• [00:30:50] * prpplague^2 (n=dave123_@ppp-70-242-123-131.dsl.rcsntx.swbell.net) has joined #beagle
  • [00:30:50] * prpplague (n=dave123_@ppp-70-242-123-131.dsl.rcsntx.swbell.net) Quit (Read error: 104 (Connection reset by peer))
  • [00:34:11] * prpplague^2 is now known as prpplague
  • [00:37:00] * rsalveti (n=salveti@200.184.118.132) Quit (Read error: 113 (No route to host))
  • [00:43:12] * rsalveti (n=salveti@189.70.212.146) has joined #beagle
  • [00:56:01] * Olipro (n=Olipro@unaffiliated/olipro) has joined #beagle
  • [00:56:01] * prpplague (n=dave123_@ppp-70-242-123-131.dsl.rcsntx.swbell.net) Quit (Read error: 104 (Connection reset by peer))
  • [00:56:02] * prpplague^2 (n=dave123_@ppp-70-242-123-131.dsl.rcsntx.swbell.net) has joined #beagle
  • [01:06:13] <JoeBorn> jkridner: do you have any dead boards? they might be useful for thinking about housings.
  • [01:34:18] * BThompson (n=BThompso@cpe-76-185-93-11.tx.res.rr.com) has joined #beagle
  • [01:37:06] * JoeBorn is now known as JB_out
  • [01:54:04] <jkridner> hmmm.... yeah, we have some that are less than functional.
  • [01:54:38] <jkridner> need to think about a few things, including if we are going to shift back to a DVI-D connector.
  • [01:54:57] <jkridner> we are worried that the current connector gives the wrong impression and might be harder to find cables for.
  • [01:55:56] * mru doesn't see enough room on the board for a dvi connector
  • [01:56:43] <jkridner> it would mean changing the expansion connector pins.
  • [01:57:52] <ds2> U-boot on the EVM is EVIL!
  • [02:00:11] <ds2> leaving clocks on willy nilly :(
  • [02:48:38] * RobotGuy (n=robotguy@dsl093-038-072.pdx1.dsl.speakeasy.net) has joined #beagle
  • [03:04:42] <BThompson> jkridner: for people looking for a HDMI to DVI solution, I use the adapter here http://www.monoprice.com/products/product.asp?c_id=104&cp_id=10419&cs_id=1041902&p_id=2080&seq=1&format=2 ~$3 bucks and any beagle can have DVI :)
  • [03:06:29] <BThompson> I actually use one for my PS3 to my projector, but this is the cheapest hdmi -> dvi adapter ive seen, I figure I will get another one when I actually get a board :)
  • [03:37:09] * roxfan (n=dunno@156.99-201-80.adsl-dyn.isp.belgacom.be) Quit (Remote closed the connection)
  • [03:37:47] * roxfan (n=dunno@156.99-201-80.adsl-dyn.isp.belgacom.be) has joined #beagle
  • [03:43:58] * khasim (n=a0393720@192.163.20.231) has joined #beagle
  • [03:53:56] <jkridner> good morning khasim
  • [03:54:58] <jkridner> my attempts at rebuilding beagleboard-demo-image have been frustrating. it requires waiting a really-long-time for many operations.
  • [03:55:47] <jkridner> when I did 'mtn --db=~/stuff/OE.mtn pull', it was enough to convince me that monotone is much slower than git.
  • [03:56:26] <jkridner> I did my update, then bitbake had to rebuild my entire toolchain. :(
  • [03:56:40] <jkridner> build of libtool-cross failed.
  • [03:57:35] <keesj> jkridner: this is the joy of OE
  • [03:57:58] <jkridner> some tested labels would be a huge improvement.
  • [03:58:02] <keesj> jkridner: you also did a mtn update i guess?
  • [03:58:15] <jkridner> I would love to dedicate some CPU cycles to auto-building and reporting status.
  • [03:58:19] <jkridner> yes.
  • [03:58:47] <jkridner> I'm getting armv7a output now, which is new since the last time I updated.
  • [04:00:29] <jkridner> how soon before the shift to git?
  • [04:00:46] <jkridner> do you think anyone will create some labels?
  • [04:01:24] <jkridner> would it be a useful service to run 'updates/builds' via a cron job so that everyone knows the current state?
  • [04:01:29] <keesj> myself I have only been really bussy on mamona and we swiched to git a few month ago
  • [04:04:02] <keesj> I guess it would be very good to have known working stets and that mean branching. I guess it is really up to others to decide. but having different images ready for testing can't be a bad thing
  • [04:05:40] <jkridner> If anyone wants to suggest a particular build script to run and when to run it, I can do it with the angstrom@beagleboard.org account (http://www.beagleboard.org/~angstrom).
  • [04:06:19] <keesj> cool
  • [04:06:21] <jkridner> darn it, I really need to learn not to type e-mail addresses. all the scraping tools probably already have them.
  • [04:06:52] <jkridner> fixing e-mail addresses is something I need to teach the BeagleLogBot.
  • [04:07:36] <jkridner> please let me know. there are going to be a bunch of boards out there in the coming months and I'd like to make sure they can get a working distro.
  • [04:07:55] <jkridner> I don't mind pointing them all to koen's builds, which may be the right answer.
  • [04:10:39] <RobotGuy> Are GNU GCC tools usable with Beagle?
  • [04:11:09] <jkridner> yes, that is the common approach today.
  • [04:11:20] <jkridner> Code Sourcery 2007q3 seems to work the best.
  • [04:11:39] <RobotGuy> I couldn't see where GCC tools are downloadable.
  • [04:11:43] <jkridner> There has been some luck with user-space code on newer GCC versions coming from GNU.
  • [04:12:07] <jkridner> there are links at http://code.google.com/p/beagleboard/wiki/BeagleSourceCode
  • [04:13:10] <jkridner> It looks like Code Sourcery has adjusted the links on their site.
  • [04:13:11] <jkridner> http://www.codesourcery.com/gnu_toolchains/arm/portal/package1786?@template=release
  • [04:13:44] <RobotGuy> I am talking about being able to use 100% Open Source tools.
  • [04:13:57] <jkridner> the source is available for Code Sourcery.
  • [04:14:17] <jkridner> build your own tools with http://www.codesourcery.com/gnu_toolchains/arm/portal/package1786/public/arm-none-linux-gnueabi/arm-2007q3-51-arm-none-linux-gnueabi.src.tar.bz2
  • [04:14:20] <RobotGuy> That doesn't necessarily mean it's Open Source.
  • [04:14:45] <jkridner> uh, that is a strange definition to not call GPL code Open Source.
  • [04:15:01] <jkridner> it isn't all upstream yet, but it is getting there.
  • [04:15:18] <RobotGuy> You did not mention GPL. Just because source is available, doesn't mean it is GPL.
  • [04:15:33] <jkridner> ah, but GCC does.
  • [04:15:44] <RobotGuy> GCC is GPL
  • [04:15:46] <jkridner> this is, after all, just GCC.
  • [04:16:47] <RobotGuy> Can the GNU GCC be used? I mean pure GNU GCC?
  • [04:17:08] <jkridner> you can also use upstream GCC, but there have been some bugs in GCC that show up when building the kernel.
  • [04:17:29] <jkridner> at least, I believe, it is suspected the bugs are in GCC.
  • [04:18:00] <RobotGuy> Can, say, GCC 4.1.2 be used to build software for BeagleBoard?
  • [04:18:03] <jkridner> also, Code Sourcery 2008q1 seems to have some NEON auto-vectorization bugs.
  • [04:18:32] <jkridner> but, I heard that they have some fixes to 2008q1 that haven't been made public yet.
  • [04:18:47] <RobotGuy> You are not answering questions very well.
  • [04:19:03] <jkridner> I guess I'm telling you more than you want to hear.
  • [04:19:19] <RobotGuy> You aren't answering my questions.
  • [04:19:21] <jkridner> do I know that every single version of GCC works? no.
  • [04:20:01] <RobotGuy> Pick any version of GNU GCC. Will it work to build software for BeagleBoard?
  • [04:20:05] <jkridner> if you plan on running unoptimized code, then I would suspect that GCC 4.1.2 would work fine.
  • [04:20:44] <jkridner> I'll have to check which version it is that runs the user-space code fine. it is much newer than 4.1.2.
  • [04:21:04] <jkridner> for the kernel, I recommend the Code Sourcery 2007q3 version.
  • [04:21:20] <jkridner> I'll check which version I'm using of GNU GCC right now.
  • [04:22:01] <RobotGuy> I must be able to build software using a "pure" version of GNU GCC.
  • [04:22:15] <jkridner> ngstrom@domU-12-31-38-00-9C-98 ~/stuff/tmp/cross $ arm-angstrom-linux-gnueabi/b
  • [04:22:15] <jkridner> in/gcc --version
  • [04:22:15] <jkridner> gcc (GCC) 4.3.1
  • [04:22:48] <RobotGuy> What does your gcc -V show?
  • [04:23:10] <jkridner> angstrom@domU-12-31-38-00-9C-98 ~/stuff/tmp/cross $ arm-angstrom-linux-gnueabi/bin/gcc -V
  • [04:23:11] <jkridner> gcc: '-V' option must have argument
  • [04:23:20] <RobotGuy> Must be -v then
  • [04:24:16] <jkridner> pastebin.com is slow.
  • [04:25:24] <jkridner> http://www.beagleboard.org/~angstrom/tmp/cross/gcc-v.txt
  • [04:26:26] <jkridner> does that tell you something more?
  • [04:27:05] <jkridner> I don't mind answering your questions, but you might get better answers if you give me more background.
  • [04:27:45] <jkridner> Beagle is still in relatively early stages and most of the people around here are reasonably patient.
  • [04:27:51] <RobotGuy> I'll just pass for now. I guess I am just not communicating well.
  • [04:28:10] <jkridner> I can give you the URL where that source was pulled.
  • [04:28:43] <jkridner> it came directly from GNU. There might be some OpenEmbedded patches applied to it at most.
  • [04:29:14] <jkridner> ARM Cortex-A8+NEON support is very new. You can use older tools just fine, but they won't be as optimized.
  • [04:29:46] <jkridner> In general, with the super-scaler ARM, people are seeing 1.5X+ improvement in instructions per cycle.
  • [04:30:16] * RobotGuy (n=robotguy@dsl093-038-072.pdx1.dsl.speakeasy.net) has left #beagle
  • [04:30:23] <jkridner> ouch.
  • [04:32:00] <jkridner> All, let me know if could have handled that better.
  • [04:32:23] <keesj> /ignore :p
  • [04:33:49] <jkridner> back to reading binutils-cross error messages.
  • [04:34:18] <keesj> thinking of it we will not want everybody to build an image from scratch.
  • [04:35:47] <keesj> perhaps an appliance is a good idea, so we can point to that.
  • [04:35:48] <jkridner> probably not.
  • [04:36:31] <jkridner> at one point, I started to create a shell script that would build an Amazon EC2 machine image...
  • [04:36:43] <jkridner> I'm thinking that could be an OE target.
  • [04:36:53] <jkridner> sorry, just had to go on a day dream vector.
  • [04:37:02] <keesj> that sound like what openmoko is doing
  • [04:37:06] <jkridner> how do I find the status of the last build?
  • [04:37:13] <jkridner> oh? any good starting point?
  • [04:37:30] * jkridner has to step away from the computer.
  • [04:37:48] <keesj> I don't follow you
  • [05:07:54] * moenicke (n=moenicke@unaffiliated/moenicke) has joined #beagle
  • [05:14:33] * Olipro (n=Olipro@unaffiliated/olipro) Quit (Connection timed out)
  • [05:23:23] * JoeBorn_ (n=jborn@adsl-75-2-240-30.dsl.chcgil.sbcglobal.net) has joined #beagle
  • [05:27:32] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
  • [05:37:40] * moenicke (n=moenicke@unaffiliated/moenicke) Quit ()
  • [06:14:18] * BThompson (n=BThompso@cpe-76-185-93-11.tx.res.rr.com) Quit ("Trillian (http://www.ceruleanstudios.com")
  • [07:07:56] * shoragan (n=shoragan@debian/developer/shoragan) Quit ()
  • [07:09:11] * shoragan (n=shoragan@debian/developer/shoragan) has joined #beagle
  • [08:16:28] * likewise (n=chatzill@82-171-51-231.ip.telfort.nl) has joined #beagle
  • [08:16:32] <likewise> gm all!
  • [08:20:05] <koen> hey likewise !
  • [08:21:47] <likewise> hey koen, up early and committing already? :-)
  • [08:22:39] <koen> yeah
  • [08:57:32] * ldesnogu_ (n=ldesnogu@ven06-2-82-247-86-183.fbx.proxad.net) has joined #beagle
  • [09:27:24] * khasim (n=a0393720@192.163.20.231) has joined #beagle
  • [09:45:42] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
  • [10:36:37] <Crofton|work> jkridner, my demo-imae build succeeded
  • [10:38:36] <mru> hi guys
  • [10:52:10] <Crofton> gm
  • [10:53:00] <mru> koen: you around?
  • [11:16:05] <koen> yes
  • [11:17:16] <mru> koen: I heard you (might?) be coming this lugradio event in july
  • [11:17:28] <mru> have you given it any more thought?
  • [11:17:55] <koen> I'm planning to go there
  • [11:18:06] <koen> AIUI the booth isn't 100% certain yet
  • [11:21:52] <jkridner> any Qs I can answer?
  • [11:22:04] <Crofton|work> airfare to .eu in the summer is outrageous
  • [11:22:24] <Crofton|work> What is your favorite colour?
  • [11:22:47] <jkridner> :)
  • [11:23:04] <jkridner> I mean, good morning.
  • [11:23:14] <koen> mru: is your NEON ffmpeg stuff online somewhere?
  • [11:23:38] <mru> it's in my git repo
  • [11:23:49] <mru> git://git.mansr.com/ffmpeg.mru
  • [11:23:53] <mru> branch arm-neon
  • [11:24:10] <mru> there's a gitweb interface too
  • [11:28:49] <mru> so far, idct on an mpeg2 sample is down from 33.5% to 20%
  • [11:28:53] <mru> according to oprofile
  • [11:28:58] <koen> nice
  • [11:29:16] <mru> and I haven't even tried to optimise instruction scheduling
  • [11:35:46] <jkridner> Crofton|work: is there a label I can checkout?
  • [11:36:35] <jkridner> do you have glibc 2.6.1, or do you get that warning at the start?
  • [11:38:29] <jkridner> http://www.beagleboard.org/~angstrom/tmp/work/armv7a-angstrom-linux-gnueabi/libtool-cross-2.2.4-r23/temp/log.do_stage.28886
  • [11:38:54] <jkridner> I guess that is mostly self explanatory, but I don't know why I don't have libtool.
  • [11:39:46] <Crofton|work> when did you pull last?
  • [11:40:18] <Crofton|work> http://tinderbox.openembedded.net/builds/8721/
  • [11:40:27] <Crofton|work> the rev I built is there
  • [11:40:47] <jkridner> yesterday.
  • [11:41:08] * shoragan (n=shoragan@debian/developer/shoragan) Quit (kubrick.freenode.net irc.freenode.net)
  • [11:41:16] * shoragan (n=shoragan@debian/developer/shoragan) has joined #beagle
  • [11:41:27] <Crofton|work> try pulling
  • [11:41:34] <jkridner> ah! I didn't know about tinderbox.
  • [11:41:46] <Crofton|work> it is pretty cool
  • [11:42:23] <koen> I need to poke jeremy about getting the rev from git
  • [11:42:35] <Crofton|work> mtn update -r c7403af10901017ee2fc0ec3e0caaed615cb6687
  • [11:43:29] <Crofton|work> although the only difference between that rev and HEAD are some EVm kernel updates
  • [11:43:45] <Crofton|work> I'm still confused by the libtool problem though
  • [11:44:29] <jkridner> http://tinderbox.openembedded.net/search/?machine=beagleboard is almost exactly what I wanted to do.
  • [11:45:05] <Crofton|work> I've commented out the apps that fail in demo image
  • [11:45:11] <Crofton|work> until fixes are committed
  • [11:46:39] <jkridner> http://mojo.handhelds.org/grumpy-armv6el-vfp has a bit more of a summary view.
  • [11:46:48] <jkridner> (for the wrong distro)
  • [11:47:53] <mru> hint: avoid using vfp on cortex-a8
  • [11:47:56] <mru> it's very slow
  • [11:48:06] <mru> floating point code should use neon
  • [11:49:22] <jkridner> wouldn't it still be faster than ARMv5EL or plain ARMv6?
  • [11:49:53] <mru> those would probably be using softfloat
  • [11:50:02] <mru> but vfp on cortex-a8 isn't pipelined
  • [11:50:34] <mru> it's a heavily simplified unit, only there for compatibility
  • [11:51:45] <jkridner> I just mean that shouldn't the floating point libs still be faster than softfloat?
  • [11:52:23] <jkridner> or does the code optimizer start adding a bunch of floating-point operations?
  • [11:52:49] <jkridner> I would think that even an unpipelined floating point operation is still faster than emulation.
  • [11:56:35] <jkridner> koen: why did you remove the smartreflex patch?
  • [11:57:02] <koen> jkridner: because it's upstream now?
  • [11:57:12] <jkridner> good reason. :)
  • [11:57:51] <koen> jkridner: http://source.mvista.com/git/gitweb.cgi?p=linux-omap-2.6.git;a=commit;h=3f42fa9cb48a25a52a62a3fe1ff3aee7bd564d7b
  • [11:58:07] * jkridner continues to fumble around trying to learn.
  • [12:03:12] <koen> -mfpu=neon should take care of it, right?
  • [12:03:18] <mru> yes
  • [12:03:34] <mru> but beware, only gcc-csl 2007q3 seems to work with that flag
  • [12:03:46] <mru> stock gcc dies immediately
  • [12:04:34] * Crofton|work has a leaky IV bag, must go get a replacement from the vet
  • [12:05:10] <mru> sounds like your cat isn't well at all... hope he gets better
  • [12:09:15] <koen> mru: http://gitweb.openembedded.net/?p=org.openembedded.dev.git;a=blob_plain;f=packages/gcc/gcc-4.3.1/gcc-release-branch/PR-36500-fix-neon.patch;hb=HEAD
  • [12:09:42] <jkridner> Crofton|work: I still have the libtool issue.
  • [12:10:04] <mru> koen: yes, I've seen something like that patch before
  • [12:10:05] <koen> jkridner: bitbake -c clean libtool libtool-cross
  • [12:10:11] <mru> but gcc 4.3 still miscompiles the kernel
  • [12:10:15] <mru> so I don't trust it
  • [12:11:21] <koen> /tmp/cca7mIKa.s:26: Error: bad instruction `vadd.i16 q0,q0,q0'
  • [12:11:23] <koen> hmm
  • [12:11:48] <mru> did you use -mcpu=cortex-a8?
  • [12:12:15] <koen> arm-angstrom-linux-gnueabi-gcc -mcpu=cortex-a8 -mfpu=neon -o test.o test.c
  • [12:12:52] <mru> maybe you need newer binutils
  • [12:12:57] <koen> and test.c is: int foo(void){ asm volatile("vadd.i16 q0, q0, q0"); }
  • [12:13:42] <koen> mru: 2.18.50.0.7 is too old?
  • [12:14:06] <mru> I'm using 2.18.50.0.7.20080502
  • [12:14:19] <mru> yours seems fairly recent
  • [12:15:27] <mru> your little program compiles here with those flags
  • [12:15:56] <koen> it's your little program :)
  • [12:16:04] <koen> it's the ffmpeg configure test for neon :)
  • [12:16:08] <mru> oh, I see
  • [12:16:15] <mru> well, that one works here too
  • [12:17:29] <mru> run gcc with -S and let's have look at the assembler it produces
  • [12:19:48] <koen> mru: http://rafb.net/p/17HtN045.txt
  • [12:20:18] <mru> .fpu softvfp
  • [12:20:21] <mru> that explains it
  • [12:20:42] * koen wonders where it gets that from
  • [12:20:52] <mru> that's probably the default
  • [12:21:01] <mru> maybe it ignored -mfpu=neon
  • [12:23:29] <koen> it must have
  • [12:23:45] <mru> try -mfpu=vfp and see if that makes it through
  • [12:25:57] <koen> still softvfp
  • [12:26:08] <mru> I call that gcc broken
  • [12:26:08] <jkridner> I still get "arm-angstrom-linux-gnueabi-libtool does not exist". investigating...
  • [12:26:30] <mru> I've been struggling for years to figure out what libtool is supposed to be good for
  • [12:26:36] <koen> aha
  • [12:26:36] <mru> all it's ever done for me is pain
  • [12:26:40] <koen> -mfloat-abi=softfp
  • [12:26:58] <Crofton|work> jkridner, try removing tmp and starting over
  • [12:27:08] <Crofton|work> also what are you building on?
  • [12:27:28] <jkridner> I'm building on Gentoo.
  • [12:27:32] <jkridner> 2007.0, I believe.
  • [12:27:41] <jkridner> many recent updates.
  • [12:27:45] <Crofton|work> I think plenty of people have tried that
  • [12:27:48] <mru> I'm running gentoo on my beagle
  • [12:28:17] <jkridner> that was the first distro I ran on my beagle, but I had headaches with getting X up-and-running.
  • [12:28:33] <mru> I haven't tried that yet...
  • [12:28:48] <jkridner> koen had X up on OE before I could get there, so I've tried to switch.
  • [12:29:20] <jkridner> I'm still interested in getting a good Gentoo build too.
  • [12:30:34] <jkridner> of course, the first X we had up on Beagle was from a guy that never has any time but did it directly, w/o a distro.
  • [12:30:43] <jkridner> that is just annoying because it is so hard to reproduce.
  • [12:31:24] <jkridner> I'm in the process of getting him to move to OE.
  • [12:31:50] <koen> ARMv5TE enabled yes
  • [12:31:50] <koen> ARMv6 enabled yes
  • [12:31:50] <koen> ARM VFP enabled yes
  • [12:31:50] <koen> IWMMXT enabled no
  • [12:31:52] <koen> NEON enabled yes
  • [12:31:55] <koen> there we go
  • [12:32:06] <koen> a case of fixing CFLAGS :)
  • [12:34:13] <jkridner> btw, http://oe.linuxtogo.org/wiki/GettingStarted looks like it has a CR/LF issue.
  • [12:34:35] <Crofton|work> need to pass that to the wiki team :)
  • [12:34:40] <Crofton|work> er website team
  • [12:35:14] <koen> jkridner: http://wiki.openembedded.net/index.php/Getting_Started
  • [12:35:36] <jkridner> much prettier. :_
  • [12:35:38] <jkridner> :)
  • [12:35:57] <Crofton|work> what is amusing is openembedded.org is owned by an ex-to guy :)
  • [12:36:02] <Crofton|work> er ex-ti
  • [12:36:17] <Crofton|work> Hopefully, we get that transfered to the eV soon
  • [12:37:27] <mru> koen: are you using my ffmpeg tree?
  • [12:39:18] <koen> mru: yes
  • [12:40:27] <jkridner> old joke around TI is that it stands for "training institute".
  • [12:40:36] <jkridner> Compaq was started by ex-TI folk.
  • [12:40:52] <jkridner> I don't know all the examples, but I know there are many.
  • [12:41:24] <Crofton|work> http://www.roanoke.com/news/nrv/wb/167451
  • [12:41:37] <Crofton|work> They got their money out of this bridge :)
  • [12:42:34] <jkridner> $469 doesn't go as far as it used to.
  • [12:42:44] <mru> ;-)
  • [12:42:49] <Crofton|work> yeah, that's for sure
  • [12:42:54] <mru> but will get you a few beagle boards
  • [12:43:45] <sakoman> good morning!
  • [12:43:51] <Crofton|work> gm
  • [12:43:57] <Crofton|work> fire update?
  • [12:44:12] <sakoman> pretty much the same
  • [12:44:40] <sakoman> seems they are so swamped with fires there's not much time to update status
  • [12:45:06] <sakoman> The Governator was here yesterday
  • [12:45:31] <Crofton|work> http://money.cnn.com/magazines/fortune/fortune_archive/2006/09/04/8384732/index.htm
  • [12:45:33] <Crofton|work> awesome
  • [12:45:39] <sakoman> Trying to pump up the firefighters and get federal disaster area status
  • [12:45:58] <Crofton|work> last year when we were in .au I saw a paper with a headline refering the governator
  • [12:46:53] <Crofton|work> I saw a really funny photo of the entrance to the governator mansion
  • [12:47:09] <Crofton|work> both "McCain" and "Obama" signs
  • [12:48:37] <sakoman> got to keep the wife happy!
  • [12:49:12] <Crofton|work> yeah, good situation though, connections to both candidates :)
  • [12:49:36] <sakoman> he still can't pronounce California
  • [12:49:42] <Crofton|work> heh
  • [12:50:06] <keesj> mru: funny
  • [12:50:12] <sakoman> ends up sounding something like "cauliflower"
  • [12:50:28] <keesj> oop sorry i was read past log
  • [12:51:23] <sakoman> jkridner: there are a number of OE autobuilder scripts floating around
  • [12:51:43] <jkridner> are those necessary/helpful?
  • [12:51:53] <jkridner> I thought it was already pretty well...
  • [12:51:56] <jkridner> oh... for the server.
  • [12:51:59] <sakoman> I find them amazingly usefull
  • [12:52:09] <jkridner> can you recommend one?
  • [12:52:24] <jkridner> I'd shop around, but I wouldn't know what to look for yet.
  • [12:52:48] <Crofton|work> jkridner, you should get a manual build to succeed first
  • [12:52:59] <sakoman> Good advice :-)
  • [12:53:08] <jkridner> :)
  • [12:53:18] <Crofton|work> cart, horse
  • [12:53:27] <sakoman> You will likely have to modify any you find to suit your purposes
  • [12:53:31] <jkridner> on task 89 out of 4815. :)
  • [12:53:51] <Crofton|work> are you running || make?
  • [12:53:52] <sakoman> There is one in the contrib directory if I recall correctly
  • [12:53:59] <jkridner> I must do something while waiting on the machine. :)
  • [12:54:06] <Crofton|work> :)
  • [12:54:08] <sakoman> I'd be happy to send you mine when you get to that point
  • [12:54:15] <jkridner> Crofton|work: no.
  • [12:54:48] <sakoman> || make makes a huge difference in my experience
  • [12:55:07] <jkridner> k. when there is one ready that folks would like to see hosted, I'd be anxious to do that.
  • [12:55:11] <jkridner> even on a single CPU machine?
  • [12:55:20] <sakoman> jkridner: yes
  • [12:55:42] <sakoman> and really big time on multi-core
  • [12:56:43] <cwz> thanks for entering the 'issue' for me, but I would call it more of a defect than an enhancement
  • [12:56:51] <cwz> er, that's wrt to the power supply
  • [12:56:55] <sakoman> jkridner: my autobuilder is doing several builds per day for beagle and evm with multiple images and a huge list of packages
  • [12:56:59] <jkridner> just 'PARALLEL_MAKE = "-j 1"?
  • [12:57:10] <sakoman> I use 4
  • [12:57:15] <jkridner> k.
  • [12:58:06] <sakoman> to be precise, I use:
  • [12:58:08] <sakoman> PARALLEL_MAKE = "-j 4"
  • [12:58:09] <sakoman> BB_NUMBER_THREADS = "4"
  • [13:02:53] <sakoman> jkridner: I don't get video out on my beagleboard. tried 2 different monitors
  • [13:03:16] <jkridner> I believe it is currently hardcoded to 1024x768.
  • [13:03:29] <sakoman> could there be some sensitivity to the hdmi/dvi adapter brand?
  • [13:03:45] <jkridner> possible, but I haven't seen it before.
  • [13:03:53] <koen> sakoman: I think we use a weird refresh rate due to the kernel only doing 48Mhz for pixclock
  • [13:04:09] <sakoman> Tried a Viewsonic and an Apple 20"
  • [13:04:12] <jkridner> I have seen TVs that didn't work with the precompiled video settings and some monitors that didn't handle certain resolutions.
  • [13:04:22] <mru> if anyone needs I have the hdmi video timing specs here
  • [13:04:34] <sakoman> The Viewsonic is amazingly adaptable
  • [13:04:49] <sakoman> mru: pointer?
  • [13:04:58] <sakoman> Time to bust out the scope again
  • [13:05:40] <sakoman> The beagle is getting much more reliable with git kernel, but i'm still nowhere near happy with the reliability
  • [13:05:47] <cwz> jkridner: I'm surprised that there weren't I2C pull ups on 2 of the busses -- I2C2 and I2C3
  • [13:06:03] <mru> what kind of resolution/refresh rate are you looking to get?
  • [13:06:04] <cwz> That's just not right
  • [13:06:05] <sakoman> There are still broken boots with i2c1 timeouts
  • [13:06:13] <jkridner> cwz: I believe the point was to provide the most flexibility to the expansion board.
  • [13:06:48] <cwz> It's not going to hurt anything in terms of them being GPIOs
  • [13:06:49] <sakoman> mru: anything standard to start with :-) 1024x768, 640 x 480 :-0
  • [13:06:49] <jkridner> I believe that I2C3 has the pull-ups in the driver, but I haven't looked myself.
  • [13:07:07] <cwz> jkridner: I saw that note, but I don't understand how that would work
  • [13:07:08] <mru> those are not in the hdmi spec
  • [13:07:17] <cwz> It would violate basic bus integrity AFAIK
  • [13:07:26] <cwz> The 'drivers' all expect open drain
  • [13:07:43] <sakoman> jkridner: in my latest u-boot I enabled the internal pullups on all i2c channels
  • [13:07:48] * jkridner opens up the HWRM.
  • [13:08:14] <cwz> I'm sure there are internal pullups, but I wouldn't rely on those
  • [13:08:22] <cwz> They're just usually to keep pins from floating/oscillating
  • [13:08:23] <sakoman> No more long timeout errors, but you do get (quick) arb errors
  • [13:08:42] <sakoman> cwz: external pullups remove all errors
  • [13:08:49] <cwz> sakoman: As expected :)
  • [13:09:20] <sakoman> yes, but the hw is what it is. this makes the best of the situation
  • [13:09:34] <cwz> Should be fixd for Rev C though
  • [13:09:47] <cwz> These boards are relatively cheap to make
  • [13:10:09] <cwz> SW workarounds for easily fixable HW bugs are lame
  • [13:10:41] <sakoman> remember this is a skunkworks project -- I think the beagle team has done great work on a shoestring!
  • [13:10:49] <cwz> If the schematics, layout, and gerbers were available, we can throw some $$$ at it
  • [13:11:08] <sakoman> I know how hard it is to get something like this done in a big company
  • [13:11:10] <cwz> Well, don't really need the gerbers per se
  • [13:11:31] <cwz> Gerard (TI OMAP GM) was in the other week and told me he's 100% behind this
  • [13:11:40] <cwz> This is not a huge resource consumer though
  • [13:13:03] <jkridner> cwz: I'm sure we can share the design files.
  • [13:13:42] <cwz> jkridner: I think I asked before, but just for memory's sake, the layout was done in Allegro, right?
  • [13:13:47] <jkridner> Gerald (not Gerard) has some concern that we share too much before all of the little issues get worked out, to avoid disrupting customers. After all, the point is to provide something that OMAP customers can use.
  • [13:13:51] <jkridner> yes.
  • [13:14:14] <cwz> jkridner: We have an NDA w/ you guys, so I would treat anything received as such until told otherwise
  • [13:14:15] <jkridner> but, we can definitely approach that hurdle.
  • [13:14:52] <jkridner> as long as you know what you are getting, I'm sure Gerald would share.
  • [13:14:59] <jkridner> his e-mail is in the HWRM.
  • [13:15:08] <jkridner> http://www.beagleboard.org/uploads/Beagle_HW_Reference_Manual_A_5.pdf
  • [13:15:24] <sakoman> mru: I gues the hdmi timing spec isn't really the issue since we immediately transition to dvi with the adaptor
  • [13:16:01] <jkridner> It is a requirement of the Beagle project that all of the design files get shared.
  • [13:16:22] <mru> sakoman: someone mention picky TVs
  • [13:16:32] <jkridner> the schematics and layout are available in PDF form today, though a tiny bit stale (too busy working on fixes).
  • [13:16:40] <sakoman> I'll fiddle with the fb timing setup to try to get to something more standard
  • [13:18:15] <sakoman> koen: what type of monitor do you use?
  • [13:22:17] <koen> sakoman: IIyama AU5131
  • [13:22:30] <sakoman> is that crt or lcd?
  • [13:23:03] <koen> tft
  • [13:23:06] <jkridner> sakoman: are you using a u-boot that outputs the splashscreen?
  • [13:23:34] <jkridner> I believe the splashscreen patch for u-boot outputs 1280x1024.
  • [13:23:56] <koen> 1280x1024 is such a annoying resolution
  • [13:24:08] <sakoman> I just tried that and get nothing there either :-( So it likely isn't timing -- perhaps the adapter??
  • [13:24:13] <koen> it's not 4:3 aspect ration
  • [13:24:34] <koen> sakoman: do your screens detect input?
  • [13:25:06] <sakoman> the Viewsonic seems to -- it's status light switches from yellow to green
  • [13:25:14] <sakoman> but the screen stays black
  • [13:25:41] <sakoman> the Apple monitor doesn't give too much indication
  • [13:27:48] <jkridner> koen: I'd be happy to see it changed.
  • [13:28:15] <jkridner> in fact, the monitor I use the most with Beagle is only 1024x768 (picked it up for $99).
  • [13:29:29] <sakoman> I think i may start by running into town and picking up a different brand of adapter
  • [13:30:08] <sakoman> After eliminating that as the source of the issue then I'll look at timing
  • [13:32:13] <sakoman> koen: do you use an adapter cable or a one piece molded adapter
  • [13:32:24] <shoragan> jkridner, are the board already on the way to the winners? :)
  • [13:34:24] <jkridner> 9 went out on Friday.
  • [13:34:39] <jkridner> when they are picked up, I expect to get tracking numbers to share.
  • [13:34:48] <jkridner> I don't know which 9, but most were in Germany.
  • [13:35:29] <shoragan> great :)
  • [13:35:37] * shoragan is eager to start working
  • [13:36:09] * jkridner really, really hates international shipping.
  • [13:36:37] * jkridner cannot wait until we have boards at the distributor.
  • [13:37:33] <jkridner> well, another morning has gone by where I got almost nowhere on many parallel tasks, but now it is time for hockey.
  • [13:41:02] <koen> sakoman: cable
  • [13:41:44] <koen> sakoman: it was ???4 or something, so I suspect it's completely passive
  • [13:42:54] <koen> (if "active" ones actually exist)
  • [13:43:09] <mru> there is no need
  • [13:43:12] <cwz> I see that the TXS0102 does indeed contain 10K pull ups
  • [13:43:18] <mru> hdmi and dvi are electrically the same
  • [13:43:20] <cwz> That sould be plenty for I2C
  • [13:43:53] <cwz> HDMI requires HDCP though AFAIK
  • [13:44:20] <mru> hdcp is an addon that works on both hdmi and dvi
  • [13:44:36] <cwz> Yes, but I believe HDMI requires it whereas DVI does not
  • [13:44:47] <cwz> Hence why HDMI isn't supported
  • [13:45:14] * mru checks the hdmi spec
  • [13:45:34] <mru> hdcp is recommended
  • [13:45:38] <mru> not required
  • [13:45:42] <cwz> ok
  • [13:45:56] * cwz stands corrected
  • [13:46:26] <mru> that's hdmi 1.2a
  • [13:49:00] <cwz> http://www.hdmi.org/learningcenter/kb.aspx?c=16#83
  • [13:51:08] <cwz> So who knows, maybe the HDMI inteface will work
  • [13:51:10] <cwz> *shrug*
  • [13:51:32] <mru> the trouble is the timing parameters
  • [13:51:43] <mru> many TVs refuse anything that isn't on the list
  • [13:51:47] <mru> and the list is short
  • [13:53:24] <sakoman> heading into town now. I'll pick up an adapter cable & see where that gets me.
  • [13:53:26] <sakoman> later!
  • [13:55:27] <cwz> BTW, w/ regards to the earlier comments of 'RobotGuy' -- I can't understand what the nonsense was all about
  • [13:55:37] <cwz> CodeSourcery tool chains aren't non 'pure'
  • [13:55:57] <cwz> They are as open source as anything else -- they come with full source
  • [13:56:14] * Crofton|work checks "Giving cat's fluids with needles" of his list
  • [13:56:32] <Crofton|work> koen, any soreness from getting hit?
  • [13:56:38] <Crofton|work> sometimes day 2 is worse
  • [14:01:14] <koen> Crofton|work: a bit sore, but that might also be from managing the end-of-year thing at the uni the 15 hours before the accident
  • [14:01:42] <Crofton|work> end-of-the year thing?
  • [14:01:51] <koen> cwz: I only use CSL toolchains I have built myself from source :)
  • [14:01:53] <Crofton> serving beer again?
  • [14:02:11] <koen> I'm in the technical department
  • [14:02:22] <koen> rigging, bands, etc
  • [14:03:20] <Crofton> cool
  • [14:03:36] <Crofton> My FPGA knowledgable friend is an "Actor"
  • [14:03:41] <koen> I was getting some stuff from home to build a mini-jack -> XLR cable for the DJs that only use a single laptop
  • [14:03:51] <likewise> Crofton|work: the last my girlfriend used needles on little cats she ended up in the hospital for being attacked by our cat, who wanted to protect the yelling/mi-awing little kittens.
  • [14:04:04] <Crofton> ouch
  • [14:04:27] <Crofton> ok, I have text login on my beagle via monitor
  • [14:04:43] <koen> Crofton|work: yeah, I fixed up textmode last week
  • [14:04:46] <Crofton> what do I do know?
  • [14:04:54] * Crofton looks for usb kb
  • [14:04:58] <koen> type "root" :)
  • [14:06:32] * Crofton types root
  • [14:07:03] <Crofton> how do I start X?
  • [14:07:07] <Crofton> startx?
  • [14:07:11] <Crofton> telinit 6?
  • [14:07:59] * Crofton ssh's to another machine
  • [14:10:10] * Crofton observes he has not copied demo-image to the flash
  • [14:10:25] <koen> Crofton|work: X should start automagically
  • [14:10:30] * Crofton needs another monitor
  • [14:10:46] <Crofton> koen, I am an idiot
  • [14:11:07] <Crofton> my USB NIC disconnects and reconnects frequently
  • [14:11:08] <DJWillis> nanpa
  • [14:11:14] <Crofton> hmmm
  • [14:11:15] <DJWillis> ooops
  • [14:11:41] <DJWillis> wrong window (and thinking the PC was locked on screen off ;-))
  • [14:11:52] <Crofton> I need another monitor since my card reader is on the machine whose monitor I am using
  • [14:14:44] * Crofton|work (n=balister@66-207-66-26.black.dmt.ntelos.net) Quit (Remote closed the connection)
  • [14:15:10] <Crofton> I am going to end up with a widescreen monitor, which I hate
  • [14:22:47] * Crofton|work (n=balister@66-207-66-26.black.dmt.ntelos.net) has joined #beagle
  • [14:22:51] <Crofton|work> x
  • [14:27:27] * esden is now known as esden`away
  • [14:36:02] <Crofton> Demo Image is booting very slowly
  • [14:36:07] <Crofton> I do have display
  • [14:42:13] <Crofton> hmmm, screen went blank?
  • [14:42:39] <Crofton> I had Angstrom progress screen, then screen balnkes
  • [14:42:52] <Crofton> moving stuff does not get it back
  • [14:43:05] <Crofton> serial port is still showing
  • [14:45:17] * Crofton runs firefox on the beagle
  • [14:48:15] <koen> Crofton: that's not firefox :)
  • [14:48:53] <koen> some recent images are at http://www.angstrom-distribution.org/unstable/autobuild/beagleboard/
  • [14:49:42] <koen> Crofton: http://gitweb.openembedded.net/?p=org.openembedded.dev.git;a=blob;f=packages/angstrom/epiphany-firefox-replacement.bb;h=2ced2c9c06ac6862b046a935d798fc240f849d7d;hb=HEAD
  • [14:51:24] <Crofton> heh
  • [14:51:31] <Crofton> the button said firefox
  • [14:53:24] <koen> I know
  • [14:53:41] <koen> e17 doesn't check if firefox is actually installed and always has a firefox button
  • [14:59:20] <Crofton> sakoman, demo-image runs fine for me
  • [14:59:54] <Crofton|work> http://www.flickr.com/photos/32615155@N00/2617726155/
  • [15:06:01] <Crofton> I got noise from the speakers
  • [15:06:37] <Crofton> and seeing some oopese
  • [15:07:12] <Crofton> off to do some real world things
  • [15:10:11] * esden`away is now known as esden
  • [15:41:45] * DaQatz (n=db@c-66-30-48-54.hsd1.nh.comcast.net) Quit (Read error: 110 (Connection timed out))
  • [15:51:00] * BThompson (n=BThompso@cpe-76-185-93-11.tx.res.rr.com) has joined #beagle
  • [16:01:55] * Olipro (n=Olipro@unaffiliated/olipro) has joined #beagle
  • [16:15:09] * intu1 (n=intu@ubr1-201-91.ubr.tartu.stv.ee) Quit (Read error: 104 (Connection reset by peer))
  • [16:15:10] * intu (n=intu@ubr1-201-91.ubr.tartu.stv.ee) has joined #beagle
  • [16:28:18] * JoeBorn_ is now known as JoeBOrn
  • [16:28:24] * JoeBOrn is now known as JoeBorn
  • [17:12:53] * calculu5 is now known as calculus
  • [17:22:10] * prpplague^2 is now known as prpplague
  • [17:34:02] * Olipro_ (n=Olipro@unaffiliated/olipro) has joined #beagle
  • [17:37:58] * Olipro (n=Olipro@unaffiliated/olipro) Quit (Read error: 104 (Connection reset by peer))
  • [17:49:38] * khasim (n=a0393720@192.163.20.231) has joined #beagle
  • [18:15:31] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
  • [18:39:05] * khasim (n=a0393720@192.163.20.231) has joined #beagle
  • [18:45:38] <koen> mru: what configure options do you pass to ffmpeg configure?
  • [18:47:44] <mru> koen: --cross-prefix=arm-omap3-linux-gnueabi- --arch=arm --cpu=cortex-a8 --disable-encoders --enable-encoder=rawvideo
  • [18:47:56] <mru> replace cross-prefix as appropriate for your system
  • [18:48:49] * JoeBorn (n=jborn@adsl-75-2-240-30.dsl.chcgil.sbcglobal.net) Quit ("rebootus maximus")
  • [18:58:14] <mru> moving data from neon to arm is annoyingly slow
  • [18:59:03] * likewise (n=chatzill@82-171-51-231.ip.telfort.nl) Quit ("ChatZilla 0.9.83 [Firefox 3.0/2008061015]")
  • [19:02:56] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
  • [19:18:28] <Crofton|work> sakoman, ping
  • [19:25:08] <koen> mru: what's the deal with swscale?
  • [19:25:14] <mru> deal?
  • [19:25:44] <Crofton|work> anyone know if sakoman solved his monitor issue?
  • [19:26:13] <koen> I can't enable it, and typing 'make' inside libavcodedec/ uses swscale regardless of its enabled/disabled state
  • [19:26:13] <sakoman> Crofton: pong
  • [19:26:40] <sakoman> no, local store didn't have adapter cable
  • [19:26:48] <sakoman> so still no monitor
  • [19:26:56] <mru> koen: can you give more details? libavcodec doesn't use swscale at all
  • [19:27:23] <Crofton|work> I can ship you one
  • [19:27:35] <Crofton|work> our local PC store had just the right cable
  • [19:27:57] <koen> mru: http://rafb.net/p/QttMMf99.txt
  • [19:28:53] <koen> I don't care about swcale, but all the error about missing .h files when running configure makes me think someone did a half-assed commit to remove it
  • [19:29:20] <mru> oh, you're using my git tree, right?
  • [19:29:39] <koen> yes
  • [19:29:41] <koen> License: unredistributable
  • [19:29:41] <koen> Creating config.mak and config.h...
  • [19:29:41] <koen> grep: /OE/angstrom-tmp/work/armv7a-angstrom-linux-gnueabi/ffmpeg-0.4.9+r0+gitr68893a3bf603de31c100fd39306b35e984292605-r0/git/libswscale/swscale.h: No such file or directory
  • [19:29:56] <mru> well, for "historical reasons" libswscale is in a separate repo
  • [19:30:24] <mru> do a git clone git://git.mplayerhq.hu/libswscale in your ffmpeg dir to fetch it
  • [19:30:26] <koen> the build stops dead inside libavcodec without an error and typing 'make' again results in no output
  • [19:31:12] <sakoman> Crofton: I have an adapter I bough online. Not sure why it doesn't work!
  • [19:31:39] <sakoman> I was just going to try a different monitor
  • [19:32:11] <Crofton|work> ok
  • [19:32:32] <Crofton|work> It's beginnning to look like it would be handy to have another monitor for the beagle
  • [19:33:14] <koen> I wish the s-video work work with linux-omap git
  • [19:33:17] <sakoman> Mine is a one piece moulded thingy with male hdmi on one side and female dvi on the other
  • [19:33:28] <sakoman> Would seem to be just what is needed
  • [19:35:03] <Crofton|work> same here
  • [19:35:47] <Crofton|work> if I run out and look into monitors, I'll see if I can get the manufacturer info
  • [19:40:09] * esden is now known as esden`away
  • [19:45:25] <koen> mru: doesn't git support something like 'svn externals'?
  • [19:45:47] <mru> maybe it does now
  • [19:46:19] <koen> I remember linus mentioning that is was "being worked on" in his google presentation
  • [19:46:47] <mru> I haven't had time and motivation to investigate
  • [19:46:50] <sakoman> Crofton: rats, that monitor doesn't work either
  • [19:47:02] <sakoman> time to looks at the signals with a scope
  • [19:47:33] <koen> mru: anyway, checking out libswscale helped
  • [19:48:58] <koen> Ugly OE recipe for it: http://amethyst.openembedded.net/~koen/ffmpeg_git.bb
  • [19:49:58] <mru> you're enabling an awful lot of stuff
  • [19:50:43] <koen> it's what was there for the svn version
  • [19:50:56] <mru> ok
  • [19:51:10] <mru> but using liba52 with ffmpeg is discouraged
  • [19:51:18] <mru> there is a builtin ac3 decoder
  • [19:51:27] <mru> and vorbis
  • [19:54:09] <koen> I remember reading something about liba52 breaking stuff with ffmpeg
  • [20:00:43] <sakoman> Crofton: interesting. don't see any video signals at the connector. a couple pins are 5V, the rest are all gnd
  • [20:01:31] <Crofton|work> thats no good
  • [20:01:48] <Crofton|work> Do you want me to get you a cable?
  • [20:02:02] <Crofton|work> Should I break down and get a widescreen monitor?
  • [20:02:36] <sakoman> no, I think it wouldn't do any good since there is nothing on my hdmi connector
  • [20:02:48] <sakoman> need to figure out why
  • [20:03:53] <Crofton|work> I just loaded demo-image and it worked
  • [20:13:07] * esden`away is now known as esden
  • [20:18:15] <sakoman> Crofton: does your 4030 get too hot to touch?
  • [20:18:47] <Crofton|work> the big chip by the connector?
  • [20:18:53] <Crofton|work> my beagle has been off
  • [20:19:26] <Crofton|work> I turned it on
  • [20:19:30] <mru> the 4030 is the small chip
  • [20:19:49] <mru> on the opposite side of the cpu from the hdmi connector
  • [20:20:32] <mru> and mine is well above room temperature
  • [20:20:37] <mru> but not too hot to touch
  • [20:20:47] <Crofton|work> warm atm
  • [20:20:50] <sakoman> mine reads 150F
  • [20:20:58] <Crofton|work> small bga near CPU?
  • [20:21:02] <sakoman> just curious
  • [20:21:14] <Crofton|work> warm
  • [20:21:21] <sakoman> yes, near poer jack
  • [20:21:24] <sakoman> power
  • [20:21:56] <mru> I'd estimate mine at maybe 50C
  • [20:22:09] <mru> I don't have an IR camera...
  • [20:22:37] <sakoman> I get no signals on my hdmi connector -- the only pins that have anything other than ground are the 5v pins
  • [20:24:58] <sakoman> well actually the 5v pin and the scl and sda pins
  • [20:25:11] <sakoman> so perhaps my board is broken
  • [20:25:36] <sakoman> need to try the old u-boot one more time to see if it produces any signals
  • [20:25:43] <Crofton|work> inspect the connector attachment to the board?
  • [20:26:10] <sakoman> looks fine, i probed at the point where the pins solder to the board
  • [20:26:40] <sakoman> they match the signals at the other side of the adapter, so i'm pretty sure the adapter is good
  • [20:38:32] <sakoman> Crofton: no signals on the connector with u-boot 1.1.4 (which has splash screen and audio tone)
  • [20:38:49] <sakoman> my board may have broken video out :-(
  • [20:41:55] <mru> damn it, crashed again
  • [20:51:45] <sakoman> Crofton: on your board are the two components loaded which are directly above the U8 silkscreen
  • [20:58:40] <sakoman> I've got empty spots there
  • [21:03:50] <mru> I have two pairs of solder pads there
  • [21:05:01] <jkridner> they should be empty.
  • [21:05:17] <mru> what are they for?
  • [21:06:56] <sakoman> just searching for a reason why I'm getting no video :-(
  • [21:07:34] <sakoman> jkridner: any ideas on what might be wrong?
  • [21:08:07] <jkridner> no, not if the 1.1.4 u-boot image doesn't cause some signal wiggling.
  • [21:08:24] <jkridner> EDID reads work fine, right?
  • [21:09:35] <sakoman> jkridner: haven't tried the EDID reads
  • [21:09:42] <sakoman> how does one do that?
  • [21:09:49] <jkridner> it has been so long since I really did hardware.... which direction do the pin numbers go? clockwise or counter-clockwise?
  • [21:11:01] <jkridner> I believe koen put something in the kernel to spit out EDID values. There is also a way to do it using u-boot via 'ibus'. I believe the easiest way, however, was described by dirk2 via the mailing list....
  • [21:11:14] <koen> sakoman: check /sys/bus/i2c
  • [21:11:43] <jkridner> he just uses a recent kernel's ability to read the i2c bus and apparently there is an 'eprom' entry at the path that koen just mentioned.
  • [21:11:52] <jkridner> just 'cat' it.
  • [21:12:20] <koen> sakoman: http://rafb.net/p/WLUSRu26.txt
  • [21:13:15] <koen> it should be on 3-0050 (bus 3, device 50)
  • [21:13:20] <jkridner> BTW, the "DNI" on the schematics indicates what devices aren't populated.
  • [21:14:56] <jkridner> R103, R104, and R105 are all "DNI".
  • [21:17:22] <koen> jkridner: I just used the builtin i2c support in the kernel :)
  • [21:17:37] <jkridner> it seems like the EDID value should be a bit more readable, but it looks like you'd need to parse it with something. could probably try parse-edid
  • [21:17:42] <koen> jkridner: the kernel patch was an idea to make omapfb smarter
  • [21:17:50] <jkridner> k
  • [21:18:08] <koen> parse-edid is a perlscript, and hence needs some megs of perl support
  • [21:18:25] <sakoman> koen: root@beagleboard:/sys/bus/i2c/devices# ls .
  • [21:18:25] <sakoman> 1-0048 1-0049 1-004a 1-004b
  • [21:18:40] <sakoman> so it doesn't seem to be working
  • [21:18:43] <koen> so nothing on bus 3
  • [21:18:50] <jkridner> ah. well, I've been using perl and rather like having it, even on beagle.
  • [21:19:30] <koen> it's available
  • [21:19:37] * jkridner makes a run to Lowes/HomeDepot for a new ceiling fan.
  • [21:19:40] <koen> I've used parse-edid on the beagle :)
  • [21:19:43] <sakoman> koen: I always get this on boot on my board: i2c_omap i2c_omap.3: Arbitration lost
  • [21:20:09] <sakoman> I don't think you get that right?
  • [21:20:25] <koen> nope, only bus 2 is troublesome
  • [21:20:51] <sakoman> have you tried my u-boot git?
  • [21:21:10] <koen> I think I'm on r1 of your uboot
  • [21:21:13] <sakoman> bus 2 timeouts go away even without external pullup mod
  • [21:22:43] <sakoman> well, enough playing around with video. time for something less frustrating :-)
  • [21:22:52] <koen> ASoC?
  • [21:23:11] <sakoman> yeah, I've been working on that in the background
  • [21:23:33] <sakoman> Have you had any issues with the current version?
  • [21:23:42] * koen notes the National Geographic has a forest fire front page
  • [21:24:01] <sakoman> The air is *really* bad today
  • [21:24:16] <sakoman> can't go outside without burning eyes and lungs
  • [21:25:06] <sakoman> I stopped at the farmers market this morning and had to crack up at a sign that they always have at the entrance" "This is a smoke free event"
  • [21:25:23] <koen> heh
  • [21:25:36] <sakoman> This when you can't see 40 meters in front of your face for all the smoke!
  • [21:26:05] * Olipro_ is now known as Olipro
  • [21:26:07] <koen> in 3 days the smoking ban for restaurants and bars kicks in in .nl
  • [21:26:32] <sakoman> asoc has been reliable for me so I've been working on cleanup and the power managment hooks
  • [21:27:00] <sakoman> let me know if you run into any issues with the cureent version and I'll try to fix them while I am at it
  • [21:27:58] <koen> I can check tomorrow after I'm done booting up my avr32 boards
  • [21:28:24] <sakoman> ok
  • [21:28:27] <sakoman> ttly!
  • [22:00:43] * DaQatz (n=db@c-66-30-48-54.hsd1.nh.comcast.net) has joined #beagle
  • [22:08:54] * bazbell (n=a0192809@12.165.163.171) has joined #beagle
  • [22:40:43] <mru> why would a vext instruction take silly amounts of time?
  • [22:54:13] <ldesnogu_> mru: it should not
  • [22:54:29] <mru> maybe oprofile is lying
  • [22:54:32] <ldesnogu_> are you sure it's the vext instruction itself that's slow?
  • [22:54:42] <mru> I'm not sure of anything
  • [22:55:56] <ldesnogu_> I guess vext is part of a function that oprofile showed you
  • [22:56:15] <mru> I used opreport -d
  • [22:56:27] <mru> which should give a per-instruction breakdown
  • [22:56:32] <ldesnogu_> I don't know anything about oprofile :/
  • [22:56:36] <mru> I guess it's slightly inaccurate
  • [22:56:56] <mru> there's a load nearby that would be more likely to stall
  • [22:57:23] <mru> and now my kernel oopsed
  • [22:57:25] <mru> again
  • [22:57:26] <ldesnogu_> I you want to know if the vext itself is killing you, put another dumb instruction
  • [22:57:29] <ldesnogu_> arf
  • [22:58:47] <ldesnogu_> wish you the best; time for me to sleep :)
  • [22:58:52] * ldesnogu_ (n=ldesnogu@ven06-2-82-247-86-183.fbx.proxad.net) Quit ()
  • [23:11:01] * Crofton|work is done injecting the cat for the day ....
  • [23:11:16] <mru> how's he doing?
  • [23:11:21] <Crofton|work> better
  • [23:11:28] <Crofton|work> holds his head up
  • [23:11:35] <mru> that's good
  • [23:11:37] <Crofton|work> no idea what the long term outlook is
  • [23:11:46] <mru> what's wrong with him?
  • [23:11:53] <Crofton|work> he got really fehydrated and has bad kidneys
  • [23:12:02] <Crofton|work> the kidneys will get him in the end
  • [23:12:09] <Crofton|work> just a question of when :(
  • [23:12:21] <mru> yes, that's a weakness in cats
  • [23:12:31] <Crofton|work> hopefully he stabilizes and we can et him off fluids
  • [23:12:55] <Crofton|work> I hate needles, stinking them in him makes my skin crawl
  • [23:13:02] <Crofton|work> he is really good about it
  • [23:13:29] <Crofton|work> gonna go be ncie to him for a bit