#rockbox 2014-10-08

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 2014-10-08

00:01:21		Nick DormantBrain is now known as SuperBrainAK (~andy@2001:470:8:a61::5f92:59a1)
00:05:49	[Franklin]	weird... 2048 is only drawing the left edge scaled
00:05:52	[Franklin]	left 4 tiles
00:06:00	[Franklin]	everything else is flickering
00:06:40		Quit io53 (Ping timeout: 260 seconds)
00:08:50		Quit pit (Quit: Page closed)
00:11:39		Quit lebellium (Quit: ChatZilla 0.9.91 [Firefox 33.0/20141002185629])
00:16:39	foolsh	It's a mess as far as white space, it could be PLA-ified but I don't think that will help game play
00:17:18	[Franklin]	and there are some erroneous yield and sleep calls in there, but really it looks good to me
00:17:32	[Franklin]	but PLA-fying wouldn't be that easy
00:17:43	[Franklin]	most targets don't have enough keys
00:17:48	foolsh	It would hurt game play IMO
00:17:53	[Franklin]	PLA doesn't have enough keys
00:18:07	[Franklin]	Up/down/left/right/select/quit/cancel
00:18:11	[Franklin]	not that many
00:18:22	[Franklin]	so yeah, I think keymaps are the way to go with this one
00:18:25	foolsh	up down left right fire and menu are all thats really needed for xrick
00:18:37	[Franklin]	dynamite?
00:18:40	[Franklin]	stick?
00:18:58	foolsh	Oh I couldn't play that far in on my fuze+
00:19:05	foolsh	didn't know
00:19:26		Join io53 [0] (~berg@k24.ip4.netikka.fi)
00:21:03	[Franklin]	my hold button's broken on my ipod
00:21:07	foolsh	but still yeah keymaps
00:21:19	[Franklin]	the GPIO flips, but nothing happens
00:27:45		Join chrisb [0] (~chrisb@pool-71-175-243-4.phlapa.east.verizon.net)
00:37:37		Quit bertrik (Remote host closed the connection)
00:40:13		Quit ender` (Quit: Sometimes I lie awake at night, and I ask, "Where have I gone wrong?" Then a voice says to me, "This is going to take more than one night.")
00:45:18		Quit xorly (Ping timeout: 260 seconds)
00:48:56		Quit chrisb (Remote host closed the connection)
00:53:42		Quit io53 (Ping timeout: 260 seconds)
00:55:52	[Franklin]	ahh
00:55:57	*	[Franklin] sees the bug
00:56:10	[Franklin]	width instead of height
01:00
01:08:58	*	[Franklin] wonders if it's worth it to provide pre-scaled bitmaps
01:09:08	[Franklin]	probably not... it'd double the binary size
01:09:33	[Franklin]	perhaps I could provide a large bitmap and scale it down, though
01:13:30	[Franklin]	though binary size really isn't an issue
01:17:43	[Franklin]	it's already the 3rd largest plugin
01:17:43	[Franklin]	in terms of bin size
01:18:00	[Franklin]	only behind blackjack and solitaire
01:18:06	[Franklin]	but not by much
01:19:25		Join krabador [0] (~krabador@unaffiliated/krabador)
01:21:21	[Franklin]	so more bitmaps wouldn't be wise
01:23:16	[Franklin]	or maybe some kind of RLE for bitmaps in the future?
01:24:29	[Franklin]	nothing fancy
01:29:17		Quit foolsh (Remote host closed the connection)
01:35:38	***	Saving seen data "./dancer.seen"
01:40:56	*	[Franklin] sees a very crude method
01:48:10	[Franklin]	but it could work
01:52:05	[Franklin]	just use an escape char
01:54:59	[Franklin]	and a call to bitmap_rle_decode or something like that to generate the output
01:55:26	[Franklin]	but that leads to an interesting point: does allocating an empty array affect the bin size?
01:55:53	[Franklin]	say I have char array[10000] and no code that touches it. will that affect the bin size?
01:56:13	[Franklin]	(statically allocated)
02:00
02:08:54	*	[Franklin] wants to do it without an escape char, actually
02:26:15		Join io53 [0] (~berg@k24.ip4.netikka.fi)
02:36:27		Quit dys (Ping timeout: 260 seconds)
02:46:55		Quit Aldem (Quit: Leaving)
02:48:20		Quit [Franklin] (Ping timeout: 258 seconds)
03:00
03:00:00		Quit AlexP (Remote host closed the connection)
03:16:56		Join ikeboy [0] (~ikeboy@ool-435622d3.dyn.optonline.net)
03:35:42	***	Saving seen data "./dancer.seen"
03:39:50		Quit ikeboy (Quit: Leaving)
03:53:53		Join ikeboy [0] (~ikeboy@ool-435622d3.dyn.optonline.net)
04:00
04:12:45		Part fling
04:24:22		Join Adm [0] (~chatzilla@unnum-78-27-173-27.domashka.kiev.ua)
04:25:05		Nick Adm is now known as djzloy (~chatzilla@unnum-78-27-173-27.domashka.kiev.ua)
04:26:05	djzloy	hello! How i can recover clip+ with 32 mb RAW disc?
04:27:00	[Saint]	Check our wiki for AMS Unbrick
04:27:10	djzloy	was readed and tryed all from AMS brick
04:27:22	[Saint]	But...its almost certainly dead.
04:27:23	djzloy	have linux and win
04:27:28	[Saint]	Its cheap. Move on.
04:27:43	djzloy	no ways more ?
04:27:47		Join ungali [0] (~ungali@S010600226b6da694.cg.shawcable.net)
04:27:47		Quit ungali (Changing host)
04:27:47		Join ungali [0] (~ungali@unaffiliated/ungali)
04:27:49	[Saint]	No.
04:28:35	djzloy	maybe can unsolder something?
04:28:42	[Saint]	No.
04:28:53	djzloy	sad
04:29:05	djzloy	was with rockbox installed
04:29:20	[Saint]	Thats nice.
04:29:32	djzloy	bootloader maybe from dual boot?
04:29:38	djzloy	can this help?
04:29:46	[Saint]	No.
04:30:21	djzloy	and no ways to reflash all mem ?
04:30:34	[Saint]	No.
04:30:53	djzloy	sad
04:31:01	djzloy	thx
04:31:17	[Saint]	Asking the same question multiple times wont change the answer I'm afraid.
04:31:35	djzloy	yes i know
04:31:50	djzloy	thx bro
04:32:39	[Saint]	All we know about recovering these devices is listed in the wiki page you read.
04:32:57	[Saint]	If that didn't work, its dead.
04:33:13	djzloy	got it
04:34:49	djzloy	sad( seems its moment dead after files copping
04:43:05		Join jhMikeS [0] (~jethead71@c-68-43-2-35.hsd1.mi.comcast.net)
04:43:05		Quit jhMikeS (Changing host)
04:43:05		Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
04:44:41		Quit krabador (Remote host closed the connection)
04:48:09		Join Guinness` [0] (Slayer@c-69-143-187-144.hsd1.va.comcast.net)
04:50:42		Quit Guinness (Ping timeout: 244 seconds)
04:50:43		Nick Guinness` is now known as Guinness (Slayer@c-69-143-187-144.hsd1.va.comcast.net)
04:56:28		Quit amiconn (Disconnected by services)
04:56:28		Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:56:29		Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:56:29		Quit pixelma (Disconnected by services)
04:56:31		Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:56:31		Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
05:00
05:28:19		Quit TheSeven (Ping timeout: 272 seconds)
05:28:55		Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
05:32:48		Quit ungali (Remote host closed the connection)
05:35:44	***	Saving seen data "./dancer.seen"
05:55:43		Quit the-kyle (Remote host closed the connection)
05:59:28		Join the-kyle [0] (~kyle@kyle.tk)
06:00
06:19:53		Quit ikeboy (Quit: Leaving)
06:40:16		Quit shamus (Read error: Connection reset by peer)
06:40:40		Quit Catelite (Read error: Connection reset by peer)
06:41:03		Join shamus [0] (~shmaus@ip-206-192-193-180.marylandheights.ip.cablemo.net)
06:41:53		Join Catelite [0] (~Deluge@Powder/Staff/Catelite)
07:00
07:01:50		Join dfkt_ [0] (~dfkt@unaffiliated/dfkt)
07:01:59		Quit dfkt (Disconnected by services)
07:02:14		Nick dfkt_ is now known as dfkt (~dfkt@unaffiliated/dfkt)
07:12:04		Join pamaury [0] (~quassel@rockbox/developer/pamaury)
07:26:20		Join mortalis [0] (~kvirc@212.44.150.238)
07:30:57		Join Scromple_ [0] (~Simon@27.127.199.230)
07:33:16		Quit Scr0mple (Ping timeout: 260 seconds)
07:35:45	***	Saving seen data "./dancer.seen"
07:43:14		Join kugel_ [0] (~kugel@rockbox/developer/kugel)
07:46:04		Quit kugel_ (Remote host closed the connection)
07:46:12		Join kugel_ [0] (~kugel@rockbox/developer/kugel)
07:58:01		Quit pamaury (Ping timeout: 272 seconds)
08:00
08:09:01		Quit Provel (Read error: Connection reset by peer)
08:11:30		Join Provel [0] (~Provel@75-132-30-64.dhcp.stls.mo.charter.com)
08:14:25		Join xorly [0] (~xorly@m180.dkm.cz)
08:20:54		Join wodz [0] (~wodz@iwl138.internetdsl.tpnet.pl)
08:23:42		Join ender` [0] (krneki@foo.eternallybored.org)
08:27:43	yuriks	how can I get the logging output on a Clip+ if I enabled "log to serial port"? Does it require actually hooking up to an internal serial port or does it log over USB or something?
08:28:25		Quit xorly (Ping timeout: 272 seconds)
08:32:13	wodz	maybe funman or bertrik will remember
08:36:27	yuriks	looks like it's not even building in debug mode, eh
08:38:59		Quit amiconn (Remote host closed the connection)
08:38:59		Quit pixelma (Remote host closed the connection)
08:40:40		Join amiconn [0] (amiconn@rockbox/developer/amiconn)
08:40:41		Join pixelma [0] (pixelma@rockbox/staff/pixelma)
08:44:44		Quit [Saint] (Ping timeout: 246 seconds)
08:44:55	wodz	pamaury: (log) I tried qeditor diff patch using checkout from gerrit to pull all dependencies. 1) in lib there is missing '}' at the end. 2) UI wise this needs substantial work. The tree hints you where are the changes but when you click on register it is really hard to find the difference
08:45:46	wodz	pamaury: (log) Also it is not clear at the first sight which dump coresponts to which values in diff view
08:46:41	wodz	anyway that is great progress from the usability point of view
08:47:17	yuriks	it doesn't look like anyone ever builds with DEBUG enabled
08:47:27	yuriks	there are lots of bitrotten logging statements that don't compile anymoer'
08:47:34	yuriks	anymore*
08:51:14		Join pamaury [0] (~quassel@rockbox/developer/pamaury)
08:57:26		Quit dfkt (Remote host closed the connection)
08:58:43		Join dfkt [0] (dfkt@unaffiliated/dfkt)
09:00
09:05:10		Join [Saint] [0] (~saint@rockbox/staff/saint)
09:07:32	wodz	pamaury: see logs
09:13:47		Quit TheSeven (Ping timeout: 272 seconds)
09:14:36		Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
09:16:39		Join pamaury_ [0] (~quassel@sphinx.lix.polytechnique.fr)
09:19:41		Join dys [0] (~user@2a01:1e8:e100:8296:21a:4dff:fe4e:273a)
09:27:52		Join petur [0] (~petur@rockbox/developer/petur)
09:30:52		Join AlexP [0] (~alex@rockbox/staff/AlexP)
09:35:46	***	Saving seen data "./dancer.seen"
09:39:19	wodz	[Saint]: ping
09:56:55		Quit mg_ (Ping timeout: 260 seconds)
09:59:03		Join bertrik [0] (~quassel@rockbox/developer/bertrik)
10:00
10:14:26		Join LinusN [0] (~linus@giant.haxx.se)
10:23:41		Quit pamaury (Ping timeout: 272 seconds)
10:24:05		Join pamaury [0] (~quassel@rockbox/developer/pamaury)
10:31:04	[Saint]	wodz: yo
10:32:17	wodz	[Saint]: added comment on gerrit - basically there might be difference between earlier version and this with timeouts. I posted small change which should fix this.
10:32:23	wodz	I mean in comment
10:33:55	[Saint]	I can recheck some hours from now, I'll cherrypick and start the build process now but I only have my 5.5G on me presently.
10:34:31	[Saint]	Also, thanks.
10:34:56	wodz	[Saint]: basically you need to manually tweak one line in your yesterday work
10:36:21		Quit megal0maniac (Ping timeout: 272 seconds)
10:36:42		Join megal0maniac [0] (~megal0man@unaffiliated/megal0maniac)
10:39:06		Nick SuperBrainAK is now known as DormantBrain (~andy@2001:470:8:a61::5f92:59a1)
10:39:26	[Saint]	Right. Yeah. I just saw the trivial difference.
10:42:41		Quit megal0maniac (Ping timeout: 272 seconds)
10:43:51	yuriks	been reading the USB code in RB. Not that complicated
10:44:10		Join einhirn [0] (~Miranda@bsod.rz.tu-clausthal.de)
10:44:19	yuriks	I guess I'll try to implement something dumb like a keyboard or mouse to get familiar with it
10:44:27	wodz	basic usb stuff is not that hard as long as you don't read the spec :P
10:44:51	wodz	yuriks: I guess you are aware that rb has HID support
10:45:17	yuriks	wodz: yeah, though I haven't quite figured out what it's used for (was going to look into it now. I judget the MSC code more interesting)
10:45:43	[Saint]	We can do mouse and multimedia buttons.
10:45:55	wodz	yuriks: you can control PC from rb device
10:45:55	[Saint]	Um...two other things i forget.
10:46:09	yuriks	wodz: ah, so a keyboard :P
10:46:23	wodz	yuriks: sort of, yes
10:46:28	yuriks	I guess I'll try enabling that and playing around with it
10:46:42	wodz	not all devices support HID though
10:46:44	[Saint]	You could made a morse HID keyboard
10:46:46	[Saint]	:p
10:46:57	wodz	yeah, truly hardcore
10:47:31	yuriks	wodz: why not? Do some USB controllers implement the class directly in hardware?
10:48:10	wodz	yuriks: no, interrupt endpoints are needed for this
10:48:25	wodz	yuriks: we can't make it work on some controllers
10:48:27	yuriks	ah, and some can only do Control and Bulk
10:48:40	gevaerts	I'm fairly sure the issue is actually bugs in our drivers
10:49:09	wodz	thats why I said 'we can't make it work'
10:49:15	yuriks	from what I understand the real distinction between the transfer types is actually on the host and how it schedules bandwidth
10:49:23	gevaerts	bulk and interrupt are exactly the same on the wire, the only difference is in the descriptors
10:49:33	gevaerts	Between bulk and interrupt, yes
10:49:42	gevaerts	Control and isochronous are actually different
10:54:00	yuriks	seems like the HID driver supports dumping data via typing
10:54:02	yuriks	neat.
10:54:14		Join megal0maniac [0] (~megal0man@unaffiliated/megal0maniac)
10:54:30	yuriks	er, other way around, nevermind :P
10:54:36	yuriks	was thinking of a piezo-style hack there
11:00
11:04:42	yuriks	seems like all the easy standard classes are implemented already (serial too), so maybe I'll have to just jump right into isochronous audio? :P
11:18:17		Quit pamaury (Remote host closed the connection)
11:18:47		Quit Cinos (Ping timeout: 272 seconds)
11:19:16		Join maruk1 [0] (~papier@titanium.v6.sdv.fr)
11:19:21		Join xorly [0] (~xorly@wced-94-218-32-147.feld.cvut.cz)
11:20:30	pamaury_	wodz: just read the logs, yeah at the moment register view doesn't highlight the differences, it will implemented soon. How do you suggest to differentiate between the two values: I assumed it would be clear the first column is the first file, but I can put "value A" and "value B" and "file A"/"file B" if you think it helps
11:20:44	pamaury_	strange I didn't see the missing } in the lib, thanks
11:21:06		Nick pamaury_ is now known as pamaury (~quassel@sphinx.lix.polytechnique.fr)
11:22:00	wodz	pamaury: maybe the file chooser should be on the left/right instead of one below the other
11:22:33	yuriks	so what actually changes driver-side for isochronous? I'm taking a look and it doesn't seem like any of the code specifically checks for any kind of transfer type
11:22:44	wodz	pamaury: in the reg values view it would be nice to have bold line separating the two
11:22:54	pamaury	ok
11:23:20	wodz	yuriks: pamaury said this before - the main difference is that you simply can't have interrupt for each microframe
11:23:21	pamaury	maybe we can have two colorsets: one for A and one for B (like light red, light green) ?
11:23:51		Quit kugel_ (Ping timeout: 272 seconds)
11:23:58		Join Cinos [0] (Cinos@cinos.biz)
11:24:00	yuriks	wodz: yeah, but the actual handling doesn't change? You can't have all those interrupts simply for performance reasons
11:24:01	pamaury	yuriks: yeah that's the main difference, you need to somehow queue several transfers
11:24:15	pamaury	but the handling is usually the same
11:24:16	wodz	pamaury: I'd leave color to emphasis the difference in reg
11:24:24	pamaury	ok
11:25:07		Quit Scall (Ping timeout: 272 seconds)
11:25:11	yuriks	and since the hardware fires the interrupt we need either: a) hw support for queuing, b) a fast path for enqueing/dequeing isochronous transfers
11:25:12	yuriks	correct?
11:25:14	pamaury	yuriks: beware though, the RB trunk arc driver assumes one transfer per interrupt, whereas the code in my branch handles several transfer / interrupt
11:25:38	pamaury	yuriks: basically yeah, either hardware queue (like on arc), or maybe FIQ for controllers without it
11:26:22	wodz	pamaury: that assumes you can do fiq. rk27xx can't
11:27:05	pamaury	yeah, irq might be do the trick if ran sufficiently fast
11:27:13	pamaury	*might do
11:27:41	wodz	the only difference between fiq and irq is register window
11:27:54	wodz	and possibly the separate stack
11:28:07	pamaury	yeah and priority too, you really want nested interrupts in this case
11:28:13	yuriks	a microframe is 125us, that's 8kHz, it doesn't sound too bad
11:29:09	wodz	once needs to check to judge
11:29:40	wodz	but I'd suspect 8kHz interrupts to f*** up tick timer and stuff
11:30:02	wodz	ah and dma & i2s transfers
11:30:12	yuriks	oh, there's that, hmm
11:30:29	yuriks	can't be blocked by DMA
11:31:19	wodz	well dma runs asynchronously usually BUT it needs to fire irq from time to time and some hardware have really short i2s fifo
11:32:16	pamaury	that's still a lot: each interrupt needs to push many registers and do some stuff, you quite easily execute ~250 instructions in the interrupt handle even in trivial cases, and if you have more complicated handles which touch kernel dara structures (semaphore for example), I suspect you easily do 1000 instructions/interrupt, that's already 8KHz*1000=8Mips, so one needs to be careful but maybe it can be done
11:32:54	pamaury	anyway, the ARC controller has hardware queues so it solves the problem for this controller at least
11:33:06	pamaury	I'm not even sure rk27xx controller can do isochronous transfers
11:33:25	wodz	it can
11:33:53	wodz	or maybe I mix this up with int
11:34:24		Join Rower [0] (~husvagn@h176n2-aeg-a11.ias.bredband.telia.com)
11:35:01	wodz	yeah no isochronous eps
11:35:06	wodz	so rk27xx is out
11:35:26		Join Scall [0] (~chat@unaffiliated/scall)
11:35:47	***	Saving seen data "./dancer.seen"
11:36:45	yuriks	I noticed that the as3525 seems to use the s3c6400x driver. What gives?
11:37:19	gevaerts	Naming
11:37:47	gevaerts	AFAIK it's called "s3c6400x" because that's the SoC with the datasheet that had a proper description of the controller
11:37:47	yuriks	hm?
11:37:52	yuriks	oh
11:38:01	yuriks	so it's a generic controller used by many ARM SoCs
11:38:05	gevaerts	Yes
11:38:07	gevaerts	They all are
11:38:21	yuriks	that's mighty convenient
11:38:32	gevaerts	There are a few USB controller cores around, and SoC vendors pick one of those
11:38:50	gevaerts	The same goes for most other bits on the SoC, usually
11:39:59	gevaerts	It tends to mean that (a) there's a reasonable chance you already have the driver, and (b) if you don't have the datasheet (or if the datasheet you have is vague) with a bit of luck you can find another datasheet that matches that bit of hardware
11:40:20	gevaerts	Of course, that other datasheet might describe a different revision of the hardware
11:40:29		Quit wodz (Read error: No route to host)
11:41:56	gevaerts	And there will be some specific bits of course, such as where in memory it's mapped, which GPIO (or even which mechanism) detects the cable connection, how you set the power, ...
11:44:00	yuriks	right, there's a as3525 specific file with a few bits
11:44:44	*	yuriks will sleep on all this
11:45:06	yuriks	if it comes to implementing queuing on hardware I'm not sure if I could do it
11:45:20	yuriks	though since it's only a few controllers the work might be worth it
11:46:14	gevaerts	You don't have to do it all
11:46:44	gevaerts	I mean, if you have a clean implementation on one controller, other people can do others
11:47:05	gevaerts	And even then, there are some people around who can help with this
11:47:41	gevaerts	Hmmm
11:48:25	gevaerts	There's some stuff on gerrit about some of the controllers (things like an entirely new driver implementation...), and s3c6400x rings a bell there
11:49:04	gevaerts	It probably makes sense to look into that a bit first before you satrt implementing isochronous in a driver that might soon be replaced anyway...
11:49:29	gevaerts	TheSeven: that's your work I believe. Do you know which SoCs are affected?
11:49:34		Join wodz [0] (~wodz@iwl138.internetdsl.tpnet.pl)
11:49:38	yuriks	gevaerts: ah, thanks for the heads up
11:50:02	yuriks	hm, if I'm looking at the right manual, the s3c6400x seems to have generous 4kb fifos for each endpoint
11:54:37		Join mg_ [0] (martigam@cassarossa.samfundet.no)
11:58:27	yuriks	just did the math, at 125us per microframe, stereo 16-bit 48kHz audio, you git 6 samples per USB packet :P
11:58:29	yuriks	how wasteful
11:58:34	yuriks	you fit*
11:58:41	yuriks	micropacket*
11:59:58		Join foolsh [0] (~bbrown@c-67-174-138-234.hsd1.in.comcast.net)
12:00
12:02:04	gevaerts	You don't have to use every microframe
12:02:15	gevaerts	You can set things up to use every 32nd or so
12:02:27		Quit jhMikeS (Ping timeout: 260 seconds)
12:03:12	gevaerts	That increases latency of course, but if people really need extra-low latency they should just buy a proper DAC and not try to get away with a cheap clip :)
12:03:46	gevaerts	Oh, bonus points if you implement MIDI and pass that through to the MIDI player plugin :)
12:04:01	yuriks	gevaerts: the part of the spec I read says that the variance between packets can only be of max. +-1.5 sample
12:04:40	yuriks	though maybe I misinterpreted it and it's talking about packets with data in them only
12:06:32	yuriks	this spec sucks :/
12:06:52	yuriks	it spends a lot of time talking about Chorus and Reverberation processors and too little about the actual core competency of the protocol
12:16:31	gevaerts	yuriks: do you have a USB audio device?
12:17:27	gevaerts	If not, I'd recommend getting one. Looking at how a real device behaves can help a lot in understanding the spec
12:17:38	yuriks	gevaerts: no, if I had one I would capture packets to check this
12:17:41	gevaerts	I think I once got one for $3 on dealextreme
12:17:51	yuriks	a very good idea :)
12:18:03	yuriks	I just found that you can set a data rate for isochronous devices
12:18:10	yuriks	I thought they had a fixed packet rate
12:18:38	foolsh	theres also the usb gadget sources in the linux kernel to look at too
12:20:36	yuriks	so if you choose the largest available interval you get 2^15*125us = 4.096s
12:20:54	foolsh	latency ?
12:21:20	yuriks	exactly. Latency is good to avoid interrupt starving things :)
12:21:36	yuriks	there's no point is having a smaller interval if you're just queuing the packets anyway
12:21:41	yuriks	point in*
12:26:15		Quit copper (Remote host closed the connection)
12:27:19		Join copper [0] (~copper@unaffiliated/copper)
12:27:31	yuriks	for 16-bit 48kHz 2ch, fitting the largest amount of data per microframe you can (and thus getting the smallest packet rate) you want an interval of 2^7*125us = 16ms
12:27:50	yuriks	I think this can be easily handled manually by the CPU without queuing
12:28:50	yuriks	(it's ~62Hz)
12:29:52	yuriks	as a baseline. I think 4ms would be a good latency to shoot for
12:36:36		Quit xorly (Ping timeout: 260 seconds)
12:43:03	mg_	yuriks: talking about usb audio now?
12:51:15		Quit Catelite (Ping timeout: 250 seconds)
12:51:50		Quit djzloy (Remote host closed the connection)
12:52:31	[Saint]	wodz: I wish I had good news.
12:52:33	[Saint]	http://pastebin.com/jf3kwscm
12:53:24	[Saint]	that's the exact same behavior from last night.
12:54:18		Nick megal0maniac is now known as Guest67744 (~megal0man@unaffiliated/megal0maniac)
12:54:18		Quit Guest67744 (Killed (sendak.freenode.net (Nickname regained by services)))
12:54:21		Join megal0maniac [0] (~megal0man@105.184.58.28)
12:55:35	wodz	[Saint]: now change #define WHILE_WITH_TMO(cond, us) to this: #define WHILE_WITH_TMO(cond, us) while(cond)
12:56:13	wodz	[Saint]: this should basically revert my timeout change
12:58:05	wodz	[Saint]: and retest of course
12:58:58	[Saint]	so, wait...so change the entire WHILE_WITH_TMO define to that, or just that small section?
12:59:17	[Saint]	and on top of the past attempt, or clean?
12:59:18	wodz	[Saint]: whole define
12:59:19		Nick megal0maniac is now known as Guest89286 (~megal0man@105.184.58.28)
12:59:19		Quit Guest89286 (Killed (sinisalo.freenode.net (Nickname regained by services)))
12:59:22	[Saint]	aha.
12:59:23		Join megal0maniac [0] (~megal0man@105.229.172.191)
13:00
13:05:03	[Saint]	wodz: http://pastebin.com/yTMZacDs
13:05:16	[Saint]	different error, entirely similar behavior.
13:06:12	[Saint]	device hard locks (but the backlight thread and buttons at least seems running)
13:07:04	wodz	[Saint]: so I honestly don't know how this worked for you the first time (I mean the first patchset). With the last try both codes should be exactly the same
13:08:52	[Saint]	WHen I had this patch series applied last I also had a fair amount of other shit going on in my local tree, and while I'm /fairly/ confident I had it applied correctly, I may indeed not have.
13:09:27	[Saint]	I'm about 99.99% certain I did revert my local changes prior.
13:11:04		Quit mg_ (.net .split)
13:11:05		Quit mikroflops (.net .split)
13:11:05		Quit varogami (.net .split)
13:11:05		Quit WakiMiko (.net .split)
13:11:06		Quit maraz (.net .split)
13:11:06		Quit Marex (.net .split)
13:11:07		Quit igitoor (.net .split)
13:11:07		Quit gevaerts (.net .split)
13:11:08		Quit froggyman (.net .split)
13:11:08		Quit Elfish (.net .split)
13:11:08		Quit Topy44 (.net .split)
13:11:09		Quit davidfg4 (.net .split)
13:11:09		Quit Ketturi (.net .split)
13:11:12		Join Elfish [0] (amba@2001:1608:12:1:13:3:3:7)
13:11:13		Join maraz [0] (maraz@kapsi.fi)
13:11:15		Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
13:11:23		Join mikroflops [0] (~yogurt@178.174.137.62)
13:11:26		Join WakiMiko [0] (~WakiMiko@unaffiliated/wakimiko)
13:11:31		Join froggyman [0] (~frogs@unaffiliated/froggyman)
13:11:34		Join Topy44 [0] (topy@cl-406.cgn-01.de.sixxs.net)
13:11:37		Join igitoor [0] (igitur@2a00:d880:3:1::c1ca:a648)
13:11:37		Join Marex [0] (~Marex@195.140.253.167)
13:11:39		Join Ketturi [0] (ketturi@hilla.kapsi.fi)
13:11:42		Join varogami [0] (~varogami@dynamic-adsl-94-34-4-47.clienti.tiscali.it)
13:11:44		Join mg_ [0] (martigam@cassarossa.samfundet.no)
13:11:46		Join davidfg4 [0] (david@pdpc/supporter/active/davidfg4)
13:12:08	[Saint]	What I would like to comprehend is why GCC so badly brutalizes the original patch set, whereas binaries I've been supplied by TheSeven using his fancy-pants toolchain work perfectly.
13:12:33	[Saint]	Is one more or less permissive, or aggressive, then the other?
13:13:56	[Saint]	do you have any insight there, wodz? I'm naive as to such things.
13:14:23		Join nk2032 [0] (~nick001@ANantes-652-1-275-161.w2-8.abo.wanadoo.fr)
13:15:04	wodz	[Saint]: I don't unederstand the question, honestly
13:15:54	*	foolsh is revealed, thought was just him
13:15:55		Quit igitoor (Changing host)
13:15:56		Join igitoor [0] (igitur@unaffiliated/contempt)
13:19:09	[Saint]	My naivety showing obviously. What is it that's so different about TheSeven's setup that allows his original patch set to compile against git head and function perfectly, and yet fall over completely in the Rockbox default build environment?
13:20:42	wodz	[Saint]: his patchset doesn't build with rb toolchain so I still fail to understand your question
13:21:20	wodz	[Saint]: 7's test build uses gcc 4.8 or something
13:21:30	[Saint]	He doesn't use our toolchain - is the point - and I wonder what it is that's so different about his environment
13:21:48	wodz	[Saint]: gcc?
13:21:55	wodz	:-)
13:21:56	[Saint]	sorry that took so long to convey.
13:22:44	wodz	gtf
13:22:51	wodz	*gtg
13:22:54		Quit wodz (Quit: Leaving)
13:23:03	foolsh	Really? Thats cool, gcc is little bitch if newlib or something else isn't "just right"
13:23:56	foolsh	we need TheSeven to write a new rockboxdev.sh
13:25:50	[Saint]	IIUC, he's rather fucked for time presently.
13:26:54	*	[Saint] is reminded he wants to shoehorn setting up for Android into rockboxdev.sh
13:27:28	foolsh	yeah, it needs it, it would help keep it from falling down so much
13:35:48	***	Saving seen data "./dancer.seen"
13:35:51		Join krabador [0] (~krabador@unaffiliated/krabador)
13:44:59		Quit foolsh (Remote host closed the connection)
13:51:58		Join kugel_ [0] (~kugel@rockbox/developer/kugel)
13:57:32		Join kugel__ [0] (~kugel@rockbox/developer/kugel)
14:00
14:00:41		Quit kugel_ (Ping timeout: 272 seconds)
14:02:28		Quit kugel__ (Read error: Connection reset by peer)
14:02:35		Join kugel_ [0] (~kugel@rockbox/developer/kugel)
14:03:44		Quit kugel_ (Remote host closed the connection)
14:03:52		Join kugel_ [0] (~kugel@rockbox/developer/kugel)
14:04:31		Quit kugel_ (Remote host closed the connection)
14:04:38		Join kugel_ [0] (~kugel@p5DDB7DFB.dip0.t-ipconnect.de)
14:04:38		Quit kugel_ (Changing host)
14:04:38		Join kugel_ [0] (~kugel@rockbox/developer/kugel)
14:10:28		Quit kugel_ (Remote host closed the connection)
14:25:12		Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
14:27:57		Join xorly [0] (~xorly@wced-94-218-32-147.feld.cvut.cz)
14:43:29		Quit cmhobbs (Ping timeout: 246 seconds)
14:46:42		Quit mortalis (Ping timeout: 258 seconds)
14:47:34		Quit dfkt (Remote host closed the connection)
14:48:43		Join dfkt [0] (dfkt@unaffiliated/dfkt)
14:58:03	pamaury	yuriks: my memory might be wrong but I think isochronous endpoints have to use a latency of 1 (ie bInterval = 1)
14:59:17	pamaury	but maybe I'm wrong, in which it is indeed a good idea to slow it down to avoid queuing :)
15:00
15:00:52		Quit xorly (Ping timeout: 240 seconds)
15:01:41		Join amayer [0] (~amayer@mail.weberadvertising.com)
15:03:03		Quit TheSeven (Ping timeout: 260 seconds)
15:04:17		Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
15:07:37		Quit efyx_ (Quit: Quitte)
15:08:50	pamaury	yuriks: indeed one can set bInterla for iso endpoints, don't know why I thought different, so yeah that's definitely the way to go, so you don't even have to modify the usb drivers
15:11:10		Join pamaury_ [0] (~quassel@rockbox/developer/pamaury)
15:15:27		Join xorly [0] (~xorly@wced-94-218-32-147.feld.cvut.cz)
15:18:22		Quit pamaury_ (Ping timeout: 240 seconds)
15:18:55		Quit shamus (Read error: Connection reset by peer)
15:24:59		Quit xorly (Ping timeout: 260 seconds)
15:35:51	***	Saving seen data "./dancer.seen"
15:49:37		Join shamus [0] (~shmaus@ip-206-192-193-180.marylandheights.ip.cablemo.net)
16:00
16:05:52		Join chrisb [0] (~chrisb@pool-71-175-243-4.phlapa.east.verizon.net)
16:21:29		Quit Scall (Ping timeout: 244 seconds)
16:26:33		Join Scall [0] (~chat@unaffiliated/scall)
16:28:14		Quit chrisb (Ping timeout: 272 seconds)
16:28:46	fs-bluebot	Build Server message: New build round started. Revision 40a2ac4, 255 builds, 32 clients.
16:33:49	fs-bluebot	Build Server message: Build round completed after 303 seconds.
16:33:58	fs-bluebot	Build Server message: New build round started. Revision aa8ba60, 255 builds, 29 clients.
16:44:26		Join chrisb [0] (~chrisb@li482-205.members.linode.com)
16:50:35	fs-bluebot	Build Server message: New build round started. Revision 877bd98, 255 builds, 33 clients.
17:00
17:05:47		Quit krabador (Ping timeout: 260 seconds)
17:35:55	***	Saving seen data "./dancer.seen"
17:54:05		Quit pamaury (Read error: Connection reset by peer)
18:00
18:02:16		Quit mc2739 (Ping timeout: 245 seconds)
18:04:22		Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
18:06:36		Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)
18:07:36		Join SteveTh3Pirate [0] (586bf76d@gateway/web/freenode/ip.88.107.247.109)
18:40:51		Quit maruk1 (Quit: Leaving.)
18:44:39	mg_	hm, there is a .flac file that my computer can play (with mpv for example) that the clip+ won't, it just stops in the same spot every time.. Is this a known issue?
18:46:10	copper	mg_: run flac -t file.flac on your computer
18:47:24	mg_	get "ok" on two files, both of which have this problem (from the same album)
18:48:19	copper	ok, connect your clip+ to your computer over USB, then run flac -t on the file that's on the clip+
18:49:52	mg_	good plan
18:49:58	mg_	yeah, corrupted somehow
18:50:37	copper	There you go.
18:50:50	mg_	thank you :)
18:50:56	copper	"flac -t" tests for file integrity
18:51:23	mg_	yeah I figured. Good tip!
18:51:23	copper	some decoders are more forgiving with corrupted files
18:51:42	copper	I suggest you fsck your clip+
18:52:10	copper	fsck.vat -a -v
18:52:15	copper	er
18:52:18	copper	fsck.vfat -a -v
18:53:01	copper	sudo fsck.vfat -a -v /dev/sdX1
18:55:40	mg_	Hmm, no errors, but that might be expected? got this though
18:55:42	mg_	FATs differ but appear to be intact. Using first FAT.
18:58:46	copper	now you can delete corrupted files, copy them again, and run flac -t on the copies
18:59:04	copper	you might want to run it on all your files (on the Clip+)
18:59:40	copper	find <MOUNT_DIR> -type f -iname '*.flac' -exec flac -t '{}' '+'
19:00
19:00:38		Join y4n [0] (~y4n@unaffiliated/y4ndexx)
19:01:54	mg_	yeah already on it
19:11:02	copper	anyone: how's the Samsung YP-R0 with Rockbox?
19:11:20	copper	Does the SD card work? The wiki isn't very clear on that point.
19:14:38	copper	never mind
19:14:42	copper	http://rmaa.dfkt.tk/Comparisons/Samsung%20YP-R0%20Loads.htm
19:14:43	copper	:-/
19:16:17	copper	hmmm
19:16:39	copper	actually I'm still looking for information on it
19:17:01	copper	I could use my Bose QC15s with it (output impedance doesn't matter with those)
19:17:37	copper	where's lebellium when you need him!
19:20:04	copper	http://www.anythingbutipod.com/forum/showthread.php?t=58642 suggests the card reader works
19:23:41		Quit zoktar (Quit: -)
19:29:12	SteveTh3Pirate	Okay so i kinda need help with my ipod after installing rockbox, i have a white screen, with the following codes. PANIC mount:0 pc:00051C9C sp:000D8350 bt end. Can anyone help me get my ipod back to normal please
19:30:50		Join zoktar [0] (~zoktar@unaffiliated/zoktar)
19:32:47		Quit Rower (Ping timeout: 260 seconds)
19:34:16		Join lebellium [0] (~chatzilla@89-93-178-161.hfc.dyn.abo.bbox.fr)
19:35:59	***	Saving seen data "./dancer.seen"
19:41:19	copper	lebellium: how's the Samsung YP-R0 with Rockbox?
19:41:26	copper	(hi)
19:41:50	copper	I probably asked you already, a while ago, but I can't remember
19:42:02	lebellium	what's the exact question? 'cause you already know that the YP-R0 is the best mp3 player ever :)
19:42:07	copper	lol
19:42:10	copper	WHY
19:42:41	copper	what are the quirks?
19:42:55	SteveTh3Pirate	Anyone see my question?
19:43:28	copper	SteveTh3Pirate: yes, people who know will answer whenever they're available
19:43:53	lebellium	because it has a nice aluminum design with tactile buttons, microSD slot, a vertical screen, and a nice hardware (IMX37 532Mhz)
19:44:11	SteveTh3Pirate	ah okay, ive tried looking online but i cannot find a solution anywhere :(
19:44:20	copper	how's the display?
19:44:38		Join krabador [0] (~krabador@unaffiliated/krabador)
19:44:51	lebellium	quite similar to Fuze+'s display
19:45:01	lebellium	but it doesn't have a shitty plastic case and touch buttons
19:45:04	lebellium	here is the difference :)
19:45:13	copper	yeah
19:45:18	copper	hmmm
19:45:41	copper	lebellium: would you sell me one of yours? I see you have 5 of them!
19:45:58	lebellium	copper: no, but I would buy another if you sell
19:46:03	lebellium	it to me
19:46:04	copper	lol
19:46:12	lebellium	5 is not enough
19:46:18	copper	what the hell do you need 5 of them for?
19:47:17	lebellium	Now for collection and backup (if one dies), before because I worked both with Samsung R&D on OF and on Rockbox developement so I needed several devices
19:47:50	copper	how's scrolling in the file or database browser?
19:48:03	copper	long press down, I presume?
19:48:07	lebellium	yep
19:48:22	SteveTh3Pirate	copper: i have managed to get it to boot normally, thank you anyway :)
20:00
20:00:14		Join xorly [0] (~xorly@m180.dkm.cz)
20:01:21		Quit SteveTh3Pirate (Quit: Page closed)
20:07:37		Join n1s [0] (~n1s@rockbox/developer/n1s)
20:15:23		Quit bluebrother (Disconnected by services)
20:15:29		Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
20:16:51		Quit fs-bluebot (Ping timeout: 245 seconds)
20:21:50		Quit n1s (Quit: Ex-Chat)
20:23:46		Join fs-bluebot [0] (~fs-bluebo@g225252163.adsl.alicedsl.de)
20:38:36		Quit chrisb (Ping timeout: 245 seconds)
20:43:54		Join einhirn [0] (~Miranda@bsod.rz.tu-clausthal.de)
20:44:10		Quit megal0maniac (Killed (kornbluth.freenode.net (Nickname regained by services)))
20:44:14		Join megal0maniac [0] (~megal0man@ti-227-3-48.telkomadsl.co.za)
20:44:24		Join chrisb [0] (~chrisb@li482-205.members.linode.com)
21:00
21:28:26		Quit TheSeven (Ping timeout: 272 seconds)
21:29:25		Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
21:36:00	***	Saving seen data "./dancer.seen"
21:39:59		Quit kvieta (Quit: okay)
21:40:44		Join lycide [0] (~sneak.znc@149.255.100.107)
21:43:23		Part LinusN
22:00
22:04:32		Join foolsh [0] (~bbrown@c-67-174-138-234.hsd1.in.comcast.net)
22:07:52		Quit y4n (Quit: MOTHER EUROPA CALLING ME!)
22:11:15		Join rela [0] (~x@pdpc/supporter/active/rela)
22:15:37		Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)
22:22:07		Join Pierluigi [0] (~quassel@host-84-13-168-162.as13285.net)
22:27:31		Quit lycide (Quit: greedo shot first)
22:27:50	Pierluigi	hi guys, can anyone give me some extra info about gerrit?
22:28:48	yuriks	like for example?
22:37:26		Quit rela (Read error: Connection reset by peer)
22:40:12	Pierluigi	was wondering about the kind of git command people use when submitting patches
22:40:34	Pierluigi	do people use git commit "−−amend" every time, even for incremental patches?
22:41:01	Pierluigi	I mean, even for commits that would normally be done as a plain "git commit" (no amend)
22:42:02	foolsh	either use amend of squash your commits or gerrit will open separate task for each one
22:43:38	*	foolsh found out the hard way
22:43:57	foolsh	I clobbered gerrit the first time I used it
22:44:12		Quit krabador (Ping timeout: 260 seconds)
22:45:19	Pierluigi	:)
22:45:29		Join fml [0] (~chatzilla@p5DCB6565.dip0.t-ipconnect.de)
22:45:44	Pierluigi	I thought I could do a "git commit" and copypaste previous changeid in commit message
22:46:11	Pierluigi	so that I could generate a patch set with only relevant changes visible
22:46:28		Quit nk2032 ()
22:47:09	fml	Ehm... The alignment in keymap-yh8xx_yh9xx.c is a bit weird IMO. I think the goal is to have the columns left aligned. Aligned closing braces are not so important. Their alignment can be achieved by adding extra spaces at line ends (before braces).
22:47:42	fml	As of now, the right column is a bit hard to read
22:47:57	Pierluigi	from what I can see, with git commit −−amend, the web interface clutters the patch with all files, even the previous ones that I haven't touched (in my previous commit)
22:48:26	Pierluigi	of course this makes sense from a git standpoint of view, because I'm amending previous commit
22:49:08	Pierluigi	so I was wondering whether or not people use a plain "git commit" with previous changeid to generate only relevant changes
22:49:12	foolsh	Pierluigi: I've only messed with change id when hijacking another's task, I usually use −−amend, but squash really big projects before I submit it to gerrit to be safe, so I really don't know
22:50:28		Join lycide [0] (~sneak.znc@149.255.100.107)
22:50:28		Quit lycide (Excess Flood)
22:50:29	Pierluigi	ok foolsh thanks anyway
22:50:58		Join lycide [0] (~sneak.znc@149.255.100.107)
22:50:58		Quit lycide (Excess Flood)
22:51:28		Join lycide [0] (~sneak.znc@149.255.100.107)
22:51:28		Quit lycide (Excess Flood)
22:52:58		Join lycide [0] (~sneak.znc@149.255.100.107)
22:53:06	foolsh	fml: I think I see what you mean, in some places its hard to tell there is even a right column at all
22:54:58		Quit lycide (K-Lined)
23:00
23:09:40		Join Catelite [0] (~Deluge@Powder/Staff/Catelite)
23:12:51		Quit xorly (Ping timeout: 240 seconds)
23:13:33		Quit Strife89 (Quit: Leaving)
23:14:50		Quit fml (Quit: ChatZilla 0.9.90.1 [Firefox 31.0/20140717132905])
23:25:17		Quit Provel (Ping timeout: 272 seconds)
23:25:26		Join Provel [0] (Provel@75-132-30-64.dhcp.stls.mo.charter.com)
23:36:02	***	Saving seen data "./dancer.seen"
23:38:15		Quit amayer (Quit: Leaving)
23:41:58		Quit shamus (Ping timeout: 260 seconds)
23:46:43	Pierluigi	ok I think I've found an answer to my git question
23:47:24	Pierluigi	found someone on stackoverflow asking the same thing: http://stackoverflow.com/questions/8662961/is-there-a-way-to-generate-incremental-patch-set-in-gerrit-using-git
23:48:06	Pierluigi	but basically no satisfying answer
23:49:11	Pierluigi	I guess this link contains a good hint at the answer: http://stackoverflow.com/questions/6357187/is-it-possible-to-add-patch-sets-to-a-gerrit-review-without-ammending-squashing
23:49:15	[Saint]	There are very few satisfying answers in git.
23:49:34	Pierluigi	so basically −−-Gerrit "patches" are meant as a replacement for the original change being reviewed, not a child commit. The 2nd and subsequent patches are not rewriting history, they are replacements for the present. As such an "amendment" is suitable because history has not been written yet.−−−−
23:49:37	[Saint]	It seems to be one of those pieces of software you can either use well, or, will struggle with for all eternity.
23:50:59	Pierluigi	at the end of the day gerrit doesn't care because the way to see changes between patches is via webui itself, not by issuing a different commit command
23:53:36	Pierluigi	on the web page u simply select the previous patch from the "Reference Version" the drop down list, then click on the wanted patch and diff it
23:54:38	Pierluigi	so basically you always want to use "−−amend" so that git automatically puts the previous changeid for you
23:55:15	Pierluigi	or at least....this is my interpretation of it...
23:55:23		Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)

Previous day | Next day