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).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2011-10-27

00:00:04semitonessaratoga: i dd'd the entire disk to a file and md5'ed that
00:05:18***Saving seen data "./dancer.seen"
00:09:38saratogaand that matched the md5 from the sandisk firmware?
00:19:40 Quit domonoky (Read error: Connection reset by peer)
00:22:20 Quit z180 (Remote host closed the connection)
00:23:46 Quit tjb0607 (Ping timeout: 255 seconds)
00:24:01semitonessaratoga: no, because it is bigger than the firmware file
00:24:14saratogaso how were you comparing?
00:24:36 Quit bertrik (Quit: And That, My Liege, Is How We Know the Earth to Be Banana Shaped)
00:25:03semitonessaratoga: I was comparing the disk after I dd'd the OF to it, and after I turned it off and on again.
00:25:16 Quit Jerom (Quit: Leaving.)
00:25:20semitonesand again after I dd'd /dev/null and then the OF
00:25:36saratogadid you check that the disk actually contained the firmware?
00:25:51semitonessaratoga: no. I don't know how I would do that
00:26:12saratogawrite the OF and then read it back after a reboot
00:26:16saratogasee if its the same
00:26:26semitonessaratoga: I did that
00:26:29semitonesit was the same
00:26:46saratogawhat did you do?
00:27:50 Quit ender` (Quit: Our chances of being caught by the RIAA or IFPI are somewhat less than being hit by lightning - or choking on a wasabi-flavoured peanut.-- TheRegister)
00:28:21semitonesdd'd the OF to the disk. md5 the disk. reboot. md5 the disk. they were the same
00:28:47semitonesby md5 the disk, i mean I dd'd the whole disk to a file and md5'ed that
00:29:02saratogawhat i'm suggesting is that you check that what youv'e read back actually contains the file you wrote to it
00:29:11saratogafor all you know the file you're reading back is all zeros correct?
00:29:46semitonessaratoga: yes. I could dd /dev/null to the disk, and see if that returns a different checksum
00:30:07saratogawhy not just look at what you've read back and see if it matches the OF bin file?
00:30:28semitonessaratoga: using what? just open the file in less or something?
00:30:39saratoganotepad for all i care . . .
00:30:47 Quit liar (Remote host closed the connection)
00:31:08semitonessaratoga: sorry, i didn't realize you could just open it and read it
00:32:45semitonessaratoga: it's different. less shows that the sansa disk was full of these: @^
00:36:17 Join Scromple [0] (
00:36:31semitoneswhy was dd reporting success but not succeeding
00:38:23saratogaprobably because theres nothing actually connected
00:39:39semitonesso fdisk reports the disk, but I can't connect to it
00:43:20saratogajust because you give a USB device data doesn't mean its backed by storage
00:44:06 Join zchs [0] (
00:50:54 Join tjb0607 [0] (
01:01:49 Quit tjb0607 (Ping timeout: 259 seconds)
01:03:51 Quit AlexP (Remote host closed the connection)
01:11:28saratogadoes anyone have any idea why dircache seems so sensitive to file system corruption?
01:11:49saratogait seems to crash readily on disks that rockbox would happily read/write without incident
01:15:50 Quit hilbert (Quit: Computer has gone to sleep.)
01:17:05 Join hilbert [0] (
01:17:48 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
01:23:03JdGordonsaratoga: because it reads every file?
01:23:39saratogabut why does an error in the file system crash the parser? shouldn't it just find a file that has the wrong size or something similar?
01:24:28saratogai mean its just walking the allocation table right, not actually scanning the disk?
01:24:40JdGordondunno, depends on the corruption probably
01:37:56 Join DerPapst [0] (
01:40:01 Join Strife89 [0] (~Strife89@
01:40:44DerPapstjust on a side note... the multipart/mixed "server push" method works fine again in firefox version 7 and up (probably for older versions except ff4 as well). it would be nice if someone could reenable this feature for the irc reader :)
01:43:01 Quit DerPapst (Quit: CGI:IRC)
01:49:25 Quit hilbert (Quit: Computer has gone to sleep.)
01:51:17 Join mshathlonxp [0] (msh@
01:52:33saratogaJdGordon: just skimming the parser it does things like run strlen and then use the result to increment pointers, so I suspect that could end badly if the entry is corrupt
01:52:35 Quit GodEater (Ping timeout: 240 seconds)
01:53:08 Join Acorn [0] (2e4019ad@gateway/web/freenode/ip.
01:53:50AcornIs there a quick way to get back to the music playback screen without going via the main menu?
01:54:03AcornAnd is there a way to get back to that screen without resuming playback?
01:54:24saratogadepends on the target
01:56:45 Quit Stummi (Quit: Leaving...)
01:56:58Acornsansa clip+ in this case
01:58:30JdGordonmenu+select iirc
01:59:33JdGordonsaratoga: yeah, sounds likely... I know nothing about dircache and the FS layer, just seems obvious to me that dircache stresses the fs more than normal browsing might, so with some corruption it oculd kill it
02:05:19***Saving seen data "./dancer.seen"
02:05:24saratogai don't know anything about fat32, but it seems logical to me that the system should at least check that the lengths of strings and such are consistent with the maximum values allowed by the file system and if they aren't stop trying to build the cache
02:05:59saratogaunless the file system driver is already doing that somewhere
02:06:46AcornJdGordon: hmm, looks like you can't get to the playback control with a plugin running..
02:07:02JdGordonsome plugins provide it in a menu
02:07:17Acornyeah, noticed the blackjack plugin had a menu
02:07:29Acornshame you can't have a plugin running in the background
02:08:02JdGordonyou can
02:08:17JdGordononly battery bench supports that though
02:08:17Acornah, that's good news
02:08:27 Join GodEater [0] (
02:08:27 Quit GodEater (Changing host)
02:08:27 Join GodEater [0] (~bibble@rockbox/staff/GodEater)
02:09:31AcornI was going to write a simple plugin to make beeping noises at specific intervals to help with jogging, probably in LUA, and it would be nice to be able to use the proper playback control interface while it was running
02:10:11JdGordonthe LUA pluign probably doesnt support it either
02:24:25 Quit ReimuHak_ (Quit: Leaving...)
03:04:21 Quit Strife89 (Quit: Reboot.)
03:04:31saratogaFS #12350 is really weird
03:04:33fs-bluebot 3Clipv2 - Rockbox duplicates samples - effectively halves samplerate (bugs, unconfirmed)
03:05:06saratogawhen you record in stereo the left channel gets sample 1 two times, then the right channel gets sample 2 two times, and so on
03:05:14saratogalike the interleave logic got reversed
03:05:21saratogaanyone know where that happens in the recording code?
03:14:00 Join ChickeNES [0] (
03:15:40 Quit Zarggg (Quit: Rebooting client...)
03:15:56 Join Zarggg [0] (
03:17:58 Join ReimuHak_ [0] (
03:26:53 Nick scorche` is now known as scorche (~scorche@rockbox/administrator/scorche)
03:32:20 Quit ReimuHak_ (Quit: Leaving...)
03:32:52 Join ReimuHak_ [0] (
03:35:30 Quit ChickeNES (Quit: Computer has gone to sleep.)
03:53:58 Join Lars_G [0] (~Lars@unaffiliated/lars-g/x-000001)
03:54:36Lars_GHi all.
03:55:54Lars_Ghi. I think I'm better at community :D
04:05:23***Saving seen data "./dancer.seen"
04:18:16 Quit ageis (Quit: Leaving)
04:20:45 Quit [7] (Disconnected by services)
04:20:57 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:24:24 Join tjb0607 [0] (
04:24:28 Join ageis [0] (
04:27:33 Quit tjb0607 (Remote host closed the connection)
04:29:57 Join tjb0607 [0] (
04:30:15 Quit linuxstb (Ping timeout: 260 seconds)
04:32:53 Join linuxstb [0] (
04:32:53 Quit linuxstb (Changing host)
04:32:53 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
04:37:02 Quit saratoga (Quit: Page closed)
04:47:23 Quit amiconn (Disconnected by services)
04:47:24 Quit pixelma (Disconnected by services)
04:47:24 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:47:26 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:47:28 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:47:46 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:51:35 Quit sinthetek (Read error: Connection reset by peer)
04:56:05 Join sinthetek [0] (~sinthetek@unaffiliated/sinthetek)
05:04:30 Quit Lars_G (Ping timeout: 252 seconds)
05:32:16 Quit linuxstb (Read error: Operation timed out)
05:33:54 Join linuxstb [0] (
05:33:54 Quit linuxstb (Changing host)
05:33:54 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
05:34:12 Quit ps-auxw (Ping timeout: 252 seconds)
05:34:32 Quit MethoS- (Remote host closed the connection)
05:36:03 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
05:37:15 Join Rob2223 [0] (
05:41:02 Quit Rob2222 (Ping timeout: 255 seconds)
05:44:11 Quit Horschti (Quit: Verlassend)
06:05:24***Saving seen data "./dancer.seen"
06:06:53 Quit Acorn (Ping timeout: 265 seconds)
06:26:25 Quit tjb0607 (Quit: kthxbai.)
06:34:19 Quit linuxstb (Ping timeout: 256 seconds)
06:35:26 Quit robin0800 (Quit: Leaving)
06:35:33 Join linuxstb [0] (
06:35:34 Quit linuxstb (Changing host)
06:35:34 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
07:21:28 Quit Scromple (Quit: Leaving)
07:25:27 Join Scromple [0] (
07:58:45 Join Prodicus [0] (
08:03:55 Join Stummi [0] (~Stummi@rockbox/developer/Stummi)
08:04:14 Join Kuitsi [0] (
08:04:22ProdicusNew to Rockbox here- what communication channels are best for feature requests and discussing them? Forums, user list, dev list, tracker?
08:05:11 Quit Stummi (Client Quit)
08:05:27***Saving seen data "./dancer.seen"
08:05:38[Saint_]I'm not sure we really *do* feature requests...
08:05:52[Saint_]We have "feature ideas", that's it.
08:06:10[Saint_]a request kinda implies that someone will do something about it :)
08:06:32ProdicusWell I didn't say feature demands :)
08:06:50JdGordonthe forum is probably best for feature ideas
08:06:51JdGordonor here
08:07:40JdGordonbut yeah, very few feature ideas make it anywhere unless someone comes up with a patch from the start
08:07:54ProdicusCouple thoughts then- first off, has anybody looked into doing CELT/Opus? You already do speex, CELT/Opus has a fixed-point decoder, and it outperforms speex at every bitrate
08:08:58ProdicusI don't know how much work they've done on ARM optimization yet though, probably not much
08:09:33KuitsiI did browse my music via "Database -> Year" and in some cases it shows pretty strange years, like 204 or 3107. why? in amarok those tags seems to be correct. i'm using that "unusable" ipod classic port which otherwise seems to be working fine exept much shorter battery life
08:09:41JdGordonespecially with audio codecs.. if the license is acceptable and there is a patch, it will probably be accepted
08:09:55JdGordonI dont tihnk a codec has ever been rejected because it wasnt wanted
08:10:02JdGordonwe *even* have wma :p
08:13:42 Join nosa [0] (
08:14:43ProdicusOk; good to know. CELT/Opus is a collaboration between Xiph (esp. jmvalin, the guy behind speex) and Skype, and its license is a BSD variant. Should be accepted as an IETF standard pretty soon here.
08:15:11 Quit nosa-j (Ping timeout: 240 seconds)
08:15:11 Nick nosa is now known as nosa-j (
08:16:24 Join fyre^OS [0] (
08:17:10 Quit fyrestorm (Read error: Connection reset by peer)
08:17:58ProdicusSecond- I did a the default/recommended "complete" install on a Sansa Clip+, and after finding that Doom took up as much space as everything else combined, I removed it. Tossing the other extras into a default install comes at very little cost but I wonder whether you might want to rethink putting Doom in the defaults.
08:18:44 Join GodEater_ [0] (93722cd0@rockbox/staff/GodEater)
08:19:25 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
08:21:24ProdicusThird, I noticed that the etiquette page linked from your mailing list page is completely empty. While I suppose that could be construed to mean something about expected etiquette I doubt it's the message you mean to send :)
08:24:54 Quit fyre^OS (Read error: Connection reset by peer)
08:34:23 Quit factor (Ping timeout: 260 seconds)
08:37:29JdGordongrr... there is no way to know if the user left the filebrwoser with the menu button or by backing out :/ both cases return GO_TO_ROOT :(
08:38:34 Join factor [0] (~factor@
08:46:11 Join einhirn [0] (
08:55:49 Join wodz [0] (
08:58:02wodzyesterday I hit rather odd behaviour: from file browser select file for playback -> change repeat mode to one -> play a few loops -> change repeat mode to off -> and skip (short ff) do nothing
08:58:39ProdicusPerhaps sometime soon I may find time to set up a build environment and see whether I can manage to get opus to build under the cross compiler. TTFN.
08:59:07 Quit Prodicus (Quit: CGI:IRC 0.5.9 (2006/06/06))
08:59:26 Join nick-p [0] (
09:00:57 Join ender` [0] (
09:05:53 Join hilbert [0] (
09:21:20 Join mortalis [0] (~mortalis@
09:21:53 Quit Scromple (Quit: Leaving)
09:23:35kugelmight do a call for translators as we're still frozen?
09:23:54Zagorgood idea
09:26:15 Quit mshathlonxp ()
09:26:25 Join Kiwi_CAM [0] (
09:34:03GodEater_whats actually at the root of your questions KiwiCam ?
09:34:10GodEater_whats actually at the root of your questions Kiwi_CAM ? :)
09:34:18*GodEater_ gets the right nick finally
09:38:42 Join LinusN [0] (
09:40:06Kiwi_CAMGodEater: That most iPod users won't have even heard of Rockbox. I can understand that, but it's a shame when you consider the potential number of RB users out there.
09:43:05 Join bluefoxx [0] (
09:43:30wodzKiwi_CAM: most iPod users are special kind of people not really interested in anything not Apple branded
09:43:51 Quit wodz (Quit: Leaving)
09:47:43Kiwi_CAMSo, the cool kids come here? ;)
09:49:35Zagorthe kids who want more than came in the box, yes
09:50:03Zagorif that is "cool" or "stupid" depends on your perspective
09:50:42Kiwi_CAMI was fantasizing about spreading the RB word, by the use of stickers. Sort of like the ones you find of your laptops. Mine say Mint. They would be wasted on iPod users though, unless somehow, you could make it a cool thing to do.
09:51:41Kiwi_CAMI suppose, for your guys, it's just doing this, is what's important.Not the number of RB users.
09:52:12Zagorpretty much, yes. we have no reason to convince people to use rockbox.
09:52:41Kiwi_CAMNah. Stickers wouldn't work. I could never get one on my Clip ;)
09:52:53Zagorwe do want users, because a percentage of users are potential developers
09:53:25Kiwi_CAMMmm. The developers would know where you were.
09:53:53GodEater_I think we're victims of our own success in a lot of ways
09:54:07GodEater_I know a reasonable number of people here at work who are ex-rockbox users
09:54:08Kiwi_CAMHow so?
09:54:13JdGordonKiwi_CAM: you mean like ?
09:54:13GodEater_who are surprised to find the project is still going
09:54:18Kiwi_CAMYou can brag. I don't mind.
09:54:37GodEater_they came here "back in the day", installed rockbox, loved it, but never came back
09:54:44GodEater_it already did so much for them they never bothered with upgrades
09:54:57Kiwi_CAMJdGordon: I got a 404 there.
09:55:06GodEater_and then when they moved on to new devices, they figured the thing was so old now, there was no point going back to see if there were ports for their new toy
09:55:19GodEater_a lot of them were absolutely staggered to discover the android port
09:55:34JdGordonok, i fail at g+
09:59:23Kiwi_CAMThat's the beauty of Rockbox. It evolves. You can always come back. There are always new players. But, I can see why you're doing Android. I have a droid, but I'm not giving up my MP3 Player.
10:01:52Kiwi_CAMWTF is that?!? Scale's hard to tell. A subwoofer? You could make a small sticker with a capital RB (grey B). "I've been Rockboxed" or Rockbox is Me!"
10:03:01JdGordonthats my laptop
10:05:28***Saving seen data "./dancer.seen"
10:06:05Kiwi_CAMI should have known by the bulge ;) Too much heat isn't good for the family jewels mate. Where can one get a Rockbox sticker then? Pretty please?
10:19:24*Kiwi_CAM sulks in the RB non community corner.
10:23:58 Join pamaury [0] (~quassel@
10:23:59 Quit pamaury (Changing host)
10:23:59 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
10:32:52 Quit linuxstb (Ping timeout: 276 seconds)
10:34:40*[Saint_] wonders if a: its possible to, and b: if there's ay reason not to allow RaaA to rotate 180 degrees on devices that can support it.
10:35:17[Saint_]I end up using my phone upside down quite a bit.
10:35:21Zagorit's not possible today, since they are hardcoded to a specific resolution and orientation
10:35:28Zagoroh, 180
10:35:40Zagorthat should be perfectly doable
10:36:51Zagorwe already have an upside down setting for other targets
10:37:18[Saint_]Not all devices support app rotation at 180 degrees, but I know the CM ROMs defintely do, for apps that allow it.
10:37:33[Saint_]And it could also be an actual on/off/sensor setting presumably.
10:40:05 Join linuxstb [0] (
10:40:05 Quit linuxstb (Changing host)
10:40:05 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
10:50:22 Quit linuxstb (Ping timeout: 245 seconds)
10:54:14 Quit bluefoxx (Quit: bluefoxx)
11:04:00 Join bluefoxx [0] (
11:06:00 Join linuxstb [0] (
11:06:00 Quit linuxstb (Changing host)
11:06:00 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
11:15:22pixelmaKiwi_CAM: scorche and bluebrother made some stickers but I don't know if any are left
11:19:39kugelpixelma: I pinged because I wanted to ask you about hte german transaltion on the tracker
11:22:25kugelmc2739: ping
11:24:22 Quit scorche (Ping timeout: 252 seconds)
11:26:55 Quit Kiwi_CAM (Quit: Leaving)
11:32:42 Join scorche [0] (~scorche@rockbox/administrator/scorche)
11:35:59pamauryjhMikeS: ping
11:41:43 Join Topy [0] (
11:42:34 Quit T44 (Read error: Connection reset by peer)
11:49:03pamauryI wonder one again why *does the fucking hell* kernel.c contains the documentation of the kernel function but NOT the header kernel.h, this makes no sense !
11:50:07 Join dfkt [0] (~dfkt@unaffiliated/dfkt)
11:51:14Zagorprobably because they were written while writing the code
11:51:35 Quit bluebrother (Ping timeout: 240 seconds)
11:51:59 Quit fs-bluebot (Ping timeout: 240 seconds)
11:53:39 Join fs-bluebot [0] (
11:53:41 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
11:56:36 Quit hilbert (Ping timeout: 258 seconds)
11:59:28 Quit linuxstb (Ping timeout: 258 seconds)
11:59:58 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
12:05:31***Saving seen data "./dancer.seen"
12:05:40pamauryyes but this is weird
12:07:34Torne"Topit" is a forum spammer who has gotten past the filter
12:07:39Tornebumped a load of threads with obvious junk
12:07:59Tornealso Tonglimit
12:08:00 Quit linuxstb (Ping timeout: 252 seconds)
12:14:44 Join hilbert [0] (
12:18:22 Quit mortalis (Ping timeout: 248 seconds)
12:20:18 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
12:20:22pamaurythat's really strange, on my fuze+ the usb thread is created but does not appear in the os stack list :-/
12:20:33 Join mortalis [0] (~mortalis@
12:26:05 Join lorenzo92 [0] (
12:28:10pamauryapparently something is calling usb_close
12:30:28pamaurylol, there is a buggy switch fallback in usb.c !!
12:30:40pamauryreally really weird
12:31:24 Join MethoS- [0] (~clemens@
12:31:52 Quit linuxstb (Ping timeout: 258 seconds)
12:35:47CIA-14New commit by 03pamaury (r30843): usb: add forgotten break in usb switch, this could lead to unwanted code being execute like USB_QUIT to be sent on hotswap !
12:38:01CIA-14r30843 build result: All green
12:38:20CIA-14New commit by 03pamaury (r30844): fuzeplus: rework button handling to use a queue instead of a blocking semaphore in the thread
12:39:42GodEater_I thought we were in feature freeze?
12:39:57CIA-14r30844 build result: All green
12:39:57GodEater_oh, sorry - new port
12:39:59GodEater_ignore me
12:44:01pamauryand bugfix
12:44:15pamauryr30843 is clearly a bug
12:44:38 Quit nick-p (Quit: Leaving)
12:49:09 Join linuxstb [0] (
12:49:09 Quit linuxstb (Changing host)
12:49:09 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
12:57:56 Quit linuxstb (Ping timeout: 258 seconds)
12:58:18 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
13:08:24kugelpamaury: well, we have no established policy regarding that, so everything is correct (unfortunately including no documentation at all) :(
13:09:06pamauryyes I know, it's just that it's not the first time I look at kernel.h, grumbles because there is no, so I go to kernel.c to read it and bam ! there is doc :-/
13:09:10kugelI think documentation in headers has one major flaw. implementation and its documention are separated, which leads to outdated documenation
13:09:30pamaurybut in this case, the doc is not even complete
13:12:59 Quit Bagder (Ping timeout: 240 seconds)
13:13:46Zagorpamaury: exactly. it's not doc as much as a few descriptive code comments.
13:13:48 Quit hilbert (Quit: Textual IRC Client:
13:16:13kugellorenzo92: let's talk here
13:19:22lorenzo92so r0 application manages everything
13:19:28 Join HaimN [0] (~HaimN@
13:19:32lorenzo92the kernel is not supposed to be used with other apss
13:19:56kugelI can believe that
13:19:57lorenzo92that's why we need to manage everything in rockbox too
13:20:32kugelit seems strange they don't use a sensible cpu governor
13:20:57lorenzo92eh there are a lot of strange things in their me hehe. Just ask lebellium too :)
13:21:33lorenzo92even keypad for example
13:21:49lorenzo92they didn't implement it as a normal input device in!...they used a module
13:22:01lorenzo92polling everytime /dev/r0Btn
13:22:27lorenzo92mgue has tried to build a kernel module to implement it as keyboard but he failed...
13:24:07kugelI cant find sys/platform/cpu/.../cur_cpu_freq on my system
13:24:16lorenzo92wait I give you the exact path
13:24:44lorenzo92anyway here the doc too:
13:25:29lorenzo92hit ls on there and see...
13:25:52kugelI looked there :)
13:26:02Zagorwhat are you discussing?
13:26:12lorenzo92YP-R0 port
13:26:36Zagordoes it run linux?
13:26:44lorenzo92in 2 words: the player runs a linux kernel, so I'm porting rb as an application, but we need to control HW directly like in a normal rb port
13:27:01Zagoryay, more hybrids :-)
13:27:06lorenzo92yes :D
13:27:19kugellorenzo92: I found scaling_cur_freq there
13:27:31kugelthat never gives the lowest frequency for me
13:27:43kugelprobably because reading it means load
13:27:48lorenzo92eh it depends on config and acrive governor
13:27:59kugelondemand here
13:28:19lorenzo92uhm ok...well so it depends on the cpu type
13:28:35lorenzo92for example here on r0 we have 200-400-532 MHz only
13:28:36kugeldont think so
13:28:55kugelI see various higher frequencies but never the lowest one
13:29:06 Quit linuxstb (Ping timeout: 256 seconds)
13:29:18kugelI'm thinking reading that file isn't the best measure to read the current frequency
13:29:20lorenzo92uhm try to cat scaling_min_freq or something similar
13:29:27lorenzo92yes it is
13:29:52lorenzo92at least on my player...
13:29:54kugelscaling_min_freq gives the lowest
13:30:12kugelthat gives 532 on the r0?
13:30:28lorenzo92200 mhz lowest, 532 max
13:30:41lorenzo92I've seen the kernel sources too, that's right
13:31:09lorenzo92wait I turn on rockbox to see values with text viewer
13:31:26kugelthe text viewer doesnt refresh I think
13:31:42lorenzo92yes indeed...everytime i need to do manually :)
13:32:11kugelI'm questioning that it runs at 532 all the time, and that reading that file is a good way to find out
13:32:50lorenzo92yes it is. trust me hehe. it runs everytime at 532 by default. Samsung developers are not so clever indeed
13:33:08kugelwhat does scaling_min_freq give you?
13:33:18kugeland which governor?
13:33:26lorenzo92by default userspace
13:33:37lorenzo92now I setted it to conservative to test
13:33:43lorenzo92and freq is scaled down to 200 mzh
13:33:55lorenzo92(i can see because rb is a bit laggy)
13:34:03lorenzo92brickmania for example
13:34:05kugelah okay. if userspace is the governor than it is believable
13:34:41lorenzo92oh I just discovered we have stats too
13:34:45 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
13:34:54kugelI personally think ondemand is the best; but yea, if userspace is the default governor we should do something about it
13:35:37lorenzo92I can set it to ondemand
13:35:41lorenzo92before loading rb...
13:36:02kugelwith something could mean to set ondemand :)
13:37:54kugelZagor: looks like we need to do power management, backlight, sound output (i.e. not using alsa/sdl/whatever) in rockbox
13:38:07kugelso not quite an normal app port
13:38:35*kugel almost doesnt dare to ask
13:38:41kugellorenzo92: what about usb?
13:40:20lorenzo92btw sound is ok throug sdl (i'm listening fine)'s volume that we need it to be hw (as3543 codec)
13:41:59Zagoris this really going to be any easier than a native port?
13:42:46HaimN Hi, Can I interrupt you for a minute? I can't download rockbox for android from"> it's always stop at the middle, there is another way to download it?
13:52:00pixelmakugel: I may have some time today in the evening
13:52:41ZagorHaimN: it works fine for me
13:53:45 Quit ReimuHak_ (Quit: Leaving...)
13:55:48lorenzo92native port would be quite impossible: bootloader is not possible to be flashed without a samsung tool that's no public
13:56:42lorenzo92but sound is working and backlight too (I've already found the ioctl code to manage ascodec registers through afe.ko module)
13:57:06lorenzo92we need "just" to understand how to implement ascodec full management for hw audio (to keep sdl volume at 100%)
13:58:32lorenzo92...and for other goodies such as rtc alarm + time management
13:59:04lorenzo92ah yeah another issue...time isn't managed by the kernel app saves a value in /dev/stl1, like a time diff
13:59:14lorenzo92I now go to understand it
14:00:46lorenzo92for example, in the last dump I have this value
14:01:04lorenzo9288293553273...maybe they are milliseconds or what else uhm uhm
14:01:52 Join Thra11 [0] (~thrall@
14:03:19mc2739kugel: pong
14:03:27HaimNZagor: strange, Is there a site that I can upload the file directly from the Internet and then download it? Maybe this will solve my problem
14:03:31lorenzo92kugel: usb is easy to make it work, they are usb gadgets. The fact I cannot understand is that sharing partitions with usb in a script works in the original kernel, but not in the recompiled one
14:03:44kugelmc2739: I posted a patch to the usb screen corruption task
14:04:13Zagorlorenzo92: can you load a custom kernel?
14:04:25lorenzo92via kexec you mean?
14:04:31lorenzo92ah flashing it? yes...
14:05:10mc2739I saw it. I will try to test, but am leaving for work shortly
14:05:16Zagorso what stops you doing a native port again? :-)
14:05:32***Saving seen data "./dancer.seen"
14:05:53lorenzo92knowledge hehe
14:06:06 Quit bluefoxx (Ping timeout: 248 seconds)
14:06:22lorenzo92I'm still on high-level things :D
14:07:09pamauryZoiah: did you think about the decryption/encryption method I told you about in my mail ?
14:07:15 Join bluefoxx [0] (
14:10:07lorenzo92zagor: and of course to flash something you need a special cable + special software if you want to be sure...
14:10:07Zoiahpamaury: we're missing each nicely every day. ;)
14:10:07Zoiahpamaury: you mean in your mail from 25-10-2011 20:39?
14:10:07pamauryI think so yes
14:10:07pamauryyes, this week I'm in holydays in a foreign conntry so I'm not online quite often, except today
14:10:07ZoiahI responded with 'I'm totally up for it. :)', etc. right?
14:10:10 Quit LinusN (Quit: disconnecting from stoned server.)
14:10:27 Join LinusN [0] (
14:10:34pamaurystrange, I didn't get answer, let me check
14:11:14pamauryoh strange, I got your answer but gmail never told me so :(
14:12:09pamauryok, then I'll try to do that, modify my sbtools to work with a device to encrypt/decrypt, that will be fun :)
14:12:27pamauryI'm just curious: do you plan to port rockbox to it if it works ?
14:13:37 Quit HaimN (Remote host closed the connection)
14:14:26ZoiahI could give it a try.
14:14:46ZoiahI'm not very skilled at this stuff, but is an example of some other (non-finished) work I did.
14:15:53pamauryI already wrote much of the code for the imx233/stmp3780 so you "just" need to reverse engineer the missing part: LCD init is the first major step. Actually, it would be awesome it you could put photos of the inside of your device on the wiki, if you can open it
14:16:21pamauryso that we know what is the hardware
14:16:30ZoiahBut encryption, let alone OTP, fuse, chip, etc. stuff is way out of my league. I think I kinda understand what you're getting at, but I don't think I would be able to think of it, or reproduce it. :)
14:17:03ZoiahI could try cracking it open, I don't think there is a teardown on the net.
14:17:55pamauryit doesn't really matter, I know it because I spent hours reversing the ROM, reading the manual and reversing the fuze+ firmware :) Yes, with nice picures of it.
14:17:59lorenzo92so for my R0 the things to be discussed are: how to implement the ascodec management in a clean way?
14:18:46lorenzo92I've already prepared basic things for reading/writing registers. And this is working like a charm for the backligh
14:18:47pamauryI'll got out for an hour, to put things on the wiki you need to register and then ask for write access I think, just ask on the channel
14:19:23ZoiahAny idea why the fuze+ wasn't encryped?
14:19:32ZoiahOr encrypted with zeros AES key, actually, right?
14:19:36pamauryis was, but with default key (0) :)
14:19:45pamauryjust lazyness I guess
14:20:10pamauryor perhaps it required something special during manufacturing process ?
14:20:12ZoiahDidn't Sandisk support Rockbox in the past?
14:20:55pamaurythey kind of supported us without support :) They offered us to send free fuze+ but not the documentation so they might want to see a Rockbox port so that people buy more of they devices :D
14:20:58ZoiahMaybe Sandisk supports freedom. ;)
14:21:31pamauryActually I won't complain, I see this as an opportunity :)
14:21:41ZoiahI understand why the OTP bits don't want to be read, but I don't understand how you want to try to make the thing encrypt a binary.
14:21:57ZoiahEmbedded hardware is cool though, it's always been a random interest of me.
14:22:25pamauryThat's the nice point: AES is so strong that even if you have a black box that encrypt/decrypt, it will take you years to find the key
14:22:56ZoiahYup, there are some attacks that make the searchable keyspace less, but still more than the heatdeath of our universe. ;)
14:22:57pamauryso you write the key somewhere, lock it so noone can read it except the hardware and then you use the hardware to encrypt/decrypt (which is the point of AES)
14:23:06ZoiahHow come you can run your code on it?
14:24:24ZoiahAlso,I thought you needed to go? ;)
14:25:18pamaurybecause the ROM is buggy. The encrypted with two keys, knowing one key is sufficient to encrypt/decrypt, except that if you do this, you have to be able to produce a file with a known SHA-1 cipher which is at least as hard as breaking AES I think. Here is the point: the SHA-1 is a control sum of the *whole* image the ROM doesn't check it :D
14:25:34pamauryyes, I'll go, I'll give you the details later if you want
14:25:41ZoiahYes, I'm interested in this. :)
14:29:11 Quit lorenzo92 (Ping timeout: 240 seconds)
14:30:16 Quit zchs (Ping timeout: 252 seconds)
14:36:47 Quit GeekShad1w (Read error: Connection reset by peer)
14:38:03 Join chkktri_ [0] (
14:39:22 Quit chkktri (Ping timeout: 256 seconds)
14:40:31 Join GeekShadow [0] (
14:43:06 Join lorenzo92 [0] (
14:43:21 Join zchs [0] (
14:50:51 Join tails___ [0] (
14:50:51 Quit chkktri_ (Read error: Connection reset by peer)
14:52:32lorenzo92okay the governor settings are quite fine now
14:52:45lorenzo92just need to lower the frequency of control
14:55:59 Quit tails___ (Read error: Connection reset by peer)
14:56:06 Join chkktri_ [0] (
14:58:37lorenzo92anyone knows if imx37 display can be switched off? I mean, not only backlight...because I'm guessing that
15:00:10 Quit Farthen (Excess Flood)
15:00:26 Join Farthen [0] (~Farthen@2a01:4f8:101:2a4:0:bc28:b2e1:9)
15:00:59 Quit antil33t (Read error: Connection reset by peer)
15:01:22 Join antil33t [0] (
15:07:03JdGordonZagor: the restarting playback of the current track *does* still for sure happen in svn
15:07:06JdGordonthe crashes dont though
15:07:16Zagorargh, ok
15:07:17JdGordonso not sure if that should be reopened
15:07:31JdGordonprobably a new task made to more explain it
15:08:48ZagorI think the task is fine. I'll just downgrade it.
15:09:21 Quit Farthen (Excess Flood)
15:10:56 Join Farthen [0] (~Farthen@2a01:4f8:101:2a4:0:bc28:b2e1:9)
15:13:30lorenzo92okay to put framebuffer in sleep mode this should do the job:
15:13:31lorenzo92 # echo 1 > /sys/class/graphics/fb0/blank
15:14:59 Quit nosa-j (Ping timeout: 240 seconds)
15:15:35 Quit zchs (Read error: Operation timed out)
15:16:17 Quit linuxstb (Read error: Connection timed out)
15:17:53 Join nosa-j [0] (
15:22:59lorenzo92I go now see you all ;)
15:23:03 Quit lorenzo92 (Quit: ChatZilla 0.9.87 [Firefox 7.0.1/20111008085652])
15:27:32 Join matze` [0] (
15:28:14 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
15:32:22 Quit antil33t (Read error: Connection reset by peer)
15:32:48 Join antil33t [0] (
15:38:14 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
15:40:43 Join WalkGood [0] (~4@unaffiliated/walkgood)
15:41:32pamauryZoiah: still here ?
15:47:19 Join TheLemonMan [0] (
15:48:19 Quit GodEater_ (Ping timeout: 265 seconds)
15:50:41 Join zchs [0] (
15:50:57Zoiahpamaury: yeah. :)
15:51:04Zoiahpamaury: but kinda busy during the day usually.
15:53:47pamauryOk, so here is basically how it works. The SB file format has a small weirdness: it doesn't use the OTP key to encrypt the file. Instead it uses a random key. So that it does is the following: generate a random key, encrypt it using the OTP key, write the encrypted key in the image and then encrypt everything with the real random key. So far so good except that the SB format has support for *several* OTP keys and here is the flaw: if you have N
15:53:47pamaury keys, you encrypt the *same random key* with the N keys and generate N encrypted form of the key. The consequence is that if you know any of the key, you can recover the real key
15:55:01Torneyou can? it's doing XOR?
15:55:01 Join GodEater_ [0] (93722cc8@rockbox/staff/GodEater)
15:55:03Torneor similar?
15:56:36TorneOh, wait
15:56:40TorneSorry, yes, I see what you mean
15:57:21Torneyou can recover the "session" key, and thus encrypt other files and just copypaste the encrypted bit we don't have the key for
15:57:35pamauryThat's idea, details are coming :)
15:58:54pamauryYou can call it a feature but this is actually a flaw because of the following fact. Assume now that you have an encrypted image with two keys: K1 and Z, you don't know K1 but you know Z (before it's 0 for example). Then with Z you decrypt the real key and then use it to encrypt your data. Then at the end you don't use the header you should have constructed but the one of the image you've got. Of course the header doesn't fully match the file
15:58:54pamaurycontent because it doesn't have the same length and same section but here comes the second, fatal flaw of the SB format. Is was designed to be streamed and not access randomly, which means that even though the mastee header tells you everything about the file, most of information is duplicated later for the ROM to stream everything and avoid keeping information. So once the header has been read and validated, it's forgets everything and just
15:58:54pamauryapply the commands as long as there are properly encrypted :D
16:00:48pamauryUsing this technique, you can craft a file that get data through USB, use the hardware to encrypt/decrypt using the OTP key K1 (that you don't know and will never know) and then give it back through USB.
16:01:32pamauryThen you're done because you can encrypt any file you want
16:03:49 Join GeekShad1w [0] (
16:04:09 Quit GeekShadow (Remote host closed the connection)
16:04:14pamauryThat's theory, now the last step needs to be done: write the magic encrypter/decrypter. But the inital mistake was to release a file with two keys: the OTP and zero.
16:05:35***Saving seen data "./dancer.seen"
16:09:59Zoiahpamaury: I'll save this text for future reading, it looks like I should understand it, but it's not hitting me yet. ;)
16:10:16ZoiahAnyhow, I understand how you can run your own code now.
16:10:37ZoiahAny code can encrypt stuff?
16:12:20ZoiahAs a feature?
16:12:49pamaurywhat do you mean ?
16:13:10ZoiahDo I understand correctly that the flaw you explained allows you to run your own code on ARM core?
16:14:01pamauryyes, you can send a file and the ROM will think it's valid and run it
16:14:22ZoiahIn theory, you could always use this to bootstrap rockbox?
16:14:49pamauryexcept there is a little problem
16:15:21pamauryyou need to reverse engineer the OF to be able to port rockbox to this target (thing like LCD for example).
16:15:33pamauryand to reverse engineer the OF you need to decrypt it :)
16:16:35pamauryThat's why I want to write the tool that decrypt/encrypt, I'll start it in a minute, it *shouldn't* be too long to get it work
16:16:44ZoiahBut the LCD is probably some standard component? So opening up and identifying stuff is probably a step in the right direction?
16:17:09 Join hilbert [0] (
16:17:21Torneidentifying stuff is useful, but pretty much every target has a number of things that either don't have public docs, or are sufficiently complicated that things like "how exactly is it wired up" are significant
16:17:32Torneso reverse engineering the OF is pretty much always also required
16:18:06ZoiahBut it probably helps that the i.mx233 has a LCD controller integrated, so at least this part is standard?
16:18:19pamaurythis is partly standard yes
16:18:39pamaurybut things like backlight, chip enable might not be and even then you usually can't tell what the lcd model s
16:18:47ZoiahBut there seems to be a few different ways to talk to the LCDs.
16:19:03pamaurytalking is one thing
16:19:06TorneLCD panels have a nasty habit of the panel/controller mdoel numbers and so on facing inwards :p
16:19:15Tornesuch that you can only raed them if you pry the LCD module apart
16:19:24Tornewhich is usually destructive due to thin ribbon cables
16:20:35pamauryand lcd is not the only problem
16:20:43ZoiahIt's a 2.4", 320x240, how many of those could there be. ;)
16:21:11 Join y4n [0] (y4n@unaffiliated/y4ndexx)
16:21:44pamaurythings like buttons, chip enable, radio enable, ... are difficult to find out without the schematics. And if by chance it uses a raw nand with a FTL, you need to figure out the FTL as well :-/
16:22:29pamauryI think there are a number of them ?
16:22:37Tornethere are quite a lot
16:22:50Torneand people often invent their own
16:22:52Tornebecause they are idiots
16:23:25GodEater_pamaury: which we've never successfully done with any great degree of reliability have we?
16:23:28pamauryif they use the standard i80 protocol, you can read register 0 which might help you figuring out the model
16:23:32GodEater_(re: FTL layer)
16:23:41pamauryGodEater_: don't think so, indeed :)
16:24:09pamauryit's just too complicated to understand without the doc
16:24:20GodEater_black and evil magic indeed
16:24:37Zoiahpamaury: how did you end up disassembling the bootrom of the Sansa?
16:24:48Torneone day if you all are very, very nice to me i might try reversing one of the FTLs ;P
16:24:58GodEater_how nice?
16:25:13ZoiahWere you able to find a copy? Or did you dump it from flash with a in-system-header-thing?
16:25:18GodEater_are we talking a bottle of champagne, or a weekend in Vegas nice?
16:25:51pamauryZoiah: I dump it because it's at a know address
16:25:55TorneI have quite a list of things to be reversing :p
16:26:06Tornei've been fiddling with ipodvideo diagmode on and off lately
16:26:14Tornei should probably actually have a go at implementing positive charger detection
16:26:19TorneI think it's only going to work on ipodvideo, though
16:26:27Torneit looks like none of the other models do it the same way :/
16:26:43Torneit's pretty trivial there, though
16:26:44Zoiahpamaury: so you were already able to run your own code before you started disassembling this?
16:28:34kugelGodEater_: the nano2g has ftl
16:28:34pamauryZoiah: yes :) Because even before I disassembled the Freescale tool which build SB file so I was able to figure out the SB format before Freescale put it online and then I was to decrypt the fuze+ OF and that was enough information. I only dumped the ROM because I needed to undersand a precise detail (which I did) but now it's seems rather pointless because Freescale released some doc
16:28:56kugelreversed from the apple os, and it works well afaik
16:29:07pamauryThe only useful knownledge was that the ROM load the file in streaming mode and doesn't check SHA-1 sum at the end
16:29:24Zoiahpamaury: ahh, ok, that makes sense.
16:29:31Zoiahpamaury: IDA?
16:29:42pamaurykugel: doesn't the apple os erase itself sometimes and ask for recover ?
16:29:44pamauryZoiah: yes
16:29:56kugeldont know
16:30:08kugelthe nano2g port is stable, though
16:31:10pamaurywow, just noticed in the imx manual that they derived their AES hardware from an opencores implementation
16:33:16 Quit Thra11 (Ping timeout: 260 seconds)
16:33:22ZoiahFreescale published most of the original Sigmatel docs?
16:33:59 Quit antil33t (Read error: Connection reset by peer)
16:34:22 Join antil33t [0] (
16:34:47Zoiahpamaury: I'll be at home in 1-1.5 hours or so, and then I'll quickly try to pop it open before I have to leave again. ;)
16:35:06ZoiahThere don't seem to be any visible screws on the outside, so this should be fun.
16:36:25pamauryA knife is a good tool to open players :)
16:36:55pamauryZoiah: they published the SoC manual, the elftosb and sbtools source and the doc of the tools/format
16:37:08pamaurythey didn't published some things related to the ROM though
16:37:15 Join liar [0] (
16:41:22ZoiahNot the SDK, right?
16:45:23pamauryno :) I would dream of it since I spent hours and hours reverse engineering it
16:50:43ZoiahSTMP36xx is probably too different? Because that SDK seems leaked.
16:52:07 Part LinusN
16:58:44pamauryit's quite different
17:02:45pamaurythe sdk of the stmp37xx is based on ThreadX
17:02:50ZoiahIt's atleast an ARM core, compared to the MP35xx CPU craziness.
17:03:01 Part Zagor
17:05:26ZoiahSigmaTel uses Express Logic's ThreadX® real-time operating system (RTOS) in its STMP3600 System on Chip (SoC) which provides …/,
17:09:46pamauryreally ? Didn't know that. In any case, there are lots of hardware differences and it's doesn't solve the fundamental problem that it's a sdk so it's meant to be modified and you are interested in the modification :)
17:24:53 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
17:25:08 Join Darko [0] (5fb434a9@gateway/web/freenode/ip.
17:25:26DarkoI have compiled my Tutorial here
17:25:34Darkoand registered to Wiki
17:25:45Darkocan someone grant me access to write/editi Wiki pages
17:26:07Darkoalso, do you have any suggestion where to put this Tutorial in? What section is best?
17:26:54 Quit GodEater_ (Quit: Page closed)
17:28:48 Quit chkktri_ (Read error: Connection reset by peer)
17:29:09 Join tails___ [0] (
17:34:20 Join chkktri_ [0] (
17:36:04 Quit hilbert (Ping timeout: 255 seconds)
17:36:09 Join ChickeNES [0] (~ChickeNES@
17:36:22 Quit Llorean (Ping timeout: 240 seconds)
17:37:51 Quit tails___ (Ping timeout: 256 seconds)
17:37:52 Quit chkktri_ (Read error: Connection reset by peer)
17:38:02 Join chkktri_ [0] (
17:41:03 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
17:43:21 Quit chkktri_ (Read error: Connection reset by peer)
17:46:21mortalisDarko: What's your wiki name?
17:52:32 Quit Darko (Ping timeout: 265 seconds)
17:52:40 Quit MethoS- (Read error: Connection reset by peer)
17:55:11Zoiahpamaury: with 'random key' in your sb 'exploit', does this mean IV?
17:55:29ZoiahBear with me if I'm asking stupid questions. ;)
17:56:53 Join Bagder [0] (
17:56:54 Quit Bagder (Changing host)
17:56:54 Join Bagder [241] (~daniel@rockbox/developer/bagder)
17:57:22 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
18:02:31 Join Darko [0] (5fb434a9@gateway/web/freenode/ip.
18:02:59Darkoit is Darko
18:03:01Darkomy Wiki name
18:05:38***Saving seen data "./dancer.seen"
18:05:43 Join fyrestorm [0] (
18:07:47mortalisDarko: your wiki name should be first name + last name
18:08:02DarkoDarko Angeleski
18:09:24mortalisDarko: access granted
18:09:50Darkocan you recommend section where to put this tutorial in?
18:10:13DarkoI will work on it later this evening
18:11:10Darkoprobably as a branch of SansaAMSunbrick
18:12:15 Join dv_ [0] (
18:12:48dv_I have a sansa clip+, and want to put rockbox on it
18:13:13krazykitok, so go ahead
18:13:30dv_however, this wiki page speaks of USB problems: according to this, I could get the svn code and apply the USB patch,
18:13:44dv_but it is not recommended, because it gives me only USB 1.1 on windows.
18:13:48dv_is this the only problem?
18:16:33krazykitwell, you can always boot into the original firmware for usb, if it really bothers you
18:18:20 Join The_Seeker [0] (
18:20:52 Quit Darko (Quit: Page closed)
18:25:47 Join hilbert [0] (
18:26:41 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
18:31:28 Quit WalkGood (Quit: ♪ ♫ ♪ ♫ ♪ ♫ ♪ ♫)
18:32:49pamauryZoiah: IV and key are two separate things in AES, I don't really have time to explain I need to leave, let's say it's a tricky detail of the format :)
18:32:52 Quit pamaury (Remote host closed the connection)
18:34:24 Quit The_Seeker (Quit: I know your passwords)
18:34:26kugelmc2739: bummer!
18:37:45 Join Jerom [0] (~jerome@
18:37:45 Join Thra11 [0] (~thrall@
18:45:14mortalisI want to start hm801 port. Can i get svn access? Or I should use FS for posting patches? svn access would be much more handy for me as i plan to continue work on RB for hifimans.
18:49:09Tornepeople are invited to become committers after we have seen their work
18:49:23Tornenot because they are doing some work in future :)
18:49:41TorneWe're trying to transition to using git, which will make this easier
18:49:57Tornethere is a git mirror of svn at the moment that you could use.
18:51:01 Quit ChickeNES (Quit: Computer has gone to sleep.)
19:00:32 Join acorn [0] (2e4019ad@gateway/web/freenode/ip.
19:01:50acornis it possible to write a playlist that specifies start end times within mp3 files that can be played in rockbox?
19:02:47 Join thomasjfox [0] (~thomasjfo@rockbox/developer/thomasjfox)
19:03:19acornUsing a cue sheet wouldn't help?
19:03:41LloreanA cue sheet is for subtracks within a single file. It's a bit different than a playlist.
19:04:03acorncan you refer to subtracks within a file from a playlist?
19:04:10LloreanNo, as I said.
19:04:28LloreanThis isn't defined within standard M3U as far as I know, and this is what we currently support.
19:04:42mortalisTorne: I've done some work allready. hifiman hm-601 initial port, which was commited by wodz. Do i need some permissions on git mirror to commit my work so everyone could see it?
19:04:53acornI suppose it could be done with a plugin using the playback controls API
19:05:04 Quit tmzt (Remote host closed the connection)
19:05:23 Join lorenzo92 [0] (
19:07:23 Quit fyrestorm (Ping timeout: 256 seconds)
19:10:19 Join saratoga [0] (98034408@gateway/web/freenode/ip.
19:16:36 Quit lorenzo92 (Ping timeout: 245 seconds)
19:19:46saratogamortalis: usually people doing a new port get svn invites pretty quickly
19:20:06 Quit linuxstb (Ping timeout: 258 seconds)
19:27:14 Quit fs-bluebot (Quit: So long, and thanks for all the fish.)
19:28:39 Join ReimuHak_ [0] (~reimu@
19:29:25 Join fs-bluebot [0] (
19:31:46 Quit y4n (Quit: To those frustrated with their marginal existence, we transmit: your time is now.)
19:32:27 Join y4n [0] (y4n@unaffiliated/y4ndexx)
19:34:42 Join Horscht [0] (
19:34:42 Quit Horscht (Changing host)
19:34:42 Join Horscht [0] (~Horscht@xbmc/user/horscht)
19:36:57 Join linuxstb [0] (
19:36:58 Quit linuxstb (Changing host)
19:36:58 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
19:37:35*user890104 wonders why shows changes since 3.8 and not since the last release
19:38:48bluebrotherprobably because nobody updated it :)
19:41:23 Quit linuxstb (Ping timeout: 256 seconds)
19:42:53 Join chkktri [0] (chikakitaa@unaffiliated/chkktri)
19:43:15 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
19:45:23 Join T44 [0] (
19:45:49 Join robin0800 [0] (~robin0800@
19:48:44 Join lorenzo92 [0] (
19:49:03 Quit Topy (Ping timeout: 258 seconds)
19:57:34 Quit robin0800 (Ping timeout: 248 seconds)
19:59:19 Join robin0800 [0] (~robin0800@
20:05:41***Saving seen data "./dancer.seen"
20:07:10 Quit robin0800 (Ping timeout: 248 seconds)
20:08:42 Join robin0800 [0] (~robin0800@
20:10:17 Quit matze` (Remote host closed the connection)
20:12:13mc2739kugel: no longer a bummer!
20:13:33 Join Buschel [0] (
20:13:49kugelmc2739: hmm; I also tried the mass-files trick (my dircache was ~3MB)
20:13:53kugelwithout success
20:17:01 Join Buschel_ [0] (
20:18:24 Quit Buschel (Ping timeout: 260 seconds)
20:18:29 Nick Buschel_ is now known as Buschel (
20:18:57 Quit robin0800 (Ping timeout: 252 seconds)
20:20:29 Quit Thra11 (Ping timeout: 256 seconds)
20:20:29 Join robin0800 [0] (~robin0800@
20:23:19 Quit ReimuHak_ (Ping timeout: 256 seconds)
20:25:11bertrikonce I have the clip zip (and it has a rds capable fm tuner indeed) I can finally start to work on RDS support for real
20:25:58amiconnsaratoga: Please remind me, what's this channel swap good for?
20:26:17bertrikit looks like a GPIO is used as interrupt to notify receipt of a new RDS packet
20:27:42bertrikI think I'll use a thread to read the RDS data over i2c and use the interrupt to wake up the thread
20:28:02 Quit Zambezi (Quit: Reconnecting)
20:28:12 Join Zambezi [0] (Zulu@
20:28:37*jhMikeS doesn't think "channel swap" should be a stereo mode at all, nor should it need to copy anything.
20:32:43 Join Thra11 [0] (~thrall@
20:33:50 Quit Zambezi (Changing host)
20:33:50 Join Zambezi [0] (Zulu@unaffiliated/zambezi)
20:34:06jhMikeSits' not really a stereo channel mode like karaoke but that's how the patch was implemented, as part of the stereo mode menu
20:34:26jhMikeSsecond, pointer swapping should be all that's needed, not swapping the channel data
20:34:56 Quit robin0800 (Ping timeout: 245 seconds)
20:35:10amiconn(a) it's an additional option. (b) if we want it at all, on hwcodec the only way to implement it is to use the stereo matrix
20:35:32jhMikeSas for what it's good for, who knows. I suspect it's to get around some PCM driver bug on one target.
20:37:17amiconnIf it's only that, I'd say it's silly to add such an option. Instead of that, the bug should be fixed
20:37:42saratogaamiconn: dfkt and linuxstb came up with it ages ago, i'm not sure what the use case was
20:37:52saratogai don't think it was a device bug though
20:38:16kugel"don't know what this ancient patch is for...let's commit and see!"
20:38:37jhMikeSwas it? I remember it being suggested due to channel swap on newer ipods. I also recall dicussing it here briefly, but noone was exactly enthusiastic.
20:39:17saratogawell we have mono and karaoke modes, so channel reverse seems sensible enough
20:39:21jhMikeSthe former was recent, the latter years ago
20:39:39 Quit benedikt93 (Quit: Bye ;))
20:39:43mortalisFS #12355. One more HiFiMAN. Comments are welcome.
20:39:45fs-bluebot 3First patch for HiFiMAN HM-801 target (patches, unconfirmed)
20:39:49saratogaoh nice
20:40:04jhMikeSyeah, but channel swap can be applied along with the other modes
20:40:19saratogawhich other modes would it make sense in?
20:40:58saratogai guess the cross feed type mode?
20:41:34jhMikeSstereo, karaoke, custom
20:42:03jhMikeSI was thinking it should just be a separate step if it's to be there at all
20:42:43saratogai guess the custom mode could be used in parallel with some of the others as well
20:42:47dfktsaratoga, the case for channel swap was that some in-ear phones are designed for use with the cables hanging down, which is uncomfortable, insecure, and adds cable noise. with channel swap one can reverse them and use them over the ear
20:43:11dfktlike radius ddm, jays q-jays, etc
20:43:22saratogaah yes, i had a pair of sony headphones like that years ago
20:43:43amiconnjhMikeS, saratoga: Depending how the stereo mixing stage is implemented, adding channel swap should be easy
20:44:14 Quit kugel (Remote host closed the connection)
20:44:19saratogai'm actually surprised it doesn't already
20:44:20 Join kugel [0] (~kugel@
20:44:20 Quit kugel (Changing host)
20:44:20 Join kugel [0] (~kugel@rockbox/developer/kugel)
20:44:39amiconnThe MAS has a stereo matrix with 4 values, L -> L gain, R -> R gain, L -> R gain and R -> L gain
20:44:43saratogaa gain of 0 and a cross of 1 ought to flip the channels (although with substantially more overhead then pointer swapping)
20:44:53dfktif the channel swap patch gets committed, i would think it would be better to put it above karaoke mode in the list menu, below the mono modes. karaoke is rather "special", while swap is more related to regular channel modes
20:44:55amiconnChannel swap means just swapping the straight gains with the cross gains
20:45:00 Quit mortalis (Quit: KVIrc 4.1.1 Equilibrium
20:45:12 Join advcomp2019_ [0] (
20:45:13 Quit advcomp2019_ (Changing host)
20:45:13 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
20:45:40amiconndfkt: Making it a mode in that list doesn't make sense. It should be a separate option that can be used in conjunction with any of these modes
20:45:52 Join mortalis [0] (
20:45:54saratogaamiconn: while I have your attention, another dsp related question: I'd like to add L/R volume balance to the DSP engine
20:46:03amiconn(technically it doesn't make sense for karaoke and pure mono though)
20:46:23saratogaright now its done in hardware, which has some drawbacks (e.g. RAAA)
20:46:39dfktusing it in conjunction with mono mode doesn't make much sense either, methinks?
20:46:41jhMikeSsort of does in karaoke since one channel is the inverse of the other
20:46:44saratogai'm guessing this is because the dsp code evolved from hwcodec?
20:47:38jhMikeSwhy add volume balance to the DSP engine
20:48:44saratogaso that it works on devices that can't do it in hardware (e.g. RAAA, various line out ports, etc)
20:49:04 Quit advcomp2019 (Ping timeout: 276 seconds)
20:49:08jhMikeSahh, nvm. it'd be better done in the pcm driver (as well as sw volume control)
20:49:27saratogaalso i kind of dislike doing it in hardware in general just because it means that on some devices you'll get different outputs between the line out and headphone jack which is kind of confusing
20:49:39dfktyeah, balance on RaaA would be awesome... my ears are imbalanced, and i have to use a bridged inline volume control to get balanced sound
20:49:40amiconnWhy can't raaa do it in hardware? It's just outside rockbox...
20:49:43saratogaright now its kind of random weather or not balance will actually do anything depending on which device and output you're using
20:50:33*amiconn actually prefers using the hardware controls over software where possible
20:51:07saratogayeah thats what i was going to ask you, whats your preferred way to handle this on devices like the X5 where the headphone jack does balance but the line out jack does not
20:51:44saratogai'm leaning towards just doing software on devices where not all outputs can handle it in hardware so that output is consistent
20:52:33amiconnWell line out is line out - it shouldn't be affected by any of the sound controls imo
20:53:08saratogait shouldn't be affected by volume adjustments, but balance is something else
20:53:28saratogayou might reasonably have a stereo where you need more power on one side of the room then the other
20:53:50saratogajust as you might want more power at lower frequencies, etc
20:54:02amiconnLine shouldn't be affected by balance, tone controls, volume, anything
20:54:24saratogaand EQ?
20:54:45amiconnSame (not possible with sw eq, unfortunately)
20:55:04saratogacould you explain that point of view to me?
20:55:13amiconnLine out is for connecting better equipment. If you feed the signal to an amp for listening in a large room and you need to adjust balance, you should do so using the amp controls
20:55:22amiconnSame for tone etc
20:55:50amiconnThe controls on the device are for listening using the built-in amplifier
20:56:19saratogaits relatively unlikely that built in controls are any better though for a great many of these things
20:56:25saratogaif they are even available
20:56:43saratogae.g. how many speaker docks have a parametric eq?
20:57:05amiconnIf the amp doesn't have its own control, it should be connected to headphone out instead of line out
20:57:43saratogabut line out makes more sense since it won't be affected by volume control
21:00:04saratogabasically on these devices line out is just headphone out with fixed volume (and sometimes a dock connector) and higher impedance so its not clear to me why we should be telling users which one they should be using for a given task, imo we should try and make them work the same and let each user pick the right one for a job
21:01:42amiconnIf line out doesn't respond to volume controls but other sound controls, that's inconsistent
21:02:02saratogaits consistent with what a line out is supposed to do
21:02:18saratogaits supposed to be an output with fixed level and impedance
21:02:24saratoganothing more, nothing less
21:02:33amiconnA line out should supply a linear signal response
21:02:45saratogaso should all outputs
21:02:59saratogaa nonlinear response would be useless in audio
21:03:26amiconnTone controls aren't linear
21:03:34 Quit mortalis (Quit: CGI:IRC)
21:03:36 Quit Buschel (Quit: ChatZilla 0.9.87 [Firefox 7.0.1/20110928134238])
21:03:55saratogatechnically frequency adjustmnet is linear, but i suspect thats not what you're trying to tell me
21:04:25 Quit thomasjfox (Remote host closed the connection)
21:04:50amiconnImo the rule is simple. If one person listens using the built-in headphone output and (an)other person(s) listens using an amplifier connected to line out. Person #1 adjusting tone/ volume/ balance shouldn't affect the others
21:04:53saratogabasically you're saying that line out should disable as much of the effects as possible rather then expect the user to pick the correct settings himself?
21:05:40saratogawhile i can see how that would be useful in some cases, I don't think thats a good idea in general
21:06:33amiconnThat's how any line out is supposed to work, otherwise it should not be called line out
21:06:42saratogai also disagree that theres any expectation that a line out will somehow have less processing then a headphone out (or any other output)
21:10:03amiconnWhat if person #2 is not actually a person, but e.g. a recording device? Okay, analog recording is probably not used often these days, but sometimes it still is...
21:10:32amiconnThe same reasoning applies to digital outputs, btw
21:13:30amiconnI'm not saying that an adjustable output doesn't make sense, but it should not be limited to this mode, because it reduces versatility
21:13:53amiconnAlso if it's called line out it should not default to adjustable mode
21:16:00dfktas far as i know, mp3 player manufacturers use the line-out definition rather loosely anyways. i don#t know of any device that puts out -10dbV or 0.3 VRMS
21:18:24saratogaamiconn: if an analog device is hooked up I expect it to record what I hear, transmitted at the correct level and impedance
21:18:50saratogai don't think line output typically work in the way you seem to imply they should
21:26:31 Join fdm [0] (
21:27:10 Quit linuxstb (Ping timeout: 248 seconds)
21:28:15amiconnNearly all lineouts I dealt with so far do work that way
21:29:34 Join Strife89 [0] (~Strife89@
21:30:38 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
21:36:41 Join robin0800 [0] (~robin0800@
21:44:09 Quit linuxstb (Ping timeout: 260 seconds)
21:45:28 Join linuxstb [0] (
21:45:30 Quit linuxstb (Changing host)
21:45:30 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
21:48:48saratogaamiconn: which devices are you thinking of?
21:50:16 Quit Jerom (Quit: Leaving.)
21:56:14 Quit Thra11 (Quit: kthxbai)
21:59:10 Quit linuxstb (Ping timeout: 260 seconds)
22:00:15 Quit saratoga (Quit: Page closed)
22:02:12 Join AlexP [0] (~alex@rockbox/staff/AlexP)
22:04:05 Join webguest15 [0] (
22:05:10webguest15Will the new <8and temporarily reverted) swap channels feature allow to change the deepness? I.e. combine "custom" and "swap"?
22:05:45***Saving seen data "./dancer.seen"
22:07:16 Join Horschti [0] (~Horscht@xbmc/user/horscht)
22:07:50*jhMikeS was arguing in that direction that it should and be a separate setting if anything
22:10:59 Quit Horscht (Ping timeout: 256 seconds)
22:19:41 Quit Kuitsi (Ping timeout: 258 seconds)
22:23:52 Quit mystica555 (Remote host closed the connection)
22:24:22 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
22:27:06 Join Kuitsi [0] (
22:27:34 Join Darko [0] (5fb434a9@gateway/web/freenode/ip.
22:27:54Darkocan someone help me how to start new Wiki page?
22:28:09 Quit webguest15 (Quit: CGI:IRC)
22:28:28 Join wodz [0] (
22:29:01wodzjhMikeS: Is it correct that playback is always 44.1kHz and anything else is resampled?
22:29:40DarkoI want to create page that branches from this section
22:30:38wodzDarko: you start new wiki page by entering correct wiki address which does not exist yet
22:31:08Darkook thanks
22:36:31 Join linuxstb [0] (
22:36:32 Quit linuxstb (Changing host)
22:36:32 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
22:43:08 Join ReimuHak_ [0] (
22:52:31Darkoalright, the core page has been created with some basic editing, still needs formating of pictures and url's
22:53:47 Quit TheLemonMan (Quit: WeeChat 0.3.6)
22:54:38 Quit ReimuHak_ (Quit: Leaving...)
22:54:56 Join ReimuHak_ [0] (
22:57:47 Quit robin0800 (Ping timeout: 240 seconds)
23:03:20 Quit lorenzo92 (Quit: ChatZilla 0.9.87 [Firefox 7.0.1/20111008085652])
23:04:08 Join dhrasmus [0] (
23:06:03 Quit Strife89 (Remote host closed the connection)
23:06:58 Join robin0800 [0] (~robin0800@
23:09:14 Quit domonoky (Read error: Connection reset by peer)
23:10:11 Join MethoS- [0] (~clemens@
23:17:55 Quit [Saint_] (Ping timeout: 276 seconds)
23:24:52 Join z180 [0] (
23:29:56 Quit ender` (Quit: /* Return code=1: generic error condition, Return code=2: all other error conditions */)
23:34:16 Quit bertrik (Ping timeout: 255 seconds)
23:45:55 Quit Darko (Quit: Page closed)
23:48:49 Quit Llorean (Read error: Connection reset by peer)

Previous day | Next day