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 2010-12-11

00:01:39 Quit yosafbridge (Ping timeout: 272 seconds)
00:06:15 Quit bertrik (Quit: :tiuQ)
00:07:22 Join yosafbridge [0] (
00:08:53 Quit simonrvn (Ping timeout: 240 seconds)
00:10:45***Saving seen data "./dancer.seen"
00:11:30 Quit ender` (Quit: The word 'politics' is derived from the word 'poly', meaning 'many', and the word 'ticks', meaning 'blood sucking parasites'. -- Larry Hardiman)
00:11:57 Quit tchan (Ping timeout: 265 seconds)
00:16:58 Quit kugel (Ping timeout: 245 seconds)
00:18:29 Quit evilnick_B (Quit: Page closed)
00:18:43 Quit {phoenix} (Remote host closed the connection)
00:24:52 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
00:25:32 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
00:38:07 Quit pamaury (Remote host closed the connection)
00:39:43 Quit casainho (Remote host closed the connection)
00:41:19 Join simonrvn [0] (
00:52:44 Quit liar (Quit: Leaving)
00:58:10 Join kugel [0] (~kugel@rockbox/developer/kugel)
00:59:40 Quit GodEater (Read error: Operation timed out)
01:08:48 Join krazykit` [0] (
01:09:00 Quit krazykit (Ping timeout: 265 seconds)
01:15:48 Join GodEater [0] (
01:15:48 Quit GodEater (Changing host)
01:15:48 Join GodEater [0] (~bibble@rockbox/staff/GodEater)
01:18:09 Quit Horschti (Ping timeout: 264 seconds)
01:36:24 Join saratoga [0] (9803c6dd@gateway/web/freenode/ip.
01:36:46saratogawhat the hell, this rm file is detected as unsupported by the metadata parser, but playback STILL loads a (completely random) codec to try and decode it
01:37:05 Quit mc2739 (Quit: leaving)
01:37:45 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
01:38:05saratogaand grep tells me playback.c does not check the return value on get_metadata
01:38:08saratogahow is this possible
01:40:03saratogais there anyone that knows how metadata parsing is supposed to work
01:42:24 Join kugel_ [0] (
01:42:36 Quit kugel (Disconnected by services)
01:43:10 Quit kugel_ (Remote host closed the connection)
01:43:20 Join kugel [0] (~kugel@rockbox/developer/kugel)
01:48:28 Quit hebz0rl (Quit: night)
01:49:33 Join shai_ [0] (
01:52:19saratogahmm no maybe thats ok in buffering.c
01:53:23 Quit shai (Ping timeout: 260 seconds)
01:53:29saratogaah got it
01:57:13 Quit kugel (Remote host closed the connection)
01:58:22 Quit t0rc (Quit: Give someone code, help them with one project. Teach someone to code, help them rule the world.)
01:58:33 Quit Judas_PhD (Quit: This is a quitting message)
01:59:59CIA-7New commit by saratoga (r28788): Correct mistake in the RealMedia parser that prevented rejecting files with unsupported codecs.
02:01:56CIA-7r28788 build result: All green
02:02:48saratogarasher: no commit logging so that i get more test files to fix
02:03:29saratoganow commit i mean to say
02:04:03ej0rgesaratoga: cute - congrats on the quick fix too
02:04:38the_Kylesaratoga: 11541 looked like a really good patch. What would need to be done to try to get this patch into my custom build? Would I have to try to apply each piece manually?
02:10:46***Saving seen data "./dancer.seen"
02:15:44 Join shai__ [0] (
02:18:27 Quit GodEater (Ping timeout: 272 seconds)
02:19:04 Quit shai_ (Ping timeout: 245 seconds)
02:25:44saratogathe_Kyle: yes
02:26:13saratogaej0rge: if you still have db problems let me know
02:27:28ej0rgesaratoga: will do
02:29:11the_KyleWell, I'm up to the challenge, I guess. Should I resubmit against latest trunk if I can get it working? Seems it needs a bump.
02:32:23 Join CaptainKewl [0] (
02:32:41the_KyleI would of course resubmit as a comment to 11541 rather than starting a new task for it.
02:32:50 Join [Saint] [0] (cbb802e9@gateway/web/freenode/ip.
02:33:11 Quit Rob2222 (Ping timeout: 240 seconds)
02:35:31 Quit dfkt (Quit: -= SysReset 2.53=- Sic gorgiamus allos subjectatos nunc.)
02:37:56 Quit mortalscan (Ping timeout: 265 seconds)
02:39:50 Quit [Saint] (Ping timeout: 265 seconds)
02:44:43 Quit BHSPitMonkey (Ping timeout: 260 seconds)
03:07:47preglowhow does one exit the android app?
03:12:08 Quit moos (Ping timeout: 260 seconds)
03:26:23 Quit factor (Ping timeout: 240 seconds)
03:26:30 Join Judas_PhD [0] (
03:29:21 Join GodEater [0] (
03:29:25 Quit GodEater (Changing host)
03:29:25 Join GodEater [0] (~bibble@rockbox/staff/GodEater)
03:35:19 Nick krazykit` is now known as krazykit (
03:36:19 Quit pixelma (Disconnected by services)
03:36:22 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
03:36:24 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
03:36:36 Quit amiconn (Disconnected by services)
03:36:37 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
03:36:39 Quit Sudos (Ping timeout: 255 seconds)
03:36:57 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
03:41:15 Join simabeis_ [0] (
03:42:22 Join Sudos [0] (
03:43:49 Quit simabeis (Ping timeout: 255 seconds)
03:44:14 Join ZhangNing [0] (~ZhangNing@
03:45:46 Quit MethoS- (Remote host closed the connection)
03:50:19 Quit Judas_PhD (Quit: This is a quitting message)
03:51:04saratogapreglow: i don't think you can
03:52:54 Join T44 [0] (
03:56:47 Quit Topy44 (Ping timeout: 240 seconds)
03:57:52 Join factor [0] (
03:58:17 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
04:01:22 Quit yosafbridge (Quit: Coyote finally caught me)
04:10:47***Saving seen data "./dancer.seen"
04:11:04 Join S_a_i_n_t [0] (S_a_i_n_t@
04:24:22 Join JesusFreak316 [0] (
04:33:08 Quit amiconn (Disconnected by services)
04:33:09 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:33:19 Quit pixelma (Disconnected by services)
04:33:21 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:33:24 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:33:29 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:37:04 Quit TheSeven (Ping timeout: 272 seconds)
04:40:15 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:42:48 Join Barahir_ [0] (
04:45:38 Quit krabador (Quit: Sto andando via)
04:46:16 Quit Barahir (Ping timeout: 264 seconds)
04:56:02 Quit sasquatch (Quit: WeeChat 0.3.2)
04:56:27 Join sasquatch [0] (
04:59:32 Join fdinel [0] (
04:59:52 Quit efyx (Remote host closed the connection)
05:00:27 Quit fdinel (Client Quit)
05:04:53 Quit Bushmills (Ping timeout: 264 seconds)
05:14:19 Quit ZhangNing (Ping timeout: 240 seconds)
05:24:39 Quit eWill (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
05:28:11 Quit antil33t ()
05:30:57CIA-7New commit by saratoga (r28789): Commit FS #11776 by Jonas Haggqvist. Enables option to log building the database in order to trouble shoot crashes.
05:32:54CIA-7r28789 build result: All green
05:33:48 Quit InsDel (Read error: Connection reset by peer)
05:47:08 Quit LambdaCalculus37 (Quit: zoom)
05:58:30CIA-7New commit by saratoga (r28790): Commit small piece of FS #11748 by Michael Hohmuth. Disables database specific fields if the database is not compiled in.
06:00:55CIA-7r28790 build result: All green
06:01:29 Quit simonrvn (Ping timeout: 240 seconds)
06:06:04 Quit JesusFreak316 (Ping timeout: 245 seconds)
06:10:48***Saving seen data "./dancer.seen"
06:12:46 Join Bushmills [0] (
06:12:46 Join The_Pwny [0] (
06:35:01 Join antil33t [0] (
06:35:25 Quit antil33t (Client Quit)
06:35:40 Join antil33t [0] (
06:44:02 Quit panni_ (Read error: Connection reset by peer)
06:45:20 Quit ReimuHakurei (Quit: Stand by, Ready!)
06:46:46 Join ReimuHakurei [0] (~reimu@
06:48:16 Quit ReimuHakurei (Read error: Connection reset by peer)
06:48:30 Join ReimuHakurei [0] (~reimu@
06:48:53 Quit ReimuHakurei (Read error: Connection reset by peer)
06:49:12 Join ReimuHakurei [0] (~reimu@
06:52:00 Quit ReimuHakurei (Read error: Connection reset by peer)
06:52:08 Join ReimuHakurei [0] (~reimu@
06:56:57 Join yosafbridge [0] (
06:58:24 Quit ReimuHakurei (Read error: Connection reset by peer)
07:14:31 Quit yosafbridge (Quit: Coyote finally caught me)
07:14:42 Join yosafbridge [0] (
07:38:42 Quit factor (Read error: Connection reset by peer)
07:40:57 Join simonrvn [0] (
07:56:04 Join factor [0] (
08:06:13 Join Horscht [0] (~Horschti@xbmc/user/horscht)
08:10:49***Saving seen data "./dancer.seen"
08:24:46 Join mortalscan [0] (~mortalsca@
08:27:18 Quit BHSPitMonkey (Read error: Connection reset by peer)
08:36:06 Quit shai__ (Quit: Leaving)
08:59:43 Join bmbl [0] (
08:59:43 Quit bmbl (Changing host)
08:59:43 Join bmbl [0] (~bmbl@unaffiliated/bmbl)
09:02:37 Join Judas_PhD [0] (
09:31:42 Join user890104 [0] (
09:34:46 Quit Stummi (Excess Flood)
09:34:55 Join Stummi [0] (Stummi@rockbox/developer/Stummi)
09:34:58 Quit Stummi (Excess Flood)
09:35:23 Join Guest8118 [0] (
09:35:27 Quit Guest8118 (Excess Flood)
09:35:59 Join Stummi_ [0] (Stummi@rockbox/developer/Stummi)
09:47:05 Join Buschel [0] (
10:00:13 Join slooopy [0] (
10:00:18 Quit Judas_PhD (Quit: This is a quitting message)
10:05:51 Quit linuxstb (Read error: Connection reset by peer)
10:06:02 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
10:08:23 Join Judas_PhD [0] (
10:10:53***Saving seen data "./dancer.seen"
10:26:25 Join bertrik [0] (
10:26:25 Quit bertrik (Changing host)
10:26:25 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
10:39:37CIA-7New commit by Buschel (r28791): Derive clock and timer defines from frequency of external source.
10:41:26CIA-7New commit by Buschel (r28792): Detail comment in timer configuration of S5L870x.
10:41:50CIA-7r28791 build result: All green
10:42:44CIA-7New commit by Buschel (r28793): Set DRAM configuration for iPod nano 2G in crt0.s
10:43:58CIA-7r28792 build result: All green
10:46:16CIA-7r28793 build result: All green
10:55:53 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
10:57:43 Part Bushmills
10:59:13 Nick S_a_i_n_t is now known as [Saint] (S_a_i_n_t@
10:59:50 Quit bmbl (Ping timeout: 240 seconds)
11:01:20 Quit Judas_PhD (Quit: This is a quitting message)
11:02:55 Join bmbl [0] (~bmbl@unaffiliated/bmbl)
11:14:02 Join dfkt [0] (dfkt@unaffiliated/dfkt)
11:21:36 Join pamaury [0] (
11:21:36 Quit pamaury (Changing host)
11:21:36 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
11:22:20bertrikYuk, several tricky problems have been appearing lately on the ams sansas: 1) fuze v2 not recognising the tuner chip and hanging 2) (big) SD cards not working on the clip+ 3) clip+ display being dim and player getting warm
11:22:37bertrik(only on some players apparently)
11:23:21 Join bimbel [0] (~bmbl@unaffiliated/bmbl)
11:24:51 Quit bmbl (Ping timeout: 245 seconds)
11:27:43 Join TheLemonMan [0] (
11:29:11 Nick bimbel is now known as bmbl (~bmbl@unaffiliated/bmbl)
11:31:26 Join {phoenix} [0] (
11:35:48[Saint]Is anyone able to tell me what the largest capacity disk is that I could expect to put in an iPod Colour/Photo?
11:38:42CIA-7New commit by Buschel (r28794): iPod nano 2G: Call lcd_update_rect() in lcd_update() instead of using an own implementation.
11:40:35CIA-7r28794 build result: All green
11:44:08 Join stoffel [0] (
11:49:38preglowsaratoga: apparently i can't at all, since i got 2.2
11:49:43preglowtask killer leaves it be now
11:50:11JdGordonkugel: I just tried your patched android build on my stock g2 and it is 0% cpu and doesnt appear to be waking up at all which is good, so maybe its a cyanogenmod thing, that said I still don't really like the way ticks are done
11:51:37amee2kkugel is here? 0.0
11:52:29[Saint]amee2k: logs
11:52:41amee2kooh, right
11:52:58amee2ki think he is the only one i know who uses logs as bnc replacement >_>
11:54:15[Saint]sliding offtopic here, but lots of people use the logs instead of idling in channel.
12:00:12amee2komg... the network card on my rockbox laptop is so going on my ***** >_<
12:02:27amee2kis there a way to redirect kernel messages like "[ 1234.5678] eth0: coax cable problem" to something other than the currently selected terminal?!
12:02:51pixelmacan a thread in our forums be split so that you can also split posts in themselves?
12:03:17amee2kthe driver for some reason keeps insisting that the non-existent coax interface on my network card is, well, non-existent
12:03:22pixelmaI mean if a post contains statements about two different topics
12:06:38 Join moos [0] (moos@rockbox/staff/moos)
12:08:13 Quit slooopy (Ping timeout: 265 seconds)
12:09:27bertrikI have no idea
12:10:55***Saving seen data "./dancer.seen"
12:21:31 Join slooopy [0] (
12:38:32 Join kugel [0] (~kugel@rockbox/developer/kugel)
12:39:31 Join kevku [0] (~kevku@2001:7d0:0:f000::135d)
12:49:47 Quit The_Pwny (Quit: Oops. My brain just hit a bad sector)
12:55:27 Join stripwax [0] (
12:56:11 Join DerPapst [0] (
12:56:58 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
13:10:20 Join shai [0] (
13:11:09*Buschel just sped up nano 2g lcd by 50% :)
13:12:22bertrikkugel, are you familiar with the ams/amsv2 sd driver?
13:13:21bertrikseveral people seem to have problems with sd cards on the clip+
13:14:19bertrikso I'm looking a bit into the driver. I plan to make the amsv1 and amsv2 drivers more similar, the amsv1 sd driver seems to do some things that the amsv2 sd driver does not.
13:14:51kugelbertrik: no sorry
13:14:52bertrikAlso I'm still a bit puzzled by the mci_delay()s a bit, they seem to be present whenever we *dont'* expect a response from the card
13:15:04bertrikbut not consistently
13:15:11kugelbut FlynDice worked a lot on the v1 driver, which he did not on the v2 one IIRC
13:15:40bertrikyeah, I guess I need funman and FlynDice, but they both haven't been here for a while
13:15:52kugeldidn't the v1 one change all commands to expect a response which removed the need for the delays?
13:16:37bertrikI indeed remember a discussion about that, but I'm not sure if that was ever implemented
13:21:23bertrikah yes, this was implemented for amsv1 sd driver in 27066, but there are still quite a few mci_delay()s left
13:22:59ranmachanYeah, I did some changes to the v1 driver that removed a few mci_delay()
13:26:32[Saint]Buschel: Oh? How so?
13:28:17BuschelSaint: playing a bit with the FIFO status :)
13:28:18bertrikranmachan, I'd like to apply that to the ams v2 driver too. What do I need to test to make sure that it doesn't break anything?
13:28:27 Quit stripwax (Quit:
13:28:50[Saint]Buschel: You work on Nano2G a fair bit, do you know what's wrong with "System - Debug (Keep Out!) - View disk info"?
13:29:10[Saint]it seems to repeat the same set of info ~5 times
13:29:35Buschel[Saint]: never looked at it...
13:30:17Buschelyes, it's repeated
13:31:13[Saint]I had a quick look, but couldn't see why it does that.
13:31:15ranmachanbertrik: You could just try if the r27066 patch applies similarily to the amsv2 code and sd cards still work fine with that
13:34:01ranmachanThe Physical Layer Simplified Spec should list the response types to various commands
13:34:12ranmachanOnly a few commands don't send a response back
13:34:54Buschel[Saint]: I cannot see this at a first glance as well
13:35:06Buschel[Saint]: what LCD type does your nano have?
13:35:18bertrikranmachan, ok, then it sounds like a very good thing to actually receive the response instead of inserting some random delay :)
13:37:01 Quit bluebrother (Disconnected by services)
13:37:02 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
13:37:17[Saint]Buschel: I have mostly type 1, (7) LSD176 LCDs in my Nano2Gs
13:37:52Buschel[Saint]: would be great, if you could test my lcd driver changes on your nanos as well
13:38:11[Saint]yeah, that's sweet.
13:38:22[Saint]Is there an FS for it yet?
13:39:17Buschelnot yet ->
13:40:09 Join Luca_S [0] (
13:40:18 Join MethoS- [0] (~clemens@
13:41:16Luca_Shello - re: SD driver on ams, I can test patches on a fuzev2 if needed
13:43:36Luca_SI remember back from the initial porting days that my uSD did not work with the initial delays and started working when changing a couple of commands to expect a response
13:44:45Luca_Stalking about fuzev2, enabling USB breaks the cpu frequency switching :(
13:45:06Luca_S(i mean, the CPU frequency switching patch that's on the tracker)
13:45:08 Join liar [0] (
13:45:08bertrikLuca_S, I plan to put some patches in
13:45:39bertrikyou know how to patch, build and test, right?
13:47:35Luca_Suhm, my linux vm is borked... I can test as soon as I restore an ubuntu vm :D
13:51:38*amee2k nervously rolls kugel around the channel
13:51:44amee2ki hate my life >_<
13:52:16amee2kthe es18xx sound driver in deb testing is broken
13:52:41[Saint]amee2k: I can't imagine a channel that is appropriate for, but it isn't this one.
13:53:36amee2k[Saint]: it is appropriate in so far as i'm trying to get rb to work on the box and wanted to give kugel a status up on it
13:53:55[Saint]"<amee2k> i hate my life >_<"
13:54:45amee2kalso, please just tell me to fsck off in my face, rather than beating around the bush like that
13:57:20 Join _jhMikeS_ [0] (
13:57:20 Quit _jhMikeS_ (Changing host)
13:57:20 Join _jhMikeS_ [0] (~jethead71@rockbox/developer/jhMikeS)
13:57:20 Quit jhMikeS (Disconnected by services)
14:01:53 Quit antil33t (Read error: Connection reset by peer)
14:02:02 Join antil33t [0] (
14:10:56***Saving seen data "./dancer.seen"
14:13:49 Join Kupop [0] (
14:21:38 Quit antil33t (Read error: Connection reset by peer)
14:21:47 Join antil33t [0] (
14:22:04Buschel[Saint]: FS #11807
14:30:28 Join efyx [0] (
14:43:40 Quit slooopy (Ping timeout: 240 seconds)
14:45:50kugelamee2k: what's up?
14:46:42amee2kkugel: the sound driver for my card is screwed up in deb testing :/
14:47:39 Quit Kupop (Ping timeout: 255 seconds)
14:48:43amee2kexactly what i thought, just somewhat less polite
14:57:21 Quit linuxstb (Ping timeout: 245 seconds)
14:57:27 Join slooopy [0] (
14:58:20kugeldid the idle usage change?
15:06:09amee2ki only gave it a very quick try so far, but first indication is no
15:07:01amee2kkugel: but htop now seems to show proper CPU% figures per-process
15:07:17amee2kdidn't try powertop yet though
15:08:46 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
15:12:55pamaurybertrik: did you continue to disassemble the fuze+ firmware ?
15:13:36 Quit linuxstb (Ping timeout: 245 seconds)
15:25:35 Join einhirn [0] (
15:28:30 Quit einhirn (Client Quit)
15:40:18pamauryI see some interesting strings in the fuze+ last stage bootloader
15:40:31pamauryand some C++ things too :(
15:43:19 Quit slooopy (Ping timeout: 265 seconds)
15:45:31 Quit shai (Ping timeout: 240 seconds)
15:48:14bertrikpamaury, no
15:48:35bertrikI did spot a c library, forgot which one it was
15:49:53pamauryok, when I'm done with the bootloader, do you think I should try host or play ?
15:51:49 Join shai [0] (
15:53:57 Quit GeekShadow (Ping timeout: 240 seconds)
15:54:00bertrikpamaury, actually I didn't get that far and am not planning to do more in the near future, so just pick the one you like
15:54:27 Join einhirn [0] (
15:54:38 Quit einhirn (Client Quit)
15:56:34 Join slooopy [0] (
15:57:13 Join einhirn [0] (
16:05:36 Join casainho [0] (~chatzilla@
16:10:57***Saving seen data "./dancer.seen"
16:25:22 Quit casainho (Read error: Operation timed out)
16:32:13 Join noamsml_ [0] (
16:33:03 Join hebz0rl [0] (
16:34:40 Quit noamsml (Ping timeout: 255 seconds)
16:35:58 Quit jfc (Ping timeout: 250 seconds)
16:38:07 Quit Strife89 (Quit: Reboot.)
16:40:42 Quit mortalscan (Ping timeout: 245 seconds)
16:42:07 Quit Feisar- (Remote host closed the connection)
16:43:19 Join Strife89 [0] (
16:43:50 Join ender` [0] (
16:44:06 Join Feisar [0] (
16:44:17 Join jfc [0] (
16:44:32 Nick Feisar is now known as Guest24228 (
16:51:48 Quit Bawitdaba (Read error: Connection reset by peer)
16:52:07 Join Bawitdaba [0] (
16:55:40amee2kkugel: do you think testing performance on the new system has a point if the soundcard isn't working?
16:56:38amee2kbecause i just stalled and dropped out of the sky like a rock, as to what the problem with the card is
16:57:14amee2kso my working theory about it right now is that the driver build is simply screwed up
16:57:51amee2kafter all, its an ISA-PNP device. so it was probably never tested on real metal until now
16:58:14 Quit Strife89 (Quit: Reboot.)
17:00:29amee2ktesting comes with powertop 1.11 so i'll compile the new one too
17:05:37 Join Strife89 [0] (
17:10:44 Join casainho [0] (~chatzilla@
17:10:45 Quit ender` (Quit: If man could be crossed with the cat it would improve man, but it would deteriorate the cat. -- Mark Twain)
17:15:12 Join ender` [0] (
17:27:44pamauryThe last stage bootloader of fuze+ seems to support debug uart, don't know what you can do with it though
17:28:38TheLemonMantheres usually a terminal console over that
17:33:13amee2kwhat keybinding is quit on the SDL app btw?
17:34:54gevaertsI don't think stopping supported :)
17:35:43amee2kkugel: no noticable improvement. htop is still showing 60..70% idle load
17:36:16amee2kthe process list attributes ~50% CPU load to the main process and the rest to its threads
17:36:57amee2k << deb testing, not patched, idling in main menu
17:37:06amee2krecompiling with your patch right now
17:37:46gevaertshm, does the SDL app support "LCD shutdown"? i.e. can it stop repainting the screen
17:38:22 Join shai_ [0] (
17:42:26 Quit shai (Ping timeout: 260 seconds)
17:45:17kugelamee2k: how do you identify the main process?
17:46:32amee2ki switch htop into tree mode (F5)
17:46:43 Quit stoffel (Remote host closed the connection)
17:46:48kugeloh, nice tip :)
17:46:55amee2kit shows one rockbox process on top and like 6 or 7 below it
17:47:07amee2ki suppose the subprocesses are scheduling entities of the threads
17:47:53 Quit TheLemonMan (Quit: free(me))
17:49:18 Join GeekShadow [0] (~Antoine@
17:49:22 Quit GeekShadow (Changing host)
17:49:22 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
17:49:23amee2k (the load figures are off because the compiler is running right now)
17:49:26kugelamee2k: we use sdl thread (which map to pthreads) for the rockbox threads
17:49:37kugelwith my patch we implement a userspace threads
17:50:40kugelwhich reduces the 7+ pthreads to 3 or 4
17:50:49kugel(also cheaper context switching)
17:51:43kugelamee2k: I'm surprised the main thread puts so much load
17:51:50kugelit's mostly sleeping
17:52:15kugelperhaps sem_wait() doesn't work correctly on your system, that's used all over the place for idling
17:52:44amee2kmmh, didn't someone quote the docs a few days ago that under certain conditions it reverts to polling?
17:53:06amee2kand err... busy waits
17:53:28pamauryyeah, someone said on some systems it was implemented as polling every x msec
17:53:34pamaury(from the doc)
17:54:34 Quit ender` (Quit: Where there's a will, there's an inheritance tax.)
17:54:36*TheSeven now has a generic copyright-free exploit for the s5l8702 :)
17:54:48pamaurybut I think is was when there is a timeout
17:54:58pamauryTheSeven: exploit to do what
17:57:25TheSevenrun arbitrary code from DFU mode
17:58:04pamaurydoes this need an exploit ?
17:58:17TheSevenyep :/
17:58:40pamauryis it target specific ? I though a DFU mode would be design to run code :/
17:59:00TheSevenyes, code signed by apple using an RSA certificate
17:59:17TheSevennor our code :)
18:00:02 Join ender` [0] (
18:00:06pamauryhow do you do that ?
18:10:58***Saving seen data "./dancer.seen"
18:10:59Buschel[Saint]: could you already test the lcd patch?
18:12:05amee2kkugel: where in the source is the sem_wait() call in question?
18:12:27amee2ki just tried making a testcase, but it appears to work correctly
18:13:32 Join InsDel [0] (
18:14:01TheSevenpamaury: they have a nice buffer overflow vulnerability in the certificate parser :)
18:14:46pamaurylol, makes the whole thing useless
18:20:24TheSevenyes, but exploiting that wasn't exactly trivial :)
18:21:41*TheSeven thinks it's time to start the rockbox port...
18:22:07TheSevenwho wants to hold hands? :)
18:22:08pamauryI trust you ;)
18:22:14*TheSeven hasn't created a port from scratch before
18:22:31pamaurywhich device ?
18:22:53*pamaury is too busy disassembling the fuze+ for now
18:24:10kugelamee2k: wait_for_interrupt() in my patch, thread-sdl.c in svn
18:25:15casainhowhile trying build my port, I get this error: Error: selected processor does not support `clz r2,r1'
18:25:26amee2koh, thats in the patch?
18:25:44casainhoin support-arm.S
18:25:59amee2ki thought the regular pthreads implementation is using it too
18:26:12amee2kbut grepping the source tree doesn't come up with any matches for me
18:26:56 Quit FOAD (Quit: I'll be back)
18:27:19 Join FOAD [0] (~dok@
18:29:15 Quit kugel (Remote host closed the connection)
18:30:04TheSevenamee2k: which ARM core is that? does it actually support that instruction?
18:30:26TheSevenif yes, there's probably a wrong -mcpu= whalue in tools/configure
18:30:51TheSevenif not, it's probably some wrong define
18:31:00amee2khu? wrong nick?
18:31:13amee2ki'm with the SDL guys here >_>
18:31:28TheSevenah, damn, wanted to ping casainho
18:31:29pamauryyou want to talk with casainho
18:33:30casainhoTheSeven: well, I need to see on datasheet −− 2 minutes
18:34:57casainhoTheSeven: it's ARM926EJ-S
18:35:49TheSeventhat's ARM9e, so it should support clz
18:35:55TheSevenlook at tools/configure in that case
18:37:44casainhoTheSeven: I have this there: arm9tdmicc
18:39:07TheSevenshould be arm926ejscc
18:39:42casainhoTheSeven: thanks ;-)
18:49:40casainhoTheSeven: that was the problem - now resolved, thanks.
18:50:20 Quit MethoS- (Remote host closed the connection)
18:50:47 Quit liar (Quit: Leaving)
18:51:26 Join Strife1989 [0] (
18:52:55CIA-7New commit by Buschel (r28795): Fix typo in comment.
18:53:00 Quit Strife89 (Ping timeout: 240 seconds)
18:54:50CIA-7r28795 build result: All green
18:56:14 Join panni_ [0] (
19:02:34 Quit Luca_S (Quit: CGI:IRC (EOF))
19:06:07 Quit Buschel (Ping timeout: 245 seconds)
19:07:05 Join krabador [0] (
19:08:08 Quit bmbl (Read error: Connection reset by peer)
19:17:57TheSevenwhat scramble magic number do we want to have for the ipod classic?
19:18:08TheSevenip6g? ipcl? something else?
19:20:02LloreanI think 6g is the best, least ambiguous
19:20:14LloreanSince they may call future models the "classic" as well
19:22:31TheSevenwell, there are already three versions of the ipod classic, which we can probably handle with one port
19:22:52TheSevenand it looks like they're dropping the HDD-based series altogether soon
19:23:03 Nick Strife1989 is now known as Strife89 (
19:26:07LloreanDoes any of that make "ipcl" better than "ip6g" though?
19:26:23amee2kkugel: << deb testing, with your patch, idling in main menu
19:27:16amee2kits down to one main process and 3 sub-processes now, with the main one still causing 40..50 CPU% load
19:27:28TheSevenLlorean: yeah, but what about the names in the target tree? target/arm/s5l8702/ipod6g? target/arm/s5l8702/ipodclassic? or even in target/arm/s5l8700?
19:28:01amee2kof the subprocesses, the first one is floating very stable around 7%, the last one around 3%, and the middle one bottomed out at 0%
19:29:02 Join dantje_ [0] (
19:30:13 Join liar [0] (
19:30:42pamauryTheSeven: arm/s5.../ for generic things (if ther are any) and then arm/s5.../ipodxxx/ for ipod specific
19:30:55TheSevenyeah, that's clear :)
19:31:00LloreanTheSeven: I don't see why you'd use "ipodclassic" in it if you use "ip6g" for the magic number. I think using "ipod classic" as the name for it internally is just adding a risk of having a different ipod classic later if apple changes their mind.
19:31:34*TheSeven thinks that ipod6g might suggest similarity to the ipod 5g/5.5g
19:31:35LloreanI don't really see a disadvantage to ipod6g internally, while iPod Classic doesn't have an immediate disadvantage (other than that people already think of all the iPods as "classic" sometimes already) other than in hypothetical futures.
19:31:49TheSevenalso, this definitely won't go into arm/ipod because it's just completely different
19:31:51LloreanI don't see how "6g" suggests similarity in any way.
19:32:02Lloreanother than it's a generation of ipod.
19:32:05 Join Kitar|st [0] (Kitarist@
19:32:58amee2kwhere does the SDL app have its button map?!
19:33:42TheSevenhaving things in /target/arm/ipod/video for the 5g but /target/arm/s5l870x/ipod6g just seems to be a bit confusing to me
19:34:05pamauryarm/ipod is the old style tree no ?
19:34:22TheSevenarm/ipod are the portalplayer-based ones
19:34:59TheSevenanother issue is having the s5l8701 (nano2g) in the s5l8700 dir, but the s5l8702 separately
19:35:26TheSeventhere are *some* similarities between the 8701/8702, but not that many
19:35:37TheSeventhe 8702/8720 seems to be more similar
19:36:37pamaurythe wiki has a TargetTree entry which states firmware/target/<cpu>/<manufacturer>/<model>
19:36:50TheSevenyeah, but what is the model?
19:37:04TheSevenpeople tend to merge that into whole series of chips if they are similar enough to share most things
19:37:48TheSevenbut the odd relation between the samsung model numbers and their similarities will lead to confusion here
19:39:14TheSevenif we would really follow what the wiki says, it should be firmware/target/arm/samsung/ipod6g or even firmware/target/arm/samsung/s5l8702/ipod6g
19:39:25pamauryperhaps the problem comes from this s5l8700 dir ? Does this chip exist or is it family ?
19:39:56pamauryyeah you're right, I think lately we used something more closer to target/<cpu>/<soc>/
19:39:57TheSevenor is the player manufacturer being meant? then it would be target/arm/apple/ipod6g, but no other target seems to interpret it that way
19:40:22TheSeventhere are four s5l chips we know:
19:40:40TheSeven- s5l8700 (various meizu players, samsung yp-s3)
19:40:47TheSeven- s5l8701 (ipod nano 2g)
19:40:54TheSeven- s5l8702 (ipod nano3g, ipod classic)
19:41:02TheSeven- s5l8720 (ipod nano 4g, ipod touch 2g)
19:41:22TheSeventhe 8700/8701 have many things in common, and the 8702/8720
19:41:42pamaurythat's a weird numbering :-/
19:42:16pamauryor do something like s5l8700v2 :)
19:42:20TheSevento make it even worse: ipod nano 5g is s5l8730, but ipod nano 5g is s5l8723
19:42:40TheSeventhe latter one should be 6g of course
19:43:06pamaurydo 730 and 723 share something in common ? with other series ?
19:43:19TheSevenno idea, we haven't seen unencrypted code for these yet
19:43:48pamauryand nobody has the doc ?
19:44:11TheSeventhe only thing that's in the wild is a leaked "samsung confidential" preliminary datasheet for the s5l8700
19:44:16TheSeveneverything else was reverse-engineered
19:44:35pamauryI guess there are unwilling to share it ? ;)
19:44:48TheSevenand that datasheet doesn't even fully match our reverse-engineering results of the 8700
19:45:01 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
19:45:57pamauryI would personally go for arm/s5l8702 and put s5l8720 with it, they share the "2" ;)
19:46:23pamauryor s5l8702_8720
19:47:15TheSevenactually i'd like to identify all the damn cores they're using, and name them according to their manufacturer
19:47:24TheSevennot caring about socs at all
19:48:11TheSevene.g. synopsys/usb-otg.c, arm/primecell/vic-pl192.c, ...
19:48:17pamaurybut do you have this information ?
19:48:28TheSevenonly for some small parts of it :/
19:49:19pamauryour tree is a total mess :|
19:49:29gevaertsTheSeven: but what about cores that were from company A when SoC X was built but from company B when SoC Y was built?
19:49:58pamaurywhat do you mean ?
19:50:42TheSevengevaerts: i would go for the core's designer, not the soc manufacturer
19:51:03TheSevenbut i'd guess that we won't ever manage to figure this out for lots of cores
19:51:29gevaertsTheSeven: those get taken over every year or so
19:51:46 Quit AlexP (Quit: Please insert girder)
19:52:42gevaertsthe USB core in the PortalPlayer 502x and imx31 SoCs is from ARC, Synposys, and possibly ChipIdea (not sure about that last one)
19:53:56gevaertsWe call it ARC because that name appears somewhere in a datasheet, but...
19:55:26gevaertsAnd in the end I think Synposys has something like seventeen USB cores thanks to their acquisitions
20:03:52 Quit Llorean (Quit: Leaving.)
20:05:49 Quit einhirn (Read error: Connection reset by peer)
20:06:33 Join ReimuHakurei [0] (~reimu@
20:10:47 Join AlexP [0] (~alex@rockbox/staff/AlexP)
20:11:00***Saving seen data "./dancer.seen"
20:11:49 Join Judas_PhD [0] (
20:32:33 Join AlexP_ [0] (~alex@rockbox/staff/AlexP)
20:33:09 Quit AlexP (Ping timeout: 240 seconds)
20:33:27 Quit ReimuHakurei (Quit: Stand by, Ready!)
20:33:33 Join ReimuHakurei [0] (~reimu@
20:54:14pamaurybertrik: I have an explaination for the SGTL in sb file format
20:54:36bertrikSiGmaTeL I mean
20:56:35pamauryin the imx23 manual there is a register called SGTL which is supposed to returned "SigmaTel..." and is used by the debugger to check it's a genuine freescald chip. They used the same in the format. I particularly like the comment:
20:57:07pamauryThe debugger does a string compare on these 12 successive little endian bytes. Any chip that reads back these values is either a Freescale chip or it is a competitors chip that is violating Freescale registered trademarks and or copyrights.
20:58:40bertrikah, interesting
21:02:42pamauryif one day we create a chip, we'll design a nice register which prints rockbox and put a copyright on it ;)
21:15:51casainhopamaury: any advances?
21:16:07casainhopamaury: on imx23 port?
21:17:25pamauryI'm in the last stage of the bootloader, there are lots of interesting things; probably some LCD code so perhaps I'll figure the pins, if bertrik don't already have them already. But there is quite a bunch of code, I don't really know what it does. when I'm done, I'll move on to the firmware itself, to figure out the pins
21:17:40 Quit slooopy (Ping timeout: 240 seconds)
21:18:08pamauryI'm not quite ready to write code :)
21:18:15 Nick noamsml_ is now known as noamsml (
21:19:23casainhopamaury: you mean you have for example, rockbox kernel() working?
21:19:49pamauryno, I didn't write any line of code, nor I have a fuze+ for now ;)
21:20:00pamauryI'm disassembling the fuze+ firmware
21:20:06casainhoah, ok
21:20:32casainhobecause I started the port, I am trying to build bootloader, using empty stubs
21:20:51casainhoI am looking for the help of BobC
21:21:37casainhohe wrote the code for last port, of s3c2440/mini2440
21:21:47casainhowe both have now the same hardware/imx233
21:22:08pamauryI don't know how we will do for the fuze+ port (if there is one); but my guess is that reusing the original firmware bootloader since it's doing lots of hardware initialisation
21:22:12 Join fml [0] (
21:22:26pamaurys/reusing/we will reuse
21:22:58casainhowe have initizlization code from Chumby code :-)
21:23:02casainhodo you know Chumby?
21:23:25fmlHello. FS #11785 contains a rather trivial fix for 14-Nimbus. But we don't have the real name of the contributor. How do we deal with such patches? I was to fix the font I would do it in the very same way.
21:24:02pamauryyes I know Chumby
21:24:30pamaurythen it's the same for you, you won't have to write all the really low level init stuff probably
21:25:25TheSevenfml: why not just fix it yourself then? :)
21:25:43pamaurybut contrary to you, we don't don't know the pins / have no control over them ;)
21:26:04casainhopamaury: I already used the chumby code to initialize the SDRAM and write/read to it :-)
21:26:29casainhopamaury: well, the SDRAM are the same pins for sure :-)
21:26:39pamauryyeah, you don't have much choice :)
21:26:39fmlTheSeven: because he/she was the one who noticed and fixed it (the former was the harder part).
21:26:55casainhopamaury: even the LCD, because of LCD controller
21:27:15casainhopamaury: ok, good luck
21:30:00 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
21:30:10 Join slooopy [0] (
21:32:49TheSevenso should I go for "ip6g" and firmware/target/arm/s5l8702/ipod6g?
21:33:22 Quit ReimuHakurei (Quit: Stand by, Ready!)
21:33:32 Join ReimuHakurei [0] (~reimu@
21:33:32 Quit ReimuHakurei (Client Quit)
21:34:40 Join ReimuHakurei [0] (~reimu@
21:35:12 Quit krabador (Ping timeout: 265 seconds)
21:38:24 Join krabador [0] (
21:40:26 Quit liar (Quit: Leaving)
21:41:46TheSevenhm... ipco, ip3g, ip4g, ipvd, ip6g, mini, mn2g, nano, nn2g, ... very consistent...
21:43:51 Quit krabador (Quit: Sto andando via)
21:45:24 Join liar [0] (
21:45:58 Quit GeekShadow (Quit: The cake is a lie !)
21:46:11 Join Buschel [0] (
21:52:01TheSevenipod 6g feels really silly if the previous model was called "ipodvideo" everywhere
21:53:08TheSevenhm, very good variable name: t_manufacturer="s5l8702"
21:54:10 Join JesusFreak316 [0] (
22:01:08*Buschel needs some testers with nano 2G's for FS #11807
22:05:23CIA-7New commit by alle (r28796): Fix the eogonek glyph in 14-Nimbus (the credit goes to the author of FS #11785)
22:06:49TheSevenBuschel: pasting the flyspray's title might attract more people :)
22:07:09CIA-7r28796 build result: All green
22:07:45TheSevenhow does this compare to DMA?
22:08:19BuschelI thought most people would just use the link :o) but here is the description: "Major speed up of iPod nano 2G LCD"
22:08:50BuschelTheSeven: I don't know how this compares to DMA. did you bring this to wokr?
22:09:05TheSevenmy embios-based benchmarks are using DMA
22:09:22Buschel:o) what do the numbers tell you?
22:09:35TheSeveni posted them some weeks ago
22:09:41TheSevenlet me check the logs
22:09:51Buschelahhh, wait. you had the strange LCD that was limited to 50-60 fps?
22:09:55TheSeventhe reason i am doing this from emBIOS is that the rockbox LCD driver architecture doesn't work well for DMA
22:10:15TheSevenBuschel: I have the "type 7" one, which IIUC is the one you also have
22:10:22TheSeven(LCD_DETECT0: 1, LCD_DETECT1: 0)
22:11:02TheSevenand the limit seemed to come from the system bus actually, not the LCD itself
22:11:03***Saving seen data "./dancer.seen"
22:11:12BuschelTheSeven: we talked a while ago and you had measured there was a limitation in fps which seemed to be connected to HCLK
22:11:26*Buschel is too slow
22:11:48TheSevenwhat are your LCD_DETECT values?
22:12:02TheSeven(diagmode => others => status => second page IIRC)
22:12:41Buschel? how do I enter diagmode?
22:13:25TheSevenprev+select during boot
22:13:38TheSeven[01:11] <TheSeven> 25.8fps full RGB, 103.5fps quarter RGB, 22.5fps full YUV, 89.5fps quarter YUV at 48MHz
22:13:38TheSeven[01:13] <TheSeven> 51.7fps full RGB, 206.5fps quarter RGB, 51.7fps full YUV, 206.5fps quarter YUV at 192MHz
22:13:44the_KyleI made a recording using Rockbox on my Clip+. Mic gain by default is set at 23. I increased it to 29. During an unusually loud portion of the recording, I had to decrease the level back to 25 due to distortion.
22:13:45TheSeventhose are DMA measurements IIRC
22:14:01 Quit krazykit (Quit: awe yeeeeeee)
22:14:11TheSevenhm no, that were rockbox measurements
22:14:15the_KyleBut I can use SoX to increase the volume to four times the original without clipping a single sample.
22:14:50the_KyleWhat did I do wrong to make the volume still so quiet even though I was getting distortion in the recording?
22:14:59TheSeven[01:20] <TheSeven> 51.7fps in emBIOS using DMA
22:15:02Buschelmine is DETECT0: 1 DETECT1: 0
22:15:09TheSeven[01:27] <TheSeven> Buschel: 25.8fps using DMA at 48MHz :/
22:16:17Buschelhow do I exit diagnosis mode? resetting?
22:16:32TheSevenyou can also use the "deep sleep" or "reset" tests :)
22:16:40 Join krazykit [0] (
22:16:42the_KyleI also get a hissing sound in the recording whenever the buffer is written to flash, but I think this is a hardware issue, and there's probably not much I can do about that. It's not particularly noticeable all the time anyway.
22:16:43TheSevenand as we apparently have the same LCD: what values are you getting with and without your patch in rockbox?
22:17:13pamaurywhat is Texas Instrument SSI useful to ?
22:17:49BuschelTheSeven: w/o 192 MHz: 86.5 / 343.5 48 MHz: 36.3 / 144.5
22:17:55BuschelTheSeven: v=
22:18:27BuschelTheSeven: v03 192 MHz: 129.0 / 515.5, 48 MHz: 64.5 / 228.5
22:18:38Buschel50% faster
22:18:57TheSevenwhy are you getting more fps than me on the same LCD type?
22:19:17Buscheldifferent configuration?
22:19:23TheSeveni'm using HCLK=96MHz/48MHz, LCDCON=0xc01
22:19:51Buschelwhat DRAM speeds do you measure? and what is LCD_CON set on your nano?
22:20:00Buscheltoo slow again... :/
22:20:10BuschelLCD_CON is 0xc01 here as well
22:20:24*TheSeven suspects this might be bootloader-related
22:20:27 Quit fml (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
22:20:28BuschelHCLK is also 96 or 48 MHz
22:21:10TheSevenyou are using the apple bootloader, right?
22:22:40TheSevenhm, in this case emBIOS loader might be the culprit
22:24:16BuschelTheSeven: do you see any way to use 16 bit register writes instead of those damned 8 bit writes?
22:24:44TheSevenno idea
22:24:54BuschelTheSeven: I played around with LCD_CON but I don't think the spec is correct
22:25:11TheSevenhowever they are doing 16bit writes on the 8702/8720, so there might be a way to make it work on the 8701, too
22:25:55TheSevenfun fact: on the 8702 they're doing 16bit writes when writing directly, however for DMA they set up 8bit transfers IIUC :)
22:25:56Buschelthe spec says you can set APB bus width and set endianess
22:26:21*TheSeven thinks there was a reason why apple chose 8bit on the 8701
22:27:33Buschele.g. architectural reasons. I don't think they thought about performance...
22:27:48 Quit casainho (Ping timeout: 245 seconds)
22:27:51Buschelbecause it's fast enough anyway :)
22:31:00pixelmathe_Kyle: what codec did you use for recording?
22:31:25pixelmaI mean your encoder settings
22:33:10the_Kylepixelma: I'm using Wavpack. It doesn't have any additional encoder settings. Frequency is 44100 and channels is set to mono.
22:33:42saratogathe AMS chips have a 14 bit ADC, your PC uses 16 bit, and 4x louder is 2 bits
22:33:57saratogatheory: we don't shift the 14 bit output 2 bits to the left when saving it to wav
22:34:31pixelmathat was to saratoga
22:35:05saratogapresumably the input the wavpack encoder is functionally equivalent to wav
22:35:17saratoga'to the wavpack encoder'
22:37:35the_KyleThis seems to explain it quite well. Looks like I need to adjust the volume on the PC to get the desired result and just keep my mic gain somewhere around 23 to 25.
22:39:08the_KyleSince I'm working with Wavpack, this shouldn't reault in any loss of quality, and SoX does this and other things very well.
22:40:49 Quit moos (Ping timeout: 240 seconds)
22:40:53 Join AlexP [0] (~alex@rockbox/staff/AlexP)
22:43:39 Join BigBambi [0] (~alex@rockbox/staff/AlexP)
22:44:35 Quit AlexP (Read error: Operation timed out)
22:45:04 Join moos [0] (moos@rockbox/staff/moos)
22:45:08 Quit AlexP_ (Ping timeout: 264 seconds)
22:47:45 Join wodz [0] (
22:51:23 Quit dickhead ()
22:52:20 Join casainho [0] (
22:54:14casainhoI am getting this warning while trying to port rockbox: warning: implicit declaration of function ‘lcd_set_backdrop’
22:54:28casainhoseen Bagder
22:54:43casainho\seen Bagder
22:55:13pixelmacasainho: msg logbot
22:55:27pixelmait'll tell you that you need glasses though
22:55:30Bagdercasainho: I'm here
22:59:23casainhoBagder: i were trying to know last time yo were active.
23:00:43 Quit JesusFreak316 (Ping timeout: 245 seconds)
23:01:48Bagderthat'd be 1987, in the spring B)
23:01:50casainhoany one cann tell me if I need to build lcd-2bit-vert.c with my bootloader files?
23:02:22casainhoBagder: I didn't saw you talking since a long time :-)
23:02:31casainhoBagder: I am back to do another port...
23:02:39BagderI'm lurking in the shadows
23:02:51Bagderyou need that file if you have a LCD using that format
23:03:28Bagderso what's the target this time?
23:04:15casainhothis one:
23:04:26casainhohey, I know, very crap LCD
23:05:01 Quit hebz0rl (Remote host closed the connection)
23:05:06Bagderlots of CPU for such an LCD...
23:05:24casainhobut I need something I can build, I can source
23:05:43*TheSeven realizes that config.h and SOURCES are just a *damn* *huge* *MESS*!
23:05:50 Join JesusFreak316 [0] (
23:07:16casainho#define HAVE_LCD_BITMAP −−- my LCD is 2 bits, black and white... what should I define?
23:07:52 Part DylanJ
23:11:22 Quit Buschel (Ping timeout: 276 seconds)
23:12:21 Join Buschel [0] (
23:13:50TheSevencasainho: black&white is 1 bit :)
23:13:57saratogaBuschel: if i ever work out this new libmad filterbank, would you be able to use it in mpc, or are the two codecs too different?
23:14:41casainhoTheSeven: ok, so what should I do?
23:15:18TheSeven#define LCD_DEPTH 1
23:15:44Buschelsaratoga: it is possible. the array and some algorithms need to be resorted though
23:16:01casainhodo you know if there is any target like my? with same kinf of LCD?
23:16:07saratogayou mean the D coefficients?
23:17:08Buschelsorry, was a typo: I wanted to tpye "arrays". the subband samples must be resorted to be able to use the combination of dct/dewindowing/D
23:17:40Bagdercasainho: not with the same one, no, but there are targets with 1bit lcd
23:17:53Bagderso the lcd frame buffer routines can be used
23:18:00saratogathats better then the D coefficients though :)
23:18:15saratogaits been a month and i'm still toying with those!
23:18:33BuschelI can imagine
23:19:00casainhoBagder: can you suggest me a target so I can look at it code?
23:19:09Bagderthe recorder?
23:19:52casainhoBagder: firmware/config/export/???.h ???
23:20:24casainhoBagder: ok, found it, Archos :-)
23:22:25saratogai have this terrible feeling that i'm missing some obvious underlying simplification to this filter because sometimes i'll make a simple mistake (flip to pointers) and the output doesn't change
23:22:28BuschelTheSeven: why are all the LCD registers defined as 16bit? all are 32bit
23:22:41Buschelwas this intentional?
23:24:18TheSevenno idea, IIRC bertrik did this
23:24:27bertrikdid I?
23:25:24TheSevenat least it wasn't me who wrote s5l8700.h :)
23:25:52bertriksvn ann firmware/export/s5l8700.h
23:26:10BuschelTheSeven: interesting is that you can use LCD_WDATA and the following registers to write data to the LCD IF. so, you are able to use stmia
23:27:00BuschelTheSeven: bad news is that we still have a 8bit configuration and therefore need 4x the data accesses that we would need
23:27:48bertrikI don't know the reason why they are defined as 16-bit
23:28:16Buschelthen let's correct them
23:28:25*Buschel just compiles
23:28:34TheSevenI know that apple is also using STRH instructions on them, and only actually using 8 bits :)
23:29:42bertrikI thought that, generally, on arm all accesses are 32-bit anyway
23:30:27 Nick BigBambi is now known as AlexP (~alex@rockbox/staff/AlexP)
23:31:35TheSeveni'm pretty sure that 16 vs. 32 bit accesses on the LCD regs don't make any difference at all
23:32:45BuschelTheSeven: correct. but I had the hope using 1x stm instead of 4x str would make a difference
23:32:48TheSevendamn, what should I do about the ATA driver?
23:32:58Buscheland it doesn't... :/
23:34:01Buschelit would again be interesting to find a way to be allowed to write 16 or 32 bits to the LCD FIFO and let the LCF IF do the rest (8 bit transfer to LCD itself)
23:34:16Buschel*that* should make a differencre
23:34:35 Quit slooopy (Ping timeout: 265 seconds)
23:34:51bertrikBuschel, we do something like that for the ams sansas IIRC, so the FIFO converts from 32-bit input to 2x16-bit output
23:34:55TheSevenis there anybody around who knows how this weird semi-generic ATA driver works?
23:35:06TheSeveni.e. how do I implement a lowlevel driver for that?
23:35:28Buschelbertrik: problem is the missing (or incomplete/buggy/preliminary) spec
23:35:46bertrikI think we also found that polling often the FIFO for full-status seemed to slow it down somehow
23:36:30Buschelbertrik: my patch also reduces the polls -> +50% speed
23:36:30CIA-7New commit by wodz (r28797): HD300 - tweak lcd_update() (4-5% speedup)
23:37:03wodzYou have inspired me :-)
23:38:20CIA-7r28797 build result: All green
23:38:26 Join rasher_ [0] (
23:38:26 Quit rasher_ (Changing host)
23:38:26 Join rasher_ [0] (~rasher@rockbox/developer/rasher)
23:38:57*TheSeven throws a rotten banana at whoever designed that ATA driver
23:43:25 Quit rasher (Quit: leaving)
23:43:25 Nick rasher_ is now known as rasher (~rasher@rockbox/developer/rasher)
23:47:08 Join slooopy [0] (
23:47:56saratogaon arm9 theres no speed advantage to STM over STR, although the code size is smaller for the former
23:48:11saratogaarm7 and arm11 are faster with stm and ldm though
23:48:40Buschelat least code readability is better :)
23:48:42Strife892010-12-11 17:48:19 Starting client StrifeBox, revision 35, cores 1
23:48:42Strife892010-12-11 17:48:19 HELLO 35 arm-eabi-gcc444,sdl Strife89:randomshit StrifeBox i686 32 Linux
23:49:48the_KyleIf I have the USB patches applied, is there any way to keep my clip+ from booting into the OF if it's powered off before I plug in the USB cable?
23:50:06TheSeventhe_Kyle: yes, but using a usb-aware bootloader
23:51:03the_KyleDid my bootloader not rebuild and become USB aware when I built the zip file? Or do I need to patch the OF again?
23:55:21pixelmathe latter
23:56:15CIA-7New commit by Buschel (r28798): S5L870x LCD interface registers are 32 bit.
23:57:39saratogathere is no USB aware bootloader for AMSv2 (and technically its not the bootloader that decides but rather the patcher code in mkamsboot)
23:57:57CIA-7r28798 build result: All green
23:58:08 Quit liar (Quit: Leaving)
23:58:52 Quit simonrvn (Remote host closed the connection)

Previous day | Next day