#beagle IRC Log on BeagleBoard.org

Join the chat at beagleboard.org/chat

IRC Log for 2009-03-20

Timestamps are in UTC.

[00:04:30] <dioioib> if I am going to install a linux distro from here http://people.ubuntu.com/~ogra/arm/babbage/ and have a 16gb SDHC card formatted with 2 partitions what are my next steps. I attempted to use the apt-get method but that yielded no results, so I am attempting this method, but I have never installed a linux distro like this before.
[00:04:40] <akram> ds2: You sure it's not just ARM/Cortex thing?
[00:04:57] <ds2> akram: eh?
[00:05:32] <akram> The commercial win CE platform/
[00:05:45] <akram> I remember seeing it in the RSS feed on the beagleboard.org front page.
[00:05:48] * fraz_ (n=fraz@206.231.255.81) has joined #beagle
[00:05:53] * fraz_ (n=fraz@206.231.255.81) has left #beagle
[00:05:57] <akram> I recall they had tested it on beagle.
[00:06:05] <ds2> no, it might have ben generic OMAP3
[00:06:20] <akram> Hm, that makes more sense.
[00:07:35] <ds2> I see the EVM one
[00:07:48] <ds2> I am almost sure I saw the Beagle one somewhere, might have been at a show
[00:07:50] <akram> This news page: http://www.mpc-data.co.uk/about-us-news/
[00:07:58] <akram> MPC has a custom build on beagle :)
[00:08:26] <ds2> BSQuare is the company I am thinking of
[00:08:35] * florian (n=fuchs@f049168213.adsl.alicedsl.de) Quit ("Verlassend")
[00:08:40] <akram> Their's is generic OMAP3, yes?
[00:08:54] <dioioib> why would you want to run windows CE
[00:09:15] <akram> <-- doesn't, was performing discussion necromancy on something people were talking about earlier.
[00:09:20] <dioioib> it was a horrible os on my toshiba e800
[00:09:44] <akram> I use Angstrom, it is win :)
[00:10:05] <dioioib> really? I heard the audio support wasn't that great, is this true?
[00:10:10] * wiking (n=wiking@info2k1.hu) has left #beagle
[00:11:54] <akram> I had some issues with earlier kernel version, but ALSA works great now.
[00:12:10] <akram> Which is good, b/c the system I have built is primarily interested in audio processing ;)
[00:12:34] <akram> Right now I am trying to figure out how to get expansion header mux to let me use McSPI3 without touching my u-boot.
[00:12:36] <dioioib> ok you are the guy I need to talk to then
[00:12:53] <akram> What are you trying to do?
[00:12:56] <ds2> akram: fix the mux code
[00:12:58] * fulgas is now known as FuL|OUT
[00:13:31] <akram> ds2: am I right when I assume I can't mess with it from userspace? I guess I have to fix the padconf stuff in kernel.
[00:13:53] <ds2> akram: yes, that is right... feel free to implement something, prehaps with sysfs
[00:14:11] <dioioib> well for starters get a linux distro up on the board. I want to run a variety of audio related apps. I code some stuff in processing.org, Max/MSP, and have started to build my own midi controller. I am thinking of the beagle as a sequencer, audio processor and live performance tool
[00:14:41] <akram> ds2: I'm on it ;-)
[00:14:53] <akram> dio: what problems have you run into?
[00:15:42] * a_way (n=ali_as@ambix.plus.com) Quit (Read error: 110 (Connection timed out))
[00:16:03] <dioioib> well it mainly has to do with the creation of the file system I can't wget the files through ubuntu so I downloaded a premade fs and I'm not sure if I can just drop it in. seems too easy
[00:16:06] * Batko_Marto (n=Batko_Ma@141.117.181.96) Quit (Remote closed the connection)
[00:16:13] * mib_sktqym (i=60fc7a40@gateway/web/ajax/mibbit.com/x-5df0b243ab627b2e) has joined #beagle
[00:16:15] * TAK2k4 (n=thomas@dslb-088-072-223-252.pools.arcor-ip.net) Quit (Remote closed the connection)
[00:16:43] <akram> I would recommend installing open-embedded into your ubuntu.
[00:16:53] <akram> If only so you can use patches others have made regarding audio.
[00:17:01] <akram> The pre-made FS should work OK for alsa, though...
[00:17:14] <dioioib> ok... so here is the link i found http://people.ubuntu.com/~ogra/arm/OMAP35x_EVM/
[00:17:32] <akram> ah, you want to use Ubuntu on beagle?
[00:17:51] <dioioib> well I was thinking yes. but I'm not apposed to something else
[00:18:18] <akram> I'm sorry, no idea how to tweak things in Ubuntu -- I got it working nicely with Ogra's instructions, but had issues with apt not working properly.
[00:18:34] <dioioib> ok.
[00:18:41] <akram> I would wait a bit for the armel stuff to mature TBH.
[00:19:51] <dioioib> ok.. so the other method is there anything different I need to do.. obviously you dont need a linux file system right
[00:20:56] * kao (i=577b5e9d@gateway/web/ajax/mibbit.com/x-0960aca4f1ae28de) has joined #beagle
[00:21:08] <kao> hi
[00:21:11] * guillaum1 (n=Guillaum@AMontsouris-153-1-91-205.w90-2.abo.wanadoo.fr) Quit ("Leaving.")
[00:21:33] * guillaum1 (n=gl@AMontsouris-153-1-91-205.w90-2.abo.wanadoo.fr) has joined #beagle
[00:21:44] <dioioib> ok I found the Angstrom link i am going to read through it
[00:23:23] <kao> i'm considering hooking up a beagleboard to my tv, wifi and run mythtv on it...is there any way to get audio out the hdmi connector or would i have to go for analog?
[00:23:33] * mib_8iews2 (i=4dfc12d9@gateway/web/ajax/mibbit.com/x-c64abbe468dd5751) Quit ("http://www.mibbit.com ajax IRC Client")
[00:23:55] <dioioib> the HDMI is not HDMI compliant
[00:24:19] <dioioib> you have to connect to DVI port and run 3.5mm audio jacks for your audio
[00:24:24] <kao> yeah i read that but thought maybe there's some hack anyway
[00:24:48] <dioioib> could be ... not too sure.. I just got mine to be honest and I am no expert
[00:25:01] <dioioib> are you making a media center?
[00:25:17] <kao> don't think my tv set will play analog audio when using one of the hdmi inputs
[00:25:21] <kao> that's the idea
[00:25:54] <kao> or rather just a client. recording and storage on the desktop and playback on a small gadget to reduce cabling
[00:26:07] * cgs (n=carl@130.120.124.202.static.snap.net.nz) has left #beagle
[00:26:10] <dioioib> sounds like a great idea.
[00:27:07] <Russ> dioioib: just because it doesn't carry audio doesn't mean its not "compliant"
[00:27:16] <kao> thx :)
[00:27:47] <kao> guess i'll just have to try and share the results then
[00:27:54] <dioioib> well HDMI compliance requires audio since 2.0
[00:28:14] <Russ> ok, then its not HDMI 2.0 compliant
[00:28:32] <dioioib> ok i will meet you there, its not 2.0
[00:28:53] <dioioib> but couldn't you just jumper from the 3.5mm to the HDMI pin out.
[00:29:08] <Russ> hmmm...I don't think there is a 2.0
[00:29:24] <kao> hmmmm wouldn't i need digital audio?
[00:29:28] <dioioib> well its 1.1 which is the second version
[00:29:34] <kao> interlaced with the video stream?
[00:29:36] <Russ> I think 1.3 is as high as it goes
[00:32:34] <dioioib> the standard talks about uncompressed MPCM
[00:33:18] <dioioib> which "is only supported by some recievers" and "must be set as it is not defult."
[00:33:29] <dioioib> default rather
[00:33:30] * gcohler (n=gcohler@nat/ti/x-539340690626fc0f) Quit (Remote closed the connection)
[00:36:44] <akram> phew, sorry dio -- afk phone.
[00:37:06] <akram> dio: Unless you're going to go crazy and build all your kernel modules etc into the kernel...I think you will want a rootfs.
[00:37:37] <akram> For example, if you use bitbake and customise your Angstrom kernel -- perhaps you will include ALSA driver in your kernel instead of as a driver.
[00:37:43] <akram> (module form)
[00:37:45] <dioioib> np. ok so I have a question where do I find the MLO file
[00:37:59] <akram> http://www.angstrom-distribution.org/demo/beagleboard/
[00:38:36] <akram> I would suggest using 'Narcissus' , however ;-)
[00:38:53] <akram> http://amethyst.openembedded.net/~koen/narcissus/
[00:39:38] <dioioib> build as omap3evm?
[00:40:25] * julianol1ver (n=julian@164.pool85-53-3.dynamic.orange.es) has joined #beagle
[00:40:26] <dioioib> oh wow lots of choice
[00:40:47] * julianoliver (n=julian@164.pool85-53-3.dynamic.orange.es) Quit (Read error: 113 (No route to host))
[00:40:53] <akram> If you are at Narc, choose beagleboard.
[00:41:03] <dioioib> ok
[00:41:35] <dioioib> any packages recommended
[00:43:04] <akram> Options are pretty self-explanatory, but if you want to look them up you can check out package repository.
[00:43:17] <akram> http://www.angstrom-distribution.org/repo/
[00:43:33] <akram> For example, I have no desire for GUI & X, just waste of space & memory.
[00:43:44] <akram> So when I bitbake, I make a console image.
[00:45:35] * kao (i=577b5e9d@gateway/web/ajax/mibbit.com/x-0960aca4f1ae28de) Quit ("http://www.mibbit.com ajax IRC Client")
[00:47:18] <ds2> you would most likely need to use a McBSP to generate the audio cuz I don't think HDMI audio is analog (could be wrong)
[00:47:22] <dioioib> well I have a 16gb card I am more concerned with ram than space
[00:48:42] <akram> dio: the sd-card image that you can get from Narc is only 1GB iirc.
[00:48:45] <dioioib> ds2: I think you are correct the only one that seems even remotely analog is MPCM - the PCM as in PCM Wave, but with M standing for multi channel you are probably 10% correct
[00:49:09] <dioioib> ok so I will have to build my own?
[00:49:22] <akram> you'll need to set card geometry.
[00:49:32] <dioioib> i did that all ready
[00:49:58] <dioioib> I have the fat32 partition set as boot. etc. just no files in them
[00:50:12] <akram> OK.
[00:50:33] <akram> you can use the files from here http://www.angstrom-distribution.org/demo/beagleboard/
[00:50:40] <akram> together with your Narc built FS.
[00:50:54] <dioioib> ok i have them. unpacked.
[00:51:19] <akram> (so, get MLO, u-boot.bin and uImage & x-load)
[00:51:37] <akram> Copy MLO to boot partition first...it MUST be first file copied.
[00:51:48] <akram> after that, u-boot, uImage and x-load in no particular order.
[00:52:00] * davidm (n=David@nat/ti/x-83560c4d3f07898a) has joined #beagle
[00:52:23] <akram> then is quite simple, put your rootfs archive onto your ext partition and unarchive it.
[00:52:29] * davidm (n=David@nat/ti/x-83560c4d3f07898a) has left #beagle
[00:52:50] <dioioib> and the MLO is the only thing I am not clear on
[00:53:20] <dioioib> is that generated by Narciss..
[00:53:45] <akram> rootfs is generated by Narc, yes.
[00:53:54] <akram> Narc will also give you a link to a sample sd-card image.
[00:53:59] <akram> Don't use that...
[00:54:06] <akram> What's not clear about MLO?
[00:54:17] <dioioib> where is it
[00:54:58] <dioioib> never mind
[00:55:03] <dioioib> found it
[00:55:06] <akram> in the Angstrom demo website.
[00:55:10] <akram> Oh good ^^
[00:55:14] <dioioib> lol
[00:58:20] <fenn> my pig will be suspended in neutrally buoyant oxygenated perfluorocarbon solution
[00:58:22] <dioioib> ok so MLO is copied. u-boot u-image
[00:59:12] <dioioib> the x-load.bin.ift should that name remain unchange
[00:59:13] <dioioib> d
[00:59:35] * scary (n=scary@76.14.22.230) has joined #beagle
[00:59:40] <akram> yep.
[01:01:50] <ds2> fenn: stop wasting good bacon!
[01:02:18] <fenn> it's a super-intelligent cyber-pig
[01:02:25] <fenn> anything less would be uncivilized
[01:02:42] <dioioib> akram: what do you use to burn the image?
[01:03:34] <dioioib> you said something about bitbake?
[01:04:27] <akram> bitbake is a build system.
[01:04:36] <akram> rootfs is just a file.
[01:04:44] <akram> so, download it onto your ubuntu.
[01:04:52] <akram> Mount your SD card that you have partitioned.
[01:05:03] <akram> and copy rootfs onto the ext3/2 partition.
[01:05:12] <akram> once there, just extract using tar.
[01:05:21] <dioioib> so its that easy...
[01:05:25] <akram> (tar -xjvf filename probably).
[01:05:26] <akram> Yep.
[01:05:34] <akram> Once it is on, you can attempt to put it into the beagle.
[01:06:28] <akram> If you want to use alsa-aplay, you will need to install alsa-utils-aplay package using the 'opkg' package manager.
[01:06:42] <akram> (aplay package also gives you arecord...)
[01:07:27] <dioioib> so I am going to need a USB network solution for sure.
[01:10:23] <akram> you can configure usb0 to speak ethernet if you need to ;-)
[01:10:40] <akram> But otherwise, yes, you'll need an adaptor.
[01:11:17] <akram> Most usb-2-ethernet adapters that work with the Nintendo Wii also work with Angstrom -- chipset is supported by ASIX driver :)
[01:11:58] <dioioib> ok. I have an old 3-com USB ethernet adapter. but I dont have the correct connection yet
[01:13:24] <akram> Might as well try it, you may be lucky and it will work!
[01:16:58] * mib_sktqym (i=60fc7a40@gateway/web/ajax/mibbit.com/x-5df0b243ab627b2e) Quit ("http://www.mibbit.com ajax IRC Client")
[01:20:29] <dioioib> akram: thanks for the help. I am going to give this a shot after I make my own USB Mini-A to A cable.
[01:22:07] <akram> dio: no problem. Good luck with cable -- they are very cheap from digikey, so I wouldn't bother if I were you :P
[01:29:59] <MostAwesomeDude> Hey, sorry to be a bother. Does anybody have a link to the Nokia-authored DRM code for PowerVR/SGX?
[01:32:29] * joesensport (n=joesensp@218.242.229.234) has joined #beagle
[01:38:05] * khasim (n=a0393720@192.163.20.231) has joined #beagle
[01:44:54] * akram (i=cb2d168a@gateway/web/ajax/mibbit.com/x-38e97565081988f7) Quit ("http://www.mibbit.com ajax IRC Client")
[01:45:14] * joesensport (n=joesensp@218.242.229.234) Quit (Read error: 104 (Connection reset by peer))
[01:45:34] * joesensport (n=joesensp@218.242.229.234) has joined #beagle
[01:48:38] * emeb (n=ericb@ip72-223-90-212.ph.ph.cox.net) has left #beagle
[01:52:23] * lool (n=lool@debian/developer/lool) Quit (Read error: 113 (No route to host))
[02:00:31] * bmxr (n=bmxr@S01060018f3b11a22.vf.shawcable.net) has joined #beagle
[02:10:13] * Wowbagger_ (n=wowbagge@d154-20-187-250.bchsia.telus.net) has joined #beagle
[02:13:16] * akram (i=cb2d168a@gateway/web/ajax/mibbit.com/x-7922d2be2f8206cb) has joined #beagle
[02:19:33] * lool (n=lool@debian/developer/lool) has joined #beagle
[02:21:53] * davidcb (n=davidcb@c-76-124-23-143.hsd1.nj.comcast.net) has left #beagle
[02:24:49] * Olivier (i=4bb6664b@gateway/web/ajax/mibbit.com/x-ba32c44db3dcbf12) has joined #beagle
[02:26:45] <Olivier> Hi! I build the kernel I grab from Android (and from some other places too). I can play audio fine (aplay). But arecord records only noise.... Any idea what I might be missing?
[02:28:30] * Olivier (i=4bb6664b@gateway/web/ajax/mibbit.com/x-ba32c44db3dcbf12) Quit (Client Quit)
[02:29:04] * leslie (n=leslie@116.228.58.74) has joined #beagle
[02:33:55] * mib_dhmr7m (i=4cbadebf@gateway/web/ajax/mibbit.com/x-0c171db0bc8eaa47) has joined #beagle
[02:34:04] * mib_dhmr7m (i=4cbadebf@gateway/web/ajax/mibbit.com/x-0c171db0bc8eaa47) Quit (Client Quit)
[02:34:36] * Tracy (i=4cbadebf@gateway/web/ajax/mibbit.com/x-92960f04298f7e4a) has joined #beagle
[02:35:51] <Tracy> I'm trying to plug in a USB hub to hookup my keyboard/mouse to BB but they dont seem to work. USB hub is powered separately.
[02:36:38] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
[02:37:03] <Tracy> I'm a bit confused about if I need a mini A or mini B connector hooking up to the BB
[02:39:50] <Tracy> Are there any MiniA-to-FemaleA type adapters out there where I can purchase from some store (cant wait for online orders)?
[02:46:27] <Tracy> anyone out there?
[02:49:53] * brolin (n=brolin@190.157.13.75) Quit (Client Quit)
[02:51:02] <dioioib> hello
[02:52:02] <Tracy> hi
[02:52:22] <Tracy> Any beagleboard experts out there?
[02:52:27] <dioioib> i'm not too sure about being able to buy that cable in a store.. its not that common.
[02:52:47] <akram> Tracy: do you want to run the USB port as host? If so, you can plug in miniB, I suspect.
[02:52:50] <dioioib> if you order from digikey and are in the us or canada you'll get the item in one day
[02:52:52] <akram> Otherwise, use mini-A.
[02:52:55] <Tracy> what is the cable that I would need?
[02:53:09] <akram> I use a mini-A (beagleboard side) to mini-B (usb hub side)
[02:53:28] <dioioib> akram do you have a part number on digikey
[02:53:44] <akram> http://code.google.com/p/beagleboard/wiki/BeagleBoardShoppingList
[02:55:27] <dioioib> doesn't have digikey part number I'll find it
[02:56:09] <akram> That cable is not on the list, I think. I have no idea where my last order details are...:-/
[02:56:43] <dioioib> that other place special computing has ridiculous shipping charges
[02:58:44] <dioioib> http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=WM17494-ND
[02:59:58] <dioioib> hope that helps
[03:02:43] <Tracy> Thanks! how long does digikey take to ship?
[03:03:24] <dioioib> 1 business day
[03:03:38] <dioioib> so you probably wont get it until monday
[03:03:46] <dioioib> but shipping is cheep
[03:05:20] * biliquai (n=biliquai@61.6.64.6) has joined #beagle
[03:05:22] <dioioib> keep in mind thats for connecting the beagle to a hub.
[03:05:55] * kozak (n=subbu@117.192.6.225) has joined #beagle
[03:06:05] <dioioib> akram: I am having issues with permissions on the linux file system of the SD card
[03:06:21] <dioioib> it wont let me copy do I need to untar using sudo
[03:10:47] <Tracy> yes untar using sudo
[03:10:57] <dioioib> thanks
[03:12:03] <Tracy> make sure that you untax your fs directly in to your SD card. I untared once to my ubuntu machine and copied them to sdcard and that did not work.
[03:13:52] <dioioib> what was your command line for that. I just tried and its hung up on it
[03:16:52] <Tracy> I think I ued: sudo tar -xvjf fs.tar.bz2
[03:17:50] <Tracy> your partition was correctly created and its ext3?
[03:18:28] <dioioib> yes. i just opened sudo nautilus and it seems to be working
[03:19:28] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
[03:19:30] <dioioib> tracy what version of linux are you running on beagle
[03:20:03] <Tracy> I'm running an angstrom distro...can't recall the version
[03:20:28] <Tracy> I got it from the beagleforbeginners wiki
[03:22:08] <Tracy> http://www.elinux.org/BeagleBoardBeginners (look for angstrom demo images)
[03:22:55] <dioioib> I'm going to try out ubuntu jaunty
[03:23:13] <Tracy> what are you trying to do with beagle?
[03:24:15] <dioioib> well right now play. but I want to use it for audio performance works, I need to install java and processing, and hopefully run max patches on it,
[03:24:27] <dioioib> those are the ideas for now. I will see how it goes
[03:24:53] <dioioib> how about you what are you working on?
[03:25:17] <Tracy> 1080p video decoding...I want to see how BB will handle it
[03:26:28] <Tracy> Since it has a DSP core I'm hoping that I could find some codecs that has HW acceleration so that I can offload the ARM.
[03:27:35] <dioioib> cool, yeah that would be awesome. I wonder if you could cook up a small expansion using the header to do that. run an additional DSP.
[03:28:30] <dioioib> what I would really like to do is some augmented reality stuff.
[03:28:49] <Tracy> maybe...but at least 720p decoding should be possible on BB alone (with no other processor) from what I've read.
[03:29:03] <Tracy> augmented reality?
[03:29:28] <Tracy> ok.....Obama on Jay Leno gotta go...will catch you later
[03:30:17] * Tracy (i=4cbadebf@gateway/web/ajax/mibbit.com/x-92960f04298f7e4a) Quit ("http://www.mibbit.com ajax IRC Client")
[03:30:19] <dioioib> I ordered a bunch of stuff to interface into the bb. I'm a new media artist and the idea of being able to augment reality, through geo-location really interests me. So not VR but an overlay on top of reality
[03:34:55] * ZeZu (n=null@c-98-227-56-17.hsd1.in.comcast.net) Quit (Read error: 104 (Connection reset by peer))
[03:37:08] <fenn> dioioib: i'm right there with ya buddy
[03:37:30] <fenn> i want inventory management and pin labeling for IC's
[03:37:35] <dioioib> there is just so much you can do with this little thing
[03:38:35] <fenn> and all the usual GPS/star identification/wifi radar
[03:39:10] <dioioib> has anyone started development on expansion boards?
[03:39:20] <fenn> why do you need an expansion board?
[03:40:51] <dioioib> well I could use something like a pic 32 to interface some things, I want to make a live performance interface that I can just connect to the bb. I guess I could do it USB but I have the thing partly built all ready. I am just interested in seeing what others have started to do with the BB
[03:43:40] <dioioib> do you know what the polarity of the 5v barrel is center-pos?
[03:43:54] * BThompson (n=Bernie@host-69-146-39-154.static.bresnan.net) Quit (Read error: 104 (Connection reset by peer))
[03:43:59] <fenn> barrel plugs are almost always + center tap
[03:46:54] <fenn> page 86 of the manual shows ground as outside and DC as the center pin
[03:47:53] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
[03:58:28] * kozak (n=subbu@117.192.6.225) Quit (Remote closed the connection)
[04:12:52] <ds> anyone hooked up a gigabit ethernet usb adapter to a beagleboard?
[04:13:34] <ds> at what speed does it max out, and is it cpu or usb limited?
[04:14:26] * bazbell (n=a0192809@nat/ti/x-5fd5a74433fa4f7c) Quit ("Leaving.")
[04:17:53] * cbrake_ (n=cbrake@oh-69-34-21-229.sta.embarqhsd.net) has joined #beagle
[04:22:04] * cbrake (n=cbrake@oh-69-34-21-229.sta.embarqhsd.net) Quit (Read error: 110 (Connection timed out))
[04:22:50] <dioioib> are there any portable screens that are compatible with the BB
[04:23:23] <dioioib> i heard someone mention the PSP screen
[04:42:47] <jkridner> dioioib: there is talk about someone selling one on the mailing list.
[04:43:15] <dioioib> oh... is it touch screen?
[04:48:46] * jsync (n=jess@59.160.172.220) has joined #beagle
[04:59:54] * bmxr (n=bmxr@S01060018f3b11a22.vf.shawcable.net) Quit (Remote closed the connection)
[05:08:13] * Wowbagger_ (n=wowbagge@d154-20-187-250.bchsia.telus.net) Quit ("Leaving")
[05:08:29] <ds2> blah blah blah
[05:08:47] <jkridner> Crofton|work: is this useful http://dsignalp.blogspot.com/2009/03/sca-frameworks.html ?
[05:08:51] <jkridner> hi ds2
[05:17:09] <ds2> hi jkridner
[05:20:32] * khasim (n=a0393720@192.163.20.231) has joined #beagle
[05:39:38] * akram (i=cb2d168a@gateway/web/ajax/mibbit.com/x-7922d2be2f8206cb) Quit ("http://www.mibbit.com ajax IRC Client")
[05:44:19] * jkridner|work1 (n=a0321898@nat/ti/x-06df02e956c13354) has left #beagle
[05:47:08] * robclark (n=robclark@166.128.145.118) has joined #beagle
[05:56:15] <ds2> hmmmm SPI module is confirmed to be seeing the line
[06:03:35] * uwe_ (n=uwe_@dslb-084-056-024-012.pools.arcor-ip.net) Quit (Read error: 110 (Connection timed out))
[06:08:59] * rupeshgujare (n=rupesh@59.160.172.220) has joined #beagle
[06:17:11] * uberfry (n=spinl0ck@vodsl-10890.vo.lu) has joined #beagle
[06:23:51] * ZeZu (n=null@c-98-227-56-17.hsd1.in.comcast.net) has joined #beagle
[06:34:41] * gregoiregentil (n=zonbu@adsl-71-135-118-139.dsl.pltn13.pacbell.net) has joined #beagle
[06:35:41] * uwe_ (n=uwe_@dslb-084-056-024-012.pools.arcor-ip.net) has joined #beagle
[06:40:06] * geckosenator (n=sean@71.237.94.78) has joined #beagle
[06:45:42] * julianol1ver (n=julian@164.pool85-53-3.dynamic.orange.es) Quit (Read error: 113 (No route to host))
[07:06:01] <ds2> ARRRRRRG
[07:06:15] <ds2> is there a errata for 3530 ES3?
[07:08:21] * mib_2kj7oj (i=5640c7e4@gateway/web/ajax/mibbit.com/x-869be8b967ff6015) has joined #beagle
[07:10:34] <AV500> ds: what are you looking for?
[07:10:57] <ds2> the erratas for the ES3 silicon
[07:11:05] <AV500> yes, what area? SPI?
[07:11:20] <ds2> yeah SPI
[07:12:31] <AV500> Section 1.18, ???Data transmitted by McSPI module is 1-bit clock delayed in slave mode???.
[07:12:40] <AV500> Section 1.57, ???SPI Dummy DMA RX request generation???.
[07:12:55] <AV500> Section 1.127, ???McSPI can generate a wrong underflow interrupt???.
[07:13:29] * pmeerw_ (n=pmeerw@195-70-114-97.stat.salzburg-online.at) has joined #beagle
[07:13:33] <MostAwesomeDude> :c
[07:13:44] <AV500> these are the erratas for the McSPI
[07:13:45] <MostAwesomeDude> Are these from the non-free stuff?
[07:13:52] <AV500> ?
[07:14:20] <ds2> none of that explains what I am seeing on McSPI3... settings of MOSI (an output line) to INPUT disable in PADCONF results in a failure to receive on the MISO line but if I turn on thepull up on it, it works
[07:14:21] * AudioFreak (n=hugo@84.27.136.88) has joined #beagle
[07:15:24] <AV500> ds2: then it is either you who found a new one, or you are doing smthg wrong
[07:15:27] <AV500> :-)
[07:16:04] <MostAwesomeDude> AV500: I'm not sure what you're pasting from, but please don't paste non-free stuff verbatim?
[07:16:15] <ds2> McSPI4 is doing similar things
[07:16:22] * denix0 (n=denix@pool-71-255-243-27.washdc.east.verizon.net) has left #beagle
[07:16:39] <AV500> MostAwesomeDude: well, this is the most I would paste
[07:16:51] * rsalveti (n=salveti@pc-a81040.wlan.inet.fi) Quit (Read error: 110 (Connection timed out))
[07:17:07] <ds2> i am throughly irked by this...
[07:17:49] <AV500> ds2: you have to disable MOSI?
[07:17:52] <ds2> disabling the input receiver from PADCONF for output only lines apparently screw up the McSPI
[07:18:30] <ds2> AV500: I have to enable the receiver on at least all 3 McSPI lines or MISO won't work even though MISO reports the correct info in SYS Test mode
[07:19:08] <ds2> this is weird but I have a looping McSPI4 and McSPI3 now
[07:20:48] <MostAwesomeDude> AV500: Ah. 'k. I really can't look at the non-free stuff, and I really don't want to have to stop hanging out here. :3
[07:21:17] <ds2> MostAwesomeDude: why's that?
[07:22:38] <MostAwesomeDude> I'm an X.org developer and hopefully at some point I'll be working on open-source SGX efforts.
[07:23:41] <ds2> ah
[07:24:38] <AV500> MostAwesomeDude: ok, I will not paste any non free SGX info here :-)
[07:30:13] <Russ> AV500: you have some?
[07:31:01] <AV500> Russ: :-)
[07:31:08] * colla (n=ACollama@host198-216-static.52-88-b.business.telecomitalia.it) has joined #beagle
[07:33:53] <raster> MostAwesomeDude: how do u plan to do that? reverse engineer it?
[07:34:52] <MostAwesomeDude> raster: Probably. I'll cross that bridge when I come to it.
[07:36:30] <AV500> note the current trouble with the Poulsbo SGX Driver...
[07:36:43] <Russ> AV500: is the pasu or whatever stuff that greg kh touching related in any way?
[07:36:52] <AV500> pasu?
[07:37:07] * sauravbh_gandhi (i=3ba217dd@gateway/web/ajax/mibbit.com/x-0f7e51eb66f1caa6) has joined #beagle
[07:37:14] <Russ> paulsbo
[07:37:31] <Russ> I can never remember the word, much less the abbreviation in xf86-driver-puloerho
[07:37:33] <Russ> whatever
[07:37:36] <AV500> yes, poulsbo graphics is SGX as well
[07:37:45] <Russ> but how similar
[07:37:46] <AV500> just a different version
[07:38:57] <AV500> intel is very much behind Linux in the atom/poulsbo scheme, so It will be interesting to see how they will handle the SGX issue
[07:39:02] <AV500> for a linux driver
[07:39:47] <Russ> so would that be a based gallium effort?
[07:40:10] <MostAwesomeDude> I'd do Gallium, yeah.
[07:40:17] <MostAwesomeDude> But that's mostly because I'm tired of classic Mesa.
[07:40:41] <Russ> would there be a 2D state tracker, or is NEON faster at that stuff?
[07:42:13] <MostAwesomeDude> Yes and yes.
[07:42:41] <Russ> Maybe I'm asking that question wrong
[07:42:49] <Russ> would the SGX core be involved in any of the 2D state tracker operations?
[07:42:50] <MostAwesomeDude> The driver would support 2D ops on the 3D engine, and stuff like VDPAU or XvMC could be done through Gallium.
[07:44:40] <MostAwesomeDude> If you're thinking in terms of Xserver, we've got an xorg state tracker that turns any KMS-capable driver into a modesetting DDX.
[07:45:27] <MostAwesomeDude> But I'm getting ahead of myself. The first thing to do would be to consolidate the omapfb, Poulsbo, and SGX DRM code into one KMS-based DRM module. That'd be a lot of work in and of itself.
[07:46:26] <Russ> I really need to get paid to work on stuff like that
[07:46:53] <MostAwesomeDude> Me too.
[07:47:38] <Russ> I don't do any cool stuff in my current job, mainly system engineering and some small amount of code, 99% of which gets thrown away
[07:47:47] * sauravbh_gandhi (i=3ba217dd@gateway/web/ajax/mibbit.com/x-0f7e51eb66f1caa6) Quit ("http://www.mibbit.com ajax IRC Client")
[07:48:46] <MostAwesomeDude> I get paid very small amounts of money to make things work. If I got paid more, or if things worked, I'd have time for cool stuff.
[07:48:52] <Russ> maybe I could get redhat to pay me to work on poulsubo...it would probably help if I could spell it
[07:52:47] <AV500> http://poulsbo.net/
[07:53:52] <AV500> see what you are missing when not working [in|on] Poulsbo: http://poulsbo.net/photos/bunads_norway06.jpg
[07:54:26] <Russ> heh
[07:54:38] <Russ> so is GL ES also not an open thing?
[07:55:35] <MostAwesomeDude> GLES is an open standard with (AFAIK) no open implementations.
[07:55:54] <Russ> hmmm...its bad when you do a modprobe remotely on a distant machine and hear a distant beep...
[07:56:20] <MostAwesomeDude> XFD
[08:01:52] <kulve> MostAwesomeDude: there's open source sw implementation for open gl es 1.1 (but nothing to 2.0 afaik)
[08:02:41] <Russ> would there be much point to an open gl es state tracker?
[08:02:58] <MostAwesomeDude> Sure. If done in Gallium, it would be usable on all Gallium drivers.
[08:03:11] <MostAwesomeDude> (Sure, there's only one complete Gallium driver so far, but still...)
[08:03:57] <Russ> I know, but the only software I know of that uses GL ES are the SGX demos and games ported for use with the SGX GL ES lib
[08:04:23] <raster> MostAwesomeDude: poor bugger. reverse engineering is.. slow. but..., good luck!
[08:04:25] <MostAwesomeDude> There's lots of proprietary code that deals in GLES.
[08:04:28] <kulve> Russ: and maemo 5 alpha sdk
[08:04:36] <raster> Russ: i have a gles2 engine for evas
[08:04:39] <MostAwesomeDude> raster: If we can RE nVidia, we can RE anything.
[08:04:45] <raster> Russ: and clutter too
[08:04:56] <raster> MostAwesomeDude: sure! it's just not too fast
[08:05:13] <Russ> raster: well, at least the nueavo people have come up with a lot of code to assist with reverse engineering
[08:05:33] <raster> MostAwesomeDude: (compared to writign based off specs)
[08:05:42] <raster> Russ: sure!
[08:06:12] <raster> right now though.. i'm a little devoid of sgx's to poke as my bb died
[08:06:20] <raster> so i'm just poking neon in the meantime
[08:06:42] <Russ> raster: did you send it back yet?
[08:06:50] <raster> Russ: no...?
[08:06:54] <raster> why?
[08:06:59] <Russ> to get a replacement
[08:07:02] <raster> my guess it was the cat sleeping on it that blew it up
[08:07:10] <raster> i dont think its a warranty issue
[08:07:23] <raster> i could order another
[08:07:29] <raster> but with revc around the corner.. why?
[08:07:34] <raster> and i have a pandora on order
[08:07:35] <MostAwesomeDude> Russ: renouveau, the tool used to RE 3D registers, works on Radeons and probably can be ported to SGX as well.
[08:07:43] <Russ> http://groups.google.com/group/beagleboard/topics
[08:07:43] <raster> so by the time i can get a revc.. a panny will turn up
[08:07:46] <raster> or should
[08:08:52] <Russ> "You should have gone the RMA route before taking the part off and been honest up front with what you had done. I would have had them send you a replacement board where the total cost to you would have only been time." -- Gerald Coley <gerald@beagleboard.org>
[08:09:14] <Russ> RE: "[HW] 12V Power -> Smoke on TPS2141"
[08:10:04] * RobertK (i=3edc0622@gateway/web/ajax/mibbit.com/x-dc62d6d69cd9db22) has joined #beagle
[08:10:09] <RobertK> good morning
[08:10:34] <RobertK> I always used IMAGE_INSTALL += ... in a recipipe to install additional software into an image.
[08:11:02] <RobertK> but now bitbake fails: "Cannot find package setterm."
[08:11:06] <RobertK> What can I do?
[08:11:09] <raster> Russ: dunno. doesnt seem right. it shouldnt be the cost to the maker that my can blew the bb up
[08:11:25] * methril|away is now known as methril|work
[08:11:37] <Russ> If I was running beagleboard, I'd send you a new board
[08:11:38] <RobertK> example: http://widget.mibbit.com/pb/72odMP
[08:11:52] <raster> (well to be honest i dont know what did - but i was away for a week, bb stayed plugged in on my desk - desk is glass so wont conduct, and when i came back bits of my bb were burnt out)
[08:12:19] <methril|work> good morning
[08:12:25] <Russ> RobertK: there was a massive rename in oe recently
[08:12:25] <Russ> RobertK: go to the oe channel and/or lists
[08:12:27] <raster> (my guess is my cat did it - i've shoed him off my bb before as he decided to nap on it)
[08:12:39] <Russ> packages -> recipes
[08:12:59] <Russ> then maybe I'd send you a beagleboard along with a plexiglass case
[08:13:41] <raster> hehehhee
[08:13:59] <raster> i think i temporarily blew an overo - tyho it seems to have come back to life
[08:14:07] <raster> via powersupply
[08:14:12] <raster> pluggined 110v into 240v
[08:14:22] <Russ> raster: If you want to pay for one, but want it to be a revc, then email the beagle people and explain your situation, I bet they'd send you an early one
[08:14:34] <raster> but thats kind of odd as since i'm in .au i should have been sent a 240v capable supply anyway
[08:14:46] <raster> so i assumed it was at least universal - didnt read the fineprint
[08:15:02] <raster> anyway - for now i'm fine - i have an overo
[08:15:09] <raster> its a 3503 and gets me to play with neon stuff
[08:15:10] * FuL|OUT is now known as fulgas
[08:15:13] <raster> thats good enough for now
[08:15:24] <raster> i'll worry about the sgx / dsp stuff later
[08:15:31] <raster> got a todo list thats long enough :)
[08:15:56] <koen> raster: http://www.catswhothrowupgrass.com/kill.php
[08:16:22] <koen> raster: the "sleeping on your electronics" section
[08:17:17] <raster> koen: HAHAHAHHAA
[08:17:20] <raster> just saw that
[08:17:26] <Russ> my rabbits eat my cables
[08:17:44] <pmeerw_> hi
[08:17:45] * mckoan|away is now known as mckoan
[08:17:48] <Russ> I now have bluetooth headphones
[08:18:01] <raster> hahahahahhaha
[08:18:05] <pmeerw_> trying to build console-image with openembedded
[08:18:15] <pmeerw_> ERROR: Nothing PROVIDES 'console-image', any idea?
[08:18:43] <Russ> there was a massive rename in oe recently
[08:18:43] <Russ> go to the oe channel and/or lists
[08:20:03] * a_way (n=ali_as@ambix.plus.com) has joined #beagle
[08:20:15] * mpoullet is now known as mpoullet|work
[08:20:31] <pmeerw_> thx, oe is a mess
[08:20:38] <mpoullet|work> good morning
[08:31:22] * as_leep (n=ali_as@ambix.plus.com) Quit (Read error: 110 (Connection timed out))
[08:38:59] * recalcati (i=5e51e963@gateway/web/ajax/mibbit.com/x-699e2abcdfdde154) has joined #beagle
[08:41:03] <RobertK> argh, there is no setterm in angstrom anymore...
[08:43:44] <RobertK> I have to do a 'bitbake lsof' before I can use lsof in IMAGE_INSTALL += ! Amazing,
[08:43:51] * koen_ (n=koen@s55917625.adsl.wanadoo.nl) has joined #beagle
[08:47:31] * mib_vzefm6 (i=7dbe2633@gateway/web/ajax/mibbit.com/x-3a85408c0a7cfaeb) Quit ("http://www.mibbit.com ajax IRC Client")
[08:47:53] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
[08:54:15] * ceyusa (n=ceyusa@cm216016.red83-165.mundo-r.com) has joined #beagle
[08:55:16] * koen (n=koen@s55917625.adsl.wanadoo.nl) Quit (Read error: 110 (Connection timed out))
[08:57:43] * koen_ is now known as koen
[09:00:18] * RogerMonk (n=a0740758@nat/ti/x-687f1fca73c4b5b6) has joined #beagle
[09:01:22] * likewise (n=likewise@atwork-193.r-212.178.107.atwork.nl) has joined #beagle
[09:03:16] <like|away> good morning
[09:06:43] * khasim (n=a0393720@192.163.20.231) has joined #beagle
[09:09:42] <mru> morning
[09:10:21] * PhastPhrog (n=chatzill@194.193.86.112) has joined #beagle
[09:11:06] <koen> good morning all
[09:12:49] * khasi1 (n=a0393720@192.163.20.231) has joined #beagle
[09:12:54] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
[09:13:11] <raster> mru: boo
[09:13:17] <raster> koen: crazy dutchman
[09:13:30] <raster> mru: its a 20 cycler hit to go neon -> arm, but not the other way - right?
[09:13:54] <AV500> gm
[09:13:56] <mru> right
[09:14:17] <mru> it's 20 cycles during which you can continue executing neon instructions
[09:14:22] <mru> but not arm
[09:14:41] <raster> hmm
[09:14:52] <raster> looks like i am going to have to load all my data arm-side
[09:14:58] <raster> as i need to be able to branch based on input data
[09:15:20] <raster> and loading to neon then moving to arm for a cmp just wont be happy-lucky-fun-times.
[09:15:24] <raster> AV500: boo!
[09:15:58] <Russ> raster: don't forget the dsp...oh, wait, the cat ated it
[09:16:47] <raster> Russ: see what trouble u get into when you mix electronics and pussy?
[09:17:08] <Russ> raster: usually if you do a cmp, it precedes a memory load or branch, maybe you could do a prefetch doing those 20 cycles?
[09:17:18] <Russ> or you can't prefetch with NEON..?
[09:17:56] <raster> prefetch is either pld or ple land
[09:18:01] <raster> havent played with ple yet
[09:20:09] <raster> but in this case
[09:20:25] <raster> i'm actualyl trying to avoid loads and/or stores by the cmp
[09:20:37] <raster> as well as avoid a chunk of computing
[09:20:49] <raster> i have 1 load i have to do no matter what
[09:20:56] <raster> but based on that i could take 1 of 3 paths
[09:21:14] <raster> 1 with another load and compute and store, 2 with just a simple store, and one with a nop
[09:21:27] <AV500> raster: you take the alpha=0|255 decision per pixel?
[09:21:33] <raster> at least in c.. it brings real speedups in real life cases
[09:21:37] <raster> AV500: yup
[09:21:45] <raster> it actually really helps
[09:21:53] <raster> in "realistic image" cases
[09:22:07] <raster> of course if all alpha >0 && < 255
[09:22:13] <raster> you end up in the pathological case
[09:22:20] <raster> and it hurts
[09:22:54] <AV500> so, you make a heuristic that guesses alpha :-)
[09:23:05] <AV500> for at least 16 pixels or so
[09:23:22] <raster> oph for neon i'd make this decision for a 4 pixle run
[09:23:23] * tzhau (n=tzhau@dsl-245-92-01.telkomadsl.co.za) Quit (Read error: 104 (Connection reset by peer))
[09:23:39] <raster> or more as its actually sanest to use neon to do blobs of pixels at once
[09:23:48] <raster> for c i do it per pixel
[09:24:09] <raster> what i really need is actually to split images into tiles
[09:24:12] <raster> and have a tile flag
[09:24:42] <raster> 16x16 tile is 100% solid, 100% trans, or its a "mix" and if a mix - just count the # of pixels that are > 0 && < 255
[09:24:53] <raster> and based on that number give it a "pathological" or not flag
[09:25:04] <raster> the problem here is. that works... if u never do a transform
[09:25:12] <raster> u can generate such a tile array on load
[09:25:38] <raster> if u do transforms (scaling for example) this all g3ets mighty-hairy
[09:25:50] <raster> of course.. a scale cache (keep a scaled copy) helps
[09:25:56] <raster> dont have that currently
[09:25:56] <raster> :)
[09:26:08] <AV500> I see.
[09:26:30] <AV500> but with neon doing only the path case, how is it compared to C and the if(0|255)
[09:27:53] <raster> well brute-force neon right now is much slower than c
[09:28:09] <raster> and even if all i do is do a simple copy and skip the blend calcs
[09:28:12] <raster> neon is slower than c
[09:28:39] <koen> you'd expect neon to be good at blend stuff
[09:28:40] <raster> the c i think is beating the neon as it gets the if(0|255) case to save it work at least in the if (0) case
[09:28:49] <raster> i havent tested neon vs "brute force" c
[09:28:54] <raster> koen: i would
[09:29:05] <raster> whihc is why currently i just think i have something wrong in my code
[09:29:19] <koen> have you checked for pipeline stalls?
[09:29:35] <raster> but... i'm distracted off in implementing select/cut/copy/paste on entry widgets in elementary for a touchscreen
[09:29:46] <raster> (so u can still finger-scrol land select and copy/paste etc.)
[09:29:53] <raster> koen: nup. havent yet
[09:30:07] <raster> i need to get this done first tho
[09:30:19] <Russ> gcc is pretty good at re-ordering around hazards
[09:30:45] <koen> raster: did you see the iphone OS 3.0 copy/paste stuff?
[09:30:47] <raster> gcc does a pretyt good job of dealing with the evil macro hell that is the c blender
[09:31:01] <raster> koen: i havent yet. i dont want to pollute myself
[09:31:06] <raster> i am doing this "cleanroom"
[09:32:06] <koen> thankfully people can't cleanroom RE awesome air
[09:32:38] * Russ wonders if apple has any touchscreen copy/paste patents
[09:33:00] * like|away is now known as like2wise
[09:33:04] <raster> koen: awesomeair is a genius idea
[09:33:26] <raster> Russ: i dont want to know! dont tell me!
[09:33:36] <Russ> raster: it doesn't matter if you don't know
[09:33:53] <Russ> thats what sucks about software patents, not knowing is actually worse
[09:33:59] <raster> difference between triple damages or not
[09:33:59] <raster> :)
[09:34:26] <raster> thus i happily wish ti remain ignorant
[09:34:30] <raster> :)
[09:35:44] <AV500> raster: my resident Mr brains here has some ideas:
[09:36:05] <AV500> with neon, you have to interleave neon and arm instrcutions
[09:36:23] <AV500> so, let the neon calc the 4pel 0|255 decision result
[09:36:37] <AV500> at the same time, let the arm speculatively calc the 0|255 case
[09:36:50] <raster> oh - my code doesnt even have the compare yet
[09:37:02] <AV500> then pass the neon result back to the arm, but continue to calc the next 4pel on neon
[09:37:09] <raster> i need to add it to give the neon a chance to be "on par"
[09:37:26] <AV500> then let the arm eventually get the result and use the speculative thing or call the neon for real alpha
[09:37:29] <raster> hmm
[09:37:36] <raster> passing neon result back to arm will be a 20 cycle hit
[09:37:40] <raster> on the arm
[09:37:47] <AV500> all this of course interleave so that the 20 cylce hit is masked by the interleaving
[09:37:52] <raster> hmmm
[09:37:56] <raster> tricky
[09:38:10] <AV500> raster: yes, but if you do not use the result immediately, the 20 cycle hit might be hidden, no?
[09:38:16] <raster> as there is 0 hit arm -> neon
[09:38:20] <raster> i was more thinking
[09:38:25] <AV500> doesnt it happen in "the background"?
[09:38:28] <raster> load 4 pixels inot arm regs
[09:38:30] <raster> (movm)
[09:38:47] <raster> then move them to neon
[09:39:35] <raster> bitshift right 24 each and add them
[09:40:01] <Russ> a diagram of the structural units in the pipeline would be helpful
[09:40:21] <raster> cmp them to 0 or 765
[09:40:33] <raster> (in the meantime do some of the first stages of the neon blend op)
[09:40:44] <raster> tho the first stage involves a dst fecth - whihc i'm trying to avoid
[09:41:17] <AV500> raster: with all the data likely in cache, wouldnt a two pass approach work?
[09:41:17] <raster> this way i have no 20 cycle hit anywhere
[09:41:47] <AV500> let the neon just calc the 0|255 decision in 1st pass, then in the 2nd pass call the neon only for the pel groups that need it
[09:42:09] <raster> hmm
[09:42:22] <raster> that'd kill cache as u load in a whole cachline which loads in consecutive pixels
[09:42:24] * tzhau (n=tzhau@dsl-245-92-01.telkomadsl.co.za) has joined #beagle
[09:42:25] <raster> u will do this twice
[09:42:29] <raster> unless u mean for a scanline
[09:42:33] <raster> then it might be possible
[09:42:39] <raster> (not a whole source)
[09:42:52] <AV500> raster: yes, but there is more than one cacheline in the cahce
[09:43:02] <AV500> use another one for the intermediate result
[09:43:11] <raster> hmm
[09:43:28] <raster> that may work...
[09:43:30] <AV500> all your pixels stay in cahce
[09:43:41] <raster> i'm wondering how i should do this
[09:43:48] <AV500> no idea :-)
[09:44:14] <raster> run over it with c just to make the 0|255|other decision per blokc of 4 src pixels
[09:44:22] <raster> but how to store such a result efficiently
[09:44:32] <raster> (for the next pass with neon)
[09:44:42] <AV500> hmmm
[09:44:52] <raster> a pointer per block of 4 would work
[09:44:58] <koen> "Also, I really hate how embedded Linux takes away one of coolest things about Desktop Linux: ability to compile own kernel."
[09:45:03] <raster> with the ptr simply being the appropriate neon routine
[09:45:03] <koen> mozilla people are funny
[09:45:24] <raster> koen: bizarre mozilla people indeed
[09:45:28] <raster> where did u see that?
[09:45:34] <koen> http://blog.mozilla.com/tglek/2008/12/23/fennec-a2-performance/
[09:45:44] <raster> AV500: i actually wonder.. a single run should fit in l1 cache most of the time
[09:45:57] <raster> src pixels anyway
[09:46:02] <raster> 32k l1
[09:46:17] * mib_8iews2 (i=4dfc12d9@gateway/web/ajax/mibbit.com/x-d500822b55fdbe8e) has joined #beagle
[09:46:19] <Pavlov> compiling your own kernel for the n810 is rocket science
[09:47:26] <AV500> Pavlov: so, at least its not rocket surgery
[09:47:47] <raster> hmmm
[09:47:48] <Stskeeps> Pavlov: nah, it's not -that- bad. it could be worse :P
[09:47:53] <Pavlov> it is pretty bad
[09:47:57] <raster> actually this address thing seems a good call
[09:47:58] <uberfry> anyone here got ADD?
[09:48:06] <raster> either that or a byte array of values 0/1/2
[09:48:18] <AV500> raster: yes,
[09:48:39] <raster> uberfry: all the time. they oftne come along with SUB, MUL, XOR and MOV... and other friends
[09:48:39] <AV500> thisis like your tile meta info
[09:48:45] <uberfry> :P
[09:48:52] <raster> AV500: its an interesting idea
[09:48:54] <uberfry> alright: anyone got ADHD?
[09:48:56] <AV500> advanced dungeons and dragons?
[09:48:57] <raster> c seeds the cache
[09:49:02] <raster> neon just bails thru it
[09:49:12] <Stskeeps> Pavlov: qwerty12 from #maemo is pretty capable for questions if you have any :)
[09:49:15] <raster> in fact i could also queue fetches of dest too
[09:49:58] * Pavlov has enough other problems -- doesn't need to try recompiling his n810's kernel again anytime soon
[09:49:59] <raster> uberfry: you just were in at a time when theres a topic of assembly7 going around :)
[09:50:07] <uberfry> lol
[09:50:23] <uberfry> I know enough assembly, thanks ;)
[09:50:32] <uberfry> for multiple architectures, too
[09:50:46] <uberfry> but I need to ask someone with ADHD some questions
[09:50:50] <AV500> uberfry: you want ADHD or you have it and want to sell?
[09:51:22] <uberfry> need the drugs ;P
[09:51:28] <uberfry> but they're illegal here
[09:51:41] <uberfry> so even if you have ADHD, you can't get the drugs here :(
[09:53:10] <koen> vlad_: ping
[09:54:12] * Pavlov is sure he's asleep
[09:55:56] <uberfry> AV500: so you have ADHD?
[09:56:12] <AV500> uberfry: not that I know of
[09:56:27] <uberfry> ugh, I need to get some, asap
[09:56:44] <uberfry> I fucking hate this country, can't even get any kind of acids to decap chips
[09:57:02] <uberfry> yet, kids are allowed to drink at the age of 16...
[09:57:11] <AV500> .de and .fr should be pretty close, no?
[09:58:22] * mib_8iews2 (i=4dfc12d9@gateway/web/ajax/mibbit.com/x-d500822b55fdbe8e) Quit ("http://www.mibbit.com ajax IRC Client")
[09:59:13] <uberfry> yes
[09:59:19] <uberfry> but .fr it's illegal aswell...
[09:59:22] <uberfry> *in
[09:59:38] * Russ (i=foobar@ip70-176-253-20.ph.ph.cox.net) Quit (Read error: 110 (Connection timed out))
[09:59:42] <uberfry> benelux and .fr made amphetamines and the likes illegal apparently...
[09:59:47] * khasi1 (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
[10:00:01] <uberfry> or maybe the doctor I consulted was just a french bastard talking outta his ass
[10:00:35] <uberfry> it's not like I'd use that stuff forever, it's just that I have 2 exams coming up
[10:00:45] <uberfry> and if I fail either one of them, I'm out of that university
[10:01:05] <uberfry> and if I don't go to that university ,I can never study IC design in my life
[10:01:41] * likewise (n=likewise@atwork-193.r-212.178.107.atwork.nl) Quit ()
[10:02:46] * khasim (n=a0393720@192.163.20.231) has joined #beagle
[10:03:18] * colla1 (n=ACollama@94.160.8.44) has joined #beagle
[10:03:49] <koen> uberfry: do a zero-g flight
[10:04:07] <koen> uberfry: in france they add amphetamines to the anti-nausea stuff to keep you sharp
[10:04:39] <koen> uberfry: in the US they don't do that, so the people doing the experiments are all groggy from the anti-nausea meds
[10:05:21] * colla2 (n=ACollama@host198-216-static.52-88-b.business.telecomitalia.it) has joined #beagle
[10:05:21] <uberfry> ic...
[10:05:42] <uberfry> hrm, I just took 2 guarana bottles
[10:05:44] <uberfry> hope it works
[10:05:54] * koen would like some anti-nausea medicine since he's looking at mozilla code for arm
[10:06:11] <koen> drat, I forgot to bring back some guarana from brazil
[10:07:45] <gletelli_> koen: anything wrong with the the mozilla code?
[10:09:59] <Pavlov> koen just likes to complain
[10:10:17] <koen> have a look at this bug https://bugzilla.mozilla.org/show_bug.cgi?id=476903
[10:10:19] * mib_g4pj0i (i=58aac895@gateway/web/ajax/mibbit.com/x-9216b751e9e10961) has joined #beagle
[10:10:45] <koen> that patch only compiles if your compiler is broken
[10:11:11] * mib_g4pj0i (i=58aac895@gateway/web/ajax/mibbit.com/x-9216b751e9e10961) Quit (Client Quit)
[10:11:19] * ogra (n=ogra@ubuntu/member/ogra) Quit (Read error: 104 (Connection reset by peer))
[10:11:25] <gletelli_> :/
[10:11:31] <Pavlov> so, gcc and visual studio?
[10:11:40] <Pavlov> i guess realview works?
[10:11:55] * ogra (n=ogra@ubuntu/member/ogra) has joined #beagle
[10:12:57] <gletelli_> I don't think this patch has been tried with Realview armcc
[10:13:08] <gletelli_> only a previous version of mozilla
[10:13:36] <Pavlov> given the stuff checked in does work everywhere, not sure what compiler that isn't broken it works on
[10:13:53] <Pavlov> er, doesn't work on.
[10:14:09] <koen> mru: I can confirm your findings of revB<7 being flaky compared to B7 and C
[10:14:29] <mru> in what respect?
[10:14:40] <koen> random errors
[10:14:52] <mru> funnily enough, I get no crashes
[10:14:55] <mru> just random output from ffmpeg
[10:15:09] <mru> and usually only with a few specific tests
[10:15:36] <koen> my packagemanager seems to be ably to trigger is (and usually recover)
[10:16:32] <gletelli_> koen: what is your (not-broken) version of compiler that doesn't work with the patch?
[10:17:54] <mru> the patch is just changing some types, no?
[10:18:29] <Pavlov> just casting them differently
[10:18:38] * joesensport (n=joesensp@218.242.229.234) Quit (Read error: 113 (No route to host))
[10:19:03] * colla (n=ACollama@host198-216-static.52-88-b.business.telecomitalia.it) Quit (Read error: 110 (Connection timed out))
[10:24:29] <mru> so how does it not work?
[10:24:42] <mru> granted, there could be alignment issues
[10:24:51] <mru> but that was already the case
[10:24:57] <Pavlov> it does work
[10:25:10] <Pavlov> it has been checked in for a while, and is built on a number of different compilers
[10:25:22] <Pavlov> don't know what koen was on about
[10:25:26] <mru> koen: how does it break?
[10:26:46] * colla1 (n=ACollama@94.160.8.44) Quit (Read error: 110 (Connection timed out))
[10:29:01] * leslie (n=leslie@116.228.58.74) Quit (Read error: 104 (Connection reset by peer))
[10:30:35] <mpoullet|work> I've built opkg from scratch on host to be able to look at the content of the ipks, but I get * Could not obtain administrative lock, any idea?
[10:31:10] <mru> probably no write permission in /var/lock/something
[10:31:10] <Pavlov> mru: if you find out, let me know -- would love to fix whatever!
[10:31:20] <mru> it's koen saying it's broken
[10:31:39] <Pavlov> well, yes, but i think you're more likely to get an answer from him than I am ;)
[10:32:09] * TAK2004 (n=thomas@dslb-088-074-044-174.pools.arcor-ip.net) has joined #beagle
[10:32:49] <mpoullet|work> mru: nothing in /var/lock that looks like opkg, only apache2 and .ramfs...
[10:32:52] <mru> he's probably afk or something
[10:33:05] <mru> mpoullet|work: well, something else then
[10:33:09] <mru> strace it
[10:34:13] <mpoullet|work> mru: good idea :)
[10:36:42] * acmeguy (i=d5b09766@gateway/web/ajax/mibbit.com/x-d3f82414c29b196a) has joined #beagle
[10:38:58] * RobertK (i=3edc0622@gateway/web/ajax/mibbit.com/x-dc62d6d69cd9db22) Quit ("http://www.mibbit.com ajax IRC Client")
[10:39:48] <koen> /OE/angstrom-dev/work/armv7a-angstrom-linux-gnueabi/fennec-1_0.9+1.0b2pre-r3/mozilla-central/xpcom/reflect/xptcall/src/md/unix/xptcinvoke_arm.cpp:134: error: size of array 'arg' is negative
[10:39:55] <koen> that's from the first static assert
[10:41:02] <koen> mpoullet|work: it tries to access /usr/lib/opkg, but an .ipk is a .deb, so dpkg-deb -I foo.ipk and dpkg-deb -C foo.ipk work as well
[10:41:09] <mpoullet|work> mru: strace solved it: sudo mkdir /usr/lib/opkg/ thanks :)
[10:41:38] <mpoullet|work> koen: ah ok, dunno that, good idea too!
[10:42:24] <Pavlov> koen: there is no way for those expressions to be negative
[10:42:31] <Pavlov> sounds like your compiler is hosed ;/
[10:43:23] <Pavlov> http://mxr.mozilla.org/mozilla-central/source/nsprpub/pr/include/prlog.h#259
[10:43:57] <Pavlov> (condition) there has to either be 1 or 0
[10:46:33] <koen> well I know my compiler isn't hosed since the whole system is compiled with it and functioning properly
[10:48:02] <Pavlov> heh
[10:48:02] <Pavlov> OK!
[10:48:20] <geckosenator> I can't start firefox on a 480x272 display
[10:48:36] <geckosenator> since I can't click the ok button which is off the screen to accept the license agreement
[10:49:44] <mru> geckosenator: ;-)
[10:50:05] <AV500> geckosenator: --accept-license ?
[10:50:32] <mru> Pavlov: wtf is that supposed to do?
[10:50:45] <mru> ah, I see
[10:50:48] <Pavlov> compile time fail
[10:50:50] <Pavlov> if not true
[10:50:50] <mru> it's supposed to be an error
[10:50:55] <mru> but that's not valid c
[10:51:18] <mru> it relies on the compiler optimising the constant expression
[10:53:40] <geckosenator> AV500: i have no keyboard only mouse
[10:54:12] <geckosenator> I must have wired something wrong.. tux's feet are blue I think they are supposed to be yellow
[10:54:32] <mru> extern char pr_static_assert[1 - !(condition)] should work
[10:55:49] <mru> or maybe the condition is failing
[10:56:35] <mru> koen: does it work -fno-signed-char ?
[10:56:35] <AV500> geckosenator: maybe they are blue from the cold?
[10:57:58] <AV500> geckosenator: about --accept-license,sorry, I made that up :-)
[10:58:01] <geckosenator> AV500: no
[10:59:17] <koen> mru: trying that now
[11:02:10] * acmeguy (i=d5b09766@gateway/web/ajax/mibbit.com/x-d3f82414c29b196a) Quit ("http://www.mibbit.com ajax IRC Client")
[11:06:09] <koen> mru: mozilla passes in -fsigned-char by itself
[11:06:56] <Pavlov> weird
[11:07:03] <Pavlov> i don't see anything referencing signed-char
[11:07:06] * Pavlov looks
[11:07:38] <koen> it's not in any of my C/LDflafs
[11:07:47] <koen> flags*
[11:09:55] <Pavlov> yeah. i don't see it anywhere in our source tree
[11:10:10] <Pavlov> different gcc's seem to default it differently
[11:11:10] <Pavlov> anyways, i gotta be up in 4 hours
[11:12:44] * maelcum|konv (n=horst@natr.physik.hu-berlin.de) has joined #beagle
[11:19:37] <koen> + for i in $(find ${S} -name "autoconf.mk") ; do
[11:19:37] <koen> + sed -i -e s:fsigned-char:fno-signed-char:g $i
[11:19:37] <koen> + done
[11:19:39] <koen> there we go
[11:19:47] <koen> mru, Pavlov: thanks for the hints
[11:20:02] * koen attempts to remember his mozilla bugzilla credentials
[11:20:49] <recalcati> hi everybody
[11:21:21] <recalcati> anybody is working on DM365 ?
[11:21:23] <koen> hey recalcati
[11:21:36] <koen> recalcati: I'm currently working on dm355
[11:21:54] <koen> recalcati: but RogerMonk mentioned he received a dm365 board, no idea if he's working on it
[11:22:02] <mru> Pavlov: properly written code doesn't make assumptions about the signedness of char
[11:22:15] <recalcati> interesting. but we are starting a project and probably we'll start on DM365
[11:22:41] <koen> recalcati: do some research first, since dm355 only seems to support 2.6.10/OABI
[11:22:53] <koen> I hope dm365 supports something more recent
[11:23:12] <koen> and I *really* hope I can get the dm355 codecs to work on 2.6.29rc7/EABI
[11:23:22] <mru> Pavlov: use explicit signed/unsigned whenever it matters
[11:23:27] <koen> since they only hook into the hardware engines
[11:23:29] <maelcum|konv> mru: i would rather say that properly written code forces signed or unsigned char... or is there a situation in which it makes sense to leave signedness to the platform?
[11:23:31] <maelcum|konv> okay
[11:23:39] <recalcati> you mean 2.6.10 vanilla ? not montavista
[11:23:48] <koen> recalcati: 2.6.10mvista
[11:23:58] <recalcati> ok, thx
[11:24:04] <recalcati> 9000k???
[11:24:12] <recalcati> ops 9k???
[11:24:27] * davidcb (n=Spot@m0f5e36d0.tmodns.net) has joined #beagle
[11:25:06] <koen> recalcati: getting a recent kernel running is fairly trivial (took me 2 hours or so, and I'm not a coder), but I seem to be hitting a wall inside TI for updated codecs
[11:25:22] <maelcum|konv> is there any cool new stuff in 2.6.29(-omap)?
[11:25:25] <koen> to be fair, people at TI are busy preparing for ESC Boston
[11:26:01] <koen> recalcati: http://beagleboard.org/leopard is dm355 based
[11:26:13] <recalcati> thx a lot
[11:27:19] <geckosenator> the hand print in the lower left on the angstrom login screen.. it's supposed to be blue right?
[11:27:51] <geckosenator> I have red and blue flipped.. but maybe you can flip them in software too
[11:30:16] * __alanc__ (n=a-campbe@nat/ti/x-fdfd7cb347c12e0f) has joined #beagle
[11:30:48] <koen> geckosenator: yes, it's supposed to be blue
[11:32:36] <koen> recalcati: the dm3x5 lineup looks really nice if you like 720p mpeg4 playback and encode so I really hope TI gets its act together and starts supporting newer kernels with the codecs
[11:32:49] * mib_2kj7oj (i=5640c7e4@gateway/web/ajax/mibbit.com/x-869be8b967ff6015) Quit ("http://www.mibbit.com ajax IRC Client")
[11:33:19] <koen> recalcati: TI is currently sending patches for the video capture drivers to the linux-davinci mailinglist, so that's a good sign
[11:33:32] <maelcum|konv> btw, does anyone know if TI/the bb is going to be present at linuxtag?
[11:33:41] <maelcum|konv> it's how i first heard of it
[11:33:49] <maelcum|konv> (last year's linuxtag)
[11:34:15] <koen> maelcum|konv: florian is going to be at linuxtag, so atleast one beagle will be there :)
[11:34:27] <koen> no idea on 'official' TI presence
[11:34:29] <maelcum|konv> florian who? :)
[11:34:35] <koen> florian boor
[11:34:41] <__alanc__> koen - we're looking at it...(codecs/git)
[11:34:56] <koen> __alanc__: ah, cool
[11:35:03] <maelcum|konv> okay, i have no idea who he is but i'll probably find him
[11:35:31] <koen> __alanc__: I know I can rebuild the codecs, but then I'd have to unseal the cflags and things
[11:36:03] <koen> __alanc__: which enforces my point that codecs shouldn't be having their own CPU side cflags
[11:36:17] <koen> at least, for a distribution POV
[11:36:37] <koen> from a TI pov it makes sense since 3rd party codecs work on the 'supported' platforms
[11:37:34] <koen> so beagle rev C4 will be A3 sized
[11:37:36] <__alanc__> needed a minor patch to dm355mm (src is avail) module but apart from that was fairly clean
[11:37:49] <koen> with a large FIVE VOLTS ONLY!!! message :)
[11:37:51] <__alanc__> I'll talk to Niclas about pushing it out
[11:38:32] <__alanc__> koen - which tree to push to (Kevin's, your OE baseline)?
[11:39:36] <mru> maelcum|konv: ffmpeg should be at linuxtag, and I plan to bring a beagle
[11:39:37] <koen> __alanc__: I use kevin's tree + vfpe patches from Murali Karicheri
[11:39:56] <maelcum|konv> mru: cool. i will visit you.
[11:40:07] * cbrake_ is now known as cbrake
[11:40:21] <koen> mru: you got your first beagle at LT, right?
[11:40:26] <mru> yes
[11:41:01] <__alanc__> koen : thks
[11:42:44] <uberfry> are you guys any good with binomial heaps n shit?
[11:42:47] <uberfry> sorry for the shit part
[11:42:49] <uberfry> it just slipped out
[11:42:54] <uberfry> I'm kinda hyperactive right now
[11:42:57] <uberfry> :X
[11:42:57] <__alanc__> koen, all - is there a decent calendar of the various events (linuxtag, ESC...) we should have a TI presence at? If yes, we'll look at having ppl at them
[11:45:55] * abitos (n=nixgibts@p5B2E74BC.dip.t-dialin.net) has joined #beagle
[11:49:19] * fulgas is now known as FuL|OUT
[11:53:40] <mpoullet|work> koen: thanks for helping me with my packaged-staging issue, I haven't saved the rest of the deploy dir because it wasn't mentionned in http://pokylinux.org/blog/index.php/2008/07/packaged-staging-and-what-it-gives/
[11:55:38] <__alanc__> Crofton|work - i presented your wife's "if this is the dummies guide I'd hate to see the docs" comment at our Staff Mtg yesterday - everybody got a kick out of it :-)
[11:56:31] * FuL|OUT is now known as fulgas
[11:57:32] <mru> __alanc__: you with TI?
[11:57:49] <Stskeeps> __alanc__: from TI? happen to know why the european version of dummies book ordering ends up with a "Sorry, we can't find that page"? :P
[11:58:51] <__alanc__> Stskeeps : argh - alan silently fumes - I'll ping Marketing (again)
[11:58:59] <__alanc__> mru - yes
[11:59:14] <dioioib> hello guys, I was wondering if someone could direct me to a small portable screen that will work with the BB. Something like the PSP touch screen
[11:59:33] <mpoullet|work> mru: according to the dummy book is __alanc__ is the SDO Applications Manager at TI
[11:59:37] * gcohler (n=gcohler@nat/ti/x-810ccef953f07c46) has joined #beagle
[12:00:04] * mru doesn't have the dummy book
[12:00:10] * mru prefers the TRM
[12:00:18] * maelcum|konv (n=horst@natr.physik.hu-berlin.de) Quit (Remote closed the connection)
[12:00:55] <mru> __alanc__: do you know of any TI plans for linuxtag?
[12:01:02] * mpoullet|work prefers the TRM too but the dummy book is great too
[12:01:22] * AV500 would have loved the dummies book 3 years ago
[12:01:53] * mpoullet|work too as we start working on the davinci
[12:02:11] <koen> __alanc__: one comment I got about the dummies book is that is was completely DSP centered, no real mention or CPU side codecs and workload splitting
[12:02:13] <__alanc__> mru : will talk w/ the guys (and girls!)
[12:02:58] <__alanc__> koen : yep fair point
[12:03:08] <mpoullet|work> koen: yes that's true, the dummy book is too TI tools (link/codec engine) oriented
[12:03:24] <__alanc__> focus was to try & get decent acceleration out of the dsp
[12:03:28] <AV500> mpoullet|work: ??? the book is about TI tools, no?
[12:03:40] <AV500> it is not an ARM9/OMAP3 book
[12:03:43] <uberfry> what differs a DSP from a CPU? FP operations?
[12:04:01] <uberfry> I AM STUPID SORRY
[12:04:21] <__alanc__> uberfry : c64+ dsp has a nice VLIW arch
[12:04:37] <__alanc__> runs vid/img/aud codecs nicely
[12:04:44] <uberfry> ahhhh ok thanks for clearing that up
[12:05:18] <uberfry> I bet it can't make inward singing possible though...
[12:05:26] <mpoullet|work> AV500: well the title is "OMAP and DaVinci Software for Dummies" but it should be "TI OMAP and DaVinci Software for Dummies"...
[12:05:28] <koen> __alanc__: with davinci it didn't matter since the CPU is quite slow, but with omap3 the CPU can do 720p24 :)
[12:05:30] <uberfry> you know inward singing?
[12:05:59] <AV500> mpoullet|work: with TI being the only OMAP and DV vendor, I assumed that :-)
[12:06:41] <koen> the book does a good job at showing why things like RTSC exist
[12:07:06] <koen> it also shows people at TI are bored enough to come up with puns like 'QualiTI' :p
[12:07:49] <__alanc__> koen : yes & no : we could do more w/ the accelerators on the dsp-side from 720p POV
[12:11:59] * recalcati (i=5e51e963@gateway/web/ajax/mibbit.com/x-699e2abcdfdde154) Quit ("http://www.mibbit.com ajax IRC Client")
[12:12:56] * spinl0ck (n=spinl0ck@vodsl-10890.vo.lu) has joined #beagle
[12:13:35] <gcohler> __alanc__, koen: There's a hierarchy: 3503 easiest to use, cheapest, reasonable video performance / 3530 (with video acceleration) harder to use, more expensive, better video performance. TI documentation tends to push the 3530 viewpoint. :-)
[12:14:32] <gcohler> What's incredible to me is how great the performance is on 3503. Because that means that more people can take advantage of the product.
[12:16:05] * ogra (n=ogra@ubuntu/member/ogra) Quit (Read error: 113 (No route to host))
[12:16:10] * ogra (n=ogra@ubuntu/member/ogra) has joined #beagle
[12:18:39] * Openfree (n=dennis@222.65.175.96) has joined #beagle
[12:19:59] * tzhau (n=tzhau@dsl-245-92-01.telkomadsl.co.za) Quit (Read error: 110 (Connection timed out))
[12:20:13] <gcohler> Has anyone tried to run Angstrom on the OMAP L137 board yet?
[12:22:55] * tzhau (n=tzhau@dsl-245-92-01.telkomadsl.co.za) has joined #beagle
[12:23:27] <koen> gcohler: I know of people that did
[12:24:13] <koen> gcohler: but they put the project on hold since their TI supported could only speak sentences containing "only mvista supported"
[12:24:48] <gcohler> Cool. Because I'm finding that while the OMAP 35x is the shiny processor to show off, many people settle for the utilitarian L137. So it would be great to be able to paint a software compatibility train of thought between the two products.
[12:25:26] <koen> the l137 is nice since it has a floating point DSP
[12:25:28] <raster> bah!
[12:25:52] <raster> NOT wanting to show off the shiny bling of the 3530!
[12:26:00] <gcohler> Yes, many people like the floating point.
[12:26:31] <gcohler> raster: Not saying that I don't show off the bling. But many times bling brings them into the store, but they buy the meat and potatos.
[12:26:33] <AV500> raster: show off the smoking riuns of YOUR 3530?
[12:27:10] <raster> AV500: it lost its magic blue smoke! :(
[12:27:29] <raster> gcohler: dunno - but i tend to stick to nice benchmarks
[12:27:41] <raster> and if the bling got me to the store - i'd want the thing thats blinging
[12:27:45] <raster> :)
[12:27:58] <koen> gcohler: l137 is primus or freon?
[12:28:06] <raster> then again... i'm one of these people - if you give me a foot, i'll take a mile :)
[12:28:23] <raster> i.e. - the mroe grunt.. the better
[12:28:27] <gcohler> L137 is primus. L1x8 is Freon.
[12:28:54] * uberfry (n=spinl0ck@vodsl-10890.vo.lu) Quit (Read error: 110 (Connection timed out))
[12:29:15] <gcohler> raster: Yes, but if the bling costs half again as much, and you don't need to drive above 60 mph, then you might choose to save your money to buy food. But that's just me. :-)
[12:29:55] <gcohler> And I believe that a full up 3530 is more than double the price of the L137.
[12:30:32] <raster> gochsince no one put speed limits on my highways.. i'll do the 120mph thanks :)
[12:31:01] <AV500> so, you really want to run X and all on a 300mhz arm9?
[12:31:07] <gcohler> Ah. So you're the one creating 4 kg of CO2/day for the fun of it.
[12:31:37] <gcohler> I'm not sure what I want. Just thinking. Maybe no X. But console might be enough.
[12:31:44] <raster> gcohler: and feeding it to my pet triffid :)
[12:32:08] <raster> though on a serious note
[12:32:13] <AV500> gcohler: don't worry about raster, UV will take care of him eventually
[12:32:13] <raster> what else will be in the system
[12:32:27] <raster> double price for the soc may make little difference if theres enough other stuff
[12:32:34] <raster> and you need to evaluate what u will do
[12:32:35] <raster> and how
[12:35:45] <gcohler> Yes I fully agree. There are cases where the extra cost is minimal compared to the system cost. But there are plenty of folks for whom their projected volumes make cost savings worthwhile.
[12:35:55] * RobertK (i=3edc0622@gateway/web/ajax/mibbit.com/x-75b6d7599b780ab3) has joined #beagle
[12:36:10] <RobertK> good morning again.
[12:36:34] <raster> gch what resolution display will u have?
[12:36:40] <raster> and how do u plan on filling it?
[12:36:50] <raster> (and what are you planning on filling it with)?
[12:37:58] * robclark_ (n=robclark@nat/ti/x-c7429badd60aaa46) has joined #beagle
[12:38:43] <gcohler> Oh, I don't have any real plans. My real goal is to be able to tell people, you can buy L1 or OMAP 35x and either way run the same software base. So your efforts in software development are preserved across the boundaries. Hence, you should have no fear about committing to designing in L1 now to get started and you can always upgrade to 35x when you need it.
[12:38:52] <koen> raster: remember the strongarm? X in 200MHz and quite fast as well :)
[12:39:41] <raster> koen: only if u do qvga :)
[12:40:21] <gcohler> qvga is ok for many embedded devices.
[12:40:29] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
[12:40:32] <koen> keesj, GeneralAntilles: congrats
[12:40:37] <raster> gcohler: hmm ok. so basically different $ for different performance levels - that's all. ie offer both - but pay for the privilege of more mips
[12:40:38] <gcohler> although I do admit most are heading to vga these days.
[12:40:51] <AV500> gcohler: wide VGA please :-)
[12:40:53] <koen> raster: fullscreen qvga is the only thing there is, just look at the sgx stuff :)
[12:40:55] <raster> it really depends what u want to do
[12:41:08] <raster> qvga is doable on even old cpu's - unless they really sucked
[12:41:29] <AV500> we used the omap1 5910 on 320x240 at 150MHz
[12:41:29] <raster> vga is beginning to push it - of course... depending what u want to do. if u keep things ultra-plain/simple... sure.
[12:41:39] <raster> koen: HAHAHAHAHHAAH rotfl!
[12:41:40] * robclark (n=robclark@166.128.145.118) Quit (Read error: 110 (Connection timed out))
[12:41:40] <raster> :)
[12:42:15] <raster> gcohler: mind u.. there are some nice 800x480 (QVGA) screens... this 3.4" WVGA is rather yummy.
[12:42:21] <RobertK> koen: When I dhcp with IMAGE_INSTALL to an image I get add: "Cannot find package dhcp."
[12:42:25] <raster> actually its dpi is stupidly high
[12:42:33] <raster> i actually would opt for a lower res
[12:42:35] <raster> HVGA
[12:42:36] <RobertK> koen: but I can do a bitbake dhcp
[12:42:50] <raster> or maybe 640x320 (one older nokia phone did that)
[12:42:55] <koen> RobertK: IMAGE_INSTALL doens't work as you expect
[12:43:12] <RobertK> koen: okay, how can I add dhcp to an image?
[12:43:13] <gcohler> I think that there is no "right" answer. So offering a breadth of entry points is the key.
[12:43:16] <koen> RobertK: if you want stuff in an image, but them in a task, and include that task in image_install
[12:43:27] <RobertK> ls
[12:43:39] <RobertK> sorry, that should go to my terminal
[12:43:53] <AV500> mru: what is the NEON option in ffmpeg configure?
[12:45:12] <koen> AV500: --cpu=cortex-a8 and it should autodetect from there on
[12:46:50] <kulve> AV500: i.e. you need to have your gcc options set right and ffmpeg will detect them
[12:47:12] <AV500> NEON enabled no :-(
[12:47:17] <raster> do u need to do that?
[12:47:32] <AV500> ARM VFP enabled yes
[12:47:48] <kulve> NEON enabled yes
[12:48:07] <raster> evas builds in neon if on
[12:48:10] <raster> #if defined(__ARM_ARCH_7__)
[12:48:24] <raster> i'm actually unsure what sets that....
[12:48:31] <AV500> kulve: what is your gcc --v
[12:49:37] <kulve> http://pastebin.com/d7df7a810
[12:49:44] <kulve> AV500: but do note that I'm not using OE
[12:51:08] <raster> heretic!
[12:51:20] <AV500> kulve: neither am I
[12:51:28] <AV500> I have stuff like: --with-tune=cortex-a8 --with-cpu=cortex-a8 --with-arch=armv7-a --with-fpu=vfp3 --with-float=softfp
[12:51:31] <AV500> in gcc -v
[12:51:53] <kulve> fpu=neon
[12:52:17] <kulve> http://linux.onarm.com/gitweb/?p=beagleboard/config/beagleboard.git;a=blob;f=board/beagleboard;h=ea87b6b55f192a0b1652a01b051a0a64417722a1;hb=beagleboard-1
[12:52:23] <kulve> that's the gcc options I use compile time
[12:52:33] <kulve> s,that's,those are,
[12:54:56] <AV500> kulve: Ok, I understand now. actually we did not dare enable fpu=neon, but still I want to compile ffmpeg using the NEON inline assembly
[12:55:35] <raster> AV500: neon as fpu might be nice...
[12:55:47] <raster> if gcc produced good neon asm for it...
[12:55:52] * geckosenator (n=sean@71.237.94.78) Quit (Read error: 110 (Connection timed out))
[12:55:57] <raster> and u didnt convert back to ints a lot
[12:56:01] <AV500> yes, but so far we did not trust GCC to emit working neon code
[12:56:20] * RobertK (i=3edc0622@gateway/web/ajax/mibbit.com/x-75b6d7599b780ab3) Quit ("http://www.mibbit.com ajax IRC Client")
[12:57:13] <kulve> AV500: 2007q3 seems to produce proper code with mfpu=neon
[12:57:16] <raster> take a chance!
[12:57:17] <raster> :)
[12:57:26] <kulve> proper meaning working, not efficient
[12:57:29] <AV500> I will, I am the daring kind
[12:58:16] <kulve> AV500: but there are some issues with 2008qX toolchains
[12:59:22] * massiveRobot (n=massiveR@75.149.106.130) Quit ()
[12:59:50] <AV500> kulve: yes, I followed that
[13:00:12] <AV500> still, I should be able to compile ffmpeg without a gcc that has fpu=neon
[13:00:19] <AV500> and using NEON inline asm
[13:00:42] <kulve> AV500: I guess you need to force it on then somehow..
[13:00:47] <AV500> yes
[13:02:39] <khasim> is there any further news on eclipse IDE for ARM / OMAP3
[13:03:00] * khasim feels odd to ask the question that he was supposed to answer
[13:03:31] <AV500> khasim: no idea, ask khasim
[13:03:42] <khasim> :)
[13:03:58] <AV500> will the OMAP3 run eclipse?
[13:07:27] <__alanc__> would be good to get some community fdbk on this : http://wiki.davincidsp.com/index.php?title=Linux_Aware_Debug
[13:11:31] <AV500> you give free CCS and JTAG with that :-)
[13:13:08] * pmeerw_ (n=pmeerw@195-70-114-97.stat.salzburg-online.at) Quit (Remote closed the connection)
[13:13:35] * JuanG (n=Juan@nat/ti/x-5a702c5b1f189154) has joined #beagle
[13:13:41] <AV500> kulve: hmm, it fails because my GCC does not like vadd.i16 q0, q0, q0
[13:13:45] <__alanc__> AV500 - there is a strat for getting to lo-cost tools : see http://wiki.davincidsp.com/index.php?title=XDS100 (it at least helps the Primus dudes)
[13:13:45] * JuanG (n=Juan@nat/ti/x-5a702c5b1f189154) has left #beagle
[13:14:03] * RobertK (i=3edc0622@gateway/web/ajax/mibbit.com/x-88ebb9c1d4ee2644) has joined #beagle
[13:15:03] <AV500> __alanc__: right, forgot about that one
[13:15:24] <RobertK> koen: I did what you suggested: IMAGE_INSTALL += "task-RWK"
[13:15:34] <RobertK> koen: and my task: http://widget.mibbit.com/pb/Seqrbp
[13:15:49] <RobertK> but: * Cannot find package task-RWK.
[13:16:04] <koen> uppercase isn't allowed in package names
[13:18:36] <RobertK> koen: now Cannot find package task-rwk.
[13:18:53] <RobertK> koen: isn't a touch conf/local.conf enough?
[13:19:36] <__alanc__> AV500 - u can always Leave A Comment (link @ bottom) on that wiki page requesting beagle/A8 xds100 support (no promises!)
[13:19:46] * recalcati (i=5e51e963@gateway/web/ajax/mibbit.com/x-ba1f39c1bf739170) has joined #beagle
[13:19:51] <koen> did you add task-rwk to DEPENDS on the image recipe?
[13:20:16] <RobertK> koen: no, to image_install
[13:21:22] <RobertK> just try DEPENDS...
[13:21:47] <RobertK> ...the computer is calculating...
[13:21:53] <RobertK> ...NOTE: package rwk-image-1.0: completed .... :-)
[13:22:38] <AV500> __alanc__: oops, so BB is not yet supported?
[13:23:11] <RobertK> koen: but no /etc/dhcp in the image :-(
[13:23:26] * felipec (i=c0647cda@gateway/web/ajax/mibbit.com/x-bb438b6a02391d44) has joined #beagle
[13:24:25] <RobertK> koen: my task-rwk.bb: http://widget.mibbit.com/pb/QC20vA
[13:24:38] <RobertK> koen: isn't rdepends correct?
[13:25:41] <koen> anything in rdepends will end up in the image
[13:25:54] <koen> RobertK: check the testlab dir in deploy/glibc/beagleboard
[13:26:01] <koen> that should what ended up in the image
[13:26:23] <jkridner> good morning all
[13:26:46] * L84Supper (n=ly@c-68-53-183-162.hsd1.il.comcast.net) Quit ("really gone")
[13:29:06] <__alanc__> AV500 - not yet - no harm in adding Leave A Comment though on the page...
[13:29:42] * L84Supper (n=ly@c-68-53-183-162.hsd1.il.comcast.net) has joined #beagle
[13:29:51] <RobertK> files-in-image.txt: no dhcp
[13:30:47] <koen> RobertK: don't look at files, look at installed packages
[13:30:53] <AV500> kulve: I got it, configure gives me NEON now
[13:31:06] <kulve> great
[13:31:14] <AV500> enabled neon && check_asm neon '"vadd.i16 q0, q0, q0"' -mfpu=neon
[13:31:22] <AV500> had to add -mfpu neon for this check
[13:32:12] <RobertK> koen: my error, I putted task-rwk in DEPENDS, not in IMAGE_INSTALL
[13:32:48] <RobertK> but now: * Cannot find package task-rwk.
[13:34:32] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
[13:35:47] <RobertK> koen: how can I achieve the oe rebuilds its cache?
[13:37:28] <koen> RobertK: why would you need to force that?
[13:38:10] <RobertK> koen: cause oe didnt find task-rwk
[13:38:35] <koen> and why whould you need to force a cache rebuild for that?
[13:39:08] <RobertK> koen: I thought so. I thought in a rebuild oe scans all directories for bb files?!?!
[13:39:54] <RobertK> I put task-rwk in $oe/openembedded/recipes/tasks. Isnt this correct?
[13:40:00] * brijesh (n=bksingh@nat/ti/x-8961238c6aad66a2) has joined #beagle
[13:41:00] <RobertK> koen: why oe didnt find my task-raw in recipes/tasks?
[13:43:15] <koen> it scans for all *.bb files in BBPATH when you start a build
[13:45:30] <koen> FFS
[13:45:36] * koen shoots mozilla people
[13:46:01] <koen> if something doesn't show up in their tinderbox it can't be a bug they say
[13:46:32] * koen adds mozilla to the list of "braindead upstrem list"
[13:46:40] <koen> +a
[13:48:27] <mru> AV500: did you figure out your ffmpeg flags?
[13:53:23] <ddompe> koen: did you recall the name of the linux app that is used to read u-boot env partitions
[13:53:24] <ddompe> ?
[13:53:27] * abitos (n=nixgibts@p5B2E74BC.dip.t-dialin.net) Quit (Read error: 110 (Connection timed out))
[13:54:44] <koen> ddompe: u-boot-utils?
[13:55:15] <ddompe> koen: mmm, nop
[13:55:22] <sakoman__> ddompe: fw_printenv
[13:55:37] <sakoman__> fw_setenv
[13:55:37] <ddompe> sakoman__: thanks
[13:57:51] <RobertK> koen: http://widget.mibbit.com/pb/2GVdDT any idea?
[14:00:14] <koen> hmm and mozilla is using sqlite api that's marked as 'experimental'
[14:00:46] <koen> RobertK: what does 'bitbake task-rwk' say?
[14:01:21] <RobertK> koen: isc "bitbaking" - it compiles joe and dhcp
[14:01:27] <RobertK> isc = its
[14:02:27] <RobertK> the rwk-image.bb: http://widget.mibbit.com/pb/yzgJez task-rwk is in IMAGE_INSTALL
[14:02:32] * fagius (n=fagius@static-71-111-254-119.rlghnc.dsl-w.verizon.net) has joined #beagle
[14:03:17] <koen> RobertK: http://pastebin.com/d490edb44
[14:03:58] <RobertK> koen: I try it...
[14:04:49] <sakoman__> gcohler: BTW, if you want to play around with the Applications menu without editing xml, you could use the xdg tools
[14:05:23] <sakoman__> try: 'xdg-desktop-menu --manual' for details
[14:05:35] <sakoman__> xml might be simpler though :-)
[14:08:24] * mib_vto769 (i=7aa62e54@gateway/web/ajax/mibbit.com/x-5e403f97dc2a7588) has joined #beagle
[14:10:29] <gcohler> sakoman__: thanks. will do.
[14:11:33] <ldesnogu> AV500: just in case this is the line I use for ffmpeg ./configure --enable-cross-compile --cc=/path/to/arm-none-linux-gnueabi-gcc --arch=arm --cpu=cortex-a8 --extra-cflags='-mfpu=neon -mfloat-abi=softfp' --enable-swscale --enable-gpl
[14:12:26] <mru> ldesnogu, AV500: it's recommended to use --cross-prefix=arm-none-linux-gnueabi- (or similar)
[14:12:28] * florian_kc (n=fuchs@port-217-146-132-69.static.qsc.de) has joined #beagle
[14:12:48] <mru> the native binutils often can't handle arm objects
[14:12:54] <mru> ar, strip, etc.
[14:13:05] <ldesnogu> is this some recent change?
[14:13:12] <mru> no, it's been there forever
[14:13:14] <ldesnogu> I never had issue with my configure flags
[14:13:17] * florian_kc is now known as florian
[14:13:19] <mru> you got lucky
[14:13:41] <ldesnogu> I'm usually unlucky so that's strange
[14:13:45] <mru> it depends on how your binutils is configured
[14:13:58] <mru> if it's built with multitarget support, it will work
[14:14:17] <mru> otherwise it will choke saying the format is not supported
[14:19:47] * massiveRobot (n=massiveR@75.149.106.130) has joined #beagle
[14:23:54] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
[14:25:11] * NishanthMenon (n=nmenon@nat/ti/x-17c1b5fc04f2378d) has joined #beagle
[14:25:45] <vlad_> koen: pong
[14:26:24] <mru> ldesnogu: oh, --enable-swscale is gone now
[14:26:57] <ldesnogu> so I'm obsolete :)
[14:29:10] <koen> vlad_: mozilla-central arm build is broken (executive summary)
[14:29:18] <koen> vlad_: and Ted is an asshat
[14:29:26] <vlad_> broken how?
[14:29:28] <vlad_> and asshat how? :)
[14:30:14] <vlad_> http://tinderbox.mozilla.org/showbuilds.cgi?tree=Mobile are all arm builds
[14:30:17] <koen> vlad_: the PR_STATIC_ASSERT you added to xpcom only work with -fno-signed-char, and over here your mozilla configure is adding -fsigned-char hance breaking the test
[14:30:42] <vlad_> hmm
[14:30:57] <AV500> mru: thanks, I think I have it compiled now
[14:31:10] <koen> vlad_: Teds response can be summarized as "tinderbox is green, we have nog bugs"
[14:31:17] <vlad_> heh
[14:31:25] <koen> vlad_: when you tinderbox only builds for maemo by the looks of it
[14:31:33] <vlad_> that is a kind of asshat response
[14:31:42] <vlad_> yeah, and for Windows CE
[14:32:35] <vlad_> well, the good news is that the STATIC_ASSERT is doing what it needed to do
[14:32:46] <vlad_> but let me look at that code again
[14:32:51] <koen> I have no idea why -fsigned-char gets added, my gcc dumpspecs leaks it in
[14:33:11] <mru> it *is* the default pretty much everywhere
[14:33:12] <koen> but I'm not adding -fsigned-char, so *mozilla* is putting it in
[14:33:34] <vlad_> mru: signed-char is?
[14:33:52] <koen> for i in $(find ${S} -name "autoconf.mk") ; do
[14:33:52] <koen> sed -i -e s:fsigned-char:fno-signed-char:g $i
[14:33:53] <koen> done
[14:33:56] * eFfeM (n=frans@195-241-226-180.ip.telfort.nl) has joined #beagle
[14:33:58] <koen> that's how I 'fix' it in OE
[14:34:25] * gregoiregentil (n=zonbu@adsl-71-135-118-139.dsl.pltn13.pacbell.net) has left #beagle
[14:34:30] <koen> vlad_: also, fennec depends on sqlite3 >=3.6.4, but doesn't check for it. I was using 3.6.2...
[14:34:41] <mru> vlad_: char is usually signed by default on most targets
[14:34:42] <koen> (filed as bugs)
[14:34:51] * methril|work is now known as methril|lunch
[14:35:05] <vlad_> koen: I think we got rid of (or were about to get rid of?) support for using the system sqlite
[14:35:09] <vlad_> though maybe I'm misremembering that
[14:35:27] <mru> why the hell do that?
[14:35:37] <koen> vlad_: my checkout is ~6 hours old :)
[14:35:52] <koen> 1.0b2pre
[14:35:53] <vlad_> mru: because we have some patches/changes to ours that aren't in upstream yet
[14:36:19] <mru> 1) are they good? 2) have you submitted them?
[14:36:22] <vlad_> though the system stuff tends to not work until actual release points
[14:36:35] <vlad_> yes and yes, afaik; though it's 7:30am for me, I'm only half awake :)
[14:36:38] * julianoliver (n=julian@164.pool85-53-3.dynamic.orange.es) has joined #beagle
[14:36:46] * lcuk (i=lcuk@cpc3-oldh7-0-0-cust590.manc.cable.ntl.com) Quit ("Leaving")
[14:36:48] <mru> it's such a waste of space when packages include random versions of all their dependencies
[14:37:18] <vlad_> sure, which is why we try to sync up release versions at our own release points
[14:37:22] <eFfeM> yes ! my self crafted serial cable works right away
[14:37:24] <vlad_> but not during alpha/beta/etc. periods
[14:38:13] <vlad_> koen/mru: is there a gcc #define for whether char is signed or not?
[14:38:14] <mru> but you said you were dropping support for using the system sqlite
[14:38:41] <vlad_> mru: yeah, that's the part I'm not sure about; I think there was a reason, but I could be wrong since the option is obviously in there still
[14:38:43] <mru> there's no standard C way to know
[14:38:45] <vlad_> I'll check when I get to the office
[14:38:49] <mru> so you're not allowed to know
[14:38:52] <vlad_> gcc-specific would be fine :)
[14:38:53] <mru> code portably
[14:39:11] * jrmuizel (n=jrmuizel@mozca02.ca.mozilla.com) has joined #beagle
[14:39:13] <mru> use explicit signed/unsigned where it matters
[14:39:23] <vlad_> + case nsXPTType::T_CHAR : *((PRUint32*)d) = s->val.c; break;
[14:39:41] <vlad_> I need to know whether to cast to int32 or uint32
[14:40:19] <vlad_> and the c in the union is just the base char type, or else we'd break the xpcom abi (yes, this is probably a bug that should've been dealt with a long time ago in the abi)
[14:40:39] <mru> portable code...
[14:41:10] <mru> actually casting to signed int should always work
[14:41:17] <mru> if char is signed it will sign-extend
[14:41:23] <mru> if it's unsigned it will still word
[14:41:25] <mru> work
[14:41:39] <mru> signed int can represent every value of unsigned char
[14:41:40] <vlad_> mm
[14:41:43] <vlad_> yeah, you're right
[14:41:46] <mru> the C standard says so
[14:41:48] <AV500> mru: now that I try to link against lavc, I get: ..../libavcodec.so: undefined reference to `q12'
[14:41:50] <vlad_> ok, that's easy, I'll fix that in a bit
[14:41:56] <mru> AV500: broken binutils
[14:42:01] <mru> don't use the codesourcery ones
[14:42:03] <vlad_> brb
[14:42:11] * mib_rhpf2y (i=5076a30c@gateway/web/ajax/mibbit.com/x-184e25c0d3ffee61) has joined #beagle
[14:42:18] <AV500> mru: thx
[14:42:55] <mru> or hunt down the offending vmov instruction and replace with vorr
[14:43:11] <AV500> ok, might be easier
[14:43:13] <AV500> for now
[14:43:15] <mru> it's the same instruction
[14:43:28] <mru> but I'm keeping the code
[14:43:36] <mru> just so I can poke at codesourcery
[14:43:40] <koen> :)
[14:43:46] <AV500> mru: every vmov?
[14:43:50] <mru> just one
[14:44:05] <mru> I forgot which one
[14:44:12] <mru> there's a label in front of it
[14:44:18] <mru> might be what differentiates it
[14:44:22] <mru> towards the end of a file
[14:44:31] <mru> or used to be at least
[14:44:37] <mru> I might have added more code
[14:44:49] <koen> mru: I met another CSL guy, but he denied knowledge of toolchains, since he was in the highperformance computing dept :)
[14:45:05] <AV500> koen: where they dont use toolchains :-)
[14:45:20] <AV500> mru which binutils were ok?
[14:45:50] <mru> plain gnu
[14:46:03] <mru> 2.18 and later are known to work
[14:49:05] <RobertK> koen: I putted a 'inherit task' in task-rwk - and now bitbake finds it. But still no dhcp in the image which is built.
[14:49:17] <ldesnogu> mru: I can compile ffmpeg svn head with CSL 2007q3
[14:49:58] * methril|lunch is now known as methril|work
[14:53:19] <koen> RobertK: aaaah, inherit task adds the ALLOW_EMPTY flag :)
[14:53:30] * mib_wjbubi (i=c05e5e6a@gateway/web/ajax/mibbit.com/x-e022762c5bdf29e1) has joined #beagle
[14:53:33] <RobertK> koen: :-)
[14:53:46] <RobertK> koen: and now what to do?
[14:54:31] <mru> ldesnogu: including binutils?
[14:54:40] <mru> I always use gnu binutils
[14:54:42] <koen> dpkg-deb -I task-rwk.ipk
[14:56:05] <ldesnogu> mru: yes
[14:56:18] <mru> AV500: did you build shared libs?
[14:57:43] <RobertK> koen: there is no task-rwk.ipk, only the dev and the dbg.
[14:57:44] * emeb (n=ericb@ip72-223-90-212.ph.ph.cox.net) has joined #beagle
[14:58:11] * jkridner|work (n=a0321898@nat/ti/x-9313826aa02ebe20) has joined #beagle
[14:58:54] * su_su (i=5168217d@gateway/web/ajax/mibbit.com/x-12091c932b22db7d) has joined #beagle
[14:59:19] <koen> RobertK: try bitbake -c rebuild task-rwk
[14:59:31] <AV500> mru: yes
[14:59:31] <koen> (make sure inherit task is still inthere)
[15:00:08] <mru> ldesnogu, AV500: that could be the difference
[15:00:19] * su_su (i=5168217d@gateway/web/ajax/mibbit.com/x-12091c932b22db7d) Quit (Client Quit)
[15:01:41] <ldesnogu> mru: I'm giving shared build a try...
[15:03:19] * mib_wjbubi (i=c05e5e6a@gateway/web/ajax/mibbit.com/x-e022762c5bdf29e1) Quit ("http://www.mibbit.com ajax IRC Client")
[15:03:41] <RobertK> koen: still the same - no ipk. My task-rwk.bb is okay: http://widget.mibbit.com/pb/nJlsda ?
[15:03:42] * Batko_Marto (n=Batko_Ma@ncart-cam.scs.ryerson.ca) has joined #beagle
[15:04:16] <ldesnogu> mru: problem confirmed
[15:04:20] <koen> that should work
[15:04:26] <AV500> mru: i replaced 4 instances of vmov with vorr
[15:04:39] <AV500> hmm, one q8 left, must be more
[15:05:39] <koen> RobertK: let me try to build that recipe here
[15:06:44] <mru> AV500: the funny thing is, the code is assembled correctly
[15:06:56] <mru> it just adds a spurious symbol reference
[15:08:38] <AV500> yes, thats what I just saw
[15:08:48] <AV500> http://sourceware.org/ml/binutils/2008-03/msg00260.html
[15:09:21] * rbelem (n=rodrigo@pdpc/supporter/active/rbelem) has joined #beagle
[15:09:21] <AV500> fixed by CSL in march 08
[15:10:13] <mru> ah, ok
[15:10:28] <mru> but 2008qX releases are still broken
[15:10:39] <mru> 2008q3 currently compiles ffmpeg correctly actually
[15:10:55] <AV500> so, you "approve" it :-)
[15:10:56] <mru> but that's only because we replaced the code it was failing on for unrelated reasons
[15:11:03] <AV500> ah, ok
[15:11:05] <mru> the bug is still there
[15:11:11] <mru> it's in fsf gcc 4.3.3 too
[15:11:14] <mru> and on ppc as well
[15:11:25] <AV500> in the end there were a lot of vmov to replace, mostly in the h264 stuff
[15:11:49] <mru> I didn't know I had that many
[15:11:52] <mru> I try to avoid it
[15:12:03] <mru> since it doesn't actually do anything
[15:12:19] <mru> iirc most of them should dual-issue with a load/store
[15:12:29] <mru> and there was nothing else to do there anyway
[15:14:14] <RobertK> koen: I have a ipk in all: http://widget.mibbit.com/pb/MzuExO
[15:14:50] * mpoullet|work (n=mpoullet@proxye.avm.de) has left #beagle
[15:15:52] <ZeZu> bringing old topic back up :: that 20cycle hit is ......
[15:16:02] <mru> gone in a9
[15:16:20] <ZeZu> but that is for moving regs isn't it ?
[15:16:37] <jkridner> http://makerfaire.com/bayarea/2009/callformakers/
[15:16:40] <mru> only 3 cycles in a9
[15:17:03] <ZeZu> i think it is what is killing quite a good ammount of speed in my emulator
[15:17:48] <ZeZu> i get 100% emulation much of the time now if there are no floating point ops
[15:18:11] <ZeZu> a game that processes a lot of vertices ... < 10%
[15:18:18] <mru> do you use vfp?
[15:18:35] <koen> RobertK: tthat looks ok
[15:18:39] <ZeZu> neon where i can but vfp for div
[15:18:49] <ZeZu> some of it is still in C
[15:18:51] <mru> neon how?
[15:18:58] <ZeZu> 1s
[15:19:33] <mru> e.g. vmul s0, s1, s2 will run in vfp if rounding/exception modes are not set right
[15:20:49] <ZeZu> well this is in my compiler
[15:21:17] <ZeZu> http://rafb.net/p/logbFP43.html
[15:21:38] * kulve (i=0a5a46f9@emperor.pingu.fi) Quit (Read error: 104 (Connection reset by peer))
[15:21:47] * RobertK (i=3edc0622@gateway/web/ajax/mibbit.com/x-88ebb9c1d4ee2644) Quit ("http://www.mibbit.com ajax IRC Client")
[15:21:59] <ZeZu> its typed, eFPR is just enum of u32
[15:22:10] * scary (n=scary@76.14.22.230) Quit (Operation timed out)
[15:22:10] * zuh (i=66a2292a@emperor.pingu.fi) Quit (Remote closed the connection)
[15:22:17] * mib_qh5jbj (i=c05e5e69@gateway/web/ajax/mibbit.com/x-22452b8e66968c6c) has joined #beagle
[15:22:18] * kulve (n=kulve@emperor.pingu.fi) has joined #beagle
[15:22:18] <ZeZu> dpr can be ignored
[15:22:52] * recalcati (i=5e51e963@gateway/web/ajax/mibbit.com/x-ba1f39c1bf739170) Quit ("http://www.mibbit.com ajax IRC Client")
[15:22:54] * mib_qh5jbj (i=c05e5e69@gateway/web/ajax/mibbit.com/x-22452b8e66968c6c) Quit (Client Quit)
[15:22:57] * recalcati (i=5e51e963@gateway/web/ajax/mibbit.com/x-3a87725d9bc4a9cc) has joined #beagle
[15:23:03] <ZeZu> http://rafb.net/p/6W7hGp54.html
[15:23:29] <ZeZu> this is (very very WIP) on how the compiler implements a few binary fp operations
[15:23:51] <ZeZu> its passed a function pointer to sub/add/mul/div
[15:24:18] <ZeZu> const FPBinOP* opcds[]= { VADD_NFP,VSUB_NFP,VMUL_NFP,VDIV_VFP };
[15:24:20] <ZeZu> ngen_fp_bin(op, opcds[op->op-shop_fadd]);
[15:24:30] * scary (n=scary@76.14.22.230) has joined #beagle
[15:24:35] <ZeZu> from the first pastebin
[15:25:28] <ZeZu> but its mixed in randomly, i have no chance to group (yet / in most cases anyhow)
[15:25:32] <AV500> mru: this 20cylce hit, is it "syncronous", or can it be masked by doing other things on the ARM/NEON in the meantime?
[15:25:58] <mru> you can do other neon stuff in the meantime
[15:25:59] <ZeZu> looks like the arm is at a firm 20cycle hit but neon can still execute
[15:26:08] <mru> if you go via memory you can do other arm stuff too
[15:26:15] <mru> if you stay off that cache line
[15:27:22] * kulve (n=kulve@emperor.pingu.fi) Quit (Remote closed the connection)
[15:29:33] <ZeZu> i think it will be quite some time before the neon stuff is finalized, i'm having worse luck with it in the assembler than I am with my own compiler
[15:29:41] <ZeZu> its a mess trying to use it in the assembler actually
[15:30:29] <mru> how so?
[15:30:53] <mru> I don't recall having any serious trouble with it
[15:31:10] <mru> took a while to get a feel for what's fast and such of course
[15:31:14] <ZeZu> user error i'm sure
[15:32:16] <ZeZu> perhaps not the best handle on some of the instructions or i'm not sure, just got odd results .. do the same thing in three different ways and one of the worked partially , the other two not at all :)
[15:33:17] <ZeZu> the code in the compiler worked the first try though, except for vdiv, which did/does not work
[15:33:29] <ZeZu> I expect that is due to difference of it being vfp
[15:34:15] * nabax (n=nabax@89.129.96.71) has joined #beagle
[15:42:12] * rupeshgujare (n=rupesh@59.160.172.220) has left #beagle
[15:43:51] * like2wise is now known as like|away
[15:44:05] <eFfeM> hmm.apparently today is not my day
[15:44:25] <eFfeM> the serial cable i made does receive but not transmit
[15:44:40] <mru> eFfeM: today is *my* day, and don't you dare try to take it
[15:44:42] <eFfeM> the zd1211 dongle i have used to work under 2.6.27 but does not work on 2.6.28
[15:45:14] <eFfeM> and my pvrusb2 which works on opensuse and directily on usb host does not work on my hub
[15:45:20] <eFfeM> mru lol
[15:51:55] * recalcati (i=5e51e963@gateway/web/ajax/mibbit.com/x-3a87725d9bc4a9cc) Quit ("http://www.mibbit.com ajax IRC Client")
[15:54:04] <AV500> eFfeM: if you have a 2nd cable, that does TX, but not RX, you could combine them
[15:54:20] <AV500> parallel, not in series!
[15:55:30] <eFfeM> AV500: yes, probably first test it, just made it from a flat cable a db9 and an idc10 connector for flat cable
[15:55:52] <eFfeM> they are not that easy to check so tried in the field
[15:56:00] <eFfeM> issue is that it was my last idc10 connector
[15:56:27] <eFfeM> at least I can see the error messages :-)
[15:57:32] <AV500> is there a pastebin for images?
[15:57:42] <eFfeM> www.filebin.ca
[16:00:41] <AV500> http://imagebin.org
[16:01:08] <GeneralAntilles> Thanks, koen! :)
[16:01:15] <eFfeM> is there a prize for the weirdest beagle application ?
[16:01:25] * eFfeM just made a beagle powered vacuum cleaner :-)
[16:03:18] <AV500> koen: http://imagebin.org/42093
[16:04:48] <eFfeM> why on earth would a device refuse to work on a hub but work when directly connected ???
[16:06:46] <AV500> the vacuum cleaner?
[16:07:14] <eFfeM> no, my pvrusb2 usb tv receiver
[16:07:17] <eFfeM> it is self powered
[16:07:48] <eFfeM> and works on the usb host port (c2 hw) but not if there is a hub on that port and I connect to that hub
[16:08:04] <eFfeM> although other devices like my network dongle do work on that hub
[16:08:17] <eFfeM> nothing interesting in dmesg
[16:08:26] <eFfeM> tried 3 different brands usb hubs
[16:12:42] * pcgeil (n=steffen@p5B17F985.dip.t-dialin.net) has joined #beagle
[16:14:30] <eFfeM> does 2.6.29rc6 work for beagle ?
[16:18:59] * scary_ (n=scary@76.14.22.230) has joined #beagle
[16:19:03] <AV500> mru: 1st try with lavc hangs the kernel hard, trying now a versiob without neon
[16:19:12] * jsync (n=jess@59.160.172.220) has left #beagle
[16:19:34] * scary (n=scary@76.14.22.230) Quit (Read error: 131 (Connection reset by peer))
[16:20:39] * RobertK (i=3edc0622@gateway/web/ajax/mibbit.com/x-8d1e31e01c1dd2d4) has joined #beagle
[16:20:47] <RobertK> good mornig again, gentlemen
[16:21:00] <RobertK> After the gret OE rename: which package provides libstdc++.so.6
[16:21:14] <mru> AV500: that's worrying
[16:21:52] <AV500> well, it is the 1st time I try lavc video decoding in my system, so I want to have a non neon baseline reading 1st
[16:22:06] <AV500> could be the neon cache bug, no?
[16:22:15] <mru> which u-boot?
[16:22:38] <AV500> no uboot at all :-)
[16:22:40] <AV500> avboot
[16:23:03] <mru> you need to set the L1NEON bit in the boot loader
[16:23:06] <mru> see u-boot
[16:23:31] <AV500> yes, I know about that, my guys told me it was set
[16:23:56] <mru> which omap rev?
[16:24:33] <AV500> 2.0 or 2.1
[16:25:58] <koen> AV500: :)
[16:26:42] <AV500> cannot show to my boss, 10s to load Mahjongh is UNACCEPTABLE
[16:30:02] <koen> RobertK: the rename has no runtime impact
[16:30:26] <koen> RobertK: nor buildtime impact actually, only impatch is pre parsing
[16:32:16] <AV500> mru: 2.1
[16:32:32] <AV500> Revision : 34301333
[16:34:41] <AV500> mru: it works without NEON
[16:38:48] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-0d33dde421bb6410) has joined #beagle
[16:39:48] * abitos (n=nixgibts@p5B2E74BC.dip.t-dialin.net) has joined #beagle
[16:40:27] * guillaum1 (n=gl@AMontsouris-153-1-91-205.w90-2.abo.wanadoo.fr) Quit ("Leaving.")
[16:40:54] <RobertK> koen: Yesm but now I want to manually install all needed packages.
[16:41:29] <RobertK> koen: oh no: bzcat: Can't open input file uClibc++-0.1.9.tbz2: No such file or directory. - but it is there
[16:41:56] * guillaum1 (n=Guillaum@AMontsouris-153-1-91-205.w90-2.abo.wanadoo.fr) has joined #beagle
[16:42:17] * zedstar (n=john@fsf/member/zedstar) has joined #beagle
[16:42:33] <RobertK> NOTE: package uclibc++-0.1.9-r1: task do_unpack_real: failed
[16:43:17] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-0d33dde421bb6410) Quit (Client Quit)
[16:43:18] * pcgeil (n=steffen@p5B17F985.dip.t-dialin.net) Quit ()
[16:43:40] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-ceccb3396d8a0d2d) has joined #beagle
[16:43:42] <mru> AV500: btw the cache bug on ES2.x is *very* hard to hit
[16:43:58] <AV500> ok, so most likely not this
[16:44:01] <mru> with L1NEON set, I've never seen it
[16:44:15] <AV500> and without this set?
[16:44:34] <mru> it crashes because of the other bug
[16:44:51] <AV500> ok, so could be my boot loader is not OK
[16:45:22] <AV500> btw, MPEG4 QVGA with lavc and a totally generic color conv runs in 20ms/frame
[16:45:27] <AV500> without neon
[16:45:28] * mckoan is now known as mckoan|away
[16:45:51] * RobertK (i=3edc0622@gateway/web/ajax/mibbit.com/x-8d1e31e01c1dd2d4) Quit ("http://www.mibbit.com ajax IRC Client")
[16:46:18] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-ceccb3396d8a0d2d) Quit (Client Quit)
[16:46:26] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-c5cdb55e7e8683cb) has joined #beagle
[16:47:02] <mru> usually it takes a few minutes to lock up
[16:47:08] <mru> or more
[16:49:19] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-c5cdb55e7e8683cb) Quit (Client Quit)
[16:49:20] * rbelem (n=rodrigo@pdpc/supporter/active/rbelem) Quit (No route to host)
[16:49:22] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-f2d4464e5cf1428e) has joined #beagle
[16:49:38] <AV500> mru: i'll now try to do the yuv conversion with neon
[16:51:39] * fsk141 (n=jg@adsl-99-162-208-254.dsl.lsan03.sbcglobal.net) has left #beagle
[16:52:20] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-f2d4464e5cf1428e) Quit (Client Quit)
[16:52:23] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-2f8c7808783a9762) has joined #beagle
[16:55:21] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-2f8c7808783a9762) Quit (Client Quit)
[16:55:31] <mru> AV500: which yuv conversion? 420 to 422?
[16:55:51] <mru> or to rgb?
[16:56:14] <AV500> 420 to 422
[16:56:20] <AV500> it goes to video, not gfx
[16:57:07] * valhalla (n=valhalla@81-174-22-44.dynamic.ngi.it) has joined #beagle
[16:57:38] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-6a2a33f5342da198) has joined #beagle
[16:58:55] <AV500> hmm, same thing, crashes if using neon
[16:59:30] <mru> are you using my code from omapfbplay or writing your own?
[16:59:37] <AV500> yours
[16:59:41] <AV500> yuv.S
[16:59:49] <mru> that's a bsd-like licence, isn't it?
[16:59:53] * TAK2004 (n=thomas@dslb-088-074-044-174.pools.arcor-ip.net) Quit (Remote closed the connection)
[17:00:03] <AV500> yes, it has a most convenient license :-)
[17:00:22] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-6a2a33f5342da198) Quit (Client Quit)
[17:00:33] <mru> and of course you're welcome to profit from it as much as you like ;-)
[17:00:42] <AV500> PROFIT$$$$
[17:01:09] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-b80e07272cc41f31) has joined #beagle
[17:02:25] <AV500> mru: I just added some timing measures, for QVGA without neon, decode takes ~8ms, yuv conv takes 12
[17:02:40] <AV500> but this yiuv conf is plain C, no optim at all
[17:02:50] * colla2 (n=ACollama@host198-216-static.52-88-b.business.telecomitalia.it) has left #beagle
[17:03:03] <AV500> btw, we released MPEG4 and WMV9 720p to the masses this week
[17:03:47] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-b80e07272cc41f31) Quit (Client Quit)
[17:04:15] <mru> we know the neon yuv converer can do a 720p frame in about 10 ms
[17:04:40] <mru> or maybe a bit less even
[17:04:46] <mru> I don't remember the exact numbers
[17:05:51] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-1c55e4d54b4f8526) has joined #beagle
[17:06:29] * PhastPhrog (n=chatzill@194.193.86.112) Quit ("ChatZilla 0.9.84 [Firefox 3.0.7/2009021910]")
[17:08:48] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-1c55e4d54b4f8526) Quit (Client Quit)
[17:08:56] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[17:09:19] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit (Client Quit)
[17:09:29] * magnet_ (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit ("leaving")
[17:09:59] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[17:10:33] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-ce1f19a385a0f0f0) has joined #beagle
[17:12:17] * methril|work (n=Methril@213.27.233.98) Quit ("Saliendo")
[17:12:49] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-ce1f19a385a0f0f0) Quit (Client Quit)
[17:13:48] * magnet_ (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[17:13:48] * magnet_ (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit (Client Quit)
[17:15:23] * magnet_ (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[17:15:26] * magnet_ (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit (Client Quit)
[17:17:09] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-0e5a26721e6b9045) has joined #beagle
[17:19:50] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-0e5a26721e6b9045) Quit (Client Quit)
[17:19:53] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-9fba0531b3629cb0) has joined #beagle
[17:22:51] * robtow (n=rtow@nat/montavista/x-cc34bc99a5de6b92) has joined #beagle
[17:22:51] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-9fba0531b3629cb0) Quit (Client Quit)
[17:24:58] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-6177d32c0d2b5a2b) has joined #beagle
[17:27:52] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-6177d32c0d2b5a2b) Quit (Client Quit)
[17:27:55] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-5ce4f81c0b3b9915) has joined #beagle
[17:28:40] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit ("Lost terminal")
[17:28:50] * davidcb (n=Spot@m0f5e36d0.tmodns.net) Quit ("Leaving.")
[17:28:57] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[17:29:54] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit (Client Quit)
[17:30:11] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[17:30:48] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit (Remote closed the connection)
[17:30:53] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-5ce4f81c0b3b9915) Quit (Client Quit)
[17:30:59] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-86965bf47283698a) has joined #beagle
[17:31:01] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[17:31:43] * magnet_ (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[17:33:38] * alecrim (n=alecrim@189.2.128.130) Quit (Remote closed the connection)
[17:33:54] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-86965bf47283698a) Quit (Client Quit)
[17:34:33] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-4d34785d0236934b) has joined #beagle
[17:35:19] * PhastPhrog (n=chatzill@87.112.0.241.plusnet.ptn-ag1.dyn.plus.net) has joined #beagle
[17:35:54] * mib_rhpf2y (i=5076a30c@gateway/web/ajax/mibbit.com/x-184e25c0d3ffee61) Quit ("http://www.mibbit.com ajax IRC Client")
[17:36:07] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit ("Lost terminal")
[17:36:07] * magnet_ (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit (Client Quit)
[17:36:23] * likewise (n=chatzill@82-171-51-231.ip.telfort.nl) has joined #beagle
[17:36:23] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[17:36:33] <likewise> hi fooks
[17:36:45] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit (Client Quit)
[17:36:55] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-4d34785d0236934b) Quit (Client Quit)
[17:37:12] <AV500> hi
[17:38:00] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[17:38:10] <eFfeM> hi likewise
[17:38:48] * alecrim (n=alecrim@189.2.128.130) has joined #beagle
[17:38:53] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit (Client Quit)
[17:38:55] <likewise> eFfeM: hi, otg cable is in the shop. will pick it up tomorrw and send one to you if you want
[17:39:07] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[17:39:10] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-16a1955d608813dc) has joined #beagle
[17:39:22] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit (Client Quit)
[17:39:52] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[17:40:40] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit (Client Quit)
[17:41:25] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[17:41:37] <ds2> morning
[17:41:56] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-16a1955d608813dc) Quit (Client Quit)
[17:41:59] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-ac448d621238cbd5) has joined #beagle
[17:42:28] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit (Client Quit)
[17:42:50] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[17:43:13] * magnet_ (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[17:44:11] * rbelem (n=rodrigo@189.2.128.130) has joined #beagle
[17:44:31] <eFfeM> likewise, ok
[17:44:57] <AV500> mru: I fixed the boot loader, still crashes in neon yuv code...
[17:44:57] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-ac448d621238cbd5) Quit (Client Quit)
[17:45:22] <koen> AV500: neon is enabled in your kernel?
[17:45:23] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-69711961b7222d4c) has joined #beagle
[17:47:58] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-69711961b7222d4c) Quit (Client Quit)
[17:48:30] <AV500> koen: yes, I just ran a very short neon test file OK
[17:48:37] <AV500> (at least it did not crash)
[17:49:38] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit ("Lost terminal")
[17:49:38] * magnet_ (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit ("Lost terminal")
[17:50:09] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-9c89ffac4b8d7db5) has joined #beagle
[17:51:05] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[17:51:11] * abitos (n=nixgibts@p5B2E74BC.dip.t-dialin.net) Quit (Read error: 110 (Connection timed out))
[17:51:57] * dasnake (n=dasnake@79.171.56.70) has joined #beagle
[17:52:04] <dasnake> hello
[17:52:59] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-9c89ffac4b8d7db5) Quit (Client Quit)
[17:54:05] <dasnake> I'm struggling to get beagleboard+dss2 work on an hdtv
[17:54:15] <dasnake> I've made some posts on ml, but no reply
[17:54:21] <dasnake> my current status is:
[17:54:49] <dasnake> u-boot can drive the dvi output at 1280x720@60 without problems (beagle is displayed correctly as splashscreen)
[17:54:51] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-7ab0bb0288143699) has joined #beagle
[17:55:35] <dasnake> if I compile the kernel to check graphics bootloader init I can see also linux driving dvi output, at least until I try to use it (for example X or omapfbplay or mplayer)
[17:56:05] <dasnake> if I compile the kernel the standard way as soon as linux boot and try to control the output everything goes black and "mode not supported"
[17:56:08] <dasnake> now
[17:56:21] <dasnake> what may be the difference between u-boot handling and linux handling of dvi output?
[17:56:59] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-7ab0bb0288143699) Quit (Client Quit)
[17:58:35] * scary__ (n=scary@76.14.22.230) has joined #beagle
[18:00:00] * scary_ (n=scary@76.14.22.230) Quit (Read error: 104 (Connection reset by peer))
[18:01:47] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-71c9ded29fdc6dc5) has joined #beagle
[18:03:50] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit ("Lost terminal")
[18:04:03] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-71c9ded29fdc6dc5) Quit (Client Quit)
[18:04:06] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[18:04:18] <tharvey> what exactly is MLO vs x-loader? is MLO a specific configuration of x-loader?
[18:04:45] <eFfeM> tharvey: if i recall correctly mlo is a signed version of x-loader
[18:04:51] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit (Client Quit)
[18:05:01] <Crofton|work> signed version of xloader i think
[18:05:01] <adj> just signed and renamed
[18:05:07] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[18:05:25] <dasnake> ...
[18:05:43] <dasnake> it's ok to have a lot of clockdomain/powerdomain messages like:
[18:05:45] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit (Client Quit)
[18:05:49] <dasnake> clockdomain: adding iva2_pwrdm sleepdep/wkdep for pwrdm dss_pwrdm
[18:05:49] <dasnake> powerdomain: will prevent dss_pwrdm from sleeping if iva2_pwrdm is active
[18:05:52] <tharvey> signed because the OMAP's internal bootloader won't boot unsigned code?
[18:06:20] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[18:06:53] * magnet (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit (Client Quit)
[18:07:26] <dasnake> ?
[18:09:05] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-3482c69b04142fad) has joined #beagle
[18:09:17] * magnet_ (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[18:10:30] * magnet_ (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit (Read error: 104 (Connection reset by peer))
[18:10:34] * jrmuizel (n=jrmuizel@mozca02.ca.mozilla.com) Quit ()
[18:11:19] * jrmuizel (n=jrmuizel@mozca02.ca.mozilla.com) has joined #beagle
[18:11:51] * magnet_ (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[18:13:04] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-3482c69b04142fad) Quit (Client Quit)
[18:15:16] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-ea7af15dd48aff3f) has joined #beagle
[18:17:05] * magnet_ (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) Quit ("Lost terminal")
[18:17:14] * Russ (i=foobar@ip70-176-253-20.ph.ph.cox.net) has joined #beagle
[18:17:26] * magnet_ (n=magnet@AMontpellier-258-1-108-123.w90-27.abo.wanadoo.fr) has joined #beagle
[18:18:06] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-ea7af15dd48aff3f) Quit (Client Quit)
[18:18:17] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-9d2f2ad0691006c2) has joined #beagle
[18:19:32] <jkridner> tharvey: the internal bootloader looks for some placeholder information for the signature, even on non-secure devices.
[18:21:07] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-9d2f2ad0691006c2) Quit (Client Quit)
[18:21:18] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-8645775c957b61b3) has joined #beagle
[18:23:09] <mru> AV500: how does it crash?
[18:23:13] <mru> and what kernel are you running?
[18:24:08] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-8645775c957b61b3) Quit (Client Quit)
[18:24:22] <dasnake> omapdss OVERLAY: check_overlay 0: (0,0 1280x614 -> 1280x614) disp (1280x720)
[18:24:31] <dasnake> why 1280x614 ? why 614?
[18:26:17] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-463074df0defb1cc) has joined #beagle
[18:26:59] <ds2> I think I have a conclusion about McSPI's... there might have been some last minute duct tape style changes in the silicon relating to it.
[18:27:50] <ds2> that has to be the only explanation for the convoluted behavior (barring me missing something the TRM)
[18:29:07] <keesj> koen: Thanks! I will try my best to make this all work
[18:29:09] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-463074df0defb1cc) Quit (Client Quit)
[18:29:23] * lcuk (i=lcuk@cpc3-oldh7-0-0-cust590.manc.cable.ntl.com) has joined #beagle
[18:29:42] <ds2> (write ups after ESC)
[18:29:59] * valhalla (n=valhalla@81-174-22-44.dynamic.ngi.it) Quit ("Leaving")
[18:31:18] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-3d45b5518550841b) has joined #beagle
[18:32:00] * alecrim (n=alecrim@189.2.128.130) Quit (Read error: 113 (No route to host))
[18:32:13] <Crofton|work> ds2, what is the problem you are seeing?
[18:33:49] <ds2> Crofton|work: the McSPI apparently uses the Pinmux input drivers to loop back a clock to their receive shift register
[18:34:11] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-3d45b5518550841b) Quit (Client Quit)
[18:34:12] <ds2> but to answer your question - make SPI_CLK a output only line, it does not work
[18:34:34] * eFfeM (n=frans@195-241-226-180.ip.telfort.nl) Quit ("Leaving.")
[18:34:36] <ds2> make it an input capable line, it works fine; MISO was tested using SYST to verify it is functional
[18:34:43] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-6361599d3b630160) has joined #beagle
[18:35:49] <ds2> of course on "normal" SPI implementations, there is only 1 clock and 1 shiftregister so this should not be an issue; it is as if the McSPI was originally designed by someone to have a separate input and output clock but someone (his boss?) insisted on only exposing one clock. So rather then reworking his stuff, he just tied the two signals at the pin
[18:35:56] <ds2> (just speculation of course)
[18:36:04] * florian (n=fuchs@port-217-146-132-69.static.qsc.de) Quit (Remote closed the connection)
[18:36:04] <Crofton> heh
[18:36:15] * alecrim (n=alecrim@189.2.128.130) has joined #beagle
[18:37:12] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-6361599d3b630160) Quit (Client Quit)
[18:38:03] * ceyusa (n=ceyusa@cm216016.red83-165.mundo-r.com) Quit ("Ex-Chat")
[18:38:24] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-5b108d13ae5869fe) has joined #beagle
[18:41:38] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-5b108d13ae5869fe) Quit (Client Quit)
[18:41:40] * PhastPhrog (n=chatzill@87.112.0.241.plusnet.ptn-ag1.dyn.plus.net) Quit (Read error: 104 (Connection reset by peer))
[18:45:51] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-29b14a4854a759ea) has joined #beagle
[18:47:30] * jkridner|work1 (n=a0321898@nat/ti/x-b4e15a953c129018) has joined #beagle
[18:48:40] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-29b14a4854a759ea) Quit (Client Quit)
[18:48:43] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-2875b7403e407874) has joined #beagle
[18:51:42] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-2875b7403e407874) Quit (Client Quit)
[18:52:30] <dasnake> it's x-loader or u-boot that initialize display to output splashscreen?
[18:53:24] * NishanthMenon (n=nmenon@nat/ti/x-17c1b5fc04f2378d) Quit ("Leaving")
[18:53:41] <ZeZu> probably neither if your talking about beagle screen
[18:53:48] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-6cb5c2e6ea05f558) has joined #beagle
[18:56:42] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-6cb5c2e6ea05f558) Quit (Client Quit)
[18:58:13] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-d7962bd3f0afceef) has joined #beagle
[18:59:02] * orifice_work (n=aho@67.71.223.35) Quit (Remote closed the connection)
[18:59:07] * felipec (i=c0647cda@gateway/web/ajax/mibbit.com/x-bb438b6a02391d44) Quit ("http://www.mibbit.com ajax IRC Client")
[19:00:04] * rsalveti (n=salveti@pc-a81040.wlan.inet.fi) has joined #beagle
[19:00:43] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-d7962bd3f0afceef) Quit (Client Quit)
[19:01:03] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-2967b4f26080bb0a) has joined #beagle
[19:01:21] * orifice_work (n=aho@67.71.223.35) has joined #beagle
[19:02:57] * jkridner|work (n=a0321898@nat/ti/x-9313826aa02ebe20) Quit (Remote closed the connection)
[19:03:18] * tharvey (n=tharvey@adsl-76-205-222-173.dsl.snlo01.sbcglobal.net) Quit ("Leaving")
[19:03:24] * alecrim_ (n=alecrim@189.2.128.130) has joined #beagle
[19:03:44] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-2967b4f26080bb0a) Quit (Client Quit)
[19:05:36] * mib_8iews2 (i=4dfc12d9@gateway/web/ajax/mibbit.com/x-12e1e2c2ae75191a) has joined #beagle
[19:09:08] * _don_ (n=ddarling@nat/ti/x-bbdb5cc84b6c23e5) Quit ()
[19:10:08] * orifice_work (n=aho@67.71.223.35) Quit (Remote closed the connection)
[19:10:22] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-e317bf142f27d364) has joined #beagle
[19:11:17] * mib_8iews2 (i=4dfc12d9@gateway/web/ajax/mibbit.com/x-12e1e2c2ae75191a) Quit ("http://www.mibbit.com ajax IRC Client")
[19:13:17] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-e317bf142f27d364) Quit (Client Quit)
[19:14:03] * mib_8iews2 (i=4dfc12d9@gateway/web/ajax/mibbit.com/x-1f01eb1d2ffd147e) has joined #beagle
[19:17:40] * RogerMonk (n=a0740758@nat/ti/x-687f1fca73c4b5b6) Quit (Remote closed the connection)
[19:17:43] * alecrim (n=alecrim@189.2.128.130) Quit (Read error: 110 (Connection timed out))
[19:19:25] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-81cfbdd07cf61ed2) has joined #beagle
[19:20:12] * _don_ (n=ddarling@nat/ti/x-25903084584a1469) has joined #beagle
[19:20:54] * tharvey (n=tharvey@adsl-76-205-222-173.dsl.snlo01.sbcglobal.net) has joined #beagle
[19:21:35] * fulgas is now known as FuL|OUT
[19:22:20] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-81cfbdd07cf61ed2) Quit (Client Quit)
[19:23:00] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-8a5803eebfee77c9) has joined #beagle
[19:25:52] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-8a5803eebfee77c9) Quit (Client Quit)
[19:25:56] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-5f81fd0af639cd00) has joined #beagle
[19:30:42] * alecrim_ (n=alecrim@189.2.128.130) Quit ("Leaving")
[19:31:53] * alecrim (n=alecrim@189.2.128.130) has joined #beagle
[19:33:29] * florian (n=fuchs@f048070029.adsl.alicedsl.de) has joined #beagle
[19:34:24] <florian> re
[19:39:57] <koen> raster: http://openbossa.blip.tv/#1901027
[19:41:53] * orifice_work (n=aho@67.71.223.35) has joined #beagle
[19:52:18] * tequilaworm (n=chatzill@212-198-79-134.rev.numericable.fr) has joined #beagle
[19:53:06] * jkridner|work (n=a0321898@nat/ti/x-c3a8f051972f194c) has joined #beagle
[19:55:29] <jkridner|work> koen: when is your video going to show up?
[19:57:16] * jkridner|work1 (n=a0321898@nat/ti/x-b4e15a953c129018) Quit (Remote closed the connection)
[19:58:03] <tequilaworm> Hi, I am looking for the whole 34XX clock tree drawing. I downoad the TRM 34XX and it is not there any idea ?
[20:00:08] <Russ> tequilaworm: there are many sections of the manual
[20:00:21] <Russ> which section did you download?
[20:00:36] <tequilaworm> yes the PRCM shall be the right one
[20:01:23] * geckosenator (n=sean@71.237.94.78) has joined #beagle
[20:01:47] <Russ> It says in the title, OMAP34xx Power, Reset, and Clock Management Reference Guide?
[20:02:21] * mpoullet (n=mpoullet@77-22-121-59-dynip.superkabel.de) has joined #beagle
[20:02:39] <tequilaworm> the TRM is a doc that includes all the IP's including the PRCM
[20:02:43] * Batko_Marto (n=Batko_Ma@ncart-cam.scs.ryerson.ca) Quit (Read error: 110 (Connection timed out))
[20:03:29] <Russ> then you have the overview document
[20:04:02] <tequilaworm> not this guy has 365 pages
[20:04:22] <tequilaworm> the TRM = Technical Reference Manual
[20:04:41] <Russ> yes, that is the overview document
[20:04:47] <Russ> there are thousands of pages of documentation
[20:05:27] <mpoullet> TRM available here: http://focus.ti.com/lit/ug/spruf98b/spruf98b.pdf
[20:05:37] <tequilaworm> thx I have it
[20:05:52] * lcuk (i=lcuk@cpc3-oldh7-0-0-cust590.manc.cable.ntl.com) Quit ("Leaving")
[20:06:02] <tequilaworm> So if that is the case where can I find a more detailed PRCM doc?
[20:06:43] <ZeZu> hmm, what is the other common disk partitioning util that is all command line / not menu based like fdisk ?
[20:07:18] <Russ> I think you are looking for sprufa5a.pdf
[20:07:21] <tequilaworm> I used to have a nice drawing with the whole clock tree, but I can-t find it any more :(
[20:08:11] <Russ> http://focus.ti.com/docs/prod/folders/print/omap3530.html
[20:08:13] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-5f81fd0af639cd00) Quit (Read error: 104 (Connection reset by peer))
[20:08:13] * mib_8iews2 (i=4dfc12d9@gateway/web/ajax/mibbit.com/x-1f01eb1d2ffd147e) Quit (Read error: 104 (Connection reset by peer))
[20:08:13] * mib_vto769 (i=7aa62e54@gateway/web/ajax/mibbit.com/x-5e403f97dc2a7588) Quit (Connection reset by peer)
[20:08:13] * mib_7hozoy (i=7c0a5030@gateway/web/ajax/mibbit.com/x-27594cdf847a45b9) Quit (Connection reset by peer)
[20:08:17] <Russ> ok, but now I'm confused
[20:08:33] <Russ> what is the relation between omap3430 and omap3530
[20:09:12] <ZeZu> ah sfdisk
[20:09:16] <mpoullet> omap3430 is the version used in mobile phone, like in Nokia's phones
[20:09:36] <tequilaworm> 34XX is the father and 35xx is a variation
[20:09:40] <mpoullet> omap3530 is the catalog version available for all customers
[20:09:48] <tequilaworm> yep
[20:09:50] <Russ> why do parts of the TI pages seem to claim that the 35xx doesn't exist
[20:10:12] <Russ> tequilaworm: on the link I pasted, scroll down and get ch3
[20:10:13] <tequilaworm> that is for customers and 34xx is for internal
[20:10:14] <Russ> er, ch4
[20:10:39] * mpoullet is now known as mpoullet|home
[20:12:17] <tequilaworm> russ: the doc that you are talking about is already included in the TRM, the TRM is a doc that has all the information for a given TI product
[20:12:35] <Russ> how many pages is it?
[20:12:46] <mpoullet|home> Russ: exact the TRM has all the TI docs in it
[20:12:47] <tequilaworm> the trm more that 3000
[20:12:53] * amit (i=816ef12f@gateway/web/ajax/mibbit.com/x-591e9daf6c75e378) has joined #beagle
[20:12:56] <mpoullet|home> 3427 pages
[20:13:09] <Russ> dunno, I think you might has been mislead
[20:13:17] <tequilaworm> sorry 3000 + 427
[20:13:19] <Russ> ch 4 alone is 436 pages, and there are more than 20 chapters
[20:13:21] * amit is now known as Guest94528
[20:13:47] <tequilaworm> russ: if you want to get the all the spec get the TRM
[20:14:06] <Russ> the document I'm telling you to get is a TRM
[20:14:14] * mpoullet|home recommends getting the TRM too
[20:14:36] <tequilaworm> You are right
[20:14:51] * spinl0ck (n=spinl0ck@vodsl-10890.vo.lu) Quit (Read error: 110 (Connection timed out))
[20:14:57] <tequilaworm> this doc shall be called TRM PRCM
[20:15:26] <geckosenator> there is no dhclient or dhcpcd.. what can I use?
[20:15:51] <Russ> udhcpc
[20:16:10] <Russ> my samsung TV uses udhcpc :)
[20:18:25] * mib_7hozoy (i=7c0a5030@gateway/web/ajax/mibbit.com/x-aafba2d54cb6a59c) has joined #beagle
[20:18:39] <Russ> tequilaworm: ok, I was wrong
[20:18:51] <ZeZu> the bootrom doesn't care at all for flash, it just wants MLO in the bootsector right ?
[20:18:54] <ZeZu> err
[20:18:58] <geckosenator> Russ: udhcpc: SIOCGIFINDEX: No such device
[20:18:59] <geckosenator> hmm
[20:19:04] <ZeZu> doesn't care at all for filesystem * (hah)
[20:19:13] <Russ> geckosenator: what interface?
[20:19:22] <geckosenator> wlan0
[20:19:33] <Russ> udhcpc -i wlan0
[20:19:35] <geckosenator> I can see it in ifconfig
[20:19:45] <geckosenator> and I can scan networks and see them with iwlist
[20:20:01] <geckosenator> but I keep getting this: DMA misaligned error with device 0
[20:20:10] <Russ> thats not happy
[20:20:26] <Russ> is this with the z1211?
[20:20:28] <geckosenator> well the wifi works with 2.6.27 as someone mentioned on the email list
[20:20:41] <geckosenator> once I try 2.6.28 or 2.6.29 it can't load the firmware anymore
[20:20:51] <geckosenator> and this person has the same exact problem
[20:20:57] * likewise (n=chatzill@82-171-51-231.ip.telfort.nl) Quit (Read error: 54 (Connection reset by peer))
[20:22:52] <geckosenator> when I unplug the wifi usb dongle: zd1211rw 1-1.3.3:1.0: error ioread32(CR_REG1): -22
[20:24:02] <Russ> that last part is normal
[20:24:08] <geckosenator> heh ok
[20:24:27] <geckosenator> well I"m not sure why I can't use udhcpc
[20:24:42] <geckosenator> do you get the DMA misaligned errors ?
[20:24:47] * lcuk (i=lcuk@cpc3-oldh7-0-0-cust590.manc.cable.ntl.com) has joined #beagle
[20:24:51] <Russ> you could trying switching the MUSB controller to PIO only mode to see if that works around the problem
[20:24:57] <geckosenator> how can I do that?
[20:25:04] <Russ> I don't have a zd1211rw
[20:25:08] <Russ> in the kernel config
[20:25:09] <geckosenator> ok
[20:25:15] <geckosenator> oh, I need to recompile ok
[20:25:27] <geckosenator> I just downloaded the 2.6.27 binary real quick to see if it would fix anything
[20:25:44] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-3f0f77443616e0b8) has joined #beagle
[20:25:46] <geckosenator> tomi's git tree doesn't pull, so I have to use the one without dss2
[20:25:53] <Russ> if it does work, you'd just need to debug what the driver is asking DMA to do that its balking at
[20:25:58] <geckosenator> but I guess I can checkout the 2.6.27 revision maybe
[20:26:21] <geckosenator> well first thing is I want to get wifi working
[20:26:45] <geckosenator> then yes, debug the dma and also figure out what happened to the newer kernels to make them barf sicne there are at least 3 people with this problem now
[20:27:03] <Russ> http://www.bat.org/~tomba/git/linux-omap-dss.git <- isn't working for you?
[20:27:14] <geckosenator> it works.. sort of
[20:27:21] <geckosenator> but it takes hours and hours to sync
[20:27:28] <geckosenator> and I have never successfully completed
[20:27:48] <Russ> its because its dumb
[20:27:53] <Russ> are you pulling from other places into the same tree?
[20:28:04] <geckosenator> no
[20:28:07] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-3f0f77443616e0b8) Quit (Client Quit)
[20:28:10] <geckosenator> it's going about 20 bytes per second
[20:28:10] <Russ> there's your problem
[20:28:10] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-0088002c85bca4b4) has joined #beagle
[20:28:21] <geckosenator> I need to pull from other places into the same tree?
[20:28:28] * tequilaworm (n=chatzill@212-198-79-134.rev.numericable.fr) Quit ("ChatZilla 0.9.84 [Firefox 3.0.7/2009021910]")
[20:28:35] <Russ> you are trying to pull down kernel.org linus+omap patches+dss2 all from poor bat.org via dumb http
[20:28:42] * FuL|OUT is now known as fulgas
[20:28:59] <geckosenator> well do you have a better way?
[20:29:00] * gcohler (n=gcohler@nat/ti/x-810ccef953f07c46) Quit (Remote closed the connection)
[20:29:05] <Russ> you can either add a bunch of remotes to one tree, or add your other trees as alternates
[20:29:10] <Russ> start from a linus 2.6 tree
[20:29:20] <Russ> and then do git remote add <omap tree>
[20:29:27] <geckosenator> this day is getting complicated.. but ok
[20:29:27] <Russ> git remote add <tomba dss2 tree>
[20:29:33] <geckosenator> I'll start with an omap 2.6 tree
[20:29:47] <Russ> then it will pull 99% of the code from kernel.og
[20:29:57] <Russ> another .8% from the omap tree
[20:30:05] <Russ> and the last .2% from bat.org
[20:30:05] <geckosenator> I already have omap tree pulled
[20:30:08] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-0088002c85bca4b4) Quit (Client Quit)
[20:30:23] <geckosenator> does it share disk space like that too?
[20:30:29] <Russ> yes
[20:30:36] <geckosenator> good
[20:30:51] <Russ> you can share between repositories too by cloning with the appropriate options
[20:31:02] <geckosenator> oh, I don't think I have done this yet
[20:31:32] <Russ> just have one primary repo where you have all your remotes added, and do git remote update from there
[20:32:18] <Russ> then clone that repo for individual projects, but with the --shared option
[20:32:36] <Russ> there are a million ways to do it
[20:33:29] <geckosenator> thanks
[20:34:20] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-e11eff73627b5e87) has joined #beagle
[20:34:49] <Russ> once you have remotes setup, 'git branch <remote>/master --track <remote>' is really useful
[20:35:09] <Russ> er, thats backwards
[20:35:22] <Russ> git branch <remote> --track <remote>/master
[20:35:27] * john3909 (n=jsynesio@76-216-254-210.lightspeed.dlmrca.sbcglobal.net) Quit (Read error: 110 (Connection timed out))
[20:35:32] <geckosenator> by the time I get all this figured out, the day will be over :-P
[20:35:58] <geckosenator> but hey.. I got my lcd working
[20:36:15] <geckosenator> although I think I wired some of the datapipns out of order
[20:36:29] <geckosenator> so stuff looks screwy
[20:36:45] <ZeZu> :)
[20:37:07] <geckosenator> I wish I had not ripped the ribbon cable for my 7" lcd though
[20:37:10] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-e11eff73627b5e87) Quit (Client Quit)
[20:37:14] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-6afdd2f555896474) has joined #beagle
[20:37:27] <geckosenator> I think I would have it working :-/
[20:40:11] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-6afdd2f555896474) Quit (Client Quit)
[20:41:51] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-24c9d04d579e104b) has joined #beagle
[20:43:56] <ds2> geckosenator: what data voltage are you running the LCD at?
[20:44:09] <geckosenator> 3.3v
[20:44:22] <geckosenator> do you think it's a problem?
[20:44:31] <ds2> geckosenator: are you sequencing it with the 5V line?
[20:44:46] <ds2> it is out of spec unless you use a precision 3.3V source
[20:44:50] <geckosenator> acutally there might be a problme
[20:45:00] <ds2> which might be a problem?
[20:45:02] <geckosenator> there are two power sources for the lcd
[20:45:11] <geckosenator> 5v and the other is from a 3v ldo
[20:45:16] <geckosenator> then the data lines are actually 3.3v
[20:45:28] <ds2> so you are NOT sequencing it ?
[20:45:31] <geckosenator> no
[20:45:39] <geckosenator> well manually sequencing
[20:45:50] <geckosenator> what are you supposed to do?
[20:45:52] <ds2> good enough for me, I'll turn both on at the same time and cross my fingers
[20:46:03] <geckosenator> well did you see my email?
[20:46:05] <ds2> you are suppose to apply the 2.5V first then apply 5V per the datasheet
[20:46:08] <geckosenator> I actually have problems
[20:46:12] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-24c9d04d579e104b) Quit (Client Quit)
[20:46:19] <geckosenator> I have to pull DE high while in u-boot
[20:46:30] <geckosenator> otherwise the lcd is way too bright
[20:46:37] <ds2> Hmmm
[20:46:51] <geckosenator> I think it's because u-boot supplies clocks too, but with different timings
[20:46:53] <ds2> is DE a pass through signal for you or does that get the TFP410 to do something else?
[20:46:55] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-29a46b8b49c2fee2) has joined #beagle
[20:47:05] <geckosenator> I had it going to the tfp401a
[20:47:10] <geckosenator> which comes out of hdmi
[20:47:13] <geckosenator> but, that never works
[20:47:19] <ds2> Hmmm
[20:47:22] <geckosenator> that de signal is like anding hsync and vsync
[20:47:34] <geckosenator> so instead I have to pull it high
[20:47:39] <ds2> I wonder what difference will be for me with a direct connection
[20:47:52] <geckosenator> I would love to hear
[20:47:57] <ds2> think I am making DE a GPIO and using that drive EN
[20:48:13] <ds2> take that back
[20:48:33] <ds2> EN is driven by a ! of the HDMI enable signal and DE drives the backlight control for me
[20:48:35] <geckosenator> there is a DE line already going to the tfp410 right?
[20:48:46] <geckosenator> oh, I have backlight control on a manual switch
[20:48:48] <geckosenator> whcih works fine
[20:48:56] <geckosenator> it's a completely separate circuit
[20:49:00] <ds2> yes, but I am intercepting it before it gets through... no HDMI for me
[20:49:13] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-29a46b8b49c2fee2) Quit (Client Quit)
[20:49:19] <ds2> well, I want PM eventually since I am already setting it up for Li battery power
[20:49:25] <geckosenator> well I would like to see how the ! of the hdmi enable signal goes
[20:49:38] <geckosenator> I'm already running on battery power :-P
[20:49:50] <ds2> oh then isn't manual backlight bad?
[20:49:55] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-7745ec36ca99ef27) has joined #beagle
[20:50:03] <ds2> I intend to tie it into the blanking routine
[20:50:03] <geckosenator> well maybe not ideal
[20:50:12] <geckosenator> I can always run it to a gpio
[20:50:21] <ds2> you got enough spare GPIOs?
[20:50:22] <geckosenator> but I'm going to switch it off when I need to, manually
[20:50:28] <geckosenator> on the beagle, sure
[20:50:33] <geckosenator> I'm not using the expansion header at all
[20:50:43] <ds2> oh heh... I am almost out of GPIOs
[20:50:47] <geckosenator> I was going to wire up another uart, that's about it
[20:51:22] <geckosenator> so.. do you have your boards yet? what is stopping you from testing it?
[20:51:38] <sakoman__> ds2: glad to hear you resolved the SPI issue. if you could pastebin your final pinmux some time I'll update u-boot
[20:51:49] <ds2> geckosenator: I been bring it up since Saturday!
[20:52:14] * may (i=8970728d@gateway/web/ajax/mibbit.com/x-7745ec36ca99ef27) Quit (Client Quit)
[20:52:15] <ds2> sakoman__: will do... will also do a write up. I just can't believe CLK needs to have IEN
[20:52:43] <ds2> geckosenator: I got 2 GPIO expanders; 1 TSC controller, 1 UART, and a MMC slot to validate before getting to the LCD
[20:53:10] <ds2> sakoman__: should I base it off your tree or is there a "proper" u-boot tree to work against?
[20:53:36] <sakoman__> ds2: most of the u-boot work is being based off my omap3-dev branch
[20:53:46] <sakoman__> we stage things there and then submit upstream
[20:54:07] <sakoman__> it typically takes months to get things upstream, so most folks just build off my branch
[20:54:18] <ds2> sakoman__: okay, got git URL for me to clone? I am working off an omap3-dev that was pointed to by the RevC validation page but been told that is out of date
[20:54:45] <ds2> cuz I also would like to have second MMC slot support in U-boot
[20:54:52] <geckosenator> I'm trying to figure out how to clone from kernel.org, then just clone the differences from another url..
[20:54:55] <sakoman__> ds2: I've found it is usually best to always enable the pin input, unless there is a real downside to doing so
[20:55:14] <ds2> geckosenator: look up remote branches
[20:55:22] <geckosenator> ds2: yeah doing it
[20:55:26] <sakoman__> ds2: clone url's are here: http://www.sakoman.net/cgi-bin/gitweb.cgi?p=u-boot-omap3.git;a=summary
[20:55:29] <ds2> sakoman: but doesn't that burn more power leaving the input receiver on?
[20:56:00] <sakoman__> a tiny amount, yes
[20:56:14] <ds2> so with 300-400 of them, it would add up
[20:56:42] <ds2> what irks me is no where does it mention that the receive clock is internally looped by something outside of the McSPI but inside the chip
[20:57:09] <sakoman__> I'm talking about the one'sy two'sy spi, gpio kinds of things
[20:57:59] <ds2> they still add up; also i don't think they get shut off in the PM modes either
[20:58:07] <geckosenator> sakoman__: does that url have dss2?
[20:58:20] <sakoman__> It's the kind of thing that you could fine tune after you get things working if you really care about power
[20:58:31] <geckosenator> sakoman__: we really care :-P
[20:59:11] <sakoman__> geckosenator: I usually care about function first, then power performance tuning later
[20:59:26] <sakoman__> geckosenator: dss2 for u-boot?
[20:59:29] <ds2> that is a fair argument
[20:59:39] <geckosenator> sakoman__: heh, sorry I realize it's u-boot
[20:59:47] <sakoman__> do you mean revised u-boot env for dss2?
[21:00:15] <sakoman__> geckosenator: I don't have a public kernel git for current dss2
[21:00:23] <geckosenator> yeah I realize that now
[21:00:24] <sakoman__> tomba is the official source
[21:00:39] <geckosenator> but when I clone from him it goes 20 bytes a second and never finishes
[21:00:42] <sakoman__> I just pull from him and generate an OE patch
[21:00:54] <geckosenator> so I'm going to have to figure out git-remote
[21:01:15] <geckosenator> also.. something in the wifi stuff broke from 2.6.27 to 2.6.28
[21:01:16] <ds2> git remote add LocalNameToUse GitRepoURL
[21:01:22] <ds2> git remote update
[21:01:28] <geckosenator> since my wifi card cannot load firmware with 2.6.28 or later, but 2.6.27 works
[21:01:31] <ds2> and you can see the remote branches with "git branch -a"
[21:01:35] <sakoman__> geckosenator: tomba talked about doing a gitorious mirror, don't know that he ever got around to doing it
[21:01:37] <dasnake> if neither x-loader nor u-boot initialize hdmi output when the beagle is displayed at boot, who is initializing video at that point?
[21:01:51] <dasnake> because that's exactly 1280x720@60 that I would like to use
[21:02:09] <geckosenator> ds2: I did that but I don't see a dss branch
[21:02:10] <sakoman__> my first clone from him took forever, but normal pulls aren't too bad
[21:02:39] <geckosenator> well I let it clone for 8 hours and it did not complete
[21:03:02] <sakoman__> dasnake: beagle's custom u-boot intializes dvi for the splash screen
[21:03:06] <geckosenator> ds2: oh I missed git remote update
[21:03:46] <dasnake> sakoman, so whatever magic u-boot is doing (and dss2 fails to do) is in the u-boot sources, right?
[21:03:58] <ds2> geckosenator: remote update is what downloads it
[21:04:21] <geckosenator> ds2: ok, makes sense because it's going really slow again :-P
[21:04:24] <ds2> I wonder what surprises are waiting on I2C2 ;)
[21:04:31] <sakoman__> dasnake: I'm not sure what you mean
[21:04:38] <geckosenator> I have a realtime clock for i2c I would like to hook up actually...
[21:04:42] <ds2> since I2C doesn't quote loop
[21:04:52] <dasnake> sakoman, I've troubles with beagle display on hdtv
[21:05:06] <ds2> geckosenator: you have no network capability in your setup?
[21:05:06] <dasnake> sakoman, u-boot displays the beagle correctly at 1280x720
[21:05:20] <geckosenator> ds2: unless I use usb in device mode
[21:05:31] <ds2> I see
[21:05:33] <dasnake> sakoman, linux pulls a "mode not supported" whatever I try (dss1,dss2,tons of different options, ...)
[21:05:35] <geckosenator> ds2: but I did get to scan wifi networks and found them with 2.6.27 kernel
[21:05:48] <geckosenator> ds2: but then I got dma errors and I could not run udhcpc
[21:05:48] <ds2> geckosenator: why not drop in a BT unit?
[21:05:56] <geckosenator> I only have 1 bt unit
[21:06:09] <geckosenator> I have bt devices.. but I need 2 bt usb to do networking
[21:06:12] <dasnake> geckosenator: for the dss2 patches I do the following way: pull tree from git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap-2.6.git then add as remote tomba and the git remote update .. it's fairly fast
[21:06:20] <ds2> uart units aren't expensive and according to sakoman ttyS1 can do the near 1M mode
[21:06:30] <geckosenator> dasnake: thanks, I'm trying that now
[21:06:40] <geckosenator> I had a uart unit
[21:06:50] <geckosenator> but I managed to "brick" it by setting the wrong parameters
[21:06:54] <geckosenator> and it can only recover via spi
[21:07:02] <sakoman__> dasnake: dss2 is still under development and the option syntax changes a lot. you need to make sure your boot args match the version of dss2 that was built into your system
[21:07:04] <ds2> use a McSPI then ;)
[21:07:36] <geckosenator> ds2: maybe... but I'm in your situation.. too much stuff to do
[21:07:36] <dasnake> sakoman, at boot shows no errors and the debug messages doesn't tell anything revealing of the problem, at least to me
[21:07:57] <dasnake> sakoman, I tried to ask in ml several times, with logs, but I couldn't catch attention
[21:08:20] <dasnake> the only thing I figured out is that u-boot is able to initialize dvi output in a reasonable way
[21:08:23] <dasnake> and linux don't
[21:09:11] <geckosenator> ds2: I want to get wifi working more than bt networking anyway
[21:09:17] <dasnake> if I compile framebuffer to check if bootloader initialized graphics I'm able to see linux output the angstrom splashscreen, but as soon as I try to use fb I got "mode not supported" again
[21:09:31] <dasnake> I guess I'm doing something foundamentally wrong, but I can't find what
[21:09:41] <sakoman__> dasnake: sorry, I am not up on what all the various kernels out there are using as far as dss2 rev, so I can't really offer any specific advice
[21:09:42] <geckosenator> dasnake: can you try different modes?
[21:10:40] <sakoman__> dasnake: if you are building the kernel yourself, you can just look in the Documentation/omap/dss text file for what your specific version is expecting
[21:10:49] <dasnake> geckosenator: I tried everything from 320x200 to 1280x720 with 2.6.27 (android), 2.6.28 (angstrom oe), 2.6.29 (mru), 2.6.29 (tomba, <18 march and >18 march revisions)
[21:11:09] <sakoman__> if you are using someone else's pre-built kernel, then only they can help you
[21:11:42] <ds2> just trace through the source; takes a bit of time but works
[21:12:08] <dasnake> sakoman, I guess I'm passing the right thing, omapdss debug output is coherent to what I ask, also fbset is coherent, it just doesn't work on my hdtv
[21:12:09] <mru> I don't provide prebuild kernels
[21:12:19] <dasnake> but it's not a cable/htdv problem, because u-boot works
[21:12:24] <mru> I guess he took something from my git
[21:13:04] <geckosenator> I wonder if I can use git-bisect to figure out why 2.6.28 broke the firmware upload for my wifi
[21:13:51] <sakoman__> geckosenator: only if you are a very patient man :-)
[21:14:11] <geckosenator> well it's doing a binary search...
[21:14:29] <geckosenator> I need to tell it only to deal wtih commits in the usb folder
[21:15:09] <geckosenator> and it has that ability
[21:16:15] <sakoman__> I've always found it a frustrating experience, especially when you run into commits that result in a broken state
[21:16:57] <geckosenator> yeah.. I was afraid of that
[21:17:17] <dasnake> anyhow, how does u-boot figure out the perfect display size for my hdtv?
[21:17:23] <sakoman__> dasnake: have you examined drivers/video/modedb.c to see how to concoct video mode strings?
[21:17:28] <dasnake> ask for resolutions via dvi?
[21:17:35] <sakoman__> dasnake: it doesn't figure it out, it is hard coded
[21:17:37] <geckosenator> dvi can't ask for resolutions
[21:17:42] <dasnake> uhmm
[21:17:44] <mru> u-boot has some magic hardcoded
[21:17:52] * mib_wa8h8l (i=c05e5e69@gateway/web/ajax/mibbit.com/x-bb2a20498a5b2aa0) has joined #beagle
[21:17:55] <mru> geckosenator: dvi can ask for resolutions, but u-boot can't
[21:18:12] <geckosenator> mru: oh, is that via i2c?
[21:18:16] <mru> yes
[21:18:19] <geckosenator> oh ok
[21:18:22] <ds2> geckosenator: firmware loading is handled outside of the usb directory
[21:18:26] <mru> same as vga
[21:18:35] * mib_wa8h8l (i=c05e5e69@gateway/web/ajax/mibbit.com/x-bb2a20498a5b2aa0) Quit (Client Quit)
[21:18:47] <geckosenator> ds2: thanks, that should save me some time.. do you know where it is handled?
[21:18:47] <dasnake> sakoman, yes the latest kernel I used (2.6.29 with >18 march tomba) parse resolutions in the form of 1280x720MR-24@60
[21:19:03] <dasnake> sakoman, I checked in the sources, and works, I have lines like:
[21:19:07] <sakoman__> geckosenator: ds2 is correct, I believe it is udev loading firmware
[21:19:15] <dasnake> fbcvt: 800x600@60: CVT Name - .480M3-R
[21:19:27] <dasnake> that tells me the kernel understood what I tried
[21:19:42] <geckosenator> sakoman__: isn't udev a package?
[21:20:02] <ds2> the loader code is in driver/firmware
[21:20:17] <ds2> yes udev does the sysfs acrobatics to answer the call for firmware
[21:20:19] <dasnake> mru, let me understood: u-boot doesn't ask for resolution via i2c, but magically figure out that hdtv would like 1280x720@60 ?
[21:20:20] <sakoman__> dasnake: the "R" means reduced sync, which many TV's do not like
[21:20:31] <dasnake> ook
[21:20:49] <geckosenator> ds2: well I only changed the kernel to fix it, so I'm hoping the problem is in drivers/firmware then
[21:20:51] <ds2> dasnake: I believe the U-boot mode is specially choosen to work on many devices
[21:20:55] <dasnake> but I read that beagle isn't able to do things different from reduced resolutions
[21:20:55] <Russ> I had a lot of trouble getting 1280x720M-24@60 working
[21:21:02] <sakoman__> dasnake: u-boot doesn't "figure out" anything. It is a hard coded value!
[21:21:06] <mru> dasnake: no, u-boot has hardcoded magic settings for 720p
[21:21:33] <dasnake> mru, really magic :)
[21:21:42] <Russ> hsw: 0x40, hfp: 0x0c0, hbp: 0x040, vsw: 0x05, vfp: 0x003, vbp: 0x014 worked for me
[21:21:54] <mru> you can get the same modes with a properly configured fb driver
[21:22:09] <dasnake> ok
[21:22:42] <geckosenator> ds2: do you adjust the backlight brightness?
[21:22:43] <Russ> I think that the HDTV video modes need to get added to modedb.c for stuff to work properly
[21:22:59] <dasnake> russ, I'll try to add your magic to modedb.c
[21:23:15] <ds2> geckosenator: resistor on the backlight or PWM your backlight boost controller
[21:23:21] <Russ> I got mine in a bizzare way, and its not even close to what HDTV's really expect
[21:23:26] <Russ> you might try a real 720p mode first
[21:23:59] <geckosenator> ds2: I was going to put a potentiometer in series with the resistor
[21:24:08] <geckosenator> ds2: but maybe pwm is better
[21:24:16] <Russ> http://www.3dexpress.de/displayconfigx/timings.html
[21:24:27] <ds2> geckosenator: what booster are you using?
[21:24:35] <dasnake> russ, what hdtv panel is yours?
[21:24:58] <geckosenator> ds2: TB62752FUG
[21:25:05] <Russ> I have a samsung LN46A750
[21:25:16] * john3909 (n=jsynesio@76-216-254-210.lightspeed.dlmrca.sbcglobal.net) has joined #beagle
[21:25:25] <ds2> geckosenator: no usable matches on that P/N
[21:25:30] <Russ> http://www.samsung.com/us/consumer/detail/detail.do?group=televisions&type=televisions&subtype=lcdtv&model_cd=LN46A750R1FXZA
[21:26:04] <dasnake> uhmm, I've a samsung too, but a le32r86
[21:26:13] <geckosenator> ds2: http://doc.chipfind.ru/marktech/tb62752afug.htm
[21:26:56] * scary__ (n=scary@76.14.22.230) Quit (Remote closed the connection)
[21:27:05] <mru> my LE40A646 seems quite happy with the beagle
[21:27:28] <geckosenator> ds2: it says you can use pwm to the shutdown
[21:27:30] <ds2> geckosenator: read the section "Setting the Io"
[21:27:37] <ds2> that would be the best palce to put the pot
[21:27:39] <dasnake> I'm sure that also my beagle can be friend of my samsung ..
[21:27:51] <ds2> do not exceed the backlight current spec
[21:28:07] <Russ> in the future, the beagle will supposively talk to the display and obtain modeline info and things will be easier
[21:28:12] <geckosenator> ds2: yeah, that's why I would put the pot in series with a resistor
[21:28:49] <geckosenator> I got a 20 ohm potentiometer.. it should work
[21:28:49] * Russ wonders if given the current exchange rate, a trip to Korea to pick up a LCD TV would be advantageous
[21:29:29] * fagius (n=fagius@static-71-111-254-119.rlghnc.dsl-w.verizon.net) Quit ("Leaving")
[21:29:55] <ds2> that'll get you a manual control
[21:30:04] <geckosenator> yeah, that's probably ok too
[21:30:10] * RogerMonk (n=a0740758@nat/ti/x-44e8d99c013bf27d) has joined #beagle
[21:30:16] <ds2> GPIO to the SHDN line is what I did
[21:30:19] <geckosenator> I would rather have it working than have the beagle not outputting pwm yet
[21:30:24] <mru> Russ: shipping it back might be tricky/expensive
[21:30:29] <ds2> heheh
[21:30:36] <geckosenator> ds2: so you have to set up software before you see anything
[21:30:44] <Russ> mru: what are you talking about, my 55 inch display is a carry on item
[21:30:49] <geckosenator> can uboot display the console on the lcd?
[21:31:01] <mru> otherwise a few items bought over there can easily make up for the plane ticket in saved money
[21:31:02] <ds2> yes, you need to setup software first
[21:31:12] <ds2> u-boot doesn't do console but you can use fbconsole in Linux
[21:31:18] * mru saved half a plane ticket on a single camera lens
[21:31:40] <dasnake> uhmm, any idea about the fact that if I set mode 1280x720MR-24@60 fb is set at 1280x614 ?
[21:31:45] <geckosenator> ds2: oh framebuffer console is cool.. but u-boot console might be nice too in case you screw stuff up and don't have another machine for serial console
[21:32:12] <Russ> geckosenator: but very complex...a kexec loader would be a much better solution
[21:32:17] <geckosenator> does anyone know what the branches in dss2 mean "pub1" "pub2" etc.. which should I use?
[21:32:39] <geckosenator> Russ: kexec loader? or do you mean you modify the file uboot uses in linux, then just reboot and pray?
[21:32:53] <Russ> http://www.flickr.com/photos/koenkooi/3300102390/ <- kexecboot is friggin awesome
[21:33:14] <Russ> the kernel has a way to start another kernel
[21:33:29] <geckosenator> oh I see
[21:33:43] <Russ> so you load a minimal kernel with a minimal fs with the bootloader, and then use that environment to load the real kernel
[21:33:45] <geckosenator> so you boot one kernel with uboot which gives a way to select a different one
[21:34:02] <geckosenator> cool
[21:34:05] <Russ> right, with full support for devices supported by the kernel
[21:34:13] <Russ> is that a picture of koen with an owl, or some other guy
[21:37:16] <ds2> geckosenator: U-boot USB tty support
[21:37:32] <geckosenator> ds2: that only works if you have another machine right?
[21:37:42] <ds2> yeah
[21:37:48] * massiveRobot (n=massiveR@75.149.106.130) Quit ()
[21:37:58] <ds2> solution of course is to never screw things up that bad if there is no second machine :P
[21:38:04] <mru> dev boards usually assume you have another machine
[21:38:14] <mru> get two beagles
[21:38:20] <mru> hack on one at a time
[21:38:29] <Russ> heh
[21:38:57] <Russ> just have a rescue SD card
[21:39:06] <Russ> then you can boot it with the user button
[21:40:01] <geckosenator> mru: yes I think I will get a second beagle once revc is out
[21:40:10] <geckosenator> Russ: yeah another good idea
[21:40:41] <Russ> oh, and buy a label maker for all those SD cards
[21:40:50] <geckosenator> I have a pen
[21:40:54] <Russ> my label maker can actually make an A with the circle on top
[21:41:06] <tomba> geckosenator: master
[21:41:15] <geckosenator> tomba: ok, that's what I'm doing thanks
[21:41:45] <Russ> geckosenator: are you some kind of amish developer
[21:42:19] <geckosenator> uh.. I'm not amish, but why do you ask?
[21:42:34] <Russ> using pens to label SD cards...
[21:42:54] <geckosenator> a pen has many more purposes than some special label making machine
[21:43:03] <Russ> bah
[21:43:04] <geckosenator> and it's lighter weight, cheaper, easier to replace etc..
[21:43:21] <mru> I'm with geckosenator on this one
[21:43:23] <Russ> I suppose next you'll tell me it doesn't need batteries
[21:43:45] <mru> actually, I have an even simpler method: I remember what's on the cards
[21:43:53] <mru> if I can't remember, it's not important
[21:43:59] <geckosenator> right
[21:44:05] <geckosenator> for recovery purposes, you could just try them all
[21:44:30] <mru> and for other purposes you can check what's there before writing
[21:44:45] <geckosenator> they have digital labels too
[21:45:33] <mru> I sometimes use filesystem labels to mount a specific card in the same location regardless of what random scsi name it gets
[21:47:27] <geckosenator> like /media/mystuff instead of /media/disk-11
[21:47:36] <geckosenator> the problem is sometimes I end up with /media/mystuff-3
[21:49:20] * brijesh (n=bksingh@nat/ti/x-8961238c6aad66a2) Quit ()
[21:51:15] * sakoman (n=sakoman@static-74-41-60-154.dsl1.pco.ca.frontiernet.net) has joined #beagle
[21:53:29] * sakoman__ (n=sakoman@static-74-41-60-154.dsl1.pco.ca.frontiernet.net) Quit (Read error: 60 (Operation timed out))
[21:54:20] * lcuk (i=lcuk@cpc3-oldh7-0-0-cust590.manc.cable.ntl.com) Quit ("Leaving")
[21:57:12] * cbrake is now known as cbrake_away
[22:02:50] * uberfry (n=spinl0ck@vodsl-10890.vo.lu) has joined #beagle
[22:06:12] <AudioFreak> mru: but what if you forget which one is the rescue card?
[22:06:53] <mru> it's the 64MB one
[22:06:57] <mru> useless for anything else
[22:06:59] <AudioFreak> lol :P
[22:12:04] <Russ> or like geckosenator, just keep trying till one boots
[22:13:23] * ddompe (n=ddompe@200.122.155.113) Quit ("ddompe is leaving now")
[22:14:33] * lcuk (i=lcuk@cpc3-oldh7-0-0-cust590.manc.cable.ntl.com) has joined #beagle
[22:14:58] * jrmuizel (n=jrmuizel@mozca02.ca.mozilla.com) Quit ()
[22:16:01] <AudioFreak> I think I'll just do my best to keep my laptop alive, then i'll probably be able to fix the beagleboard ;)
[22:16:28] * zuh (i=344cbe8a@emperor.pingu.fi) has joined #beagle
[22:16:40] * kulve (i=6226ddab@emperor.pingu.fi) has joined #beagle
[22:17:06] * adj_ (i=antti@hervanta.com) has joined #beagle
[22:28:05] * rsalveti (n=salveti@pc-a81040.wlan.inet.fi) Quit (Read error: 110 (Connection timed out))
[22:28:17] * adj (i=antti@hervanta.com) Quit (Read error: 110 (Connection timed out))
[22:28:32] * rsalveti (n=salveti@pc-a81040.wlan.inet.fi) has joined #beagle
[22:36:37] <geckosenator> AudioFreak: laptops are heavy though
[22:36:47] <geckosenator> I think having 2 beagle boards would be the obvious
[22:37:28] <ds2> laptops are such ancient technology
[22:37:33] <mru> touch book !
[22:37:49] <mru> it's a beagle *and* a laptop
[22:38:15] * mru wants
[22:42:56] <geckosenator> what do you want?
[22:46:09] * Russ (i=foobar@ip70-176-253-20.ph.ph.cox.net) Quit (Remote closed the connection)
[22:46:30] * ZeZu also
[22:46:57] <ZeZu> cortex-a9 w/ a ton of cores ~1ghz would be a lot nicer though
[22:47:34] * russ (i=foobar@ip70-176-253-20.ph.ph.cox.net) has joined #beagle
[22:50:24] <mru> it only supports up to 4 cores iirc
[22:53:03] <AudioFreak> geckosenator: as soon as rev c is out and my bankaccount permits me to, i'll buy a second beagle :P
[22:53:44] <AudioFreak> anyone ever tried mpd on a beagleboard?
[22:53:57] * russ (i=foobar@ip70-176-253-20.ph.ph.cox.net) Quit ("Ex-Chat")
[22:54:17] * russ (i=foobar@ip70-176-253-20.ph.ph.cox.net) has joined #beagle
[22:54:25] * russ is now known as Russ
[22:54:47] <AudioFreak> i love it on my laptop, but it would be better to have a beagle at home to organise my music and stream it to icecast and local soundset :P
[22:56:53] <geckosenator> all over wireless?
[22:58:23] <AudioFreak> hmm, well mpd needs the source files local, and when i listen to it remotely i just need a 128kbit stream
[22:58:35] <AudioFreak> that should be doable over wireless :P
[22:58:37] <geckosenator> yeah
[22:58:50] <geckosenator> or bluetooth even
[22:58:55] <geckosenator> since bluetooth can carry sound
[22:59:02] <geckosenator> I don't know if icecast supports that
[22:59:08] <AudioFreak> hmm, don't know
[22:59:15] <AudioFreak> would be cool though
[22:59:18] <geckosenator> yeah
[22:59:27] <geckosenator> then you would just need bluetooth headphones, not a laptop
[22:59:41] <AudioFreak> buy a stereo headset and just listen everywhere in you house :D
[22:59:49] <geckosenator> yeah
[23:00:07] <AudioFreak> gonna write that down before i forget :P
[23:01:04] <AudioFreak> oh and mpd can probably output "directly" to the bluetooth device
[23:01:16] <AudioFreak> so icecast isn't used then
[23:01:38] <AudioFreak> mpd can have as many outputs at the same time as your hardware can support
[23:01:45] <geckosenator> I'm thinking about using two lithium cells in series
[23:02:16] <geckosenator> then buck that to get 5v
[23:02:41] <AudioFreak> so i'd just configure a local output to my soundset, a bluetooth output to my headset and a icecast for remote music :)
[23:02:54] <geckosenator> you could do that
[23:03:02] <geckosenator> but then multiple people won't be able to tune in
[23:03:14] <AudioFreak> why not?
[23:03:48] <AudioFreak> only problem would be the bandwidth of my wireless/internetconnection as far as i can see
[23:04:12] <geckosenator> I mean over bluetooth
[23:04:30] <Russ> bluetooth audio doesn't do well through walls
[23:04:30] <AudioFreak> i guess it would be interesting to do some research on utilizing the dsp for decoding and encoding the audio
[23:04:38] <Russ> its not very accepting of loss
[23:04:57] <AudioFreak> guess the walls at my place won't be too much of a problem
[23:05:13] <AudioFreak> my dsp comment was about patching/rewriting mpd to so
[23:05:22] <Russ> and I think most headphones are probably class 2
[23:05:30] * zedstar (n=john@fsf/member/zedstar) Quit (Remote closed the connection)
[23:06:26] * bmxr (n=bmxr@S01060018f3b11a22.vf.shawcable.net) has joined #beagle
[23:06:37] * RogerMonk (n=a0740758@nat/ti/x-44e8d99c013bf27d) Quit (Remote closed the connection)
[23:07:11] <AudioFreak> hmm, is it posible to connect solid state disks to the beagle? i guess the outputs are not too much of a problem
[23:07:23] * _don_ (n=ddarling@nat/ti/x-25903084584a1469) Quit ()
[23:07:32] <AudioFreak> but old fashioned magnetic disks are just too big :P
[23:07:59] <Russ> AudioFreak: just put an SSD in a USB enclosure
[23:08:03] <AudioFreak> i'm thinking of about 2 terabytes of FLAC music :P
[23:08:13] <Russ> but really, SD cards are so large now
[23:08:19] <Russ> but perhaps I am underestimating your collection
[23:08:22] <Russ> yes, I am
[23:08:26] <AudioFreak> :P
[23:08:38] <AudioFreak> well it's not exactly 2 terabytes yet
[23:08:53] <Russ> I don't think 2TB of SSD would come cheap
[23:08:59] <AudioFreak> nope :P
[23:09:14] <Russ> Buy a NAS with 3 or 4 bays
[23:10:40] * _don_ (n=ddarling@nat/ti/x-f531b3bb0cec4be1) has joined #beagle
[23:10:48] <Russ> holy crap, I have an important announcement
[23:10:55] * __alanc__ (n=a-campbe@nat/ti/x-fdfd7cb347c12e0f) Quit ()
[23:11:00] <Russ> I just got my OFFICIAL NOTIFICATION from the MICROSOFT LOTTERY
[23:11:08] * TAK2004 (n=thomas@dslb-088-072-216-168.pools.arcor-ip.net) has joined #beagle
[23:12:33] <Russ> aparently, I can come "...Down to Uk Personally to Pick [My] Price"
[23:13:06] <mru> so you didn't win a day without crashes?
[23:13:29] <Russ> no, but I think they are sending me a CE 4.0 SDK
[23:13:41] <mru> I pity you
[23:16:11] <AudioFreak> Russ: 2 TB of SSD would be at least 4000 euro :P
[23:16:26] <AudioFreak> but it makes less noise than a 4 bay NAS :P
[23:16:49] <Russ> right, but the NAS you can stash anywhere
[23:18:08] <AudioFreak> guess i'll drop the wifi and go for a cabled network
[23:18:19] <Russ> for 4000 euro, you can rent rack space for your NAS
[23:18:32] <Russ> just have a bridge somewhere
[23:18:57] <AudioFreak> lol, i guess i'd talk to my boss then
[23:19:27] <AudioFreak> work for a local ict company hosting terminal server, exchange etc.
[23:20:44] <Russ> I'm waiting for some good 2TB drives so I can get a small 4TB array for myth
[23:20:58] <AudioFreak> :)
[23:21:00] * mpoullet|home (n=mpoullet@77-22-121-59-dynip.superkabel.de) has left #beagle
[23:21:10] * florian (n=fuchs@f048070029.adsl.alicedsl.de) Quit ("Verlassend")
[23:23:02] <AudioFreak> but all these media player plans have to wait :(
[23:23:23] <AudioFreak> i first need a pre-amp, which i bought the beagle for
[23:25:34] <AudioFreak> though depending on the hardware resources left over i might be able to stash in the player as well
[23:33:42] * brolin (n=brolin@190.157.13.75) has joined #beagle
[23:36:28] * likewise (n=likewise@82-171-51-231.ip.telfort.nl) has joined #beagle
[23:36:53] * rbelem (n=rodrigo@pdpc/supporter/active/rbelem) Quit ("Leaving")
[23:40:16] * bmxr (n=bmxr@S01060018f3b11a22.vf.shawcable.net) Quit (Remote closed the connection)
[23:40:26] * robtow (n=rtow@nat/montavista/x-cc34bc99a5de6b92) Quit (Read error: 110 (Connection timed out))
[23:41:03] <AudioFreak> hmm, it getting late
[23:41:41] <AudioFreak> don't understand a bit of the DSP tutorial, so i'm going to bed
[23:42:10] <AudioFreak> hope i get it tomorrow :P
[23:42:42] <AudioFreak> good night everyone
[23:42:49] * AudioFreak is now known as AudioFreak|sleep
[23:47:40] * mib_vto769 (i=7aa62e54@gateway/web/ajax/mibbit.com/x-0f075b84e1c44996) has joined #beagle

These logs were automatically created by BeagleLogBot on irc.freenode.net using the Java IRC LogBot.

See http://microformats.org/wiki/mflogbot for more information.