[00:00]	<JoeBorn>	kind of an ftp based location free tv type of thing
[00:00]	<[g2]>	I'd scp if you're copying files
[00:01]	<[g2]>	scp with public key and no passwords allowed
[00:02]	<[g2]>	copying files down from the Internet won't be as much of an issue, getting out the internet may be (unless you've got something like FIOS(
[00:02]	<JoeBorn>	scp into the NSLU2?
[00:02]	<[g2]>	yeah
[00:03]	<[g2]>	the "unslung" firmare is the basic firmware+ extra packages
[00:03]	<[g2]>	so adding ssh gives you ssh+scp
[00:03]	<JoeBorn>	how hard is that to setup?
[00:03]	<JoeBorn>	i'll still need to solve the port forwarding issue, right?
[00:04]	<[g2]>	it can run iptables too
[00:04]	<[g2]>	when are you leaving ?
[00:05]	<[g2]>	a box I sell is like a slug on steroids
[00:05]	<[g2]>	runs debian and can scp at 2MBs
[00:05]	<[g2]>	does 100Mbs filtering and forwarding
[00:07]	* [g2]  's media file sync finished. 91.6GB in 82.5 minutes
[00:07]	<[g2]>	sweet
[00:18]	<JoeBorn>	had no idea you actually sold them, do you have any?
[00:18]	<[g2]>	JoeBorn yeah I've got a few
[00:19]	<JoeBorn>	unfortunately I'm leaving sunday, but I'd be interested in one.
[00:19]	<[g2]>	it's a shame I didn't know earlier
[00:19]	<JoeBorn>	yea, I got the idea in a hurry and you weren't around that day...
[00:19]	<[g2]>	I can have you setup with Debian running on it from CF
[00:21]	<[g2]>	there are lots of options
[00:22]	<[g2]>	JoeBorn we'll have to chat after CES, but it's a low power fanless box
[00:23]	* [g2]  has gone running and biking with GPS logging running from 8AAs on it
[00:23]	<[g2]>	pretty amazing as it's a 533Mhz with 3 network processors on chip
[00:23]	<[g2]>	and that powered the GPS too
[00:25]	<JoeBorn>	holy smokes
[00:30]	<[g2]>	nite all
[00:31]	<JoeyBorn>	night
[00:37]	<JoeyBorn>	it's working!!!
[00:49]	<JoeyBorn>	fellas, I am watching Dr. Phil crap on some young gal on my laptop, this is huge!
[03:32]	<nerochiaro>	jborn__: i hope you didn't mean that literally
[03:33]	<nerochiaro>	jborn__: how's CES ?
[04:04]	<nerochiaro>	hey crweb, you there ?
[04:05]	<shirour>	hi everybody
[04:05]	<nerochiaro>	hey
[04:06]	<shirour>	nerochiaro: what's up?
[04:06]	<nerochiaro>	messing around with enlightenment libraries, fixing them up to work on the osd
[04:10]	<crweb>	yo
[04:10]	<crweb>	i am here
[04:11]	<nerochiaro>	crweb: yo, did you see the post from nathan on the imem stuff ?
[04:11]	<crweb>	i haven't
[04:11]	<nerochiaro>	go see that, then
[04:12]	<shirour>	hi crweb
[04:12]	<crweb>	oh yes
[04:12]	<shirour>	why enlightenment?
[04:12]	<crweb>	I totally understand the issue
[04:12]	<crweb>	they keep talking about the audio/video codecs
[04:12]	<crweb>	thats not our issue
[04:13]	<nerochiaro>	crweb: that's what i was going to point out in the reply i'm writing
[04:13]	<crweb>	our issue is ALL hardware being allocated statically
[04:13]	<nerochiaro>	crweb: that core.c file defines lots of other static allocations
[04:13]	<nerochiaro>	indeed
[04:13]	<nerochiaro>	shirour: one minute
[04:14]	<shirour>	where can i find the posts you're talking about? didn't see any in gmail's group
[04:14]	<crweb>	yeah, i totally userstand the audio/video stuff
[04:14]	<nerochiaro>	shirour: it's there. use the group search for "imem"
[04:14]	<crweb>	its stupid, and wrong, but thats acceptable
[04:15]	<nerochiaro>	crweb: well, it might be a limitation of the ARM/DSP interface
[04:15]	<nerochiaro>	but only for av
[04:15]	<crweb>	that very well may be
[04:16]	<crweb>	but last i checked, why include a memory manager in linux if nothing can use it
[04:17]	<nerochiaro>	crweb: wheres' the header where HDD_BASE, CFC_BASE etc are defined ?
[04:17]	<shirour>	don't you allocate it by giving the kernel only a small portion of the whole memory ? i tought imem doesn't touch the kernel's mem
[04:17]	<shirour>	nerochiaro: these are dm320's things, as far as i know
[04:17]	<crweb>	nerochiaro: i don't know. I've was basing of private emails
[04:18]	<nerochiaro>	crweb: no wait, i mean, in that thread there's a reference to a core.c file in linux/arch/arm/.../core.c ... you were looking at that the other day
[04:19]	<crweb>	oh
[04:19]	<shirour>	nerochiaro: the dm320's base mem arch defines these addresses. the kernel later changes these a bit when turning on the mmu
[04:19]	<crweb>	hardware.h
[04:20]	<crweb>	include/asm/arch-ntosd-dm320/hardware.h
[04:20]	<nerochiaro>	shirour: crweb: aren't these the definitions of the fixed memory areas ?
[04:20]	<crweb>	i believe so
[04:20]	<crweb>	thats what I gathered
[04:21]	<nerochiaro>	however, see the comment at the top of the hardware file
[04:21]	<nerochiaro>	it talks about virtual memory
[04:21]	<crweb>	bsp/kernel/linux/arch/arm/mach-ntosd-dm320/core.c
[04:22]	<crweb>	yeah, this is a damn inconsistant
[04:22]	<crweb>	since, pre 0.46 the drivers worked fine with the memory manager
[04:22]	<crweb>	why they decided to hard code a ethernet driver, that was working fine, is beyond me
[04:23]	<nerochiaro>	crweb: wait a second. pre .46 in svn, that core.c file is different ? i don't think so
[04:23]	<nerochiaro>	since i'm on .35 and i'm looking at it and it seems to hardcode the stuff in the same way
[04:24]	<crweb>	all I know that is for sure fact is: before i upgraded to .46, I was able to specify 62M of ram and everything work. and now If i specify greator than 16M NOTHING works
[04:25]	<nerochiaro>	that's what worth investingating
[04:25]	<crweb>	it very well could be terry didn't answer my question.
[04:25]	<crweb>	cause they seem to totally miss the point most of the time
[04:25]	<crweb>	i think the default env config changed
[04:26]	<crweb>	err, the real defaults
[04:26]	<crweb>	not just uboot env's
[04:26]	<crweb>	i can't remember what its called
[04:26]	<crweb>	it is 4:30am
[04:26]	<crweb>	;)
[04:27]	<nerochiaro>	i'm not sure what defaults you're talking about
[04:27]	<crweb>	me either
[04:27]	<crweb>	scratch that
[04:27]	<crweb>	something changed
[04:27]	<crweb>	and it was perm
[04:28]	<crweb>	it doesn't matter what kernel I boot with now
[04:28]	<nerochiaro>	so either flash or uboot
[04:28]	<crweb>	It was still working with they went to only 1mb rw
[04:30]	<crweb>	and eithernet still works for tftp and uboot
[04:30]	<crweb>	it fails when the ethernet driver loads
[04:31]	<crweb>	as you can see from the dmesg I posted
[04:31]	<crweb>	all those release() lines
[04:31]	<crweb>	Badness in device_release at drivers/base/core.c:83
[04:32]	<crweb>	Device 'dm9ks.0' does not have a release() function, it is broken and
[04:32]	<crweb>	must be fixed.
[04:32]	<crweb>	these things started sometime between 0.34 and 0.46
[04:32]	<crweb>	they did not occur with mem = 62M on 0.34
[04:32]	<crweb>	Device 'IDE-CFC.0' does not have a release() function
[04:33]	<crweb>	Device 'IDE-HDD.0' does not have a release() function
[04:33]	<nerochiaro>	see the row before that: "imem.0: failed to claim resource 0"
[04:33]	<crweb>	Device 'dm9ks.0' does not have a release() function
[04:33]	<crweb>	those are the 3
[04:33]	<crweb>	yeah
[04:33]	<nerochiaro>	yeah, but the root cause to me seems that imem error message before
[04:33]	<crweb>	wonder what thats about
[04:34]	<crweb>	it does it even with 17M
[04:34]	<crweb>	didn't before..
[04:34]	<nerochiaro>	my guess is: it tries to manage some physical memory, and that memory is already managed by the kernel (per boot settings)
[04:34]	<nerochiaro>	so it chockes
[04:34]	<crweb>	right, but if the file didn't change
[04:35]	<crweb>	why now?
[04:35]	<nerochiaro>	no idea
[04:36]	<crweb>	yeah.. so, you can't go back
[04:36]	<crweb>	everything is invalid package
[04:36]	<crweb>	can you do something for me?
[04:36]	<nerochiaro>	depends ;). try asking
[04:37]	<crweb>	in uboot
[04:37]	<crweb>	wait
[04:37]	<crweb>	/etc/init.d/rcS
[04:37]	<crweb>	change last like to 2
[04:37]	<crweb>	then in uboot, setenv mem_reserve 32M
[04:37]	<crweb>	type boot
[04:38]	<crweb>	and post the bootup
[04:38]	<crweb>	anyone else in this channel, that HAS NOT updated to 0.46 and is still on 2.26-0.34 or older, please contact me.
[04:41]	<nerochiaro>	crweb: i'll do that in a minute. let me finish the post for the ML first to reply to nathan asking about the other non-av devices
[04:41]	<nerochiaro>	i want to uderstand what's going on.
[04:41]	<crweb>	i'm hoping that someone didn't understand my problem and this was a miss communication and I happened onto a bug
[04:42]	<crweb>	but, false hope is a easy wish
[04:42]	<nerochiaro>	even if that was so, i still want to understand that imem stuff
[04:42]	<crweb>	i aswell
[04:42]	<crweb>	i kinda understand what he is saying
[04:42]	<nerochiaro>	but only makes sense for av
[04:42]	<crweb>	exactly
[04:43]	<crweb>	and it it needs 48 mb of buffers
[04:43]	<nerochiaro>	and what terry said in the post before ... nathan seems to fail to comment on that, so i'm going to try and get more explanation
[04:43]	<crweb>	then, we need to find a new codec vendor, cause damn
[04:43]	<crweb>	i have a dvd-rom divx player that has 4mb
[04:43]	<crweb>	and reads from a slow ass dvd
[04:44]	<nerochiaro>	more than 48. it's at least 16MB x 4 for just usb,cf,flash and hdd ... plus all the imem size
[04:44]	<crweb>	yeah
[04:44]	<nerochiaro>	this is what seems strange to me. we don't even have all that mem
[04:45]	<crweb>	nathan is the ingenient guy right?
[04:45]	<crweb>	hes got a good explination
[04:45]	<nerochiaro>	yeah
[04:45]	<crweb>	i totally understand
[04:45]	<crweb>	just wish it was for our issue...
[04:46]	<nerochiaro>	maybe it is, but we dont' understand how. one minute and i'll be done
[04:46]	<crweb>	i'd be willing to give up 32M to a/v buffers
[04:47]	<crweb>	but, it doesn't do much good if we don't have enough mem to load the player...
[04:48]	<crweb>	you know, another thing i've been noticing is
[04:48]	<crweb>	I'm having a lot of net card lockups
[04:48]	<crweb>	and sometimes, the card won't come on unless i unplug and replugin the ethernet
[04:49]	<nerochiaro>	never happened here
[04:49]	<crweb>	it didn't happen pre 0.46 either
[04:50]	<crweb>	and I've had it unplugged so it isn't some random damage
[04:50]	<crweb>	I upgradded to 0.46 my first boot
[04:53]	<crweb>	i really wish i could downgrade...
[05:00]	<crweb>	nerochiaro: why would each codec need its own buffer?
[05:00]	<crweb>	you can only play 1 file at a time..
[05:00]	<nerochiaro>	why do you say it does ?
[05:00]	<crweb>	and you can only have one codec loaded on the dsp
[05:01]	<crweb>	well.. 32MB is one hell of a buffer
[05:01]	<crweb>	especially when an average recording is only 90-180 mb to beginwith
[05:01]	<crweb>	in divx
[05:02]	<crweb>	i know for a fact it can be done with 4mb
[05:02]	<nerochiaro>	crweb: worth asking too, but later
[05:02]	<nerochiaro>	crweb: just posted
[05:03]	<nerochiaro>	crweb: now i'm trying what you asked
[05:03]	<crweb>	cool
[05:03]	<nerochiaro>	that ingenient guy is a blessing, however. i'm really glad he came on board
[05:03]	<crweb>	i mean, if you're filling 32mb of buffer during recording
[05:04]	<crweb>	you're eventually going to start frame dropping
[05:04]	<crweb>	cause you can't keep up
[05:04]	<nerochiaro>	why not ?
[05:04]	<crweb>	it might be 6+ hours of recording
[05:05]	<crweb>	unless they are using buffer for "smoothness"
[05:05]	<crweb>	its going to over flow
[05:05]	<crweb>	and if they are using it for smoothness.. they don't need 1/3 of a recording (speaking for divx)
[05:06]	<crweb>	my decent quality divx tv shows end up at 145mb for 400x300
[05:06]	<crweb>	for 1 hour
[05:07]	<crweb>	and they have to stop advertizing/saying there is 64mb of ram
[05:07]	<crweb>	thats for sure
[05:08]	<crweb>	memory dedicated for a/v buffering shouldn't be included as "system ram"
[05:08]	<nerochiaro>	i can't really comment on the buffers if they are too large or not, but the advertising of 64 you're right
[05:08]	<nerochiaro>	anyway
[05:08]	<nerochiaro>	i tested the stuff
[05:08]	<nerochiaro>	it gives me the same errors as you
[05:08]	<crweb>	damnit
[05:08]	<crweb>	you're too new
[05:09]	<crweb>	0.23 is what i had
[05:09]	<nerochiaro>	i can still try to downgrade
[05:09]	<crweb>	would you please
[05:09]	<crweb>	you don't have to now
[05:09]	<crweb>	I'm just trying to verify
[05:09]	<crweb>	i'm starting to look like a crackpot
[05:10]	<crweb>	i swear I had it working with 62M
[05:10]	<nerochiaro>	maybe they used to reserve less memory before
[05:10]	<crweb>	i never took it above 62
[05:10]	<crweb>	remember we had 14
[05:11]	<crweb>	now we have 16
[05:11]	<crweb>	so what i did was
[05:11]	<nerochiaro>	i still have 14
[05:11]	<crweb>	i saw that we were using the 64M codecs
[05:11]	<crweb>	and i figured well, the codecs need 2mb
[05:11]	<crweb>	so i put it at 62M
[05:11]	<nerochiaro>	why 2M ?
[05:11]	<crweb>	several others were running find with 64M
[05:11]	<crweb>	16-2 = 14
[05:12]	<crweb>	i thought the very first gamma unit was 16M
[05:12]	<nerochiaro>	32
[05:12]	<nerochiaro>	maybe the initial dev boards
[05:12]	<crweb>	yes, i realize that now ;)
[05:12]	<nerochiaro>	but i think they were 32 too
[05:12]	<nerochiaro>	oh wait a sec
[05:12]	<nerochiaro>	i set mem_reserve wrong, actually disabling it
[05:13]	<nerochiaro>	i'll give it another shot
[05:13]	<crweb>	i always leave out the e
[05:13]	<crweb>	or put mem_reserve=
[05:13]	<crweb>	but if you got that error, its mostlikely going to happen again
[05:13]	<nerochiaro>	isn't it "setenv mem_reserve=mem=32M" ?
[05:13]	<crweb>	no
[05:14]	<crweb>	setenv mem_reserve 32M
[05:14]	<nerochiaro>	nope
[05:14]	<crweb>	the nfs boot command is different than the normal
[05:14]	<nerochiaro>	"printenv mem_reserve" => "mem_reserve=mem=14M"
[05:14]	<crweb>	what the hell
[05:15]	<crweb>	thats new
[05:15]	<crweb>	ahh
[05:15]	<crweb>	damnit
[05:15]	<crweb>	damnit damnit
[05:15]	<crweb>	mem_reserv
[05:15]	<nerochiaro>	i tried both, anyway. and none works. since the kernel believes it has 64M anyway
[05:15]	<crweb>	nfs boot uses mem_reserv
[05:15]	<crweb>	crap..
[05:16]	<crweb>	or used to
[05:16]	<nerochiaro>	reserv without the e works with 32M
[05:16]	<crweb>	it didn't change the mem though
[05:16]	<crweb>	like it used to
[05:17]	<crweb>	they've chaned the boot lines
[05:17]	<nerochiaro>	no wait. crap. give me a minute
[05:17]	<nerochiaro>	how do i set minicom to log to a file ?
[05:17]	<crweb>	it used to be mem_reserv 32M but now they've changed nfs_cmd
[05:17]	<crweb>	i dunno, i copy and paste my terminal output
[05:18]	<nerochiaro>	nfs_cmd uses $(mem_reserve) here
[05:18]	<crweb>	yeah, if you go back 1 version, it won't
[05:18]	<nerochiaro>	yeah, but it doesn't line wrap
[05:18]	<crweb>	yeah, i have 2 1280x1024 displays side by side
[05:19]	<crweb>	i was setting it wrong too
[05:19]	<crweb>	it was using all 64
[05:19]	<crweb>	i think
[05:19]	<nerochiaro>	yes, i just confirmed that. now i'm positive i had it use 32M and it works
[05:19]	<anders_>	^A w to enable linewrap in minicom.
[05:20]	<crweb>	well..
[05:20]	<crweb>	damn
[05:20]	<nerochiaro>	anders_: thanks
[05:20]	<crweb>	nerochiaro: now try 62M
[05:20]	<nerochiaro>	crweb: Memory: 32MB = 32MB total
[05:20]	<nerochiaro>	crweb: this is the line you should watch for during boot
[05:20]	<crweb>	yes, 32M is working for me
[05:20]	<nerochiaro>	trying 62M
[05:20]	<crweb>	crap, i think this was all my fault..
[05:20]	<crweb>	i didn't know the changed the mem_reserv line
[05:20]	<crweb>	and didn't read the boot lines
[05:21]	<crweb>	nfs_cmd used to be mem=($mem_reserv)
[05:21]	<crweb>	now its $mem_reserve
[05:21]	<nerochiaro>	62M works
[05:21]	<crweb>	crap
[05:21]	<crweb>	i'm an idiot
[05:22]	<nerochiaro>	i mean, it doesn't give me the errors you had
[05:22]	<crweb>	its not giving them to me either
[05:22]	<crweb>	set it to 64M
[05:22]	<crweb>	and it will
[05:22]	<crweb>	i figured they were only using 2mb for buffers and stuff
[05:22]	<nerochiaro>	if you are thinking about posting excuses and stuff in ML, wait a sec though ;)
[05:22]	<crweb>	psh
[05:22]	<crweb>	let them sweat it out
[05:23]	<crweb>	my error was my own fault though
[05:23]	<crweb>	and for those irc loggers, scratch that i said i'd gladly give up 32M
[05:23]	<nerochiaro>	lol
[05:23]	<crweb>	i want all 64 now damnit
[05:24]	* chreekat  decides it's too silly a time for a Californian to be awake
[05:24]	<crweb>	I'm an illinoian and I'm up
[05:25]	<crweb>	hmmm...
[05:25]	* crweb  is slightly embarrased..
[05:25]	<crweb>	I'll own up to it after CES
[05:26]	<crweb>	.. since everything else is waiting till after CES
[05:26]	<nerochiaro>	"own up"
[05:26]	<crweb>	?
[05:27]	<nerochiaro>	you said "i'll own up". what does that mean ?
[05:27]	<crweb>	oh
[05:27]	<crweb>	take the blame for ones own mistake
[05:27]	<crweb>	so yeah, Qt4 interface, runs fine with 32M of ram available
[05:27]	<nerochiaro>	too late i think. take a peek on mailing list on that thread. i had to post that to clarify on my previous post.
[05:29]	<crweb>	ok, i'll explain what happened ;)
[05:29]	<crweb>	heh
[05:29]	<crweb>	now i understand what terry was saying
[05:29]	<nerochiaro>	would you wait for that after nathan replies ? so there's less confusion ?
[05:29]	<crweb>	yeah
[05:29]	<nerochiaro>	now wait, what was he saying ?
[05:29]	<crweb>	he said my env didn't match his
[05:30]	<nerochiaro>	that imem releates to cf,usb etc ? how ?
[05:30]	<crweb>	he was talking about the txt, not just the 32M part
[05:30]	<nerochiaro>	ah, ok
[05:30]	<crweb>	he was talking about me corrupting my env's
[05:30]	<crweb>	cause I erased mem=
[05:30]	<crweb>	man i feel like a jackass
[05:31]	<crweb>	though
[05:31]	<nerochiaro>	crweb: yeah, but i still don't understand "This will not only effects the network card, it also will effect the
[05:31]	<nerochiaro>	media cards. As they are all added in that core.c"
[05:31]	<nerochiaro>	from terry
[05:31]	<crweb>	excactly
[05:31]	<crweb>	my point is still valid
[05:31]	<crweb>	just.. not as extreme as I made it out to be
[05:31]	<nerochiaro>	heh
[05:32]	<crweb>	i have my ram back, YAY
[05:32]	<nerochiaro>	feel like a jackass too for jumping in, actually :P
[05:32]	<crweb>	sorry bout that
[05:32]	<nerochiaro>	my fauly
[05:32]	<crweb>	thats twice now
[05:32]	<crweb>	i did fix my segfaulting arm-linux-gcc compiler though
[05:33]	<crweb>	it is pretty sensitive and doesn't like my overclocked cpu
[05:33]	<crweb>	our point is still valid
[05:33]	<crweb>	netcard, ide cf and stuff, shouldn't be hardcoded anywhere
[05:33]	<crweb>	not even in imem
[05:34]	<crweb>	theres nothing special about that network chip
[05:35]	<crweb>	but yeah, if we backoff now, we won't get our answers ;)
[05:36]	<crweb>	it looks to me like they are using less than 1MB
[05:36]	<crweb>	i jsut set for 63M
[05:38]	<crweb>	sweet
[05:38]	<crweb>	my qt is so fast now
[05:38]	<crweb>	loads in 1/2 the time, and scrolls just fine
[05:39]	<nerochiaro>	cool. i'll go get something to eat. be back in a while
[05:45]	<crweb>	Memory: 62464KB available
[05:45]	<crweb>	that was throwing me off
[05:45]	<crweb>	i thought my 62M was working
[07:09]	<allyourrejects>	nerochiaro: quick note
[07:09]	<allyourrejects>	nerochiaro: the remote keycodes change
[07:09]	<allyourrejects>	nerochiaro: if you hold the button down
[07:32]	<crweb>	nerochiaro: also, remember how I said an "unknown" key is read as the home key. I was wrong. a unknown key, is returned as the last key pressed.
[07:32]	<crweb>	so other remotes return as the last OSD remote button pushed
[08:26]	<ymo000>	howdy
[08:30]	<nerochiaro>	crweb: i can't account for the other remotes. as for the keycode that changes on button hold, i was talking about that with Gao. there are 3 different codes for each key: "press", "press and hold for 2 seconds (or so)" and then a third keycode each x msec as long as it stays pressed. as far as i'm concerned that's a brain damaged scheme. there should just be one code for key press and one for key release. the app should handle the rest.
[08:30]	<crweb>	i agree
[08:31]	<ymo000>	its better i think you its handled for you ...
[08:31]	<crweb>	you're going to flood the dev node with hold down events
[08:31]	<crweb>	and it majorly slows down
[08:31]	<nerochiaro>	on my driver i send out one key down and one key up event for each key code i receive. if they are 2nd and 3rd keycodes, then i just send the 1st keycode (2nd and 3rd are just 1st masked)
[08:32]	<nerochiaro>	ymo000: can you repeat what you said ? i didn't uderstand it
[08:32]	<nerochiaro>	sorry
[08:33]	<nerochiaro>	crweb: however that will change once i can get this issue straightened out with Gao
[08:33]	<nerochiaro>	crweb: i'm stil