Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2010-11-29

00:00:01 Join markun [0] (~markun@rockbox/developer/markun)
00:00:34eWillI thought maybe. I'm firing-up a disk-hex editor...???
00:06:45 Quit elcan (Read error: Connection reset by peer)
00:06:53 Join simonrvn [0] (
00:10:12CIA-7New commit by wodz (r28700): Rename bootloader file for MPIO players as it supports both HD200 and HD300.
00:12:22CIA-7r28700 build result: All green
00:13:25 Join elcan [0] (
00:14:29 Quit evilnick_ (Read error: Connection reset by peer)
00:15:54 Join JdGordon| [0] (
00:15:55 Quit JdGordon| (Changing host)
00:15:55 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
00:19:14 Join mortalscan [0] (~mortalsca@
00:21:42 Quit AlexP_mob (Quit: Cheerio)
00:24:58 Quit Kitr88 (Ping timeout: 245 seconds)
00:30:14 Part Zagor
00:30:27 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
00:31:30 Quit ReimuHakurei (Read error: Connection reset by peer)
00:31:34 Join Kitar|st [0] (
00:32:51 Part Zagor
00:35:50 Quit InsDel (Read error: Connection reset by peer)
00:40:20*TheSeven just got an ATA IDENTIFY packet from his drive :)
00:41:25TheSeveneWill: which target did that happen on?
00:41:31TheSeveni've had that several times on nano2g
00:43:42eWillfuze v2 (NOTE: i'm using RB USB patch)
00:43:42eWillI can't find it with a hex editor... :(
00:44:14 Join ReimuHakurei [0] (~ReimuHaku@
00:45:08TheSevenso there is not even a directory entry?
00:45:32TheSevenyou might want to have a look at the first bunch of free clusters
00:46:35eWillI don't know how the file system is constructed. I'm just searching for text that is in all 'battery_bench.txt' files.
00:47:21TheSevenok, if you scanned the whole flash for that, it just isn't there.
00:51:37 Join Nerdy3_14159265 [0] (
00:52:35 Quit wodz (Quit: Leaving)
00:57:36***Saving seen data "./dancer.seen"
01:00:11 Quit Nerdy3_14159265 (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101027124735])
01:02:25 Join InsDel [0] (
01:03:11 Quit advcomp2019_ (Ping timeout: 276 seconds)
01:05:24 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
01:05:28 Quit nls (Quit: Lämnar)
01:08:15 Quit sasquatch (Ping timeout: 255 seconds)
01:08:34JdGordon|can anyone do me a logf mini2g build please?
01:18:52 Quit efyx (Remote host closed the connection)
01:18:54 Quit dfkt (Quit: -= SysReset 2.53=- Sic gorgiamus allos subjectatos nunc.)
01:20:09 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
01:21:03 Join sasquatch [0] (
01:21:43 Quit ender` (Quit: The shortest unit of time in the multiverse is the New York Second, defined as the period of time between the traffic lights turning green and the cab behind you honking. -- Terry Pratchett)
01:24:09 Join froggyman [0] (~seth@
01:24:09 Quit froggyman (Changing host)
01:24:09 Join froggyman [0] (~seth@unaffiliated/froggyman)
01:35:55TheSevenany ata guru around?
01:36:16TheSevenhow do i tell a drive to switch to ultradma mode? or do i even need to do so?
01:40:12pamauryNot sure you have to tell the drive, most probably the dma controller if there is one decidated to the ata drive
01:43:41 Nick JdGordon| is now known as krazykit (~jonno@rockbox/developer/JdGordon)
01:43:49 Quit krazykit (Disconnected by services)
01:43:55TheSeventhe ATA core is an AHB master that will even do PIO mode transfers internally via DMA
01:44:20 Join JdGordon| [0] (
01:44:20 Quit JdGordon| (Changing host)
01:44:20 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
01:54:01 Quit pamaury (Quit: - Discuter simplement. Partout.)
02:05:06eWillWTF??? I just notice my Fuze v2 was charging REALLY fast. So I hold select button, and connect. System > Debug > View Battery > "Charger: 3C" ????????????????
02:07:19eWillDid I just ruin my battery? (charging ~1 hour)?
02:07:50eWilli'm not mad, i just wanna know.
02:10:12eWillThese things VENT WITH FLAME when damaged.
02:12:15 Join UnclePervyJesus [0] (~ubernoob@2001:c08:3700:ffff::13:ef7c)
02:21:12 Nick JesusFreak316_ is now known as JesusFreak316 (
02:22:39eWillmaybe i wasn't clear (sorry) −− _Rockbox_ is what what charging my player too fast, not the OF.
02:24:41UnclePervyJesusanybody got a link to runtimes for the h10 (iriver) ?
02:26:33UnclePervyJesusthanks, i just got a 20gb for 25bucks today :D
02:28:03eWillyesterday my Fuze v2 took ~3 hours to reach 100%. Today RB was at 87% in less than an hour.
02:29:04UnclePervyJesusyeah that can happen, not sure why, but i do know that i have to fully discharge mine every couple months to keep the meter accurate
02:34:48 Quit mortalscan (Ping timeout: 245 seconds)
02:36:19 Join mortalscan [0] (~mortalsca@
02:46:08 Quit GeekShadow (Quit: The cake is a lie !)
02:57:39***Saving seen data "./dancer.seen"
03:07:52 Quit fyre^OS (Read error: Connection reset by peer)
03:10:20 Join T44 [0] (
03:13:35 Quit Topy44 (Ping timeout: 240 seconds)
03:26:53 Join krabador [0] (
03:35:44eWillanyone ever ported a file compressor to RB?
03:48:11 Quit factor (Quit: Leaving)
03:54:30CIA-7New commit by rmenes (r28701): FS #11780 by Melba Sitjar: add missing strings to the Tagalog ...
03:56:52CIA-7r28701 build result: All green
04:06:21 Quit InsDel (Read error: Connection reset by peer)
04:06:58 Quit bluebroth3r (Read error: Connection reset by peer)
04:07:20 Join bluebrother [0] (
04:07:20 Quit bluebrother (Changing host)
04:07:20 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
04:12:46 Join DerPapst [0] (
04:14:34 Join hebz0rl [0] (
04:19:42 Join cjcopi [0] (
04:20:04 Quit TheSeven (Ping timeout: 276 seconds)
04:23:07 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:26:40 Quit Barahir (Ping timeout: 265 seconds)
04:28:20 Join Barahir [0] (
04:33:58 Join factor [0] (
04:45:47 Quit krabador (Quit: Sto andando via)
04:47:01 Join plombier [0] (
04:49:19 Quit amiconn (Disconnected by services)
04:49:21 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:49:40 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:50:17 Quit DerPapst (Quit: Leaving.)
04:50:20 Quit pixelma (Disconnected by services)
04:50:23 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:50:25 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:52:35plombierFirst greetings to #rockbox - Cheers to wodz - just compiled/installed/ran rockbox for the first time on MPIO HD300.
04:53:15plombierGreat porting job, and great job all on rockbox. Am currently enjoying 20G of fine oggs.
04:54:23 Quit factor (Read error: Connection reset by peer)
04:57:42***Saving seen data "./dancer.seen"
04:59:22 Nick scorche` is now known as scorche (~scorche@rockbox/administrator/scorche)
05:00:51 Quit Judas_PhD (Quit: This is a quitting message)
05:09:23 Part S00row1
05:10:20 Join factor [0] (
05:11:09 Join Judas_PhD [0] (
05:14:27 Quit clone4crw (Remote host closed the connection)
05:16:50 Quit Judas_PhD (Quit: This is a quitting message)
05:16:56 Quit mystica555_ (Ping timeout: 265 seconds)
05:28:24 Join MethoS- [0] (~clemens@
05:30:59 Quit anewuser ()
05:31:12 Quit ReimuHakurei (Remote host closed the connection)
05:31:27 Join ReimuHakurei [0] (~reimu@
05:35:52 Quit Horschti (Quit: Verlassend)
05:37:12 Quit panni_ (Quit: ( :: NoNameScript 3.81 :: ))
05:53:31 Nick shai_ is now known as shai (
05:57:30 Quit Keripo1 (Ping timeout: 240 seconds)
06:01:04 Join Keripo [0] (
06:12:23 Join DSStrife89 [0] (
06:16:31 Quit madalu (Remote host closed the connection)
06:33:29 Quit JesusFreak316 (Ping timeout: 245 seconds)
06:50:20 Join dantje [0] (
06:56:58 Quit Kitar|st (Ping timeout: 250 seconds)
06:57:14 Quit sasquatch (Quit: new kernel worship reboot)
06:57:45***Saving seen data "./dancer.seen"
07:01:28 Join sasquatch [0] (
07:16:38 Quit JdGordon| (Quit: leaving)
07:19:12 Quit BHSPitMonkey (Quit: Ex-Chat)
07:32:19 Join Buschel [0] (
07:40:07BuschelTheSeven: hmm, the nano 2g crashed again this night. this was with 24 MHz and changed refresh settings to cope with 24 MHz
07:40:51 Join stoffel [0] (
07:44:27BuschelTheSeven: the battery bench itself is ranging down to 75%. from this few data it seems the downclocking might bring another 10% more runtime −− compared to svn
07:51:04Buschelbtw, if it hangs up it seems to be frozen in a state with higher current consumption (boosted?).
07:57:47 Quit MethoS- (Remote host closed the connection)
07:59:40 Quit Keripo (Read error: Connection reset by peer)
08:03:38 Quit Stummi (Excess Flood)
08:04:05 Join Stummi [0] (Stummi@rockbox/developer/Stummi)
08:04:06 Quit Stummi (Excess Flood)
08:04:35 Join Guest78023 [0] (
08:05:38 Join Keripo [0] (
08:12:25 Quit xavieran (Ping timeout: 245 seconds)
08:14:14 Join Keripo1 [0] (
08:15:09 Join Judas_PhD [0] (
08:16:29 Quit Keripo (Ping timeout: 260 seconds)
08:25:22 Join xavieran [0] (
08:27:17 Join LinusN [0] (~linus@rockbox/developer/LinusN)
08:28:48 Join BlindWanderer [0] (
08:30:33BlindWandereranyone about?
08:31:08 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
08:32:20BuschelTheSeven: hmm, when reading the DRAM spec in detail the row address bit widthis 13, not 12. in this case svn's config is correct for 96MHz and we are lucky when running the nano at 48 MHz. I will do another test run with a configuration for 24 MHz based on 13 bit row address width.
08:34:31 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
08:35:34BuschelTheSeven: we could use different DRAM settings for the different clocks. this could allow further speed-up, not sure though −− will also test this some day
08:35:39 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101026210630])
08:37:43BlindWandererI think r28684 has propagated a bug: The loop in find_extension skips the first item. If you look at the code removed it shows disagreement as to which index to start on.
08:42:09 Join einhirn [0] (
08:43:12 Quit BlindWanderer (Quit: CGI:IRC)
08:48:54 Join ender` [0] (
08:48:54 Quit Judas_PhD (Read error: Connection reset by peer)
08:50:02 Join Judas_PhD [0] (
08:57:47***Saving seen data "./dancer.seen"
09:00:10 Join xxcv [0] (
09:13:18 Quit stoffel (Ping timeout: 265 seconds)
09:14:41 Join silver_hook [0] (~silver_ho@
09:19:27 Join Rob2223 [0] (
09:20:09silver_hookI'm at the brink of switching my -old- ancient iRiver iFP-895 for a newer and since I wanted to put Rockbox on it (sadly iFP is not supported, so I didn't have the honour yet), I'd appreciate any suggestions.
09:20:58silver_hookSo far I'm looking at Sansa Clip+…
09:22:30 Quit Rob2222 (Ping timeout: 240 seconds)
09:28:04 Quit bertrik (Ping timeout: 245 seconds)
09:28:49eWillI just bought a re-certified Fuze v2 4GB off Newegg for $30 (I think they bumped it to $40 now). The battery seems brand new.
09:29:02 Quit sasquatch (Quit: WeeChat 0.3.2)
09:29:27 Join sasquatch [0] (
09:36:57silver_hookeWill: And happy with it?
09:37:19 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:37:28 Join swilde [0] (
09:38:50eWillI've had it 3 days. All seems well. Around 22 hour battery when running Rockbox. I'm not an audiophile.
09:39:23eWillbattery in 'battery bench" −− NOT real life use.
09:39:48eWillMP3 V0
09:43:21eWillsilver_hook: hey
09:43:33silver_hookeWill: Sorry, didn't see it.
09:43:40eWillone thing: The USB jack on the device and...
09:43:58eWillthe headphone jack are right next to each other which means
09:44:20silver_hookOuch :\
09:44:44eWillfor me that means I had to wrap wire around the plug so I don't have to push in the buttons on the side
09:45:01eWillotherwise I'd have to remove the headphone jack every time I charge
09:45:10eWill*headphone plug
09:45:19silver_hookCould be worse, but it's still annoying yeah :\
09:45:24eWillbut it's not loose
09:45:56 Quit ReimuHakurei (Read error: Connection reset by peer)
09:46:34eWillI just used a twist-tie from a loaf of bread.
09:48:37silver_hookHehe, that's a quick hack ;)
09:51:38silver_hookeWill: And I suppose the question whether Rockbox is better then original firmware, right? ;)
09:52:26eWilli never use the OF (now with patch to use USB in Rockbox −− not yet commited).
09:53:34eWillexcept, I MAY have a look at the video player on the OF. I don't know yet if divx files will be much smaller than the mpg's that RB plays.
09:54:39silver_hookBTW, what is this issue with the USB + Sansa AMS + Rockbox anyway?
09:55:41eWillI think AMSv1 is stable. I've been using the patch (combined with another one that is now commited) for a few days, with no problems.
09:56:24silver_hookWhat's the issue anyway?
09:56:55eWillI forgot. (hold on a sec...)
09:57:29eWill"usb random failure"
09:57:43 Join Kitar|st [0] (
09:58:16eWilli've had no problems with that patch (the second .diff) (the one for Fuze v2 is further down the page)
09:58:51eWillby "second diff" i mean "as3525v2-usb.diff"
09:58:54silver_hookAha, OK :)
09:59:23 Join petur [0] (d408b802@rockbox/developer/petur)
09:59:52eWillThe only compliant i have on personal audio players today is surface-mounted (is that the word?) head phone jacks.
10:00:54eWillsilver_hook: they seem to become loose/unstable from simple plugging/unplugging head phones. BUT...
10:01:16eWillthe fuze has placed the jack in the center of the PCB
10:01:34eWillso I'm hopping this one will work.
10:02:03eWillnot break
10:03:43silver_hookOoooh, OK.
10:07:16eWillI don't mean "center" length or width-wise −− I mean height wise. You might have to look a couple pics of an opened-up Fuze to see what I mean.
10:07:30silver_hookI think I get what you mean…
10:07:46 Quit JdGordon (Ping timeout: 240 seconds)
10:13:04amee2kis there a way to run RB in a simulator? so i don't need to have a real ipod to try something?
10:14:30Zagoramee2k: yes
10:14:49Guest78023good morning
10:14:50eWillamee2k: there is a link on that page to "unoffical, but trusted" builds, if you don't want to build it yourself.
10:15:59 Nick Guest78023 is now known as Stummi (
10:16:00 Quit Stummi (Changing host)
10:16:00 Join Stummi [0] (Stummi@rockbox/developer/Stummi)
10:16:20 Join wodz [0] (~wodz@
10:17:27gevaertseWill: you know charging time isn't linear, right?
10:18:11eWillare you referring to RB charging faster than the OF?
10:18:58gevaertsI'm referring to your claim that rockbox charges faster than the OF, yes
10:19:00amee2keWill: hmmm looks like the precompiled versions are only for legacy os stuff
10:19:42gevaertsCharging to 87% is going to take much less than 87% of the time needed to charge to 100%
10:19:54amee2kbut the simulator sounds kinda cool. i think i'll try building one until i have the ipod fixed
10:20:52eWillgevaerts: ooooh, and IIRC lipo's never need a full charge to stay healthy. Cool. Thanks.
10:22:54wodzDo we have some driver which has nice crossplatform api? I am looking for en example how to shape i2c driver cleanly
10:24:04silver_hookHmmm, how often does RB release?
10:24:34Zagorsilver_hook: feature releases (x.y) every three months. bugfix releases (x.y.z) in between.
10:24:51silver_hookZagor: Cool, thanks :)
10:24:58Zagordaily builds daily :-) bleeding edge builds for every single commit.
10:25:08Zagortake your pick :)
10:26:50silver_hookOK, so I suppose I'll get meself a Sansa Clip+ for Christmas and flash it with RB as soon as the USB issue is solved in stable.
10:27:08gevaertsWhy wait for that?
10:27:10silver_hook…this way I'll be able to see the diff. between OF and RB as well ;)
10:27:44silver_hookgevaerts: Em, doesn't the USB issue make the USB connection unpredictable?
10:28:20gevaertsIf our USB stack were enabled, yes, but it isn't
10:28:23eWillYou just connect in OF to do any file transfers.
10:28:32gevaertsIt just means you have to reboot to the OF to copy stuff
10:28:43silver_hookeWill: You're dualbooting the bugger?
10:28:54eWilleveryone is
10:29:26silver_hookI can't decide whether this is mega ultra awesome or just stupid waste of space…
10:29:59gevaertsIt's neither
10:30:09wodzis it sufficient to edit buildserver/builds to add target to build system?
10:30:18silver_hookgevaerts: What then?
10:30:29Zagorwodz: yes
10:30:40gevaertsDualboot isn't rocket science, and the OF doesn't take much space
10:30:46*eWill thinks Rockbox _is_ mega ultra awesome
10:30:50wodzand what about tools/ ?
10:31:37gevaertswodz: that one gets important as soon as you're unstable
10:31:42Zagorwodz: yes, you need tools/configure and tools/ done before anyone can build them. buildserver/builds is the last step, to instruct the farm to include that target in the commit builds.
10:32:00gevaertsIs needed to just build?
10:32:24wodzit isn't
10:32:52silver_hookgevaerts: Oh, OK. I just thought that there was very limited space to flash. Like e.g. in routers where you have barely space for one firmware.
10:33:09Zagorno it isn't, but it should be there to make it possible to include in the web site
10:33:29CIA-7New commit by wodz (r28702): Add HD300 to
10:34:16eWillsilver_hook: also, I saved ~200 MB of space by deleting a folder that the OF uses for the database. Though now OF won't boot (only lets me charge/transer-files, then shuts down). In case anyone is wondering −− it's the "##PORT#" dir, and you have almost fill the device so there isn't enough space to re-create it.
10:34:41TheSevenBuschel: "svn's config is correct for 96MHz and we are lucky when running the nano at 48 MHz" << that's what I expected as I copied those settings from apple, which isn't doing CPU frequency scaling
10:34:54silver_hookeWill: Cool tip.
10:35:24CIA-7r28702 build result: All green
10:36:27TheSevenBuschel: I still wonder why the CPU frequency has such a great impact on Rockbox's (idle) current consumption, why it doesn't seem to make a noticable difference in emBIOS if the CPU is idling
10:36:29CIA-7New commit by wodz (r28703): Add MPIO HD300 to build system, list the target on the frontpage.
10:36:41wodzZagor: may I ask to update www?
10:37:05Zagorwodz: done
10:38:35 Join DerPapst [0] (~Alexander@
10:38:40eWillsilver_hook: Oh, and preventing that folder from creation also by-passes "refreshing your database" every time you connect with OF.
10:38:43 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
10:39:30*wodz is looking for volunteer to prepare HD300 svg
10:43:33 Join ReimuHakurei [0] (~reimu@
10:43:49silver_hookeWill: But if you need OF only for file transfer and charging …and if I got it right, RB will solve that issue soon.
10:43:50 Join sideral [0] (~sideral@unaffiliated/sideral)
10:44:43eWillsilver_hook: Oh yeah −− I'm not complaining.
10:45:00eWillit's a plus
10:45:45eWillDB refresh in the OF can take a loooong time if you're tags aren't "perfect"
10:54:45 Quit sideral (Ping timeout: 245 seconds)
10:55:27 Join efyx [0] (
10:57:48***Saving seen data "./dancer.seen"
10:57:52pixelmapetur: do you remember what "Warning: 00000002" in the rec screen (in the "Time:" line) is?
10:58:45peturno, I'd have to look at the code (jhMikeS wrote that), but it is probably a buffer overflow or so
10:59:13peturlike not being able to flush to disk fast enough
10:59:52pixelmaok, thanks. I guess I'll repeat that recording then...
11:00:18 Join sideral [0] (~sideral@unaffiliated/sideral)
11:01:13peturwould be interesting to know what went wrong... disk almost full or so?
11:02:13pixelmashouldn't but I'll check
11:02:24peturI have also seen it happening when there is too much vibration (recorder lying on the stage at a rock concert), causing the disk to fail writing... another reason to move to CF :)\
11:04:11pixelma25GB for a 4 minute recording should be more than enough and this is an SSD ;)
11:04:21pixelmaand I'm at home
11:05:31peturweird to see that warning then
11:06:06amee2k << this article references a file "uisimulator/sdl/button.c" that i can't find in my source tree (that i downloaded with svn 10 minutes ago). am i supposed to create it?
11:06:39pixelmapetur: indeed, I already did some recordings with the same setup and didn't see the warning before
11:07:11peturlet me see what the warning means...
11:07:21pixelmaI'll keep the file for now
11:07:30gevaertsamee2k: no. That page is apparently outdated
11:09:16amee2kyeah... i can't find it anywhere, but lots of files in uisimulator/buttonmap that are apparently per-device type
11:09:49 Quit sideral (Remote host closed the connection)
11:10:54 Join sideral [0] (~sideral@unaffiliated/sideral)
11:14:03peturpixelma: warning 2 = PCMREC_W_ENC_BUFFER_OVF so the buffer of the encoder overflowed. That normally means it didn't flush to disk in time. warning 1 is the PCM buffer overflow that happens if the encoder didn' t encode fast enough. So it looks like a flushing to disk issue. Or maybe the watermarks are set too critical
11:14:05 Quit UnclePervyJesus (Ping timeout: 245 seconds)
11:14:16 Quit eWill (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101026210630])
11:16:11pixelmaI listened to the recording and there is one occasion which sounds a bit like a skipping needle on vinyl
11:16:17pixelmaone spot
11:16:29 Join UnclePervyJesus [0] (
11:16:51pixelmaI record to wave so there shouldn't be an encoder issue
11:17:19peturindeed, it needs to flush a lot of data then...
11:17:29wodzwave also means higher demand for disk throughput
11:17:56silver_hookWell thanks for explaining everything guys!
11:18:19silver_hookI'll make sure my next player (Sansa or otherwise) will rock with RockBox ;)
11:18:35wodzpixelma: have you tried teru's patch to remove this silly // in front of recording files?
11:19:18peturwhat silly // ?
11:19:23pixelmasure, as I said I didn't have an issue before though. I guess I'll also be checking the disk for errors and if there aren't, I'll put it down to bad luck
11:20:23pixelmapetur: in the rec screen, if your recording dir is set to the default which is root, you'll be shown //Rec_* for the file name
11:20:35 Quit UnclePervyJesus (Ping timeout: 245 seconds)
11:21:25pixelmawodz: no, forgot about it :( and my recording dir is currently not the root
11:21:47wodzok, I'll check this on HD300 than
11:22:42pixelma:) congrats for the quick port progress btw.
11:23:59 Join InsDel [0] (
11:24:10wodz:-) I even have first 'success' report from end user
11:24:32pixelmapetur: chksdk didn't find errors and repeating the recording was successful too. I'll keep an eye on it now though.
11:25:18peturpixelma: ok... I wouldn't expect such an issue on an SSD unless it were a slow old CF card...
11:25:45pixelmadidn't see major slowness so far
11:27:07 Join eWill [0] (
11:29:16eWillI just tried to make a voice file with RButil. −−- Encoding of C:/Users/<user name>/AppData/Local/Temp/rbvoice//LANG_VIEW_DYNAMIC_PLAYLIST.wav failed > Encoding aborted −−- (is there supposed to be two slashes before the filename?)
11:29:48 Part silver_hook ("Konversation terminated!")
11:30:45pixelmaI've seen a similar "report" in the forums where the poster said it would fail on different files everytime he tries. Maybe a timing issue?
11:31:48 Quit liar (Ping timeout: 255 seconds)
11:34:02TheSevengevaerts: can you explain how rockbox's ATA infrastructure works?
11:34:31 Join liar [0] (
11:34:31TheSevenis the whole protocol stack handled by the target tree, or only up to the point needed to access the task file?
11:34:40gevaertsTheSeven: not really. I know how things work (more or less) starting from the storage API and up
11:35:14wodzpixelma, teru - the patch to remove // from recording files works for me
11:40:03amee2kwhere do i find a current version of the API documentation for plugins?
11:41:49gevaertsplugin.h *and* grep?
11:41:50*Stummi just asked the same yesterday and gots the same answer
11:42:06amee2kdocumentation != prototypes
11:42:24wodzamee2k: you are free to fix this
11:43:04amee2kto fix it, i need to learn it. to learn it, i need something to learn it from
11:43:22amee2kand prototypes aren't helping if you're starting from scratch ;)
11:43:32Stummiamee2k, look at source or ask here what all the methods are doing ;)
11:43:55Zagoramee2k: docs/API is all we have. yes it's out of date. most people prefer writing code over writing documents.
11:44:14Stummibut you are right. Its a bit anoying for new users (like i am) that there is no usable documentation
11:45:03amee2kwell, if bugging people here about it works i'm fine with that
11:45:34amee2kits just ... an unusual way to document it ;)
11:46:59pixelmayou get the answer "code is the documentation" in quite a few open source project... ;)
11:47:20wodzfortunately codebase for rockbox is not that big
11:47:30amee2kmmmh, yeah i noticed.
11:48:47 Join dfkt [0] (dfkt@unaffiliated/dfkt)
11:48:57 Quit eWill (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101026210630])
11:50:19wodzI'll ask again - do we have some driver with nice crossplatform api?
11:52:12Zagorwodz: I don't understand the question. which platforms are you referring to?
11:52:12 Join PurlingNayuki [0] (~PurlingNa@
11:52:17 Part PurlingNayuki
11:54:28wodzI am thinking how to reshape i2c to be handled uniformly across targets. I mean something like (pseudocode) i2c_write(I2C_INTERFACE_HANDLE, address, buffer, count), i2c_read(I2C_INTERFACE_HANDLE, address, buffer, count), i2c_register(I2C_INTERFACE_HANDLE, i2c_struct) and make i2c_register call platform specific bits of initialization
11:55:57wodzcurrently i2c has a few incompatiblile implementations which forces to use weird wrappers (fmradio, wmcodec) to name a few
11:58:29tmzt_g2rootuse smbus :)
12:01:35 Quit linuxstb_ (Read error: Connection reset by peer)
12:01:53wodztmzt_g2root: no need for, simple unification and standardisation of i2c is enough
12:05:30 Quit sideral (Ping timeout: 259 seconds)
12:10:19 Part LinusN
12:15:55 Join teru [0] (
12:16:15 Quit teru (Client Quit)
12:17:37 Join teru [0] (
12:18:01TheSevenwodz: but *please* don't produce another IF_MD2(whatever, ) hell :)
12:21:38 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
12:28:55CIA-7New commit by teru (r28704): cahnge button mapping in list cotext so that it ignores button release.
12:30:34wodzteru: ping
12:30:43teruwodz: pong
12:31:00wodzyour fix for // in recording filenames works for me
12:31:21CIA-7r28704 build result: All green
12:32:07teruwodz: thank you. i'll commit it then..
12:32:19wodzyou wellcome
12:32:20JdGordonteru: I'm a bit concerned with that commit... you couldnt have possibly tested all those changes and didnt mention it on the mailing list....
12:33:53wodzhmm maybe it would be nice if CIA print link to diff for commit
12:34:44amee2khm... this is odd. plugin.h includes "lcd.h", yet after including the former i don't seem to have prototypes in lcd.h available
12:34:56amee2kbut there only exists one lcd.h in the source tree either
12:38:10JdGordonteru: if that ? was to me, those changes are non trivial and unless they were actually tested on target they could potentially break other actions
12:38:10JdGordonthat *really* needed to get discussed and widely tested
12:38:10 Part Zagor
12:38:10 Join LinusN [0] (~linus@rockbox/developer/LinusN)
12:38:10 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
12:38:22amee2kmmmh, wtf is HAVE_LCD_CHARCELLS?
12:38:57wodzthat's define for charcell lcd (not graphic one)
12:39:05wodzit is used in early archoses
12:39:29JdGordona pox on all our houses is what it is
12:39:40amee2khm. then i kinda get why its not defined for my ipod
12:39:55JdGordonits for the charcell lcd type which is used by one target (archos player) of which there are about 3 working ones in existance
12:39:58amee2kbut isn't graphic LCD kind of a superset of character LCD functionality wise?
12:40:36amee2ki mean, you can always subdivide a graphical LCD into character cells and fake it
12:41:12amee2kso i'm not entirely sure how not having a character display makes API parts unavailable
12:41:23amee2kwhich is kind of a bad thing, eh?
12:41:29Tornewhy is it a bad thing?
12:41:44 Quit InsDel (Read error: Connection reset by peer)
12:41:49Tornethose APIs are useless on pixel displays
12:42:06Tornethe things which make sense for both are available on both
12:42:25Tornee.g. lcd_puts
12:42:36amee2kthen i'm probably making sense where there isn't
12:43:57Tornethe charcell-only APIs are all relating to defining custom characters for the LCD hardware
12:44:18Torneemulating that for things with a graphical display would be a waste of space when those devices can just display bitmaps
12:44:36Tornenobody would want to print monotone fake-graphical icons on a real pixel display :)
12:55:09pixelmateru: I also don't understand the list context commit, I'm not sure what it is for (shall it fix something) and I doubt it has been tested on all the targets it affects
12:56:20teruJdGordon: i checked the code and it doesn't look to break other actions.
12:57:53***Saving seen data "./dancer.seen"
12:59:40pixelmait's not easy to spot sometimes and being affected by button action problems a few times, I'm always in high doubt when things change there. E.g I could imagine that now an action in a linked context could happen on the button release which didn't before because the tree context already "took" the button action
12:59:57pixelmaon some targets
13:05:19 Quit antil33t (Read error: Connection reset by peer)
13:05:29 Join antil33t [0] (
13:19:13amee2ki think i'll just ignore everything that needs charcells then for starters
13:30:49 Quit GeekShadow (Ping timeout: 240 seconds)
13:31:48 Quit stacker55 (Quit: ircN 7.27 + 7.0 for mIRC (2002/01/10 00.00))
13:34:11 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
13:45:19TheSevengevaerts: are there still drives in the wild that don't support at least udma0?
13:49:31wodzdrives or controllers?
13:49:49wodzI don't think so
13:50:00wodzoh wait mybe CFs?
13:51:31wodzat leas CF in my mini reports it is UDMA2 capable
13:51:47CIA-7New commit by teru (r28705): FS #11725: fix duplicationg slash indicating root in filename.
13:53:32JdGordonmy CF card is a druggy!
13:53:37JdGordonMDMA support 0,1,2
13:53:51CIA-7r28705 build result: All green
13:58:53TheSevenMDMA != UDMA
13:59:17TheSevenaccording to wikipedia, UDMA CF cards are available since ~2004, which is not that long
14:03:19TheSevenwodz: once you're done with I2C, you can continue with ATA :/
14:03:41TheSevenfirmware/drivers/ata.c smells PP-ish
14:05:58gevaertsTheSeven: that sounds unlikely. It dates back to the Archoses
14:06:19TheSevenhm, at least it smells target-specific
14:06:36 Quit user890104 (Ping timeout: 272 seconds)
14:07:21wodzthat's first draft - it borrows heavily from generic_i2c.c
14:08:18 Join kevku [0] (~kevku@2001:7d0:0:f000::135d)
14:08:42wodzI don't know if this will work properly for Coldfire in this form as i2c controller in CF is weird in that you should set STOP before reading last byte and disable ACK 2 bytes before end (according to datasheet)
14:08:44 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
14:09:38 Quit xxcv (Ping timeout: 276 seconds)
14:12:42 Quit JdGordon (Ping timeout: 264 seconds)
14:17:59 Join krabador [0] (
14:21:02 Join user890104 [0] (
14:28:27CIA-7New commit by wodz (r28706): MPIO - drop wait states for LCD access
14:28:31CIA-7New commit by wodz (r28707): HD300 - decrease touchstrip sensitivity
14:29:13 Quit factor (Ping timeout: 265 seconds)
14:30:37CIA-7r28706 build result: All green
14:32:41CIA-7r28707 build result: All green
14:34:16 Join saratoga [0] (9803c57f@gateway/web/freenode/ip.
14:35:45saratogaamiconn: I just noticed you reverted r27289, was there some reason to do that?
14:43:50amee2kweee... i've segfaulted the simulator
14:45:26wodzonly once? you should practice more :-)
14:45:28 Quit efyx (Remote host closed the connection)
14:45:44amee2kfunny thing is i didn't even make any modifications to the source yet
14:46:07 Join efyx [0] (~efyx@
14:46:47amee2kand its reproducible. i tried playing an mpeg video... starts decoding audio, then goes mad_header_decode failed, mad_stream-sync failed, out of memory
14:47:20amee2klots of "SDL_Wait_Event() error" messages, then it dies
14:48:14amee2khm. its only a segfault if you have something else playing before you switch to the video
14:48:36amee2kotherwise its an assertion failure
14:49:19wodzgdb it and take the backtrace
14:51:09amee2k << thats the point where it fails... how do i get a backtrace out of this gdb thing?!
14:52:05amee2k << this one?
14:52:47*amee2k can count the number of times he has used gbd before off on his balls
14:54:34 Quit saratoga (Quit: Page closed)
14:55:20 Quit timccc (Ping timeout: 265 seconds)
14:56:45amee2kdid i find something worth fixing?
14:57:57***Saving seen data "./dancer.seen"
15:01:07 Quit plombier (Quit: plombier)
15:04:39 Quit GeekShadow (Ping timeout: 255 seconds)
15:06:27 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
15:09:49 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
15:10:01 Part Zagor
15:10:36 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
15:11:03Zagoramee2k: it depends. can you make it happen again?
15:11:18 Quit TBCOOL (Read error: Operation timed out)
15:14:06TheSevenZagor: he said so above
15:14:44amee2kyeah, i tried it 5 or 6 times now
15:14:47*TheSeven wonders if anyone can explain to him what the actual differences between PIO, MDMA and UDMA from a drive's perspective are
15:15:11CIA-7New commit by pamaury (r28708): sbinfo: add full support for various kind of cryptographic checks ...
15:15:21 Join timccc [0] (~timccc@
15:15:36 Join komputes [0] (~komputes@ubuntu/member/komputes)
15:16:08amee2kif i go for the video right after start up, it sometimes fails with an assertion failure (not very reproducible), but it *always* crashes on this video
15:16:14amee2keither assertion failure or segfault
15:16:23wodzdma transfers utilize more signal lines (at least on PATA)
15:16:40amee2kcould be corrupteed video file, but then it should refuse playback instead of crash
15:16:47amee2kshould i upload the video in question somewhere?
15:17:01CIA-7r28708 build result: All green
15:17:13TheSevendoes it happen with other videos as well?
15:17:52amee2khm... no, a different one plays fine
15:19:04pamauryI think we now have nearly all the information to write an open source version of elftosb2 for the fuze+ if needed \o/
15:19:09amee2kboth are MPEG-1 video and MPEG-1 layer 2 audio. the one that crashes is 640x480 @ 24fps
15:19:18amee2kthe one that works is 320x180 @ 30fps
15:20:01amee2kboth play fine on my desktop system in mplayer
15:20:05 Join anewuser [0] (anewuser@unaffiliated/anewuser)
15:21:42gevaertsYou know that the mpegplayer plugin doesn't scale, right?
15:21:59amee2kyeah, i only see the top left corner of the one that plays
15:22:16 Quit antil33t (Read error: Connection reset by peer)
15:22:25 Join antil33t [0] (
15:22:50 Join TBCOOL [0] (
15:28:20 Join saratoga [0] (9803c6dd@gateway/web/freenode/ip.
15:28:39saratogamaybe we need to do bounds checking and refuse to play very large videos
15:30:05amee2khmmm you think its a size issue?
15:30:57amee2ki don't know much about video compression but i think you can't decompress a frame partially because the next incremental frame can move off-screen stuff into the visible area
15:31:18amee2kand going back to render the missing part would suck big time
15:31:31saratogatheres limited memory available for decoding frames, presumably if you try and watch a very large one you will run out of memory and crash
15:31:40saratogaalthough i don't know for sure thats what your problem is
15:31:51amee2kwell, it did complain about memory before crashing
15:33:00saratogaapps/plugins/mpegplayer/alloc.c:91:static void * mpeg_malloc_internal (unsigned char *mallocbuf,
15:33:18saratogayou could put a DEBUGF there and see if it runs out of malloc space when it crashes
15:33:37saratogaor even just check to see if malloc ever returns -1 and print a message if it does
15:33:43saratogathat should tell you if its lack of memory
15:34:13amee2k << line 245
15:35:24amee2khmmm and i randomly get this assertion failure sometimes instead of the segfault
15:35:28amee2krockboxui: ../../src/xcb_lock.c:77: _XGetXCBBuffer: Assertion `((int) ((xcb_req) - (dpy->request)) >= 0)' failed.
15:35:43 Quit teru (Quit: Quit)
15:35:56amee2khappened like two out of 7 or 8 tries
15:36:19saratogawell that error looks pretty clear
15:40:03amee2kso its trying to bite off a larger frame buffer than it can handle?
15:41:05saratogai don't know what its actually doing, but it does try to request more memory then is available
15:41:11saratogafor some reason
15:42:30amee2kand that kills the kernel?
15:42:56TheSevenit probably starts doing nonsense once it ran out of memory
15:43:12amee2ki haven't found where the whole memory thing goes down, but can't it just refuse the memory request
15:43:27amee2kand let the plugin that requested it die on its own in a corner
15:43:42TheSevenwell, i'd guess it refused the memory request, got a null pointer, and later tried to access that
15:43:50gevaertsYou're assuming a system with MMU and things like that
15:43:50TheSeven(not realizing that the allocation had failed)
15:44:31TheSeventhe plugin malloc could of course bail out directly if it runs out of memory and kill the plugin
15:44:42amee2kwell, if it can't make room for the request it can still just kick out the plugin
15:44:49amee2kwhere does it need an MMU for that
15:45:01saratogathe plugin is the system
15:45:07saratogaso when it does theres nothing left to run
15:45:31amee2kwell, the exchangable part of the plugin
15:45:38TheSevenif the plugin would be a bit more cautious about that, it could probably terminate safely in an out of memory condition
15:45:44Tornesaratoga: the malloc() could catch being out of memory, though, and exit the plugin with longjmp
15:45:46amee2kthe image of the system i have in my head right now kinda looks like a DOS system
15:45:53Torneamee2k: yeah, it's not like that
15:45:55gevaertsTheSeven: sure, but that's still inside the plugin
15:46:00saratogaunless the plugin detects the failed malloc theres nothing to do but crash
15:46:08TheSevengevaerts: that's what i'm saying
15:46:17amee2kwhere the plugin is a running program, and the kernel is like the BIOS and other interrupt services
15:46:24Torneamee2k: it's not like that
15:46:26Torneat all.
15:46:33Tornethe plugin *and* the rockbox kernel are both the BIOS
15:46:36Tornethere is nothing else
15:46:44Tornethere is no separation/abstraction at all
15:47:12amee2kbut how does it load anything then?
15:47:15Tornethe plugin API calls are just function calls
15:47:21Tornethey're not anything like an interrupt service
15:47:27Tornesame as any computer loads anything
15:47:30Tornecopy it into ram, run it
15:47:32wodzwelcome to the embedded systems RTOSes :-)
15:47:43amee2ki mean, i understand that you can't fit fully featured privilege separation into an embedded system
15:47:46*TheSeven thinks the bios/dos comparison isn't too bad. but what happens if a dos program goes nuts? yes, the system locks up, resets, or whatever
15:47:59Torneall the plugins are linked as if they were part of the main binary
15:47:59amee2kbut not even having a logical separation kinda makes me go 0.0
15:48:05Tornebut they're all linked as if they were the same address
15:48:10Torneso, only one gets loaded into that location at a time
15:48:19Torneit's a different locatoin than the core, though
15:48:22Tornethe plugin buffer, in fact :)
15:48:37Torneour plugins are more like overlays :)
15:48:44Torneif you are used to DOS stuf.
15:48:48amee2khmmm i see
15:48:57Tornethere is a *logical* separation
15:49:00Tornein teh source code.
15:49:16Torneplugins only get to call stuff in pluginapi, they are not linked directly to core in a way that allows them to see core's symbols
15:49:29amee2kwell, dos is a few years back for me too but concepts i grew up with kinda stick lol
15:49:36saratogaamee2k: believe it or not you actually need hardware support to do things like you're suggesting
15:50:10saratogayou can't just tell the code to be separated, instead thats something that hardware must enforce
15:50:17amee2kbut the permanent half of the kernel is still responsible for handling the memory requests?
15:50:23Torneno, not at all
15:50:31Tornethe core doesn't have dynamic memory allocation either
15:50:36Torneall the core's memory is statically allocated
15:50:45saratogaamee2k: you need virtual memory to do that kind of thing
15:50:51amee2kso every plugin has like a statically linked c library on its back?
15:50:52Tornemalloc() for the mpeg player plugin is just a function compiled into the plugin
15:51:05Torneany plugin that needs malloc has a malloc implementation compiled into it, yes
15:51:22Tornesome C library functions are in core, most are just not present at all, in core *or* in plugins
15:51:29Tornewe don't have a full C library at all, in any form
15:51:37Torneonly the bits we have had to implement to make specific core/plugin featurs work
15:52:03*gevaerts thinks we should reimplement rockbox in python
15:52:10Tornegevaerts: i'll get right on that
15:52:11 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
15:52:16*TheSeven falls off his chair
15:52:37gevaertsTorne: I knew we could count on you :)
15:53:46amee2kooh... i can SO identify with point 4 on that list
15:53:46wodzpython is not sexy, let's do it in java!
15:54:56amee2kperl ftw
15:55:08amee2kand bash right behind it
15:55:21saratogai should probably update that nomalloc pge
15:56:01wodzbrainfuck :-P
15:56:29tmzt_g2rootwodz: then you could program on a sansa, awesome
15:56:42amee2kbut only if we get fancy ubuntu-style codenames for the version numbers too
16:00:26*pixelma is reminded of that LANG_DISK_NAME_MMC string again
16:00:48wodzI propose Braindead Sucker for the 1.0 of rockbox reimplementation in brainfuck
16:00:53amee2k"15.4 wanking weasel" ?
16:10:48pixelmarasher: you said " −−changetarget −−from target −−to target −−id ID1 −−inplace *.lang". My commandline knowledge is very coarse... I guess "target" is the features part that needs changing - do I have to put the string in "" or '' and the same with the lang ID?
16:12:08rasherpixelma: quoting everything doesn't hurt
16:12:30 Join MethoS- [0] (~clemens@
16:12:51rasherpixelma: Something like −−from "ipodvideo,sansae200" −−to "ipodvideo,sansae200,somenewtarget"
16:13:43pixelmaah, I think I understood. Thanks for confirmation. I'll try
16:14:10rasherJust make sure to check the diff
16:16:01 Part LinusN
16:17:08 Join ReimuHakurei_ [0] (~reimu@
16:17:14 Quit ReimuHakurei (Read error: Connection reset by peer)
16:18:31pixelmaooh, seems I got it right, now I just need someone to try on a Clip+ :\
16:20:44pixelmaand one needs to try the list context commit, I'm very doubtful that it didn't break something on the different targets
16:24:00wodzthis may introduced bugs which will take some time to spot unfortunately
16:24:28 Quit efyx (Remote host closed the connection)
16:24:43 Join efyx [0] (
16:24:55 Join JesusFreak316 [0] (
16:26:41pixelmaI'm afraid too and I'm still curious why
16:26:58pixelmaguess I need to pick up the started thread in the dev ml
16:27:40 Join panni_ [0] (
16:33:11 Quit dantje (Quit: Ex-Chat)
16:34:54 Join sideral [0] (~sideral@unaffiliated/sideral)
16:42:41 Quit mortalscan (Ping timeout: 245 seconds)
16:42:43 Quit wodz (Ping timeout: 264 seconds)
16:53:04 Quit T44 (Read error: Connection reset by peer)
16:53:05CIA-7New commit by pamaury (r28709): sbinfo: add a file containing key used to encrypt the Fuze+ firmware (it uses the zero key)
16:54:25 Quit benedikt93 (Quit: Bye ;))
16:55:09CIA-7r28709 build result: All green
16:57:57 Quit tchan1 (Quit: WeeChat 0.3.3-dev)
16:57:59***Saving seen data "./dancer.seen"
16:58:34 Part Zagor
16:59:12 Join mortalscan [0] (~mortalsca@
16:59:17 Join toffe82 [0] (
17:02:34 Quit sideral (Remote host closed the connection)
17:03:18 Join sideral [0] (~sideral@unaffiliated/sideral)
17:07:58 Quit GeekShadow (Ping timeout: 255 seconds)
17:14:24 Quit mikroflops (Remote host closed the connection)
17:14:36 Join mikroflops [0] (
17:17:36 Join wodz [0] (
17:26:31 Quit sideral (Ping timeout: 264 seconds)
17:27:36 Join sideral [0] (~sideral@unaffiliated/sideral)
17:30:22*TheSeven wonders what's the point of the READ MULTIPLE SECTORS command, if READ SECTORS can read multiple sectors as well, or even more of them at a time
17:31:08pamauryscatter-gatter ?
17:31:19TheSeveni don't think so
17:31:52TheSeventhey basically seem to do the same thing, with the MULTIPLE variant being limited on the maximum number of sectors at a time
17:32:15TheSevenat least that's how i interpret the ata spec
17:33:52pamauryI don't remember the difference, I haven't read the ata spec from a long time
17:34:19gevaertsDidn't Torne add this?
17:34:55TheSeveni just saw it in the spec, no idea what rockbox is doing
17:36:39pamaurythere is a difference for the interrupts generated (cf spec)
17:37:43pamaurybasically READ SECTOR(S) generate one interrupt per sector and READ MULTIPLE one for all sectors, if I read correctly
17:45:00 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
17:46:14 Quit linuxguy4 (Quit: leaving)
17:48:12 Join Topy44 [0] (
17:49:08 Quit ReimuHakurei_ (Read error: Connection reset by peer)
17:50:10 Part wastrel ("l8r")
17:53:31 Quit antil33t (Read error: Connection reset by peer)
17:53:40 Join antil33t [0] (
18:08:32 Quit petur (Quit: Page closed)
18:08:54 Join stoffel [0] (
18:13:09 Quit mikroflops (Ping timeout: 240 seconds)
18:18:58 Join bertrik [0] (
18:18:58 Quit bertrik (Changing host)
18:18:58 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
18:19:51pamaurybertrik: I nearly completed the analysis of the sb format; even the checksums and all that; the only unknown point is the data section (the 'rsrc')
18:20:23bertrikit contains (among other things) the graphics for the app as far as I could see
18:20:41pamauryyes but is there a format ? How, when, where is it loaded ?
18:22:54bertrikI don't know
18:23:15bertrikthere is a RSRC text in there too, and possibly some version numbers (similar to the STMP header)
18:24:12pamauryI'll try to have a look at elftosb2 but since we don't know the format of the command file...
18:24:19bertrikI can check if I see some kind of sub-block in that chunk
18:24:49bertrikpamaury, huh? I do have an example command file from the chumby project
18:25:20pamauryreally ? how do you specify some data ?
18:25:34pamauryI only focused on the boot sections to be honest
18:25:49bertrikload power_prep;
18:25:59bertrikcall power_prep; etc.
18:26:13bertrikthe last command in the script is a jump
18:26:36pamauryoh you mean if the file is not an executable it is loaded in the data section ?
18:27:19bertrikthis is what I have as an example:
18:28:18pamaurybut these are all executables, so they will be loaded in boot sections no ?
18:29:22bertrikok, I see what you mean, but I don't know
18:30:17pamauryI don't know how to generate a section like the "rsrc" for example; perhaps from the elftosb2 code I can find the syntax but it doesn't help
18:30:37pamauryBut I guess it's not too important
18:34:04 Quit DerPapst (Quit: Leaving.)
18:39:05TorneTheSeven: READ SECTORS does one transfer for each sector, with each one generating an interrupt and each one involving waiting for ready/etc
18:39:24TorneTheSeven: READ MULTIPLE SECTORS does one transfer for them all, you just read N times as much data from the single transfer
18:39:35TorneI added WRITE MULTIPLE to our driver, it was already doing READ MULTIPLE.
18:39:50TheSevenand READ DMA roughly behaves like READ MULTIPLE?
18:40:02Tornethe limit on READ MULTIPLE is because the drive's buffer for outgoing data is finite
18:40:23Torneit also does one actual transaction on the disk heads, on most drives
18:40:43TheSeveni'm still a bit puzzled that my drive only supports 16 sectors at a time, as those drives usually have several megabytes of cache
18:40:49Torne16 is the normal limit
18:41:10 Join Strife89TX [0] (
18:41:32Tornei'm not sure if there's some actual reason
18:41:40Tornegenerally you use DMA so it doesn't matter
18:41:43Torneit's onyl relevant for PIO
18:41:53TheSevenhm, apparently not all CF cards support DMA :/
18:42:23Tornewith DMA there's no ATA programmed transfer cycle at all, so it doesn't matter how many sectors it is
18:42:34Tornethe controller just streams the data as and when the disk makes it available directly to the DMA system
18:42:57Tornelike the normal READ SECTORS but without having to involve the host CPU at all for anything
18:43:06Tornebut yeah, CF cards often don't do DMA.
18:43:13 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
18:44:02TheSevenhm. what's the width of the task file registers? i'm seeing lots of funny code to put 16 bits into 8bit registers in a weird order...
18:44:22TorneATA registers are all 8 bit
18:44:35 Quit swilde (Remote host closed the connection)
18:44:37Torneaddresses/counts are split weirdly because of legacy CHS compatibility :)
18:44:39TorneOh, er
18:44:52Tornebtw, READ MULTIPLE you can still ask for just as many sectors at once
18:45:05Torneif the multisector count is 16 it means it'll give you 16 sectors worth of data per transfer, is all
18:45:26Torneyou can still ask for more than 16 sectors in one command, you just have to do the transfer loop multiple times to get all the data
18:45:58TheSevenso how does lba48 work? just by writing the task file regs exactly twice?
18:46:31Tornedoesn't actually have to be exactly
18:46:38Tornethe registers just all have invisible shadows which contain the *previous* written value
18:46:43Tornewhich is used as a parameter for LBA48 commands
18:46:53Torneyou can write more than twice, and the current and previous values are the ones used. older values are discarded.
18:47:22TheSevenok, so the order doesn't matter
18:47:22Torneit's *moderately* insane
18:47:26Tornethe order does matter
18:47:44Torneiirc the previous values are the top 24 bits and the current values are the bottom 24 bits
18:47:46TheSeveni meant the order of the different registers
18:47:50TorneOh, no.
18:47:57Torneother than the command register
18:48:02TheSevenyes, of course :)
18:48:09Torneyou just have to write each of the params twice, in whatever order, then the command
18:48:38TheSevenand this means that one can basically write an lba48 task file with an lba24 command, and it will work just fine as long as the LBA was <=24 bits?
18:49:01Torneexactly :)
18:49:07 Join ReimuHakurei [0] (~reimu@
18:49:12Tornecunning eh
18:49:19Tornealso insane and ugly. but cunning!
18:49:43TheSevenhm. any particular reason they didn't just use the other 8 bits of the bus?
18:49:47 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
18:49:54TheSevento allow addressing >120GB drives with stoneage controllers?
18:50:16Torneno idea.
18:50:29*Torne hugs his breadboard ATA controller made of switches and LEDs. :)
18:51:05TheSevenoh yeah, this is a fully asynchronous protocol, right?
18:51:38TheSevenso you can poke in the bits one after each other and then toggle the strobe once :)
18:55:39 Quit ReimuHakurei (Ping timeout: 276 seconds)
18:56:13 Quit scorche|sh (Changing host)
18:56:13 Join scorche|sh [0] (~scorche@rockbox/administrator/scorche)
18:56:54Tornegevaerts: ok i don't have it made up at the moment but i have done in the past because someone didn't believe me
18:57:06Torneyou just need the switches and leds and a strobe button :)
18:57:22Tornewith a big enough quick reference sheet you can indeed do ATA by hand
18:57:53Torne(more usefully you can do it with a uC with enough GPIOs)
18:58:03***Saving seen data "./dancer.seen"
18:58:39gevaertsWriting an entire bootsector without errors might be a bit tricky
18:59:04tmzt_g2rootTorne: you can get UDMA CF cards though, for cameras, wouldn't imagine most controllers support it though
18:59:54gevaertstmzt_g2root: the point was that it's not realistically possible yet to require UDMA, so the existence of cards that can do it is somewhat irrelevant here
19:00:15alexbobPtmzt_g2root: hehe, nice name
19:00:26alexbobPI gots me a rooted g2 as well
19:00:44gevaertsyes, as in not support devices that can't handle it
19:00:48tmzt_g2rootalexbobP: I was kind of involved and playing a trick on somebody else and it stuck
19:00:52scorche|shalexbobP: this sort of thing is what #rockbox-community is for ;)
19:00:57tmzt_g2rootand yeah
19:01:30Tornetmzt_g2root: it means that TheSeven has to write the code for non-DMA reads for whatever he's doing, because DMA might not be there. and that's tedious, DMA is easier ;)
19:01:58gevaertsYes, DMA doesn't require little imps togling switches
19:02:04tmzt_g2rootright okay, so PIO as fall back
19:02:17alexbobPscorche|sh: I know, I would have used that if tmzt were there...
19:02:27alexbobPsorry for the offtopicness
19:02:30tmzt_g2rootalexbobP: #g2root if you want to discuss
19:02:41 Quit Keripo1 (Quit: Leaving.)
19:06:32TheSevenTorne: do i understand correctly that MDMA vs. UDMA doesn't make a difference for the highlevel protocol?
19:07:49TheSevenso to read something I basically wait for BSY (CSD bit 7) to go low, set up the task file, and then wait for RDY (CSD bit 6) to go high, and kick off the DMA transfer?
19:08:05TorneTheSeven: yes
19:13:30 Join TheLemonMan [0] (
19:20:04TheSevenTorne: do i even need to wait for RDY when doing MDMA/UDMA?
19:24:04 Quit Judas_PhD (Quit: This is a quitting message)
19:26:50*TheSeven just realized that non-lba48 is actually lba28 and not lba24 :/
19:27:37TheSevencan i even assume that those weird drives support LBA at all!?
19:28:01pamauryhehe, chs is quite a nightmare
19:28:19gevaertsDon't you need LBA starting from a really low number (in modern terms)?
19:28:19TheSevenit isn't much worse than lba28
19:28:47TheSeventhat depends on the number of cylinders that ata supports
19:29:03 Join Sajber^ [0] (~Sajber^
19:29:38 Quit saratoga (Quit: Page closed)
19:30:25TheSevendid they really keep all this insanity for sata?
19:30:58 Join Judas_PhD [0] (
19:31:25 Quit Judas_PhD (Client Quit)
19:32:23 Join kugel [0] (~kugel@rockbox/developer/kugel)
19:34:29 Join DerPapst [0] (
19:35:33pamauryno idea, I never read the sata spec
19:36:55 Join Horscht [0] (~Horschti@xbmc/user/horscht)
19:44:31 Join Judas_PhD [0] (
19:50:38 Join Alchimysta [0] (
19:51:13 Join ReimuHakurei [0] (~reimu@
19:53:22 Quit Alchimysta (Client Quit)
19:53:48 Join Keripo [0] (
19:57:32 Join paulk_ [0] (
19:58:26paulk_hello! The other day, I had problems with my neo freerunner (I couldn't boot the uSD) and I wanted to fix it with my rockbox-ed sans e250, but it wasn't able to read the ext2 partitions…
19:58:33paulk_Is any ext2 planned for rockbox ?
19:59:24 Join ppawel [0] (~ppawel@
19:59:30paulk_that's sad :(
19:59:47gevaertsIt's actually explicitely not wanted
20:00:09paulk_it's too fat code or something ?
20:00:36gevaertsWell, that's an unhappy choice of words (given that we do support FAT), but yes, that's part of it
20:01:06gevaertsAnother part is the code maintenance burden
20:01:57paulk_(yes I should better not use fat when talking about ext2 :-)
20:02:18 Join funman [0] (~fun@rockbox/developer/funman)
20:03:42*gevaerts waves to funman
20:04:00 Join factor [0] (
20:05:08*funman waves back
20:06:06funmanchecking what's new in rockbox, it seems you are busy hacking the fuze+, nice!
20:06:24bertrikpamaury has been busy :)
20:06:57bertrikwe are able to fully decrypt the fuze+ firmware file now
20:08:24funmanhow did you find the key?
20:08:33bertrikand probably able to generate a firmware file ourselves that the fuze+ will accept
20:08:49bertrikit uses the zero key 00..00
20:08:59funmanencryption at its best!
20:09:22 Quit Dreamxtreme (Ping timeout: 260 seconds)
20:09:52bertrikI don't know if there is some kind of recovery in case a bad firmware is sent
20:10:19gevaertsDo any of the people working on this actually have a fuze+?
20:10:36bertrikno, as far as I know
20:12:12bertriknot a rockbox dev, anyway. But there is someone who already took it apart and took pictures.
20:12:54 Quit paulk_ (Quit: leaving)
20:14:10funmanaccording to git log, nobody is working on the meizus :(
20:14:38bertrikyeah, they're stalled
20:15:14funmani spent some nights on m6sl LCD but i didn't get anywhere
20:15:45gevaertsAh, I feel less stupid now :)
20:16:09*amiconn still didn't figure out the Elio LCD either :(
20:16:36 Join InsDel [0] (
20:16:42*TheSeven is still fighting with the ipod classic
20:17:01funmanTheSeven: at which stage?
20:17:19TheSeventrying to get ATA running right now
20:17:28TheSevengot an identify packet yesterday :)
20:17:52TheSevenLCD, backlight, USB, SDRAM etc. are working
20:18:24 Quit krabador (Ping timeout: 272 seconds)
20:18:59funmanwho needs storage, put all your songs in ram ;)
20:19:15TheSevenyeah, one could almost do that
20:19:38funmanpamaury: i've been running your patch for amsv2 usb and it works fine for me
20:19:42TheSevenbtw, with a bit of firmware hacking, one could hook up SD cards to an ipod classic
20:23:33 Join Dreamxtreme [0] (~Dre@
20:24:16kugelhm, mp3data.c uses audiobuf in a nasty way
20:25:11 Quit bluebrother (Ping timeout: 265 seconds)
20:25:52 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
20:25:57TheSevenTorne: do i need to care about CRC at any point, or is that completely handled by the hardware? (usually)
20:26:18 Quit factor (Read error: Connection reset by peer)
20:26:32 Join factor [0] (
20:34:15 Part gammy_
20:35:01*kugel built the metadata parsers as a library
20:38:04 Join Buschel [0] (
20:39:21 Join mikroflops [0] (
20:41:51 Quit sideral (Quit: Leaving.)
20:46:27 Quit TheLemonMan (Ping timeout: 245 seconds)
20:51:16 Quit Keripo (Quit: Leaving.)
20:53:23*Buschel gives up...
20:53:50Buschelthe nano 2g hung up with the settings which I expected to be correct for 24 MHz HCLK
20:54:03 Quit funman (Quit: free(random());)
20:58:07***Saving seen data "./dancer.seen"
20:58:33 Join Keripo [0] (
20:59:03 Join TheLemonMan [0] (
21:05:10 Quit mikroflops (Ping timeout: 240 seconds)
21:07:18kugelGodEater: you were interested in the metadata parsers as standalone library?
21:19:09TheSevenis there any document that describes, which status register bit (mainly BSY, RDY, DRQ, ...) gets set in which condition, and which bits one needs to check for to detect completion of various command types?
21:22:44 Quit kevku (Quit: KVIrc 4.0.2 Insomnia
21:22:51amiconnThe ata specs... btw, why are you doing this?
21:23:02amiconnIs our ata driver not sufficient?
21:29:20 Join Kupop [0] (
21:30:45 Quit hebz0rl (Quit: Leaving)
21:30:53 Quit ppawel (Quit: Leaving)
21:33:52 Join GeekShad0w [0] (
21:34:06 Quit GeekShadow (Ping timeout: 265 seconds)
21:41:21 Join ReimuHakurei_ [0] (~reimu@
21:42:46 Quit ReimuHakurei (Read error: Connection reset by peer)
21:42:57 Join robin0800 [0] (
21:44:22*Buschel played with DRAM settings again
21:46:42Buschelchanging the timings with HCLK can speed up DRAM by 20-30% (write) for boosted and 40% (write).
21:48:11*TheSeven hates samsung datasheets
21:48:28TheSevenamiconn: because i didn't find a target-independent ata driver
21:50:34 Join Buschel_ [0] (
21:50:59 Quit Buschel (Ping timeout: 250 seconds)
21:51:01 Nick Buschel_ is now known as Buschel (
21:54:21Buschelfor mpc this equals +1-2% speed in unboosted state, nearly unmeasurable for boosted
21:58:39 Join saratoga [0] (9803c6dd@gateway/web/freenode/ip.
22:06:55 Quit Keripo (Quit: Leaving.)
22:07:09 Quit Stummi (Excess Flood)
22:07:39 Join Stummi [0] (Stummi@rockbox/developer/Stummi)
22:11:24GodEaterkugel: I was - I started trying to port them into one in January I think
22:11:34GodEaterbut I've had a complete lack of time to continue it
22:18:00 Quit DSStrife89 (Quit: ClIRC - IRC client for Nintendo DS)
22:19:53kugelGodEater: I've managed to make one with a medium amount of hackery
22:25:57 Join Keripo [0] (
22:27:17 Quit stoffel (Remote host closed the connection)
22:32:20 Quit komputes (Quit: I haven't slept for ten days, because that would be too long.)
22:33:25 Nick jepler- is now known as jepler (~jepler@emc/developer/pdpc.professional.jepler)
22:37:04 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
22:42:59 Quit TheLemonMan (Quit: Help me, i got shot! *DIES*)
22:45:13 Quit Sajber^ (Quit: Leaving.)
22:45:40 Join Sajber^ [0] (~Sajber^
22:50:05 Quit liar (Quit: Leaving)
22:51:20 Quit Sajber^ (Read error: Connection reset by peer)
22:51:40 Join Sajber^ [0] (~Sajber^
22:54:10 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
22:56:54 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
22:57:03 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101026210630])
22:57:19 Quit dfkt (Quit: -= SysReset 2.53=- Sic gorgiamus allos subjectatos nunc.)
22:58:11***Saving seen data "./dancer.seen"
22:58:58 Quit markun (Read error: Connection reset by peer)
23:00:09 Quit Sajber^ (Quit: Leaving.)
23:00:17 Join ppawel [0] (
23:07:06pamauryThe fuze+ seems to be a nice player, I would like to see a port happen but I don't know I would be able to port it myself...
23:07:13pamaury*if I
23:08:56 Join JdGordon1 [0] (~jonno@
23:09:00 Quit JdGordon (Read error: Connection reset by peer)
23:11:54 Quit ppawel (Quit: Leaving)
23:13:50saratogai think i'm going to commit FS #11235 - libmad asm tweaks for ARM9 and above
23:16:24 Quit robin0800 (Remote host closed the connection)
23:19:52 Join markun [0] (~markun@rockbox/developer/markun)
23:22:34 Quit JdGordon1 (Ping timeout: 265 seconds)
23:22:56 Quit markun (Read error: Connection reset by peer)
23:23:07 Join LambdaCalculus37 [0] (~rmenes@rockbox/staff/LambdaCalculus37)
23:23:27kugelsaratoga: great, I guess we can enable asm on android then
23:23:50kugel(assiming it's the patch you once gave me, I can't remember)
23:23:58saratogayeah i think so
23:24:08saratogajust going to make sure the output is identical
23:24:09 Quit benedikt93 (Quit: Bye ;))
23:24:37 Join robin0800 [0] (
23:24:46saratogahopefully i'll have the arm11 stuff for mp3 decoding ready soonish
23:25:06 Quit robin0800 (Remote host closed the connection)
23:25:27*LambdaCalculus37 catches up with the logs
23:25:30rasherIs this for mp3 or all codecs getting enabled?
23:25:32 Join robin0800 [0] (
23:25:33 Quit bertrik (Quit: :tiuQ)
23:25:47 Nick magnetodyne is now known as krazykit (
23:26:11kugelrasher: all, it's just that mp3 is currently slower with asm
23:26:39LambdaCalculus37TheSeven: Nice work on the Classic so far. :)
23:26:52saratogahopefully the arm11 version will be a lot faster on android
23:27:07kugelwe don't build for armv6 presently
23:27:40kugelI took the defaults of the ndk, perhaps there are armv5 android devices
23:28:13kugelbuilding for armv6 is only very slightly faster, unless hand crafted asm is involved
23:28:24saratogaits possible people have hacked android onto armv5 devices, but all phones are armv6
23:28:27saratogaeven my old g1 is
23:28:46 Join markun [0] (~markun@rockbox/developer/markun)
23:28:47kugelpeople have hacked it onto armv4 ones too
23:28:56saratogahave you tried optimizing for arm11 or cortex?
23:29:05saratogai expect that to make more of a difference then armv6
23:30:07kugelI've built with -march=armv5 -tune=arm11e-js (or something like that), no noticeable difference except in ape which has asm to exploit simd on armv6
23:30:51saratogathis is with ASM enabled I guess?
23:31:19kugelwell, ape didn't have the advantage in the C run of course
23:31:26saratogaits possible the NDK defaults to arm11 or similar
23:31:33saratogai guess that probably makes sense
23:32:25kugelwell, it doesn't
23:34:56CIA-7New commit by saratoga (r28710): ARM9 optimized synth_full for libmad. Speeds up mp3 decoding by an even 2 MHz on all ARM9 and later devices. Note this is only optimized for arm9 ...
23:35:15saratogai mean if you don't give it a flag it may schedule for something like arm11
23:35:28saratogathat would make sense given that its the ndk's lowest common denominator
23:36:50CIA-7r28710 build result: All green
23:37:15*TheSeven hates nothing more than the transfer succeeding but no data arriving
23:37:58LambdaCalculus37TheSeven: Would you like an additional tester for any of the early code?
23:38:27*TheSeven is still busy making things work on his ipod at least :)
23:38:30TheSevenwhich generations do you have?
23:39:28 Join T44 [0] (
23:39:47 Quit Topy44 (Ping timeout: 245 seconds)
23:41:21LambdaCalculus37TheSeven: 80GB Classic; 1st gen, you could say.
23:41:43*TheSeven has the same one
23:43:41kugelsaratoga: ? without flags or any changes to the default make files it builds for armv5
23:43:56saratogayeah but what does it schedule for?
23:44:44kugelmtune=xscale, if you mean that
23:45:08kugelbut I doubt gcc is capable of optimizing for armv6 if march is armv5
23:45:24saratogawhys that?
23:45:41 Join noamsml_ [0] (
23:45:54saratogai assume you can schedule for any processor capable of running armv5 ops, just it won't produce code using the new ops in armv6
23:46:03 Quit noamsml (Ping timeout: 265 seconds)
23:46:23LambdaCalculus37TheSeven: Besides that I'm quite excited at the prospect of hacking about on my Classic as well. :)
23:46:23kugelyou probably can, but I assume gcc can't
23:47:01 Quit Kupop (Ping timeout: 265 seconds)
23:57:08 Quit LambdaCalculus37 (Quit: This computer has gone to sleep)
23:58:05 Quit pamaury (Remote host closed the connection)

Previous day | Next day