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 2004-09-28

00:04:52amiconnFound my mistake with the mmc...
00:06:59 Quit AciD (Read error: 232 (Connection reset by peer))
00:10:07PaulSelinebe: Thanks!
00:10:51bagawkis it just me are does the player w32 simulator not build?
00:11:02bagawki have tons of errors on tree.c
00:11:27amiconnmaybe you didn't recpompile convbdf?
00:11:36bagawkno, convbdf is new
00:11:43bagawkhold on, i try x11 version
00:11:58elinenbePaulS: you didn't answer my question yet... have you gotten anything of substance to run yet?
00:12:07PaulSelinenbe: As I said in the wiki, I used something very similar to patch the ID3 tag processing code so that it called a function that spit the tags out the UART port. I'm fairly confident about the hex file output of the script in the Wiki, but I'm too chicken to try it out without a backout strategy. B)
00:12:22bagawkthis source is a little modified, but i have done nothing to tree.c
00:12:32bagawkill grab new source off cvs brb
00:13:21amiconnbagawk: wfm
00:14:17bagawkamiconn: wfm?
00:14:22amiconnworks for me
00:19:03bagawkhumm works this time...
00:19:30bagawki must have made a bad change in a file that broke the player
00:23:15 Join gromit`` [0] (
00:24:50bagawkarghh it is hard to find a bug when you have lots of #ifdef's
00:25:33 Quit gromit` (Read error: 110 (Connection timed out))
00:47:05 Join MisticJeff [0] (
00:53:14 Join man|e [0] (
00:54:13man|ecan someone help me, plz ? :/
00:54:56***Saving seen data "./dancer.seen"
00:55:26man|eI have a problem with my av320 joystick :(
00:57:36PaulSSounds like a hardware problem.. It's for hardware that RockBox hasn't even started to write code for. Are there specifics for the problem?
00:58:23man|ewell, the joystick doesn't respond to right and down movements, only up and left
00:58:42man|eI assume its hardware problem, but I don't know where to fix it :(
00:58:43bagawkman|e: oyu can ask in another channle
00:58:49bagawkeither #linav, or #avos
00:59:04bagawki would not know about how to fix it
00:59:04man|eno one is in #linav nor #avos right now :(
01:15:00bagawktime to go bye
01:15:15bagawkZagor: or adjö to you :)
01:15:22amiconnThe MMC manufacturers do the same cheating as the hd manufacturers :(
01:16:05 Quit bagawk ("umount /dev/brain")
01:16:19amiconnMy 256 MB card has exactly 256835584 bytes, which equals a little less than 245 MB
01:16:53amiconnBtw: The values is read from the card by my code :) :) :)
01:45:54 Quit Zagor ("Client exiting")
01:46:30 Part man|e ("Leaving")
02:04:26 Quit MisticJeff (Read error: 238 (Connection timed out))
02:04:31 Quit mecraw ("Trillian (")
02:32:55 Join plok [0] (
02:37:59*plok is away - Automatically set away. - messages will be saved.
02:39:28 Part PaulS ("Off I go")
02:45:53amiconnFirst real MMC file system access on Ondio!">
02:54:59***Saving seen data "./dancer.seen"
02:56:49 Part amiconn
03:12:33 Quit plok ("I'm outta here!")
04:49:23 Join Byron [0] (
04:53:49Byroncan someone tell me what CVS is, I wanted a talking clock on my archos but I can't find it, I installed a Daily Build but I don't see a clock in the info menu
04:55:00***Saving seen data "./dancer.seen"
05:15:22 Join darryl-upstairs [0] (
05:15:44darryl-upstairsDoes anyone have instructions on upgrading the hard drive on the FM Recorder
05:16:11 Nick darryl-upstairs is now known as Byron-2 (
05:16:21Byron-2oops, stupid ghost
05:16:41Byron-2friend used my computer and it still had his nick on it
05:21:52 Join ashridah [0] (
05:35:35 Quit Byron (Read error: 110 (Connection timed out))
06:49:46 Join LinusN [0] (
06:55:02***Saving seen data "./dancer.seen"
07:10:28 Quit Byron-2 (Read error: 238 (Connection timed out))
07:15:43 Join PaulS [0] (
07:16:23PaulSWell, I got too cocky, and now I need a wiggler to dig myself out of a jam I got myself in...
07:20:22LinusNsaw tat :-)
07:20:38LinusNand i haven't even got the wiggler to work yet ...
07:21:47PaulSThem's the breaks. There aren't any bad short-term side effects. The sad part is that the code I have now isn't the one with my ID3 hack, so I can't continue with the other stuff I was thinking of messing with.
07:22:39LinusNanother sad thing is that the bdm connection process is rather involved
07:22:48PaulSI'm considering seeing if the chip works in JTAG boundary scan mode. If it does, I can bit-bang a flash programming routine.
07:23:48PaulS(I did this sort of thing with the StrongARM a couple years ago.)
07:24:02LinusNso you haven't seen any evidence in the original firmware of a boot monitor of any kind
07:24:29LinusNlike waiting for a character on the serial port while booting
07:24:46PaulSThe chances of that are pretty dismal.
07:25:35LinusNso your firmware can't load ram images either?
07:26:37PaulSNot as far as I can tell. I'm hard pressed to say why. I got hung out to dry when I put some debug "putc" calls in my wrapper routine.
07:27:40PaulSI know that there's code that makes sure the file size is > 0x200, but padding the RAM image didn't change its behavior significantly.
07:28:16PaulS(But that could have been after I messed up the wrapper function with the putc calls...)
07:28:45LinusNreally annoying
07:29:10PaulSIt's possible that the original image would have worked if I'd padded the RAM image, now that I think of it. The putc themselves could be killing me for some reason.\
07:31:16PaulSWhat sort of troubles are you having with the BDM drivers?
07:31:42LinusNthere is a kernel driver and a userspace library to communicate with it
07:32:20LinusNand the user space lib is not behaving correctly, it doesn't talk to the driver at all, it thinks it connects but it doesn't
07:32:41 Quit midk (Read error: 104 (Connection reset by peer))
07:32:54 Join midk_ [0] (
07:35:25uskiLinusN: u there ? (i don't think so :D)
07:35:38PaulSHmm.. I'll hold off ordering the wiggler until you get a little further. It might be better for me to work on the boundary scan in parallel.
07:36:46LinusNuski: i'm here
07:37:21LinusNPaulS: indeed, that might be a better way of salvaging a bad flash, since there are fewer wires to solder
07:39:22PaulSLooks like all the JTAG pins are there, as long as you can match the silkscreen up with the right pads.
07:40:13LinusNit's not trivial, but i know which one is which
07:45:58PaulSHmm.. Any chance you could mark-up a JPEG with nice pointy arrows? B)
07:46:08LinusNi could do that
07:46:20LinusNbut not until tonight :-(
07:47:28PaulSI think that if I stare at it hard enough I can argue that there's only one set of assignments that make sense.
07:47:54PaulSI prolly won't be opening my iRiver until tomorrow, if then anyway.
07:50:24LinusNi can tell that the staring method won't work
07:50:34LinusNbeen there, tried that :-)
07:51:08LinusNwhen i later checked with the multimeter on the naked iriver pcb, i had to move quite a few wires
07:54:07PaulSHm. Well, just bear in mind that inquiring minds want to know the correct pin assignments. When you can, of course. I'm not sitting here with a non-functional unit.
07:54:19PaulSMoreover my own hubris got me here. :-)
07:56:06LinusNdon't worry, i'm on the case
07:58:51 Join [IDC]Dragon [0] (
07:59:53[IDC]Dragonlogpeeker asks: what did PaulS do?
08:00:44LinusNpatched the original firmware to run images directly from disk
08:01:11[IDC]Dragonoh, wow
08:01:12LinusNs/patched/tried to patch/ :-)
08:01:29[IDC]Dragonoh, and now it's dead?
08:01:45LinusNno, it just can't reflash anymore :-(
08:01:58[IDC]Dragonalmost the same
08:02:28[IDC]Dragonis there a check in the firmware?
08:02:54 Join amiconn [0] (
08:03:06[IDC]Dragonhey, Jens is the man!!!
08:03:34PaulSAh.. Here, a first clue: TRST == DSTCLK. That's not a typo −− those pins are the same thing.
08:03:39amiconn[IDC]Dragon: Seen my lo entry/ photo link ;-)
08:03:49[IDC]Dragonyes, terrific!
08:03:57[IDC]Dragoncommit, commit
08:04:32amiconn[IDC]Dragon: (1) It is very preliminary: rudimentary error checking, and currently read-only
08:05:00[IDC]Dragonbetter than cvs, still
08:05:01amiconn(2) You have to disable _a whole lot_ of inits to prevent rockbox from hanging
08:05:13PaulS[IDC]Dragon: To be clear, my iRiver is completely functional except for the ability to reflash.
08:05:32[IDC]DragonPaulS: strange
08:05:52PaulSNo, not too strange, considering the bits of code I was toying with.
08:06:27amiconnLinusN: We do indeed have a problem: The MMCs (at least the internal one) don't use the (logical) 63 sectors/tracks, but only 32. So if I would have enabled writing, I would have trashed the fat...
08:06:41[IDC]Dragonvery couragous, to flash some firmware you don't know if it could work
08:06:44amiconn...with the config sector
08:07:35[IDC]Dragonamiconn: per track? Don't we do LBA?
08:08:43amiconn[IDC]Dragon: Yes we do LBA, but there is still that virtual track blocking. Partitions can only start/ end at multiples of sectors_per_track
08:09:09amiconnThe card uses 32 sectors/track, so the first partition starts at sector 32...
08:09:39amiconnLinusN: Am I correct that the only used sector on track 0 is sector 1 (MBR)?
08:09:47[IDC]Dragonand the config sector is where?
08:09:54PaulS[IDC]Dragon: "courageous" −− some would use less flattering words. :-)
08:10:14[IDC]Dragonwe can easily move it somewhere else
08:10:31amiconn[IDC]Dragon: config sector is 61 (HDs usually use 63 sectors/track)
08:11:03amiconnLinusN: In that case we should make the config sector dynamic, calculating first_partition_start -2
08:11:20[IDC]DragonPaulS: when I did the Archos flash thing, I've spent most of the effort into the UART boot
08:11:26LinusNamiconn: it depends on how you have partitioned
08:11:47LinusNif you have no logical partitions, only the MBR is used
08:12:21amiconnAnd logical partitions use one additional sector per partition?
08:12:30LinusNdon't remember
08:12:40amiconnHmm, ok
08:12:58LinusNbut we can find out the best location for the config sector
08:13:06LinusNby reading the partition tables
08:13:26amiconnLinusN: <amiconn> LinusN: In that case we should make the config sector dynamic, calculating first_partition_start -2
08:16:38[IDC]Dragonamiconn: did your Ondio play?
08:16:50[IDC]Dragonor is the MAS too diffderent?
08:17:14amiconnNo, and this can't work. As I said: [08:05:02] <amiconn> (2) You have to disable _a whole lot_ of inits to prevent rockbox from hanging
08:18:25[IDC]Dragonok, I didn't know _that_ lot
08:18:26LinusNamiconn: yes
08:19:24amiconn[IDC]Dragon: I didn't figure out exactly which inits do hang (at 3 a.m.), but I disabled settings_apply() through playlist_init() and mp3_init() through talk_init() in main.c to take the photo.
08:20:21amiconnsettings_apply() definitely does hang
08:23:29[IDC]Dragonmore than fair, at 3 am :-/
08:25:12[IDC]Dragonoh, thanks
08:25:50amiconnYou'd have to do the adaptions of main.c yourself for now. Beware: if you do this, the menu doesn't work anymore. Dunno why yet...
08:26:33[IDC]Dragonvery rough and strange, indeed
08:27:37amiconnMaybe there's a bug in the new button handling on the Ondio (menu key in browser not catched?)
08:30:17amiconnBtw: We can't do an early return from writing. After transmitting the (last) data block to the card, we have to wait until the block got written. Then there is a "data response" from the card, indicating whether the block got written correctly.
08:30:51amiconnThe timeouts are _very_ long with some cards
08:31:57[IDC]Dragonwhich timeouts?
08:32:00amiconnPerhaps this is the reason why Antonius' 1 GB card doesn't work for playback - the archos fw uses a fixed timeout, which is too short for the worst case even with my 256 MB card
08:32:54amiconn(1) Read timeout: After sending the read command, you'll have to wait for the data block to arrive
08:33:33amiconn(2) Write timeout: After sending the write command and the data block, you'll have to wait until the data got written, and the card tells you the status
08:34:10[IDC]Dragonah, ok, so this is not done "on the fly"
08:34:38[IDC]DragonI naively thought this transmission is slow enough
08:35:38amiconnRead timeout with the internal flash is 37,500*8 clocks, and write timeout is 150,000*8 clocks (worst case).
08:36:17[IDC]Dragonis it written on the card?
08:36:27[IDC]Dragonor just the datasheet?
08:36:38amiconnYes, you have to calculate this from the CSD register data
08:37:24amiconnThe archos fw does not do this, but uses a fixed timeout of 150,000*8 clocks. Some cards may want more...
08:37:33amiconnUsing the multi-block read and write commands is really useful with mmc
08:39:42amiconnThere is also a operating voltage range coded into the OCR register (for accessing the flash core, accessing the logic works with anything 2.0...3.6 V)
08:40:09amiconnIf we want to be cautious, we'd have to evaluate this too. (Not done yet)
08:42:25[IDC]Dragonisn't it too late then, the card is already plugged?
08:43:07amiconnNo, as long as you don't access the flash core (read/ write data blocks)
08:43:25[IDC]DragonOK, then we should do so
08:43:54amiconnMaybe the archos fw does not do this though, so it accesses the card regardless while booting
08:44:24[IDC]Dragonbut only the internal, I'd reckon
08:45:03amiconn(As long as it finds an ajbrec.ajz there)
08:45:29dwihnoDo you guys know if it's a bad idea to remove a flash cart while a unit (disregarding flash type) is powered on?
08:45:48LinusNthe BDM driver is working
08:46:11 Join Zagor [242] (
08:46:18LinusNtoo bad the target system is 50km away, so i can't turn it on :-)
08:46:57LinusNcoffee time
08:46:59dwihnoLinusN: well, use the remote debugging feature(?) :)
08:47:09amiconnGotta run.
08:47:31PaulSLinusN: Awesome! Teach that thing some manners. :-)
08:47:51[IDC]Dragonamiconn: cu
08:49:28[IDC]DragonZagor: I'm reaching the point where multi-choice config definitions are useful
08:49:47[IDC]Dragonfor tuner, hardware codec, battery, etc.
08:51:57Zagori'm fine with that
08:52:23[IDC]Dragonhow would you like the "none" case?
08:52:33[IDC]Dragontake the tuner, for example
08:52:54Zagor#ifndef TUNER
08:53:02[IDC]Dragonsay, we have CONFIGURATION_TUNER
08:53:17[IDC]Dragonwe set it to samsung, philips, etc.
08:53:32[IDC]Dragonor to NONE, or not define it at all?
08:53:35ZagorI prefer CONFIG_TUNER, since the header files are called config.h
08:53:51[IDC]DragonI like shorter, too
08:53:51Zagorthen all targets need all defines, which is bad
08:54:51[IDC]Dragonwill #if (CONFIG_TUNER == TEAxx) not complain if undefined?
08:55:05***Saving seen data "./dancer.seen"
08:55:14[IDC]Dragon(simple preprocessor question, soory for rtfm)
08:56:44Zagorit will complain, but we just put an #ifdef TUNER around the whole file
08:56:44[IDC]Dragonin other words, would we need a #ifdef CONFIG_TUNER around the whole thing, too?
08:57:06[IDC]Dragonhmm, doesn't look nice
08:57:22Zagoradding all (future) config options to all targets isn't very nice either
08:58:26[IDC]Dragonso we could do #if defined(CONFIG_TUNER) && (CONFIG_TUNER==TEAxx)
08:58:26LinusNdwihno: i am debugging remotely from work, but i need to connect and turn on the iriver...
08:59:02dwihnoLinusN: aah. snibaaz! :/
08:59:21LinusN[IDC]Dragon: so your tuner configured as I2C?
09:00:02Zagoryes we could. but after all fmradio.c depends on the presence of a tuner in the first place and so a single #ifdef around the file is sufficient
09:00:21[IDC]DragonLinusN: haven't tried it yet, but will soon
09:00:36LinusNbut are you sure it's i2c and not 3-wire?
09:00:54[IDC]DragonZagor: for that place, yes, but there may be others.
09:01:07[IDC]DragonLinusN: pretty much, yes.
09:02:04[IDC]Dragonbut it's good to be unspecific in the app (#ifdef CONFIG_TUNER) and more specific in the driver (#if CONFIG_TUNER == xx)
09:02:26Zagori agree
09:03:31[IDC]Dragonone catch: I suspect the Ondio may have the Samsung tuner on some models, versus I have Philips
09:03:52[IDC]Dragonso we'd have to runtime decide
09:04:01LinusNi think the tuner code shouldn't be compiled at all if there is no tuner
09:04:06Zagorouch. let's solve it when we encounter it.
09:04:13ZagorLinusN: yes, that is the next step.
09:04:16[IDC]DragonZagor: yes, definitely
09:04:50[IDC]Dragonthe OndioSP build is different anyway, it has no tuner defined.
09:09:33Zagoryes but that is already covered by the configure script
09:10:17[IDC]Dragonyes, I know
09:15:26LinusNnow i know how people must feel when they send patches to this project, and they never get applied
09:15:46[IDC]Dragonwhy now?
09:15:56LinusNi've sent my second patch to the binutils project, and i'm completely ignored
09:16:27[IDC]Dragonperhaps you haven't waited 6 month yet
09:16:35Bagder_well, we usually at least comment the patch
09:16:37LinusNi guess i'll have to :-)
09:16:41 Nick Bagder_ is now known as Bagder (
09:20:42LinusNPaulS: have you disassembled the startup code to any extent?
09:22:06LinusNi'm interested in what it does to stay on when it is started with the Play button
09:23:55PaulSLinusN: I'm not quite sure about what you mean by "what it does to stay on"
09:25:12LinusNi haven't investigated the power supply yet, but i assume that the firmware has to do something (set a port bit) to make the regulator continue feeding power to the cpu when the user releases the button
09:26:30[IDC]Dragoncu later
09:26:38 Quit [IDC]Dragon ()
09:27:03PaulSWell, there's a bunch of port accesses in the code in flash. If you have a guess as to which GPIO ports are attached to the power supply, I can tell you more about what the state of those pins are.
09:31:55ZagorLinusN: can you commit the recording buffer change we discussed?
09:32:42ZagorBagder: excellent
09:33:15Bagderthe email@ reaches all three of us
09:33:52Bagderthe /etc/mail/virtusertable file is what needs editing to add people
09:40:30LinusNZagor: sure, but are you sure that it will solve anything for anyone?
09:54:36LinusNdoes the 1.8" hd run with 3.3 or 5 volts?
09:56:58LinusNbtw, it looks like GPIO51 is the one i'm looking for
09:57:14LinusNgoes to what looks like a voltage regulator
09:57:18 Join amiconn_ [0] (
09:58:07PaulSOkay, let me take a look.
09:58:31LinusNyour wiki table says that it is set to 0 at power off
10:01:15 Join [ [0] (
10:01:37 Part [
10:01:52 Join [IDC]Dragon [0] (
10:02:53[IDC]DragonI didn't notice until yesterday that rockbox has a domain
10:03:09ZagorLinusN: of course i'm not sure. feel free to make a special build and send to whoever had the problem. i just want to release 2.3.
10:03:34LinusNZagor: it is unlikely that it solves Paul's problem, since it works with 2.2
10:04:42Zagorcan we please test it before rejecting it?
10:05:12PaulSLinusN: A 1 gets written to GPIO51 pretty early on in the flash.
10:06:47[IDC]Dragonare those new forums replacing the mailing list?
10:06:58Zagor[IDC]Dragon: only for user issues
10:07:23[IDC]Dragonwhich is 90%
10:08:42Zagorthe forum is a complement for those who are afraid of mailing lists. mostly that means non-technical users.
10:09:11Zagorthe mailing list has always been about development, not user issues. we've been referring to other sites for pure user interaction.
10:11:54Bagderyeps, the MLA (Mailing List Afraid) people have been using the yahoo group and newmp3 tech instead
10:12:02Bagderthis is just Yet Another User Forum
10:12:04[IDC]Dragonit is very diversed, are you sure this is helpful?
10:12:25[IDC]Dragonwho should read all forums, or people need to
10:12:51[IDC]Dragone.g. the V2 recorder has its own forum
10:12:52BagderI surely will never read any of the forums close enough
10:13:48[IDC]DragonI'd merge recorder, FM, V2
10:14:05PaulSLinusN: It really doesn't look like they're in a big rush to assert GPIO51, though. It's not like the first thing it does. It's the 4th or so function call in the block of initialization calls. There must be something lofting that signal afloat during the interim.
10:14:09[IDC]Dragonbut miss the Ondio forum ;-)
10:15:24 Quit amiconn (Read error: 110 (Connection timed out))
10:15:25 Nick amiconn_ is now known as amiconn (
10:15:26[IDC]DragonI haven't found any Ondio forum anywhere, so it would be a premiere
10:15:54 Join Chronic007 [0] (~Miranda@
10:23:08Zagorsend all feedback to jeff. the forum is his creation.
10:28:19LinusNPaulS: they aren't in a hurry, in fact, they should wait so that a tiny glitch doesn't start the device
10:32:21PaulSOh, so the play button holds the power supply on long enough for us to get to this code, or the machine goes back to sleep..
10:33:08 Join webguest74 [0] (
10:33:14webguest74wow !
10:33:20PaulSThis should imply that I can keep the machine on by holding down the play button? I'll try that..
10:33:22webguest74didn't excpected to see such a crowd
10:33:26webguest74hello :)
10:33:34Zagorhi there
10:33:51webguest74I've just discovered your project
10:33:52*Bagder waves
10:34:07webguest74and I've got a very very very important question to ask you
10:34:14Bagderwelcome to our merry corner of the world
10:34:29webguest74u can actually see it on
10:34:38PaulSHmm.. The box can turn itself off even with the play button held down.
10:34:39webguest74I'm the one who started the topic
10:35:23Bagderand what exactly is the question?
10:36:00webguest74the questio was about playing ym or sndh files (atari St sound processor) or sid ones on Iriver IFP players ?
10:36:29 Join pillo [0] (
10:36:30Bagderis it very CPU intensive?
10:36:55webguest74i asked for it and a friend (DMA/Sector one ) told me about your rockbox project
10:37:12PaulSBadger: Prolly not −− I think an iRiver outclasses an Atari ST in CPU power, unless there were some serious ASICs involved.
10:37:36Bagderright, I don't think the ST had much extra power than the CPU itself
10:37:40Zagorwebguest74: if you are referring to tracker "mod" files, yes it is possible to make a player for them on the iriver
10:38:26webguest74so i wanted to invite you to discuss this topic on the atari-forum, as everyone seems to be pretty much interested in such a player
10:38:53webguest74we'd love to hear from the rockbox programmers about that question
10:39:07Zagorthanks, but we're still in the infant stages of the port and haven't got any running code yet
10:39:10Bagderwe don't even have it running on the iriver yet
10:39:28Bagderand the Archos is too weak to do it
10:40:09webguest74ok, badger... I saw you were talking about iriver with Hdds but do u think it would be possible on models without Hdd like the IFP series ?
10:40:10Bagderwebguest74: it seems more reasonable to rever the invite: you join a suitable rockbox forum/list and help us develop
10:40:49Zagorwebguest74: that depends on the hardware. if it's coldfire-based, it is not impossible. but we're focusing on harddisk-based players.
10:40:58webguest74it is coldfire based
10:41:26Bagderthen it depends on how helpful people with such players are
10:42:04webguest74ok badger and zagor, i just wanted to ask... it's true that i'd LOVe to see a port for my iriver, so i think if it could be opssible i could help
10:42:14Bagderwe don't get any help from any manifacturer, we have ONLY volounteers
10:42:19webguest74i'm no programmer but, i don't know
10:42:32webguest74about YM format, everything is on
10:42:32Zagorall platforms anyone wants rockbox on must be thoroughly documented first (and that is *thoroughly*, as in every single chip and even the board schematics)
10:42:58webguest74yes i've seen the huge work on the site, it's impressive
10:45:11Zagorwhat I mean is that we are only a few people, we cannot do all work for all platforms. if you want rockbox on a specific platform, doing the research is the best way to increase the chances of that happening
10:46:23amiconnBagder: Iirc the Atari ST features a DSP in addition to its m68k CPU
10:47:02Zagorhmm, ym is YM2149, a dedicated sound chip. this must be emulated, which means it will require a lot more cpu power than mod files.
10:48:27Zagorthat puts it on par with sidplay, in the "maybe" area.
10:48:37webguest74yes, it's quite similar to sid
10:49:24webguest74but as for the harddisk issue only, nothing would prevent rockbox to run on iriver iFp series rather than Ihd ones ?
10:49:53Zagorlots of things would prevent it, since the hardware is different. code must be written to support the hardware.
10:50:16webguest74:/ ok ok
10:50:41webguest74thanks a lot for the time you spend answering my questions
10:55:06***Saving seen data "./dancer.seen"
11:00:14PaulSYM2149 datasheet: We're talking about tone generators and envelope registers. I wouldn't want to simulate the DSP, but even that probably wouldn't sink the iRiver.
11:01:22webguest74that's a pretty precise documentation! :-)
11:03:52[IDC]Dragonthere is no DSP, those weren't invented
11:04:03[IDC]Dragonit is FM synthesis
11:04:15[IDC]Dragon(analogue, iirc)
11:05:07amiconnI remember having read something about a 56000 DSP in conjunction with the Atari. Maybe this was about an extension board...
11:06:32webguest74i think the first ataris ST only had the YM
11:06:41webguest74and then the STE series had a dsp
11:06:52webguest74but it was not used by anyone
11:06:59webguest74i dont ' know
11:07:08webguest74if this page could help :
11:07:49PaulSI think the 56000 came a lot later.
11:10:04webguest74but in fact, the good format to play atari music is the sndh one
11:10:18 Join MooMaunder [0] (~me@
11:10:27 Join kurzhaarrocker [0] (
11:11:05 Join MisticJeff [0] (
11:11:47MisticJeff[IDC]Dragon: you now have your Ondio Forum ;)
11:11:58webguest74The SNDH files can include any kind of YM-2149 playing
11:11:59webguest74 routine. Including SIDvoices, Syncbuzzers and Digidrums.
11:19:43 Join tester [0] (
11:20:17 Quit tester (Client Quit)
11:21:21 Join CGI713 [0] (
11:21:41amiconnMisticJeff: On the forum pages, the side menu does look a bit odd. Perhaps a css problem.
11:22:04 Part CGI713
11:22:16MisticJeffyes, me messing around this evening and not paying attention. i'll fix in the morning
11:22:19kurzhaarrockerI read something about a recording buffer change by LinusN. What is it supposed to fix?
11:23:41 Quit webguest74 ("CGI:IRC (EOF)")
11:23:42Zagorkurzhaarrocker: i want to test lowering the recording buffer watermark so we write out to disk long before the buffer is full, and see if that makes any difference on the reported problems. some have said there are more problems with full disks and/or long files.
11:24:14kurzhaarrockerHm. Might even be related to a problem I recently had: stuttering recordings.
11:26:57 Part Chronic007
11:27:11 Join ripnet [0] (
11:33:47[IDC]DragonMisticJeff: thanks!
11:38:09PaulSLinusN: I'm guessing the iRiver uses the 160 pin version of the 5249 part?
11:38:32[IDC]Dragonamiconn: did you commit bitswap.h, too?
11:42:57ripnetare you going with what PaulS is trying to do, ie moding the iRiver firmware (maybe via a .ips patch) to include a boot loader? im guessing the idea is that if a certain file exists and/or a key is pressed during boot, it loads the file into memory and executes it?
11:48:11amiconn[IDC]Dragon: Ahem, 2 days ago, yes.
11:48:53PaulSNobody's going with anything yet. Essentially all my efforts are premature, and I'm now paying for it. :-) The "keypress during boot" option, while attractive, isn't an option we know how to do yet with the iRiver firmware. I was trying to override the functionality of the "Firmware Upgrade" menu option to support RAM-loaded files.
11:52:22PaulSIt's sleepytime over here in PST. Goodnight.
11:53:00 Part PaulS ("ZzzzzZzzzz...")
12:00:22[IDC]Dragonamiconn: ah, that's why I didn't see it there. sorry.
12:02:12MisticJeffGoodnight everyone...
12:02:15 Quit MisticJeff ()
12:07:03 Quit pillo (Read error: 54 (Connection reset by peer))
12:07:36amiconn[IDC]Dragon: It's still on the rockbox front page...
12:19:30 Quit ripnet ("CGI:IRC")
12:55:10***Saving seen data "./dancer.seen"
12:57:15[IDC]Dragonamiconn: yes, yes, I was only looking at the commit of this morning :-/
13:02:47amiconnYour swapbyte macro idea is not yet implemented
13:05:13 Join Schoki2_ [0] (
13:33:41[IDC]Dragonhow ironic, I found a wiggler here!
13:33:57[IDC]DragonMcCraigor or so, is that the one?
13:42:13LinusNunfortunately not
13:42:41 Part Schoki2_
13:42:57LinusNthe macraigor wiggler has too few i/o:s, and lacks logic for clock synchronization
13:43:08LinusNi have two macraigor wigglers here too :-)
13:45:14amiconn[IDC]Dragon: Are you preparing to do the same thing with the iRiver flash players that we are doing with the Ondio now? ;-)
14:10:26 Join elinenbe_ [0] (~elinenbe_@
14:23:49kurzhaarrockerIs there special software for those wigglers?
14:26:14LinusNkurzhaarrocker: yes
14:28:56kurzhaarrockerIs that a dedicated program with a fancy gui or more something like a api?
14:37:44LinusNit's gdb
14:37:58LinusNwith a dedicated wiggler driver
14:38:33LinusNthe wiggler driver is a kernel driver and a user space library
14:38:53LinusNand the library can be used for more "general" wiggling, like programing flash etc
14:43:15 Nick midk_ is now known as midk (
14:43:40kurzhaarrockerOk, now I get the idea. So the wriggler is specialized for communicating with digital hardware. Not just a bunch of general purpose i/o pins.
14:45:23LinusNit's mostly general purpose pins
14:45:36LinusNthe macraigor wiggler is only general purpose pins
14:45:55LinusNbut the p&e wiggler contains some extra logic
14:46:39LinusNthe macraigor wiggler is nothing more than a buffered parallel port
14:49:45kurzhaarrockerCan you debug systems only if they provide some single-step-capabilities? Otherwise I'd be rather surprised if the speed of a parallel port would be fast enough to trace/control the signals in the real world.
14:50:07midkwee! forums! :)
14:50:16LinusNthe wiggler is designed for BDM, Background Debug Mode
14:50:41LinusNi.e you stop the CPU and communicate via a bit-banged serial interface
14:52:16kurzhaarrockerDoes that include modifying the hardware? For example disconnecting some clock signal generators or handshake lines?
14:54:17LinusNthe coldfire has a dedicated bdm port
14:55:12***Saving seen data "./dancer.seen"
15:00:43kurzhaarrockerbtw: is fine but just results in "Hepp" :)
15:01:05Zagoryup, no server there :)
15:01:44kurzhaarrockerIs that a bug or a feature?
15:01:48Bagderlet's fix that
15:01:57*Bagder will fix
15:02:03midkpeople: doesn't work, you have to have "www." in front of it?\
15:02:27midkis that changeable?
15:02:32*Bagder will fix
15:02:37Zagoruse the www address
15:03:02kurzhaarrockerMaybe we should replace "Hepp" with "use the www adress" :)
15:03:21midkzagor: i will if i have to, i'd just continue using '' in that case since i'm used to it :)
15:03:32 Quit [IDC]Dragon ("CGI:IRC (EOF)")
15:03:45Bagdermidk: try now
15:03:56Zagormidk: well it redirects you to anyway
15:04:01midkthere it goes
15:04:12BagderZagor: you mind if I change how that is done?
15:04:15kurzhaarrockercool :)
15:04:16midkZagor, yeah, i realized that :)
15:04:27ZagorBagder: sure, as long as the result is the same
15:05:18BagderZagor: I think its nicer to use RedirectPermanent from a separate virtualhost
15:05:25Bagderdone now
15:06:05Zagoryou forgot a slash somewhere: http://www.rockbox.orghistory.shtml/
15:06:54Zagoryup, works
15:10:47 Part LinusN
15:10:50 Join LinusN [0] (
15:11:01 Part Zagor
15:11:09 Part kurzhaarrocker
15:14:35 Join Zagor [242] (
15:27:19 Nick ashridah is now known as Lost-tv (
15:44:51 Join R3nTiL [0] (
15:51:50 Quit ze (Read error: 110 (Connection timed out))
16:00:04LinusNtime to go
16:00:06 Part LinusN
16:03:27 Join Hellish [0] (
16:03:37HellishHi everyone
16:04:14Hellishdoes anyone know the current development process of the new rockbox firmware for the ihp-120
16:04:25Hellishive been reading loads about it and im getting excited
16:13:09 Join methangas [0] (
16:14:19 Quit Hellish ("CGI:IRC (EOF)")
16:21:23 Nick Lost-tv is now known as ashridah (
16:29:09 Part Zagor
16:29:51ashridahyeah. so. is the iriver port going to be finished tomorrow?
16:29:56ashridahwhat about the next day? :)
16:32:44 Quit R3nTiL ()
16:55:16***Saving seen data "./dancer.seen"
17:00:23 Join ze [0] (
17:00:41 Join mecraw [0] (~lmarlow@
17:08:15 Join [IDC]Dragon [0] (
17:09:01[IDC]Dragonserver fiddling kills cgiirc, I guess
17:14:33 Quit ze (Read error: 60 (Operation timed out))
17:43:02 Join AciD [0] (
17:52:16 Quit ashridah ("sleep")
17:54:44 Join oxygen77 [0] (
17:54:55 Part oxygen77 ("Cho")
18:01:16 Quit AciD (Read error: 104 (Connection reset by peer))
18:03:53 Join AciD [0] (
18:19:47 Join ze [0] (
18:44:23 Quit [IDC]Dragon ("CGI:IRC")
18:55:19***Saving seen data "./dancer.seen"
19:01:24 Join R3nTiL [0] (
19:02:36 Quit R3nTiL (Client Quit)
19:05:13 Join webguest35 [0] (
19:05:32webguest35hi all
19:05:59webguest35stupid question: where are the .bin files (for flashing) located?
19:06:21webguest35I've got rockbox loaded
19:06:31webguest35but I want to flash the rom
19:06:46amiconnBagder: I'm looking for a place to put the config-block sector calculation into without creating too much cross dependency.
19:06:47webguest35the files aren't in the .zip (or I missed them)
19:07:05Bagderwebguest35: you flash .ucl, not .bin
19:07:28webguest35maybe I misread the howto.
19:07:31amiconnwebguest35: First-time flash, or flash a newer build?
19:07:36webguest35first time
19:08:02webguest35if i try the flash util, it says the file is not found
19:09:18webguest35am i VERY lost?
19:09:30amiconnFollowing the howto, there are links way down where the first-time flash packages can be downloaded
19:10:19webguest35I was thinking it'd be part of the main download.
19:11:17amiconnBagder: Got my question?
19:12:45Bagderyes, but I have no time to give a good answer now
19:13:24amiconnOkay, so I'll have to find it myself. I need this before I write-enable the flash driver...
19:13:39BagderI trust your ability ;-)
19:21:51webguest35Looks like all went well. much faster boot. thanks
19:22:22 Quit webguest35 ("CGI:IRC (EOF)")
19:48:51 Quit pike (Read error: 232 (Connection reset by peer))
19:49:04 Join pike [0] (
19:59:30amiconnUrgs, more #ifdefing
20:55:23***Saving seen data "./dancer.seen"
21:48:39 Join scott666_ [0] (
21:56:42 Join [IDC]Dragon [0] (
22:03:56amiconnhi again Jörg
22:04:10[IDC]Dragongood evening
22:04:31[IDC]DragonI'm starting the multiple choice config
22:04:44[IDC]Dragon*lots* of file to touch
22:04:58[IDC]Dragonexpect a major, colorful commit
22:06:45amiconnHmm. I'm currently trying to figure out why the mmc driver works for a number of tries and then suddenly locks up. This does happen with the internal card only; with the external it will simply fail at some point, but not lock up.
22:06:55 Quit Nibbler (Read error: 110 (Connection timed out))
22:07:17amiconnI hesitate to add writing until I get it running stable...
22:07:30[IDC]DragonI understand
22:08:23dwihnoYou guys got MMC reading working already?
22:08:27amiconnBtw, did I say I hate that internal flash chip?
22:09:33[IDC]Dragonno, you didn't
22:10:54amiconndwihno: Yes, although read-only for now and obviously a bit unstable. See today's cvs activity log
22:11:54amiconn[IDC]Dragon: I do know now why the menu isn't callable from the browser, but have currently no idea how to solve this.
22:12:11dwihnoYou guys rock!
22:12:41[IDC]Dragonwhy is it not?
22:13:32amiconnWith Zagor's new button code, the menu key alone is defined to be shift, with repeat it should call the menu. However, this can't work, because triggering shift jumps away into a separate function, so menu+repeat is never caught in the main loop...
22:14:35[IDC]Dragonwith shift you mean like On+Up/Down for the recorder?
22:14:48amiconnA similar problem exists with the onplay (on Ondio == menuright) menu
22:15:35amiconnPlus, on its own (without using up/down) this is the browser<->wps toggle
22:16:03amiconnThere are definitely too few keys...
22:16:49[IDC]Dragonsigh, yes
22:17:03[IDC]Dragonor, Rockbox is overpowered
22:17:24[IDC]Dragonwe should not assign shift for now
22:17:45amiconnI would like to switch the browser display to "supported" in order to use RoLo, but obviously can't do so with the menu unavailable...
22:17:50[IDC]Dragonmenu is more important
22:18:45[IDC]Dragonyou can hard-coded "supported" as default, for your local build
22:33:47[IDC]Dragonamiconn: how is your MAS called again?
22:34:05amiconnMAS3539F, see wiki ;)
22:34:21[IDC]Dragonah, I forgot wiki :-(
22:34:37amiconnThis is controlled the same way as the 3587F, with 2 differences:
22:34:45amiconn(1) There is no recording
22:35:05amiconn(2) The memory cells have different addresses
22:35:12amiconnThe registers are all the same
22:35:57[IDC]Dragonwe'll have to differentiate that
22:36:09amiconnIt offers an additional, more efficient transfer mode, but we are not forced to use this...
22:36:24[IDC]DragonI'm introducing CONFIG_HWCODEC
22:36:48[IDC]Dragonwith the MAS types as values
22:36:58amiconnI intended to handle the different memory cells via #defines, so the access code itself can be shared
22:37:22[IDC]Dragonyes, makes sense
22:39:40amiconnFor instance, there is a cell called AppSelect in both MAS's, which is D0:7F6 for 3587 and D0:34B for 3539
22:40:27[IDC]Dragonlater, later ;-)
22:41:22 Quit uski ()
22:41:36amiconnImho this will be the next thing I'll have to do as soon as MMC is working. I want to listen to some music at last. Of course you don't need this...
22:42:26[IDC]Dragonthe debug menu also doesn't show the MAS in good shape for me
22:42:57amiconnPerhaps you have to init the SIBI pin properly...
22:43:00[IDC]Dragonperhaps some port init problem
22:43:24amiconnOther than that, the MAS control looks identical to the recorders
22:43:49[IDC]DragonI need to fix the polarity
22:44:07[IDC]DragonI'm not sure which one is for directly connected
22:44:49[IDC]Dragonthat was a mask bit setting for recorders
22:44:59amiconnAh ok.
22:45:08[IDC]Dragonwhether there's an inverter or not
22:45:27amiconnSet PB8 to GP out & low.
22:46:06amiconnWe need a mas_init() ...
22:47:47amiconnGrr, this simply gets stuck without showing even one of the splash()es I added all over...
22:55:26***Saving seen data "./dancer.seen"
22:56:04 Join Sebulba03 [0] (
22:58:45amiconn[IDC]Dragon: For your config method change, that the current HAVE_BATTERIES is not the best solution, since it means there are both batteries and a charger. The latter is obviously wrong for the Ondio. So this should be split into something like HAVE_BATTERIES and HAVE_CHARGING.
22:59:00amiconn*keep in mind, that
22:59:24[IDC]Dragonyes, that's on my list, too
22:59:40[IDC]DragonI also want to specify what kind of batteries
23:00:00[IDC]Dragonto avoid #ifdef HAVE_LIION
23:00:01amiconnThen there is that cross-dependency between recording and bitmap display....
23:00:12[IDC]Dragonlots of them
23:00:24[IDC]DragonI'm not starting there :(
23:00:55amiconnThere is even a cross-dependency between bitmap display and the firmware file name/ extension
23:01:54amiconnWe need a typical alkaline discharge curve to get the battery display right on the Ondio
23:02:59amiconnDon't forget the F buttons - certainly we don't want the button bar on the Ondio
23:08:25 Quit methangas (" HydraIRC -> <- The dawn of a new age")
23:18:56 Join Bagder_ [0] (
23:19:11 Quit Bagder (Read error: 104 (Connection reset by peer))
23:20:22 Nick Bagder_ is now known as Bagder (
23:28:57 Join NibbIer [0] (
23:33:39amiconnDoes somebody know whether the ata functions are called from more than one thread? It seems that the ata_mutex causes the sporadic hangs...
23:34:37*Bagder has no clue
23:35:29[IDC]DragonI think they are, else a mutex makes no sense
23:36:06 Quit Sebulba03 ("brb")
23:37:02 Join Sebulba03 [0] (
23:37:21amiconnHmm, this leaves the questions (1) which threads could do this on bootup and (2) why there is no such race with the hd models. The bad thing is that now that I have splash()es both before and after the mutex_lock(), the hang does no longer happen...
23:38:03amiconnA splash is slow, of course. A led on a port pin would come in handy...
23:39:12 Quit Sebulba03 (Client Quit)
23:40:16[IDC]Dragonif you don't mind some soldering...
23:41:39amiconnIsn't PB4 available at the tuner board header? It does have no function on the Ondio SP... I could simply set/reset it,and measure the voltage in case of a hang
23:42:46 Join Sebulba03 [0] (
23:43:34amiconnRoLo hangs... :(
23:45:15[IDC]Dragonyes, PB4 on the tuner board is free
23:45:38[IDC]Dragonrolo maybe does pad port inits
23:46:35[IDC]Dragonthe backlight pin is free as well
23:53:19amiconnIt's not the mutex...
23:57:24amiconnThe flash is specified for 100.000 writes per sector. I wonder how many of them I did already use up ;-/
23:57:57[IDC]Dragondon't worry, it has replacement sectors
23:58:16[IDC]Dragonrolo would really help...
23:58:18amiconnYes I know. It has very sophisticated error handling
23:58:43amiconnYes, especially as the archos usb mode is really sloow compared to rockbox usb mode

Previous day | Next day