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 2012-10-25

00:01:16 Quit prof_wolfff (Ping timeout: 244 seconds)
00:04:55***Saving seen data "./dancer.seen"
00:06:55 Quit bertrik (Ping timeout: 255 seconds)
00:11:30 Join georgehank [0] (
00:17:53 Join Syconaut [0] (
00:18:39 Quit Syconaut^ (Ping timeout: 244 seconds)
00:22:46 Quit mirak (Remote host closed the connection)
00:25:01 Quit lebellium (Ping timeout: 244 seconds)
00:29:22 Join Clear_runway [0] (~Clear_run@
00:31:39 Quit nosa-j (Ping timeout: 256 seconds)
00:36:45 Quit Scromple (Read error: Connection reset by peer)
00:45:00 Join Scromple [0] (~Simon@
00:50:28 Quit ender` (Quit: You do not need a parachute to skydive. You only need a parachute to skydive twice.)
00:56:01 Join Deathrow [0] (
01:08:20 Quit KiwiCam (Remote host closed the connection)
01:14:19 Join KiwiCam [0] (~quassel@
01:21:13 Quit Clear_runway (Ping timeout: 240 seconds)
01:27:37 Join nosa-j [0] (~m00k@
01:33:11 Quit Xerion (Ping timeout: 272 seconds)
01:51:46 Quit factor (Ping timeout: 255 seconds)
01:52:32 Join factor [0] (
02:04:55 Part jkonami
02:04:56***Saving seen data "./dancer.seen"
02:14:01 Join alcides [0] (~alcides@unaffiliated/alcides)
02:14:12 Part alcides
02:17:35 Quit Deathrow (Ping timeout: 240 seconds)
02:35:07 Join eckoit [0] (
02:51:06 Join MasT3R [0] (
03:09:18 Quit amithkk (Remote host closed the connection)
03:12:42 Join tjb0607 [0] (~tjb0607@
03:15:58 Quit eckoit (Remote host closed the connection)
03:22:57 Join amithkk [0] (u4289@2buntu/writers/amithkk)
03:23:31 Quit MasT3R (Quit: CGI:IRC (EOF))
03:24:04 Join japc [0] (
03:24:41 Quit scorche (Disconnected by services)
03:24:45 Join scorche` [0] (~scorche@rockbox/administrator/scorche)
03:33:29 Join Clear_runway [0] (~Clear_run@
03:48:23 Join Totalled [0] (
04:00:01 Quit tjb0607 (Ping timeout: 255 seconds)
04:03:47 Join Xerion [0] (
04:04:59***Saving seen data "./dancer.seen"
04:05:13 Join eckoit [0] (~ryan@
04:06:01 Quit Scromple (Quit: Leaving)
04:11:03 Join Scromple [0] (~Simon@
04:13:25 Join Deathrow [0] (
04:14:59 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:15:00 Quit amiconn (Disconnected by services)
04:15:00 Quit pixelma (Disconnected by services)
04:15:00 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:15:02 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:15:03 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:24:46 Join tjb0607 [0] (~tjb0607@
04:25:16 Join TheSphinX_ [0] (
04:28:10 Join perrikwp [0] (
04:28:47 Quit TheSphinX^ (Ping timeout: 265 seconds)
04:59:35 Quit tjb0607 (Ping timeout: 240 seconds)
05:04:51 Quit Rower85 (Quit: Hmmm...)
05:05:35 Quit japc (Quit: Ex-Chat)
05:06:38 Quit kevku (Ping timeout: 272 seconds)
05:10:40 Quit nateloaf (Quit: Leaving.)
05:16:26 Quit [Saint_] (Ping timeout: 244 seconds)
05:18:04 Quit scorche` (Read error: Connection reset by peer)
05:22:22 Quit Clear_runway (Ping timeout: 255 seconds)
05:27:23 Join scorche [0] (~scorche@rockbox/administrator/scorche)
05:29:22 Join tjb0607 [0] (~tjb0607@
05:41:21 Quit shai_ (Read error: Connection reset by peer)
05:41:47 Join shai_ [0] (
05:43:03 Quit Gallomimia (Ping timeout: 252 seconds)
05:57:03 Quit [7] (Disconnected by services)
05:57:11 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
06:05:01***Saving seen data "./dancer.seen"
06:09:02 Join [Saint] [0] (~quassel@rockbox/user/saint)
06:14:29 Join brkfstsqd [0] (
06:14:50brkfstsqdhello all
06:18:58brkfstsqdi was reading this page and saw my device, does this page have to do with rockbox ports that are in development or is it just research?
06:22:19JdGordonjust research
06:56:24 Join reserve [0] (
06:58:12 Quit KiwiCam (Ping timeout: 245 seconds)
06:59:25 Quit reserve (Client Quit)
07:00:21 Join reszerve [0] (
07:04:36reszerveHas anyone noticed the Rockbox main web page Project news latest says 2012-04-12 Rockbox 3.11.2 is released?
07:15:51 Quit reszerve (Quit: CGI:IRC (EOF))
07:20:15 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
07:32:29 Quit alexbobp (Ping timeout: 264 seconds)
07:47:22 Join mortalis [0] (
07:48:28 Quit brkfstsqd (Remote host closed the connection)
07:49:50 Join alexbobp [0] (
08:05:05***Saving seen data "./dancer.seen"
08:12:10 Quit Provel (Read error: Connection reset by peer)
08:12:37 Join Provel [0] (
08:12:44 Join liar [0] (
08:22:07 Join KiwiCam [0] (~quassel@
08:24:53 Quit Provel (Ping timeout: 268 seconds)
08:40:46 Quit XavierGr (Ping timeout: 240 seconds)
08:41:28 Join ender` [0] (
08:43:18 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
08:52:57 Join Provel [0] (
09:03:36 Join einhirn [0] (
09:04:13 Quit einhirn (Client Quit)
09:08:19 Join einhirn [0] (~Miranda@2001:638:605:4:b9b9:2fc1:74a8:3bef)
09:20:49 Join petur [0] (~petur@rockbox/developer/petur)
09:23:07 Quit bertrik (Ping timeout: 272 seconds)
09:26:15 Join Zagor [0] (
09:26:15 Quit Zagor (Changing host)
09:26:15 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
09:35:53 Join pamaury [0] (
09:35:53 Quit pamaury (Changing host)
09:35:53 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:48:16 Join LinusN [0] (
09:58:29 Join wodz [0] (
10:03:00 Quit pamaury (Ping timeout: 244 seconds)
10:05:09***Saving seen data "./dancer.seen"
10:26:34 Join PathosesNorms [0] (
10:31:09PathosesNormsHi, im trying to install Rockbox on my Sansa Clip Zip but i cant find the device in Rockbox Utility
10:32:12PathosesNormsI can see it under Portable Devices in explorer but nothing in Utility
10:32:18PathosesNormsanyone know what i can do?
10:34:38gevaertsYou need to set it to MSC
10:36:15PathosesNormsI'll look in to that, thank you
10:40:03wodzamiconn: ping
10:40:12 Join pamaury [0] (
10:40:13 Quit pamaury (Changing host)
10:40:13 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
10:42:21wodzmortalis: so you state that inserting any instruction instead of DEVID &= ~(1<<31); DEVID |= (1<<31); helps?
10:42:32wodzor I read it all wrong?
10:43:02mortalisinserting more then 1 nop in commit_discard cache helps
10:43:25mortalis1 nop doesn't work
10:47:51wodzmortalis: I would say we don't know whats going on to be honest
10:51:00wodzanyway I would commit DEVID &= ~(1<<31); DEVID |= (1<<31); stuff and wait what will happen
10:51:37mortalisi don't mind
10:56:54wodzits your founding - go ahead and commit :-)
10:58:46mortalislittle later, gtg now
11:06:58 Join Rower85 [0] (
11:35:15wodzTorne: regarding elfloader size impact - we could use non standard ELFs for codecs and plugins - we don't use .strtab at all, we could retain only symbols needed for relocation and we could remove unused types of relocs (like R_ARM_V4BX)
11:35:41wodzthis will shrink codecs and plugins considerably
11:36:30wodzbasically this means we will need to use custom version of strip-like utility to process regular ELFs
11:43:57Tornedoesn't strip already know how to remove symbols not needed by relocations?
11:44:22Torneisn't that exactly what −−strip-unneeded is supposed to do?
11:44:50Tornealso strip -R .strtab should work, no?
11:59:00kugelbtw, why isn't objdump used
11:59:56 Quit fs-bluebot (Ping timeout: 244 seconds)
12:01:11 Quit bluebrother^ (Read error: Operation timed out)
12:01:14 Join fs-bluebot [0] (
12:03:04 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
12:05:10***Saving seen data "./dancer.seen"
12:19:16 Quit Xerion (Ping timeout: 264 seconds)
12:26:57 Join newt66ssa [0] (~bdnewton@
12:28:52 Quit Topy (Ping timeout: 244 seconds)
12:37:24 Join Topy44 [0] (
12:40:41 Join thomasjfox [0] (~thomasjfo@rockbox/developer/thomasjfox)
12:41:20 Part thomasjfox
12:42:51 Join GeekShad1w [0] (
12:45:09 Quit GeekShadow (Ping timeout: 272 seconds)
13:02:09 Join sciopat [0] (
13:05:05 Quit petur (Quit: *plop*)
13:06:02 Quit sciopath (Ping timeout: 255 seconds)
13:08:05 Join lebellium [0] (
13:16:44 Quit newt66ssa (Quit: Leaving)
13:22:05 Quit AlexP (Ping timeout: 244 seconds)
13:22:12 Join AlexP_ [0] (~alex@rockbox/staff/AlexP)
14:00:01 Join kevku [0] (
14:01:16 Join japc [0] (~japc@
14:05:12***Saving seen data "./dancer.seen"
14:10:35 Join dfkt [0] (dfkt@unaffiliated/dfkt)
14:18:40 Quit Barahir (Ping timeout: 264 seconds)
14:19:25 Join Barahir [0] (
14:23:18wodzstrip -R .strtab is ignored :-)
14:30:51PathosesNormsso.. right now I'm getting 'Undefined instry (all that i can see of first line) pc:30800000 sp:3 on my Sanca Clip Zip with new Rockbox
14:31:24 Join amayer [0] (
14:31:27PathosesNormsi havent done anything, i installed and disconnected the player and after a while when i was about to upload some musik that is what i found
14:31:40PathosesNormscant turn it off
14:32:55gevaertsPathosesNorms: have you tried holding the power button for at least 30 seconds?
14:33:11PathosesNormsnot 30 seconds, no
14:33:49PathosesNormsthank you
14:35:31gevaertsIf you get this sort of issue regularly, you might want to consider booting to the original firmware for USB access
14:35:35amayercan someone unspam my post:
14:35:39amayerplease and thank you
14:53:23 Quit funman (Ping timeout: 260 seconds)
15:01:53 Join Xerion [0] (
15:04:01 Join funman [0] (~fun@rockbox/developer/funman)
15:14:40 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
15:16:12 Quit wodz (Quit: Leaving)
15:38:19 Join dfkt [0] (dfkt@unaffiliated/dfkt)
15:49:13 Quit guymann (Ping timeout: 255 seconds)
15:50:41 Join guymann [0] (~c@unaffiliated/guymann)
15:57:17 Quit guymann (Ping timeout: 260 seconds)
15:58:45 Quit mortalis (Quit: Leaving)
15:59:05 Join guymann [0] (~c@unaffiliated/guymann)
15:59:46amayerI have a question. this might be for someone like TheSeven.
15:59:48amayerhow does the iPod Classic know when a button is pressed when it is off?
15:59:50amayeris that a hardware thing? Im asking because ive been thinking about the way the OF kinda of "sleeps" or "suspends" but wakes up instantly.
15:59:52amayerI was talking with [Saint] a couple of weeks(months?) ago about how one would go about implementing this feature into rockbox.
15:59:54amayerIIRC it was said that you cant shut the device down but let the RAM state intact without letting the proccesor on to catch a key press.
15:59:56amayerwould it be possable to somehow capture the keypress the same way as when the device is shut down?
16:00:31Tornethe buttosn that wake it up are connected to the power management unit
16:00:38Torneand just pressing them causes it to power on and wake
16:00:46Tornewe use exactly the same mechanism to wake from "off"
16:00:55Torneit's not anything special
16:01:11Tornethe thing we don't have is the ability to actually go to sleep in a way that we can wake up from :)
16:03:23pamaurywow, I've never a hardware security module as complicated as the one in mp200/em1
16:03:27pamaury*never seen
16:04:08Torneamayer: i don't know about the classic specifically, but the way sleep work son the PP ipods is that the state of the CPU is written out to memory somewhere in some fixed location
16:04:31Tornethen the CPU shuts off entirely, but leaves the RAM powered so the entire contents of RAM are preserved (including, say, the loaded firmware image, and the state that was saved before sleeping)
16:04:45Tornewhen you press a button the PMU takes that as a poweron signal and the CPU starts executing from the bootloader again
16:04:54Tornebut the first thing the bootloader does is look at that special location in RAM to see if it was in sleep
16:05:02Torneand if so just uses that info to jump back to what it was doing.
16:05:13Torneit's not anything like the way a PC goes to sleep, with explicit hardware support
16:05:16***Saving seen data "./dancer.seen"
16:05:24Torneother than the RAM it really is actually off inbetween
16:05:36Torneand the "wake" process is actually just "boot, but using the previous contents of RAM to make bootin gsuper fast" :)
16:06:01Torneto make rockbox do that we'd have to reverse engineer all the details of what it is that the boot rom does on poweron to detect that the device was asleep and restore state
16:06:13Torneand then make rockbox exactly copy the state-preservation process including the exact memory locations involved.
16:06:28Tornei assume that the classic works similarly, because it's pretty standard for embedded devices to do it this way :)
16:20:56 Join [Deathrow] [0] (
16:24:09 Quit Deathrow (Ping timeout: 246 seconds)
16:27:46 Join XavierGr [0] (~xavier@rockbox/staff/XavierGr)
16:28:43 Quit tjb0607 (Ping timeout: 272 seconds)
16:29:49 Join tjb0607 [0] (~tjb0607@
16:36:47 Join pretty_function [0] (~sigBART@
16:45:51 Quit pamaury (Ping timeout: 246 seconds)
16:46:35 Quit mystica555 (Read error: Connection reset by peer)
16:47:49 Join mystica555 [0] (
16:52:54 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
16:59:45 Quit pretty_function (Ping timeout: 272 seconds)
17:12:11 Join nateloaf [0] (
17:26:21 Join saratoga [0] (123e0cfc@gateway/web/freenode/ip.
17:26:26saratogawodz: were you able to get the logging code working?
17:28:55 Join wodz [0] (
17:29:15wodzsaratoga: yes
17:29:47saratogasorry i wasn't able to work on that more but my work has me really bogged down these days
17:30:09wodzyeah, reallife is annoying :-)
17:34:38 Join n1s [0] (
17:34:39 Quit n1s (Changing host)
17:34:39 Join n1s [0] (~n1s@rockbox/developer/n1s)
17:42:09 Quit Zagor (Quit: Clint excited)
17:43:06 Join pretty_function [0] (~sigBART@
17:44:19 Quit wodz (Ping timeout: 265 seconds)
17:49:44 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
17:58:06 Quit n1s (Read error: Connection timed out)
17:59:03 Join n1s_web [0] (
18:00:19 Quit tjb0607 (Ping timeout: 248 seconds)
18:03:58 Quit n1s_web (Quit: CGI:IRC (Ping timeout))
18:05:14 Join n1s_web [0] (
18:05:17***Saving seen data "./dancer.seen"
18:24:37 Quit pretty_function (Ping timeout: 272 seconds)
18:27:13 Quit FOAD (Quit: I'll be back)
18:27:24 Join FOAD [0] (~foad@
18:29:52 Join tjb0607 [0] (~tjb0607@
18:36:15 Quit nateloaf (Quit: Leaving.)
18:38:23derfn1s_web: BTW, if you didn't see in #opus, I have an updated global stack usage analysis (based on some recent changes we made):
18:38:44 Part LinusN
18:39:40 Join pretty_function [0] (~sigBART@
18:42:50 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
18:44:24n1s_webderf: nice, so since we use opus_decode the new worst case (with your other changes) is 37590?
18:44:36derfYes, I believe that is correct.
18:44:45derfI've merely done the math, I haven't tested it.
18:44:55derfThe change you need is 8ccaefd3 upstream.
18:46:45n1s_webi guessed as much
18:47:59 Join n1s [0] (
18:48:00 Quit n1s (Changing host)
18:48:00 Join n1s [0] (~n1s@rockbox/developer/n1s)
18:49:16n1sderf: i'd guess the changes to reduce the pointer indirection will reduce the code size a bit too so syncing our codec to upstream will be nice for getting the lowmem targets working
18:49:26 Quit n1s_web (Quit: CGI:IRC)
18:55:03 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
19:10:08 Join prof_wolfff [0] (
19:17:38 Quit japc (Ping timeout: 260 seconds)
19:25:42 Join wodz [0] (
19:30:46n1sok, so the glitch with the very long flac file is because the calculation of elapsed time overflows in the flac codec
19:34:26amayern1s: so its an easy fix?
19:39:10bertrikwe've had that before in several other codecs, right?
19:39:25n1samayer: i don't really see a good way except doing the calculation in 64bit
19:39:38n1sbertrik: probably
19:40:09amayerso a codec can cause the time to display wrong in the WPS?
19:40:41n1saac does it in 64 bit like i suggested so i guess that's fine
19:41:04n1samayer: yes, the codec tells the playback engine how much time has elapsed
19:41:46amayeri see
19:42:02saratoga64 bit is fine so long as its not done in a critical loop
19:42:07saratogaduring initialization doesn't matter at all
19:42:12saratogajust a few extra bytes of memory
19:42:36bertrikn1s: flac.c line 486?
19:44:03n1sbertrik: mostly line 527 but they do the same thing
19:44:05amayeris there any advantage to using mp3 over mpeg audio or vise versa when transcoding videos for use on rockbox?
19:44:08amayer(ie does one use less system resourcse?)
19:46:22saratogamp3 is mpeg audio
19:48:10bertrikit looks like someone already recognized the potential for overflow there, although insufficiently
19:50:51n1sbertrik: yep
19:51:22amayersaratoga: hmm... it gives me those as 2 different options in VLC
19:51:38saratogacheck the vlc documentation
19:51:46saratogampeg audio might just mean mp2
19:52:06amayerI just used the settings on the mpeg plugin page
19:52:46amayerit says for the plugin that more then 200kbps will skip frames.
19:52:48amayeris this old data? does it varry on different players?
19:53:09bertrikn1s: several other codecs have the same kind of formula :)
19:53:31n1sbertrik: not surprising
19:58:02amiconnwodz: pong
19:58:35amiconnwodz: So the elf loader is less than 2KB on cf? If so, this is nice - I would have expected more
20:02:05bertrikn1s: the track has to be exceptionally long to trigger that overflow, right?
20:02:22bertrikI'm calculating about 2h42m for a 44100 Hz track
20:02:46bertrikshorter for a faster sample rate I guess
20:03:07n1syep, this file is 22050Hz and it happened at 5:25:20 ish
20:04:57 Join Horscht [0] (
20:04:57 Quit Horscht (Changing host)
20:04:57 Join Horscht [0] (~Horscht@xbmc/user/horscht)
20:05:18***Saving seen data "./dancer.seen"
20:08:50 Quit akaWolf (Ping timeout: 265 seconds)
20:12:21 Quit georgehank (Quit: leaving)
20:12:59 Join georgehank [0] (
20:17:54 Quit n1s (Read error: Operation timed out)
20:30:18wodzamiconn: and <- which distance should be 255/256 to hit the bug?
20:33:04 Quit ender` (Quit: drug, n: A substance that, injected into a rat, produces a scientific paper.)
20:34:57 Join mortalis [0] (~mortalis@
20:35:03 Join ender` [0] (
20:42:04 Quit pretty_function (Ping timeout: 240 seconds)
20:47:54wodzTorne: to answer your question about symbols - I was a bit surprised but test shows that about 1/3 of all symbols in .symtab is actually used for relocation even after striping with −−strip-unneeded
20:48:46Torneso it is stripping the unneeded ones, but loads are needed?
20:49:02 Quit mortalis (Quit: Leaving)
20:49:45wodzfor n2g build for example total number of symbols for cube.rock is 78 of which 23 unique are used for relocation
20:52:31wodzdoom has total of 3838 symbols of which 1589 are used for relocation
20:52:48wodzso the symtab bloat is massive
20:52:51 Join japc [0] (
20:53:55wodzthis in turn implies .strtab size as each entry in .symtab has associated string stored there
20:54:35wodzso either strip-unneeded is buggy or I don't understand how to use it
20:55:28Tornei don't see how it oculd be buggy, it's trivial
20:55:42Tornei'm not surprised that lots of the symbols are used for relocation
20:55:51Tornein fact i wouldn't be surprised if most of them were
20:56:05Tornehow are you determining that the things left over aren't used?
20:57:22wodzI count unique symbols used in whole process of relocation
20:57:54Torneis that including relocatoin types we don't use, though?
20:57:55Tornelike v4bx?
20:58:20Tornebasically every function is going to have a v4bx reloc
20:59:01Torneanyway, you can just look in the elf, no?
20:59:15wodzwhat you mean?
20:59:24Tornei mean, dump the relocations and just see what symbols they refer to
20:59:42Torneand what?
21:00:18wodzand what to do with that knowledge?
21:00:18Tornesee if it thinks they are referred to or not.
21:00:58wodzhave you seen the pastie? thats what I did to judge if all/most symbols are used
21:01:26Tornewhat i'm suggesting is rather than rely on your code to decide i they are used, look and see what the toolchain thinks
21:01:29wodzI relocate the file and see how many unique symbols were referenced
21:02:24 Quit [Deathrow] (Ping timeout: 246 seconds)
21:02:51Torneanyway, what about throwing away strtab? does that not work?
21:03:23wodzyou cannot throw away .strtab with strip - that is simply ignored
21:04:31wodzI don't quite understand what you propose to see 'what toolchain thinks'
21:05:14Torneuse the normal tool sin the toolchain to look at whatever the smallest plugin is, and see if the set of symbols left behind after stripping what it thinks is unneeded is actually reasonable compared to what the relocations say
21:06:35wodzbtw. R_ARM_V4BX reference NULL symbol
21:07:27 Quit sciopat (Read error: Connection reset by peer)
21:07:52 Join sciopath [0] (
21:11:14wodzhmm, it seems −−strip-unneeded doesn't like to be combined with −−strip-debug
21:21:36wodzyes, striping with −−strip-unneeded only helps a bit
21:23:37kugeli suspect stripping using objdump gives the same results?
21:24:49wodzI didn't try to get rid of .strtab with objdump - lets see
21:25:54wodzeee objcopy?
21:27:16kugelyes, objcopy :)
21:27:24wodzobjcopy -R .strtab doesn't work either
21:28:14 Join nateloaf [0] (
21:31:09wodzexplicitly selecting sections with -j also outputs .strtab so this is simply not possible with standard tools
21:33:04 Join brkfstsqd [0] (
21:41:48 Nick AlexP_ is now known as AlexP (~alex@rockbox/staff/AlexP)
21:46:23kugeltime for another bug report?
21:46:57kugelbut, is the plugin file size really that much of a problem?
21:53:02wodzI don't think so personally
22:00:51 Quit dfkt (Remote host closed the connection)
22:00:58 Join dfkt [0] (dfkt@unaffiliated/dfkt)
22:01:21 Quit wodz (Quit: Leaving)
22:05:19***Saving seen data "./dancer.seen"
22:05:24 Quit y4n (Quit: Today is the perfect day for a perfect day.)
22:16:38 Quit XavierGr (Read error: Connection reset by peer)
22:16:38 Join XavierGr [0] (
22:16:38 Quit XavierGr (Changing host)
22:16:38 Join XavierGr [0] (~xavier@rockbox/staff/XavierGr)
22:16:46 Join Horschti [0] (
22:16:47 Quit Horschti (Changing host)
22:16:47 Join Horschti [0] (~Horscht@xbmc/user/horscht)
22:17:24 Join Ward [0] (
22:17:48 Nick Ward is now known as Guest37912 (
22:19:52 Quit Horscht (Ping timeout: 264 seconds)
22:20:23amayerthis is purely out of curiosity but is it possable to access the harddrive and continue playback at the same time?
22:20:25amayer(i know this feature is not implemented im just wondering if it theoreticaly could be)
22:22:06Torneyou could keep playing until the buffer ran out in theory
22:22:24Torneactually buffeirng more data while the disk is mounted over usb would be considerably more complicated
22:23:24gevaertsOn FAT, I'd say it's not possible
22:23:40Tornegevaerts: you could do it if you were.. especially brave.
22:24:11gevaertsTorne: the other side could be running a defragmenter
22:24:41Tornethey could, but they probably aren't, and even if they are if you stop caching everything your odds are still not bad :p
22:24:57Torneour FAT implementation is hardly robust enough to even pretend to do this, though :p
22:25:29gevaertsI mean, sure, you could be lucky if the other side is writing in directories you're not using, and not touching the FAT chains of the files in your playlist, but they could just as well be running a FAT to NTFS conversion tool :)
22:25:59Tornein general we could do a better job of handling filesystems being in weird conditions gracefully :p
22:26:03gevaertsIt would be a good way to test your fault tolerance, true
22:26:25gevaertsBut there's a difference between not crashing and playing audio sensibly :)
22:27:16gevaertsUse MTP if you want that, really
22:28:02Tornewell yes
22:30:03amayerthank you :)
22:37:14 Quit liar (Remote host closed the connection)
22:45:52 Quit bertrik (Ping timeout: 240 seconds)
22:50:03 Quit Horschti (Quit: Verlassend)
22:57:59TheSevenamayer, Torne: it works roughly like that on the classic as well
22:58:10TheSevenexcept that the bootrom doesn't deal with external RAM or resuming at all
22:58:26TheSeventhat's done by the second level boot loader
22:59:28TheSeventhe bootrom basically just does some SoC init, checks for the DFU case, loads the second stage bootloader from NOR, decrypts and sigchecks it, and executes it from IRAM
22:59:31 Join Deathrow [0] (
22:59:47TheSevenDRAM will get initialized by the second stage, which also handles resuming
23:00:09Torneyah, becaus eit actually has a proper bootrom
23:00:15Tornewhereas the PP ipods boot directly from NOR
23:00:26TheSevenwell we have full control about the relevant code in both cases
23:00:31TheSevenso this wouldn't be particularly hard from a reverse engineering point of view
23:00:39Tornesure, but nobody has done it :)
23:00:43TheSeventhe harder part is to make rockbox go into a sleep state
23:02:04TheSeveni.e. waiting for disk accesses and LCD transfers to finish, pausing PCM, saving the hardware state and copying IRAM
23:02:21TheSevenand the relevant code for restoring that
23:02:57TheSevenmaking emcore loader detect that case and jump into DRAM should be fairly easy
23:03:24TheSevenso it's mostly a rockbox infrastructure thing, not so much a hardware problem
23:04:12TheSeventhe samsung-based ipods don't use suspend to disk anymore btw, just suspend to ram
23:04:50 Quit Totalled (Remote host closed the connection)
23:05:00 Quit Guest37912 (Quit: Blarglarg)
23:05:08 Join Totalled [0] (
23:05:27amayerTheSeven: yeah thats what i was talking about suspend to ram.
23:05:30amayercould you also wake on usb connectiong?
23:06:10TheSevenIIRC the portalplayer ipods used suspend to ram and then suspended to disk if they were in suspend to ram for a couple of hours
23:06:23TheSeventhat has changed
23:06:25Torneonly the later ones suspended to disk
23:06:36amayerdo any ipods(rockboxed) do this already?
23:06:51TheSevenrockbox has no infrastructure for suspending (yet?)
23:07:13 Join Ward [0] (
23:07:22amayeroh ok. is this something that we could implement on all players? or is this specific to apple/players with disks?
23:07:37 Nick Ward is now known as Guest97581 (
23:08:50TheSevencertainly depends on the hardware, but there are at least several devices that would support this
23:08:56TheSevenIIUC all apple devices, and possibly some others
23:09:14TheSeventhere are two basic requirements:
23:10:14TheSeven1. the ram must be powered separately from the cpu, must support self refresh, and there must be a way to keep it powered on without noticably draining the battery
23:10:25TheSeven2. the firmware may not trash the ram contents before we take over control during boot
23:11:48amayerwould suspend to ram/disk make a great improvement on players that dont have hard disks?
23:12:01TheSevenwould possibly speed up boot by about a second
23:12:22amayeri know it would make a huge difference on players with disks
23:12:53TheSevenyeah the impact would be much more noticable on hdd based players than on flash based ones
23:13:48TheSevenwith some tricks it might be possible to do some windows 8 like boot acceleration on hdd based players though
23:14:29TheSevenit could also help to defragment the hdd in a way that makes the files contiguous in the order they are read during boot
23:14:46amayerwhat do you mean by boot acceleration?
23:14:48amayerwhat does it technically do?
23:15:20amayerTheSeven: i would like to continue this conversation later. work just let out so im headed home. ttyl
23:15:22TheSevenbasically write a ram image of a mostly booted state to the hdd and just load that during the next boots
23:15:23amayerill read the logs when i get home
23:15:33amayerthat sounds cool
23:15:41TheSeveni.e. with the theme and that kind of stuff loaded already
23:15:52TheSevenwhich is what I think causes most seeks during boot (but I might be wrong)
23:16:11 Part amayer
23:30:32 Join stripwax [0] (~Miranda@rockbox/developer/stripwax)
23:36:42 Quit stripwax (Read error: Connection reset by peer)
23:37:06 Join Clear_runway [0] (~Clear_run@

Previous day | Next day