• [00:12:49] <mru> hey guys, does someone have a minute for a stupid question?
  • [00:13:34] <mru> my board is booting, but it doesn't seem to listen to the serial port
  • [00:13:55] <mru> I get normal-looking output, but typing has no effect
  • [00:14:09] <mru> I'm probably missing something trivial...
  • [00:16:15] <Crofton> what is normal looking output?
  • [00:16:29] <BThompson> not sure what it is, but you could try a different cable, or a different telnet client
  • [00:16:47] <BThompson> yea by normal looking do you mean you can see UBoot?
  • [00:16:56] <Crofton> it sounds like you are missiung the data to beagle wire ....
  • [00:17:03] <Crofton> gotta run dinner time here
  • [00:19:38] <mru> I see uboot stuff, and then it boots the linux kernel
  • [00:26:56] <ds2> check your flowcontrol settings
  • [00:27:02] <ds2> think it should be none
  • [00:28:57] * ldesnogu (n=ldesnogu@fw-tnat.cambridge.arm.com) Quit (Read error: 113 (No route to host))
  • [00:30:14] * ldesnogu (n=ldesnogu@fw-tnat.cambridge.arm.com) has joined #beagle
  • [00:31:45] <mru> ah, that did it
  • [00:31:58] <mru> I knew it would be something trivial
  • [00:46:14] <mru> now on to the fun stuff...
  • [00:46:51] <mru> is there any good summary of driver support and such somewhere?
  • [00:54:45] * bazbell (n=a0192809@nat/ti/x-5a1e72439940cb2c) has joined #beagle
  • [01:39:49] * khasim (n=a0393720@192.163.20.231) has joined #beagle
  • [02:44:53] * aleij (n=ad@adsl-84-226-97-190.adslplus.ch) Quit ("done!")
  • [03:09:53] * BThompson (n=BThompso@cpe-76-185-93-11.tx.res.rr.com) Quit ("Trillian (http://www.ceruleanstudios.com")
  • [03:41:46] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
  • [03:54:57] <jkridner> mru, did you get your serial port issues resolved?
  • [03:55:14] <jkridner> typical problem is incorrectly wired adapter.
  • [03:55:21] <jkridner> oh...
  • [03:55:24] <jkridner> I should read better.
  • [03:55:29] <jkridner> issue was flowcontrol.
  • [04:09:45] <ds2> yipeee signs of sanity... the various uboots may become one once again
  • [05:17:01] * khasim (n=a0393720@192.163.20.231) has joined #beagle
  • [05:29:09] * bazbell (n=a0192809@nat/ti/x-5a1e72439940cb2c) Quit (Remote closed the connection)
  • [05:36:24] * bazbell (n=a0192809@nat/ti/x-4fc7f1b965916b23) has joined #beagle
  • [07:39:30] * Guizmo` (n=guizmo@41.92.198-77.rev.gaoland.net) Quit (Read error: 113 (No route to host))
  • [08:40:32] * Guizmo` (n=guizmo@81.255.46.1) has joined #beagle
  • [08:44:16] * Guizmo_ (n=guizmo@81.255.46.1) has joined #beagle
  • [08:51:47] <koen> good morning all
  • [08:59:14] <ldesnogu> hi koen
  • [09:02:34] * Guizmo` (n=guizmo@81.255.46.1) Quit (Read error: 113 (No route to host))
  • [09:09:07] <likewise> morning koen, lal
  • [09:09:09] <likewise> all
  • [09:12:03] * Guizmo_ (n=guizmo@81.255.46.1) Quit (Read error: 113 (No route to host))
  • [09:39:02] * Guizmo` (n=guizmo@81.255.46.1) has joined #beagle
  • [10:15:40] * Guizmo` (n=guizmo@81.255.46.1) Quit (Read error: 113 (No route to host))
  • [10:20:44] * Guizmo` (n=guizmo@81.255.46.1) has joined #beagle
  • [10:48:25] * Guizmo` (n=guizmo@81.255.46.1) Quit (Read error: 113 (No route to host))
  • [10:53:08] * chadnickbok (n=Owner@koala.ozonline.com.au) has joined #beagle
  • [11:18:42] <chadnickbok> hrm, so the beagle has a tv-out-ish port? Is it just a standard VGA out?
  • [11:19:49] <nomis> chadnickbok: there is a DVI output and a S-Video output, the fate of the latter apparently is not set in stone yet.
  • [11:20:19] <nomis> the DVI output is somehow in an HDMI-plug (for the small form factor) and you need an adaptor to plug it into a regular LCD.
  • [11:28:18] <chadnickbok> Hrm, an hdmi plug hey? Thats _awesome_
  • [11:29:56] <theoddbot> yeah dvi-d only
  • [11:30:09] <theoddbot> nomis: why not set in stone ? :(
  • [11:34:07] <nomis> theoddbot: from what I understood on the Linuxtag booth they are pondering to replace it with an ethernet jack.
  • [11:34:31] <nomis> (which would be beneficial for me)
  • [11:34:41] <theoddbot> hmm that could be annoying, since i have a nice little s-video lcd i snagged on ebay
  • [11:34:45] <theoddbot> oh well :)
  • [11:35:12] <chadnickbok> Hrm, I must say, an ethernet connection would be pretty hot
  • [11:35:21] <chadnickbok> But poor little lcd :(
  • [11:35:26] <nomis> theoddbot: they apparently don't really know yet - an ethernet jack probably is a bit more expensive and the 149$ *is* set in stone :)
  • [11:35:43] <theoddbot> yeah ethernet + phy is annoying
  • [11:39:26] <Crofton|work> ok the vfp unit only works on floats and NEON adds integer data types
  • [11:40:59] <koen> doesn't neon to float simd as well?
  • [11:41:07] <Crofton|work> yes
  • [11:41:22] <Crofton|work> float and int SIMD
  • [11:41:32] <Crofton|work> but vfp is absicaly simb also
  • [11:42:49] <khasim> You can still get ethernet of USB OTG port using USB to ethernet dongle
  • [11:43:54] <Crofton|work> hmm, maybe the vfp unit does not do vector ops ...
  • [11:44:16] <jkridner> hi khasim. good morning all.
  • [11:44:50] <theoddbot> khasim: or wifi :)
  • [11:47:32] <Crofton|work> anyone know if the A8 contains a vfp9, 10, or 11?
  • [11:48:25] <jkridner> chadnickbok: tv-out-port is S-Video.
  • [11:48:29] <Crofton|work> ah they all do "short vectors"
  • [11:55:08] <jkridner> sakoman: it looks like you've removed your OE for the OMAP35x EVM. too much traffic?
  • [11:57:41] <Crofton|work> jkridner, images?
  • [11:58:04] <jkridner> Crofton|work: I'm not familiar with the 9,10, 11 versions. The VFP instruction set is VFPv3.
  • [11:58:14] <Crofton|work> I think I figured it out
  • [11:58:29] <Crofton|work> I need to talk about neon and vfp
  • [11:58:38] <Crofton|work> and know what the differences are
  • [11:59:06] <Crofton|work> I don't need to be perfect, just need to get some people interested in prototyping some radios on a beagle :)
  • [11:59:52] <jkridner> by "images" you mean my question to sakoman about OE for the OMAP35x?
  • [12:00:13] <jkridner> I'd pointed some TI people there for demos and they found it isn't there anymore.
  • [12:01:13] <Crofton|work> yeah
  • [12:01:29] <Crofton|work> we can see about putting them on amethyst
  • [12:02:47] <chadnickbok> hrm
  • [12:03:09] <chadnickbok> It occurs to me that the processing power of the beagle could land a man on the moon...
  • [12:04:37] <jkridner> hrm?
  • [12:05:49] <Crofton|work> it occurs to me the processing power of an early 80's calcualtor could land a man on the moon :)
  • [12:23:13] <ldesnogu> Crofton|work: I posted several links about VFP and NEON on Cortex-A8 a few days ago; look in the IRC logs
  • [12:23:57] <Crofton|work> infocenter.arm.com? They are open in my browser :)
  • [12:25:30] <ldesnogu> yes :)
  • [12:25:50] <ldesnogu> if that doesn't answer your questions, ask and I'll try to find the answer
  • [12:26:07] <Crofton|work> I think I have enough for tomorrow from these
  • [12:26:20] <Crofton|work> I basically need to explain why SDR people care
  • [12:26:33] <ldesnogu> just remember something : VFP is shitty on Cortex-A8
  • [12:26:42] <ldesnogu> always use NEON unless you need doubles :)
  • [12:26:52] <Crofton|work> ok
  • [12:27:13] <Crofton|work> A2D converters limit the number of bits we have to process
  • [12:27:26] <Crofton|work> saturating 16 bit math covers many apps
  • [12:27:45] <ldesnogu> so you don't even need FP ?
  • [12:27:53] <Crofton|work> can we shut off vfp and neon?
  • [12:27:56] <Crofton|work> not really
  • [12:28:05] <Crofton|work> but it is easier if you can use it :)
  • [12:28:34] <ldesnogu> I think you can power down (or stop the clock) the VFP/NEON block
  • [12:28:44] <Crofton|work> also if you are doing something that needs a lot of dynamic range, float is handy
  • [12:28:54] <Crofton|work> but both go down when you do
  • [12:29:00] <ldesnogu> yes IIRC
  • [12:29:06] <Crofton|work> that makes sense
  • [12:34:42] * chadnickbok (n=Owner@koala.ozonline.com.au) has left #beagle
  • [12:37:56] * ahare (n=ahare@ahare.gotadsl.co.uk) Quit (Read error: 110 (Connection timed out))
  • [12:46:28] <sakoman> jkridner: they were getting a bit old and I didn't realize anyone cared!
  • [12:46:41] <sakoman> I'd be happy to post some fresh ones
  • [12:47:15] <jkridner> love to see them back again. :)
  • [12:49:46] <sakoman> I'm setting up an autobuilder to start building twice daily for evm and beagle
  • [12:50:00] <sakoman> It will automatically post fresh images
  • [12:50:14] <sakoman> I'll post the link once it is going
  • [12:53:18] * bazbell (n=a0192809@nat/ti/x-4fc7f1b965916b23) Quit (Remote closed the connection)
  • [13:04:28] <jkridner> thanks
  • [13:06:53] <Crofton|work> wtf does liboil stand for?
  • [13:07:23] <sakoman> optimize inner loops ??
  • [13:07:29] <Crofton|work> ah
  • [13:07:35] <Crofton|work> they don't say on the wiki
  • [13:07:37] <sakoman> I think
  • [13:08:00] <sakoman> too many libs to keep track of!
  • [13:08:51] <Crofton|work> no, just trying to list some projects that would be useful for SDR work on the beagle
  • [13:09:52] * prpplague (n=dave@mail.americanmicrosystems.com) has joined #beagle
  • [13:14:57] <jkridner> sakoman: apparently http://linux.omap.com/pub/kernel/3530evm/ has the 0.9.7 release. This supposedly has some patches to help with the touchscreen noise issue. I haven't had a chance to try it yet.
  • [13:15:21] <jkridner> I'm still confused why it is there first instead of http://www.ti.com/omapsoftwareupdates.
  • [13:20:34] <kulve> What's http://www.ti.com/omapsoftwareupdates?
  • [13:21:18] <jkridner> when you register the EVM, you are meant to get access to that page.
  • [13:21:29] <kulve> ok
  • [13:21:43] <jkridner> most of the EVM SDK content, however, is open source and available at http://linux.omap.com
  • [13:24:06] * bazbell (n=a0192809@nat/ti/x-2e6929d305b3246c) has joined #beagle
  • [13:41:21] * BThompson (n=BThompso@nat/ti/x-356af043a9181a38) has joined #beagle
  • [13:45:07] <koen> drat, khem's patch introduced a c++ change
  • [13:58:51] <khasim> I had announced the 0.9.7 availability on linux omap vger list (after uploading the same)
  • [13:59:12] <khasim> on linux.omap.com
  • [13:59:51] <jkridner> k. I had never promoted it here or on discussion@beagleboard.org....
  • [14:00:20] <jkridner> ugh, didn't mean to do that (give actual e-mail address). I make hack that in the log to avoid spambots.
  • [14:01:00] <jkridner> I think that there are some features in that kernel that might be useful for people to want to merge into the Beagle kernel.
  • [14:01:02] <koen> sakoman: did disabling the clocks make a differnece with your bench supply?
  • [14:03:31] * bazbell (n=a0192809@nat/ti/x-2e6929d305b3246c) Quit ("Leaving.")
  • [14:05:09] <khasim> sakoman: have you stopped working on ASOC ?
  • [14:05:13] <sakoman> koen: haven't tried that yet. I'll let you know.
  • [14:05:32] <koen> 'm waiting for Kalle to update his smartreflex patch set
  • [14:05:37] <sakoman> khasim: no going back at it today. go t distracted by beagle & u-boot
  • [14:06:24] <khasim> sakoman: If you are working on u-boot then probably I can look at ASOC, but I would seriously want to explore CE/Link and GFX drivers for open community. with sample applications/demos for each of these
  • [14:06:42] <khasim> sakoman, will I be able to pull your u-boot GIT
  • [14:07:06] <sakoman> khasim: at this point uboot will be a background task since it basically works
  • [14:07:13] <sakoman> yes you can pull it
  • [14:07:40] <sakoman> So SoC will move to foreground task
  • [14:08:15] <sakoman> You should go ahead and do the CE/Link GFX stuff
  • [14:08:27] <sakoman> It will make a lot of people happy
  • [14:08:52] <sakoman> I'll ping you if I need help with SoC questions
  • [14:08:52] <khasim> sakoman: what is the link for downloading the u-boot, I would like to add the NAND bad block to it and send just the patch to you
  • [14:09:38] <khasim> I need a http link to pull your u-boot GIT
  • [14:11:21] <sakoman> Well, it will be http://www.sakoman.net/git/u-boot-omap3
  • [14:12:02] <sakoman> But I've got something wrong in the config cause http clones seems to fail
  • [14:12:21] <jkridner> btw, was there no u-boot git for omap already?
  • [14:12:26] <sakoman> Any git http export experts here?
  • [14:13:21] <sakoman> jkridner: I have no idea. I set it up for my own purposes. Never thought to check if one existed
  • [14:13:25] <khasim> Jason, we never had a OMAP specific u-boot git
  • [14:14:55] <jkridner> sakoman: I'm nowhere close to an expert, but i've done it.
  • [14:15:21] <jkridner> magic for me is 'git --bare gc', frequently. :)
  • [14:15:46] <jkridner> Theoretically, it is possible to trigger functions on updates, but I haven't been able to get that to work.
  • [14:16:00] <sakoman> how do you do it if the repo isn't bare?
  • [14:16:42] <jkridner> hmmm.... I'm not sure. Are there more than one .git/ folders in a git respository?
  • [14:16:57] <jkridner> I'm not sure if the format for the .git folder is the same as for a bare repo.
  • [14:17:28] <jkridner> the ones I've shared are bare. you could always just share a bare clone via http.
  • [14:17:34] <sakoman> I believe it is. I tried using git update-server-info and it touched the files that the docs indicate it would touch
  • [14:18:07] <sakoman> But it is not clonable, though the repo files are browwsable
  • [14:18:12] <jkridner> I could try to clone if you share a link to your .git folder.
  • [14:18:30] <jkridner> ? any git repo should be clonable.
  • [14:18:45] <jkridner> do you mean via http?
  • [14:18:49] <sakoman> not clonable via http, git clones work fine
  • [14:19:02] <jkridner> k. try 'git gc'.
  • [14:19:11] <sakoman> ok
  • [14:19:16] <jkridner> 'update-server-info' doesn't seem to do enough.
  • [14:19:47] <Crofton|work> you could talk to wolfgang about an omap maintainer
  • [14:19:59] <Crofton|work> the arm u-boot guy is slow ...
  • [14:20:22] <Crofton|work> I think there is already a specific maintainer for samsung parts
  • [14:21:05] <koen> it would be nice to have dfu support
  • [14:21:25] <jkridner> nice is an understatement.
  • [14:21:26] <koen> I love how fast the n800 can flash a rootfs
  • [14:21:38] <sakoman> jkridner: same issue after git gc
  • [14:21:40] <Crofton|work> dfu?
  • [14:21:47] <koen> like 6MiB/s
  • [14:21:52] <jkridner> device firmware update.
  • [14:21:53] <koen> Crofton|work: device firmware upload profile
  • [14:21:54] <jkridner> it is a usb class.
  • [14:21:59] <Crofton|work> ah
  • [14:22:26] <jkridner> I believe the openmoko guys were the first to put support into u-boot--at least that is where I've seen it.
  • [14:22:41] <koen> Crofton|work: http://wiki.openmoko.org/wiki/Dfu-util
  • [14:23:18] <jkridner> RidgeRun added the support for DM355, but I don't believe they've pushed it upstream yet. I think the hassle for them is the state of the feature in u-boot.
  • [14:23:50] <jkridner> DM355 uses the musb controller, so it may provide a starting point for OMAP, ignoring the DMA.
  • [14:24:24] * DarrenEtheridge (n=a0867391@nat/ti/x-a71b83b68d5ac3fb) has joined #beagle
  • [14:25:18] <jkridner> sakoman: for a bare repo, I don't recall doing other steps than 'git --bare gc', but I might have done something in fumbling around.
  • [14:26:01] <jkridner> sakoman: what error are you getting when you try to clone?
  • [14:26:53] <sakoman> sakoman@tera:~/source$ git clone http://www.sakoman.net/git/u-boot-omap3 test-uboot
  • [14:26:53] <sakoman> Initialized empty Git repository in /home/sakoman/source/test-uboot/.git/
  • [14:26:53] <sakoman> Cannot get remote repository information.
  • [14:26:53] <sakoman> Perhaps git-update-server-info needs to be run there?
  • [14:26:58] <jkridner> the generic, and least useful error, is "Perhaps git-update-server-info needs to be run there?"
  • [14:27:11] <sakoman> There you go :-)
  • [14:27:30] <koen> sakoman: git clone http://www.sakoman.net/git/u-boot-omap3/.git
  • [14:27:34] <koen> that seems to work for me
  • [14:28:03] <koen> koen@lieve:/OE$ git clone http://www.sakoman.net/git/u-boot-omap3/.git
  • [14:28:03] <koen> Initialized empty Git repository in /OE/u-boot-omap3/.git/
  • [14:28:03] <koen> Getting alternates list for http://www.sakoman.net/git/u-boot-omap3/.git
  • [14:28:03] <koen> Getting pack list for http://www.sakoman.net/git/u-boot-omap3/.git
  • [14:28:03] <koen> Getting index for pack c9d5333c68453a637bd2d5a3fbe47aaae5956d36
  • [14:28:05] <koen> etc.
  • [14:28:23] <jkridner> ah, yeah, that error is pretty useless. it could be a bad url, server down, bad repo, and many other things.
  • [14:28:46] <jkridner> clone seems to be working for me as well.
  • [14:29:25] <jkridner> looks like you were missing /.git in your URL.
  • [14:29:34] <sakoman> yeah, works for me as well. that's what I get for trusting the git docss!
  • [14:29:57] <koen> khasim: git clone http://www.sakoman.net/git/u-boot-omap3/.git
  • [14:29:57] <jkridner> in a bare repo, you wouldn't have the .git folder.
  • [14:30:33] <jkridner> shall I mirror this on http://www.beagleboard.org/gitweb ?
  • [14:30:50] <sakoman> khasim: after cloning, check out the test branch, that's where the changes are
  • [14:31:01] <sakoman> jkridner: sure, no problem
  • [14:31:19] <khasim> ok, I am able to clone the git now
  • [14:37:09] * DarrenEtheridge (n=a0867391@nat/ti/x-a71b83b68d5ac3fb) has left #beagle
  • [14:37:44] <sakoman> koen: is it possible to have the oe git fetcher use a branch?
  • [14:37:57] <koen> ;branch=test
  • [14:38:18] <koen> as in: SRC_URI = "git://git.neurostechnology.com/git/linux-davinci-2.6;protocol=git;branch=neuros"
  • [14:38:21] <sakoman> ah, nice and obvious. who would have thought!
  • [14:39:29] <koen> I looked it up last night when working on the neuros kernel recipe :)
  • [14:39:44] <sakoman> It's interesting trying to figure out an efficient workflow using oe and git repo's
  • [14:41:06] <sakoman> Especially when you are doing lots of quick change, rebuild, test cycles
  • [14:41:49] * bazbell (n=a0192809@nat/ti/x-e2bfb64d3e358d6f) has joined #beagle
  • [14:41:51] <khasim> For some reason the git clone is hung.... :(
  • [14:41:52] <koen> I'd do a manual checkout into ${S}
  • [14:41:55] <Crofton|work> I just put the tool chain in my path
  • [14:41:58] <koen> so 'git pull' works
  • [14:42:02] <Crofton|work> and work in local repo
  • [14:42:17] <jkridner> khasim: are you sure it is hung? some of the patch sets are quite large.
  • [14:42:23] <Crofton|work> easy for kernel/u-bbot, since they are not dependent on otherpackages
  • [14:43:51] <khasim> I think it is quite large and not hung
  • [14:43:55] <khasim> will wait
  • [14:44:29] <sakoman> I have the added constraint of wanting to facilitate upstream patch submission, so I delay commits until I have functionality chunks debugged
  • [14:45:22] <khasim> sakoman: we have to cleanup a lot before pushing them to upstream,
  • [14:45:39] <sakoman> So I'm always running off the current upstream git with a local patch generated by git diff master
  • [14:46:11] <sakoman> khasim: I realize that for u-boot! I was actually referring to kernel work flow
  • [14:46:22] <khasim> sorry :)
  • [14:46:28] <sakoman> u-boot needs massive amounts of work
  • [14:46:36] <khasim> true :(
  • [14:47:01] <sakoman> I think we'll get there if we all work together on it
  • [14:47:24] <Crofton|work> we could give up and have a meeting about how to solve the problem!
  • [14:47:33] <khasim> I prefer to totally switch to latest u-boot 1.3.3 and support only that version and would like to remove 1.1.4 dependency
  • [14:47:47] <sakoman> khasim: agreed
  • [14:47:54] <jkridner> we have a pretty big volunteer group: sakoman, dirk, khasim, bazbell, JonHunter, ...
  • [14:48:10] <jkridner> I hope that we can coordinate well. (sorry for those I left off)
  • [14:48:32] <khasim> I also see v2 patches from Nishanth menon getting very minor comments
  • [14:48:45] <khasim> we should have u-boot v2 on GIT soon
  • [14:49:23] <khasim> this is tremendous effort :)
  • [14:49:24] <sakoman> khasim: I wish I had time for v2, need 1.3.x for paying customers though
  • [14:49:41] <khasim> sakoman: yes
  • [14:49:55] <jkridner> Will v2 have an easier time with the USB features?
  • [14:51:08] <sakoman> khasim: is your clone still progressing? network led on my server is flashing furiously :-)
  • [14:51:36] <khasim> yes, its still going on
  • [14:51:48] <koen> sakoman: do you have an x-loader and u-boot binary available that I can put into nand?
  • [14:52:11] <khasim> koen: I have forwarded one to the list
  • [14:52:44] <jkridner> I'm also still in the middle of a clone.
  • [14:52:53] <sakoman> koen: yes. I will post in a moment. xloader is identical to khasim's
  • [14:53:57] <jkridner> I wish I made a bare clone. If the b/w is an issue, I'll clone this copy to the final mirror location and change the pull URL.
  • [14:55:39] <sakoman> jkridner: my dsl uplink is only 1mb, not sure how many people are pulling right now
  • [14:56:04] <jkridner> mine is done.
  • [14:58:37] <koen> Receiving objects: 81% (49547/61169), 19.73 MiB | 36 KiB/s
  • [14:58:41] <koen> almost done
  • [14:59:25] <jkridner> I need to do some cleanup, but I have a clone at http://www.beagleboard.org/u-boot.git/.
  • [15:00:23] <khasim> oh, I should have tried later then :) - mine is still going on
  • [15:01:32] <sakoman> koen: u-boot binary is in http://www.sakoman.net/omap3/
  • [15:01:54] <koen> thanks!
  • [15:02:24] <sakoman> flash procedure:
  • [15:02:27] <sakoman> loadb 80200000
  • [15:02:27] <sakoman> <send uboot>
  • [15:02:27] <sakoman> nand unlock
  • [15:02:27] <sakoman> nand ecc sw
  • [15:02:27] <sakoman> nand erase 80000 160000
  • [15:02:28] <sakoman> nand write 80200000 80000 160000
  • [15:04:25] <koen> can't I just flashcp it with linux?
  • [15:04:31] <sakoman> sure
  • [15:05:16] <sakoman> jkridner: the test branch doesn't show up in that clone
  • [15:05:24] <jkridner> doh!
  • [15:05:38] <jkridner> did I clone wrong?
  • [15:06:12] <sakoman> try git checkout test
  • [15:07:07] <jkridner> error: pathspec 'test' did not match any file(s) known to git.
  • [15:07:07] <jkridner> Did you forget to 'git add'?
  • [15:07:59] <sakoman> bummer. wish I was i git expert :-(
  • [15:08:03] <khasim> git-checkout --track -b tosakoman origin/test
  • [15:08:21] <khasim> this worked for me, I can see the changes made by sakoman
  • [15:10:09] <jkridner> domU-12-31-38-00-9C-98 u-boot-omap3 # git-checkout --track -b tosakoman origin/test
  • [15:10:09] <jkridner> Branch tosakoman set up to track remote branch refs/remotes/origin/test.
  • [15:10:09] <jkridner> Switched to a new branch "tosakoman"
  • [15:11:07] <jkridner> I don't see how "checkout" impacts what branches exist in the repository.
  • [15:11:37] <sakoman> jkridner: agreed. this git transition has a pretty steep learning curve :-(
  • [15:11:51] <sakoman> svn was so familiar :-)
  • [15:12:36] <sakoman> Has to be done to live in the linux kernel world though
  • [15:16:43] <bazbell> http://www.sourcemage.org/Git_Guide has some FAQs on usage. The command from khasim allows you work on an upstream branch other than "master".
  • [15:17:58] * RogerMonk (n=a0740758@nat/ti/x-bd90e816abcb6fe8) has joined #beagle
  • [15:22:15] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
  • [15:30:56] <koen> hrm
  • [15:31:00] <koen> that doesn't seem to work
  • [15:31:25] <koen> x-load (from flash) says: Texas Instruments X-Loader 1.41
  • [15:31:25] <koen> Starting on with MMC
  • [15:31:42] <koen> and then nothing
  • [15:33:33] <sakoman> koen: let me try it here
  • [15:33:53] <sakoman> maybe something got corrupted??
  • [15:36:40] <sakoman> koen: downloaded the u-boot image and transferred to beagle. works fine here
  • [15:37:10] <sakoman> did you put in flash or on mmc?
  • [15:38:01] <koen> f97e0190e67782827d15666c34315824 u-boot-beagleboard-git-r1.bin
  • [15:39:07] <sakoman> f97e0190e67782827d15666c34315824 u-boot-beagleboard-git-r1.bin
  • [15:39:28] <koen> I did this: http://rafb.net/p/R7IJAb43.txt
  • [15:41:32] <sakoman> looks right
  • [15:42:09] <sakoman> try doing a fatload of u-boot, then "go 80200000"
  • [15:42:15] <sakoman> see if it runs from ram
  • [15:42:49] <koen> U-Boot 1.3.3 (Jun 2 2008 - 15:47:15)
  • [15:42:53] <koen> seems to work
  • [15:44:41] <sakoman> now "nand read 80200000 80000 160000" and "go 80200000"
  • [15:46:18] <sakoman> if that works it might be a xload issue
  • [15:47:05] <koen> tat
  • [15:47:08] <koen> that works
  • [15:47:41] <koen> http://rafb.net/p/4JrZZX18.txt
  • [15:48:24] <koen> looks like an xload issue then
  • [15:48:58] <koen> anyway
  • [15:49:01] * koen heads of to a BBQ
  • [15:54:16] * RogerMonk (n=a0740758@nat/ti/x-bd90e816abcb6fe8) Quit (Remote closed the connection)
  • [16:23:33] * BThompson (n=BThompso@nat/ti/x-356af043a9181a38) Quit (Remote closed the connection)
  • [16:41:25] <jkridner> sakoman: I tried a different approach for cloning...
  • [16:41:34] <jkridner> git --bare init
  • [16:41:55] <jkridner> git remote add -t master -t test git://www.sakoman.com/git/u-boot-omap3.git
  • [16:41:57] <jkridner> get fetch
  • [16:42:05] <jkridner> er, git fetch
  • [16:42:22] <jkridner> let me know if it starts to hit your system, as it doesn't seem to be doing anything.
  • [16:43:00] <jkridner> over on #git, I got the advice to do 'git remote add --mirror' in a bare repo
  • [16:43:30] <jkridner> that advice doesn't seem to work for my version of git, and I don't know what other args to provide.
  • [16:43:38] <jkridner> so, above is the command I tried.
  • [16:45:38] <jkridner> doh. connection timed out.
  • [16:45:52] <jkridner> hope we didn't bombard you.
  • [16:51:48] * ldesnogu_ (n=ldesnogu@ven06-2-82-247-86-183.fbx.proxad.net) has joined #beagle
  • [17:22:28] <sakoman> jkridner: didn't notice a thing
  • [17:22:51] <jkridner> I made a sakoman.com vs. sakoman.net mistake.
  • [17:22:54] <jkridner> I seem to do that a lot.
  • [17:23:06] <jkridner> latest method doesn't seem to work well.
  • [17:23:13] <jkridner> last thing I tried was...
  • [17:23:16] <jkridner> git --init bare
  • [17:23:38] <jkridner> git remote add -t master -t test sakoman http://www.sakoman.net/git/u-boot-omap3/.git
  • [17:23:51] <sakoman> and it didn't work?
  • [17:23:55] <jkridner> git fetch sakoman
  • [17:24:07] <jkridner> not what I'd want...
  • [17:24:17] <jkridner> didn't seem to keep the right history and has permissions problems.
  • [17:24:19] <jkridner> see http://www.beagleboard.org/gitweb/?p=u-boot.git;a=summary
  • [17:25:28] <sakoman> jeeze it doesn't seem like it ought to be that hard
  • [17:25:30] <jkridner> I need to figure how how to get the right 'shortlog' and 'heads'.
  • [17:25:40] <sakoman> maybe I'm screed up on my end
  • [17:25:46] <sakoman> screwed up
  • [17:26:02] <jkridner> arg, btw, git --init bare was actually 'git --bare init'.
  • [17:26:07] <jkridner> doubt it...
  • [17:26:16] <jkridner> it looked better to me when I just cloned...
  • [17:26:30] <jkridner> but, the heads section didn't include test.
  • [17:26:57] <jkridner> It seems that you must be a master of the text files in the repository to be able to use git. :(
  • [17:27:04] <jkridner> well, host git anyway.
  • [17:27:13] <jkridner> using git locally is really nice and easy.
  • [17:27:48] <jkridner> the 'remotes' concept is what doesn't come to me easily.
  • [17:27:56] <jkridner> seems like it is easy, but it keeps surprising me.
  • [17:27:58] <sakoman> same here
  • [17:44:46] * BThompson (n=BThompso@nat/ti/x-691ea9c24adf471c) has joined #beagle
  • [17:52:40] * GAN8001 (n=GAN8001@pdpc/supporter/active/generalantilles) has left #beagle
  • [18:02:13] <keesj> I need a beagleboard yesterday!
  • [18:17:59] * NishanthMenon (n=nmenon@nat/ti/x-cbe1fdc2a958f5e1) has joined #beagle
  • [18:59:09] <DJWillis> I will settle for one tomorrow ;-)
  • [19:00:39] <mru> I have one today, but my newly built kernel doesn't want to boot :-(
  • [19:03:32] <DJWillis> mru: what sort of issues?
  • [19:03:48] <mru> Unable to handle kernel NULL pointer dereference at virtual address 00000000
  • [19:03:55] <DJWillis> hmmm
  • [19:05:57] * like2wise (n=likewise@82-171-51-231.ip.telfort.nl) has joined #beagle
  • [19:06:27] <ds2> just realize this is june... is there a shipping date for boards yet? :)
  • [19:15:31] <Crofton|work> there have been some emails on the list, I forget the specifics
  • [19:15:42] <ds2> I'll look at the archives
  • [19:17:24] * Crofton|work is doing a presentation on the beagle tomorrow
  • [19:18:07] <keesj> Crofton|work: where?
  • [19:18:33] <Crofton|work> http://www.wireless.vt.edu/symposium.html
  • [19:18:40] <Crofton|work> about 500 yards from my house :)
  • [19:19:47] <Crofton|work> http://www.wireless.vt.edu/images/symposium/Symposium203.jpg
  • [19:20:06] <Crofton|work> I'm closest to the camera and there is an OSK in front of me
  • [19:20:45] * like2wise (n=likewise@82-171-51-231.ip.telfort.nl) Quit ()
  • [19:28:17] <keesj> Well it is hidden between the computer and the mesuring device :p but apparently in the wrong continent for me to hop in
  • [19:28:29] <Crofton|work> :)
  • [19:28:30] <Crofton|work> sorry
  • [19:29:26] <keesj> it's ok , you are probably going to tell that the beagle's ears are to small to move over the air and that you need more hardware
  • [19:30:53] <Crofton|work> :)
  • [19:57:22] * BThompson (n=BThompso@nat/ti/x-691ea9c24adf471c) Quit ("Trillian (http://www.ceruleanstudios.com")
  • [20:01:49] <koen> re
  • [20:09:58] * ldesnogu_ (n=ldesnogu@ven06-2-82-247-86-183.fbx.proxad.net) Quit ("Leaving")
  • [20:12:12] * theoddbot (n=brendan@203-214-112-43.dyn.iinet.net.au) Quit (Read error: 101 (Network is unreachable))
  • [20:31:23] * jkridner (n=jason@c-76-31-18-64.hsd1.tx.comcast.net) Quit ()
  • [20:31:37] * jkridner (n=jason@c-76-31-18-64.hsd1.tx.comcast.net) has joined #beagle
  • [20:32:17] * koen removed x-loader from NAND
  • [20:32:30] <jkridner> and you are booting?
  • [20:35:19] <koen> yes
  • [20:35:28] <koen> x-loader and sakoman's uboot from sd
  • [20:36:53] <koen> sakoman: your u-boot is working pretty well :)
  • [20:37:44] <Crofton> whats the TRM url?
  • [20:38:18] <koen> spruf98.pdf
  • [20:38:43] <Crofton> Your search - spruf98.pdf - did not match any documents.
  • [20:39:49] <jkridner> which TRM? OMAP35x?
  • [20:39:58] <koen> http://www.ti.com/lit/gpn/omap3530
  • [20:40:10] <koen> wrong link, sorry
  • [20:40:17] <Crofton> we need an important link page ,,,,
  • [20:40:39] <koen> http://www.ti.com/litv/pdf/spruf98a
  • [20:40:41] <jkridner> doc number is spruf98a now.
  • [20:40:57] <Crofton> got it thanks
  • [20:40:58] <jkridner> elinux.org/BeagleBoard has many important links.
  • [20:41:03] <jkridner> I suggest consolidating there.
  • [20:41:49] <Crofton> yeah
  • [20:41:58] <sakoman> koen: I'm running into a tiff build issue. Error is:
  • [20:41:58] <sakoman> | ../libtool: line 784: X--tag=CC: command not found
  • [20:42:20] <sakoman> You seem to be the last one to touch tiff :-)
  • [20:42:48] <sakoman> koen: strange x-loader doesn't work from nand for you
  • [20:42:58] <koen> and that u-boot doesn't work for dirk
  • [20:45:24] <sakoman> koen: I thought he said it basically works
  • [20:45:34] <sakoman> He is getting i2c errors
  • [20:46:15] <sakoman> and the patch he used did *not* have the mmc fix, but mmc worked for him!
  • [20:46:44] <koen> sakoman: mtn pull for the libtiff fix
  • [20:46:45] <sakoman> I suspect that it is related to us using different compilers and the code not being quite "right"
  • [20:46:58] <sakoman> koen: thanks :-)
  • [20:47:03] <koen> csl compilers are more trouble thatn they are worth
  • [20:47:23] <koen> every build engineer I know hates them
  • [20:47:28] <sakoman> I think it might make sense to blame the code in this case
  • [20:47:34] <jkridner> without them, would the ARM features ever get done?
  • [20:48:12] <koen> sure, arm would just pay people to put the code into mainline gcc
  • [20:48:18] <sakoman> I think some of the constructs in beagle u-boot work by luck
  • [20:48:39] <koen> the trouble with csl compilers is that the line between fixes, features and black magic isn't clear
  • [20:50:35] <koen> in the old days it was "get EABI" vs "C++ apps crash randomly"
  • [20:50:55] <sakoman> I should probably look through the i2c code and see if anything looks funky there. Could be something similar to the mmc issue.
  • [20:51:22] * koen is holding his breath for gcc 4.3.2
  • [20:52:16] <sakoman> koen: are you sure you used the right x-load for nand?
  • [20:52:31] <sakoman> I'm bothered by the fact that it didn't work for you
  • [20:52:41] <koen> 37e1647c5b92e0f3016b4b5e5f925b44 x-load
  • [20:53:21] <koen> drat
  • [20:53:29] <koen> 301ece8b7e87bdef210ebb6f592b1f80 x-load.bin.ift
  • [20:53:39] <jkridner> :)
  • [20:54:52] <sakoman> 301ece8b7e87bdef210ebb6f592b1f80 /home/sakoman/Desktop/x-load.bin.ift
  • [20:54:56] <sakoman> :-)
  • [20:55:49] <koen> I had another x-load.bin.ift :(
  • [20:56:28] <sakoman> I'll stash a copy with my u-boot binary
  • [20:57:13] <koen> hmmm
  • [20:57:18] <jkridner> koen: I thought you were x-load free now? I guess I misunderstood.
  • [20:57:22] <koen> 'nand ecc hw' doesn't work in your uboot
  • [20:59:11] <koen> crap, I need uboot 1.1.4 to flash x-load
  • [20:59:15] <sakoman> good catch! I hadn't tried that yet
  • [21:00:17] <koen> and 'run' doesn't seem to work
  • [21:02:56] <koen> ah, now it works
  • [21:03:25] <koen> uboot 1.3.3 loads uboot 1.1.4 into ram, go 802000, flash x-loader and reboot
  • [21:03:56] <sakoman> koen: it seems you can skip the ecc step with 1.3.3
  • [21:04:29] <sakoman> That's what I've been doing without even realizing it
  • [21:05:43] <sakoman> koen: run command wasn't included in the build. I'll add it
  • [21:06:31] <koen> I now have x-load, u-boot and a rootfs in NAND
  • [21:06:39] <sakoman> me too :-)
  • [21:07:02] <sakoman> now you can use mmc for other interesting things
  • [21:07:18] <koen> I am still SDIO-less :)
  • [21:07:34] <sakoman> one step at time ;-)
  • [21:08:00] <koen> I really need to finish those neuros bounties to get some cash for gadgets :)
  • [21:10:21] <sakoman> koen: any other favorite u-boot commands you would like enabled while I am at it?
  • [21:11:06] <koen> ext2load, fatload and jffs2 stuff is all I can think of right now
  • [21:13:01] <sakoman> I suppose the net commands don't make sense :-(
  • [21:13:33] <jkridner> loadb/loady.
  • [21:13:39] <jkridner> other serial stuff for now.
  • [21:14:59] * NishanthMenon (n=nmenon@nat/ti/x-cbe1fdc2a958f5e1) Quit ("Ex-Chat")
  • [21:15:00] <koen> while ethernet on the beagle would be neat, I still rather have the s-video out
  • [21:15:28] <koen> (which still doesn't work in linux-omap git, though)
  • [21:15:51] <ds2> with all this uboot talk, does this mean eventually, all the other platforms will (Zoom, SDP, EVM) will gain the ability to MMC/USB boot too? along with the full fs support?
  • [21:16:23] <koen> root@beagleboard:~# df -h | grep mtd
  • [21:16:23] <koen> mtd3 4.0M 2.5M 1.5M 62% /boot
  • [21:16:23] <suihkulokki> koen: ack.
  • [21:17:03] <suihkulokki> there is already quite few boards with ethernet sockets - but pretty much none with SGX and tv-out :)
  • [21:17:04] <koen> I don't know of any usb->svideo dongles
  • [21:17:27] <ds2> koen: there are a few... the SiS is supported by Linux currently
  • [21:17:40] <koen> the sis one I have has vga only AFAIK
  • [21:17:42] <ds2> oops you said svideo
  • [21:17:44] <ds2> nevermind
  • [21:17:47] <koen> and is slooooooooooow
  • [21:17:48] <sakoman> jkridner: loadb and loady are already there
  • [21:18:18] <ds2> been toying with the idea of putting an Analog devices chip to give me svideo out (assuming the clocks on it can be reprogrammed)
  • [21:18:42] <koen> sakoman: do you by any change know how I can teach the kernel that 1024x786 (in the lcd file) is just a default and that it supports multiple resolutions?
  • [21:19:01] <jkridner> ds2 ? Beagle has S-Video out.
  • [21:19:29] <ds2> jkridner: so there is no truth to the rumor that S-Video out is getting dropped in favor of ethernet?
  • [21:19:34] <koen> jkridner: there was talk on the mailinglist that people at linuxtag said that rev C might replace s-video with ethernet
  • [21:19:44] <sakoman> koen: not offhand. the kernel has changed since the last time i did that.
  • [21:19:50] <jkridner> totally wrong. didn't see that on the mailing list.
  • [21:19:58] <koen> great!
  • [21:20:05] <ds2> dropping s-video would be the push for me to goto openpantera instead
  • [21:20:05] * koen hugs svideo port
  • [21:20:10] <sakoman> do any of the other video drivers do that? that's where I would start :-)
  • [21:20:33] <ds2> of course, will the linux-omap tree support s-video...that would be the important question ;)
  • [21:20:47] <jkridner> I don't see that thread on the e-mail lsit.
  • [21:20:55] <koen> ds2: IIRC the dispc driver already does
  • [21:21:04] <jkridner> is that on linux-omap, rather than beagle mail list?
  • [21:21:19] <koen> ds2: it's just a matter of hooking up the beagle board file to that
  • [21:21:27] <ds2> koen: when I looked at it 6 months ago, I think the s-video out didn't work... had to use the TI kernel for that
  • [21:21:34] <ds2> this was for the SDP
  • [21:21:37] <koen> 13:34 < nomis> theoddbot: from what I understood on the Linuxtag booth they are pondering to replace it with an ethernet jack.
  • [21:21:46] <koen> ~10 hours ago on irc
  • [21:22:33] <jkridner> k. I missed that. nomis, theoddbot, this is not currently under consideration at all. Ethernet will be via USB for Beagle.
  • [21:22:56] <ds2> whew
  • [21:23:06] <jkridner> if we were to put Ethernet on the board, it would probably make the most sense to connect it by USB anyway, which would add a lot of cost for those who want WiFi instead.
  • [21:23:39] <ds2> heh.. I question why ethernet was even put on the zoom
  • [21:24:24] <Crofton|work> why can't they give rough clock numbers and data transfer rates in the TRM?
  • [21:25:21] <Crofton|work> ball park number, how many k byte per second can I do on McBSP?
  • [21:25:31] <Crofton|work> ball park number, how many M byte per second can I do on MMC?
  • [21:25:45] <ds2> can't you deduce the McBSP figure from the max clock speed?
  • [21:25:51] <koen> mmc is 8bits and 48MHz, right?
  • [21:26:00] <Crofton|work> not in 30 minutes
  • [21:26:52] <ds2> 1 bit
  • [21:27:41] <Crofton|work> I'm not familair with the interfaces and the TRM to navigate through the clock sources ...
  • [21:28:05] <ds2> now add to that, power management
  • [21:28:16] <Crofton|work> I think I'll just say McBSP is kbytes per secind
  • [21:28:25] <ds2> bits is more accurate
  • [21:28:27] <Crofton|work> er 100's of kbs
  • [21:28:36] <Crofton|work> and mmc is 10's of MBS
  • [21:39:08] <sakoman> current u-boot default environment is set up for mmc initrd boot
  • [21:39:20] <sakoman> should we change this to nand boot?
  • [21:39:46] <sakoman> suppose that only would make sense if the beagleboards are going to ship with linux/rootfs in nand
  • [21:40:46] <koen> and kernel in jffs2 on mtd3, right?
  • [21:42:39] <ds2> why not yaffs2?
  • [21:43:46] <koen> yaffs2 has more bugs than lines of code
  • [21:43:55] <koen> and it will never be in the upstream kernel
  • [21:44:04] <koen> logfs or ubifs would be a far better choice
  • [21:44:10] <ds2> yes but there are things that work in yaffs2 but not in jffs2
  • [21:44:18] <ds2> do either of those 2 allow memmap?
  • [21:44:40] <ds2> there is a certain not fully released phone stack that will not work in jffs2
  • [21:47:39] * koen makes no comments about android or its future
  • [21:53:56] <sakoman> well, since there is no support in u-boot for yaffs2, logfs, or ubifs it makes the decision fairly simple :-)
  • [21:55:10] <sakoman> koen: yes kernel should be in a jffs2 partition on mtd3
  • [21:55:40] <sakoman> does linux.inc know how to build an image like that?
  • [21:57:36] <koen> an image like what?
  • [21:57:43] <koen> uImage in jffs2?
  • [21:57:47] <sakoman> yes
  • [21:57:49] <sakoman> :-)
  • [21:57:59] <koen> I'd add an image description for that :)
  • [21:58:18] <Crofton|work> why do you want to put the kernel in a file system?
  • [21:58:19] <sakoman> what a good man!
  • [21:58:24] <sakoman> bad blocs
  • [21:58:28] <koen> Crofton|work: bad block handling
  • [21:58:33] <Crofton|work> doh
  • [21:59:05] <koen> sakoman: maybe a do_deploy_append would work as well
  • [21:59:07] <ds2> easier to boot off
  • [21:59:11] <ds2> can list available kernels
  • [21:59:31] <sakoman> I'm still concerned that the u-boot partition is required to be good blocks
  • [21:59:53] <sakoman> Fixing that will require beefing up x-load
  • [21:59:56] <koen> sakoman: I'm curren;y thinking of writing a uboot script that tries to load from mmc first and falls back to NAND
  • [22:00:26] <sakoman> good idea. that's what we do on the gumstix boards
  • [22:00:45] <sakoman> only there we try both mmc and cf
  • [22:01:13] <ds2> though being able to do bootm 0x..... when a kernel is in NOR is nicer then JFFS...but beagle has no NOR
  • [22:02:36] <sakoman> koen: I enabled autoscript support in the version of u-boot I am building right now
  • [22:02:50] <sakoman> as well as ext2 and a couple of other things
  • [22:02:51] * koen needs to google that first
  • [22:03:50] <koen> sweet, mkimage can include a script
  • [22:05:23] <sakoman> very useful feature
  • [22:05:49] * koen tries to figure out how to get the angstrom banner into an uImage
  • [22:06:22] * koen vi classes/kernel.bbclass
  • [22:16:36] <koen> hmmmm
  • [22:16:51] <koen> it seems mkimage can only build _seperate_ script images
  • [22:20:23] <sakoman> OMAP3 beagleboard.org # ext2ls mmc 0:2 etc
  • [22:20:23] <sakoman> <DIR> 4096 .
  • [22:20:23] <sakoman> <DIR> 4096 ..
  • [22:20:23] <sakoman> 30 ld.so.conf
  • [22:20:23] <sakoman> 428 issue.net
  • [22:20:24] <sakoman> 12 hostname
  • [22:20:26] <sakoman> 433 issue
  • [22:20:40] <sakoman> so ext2 works now :-)
  • [22:21:03] <koen> yay!
  • [22:22:08] * bazbell (n=a0192809@nat/ti/x-e2bfb64d3e358d6f) Quit (Remote closed the connection)
  • [22:24:22] * theoddbot (n=brendan@124-171-194-198.dyn.iinet.net.au) has joined #beagle
  • [22:24:46] <sakoman> you can pick up a binary at http://www.sakoman.net/omap3
  • [22:25:03] <sakoman> it's the r2 version
  • [22:25:12] <koen> I guessed as much :)
  • [22:25:15] <sakoman> only 10K larger
  • [22:25:19] * RogerMonk (n=a0740758@nat/ti/x-5d85d52fc42c9828) has joined #beagle
  • [22:25:54] <sakoman> the autoscript stuff is there as well as run and a few other goodies
  • [22:25:59] <sakoman> type help for details
  • [22:26:02] * koen heads to bed
  • [22:26:04] <koen> 'night all
  • [22:26:07] <Crofton> gn
  • [22:26:09] <sakoman> gn
  • [22:30:31] * prpplague (n=dave@mail.americanmicrosystems.com) Quit ("Leaving")
  • [22:31:01] * RogerMon1 (n=a0740758@nat/ti/x-689917b64a8282e4) has joined #beagle
  • [22:34:46] * RogerMonk (n=a0740758@nat/ti/x-5d85d52fc42c9828) Quit (Remote closed the connection)
  • [22:36:28] * Guizmo` (n=guizmo@41.92.198-77.rev.gaoland.net) has joined #beagle
  • [22:57:14] * BThompson (n=BThompso@cpe-76-185-93-11.tx.res.rr.com) has joined #beagle
  • [23:03:12] <mru> hey, what's a good kernel version to use?
  • [23:09:56] * bazbell (n=a0192809@nat/ti/x-f782e8c142f41534) has joined #beagle
  • [23:35:05] <sakoman> mru: TI provides 2.6.22 in both source and binary
  • [23:35:38] <mru> that's getting a bit long in the tooth...
  • [23:36:02] <sakoman> some of us are living dangerously and using 2.6.26rc4 from linux-omap git
  • [23:36:08] <ds2> 2.6.24 ;)
  • [23:36:32] <mru> I tried building from linux-omap git and that's crashing during bootup
  • [23:36:38] <ds2> what hardware?
  • [23:36:57] <mru> beagle rev b
  • [23:37:19] <ds2> is it showing anything at all?
  • [23:37:41] <mru> yes, Unable to handle kernel NULL pointer dereference at virtual address 00000000
  • [23:37:57] <mru> PC is at __rcu_process_callbacks+0x1ac/0x244
  • [23:38:56] <ds2> got the entire traceback?
  • [23:41:09] <mru> [<c0073abc>] (__rcu_process_callbacks+0x0/0x244) from [<c0073d18>] (rcu_process_callbacks+0x18/0x38)
  • [23:41:09] <mru> r6:0000000a r5:00000001 r4:c032b5a4
  • [23:41:09] <mru> [<c0073d00>] (rcu_process_callbacks+0x0/0x38) from [<c0052718>] (__do_softirq+0x60/0xc0)
  • [23:41:09] <mru> [<c00526b8>] (__do_softirq+0x0/0xc0) from [<c00527c0>] (irq_exit+0x48/0x88)
  • [23:41:09] <mru> [<c0052778>] (irq_exit+0x0/0x88) from [<c0029050>] (__exception_text_start+0x50/0x68)
  • [23:41:10] <mru> r4:c030f530
  • [23:41:12] <mru> [<c0029000>] (__exception_text_start+0x0/0x68) from [<c0029710>] (__irq_svc+0x30/0x80)
  • [23:41:38] <ds2> seems like something registered a bad interrupt handler
  • [23:41:47] <sakoman> mru: if you want to try one that I built, look in http://www.sakoman.net/omap3/beagle/
  • [23:44:23] <mru> that boots ok
  • [23:44:43] <mru> maybe my compiler is bad
  • [23:44:57] <sakoman> I used gcc 4.2.2
  • [23:45:06] <mru> yes, I can see that
  • [23:45:26] <sakoman> There's a 1.3.3 u-boot binary in there too
  • [23:45:39] <sakoman> the r2 version has ext2 support enabled
  • [23:45:56] <sakoman> seems to work ok
  • [23:49:28] <mru> I'm trying to build a toolchain with neon instruction support
  • [23:49:31] <mru> it's proving difficult
  • [23:54:31] <sakoman> mru: I'm sticking to the basics for now :-) You might talk with koen about his compiler experiences.