#rockbox 2021-09-24

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 2021-09-24

00:59:07		Quit advcomp2019_ (Read error: Connection reset by peer)
00:59:30		Join advcomp2019_ [0] (~advcomp20@user/advcomp2019)
01:00
01:26:16		Join ZincAlloy [0] (~Adium@ip5f5abcae.dynamic.kabel-deutschland.de)
01:30:47		Quit ZincAlloy (Ping timeout: 265 seconds)
01:46:29	***	Saving seen data "./dancer.seen"
01:47:27		Quit pixelma (Quit: .)
01:47:27		Quit amiconn (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.)
01:53:17		Join amiconn [0] (jens@p200300ea8723ee00305e95fffec66ff3.dip0.t-ipconnect.de)
01:53:17		Join pixelma [0] (marianne@p200300ea8723ee00305e95fffec66ff3.dip0.t-ipconnect.de)
01:53:51		Join ZincAlloy [0] (~Adium@ip5f5abcae.dynamic.kabel-deutschland.de)
01:58:04		Quit ZincAlloy (Ping timeout: 252 seconds)
03:00
03:46:33	***	Saving seen data "./dancer.seen"
05:00
05:46:34	***	No seen item changed, no save performed.
06:00
06:52:09	speachy	anyone here object if I take things down for an hour or so this coming Sunday?
06:52:42	speachy	(everything on www + builds + buildmaster)
07:00
07:12:22	speachy	DNS is likely to be the long tail again, but there are a lot of moving parts this time.
07:19:49		Quit spork (Ping timeout: 252 seconds)
07:25:08	braewoods	speachy: probably not
07:33:32		Join spork [0] (topic@31-151-2-135.dynamic.upc.nl)
07:46:37	***	Saving seen data "./dancer.seen"
07:47:39		Join massiveH [0] (~massiveH@ool-18e4e82f.dyn.optonline.net)
07:57:08		Join danwellby [0] (~danwellby@88.97.8.245)
08:00
08:45:59		Quit massiveH (Quit: Leaving)
09:00
09:46:41	***	Saving seen data "./dancer.seen"
11:00
11:36:15		Join amachronic [0] (~amachroni@user/amachronic)
11:38:01	amachronic	looks like I was right about needing post-DMA cache flush on the x1000. I caught it red-handed in the USB storage driver.
11:38:43	amachronic	Linux also does post-DMA flush: https://github.com/torvalds/linux/blob/master/arch/mips/mm/dma-noncoherent.c#L15
11:39:41	amachronic	given that the jz47xx are also xburst CPUs, I wonder if their drivers should be doing this as well.
11:46:45	***	No seen item changed, no save performed.
12:00
12:00:16		Quit amachronic (Quit: amachronic)
12:28:03		Join lebellium [0] (~lebellium@2a01cb04012c0900dd6ffdf0b12623c7.ipv6.abo.wanadoo.fr)
13:00
13:46:48	***	Saving seen data "./dancer.seen"
14:00
14:17:10		Join ZincAlloy [0] (~Adium@ip5f5abcae.dynamic.kabel-deutschland.de)
14:30:43		Quit dys (Remote host closed the connection)
14:33:58		Join dys [0] (~dys@user/dys)
14:59:47	_bilgus	amachronic, likely..
15:00
15:00:34	_bilgus	I dont think we ever got the supplemental datasheet covering the dmas
15:32:14		Join tomato3 [0] (~tomato@user/tomato)
15:33:22		Quit tomato (Ping timeout: 260 seconds)
15:33:23		Nick tomato3 is now known as tomato (~tomato@user/tomato)
15:46:51	***	Saving seen data "./dancer.seen"
16:00
16:41:52		Quit ZincAlloy (Quit: Leaving.)
17:00
17:39:58		Join MarcAndersen [0] (~no_znepna@85.218.172.116)
17:41:03	MarcAndersen	Hi. Is it just me, or is the gigabeat s transfer speed very slow?
17:46:54	***	Saving seen data "./dancer.seen"
17:52:32	braewoods	MarcAndersen: when are you initiating the transfer
17:55:09	MarcAndersen	What do you mean?
17:56:43	MarcAndersen	And also it's shown as a ZF18-128 USB Device, shouldn't it be shown as rockbox intarnal storage usb device, or is that not on the gigabeat?
18:00
18:02:04	braewoods	MarcAndersen: probably the name of the internal drive
18:04:08	MarcAndersen	There are a lot of things on the internal storage: Content, firmware, devcert.dat, drmstore.dat, MediaLibrary.edb and MediaLibrary_thumbs.edb. Can I delete all of that junk or is some of it needed for rockbox?
18:05:14	braewoods	probably OF files. i would leave them.
18:05:59	braewoods	if you try to transfer files, what do you get?
18:06:17	braewoods	it may also help to try it from within rockbox vs the bootloader
18:07:20	braewoods	you should get between 10 and 20 megabytes a second
18:08:27	MarcAndersen	Hmm, this time it transfers very quickly, it was only the first time it took 4 minutes to transfer a 200 mb file
18:08:41	braewoods	if performance is still bad it might help to connect it to power while doing usb
18:08:48	braewoods	this unit can't charger from the usb port
18:09:22	MarcAndersen	Not if I set it to force?
18:09:50	braewoods	won't matter afaik. the OF only charged from barrel jak.
18:09:54	braewoods	jack
18:10:06	braewoods	you're free to try it though
18:11:05	braewoods	normally i'd suggest using a power meter but that requires the ability to read it...
18:11:25	braewoods	but afaik the only power drawn in this case is to power the usb circuitry
18:11:36	MarcAndersen	This is strange. Writing is super fast but reading is just crap
18:11:48	braewoods	probably caching
18:11:52	braewoods	you're on Linux?
18:11:57	MarcAndersen	Windows
18:12:01	braewoods	huh.
18:12:24	braewoods	well let me see how mtp fares
18:18:32	MarcAndersen	I have a read speed of 0.70 MB/Sec
18:20:21		Quit lebellium (Quit: Leaving)
18:22:55	braewoods	hm
18:23:06	braewoods	10MB/s here from Linux, booted into RB
18:23:18	MarcAndersen	Done
18:24:24	MarcAndersen	Sorry I missunderstood your message, I am also in rb
18:27:54	braewoods	strange. 2MB/s read? i think i found the explaination.
18:28:14	braewoods	it also sucks under Linux unless i specify a larger block size.
18:28:26	braewoods	(with an unmodified gigabeat S no less)
18:28:48	MarcAndersen	How can I set the block size?
18:28:53	braewoods	sounds like windows is performing inefficient accesses
18:30:05	MarcAndersen	And why is it fast at writing and not reading? That makes no sense
18:31:30	braewoods	actually it makes perfect sense. windows is probably buffering the writes.
18:31:40	braewoods	that means they aren't written to disk immediately
18:31:48	braewoods	so it returns before it's fully written
18:32:00	MarcAndersen	Can I do anything about it?
18:32:24	braewoods	i'm looking. under linux i could deal with it by specifying the IO size used for file transfers.
18:32:32	braewoods	in my tests at least
18:33:09	braewoods	long term this might be best solved with MTP since rockbox can at least choose a suitable IO configuration but for now this is what we have
18:34:36	MarcAndersen	How do I get it into mtp?
18:35:04	braewoods	you don't. we don't have an MTP driver right now.
18:35:33	MarcAndersen	So I can't do anything about it?
18:36:44	braewoods	there might be a solution though i know more about how Linux handles IO
18:40:03	braewoods	hm
18:41:48	braewoods	MarcAndersen: what happens if you do this from a Linux host
18:45:25	MarcAndersen	I don't have one right here, and it's getting late here. I can try to set something up tomorrow.
18:48:25	braewoods	ok. i'll look into it.
18:52:06		Quit MarcAndersen (Quit: I was using NightOwl 0.2.)
18:59:21		Join MarcAndersen [0] (~no_znepna@85.218.172.116)
19:00
19:46:55	***	Saving seen data "./dancer.seen"
20:00
20:05:13		Join massiveH [0] (~massiveH@ool-18e4e82f.dyn.optonline.net)
20:47:53	MarcAndersen	Sorry if this is a stupid question, but why are some rock files only included with some player builds? For example there are many more on the gigabeat s than on the xduoo x3?
21:00
21:05:31	_bilgus	MarcAndersen, generally plugins are included / excluded based on criterion such as [screen size, color/ bit depth, plugin buffer size]
21:05:45		Join dconrad [0] (~dconrad@208.38.228.17)
21:07:36	MarcAndersen	Oh, that's the reason, thnks
21:22:32	dconrad	I did a discharge batterybench with gerrit patch 3822 applied, but apparently it didn't write to disk before shutting down :-(
21:22:56	dconrad	trying again, hopefully I get the full file this time
21:24:39	dconrad	I only got 16 hours :-\
21:46:56	***	No seen item changed, no save performed.
21:59:10		Join Moriar [0] (~moriar@107-200-193-159.lightspeed.stlsmo.sbcglobal.net)
22:00
22:04:21		Quit MarcAndersen (Read error: Connection reset by peer)
22:32:54	_bilgus	perhaps the low battery shutdown didn't leave enough time to write the file? I thought it did all or nothing not in chunks
22:44:17	dconrad	I think it writes at 1000 entries (~16 hours) and then when it shuts down
22:56:36		Quit dconrad ()
23:00
23:46:58	***	Saving seen data "./dancer.seen"
23:48:46		Quit j-r (Ping timeout: 252 seconds)
23:49:23		Join j-r [0] (~j-r@p2003000623793827404207fffefd0a65.dip0.t-ipconnect.de)

Previous day | Next day