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 2004-09-30

00:00:17[IDC]Dragonthen it's just too slow
00:00:59midkhardware poweroff is (for the fm/v2) 5 seconds...
00:01:14[IDC]Dragongoodnight folks
00:01:20midknite idc
00:01:45 Quit [IDC]Dragon ()
00:04:15webguest71bye ...
00:04:21 Quit webguest71 ("CGI:IRC (EOF)")
00:14:06***Saving seen data "./dancer.seen"
00:15:42 Quit _aLF ("bye")
00:21:40amiconnI killed the file system...
00:22:32amiconnUhh, not exactly. Let's see what is wrong...
00:25:31amiconnDunno what went wrong. Archos firmware says there is no valid partitio, but windows does read it fine. Anyway, I just reformatted.
00:31:44amiconnRoLo works on the Ondio!
00:31:57midkwouh! nice work! :)
00:32:37amiconnFonts and languages too (not from the menu though)
00:34:29midkthat's great
00:34:31midkpost pics :)
00:38:02amiconnTrying to dump the rom into files doesn't work. All I get are zero length files, and the first try caused the file system corruption
00:38:35midki see.. :\
00:39:02amiconnAnyway, I need to sleep now.
00:39:26 Part amiconn
01:01:38 Join NelsonM [0] (
01:01:54NelsonMhelp, my rockbox won't boot anymore!
01:02:12midkplease describe the problem a bit. i've got no idea what has happened or could be wrong.
01:02:22NelsonMI can't turn it on in play mode or USB mode anymore. Is there a way to disable the rockbox bios and just boot from firmware?
01:02:33NelsonMI have a 20 gig recorder that's been working for a year. I haven't tried to flahs the bios.
01:02:45NelsonMwhen I turn it on it comes on for a few seconds, then turns off and reboots. I don't think it's out of power.
01:02:56midkcould you plug it in to be sure?
01:03:20NelsonMyeah, I should do that, I don't have the charger here.
01:03:58NelsonMis there some way to tell the Archos firmware not to load an OS from the hard drive? I thought F1+on may do that, but I'm not sure.
01:04:32NelsonMf1+on doesn't do that?
01:09:17midkback, rather
01:09:25NelsonMthanks for the help. I love rockbox.
01:09:55midkis the backlight or lcd dim? that would indicate very low batteries..
01:10:11NelsonMyeah, it's not, and I just charged it up. But maybe the batteries have gotten old, I should try again.
01:10:44NelsonMI sure wish Rockbox worked on hardware that had a Li-Polymer battery. These NiCd batteries are junk.
01:11:03midkNiMH, actually. if you get higher capacity you get better runtime
01:11:15NelsonMwill AA NiMH work fine in the rockbox charger circuit?
01:11:22midkships with 1500mah, ~8h. you can get 2100 or 2200mah, ~12h.
01:11:28midkyep. any nimh.
01:11:42NelsonMWhat do rockbox developers do if they put a broken OS on the hard drive? Is there any way to recover it other than plugging the drive into a laptop?
01:12:15 Join bagawk [0] (Lee@bagawk.user)
01:12:21midkno, you can go into usbmode without booting the firmware really
01:12:27midkto delete rockbox if needed
01:12:38midkplug in usb, then turn it on. as opposed to booting it then plugging in usb.
01:12:44NelsonMah, I thought the rockbox OS got involved in USB mode too.
01:13:04midkit does.
01:13:22midkbut archos firmware starts before rockbox. and if usb is in right when it's started it automatically does that instead of harddisk firmware
01:13:39NelsonMoh, cool
01:13:58midkyou can try it, not sure if it will help.
01:14:04midksounds like the batteries may be the problem.
01:14:10midkcan you feel the harddisk spin up?
01:14:54NelsonMjust barely, a bit of spin up but no seek
01:15:03NelsonMquarter second or so.
01:15:43midkok, that must be the batteries.
01:15:51NelsonMI hope so, that's an easy fix :-)
01:15:58NelsonMjust have to go home and plug it in.
01:16:10midkyeps.. hopefully that's it :)
01:16:11NelsonMbtw, rockbox got 10x more useful to me when it implemented bookmarks. Great for audiobooks
01:16:24 Join edx [0] (
01:16:45midkyeah, they are
01:19:25 Quit mecraw ("Trillian (")
01:23:51 Quit NelsonM ("Leaving")
01:46:21 Quit edx (Read error: 238 (Connection timed out))
01:51:10 Quit bagawk ("umount /dev/brain")
02:00:22 Join ashridah [0] (ashridah@
02:14:08***Saving seen data "./dancer.seen"
02:14:12 Join Chronic007 [0] (~Miranda@
02:38:36 Quit Chronic007 ("Miranda IM! Smaller, Faster, Easier.")
02:51:42 Join scott666_ [0] (
03:23:58 Join edx [0] (
04:14:11***Saving seen data "./dancer.seen"
04:50:42 Quit scott666_ ("i'll be back...eventually...")
05:10:00 Quit pyros- (Read error: 242 (No route to host))
05:38:07 Join LinusN [0] (
06:14:14***Saving seen data "./dancer.seen"
06:23:33 Join AciD [0] (
06:36:13 Quit ashridah ("gone")
06:49:07 Quit edx ()
07:01:03 Quit dwihno (Remote closed the connection)
07:08:49 Join DomasoFan [0] (
07:09:10 Quit DomasoFan (Client Quit)
07:10:58 Join PaulS [0] (
07:11:25LinusNhi PaulS
07:11:30PaulSHey, Linus. My first attempt at JTAG was met with mixed success.
07:12:04LinusNit struck me that the pins are configured as bdm pins
07:12:54PaulSMy first realization (even before I started) is that the TEST[3:0] pins need to be set correctly for the Coldfire to go into JTAG vs BDM mode. Thanks to you nudie pics, I found the 10K resisitor that pulled the chip into BDM mode and shorted it to ground like TEST[3:1] were.
07:13:08LinusNok good
07:15:09PaulSMy second challenge was that I couldn't actually find a test pad for TCK like the other JTAG pins hanging out by the BDM pads. I did a little light probing (don't want to do too much on a populated board), and still couldn't find it. Again, thanks to your nudie pics, I was able to find the backside of a via that carries TCK. It's strange, though. I'd imagine that if there's a via, that signal must be going somewhere.
07:17:17PaulSMy next challeng was finding software that understood the Coldfire in debug mode. I picked up the OpenWinCE jtag tools, and wrote a part definition based on the BSDL file from Freescale, and then wrote a bus definition based on the datasheet.
07:19:06PaulSMy current challenge is to get my laptop parallel port working with either ppdev or direct iopl() so I can test all of the above. Hopefully I'll get some time tomorrow night to try that out.
07:24:30LinusNa qick search for the TCLK signal gave me nothing on the PCB
07:25:26LinusNi.e no test pad or pullup resistor...
07:25:48PaulSI also tried some of the pins of the (unpopulated) flex connector on the backside of the board (the ones that weren't obscured by the LCD padding) and still couldn't find TCK.. Odd. It seems, BTW, that anyone who's serious about playing with BDM/JTAG on the iRiver would pick up that flex connector. There are little cable assemblies you can buy to stick into them, so all you have to do is solder your wiggler to that −− of course this is all modulo sig
07:26:16 Join dwihno [0] (~dw@
07:27:07LinusNi'd like to find a good place to buy such a connector
07:27:33PaulSIsn't it odd that there's still a via next to the pad though?
07:29:00LinusNfound it, R35
07:30:09LinusNpullup to VCC it seems
07:30:24 Join amiconn [0] (
07:30:41LinusNR35 is located right below the CPU
07:32:23 Join [IDC]Dragon [0] (
07:32:43PaulSWhat pitch would you say those debug connector pins are? I'm guessing something along the lines of would work, which would make be a reasonable source.
07:33:00PaulSLinusN: R35: Oo. Let me take a peek.
07:34:23PaulSI tested the other pads from the vias over to the test pads and I didn't get any nasty surprises, so I think I'm in good shape otherwise.
07:34:36LinusNbut that connector has two rows of pins, doesn't it?
07:35:39PaulSWhups.. Let me find the right part..
07:37:46 Part LinusN
07:37:57 Join LinusN [0] (
07:40:36PaulSCheck out Unfortunately the 24 pin version is not in stock at Digikey. Also, you need to confirm that the pin pitch is actually 0.5mm on the iRiver. That was just my guess.
07:41:01 Join DomasoFan [0] (
07:42:00LinusNtoo bad the LCD connector is only 18 pins, otherwise i could just have used the one from the naked board :-(
07:43:24 Quit DomasoFan (Client Quit)
07:48:48 Quit PaulS ("CGI:IRC (EOF)")
07:49:29 Join PaulS [0] (
07:50:46PaulSHere's a part DigiKey has in stock: XF2H-2415-1
07:53:29LinusN24 contacts?
07:54:00PaulSThen, say you need a 6" flex jumper. In stock too. 050-24-152B
07:54:10PaulSWas I wrong about the number of contacts?
07:55:08LinusNshould be 20
07:55:25LinusNhow about this one: HFJ20TR-ND
07:55:45PaulSAh. I hadn't tried to count them since before the nudes..
07:57:22PaulSHey, you're getting the hang of it!
07:58:16PaulSActually, Digikey's site puts a reasonable amount of pleasure into the act of finding the right part. I get the catalog in the mail but I hardly ever use it.
07:58:57LinusNas opposed to Farnell, where the web search is a real pain
08:01:06PaulSThe Omron XF2J-2024-11 connector is $1.03, compared to the Hirose $2.55. How could you justify the extra cost, especially in single unit quantities? :-)
08:01:38PaulSActually, I think Digikey has a $20 minimum for web orders.
08:02:18LinusNthe shipping to sweden will probably cost the double :-)
08:02:55PaulSWell, so far you don't have a lot of shipping weight, which works to your advantage.
08:03:21LinusNbtw, i'm a little curious about the communication with the tuner
08:03:52PaulSYeah? Anything I can help with? I haven't stared at that code too much, but now is a good time to ask me to do that.
08:03:59LinusNyour table seems to imply that the GPIO3 and GPIO55 pins are configured as GPIO
08:04:39LinusNbut they are definitely more likely to be configured as I2C pins
08:05:21LinusNor could it be that they wanted to use the internal i2c, but found it easier/better to bit-bang instead?
08:05:34PaulSActually, I've got them listed as "F", function pins.
08:05:50LinusNnot before my latest change in wiki
08:06:01LinusNi changed them to "F"
08:06:02PaulSOh, actually, you did that. B)
08:07:07PaulSUmm... Let me see what the register accesses look like. Definitely GPIO3 confused me as to how they were using it, hence the question mark.
08:10:19 Join DomasoFan [0] (
08:10:46PaulSAlso, the pin is a function pin until programmed otherwise. Since the code is structured so that they perform incremental ors and ands to the function bits, sometimes it's hard to say how the pin is at "steady state". It's concievable, for example, that when I2C isn't active they switch to GPIO mode so they can force their own values.
08:13:07PaulSGPIO3 is never touched, so I concur that it is a function pin.
08:13:17PaulS(Rather FUNCTION_GPIO3)
08:13:53LinusNgpio3 is the i2c clock
08:14:15***Saving seen data "./dancer.seen"
08:16:53[IDC]Dragonamiconn: I got the same file system corruption with the new code
08:22:08PaulSAs for GPIO55, it _looks_ pretty cut and dry that it's set to be GPIO. The entry point calls this init routine that calls this function that says GPIO_FUNCTION |= 0x800000.
08:24:47PaulSOo. This function also sets GPIO_FUNCTION0 |= 0x8. Missed that one. I bet they wanted to leave the option open to use the builtin I2C but they couldn't get it to work right so the punted.
08:25:19 Quit DomasoFan ("CGI:IRC (Ping timeout)")
08:26:54LinusNso my guess was right
08:27:03LinusNthey bitbang the i2c
08:28:34LinusNinteresting, that has happened to a few other projects i've been working in
08:28:51LinusNthey wanted to use the internal controller, but ended up bitbanging instead
08:31:20PaulSOnce you write the code, life turns out kinda okay, if you've got CPU left over. The one that i think hurt them the most is how they bitbang the SPI for the remote. That must have hurt.
08:31:59midkleaving for ~5 days tomorrow, so if i don't catch you, see you later :)
08:34:23LinusNcu midk, have fun
08:34:43midklater LinusN, thanks :)
08:35:00PaulSLater midk.
08:35:25midksee you PaulS.. now, bed :)
08:52:00amiconn[IDC]Dragon: The funny thing with the file system corruption is: I got it with the internal flash first, the Ondio couldn't read it any more. Windows was able to access it, and repairing with Windows yielded a number of lost clusters.
08:52:18amiconnThe Ondio still couldn't access it, so I reformatted.
08:52:48amiconnAfter this first reformat, I couldn't make the corruption happen again for the internal flash!
08:53:22amiconnThis morning I tried the same with my external card, with the same results
08:54:26amiconnWriting seems to work okay when it has to write single blocks (config block and .playlist_control work fine), but chokes on larger block counts (i.e. ROM dump)
08:57:23LinusNso it might be a timing error between writes?
09:00:30[IDC]Dragonamiconn: I'm working on the port inits (made a debug output), but not yet done
09:00:49[IDC]Dragonwhen cold-started, we have many wrong pin functions
09:05:01 Join amiconn_ [0] (
09:10:33 Quit [IDC]Dragon ()
09:15:55 Quit amiconn (Nick collision from services.)
09:15:55 Nick amiconn_ is now known as amiconn (
09:16:45 Join Zagor [242] (
09:28:57LinusNfirst alpha iriver schematics are in the wiki
09:29:46 Join ashridah [0] (
09:32:12PaulSOMG LinusN. That's the hotness!
09:33:10LinusNi'm a bit unsure about the disk led though
09:38:30Zagorradio chip is marked U$1
09:39:41 Join pyros [0] (
09:39:43dwihnoLinusN: Donna Sunner doesn't even have this hot stuff. Great work!
09:41:58LinusNcan anybody see which U-number the fm chip is?
09:42:30LinusNit's really crowded around that little sucker
09:44:46PaulSUnless it's U8, I don't see it.
09:46:33ZagorI don't think it's U8
09:47:49PaulSThe "U8" is parked out by a capacitor. There's nothing in that area that looks like it should belong to U8. It's possible that the guy doing the layout lost track of where the label was.
09:49:36ZagorU8 is the mic, isn't it?
09:52:14PaulSHmm.. It's possible. Personally I wouldn't give a microphone a U-number, but that could be argued as a personal taste thing.
09:53:26Zagori agree, but I saw the line in/our and power connectors have U-numbers too
09:54:04 Join Chronic007 [0] (~Miranda@
09:54:10PaulSWell, some of those have bunches of pins, and in the line-in/line-out cases they're arguably not-completely-passive.
09:55:30Zagorheadphone connector is U5 :)
09:56:19Zagoranyway, I can't find a designation for the radio chip either
09:59:57LinusNi have set it to "U???" for now :-)
10:07:36PaulSIt's bedtime for me. Good work all. Wish me an uneventful JTAGging tomorrow if I get the time.
10:08:25LinusNPaulS: good luck
10:10:05Bagderonce my CD drive fails to rip a CD and I get kernel error messages about ATAPI
10:10:13Bagder... it then stops being able to read from it
10:10:19BagderI have to... REBOOT
10:10:33*Bagder shrugs
10:11:40ZagorBagder: kernel version?
10:12:18ashridahBagder: ide cdrom? tried using hdparm to reset it completely?
10:12:21ZagorI suggest an upgrade before you reboot
10:12:24PaulSI've had that happen in some of the 2.4 kernels with a specific NEC CD-RW drive. Very annoying.
10:12:28ashridahi've foudn that works for all sorts of fun ide cdrom issues
10:12:32PaulSI think the drive itself goes nuts.\
10:13:11Zagori don't think I've ever encountered a failed rip, come to think of it.
10:13:19Bagderyes, it might be the drive itself
10:13:32Zagorwhat drive is it?
10:13:34ashridahPaulS: tried looking for a firmware update?
10:13:35BagderZagor: you should try some of my really old and crappy ones
10:13:39BagderCDs that is ;-)
10:13:48BagderPLEXTOR DVDR PX-708A
10:13:55PaulSI think just maybe I've got around the problem in the past with the right set of eject/inserts and hdparm ata resets, but I might me mistaken. I might have just switch to using the other recorder. (I got these errors with cdrecord).
10:14:17***Saving seen data "./dancer.seen"
10:14:17Bagderdmesg reports numerous ATAPI resets
10:14:28Bagderand they don't seem to make it better
10:14:37PaulSashridah: I haven't looked for a firmware update. That's a good idea. I mostly just stay away from that drive.
10:15:20PaulSAnyway, the pillow bellows.
10:15:39 Part PaulS
10:18:59BagderI better reboot then
10:19:02 Quit Bagder ("Off to search for that connect-resetting peer guy!")
10:29:42 Quit kaouete (
10:29:47 Join kaouete [0] (
10:51:13 Part Chronic007
10:57:13 Join Chronic007 [0] (~Miranda@
11:44:31amiconnZagor: r u there?
11:45:20amiconnMore button woes, even with the classic jukeboxes :(
11:46:44amiconnIf you go into the browser functions available from the menu (plugins/ fonts/ langs/ .wps) and then leave with "left", this is also catched by the menu code afterwards, bringing you one level back (unwanted)
11:47:46amiconnThen there are the button handling issues on the Ondio. Currently, neither the main menu is callable, nor the onplay menu works.
11:49:03Zagorright, I'm planning to redo some code to avoid the trailing release events.
11:50:02Zagorand for the ondio I will simply remove the shift code and call menu and on-play separately (on-play with right+repeat)
11:54:47amiconnIt's probably a good idea for all architectures to use some delay (i.e. button repeat) for more xotic functions instead of button combos. It makes one-hand operation a lot easier
11:54:47ZagorReal Soon Now ;)
11:56:10amiconnAnd on the Ondio it is possible to use the OnOff button for short keypresses (not triggering poweroff).
11:57:00amiconnThe browser<->wps switching and resume comes to mind
11:57:43Zagoryes but the onoff button is not in a very nice position. i think using the option button is better.
12:02:39 Join [IDC]Dragon [0] (
12:04:08 Quit Chronic007 (Read error: 110 (Connection timed out))
12:08:12[IDC]Dragonamiconn: do you read?
12:14:20***Saving seen data "./dancer.seen"
12:14:52 Quit gromit` ("Client exiting")
12:18:00amiconn[IDC]Dragon: yes
12:19:05 Quit pyros (Read error: 54 (Connection reset by peer))
12:19:28 Join pyros [0] (
12:19:41[IDC]Dragonone more addition to the f/s corruption:
12:20:11[IDC]DragonI got the very same thing, looked well under Windows, but refused to boot from there
12:20:35[IDC]Dragonthen I did a quickformat, which fixed it
12:21:38[IDC]DragonI still have a dump of the partition before all these experiments
12:21:57[IDC]Dragonso I could restore that and try if it happens again
12:22:21amiconnDid you check _what_ got corrupted? I didn't have an idea how to do this
12:22:45[IDC]Dragonif it's reproduceable, I could diff the dump
12:23:02 Join gromit` [0] (
12:26:59amiconn[IDC]Dragon: The config sector (both calculation and usage) works well. It is sector 30 (on the internal flash at least). All settingspersist
12:28:25[IDC]DragonI noticed that in case of error your calculation function returns 0.
12:28:45[IDC]Dragondo you catch that from the caller?
12:28:59[IDC]Dragonwe don't want the config in sector 0...
12:29:06amiconnYes, that's intended. I changed the the settings_load() / settings_save accordingly
12:29:55[IDC]Dragonelse it would be catastrophic
12:30:23amiconn..replacing the old check, which checked whether the partition started at sector 0 (floopy-like disk layout without partition table)
12:30:49[IDC]Dragonohh, I liked that
12:31:16[IDC]Dragonso I could test the fat code with a partition dump
12:32:18[IDC]Dragonsorry, ignore me, that's in the test code only
13:06:54 Join MooMaunder [0] (~me@
13:16:13Zagorhas anyone ever seen/written a nice little hack that turns a numberic range into a regex? i.e. 5-12 => ([5-9]|1[0-2])
13:20:36LinusNZagor: why?
13:21:50Zagorstreet numbers...
13:23:13 Join Bagder [0] (
13:28:09 Join R3nTiL [0] (
13:31:00LinusNZagor: aha
13:31:46Bagderseen it, loved it!
13:32:05Bagderzoom zoom zoom
13:33:01gromit`Zagor: /(37.3,200)/; # matches any number x, 37.3 < x < 200
13:34:09gromit`is that what you want ?
13:36:01Zagoryes it is. now can you make it match only odd or even numbers?
13:39:03 Quit [IDC]Dragon ("CGI:IRC (EOF)")
13:54:09Bagderrumours have it RMS was in a car crash today
14:00:07 Quit R3nTiL ()
14:01:45Bagderapparently RMS survived, but at least one died
14:02:28Bagderbut this is "a user on another channel confirmed this" and I read this in another channel ;-)
14:02:41LinusNRichard Stallman
14:10:52 Quit Bagder ("Off to search for that connect-resetting peer guy!")
14:14:23***Saving seen data "./dancer.seen"
15:01:54 Join [IDC]Dragon [0] (
15:04:34 Part LinusN
15:41:08 Join edx [0] (
15:42:01 Join methangas [0] (
15:50:12 Quit ashridah ("sleep.")
15:55:04 Join mecraw [0] (~lmarlow@
16:01:23 Join Sporta [0] (
16:01:28Sportahi guys
16:01:30Zagoryay, microsofts vfat patent has been rejected
16:02:47Sportahey, I'm having a weird problem. All of a sudden, when I boot up my FM Recorder, there's no directory list or files or anything −− it just shows the arrow cursor and the header with the battery, time, etc. then later on, it'll say "Dir buffer full" on the screen... anybody run into this?
16:05:34NibbIerSporta: how many files do u have in your root dir?
16:05:50Sportanot many. I have maybe 15 files and another 20 or 30 folders
16:06:12NibbIerunless thos files/dirs have extremely long names i have no idea, sorry
16:06:45Sportaall right. I did try to put on one of the daily builds the other day, though
16:08:50Sportafor some reason, though, it didn't overwrite the version 2.2 of Rockbox. The only way I could boot the daily build was by selecting ajbrec.ajz and hitting play. is that not right?
16:14:25***Saving seen data "./dancer.seen"
16:19:32 Quit [IDC]Dragon ("CGI:IRC (Ping timeout)")
16:26:19 Quit Sporta ()
16:59:45 Quit elinenbe (" HydraIRC -> <- Get hot chicks here!")
17:02:27 Part Zagor
17:28:54 Join Tim[RIP] [0] (
17:29:09 Part Tim[RIP]
17:48:50 Quit kaouete ("Changing server")
17:48:54 Join kaouete [0] (
17:55:32 Nick gromit` is now known as gromit\party (
17:59:43 Join _aLF [0] (
18:01:16 Join Zagor [0] (
18:11:05 Join dwihno_ [0] (~dw@
18:12:48 Quit MooMaunder (
18:12:48 Quit dwihno (
18:13:20 Quit AciD (Read error: 104 (Connection reset by peer))
18:14:29***Saving seen data "./dancer.seen"
18:30:28 Join AciD [0] (
18:36:44 Quit pike (Read error: 104 (Connection reset by peer))
18:37:31 Join pike [0] (
19:36:12 Join webguest64 [0] (
19:36:24 Part webguest64
20:14:32***Saving seen data "./dancer.seen"
20:46:30 Join adi|blah [0] (
20:47:03 Quit adi|blah (Client Quit)
20:53:33 Join scott666_ [0] (
21:14:45 Join zeekoe [0] (
21:27:18 Quit methangas (Read error: 232 (Connection reset by peer))
21:29:50 Join Bagder [0] (
21:51:04 Quit Zagor ("Client exiting")
21:52:47amiconnThere seems to be at least one bug in the FAT16 support :(
22:10:04 Join Toni [0] (
22:10:32ToniHi guys
22:10:47ToniYou do a great job!
22:11:19ToniDoes anybody know a freeware mp3-transcoder?
22:11:41Bagdertranscode between what?
22:11:53ToniMp3 to Mp3
22:12:24Bagderlame can do it
22:12:38ToniI mean, lame dos this job, but no UI.
22:13:35Bagdertranscoding like that is a terrible idea anyway ;-)
22:14:36***Saving seen data "./dancer.seen"
22:15:22ToniBut I have a big list of mp3 with high bitrate, which I like to shrink to get it fit into my MMC card
22:16:07ToniHow about the idea, I make a Windows-UI and put it to the rockbox page as a goody?
22:16:26amiconnToni: You are the Ondio user with FMR and 1 GB CF, right?
22:16:59Bagder <= it would fit there
22:17:20ToniOK, thanks.
22:17:30amiconnToni: Got my reply mail?
22:17:51ToniNot yet, I look.
22:21:12ToniMAS shows also only zeros
22:22:05amiconnHmm. So it is really strange that you are able to play music. I'm not :(
22:23:07amiconnOkay, maybe I should try it with low bitrate files. Usually my files are vbr around 200 kbps
22:24:40ToniI can hear a small click when firmware is loaded, but I donīt know, wether it comes from original or new.
22:25:37amiconnYou can set "Show files" to "supported" or "all" and then try to RoLo. If you hear the click again, it is produced by rockbox
22:27:13amiconnI don't hear a click :(
22:27:34ToniSorry, I am a newcomer, what is RoLo?
22:27:45 Join [IDC]Dragon [0] (
22:28:20amiconnToni: Oh ok, RoLo means "rockbox loader". You can load a (perhaps different) firmware without switching off & on again
22:28:57amiconnIf you see a firmware file in the file browser (extension .ajz), simply "play" it.
22:29:15amiconn[IDC]Dragon: Bad news: FAT16 seems to have at least one bug :(
22:30:18ToniItīs definitely Rockbox.
22:30:18[IDC]Dragonit went through Zagors' tests all fine
22:30:19amiconnI checked my writing, and found that it always returns success. Then I did the following, which should definitely work if FAT16 wouldn't have a bug:
22:31:09amiconn(1) I made a player build with 2 modifications: (a) compiled with HAVE_FAT16_SUPPORT, (b) modified init() in main.c to only mount partitions with type == 6
22:32:07amiconn(2) The I repartitioned my player: 1st partition ~3 GB, FAT32, containing archos.mod only. 2nd partition ~260 MB FAT16, with archos.mod and .rockbox dir
22:33:11amiconnThen I booted with the modified build. The FAT16 partition gets mounted. If you then try to dump the ROMs -> *PANIC* Writing before...
22:34:32amiconnBtw: it's nice to have a spare box as a developer...
22:36:11[IDC]Dragonhmm, interesting
22:36:29[IDC]Dragon(I don't have a spare box)
22:37:09amiconnThen just tell me what additional tests I could perform
22:37:51amiconn(You mean: you don't have a _working_ spare box ;) )
22:37:56[IDC]DragonI have a look at the code first
22:38:28 Part Toni
22:38:52[IDC]DragonI could need a disk image which is before the problem
22:40:11amiconnDo you mean from the Ondio or my player test? On the player, it didn't corrupt the fs, it just refused to wite (with that panic)
22:42:09amiconnSome interesting info about the MMC driver: I measured the raw read speed, got ~135 KByte/s with the internal flash and ~155 KByte/s with my 256 MB card
22:47:30[IDC]Dragondifferent response time, right?
22:49:25amiconnThe theoretical maximum we could achieve if there would be zero response time is ~350 KByte/s. With response times I expect a maximum of ~250 KByte/s
22:49:50amiconnSo there is room for optimization :)
22:51:53[IDC]Dragonmax is theoretical, or with your code?
22:55:02amiconnThe ~250 KByte/s is derived from the assumption that the read speed scales somewhat linear with the card clock, which is backed by the fact that reading the Ondio (both int & ext card) from the PC (bridge clock 12 MHz) delivers ~1 MByte/s, while reading an MMC with my card reader (most likely using the max possible clock, 20 MHz) delivers ~1.6 MByte/s
23:00:43amiconnAnother interesting fact: Early return from writing the last block is possible, if we don't want to check card status. My first interpretation of the datasheet was incorrect (That's why my first version of write_sectors did only work for single sectors)
23:01:57[IDC]Dragonsaving the last sector, hmm
23:02:07[IDC]DragonI'm very sleepy today
23:02:50[IDC]Dragonport inits didn't like me
23:03:00[IDC]DragonI call it a day, cu
23:03:11 Quit [IDC]Dragon ()
23:16:38 Join gromit`` [0] (
23:30:35amiconn[IDC]Dragon: (for the log). It's definitely the file system and not the MMC driver. I reformatted my ext MMC with FAT32, and ROM dump did work onto that card!
23:37:31 Quit gromit\party (Read error: 110 (Connection timed out))
23:46:03 Quit edx (Read error: 60 (Operation timed out))
23:53:45amiconnSomeone here with a player?
23:55:17 Join webguest61 [0] (
23:56:14 Quit webguest61 (Client Quit)

Previous day | Next day