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 2010-03-07

00:02:43 Quit Highlander (Quit: Quitte)
00:02:58liaritunes wants to restore the ipod if a custom bootloader is installed
00:03:42TheSeventhat's another reason for a nor-based installation then
00:03:58TheSeventhat way we can leave the firmware partition completely untouched)
00:09:40 Quit anewuser ()
00:14:12 Quit froggyman (Ping timeout: 245 seconds)
00:15:02 Quit Adnyxo (Read error: Connection timed out)
00:15:43 Join Adnyxo [0] (
00:16:49 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
00:20:04 Quit m3dlg (Ping timeout: 240 seconds)
00:20:47 Quit bluebroth3r (Ping timeout: 276 seconds)
00:20:53 Join m3dlg [0] (
00:26:52 Join arbingordon [0] (
00:27:02 Join planetbeing_ [0] (
00:27:09 Quit planetbeing (Read error: Connection reset by peer)
00:27:09 Nick planetbeing__ is now known as planetbeing (
00:29:20 Quit Buschel ()
00:30:13 Quit ender` (Quit: I contend that we are both atheists. I just believe in one fewer god than you do. When you understand why you dismiss all the other possible gods, you will understand why I dismiss yours. -- Stephen Roberts)
00:37:05amiconnTheSeven: Did you also consider (or even test) flashing rockbox in addition to the loader?
00:37:05 Quit planetbeing_ (Read error: Connection reset by peer)
00:37:09 Join planetbeing_ [0] (~planetbei@
00:37:17 Join planetbeing__ [0] (
00:38:23TheSevenyou mean booting directly from bootrom to rockbox?
00:38:51TheSeven(with only the recovery stub in between)
00:38:52 Join planetbeing___ [0] (
00:39:07TheSeventhat *could* work, but i haven't tested it yet
00:39:26 Quit perfectdrug (Quit: perfectdrug)
00:39:28TheSevenflashing a rockbox binary *additionally* to iloader works
00:40:04amiconnSo you can select (with a keypress) what to boot?
00:40:18 Join perfectdrug [0] (
00:40:25TheSeveniloader is fully configurable
00:40:28 Quit planetbeing (Ping timeout: 252 seconds)
00:40:28 Nick planetbeing__ is now known as planetbeing (
00:40:39 Quit planetbeing_ (Read error: Connection reset by peer)
00:41:28amiconnOn the archoses, flashing works like that. There is a tiny (~2.5KB) flash loader, which only has two functions: Selecting one of two firmware images to boot (or a recovery serial "minimon"), and booting the selected image
00:41:40TheSevenyou can make it boot something automatically (or something different if you hold some key) or have a menu where each button is associated with an image, or have a scrollable list as a boot menu, or any combination of the above
00:41:59 Part orangey
00:42:12amiconnEach of the two images can be either ucl compressed (then it's decompressed to ram before execution) or uncompressed (then it executes directly from the nor flash)
00:43:08amiconnThe first image is permanent, i.e. it doesn't change anymore unless you do a full reflash (which is also needed to update the flash loader, and is rarely needed)
00:43:12 Quit dfkt_ (Read error: Connection reset by peer)
00:43:28TheSevenon the nano, there is a bootrom, which will load "norloader" (8KB, which will initialize the hardware, and boot the next loader stage, or drop into a built-in ibugger if the hold switch is on)
00:43:45amiconnThe second image (which is booted by default) can be updated easily with a flash plugin. This is the main rockbox binary
00:44:27TheSevennorloader will then usually load iloader, which is the configurable part
00:44:36amiconnYeah, the SH7034 also has a built-in boot rom (mask rom, not changeable in circuit, 64KB) which runs before the flash
00:44:46 Quit moos (Quit: ChatZilla 0.9.86 [Firefox 3.6/20100115144158])
00:44:48TheSevensame here
00:45:14TheSeveniloader will search on the data partition for a config file, and if there is none, look for one on the nor. the config file will then define how it behaves
00:45:16amiconnJust that there is no real encryption involved, just scrambling
00:45:34 Join Rob2222 [0] (
00:45:35TheSeventhe only part we still need to encrypt is norloader, the rest is unencrypted
00:47:13TheSevenso what did you mean above? replacing iloader with a rockbox binary, or just flashing rockbox into the nor and having it booted by iloader?
00:47:37pamauryhehe, I have a new sound card on my computer, it's called "e200 device" :) It's a shame it doesn't know how to playback sound yet
00:47:42amiconnOn archos we do no disk access before the main rockbox binary. (1) This reduces the risk of bricking (well, semi-bricking for ~99% of the devices - those with the 64KB boot rom can be recovered with a serial debug connection (RS232 + level converter))
00:48:01amiconn(2) It speeds up booting, especially on hdd based devices
00:48:24 Quit Rob2223 (Ping timeout: 240 seconds)
00:49:14TheSevenwell, rockbox on nand = 2 seconds total boot time
00:49:40amiconnIiuc it would (roughly) compare to loading rockbox from nor by iloader - with iloader's disk access cut away
00:49:43TheSevenhaving it on the nor could potentially bring that down to 1.5 seconds, but only if we drop support for a config file on the nand
00:50:26amiconnJust boot rockbox by default, and one or more other options on button press
00:50:57TheSeveni don't think that that fraction of a second is worth the loss of functionality though
00:52:27amiconnThat's what we do on archos (buttons given for recorders: (a) F1 boots the backup image (bootbox these days, which is a cut-down rockbox allowing disk access and charging. (b) F2 or no button at all boots the main image (rockbox). (c) F3 enters minimon
00:53:22amiconnThe Player and the Ondios have different buttons of course, but use the same system
00:55:23 Quit pamaury (Quit: Page closed)
00:55:27amiconnOn a flash based target a disk access is cheap, while on hdd it's not. Trying to access the hdd too early can even bring the dap into a state where it can't be used anymore without "external" help
00:57:25***Saving seen data "./dancer.seen"
00:57:29amiconnAt least that applies to recorder v1: If batteries are depleted, powering up the hdd may drop voltage below the level needed to keep it running, and then you can't charge anymore... (recorder v1 has software controlled charging)
00:58:23amiconn-> take out batteries and charge externally (hopefully you'll have an external AA NiMH charger around...)
00:58:52 Quit TheSeven (Read error: Connection reset by peer)
00:59:05amiconnI guess on the Nano G2 there is no such problem, because it's flash based
00:59:30 Join TheSeven [0] (~theseven@rockbox/developer/TheSeven)
01:18:17 Join shaggy-h [0] (
01:23:55 Quit liar (Ping timeout: 252 seconds)
01:25:31 Join planetbeing_ [0] (~planetbei@
01:29:20 Quit DerPapst (Ping timeout: 248 seconds)
01:29:23 Quit planetbeing___ (Ping timeout: 260 seconds)
01:31:17 Quit m3dlg (Ping timeout: 245 seconds)
01:31:42 Join DerPapst [0] (
01:31:53 Join m3dlg [0] (
01:35:58 Quit m3dlg (Remote host closed the connection)
01:36:24 Join m3dlg [0] (
01:41:04 Quit bertrik (Ping timeout: 258 seconds)
01:42:17 Join Rob2223 [0] (
01:42:31 Quit perfectdrug (Quit: perfectdrug)
01:43:09AlexPBuschel (logs) Yes sure, just highlight me and if I'm about I'll reply :)
01:45:27 Quit Rob2222 (Ping timeout: 245 seconds)
01:46:27 Join perfectdrug [0] (
01:51:38 Quit bzed (Remote host closed the connection)
02:04:25 Quit MethoS- (Remote host closed the connection)
02:04:58 Join Tomis2 [0] (~Tomis@
02:05:45 Quit Schmogel (Quit: Miranda IM! Smaller, Faster, Easier.
02:06:21 Quit Tomis (Read error: Operation timed out)
02:06:22 Nick Tomis2 is now known as Tomis (~Tomis@
02:14:59 Join bzed [0] (
02:16:12 Quit GeekShadow (Quit: The cake is a lie !)
02:29:47 Quit CGL (Ping timeout: 264 seconds)
02:32:13 Join CGL [0] (~CGL@
02:37:12 Join fyrestorm [0] (
02:53:35 Join n1s [0] (~n1s@rockbox/developer/n1s)
02:55:59 Quit Sajber^ (Read error: Connection reset by peer)
02:57:27***Saving seen data "./dancer.seen"
03:12:34 Quit CGL (Ping timeout: 276 seconds)
03:23:20 Join CGL [0] (~CGL@
03:36:06 Quit DerPapst (Quit: Leaving.)
03:42:33CIA-5New commit by funman (r25049): Sansa Clip: reduce current consumption for more correct battery life estimation after r25030
03:57:30 Quit CGL (Quit: Saliendo)
04:07:35 Quit n1s (Quit: Lämnar)
04:07:48JdGordonr25042 looks wrong....
04:13:38JdGordoncan someone forward me the svn commit email for 25042 please?
04:16:42JdGordonah screw it, I'll just fix it
04:20:37CIA-5New commit by jdgordon (r25050): revert part of r25042 so boot error splashes are localised, and so there isnt a boot slowdown (which was the origional reason for r24922)
04:20:37 Join Strife89 [0] (
04:28:44 Quit Barahir (Ping timeout: 265 seconds)
04:29:34 Join Barahir [0] (
04:30:30 Join dys` [0] (
04:33:11 Quit dys (Ping timeout: 276 seconds)
04:35:29 Quit Adnyxo (Ping timeout: 248 seconds)
04:35:33S_a_i_n_tTheSeven: Hey man...sorry, I slept in.
04:35:59*S_a_i_n_t relises this is a *late* response... :[
04:36:36 Quit planetbeing_ (Read error: Connection reset by peer)
04:36:49TheSevenS_a_i_n_t: i have an installer ready :-)
04:37:04 Join planetbeing_ [0] (~planetbei@
04:38:27 Join planetbeing___ [0] (
04:40:49 Quit planetbeing (Ping timeout: 248 seconds)
04:40:49 Nick planetbeing___ is now known as planetbeing (
04:43:37 Quit TheSeven (Disconnected by services)
04:43:49 Join The_Seven [0] (~theseven@rockbox/developer/TheSeven)
04:43:59 Nick The_Seven is now known as TheSeven (~theseven@rockbox/developer/TheSeven)
04:57:29***Saving seen data "./dancer.seen"
05:10:26 Quit planetbeing_ (Read error: Connection reset by peer)
05:35:07 Quit antil33t ()
05:43:58 Join Guest27451 [0] (
05:45:39 Part Guest27451
05:47:38 Join Guest27451 [0] (
05:48:07 Nick Guest27451 is now known as froggymana (
05:49:59 Join FlynDice [0] (~FlynDice@
05:51:44 Quit m3dlg (Ping timeout: 240 seconds)
05:52:12 Join m3dlg [0] (
05:52:40 Nick froggymana is now known as froggyman (
06:15:45 Quit shai (Read error: Connection reset by peer)
06:16:10 Join shai [0] (
06:35:23 Quit Strife89 (Ping timeout: 276 seconds)
06:36:44 Join FlynDice_ [0] (~FlynDice@
06:37:37 Quit FlynDice (Ping timeout: 248 seconds)
06:44:13 Join kaniini [0] (
06:57:32***Saving seen data "./dancer.seen"
07:11:33 Quit froggyman (Quit: he's got a towel! AHHH)
07:16:12 Join krazykit` [0] (
07:17:26 Quit krazykit` (Client Quit)
07:17:36 Join krazykit` [0] (
07:18:43 Quit krazykit (Disconnected by services)
07:18:53 Nick krazykit` is now known as krazykit (
07:21:56 Join Horschti [0] (~Horscht2@xbmc/user/horscht)
07:23:09 Quit FlynDice_ (Ping timeout: 264 seconds)
07:25:02 Quit Horscht (Ping timeout: 268 seconds)
07:31:40 Join antil33t [0] (
07:32:27 Quit karashata (Quit: The fluffy dragon has left completely!)
08:15:23CIA-5New commit by uchida (r25051): wave codec ...
08:27:47CIA-5New commit by uchida (r25052): ima adpcm/swf adpcm: corrects the problem the noise occurs after the play ends.
08:32:31 Join shai_ [0] (
08:34:07 Join ender` [0] (
08:34:38 Quit shai (Ping timeout: 245 seconds)
08:37:00 Quit arbingordon (Ping timeout: 264 seconds)
08:38:09 Quit m3dlg (Ping timeout: 248 seconds)
08:38:20 Join m3dlg [0] (
08:38:39 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
08:38:56 Quit Llorean (Quit: Leaving.)
08:41:48 Quit advcomp2019_ (Ping timeout: 264 seconds)
08:54:01 Quit shai_ (Quit: Leaving)
08:57:35***Saving seen data "./dancer.seen"
09:00:50 Join Strife89 [0] (
09:12:10 Join shai [0] (
09:17:54 Quit CaptainKewl (Remote host closed the connection)
09:20:23 Join teru [0] (
09:27:47 Join godoitis [0] (
09:28:05 Nick godoitis is now known as patgodo (
09:38:43 Join GodEater [0] (~bibble@rockbox/staff/GodEater)
09:39:55 Quit advcomp2019 (Read error: Connection reset by peer)
09:40:16 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
09:44:03 Join flydutch [0] (
09:50:36 Join Buschel [0] (
09:55:27 Quit mt (Ping timeout: 265 seconds)
10:03:16 Join liar [0] (
10:08:06 Join Zack [0] (~chatzilla@
10:10:50 Quit advcomp2019 (Ping timeout: 252 seconds)
10:11:03Zackis there any kind of hack for the zune
10:12:04 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
10:12:05 Quit planetbeing (Quit: planetbeing)
10:13:26 Join planetbeing [0] (
10:16:30BuschelAlexP: are you there?
10:19:22 Join MethoS- [0] (~clemens@
10:20:26 Quit planetbeing (Quit: planetbeing)
10:22:02 Part patgodo
10:24:13 Join planetbeing [0] (
10:24:50 Quit Zack (Quit: ChatZilla 0.9.86 [Firefox 3.6/20100115144158])
10:35:37AlexPBuschel: Just about :)
10:37:17 Join planetbeing_ [0] (~planetbei@
10:37:17 Quit planetbeing (Read error: Connection reset by peer)
10:37:18 Nick planetbeing_ is now known as planetbeing (~planetbei@
10:38:18BuschelI have something for you :o) ->
10:38:55AlexPRighto :)
10:39:29Buschelit's just about speed, functional changes
10:39:40 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
10:42:37BuschelI found another piece of data that has not been put to iram yet
10:42:58 Quit advcomp2019 (Ping timeout: 245 seconds)
10:43:58 Join DerPapst [0] (
10:47:21 Join advcomp2019__ [0] (~advcomp20@unaffiliated/advcomp2019)
10:48:02AlexPBuschel: 418.5%, 29.67MHz
10:48:21Buschelat least reasonable faster
10:48:40AlexPyes, creeping closer :)
10:49:16Buschelcan you make some local changes on the patched version? afterwards it won'z play music but we can measure some things with the codec
10:50:26 Quit advcomp2019_ (Ping timeout: 252 seconds)
10:51:20AlexPyep, sure
10:51:26Buschelsynthfilter.c, line 540: comment line 550 (//mpc_decoder_windowing...)
10:52:53AlexPcomment out line 550 only, or 540 too?
10:53:18Buschelonly the call of mpc_decoder_windowing
10:53:47AlexPanything else?
10:54:17Buschelno. just compiling, no clean make needed. yu can just copy the mpc.codec file to the codecs path on you dap
10:54:40Buschelwell, typomania...
10:54:53*Buschel get some more coffee
10:57:38***Saving seen data "./dancer.seen"
10:58:22AlexPBuschel: 556.36%, 22.32 MHz
10:58:51Buschelok, now also comment the line above (mdct32)
11:00:33AlexP895.63%, 13.86 MHz
11:00:34 Quit fyrestorm (Read error: Connection reset by peer)
11:01:51 Join fyrestorm [0] (
11:03:32Buschelok, now comment line 203 and 204 in mpc_decoder.c (call of mpc_decoder_requantisierung and mpc_decoder_synthese_filter_float)
11:04:02Buschelbtw, the results are totally different to arm
11:04:35AlexPJust to be useful :)
11:06:08AlexP1097.86%, 11.31 MHz
11:06:30Buschelperfect. now we need to do the same with svn :o)
11:07:12AlexPyeah :)
11:08:02Buschel11.31 MHz is what the bitstream demuxing needs. It is not comparable to arm numbers, but this seems too high... Lets see
11:10:34 Join karashata [0] (
11:14:48 Join S_a_i_n_t_ [0] (S_a_i_n_t@
11:16:11 Quit S_a_i_n_t (Ping timeout: 276 seconds)
11:21:06BuschelAlexP: the changes in synthfilter.c should be nearly the same (except detailed location)
11:33:38 Join patgodo [0] (
11:37:10BuschelAlexP: any results yet?
11:37:33AlexPWhat from?
11:37:50Buschelsame test with local changes to svn? ;o)
11:38:14AlexPOh, sorry - I didn't realise you asked for those :)
11:38:20AlexPGive me a few minutes :)
11:38:33Buschelwe misunderstood and lost 30 minutes now :o)
11:38:56pixelmaI wouldn't have understood this either
11:39:19*Buschel needs to speak more clearly
11:41:05 Quit leavittx (Ping timeout: 248 seconds)
11:44:00 Quit Strife89 (Ping timeout: 260 seconds)
11:45:46 Join bertrik [0] (
11:54:25AlexPBuschel: Could you confirm line numbers in svn for me? I think I'm getting the correct ones, but just want to be sure
11:55:08AlexPI commented line 549 in synthfilter.c, then after 548
11:55:22Buschelyes, just wanted to state this.
11:55:58Buschelafterwards line 441 and 442 in mpc_decoder.c
11:56:06AlexPOK, starting from SVN 483.94%, 25.66MHz - commented line 549 690.21%, 17.99 MHz - commented line 548 as well 1314.15%, 9.44 MHz
11:58:53AlexPno change with the change in mpc_decoder.c
11:59:07AlexPyeah, somewhat odd
11:59:13AlexPBut I did it twice
12:00:29AlexPI see mpc_decoder_requantisierung4 and 516 and mpc_decoder_synthese_filter_float also at lines 51
12:00:41AlexPer, 514 and 516 not 51
12:01:10 Part patgodo
12:01:17Buscheljust comment all of them
12:01:37AlexPall of each, or all of when I see the two together?
12:01:42 Join Sajber^ [0] (~Sajber^
12:01:46Buschelall of them :)
12:01:51AlexPe.g. line 462 too?
12:02:39Buschelthe old sv7 decoder is a bit fuzzy compared to new one
12:04:00AlexPI can't comment 604 otherwise it won't build
12:04:14AlexPis that a declaration or somesuch?
12:04:23*AlexP has next to zero knowledge here :)
12:04:28Buschelyes, that is the declaration
12:04:58Buschelyou do _not_ have zero knowledge
12:05:18Buscheli predicts a result of 6.5-7.0 MHz w/o the filter/requantization
12:05:49AlexPhaha - 6.87 MHz
12:07:03Buschelthis bad though... might be hard to optimize. it is the bitstream reading...
12:07:30AlexPyeah, the difference there is the overall difference
12:09:11BuschelI assume this is an effect of the sv8 decoder using unit8 buffers instead of uint32. More loads on the non-iram'ed main bitstream buffer...
12:10:36Buschelplease test for speed
12:11:04Buschelif this leads to the right direction, i can do some further changes
12:12:03 Nick S_a_i_n_t_ is now known as S_a_i_n_t (S_a_i_n_t@
12:13:48 Quit Farthen (Ping timeout: 245 seconds)
12:14:31 Quit perfectdrug (Quit: perfectdrug)
12:15:14AlexPBuschel: 414.46%, 29.96 MHz
12:15:30AlexPslightly slower than v15
12:15:32Buschelmore or less no effect
12:15:38AlexPyeah, no real change
12:16:48Buschelbtw, can you try to change line 25 in libmusepack.make to -O1, rebuild and test?
12:26:32AlexPBuschel: 377%, 32.94 MHz
12:26:51Buschelok, so I better leave this at -O2
12:27:00AlexPlooks like it :)
12:28:29 Join stooo [0] (
12:49:47 Join pamaury [0] (~c2c7a50a@gateway/web/freenode/x-sctudgqaprzvqsya)
12:50:18 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
12:53:23 Join petur [0] (
12:53:24 Quit petur (Changing host)
12:53:24 Join petur [0] (~peter@rockbox/developer/petur)
12:55:08 Part domonoky
12:55:13BuschelAlexP: can you please test ?
12:56:48AlexPBuschel: yep, it'll be 20 minutes or so - just haveing a spot of lunch :)
12:57:04BuschelAlexP: great. thanks in advance
12:57:14 Quit stooo (Read error: Connection reset by peer)
12:57:39 Join stooo [0] (
12:57:42***Saving seen data "./dancer.seen"
12:57:59Buscheldo we have an overview of the iram usage for all targets?
13:06:15BuschelAlexP: sorry did a mistake. please use the v18 version ->
13:14:28 Join robin0800 [0] (
13:17:45 Join Farthen [0] (
13:18:05 Join krzakx [0] (~krzakx@
13:19:49 Part krzakx
13:20:13 Join kugel [0] (~kugel@rockbox/developer/kugel)
13:21:34AlexPBuschel: region IRAM is full (/home/alex/rockbox/build/H120-mpc/rockbox.elf section .stack)
13:21:59AlexPwith v18
13:22:38 Quit shaggy-h (Ping timeout: 240 seconds)
13:24:20 Quit kugel (Disconnected by services)
13:24:25 Join kugel [0] (
13:24:35 Quit kugel (Changing host)
13:24:36 Join kugel [0] (~kugel@rockbox/developer/kugel)
13:36:04 Join domonoky1 [0] (
13:39:47BuschelAlexP: back again.
13:44:34AlexPdid you see the iram full issue?
13:45:00Buschelyes, i am rebuilding for h100 now.
13:45:25Buschelok, got the same now
13:50:41 Join moos [0] (moos@rockbox/staff/moos)
13:51:26CIA-5New commit by teru (r25053): Pass width instead of x+width to ab_draw_markers()/cue_draw_markers() and don't recalculate the width in these functions.
13:52:04BuschelAlexP: that is really baaad... I need another solution....
14:02:57 Part domonoky1
14:08:14 Quit advcomp2019__ (Ping timeout: 265 seconds)
14:09:20 Join gill0r [0] (
14:23:33 Join mt [0] (~mtee@rockbox/developer/mt)
14:26:19 Quit kugel (Remote host closed the connection)
14:39:21CIA-5New commit by teru (r25054): Slightly change how %Li works when used as a conditional. last part is selected if and only if icon is not presented (i.e. is no icon) so that it can ...
14:39:59 Join Strife89 [0] (
14:40:17S_a_i_n_tDammit! That better not break the theme I was just about to upload :P
14:40:56 Quit guymann (Ping timeout: 276 seconds)
14:42:49 Join Schmogel [0] (
14:48:29S_a_i_n_tteru: So, does that last commit still work if I want to supply an icon for cases where there is none? As, that was one of the things I liked about the way it was prior.
14:50:52teruif you set all 33 items explicitly, there shouldn't be any difference.
14:51:38 Join einhirn [0] (
14:51:46Buschelamiconn: how many cycles does a load from iram vs dram take on coldfire? so: what is the penalty?
14:53:08S_a_i_n_tteru: Thanks.
14:56:26 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
14:56:26 Quit pixelma (Disconnected by services)
14:56:42 Quit simabeis (Remote host closed the connection)
14:56:43 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
14:57:00 Quit amiconn (Disconnected by services)
14:57:02 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
14:57:26 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
14:57:46***Saving seen data "./dancer.seen"
15:05:16 Quit karashata (Quit: The fluffy dragon has left completely!)
15:11:19 Quit petur (Remote host closed the connection)
15:12:13 Join doomcup [0] (
15:12:21doomcupHi there
15:12:51doomcupI'm having some problems with rockbox on my 5th generation ipod
15:12:59 Quit ecio (Ping timeout: 265 seconds)
15:13:10doomcupI just tried copying some files over, and during the transfer the ipod froze
15:13:25doomcupI had to disconnect it and manually reset it
15:13:39AlexPWhat version of Rockbox?
15:16:01doomcupIs there a way to check on the device? It's been a long while since I made the install
15:16:17doomcupNevermind, found it
15:16:21doomcupIt's version 3.3
15:17:00gevaertshm, 3.3 on ipod uses the Apple ROM USB mode
15:17:23doomcupSo it might be a hardware issue?
15:19:18AlexPor disk corruption or an OS issue
15:19:31AlexPI assume you are using Windows?
15:19:44doomcupI'm actually using Ubuntu
15:19:48doomcup9.10, IIRC
15:20:03doomcupYes, 9.10
15:20:11gevaertsIt could be. I've never really heard of issues with that mode. Just to confirm, when you connect, does it show anything like (without the pink), or a blacn and white screen saying something like "Do not disconnect"?
15:20:11AlexPOK, so I'd try running fsck.vfat on it
15:20:56doomcupgevaerts, yeah, it shows the no-sign with do not disconnect, which quits flashing when it freezes
15:21:05doomcupAlexP, alright, I'll get right on that
15:21:16gevaertsok, so it's definitely the Apple mode. That means that this is not my fault :)
15:22:16*gevaerts wouldn't really expect file system corruption to make the *device* freeze
15:22:32AlexPme neither, but it is an easy first step
15:23:10doomcupThere does seem to be some problems
15:23:20 Quit einhirn (Read error: Connection reset by peer)
15:23:22doomcuppastebin forthcoming
15:23:59gevaertsyes, and it has been known to cause just about every other sort of issue
15:24:27doomcupIt's prompting me to correct it, should I?
15:24:54AlexPyes :)
15:25:03gevaertsDo you have backups of everything on it?
15:25:15doomcupYeah, fortunately
15:25:19doomcupOkay, correcting it
15:25:58doomcupThat is wierd
15:26:14doomcupCorrecting it seemed to have no discernable effect
15:26:21AlexPYou need -r (or -a)
15:27:30doomcupOkay, trying the transfer again
15:27:39doomcupLet's see if it freezes this time as well
15:29:12doomcupAlso, now that I notice I'm two versions behind, would upgrading be a good idea, you think?
15:29:24AlexPcertainly :)
15:30:56 Join simabeis [0] (
15:32:18 Join einhirn [0] (
15:32:27 Join FlynDice_ [0] (~FlynDice@
15:32:53 Nick FlynDice_ is now known as FlynDice (~FlynDice@
15:33:04doomcupAlright, I'll just grab that while this is going on
15:33:16 Quit einhirn (Client Quit)
15:39:01 Part TMM ("Ex-Chat")
15:48:32 Quit Sajber^ (Ping timeout: 252 seconds)
15:51:59 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
15:52:36 Quit Horschti (Ping timeout: 264 seconds)
15:55:00JdGordonteru: why the heck did you change %li?
15:55:14 Join stoffel [0] (
15:55:24 Join ecio [0] (
15:55:44teruis there any problem?
15:55:59JdGordonwhat was the point?
15:56:18JdGordonit is also not consistant with the other tags now
15:57:44terui just don't want to do %?Li<%Vda|%Vda|%Vda|...|%Vda||%Vdb> to display something different whether icon is presented or not.
15:57:45JdGordonyou also didnt modify the wiki or manual
15:58:42AlexPThe manual needs it adding in the first place...
15:59:39JdGordonteru: now you HAVE to do some really long line for that to work
16:00:04JdGordonyou used to be able to do %?Li<|%vda> if you wanted a shown when an icon is there and nothing when there isnt
16:01:37JdGordonthe tags are all %?xx<no|yes> or %?<yes> so you should revert that change
16:02:18teruhave you actualy tested? things didn't work like that to me.
16:03:03JdGordonskin_display.c:581 else if (intval > num_options || intval < 1) intval = num_options;
16:03:29JdGordonintval would be 2+ num_options would be 2
16:03:29soapwhen encouraging iPod users to upgrade from a Rockbox release using Apple's USB mode to one using Rockbox's USB mode, should we not explicitly mention the charging issue? I know it is documented, but even someone familiar with 3.3's manual (for example) is unlikely to pan out that nugget.
16:03:41AlexPsoap: 3.5 uses apple usb
16:04:03soapthen there isn't an issue!
16:04:06teruand according to, tags are %?xx<yes|no> or %?<yes>.
16:04:07*soap slowly backs away.
16:04:07doomcupAnd it finished with no problems
16:04:17doomcupThanks for your help guys
16:04:23AlexPcool, no problem
16:04:29doomcupI'll give the new version a shot
16:05:04 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
16:07:39JdGordonrg, mm, mp, xf, and others all say otherwise
16:07:44doomcupKeeps giving a permission denied error
16:07:51doomcupIs there something I'm forgetting to do?
16:08:16doomcupAh, right, as root
16:08:25doomcupWHen I did this last time I was in windows :p
16:08:54JdGordonsx pparently is half and haf
16:09:27 Quit MethoS- (Remote host closed the connection)
16:10:49 Quit FlynDice (Remote host closed the connection)
16:11:11 Join dfkt [0] (dfkt@unaffiliated/dfkt)
16:11:30 Join FlynDice [0] (~FlynDice@
16:12:06S_a_i_n_tI'm personally with reverting the %Li change as well...
16:12:16S_a_i_n_tIt worked fine for me as it was.
16:15:25JdGordonteru: that 2 option didnt work because the no icon one never worked because line 61 in stausbar_skinned.c is wrong
16:15:58JdGordon1 is the first option, not 0
16:16:57*S_a_i_n_t has/had it working (and thinks he still does) as o=no icon, 1, 2, 3 etc...
16:17:10S_a_i_n_t*0 rather
16:20:14JdGordonI don't believe you :)
16:24:37 Quit stoffel (Ping timeout: 246 seconds)
16:24:39teruthis is one i made to test and it seems to show correct icons (with my change):
16:25:57 Join anewuser [0] (anewuser@unaffiliated/anewuser)
16:28:15JdGordonstatusbar_skinned.c line 61 just needs to be changed to +2 instead of +1 for it to work how it should
16:28:29JdGordoneverything where its off and a list of values has off first
16:29:09 Quit Farthen (Quit: ChatZilla 0.9.86 [Firefox 3.6/20100115144158])
16:29:32 Join Adnyxo [0] (
16:32:48 Quit FlynDice (Ping timeout: 264 seconds)
16:33:27S_a_i_n_tWhy does checkwps spit now when a font is "included in the fontpack"...I don't like the idea that my theme won't work if someone forgets to read that they have to download the font pack also...
16:33:49S_a_i_n_tfor the sake of 2 fonts :[
16:34:24gevaertsS_a_i_n_t: two fonts *per theme*...
16:34:43 Quit DerPapst (Quit: Leaving.)
16:34:51rasherS_a_i_n_t: Plus when the font pack is updated and people install your theme, the new version of the font is overwritten by the old version in your theme
16:35:29S_a_i_n_tCan I include bold type in my theme description?
16:35:45S_a_i_n_tI can imagine a lot of people expecting it to "just work"
16:37:27S_a_i_n_tbah...shit. Even if I link to the font pack, they'll have to cut&paste the link...Hmmm.
16:37:45AlexPThe manual links to it, and suggests they install it
16:37:54AlexPRockbox Utility will also install it
16:38:41AlexPJust say something like "Requires some fonts from the font pack"
16:38:51S_a_i_n_tI'm trying to figure out a clear enough way to tell people that they MUST download the fontpack for it to work...
16:39:02S_a_i_n_tyeah, I was thinking along those lines.
16:41:53 Join CGL [0] (~CGL@
16:46:43 Quit n17ikh (Ping timeout: 245 seconds)
16:46:59 Join FlynDice [0] (~FlynDice@
16:51:30 Join n17ikh [0] (
16:52:03S_a_i_n_t <−− Idiot proof description?
16:56:22 Quit antil33t (Read error: Connection reset by peer)
16:56:28 Join antil33t [0] (
16:56:53pamaurygevaerts: did you see my comments about usb core this night ?
16:57:21gevaertsI did, yes
16:57:49***Saving seen data "./dancer.seen"
16:58:39gevaertsThe short answer is that anything that isn't needed by MSC or HID will be either missing or buggy :)
16:58:57pamauryI think some should be fixed right now. For example the endpoint requesst that are all interpreted as standard requests. I also, correct me if I'm wrong but SET_INTERFACE and GET_INTERFACE is invalid for interface with only one possible setting so it should fail instead of being ignored
16:59:36pamaury*I also -> And also
17:01:08pamauryThe rest is not a problem and will eventually be fixed if I manage to have usb audio work
17:02:08pamauryAh yes, there is another problem. iirc in the endpoint handler, the core call ep_data[EP_NUM(req->wIndex)].control_handler[EP_CONTROL] which is typo:
17:02:23pamauryis should be EP_DIR(req->wIndex) instead of EP_CONTROL
17:02:34 Join CaptainKewl [0] (~jason@
17:02:44pamauryI don't have the code to check but I think I corrected it my usbaudio code
17:03:34pamauryIf you don't mind, I'll commit a little fix tomorrow
17:04:06gevaertsSo what you're saying is that request_handler_endpoint() should be more like request_handler_interface()?
17:04:37gevaertsand yes, that EP_DIR(req->wIndex) does indeed look wrong
17:04:53gevaertshm, the EP_CONTROL of course :)
17:05:44pamauryyes, basically request_handler_endpoint should first check for the request type and then call something like request_handler_standard_endpoint
17:06:55pamauryFor usbaudio, this is a must have because otherwise there is a conflict. But more generally this is better sanity check, even though the set of endpoint request is quite limited and doesn't do anything important
17:06:57gevaertsabout SET_INTERFACE/GET_INTERFACE with only one available interface, I don't actually know.
17:07:24gevaertsin general, if you see something wrong, please fix it :)
17:07:25pamauryI think I read it in the spec, I can try to find it
17:08:14pamauryyes but I can't commit from here. The "université de Paris" nicely blocks all ports except 80 :(
17:08:27pamauryI'll commit from work :)
17:12:54pamaurygevaerts: usb 2.0: 9.4.10 (SET_INTERFACE): If a device only supports a default setting for the specified interface, then a STALL may be returned in the Status stage of the request.
17:13:21*gevaerts doesn't really like that word in specs...
17:13:26 Join Sajber^ [0] (~Sajber^
17:13:34pamauryyes, that's error prone
17:13:48 Join FlynDice_ [0] (~FlynDice@
17:14:29pamauryperhaps they define what may means and it means "must" in reality :)
17:14:40 Quit FlynDice (Ping timeout: 246 seconds)
17:15:55gevaertsI don't know what we should do then. I'd expect stalling to trigger more OS bugs, but I can easily be wrong there
17:16:23pamauryNo, I think it should stay like that then (until we support multiple settings).
17:18:25 Quit teru (Quit: Quit)
17:18:40AlexPBuschel: v20 450.32%, 27.57 MHz - getting there :)
17:19:01Buschelwild caching of buffers
17:27:59 Quit CGL (Ping timeout: 276 seconds)
17:28:30AlexPJdGordon: Thanks for FS11080 :)
17:31:08 Join domonoky1 [0] (
17:41:09AlexPBuschel: v21 466.21% 26.64 MHz
17:41:57BuschelAlexP: \o/
17:42:41AlexPJust 1.3 MHz off svn now :)
17:43:09Buschelthe last 1.3 MHz will be hard...
17:43:30BuschelI guess there won't be much gain from now on.
17:43:46AlexPThat's not bad at least :)
17:44:15 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
17:44:48 Quit advcomp2019 (Ping timeout: 264 seconds)
17:45:11AlexPBuschel: The Glinka sv8 sample is 446.57%, 27.80 MHz
17:48:33 Quit FlynDice_ (Ping timeout: 248 seconds)
17:48:57 Join FlynDice_ [0] (~FlynDice@
17:54:41 Join kugel [0] (~kugel@rockbox/developer/kugel)
17:55:27*kugel tests gcc 4.4.3+binutils 2.20.1 (eabi)
17:56:22 Quit FlynDice_ (Ping timeout: 265 seconds)
17:58:24 Join Barahir_ [0] (
17:58:55TheSevenNotice: Smarty error: html_image: 'themes///' is not a valid image file in /usr/share/php/smarty/libs/Smarty.class.php on line 1092
17:59:02TheSevenWarning: zip_read() expects parameter 1 to be resource, integer given in /home/themes/private/themesite.class.php on line 166
18:00:53 Quit Barahir (Ping timeout: 245 seconds)
18:01:40TheSevenwhat does this in the top right corner mean?
18:02:09gevaertsI think it means that there are themes in "reported" status
18:02:11AlexPThat an admin needs to look at something
18:02:25AlexPSomeone has reported (a) theme(s)
18:10:40 Join Strife1989 [0] (
18:11:30AlexPBuschel: v22 459.74% 27.01 MHz
18:12:05BuschelAlexP: ok, so the last was a bit faster...
18:12:20AlexPyep, v21 was a little fasteer
18:14:08 Quit Strife89 (Ping timeout: 260 seconds)
18:14:17 Quit m3dlg (Ping timeout: 252 seconds)
18:14:46 Join m3dlg [0] (
18:20:32kugelwhat happened to my Kore theme?
18:21:03kugelit's gone! :?
18:22:32 Join S_a_i_n_t_ [0] (S_a_i_n_t@
18:23:22 Join tuma [0] (
18:25:10 Quit S_a_i_n_t (Ping timeout: 276 seconds)
18:26:22 Quit kugel (Remote host closed the connection)
18:26:23AlexPBuschel: 464.01% 26.76 MHz
18:26:40AlexPfor v23 that is
18:26:47AlexPv21 is still just winning :)
18:28:37 Join ecio1 [0] (
18:28:58 Join FlynDice_ [0] (~FlynDice@
18:29:08 Join advcomp2019__ [0] (~advcomp20@unaffiliated/advcomp2019)
18:30:16 Quit ecio (Ping timeout: 246 seconds)
18:30:28 Quit Strife1989 (Ping timeout: 260 seconds)
18:31:11 Quit pamaury (Ping timeout: 252 seconds)
18:31:18 Quit advcomp2019_ (Ping timeout: 245 seconds)
18:34:33 Quit advcomp2019__ (Ping timeout: 265 seconds)
18:34:49 Join kugel [0] (~kugel@rockbox/developer/kugel)
18:36:41kugelis there a way to batch download the test_files?
18:38:09 Join merbanan [0] (
18:39:31amiconnBuschel: Loading from IRAM takes as many cycles as given in the SoC documentation (iirc 2 cycles). Loading from DRAM takes many many cycles (no data cache!)
18:40:07kugelUnhelpful: how exactly? I mean without necessarily knowing the files in the dir
18:41:19 Join Llorean [0] (
18:41:19 Quit Llorean (Changing host)
18:41:19 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
18:42:35Buschelamiconn: ok, that explains everything :o) I just wrote a small wrapper to achive better performance on Coldfire. Works quite wll so far
18:45:44AlexPBuschel: v25 470.63%, 26.38 MHz
18:45:55AlexPquickest yet!
18:45:57Buscheleven faster! :)
18:46:17kugelfor the record, wget -r −−no-parent works
18:46:30 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
18:46:37BuschelAlexP: that is +1 MHz against svn, correct?
18:47:24AlexPyes, SVN is 490.49% realtime, 25.31 MHz
18:48:19 Quit anewuser ()
18:48:31 Join pamaury [0] (~c2c7a50a@gateway/web/freenode/x-fklgdiypbghiancp)
18:48:41pamauryDoes anymore here has a usbaudio device
18:49:08gevaertspamaury: I have one at home :)
18:49:47pamauryAh, could you please give me the result of a complete lsusb on that device (use the verbose mode) ?
18:50:18gevaertsyes, next saturday :)
18:50:28Tornei have an xbox pad, it has usbaudio in it
18:50:53Torneis that useful?
18:51:05Tornedon't recall if it has a hub or if it's a composite device
18:52:00pamauryTorne: you can try, I'll tell you if it's useful when I'll have it :)
18:52:09Torneok, one sec
18:55:10Torneah, i don't think it's actually a standard audio device..
18:55:20pamaurywhy ?
18:55:56Torneall the itnerfaces are class 255
18:56:07 Quit kugel (Ping timeout: 260 seconds)
18:56:37pamauryah, indeed, I guess lsusb is not that verbose then
18:57:50***Saving seen data "./dancer.seen"
18:59:13 Join kugel [0] (~kugel@rockbox/developer/kugel)
19:01:32pamauryyes, not very useful :(
19:03:14 Join Kitr88 [0] (
19:06:19 Quit Kitar|st (Ping timeout: 256 seconds)
19:07:22 Quit Kitr88 (Ping timeout: 246 seconds)
19:09:09 Quit FlynDice_ (Ping timeout: 256 seconds)
19:13:27 Join Kitar|st [0] (Kitr88@
19:14:18 Join guymann [0] (
19:14:30 Join perfectdrug [0] (
19:15:56pixelmadomonoky1: did you experiment with rbspeexenc on a Mac? I seem to remember you investigating broken voice files or something generated with RbUtil but am not sure
19:17:13domonoky1not as far as i remember, testing rbutil over ssh is difficult :-) Maybe bluebrother did ?
19:18:21domonoky1but i think i had reports of working voice generation on mac, atleast with the opensource tts (flite espeak or festival)
19:18:26pixelma I thought I remembered you looking at the voice files themselves
19:19:18domonoky1as i did the voicesystem in rbutil, i might have done that, but cant remember any exact results :-)
19:20:36pixelmaI'm testing bluebrother's work to make use of Apple's TTS on MacOS 10.4 - and it works on Intel Mac but only for my Ondio (MP3 files), the speex talk clips for my M5 just sound gibberish and the voice file doesn't want to work at all
19:21:37pixelmaa PPC Mac has a different problem - there RbUtil hangs completely when it has to use the TTS in some way
19:22:29domonoky1i would suspect that the generated wav file is somehow in a format rbspeex doesnt understand.
19:22:34pixelma10.5 (on an Intel Mac) doesn't have any problems
19:23:17pixelmano, intermediate aiff and wav files are fine, I tested a version which didn't remove those and checked
19:24:17domonoky1if you have such a intermediate wav file, you could encode/decode it with rbspeexenc /rbspeexdec and test how it sounds.
19:25:02kugeldomonoky1: any idea what happened to my Kore theme?
19:25:05domonoky1also testing with filte/espeak or similar on this 10.4 Macs would perhaps be helpfull to locate the error.
19:25:13kugelit seems to be completely gone from the theme site
19:25:13pixelmaI'd need an rbspeexenc for MacOS 10.4 right?
19:27:02domonoky1pixelma: jup, or at least a rebspeexenc/dec compiled against a 10.4 sdk.
19:27:12 Join arbingordon [0] (
19:28:05 Quit tuma (Quit: leaving)
19:28:10domonoky1if filte/espeak work with rbutil on mac 10.4, it has to be the wav files.. if not, the error could be somewhere else.
19:28:32domonoky1kugel: do you know the theme id of this theme ?
19:28:52pixelmadomonoky1: ok, some more things to test then. I'm not sure I can compile rbspeexenc (or RbUtil) there because those are work Macs and I don't want to mess around too much with them
19:29:40domonoky1pixelma: maybe bluebrother or someone else can compile one for you ? (i dont have access to a mac at moment).
19:30:40pixelmayeah, maybe he sees our small discussion in the logs
19:31:23kugeldomonoky1: luckily I have posted the link at the abi forums. theme id is 549
19:31:31kugel shows it actually.
19:31:42domonoky1kugel: at least the theme is still in the db :-)
19:31:51kugelbut it's not under the theme listings (not even under admin -> browse all -> showing all)
19:32:29domonoky1kugel: its probably invisible because it fails all checkwps versions ?
19:32:40rasherShould be in the admin interface though
19:32:47domonoky1but admin page should show them, unless there is a bug.
19:32:50kugelshouldn't I be able to see those in the admin panel?
19:33:21*domonoky1 goes checking code.
19:34:18kugelcould it fail through checkwps because the "statusbar: custom" setting has been removed?
19:35:10kugelif yes, that would be bad IMO, since that line doesn't have any impact but would even though break all existing sbs enabled themes
19:35:40domonoky1thats a setting in the config ? i think only sbs/wps are checked in the checkwps runs..
19:35:55kugelah yes, cfg checking was a sperate beast :)
19:35:56 Join FlynDice_ [0] (~FlynDice@
19:36:09 Join advcomp2019__ [0] (~advcomp20@unaffiliated/advcomp2019)
19:37:15domonoky1hm, the themepage always list only themes which pass the checkwps check. i blame rasher for this code :-)
19:37:30rasherThis is clearly a mistake
19:37:37rasher(for the admin pages)
19:38:09domonoky1jup, that should be corrected. admin page could generally need more work.
19:38:30gevaertskugel: the %Fl tag was changed to require the extension
19:39:05gevaertsyou just need to add .fnt
19:39:14kugelah, I missed that commit
19:40:12domonoky1we could need more people hacking on the theme page :-)
19:41:10 Join dfkt_ [0] (~dfkt@unaffiliated/dfkt)
19:41:13kugelI thought themes failing checkwps are marked hidden?
19:41:43rasherNo. checkwps failure is different
19:42:17 Quit bmbl (Ping timeout: 252 seconds)
19:42:46 Join Farthen [0] (
19:44:38 Quit dfkt (Ping timeout: 265 seconds)
19:49:07 Quit arbingordon (Changing host)
19:49:07 Join arbingordon [0] (~w@unaffiliated/arbingordon)
19:53:17 Join efyx_ [0] (
20:01:28 Join planetbeing_ [0] (
20:02:58CIA-5New commit by Domonoky (r25055): show also themes which fail all checkwps in the admin interface.
20:12:36Buschelanything speaking against submitting the mpc sv8 decoder)? the decoder does not fully reach svn's performance for sv7, but the missing 2-4% will be hard to tweak due to decoder internal restructuring.
20:12:44Buscheland of course we will have sv8 then :)
20:13:12 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
20:13:15*domonoky1 wants to propose website work, ie improved themepage, a fm preset db, translation services or online voice generation for rockbox gsoc. What do people think ?
20:16:19 Join DerPapst [0] (
20:21:28kugelBuschel: i don't use mpc, so I'm all for supporting more codecs :)
20:22:18 Quit domonoky1 (Quit: Leaving.)
20:22:39 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
20:22:48 Quit robin0800 (Remote host closed the connection)
20:23:16Buschelok, then I will at least no brake the playback for _you_ :o)
20:31:43 Quit advcomp2019__ (Ping timeout: 245 seconds)
20:33:21 Quit merbanan (Ping timeout: 248 seconds)
20:33:32domonokyBuschel: new codecs are always good :-)
20:33:34 Join MethoS- [0] (~clemens@
20:34:47CIA-5New commit by Buschel (r25056): Major change to musepack decoder: Import v1.3.0 (r458 from to rockbox. Several adaptions in the musepack decoder were made to get ...
20:34:50*Buschel crosses fingers
20:37:58 Join FOAD_ [0] (
20:40:57Buschelnot too bad :)
20:41:37 Quit FOAD (Ping timeout: 240 seconds)
20:41:38 Nick FOAD_ is now known as FOAD (
20:44:23 Join leavittx [0] (~leavittx@
20:46:34 Join merbanan [0] (
20:48:00 Join punkt [0] (
20:49:54 Join stripwax [0] (
20:50:01 Quit Barahir_ (Ping timeout: 256 seconds)
20:57:53***Saving seen data "./dancer.seen"
20:58:25 Join Barahir [0] (
20:59:31*Buschel is searching for elegant way to solve the iram issue in svn
21:01:05kugelBuschel: you're not the first one :)
21:01:57kugelideally there should be a #define in some .h saying how much iram can be used for core/codecs. as of now the splitting is in the linker scripts where the rest of the code doesn't have access to
21:04:36Buschelbtw, why is IRAM for M5/X5 splitted to 64KB/64KB for apps/plugins? Other targets with large IRAM still use 48KB for apps and give the additional IRAM to the plugins. This way the codecs can be further optimized −− especially as Coldfire are very IRAM-hungry in terms of speed
21:05:32 Quit stripwax (Ping timeout: 276 seconds)
21:17:47 Quit Adnyxo (Remote host closed the connection)
21:20:33 Join stripwax [0] (
21:20:37 Join Adnyxo [0] (
21:22:50CIA-5New commit by Buschel (r25057): Set IRAM split to 48KB / 80KB (apps / plugins) for M5 and X5. This way codecs can be optimized further.
21:23:01 Quit m3dlg (Ping timeout: 240 seconds)
21:23:38 Join m3dlg [0] (
21:29:46 Quit stripwax (Ping timeout: 276 seconds)
21:35:22CIA-5New commit by Buschel (r25058): Move WMA decoder array to IRAM on M5/X5 for better performance.
21:38:08 Join petur [0] (~petur@rockbox/developer/petur)
21:41:26kugelwhy do the recent metadata commits lead to such exessive binsize increases?
21:42:57kugelBuschel: do you introduce floating point operations?
21:43:27kugel"gain = (long)(SV8_TO_SV7_CONVERT_GAIN - (gain*100./256.));" ?
21:44:51CIA-5New commit by Buschel (r25059): Fix red.
21:45:26 Quit bluebrother (Disconnected by services)
21:45:26 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
21:45:30Buschelkugel: yes, this is called twice per track.
21:45:57Buschelkugel: i have an idea where the bin increase comes from. give me minutes to check.
21:46:14kugelwell, linking software fp into rockbox could explain it
21:46:42kugelcheck out the scripts in utils/analysis/
21:46:57BuschelI did it when researching the gain calculation...
21:51:02bluebroth3rdomonoky: I was thinking about some fm preset website a while ago. Haven't done any work on that though.
21:51:50bluebroth3rpixelma: I'll try to build rbxpeexenc / rbspeexdec on mac. Though I'm wondering if it will make a difference if built with rbutil or not. Need to check that, it might be an issue.
21:55:41CIA-5New commit by Buschel (r25060): Reduce binsize again.
21:56:58 Quit Llorean (Quit: Leaving.)
21:58:40bluebroth3rpixelma: rbspeex binaries uploaded to −− get rbspeexenc-mac and rbspeexdec-mac. Those are uncompressed binaries. You need to make them executable after downloading (chmod +x rbspeex*-mac)
22:00:29 Join Zacam-Mobile [0] (
22:02:48Buschelbinsize table looks much better now
22:02:56CIA-5New commit by Buschel (r25061): Use larger IRAM for M5/X5 in vorbis.
22:03:20Zacam-MobileIf I have a sansa v2 runing r23165-091013 rockbox, do I need to do anything special to upgrade it to r25057?
22:04:55CIA-5New commit by Domonoky (r25062): always show checkwps status.
22:07:43CIA-5New commit by jdgordon (r25063): revert r25054 - tags which have an off and a list of option should have off as the first not last option. also fix it so the no icon position actually ...
22:08:28 Join freddyb [0] (
22:13:25 Join Strife1989 [0] (
22:13:58 Quit m3dlg (Ping timeout: 276 seconds)
22:14:33 Join m3dlg [0] (
22:15:14 Join Zagor [0] (
22:15:14 Quit Zagor (Changing host)
22:15:14 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
22:16:06 Nick Strife1989 is now known as Strife89 (
22:17:24 Quit domonoky (Read error: Connection reset by peer)
22:20:41 Quit m3dlg (Ping timeout: 252 seconds)
22:21:16 Join m3dlg [0] (
22:21:31 Quit stooo (Quit: Leaving.)
22:22:37 Join Zacam [0] (
22:22:40 Quit Zacam-Mobile (Ping timeout: 246 seconds)
22:23:23 Quit dfkt_ (Ping timeout: 245 seconds)
22:24:05 Join dfkt [0] (~dfkt@unaffiliated/dfkt)
22:25:14ZacamAnywho. To verify before I do this, is there anything special I need to do when upgrading from the dev r23165-091013 on a Sansa v2 to the latest r25057 or stable?
22:26:33 Quit punkt (Ping timeout: 265 seconds)
22:27:23gevaertsZacam: I don't think so. As far as I know your current bootloader should be new enough
22:28:07ZacamCool. Well, I just got my taxes back, so if it blows up, I'll just buy a new 16 to replace the 8. Should I leave the sansa firmware alone?
22:29:08 Quit bmbl (Quit: Bye!)
22:29:22gevaertsjust install a new and you'll be fine
22:30:26 Join froggyman [0] (
22:31:52 Join punkt [0] (
22:36:00 Quit m3dlg (Ping timeout: 240 seconds)
22:36:33saratogaBuschel: I was thinking about putting IRAM rules into the codeclib
22:36:55saratogabasically defining IRAM_48K, IRAM_80K, IRAM_128K etc
22:37:11saratogamaybe in apps/codecs/lib/iram.h or something similar
22:37:17saratogathen just let the codecs import that
22:37:53saratogawe still need to add some IRAM stuff for the nano2g which has 192k total IRAM (IIRC)
22:38:44CIA-5New commit by Buschel (r25064): Use more IRAM for musepack on S5L870X CPUs.
22:38:58Zacamgevaerts: Thanks for the info. 5x5 here.
22:40:19Buschelsaratoga: sounds like a good idea. we should just agree to some fixed defined sizes like you have mentioned.
22:42:30 Join Casainho [0] (
22:44:45Buschelsaratoga: wow. just checked the *.lds files -> nano2g (S5L8701) has 176 KB iram. the *.lds files for S5L8700 are weird ( 64/64, 128)
22:57:56***Saving seen data "./dancer.seen"
23:04:14 Quit JdGordon (Quit: Leaving.)
23:04:50 Join dfkt_ [0] (~dfkt@unaffiliated/dfkt)
23:05:00 Quit FlynDice_ (Remote host closed the connection)
23:06:24 Quit dfkt (Ping timeout: 248 seconds)
23:06:41saratogaBuschel: the samsung chips are weird, I think the customer can choose how much IRAM each has
23:08:13saratogai think right now we basically assume 48K for everything except the few PP and CF targets with 128k total, in which case codecs get 80KB
23:08:29saratogathough maybe a few targets still have 64KB if you haven't just changed them all
23:12:19CIA-5New commit by alle (r25065): Remove the caption and label parameters from the tagmap macro. They are never used. Also don't embed tagmaps into a table environment since we want ...
23:12:34 Join fml [0] (
23:13:31 Quit gill0r (Remote host closed the connection)
23:13:57 Quit punkt (Ping timeout: 256 seconds)
23:14:00saratogarockbox 3.6 is shaping up nicely, we've already got nearly as many major changes as all of 3.5
23:14:26 Join punkt [0] (
23:14:35 Part punkt
23:15:12Buscheland 8 / 14 major changes are codec related :o)
23:15:20 Quit dfkt_ (Ping timeout: 240 seconds)
23:15:34 Join gill0r [0] (
23:15:50 Join dfkt [0] (~dfkt@unaffiliated/dfkt)
23:15:52saratogaBuschel: are there any 64/64 IRAM targets left?
23:18:39Mode"#rockbox +o TheSeven" by ChanServ (ChanServ@services.)
23:18:46Topic"3.5.1 has been released! | Please read before speaking: | Please direct offtopic/social chat to #rockbox-community | This channel is logged at" by TheSeven (~theseven@rockbox/developer/TheSeven)
23:19:05Mode"#rockbox -o TheSeven" by TheSeven (~theseven@rockbox/developer/TheSeven)
23:20:12 Join anewuser [0] (anewuser@unaffiliated/anewuser)
23:20:37 Quit petur (Ping timeout: 240 seconds)
23:26:48Buschelsaratoga: not that I know of.
23:30:56saratogagood, even easier then
23:32:35 Quit freddyb (Remote host closed the connection)
23:32:37 Quit dys` (Ping timeout: 276 seconds)
23:32:53 Quit pamaury (Quit: Page closed)
23:34:26 Quit flydutch (Quit: /* empty */)
23:35:28*Buschel wants to bring up the clock discussion again as we now optimized so many codec to <<30 MHz
23:36:00Buschel(for the MHz speaking of arm)
23:36:16TheSevenyou mean going even below 50MHz?
23:36:45Buscheli mean using 24 MHz as default clock for PP processors
23:37:24Buschelthose CPUs can save quite some power when using lower clock
23:37:27TheSevencan a lower clock speed even compensate the reduced amount of time it will be in low power states?
23:38:22Buschelthere are at least two patches that show the positive and measurable effetc on runtime
23:39:16Buschelfs#8668 and fs#9800
23:39:39 Join planetbeing__ [0] (
23:40:45amiconnBuschel: Too slow UI especially on iPod Video
23:42:24 Quit planetbeing_ (Ping timeout: 264 seconds)
23:42:24Buschelamiconn: neverending story. use a gui boost like in fs#8668 and gui is more than fine −− even better than before. of course this will have some effect on our plugins ans well.
23:43:56Buschelamiconn: I am not convinced of throwing away runtime for keeping some gaming/demo plugins run nicely at stable speed. but of course this is _my_ opinion
23:44:17pixelmaBuschel: I'm a bit curious if those IRAM changes were tested and what effect they might have. Unfortunately I don't understand the technical part of it but would have thought that there was a reason why things were the way they were
23:44:54pixelmafor X5/M5
23:46:49Buschelpixelma: if it compiles it is fine. the 64/64 was based on the good old 50% rule. as playing music files is one the most important functionalities of rockbox it is reasonable to give the fast iram to codecs that desperately need them −− espcecially on coldfire
23:47:24amiconnThe 64/64 split on MCF5250 had a (weak) technical reason
23:47:36Buscheland that was?
23:48:00amiconnThe IRAM on both MCF5249 and 5250 is not uniform. There are two blocks, one that is DMA capable and one that's not
23:48:32amiconnOn MCF5249 the blocks are 64KB and 32KB, on MCF5250 they are 64KB and 64KB
23:48:52amiconnSo the core IRAM is entirely DMA capable, the codec/plugin IRAM is not
23:49:06Buschelcannot remember to see any comment on that fact :/
23:49:15amiconnOn MCF5249 this doesn't hold, that's why I said it's a weak reason
23:49:25Buschelnevertheless it wasn't used by core
23:50:33amiconnHmm, actually both are handled more similar now. The core IRAM is still fully DMA capable, whereas codec/plugin IRAM is DMA capable in the first 16KB only
23:50:40Buschelto be precise: it was/is used to <48KB by core
23:51:26amiconnWe once had a bug related to that: Doom had its audio buffer in IRAM - and pcm output uses dma on coldfire
23:51:50amiconnThen something was added, and the buffer moved into the non-dma capable part of iram -> pcm hung
23:52:12amiconnI don't think it's a problem - but has that change been tested before commit?
23:52:13 Join karashata [0] (
23:52:57Buschelto be honest, i did not think of such bugs...
23:53:13Buschelas you said yesterday: i need a coldfire target ;)
23:53:47Buschelyou have a x5 or m5?
23:53:58gevaertsBuschel: to be precis, you need two, a 5250 and a 5249 :)
23:54:48*kugel has a 5250 :)
23:54:53kugelit doesn't run rockbox though :'(
23:55:18 Quit fml (Quit: CGI:IRC)
23:55:55Buschelgevaerts: i have a family! i cannot have more targets... :)
23:58:05*gevaerts spots the problem :)
23:58:41pixelmaBuschel: if you have children you can declare the DAPs as presents to them ;)

Previous day | Next day