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).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2005-02-06

00:00:11Digital007I saw that iRiver work is progressing, but no updates on the iriver wiki page
00:00:20Cassandraamiconn: problem is safe_shutdown is in app, and the idle timer is in firmware.
00:00:24[IDC]Dragonamiconn, about a UI aspect of charging: I'd prefer to indicate charging by a blinking plug symbol instead of the animation
00:00:38amiconn[IDC]Dragon: Imho the problem is (1) Charge with full current, to fill up the cells as quickly as possible, but don't overcharge (2) Start charging as soon as possible, to avoid the flat battery problem
00:00:48[IDC]Dragonso the battery symbol can show the approximate battery state
00:01:01amiconnThe "don't overcharge" is the most problematic part
00:01:53amiconnSo instead of tweaking the delta-v with all its problems caused by the variable load, I'd rather propose to turn the variable load into an advantage for the charging algo
00:02:29amiconnThe variable load allows us to know the inner resistance, which gives more precise info about charging state than the voltage
00:02:35[IDC]Dragonto measure explicitely under load?
00:03:09amiconnYes, that's why I think the power management needs to know the power state (hd, backlight)
00:04:15[IDC]Dragonthe HD, mostly
00:04:26amiconnHD needs the highest current, but it's not very predictive
00:04:40[IDC]Dragoncurrently, measurements with the HD are taken out
00:04:47amiconnI measured surges up to 0.9 A when hd starts
00:05:02[IDC]Dragonbut I always had the feeling that these are rather interesting
00:05:18amiconnAverage is 0.5 A when the disk is accessed, and ~0.25 A when spinning only
00:05:26BagderDigital007: reload now
00:05:57amiconnOf course this also depends on the hd model
00:07:08amiconnBacklight current is much more predictive, but I doubt it's high enough to see a voltage difference
00:07:51[IDC]Dragonour backlight needs more current ;-)
00:08:12amiconnThe adc gives us a precision of ~5 mV, but the jukebox environments seems to be rather noisy
00:08:17[IDC]Dragonthe original is very low power
00:09:09[IDC]Dragonwhat a pity that we can't measure our own battery current
00:09:14amiconnI observe fluctuations of up to +/- 50 mV in the debug display
00:09:25[IDC]Dragonfor backlight?
00:09:43amiconnNo, caused by the noise
00:10:04amiconnBtw, the Ondio does not show this effect
00:10:11amiconnNice stable readings there
00:10:31[IDC]Dragonno HD
00:10:50amiconnIt also happens when the HD is off
00:11:12amiconnStart your box, go to debug->view i/o ports
00:11:49CassandraOh rats. The sleep timer is handled the same way as the idle poweroff. Calling safe_shutdown() from powermgmt.c is definitely a no-no, right, due to that being a callback from the firmware to the application?
00:11:51amiconnOn the player, I get readings ranging from 5.13..5.26 V right now
00:12:28amiconnCassandra: That should be avoided, yes
00:13:17Cassandrahmmm. Not really sure what to do about it then.
00:13:18amiconnThe recorder fluctuates less, I get 5.20..5.22 V
00:13:47amiconnOf course, recorder is the most important
00:13:52Digital007cheers Bagder
00:14:14amiconn[IDC]Dragon: For the charging animation, the method my cellphone uses is quite good
00:14:56[IDC]DragonI don't know your cellphone ;-)
00:15:12einhirnAbout these voltage fluctuations: Maybe they just did a better job with stabilizing the voltage in the Ondio... A nice cap near the variable loads (ie, processor) can do wonders...
00:15:13amiconnIt shows an animation, filling up pixel by pixel, but not always from empty to full, but from the current charging state to full
00:16:23amiconnMy cordless phone does pretty much the same, only that it doesn't work pixel by pixel, but has 4 segments
00:17:12amiconnAs long as the battery is almost-but-not-yet full, the last pixel column/last segment effectively flashes
00:19:14amiconnWith this system, you can estimate how much charging time remains until full charge
00:20:35[IDC]DragonI understand
00:20:51CassandraRight - I think half on clean_shutdown needs to go into firmware.
00:20:55[IDC]Dragonbut the flashing plug symbol would do it for me
00:22:38amiconnIn what respect would that be different to the current implementation?
00:23:02[IDC]Dragonthat you see the fill level
00:23:09amiconnAh, ok
00:23:32CassandraActually, can anyone think of a good reason why clean_shutdown shouldn't *all* move into firmware?
00:23:33[IDC]Dragonand I want that in USB mode, too
00:24:03[IDC]DragonCassandra: saving the settings is app code
00:24:30amiconnThe current implementation does work in usb mode, so no change for that.
00:25:03amiconnThe fill-from-current-state animation could be improvend a bit further, by using a different pattern for the fill up (not completely black)
00:25:09CassandraHmmm. That must happen in ata_flush(), which is already in the firmware.
00:26:07amiconnThis way the current state would always be visible, yet the animation being intuitive
00:26:42amiconnAll this shouldn't be hard to implement, much easier than the overhaul of the charging algo itself.
00:26:49CassandraNo, I'm wrong.
00:26:59CassandraWhere the hell are the settings actually saved?
00:27:51einhirnamiconn: That would be a nice to have, but a gauge (or as I prefer the numeric display) showing the current level and the blinking Plug symbol indicating "I'm charging" would surely be even easier to implement, wouldn't it?
00:28:47[IDC]Dragoneinhirn: yes, for numeric you don't want this to count up or so for "animation"
00:28:50amiconnYes, but it wouldn't be as intuitive
00:29:57amiconnDoing something with the _battery_ symbol to indicate charging the _battery_ is intuitive imho, while blinking the plug is not.
00:30:08einhirn[IDC]Dragon: I've already seen this with my box: When charging, it sometimes shows the gauge animation, sometimes it shows the gauge animation numerically. Quite weird...
00:30:35 Quit Digital007 ("CGI:IRC (EOF)")
00:30:37amiconnIn fact, if the plug symbol blinks one could think this indicates something is wrong with the charger, and disconnect it...
00:31:01amiconneinhirn: If so this is definitely a bug
00:31:30amiconnEven when numeric battery display is enabled, the charging animation should always be graphic
00:31:31[IDC]Dragonamiconn, how do you indicate charging for numeric display then?
00:31:55[IDC]Dragonyeah, not very stringent
00:32:17amiconn[IDC]Dragon: You could either do the same as it is now (or should be) - switch to graphic display
00:32:51amiconnOr you could invert (or invert-blink) the numeric display
00:32:53CassandraHow about inverting the charge indication area when the battery is charging.
00:33:10CassandraSo charging symbol is white on black when mains connected.
00:33:54amiconnCassandra: There is almost no frame around it.
00:33:58einhirnWell, the gauge has one big problem for me: It can't tell me wether I have 10 or 30 Percent charge left - at least I won't see it because I'm not going to count pixels.
00:34:20[IDC]DragonI don't want to fuel a debate, but I'd prefer the gauge (numeric or graphically) show the level all the time, and use the charger plug symbol for plugged and charging
00:34:38*Cassandra nods. Fair point.
00:34:43[IDC]Dragonsince we have the luxory of it
00:34:50amiconneinhirn: Correct, but then knowing the exact percentage doesn't tell you much more, at least with the current implementation
00:34:54BagderI think I'm with [IDC]Dragon on this
00:36:13[IDC]Dragonaccuracy is an unrelated issue...
00:36:28CassandraYou know, as far as I can tell, clean_shutdown doesn't flush the settings to disk at all. I've got to be wrong about that, haven't I?
00:36:31[IDC]Dragonto be solved at a diferent place
00:37:52amiconnCassandra: It does, by calling ata_flush()
00:38:17CassandraOh, yes. The change is in the ata buffer already.
00:38:26amiconn[IDC]Dragon: Yes of course
00:39:25amiconnCassandra: To be precise, it's in the delayed_sector buffer. The ata driver does no write buffering otherwise, only for the config sector
00:39:34HClwhats up?
00:39:36CassandraI think, except for the onscreen message, all of that could safely be moved to say hw_clean_shutdown in firware so we can have the sleep timer and idle timer do a clean shutdown too. Would people be OK with me doing that?
00:40:00einhirnamiconn: Of course it isn't precise, but that doesn't matter. Lets take another example: I like "analog" watches better than digital ones. But only if I can grasp the current time with a blink of the eye. If its display is too crowded or too small, I'd like the digital watch better. Same for the Graphical gauge: It's so small that I won't get it with a blink of the eye. So I like the numeric displ
00:40:00einhirnay better, because I can read it and know the battery percentage my box currently guesses...
00:41:16einhirnWell, anyway, I'm with [IDC]Dragon on the "I'm charging"-display too...
00:41:25amiconnFunny, for me it's the other way round. I always prefer digital watches over analog ones, because I want the exact time. However, I have set both battery & volume display to graphic
00:42:50einhirnk ;) I know that the battery gauge is just a guess, and can't be really exact... So, I'd rather read the "exact" guess, if I'm interested in it...
00:43:00amiconn...I think this is because I don't want the exact values (they're of little use), but want an estimation at quick glance
00:43:10[IDC]Dragonafter the 80th, I got an analog watch again
00:43:12amiconn...which isn't possible with numeric display
00:44:00einhirnThats right - and thats why I like analog watches more than digital ones ;)
00:44:34amiconnSo you are more interested in the exact charging state of your jukebox than in the exact time ?! ;-)
00:46:18einhirnHehe ;) As I already said: If I'm interested in knowing rockbox'es guess about battery level, I will read the numbers, since they offer a bigger resolution than the graphic gauge...
00:46:43einhirnEmphasize: _If I'm interested_
00:47:07amiconnThen you could still use Info->Rockbox Info...
00:47:45amiconn[IDC]Dragon: The "blinking plug" method won't work on the player. There is no plug icon...
00:47:55einhirnOk, but for that I'd need to navigate the menus, which, for me is a bother rather than a joy...
00:48:45einhirnAnyway, just make it configurable (a little) and everybody will find the way he/she likes...
00:48:56einhirnNow to something completely different ;)
00:49:01[IDC]Dragonplayer, ohh
00:49:49Bagderthe only people still using players are blind ;-)
00:49:52[IDC]Dragonconfigurable? please no further option clutter
00:50:04einhirnCouldn't the voltage fluctuations be circumvented by taking the average of the last three or five values?
00:50:05amiconn..while the fill-from-current state would work (just not very precise though, like the state display itself)
00:50:12***Saving seen data "./dancer.seen"
00:50:52amiconnYes of course. But the battery voltage alone doesn't tell us that much
00:51:08amiconn...because the discharge curve of NiMH is rather "flat"
00:51:33amiconn...and the tolerance between units may hit
00:51:50einhirn[IDC]Dragon: No, thats not what I want, I'd stay with how it works now. Like "Fancy animation" <-> "Numeric Display"
00:52:50amiconnJörg's and my Ondio differ by about 1.5 % on voltage display, that makes a difference in charging state of ~30 % in the center of the discharge curve...
00:53:57amiconn...when using NiMHs, that is
00:54:11einhirnamiconn: Ok, you want to guess Battery level using a change in load... Just turning on the Backlight won't do the trick, since the fluctuation is rather large. But maybe turning on the backlight and taking an average of the voltage would do?
00:54:41einhirnOf course I get the point that the guess still would be rather crude...
00:54:55[IDC]Dragonamiconn: the player can go without a plug icon, since the charging is done in hardware
00:55:46amiconnWhile the charging is done in hardware, I still want to know whether the cells are already full or not
00:56:10[IDC]Dragonthe gauge is there
00:56:11einhirnYou can still see that on the gauge...
00:56:14amiconn(that may be even more difficult than the recorder, as we have no control over the current
00:56:45amiconnRockbox currently doesn't do this on the player. It always shows the animation
00:58:03einhirnhmm... I'd need a confirmation that the charger is plugged in and is supplying voltage... I'd leave the rest to the charging electronic...
00:59:40amiconnHow do you know the cells are full then, so you can disconnect the charger and take the unit with you, having maximum possible runtime?
00:59:57einhirnOf course I'd also want to know what level my batteries have, so even during charging I'd like the box to do an educated guess about level ;)
01:00:51amiconnYups, and that's why I think the fill-from-current-state animation is useful.
01:01:30amiconnThe player can only do graphic display of course
01:02:02einhirnWhat I meant with confirmation: I don't need to know _if_ it is charging ATM, but that it _can_ charge if it wants to.
01:02:15amiconnI think using all the techniques proposed (averaging the voltage measurements to get rid of the fluctuations, measuring input voltage to calculate the current, measure much more often, and take the load into account by knowing the hd & backlight state), we could come up with a really decent charging algorithm and state estimation
01:03:17einhirnAnd the other thing, knowing when it's ready: Of course you could do the animation. And you could also do an equivalent of that in numeric mode: Toggle between 100 and current value ;)
01:05:06amiconn[IDC]Dragon: If the charging algo would always work correctly, one wouldn't need a separate charging indication. Having the charger connected (plug symbol on), and battery state < 100% implies charging, because reaching 100% is the goal of charging
01:06:02einhirnAmiconn: Thats what I meant with "I need to know that it _can_ charge _if_ it needs to" ;)
01:06:02einhirnCan't we get to know something about the cells from "Charger-On-Off"-Behaviour?
01:06:46einhirnOr tune the Trickle PWM to equalize the System voltage?
01:07:27einhirnLike the switching voltage regulator does it's job...
01:07:38amiconnThe problem with regulating voltage is that you don't know which voltage is "right", as I already said.
01:08:29einhirnValid point. And guessing the "right" voltage could be fatal to the Batteries.
01:10:35einhirnBack to "Charger On Off": I don't know if it will kill the cells, but try to charge the Batteries for lets say 10 sec, then measure the voltage decay afterwards. This way one could guess where in the discharge curve we are...
01:10:35amiconnI think I should put together a tsr plugin that monitors voltage and logs it, together with hd spinning/not spinning, backlight on/off and charger on/off info, and run this as autostart.rock for a while.
01:11:05einhirnamiconn: So that we have values to talk about - good Idea ;)
01:11:27lostlogicsome of the supported players use software controlled charging?
01:11:59einhirnlostlogic: One of them: Recorder V1
01:12:15amiconnFirst I need to purchase a new set off cells - mine seem to have problems (probably one cell having a high self-discharge rate, and/or increased inner resistance)
01:12:16[IDC]Dragonamiconn, how do you know the disk is spinning?
01:12:51[IDC]Dragonby the voltage dip?
01:13:17[IDC]Dragonand the backlight by dirty port probing?
01:13:29amiconnI would hook into the backlight driver
01:13:46[IDC]Dragonoops, it's I2C, forget that
01:13:49amiconn(via a modified plugin api)
01:14:28[IDC]Dragonyou may even notify disk useage, but not in USB mode
01:14:29amiconnI would also log the ata power on/off, but that alone doesn't tell me whether the disk is actually spinning
01:16:05[IDC]Dragonport AN3 use is unknown for recorders
01:16:06amiconnThe logging should of course interfere with normal operation as little as possible, so logging into (plugin) ram as much as possible, only fushing if it gets full
01:16:34amiconnOr better still, synchronize the flushing with other ata accesses
01:16:45[IDC]Dragonbut this doesn't seem to be the desired, undocumented current probe
01:16:45amiconn(that could be tricky)
01:17:10[IDC]Dragon(just tried)
01:18:59amiconnAN3 goes to 0x3FF if USB is connected
01:19:28[IDC]Dragonif you log only significant changes, with a timstamp, 32k may be sufficient
01:19:36einhirn[IDC]Dragon: It goes up from 6..7 to A..C if I press the "down"-Button in "View IO-Ports"
01:20:41amiconnMine shows ~0x024 when idle
01:20:48[IDC]DragonAN2 should be USB...
01:22:15amiconnAN2 does the same as AN3 for me...
01:23:59[IDC]Dragonwhat a waste
01:24:32amiconnThey don't seem to be directly connected though, the fluctuation without USB connected is different
01:24:48einhirnPerhaps they "want" to show 0 but are prohibited to do so by noise?
01:27:24amiconn[IDC]Dragon: Regarding the player, the archos firmware changes from "battery: charging" to battery: charged" in its charging screen after some time. I did not yet check whether this is onyl time controlled though, unrelated to the real charging state.
01:28:13einhirnanyway, I don't see much capacitor around the processor on the schematics - thats for the fluctuations...
01:29:37einhirnamiconn: Wasn't there this pin for the firmware to monitor charging state? Why shouldn't archos use it?
01:29:43*[IDC]Dragon says goodnight
01:29:49amiconnNite [IDC]Dragon
01:29:54 Part [IDC]Dragon
01:30:14*einhirn was too slow - again ;)
01:30:53*einhirn thinks about going to bed now, too
01:31:12einhirnGood Night everyone ;)
01:31:22 Part jyp ("poof!")
01:31:27amiconneinhirn: Not on the player
01:32:04amiconnThis is possible on the v2/fm though, as there is AN7 to monitor charging current.
01:32:19einhirnk. I didn't look into the schematics for the player...
01:32:25amiconnIf it goes below a certain threshold, the batteries are considered full
01:32:50amiconn(Charging of LiIon is entirely different from charging of NiMH)
01:33:05einhirnamiconn: Yup...
01:33:43amiconnI *will* check whether archos only uses a timer, I already started...
01:34:25einhirnThat and trying not to blow your LiIon up is the cause for always using a "charging processor" and not trying to do it yourself ;)
01:34:45 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
01:35:44amiconnIn fact, charging a LiIon is easier than charging NiMH, you only need a very precise voltage regulator
01:41:11lostlogicand a proper cutoff when the current drops below 3% of peak to avoid damaging the battery...
01:44:03amiconnThe cutoff is only an additional safety measure, depending on the regulated voltage.
01:45:15amiconnIf this is chosen a few mV "too high" in order to speed up charging, the cutoff is indeed necessary
01:45:28lostlogicamiconn: from what I've read about li-ion, holding the cell voltage over 4.05 can damage the battery...
01:45:31lostlogicI could of course be wrong
01:47:16amiconnHmm, maybe, I'd need to check myself. It could also be dependent on the exact type (LiIon vs. LiPoly)
01:47:43lostlogicthat could be...
01:47:49*amiconn checks the fm/v2 charger's datasheet
01:47:50lostlogicI got my infoz from
02:02:43amiconnThe fm/v2 charger chip doesn't do this shutdown. It can be shut down externally, but the "typical application" circuit doesn't use this feature
02:32:11 Quit XShocK (" HydraIRC -> <- The future of IRC")
02:37:27 Part amiconn
02:39:26 Join DMJC-L [0] (
02:40:43 Quit Trevmar (Read error: 60 (Operation timed out))
02:47:06 Quit cYmen ("leaving")
02:50:16***Saving seen data "./dancer.seen"
02:56:19 Join LinusN [0] (
02:56:57*LinusN is browsing directoried with rockbox on his iriver
02:57:39rasherwhat's next?
02:58:27LinusNsound, i guess
02:58:31rasherI'm getting all excited here everytime I hear about progress.. it appears to be a lot faster than what I had expected
02:58:31lImbushi LinusN
02:58:39lImbusbig congratz and kudos
02:58:51rasherAre you still the only only one who has flashed?
02:59:07rasherAnd yes, congrats indeed :)
02:59:40LinusNyes, i am the only one who has flashed
03:01:28preglowgreat, linus!!!
03:01:49preglowi'm multiple-exclamation-mark-excited!
03:03:10preglowwhat was wrong?
03:03:17LinusNthe hang?
03:03:34LinusNi think i called it "crash" last time...
03:03:51preglowno, you did say hang
03:03:52LinusNit was daniels fix in mpeg.c to make it compile
03:04:09LinusNthe mpeg thread terminated
03:04:29LinusNand the threads aren't supposed to terminate
03:04:31preglowand why does that cause a hang?
03:04:53preglowbut this is great news, bagder managed to compile libmad as well
03:04:54LinusNthey should yield() in a while() loop
03:08:05LinusNdamn, ata driver bug
03:08:58LinusNsoft reset doesn't work
03:14:59 Join XShocK [0] (
03:15:02LinusNah, easily fixed
03:15:07preglowso you can soft reset?
03:15:41preglowso: how long till you'll have tested the loader on another unit?
03:15:47preglowi'm more than a little tempted to flash my player right now
03:16:17LinusNthe m68k compiler aligns things a little differently than i'm used to
03:16:57preglowaligns what? structs?
03:17:09LinusNchar arrays
03:17:15preglowhow does that matter in what you're doing?
03:17:28LinusNthe stacks are char arrays
03:17:44LinusNand the compiler places then at odd addresses
03:17:59LinusNsloppy programming
03:18:21preglowwell, people do say that the 68k backend is a bit old and crummy
03:26:46 Join DMJC [0] (
03:36:00preglowwoot, libmad.a
03:36:04preglowi'll go to bed
03:37:20 Quit preglow ("off")
03:50:57 Join CrunchyWhiteMeat [0] (
03:57:40 Part CrunchyWhiteMeat
03:59:31 Join CrunchyWhiteMeat [0] (
04:03:59 Part CrunchyWhiteMeat
04:06:07 Quit DMJC ("Leaving")
04:09:58 Quit lImbus ("bed")
04:12:33 Join muesli_- [0] (
04:12:44muesli_-g'day mates
04:12:56LinusN"day"... :-)
04:13:16LinusNit's 4am in sweden :-)
04:13:28muesli_-same for germany ;)
04:13:56muesli_-4.13am to say it exactly
04:15:04muesli_-from which city are you coming from?
04:15:30muesli_-i have been in wasteros (dont ask me for spelling) years ago
04:16:04muesli_-yeah, it was around 50-100km west of stockholm
04:16:16muesli_-maybe even shorter
04:20:50 Quit QT (Read error: 113 (No route to host))
04:21:16rashergosh, it really *is* late :|
04:40:52muesli_-will roll into my bed...
04:41:02muesli_-n8 fellows...
04:41:12 Nick muesli_- is now known as muesli_zZ (
04:50:20***Saving seen data "./dancer.seen"
04:51:45LinusNtime to sleep
04:51:47 Part LinusN
05:00:42 Quit XShocK (" HydraIRC -> <-")
05:29:40SoulEataoh damn i missed linus
05:43:47lostlogic(hopefully this isn't distasteful) what is it with swedes named Linus being low level code gods?
05:44:13rasherActually the Other Linus isn't Swedish
05:44:39lostlogicFinish, but swedish speaking... yeah yeah yeah
05:45:25rasherAnd.. I doubt it :)
05:45:38lostlogicyou doubt which?
05:45:59rasherthat there "is something"
06:10:05 Join ashridah [0] (
06:31:27 Quit SoulEata ("CGI:IRC (EOF)")
06:50:21***Saving seen data "./dancer.seen"
07:14:12 Join ashridah_ [0] (
07:16:34 Quit ashridah (Read error: 110 (Connection timed out))
07:25:07 Quit ripnetuk (Read error: 60 (Operation timed out))
07:26:12 Nick ashridah_ is now known as ashridah (
07:33:30 Join Trevmar [0] (
07:33:36 Quit Trevmar (Client Quit)
08:02:53 Quit ashridah (Read error: 60 (Operation timed out))
08:06:11 Join XShocK [0] (
08:10:35 Quit XShocK (Client Quit)
08:15:20 Join ashridah [0] (
08:18:18 Quit gromit` (Remote closed the connection)
08:40:42 Join midk [0] (
08:42:32 Quit midk (Client Quit)
08:42:35 Join midk [0] (
08:50:24***Saving seen data "./dancer.seen"
09:24:57 Join Cassandra_ [0] (~christi@
09:25:29 Join ripnetuk [0] (
09:37:51 Join R3nTiL [0] (~zorroz@
09:38:39 Quit midk ("Leaving")
09:39:13 Join amiconn [0] (
09:41:59 Join midk [0] (
09:42:31 Join methangas [0] (
09:43:01 Quit Cassandra (Read error: 110 (Connection timed out))
10:04:46 Join LinusN [0] (
10:05:32amiconnhi LinusN
10:05:54amiconnThe iriver port seems to be progressing nicely. Congrats!
10:06:02LinusNi just installed the boot loader on my second H140, works perfectly
10:07:51LinusNnext task for me is to verify with different iriver firmware versions
10:09:28amiconnLinusN: (unrelated nag) What about flashing the oldplayer?
10:11:24LinusNok, i'll send it to you
10:11:42Cassandra_LinusN: I'd like to move most of clean_shutdown into firmware, so that the idle timer and sleep timer can call it too. Are you OK with that?
10:12:35LinusNof course. How would you do it?
10:13:04LinusNa firmware->app call to save settings?
10:13:15Cassandra_Just have a shutdown_hw routine that's called instead of calling power_off directly.
10:13:37 Join Marder [0] (
10:14:14Cassandra_Which would contain most of the old clean_shutdown code, which is just firmware functions.
10:17:11MarderCongrats Linus, for flashing your H140 (and the rest of your work of course) !
10:17:30MarderDo you consider the bootloader as safe enough now ?
10:17:46Cassandra_I'm really looking forware to putting Rockbox on my 140.
10:18:13MarderMe too..
10:18:52LinusNMarder: i think i do, but you should wait for me to put up tested binaries in the wiki
10:20:51MarderOk, that is really great news :)
10:21:48Cassandra_I'm in no hurry, at least not until we have MP3 playback working.
10:22:03Cassandra_By we, I of course mean you, Linus. ;)
10:22:49MarderWell, with a safe bootload, other developpers can work on it
10:22:54amiconnLinusN: Different thing regarding oldish archos jukebox: I don't know whether you've read the logs, but do you have an idea why my player doesn't seem to suffer from rld? It has the "right" hd...
10:22:55LinusNwell, i hope it will be "we", it's been pretty lonely
10:23:13LinusNamiconn: i dunno
10:23:32Cassandra_*nod* Well, the higher up it gets, the more people will feel competent to get involved hopefully.
10:23:50amiconnDo you remember where there were rld reports for players, or is it a recorder only issue?
10:24:11LinusNi *think* there has been reports from player owners as well
10:24:12Cassandra_Me, I'm OK at tinkering round the edges of Rockbox. I'm not sure I'd trust me to do something major like implement MP3 playback.
10:25:02LinusNCassandra_: btw, i really appreciate your work on the rtc/alarm stuff, it's something i've been wanting to do for a long time
10:25:24Cassandra_You're welcome. It's something I've wanted for a while.
10:25:47LinusNi see in the legs that you have the same approach as i
10:26:15Cassandra_Well, given the brokenness of the hardware, it seems like the best way to handle it.
10:26:34LinusNbrokenness, as in no irq?
10:27:12Cassandra_As in if you get an alarm while powered on, you can't turn the jukebox off.
10:27:58Cassandra_Otherwise it might be nice to say beep if you got an alarm while powered on, or something,
10:28:02LinusNyou can call it hw brokenness, but it's eailly fixed in hardware
10:28:51LinusNor did i miss something?
10:29:08LinusNi didnät know it was impossible to clear the alarm
10:29:22 Quit R3nTiL ()
10:29:31Cassandra_I don't know. The only fix I came up with was to implement a seperate alarm thread, which seemed kind of resource intensive.
10:29:35amiconnThe problem is that you need to notice it turned on
10:29:55LinusNamiconn: yes, you will have to poll the rtc
10:30:14LinusNbut we do that in the status code anyway, to display the time
10:30:32LinusNjust add a SYS_ALARM event
10:30:38Cassandra_That's useful to know.
10:30:44LinusNlet the status code broadcast the event
10:30:49amiconn...which only gets called _if_ there is a status display, e.g. not in plugins. Am I wrong?
10:31:01LinusNamiconn: you are right
10:31:23Cassandra_Ah. :( Then we're back to the separate thread.
10:31:40LinusNi have no problems with a separate thread
10:31:44Cassandra_Shame the iRiver has no RTC.
10:31:48LinusN300 has
10:32:07Cassandra_Oooh. Shiny.
10:32:19amiconnImho the power thread might do this as well.
10:32:31Cassandra_(Although I'm definitely not getting one of those until Rockbox is already running on it.
10:32:46LinusNamiconn: sure
10:33:23Cassandra_That's true, actually. Although then you have the problem of a callback to handle the alarm.
10:34:31amiconnIf the alarm flag gets set, do 2 things in the power thread: (1) Send the SYS_ALARM event. (2) Clear the flag immediately
10:34:34Cassandra_You probably want the application to do something if you receive an alarm while the power is on.
10:34:54amiconnYes, exactly, by handling the SYS_ALARM event
10:35:05Cassandra_Ah, there's an event queue somewhere? I missed that.
10:35:14LinusNjust like SYS_USB_INSERTED and SYS_POWEROFF
10:35:27Cassandra_Oh, right. Yes.
10:35:27amiconnIn fact this could also be useful on units which cannot power on from the alarm (unmodded v1 recorder)
10:35:57Cassandra_OK. Let me get the basic approach working first, then we can look at that.
10:35:58LinusNfor timed recordings
10:36:28amiconnThis way, you could use a v1 recorder for timer recording as well. Just leave it plugged to the charger, so that it doesn't shut down
10:42:30LinusNgotta go, cu folks
10:42:42 Part LinusN
10:42:43Cassandra_Oh, I have a slightly weird effect, btw. If I set the idle power off to 1 min and ROLO / F1 + boot into my test firmware, it still seems to take about 10 mins (possibly coincidentally my previous value) for the idle code to kick in. I've verified this with vanilla code, so my mods aren't causing it. Is this known weirdness with working from a non-flashed image?
10:43:31amiconnHmm, I don't know. I have idle poweroff set to 3 mins, and it always worked for me.
10:43:39amiconnLemme check...
10:46:58amiconnHmm, works both on the player and the v1. Did you press any buttons, or did you leave it at the resume? prompt?
10:48:19Cassandra_Left it sitting in the browser.
10:48:40 Join LinusN [0] (
10:48:48Cassandra_This is on a v2.
10:48:54LinusNCassandra_: are you sure that the settings aren't reset when you rolo?
10:49:17Cassandra_If I check the setting in the ROLOed f/w it still claims to be 1min.
10:49:21LinusNbtw, are you sure that the settings are properly saved?
10:49:22 Quit midk ("Leaving")
10:50:15 Part LinusN
10:50:20Cassandra_I think I often experience weirdness because my standard firmware is flashed but I work from a hd image for testing.
10:50:28***Saving seen data "./dancer.seen"
10:50:56Cassandra_6 mins and counting so far.
10:50:57amiconnHmm. This will cause settings problems if something witihn the settings bits changed
10:51:35amiconnI flash rather frequently when developing. Sometimes flashing > 10 versions a day.
10:51:37Cassandra_Ah, yes. That's probably it then. Since the daily builds have the start in recording screen option.
10:52:12amiconnIf it's only a setting added at the end, this shouldn't cause problems though
10:52:21Cassandra_It is. Hmm.
10:52:59amiconnWorks perfectly here, both with F1-boot ("-"-boot on player) and with rolo
10:53:25Cassandra_8 mins.
10:53:28amiconnHowever, my .ajz/.mod is the same version as flashed rockbox
10:54:59Cassandra_Yep - definitely powered off at 10 mins.
10:55:26Cassandra_I'm going to assume that's weirdness related to sick things I'm doing with Rockbox rather than a bug, I think.
10:57:24amiconnFrom time to time, I also get weird things in settings etc. A settings reset & reload of my favourite config file always fixed it.
10:58:08Cassandra_Makes sense.
10:58:26amiconnHmm, hold on, it seems I was wrong.... I'm confused.
10:59:48Cassandra_That makes two of us, I guess.
11:00:06amiconnAnyway, gotta go now. cu.
11:00:12Cassandra_Take care.
11:00:18 Part amiconn
11:00:53 Join LinusN [0] (
11:01:05LinusNCassandra_: are you really sure the settings are saved?
11:01:48LinusNhmm, they should be, as they are in rtc
11:03:29 Part LinusN
11:09:58Cassandra_I thought so, yes.
11:14:31 Join midk [0] (~midk@
11:23:04 Join midk_ [0] (
11:29:52 Quit midk (Read error: 60 (Operation timed out))
11:56:49 Join midk [0] (~midk@
11:59:12 Join ripnet [0] (
11:59:35ripnetMorning :) i see you have flashed your 140 Linus - Grrreeaat :)
12:00:14ripnetwould you be prepared to tell me the md5 of bootloader.bin and which firmware you injected? was it the 1.63EU version?
12:00:25ripnetand the md5 of the injected firmware?
12:00:55ripneti make bootloader.bin (cvs 5 mins ago) e1028ec522b6570daeb9b04a0cd27f5e
12:05:21 Quit Marder (Read error: 60 (Operation timed out))
12:09:02ripneteveryone must still be in bed after that late night last night :)
12:14:14 Quit midk_ (Read error: 110 (Connection timed out))
12:16:24 Quit ripnet ()
12:44:58 Join DMJC [0] (
12:50:32***Saving seen data "./dancer.seen"
12:55:53 Join bobTHC [0] (
12:55:59bobTHChi all!
13:10:54 Join amiconn [0] (
13:20:10 Join preglow [0] (
13:21:50preglowso, linus has flashed his other player, has he
13:21:58preglowi take that as good reason to flash my own
13:22:03muesli_zZmorning ladies
13:24:21DMJC-Llinus flashed both irivers?
13:25:04DMJC-Lmight flash mine and start dev work then
13:26:02 Join muesli|tarn [0] (
13:29:15DMJC-Lhas anyone got a prebuilt version of the crosscompiler?
13:31:14preglowDMJC-L: the logs do say so, yes, think i'll wait for him to confirm the procedure, then i'll flash it
13:31:20preglowDMJC-L: well, yes, for linux
13:31:48DMJC-LI'm running gentoo here
13:33:28preglowthat's exactly that i'm running as well
13:33:33preglowbut bullding yourself is really easy
13:34:38preglowbut sure, i can tar up the bins, just don't chop my head off when they appear to be making corrupt binaries ;)
13:44:57DMJC-Lwhich architecture are you using?
13:45:11DMJC-Lintel, amd... although i guess that won't matter
13:46:23 Quit muesli_zZ (Read error: 110 (Connection timed out))
13:47:33preglowright now, athlon64
13:47:48preglowshouldn't matter, no, it's not compiled with anything specific
13:47:54preglowi didn't ask for it, at least
13:48:10preglowahh, that's bs, btw, i compiled it on a duropn
13:48:24preglowi forgot it's my server we're talking about, not this machine
13:49:03DMJC-Lheh cool
13:49:08preglowthe bins are here if you want them:
13:49:09DMJC-Lmy old box was a duron
13:49:14DMJC-Lmy desktop is an athlon
13:49:28preglowmy desktop is a spanking new athlon64 3200+
13:52:21muesli|tarnspank my m**ey :D
13:52:33 Nick muesli|tarn is now known as muesli_ (
13:52:57preglowi refuse
13:58:13 Join LinusN [0] (
13:58:26LinusNmd5sum ihp_120.hex
13:58:26LinusN9639f6019b932b9d9bdcbbab476f6ec1 ihp_120.hex
13:58:36LinusNlatest cvs
14:00:03DMJC-Lthat flashes 140s properly?
14:01:08preglowi'll have a eat, then flash
14:01:18preglowthat's european firmware, yes?
14:03:03 Join quel|out [0] (~kvirc@
14:03:13 Nick quel|out is now known as quelsaruk (~kvirc@
14:04:48DMJC-Lmake[110]: Entering directory `/home/james/development/rockbox/rockbox/firmware'
14:04:59DMJC-Lshould I be getting this?
14:05:17amiconnLinusN: Interesting news - obviously ata poweroff does influence the occurence of rld. When I went out a short time ago, I deactivated ata poweroff. Then I got rld within 2 minutes after starting playback....
14:06:28amiconnHowever, I only got it once. I have an idea why this is, have to check the ata code.
14:06:40LinusNpreglow: yes, it's 1.3e
14:07:07 Join Digital007 [0] (
14:07:22 Quit Digital007 (Client Quit)
14:07:45DMJC-Lmake[1650]: Entering directory `/home/james/development/rockbox/rockbox/firmware'
14:07:46LinusNDMJC: "entering directory"? yes, that's normal
14:07:48amiconnCurrently, I'm looking into the player's volume/balance/treble/bass handling. I found a bug...
14:07:55DMJC-Llook at the number..
14:08:21DMJC-Land it keeps saying nothing to be done for all
14:08:27LinusNnot normal
14:08:46LinusNwhich target?
14:09:30LinusNin which dir do you type "make"?
14:10:57LinusNgotta go
14:11:03 Part LinusN
14:25:11preglowdidn't get the same md5sum
14:35:25HClhow goes?
14:35:34*HCl is on a wireless network he hacked into o.o
14:35:49preglowfree net!!
14:35:51HClbut since my irc goes over an ssh connection, its safely to talk about it here
14:36:02HCli stole the mac address of my target
14:36:07HClso i'm literally hitch hiking
14:36:08HClon his ip
14:36:22HClhe's using his laptop on the same ip as mine at the moment.
14:36:37HClwhats up? :)
14:37:05preglowplanning to flash my player
14:37:10HClwhat kind
14:37:25preglowbut the md5 i got doesn't match what linus gave before he ran away
14:37:28HClanything new?
14:37:39HCllinus compiles with 3.4.2, i think
14:37:47 Join lImbus [0] (
14:37:47HClwhatcha using?
14:37:58lImbushi all
14:38:22HCli'd say try 3.4.2
14:38:30HCland if that doesn't work, the recommended 3.3.4
14:38:48HClany progress on rockbox on iriver? aside from the cookie bootloader thing
14:39:19preglowit runs
14:39:23preglowrockbox runs
14:39:26preglowmenus and everything
14:39:38preglowso sound will probably be next
14:39:48HClso i could start on gameboy plugin?
14:40:10HClwell, obviously i want someone to test it on h140 first..
14:40:32 Join R3nTiL [0] (~zorroz@
14:40:51Bagderthe CPU is also clocked too low still
14:40:57muesli_do i undestand you correct that your implantating games on the iriver?
14:41:04preglowHCl: he did try it on his 140
14:41:11HClpreglow: did it work?
14:41:16preglowHCl: he said no problems
14:41:18lostlogicBagder: the clock speed onthe iRiver is software adjustable?
14:41:20HClpreglow: nice.
14:41:32muesli_how much does those games affect firmwire file size?
14:41:33HClso i could actually increase its clock
14:41:38HClif i need more speed for the gameboy emu?
14:41:38lostlogicBagder: sweet, now we just need to be able to software control it :-D
14:42:09ashridahlostlogic: it's done by the firmware to reduce power consumption.
14:42:16ashridahthere's really no point in overclocking it for no reason
14:42:18lostlogicashridah: oh, I figured that :)
14:42:30HClexcept if you want to run gameboy P
14:42:31HCl :p
14:42:46muesli_overclocking an iriver LOL ;)
14:42:51lostlogicashridah: which would explain why different decoding tasks require more battery power on the H3x0 series
14:43:10muesli_how much does those games affect firmware file size?
14:43:16ashridahlostlogic: oggs tend to use battery faster than mp3 in general
14:43:25Bagdermuesli_: they're plugins, loaded on demand
14:43:38lostlogicashridah: which I assume is algorithmic complexity speaking?
14:43:51ashridahlostlogic: pretty much
14:44:06ashridahyou'd calculate the required frequency plus a margin based on the bitrate
14:44:32muesli_how much memory is reserved for the firmware? i mean, is there a limit?
14:45:11Bagderthe firmware is right now ~150K
14:45:17Bagderon disk
14:45:49HClthe iriver has 32mb ram
14:45:57HClwhich is more than enough
14:46:02HClfor almost anything o.o;
14:46:15HCli wonder if we can make it run windows 98 xD
14:46:22HClnah, but linux should work o.o
14:46:34muesli_yes, but this ram isnt reserved for caching media files?
14:46:40ashridahHCl: uh. does the coldfire cpu have an MMU?
14:46:43preglowmuesli_: not reserved, no, you can use as much as you like
14:46:49HClashridah: i have no idea o.o
14:46:51ashridahmuesli_: the stock firmware decodes itself into ram
14:46:53preglowi don't think coldfires have mmu's
14:47:14preglowdamn, the stock 68k didn't even have an mmu
14:47:15HClso you can't do paging, thats not a disaster :x
14:47:15ashridahit'll be faster to execute than the flash will, and the images are encoded anyway, apparently.
14:47:26muesli_playing music and having a game at the same time would be kewl :-)
14:47:30ashridahHCl: uh. win98 kinda requires protected mode, last i checked.
14:47:35HClashridah: ok :P
14:47:36ashridahwin3.1 maybe
14:47:39HClit was a joke anyways
14:47:39preglowyes, and it's x86 code :VV
14:47:41Bagdermuesli_: it does that
14:47:44HClohyea xD
14:47:53 Join ripnet [0] (
14:47:55ashridahpreglow: no-one was suggesting it'd run FAST
14:47:57ripnetpreglow - you there?
14:48:00preglowripnet: sure
14:48:15HClmuesli_: i don't think (at least for gameboy) that the iriver will have much cpu left to play mp3
14:48:20ripnetdid you get be0bf0e1df367e5cffd9e27eff988ef1 new.hex ?
14:48:40HClyou'll want to check checksums for bootloader.bin
14:48:53ripnetmore important to check new.hex :)
14:49:00HClokay :x
14:49:06muesli_HCl: yeah :-/ but even the possibility sounds nice ;-)
14:49:08preglowbe0bf0e1df367e5cffd9e27eff988ef1 ../build/rockbox/build/new.hex
14:49:22ripnetwe get the same... we are both using 4.3.4
14:49:23HClsounds good
14:49:33preglowripnet: i'm building 3.4.2 now, to test if i get the same as linus
14:49:54ripnetdammit! i build 3.4.3 and 3.3.4 and thought i had all bases covered :)
14:50:02preglowif it's the same, i'm flashing this mutha
14:50:06ripnetme too
14:50:19HClripnet: do you have h120 or h140
14:50:34***Saving seen data "./dancer.seen"
14:51:32preglowahh, great, the gcc build vomited
14:51:44lostlogicI wonder why it allows higher bitrate oggs than mp3s then...
14:52:13preglowvorbis files can have more or less any bitrate they want, mp3s are max 320
14:52:57*ripnet strokes my 140 and wonders if it will be dead later :)
14:53:14ripnethows the gcc build going preglow?
14:53:20preglowripnet: it crashed, but i'm working on it
14:53:58preglowi deleted the cvs binutils
14:54:15preglowhell, i'm more than a little tempted to just go ahead
14:54:17ripneti deleted all my source (for toolchain) as well.. its MASSIVE
14:54:20ripnetme too
14:54:34ripnetshame Linus isnt around to send us the known good bootloader.bin
14:54:47HClwhat was his md5?
14:54:48preglowbut i'll gladly invest some more minutes to get it right
14:54:50HCllet me try my build
14:54:51preglowi'll just go from scratch
14:54:59preglowHCl: what gcc have you got?
14:55:08preglow13:58 < LinusN> 9639f6019b932b9d9bdcbbab476f6ec1 ihp_120.hex
14:55:10ripnethe posted 9639f6019b932b9d9bdcbbab476f6ec1 ihp_120.hex
14:55:12lostlogicOk, this is getting irritating, whatver IO scheduler I'm using on this computer, it is scheduling my interractive IOs like literally minutes after the noninteractive process of copying via wireless network. so as I type this nothing is showing up, because my IRC client is in IO wait for a couple minutes.
14:55:30ripnetim gonna go eat... talk to you in a bit
14:55:30preglowlostlogic: what os?
14:56:48lostlogicpreglow: linux-2.6
14:56:54lostlogicLinux lost 2.6.9 #2 SMP Wed Oct 27 20:33:21 CDT 2004 i686 AMD Athlon(tm) MP 2100+ AuthenticAMD GNU/Linux
14:57:26lostlogicI blame all the IO scheduler stuff that AKPM et al have done lately.
14:58:04preglowwell, 2.6 isn't a stable tree
14:58:25lostlogicpreglow: it is the stable tree... (even numbers are)
14:59:08ashridahlostlogic: you haven't been reading up about linus's policies recently have you? :)
14:59:52preglowlostlogic: ashridah has a point, they don't see it as their responsability to keep the tree stable anymore
14:59:54lostlogicashridah: You mean WRT not branching despite intrusive changes? Oh, I'm up on it... I used to be the kernel patchset guy for Genwoo...
15:00:39lostlogicso, I made it let me talk and copy at the same time... by starting a compile in another terminal. LInux is dumb. (s/Genwoo/Gentoo/)
15:01:43preglowhaha, no, that particular kernel is dumb
15:01:50preglowthe linux scheduler is pretty good, most of the time
15:02:18lostlogichehe, I know... that was just way up there on the irritation scale... some less forgiving IRC networks disconnected me for it
15:02:53izzy_preglow: I got the same md5sum as you did, with 3.4.2 gcc
15:03:15ripnetare we using the same original firmware as Linus?
15:03:35ripneti downloaded from that linkon the wiki
15:03:39preglowwell, then i think i'll wait a bit
15:03:45preglowi don't want to brick it quite yet
15:03:57ripnetyes... ideally we would lift a working bootloader.bin from Linus
15:04:06HClwhy do i have a different md5sum for my iriver firmware?
15:04:21preglowi don't know, broken compiler?
15:04:24HCl14488347a171480c63c94bc7b885225d ihp_120.hex
15:04:25HCl14488347a171480c63c94bc7b885225d ihp_120.hex
15:04:30HClno, this is the original..
15:04:37HCllet me unpack it again..
15:04:37preglowwell, i haven't tested THAT
15:04:55HClohh. the md5sum linus gave was of the patched?
15:05:08ripnetyes thats what i got for the original
15:05:13preglowyou've got a correct firmware
15:05:26HClhave either of you tried 3.4.2?
15:05:30HCli have 3.3.4
15:05:31izzy_I have the same
15:05:34ripnetme 2
15:05:35izzy_I tried with 3.4.2
15:05:47preglowHCl: try with 3.3.4, please
15:05:49HCl66541d4ea11332464dedd68694f357e3 new.hex
15:05:49preglowhe might have that
15:05:56HCl :/
15:06:07HClwhat about md5sums of the bins?
15:06:19HCl2826ba6cb6168f5ca99fd30c12360a03 bootloader.bin
15:06:19preglowe1028ec522b6570daeb9b04a0cd27f5e bootloader.bin
15:06:28HCl :x
15:06:33HClbut this is 3.3.4...
15:06:38izzy_e1028ec522b6570daeb9b04a0cd27f5e bootloader.bin
15:06:39preglowyes, but with the different compiler version, that's no surprise
15:06:43preglowthey're correct
15:06:45ripnete1028ec522b6570daeb9b04a0cd27f5e bootloader.bin on 3.4.3
15:06:48preglowthere's just something else going on
15:06:50preglowi'll wait for linus
15:07:13lostlogic*tries out his cross compiler at this
15:07:30izzy_I guess we are a bit impatient ;-)
15:07:33HCl :P
15:07:34HClmaybe :P
15:07:57ripneta little
15:08:28preglowhaha, you think?
15:08:32preglowbut yes, i've got other stuff to do
15:08:40preglowwhich quite lukcily also involves coding
15:08:44HClwhat do you guys think, should i generate a little bit more traffic in order to be less bored? or not take the risk of getting noticed by the owner of this network?
15:08:45preglowso brb
15:09:07dwihnoif the charger polarity is reversed for a couple of seconds (thank you, gf), and the unit starts to smoke, do you guys think the unit will be damaged in any way? seems to work at the moment.
15:09:32HClsmoke, and it still works?
15:09:48ashridahdwihno: can't have been essential magic smoke then
15:10:24muesli_harry potter is greating :d
15:10:37dwihnohci, yes
15:12:08 Join Cassandra [0] (~christi@
15:12:12dwihnoash, thanks for being serious
15:12:30preglowwhat is it with girlfriends and iriver chargers? that's the third one i know of that managed that
15:12:38preglowboth of the other ones ended up never starting again
15:12:50HClhow do you even manage that...?
15:12:59preglowby not using the original charger,i guess
15:13:13DMJC-LI never use my charger..
15:13:21HClwhat do you use?
15:13:27DMJC-Lhacked a usb cable to do it
15:13:32DMJC-Lworks brilliantly
15:13:38HClaha o.o;
15:13:58CassandraWouldn't that be slow?
15:14:39DMJC-Lusb carries 5volt DC charge..
15:14:49CassandraBecause your max current on USB is 500mA
15:15:11DMJC-LI don't care about speed, it recharges while I sleep
15:15:30DMJC-Ldoesn't seem to take long though
15:15:47CassandraOooh-hoo, that's shiny.
15:16:13CassandraI just got the sleep timer and the wake up alarm working together.
15:17:11ripnetso, have we established that 4.3.2 and 3.3.4 give the same bootloader.bin, but we end up with a different .hex from Linus?
15:17:21ripnet3.4.2 and 3.4.3 ?
15:17:38dwihnocass, you did the alarm mod?
15:18:22preglowripnet: correct
15:18:32preglowripnet: seems 3.3.4 gives different md5 as well
15:18:54preglowso either he's done some additional magic, or we're idiots :P
15:20:21CassandraDamn - how do I get to build my own UCL file?
15:20:28ripnethmmm... i already knew that 3.3.4 and 3.4.3 gave different md5s
15:20:50ripnethave we checked that nothing has gone into cvs since Linus did his?
15:21:18preglowbe0bf0e1df367e5cffd9e27eff988ef1 new.hex
15:21:18CassandraThere have been cvs changes since then, yes.
15:22:26ripnetwell at least we know that those 2 3.4.x versions give the same md5
15:22:28preglowi just updated both rockbox and bootloader
15:22:49ripnetwe await the return of our leader for guidance :)
15:23:44 Quit ashridah ("sleep")
15:25:23 Quit Cassandra_ (Read error: 110 (Connection timed out))
15:26:54ripnetcan we book out from cvs 'as it was then'?
15:27:07ripneti know you can do -d to do date, but not ssure about time
15:27:50preglowyes, you can, but i don't know about cvs, i'm more used to svn now
15:28:13lostlogicripnet: yes
15:28:22ripnetactually, i just thougt of something... when you do version in rockbox on a cvs version, it tells you what date it was booked from cvs. Therefore, I reckon there is a cvs version tag in the object code somewhere, which is causing the md5 difference
15:28:27lostlogiccvs up -D"Datespec"
15:29:07preglowcvs version tags should not make it to the object code?
15:29:27 Quit DMJC-L ("Leaving")
15:29:29ripnethow does it do the 'cvs 20040404' thing then?
15:29:42preglowlike that
15:29:49preglowthat might be so
15:30:05lostlogicif you put $Header: $ in your CVS code, the CVS header info is put there...
15:30:18lostlogiccan also do $id: $ and some other ones
15:41:23preglowwell, i can't find any cvs string in the object files
15:42:08ripnetmaybe thats not it... looks like only a couple of souce files have changed this afternoon
15:42:36ripnetsoundmenu and mp3playback... do they compile into the boot?
15:43:28ripnetmp3playback.c seems to
15:43:57ripnetso im guessing if we book out mp3playback from yesterday, we get the same md5
15:45:34preglowwell, try it, if that's the case, then i'll just smack the patched firmware i've got in
15:46:28ripnethmmm... unless i got the syntax for the cvs wrong, the file hasnt in fact changed
15:47:25ripnetor maybe ive still got the old one
15:47:31ripnetrebuilding from cvs
15:48:14lostlogicrevision 1.38
15:48:14lostlogicdate: 2005-02-06 07:47:21 -0600; author: amiconn; state: Exp; lines: +31 -40
15:48:19lostlogicon mp3_playback.c
15:48:36ripnetyes, i was building from old(er) cvs anyway... the one altered on the 3rd
15:48:36amiconnI changed only player specific code
15:48:58ripnetthat'll be why then...
15:49:18preglowlinus said 'latest cvs' anyway
15:49:30preglowthe last changes took place before he sais that
15:49:44preglowthat might be wrong, i'm gmt+1
15:50:39CassandraIt's possible some of my changes would've affected the iRiver build.
15:50:47CassandraI can't remember what I did when now.
15:51:25ripnetthey are all 10:16 this morning or before
15:51:29 Join linuxstb [0] (
15:51:40ripnetim going to switch off my computer, and go out before I get tempted to do something stupid :) see you all later
15:51:41 Part amiconn
15:53:01 Join muz [0] (
15:53:34muzhey does any one have a picture of rockbox running the iriver ( like a new one from recent developments)
15:53:48CassandraNot seen one, sorry.
15:53:56CassandraAsk Linus when he turns up again.
15:55:06muzit cant be too much longer now till rockbox is finished for the iriver, i mean once some formats are done a little testing....
15:55:09preglowno, but i bet you'll se a ton once we get it up later
15:55:38preglowmuz: might be, depends on how hard it'll be to port the codecs
15:55:54muzso linus is still the only one to have flashed
15:56:20preglowi'm about to flash, but need to talk to linus first
15:56:30preglowalso a couple of others are about to flash
15:56:34muzabout md5s?
15:56:43muesli_how do you flash back?
15:56:59preglowwell, by booting the original firmware and just flash the old one?
15:57:05lostlogic:-( my binutils for m68k appear to be broken
15:57:19muzdo things like config menus come up or is just a dir browser
15:57:37lostlogicLinus said all menus and directory listing worked
15:57:38preglowmuz: dir browsers, config menus
15:58:11muzso when a file is played does it just crash
15:58:11 Join CrunchyWhiteMeat [0] (
15:58:41 Join Lynx_ [0] (HydraIRC@
15:59:08preglowit can't play files
15:59:10preglowthere are no codecs
15:59:31CassandraBear in mind that a lot of hardware support is missing. It's still not near a functional build.
15:59:43preglowsound support isn't even programmed yet
15:59:49Cassandra(But cool nonetheless)
16:00:28lostlogicI guess the nice thing about the iriver is that it is a general purpose CPU and has enough ram to do stuff... something the archos can really only dream of...
16:01:17linuxstbHi all - just wanted to stick my head up and say that I'm also wiating patiently in the queue to flash my H140 - I'm getting the same md5sums for my "new.hex" as others have posted today (but not the same as Linus's).
16:01:30CassandraOh the SH1 does pretty well. You can play minesweeper on it. ;)
16:02:00muzlostlogic: does that mean that better plugins will be available
16:02:48CassandraIt's possible rockbox will do its own ID3 indexing as a plugin, for example.
16:02:56CassandraYou could never do that on an SH1.
16:03:55lostlogicmuz: someone was just talking about doing a gameboy emulator... so I'd guess so... 32M of ram is a lot to play with.
16:04:02preglowyes, as is 140 mhz
16:04:18muzwhats the archos
16:04:28preglowCassandra: rockbox has minesweeper????
16:04:42preglowthat's great, i love minesweeper
16:04:43CassandraAnd klondike solitaire.
16:04:47preglowonly windows game i can stand
16:04:57muesli_i am keen on salitaire :)
16:05:19muzu shud have 2 player games one using main unit other usin remote
16:05:26CassandraI think the plugins are cool, but I don't really use them much.
16:05:33lostlogicthe archos is 11-12mhz
16:05:40CassandraI pretty much want my MP3 player to play music at me.
16:06:15muzotf playlists, flac support, and vidoes are quite cool for me
16:06:40muzwhere can i find pics/videos of rockbox runing ?
16:06:49muzon an archos or nething?
16:06:59muesli_gapless is what i am yearning to..
16:07:02CassandraI could live with flac and ogg, yes. But that's codecs rather than plugins.
16:07:17Cassandramuz: The manual is full of screenshots. Would that help?
16:07:42lostlogicmuesli_: yeah, gapless... should be possible on iRiver... just by moving buffer pointers around... probably not even need two decode threads to do it.
16:07:44muzis there a pic of a video bein played
16:08:14CassandraI don't think so. Uses the greyscale engine, so it's kind of hard unless you actually take a photo.
16:08:52muzyea thats wat i mean is there a photo
16:09:43CassandraNot that I know of. I *think* there's a photo of the greyscale engine on the Mandelbrot plugin page, but that's about the closest you'll get.
16:09:52CassandraOn the wiki.
16:10:00CassandraWoohoo. It's confirmed.
16:10:14CassandraI just used my Jukebox as an alarm clock.
16:10:30preglowhah, congrats
16:10:40Cassandra(Not so useful for the iRiver, but cool for Archos users)
16:10:43preglowtoo bad the h120 doesn't have an rtc
16:10:50preglowCassandra: the h3x0 has rtc's
16:11:03preglowi'd love to use my h120 as an alarm clock
16:11:04muzill get over that
16:11:07Cassandrapreglow: So I'd heard.
16:11:20preglowi positvely hate getting woken up by something that's not music
16:11:28muesli_lostlogic: why never couldnt iriver nor apple nor others implant gapless? its so hard or what?
16:11:34lostlogicthe h3 has an rtc? whee!
16:12:07lostlogicmuesli_: it is a bunch more code than they want to write is my feeling... or maybe it is harder than I'm giving it credit for...
16:12:10muzis there any chance i cud watch tv episode on rockbox
16:12:12preglowmuesli_: no, it's not hard
16:12:28preglowmuesli_: their file buffering engine would have to be partially rewritten, that's more or less why they're not doing it
16:12:52muesli_so its a question of efforts?
16:12:57preglowand, it's not much in demand
16:13:15preglowit seems most people are used to gaps or don't listen to music that has seamless transitions
16:13:26muesli_obviously quit dumb..thats the feature 99% of all users are yearning to..
16:13:41preglowthat's an overstatement, heh
16:13:46preglowabout 10% if you ask me
16:14:02muzdoes rockbox have a browser like the ipod
16:14:09muzone which is as easy to use and fast
16:14:35preglowhow can a browser not be easy to use? :P
16:14:42muesli_i love the folder structure as in windows
16:14:47preglowso do i
16:14:56muesli_nothing to improve for me
16:15:05muzirivers browser is awful
16:15:22muzhalf the screen is a stupid logo
16:15:25 Quit Lynx_awy (Connection timed out)
16:15:25 Nick Lynx_ is now known as Lynx_awy (HydraIRC@
16:15:35muzand there is no accelerated scrolling
16:15:53muesli_yepp, that would be fine
16:16:06muesli_but which logo do you mean?
16:16:17muzthe filetree picture
16:16:27CassandraHmm if you declare a C function static, that makes it local to the current file, yes?
16:16:34muzit says files and its so pointless to have it there
16:16:40preglowCassandra: yes
16:16:46preglowCassandra: it'll even perhaps be inlined
16:16:48muesli_muz you mean on the left?
16:16:51CassandraThat's what I thought.
16:17:05muesli_yepp, absolutely needless
16:17:31muzbut i read about an id3 browser
16:17:42muzand that you cant actually play files from it
16:18:11muesli_you mean the data base?
16:18:34muzim talking about on rockbox
16:18:37muesli_dunno, never used that stuff ;)
16:18:58CassandraI really need to have a look at the database stuff. It sounds cool.
16:19:03muesli_you was asking for pictures, are there any?
16:21:39muzis the tagdatabase good on rockbox?
16:22:37CassandraIt's a fairly new feature.
16:22:43CassandraI've never used it.
16:23:03muzit seems quite useful
16:23:14Cassandra(I have my MP3s laid out in a directory structure by artist and album anyway.
16:23:40muzi mean my music collection has loads of compilation albums, so if i can choose just on artist it would help loads
16:24:41 Quit muz ("CGI:IRC")
16:30:36 Join gromit` [0] (~gromit`
16:31:47 Join webguest87 [0] (
16:31:56lostlogicpreemptable kernel is off on this kernel. so long writes (16M over 802.11G NFS, sync) cannot be interrupted by a small write. causes my current pausing problem (for anyone who was reading that earlier)
16:33:16preglownot very impressive
16:33:43lostlogicpreglow: should it have been?
16:34:11preglowwell, i can't say i think that is the behaviour i'd have wanted
16:35:20lostlogicpreglow: yeah, but it's my own damn fault for doing synchronous NFS with a wsize of 16384 over 802.11G and turnign preempt off. *shrug* next kernel I'll have preempt on, and use 2048 wsize.
16:36:25 Quit webguest87 (Client Quit)
16:40:48preglowi think i'll try installing amd64 debian
16:41:15CassandraYou might find ubuntu easier.
16:41:24preglowany particular reasons?
16:41:50CassandraWell, there is a supported ubuntu release for amd64, and it's just Debian under the cover anyway.
16:42:10preglowhaven't got a working cd burger right now, so i haven't got much of a choice
16:42:17preglowcd burger, yes
16:42:19preglowcd burger
16:42:23preglowlet's try again, cd burner
16:42:55lostlogicGentoo Gentoo Gentoo :):)
16:43:15preglowlostlogic: i need this done quick, heh, i don't want to spend several hours compiling
16:43:23preglowmy server is gentoo, i love it, but not when just testing
16:43:24lostlogicfine :-D
16:43:41lostlogicall my boxen run Gentoo, but that's because I dev for Gentoo :)
16:45:06preglowthen again, it would be fun to see this mad monster of a cpu chug away at the source files
16:45:37lostlogicdistcc it with your other box :) −− that's what I love seeing... all 4 of my CPUs (3 machines) working on compiling a kernel or something :)
16:46:06preglowmy other box has stability problems, so i try not to pressure it too much
16:46:20preglowit reboots every five days or so
16:46:31preglowthink it's a mother board problem
16:46:50preglowthough i'll do a memtest on it some day soon
16:47:58 Quit R3nTiL ()
16:48:44lostlogicone of these days I need to get an amd64 box... but 2x2100mp is serving well enough for now.
16:49:10lostlogicwhat amd64 do you have?
16:49:17preglowathlon64 3200+
16:49:21preglowthe 1 meg level2 cache version
16:49:27lostlogic939 or 754?
16:49:36preglowdidn't want to ruin myself
16:50:37***Saving seen data "./dancer.seen"
16:51:34lostlogiccool −− not that 939 makes much diff when you have 1M l2
16:53:24preglowit should mean less, at least
16:55:07 Join cYmen [0] (
17:10:31 Join LinusN [0] (
17:10:35 Join ze__ [0] (
17:10:53 Quit ze (Nick collision from services.)
17:11:01 Nick ze__ is now known as ze (
17:12:23 Nick ze is now known as ze_test (
17:14:14 Part LinusN
17:15:37 Join ze_test_ [0] (
17:15:39 Nick ze_test_ is now known as ze (
17:28:14CassandraHi Linus
17:28:23CassandraWe have a working alarm clock! ;)
17:28:46CassandraAh, damn.
17:28:51preglowdamn inded
17:29:40 Quit ze_test (Read error: 110 (Connection timed out))
17:42:44 Join thegeek [0] (
17:48:17 Join Marder [0] (
17:57:03MarderDarn, I get a different md5 from everybody else's for the iRiver bootloader.bin
17:58:29linuxstbMarder: Using gcc-3.4.2 and today's cvs of binutils, I'm getting be0bf0e1df367e5cffd9e27eff988ef1 for "new.hex". This seems to match a couple of other peoiple here, but not Linus.
17:59:24MarderWhich gcc is it ? The "regular" one or the m68k-elf one ?
18:00:26linuxstbI compiled gcc-3.4.2 from source using "−−target=m68k-elf"
18:01:17preglowm68k-elf one, doesn't make sense to use any other one
18:01:17linuxstbI'm also using today's CVS of rockbox - Linus checked in some changes at about 3am this morning.
18:02:13 Join amiconn [0] (
18:02:43linuxstbMarder: what is the md5sum of your bootloader.bin?
18:03:50MarderI compiled the latest binutils, but my gcc still seems to be 3.3.5. Where do you get 3.4.2 ?
18:04:15preglowfrom gnu?
18:04:18preglowit's on the siute
18:05:09MarderThx, cygwin site only lists 3.4.1.
18:05:38preglowahh, i don't use cygwin
18:05:51MarderFor the records, I get ee4559d6164f797b9248ac16c1c3d04e for bootloader.bin
18:05:51 Join Yono [0] (
18:06:17preglowi don't
18:06:26preglowe1028ec522b6570daeb9b04a0cd27f5e bootloader.bin
18:06:43linuxstbIt shouldn't make any difference that you're using cygwin. You need to get the source for 3.4.2 and configure it as a cross-compiler. I'm assuming cygwin doesn't have pre-packaged cross compilers.
18:07:14linuxstbpreglow: same here for the bootloader.bin md5sum
18:09:50MarderOk, I will try ot get 3.4.2 to compile. Thx
18:21:29ripneti thought we established that 3.4.2 and 3.4.3 gave the same md5?
18:22:46izzy_yep, I have 3.4.2
18:25:21linuxstbWhat features are people planning on working on once they get the bootloader flashed? FLAC decoding will be my first contribution (unless anyone else is working on it).
18:26:18preglowlinuxstb: audio and libmad, i guess
18:26:20preglowi don't really know
18:26:37preglowlibmad does compile, but it should be adapted to use the emac unit on the coldfire
18:26:57linuxstbSorry, but what's the emac unit?
18:27:52preglowextended multiply accumulate
18:28:04preglowif you know dsp, you should know that's useful
18:28:14preglowif not, it's an extended precision unit to do basic dsp stuff
18:28:35linuxstbIs there documentation and compiler support available?
18:29:13preglowcompiler support: no, documentation: yes, lots
18:30:53linuxstbJust found this link: - Motorola have developed an MP3 decoder that works in 37kbyte of RAM and needs 19MHz on a ColdFireV2
18:31:34preglowi know, but that doesn't help us much
18:31:58preglowthat's probably pure assembler
18:32:06linuxstbThey also have a realtime mp3 encoder... Yes - obviously no use to Rockbox, but it's a benchmark to aim for.
18:32:25preglowa pretty unattainable one, i'd think
18:32:29preglowbut we'll see
18:33:09lostlogicwhy is realtime mp3 encoder useless for rockbox?
18:33:15 Join Rhunak [0] (
18:33:29ripnetthe fact one exists is useless
18:33:29preglowlostlogic: not the concept, just that particular implementation
18:33:33ripnetas its not gpl (i assume)
18:33:35lostlogicoh, gotcha
18:33:47ripnetif it were gpl, it would be perfect
18:33:52preglowrealtime mp3 encoders aren't very good unless they're hardware anyway
18:34:01ripnetarent they?
18:34:04preglowthe stock one that comes in the h120 sucks immensly
18:34:13ripneti dunno, ive never recorded on it
18:35:27preglowi wouldn't consider recording anything but wav
18:35:31preglowor at least flac
18:35:44Rhunakflac would be amazing
18:35:54lostlogicI'm confused, I've heard great thinga bout the stock mp3 encoder on the H series... being used for all kinds of make-shift studio recording
18:35:56preglowlinuxstb: how's things with flac and no floating point unit?
18:36:02ripnetwould be cool, but surely 40gigs is enough to use wav?
18:36:12ripnetthat would give you, what, 60 hours recording?
18:36:27Rhunakif it's empty, yes :D
18:36:47preglowit'd drain its batteries long before that
18:37:06ripnetunless you had it on apower source
18:37:55linuxstbpreglow: I think only the encoder needs floating point.
18:37:56 Join Yono_ [0] (
18:38:22preglowlinuxstb: does it truly _need_ it? can't it be rewritten to use fixed point?
18:38:27ripnetif a architeture is missing floating point, doesnt the compiler make up the gap with software routine?
18:38:36preglowripnet: if you want to, but that's extremely slow
18:38:41preglowripnet: better to use fixed point from the start
18:39:00ripnetextrememly as in too slow to work at all on our coldfire?
18:39:17preglowbut why do that anyway?
18:39:24preglowit would drain batteries at an alarming rate
18:39:38preglowand fixed point is more or less just as easy, as long as you know what you're doing
18:39:42ripnetmy style of programming is to get it working, and optimize later...
18:39:58linuxstbBut the good news is that libmad is fixed point, so that's already perfect for us.
18:40:04preglowthat's not a problem
18:40:09preglownor for tremor, which is also already fixed point
18:40:26preglowthere a fixed point musepack implementation as well
18:40:37ripnetim not familar with musepack
18:40:40lostlogicfixed point ogg encoding is what I want
18:41:31preglowlostlogic: that's not yet written
18:41:56linuxstbMost codecs are designed to be hard (i.e. slow) to encode, but easy (i.e. fast) to decode. So we are going to be struggling to get any realtime encoder for the iRiver.
18:42:19preglowand for me it's not even a priority
18:42:23lostlogicnod nod nod
18:42:31preglowif i'm going to encode anything, i'll use the best encoder i can damn well get my hands on
18:42:35preglowwhich would be lame
18:42:47preglowbut flac recording support would be awsome
18:42:51ripnetwindows media encoder
18:42:54ripnetjust joking
18:42:56preglowlostlogic: of course, for some reason i was thinking mp3, heh
18:43:34linuxstbI think recording to wav will always be the most reliable option. But we could have a non-realtime flac encoder that encoded to flac after the recoridng finishes.
18:44:07ripnetcan we vary the clock speed of the coldfire during use?
18:44:16ripnetor do we have to fix it as we init?
18:44:16preglowflac encoding isn't very intensive, so might be an option for realtime
18:44:27preglowmp3 encoding sure as hell is more intensive, and they've managed that on the iriver
18:44:49linuxstbpreglow: but at what quality?
18:44:56preglowlinuxstb: lousy, yes, but still
18:46:35linuxstbAnyway, decoding is obviously the first priority. My first test with flac will simply be to confirm (I hope) that the iRiver is fast enough to decode in realtime.
18:46:47preglowi don't think that'll be a problem
18:46:49 Join XShocK [0] (
18:47:10preglowbut you can start with that right away, you'll need to fixed point-ify it
18:47:17preglowand that you can do on an ordinary computer
18:47:37preglowif you depend on the floating point emulation, it most certainly will not be able to do it realtime
18:48:18linuxstbI'm only talking about the decoder, which (AFAIK) doesn't need floating point.
18:48:38linuxstbBut yes, I'm working on flac at the moment.
18:50:05preglowahh, of course
18:50:08preglowit's lossless
18:50:15preglowit might very well be integer based already, then
18:50:38***Saving seen data "./dancer.seen"
18:51:52 Quit Yono (Read error: 110 (Connection timed out))
18:51:53 Nick Yono_ is now known as Yono (
18:52:17preglowand it seems the only fpu part in the encoder, is the lpc part
18:52:26preglowwhich should be doable fixed point
18:53:35preglowbut i'll go install a linux, later
18:53:38 Quit preglow ("wop")
19:02:57 Join preglow [0] (
19:03:03preglowahhh, faulty installation media
19:12:17 Nick lImbus is now known as lImbuAWAY (
19:12:34 Quit Yono ("Leaving")
19:33:17XShocKwhat are the differences between US and EU version of the firmware in ihp1x0 ?
19:33:17 Join Stryke` [0] (~Chairman8@resnet-241-86.resnet.UMBC.EDU)
19:36:31muesli_imho tuner region settings and specific language chars for correct id tags only
19:41:42XShocKgood then
19:41:43 Quit Marder (Read error: 60 (Operation timed out))
19:49:00linuxstbMy compilation of libFLAC isn't going very quickly - for a start, libFLAC is using Stream I/O (FILE*, fopen, fread etc), but Rockbox has only implemented file I/O (open, read, write). Lots of other missing library functions as well (e.g. stat).
19:50:06linuxstbI would have prefered not to have had to hack the libFLAC sources, but it seems it will be necessary.
19:50:09pregloware you sure there isn't an api where you just feed it data raw?
19:50:14 Quit ripnet (Read error: 60 (Operation timed out))
19:50:35preglowit has to be that way, rockbox itself has to take care of the file buffering
19:51:52preglowso all file io should probably be ripped out of it, if there isn't another api
19:51:56 Part amiconn
19:52:18linuxstbpreglow: good point, I'll check the API
19:55:07linuxstbpreglow: yes, there is a file decoder, which depends on the seekable stream decoder, which depends on the stream encoder. So I should strip out the file decoder module.
19:56:07preglowyes, i more or less expected there to be a more low level one
19:59:04linuxstbNot sure how the metadata decoding fits in, hopefully this is easy to incorporate without the file_decoder
20:00:44preglowi'm note sure either, i'm relatively new to rockbox
20:03:26preglowCassandra: i'll take your advice and try ubuntu
20:03:56 Nick lImbuAWAY is now known as lImbus (
20:07:18 Quit ripnetuk (Read error: 110 (Connection timed out))
20:11:24 Quit preglow ("n00bix")
20:13:58 Join Patr3ck_ [0] (
20:24:41 Nick quelsaruk is now known as Quel|away (~kvirc@
20:26:54 Join hubble [0] (
20:28:28hubblehas anyone looked into how iRivers codec (UDA1380) is connected to the CPU?
20:29:30hubbleI tried looking at the board scans but couldn't follow the pinouts very far :(
20:31:52 Quit Patr3ck (Read error: 110 (Connection timed out))
20:38:42 Join preglow [0] (
20:42:15 Part CrunchyWhiteMeat
20:50:41***Saving seen data "./dancer.seen"
20:55:20hubbleLike L3DATA and L3CLOCK terminates at a yellow small circle, which I guess continue at the other side
20:56:52hubblebut the dont.. I can see the matching yellow hole (using photoshop's layers) but the hole isn't connected?
20:57:00hubbleis the board using multiple layers?
20:57:37preglowperhaps, it's pretty common
21:05:15 Join jyp [0] (
21:05:58 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
21:15:34 Join JJ_ [0] (
21:15:38 Join Cassandra_ [0] (~christi@
21:20:50JJ_does that "lock" error appear on the studio 10 version?
21:20:55JJ_I'm kinda paranoid
21:21:18JJ_"None of us developers have experienced this problem since march 2002." = fixed?
21:21:34Cassandra_I would think so.
21:21:53Cassandra_People have occasional lockup problems with Rockbox. I've never had one.
21:22:12Cassandra_If I did, I'd turn the Jukebox off and on again, and all would be fine.
21:25:35 Quit JJ_ ("Chatzilla 0.9.66 [Mozilla rv:1.7.5/20041107]")
21:27:00HClwhats new?
21:27:04HCllinus back yet?
21:27:32Cassandra_I think he's hiding behind that pixel over there.
21:28:14XShocKppl, please help with compiling gcc. what do i miss if i get "as: unrecognized option `-mc68020'" while compiling?
21:29:54Cassandra_I'd guess that you're trying to use gcc for the wrong architecture, and you need a gcc specifically built for cross-compiling. But I am not an expert.
21:30:16XShocKmmm... I used " ../../gcc-3.3.4/configure −−target=m68k-elf −−prefix=/home/shock/m68k-compiler −−enable-language=c"
21:30:29XShocKand make.
21:30:32HCldid you compile a cross binutils first
21:30:51XShocKit built fine.
21:31:01HClwell, should work then o.o
21:31:15XShocKi did like in ""
21:31:19linuxstbXShock: is it installed, and is /home/shock/m68k-compiler/bin/ in your path (assuming you put the binutils there as well)
21:31:25XShocKexcept that used m68k-elf target
21:31:27HClyea, thats a vital part
21:31:32HClit needs to be in your path
21:32:46 Quit Cassandra (Read error: 110 (Connection timed out))
21:32:56XShocKPATH is /home/shock/rockbox/m68k-compiler/bin:/usr/local/bin:/usr/bin:/bin:/usr/X11R6/bin:/cygdrive/g/
21:33:02 Nick Cassandra_ is now known as Cassandra (~christi@
21:33:52linuxstbXShock: which version of binutils did you use? I used today's CVS version, and it worked fine.
21:34:32HClyou need cvs
21:34:35HClfor iriver toolchain.
21:34:57XShocKaah. ok.
21:35:06XShocKthx. :)
21:35:42linuxstbHCl: Is gcc-3.4.2 required for the iRiver as well? XShock is using 3.3.4.
21:35:57preglowthere's no requirement on that side
21:36:03preglowi think most of them should work
21:36:24linuxstbDo we know what Linus is using?
21:38:53HCli use 3.3.4, since it was the recommended in the text
21:40:51 Join CrunchyWhiteMeat [0] (
21:41:17linuxstbThe CrossCompiler Wiki page says gcc-3.4.2 and CVS binutils for the Coldfire
21:41:45HClit does?
21:42:36HClthats the compiler used to compile the cross gcc
21:43:00HCli think
21:43:11HClah well.
21:43:17HClwhere is linus!? :3
21:43:27HCldarnit :x
21:43:28linuxstbIt's ambiguous, but I think it's saying compile 3.4.2 as a cross-compiler.
21:44:34 Join ripnetuk [0] (
21:44:41HClhey rip
21:44:43HClanything new?
21:45:23preglowit doesn't matter what compiler you use to compile the cross compiler
21:45:33preglowas long as it's bug free
21:47:09linuxstbI'm progressing well getting libFLAC (and an associated test player) compiling for the iRiver, but wonder if anyone can help with the following gcc warning:
21:47:09linuxstb warning: inlining failed in call to 'FLAC__bitbuffer_write_raw_uint32': −−param inline-unit-growth limit reached
21:47:56preglowyou need to increase an inline limit in gcc
21:48:27preglowbut unless inlining is critical, you can ignore it
21:48:39preglowyou're using gcc 3.4, yes?
21:48:52linuxstbYes, 3.4.2
21:49:00preglowthat's got inliner bugs, i bumped into one of them the other day while coding a vst plugin
21:49:29preglowbut it's not critical
21:50:02linuxstbWhat do you mean by "not critical" - it is the kind of bug that would stop my FLAC decoder working?
21:50:25preglowthe opposite, it will not stop it working
21:50:29preglowit just wont inline the call
21:50:38linuxstbI can live with that then.
21:51:00linuxstb... until I get to the optimisation stage.
21:51:43preglow3.3.x seems like the best choice if you depend on inlining, i actually had to revert to 3.3.1 for that project myself
21:52:21rasherIs anyone working on vorbis? Or is that just gently beating tremor into submission?
21:52:40preglowthat's right
21:53:00rasherSo (hopefully) no biggie then?
21:53:10preglowright, but it'll still take some work, of course
21:53:22preglowit's hard to say, we haven't ported any codecs yet, so don't know how hard it'll be
21:54:29rasherTrying not to get my hopes too high :)
21:54:47preglowno reason not to, it might just take time
21:56:29 Quit ripnetuk (Read error: 60 (Operation timed out))
21:58:30XShocKcan someone check if it is only me who can't access website
21:59:04preglowi can't access any .coms right now
21:59:13preglowseems like some crucial connection went down
21:59:26jypI've fixed gcc/calmrisc16 ICEs ...
21:59:48jypBagder, (or someone else), would you try it out ?
22:00 doesn't work for me at the moment (I'm in the UK), but it was working about 8 hours ago when I checked out binutils from CVS
22:09:07linuxstbI can put a .tar.bz2 file of today's binutils CVS on my website if you want - it's about 12MB.
22:09:08preglowwell, it doesn't work here
22:09:16preglowneither do any of the boxes i know are on that network is horrid.
22:10:51HCli had to retry like, at least 25 times
22:10:56HClbefore it'd finally let me log on
22:10:59HCland download cvs binutils
22:11:14HClkeep getting connection reset by peer
22:11:44XShocKlinuxstb: yep. please publish it
22:12:30linuxstbOK, I'll start it uploading now - should take about 5-10 minutes. I
22:22:48linuxstbSorry, I think I spoke too soon - I can't access my own website any more.... The scp worked though.
22:23:36XShocKi can't do it too. :)
22:28:14 Quit preglow ("ble")
22:28:29XShocKit says 404 not found
22:29:15 Join Digital007 [0] (
22:29:36linuxstbXShocK - I'm trying to find somewhere else to put it
22:30:31 Join preglow [0] (
22:31:35DMJCI've built it
22:31:48DMJCbe0bf0e1df367e5cffd9e27eff988ef1 md5sum of the new hex
22:32:12DMJCab19b8654696e8ac77b8d299cc9a07c2 md5sum of the .bin file before encoding
22:32:25preglownew.hex is same as mine
22:36:12 Join amiconn [0] (
22:37:07DMJCloading firmware... result -1
22:37:31DMJCit works
22:37:39preglowyou flashed it?
22:37:41preglowi'm going for it!
22:37:51DMJCI held record at boot
22:37:55DMJCstraight in
22:37:59DMJCiriver firmware loads fine
22:38:09preglowyou didn't put in a rockbox yet?
22:38:09rasherwhat was that: [22:37] <DMJC> loading firmware... result -1
22:38:36rasheroh, it was looking for rockbox on the harddisk?
22:38:44amiconnI think that means it didn't find a loadable firmware, i.e. rockbox
22:39:08DMJCI don't have rockbox on there
22:39:10DMJCso it's fine
22:39:37preglowok, here we go
22:39:43DMJCso that's the evil part done
22:40:01DMJCI'm guessing rockbox updates are bootloader neutral
22:40:10preglowshould be
22:40:46 Join ripnet_ [0] (
22:40:55 Quit methangas (" HydraIRC -> <- IRC with a difference")
22:41:26ripnet_wow! i just flashed and it works :)
22:41:30preglowok, what's it supposed to do?
22:41:39preglow'cause mine's less than responsive
22:41:48DMJCso what do I do now?
22:41:49preglowwhich would be pretty typical
22:42:04DMJCok when you flash and put it on it should say loading firmware result -1
22:42:09DMJCthen turn it off
22:42:14DMJCand turn it on and hold record
22:42:19DMJCshould boot iriver firmware
22:42:21ripnet_i just did a make zip, and unzipped the zip onto the root. included a firmware and a .rockbox
22:42:24preglowthere it comes
22:42:31preglowthe protection worked
22:42:34ripnet_and original firmware still works :)
22:42:36preglowdidn't even have to push record
22:42:48ripnet_big round of applause for LinusN :)
22:42:49preglowthe hdd led just blinked like hell
22:42:59DMJChow do we install rockbox? heh
22:43:10preglowwell, you compile it first :P
22:43:21preglowcopy it to root directoyy
22:43:24linuxstbCongratulations - you're braver than I am.
22:43:32preglowdamn, you've got balls, man, i was planning to wait for linus :P
22:43:48ripnet_rockbox boots :)
22:43:52ripnet_can browse dirs
22:43:58preglowcould someone put up a working rockbox image?
22:44:03ripnet_have crashed it already
22:44:03preglowi've got tool problems atm
22:44:14DMJCI guess this makes my 2 year warranty void..
22:44:17ripnet_preglow - do you want me to send you the entire zip with plugins and stuff
22:44:22preglowripnet_: would be cool
22:44:27ripnet_hang on a sec
22:44:30 Quit Cassandra (Read error: 110 (Connection timed out))
22:44:34preglowripnet_: no rush, i've got to eat
22:44:36rasherDMJC: just flash it with a virgin firmware - they'll never know
22:44:56rasherunless of course it crashes with dirty firmware on it
22:45:06DMJC"what are these flac files doing on here...?"
22:45:11ripnet_go to">
22:45:18rasherwell you can always wipe the harddrive
22:45:19 Join Cassandra [0] (~christi@
22:45:19ripnet_there is a zip there, just unzip it into root of iriver
22:45:35DMJCanyway.. what do you put on it?
22:45:39DMJCfor the rockbox stuff
22:48:35HClDMJC / preglow: what gcc version did you use
22:48:48DMJCsomeone zipped/sent me their compiler
22:48:57HClDMJC: can you send the .hex to me?
22:49:05Digital007so rockbox for iriver works now?
22:49:17HClDigital007: bootloader works, dir browsing works, menu works.
22:49:27HClno sound, no plugins
22:49:34Digital007how do u get back to iriver f/w?
22:49:42HClpress record while booting
22:49:43DMJCrecord button
22:49:45ripnet_you hold record while reseting
22:49:52rasheryou sure told him
22:49:54HCllet me get 3.4.3
22:49:58DMJCso hold off button to turn off
22:50:02Digital007can u send me it as well please?
22:50:17DMJCthe bootloader?
22:50:23Digital007the .hex
22:50:34DMJCdcc reject send
22:50:44***Saving seen data "./dancer.seen"
22:51:13Digital007looking forward to see the first official f/w for iRiver
22:51:23HCli think i'll try and build 3.4.3 first
22:51:26HClsee if i get the same hex
22:51:27HClif i do
22:51:32Digital007hopefully it won't be long now
22:51:34HCli'm gonna change its bootup sequence slightly
22:51:36HClthen flash
22:51:58HClDigital007: well, the bootloader isn't gonna change much..
22:52:12HClso you can pretty much flash already, since future flashes are prolly not needed
22:52:25Digital007i mean, the first official build available from the rockbox website
22:52:32Digital007im looking forward to try it out myself
22:53:51preglowStarting original firmware
22:53:54preglowthis rocks immensly
22:54:04HClwell, it *is* rockbox :P
22:54:21preglowwhen it fails here, i don't get message
22:54:27preglowthe hdd led just flashes and nothing happens
22:54:42HClwhats it doing?
22:54:56preglowhow should i know :PP
22:55:00preglowthe led just flashes
22:55:05HClno, i mean, what did you do and what is it doing now? :P
22:55:05preglowi need a rockbox to test if that part works
22:55:11preglowi reset it
22:55:13 Quit midk ("Leaving")
22:55:14hubbleanyone got a archos fm recorder 20gb (v2) for sale?
22:55:14HClyou didn't press rec?
22:55:20ripnet_sookoban works :)
22:55:21HCl :p
22:55:24HClripnet_: :P
22:55:28preglowripnet_: sendsendsend!
22:55:31HCli thought plugins didn't work yet?
22:55:46ripnet_most dont seem to
22:55:52ripnet_preglow - send what?
22:56:04ripnet_the bootloader or the binary?
22:56:13ripnet_i put it on my site for you
22:56:42preglowmissed that, obviously
22:57:10ripnet_star works as well
22:58:12preglowhow long time is it supposed to take before it boots?
22:58:21ripnet_rockbox or a plugin?
22:58:31ripnet_not long
22:58:38preglowwell, then something glitches here
22:58:41ripnet_you get the splash almost straight away
22:58:45preglowi get the hdd led thing now as well
22:59:35ripnet_what happens when you press reset?
22:59:49preglowyou mean the hard reset at the bottom, yes?
22:59:56preglowit them resets and loads the iriver firmware
23:00:05HClmust be the cookie protection?
23:00:22ripnet_have you unzipped that firmware zip i have on my site (or built your own)
23:00:24preglowbut what happend with you guys when you load with no rockbox on it?
23:00:27preglowripnet_: yours
23:01:05linuxstbCan someone confirm the md5sum of the known good "new.hex" ? I just want to double-check before I go ahead.
23:01:08ripnet_did you check you got rockbox.iriver in the root of the jjukebox
23:01:14preglowripnet_: yes
23:01:35preglowit's there, along with .rockbox
23:01:44ripnet_be0bf0e1df367e5cffd9e27eff988ef1 new.hex
23:02:07preglowbe0bf0e1df367e5cffd9e27eff988ef1 *ihp_120.hex
23:02:21ripnet_and the flash went ok?
23:02:37HClpreglow: h120 or h140?
23:02:38ripnet_you havent still got the cookie there have you?
23:02:40preglowHCl: h120
23:02:47HClripnet_: h140 right?
23:02:50HClDMJC: h140 too, right?
23:02:55ripnet_if you had a failed attempt (because of missing rockbox.iriver) it may have put the cookie in place
23:03:02linuxstbOK, I'll give it a shot. Without FLAC or MP2 playback, my iRiver is just a paperweight anyway...
23:03:03ripnet_which can survive a soft reset
23:03:20ripnet_did you have a 'bad start' first?
23:03:32preglowbut it's been power off now
23:03:33HClmy 3.4.3 is done..
23:03:38preglowthe cookie does not remain there after that
23:03:47ripnet_i dont know if power off would clear the ram or not
23:03:50DMJCnow if rockbox hangs
23:03:53preglowbut i guess correct behaviour is anyway not just blinking the hdd led when something is wrong
23:03:54DMJCdo I just reset the player
23:04:13ripnet_h140 for me
23:04:33ripnet_pre - does record+on boot original soft?
23:04:44preglowripnet_: yes, i get 'boot original soft' message
23:04:58DMJCrockbox loads!
23:04:59 Quit Cassandra (Read error: 60 (Operation timed out))
23:05:01preglowbut if i don't, the player spins up the hdd, and then jsust blinks the led
23:05:12preglowi thought someone mentioned an error message if it couldn't find rockbox
23:05:16preglowi do not get one of those
23:05:25ripnet_i never tried it with no rockbox.iriver
23:05:41preglowDMJC: but you did, right? and you got an error message?
23:06:05preglowlooks like it's stuck in a loop, the hdd blinking pattern is the same
23:06:09DMJCIF you don't have rockbox on the iriver it'll give an error of -1
23:06:18preglowDMJC: i don't get that
23:06:19DMJCif you do have rockbox it should boot it
23:06:26preglowit doesn't do that either
23:06:37preglowgod damn, i'm glad this baby still has a working os on it
23:07:08DMJCI'm testing on my ihp-140
23:07:30linuxstbLooking at the code in bootloader/main.c, it looks like it loops until you press a button, and then powers off (if there is no rockbox.iriver file)
23:08:06ripnet_pre - thats the same zip i unzipped onto mine... wanna do a md5sum on the rockbox.iriver to check?
23:08:22ripnet_6d9989364eaf8d4aff12ab092f7213d0 rockbox.iriver
23:08:26preglowlinuxstb: well, it's not responsive, i have to use the reset button on the bottom of the player
23:08:32HClDMJC: it worked for you ?
23:08:37preglowthen the cookie triggers, and it doesn't even check for the record button
23:09:07preglowripnet_: that's right, same here
23:09:26HCllinuxstb: do you have an h120 or h140?
23:09:39preglowit looks like there are still bugs in the bootloader, 'cause i've got the exact same md5 as you guys, and everything BUT the rockbox booting part seems to function
23:09:41DMJCoh crap
23:10:03DMJCI03:AddrErr at 31ff89b0
23:10:13preglowreset it, then :PP
23:10:14ripnet_i got some of them
23:10:17ripnet_in plugins
23:10:22preglowit shouldn't suprise you that there are still bugs
23:10:28linuxstbI can't find my iRiver manual - but I'm assuming I should have the charger plugged in when reflashing?
23:10:33DMJChow do I reset it?
23:10:44DMJCjust the button on the base?
23:10:46HClDMJC: reset hole at the bottom
23:10:46ripnet_press the reset button on the bottom
23:11:21 Quit bobTHC (Read error: 110 (Connection timed out))
23:11:41preglowwell, shit
23:11:43preglowthis was not what i planned
23:11:50HClyou can reflash it with the original
23:11:54HCltill linus looked at it?
23:12:03preglowi'll wait for linus
23:12:08preglowperhaps he can figure out what's wrong
23:12:09 Join bobTHC [0] (
23:12:11ripnet_it might not be a prob with the bootloaader
23:12:20HCli compiled, and i got the same md5sum
23:12:31HCli'm just kind of waiting on a reply from DMJC on how well it works for him
23:12:35HCland ditto for linuxstb
23:12:54ripnet_pre - does yours even say loading firmware?
23:12:57linuxstbI still haven't summoned the courage to do it.
23:13:04preglowripnet_: no
23:13:13preglowripnet_: only if i press rec, then it says 'loading original firmware'
23:13:39preglowthis happens: backlight comes on, drive spins up, red hdd led blinks erratically in a pattern about a second long
23:13:42preglowthat's that
23:14:22ripnet_i guess we are needing Linus to work out whats happening. Its encouraging that it can display text at least :) and very encouraging that you can boot original
23:14:35preglowi quite enjoy that fact
23:15:07linuxstbOK, I'm flashed, about to turn back on....
23:15:13*HCl prods DMJC
23:15:14ripnet_have you checked there is nothing funny in your root dir pre?
23:15:41ripnet_hcl - prodding is here usually seems to be with a stick :)
23:15:43preglowripnet_: like what?
23:15:53ripnet_dunno, loads of files perhaps?
23:15:57preglowripnet_: there's the firmware and rockbox and a bunch of dirs
23:15:57DMJCdamn mother
23:16:04linuxstbEverything seems fine - I held down record+on, a message came up saying loading original firmware, and then the original firmware loaded. :-)
23:16:09DMJCtrying to come in here and clean up my room..
23:16:13ripnet_similar to mine.
23:16:24DMJCeven my girlfriend recognizes that computer desks need to be messy
23:16:35preglowit acutally seems like stuff hangs in ata_init()
23:16:45DMJCthen again... she's a bs user heh
23:17:02HCllinuxstb: h120 or h140?
23:17:17linuxstbh140 - I bought it about 4 weeks ago from
23:17:19preglowso, i'm the only one with a h120 here?
23:17:23preglowthat might of course be it
23:17:35preglow'cause then i'm the first one to actually flash a h120
23:17:43DMJClinus flashed his 120
23:17:45HClpreglow: well, linus works on a h120 right?
23:17:50HClbut yea
23:17:53ripnet_the first one to go was a 120
23:17:54rasherLinus is cheating though :)
23:17:54preglowwell, did he flash it the ordinary way?
23:17:55HClit seems to be the difference so far
23:18:00preglowhe might have used bdm
23:18:01HCli think i'll flash :3
23:18:07preglowbut anyway, enough speculation
23:18:11 Join Cassandra [0] (~christi@
23:18:12ripnet_i think he did it with bdm and then 'normally'
23:18:15preglowi just hope he comes online soon :P
23:18:34rasherAnywhere to get a firmware image without going through a compiling hell?
23:18:34DMJCok how the hell do i turn rockbox off?
23:18:36linuxstbSo who's flashed their iRiver now, and which model?
23:18:46linuxstbI've flashed a H140.
23:18:57preglowh120, yes
23:19:15DMJCI've done a 140
23:20:12HClwe have 3 h140s working
23:20:15DMJCI can't turn it off..
23:20:22DMJChow do you stop it playing
23:20:24HCldmjc, ripnet_ and linuxstb
23:20:28HClhard reset
23:20:44DMJCI just want to tell it to shutdown which should be the stop button
23:20:56preglowwell, it's not yet
23:21:00preglowhard reset is the way to go
23:21:02HCli don't think power is working yet
23:21:17 Join mrmags [0] (
23:21:31DMJCisn't there a reboot function?
23:21:32linuxstbHCl - do you mean that you can't turn your iRiver off running the rockbox firmware?
23:21:52linuxstbI haven't tested a firmware yet...
23:22:16DMJCyou can
23:22:17ripnet_you can turn it off
23:22:20DMJCI just did it
23:22:30DMJCbut how do you turn it off when it's playing?
23:22:34DMJCI couldn't make it stop
23:22:47rashercould it have been crashed?
23:23:02DMJCwhen i tell it to play mp3s it let's me navigate around still
23:23:15 Nick lImbus is now known as lImbuAWAY (
23:23:56 Nick lImbuAWAY is now known as lImbus (
23:24:26XShocKlinuxstb: can you upload binutils-cvs to ?
23:25:03linuxstbXShocK: Yes - I'll start it now.
23:25:11DMJCso.. how long do they reckon till mp3 playback works?
23:25:33linuxstbXShocK: My website now seems to be working - can you check if the link I posted works?
23:25:50XShocKaah it works. :)
23:26:10XShocKyep. downloaded fine.
23:26:21preglowok, so everyone who has just flashed a h120, raise your hand
23:26:53*HCl copied the new firmware.. goes to flash :X
23:27:49DMJCis there a setting in rockbox to make the bootloader load the original firmware?
23:27:56ripnet_dm - no
23:28:04ripnet_you would have to mod the bootloader source
23:28:05preglowthere can't be, the bootloader's in flash
23:28:06HClDMJC: you can change the bootloader...
23:28:08rasherpreglow: what was your md5sum of the .hex ?
23:28:19DMJCreflashing sucks
23:28:21preglowbe0bf0e1df367e5cffd9e27eff988ef1 *ihp_120.hex
23:28:22ripnet_i guess what we need is a dummy rockbox.iriver that gives a menu and boots the original
23:28:23DMJCI'll leave it as is
23:28:32HClripnet_: yup.
23:28:38ripnet_or the new
23:28:46ripnet_shouldnt be too hard to code, a
23:28:47HClfirmware upgrade
23:28:48rasherripnet_: alright, I'm on my way to flash my h120
23:28:48preglowrasher: but hell, don't ask me for the md5sum, mine player was nearly bricked :P
23:28:50HClare you sure :x
23:29:00rasherpreglow: but it wasn't though :)
23:29:05preglowno, just nearly
23:29:06DMJCone sex
23:29:08DMJCone sec
23:29:14DMJCI'll give you my md5s
23:29:32rasherAnd you're the only one with a flashed h120 so far I think?
23:29:37preglowso it seems
23:29:48rasherI'd rather have a md5 for something that didn't break, than something that worked on another player :)
23:29:53HClits supposed to turn off after flashing right?
23:29:56DMJCbe0bf0e1df367e5cffd9e27eff988ef1 new.hex
23:30:00DMJCthat didn't break
23:30:04preglowHCl: yes
23:30:10preglowHCl: now turn on
23:30:20rasherAnd hold your breath!
23:30:21HClloading firmware
23:30:22HClresult -1
23:30:39ripnet_hcl - im guesing you havent installed the actual rockbox yet?
23:30:49HCli'm guessing i need to reset it now?
23:30:52preglowHCl: and this was h120?
23:31:05preglowrasher: get it on already! ;)
23:31:12rasheryup, hang on!
23:31:52ripnet_darn! rockBLOCKS (tetris) doesnt run
23:32:14rasherbe0bf0e1df367e5cffd9e27eff988ef1 /mnt/iRiver/ihp_120.hex
23:32:23mrmagsif someone could be so kind as to send me their new.hex, I'll have a go at blowing up my 140 too
23:32:26HClits supposed to run the original firmware when you reset it and it hasn't loaded rockbox
23:32:32HClmrmags: hold on
23:32:49Digital007can someone send me this hex for me to try please?
23:33:15ripnet_i would put it on my web site if Linus didnt object
23:33:27ripnet_but its only fair to ask him first
23:33:34mrmagsmany thanks
23:33:39ripnet_can someone make a .ips patch to convert original to new?
23:33:41*rasher holds his breath
23:34:03DMJCremember to rename it to ihp_120.hex
23:34:06HClDigital007: the one on my ftp is the one i flashed my iriver with
23:34:18HClwhich is the one dmjc sent me
23:34:21DMJCsame with the one I'm sending
23:34:37rasherokay, turn on, hold rec?
23:34:46DMJChold record for iriver firmware
23:34:56DMJCyou should get a -1 error
23:35:01rasherI'm not
23:35:03DMJCif you don't have rockbox on it
23:35:08preglowrasher: blinking red led?
23:35:13rasherI'm getting blinking harddisk
23:35:13rasherand I do have rockbox on it
23:35:13HClworks like a charm
23:35:19preglowrasher: then it's obviously a h120 problem
23:35:21rasherpreglow: what now, hard reset?
23:35:23preglowrasher: i get the EXACT same thing
23:35:25preglowrasher: yes
23:35:31HClsince it was developed on a h120
23:35:33DMJCask linus
23:35:35HClyou'd say it ran on a h120
23:35:35preglowHCl: yes, indeed
23:35:53DMJCI think I know what's happenned
23:36:00DMJClast I heard linus was testing on his 140
23:36:09DMJChe may have broken something without realizing
23:36:15rasherpreglow: iRiver firmware booted fine
23:36:23rasherno rockbox, but no brick either
23:36:30preglowrasher: nor is mine
23:36:33HCltime to put the rockbox files on it...
23:36:35preglowrasher: we have the same problem, that's all
23:36:43rasherI guess that's a good thing
23:36:52preglowfor diagnostics, at least
23:37:24rasherNo point in flashing more h120s I guess
23:37:25ripnet_i thought 120 and 140 were identical... maybe there is somethign strange about the hdd in the 120s rasher and preglow have?
23:37:36preglowripnet_: perhaps
23:37:58Digital007how long does rockbox take to boot up?
23:38:06ripnet_digital - not long at all
23:38:08preglowripnet_: but the fact that these are the only two h120s that have been flashed, and that linus' h120 is bdm'd does make it a bit trickier to conclude
23:38:09mrmagssweet, it totally worked.
23:38:13DMJClike 3 secs
23:38:16DMJCit's hell fast
23:38:21DMJCmakes iriver look gay
23:38:35ripnet_< 10 secs to 'resume yes/no' screen
23:38:42rasherI'd say their firmware makes them look gay :)
23:38:46ripnet_the display is amazing, loads of lines
23:38:52Digital007mine is just showing a blank screen and the hd light flashing
23:38:53preglowyse, quite
23:38:58rasherDigital007: h120?
23:38:58preglowi noticed the font was nicely sized
23:39:06linuxstbYes, I make it about 4 seconds to the Rockbox logo, and another 2 or 3 seconds until the browse screen appears
23:39:12preglowit's a h120 problem
23:39:16rasherAlright, we definately have a h120 problem
23:39:21preglowif you've got a h120 and want to flash it, don't flash it
23:39:30ripnet_thats 3 120s not working?
23:39:32DMJCyeah it's nice being able to actually read the damn id3 tags
23:39:36ripnet_how may 140s working?
23:39:43preglowfour or five, i'd guess
23:39:43DMJCat least 3
23:39:49linuxstbCan someone summarise the H120 problem?
23:39:49Digital007so how do i get back to iriver?
23:39:51rashersomeone should be keeping stats in the topic :)
23:40:01rasherDigital007: hard reset.. bottom of the unit
23:40:08DMJCH120 problem: IMHO sounds like rockbox has an ata driver issue
23:40:17ripnet_digital - and hold record when pressing reset
23:40:22DMJCI think the bootloader itself may be fine
23:40:23*HCl unzips ripnet's zip
23:40:34rasherripnet_: no, that'll just try and boot rockbox again, and fail
23:40:46linuxstbDMJC: So you load the original firmware, but it won't read rockbox.iriver from the disk?
23:40:48ripnet_record = original soft
23:41:00DMJCthat's what some people were saying
23:41:02preglowrasher: but if you start with the record button pressed, you do get a message from the bootloader that it's loading the original firmware, yes?
23:41:11ripnet_dm - the ata driver is in the boot loader
23:41:16DMJCah... k
23:41:19 Join midk [0] (
23:41:27DMJCso I guess the bootloader ata driver is b0rked
23:41:28rasherpreglow: ah, yes
23:41:31Digital007its alright i have fixed my h120
23:41:32rasherpreglow: sorry
23:41:38preglowrasher: what sorry?
23:41:45HCli think i made it crash just by scrolling o.o;
23:41:52preglowi think the error is in ata_init() somewhere
23:42:04rasherpreglow: I was talking nonsense about holding in record
23:42:17preglowrasher: i didn't read it anyway, so no biggie ;)
23:42:30ripnet_it could be a non-120 problem - did you all get the soft from the same source? do we have different independent source?
23:42:39preglowi compiled mine myself
23:42:49preglowand noone has gotten it from me
23:42:49rasherripnet_: I have the rockbox from your site
23:42:51HClripnet_: the md5 sums are the same
23:43:11rasherHCl: that's the firmware, not the rockbox software
23:43:11preglowwe've all flashed with identical md5s
23:43:13preglowso it's not that
23:43:22ripnet_i guess its wait for Linus time again... sorry i cant help any more
23:43:22DMJCdid anyone compile their own h140 firmware?
23:43:31HClmy rockbox software comes from you
23:43:33ripnet_dmjc - i compiled everything myself
23:43:33DMJCor did everyone use mine?
23:43:36HClmy make zip build failed on fonts something
23:43:48ripnet_hcl - also checkout the fonts from cvs
23:44:07HClis it me or is it scrolling really slowly?
23:44:09HCland like
23:44:10HClif you hold it
23:44:17HClit'll just remember keystrokes and keep scrolling
23:44:19Bagderit is slow
23:44:24mrmagssame here
23:44:31*rasher watches his hdd led blinking
23:44:35Bagderit only runs at 11mhz
23:44:45ripnet_its not finished... the ata driver is slow atm, but it will get a lot bettter - the processor is a beast compared to the archos
23:44:46HClwhy is its battery blinking?
23:45:05Digital007so why shouldn't it work on the 120 if it works on the 140?
23:45:24HClhow do i access its menus?
23:45:27DMJCit (should)
23:45:35ripnet_hcl - a-b key
23:45:37DMJCthere is one difference between the two models
23:45:43DMJCthe hard drives themselves
23:45:45Digital007all i got was a blank screen and the HD light flashing
23:45:47DMJCone is single platter
23:45:53DMJCthe other is dual platter
23:46:00HClpretty is different
23:46:02HClbut at least its functional
23:46:03DMJCobviously there are gonna be differences there
23:46:19 Quit midk ("Leaving")
23:46:38rasherpreglow: what happens if you boot without .rockbox + rockbox.iriver
23:46:46preglowrasher: same as if i boot with
23:46:53rasherSo it's before loading
23:47:05rasherI guess we knew that since we get no splash/text
23:47:05Digital007im just gonna wait for the "official" builds to appear on the rockbox website
23:47:07preglowi think it is in the hard drive init
23:47:10HCli made it crash earlier when i was holding the button down for a long time
23:47:17HClcan anyone reproduce that?
23:48:17HCljust while scrolling directories
23:48:24mrmagscute, upside-down works
23:49:09 Quit Patr3ck_ ("User pushed the X - because it's Xtra, baby")
23:49:17*HCl grins
23:49:26rasheris someone updating the wiki?
23:49:31HClbut we should.
23:49:36HCli'll make an attempt
23:49:39HClprovided i can remember my login
23:49:55ripnet_should be list model, and md5 of hex file flashed, along with success?
23:50:00amiconnw00t - new virtual keyboard for player works! While now voiced like the recorder keypad and much more user friendly, it's also more compact and straightforward code.
23:50:25amiconnI'll apply some of the ideas to the recorder keypad as well.
23:51:16DMJCheh pity we can't plug a keyboard into the usb port and use it as a pda..
23:51:32ripnet_its going wrong somewhere in the lines
23:51:33ripnet_ system_init();
23:51:33DBUGEnqueued KICK ripnet_
23:51:33ripnet_ kernel_init();
23:51:33ripnet_ backlight_init();
23:51:33***Alert Mode level 1
23:51:33ripnet_ set_irq_level(0);
23:51:35ripnet_ lcd_init();
23:51:37ripnet_ font_init();
23:51:39ripnet_ adc_init();
23:51:41ripnet_ button_init();
23:51:43ripnet_ lcd_setfont(FONT_SYSFIXED);
23:51:45ripnet_ sleep(HZ/10); /* Allow the button driver to check the buttons */
23:51:47ripnet_(sorry for flood)
23:51:50HClDMJC: actually.
23:51:57HClDMJC: the connector for the remote is a generic interface
23:52:00ripnet_and you get the backlight, so im guessing its after backlight_init
23:52:08HCllinus said its possible to hook a keyboard up to it
23:52:15rasherripnet_: nah, I don't think we get backlight
23:52:28rasherhang on
23:52:28preglowripnet_: it is in ata, the drive spins up
23:52:50preglowripnet_: and the earliest ata that happens, is ata_init, and right after it, shit is printed to the screen
23:52:57preglowripnet_: hence, i believe ata_init is the culprit
23:53:17rasherripnet_: we get backlight, nevermind
23:53:25 Join amiconn_ [0] (
23:53:30ripnet_pre - that makes sense
23:53:42preglowripnet_: i agree ;)
23:53:47 Nick DMJC is now known as DMJC-SUPERBOWL (
23:54:07HClfor some reason
23:54:09ripnet_looking at code, i agree with you agreeing with me agreeing with you :)
23:54:11HCli can't even log in to the wiki
23:54:27HClit simply says authentication failed without even asking for a password
23:54:57rasherI'm logged in
23:54:59hubblerasher: i'm writing some bootloader status in the wiki
23:55:11rasherhubble: good
23:55:32hubblemrmags: did you flash? model/md5/compiler/result ?
23:56:08mrmagsyes, 140 / be0bf0e1df367e5cffd9e27eff988ef1 / ripnet's / success
23:56:14CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
23:56:14*HCl got his wiki account to work
23:56:26 Quit jyp ("poof!")
23:57:01HCli'm guessing rockbox has a limited gui because archos aren't powerful enough?
23:57:10 Join jyp [0] (
23:57:23rasherhubble: 120 / be0bf0e1df367e5cffd9e27eff988ef1 / ripnet's / rockbox gives blinking hdd, can boot iriver firmware by holding rec
23:57:44preglowhubble: yes, and crash detection works
23:58:02 Quit jyp (Client Quit)
23:58:04linuxstbh140 / be0bf0e1df367e5cffd9e27eff988ef1 / gcc-3.4.2, binutils-cvs-2005-06-02 / success
23:58:18preglowwe've all got the same image, i'm pretty certain about that
23:58:20muesli_n8 ladiez
23:58:33Rhunakhey, just wanted to say that you guys are doing amazing work... good n8 @ all
23:58:34HClpong dies.
23:58:39HCl :/
23:58:43ripnet_h140 / be0bf0e1df367e5cffd9e27eff988ef1 / complete success

Previous day | Next day