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 2009-01-15

00:01:14 Quit bmbl ("Woah!")
00:01:46ZagorFS #9800 - Zero-wait boost for PP502X
00:02:37 Join kachna|lappy [0] (
00:03:13 Quit bluebrother ("leaving")
00:04:05 Join solexx [0] (
00:04:21 Quit kachna (Read error: 104 (Connection reset by peer))
00:07:06 Part red0nkulous
00:10:30kugelZagor: compiling
00:10:41 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
00:14:36kugelZagor: updating bootloader isn't needed, is it?
00:15:28 Quit ibseco (Read error: 60 (Operation timed out))
00:16:15 Join itcheg [0] (i=62db4767@gateway/web/ajax/
00:18:04kugelJdGordon: your patch works good so far
00:18:46kugeland the lag I was reporting was actually svn, I've never noticed before :S
00:19:42JdGordonpixelma: did you get to test it on hwcodec?
00:19:56 Quit solexx_ (Read error: 110 (Connection timed out))
00:21:45 Quit Llorean (Read error: 104 (Connection reset by peer))
00:22:09 Join Llorean [0] (
00:22:53 Join fdinel [0] (
00:24:02 Quit MethoS (Remote closed the connection)
00:26:38 Join akur [0] (
00:26:43 Part akur
00:27:34 Quit Zagor ("Bed time")
00:30:18 Quit tim__ (Read error: 104 (Connection reset by peer))
00:31:13 Join tim__ [0] (n=aoeu@
00:33:33 Quit J-23 ("ZNC -")
00:33:43 Join Davide-NYC [0] (
00:34:02 Quit gevaerts (Nick collision from services.)
00:34:12 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
00:35:34kugeljhMikeS: I'm hoping you can help me a bit with setting the cache and mmu up. It's exactly the same procedure as it's the (almost) the same SoC. I copied the mmu setup from the f/x crt0.S, and the bootloader does boot, but not the main build, so there's probably something missing
00:35:57 Join J-23 [0] (
00:38:10kugelJdGordon: I just had the next track info not showing up
00:38:46kugelI wanted 20s for it to show up, then skipped. At the next song, the next track info was good again
00:39:38JdGordonwas it actually in the buffer?
00:40:01JdGordonI think some debug wps tags might be useful here :p
00:40:02kugelmeh, forgot to look
00:40:21JdGordoni doubt it was... next track seems to be fairly reliable now, except for PaulJam :/
00:40:55kugelah here, I have a song playing now, buffer is nearly empty, and next track info is missing
00:41:05kugelstill not there after 30s
00:42:11PaulJamkugel: you can look in the buffering debug menu, if track count is 1, then the next track isn't buffered yet.
00:42:55kugelit is
00:43:09 Join Davide-NYC_ [0] (
00:43:18 Join neddy [0] (n=john@nat/sun/x-ce5518c2ead53692)
00:43:58JdGordonLlorean: you have an archos yeah?
00:44:39Davide-NYC_Does anyone know if MarcGuay recovered his Gigabeat S just to v1.2 OF or if he got Rockbox on there proper.
00:45:22kugelwow, the buffer was <10k before refilling. Nice watermark rework
00:46:10JdGordonyeah.. i wonder if thats going to be too tight
00:46:23kugelI didn't hear skipping
00:46:42JdGordonbut you didnt have next track info! :D
00:47:13kugelblame Zagor? ;)
00:48:10 Quit Davide-NYC (Read error: 145 (Connection timed out))
00:48:22 Nick Davide-NYC_ is now known as Davide-NYC (
00:48:26kugelJdGordon: is it possible to get the next track info from playlist (instead of mob) if the next track isn't buffered?
00:48:53PaulJamJdGordon: isn't it expected that there is no next track info when the track isn't buffered yet? The comment in playback.c somehow sounds as if this isn't implemented yet (// TODO: fix this.. )
00:48:57JdGordonim not sure.. but that isnt ideal.. it would force a spin up which we dont want
00:49:24kugelJdGordon: but it's determined at refilling
00:49:43JdGordonPaulJam: yeah, that fixme is more figure out how what should actually be there.. that was my first guess but may be wrong
00:49:44 Quit Schmogel ("Miranda IM! Smaller, Faster, Easier.")
00:49:46kugeli.e. you can ask the playlist for the missing track info already at refilling time
00:49:59kugelwithout extra spinup
00:50:14gevaertsYou can ask about the filename, but not much more
00:50:31 Part toffe82
00:50:37kugelif you have the filename you can to get_metadata()..
00:50:39saratogaZagor: using fs9800 on e200v1 now. No apparent issues, gui seems responsive enough
00:51:01saratogaaverage clock speed is below 30MHz too, so we're saving battery
00:51:06*pixelma wonders why one gets the warning abour reconfiguring after "make && make zip"
00:51:13pixelma*about too
00:51:20kugelconfigure is newer than the makefile
00:51:28saratogai believe that happens if the configure script changes and you haven't rerun it
00:52:00JdGordonkugel: the second the metadata is avilaable for the next track, the WPS will update... this will happen automatically when its refilled.. we dont want to cause extra spinups to get this
00:52:15pixelmayes, but why does it tell me after the successfull nuild?
00:52:28kugelJdGordon: :? I just explained that you don't need an extra spin up
00:52:43*pixelma joins typing school
00:53:15JdGordonhow would you get the metadata if you only have a filename?
00:53:45 Quit _lifeless (Remote closed the connection)
00:53:47kugelthat's used in pictureflow to get the metadata for the current slide
00:53:53JdGordondoes that magically pull the info from thin air?
00:54:12kugelI don't understand
00:54:29JdGordonPaulJam: I cant reproduce your next track bug
00:54:33PaulJamJdGordon: why don't you do it like it is currently done in SVN? if i understand correctly there is an extra struct (lasttrack_id3) that holds the metadata of the first unbuffered track.
00:55:05***Saving seen data "./dancer.seen"
00:55:08 Quit Davide-NYC ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
00:55:13kugelI can :S
00:56:12Lloreankugel: How will get_metadata get metadata without a spinup if the data isn't on the buffer already?
00:56:40kugelthe disk is still spinning from the (re)buffer?
00:57:18 Quit ender` (" A man without religion is like a fish without a bike.")
00:57:23JdGordonPaulJam: I guess thats what the fixme is for...
00:57:33kugelas I said, you can determine the missing track at refilling time. so once refilling is finished, you can ask the playlist for the filename and parse the metadata
00:57:35 Join _lifeless [0] (n=lifeless@
00:57:43 Quit _lifeless (Remote closed the connection)
00:57:46kugelthen the spins down normally
00:58:22 Join _lifeless [0] (n=lifeless@
00:58:51Lloreankugel: On all of our disk based targets the disk should long since be spun down by the time you're at the end of the buffer and the next track is off-buffer, except with long FLAC files or something
00:59:28kugelLlorean: I think you don't quite understand what I mean
00:59:55kugelwhen the buffer is empty (or you start something new), the buffer is refilling
01:00:31kugelthis is the point where you can determine the missing track. and within this refilling you can parse the missing track (which won't fit into the buffer) for metadata
01:00:58kugelsplit seconds after refilling finished
01:01:37pixelmaisn't that what's currently happening?
01:01:50 Quit J-23 (Read error: 101 (Network is unreachable))
01:01:57kugelapparently not in JdGordon's patch
01:02:09*kugel doesn't know how it's exactly done in svn
01:02:15 Join mrkiko [0] (
01:02:28JdGordonapparently its done like that in svn
01:02:50JdGordonwhich is a bit i didnt understand fully, but will readd
01:03:33kugelJdGordon: I think that should be working with your patch too, and I suspect you possibly don't even need a temporary second id3 struct
01:04:23JdGordonyou're right... it needs the 3rd static id3 sturct put back
01:04:32kugelunless it's invalid to put the next track info for the unbuffered track into the MoB
01:04:55JdGordonit wont fit in Mob!
01:04:57JdGordonthat the point
01:05:05JdGordononce bufffering finishes... there is no room left
01:05:36kugelhm? after 0.1s there should be enough place for a id3 struct, not?
01:06:04JdGordonby then the disk should be spun down
01:06:06kugelor maybe not, I haven't calculated
01:07:11pixelmaJdGordon: currently testing on my Ondio, no next track info before the first track change
01:08:19JdGordonat all?
01:08:59JdGordondoes the wps show correct info for the current track?
01:09:15kugel"/* prev track during track-change-transition, or end of playlist, * next track otherwise */"
01:09:22kugeldidn't that work out well?
01:10:14kugelI don't see where you need 3 id3 structs if you use your approach of user "other_track" for both prev and next track
01:10:56JdGordonthat would only work if the buffer holds only 1 track
01:10:59pixelmatried before with SVN and it worked as it used to. No next track at all (in my WPS I check for the presence of a next track's filename and of no other metadata is found it'll display that but it didn't)... same on second try, just letting it play
01:11:18pixelmacurrent track's metadata is displayer correctly
01:11:53kugelI think you can have 2 id3 structs. the one holding the current track, one holding the other. you never need prev and next track at the same time, do you?
01:12:59JdGordonpixelma: ok, I guess this means I havnt modified mpeg.c properly.... as long as it doesnt crash or anything I'm happy... someone with a actual target might need to fix this :/
01:13:05kugelso you can just swap "next track" with "current track" on track transition, and then write the next track info into that that held current_trackbefore
01:13:07JdGordonkugel: comments are misleading...
01:13:51pixelmaJdGordon: it doesn't even help when stopping and resuming (forcing a buffer) within the last half a minute of the track which should be a time next track is in buffer with the 128kbps MP3s I'm testing with
01:13:53kugelbut I imagine it can actually work this way
01:15:43JdGordonpixelma: ok.. thanks for testing
01:16:33JdGordonkugel: no.. it cant... there are 2 id3 structs which get round-robbined between the current track and the next track... there is nowhere to store info for the unbuffered track just yet
01:17:55 Quit mrkiko ("leaving")
01:18:18kugelJdGordon: well, actually it doesn't matter anyway
01:18:49kugeleither way, if you put it into the MoB, or in a static struct, you'll need 1 extra id3 struct in both cases
01:20:44 Quit faemir ("Leaving")
01:20:57 Join J-23 [0] (n=zelazko@
01:23:46 Quit tim__ (Connection timed out)
01:24:04pixelmaJdGordon: what did Zagor say for testing the coldfire patch - just trying with the new firmware file is enough, changing the build completely is not needed?
01:25:30 Join tim__ [0] (n=aoeu@
01:27:29JdGordonPaulJam: good work... I've put that back in, so i now have next track info with only 1 track buffered
01:28:30JdGordonthis probably makes the next id3 avialable event useless...
01:28:54TheSkunkManif i wanted to try to decrypt the iriver lplayer and e100 firmware where should i start? is there a program that simplifies comparing the .hex files at all?
01:29:23kugelJdGordon: probably ;)
01:29:23JdGordonTheSkunkMan: a short course in cryptonalyses...
01:29:53saratogakugel: was there ever agreement that we should change the power button the sansas to not go up one level when you hit it?
01:29:54kugelso, lets sum up, I was twice suggesting things that are already done this way today :/
01:29:55JdGordonTheSkunkMan: we need to get in touch with the guy who cracked the older iriver firmwares...
01:30:04saratogai'm 99% sure i remember that being a bug but i want to double check
01:30:25kugelsaratoga: I think there was, and I made the patch (not entirely finished since I'm waiting for Llorean to test it)
01:30:57saratogakugel: what was the fs number for it?
01:31:44 Join JackZhou [0] (n=cef8a591@gateway/web/cgi-irc/
01:32:03kugelthe result is, that power cannot really be used for canceling (except for plugins), but that's supposed to be mapped to |<< anyways
01:32:10saratogakugel: this issue has been driving me crazy ever since I tried your first version of that patch and noticed it
01:32:45 Join Sonicjosh [0] (
01:32:50saratogawell i'm strongly in favor of commiting it
01:32:59saratogaLlorean wanted to look at it before we do?
01:33:10kugelI think so yes, I talked to him
01:33:34saratogaLlorean: what are you thoughts on commiting fs9662?
01:33:37 Quit flydutch ("/* empty */")
01:33:38kugelbut before committing it should be noted that langs need fixing (yesno screen, time&date)
01:34:31 Join SUSaiyan [0] (
01:35:03kugelsince cancel is mapped to |<< consistently (it always was), but yesno screen and time and date ask the user for power nevertheless
01:35:57kugeland keymap is mixed with langs here :/
01:36:21 Join Makuseru [0] (
01:36:32Sonicjoshwhere can i download rockboy?
01:36:41kugelsaratoga: if you test it, would you please comment on the time & date screen? if that's still usable sufficiently? it changed a bit
01:37:01kugelSonicjosh: it's built-in (as a plugin)
01:37:13rasherSonicjosh: check the manual
01:37:16kugelyou start it by "clicking" on a gb(c) file
01:37:24pixelmaSonicjosh: maybe download the manual instead... ;)
01:37:26Unhelpfulis there a way to get diff-stat output from svn, instead of an actual diff?
01:37:40*pixelma too slow
01:38:07Unhelpfulthat might belong in the other channel, but it relates to the build-info.txt some of use were talking about a while ago...
01:38:18kugelsvn diff|diff-stat?
01:38:41kugeli.e. piping, if diff-stat exists as an external tool that is
01:39:28Unhelpfulright... i don't have it here, but i guess we can check for it before trying to use it.
01:39:33JdGordonPaulJam: kugel: new version is on the tracker
01:39:44JdGordonexcept i forgot to remove the CF clock patch...
01:40:36JdGordonok, its safe now
01:43:55rasherif [ -x "`which diffstat 2> /dev/null`" ]; then svn diff -r 19764:19765|diffstat; fi
01:45:09 Quit itcheg (" ajax IRC Client")
01:45:46 Quit agaffney (Read error: 104 (Connection reset by peer))
01:46:59JackZhouHi, I'm still requesting edit permissions to the wiki to upload a wps.
01:47:48 Quit Makuseru (Read error: 104 (Connection reset by peer))
01:48:09 Join ae88925 [0] (
01:48:53rasherJackZhou: I assume you're JackZhou?
01:50:08JackZhourasher: Yup.
01:50:16rasherAdded, now please don't spam.
01:50:27Unhelpfulrasher: yes, basically. it would be a bare svn diff, though... the idea is that build-info.txt would have "detailed" info, whatever that means, on the source from which the build was made. if building from svn, i'm not sure we can get more than a diff-stat of the local changes. if building from git, probably the diff-stat, a commit-ish for the current HEAD, and a note if the working tree is dirty.
01:50:46JackZhourasher: Okay, thanks.
01:50:48 Quit JackZhou ("CGI:IRC")
01:51:14 Join agaffney [0] (n=agaffney@gentoo/developer/agaffney)
01:51:40rasherUnhelpful: ah, I didn't know what the intention was, yeah, that makes sense
02:01:38 Join mhydronic [0] (
02:01:41 Part Sonicjosh
02:02:13mhydronichi, my calendar is off. it says tuesday the 14th. could this be a wps issue?
02:02:57saratogawrong year probably
02:03:20mhydroniccan you set the calendar?
02:04:29saratogapresumably it uses the date in rockbox
02:04:54mhydronicwhich can be set?
02:05:40ae88925It's under "System" at the main menu, I think
02:05:46JdGordonhmm... that new version of the playback patch may have made things worse... I just saw gobledygook info for the current track which I hadnt seen in a whil :(
02:06:01 Quit tvelocity (Remote closed the connection)
02:06:09kugelJdGordon: looks perfect now
02:06:13kugelthe patch that is
02:06:45PaulJami havent seen any issues with the new version of the patch
02:07:15 Quit saratoga ("CGI:IRC (EOF)")
02:08:11mhydroniccan someone guide me in changing the time/date ?
02:08:34 Quit Thundercloud (Remote closed the connection)
02:09:11Unhelpfulae88925 already gave you a pointer, you could also try reading the manual.
02:10:11mhydronicim trying that
02:10:16mhydronici don't see that option in system
02:10:24mhydroniclet me try switching wps
02:10:44Unhelpfulwhy would that help? wps doesn't change anything in the menu.
02:10:57Unhelpfuli think it might actually be under settings... but really, you could look at the manual.
02:11:05mhydronicyes, i'm doing that
02:11:17mhydronicand its supposed to be under system, yet its nto
02:11:20kugelIt's under system in the main menu
02:13:04mhydronicthis type of question is prolly something you guys hear a lot, but do you know if there will be support for ipod classic? those things are cheap on amazon right now
02:17:28PaulJammhydronic: as far as i know no one with the neccessary knowledge is workin on that port. And the encryption and the undocumented hardware don't really raise the chances of a port happening.
02:20:38mhydronicok, thanks PaulJam
02:20:45 Quit tessarakt ("Client exiting")
02:20:47 Quit mhydronic (), Sandisk sansa e260 4gb @ 64eu
02:22:00JdGordonthat sounds expensive...
02:22:49at0mi'm unexperienced w that model, but cnet price comparison was from 99$
02:27:09 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
02:47:21 Quit soap (Remote closed the connection)
02:49:56 Join soap [50] (n=soap@rockbox/staff/soap)
02:50:33 Join Willwolfe [0] (
02:55:07***Saving seen data "./dancer.seen"
03:05:51 Quit culture (Connection timed out)
03:06:01 Quit kugel (Read error: 104 (Connection reset by peer))
03:07:11 Quit XavierGr ()
03:08:53 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
03:11:55 Quit PaulJam (".")
03:16:24 Quit cool_walking_ (" ajax IRC Client")
03:24:53 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
03:33:36 Join nuonguy [0] (
03:37:04 Quit fdinel (Read error: 104 (Connection reset by peer))
03:44:49 Quit Willwolfe ("ChatZilla 0.9.84 [Firefox 3.0.1/2008070208]")
03:52:32Unhelpfulany opposition to renaming resize_on_road, and possibly resize.c as well? if it's going to happen, it should be before other things are using it... but i think that scaler.c and scale_image_callback would be better names.
03:53:11Unhelpfulor just scale_image, but it might be wise to reserve in case we need/want a non-callback-based version, at some point.
03:56:38 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
03:57:20 Quit blithe (Read error: 60 (Operation timed out))
03:57:58 Join Esdian [0] (n=Esdian@
04:00:12EsdianSorry, to bother everyone, but how do you Restart a rockbox'd Sansa e200 In safe mode. Cause I thought It was off, but it was not oops, when I pluged in the USB cord the screen went all blocky and I can't turn it off now.
04:01:58Unhelpfuli don't know what you mean by "safe mode"... but if you hold the power button long enough, it should turn off, regardless of how it ended up locked up.
04:03:00EsdianGot it off thanks.. would have hated to brick it in 2 weeks of use.
04:12:03 Join fyrestorm [0] (n=fyre@
04:13:19 Join blithe [0] (
04:13:40 Quit Nico_P (Remote closed the connection)
04:13:49 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
04:16:51 Join Barahir_ [0] (
04:19:47 Part Esdian
04:25:07 Quit Barahir (Read error: 145 (Connection timed out))
04:25:59 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
04:26:04 Join lymeca [0] (
04:28:18 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:30:31TheSkunkMandoes anyone have a favored cryptanalysis program for comparing encrypted files?
04:31:25*Unhelpful wonders if this is the right channel for that question?
04:32:13TheSkunkManwell i was wondering because im interested in trying to decrypt firmware
04:33:02TheSkunkManand you guys might prefer some over others for that kind of stuff
04:35:23saratogai wasn't aware such general programs even existed
04:37:25TheSkunkManmaybe not i just assumed there was something out there ill just try to write my own scripts
04:38:28Unhelpfulgeneral cryptanalysis isn't readily automated... you will need to know what cipher is in use, or to reverse-engineer it, before you can start thinking about keys.
04:47:45 Join toffe82 [0] (n=chatzill@
04:52:25 Quit colesterol_dog ("sleep")
04:54:16 Join blkhawk- [0] (
04:55:10***Saving seen data "./dancer.seen"
04:56:52 Quit tim__ (Connection timed out)
04:59:36 Join tim__ [0] (n=aoeu@
05:05:41 Quit at0m ("zzZ")
05:11:13 Quit blkhawk (Read error: 110 (Connection timed out))
05:11:14 Nick blkhawk- is now known as blkhawk (
05:14:40 Join itcheg [0] (i=62db4767@gateway/web/ajax/
05:19:24 Quit itcheg (Client Quit)
05:27:56 Quit BXCracer (Remote closed the connection)
05:28:20 Quit lymeca (Read error: 110 (Connection timed out))
05:28:30 Join lymeca [0] (
05:37:41 Quit fyrestorm ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
05:45:38 Quit Horscht ("Snak 5.3.3 Unregistered copy. Evaluation period is over. Program will now quit. Thanks for using Snak.")
05:47:41 Quit TheSkunkMan ("Leaving")
05:51:45 Quit Dhraakellian (Read error: 104 (Connection reset by peer))
05:53:42 Join tajlero [0] (
05:57:47 Join Makuseru [0] (
05:58:56 Quit Seed ("cu, Andre")
05:59:31 Quit Aurix_Lexico ("Leaving.")
05:59:43 Join Aurix_Lexico [0] (
06:02:31 Quit amiconn (Nick collision from services.)
06:02:33 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
06:03:22 Quit tim__ (Read error: 60 (Operation timed out))
06:03:32 Quit Makuseru (Read error: 104 (Connection reset by peer))
06:04:43 Join tim__ [0] (n=aoeu@
06:07:16 Join Alystair [0] (
06:07:32AlystairWhere do I find Windows drivers for rockbox once it's installed?
06:08:07AlystairI plugged in the player and now windows says it detected new hardware heh
06:08:08cool_walking_You don't need them. Ignore the message.
06:08:12Alystairah ok
06:08:41cool_walking_Rockbox doesn't have it's own USB mode implemented yet, so you need to use the Original Firmware for transferring files.
06:08:47Alystairnow it doesn't show up as a drive anymore? :/
06:08:58cool_walking_I can see the future.
06:09:00Alystairhow in the world do I put new music on it then...
06:09:19scorche"so you need to use the Original Firmware for transferring files."
06:09:53Alystairis there a magical key-code to temporarly revert to original firmware or do I have to do it by hand every time I put new music on the thing
06:10:10cool_walking_Yes, there's a button combination, different for each device. Read the manual.
06:11:21saratoga_labmost devices boot automatically if you power them off and then plug in USB
06:11:25 Quit toffe82 (Read error: 110 (Connection timed out))
06:11:49 Quit Aurix_Lexico ("Leaving.")
06:13:55 Quit lymeca (Read error: 110 (Connection timed out))
06:14:37cool_walking_You can enable Rockbox's (work in progress) UMS mode if you compile your own build ( I've been using it on a Gigabeat S with no problems for quite a while, but there might be some issues depending on what target you're using.
06:15:50AlystairI'm using windows and don't know a compiler from a command prompt :V
06:16:16Unhelpfulcool_walking_: ...beast is not portalplayer, and compiles with MSC on by default... since there is no MSC mode in the OF
06:16:32cool_walking_I didn't know that
06:17:14AlystairI use an iRiver H10
06:17:14cool_walking_Wait, what?
06:17:44Unhelpfulalso, i think the USB issues don't affect beast
06:18:14cool_walking_Oh you just mean you don't have to specifically turn UMS on for the beast, yeah I know that.
06:18:33cool_walking_But it's the same code isn't it?
06:19:17saratoga_labyes but different hardware
06:19:20Unhelpfuli have no idea how much of it is shared, i didn't write any of that
06:19:51cool_walking_So the remaining issues with the USB stack are due to hardware?
06:20:15 Join pixelma_ [0] (n=pixelma@rockbox/staff/pixelma)
06:20:26 Quit pixelma (Read error: 110 (Connection timed out))
06:20:46Unhelpfulbut sansa seems to know better work-arounds than we do...
06:24:41Alystairhah now I can't even update rockbox because it doesn't show up as drive :D
06:25:03Alystairemergency connection works we suggested before to use ;0
06:25:44Alystairdoes RB care how files are put on the disk?
06:25:55Alystaireg. if I throw all mp3s into a single folder will it puke
06:25:57cool_walking_other than it's own files, nope.
06:26:03Alystaircool beans
06:26:48cool_walking_Well, there's a limit to how many files you can see in a folder. There's a setting you have to increase to show more files.
06:28:11cool_walking_This one:
06:29:05scorchethough if you havent yet, you probably should have a look through the manual anyway, as you will likely find out many interesting things as well as answering a few other questions you may have
06:30:43AlystairI can't believe how cool rockbox is tho'
06:30:53Alystairthis is some handy stuff, totally refreshes this hardware :D
06:31:12Alystairrecording with visual audio bars :D
06:33:30Alystair! album art
06:34:50Alystairsome of these themes are showing up empty when I try to extract them
06:35:04cool_walking_What do you mean?
06:35:58Alystairshows up as a bunch of empty folders
06:36:18Alystairall of them do...
06:36:26cool_walking_Are you on a Mac or Linux?
06:36:31Alystair... windows
06:36:39Alystairusing ALZIP
06:37:26scorchethat zipping program likely hides folders beginning with a "."...tell it to extract anyway, as the folder is still there, and it should work...otherwise, you may need to use a different program
06:37:27cool_walking_The one you linked shows a ".rockbox" directory for me, and in there are some files.
06:38:15Alystairnevermind, ALZIP is really dumb and can't see the files
06:38:39cool_walking_It's probably done on purpose. It probably has a setting to see them.
06:42:35Alystairwhat format are the .fnt files?
06:43:15scorcheplease do a bit of searching before asking...
06:52:28 Quit _lifeless (Remote closed the connection)
06:53:52 Join _lifeless [0] (n=lifeless@
06:55:12***Saving seen data "./dancer.seen"
06:55:24 Quit saratoga ("CGI:IRC (EOF)")
07:10:54 Quit tim__ (Read error: 145 (Connection timed out))
07:11:11 Quit z35 ("Leaving")
07:12:55 Quit Alystair ()
07:16:38 Quit CaptainKewl (Read error: 60 (Operation timed out))
07:17:40 Join __lifeless [0] (n=lifeless@
07:19:37 Join tim__ [0] (n=aoeu@
07:20:26 Quit agaffney (Read error: 110 (Connection timed out))
07:27:27 Quit parafin (Read error: 60 (Operation timed out))
07:33:24 Join toffe82 [0] (
07:34:44 Quit tajlero (Read error: 110 (Connection timed out))
07:34:54 Quit _lifeless (Read error: 110 (Connection timed out))
07:35:23 Part toffe82
07:35:58 Join parafin [0] (
07:36:39 Quit saratoga_lab ("CGI:IRC (EOF)")
07:47:32 Nick fxb__ is now known as fxb (
07:49:38 Nick fxb is now known as fxb__ (
07:52:11 Join timc`` [0] (n=aoeu@
07:52:47 Nick Barahir_ is now known as Barahir (
07:58:19 Quit tim__ (Connection timed out)
07:59:35 Join GodEater_ [0] (i=c2cbc962@rockbox/staff/GodEater)
08:19:05 Quit japc (Read error: 110 (Connection timed out))
08:21:08 Join PaulJam [0] (
08:24:12 Quit JdGordon (Remote closed the connection)
08:24:24 Join Rob2223 [0] (
08:24:59PaulJampixelma_: concerning your comment in FS #9797, i thought this patch currently only affects the iriver coldfire targets.
08:26:06 Join JdGordon [0] (
08:29:43 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
08:30:34 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
08:38:53 Join ender` [0] (
08:42:14 Quit BHSPitLappy (Remote closed the connection)
08:42:48 Quit Rob2222 (Read error: 110 (Connection timed out))
08:45:27 Join troy_ [0] (n=toppy@
08:50:02UnhelpfulZagor: you still about 3/4 in favor of FS #9770? and is the other quarter something that i might be able to "fix"?
08:51:38 Quit BigBambi (Read error: 113 (No route to host))
08:51:53 Quit ze ("bleh")
08:52:02ZagorThe other 25% is the part of me saying "don't hide the code". And since that is the whole purpose of your patch I doubt it can be fixed. :-)
08:52:06amiconnHmpf, saratoga's gone
08:52:20 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
08:53:16UnhelpfulZagor: it's not *entirely* hidden, it's declared extern in plugin.h... maybe it would be better to think of it as standardizing what most plugins were doing already?
08:53:34ZagorUnhelpful: yes, hence the 75% :)
08:55:16***Saving seen data "./dancer.seen"
08:55:27*amiconn thinks that global api pointer is a good thing, because it gets rid of a lot of multiplicated (not only duplicated) code
08:55:28Unhelpfulright... i just don't really think of it as "hidden", any more than some symbol that's defined in another object file. would it be less "hidden" if the plugin had to declare it, instead of having it in the PLUGIN_HEADER macro? that seems more error-prone to me, rather than less...
08:55:48 Join Bagderr [241] (n=daniel@rockbox/developer/bagder)
08:56:11amiconnAnd it also solves a potential race condition if gcc decides to use a mem* function by itself
08:56:13 Nick Bagderr is now known as B4gder (n=daniel@rockbox/developer/bagder)
08:56:59amiconnRight now the mem* function wrappers rely on a properly initialized 'rb', but that is done *in* the plugin, so if that init is too late for the first mem* use -> boom
08:57:21amiconnThis cannot happen if 'rb' is initialized by the loader
08:57:33ZagorUnhelpful: I don't think my 25% demon can be silenced. Better just ignore him.
08:58:35Unhelpfulamiconn: that's pretty unlikely, unless you're copying structs before copying the api pointer... i also noticed that either resize.c or bmp.c (can't recall which, now) was getting a memset inserted on *some* targets, though i have no idea what for...
08:58:40amiconnIt's just that the double indirection (while being the most readable implementation) is a little awkward
08:59:47Unhelpfulif we could use a linker script to alias the symbols on sim, we could kill the double indirection. i don't *like* the pointer-to-pointer, either :/
09:03:07amiconnYeah, but the linker script only solves it for target, not for the sim
09:03:58 Join petur [50] (n=petur@rockbox/developer/petur)
09:04:30Unhelpfulis there any way to do something similar on the target? if we want to tell the linker where to put a few things, would we be forced to duplicate the entire binutils-provided linker script for dynamic libraries?
09:08:58 Quit JdGordon (Read error: 54 (Connection reset by peer))
09:11:22 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
09:12:07Unhelpfuls/target/sim/ of course.
09:15:55 Quit jhulst (Read error: 113 (No route to host))
09:16:16 Join Alystair [0] (
09:16:32Alystairhrm so as soon as I try changing to a different theme on rockbox, it reverts to a no-theme type setup... very ugly
09:16:38Alystairthen I can't return to default :/
09:17:03cool_walking_That's the default, built-in theme. You get that when your theme is broken.
09:17:06JdGordonload the cabbiev2 theme to go back to "default"
09:17:12Unhelpfulperhaps something is wrong with the theme you tried?
09:17:50amiconnUnhelpful: Not sure whether it is possible, it probably is, but then we'd have to make sure the script works for all kinds of sim hosts
09:18:32*amiconn would prefer to stay away from that
09:18:39Alystairhow do I enable album art view?
09:18:45lucentuhh... microSD is working on my fuze 8gb without any delay
09:18:49cool_walking_Use a theme which uses album art.
09:18:57lucentwhat changed with the fuze ams stuff since yesterday?
09:19:02*lucent pokes svn log
09:19:07UnhelpfulAlystair: use a them that uses album art, and make sure that you have proper album art on your player.
09:19:19Alystairthe one that shows album art breaks :/
09:19:40cool_walking_You can try to fix it if you want.
09:19:41lucentr19766 by Zagor
09:20:31Zagorlucent: that wasn't yesterday :)
09:21:05GodEater_Alystair: the default theme (cabbiev2) shows album art just fine
09:21:10lucentI don't understand why my microSD card didn't cause a panic though
09:21:32 Join DerDome [0] (
09:21:36Unhelpfulamiconn: i'm inclined toward not blocking commit over the double indirection, but that could be "i-wrote-it-" bias talking. but, it's something that may not *be* possible, and is probably quite tricky if it *is* possible...
09:22:27lucentZagor: the !@#$ thing magically works now
09:22:40AlystairGodEater: really?
09:22:55AlystairI have embedded covers and they are not showing up
09:23:00cool_walking_Alystair: See to make sure you're putting the images in the right place.
09:23:04amiconnWe could of course introduce another symbol lookup for the sim, but that wouldn't solve the target<->sim difference, and it adds the question how to return two pointers from sim_plugin_load()
09:23:05Alystairdo they have to be the exact right resultion to be displayed
09:23:06cool_walking_embedded won't work
09:23:10Zagorlucent: nice!
09:23:38cool_walking_No they are resized automatically.
09:23:39ZagorUnhelpful: what is this double indirection?
09:23:52GodEater_Alystair: Rockbox does not support embedded Album Art.
09:24:36UnhelpfulZagor: the plugin header has a pointer to rb in it, which the loader uses to store the correct value to rb. it's possible to alias rb to a field in the headin directly with a linker script, but we don't know how to do so on sim.
09:25:47Alystairw/e I'm not going to fret over it then
09:25:51cool_walking_Alystair: this may be useful:
09:26:08Alystairtoo much hastle for such a minimally used feature, not going to be looking at the damn thing half the time anyway..
09:26:16AlystairI love the fact that the hold button instantly turns off the screen
09:26:24Unhelpfulif you look in the patch, the changes in plugin.h to PLUGIN_HEADER and in plugin.c are all you need to see "how it works"
09:27:04Alystairoh man my english voice in this sounds so great now
09:27:39Alystairhey I found a bug!
09:27:54Alystairpausing the music stops voice fx from playing until unpausing music
09:28:22GodEater_That's not a bug, that's a feature ;)
09:28:32 Join japc [0] (n=japc@
09:28:41Unhelpfulamiconn: i had a semi-solution for passing the second pointer back on target, actually. if you use dlsym, you're still only going to get a pointer to rb, anyway. i had a version in which target uses __header.api, which is aliased directly to rb, and sim uses __header.api_ptr, which works like the pointer-to-rb that the patch uses now.
09:28:41ZagorUnhelpful: why the explicit .data section attribute?
09:29:10Alystairat which point you get a cacaphony of SETTINPLUFILESDATABASE comes on
09:29:12UnhelpfulZagor: amiconn caught that one - if rb is in .bss, it can be overwritten by PLUGIN_IRAM_INIT
09:29:15 Join JdGordon|uni [0] (i=7bf38c1f@rockbox/developer/JdGordon)
09:29:37 Quit scorche|sh (Read error: 60 (Operation timed out))
09:29:40 Quit JdGordon (Read error: 104 (Connection reset by peer))
09:29:46*GodEater_ tighten's his netsplit seatbelt.
09:29:46 Nick JdGordon|uni is now known as JdGordon (i=7bf38c1f@rockbox/developer/JdGordon)
09:30:18*lucent follows suit
09:30:24 Join scorche|sh [0] (
09:30:43Unhelpfulhere's the direct-alias-on-target, pointer-to-api-on-sim version, but it's out of sync, now:
09:31:03 Quit DerDome ("Leaving.")
09:31:30amiconnOf course we could just #define rb __header.api and be done. But I think this actually needs more code for api access than a pure rb pointer, initialized in whatever way
09:32:02amiconnI think the double indirection is the way to go for now, even if it's "wasting" 4 bytes
09:33:33Zagorwhy does plugin_iram_init() clobber .bss?
09:34:59 Quit cool_walking_ (" ajax IRC Client")
09:36:04 Join troy__ [0] (n=toppy@
09:37:38*Unhelpful didn't even know that it did, which is why somebody else had to figure out why mpegplayer crashed.
09:38:08amiconnZagor: It doesn't. The problem is not the iram init, but rather that .bss reuses the iram section. So we must not write to .bss *before* the iram init is done
09:38:28 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
09:40:03pixelmaPaulJam: the patch description just read "coldfire" so I didn't look closer but I see now that there are some changes only to system-iriver.c...
09:42:11pixelmathe patch definitely changes behaviour here still ;)
09:47:06 Join Thundercloud [0] (
09:49:06 Join ze [0] (i=ze@
09:54:22 Quit troy_ (Read error: 110 (Connection timed out))
09:55:18 Quit JdGordon (" ajax IRC Client")
09:55:38lucentZagor: can you describe for me what r19766 is?
09:55:56lucentI'm wondering whether or not it is the reason microSD suddenly works
09:56:04lucentit only works after boot
09:58:10Zagorlucent: it changed the way cpu clocking flags were set. the old way used xor instead of or for some reason.
09:58:35lucentZagor: does it affect the bootloader?
09:59:26Zagorlucent: no, the bootloader does not use this function
09:59:56lucentsansafuze has some weird bugs
10:00:06DEBUGEOF from server (Success) (snapshot: netstuff.c line 545)
10:00:06***Saving seen data "./dancer.seen"
10:00:07***Started Dancer V4.16
10:00:07***Connected to on port 6667
10:00:07***Logfile for #rockbox started
10:00:11***Server message 501: 'logbot :Unknown MODE flag'
10:00:11Mode"logbot :+i" by logbot
10:00:11 Join logbot [0] (n=bjst@gateway/web/cgi-irc/
10:00:11 Join maraz [0] (
10:00:11 Join ze [0] (i=ze@
10:00:11 Join Thundercloud [0] (
10:00:11 Join troy__ [0] (n=toppy@
10:00:11 Join scorche|sh [50] (n=scorche@rockbox/administrator/scorche)
10:00:11 Join japc [0] (n=japc@
10:00:11 Join Alystair [0] (
10:00:11 Join petur [50] (n=petur@rockbox/developer/petur)
10:00:11 Join B4gder [241] (n=daniel@rockbox/developer/bagder)
10:00:11 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
10:00:11 Join ender` [0] (
10:00:11 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
10:00:11 Join Rob2223 [0] (
10:00:11 Join PaulJam [0] (
10:00:11 Join GodEater_ [0] (i=c2cbc962@rockbox/staff/GodEater)
10:00:11 Join timc`` [0] (n=aoeu@
10:00:11 Join parafin [0] (
10:00:11 Join __lifeless [0] (n=lifeless@
10:00:11 Join pixelma [0] (n=pixelma@rockbox/staff/pixelma)
10:00:11 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
10:00:11 Join blkhawk [0] (
10:00:11 Join Barahir [0] (
10:00:11 Join blithe [0] (
10:00:11 Join nuonguy [0] (
10:00:11 Join soap [50] (n=soap@rockbox/staff/soap)
10:00:11 Join ae88925 [0] (
10:00:11 Join SUSaiyan [0] (
10:00:11 Join J-23 [0] (n=zelazko@
10:00:11 Join neddy [0] (n=john@nat/sun/x-ce5518c2ead53692)
10:00:11 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
10:00:11 Join Llorean [0] (n=DarkkOne@rockbox/administrator/Llorean)
10:00:11 Join solexx [0] (
10:00:11 Join kachna|lappy [0] (
10:00:11 Join undertakingyou [0] (
10:00:11 Join n17ikh|Lappy [0] (
10:00:11 Join HellDragon [0] (n=jd@Wikipedia/HellDragon)
10:00:11 Join suom1 [0] (i=markus@2001:4c40:1:0:0:0:0:11)
10:00:11 Join FlynDice [0] (
10:00:11 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
10:00:11 Join AndyI [0] (i=AndyI@
10:00:11 Join thegeek [0] (
10:00:11 Join goffa_ [0] (n=goffa@
10:00:11 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
10:00:11 Join Zambezi [0] (i=stolgfor@
10:00:11 Join preglow [0] (
10:00:11 Join jfc [0] (
10:00:11 Join gromit`` [0] (
10:00:11 Join crwl [0] (
10:00:11 Join Xerion [0] (
10:00:11 Join shodanX [0] (
10:00:11 Join reacocard [0] (
10:00:11 Join axionix [0] (
10:00:11 Join Acksaw [0] (
10:00:11 Join markun [50] (n=markun@rockbox/developer/markun)
10:00:11 Join HBK [0] (
10:00:11 Join alexbobp [0] (n=alex@
10:00:11 Join midkay [0] (n=midkay@rockbox/developer/midkay)
10:00:11 Join havien [0] (
10:00:11 Join Slack_ [0] (
10:00:11 Join kadoban [0] (
10:00:11 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
10:00:11 Join Anges [0] (
10:00:11 Join Tuplanolla [0] (n=jani@unaffiliated/tuplanolla)
10:00:11 Join Dieterbe [0] (
10:00:11 Join DaCapn [0] (
10:00:11 Join tim__b [0] (
10:00:11 Join Kopfgeldjaeger [0] (
10:00:11 Join obo [0] (n=obo@rockbox/developer/obo)
10:00:11 Join kkurbjun [0] (
10:00:11 Join Tristan [0] (
10:00:11 Join amigan [0] (
10:00:11 Join ajb [0] (
10:00:11 Join BlakeJohnson86 [0] (
10:00:11 Join sbhsu [0] (
10:00:11 Join feisar [0] (
10:00:11 Join merbzt [0] (n=benlar@
10:00:11 Join idshark [0] (
10:00:11 Join GodEater [0] (n=ge@rockbox/staff/GodEater)
10:00:11 Join dionoea [0] (n=dionoea@videolan/developer/dionoea)
10:00:11 Join jon-kha [0] (
10:00:11 Join _Auron_ [0] (
10:00:11 Join killan [0] (
10:00:11 Join Beta2K_ [0] (
10:00:11 Join Hadaka [0] (
10:00:11 Join daurn [0] (n=daurnima@unaffiliated/daurnimator)
10:00:11 Join daurnimator [0] (
10:00:11 Join liiwi [0] (
10:00:11 Join lostlogic [50] (n=lostlogi@rockbox/developer/lostlogic)
10:00:11 Join balou [0] (
10:00:11 Join shadearg [0] (
10:00:11 Join DataGhost [0] (i=dataghos@unaffiliated/dataghost)
10:00:11 Join avacore [0] (
10:00:11 Join lucent [0] (
10:00:11 Join krazykit [0] (
10:00:11 Join Unhelpful [0] (n=Militant@rockbox/developer/Unhelpful)
10:00:11 Join Bagder [241] (n=daniel@rockbox/developer/bagder)
10:00:11 Join tchan [0] (
10:00:11 Join chrippa [0] (
10:00:11 Join yosafbridge [0] (
10:00:11 Join Neovanglist [0] (i=Neovangl@
10:00:11 Join Slasheri [0] (i=miipekk@rockbox/developer/Slasheri)
10:00:11 Join larstobi [0] (n=larstobi@
10:00:11 Join desowin_ [0] (n=desowin@
10:00:11 Join fred_2 [0] (
10:00:11 Join lastebil [0] (
10:00:11 Join flux [0] (
10:00:11 Join crashd [0] (
10:00:11 Join freqmod_qu [0] (i=quassel@2001:700:300:1430:213:d3ff:fee9:5ed0)
10:00:11 Join rasher [50] (n=rasher@rockbox/developer/rasher)
10:00:11 Join courtc [0] (n=court@unaffiliated/courtc)
10:00:11 Join trisiak [0] (
10:00:11 Join pabs [0] (
10:00:11 Join ch4os [0] (n=ch4os@gentoo/user/ch4os)
10:00:11 Join TMM [0] (
10:00:11 Join tmzt [0] (n=tmzt@
10:00:11 Join martian67 [0] (i=lol3izer@about/linux/regular/martian67)
10:00:11 Join synergist [0] (
10:00:11 Join offset [0] (
10:00:11 Join fxb__ [0] (
10:00:11 Join @ChanServ [0] (ChanServ@services.)
10:00:11 Join Galois [0] (
10:00:11 Join romain [0] (
10:00:16*lucent frantically grabs a helmet
10:00:17 Join Kohlrabi [0] (
10:00:17 Join FOAD [0] (
10:00:17CtcpVersion from freenode-connect!freenode@freenode/bot/connect
10:00:20 Join Baw [0] (
10:00:46 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
10:00:48 Join tarbo [0] (n=me@unaffiliated/tarbo)
10:01:02lucentZagor: I confirm, your patch changes behavior on 8gb fuze
10:01:04 Join cool_walking_ [0] (
10:01:11lucentI should say commit r19766
10:01:20 Join scorche [0] (i=Blah@rockbox/administrator/scorche)
10:01:30Zagorlucent: well, yes. but how exactly do you mean? in unexpected ways?
10:02:01lucentZagor: prior to r19766, insertion of microSD card to fuze 8gb meant a panic
10:02:38lucentI had figured out that adding extray delays in certain places in the code avoided the panic
10:03:32lucentright now if the 8gb fuze boots with a microSD card inserted, then there is a panic
10:04:06lucentif it boots without a card inserted, and I later insert / eject a card with the device powered on, then there is no problem and it works really good
10:04:50 Join lightbulbjim [0] (
10:05:57lucentZagor: could whatever you bumped be duplicated to run in the bootloader?
10:06:03lucentor is that nonsense
10:06:26 Quit tarbo (Read error: 104 (Connection reset by peer))
10:06:40 Join tarbo [0] (n=me@unaffiliated/tarbo)
10:06:51Zagorlucent: does the bootloader panic/crash with SD inserted?
10:07:18lucentZagor: I am not certain, I think it might be the firmware or the bootloader
10:07:26lucentZagor: how to know which one it is?
10:08:00Zagorlucent: depends on how the error manifests. does it freeze, reboot or panic?
10:09:41Zagorin any case, the clock change is merely hiding the problem. the proper fix is to make SD work with any clock spped.
10:10:57 Join BigBambi [0] (i=86ceaf37@rockbox/staff/BigBambi)
10:11:11lucentreally don't know what the bug is event
10:11:24Zagorwhat does the panic say?
10:11:27lucenteven*... kugel has 4gb fuze and says none of these problems bite him
10:11:42lucentZagor: I'm almost certain it's a failure of the CID command
10:11:44Zagorand I have the 2GB... fun fun :-)
10:11:48lucentI can verify in a minute
10:12:23 Quit tarbo (Read error: 104 (Connection reset by peer))
10:12:37 Join tarbo [0] (n=me@unaffiliated/tarbo)
10:14:08pixelmaZagor: I'm a bit confused as the description and the player-type listing of FS #9797 only mention "coldfire" but from looking at the changed files it seems like it's only for coldfire Irivers. What's correct now?
10:14:48Zagorpixelma: iriver is correct. I'll fix the description.
10:15:05Zagorfirst comment: "The patch actually only affects iriver currently. I'll update it for iAudio if anyone wants to test on those aswell."
10:15:51pixelmawell, I saw an effect already...
10:16:26pixelmaand didn't read close enough at half past one or so ;\
10:16:45 Join gartral [0] (
10:17:22gartralstats.rock in application does not scroll if the text goes "off" the side of the screen
10:17:47 Quit SUSaiyan (Read error: 54 (Connection reset by peer))
10:18:15 Join SUSaiyan` [0] (
10:18:44Zagorpixelma: that effect is probably because you are running with the new slower speed #defines but with code that actually sets other speeds. I would expect that to cause odd behaviour.
10:18:49 Quit tarbo (Read error: 104 (Connection reset by peer))
10:19:03 Join tarbo [0] (n=me@unaffiliated/tarbo)
10:22:22lucentZagor: -5 is the result from sd_card_init
10:22:43lucentsd_init_card() I meant
10:22:52lucent"card not initialised"
10:23:00 Quit nuonguy ("This computer has gone to sleep")
10:25:10 Quit tarbo (Read error: 54 (Connection reset by peer))
10:25:23 Join tarbo [0] (n=me@unaffiliated/tarbo)
10:30:01*lucent notes "sleepy time"
10:30:18 Quit kachna|lappy (No route to host)
10:31:16*Zagor is not thrilled about all the #defines in storage.h
10:31:38 Quit tarbo (Read error: 131 (Connection reset by peer))
10:31:51 Join tarbo [0] (n=me@unaffiliated/tarbo)
10:32:48B4gderisn't that to cause as little memory overhead as possible?
10:32:48 Quit Thundercloud (Remote closed the connection)
10:33:45ZagorB4gder: why don't simply rename ata_init to storage_init?
10:34:22ZagorI was rather flummoxed by not finding storage_init when greping all c files
10:34:49Zagorthat's how we do all other target-specific implementations. system_init, set_cpu_frequency etc.
10:36:02linuxstbgevaerts is the person to talk to, but wasn't the idea to be able to support multiple types of storage on the same target?
10:36:49B4gderI think so, yes
10:37:17Zagorlinuxstb: that will require refactoring too. I should look at gevaerts' patch to see how he does that.
10:37:18 Quit tarbo (Read error: 104 (Connection reset by peer))
10:37:31 Join tarbo [0] (n=me@unaffiliated/tarbo)
10:40:07 Quit daurn (Remote closed the connection)
10:40:10Zagorlucent: are you really getting a panic, or an "ATA error"?
10:40:29 Quit daurnimator (Client Quit)
10:42:17Zagorif it's a panic, it is in the application and not the bootloader. which makes sense since the bootloader is not multivolume enabled.
10:43:37 Quit tarbo (Read error: 104 (Connection reset by peer))
10:43:52 Join tarbo [0] (n=me@unaffiliated/tarbo)
10:45:34 Join Darksair [0] (n=user@
10:49:58 Quit tarbo (Connection reset by peer)
10:50:11 Join tarbo [0] (n=me@unaffiliated/tarbo)
10:54:15ZagorI wonder if the as3525 PLL setup is correct
10:56:18 Quit tarbo (Connection reset by peer)
10:56:32 Join tarbo [0] (n=me@unaffiliated/tarbo)
11:02:05 Quit tarbo (Connection reset by peer)
11:02:18 Join tarbo [0] (n=me@unaffiliated/tarbo)
11:08:33 Quit tarbo (Read error: 104 (Connection reset by peer))
11:08:47 Join tarbo [0] (n=me@unaffiliated/tarbo)
11:15:08 Quit tarbo (Read error: 104 (Connection reset by peer))
11:15:20 Join tarbo [0] (n=me@unaffiliated/tarbo)
11:17:40Unhelpfulhrm... we have sim plugins building with rb aliased to __header.api, i can verify that by running under gdb and getting the addresses of both after the plugin's been loaded, before it's called
11:17:45Unhelpfulbut, the plugin crashes. :/
11:20:56 Join tajlero [0] (
11:21:43 Quit tarbo (Read error: 104 (Connection reset by peer))
11:22:00 Join tarbo [0] (n=me@unaffiliated/tarbo)
11:22:20gartralis there a WPS tag for displaying an alternitive file, or say lyrics off of the comments?
11:24:33Unhelpfuli'm pretty sure there's not, but has most of what you'd need to know about wps tags.
11:25:37gartral%iC will do nicely
11:27:56 Join robin0800 [0] (
11:28:13 Quit tarbo (Read error: 104 (Connection reset by peer))
11:28:25 Join tarbo [0] (n=me@unaffiliated/tarbo)
11:29:48 Join kachna|lappy [0] (
11:34:25 Quit tarbo (Read error: 104 (Connection reset by peer))
11:34:37 Join tarbo [0] (n=me@unaffiliated/tarbo)
11:34:53 Join flydutch [0] (
11:36:13 Quit tarbo (Client Quit)
11:39:55 Quit Alystair (Read error: 60 (Operation timed out))
11:39:57 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
11:42:14 Quit ze (Read error: 60 (Operation timed out))
11:42:18 Join tarbo [0] (n=me@unaffiliated/tarbo)
11:42:36gartralunder notes in the custom WPS wiki page, there is a very confusing entry... it say "The image tag must be on its own line. " but theres no real reference
11:43:59 Quit linuxstb (Read error: 110 (Connection timed out))
11:48:12 Quit tarbo (Read error: 104 (Connection reset by peer))
11:48:25 Join tarbo [0] (n=me@unaffiliated/tarbo)
11:54:04 Join culture [0] (
11:54:10 Quit tarbo (Read error: 104 (Connection reset by peer))
11:54:24 Join tarbo [0] (n=me@unaffiliated/tarbo)
11:54:28 Quit cool_walking_ (Remote closed the connection)
11:54:45 Quit tarbo (Client Quit)
12:00:11***Saving seen data "./dancer.seen"
12:02:47 Join agaffney [0] (n=agaffney@gentoo/developer/agaffney)
12:03:23 Join ze [0] (i=ze@
12:12:50 Join kugel [0] (n=chatzill@unaffiliated/kugel)
12:19:34*kugel is curious about lucent's fixed microsd
12:22:40 Quit kachna|lappy (Read error: 113 (No route to host))
12:28:25 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
12:36:27 Quit XavierGr ()
12:37:22 Join Xerion_ [0] (
12:38:23 Quit robin0800 (Success)
12:40:03 Join moos [0] (i=Mustapha@rockbox/staff/moos)
12:46:00 Part gartral
12:51:30 Quit BigBambi (" ajax IRC Client")
12:52:49 Quit tajlero ("Leaving")
12:53:52 Quit Xerion (Connection timed out)
12:53:53 Nick Xerion_ is now known as Xerion (
12:55:05 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
12:56:05 Nick fxb__ is now known as fxb (
13:00:28 Join robin0800 [0] (
13:06:50 Quit robin0800 (Remote closed the connection)
13:07:56 Quit linuxstb (Read error: 113 (No route to host))
13:08:41 Join robin0800 [0] (
13:08:50 Join tajlero [0] (
13:09:09 Join tarbo [0] (n=me@unaffiliated/tarbo)
13:11:53 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
13:15:00 Quit robin0800 (Read error: 104 (Connection reset by peer))
13:15:55 Join robin0800 [0] (
13:22:28 Quit tajlero ("Leaving")
13:23:09 Join miepchen^schlaf [0] (
13:23:29 Quit PaulJam (Read error: 60 (Operation timed out))
13:25:45 Quit robin0800 (Read error: 104 (Connection reset by peer))
13:29:19 Join Seed [0] (
13:29:23 Join AndyIL [0] (i=AndyI@
13:33:57 Join tajlero [0] (
13:39:20 Nick tajlero is now known as thisistajlero (
13:39:53 Part B4gder
13:40:44 Quit AndyI (Read error: 110 (Connection timed out))
13:40:47 Join PaulJam [0] (
13:42:38 Quit kugel ("ChatZilla 0.9.84-rdmsoft [XULRunner]")
13:46:30 Join robin0800 [0] (
13:57:01 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
13:59:23 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
14:00:14***Saving seen data "./dancer.seen"
14:04:40 Join nplus [0] (n=nplus@243.131.Globcom.Net)
14:06:23 Quit bmbl ("Woah!")
14:06:35 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
14:07:26 Quit Horscht (Client Quit)
14:14:31 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
14:29:18 Quit XavierGr ()
14:30:17 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
14:31:51Zagordid flyspray rss never work?
14:32:23 Quit FlynDice ()
14:34:39Zagorhmm, it does work but my rss readers barf on the ^L chars sprinkled in it
14:40:49 Quit timc`` ("Leaving")
14:42:48JdGordonLinusN: hey, you round?
14:45:43GodEater_Zagor: works fine for me, I use Google Reader for my RSS client
14:46:06 Quit Zagor ("Client exiting")
14:46:20 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
14:46:43Zagormemo to self: emacs escape key does not work in xchat
14:47:59ZagorGodEater: yeah it turns out there was just a single task with ^L in it, which confused 'liferea' that I decided to try
14:48:16ZagorI simply want a small popup when tasks are added or modified
14:48:51JdGordonpixelma: you round?
14:49:21 Join itcheg [0] (i=41d59de2@gateway/web/ajax/
14:49:30jhMikeSgevaerts: Is there a time limit specified between getting the bus reset and continuing communication? It seems to work just intially enabling the driver for the reset and then enabling other interrupts after everthing acked, diconnecting otherwise if enough time passes.
14:55:25JdGordonlinuxstb: was it you who said you use one of the sims as your audio player on the comp?
15:05:47JdGordonLinusN: (or anyone...) I need to fix up hwcodec mpeg.c to send an event when the next tracks' id3 info is available. FS #9795 tries to do this but apparently it doesnt work.. Can you have a quick look and see if there is a better place I should be puttting this send_event() ?
15:06:49Zagoranyone want to guess what the "EAH2" component next to the cpu in this pic is?
15:09:20jhMikeSIs that a +/- on it?
15:09:51jhMikeScould be a tiny battery/cap?
15:10:06Zagorit looks like a -, yes
15:11:47 Join CaptainKewl [0] (
15:17:41 Join LambdaCalculus37 [0] (i=44a04303@rockbox/staff/LambdaCalculus37)
15:18:30 Quit thisistajlero ("Leaving")
15:22:10 Join tajlero [0] (
15:22:30 Nick tajlero is now known as rocko (
15:22:43 Nick rocko is now known as rocko_ (
15:24:19 Join akahhannah [0] (n=be190c9b@gateway/web/cgi-irc/
15:25:57 Nick rocko_ is now known as rockos (
15:26:00robin0800Zagor: I think they are heads of screws holdibg it in place
15:26:44Zagoryou don't use screws for such small components
15:28:52Zagorwhat I'm fishing after is if there is an external clock source on the board. because if there isn't and we are using the internal 24 MHz clock, then it looks like we are boosing at 124 MHz instead of 248.
15:29:25akahhannahJdGordon: hey! I have a question about your bmp numbers patch
15:29:57robin0800Zagor: sorry battery yes google
15:32:22Zagorwell those hits aren't exactly conclusive...
15:32:33 Quit Horscht ("")
15:32:59JdGordonakahhannah: dont ask to ask.... just ask
15:34:22akahhannahok just wanted to know if you were there... well I have been only able to use on the ipod video and mini 2g sims
15:35:23akahhannahas i did put on the tracker and the forum, I get a data abort crash on the actual targets
15:36:21JdGordonright... like i said in the forum... i've abandoned it for the time being...
15:36:46 Join archivator [0] (i=foobar@
15:37:14 Nick archivator is now known as Guest42730 (i=foobar@
15:38:01 Nick Guest42730 is now known as archivator (i=foobar@
15:38:07akahhannahoh... ok I didn't realize it then.. I still think it's great, hope it makes it through at some point... thanks
15:38:49LinusNZagor: that definitely looks like an oscillator
15:42:11 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
15:43:09ZagorLinusN: ok
15:43:18 Quit akahhannah ("CGI:IRC (Ping timeout)")
15:46:10PaulJamZagor: with the zero wait boost patch that sets the normal frequency to 56MHz i haven't noticed any audio droputs. but i'm not sure it this is useful for testing stability of clock switching, because even though i have enabled crossfeed, eq and replaygain, the cpu doesn't seem to boost at all with mp3 files.
15:47:54ZagorPaulJam: good point. although it always boosts on buffer refill, so it gets at least a little excercise anyway
15:48:48ZagorI think the pcm starve flag would be interesting
15:53:01PaulJamhmm, maybe i can create some aac files. they should definitely need boosting.
15:53:21ZagorPaulJam: ogg is quite demanding too
15:54:58 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
16:00:15***Saving seen data "./dancer.seen"
16:08:15 Quit Darksair (Connection timed out)
16:08:54 Join n17ikh [0] (
16:18:04 Quit n17ikh|Lappy (Connection timed out)
16:23:11Zagorthe more I think about boosting, the more sense it makes to run ALL code boosted and then go to sleep when no code is running
16:24:02 Join B4gder [241] (n=daniel@rockbox/developer/bagder)
16:24:52ZagorSince the cores are static, power is spent by the cycle. So if we make sure we don't do boosted idling/waiting, toggling all/nothing is more efficient than all/some.
16:26:52 Join timc`` [0] (n=aoeu@
16:29:43 Join toffe82 [0] (
16:29:56 Join tyfoo [0] (
16:32:04ZagorOn the other hand I haven't looked into what kind of sleep and wakeup options we have on the different chips. Very Low Clock could of course do the same trick.
16:38:08 Join MethoS [0] (
16:47:39 Join kugel [0] (n=chatzill@unaffiliated/kugel)
16:49:50 Join midkay_ [0] (
16:50:02 Quit ajb (Remote closed the connection)
16:50:12 Join ajb [0] (
16:51:41 Join Horschti [0] (
16:51:57 Quit Horscht (Nick collision from services.)
16:54:54kugelLinusN: hey, I'm going to add the setting for "natural sorting"
16:55:31kugelI'm just curious about how to call it, I don't really like natural
16:56:06B4gder"assumption based" ? B)
16:56:35 Join Rob2222 [0] (
16:56:49rasherHuman sorting?
16:57:19kadoban"natural sorting" is the only thing i've ever heard it called
16:57:29kugelI'd rather call it "smart sorting", since no matter of the personal preferences, it's sorting smarter (i.e. by some algorithm vs plainly looking up the ascii table)
16:57:32 Quit Rob2223 (Read error: 104 (Connection reset by peer))
16:57:37kugelor human
16:57:58kugel"human understandable" even? ;)
16:58:00rasherYeah, I've never heard it called anything but natural sort
16:58:12Zagorfun fact: ARM922T draws 250 picowattseconds per cycle
16:58:21kugelerr, "non-engineerer understandable"
16:58:56kugelrasher, kadoban: I just think "the nature doesn't sort"
16:59:10B4gdernatural is a weird term for it
16:59:24kadobanyeah, it's kind of weird...but it's already known as that :)
16:59:31rasherkugel: I agree, and I'm not terribly opposed to calling it something else
16:59:39rasherIt's not like it's a standard or anything
16:59:51B4gderI don't have any strong opinion either
16:59:56kugelyo, I'm just hoing to find an agreement on the name
17:00:10kadobani think any that have been suggested would probably be fine...
17:00:16kugelso, smart and human are my favorites
17:00:20B4gderI won't use it anyway :-)
17:01:02 Join Jaykay [0] (
17:01:26Zagorwhat does it do? (/me crawls out from under his rock)
17:01:42*rasher really thinks it should default to enabled though
17:02:08kugelrasher: well, that's very controversial
17:02:09B4gderZagor: sorts 2 before 10
17:02:51kugelI'd also have it as default. but on the other hand, it's likely to change behavior without people doing anything but updating (i.e. if they didn't change the sort setting at all)
17:03:17Zagoris that the same as /usr/bin/sort calls "numeric sort"?
17:03:18rasherkugel: I know. Doesn't change my opinion though. For the average, new user (which is what defaults are for), he's not likely to understand why 2 is sorted after 10, whereas someone who actually *wants* that behaviour, will know what's going on
17:03:25B4gderZagor: I'd guess so
17:03:27kugeltherefore I'd rather not change the defaults, just to not unexpectedly change things
17:03:31kadobanZagor: yes
17:03:34Zagorthen I vote for that name
17:04:00*rasher thinks that's a bit weak
17:04:17rasherThe defaults should be sane, even if changing them might change the settings for some people
17:04:59 Quit Zagor ("Client exiting")
17:05:07rasherThey can always change them back if they prefer asciibetical sorting
17:05:31kugelbut even the average joe that used rockbox for some time is used to the old sorting, and will be curious (even if he likes the new sorting)
17:05:51rasherSo be it, says I
17:06:31LinusNi vote for numerical sort as the default
17:07:19 Quit B4gder ("It is time to say moo")
17:07:29*kugel remembers the last discussion on this topic :/
17:07:47LinusNi still wake up at night because of it
17:07:56kugelI'm fine with either
17:08:03 Quit midkay (Read error: 110 (Connection timed out))
17:08:31rasherCan we agree that when chosing a default, we should consider new users, and in particular, new less clued users?
17:08:41 Join webguest56 [0] (n=3a6a711d@gateway/web/cgi-irc/
17:08:45LinusNrasher: yup!
17:09:21LinusNthat's what "sane defaults" is to me
17:09:25rasherI can't imagine that that group of users would prefer ascii-sorting
17:09:45 Part LinusN
17:09:47rasherPreaching to the choir here
17:09:54 Quit webguest56 (Client Quit)
17:10:54*kugel isn't really happy with "numeric sorting"
17:11:16kugelsounds like it would only sort numeric filenames, the rest is random
17:11:25Jaykayi suggest closing of, cheats for doom. its incomplete, no work for nearly 2 years, doesnt patch, and cheats are usually reducing the fun after some time...
17:11:27 Quit Horschti ("I got raided by the FBI and all i got is this lousy quit message")
17:11:36Jaykaysorry for breaking in your discussion ;)
17:15:07LambdaCalculus37Jaykay: Have you tried it recently?
17:15:38Jaykaypatching? yes, 2 of 6 hunks failed... should i try to sync it?
17:16:12 Join MethoS- [0] (
17:16:13LambdaCalculus37Yes, try syncing it and see what it does. If you're registered on Flyspray, it would also be a good idea to comment on the patch as well.
17:16:15 Join nuonguy [0] (
17:16:42LambdaCalculus37If the patch still fails after syncing, I'll go ahead and close it. It's rotting in the tracker, anyway.
17:17:30 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
17:19:14LambdaCalculus37Speaking of bit rot, there are quite a lot of very old patches in the tracker that should either be closed or tested to see if they sync with current SVN.
17:22:10JaykayLambdaCalculus: should i just apply the failed changes in the .rej? and the "rotten" patches, i could test them. if you had some patches in your mind (?), tell me....
17:23:52LambdaCalculus37Jaykay: I think it's safe to say that the Doom cheats patch is not going to work at all with current SVN.
17:25:17Jaykayok i synced it succesfully but its not working anyway ;) why should i test it when its safe that......
17:25:45Jaykay(actually i didnt manage it, i dont understand the .rej)
17:28:18 Quit MethoS (Read error: 113 (No route to host))
17:29:31Jaykaylambdacalculus37: sorry what should i do now?
17:31:20 Join Xerion_ [0] (
17:33:36kugelrasher: am I supposed to add new lang strings at the end of the .lang, or doesn't that really matter?
17:34:19rasherkugel: I'm pretty sure you are, yes - only matters for english.lang though, iiuc
17:35:39Jaykayrasher: is it possible to use the form at the bottom of"> anywhere with own lang files?
17:35:52rasheronly if I code it up
17:36:56LambdaCalculus37Jaykay: Can you show me the output you're getting from the patching with a pastebin?
17:39:10kugelLambdaCalculus37: "Stripping traling CR"?
17:39:30kugelall those paste sites add CRs to the paste, for some reason
17:40:44Jaykaylambdacalculus37: i dont get the text out of vmware, im sorry
17:41:51Jaykayhow can i store text copied with strg-c in a txt file?
17:41:55*kugel suggests using virtual box
17:42:11*LambdaCalculus37 seconds kugel's suggestion
17:42:39 Join evilnick [0] (i=0c140464@gateway/web/ajax/
17:43:12Jaykayi tried this some time ago, somehow it didnt work
17:43:37Jaykayiirc the image was only for vmware and i didnt know how to convert it
17:43:49kugelwell, set it up yourself
17:44:04kugelinstall virtual box, install ubuntu, run
17:44:32kugeleach step is very user friendly
17:45:03 Quit petur ("work->recording")
17:45:05Jaykaywhat is
17:45:29rasherIt compiles the crosscompilers you need to compile Rockbox
17:45:33kugelthe tool that automatically builds the cross compilers
17:47:30 Quit Xerion (Connection timed out)
17:47:31 Nick Xerion_ is now known as Xerion (
17:49:12Jaykaytoo lazy, im sorry. can anybody say me how to store the text copied with strg-c in a file?
17:57:30 Quit rockos ("Leaving")
17:59:23 Quit Horscht ("I am root. If you see me laughing, you better have a backup")
18:00:16***Saving seen data "./dancer.seen"
18:04:28kugelJaykay: sorry, too lazy
18:09:45 Join tajlero [0] (
18:11:05 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:15:10 Quit tajlero ("Leaving")
18:15:38Jaykaykugel: the only difference is that it would take a few seconds for you...
18:17:17 Quit domonoky ("Leaving.")
18:18:35 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:42:41 Quit Seed ("cu, Andre")
18:45:40 Quit robin0800 (Read error: 104 (Connection reset by peer))
18:47:06 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
18:49:53 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
18:51:47 Join mrkiko [0] (
18:56:54 Quit toffe82 (Read error: 110 (Connection timed out))
18:58:22 Quit mrkiko ("leaving")
19:08:56 Quit Jaykay ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
19:09:21 Join toffe82 [0] (
19:10:18kugelBagder: is it possible to get rockbox-info.txt from make bin too?
19:10:52kugelsucks a bit for my slow laptop to build everything if I only want to check bin&ram sizes
19:11:37kugelhm, maybe it's not possible without making LANGs at least?
19:22:08lucentkugel: hi there
19:22:39lucentkugel: I can't explain why Zagor's commit fixed microSD after power on, but it does
19:24:09lucentadding a couple of delays to the inner while() loop for acmd41 thing, now everything is stable on power-on and during operation for external microSD insert / eject
19:24:11 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
19:25:00kugellucent: well, who knows what the piece of code before his commit did
19:25:20lucentoh? was it very weird?
19:25:24kugelI thought no more delays are needed?
19:25:37lucentno delays are needed if the device is already booted up and powered
19:25:44kugelno idea, that's the point ;) seems it wasn't doing good things if it broke some microsds
19:25:58lucentI still get a panic if microSD is inserted on power up
19:26:08kugelwell, have you updated the bootloader?
19:26:22lucentbootloader was not touched by Zagor's commit
19:26:27kugelit was
19:26:34lucentthat's in his words, I asked him... so I don't know
19:26:54kugelsurely it's touched by the commit
19:27:35kugelbut the bootloader doesn't support microsds (and thus should look for those) so maybe he meant that
19:27:59lucentoh, maybe
19:28:47lucentI uploaded the bootloader anyways, and it does panic with microSD connected
19:28:59lucentbut I don't think it's the bootloader that panics
19:29:07 Join gregorovius [0] (
19:29:19lucentit should be the firmware that is panic'ing
19:29:59kugelwait a second
19:32:23lucentI'm not understanding how to say "we are bootloader" or "we are firmware" when I see a panicf()
19:33:37lucentit's the same panic as I always know on 8gb fuze, sd_init_card() comes back with value -5 and the struct is not filled out completely so the code does panicf() not initialised
19:34:18 Quit linuxstb (Read error: 110 (Connection timed out))
19:36:00kugellucent: try this
19:36:37 Join BXCracer [0] (n=bxcracer@
19:36:43lucentokay, it will be a moment while I test
19:36:50kugelthis causes rockbox to not init the sd at startup, but slightly later
19:37:23kugelslightly later, as in once the sd thread is running
19:37:29kugelsd tick task*
19:38:03 Quit preglow (Read error: 110 (Connection timed out))
19:39:57lucentkugel: update my bootloader with this? firmware? firmware + bootloader?
19:41:10kugelyea, for the sake of making sure
19:41:12 Join The-Compiler [0] (n=dsmoddin@unaffiliated/the-compiler)
19:42:10 Join webguest97 [0] (n=cd7c35fe@gateway/web/cgi-irc/
19:42:37 Quit MethoS- (Read error: 104 (Connection reset by peer))
19:42:58 Quit timc`` (Read error: 110 (Connection timed out))
19:43:23 Join timc`` [0] (n=aoeu@
19:43:25 Quit webguest97 (Client Quit)
19:44:29 Join Thundercloud [0] (
19:44:40 Quit TMM (Read error: 104 (Connection reset by peer))
19:46:32 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
19:47:03lucentkugel: without microSD inserted, it boots fine
19:47:21lucentkugel: with microSD inserted, it doesn't boot past ROCKbox graphic screen
19:48:22 Join MTee [0] (n=mtee@
19:48:37saratogado floating point operations on the beast use software emulation or the VFP?
19:48:53 Join TMM [0] (
19:49:51 Quit japc (Read error: 60 (Operation timed out))
19:57:29kugellucent: and if you insert after boot?
19:57:38lucentkugel: works good if I insert after boot
19:57:47amiconnsaratoga: I would think they use software emulation
19:57:59*amiconn sincerely doubts that gcc has VFP support
19:59:45amiconnsaratoga: Btw, the problem with putting codecs on the cop is not only thread safeness, but also cache coherency between CPU and COP
20:00:20***Saving seen data "./dancer.seen"
20:00:23saratogaamiconn: isn't VFP the main FP model in ARM?
20:00:28saratogawhy wouldn't gcc support it
20:00:47saratogaalso, yes I realize the coherrency issue, but thats much easier to solve then thread safety
20:00:59amiconnNo, it's an extension
20:03:02Bagderarm-elf-gcc -mfpu=vfp
20:05:00 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
20:05:21kugellucent: weird that it hangs
20:07:06saratogaBagder: so fp codecs should work on the beast?
20:07:56lucentkugel: I can make r19773 work on boot by adding some delays to the while() loop with acmd43
20:08:05Bagdersaratoga: afaiu, yes
20:08:37Bagderhm, possibly not in the 4.0.x series though
20:09:44amiconnFixed in 4.1.1, aha
20:11:21kugellucent: I noticed that
20:11:41*amiconn isn't sure whether using VFP is generally faster than emulated FP on arm11
20:12:27saratogait seems the newer gcc versions have a lot of improvements for advanced arm
20:12:34amiconnVFP should speed up vector operations a lot (hence the name), but for single operations the overhead might be significant
20:13:21amiconnA test should tell...
20:13:22saratogai think it implements standard ieee754 scalar fp too
20:13:32kugelI've managed to build rockbox with arm-elf-gcc v4.1.x, so a compiler change may could be considered (and manageable) I guess?
20:14:28kugelwithout editing rockbox, I might add
20:15:57amiconnDid it work properly?
20:16:11saratogapeople have benchmarked codecs on PP in gcc 4.1.x before
20:16:36saratogano better then 4.0.x, but for armv5+ that might not be so
20:16:46amiconnEarlier gcc version switches sometimes had the problem that the build(s) compiled ok, but the resulting binaries crashed
20:16:57 Join sarixe [0] (
20:17:07Lloreankugel: You'd basically need to test bootloaders and binaries on all ARM targets with gcc 4.1.x
20:17:33kugelamiconn: I didn't test every part, but it booted and played music fine (that's what I remember)
20:17:48kugelLlorean: yea, that's true
20:18:12saratogacouldn't we just use a different gcc for the beast and newer targets?
20:18:21LloreanWe've seen some strangeness with GCC versions. The Nano bootloader acted very strange back in the early days which is why we ended up using a different GCC than iPL
20:18:21kugeljust saying, that it may work, if the benefits are worth it
20:19:13Lloreansaratoga: I'd say that should more be a "last resort" than "first plan" :)
20:19:14 Join karashata [0] (n=karashat@
20:19:28LloreanBut it may also be a valid temporary measure if it really does offer measurable benefits
20:19:34saratogatheres posts on rockbox-dev about using gcc 4.2
20:19:43amiconnsaratoga: I'd rather want to reduce gcc version diversity that increase it
20:20:19kugelsaratoga: I remember there was also a tracker entry about it
20:20:20amiconnSo if we do a switch, we should check whether it's possible/ desirable to switch the other architectures as well
20:20:32saratogaamiconn: yeah that makes sense
20:20:55amiconnRight now we're using 2 versions, 4.0.3 for ARM and SH1, and 3.4.6 for coldifire
20:21:13lucentwhy 3.4.6 for coldfire?
20:21:26amiconnI know that 4.0.3 works for coldfire too, but that switch wasn't done because there were some performance reductions observed
20:21:34Lloreansaratoga: It really needs a _lot_ of testing, unfortunately. If I recall, with 2.95 the Rockbox binary would work (with a hacked bootloader to remove dual boot) on the Nano, the 4G (or was it 5G) bootloader and binary worked fine, but if you left the "if" in for detecting the button press for dual boot on the nano, the bootloader failed completely
20:21:48LloreanIt was very strange. Basically there have been some very target-specific quirks with GCC versions in the past.
20:21:56amiconnI couldn't reproduce them, in my tests almost everything performed the same, and wavpack saw a slight speedup
20:22:27amiconnNewer versions than 4.0.x apparently don't work for coldfire
20:22:46amiconn(weird binutils <-> gcc parameter clashes)
20:23:20 Join tyfoo2 [0] (
20:23:55amiconnI didn't try newer gcc for arm or sh1 yet
20:24:23 Join bertrik [0] (
20:24:28saratogatheres a message in mailing list from Nils saying 4.3 works with some effort
20:24:33saratogafor coldfire that is
20:24:46 Quit tyfoo (Read error: 60 (Operation timed out))
20:24:46 Nick tyfoo2 is now known as tyfoo (
20:27:34saratogaoh wow FS #7832 - Building with gcc 4.3
20:30:00amiconnHmm, already more than a year old
20:30:09amiconnBut it's coldfire only.
20:30:21amiconnNow that binutils 2.19 are out, it should be easier
20:30:22saratogathe changes look pretty simple, and one of them [tremor floor0] should be commited regardless
20:30:49amiconnSomeone should try binutils 2.19 + gcc 4.3.2 for all architectures
20:31:52saratoga"Some quick testing (with a GCC snapshot from a couple of weeks ago) suggests that AAC is much faster (about 30%) while Vorbis is slower (10-15%). MP3 is pretty much the same. "
20:32:09saratogathrough i bet anything thats due to the weird fft AAC no longer uses
20:32:43amiconnIt might also need -O tweaking
20:33:12amiconnFrom what I've seen so far, -O3 is sometimes significantly slower than -O2
20:34:31saratogais building a different gcc as easy as changing the version numbers in
20:35:12amiconnIt depends. If gcc needs patches for proper rockbox support, it's more than that
20:36:19*amiconn would build gcc manually for those tests, adapting patches as necessary, and do the necessary changes to only when everything works
20:37:06*amiconn is still missing a standard test track which has a free license and comes as lossless file
20:37:21kugellucent: please test
20:38:24 Join gregzx [0] (
20:38:43saratogaamiconn: what about the one in CodecPerformanceComparison
20:40:39amiconnOh, didn't notice there was a link...
20:40:56*amiconn thinks that wiki page is not very helpful though
20:41:21amiconnThe performance tables are too detailed and not easily comparable
20:41:38 Join at0m [0] (n=a548c80b@gateway/web/cgi-irc/
20:46:43amiconnAlso some audio formats are missing: ac3, speex, alac (probably due to lack of free encoder), shorten, ape
20:48:32lucentkugel: okay, testing now
20:49:10gevaertsjhMikeS: I don't know
20:50:31lucentkugel: (boot without microSD and then insert after boot) = OK
20:50:51lucentkugel: (boot with microSD connected) = PANIC card not inserted
20:51:09lucentkugel: oh wait, I messed up
20:51:34 Join Seed [0] (
20:52:00lucentkugel: I just tested plain SVN r19773, gotta retry with your patch
20:54:05lucentkugel: bootloader won't build with your patch applied
20:54:52lucentata_sd_as3525.c:(.text.sd_init+0xc0): undefined reference to `sd_enable'
20:59:22kugellucent: yea, don't build the bootloader :p
20:59:50lucentkugel: holy ... hey guess, this is crazy... now my wheel light is working with your patch
20:59:56*lucent looks confused
21:00:39lucentkugel: microSD panic's on boot though, works fine otherwise
21:00:53lucentno change
21:00:57lucenterror is -5
21:03:08 Quit Acksaw ("Erection reset by queer")
21:03:48lucentkugel: I've not seen wheel light work in rockbox since a long long time
21:05:01lucentkugel: I'm distracted by this. Let me describe... wheel light turns on when I tap power button momentary, it turns on, and the timeout works, it turns back on after the timeout if I use the wheel or any button
21:05:16lucentafter backlight powers off for the LCD, the wheel light is dark and does not activate
21:05:24 Join jgarvey [0] (
21:05:27lucentif I hit power button, it starts working again
21:06:32lucentmaybe I describe some things wrong
21:06:40lucentany button press makes the wheel work
21:06:53lucenterr... any button press makes the wheel *light* work :)
21:07:27kugellucent: yea, buttonlight doesn't seem to work without backlight it seems
21:08:05lucentkugel: does the code turn on the wheel light when moving the wheel?
21:08:40 Join PaulJam_ [0] (
21:09:19kugellucent: not sure, the wheel is a bit special
21:09:27lucentsorry to get distracted
21:09:34 Quit obo (Read error: 110 (Connection timed out))
21:09:39lucentmicroSD panic returns -5 if microSD is connected on boot
21:10:43 Quit jhulst (Read error: 110 (Connection timed out))
21:10:51lucent"card not initialised (-5)"
21:14:56kugellucent: but it still works when inserting after boot?
21:16:04lucentkugel: yes it does, after boot is okay
21:17:06lucentkugel: microSD insert/eject after boot, this is now reliable since r19773
21:17:09 Quit Horscht ("User was distributing pornography on server; system seized by FBI")
21:17:13kugellucent: ok, then show me how you inserted the delays, and then we look if we can commit this
21:17:27 Join Acksaw [0] (
21:17:31lucentokay I'll make a patch to share
21:17:48lucentkugel: but your patch you just made "fixes" buttonlight on 8gb fuze
21:17:57lucentI'm going to be sad to see it go away =P
21:18:15lucentbuttonlight/wheel light
21:18:25kugellucent: well, I'm doing buttonlight work parallel, seems the patch contained parts of it
21:18:43 Quit GodEater_ (Remote closed the connection)
21:18:43 Quit LambdaCalculus37 (Remote closed the connection)
21:18:43 Quit evilnick (Remote closed the connection)
21:19:26 Join evilnick [0] (i=0c140464@gateway/web/ajax/
21:19:51 Join Strife89 [0] (n=michael@
21:20:48lucentI have not regression tested since the fixed boost behavior revision in SVN
21:20:57lucentso... I tried what worked before
21:21:13lucentit works now but I haven't tried i.e. only one delay or moving delay around
21:22:25kugellucent: better to much delays than to less. they're at init only anyway
21:22:46 Join LambdaCalculus37 [0] (i=44a04303@gateway/web/ajax/
21:23:09kugeli.e. not while doing transfers, so they don't really matter
21:23:14kugelbetter we play safe
21:24:40 Quit PaulJam (Read error: 110 (Connection timed out))
21:25:00lucentmy 2GB microSD card arrives today, I will test with normal class microSD then
21:25:17kugelsaratoga: I have a patch to enable backlight fading, sd driver clean ups, and fix buttonlight for fuze/ams sansa. You want seperate patches or is 1 ok?
21:25:40kugellucent: lets just insert the delays if they work
21:25:48lucentyes, they do work
21:25:55kugelthey really don't hurt
21:27:44 Join Aurix_Lexico [0] (
21:32:07lucentkugel: if (while rockbox is in the process of booting) I insert microSD, it panics
21:32:13lucentbut I think that's normal, right?
21:35:22kugellucent: why should it?
21:35:32kugelwith or without delay?
21:35:38kugelor delays
21:36:46lucentkugel: if you boot your 4gb fuze, and insert microSD immediately after the bootloader flashes, but just before the firmware loads
21:36:51lucentit would panic right?
21:37:07 Join Schmogel [0] (
21:37:41lucentbootloader graphic (flashes)
21:37:47kugelit doesn't panic at all here
21:38:29lucentoh, then I don't think the delay is a solution
21:38:58lucentI want 8gb fuze to work as well as 4gb fuze with microSD
21:39:11kugelI think it does with delays?
21:39:18kugelwhy are you so opposed to delays?
21:39:30kugelrockbox is full of those small delays
21:40:01lucentI want to do more testing so we are not hiding a different problem
21:40:13kugelit's clearly related to your card. and we want to have all cards working, so we add those delays
21:40:31lucentwhen I have a non-SDHC card, I will test these things again
21:40:42lucentit could be the 8gb fuze and not the card
21:40:51kugelso can you give me a patch with working delays anyways?
21:41:04kugelyou can still test if you want
21:41:05 Join akahhannah [0] (n=be190c9b@gateway/web/cgi-irc/
21:41:19 Join japc [0] (
21:41:40lucentit still panics if you mess with microSD card between bootloader and firmware loading
21:42:07lucentif you have microSD connected then there's no problem, or if you connect it after everything loads up, then it is also good
21:42:47 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
21:43:12kugellucent: did you insert the card fully or just shortly?
21:43:26kugelit still panics if you have the card inserted only very shortly
21:43:35akahhannahhello, anyone knows if it's possible to identify (display) in the WPS the config file loaded? maybe through the %St tag?
21:43:49kugeli.e. if you pull it out during init
21:43:56Lloreanakahhannah: Rockbox doesn't keep track of the .cfg file you last loaded.
21:44:05LloreanAll it cares about are the settings that are changed, not the filename.
21:44:44akahhannahok thanks
21:44:47kugellucent: well, thanks a lot for your effort
21:45:42lucentkugel: I apprecaite your time helping me walk through these efforts
21:46:21akahhannahLlorean: is there maybe a setting you could think of, that could be changed when loading a .cfg file, and then displayed using %St, and so identify the "settings group"?
21:46:43lucentkugel: before the fixes to boost init, I had 4 places where I inserted mci_delay()'s to make microSD insert / eject work
21:46:50lucentI'm testing those other extra delays now
21:47:04 Quit karashata ("G'bye everyone!")
21:47:05Lloreanakahhannah: If you have the equalizer disabled, you could always change some of the equalizer values since they're not in use anyway.
21:47:50LloreanWhatever you use, you'd have to make sure that *all* of your config files change it, because nothing will automatically reset when you change configs.
21:48:49akahhannahthat idea might work, thanks, I'll see if i can make it that way
21:48:51lucentkugel: I think those other delays fix panic on my abuse of insert/eject on boot :)
21:49:23lucentkugel: nevermind that
21:49:53 Quit Strife89 ("Packing for a trip.")
21:50:09lucentkugel: okay, I think there is another problem but I agree that committing the 2 extra mci_delay()'s in won't hurt anything
21:51:28 Quit akahhannah ("CGI:IRC (EOF)")
21:52:45 Quit Acksaw ("Erection reset by queer")
21:53:03 Join Acksaw [0] (
21:55:06 Quit jhulst (Read error: 104 (Connection reset by peer))
21:56:33 Quit LambdaCalculus37 (" ajax IRC Client")
21:59:26lucentkugel: maybe the device is initializing the internal SD and that's why inserting an external SD causes it to panic?
21:59:33lucentkugel: you said it doesn't happen on your device though
21:59:59kugelis it reproducable?
22:00:21***Saving seen data "./dancer.seen"
22:00:35 Join tvelocity [0] (
22:01:03kugelno idea, but it's possible
22:01:27lucentno delays added or removed affect it though
22:02:46 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
22:03:24 Quit The-Compiler (Read error: 104 (Connection reset by peer))
22:03:24kugelI got the buttonlight working ok
22:03:30*lucent cheers
22:06:32lucentkugel: one extra mci_delay() is sufficient, I'm going to try some more things
22:07:33 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
22:11:24 Quit nplus (Remote closed the connection)
22:20:33lucentkugel: I have a cleaner patch, I think, it only delays more if sdhc is true
22:22:14kugeladding dozens of ifs to make 1 tiny delay conditional isn't what I consider as cleaner
22:22:33lucentwell, I mean it is more obvious what I am trying to do
22:22:39kugelbut it adds 2 delays again? I thought 1 is enough?
22:22:49lucent1 delay is already there
22:22:59lucentI was making 1 delay into 3 delays before
22:23:03lucentthat's unnecessary
22:23:03kugeloh, sorry, I oversaw the "-"
22:23:32kugelwhy do you think it's only on SDHC?
22:23:45lucentwell, do you test with SDHC or regular card?
22:25:20lucentI have SDHC card here, but no regular card, so I am only making a guess. When my package arrives today I will test with regular card
22:26:12 Join Zagor [242] (
22:26:19lucentkugel: I'll be quiet now. I'm just wasting time, sorry
22:26:20saratogakugel: seperate patchs are strongly encouraged
22:27:06kugellucent: I have 4GB, so HC
22:27:10amiconnZagor: Running boosted all the time and sleeping would waste lots of energy vs. lowering the clock
22:27:41amiconnThe cpu core(s) might be fully static, but our SoCs contain a lot more than just the cpu core(s)
22:27:52kugelsaratoga: I thought so too, will take some minutes to seperate them though
22:28:02 Quit archivator ("<>")
22:28:02amiconnsaratoga did some interesting measurements regarding this
22:28:24 Join Jaykay [0] (
22:29:12amiconnThat's in fact another reason not to go too low with the unboosted clock
22:29:43amiconnHmm, and when thinking a bit more, it would also be a reason for multi-level boost even if the target doesn't have variable core voltage..
22:29:54JaykayZagor: should i do a battery bench with fs 9800?
22:30:26kugelgevaerts: can we use sd_enable for sd_spin and sd_sleep?
22:30:34amiconnHmm, or not. The consumption will be linear for the other parts of the SoC as well
22:30:38 Join miepchen^schla [0] (
22:30:52amiconnBut it's a reason to sleep the core as little as possible - reduce clock instead
22:30:57 Join webguest35 [0] (n=453ceedb@gateway/web/cgi-irc/
22:30:57saratogaamiconn: i don't think we ever explored the power consumption at various clock speeds
22:31:04saratogawe've just assumed its close to linear
22:31:07gevaertskugel: I'll answer as soon as I understand the question ;)
22:31:11saratogaso perhaps i should look into that
22:31:14Zagoramiconn: what is a reason not to sleep?
22:31:32 Quit bmbl ("Woah!")
22:31:34amiconnSleeping at high clock wastes energy in the internal peripherals of the SoC
22:31:39kugelgevaerts: I saw "void sd_spin(void)" and "void sd_sleep(void)" stubs in our sd driver
22:31:48Zagorand we don't clock the other parts of the soc. I'm talking about core clocking. if everything clocked up with the core, we wouldn't need waitstates.
22:32:02 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
22:32:12webguest35i was in here like 2 months ago...and I asked for a link to theme site
22:32:14Zagorhowever waitstates do mean that high speed is more expensive than low
22:32:28amiconnsaratoga did some measurements of PP power consumption boosted and unboosted, sleeping and busy
22:32:28Zagorso yes, generally I agree with you
22:32:35gevaertskugel: I still don't understand what you're getting at
22:32:38webguest35because the link thats on the main site doesn't work on mine because I have the new rockbox
22:32:41*gevaerts may be too tired
22:32:47amiconnPP draws more power when sleeping at 80MHz than when busy at 30MHz
22:32:47ZagorPP is a poor example, since we have zilch docs for it
22:32:59webguest35can anyone link me up to the new themes again?
22:33:06kugelgevaerts: we have sd_enable(bool on), so we can power down the sd controller if unused
22:33:36kugelI'm wondering if basically sd_spin() == sd_enable(true) and sd_sleep() == sd_enable(false)
22:34:52Zagoramiconn: I strongly suspect that is due to us not configuring it right. which will always be a problem with PP.
22:34:54gevaertsah ok. I'm not sure actually
22:35:27amiconnI would expect the other SoCs to behave similarly
22:35:38Zagoramiconn: I wouldn't
22:37:37kugelgevaerts: I'd think so if sd_sleep (which is storage_sleep later) is meant to save power
22:37:42Jaykaydoes anyone have a clue whether a bench with flac or mp3 is better?
22:37:44webguest35I need the newer themes the ones that work with the updated WPS
22:37:59webguest35does anyone know where that page went? I lost it
22:38:46gevaertskugel: sd_sleep() seems safe enough, but I'm not sure if sd_spin() is called before all operations
22:38:50kugelJaykay: what do you mean by better?
22:39:14kugelgevaerts: well, the transfer function calls sd_enable anyways
22:39:21gevaertswebguest35: have you tried the link "extras" on the website?
22:39:31domonokywebguest35: take a look at the wiki, WpsGallery
22:39:35kugeli.e. sd_enable(true) before transfering, and sd_enable(false) when finished
22:39:50Jaykaykugel: more meaningful (?) [aussagekräftiger]
22:40:05 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
22:40:08gevaertskugel: in that case, what would calling them again from sd_sleep() and sd_spin() gain you?
22:40:12kugelJaykay: neither of those are good for that, as they run unboosted
22:40:21kugelgevaerts: not sure, probably nothing
22:41:16amiconnata_spin() is for *keeping* the disk spinning. SPinup on access when it's sleeping is automatic
22:41:22Jaykaykugel: sorry i mean ogg, when i did my first benches you said they run boosted
22:41:41amiconnSo ata_spin() (and hence sd_spin() ) is only called in special situations
22:41:58gevaertsAh ok. That seems to match my recollections :)
22:42:05amiconnSleeping is automatic as well, after spindown delay. ata_sleep() is for quick-sleeping
22:42:40ZagorJaykay: sorry for late response. thanks, but a battery bench of #9800 won't be very interesting. the purpose of that patch is not to save power.
22:42:42kugelJaykay: flac and mp3 run unboosted, so ogg will yield more accurate results
22:42:43 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
22:43:38kugelsaratoga: here's the first one
22:44:40webguest35i know the wps gallery
22:44:48webguest35but I need the updated WPS
22:45:01kugelsaratoga: hm, I'm gonna test it a bit more
22:45:09JaykayZagor: if you dont mind i'll do it anyway, unless you have a more helpful bench for me
22:45:15webguest35there was a page with just themes with the new WPS themes that work with the new Rockbox version
22:45:26ZagorJaykay: sure, have fun :-)
22:46:27Zagorlucent: haha, guess if sd-at-boot works for me :-)
22:46:50kugelZagor: but it would surely be interesting if the bench would confirm Buschel
22:47:20lucentZagor: I'm afraid. Tell me
22:47:32Zagorlucent: it works
22:47:33kugeli.e. that it actually increases power consumption
22:47:57lucentZagor: it could just be my card then, what brand / size / class card are you using?
22:48:34Zagorlucent: sandisk 1GB
22:49:18 Quit Jaykay ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
22:49:41lucentZagor: mine is Transcend 8gb class
22:49:55*lucent wonders where the "class 6" got cut off
22:50:16Zagorkugel: true, that would be interesting. I plan to add wires to my c200 battery pins so I can measure the current used with different settings.
22:51:24ZagorI guess I should do it to the H140 and one of the clips while I'm at it
22:51:28webguest35does anyone know where this page
22:51:34webguest35is at :P
22:52:19lucentwebguest35: TWiki doc index?
22:52:19Lloreanwebguest35: The most up to date page *should* be the WPS gallery unless people are intentionally choosing not to post their WPSes there.
22:56:11kugelZagor: what needs to be built to report a accurate rockbox-info.txt? LANGs and main binary, or more?
22:56:44kugelit sometimes drives me crazy that I need to build everything just to check bin&ram sizes
22:56:55 Join kachna [0] (
22:56:55saratogakugel: i'll be in and out today, just post what you have when you're done and i'll try and commit in a few hours
22:57:17 Join {phoenix} [0] (
22:57:35Zagorkugel: the binary
22:57:36Bagderkugel: rockbox-info.txt is made by, you can run that manually
22:58:20Zagorsaratoga: did you write down your clock vs current findings on PP?
22:58:36kugelthe langs not (for the buffer)?
22:58:52kugelBagder: thanks for the hint, what does that build?
22:58:57Bagderthe binary won't be right without the langs, correct
22:59:08Bagderkugel: the .txt file only
22:59:31saratogaZagor: some of them outputfile
22:59:43saratogai lost one notebook, but still have some of the graphs I made
22:59:47kugelhm, well, I was looking for a make target which builds everything what's needed to get rockbox-info.txt properly
22:59:49saratogaand i can make more of course
23:00:08Bagderand I offered a work-around
23:00:13webguest35:O I think the TWiki doc page might have been it! can someone link me?
23:00:32Bagdermake bin must build the langs too, doesn't it do that?
23:00:53kugelI think so yes
23:01:04Lloreanwebguest35: The only page on the wiki with WPSes is the WpsGallery
23:01:07kugelso make bin && ../ ?
23:01:20Bagderkugel: yes, I think that should work just fine
23:01:21webguest35oh -_-
23:01:38webguest35i know there is one...its hard for me to explain cuz im not very smart :P
23:01:56Zagorsaratoga: how much documentation do you have? for example do you/we know what the bits in DEV_TIMING1 do?
23:02:00webguest35its not the main page, it only had themes with 3.0 WPS
23:02:01Lloreanwebguest35: Many of the WPSes on the gallery page should work just fine.
23:02:23LloreanThere is no official "just 3.0 themes" page.
23:02:34LloreanIf someone's hosting them on some random website elsewhere, we can't help you.
23:02:36webguest35how do I know which ones will work
23:02:59LloreanIf they don't say, try them.
23:03:17kugelBagder, Zagor: how about adding info: bin $(RBINFO) to root.make?
23:03:27 Join sputnik [0] (
23:03:32*Zagor again urges amiconn to post what whatever docs he has
23:03:34kugelI just tried and it works
23:04:48saratogaZagor: I only really looked at DEV_EN registers, I don't believe I tested that one
23:04:58Zagorkugel: no need. just run: make bin `pwd`/rockbox-info.txt
23:05:14saratogawould power measurements be useful?
23:05:21kugel"make info" is much easier :(
23:05:53Zagorkugel: I don't understand what you are after. why aren't running a full make?
23:05:59gevaertskugel: "man alias" :)
23:06:32kugelbecause my laptop is slow, and building zbox and doom alone needs almost 5 minutes
23:07:08Zagorkugel: and why do you need rockbox-info.txt?
23:08:13sputnikdo i have to reboot my 3g ipod to original firmware to load music?
23:08:14Zagorsaratoga: perhaps not at the present state
23:09:17kugelchecking bin and ramsize easily
23:10:26webguest35nvm I don't any of you guys are the guys that helped me last you probably donno of the page :( or if you do I can't explain it well enough >_<
23:10:31webguest35thanks for your help anyways
23:10:53 Quit {phoenix} (Remote closed the connection)
23:10:55Zagorkugel: well, I don't think we should add a target just to save you typing 20 characters :-)
23:11:18Bagder... that could be made an alias or script and become one character
23:12:19kugelwell, I'll learn to live with it ;)
23:12:40kugelnow that i know how it works without building everything it's not a big deal
23:12:42 Quit japc (Read error: 110 (Connection timed out))
23:13:13 Join japc [0] (
23:13:17 Quit Seed ("cu, Andre")
23:15:15Zagoramiconn: did you see my question about the "SLOW" clock source? do you know if we can use that?
23:15:23saratogaZagor: i suppose I should compute watts/mull for a few different clock speeds though to see if power consumption really does change linearly
23:15:52 Join obo [0] (
23:16:07 Quit japc (Read error: 60 (Operation timed out))
23:16:07Zagorsaratoga: that would be interesting. can you measure on other targets than PP?
23:17:14saratogaZagor: I have a fuze, clip, e200v1, gigabeat F/S and a mostly broken ipod 3G
23:17:24 Join MethoS [0] (
23:17:34Zagorsaratoga: all prepared for measuring?
23:17:46saratogano but its easy enough to prepare them
23:17:48 Quit domonoky (Read error: 104 (Connection reset by peer))
23:18:07saratogaat most I'd have to cut a USB cable and splice in a 0.5 ohm resistor
23:18:33Zagorsaratoga: excellent. I think an AMS target would be interesting as comparison.
23:18:49saratogaare they clocked well enough to be interesting yet?
23:19:00saratogait sounds to me like they're still not working properly
23:20:07kugelZagor: and how about a make target, that installs to the device directly without needing to create a tar/zip/whatever and going through the extracting procedure?
23:20:15*kugel has some funky ideas today :/
23:20:30 Nick fxb is now known as fxb__ (
23:20:55 Join Makuseru [0] (
23:20:57saratogathat'd be nice
23:20:58Zagorwell, we at least know that we are clocking the core separately and that boost is 8x the normal freq.
23:21:10saratogaand probably pretty easy to edit
23:21:38saratogayeah but if we're not clocking the peripherals right the power measurements won't mean much
23:23:24 Join Seed [0] (
23:23:47 Join aneqrs [0] (
23:23:52ZagorIt depends how much work it is for you. I think it would still be interesting to see.
23:24:06saratogaif i'm going to do PP its not a big deal to do AMS too
23:24:30 Quit aneqrs (Connection reset by peer)
23:25:03saratogajust curious what you're thinking
23:25:09 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
23:25:13ZagorI didn't even know these players could run without battery.
23:26:03saratogathe V1s can
23:26:20saratogathe fuze and clip cannot easily have the battery removed actually
23:26:27saratogaso measurements on them will be complicated
23:26:36saratogathats not something i'd considered
23:26:59 Quit itcheg (" ajax IRC Client")
23:27:37Zagorright. we'll just delay that a bit then :)
23:30:39kugelsaratoga: tested on fuze+clip; main+bootloader
23:31:07 Quit webguest35 ("CGI:IRC (EOF)")
23:33:05saratogakugel: i haven't read todays logs yet, how much have you changed?
23:34:14kugelmake the retry volatile, since it's potentially be changed in the interrupt handler, to make sure gcc doesn't optimize it away
23:34:36saratogaso this mostly fixes crashes with some cards i guess?
23:34:41kugelsome cleanup here and there, and add the delays which make lucent's microsd card working
23:34:46saratogaand of course the cleanup
23:34:51saratogaok sure i'll commit it tonight
23:34:53saratogahave to run
23:35:30kugeloh, I also changed some panics to be more helpful, i.e. printing the error code
23:37:28kugelsaratoga: ok cool. then I'll upload the patches for buttonlight and backlight fading to the tracker entry which already exists for those tomorrow
23:38:25 Join TheSphinX^ [0] (
23:41:50kugelZagor: did you already experiment with the clipv2?
23:42:37Zagorkugel: no, I got caught up in other interesting things :)
23:42:59 Join tim__ [0] (n=aoeu@
23:43:18*kugel hopes he can have a talk with jhMikeS soon regarding the MMU
23:44:03Zagordid you download the ARM922T to read up on the COP15?
23:44:27 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
23:44:42kugelZagor: yep
23:44:58kugelI didn't understand fully though
23:45:01Zagorsaratoga: can you paste your ams perfcheck() somewhere? (or even commit)
23:45:14 Quit TheSphinX^ ("XChat@Linux")
23:45:54kugelbut I understood enough to see that it's exactly the same as on F/X, amiconn also said it's nearly the same CPU (arm922T vs arm920T), so it makes sense
23:47:04kugelso I'm at the point of copying F/X start up to ams, and I think I got the MMU running (the bootloader boots). the main doesn't start though, so there's something missing
23:47:26kugelI guess I didn't map the memory regions correctly
23:48:53kugelor others things I didn't understand enough, so I hope someone can help me
23:49:17 Quit bertrik ("Leaving")
23:49:33kugelanyway, I was reading crt0.S yesterday, a file I always found totally scary, but I actually understood it pretty good
23:50:10amiconnOnce you enabled the caches, you need to make sure to flush & invalidate them in the correct order every time you load some code
23:50:32amiconnThis is because of those CPUs having separate instruction and data caches
23:51:34amiconnThis e.g. applies to the bootloader. After loading the main binary, you need to flush dcache and then invalidate icache before running it
23:51:54kugelhm, I think I didn't do that
23:52:10amiconnGigabeat F/X has to do that somewhere, unless it doesn't enable the caches in the bootloader
23:52:19kugelbut our bootloader is in IRAM, so there's nothing in the SDRAM before executing the main
23:52:36amiconnWell, you load the main binary into sdram
23:52:44amiconnWhat was there before doesn't matter
23:53:00kugelso, load binary into ram, flush&invalidate, then execute?
23:53:21amiconnIf you don't flush the data cache, some data will not end up in sdram, hence the cpu would fetch what was there before
23:54:09amiconnAnd if you don't invalidate the instruction cache after flushing the data cache, the cpu might not read sdram because the memory area was already cached before. This shouldn't apply at boot, but better play safe...
23:54:32kugelok, thanks, I'll have a look
23:54:42 Join jhulst_ [0] (n=jhulst@unaffiliated/jhulst)
23:54:50 Quit tyfoo ("Carpe diem")
23:55:06kugelcan you also tell me what this "map_section(0, 0, 0x1000, CACHE_NONE); /* map every memory region to itself */" (from F/X) means? also what's the 0x1000 referring to?
23:55:11amiconnCaching in systems with more than one cache (be it like on arm9 with separate icache and dcache, or like on PP with its 2 cores, each having its own cache) tends to cause major headaches
23:55:27kugelyea, I noticed

Previous day | Next day