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 2004-09-16

00:00:43 Quit silencer_ (Remote closed the connection)
00:02:16amiconnSeems I was a bit confused last night.
00:02:53amiconnYes, I confused PA and PB (RTW and EOD).
00:03:10amiconnRockbox used indeed PB14
00:04:21amiconnAnyway, I implemented my modifications into the recording transfer (first version is not asm, but plain C)
00:04:41amiconnCurrently doing a longer test recording
00:05:54amiconnI also implemented (quick hack) displaying the spdif crc status in the recording screen
00:06:34amiconnThe 2 data sheets disagreed on the correct register number, but I found the right one
00:08:02amiconnWith several short test recordings I was able to trigger the spdif synchronization bug. This gives totally distorted recordings, but not more bad frames than usual
00:09:00amiconnHowever, I found another rockbox bug :(
00:09:31amiconnYou can call the recording menu from the recording screen via F1, even when currently recording
00:09:44amiconnIf you go back, the recording is stopped!
00:10:01midkamiconn, actually stopped or just shows the stopped icon?
00:10:50amiconnIt's actually stopped, the recording size displays as 0 bytes. "Play" will start a new recording
00:11:18amiconn(At least this happens when source == spdif)
00:12:01midks/bad/bad Linus.
00:12:07LinusNthe lcd on the iriver is parallel
00:12:25LinusNi have identified d0-d7
00:12:32LinusNand the cs
00:12:57LinusNLA hhokup is next
00:25:33 Join ripnetUK [0] (
00:27:35 Join mecraw_ [0] (~lmarlow@
00:28:13 Join silencer [0] (
00:34:55LinusNthe bdm wiggler is now in Paris, coming closer... :-)
00:35:52bagawkwhats what?
00:35:58ripnetUKwhere does it have to get to??
00:36:05ripnetUKthe wiggler that is :)
00:36:35LinusNit's a debugger interface for the coldfire cpu, and it's on its way to me in stockholm
00:36:54bagawktime to go do school work
00:36:56bagawksee you :)
00:37:09 Quit bagawk ("Leaving")
00:38:10ripnetUKits 2008km by road from Paris to Stockholm :)
00:40:34 Quit zeekoe ("bed time")
00:41:42amiconnLinusN: Why does usb_slave_mode(true) call ata_init() before handing over control to the usb bridge? Thinking about the mmc driver...
00:43:59 Quit mecraw (Read error: 110 (Connection timed out))
00:48:35***Saving seen data "./dancer.seen"
00:49:17 Join webguest93 [0] (
00:50:46 Join PaulS [0] (
00:51:42PaulSIs Dave Hooper around?
00:53:30 Quit webguest93 (Client Quit)
00:54:12LinusNamiconn: i din't remember
00:54:42LinusNit may not even be necessary...
00:54:55amiconnThe only purpose I can imagine is switching on the ata power...
00:54:58LinusNcould be some old obsolete workaround
00:55:49amiconnAlthough I don't know if it may happen that ata is not yet initialized when entering usb mode
00:56:23amiconnIn that case, ata_standby(15) wouldn't work without ata_init()
01:04:15LinusNata_standby(15) should be removed and replaced with changing the powersave mode instead
01:05:06LinusNsince ata_standby attempts to spin down immediately
01:05:15LinusNgotta reboot
01:05:17LinusNcu soon
01:05:20 Part LinusN
01:08:06 Join LinusN [0] (
01:08:32PaulSI'm having all sorts of troubles getting my attachment in IriverPort up... It claims to have attached the file but browsing to it returns "Internal Error" from the server.
01:08:33amiconnhi again
01:09:15PaulS(Oo.. Hi LinusN! How's the River hangin'? :-)
01:10:13LinusNPaulS: i think the web server is confused, and tries to run the script as a cgi
01:10:22LinusNcool exploit!
01:10:37LinusNchange the extension
01:11:41PaulSOh, BTW, LinusN, you need an A0 pin on that LCD display pinout, basically the Command/Data switch.
01:14:30PaulSShweet.. Now I'll fixeup the links
01:16:49 Join webguest00 [0] (
01:17:00 Quit _aLF (".")
01:17:53 Quit webguest00 (Client Quit)
01:22:43 Quit Headie (Read error: 54 (Connection reset by peer))
01:23:14 Join Headie [0] (
01:23:39LinusNPaulS: i know, i just don't know which yet
01:25:04 Quit mecraw_ ("Trillian (")
01:28:17PaulS(More about A0) If you're using a scope to look at it, you'll find that the A0 signal goes active (high) for multple CS cycles during long data bursts. least that's my assessment from looking at the code.
01:31:49PaulSThe CS is driven by the ColdFire internals, and the A0 is driven by GPIO, and they just leave A0 high during the course of a data burst.
01:37:58LinusNhmmm, tracing with a logic analyzer, and it doesn't make any sense...
01:46:29LinusNanyway, i gotta sleep now, cu around
01:46:36midknite LinusN
01:46:41 Part LinusN
01:48:12PaulSAwww... He left, and I had more suggestions.. Aie well.
01:48:26midksuggest them now, there are logs
01:50:44PaulSWell, if he's using a analyzer, he should just trigger on one of the control words like 0xB1, and follow the trace forward. The B1 byte should have A0 inactive. The next byte should have A0 active, and the next byte will be a 0x13 with A0 inactive again.'
01:52:30PaulSFrom a software point of view, though, I'm not sure this is very important, since we already know where (which GPIO pin, what register address) A0 is and what it does.
02:06:27 Join _Headie [0] (
02:06:28 Quit Headie (Read error: 54 (Connection reset by peer))
02:09:55PaulSAnyways, I'm off as well. Later, folks.
02:11:39 Quit PaulS ("CGI:IRC")
02:27:49 Join Headie [0] (
02:27:49 Quit _Headie (Read error: 54 (Connection reset by peer))
02:41:53 Join bagawk [0] (
02:45:50 Quit amiconn (" pow")
02:48:37***Saving seen data "./dancer.seen"
03:25:24 Quit gromit` (Remote closed the connection)
03:38:06 Join gromit`` [0] (
03:38:06 Quit bagawk (Read error: 104 (Connection reset by peer))
03:41:46 Join BC [0] (
04:05:15 Part BC
04:48:38***Saving seen data "./dancer.seen"
05:38:19 Quit MrMoo (Read error: 110 (Connection timed out))
05:51:58 Join maikeul [0] (
05:58:14dwihnobaguette avec saucisson et fromage pour moi, s'il vous plait
06:02:14 Quit gromit`` (Read error: 110 (Connection timed out))
06:12:39 Quit scott666_ ("i'll be back...eventually...")
06:13:18 Join flabat [0] (
06:13:44flabati just want to say : thank you!
06:14:36flabatany developers online?
06:15:57midkme.. somewhat
06:16:50flabati just installed rockbox in my studio20, upgrade the hdd to 80gb
06:17:02flabatthis is awesome software
06:19:40dwihno80gb? nice
06:20:24dwihnodoes the studio have a 2 line charcell or a lcd?
06:20:29dwihnoI never got the hang of that
06:21:38dwihnothen flabat is missing the lcd ride :(
06:23:07midk:\ yeah
06:23:17dwihnothe best improvement, imho
06:23:31flabat2 lcd lines
06:23:58flabatwhats the lcd ride?
06:24:27midkbitmap lcd: ~8 lines, depending on font you're using
06:24:31midkgraphics, etc
06:24:38midkjpeg viewer, video player
06:24:53flabatno, just text on the studio 20
06:25:06midkthe bitmap lcd on the recorders does the above
06:25:09flabatno money for that
06:25:23dwihnobut you have money for a 80 gbyte disk ;)
06:25:32midki recall they were about the same price
06:25:49flabatthis was a unit with bad hard drive and the hard drive i got it for free
06:25:54dwihnoah, okay
06:25:59dwihnofreebies for all! :)
06:26:03midkcool cool :)
06:26:04flabatso is a 80gb mp3 player for free
06:27:28midkboth the player and hard drive were free
06:27:46midkcertainly not bad
06:28:56flabatand with the firmware update even better, the original firmware is really bad
06:29:37dwihnoflashing the firmware is also great
06:29:51dwihnodecreasing from a 15 sec boot time to ~1 sec
06:30:00flabatis not dangerous
06:30:17dwihnoShould not be if you know what you are doing
06:30:49dwihnoAlso, flashing the so-called rombox frees an additional ~160 kbyte for the mpeg buffer
06:31:03flabatmmm, i can wait 15 secs i guess
06:31:40flabatso whats the benefits with thw buffer
06:32:09dwihnoI've read something about a 4% runtime increase
06:32:49flabatwith the old player also
06:32:52midkcan't do rombox/flashing on player
06:32:58dwihnomidk: oh.
06:33:05dwihnomidk: Well, didn't know that :)
06:33:11dwihnomidk: Another reason for getting the recorder ;)
06:33:37midkdwihno, high five :]
06:33:40*midk high fives
06:33:47*dwihno high fives
06:34:09midk o
06:34:11midk / \
06:34:40dwihnoSo, what do you think about rockbox on ihp?
06:34:49dwihnowill it see the light of day?
06:34:54midksounds very interesting actually
06:35:02midkhmm. it will take a while but i think they can do it
06:36:37dwihnoI will never doubt any project with the team members Zag/Bag/Lin
06:37:10midksame reasoning here :)
06:37:27midki definitely don't doubt that if they want to they could do it.. and i think it would be very cool to have that running
06:37:39midkinstead of it dying out with the old archos series it could live on in the newer ihp models
06:38:13dwihnoA successful port on the ihp will open the doors, I think
06:38:53dwihnoand perhaps make the iriver guys uncomfortable
06:39:13dwihnoI guess it's not all sugar and cake
06:39:22dwihnoit's salt and vinegar too
06:39:27midkthey're cool imo.. they replace units basically no matter how they broke down
06:39:29flabatok guys, thanks again and good night
06:39:40midktheir firmware/support isn't bad compared to most companies
06:39:49 Part flabat
06:41:10dwihnoIt's a tough market, so I guess their high prices perhaps increases the support and software divisions of the company
06:41:36midkthey are relatively high priced, true
06:41:44midkbut nice hardware/abilities make up for it
06:42:00midkfm tuner, 320kbps cbr recording, optical in/out, nice large lcd, etc
06:42:04dwihnothey use 2.5" or 1.8"? Can't remember
06:42:27dwihnoDid you see the screenshot Zagor showed yesterday?
06:42:28midk1.8 i believe
06:42:32midkyeah, of the sim?
06:42:34midkthat was really cool
06:42:38midkand my clock was in it of course :)
06:42:44dwihnoit was?
06:42:45midkadvertisement is good
06:42:48dwihnoDidn't see that :)
06:42:49midkthe filename at least
06:43:19dwihnoUsing the prop font, most titles will fit the screen
06:44:35midkyeah, that is really cool
06:45:54dwihnoMy first archos' disk died a couple of weeks ago
06:46:10dwihnoDo you think rapid cold and heat changes can damage the disk?
06:46:16midk:\ sorry about that. hopefully it's backed up?
06:46:22midkit's possible, but doubtful...
06:46:28midkif you keep it cool you may be able to boot it
06:46:49dwihnoI mean, swedish winters can be quite cold
06:46:53dwihno(which sucks, btw)
06:47:04midki bet
06:47:17plokiRiver has a 1.8 drive
06:47:23midkso.. it was in the snow then you grabbed it and put it in the oven to dethaw it? :)
06:47:32dwihnohaha :)
06:47:39dwihnothanks plok
06:47:56dwihnoI always keep backups
06:48:06dwihnosince I had a disk failure a couple of years ago
06:48:09midkgood boy.
06:48:17dwihnoLost a ton of pre-2k sourcecode
06:48:20midkyeah, i had a disk failure. no, i don't back up.
06:48:40***Saving seen data "./dancer.seen"
06:48:50dwihnoIncluding a rather nice VESA2 engine (first time I ever read docs to accomplish anything) :)
06:49:16midkhmm. well, that sucks, but i guess you learned a lesson
06:49:30midkand just remember, it was destined to happen.
06:49:40midkwell, no, if you would have backed up it would have been avoided. ;)
06:50:20dwihnoThe great giant fork in the sky has already foreseen my future :)
06:50:37midkyou sure that's not a spork? ;p
06:53:46dwihnochocolate candy
06:54:07midkyum, chocolate covered raisins.
06:54:29dwihnoyuck :P
06:54:35dwihnothose taste like...
06:54:40dwihnochocolate covered raisins :P
06:54:43midkhaha. :)
06:59:11 Join LinusN [0] (
06:59:20dwihnoWelcome back!
06:59:31dwihnoWe were busy praising the project ;D
07:01:05midkhaha! yeah. :)
07:01:24midki recall something about a spork too.
07:01:35dwihnoit was a FORK goddamnit! :)
07:02:01dwihnoI actually wrote an ultra light encryption thingie which I named "sporkel"
07:02:02midkspork? fork? fillet knife? what's the difference.
07:02:28midkwhen i was trying to save tetris high score to disk i had to have an encrypted number to make sure it was unmodified..
07:02:44midkthe algorithm was something like "level*rows/blocks"
07:02:56midki just sort of scrapped the idea.
07:03:07dwihnodoesn't really matter
07:03:14dwihnopeople who cheat sucks anyhow :)
07:03:23midkyeah, that's it
07:04:13dwihnoIt would be really swell if I could make windows start my processes a bit different
07:04:30midkhm? how?
07:04:39dwihnodisk i/o intense applications would've been given lower priority
07:04:50midkbootvis, if it works, is great.
07:05:02midki used it once a while back.. it shaved ~15s off my booting time
07:05:24midkit analyzes which drivers are loaded when, how long etc, then rearranges the order and stuff to make it quicker
07:05:32midkyeah.. google for it, ms took it off their site
07:05:33dwihnoaaah... cool stuff!
07:06:17dwihnowonder why
07:06:23dwihnoxp still boots quite fast
07:06:53midkyeah, it just went faster for me :)
07:06:59midkit overlaps some and stuff
07:07:03midkloading earlier
07:07:05midknot really sure.
07:10:09dwihnoI read something about being able to replace some files on win2k with the winxp equivalents to increase boot time
07:11:10midkwiat, to increase?
07:11:59dwihnodecrease ;)
07:12:10dwihno"I think my computer runs too fast! I need to slow it down!" :)
07:12:22dwihnoThat reminds me of the turbo buttons on those old 486 machines
07:12:32midki never understood the purpose.
07:15:54LinusNeasy, many programs couldn't handle the extra speed, like Tetris for example
07:16:10LinusNso you had an option to run at a compatible speed
07:16:26dwihnowhat's with your tetris obsession, midk? :)
07:16:35midktetris obsession?!
07:16:38midkTETRIS obsession?~
07:16:55LinusNever tried to run the original PC version of tetris on a modern pc?
07:17:14midki imagine it would occur such like that as... er... no.
07:17:33LinusNthe timing is done with delay loops
07:17:47LinusNso it runs faster on faster pc:s
07:18:42dwihnoYay! Firefox 0.10 is out!
07:18:50dwihnoI like Firefox and Thunderbirds.
07:19:17midk1.0 you mean.
07:19:38midkand it's a preview release, not official 1.0,
07:19:43midkand it's been out for about 2 days. :)
07:22:06dwihnoWell, I'm living under a rock, okay?! :)
07:24:32midkyeah, i know.
07:25:35 Join AciD [0] (
07:26:31dwihnotime to test this stuff
07:26:44midkWHICH STUFF?!
07:30:09dwihnoJust a plain reboot
07:30:47midkhm, ok.
07:30:55midkpermission granted.
07:31:49midkbedtime anyways, nite
08:41:56 Quit Hadaka (Read error: 60 (Operation timed out))
08:48:41***Saving seen data "./dancer.seen"
08:54:14 Join [av]bani [0] (
08:55:40 Part LePoulpe303
08:57:35[av]banino dead ihp's yet?
08:59:57[av]banire: rng
09:00:18[av]baniyou can collect entropy from user button presses and ide interrupt timing
09:00:27[av]banisave it when you shutdown and restore when you boot back up
09:00:32 Join Zagor [242] (
09:03:50[av]baniany idea wtf that whine is about?
09:07:56Zagorthey got sour when we quoted information from their forum in our wiki
09:07:56Zagorwe gave proper credit and references, but they still didn't like it
09:07:57Zagorso I removed the information again
09:08:04 Join Fusen [0] (~Fusen@fusen.user)
09:08:08Zagorbut that wasn't enough either!
09:08:18Zagorso I gave up and ignored them
09:08:54 Part Fusen
09:10:50LinusN[av]bani: we have no rng entropy problems in rockbox
09:21:38[av]banire: an ihp to mangle
09:22:00[av]banimaybe start a donation pool colecting funds for a sacrificial ihp-120
09:22:15[av]banianyone who donates >$10 wil get early access to rockbox beta testing
09:22:21LinusNwe can use the regular donation fund
09:22:25Zagorwe have a donation fund already. it just feels bad to buy a brand new box and destroy it...
09:22:59[av]baniwell, did you have to destroy any archos'en ?
09:23:13Zagorno, since they didn't use bga
09:24:37Zagorthat was fast :)
09:24:40[av]baniah, did bdm wiggler come out of the fund?
09:25:13[av]banithat works :)
09:25:39 Join webguest46 [0] (
09:26:14 Join Naked [0] (
09:26:24 Nick Naked is now known as Hadaka (
09:27:43 Quit webguest46 (Client Quit)
09:29:23[av]baniah so lcd is parallel after all
09:29:25[av]banithat is good :)
09:29:56LinusNi hooked up my logic analyzer yesterday, but didn't get any wiser, will work some more on it
09:30:14 Join Bagder [0] (
09:30:39[av]banii imagine many lcds even from different vendors share similar command sets
09:31:07[av]banisince companies may want to multiple source parts
09:33:04 Join pillo [0] (
09:33:54 Join PaulS [0] (
09:35:33PaulSHullo guys.
09:35:38[av]baniWhen I hit track-forward it sends any of a few strings: "1134-1", "1279", "134-2A", "13589". None of the strings are terminated by whitespace/cr/nl. I haven't yet figured out the rhyme or reason for each, although I think one of these means "went to next album". At startup I got a "674".
09:35:54[av]baniif they code like me, each digit probably indicates a function/routine
09:36:20Zagor[av]bani: are you talking about the remote?
09:36:32[av]banino, thats re: the rs232 output
09:36:41Zagoron the rx/tx pads?
09:37:01Zagorok. free debug output :)
09:37:08PaulS[av]bani −− that's correct. I've seen where in the code that lives, and I know where the "putc" routine is.
09:37:10[av]baniyou havent kept up with the wiki?
09:37:21Zagorapparently not :)
09:37:30[av]banipauls, my guesses have been good so far :)
09:37:57[av]bani1 is probably "main"
09:38:01[av]banithe rest, shrug
09:38:22[av]baniA maybe ata :)
09:40:24LinusNmaybe it's similar to rockbox's panic error codes
09:40:33PaulS"1 is probably 'main'" −− not quite. It's debugging different spots in the track access routines. The numbers are pretty much linear 0-9 in the code, and depending on the path through the branches different digits get triggered.
09:41:45[av]baniprobably something which gave them a lot of trouble
09:41:52[av]baniso they put in debug for it
09:41:58[av]banithough they never took it out, interesting
09:45:09plokIs there anything wrong with, or is it just my web proxy?
09:45:21[av]banijust you
09:45:34PaulSIn the 1.40 image, putc is 0x2CBE0. The byte pushed on the stack before the JSR is the ASCII byte that'll spit out the TX.
09:47:18PaulSThe site has been slow for me, but I don't have a lot of context.
09:48:27LinusNit should be faster now
09:51:41PaulSThe other thing they failed to take out was all the calls to debug prints. They ifdeffed out the actual content of the debug print, but the strings are there and the call to the prints are there. There are different kinds of debug print functions −− each takes a string arg, but some act like printf("%s %d"), while others are string-only.
09:53:17[av]banisounds like someone forgot to remove -DDEBUG :)
09:53:26[av]banior theyre too noobish to have #ifdef'd it
09:53:52PaulSDon't complain. :-)
09:54:06[av]banitbh the iriver firmware is pretty bad, i suspect most of their code is purchased 3rd party that they threw together
09:54:25[av]baniwell i complain when _I_ know i do better
09:54:42dwihno[av]bani: from what I've heard, the iriver firmware is supposed to be quiote good
09:55:07plokhmm, I just tried building the cross compiler GCC again (with −−enable-languages=c only!) and I'm getting the same error again
09:55:26PaulSBut... but... those debug strings are going to tell me how the whole thing works!
09:55:26plokdwihno: The iRiver firmware could be a LOT better, so easily!
09:55:33Zagorplok: you need to use the same −−prefix for binutils and gcc
09:55:54plokZagor I'm sure that I did. I get this error
09:56:07dwihnoIf there was rockbox for iriver, I would have to buy one just for the cause of it :)
09:56:32plokinsn-emit.c:8549: Internal compiler error in simplify_subreg, at simplify-rtx.c:2452
09:56:48Zagorwow. I haven't seen that problem before.
09:57:01Zagorwhat version is your native compiler?
09:57:08plokgcc version 3.2 20020903 (Red Hat Linux 8.0 3.2-7)
09:57:15Zagorshould be fine
09:57:31dwihnohave red hat stopped their custom gcc patching?
09:57:46Zagorno idea. don't they patch everything?
09:57:50 Join MrMoo [0] (~me@
09:58:00plokI'll blow it all away and start from scratch again just to make sure
09:58:00*Zagor uses debian
09:58:08[av]baniredhat hacks everything
09:58:22[av]banitbh what they patch was generally broken to begin with though
09:58:37[av]baniso its generally not worse :)
09:59:20BagderI disagree with that description
09:59:26dwihnoZagor: probably... I guess you should use one of the "vanilla" distrubitions... (are there any left?)
09:59:31Bagderbut nevermind
10:00:38[av]baniis the closest
10:00:52Bagderdebian is the coolest ;-)
10:01:04*Bagder joins the distro wat
10:02:20plokMaybe I should try building on windows instead...
10:03:28pilloI use gentoo too ;)
10:03:43pillohi all, sorry to say hello this way but couldn't resist ^_^
10:04:22Zagorplok: write a doc if you succeed ;)
10:04:28 Part [av]bani
10:04:46pilloI'm watching this channel these days as I'm interested in the iriver port too, and hope to start contributing soon.
10:05:37 Quit ripnetUK ()
10:05:39 Join webguest72 [0] (
10:06:22plokZagor: Sure :) Before or after I manage to build it with a c++ compiler?
10:06:42plokI've rebuilt binutils from scratch and am rebuilding gcc now
10:08:38LinusNplok: don't bother, we don't need a c++ compiler
10:10:47plokI was joking, I got the messgae yesterday :)
10:11:21*dwihno uses slackware
10:11:29dwihnoUntil I get my 250 gig disk working under FreeBSD ;D
10:11:47dwihnoIt's some kind of fat32 file system issue
10:12:28 Quit Bagder ("Leaving")
10:12:43plokgrr... same error again :(
10:14:11Zagormaybe we should tar up our build
10:15:12plokI have another linux box with an earlier version of redhat. I could try that one first
10:15:20LinusNdwihno: 250gig disk, are you sure that your ATA controller can handle that?
10:15:27plokAll we have at work are redhat machines
10:16:02Zagorthe rockbox build server is currently an old redhat too
10:16:16 Quit webguest72 ("CGI:IRC (EOF)")
10:17:09Zagor (22MB)
10:18:07 Quit PaulS ("CGI:IRC (EOF)")
10:18:39Zagoruh, bad idea to restart apache right now. sorry :)
10:19:06plokI was going to say, I only got 19k?!
10:28:18plokThis other machine is so slow :(
10:42:08dwihnoLinusN: I connect it using USB...
10:42:36dwihnoLinusN: Works flawlessly using Wintendo and Linux... The fat32 driver for FreeBSD has some limits.
10:42:57Zagorboo :)
10:43:09dwihnoPerhaps you should patch it, Zagor? ;)
10:43:19dwihnomy co-worker brought senap!
10:43:27dwihnoGuess who's gonna have ärtsoppa deluxe for lunch?
10:48:42***Saving seen data "./dancer.seen"
10:53:29 Part pillo
10:54:58plokCool, I have the archive. Thanks Zagor!
11:36:18 Join ripnetUK [0] (
11:37:39 Join amiconn [0] (
11:45:46HeadieÄrtsoppa and pannkaka.. Yummie
11:57:29 Join Lynx_ [0] (lynx@
12:11:30 Join ashridah [0] (
12:23:44 Join Bagder [0] (
12:32:20BagderZagor: is brokeb
12:32:38Bagderand the source one too
12:32:47BagderI bet you've removed my symlinks
12:33:04Bagderor something
12:35:23Bagderthe script checks in the dir named as the 'bin' parts says
12:35:48Zagorhmm, those dirs don't exist
12:35:55BagderI hade them as symlinks
12:36:05Bagderln -s . install
12:36:09Zagori haven't removed them afaik
12:36:25Bagderperhaps the auto-remover removes them...
12:36:29BagderI'll check
12:37:21Bagderoh yes it does!
12:37:24Bagderhow silly
12:38:16Bagderbtw, should we keep files longer now when they're no longer fill the daily-build page?
12:38:45Zagorwhy not
12:38:54Zagora month perhaps
12:39:02Bagderok, increasing to 30 days
12:40:13Bagdercurrently the files use 29MB for 6 days
12:41:34Zagorcurrently we have 7.5G free :)
12:41:46Bagderand the auto-cvs-build dir takes 110MB
12:42:14Bagderroom to grow indeed! ;-)
12:48:44***Saving seen data "./dancer.seen"
12:55:08 Join dwihno_ [0] (~dw@
12:55:08 Quit dwihno (Read error: 104 (Connection reset by peer))
13:22:04 Quit AciD (Read error: 60 (Operation timed out))
13:25:23 Join [IDC]Dragon [0] (
13:28:49amiconnhi Jörg!
13:30:21amiconn[IDC]Dragon: Yesterday I digged a bit into the beginnings of an mmc driver. However, I have some questions you could help me to solve.
13:31:52[IDC]Dragonhi jens
13:32:01[IDC]Dragonstandby, I'm busy
13:33:15amiconnAnyway, here are the questions. No need to hurry though.
13:34:03amiconn(1) When not in usb mode, does the SCK1 output continuously output the clock (~1.5 MHz) or bursts?
13:34:27amiconn(2) I suspect that there is actually something connected to PA13. Could you check this?
13:42:14 Quit ashridah ("Client exiting")
14:03:02 Join ashridah [0] (
14:10:40 Join R3nTiL [0] (
14:23:37 Join lImbus [0] (
14:31:07[IDC]DragonJens: now is better
14:31:38[IDC]Dragon(1) The Archos firmware generates burst, longer than 1 byte
14:32:24[IDC]Dragon(2) I don't know about PA13, haven't found a connection yet. In which context is this used?
14:33:20[IDC]DragonBagder: I've pushed some OndioFM pictures in your upload area, since my email wasn't working this morning.
14:36:45[IDC]Dragonconveniently made them the size used on your page
14:36:52BagderI noticed
14:37:20[IDC]Dragonexcept for the big one, of course
14:37:22BagderI'll make use of them when we start providing daily builds
14:37:35[IDC]Dragonyes, no hurry
14:38:14Bagderthere here now:
14:38:43[IDC]Dragonmaybe rename them to OndioFM_xxx
14:38:51[IDC]DragonI forgot to do so
14:39:02[IDC]Dragonor whatever the naming convention is
14:39:05Bagderthat whole dir is only temporary
14:39:19BagderI'll rename and fix when the time comes
14:39:32 Join JoeBorn [0] (
14:39:42[IDC]Dragonoh, they seem to be progressive jpeg
14:39:48[IDC]Dragondid I do that?
14:40:37Bagderit doesn't matter
14:44:23 Join webguest88 [0] (
14:48:48***Saving seen data "./dancer.seen"
14:50:07[IDC]Dragonamiconn: have to reboot, see you
14:50:16[IDC]Dragon(and see above)
14:50:22 Part [IDC]Dragon
14:51:13 Quit Bagder ("Leaving")
14:51:57JoeBornhello everyone, does anyone mind talking about hardward generally for a bit?
14:52:12Zagornot a bit
14:52:42JoeBornwell, I guess as most of you probably know, we've made some mistakes in some of our choices along the way :(
14:53:03webguest88is it possible that I use my JukeBox 20 with the Jukebox 5000-6000 software?
14:53:12JoeBornin going forward, I'm just looking to make sure that we don't keep repeating those
14:53:57Zagorwebguest88: yes
14:54:21webguest88okay, thanks
14:54:43ZagorJoeBorn: we'd like that too :)
14:55:15JoeBornI guess I'll start with a couple questions. It's obvious that the ARM cores have the Gcc tools and the benefit of good support generally
14:55:26Zagor(for those who don't know, JoeBorn is talking about the Neuros player)
14:56:02JoeBornbut it's also obvious that the DSP side is of great interest to developers and it seems that these have often been less open,
14:56:19JoeBornwhat's your take on availability of tools for the DSPs?
14:56:22Zagoryes, that has been a problem
14:56:40JoeBornwhat processor does the iRiver use?
14:56:43Zagorwell gcc supports some dsps, not sure exactly which though
14:56:46Zagorcoldfire 5249
14:57:05Zagorwith an EMAC, so not a full DSP but still enough for fast signal processing
14:57:43JoeBornand those have complete support tool-wise?
14:58:01LinusNnot that i know of
14:58:13Zagorthe EMAC needs to be assembly-programmed, but being very restricted (just a few instructions) it's not a problem
14:58:47Zagorthe coldfire core itself is quite well supported though, since it builds on the m68k core
14:59:39JoeBornso have you guys researched processors much or you pretty much just looked at the handful of players out there and then looked at the processors for those?
14:59:57Zagorthe latter
15:00:14Zagoralthough we work with stuff like this for a living, so we know a lot of different cpus
15:00:20JoeBornhave you looked at any of the TI products?
15:00:27LinusNwe have of course worked on quite a few cpu's, but not in the scope of the rockbox project
15:00:42LinusNno, not ti
15:00:51Zagoryes, when I was looking at the archos multimedia players. TI is notorious about hiding documentation though so it quickly gets boring.
15:01:37JoeBornTI is appealing for us, because we have so much code written for it already.
15:02:03ZagorI can understand that
15:02:13JoeBornFM stereo algorithms, etc, etc.
15:02:20Zagorbut their attitude to non-commercial development is rather hostile
15:02:33LinusNi'd stay away from the ti stuff
15:02:36JoeBornso how much of an impediment is that "hostility"?
15:02:48LinusNthe lack of docs is a real problem
15:02:51Zagora rather big one. we simply cannot get docs.
15:02:52JoeBornwhere does that hostility stem from? what form?
15:03:14JoeBornoh, and the docs are all protected so we can't distribute them either, I assume?
15:03:31Zagorif I ask for docs, I get the response "how many chips are you planning to buy?"
15:03:34Zagoryes, generally
15:03:50JoeBornfunny you say that, we get that too.
15:04:12Zagoryeah, but you can answer :)
15:04:25LinusNthat's the first signal that tells you to stay away
15:04:27JoeBornwhat about the gcc project to port to TI
15:04:37LinusNdon't hold your breath
15:04:55Zagori've only seen the sourceforge project page, but no content
15:04:55JoeBornyou think it's just too difficult and big a project?
15:05:14JoeBornright, it clearly needs a push.
15:05:19LinusNi think the TI secrecy will be a problem for that project as well
15:06:03JoeBornhave any silicon manufacturers been particuliarly supportive?
15:06:06LinusNas i see it, they don't deserve the gcc port
15:06:13JoeBornor open?
15:06:40Zagormotorola is pretty open in that all their docs are available on their website
15:06:49Zagorthat's all we ask
15:07:10LinusNand one should also choose a common architecture
15:07:13Zagori think others do that too, I just can't think of any names right now
15:07:30JoeBornwhat do you mean "common architecture?"
15:07:39LinusN68k, ppc, arm etc are very popular
15:07:48JoeBornoh, right.
15:07:53LinusNso there are good tools for those
15:08:21Zagoryes. the more popular a platform is, the better the tools are
15:09:04JoeBornso in a nutshell, on the DSP side, although there may not be an available compiler, you'd probably use assembly anyway.
15:09:14LinusNand if you want to benefit from opening the source, virtually everyone should be able to compile and link the firmware
15:09:30Zagorit depends. for the coldfire we use DSP since it's only a few instructions.
15:09:46Zagorbut for a case like yours where the entire system runs on a DSP I'd want C
15:09:56 Quit midk (Read error: 54 (Connection reset by peer))
15:09:59JoeBornsure, that makes sense
15:10:06ZagorLinusN: right
15:10:13 Join midk [0] (
15:11:06JoeBornbut looking at some of your discussions, it seems clear that you were facing resource issues with the Archos and that's one of the big reasons for the move?
15:11:47JoeBornobviously some of these issues would still be there with a processor like we're currently using, perhaps not quite as severe, but still there.
15:11:58Zagorwell the archos is an old platform that is becoming obsolete. we wanted to follow the times and not stay lock in to a dying product.
15:12:02LinusNwell, they don't manufacture the archor jukebox anymore
15:12:28ashridahZagor: you do realise they don't manufacture the 1xx series anymore either, iirc.
15:12:30LinusNand it's nice to have the codec in s/w
15:12:41ashridah(iriver H1xx series i mean)
15:12:43Zagorashridah: really? interesting...
15:12:49LinusNashridah: :-)
15:12:53ashridahZagor: yeah, they're pretty much pushing the 3xx
15:12:59LinusNof course
15:13:40Zagorwell the h100 isn't exactly hard to get yet anyway. the archos is.
15:14:09Zagorbut sure, we won't be able to stay with the h100 forever either
15:14:42LinusNand it shouldn't be *that* hard to port to the 300, afaics
15:14:57*ashridah nods
15:14:57JoeBorn3xx uses the same 5249?
15:14:59Zagorif joe makes a good platform we can use, at least we'd know when it stopped being manufactured ;)
15:15:27ZagorJoeBorn: as far as I know, yes
15:15:28LinusNi believe the cpu is the same, yes
15:16:02JoeBornso what about the rest of the hardware. I'm absolutely amazed that you guys can do so much without support from teh mfg.
15:16:19LinusNlots of experience
15:16:23JoeBornyou might have specs on the CPLD, but no information on the code, etc
15:16:36JoeBornhow much of an impediment is that?
15:16:39Zagorwell one thing we've found useful is the separate usb/ata bridge. it means we don't have to write and debug a complex usb-storage driver.
15:16:52LinusNwell, all hardware must have public data sheets
15:16:56Zagorit also means the disk always works, even if we fubar the firmware
15:16:58 Quit webguest88 ("CGI:IRC (EOF)")
15:17:02Zagor(nearly always... ;)
15:17:58Zagormake sure you line up the MSBs and LSBs on your serial connections... :) (internal joke, archos players needs a lot of byte swapping)
15:18:40Zagorparallell connection to the display, preferrably, for fast access
15:19:10JoeBornI'm not really a hardware engineer, I just play one on TV, inside Hw jokes will surely go over my head :(
15:19:47JoeBornfast access for video? animation?
15:20:21LinusNwell, a fast lcd connection leaves more power for the codec
15:20:33JoeBornoh! I get it
15:20:36LinusNsame goes for all peripherals
15:21:13LinusNconnecting the hard drive correctly, so you can use the internal DMA in the cpu (if any)
15:22:09JoeBornno those are good tips.
15:22:31LinusNthere are so many small details that can make a huge difference
15:22:40JoeBorn so, it's clear that you guys have been able to figure out most everything without manufacturer support, so besides just designing good HW, what else do you need from us?
15:23:01JoeBorn(us meaning the manufacturer generally)
15:23:27LinusNa dialog, and hardware to develop on
15:23:31 Join AciD [0] (
15:24:10JoeBornwhat about hardware tools, emulators and such?
15:24:39LinusNif you choose your platform wisely, you may not need emulators
15:25:06LinusNonly a cheap BDM/JTAG wiggler
15:25:57LinusNand the gnu debugger
15:26:53JoeBornwhat platforms support that?
15:27:24LinusNat least arm and coldfire
15:27:27JoeBornwhat about the mp3 codec? Do you have a source for that?
15:27:44LinusNlibmad is gpl
15:27:59LinusNwe will probably use it for the iriver
15:28:14 Join Lynx [0] (lynx@
15:28:33LinusNi don't know much about free realtime encoders though
15:28:56LinusNwe will probably have to do some "inhouse" developing for that
15:29:08JoeBornit seems like writing one is quiet an effort.
15:29:13Zagornot the entire codec, just optimising an existing
15:29:56JoeBornwhen we asked monty about writing an Ogg encoder for our machine, he said he heard it took TI 3 man years to write their mp3 encoder
15:30:13Zagori can imagine
15:30:47Zagorbtw can I ask if the ogg codec he wrote for you is (or will be) released?
15:31:26JoeBornI just haven't looked the contract, but I'm 99% sure from my recollection that it's fine.
15:31:45JoeBornxiph wrote our codec for us.
15:32:07Zagori guessed that :)
15:32:31JoeBornbut I'm 99% sure its actually just a BSD license and there's no problem.
15:33:04Zagoryeah but it depends on the terms of the contract, as you say. with bsd you can choose to make the result proprietary.
15:33:26JoeBornso Zagor, what do you think you need from the manufacturer?
15:33:41JoeBornwhat are the impediments that you see?
15:33:43LinusNloadza cash!
15:34:04Zagordocs, mostly. and as linus said, a dialog with someone we can ask about stuff we can't figure out
15:34:11 Join R3nTiL1 [0] (
15:34:16Zagorsome development hardware, of course
15:34:17JoeBornwell, you know we are a for profit company (at least trying to be)
15:34:52LinusNwhen i said "dialog", i also meant regarding the hardware design
15:35:00JoeBornbut honestly supporting open source with some hardware is obviously a lot cheaper than hiring internal developers
15:35:43JoeBornwell, this is that start of that dialog.
15:36:09ZagorI just need to raise a warning flag here: you cannot depend 100% on us writing the firmware for you for free. we may not do it in the speed you want it, or with the features you'd like.
15:36:25JoeBornwhat about installed base? I know that's an issue for you guys, and would be a disadvantage of anything we do, at least for the time being
15:36:46LinusNrelying on programmers that do it for free on their spare time isn't all that good
15:37:15JoeBornof course, what I'm imagining is doing our development with internal developers in an open environment
15:37:38ZagorJoeBorn: ok, good. just wanted to avoid any misunderstandings
15:37:39JoeBornwith support from outsiders as a bonus
15:37:46LinusNsounds good
15:38:12JoeBornwe've seen that in our own experience those issues.
15:38:22LinusNbasically, make the hardware like we want it, and we port rockbox to it -)
15:38:24JoeBornit's impossible to control.
15:38:38Zagorinstalled base is obviously an issue, since we as a project are always looking for contributors and testers. if only a dozen people have the hardware it's not going to be as much fun.
15:39:15LinusNon the other hand, if the hw is sexy, and the publicity is right, it could very well be a hit
15:39:26JoeBornwell, it won't be a dozen people (or we've got some other big issues)
15:39:54JoeBornyes, I think it could be a big hit and it's obvious that the things you want are the same things that other folks want too.
15:40:09JoeBornbut it's going to be a new product from a small brand
15:40:11LinusNnot the ipod herd...
15:40:26Zagori think the xclef model of having both an 1.8" and a 2.5" player with similar hardware is a good idea to consider.
15:40:55Zagorone small&sexy for the pocket and one which can fit huge 100gig drives, for the car or bag
15:40:58 Join Bagder [0] (
15:41:00JoeBornjust so you can have an 80GB if you want?
15:41:03 Quit R3nTiL (Read error: 110 (Connection timed out))
15:41:33JoeBornwhat about wireless? does anyone care?
15:41:42LinusNi dont
15:41:59JoeBornI'm referring to WiFi etc, not FM broadcasting
15:42:03LinusNi know
15:42:22JoeBornanyone else have an opinion on that?
15:42:25Zagori don't think it's worth it having it built-in. those things are easy to add with external hardware.
15:42:53LinusNbtw, S/PDIF is a must imho
15:42:55Zagorbut I'm not pretending to "know the market"
15:43:02JoeBornnow that's a very interesting point, what's the interface for external hardware?
15:43:25LinusNactually, an ethernet connection might be c00l
15:43:57JoeBornCF is big
15:43:58LinusNbut i think usb2.0 is good enough
15:44:01Zagorrio karma apparently has it
15:44:15ZagorCF is a good interface
15:44:23 Quit Lynx_ (Read error: 110 (Connection timed out))
15:44:24 Nick Lynx is now known as Lynx_ (lynx@
15:44:31JoeBornbut it adds considerable size
15:44:44LinusNcompared to what?
15:44:44Zagoryes. i'm not sure the tradeoff is worth it.
15:45:44LinusNthe 2.5" version could have more "advanced" connectivity features, while the 1.8" version could have only USB
15:46:00ZagorLinusN: sounds good
15:46:25*Zagor ducks
15:46:26 Quit MrMoo (Read error: 104 (Connection reset by peer))
15:46:32*Bagder swings for Zagor
15:46:34LinusNthe 2.5" version would be more like a portable music collection, much like i use my archos today
15:46:49JoeBornI take it you've already had the wireless conversation many times?
15:46:58LinusNnot really
15:47:24JoeBornoh it sounded like an inside joke
15:47:35LinusNrockbox is not that much about developing hardware, we have to settle with what the player has
15:47:41Zagorwell bluetooth *is* an inside joke, isn't it? ;)
15:48:07JoeBornI'm sorry to dominate the discussion, but one final question about licensing
15:48:20Zagordon't worry, this is a very interesting discussion
15:48:30JoeBorngood, thanks.
15:48:47LinusNwe'll kick you out when we find something else to talk about :-)
15:49:07JoeBornyou've never been shy
15:49:09 Join MrMoo [0] (~me@
15:49:36JoeBornso let's say we do what we're talking about and release a piece of hardware and do the development in an open environment
15:50:10JoeBornsourceforge or hosted internally, but with open version control, bugzilla, etc.
15:50:47JoeBornone part of our business model would be to ultimately license or OEM to other manufacturers
15:51:03JoeBornI know this is a pretty sensitive subject, but I have to bring it up.
15:51:16JoeBornso there are a couple ways for us to deal with this.
15:51:51JoeBornwe can fork the code into a version where we're the copyright holder or we can do a dual license, etc.
15:52:07JoeBornany particular thoughts on this? Does it make sense what I'm asking?
15:52:33LinusNi see two issues, one about copyright and one about money
15:52:39 Quit R3nTiL1 ()
15:53:25JoeBornI'd say that's right.
15:54:21Zagorthe question is how would you prevent other manufacturers from using the code unlicensed?
15:54:22JoeBornif money were no issue, then we could just release the code under BSD and contributors could contribute under that license
15:54:40JoeBornbut then why would anyone pay us?
15:55:15LinusNi also see an issue with fellow developers contributing to the project
15:55:16JoeBornand essentially we'll be paying developers to develop code for all manufacturers to use for free.
15:55:39JoeBornand of course the same issue you've faced.
15:55:44LinusNbasically, "why should i contribute, what's in it for me?"
15:55:50 Join [IDC]Dragon [0] (
15:56:12JoeBornright, that's a great question.
15:56:15[IDC]DragonHi again! (took a bit more than a reboot)
15:56:41ZagorJoeBorn: there is an ethical dilemma in selling contributed code
15:57:04JoeBornso what's the solution?
15:57:38ZagorI'm afraid I don't see one. Open Source and exclusive licensing doesn't mix well.
15:57:55BagderI don't think there's any shortcut. The source need to be "free enough" to attract hackers to it.
15:58:20LinusNand then anyone can clone the hardware and run the code on it
15:58:23JoeBornwell, that solution means forking the code into a plain old GPL
15:58:39Bagderyes, if someone makes a clone of the hw, they get code for free
15:58:47Zagorwould licensing only the hardware design not be enough, you think?
15:58:59JoeBornI don't have any issue with anyone cloneing the hardware and running the code on it.
15:59:26JoeBornremember they'd have to adhere to the GPL themselves. not many manufacturers would be willing to do that.
15:59:36Zagorexcellent point
16:00:00Bagderand if they'd do, they'd cooperate on improving the code
16:00:10Zagoreverybody wins
16:00:37LinusNexcept the company who has the most manufacturing costs, losing sales
16:00:38JoeBorneventually there will be a lot of hardware companies that will do that, but it will take time
16:00:55ZagorLinusN: why?
16:01:05Zagoroh, ok i get it
16:01:27Zagorhowever cost is only one variable. marketing is a powerful thing.
16:01:29Bagder... and there's money to be made by the people who know the sw to write adaptions for new hw
16:01:50LinusN...and who might that be? :-)
16:02:24JoeBornZagor: licensing HW might be enough, it depends how clever and unique or even *gasp* proprietary it is
16:02:59ZagorJoeBorn: yeah and we (rockbox) of course don't want it to be proprietary at all...
16:03:11JoeBornI know
16:03:35Zagorhowever do you see any problem with using the GPL, thus avoiding the "free clones" problem?
16:04:03JoeBornbut if it's completely a commodity, just a version based off the reference design then obviously it's not going to be worth much.
16:04:31JoeBornif you have a clever patentented wheel, etc. with a cool design, then maybe
16:05:06Zagoryes, design is very important. two devices with identical schematics can differ substantially in attractiveness.
16:05:09JoeBornI don't see a problem with the GPM and free clones problem
16:05:23JoeBornGPL I mean,
16:05:59JoeBornwhat I beleive will happen eventually is that the asian manufacturers will eventually realize the "free clones" opportunity and start making them.
16:06:18 Join Nibbler [0] (
16:06:36JoeBornalthough, it's an interesting question because will a "free clone" really work perfectly?
16:07:01JoeBornI mean the schematics are copyrighted themselves so you can't just copy them.
16:07:25JoeBornso how do you get them perfect when all the debugging has been done on the original device?
16:07:31Zagoreven "clones" would most likely have small differences tthat need code changes
16:07:47Bagderlike PCs running linux
16:07:53JoeBornRight, and that's an impdediment.
16:08:03Zagornot if the code is GPL
16:08:32JoeBornwell, it's still an impediment because you can't just drop the existing code into the device.
16:08:51LinusNyou would at least change the startup logo :-)
16:09:17JoeBornin our example, the manufacturer, whose english is not that good would have to contact you to have changes incorporated
16:09:18Zagorright, but the manufacturer would probably spend a month or however much it takes to adapt the software. rather than commission a complete 5 man-year rewrite .
16:09:33Zagor...or you
16:10:01JoeBornsure, but now they fork the code and when the official rockbox releases a cool new feature, they don't have it.
16:10:20Zagorif they want rockbox to work, they will submit their changes to us
16:10:27JoeBornas a user, now my clone just bit me, and I'd sure rather have "the real thing"
16:10:46Zagor...and thus they buy "the original", the neuros
16:11:20JoeBornright, wouldn't you?
16:11:23LinusNyes, so neuros can be proactive and contribute a lot
16:11:38Zagori probably would
16:11:47JoeBornexactly. that's why I'm not worried about the free clone issue.
16:11:56LinusNgood thought
16:12:20JoeBornthe issue is that we'd like to preserve the ability to license to someone else.
16:12:42JoeBornI wouldn't say it's a must have, but it would sure be nice.
16:13:11Bagderyou mean selling the code without GPL stamped on it?
16:13:23ZagorPerhaps, but I think it would be a hard sell to contributors.
16:13:46Bagderit would require a TrollTech/MySQL setup
16:13:58JoeBornI think if you say it would be a hard sell to contributors, then we can be pretty confident it's true:)
16:14:03LinusNa norwegian web site picked up on the latest slashdot news about rockbox being ported to iriver, saying thet rockbox coule "very well be the linux of media players"
16:14:14ZagorLinusN: link?
16:14:16JoeBornBadger: that's exactly right
16:14:19LinusNhang on
16:14:38BagderJoeBorn: it would "only" require that you are the copyright owner of all the code...
16:14:55Zagoryes, but I know a lot of people who won't contribute to qt for precisely that reason
16:15:03Bagderor have an agreement with the copyright holders
16:16:18ZagorI would suggest to drop the dual-licensing. I'd say the risk (of alienating contributors) is greater than the chance (of licensing the code for profit).
16:16:30ripnetUKhow the pan-european wiggler journey going ;-)
16:16:37LinusNmy mistake, it wasn't the norwegian web mag, it was a blog:
16:16:41JoeBornhave you guys considered making an "official organization" out of rockbox?
16:17:05ZagorJoeBorn: no :)
16:17:44LinusNripnetUK: holy moses, it's in Stockholm!!!
16:17:52JoeBornthat's one possible solution
16:18:00ZagorJoeBorn: how do you mean?
16:18:18LinusNstatus: "On FedEx vehicle for delivery"
16:18:26LinusNand i'm not at home :-(
16:18:33ZagorLinusN: call them
16:18:34JoeBornwell then the organization can own the code
16:18:47ripnetUK:) wiggler hits home :)
16:19:41LinusNmy wife has received the package! :-)
16:19:46JoeBorn"I wonder how long until companies stop making (and maintaining) their own firmware and start building players around Rockbox" :)
16:20:00ZagorJoeBorn: sure, but it's not a legal problem. it's an ethic/trust/goodwill issue
16:20:15ripnetUKnice... uk couriers have a habbit of sending deliveries to a depot MILES away
16:20:22Zagormaking rockbox a formal organisation won't help with that
16:20:45ZagorLinusN: excellent!
16:21:06LinusNZagor: that means i won't get any sleep this night either :-)
16:21:22JoeBornwell, it relies on teh continued trust and ethics and goodwill of Rockbox
16:21:38JoeBornit would just provide a mechanism for you to officially hold resources.
16:21:53LinusNi have to go now
16:21:57ZagorLinusN: go go go!
16:22:10LinusNJoeBorn: nice talking to you, keep in touch
16:22:14LinusNcu all!
16:22:19ZagorJoeBorn: ok I see what you mean
16:22:21JoeBornnice talking to you
16:22:24 Part LinusN
16:22:26 Join kurzhaarrocker [0] (
16:22:45JoeBornIt might be a horrible idea for all I know, I was just curious.
16:22:46Zagorwe don't have anything against formalizing rockbox, there just hasn't been any reason to
16:22:59Zagorand we hate paperwork :)
16:23:19JoeBornI understand that.
16:23:57kurzhaarrocker(completely ot) Zagor: I want to request a web page with a form and parse it. Is curl the way to go?
16:24:19JoeBornI need to run myself, thanks for your time everyone.
16:24:21Zagorcurl can do anything :)
16:24:38kurzhaarrockerOk, Ill try to make it make coffee.
16:24:45ZagorJoeBorn: ok, thanks for the discussion
16:26:33dwihno_Yay for the wiggler!
16:26:35 Nick dwihno_ is now known as dwihno (~dw@
16:27:18ripnetUKlets hope it wiggles ok with the iRiver
16:36:13Zagorbrace for commit
16:39:41 Quit ashridah ("crash")
16:40:13*dwihno bucles his safety belt
16:40:19*dwihno braces for impact
16:40:55*kurzhaarrocker tries to curl up :)
16:41:33Zagornow you can all build iriver simulators and play with :)
16:42:32 Join methangas [0] (
16:42:49Bagderfinally! :-P
16:43:32Bagderwant that in the build table?
16:43:59kurzhaarrockerI wonder when the ideas about supporting other platforms than the recorder/player became serious.
16:44:32Bagderkurzhaarrocker: when they discovered enough details on the iRiver
16:44:43 Quit Nibbler (Read error: 110 (Connection timed out))
16:45:11Zagoractually the *idea* came earlier than that
16:45:27kurzhaarrockerBagder soon a table won't be sufficiant any more, prepare to build a build cube :)
16:46:05kurzhaarrockerYes but usually the idea was considered as a topic of the nodo.
16:46:29 Quit JoeBorn ("CGI:IRC (EOF)")
16:47:50kurzhaarrocker:( my web hoster doesn't support curl.
16:48:22Bagderbuild it yourself
16:48:52***Saving seen data "./dancer.seen"
16:49:11 Join Nibbler [0] (
16:51:07kurzhaarrockerI'm not sure but I believe that if php says it was configured with "−−without-curl" I'm in a dead end - no matter how many curl libs I build...
16:51:35Bagderright, but you can use the command line tool
16:51:36Zagoryes, if you must use it inside php. but you can always use the command-line program
16:52:13*kurzhaarrocker considers that
16:53:08Bagderit might have the tool installed
16:58:02 Join midk_ [0] (
16:58:12*[IDC]Dragon missed the interesting-looking discussion with Joe
16:58:30[IDC]Dragon(and had not enough time yet to read it)
16:59:01 Join mecraw_ [0] (~lmarlow@
17:07:42[IDC]Dragonamiconn: do you read?
17:07:50Zagoryeah it was a nice chat
17:08:07[IDC]Dragonhow do you know that guy?
17:08:28[IDC]Dragonhe's from Neuros s/w development?
17:08:31Zagorwe've mailed a bit about their open source effort
17:08:36Zagorhe's their ceo
17:09:45Zagorsorry, "only" chairman and cto
17:10:15[IDC]Dragonpoen source witout open development tools :(
17:10:47Zagoryeah. that's why he was here talking to us today, to avoid doing such a mistake next time
17:11:01[IDC]Dragonmistake? ;-)
17:11:46Zagorwell what else would you call it? if they wanted it proprietary, they wouldn't have opened the source would they?
17:13:10[IDC]Dragonno, I mean because it has that accidental touch, ironically
17:13:36[IDC]Dragonby mistake practically nobody can build it, sorry
17:14:25 Quit midk (Read error: 110 (Connection timed out))
17:18:30[IDC]Dragon"Added Iriver H100 to tools/configure", hehe
17:18:55Zagoryeah, so we can build the h100 simulator
17:20:41[IDC]Dragonapart from that, it's early and optimistic :-)
17:21:27Zagoryup :)
17:28:19ripnetUKjust built the UI sim for iRiver - looking kinda cool
17:28:33ripnetUKalthough still green :)
17:31:28Zagorgotta go
17:31:30 Part Zagor
17:36:18 Quit AciD ("how much wood would a wood chuck chuck if a wood chuck could chuck wood ?")
17:40:41 Quit [IDC]Dragon ("CGI:IRC")
17:48:10 Part kurzhaarrocker
17:55:43 Quit Bagder ("Leaving")
17:58:41ripnetUK−−- uisimulator/x11/uibasic.c 20 Jan 2003 09:39:37 -0000 1.13
17:58:41ripnetUK+++ uisimulator/x11/uibasic.c 16 Sep 2004 15:58:16 -0000
17:58:41ripnetUK@@ -61,7 +61,7 @@
17:58:41DBUGEnqueued KICK ripnetUK
17:58:41ripnetUK char *progclass = "rockboxui";
17:58:41ripnetUK char *defaults [] = {
17:58:41***Alert Mode level 1
17:58:41ripnetUK- ".background: lightgreen",
17:58:43ripnetUK+ ".background: lightblue",
17:58:45ripnetUK ".foreground: black",
17:58:47ripnetUK "*help: false",
17:58:49ripnetUK 0
17:58:53ripnetUKmuch more iRiver
18:08:24 Join LinusN [0] (
18:08:42***Alert Mode OFF
18:09:41 Join AciD [0] (
18:11:35 Part LinusN
18:22:51 Join _aLF [0] (
18:37:22 Join webguest07 [0] (
18:38:05 Quit webguest07 (Client Quit)
18:48:56***Saving seen data "./dancer.seen"
19:03:41 Quit AciD (Read error: 104 (Connection reset by peer))
19:06:55 Quit Lynx_ (" WOW! This IRC Client ownz! HydraIRC -> <-")
19:11:45 Join AciD [0] (
19:14:50 Join srn [0] (
19:16:49srni just added some pages to the iriver wiki. unfortunately, i positioned a page at the wrong level. IriverPort > IriverBDM > IriverMemoryMap −− should have been −−> IriverPort > IriverMemoryMap. can any one fix this?
19:17:16 Join [IDC]Dragon [0] (
19:18:03amiconnhi. Now I'm around
19:18:20[IDC]Dragonseen my "answers"?
19:18:21amiconn(Sometimes I even have to work at work ;) )
19:18:31[IDC]Dragonthat sucks :(
19:22:12amiconnYes, found your answers
19:22:39[IDC]Dragonthey seem to "burst" a sector or so
19:23:11[IDC]DragonI didn't measure it, but could do so for an estimate, if that helps
19:23:16amiconnOn PA13: It is used both in the serial transmit and receive, switched between GPIO and IRQ1
19:23:45[IDC]Dragonwith serial you mean SPI for MMC?
19:24:10[IDC]DragonI will go hunting for that signal
19:28:08 Join R3nTiL [0] (
19:30:47amiconn[IDC]Dragon: before transmit/ receive: PACR1 &= 0xF3FF (PA13->GPIO). After transmit/receive: PACR1 |= 0x0400 (PA13->IRQ1)
19:31:03dwihnoare you fiddling with the ondio?
19:31:51dwihnoAre you trying MMC communication with the on-board chip?
19:33:25amiconnNot for real yet.
19:34:10amiconnCurrently looking into various data sheets and at the disassembled archos fw to get a clue how this is supposed to work
19:35:14 Join webguest08 [0] (
19:36:17 Quit webguest08 (Client Quit)
19:36:43amiconnThe SH1 hardware manual is not very precise about the serial communication. E.g. with synchronous mode, what could cause a receive overrun error? Iiuc, synchronous mode does only output the clock signal if it is ready to receive, so in theory overrun shouldn't be possible.
19:37:15[IDC]Dragonwith exteral clock, I think
19:39:49amiconnI already have a serial init, and will try to check if enabling serial causes continuous clock output. It's a bit difficult for me though (neither a LA nor a digital scope, only a simple analog scope)
19:40:10[IDC]Dragonanalog is fine
19:40:16[IDC]DragonI do most with it
19:40:39amiconnA logic analyzer could actually prove valuable here
19:41:37[IDC]Dragonc'mon, you only want to check for your piulses, nothing complx
19:41:59[IDC]Dragon...pulses, nothing complex
19:42:19amiconnNot for this very job, but for capturing e.g. a data transfer burst
19:42:59[IDC]DragonI can do that, if necessary
19:43:10[IDC]Dragon(but mostly, it isn't)
19:43:35[IDC]Dragonthe SCI itself should not need debugging
19:44:34amiconnYes, like the recording data transfer. Still no luck :(
19:44:47[IDC]Dragonwe got out LCD working, too, where every transition is done "by hand"
19:45:34amiconnYes. In fact I think that's easier than dealing with the SH1 SCI
19:45:56amiconnIt may even prove to be faster
19:46:03[IDC]Dragonfor a first shot, you could do so as well
19:46:31[IDC]Dragonno, it won't be faster, methinks
19:47:11[IDC]Dragonwith the LCD, we got an overall ~900kBit/s
19:47:18amiconnThe highest serial speed we can set for continuous transfer is 1.5 MBit/s, which equals 8 cpu cycles per bit.
19:47:29[IDC]Dragon112*64*130 fps
19:47:50[IDC]Dragonand the carry trick won't work here
19:48:05[IDC]Dragonfor bad data bit position
19:48:18[IDC]Dragon(I already checked this mentally)
19:49:37amiconnAh ok. I didn't take into account that each data access to the ports needs 2 wait states
19:50:25amiconnWhere do I have to measure the clo
19:51:01amiconnIs it possible at the 74AC08?
19:51:17amiconnSo it's not that difficult
19:51:43[IDC]Dragonsee twiki table
19:52:16[IDC]Dragonbtw, did you see my new photo of the "naked" board?
19:53:00 Quit R3nTiL ()
19:56:24 Quit maikeul (Remote closed the connection)
20:01:29amiconn[IDC]Dragon: I have seen the bursts of the original fw on my scope now.
20:02:13amiconnThe bridge clock is too high for my scope though
20:02:48[IDC]DragonI found PC6 instead
20:02:58[IDC]Dragonmore later, cu
20:03:11 Quit [IDC]Dragon ()
20:04:53 Join gromit`` [0] (
20:49:00***Saving seen data "./dancer.seen"
20:50:17 Quit Nibbler (Read error: 60 (Operation timed out))
20:56:24 Join RyanA [0] (
20:56:49RyanAdoes anyone know where to get a replacement AC adapter cord real cheap?
20:58:39 Quit RyanA (Client Quit)
21:30:35 Join mecraw__ [0] (~lmarlow@
21:41:03 Quit srn ("CGI:IRC (EOF)")
21:48:32 Quit mecraw_ (Read error: 110 (Connection timed out))
21:56:48 Join [IDC]Dragon [0] (
21:57:13[IDC]Dragonhi again
21:57:23[IDC]Dragonamiconn: I found PA13
21:57:33amiconnWhat is it?
21:57:36[IDC]Dragonit disappears under the flash
21:57:50[IDC]Dragonand goes to a test pad
21:58:09[IDC]Dragonthe flash is BGA, I can't beep that
21:58:16amiconnI know.
22:00:49 Join Zagor [0] (
22:01:56amiconn[IDC]Dragon: Maybe it is connected to the RDY/BSY interrupt signal
22:02:32amiconnPA13 is always input
22:02:56amiconnWe won't need that for now
22:03:10ZagorripnetUK: committed
22:03:15amiconn[IDC]Dragon: What about PC6?
22:03:42amiconnAh, found it.
22:03:58amiconnNeeded to reload the wiki page :)
22:04:22[IDC]Dragondidn't you say it switches direction?
22:04:59amiconnNo, it switches function: GP in <> IRQ1
22:05:39[IDC]Dragonah, never mind.
22:06:25amiconnI tried to output some clocks (writing 0xFF == dummies). No luck yet :(
22:06:35[IDC]Dragonmaybe it's the ready/busy signal
22:09:08 Join scott666_ [0] (
22:14:57[IDC]Dragonthat's the only thing which makes sense
22:16:32 Join Bagder [0] (
22:27:21amiconnI don't seem to get that dreaded SCI setup right. Both transmit and receive don't output any clocks; transmit ends much faster than it should, receive hangs :((
22:28:10[IDC]Dragonhave you set the pin functions correct?
22:28:20amiconnYup (at least I think so)
22:28:56[IDC]Dragonpeeking at the MAS code doesn't help?
22:29:08amiconnI did peek there
22:29:19[IDC]Dragonour MAS code, I mean
22:30:17 Join [av]bani [0] (
22:30:30[av]baniRe: JoeBorn
22:30:38[av]banithey should either focus on the hardware, or the software
22:30:42[av]banibut not both
22:30:52[av]banieg, make killer hw and make it easy to code for
22:31:11[av]banimake a generic clonable/licensable open hw, and sell killer sw for it
22:31:21Zagorwell even if they make killer hardware, they have to make decent firmware too.
22:31:32[av]baniif they pull a zaurus, it will flop
22:31:40[av]baninice hw, but poorly supported
22:31:53Zagoryes, but you also can't bet the company that an open source project will do everything
22:32:26[av]baniwell, take rockbox for an example, the opensource fw turned out much nicer than stock, because you were able to hack it
22:32:38[av]banibut imagine how much ebtter it would have been had it been openly supported from the start
22:32:45[av]baniinstead of making you jump through hoops
22:33:00[av]baniits like the pc market
22:33:08[av]baniyou either focus on the hw, or focus on the sw
22:33:13[av]baniyoure either dell or microsoft
22:33:19[av]banidont try to do both
22:33:21Zagorpcs are commodities. mp3 players aren't (yet)
22:33:33[av]banisure they are, mp3 players are definitely commodities
22:33:43[av]baniapple made sure of that
22:34:04[av]banionly a few companies can really pull off both, and it takes a lot of effort
22:34:07Zagormp3 player hardware is not as interchangeable as pc hardware, is what I mean
22:34:10[av]baniapple ipod, rio
22:34:20Zagoreach hardware manufacturer has to also make the software to run on it
22:34:24[av]banilook at how many companies flopped because either the hw was inadequate or the firmware was
22:34:29Zagormuch like the pc business in the early 80s
22:34:47[av]baniwell in the case of being open
22:34:52[av]banieither focus on the hw or focus on the sw
22:35:03[av]banione or the other
22:35:12Zagoryou are right in principle, but we aren't there yet
22:35:15[av]baniif youre totally closed/proprietary, you dont really have a choice
22:35:24[av]baniwell, i'm just saying what Joe should do
22:35:26Zagorjust making good hw will, as you say, risk a zaurus flop
22:35:30[av]baniif they want an open player
22:35:49[av]banidecide which side they will take
22:36:06[av]banimake killer hw and work closely with rockbox for good support
22:36:29[av]baniboth sides win - they get nice fw and you get a nice player
22:36:45Bagderbut that _is_ making good hw and good sw
22:36:58Bagderonly open sw
22:37:02[av]banithey wouldnt be writing the sw though
22:37:09[av]banitheyd be making it easy for you to do so
22:37:42amiconn[IDC]Dragon: Many of your &= / |= settings for the PFC change only either the upper or lower half of the word, so could be replaced with and_b() / or_b(). Not done yet.
22:37:49Zagortoo risky. they have to write some software themselves too.
22:37:53amiconnAnd it seems I found a mistake
22:38:01Zagormaybe the next company doesn't, but Joe definitely has to
22:38:32[av]banirockbox already works, if they make it easy for you to poke the hw then a port should be easy
22:38:39[av]baniand faster than them writing their own
22:38:41Zagorat least I wouldn't want to bet my company on an bunch of volunteers who are under no obligation to do anything
22:38:48[av]baniplus, they can plug it that way, opensource player
22:38:56[av]baniget all the /.'ers drooling etc
22:38:57Zagoryes, we all agree about that
22:39:01Zagoreven joe
22:39:12[av]banibecause writing fw from scratch would likely take longer and cost more
22:39:19[av]baniso they can focus more on investing in hw
22:39:23[av]bani= better player
22:39:45[IDC]Dragonamiconn: yes, I know. didn't bother for now, it reads nicer while experimenting.
22:39:47[av]banifew vendors actually _sell_ the firmware
22:39:55 Join Nibbler [0] (
22:40:20[av]baniin the sense that most players are built with off the shelf parts
22:40:26[av]baniwell, rockbox is one of those parts :)
22:40:36Zagor[av]bani: you don't need to sell *us* on the benefits of open source :)
22:40:52[av]banidid he say when he would be back?
22:40:57[av]baniso i can pitch to him :)
22:41:14[av]banii just think that an openly clonable hw platform is the wrong way to go for them
22:41:25[av]baniwhich was what he was proposing
22:41:26Zagoryou don't need to pitch to him either. the sole reason he was here was to find out how he could make their next product more rockbox-friendly
22:41:57[av]banischematics and devkits with bdm support would be the best way :)
22:42:40[av]baniretail box, and dev sdk
22:47:16[av]baniwiggler@home ?
22:47:48Bagder@linus' home
22:49:04***Saving seen data "./dancer.seen"
22:49:07amiconn[IDC]Dragon: Although I made SCK1 an output now, still no clock pulses :(
22:50:15[IDC]DragonI once got pulses, a steady frequency
22:50:24[av]banipulses are good
22:50:46[IDC]Dragon(even without outputting data)
22:50:53[IDC]Dragondo you writedata?
22:51:20[IDC]Dragontransmit enabled, and everything?
22:51:36amiconnI try to write 15625*16 bytes @400 kBit/s. Should take around 5 secs
22:53:38amiconn@375 kBit/s to be exact
22:54:38[IDC]Dragonhow do you init SCI1?
22:54:49amiconnvoid setup_sci1(int rate) /* in bit/s, rounds down to the next possible value */
22:54:50amiconn int i;
22:54:50DBUGEnqueued KICK amiconn
22:54:50amiconn int divider = (FREQ/4 - 1) / rate;
22:54:50***Alert Mode level 1
22:54:50amiconn SCR1 = 0x00; /* disable serial port */
22:54:52amiconn SMR1 = 0x80; /* Synchronous, no prescale */
22:54:54amiconn BRR1 = divider;
22:54:56amiconn SCR1 = 0x01; /* clock output */
22:54:58amiconn and_b(0xC7, SSR1); /* clear ORER, FER and PER */
22:55:02amiconn for (i = 0; i < divider; i++); /* wait at least 1 bit time */
22:55:02amiconn or_b(0x30, SCR1); /* enable TxD and RxD */
22:56:06amiconnThis is called from ata_enable(true) with setup_sci1(400000); after setting SCK1, TxD1 and RxD1 pin functions, and making SCK1 and TxD1 outputs
22:56:38[IDC]Dragonthe output part is only relevant if GPIO
22:56:58amiconnThe divider evaluates to 7 for rate = 400000, verified with readback
22:57:58amiconnThe output part is done in the mas serial setup (at least for SCK0). It shouldn't hurt anyway
23:01:11 Part [av]bani
23:01:33[IDC]Dragonhow do you wait for your output?
23:01:53amiconn for (i = 0; i < len; i++)
23:01:53amiconn {
23:01:53amiconn while (!(SSR1 & 0x80)); /* wait for TDRE = 1 */
23:01:53***Alert Mode level 2
23:01:53amiconn TDR1 = buf[i]; /* write byte */
23:01:54***Alert Mode level 3
23:01:54amiconn and_b(~0x80, SSR1); /* start transmitting */
23:01:54***Alert Mode level 4
23:01:54amiconn }
23:01:56amiconn while (!(SSR1 & 0x04)); /* wait for TEND = 1 */
23:02:13amiconnThis is part of my mmc_write_transfer function
23:02:41amiconn(Well, most of it for now. I don't bother trying DMA for now)
23:02:57[IDC]Dragonyou have to clear TDRE, i think
23:03:21amiconnI do this: and_b(~0x80, SSR1); /* start transmitting */
23:04:02[IDC]Dragonah, yes
23:04:34[IDC]Dragonhave you tried the reverse order?
23:04:53[IDC]Dragonclear TDRE first, then write new data
23:05:44amiconnThis would transmit the data that was in TDR1 before. The sequence is from the sh datasheet. I can't see what I am doing wrong here, maybe I'm blind.
23:07:59[IDC]DragonI don't see it either
23:08:44[IDC]Dragonhow about a close comparison withthe MAS play code?
23:09:09amiconnThe MAS play code uses dma & interrupts...
23:10:30BagderZagor: I can't build the iriver sim
23:11:11Bagderoh never mind
23:11:15 Quit silencer (Nick collision from services.)
23:11:18 Join silencer [0] (
23:11:19Zagorok :)
23:11:27 Quit silencer (Nick collision from services.)
23:11:55***Alert Mode OFF
23:11:56Bagderbiiig display ;-)
23:12:38Zagorthe text reader becomes very nice on this display
23:13:26Zagorhmm, did I break something? "make install" does not install the fonts
23:14:10[IDC]Dragongotta sleep, c u
23:15:20 Quit [IDC]Dragon ()
23:15:41BagderZagor: I think the buildzip needs a little pat on the shoulder
23:16:17 Join silencer_ [0] (
23:16:43Bagdercheck the last 5 lines of it
23:20:07Bagderyou fix or I?
23:21:25Zagorwe should remove all line wraps from the docs when we add it to the zip. they are ugly to read like this.
23:22:18elinenbewhat's the status with the iriver? has much/any progress been made since last week?
23:22:41Zagorlots of new information, no code ran yet
23:23:19Zagorbut our bdm pod arrived to linus today so things will hopefully take off a little soon
23:23:53elinenbeis everything in the wiki?
23:24:01elinenbewhat is the bdm pod used for?
23:24:55Zagorit's for interfacing with the cpu's debug module. that way we can run gdb against it without running code. also it will help us flash it.
23:27:17elinenbewow... nice! how much did that cost?
23:32:14CtcpIgnored 2 channel CTCP requests in 4 hours and 56 minutes at the last flood
23:32:14*Zagor got a mail with a link for a broken iriver on ebay
23:32:46Bagderlemme guess!
23:32:53Bagder"deliver in US only" ;-)
23:33:01Zagornope, located in ireland
23:33:09Bagderhow much?
23:33:18Zagor$15 so far
23:33:35 Nick silencer_ is now known as silencer- (
23:33:36Bagderand long time time is it left?
23:33:40Zagor3 days
23:34:02Zagor4 days have passed, 3 bidders (started at $5)
23:34:24Bagderok, so it could be worth making a bid in a few days
23:38:45 Quit ripnetUK ()
23:51:07 Quit scott666_ (Read error: 110 (Connection timed out))
23:51:42 Join scott666_ [0] (

Previous day | Next day