• [00:00:46] <ds2> if you got a 1024x768 HMD, you have the budget for a EVM!
  • [00:01:06] <ds2> and I have priced out HDMs
  • [00:04:09] * Thanatos (n=rwvens@a83-132-76-156.cpe.netcabo.pt) has joined #beagle
  • [00:17:34] * Thanatos_Wibrain (n=rwvens@a83-132-76-156.cpe.netcabo.pt) Quit (Read error: 110 (Connection timed out))
  • [00:56:07] * bazbell (n=a0192809@nat/ti/x-0d97759639aebf08) has joined #beagle
  • [01:09:52] * Beagle8 (n=Beagle8@user-0c93qer.cable.mindspring.com) has joined #beagle
  • [01:11:16] * Beagle8 (n=Beagle8@user-0c93qer.cable.mindspring.com) Quit (Client Quit)
  • [02:37:42] * MattJ (n=MattJ@88-111-173-221.dynamic.dsl.as9105.com) Quit (Read error: 110 (Connection timed out))
  • [03:10:25] * Thanatos (n=rwvens@a83-132-76-156.cpe.netcabo.pt) Quit (Read error: 110 (Connection timed out))
  • [03:41:02] * BThompson (n=BThompso@cpe-76-185-93-11.tx.res.rr.com) Quit ("Trillian (http://www.ceruleanstudios.com")
  • [03:47:08] * zarquin (n=zarquin@bright-snat.ucc.asn.au) has joined #beagle
  • [03:47:30] <zarquin> hi
  • [05:42:44] * banderson (n=irc@69.71.183.7) Quit (Read error: 110 (Connection timed out))
  • [06:24:17] * hadara (i=hadara@bsd.ee) Quit (Read error: 104 (Connection reset by peer))
  • [06:24:21] * hadara (i=hadara@bsd.ee) has joined #beagle
  • [07:12:36] * trickie (n=trickie@basesoft.xs4all.nl) has joined #beagle
  • [07:27:22] * RogerMonk (n=a0740758@nat/ti/x-1f98f7d0f3a390ca) Quit (Remote closed the connection)
  • [07:34:16] * banderson (n=irc@69.71.183.7) has joined #beagle
  • [07:47:46] * methril|away is now known as methril
  • [08:43:40] * Olipro (i=Olipro@unaffiliated/olipro) has joined #beagle
  • [09:13:28] * RogerMonk (n=a0740758@nat/ti/x-a63bc3c69134544f) has joined #beagle
  • [09:28:09] * koen watches myth 0.21 + svn fixes compile
  • [10:14:13] * RogerMonk (n=a0740758@nat/ti/x-a63bc3c69134544f) Quit (Remote closed the connection)
  • [10:19:37] <koen> there, myth 0.21 compiled for armv5te
  • [10:19:44] * koen moves on to armv7a
  • [10:22:32] * MattJ (n=MattJ@88-111-173-221.dynamic.dsl.as9105.com) has joined #beagle
  • [10:33:39] * RogerMonk (n=a0740758@nat/ti/x-8e6c2ccf7854e7e2) has joined #beagle
  • [10:52:26] * Olipro_ (i=Olipro@unaffiliated/olipro) has joined #beagle
  • [10:58:32] * RogerMonk (n=a0740758@nat/ti/x-8e6c2ccf7854e7e2) Quit (Remote closed the connection)
  • [11:03:16] * RogerMonk (n=a0740758@nat/ti/x-4d16e5507e9825d0) has joined #beagle
  • [11:07:25] * Olipro (i=Olipro@unaffiliated/olipro) Quit (Nick collision from services.)
  • [11:07:28] * Olipro_ is now known as Olipro
  • [12:07:10] <koen> RogerMonk: have a look at http://scap.linuxtogo.org/ :)
  • [12:08:45] <koen> RogerMonk: mythtv 0.21 running on beagle
  • [12:35:43] <Crofton> koen, nfs-utils-1.1.3 creates mount.nfs :)
  • [12:37:02] * jkridner|work (n=a0321898@nat/ti/x-73c2887ba55548e3) has joined #beagle
  • [12:37:17] <jkridner|work> good morning
  • [12:37:27] <Crofton> gm
  • [12:48:27] * JuanG (n=Juan@nat/ti/x-07a91c160fab31d9) has joined #beagle
  • [12:48:36] * JuanG (n=Juan@nat/ti/x-07a91c160fab31d9) has left #beagle
  • [12:57:50] * ldesnogu_ (n=ldesnogu@fw-tnat.cambridge.arm.com) has joined #beagle
  • [12:57:50] * ldesnogu (n=ldesnogu@fw-tnat.cambridge.arm.com) Quit (Read error: 104 (Connection reset by peer))
  • [12:58:18] * BThompson (n=BThompso@nat/ti/x-669d69ec18a4a30f) has joined #beagle
  • [12:59:22] * ldesnogu_ is now known as ldesnogu
  • [13:11:14] <sakoman_> gm
  • [13:14:39] <koen> Crofton: 1.1.2 as well, and that's in OE now :)
  • [13:14:42] <koen> hey sakoman_
  • [13:15:03] <Crofton> er, I think that is what I meant :)
  • [13:18:30] <koen> I had to edit the recipe to install mount.nfs, though
  • [13:24:08] <Crofton> I think there needs to be a link from mount.nfs3 to mount.nfs
  • [13:24:18] <Crofton> I'll check that when I can
  • [13:24:22] <Crofton> hopefully soon
  • [13:24:41] <Crofton> trying to work out how to get gnu radio test programs to run
  • [13:30:45] <koen> http://www.flickr.com/photos/koenkooi/2695061759/in/pool-beagleboard
  • [13:32:22] <jkridner|work> sweet!!!
  • [13:32:31] <jkridner|work> I have to show that at LinuxWorld!
  • [13:32:32] * prpplague (n=dave@mail.americanmicrosystems.com) has joined #beagle
  • [13:32:55] <jkridner|work> using a keyboard for controls?
  • [13:33:29] <jkridner|work> drool.... Need DSP/FFmpeg integration.... drool.
  • [13:33:53] <koen> jkridner|work: keyboard for control, yes
  • [13:34:12] <koen> jkridner|work: should be fairly easy to drop in mrus NEON stuff
  • [13:34:31] <koen> since myth has an internal copy of ffmpeg sources
  • [13:34:58] <ldesnogu> I guess jkridner|work would like FFmpeg to use c64x+ :-)
  • [13:35:03] <koen> jkridner|work: but I think having an XV driver for the omapfb overlay would make another big impact
  • [13:36:26] <ldesnogu> koen: isn't xv supposed to use hardware for reaching high-speed? and IIUC the hardware doesn't have the right yuv support
  • [13:38:53] <jkridner|work> ldesnogu: you know I would, but I'm happy with NEON for now.
  • [13:39:33] <jkridner|work> Once there are plenty of developers, I suspect some will look to off-load the ARM or add multi-channel or higher-resolution support.
  • [13:40:11] <ldesnogu> I wonder how hard it would be to have yuv2yuv *and* audio decoding on the DSP
  • [13:40:44] <jkridner|work> I need to escalate the issue of finding out about YUV420 support vs the YUV422 support that is in the driver today.
  • [13:41:05] <ldesnogu> that's only a driver limitation, not a hw one?
  • [13:41:13] <jkridner|work> yuv2yuv and audio decoding would be minimal processing overhead for the DSP.
  • [13:41:29] <jkridner|work> the trick is getting the right software for ARM/DSP communications.
  • [13:41:38] <ldesnogu> jkridner|work: indeed, I am only wondering how multiple processes run on the DSP ;)
  • [13:42:04] <jkridner|work> well, if you download CCS, you get a copy of DSP/BIOS.
  • [13:42:36] <jkridner|work> (DSP/BIOS is an RTOS)
  • [13:43:06] <ldesnogu> but is that compatible with the various dsp gateway/link/whatever?
  • [13:43:06] <jkridner|work> CCS can be downloaded for free, but the debugger will stop working after 120 days.
  • [13:43:33] <jkridner|work> DSP/BIOS is used with Link and Bridge. I don't know the Gateway architecture.
  • [13:43:44] <ldesnogu> jkridner|work: I guess not only the debugger will stop working, the compiler tool will too...
  • [13:43:59] <ldesnogu> ok good to know :)
  • [13:45:04] <ldesnogu> jkridner|work: BTW I got no answer for my question about open source usage for OpenMAX DL NEON library
  • [13:45:21] <ldesnogu> I probably didn't find the right person to ask to...
  • [13:45:48] <jkridner|work> are you talking about within FFmpeg?
  • [13:46:10] <ldesnogu> within anything that could make use of that code
  • [13:46:11] <jkridner|work> I probably missed the question.
  • [13:46:27] <ldesnogu> it's an ARM internal question so it's good you missed it :-)
  • [13:46:35] <jkridner|work> ah! :)
  • [13:47:02] <jkridner|work> *can* it be used, vs. is anyone interested in using it.
  • [13:47:17] <ldesnogu> well if it can't be used, the subject is closed
  • [13:47:37] <ldesnogu> if it can be used, it's up to someone to pick code and copy it where needed
  • [13:47:51] <jkridner|work> I heard that some ARM folks had some ideas on improving mru's DCT. any idea if those will be shared to see if they are good ideas?
  • [13:48:25] <ldesnogu> who told you that?
  • [13:48:37] <ldesnogu> oh well I guess it's not a big secret...
  • [13:48:53] <ldesnogu> ARM IDCT is faster but less accurate as far as I understood
  • [13:49:07] <koen> jkridner|work: tried the TI neon compiler on ffmpeg yet?
  • [13:49:22] <ldesnogu> I am not sure mru wants to trade accuracy for speed
  • [13:50:18] * ldesnogu didn't know TI had a NEON compiler
  • [13:50:55] <jkridner|work> I have not, but I understand there is a meeting coming up to talk about that.
  • [13:51:01] <jkridner|work> TI does make a NEON compiler. :)
  • [13:51:26] <jkridner|work> ARM has one too, which should be interesting to try.
  • [13:51:49] <koen> AIUI the TI one is better at compiling and vectorizing things with loops
  • [13:51:57] <ldesnogu> I don't trust a compiler for vectorization...
  • [13:51:59] <jkridner|work> You can download and try TI's NEON compiler with CCS.
  • [13:52:03] <koen> coming from a DSP company and all
  • [13:52:12] <Crofton> hmm, the gnuradio benchmark dot product benchmark is way slower on the beagle than my laptop ....
  • [13:52:23] <koen> Crofton: double precision floats?
  • [13:52:37] <Crofton> I do not think so
  • [13:52:39] <ldesnogu> if you laptop is a Core thingie that's no surprise :)
  • [13:52:46] <Crofton> pretty sure single precision
  • [13:52:51] <Crofton> it is :)
  • [13:52:52] <koen> NEON should be able to do pretty fast single precision dot products
  • [13:53:23] * Crofton is just hoping for results atm
  • [13:53:41] <koen> ldesnogu: the issue with mythtv and XV is that it will spend eons in colourspace conversion
  • [13:53:50] <jkridner|work> Crofton: which compiler and which settings?
  • [13:53:54] <Crofton> hopefully, we can interest some of the gnuradio crowd when the boards go on sale
  • [13:53:59] <jkridner|work> GCC? CS2007q3?
  • [13:53:59] <Crofton> OE default :)
  • [13:54:03] <koen> ldesnogu: mru has a NEON based colourspace converter, but even that takes 25% of cpu time
  • [13:54:04] <jkridner|work> ah.
  • [13:54:19] <Crofton> right now, I am working through build issues :)
  • [13:54:23] <ldesnogu> koen: are you sure the hardware can handle that?
  • [13:54:28] <Crofton> and how to get the make check target to work
  • [13:54:50] <ldesnogu> Crofton: you probably ended up running on the VFPlite, which is bad :)
  • [13:54:52] <prpplague> morning folks
  • [13:55:03] <ldesnogu> hello
  • [13:55:06] <koen> jkridner|work: a journalist from http://www.heise-online.co.uk/ was interested in the beagleboard stuff, RogerMonk shoudl have his contact sheet
  • [13:55:26] <koen> ldesnogu: -mfpu=neon shouldn't fall back to vfp, right?
  • [13:55:30] <koen> hey prpplague
  • [13:55:38] <ldesnogu> koen: it is IIRC
  • [13:56:39] <Crofton> excellent, results
  • [13:58:24] <Crofton> http://pastebin.ca/1080760
  • [13:58:53] <Crofton> hmm, I worry about the ads during US TV tour coverage
  • [13:59:11] <Crofton> why would people interested in cycling need "natural male enhancement"?
  • [13:59:28] <koen> cyclists do dope, so why shouldn't the fans?
  • [14:00:06] <Crofton> natural dope :)
  • [14:00:09] <koen> Crofton: so generic is 30 times slower
  • [14:00:16] <Crofton> soemthing like that
  • [14:00:26] <koen> and 300 times slower than sse
  • [14:00:30] <Crofton> we need to add neon to generic
  • [14:00:32] <koen> try doing a NEON version :)
  • [14:00:40] <Crofton> yeah
  • [14:00:42] <ldesnogu> koen: did you try and see if the code changes? :)
  • [14:00:54] <koen> ldesnogu: which code?
  • [14:01:03] <ldesnogu> -mfpu=neon vs vfp
  • [14:01:13] <ldesnogu> according to my tests it did not change anything
  • [14:01:15] <koen> gcc -S shows it puts fpu=neon on top
  • [14:01:31] <koen> didn't disassemble the results
  • [14:01:34] <ldesnogu> well if it still generated vfp instructions that doesn't change anything
  • [14:01:49] <ldesnogu> I will check again :)
  • [14:04:43] <koen> ldesnogu: I needed -mfloat-abi=softfp to get the fpu-neon with gcc -S
  • [14:07:56] <koen> ldesnogu: cortex-a9 has vfp-heavy?
  • [14:08:09] * prpplague^2 (n=dave@mail.americanmicrosystems.com) has joined #beagle
  • [14:08:43] <koen> RogerMonk: does usb otg work on the rev A boards that arrived on monday?
  • [14:08:46] <ldesnogu> koen: :P
  • [14:08:57] <ldesnogu> the C-A9 VFP is fully pipelined
  • [14:12:42] * methril is now known as methril|away
  • [14:12:58] <koen> jkridner|work: a lot of people were asking how beagle compared to the arduino
  • [14:13:15] <ldesnogu> koen: with 2007q3 -mfpu neon or vfp doesn't change anything
  • [14:14:07] <koen> and adding -mfloat-abi=softfp?
  • [14:14:16] <jkridner|work> totally different class of capability/performance.
  • [14:14:21] <ldesnogu> that's what I did
  • [14:14:27] <ldesnogu> same with 2008q1
  • [14:14:40] <jkridner|work> Beagle is not something that can be built in a garage.
  • [14:15:05] <Crofton> what? You can't do the bga and pop in a toaster oven?
  • [14:15:36] <koen> jkridner|work: I said something like "arduino is good for making leds blink, beagle does HD video"
  • [14:16:10] * prpplague (n=dave@mail.americanmicrosystems.com) Quit (Nick collision from services.)
  • [14:16:13] * prpplague^2 is now known as prpplague
  • [14:16:23] <jkridner|work> Beagle looks to pull development environments from desktop code, rather than building up a lightweight programming language/system.
  • [14:16:38] <jkridner|work> koen: nice summary!
  • [14:17:00] <jkridner|work> your little LED blinking video might give the wrong impression. ;)
  • [14:17:26] <koen> people didn't believe that the 24" screens were being driven by the beagle
  • [14:17:30] <jkridner|work> Having a video of the video playback would be nice. I just have no time to make one.
  • [14:17:37] <jkridner|work> :)
  • [14:17:59] <jkridner|work> need to figure out how to get them convinced.
  • [14:18:14] <jkridner|work> btw, this is great feedback. thanks!
  • [14:18:19] <prpplague> jkridner|work: we have the same problem with the hammer
  • [14:19:38] <jkridner|work> (oh, won't someone please get the Flyswatter to work with Beagle)
  • [14:31:29] <koen> RogerMonk: could you hop on to #neuros?
  • [14:32:59] * docelic (n=docelic@78.134.199.245) has joined #beagle
  • [14:41:22] * bazbell (n=a0192809@nat/ti/x-0d97759639aebf08) Quit (Remote closed the connection)
  • [14:44:03] * bazbell (n=a0192809@nat/ti/x-3f1b266c19d2c302) has joined #beagle
  • [14:46:29] <Crofton> fyi, here is the code for the generic dot product!
  • [14:47:35] <koen> where?
  • [14:47:48] <Crofton> http://gnuradio.org/trac/browser/gnuradio/trunk/gnuradio-core/src/lib/filter/complex_dotprod_generic.cc
  • [14:47:52] <Crofton> oops :)
  • [14:48:07] <Crofton> there are a number of forms
  • [14:48:17] <Crofton> depending on input and taps types
  • [14:50:39] <koen> and the sse version: http://gnuradio.org/trac/browser/gnuradio/trunk/gnuradio-core/src/lib/filter/ccomplex_dotprod_sse.S
  • [14:51:04] <Crofton> yeah
  • [14:56:09] <Crofton> http://gnuradio.org/trac/changeset/8976
  • [14:56:15] <Crofton> need to do this for NEON
  • [14:56:39] <Crofton> need to find someone willing to pay for that :)
  • [15:04:31] <ldesnogu> it looks like they are using intrinsics for Altivec
  • [15:04:57] <Crofton> He did a quick job
  • [15:05:02] <koen> hence the "we can make it faster" comment :)
  • [15:05:07] <Crofton> yeah
  • [15:05:09] <ldesnogu> ^^
  • [15:05:28] <Crofton> bother nfs is stuck again
  • [15:07:25] <Crofton> oh well, these crazy guys are going to ride uphill in a minute ...
  • [15:08:07] * dirk2 (n=dirk@F301f.f.strato-dslnet.de) has joined #beagle
  • [15:13:44] <jkridner|work> I cannot believe it took me this long to find this link: http://focus.ti.com/dsp/docs/dspsupportaut.tsp?sectionId=3&tabId=416&familyId=44&toolTypeId=30
  • [15:13:47] <jkridner|work> CCS eval version.
  • [15:17:36] <jkridner|work> not sure if it includes all of the Cortex-A8/NEON updates.
  • [15:18:01] <ldesnogu> I don't think so
  • [15:18:27] <jkridner|work> k. service release 9 (SR9) should have all the updates.
  • [15:18:29] <ldesnogu> I installed it a few months ago and there was no mention of it
  • [15:19:05] <ldesnogu> anyway this is useless for us poor Linux users as it refuses to install with Wine :(
  • [15:24:07] * denix is back
  • [15:25:09] * trickie is now known as trickie|away
  • [15:25:26] <denix> koen: woot! myth is here, congrats! :)
  • [15:36:11] * Olipro is now known as \\TEST\
  • [15:36:22] * \\TEST\ is now known as Olipro
  • [15:44:51] <koen> denix: I put 0.21-fixes into the feeds
  • [15:45:24] <koen> http://www.angstrom-distribution.org/repo/?action=details&pnm=mythtv-frontend
  • [15:45:41] <koen> probably needs http://www.angstrom-distribution.org/repo/?action=details&pnm=mythtv-theme-g.a.n.t
  • [15:45:41] <denix> thanks!
  • [15:46:02] <denix> and qt3?
  • [15:47:45] <koen> in there as well
  • [15:48:02] <koen> 'opkg install mythtv-frontend mythtv-theme-g.a.n.t' should do the trick
  • [15:49:25] * Thanatos_Wibrain (n=rwvens@a83-132-76-156.cpe.netcabo.pt) has joined #beagle
  • [15:49:34] * Thanatos_Wibrain (n=rwvens@a83-132-76-156.cpe.netcabo.pt) Quit (Client Quit)
  • [15:58:32] * cian (n=cian@cian.ws) has joined #beagle
  • [15:59:16] * Thanatos (n=rwvens@a83-132-76-156.cpe.netcabo.pt) has joined #beagle
  • [15:59:43] * Thanatos is now known as Thanatos_Wibrain
  • [16:01:28] * Paulius (n=Paulius@unaffiliated/paulius) has joined #beagle
  • [16:01:39] <Paulius> When is the board going to be available/on sale?
  • [16:06:37] <dirk2> Paulius: There are some reports that you might be able to order rev B boards starting end of next week from digi key
  • [16:06:44] <Paulius> ah, cool.
  • [16:08:08] <dirk2> Paulius: You should join beagle mailing list to get more info when sale starts
  • [16:08:13] <Paulius> gotcha
  • [16:09:43] <docelic> Do you guys notice that Login button on beagleboard.org is not working?
  • [16:10:50] <dirk2> docelic: Just checked, its working, I'm logged in
  • [16:11:08] <docelic> how do you make it work? I'm clicking it and nothing is happening
  • [16:11:21] <docelic> just refreshes the page
  • [16:13:01] <dirk2> docelic: I entered my openID "link" http://xxxxxx.myopenid.com/, then I'm redirected to openID login, enter there the password and then I'm logged in
  • [16:15:00] <dirk2> docelic: See description at http://beagleboard.org/login
  • [16:15:19] <docelic> you're right, thanks
  • [16:22:56] * docelic (n=docelic@78.134.199.245) Quit ("http://www.spinlocksolutions.com/")
  • [16:50:24] * Thanatos_Wibrain (n=rwvens@a83-132-76-156.cpe.netcabo.pt) Quit (Read error: 110 (Connection timed out))
  • [17:04:04] <koen> http://www.flickr.com/photos/koenkooi/2692388640/ now has some notes
  • [17:04:36] <prpplague> koen: why don't you use the elinux wiki to post that kind of stuff?
  • [17:04:59] <koen> because I'm used to flickr
  • [17:23:49] * dirk2 (n=dirk@F301f.f.strato-dslnet.de) has left #beagle
  • [17:26:35] * mickey|zzZZzz (i=mickey@openmoko/coreteam/mickey) has left #beagle
  • [17:28:36] * jkridner|work (n=a0321898@nat/ti/x-73c2887ba55548e3) Quit (Remote closed the connection)
  • [17:29:16] * jkridner|work (n=a0321898@nat/ti/x-9316420650938f8b) has joined #beagle
  • [17:42:16] * cian (n=cian@cian.ws) Quit ()
  • [17:59:55] * JuanG_ (n=Juan@nat/ti/x-69e6f2ba6eb48698) has joined #beagle
  • [18:05:17] * JuanG_ (n=Juan@nat/ti/x-69e6f2ba6eb48698) Quit (Remote closed the connection)
  • [18:12:50] <mru> good afternoon
  • [18:13:26] <Crofton|work> gm
  • [18:43:15] <koen> mru: I see you did some commit to your omap tree
  • [18:43:29] <koen> mru: does it solve the crashes, or is it just cleaning up patches?
  • [18:48:24] <mru> I've only cleaned up the clock setting a bit
  • [18:48:47] <mru> no change in actual runtime settings with my config
  • [18:51:20] <ds2> what happens on monday?
  • [18:51:33] <mru> weekend ends
  • [18:52:06] <ds2> it almost sounds like the voyage of the TI Beagle begins
  • [18:54:21] <koen> I suspect there will be a press event on monday
  • [18:54:39] * koen suspects reuters HQ needs to look out for TI employees dressed in beagle suits
  • [18:57:14] <ds2> Hmmm
  • [19:01:03] * RogerMonk (n=a0740758@nat/ti/x-4d16e5507e9825d0) Quit (Remote closed the connection)
  • [19:05:52] <bjdooks> hmm, must get boss to place digikey order before he goes on holiday friday
  • [19:06:02] * bjdooks has components he needs to complete projects
  • [19:14:35] <sakoman_> why the speculation about Monday? Did I miss some announcement?
  • [19:14:51] <ds2> read your mail?
  • [19:16:57] <sakoman_> Ah, the Digikey stuff?
  • [19:17:03] <ds2> *nod*
  • [19:18:18] <sakoman_> Has anyone here successfully built a custom x-load from sources?
  • [19:18:45] <ds2> has anyone drunk frong the holy grail?
  • [19:18:49] <ds2> =)
  • [19:18:55] <ds2> s/frong/from/
  • [19:20:17] <mru> I've built x-load from source
  • [19:20:40] <mru> I couldn't figure out why it kept trying to load uboot from sd, so I disable sd support
  • [19:27:13] <sakoman_> mru: any issues?
  • [19:27:27] <mru> with x-load? no, none at all
  • [19:27:48] <mru> except the non-ability of then-current uboot to flash it
  • [19:28:32] <sakoman_> mru: I'm working with some new hw, so I'm trying to build a version to use with USB peripheral boot mode
  • [19:28:51] <mru> oh, I didn't try anything fancy like that
  • [19:29:08] <prpplague> i keep telling you guys, you should try apex, the nand boot features are pretty solid
  • [19:29:31] <mru> what's apex?
  • [19:29:44] <ds2> he he he
  • [19:29:49] <prpplague> http://www.elinux.org/APEX
  • [19:30:38] <mru> I see no mention of omap there...
  • [19:31:03] <mru> and besides, x-load/u-boot is working fine
  • [19:31:05] <ds2> did someone just volunteered to write the OMAP init code for APEX? ;)
  • [19:31:09] <sakoman_> Things aren't well documented on the peripheral boot side. I *think* it wants an unsigned binary, but I'm not sure :-(
  • [19:32:27] * bernrdo (n=bfanti@71-212-209-219.hlrn.qwest.net) has joined #beagle
  • [19:38:48] <prpplague> mru: yea there is no cortex-a8 support currently, but porting it is soooo easy
  • [19:39:09] <prpplague> ds2: as soon as i have time to get a beagle board, that and openocd support are on my agenda
  • [19:39:53] <prpplague> ds2: just have no time right now with the m8050 soon to be released
  • [19:40:03] <mru> I'll stick with what works, and the things I can fix
  • [19:40:21] <mru> I have a working boot loader, so I don't need another just now
  • [19:40:31] <mru> but more choice can't be bad
  • [19:41:05] <prpplague> mru: i did the apex support for the s3c2410 and s3c2440
  • [20:06:47] <denix> sakoman_: are you building x-loader for beagle or for evm?
  • [20:07:46] <sakoman_> denix: actually, neither. I am doing board bringup on a clients omap3 boards
  • [20:10:00] <denix> oh, I see. we've done some changes to x-loader recently, like adding kermit serial download support.
  • [20:11:17] <denix> it won't work on beagle, as it can't do serial boot, though
  • [20:13:33] <sakoman_> denix: why can't beagle do serial boot?
  • [20:14:07] <sakoman_> In any event, I am using usb peripheral boot at the moment
  • [20:15:23] <denix> sakoman_: uart3 is missing :)
  • [20:15:36] * cian (n=cianh@cian.ws) has joined #beagle
  • [20:15:58] <sakoman_> I am booting Nishanth's uboot-serial binary and using that to kermit download my u-boot 1.3.4 code
  • [20:16:26] <sakoman_> denix: I thought beagle had *only* uart3!
  • [20:16:51] <denix> how do you load the initial uboot?
  • [20:17:04] <denix> can you load x-loader over the serial?
  • [20:17:05] <sakoman_> usb peripheral boot
  • [20:17:37] <sakoman_> so rom code loads uboot-serial over usb
  • [20:18:04] <denix> but not over serial
  • [20:18:12] <sakoman_> then I run kermit on serial 3 and interact with uboot-serial to download my larger u-boot 1.3.4 code over serial
  • [20:19:18] * Paulius (n=Paulius@unaffiliated/paulius) Quit ()
  • [20:19:35] <denix> so, the first stage is done over usb, not serial. evm can do all of it over serial w/o usb
  • [20:19:36] <sakoman_> denix: I've tried doing usb peripheral boot with x-load images (both signed and unsigned) and they do nothing!
  • [20:20:00] <sakoman_> just like the bits go into a black hole :-(
  • [20:20:45] <denix> afaik, beagle shares some uart3 lines with usb, while uart1 is wired out
  • [20:21:09] <sakoman_> denix: no uart3 is wired out on beagle
  • [20:21:27] <sakoman_> that is the standard serial port for beagle
  • [20:22:14] <denix> ok, I'd need to check the hw manual... :)
  • [20:22:43] <sakoman_> denix: do you know why xload images downloaded with usb peripheral boot won't give any signs of life? Not even the Xload banner!
  • [20:23:26] <Crofton|work> I wonder if the opening of the beagle board, will have better sales than the batman movie .....
  • [20:29:43] <denix> hmm, not sure if it's changed, but early revs of beagle had uart3 muxed with ULPI, which screws with ROM boot...
  • [20:30:50] <denix> sakoman_: which uart those x-loader images use for the console?
  • [20:31:11] <sakoman_> uart3
  • [20:31:26] <sakoman_> And I am looking on uart3
  • [20:32:14] <sakoman_> If I download Nishant's uboot v2.0 via USB peripheral boot it works perfectly and I can interact with it on uart3
  • [20:32:41] <sakoman_> But xloads just seem to die
  • [20:32:48] <denix> you mean, interact over the serial port...
  • [20:32:51] <sakoman_> yes
  • [20:33:04] <sakoman_> USB just gets the code into sram
  • [20:33:49] <sakoman_> then I switch to serial
  • [20:37:01] <denix> have you built your own x-loader images?
  • [20:38:27] <sakoman_> yes, I have. I've also tried to use the pre-built evm x-load with usb peripheral boot (on the evm) with exacltly the same result
  • [20:39:17] <sakoman_> oops - take that back. I did the pre-built binary experiment on beagle
  • [20:47:00] * cian (n=cianh@cian.ws) Quit ()
  • [20:50:00] * cian (n=cianh@cian.ws) has joined #beagle
  • [20:52:31] <denix> double-checked the hw ref and indeed it says serial port/header is wired to uart3...
  • [20:52:40] <sakoman_> yup
  • [20:54:09] <denix> need jkridner to clarify... :)
  • [20:56:07] <Crofton|work> http://www.balister.org/~balister/shoulder-xray/01.jpg
  • [20:56:14] <Crofton|work> this afternoons project
  • [20:59:59] * maelcum (n=maelcum@e178145099.adsl.alicedsl.de) has joined #beagle
  • [21:00:06] * DJWillis (n=djwillis@82-46-19-72.cable.ubr02.bath.blueyonder.co.uk) Quit (Read error: 104 (Connection reset by peer))
  • [21:00:27] <maelcum> is the usb host port going to come back in future board revisions?
  • [21:00:46] * DJWillis (i=djwillis@82-46-19-72.cable.ubr02.bath.blueyonder.co.uk) has joined #beagle
  • [21:03:03] * BThompson (n=BThompso@nat/ti/x-669d69ec18a4a30f) Quit ("Trillian (http://www.ceruleanstudios.com")
  • [21:04:44] * bazbel1 (n=a0192809@nat/ti/x-f70a81fa79187ad7) has joined #beagle
  • [21:09:28] <ds2> is the USB host port really needed? what application do you have that needs both ports?
  • [21:10:53] * bazbell (n=a0192809@nat/ti/x-3f1b266c19d2c302) Quit (Remote closed the connection)
  • [21:13:54] <maelcum> this board could be used for a small server. servers need hard disks and network, for example.
  • [21:14:45] <sakoman_> Crofton: looks broke to me!
  • [21:14:58] <sakoman_> looks "ouch" too!
  • [21:15:08] <sakoman_> bike crash?
  • [21:18:50] <ds2> maelcum: the host-only port is not needed for that
  • [21:19:21] <ds2> hub on the MUSB port will do that just fine
  • [21:19:40] <ds2> anyone know what speed the EVM OMAP is rated for?
  • [21:20:25] <denix> runs at 600
  • [21:20:39] <Crofton> yeah bike wreck
  • [21:20:52] <Crofton> I can't find anything good to compare with
  • [21:21:00] <Crofton> I'll have to wait on real doctors
  • [21:21:14] <sakoman_> Crofton: you really need to stop doing that!
  • [21:21:19] <Crofton> yeah
  • [21:24:56] <ds2> denix: but is it rated for 600?
  • [21:26:53] <denix> what's the difference? :)
  • [21:28:38] <maelcum> ds2: ok, nice. i didn't know. so the port that is available can do host *and* gadget? two ports would still be better for performance, though :/
  • [21:28:50] <Crofton> sakoman, only the one wreck, the shoulder is still painful for some usage patterns
  • [21:29:29] <ds2> yes, the existing port can do host OR gadget; just not simultaenously
  • [21:29:30] * prpplague (n=dave@mail.americanmicrosystems.com) Quit ("Leaving")
  • [21:30:03] <ds2> the host only port has other restrictions on it ... like it is HS only, no FS/LS support
  • [21:31:53] <denix> ds2: http://elinux.org/index.php?title=BeagleBoard#Clocking
  • [21:32:01] <bjdooks> maelcum: not necessarily, if there's only one EHCI controller then it'll be bandwith shared between two ports
  • [21:33:06] <ds2> denix: thanks... guess all OMAP35xx's are rated at max 600MHz? no such thing as a part only good for 400MHz, etc like the x86 world?
  • [21:33:48] * prpplague (n=dave@mail.americanmicrosystems.com) has joined #beagle
  • [21:34:35] <denix> 600 MHz / 430 MHz @ 1.35V
  • [21:34:46] <denix> 550 MHz / 400 MHz @ 1.27V
  • [21:34:53] <denix> 500 MHz / 360 MHz @ 1.2V
  • [21:35:03] <maelcum> i see that the loss of the host-only port is not really a loss.
  • [21:35:05] <denix> that is ARM/DSP
  • [21:35:06] <ds2> I'm used to the x86 world where they play games with labeling
  • [21:35:45] <denix> I though 3503 was rated lower that 3530, but it doesn't look like it from the specs...
  • [21:36:27] <sakoman_> Crofton: did they tell you your clavicle is broken?
  • [21:36:41] <maelcum> does anybody have an idea how well the opengl es hardware can be used for general purpose computations?
  • [21:36:42] <Crofton> now
  • [21:37:04] <Crofton> I made an appointment and they sent me for xrays first
  • [21:37:06] <ds2> unfortunately, the OpenGL stuff is under wraps at the moment
  • [21:37:24] <Crofton> I just got them done, and like any nerd, I post them on the internet :)
  • [21:37:43] <sakoman_> That has got to be really painful!
  • [21:37:55] <Crofton> The cardiologist refuses to diagnose :)
  • [21:37:58] <Crofton> not really
  • [21:38:08] <maelcum> hum. would be awesome to have aes acceleration on the graphics hardware, you know :)
  • [21:38:18] <Crofton> basically, overhead work, some arm motions and combing my hair
  • [21:38:19] <sakoman_> My wife's a nurse. She walked by and said "ouch! broken clavicle"
  • [21:38:23] <Crofton> hahahahaha
  • [21:38:39] <Crofton> I know how to make it hurt a lot :)
  • [21:39:04] <sakoman_> Yeah, just get those two bone ends rubbing!
  • [21:39:07] <Crofton> I sent it to a list with a couple of med students :)
  • [21:40:07] <sakoman_> It might not be broken -- maybe just ripped ligaments where things attach ;-)
  • [21:40:13] <Crofton> yeah
  • [21:40:28] <Crofton> I can't find good comparisons, this internet thing is useless
  • [21:40:52] <Crofton> we should pull that up on a beagle
  • [21:41:52] <Crofton> the stupid CD had to be run in a windows box with IE only
  • [21:43:34] <Crofton> it would be much easier to self diagnose if I had an xray of the other side :)
  • [21:52:56] <sakoman_> Crofton: I think you can be pretty sure that the gap between the two bone ends isn't normal
  • [21:53:19] <Crofton|work> heh
  • [21:53:44] <Crofton|work> most collarbone fractures are out of the joint
  • [21:53:55] <Crofton|work> at the latest I should have an answer Aug 7
  • [21:55:41] <Crofton|work> My friend the ER doctor should have some good comments :) He's always willing to stick his neck out
  • [22:00:24] * dschaeffer (n=daniel@timesys-gw0.cust.expedient.net) has left #beagle
  • [22:04:33] * Paulius (n=Paulius@unaffiliated/paulius) has joined #beagle
  • [22:05:22] * BThompson (n=BThompso@cpe-76-185-93-11.tx.res.rr.com) has joined #beagle
  • [22:06:02] * methril|away (n=Methril@213.27.233.98) Quit (brown.freenode.net irc.freenode.net)
  • [22:06:02] * bjdooks (n=ben@trinity.fluff.org) Quit (brown.freenode.net irc.freenode.net)
  • [22:06:17] * bjdooks (n=ben@trinity.fluff.org) has joined #beagle
  • [22:06:17] * methril|away (n=Methril@213.27.233.98) has joined #beagle
  • [22:34:02] * prpplague (n=dave@mail.americanmicrosystems.com) Quit ("Leaving")
  • [22:53:25] * banderson (n=irc@69.71.183.7) Quit ("Leaving")
  • [23:00:08] * rsalveti (n=salveti@200.184.118.132) Quit (Read error: 113 (No route to host))
  • [23:00:26] * jkridner_ (n=jason@c-76-31-18-64.hsd1.tx.comcast.net) has joined #beagle
  • [23:02:53] * bernrdo (n=bfanti@71-212-209-219.hlrn.qwest.net) Quit ()
  • [23:15:05] * jkridner (n=jason@c-76-31-18-64.hsd1.tx.comcast.net) Quit (Read error: 110 (Connection timed out))
  • [23:35:33] * jkridner|work (n=a0321898@nat/ti/x-9316420650938f8b) Quit ("Leaving.")
  • [23:51:25] * banderson (n=irc@69-71-183-7.mammothnetworks.com) has joined #beagle
  • [23:57:01] * rsalveti (n=salveti@189.70.143.203) has joined #beagle