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 2011-01-01

00:04:33 Quit feisar- (Ping timeout: 240 seconds)
00:17:08 Quit preglow (Changing host)
00:17:08 Join preglow [0] (thomj@rockbox/developer/preglow)
00:19:03 Quit froggyman (Quit: Ex-Chat)
00:23:06***Saving seen data "./dancer.seen"
00:24:29 Quit kevku (Quit: KVIrc 4.0.2 Insomnia
00:29:26 Join mystica555_ [0] (
00:43:35 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
00:46:11 Quit BHSPitMonkey (Remote host closed the connection)
00:47:59TheSevenwhat's the maximum chunk size a PCM driver needs to be able to handle?
00:57:43 Quit Battousai_ (Changing host)
00:57:43 Join Battousai_ [0] (~bryan@gentoo/developer/battousai)
00:57:46 Nick Battousai_ is now known as Battousai (~bryan@gentoo/developer/battousai)
00:59:01TheSevencan i safely assume that we'll only ever do 16bit stereo PCM transfers?
00:59:08TheSevensome other drivers seem to rely on that as well
00:59:46TheSevenhm, "16 bit samples, and an even sample count" fits it better
01:00:53 Quit kadoban (Ping timeout: 240 seconds)
01:04:43 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
01:13:19 Join anewuser [0] (anewuser@unaffiliated/anewuser)
01:18:39 Quit robin0800 (Remote host closed the connection)
01:20:50 Join evilnick_ [0] (
01:23:58 Quit evilnick (Ping timeout: 260 seconds)
01:25:08 Nick evilnick_ is now known as evilnick (
01:25:17 Quit evilnick (Changing host)
01:25:17 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
01:37:43 Join Keripo1 [0] (
01:40:56 Quit Keripo (Ping timeout: 276 seconds)
01:43:51 Quit JesusFreak316_ (Ping timeout: 240 seconds)
01:52:13TheSevenseems like there are only two drivers left before we can start testing: LCD and ATA
01:52:53TheSevenlinuxstb: would you like to have an updated patch? or are you sure you won't have time anyway?
01:53:55 Join DerPapst [0] (
02:10:13 Join kadoban [0] (
02:12:41 Quit factor (Remote host closed the connection)
02:13:36 Join factor [0] (~factor@
02:16:18 Quit kadoban (Ping timeout: 240 seconds)
02:23:07***Saving seen data "./dancer.seen"
02:27:48 Join designate72 [0] (
02:57:30 Part toffe82
03:03:35 Quit DerPapst (Quit: Leaving.)
03:06:33 Join feisar- [0] (
03:26:32 Quit GuySoft (Ping timeout: 265 seconds)
03:27:06 Join merbanan [0] (
03:28:42soapBuschel, I'll be up and awake in about 9 hours and should be ready to test then. Sorry for the 24 hour delay.
03:30:31 Join Keripo [0] (
03:32:53 Quit Keripo1 (Ping timeout: 276 seconds)
03:39:07 Join GuySoft [0] (
03:44:19 Join kadoban [0] (
03:45:00 Quit designate72 (Read error: Connection reset by peer)
03:56:18 Quit GuySoft (Ping timeout: 240 seconds)
03:59:42 Quit Keripo (Ping timeout: 255 seconds)
03:59:45 Join Keripo1 [0] (
04:06:08 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
04:09:32 Join GuySoft [0] (
04:19:49 Quit GuySoft (Ping timeout: 240 seconds)
04:20:16 Join Keripo [0] (
04:21:09 Join Barahir [0] (
04:22:31 Quit Keripo1 (Ping timeout: 265 seconds)
04:23:09***Saving seen data "./dancer.seen"
04:24:39 Quit Barahir_ (Ping timeout: 276 seconds)
04:26:36 Join PurlingNayuki [0] (~PurlingNa@
04:35:04 Join Keripo1 [0] (
04:35:22 Quit Keripo (Ping timeout: 240 seconds)
04:44:12 Join bmbl [0] (~bmbl@unaffiliated/bmbl)
04:52:12 Quit amiconn (Disconnected by services)
04:52:14 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:52:19 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:53:19 Quit pixelma (Disconnected by services)
04:53:21 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:53:23 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:54:03 Join GuySoft [0] (
04:55:03 Quit TheSeven (Ping timeout: 260 seconds)
04:58:39 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:59:27 Quit bmbl (Quit: Verlassend)
05:04:34 Quit [Saint] (Ping timeout: 265 seconds)
05:04:38 Join [Saint] [0] (S_a_i_n_t@
05:07:27 Quit GuySoft (Ping timeout: 248 seconds)
05:10:47 Part PurlingNayuki
05:17:47 Quit Rob2222 (Ping timeout: 240 seconds)
05:17:57 Join Keripo [0] (
05:19:23 Quit Keripo1 (Ping timeout: 240 seconds)
05:26:26 Join GuySoft [0] (
05:35:02 Join Keripo1 [0] (
05:36:28 Quit Keripo (Ping timeout: 260 seconds)
05:47:28 Quit GuySoft (Ping timeout: 264 seconds)
05:49:57 Join GuySoft [0] (
05:59:56the_KyleHappy New Year!
06:13:05 Quit parafin (Ping timeout: 276 seconds)
06:14:07 Join parafin [0] (
06:16:38 Quit GuySoft (Ping timeout: 250 seconds)
06:23:12***Saving seen data "./dancer.seen"
06:23:42 Join Keripo [0] (
06:24:31 Quit Keripo1 (Ping timeout: 255 seconds)
06:27:48 Join Keripo1 [0] (
06:29:20 Quit dfkt_ (Quit: -= SysReset 2.53=- Sic gorgiamus allos subjectatos nunc.)
06:30:09 Quit Keripo (Ping timeout: 240 seconds)
06:31:57 Join GuySoft [0] (
06:32:50 Quit parafin (Remote host closed the connection)
06:35:03 Join parafin [0] (
06:38:20 Join dfkt [0] (dfkt@unaffiliated/dfkt)
06:39:15 Quit mortalscan (Ping timeout: 246 seconds)
06:42:31 Join mortalscan [0] (~mortalsca@
06:44:17[Saint]soap: ping?
06:45:09 Quit t0rc (Quit: Give someone code, help them with one project. Teach someone to code, help them rule the world.)
06:45:22[Saint]Is FS #11830 v7 the ATA patch in the iPod Colour build you built for me?
06:45:50[Saint]I'm compiling my own builds, and I'd like to keep testing it for testings sake.
06:49:38 Join Keripo [0] (
06:51:04 Quit Keripo1 (Ping timeout: 276 seconds)
06:53:15 Quit GuySoft (Ping timeout: 246 seconds)
06:53:51 Join GuySoft [0] (
06:55:53 Quit dfkt (Quit: -= SysReset 2.53=- Sic gorgiamus allos subjectatos nunc.)
07:15:54 Join fyrestorm [0] (
07:16:14 Quit balintx (Remote host closed the connection)
07:16:32 Join balintx [0] (
07:20:09 Quit fyrestorm (Client Quit)
07:21:29 Quit mortalscan (Ping timeout: 240 seconds)
07:27:59[Saint]TheSeven: Ping?
07:29:16 Quit GuySoft (Ping timeout: 276 seconds)
07:30:49 Quit balintx (Remote host closed the connection)
07:32:24 Quit parafin (Ping timeout: 276 seconds)
07:34:33 Join parafin [0] (parafin@2001:470:1f0b:81::1)
07:40:50 Join GuySoft [0] (
07:43:35 Join dfkt [0] (dfkt@unaffiliated/dfkt)
07:46:26 Quit CaptainKewl (Ping timeout: 255 seconds)
07:46:37 Quit [Saint] (Read error: Connection reset by peer)
07:47:17 Join [Saint] [0] (S_a_i_n_t@
07:49:46 Join Keripo1 [0] (
07:51:49 Quit Keripo (Ping timeout: 260 seconds)
07:53:48 Join Horscht [0] (
07:53:48 Quit Horscht (Changing host)
07:53:48 Join Horscht [0] (~Horschti@xbmc/user/horscht)
07:54:47 Quit dfkt (Read error: Connection reset by peer)
07:56:29 Quit Horschti (Ping timeout: 260 seconds)
08:01:52 Join Keripo [0] (
08:03:23 Quit Keripo1 (Ping timeout: 240 seconds)
08:10:51 Join Keripo1 [0] (
08:12:46 Quit Keripo (Ping timeout: 250 seconds)
08:18:50 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
08:23:13***Saving seen data "./dancer.seen"
08:41:15 Join Keripo [0] (
08:44:01 Quit Keripo1 (Ping timeout: 276 seconds)
08:47:42 Quit GuySoft (Ping timeout: 246 seconds)
09:01:20 Quit shai (Read error: Connection reset by peer)
09:03:30 Quit Keripo (Ping timeout: 240 seconds)
09:04:55 Join Keripo [0] (
09:08:24 Join GuySoft [0] (
09:16:03 Join Keripo1 [0] (
09:17:20 Quit Keripo (Ping timeout: 255 seconds)
09:31:45 Quit kadoban (Read error: Operation timed out)
10:23:17***Saving seen data "./dancer.seen"
10:35:29 Quit BHSPitMonkey (Remote host closed the connection)
10:40:03 Quit evilnick (Ping timeout: 255 seconds)
10:57:54 Join Buschel [0] (
10:59:12 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
11:03:03 Join stoffel [0] (
11:05:22Buschel[Saint]: you had the chance to test FS11843 v11 on one of your ipod color's ? or are the devices still in the locker ;)
11:06:54TheSeven[Saint]: gnip
11:17:30 Quit Buschel (Ping timeout: 264 seconds)
11:21:11[Saint]TheSeven: I noticed a change to enable USB interrupts snuck into the piezo patch, was that intentional?
11:21:35TheSevenprobably not
11:21:41[Saint]I'm rebuilding my tree, and I couldn't figure out why I was getting HID enabled Nano2G builds. ;)
11:23:23 Join DerPapst [0] (
11:28:31 Join robin0800 [0] (
11:45:09 Join Keripo [0] (
11:45:48 Quit [Saint] (Quit: I'm only going to Heaven if it feels like Hell, I'm only going to Heaven if it tastes like caramel...)
11:47:24 Quit Keripo1 (Ping timeout: 240 seconds)
11:48:00*TheSeven decides to do some ATA driver rework
11:55:51 Join kevku [0] (~kevku@2001:7d0:0:f000::135d)
11:56:07 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
11:56:09 Quit robin0800 (Read error: Connection reset by peer)
11:59:41 Quit bluebroth3r (Ping timeout: 276 seconds)
12:02:25 Join n1s [0] (~n1s@rockbox/developer/n1s)
12:14:07CIA-7New commit by mt (r28937): Fix FS #11845 by rejecting unknown header signatures in the audio stream info block. RALF uses a different header signature than ('.', 'r', 'a', 0xfd). ...
12:14:58*TheSeven throws into the room
12:15:14TheSevendo we want such a change in general?
12:15:25TheSevendid i get all the endianness logic right?
12:15:47TheSevenwho wants to test it?
12:16:28CIA-7r28937 build result: All green
12:21:14 Quit DerPapst (Read error: Connection reset by peer)
12:21:42 Join DerPapst [0] (
12:23:23***Saving seen data "./dancer.seen"
12:32:33*TheSeven will probably just commit it if nobody bothers to even reply :)
12:48:11AlexPIt is new year's day, I suspect there are a few sore heads around (including mine)
12:50:59 Join Buschel [0] (
13:02:08 Join MethoS- [0] (~clemens@
13:24:38 Join thomasjfox [0] (
13:25:04thomasjfoxhappy new year to everyone
13:36:26pixelmasame to you and all the others :)
13:42:40Dreamxtremeanyone get a change to look at the bug i said about ?
13:42:48*TheSeven wonders if Torne or some other ATA guru is awake :)
13:45:39 Join Topy [0] (
13:49:24 Quit T44 (Ping timeout: 240 seconds)
14:10:36thomasjfoxkugel: As you usually read the IRC logs, happy new year to you and I've fixed the database duplication issue for the application.
14:10:45Buschelsoap: does FS #11843 v11 work for you?
14:11:14soapI'm about to test
14:12:50soap94.0 and 376.0 were the numbers for v10 with FORCE FIFO WAIT commented out IIRC.
14:13:33 Join user890104 [0] (
14:16:19soap83.5 / 335.0 at 80Mhz, Buschel
14:16:44soapand I caught a brief glimpse of screen corruption of a splash when launching the plugin.
14:17:15soapyes, corruption of all in test_viewports
14:17:57Buschelhmm, is this also happening with v09?
14:18:14soapwas about to build 10, let be roll back again and check
14:21:43soaptest-viewports looks fine, splashes look fine with v09
14:22:38Buschelok, will take a look at this after lunch
14:23:27***Saving seen data "./dancer.seen"
14:33:10 Quit JdGordon| (Ping timeout: 255 seconds)
14:38:04Buschelsoap: this takes back the change I think is the cause ->
14:38:12Buschelcan you verify?
14:46:09soap83.5 / 335.0 fps boosted (YUV) with this one (v11), hmm, a minor issue with test_viewports. Might be something i missed previously with stock
14:47:12Buschelok, now just change the following manually in target\arm\ipod\lcd-color_nano.c
14:47:23Buschelline 316, change to "#if 1"
14:47:52Buschelline 320, change to "addr += (LCD_WIDTH)/2;
14:47:59soapin a sec, rebuilding v09 to verify the test_viewports odd hang isn't somehow related to v11
14:48:52soapjust those two changes?
14:50:24 Join dantje [0] (
14:51:22 Quit feisar- (Ping timeout: 246 seconds)
14:55:03soapok the slight viewport glitch verified as having existed with v09, so ignore that as it isn't new. V12 = 83.5 and 335.0 again.
14:55:58Buscheldid you already compile and test with local manual changes?
14:56:17soapthat's what I'm calling v12
14:56:27soapso yes.
14:57:08Buschelok, then this issue was dumb error :)
14:57:24Buschel(which is good as this can be easily fixed)
15:01:14Buschelso, this should work properly now ->
15:08:17soapv13 = 83.5 / 335.0 fps again, no text corruption.
15:16:30Buschelgood, updates FS #11843. now I only need a verification for iPod color (by [Saint]?)
15:18:06 Join feisar- [0] (
15:22:31 Quit feisar- (Ping timeout: 240 seconds)
15:22:33 Quit DerPapst (Quit: Leaving.)
15:23:04 Quit parafin (Quit: So long and thanks for all the fish)
15:23:14 Join parafin [0] (
15:33:30 Join domonoky [0] (
15:33:34 Quit domonoky (Changing host)
15:33:34 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
15:40:16 Quit kevku (Ping timeout: 260 seconds)
15:43:07 Join CaptainKewl [0] (
15:56:25 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
16:04:41Buschelsoap: another refactoring (simplify loops for YUV updates) ->
16:04:52Buschelsoap: do you still have got some time for this?
16:05:20soapjust a few more minutes, I need to go to the gym
16:05:47Buschelthen let's hope there is no issue ;)
16:09:23soapv14 result: 84.0, 335.5
16:10:08soapI forget which one now was "too fast for [Saint]'s ipod screen", but it was still significantly faster.
16:10:08Buschelmovies play fine?
16:10:49soapI believe so. We need a new test file.
16:10:52Buschelsoap: v10 was too fast
16:11:12Buschelyou have the sample videos (elephants dream)?
16:11:14soapthe current encodes of elephant's dream have too many artifacts.
16:11:33soapfine for legacy testing of fps, but worthless for looking for flaws.
16:11:45 Join kadoban [0] (
16:12:05Buschelwell, if the code is not working properly the artifacts would be obvious and distracting
16:12:30Buschellike the screen shots you've sent me in the past days
16:12:53soapI hear you and know what you mean, but I consider the current state "obvious and distracting"! ;)
16:18:13Buschelyou're talking of elephants dream in general and not of the last patch?
16:18:25 Join CaptainKwel [0] (
16:20:31 Quit CaptainKewl (Ping timeout: 260 seconds)
16:22:33 Quit CaptainKwel (Ping timeout: 240 seconds)
16:23:31***Saving seen data "./dancer.seen"
16:27:18 Join xavieran [0] (
16:27:25 Join mortalscan [0] (~mortalsca@
16:29:34soapyes, sorry, did not intent to confuse. I find the Nano screen irritating to watch because of its poor off-axis response and I find (at least the) Rockbox Nano encodes of Elephant's Dream very distracting to watch carefully because of the significant artifacting.
16:53:13 Join evilnick [0] (
16:59:13 Join robin0800 [0] (
16:59:25 Quit mortalscan (Ping timeout: 240 seconds)
17:00:58 Quit robin0800 (Remote host closed the connection)
17:01:27 Join robin0800 [0] (
17:19:27 Quit dantje (Quit: Ex-Chat)
17:19:49BuschelSaint: (for the logs) can you please check what lcd_type your iPod color has? this can be seen in the debug menu under "View HW info"
17:21:34 Quit factor (Ping timeout: 255 seconds)
17:24:00 Join mortalscan [0] (~mortalsca@
17:25:24 Join factor [0] (~factor@
17:26:26 Quit Buschel (Ping timeout: 260 seconds)
17:29:40 Quit robin0800 (Remote host closed the connection)
17:30:05 Join feisar- [0] (
17:34:11 Quit feisar- (Ping timeout: 240 seconds)
17:35:10 Join DerPapst [0] (
17:41:02 Quit GuySoft (Ping timeout: 265 seconds)
17:45:40 Quit t0rc (Remote host closed the connection)
17:53:27 Join GuySoft [0] (~guysoft@
18:02:14 Quit DerPapst (Quit: Leaving.)
18:02:38*TheSeven starts playing around with an iPod Classic Rockbox build! \o/
18:03:04TheSevenit's still locking up before I see anything on the LCD
18:03:15TheSevenbut I'm pretty sure I'll figure out the cause soon
18:03:25 Quit krazykit (Ping timeout: 255 seconds)
18:04:33 Quit evilnick (Quit: Leaving)
18:05:44TheSevenhm, it reaches main()
18:05:49 Join Buschel [0] (
18:10:37 Join krazykit [0] (
18:23:34***Saving seen data "./dancer.seen"
18:24:36 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
18:29:39*TheSeven swears at Buschel's code
18:44:22 Quit Keripo (Quit: Leaving.)
19:06:08BuschelTheSeven: what is so bad with it?
19:06:54TheSevenif i knew that
19:07:12TheSevensomething in your LCD asm locks up on the ipod classic, and i have absolutely no idea why
19:07:36TheSeveni bet the root cause isn't in your code though :)
19:07:46Buschelwhich function in detail? and copied from which driver?
19:07:54Dreamxtremebest update the status on the home page then :D
19:08:02TheSevenlcd_write_line from nano2g
19:08:37Buschelthe C code is comparable for both targets?
19:10:05TheSeventhe LCD controller core seems to be identical, just at a different AHB address
19:10:11Buschelsomething wrong with the alignment?
19:10:16 Quit Slasheri (Ping timeout: 260 seconds)
19:10:17TheSevenno idea
19:10:38Buschelis there working c code available?
19:10:40TheSevenwith caches enabled, it seems to lock up at "ldr lr, [r2, #0x1C]", with caches disabled it doesn't lock up but doesn't influence the lcd contents either
19:10:57TheSevenyes, but that C code works in a totally different way
19:11:40TheSeven(that's using DMA)
19:12:45 Join Slasheri [0] (miipekk@rockbox/developer/Slasheri)
19:14:29Buschelok, totally different approach...
19:14:43TheSevenyep, i just offload it all to the hardware
19:15:45Buschelwhy not keep using DMA for the classic port?
19:16:07TheSevenbecause we would need to block for DMA completion anyway
19:16:34TheSevenrockbox can't handle double buffering right now, neither for audio nor video
19:16:43TheSevenfor audio i used some hack, but for video it isn't that easy
19:17:09Buschelyou're right
19:18:57Buschelare you sure the lock up is in "ldr lr, [r2, ...]" and not in the command before?
19:22:36TheSeveni branched out of the code immediately before and after that instruction, and if the branch was before it, it worked (and resetted the SoC), if it was after the instruction it didn't seem to be reached
19:28:23TheSevenwell, as the caches had an impact, I'd expect something with my MMU setup to be wrong
19:28:36TheSevenbut even with a disabled MMU it still doesn't show anything on the LCD
19:47:20 Quit mortalscan (Quit: Leaving)
19:52:19 Join Keripo [0] (
19:59:16 Join bertrik [0] (
20:00:47 Quit xavieran (Ping timeout: 255 seconds)
20:05:01 Quit stoffel (Ping timeout: 240 seconds)
20:10:39 Join leavittx [0] (~lev@MS-10-112.dyn-ip.SPb.SkyLink.RU)
20:11:23 Quit liar (Quit: Leaving)
20:14:25 Join xavieran [0] (
20:23:37***Saving seen data "./dancer.seen"
20:25:11CIA-7New commit by Buschel (r28938): Save some binsize in LCD driver of iPod nano 2G. No impact to speed.
20:28:30 Quit factor (Quit: Leaving)
20:28:51 Join factor [0] (~factor@
20:28:54 Quit factor (Read error: Connection reset by peer)
20:31:16 Quit leavittx (Ping timeout: 276 seconds)
20:35:29 Quit bug2000 (Read error: Operation timed out)
20:37:01 Join bug2000 [0] (
20:37:01 Quit bug2000 (Changing host)
20:37:01 Join bug2000 [0] (~bug@unaffiliated/bug2000)
20:39:32 Join leavittx [0] (~lev@MS-10-112.dyn-ip.SPb.SkyLink.RU)
20:54:02CIA-7New commit by Buschel (r28939): Higher precision for test_mem plugin.
20:56:01CIA-7r28939 build result: All green
20:58:04 Quit leavittx (Ping timeout: 276 seconds)
21:00:09 Join leavittx [0] (~lev@MS-10-112.dyn-ip.SPb.SkyLink.RU)
21:02:20*Buschel wonders if this a good or bad sign
21:02:20TheSevenrockbox logo on the screen, followed by "ATA error: -22\n\nPress ON to debug"
21:02:29Buschelgood sign :)
21:03:09Buschelwhat was the trick?
21:03:12TheSevenhowever only if i replace your ASM code with something more stupid
21:03:30TheSeven for (y = y0; y <= y1; y++)
21:03:30TheSeven for (x = x0; x <= x1; x++)
21:03:30TheSeven s5l_lcd_write_data(lcd_framebuffer[y][x]);
21:03:33TheSeventhis does the trick
21:03:51Buschelthe old code
21:04:07Buschelisn't it?
21:04:12TheSevenbasically yes
21:04:21TheSeveni can't see any obvious difference though
21:04:57Buschelwell, the alignment is critical. x must be even and width must be a multiple of 4 to use my asm code
21:05:08TheSevenbut i'll probably go after the ATA error first, no that i have a working lcd driver to print debug info
21:05:29TheSevenBuschel: if that doesn't hold true, it should at least do *something*
21:05:36TheSeventhe LCD didn't show any reaction at all
21:05:44Buschelbut this should not lead to a crash like you described...
21:08:55Buschelhmm, and s5l_lcd_write_data polls for (STATUS & 0x10) each pixel and not for (STATUS & 0x08) each 4 pixels.
21:09:25*Buschel spots an error in a comment
21:10:20CIA-7New commit by Buschel (r28940): Fix comment.
21:11:32Buschelhowever, you're able to proceed with other stuff now
21:12:15CIA-7r28940 build result: All green
21:14:16TheSevenBuschel: but even if that fifo check doesn't work right, i would at least expect some garbage on the screen
21:14:33TheSeventhe CPU is running with caches disabled, so at a ridiculously slow speed
21:14:41TheSevenyou could remove the fifo checks altogether and it wouldn't matter
21:15:03Buschelif the FIFO registers work the same...
21:19:10Buschelwhat does the S5L8702 do, if you access a 2-byte aligned address with a 4-byte read?
21:27:42 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
21:32:07 Join saratoga [0] (9803c6dd@gateway/web/freenode/ip.
21:33:31 Join telliott [0] (
21:33:37saratogaBuschel: IIRC unaligned loads with ldr generate an error, while unaligned loads with ldm just return the lower aligned word on armv5
21:34:55Buschelok, would also not explain TheSeven's experiences (as your 2nd case is relevant)...
21:35:44TheSevendepends on some cp15 setting IIRC
21:36:29Torneif you have alignment faults on it faults, otherwise it returns the wrong data rotated
21:37:10Torneyou get the wrong data whether you se ldr or ldm
21:37:31TheSevenTorne: what do you think about the ATA changes btw?
21:38:25Torneseems sensible..
21:38:57TheSeveni might have broken it in some places wrt. endianness
21:39:42TheSevenespecially line 365 of the paste
21:40:23Torneer, i can't really have a thorough look atm, just passing throgh :)
21:44:24thomasjfoxcould any of the devs have a quick look at a patch I'd like to upstream?
21:45:30thomasjfoxIt's really small
21:50:11gevaertsthomasjfox: I'm not sure if I like the general idea
21:50:52thomasjfoxgevaerts: What do you prefer?
21:51:11thomasjfoxgevaerts: Right now we have non-working menus in the app build
21:51:16gevaertsI know
21:51:35thomasjfoxOr non-working picture flow menu entries
21:51:42gevaertsBut I think that technically a rockbox build with non-working credits is a license violation
21:52:24gevaertsAs long as this is a temporary issue, I don't think people will have a problem with that, but this patch seems to make that a permanent "feature
21:52:50thomasjfoxso maybe credits should be a core function instead of a plugin?
21:52:56thomasjfoxthere are other targets without plugins, too
21:53:44saratogai thought we wanted at least some plugins on app builds (e.g. fft, etc)
21:54:08gevaertsWell, as long as it's "plugins don't work yet", I think that's fine. And not having games is also fine
21:54:23gevaertsBut "no plugins" as an end goal isn't
21:54:39thomasjfox"No plugins" is not the desired end goal, that's for sure
21:54:49thomasjfoxI already got them working somewhat, it's still very broken
21:55:09TheSevenhm. our arm-elf-eabi-gcc isn't exactly clever
21:55:31gevaertsSo the question is "Do we want in-progress ports to be able to look complete?"?
21:55:37soapBuschel, latest YUV = 84.0 / 335.5 / no corruption
21:55:42gevaertsMy personal answer would be "Why?
21:56:07thomasjfoxThe more it looks working, the better?
21:56:16gevaertsIs it?
21:56:28thomasjfoxI don't want to show features to the user which just crash. Users will click on it
21:56:30gevaertsIf it *looks* finished, why would anyone try to fix the remaining bits?
21:56:53gevaertsDoes it crash?
21:56:53Buschelsoap: v16 ?
21:57:08thomasjfoxThe picture flow menu outputs a very ugly error message
21:57:15thomasjfoxSame for the credits screen
21:57:41gevaertsWhat does it say? A "plugin error" splash?
21:57:56thomasjfoxLet me just re-enable plugins ;)
21:59:32soapBuschel, yes, v16
21:59:42thomasjfox"failed to load /opt/rockbox/lib/rockbox/rocks/demos/pictureflow.rock"
22:00:30 Quit xavieran (Ping timeout: 255 seconds)
22:00:48Buschelsoap: thanks! I now _really_ need Saint for further testing on iPod Color/Photo... I would like to submit this stuff :)
22:01:53gevaertsI can see your point for the pictureflow thing, but I'd say a "DISABLE_PICTUREFLOW_INTEGRATION" define in config/application.h (possibly #ifdeffed) is cleaner
22:02:19gevaertsI don't think having the Credits menu item go away is a good idea though
22:02:29gevaerts(even if it gives an error)
22:02:59thomasjfoxOk. Almost the same idea: Keep the patch like it is except for the credits part?
22:04:51thomasjfoxIdea for the future: Provide a minimal non-plugin credits screen
22:05:32gevaertsWhy? What's wrong with having it as a plugin?
22:05:57thomasjfoxSo we have a credits screen if for whatever reason the plugins are disabled?
22:06:10thomasjfoxJust as an alternative
22:07:06gevaertsDisabled plugins should really be a temporary state while developing, just like e.g. having no LCD driver yet
22:07:49gevaerts(note: "enabled plugins" is not the same as "we build all games")
22:08:38thomasjfoxThat is true, see this one (which is hopefully temporary):
22:09:16gevaertsYou're allowed to reorder entries if that makes things more readable :)
22:10:43 Quit thomasjfox (Remote host closed the connection)
22:11:05 Join Ogham [0] (~Ogham@unaffiliated/ogham)
22:11:19OghamHi all, I don't suppose anyone here owns a Sansa Clip+ ?
22:11:53AlexPjust ask any questions
22:13:11OghamJust a random query, my Clip+ makes a loud creaking noise when I hold/press the right hand corner lightly.. I was just wondering if other people have noticed this on theirs, or if mine may have been damaged in transit?
22:13:33gevaertsAlexP: did you see ?
22:13:35 Join xavieran [0] (
22:14:33AlexPgevaerts: no
22:14:57AlexPgevaerts: But I'd say if it works, get it in, do a 3.7.2 and then that is probably the last before 3.8
22:15:04AlexPunless anything massive comes up
22:15:20gevaertsbertrik: ping
22:15:20LloreanOgham: I don't know what you mean by "right hand corner" since there's two corners on the right side, but neither of them make any loud noises when I squeeze lightly.
22:15:36 Join thomasjfox [0] (
22:16:16 Join Tim_Elliott [0] (
22:16:31OghamLlorean: Oops, yes I meant to add 'bottom' to that, thanks for checking.. I'm not sure if I should return it or not, it works without any issue, rockbox is great on it :)
22:17:26OghamI'm a little concerned about having my IE-8's plugged in if the unit has taken a beating!
22:18:46gevaertsOgham: just to clarify, does the case (i.e. the plastic) make this noise, or is it on the headphones?
22:19:33 Quit telliott (Ping timeout: 250 seconds)
22:19:36gevaertsIf it's in anything electrical, and you can return it, I think that's the best option. This is the sort of thing that may easily get worse over time
22:19:59Oghamgevaerts: Its the casing of the Clip+, just holding it there in normal use causes a nasty creaking sound and there is some movement..
22:21:24Buschelanother candidate for a 3.7.2 is FS #11830. it can be submitted with some proper nano1g-ifdefing to not touch other targets.
22:21:32Oghamother than that there does not seem to be anything wrong, but if Llorean does not notice the same thing.. I think I'll see if I can return it :(
22:22:19gevaertsBuschel: that's not in trunk yet I think?
22:22:40LloreanOgham: There's not really any movement in mine either.
22:22:43Buschelnot yet. might submit this next week, if there are no side effects reported
22:23:41***Saving seen data "./dancer.seen"
22:23:45OghamLlorean: So the whole thing feels pretty solid to you?
22:24:09LloreanOgham: Yeah, no real give anywhere.
22:24:34OghamLlorean: ah well, better see if I can get this one exchanged :( Thanks for checking for me!
22:24:35LloreanI mean the plastic flexes a little if I press down on the word "sansa" but that's not what you're talking about, I'm sure
22:24:43Buschelgevaerts: when would 3.8 most likely be branched off?
22:24:48saratogaanyone notice that none of the gigabeat players can be "ejected" in windows, but the sansa players can (when using rockbox USB)?
22:24:49saratogawhy is that
22:25:19AlexPBuschel: See the Rockbox calendar!
22:25:37 Join wah_wah_69 [0] (~io@
22:26:10gevaertssaratoga: the last few lines of firmware/usbstack/usb_storage.c
22:26:48BuschelAlexP: cool, didn't knwo about this
22:26:58gevaertsPersonally I think any drive that's not hotswap in rockbox itself shouldn't be marked as removable, but I seem to be the only one in the world to think that way
22:26:58AlexPBuschel, gevaerts: And yeah, I think FS #11830 should be OK, but it should go in trunk for a while first to check it all out :)
22:27:08OghamLlorean: nope, just the bottom right corner, I'm guessing a clip inside is broken..
22:27:10AlexPBuschel: I sent two emails to -dev about it! :P
22:27:15wah_wah_69Hi everyone! I've just found that my player (packard bell vibe 300) uses
22:27:26OghamThanks again, have a good new year all :)
22:27:26wah_wah_69crap this irc client
22:27:32saratogagevaerts: well its nice to be able to unmount the disk easily in windows
22:27:40Lloreangevaerts: That makes sense to me. If you can't re-insert it without replugging the cable, don't mark it as "removable" basically?
22:27:46 Part Ogham
22:28:08Lloreansaratoga: I can still safely remove hardware on my gigabeat, I just don't have an "eject" option in the right click menu.
22:28:20wah_wah_69I mean my packard bell player seems to use portal player 5020, or at least has mi4 file with the name pp5020.mi4 on the system directory
22:28:20gevaertsWell, that "removable" bit really is about the media, not the SCSI device
22:28:42saratogaLlorean: yes thats what i'm refering to
22:28:49wah_wah_69I've open it as text
22:28:56gevaertsSCSI hotswap (which USB basically provides) is really distinct for removable media drives
22:30:11wah_wah_69It seems to use PP5020AF, the thing is I'd like to post about this in the forum, and I'm not sure where to post
22:30:27saratogaalso, does the GBS charge off USB?
22:30:39Tornebut not if it's flat
22:30:46LloreanIt has to be "on" to charge.
22:30:46Torneit has to boot up successfully first
22:30:53LloreanUnlike the F. =/
22:31:20gevaertswah_wah_69: a new thread in New Ports I'd say (unless there is an existing one, but I don't think so)
22:32:20wah_wah_69gevaerts, I must be blind or something, I didn't notice that subforum
22:33:09wah_wah_69I'm tempted to try the mi4 from iriver h10, there's no risk as long as I have a backup of the original mi4, right?
22:33:38AlexPUnless it is identical hardware it won't work
22:35:02wah_wah_69it uses the same chip from portal player, 5020, maybe a different revision hence the AF ending in the text from the firmware
22:35:10AlexPand all the other hardware?
22:35:23AlexPAnd is everything connected up the same?
22:35:52AlexPAnd ...
22:36:10wah_wah_69I'm not sure, haven't even tried to open it.
22:36:44Torneit will only work if the player is literally identical apart from the plastics
22:37:04Tornebeing similar makes it less work to do a port, but you still need to do a port
22:37:05gevaertsthomasjfox: that (well, I'm looking at 229da89f31cbe13bf5b32e9ce559e9c27563105f..3fae73d38325395cb4a3512d37f0ffd72c676d71 now) looks reasonable. I'm still not sure about it (the entire HAVE_PLUGINS thing) though, but maybe I'm alone there...
22:37:05wah_wah_69I'll make a post in the forum with as many details as I can about this player (vibe 300), i'll try to open it later
22:37:09saratogawah_wah_69: do you have any experience with programming hardware devices?
22:37:38wah_wah_69I'm finishing my degree in computer engineering
22:37:50saratogathen this should be a fun project
22:38:02saratogabut you've picked a really obscure device so you're probably on your own
22:38:44gevaertsIf you're lucky, it might be similar to the vibe 500?
22:39:49wah_wah_69lets suppouse the vibe 300 bootloader is similar to the vibe 500, so its already known, and the portal player chip is well understood, apart from that what's left to figure out? all the remaning glue logic?
22:39:52thomasjfoxgevaerts: Don't keep your mind occupied with HAVE_PLUGINS. In the end that patch is going to die, atleast for maemo
22:40:20gevaertsthomasjfox: I hope so :)
22:40:20thomasjfoxgevaerts: I'm just doing another binary .deb build with plugins enabled. If all goes well (doesn't crash/hang), I'll push it out soon.
22:40:40thomasjfoxPeople are really happy with rockbox on the n900
22:41:05thomasjfoxStuff like "Where should I send my firstborn?" is mentioned ;)
22:42:56thomasjfoxgevaerts: Maybe I asked this before, do you own a n8xx device?
22:43:01gevaertsI don't
22:43:11thomasjfoxI'm wonderin how high the CPU usage might be
22:43:17gevaertsI know petur has one though, and IIRC soap as well
22:51:20soappeople on are happy whenever the benevolent gods drop crumbs in their direction. ;)
22:52:24gevaertsWell, only if you keep dropping crumbs
22:52:42gevaertsAs soon as you don't drop a crumb one week, you'll know what they *really* think :)
22:53:45thomasjfoxsoap: Would you have time to play around with the maemo port on a n8xx?
22:53:55thomasjfoxI'll have a new build up in about 20min
22:54:05soapsure, If I can recall how to install apps on it!
22:54:10thomasjfoxdpkg -i ;)
22:54:28soapI need to run an errand, I'll be back in about 90 minutes, I'll highlight you then.
22:58:13 Join kugel [0] (~kugel@rockbox/developer/kugel)
23:14:40thomasjfoxkugel: Seen my message in the irc log? ;)
23:15:41 Quit thomasjfox (Remote host closed the connection)
23:16:13 Join thomasjfox [0] (
23:17:36thomasjfoxkugel: The app_* and sim_* wrapper layer leave room for improvement. The plugins have again trouble saving their config as they want to write to the "rocks" dir...
23:17:55thomasjfoxkugel: Maybe we can unify the file functions?
23:21:54kugelunify how?
23:23:09thomasjfoxmaybe have just one open function which does the /.rockbox/ path expansion?
23:23:24kugelare there more than one?
23:23:43thomasjfoxapp_open calls sim_open which also does some kind of path transformation IIRC
23:23:56kugelonly on the sim though
23:24:40kugeland android doesn't call sim_open()
23:24:45 Join tuxifier [0] (
23:24:51thomasjfoxkugel: Have a look at firmware/common/rbpaths.c
23:25:18thomasjfoxFor PLATFORM_SDL it also calls sim_open
23:25:34kugeli mean sim_open() only transforms paths on the sim
23:26:50thomasjfoxso get_sim_pathname() is not used if APPLICATION is defined
23:29:14thomasjfoxhmm. Anyhow, I manage to get the plugins running and they also need path translations as I could determine via strace
23:30:04 Join mortalscan [0] (
23:30:30kugeli know
23:31:13kugelwe thought some PLUGIN_DATA_DIR where plugins need write access could work
23:32:30thomasjfoxso that would map to $(HOME)/.config/
23:34:12thomasjfoxsoap: the n8xx build can be found here:
23:35:03CIA-7New commit by jethead71 (r28941): Gigeabeat S: Reset DMA size info when stopping audio playback and recording transfers so that size remaing/peak buffer calls return 0/NULL when ...
23:36:58CIA-7r28941 build result: All green
23:41:00kugelthomasjfox: it's wrong to call it 3.7.2 imo
23:41:15thomasjfoxkugel: I know. 3.7.1 wasn't right either
23:41:34thomasjfoxkugel: Dunno how I can set svn version numbers with debian packaging
23:41:57thomasjfoxSo many things to fix, so little time :)
23:42:36kugelcan you run
23:42:57thomasjfoxI need to enter the version number manually into debian/changelog
23:43:19thomasjfoxSo I would need another script that automates debian packaging
23:43:33 Quit pamaury (Remote host closed the connection)
23:44:32gevaerts3.7.2 and 3.7.1 are equally wrong I think :)
23:45:06 Join leavittx_ [0] (~lev@MS-55-111.dyn-ip.SPb.SkyLink.RU)
23:45:32thomasjfoxI'll see what I can do about for the next build
23:46:44thomasjfoxbtw: Would be nice to run "make veryclean" without a full build setup
23:47:12thomasjfoxTook me a while to find out there are tools which don't get removed by "make clean"
23:47:15Tornethere are debhelper scripts to do svn/cvs/etc version numbering, i'm pretty sure
23:47:57 Quit leavittx (Ping timeout: 246 seconds)
23:48:59thomasjfoxThough the svn # is wrong also as the maemo port contains code that is not yet in subversion :o)
23:49:11AlexPgive it your own random one for now
23:49:40AlexPImplying it is either an existing release or a release that hasn't yet happened that it won't be in is not quite right :)
23:49:41gevaertsA date-based one may be best
23:49:54AlexPyes, good idea
23:50:07thomasjfoxSimple as that
23:51:46 Quit mystica555_ (Ping timeout: 265 seconds)
23:53:56 Join feisar- [0] (
23:58:47tuxifierI don't get it - got 2 identical sansa clip+ with rockbox r28932-101229 and two identical 16GB microSD cards

Previous day | Next day