 :-)
15:42:31funmanI'll try to measure RAM performane with the RTC (1 second precision) because I can't use TIMER1 :/
15:42:55funmandomonoky: there is 2 registers no ? (10bits)
15:43:08domonokykugel: you have to vary the "2" in the codec_write to measure other channels..
15:43:26 Quit pixelma (" bbl")
15:43:27 Join pixelma2 [0] (n=marianne@rockbox/staff/pixelma)
15:43:28domonokykugel, yes. but for a first test i was happy with the first 8 bits :-)
15:44:02kugelhow many channels are there?
15:44:21domonokyoh, my code is wrong... :-/ i am reading the wrong register... lets retest.. :-)
15:44:50domonokythe datasheet states 16 channels, but only 12 are used..
15:45:08funmankugel: check page 162- of the datasheet
15:45:34*kugel should really take a look at the datasheet sometime
15:46:03funmanI think 1011b (= 11) should be the buttons
15:46:15 Quit Zarggg ()
15:47:02domonokychange the 0x2E in the codec read to 0x2F to really read the first 8 bits of the ADC :-)
15:47:50*funman wonders who broke the incomplete clip simulator
15:48:36domonokyto make it perfect, you should read both 0x2E and 0x2F and combine them (0x2F is the bottom word, and 0x2E the top)
15:49:02domonokymaybe i should just put this into a adc driver :-)
15:49:50domonoky(but it would need interrupt support to properly work..)
15:49:59funmanoh really, why ?
15:50:26domonokyconversion ready is only signaled by a interrupt, as far as i understand the datasheet..
15:51:28domonokyfor first tests its enough to just wait a bit.. but for proper adc support we would need them i think..
15:54:26funmanhmm .. again a data abort .. I wonder where it comes from
15:55:25domonokykugel: improved adc test code:
15:55:32funmanlooks like the code branches to an incorrect location
15:56:02 Join Zarggg [0] (
15:56:25funmanperhaps I should print a bit of the stack in the data abort handler to understand
15:58:04domonokyfunman: maybe its the cache ? did you try without enabled caches ?
15:58:38funmanyes it seems to prevent that, but since the data abort is not reproducable everytime I don't know if it won't happen later
15:59:01funmanthe cache is invalidated before being enabled, isn't that enough to use it ?
15:59:14domonokymaybe we miss some cache invalidate/ flush somewhere ?
15:59:53domonokyinvalidating it before using it should be fine, but maybe it gets confused as we init the sd-ram afterwards ?
16:00:29funmanpossible, I'll move the sd-ram init before and see
16:04:46bertriklinuxstb, yes I think it would be possible. The codec read/write stuff only ever reads 1 bytes at a time (also in the OF it seems), so it will get a little more complex to do more, but not much I think.
16:05:20bertrikthe i2c bus for the as3525 is an internal i2c bus as far as I understand from the datasheet
16:06:23bertrikdomonoky, I think we can re-use a lot of the code from the e200 for the things that access the codec part, like rtc and adc
16:06:23funmandomonoky: I get the same problem if I move cache enabling after sdram init
16:08:15bertrikfunman, I got data aborts too when accessing some registers of the PL180 too quickly in succession, the PL180 datasheet also mentions something like that
16:10:05*bertrik wonders how the radio of the clip and e200 is connected to the as3525
16:11:28funmansame :/
16:11:44funmandid you find the function which enable the peripherals ?
16:12:20bertrikcan't remember if I found a universal function
16:12:49funmanwell it only uses cgu_* so it may be only for as3525 peripherals
16:14:04funmanargh if I use the same test for IRAM and SDRAM, I got dataabort for SDRAM :'(
16:14:24funmanby the way if we don't use threads in the bootloader, is there a way to not compile the code ?
16:14:45 Join einhirn [0] (
16:16:50kugelfunman domonoky: I can
16:16:57kugelI can't seem to find any buttons
16:17:06kugelthe printed value hardly changes
16:17:14funmantry other sources
16:17:20domonokykugel: which channels have you tried ?
16:17:28kugelI tried 2, 9,10,11,12,13 now
16:17:55*domonoky would guess for channel 7
16:18:13domonokychannel 2 should change, when you insert usb..
16:18:14bertrikthe radio is of course not very essential in bringing the port forward, but sure looks like low-hanging fruit
16:18:25kugelalso, shouldn't high_res only be 2 bits?
16:18:32kugelres_high even
16:18:46domonokykugel: yes, its only the first 2 bits, but the others are 0...
16:19:31kugelsounds 3840 like a reasonable number for high_res (I wouldn't think so)
16:19:39kugelres_high again!
16:19:44funmanperhaps the RAM clock frequency is too high
16:19:44 Join massiveH [0] (
16:22:16funmanif I use 24MHz I have still the problem ..
16:22:36domonokykugel: 3840 dezimal in res_high ? that doesnt sound correct (thats more than 8bits)
16:23:09kugeldomonoky: yep
16:23:19kugelthat's why I wondered
16:25:21kugelfor channel 2: I get ~60(decrementing slowly over time) (res_low) and 768 (res_high), res_low changes to 63 after USB insert
16:26:09kugel768 is still too high
16:26:45funman(res_high & 0x3)<<8 should be more correct
16:27:39kugelthere's no &0x3 in domonoky code
16:27:58funmanI suggested you to add it
16:28:01domonokyres_low+(res_high&0x3) << 8 is working better...
16:28:09funmanto keep only the 2 last bits of res_high
16:28:29funmansaratoga: IRAM is 33% faster than SDRAM
16:28:30domonokyalso the channel number should be shifted by 2...
16:29:07 Join easy_way [0] (n=dave@unaffiliated/easyway/x-973744)
16:29:20domonokyso to read the USb voltage it would be as3525_codec_write(0x2E,0x2 << 2);
16:30:14kugeldomonoky: 2 or 0x2? in your code it was 2
16:30:15funmansaratoga: IRAM is 33% faster than SDRAM (on the AS3525, with pclk and fclk at 24MHz)
16:30:34*funman hands a hexadecimal howto to kugel (2 == 0x2)
16:30:41bertrikyou could also have a look at adc-as3514.c
16:30:41domonokykugel: thats the same hexadezimal 2 = decimal 2 :-)
16:30:49kugelof course, sorry
16:31:49funmandionoea: did you install the bootloader that shows the rockbox logo on your sister's clip ? ;)
16:36:40funmankugel: try adding that before checking the adc: as3525_codec_write(0x14, (1<<1));
16:36:49funmanit enables 'microphone 2 input'
16:38:59linuxstbbertrik: Did you look at my "ascodec" commit for the existing as3514 targets? (e200v1, c200v1, philips sa9200)
16:39:45 Join Darksair` [0] (n=user@
16:41:19linuxstbAre you going to be the Mr Someone who implements it for the as3525? ;)
16:42:00*funman looks for Mr Someone who finds why the sansav2 embedded SD card is not powered up
16:42:11n1sfunman: what do you think about fs#9522, should both patches go in?
16:42:14 Quit Darksair (Read error: 54 (Connection reset by peer))
16:43:07funmann1s: if the patch I posted is not a concern performance-wise, then yes
16:44:59kugelfunman: on channel 11 (neither with nor without left shifting the channel) I get no reaction on button pressing
16:45:11kugelbut with shifting, the number quite changes upon usb insert
16:45:23funmanand on channe 7 ?
16:45:54n1sfunman: I don't se why it would be a concern, meta data reading is hardly time critical but maybe we should wait for Lear or someone to comment
16:46:28 Quit PaulJam_ (".")
16:47:34 Nick mofux_ is now known as mofux|bbl (
16:48:05funmanI'll try to implement watchdog for sansav2, to prevent from crashes when the power-off is disabled
16:50:23kugelnothing on channel 7
16:51:00 Join lee000000 [0] (n=48fb23c7@gateway/web/cgi-irc/
16:51:16funmanthen you'll have to go through disassembly ..
16:51:16bertrikkugel, what are you trying to measure anyway?
16:52:55lee000000Hey guys. I'm in Ubuntu trying to run '', but am reciving this message "ROCKBOXDEV: patch is required for this script to work."
16:53:08 Join {phoenix} [0] (
16:53:49 Quit Horscht ("We don't make mistakes, we just have happy little accidents")
16:54:55kugellee000000: sudo aptitude install patch
16:55:01kugelbertrik: buttons
16:55:42lee000000sorry, I was thinking "a" patch (new to Ubuntu) thanks.
16:56:55kugelno problem
16:56:55 Quit lee000000 ("CGI:IRC (EOF)")
16:57:13kugelfunman: the ram on the fuze is 8MB (the of tells that)
16:59:03funmankugel: then I don't know why it uses the 2MB model in the init routine
16:59:10linuxstbkugel: That would make more sense - 2MB for a relatively large colour LCD isn't very much...
16:59:10soapJdGordon|zzz, you could always show just the EQ slider in question + the neighboring (max two) sliders. This would remove two from the screen and give you your frame of reference.
17:00:17kugelcan we disassemble the diagnosis mode (which is enabled when you name the firmware with t at the end)?
17:00:25kugelthere's 1 test for all buttons
17:00:40funmanyou can disassemble everything
17:00:59kugelok, let me rephrase
17:01:00funmanI couldn't find the button code when I tried however
17:01:02kugelcan you do that? :)
17:01:17funman:p no I prefer to spend my time on the Clip OF
17:01:42funmanbut linuxstb seems to like writing code for the models he doesn't own :))
17:01:58*linuxstb is just sticking with easy lcd code
17:02:04kugeljust by disassembling the OF! Yes, I should ask him ;)
17:04:30 Quit kugel ("ChatZilla 0.9.83 [Firefox 3.0.3/2008101315]")
17:04:43 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
17:04:44 Join kugel [0] (n=chatzill@unaffiliated/kugel)
17:05:34kugelone part of the test is filling the display with red. If we found that in the disassembly, that'd mean we're in the diagnosis mode?
17:05:39 Quit Darksair` (Read error: 54 (Connection reset by peer))
17:05:59 Join Darksair` [0] (n=user@
17:23:39 Join stoffel_ [0] (
17:38:39***Saving seen data "./dancer.seen"
17:45:50funmanI wonder what the watchdog isr is supposed to do
17:46:14funmaneven if the system is in a 'bad' state, the isr would still be called, so only clearing the watchdog register without any check is useless no ?
17:47:14funmanthe OF doesn't seem to use watchdog by the way
17:49:21funmanI think system_reboot() could be implemented by enabling watchdog and disabling interrupts
17:54:12funman < watchdog & interrupts for as3525
17:56:23 Join kkurbjun [0] (
18:00:06funmanin fact I would like to know if this isr effectively prevents the system to be locked before trying to lock my Clip ;)
18:02:27*domonoky tries to understand the code.. you set up the watchdog to rise a IRQ when it timeout ? and what does this irq do then ?
18:03:18funmanprevents the watchdog unit from resetting the system
18:03:25domonokycouldnt you just setup the watchdog without IRQ (so it reboots on timeout), and reset the watchdog regularly ?
18:03:46funmanright .. that makes more sense
18:05:46*domonoky thinks thats the common use of watchdogs. If the code fails to reset the wd in time, the system is "out-of spec" and reboots.. even the marsrovers do this :-)
18:07:29*n1s thinks marsrover code is probably like 20% do stuff 80% make sure suff is done right
18:11:30pixelma2suff = German "getting drunk" ;)
18:11:58funmanI'll keep watchdog local, but I've committed interrupts code
18:13:23 Join fml [0] (n=4fd3cbb4@gateway/web/cgi-irc/
18:15:06fmlNow that there is a plugin that spits out all the available config options to a file, couldn't we introduce some settings without an entry in the menu? I.e. such settings would only be settable through direct editing of the config file. I'm thinking of many "set once and forget" settings.
18:15:13 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
18:15:37fmlThus the menus would be made less cluttered, and we'd save some bin size as well.
18:15:43funmanwell the watchdog unit only works with interrupt, so I'd have to regularly toggle some value in memory and make the isr read it
18:16:10n1sfml, also changing settings would be more difficult an inconvenient.
18:16:55fmln1s: we'd have to agree on what settings would stay in the menus. But some can be excluded IMHO
18:17:47n1sif they are not worth having in the menu i think they are not worth having at all (btw we already have at least one such setting, probably more)
18:18:14pixelma2I don't think hiding settings is a good idea, there already is one (or was?) to switch off the left/right scrolling when long pressing left or right
18:18:40pixelma2I guess not many know about it
18:18:52fmln1s: having a thing as a setting and having a menu entry for it isn't the same IMHO.
18:19:30n1sfml of course it's not the same thing, i just don't think having hidden settings is a good idea
18:20:01fmlSince, for some values, it would be nice to be able to change the setting, but it's mostly only done once to one's own taste and then never changed.
18:20:18n1sit is a) less convenient to change b) less likely users will know about c) will get less testing etc.
18:21:03n1sfml: well, i don't think such settings should be removed from the menu because they are not changed _often_
18:21:48fmln1s: re a) they are documented and exposed in the automatically generated file. We could also make a plugin for editing a cfg file (based on the file produced by that plugin)
18:22:37 Quit jon-kha (Remote closed the connection)
18:23:20n1sfml: many users don't read the manual, do you think they will launch some obscure plugin that gives them a list of settings? How is a cfg editing plugin different from the "settings menu in a plugin" idea that has been discussed over and over
18:24:05fmlThis would on the one hand, allow to have many settings thus making RB very customizable but on the other hand not strike a new user with too many settings. After having used RB for some time, a ser would then look into the manual / automatically generated file and fiddle with the other settings.
18:25:03 Join jon-kha [0] (
18:25:26jchillerupWhat's the channel for the IPL for nano project?
18:25:45kugelfunman, domonoky: this should work yes? I'm getting values of ~15000
18:26:01*domonoky thinks about something like about::config for rockbox, but only for thing where we dont want to spend a setting...
18:26:07 Quit einhirn (Read error: 104 (Connection reset by peer))
18:26:20kugel+" suggest parentheses around + or - inside shift" warning
18:26:43domonokykugel: could you paste it into another pastbin ? this one doesnt work for me..
18:26:58kugelwhat doesn't work? is sloooooooow
18:27:20domonokythe website.. it always timesout for me..
18:27:24funman((res_high....) << 8) , use these parenthesis
18:27:27n1sIMHO, if something is settable it's a setting, and settings should be in the settings menu, hiding them solves nothing except maybe reducing the sheer abount of settings in the menu but leads to having settings in different less convenient places.
18:27:40pixelma2jchillerup: I guess everyone around here would have to google too...
18:27:42kugelworks fine for me..
18:27:43 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
18:27:44*linuxstb has never had problems with - it's working fine now
18:27:54jchilleruppixelma2: tried that
18:28:01fmlBased on the work done by JdGordon, I think it would not be very hard to write a generic cfg editor (as a plugin).
18:28:11kugeldomonoky: ^
18:28:35jchilleruppixelma2: I want to read about the encrypted firmware
18:29:04fmlThis might make the threshold for new settings a bit lower. Because the main reason for rejecting them IIUC are the cluttered menues and bin size, not the settings themselves.
18:29:26domonokykugel: shouldnt the enabling of the mic_input (the 0x14) be before the start of conversion ?
18:29:49linuxstbjchillerup: It wasn't hard to find -
18:30:00kugelfunman: gcc doesn't warn because the parameter isn't in parenthesis, it rather warns because it's not clear whether to shift the whole sum or only the second operand
18:30:09n1sfml: the only difference between that and the "settings menu in a plugin" is that you keep some of the settings menu in the core and hide some of it away in a plugin so you lose the "benefit" of lower binsize and add another place to search for the setting to change...
18:30:11funmankugel: so make it clear for gcc
18:30:30kugelI did if I knew which is really intended
18:30:41jchilleruplinuxstb: well, then your google foo must be better than mine. Thank you very much.
18:30:49kugelI'm not much more intelligent than gcc for code you give me
18:30:54funmanyou shouldn't run code if you don't understand it ..
18:31:44domonokykugel: res_high should be shifted (after the masking) so it gives the top bits for the 10bit adc value..
18:32:24*domonoky tries to correct the code... the delays are worng..
18:32:32fmln1s: no. Core setting would stay in the main RB thus allowing e.g. live setting change. But rarely used settings could go away from the menues.
18:32:46kugelfunman: I understood it, and I put parenthesis around ((res_high&0x3) << 8), but since it didn't work I thought it could be the other way
18:32:59funmanwhat didn't work ?
18:33:34J-23hmm, what does as3525_codec_read function do?
18:33:39kugelI got no response for button presses (not necessarily the code's fault, I know)
18:34:30 Join toffe82 [0] (
18:34:58domonokyforgot the parenthesis:
18:35:22 Quit Darksair` (Remote closed the connection)
18:35:23n1sfml: that is _exactly_ what i said, you keep some in the core and have some in a plugin
18:35:46 Join Darksair` [0] (n=user@
18:35:46domonokykugel: it could also be, because your delay loops where worng, so maybe it didnt wait long enough for the conversion..
18:36:27kugeldomonoky: I didn't know I need to enable the mic input before starting the conversion (funman said I should put that before the read)
18:36:40domonokyJ-23: as3525_codec_read/write speak via i2c with the audio/pmu unit of the cpu in the v2s...
18:36:50 Join Lear [0] (i=chatzill@rockbox/developer/lear)
18:37:20domonokykugel: i dont know for sure, but it would make sense, as the write starts the adc sampling, the read only fetches the results..
18:37:21pixelma2fml: I very much dislike hidden settings in general (e.g. I always switch off those "intelligent" menus in MS Office in case you know those). I also dislike relying on a plugin for settings which are core functionality - what if your install went wrong?
18:38:32pixelma2and maybe you only realise later when you're not at the computer anymore
18:39:09 Join nuonguy [0] (
18:39:11 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
18:39:38kugelI don't consider it bad when rockbox relies on the fact that the installation didn't go wrong
18:39:49 Quit jhMikeS (Read error: 110 (Connection timed out))
18:40:03 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
18:40:20*domonoky wouldnt object against something like about::config in rockbox, but only for obscure settings, like the spacing chars for scrolling lines..
18:41:42kugelwhy not for all settings?
18:41:58kugelwhy have even more special cases?
18:42:03 Join fredddy [0] (
18:42:20bertrikehm, why rely on random delays for adc if you can also poll a register?
18:43:07bertrikbit 0 of register irq_enrd_2 should tell you if the adc is done
18:43:28funmanWriting 1 to GPIOA_PIN7 (which is also the power button when read) shuts down the Clip
18:43:32domonokybertrik: it was just a first test (which worked for me). which register can you poll for conversion ready ?
18:44:17bertrikfunman, I think a long pulse (like a person holding the power button) shuts it down, with a short pulse the software polls the power button state
18:45:27funmanoh ok
18:45:34funmanso if we leave it to 1 it will be a long pulse
18:47:21pixelma2kugel: really? I'll ask you again after that happened to you (ending up in a position where you can't change the setting for a whole day mayb)
18:47:34pixelma2*maybe too
18:47:35kugelbtw: the fuze shuts down pretty fast, I only need to hold power for a split second
18:48:07kugelpixelma2: Ok, then lets support failed installations
18:48:09funmansame thing for the clip
18:48:33domonokypixelma2: wouldnt you just boot with the "reset settings" button pressed ?
18:48:48 Join PaulJam [0] (
18:49:02pixelma2domonoky: if the default settings are not what I want?
18:49:14kugelpixelma2: Also, you assume the installation went good enouigh to copy the whole rockbox binary and only some/all plugins aren't on the player
18:49:18domonokythe shutdown time seems to be configurable, its longer in the OF..
18:49:33funmanit's configurable in a i2c register - don't remember which
18:49:40funman0x21 maybe
18:49:47domonokypixelma2: but resetting should be enough to get it usable again ?
18:49:52funmanit's in 0x15 0x16 0x20 or 0x21 for sure
18:50:23 Join tyfoo2 [0] (
18:50:44BigBambiI would object loud and long to needing a PC to change settings
18:50:46 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
18:50:59 Quit tyfoo2 (Client Quit)
18:51:28*domonoky also woudnt see the need of PC for setting "hidden" settings, what about a settings editor, or just the text editor ?
18:51:35 Join tyfoo2 [0] (
18:51:43pixelma2domonoky: I didn't talk about settings being screwed up so you can't use Rockbox at all, just something like changing a setting at all (maybe turning off repeat or so)
18:51:50kugelthe text editor is a plugin too
18:52:11*domonoky wouldnt put such important settings into the "hidden" section...
18:52:12kugeland in this case, the rockbox binary is copied fine, but the plugins not
18:52:46pixelma2domonoky: referring to the general idea of having settings in a plugin
18:52:47domonokyonly settings which arent really needed, but some people might like to change. like this space charactor for the scrolling line...
18:53:21pixelma2all settings
18:53:24 Join mailman1175 [0] (
18:53:33*domonoky is confused... we surely dont want all settings in a plugin..
18:53:52 Join mc2739 [0] (
18:54:10pixelma2the idea was brought up
18:54:17domonokybut haveing a way to introduce many less used features without harming usablility would be good...
18:54:27kugeldomonoky: sounds 800 like a reasonable value? (back to the adc test)
18:54:52domonokykugel: thats a more sane value...
18:55:06mailman1175hello all: potential rockbox user w/ a question (if I can interrupt topic in progress)
18:55:23bertrikmailman1175, sure, just ask
18:55:33 Join MethoS- [0] (
18:56:00mailman1175I have an automotive headunit that has iPod control.. what are the chances of rockbox confusing/not working with it?
18:56:38 Quit massiveH ("Leaving")
18:57:35domonokymailman1175: current rockbox probably doesnt work with this thing (maybe lineout works, but no controlling).. but there is a patch on the tracker about apple accessoary support...
18:57:56linuxstbmailman1175: You can easily "dual-boot" into the original Apple firmware when in your car - so if Rockbox doesn't work with it, you still have that option.
18:57:58funmandoesn't it use a USB connection ?
18:58:18funmani.e. mass storage usb, not using any apple specificities
18:58:49mailman1175funman: It uses the standard iPod connector, but I don't know if it uses apple specifics or not
18:58:51domonokyfunman: ipods a have a serial connection in the dockport , which is used by many "ipod accessorys"..
18:59:28domonokyso if the automotive unit allows controlling of the ipod, it surely uses the serial port, and not USB...
18:59:42mailman1175domonoky: that's correct
19:00:11mailman1175is dual boot pretty easily configured for a linux-noob?
19:00:33funmanwith all the GPIO pins set to output and high (except A7-power, and A4-which crashes the Clip) the is still not powered up
19:00:50funmanmailman1175: yes you just have to press some key combination
19:01:05domonokymailman1175: you dont have to configure it, its just a button press while swithcing the dap on...
19:01:20mailman1175cool; thx guys
19:02:04 Part mailman1175
19:02:28*domonoky thinks there is even a custom build in the forums to test the accessory patch.. it might work with this.
19:02:51pixelma2fml: by the way - you'll have a second problem to solve when you introduce hidden settings: who decides which settings is worth having in the core and which is a "second class citizen"...
19:03:37J-23can I help somehow with fixing USB drivers?
19:04:41 Quit tyfoo (No route to host)
19:05:24linuxstbdomonoky: Yes, Soap maintains one, but mailman1175 left....
19:05:49*domonoky is too slow again :-)
19:12:16fmlpixelma2: sorry, I've been away for a couple of minutes. My idea was not to have _all_ setting menues in a plugin, only rarely used ones. Core settings like sound settings, repeat mode etc. would probably stay in the core.
19:14:29amiconnfml: So which settings are rare?
19:14:38*J-23 tests logf via USB :)
19:14:47pixelma2yes, I know (was just generalising a bit) - that's why I added the second class citizen statement later
19:14:50amiconnImo all settings must be accessible on device, through the menus
19:15:01BigBambiIs there any real reason that the apple accessory patch hasn't been committed, other than nobody has?
19:15:03fmlA cfg editor plugin would not be just a text editor, it would be much smarter since we have the information about the possible values for each setting. So for each setting we could offer a menu with the possible values. The default value would be marked in a special way. But there would be no live reaction. I.e. you'd have to load the created cfg file to activate those settings.
19:15:12*BigBambi agrees with amiconn
19:16:05fmlamiconn: e.g. ff/rew acceleration and minimal step; list cursor (arrow/block); bidirectional scrolling %. Justt off the top of my head. There are more, I'm sure
19:16:06amiconnThe settings UI could be moved to a plugin, but there are prerequisites which aren't met yet
19:16:51pixelma2fml: so instead of one discussion you have to lead, there are two... ;)
19:17:31fmlpixelma2: even three: n1s, amiconn, and you! :-)
19:18:22pixelma2no, I meant the second class question
19:18:58fmlamiconn: do you mean that once we're able to localize and voice the plugins, nothing would be against moving all the menus to a plugin?
19:19:25amiconnDefinitely not all the menus. Just settings which need to hit the disk anyway
19:20:02amiconnI don't want to wait extended times just to see the menu
19:20:39 Quit soap (Remote closed the connection)
19:20:54amiconnMaybe it's not even a good idea to do this in general, but it might be necessary for target with really tight ram (Clip...)
19:21:22 Join Strife89 [0] (n=michael@
19:21:27kugelthis discussion makes me tired. I see no hope of a consensus at all, JdGordon's mail and the discussion afterwards didn't bring a consensus either. So we should actually stop discussion and ask the rsb maybe
19:21:29*amiconn thinks that it might even enable a rockbox port to that old Neo thingy...
19:21:49linuxstbamiconn: The Clip isn't that tight - it's now been discovered to have 2MB SDRAM, plus 320KB IRAM...
19:21:56kugelamiconn: the clip has more then 2MB ram, so not that tiny
19:22:30pixelma2kugel: did you check how much RAM you use on your e200?
19:22:36*kugel notices that fact didn't spread enough, people are still arguing with the clips tiny ram against any sort of bloat
19:22:39*linuxstb thinks we need to heavily reduce that 1MB codec buffer soon though...
19:22:50kugelpixelma2: why my custom build or without?
19:23:04amiconn2MB on swcodec is still tight
19:23:13kugelabout ~3MB, including dircache and load to ram
19:23:26linuxstbkugel: 2MB is still small for a swcodec target - once you think of voice files, database in RAM, dircache, codecs, ....
19:23:57domonokyif we shrink the codec buffer a bit, current rockbox should fit into that space...
19:24:16amiconnA typical swcodec target needs ~3MB ram with my rather conservative settings
19:24:30*domonoky s iriver shows it needs 2,8 MB ram..
19:24:31kugellinuxstb: well, but the clips monochrome display already saves a lot of ram
19:24:43amiconnkugel: s/a lot/a bit/
19:24:56funmannot to mention how tiny it is, even if it's bigger than archos
19:24:59kugelyep, a bit*
19:25:27pixelma2kugel: I don't see a big difference between my greyscale M5 and the colour c200
19:27:12 Quit easy_way (Remote closed the connection)
19:27:52*pixelma2 goes to compare an X5 with the M5
19:28:50 Join mf0102 [0] (
19:31:41pixelma2free buffer with default settings: 13.2MB on M5 vs. 12.9MB on X5
19:32:13funmanhum .. why the 10MB difference with the ~3MB usage mentioned before ?
19:32:52kugelfunman: have you tried putting a firmware with t suffix on your clip if it has a diagnosis modeß
19:33:57kugeldoes it have one?
19:33:58 Quit fml ("CGI:IRC 0.5.9 (2006/06/06)")
19:34:07funmanyes - I mentioned it in the forum thread
19:34:26 Quit stoffel_ (Read error: 113 (No route to host))
19:34:32kugeldid the diagnosis mode show you the ram? it showed 2.3 probably
19:34:55funmanit shows 2MB
19:36:59kugeldomonoky: just curious, what was the reason to shift the adc channel number?
19:37:21funmankugel: are you reading the as3525 datasheet ?
19:37:45domonokykugel: if you look at the datasheet, you see that the lowest 2 bits in this registers are bit 8 and 9 of the result..
19:37:52funmanthe source is set on bits 7:4 of register 0x2e
19:38:08domonokyso the channel number beginns at bit 3
19:38:31funmanbits 3:2 are described as '0, do not change'
19:38:44***Saving seen data "./dancer.seen"
19:38:56kugelbut if I shift << 2, I end up in 3:2
19:39:10domonokyso the channel number should be shifted by 4... :-/
19:39:29kugelyes, that's why I asked :)
19:51:24n1skugel: the mailing list discussion was AFAICT not about a hidden settings menu or even about settings in a plugin (that was discussed a bit but was really off-topic) and many of us have pretty firm views of things so we usually have to compromise.
19:54:28 Quit fdinel (Read error: 110 (Connection timed out))
19:56:02 Quit fredddy (Read error: 110 (Connection timed out))
19:56:29 Join fredddy [0] (
19:59:48 Join mofux_ [0] (
20:01:30 Join kushal_12_27_200 [0] (n=kushal@
20:01:33 Join stoffel_ [0] (
20:01:58 Join soap [50] (n=soap@rockbox/staff/soap)
20:02:05kugelfunman: I don't think I need to test other channels than 7 and 11, do I?
20:02:27funmankugel: just read the datasheet, I don't have more information than you
20:04:05kugelfunman: yes, I'm reading. I doubt they override checking voltages and temperatures
20:05:21domonokychannel 6 could also be used..
20:11:04 Join m0f0x [0] (
20:11:21 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
20:12:06 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
20:14:21kugeldomonoky: can't we use while(i2c_busy()); instead of delays?
20:15:12domonokykugel: thats not the correct check. we are not waiting for the i2c but for the adc conversion...
20:15:22kugelah ok
20:15:49 Quit Darksair` (Client Quit)
20:15:56domonokybut it should be possible to check bit 0 of IRQ_ENRD_2
20:15:58funmanthe as3525_codec_*() functions do that already I guess
20:16:19 Quit mofux|bbl (Read error: 113 (No route to host))
20:16:30kugelfunman: yes, but they return instead of waiting
20:17:03domonokyso it would be (while(as_codec_read(0x27) & 0x1 != 0)
20:17:06funmanah right
20:20:27kugeldomonoky: read returns a negative value on non-success
20:21:17 Join mofux [0] (
20:21:35 Quit mofux_ (Read error: 60 (Operation timed out))
20:22:12 Join merbanan [0] (n=banan@
20:22:12domonokykugel: yes, but that only happens, when somehow a i2c transfer is already running.. they wait at the end for transver finish... so normaly no transver should be running,when you enter this functions..
20:22:58kugelah right, i see
20:23:21 Quit martian67 (SendQ exceeded)
20:25:16domonokyand we have to wait for adc conversion to finish, not i2c transfers... so the above should work for the delay between the adc write and read... all other delays are not really needed...
20:25:37 Join martian67 [0] (i=lol3izer@about/linux/regular/martian67)
20:26:08 Quit fredddy ("Verlassend")
20:26:27*funman wonders if the value 0x55 is in the i2c registers table, or just after it ..
20:29:13 Quit stoffel_ ("leaving")
20:34:33 Quit MethoS- (Remote closed the connection)
20:35:15domonokyfunman: as far as i can see from the datasheet, the registers go from 00h - 3Fh, so 0x55 is way to high...
20:36:16funmandomonoky: something could not be printed in the datasheet, but I just tried writing to this register without success
20:36:40funmanstrange that this 0x55 appears in the middle of a table used to store i2c registers
20:36:48 Join MethoS- [0] (
20:37:43domonokymaybe sandisk added something there ?
20:38:21funmanjust like they added a pl180 controller on top/instead of the AS3525 NAND controller ? maybe ..
20:39:22pixelma2moos: while you are at it - I just noticed that in LANG_OFF_ABORT there is a missing space between gigabeatf: and the following " (already in english.lang). Not sure how important that is but maybe you can change that
20:40:09 Join robin0800 [0] (
20:40:45moospixelma2: sure I will, don't esitate if you spot something else...
20:43:55 Join petur [50] (n=petur@rockbox/developer/petur)
20:45:59kugelfunman: so, the buttons are not on adc. does it mean they're on gpio or is there another possiblity?
20:46:32funmanno idea
20:46:49funmanthere could very well be an undocumented controller
20:46:54bertrikthe fuze isn't so different from an e200 right?
20:47:14funmanexcept it is based on the same SoC, you can imagine everything
20:47:56bertrikso couldn't it also use a keyscan matrix?
20:48:31funman_could_ , of course
20:49:13kugelI'd like to find out
20:49:16bertrikmost probably I would think
20:50:42kugelI don't know where to start though
20:52:06bertriklook in the OF with a disassembler, I estimate this would take about a day
20:52:32 Join Thundercloud [0] (
20:52:34funmanbertrik: hum did you find the nand/sd functions table in clip firmware v30 ? the offset I had given you was wrong ..
20:53:09 Join Schmogel [0] (
20:53:24kugeldo the e200 button driver work in svn?
20:53:51 Quit martian67 (SendQ exceeded)
20:54:21soapOk, the Apple Accessory patch appears to be out of sync, and beyond my skillset. I thought I manually merged the one hunk which was failing - but that does not seem to be enough to make it work.
20:54:49linuxstbWhere are the problems/
20:55:13soapone sec
20:55:41funmanbertrik: ah it starts at 0x6E04
20:55:59 Join martian67 [0] (i=lol3izer@about/linux/regular/martian67)
20:56:02funmanand in Clip v29 0x6E08 - not big difference ..
20:57:22soapOnce again I cry 'Wolf''. :( I though I had all traces of the prior patch cleaned, but now it appears that I didn't.
20:57:37 Quit Seed ("cu, Andre")
20:58:54funmanbertrik: at 0x6EE4 and 0x6F0E there seems to be bytes tables which are used as addresses on the i2c bus, and 0x55 appears in them - maybe it's a kind of terminator, or INVALID_ADDRESS
21:00:07bertrikhmm, I don't know, I think on the as3514 the codec addresses above 0x3F were aliased back to 0
21:00:22 Quit moos (Read error: 110 (Connection timed out))
21:00:33 Join Seed [0] (
21:00:41 Join stoffel_ [0] (
21:00:50funmanwell when I try to use 0x55 my Clip freezes it seems
21:03:54 Quit martian67 (SendQ exceeded)
21:04:50 Quit {phoenix} (Remote closed the connection)
21:04:57 Quit mofux (Read error: 110 (Connection timed out))
21:05:48 Join mofux [0] (
21:06:29 Join martian67 [0] (i=lol3izer@about/linux/regular/martian67)
21:07:14bertrikoh, the sansa fuze buttons and layout looks very similar to the layout of the clip
21:08:22 Quit nuonguy ("This computer has gone to sleep")
21:09:36funmanthe screen (and the device) is bigger, and there is a scrollwheel
21:09:54funmanto me it looks exactly the same
21:10:47linuxstbDoes the fuze have any buttons on the sides?
21:10:51bertrikthe hold/power button is on the right side instead of the left side (can't find the volume control)
21:11:41*funman calls for an e200v2 owner who is ready to open it for unbricking
21:12:27*mc2739 is here
21:12:34bertrikkugel, what kind of volume control does the fuze have?
21:13:07funmanmc2739: I think we should try to hijack the OF initialization with our custom routine, which would print the values of all i2c registers, what do you think of that?
21:13:51funmanthere is 64 1 byte registers it could even fit on the whole screen
21:14:08kugelthe fuze pretty much combined the controls of the clip and e200
21:14:15funmanwhen we wanted to do that with daniel_at the problem was to find where this routine would fit
21:14:43funmanhum the problem is that we don't have printf() in the OF :/
21:17:24bertrikis it really a wheel that you can rotate, or just 4 buttons (on for each direction)?
21:17:34n1sfunman: if you have enough space you could just jump to the rockbox bootloader for example (by patching in a suitable jump somewhere) to get lcd, etc drivers
21:18:12n1s(removing and hw init from the bootloader first of course
21:18:14funmanyes maybe we can tweak a custom bootloader
21:18:26funmanhoping that the OF doesn't bzero() the whole SDRAM
21:19:28funman1/ decompress our bootloader to SDRAM, 2/ decompress OF in IRAM, 3/ add a branch at the right place in the OF (tm) 4/ access i2c bus and printf it
21:19:43funmanmc2739: do you feel confident to code that ?
21:20:42funmanlet me try to do that then
21:20:50funmando you already have opened your e200 ?
21:20:58 Quit Strife89 ("Bah, got to unload the van now.")
21:21:04funmangood :)
21:21:06kugelbertrik: it's a wheel. it rotates nicely (better than the e200 one). The directional buttons are integrated
21:21:09n1sI assume this can't be done by disassembly? maybe looking at emulators is a good idea? (a simple one was used in the early stages of the e200v1 port)
21:21:33funmann1s: well I've been reading disassembly for months already
21:21:40mc2739I also have verified that I can get into the recovery mode
21:21:53funman+ we would need to write all hardware devices for the emulator
21:22:47n1sfunman: if it is only an init you could just log what it writes to different addresses when booting maybe?
21:23:03funmanit is 'an' init made of zillions functions
21:23:48funmanluckily for lcd / buttons / SDRAM there was only 3/4 functions and most of the code was sequential so it was easy to reverse engineer
21:23:56funmanfor i2c access it looks more complex
21:24:13domonokydumping the registers after the of init, is probably a good idea. and if the sd-ram trick for lcd doesnt work, we could also use led-dumping...
21:25:43 Join {phoenix} [0] (
21:26:03 Quit n1s ()
21:32:37 Quit amiconn (" bbbl")
21:33:12 Quit kugel ("ChatZilla 0.9.83 [Firefox 3.0.3/2008101315]")
21:34:13 Quit nplus (Remote closed the connection)
21:35:10 Quit Thundercloud (Remote closed the connection)
21:36:50 Quit bmbl (Read error: 104 (Connection reset by peer))
21:37:21 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
21:37:54 Join Thundercloud [0] (
21:38:30 Quit J-23 ("Flying cow pressed ^D on my keyboard.")
21:38:47***Saving seen data "./dancer.seen"
21:39:27funmanlinuxstb: hum .. we don't setup sp register in dualboot.S ...
21:39:30bertrikanyone have a link to a fuze firmware?
21:40:05linuxstbfunman: We don't use the stack either... Or do we?
21:40:32bertrikI think we should start a sansa fuze v2 wiki page
21:40:50funmanoh the decompressor uses stack only for (disabled) extra checks
21:40:56linuxstbbertrik: On Bagder's V2 page -
21:40:57funmanbertrik: check the sansa forums
21:44:56 Join tvelocity [0] (
21:52:17 Join lee000000 [0] (n=48fb22de@gateway/web/cgi-irc/
21:52:31 Quit HBK (Read error: 104 (Connection reset by peer))
21:52:58 Join HBK [0] (
21:56:54 Quit toffe82 (Read error: 110 (Connection timed out))
21:57:11 Quit lee000000 (Client Quit)
21:58:31 Quit robin0800 (Remote closed the connection)
21:58:51 Join robin0800 [0] (
22:01:20 Join webguest20 [0] (n=3e2fb2fb@gateway/web/cgi-irc/
22:06:37 Quit webguest20 ("CGI:IRC")
22:12:42funmanmc2739: can you try ? for now it only decompresses & execute bootloader from SDRAM (conditionally on C3)
22:14:19funmanhmm wait
22:14:52funmanno it's ok, pop/push use full descending stack, just like stmfd/ldmfd
22:15:13 Quit aarcane (Read error: 113 (No route to host))
22:16:09mc2739funman: trying now
22:17:07 Join nuonguy [0] (
22:21:36 Quit Thundercloud (Remote closed the connection)
22:24:13 Join Guest24535 [0] (
22:26:35 Nick tyfoo2 is now known as tyfoo (
22:27:47mc2739funman: no good - no response on powerup
22:28:31funmannot even while keeping left pressed ?
22:29:36 Join JThundley [0] (
22:30:16mc2739bbl - unbricking
22:30:37funmansorry ^^
22:30:49JThundleyis there a way to disable booting to my ipod's original firmware? Can't I just move some file off?
22:31:00JThundleyI want to boot to rockbox and nothing else :)
22:31:01mc2739no problem
22:31:22 Quit mc2739 ()
22:32:04soapJThundley, see the bottom of the page:
22:32:44 Quit pixelma2 ("-")
22:32:53JThundleysweet I'll read up, thanks!
22:32:58 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
22:35:04LearBagder: I think I have an ugly fix for the failed sim builds on cygwin: remove any generated dependencies for rockboxlogo.h, usblogo.h and default_icons.h. The files including these also include recorder/icons.h, so the icons.h dependency on the bitmap libs is enough. Don't know why these dependencies aren't generated by the cross compilers (gcc-m68k 3.4.6 and gcc-arm 4.0.4)...
22:35:53peturis there a way to AND features required for LANG entries? Or should I just create a new feature?
22:37:31funmanah I messed up ..
22:39:39 Join lard-Vader [0] (
22:40:02 Join Acknix [0] (
22:40:41 Join webguest49 [0] (n=3e2fb2fb@gateway/web/cgi-irc/
22:40:59 Quit webguest49 (Client Quit)
22:41:12lard-VaderHi, can someone explain to me the different repeat modes, especially "shuffle"
22:41:23lard-Vaderisn't there already an option to turn shuffle on and off?
22:42:00domonokymaybe take a look at the manual ? :-)
22:42:28lard-VaderI couldn't find anything with google
22:42:40lard-VaderI will try
22:42:44domonokynot google, the rockbox manual..
22:42:57 Quit {phoenix} (Remote closed the connection)
22:43:58kushal_12_27_200Hello, I would like to resolve the issue of "Refreshing database" on my sansa C250. I am working on an Intel Mac 10.4.11. sudo mount -t msdos /devdisk1 /tmp/sansa gives me error. Any ideas?
22:43:59domonokysection 7.2 ...
22:45:55 Join Metamorphus [0] (
22:46:06funmankushal_12_27_200: /dev/disk....
22:46:18funman/dev is a directory
22:47:36kushal_12_27_200thank you , funman.
22:48:43 Quit stoffel_ ("leaving")
22:49:22 Join toffe82 [0] (n=chatzill@
22:51:18Guest24535Hi, for some very useful reason, I have some ogg files ending with .wav. rockbox tries to play them with the wav decoder and fails horrible. Is rockbox really that focused on filename and does not parse the content or did I miss something?
22:51:35 Part lard-Vader
22:52:35bluebrotherRockbox relies on the filename extension
22:52:41domonokyGuest24535: how horrible is it failing ?
22:52:52bluebrotherbut for what reasons does it make sense using wav for ogg files?
22:53:00domonokyrockbox should crash if it finds a broken file...
22:53:05*kushal_12_27_200 is a bit annoyed by the Refresh Database on his sansa c250 which did not go away even after going to recovery mode to put fresh copies of mi4 and rom
22:53:14MetamorphusIf I want to buy a player for running rockbox on, which one would be recommended?
22:54:00domonokyMetamorphus: take a look at:
22:54:17Metamorphusdomonoky: thank you
22:54:31bluebrotherdomonoky: Rockbox _should_ crash? ;-)
22:54:42 Join lee000000 [0] (n=48fb22de@gateway/web/cgi-irc/
22:54:45*domonoky reverts his earlier statement... it misses a "not" :-)
22:54:52Guest24535horrible is a little dramatic. As it skips all the files, it just goes to the next folder it can play.
22:55:03Guest24535fails gracefully...
22:55:12BigBambiThat is quite different :)
22:55:24domonokyGuest24535: thats correct behavior, it just skips files, it cannot read...
22:55:37Guest24535Which shows me, that the decoder knows how not to handle unknown files.
22:55:50BigBambiEr yes, if it is unknown how can it handle them?
22:56:02Guest24535But detecting the right format and totally ignoring the extension does not seem that complicated
22:56:18BigBambiKeep in mind these are very limited resources platforms
22:56:19 Quit merbanan (Read error: 104 (Connection reset by peer))
22:56:21bluebrotherGuest24535: patches are always welcome ;-)
22:56:25lee000000I just built on Ubuntu (first time not in Cygwin), and when I start my Sansa c240 I get "Prefetch abort at FFFFFFFe (0)" −− any ideas?
22:56:26domonokyguess how bad it would be to not go by the file extension. We would have to probe with every decoder if it works... and we support many formats...
22:57:13bluebrotherwell, possibly this could be done by simply searching for some magic bytes at the beginning of the file
22:57:40bluebrotherthough I guess this wouldn't be enough to distinguish all supported formates
22:57:44Guest24535I guess just giving the file to all decoders available and just testing, which ones likes the data is kind of a brute force attack on files with wrong extensions.
22:57:49BigBambiI guess the question is that if the file name suggested decoder fails, then how feasible is it to try others.
22:57:57*domonoky wants see code from Guest24535 which can reliably distinguish all of our 20-30 supported music formats..
22:58:36*BigBambi suggests giving files their correct extensions :)
22:58:57 Join mc2739 [0] (
22:59:33lee000000I just built on Ubuntu (first time not in Cygwin), and when I start my Sansa c240 I get "Prefetch abort at FFFFFFFe (0)" −− any ideas?
22:59:37mc2739funman: I'm back
22:59:55BigBambiGuest24535: If you can come up with a good and fast way of checking reliably each of the formats Rockbox supports, I'm sure people would be interested :)
23:00:00kushal_12_27_200hello, I have a sansa c250 using the latest rockbox nightly (October 14 actually). When loading the original firmware, the sansa seems to be perpetually stuck in Refresh Database screen. I will not be able to get anything into or out of the sansa as the OF never finishes booting. (I could charge it, but that's all) Please advise.
23:00:11*bluebrother still wonders what might be a good reason to name an ogg file "wav2
23:00:29bluebrotherlee000000: no need to repeat, we heard you.
23:00:46 Quit mf0102 ("Ex-Chat")
23:00:50bluebrotheryou need some patience on irc. At least usually.
23:01:02bluebrotherhow did you set up the toolchain?
23:01:08BigBambilee000000: Can you try the same revision downloaded from the Rockbox site?
23:01:24Guest24535As the only useful way seems to be some magic byte parsing and there is no support for this kind of stuff yet in the codebase... Maybe its easier to find a solution for the wrong extension problem locally here. It wouldnt be that complicated, just some script doing renames back and forth.
23:01:33lee000000they only thing I did was run
23:02:00mc2739kushal_12_27_200: have you tried the c22 erase -
23:02:27Guest24535But as you mentioned and that is why I wondered: Parsing for format would be quite smart and I would like it. But maybe it is not worth the work.
23:02:31bluebrotherGuest24535: please enlighten me why ogg files should be named wav in the first place at all
23:03:27Guest24535The unix command "file" recognizes all file formats I use perfectly. Do not know if all. And surely do not know how much work a port to rockbox would be.
23:03:54MetamorphusI'm a bit confused, is Sansa E280 supported?
23:04:14mc2739Metamorphus: the v1 is, the v2 is not
23:04:16bluebrotherMetamorphus: yes. e200 is the series
23:04:19domonokyMetamorphus: yes, but only the "v1" ones...
23:04:34MetamorphusHow do I know if I'm being v1 or v2?
23:04:42funmanmc2739: I have found some omissions in my code
23:04:50kushal_12_27_200I don't have access to a Linux box atm ... However, I did something else that seems to have worked. I put sansa in recovery mode and put in fresh versions of mi4 and rom. then I let the OF boot. It finished just now and I got into the OF.
23:04:54mc2739look at the firmware version v1.xx = v1
23:04:56funmannow I try to run it in the emulator first to see if it's alright but I have some trouble .. :o
23:05:02BigBambiMetamorphus: What is the version of the Sansa firmware?
23:05:35bluebrotherGuest24535: well, "file" is not running on a low-mem, low-speed cpu and it has no problems looking at the complete file if necessary
23:05:41mc2739kushal_12_27_200: good
23:06:03MetamorphusI haven't bought it yet, isn't there another way?
23:06:11BigBambiNot reliably
23:06:13bluebrotherit might be an interesting starting point, however. In fact I'm trying to figure what bytes file is looking at right now ;-)
23:06:21BigBambiMetamorphus: But if it is new, it is probably v2
23:06:33MetamorphusOkay, I guess it is then
23:07:02domonokyMetamorphus: if you are in the us, you could buy from froobi they have garantied rockboxable sansas...
23:07:59kushal_12_27_200Thank you mc2739. The OF on my sansa is 1.00.06P. I formatted the sansa just in case. (Rockbox was unable to delete one directory inside "Photos" which could be related to why OF was unable to refresh database.) Do you think it would help/hurt me if I updated the OF using sansa software before putting rockbox again?
23:08:11Metamorphusdomonoky: Unfortunately, I don't live in the US
23:08:24Guest24535Oh, the source of all the problem: My symbian device saves all kinds of audio stuff in wav files. However a ogg decoder is installed and all programs that want to play files use the file system supported decoders. As symbian and epoc never look on file extension, it is therefor easy to recode all wavs to oggs on that platform transparently. The app just loads it file, ending in .wav, also if us ogg.
23:08:59mc2739kushal_12_27_200: sorry, but I do not know which version OF is best for the c200
23:09:05Guest24535But I do backup the symbian smartphone on my rockbox hd. And when I try to play the audio files inside the backup...
23:09:07 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
23:09:57kushal_12_27_200mc2739, do you think updating would break rockbox compatibility?
23:10:09Guest24535Its easy to rename of course. But to be a real backup, it is a copy to another directory or I have to rename backwards before restoring the data.
23:10:52lee000000I also ran this 'echo "export PATH=$PATH:/usr/local/arm-elf/bin" >> ~/.bashrc' at the terminal (is that part of setting up a toolchain?)
23:10:55Guest24535No big deal at all. But of course not as elegant, as if rockbox just did play the stuff.
23:11:09mc2739kushal_12_27_200: probably not
23:11:46kushal_12_27_200mc2739, will I ever be able to roll back the firmware if it does?
23:12:24linuxstbGuest24535: I can understand why you would want Rockbox to do this, but I don't think anyone else has ever asked for Rockbox to play files with wrong extensions...
23:13:13 Quit LambdaCalculus37 ("Ka-chunka")
23:14:46lee000000kushal_12_27_200: I'm pretty sure all v1 firmwares are Rockbox-compatible, but that above a certain number you can't put the player's USB mode to MSC.
23:15:05Guest24535Maybe it is not the most popular feature. And in fact: If a developer would ask me, if she should implement this or some other feature, most possibly I would vote for that other feature.
23:15:08mc2739kushal_12_27_200: as far as I know, you can just put the older version in the root directory and it will "upgrade" to that version
23:15:38lee000000kushal_12_27_200: I'm pretty sure you downgrade too.
23:15:47lee000000you = you can
23:15:56kushal_12_27_200thanks to both of you.
23:17:40 Quit petur ("*plop*")
23:19:08Guest24535I was a little shoked at first, btw. Because I did never understand, why microsoft outlook opens an pdf email extension of mime type pdf, with a filename ending in wav with the media player. Ir is just plain wrong. And now my beloved rockbox does something similar. :-)
23:19:13funmanmc2739: this works fine in the emulator so at least you will be able to boot OF if SDRAM init works ;)
23:20:39mc2739funman: ok - trying
23:20:40 Quit Lear ("ChatZilla 0.9.83 [Firefox 3.1b2pre/20081031032638]")
23:21:14lee000000i'm gonna download the latest pre-built...
23:21:25 Quit domonoky (Read error: 104 (Connection reset by peer))
23:21:27 Quit lee000000 ("CGI:IRC (EOF)")
23:23:05 Join webguest19 [0] (n=48fb22de@gateway/web/cgi-irc/
23:23:08funmanmc2739: do you use firmware V03.01.16A ?
23:23:39 Join Razvan [0] (n=Razvan@
23:23:46 Join Thundercloud [0] (
23:24:06webguest19I think there's a typo at −− "If you have a c250, download c240erase.bin" −− should be c240?
23:24:25mc2739funman: yes, that is the version I am using
23:24:26 Quit webguest19 (Client Quit)
23:24:41Razvani wanna know if rockbox wil work with a Philips Go Gear SA28xx ?
23:26:10funmanRazvan: if I look here I don't see this model, so no
23:26:12linuxstbRazvan: All devices Rockbox works on (even partially, or even just people thinking about it) are here - If it's not there, Rockbox doesn't work on it.
23:26:35Razvanok thank you
23:26:51 Quit robin0800 (Read error: 104 (Connection reset by peer))
23:27:32 Quit Razvan (Client Quit)
23:27:49 Quit Metamorphus (Read error: 60 (Operation timed out))
23:32:12mc2739funman: still no good - no response
23:33:12funmandid you try both with and without button press ?
23:33:26funmanperhaps I should switch the butotn led on to be sure if the device is running or not
23:33:31mc2739yes - left button
23:35:35mc2739funman: can you leave in the usb check to get back to the OF?
23:35:58funmanhum yes
23:37:34funmanargh I messed the button check
23:38:04funmanmc2739: if you go to next2: you see we enable gpio clock
23:38:14funmanbut I forgot ldr r0, =GPIOC
23:38:20mc2739yes, I see
23:38:50***Saving seen data "./dancer.seen"
23:39:06 Quit bluebrother ("leaving")
23:40:38funmanif we can't execute properly from SDRAM we'll copy back the code to read & display i2c into IRAM
23:41:06 Quit Acknix (Read error: 104 (Connection reset by peer))
23:42:50 Part Guest24535
23:44:03 Join lee000000 [0] (n=48fb22de@gateway/web/cgi-irc/
23:46:11mc2739back in a bit - unbricking again
23:46:21 Quit mc2739 ()
23:46:53funmanweird, the sdram_init() function I build doesn't restore the registers it modifies
23:47:14funmanwell, not r0, r1, r2, r3 at least (the arguments registers), even if I declared it void
23:47:28lee000000BigBambi: I asked a little while ago about getting this 'Prefetch abort at FFFFFFFe (0)' on booting into Rockbox. I fixed it by completely deleting the '.rockbox' folder on my player before putting the new one on it. Is there an obvious reason I'd get the error?
23:48:07BigBambiI don't know I'm afraid, but you shouldn't have to do that
23:48:25BigBambiMaybe something got corrupted? Perhaps a disk scan would help?
23:49:26lee000000BigBambi: I tried a pre-built .rockbox of the website, and got the same error. Could it have to do with the fact that I haven't upgraded for over a month?
23:49:37BigBambiShouldn't do, no
23:49:49BigBambiWhat OS are you on?
23:50:16lee000000I'm on Ubuntu right now (but have XP on same computer)
23:50:25 Quit lasser ("ChatZilla 0.9.83 [Iceweasel 3.0.3/2008092816]")
23:50:57lee000000how do you run a disk scan in a Unix system?
23:51:04BigBambiIt might be worth running a disk check - perhaps one of Rockbox's control or settings files that don't get updated are corrupted
23:51:48lee000000< running scan...
23:54:43 Quit Schmogel (Read error: 104 (Connection reset by peer))
23:56:46lee000000fsck.vfat "SANSA C240" −− takes only a split-second to complete, and doesn't saything like "OK" or "bad"...?
23:57:02 Quit toffe82 (Read error: 110 (Connection timed out))
23:57:41 Quit bmbl ("Woah!")
23:57:52 Quit Llorean (Read error: 104 (Connection reset by peer))
23:58:13 Join Llorean [0] (
23:58:53lee000000my current dir should be "will@Presario:/media$ " right?

Previous day | Next day