• [00:22:01] * abitos (n=nixgibts@dslb-084-057-195-145.pools.arcor-ip.net) Quit (Read error: 113 (No route to host))
  • [00:23:43] * felipec (n=felipec@a88-113-6-167.elisa-laajakaista.fi) Quit ("Leaving")
  • [00:29:53] * jrmuizel (n=jrmuizel@mozca02.ca.mozilla.com) Quit ()
  • [00:34:29] * khasim (n=a0393720@192.163.20.231) has joined #beagle
  • [00:39:35] * emeb_mac (n=ericb@ip72-223-84-209.ph.ph.cox.net) has joined #beagle
  • [01:00:22] * emeb_mac (n=ericb@ip72-223-84-209.ph.ph.cox.net) Quit ()
  • [01:07:29] * emeb (n=ericb@ip72-223-84-209.ph.ph.cox.net) has left #beagle
  • [01:21:18] * emeb_mac (n=ericb@ip72-223-84-209.ph.ph.cox.net) has joined #beagle
  • [01:30:46] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
  • [01:38:06] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
  • [01:47:03] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
  • [02:03:46] * gcohler (n=gcohler@pool-98-110-251-38.bstnma.fios.verizon.net) has left #beagle
  • [02:15:02] * tulpe (n=ask@p54AF7190.dip.t-dialin.net) Quit (Read error: 60 (Operation timed out))
  • [02:15:14] * tulpe (n=ask@p54AF704F.dip.t-dialin.net) has joined #beagle
  • [02:15:30] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
  • [02:18:45] * keesj (n=keesj@ip49-193-210-87.adsl2.static.versatel.nl) Quit (Read error: 110 (Connection timed out))
  • [02:22:13] * _ZeZu_ (n=null@c-24-14-8-40.hsd1.in.comcast.net) has joined #beagle
  • [02:23:49] * FuL|OUT (n=fn@81.84.214.243) Quit (kubrick.freenode.net irc.freenode.net)
  • [02:23:49] * ZeZu (n=null@c-24-14-8-40.hsd1.in.comcast.net) Quit (kubrick.freenode.net irc.freenode.net)
  • [02:23:49] * denix (n=denys@nat/ti/x-eb775eba114fbd96) Quit (kubrick.freenode.net irc.freenode.net)
  • [02:23:49] * philips_ (n=brandon@ash.osuosl.org) Quit (kubrick.freenode.net irc.freenode.net)
  • [02:24:40] * FuL|OUT (n=fn@81.84.214.243) has joined #beagle
  • [02:24:40] * ZeZu (n=null@c-24-14-8-40.hsd1.in.comcast.net) has joined #beagle
  • [02:24:40] * philips_ (n=brandon@ash.osuosl.org) has joined #beagle
  • [02:24:40] * denix (n=denys@nat/ti/x-eb775eba114fbd96) has joined #beagle
  • [02:28:38] * denix (n=denys@nat/ti/x-eb775eba114fbd96) Quit (kubrick.freenode.net irc.freenode.net)
  • [02:28:38] * ZeZu (n=null@c-24-14-8-40.hsd1.in.comcast.net) Quit (kubrick.freenode.net irc.freenode.net)
  • [02:28:38] * philips_ (n=brandon@ash.osuosl.org) Quit (kubrick.freenode.net irc.freenode.net)
  • [02:28:38] * FuL|OUT (n=fn@81.84.214.243) Quit (kubrick.freenode.net irc.freenode.net)
  • [02:29:45] * FuL|OUT (n=fn@81.84.214.243) has joined #beagle
  • [02:29:45] * philips_ (n=brandon@ash.osuosl.org) has joined #beagle
  • [02:29:45] * denix (n=denys@nat/ti/x-eb775eba114fbd96) has joined #beagle
  • [02:36:20] * nathanm (n=nathan@adsl-69-106-228-95.dsl.pltn13.pacbell.net) Quit (Read error: 113 (No route to host))
  • [02:36:56] * BThompson (n=BThompso@cpe-76-185-93-11.tx.res.rr.com) Quit ("Trillian (http://www.ceruleanstudios.com")
  • [02:41:18] * keesj (n=keesj@ip49-193-210-87.adsl2.static.versatel.nl) has joined #beagle
  • [02:50:09] * bullwood42 (n=bullwood@adsl-99-171-171-142.dsl.aus2tx.sbcglobal.net) Quit ("Powered by OE: www.openembedded.org")
  • [03:15:53] * khasim (n=a0393720@192.163.20.231) has joined #beagle
  • [03:25:33] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
  • [03:42:19] * Beagle1 (n=Beagle1@97-114-87-186.spkn.qwest.net) has joined #beagle
  • [04:10:33] * Beagle1 (n=Beagle1@97-114-87-186.spkn.qwest.net) has left #beagle
  • [04:27:01] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
  • [04:30:49] * soman (n=somnath@61.16.248.242) has joined #beagle
  • [04:32:52] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
  • [04:33:05] * jkridner|work (n=a0321898@nat/ti/x-b1fb4f1f75a59763) has joined #beagle
  • [04:34:57] * jkridner|work (n=a0321898@nat/ti/x-b1fb4f1f75a59763) Quit (Remote closed the connection)
  • [04:35:24] * shiv (n=root@59.160.172.220) has joined #beagle
  • [04:45:37] * emeb_mac (n=ericb@ip72-223-84-209.ph.ph.cox.net) Quit ()
  • [04:45:58] * jkridner|work (n=a0321898@nat/ti/x-760b3660db5b997d) has joined #beagle
  • [05:04:13] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
  • [05:16:14] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
  • [05:20:27] * DaQatz (n=db@c-66-30-48-54.hsd1.nh.comcast.net) has joined #beagle
  • [05:20:56] * DaQatz (n=db@c-66-30-48-54.hsd1.nh.comcast.net) Quit (Remote closed the connection)
  • [05:23:47] * khasim (n=a0393720@192.163.20.231) has joined #beagle
  • [05:32:20] * emeb_mac (n=ericb@ip72-223-84-209.ph.ph.cox.net) has joined #beagle
  • [05:33:12] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
  • [05:44:27] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
  • [06:12:59] * jsync (n=jess@59.160.172.220) has joined #beagle
  • [06:13:50] * khasim (n=a0393720@192.163.20.231) has joined #beagle
  • [06:21:51] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
  • [06:23:28] <jkridner|work> jffs2 in koen's 2.6.27 build seems to be busted for me.
  • [06:24:15] * khasim (n=a0393720@192.163.20.231) has joined #beagle
  • [06:25:51] * khasi1 (n=a0393720@192.163.20.231) has joined #beagle
  • [06:25:51] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
  • [06:26:56] <ldesnogu> mru: did you try your memcpy_armneon with overlapping memory areas to see if it hits some famous A8 errata?
  • [06:28:31] * khasim (n=a0393720@192.163.20.231) has joined #beagle
  • [06:28:32] * khasi1 (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
  • [06:33:19] * Beagle2 (n=Beagle2@130.231.50.117) has joined #beagle
  • [06:35:03] * nathanm (n=nathan@adsl-69-106-228-95.dsl.pltn13.pacbell.net) has joined #beagle
  • [06:49:37] * methril|gone is now known as methril
  • [06:50:50] * emeb_mac (n=ericb@ip72-223-84-209.ph.ph.cox.net) Quit ()
  • [06:55:15] * Beagle2 is now known as ewerw
  • [06:56:34] * koen starts cleaning up the dsplink hacks from the past few days
  • [07:00:16] <jkridner|work> simplified my u-boot hack: http://www.beagleboard.org/uploads/u-boot-omap3/ignorenv.txt
  • [07:00:35] <jkridner|work> looks like there are some other hacks in u-boot along similar lines, but I haven't figured out how to use any of them.
  • [07:01:11] <jkridner|work> I'm struggling to figure a way to try 'fatload mmc 0 ...' when no card is present to see if I can fall-back to booting from NAND flash.
  • [07:02:09] <jkridner|work> jffs2 stuff doesn't seem very solid in "2.6.27-rc7-omap1 (koen@dominion) (gcc version 4.2.1) #1 Thu Oct 9 09:35:52 CEST 2008"
  • [07:04:12] <koen> never tried jffs2 on that kernel
  • [07:06:06] <ds2> what does jffs2 do right now?
  • [07:06:24] * trickie|work (n=trickie@a5061.upc-a.chello.nl) has joined #beagle
  • [07:06:24] <koen> waiting to get killed by ubifs
  • [07:06:30] <jkridner|work> mounts for a while, then starts spitting out register dumps.
  • [07:06:47] <ds2> I suspect fatload will hang if there is no card unless they made a big change
  • [07:07:06] <jkridner|work> koen: have you patched in ubifs into any of the kernels?
  • [07:07:12] <jkridner|work> ds2: it does.
  • [07:07:17] <ds2> ewwww ubifs
  • [07:07:24] <jkridner|work> ds2: any way I can test for card presence?
  • [07:07:27] <ds2> jkridner|work: read a raw image w/o a FS
  • [07:07:34] <koen> jkridner|work: isn't ubifs in .27rc already?
  • [07:07:39] <ds2> jkridner|work: none that I can tell
  • [07:08:07] <jkridner|work> ds2: were those both answers to the same question?
  • [07:08:37] <ds2> jkridner|work: I don't know of a way to check for card presense
  • [07:09:07] <jkridner|work> so, reading a raw image w/o a FS wouldn't work? (not sure how you'd try that anyway)
  • [07:09:58] <ds2> reading a raw image should work.... it might get tricky with bad blocks (donno if nandwrite in Linux handles bad blocks the same as nand write.jffs2 in U-boot)
  • [07:10:23] <ds2> that's what the default boot command does...except they dont' seem to check for bad blocks
  • [07:10:47] <jkridner|work> oh, reading a raw file from NAND is direct, reading one from SD/MMC is an unknown to me.
  • [07:12:07] <jkridner|work> I've got a config that will load a ramdisk image from NAND flash now, even with the 2.6.27 kernel, but I wanted to boot from the SD card if it was installed.
  • [07:12:25] <jkridner|work> (it doesn't do anything with flash access in the kernel)
  • [07:13:04] <jkridner|work> struggle is that my 32MB ramdisk image is a bit small for some of the apps I'd want to load.
  • [07:13:22] <jkridner|work> i may switch to a kernel that I know was working with JFFS2 tomorrow.
  • [07:14:05] <ds2> YAFFS if you are adventurous
  • [07:14:27] <ds2> actually, I think the kernel I built has a working JFFS2.... at least it was working as of a few months ago
  • [07:14:41] <jkridner|work> k
  • [07:15:32] * and-ri (n=x0103690@nat/ti/x-d1eedad5bf35ec49) has joined #beagle
  • [07:15:50] <jkridner|work> you might look to see if you can adjust http://www.beagleboard.org/uploads/u-boot-omap3/mkubootcmd to make http://www.beagleboard.org/uploads/u-boot-omap3/u-boot.bin work for you.
  • [07:16:36] <jkridner|work> at this point, executing arbitrary u-boot commands via u-boot.img is reliable.
  • [07:16:51] <jkridner|work> so, I can configure the flash environment however we'd like.
  • [07:17:20] <jkridner|work> I may have to have a collection of u-boot.img files for the various classes.
  • [07:17:39] <jkridner|work> I also have a ramdisk image that automatically executes boot.sh on the SD card.
  • [07:18:22] <ds2> the only thing I found so far is I either need to change ~/.profile or put in a symlink
  • [07:18:36] <jkridner|work> it has mtd-utils and most of the flash utils I need, though I do need to add mkfs.jffs2 (though flash_eraseall and mount -t jffs I think should do the trick).
  • [07:19:27] <jkridner|work> if you need to have an SD card inserted to run this environment config, how will you get it to alter the SD cards?
  • [07:19:58] <jkridner|work> do we need to show up early and start overwriting >300 SD cards? :)
  • [07:20:22] <ds2> I can't think of a good way to do this
  • [07:20:37] * Spexius (n=spex@gb-r9-01.pixelpark.net) has joined #beagle
  • [07:21:12] <ds2> current plan is to announce a 'errata task' just after going through the theory before the lab portion starts
  • [07:21:30] <jkridner|work> well, if I at least get everyone to be able to boot into the Angstrom ramdisk from their NAND flash, they can manually add the symlink.
  • [07:22:13] <jkridner|work> then, they would need to run another magic card to change the bootcmd back to SD card boot.
  • [07:22:14] <ds2> if there was NOR, we could stash stuff in raw flash
  • [07:22:46] <jkridner|work> will they be able to get to the point to execute the errata task?
  • [07:22:49] <ds2> actually, they can boot into SD card, wait for a prompt on the console; swap cards; hit enter and proceed
  • [07:23:05] <jkridner|work> that's the issue, isn't it? if it is possible to execute the fix w/o a serial cable?
  • [07:23:10] <ds2> for me, yes. so I have a viable workaround...
  • [07:23:24] <ds2> sure it is, you do this over the framebuffer console
  • [07:23:37] <jkridner|work> k
  • [07:23:52] <wiking> anybody knows which version of gcc started to have support for armv7-a ?
  • [07:23:54] <ds2> it'll be something like - boot to ramdisk; show prompt to change cards; swap cards; click okay or hit enter
  • [07:23:54] <jkridner|work> got all the tools in place for that now?
  • [07:24:00] <jkridner|work> does this little u-boot hack help?
  • [07:24:05] <jkridner|work> k
  • [07:24:15] <jkridner|work> yeah, that should work.
  • [07:24:18] <ds2> this u-boot hack is overkill for what I need.
  • [07:24:34] <jkridner|work> oh, b/c it has bootargs replacement?
  • [07:24:44] <ds2> because you are loading all the other stuff
  • [07:24:52] <jkridner|work> you still needed something that would automatically boot, ignore bootcmd, and be able to write bootcmd.
  • [07:25:01] <jkridner|work> oh, just the script itself.
  • [07:25:04] <ds2> all I need is to get to the login prompt with stock image
  • [07:25:18] <jkridner|work> that is just experimentation... you can use a much simpler script.
  • [07:25:50] <ds2> true; but if your script leaves the board in a state that it can boot the stock flash disk, I can work from there
  • [07:26:36] <ds2> my main concerns right now are else where - will the parts be there and will the room be too big
  • [07:28:24] <jkridner|work> http://gcc.gnu.org/gcc-4.3/changes.html ?
  • [07:28:30] * rupeshmg (n=rupesh@59.160.172.220) has joined #beagle
  • [07:28:36] <jkridner|work> that was to wiking.
  • [07:29:24] * rupeshmg (n=rupesh@59.160.172.220) has left #beagle
  • [07:29:46] <wiking> jkridner|work, thnx a lot
  • [07:30:25] <jkridner|work> too bad GCC is still using SVN.
  • [07:30:42] <ds2> better then CVS
  • [07:30:47] <jkridner|work> by flash disk you mean SD card?
  • [07:31:07] <jkridner|work> did you get the BT dongles?
  • [07:31:16] <ds2> yeah SD card
  • [07:31:29] <ds2> they are suppose to be there per the "authority"
  • [07:31:41] <ds2> no choice but to have faith as I am 3000+ miles away
  • [07:31:42] * jkridner|work crosses fingers.
  • [07:32:04] <jkridner|work> similar with the projectors should work with Beagle. :)
  • [07:32:28] <ds2> I am 100% counting on that cuz my slides are on the beagle!
  • [07:33:08] <ds2> actually, worse case you can use the S-Video port... I think
  • [07:33:21] <jkridner|work> let me know if you think of a way for me to default to booting from the SD card, but allows people to still optionally boot from the NAND.
  • [07:33:32] <ds2> untested but no time better to try it then 5 minutes before and you have no other choice
  • [07:34:06] <ds2> I can think of several ways, just none that I would recommend doing at this time ;)
  • [07:34:54] <ds2> ie. you can have U-boot read the state of the User button and change defaults based on that but in 2 days, that could be risky
  • [08:01:52] * RogerMonk (n=a0740758@nat/ti/x-f07db979f44ce69c) has joined #beagle
  • [08:03:49] * adj_ (i=antti@hervanta.com) has joined #beagle
  • [08:06:16] <RogerMonk> Koen - morning!
  • [08:06:24] <koen> good morning
  • [08:14:40] * [X]Spoty is now known as Stasi
  • [08:14:58] <Stasi> RogerMonk good morning ;)
  • [08:15:22] <mru> morning
  • [08:19:50] <koen> hey mru
  • [08:36:03] <and-ri> do anybody work with the beagleboard which is connected via DVI to a TFT? How to configure the resultion? I found a document which explain the hard coded resultion in the kernel, but the changes doesn't work
  • [08:45:06] <koen> if you have the angstrom .27rc testing kernel you can select different modes with uboot
  • [08:45:16] <koen> otherwise you'll need a better monitor ;)
  • [08:50:07] <wiking> hmmm the video accelerator part of iva2.2 in omap3 will be every be available for use?
  • [08:56:24] <_AV500_> depends on whether TI releases the specs. TI?
  • [08:57:14] <wiking> doh... again a part of the chip that is not available :(
  • [08:57:48] <koen> wiking: you mean the motion estimation engine?
  • [08:59:00] <_AV500_> oops, looks like this is in the specs actually
  • [08:59:15] <_AV500_> iME, ILF and iVLCD
  • [08:59:53] * florian__ (n=fuchs@217.146.132.69) has joined #beagle
  • [09:00:18] <koen> hey florian__
  • [09:00:31] * florian__ is now known as florian
  • [09:00:32] <wiking> i'm just reading the sprufa3b.pdf specs of IVA and there is a submodule Video Hardware Accelerator
  • [09:00:42] <_AV500_> btw, there is another ARM968E-S core called the "video sequencer" inside the IVA :-)
  • [09:00:44] <florian> good morning
  • [09:04:16] * RogerMon1 (n=a0740758@nat/ti/x-b15f492f63db1b98) has joined #beagle
  • [09:04:34] <wiking> but yet i haven't seen any specs about it's registers
  • [09:09:48] * RogerMonk (n=a0740758@nat/ti/x-f07db979f44ce69c) Quit (Remote closed the connection)
  • [09:09:53] <wiking> koen, just found your comment in irc logs about this... i guess the situation hasn't changed since beginning of september
  • [09:12:57] <koen> dsplink built for davinci in OE :)
  • [09:13:04] <koen> </random comment>
  • [09:13:52] <florian> koen: you always have good news :-)
  • [09:15:23] <florian> koen: do we have all the musb patches in oe metioned on the list this morning?
  • [09:18:25] <_AV500_> wiking: I have the OMAP3 TRM that documents this, but I am also under NDA.
  • [09:18:43] <wiking> lol
  • [09:18:45] <wiking> GIMMEEE!!!
  • [09:18:46] <wiking> :D
  • [09:19:17] <wiking> _AV500_, do you know if there's any plan to make this public?
  • [09:19:32] <wiking> or it will be as with the gfx accel part of the chip?
  • [09:20:23] <koen> florian: the linux-omap recipe has them all
  • [09:21:40] <_AV500_> wiking: the sprufa3b seems to describe the generic iva2.2, not the "special" one in the OMAP3 which also has the accels
  • [09:21:51] <_AV500_> I have no idea whether TI will release that info
  • [09:22:30] <florian> koen: very good... or no, not that good because it still does not work :-)
  • [09:23:01] <wiking> _AV500_, :(
  • [09:30:28] <_AV500_> but to use the accels, you would 1st need to make the DSP work for you, you already that far?
  • [09:30:39] <wiking> yep
  • [09:30:44] <wiking> i mean the c64x
  • [09:31:07] * igor321 (n=igor@93-138-101-153.adsl.net.t-com.hr) has joined #beagle
  • [09:31:33] <wiking> _AV500_, currently trying to port a decent linux distro for that board - aka owrt :D
  • [09:32:53] * abitos (n=nixgibts@dslb-084-057-206-232.pools.arcor-ip.net) has joined #beagle
  • [09:34:46] * aleph_6510 (n=aleph@trinity.develer.com) has joined #beagle
  • [09:55:15] * wiking (n=wiking@91.83.6.179.pool.invitel.hu) Quit ("Leaving")
  • [09:59:16] <kulve> what kernels people are runnin on EVM boards? I got linux-omap HEAD booted up, but X shows weird colors. And with the v2 kernel from linux.omap.com I'm having troubles to even to get X started up..
  • [10:01:36] <koen> I'm having troubles with getting the lcd to turn on on the omap3evm with a git kernel
  • [10:01:45] <koen> haven't checked latest git or latest uboot
  • [10:02:49] * koen is now playing with a davinci evm
  • [10:04:33] <kulve> hmm.. I actually used the u-boot-v2 from linux.omap.com and the kernel from the linux-omap tree. What u-boot should have I tried with the linux-omap kernel?
  • [10:05:51] <koen> the one from sakomans tree ( or upstream + patches from the uboot ml)
  • [10:06:58] <kulve> I'll try the sakoman's tree next
  • [10:07:00] <kulve> thanks
  • [10:23:48] * aleph_6510 (n=aleph@trinity.develer.com) has left #beagle
  • [10:30:32] * aleph_6510 (n=aleph@trinity.develer.com) has joined #beagle
  • [10:43:05] <koen> florian: could you restart apache on ltg?
  • [10:43:56] <kulve> koen: I git clone the sakoman's u-boot but I don't see any omap3 boards there..?
  • [10:44:07] <kulve> should I have used some branch or something..?
  • [10:44:49] <koen> kulve: the 'common' or 'overo' branch
  • [10:45:26] <koen> steve will wakeup in ~2 hours I guess
  • [10:46:27] <kulve> thanks, I found the omap3/evm config
  • [10:49:24] * florian (n=fuchs@217.146.132.69) Quit (Read error: 113 (No route to host))
  • [10:55:52] * florian (n=fuchs@217.146.132.69) has joined #beagle
  • [11:09:26] * koen wonders why people expect that random (shared library!) binaries work without installing their deps
  • [11:09:45] * koen also wonders why people expect random binaries to work on a random arm system
  • [11:23:33] * ldesnogu wonders why people expant random arm system to even work
  • [11:33:55] * ewerw (n=Beagle2@130.231.50.117) Quit (Remote closed the connection)
  • [11:35:29] * [X]Spot (n=stancho@78.90.115.116) has joined #beagle
  • [11:35:29] * Stasi (n=stancho@78.90.115.116) Quit (Read error: 104 (Connection reset by peer))
  • [11:44:54] * shiv (n=root@59.160.172.220) Quit (Remote closed the connection)
  • [11:59:59] * robclark (n=robclark@nat/ti/x-52b98e00219ae9e1) has joined #beagle
  • [12:04:29] * khasim wants to know more about openmax - any good tutorials
  • [12:11:11] * felipec (i=c0647cdb@gateway/web/ajax/mibbit.com/x-ffb8e436dc758d50) has joined #beagle
  • [12:20:25] <khasim> will beaglegame run on beagleboard itself. Did some one try it with angstrom rootfs?
  • [12:20:31] <khasim> koen: ?
  • [12:21:17] <Crofton> koen, please remove 128M of ram from the evm and re-test :)
  • [12:26:39] <_AV500_> khasim: the game asks for Flash9, it does not run with Flashlite 3.1 :-(
  • [12:34:56] <khasim> _AV500_: thanks for the update
  • [12:35:21] <khasim> Today we were having a big discussion about open max v/s TI APIs for DSP access
  • [12:35:37] <khasim> which is the way for future?
  • [12:36:49] <khasim> I was arguing in favour of openmax, now I have to get some data to show how going with openmax is much better compared to VISA APIs, there are also some good stuff that I found in VISA APIs that is not part of Open Max, but that could be ported I believe
  • [12:37:33] <khasim> Does any one know how other Operating Systems are using Open Max, is this really a defacto standard that could be considered or should we look more into VISA APIs
  • [12:38:22] <khasim> I am pulling for a DSP Foundation that comprises of TI, Open Community and few others to define standard interface for DSP (usage) from Linux. will update if I get a green signal on doing the same
  • [12:38:32] <_AV500_> to me OMX looked like a big layer doing nothing
  • [12:38:48] <_AV500_> but then I just don't want to port all my CE stuff to OMX
  • [12:38:51] <_AV500_> :-)
  • [12:39:13] <khasim> CE will still exit below OMX
  • [12:39:29] <_AV500_> I thought OMX loves bridge. no?
  • [12:39:36] <_AV500_> I thought OMX loves bridge more, no?
  • [12:40:30] <khasim> there should be one solution, at least in future I am looking at one layer that acts as a bridge/CE above Link APIs.
  • [12:41:10] <koen> Crofton: actually CE on evm is expecting 128MB now...
  • [12:41:17] <_AV500_> when I looked at OMX I was immediately put off by seeing that it defines totally different metadat structures for EACH codec
  • [12:41:26] <Crofton> great
  • [12:41:27] <koen> Crofton: so the beagle and the dvevm are pretty crippled at the moment
  • [12:41:29] * Zedstar_ (n=user@217.41.237.1) has joined #beagle
  • [12:41:32] <Crofton> heh
  • [12:41:38] <Crofton> what about link?
  • [12:41:46] <khasim> koen: for which application
  • [12:41:52] <_AV500_> ??? beagle has 128M, no?
  • [12:41:54] <Crofton> or is CE building its own link?
  • [12:42:01] <koen> Crofton: link has no memory requirements IIRC
  • [12:42:03] <khasim> koen: I use 80M for my MP4 demo
  • [12:42:16] <koen> _AV500_: my beagle has 256MB ram :)
  • [12:42:50] <_AV500_> yes, still I don't understand: actually CE on evm is expecting 128MB now...
  • [12:43:07] <khasim> My bootargs to run kingkong video using CE/Link has mem=80M in bootargs
  • [12:43:47] <RogerMon1> Koen/Crofton/khasim - link and ce have configurable memory maps - they just have some defaults.
  • [12:43:47] * bazbell (n=a0192809@nat/ti/x-d693f63291321619) Quit ("Leaving.")
  • [12:43:53] <koen> I have mem=120M on the davinci evm now
  • [12:43:58] <koen> right
  • [12:44:26] <RogerMon1> dsplink examples require 1MB of ram, on davinci, the default is using 255->256MB
  • [12:44:35] <Crofton> does link now have a default for 128M max systems?
  • [12:44:37] <RogerMon1> on omap, default is 127->128
  • [12:44:41] <RogerMon1> no
  • [12:44:59] <RogerMon1> Crofton, but it's easy to change - we'll test that this afternoon
  • [12:45:05] <_AV500_> but these are defaults anyway
  • [12:45:11] <_AV500_> just change them
  • [12:45:15] <RogerMon1> _AV500_ exactly
  • [12:45:26] <RogerMon1> there's no issue here - just defaults
  • [12:45:38] <koen> Packaged contents of dsplink-lpm-module into /OE/angstrom-dev/deploy/glibc/ipk/davinci-dvevm/dsplink-lpm-module_221-r0.1_davinci-dvevm.ipk
  • [12:45:41] <koen> Packaged contents of dsplink-cmemk-module into /OE/angstrom-dev/deploy/glibc/ipk/davinci-dvevm/dsplink-cmemk-module_221-r0.1_davinci-dvevm.ipk
  • [12:46:30] <Crofton> RogerMon1, thanks
  • [12:47:33] <Crofton> I am working out of town this week, otherwise I would be testing on the sffsdr board
  • [12:48:13] * _AV500_ is now known as AV500
  • [12:49:46] <khasim> do we know who fixed the software reset issue on beagle?
  • [12:50:20] <koen> with .27rc?
  • [12:50:26] <AV500> khasim: if you are looking into a "clean" API for audio/video decode, make it compatible to ffmpeg :-)
  • [12:51:32] <AV500> in my codebase, the "driver" for using CE is about 10x larger than the one for ffmpeg
  • [12:51:39] <AV500> I fear, OMX will not be much different
  • [12:51:53] <khasim> koen: earlier there was an issue that reboot/reset used to fail, I think reboot works now, I still see some hang with IRQ -33 with reset on board - sometimes
  • [12:52:29] <koen> there are patches to solve the -33 issue
  • [12:52:34] * rsalveti (n=salveti@189.70.63.219) Quit (Read error: 113 (No route to host))
  • [12:52:36] <koen> the angstrom .27rc kernel has one
  • [12:52:48] <koen> iirc tony applied another one to git yesterday
  • [12:53:13] <khasim> AV500: I can propose now, but I need to make a formal proposal, as mentioned I was pushing towards OMX as its more open standard than VISA or Bridge APIs
  • [12:53:33] <khasim> koen: thanks will look into this
  • [12:54:39] <felipec> khasim: I know some solutions using openmax already
  • [12:54:46] <AV500> khasim: yes I see that you introduced IUNIVERSAL, so you gave up on VISA in the end :-)
  • [12:55:17] <felipec> khasim: TI, STMicro, Marvell, Hantro
  • [12:55:27] <AV500> android
  • [12:55:54] <khasim> I know TI's solution so far, but I think there is lot of redundancy like Tunnelling when integrated with gStreamer
  • [12:56:24] <AV500> yes, it seemed to me that both gstreamer and OMX implement buffer handling/sharing
  • [12:56:28] <felipec> khasim: yes, but you don't need to use the full standard
  • [12:56:43] <khasim> felipec: like?
  • [12:56:53] <AV500> felipec: but then you end up with a glorified way to call a codec, no?
  • [12:56:56] <felipec> khasim: tunneling components are part of the interop specification, you can have components that are not compliant with that spec
  • [12:57:35] <felipec> AV500: yes, but it's a standard API
  • [12:57:36] <khasim> felipec: I am more worried about not following a proper standard, will it again not deviate us from a standard interface
  • [12:57:54] * john|new (n=a0270504@nat/ti/x-c2c58e784837e404) has joined #beagle
  • [12:58:07] <khasim> felipec: How easy is it to get some standards/APIs rectified in OpenMax 2.0?
  • [12:58:19] <AV500> yes, but then CE was supposed to be "standard", yet no a single codec I touched cared
  • [12:58:53] <felipec> AV500: who managed that CE standard?
  • [12:59:01] <AV500> TI, I guess :-)
  • [12:59:18] <AV500> but CE did not "enforce" it
  • [12:59:24] <felipec> khasim: you would need some contacts in the working group =/
  • [12:59:59] <khasim> AV500: There is a VISA layer on top of CE, that is public and a standard. but it is more from TI side, it will make more sense to follow some open standard where a community is focused today -
  • [13:00:15] <felipec> AV500: well, OpenMAX IL is managed by Khronos, so it can be a standard as OpenGL, which is not to say much, but it's something
  • [13:00:16] * ful (i=c2913c85@gateway/web/ajax/mibbit.com/x-ed74ad9f81713e74) has joined #beagle
  • [13:00:18] <khasim> TI has a member in working group, I can check with them
  • [13:00:39] <AV500> khasim: yes I know VISA in and out, but still I had to tweak my VISA "driver" for every codec I got
  • [13:00:41] <khasim> Is OpenMax used with other operating system as well
  • [13:00:55] <AV500> because CE/VISA does not "enforce" anything
  • [13:01:29] <felipec> khasim: it's OS agnostic, I know at least S60, PS3, Linux, and Windows CE
  • [13:01:31] <khasim> AV500: that is bad we need to address such issues
  • [13:01:35] <AV500> in the end, a lot of codecs did it the IUNIVERSAL way, by just adding adding random stuff to IVIDDEC etc.
  • [13:01:39] <felipec> whatever PS3 is using =/
  • [13:01:41] * vu3rdd (n=user@117.196.133.235) has joined #beagle
  • [13:02:39] <AV500> khasim: some codec reorder internally, some reorder using the outbufs, some need the extradata per PARSE_HEADER, other can consume it inline etc.
  • [13:03:22] <khasim> felipec: I know other its os agnostic, but are others using it? it shouldn't be a dead case :)
  • [13:03:33] <AV500> some use the standard error code, others use some vendor specific
  • [13:03:57] <khasim> AV500: was this on Davinci?
  • [13:04:04] <AV500> and OMAP3
  • [13:04:32] <AV500> the "cleanest" codecs were from TI, but 3rd parties bent the API a lot
  • [13:04:45] <felipec> khasim: TI provides solutions for Symbian, Linux and Windows CE
  • [13:05:11] <khasim> felipec: I know Symbian team that is working on here, will get in touch with WinCE
  • [13:05:47] <khasim> felipec: again I don't want to get into TI's solution, but it should be a generic solution that gets sustained and evolves with time...
  • [13:06:48] <khasim> felipec / AV500: Do you guys suggest OpenMax as the ideal way for us to go with Multimedia and other apps on DSP?
  • [13:07:12] <khasim> how about the other side, do we have enough apps that use Open Max today?
  • [13:07:38] <AV500> khasim: does OMX reayll support any app? Or is it media centric?
  • [13:07:48] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
  • [13:07:48] <felipec> khasim: so you have providers and users, there are many providers in different platforms, but unfortunately there aren't many users yet
  • [13:08:24] <AV500> I fear if you want it universal, it will be just a RPC mechanism
  • [13:08:35] <felipec> khasim: ultimately OMX is just a bunch of headers
  • [13:09:06] * AV500 does not like mure bunches of headers
  • [13:09:28] * JuanG (n=Juan@nat/ti/x-7592aa7c58245b8e) has joined #beagle
  • [13:09:37] * JuanG (n=Juan@nat/ti/x-7592aa7c58245b8e) has left #beagle
  • [13:11:15] <felipec> AV500: the less there is to agree, the easier
  • [13:11:55] <khasim> it is good time to define a strategy where we should go
  • [13:12:10] <koen> if e.g. gstreamer and ffmpeg support openmax backend you gain tons of apps for free
  • [13:12:24] <koen> since pretty much every multimedia app on linux uses ffmpeg or gstreamer
  • [13:12:31] <koen> (or xine, but that depends on ffmpeg as well)
  • [13:12:43] <khasim> koen: I remember there were lots of issues in integrating gstreamer with Open Max IL
  • [13:12:44] <AV500> if you make the API ffmpeg compatible, it is even easier to plug it :-)
  • [13:13:52] <AV500> and you keep mru happy
  • [13:14:06] <felipec> khasim: there were issues, but not that many
  • [13:14:12] * khasim thinks that is important
  • [13:14:16] <felipec> it has been integrated to Helix too
  • [13:14:18] <khasim> felipec: ok
  • [13:14:45] <AV500> felipec: at what level is the integration? At the level to open/decode/close a codec?
  • [13:15:13] <khasim> felipec: can you give some info on it's integration to gstreamer and ffmpeg
  • [13:15:56] <felipec> AV500: play/pause/stop, seek, tunneling, zero-copy
  • [13:15:59] * abitos (n=nixgibts@dslb-084-057-206-232.pools.arcor-ip.net) Quit (Read error: 113 (No route to host))
  • [13:16:51] <AV500> felipec: www.celinux.org/elc08_presentations/gst-openmax.pdf I guess this is you :-)
  • [13:17:26] <felipec> AV500: yeap :)
  • [13:17:57] <felipec> unfortunately there was no beagle to show a fancy demo :( I just ran it in my laptop with Bellagio
  • [13:19:07] <khasim> koen/felipec/AV500: as I mentioned we need to define a way to go forward other wise we will be again spending enormous time on re inventing the work, I will need your help in framing this direction.
  • [13:20:34] <felipec> khasim: do you have information about your solution?
  • [13:21:17] <khasim> nothing planned yet, I am still at proposing the direction.
  • [13:21:48] <khasim> I think open max, but again bit hessitant as it doesn't fit with ffmpeg yet
  • [13:22:07] <AV500> felipec: when I looked at OMX, I had the impression that it is not really codec agnostic, for each codec there was a separate property struct defined
  • [13:22:27] <AV500> although they would share 90% of the props.
  • [13:22:34] <felipec> AV500: no, it's not codec agnostic, it's not extensible to other codecs
  • [13:22:35] <AV500> although they would share 90% of the props in real life
  • [13:22:48] <AV500> ??
  • [13:22:55] <felipec> AV500: if you implement an extension, that's outside the spec
  • [13:22:58] <khasim> AV500: doesnt this get addressed by CE?
  • [13:23:29] <khasim> I thought CE should handle all the Codec accesses
  • [13:24:21] <khasim> shall we list down issues:
  • [13:24:34] <AV500> felipec: look at page 18 of your slides, I handle about twice the codecs today, so what do I do with OMX?
  • [13:24:39] <khasim> 1) Codec extensions are not supported in openmax
  • [13:25:21] <felipec> AV500: those are the ones implemented in gst-openmax, there are many more in the spec
  • [13:25:41] <AV500> OK, but even the spec missed some IIRC
  • [13:26:06] <khasim> felipec: how about adding a new generic algo like opensdr....
  • [13:26:33] * igor321 (n=igor@93-138-101-153.adsl.net.t-com.hr) Quit ("leavin'")
  • [13:27:29] <and-ri> how to empty the linux framebuffer in the console? or under C?
  • [13:27:47] <koen> cat /dev/zero > /dev/fb
  • [13:27:54] <felipec> AV500: yes, but for example, a FLAC decoder, you don't need to know that it's a FLAC decoder, you can threat it as a generic audio decoder
  • [13:28:11] <khasim> felipec: in case of spec voilation can't the spec be upgraded to handle new inputs
  • [13:28:17] <felipec> AV500: you just dump input buffers, and get pcm data
  • [13:29:05] <felipec> khasim: yes, other formats are continually being integrated, but it's a process, other parties need to agree on the fields passed through the standard structs
  • [13:29:45] <khasim> felipec: I think its fine...unless it doesn't drive every crazy with unnecessary delays
  • [13:29:47] <AV500> felipec: but the same is true for most of the video decoders, no?
  • [13:30:02] <AV500> you can treat them as generic video decoders
  • [13:30:09] <felipec> AV500: yes, exactly
  • [13:30:32] <koen> hmmm
  • [13:30:41] <koen> sound is broken on dvevm, but we already knew that
  • [13:30:42] <khasim> how about ffmpeg support?
  • [13:30:49] <khasim> where do we stand on this?
  • [13:31:01] <AV500> trouble with ffmpeg is that the API might "evolve"
  • [13:31:36] <koen> both gstreamer and openmax *L have an ffmpeg backend
  • [13:32:44] <kulve> sakoman_: ping
  • [13:33:25] <khasim> AV500/felipec: ok, then I will go with OpenMax as the proposal, I will include you both in further discussions
  • [13:33:50] * Geicus (n=afsd@64.132.1.226) has joined #beagle
  • [13:34:28] <Geicus> I am looking at this beagleboard b4 on ebay
  • [13:34:47] <Geicus> do you all think its worth it, since its a B4 and i've read that there are some bugs
  • [13:34:49] <Geicus> http://cgi.ebay.com/TI-Beagleboard-DSP-ARM-Core-Embedded-Linux_W0QQitemZ250312372058QQihZ015QQcategoryZ97184QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
  • [13:35:38] * soman (n=somnath@61.16.248.242) Quit ("Leaving")
  • [13:35:50] <AV500> so, they 1st person gave up? :-)
  • [13:35:56] <and-ri> koen: IF I TRY "cat /dev/zero > /dev/fb" HE SAID "WRITE ERROE: NO SPACE LEFT ON DEVICE
  • [13:35:59] <AV500> or is this a way around "export" control?
  • [13:36:25] <AV500> and-ri: your frame buffer is full, you have to delete some pixels first :-)
  • [13:36:41] * prpplague (n=dave@mail.americanmicrosystems.com) has joined #beagle
  • [13:37:26] <and-ri> AND HOW DO I THAT?
  • [13:37:44] * rsalveti (n=salveti@200.184.118.132) has joined #beagle
  • [13:37:44] <and-ri> I WANT TO EMPTY THE WHOLE BUFFER
  • [13:37:46] <LoneMeow> maybe start by turning off capslock
  • [13:37:46] <Spexius> and-ri: framebuffers have a limited size, while /dev/zero is an "endless" source of 0-information
  • [13:38:13] <Spexius> imagine writing to a file which has a max size, and you trying to write more data to it
  • [13:38:29] <and-ri> AH YEAH I UNDERSTAND
  • [13:38:31] <prpplague> and-ri: using capital letters is the irc equivilant of screaming
  • [13:38:53] <and-ri> BUT DO IT WHAT IT SHOULD DO, EMPTY THE BUFFER
  • [13:39:03] <and-ri> sorry
  • [13:39:45] <felipec> AV500/khasim: http://pastie.org/298873
  • [13:39:47] <Spexius> and-ri: to say otherwise: Your command works as in it clears the framebuffer (the default / fb0 one); just ignore the error message. Alternatively you could limit the data written to the framebuffer
  • [13:40:01] <and-ri> prpplague: i do not want scream, i just push the capslock and forgot it
  • [13:40:35] <felipec> that's for a simple audio decoder
  • [13:40:36] <and-ri> ah kk maybe this is what i needed
  • [13:40:45] <khasim> felipec: thanks.
  • [13:41:06] * vu3rdd (n=user@117.196.133.235) Quit (Read error: 110 (Connection timed out))
  • [13:41:11] <felipec> (client)
  • [13:41:19] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
  • [13:41:56] <khasim> felipec: one general question, how is community going with OpenMax? is there enough interest towards this? I know even ARM is pushing for OpenMax.
  • [13:41:57] <Geicus> how bad is rev B4? i've read the EHCI and 32khz are not working, any other problems?
  • [13:42:33] <kulve> khasim: community doesn't really have any reason to use OpenMAX atm.
  • [13:42:34] <felipec> khasim: you mean industry support or FOSS community?
  • [13:42:35] <khasim> Geicus: 32khz issues is simple to fix, just remove one resistor
  • [13:42:51] <AV500> felipec: OK; how do you pass the bitrate to the audio encoder?
  • [13:42:56] <khasim> FOSS community
  • [13:42:59] <felipec> khasim, AV500: and here is a minimal OpenMAX IL implementation with a single component: http://github.com/felipec/gst-openmax/tree/master/tests/standalone/core.c
  • [13:43:32] <khasim> kulve: even if DSP codecs are available?
  • [13:44:37] <and-ri> ok and the same thing in C?
  • [13:45:00] <kulve> khasim: once they are available there might be interest (just a guess). I don't see any reason why not. But for now there's one video codec from TI that runs on a DSP and I didn't even get it working..
  • [13:45:17] <khasim> :)
  • [13:45:26] <felipec> AV500: you would do something like: OMX_AUDIO_PARAM_MP3TYPE param, param.nBitRate = 128000, OMX_SetParameter (omx_handle, OMX_IndexParamAudioMp3, param)
  • [13:45:34] <kulve> khasim: but I plan to get it working :)
  • [13:45:43] <khasim> felipec: any comment on kulve's input?
  • [13:46:04] <AV500> felipec: that's what I feared, I have to have a big switch() for each codec type, to pass something simple and universal as a bitrate
  • [13:46:05] <kulve> felipec could try to reproduce my environment and debug out the reason for me ;)
  • [13:46:48] <felipec> khasim: yeah, the FOSS community doesn't seem to be too interested in OpenMAX, but I don't see any valid reason
  • [13:46:49] <khasim> felipec: my question was towards: once they are available there might be interest (just a guess). I don't see any reason why not.
  • [13:47:00] <khasim> felipec: ok
  • [13:48:00] <khasim> felipec: I believe in kulve's message, there is nothing much available at DSP side to really motivate any one to try for highperformance acceleration, when we make such things available the FOSS community should get built around Open Max
  • [13:49:25] <felipec> AV500: yes, but not all audio encoders support the same kind of bitrate
  • [13:50:41] <felipec> AV500: in the spec for g723 has bit rate as an enum: Unused, Low, High, for example
  • [13:50:48] <khasim> felipec: the openmax concepts have been introduced to community much earlier with maemo and nseries, bit confused on why it didnt get picked up
  • [13:51:39] <khasim> kulve: with ARM releasing lot of free codecs on Neon and promoting open max do you think FOSS community will get interested here atleast?
  • [13:51:59] <khasim> I dont know if it is lot at least few for sure :)
  • [13:52:52] <felipec> khasim: Maemo still hasn't introduced OpenMAX, and ARM is providing OpenMAX DL, not IL, so you can't really use those codecs right away
  • [13:52:57] <AV500> felipec: yes, but projects like ffmpeg still manage to have a largely unified codec API
  • [13:53:09] * BThompson (n=BThompso@nat/ti/x-d584b57d2ad49fc8) has joined #beagle
  • [13:54:00] <AV500> e.g. I would assume *every* video codec would have width/height
  • [13:54:10] <khasim> felipec: do you have a site that quickly points to the difference between DL and IL :(
  • [13:54:43] <AV500> http://www.khronos.org/openmax/
  • [13:54:45] * abitos (n=nixgibts@dyn-k2-211.faveve.uni-stuttgart.de) has joined #beagle
  • [13:54:52] <jkridner|work> anybody now a good way to sample the user button within u-boot?
  • [13:54:55] <felipec> AV500: yes, width and height are generic
  • [13:54:59] <felipec> in omx
  • [13:55:16] <jkridner|work> or, even better, simply detect if the SD/MMC is there?
  • [13:55:27] * AV500 rejoices
  • [13:55:43] <AV500> and a byte has 8 bit, I hope :-)
  • [13:56:23] <felipec> khasim: DL gives you the building blocks, the things to optimize in each platform, IL handles whole elements, encoders, decoders, etc.
  • [13:57:35] * methril is now known as methril|gone
  • [13:59:05] <khasim> felipec: Will I be able use ARM codecs on Neon atleast (with OpenMax dl)?
  • [13:59:49] <felipec> khasim: for example, ideally you would have an mp3decoder in IL, which does the header parsing, and things that pretty much every mp3decoder does again and again, and then in DL you have the core of the algorithm, that is optimized for ARM, x86, and so on
  • [14:00:33] <felipec> khasim: you would need to write the decoder itself, but the NEON optimized pieces are already provided by ARM, in OpenMAX DL
  • [14:02:34] <khasim> felipec: how about the DL interface to gstreamer and ffmpeg, are they still the same?
  • [14:03:21] <sakoman_> jkridner|work: look at what I did in x-load to detect/load from mmc. place to look is start_armboot function.
  • [14:03:23] <sakoman_> http://www.sakoman.net/cgi-bin/gitweb.cgi?p=x-load-omap3.git;a=blob;f=lib/board.c;h=a3b83694345f3f98540f8fa2ec67a315d1a1cfce;hb=65ed00323f3807197a83abc75d62ed2a8d3f60de
  • [14:03:52] <Crofton> sakoman_, gm
  • [14:04:34] <sakoman_> Crofton: good morning!
  • [14:05:49] <Crofton> trying to stay partly on east coast time ...
  • [14:06:01] <felipec> khasim: you are not getting it, DL provide things like omxVCM4P2_IDCT8x8blk, it's a function, you would need to write a GStreamer element that uses that function
  • [14:06:06] <Crofton> no auto first class upgrades on the flight home :(
  • [14:06:10] <sakoman_> Not a bad idea for a short stay
  • [14:06:11] <kulve> sakoman_: I tried linux-omap HEAD kernel on omap3evm with u-boot v2 from linux.omap.com and I got weird colors on X. When I updated to u-boot HEAD from your git, I don't get image at all. Any hints what I should try?
  • [14:06:14] * cringthis (n=cringthi@nat/ti/x-c11bf9da3bbab2de) has joined #beagle
  • [14:06:15] * jsync (n=jess@59.160.172.220) has left #beagle
  • [14:07:59] <sakoman_> kulve: are you using the common branch from my u-boot git?
  • [14:08:04] <kulve> yes
  • [14:08:15] <kulve> the master branch didn't seem to have anything related to omap3
  • [14:08:39] <sakoman_> kulve: yeah, it is just a tracking clone of upstream u-boot git
  • [14:09:08] <sakoman_> I haven't tried omap3evm in quite some time, been swamped with other things
  • [14:09:36] <sakoman_> Do your boot messages look reasonable as far as pixel clock rate, etc. ?
  • [14:09:59] * jsync (n=jess@59.160.172.220) has joined #beagle
  • [14:10:08] <felipec> sakoman_: regarding ALSA, so are you going to send the patches to linux-alsa?
  • [14:10:30] <khasim> felipec/AV500: thanks for your time
  • [14:11:09] <sakoman_> felipec: yes, the plan is that I will resubmit them to linux-alsa now that twl4030 driver is upstream ready
  • [14:11:14] <kulve> sakoman_: <6>omapfb: Pixclock 25411 kHz hfreq 48.9 kHz vfreq 74.4 Hz
  • [14:11:17] <felipec> cool
  • [14:11:50] <sakoman_> felipec: no significant changes from what is in linux-omap right now
  • [14:12:18] <sakoman_> just the boot message patch from David Browneel
  • [14:12:49] <sakoman_> and potentially a one bit change in a register setup (pending some more testing)
  • [14:12:52] <felipec> sakoman_: you mean for overo? because there's no support for beagle in linux-omap right now :)
  • [14:13:07] <koen> sakoman_: there was a patch for lcd for evm uboot on the beagle list a while ago, is that already in, or are you waiting for the patch to get split up?
  • [14:13:11] <sakoman_> I'll submit the beagle patch too :-)
  • [14:13:15] <koen> sakoman_: I'm a bit fuzzy on the details, sorry
  • [14:13:58] * abitos (n=nixgibts@dyn-k2-211.faveve.uni-stuttgart.de) Quit ("Chatzilla 0.9.77 [SeaMonkey 1.1/2007011111]")
  • [14:14:29] <sakoman_> koen: dirk2 and I have agreed to keep the common branch unchanged until we get the patches upstream
  • [14:14:31] <felipec> sakoman_: which is a significant change from linux-omap :)
  • [14:14:37] <sakoman_> we don't want a moving target
  • [14:14:57] <felipec> at least for beagle users
  • [14:15:01] <sakoman_> I don't recall an lcd patch -- do you have a link?
  • [14:15:07] * odesus (n=OMAR@148.223.106.43) has joined #beagle
  • [14:16:20] * sakoman_ really needs to get his autobuilder re-enabled to build beagle and ompa3evm regularly too
  • [14:17:26] <sakoman_> koen: I do recall that a couple patches came in from Mani that had multiple feature changes and we asked him to split it up
  • [14:20:06] <khasim> sakoman_: Any updated on overo availability in market
  • [14:21:32] <sakoman_> khasim: I expect that Gumstix should be announcing availability "Real Soon Now"
  • [14:21:39] * jsync (n=jess@59.160.172.220) has left #beagle
  • [14:23:38] * emeb_mac (n=ericb@ip72-223-84-209.ph.ph.cox.net) has joined #beagle
  • [14:23:44] <sakoman_> koen: linux-omap head is working *so* much better now! Last few weeks have been hell :-(
  • [14:24:58] * felipec agrees
  • [14:26:03] <sakoman_> even musb :-)
  • [14:27:59] * jrmuizel (n=jrmuizel@mozca02.ca.mozilla.com) has joined #beagle
  • [14:28:26] * massiveRobot (n=massiveR@75.149.106.130) Quit (Read error: 60 (Operation timed out))
  • [14:32:30] <felipec> sakoman_: hmm, are you confusing me with Felipe Balbi? or why specifically mention musb?
  • [14:33:16] * massiveRobot (n=massiveR@69.241.19.12) has joined #beagle
  • [14:36:17] * emeb_mac (n=ericb@ip72-223-84-209.ph.ph.cox.net) Quit ()
  • [14:38:31] <koen> sakoman_: I think it was indeed in some mani super-patch
  • [14:43:05] <sakoman_> felipec: the musb comment wasn't directed at you -- it was just a general comment about a previously troublesome area
  • [14:43:38] <khasim> :)
  • [14:47:30] * dirk2 (n=dirk@F31d5.f.strato-dslnet.de) has joined #beagle
  • [14:47:40] <kulve> sakoman_: did you notice the pixclock I pasted above? I don't see anything odd in the boot logs. And looks like X starts up, I just don't see anything on the lcd..
  • [14:49:12] * koen has the same problem as kulve
  • [14:49:15] <koen> (on the evm)
  • [14:49:24] <sakoman_> kulve: looks reasonable to me
  • [14:49:25] <koen> and I want to test zuh's awesome xv driver on it
  • [14:49:29] <kulve> here's the boot log: http://pastebin.com/d667e36b5
  • [14:49:56] <sakoman_> kulve: it used to work fine for me with X, but that was a few months back
  • [14:50:37] <sakoman_> kulve: mani made a few patches to the evm u-boot in the interim
  • [14:50:39] <kulve> sakoman_: it works "ok" with X, but I don't get anything on the LCD. I guess I should see the penguin logo there while the kernel boots?
  • [14:51:13] <sakoman_> kulve: depends on your kernel build options whether you see the penguin
  • [14:51:25] <sakoman_> what kernel are you using?
  • [14:51:26] * BeppoS (n=BeppoS@port-92-202-18-241.dynamic.qsc.de) has joined #beagle
  • [14:51:38] <kulve> HEAD
  • [14:51:42] <kulve> from linux-omap
  • [14:51:44] <BeppoS> Hello
  • [14:51:45] <kulve> CONFIG_LOGO=y
  • [14:51:55] <Spexius> hmm, is CONFIG_FRAMEBUFFER_CONSOLE set?
  • [14:51:58] <sakoman_> yeah, then I would expect to see the logo
  • [14:52:13] <kulve> CONFIG_LOGO_LINUX_MONO, _VGA16 and _CLUT224 are yes also
  • [14:52:17] <koen> sakoman_: I suspect TI changed LCD panels, but I'm not sure
  • [14:52:50] <kulve> Spexius: no, but I don't need the framebuffer console
  • [14:53:09] <sakoman_> koen, kulve: it will be a couple of days before I will get time to try a build on my evm
  • [14:53:16] <Spexius> it will display the logo without a fb console?
  • [14:53:17] <kulve> koen: what revision you have of the evm board?
  • [14:53:27] <kulve> Spexius: I think so, but I'm really not sure
  • [14:54:00] <LoneMeow> IIRC that should work
  • [14:54:15] <LoneMeow> this board I have here has no fb console and shows a logo during boot anyway
  • [14:54:16] <kulve> sakoman_: you don't happen to know which u-boot version you tested when it worked? I could try some older version..
  • [14:55:24] <LoneMeow> hmm, it actually has CONFIG_FRAMEBUFFER_CONSOLE=y but uses serial console, that might have some effect (or not)
  • [14:55:56] <Spexius> well, you should be able to copy some binary data to /dev/fb
  • [14:56:00] <Spexius> and see something
  • [14:56:03] <sakoman_> kulve: don't know exactly. you might look at the history for the omap3evm board file and pick something prior to Mani's patches
  • [14:56:32] <LoneMeow> spexius, assuming the fbdev mode is set to something half reasonable...
  • [14:56:46] <sakoman_> kulve: Spexius has good advice -- see if the hardware is working by copying something to the frame buffer
  • [14:57:24] <LoneMeow> kulve, does fbset report some halfway reasonable mode being active?
  • [14:57:31] <kulve> sakoman_: I do get wrong colors with other kernels/u-boots and I do get proper colors with the preflashed test software so the hardware is ok
  • [14:58:16] <sakoman_> kulve: I meant to try it with the u-boot version where x doesn't work, just to see if you can light up some pixels
  • [14:58:52] * aleph_6510 (n=aleph@trinity.develer.com) has left #beagle
  • [15:00:40] <koen> kulve: rev D iirc
  • [15:00:56] <kulve> ok, I have some older C version..
  • [15:01:14] <kulve> I wonder if newer boards actually works with the latest kernel and u-boot..
  • [15:02:14] * odesus (n=OMAR@148.223.106.43) has left #beagle
  • [15:02:22] <sakoman_> kulve: it could be that the pin mux setup for evm got screwed up in the grand unification
  • [15:02:58] <sakoman_> could be something as simple as the LCD enable gpio not being set up properly
  • [15:03:11] <kulve> it sounds something like that
  • [15:03:13] <kulve> or looks like
  • [15:03:27] <khasim> see you all. good night
  • [15:03:35] <Spexius> sleep well
  • [15:03:37] <koen> does the sdk kernel work on the evm?
  • [15:03:41] <sakoman_> khasim: gn!
  • [15:03:47] <kulve> koen: sdk kernel=?
  • [15:03:50] <koen> if the sdk kernel doesn't work, it's a uboot problem
  • [15:03:55] <koen> kulve: the one in the mistral BSP
  • [15:04:21] <koen> the 2.6.22.ti.evil ine
  • [15:04:51] <kulve> I took the v2 kernel and u-boot from the linux.omap.com and I think the lcd starts up ok with those, but the X doesn't.
  • [15:06:04] <sakoman_> kulve, koen: looks like Mani did submit some pinmux changes, but they were in a massive patch that was a mix of changes that were a mix of omap specific and general changes that should have been submitted upstream directly
  • [15:07:01] <sakoman_> I'll try to make some time in the next day or two to see if I can figure out what is broken
  • [15:08:53] <kulve> sakoman_: thanks! I'm happy to test anything you might come up
  • [15:09:46] <kulve> hmm. I took some version from around Aug 14 but same thing..
  • [15:10:44] <koen> kulve: try sakoman uboot with v2 kernel and see if the lcd turns on
  • [15:11:59] <kulve> I will
  • [15:12:49] * felipec (i=c0647cdb@gateway/web/ajax/mibbit.com/x-ffb8e436dc758d50) Quit ("http://www.mibbit.com ajax IRC Client")
  • [15:13:01] <koen> if it doesn't, it's a uboot pinmux problem
  • [15:13:14] <koen> I need sakomans uboot for sd support
  • [15:13:20] <koen> teh TI one lacks that
  • [15:13:38] * florian (n=fuchs@217.146.132.69) Quit (Remote closed the connection)
  • [15:14:27] <kulve> it doesn't. And if I then copy the v2 u-boot there I get immediately the penguin logo
  • [15:16:41] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
  • [15:17:10] <koen> right
  • [15:17:18] <koen> so a u-booot pinmux problem
  • [15:20:31] <kulve> and if I take the v2 u-boot and the linux-omap HEAD I get weird colors but otherwise it's working
  • [15:20:50] * jkridner|work (n=a0321898@nat/ti/x-760b3660db5b997d) Quit (Remote closed the connection)
  • [15:22:18] <sakoman_> if someone wants to take a crack at fixing this before I get a chance, the right place to look is in include/asm-arm/arch-omap3/mux.h
  • [15:22:50] <sakoman_> I believe that evm board file uses the MUX_DEFAULT_ES2 settings
  • [15:23:33] <sakoman_> check to make sure that the GPIO pin(s) needed for the display are enabled in the proper mode/direction
  • [15:23:37] * BeppoS (n=BeppoS@port-92-202-18-241.dynamic.qsc.de) has left #beagle
  • [15:24:56] * RogerMonk (n=a0740758@nat/ti/x-a045b973beea41f5) has joined #beagle
  • [15:26:38] <dirk2> sakoman_: Yes, Mani has a patch to introduce MUX_DEFAULT_EVM. I'm looking for the link ...
  • [15:27:02] * RogerMonk (n=a0740758@nat/ti/x-a045b973beea41f5) Quit (Remote closed the connection)
  • [15:27:16] <dirk2> http://beagleboard.googlegroups.com/attach/bf1d34a96811876f/u-boot-mux-p1.txt?view=1&part=2
  • [15:28:09] <dirk2> http://groups.google.com/group/beagleboard/msg/bf1d34a96811876f
  • [15:28:12] * trickie|work (n=trickie@a5061.upc-a.chello.nl) Quit ("Leaving")
  • [15:28:38] <sakoman_> dirk2: yeah lots of changes in that patch!
  • [15:29:07] <dirk2> ... just in case somebody likes to have a look to it ;)
  • [15:29:44] <sakoman_> dirk2: My guess is that a one or two line change will fix the issue
  • [15:30:58] * SimonAdameit (n=simon@rlh.mediascape.de) has joined #beagle
  • [15:30:59] <dirk2> Having MUX_DEFAULT_xxx() for each new board isn't a good solution. If anybody has an idea how to do this better, let us know.
  • [15:31:59] <SimonAdameit> Hi, where can I find more about beaglefs, there does not seem to be much on the net in forms of an presentation of it, what it does, how that would look, etc.
  • [15:32:17] <dirk2> sakoman_: Yes, think so, too. Unfortunately at the moment you need a complete new MUX_DEFAULT_xxx() if you only one to change one pin and not influence the other boards.
  • [15:32:29] <kulve> SimonAdameit: beaglefs?
  • [15:32:34] <dirk2> SimonAdameit: Do you now http://elinux.org/BeagleBoard ?
  • [15:32:42] <dirk2> s/now/know/
  • [15:33:26] <and-ri> thanks for all zour help bye
  • [15:33:31] <and-ri> your
  • [15:33:45] <SimonAdameit> kulve: a file system that represents a live beagle query
  • [15:33:49] * and-ri (n=x0103690@nat/ti/x-d1eedad5bf35ec49) Quit ()
  • [15:33:54] <SimonAdameit> dirk2: no, thanks
  • [15:34:14] <suihkulokki> SimonAdameit: see the topic of this channel
  • [15:35:41] <SimonAdameit> Oh
  • [15:35:58] <SimonAdameit> So I seem to be in the wrong channel here :)
  • [15:36:38] * abitos (n=nixgibts@dyn-225-10.rus.uni-stuttgart.de) has joined #beagle
  • [15:38:53] * RogerMon1 (n=a0740758@nat/ti/x-b15f492f63db1b98) Quit (Remote closed the connection)
  • [15:40:46] * Spexius (n=spex@gb-r9-01.pixelpark.net) Quit (Read error: 113 (No route to host))
  • [15:40:50] * SimonAdameit (n=simon@rlh.mediascape.de) has left #beagle
  • [15:44:00] * felipec (i=c0647cdb@gateway/web/ajax/mibbit.com/x-4249d2fb48c25350) has joined #beagle
  • [15:44:38] * vu3rdd (i=79f79106@gateway/web/ajax/mibbit.com/x-272e81f4348dbc51) has joined #beagle
  • [15:48:53] * trickie (n=trickie@hgaulton.xs4all.nl) Quit ("Leaving")
  • [15:49:04] * NishanthMenon (n=nMenon@nat/ti/x-402649a3c6f3ec11) has joined #beagle
  • [15:49:24] * trickie (n=trickie@hgaulton.xs4all.nl) has joined #beagle
  • [15:50:45] * massiveRobot (n=massiveR@69.241.19.12) Quit ("The computer fell asleep")
  • [15:54:33] * ful (i=c2913c85@gateway/web/ajax/mibbit.com/x-ed74ad9f81713e74) Quit ("http://www.mibbit.com ajax IRC Client")
  • [15:57:26] * robclark (n=robclark@nat/ti/x-52b98e00219ae9e1) Quit (kubrick.freenode.net irc.freenode.net)
  • [15:57:26] * keesj (n=keesj@ip49-193-210-87.adsl2.static.versatel.nl) Quit (kubrick.freenode.net irc.freenode.net)
  • [15:58:31] <kulve> koen: who could try to check why the zuh's XV's scaling goes boom on beagle?
  • [16:00:54] * robclark (n=robclark@nat/ti/x-52b98e00219ae9e1) has joined #beagle
  • [16:00:54] * keesj (n=keesj@ip49-193-210-87.adsl2.static.versatel.nl) has joined #beagle
  • [16:01:10] * ful (i=c2913c85@gateway/web/ajax/mibbit.com/x-f5e71549f820b9f9) has joined #beagle
  • [16:10:39] <koen> kulve: I'll queue it up for the weekend
  • [16:12:18] * Geicus (n=afsd@64.132.1.226) has left #beagle
  • [16:12:36] <koen> roger has left the building
  • [16:14:10] <koen> jkridner: link and CE also build for neuros-osd2
  • [16:14:49] <Crofton|road> having paying work is annoying
  • [16:15:07] * jkridner|work (n=a0321898@nat/ti/x-06ff6fb92c0b36ee) has joined #beagle
  • [16:15:09] <Crofton|road> at least when it is unrelated to beagle/davinci/oe ...
  • [16:16:17] <kulve> it sure is :)
  • [16:17:33] <jkridner|work> thanks sakoman, looking now.
  • [16:20:41] * dirk2 (n=dirk@F31d5.f.strato-dslnet.de) has left #beagle
  • [16:20:51] <kulve> does that Tomi Valkeinen's dss patch support sdp only..?
  • [16:38:53] * valhalla (n=valhalla@81-174-21-31.dynamic.ngi.it) has joined #beagle
  • [16:42:37] * [X]Spot (n=stancho@78.90.115.116) Quit (Read error: 113 (No route to host))
  • [16:52:37] * ful (i=c2913c85@gateway/web/ajax/mibbit.com/x-f5e71549f820b9f9) Quit ("http://www.mibbit.com ajax IRC Client")
  • [16:57:11] * massiveRobot (n=massiveR@75.149.106.130) has joined #beagle
  • [16:58:52] * Xenion (n=robert@p579FC133.dip.t-dialin.net) has joined #beagle
  • [17:11:18] * Geicus (n=afsd@64.132.1.226) has joined #beagle
  • [17:14:38] * koza1 (n=subbu@117.192.6.239) has joined #beagle
  • [17:18:53] * chakie (n=chakie@dsl-tkubrasgw1-fe24fa00-230.dhcp.inet.fi) has joined #beagle
  • [17:21:40] * sweetlilmre (i=sweetlil@dsl-244-129-205.telkomadsl.co.za) has joined #beagle
  • [17:24:01] <Crofton|road> koen: next week I'll try and line up some people to use dsplink/CE
  • [17:24:16] <koen> Crofton|road: I hope those people have NDAs
  • [17:24:24] <koen> Crofton|road: since all the stuff in OE isn't public yet
  • [17:24:25] <Crofton|road> heh
  • [17:24:30] <sweetlilmre> Crofton|road: got time for a question or two?
  • [17:24:38] <Crofton|road> that problem can be sovled
  • [17:24:44] <Crofton|road> sweetlilmre: not really
  • [17:25:00] <koen> Crofton|road: Roger hopes to get the patches we need into the public releases
  • [17:25:01] <Crofton|road> nothing in depth
  • [17:25:07] <sweetlilmre> Crofton|road: no worries, thats why I asked :)
  • [17:27:04] * khasim (n=a0393720@192.163.20.231) has joined #beagle
  • [17:27:40] <Crofton|road> koen: do the public version have recipes that work on davinci?
  • [17:28:50] <koen> Crofton|road: the recipes in OE already work on davinci and beagle :)
  • [17:28:53] * _aleph (n=aleph@adsl-ull-140-26.49-151.net24.it) has joined #beagle
  • [17:29:00] <koen> there's now one codec-engine to rule them all :)
  • [17:29:07] * _aleph (n=aleph@adsl-ull-140-26.49-151.net24.it) has left #beagle
  • [17:29:13] <koen> (except 5912osk, but that's being looked at)
  • [17:29:36] <Crofton|road> I'll need to figure this out when I have time
  • [17:30:01] <Crofton|road> we are not using/supporting seperate dsplink
  • [17:30:06] * abitos (n=nixgibts@dyn-225-10.rus.uni-stuttgart.de) Quit (Read error: 110 (Connection timed out))
  • [17:30:13] <ds2> jkridner|work: you might be able to detect an SD card but you need to have I2C working to talk to the TWL4030 and read GPIO0 on there
  • [17:30:34] <koen> Crofton|road: you can still use a seperate dsplink
  • [17:30:52] <koen> Crofton|road: at the external dsplink is identical to the one in the ce tarball
  • [17:30:56] <Crofton|road> the dsplink 1.50 recipe works on davinci?
  • [17:31:07] <Crofton|road> I think that is the public version ....
  • [17:31:30] <ds2> jkridner|work: reading the user button in u-boot should be simple... check the pinmux (think it is setup right); then read the GPIO; it is a memory read
  • [17:31:30] <koen> 1.50 is for davinci
  • [17:31:44] <Crofton|road> yes
  • [17:31:49] <koen> 1.51 is a fork and has davinci removed, but omap added
  • [17:31:53] <koen> 1.60 works on both
  • [17:31:58] <Crofton|road> I'll be talking to some davinci people next week
  • [17:32:26] <Crofton|road> crazy :)
  • [17:32:42] <koen> I had access to a davinci dvevm to test
  • [17:34:59] <Crofton|road> TI should pay you :)
  • [17:36:50] * john|new (n=a0270504@nat/ti/x-c2c58e784837e404) has left #beagle
  • [17:37:26] * gregoiregentil (n=zonbu@adsl-71-135-119-175.dsl.pltn13.pacbell.net) has joined #beagle
  • [17:38:23] * robclark (n=robclark@nat/ti/x-52b98e00219ae9e1) Quit (Remote closed the connection)
  • [17:38:59] * robclark (n=robclark@nat/ti/x-cf3a518b655db062) has joined #beagle
  • [17:42:22] * vu3rdd (i=79f79106@gateway/web/ajax/mibbit.com/x-272e81f4348dbc51) Quit ("http://www.mibbit.com ajax IRC Client")
  • [17:43:21] * aleph_6510 (n=aleph@adsl-ull-140-26.49-151.net24.it) has joined #beagle
  • [17:47:26] * aleph_6510 (n=aleph@adsl-ull-140-26.49-151.net24.it) has left #beagle
  • [17:52:26] * khasim (n=a0393720@192.163.20.231) Quit (Remote closed the connection)
  • [17:53:52] * sweetlilmre (i=sweetlil@dsl-244-129-205.telkomadsl.co.za) has left #beagle
  • [18:00:02] * gregoiregentil (n=zonbu@adsl-71-135-119-175.dsl.pltn13.pacbell.net) has left #beagle
  • [18:12:17] * koza1 (n=subbu@117.192.6.239) Quit (Read error: 104 (Connection reset by peer))
  • [18:16:21] * Zedstar_ (n=user@217.41.237.1) Quit (Read error: 110 (Connection timed out))
  • [18:20:53] * FuL|OUT is now known as fulgas
  • [18:30:51] * kozak (n=subbu@117.192.3.154) has joined #beagle
  • [18:35:48] * Geicus (n=afsd@64.132.1.226) has left #beagle
  • [18:40:57] * nathanm (n=nathan@adsl-69-106-228-95.dsl.pltn13.pacbell.net) Quit (Read error: 113 (No route to host))
  • [18:43:09] * Beagle6 (n=Beagle6@3e6b6820.rev.stofanet.dk) has joined #beagle
  • [18:44:25] <Beagle6> hallo anybody here - Just checking if I'm properly connected
  • [18:44:38] * Beagle6 (n=Beagle6@3e6b6820.rev.stofanet.dk) Quit (Client Quit)
  • [18:45:00] * Beagle6 (n=Beagle9@3e6b6820.rev.stofanet.dk) has joined #beagle
  • [18:46:27] <Crofton|road> Beagle6: I hear you
  • [18:46:28] * Beagle6 (n=Beagle9@3e6b6820.rev.stofanet.dk) Quit (Client Quit)
  • [18:46:40] * Beagle6 (n=Beagle6@3e6b6820.rev.stofanet.dk) has joined #beagle
  • [18:49:06] <Crofton|road> Beagle6: I hear you
  • [18:49:17] <geist> I hear you
  • [18:50:23] * RogerMonk (n=a0740758@trillian.kpn-cc.nl) has joined #beagle
  • [18:51:22] <Beagle6> Great - Thanks - Any idea, how I change my name in this java IRC client, which is the only thing I have currently available on this PC
  • [18:51:32] <koen> RogerMonk: good evening!
  • [18:51:55] <Crofton|road> Beagle6: try typing /nick newnick
  • [18:52:16] <Crofton|road> RogerMonk: thanks for the help with dsplink and CE
  • [18:52:50] * Beagle6 is now known as SSC
  • [18:54:56] * RogerMon1 (n=a0740758@nat/ti/x-7e8fa0e27e18b65d) has joined #beagle
  • [18:55:37] * [X]Spot (n=stancho@78.90.115.116) has joined #beagle
  • [18:55:51] <koen> and a vpn'ed version of RogerMonk appears :)
  • [18:56:52] * SSC is now known as SSC-Denmark
  • [18:57:14] <SSC-Denmark> Great - I managed to change my name :-)
  • [18:58:19] * RogerMon1 (n=a0740758@nat/ti/x-7e8fa0e27e18b65d) Quit (Remote closed the connection)
  • [18:59:28] * RogerMon1 (n=a0740758@nat/ti/x-5f306c68f33b4a14) has joined #beagle
  • [19:00:40] * SSC-Denmark (n=Beagle6@3e6b6820.rev.stofanet.dk) Quit ()
  • [19:05:00] * RogerMon2 (n=a0740758@nat/ti/x-d24fddf3e4cf1bb8) has joined #beagle
  • [19:08:28] * RogerMon1 (n=a0740758@nat/ti/x-5f306c68f33b4a14) Quit (Remote closed the connection)
  • [19:08:38] * RogerMon1 (n=a0740758@nat/ti/x-b468067cb070b41f) has joined #beagle
  • [19:10:25] * RogerMonk (n=a0740758@trillian.kpn-cc.nl) Quit (Read error: 110 (Connection timed out))
  • [19:12:25] * RogerMonk (n=a0740758@nat/ti/x-e7d71c97adb3e518) has joined #beagle
  • [19:16:13] * RogerMon3 (n=a0740758@nat/ti/x-8d86b006f21e02f3) has joined #beagle
  • [19:16:16] * RogerMon1 (n=a0740758@nat/ti/x-b468067cb070b41f) Quit (Remote closed the connection)
  • [19:16:21] * RogerMon3 (n=a0740758@nat/ti/x-8d86b006f21e02f3) Quit (Remote closed the connection)
  • [19:17:31] * RogerMonk (n=a0740758@nat/ti/x-e7d71c97adb3e518) Quit (Remote closed the connection)
  • [19:18:33] * hli (i=chaton@vig91-2-82-232-97-149.fbx.proxad.net) has joined #beagle
  • [19:20:02] * RogerMon2 (n=a0740758@nat/ti/x-d24fddf3e4cf1bb8) Quit (Remote closed the connection)
  • [19:22:15] <mru> lots of rogers today...
  • [19:22:48] <geist> an army of rogers is gearing up for an invasion
  • [19:23:05] <mru> are they jolly?
  • [19:23:22] <geist> we can only hope...
  • [19:29:41] * chelli (n=chelli@p54B86B59.dip.t-dialin.net) has joined #beagle
  • [19:32:03] * felipec (i=c0647cdb@gateway/web/ajax/mibbit.com/x-4249d2fb48c25350) Quit ("http://www.mibbit.com ajax IRC Client")
  • [19:42:44] * kozak (n=subbu@117.192.3.154) Quit (Read error: 110 (Connection timed out))
  • [19:44:55] * kozak (n=subbu@117.192.2.97) has joined #beagle
  • [19:53:02] * garren (n=garren@dsl-243-111-39.telkomadsl.co.za) has joined #beagle
  • [19:53:12] <garren> hi all
  • [19:53:38] * aleph_6510 (n=aleph@adsl-ull-140-26.49-151.net24.it) has joined #beagle
  • [20:01:11] * hex4def6 (n=hex4def6@adsl-68-122-40-18.dsl.pltn13.pacbell.net) Quit ("Leaving")
  • [20:06:07] * kozak (n=subbu@117.192.2.97) Quit (Read error: 110 (Connection timed out))
  • [20:06:33] * hex4def6 (n=hex4def6@adsl-68-122-40-18.dsl.pltn13.pacbell.net) has joined #beagle
  • [20:08:33] <hex4def6> Question: How do I change the pin assignments on the expansion port connector under linux on the beagle? For instance, I want access to UART2
  • [20:11:27] * kozak (n=subbu@117.192.0.220) has joined #beagle
  • [20:33:51] * kozak (n=subbu@117.192.0.220) Quit (Read error: 110 (Connection timed out))
  • [20:34:34] * ssvb (n=ssvb___@a88-112-4-224.elisa-laajakaista.fi) has joined #beagle
  • [20:37:03] * kozak (n=subbu@117.192.1.66) has joined #beagle
  • [20:41:04] <robclark> hex4def6: doesn't the TRM have info about pin-muxing?
  • [20:41:32] <garren> quiet in the chat rooms tonight...
  • [20:42:06] <mru> ssh, you might wake them...
  • [20:44:21] <garren> LOL
  • [20:44:45] <garren> cheers all I'm off
  • [20:44:47] * garren (n=garren@dsl-243-111-39.telkomadsl.co.za) Quit ("Ex-Chat")
  • [20:45:00] <NishanthMenon> robclark, see the datamanual also - that has the ball out to reg name map
  • [20:46:20] <robclark> yeah, omap3530.pdf has the ball<->signal mapping.. there should be a bigger doc that has all the register addresses so you know how to configure the pin-muxing though (I'm just trying to remember where it is)
  • [20:46:33] <robclark> (but fwiw, it was hex4def6 who was curious)
  • [20:52:54] <robclark> ahh, here it is: http://focus.ti.com/lit/ug/spruf98b/spruf98b.pdf
  • [20:54:33] * Beagle9 (n=Beagle9@rescomp-08-107429.Stanford.EDU) has joined #beagle
  • [20:56:34] <Beagle9> Hi everybody, Does anybody know much about the capabilities of the linux distributions that currently run on the Beagle board?
  • [20:57:31] <Beagle9> *like Ubuntu or Angstrom
  • [20:57:32] <prpplague> Beagle9: i don't think there is an official distro for the beagle, however alot of people use OE to generate packages for the beagle
  • [20:58:37] <Beagle9> what are the limitations for these packages?
  • [20:59:03] <Beagle9> I've just been poking around and have seen a few demos that show the board running desktop applications
  • [21:01:04] <Beagle9> Like for instance, would something like Video Lan Client work on a linux distro on the beagle? If not, is it a software limitation or more of a hardware limitation (not enough processing power for uncompressing video)?
  • [21:01:16] * kozak (n=subbu@117.192.1.66) Quit (Read error: 110 (Connection timed out))
  • [21:01:31] <mru> depends on what video you'd like to watch
  • [21:01:53] <mru> the beagle is not as fast as a core2 duo 2GHz...
  • [21:02:04] <Beagle9> definately not.
  • [21:02:26] <Beagle9> but I'm thinking about something small on the order of 640x480' max
  • [21:02:30] <hex4def6> oh, thanks robclark
  • [21:02:38] <mru> that's no match for a beagle
  • [21:02:53] <Beagle9> no match in what sense?
  • [21:03:04] <Beagle9> like the board should be able to handle it?
  • [21:03:08] <mru> you could play two of those videos and still have time over
  • [21:03:22] * kozak (n=subbu@117.192.4.112) has joined #beagle
  • [21:03:40] <Beagle9> if it was streamed off of the SD card right?
  • [21:03:50] <Beagle9> how about this?
  • [21:04:06] <Beagle9> I'm looking for live wireless video streaming solution
  • [21:04:06] <mru> I'm running a test right now with an 854x480 source
  • [21:04:14] <mru> it's cranking out about 50fps
  • [21:04:36] <Beagle9> Is that running in the desktop environment?
  • [21:04:47] <hex4def6> eek -- that's a big document... :)
  • [21:04:58] <mru> that's running nothing but ffmpeg
  • [21:05:12] <Beagle9> gotcha
  • [21:05:17] <mru> it would drop a bit if the video were being displayed...
  • [21:05:55] <Beagle9> um... do you mind if I ask you a quick feasibility question?
  • [21:06:11] <robclark> hex4def6: yeah
  • [21:06:12] <mru> just ask, if I mind I won't answer ;-)
  • [21:07:02] <robclark> hex4def6: ok, have a look at 7.4.4.3
  • [21:07:08] <Beagle9> I plan to have a video server streaming a live feed.... I need this board to be able to act as a client that can display that stream on the screen
  • [21:07:32] <Beagle9> *dvi-d or vga
  • [21:07:52] <mru> vga is not an option
  • [21:08:06] <mru> there's only dvi-d and s-video output
  • [21:08:13] <Beagle9> well... dvi-d is fine... I have converters to go from dvi-d to vga
  • [21:08:22] <mru> really?
  • [21:08:30] <Beagle9> well not 1:1
  • [21:08:35] <mru> such things are big and expensive
  • [21:08:42] <Beagle9> it'll letter box things
  • [21:09:02] <mru> ok, just making sure you weren't looking at a dvi-a to vga adaptor
  • [21:09:07] <Beagle9> no no
  • [21:09:30] <mru> dvi-a is a contradiction in terms
  • [21:09:36] <Beagle9> -d
  • [21:09:40] <Beagle9> typo
  • [21:10:04] <Beagle9> well forget about the vga thing
  • [21:10:13] <Beagle9> that's not important to me
  • [21:10:17] <mru> there are dvi-d to vga converters
  • [21:10:29] <mru> basically a graphics card in a box
  • [21:10:50] <Beagle9> yea, I've been looking at those
  • [21:11:03] <mru> I've never used one
  • [21:11:13] <Beagle9> you'd be surprised, there are some small form factor solutions that are fairly cheap
  • [21:11:43] <mru> a tdms receiver and a few DACs should do the job actually
  • [21:11:46] <Beagle9> but anyways, I'm running away from my main question
  • [21:11:54] <mru> ok, back to it
  • [21:12:25] <Beagle9> do you think it's possible to run a video media client on this board?
  • [21:12:32] <mru> of course it's possible
  • [21:12:36] * Xenion (n=robert@p579FC133.dip.t-dialin.net) Quit ("Verlassend")
  • [21:12:49] <Beagle9> how about in the application layer in linux?
  • [21:13:03] <Beagle9> I'd like to avoid low level dsp programming as much as possible
  • [21:13:18] <mru> you can play video on the beagle right now
  • [21:13:39] <mru> though there's driver work underway that will improve things
  • [21:13:45] <Beagle9> so it has no problems decoding videos on the fly?
  • [21:13:46] <mru> like and Xvideo driver
  • [21:13:50] <mru> *an
  • [21:14:21] <mru> it depends on the size, codec, and bitrate
  • [21:15:27] <Beagle9> I'm thinking of H.264 (640x480 resolution) and 30fps max
  • [21:16:07] <mru> that shouldn't be a problem
  • [21:16:08] <Beagle9> codec and resolution are both things that I can change.
  • [21:16:15] <mru> unless you crank up the bitrate way too high
  • [21:16:37] <Beagle9> no way, I'm trying to keep wireless traffic down
  • [21:16:55] <Beagle9> this is going to be for a small handheld screen
  • [21:17:31] <Beagle9> so bitrate and resolution will be fairly low
  • [21:17:52] <mru> then you shouldn't have too much trouble
  • [21:18:03] <Beagle9> sweet.
  • [21:18:08] <mru> you won't have a *lot* of time left for other processing
  • [21:18:10] <Beagle9> one last question if you don't mind
  • [21:18:17] <mru> what do you plan to use for audio?
  • [21:18:21] <mru> codec etc
  • [21:18:30] <Beagle9> *it doesn't need to do anything else
  • [21:18:46] <Beagle9> audio isn't needed
  • [21:18:53] <mru> ok, even better
  • [21:18:55] <Beagle9> this is a specific application
  • [21:19:17] <mru> what I mean is, there won't be time for any fancy flash animations and such while playing video
  • [21:19:32] <Beagle9> so, the version of linux on this board.... can it run most things that normally run in ubuntu or other linux distros?
  • [21:19:47] <mru> you can run ubuntu on the beagle if you like
  • [21:19:54] <mru> it doesn't come with anything installed
  • [21:20:02] <mru> and there is no official distro
  • [21:20:11] <mru> and no distro officially supports it
  • [21:20:22] * felipec (n=felipec@a88-113-6-167.elisa-laajakaista.fi) has joined #beagle
  • [21:20:34] <Beagle9> true, but there seems to be groups that are working on versions that work on it
  • [21:20:46] <Beagle9> I've been looking at the beagleboard.org website
  • [21:20:50] * chelli (n=chelli@debian/developer/tschmidt) Quit ("IRC is just multiplayer notepad")
  • [21:21:04] <Beagle9> and they've been touting Angstrom as something that comes with the board now
  • [21:21:50] <mru> angsrom is probably the closest you'll get to anything that officially runs on the board
  • [21:22:19] <Beagle9> thanks alot, I think you've sold me on this
  • [21:22:22] <vlad_> I'd go with the mojo stuff myself if you want a distro that you can quickly install stuff on, as opposed to just building an image
  • [21:22:34] <vlad_> angstrom has lots of problems (crasy package managers in particular)
  • [21:22:37] <vlad_> (er, crashy)
  • [21:23:16] <Beagle9> ahh you're talking about handheld.org's mojo ubuntu port?
  • [21:23:28] <vlad_> hm, their armv6el-vfp rebuild looks to be ready
  • [21:23:37] <ds2> to switch in the UART, you need to config the pinmux's in Linux
  • [21:23:39] <vlad_> yeah; i'm running just debian atm, but it's compiled for an ancient arm architecture
  • [21:23:43] * gerrynjr (n=gerrynjr@gentoo/user/gerrynjr) has joined #beagle
  • [21:23:52] <vlad_> I'll probably switch to the mojo hasty armv6el-vfp
  • [21:24:06] <mru> not a good idea on beagle
  • [21:24:15] <Beagle9> and you've installed other applications on that and it works
  • [21:24:21] * gerrynjr (n=gerrynjr@gentoo/user/gerrynjr) has left #beagle
  • [21:24:23] <ds2> they are not defined so you will need to add them then add calls to omap_fg_reg() to set them
  • [21:24:26] <vlad_> though they still have the wrong-arch problem (they use "arm" instead of "armel")
  • [21:24:29] <ds2> once you do that, UART2 should just work
  • [21:24:36] <mru> code optimised for armv6-vfp is far from optimal on cortex-a8
  • [21:24:47] <vlad_> Beagle9: sure, with the debian armel dist you can just apt-get install whatever
  • [21:25:02] <vlad_> mru: sure, but it's bound to be better than armv4 code
  • [21:25:07] <mru> probably
  • [21:25:14] <mru> yes, almost certainly
  • [21:25:20] <vlad_> which is what the debian stuff is :)
  • [21:25:24] <vlad_> it's by no means ideal
  • [21:25:34] <mru> but I'd expect a significant improvement by optimising for cortex-a8
  • [21:25:43] <mru> debian is simply not ideal
  • [21:25:45] <vlad_> and a custom build with angstrom or whatever will give you better perf for sure
  • [21:25:59] <mru> it may be idealistic, but it is not ideal
  • [21:26:05] <vlad_> yeah, I agree
  • [21:26:08] <vlad_> it is easy though
  • [21:26:20] <vlad_> which has some benefits :)
  • [21:26:40] * mru stabs debian "security" team for f*cking up lighttpd yet again
  • [21:26:57] * guillaum1 (n=gl@AMontsouris-153-1-18-204.w86-212.abo.wanadoo.fr) Quit ("Leaving.")
  • [21:27:00] <Beagle9> time & money
  • [21:27:15] * guillaum1 (n=Guillaum@AMontsouris-153-1-18-204.w86-212.abo.wanadoo.fr) has joined #beagle
  • [21:27:19] * kozak (n=subbu@117.192.4.112) Quit (Connection timed out)
  • [21:27:25] <mru> debian is not easy... in fact, it's the only distro I've failed to install on a machine
  • [21:27:40] <mru> to be fair, that was a few years ago
  • [21:28:24] <Beagle9> I know I can just program the DSP in c with JTAG and probably get the video client and udp protocol up and running
  • [21:28:49] <Beagle9> but that's really time consuming and a pain to do
  • [21:29:10] <mru> you don't need jtag
  • [21:29:19] <mru> the dsp runs as a slave to the arm
  • [21:29:43] <mru> you boot the arm first
  • [21:29:49] <Beagle9> but how difficuilt is it really to setup a linux port on this device?
  • [21:30:16] <mru> depends on your prior experience of course
  • [21:30:38] <Beagle9> I've setup a linux desktop and server once, but that was a while ago
  • [21:30:47] <Beagle9> and it was on a simple desktop pc
  • [21:31:10] <Beagle9> I don't know how different and/or difficult it would be on this device
  • [21:31:28] <mru> to boot the beagle, you compile a kernel and filesystem, copy them to an SD card, and power it up
  • [21:31:55] * fulgas is now known as FuL|OUT
  • [21:32:15] <Beagle9> but the settings you have to change to compile the kernel is going to different right?
  • [21:32:47] <mru> yes
  • [21:32:53] <Beagle9> I mean, it can't possibly be that simple
  • [21:33:06] <mru> openembedded has configurations ready to use
  • [21:33:10] * kozak (n=subbu@117.192.4.99) has joined #beagle
  • [21:33:15] <Beagle9> really?
  • [21:33:23] <mru> yes, really
  • [21:33:26] <Beagle9> openembedded.org?
  • [21:33:29] <Beagle9> or .com
  • [21:33:35] <mru> .org
  • [21:33:47] <Beagle9> let me check that out then
  • [21:34:43] <Beagle9> it takes you to a wiki pages?
  • [21:34:45] <hex4def6> look at the page on openembedded on the wiki for how to set up the environment btw.
  • [21:34:59] <hex4def6> I think it's openembeddedandGit or something like that
  • [21:35:34] <hex4def6> http://elinux.org/BeagleBoardAndOpenEmbeddedGit
  • [21:36:59] <Beagle9> nice
  • [21:37:03] <Beagle9> thanks for the link
  • [21:38:10] <Beagle9> alright guys
  • [21:38:25] <Beagle9> thanks for all the help and answering all my questions
  • [21:38:53] <Beagle9> I've got to go, but I'll probably put in a order this friday for a board to mess around with.
  • [21:40:04] <hex4def6> I got one last weekend, and have been learning how to use it; they really are amazing little boards
  • [21:41:16] * Beagle9 (n=Beagle9@rescomp-08-107429.Stanford.EDU) Quit ()
  • [21:43:21] * guillaum1 (n=Guillaum@AMontsouris-153-1-18-204.w86-212.abo.wanadoo.fr) Quit (Read error: 104 (Connection reset by peer))
  • [21:44:37] * guillaum1 (n=gl@AMontsouris-153-1-18-204.w86-212.abo.wanadoo.fr) has joined #beagle
  • [21:55:14] * kozak (n=subbu@117.192.4.99) Quit (Read error: 110 (Connection timed out))
  • [21:57:53] * kozak (n=subbu@117.192.1.164) has joined #beagle
  • [22:02:59] * valhalla (n=valhalla@81-174-21-31.dynamic.ngi.it) Quit ("Leaving")
  • [22:03:34] * jrmuizel (n=jrmuizel@mozca02.ca.mozilla.com) Quit ()
  • [22:10:59] * robclark (n=robclark@nat/ti/x-cf3a518b655db062) Quit ()
  • [22:14:02] * prpplague (n=dave@mail.americanmicrosystems.com) Quit ("Leaving")
  • [22:24:15] * kozak (n=subbu@117.192.1.164) Quit (Read error: 110 (Connection timed out))
  • [22:26:23] * kozak (n=subbu@117.192.0.132) has joined #beagle
  • [22:26:47] * hli (i=chaton@vig91-2-82-232-97-149.fbx.proxad.net) Quit ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
  • [22:27:13] * BThompson (n=BThompso@nat/ti/x-d584b57d2ad49fc8) Quit (Remote closed the connection)
  • [22:45:24] * kozak (n=subbu@117.192.0.132) Quit (Read error: 110 (Connection timed out))
  • [22:46:32] * aleph_6510 (n=aleph@adsl-ull-140-26.49-151.net24.it) has left #beagle
  • [22:47:25] * kozak (n=subbu@117.192.4.168) has joined #beagle
  • [22:54:15] * ssvb (n=ssvb___@a88-112-4-224.elisa-laajakaista.fi) Quit ("Leaving")
  • [22:56:41] <_ZeZu_> is it just me or does the layout of TI www page just suck ?
  • [22:56:50] * _ZeZu_ is now known as zezu
  • [22:57:06] <mru> which page?
  • [22:57:23] <zezu> all of them
  • [22:57:28] <mru> the front page of ti.com sucks about as much as that of any big company
  • [22:57:39] <mru> the product catalogue pages are better
  • [22:58:11] <zezu> now i can't find where i downloaded the spruf* from
  • [22:58:27] <mru> google it ;-)
  • [22:58:32] <zezu> i am :)
  • [22:59:33] <mru> http://focus.ti.com/docs/prod/folders/print/omap3530.html
  • [22:59:40] <mru> three clicks from the front page
  • [22:59:45] <mru> not bad for a company that size
  • [22:59:59] <mru> try finding anything on intel's website
  • [23:00:44] <zezu> intels www site i'm familiar with unfortunately hehe
  • [23:00:50] <zezu> thanks
  • [23:00:59] <mru> there's the pdf: http://www.ti.com/litv/pdf/spruf98b
  • [23:01:26] <mru> whenever I try finding anything from intel I always end up going in circles
  • [23:01:27] <zezu> does ti not provide any code samples for it (even on dev. network) ?
  • [23:01:39] <mru> code samples for what?
  • [23:01:41] <zezu> (other than the beagleboard code on google code)
  • [23:01:47] <zezu> omap3
  • [23:01:53] <mru> arm or dsp?
  • [23:01:57] <zezu> arm
  • [23:02:06] <mru> try arm for samples
  • [23:02:13] <zezu> bootstrapping code etc
  • [23:02:37] <mru> check what u-boot and linux do
  • [23:03:03] <mru> it's not complicated, just tedious
  • [23:03:50] <zezu> yea, powervr access is all through binary drivers as well isn't it ?
  • [23:04:01] <mru> yes
  • [23:04:42] <zezu> for all the windows support, i haven't seen any prebuilt cetk or the like either, only thing i see is BSquare selling a devkit
  • [23:04:55] * cringthis (n=cringthi@nat/ti/x-c11bf9da3bbab2de) Quit (Remote closed the connection)
  • [23:05:44] <zezu> i'm just trying to line up my options, and they are getting slimmer: linux on busybox with many hours of cross compiling :)
  • [23:29:07] * felipec (n=felipec@a88-113-6-167.elisa-laajakaista.fi) Quit ("Leaving")
  • [23:31:31] * felipec (n=felipec@a88-113-6-167.elisa-laajakaista.fi) has joined #beagle
  • [23:31:49] * felipec (n=felipec@a88-113-6-167.elisa-laajakaista.fi) has left #beagle
  • [23:35:42] * jkridner|work (n=a0321898@nat/ti/x-06ff6fb92c0b36ee) Quit (Remote closed the connection)
  • [23:59:32] * BThompson (n=BThompso@cpe-76-185-93-11.tx.res.rr.com) has joined #beagle