Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

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

#rockbox log for 2010-06-22

00:00:09 Join fml [0] (
00:00:22Bagderthere's no such project within Rockbox
00:00:40 Quit r0b- (Ping timeout: 264 seconds)
00:01:03fmlAlexP: hello! The section about the UI Viewport still conains obsolete info about the font
00:01:17peturgrrrr... If I set max filesize small to debug, recordings get split, if I set it to anything near 2GB, it doesn't... WTF
00:01:23 Join r0b- [0] (
00:01:25 Quit piggz (Read error: Connection reset by peer)
00:01:29Meekrabwell, the zune project in general, you know what I mean, i think
00:01:39Bagderwhich zune project?
00:01:40S_a_i_n_tThere isn't one.
00:01:47Bagderand no I don't know what you mean
00:02:18Bagderthis is Rockbox, and there's no zune port going on that we know of
00:03:04MeekrabWell, I'm pleasured to say i will start this project.
00:03:07 Join stripwax [0] (
00:03:11fmlAlexP: there are also some places where nice apostrophes are used inside code
00:03:21 Quit fml (Client Quit)
00:03:27BagderMeekrab: welcome, we always enjoy more ports!
00:04:56 Quit flydutch (Quit: /* empty */)
00:06:46 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
00:07:34***Saving seen data "./dancer.seen"
00:08:11 Quit giovanni (Ping timeout: 240 seconds)
00:08:56JeromWill the AMSv2 port promoted to stable when USB will be working ?
00:10:17 Join dfkt [0] (~dfkt@unaffiliated/dfkt)
00:10:57bertrikI don't think so, but this will probably be re-evaluated a few weeks before the next rockbox release
00:13:05 Join Trou [0] (
00:14:16 Quit bieber (Read error: Operation timed out)
00:14:19Trouhello, I just wanted to say "thank you, a lot" for the work you're doing on the Sansa Clip+
00:14:28Trouit's really awesome.
00:15:49bertriknice to hear it's appreciated :)
00:15:59 Quit petur (Quit: Zzzzz)
00:16:08 Quit mt (Ping timeout: 258 seconds)
00:16:44Troui've been using it since the first builds, the progress is impressive :)
00:17:09 Quit domonoky1 (Read error: Connection reset by peer)
00:17:15 Join Transformer [0] (
00:19:00 Part Transformer
00:19:54bertrikTrou, if the radio on your Clip+ didn't work a week ago, it should work now
00:20:31Trouyeah i've seen the svn changes, I didn't try it though, I don't use it
00:20:39Troubut I can test if you want
00:20:57bertrikSome nice progress was made today on USB for AMSv2 (like the Clip+)
00:21:26Troufor usb storage support ?
00:21:30 Join leavittx [0] (~leavittx@
00:21:53bertrikTrou, yes, eventually
00:25:01 Join M3DLG [0] (
00:27:11Trouhmm I have no sound with FM
00:27:36Trouah now it works
00:27:45 Quit Milardo2 (Quit: Leaving)
00:29:10 Quit Jerom (Quit: Leaving.)
00:29:17 Quit ender` (Quit: Washing your car to make it rain doesn't work.)
00:29:56 Quit stripwax (Quit:
00:31:15 Join funman [0] (~fun@rockbox/developer/funman)
00:32:45 Quit bertrik (Quit: De groeten)
00:35:23 Part Trou ("Leaving")
00:37:49 Join zu [0] (
00:38:38 Part Meekrab
00:38:51 Quit pamaury (Remote host closed the connection)
00:46:02 Quit M3DLG (Ping timeout: 264 seconds)
00:50:36funmanshould lossless files be more sensitive to bitstream corruption ?
00:52:04funman(than lossy files)
00:54:50 Quit r0b- (Ping timeout: 276 seconds)
01:00:37 Quit funman (Quit: free(random());)
01:01:36 Join r0b- [0] (
01:18:21 Quit bluebrother (Ping timeout: 265 seconds)
01:19:49 Join bluebrother [0] (
01:19:49 Quit bluebrother (Changing host)
01:19:49 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
01:26:50 Join halmi [0] (
01:27:52 Join anewuser [0] (anewuser@unaffiliated/anewuser)
01:29:06 Quit amiconn (Disconnected by services)
01:29:07 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
01:29:25 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
01:29:49 Quit JdGordon| (*.net *.split)
01:29:49 Quit Guest53364 (*.net *.split)
01:29:50 Quit Gabe_G23 (*.net *.split)
01:29:50 Quit jordan` (*.net *.split)
01:29:50 Quit yosafbridge (*.net *.split)
01:29:50 Quit AlexP (*.net *.split)
01:29:50 Quit Stummi (*.net *.split)
01:29:50 Quit dionoea (*.net *.split)
01:30:41 Join jordan` [0] (
01:31:49 Join sharp [0] (
01:32:12 Join Gabe_G23 [0] (~gabe@bzflag/player/GabrielG)
01:32:17 Nick sharp is now known as Guest97482 (
01:32:42 Join JdGordon| [0] (
01:32:42 Quit JdGordon| (Changing host)
01:32:42 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
01:32:46 Join dionoea [0] (
01:33:28 Join AlexP [0] (~alex@rockbox/staff/AlexP)
01:33:34 Join Stummi [0] (
01:33:38 Quit jgarvey (Quit: Leaving)
01:33:42 Quit efyx (Remote host closed the connection)
01:38:11 Quit halmi (Read error: Connection reset by peer)
01:39:32 Quit DataGhost (Ping timeout: 260 seconds)
01:49:39 Part toffe82_
01:50:47 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
01:54:27 Quit dfkt (Ping timeout: 245 seconds)
02:07:12saratogafunman: well the files are much larger, so errors are more likely
02:07:37***Saving seen data "./dancer.seen"
02:07:46saratogamost lossy codecs are pretty robust against errors, since each block is so short and they are made to resync if packets are lost
02:07:57 Quit detaos (Remote host closed the connection)
02:08:08saratoganot sure about lossless files, probably depends on the individual format and how much thought they'd given to error robustness
02:11:03 Join detaos [0] (
02:12:57 Quit dfkt_ (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
02:13:43 Quit GeekShadow (Read error: Connection reset by peer)
02:21:20 Quit rvvs89 (Ping timeout: 252 seconds)
02:21:30 Join rvvs89 [0] (
02:23:20 Join yosafbridge [0] (
02:40:41 Quit S_a_i_n_t ()
02:40:52 Join S_a_i_n_t [0] (S_a_i_n_t@
02:43:53 Join CGL [0] (~CGL@
02:54:43CIA-8New commit by 03mc2739 (r27036): FS #11427 by Marek Salaba: Update of Czech language
02:56:36CIA-8r27036 build result: All green
03:17:48 Join Transformer [0] (
03:19:05 Part Transformer
03:30:00 Join Zigtown [0] (
03:31:02 Quit MethoS- (Remote host closed the connection)
03:39:13 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
03:47:18 Quit steve|m (Ping timeout: 276 seconds)
03:56:01 Join hatoon [0] (~musis@
03:58:53hatoonwhat rockbox does, he is a program that read mp3 and mp4, or editing music
04:02:25 Join steve|m [0] (
04:06:02 Quit r0b- (Ping timeout: 245 seconds)
04:06:13 Join r0b- [0] (
04:07:39***Saving seen data "./dancer.seen"
04:14:51 Quit Barahir_ (Ping timeout: 240 seconds)
04:16:50 Join Barahir [0] (
04:17:44 Quit hatoon (Quit: Saindo)
04:24:32 Nick keanu is now known as [keanu] (~keanu@unaffiliated/keanu)
04:27:19 Quit Zigtown (Ping timeout: 276 seconds)
04:29:26 Join bieber [0] (
04:32:01 Quit pixelma (Disconnected by services)
04:32:03 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:32:19 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:33:17 Quit amiconn (Disconnected by services)
04:33:19 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:33:37 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:38:26 Join komputes [0] (~komputes@ubuntu/member/komputes)
04:39:47 Quit BHSPitMonkey (Ping timeout: 248 seconds)
04:52:28 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
04:55:17 Quit TheSeven (Ping timeout: 260 seconds)
05:00:42 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
05:11:48CIA-8New commit by 03jethead71 (r27037): e200v1: What's the wheel deal? Rework wheel code to make it a bit more straightforward and increase control a bit. Change button interrupt to clear ...
05:13:34CIA-8r27037 build result: All green
05:14:19S_a_i_n_tComical Commit messages FTW!
05:17:11 Quit komputes (Ping timeout: 265 seconds)
05:25:18 Part AzureSky
05:25:20 Quit jfc (Read error: Connection reset by peer)
05:29:44 Join komputes [0] (~komputes@ubuntu/member/komputes)
05:35:09 Join Transformer [0] (
05:35:31 Quit Horscht (Quit: Verlassend)
05:36:55 Quit Transformer (Excess Flood)
05:38:41 Join saratoga_ [0] (9803c20d@gateway/web/freenode/ip.
05:40:00 Quit saratoga (Changing host)
05:40:01 Join saratoga [0] (9803c6dd@rockbox/developer/saratoga)
05:50:20 Quit t0rc (Quit: Leaving)
05:50:43 Quit binaryhermit (Ping timeout: 258 seconds)
05:53:00 Join AzureSky [0] (
05:55:21 Part AzureSky
05:56:21 Join binaryhermit [0] (
05:56:45 Join AzureSky [0] (
06:00:37CIA-8New commit by 03jethead71 (r27038): lv24020lp tuner: On PP targets (c200/e200), use the atomic GPIO bitwise macros for the interface since it shares GPIOH with the clickwheel interrupt.
06:02:28CIA-8r27038 build result: All green
06:07:40***Saving seen data "./dancer.seen"
06:18:08 Quit anewuser (Quit: for SELL 2 by the price of 1 now!)
06:18:36 Quit antil33t ()
06:20:49 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
06:28:35 Quit binaryhermit (Ping timeout: 265 seconds)
06:30:42 Join binaryhermit [0] (
06:56:02 Join Zigtown [0] (
07:00:03bieberJdGordon, S_a_i_n_t: ping
07:00:37AzureSkywhat was the command to compile the cross compilers again?
07:15:34 Quit kramer3d (Quit: Leaving)
07:20:30 Quit storm` ()
07:21:55 Join storm` [0] (
07:25:59 Quit panni_ (Read error: Connection reset by peer)
07:28:22 Join funman [0] (~fun@rockbox/developer/funman)
07:41:55 Join kaylinsigswort [0] (
07:41:56 Quit Zigtown (Read error: Connection reset by peer)
07:42:40 Quit binaryhermit (Ping timeout: 265 seconds)
07:46:56CIA-8New commit by 03ranma (r27039): usb_storage seems to be working now, enable USE_ROCKBOX_USB on C200v2, other AMSv1 untested.
07:47:11funmanranma: cool!
07:47:19funmani think we should enable it on all targets
07:47:26funmani'll test fuzev1/clipv1 shortly
07:48:44CIA-8r27039 build result: All green
07:50:22funmanranma: UNCACHED_ADDR is not defined for AS3525, do we need to check for it in usb_storage.c ? (I'd prefer if we remove this from common usb code btw)
07:51:06funmanbtw *_dcache_range() don't work?
07:53:24ranmafunman: Well it seems to be defined, that was the reason usb_storage was passing me uncached pointers to the buffer
07:53:35ranmaI'll try dcache_ragen now
07:54:02funmani changed it to AS3525_UNCACHED_ADDR to workaround this
07:54:43ranmaAh, ok, didn't know that
08:02:12funmanranma: running lsusb -v still causes problem
08:03:27ranmaBut mounting works?
08:03:37funmanyeah i'm trying to write some stuff
08:03:54*ranma is still recompiling
08:04:05funmandevice descriptor read/64, error -110 + reset
08:06:11funmancopying rockbox.sansa is enough to lock fuzev1
08:07:44***Saving seen data "./dancer.seen"
08:10:41 Quit kaylinsigswort (Quit: Leaving...)
08:12:51 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
08:15:55ranmafunman: I see that usb-drv-as3525.c gets the setup packet (and passes it to the usb core), but the usb core doesn't seem to want to respond?
08:16:10funmani know nothing of usb
08:17:41 Join antil33t [0] (
08:18:18 Join bieber_ [0] (
08:18:41 Quit leavittx (Ping timeout: 258 seconds)
08:19:00 Quit bieber (Ping timeout: 248 seconds)
08:20:13 Join ender` [0] (
08:22:44ranmaHmm, looks like the usb thread is stuck somehow.
08:23:24ranmaNormally it should be "ctrl received 1883" and next "ctrl handled 1883" (from the thread)
08:23:37ranmaBut in the lsusb case I just see the former
08:27:44 Join Zagor [0] (bjst@rockbox/developer/Zagor)
08:38:56S_a_i_n_tbieber_: You rang?
08:39:31funmani ran test_disk on fuzev2 while the ape decoder outputted noise : "Test passed."
08:39:34bieber_Hey, how are images and text placed in viewports?
08:39:59S_a_i_n_tI think you'll need to define the question a little better.
08:40:07S_a_i_n_tyou mean, which tags?
08:40:08bieber_Does text wrap? And how are images aligned? Do you normally mix the two together in viewports?
08:40:20bieber_I mean in terms of display
08:40:49S_a_i_n_ttext doesn't wrap, it scrolls, but only if the %s tag is present, otherwise it just exceded the viewport limit.
08:41:08S_a_i_n_timages are placed by the top leftmost pixel.
08:41:33bieber_Okay, and if you have, say, text after an image, does it just start immediately to the right of the image?
08:41:40S_a_i_n_tie. to display an image in the top left of a viewport, you'd give it x,y coordinates of 0,0
08:41:59S_a_i_n_tNope, you need to do all the alignment manually.
08:42:17S_a_i_n_totherwise text will try to draw on top of, or behind said image.
08:42:31bieber_Do you have to use another viewport, or are there text alignment tags?
08:42:41ranmafunman: Hmm, seems to be some problem with sending packets > 64 bytes on the control endpoint
08:42:43JdGordonbieber_: pong
08:42:47S_a_i_n_tthere are text alignment tags.
08:43:01bieber_Thanks, that should be plenty for now :)
08:43:02JdGordondid you get an answer to last nights q?
08:43:02S_a_i_n_tthe lines for text are divided by the font height.
08:43:19bieber_JdGordon: Yep
08:43:22S_a_i_n_tie. a 100px high viewport displaying 10px font will have 10 "lines"
08:43:44bieber_So text and images are just displayed completely independently?
08:43:49saratogafunman: the flac decoder detects the error, doing a logf build might be interesting
08:44:19bieber_brb, gonna see if I can get logged in as my normal nick
08:44:19S_a_i_n_tyou know that a viewport gets its own set of coordinates right?
08:44:25 Quit bieber_ (Remote host closed the connection)
08:44:42 Join bieber [0] (
08:44:44S_a_i_n_tI may need to clarify this...I mean, it doesn't share the x,y coordinates of the "default" viewport.
08:45:03bieberRight, they're offset by the Viewport's position
08:45:21S_a_i_n_teach viewport is like a mini "screen", so the top left of a viewport will always be 0,0
08:46:08S_a_i_n_t(some people forget that, and try to use coordinates from the default viewport and wonder why thier images/text isn't aligned/doesn't display at all.
08:46:16JdGordonfunman: what made you dig into gui_statusbar_draw?()
08:46:18S_a_i_n_tA rather common fail in my experience
08:46:21funmanranma: the driver should split packets?
08:46:33funmanJdGordon: crash in text_viewer on sansa ams + -O1 + eabi
08:46:40bieberThat's perfect, since Qt does the same thing with children of graphics items (each item has its own local coordinate system for its children)
08:47:00S_a_i_n_taha, great.
08:47:33S_a_i_n_tI'm quite impressed with what you've been able to do so far with a (not ragging on you) somewhat limited knowledge of the syntax.
08:47:38S_a_i_n_tGood work man, really.
08:47:50bieberHehe, thanks
08:48:18bieberOne of these days I shoul probably try just playing around with WPS files on my iPod for a while instead of asking you guys about everything :P
08:48:39S_a_i_n_tI don't mind, really.
08:48:46JdGordonS_a_i_n_t: it is also possible that not having any knowledge about it makes it easier.. you dont get stuck in how you expect it to work
08:48:58S_a_i_n_tyes, agreed.
08:49:33S_a_i_n_tand, you're also in a position to make things work the way you expect them to if they don't already. Developer perks ;)
08:49:35Mode"#rockbox +o pixelma" by ChanServ (ChanServ@services.)
08:49:51 Quit funman (Quit: bbl)
08:49:55Topic"Rockbox 3.6 now released! | Please read before speaking: | Please direct offtopic/social chat to #rockbox-community | This channel is logged at" by pixelma (quassel@rockbox/staff/pixelma)
08:52:30 Join funman [0] (~fun@rockbox/developer/funman)
08:52:45Mode"#rockbox -o pixelma" by ChanServ (ChanServ@services.)
08:53:28 Join flydutch [0] (
08:54:35 Quit JdGordon (Quit: Leaving.)
08:55:03 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
08:55:31 Quit saratoga_ (Quit: Page closed)
08:56:44 Join hamish_ [0] (~hamish@
08:56:52 Quit hamish_ (Client Quit)
08:58:18CIA-8New commit by 03funman (r27040): test_codec: fix DEBUGF pointer format specifier
08:59:54CIA-8r27040 build result: All green
09:00:34JdGordonfunman: and you traced the crash to statusbar_draw?
09:01:03funmanyes i must look again to see exactly what happened
09:02:17 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
09:02:55funmansaratoga: simulator and target should give the same checksum in test_codec ?
09:02:58ranmafunman: No, the hardware should handle that
09:04:10 Join petur [0] (d408b802@rockbox/developer/petur)
09:04:41 Quit dionoea (Changing host)
09:04:41 Join dionoea [0] (~dionoea@videolan/developer/dionoea)
09:05:55CIA-8New commit by 03funman (r27041): CREDITS: Capitalize FFmpeg properly
09:07:28CIA-8r27041 build result: All green
09:20:58funmansaratoga: an mp3 which decodes ok (to my ears) produce a .wav different from the sim
09:21:09 Join bimbel [0] (~Miranda@unaffiliated/bmbl)
09:25:01 Quit bmbl (Ping timeout: 258 seconds)
09:27:04TheSevenfunmen: RE FS #11429: what's that whole multivolume/multidrive thing all about? what are volumes? what are drives? what's the difference?
09:27:18funmangevaerts knows
09:27:24TheSevendo we support mounting multiple partitions on a single drive? or what's going on here?
09:27:41CIA-8New commit by 03jethead71 (r27042): Sanyo lv24020lp FM: Improve frequency measurement on PP thus improving initial frequency setting. Properly account for IF when tuning FM oscillator ...
09:28:13funmanMULTIDRIVE is for sd/mmc slots
09:28:50funmanreading config.h it seems MULTIVOLUME is for multiple partitions (up to 4 -> only primary partitions?)
09:29:06CIA-8r27042 build result: All green
09:29:07funmanso i guess my comment only applies to MULTIDRIVE
09:29:38TheSevenfunman: it seems like i have mixed up MV/MD at several points, so it probably won't build for either one
09:29:45 Join swilde [0] (
09:31:38bieberJdGordon: Skins load images from /.rockbox/wps/themename/, but how does the renderer determine the theme name? Just use the file name of the .cfg?
09:32:04JdGordonnot quite
09:32:06TheSevendoes MD depend on MV? probably yes...
09:32:20 Join einhirn [0] (
09:32:30JdGordonskin.wps will load images from wps/skin/ foo.fms will load images from wps/foo/
09:32:42funmanTheSeven: i think it's the reverse but check config.h
09:32:45bieberOh, okay
09:32:47JdGordoneach skin loads from its folder, usually all named the same
09:32:54funmanhm no sorry you're right
09:34:42 Join mt [0] (~mtee@rockbox/developer/mt)
09:35:43 Join ssorgatem [0] (
09:37:42 Quit komputes (Ping timeout: 265 seconds)
09:39:37funmanreverting the 2 sd AMS commits still give the same CRC32 - different from what the sim gives
09:40:00funmanis it expected to not have the same exact result for mp3 ?
09:40:14funman(between e.g. ARM and x86 sim)
09:40:57 Join Kitr88 [0] (
09:41:42 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:45:48 Quit Kitr88 (Ping timeout: 265 seconds)
09:45:50 Quit Kitar|st (Ping timeout: 276 seconds)
09:46:34ranmaHmm, something weird is going on, looks like the setup packet responses are skewed, the host sees them one packet too late?
09:48:21JdGordonbieber: are you going to fix up the project tree to show it more like a filetree with "directories" and stuff? and show all the required images and warn when one is missing?
09:49:29JdGordoncoolies :)
09:49:49 Quit FlynDice (Remote host closed the connection)
09:51:13 Join Kitar|st [0] (
09:52:44funmanranma: hm i can't get very far with the memdump because the registers aren't saved in the data abort handler
09:53:58funmani only know r8 because it's the one causing the abort
09:55:16 Join Rob2223 [0] (
09:55:52CIA-8New commit by 03bieber (r27043): Theme Editor: Began in implementing tag rendering, %X tag now recognized
09:57:32CIA-8r27043 build result: All green
09:58:35 Quit Rob2222 (Ping timeout: 264 seconds)
09:59:08 Quit bimbel (Quit: Bye!)
09:59:22TheSevenhm, why do changes to /trunk/utils trigger a build?
09:59:38TheSeventhey aren't built by the build system anyway, are they?
10:04:27pamauryTheSeven: when looking at your patch, I think it would be better to imtroduce a fat_closedir function instead of manually unlocking the sector cache in various places
10:05:43pamauryAnd a fat_closefile
10:06:02pamaury(anyway, a fat_closedir would translate to a fat_closefile)
10:06:20TheSevenwould it? i don't think so
10:06:28TheSevenfat files don't even have a sector buffer
10:06:50TheSevenfor files the buffer is in the highlevel struct, while for dirs it's in the lowlevel struct
10:07:46***Saving seen data "./dancer.seen"
10:08:06pamauryThat's for completeness. I don't the idea of having an open but not close. We could introduce it now.
10:09:05 Join Jaykay [0] (
10:10:44TheSeveni might do that (and also try to fix the multivolume problems) in the evening
10:13:45 Nick fxb__ is now known as fxb (
10:14:03pamauryTheSeven: this part is *really* strange:
10:14:05pamaury- update_fat_entry(IF_MV2(fat_bpb,) newdir->file.firstcluster, FAT_EOF_MARK);
10:14:06pamaury+ update_fat_entry(IF_MV2(dir->file.volume,) newdir->file.firstcluster, FAT_EOF_MARK);
10:15:05TheSevenwtf did I do there?
10:15:19JdGordonbieber: we shouldnt ever get more than 3 children levels in a tree right?
10:15:39pamauryI don't know but it doesn't sound good :)
10:15:41TheSevenbut i'm pretty sure there's quite a bunch of other screwup
10:15:59pamauryFor now that's the only one I spot but I didn't check all the details
10:16:04TheSeventhat hunk is definitely bullshit though
10:16:15bieberJdGordon: There could be an infinite number with nested conditionals
10:16:32TheSeventhere are probably IF_MD macros that actually contain volume ids
10:16:52JdGordonbieber: oh right... I hope we arnt short on stack space then :)
10:17:17bieberHaha, hopefully
10:17:24bieberI don't imagine they'll generally be nested very deep
10:17:56pixelmadid I understand correctly that %t inside a conditional works now?
10:18:05TheSevenbtw, can anybody explain to me why screendump needs a whopping 2K on the USB thread stack?
10:18:32pamauryTheSeven: I'm prettry sure all your IF_MD2(dir->file.volume,) are zrong
10:18:34JdGordonI've given up on fully reworking the current system and for now I'm doing a translation thing to go from the tree to the old system
10:18:35TheSeven(with all the sector buffers removed)
10:18:49TheSevenpamaury: yes, that's what I meant
10:19:01JdGordonpixelma: it parses... if it actually will work is another story
10:19:10TheSeveni actually need the drive the volume is on at those locations
10:19:16TheSevenhow can i get that?
10:19:20funmanJdGordon: if i make gui_statusbar_draw() return immediately the crashes are gone
10:20:12pamauryYou can get the correct bpb using the volume index
10:20:23gevaertsTheSeven: as you probably know by now, MULTIVOLUME is relevant to volumes are (basically) partitions containing filesystems (not-mounted partitions don't count), and therefore relevant to filesystem code, while MULTIDRIVE means multiple *drives* are allowed, e.g. in systems with a card slot (or mad people experimenting with ramdisks). Multidrive forces multivolume, but not all multivolume systems are multidrive.
10:20:28JdGordonfunman: easily reproducable?
10:20:50gevaertsThere's also MULTIDRIVER, but that shouldn't ever escape outside of storage.[ch]
10:20:58pamauryTheSeven: you need a course on drive vs volume in rockbox :)
10:21:07funmanJdGordon: on sansa yes: use text_viewer
10:21:15JdGordonfunman: are you in the sim? breakpoing as soon as it enters that and make sure bar isnt NULL?
10:21:31funmansim doesn't crash
10:21:40JdGordonI dont tihnk any of my sansas are up to date :)
10:22:00TheSevengevaerts: yes, but I need drives in the file.c code when I need to lock sector buffers
10:22:22gevaertsTheSeven: vol_drive[] in disk.c can help you then
10:22:23JdGordonfunman: is it an immediate crash?
10:22:47funmanJdGordon: FS #11399
10:23:04pamauryTheSeven: either use what gevaerts suggested or keep in mind that fat_bpb->drive is a valid. And you can get the fat_bpb corresponding to each volume iirc
10:24:06pamauryTheSeven: example: IF_MD2(fat_bpbs[file->volume].drive) (untested)
10:24:12gevaertsah yes, it's probably stored inside fat data somewhere as well
10:25:13pamauryYou should use it this way, that's safer than using vol_drive[]
10:25:30 Quit JdGordon (Read error: Connection reset by peer)
10:25:31 Quit JdGordon| (Read error: Connection reset by peer)
10:25:41 Join hebz0rl [0] (
10:25:58 Join JdGordon| [0] (
10:25:58 Quit JdGordon| (Changing host)
10:25:58 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
10:26:02 Join JdGordon2 [0] (
10:26:48 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
10:28:26funmanJdGordon|: crashes in drawing (after '/* only redraw if forced to, or info has changed */')
10:28:57JdGordonwhich is where?
10:29:02 Quit JdGordon1 (Ping timeout: 258 seconds)
10:29:23 Quit mc2739 (Ping timeout: 276 seconds)
10:29:49funmanbut i have an idea
10:30:06pamauryTheSeven: okay, apart from this drive/volume confusion, I can't find anything suspicious. I'll a closer look tonight.
10:30:11funmanwhere is set_viewport defined ?N
10:30:35 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
10:30:52JdGordonthe lcd driver
10:31:02JdGordonI see alot of bad code not using the multiscreen api!
10:31:17JdGordonthat shouldnt cause crashes though
10:31:32funmanhm my fuzev2 is crashed and not powering off :/
10:33:29JdGordonoh hang on.... I'm pretty sure the plugin shouldnt be drawing the bar at all.. unless it uses the list widget?
10:33:34JdGordonwhich includes menus
10:33:49funmanthere's a setting for showing the SB
10:34:48JdGordonwhere is that code?
10:36:07JdGordontv_window.c comment out the code in tv_show_header() and see if that fixes it
10:36:16 Quit TheSeven (Ping timeout: 240 seconds)
10:36:59funman fixes it
10:37:52funmanthe struct viewport must live the whole time the viewport will be used but here it can be removed/overwritten from the stack after the if (vp == NULL) { }
10:38:25funmanand even this way i'm not sure gcc couldn't optimize it away as well
10:39:26JdGordonah that makes sense
10:39:32funmansince for gcc it seems viewport won't be used after set_viewport()
10:40:14funmanpamaury: can you comment on this, since you are the compiler expert ;)
10:41:30funmanin this case we could just make it static
10:42:02JdGordonbetter to pass the viewport into gui_syncstatusbar_draw()
10:42:12JdGordonbecause it is almost never NULL
10:42:16 Quit mc2739 (Ping timeout: 240 seconds)
10:42:18JdGordonso why waste the ram?
10:42:38funmangrep tends to disagree
10:43:49JdGordondinner.. back in 20
10:44:16 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
10:45:08 Join leavittx [0] (~leavittx@
10:45:52 Quit ssorgatem (Ping timeout: 265 seconds)
10:45:59 Join ssorgatem [0] (~ssorgatem@
10:51:17funmanJdGordon|: : your hunk doesn't compile, i let you fix it and commit?
10:59:42 Quit AlexP (Ping timeout: 240 seconds)
11:01:45 Join halmi [0] (
11:03:53*funman kicks CIA-8
11:08:16 Join pyro_maniac [0] (
11:09:34funmanok, what's the next bug on sansa AMS?
11:12:22funmani have no idea on how to debug poweroff failing when enabling RTC wakeup, perhaps it's related to what pamaury & bertrik reported: sometimes the clip+ seems to shutdown but a few time after the battery is dry
11:13:09 Quit CIA-8 (Ping timeout: 245 seconds)
11:13:16Stummifunman, you could add usb-support for the fuze v2 :)
11:13:35funmanpamaury is already doing it
11:13:52Stummioh, ok
11:20:15 Join efyx [0] (
11:24:01Stummiwhat is the minimal version of qt4 which I need for building the themeeditor? It failes with "build/ui/ui_editorwindow.h:132: error: ‘class QTabWidget’ has no member named ‘setTabsClosable’" and "...‘setMoveable’" on my debian, I am thinking it could be because of an older qt-version (4.4.3)
11:24:36 Quit sbhsu (Ping timeout: 248 seconds)
11:26:07JdGordonfunman: yeah, that looks OK... not too wild about the last hunk though
11:26:59JdGordondisplay->set_viewport inside statusbar_draw() can actually be #ifdefed out for CHARCELL
11:27:50funmanjust remove it then.. i let charcell maintainers fix the charcell bugs
11:27:55pamauryfunman: you summon me for compiler things ?
11:28:11funmanpamaury: yep we went another way but i'd like to know anwyay:
11:28:29JdGordonfunman: actually I think it already is
11:28:51JdGordons/charcell maintainers/amiconn/
11:28:55funmancan the compiler reuse a stack variable after it's not used in a local scope ?
11:28:58 Join bmbl [0] (
11:29:10 Quit bmbl (Changing host)
11:29:10 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
11:29:23pamauryYou mean, reuse for another purpose ?
11:29:36funmaneg: { type variable; do_smth(&variable); do_stuf(); } <- is variable memory guaranteed to still exist while do_stuff() is executing?
11:29:45amiconnfunman: screendump allocates a buffer on stack where the data is being prepared. The actual size of this buffer depends on screen size and pixel format
11:30:08amiconnFor colour targets and horizontally packed greyscale it's a single line
11:30:15funmanamiconn: what are you talking about?
11:30:45pamauryfunman: yes until the local scope is alive. After that, no.
11:30:56amiconnErr, that should have been addressed at TheSeven
11:31:01funmanok ^^
11:31:25 Join DerPapst [0] (
11:31:28funmanpamaury: so it will never be optimized away even with -foptimize-madness -O999 ?
11:31:36 Join M3DLG [0] (
11:32:08funmanJdGordon: in that case my first patch is simpler, no?
11:33:35pamauryfunman: if the compiler does some interprocedural optimization, some more complicated stuff can happen
11:34:22JdGordonfunman: yeah, your one looks fine, except GET_RECT() is missing the statusbar_position(i) call in it
11:34:43pamauryBut, the C++ standard says that as soon as the variable address is reauired, the variable should be stored somewhere in memory. I would not bet on gcc to enforce this however
11:34:46funmanJdGordon: i just tried to hack it so it builds but there's some stuff missing (display)
11:34:46JdGordonwithout that it will put the bar in the wrong place
11:35:12funmani hope uchida doesn't get mad at me
11:35:36pamauryfunman: And of course, inline could change some things
11:35:55JdGordonfunman: use &screens[i] instead of display
11:36:01JdGordonor I can fix it all if you prefer?
11:37:45JdGordonin fact the viewport.c hunk isnt needed at all
11:37:49funmanJdGordon: yes please
11:37:56JdGordonset_viewport(NULL) will work fine on charcell for this
11:38:00JdGordon99% sure
11:38:18funmani just replaced a NULL by false (unless I missed another declaration for charcells?
11:39:14 Join CIA-6 [0] (~CIA@
11:44:56CIA-6New commit by 03jdgordon (r27045): fiddle with the code a bit so gcc doesnt optimise out a struct viewport causeing crashes in text_viewer
11:46:31CIA-6r27045 build result: 6 errors, 0 warnings (jdgordon committed)
11:48:59pixelmadon't trust JdGordon's "99% sure" ;)
11:49:00CIA-6New commit by 03jdgordon (r27046): fix charcell
11:50:37CIA-6r27046 build result: All green
11:50:45 Join Luca_S [0] (
11:51:19Luca_S[<funman] ok, what's the next bug on sansa AMS? <
11:51:41Luca_Sopss sorry, wrong ctrl+v :P
11:53:22Luca_Swhat's missing on FS #11364?
11:56:10CIA-6New commit by 03jdgordon (r27047): remove a bunch of token markers which arnt necessary with the new parser
11:56:35funmanLuca_S: saratoga just needs to commit it i think
11:57:13funmanhm just pressing power button crashed my clipv2 :/
11:57:46CIA-6r27047 build result: All green
12:00:18Luca_Smmm there's no patch on flyspray for the PLL @ 384MHz - I was willing to build the ultimate fuzev2 build (11364 + cpu frequency scaling + pll @384) :D
12:02:18 Join sbhsu [0] (
12:05:36funmanif ultimate means doesn't boot then yeah it'd be nice i guess
12:06:14 Join giovanni [0] (~giovanni@
12:07:08Luca_Slol :D it wouldn't boot? I tested them separately in the past with no problems, sure I didn't try to combine them...
12:07:47***Saving seen data "./dancer.seen"
12:08:20funmanplla at 384MHz boots only with cpufreq set to 120MHz
12:08:32funmanhm 192 worked too iirc
12:09:22Luca_Sand switching audio to PLL B @ 384?
12:11:06funmanworks afaik
12:16:13JdGordonbieber: talk me out of this... I'm seriously considering using your parser to build the tree, then walk the tree and use the existing parse functions by rebuilding each tag! (passing one at a time)
12:16:13*JdGordon wonders if that would work
12:16:14Luca_Sfunman: to switch audio on pllb i have to change AS3525_MCLK_SEL?
12:22:41funmanranma: usb writing seems to work fine on clipv1 bootloader
12:23:40funmanperhaps there's something specific to fuze ? will do more tests
12:25:57 Quit M3DLG (Read error: Connection reset by peer)
12:26:24 Join M3DLG [0] (
12:28:29funmanranma: pamaury: i just found another place for USB configuration: config.h -> USB_HAS_BULK / USB_HAS_INTERRUPT
12:28:35giovanniI would like to start to build the manual for ONDA VX777/VX747 but I need some tips
12:29:03pamauryfunman: I don't understand what you mean
12:29:47funmansome usb stuff is configured in config.h, does it matter for AMS?
12:31:02pamauryyes, the USB_HAS_BULK / USB_HAS_INTERRUPT is important. I don't know of ranma setup them correctly for AMSv1. For v2 I haven't touched them yet
12:31:41funmanthey're not set afaict
12:32:14pamaurythat would be strange. Without them, the storage and hid are not compiled in normally
12:33:06 Quit GeekShadow (Ping timeout: 276 seconds)
12:33:07CIA-6New commit by 03ranma (r27048): "lsusb -vv" now works too, but the real bug is probably still in here somewhere.
12:33:20funmanUSB_HAS_BULK is never used
12:33:30pamauryyes that what I just saw
12:33:45*JdGordon pings pixelma for giovanni
12:33:58ranmafunman: We have both bulk and interrupt capability
12:34:05*funman needs to find 6768 bytes
12:34:30ranmaAnd of course iso and control too ;)
12:34:39CIA-6r27048 build result: All green
12:35:41pamauryI would not say for control. Few controllers support control for EPi != EP0 :)
12:36:15ranmaWell, I think you only ever need control for EP0 :)
12:38:56pamauryyes :)
12:43:27giovanniWhere can I start to build a manual for a new target?
12:43:30JdGordonhehe this might be easier tan I thought!
12:43:44JdGordongiovanni: do you know LaTeX?
12:43:47CIA-6New commit by 03funman (r27049): test_codec: remove 2 static string buffers and put them on the stack
12:43:51giovanninot really
12:43:54CIA-6New commit by 03funman (r27050): test_codec: allocate wavbuffer/dspbuffer dynamically ...
12:44:21giovanniI'm reading the Latex guidelines
12:44:50JdGordonassuming you already have the svn checkout you need to (I presume) create a new platform file and fiddle with configure maybe... pixelma shuold be able to point you in the right direction
12:45:21CIA-6r27049 build result: All green
12:45:49Luca_Shey, it works (it seems...)
12:46:19giovanniThanks JD, I'll try to contact pixelma
12:46:52CIA-6r27050 build result: All green
12:47:36Luca_Sfuzev2 + volume boost (FS11364) + cpu frequency scaling (11297) + MCLK_SEL = PLLB
12:48:10ranmaHmm, OF is faster when reading internal flash over usb (4MB/s vs 3MB/s), but on write they are about the same.
12:48:32ranmaBoth about 3MB/s on write
12:49:04funmanif fs#10805 works we should be faster
12:55:33 Join robin0800 [0] (
12:59:55pixelmagiovanni: I don't have much time currently (am at work). If you haven't done anything at all yet, you probably need to set up a build environment (how you would have to do that depends on the OS your computer is running and can be found in some "developers" part of the Rockbox wiki) and for building a manual I think the ManualHowto in our wiki is good too. Once you got that, I'd first try building an existing manual and then ask again :)
13:00:16pixelmanice to see someone interested in working on the manual :)
13:00:31 Quit mt (Ping timeout: 240 seconds)
13:01:42JdGordonS_a_i_n_t: ping
13:02:16*JdGordon has a rather odd patch that needs testing
13:03:01 Join mt [0] (~mtee@rockbox/developer/mt)
13:03:12funmanhm i could crash usb again :/
13:03:35giovannipixelma: thank you
13:06:52 Quit togetic (Ping timeout: 240 seconds)
13:14:00ranmafunman: Does it just hang, or do you get a panic?
13:14:00funmanhang, but USB in bootloader seems to work perfectly
13:18:21funmanbtw i still don't see the µSD
13:19:52JdGordonS_a_i_n_t: if you happen to be bored.. can you run FS #11431 on your ocmplex themes and let me know how it goes?
13:20:05JdGordonanyone else bored and want to help test the new parser?
13:21:03 Join mt_ [0] (~mtee@
13:21:32 Quit mt (Ping timeout: 240 seconds)
13:21:47 Nick mt_ is now known as mt (~mtee@
13:23:46ranmaIndeed, I don't see the µSD either
13:28:03funman sg_luns /dev/sg6
13:28:03funmanLun list length = 16 which imples 2 lun entries
13:28:16 Quit bieber (Ping timeout: 265 seconds)
13:28:28funmangevaerts: any idea what we might miss?
13:28:33 Join bieber [0] (
13:29:30gevaertsfunman: what's happening?
13:29:44funmanµSD slot isn't present on USB mass storage
13:30:57gevaertsis it seen without usb storage?
13:31:15funmanin rockbox? yes
13:31:23gevaertsAlso, is the slot not present, or the card?
13:33:17funmanonly one scsi device is detected
13:33:41*gevaerts needs more details
13:34:15gevaertsI can't see how this might be caused in usb_storage.c without the rest of the system being wrong
13:34:48gevaertsIs this in a standard build?
13:35:23funmanwith USE_ROCKBOX_USB defined, yes
13:35:54funmancould be in as3525 SD code too but i don't see anything obvious
13:37:15gevaertsYour kernel isn't configured with CONFIG_SCSI_MULTI_LUN=n?
13:39:24 Quit sbhsu (Ping timeout: 264 seconds)
13:40:41*gevaerts reads some backlog
13:40:59gevaertsIs this in the bootloader?
13:41:25funmanbefore you ask, the USB cable is plugged ;)
13:41:33 Quit advcomp2019 (Read error: Connection reset by peer)
13:42:01 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
13:42:49funmani'll try on windows, bbl
13:42:51 Quit funman (Quit: free(random());)
13:45:01 Quit [keanu] (Quit: Leaving...)
13:45:18 Join [keanu] [0] (
13:46:08 Quit ps-auxw (Quit: leaving)
13:46:31 Join ps-auxw [0] (
13:50:24 Quit bmbl (Quit: Bye!)
13:56:42 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
13:59:05 Join mischasworld [0] (
14:03:32 Join sbhsu [0] (
14:04:02 Join DerPapst1 [0] (
14:04:55 Quit DerPapst (Ping timeout: 240 seconds)
14:05:54 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
14:07:48***Saving seen data "./dancer.seen"
14:10:20 Join funman [0] (~fun@rockbox/developer/funman)
14:10:36funmanwindows vista shows the µSD slot
14:14:34 Quit pyro_maniac (Remote host closed the connection)
14:16:29ranmaIn wireshark I get a malformed response to the "get max lun" request
14:18:57funmanscsi_mod->scsi_logging_level = 7 turns nothing out
14:21:52funmanafaics GET_MAX_LUN just sends 1 byte : number of luns - 1
14:21:57*pamaury is always suspiscous about wireshark
14:22:34*ranma doesn't have a low-level usb signal analyzer available :)
14:25:30ranmaOk, I see it, rockbox sends a 0-length reply.
14:25:54ranmaIncidentally that's just the packet I was seeing later with the packet skew in "lsusb -vv" I think
14:26:09ranmaProbably because of dma threshold settings I'd imagine
14:28:38*funman can't use wireshark to capture usb
14:29:49ranmafunman: You need /sys/kernel/debug mounted
14:30:31funmanit's there; i have some traffic but not from the sansa it seems
14:30:32ranmaUsing /sys/kernel/debug/usb/usbmon/${N}u directly is a bit more extensive than wireshark even
14:31:29ranmaTry 0u, that should capture all ports (but you can't select that one in wireshark AFAICS)
14:31:58ranmaFor me it's usbmon1 in wireshark an 1u in /sys.../usbmon/
14:32:10funmani'll try something else: forget about trying to decode these awful numbers and let you handle that :P
14:32:31 Join dfkt [0] (dfkt@unaffiliated/dfkt)
14:33:42funman(i can't decode raw usb data)
14:34:24*ranma not very well either, that's why I was looking at wireshark first and then using that to find the request in the raw log
14:35:43funmanmc2739: did you test usb on e200v2?
14:37:19mc2739funman: yes, just testing now
14:38:21funmani think we should enable it on all AMSv1 to get more testing - what do you think?
14:43:33mc2739I would probably take the safer approach - let users build with usb support if they want to test, or provide a test build through the forum
14:46:48 Quit slck (Ping timeout: 276 seconds)
14:50:01mc2739funman: did you get the jtagged e200v1 at devcon?
14:50:09 Quit M3DLG (Ping timeout: 260 seconds)
14:50:51funmanit was mainly for finding why thumb didn't work but now that's solved
14:51:16 Join panni_ [0] (
14:52:58mc2739ok, I was wondering if it could be used to find why e200v1 line out level varies with volume. I can't find why it does in Rockbox, but not in OF.
14:54:29 Join slck [0] (Venci@Slackware.SlackPix.Com)
14:55:06funmanyou could ask Bagder to ship it to you
14:59:25 Quit mischasworld (Ping timeout: 245 seconds)
15:03:37 Join antil33t1 [0] (
15:03:40 Quit antil33t (Disconnected by services)
15:07:06ranmaGreat! I found the bug!
15:07:30ranmaThe setup request _before_ the get luns is 'SET_CONFIGURATION', which is handled partly in hardware
15:08:01ranmaIn response to that rockbox sends a 0-length response, which the hardware also has already sent.
15:08:19ranmaSo the host receives that 0-length response as response for the 'get luns' command
15:08:37ranmaAnd thus the 1 packet skew on EP0 was introduced...
15:20:43ranmaWhat happens if I call wakeup_signal() _before_ wakeup_wait()?
15:24:36ranmaNever mind, I can avoid that
15:25:28 Join komputes [0] (~komputes@ubuntu/member/komputes)
15:25:41funmanit's ok
15:26:58funmanwakeup_signal just sets a variable and wakeups the thread if it was sleeping, else the thread will see that it doesn't need to sleep by reading the variable
15:27:21funmanif wakeup_signal comes from interrupt context you can't know if it'll happen before wakeup_wait() anyway
15:27:29 Join Szpila [0] (~sszpila@
15:28:01CIA-6New commit by 03ranma (r27051): Hardware too smart can complicate things... We have to intercept the usb_core reply to SET_CONFIGURATION.
15:28:31funmansomeone wants to run test_codec on a thumb PP build and compare with normal build?
15:29:42CIA-6r27051 build result: All green
15:35:46 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
15:37:08 Part Szpila
15:37:40funmanranma: PANIC ep1 CNAK needed 173 retries CTRL=420. (the . is a non printable character i think)
15:38:17funmanwriting some mp3 on fuzev1 internal, with µSD mounted
15:40:49 Quit funman (Quit: free(random());)
15:42:03CIA-6New commit by 03ranma (r27052): Threshold seems to affect the CNAK retries needed. ...
15:42:26ranmafunman: Yeah, I noticed that
15:43:41CIA-6r27052 build result: All green
15:43:41 Quit t0rc (Quit: Leaving)
15:44:11CIA-6New commit by 03ranma (r27053): Remove '\n' from panicf
15:45:59CIA-6r27053 build result: All green
15:49:00 Quit r0b- (Read error: Connection reset by peer)
15:49:16 Join r0b- [0] (~nnscript@
15:49:27 Quit r0b- (Client Quit)
15:55:17 Part Zagor
15:55:45 Quit Luca_S (Quit: CGI:IRC (Ping timeout))
15:55:50 Join Zagor [0] (bjst@rockbox/developer/Zagor)
15:58:36CIA-6New commit by 03ranma (r27054): Unconditionally boost cpu and remove CNAK retry panic
15:58:53 Join DataGhost [0] (~dataghost@unaffiliated/dataghost)
16:00:16ranmaHmm, using OF, reading from internal SD and uSD is both 4.2MB/s, using RB I get 3.4MB/s internal and 1.8MB/s from the slot
16:00:16 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
16:00:16CIA-6r27054 build result: All green
16:03:48ranmaAnd 10.9MB/s using a good card reader :)
16:06:52JdGordonso a bit of room for improvement :)
16:07:51***Saving seen data "./dancer.seen"
16:07:55 Join r0b- [0] (
16:10:53 Quit wombat23 (Ping timeout: 240 seconds)
16:11:46 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
16:12:22 Join wombat23 [0] (
16:13:00 Quit dfkt_ (Client Quit)
16:13:11 Join jfc [0] (
16:20:28mc2739ranma: should I see microsd card now? works in windows not ubuntu (e260v2 r27054)
16:21:29JdGordonbieber: hehe, I'm getting good at crashing the editor :)
16:22:33JdGordon%?mh<;|;> will crash it :) (that should be legal in the new parser, but it is illegal in the old one)
16:24:53ranmamc2739: Does it work in OF?
16:25:36ranmaIn Linux one questions is always: Do you have multiple lun support enabled?
16:28:59 Join MethoS- [0] (~clemens@
16:29:46mc2739multiple lun is enabled (e280v1 works) and yes it works in OF
16:33:35 Quit antil33t1 (Ping timeout: 245 seconds)
16:34:16 Join antil33t [0] (
16:42:59mc2739ranma: never mind - it can see microsd now - something flakey in my Ubuntu (8.04)
16:43:58 Quit robin0800 (Remote host closed the connection)
16:45:47 Join halmi_ [0] (
16:48:07 Quit halmi (Ping timeout: 240 seconds)
16:49:38 Join bmbl [0] (
16:49:38 Quit bmbl (Changing host)
16:49:38 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
16:51:47 Join anewuser [0] (anewuser@unaffiliated/anewuser)
17:03:27 Part Zagor
17:04:35 Quit pamaury (Remote host closed the connection)
17:04:41 Quit r0b- (Ping timeout: 260 seconds)
17:07:41 Join Zigtown [0] (
17:09:07 Quit ssorgatem (Remote host closed the connection)
17:12:49 Join dfkt_ [0] (~dfkt@unaffiliated/dfkt)
17:15:06 Quit dfkt (Disconnected by services)
17:15:08 Nick dfkt_ is now known as dfkt (~dfkt@unaffiliated/dfkt)
17:18:15 Join ssorgatem [0] (~ssorgatem@
17:22:30 Join joeyg [0] (
17:26:27 Quit AzureSky (Read error: Connection reset by peer)
17:26:27 Quit JdGordon2 (Quit: leaving)
17:31:33 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
17:32:49 Join r0b- [0] (
17:35:58 Join kugel [0] (~kugel@rockbox/developer/kugel)
17:44:25 Quit t0rc (Remote host closed the connection)
17:47:54 Part giovanni ("Sto andando via")
17:47:55 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
17:49:24CIA-6New commit by 03ranma (r27055): Enable USE_ROCKBOX_USB on all AS3525v1, confirmed to work on C200v2 and E200v2, I'm confident it should work on the others too.
17:51:13CIA-6r27055 build result: All green
18:02:00 Quit GeekShadow (Quit: The cake is a lie !)
18:02:17 Quit petur (Quit: *plop*)
18:03:26 Join Luigi12 [0] (
18:07:55***Saving seen data "./dancer.seen"
18:08:21 Join rick_w [0] (
18:08:27 Part rick_w
18:10:23 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
18:13:18CIA-6New commit by 03bertrik (r27056): Fix wrong comments in tuner-clipplus.c
18:14:50CIA-6r27056 build result: All green
18:15:50 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
18:16:30Luigi12I have a Gen 5 iPod (Video) with Rockbox. I tried an Oker brand FM transmitter that plugs into the dock port and it doesn't work. I am wondering if the Accessory Power brand Flexpod X2 FM transmitter (also dock port) works. Neither are listed on the Wiki.
18:16:58kugelpamaury: how far is usb on amsv2?
18:17:57gevaertsLuigi12: if it's not on the wiki, people here probably won't know. Did you turn on accessory power?
18:18:22Luigi12gevaerts: with the Oker one, yes I tried that
18:18:40*gevaerts doesn't know much about ipod accessories
18:18:56Luigi12ok, thanks. You're probably right that if it's not on the wiki it's unknown.
18:19:44 Quit dfkt (Ping timeout: 265 seconds)
18:20:29bertrikEven if it's not on the wiki and not working, it might still be worth to update the wiki with this information.
18:21:04Luigi12true, but I only tested one of them. It's also possible I just got a defective one. Hard to be sure.
18:21:30Luigi12it was only $10-$15, so it wouldn't surprise me
18:39:47 Join pamaury_ [0] (
18:42:19 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
18:43:34 Quit pamaury (Ping timeout: 248 seconds)
18:44:09pamaury_kugel: I wrote the interrupt handler, the send/recv routine. But it doesn't work yet. I have a problem witht controls transfers: rockbox only sends 8 bytes instead of [insert quantity>8 here]. But I only work one or two hours yesterday. I hope it will soon work :)
18:44:18 Nick pamaury_ is now known as pamaury (
18:44:30 Quit pamaury (Changing host)
18:44:30 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
18:45:04kugelpamaury: is it much different from amsv1?
18:45:55pamauryIt's similar to the nano2g but with a more recent controller. Perhaps the two drivers will merge one day
18:46:03 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
18:51:27CIA-6New commit by 03bluebrother (r27057): Rename SA9200 manual platform file to use the actual target name.
18:51:31CIA-6New commit by 03bluebrother (r27058): Make skin_parser Makefile more silent.
18:51:53 Join domonoky1 [0] (
18:53:04CIA-6r27057 build result: All green
18:53:33 Quit domonoky (Ping timeout: 265 seconds)
18:54:08saratogafunman: the assembly code will probably have different rounding in mp3 vs. PC
18:54:41CIA-6r27058 build result: All green
18:57:25 Quit bmbl (Ping timeout: 265 seconds)
18:57:56bertrikpamaury, are you sure that it's not the OS just sending/requesting 8 bytes?
18:58:35bertrikIIRC, windows initially requests only part of the device descriptor, just enough to discover the packet size
18:58:58pamauryit would not do so 100 times :)
18:59:33pamaurywith the analyzer, I see the host asking for more data but the device refusing to transmit so yeah, I'm sure it's a device problem
18:59:53bertrikI was referring to point 3 and 4 of
19:01:15bertrikI could have a look today in the OF to see why some fuze v2s refuse to upgrade to rockbox
19:01:33pamauryActually, the strangest thing is that the device transmit only 8 byte but reports as having transfered 18...
19:02:29bertrikHow do you know it transmits only 8, by means of the hardware USB protocol analyser?
19:02:35 Join mitk [0] (
19:04:24mitkranma: Do I have to build new bootloader to test r27055?
19:04:37mitkon fuzev1
19:04:54 Join jgarvey [0] (
19:06:45 Join stoffel [0] (
19:07:36 Quit stoffel (Remote host closed the connection)
19:07:37 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
19:08:03 Join stoffel [0] (
19:08:19 Quit stoffel (Remote host closed the connection)
19:09:21bertrikdomonoky1, how's the in-core TTS project going?
19:09:31domonoky1not good.
19:10:37domonoky1i dont hear much from this student, and sent him again a mail in which i threaden him to let him fail at midterm if he didnt improve..
19:10:51domonoky1still no answer :-/
19:10:58 Join stoffel [0] (
19:11:09 Nick domonoky1 is now known as domonoky (
19:11:21 Quit domonoky (Changing host)
19:11:21 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
19:13:06saratogamc2739: doesn't the e200v1 line out volume vary because we change the mixer gain ?
19:13:36saratogaAFAIK the line out amp should be locked, but when you change the volume in RB, you're changing the volume out of the mixer which feeds the line out amp, so it'll change anyway
19:13:40 Quit DerPapst1 (Quit: Leaving.)
19:13:55bertrikAs far as I understood from the last mails, is that the TTS project is basically stalling because of technical reasons: requiring malloc and not being real-time
19:13:58ranmamitk: No, just the rockbox binary
19:14:49mitkranma: thanks
19:15:04 Quit mitk (Quit: CGI:IRC)
19:15:06domonokybertrik: yes, thats what i know too.. but he wont revolve that if he doesnt heavily improve his communication with the community..
19:15:34bertrikIf those are blockers, maybe the goals of the project just need to be moved
19:16:53domonokyyes.. but he has to talk to me for that.. till now i had to push him for every statusmail. :-/
19:17:05domonokyand he never asked for help..
19:18:01domonokyi dont have a problem with changing goals, if there are stoppers. But i have a problem with students who dont communicate.
19:18:22 Quit MethoS- (Remote host closed the connection)
19:20:27 Join MethoS- [0] (~clemens@
19:22:27 Quit flydutch (Quit: /* empty */)
19:22:57 Nick dfkt_ is now known as dfkt (dfkt@unaffiliated/dfkt)
19:23:48ranmaAny reason that sd-as3525.c doesn't use 4bit mode?
19:24:05kugelranma: bugs
19:24:06ranmaDoing so increases the read read from 3MB/s to almost 10MB/s over USB :)
19:24:24kugelFlynDice had it enabled once but it caused problems
19:24:36kugelbut the problems might have disappeared with the recent fixes in that area
19:25:10pamaurygggr, the usb analyzer software don't like virtualbox !
19:26:52 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
19:27:14*ranma admittedly hasn't tried test_disk yet *enables test plugins in build and rebuilds*
19:28:21dfktcongratulations on usb for amsv1 :) my clip is so much more fun when it doesn't rebuild the database and doesn't burn out/in the screen pixels needlessly
19:28:37*ranma also wonders if it is a bug that sd-as3525v2.c is issuing SD_SET_CLR_CARD_DETECT after SD_SET_BUS_WIDTH and before setting the controller to 4bit
19:28:54ranmadfkt: Thanks, so clip works too?
19:28:58bertrikdomonoky, ok, I hope he starts communicating again.
19:29:01dfktclipv1, yes
19:29:06bertrikranma, seems to work here! :P
19:29:14dfktseems to charge, and transfers files just fine
19:30:26domonokybertrik: yes, lets hope he gets better with that. If not, he will be out at the 12.July :-)
19:32:16*pamaury wonders why rockbox does not detect usb disconnect... Perhaps I have to do something myself in the driver
19:32:35 Quit anewuser (Quit: Sixth edition of 25 samples + 48 hours + detroit party (july 31) (maybe streamed online!))
19:32:49 Quit ssorgatem (Quit: Konversation terminated!)
19:33:08 Join ssorgatem [0] (~ssorgatem@
19:33:36pamauryI got my first set address request \o/
19:34:47*simonrvn is impressed with all the advances "congrats people"
19:35:02bertrikIt would be a pity if he would fail just for not communicating, I can't blame him for getting stuck on a tough technical problem. Maybe we should make this clear to him and propose possible alternatives to give it a chance to succeed.
19:36:37 Join binaryhermit [0] (
19:38:30 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
19:40:06*ranma also wonders about some of the mci_delay() calls, with the previous command neglecting to set the controller flag for the card response even if the command does send a response
19:41:02ranmaThe comment "Response is requested due to timing issue" is misleading I think, most commands always send a response, you just have to tell the controller which one, you don't 'request' it
19:42:50bertrikIf you can fix something there, please do :)
19:42:55 Nick Zigtown is now known as Zigtown07 (
19:43:06 Quit Luigi12 (Remote host closed the connection)
19:44:22 Join Szpila [0] (
19:44:28*ranma just happens to have written an sd driver for his Zaurus pda (pxa255) to use SDHC cards on the stock firmware :)
19:45:23 Nick Zigtown07 is now known as Zigtowb (
19:45:26 Nick Zigtowb is now known as Zigtown (
19:47:16bertrikranma, somehow I don't see the ##MUSIC# directory on the OF USB mode, but I do see it with the RB USB mode on my clip v1
19:48:43ranmaIs that some OF-related directory? Maybe OF hides it on purpose?
19:48:58 Quit MethoS- (Remote host closed the connection)
19:49:11bertrikyes, it is, and I think it's supposed to be hidden from the OF point of view
19:49:54 Join WalkGood [0] (~FU@unaffiliated/walkgood)
19:50:11bertrikUSB MSC mode accesses the player as a block device, so I find it a bit surprising to see something different on the file/directory level
19:53:10 Join bug2000 [0] (~bug@unaffiliated/bug2000)
19:53:34 Join DerPapst [0] (
19:53:34ranmabertrik: Yep, OF mangles the root directory:
19:53:38ranma-0x0001ea80: 23 23 4d 55 53 49 43 23 - 20 20 20 10 00 00 00 00 ##MUSIC# .....
19:53:39ranma+0x0001ea80: 23 23 4d 55 53 49 43 23 - 20 20 20 18 00 00 00 00 ##MUSIC# .....
19:53:43ranma-0x0001eb80: 23 23 50 4f 52 54 23 20 - 20 20 20 10 00 00 00 00 ##PORT# .....
19:53:43ranma+0x0001eb80: 23 23 50 4f 52 54 23 20 - 20 20 20 18 00 00 00 00 ##PORT# .....
19:53:54 Join stripwax [0] (
19:54:27bertrikwow :P
19:54:46*pamaury managed to get two panic on sd on the raw about multi read that failed
19:55:02 Part AzureWurk
19:55:41ranmaflags bit 0x08 is 'volume lbel'
19:55:54TheSevenhiding things, eh?
19:56:39 Join AzureWurk [0] (
19:58:29topikgreat work on the amsv1 usb guys. seems to work fine so far on my fuze v1
19:59:29pamauryI must do something wrong on this controller. I only get a set address randomly
19:59:33topikonly thing i don't understand is that screen with the usb plug image. why not let users just go through the menu menu and such?
19:59:58pamauryBecause you can't if you device act as mass storage
20:00:20bertrikthe PC has basically locked the file system
20:00:44pamauryThe filesystem is not in a consistent state at every time so you can't access it reliably
20:00:50saratogalet users go through the menus, just disable selecting any menu items!
20:01:03topikah ok. i'd kinda like to watch the debug menu to see charging and such
20:01:11TheSevensaratoga: glyph cache misses!
20:01:38saratogadisable custom fonts in rockbox!
20:01:47*TheSeven just committed that yesterday
20:02:12pamauryI keep wondering why the usb core keeps sending the ack before it received the data or treated the request ! That's not sensical, I should have changed that before. Does someone know why we do this ? gevaerts once told me that a particular deviced required it but that seems strange
20:02:36kugelgevaerts: any suggestion for the "use rockbox c library" configure option?
20:03:34TheSevenpamaury: *send* or *receive* an ack before handling the data?
20:03:57TheSevenwe need to set up receiving the ack before we send the data associated with it to prevent a race condition
20:04:07*TheSeven spent several days debugging that crap
20:04:10pamauryyes, that's what the code basically does. At the usb level, thing happend in the right order of course
20:04:26pamauryWhich race condition ? that's nonsense
20:04:58TheSevenif the ack will come in before we set up the endpoint for it, the controller will reply with STALL
20:05:30pamauryjust set NAK until it is treated
20:05:34 Join Luca_S [0] (
20:05:43TheSevenit doesn't do that
20:05:53TheSevenNAK = no buffer space left or whatever
20:06:02TheSevenSTALL = i don't know what to do with it
20:06:19 Join anbu_yuki [0] (
20:06:22TheSeventhe latter will hit if the endpoint doesn't await a packet
20:06:27pamauryNAK - Reports that the device temporary cannot send or received data. Also used during interrupt transactions to inform the host there is no data to send.
20:06:53TheSeventhe controller doesn't know that this is temporary
20:06:55pamauryif you set NAK, the host will ask until you set stall or ack or timeout
20:07:12TheSeventhe *chip* will set stall, if we don't set up the endpoint
20:07:40TheSeventhat lovely synopsys core is doing that kind of things in hardware
20:07:48pamauryWhat do you mean by setup the endpoint ?
20:07:59***Saving seen data "./dancer.seen"
20:08:15TheSevenset the EP0 OUT transfer count reg accordingly (1 packet, 0 bytes)
20:08:29TheSevenand probably an ACTIVE flag somewhere
20:09:30pamauryYou can always setup it in the driver just after you received the setup and ask it to nak until the core finally ack, that would be strange otherwise
20:10:29pamauryAnyway, it is probably not the reason of my current bug but I'll have a close look at this issue
20:10:33TheSevenyou basically can't tell it to just NAK
20:10:43TheSevenwhat's wrong about setting up the OUT transfer before the IN one?
20:10:43pamaurythere is a set/clear NAK flag
20:11:47pamauryyou should only ack when you are done, that's just it. You should not ack before you are done with the request iirc
20:12:12pamauryAnd it ask a amusing effect: you see sequences like:
20:12:33pamauryusb_drv_recv(EP_CONTROL, NULL, 0);
20:12:33pamauryusb_drv_send(EP_CONTROL, response_data, 1);
20:13:39anbu_yukihi guys , i just wanted to voice a start of a port for symbian s60 of rockbox as an application well s60v5
20:13:50topikranma: do you have any usb writing speed numbers ams v1 ?
20:14:08pamauryWhich basically mean: wait for the data to be sent, whereas you wait for the host to ack your data. that's the other way around
20:16:25anbu_yukicurrently investigating ways of directly accessing audio without the use of api decode ...
20:16:54dfkttopik - i got about 1.4MB/s on my clipv1, according to teracopy
20:17:19topikwindows tells me ~1.8MB/s on my Fuze v1 to internal memory
20:17:41 Quit stripwax (Quit:
20:17:58topikany speed is fine though as long all bytes arrive and in the same order and i never have to see 'refreshing your media' again
20:18:13dfktindeed :)
20:19:58TheSevenpamaury: there are indeed very weird hunks in my patch
20:20:03ranmatopik: 3MB/s was my result to internal memory (both OF and RB)
20:20:10TheSevensome of which definitely weren't changed by me
20:20:13kisakhow much of the AMSv1 usb is shared with AMSv2?
20:20:29TheSevenlooks like i accidentally reverted part of another commit that was mergable without conflicts
20:20:34ranmakisak: None I think, they are too different
20:20:55kisakranma: ok
20:20:55topikranma: that's a bit faster than i'm seeing so far
20:21:06pamauryTheSeven: I didn't try to apply the patch so I didn't told you there were bad hunks :)
20:21:30TheSeveni mean that one you pointed out yesterday that was apparent nonsense
20:22:00TheSevenand i just found another one where a whole function was moved inside the file
20:22:45pamauryyes I saw that
20:23:07TheSevenI definitely didn't do that.
20:23:22*TheSeven wonders how that could have slipped in
20:23:35pamauryreally ?
20:24:09TheSevenapparently svn diffed that file to a wrong revision or something
20:24:25ranmaOn windows usb device may be mounted write-through, that might reduce write performance. How does OF perform on comparison?
20:25:09topiki don't want to see the OF ever again :(. but i'll try that now
20:25:12 Quit stoffel (Remote host closed the connection)
20:27:12 Quit Zigtown (Remote host closed the connection)
20:29:10topikranma: OF does about 5.5MB/s
20:32:18topikthat's writing. reading using the OF is almost the same, slightly faster at around 6.0MB/s
20:34:05CIA-6New commit by 03kugel (r27059): Add an advanced build option to force compiling and linking our reduced C library (probably useful for the sim).
20:34:52ranmaI may have tested before this revert, I'll have to retest.;revision=27044
20:35:55CIA-6r27059 build result: All green
20:36:19topikperhaps the lack of that 4-bit goodness uSD access is the issue, nothing USB related
20:37:04kugelthe amsv1 sd driver is *slow*
20:37:30ranmaYeah, seems like it. With 4-bit and zero copy I had 10MB/s on read, with just 4-bit I now have 4.3MB/s (same as OF), have to recompile again for 1-bit mode
20:38:33kugelthe small sector buffer also makes it slower
20:38:42topikif the sd drver barely makes 2MB/s, i can't complain usb gets 1.8
20:38:54topiknot that i would complain at any number
20:39:00kugelNUA in that pdf means num_unaligned_sectors (i.e. the size of the sector buffer)
20:39:27ranmaDown to 2.4MB/s on read in 1bit-mode
20:40:01ranmaWould be really nice to get the zero copy patch working properly
20:41:46ranmakugel: Should I try committing the 4bit enable? It has passed test_disk write+verify here...
20:42:10kugelit has passed that test before
20:42:33kugelmaybe talk to FlynDice before
20:43:14kugelranma: browsed the svn history?
20:43:31 Quit Xerion (Quit: )
20:51:46CIA-6New commit by 03bieber (r27060): skin_parser: Fixed bug handling sublines with no content
20:52:57kugel"fixed bug handling" :)
20:53:16CIA-6r27060 build result: All green
20:54:07 Join giovanni [0] (~giovanni@
20:57:46 Part anbu_yuki ("Leaving")
20:58:14*TheSeven glares at test_disk WRITE&VERIFY
20:59:10TheSeven"Data size: 307200KB Wrt 12725KB, 56426KB left"
20:59:18TheSeventhat thing can't count!
21:03:14 Quit preglow (Ping timeout: 240 seconds)
21:04:26 Quit aevin (Ping timeout: 260 seconds)
21:06:59 Join aevin [0] (eivindsy@unaffiliated/aevin)
21:13:44 Join anbu_yuki [0] (
21:19:50anbu_yukiso far i am mapping out what can be done with the symbian s60v5 sound api without calling in the actual api level multimedia decoder , which seems is what most apps that playback sound do
21:20:42 Part Szpila
21:21:26kugelanbu_yuki: how's that rockbox related?
21:21:48kugelanbu_yuki: nevermind, reading backlog helps
21:21:53anbu_yukii am wanting to port rockbox as an application to symbian s60v5
21:22:32anbu_yukiit seems the sdk does not want to allow much core level hardware control of the codec ... other then volume control
21:23:39 Join T44 [0] (
21:23:55kugelanbu_yuki: you basically only need to pass pcm data, that shouldn't need a multimedia decoder api
21:24:00anbu_yukii guess the first thing to do is make something that can send maybe a simple pcm stream to the audio codec without using the api for media
21:24:12 Quit Luca_S (Quit: CGI:IRC)
21:24:40 Quit Topy44 (Ping timeout: 260 seconds)
21:24:41anbu_yukiyes... that is the issue , it does not seem to be very well documented :P ...
21:24:44kugelthat's it basically, the sdl port doesn't do much more
21:25:03kugelanbu_yuki: maybe our symbian insider Torne has an idea
21:25:45anbu_yukithat would be great :) , is this the first such project , i have searched for a while for an s60 port ?
21:25:54anbu_yukii may look at how scummvm does sound
21:26:07kugelI'm doing rockbox as an application for gsoc currently
21:27:14anbu_yukii see , so your efforts may aid in how the ui is handled on touch screen devices like s60v5 :)
21:27:22anbu_yukithat is also in my blueprint
21:27:41anbu_yukiis how to handle the touch screen , lack of physical buttons as it seems most low cost s60 devices offer
21:28:23 Join Topy44 [0] (
21:31:34 Join Buschel [0] (
21:31:43 Join Szpila [0] (
21:31:51topikyour progress wiki page is hard to find, kugel
21:33:12anbu_yukiit seems i found mention to some accelerated sound library in s60v5 .... will start at that point then look at scumm
21:33:43TheSevenpamaury: some not too promising performance numbers:
21:33:51TheSevensome of them are quite confusing
21:34:21TheSevenremember we're using a very conservative writeback policy right now
21:34:54pamauryperhaps this policy is too strict. Notably for the FAT
21:36:05TheSevenhowever, this will only hurt write access, where it already has a positive impact
21:36:17TheSevenfor some reason we're significantly slower on the reading side though
21:36:44TheSeveni can't believe that this is just the dirty page checking code
21:38:05pamauryPerhaps the cache does not behave in the same as the previous code and this has a negative impact. Thing do not stay in the cache for long but that's strange
21:41:16Torneanbu_yuki: i know nothing about app level stuff like sound, i'm afraid :0
21:41:56Tornebut symbian foundation forums and/or forum nokia should know :)
21:42:54anbu_yukisame thing goes for me , its going to be one hell of a learning experiment , thx for the reply... I was planning on joing both the symbian group and the nokia developers group to ask around
21:43:25anbu_yukiit just seems perfect port given how bad the normal audio player is
21:44:19kugeltopik: RockboxAsAnApplication2010
21:45:27anbu_yukishould i add any findings to the rock box as and application project or start a new project given its for a non discussed target ?
21:46:25bluebrotherdomonoky: ping
21:46:34kugeltopik: I added a note to RockboxAsAnApplication
21:46:34topikkugel: yeah, i finally found it. one link on the gsoc2010 page.
21:46:50domonokybluebrother: pong
21:47:03bluebrotherdomonoky: what do you think about Rockbox Utility 1.2.7?
21:47:06bertrikwhat does do?
21:47:09amiconnTheSeven: Did you see my explanation in the log?
21:47:11domonokygo for it.. :-)
21:47:20TheSevenamiconn: explanation for what?
21:47:38bluebrotheralso, how to handle FS #11420? I'd like to include that, but there hasn't been a response yet :/
21:47:41amiconnStack usage of screendump
21:47:50TheSevenno, didn't catch it
21:48:36amiconnTheSeven: Btw, test_disk counts perfectly ok. "Data Size" is the total size. "Wrt" is the block size transferred in the current transaction. "left" is the remaining size
21:48:41domonokybluebrother: you can just code that from scratch :-)
21:48:49TheSevenamiconn: hmm, a single line are like 300 bytes, not 2K
21:49:01amiconnWhat target?
21:49:13bluebrotherdomonoky: sure, the change is rather trivial :)
21:49:16TheSeven176 x 132, 16bit
21:49:30bluebrotherI guess I'll simply redo that then and reject the patch
21:49:43 Part AzureWurk
21:49:47 Join AzureWurk [0] (
21:50:35amiconnSo the buffer is 352 bytes (16bpp)
21:50:49amiconnThere is also a MAX_PATH sized filename buffer
21:51:02amiconnÍi.e. another 260 bytes iirc
21:51:27*TheSeven will retry with aggressive caching
21:52:12amiconnSo definitely not 2K. That must be from somewhere else
21:52:38TheSeven1.5K stkov's, 2K does not.
21:53:16amiconnThe usb stack probably has more things on stack
21:53:22amiconn*usb thread
21:54:40*bertrik noticed that the clock on his clip+ was nearly exactly 187 days off
21:55:15*amiconn noticed another annoying quirk on the clip+, apart from the muffled sound
21:55:42amiconnThe voice ui is not very usable, due to shutup() not working (at least not properly)
21:56:10amiconnIf you scroll through the menu quickly, you have to wait until all the inbetween items are spoken -> slow and cumbersome
21:56:44bertrikamiconn, I though dfkt measured output and found a basically flat frequency response
21:56:59*amiconn needs to test whether this is clip+ specific, or other (older) targets are affected as well
21:57:47bertrikthere is a patch from markun floating around that increases the amplifier bias current, maybe that one helps
21:58:13amiconnDid he measure with or without load?
21:58:18dfktwith load
21:58:26*amiconn wonders whether his earphones might be the culprit
21:58:46dfkthowever, some phones don't seem to work well with the clip - like the 65 ohm koss ksc75 - they indeed sound muffled or under-powered from the clip
21:59:00dfktwhile some 300 ohm sennheiser sound nice, just a bit quiet
21:59:04 Quit kugel (Remote host closed the connection)
21:59:07*amiconn has Sony IEMs (16 ohms)
21:59:14bertrikWhat impedance is considered "normal"?
21:59:23dfkthmmm, i had no issues ever with low impedance phones
21:59:24anbu_yukiyes the hd650 actually work well with the clip , sans the volume levels
21:59:30dfktbertrik, the standard is usually 16 ohm
21:59:31amiconnWell, earphones/ IEMs are usually 16 ohms today
22:00:05amiconnLight headphones are often 32 ohms, some are 64 ohms (I have 64 ohm earphones as well)
22:00:14dfkti remember reading in the AMS specs that the chip does some "impedance matching" - maybe it's not working well with some phones?
22:00:19bertrikWeird how high-impedance phones are more affected, I would have expected the exact opposite
22:00:42bertrikHigh-impedance -> less load on the amplifier
22:01:06amiconnIiuc it seems like low impedance is more affected
22:01:13anbu_yukithe hd25 i-ii are 70 and seem to work great with clip+ via both stock and rockbox
22:01:45dfkthere's a few comparisons, with different loads:
22:02:12bertrikOr maybe it means you have to increase the volume on the player quite a bit to get comparable volume in the earphones
22:02:22topiknice exchange on the official sansa clip+ forum about this topic involving mr saratoga :)
22:02:27dfktmaybe that weird crosstalk behavior of rockbox has something to do with some phones being perceived as muffled?
22:02:54ranmaIt seems just defining 'STORAGE_WANTS_ALIGN' in system-target.h gives me playback glitches
22:03:08ranmaEven without the changes in sd-*.c
22:08:02***Saving seen data "./dancer.seen"
22:10:44 Quit Strife89 (Quit: Reboot to Linux.)
22:12:22 Join funman [0] (~fun@rockbox/developer/funman)
22:13:04funmanranma: check FS #11429
22:13:41CIA-6New commit by 03bluebrother (r27061): Don't close libusb device handle if opening failed. ...
22:14:32funmani also try the sd*.c changes alone without defining STORAGE_WANTS_ALIGN and got playback trouble too
22:15:11 Join Strife89 [0] (
22:15:14CIA-6r27061 build result: All green
22:16:20 Join Dekkard [0] (
22:17:04bluebrotherdomonoky: what do you think about merging rbutil/ and utils/? Now that we have the theme editor this separation looks a bit weird.
22:19:33funmanranma: also r24054 (revert 4 bit mode on AMSv1)
22:20:02funmanthese days it might just work
22:21:51dfktbertrik, is there a task on the tracker about your clock/samplerate fix? i couldn't find anything
22:22:01CIA-6New commit by 03bluebrother (r27062): Bump Rockbox Utility version to 1.2.7.
22:23:08bertrikdfkt, FS #10906, it should work now on both amsv1 and amsv2 (although may also break soon when PLL B is used for USB)
22:23:37 Quit Jaykay (Read error: Connection reset by peer)
22:23:39CIA-6r27062 build result: All green
22:23:44dfktthank you... will keep in mind to revert it from my builds :)
22:24:05TheSevenfunman: does STORAGE_WANTS_ALIGN actually do anything besides aligning some static buffers in memory?
22:24:11CIA-6New commit by 03bluebrother (r27063): Tag Rockbox Utility 1.2.7 release.
22:24:35funmanTheSeven: i think that's all it does
22:24:40 Join kugel [0] (~kugel@rockbox/developer/kugel)
22:24:50TheSevenso I doubt it can cause trouble
22:25:02kugelbluebrother: you are in doubt that Judas Iscariote is his real name?
22:25:14funmanit's ranma who has trouble not me :-)
22:25:37bertrikdfkt, we're now back to 240 MHz PLL now for amsv2 so 1.1% error, amsv1 is at 0.15% error
22:25:58funmani tried to use 384MHz on amsv1 but it didn't work on fuzev1
22:26:36dfktbertrik, i see... i noticed the task is about amsv1, and rather old
22:26:38bluebrotherkugel: well, I simply don't know it. And as he put the nick in the real name field in FS I simply don't consider it a real name.
22:26:43kugelbertrik: what happened to the 248 pll idea?
22:27:01funmanreverted in r26979
22:27:06bluebrotherbut that change was trivial anyway, so it's not much of a difference.
22:27:08bertrikkugel, it causes problems recognising microSD on some fuze v2
22:27:34 Quit mt (Ping timeout: 240 seconds)
22:27:39kugelbluebrother: it sounded like you accuse him of not stating his real name
22:27:51kugelbertrik: I thought that was the 384 one?
22:28:08Dekkardi tried to load a theme and got it locked up: data abort at 0005C3D) (0)
22:28:16Dekkardany ideas?
22:28:35bertrikJudas Iscariote sounds like a false name indeed
22:28:45bluebrotherkugel: well, that wasn't my intent. I just wanted to state that we do require real names (and I'm not convinced that the name he put is actually his real name). I'm happy to add him to CREDITS if he shows up to clarify this issue.
22:28:46dfktjesus would agree :p
22:28:51kugelbertrik: what about the 244 pll idea?
22:30:07bertrikkugel, that's only marginally different from 248, so will probably trigger the same problems
22:30:24bluebrotherbtw, google shows that Judas Iscariot is the name of a band.
22:31:25dfktit's also the bad guy in the bible, new testament :p
22:31:26kugelbertrik: it's as marginally different from 240 :P
22:31:33kugelworth a try IMO
22:31:52 Quit anbu_yuki (Quit: Leaving)
22:31:56bluebrotherIt's also the name of a biblical figure. While I could imagine that someone has exactly the same name I appears somewhat unlikely to me
22:31:56gevaertskugel: I see I managed to avoid answering your question by waiting long enough :)
22:32:26dfktbluebrother, especially since it's an evil figure in the bible
22:32:40bluebrotherdfkt: the biblical figure was even the first thing I was thinking about, but google turned up with the band first :o
22:32:41bertrikI'm suspecting a problem in SD driver somewhere rather than a too high clock frequency for the CPU
22:33:12TheSevenfunman: did you see my updated patch?
22:33:15kugelbertrik: maybe the udelay was too off with the new frequency?
22:33:34kugelIIRC it's now based on a (somewhat) calibrated busy loop
22:33:37funmanTheSeven: yes but didn't test it yet
22:33:47 Join Nisaea [0] (
22:33:48funmankugel: it just assumes cpu freq is set correctly
22:33:51CIA-6New commit by 03alle (r27064): Do not produce an extra space for non-albumart manuals
22:34:03Nisaeahello everyone
22:34:12amiconnIs there an ams datasheet somewhere?
22:34:14kugelmost of the sd detection problems could be fixed by introducing/enlarging delays somewhere in the past
22:34:29funmanhere the card is detected correctly but not mounted
22:35:04domonokybluebrother: moving rbutil and the patchers to utils someday sounds fine.
22:35:30CIA-6r27064 build result: All green
22:35:41 Join M3DLG [0] (
22:36:43NisaeaI'm currently having troubles creating my database on my iPod 5G, si I checked the forums to see if there was a similar problem reported, but well... not quite
22:36:49 Part WalkGood
22:37:00Nisaeaso I was wondering perhaps one of you might have an idea :)
22:37:00 Quit Buschel (Ping timeout: 260 seconds)
22:38:28NisaeaI put a database.ignore file in the root of the iPod to see if the contents of the ipod control folder would be ignored but it doesn't seem to be the case, did I do anything wrong?
22:38:35kugelgevaerts: good job :P
22:39:13 Join mt [0] (~mtee@rockbox/developer/mt)
22:39:21Nisaeamoreover the scanning ends up crashing at some point
22:41:13bluebrotherhmm, should try to build against a static libpng copy on linux. Let's see ...
22:41:24 Quit pamaury (Read error: Connection reset by peer)
22:41:25TheSevenpamaury: I'm not too convinced of the cache idea any more...
22:41:33TheSevensearching the cache seems to take quite a lot of time
22:41:58TheSevenso maybe some kind of sector buffer malloc pool could be the better approach?
22:42:13funmanamiconn: there's no "as3525v2" datasheet, we use the as3525"v1" one; but AMS sent me the as3543 datasheet which is the codec/PMU chip used in as3525v2
22:42:40amiconnah ok
22:43:04amiconnDo we konw the exact arm core in as3525v2? All I know is that it's arm9 v5
22:43:24gevaertsNisaea: if scanning crashes, anything else is going to be unpredictable. Which revision are you running?
22:43:25funmanthere is a cp15 register which gives the exact model
22:43:31funmanalthough frequency is just guessed
22:44:02amiconnas3525v1 says max. 250MHz. My guess is that as3525v2 is the same
22:44:47funmanyeah, the newer AMS SoC (as353x) which use an arm926ej-s are also maxed at 250MHz
22:45:10Nisaeagevaerts, where can I find this information? is it r27064 ?
22:45:34gevaertsNisaea: System->Rockbox Info
22:46:17Nisaeagevaerts, please wait a moment :)
22:46:33TheSevenNisaea: yes, r27064 is a revision number.
22:46:52Nisaeathanks TheSeven :)
22:47:40Nisaeagevaerts, it says version 3.6
22:48:08gevaertsok, then it's not the database issue some people have been seeing since a few days...
22:48:54Nisaeathe strange thing is that although there is a database.ignore file it stills finds files to add
22:49:41gevaertsdatabase.ignore only causes the system to not actually scan files for metadata, it still goes through the directory tree
22:49:42NisaeaI also put a database.unignore in an empty folder I plan to use to store the music I wish for the database but still
22:49:54Nisaeaah ok
22:50:25gevaertsThat's actually needed for database.unignore to work. Otherwise that file also wouldn't be seen :)
22:50:30Nisaeaso it's normal to see it count the files it finds while scanning
22:50:50gevaertsWhat's not normal is of course for the scanning to crash
22:50:55Nisaeaok I understand
22:51:17Nisaeayes, beats me, the first time I scanned the whole thing it worked fine
22:51:23gevaertshm, do you have at least one file it's allowed to scan?
22:51:48Nisaeanah, none since the folder where the unignore file is is empty
22:52:04gevaertsthat might actually be the problem
22:52:15Nisaeashould I put a random file in there and test?
22:52:18funman < remove OF boot on USB for AMSv1
22:52:30gevaertsWe did have a bug that caused problems in that case. I'm not sure if that ever got fixed
22:52:39gevaertsYes, just put any random audio file there
22:53:02Nisaeaok I'm testing right now
22:53:09Nisaeathanks a lot
22:53:11 Quit jgarvey (Quit: Leaving)
22:53:33gevaertsI hope it works. We prefer known issues with easy workarounds over new bugs :)
22:54:13Nisaeahaha sorry if I stepped on a mine :)
22:54:28DhraakellianUSB on AMSv1? keeeewl
22:54:32*Dhraakellian updates his fuze
22:54:48 Join stripwax [0] (
22:55:11Dhraakellianhmm... it changes the fonts while plugged in
22:55:12bertrikdfkt, do you still have markun's patch around somewhere?
22:56:17*Dhraakellian rm -r \##*/
22:57:06bertrikdfkt, never mind, I found the bits already in the data sheet
22:58:02bertrikamiconn, there's two bits in the codec that affect quality. The datasheet also mentions some power requirement, the differences are not that big.
22:58:06Nisaeagevaerts, it didn't crash where it used to, so it might take a while now (smells like good news anyway)
22:58:10 Quit kugel (Remote host closed the connection)
22:59:35bertrikI'm inclined to enable them for high quality and not care too much about the < 1 mA or so that it saves when disabling them.
23:00:01Dhraakelliannot quite so beers-worthy as getting USB enabled on the e200v1 (given its OF's lack of >2GB microSD support), but you guys still rock incredibly nonetheless!
23:00:19gevaertsNisaea: I'll be happy when it actually gets to the end :)
23:02:37Nisaearebooting :)
23:02:45funmanranma: did you test HID ? (USB_HAS_INTERRUPT in config.h)
23:03:21Nisaeagevaerts, IT WORKED! thank you very much for your help!
23:03:30 Quit funman (Quit: free(random());)
23:03:36*Nisaea is happy :)
23:03:46bertrikamiconn, yes, the bits are SUM_HP_HIQ in AUDIOSET2 and HP_BIAS in AUDIOSET3
23:04:11Dhraakellianactually, this might be beers-worthy, since it removes the need for very annoying select-pressing when I turn the car on
23:04:13dfktbertrik, just in case, it's here -
23:06:17 Quit komputes (Remote host closed the connection)
23:07:03amiconnbertrik: Is there a patch?
23:07:18NisaeaI'm so in love with rockbox, thank you all for all the work you've done, I'll recommend it to everyone I know!
23:07:30bertrikamiconn, what dfkt just posted
23:07:42Nisaeabye bye :)
23:07:58 Quit Nisaea (Quit: Ex-Chat)
23:08:08 Join funman [0] (~fun@rockbox/developer/funman)
23:09:52 Join kugel [0] (~kugel@rockbox/developer/kugel)
23:10:23 Quit stripwax (Quit:
23:11:23 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
23:12:48 Quit dfkt (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
23:13:29 Quit evilnick_B (Quit: Page closed)
23:13:53bertrikdfkt, I'm a bit puzzled by your results: the graphs at look nearly identical, basically no improvement with the hi quality bits set, while earlier graphs do seem to show an improvement.
23:15:22ranmafunman: I'm definitively seeing problems just from enabling STORAGE_WANTS_ALIGN, given that buffering.c looks kind of complex I can imagine that maybe it's rounding a pointer into the wrong direction somewhere?
23:15:32ranmaI haven't tried hid, basically only usb_storage
23:15:46 Join grawity [0] (~grawity@Because/DNS/abuse/is/
23:16:17grawity...I just noticed that weekdays in calendar.rock are off by one.
23:16:29funmanranma: buffering_reset() -> buffer_len use a pointer aligning macro
23:16:48gevaertsgrawity: what target?
23:17:04grawitygevaerts: ipodnano2g, r26925
23:17:21grawityThe Appleware calendar is okay.
23:18:10grawityFor example, according to calendar.rock, today is June 23, Tuesday.
23:18:29*grawity will file a bug when he's less sleepy.
23:19:12funmanranma: ?
23:19:57grawitymeh. rockboxsim seems okay :| I guess I just need to upgrade.
23:20:01grawityGood night.
23:20:08funmanalso the storage alignement shouldn't assume it must align to a cache line
23:20:11gevaertsIt might be target specific
23:20:34 Join stripwax [0] (
23:20:43gevaertsIt's fine on my gigabeat f. That's an older build, but I doubt anything changed in calendar.rock
23:20:45funmannow all targets that uses it require the same alignement than cache lines but it's just a random coincidence
23:21:08grawityAlso, weird, it shows: "Battery: 88% 0h -1m"
23:21:34*grawity is quite sure that worked before.
23:21:37 Quit grawity (Quit: Good night.)
23:21:40funmansomeone needs to make a battery bench i think
23:22:59ranmaplayback.c does a fixed align on a 'line size' of 16
23:23:13 Quit bmbl (Quit: Bye!)
23:25:52*bluebrother spots that he needs to do some cleanups on the 64bit linux rbutil build vm
23:27:42funmanranma: nice spot
23:28:09ranmafunman: I'm more suspicious of for example reset_handle()
23:28:09kugelthe calendar bug has been reported already IIRC
23:28:39ranmaSTORAGE_OVERLAP() seems on first glance to be an 'align down'
23:28:47 Quit stripwax (Quit:
23:28:49funmanranma: btw about 4bits mode i think it should just go in if you tested it, we have fixed other bugs since this was reverted. If you have a patch i can test tomorrow on clip/fuze
23:29:30bertrikthe calendar bug could actually be a bug in the RTC driver too I think
23:30:39kugelfunman: is that bug about randomly skipping to the next song?
23:31:17funmannope there are some skips/glitches but lossless files output white noise
23:31:46ranmaWell, lossless are probably not recoverable if you get a bit error
23:31:56ranmaWhere mp3 will 'blip' or just skip ahead
23:32:27kugelfunman: so yes or no? you mentioned skips
23:32:42funmannope = no
23:33:01kugelI knew that, thanks
23:34:39 Part Szpila
23:35:13bertrikWhat RTC does the nano2g use?
23:35:43bertrikAccording to ipodnano2g.h, it's a RTC_NANO2G but I can't find the driver for it
23:36:20 Join fml [0] (
23:36:56 Quit fml (Client Quit)
23:37:19 Join Luca_S [0] (
23:37:26amiconnbertrik: Seems it bit better with the patch. Hard to tell
23:39:35 Quit BHSPitMonkey (Ping timeout: 240 seconds)
23:40:05funmanbertrik: the driver is in target tree i think
23:40:11bluebrotherwtf? Switching themes causes a data abort on my mini?
23:40:26bertrikfunman, yes, I just saw that :)
23:41:17Dekkarddoes it on my nano 1g too
23:41:44Dekkardi tried to load a theme and got it locked up: data abort at 0005C3D) (0)
23:43:25bluebrotherBagder: can you move rbutil 1.2.7 files to the download server?
23:43:26 Quit efyx (Remote host closed the connection)
23:43:57*Bagder gets to work
23:44:20simonrvnanyone else getting invalid ELF header with the simdisk/_temp_codec0.dll file?
23:44:36simonrvncurrent svn build, sansafuzev2
23:44:44CIA-6New commit by 03ranma (r27065): Give 4-bit mode another chance on AMSv1 since other bugs were fixed.
23:45:33bluebrotherwow, now I get the data abort immediately after startup :(
23:45:40bertrikranma, does this also apply to internal storage?
23:45:49bluebrotherdo we have map files around for current builds?
23:46:10funmanranma: btw, we could enable it in bootloader too
23:46:22CIA-6r27065 build result: All green
23:47:33Bagderbluebrother: done!
23:48:16bluebrotherBagder: thanks!
23:48:33bluebrothergnah. Haven't seen Rockbox that unstable since years :(
23:49:27*amiconn did :\\
23:49:35CIA-6New commit by 03ranma (r27066): The mci_delay()s should be unnecessary if we properly tell the controller to wait for the response these commands send back to the host.
23:50:05ranmafunman: Yeah, I thought so too
23:50:25simonrvnsorry, in the simdisk ^^^
23:50:27bluebrotherdoesn't make the current situation better though :(
23:51:04kugelbluebrother: you know where to put bug reports ;)
23:51:08ranmabertrik: Yes. Increases read throughput over usb to 10MB/s for internal (vs. 4MB/s for OF) on my C200v2, works fine for me
23:51:19CIA-6r27066 build result: 25 errors, 79 warnings (ranma committed)
23:51:30AzureWurkcould somebody link me to the page that shows all the builds with notes about what was changed, i forgot to email it to myself so i could keep an eye on it at work :)
23:51:36ranma(Well 10MB/s with the zero copy patch, but without STORAGE_WANTS_ALIGN)
23:52:11 Quit kugel (Remote host closed the connection)
23:52:12bertrikranma, it always makes me feel a bit uncomfortable when things work due to mystery delays, nice to see them fixed! :)
23:52:33 Join kugel [0] (~kugel@rockbox/developer/kugel)
23:52:33bluebrotherhmm, seems deleting nvram.bin fixed it. At least it boots up now again
23:52:40funmanbertrik: we could name the functions do_something_magic() if it makes you feel better ;)
23:52:50ranmaFlaky build host again?
23:52:51kugelother targets use STORAGE_WANTS_ALIGN happily?
23:53:03bluebrotherha. Loading my (now incompatible) theme crashes again.
23:53:19funmankugel: ipod with hard disk & ata dma and ipod nano2g
23:53:44kugelI can't imagine buffering.c is on fault then
23:53:52 Quit joeyg (Quit: lions and tigers and bears, oh my!)
23:54:11ranmaMaybe I should check if STORAGE_WANTS_ALIGN works properly with only -O and without thumb...
23:54:30 Join blaaa [0] (
23:54:39blaaahi @ all =)
23:55:52blaaadon't wanna annoy anyone but could it be, that the "amarok_plus2" theme is broken with the current build + themetag changes?
23:56:08bluebrotherthat's possible.
23:56:16blaaagives me a "code abort blabla"
23:56:30blaaasame when I try to change language
23:56:33bluebrotherwhich player?
23:56:37blaaaipod video
23:56:41bluebrotherthat is not related to the theme.
23:57:00blaaabut its in the list
23:57:03bluebrotherI'm getting crashes all time when trying to change themes since I updated a couple of minutes ago
23:57:15bluebrotherit's related to _changing_ the theme, not the _theme_ itself.
23:57:36bluebrotherso it's not the theme that's broken but some loading code in Rockbox.
23:57:47blaaakinda sux :D
23:58:03*ranma thinks rockbox could use a common sd core instead of each platform handling it completely by itself...
23:58:16blaaaguess i'll have to keep using 3.6... But usb charging would be nice
23:58:36bluebrotherBagder: do we keep map files for current builds? What happened to the idea to put map files into the zips?

Previous day | Next day