**** BEGIN LOGGING AT Thu Nov 15 02:59:59 2012
Nov 15 05:29:58 <inteloid> Hi guys, anybody used BB Bone Battery cape?
Nov 15 05:32:28 <inteloid> HELLO!!!
Nov 15 05:34:55 <muhoo> bb bone battery is a great name.
Nov 15 08:13:01 <KotH> grüezi mitenand
Nov 15 08:13:12 <dm8tbr> mooo
Nov 15 08:37:39 <KotH> apt-get moo?
Nov 15 08:47:39 <woglinde_> foo
Nov 15 08:49:08 <panto> mourning
Nov 15 08:51:16 <LetoThe2nd> can you spell foobar? -> http://www.fuba.com/
Nov 15 08:53:28 <av500> fubber
Nov 15 08:53:32 <LetoThe2nd> flubber?
Nov 15 09:36:43 <av500> http://dan.drown.org/android/
Nov 15 09:43:55 * dm8tbr is confused. so it's bionic linked but ipk based? WTF
Nov 15 09:47:21 <av500> dm8tbr: what has libc to do with packaging?
Nov 15 09:47:37 <Lee_> Hi all, I got some errors when cross-compile the OpenCV2.4.2 for BB-xM running Angstrom. I followed this link: http://processors.wiki.ti.com/index.php/Building_OpenCV_for_ARM_Cortex-A8
Nov 15 09:47:38 <av500> it just means that use the system libc and dont bring their own
Nov 15 09:48:27 <Lee_> This is my errors: http://pastebin.com/i0tNjZVe
Nov 15 09:48:35 <Lee_> Thank you so much.
Nov 15 09:49:53 <av500> why not use opencv thats already in angstrom?
Nov 15 09:50:08 <koen> or failing that, why not build it according to the angstrom instructions?
Nov 15 09:50:38 <dm8tbr> av500: yes, it's just that most of the stuff I've seen so far was BYOLibc
Nov 15 09:51:46 <Lee_> av500: I would like to build my project on the UBUNTU host
Nov 15 09:52:05 <Lee_> after that I just copy it to BB-xM and run.
Nov 15 09:53:34 <koen> Lee_: so go to the angstrom website, click on "developers" and follow the instructions
Nov 15 09:54:26 <Lee_> thank you, koen
Nov 15 09:54:57 <koen> fwiw
Nov 15 09:55:07 <koen> on beagleboard.org boards, ignore the TI wiki crap
Nov 15 09:55:17 <koen> it's wrong on so many levels it's no so funny anymore
Nov 15 09:56:12 <Lee_> koen: is that website you give me? http://www.angstrom-distribution.org/
Nov 15 09:56:19 <koen> yes
Nov 15 09:57:41 <Lee_> koen, I cannot find any topic related to opencv
Nov 15 09:59:07 <dm8tbr> start thinking, don't expect 'exact steps'
Nov 15 10:00:22 <av500> koen: ti wiki?
Nov 15 10:00:24 <av500> elinux?
Nov 15 10:08:39 <woglinde_> gm av500
Nov 15 10:17:07 <koen> av500: the processors.ti wiki
Nov 15 10:17:17 <koen> jkridner1: I reproduced the gphoto build error
Nov 15 10:45:40 <LetoThe2nd> jkridner1: ping
Nov 15 11:32:16 <av500> http://www.amazon.com/gp/product/B000VBDIDO
Nov 15 11:35:43 <ogra-cb_> av500, only used ones
Nov 15 11:47:49 <KotH> who was it who told me at ELCE that state machines are bad?
Nov 15 11:48:07 <mdutt3_> hi
Nov 15 11:51:44 <av500> KotH: elaborate?
Nov 15 11:53:51 <av500> after all, the linux kernel is just a state machine with side effects...
Nov 15 11:54:57 <LetoThe2nd> wasn't it that side effects are bad?
Nov 15 11:57:26 <angs> when I configure expansion header pins, does pinconf[0x2168/4]= 0x001C001C refers to GPIO139 or GPIO138?
Nov 15 11:59:47 <KotH> av500: someone told me it's a bad idea to write state machines, but i dont remember why
Nov 15 12:01:45 <mru> might have been me
Nov 15 12:02:25 <panto> KotH, cause if you have more than 10 states you're pretty much screwed
Nov 15 12:02:54 <mru> I've often seen state machines used where none was needed
Nov 15 12:02:55 <KotH> mru: and what was your reasoning?
Nov 15 12:05:00 <panto> for a reason why state machines are a fail, see ITU telecom standards
Nov 15 12:07:01 <av500> if(state==1) {..} else if(state!=1 && state==2) {..} else if(state!=1 && state!=2 && state==3) {..} .......
Nov 15 12:07:22 <mru> av500: dailywtf much?
Nov 15 12:07:48 <panto> av500, when I had to do that I use a different function for each state, and an array of function pointers
Nov 15 12:08:28 <av500> panto: I would make the state selection function a function pointer and change that on each state chnage too
Nov 15 12:08:58 <LetoThe2nd> you all on dope here? ;)
Nov 15 12:09:06 <av500> wanta soma?
Nov 15 12:09:16 <LetoThe2nd> yes please. with some spice.
Nov 15 12:09:23 <mru> av500: I heard/read about someone doing just that recently
Nov 15 12:12:05 <LetoThe2nd> on behalf of weird ideas: can anyone give a pointer for docs/libraries/etc for pattern recognition, it terms of some stream of numerical values?
Nov 15 12:12:13 <av500> squint
Nov 15 12:12:23 <mru> LetoThe2nd: _everything_ is a stream of numerical values
Nov 15 12:12:48 <av500> just give it a Gödel numbering...
Nov 15 12:14:02 <av500> LetoThe2nd: but I guess the financial analyst have an excel macro for that
Nov 15 12:19:15 <angs> av500: could you tell me if pinconf[0x2168/4]= 0x001C001C refers to GPIO139 or GPIO138?
Nov 15 12:20:25 <av500> do I look like a TRM to you?
Nov 15 12:22:18 <angs> what is TRM?
Nov 15 12:22:39 <koen> technical reference manual
Nov 15 12:23:26 <av500> or trained reference monkey
Nov 15 12:23:45 <angs> I just asked it to you because you have good knowledge about it
Nov 15 12:24:12 <av500> still I dont have the register values all in my head
Nov 15 12:24:20 <angs> both gpio138 and 139's physical address is 0x4800 2164
Nov 15 12:24:37 <angs> *2168
Nov 15 12:24:55 <av500> I also dont have the GPIO register layout in my head
Nov 15 12:25:12 <angs> gpio_138 is [15:0] and gpio_139[31:16]
Nov 15 12:25:27 <av500> then what is the question?
Nov 15 12:25:52 <angs> does pinconf[0x2168/4] refers to GPIO139 or GPIO138
Nov 15 12:26:55 <av500> that depends a lot on pinconf I would say
Nov 15 12:27:27 <angs>   pinconf = (ulong*) mmap(NULL, 0x10000, PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0x48000000);
Nov 15 12:28:33 <angs> CONTROL_PADCONF_MMC2_DAT6[15:0] 0x4800 2168 GPIO138
Nov 15 12:28:42 <angs> CONTROL_PADCONF_MMC2_DAT6[31:16] 0x4800 2168 gpio_139
Nov 15 12:29:46 <angs> I wonder how do you define gpio138 and gpio139 by pinconf. is it pinconf[0x2168/4] = gpio138 and pinconf[0x2168/8] = gpio139?
Nov 15 12:30:39 <av500> do you understand why /4?
Nov 15 12:30:44 <av500> mru: ww.cnx-software.com/2012/11/15/congatec-reference-board-for-video-wall-systems-powered-by-amd-embedded-r-series-apu
Nov 15 12:32:00 <angs> av500: no. is it for the first 4 bits like in netmasking?
Nov 15 12:32:24 <av500> you are aware what "/" means in C
Nov 15 12:32:34 <angs> mod
Nov 15 12:32:40 <mru> so no
Nov 15 12:32:51 <thurbad> more likely it's due to register size
Nov 15 12:33:47 <av500> try printf("%x %x\n", 0x2168, 0x2168/4);
Nov 15 12:34:10 <LetoThe2nd> av500: das that code also run on an LPC2168?
Nov 15 12:36:04 <angs> I saw the difference but it still does not make sense what 85a represents?
Nov 15 12:36:50 <av500> a house number?
Nov 15 12:37:01 <av500> ever heard of hexadecimal numbers?
Nov 15 12:38:31 <angs> sure I know it is hex value. I mean I cold not understand how to connect 85a with GPIO138 or GPIO139
Nov 15 12:38:54 <av500> ok, now you need to understand how an "array" in C works
Nov 15 12:39:14 <av500> that thing with the "[" and the "]"
Nov 15 13:02:21 <av500> http://a-hackers-craic.blogspot.de/2012/11/mailing-lists-or-lack-of-self-censoring.html
Nov 15 13:03:46 <angs> I could not figure out where the pointer points. could you hint?
Nov 15 13:03:59 * bradfa now has 27" monitor! muhahaha!
Nov 15 13:04:11 <av500> bradfa: which one?
Nov 15 13:04:13 <mru> bradfa: how many pixels?
Nov 15 13:04:19 <bradfa> HP zr2740w
Nov 15 13:04:38 <bradfa> pilfered from former coworker's desk
Nov 15 13:05:00 <av500> nice, I have that in the dell version
Nov 15 13:05:02 <bradfa> 2560x1440 pixels me thinks
Nov 15 13:05:05 <av500> yes
Nov 15 13:05:13 <mru> 27" is too large
Nov 15 13:05:18 <bradfa> many more than my 1920x1200 had
Nov 15 13:05:25 <av500> mru: step away 3 feet and its retina :)
Nov 15 13:05:26 <mru> can't comfortably fit 2 on the desk, let alone 3
Nov 15 13:05:43 <bradfa> mru, no, but I only have 1 so I don't yet have that problem :)
Nov 15 13:05:45 <mru> now that number of pixels in a 24" case...
Nov 15 13:05:52 <av500> or in 10"
Nov 15 13:06:03 <av500> imagine how many you can fit
Nov 15 13:06:05 <mru> I'll take two 1080p monitors over one of those any day
Nov 15 13:06:10 <bradfa> mru, give apple about 8 months, I bet we see twenty-something "reasonable resolution" monitors
Nov 15 13:06:21 <bradfa> I just want 4:3 aspect ratio back...
Nov 15 13:06:29 <mru> av500: there are limits to even my visual acuity
Nov 15 13:06:30 <bradfa> this 16:9/10 stuff is for the birds
Nov 15 13:06:30 <av500> wont happen
Nov 15 13:06:51 <bradfa> two of the 16:9's in portrait is probably my only hope
Nov 15 13:07:01 <av500> I would love a square monitor
Nov 15 13:07:05 <mru> why do people complain about wide monitors?
Nov 15 13:07:06 <av500> 2560x2560
Nov 15 13:07:10 <mru> just stack more windows horizontally
Nov 15 13:07:12 <bradfa> mru, my code is 80 col wide
Nov 15 13:07:17 <av500> lol
Nov 15 13:07:30 <av500> bradfa: that'S only 80x5 pixels
Nov 15 13:07:33 <mru> bradfa: did it ever occur to you to have more than one window side by side?
Nov 15 13:07:39 <bradfa> I like putting 3 or 4 buffers across
Nov 15 13:07:47 <av500> so 640pixels
Nov 15 13:08:02 <av500> you can put 4 across :)
Nov 15 13:08:37 * bradfa is getting old (not like these young whipper snappers like mranostay), bigger fonts are nice
Nov 15 13:08:38 <mru> also, shifting eyes sideways is less straining than up and down
Nov 15 13:09:00 <bradfa> mru, I will organize my code into sideways, brb
Nov 15 13:09:18 <av500> sed can do that for you
Nov 15 13:09:34 <bradfa> av500, sed can do everything, even control awk!
Nov 15 13:10:03 <koen> bradfa: I love my 27" monitor
Nov 15 13:10:13 <koen> bradfa: especially with a terminal fullscreen
Nov 15 13:10:18 <xxiao> cut 1700 from omap dept?
Nov 15 13:10:24 <av500> yes
Nov 15 13:10:28 <bradfa> xxiao, you're doing the cutting?
Nov 15 13:10:30 <koen> bradfa: still makes me scroll too much for the bone boardfile, though :)
Nov 15 13:10:37 <bradfa> koen, portrait mode?
Nov 15 13:10:45 <av500> koen: organize it sideways
Nov 15 13:10:46 <xxiao> anyone wants my pandaboard for $50?
Nov 15 13:10:58 <koen> bradfa: no, landscape
Nov 15 13:11:00 <bradfa> xxiao, does it work?
Nov 15 13:11:04 <av500> xxiao: sure, paypal me the 50
Nov 15 13:11:11 <xxiao> maybe
Nov 15 13:11:22 <bradfa> xxiao, that's not really the best sales pitch :)
Nov 15 13:11:50 <xxiao> sigh, i actually hope omap stays forever
Nov 15 13:12:02 <bradfa> koen, I agree boardfile is too long, I cut about 2/3 of it out and my bones still work :)
Nov 15 13:12:03 <av500> stay a while, stay forever!
Nov 15 13:12:16 <bradfa> xxiao, sure, omap will still be around, so long as you don't sell your panda :)
Nov 15 13:12:23 <bradfa> then you can have it forever
Nov 15 13:12:44 <av500> no, police will go around and collect all omaps
Nov 15 13:12:58 <av500> sir, step away from that panda
Nov 15 13:12:58 * bradfa hides xM and panda... brb
Nov 15 13:13:51 <av500> they will find them through the MAC address!!
Nov 15 13:13:52 <mru> ah, found it: http://ioccc.org/1994/schnitzi.c
Nov 15 13:13:57 <bradfa> last time we heard of rumors of things going away Freescale was killing i.Mx.  that's not really happening, just i.Mx7, so maybe this whole omap thing will blow over too?
Nov 15 13:14:29 <mru> that bit of code works both horizontally and vertically
Nov 15 13:14:29 <xxiao> bradfa: imx6 is already in production so they can't really kill it
Nov 15 13:14:51 <bradfa> xxiao, they were firing all the engineers for i.Mx line was the rumor, so no new i.Mx was the fear
Nov 15 13:14:53 <xxiao> but it's going to wind down, and probably just focus on some niche market plus auto industry's infotainment
Nov 15 13:14:58 <av500> mru: I like this more: http://images6.fanpop.com/image/photos/32400000/ASCII-Art-Woman-ascii-art-32449956-636-529.gif
Nov 15 13:15:25 <bradfa> xxiao, but that rumor was not quite true
Nov 15 13:15:38 <av500> it was a rumor after all
Nov 15 13:15:43 <mru> av500: http://www.textfiles.com/artscene/asciiart/ASCIIPR0N/
Nov 15 13:15:47 <av500> yes
Nov 15 13:16:08 <bradfa> rumors are always true!
Nov 15 13:16:20 * bradfa goes to work debugging windows machine now...
Nov 15 13:16:23 * bradfa is sad
Nov 15 13:16:31 <av500> mru: damn, why does the browser show so many links as visited?
Nov 15 13:16:41 <mru> :)
Nov 15 13:19:19 <mdp> lol "all the i.mx engineers"
Nov 15 13:20:05 <xxiao> on imx, they killed imx7, no more push to consumer market that is, however they used A7/A15 license to do LS-1 and LS-2, i.e. arm-core with decent IO bandwidth
Nov 15 13:20:25 <mdp> have you guys noticed yet that TI and FSL have done the _same thing_?
Nov 15 13:20:43 <xxiao> because FSL's boss is from TI...
Nov 15 13:20:43 <mru> sold out to amazon?
Nov 15 13:20:47 <mdp> literally, exact steps
Nov 15 13:21:03 <mdp> exit the crappy mobile processor market
Nov 15 13:21:18 <mru> that market was getting overcrowded
Nov 15 13:21:26 <mdp> mru, bingo
Nov 15 13:21:42 <mru> let samsung and qc have it
Nov 15 13:21:56 <av500> andfruit company
Nov 15 13:22:03 <mru> or let them fight allwinner
Nov 15 13:22:08 <mdp> yep
Nov 15 13:22:22 <av500> damn, mdp is up, the productive part of my day is over....
Nov 15 13:22:27 <mdp> lol
Nov 15 13:22:35 <mru> the fruit company isn't really relevant in the soc market
Nov 15 13:22:41 <mdp> your worst nightmare has arrived
Nov 15 13:23:04 * av500 must not blink
Nov 15 13:24:26 <panto> hey mdp
Nov 15 13:24:35 <mdp> hello panto
Nov 15 13:24:56 * mru points a large camera flash at av500 and fires
Nov 15 13:25:33 <mdp> speaking of nightmares...I'm out of Tizen stickers
Nov 15 13:26:23 * av500 looks at his 5" stack
Nov 15 13:26:57 * mdp notes that av500 kindly translated that to amuricun
Nov 15 13:27:57 <av500> mdp: it was a made up value, the " added for extra drama
Nov 15 13:28:22 <mdp> dear lazychannel omap people, which omap/dm parts have a hdq/1wire peripheral
Nov 15 13:28:30 <av500> omap4
Nov 15 13:28:39 <av500> we even tried it I think
Nov 15 13:28:49 <mdp> does it work?
Nov 15 13:29:33 <av500> hdq did
Nov 15 13:29:43 <mdp> actually, I should just look in hwmod...that's my google for this question
Nov 15 13:30:05 <av500> except for the fact that there are no 1.8V HDQ parts...
Nov 15 13:30:10 <av500> and the HDQ pin is 1.8V
Nov 15 13:30:21 <mdp> of course
Nov 15 13:30:23 <av500> minor details
Nov 15 13:33:33 <mdp> hrm, apparently all 34xx/36xx
Nov 15 13:33:56 <mdp> and omap4, as you say
Nov 15 13:35:58 <yegorich> Hi! Anyone having audio (davinci-pcm) running on am335x?
Nov 15 13:36:09 <xxiao> arm7+3*1G(ethernet) + 2*PCIe2.0
Nov 15 13:36:15 <mdp> "It would be great if somebody would make a cross reference tool showing which parts have which IP blocks"
Nov 15 13:36:15 <av500> mdp: other SoCs might have it in the bitbang subsystem
Nov 15 13:36:22 <xxiao> arm15+5*1G(ethernet)+3*PCIe2.0
Nov 15 13:36:30 <mru> xxiao: please try to make sense
Nov 15 13:36:34 <mdp> av500, yeah, am335x has no h/w 1w
Nov 15 13:36:40 <mru> arm7+GbE does not make sense
Nov 15 13:36:52 <mdp> mru, marketecture
Nov 15 13:36:53 <xxiao> mru: not until you see freescale's DPAA
Nov 15 13:36:53 <av500> I did 1w 12ys back on an SH-1
Nov 15 13:36:59 <av500> mdp: you will manage :)
Nov 15 13:37:05 <mru> and there is no such thing as arm15
Nov 15 13:37:10 <xxiao> mru: which offloads pretty all IO from cpu
Nov 15 13:37:11 <mru> so what did you really mean?
Nov 15 13:37:23 <av500> cortex ARMv15
Nov 15 13:37:23 <xxiao> http://www.freescale.com/webapp/sps/site/taxonomy.jsp?code=QORIQ_LAYERSCAPE
Nov 15 13:37:53 <av500> ah, A7
Nov 15 13:38:03 <xxiao> mru: i was checking what FSL is going to do with its purchased A7/A15 license
Nov 15 13:38:09 <mdp> cavium is going the same way
Nov 15 13:38:13 <mru> a7 != arm7
Nov 15 13:38:13 <av500> that thing has no 3D, thus it cannot run android, thus it is doomed
Nov 15 13:38:19 <mdp> same as keystone
Nov 15 13:38:31 <mdp> av500, +1
Nov 15 13:38:43 <xxiao> shoud say A7, A15 instead
Nov 15 13:38:45 <av500> didn'T I read about android on servers recently?
Nov 15 13:38:48 <mdp> how can you run xbmc on it?
Nov 15 13:38:54 <av500> +1
Nov 15 13:39:01 <mru> av500: I've heard of headless android
Nov 15 13:39:06 <xxiao> xbmc is really for allwinner nowadays
Nov 15 13:39:19 <xxiao> i.e. $30 STB single unit
Nov 15 13:39:21 <mdp> winners choose allwinner
Nov 15 13:39:28 <mdp> everybody wins
Nov 15 13:39:34 <av500> mru: that was last year, Karim chopped it off
Nov 15 13:39:58 <av500> xxiao: $30? you are ripping us off!
Nov 15 13:40:09 <mdp> yes, Karim has made Android the new embedded Linux
Nov 15 13:40:30 <mdp> all other Linii cower in fear at the droids
Nov 15 13:40:34 <xxiao> karim is working on a new book
Nov 15 13:40:51 <xxiao> android will dominate everything in embedded world
Nov 15 13:40:57 <av500> yes
Nov 15 13:41:09 * xxiao is switching to android channel
Nov 15 13:41:22 <av500> xxiao: all the channel is android
Nov 15 13:41:59 <xxiao> shit, what happened, android master took 58 minutes to build now, used to be 32 minutes
Nov 15 13:42:58 <mdp> #tizenalltheandroids
Nov 15 13:43:11 <xxiao> was asked to get i-jetty on android so it can act as a json-server
Nov 15 13:43:13 * mdp watches av500's stack go to 1" high
Nov 15 13:43:49 <xxiao> i actually tried tizen, and webos, both suck
Nov 15 13:44:02 <xxiao> rm -rf tizenwebos
Nov 15 13:44:17 <mdp> xxiao: impossible
Nov 15 13:44:36 <av500> xxiao: you did it wrong
Nov 15 13:44:46 <mdp> there's a vibrant comm^H^H^H^Hworld around it
Nov 15 13:44:55 <av500> lol
Nov 15 13:45:01 <av500> xxiao: you are supposed to run bada on tizen
Nov 15 13:45:04 <xxiao> :)
Nov 15 13:45:19 <mru> tizen had nice booth babes at elc...
Nov 15 13:45:29 * mdp checks zazzle.com for addition tizen stockers
Nov 15 13:46:00 <mdp> mru, aka "bait"
Nov 15 13:46:44 <koen> mdp: omap3 has hdq as well
Nov 15 13:46:51 <koen> mdp: not sure about voltage
Nov 15 13:46:55 <mdp> koen, yep, see my note above
Nov 15 13:47:09 <mdp> my very poor in-house hw guy can fix that
Nov 15 13:47:15 <mdp> can haz chipz
Nov 15 13:47:16 <av500> koen: 1.8V of course
Nov 15 13:47:19 <bradfa> mdp, stop making sense! It's almost Friday
Nov 15 13:47:25 * koen didn't read complete backlog
Nov 15 13:47:37 <bradfa> mdp telling us we don't know all the things
Nov 15 13:47:37 <av500> koen: despite the 27"?
Nov 15 13:47:39 <mdp> hwmod is my cross reference tool
Nov 15 13:47:50 <koen> av500: on 12" right now
Nov 15 13:47:55 <bradfa> av500, koen has in landscape, can't read as much backlog as portrait
Nov 15 13:47:59 <mdp> bradfa, you know more the thingz that me
Nov 15 13:48:02 <koen> 1024x768 ftw
Nov 15 13:48:06 <bradfa> mdp, I know nothing!
Nov 15 13:48:08 <xxiao> webos build took 70GB, i thought OE was heavy, glad to find it out and rm -rf it
Nov 15 13:48:15 <bradfa> mdp, for fear the police take away my omaps!
Nov 15 13:48:20 <bradfa> I know nothing!
Nov 15 13:48:27 <mdp> bradfa, lies! you detected my bogus blog entry!
Nov 15 13:48:38 <bradfa> mdp, did you add my ticket close to the big board from yesterday?
Nov 15 13:48:41 <koen> bradfa: there will be genuine OMAPs in a few months produced in china
Nov 15 13:48:43 <mru> ti drops omaps, focuses on opamps?
Nov 15 13:48:53 <bradfa> with national?
Nov 15 13:48:56 <bradfa> ompamps
Nov 15 13:48:59 <mdp> bradfa, "TI would like to recall the message 'OMAP'"
Nov 15 13:49:01 <koen> seeing the word 'genuine' on chinese websites makes me think of the princess bride
Nov 15 13:49:13 <koen> "I don't think that word means what you think it means"
Nov 15 13:49:28 <mdp> bradfa, no worries, got you covered on the big board
Nov 15 13:49:33 <bradfa> mdp, +1
Nov 15 13:50:09 * bradfa wonders why evince froze trying to print 18 page pdf... goes to kick printer
Nov 15 13:50:16 <mdp> you guys are doing great, Boris is pleased with the help desk coverage despite our fixed headcount
Nov 15 13:50:32 <mdp> bradfa, I will *try* to get you a $10 gift card to the cafeteria for your hard work
Nov 15 13:50:39 <mdp> bone-uses are tight though
Nov 15 13:52:45 <bradfa> mdp, do you know when boris will loosen up our headcount restrictions?
Nov 15 13:53:05 * mdp points bradfa to latest TI PR ;)
Nov 15 13:53:19 * bradfa does a little dance in anticipation of $10 of beagle cafe food!
Nov 15 13:53:32 <bradfa> SilicaGel, you get your tiwi working yet?
Nov 15 13:53:48 <SilicaGel> sadly no
Nov 15 13:54:03 <SilicaGel> I did order a usb pc based logic analyzer that can decode SDIO though
Nov 15 13:54:05 <SilicaGel> it hasn't gotten here yet
Nov 15 13:54:17 <bradfa> which sdio ananlyer?
Nov 15 13:54:21 <SilicaGel> oh heck i don't know
Nov 15 13:54:23 <SilicaGel> let me look
Nov 15 13:54:27 <bradfa> how mcuh $$?
Nov 15 13:54:30 * bradfa can't type
Nov 15 13:54:38 <SilicaGel> like $300ish
Nov 15 13:54:42 <bradfa> ooo
Nov 15 13:54:53 <SilicaGel> It doesn't look like a super fantastic one but hopefully good enough
Nov 15 13:55:00 <SilicaGel> and I've needed something aruond the lab from time to time to do simple stuff
Nov 15 13:55:07 <SilicaGel> It can decode CAN as well (which i use extensively here)
Nov 15 13:55:19 <SilicaGel> I2C, has specific modules for I2C EEPROMS and stuff
Nov 15 13:55:34 <SilicaGel> MIl-STD_1554 ... and a bunch of other protocols
Nov 15 13:55:57 <bradfa> link?
Nov 15 13:56:29 <SilicaGel> oh
Nov 15 13:56:30 <SilicaGel> http://microcontrollershop.com/product_info.php?products_id=3371&osCsid=6tcan24pl0beud52kpfobf1kj7
Nov 15 13:56:31 <SilicaGel> found it
Nov 15 13:56:32 <SilicaGel> it's that one
Nov 15 13:57:04 <bradfa> patented waveform compression technology!
Nov 15 13:57:07 <SilicaGel> i hope it's not a piece of merde
Nov 15 13:58:02 <bradfa> if it isn't merde, let me know, looks interesting
Nov 15 13:58:09 <SilicaGel> sure will
Nov 15 13:58:15 * bradfa goes to kick printer again...
Nov 15 13:58:37 * mdp mumbles about OlS and sikgrok a bit
Nov 15 14:00:25 <woglinde> hija silicagel
Nov 15 14:00:31 <woglinde> hi bradfa
Nov 15 14:00:33 <SilicaGel> hi there
Nov 15 14:01:06 <mdp> SilicaGel: that set of protocol plugins looks very good
Nov 15 14:01:15 <bradfa> howdy woglinde
Nov 15 14:02:14 <SilicaGel> yeah, it doesn't look too bad.  I actually had trouble picking 30
Nov 15 14:02:16 <SilicaGel> I had a dream last night that I lost my job and had to go work at harris.  On my first day there I found something in my desk and I didn't know what it was, so I called security, and thye decided it was a bomb, so they evacuated the building.  But later we found out it was just a hard drive.  Then they told me it cost $1,000,000 in lost productivity to have to do the evacuation.  :(
Nov 15 14:02:36 <woglinde> lol
Nov 15 14:05:18 <bradfa> SilicaGel, harris would just use it as an excuse to lay people off, don't worry :)
Nov 15 14:05:46 <mdp> koen, my biggest concern re: the hdq/1w block is if it's brought out on either beagle/panda headers...sifting pinouts now
Nov 15 14:06:18 <mdp> looks like no on bboard
Nov 15 14:06:59 <koen> iirc it was considered "battery"
Nov 15 14:07:06 <koen> so not broken out
Nov 15 14:07:09 <mdp> ahh, but yes, on xM...thank you gcoley post from 2010
Nov 15 14:09:44 <mdp> of course the driver needs DT/pinctrl support, as usual
Nov 15 14:10:27 <koen> if you don't mind using 3 wires, use bitbang it with w1-gpio
Nov 15 14:11:07 <mdp> for my purposes, I have to use omap hdq on *something*
Nov 15 14:11:16 <mdp> dayjob stuff, nbd
Nov 15 14:12:02 <bradfa> is that mdp deciding not to bitbang?
Nov 15 14:12:15 * koen tries to remember the bug with hdq
Nov 15 14:12:24 <koen> ah
Nov 15 14:12:27 * bradfa climbs under desk for fear of world ending due to non-bitbanging
Nov 15 14:12:33 <koen> https://plus.google.com/100242854243155306943/posts/eZKLqmCsDMV
Nov 15 14:12:39 <koen> the bug was in w1-gpio
Nov 15 14:13:27 <mdp> aha, bitbang fail!
Nov 15 14:13:51 <koen> exactly!
Nov 15 14:14:17 <SilicaGel> 1 wire?
Nov 15 14:14:25 <koen> yes
Nov 15 14:14:27 <av500> yes
Nov 15 14:14:28 <mdp> yes
Nov 15 14:14:30 <koen> linux calls it w1
Nov 15 14:14:30 <SilicaGel> i'm working on this little xbee-pro project
Nov 15 14:14:34 <SilicaGel> and it can support 1wire
Nov 15 14:14:35 <mdp> no more no less
Nov 15 14:14:39 <mdp> there can only be one
Nov 15 14:14:51 <SilicaGel> but I dont' think it has any actual hardware support for it, so it must be bitbanging it.  Is that reliable?
Nov 15 14:14:58 <av500> there is 0 wire aka NFC
Nov 15 14:15:02 <SilicaGel> beacuse I notice that they sell these little 1wire to SPI, etc. chips, and theyr'e relative cheap
Nov 15 14:15:16 <mdp> I have NFC
Nov 15 14:15:22 <av500> I have OCD
Nov 15 14:15:30 <ogra-cb_> no clue ?
Nov 15 14:15:38 <SilicaGel> naming it after how many wires it has is a terrible idea.  Not only that, but the idea of renaming radio back to "wireless" is comletely stupid.
Nov 15 14:15:43 <SilicaGel> It's worse with 1wire, since it's not REALLY 1 wire
Nov 15 14:16:00 <av500> I call usb w2
Nov 15 14:16:08 <SilicaGel> why not w4
Nov 15 14:16:10 <xxiao> what's a better name for "wireless"?
Nov 15 14:16:18 <mdp> I call spi w3or4or5
Nov 15 14:16:19 <SilicaGel> "radio"
Nov 15 14:16:23 <av500> etherfull
Nov 15 14:16:24 <mdp> true story
Nov 15 14:16:27 <SilicaGel> haha
Nov 15 14:16:54 <av500> mdp: how do you call RS232 then?
Nov 15 14:17:03 <SilicaGel> Oh no, it's EIA-232 now
Nov 15 14:17:03 <mdp> "perfection"
Nov 15 14:17:04 <av500> w9orw25
Nov 15 14:17:36 <mdp> I'm probably one of 3 owners of an rs-232 cape in the world :P
Nov 15 14:17:36 <SilicaGel> bradfa: some parts of my board do work though.  The 3 port USB hub that's on the same cape works great.
Nov 15 14:17:50 <xxiao> i prefer wireless to cordless
Nov 15 14:18:18 <av500> i prefer hopeless over clueless
Nov 15 14:18:33 <SilicaGel> Is that a political statement?  :)
Nov 15 14:18:59 <mdp> I don't even know what to call quadspi
Nov 15 14:18:59 <av500> discussing politics with USians is a nono
Nov 15 14:19:20 <mdp> especially when they call it quad spi but then allow more data lines than the 4 in the name :(
Nov 15 14:19:43 * xxiao back to work to save fiscal cliff
Nov 15 14:19:46 <mdp> jcm would use spi as an example of why non-standardization is a plague on the land
Nov 15 14:19:52 <av500> mdp: with 8 lines, it should be ppi
Nov 15 14:19:57 <mdp> xxiao, jump!
Nov 15 14:20:19 <av500> xxiao: stop spending stuff on the 57%
Nov 15 14:20:22 <av500> 47%
Nov 15 14:20:23 <av500> 54%
Nov 15 14:20:28 <SilicaGel> Why are I2C barometers/hygrometers/light sensors so expensive
Nov 15 14:20:36 <av500> because they can
Nov 15 14:20:50 <mdp> the cartels tightly control the sensor mines
Nov 15 14:21:09 <av500> I worked in one, itsd gruelling
Nov 15 14:21:17 <xxiao> among 47% there are 8% are looking for jobs who did not have income, so it's 39%
Nov 15 14:21:29 <av500> xxiao: trying to distract us with facts?
Nov 15 14:21:35 <av500> how mean
Nov 15 14:21:45 <xxiao> sigh...
Nov 15 14:22:04 <mdp> we're allowing facts on fridays
Nov 15 14:22:11 <xxiao> my stock loss, givmeebak money obama!
Nov 15 14:22:32 * xxiao back to work to offset stock loss
Nov 15 14:22:38 <av500> mdp: I remember smuggling black sensors over the shannon barrier
Nov 15 14:22:48 <mdp> heh
Nov 15 14:22:59 <av500> wires all still bloody
Nov 15 14:23:00 <SilicaGel> yeah, a single channel I2C to 1 wire master is 66 cents
Nov 15 14:23:13 <SilicaGel> Is that likely to be more reliable than bit banging
Nov 15 14:23:15 <SilicaGel> I'm thinmking probably
Nov 15 14:23:23 <av500> use reliable bits
Nov 15 14:23:51 <SilicaGel> bits are inherintly unreliable, because they lack their knowledge is far too compartmentalized
Nov 15 14:24:03 <SilicaGel> wow, sentence structure
Nov 15 14:24:05 <av500> makes then perfect for a spy ring
Nov 15 14:24:13 <SilicaGel> yes it does
Nov 15 14:24:21 <SilicaGel> unless one of them is sleeping with their biographer
Nov 15 14:24:32 <av500> bitgrapher
Nov 15 14:24:37 <SilicaGel> right
Nov 15 14:25:27 <av500> I can imagine how its now in the US: "honey, somebody is writing my biography - You cheat!"
Nov 15 14:25:47 <xxiao> she means it, ALL IN
Nov 15 14:25:48 <SilicaGel> 8 independent 1-wire masters $3.77 ... why would you need 8, can't you hang a crapload of devices off of a single one
Nov 15 14:25:56 * av500 excuses himself to see his "biographer" for a while....
Nov 15 14:26:08 <av500> SilicaGel: troughput
Nov 15 14:26:21 <SilicaGel> oh right, it's fairly slow isn't it
Nov 15 14:26:33 <mdp> 16-ish kbps
Nov 15 14:26:43 * av500 puts a 64 core parallel 1wire master on kickstarter
Nov 15 14:26:55 <av500> with an A9 for housekeeping
Nov 15 14:26:56 <SilicaGel> oh man
Nov 15 14:27:05 <SilicaGel> this sounds like a pretty nice PRU application :-)
Nov 15 14:27:06 <av500> ok, tomorrow then
Nov 15 14:27:08 <mdp> I vaguely recall the protocol, but I was surprised the bitbanging works reliably
Nov 15 14:27:27 <av500> it worked for me and a DS2401 back then
Nov 15 14:27:31 <mdp> I thought there were some 5-10 uS resolution pulse requirements for the symbols
Nov 15 14:27:42 <SilicaGel> yeah.  On this xbee-programmable project I'm working on, just for fun, I'm trying to interface to some GE ColorEffects christmas lights and bit banging is proving to not work
Nov 15 14:27:43 <av500> 12mhz SH-1
Nov 15 14:27:45 <mdp> I mean, reliable in the linux context
Nov 15 14:27:57 <av500> SilicaGel: is that 1wire?
Nov 15 14:28:01 <av500> wasnt that custom bitshift?
Nov 15 14:28:02 <SilicaGel> These are christmas lights where every "bulb" is an individually addressible RGB LED
Nov 15 14:28:10 <av500> yes, saw the video :)
Nov 15 14:28:20 <SilicaGel> no its' not 1wire it's a custom thing that's ... vaguely manchester encoded in a way
Nov 15 14:28:30 <mdp> yeah, I like that hack too
Nov 15 14:28:31 <av500> http://www.youtube.com/watch?v=AySja69jvHM
Nov 15 14:28:34 <SilicaGel> Each bit has 3 quanta
Nov 15 14:28:40 <SilicaGel> one is a 'start'
Nov 15 14:28:48 <SilicaGel> the other two are either high-low or low-high (that's the mancester part)
Nov 15 14:28:50 <mdp> rgb444 color iirc
Nov 15 14:28:58 <SilicaGel> it's 666 actually (!!!!!!!!)
Nov 15 14:29:41 <mdp> there's another one of the same lights where they guy makes a 7x7 grid..much much more impressive.
Nov 15 14:29:41 <mdp> sec
Nov 15 14:29:49 <SilicaGel> the issue that I am having, really, is that the xbee-pro uC is 68HCS08 or SC08 or S08 or whatever the F they call it now
Nov 15 14:29:51 <av500> that link I gave
Nov 15 14:30:01 <SilicaGel> yeah i saw that one on hackaday, ti was neat
Nov 15 14:30:07 <SilicaGel> I managed to get two 50 bulb strands of this stuff
Nov 15 14:30:30 <av500> SilicaGel: still, this is not 1wire(tm), no?
Nov 15 14:30:31 <SilicaGel> I have an evil plan, do you watn to hear about it, it has to do with more christmas lights
Nov 15 14:30:31 <mdp> ok, yeah, his remounting and text display was a great demo
Nov 15 14:30:44 <SilicaGel> av500: correct, not 1 wire, i'm asking about 1 wire for some other craziness hehe
Nov 15 14:30:50 <av500> ok
Nov 15 14:31:01 <av500> we try to keep crazy compartmentalized here
Nov 15 14:31:04 <SilicaGel> I have been doing this stuff with these xbee modules just for the fun of it.  I was going to make an xbee cape, but somebody watching the cape registry scared me off.
Nov 15 14:32:09 <SilicaGel> so instead, I made a circuit board with four solid state relays and an xbee programmable module on it, plus a swithcing power supply ... for  a 4 channel lawn christmas light controller
Nov 15 14:32:33 <SilicaGel> my plan - if the PCBs come from china early enough - is to do a synchronized christmas light display across my entire neighborhood
Nov 15 14:32:41 <SilicaGel> 2 neighbors have said yes so far
Nov 15 14:32:45 <av500> out of?
Nov 15 14:32:52 <mdp> 732
Nov 15 14:33:29 * av500 asks Nate Silver for a poll
Nov 15 14:33:41 <SilicaGel> hm well it depends on how far I want to take it.  The whole continent would be what about ... well 350M for the U.S., add another 600 peopl ein Canada ... how many are still in mexico (or do they all count as california, arizona, texas now) ?
Nov 15 14:33:59 <SilicaGel> wow, now everybody hates me :)
Nov 15 14:34:23 <av500> nah
Nov 15 14:34:26 <SilicaGel> really, though, if I can do 6 houses total (mine plus 5 others) I'll be really happy
Nov 15 14:35:01 <SilicaGel> beyond that I'd have to obtain more solid state relays off the surplus market
Nov 15 14:36:35 <mdp> let's see, did I see a project that does something similar syncing over the inter-tubes?
Nov 15 14:36:42 <SilicaGel> probably
Nov 15 14:36:51 <SilicaGel> I can't possibly be the first person to contemplate this
Nov 15 14:36:58 <mdp> I was too lazy to click and read... -ETOOMANYPROJECTS
Nov 15 14:37:12 <SilicaGel> but I'm all about the zigbee now.  I wrote my own ZCL implementation in java and am having fun with it.
Nov 15 14:37:22 <mdp> not discouraging you, I think it's great...just thinking aloud if you could leverage that thing
Nov 15 14:37:50 <av500> mdp: maybe turn on/off S3 instances in sync
Nov 15 14:38:01 <av500> see if we can trip Amazon power supplies :)
Nov 15 14:38:13 <av500> everybody md5sum NOW
Nov 15 14:38:49 <mdp> http://www.cheerlights.com/
Nov 15 14:38:49 <SilicaGel> did you see you can how back up your s3 to amazon glacier and it will only cost you $100,000 if you have to restore it?
Nov 15 14:39:11 <av500> sure
Nov 15 14:39:32 <SilicaGel> But do they have a way to snapshot your SimpleDB to S3 yet?  Noooooooooooooooooooooooooo
Nov 15 14:39:36 <SilicaGel> beacuse SimpleDB is the bastard child
Nov 15 14:41:22 <mdp> SilicaGel: if you acquire cloudlights.com and document this idea on kickstart, you are 98% of the way to retirement
Nov 15 14:41:45 <SilicaGel> You think?
Nov 15 14:42:00 <mdp> SilicaGel: people will fund anything that well promoted on kickstarter
Nov 15 14:42:17 <SilicaGel> yeah but it turns out you have to convince a small oligarchy that your project is worth it before they'll even try
Nov 15 14:42:20 <SilicaGel> cloudlights.com is taken
Nov 15 14:42:21 <mdp> I see no reason you can't attract millions with an idea that's better than most there
Nov 15 14:42:23 <SilicaGel> cloudlights.net is available
Nov 15 14:42:29 <mdp> everything is for sale
Nov 15 14:42:46 <SilicaGel> it costs $770
Nov 15 14:43:00 <mdp> marketing is everything
Nov 15 14:43:05 <mdp> you want to be a cloud company
Nov 15 14:43:10 <SilicaGel> I do
Nov 15 14:43:26 <mdp> I almost died when I sat down in the wrong keynote at ELC
Nov 15 14:43:30 <SilicaGel> beacuse dumb people think if it has the word "cloud" in it, it will never fail
Nov 15 14:43:39 <mdp> very nearly didn't make it out of the cloud buzzword linuxcon keynote :(
Nov 15 14:43:49 <mdp> it's like tiny daggers piercing me
Nov 15 14:44:05 <SilicaGel> cloudlights.net is available
Nov 15 14:45:54 <mdp> gotta be .com
Nov 15 14:46:46 <av500> ha: http://beaglecloud.com/
Nov 15 14:47:10 <SilicaGel> it's really not a bad idea
Nov 15 14:47:10 <alan_o> mdp: I was sitting in that keynote waiting for it to be over. Really irritated that I missed mlocke's.
Nov 15 14:47:15 <SilicaGel> I'd need a wifi version
Nov 15 14:47:21 <SilicaGel> with a few SCRs on it for a few channels
Nov 15 14:47:27 <av500> alan_o: mlocke's was not stellar imho
Nov 15 14:47:27 <SilicaGel> it's have to be pretty cheap
Nov 15 14:47:28 <xxiao> SilicaGel: where did you get that info? galcier will take a few hours to retrieval but not that expensive i think
Nov 15 14:47:43 <mdp> alan_o, didn't you see panto and I stand up in the middle up near the front and walk out through the zillion people?
Nov 15 14:48:01 <av500> mdp: you did not shout NVIDIA in keynote
Nov 15 14:48:02 <SilicaGel> The worst aprt about these projects, IMO, is that if you want to sell a real product you have to pay something like $25,000 per product for FCC Part 15, plus extra for intentional radiator
Nov 15 14:48:03 <alan_o> mdp: I guess not. I was in the middle, iirc.
Nov 15 14:48:18 <mdp> av500, hehe
Nov 15 14:49:00 <alan_o> mdp: When that guy said that they immediately started pushing Rackspace away, keeping them from having too much influence, when Rackspace was paying their paychecks, I wanted to hit the guy.
Nov 15 14:49:18 <alan_o> too much sense of entitlement with some OSS people.
Nov 15 14:49:29 <SilicaGel> xxiao: i'm exaggerating a bit, but the restore prices really were high.   1 cent per gigabyte
Nov 15 14:49:45 <av500> SilicaGel: glacier is for write-only data
Nov 15 14:49:48 <alan_o> SilicaGel: what Zigbee hardware are you using?
Nov 15 14:49:56 <SilicaGel> yeah i know
Nov 15 14:50:03 <xxiao> SilicaGel: that's $10 for 1TB, i think reasonable
Nov 15 14:50:06 <SilicaGel> I heard they store their data in dirt molecules that are buried below giants stadium
Nov 15 14:50:08 <SilicaGel> right under the G
Nov 15 14:51:11 <mdp> alan_o, that reminds me...is somebody working actively on getting cc2420 in mainline? wondering if it even has a caring owner in 802,154-land
Nov 15 14:51:13 <SilicaGel> alan_o: i'm using XBee Programmable modules, they're like $30.  They are the ones that have the extra microcontroller on them that you can put your own software into.  The dev environment is basically eclipse, it's pretty neat.
Nov 15 14:51:38 <alan_o> SilicaGel: cool. messing with those is on my list at some point
Nov 15 14:51:57 <alan_o> mdp: Not yet. I'm happy to do it if some TI person will send me some samples, *cough*
Nov 15 14:52:02 <SilicaGel> They are pretty fun.
Nov 15 14:52:22 <SilicaGel> I've said this before, though: the biggest problem with zigbee is that their licensing model is retarded
Nov 15 14:52:23 <mdp> alan_o, you need a real TI contact then ;)
Nov 15 14:52:49 <alan_o> mdp: Not the calculator division?
Nov 15 14:52:54 <SilicaGel> they should have said the protocol is free, and made their money from the semiconductor manufacturers.
Nov 15 14:53:39 <mdp> alan_o, btw, coincidentally, after having been on a few customer things that actually used zigbee..I *just* saw something go by where a real customer is using 6LoWPAN..pretty cool
Nov 15 14:53:40 <alan_o> SilicaGel: yeah, indeed, that's why there's not going to be in-kernel Zigbee.
Nov 15 14:53:42 <SilicaGel> so cloudlights ... the backend that coordinates all this stuff, I'm thinking XMPP
Nov 15 14:53:51 <mdp> the other stuff had to be zigbee, of course, what energy profile related
Nov 15 14:53:56 <mdp> s/what/was/
Nov 15 14:54:09 <SilicaGel> yeah
Nov 15 14:54:20 <xxiao> SilicaGel: http://www.synapse-wireless.com/
Nov 15 14:54:23 <SilicaGel> the ZCL model isn't that bad for all sorts of things though.  I kind of like it.
Nov 15 14:54:50 <xxiao> SilicaGel: this is a very interesting zigbee module, they somehow got a python env
Nov 15 14:55:05 <xxiao> very good comparing to xbee series
Nov 15 14:55:48 <SilicaGel> yeah, that looks cool
Nov 15 14:55:56 <SilicaGel> I will feel bad bankrupting them with my superior idea!
Nov 15 14:56:17 <alan_o> My slides and resources from ELC 802.15.4 and 6LoWPAN presentation are up: http://www.signal11.us/oss/elce2012/
Nov 15 14:56:24 <xxiao> SilicaGel: you write some little python script, then upload it to the module, all in a few KB, and it works reliably
Nov 15 14:56:34 <alan_o> mdp:  (so you don't have to set that alarm)
Nov 15 14:56:36 <koen> alan_o: is the  TI CC2531 supported?
Nov 15 14:56:43 <xxiao> they got a tiny python env in each module
Nov 15 14:57:03 <alan_o> koen: I don't know that one. Is that one of the ones with a microcontroller built in?
Nov 15 14:58:01 <xxiao> alan_o: was the ELC-Europe content online yet?
Nov 15 14:58:13 <SilicaGel> huh.
Nov 15 14:58:19 <mdp> koen, 2530 is what is in the *cough* Nest thermostat sparkfun teardown
Nov 15 14:58:21 <SilicaGel> ow ow ow ow this chocolate milk was so cold it gave me an ice cream headache
Nov 15 14:58:56 <alan_o> xxiao: I haven't seen videos yet, but elinux has some of the slides. What I put up was mostly about the resources and exact steps.
Nov 15 14:59:11 <alan_o> xxiao: I think elinux has whatever slides were sent in before the conf.
Nov 15 14:59:24 <xxiao> alan_o: i hate zigbee's license terms...what about 6LoWPAN?
Nov 15 14:59:42 <mdp> alan_o, 2531 is a 2530 with usb interface
Nov 15 14:59:46 <alan_o> xxiao: It's IETF, no royalty and free
Nov 15 15:00:02 <alan_o> mdp: oooh, that sounds fun actually. I'm going to check that out
Nov 15 15:00:21 <SilicaGel> the zigbee license term compatibility with linux etc. problem is what bothers me
Nov 15 15:00:40 <SilicaGel> if making a commercial product, then $1300/yr or whatever it is to join the zigbee alliance and be able to use their specs/protocols isnt' that bad
Nov 15 15:00:43 <xxiao> alan_o: what's your take on zigbee profile? say i could not runa HA and SE profile together
Nov 15 15:00:46 <xxiao> which sucks
Nov 15 15:00:48 <koen> alan_o: not sure, it's what is in the intel ELC-E giveaway :)
Nov 15 15:00:58 <koen> alan_o: shows up as ACM according to docs
Nov 15 15:00:59 <SilicaGel> You can run HA and SE together sure
Nov 15 15:01:12 <alan_o> koen: I didn't get a giveaway....
Nov 15 15:01:17 <mdp> alan_o, they both have a uC in them though so yeah
Nov 15 15:01:21 <alan_o> koen: All I got was a shirt
Nov 15 15:01:28 <SilicaGel> Every message in ZCL includes a profile id and a cluster id
Nov 15 15:01:51 <mdp> alan_o, hope you like 8051s
Nov 15 15:02:07 <bradfa> mdp, doesn't everyone love 8051s?
Nov 15 15:02:17 <SilicaGel> I'm pretty sure ZDO lets you send back a discovery that says you support as many profiles as you wish
Nov 15 15:02:18 <xxiao> SilicaGel: have you tried in practice
Nov 15 15:02:20 <bradfa> 8051 remoteproc cape, in the works!
Nov 15 15:02:24 <SilicaGel> let me check the spec
Nov 15 15:02:27 <mdp> I have a PL/M-80 compiler for DOS around here somewhere if you need it
Nov 15 15:02:28 <alan_o> SilicaGel: I don't know much of anything about the zigbee protocol. All that stuff is for meshing if I understand right..
Nov 15 15:02:33 <alan_o> mdp: you know I do!!!
Nov 15 15:02:55 <SilicaGel> oh, well it's a layered protocol.  There's all the routing and all that crap but on top of that is a sort of application layer that's object based
Nov 15 15:03:02 <xxiao> SilicaGel: i'm pretty sure zigbee zdo/zcl/whatever does not support running multiple profile in one device
Nov 15 15:03:23 <alan_o> what do you mean by multiple profiles?
Nov 15 15:03:27 <SilicaGel> objects have attributes, in the simplest case the attributes have simple times like unsigneds, strings, floats, bit fields, etc
Nov 15 15:03:31 <xxiao> i always thought zigbee asso is good at marketing, sucks at tech insights
Nov 15 15:03:46 <SilicaGel> ok so
Nov 15 15:04:13 <SilicaGel> a cluster is like an object.  A profile is a grouping of clusters for a particular cohesive purpose
Nov 15 15:04:18 <SilicaGel> so there's a "profile" for home automation
Nov 15 15:04:37 <SilicaGel> within that HA profile there are 'clsuters" defined for things like "Light control"
Nov 15 15:04:54 <xxiao> SilicaGel: that's all correct
Nov 15 15:05:02 <xxiao> under one HA you're fine with various clusters
Nov 15 15:05:03 <mdp> xxiao: no, many zigbee devices do multiple profiles
Nov 15 15:05:03 <SilicaGel> that particular cluster has a fixed set of attributes and operations.  You can turn it on, you can turn it off, you can ask its current state, various other things.
Nov 15 15:05:11 <xxiao> mdp: show me one :)
Nov 15 15:05:14 <SilicaGel> But each individual message includes the profile ID and the cluster ID
Nov 15 15:05:23 <SilicaGel> so why couldn't a device send and respond to mesasges from two different profiles
Nov 15 15:05:32 <xxiao> mdp: i had to swap in/out HA/SE profiles for different demos last year
Nov 15 15:05:50 <mdp> xxiao, I didn't say they are on the market..but several do both smart energy and home automation profile for obvious reasons
Nov 15 15:06:17 <SilicaGel> perhaps that's a limitation of the stack you are using?
Nov 15 15:06:33 <xxiao> mdp: yeah, those are "hacks", not 'standard' zigbee way
Nov 15 15:06:34 <mdp> a lot of low-end crap s/w and mcu implementations have limitations
Nov 15 15:06:39 <mdp> what SilicaGel said
Nov 15 15:06:47 <xxiao> and zigbee is all about 'standard' so they work together
Nov 15 15:06:58 <SilicaGel> "  A device or system can also support multiple profiles, for example, maybe both the ZigBee Home Automation profile AND the ZigBee Smart Energy profile."
Nov 15 15:06:58 <xxiao> those hack will not work with anything else but their own siblings
Nov 15 15:07:07 <SilicaGel> yeah but ZDO is all about discovery
Nov 15 15:07:10 <mdp> zigbee is all "vendor code" so naturally it will be of poor quality
Nov 15 15:07:15 <SilicaGel> it's about you being able to query a device and ask it about its capabilities
Nov 15 15:07:32 <SilicaGel> so all that device has to do is to respond "I support both the HA and the SE profile!" and it would work
Nov 15 15:08:10 <mdp> koen, *evil* ZAP vendor tree ;)
Nov 15 15:08:18 <xxiao> SilicaGel: when you look at HA and SE stacks, they're totally different, the zigbee stack itself differs in profiles
Nov 15 15:08:38 <SilicaGel> I'm not buying what you're telling me
Nov 15 15:08:43 <xxiao> so absolutely no way to co-exist, unless you have lots of memory that runs both in parallel
Nov 15 15:08:52 <SilicaGel> HA and SE is only the very top of the "stack" ... it's above ZCL
Nov 15 15:09:08 <SilicaGel> Everything underneath, all the routing, addressing, ZDO, all that is identical regardless of profile
Nov 15 15:09:34 <SilicaGel> so what you're saying is kind of like saying you can have a linux server that listens on port 80 for http, or 23 for ftp, but not both at the same time
Nov 15 15:09:50 <SilicaGel> Explain to me why you think I'm wrong?
Nov 15 15:10:18 <xxiao> SilicaGel: no, it's more like you have apache listening on 80 while lighttpd on 443
Nov 15 15:10:28 <SilicaGel> ok, what's wrong with that
Nov 15 15:10:36 <SilicaGel> except maybe your microcontroller doesn't have enough memory :D
Nov 15 15:10:37 <xxiao> not enough RAM
Nov 15 15:10:50 <SilicaGel> idunno man
Nov 15 15:11:03 <mdp> we can sell you more RAM
Nov 15 15:11:09 <SilicaGel> once you can parse a ZCL message, you can parse ANY zcl message
Nov 15 15:11:14 <mdp> and oparms
Nov 15 15:11:17 <mdp> ugh
Nov 15 15:11:21 <mdp> opamps ;)
Nov 15 15:11:57 <SilicaGel> I can see a practical limitation in ZDO, in that if you have a ridiculously large set of profiles/clsuters that you support you could end up with a ZDO reponse that's too big for you to formulate
Nov 15 15:11:59 <SilicaGel> maybe that's the problem
Nov 15 15:13:03 <SilicaGel> the zigbee application layer though, it's actually pretty simple.  Every field has a one byte type identifier associated with it, and there are maybe 10 operations you can do on any cluster
Nov 15 15:13:20 <SilicaGel> the spec makes it seem a lot more complicated than it really is.
Nov 15 15:14:17 <mdp> as I mentioned, it's all implementation...managing endpoints efficiently is pretty critical
Nov 15 15:14:47 <mdp> but in any case, zigbee's legal issues make it mostly uninteresting
Nov 15 15:14:55 <SilicaGel> :( :( :(
Nov 15 15:15:19 <av500> but you run the stack on the cc2530
Nov 15 15:15:23 <av500> no legal issues
Nov 15 15:15:26 <av500> so easy
Nov 15 15:15:54 <mdp> yes, if one of the soc units
Nov 15 15:16:57 <mdp> actually, there's two versions...one is hybrid and the app layer is in linux userspace..if you are using linux
Nov 15 15:17:09 <mdp> specific to TI's vendor code, I mean
Nov 15 15:18:00 <SilicaGel> wtf there is a project on kickstarter to build a z180 cpu board?  and it's 101% funded?
Nov 15 15:18:07 <SilicaGel> what is this, 1980?
Nov 15 15:18:24 <panto> we wish!
Nov 15 15:18:33 <panto> we'd be much younger then
Nov 15 15:19:13 <alan_o> mdp: come on man..... if they can do it with a z180, you've got to be able to get funding for a 6502-duino.
Nov 15 15:19:38 <av500> what about a 4004 cluster?
Nov 15 15:19:42 <av500> 4004 of them
Nov 15 15:19:50 <dm8tbr> beowulf?
Nov 15 15:19:51 <mdp> SilicaGel: I told you you're idea had merit!
Nov 15 15:19:55 <mdp> beowulf!
Nov 15 15:19:59 <mdp> s/you're/your/
Nov 15 15:20:01 <av500> might be in fact *all* of them by now
Nov 15 15:20:06 <mru> where is my beaglewolf?
Nov 15 15:20:50 <SilicaGel> wtf kickstarter approved a wireless board for the raspberry pi?  but when i proposed an xbee board for beaglebone they said "we don't fund that type of project" ?
Nov 15 15:20:50 <alan_o> did you see that cluster of RPi's? People don't have enough to do.
Nov 15 15:20:50 <SilicaGel> wtf
Nov 15 15:20:55 <SilicaGel> i think I hate these people
Nov 15 15:21:12 <mru> SilicaGel: "that kind of project" == "not using rpi"
Nov 15 15:21:21 <alan_o> ^^
Nov 15 15:21:21 <SilicaGel> what are they communists?
Nov 15 15:21:29 <mru> no, idiots
Nov 15 15:21:36 <mru> well, one is a subset of the other
Nov 15 15:21:44 <SilicaGel> kind of :-D
Nov 15 15:21:46 <alan_o> Don't you guys know, RPi is the future
Nov 15 15:22:12 <alan_o> That and blinking multi-colored USB lights. Kickstarter people are visionaries, not concerned about the past.
Nov 15 15:22:16 <mru> note, communist _leaders_ are in fact not communists at all
Nov 15 15:22:16 <alan_o> (except the z180 board0
Nov 15 15:22:23 <mru> they are just assholes
Nov 15 15:22:27 <alan_o> +1
Nov 15 15:22:33 <SilicaGel> omg
Nov 15 15:22:41 <SilicaGel> they funded
Nov 15 15:22:47 <SilicaGel> a usb device with a single RGB LED in it
Nov 15 15:22:50 <SilicaGel> for $132,317
Nov 15 15:22:51 <SilicaGel> That's it
Nov 15 15:22:51 <alan_o> :)
Nov 15 15:22:56 <SilicaGel> it's one RGB LED that you can control via USB
Nov 15 15:22:56 <alan_o> That's the one
Nov 15 15:22:58 <SilicaGel> W.T.F.
Nov 15 15:22:59 <panto> wtf? :)
Nov 15 15:23:01 * mru looks at his baconboard...
Nov 15 15:23:01 <Crofton|work> I used to be angry all the time
Nov 15 15:23:09 <koen> SilicaGel: so contact CCO, they are happy to help with capes
Nov 15 15:23:23 <SilicaGel> my capes are a failure!
Nov 15 15:23:25 * SilicaGel cries
Nov 15 15:23:25 <Crofton|work> koen, doesn't that mean talk to you
Nov 15 15:23:49 <SilicaGel> Actually, the OBD-II one works pretty well.  Except for the 6-axis accelerometer.  Which is probably a footprint error.
Nov 15 15:24:10 <Crofton|work> is there a web page for the obdII one?
Nov 15 15:24:16 <mdp> koen, contact you, right?
Nov 15 15:24:44 <SilicaGel> no, but if somebody wanted to license the design from us and sell it I'm sure we'd be interested
Nov 15 15:25:03 <SilicaGel> but I doubt enough people would be interested in that for it to be worthwhile
Nov 15 15:25:04 <mdp> SilicaGel: that device you mentioned...makes me think of that old "diamond app" on the appstore
Nov 15 15:25:10 <Crofton|work> what does it do?
Nov 15 15:25:13 <koen> Crofton|work: I'm still the sw guy
Nov 15 15:25:15 <SilicaGel> especially sinc eit mostly matters for pre-2008 vehicles
Nov 15 15:25:23 <Crofton|work> I am thinking of a few friends of mine who like cars
Nov 15 15:25:30 <Crofton|work> ah
Nov 15 15:25:33 <SilicaGel> It implements OBD-II for the four legacy protocols J1850-VPW, J1850-PWM, ISO9141-2, and uh
Nov 15 15:25:36 * SilicaGel thinks
Nov 15 15:25:43 <SilicaGel> whatever the othe rone is
Nov 15 15:25:46 <Crofton|work> what do modern cars use
Nov 15 15:25:48 <SilicaGel> and, for 2008 and beyond vehicles, CAN
Nov 15 15:25:51 <SilicaGel> yeah
Nov 15 15:26:10 <mdp> Crofton|work: with good s/w you can do "butt dynos" and such
Nov 15 15:26:14 <SilicaGel> post 2008 it's OBD over CAN using this .. I forgot the number, some ISO standard defines the messages and the tarnsport protocol for multi-packet messages
Nov 15 15:26:56 <SilicaGel> we managed to find one vehicle of each type of bus among our employees, and they all work :-D
Nov 15 15:27:12 <mdp> vagcom for the vw/audis is a good !foss example of what you can do
Nov 15 15:27:49 <SilicaGel> yeah.  it's also the worst name of any vehicle adapter I have ever heard.
Nov 15 15:31:48 <xxiao> SilicaGel: is CAN using the same interface as OBD-II? the connector
Nov 15 15:36:01 <alan_o> xxiao: CAN spec doesn't specify a connector. Higher-level protocols built on CAN presumably do (I'm only familiar with the lower layer).
Nov 15 15:37:01 <SilicaGel> yes they do, there's an SAE standard with a parallel ISO standard for OBD-II which says which pins on that connetor are for CANH and CANL
Nov 15 15:37:10 <SilicaGel> For trucks, there's a standard connector specified by SAE-J1939
Nov 15 15:37:26 <SilicaGel> so what you said is accurate
Nov 15 15:37:36 <SilicaGel> CAN itself doesn't, but the applications of CAN do.
Nov 15 15:37:40 <koen> alan_o: there's an "informal" standard for can using db9 and 10pin headers
Nov 15 15:37:47 <SilicaGel> xxiao: yeah.  For American vehicles at least.
Nov 15 15:37:54 <SilicaGel> that's actually a formalized pinout too haha
Nov 15 15:38:06 <xxiao> good to know that
Nov 15 15:38:16 <SilicaGel> it's formalized by an industrial automation protocol named CanOPEN ... everyone pretty much standardizes on their pinout for test tools, other applications etc
Nov 15 15:38:37 <xxiao> was checking obdii about 4 years back, CAN was not mentioned then, and only one chip-vendor we found for obd-ii devices
Nov 15 15:38:56 <SilicaGel> ah. Well for US vehicles OBD over CAN became required as of model year 2008.  Some vendors adopted it much earlier
Nov 15 15:39:21 <alan_o> interesting. I've seen it on 9-pin, but didn't know there was an informal standard.
Nov 15 15:39:25 <koen> SilicaGel: by which you mean "pretty much every non-US vendor"
Nov 15 15:39:25 <SilicaGel> That's for passenger cars.  For heavy duty vehicles there's still a mix of SAE-J1939 (over CAN) and J1708 (which is not CAN)
Nov 15 15:39:52 <SilicaGel> koen: yeah. I'm not sure of the timing of them all.  But I suspect you're right
Nov 15 15:40:13 <SilicaGel> actually, I shouldn't have said suspect, I should have said "I trust you know what you're talking about, beacuse you usually do"  :)
Nov 15 15:40:27 <xxiao> anyway, android in the car will get that info for me now
Nov 15 15:40:33 <SilicaGel> YES
Nov 15 15:40:41 <SilicaGel> do you have Torque for android?  It's pretty awesome
Nov 15 15:41:32 <SilicaGel> I have it on my phone and my touchpad .. the full version is 5 bucks and it rocks.  I bought a relatively cheap bluetooth adapter to talk to it
Nov 15 15:41:32 <xxiao> need check that out
Nov 15 15:41:59 <SilicaGel> yeah, do.  i've tried a few OBD apps for android and torque is by far the most comprehensive I have tried
Nov 15 15:42:11 <SilicaGel> enough that I paid the $5 for it (and I'm a cheapskate!)
Nov 15 15:43:05 <xxiao> which obdii you're using with it
Nov 15 15:44:48 <av500> SilicaGel: which one?
Nov 15 15:44:52 <av500> the bt adaptor?
Nov 15 15:45:07 <av500> cheap chinese?
Nov 15 15:45:12 <SilicaGel> yes it is
Nov 15 15:45:17 <SilicaGel> i'm trying to remember, i bought it off of amazon
Nov 15 15:45:30 <SilicaGel> It's ELM based, and i was being cheap
Nov 15 15:45:38 <SilicaGel> one really wants one that's based on the STN1110 not ELM if at all possible
Nov 15 15:46:04 <xxiao> yes ELM, that rings a bell, is it still the _only_ obd chip vendor?
Nov 15 15:47:17 <xxiao> http://www.amazon.com/Soliport-Bluetooth-OBDII-Diagnostic-Scanner/dp/B004KL0I9I/ref=pd_bxgy_auto_text_y
Nov 15 15:47:29 <SilicaGel> ther'es the STN chip now too
Nov 15 15:47:55 <SilicaGel> The STN is compatible with the ELM protocol but it adds some enhancements that some programs like Torque can take advantage of, that give it fairly significantly better speed
Nov 15 15:48:50 <SilicaGel> i'm downloading my amazon ordre history so i can tell you what one i have but like I said
Nov 15 15:48:52 <SilicaGel> if you get an STN one
Nov 15 15:48:53 <xxiao> http://www.obdsol.com/stn1110/
Nov 15 15:48:57 <SilicaGel> they're field upgradable
Nov 15 15:49:01 <SilicaGel> and faster
Nov 15 15:49:03 <SilicaGel> mine is neither
Nov 15 15:49:57 <xxiao> it's good to see competitions there
Nov 15 15:50:13 <xxiao> STN appears to be 1/2 cheaper and better on the paper
Nov 15 15:50:25 <SilicaGel> Elandpower Bluetooth ELM 327 Scanner OBDII PC Car Diagnostic
Nov 15 15:51:30 <xxiao> that's really cool, i'm buying one
Nov 15 15:52:26 <SilicaGel> wow man
Nov 15 15:52:30 <SilicaGel> look at this one  http://www.amazon.com/Bluetooth-Supper-Compatible-Andriod-2-white/dp/B009NPB214/ref=sr_1_fkmr2_1?ie=UTF8&qid=1352994680&sr=8-1-fkmr2&keywords=Elandpower+Bluetooth+ELM+327+Scanner+OBDII+PC+Car+Diagnostic
Nov 15 15:52:32 <SilicaGel> look hwo tiny that is
Nov 15 15:54:08 <xxiao> the connector looks too big
Nov 15 15:54:18 <xxiao> that might be STN, as it says ELM-compatible
Nov 15 15:55:05 <SilicaGel> yeah, I don't know.  I wouldn't expect much fo rquality out of that one.  I just like the size.  But look, they can't even spell Android right.
Nov 15 15:55:13 <Crofton|work> can you power the bone from the obd2 or other sensor connector?
Nov 15 15:55:25 <SilicaGel> You can with my OBD-II cape :-D
Nov 15 15:55:37 <SilicaGel> It has a 12/24VDC to 5VDC converter on it.
Nov 15 15:55:56 <SilicaGel> the only thing to keep in mind is on most vehicles the power on the OBD is fused battery, meaning it's always on
Nov 15 15:56:11 <SilicaGel> which is one of the reasons I really coudl use having deep sleep / hibernate work correctly on the beaglebone
Nov 15 15:56:46 <SilicaGel> of course a clock would be nice too, but I solved that problem beacuse the cape I'm talking about also has a GPS module on it, and I put a supercapacitor on the GPS module to let it keep time even when the bone is off.
Nov 15 15:57:52 <koen> SilicaGel: do you store almanac and ephimeris as well?
Nov 15 15:58:19 <SilicaGel> yeah the gps module does that
Nov 15 15:58:42 <koen> neat
Nov 15 16:12:03 <Lee_> Hello all, I installed opencv on BB-xM running Angstrom, after that, I used this command to build a simple project:  g++  -Wall -I /usr/include -L /usr/lib `pkg-config --cflags opencv` -o contours contours.c `pkg-config --libs opencv` But I got this ERROR: /usr/lib/gcc/arm-angstrom-linux-gnueabi/4.5.4/../../../../arm-angstrom-linux-gnueabi/bin/ld: cannot find -lopencv_highgui
Nov 15 16:12:51 <Lee_> I still found lopencv_highgui in the /usr/lin
Nov 15 16:12:59 <Lee_> I still found lopencv_highgui in the /usr/lib
Nov 15 16:16:55 <av500> nobody names my ware :(
Nov 15 16:17:31 <Russ> its a box
Nov 15 16:17:51 <Russ> hopefully for some kind of alcohol
Nov 15 16:20:56 * SilicaGel puts baby in a corner
Nov 15 16:21:24 <av500> Russ: no for alcohol
Nov 15 16:21:31 <av500> not*
Nov 15 16:23:51 <mru> it's the inside of something
Nov 15 16:24:07 <Russ> and it's empty
Nov 15 16:24:35 <mru> a plastic case with faux metal finish
Nov 15 16:26:56 <mru> something has been screwed to those two posts on either side of the white disc
Nov 15 16:28:12 <mru> speaking of which, appears to be some kind of external button or lever
Nov 15 16:29:26 <SilicaGel> pull the red one
Nov 15 16:29:29 <SilicaGel> not the green one
Nov 15 16:30:00 <SilicaGel> dammit i screwed it up. it's red and blue.
Nov 15 16:30:21 <mru> "not green" still holds
Nov 15 16:31:09 <av500> extrnal button yes
Nov 15 16:31:24 <av500> and led indicator too
Nov 15 16:32:32 <mru> av500: did you use that essential tool from this morning to undo those screws?
Nov 15 16:34:22 <av500> yes
Nov 15 16:34:27 <Russ> I'm actually wondering if its aluminum and not plastic
Nov 15 16:34:30 <av500> yes
Nov 15 16:34:46 <mru> definitely plastic
Nov 15 16:35:19 <mru> aluminium wouldn't have those circular marks from the mould
Nov 15 16:35:39 <Russ> aluminum can be molded too
Nov 15 16:35:42 <av500> I can assure you its aluminum
Nov 15 16:36:07 <mru> fake plastic, wtf
Nov 15 16:37:52 <Russ> hey NishanthMenon
Nov 15 16:40:00 <Russ> the logo on the front is blue and round, which makes me think hp
Nov 15 16:41:03 <mru> what, you can see the front?
Nov 15 16:41:09 * mru turns monitor around
Nov 15 16:41:10 <woglinde> opencv again
Nov 15 16:41:14 <mru> I'll be damned!
Nov 15 16:41:16 <Russ> you are looking at the back of the front
Nov 15 16:41:30 <Russ> you don't put a logo and button on the bottom
Nov 15 16:41:52 <Lee_> av500: I installed these packets to use for opencv:
Nov 15 16:41:52 <Lee_> opkg install opencv opencv-dev opencv-samples libopencv-imgproc-dev
Nov 15 16:42:09 <Lee_> is it enough av500 ? Thanks
Nov 15 16:42:19 <KotH> n'abend
Nov 15 16:42:27 <mru> Russ: yes, obviously it's the front
Nov 15 16:42:41 <woglinde> lee there are a lot more dev packages
Nov 15 16:42:46 <KotH> so... no real reasons not to use a state machine?
Nov 15 16:42:48 <woglinde> opkg list opencv
Nov 15 16:42:52 <woglinde> ups
Nov 15 16:42:57 <woglinde> opkg list | grep opnecv
Nov 15 16:43:08 <woglinde> koth right
Nov 15 16:43:08 <Lee_> thanks wog!
Nov 15 16:43:12 <mru> KotH: things should be made as simple as possible, but no simpler
Nov 15 16:43:21 <KotH> mru: that's obvious
Nov 15 16:43:26 <mru> state machines usually fail that test
Nov 15 16:43:36 <mru> not always, but most of the time
Nov 15 16:43:39 <KotH> mru: i dont see how
Nov 15 16:43:57 <mru> KotH: you're not thinking stupid enough
Nov 15 16:44:07 <KotH> mru: it could be that i havent seen too many bad examples of fsm
Nov 15 16:44:27 <woglinde> state machine is state of the art
Nov 15 16:44:29 <KotH> mru: sorry, my psychology hat cant do programming :)
Nov 15 16:44:36 <mru> some people just love to write _everything_ as a state machine
Nov 15 16:45:03 <Lee_> woglinde: I tried: opkg list | grep opnecv .But nothing appear!
Nov 15 16:45:06 <KotH> well... in VHDL, everything is a state machine ;)
Nov 15 16:45:23 <woglinde> lee check spelling
Nov 15 16:45:23 <mru> that's because physical circuits always have a state
Nov 15 16:45:27 <woglinde> dont be stupid
Nov 15 16:45:39 <KotH> mru: so, assuming that i'm not stupid, and that i think that a fsm is the right choice, then chances are high that i'm right?
Nov 15 16:45:53 <mru> what are you building?
Nov 15 16:46:03 <Lee_> yes
Nov 15 16:46:09 <KotH> remember the acceleration sensor project i talked about?
Nov 15 16:46:10 <av500> Russ: yes
Nov 15 16:46:15 <KotH> i've to do its software design now
Nov 15 16:47:29 <KotH> basically, i've different types of events that trigger actions which can be interleaved
Nov 15 16:47:54 <KotH> and first thing that came into my mind was doing a fsm with 3-4 states
Nov 15 16:48:25 <KotH> to model the longer living "event states"
Nov 15 16:49:08 <av500> show us the UML
Nov 15 16:49:30 <KotH> uh.. you dont want that...
Nov 15 16:49:37 <woglinde> av500 its top secret
Nov 15 16:49:54 <KotH> $project_lead has done an uml which is so convoluted it is nearly worthless
Nov 15 16:52:07 <woglinde> koth than show us
Nov 15 16:52:16 <mru> so I heard you like state machines: http://pastie.org/5383234
Nov 15 16:53:41 <KotH> ^^'
Nov 15 16:54:14 <KotH> somehow, i feel like being trolled...
Nov 15 16:54:34 <alan_o> That'd be a great song parody
Nov 15 16:54:40 <mru> mdp: ^^ he wants to be trolled, I think it's your turn
Nov 15 16:54:44 <alan_o> "feel like be-ing trolled by you..."
Nov 15 16:54:52 <KotH> ^^;
Nov 15 16:56:41 <av500> mru: that can be optimized be removing a few breaks ;)
Nov 15 16:56:58 <mru> av500: please stand by for version 2
Nov 15 16:57:29 <av500> also it does not use gettext!!!
Nov 15 16:57:50 <av500> also, it does not depend on dbus to run
Nov 15 16:57:51 <woglinde> hihi
Nov 15 16:57:54 <av500> its a fail
Nov 15 16:57:54 <woglinde> and glib
Nov 15 16:58:02 <av500> yes
Nov 15 16:58:24 <av500> also, state should be a char[]
Nov 15 16:58:40 <mru> and strcmp?
Nov 15 16:58:40 <av500> to allow runtime overloading of state names
Nov 15 16:58:45 <av500> yes
Nov 15 17:00:54 <woglinde> but i like glib and dbus
Nov 15 17:02:03 <alan_o> av500: gettext, that's for foreigners :)
Nov 15 17:04:03 <av500> woglinde: please leave now
Nov 15 17:04:26 <woglinde> av500 and systemd
Nov 15 17:08:00 <mdp> mru, can you imagine...there's actually a lot of distractions at work today? :)
Nov 15 17:08:07 <mdp> darn work distractions
Nov 15 17:08:27 <woglinde> mdp first people cleaning uo the cubes?
Nov 15 17:08:53 <mdp> can't see from here
Nov 15 17:22:45 <mru> as promised: http://pastie.org/5383336
Nov 15 17:23:27 <av500> I see a turing machine coming
Nov 15 17:26:32 <av500> mru: there is redundancy in the individual strokes of the letters to be exploited too
Nov 15 17:26:43 <av500> H E and L share too many common lines
Nov 15 17:35:46 <mru> better? http://pastie.org/5383388
Nov 15 17:39:14 <KotH> mru: not convoluted enough
Nov 15 17:39:19 <KotH> mru: no use of function pointers
Nov 15 17:39:25 <KotH> mru: no nested state machines
Nov 15 17:39:39 <KotH> mru: not enough work done in the IDLE state
Nov 15 17:40:44 <mru> this is just a trivial example
Nov 15 18:46:51 <prpplague> koen: ping
Nov 15 18:49:37 <prpplague> koen: please ping me asap
Nov 15 18:52:15 <mranostay> hi prpplague
Nov 15 18:52:22 <prpplague> mranostay: hey bud
Nov 15 18:52:37 <prpplague> mranostay: i sent an email to TimB about getting your submission for ELC
Nov 15 18:53:46 <mru> prpplague: any ideas what buzzwords they'll be looking for in proposals for the next one?
Nov 15 18:54:16 <prpplague> mru: not as of yet
Nov 15 18:54:19 <panto> koen, prpplague just pinged you
Nov 15 18:54:50 <mru> do you know who'll be choosing?
Nov 15 18:55:12 <av500> mru: how bout embedded state machines
Nov 15 18:55:22 <mru> bitbanged ones of course
Nov 15 18:55:29 <prpplague> mru: the program committee has not been selected as of yet
Nov 15 18:55:40 <prpplague> mru: i assume TimB will be the program chair
Nov 15 18:56:10 <av500> will there be twister again?
Nov 15 18:56:29 <mranostay> mru: is it like the google ones? where your resume doesn't get accepted without the correct number? :)
Nov 15 18:57:29 <mru> there are usually some buzzwords that seem to vastly increase the chances of being accepted
Nov 15 18:57:41 <av500> add android
Nov 15 19:02:39 <mdp> CFP went out?
Nov 15 19:03:01 <mru> not that I know of
Nov 15 19:03:02 <prpplague> mdp: no not yet
Nov 15 19:03:08 <mru> but better to be prepared
Nov 15 19:03:20 <mru> we all know it'll go out eventually
Nov 15 19:03:27 <mdp> let's see..what does tbird want to see...
Nov 15 19:03:41 <mdp> he likes printk timings..maybe something about that
Nov 15 19:03:59 <prpplague> mdp: hehe
Nov 15 19:04:05 <mdp> somebody got that ;)
Nov 15 19:04:29 <prpplague> mdp: actually i have also recommended to tbird that your re-present your 6502 one
Nov 15 19:04:40 <mdp> really?
Nov 15 19:05:14 <mdp> I'll tell you what...if that has any interest, I am readily able to show quite a bit more on top of elc-e's work
Nov 15 19:05:16 <prpplague> between mdp and mranostay , those were some of the most discussed presentations
Nov 15 19:05:28 <woglinde> intressting
Nov 15 19:05:31 <mru> mdp: yeah, can we have it again please, but with more bitbanging
Nov 15 19:05:43 <av500> mdp: it has to run a c64 game
Nov 15 19:05:48 <prpplague> mdp: *cough* you do need to cut it down to 40 minutes
Nov 15 19:05:48 <dm8tbr> bangarang!
Nov 15 19:05:57 <av500> mdp: can the pru access linux memory?
Nov 15 19:05:57 <mdp> "I'm giving 'er all the bitbanging she's got captain!"
Nov 15 19:06:01 <av500> I mean, 64k?
Nov 15 19:06:03 <mdp> av500, indeedio
Nov 15 19:06:11 <av500> so, go, make it happenZ
Nov 15 19:06:16 <mdp> av500, you lose determinism
Nov 15 19:06:24 <av500> pah
Nov 15 19:06:32 <av500> that does not deter me
Nov 15 19:06:54 <mdp> av500, I've been researching some of the out of production "MMU" parts that have been used with 65xxx in the past too
Nov 15 19:07:03 <CareBear\> mdp : what's the worst case waitstate in the pru for accessing main memory?
Nov 15 19:07:23 <mdp> CareBear\: can't give a number..would take some analysis
Nov 15 19:07:47 <mru> there's probably _some_ upper bound
Nov 15 19:07:47 <CareBear\> mdp : I'm thinking the 1MHz 6502 is so slow that the delays might not be a problem
Nov 15 19:07:49 <mdp> tbh, I'd need to learn a bit more about interconnect priority management on our parts
Nov 15 19:08:04 <mdp> I used to have to program our plb priorities on the ppc parts I worked on
Nov 15 19:08:09 <av500> mdp: use the 1ghz A8 to predict the next access :)
Nov 15 19:08:35 <mdp> CareBear\: lyren brown ran into issues hosting his data in DRAM...so I tried to KISS to start
Nov 15 19:08:37 <av500> mdp: erm wait, the 6502 is static, no?
Nov 15 19:08:41 <av500> you can slow the clock?=
Nov 15 19:08:57 <mru> av500: the a8 could run the same code on a 6502 emulator so it would know exactly what to expect
Nov 15 19:08:59 <mdp> so, there's a few options...I can give the 6502 a full 28k of ram we have
Nov 15 19:09:05 <av500> mru: YES
Nov 15 19:09:18 <mdp> av500, yeah the WDC part can run DC->14MHz
Nov 15 19:09:26 <av500> and the SGX could compare both outputs and act as a failsafe
Nov 15 19:09:38 <av500> we get five nines
Nov 15 19:09:46 <mdp> well...we start getting close to slew rate issues when breadboarding so again..I KISS
Nov 15 19:09:59 <mru> av500: or at least nine fives
Nov 15 19:10:10 <av500> mdp: talk WDC into makeing a POP part :)
Nov 15 19:10:33 <CareBear\> now that must be a big market
Nov 15 19:10:47 <mdp> CareBear\: but one thing I'm considering is bankswitch access to DRAM buffers driven by the second PRU to handle virtio framework buffers..I think we kinda touched on that before
Nov 15 19:11:00 <mdp> av500, I might need a job at WDC soon ;)
Nov 15 19:11:21 <mdp> "Bill, I know there's few 6502 hackers, do you have any openings?"
Nov 15 19:11:25 <mdp> hehe
Nov 15 19:11:45 <av500> http://www.baltissen.org/newhtm/74ls612.htm
Nov 15 19:11:48 <av500> mdp: this?
Nov 15 19:11:57 <CareBear\> mdp : we did, but your ideas for making virtio in linux less blunt seem like the better way to me
Nov 15 19:12:02 <mdp> but I think having a real OS running on it might be more intriguing to folks
Nov 15 19:12:27 <mdp> CareBear\: yeah, I feel like it should/could be fixed in that sense
Nov 15 19:12:49 <CareBear\> ack
Nov 15 19:12:55 <mdp> we can also use the mailbox IP, btw
Nov 15 19:13:05 <CareBear\> also bringing remoteproc out to userspace makes things much more interesting
Nov 15 19:13:06 <mdp> proxied by the PRU
Nov 15 19:13:25 <dm8tbr> mailbox? kinky, which number do you need to dial?
Nov 15 19:13:33 <mdp> just call me maybe
Nov 15 19:13:50 <av500> dm8tbr: 555-FIVENINES
Nov 15 19:14:11 <av500> additional charges might apply
Nov 15 19:14:11 * dm8tbr puts av500 and mdp in a 3-way call and whistles a modem sync
Nov 15 19:14:27 * av500 adds 5400Hz tone
Nov 15 19:15:04 <mdp> you can bitbang those 9s
Nov 15 19:15:14 <av500> nein!
Nov 15 19:15:18 <mdp> heh
Nov 15 19:16:12 * av500 teleports home
Nov 15 19:16:17 <mdp> *poof*
Nov 15 19:16:28 <mdp> that's because you don't have a flying car yet
Nov 15 19:16:31 <av500> no
Nov 15 19:16:39 <av500> I have public teleport only
Nov 15 19:16:40 <mdp> .de is sooo behind
Nov 15 19:16:42 <av500> its slow
Nov 15 19:16:48 <av500> upload takes ages
Nov 15 19:16:59 <av500> and you get scrambled with other people occasionally
Nov 15 19:17:06 <av500> takes a while to defrag again
Nov 15 19:17:22 <mdp> I wouldn't get in that fool contraption
Nov 15 19:17:34 <mdp> at least go by hovercraft
Nov 15 19:17:50 <mru> mdp: I would, but my hovercraft is full of eels
Nov 15 19:17:53 <av500> its full of ells
Nov 15 19:17:55 <av500> eels
Nov 15 19:17:58 <av500> gah
Nov 15 19:18:06 <mru> av500: getting slow?
Nov 15 19:18:10 <mdp> heels?
Nov 15 19:18:10 <av500> public internet lag
Nov 15 19:18:45 <mdp> uh oh, I lost my beagles
Nov 15 19:19:21 <av500> PETA freed them
Nov 15 19:19:39 <mdp> gah, they were being good...escaped the office for a moment
Nov 15 19:20:16 <mdp> curious, is anybody around here making use of the eQEP on beaglebone?
Nov 15 19:27:06 <mranostay> mdp: aren't you glad you aren't on the panda team? those would be hard to keep as pets
Nov 15 19:28:09 <mdp> yeah, I think it shows my true dedication to the cause too
Nov 15 19:28:28 <mranostay> didn't you have those before TI even? :)
Nov 15 19:28:48 <mdp> mranostay, somewhat better than keeping pets associated with your own Tiger Team ;)
Nov 15 19:28:56 <mdp> mranostay, sshhhhh
Nov 15 19:29:44 <mdp> mranostay, I bleed TI red...like a prpplague landing party TI redshirt
Nov 15 19:30:24 <mranostay> i don't think you want to be wearing a red shirt atm reading TI press releases :)
Nov 15 19:30:52 <prpplague> mdp: hehe
Nov 15 19:32:53 <elik> hey guys
Nov 15 19:33:12 <mdp> mranostay, I still have nightmares about being turned into a cuboctahedron a crushed to dust
Nov 15 19:33:23 <mdp> I don't wear red shirts...very dangerous
Nov 15 19:37:15 * mdp suspect eQEP is too esoteric for most people to care
Nov 15 19:48:48 <prpplague> koen: ping
Nov 15 20:04:37 <alan_o> koen: ping
Nov 15 20:09:08 <woglinde> *g*
Nov 15 20:12:13 <mdp> koen, ping
Nov 15 20:13:43 <mranostay> geez everyone wants koen
Nov 15 20:14:44 <alan_o> mranostay: ping
Nov 15 20:15:08 <clockman> Hi all, I have a beaglebone running stock version of angstrom that was working perfectly, but now when i try to start it up, it boots and i can log into it for a few seconds (either serial or eth) but then it shuts down. it just says "Sending SIGTERM to remaining processes..." then continues with shutdown about 5-10 seconds after it fully boots. Any clue how i can boot into single so I can troubleshoot
Nov 15 20:15:08 <clockman> this?
Nov 15 20:17:42 <alan_o> clockman: did you try re-flashing your SD?
Nov 15 20:17:47 <alan_o> just as a sanity check
Nov 15 20:18:08 <clockman> i have an extra one at home that i'll do it to. I have some code on this sd card that i dont' want to lose
Nov 15 20:18:09 <alan_o> I don't know about single user. You'll have to check systemd docs.
Nov 15 20:18:27 <alan_o> I could tell you for sysvinit :)
Nov 15 20:19:30 <mranostay> what is wrong with init script we gots?
Nov 15 20:19:34 * mranostay pokes mdp 
Nov 15 20:19:42 * KotH lols at soekris
Nov 15 20:20:40 <dm8tbr> they do ARM now?
Nov 15 20:20:44 <KotH> after selling their newest product for over one and a half year, and many customers complaining aboug poor performance... now they figure out that they had a bug in the bios thus running the cpu at the slowest possible speed...always
Nov 15 20:21:08 <mdp> mranostay, oh, *ahem* ;)
Nov 15 20:21:16 <mranostay> KotH: nice
Nov 15 20:21:22 <KotH> dm8tbr: whatever you do, do not consider buying a soekris
Nov 15 20:21:30 <mru> I'd expect nothing less from a bios
Nov 15 20:21:42 <mru> they're _supposed_ to be buggy, I'm sure of it
Nov 15 20:21:50 <mdp> a trained bios-booting monkey
Nov 15 20:21:53 <mranostay> never touch anything with a bios attached
Nov 15 20:21:58 <dm8tbr> KotH: but that's sort of user fail too, isn't it? people could just check the cpu parameters from the actual OS
Nov 15 20:22:08 <mru> there's no other way the people writing them could still have their jobs
Nov 15 20:22:16 <KotH> dm8tbr: i've yet to see another non-chinese company with such poor designs (actually allmost all chinese designs are better) and with such a horrible user support (aka calling user who report bugs "stupid" and "full of shit")
Nov 15 20:22:25 <mranostay> mru: horrible job though :)
Nov 15 20:22:56 <dm8tbr> mranostay: accoring to todays plussing they could always switch to windows kernel driver development
Nov 15 20:22:57 <KotH> dm8tbr: apearantly, the whole acpi system was broken, thus linux reported the wrong values
Nov 15 20:24:05 <mranostay> dm8tbr: i rather flip burgers
Nov 15 20:24:33 <dm8tbr> :)
Nov 15 20:24:36 <KotH> mranostay: try cutting kebab instead.. better pay, better food
Nov 15 20:24:57 <KotH> mranostay: and you learn turkish in the process ;)
Nov 15 20:25:19 * dm8tbr doesn't like the ground meat kebab that they do around here :/
Nov 15 20:25:35 <KotH> dm8tbr: yeah.. same here in .ch
Nov 15 20:25:50 <dm8tbr> neither the fact that they like to keep the cut kebab for hours in a hot box...
Nov 15 20:25:52 <KotH> only good kebab place left on earth, is the original iskender kebab place in burca
Nov 15 20:25:56 <KotH> bursa*
Nov 15 20:29:32 <panto> hey, don't make me hungry, I just had dinner
Nov 15 20:29:57 <mranostay> panto: second dinner!
Nov 15 20:30:23 <KotH> panto: there is alsways space for a good iskender kebab
Nov 15 20:30:43 <KotH> panto: and if you take the 1.5 sized portion, you'll have enough
Nov 15 20:30:54 <KotH> :)
Nov 15 20:31:21 <KotH> http://upload.wikimedia.org/wikipedia/commons/a/ae/Bursa13.JPG
Nov 15 20:31:43 <mranostay> damn you KotH
Nov 15 20:31:51 <dm8tbr> KotH: I once or twice had 'rambo' size...
Nov 15 20:32:07 <_av500_> hmm, kebab
Nov 15 20:32:21 <KotH> though, something is wrong with that picture... it doesnt look how it is supposed to look like
Nov 15 20:32:24 <dm8tbr> heh, of course av500 shows up once we start talking tasty
Nov 15 20:32:26 <CareBear\> what do you suggest for doing ~100kSPS AD acquisition on a bone?
Nov 15 20:32:44 * _av500_ shouts bitbang and runs
Nov 15 20:32:52 <CareBear\> need only a few seconds worth of data
Nov 15 20:33:21 <xanium4333> CareBear\: would 96KHz be enough?
Nov 15 20:33:30 <CareBear\> xanium4333 : maybe.. what do you have in mind?
Nov 15 20:33:59 <xanium4333> CareBear\: an audio cape, 96KHz sampling at 16/24bits, AC only
Nov 15 20:34:20 <CareBear\> mhm
Nov 15 20:34:46 <CareBear\> except the builtin ADC should be enough..
Nov 15 20:34:47 <xanium4333> CareBear\: I've actually just fabbed a 4-channel 192KHz design, but I'm beginning (right now) to start the driver
Nov 15 20:35:12 <KotH> CareBear\: what resolution?
Nov 15 20:35:21 <KotH> 32bit?
Nov 15 20:35:21 <CareBear\> KotH : 12 bits is more than enough
Nov 15 20:35:31 <xanium4333> up to 24 bits theoretical, using a audio CODEC
Nov 15 20:35:32 <KotH> lol
Nov 15 20:35:59 <xanium4333> if less than 12 is enough
Nov 15 20:36:07 <KotH> CareBear\: 12bit@100ksps is quite a good adc.. if you really want to get out 12bit and not just 8
Nov 15 20:36:24 <panto> CareBear\, ADC driver doesn't support streaming capture IIRC
Nov 15 20:36:25 <KotH> CareBear\: look at the ones from analog and linear
Nov 15 20:36:44 <KotH> CareBear\: there are versions with spi that go up to that range iirc
Nov 15 20:36:44 <CareBear\> KotH : yes, I know. I expect the builtin will be plenty noisy at 12 bit, but 8 would also be OK
Nov 15 20:36:46 <xanium4333> Isn't there a whole framework fron analog for generic streaming sampling
Nov 15 20:36:47 <panto> the audio codec does though
Nov 15 20:37:03 <KotH> CareBear\: never use build in adcs unless you need more than 8bit
Nov 15 20:37:06 <panto> xanium4333, industrial i/o
Nov 15 20:37:12 <panto> but needs driver support
Nov 15 20:37:12 <KotH> er.. wehn you need more than 8bits
Nov 15 20:37:13 <xanium4333> panto: aaah yes
Nov 15 20:37:20 <mranostay> mru: which is worse build system or bios work? :)
Nov 15 20:37:21 <_av500_> one can implement a dual slope ADC with 2 GPIOs .... just saying...
Nov 15 20:37:26 <mranostay> kinda a semi-poll
Nov 15 20:37:26 <panto> plus is really really immature at this point
Nov 15 20:37:28 <CareBear\> KotH : yes sure, I've used ADCs before, but not the one on the bone
Nov 15 20:37:41 <KotH> _av500_: been there done that, got a sampling rate in the seconds ;)
Nov 15 20:37:48 <mdp> panto, shush..it's purrrfect
Nov 15 20:37:55 <_av500_> KotH: yep, for battery level :)
Nov 15 20:38:02 <panto> for varying values of 'purrfect' :)
Nov 15 20:38:05 <KotH> _av500_: lol
Nov 15 20:38:11 <_av500_> KotH: we did that
Nov 15 20:38:16 <KotH> _av500_: was the uC with an integrated adc too expensive? ;)
Nov 15 20:38:17 <mdp> _av500_: we have a name for that type of gpio usage...
Nov 15 20:38:35 <mranostay> hehe
Nov 15 20:38:41 <mranostay> BITBANGING!
Nov 15 20:38:41 <CareBear\> um so what is the verdict?
Nov 15 20:38:42 <_av500_> bit massage?
Nov 15 20:38:45 <mdp> there it is!
Nov 15 20:38:53 <KotH> CareBear\: use an spi adc
Nov 15 20:39:08 <xanium4333> speaking of which, if I'm about to delve into the kernel to ass support for a new codec (and require the McASP interface to be working), which kernel should I be aiming for?
Nov 15 20:39:09 <KotH> CareBear\: afaik the spi framework should be able to support you enough to get to that datarate
Nov 15 20:39:15 <ds2> so that's where more bits come from?
Nov 15 20:39:28 <CareBear\> KotH : it seems an awful lot like crossing the river to get water
Nov 15 20:39:32 <xanium4333> 3.2 / 3.6? TI offical, or the github beagle/kernel beaglebone branches?
Nov 15 20:39:46 <KotH> CareBear\: well...
Nov 15 20:39:48 <mdp> 3.2
Nov 15 20:39:50 <xanium4333> there seams to be an awful lot of patches floating around
Nov 15 20:39:58 <mdp> mcasp..3.2
Nov 15 20:39:59 <KotH> CareBear\: you are not doing audio, hence cant use alsa
Nov 15 20:40:19 <KotH> CareBear\: but afaik there are drivers for adcs in the kernel.. but i never had a look at them
Nov 15 20:40:22 <mdp> for various definitions of a working mcasp driver
Nov 15 20:40:39 <KotH> CareBear\: but then, writing an adc driver is quick.. if you dont do the protocol work from user space
Nov 15 20:40:43 <xanium4333> mdp: in the end, 4-channel 96KHz, but I'll settle with 2-channel for now
Nov 15 20:41:09 <xanium4333> mdp: also recording, not playback
Nov 15 20:41:19 <mdp> xanium4333: I highly recommend you steer clear of 3.6+ mcasp until it settles down..unless you are interested in some work there
Nov 15 20:41:26 <mranostay> i hate build systems!! what is this step doing?
Nov 15 20:41:29 <mranostay> </RANT> :)
Nov 15 20:41:31 <CareBear\> KotH : reading the TRM I think the TSC ADC should be able to stream converted data with DMA
Nov 15 20:41:35 <KotH> CareBear\: for reference: most of our designs run on bare metal, and we have written drivers for a dozen adcs.. most of them are less then 100 loc
Nov 15 20:41:52 <mdp> xanium4333: might consult koen on a mcasp capture issue he's seen...even on 3.2...nfc if it's only beaglebone specific or not
Nov 15 20:42:08 <CareBear\> KotH : yes, bare metal is certainly an option
Nov 15 20:42:29 <xanium4333> mdp: I am interested, but mainly atm I need stability in McASP to test things. In the end I'll need 4-channel sampling, something I'm willing to work on
Nov 15 20:42:34 <mdp> mranostay, I made a concise build system for the b6502 codez
Nov 15 20:42:38 * ds2 hugs the nice and mature McBSP
Nov 15 20:42:45 * _av500_ likes bearmetal
Nov 15 20:42:49 <KotH> CareBear\: believe me, doing an adc driver using spi on linux isnt much harder
Nov 15 20:42:51 * mdp EOLs McBSP
Nov 15 20:43:01 <_av500_> time for McCSP
Nov 15 20:43:02 * mdp flings McASPs at ds2, prpplague-style
Nov 15 20:43:10 <ds2> no love for the cousins?
Nov 15 20:43:11 <_av500_> or even McDsp
Nov 15 20:43:13 <mdp> let's all have a McFarmer
Nov 15 20:43:22 * mdp pokes panto
Nov 15 20:43:27 <_av500_> or, as we say in omap land: McBook
Nov 15 20:43:41 <_av500_> McEbook even
Nov 15 20:43:43 <KotH> farmer stängel?
Nov 15 20:43:43 <CareBear\> KotH : but needs more hardware, where the internal peripheral should be able to do the job..
Nov 15 20:43:45 * mranostay flips through E-ELC photobook
Nov 15 20:43:53 <xanium4333> atm I'm trying to wade through all the kernels and patches that exist. There's git repos with branches that hold patches to branches of TI's kernels
Nov 15 20:43:55 <ds2> even the rats are leaving omap land!
Nov 15 20:43:57 <_av500_> mranostay: url?
Nov 15 20:44:16 <mranostay> http://www.flickr.com/photos/13825348@N03/sets/72157632014371035/
Nov 15 20:44:23 <KotH> CareBear\: what's the spec of the adc? in number of effective bits
Nov 15 20:44:30 <panto> mdp, ouch
Nov 15 20:45:03 <CareBear\> KotH : TRM says 12
Nov 15 20:45:04 <mdp> mranostay: 3 months until the next set of memories ;)
Nov 15 20:45:12 <KotH> CareBear\: under what conditions?
Nov 15 20:45:14 <CareBear\> KotH : as I said, I expect that to be a bit noisy
Nov 15 20:45:33 <mdp> xanium4333: it's craziness, indeed
Nov 15 20:45:45 <KotH> CareBear\: adn 12bit maximum bits or effective bits?
Nov 15 20:45:48 <xanium4333> also adding Angstrom into the whole mix, is making getting a relevant kernel + defconfig impossible
Nov 15 20:45:56 <xanium4333> mdp: any clues on how to break the ice?
Nov 15 20:46:23 <KotH> CareBear\: so far i've only seen uCs from analog where maximum bits = effective bits on the build in adc
Nov 15 20:46:35 <xanium4333> (I've just today finished fabbing a new cape design with a 4-channel CODEC attached, it's now time for kernel work :)
Nov 15 20:46:44 <mdp> xanium4333: I no longer work with 3.2 so I don't have a sane link handy for you
Nov 15 20:47:21 <xanium4333> mdp: but you don't recommend 3.6? Koen said something about only being able to record for 1 second on 3.2 (I think)
Nov 15 20:47:31 <mdp> ahh, ok, that's it
Nov 15 20:47:44 <mdp> well, there's no mcasp support in 3.6/3.7 yet
Nov 15 20:47:48 <CareBear\> KotH : I don't see a clear statement in the TRM
Nov 15 20:47:56 <mdp> so I can only recommend one that kinda/sorta works a little bit
Nov 15 20:48:13 <xanium4333> mdp: that'll be a start
Nov 15 20:48:14 <_av500_> mranostay: ha, I see the master seargent
Nov 15 20:48:19 <_av500_> +spellink
Nov 15 20:48:24 <mdp> you can use mcasp on mainline with am1808 if you get one of those
Nov 15 20:48:37 <mdp> I regression test that regularly for my work
Nov 15 20:48:37 <xanium4333> mdp: shoudl I be using the Ti PSP kernels or the github/beaglebone/kernel kernels?
Nov 15 20:48:43 <KotH> CareBear\: then assume that you lose 1 bit in general use and 2 bits at max speed
Nov 15 20:48:50 <ds2> isn't the mcasp the same as on the davinci's?
Nov 15 20:48:56 <ds2> like the 6446 and friends?
Nov 15 20:49:07 <KotH> CareBear\: does the bone have an internal or external voltage reference?
Nov 15 20:49:13 <CareBear\> KotH : external
Nov 15 20:49:15 <KotH> CareBear\: if it's internal, substract another bit
Nov 15 20:49:29 <xanium4333> KotH: resolution now -2 bits
Nov 15 20:49:37 <_av500_> and mdp talking to mr v4l2
Nov 15 20:49:40 <mranostay> _av500_: master seargent?
Nov 15 20:49:57 <KotH> xanium4333: learn to count
Nov 15 20:50:07 <KotH> xanium4333: and negative numbers dont exist
Nov 15 20:50:10 <xanium4333> KotH: was meant to be a joke
Nov 15 20:50:19 <mdp> xanium4333: I recommend koen's 3.2 kernel (which is derived from psp releases with additional features) if it has a feature you need OR for most other cases, the beagleboard-kernel on github either beaglebone-3.6 or 3.7 branch
Nov 15 20:50:27 <mdp> ds2, yes
Nov 15 20:50:40 <_av500_> and alan_o
Nov 15 20:50:45 <ds2> mdp: isn't that mainlined?
Nov 15 20:50:46 <_av500_> mranostay: prpplague
Nov 15 20:50:59 <mdp> ds2, yes, that am335x support is incomplete
Nov 15 20:50:59 <mranostay> heh i see me in a few
Nov 15 20:51:02 <xanium4333> mdp: okay thank you
Nov 15 20:51:26 <ds2> so the core code is there?
Nov 15 20:51:30 <mdp> I wish joelagnel was in here now, he's been messing with mcasp stuff on something close to mainline
Nov 15 20:51:31 <mranostay> a few from the party and couple of photobombing linus :P
Nov 15 20:51:32 <mdp> ds2, correct
Nov 15 20:51:39 <ds2> that shouldn't be too bad to fix up
Nov 15 20:52:15 <mdp> there's a number of patches submitted to address various integration issues 1) DT 2) pinctrl 3) sram->genalloc stuff for build 4) DTS not submitted I think 5) register offset are different
Nov 15 20:52:22 <mdp> and probably more I don't know of
Nov 15 20:52:46 <ds2> register offsets can be hacked in if you don't care about multiplatform kernels
Nov 15 20:52:49 <mdp> I do know that joelagnel was in a thread with the TI owner of the driver and couldn't get it working
Nov 15 20:52:59 <mdp> that may have changed
Nov 15 20:53:02 <KotH> http://www.flickr.com/photos/13825348@N03/8188811932/in/set-72157632014371035 <- i know these guys!
Nov 15 20:53:05 <panto> bed time
Nov 15 20:53:06 <ds2> DT is ...
Nov 15 20:53:18 <mdp> so my recommendation is that xanium4333 may experience great pain thinking it will work
Nov 15 20:53:34 <ds2> I mean, just change it in the code... not the right fix as that prevents a multichip kernel from working (i.e. a dual bone/6446 kernel)
Nov 15 20:53:35 <mdp> to my knowledge koen has no integrated a set of working patches in his 3.7 branch yet
Nov 15 20:53:58 <mdp> sure, I don't really care what approach people take to get working code ;)
Nov 15 20:54:05 <mdp> whatever it takes!
Nov 15 20:54:07 <mdp> just do it
Nov 15 20:54:24 <ds2> I am still stuck in am35x land :(
Nov 15 20:54:34 <mdp> isn't that EOLed? ;)
Nov 15 20:54:41 <_av500_> epic: http://www.flickr.com/photos/13825348@N03/8188811932/in/set-72157632014371035/
Nov 15 20:54:42 <ds2> i only wish....
Nov 15 20:54:53 <ds2> F'ing memory controller
Nov 15 20:55:07 <xanium4333> mdp: mmm seems that way, thanks for the warning.
Nov 15 20:55:18 <KotH> good night boys
Nov 15 20:55:29 <KotH> sleep tight and keep your lunch money close!
Nov 15 20:55:47 <_av500_> we are not in barcelona
Nov 15 20:55:49 <mranostay> _av500_: it was almost troll-less at the VIP party
Nov 15 20:56:04 <alan_o> _av500_: awesome!!
Nov 15 20:56:08 <mru> mranostay: told you so
Nov 15 20:57:02 <mranostay> mru: it was fun still
Nov 15 20:57:19 * dm8tbr was trolling the vip party
Nov 15 20:57:27 <mranostay> didn't see you there
Nov 15 20:57:49 <_av500_> KotH: http://www.flickr.com/photos/13825348@N03/8188777070/in/set-72157632014371035/
Nov 15 20:58:36 <_av500_> ha: http://www.flickr.com/photos/13825348@N03/8187695885/in/set-72157632014371035/
Nov 15 20:58:41 <_av500_> next time I will shave :)
Nov 15 21:00:25 <alan_o> _av500_: I just hope more people bathe next time :)
Nov 15 21:00:42 <alan_o> (not directed at anyone here)
Nov 15 21:01:15 <mru> alan_o: I didn't see rms there...
Nov 15 21:01:24 <alan_o> mranostay: sitting next to Linus, using OSX...
Nov 15 21:01:27 * alan_o shakes head
Nov 15 21:01:40 <alan_o> mru: hah! Never met the guy. Would like to.
Nov 15 21:01:50 <_av500_> no
Nov 15 21:01:51 <prpplague^2> _av500_: you ping me?
Nov 15 21:01:58 <mru> alan_o: you'd regret it
Nov 15 21:02:07 <_av500_> prpplague^2: u in elce pics
Nov 15 21:02:17 <_av500_> alan_o: ask koen about rms
Nov 15 21:02:41 <prpplague^2> _av500_: oh?
Nov 15 21:02:43 <ds2> is that glass on the camera F10000?
Nov 15 21:02:44 <prpplague^2> _av500_: url?
Nov 15 21:02:49 <mru> dear me: http://www.flickr.com/photos/13825348@N03/8187672971/in/set-72157632014371035/
Nov 15 21:02:49 <mranostay> alan_o: i could be running linux on my MBA
Nov 15 21:02:55 <mranostay> i'm not but just saying
Nov 15 21:03:04 <alan_o> mranostay: I know otherwise :)
Nov 15 21:03:14 <mranostay> but the public doesn't
Nov 15 21:03:37 <alan_o> I am on my MBP, but it was not easy to get it working well (without running both GPUs at once)
Nov 15 21:03:40 <mru> buying a mac to run linux on it is a very stupid thing to do
Nov 15 21:03:41 <_av500_> prpplague: scroll upwards
Nov 15 21:04:04 <dm8tbr> mru: just found that too, trollphoto :)
Nov 15 21:04:14 * prpplague looks
Nov 15 21:04:31 <alan_o> mru: I support 4 platforms with my OSS stuff. Besides, I like the hardware, and I got mine before Apple was being a major tool.
Nov 15 21:04:55 <_av500_> alan_o: yeah, that absolves you
Nov 15 21:04:58 <ds2> buying a mac is a very stupid thing to do
Nov 15 21:04:59 <mdp> xanium4333: it sounds like you might experience pain via either path...I only have experience on am1808, too many issues keeping me from helping debug the am335x mcasp stuff
Nov 15 21:05:22 <KotH> alan_o: you bought your mac in the 80s?
Nov 15 21:05:33 <prpplague> _av500_: hehe got me in picture wiping the sweat off my face, hehe
Nov 15 21:05:35 <mru> alan_o: a mac is more expensive than an equivalent pc and has more issues running linux
Nov 15 21:05:46 <mru> only a hipster would buy one
Nov 15 21:05:51 <alan_o> _av500_: not saying it absolves me.... I do like it though. It's a nice machine.
Nov 15 21:06:14 <_av500_> prpplague: well, I can see why :)
Nov 15 21:06:14 <dm8tbr> http://www.flickr.com/photos/13825348@N03/8188777070/in/set-72157632014371035/ is also nice
Nov 15 21:06:23 <alan_o> mru: then there were a lot of hipsters at ELC.
Nov 15 21:06:37 <_av500_> LetoThe2nd: ^^^^^
Nov 15 21:06:50 <_av500_> there was a Berlin hipster for sure
Nov 15 21:06:51 * mdp hears hipster and looks up
Nov 15 21:07:01 <_av500_> a systematic one
Nov 15 21:07:49 <alan_o> mru: I dispute the costing very much more money if you compare laptops with comparable specs, towards the top of Apple's release cycle. When I got my first one, it was on par with a Dell of the same specs, and 200 more than a lenovo with the same specs.
Nov 15 21:08:12 <_av500_> prpplague: sweat much? http://www.flickr.com/photos/13825348@N03/8188630764/in/set-72157632014371035/
Nov 15 21:10:21 <mranostay> MY EYES!
Nov 15 21:10:24 <dm8tbr> ah, that was the lady from redhat
Nov 15 21:10:48 <mranostay> really? couldn't tell from the fedora
Nov 15 21:11:15 <mranostay> so i've heard that is part of the employee welcome package
Nov 15 21:11:48 <mranostay> does the CEO get a purple one with a feather I wonder? :)
Nov 15 21:13:36 <_av500_> mranostay: you mean this one? https://plus.google.com/+JohnMueller/posts
Nov 15 21:15:40 <Crofton|work> no Yocot dev day photos
Nov 15 21:19:40 <_av500_> no photographer with 8 arms was around?
Nov 15 21:20:33 <mru> a yoctographer?
Nov 15 21:20:59 <_av500_> yo
Nov 15 21:22:57 <prpplague> _av500_: man that is one shinny bald head in that picture
Nov 15 21:23:27 <_av500_> wax on, wax off
Nov 15 21:23:49 <prpplague> _av500_: ugh, that picture  is a good reminder that i need to get back to going to the gym
Nov 15 21:24:41 * prpplague looks like he could be the Michelin Man
Nov 15 21:25:41 <ds2> so if ELC-E was in the neighboring country, you'd be a native?
Nov 15 21:28:26 <_av500_> elce in paris would be nice too
Nov 15 21:28:39 <_av500_> show people what a real mugging is like
Nov 15 21:28:50 <Russ> c'mon, head to moscow for that
Nov 15 21:28:54 <ds2> can versalles be rented out?
Nov 15 21:29:02 <ds2> versailes? (sp)
Nov 15 21:29:02 <_av500_> ds2: yes
Nov 15 21:29:10 <_av500_> versailles
Nov 15 21:29:12 <ds2> Oohh... do that forthe party
Nov 15 21:29:23 <Russ> the hall of mirrors would make a great rave
Nov 15 21:29:53 <_av500_> theres a nice restaurant with a view to the pallace gardesn at least
Nov 15 21:31:27 <mattzz> Hi & good evening! I am running angstrom w/ connmand and I want to execute a command when the wlan0 interface comes up. Can somebody please help me? (I need to switch off power management of my wifi dongle otherwise wifi will drop frequently…)
Nov 15 21:32:20 <mattzz> with old school /etc/network/interfaces I would use if-up.d but there is no such thing anymore ;)
Nov 15 21:33:01 <ds2> pickpocketd?
Nov 15 21:33:06 * djlewis and connman are not on good terms
Nov 15 21:33:23 * mattzz neither...
Nov 15 21:35:52 <_av500_> djlewis: you got conned
Nov 15 21:38:51 <_av500_> Russ: you gave up?
Nov 15 23:11:37 <Crofton|work> http://cheezburger.com/6761679360
Nov 15 23:14:22 <_av500_> yup
Nov 15 23:14:38 <_av500_> you too are wrong
Nov 15 23:17:09 <xanium4333> does there exist some sort of 'meta-beagle' layer for OE? that inlcudes the community kernel. I know about meta-ti, but thats doesn't include the community additions
Nov 15 23:17:28 <joel_> koen, how does one go about registering lcdc (using capebus stuff) without actually having a cape plugged in?
Nov 15 23:21:38 <alan_o> joel_: look for the override sections in the DT
Nov 15 23:21:55 <joel_> alan_o, ah got it, thanks
Nov 15 23:23:31 * mranostay watchs mru 
Nov 15 23:24:17 <denix> Crofton|work: yep, that's correct!
Nov 15 23:25:21 <denix> koen: not asleep yet? or are you a bot?
Nov 15 23:27:07 <_av500_> a boet
Nov 15 23:27:31 <_av500_> mranostay: I watched mru before it was cool to watch mru
Nov 15 23:27:58 <denix> trolls watch each other?
Nov 15 23:29:08 <thurbad> it's how they breed
Nov 15 23:30:14 <mranostay> whoa whoa
Nov 15 23:44:53 <elik> Hey guys, I'm working on the bring up of a board similar to BeagleBone. I've been reading on JTAG quite a bit but am still a bit confused. It seems TI has a non standard 20 pins connector for JTAG, so an adapter board is need. There seems to be a dongle to put on that new 14 connector to connect to the PC. I also read that BeagleBone has some of that on-board. Any guidance?
Nov 15 23:46:46 <mranostay> elik: you have to rework the bone to have a real CTI jtag
Nov 15 23:46:55 <CareBear\> elik : sure
Nov 15 23:47:37 <CareBear\> elik : did you look at the documentation?
Nov 15 23:47:55 <elik> mranostay: my board has the 20 pins from the chip coming out on an extension board
Nov 15 23:48:07 <elik> I saw this CTI JTAG thing, still a bit confused
Nov 15 23:48:20 <elik> My end goal is software debug of early stages (u-boot)
Nov 15 23:49:09 <elik> through openocd
Nov 15 23:49:18 <Russ> just use the on board ftdi chip
Nov 15 23:49:35 <mru> what's broken in u-boot?
Nov 15 23:50:54 <elik> nothing broken, it's just that our board is not exactlyt the same (am3359 based, but we have ddr3)
Nov 15 23:51:23 <elik> Russ: which chip would that be?
Nov 15 23:51:48 <joel_> jkridner1, almost getting there with the display
Nov 15 23:52:03 <Russ> the usb to serial chip on the bone
Nov 15 23:52:15 <joel_> HDMI driver is getting probed from DT, and I instantiated LCDC from capebus override
Nov 15 23:52:30 <Russ> http://elinux.org/Running_OpenOCD_on_Linux_with_the_Beaglebone
Nov 15 23:52:49 <joel_> the driver gets EDID/HPD events etc so HDMI side looks fine
Nov 15 23:53:03 <joel_> Have to check LCDC signals
Nov 15 23:53:05 <elik> Russ: how is that related to the serial port?
Nov 15 23:53:23 <elik> Russ: right, got that page
Nov 15 23:53:23 <Russ> the chip that connects to the am335x serial signals also connects to the jtag port
Nov 15 23:53:23 <joel_> bbt
Nov 15 23:54:33 <CareBear\> elik : the FT2232H chip has two USB interfaces, two serial ports. one is used for serial console, the other is used for JTAG, with the simple MPSSE built into the chip
Nov 15 23:55:13 <elik> CareBear\: ok, so that would be the two com port I see when I plug it in
Nov 15 23:55:43 <CareBear\> elik : openocd will detach one of the two
Nov 15 23:55:53 <CareBear\> elik : please make sure to use the ftdi/ interface config file
Nov 15 23:56:01 <CareBear\> elik : please make sure to use openocd git code and nothing else
Nov 15 23:56:27 <elik> CareBear\: the elinux guy seemed to be saying 0.5.0 in ubuntu was working fine
Nov 15 23:56:36 <elik> but I was thinking of going with 0.6.X anyway
Nov 15 23:57:57 <CareBear\> elik : only git has the ftdi driver
Nov 15 23:58:10 <CareBear\> I note however that there is no interface config for the beagles
Nov 16 00:00:26 <elik> CareBear\: elinux refers to this location on git: tcl/board/ti_beaglebone.cfg
Nov 16 00:01:28 <CareBear\> yes, and that file uses the old sucky crappy ft2232 driver
Nov 16 00:01:29 <elik> haven't looked at the hardware side quite yet
Nov 16 00:01:32 <CareBear\> I think
Nov 16 00:01:34 <CareBear\> or no?
Nov 16 00:01:48 <elik> might be, no quite there yet
Nov 16 00:05:14 <elik> I'm still a bit confused at what JTAG is exactly. I know it's a standard for hardware and software testing, but given there are so many specifics depending on the architecture and all, it looks like a very "weak" standard
Nov 16 00:05:38 <Russ> its very flexible :)
Nov 16 00:06:01 <CareBear\> elik : I gave a talk about this at ELC-E
Nov 16 00:06:35 <CareBear\> just last week
Nov 16 00:07:08 <elik> CareBear\: interesting. Is it somewhere on the net?
Nov 16 00:07:37 <CareBear\> elik : JTAG was originally for boundary scan; to test connectivity between BGA parts instead of using bed-of-nails probing
Nov 16 00:07:56 <CareBear\> the JTAG standard is low level
Nov 16 00:08:15 <CareBear\> it's just a simple state machine which muxes an unspecified number of shift registers
Nov 16 00:08:23 <elik> right
Nov 16 00:08:35 <elik> I knew about that use, for testing waffers, specifically
Nov 16 00:08:37 <CareBear\> talk not online I think
Nov 16 00:08:53 <CareBear\> wafers as well, yes
Nov 16 00:09:01 <Russ> debug registers can also be linked into the chain
Nov 16 00:09:06 <elik> CareBear\: I see the pdfs of the presentations
Nov 16 00:09:14 <elik> They say videos come online a month after
Nov 16 00:09:22 <elik> CareBear\:
Nov 16 00:09:28 <elik>  http://elinux.org/ELCE_Europe_2012_Presentations
Nov 16 00:10:22 <elik> The only beagle talk I can find is from Ranostay, and nothing on JTAG
Nov 16 00:10:51 <CareBear\> I didn't upload PDF
Nov 16 00:10:56 <elik> ah, ok
Nov 16 00:11:17 <elik> Wish I hadn't missed it, a lot of interesting topics
Nov 16 00:11:34 <elik> Stuge, on OpenOCD
Nov 16 00:11:38 <CareBear\> me
Nov 16 00:11:50 <elik> :)
Nov 16 00:12:10 <CareBear\> so yes, different registers
Nov 16 00:12:13 <CareBear\> next layer:
Nov 16 00:12:31 <elik> right
Nov 16 00:12:44 <CareBear\> the processor has many different things on the same die
Nov 16 00:12:59 <CareBear\> they are not neccessarily connected in a chain
Nov 16 00:13:08 <CareBear\> there is a router thingy
Nov 16 00:13:10 <CareBear\> icepick
Nov 16 00:13:24 <CareBear\> it's not documented in the latest manuals anymore
Nov 16 00:15:04 <CareBear\> shifting the right bits into the right registers magically extends the shift register
Nov 16 00:15:04 <elik> I've seen references to Icepick on elinux
Nov 16 00:15:43 <elik> Is icepick itself rather standard?
Nov 16 00:15:49 <CareBear\> not the least
Nov 16 00:16:04 <elik> so icepick is arm specific?
Nov 16 00:16:12 <CareBear\> no, TI processor specific
Nov 16 00:16:17 <elik> ok
Nov 16 00:18:07 <CareBear\> after magically stroking the icepick, it routes the ARM Cortex-A8 into the chain
Nov 16 00:18:16 <CareBear\> or the chain into the A8, if you prefer
Nov 16 00:18:26 <elik> ok
Nov 16 00:18:53 <elik> just to make sure, Icepick is TI specific, but other vendors have similar mechanism
Nov 16 00:20:10 <CareBear\> I'm not sure what other vendors do
Nov 16 00:20:49 <CareBear\> I'm sure some simply connect the things together, in a chain
Nov 16 00:21:39 <elik> I see
Nov 16 00:23:39 <CareBear\> so once the A8 is reachable, ARM tells how that works in the ADIv5 documentation (ARM Debug Interface v5)
Nov 16 00:24:04 <elik> ok
Nov 16 00:24:17 <joel_> elik, http://focus.ti.com/lit/ml/sprp603/sprp603.pdf
Nov 16 00:26:16 <elik> joel_: Unless I'm mistaken, BeagleBone's cpu is neither Davinci nor OMAP (beaglebone is omap)
Nov 16 00:28:00 <joel_> elik, true but I figured the IcePick discussion might be the same as there shared IPs
Nov 16 00:28:52 <joel_> elik, and beaglebone is not OMAP based though it shares IPs
Nov 16 00:29:27 <elik> joel_: looking at it, will give me a better idea of what icepick is
Nov 16 00:29:58 <joel_> yep
Nov 16 00:30:13 <elik> joel_: still a bit confused at all the "characteristics" too. Both omap and the am335X are Sitara
Nov 16 00:30:25 <elik> or some omaps are
Nov 16 00:30:29 <ds2> it is Sitara
Nov 16 00:30:44 <ds2> the Star... LoneStar CPUs? :D
Nov 16 00:32:42 <elik> CareBear\: so how much of the JTAG standard is left standing if even the electrical interface is "flexible"?
Nov 16 00:35:04 <ds2> it is serial
Nov 16 00:35:33 <alan_o> ds2: :)
Nov 16 00:35:33 <elik> And where do the jtag emulators come in (such as this one: http://www.spectrumdigital.com/product_info.php?&products_id=230)
Nov 16 00:35:36 <elik> ds2: hehe
Nov 16 01:42:43 <xxiao> thurbad: ping
Nov 16 01:43:13 <thurbad> what's up?
Nov 16 01:43:24 <thurbad> or is pong the correct response?
Nov 16 01:43:42 <xxiao> is there a way you dropbox second partition to me? the ROOT partition?
Nov 16 01:43:57 <xxiao> when i replace MLO i mistaken the second partition and erased it all
Nov 16 01:43:59 <xxiao> :(
Nov 16 01:47:12 <thurbad> I can get you one that is similar tonight, but had some gles stuff added orthe correctone tomorrow morning
Nov 16 01:47:35 <xxiao> ok, want to run some stress test with 1x256M tonight
Nov 16 01:48:01 <xxiao> thanks
Nov 16 01:48:55 <xxiao> thurbad: i just found my copy, no need to send me
Nov 16 01:49:09 <thurbad> cool
Nov 16 01:49:17 <xxiao> my 1TB is full of stuff and i stored it to a different partition which took me a while to recall
Nov 16 01:53:43 <mru> KotH: _av500_: http://pastie.org/5385547
Nov 16 02:05:35 <alan_o> mru: are we supposed to do this by hand and write out the answer?
Nov 16 02:12:00 <ds2> I like this code
Nov 16 02:13:45 <ds2> it would break on ROM'ing if a crude crt is used
Nov 16 02:14:40 <Russ> hmmm...should I click the 'report abuse' button?
Nov 16 02:14:54 <Russ> I can figure out what it says without even looking
Nov 16 02:14:57 <Russ> or running
Nov 16 02:15:05 <Russ> hello world\n
Nov 16 02:21:55 <ds2> i
Nov 16 02:24:34 <Russ> continue
**** ENDING LOGGING AT Fri Nov 16 02:59:58 2012