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 2005-02-23

00:00:01stripwaxSPV - Allow DMA Access (only SRAM1) - 0, DMA access to memory is disabled, 0, DMA access to memory is enabled
00:00:01 Quit amiconn (Nick collision from services.)
00:00:02 Nick amiconn_ is now known as amiconn (
00:00:58hubblestripwax: oh no.. i havn't done that =)
00:01:04rasherstripwax: Aha! I had forward slashes - have to be backslashes
00:01:16rasherstripwax: thanks a lot
00:01:21hubblestripwax: thanks
00:01:27stripwaxhubble no worries!
00:01:30stripwaxrasher no worries! :-)
00:03:54hubblestripwax: although linus already did that in crt0.S
00:05:03stripwaxhm, as far as I know DMA0 and DMA1 work the same so I don't know :(
00:05:16hubble move.l #0x1000e001,%d0
00:05:16hubble movec.l %d0,%rambar1
00:06:17stripwaxhubble? bit 9 needs to be a 1.. should be 0x1000e021 ..
00:06:28hubblelinus is evil =) he says that he enables dma in the comment but bit 9 is not set =)
00:07:39hubble0x1000e201 perhaps
00:08:04stripwaxer, 1000e201 of course ;)
00:11:32hubblehm.. almost.. got some sound now but then 1ms later it crashed
00:11:38stripwaxMight also be worth marking either or both of the 32KB segments in SRAM1 as having DMA priority over CPU, but then, might not!
00:11:52stripwaxoh, crashed isn't good
00:12:02preglowwell, if it's for sound buffer, that's clever
00:13:01hubblegoing to try that
00:14:24stripwaxHmmm... should be 0x1000ce21 .. that original 'e' looks bogus ?
00:15:50hubblemaybe linus places that e one wrong.. he says in the comment both banks dma priority
00:16:05hubbleso 0x10000e01
00:16:22stripwaxnot, 0x1000ce21
00:17:38stripwaxthe 0x1000c000 controls the base address, 0xe00 controls the DMA priority and DMA enabled flags, 0x21 controls the address space masks (i.e. code/data)
00:19:27stripwaxOh, er.. which means SRAM1 base is 1000c000? So maybe it should be 10000e21 after all. What's RAMBAR0 set to?
00:20:01hubble/* 64K DMA-capable SRAM at 0x10000000
00:20:01hubble DMA is enabled and has priority in both banks
00:20:01hubble All types of accesses are allowed
00:20:01DBUGEnqueued KICK hubble
00:20:01hubble (We might want to restrict that to save power) */
00:20:59stripwaxyeah, should have been 0x10000e21 not 0x1000e021. SRAM1's been running from 1000c000 i.e. overlapping SRAM0?!
00:21:03stripwax(can't be good!)
00:21:29hubblemaybe explains one or two random crashes =)
00:25:31preglowsram isn't used much yet
00:25:58 Join ripnetUK [0] (
00:26:21hubblewell.. 0x10000e01 boots, but dma still hangs
00:26:50stripwax.. e21 right, not e01?
00:29:34hubble =)
00:30:21stripwax? still hangs? still crashes?
00:31:15hubblestill hangs
00:34:07stripwaxwhat's the value of DMAROUTE (MBAR2+0x188). dunno if that would affect anything
00:35:52stripwaxnah, ignore that. I guess just try DMA1 and see if that fixes. May be a hardware quirk?
00:37:03stripwaxThis is where you need bdm, to look at the status registers.. :(
00:37:15 Quit ripnetUK ()
00:37:18 Quit Sucka ("( :: NoNameScript 3.81 :: )")
00:37:45preglowso only channel 1 works to sram?
00:38:30hubblehave not tried that, but i have seen iriver firmware does that
00:45:14stripwaxOnly thing I can really think of is alignment of the buffer within SRAM could be different than alignment within SDRAM ..?
00:47:19stripwaxis that it?
00:47:23hubblethink so
00:52:11 Join skav [0] (
00:55:57CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
00:55:57*amiconn got the iriver x11 sim to compile
00:56:33amiconnAre there test files for the various codecs somewhere?
00:59:21preglownone that i know of
00:59:24***Saving seen data "./dancer.seen"
01:00:31amiconnWell, I have plenty of mp3s, and some oggs as well. But I'd like to test flac and ac3 as well...
01:06:41preglowonly flac i've got is a little snippet i used to test if my lpc routine worked
01:06:51preglowand it's around 5 secs, heh
01:10:25 Quit Patr3ck ()
01:12:14 Quit markun ()
01:17:45 Quit midk ("Leaving")
01:18:42 Quit preglow ("yes")
01:20:43 Quit stripwax (Read error: 110 (Connection timed out))
01:20:59 Join Goatmale [0] (
01:22:07GoatmaleI think my jukebox recorder is broken..
01:23:48 Join stripwax [0] (
01:23:59stripwaxugh. my ISP sucks
01:24:17stripwaxhubble, so what's up. did you try changing alignment
01:24:47GoatmaleNo help me
01:24:59*Goatmale jumps around for attention
01:26:41GoatmaleMy archos i think is broken
01:26:45GoatmaleI think..
01:26:55Goatmale"read access error!
01:27:08GoatmaleCan't read from hard drive"
01:28:08amiconnlinuxstb: What does the splash "Recoverable!" mean in mpa2wav?
01:29:33hubblestripwax: same as before.. but i have now discovered that it doesnt work for any static buffer either SRAM or SDRAM.. The only buffer that works is one got using "buffer_alloc"
01:29:50hubblewhich is on sdram
01:29:56hubblevery strange
01:30:37stripwaxvery very strange. definitely sounds like alignment wrt static/dynamic buffer on SDRAM
01:30:50hubblealign 16 should be enough?
01:31:04*Goatmale prepares to take apart his archos
01:31:04stripwaxI'd have thought so.. :-(
01:32:12linuxstbamiconn: It's an error from libmad - that the stream is corrupt. If you've experienced it, then it's probably due to ID3 tags - my mpa2wav program doesn't deal with any tags, it just expects a "pure" mpeg audio stream.
01:32:29amiconnAh okay, then this is the problem
01:32:41amiconnI tried mpa2wav in the win32 sim, and got this
01:33:11amiconnBtw, I can't test in the x11 sim. If I try to call the onplay menu, the sim crashes....
01:33:48linuxstbI rename the mp3 files to ".mad" and then setup viewers.config to associate .mad files with mpa2wav
01:34:13linuxstbWas it you that was working on improving the X11 sim's keyboard handling?
01:34:15amiconnDoes the onplay menu also crash for you?
01:34:34amiconnYes, I looked into it.
01:34:53linuxstbI think the keys just auto-repeat, and I can't get it to come up. Or it comes up, and then immediately selects the first item.
01:35:11amiconnFor me, the sim crashes...
01:35:40linuxstbMaybe the mouse buttons would be an easy fix - I'm guessing the mouse doesn't auto-repeat.
01:35:44amiconnAnyway, this menu works perfectly in the win32 sim, so I'll strip tags from my test mp3
01:36:07amiconnThe problem is not the auto repeat, but that the kb handling is not realtime
01:36:25amiconnSo when I simply deactivate autorepeat, the long press is still not detected
01:36:42amiconnI have an idea how to solve that, without going multi-threaded
01:37:11amiconn...xlib doesn't like multithreaded x11 apps :(
01:37:54 Quit lolo-laptop ("Client exiting")
01:39:20 Quit mecraw ()
01:40:43stripwaxhubble - there's not additional stuff making its way into SRAM1 is there? will align16 *really* align?
01:41:13 Quit Goatmale ()
01:41:46hubblestripwax: it looks like that in
01:42:37stripwaxhubble - looks like it is, or isn't, aligning correctly?
01:43:05hubblestripwax: according to .map it is aligned
01:44:31 Quit Renko (Read error: 54 (Connection reset by peer))
01:47:28amiconnlinuxstb: I now have a stripped .mp3 mpa2wav immediately exits...
01:47:40stripwaxhubble what happens if you create a static buffer which is (say) twice as large as you need?
01:51:29hubblestripwax: dont know, will try
01:52:07stripwaxhubble actually will probably make no difference, DMA shouldn't/won't care what the data in the buffer actually is..
01:53:15linuxstbamiconn: Does it work on the iRiver? Do you have an iRiver?
01:53:29amiconnI don't have an iriver.
01:55:52linuxstbCan you try the X11 sim (rename the file to, for example .mad and change viewers.config)
01:57:59amiconnI just tried that, the sim crashes when I play that file
01:58:37linuxstbI'm just testing if the X11 sim still works for me.
02:00:28 Quit Aison ("( :: NoNameScript 3.72 :: )")
02:00:29hubbletime too sleep =)
02:01:31 Part hubble
02:02:44linuxstbamiconn: Yes, mpa2wav works fine in the X11 sim on Linux. Do you want me to test your file?
02:03:50amiconnI don't think it's necessary. I am doing this because I want to get the codec plugins working on windows
02:04:17linuxstbDo any of them work?
02:04:36amiconnI tried only mpa2wav so far. Can try ogg now.
02:04:45amiconnI don't have test files for the other 2"> and - test FLAC and AC3 files I put there for Linus the other day
02:06:10 Quit pill (Read error: 104 (Connection reset by peer))
02:06:49amiconnNice, thanks. I'll try them
02:08:16amiconnHuh, funny: Deliplayer (multiformat music player for windows) can play flac.
02:09:31linuxstbOne thing I noticed, is that the API documentation for FLAC (in fact, all the documentation for FLAC) is by far the best out of any of the codecs - which is one reason why it's so widely adopted.
02:10:39amiconnHmm. None of these seem to work.
02:10:56amiconn(x11 sim on cygwin, that is)
02:11:17amiconnmpa2wav crashes the sim, the other pretend to work, and wav files are produced
02:11:46linuxstbWhy do you say "pretend" ?
02:11:59amiconnHowever, none of these play, and the files produced from .ogg and .ac3 have odd sizes
02:14:41linuxstbWhat do you mean by "odd sizes"?
02:15:48amiconn9587 KB .ogg -> 2392 bytes (yes!) .wav
02:16:10amiconn1375 KB .ac3 -> 61 KB .wav
02:16:35 Join slowdowncadet [0] (
02:16:35amiconnYes, that's the word
02:17:02amiconnThe .wav produced from .flac looks like about the correect size, however it doesn't play either
02:17:06slowdowncadeti'm stripwax, and that other stripwax should disappear shortly... ISP hell ...
02:17:21amiconn3410 KB .flac -> 7709 KB .wav
02:18:23linuxstbYes, that FLAC should be 7893356 bytes as a WAV
02:18:56amiconn(7893400 Bytes here
02:19:40linuxstbI get 7893356 bytes running it through the X11 sim.
02:20:23linuxstbYou have an extra 44 bytes which is the size of the WAV header.
02:20:28amiconnThe win32 sim produces similarly odd results
02:21:19amiconnAt least mpa2wav doesn't crash it
02:21:35 Quit stripwax (Read error: 110 (Connection timed out))
02:22:27 Nick slowdowncadet is now known as stripwax (
02:22:38linuxstbThe only "unusual" thing my plugins do is to grab the MP3 buffer using rb->plugin_get_mp3_buffer()
02:23:02amiconnWell, that's not unusual. Several other plugins do that too
02:23:57linuxstbBut looking at them, they are mainly the plugins that don't work on the simulator - video, jpeg, battery_test, plus vbrfix and sort
02:24:12amiconnWin32 sim: 3410 KB .flac -> 19 KB .wav, 16044 KB .mp3 -> 44 Bytes
02:24:28amiconnsort should work on the sim...
02:25:41amiconnYeps, sort is working
02:26:04 Join lostlogic [0] (
02:27:30 Join ashridah [0] (
02:28:08amiconnlinuxstb: A few hours ago you had a crashing simulator, and used gdb to check where it crashed. How do I do that?
02:28:50 Quit edx (Read error: 110 (Connection timed out))
02:28:52linuxstbType "gdb rockboxui" and then type "run" at the gdb prompt. Beyond that, I don't know.
02:29:27amiconnI already tried that. I get an error, but no source line or such...
02:29:49 Join gromit``` [0] (~gromit`
02:30:06amiconnIt segfaults right at the start. It doesn't do that when run without gdb...
02:31:12 Quit gromit`` (Read error: 104 (Connection reset by peer))
02:32:50ashridahamiconn: type 'bt' when it crashes and drops you back to the gdb prompt
02:32:55ashridahthat'll give you a backtrace
02:32:59ashridahparticularly if it's compiled with -ggdb
02:34:24amiconnHmm, bad. Either question marks or win kernel functions....
02:37:02ashridahmm. you may or may not have much success in some cases, particularly if it crashes inside code that you didn't compile. extra libraries can often be located that have debugging info built in, which can help
02:37:08ashridahnot sure where you'd find them for windows tho
02:42:30 Quit stripwax (Read error: 110 (Connection timed out))
02:43:26 Quit gromit``` (Read error: 104 (Connection reset by peer))
02:43:29 Join gromit``` [0] (~gromit`
02:44:22amiconnHmm. Now that is even more strange: I googled a bit, the SIGSEGV when using gdb under cygwin seems to happen now and then. The suggestion is to simply type 'continue' several times. This actually works!
02:45:04amiconnHowever, and now it gets esoteric: If I do this, and then try the codec plugins, the .wav files have reasonable sizes!
02:47:03*amiconn suddenly realises what the problem might be
02:47:30amiconnAre there any endian checks in the codec libs? I guess there are...
02:47:54linuxstbI'm running on x86 Linux, so there shouldn't be any difference because of that.
02:48:13amiconnYeah, but the endian #defines on cygwin are fucked up....
02:48:40amiconnI already had that problem, with the id3 database browser not working on cygwin...
02:48:53linuxstbI don't use them - only liba52 needs to know the endianness, and at the moment that's hardcoded to little-endian for simulators, and big-endian for targets.
02:48:55amiconnI added a workaround, but that one is not global
02:49:36linuxstbYes, we need a "TARGET_ENDIAN" or similar define in the build system. But that'a a discussion for another day.
02:53:31amiconnThe .ogg decoder is still running. Runs at 15%, that is on an 1.5 GHz Pentium M...
02:54:06linuxstbYes, something evil is happening inside that ogg decoder.
02:54:26amiconnShouldn't Tremor be fast & efficient?
02:54:37linuxstbYes, that's the theory.
02:56:00linuxstbAnyway, time for sleep. Goodnight.
02:57:16 Quit lostlogic ("Going to the moon")
02:59:27***Saving seen data "./dancer.seen"
03:08:10 Join lostlogic [0] (
03:12:32XShocKSRAM DMA works. :)
03:12:55 Quit YouCeyE (Remote closed the connection)
03:16:26 Join midk [0] (
03:19:13XShocKFull speed. :))
03:19:38XShocKand noone here.. :)
03:20:33 Join YouCeyE [0] (foobar@youceye.user)
03:20:33ashridahXShocK: nice
03:20:43ashridahgot any performance stats?
03:20:46amiconnXShocK: Nice :)
03:21:07XShocKperformance stats?
03:21:09amiconnI found at least one of the problems with th cygwin simulators: lseek() is not working.
03:21:29amiconnAnyway, I need to sleep
03:21:57ashridahXShocK: well, is whatever you're doing faster/slower/thesamebutlesscpu/etc ?
03:21:59 Quit lostlogic ("Going to the moon")
03:22:20XShocKi guess it is much faster than AUDIOTICK
03:22:38XShocKbut i didn't do any tests
03:22:45XShocKjust got it working
03:25:11XShocKand what do i do with the code? should i implant it somewhere?
03:25:21XShocKin rockbox
03:28:11XShocKanyhow, i will cleanup the code first... it veery dirty.. :)
03:29:26 Quit midk ("Leaving")
03:31:02 Join Strath [0] (
03:37:43XShocKok. see you all, i'll go to bed
03:39:11 Join StrathAFK [0] (
03:43:08 Join lostlogic [0] (
03:43:49 Quit XShocK (" Try HydraIRC -> <-")
03:47:04 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
03:52:49 Join XShocK [0] (
03:53:26XShocKupdated wiki for sound. i think 40% for sound is appropriate
03:53:43XShocKsince input is not done, and framework is not done too
03:55:13XShocKexhausted... zzzzzz
03:55:27 Quit XShocK (Client Quit)
03:58:55 Quit Strath (Read error: 110 (Connection timed out))
04:05:29 Join QT_ [0] (as@area51.users.madwifi)
04:05:36 Join Mong0 [0] (
04:10:29 Join edx [0] (
04:14:02 Quit QT (Read error: 60 (Operation timed out))
04:42:49 Join IRCMonkey [0] (
04:59:30***Saving seen data "./dancer.seen"
05:07:29 Quit Mong0 ("bye")
05:22:34 Quit IRCMonkey ("ChatZilla 0.9.52B [Mozilla rv:1.6/20040113]")
06:39:37 Join midk [0] (
06:46:01 Quit Sando (Read error: 54 (Connection reset by peer))
06:50:27 Quit skav ()
06:59:31***Saving seen data "./dancer.seen"
07:21:23 Quit cYmen ("leaving")
07:40:35 Join ze__ [0] (
07:44:12 Quit midk (Read error: 104 (Connection reset by peer))
07:52:41 Quit ze (Read error: 110 (Connection timed out))
07:52:41 Nick ze__ is now known as ze (
08:35:36 Join LinusN [0] (
08:36:43 Join jyp [0] (
08:37:38 Quit Nibbler (Remote closed the connection)
08:38:18 Join rickst131 [0] (
08:39:48 Join Nibbler [0] (
08:39:49 Join NibbIer [0] (
08:39:58 Quit NibbIer (Read error: 104 (Connection reset by peer))
08:41:02rickst131is the boatloader.bin on the iriverboot page up to date?
08:41:53ashridahit doesn't support the hold switch properly, or the remote's record button, afaik
08:42:01rickst131oh ok.
08:42:09rickst131do i need to check out to get the newest?
08:42:37ashridahyes. it's also a good idea to compare it against someone else's to ensure you've got a correct build environment
08:43:12rickst131ah ok. thanks.
08:43:57LinusNi strongly advice you to only use the bootloader.bin in the wiki
08:44:51ashridahLinusN: is it up to date with hold switch support?
08:45:16 Quit Nibbler ("blubber")
08:45:30LinusNthe bootloader shares lots of code with the rest of the rockbox code
08:45:58LinusNso changes that seem to change stuff in the rockbox code will affect the bootloader too
08:46:10ashridahhrm. that's true
08:46:21ashridahstuff we've been playing with might break it
08:46:23LinusNand we dont verify the bootloader for every change in the rockbox code
08:46:36ashridahthe actual code that's doign the button checks is done fairly early tho, isn't it?
08:46:44LinusNnot early enough
08:47:34LinusNi will address this some day, but i don't feel that it's urgent
08:48:21rickst131rockbox is in c, correct?
08:49:55LinusNmostly, yes
08:50:10 Join Nibbler [0] (
08:50:52rickst131where can i get the latest rockbox code? i am familiar with programming(C++ and java), however i dont use cvs too much... i just know enough to check out. has a wiki that discusses what you need for a build environment, and what you need to check out to get the source and rockbox's tools
08:53:10 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
08:53:23 Join einhirn [0] (
08:54:31rickst131oh wow, im an idiot... i just saw the logs on that page and somehow missed the code. sorry for wasting you tiem. :-p
08:55:26ashridahthe simulators are good value for playing with rockbox's api itself, till you get the hang of it.
08:55:35ashridahif your intention is to start poking at the code
08:56:09rickst131it indeed is my intention. thanks for the help.
08:59:32***Saving seen data "./dancer.seen"
09:01:12 Quit rickst131 ("‹‹UPP››")
09:03:41 Join midk [0] (
09:19:34ashridahLinusN: reminds me, the xshock guy got DMA working instead of AUDIOTICK.
09:19:44LinusNsaw that
09:19:58LinusNbut i can't for the life of me understand why he had to use the SRAM
09:20:06ashridahah, yes. you would have read the logs
09:20:23LinusNthet spotted a bug in crt0.S, which i fixed just now
09:23:28 Join AC [0] (
09:26:22ACLinusN: Could i get cvs access for the upcoming wavpack codec?
09:31:48 Join bobTHC [0] (
10:03:22 Quit AC ("CGI:IRC")
10:09:31 Nick QT_ is now known as QT (as@area51.users.madwifi)
10:27:07 Nick StrathAFK is now known as Strath (
10:33:53 Join Aison [0] (
10:40:11 Join R3nTiL [0] (~zorroz@
10:45:23 Join Renko [0] (
10:45:37 Join R3nTiL_ [0] (~zorroz@
10:46:15 Join Lynx_ [0] (HydraIRC@
10:50:52 Join sox [0] (
10:51:02 Quit thegeek_ (Read error: 104 (Connection reset by peer))
10:51:32 Join midk_ [0] (
10:51:36 Quit midk_ (Client Quit)
10:52:52 Join R3nTiL1 [0] (~zorroz@
10:53:12soxashridah: yesterday amiconn made some OS-aware changes to the SIM Makefiles (to solve some cygwin probs i think), I think it would be possible to get the Mac OS X specific gcc options in there the same way?
10:54:50sox21.15.18 # <sox> amiconn: couldnt that OS specific configuration be useful to get the workaround for the mac os x pragma problem as well?
10:54:50sox21.16.51 # <amiconn> Certainly, the UNAME variable could be checked for this too. The question is how to hand the option to the preprocessor...
10:54:50sox21.17.14 # <sox> amiconn: eg, cpp -xassembler-with-cpp in the Makefiles
10:54:50DBUGEnqueued KICK sox
10:54:50sox21.17.29 # <sox> amiconn: instead of gcc
10:57:11 Quit R3nTiL (Read error: 60 (Operation timed out))
10:59:35***Saving seen data "./dancer.seen"
11:00:14LinusNsox: you use os x?
11:00:39LinusNand you have the cpp problem?
11:00:54 Join Marco [0] (
11:01:30soxyes, apples gcc adds a pragma line in the beginning of each makefile
11:02:06soxashridah found some good links describing the issue yesterday, check the logs
11:12:12LinusNsox: using $(CC) instead of gcc solved it, right?
11:12:32LinusN(except for the sim build, i guess)
11:13:17soxcpp -xassembler-with-cpp works
11:13:43sox(if that's what you meant)
11:14:40soxthe sim build has been impossible to compile on mac os x for more than a year i think
11:15:33 Quit R3nTiL1 ()
11:16:21 Quit R3nTiL_ (Read error: 110 (Connection timed out))
11:16:46LinusNsox: so, in the "SRC := ...", you substitute "gcc" with "cpp -xassembler-with-cpp"?
11:16:56soxLinusN: yes
11:17:15 Join DrRick [0] (
11:20:24 Quit DrRick (Client Quit)
11:24:03 Join DMJC [0] (
11:30:36DMJCis the cpu speed working yet?
11:31:08LinusNi haven't had time to work on it yet (have soldered about half of the test leads to my iriver)
11:33:15LinusNneed to measure the ata timing before committing any code
11:33:49LinusNin fact, i need to measure *all* timing :-)
11:37:17ashridahsox: i found them, but so did amiconn, he mentioned them in channel.
11:38:02soxashridah: ok
11:38:20ashridahor was that bagder. one of the two. possibly both.
11:38:36ashridahi think i was the first one to notice why apple had done it tho
11:38:40ashridahthe tools
11:38:45LinusNsox: if i send you a patch file, could you apply it and test?
11:38:49soxashridah: is it a fix we can use permanently?
11:38:54soxLinusN: sure
11:39:51ashridahsox: using 'cpp' directly should work permanently in theory. the command line exists for all versions of cpp i've tried. but the options to the preprocessor are actually technically unstable. relying on them isn't necessarily a good thing
11:45:30 Join Zagor [242] (
11:45:35LinusNhi Zagor
11:47:30amiconnSpeaking about unstability - the simulators now compile on cygwin without warnings, but they are unstable.
11:48:28 Join Patr3ck [0] (
11:50:39Zagoramiconn: unstable how? spurious crashes, or easily repeatable?
11:56:14amiconnmpa2wav.rock crashes the x11sim. Then the behaviour of all the codec test plugins is erratic both on x11 and win32 sim - see last night's log
11:56:40LinusNhow about win32 sim?
11:56:56amiconnLoading the cygwin x11sim via gdb produces a number of SIGSEGF right at the start
11:56:58LinusNignore me
11:58:35LinusNmaybe the plugins that do file i/o have problems?
11:59:37amiconnOkay, the SIGSEGV problem also happens with x11sims build with the old build system, so it might be a cygwin/ gdb issue. I read something about that on the net
12:01:27amiconnI'll try to dig into the build details this evening. Maybe I find what is wrong
12:08:15 Quit linuxstb ("Leaving")
12:21:34 Join methangas [0] (
12:27:01soxmissing header for unified diff at line 8 of patch
12:27:01soxcan't find file to patch at input line 8
12:27:01soxPerhaps you used the wrong -p or −−strip option?
12:27:01soxThe text leading up to this was:
12:27:03***Alert Mode level 1
12:27:03sox|Index: apps/Makefile
12:27:07sox|RCS file: /cvsroot/rockbox/apps/Makefile,v
12:27:09sox|retrieving revision 1.77
12:27:11sox|diff -u -r1.77 Makefile
12:27:13sox|−−- apps/Makefile 22 Feb 2005 12:19:11 -0000 1.77
12:27:15sox|+++ apps/Makefile 23 Feb 2005 10:47:27 -0000
12:27:19soxFile to patch:
12:27:21soxoops sorry
12:28:03 Join Schnueff [0] (
12:31:07ashridahwhat command did you use to try and patch it, and what directory were you in when you ran it?
12:32:18soxashridah: problem solved, it was -p0 i should've used
12:37:04***Alert Mode OFF
12:37:56CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
12:37:56*LinusN goes to lunch
12:38:46ashridahmmm. lunch
12:38:53ashridahpity it's 10:30 pm here :(
12:51:11 Quit jyp ("poof!")
12:59:37***Saving seen data "./dancer.seen"
13:14:31 Join Patr3ck_ [0] (
13:25:06 Join webguest07 [0] (
13:25:27 Join preglow [0] (
13:32:58 Quit Patr3ck (Read error: 110 (Connection timed out))
13:44:04 Quit lostlogic ("Going to the moon")
13:50:37 Join [IDC]Dragon [0] (
13:51:54[IDC]Dragonhi guys!
13:53:28[IDC]DragonI just ebayed 23 jukeboxes:
13:55:56preglowany particularily good reason for doing so? :P
13:56:31*[IDC]Dragon is the keeper of the Archos junkyard
13:56:41dwihnothe archos graveyard? ;)
13:56:59 Join cYmen [0] (
13:57:10pregloware they really so bulky they look?
13:58:07 Join markun [0] (
13:58:24Renkoanybody sourced a bdm interface in the uk?
13:58:50[IDC]Dragonu can build one
13:58:52Lynx_[IDC]Dragon: hmm, all without HD...
13:59:11[IDC]DragonLynx_: I know
13:59:36[IDC]Dragonbut less than 10 euro apiece is reasonable, methinks
13:59:38Lynx_[IDC]Dragon: just hope it's not the same thing that's broken on all of them ;)
13:59:52markunI was trying to do some profiling on the tremor code on de simulator. I thought adding -pg to gcc when compiling Tremor would do the trick, but is doesn't
13:59:53Renko[IDC]Dragon, for iriver and has anyone had success?
14:00:08markundoes anyone here use gprof before?
14:00:31[IDC]DragonRenko: the design is public, one PAL iirc
14:02:09markunpreglow, do you know how it works?
14:02:21amiconn[IDC]Dragon: I watched that aucktion too...
14:02:29ashridahmarkun: did you add it to LDOPTS as well?
14:02:43ashridahit needs to be specified when linking as well as compiling
14:02:48[IDC]Dragonamiconn: u need some parts?
14:03:16ashridahyou'll need to ditch -fomit-frame-pointer then too
14:03:31amiconnNot yet, but maybe in the future... And it might be interesting whether it's possible to get some working units out of this.
14:03:36ashridah(from the flac makefile, by the looks of it
14:03:44[IDC]Dragonmaybe I can refurbish some studios and sell them to a blind man's association
14:04:18markunashridah: From this page I guess I also need a "profiling startup file"
14:04:27markunBut I don't know what it is.
14:04:44amiconnIirc there are 3 recorder 30 units among them. Nice USB2.0 ...
14:04:46preglowmarkun: just specifying -pg when compiling AND linking should do unless you're on a strange platform
14:04:49amiconn*recorder 20
14:05:20ashridahmarkun: worked for me without anything else. i got the gmon.out file after running it
14:05:33[IDC]Dragona low USB2 ratio
14:06:18markunashridah: Did you find out which function is making Tremor so slow?
14:06:19amiconnIs the USB bridge located on the cpu board for recorders?
14:06:33[IDC]Dragonamiconn: yes
14:06:45ashridahmarkun: well, no, since i haven't actually profiled anything yet
14:06:58ashridahand that's assuming the profiling will even be valid under a simulator.
14:07:01amiconnSo maybe if the cpu board works, you could throw taht in another recorder...
14:07:52[IDC]Dragonyes, I did that with my g/f's recorder 15
14:09:09 Join elinenbe [0] (~elinenbe_@
14:11:47ashridahhm. what do i have to do to build the plugins to be usable with the simulator?
14:12:49 Quit elinenbe (Client Quit)
14:14:49ashridahrofl. okay. in the simulator, it runs at 17% of realtime :)
14:15:20markunashridah: yes, but here I also had 82% idle..
14:15:57markunmaybe the rockbox simulator is simulating to be on a slower machine?
14:16:40ashridahmarkun: which suggests it's doing a lot of sleeping
14:16:54ashridahthe simulator doesn't do any, afaik
14:19:14LinusN[IDC]Dragon: let's hope some of them are flashable
14:19:31[IDC]Dragonthe Studios?
14:19:36ashridahprobably need to modify the simulator to replace any sleepalikes with a busyloop
14:19:47LinusN[IDC]Dragon: yes
14:20:03[IDC]Dragonmaybe we'll have an old flashable Studio then
14:20:40[IDC]DragonI have to ask Uski if he's got chips left
14:21:17amiconnashridah: There *is* something wrong with tremor. I get ~15% here (x11 sim, cygwin, pentium m 1.5 GHz)
14:21:47markunamiconn: Do you also have a lot of cpu idle time when decoding using the simulator?
14:22:00amiconn[IDC]Dragon: I would be interested if there is any player with the old lcd. Then Linus doesn't need to send his.
14:22:58[IDC]Dragoncan this be seen from just looking at it?
14:23:33amiconnashridah: rockboxui.exe draws virtually no cpu power
14:24:24ashridahamiconn: even when it should be decoding a file as fast as possible?
14:24:37 Join R3nTiL [0] (~zorroz@
14:24:54preglowbear in mind that the fixed point codecs _are_ slower on pcs
14:24:58preglowbut shouldn't be that slow
14:25:07ashridahi didn't think it'd be deliberately holding itself back, and thus, the codec's only getting 20% on a 2GHz athlon here, suggests there's something weird going on, if rockbox spends most of it's time sleeping
14:25:54ashridahthe profiling isn't very helpful. strlen and strcmp take out equal places at 22%, followed by a few other functions. of course, some of which get called a *lot* and thus are explainable.
14:25:57preglowashridah: afaik, the xxx2wav plugins don't even yield(), so that's improbable
14:25:59markunpreglow I tried the testdecoder that comes with Tremor here on my PC. autobahn decoded to HD at about 80 x real-time.
14:26:20preglowmarkun: yes, sounds plausible
14:26:25amiconn[IDC]Dragon: The only method I know of (yet) is to check the ROM version...
14:26:34preglowi remember libmad used a couple of percent cpu on my athlon tbird 1.4ghz
14:28:34amiconnashridah: Something must be *very* wrong. The 05:31 minutes ogg I use for testing takes ~30 minutes to convert with vorbis2wav, but only ~3 seconds with 'The GodFather'
14:29:24ashridahamiconn: yeah, well, as i say, this thing appears to be sleeping a good deal of the time.
14:31:47LinusNvorbis2wav uses a very small pcm buffer
14:32:06LinusNit updates the lcd every 4k samples
14:32:16Renko[IDC]Dragon, is that the design at the bottom of the iriver bdm wiki page?
14:32:18LinusNthat will eat a lot of time in the x11 sim
14:33:41 Join elinenbe [0] (~elinenbe_@
14:36:38ashridahin comparison, the flac decoder decodes in about 350% of realtime
14:37:05ashridah(possibly more, my sample is tiny, so it doesn't settle, and it doesn't show it well, since the FINISHED label covers it
14:38:29 Quit R3nTiL ()
14:38:58ashridahLinusN: i increased it 10x and it's still going at about 20% realtime.
14:40:52ashridahand as i say, if it was actually doing stuff, i'd expect it to use more than 1-2% of my cpu :)
14:41:02 Join lolo-laptop [0] (~lostlogic@
14:41:32markunashridah: how much cpu does flac decoding use? The same?
14:41:35ashridahand pcmbuf's size 100x is again no difference
14:41:55ashridahmarkun: it's kinda hard to check. most of my stuff comes out a little large when encoded as flac
14:43:08ashridahokay. at about 180% of realtime for the file i've just got.
14:43:21ashridahusing 8% cpu
14:44:08ashridahthe X server isn't stealing time either, about 3% cpu.
14:44:21ashridahso, for some reason, the codecs are sleeping when used as viewers in the simulator.
14:45:26 Quit webguest07 ("CGI:IRC")
14:45:36ashridahideally, we'd want to override anything that looks remotely like a 'sleep'alike call, and replace it with a busyloop
14:45:53ashridahthis will make it show up in a profiler
14:46:33ashridahproblem is, there doesn't look like anything obvious inside tremor's code that's sleeping. (ie, grep -ri sleep finds nothing :) )
14:46:40markunashridah: I think why increasing the size of pcmbuf doesn't do the trick.
14:46:48preglowtremor should not sleep, heh
14:46:57ashridahpreglow: well. naturally.
14:47:08markun" ndicates actual number of bytes read. ov_read() will decode at most one vorbis packet per invocation, so the value returned will generally be less than length. "
14:47:09ashridahbut SOMETHING is.
14:47:25ashridahmarkun: yeah. you're probably right there.
14:47:28preglowwell, i don't know how the sims work, so no idea
14:47:51*ashridah wanders off to attend to other stuff for a few
14:48:28markunSo it's probably just decoding the same number of samples.. We could loop ov_read() and update a pointer inside the buffer until the buffer is full.
14:55:33 Join jyp [0] (
14:57:18jyppreglow: do you read me?
14:57:30preglowjyp: what?
14:57:55jypI'm at the point where I'd like to test the gmini's dsp
14:58:08preglowwhat tools do you have?
14:58:15jypBasically I'd like to run little test programs
14:58:40jypI have the gnu assembler
14:58:59preglowahh, didn't think it supported calmac
14:59:04jypI plan to implement an emulator too
14:59:22jypIt didn't until I implemented it ;)
14:59:40***Saving seen data "./dancer.seen"
15:00:05jypSo, I gather you have experience with dsps...
15:00:18ashridahmarkun: that's not a bad idea
15:00:51LinusNmpa2wav is running at 38% real time in 140MHz
15:01:01jyppreglow: You did a iir filter for the emac, I thougt maybe I could port it to calmmac
15:01:04preglowjyp: if you mean hardware, i actually have very little experience, but should be able to piece together stuff
15:01:10preglowjyp: ahh, sure
15:01:11dwihnoas in 1/3 of full speed?
15:01:43LinusNflac2wav is 77%
15:02:15soxhoy, im trying to compile the sim build on mac os x, could anyone take a look at the errors I get?
15:02:30soxi've upped the make report on
15:02:48dwihnoSo none of the codecs is running realtime yet?
15:03:18ashridahmarkun: you're a fricking genius
15:03:43jyppreglow: Is this current ?
15:03:49ashridahrofl. actually, i better double-check this.
15:03:53ashridahit could have a nasty screwup :)
15:03:56preglowLinusN: and btw, did your preliminary iram patch actually move the flac routine to iram?
15:04:01preglowjyp: gimme sec
15:04:19ashridahhrm. no. seems okay
15:04:48LinusNpreglow: yes
15:05:17LinusNit is done on a function or object level
15:05:39LinusNthe only thing moved to iram is your emac optimized function
15:05:42preglowjyp: reload
15:05:44 Quit DMJC ("Leaving")
15:05:55preglowLinusN: didn't measure much difference :/
15:06:23preglowbut of course, there will be a gain, no question about that
15:06:36preglowprobably just need to move in some data as well
15:06:45LinusNthe sdram is dead slow, so we need to take advantage of the iram
15:07:02elinenbedead slow! that's no good!
15:07:04markunTremor is decoding at 3000%!
15:07:07LinusNwe will even gain speed if we run the code directly from flash
15:07:31preglowLinusN: ahahah, not bad
15:07:34markunWell, on the simulator at least.
15:07:49preglowjyp: you need to be aware of how the fractional emac mode works, btw, the mac instruction does a full 32x32 bit = 64 bit multiply, but just keeps the upper 40 bits
15:08:09markunashridah: it does the trick :)
15:08:36jyppreglow: ok
15:08:40soxLinusN ja
15:08:48ashridahthat's a little more like it
15:08:49preglowjyp: and GET_ACC(_CLR) instruction automatically left shifts the result once
15:08:53ashridahgetting 250% realtime
15:09:12ashridahusing multiple calls to ov_read with offsets into a larger buffer
15:09:16elinenbemarkun: 3000% on the simulator... ah!
15:09:35preglowjyp: could you paste me the calmac manual again?
15:09:38elinenbeashridah: 250% realtime for what codec?
15:09:39ashridahmarkun: rofl. how many calls to ov_read did you make?
15:10:02ashridahelinenbe: tremor, using my dodgy hack, which (as expected) got a 10x speed increase for 10x more calls to ov_read.
15:10:07markunashridah: I increased the buffer by 100 and looped until the buffer was full.
15:10:24jypit's in the datasheets wiki page now
15:10:28ashridahi'm just not sure if the calculated speed isn't skewed by this
15:11:06markunn = 0; while (n < BUFFSIZE) n += ov_read(&vf,pcmbuf + n,BUFFSIZE - n,&current_section);
15:11:50ashridahheh. that's neater than mine :)
15:12:21markunashridah: you also will have to define BUFFSIZE and use it to init pcmbuf of course..
15:12:26ashridahyeah. the only thing to check is that display_status is using the amount of data processed, and not assuming it's operating at single-frame granularity
15:12:40LinusNalso use my latest change to xxx2wav.c
15:13:16markunLinusN: What did you change?
15:16:02LinusNit only updates the lcd once per second instead of once per function call
15:16:03ashridahmarkun: now comes the fun bit. test your change on the iriver :)
15:16:16ashridahor his
15:17:12ashridahor both
15:20:19dwihnoAnyone got any good ideas how to strengthen an audio signal? My new computer sound board is too quiet
15:22:09preglowwith an amplifier? :>
15:22:10markunashridah: It doesn't make a difference on the iriver so far :(
15:23:04preglowi get 0.80% more perforamnce
15:23:26*[IDC]Dragon just noticed that today is Rockbox release day!
15:23:27markunThen I did something wrong probably.
15:23:32preglowor something like that
15:24:08[IDC]DragonZagor? ;-)
15:24:26Renkoi just managed to reset my firmware ?!
15:25:17Renkoi didn't think that was possbible???
15:26:32Renkoand now it's screwed again
15:32:56dwihnopreglow: are there really small amps available? such as 3.5mm in, 3.5mm out? :)
15:33:01dwihnoI'd love a tiny tiny tiny amp
15:33:13Zagor[IDC]Dragon: hehe, i actually thought about it yesterday.
15:33:48[IDC]Dragonlet's wrap up this mess!
15:33:49preglowdwihno: would be pretty easy to build one
15:33:49 Quit bobTHC ("( :: NoNameScript 3.81 :: )")
15:33:53 Part LinusN
15:33:55markunpreglow: What speed did you get
15:34:13preglowmarkun: 'bout 7.3-7.5-ish
15:34:59markunDecoding vorbis on iriver? I get about 0.87
15:35:17dwihnopreglow: I don't want to build, I want to buy and plug it in rightaway :)
15:35:36preglowthe gains might easily be swallowed with vorbis
15:36:18ashridahdwihno: you can buy premade, 9V cell-powered mini amps that fit inside a tin the size of a penguin mints tin :)
15:37:49Renkodwihno, sure there are no gain settings or other vol controls?
15:38:17ashridahdwihno: there's actually a guy who preassembles those. the tins are tiny.
15:38:59 Quit markun ()
15:43:53soxwould these small amps work to amp an external mic as well?
15:44:26ashridahuh. i'm not sure. i doubt it. microphone output tends to need a lot of gain
15:45:26ashridahyou'd probably be able to modify the gain of a chu-moy style amp to get the extra gain you need tho
15:45:44preglowmics typically need a ton of gain, yes
15:45:49preglowdepends on the type
15:45:59soxi was surprised to see that my iriver was lots better than my ajbr at getting a higher signal from a mic
15:46:31dwihnoRenko: just volume
15:46:41dwihnoI intend to use it with my computer
15:47:04soxcant wait to get rockbox working so that i can record with more ease, irivers rec interface is not nice...
15:48:56dwihnoashridah: looks interesting
15:49:58ashridahthe article's probably old, not sure if the guy dan bought the prebuilt amp from still does it, but plenty of other people will.
15:50:23ashridahand since it runs off a 9V amp, it won't be clipping like a 5V line-out would ;)
15:51:20soxvery stylish, anyways
15:51:49ashridahanyway, it's 2am. time to crash
15:51:53 Quit ashridah ("sleep")
16:00:30 Join markun [0] (
16:08:47preglow4pm, time for a walk accompanied by h120 in the glorious sunshine
16:44:16 Join R3nTiL [0] (~zorroz@
16:54:34 Join mecraw [0] (~mecraw@
16:59:42***Saving seen data "./dancer.seen"
17:04:51 Join bobTHC [0] (
17:06:31 Quit R3nTiL ()
17:06:37 Join Chamois [0] (
17:07:03ChamoisDLinus : did you work on the bootloader and the very short press on ON button ?
17:07:38preglowDLinus? :P
17:07:40Chamoishe is not here
17:08:00preglowno, he left, but what is this very short buttonpress you speak of?
17:08:42Chamoiswhen you press the ON button on the iriver very shortly with the HOLD button ON the iriver start instead stays off
17:09:53soxi have another question on the same topic, i thought the basic function of the remote was working now? I dont see that on my unit, do I have to reflash?
17:10:09preglowno, spi is not supported yet, afaik
17:10:17soxah ok
17:10:32preglowit supports on button, but i guess that's got it's own wire
17:11:06soxbut i think i read what chamois is talking about in some old irc log, that the remote could do the same thing as the main unit, start irivers fw with ON+rec?
17:11:28soxor is it no way to start irivers original firmw with the remote?
17:14:38amiconnpreglow: Iirc the remote buttons are supported by the driver, only that no part of rockbox uses the remote button codes
17:15:00amiconn(except ON)
17:15:27preglowyes, i see the backlight comes on when i push stuff
17:16:01soxso it's a no to my last question then?
17:16:03preglowbut the on button probably is on a wire of its own anyway, so it can be started from the remote
17:16:13preglowunless the buttons don't use spi and i've misunderstood
17:16:18amiconnYes, the backlight coming on is directly handled in the driver
17:17:01amiconnNo, the buttons don't use spi. They are hooked up to the adc the same way as the main buttons, only using a different channel
17:17:31amiconnThe remote lcd uses spi iirc
17:18:15soxis this bootloader code?
17:19:18amiconnThe button driver code is shared between main rockbox & bootloader afaik
17:20:20preglowmost code is
17:20:33preglowbootloader is just a specialized version of rockbox
17:20:39amiconn(unlike the archos bootloader, btw)
17:20:53preglowspace constraints?
17:21:35soxsomeone made a bootloader version with inversed start behaviour i remember, with iriver as default and rockbox with rec pressed
17:22:02amiconnpreglow: It is both saving space and that a full-fledged button driver isn't needed.
17:23:20soxanyone remember who?
17:23:44soxthis i would like until rockbox is more functional in everyday life
17:29:54*sox is sad because noone remembers
17:30:10preglowit's pretty trivial, sox
17:31:15soxwell then, share to the world
17:32:47preglowi'm not going to do it for you
17:33:02preglowjust reverse the button check and move some code around
17:33:09preglowi prefer mine the way it is
17:36:08soxpreglow: allright, i understand totally, and will not put my programming qualities to test in such a sensitive matter ;−−)
17:41:23preglowsox: nor would i, i intend to keep my fingers away from the bootloader more or less completely
17:41:33preglowi'm able yes, will i screw it up anyway, yes
17:46:10Renkomy bootloader gives me a blank scree
17:46:51preglowhow nice
17:46:53preglowhow'd that happen
17:47:12Renkodunno, complied from cvs and prob messed up
17:47:46Renkobut... if I reconnect the battery with my finger on play it goes back to the iriver firmware...
17:47:52preglowok, so your player is trying it's own at being a paper weight?
17:48:18preglowwhat about the reset switch?
17:48:27Renkoit's broken :S
17:48:35preglowwell, ok, then its not beyond hope
17:48:39Renkoused too many times ;)
17:48:41preglowjust flash bootloader.bin in the wiki
17:49:08Renkoyeh I have done previously and it worked
17:54:52preglowmy box is getting quite a few hits on the gameboy pictures
17:54:56preglowrockboy, i mean
17:55:26preglowabout ten thousand
17:57:05 Quit einhirn (Read error: 104 (Connection reset by peer))
18:04:52 Quit Patr3ck_ ("User pushed the X - because it's Xtra, baby")
18:09:44 Join ripnetuk [0] (
18:10:33pregloware we planning to use the eeprom, btw?
18:10:59ripnetukthe person who wanted to run iRiver firmware by default instead of rockbox - i can halp in a limited (but safe) way - i have compiled a version of rockbox.bin that just jumps straight to the original rom (on my web site at - if you put that on the hard drive, it will always boot to that, which in turn boots the original. That way you can leave the iRiver flashed, but not have to fuss about to load the origin
18:11:48ripnetukof course you cannot then run Rockbox without plugging into a computer and removing my rockbox.bin and putting the real one back
18:11:58preglowdidn't think of that
18:11:59ripnetukgotta go... cya
18:12:07 Quit ripnetuk (Client Quit)
18:19:51 Join hubble [0] (
18:25:07 Quit bobTHC ("( :: NoNameScript 3.81 :: )")
18:28:46 Join Zezayer [0] (
18:30:41 Nick Zezayer is now known as zezayer|away (
18:36:58 Nick Lynx_ is now known as Lynx_awy (HydraIRC@
18:50:39 Quit jyp ("poof!")
18:51:19 Nick preglow is now known as pregone (
18:56:24 Quit Chamois ("CGI:IRC")
18:58:11 Join XShocK [0] (
18:58:17 Join condor9 [0] (
18:58:52condor9amiconn: busy?
18:59:45***Saving seen data "./dancer.seen"
19:05:15 Quit XShocK (" Try HydraIRC -> <-")
19:22:36condor9amiconn: I have my config and font file ready to send to you when you're ready.
19:34:35 Join Ski [0] (
19:34:43Skianyone here?
19:35:24HClwe all died :/
19:35:40HCldue to square-eye disease caused by computers.
19:35:59amiconncondor9: Now I'm here
19:36:49condor9okie. I just tarballed my .rockbox with a default config and a bold font.
19:37:22SkiI guess noone can help me then
19:38:41 Quit Ski (Client Quit)
19:39:35zezayer|awayWhilst its quiet, does anyone have a version of rockbox in the .hex form, ive been tryin 2 build if from the instructions online and it just aint working [ sorry if ur sick of ppl askin this ]
19:39:57 Nick zezayer|away is now known as zezayer (
19:41:06 Quit [IDC]Dragon ("CGI:IRC")
19:43:14amiconncondor9: Now that is strange. I just put your config.cfg in /.rockbox, bold.fnt in /.rockbox/fonts and selected config.cfg from 'browse .cfg files'
19:43:40amiconnThe font loads as intended, and my recorder keeps the setting across a reboot...
19:43:53 Quit Diway ()
19:43:57condor9Strange. Any ideas?
19:44:40amiconnThe only difference between your and my .cfg files I can see is that yours uses DOS line endings (cr+lf), while mine use unix style (lf only), but that shouldn't matter...
19:45:17condor9Yeah, because all the other settings are restored across a reboot.
19:45:24condor9My config was generated from within rockbox.
19:46:17amiconnNo, because the settings are actually loaded from the .cfg. The .cfg is not involved in the reload at boot
19:46:49condor9Where is the reload at boot info stored?
19:46:54amiconnThe persistent settings are stored in a special sector of the hd, before the file system.
19:47:04condor9Maybe it's corrupted?
19:47:33amiconnSome settings are also stored in the rtc (only recorders have one)
19:48:10condor9Strange that it works for 2.2
19:48:11amiconnWhen this sector gets corrupted, the settings are reset to their defaults, so no setting will survive...
19:48:39condor9Can I re-initialize that sector?
19:49:25amiconnThis sector is rewritten completely when saving settings, so no need to initialise
19:50:35amiconnPlease go to Info->Debug->View partitions and tell me what it says
19:51:02condor9Ok. Just a sec.
19:51:08amiconnThere are 4 partition entries, P0..P3
19:51:24amiconnIf you only have one partition, P1..P3 should read:
19:51:28 Quit markun ()
19:51:39amiconnPn: S:0 T:0 0 MB
19:51:46condor9I have two partitions. Here's the info.
19:52:01amiconnTell me if they don't, and also tell me the P0 and P1 parameters
19:52:03 Join Sucka [0] (
19:52:09condor9P0: S:8 T:B 12288 MB
19:52:39condor9P1: S:1800008 T:5 25866 MB
19:52:49condor9P2: S:0 T:0 0 MB
19:53:00condor9P3: S:0 T:0 0 MB
19:54:13amiconnHmm, that partition layout is a bit unusual, in that the number of sectors per track is not 63.
19:54:50amiconnBtw, that means rockbox 2.2 (which puts the config sector unconditionally at sector 61) overwrites a sector within your first partition!
19:55:39condor9Ok. So, how should I restructure my drive?
19:55:42amiconnCurrent rockbox calculates the position of the config sector as (lowest_partition_start - 2), so it's 6 in your case
19:56:14amiconnIf that doesn't work with your layout, maybe there's a bug in rockbox...
19:56:50amiconn(Btw, I just checked, .cfg files created by rockbox use dos linefeed style)
19:57:08amiconnDo you use windows or linux?
19:57:15condor9I use OS X. :)
19:57:41amiconnAh. So you should have a unix command line as well (?)
19:57:47 Join mrelwood [0] (
19:58:12condor9Yes, I've been a unix snob for quite some time.
19:58:12 Quit mrelwood (Client Quit)
19:58:46 Join mrelwood [0] (
19:58:55amiconnCould you dump the beginning of your disk (physical disk, not partition) to a file with dd?
19:59:08condor9Sure. Gimme a minute.
19:59:12amiconnFirst couple of sectors is sufficient
19:59:37mrelwoodI see the iRiver is very tightly being developed... COOL!
20:04:37condor9Sorry, I forget the sector/block size info.
20:04:51condor9Should this be enough ... dd if=/dev/disk1 of=dd.out count=20
20:05:22amiconncount=63 would be better
20:05:30condor9Okie. Just a sec.
20:05:44condor9Oh yeah. :)
20:08:51 Quit sox ("Snak 4.13 IRC For Mac -")
20:10:24mrelwoodanyone tried to record with the iriver yet?
20:13:10amiconncondor9: I think I know the problem... and it's not a rockbox bug, but rather a rockbox design problem (the 'hidden' config sector) in conjunction with your OS... :(
20:13:57amiconnRockbox 2.2 indeed saves at sector 61, which is happily inside of what looks like a FAT of your first partition
20:14:44amiconnHowever, I guess rockbox 2.3+ tries to save correctly, but your system doesn't leave the sectors before the first partition alone
20:15:23condor9You mean it changes them after rockbox writes to them?
20:15:31amiconnUsually, sector 0 is the master boot record and holds the partition table. The first partition starts one track later (usually at sector 63, in your case at sector 8)
20:16:10amiconnFor your setup, rockbox 2.3+ tries to save at sector 6, but sector 6 gets overwritten by a sector looking identical to sector 0
20:16:36amiconnI guess you OS stores copies of the MBR in sector 1...7, thereby overwriting the rockbox config
20:16:56amiconnThe fact that a part of your settings stay is also explainable
20:17:24condor9Wow ... thanks for the diagnosis.
20:17:27amiconnSince you have a recorder, the settings stored in the RTC stay intact
20:17:53amiconnOne more reason to change the rockbox config save to use a real file...
20:18:37condor9I partitioned my drive using Disk Utility ... perhaps I should do it differently.
20:19:08amiconnBtw, the problem that happened to the font setting will also occur if you try to set a different language or wps
20:19:26condor9Yeah, I noticed the wps didn't restore either.
20:19:45amiconnThese are settings that are stored in the .cfg sector only...
20:19:45condor9Can I pad the first 62 sectors?
20:20:03condor9or 60 or whatever?
20:20:22amiconnHmm. In order to get the defaukt disk layout, you would need to repartition the drive to use 63 sectors/track
20:20:45condor9Would that fix my problem?
20:21:01amiconnHowever, I doubt that it'll help. Maybe the OS extends its copying of sector 0 to sector 1..62 then.
20:21:28amiconnYou could at least try it, it it's not too much work
20:21:42amiconn(I wouldn't recommend doing this via USB1.1)
20:22:11condor9It's not too much trouble, but I can't do it right now.
20:22:39condor9Thanks so much for your help.
20:23:07amiconnOkay. When you will get to doing it, and it actually fixes the problem, just drop a small note.
20:24:05condor9I'll let you know either way, but I'm determined to get it fixed. Thanks again.
20:32:03 Quit mrelwood ("CGI:IRC")
20:35:54 Quit condor9 ("Leaving")
20:36:00 Join Stryke` [0] (
20:54:25lolo-laptopzomg wavs work!
20:54:30lolo-laptopH3x0 time
20:54:37HClthey do?
20:54:41lolo-laptopthe wiki says so
20:55:40CoCoLUSdynarec status? ^^
20:56:18HClstalled, cause i'm having trouble calling a block
20:56:42HCland my sister got married today... so i'm tired and stuff. i'll prolly look at it tomorrow..
20:57:24CoCoLUSbest wishes :)
20:57:51HClshe's still very ill. we all hope she'll survive her current disease without too many side effects.
20:59:48***Saving seen data "./dancer.seen"
21:30:08 Quit hubble ()
21:33:02 Quit Zagor ("Client exiting")
21:39:34 Join Digital007 [0] (
21:39:43Digital007Can iriverbox now play wav files?
21:40:20midkcan it even play mp3's?
21:40:40Digital007well according to the wiki it looks like wav playback is working
21:40:49rasherwav is pretty much working I think, but not merged yet
21:41:22Digital007well i just downloaded the latest version, but couldn't get a wav file to play
21:44:17rasherIt's not in cvs yet
21:45:40 Join webguest30 [0] (
21:45:52Digital007ok fair enough
21:46:57rasherThere's code linked on the iriverport status page I think
22:00:41 Quit webguest30 ("CGI:IRC (Ping timeout)")
22:05:27 Join Tang [0] (~chatzilla@
22:05:53TangWant to congrat fo latests progress once again
22:06:12TangYou're all making wonderful job :)
22:06:41TangI'm waiting for my job pay to donate again :)
22:06:45TangCheers :)
22:09:49 Quit Tang (Client Quit)
22:18:59 Join jyp [0] (
22:22:12 Quit elinenbe (" HydraIRC -> <- The dawn of a new age")
22:22:50 Join webguest30 [0] (
22:25:10 Quit webguest30 (Client Quit)
22:29:41 Quit courtc (Read error: 110 (Connection timed out))
22:30:12 Join courtc [0] (
22:33:28 Join zedtwo [0] (
22:33:29 Nick zedtwo is now known as gfdgfsdgfdsgfdsg (
22:33:42 Quit methangas (Read error: 60 (Operation timed out))
22:33:46 Nick gfdgfsdgfdsgfdsg is now known as borock (
22:34:14borockhey there
22:36:02borockhowu doin
22:38:49 Nick pregone is now known as preglow (
22:42:36 Join Patr3ck [0] (
22:43:49 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
22:43:55 Quit Schnueff ("away")
22:51:44 Join einhirn [0] (
22:59:50***Saving seen data "./dancer.seen"
23:11:22 Join Quelsaruk [0] (~kvirc@
23:30:04 Quit borock ("—I-n-v-i-s-i-o-n— 2.0 Build 3515")
23:31:16 Quit zezayer ("ChatZilla 0.9.61 [Mozilla rv:1.7.2/20040804]")
23:39:14 Join hubble [0] (
23:39:54 Quit jyp (Remote closed the connection)
23:43:43HClnot yet.
23:43:50HClin 20 minutes it'll technically be morning.
23:45:11CoCoLUSnice timezone :)
23:45:28CoCoLUSi know, i'll live here :P
23:46:08CoCoLUShm its getting late, can't even get the tenses right
23:47:04preglowit's getting early in fifteen minutes
23:47:54CoCoLUSi don't consider 000 early
23:47:58CoCoLUSearly is 4... 5 :)
23:57:55 Quit preglow ("reboot")

Previous day | Next day