**** BEGIN LOGGING AT Sun Apr 22 03:00:01 2012
Apr 22 03:27:13 <I2c> Hi. I want to interface an i2c device to the bbxm expansion header i2c port
Apr 22 03:28:15 <I2c> I read in elinux.org about enabling the second i2c port (i2c2)
Apr 22 03:29:30 <I2c> In my kernel (2.6.37), the file beagle.h is not there but the pin mixing is done in mux34xx.c it seems
Apr 22 03:30:19 <I2c> Can somebody throw little light on this? Thanks!
Apr 22 03:34:05 <I2c> Does anybody know how to enable the 2nd i2c interface in bbxm?
Apr 22 03:34:44 <I2c> Am having 2.6.37 from TI Sri
Apr 22 03:35:04 <I2c> sdk*
Apr 22 06:30:02 <aholler> 11°C
Apr 22 06:36:45 <aholler> maybe I should think about relocating to california ;)
Apr 22 06:54:40 <_av500_> aholler: think twice http://online.wsj.com/article/SB10001424052702304444604577340531861056966.html?mod=opinion_newsreel
Apr 22 06:55:11 <_av500_> but then, you might fit the profile :)
Apr 22 06:55:39 <_av500_> ah no, they would not let you in even, burning tobacco is verboten
Apr 22 06:58:13 <aholler> strong argument
Apr 22 06:59:11 <aholler> i would have to lose my insurance against pension
Apr 22 06:59:33 <_av500_> you are insured not to have a pension? :)
Apr 22 06:59:52 <aholler> I do what's possible ;)
Apr 22 07:00:26 <aholler> but looking at our old chancelor (schmidt), that doesn't seem to work
Apr 22 07:00:52 <aholler> at least not reliable
Apr 22 07:01:11 <_av500_> true
Apr 22 07:01:39 <_av500_> and i acknowledge how you are tryin to keep your old age mediacal costs down
Apr 22 07:01:52 <_av500_> by not reaching it
Apr 22 07:13:02 <aholler> I've almost had success booting 3.3.2, had enabled CONFIG_ARM_PATCH_PHYS_VIRT and it booted a bit more. Something is really serious broken in 3.3.x.
Apr 22 07:38:35 * _av500_ only boots kernels that have Google bits stolen from Oracle
Apr 22 07:39:12 <aholler> those non-patented ones?
Apr 22 07:39:55 <_av500_> the abs() implementation that larry ellison wrote himself on his yacht
Apr 22 07:44:36 <aholler> I've smuggled it of the usa: int abs(int i) { return i<0 ? -i : i; }
Apr 22 07:45:27 <_av500_> thats not java, it does not throw an exception
Apr 22 07:45:47 <aholler> but java itself isn't written in java
Apr 22 07:45:48 <_av500_> also, no camelCasing
Apr 22 07:46:48 <_av500_> of course not, its written in pixie dust
Apr 22 07:47:47 * aholler remembers the time where pgp was faxed to foreigners ;)
Apr 22 07:48:16 <Russ> aholler, so with 3.3.x, CONFIG_ARM_PATCH_PHYS_VIRT in broken on beagle?
Apr 22 07:48:31 <aholler> Russ: no, just different broken
Apr 22 07:48:44 <aholler> I really don't know whats happening.
Apr 22 07:48:44 <Russ> share
Apr 22 07:49:03 <aholler> but most of the time 3.3.2 just comes to "booting kernel"
Apr 22 07:49:27 <Russ> do you have a .config?
Apr 22 07:49:52 <aholler> sure, many, none worked ;)
Apr 22 07:50:14 <_av500_> aholler: stop random generating them
Apr 22 07:50:40 <aholler> fuzzying them is the only solution left
Apr 22 07:50:41 <aholler> ;)
Apr 22 07:50:59 <_av500_> at least be systematic and start from allnoconfig
Apr 22 07:51:18 <_av500_> until you reach allyesconfig
Apr 22 07:53:11 <aholler> Russ: http://ahsoftware.de/config-3.3.2.beagleboard
Apr 22 07:53:34 <aholler> the uart-patch was applied
Apr 22 07:55:08 <aholler> I'fve tried to switch CONFIG_RESET_CLOCKS, NO_HZ, CONFIG_ARM_PATCH_PHYS_VIRT and CONFIG_OMAP_SMARTREFLEX
Apr 22 07:55:19 <aholler> and some combinations of them
Apr 22 07:55:35 <Russ> maybe I'll get to try ktest's config bisection
Apr 22 07:55:58 <aholler> next try will be an recent u-boot, I'm still using 2011.3 or such
Apr 22 07:56:27 <Russ> 3.3.2 with that config is building now
Apr 22 08:00:53 <Russ> hooray for remote testing http://www.flickr.com/photos/31208937@N06/6952269298
Apr 22 08:02:55 <aholler> I'm using patented gender-changers ;)
Apr 22 08:04:57 <Russ> 0-3 boots for now
Apr 22 08:05:46 <aholler> from where?
Apr 22 08:06:26 <Russ> I'm up to 5 attempts
Apr 22 08:06:31 <Russ> with no successful boots
Apr 22 08:06:35 <Russ> (on xM)
Apr 22 08:06:37 <aholler> ah, ok ;)
Apr 22 08:06:52 <aholler> I just thought you are successfull booting
Apr 22 08:07:43 <aholler> how do you reset the board remotely?
Apr 22 08:08:38 <Russ> that dongle
Apr 22 08:09:47 <aholler> using dtr or such?
Apr 22 08:09:53 <Russ> yup
Apr 22 08:10:00 <aholler> hmm, good idea
Apr 22 08:11:11 <Russ> https://github.com/russdill/dtr/wiki/dtr_std_schematic.png
Apr 22 08:12:48 <Russ> https://github.com/russdill/dtr/wiki/dtr_null_schematic.png
Apr 22 08:22:11 <aholler> send them to jkridner_ for distribution ;)
Apr 22 08:28:24 <aholler> I always forget how variant/revision maps to rNpM
Apr 22 08:28:48 <aholler> and wonder why those arm-erratas can't be enabled at runtime after cpu detection
Apr 22 08:30:55 <koen> don't some of those errata (it's already plural) affect the bootup asm?
Apr 22 08:31:25 <aholler> not sure, but thought about, therefor I'm just checking which ones I have to enable
Apr 22 08:31:33 <Russ> the kernel certainly has a lot of the framework in place to patch out unused errata
Apr 22 08:31:51 <Russ> its probably just because no one is motivated to work on it
Apr 22 08:33:04 <aholler> they should write variant x and revision y in the help texts so that people could just compare them with the output of /proc/cpuinfo
Apr 22 08:33:38 <aholler> or /proc/cpuinfo should output rNpM
Apr 22 08:33:55 * Russ notices that next doesn't build
Apr 22 08:40:27 <aholler> does an errata for r2p0 apply to r1p3?
Apr 22 08:41:03 <Russ> it depends, is it a regression?
Apr 22 08:41:12 <Russ> next boots ok
Apr 22 08:41:30 <aholler> with 3.3.2 everything is ok, so my config is a regression ;)
Apr 22 08:41:37 <aholler> s/3.3.2/3.2.x/
Apr 22 08:41:52 * Russ sends an email to keshava and samuel ortiz re next breakage
Apr 22 08:43:11 <Russ> aholler, I'll figure out what the deal is
Apr 22 08:43:31 <aholler> no hassle, not important ;)
Apr 22 08:43:50 <aholler> I'm playing with my armv5-stuff right now ;)
Apr 22 08:44:33 <aholler> and thinking about building me a qmlviewer which works with the fb only
Apr 22 08:45:29 <aholler> for scripting some simple touch-enabled guis
Apr 22 08:50:37 <koen> Russ: do you have all your patches for 3.4rc online somewhere?
Apr 22 08:50:52 <Russ> I try to email stuff out when its ready
Apr 22 08:51:01 <Russ> I don't have a github linux fork setup yet
Apr 22 11:48:22 <ssvb> koen: hi, have you tried xf86-video-omap and omapdrm on omap3 hardware?
Apr 22 12:50:26 <koen> ssvb: not yet
Apr 22 12:50:39 <koen> ssvb: it requires changes to the public sdk
Apr 22 12:51:11 <koen> ssvb: I quit before I reached that part of my TODO :)
Apr 22 13:20:29 <ssvb> koen: ok, I hope it's still in somebody's TODO and they are not going to totally abandon OMAP3 in favour of OMAP4/5 yet :)
Apr 22 13:23:09 <aholler> I work actively that they don't do that ;)
Apr 22 15:51:08 <barts_> whats the latest OS i can use?
Apr 22 15:53:14 <emeb> what's the latest OS you can use on a PC?
Apr 22 15:56:40 <koen> windows 8
Apr 22 15:57:14 <emeb> s/t/m/
Apr 22 18:22:18 <Russ> wow, this channel really is the "ask anything" channel
Apr 22 18:22:26 <Russ> "what is the latest OS I can use?"
Apr 22 18:39:42 <koen> Russ: hpux 11iv2
Apr 22 18:40:42 <Russ> why, how much over schedule were they?
Apr 22 18:41:40 <LetoThe2nd> koen: i think by "OS" he/she meant Operation System, not Onomatopoeic Strangeness.
Apr 22 18:52:03 <mranostay> morning
Apr 22 18:53:15 <koen> hey mranostay
Apr 22 18:55:15 <mranostay> how is it in koen-land?
Apr 22 19:00:34 <mranostay> god Fritling crash again
Apr 22 19:01:04 <mranostay> *Fritzing
Apr 22 19:07:23 <Russ> aholler, its 'ARM: OMAP2+: hwmod: Add a new flag to handle hwmods left enabled at init' that breaks your config
Apr 22 19:07:35 <Russ> aacf094128
Apr 22 19:15:10 <aholler> Russ: thanks a lot, will try it
Apr 22 19:15:59 <Russ> I'm not sure why that breaks things
Apr 22 19:16:10 <Russ> now I'm curious what config option it is that does th ebreaking
Apr 22 19:17:03 <aholler> sounds it like it disables something it shouldn't
Apr 22 19:17:49 <Russ> dunno, my initial bisect failed because I didn't refresh the config and something in one of the kernel versions changed it
Apr 22 19:18:02 <Russ> so all revisions were marked as good
Apr 22 19:19:20 <aholler> yes "" | make oldconfig
Apr 22 19:20:07 <aholler> is what I've used for bisecting the uart-stuff
Apr 22 19:20:50 <Russ> to bisect this, I had to recopy the .config you posted to the .config before doing an oldconfig command
Apr 22 19:21:08 <aholler> sure, that was already an 3.3.2 config.
Apr 22 19:21:31 <aholler> I could have send an 3.2.x too ;)
Apr 22 19:22:02 <aholler> anyway, good to know your automatic bisecter works ;)
Apr 22 19:22:43 <aholler> you should offer a tinderbox for that
Apr 22 19:23:59 <Russ> eh, I just need to start using ktest.pl
Apr 22 19:25:43 <mru> an automatic bisecter like this? http://farm5.staticflickr.com/4022/4656503521_57b5b2d702_z.jpg
Apr 22 19:26:19 <Russ> exactly
Apr 22 19:28:30 <aholler> but your fine remote-resetter is the magic ;)
Apr 22 19:29:19 <Russ> I think most people have the 110AC web controlled thingy
Apr 22 19:29:46 <Russ> http://www.digital-loggers.com/lpc.html
Apr 22 19:33:56 <aholler> hmm , I' thought about buying that one: (german) http://www.pearl.de/a-NC5153-5452.shtml
Apr 22 19:37:07 <Russ> dunno, I've heard a lot of good things about digital loggers, but its a problem if they only have US style outlets
Apr 22 19:37:39 <aholler> but I think I will just use the web-enabled avr + some relay I have lying around ;)
Apr 22 19:38:06 <Russ> In the past, I used some X10 stuff, but they are easily subject to interference
Apr 22 19:38:28 <Russ> http://www.digital-loggers.com/220.html
Apr 22 19:38:35 <Russ> dunno how much that is
Apr 22 19:38:53 <Russ> there it is, $249
Apr 22 19:39:53 <aholler> Hmm, or I build somne lego-stuff which presses the reset-button
Apr 22 19:40:01 <aholler> ;)
Apr 22 19:40:15 <aholler> would be much more fun
Apr 22 19:40:21 <Russ> the one with us outlets is still rated for up to 240V
Apr 22 19:40:38 <Russ> so if you get pigtails that give you the outlet you need
Apr 22 19:59:34 <_av500_> aholler: rather get the net-io from reichelt
Apr 22 19:59:40 <_av500_> than that pearl crap
Apr 22 20:11:29 <koen> mranostay: weekend, being lazy
Apr 22 20:12:45 * Russ wonders if anyone has started work on ktest.py
Apr 22 20:13:19 <koen> Russ: matt or tom did
Apr 22 20:13:30 <koen> or was that ktest.pl
Apr 22 20:13:33 * koen forgot
Apr 22 20:13:42 <Russ> ktest.pl is already done
Apr 22 20:13:54 <koen> the rostedt thing
Apr 22 20:14:52 <Russ> yes
Apr 22 20:16:18 <aholler> av500: I have a net-io, but that pearl-thingy is something ready-to-use with 220v.
Apr 22 20:16:42 <koen> Russ: your serial reset thingy looks really nifty
Apr 22 20:17:24 <Russ> I'll see if tincantools can start making something like it
Apr 22 20:20:43 <Russ> every once in a while I spend a little time trying to setup ktest.pl, but its always a square peg round hole thing
Apr 22 20:21:34 <koen> Russ: I've been thinking for a while on it, I think I'd go for a 'complete' solution with ftdi usb2serial
Apr 22 20:21:45 <koen> Russ: like a flyswatter2 on steroids
Apr 22 20:21:51 <Russ> koen, I have one of those coming in too
Apr 22 20:22:00 <Russ> I think the pcbs will arrive monday
Apr 22 20:22:11 <LetoThe2nd> oO( steroidswatter )
Apr 22 20:23:00 <Russ> http://www.flickr.com/photos/31208937@N06/7043352867 http://www.flickr.com/photos/31208937@N06/7043352833
Apr 22 20:23:50 <Russ> supports power cycling, power consumption measurements, and two soft switches (for connecting to things like user and reset)
Apr 22 20:23:52 <Russ> no jtag though
Apr 22 20:24:59 * koen doesn't really care about jtag, only about serial and reset
Apr 22 20:25:10 <Russ> me too
Apr 22 20:25:32 <koen> powerconsumption measurement is a nice extra
Apr 22 20:25:39 <Russ> ti ina219
Apr 22 20:25:48 * koen stabs rtsc
Apr 22 20:25:58 <Russ> http://www.ti.com/product/ina219
Apr 22 20:26:11 <koen> I have a tab open on the ina219 page for a while now :)
Apr 22 20:26:18 <Russ> amp hour?
Apr 22 20:26:37 <koen> _roger_ wanted to do a cape that measured the 6 beagles in the videowall
Apr 22 20:44:47 <_av500_> aholler: sorry, meant this one: http://www.reichelt.de/Messen-steuern-regeln/LAN-NETIO230B-SI/index.html?ACTION=3&GROUPID=4308&ARTICLE=102254&SHOW=1&START=0&OFFSET=16&
Apr 22 21:16:13 <aholler> eh, 4W
Apr 22 21:27:32 <aholler> _av500_: would be nice if that would measure too
Apr 22 21:28:12 <_av500_> aholler: well...
Apr 22 21:28:35 <_av500_> aholler: my boss had me order one and send to paris, so far he likes it
Apr 22 21:29:37 <Russ> aholler, but then you'd have to account for loss in the ac/dc conversion
Apr 22 21:33:35 <aholler> maybe I will build me a radio-solution. I have already devices which switch and measure by radio. Just have to check out how to switch them myself (using a rfm12 or such)
Apr 22 21:38:29 <aholler> http://nanl.de/blog/2011/01/rfm12-under-linux-and-remote-controlled-power-sockets/
Apr 22 21:40:39 <Russ> my bad experience with x10 and interference would make me worry
Apr 22 21:40:45 <_av500_> and have the neighbours garage door reset your beagle?
Apr 22 21:43:12 <Russ> I do think I might setup my beagle to control my router, cable modem, and voip box, but that'd be hard wired
Apr 22 21:49:52 <aholler> yeah, already searched for some solid state relais too
Apr 22 21:50:26 <aholler> but they seem to be pricy
Apr 22 21:50:36 <Russ> if you do it after the DC conversion, its really easy
Apr 22 21:51:12 <Russ> you can get tiny power mosfets that can handle 20A
Apr 22 21:52:34 <Russ> hmmm...maybe I should make a board for beagle that has power mosfets and some current measurement devices
Apr 22 21:52:48 <Russ> it'd be really easy since you have i2c and gpios ready to go
Apr 22 21:53:43 <aholler> an i2c-only solution would be nice
Apr 22 21:54:06 <Russ> then you'd just need a gpio expander
Apr 22 21:55:22 <aholler> that ti-chip is i2c too, so a gpio-expander that ti-chip and some mosfets ;)
Apr 22 21:56:23 <Russ> basically, with a p-channel mosfet, to turn it on, you pull the gate low
Apr 22 21:56:37 <Russ> to turn it off, you allow a pullup resistor to pull it up to the voltage of the source
Apr 22 22:11:22 <hiwaeng> any recommendation on RTOS for beagleboard?
Apr 22 22:11:49 <_av500_> yes, find out if you really need an RTOS
Apr 22 22:15:23 <emeb> Everyone needs an RTOS!
Apr 22 22:20:51 <hiwaeng> <_av500_> I think so, since I am reading some sensor data through beagleboard i2c. and the problem is it takes about 30 ms to read from 4 sensors which is too much
Apr 22 22:22:32 <Russ> I'm not really sure that an rtos would have any bearing on that
Apr 22 22:23:08 <Russ> what is the bitrate of the i2c bus and how many bits do you need to ready/write/
Apr 22 22:23:29 <hiwaeng> but I think it can reduce the time of readings I guess
Apr 22 22:24:30 <Russ> what makes you think that
Apr 22 22:24:36 <hiwaeng> bitrate is 400kHz, and the data from each axis of my sensor"accelerometer" is 16 bit "high byte and low byte"
Apr 22 22:26:02 <Russ> and have you inserted timing information to find out how long things are taking
Apr 22 22:26:17 <Russ> and just so you know, rtos's are not faster, they are only more predictable
Apr 22 22:27:14 <hiwaeng> yes I inserted timing... but I don't know why it takes so much time for such reading?
Apr 22 22:27:29 <Russ> did you insert timing into the driver?
Apr 22 22:28:18 <hiwaeng> what you mean? i used timing before starting the execution loop and after
Apr 22 22:28:32 <hiwaeng> and I am using beagleboard i2c tools
Apr 22 22:28:47 <aholler> userland
Apr 22 22:28:48 <Russ> bash? c?
Apr 22 22:28:56 <hiwaeng> c
Apr 22 22:29:58 <Russ> and you are somehow convinced that there is some part of the kernel causing latency interrupting your i2c transfer?
Apr 22 22:30:03 <Russ> how did you come to that conclusion?
Apr 22 22:30:49 <Russ> it really is probably just that the i2c driver is not designed for low latency
Apr 22 22:30:58 <hiwaeng> I surfed on interned, and at the end I came up with this
Apr 22 22:31:07 <Russ> it may even have sleeping delays purposely inserted
Apr 22 22:31:15 <Russ> you need to analyze the driver
Apr 22 22:31:20 <hiwaeng> how?
Apr 22 22:31:28 <Russ> by looking at the source
Apr 22 22:31:43 <hiwaeng> which source?
Apr 22 22:31:53 <Russ> for the i2c driver in the kernel
Apr 22 22:32:42 <hiwaeng> but I am somehow newie , can you be please more specific? which file should I check?
Apr 22 22:33:16 <Russ> probably i2c-omap.c
Apr 22 22:34:27 <hiwaeng> or I can analyse the i2c using osilliscope?
Apr 22 22:35:48 <hiwaeng> or?
Apr 22 22:37:07 <Russ> you could, it probably wouldn't tell you much
Apr 22 22:38:05 <Russ> you could start by enabling the dev_dbg messages in i2c-omap.c and that would give you some basing timing info
Apr 22 22:40:34 <hiwaeng> so that is the only way to reduce latency?
Apr 22 23:09:50 <Russ> boot
Apr 22 23:09:52 <Russ> oops
Apr 22 23:15:09 <dwery> booting not allowed. a)borte, r)etry, c)cancel
Apr 22 23:15:32 <mru> you forgot ignore and fail
Apr 22 23:15:37 <dwery> s'oh!
Apr 22 23:15:40 <dwery> d'oh!
Apr 22 23:42:57 * Russ attempts a ktest.pl config-bisect
Apr 23 00:52:57 <mranostay> Russ: loving the perl? :)
Apr 23 01:02:04 <Russ> I'm just not a perl person
Apr 23 01:49:43 <Russ> bummer, the ktest.pl config bisect feature only handles very specific situations
**** ENDING LOGGING AT Mon Apr 23 02:59:58 2012