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 2014-11-29

00:09:42 Quit pamaury (Ping timeout: 264 seconds)
00:17:14 Quit TheLemonMan (Remote host closed the connection)
00:29:35 Quit xorly (Ping timeout: 240 seconds)
00:29:58 Join trampel [0] (~trampel@c-24-22-235-214.hsd1.wa.comcast.net)
00:30:58 Part trampel
00:35:35 Join hillshum [0] (~hillshum@174-23-194-96.slkc.qwest.net)
00:38:51 Quit varogami (Ping timeout: 240 seconds)
00:50:06***Saving seen data "./dancer.seen"
00:50:09 Quit hillshum (Read error: Connection reset by peer)
00:53:56 Quit ender` (Quit: If God had intended us to go around naked, He would have made us that way. -- Olum's Observation)
00:54:43bp0I have a small problem with rockbox on sansa clip zip. sometimes if I pause and just let it turn itself off, then it wont start again
00:56:56bp0it happens with 3.13 and the dev version
00:57:53 Quit krabador (Quit: Take the time.)
00:58:45[Saint]gah.
00:58:47[Saint]he left.
00:59:17[Saint]I find it absolutely hilarious that people _still_ don't realize that that truecrypt page is a canary.
01:00
01:23:40 Join hillshum [0] (~hillshum@174-23-194-96.slkc.qwest.net)
01:42:22 Quit bertrik (Remote host closed the connection)
01:43:48[Franklin]foolsh: thanks!
01:48:16 Quit foolsh (Ping timeout: 258 seconds)
01:49:43 Join foolsh [0] (~rastlin@c-24-11-243-148.hsd1.in.comcast.net)
01:49:56[Franklin]foolsh: for circles, I could just draw a lot of triangles around a single point
02:00
02:01:54 Quit hillshum (Ping timeout: 264 seconds)
02:14:49 Join hillshum [0] (~hillshum@174-19-40-128.slkc.qwest.net)
02:18:03foolshthats one way to do it :)
02:18:33foolshbut masking a blit would be better
02:18:45foolsh/s/better/faster
02:23:23[Saint]both
02:25:42foolsha library for optimaized primatives is the place for such musings, of course
02:26:52[Saint]you're an optimized primitive...
02:27:57foolshan optimized primate ;)
02:32:11 Join cmhobbs [0] (~cmhobbs@ip98-186-66-92.fv.ks.cox.net)
02:32:11 Quit cmhobbs (Changing host)
02:32:11 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
02:36:20 Quit ungali (Remote host closed the connection)
02:40:34godzirraSo apparently I can't get windows to recognize my ipod classic with rockbox at all now. It used to work in the rockbox fallback image, but now it just keeps recognizing then dropping it.
02:42:41godzirrabcca111-141110 is what I'm running normally. The fallback image is r31516-120101. I'm not really sure what to do, as I can't copy anything to or from it now.
02:45:54[Franklin][Saint]: how?
02:48:27 Join Ivoah_ [0] (~Ivoah@p-74-209-25-117.dsl1.rtr.chat.fpma.frpt.net)
02:49:59foolshhmmm, sdl has something like what you want I'm sure of it
02:50:05foolsh[Franklin]: ^
02:50:10***Saving seen data "./dancer.seen"
02:51:06 Quit Ivoah (Ping timeout: 264 seconds)
02:59:07 Quit cmhobbs (Remote host closed the connection)
03:00
03:00:11foolshWhile it's SDL specific it has some cheap optimazations such as filling half a circle and mirroring it http://content.gpwiki.org/index.php/SDL:Tutorials:Drawing_and_Filling_Circles
03:00:40 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
03:00:44[Franklin]why not do 1/4 and mirror twice? :)
03:17:19 Quit cmhobbs (Read error: Connection reset by peer)
03:17:45 Join cmhobbs [0] (~cmhobbs@ip98-186-66-92.fv.ks.cox.net)
03:17:45 Quit cmhobbs (Changing host)
03:17:45 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
03:19:38 Quit cmhobbs (Read error: Connection reset by peer)
03:20:00 Join cmhobbs [0] (~cmhobbs@ip98-186-66-92.fv.ks.cox.net)
03:20:15 Quit cmhobbs (Changing host)
03:20:15 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
03:26:16 Quit cmhobbs (Read error: Connection reset by peer)
03:26:39 Join cmhobbs [0] (~cmhobbs@ip98-186-66-92.fv.ks.cox.net)
03:26:39 Quit cmhobbs (Changing host)
03:26:39 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
03:27:23godzirraAnyone have any suggestions?
03:28:18 Quit cmhobbs (Read error: Connection reset by peer)
03:28:29 Join cmhobbs [0] (~cmhobbs@ip98-186-66-92.fv.ks.cox.net)
03:28:29 Quit cmhobbs (Changing host)
03:28:29 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
03:28:32[Franklin]godzirra: I'd say try linux
03:34:50 Quit cmhobbs (Ping timeout: 258 seconds)
03:47:35[Franklin]Could this affect rockboy? http://techcrunch.com/2014/11/28/nintendo-patents-game-boy-emulation-for-use-in-mobile-devices-in-flight-entertainment/
03:48:44[Franklin]essentially, nintendo's patented "a software emulator" for GB* on a "low-capability target platform"
03:49:01 Quit ZincAlloy (Quit: Leaving.)
03:49:19[Franklin]I'd say that applies to rockbox
03:51:04[Franklin]but thankfully, it's just a patent, and not a trade secret, so we can use it to optimize rockboy :P
03:56:32 Quit AlexP (Remote host closed the connection)
04:00
04:06:05godzirra[Franklin]: I tried mac, I don't have a linux box handy right now. :/ I can try when I go to work monday I gues.
04:06:49[Saint]Its not going to work on a Mac.
04:06:59[Saint]You'd need a patched build for that.
04:07:42[Franklin](which would leave godzirra in a catch-22) ;)
04:11:19 Quit [Franklin] (Remote host closed the connection)
04:23:24 Join JdGordon_ [0] (~jonno@ppp118-209-247-102.lns20.mel8.internode.on.net)
04:24:55 Quit JdGordon (Ping timeout: 256 seconds)
04:28:05 Quit hillshum (Ping timeout: 255 seconds)
04:28:53 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:28:53 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:28:53 Quit pixelma (Disconnected by services)
04:28:53 Quit amiconn (Disconnected by services)
04:33:17godzirra[Saint]: Yeah, I guess I'll wait until monday to try at work hten. :/
04:33:22godzirraI wish I could just get it working with my windows box.
04:41:09 Join AlexP [0] (~alex@rockbox/staff/AlexP)
04:45:16saratogaall of the claims in that patent except one were rejected, and the one that was accepted is essentially useless
04:50:14***Saving seen data "./dancer.seen"
04:50:57saratogaalthough this looks better than the one referenced above (also by same author): https://www.google.com/patents/US8157654?dq=Patrick+J+nintendo&ei=KEJ5VIuyCZGTNr3VgpAB
04:51:00 Join ungali [0] (~ungali@S010600226b6da694.cg.shawcable.net)
04:51:00 Quit ungali (Changing host)
04:51:00 Join ungali [0] (~ungali@unaffiliated/ungali)
04:52:46 Quit foolsh (Ping timeout: 245 seconds)
04:54:10 Join cmhobbs [0] (~cmhobbs@ip98-186-66-92.fv.ks.cox.net)
04:54:10 Quit cmhobbs (Changing host)
04:54:10 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
04:54:37 Join hillshum [0] (~hillshum@174-19-40-128.slkc.qwest.net)
05:00
05:14:05 Nick Ivoah_ is now known as Ivoah (~Ivoah@p-74-209-25-117.dsl1.rtr.chat.fpma.frpt.net)
05:19:16 Join foolsh [0] (~rastlin@c-50-140-158-161.hsd1.in.comcast.net)
05:34:20 Quit cmhobbs (Remote host closed the connection)
05:34:22IvoahIs Rockboy likely to ever be ported to the monochrome iPods?
05:37:32 Quit TheSeven (Ping timeout: 265 seconds)
05:40:59 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
05:41:12[Saint]I doubt it.
05:42:07[Saint]It not happening within the last decade seems to be a clear indication that there's not much of a chance of it happening within the next decade.
05:42:25[Saint]Unless you were to do it yourself, of course.
05:42:37[Saint]And you appear to want this, so, you're a prime candidate.
05:42:50 Quit hillshum (Remote host closed the connection)
05:43:18 Join JdGordon [0] (~jonno@ppp118-209-219-3.lns20.mel8.internode.on.net)
05:44:49 Quit JdGordon_ (Ping timeout: 256 seconds)
05:47:41IvoahWhere would I find the source for rockboy?
05:57:06saratogaapps/plugins
05:58:27saratogasee: http://git.rockbox.org/?p=rockbox.git;a=tree;f=apps/plugins;h=b86f38e14ad8fb86d4668d995254411f499e3954;hb=HEAD and the rockboy folder
06:00
06:00:01saratogaIvoah: ^
06:02:00saratogai guess it could work ok with temporal dither or something like that since the gameboy had 2 bit color IIRC
06:03:24 Quit foolsh (Quit: WeeChat 0.4.2)
06:09:35 Quit AlexP (Remote host closed the connection)
06:26:41 Join idundidit [0] (~idundidit@66-215-119-163.dhcp.atsc.ca.charter.com)
06:50:17***Saving seen data "./dancer.seen"
07:00
07:27:02 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
07:30:37 Join JdGordon_ [0] (~jonno@ppp118-209-219-3.lns20.mel8.internode.on.net)
07:31:01 Join TBCOOL_ [0] (~tb@c-3d96e555.09-273-73746f44.cust.bredbandsbolaget.se)
07:31:07 Join Naked_ [0] (~naked@naked.iki.fi)
07:33:29 Join GodEater` [0] (~whoknows@ec2-54-201-224-30.us-west-2.compute.amazonaws.com)
07:33:29 Quit GodEater` (Changing host)
07:33:29 Join GodEater` [0] (~whoknows@rockbox/staff/GodEater)
07:34:13 Join Bluefoxicy_ [0] (~Bluefoxic@c-73-173-208-251.hsd1.md.comcast.net)
07:37:13 Quit JdGordon (*.net *.split)
07:37:13 Quit Bluefoxicy (*.net *.split)
07:37:14 Quit Jinx (*.net *.split)
07:37:14 Quit Naked (*.net *.split)
07:37:14 Quit GodEater (*.net *.split)
07:37:15 Quit TBCOOL (*.net *.split)
07:37:15 Nick Bluefoxicy_ is now known as Bluefoxicy (~Bluefoxic@c-73-173-208-251.hsd1.md.comcast.net)
08:00
08:02:01 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
08:02:57 Quit cmhobbs (Remote host closed the connection)
08:17:50 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
08:31:27 Join einhirn [0] (~Miranda@p4FC1221A.dip0.t-ipconnect.de)
08:35:59 Join Jinx [0] (Dojo@unaffiliated/jinx)
08:50:20***Saving seen data "./dancer.seen"
09:00
09:01:15 Join foolsh [0] (~rastlin@c-24-11-243-148.hsd1.in.comcast.net)
09:25:38 Quit cmhobbs (Remote host closed the connection)
09:35:22 Nick GodEater` is now known as GodEater (~whoknows@rockbox/staff/GodEater)
09:40:12 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)
09:45:39 Quit polemon__ (Ping timeout: 256 seconds)
09:48:10 Join varogami [0] (~varogami@dynamic-adsl-78-13-194-230.clienti.tiscali.it)
10:00
10:33:37 Quit burgobianco (Remote host closed the connection)
10:37:51 Join burgobianco [0] (~viskestel@li607-220.members.linode.com)
10:40:25 Join mortalis [0] (~mortalis@84.47.128.244)
10:50:22***Saving seen data "./dancer.seen"
10:51:23 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
10:59:22pamauryI'm not a big fan of reverting the file system commit
10:59:46pamaurywouldn't it be possible to split it to bisect the problem ?
11:00
11:03:33mortalisThat is what I'm going to do.
11:04:07mortalisSpliting this into small commits also could help to bisect/investigate any aother future problems in that rewrite.
11:04:55pamauryI'm afraid this won't be easy though
11:06:00pamauryfirst I suggest to split apart all the small unrelated changes like the ones in memcpy
11:06:17pamaurythen if it's possible, try to split apart the ones in plugins.c which touch prefixing the names
11:06:51pamaurymaybe I can help with the matter if you want
11:10:36mortalisSure, help is welcom, but we need somehow seoarte tasks and coordinate our work
11:13:15pamauryif I may suggest that, I think the gerrit should no be just reverting: it should be first a revert and then a series of commit (which depend on the revert) introducing the same code in several chunks
11:13:47pamauryso one task would be to take what is left to introduce and try to extract some independent out of it
11:13:55pamaurydid you begin doing this yet or not ?
11:15:14mortalisyes, but I'm only at the very begining didn't do much yet
11:15:51mortalisperhaps separate git branch would be more handy, rather then gerrit tasks?
11:17:22pamaurywhat I like about gerrit is that you can have several versions of patches, so it's a branch where you rewrite history and keep the old history, it's very handy
11:19:17pamaurymortalis: did you check that the code compiles when reverting this commit ?
11:19:27 Join ender` [0] (krneki@foo.eternallybored.org)
11:20:00mortalisYes, I'm also checked that problem with display disappeared after revert
11:20:31 Join Polochon1street [0] (~Polochon_@asa-eclille.ec-lille.fr)
11:21:43Polochon1streethello everybody!
11:22:11mortalispamaury: I suggest to use topic 'fs_rewrite' to separate our future patches on gerrit
11:22:12pamauryPolochon1street: hi, ask your question if you have one
11:22:47Polochon1streetI'm considering porting some code I wrote as a rockbox plugin, but does someone knows if it's possible to get sample arrays from songs with the rockbox API?
11:23:20pamaurymortalis: ok, so first do you mind if I update your gerrit task with the revert + reintroduce ? I think it's useful to always keep a commit containing what is left to introduce
11:23:22Polochon1streetI searched here http://git.rockbox.org/?p=rockbox.git;a=blob;f=apps/plugin.h, but it only seems to have things like « audio_play() », stuff like this
11:25:29mortalispamaury: go ahead. One note: I somehow didn't delete pathfuncs.c in my revert, but shouldn't hurt.
11:26:20 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
11:26:32pamauryPolochon1street: I'm no expert on our audio code but I'm not sure there is such an API, maybe there is an API to capture what is played to the speakers though. What do you want to achieve ? Maybe it's possible to create a new API for it
11:28:34 Join xorly [0] (~xorly@m180.dkm.cz)
11:29:40 Join AlexP [0] (~alex@rockbox/staff/AlexP)
11:30:34pamaurymortalis: any reason why you did so ?
11:31:37mortalisI heave no idea how that happened
11:37:03pamauryhum, so later commits also depend on this commit, did you fix the conflicts by hand when reverting ?
11:40:23Polochon1streetpamaury: I developped a C program that test songs and returns 1 if it's suitable for sleeping, and 0 if it's too hard
11:40:34Polochon1streetand I thought it would be fun to port it to rockbox
11:40:36mortalispamaury: Yes, I just dropped parts from latter commits that change code in fs rewrite commit
11:41:49Polochon1streetI believe the audio code uses ffmpeg?
11:41:55Polochon1street(rockbox's)
11:43:56pamauryPolochon1street: depends on the codec, some codecs are ffmpeg but they were rewritten for fixed precision and sometimes later upstream to ffmpeg
11:44:04pamaurysome are custom
11:45:49pamauryyeah someone who knows how the codecs work would be needed, because I'm not familar with this part of the code
11:46:31Polochon1streethm, I'm going to take a look at it anyway, thanks!
11:49:56 Join wodz [0] (~wodz@89-75-106-114.dynamic.chello.pl)
11:54:40wodzpamaury: so what is your idea for multiusers of hwstub?
11:57:43pamaurywodz: we could add support for requests sent to device instead of the interface (that would be optional support), because a driver cannot "claim" the device entirely. However there is still the problem with requests that span multiple transfers like read which is READ+READ2. So then we would need to build the concept of session in the protocol. Another possibility is for the hwstub library to claim the interface when needed and release it
11:57:43pamauryafterwards but that would rely on luck so I'm not too happy with it. A proxy process would work but that's quite a bit of work I think
11:58:17pamauryon the other hand, the proxy process as the advantage of privilege separation
11:58:20pamaury*has
11:59:15wodzI think proxy would be the best as it would open up possibility to work over network
11:59:32pamaurythat's very true, I didn't think about that
12:00
12:01:02pamauryone small nit though: how would one pick the port of the proxy ? using a hardcoded number doesn't feel very nice, and that would mean that qeditor and such need to be provided with the port
12:02:53wodzYou can have default one which could be overwritten on command line. Thats how gdb works, no?
12:03:34pamauryprobably
12:06:48pamauryok that works for me, I'll try to come up with a prototype
12:25:10 Quit ender` (Quit: Gray is a color; grey is a colour.)
12:31:22 Join ender` [0] (krneki@foo.eternallybored.org)
12:32:59 Quit lycide (Max SendQ exceeded)
12:33:37 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
12:34:25 Join lycide [0] (~kvieta@149.255.100.108)
12:50:12 Quit Polochon1street (Remote host closed the connection)
12:50:26***Saving seen data "./dancer.seen"
12:53:15 Join mirak [0] (~mirak@static-176-182-184-17.ncc.abo.bbox.fr)
12:53:48 Join TheLemonMan [0] (~lemonboy@unaffiliated/thelemonman)
12:58:27 Join RiDD [0] (~RiD@bl22-48-133.dsl.telepac.pt)
13:00
13:00:35 Quit xorly (Ping timeout: 256 seconds)
13:15:32 Quit AlexP (Remote host closed the connection)
13:16:55 Join AlexP [0] (~alex@rockbox/staff/AlexP)
13:22:38 Join krabador [0] (~krabador_@unaffiliated/krabador)
13:30:21 Quit wodz (Ping timeout: 240 seconds)
13:36:21 Quit mortalis (Ping timeout: 240 seconds)
13:38:24 Join mortalis [0] (~mortalis@84.47.128.244)
13:54:01 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
14:00
14:07:39mortalisI think we just have some stupid bug in rk27xx sd driver. Glitches on screen are only on images that AFAIK loaded from disk.
14:07:55mortalisand test_disk write&verify just doesn't pass
14:09:07mortalisfs rewrite probably just disclosed some race in the driver.
14:11:16mortalispamaury: I think we should stop that 'split fs rewrite into a small commits' thing. It's likely will be just a waste of our time.
14:11:31pamauryok
14:11:37mortalisIn the meantime I'll try to review sd driver.
14:11:54 Join wodz [0] (~wodz@89-75-106-114.dynamic.chello.pl)
14:15:13mortalisJust in case: Could someone please confirm that test_disk's test WRITE&VERIFY passes on the latest revision on not rk27xx targets?
14:16:31wodzmortalis: In theory I could do it today night
14:17:08mortalisThanks
14:17:10wodzconsidering lcd & sound problems the only explanation is that this patch somehow interfere with dma
14:17:44wodzalthough sound uses different dma engine then lcd
14:22:30mortaliswodz: sound problems only on mp3 files, I didn't noticed any problems with flacs
14:23:32wodzwell, flacs are much more 'disk intensive' so this doesn't fit sd driver bug theory. That is I am not saying that sd driver is bug free
14:24:40wodzmortalis: You could try changing sd data transfer through mem-to-mem dma with plain for() loop
14:25:14wodzthis part always was somewhat problematic
14:25:20mortalisI think it could be some race, that reveals itself only on somedisk workloads
14:25:24mortalisok will try
14:29:04wodzIIRC concurrent access is protected by mutex
14:41:56 Join xorly [0] (~xorly@m180.dkm.cz)
14:50:27***Saving seen data "./dancer.seen"
14:56:41mortaliswodz: much less glitches, But still few remains
15:00
15:24:29 Quit TheLemonMan (Remote host closed the connection)
15:34:40 Quit xorly (Ping timeout: 264 seconds)
15:42:49wodzmortalis: what about test_disk?
15:48:08mortaliswodz: compare error
15:51:58 Quit AlexP (Remote host closed the connection)
16:00
16:00:07wodzmortalis: I have no other 'easy' things to test
16:03:06 Quit krabador (Remote host closed the connection)
16:04:29 Join AlexP [0] (~alex@rockbox/staff/AlexP)
16:08:57 Join krabador [0] (~krabador_@unaffiliated/krabador)
16:11:21 Quit krabador (Remote host closed the connection)
16:12:36 Join JdGordon [0] (~jonno@ppp118-209-255-228.lns20.mel8.internode.on.net)
16:13:38 Quit JdGordon_ (Ping timeout: 244 seconds)
16:35:19 Join xorly [0] (~xorly@m180.dkm.cz)
16:45:59mortaliswodz: g#1057
16:46:03fs-bluebotGerrit review #1057 at http://gerrit.rockbox.org/r/1057 : rk27xx: sd: don't use DMA for data transfers by Andrew Ryabinin
16:47:10mortalistest_disk passes now
16:50:22wodzIIRC original firmware does something very similar.
16:50:28***Saving seen data "./dancer.seen"
16:51:10mortalisbtw, I wasn't able to make work WORD size transfers
16:52:30mortalisAnd why we had one buffer with word size transfers, and another with byte?
16:53:57wodzno OF uses word access for sure
16:54:13mortaliswodz: nope
16:54:18mortalisbyte access
16:54:22mortalisI've checked
16:54:59mortalisthey have word access commented out
16:55:00wodzmortalis: me too, in disasm of my OF and in SDK sources both use word access
16:55:23 Quit bertrik (Remote host closed the connection)
16:56:05wodzCan you bench this change? sd on rk27xx is slow like a hell
16:57:45wodzbut AFAIR switching to tran is the slowest part and transfer is relatively fast
16:57:58 Quit xorly (Read error: Connection reset by peer)
16:58:10kugelmortalis: nice catch!
16:59:11kugelbtw DMA is pretty useless if you busy-poll for the transfer to finish anyway
16:59:34mortaliskugel: Well that was wodz who suggested this
16:59:37kugel(from the cpu perspective)
17:00
17:00:24wodzkugel: back then I benched this to be faster then 1) simple for loop 2) inline asm taking 4 words with ldmia/stmia
17:01:47kugelthe point of DMA is to let the CPU do useful stuff during the transfer, so you could make use of sems to allow other threads to run
17:02:17wodzkugel: sure but if can do job faster even with busyloop that is still a win
17:02:30 Join xorly [0] (~xorly@m180.dkm.cz)
17:02:35kugelregarding the issue at hand, could the cache handling be incorrect?
17:03:02wodzkugel: I was simply not able to figure out how to setup this block to do peri-to-mem transfers
17:03:41wodzkugel: that is possible in theory
17:04:15mortaliswodz: bench http://pastebin.com/0zgVJmT5
17:04:46kugelor some IRQ during the transfer causing invalid cache?
17:05:50wodzmortalis: this looks more or less similar to previous version
17:07:14wodzkugel: that would be stupid as this would imply need for disabling irqs for the whole transfer time
17:07:22 Quit xorly (Read error: Connection reset by peer)
17:07:41mortaliswodz: Ok to push this?
17:07:51wodzmortalis: yes, I am ok with this
17:07:52kugelonly those that use overlappign cache lines
17:08:10wodzkugel: you have no control on this
17:08:26kugeli didnt think we have that anywhere, maybe we do since the filesystem rework?
17:09:39 Join polemon_ [0] (~polemon@g231126011.adsl.alicedsl.de)
17:10:00 Join xorly [0] (~xorly@m180.dkm.cz)
17:14:57wodzAnyway I am not convinced this is rk27xx sd bug. If this is something corrupting cacheline after commit_discard the driver is not guilty
17:16:06pamaurywodz: did you check the buffers received by the driver are cache aligned ?
17:16:32wodzno
17:16:50 Quit xorly (Read error: Connection reset by peer)
17:17:41 Join xorly [0] (~xorly@m180.dkm.cz)
17:18:02wodzHmm, actually this may be driver bug. I guess this is legitimate to pass unaligned buffer as a target of the transfer
17:20:36wodzits interesting how other drivers using DMA handle this
17:21:13pamaurywodz: the imx driver explicitly handles this
17:21:17pamaury(which is a pain btw)
17:24:14 Quit mc2739 (Ping timeout: 255 seconds)
17:30:42mortaliswodz: pamaury: kugel: aligned buffer made dma transfers work again
17:31:16pamaurythat's probably it then
17:31:26 Join clee290 [0] (4c42b8bb@gateway/web/freenode/ip.76.66.184.187)
17:31:54wodzmortalis: bench it - if the difference is small I'd leave pio copy to not step into this in the next change
17:38:35 Quit xorly (Ping timeout: 240 seconds)
17:39:57mortaliswodz: Read is faster. I'd say it worth to use dma. http://pastebin.com/RTpi0Ch5
17:45:59wodzyeah, thats what I remembered
17:46:14 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
17:46:46wodzmortalis: Ideally would be to figure out how to do peri-to-mem instead of this stupid mem-to-mem with busyloop
17:47:22 Quit pamaury (Ping timeout: 258 seconds)
17:50:46 Join [Franklin] [0] (~franklin@unaffiliated/franklin)
17:51:17[Franklin]foolsh: I think I've made sound fully configurable now
17:51:18 Join rela [0] (~x@pdpc/supporter/active/rela)
17:52:18[Franklin]foolsh: btw, what else is there to do for the menu?
17:52:25[Franklin]*manual*
17:52:34[Franklin]not menu
17:53:18 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
17:55:13 Join xorly [0] (~xorly@m180.dkm.cz)
17:58:01 Quit clee290 (Quit: Page closed)
18:00
18:09:57wodzmortalis: Could you show the latest fix?
18:14:00 Join franta [0] (~quassel@109.238.35.178)
18:15:48 Join ZincAlloy [0] (~Adium@pD9EEAF9C.dip0.t-ipconnect.de)
18:19:32mortaliswodz: updated G#1057. commit message left from first patch, ignore it
18:19:35fs-bluebotGerrit review #1057 at http://gerrit.rockbox.org/r/1057 : rk27xx: sd: don't use DMA for data transfers by Andrew Ryabinin
18:20:29kugelmortalis: ah nice. I think we have STORAGE_ALIGN_ATTR for this?
18:20:56kugelalthough I don't know if it's mandatory for all storage buffers or if it's merely a performance hint
18:24:55mortalisSTORAGE_ALIGN_ATTR under ifdef STORAGE_WANTS_ALIGN. I think it should be defined for rk27xx in that case
18:25:33wodzlooks reasonable
18:28:24 Quit franta (Remote host closed the connection)
18:28:58wodzdouble mem move for unaligned case is avoidable but this is rare case probably
18:43:21 Join franta [0] (~quassel@109.238.35.178)
18:45:21 Join mc2739 [0] (~mc2739@75.92.33.140)
18:45:25 Quit mc2739 (Changing host)
18:45:25 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
18:50:29***Saving seen data "./dancer.seen"
18:57:25 Quit wodz (Ping timeout: 244 seconds)
19:00
19:09:27 Quit franta (Remote host closed the connection)
19:18:28 Quit krnlyng (Quit: huiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii)
19:18:46 Join krnlyng [0] (~liar@83.175.90.24)
19:32:58 Quit [Franklin] (Ping timeout: 272 seconds)
19:48:07 Quit cmhobbs (Ping timeout: 264 seconds)
19:48:07 Quit uwe_ (Ping timeout: 264 seconds)
19:48:49 Join uwe_ [0] (~uwe_@ipservice-092-211-199-052.092.211.pools.vodafone-ip.de)
19:49:19 Quit ungali (Ping timeout: 256 seconds)
19:50:56 Join [Franklin] [0] (~franklin@cpe-071-071-039-006.triad.res.rr.com)
19:54:06 Join ungali [0] (~ungali@S010600226b6da694.cg.shawcable.net)
19:54:06 Quit ungali (Changing host)
19:54:06 Join ungali [0] (~ungali@unaffiliated/ungali)
19:55:16 Join megal0ma1iac [0] (~megal0man@105.229.160.73)
19:55:22 Nick megal0maniac is now known as Guest53942 (~megal0man@unaffiliated/megal0maniac)
19:55:22 Quit Guest53942 (Killed (card.freenode.net (Nickname regained by services)))
19:55:22 Nick megal0ma1iac is now known as megal0maniac (~megal0man@105.229.160.73)
20:00
20:01:50fs-bluebotBuild Server message: New build round started. Revision 8618f2c, 255 builds, 34 clients.
20:06:32 Quit n17ikh (Ping timeout: 250 seconds)
20:07:59 Join n17ikh [0] (~n17ikh@unaffiliated/n17ikh)
20:14:52 Quit n17ikh (Remote host closed the connection)
20:18:24 Quit mortalis (Quit: KVIrc 4.3.1 Aria http://www.kvirc.net/)
20:18:28 Join pmarty [0] (1f0bb455@gateway/web/freenode/ip.31.11.180.85)
20:25:34 Join wodz [0] (~wodz@89-75-106-114.dynamic.chello.pl)
20:25:53 Join n17ikh [0] (~n17ikh@unaffiliated/n17ikh)
20:50:30***Saving seen data "./dancer.seen"
20:58:41pmartySomeone just gave me his "does not work" 2007 iPod Classic for servicing. I rebooted it into Diagnostics Mode and checked SMART data. 768 reallocs + 304 pending sectors. The disk is pretty much dead then?
20:59:11wodzlooks like
21:00
21:02:07pmartyI guess it's not possible to verify with smartmontools without tearing down?
21:02:44pmarty(over USB)
21:05:45wodzpmarty: you should catch TheSeven and ask what are the options.
21:06:22TheSevenlatest emcore development builds can do it
21:06:42TheSevenbut the diagmode data clearly shows that the drive is bad
21:07:51saratogahas anyone tested the BBE patch recently? I'm going to give it a go and then commit
21:08:03saratogaprobably needs a lot of optimization but that can wait
21:08:59copperhow does one access the diagnostics mode on a Classic?
21:09:21TheSevencopper: only without emcore installed
21:09:30pmartycopper: same as before, unless you installed rockbox :D
21:09:31coppermeh
21:09:55TheSevencopper: if you're using an emcore dev build that has emcore disk mode already, you can use that + smartmontools to have a look at the smart data
21:11:02coppernope
21:11:53 Join Strife89 [0] (~Strife89@adsl-98-80-236-36.mcn.bellsouth.net)
21:12:28pmartyare they still producing this 1.8" thingy drives? and was there any difference in drive interface between 5th and 6th ipod generation? (i did replace drive for ipod video some years ago)
21:13:09pmartyit had this mini-PATA (called ZIF) interface
21:13:17saratoganewer virtualbox has surprisingly buggy USB, went from working fine to crashing the whole VM half the time
21:21:33TheSevenpmarty: people typically replace HDDs with SD card mods these days
21:21:39TheSevenor even mSATA SSD mods
21:21:45kugelsaratoga: I looked at it briefly
21:21:52TheSevenbut I guess the HDD interface is the same as on the ipod video
21:22:01saratogathe DX port?
21:22:11saratogai'd like to get it out of the android tree if possible
21:22:11kugeli can't comment on the algorithms, but the rest and the code quality was fine
21:22:16saratogaoh DSP
21:22:20kugelyea
21:22:52saratogayeah i think it looks good, although 26MHz hit on AMSv2 for PBE seems a bit steep, probably needs some optimization
21:23:22kugelI don't even know what AFR does
21:23:40saratogafatigue reduction, i forget the algorithm
21:23:45saratogabut probably involves crossfeed
21:24:01kugelbut the manual explains it i guess
21:25:11kugeldoes any of the patches erquire lots of memory?
21:25:34saratogathats a good question
21:25:35saratogai should check
21:25:40kugels/patches/effects/
21:28:23saratogatheres some static buffers totally maybe 2-3 thousand samples
21:28:30saratogaso probably 10KB
21:28:39saratogathats a lot on low memory targets
21:36:56 Join cmhobbs [0] (~cmhobbs@ip98-186-66-92.fv.ks.cox.net)
21:36:56 Quit cmhobbs (Changing host)
21:36:56 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
21:37:18saratogaif i applied a patch from gerrit, what do i need to do to be able to update that patch on gerrit?
21:38:51wodzpush it again with *the same* id line in commit message
21:40:57saratogaCan I just copy and paste the header from the patch?
21:42:02wodzyes
21:42:46wodzwhat I do is commit my work on top of patch from gerrit and then rebase -i and merge this with original patch
21:43:41saratogasince i just used a patch, i don't have any of the git history
21:43:57wodzerr, what?
21:43:57saratogai just did a commit of the whole patch and some minor changes using the old patch header and commit idea
21:43:59saratogaid
21:44:09pmartyTheSeven: I remember people modding with CF cards + CF-ZIF adapter. Are SD mods more popular now?
21:44:23saratogapatch -p1 < file.patch
21:44:29wodzhow did you get the patch in the first place then?
21:44:50saratogadownloaded from gerrit
21:45:06TheSevenpmarty: CF cards are rare these days, people are putting CF-SD adapters inside these CF-ZIF adapters
21:45:20 Quit cmhobbs (Remote host closed the connection)
21:46:07wodzsaratoga: ah. That will not work. As soon as you commit locally new id line will be created (your version will have both - id from gerrit and id from pre-commit hook)
21:46:29pmartyTheSeven: That means there is no direct SD-ZIF available?
21:46:41saratogaCF cards are pretty cheap on newegg
21:47:35TheSevenpmarty: I'm not aware of any
21:48:13saratogahere are the manual changes I wanted to push to gerrit: http://pastebin.com/Fvj8NPQD
21:48:18saratogai have no idea what to do with them
21:48:43pmartysaratoga: newegg does not ship internationally last time I checked
21:49:37wodzsaratoga: Do you have git repo setup to work with gerrit? http://www.rockbox.org/wiki/UsingGit has detailed instructions
21:53:09 Join natanelho [0] (4d7d68ce@gateway/web/freenode/ip.77.125.104.206)
21:55:12natanelhohi everybody!
21:55:35natanelhohow to turn *.lua into *.rock?
21:56:09wodzerr, what?
21:57:06natanelhoi know i can run lua scripts directly by selecting it in files menu, but how to compile them into .rock?
21:57:16wodzyou can't
21:58:16natanelhook
22:00
22:00:09natanelhowhen i use the "do_menu" function in lua, a "settings icon" apears near the header of it. can i change it to other icon?
22:02:30natanelho?
22:03:25wodzDon't know. Never coded anything in lua for rockbox
22:05:59natanelhoi think this function exists in c of rockbox to. the full name is "rb.do_menu()"
22:07:26natanelhois rb based on unix?
22:07:59wodzno
22:08:32 Quit fs-bluebot (Ping timeout: 256 seconds)
22:08:41wodzrb is custom rt kernel with cooperative threads. We even not pretend to be POSIX compatibile
22:09:05[Franklin]wodz: but we are, to some extent
22:09:10[Franklin]but barely
22:09:37wodzwe are *sometimes* POSIX compatible yes
22:09:52[Franklin]file I/O looks unix-like, etc.
22:09:55 Quit bluebrother (Ping timeout: 240 seconds)
22:09:56[Franklin]but it's really not
22:10:15[Franklin]put it this way: rockbox is *inspired* by unix :)
22:10:26wodzanyway we are far from based on unix
22:11:14 Join fs-bluebot [0] (~fs-bluebo@f053155018.adsl.alicedsl.de)
22:11:35natanelhoah i undarstand...
22:11:43natanelhothanks anyway!
22:11:47 Quit natanelho (Quit: Page closed)
22:12:07 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
22:31:52 Quit varogami (Ping timeout: 240 seconds)
22:33:09wodzha! It seems I am able to chat with sd card on ATJ :-)
22:34:12 Quit pmarty (Quit: Page closed)
22:46:45 Quit mc2739 (Ping timeout: 258 seconds)
22:48:50 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
22:50:34***Saving seen data "./dancer.seen"
22:51:47foolsh[Franklin]: ping
22:51:55[Franklin]pong
22:52:16[Franklin]\o foolsh
22:52:36foolshLong night for me, migrated a server, still not finished
22:52:55foolshthe manual only needs the buttons finish I believe
22:53:17[Franklin]ah ok
22:53:22*[Franklin] might just add some screenshots
22:53:23foolshI stopped there since grayscale *was* on the was, but now
22:53:38[Franklin]yeah :)
22:53:44foolshWhat is the state of xworld?
22:53:51[Franklin]pretty close to completion :)
22:54:13[Franklin]it could get committed right now :)
22:56:00[Franklin]but it needs some optimization and testing
22:56:14[Franklin]oh, and vertical stride compatibility
22:56:34[Franklin]but there's only one color device with a vertical stride
22:56:42foolshyeah
22:59:15 Quit Strife89 (Ping timeout: 252 seconds)
23:00
23:10:26[Franklin]foolsh: I guess that with vertical stride, it could be committed
23:13:15Ivoah[Franklin]: Why do you have brackets around your name?
23:16:57[Franklin][Why] [Not]?
23:25:44 Join franta [0] (~quassel@109.238.35.178)
23:37:40 Quit wodz (Quit: Leaving)
23:55:24 Nick suYin`OFF is now known as suYin (suYin@server1.bouncer4you.de)

Previous day | Next day