#beagle IRC Log on BeagleBoard.org

Join the chat at beagleboard.org/chat

IRC Log for 2009-01-07

Timestamps are in UTC.

[00:00:17] <tharvey> when I build OE for beagleboard something built 'MLO-beagleboard' in my deploy/images/... dir - what builds it and is it only appropriate for MMC?
[00:00:26] <florian> good night
[00:00:47] <tharvey> I see the terms xloader and MLO used perhaps interchangeably? are they the same?
[00:01:30] <raster> recalcati: oh - already got that
[00:01:41] * jrmuizel (n=jrmuizel@mozca02.ca.mozilla.com) Quit ()
[00:01:41] * mib_o7t44o (i=c05e5e69@gateway/web/ajax/mibbit.com/x-307379bf50db6f0a) Quit (Client Quit)
[00:02:07] <raster> in fact i built my own uboot and mlo anx even xl-oader from oe
[00:02:10] <tharvey> ah... I see my issue with the OE built MLO flashed to NAND - I was using 'nand ecc hw' when I needed 'nandecc hw' instead
[00:02:21] <raster> i just need to update my oe tree to build more.. up-to-date stuff
[00:02:24] * florian (n=fuchs@f054164253.adsl.alicedsl.de) Quit ("Verlassend")
[00:02:34] * cem8101 (n=a0271661@nat/ti/x-f9541cab582ace65) has left #beagle
[00:03:13] <jkridner|work> recalati: x-loader is the source for MLO.
[00:03:47] <jkridner|work> look at the x-load*.bb files to figure out where MLO-beagleboard came from.
[00:04:30] <tharvey> ah.. there it is - thx
[00:04:59] <tharvey> so what does MLO stand for?
[00:08:52] * phobbes (n=srussell@user-387ggve.cable.mindspring.com) Quit (Read error: 110 (Connection timed out))
[00:09:07] <Gimp__> MLO = master loader object????
[00:09:19] <Gimp__> or monkey loves oranges
[00:09:54] <tharvey> M (as in Mentor Graphics) Loader? Not clear where Mentor Graphics comes into play with regards to MUSB...
[00:11:40] * jkridner|work (n=a0321898@nat/ti/x-37ab5249d9e0f16a) Quit ("Leaving.")
[00:12:52] * ScriptRipper (n=martin@opensuse/member/MartinMohring) Quit (Remote closed the connection)
[00:14:44] <Gimp__> WOW...on the very last task of OE build....whew
[00:14:49] * aleij_ (n=ad@186-199.77-83.cust.bluewin.ch) has joined #beagle
[00:15:24] * aleij (n=ad@186-199.77-83.cust.bluewin.ch) Quit (Read error: 104 (Connection reset by peer))
[00:19:34] * bazbell (n=a0192809@nat/ti/x-0a57dd4ea56d1dff) has joined #beagle
[00:20:17] <tharvey> koen, when trying to write to jffs2/NAND to populate my filesystem I get same oops in jffs2_write_end even after updating my u-boot - http://www.pastebin.ca/1301361
[00:23:16] * __alanc__ (n=a-campbe@nat/ti/x-3bbde07b5d2df5c3) Quit ()
[00:27:12] <tharvey> is there some issue with linux-omap-2.6.27-r8 and NAND/jffs2?
[00:30:27] <recalcati> but at the boot, first of all the x-loader in nand flash starts, does it?
[00:31:40] * dskinner (i=803713d8@gateway/web/ajax/mibbit.com/x-98a5d1c41cb737e1) Quit ("http://www.mibbit.com ajax IRC Client")
[00:33:58] * zedstar (n=john@fsf/member/zedstar) Quit (Remote closed the connection)
[00:35:10] * bazbell (n=a0192809@nat/ti/x-0a57dd4ea56d1dff) Quit ("Leaving.")
[00:35:45] * NishanthMenon (n=nmenon@nat/ti/x-023f3abe3c56c443) has joined #beagle
[00:35:45] <tharvey> recalcati, yes, I'm booting xloader and uboot from NAND, then from there booting the kernel/fs from my MMC/SD in order to flash to NAND
[00:36:56] * ddompe (n=ddompe@200.122.155.113) Quit ()
[00:37:07] <recalcati> instead I press user button and, after x-loader, I get u-boot from mmc and so on, but MLO on mmc is used?
[00:37:26] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
[00:39:21] <tharvey> recalcati, well... I previously did what you are referring to... booted the MLO and u-boot/kernel/fs from my MMC/SD to test them out - when I flashed MLO/u-boot to NAND I used the images from that SD so they are one and the same
[00:41:18] <recalcati> ok, but the user button change the boot sequence in a hardware way or there another boot loader the read the user button?
[00:42:21] <tharvey> not sure what your asking
[00:42:23] * ceyusa (n=ceyusa@189.163.35.95) Quit (Read error: 110 (Connection timed out))
[00:42:51] <tharvey> boot sequence is function of the BOOTROM
[00:42:51] * dskinner (i=63ba2a07@gateway/web/ajax/mibbit.com/x-754e4f12d86274b1) has joined #beagle
[00:43:05] <recalcati> if I press user button, the cpu starts really reading from mmc?
[00:43:07] <tharvey> for me, either way now, I have the same uboot/MLO in NAND as on SD
[00:43:37] <recalcati> ok, thanks
[00:43:41] <tharvey> if you power on while holding user button the BOOTROM will read from USB->UART->MMC->NAND instead of default NAND->USB->UART->MMC
[00:45:30] <recalcati> perfect, I'll see it in the schematic and in the manual of the cpu
[00:46:18] <recalcati> I'm searching some someone that helps me for usbnet
[00:46:37] <recalcati> I'm stopped
[00:46:57] * fulgas is now known as FuL|OUT
[00:49:01] * Leon_Nardella (n=leon@200-161-14-111.dsl.telesp.net.br) has joined #beagle
[00:51:04] <tharvey> when I try to flash my jffs2 to NAND via uboot I get tons of CRC errs on it when kernel tries to mount it as well
[00:52:14] <tharvey> but I think that is an incompatibility between how uboot writes JFFS2 and how kernel expects it as I see things like: Old JFFS2 bitmask found at 0x001771dc
[00:52:14] <tharvey> You cannot use older JFFS2 filesystems with newer kernels
[00:54:19] <recalcati> you can mount on the pc the jffs2
[00:54:28] <recalcati> similar to -o loop
[00:54:58] <tharvey> there is nothing wrong with my jffs2 - there is something wrong with NAND/jffs2 in uboot or kernel that OE's building
[00:55:27] <recalcati> jffs2 is an image,so u-boot only write it on the flash, for the nor flash, if something change for nand I don't know
[00:55:52] <recalcati> try to mount on pc
[00:56:54] <tharvey> I understand what your saying but regardless uboot has a 'nand write.jffs2' command - so its doing 'something' with the jffs2 - and there is a big fat warning in http://elinux.org/BeagleBoardNAND#Writing_file_system_with_U-Boot about this
[00:57:50] <recalcati> ok, dirty sectors and so on ...
[00:58:02] <recalcati> write it with the kernel
[00:58:09] <recalcati> put on an usbpen
[00:58:22] <tharvey> also you can't mount a jffs2 via loop - its an mtd filesystem - there is a way to do it that involves creating a virtual mtd device on your host though
[00:58:56] <recalcati> yes, I told similar to -o loop, I meant what you are saying.
[00:59:24] <tharvey> I would rather just understand why writing to NAND isn't working as that would apparently be a common thing people are doing according to http://elinux.org/BeagleBoardNAND - I need to use MMC for something else and figured it should be simple to get beagle booting linux from NAND
[00:59:27] <recalcati> in my opinion we need an nfs to work in an easy way
[01:00:08] <tharvey> that requires a network.... :)
[01:00:18] <recalcati> usbnet
[01:00:38] <tharvey> and I haven't yet figured out how to get the only working USB on the beagle (the OTG) to work in host mode to use a root fs off usb mass storage or to use a usbnet
[01:01:17] <tharvey> ah... scratch that... kernel is reading USB mass storage just fine
[01:01:21] <recalcati> I think that you have to use it as usb slave only
[01:02:32] <recalcati> Or, when my usbtoeth interface arrive, maybe in org mode, but I don't know
[01:02:33] <tharvey> no, it can be used as host if you set the ID pin correctly - see http://elinux.org/BeagleBoard
[01:03:14] <recalcati> yes, ID pin is pulled down when you use miniA to miniB connector - and so you are in org mode
[01:03:17] <recalcati> otg mode
[01:03:45] <recalcati> sorry, I wrote org instead of otg
[01:04:33] <recalcati> I normally put rootfs on sdcard, but I'd like to put on nfs
[01:04:47] * Leon_Nardella (n=leon@200-161-14-111.dsl.telesp.net.br) Quit (Remote closed the connection)
[01:05:22] <recalcati> I made some tries, the interface usbnet comes up, on beagleboard and on pc, but then I have e kernel panic
[01:05:28] <tharvey> not 'otg' mode - ID pin tells OTG if it should be a host or a target
[01:06:14] * Sept (n=bakljg@c-98-240-226-129.hsd1.mn.comcast.net) has joined #beagle
[01:06:29] <recalcati> sorry, I always think that OTG means host, but it's not true. OTG means usb not complete ..
[01:06:30] * alecrim (n=alecrim@192.100.104.170) has joined #beagle
[01:09:58] <tharvey> when you say your using usbnet are you talking about connecting a USB host port on a PC to the OTG port on the beagle and using the g_ether USB gadget driver?
[01:09:59] <recalcati> my config is http://www.opensurf.it/config-2.6.27-rc5-OE-beagleboard and the boot, with a nice kernel panic is http://www.opensurf.it/config-2.6.27-rc5-OE-beagleboard_boot.log
[01:10:29] <recalcati> pc is the host, beagleboard the slave
[01:10:40] <recalcati> connected directly
[01:11:51] <recalcati> The pc create the usbnet interface correctly when it sees the slave, but than I have a nice problem on some stack or I don't know
[01:12:43] <recalcati> I have the kernel compiled with all debug stuff, but finding the pb is not easy
[01:13:45] <recalcati> see <4>g_ether gadget: using random self ethernet address
[01:14:16] <recalcati> then usb0: HOST MAC fa:eb:90:83:83:1d
[01:15:05] <recalcati> g_ether gadget: high speed config #1: CDC Ethernet (ECM)
[01:15:09] <tharvey> ah there... I can boot my kernel from NAND and filesystem from USB - at least thats a start
[01:15:43] * dskinner (i=63ba2a07@gateway/web/ajax/mibbit.com/x-754e4f12d86274b1) Quit ("http://www.mibbit.com ajax IRC Client")
[01:16:57] <recalcati> ok, good starting point
[01:17:23] <recalcati> g_ether gadget: ecm_open
[01:19:21] * NishanthMenon (n=nmenon@nat/ti/x-023f3abe3c56c443) Quit ("Leaving")
[01:23:54] * _don_ (n=don@nat/ti/x-fc6696178a90c027) Quit ()
[01:24:54] * recalcati (i=5d904010@gateway/web/ajax/mibbit.com/x-b5dd44e3dc68646e) Quit ("http://www.mibbit.com ajax IRC Client")
[01:25:24] * recalcati (i=5d904010@gateway/web/ajax/mibbit.com/x-159ac1dca83953a6) has joined #beagle
[01:25:34] * mib_l2h0s7 (i=4c11b1e4@gateway/web/ajax/mibbit.com/x-f2f8750cc8d2aac3) has joined #beagle
[01:25:43] <recalcati> bye bye, I go to sleep
[01:26:43] * recalcati (i=5d904010@gateway/web/ajax/mibbit.com/x-159ac1dca83953a6) Quit (Client Quit)
[01:26:43] * mib_c6f8of (i=558823a4@gateway/web/ajax/mibbit.com/x-c83020b4e449654f) Quit ("http://www.mibbit.com ajax IRC Client")
[01:30:29] * mib_l2h0s7 (i=4c11b1e4@gateway/web/ajax/mibbit.com/x-f2f8750cc8d2aac3) Quit (Client Quit)
[01:37:03] * alecrim (n=alecrim@192.100.104.170) Quit (Remote closed the connection)
[01:53:23] * mib_dt5l7x (i=62d4395f@gateway/web/ajax/mibbit.com/x-e7585ea91a3b3f42) has joined #beagle
[01:54:04] * mib_dt5l7x (i=62d4395f@gateway/web/ajax/mibbit.com/x-e7585ea91a3b3f42) Quit (Client Quit)
[02:07:44] * aleij_ (n=ad@186-199.77-83.cust.bluewin.ch) Quit ()
[02:26:35] * mib_yxf822 (i=d3c84ae3@gateway/web/ajax/mibbit.com/x-fcfdc0efdaae5529) has joined #beagle
[02:26:59] * leslie (n=leslie@222.66.141.182) Quit (Remote closed the connection)
[02:34:00] * gmansi (n=gmansi@190.173.76.176) has joined #beagle
[02:41:52] * cringthis (n=cringthi@nat/ti/x-015f830835fdbf82) has joined #beagle
[02:47:10] * ssvb (n=ssvb___@a88-114-221-132.elisa-laajakaista.fi) has joined #beagle
[03:07:36] * dskinner (i=456e1930@gateway/web/ajax/mibbit.com/x-337aa08d41a85839) has joined #beagle
[03:29:10] * JoeyBorn (n=jborn@dsl017-022-247.chi1.dsl.speakeasy.net) Quit ("cta bus tracker-greatest invention known to man")
[04:01:54] * dskinner (i=456e1930@gateway/web/ajax/mibbit.com/x-337aa08d41a85839) Quit ("http://www.mibbit.com ajax IRC Client")
[04:02:40] * hugsig (n=hsi@2001:16d8:ff1a:0:216:3eff:fe5a:59cd) Quit (Read error: 60 (Operation timed out))
[04:05:43] * robtow (n=rtow@nat/montavista/x-53343e5fe6a4fcbf) Quit ("Leaving.")
[04:09:14] * gregoiregentil (n=zonbu@adsl-71-135-114-242.dsl.pltn13.pacbell.net) has left #beagle
[04:11:51] * gregoiregentil (n=zonbu@adsl-71-135-114-242.dsl.pltn13.pacbell.net) has joined #beagle
[04:21:06] * bazbell (n=a0192809@nat/ti/x-ebabff3c5a5fca30) has joined #beagle
[04:21:33] * mib_yxf822 (i=d3c84ae3@gateway/web/ajax/mibbit.com/x-fcfdc0efdaae5529) Quit ("http://www.mibbit.com ajax IRC Client")
[04:22:57] * Wowbagger_ (n=wowbagge@d154-20-130-66.bchsia.telus.net) has joined #beagle
[04:24:16] * mib_9ebogd (i=4c5d8d7e@gateway/web/ajax/mibbit.com/x-030075a3921f48ee) has joined #beagle
[04:29:43] * bazbell (n=a0192809@nat/ti/x-ebabff3c5a5fca30) Quit ("Leaving.")
[04:29:48] * RyoKimball (n=Ryo_Kimb@c-69-247-180-145.hsd1.ms.comcast.net) has joined #beagle
[04:30:29] * soman (n=somnath@61.16.248.242) has joined #beagle
[04:30:45] <RyoKimball> Hey, I'm sending back my Beagle soon (in reserve for the RevC). But before I do... I have a 4 GB SD card and a HDMI/DVI cord. Is there something already compiled I can shove onto the SD and just watch it run once?
[04:33:09] * jsync (n=jess@59.160.172.220) has joined #beagle
[04:33:55] * RyoKimball (n=Ryo_Kimb@c-69-247-180-145.hsd1.ms.comcast.net) has left #beagle
[04:33:58] * dskinner (i=456e1930@gateway/web/ajax/mibbit.com/x-8e6be1ed23f4281f) has joined #beagle
[04:55:31] * rupeshgujare (n=rupesh@59.160.172.220) has joined #beagle
[04:59:11] * emeb (n=ericb@ip72-223-90-212.ph.ph.cox.net) has left #beagle
[05:00:58] * hugsig (n=hsi@2001:16d8:ff1a:0:216:3eff:fe5a:59cd) has joined #beagle
[05:02:03] * emeb_mac (n=ericb@ip72-223-90-212.ph.ph.cox.net) has joined #beagle
[05:07:14] * hugsig (n=hsi@2001:16d8:ff1a:0:216:3eff:fe5a:59cd) Quit (Read error: 60 (Operation timed out))
[05:07:29] * soman (n=somnath@61.16.248.242) Quit (Read error: 113 (No route to host))
[05:07:56] * soman (n=somnath@61.16.248.242) has joined #beagle
[05:11:58] * dskinner (i=456e1930@gateway/web/ajax/mibbit.com/x-8e6be1ed23f4281f) Quit ("http://www.mibbit.com ajax IRC Client")
[05:28:53] * dskinner (i=4bd392a4@gateway/web/ajax/mibbit.com/x-8648c818e0c0ee3e) has joined #beagle
[05:32:37] <dskinner> Any idea as to when the RC3 boards are coming out?
[05:32:41] * Shivdas (n=root@59.160.172.220) has joined #beagle
[05:35:30] * garren (n=garren@dsl-244-208-148.telkomadsl.co.za) has joined #beagle
[05:40:00] * dskinner (i=4bd392a4@gateway/web/ajax/mibbit.com/x-8648c818e0c0ee3e) Quit ("http://www.mibbit.com ajax IRC Client")
[05:51:00] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
[05:52:55] * garren (n=garren@dsl-244-208-148.telkomadsl.co.za) Quit ("Ex-Chat")
[06:02:23] * felipec (n=felipec@189.153.88.243) has joined #beagle
[06:04:36] * Wowbagger_ (n=wowbagge@d154-20-130-66.bchsia.telus.net) Quit ("Leaving")
[06:07:02] * mib_9ebogd (i=4c5d8d7e@gateway/web/ajax/mibbit.com/x-030075a3921f48ee) Quit ("http://www.mibbit.com ajax IRC Client")
[06:50:31] <raster> march?
[06:57:52] * DaQatz (n=db@c-66-30-48-54.hsd1.nh.comcast.net) has joined #beagle
[07:15:37] * cringthis (n=cringthi@nat/ti/x-015f830835fdbf82) has left #beagle
[07:27:52] * garren|work (n=chatzill@mail.dm.co.za) has joined #beagle
[07:28:31] <garren|work> morning all
[07:29:10] * sh1 (i=3d5fc3cd@gateway/web/ajax/mibbit.com/x-b95fa326721e0935) Quit ("http://www.mibbit.com ajax IRC Client")
[07:37:58] * emeb_mac (n=ericb@ip72-223-90-212.ph.ph.cox.net) Quit ()
[07:43:00] * mckoan|away is now known as mckoan
[07:44:11] <mckoan> morning
[08:04:14] <jsync> felipec: ping !!!
[08:04:46] * ssvb (n=ssvb___@a88-114-221-132.elisa-laajakaista.fi) Quit ("Leaving")
[08:07:36] <felipec> jsync: pong!
[08:10:37] <raster> moo
[08:10:42] <jsync> what is the result of running dsp-dummy application ?
[08:12:11] <jsync> I mean, after running the binary "dummy" what do you observe in the console?
[08:13:34] <jsync> felipec: i got something like this filling my screen http://pastebin.com/mc0e6b88
[08:18:33] <felipec> jsync: it depends if you have the debugging on or off, but yeah, that seems to be ok
[08:19:04] <felipec> jsync: it keeps running until you click ctrl-c
[08:20:51] <jsync> felipec: so that means, i am able to achieve arm-dsp IPC , am i right ?
[08:21:03] <felipec> jsync: right
[08:21:23] <felipec> and do DSP development
[08:21:29] <jsync> felipec: thanks for the confirmation.
[08:22:15] <jsync> felipec: i have used the 6.0.16 version of toolchain..
[08:24:06] <felipec> jsync: ah, you're the guy in the mailing list that had issues with other versions of the toolchain?
[08:24:33] <jsync> felipec: yes. ,,, i 'm that guy :-)
[08:25:13] <felipec> jsync: but you tried a 6.0 version without luck, right?
[08:25:39] <jsync> felipec: yes i tried 6.0.21.. it didnt work.
[08:26:34] <felipec> jsync: hm, weird
[08:27:18] <jsync> felipec: one more version is left for me to test.. 6.0.20 :-)
[08:27:22] <felipec> officially you should use some "xdc tools" instead of Makefiles, but I decided to do it the simple way
[08:27:35] <felipec> maybe I missed something
[08:29:23] <jsync> felipec: i prefer makefiles.. easy to understand the process...
[08:30:49] <jsync> I wonder why nobody is doing any open source dsp development !!!
[08:30:58] <felipec> jsync: yeap, but something seems to be wrong for other dsp toolchains... can you mentioned in the ml the exact version you used?
[08:31:01] <felipec> for the record
[08:31:20] <felipec> jsync: because TI makes it very difficult?
[08:31:35] <jsync> i will do that now..
[08:32:01] <felipec> it has improved for omap3, but there's still a long way to go
[08:32:11] <felipec> jsync: cool! what do you have in mind
[08:32:14] <felipec> ?
[08:32:27] <jsync> felipec: but you have already achieved the key process in arm-dsp communication with dsp-dummy
[08:33:20] <jsync> felipec: i am interested in doing some development with this
[08:36:06] * recalcati (i=5d904010@gateway/web/ajax/mibbit.com/x-e20dbe312c99fa53) has joined #beagle
[08:39:36] <felipec> jsync: open source?
[08:39:50] <jsync> felipec: yes ofcourse.
[08:41:59] <jsync> felipec: what can we further do with your sample application ?
[08:43:11] <Xenion> Guten Morgen / Good Morning folks ! :-)
[08:44:37] <felipec> jsync: not much, but I guess you could take any C algorithm, compile it for the DSP and see how it behaves
[08:47:29] * aleij (n=ad@161-230.77-83.cust.bluewin.ch) has joined #beagle
[08:48:32] <jsync> felipec: ok.
[08:49:12] <jsync> felipec: i will keep you posted.
[08:49:24] <felipec> maybe ogg vorbis :)
[08:49:26] <felipec> or speex
[08:49:54] <felipec> I think the speex code already has modification older TI's DSPs
[08:51:06] <kulve> they both have something done for the dsp in omap2
[08:51:47] <jsync> kulve: are they open source ?
[08:53:14] <kulve> yeah
[08:54:13] <jsync> kulve: where can i get some sample code ?
[08:54:18] <kulve> http://tuomas.kulve.fi/blog/2007/08/31/speex-encoder-on-n800s-dsp/
[08:54:23] <kulve> that's something I've done
[08:54:49] <mckoan> recalcati: hi
[08:55:29] <kulve> jsync: about the vorbis dsp stuff, google: "site:lists.maemo.org/pipermail/maemo-developers vorbis dsp"
[08:55:32] <kulve> bbl
[08:55:53] * valhalla (n=valhalla@81-174-21-57.dynamic.ngi.it) has joined #beagle
[08:59:37] <felipec> kulve: did you get it to work? I thought you had some issues
[09:02:25] * ldesnogu_ (n=ldesnogu@ven06-2-82-247-86-183.fbx.proxad.net) has joined #beagle
[09:02:47] * abitos (n=nixgibts@dslb-084-057-136-099.pools.arcor-ip.net) has joined #beagle
[09:08:04] <recalcati> for the third day I'll try to make usbnet working on beagleboard, I'm using this config http://www.opensurf.it/config-2.6.27-rc5-OE-beagleboard and I have this kernel panic http://www.opensurf.it/config-2.6.27-rc5-OE-beagleboard_boot.log
[09:08:40] * aleij (n=ad@161-230.77-83.cust.bluewin.ch) Quit ()
[09:11:23] <recalcati> It is really slow to develop user space without nfs stuff, I can't understand if beagleboard is really used or is better to buy directly evm
[09:12:19] <koen> root-on-sd is just as fast
[09:12:51] <koen> you can use sshfs on your host to manipulate the fs if you need to
[09:13:11] * RobertK (i=3edc0622@gateway/web/ajax/mibbit.com/x-3a68f6ef515ed911) has joined #beagle
[09:13:13] * koen uses packages for everything so everything is reproducable
[09:13:29] <RobertK> goood morning
[09:13:40] <koen> RobertK: moin!
[09:14:00] <recalcati> koen: root-on-sd is good to develop kernel, because you need reboot
[09:14:14] <recalcati> sshfs , what is ^
[09:14:17] <RobertK> recalcati: It seems the musb stuff is getting more and more buggy :-(
[09:15:16] <mckoan> recalcati: I can't reach you at office phone
[09:15:19] <recalcati> RobertK: it should be possible for someone to understand the bug
[09:16:04] <recalcati> mckoan: write me at lamiaposta71@gmail.com, I'll tell my cell, too much snow, I'm at home
[09:16:33] <RobertK> here we had RAIN! At -17??C! Rain. Not snow!
[09:16:51] <mru> RobertK: sounds like fun driving ;-)
[09:18:15] <recalcati> mckoan: you are from bg, aren't you?
[09:18:35] <RobertK> mru: I live 200 meters from work :-)
[09:19:00] <recalcati> RobertK: so you have to work everytime
[09:19:59] <recalcati> see you after, mckoan call me if you are from bg , i send you the mail
[09:20:24] <mckoan> recalcati: I'm Marco :-)
[09:20:48] <mckoan> recalcati: call me at office pls
[09:22:38] <recalcati> ok
[09:24:00] * guillaum1 (n=gl@AMontsouris-153-1-25-222.w86-212.abo.wanadoo.fr) Quit ("Leaving.")
[09:24:19] * nemequ_ (n=nemequ@ip68-111-215-155.sd.sd.cox.net) has joined #beagle
[09:24:33] * guillaum1 (n=GZBA4143@AMontsouris-153-1-25-222.w86-212.abo.wanadoo.fr) has joined #beagle
[09:25:55] * nemequ (n=nemequ@ip68-111-215-155.sd.sd.cox.net) Quit (Read error: 110 (Connection timed out))
[09:27:08] * Xenion (n=robert@p579FCDEA.dip.t-dialin.net) Quit ("Verlassend")
[09:32:23] <RobertK> Are there know problmes with tomba's git at http://www.bat.org/~tomba/git/linux-omap-dss.git/ ?
[09:32:41] <RobertK> I cannot clone the repo. It stops somewhere while downloading...
[09:33:40] * RogerMonk (n=a0740758@nat/ti/x-9fe3daa891b2e395) has joined #beagle
[09:34:02] <RobertK> tomba?
[09:34:22] <tomba> lets see. I haven't changed anything at least.
[09:34:53] <mru> hi RogerMonk
[09:34:56] * lucent (i=lucent@silenceisdefeat.com) has joined #beagle
[09:35:05] <RogerMonk> hi mru - HNY!
[09:35:28] <RogerMonk> mru - made any cookies/brownies recently?
[09:36:09] <mru> no, why?
[09:36:13] <tomba> RobertK, are you sure it isn't just slow? git over http is quite slow in my experience
[09:36:13] <mru> would you like some?
[09:36:39] <RogerMonk> mru - I heard they were famous :)
[09:36:41] <lucent> hey I'm new to Beagle discussion, I have chanced to be in a position to suggest a BeagleBoard as a suitable computing device to play back video advertisements with sound for a movie theater
[09:36:48] * felipec (n=felipec@189.153.88.243) Quit ("Leaving")
[09:36:58] <lucent> is there software already written that would loop video files?
[09:38:04] <lucent> the idea is that theater projectionists could plug in a USB thumbdrive and play video off that all day long, it has to be reliable, and cheaper than the 6-7 DVD players they're already buying each year
[09:38:06] * zedstar (n=john@fsf/member/zedstar) has joined #beagle
[09:38:48] <mru> RogerMonk: I didn't know they were that famous... rumours spread far, apparently
[09:39:25] <mckoan> lucent: you can use mplayer
[09:40:11] <lucent> mckoan: that would be great, am I correct to assume that inserting a USB flash drive and reading files off of it is supported by the hardware?
[09:40:23] <lucent> I'm reading some news about the rev C1 which supports some USB host mode
[09:40:33] <lucent> what is that and why is it different from earlier revisions?
[09:41:53] <tomba> RobertK, I was able to clone it
[09:41:55] <rupeshgujare> How can I request SGX drivers and libraries for OpenGL ES ?
[09:42:23] <adj_> lucent: in short Rev B has broken USB EHCI host implementation and that is fixed in Rev C
[09:42:38] <mckoan> lucent: as far as I tested here it is possible
[09:42:44] <lucent> okay
[09:43:39] <lucent> my situation gets a little complicated in that the theater business owner is a film guy, he just likes his collectables and happened to become a rich guy being the only one in town to show movies
[09:43:44] <lucent> ;)
[09:44:04] <lucent> I need to know "does it work?" before I spend money to develop a black box to sell them on
[09:44:42] <lucent> and it's going to be one less thing to worry about I hope! those DVD players break all the time
[09:45:01] <mckoan> lucent: I'm doing such devices using mini-ITX boards since 2007
[09:45:31] <lucent> mckoan: how reliable are your implementations using mini-ITX tech?
[09:46:27] <mckoan> lucent: I already thought to use a OMAP based device to do that, but I need an eth too so Beagle is not the right choice at the moment
[09:46:43] <mckoan> the system is *very* reliable
[09:46:55] <lucent> ah, I understand
[09:48:34] <lucent> thanks for the feedback on my question :)
[09:48:57] <mckoan> lucent: you're welcome
[09:49:13] <lucent> otherwise I am looking for an off-the-shelf device which plays video from USB sticks
[09:49:25] <lucent> I couldn't find any for less than $300usd
[09:52:58] * mib_tcr87v (i=5a0f152d@gateway/web/ajax/mibbit.com/x-f3c7b40b73d6f2bd) has joined #beagle
[09:58:45] <garren|work> what does this mean for a package if I want to use it: PACKAGE =+ 'lib-jack jack-server jack-examples" ??
[09:58:57] <mckoan> lucent: maybe you simply need a multimedia box, you can find a plenty with google for less than $100
[09:59:35] * florian_kc (n=fuchs@217.146.132.69) has joined #beagle
[10:00:06] * florian_kc is now known as florian
[10:01:24] <lucent> mckoan: yeah, maybe. I'm looking and I don't know which ;)
[10:04:21] <mckoan> lucent: WD TV HD Media Player ( WDAVN00 )
[10:04:38] <mckoan> this is the best and cheaper I've seen in action
[10:04:47] <lucent> thanks, sorry to veer off topic there - the beagle board is very nice
[10:06:58] <RobertK> tomba: its seems to be very slow. At the beginning it seems to be very fast but then after some time it stops. I will try it again after lunch.
[10:08:21] * koen discovers http://bugzilla.icculus.org/show_bug.cgi?id=3721
[10:08:45] <koen> RobertK: all DSS2 patches are already applied to the OE 2.6.28 kernel :)
[10:09:31] <koen> actually, OE is ahead of the git repo since it includes the overlay colourmode fix :)
[10:14:53] * koen looks at http://bugzilla.icculus.org/show_bug.cgi?id=3721
[10:15:18] <koen> ldesnogu_: have you managed to compile a quake3 for gles?
[10:15:24] <RobertK> koen: but in the EO 2.6.28 kernel the musb is buggy. It was only a thought.
[10:16:43] <garren|work> koen: In OE how do I tell it which kernel version to select, I use the default and get 2.6.27?
[10:17:10] <ldesnogu_> koen, no I didn't try
[10:17:35] <RobertK> garrent: you can do a "bitbake -b bb-file" with a kernel bb file located in packages/linux
[10:17:47] <ldesnogu_> koen, you had issues with q3ce compilation?
[10:18:34] <koen> ldesnogu_: yes, the makefile tries to do recursive make on non-existant targets
[10:19:00] <ldesnogu_> hu
[10:19:46] <ldesnogu_> I'm not that surprised, it looks extremely messy (renaming files from .c to .cpp is usually the sign of stupid developer)
[10:20:01] <garren|work> RobertK: do you mind giving an example for changing kernels for the beagleboard?
[10:20:13] <koen> make BuildDebug_Demo_ARM/q3pvr_ARM MODE=BuildDebug_Demo_ARM
[10:20:14] <koen> make[1]: Entering directory `/OE/angstrom-dev/work/armv7a-angstrom-linux-gnueabi/q3ce-1.34+svnr4-r0/trunk/proj_kdevelop_pvr'
[10:20:17] <koen> make[1]: *** No rule to make target `BuildDebug_Demo_ARM/q3pvr_ARM'. Stop.
[10:20:24] <koen> like that :)
[10:21:26] <RobertK> garrent: ~/oe/openembedded/packages/linux$ bitbake -b linux-omap_git.bb
[10:21:42] <koen> or 'bitbake linux-omap-2.6.28'
[10:22:13] <RobertK> garrent: then you get the uImage in ~/oe/tmp/deploy/glibc/images/beagleboard
[10:23:22] <garren|work> oh so easy thanks guys
[10:24:41] <RobertK> Why does pressing middle mouse button on windows does not paste? Using vmware on windows is so hard and confusing...
[10:25:10] <koen> DJWillis: the bitbake 1.8.12 release has a fix git fetcher
[10:26:21] <DJWillis> koen: you seem to monitor too many chans ;-).
[10:29:01] <koen> DJWillis: I skim the agstrom logs once a day to see what nonsense people say there
[10:29:27] <DJWillis> koen: not a lot I notice, not what you would call active ;-).
[10:29:32] <koen> "rm /sbin/* to make it work"
[10:30:22] <koen> DJWillis: I stopped being in #angstrom when clueless newbies started coming that didn't want to take my advice
[10:30:49] <koen> if you're asking for advice why not take it?
[10:32:08] <DJWillis> koen: I know that one, I felt I should idle there but no one seems to want to help, I guess as I am picking it up I am happy to help a little (until it gets old). ;-)
[10:45:59] * philippe (n=fille@a83-245-252-47.elisa-laajakaista.fi) Quit (Read error: 60 (Operation timed out))
[10:53:25] * ScriptRipper (n=martin@85.233.34.53.dynamic.cablesurf.de) has joined #beagle
[10:58:30] <koen> the sgx kernel module builds against 2.6.28 as well
[10:58:48] <koen> but omaplfb and bc_example doesn't like DSS2, as expected
[11:03:40] * garren|work (n=chatzill@mail.dm.co.za) Quit (Read error: 104 (Connection reset by peer))
[11:11:48] * TehUni (n=cory@c-24-30-35-55.hsd1.ga.comcast.net) Quit (Read error: 110 (Connection timed out))
[11:12:05] * TehUni (n=cory@c-24-30-35-55.hsd1.ga.comcast.net) has joined #beagle
[11:16:21] * geckosenator (n=sean@ppp-69-218-241-177.dsl.dytnoh.ameritech.net) Quit (Read error: 110 (Connection timed out))
[11:20:46] <recalcati> koen: very nice, as soon as possible I'd like to show an opengl demo to my bosses
[11:28:01] * uwe_ (n=uwe_@dslb-084-056-008-063.pools.arcor-ip.net) Quit (kubrick.freenode.net irc.freenode.net)
[11:28:01] * tharvey (n=tharvey@adsl-76-205-222-173.dsl.snlo01.sbcglobal.net) Quit (kubrick.freenode.net irc.freenode.net)
[11:28:01] * dcramer (n=davec@dcdsl.ebox.com) Quit (kubrick.freenode.net irc.freenode.net)
[11:28:01] * sakoman_ (n=sakoman@static-74-41-60-154.dsl1.pco.ca.frontiernet.net) Quit (kubrick.freenode.net irc.freenode.net)
[11:28:01] * ldesnogu (n=ldesnogu@fw-tnat.cambridge.arm.com) Quit (kubrick.freenode.net irc.freenode.net)
[11:29:25] * uwe_ (n=uwe_@dslb-084-056-008-063.pools.arcor-ip.net) has joined #beagle
[11:29:25] * tharvey (n=tharvey@adsl-76-205-222-173.dsl.snlo01.sbcglobal.net) has joined #beagle
[11:29:25] * dcramer (n=davec@dcdsl.ebox.com) has joined #beagle
[11:29:25] * sakoman_ (n=sakoman@static-74-41-60-154.dsl1.pco.ca.frontiernet.net) has joined #beagle
[11:29:25] * ldesnogu (n=ldesnogu@fw-tnat.cambridge.arm.com) has joined #beagle
[11:29:32] * mckoan is now known as mckoan|away
[11:30:17] <florian> good morning
[11:31:26] <florian> Is it possible to use the svideo port for a second framebuffer?
[11:33:17] <RobertK> florain: yes, with DSS2
[11:33:36] <RobertK> ups, florian not florain
[11:34:50] <florian> RobertK: thanks
[11:34:54] * florian needs to try this
[11:35:47] * abitos (n=nixgibts@dslb-084-057-136-099.pools.arcor-ip.net) Quit (Read error: 60 (Operation timed out))
[11:35:59] <RobertK> florian: yupp. The patches are in the OE kernel.
[11:36:50] <RobertK> florian: then you can have different displays on DVI and s-video.
[11:38:15] <florian> I have to find some piece of hardware that accepts svideo input now ;)
[11:41:24] <raster> man i have to say.. omap3 is impressive
[11:41:42] <raster> its within striking distance in performance of this via c3 @ 1ghz
[11:41:50] <raster> and thats without neon optimisations
[11:42:09] <raster> and half the clockrate
[11:42:18] * koen suspects that's mostly due to cortex-a8
[11:42:21] <raster> and slower ram since its a bb rev-b
[11:42:27] <raster> koen: true :)
[11:42:35] <koen> I wonder what a dual core a9 at 1GHz would be like
[11:42:37] <raster> but impressive
[11:42:44] <raster> as theres a few factors of power usage difference
[11:42:45] <raster> :)
[11:42:58] <raster> a9 would be.... nice
[11:43:14] <raster> i already have smp rendering in evas
[11:43:29] <raster> but i dont think i really saturate the cores as well as i could
[11:43:30] <RobertK> florian: if you find a nice device please send a link. I searched through amazon and did not find a nice, little tv with s-video in.
[11:44:47] <raster> no matter how u look at it.. impressive
[11:45:16] <raster> i need to get started on my neon fun
[11:46:00] <koen> neon is loosely coupled as well, so it can execute in parallel with the arm pipeline
[11:46:17] <raster> thats true
[11:46:45] <raster> so unless i'm saturating the mem bus - i could do evil stuff like somr arm ops betwene neon ops
[11:46:49] <raster> all doing the same pipeline work
[11:46:57] <raster> and basically parallel process
[11:47:18] <koen> you need to take care when doing that
[11:47:28] <koen> touching registers can cause a neon stall
[11:47:39] <koen> so you loose another 20 cycles or so
[11:47:49] <raster> i know that neon will stall internally
[11:47:50] * koen tries to recall the details of what mru said
[11:48:11] <florian> RobertK: there are some tft dispalys with svideo input
[11:48:11] <raster> ie write to neon reg then instantly read from it == stall
[11:48:36] <raster> so you want to round-robin thru the neon regs to avoid it
[11:48:58] <raster> ie write to d0, d1, d2, d3, d4, d5, d6, d7
[11:49:04] <raster> then read d0, d1, d2, d3, d4, d5, d6, d7
[11:49:06] <raster> etc.
[11:49:12] <raster> (extend as appropriate)
[11:50:16] <raster> if you keep your code paths separate
[11:50:45] <raster> ie arm stuff does things unrelated to or not directly/often pulling/pushing data to/from neon... i can see it being ok
[11:51:24] <recalcati> koen: sshfs you mean to have usbtoeth interface and using beagleboard usb as host ?
[11:51:38] <RobertK> florian: they are too big. A little TV with s-dieo would be nicer...
[11:52:25] <koen> recalcati: yes
[11:52:45] <raster> anyone know a good lcd wholesaler who can get samples?
[11:52:55] <recalcati> thx
[11:54:19] * FuL|OUT is now known as fulgas
[11:56:49] <florian> RobertK: Like this? http://cgi.ebay.de/Touchscreen-10-4-Zoll-TFT-LCD-Monitor-VGA-AV-Schwarz_W0QQitemZ370140031621QQcmdZViewItemQQptZAuto_TFT_Monitore?hash=item370140031621&_trksid=p3286.c0.m14&_trkparms=72%3A1239|66%3A2|65%3A12|39%3A1|240%3A1318
[12:00:16] <koen> raster: when I enable the 'flame' or 'rain' module everything slows to a crawl
[12:00:28] <koen> raster: any idea why those would be taking so much cpu?
[12:02:13] <raster> koen: flame is expensive
[12:02:35] <RobertK> florian: Nice :-)
[12:02:36] <raster> it literally is calculating all the pixels int he flame in an array based off the surrounding pvalues
[12:02:38] <raster> converting to argb
[12:02:40] <raster> then blending
[12:02:54] <raster> its an expensive thing
[12:03:11] <raster> rain goes and creates tonnes of little objects - blending them and moving them
[12:09:23] <koen> raster: xflame and xripple needed a lot less resources :)
[12:09:28] <koen> way back when
[12:11:16] * Openfree (n=df@218.82.191.188) has joined #beagle
[12:13:00] <raster> xflame missed the flame->argb+composte then argb->rgb565 + dither bit
[12:13:11] <raster> it went straight ffrom flame array -> xshm segment
[12:13:24] <raster> xripplie was a blitter trick
[12:13:54] <raster> ie copy line from above the "horizon" to below with an x or y offset based on sin() + time
[12:14:08] <raster> (and perspective adjustment)
[12:16:28] <koen> RogerMonk: I signed up for the boxee alphee, let's see if it crosscompiles :)
[12:16:37] <kulve> koen: tomba: does the DSS2 support rotate? i.e. what video=omapfb:rotate:270 did on DSS1.
[12:17:56] <tomba> kulve: no
[12:18:25] <koen> patches accepted :)
[12:18:51] <kulve> can I switch from internal lcd to DVI (on the fly)?
[12:19:04] <koen> yes(*)
[12:19:58] <tomba> kulve: I don't know what the old driver did (I don't remember there even being rotation), but it is a bit problematic
[12:20:10] <koen> kulve: http://pastebin.com/d42e55753
[12:20:16] <tomba> so I haven't used to for that yet, I don't quite know how it could be done
[12:21:00] <kulve> tomba: on the old dss, the video argument rotated the framebuffer so that the X was 640x480 instead of the 480x640 it's now on the evm
[12:21:11] * dcordes (n=dcordes@unaffiliated/dcordes) has joined #beagle
[12:22:17] <tomba> kulve: heh well that is quite obvious. I meant how it does it under the hood. DMA rotation is not good for SDRAM framebuffers, and VRFB rotation has nasty problem with the virtual size of the framebuffer
[12:22:55] <kulve> koen: #!/bin.sh ?
[12:23:08] <koen> that's a typo
[12:23:26] <kulve> tomba: I try not to look under the hood usually ;)
[12:24:06] <tomba> goddamn, my ubuntu is somehow messed up. it just broke without me doing anything. it looks like it doesn't answer to arp queries from the omap board via usb anymore, until I disconnect & connect the usb cable
[12:27:34] <kulve> koen: it did switch to DVI but with odd colors and odd resolutions
[12:27:35] <kulve> -s
[12:28:10] <koen> kulve: yes, I've been trying to debug that with tomba this week :)
[12:28:27] <koen> kulve: linux text console works, albeit with a blue background
[12:28:54] <kulve> hmm.. I booted the board and it does give the image now to both (lcd+dvi) but the dvi is very messy
[12:30:04] * mib_35n8z9 (i=5118136a@gateway/web/ajax/mibbit.com/x-952830644b23b773) has joined #beagle
[12:31:53] <tomba> both at the same time?
[12:32:01] <koen> kulve: you could ask vaibhav for patches
[12:32:45] <tomba> do the displays work if you don't switch them dynamically? if the DVI is the primary display?
[12:33:30] <kulve> tomba: yeah, both at the same time. I see the cursor moving on both. The internal-lcd looks ok
[12:33:43] <kulve> and in that case I didn't switch anything
[12:33:51] <kulve> how do I set the DVI as the primary?
[12:33:53] <tomba> uhhuh. sounds interesting, I wonder how the displays are connected
[12:34:07] <kulve> I have quite old board..
[12:34:12] <koen> I thought that sort of cloning wasn't supported
[12:34:17] * soman (n=somnath@61.16.248.242) Quit (Read error: 60 (Operation timed out))
[12:34:19] <tomba> it gets quite nasty if you display the same image on both displays, as they probably want totally different timings
[12:34:25] <kulve> koen: well, it's a "messy cloning" ;)
[12:34:53] <kulve> tomba: I haven't configured them in any way.. I'm just using the kernel from OE with OE default config
[12:34:55] <tomba> kulve, with the current public version you'll have to edit the board file, and put the dvi display before the lcd display
[12:35:48] <kulve> board file in kernel?
[12:35:48] <tomba> kulve: check omap3_evm_dss_data struct. switch omap3_evm_display_data and omap3_evm_display_data_dvi
[12:35:52] <tomba> yes
[12:36:24] <tomba> but, I'm just in the process of publishing new version of DSS. it supports command line parameters, so you can define the primary display
[12:36:36] <tomba> so you may also choose to wait =)
[12:36:44] * CarlMle (i=d5d20d1e@gateway/web/ajax/mibbit.com/x-3e67edfc76e4b379) has joined #beagle
[12:37:03] <koen> DSS3?
[12:37:24] * ddompe (n=ddompe@200.122.155.113) has joined #beagle
[12:37:57] <CarlMle> Hi, Ive just got my beagle running Angstrom, I wanted to demo the video abilitys. Basically repeat the demos on you tube.. i tried mplayer with some 480 and 720 video and its very jerky. which media player is best for beagle?
[12:37:58] <tomba> let's still call it dss2 =)
[12:38:06] <kulve> tomba: any schedule for that..?
[12:38:20] <tomba> kulve: an hour or two
[12:38:43] <kulve> tomba: oh
[12:38:54] <koen> CarlMle: try -vo xv
[12:39:07] <tomba> looking at the EVM board file, it should not be possible to enable both dvi and lcd at the same time...
[12:39:08] * koen git fetch -v tomi-dss
[12:39:47] <tomba> of course the version will probably bug like hell, but hey, that's why you guys are here, right!?
[12:39:47] <kulve> tomba: is a new set of patches or can I apply just a single patch to what OE has now?
[12:40:06] <kulve> tomba: actually, atm I'm trying to find something that would be reasonable stable :)
[12:40:09] <tomba> kulve: new set of patches. quite a lot of changes
[12:40:17] <kulve> ok
[12:40:20] <tomba> kulve: well, don't look at DSS2 if you want stable =)
[12:40:49] * soman (n=somnath@61.16.248.242) has joined #beagle
[12:40:55] <kulve> tomba: I said "reasonable stable" :)
[12:41:05] <kulve> I don't need anything product quality :)
[12:41:31] <tomba> DSS2 is very stable as long as you don't touch any settings...
[12:42:48] <CarlMle> koen: Thanks, just tried that but mplayer says theres no Xvideo support for the video card
[12:43:07] <koen> tomba: are you going to rebase and squash commmits?
[12:43:18] <koen> CarlMle: sounds like you need to update the angstrom demo
[12:43:40] <koen> http://www.angstrom-distribution.org/demo/beagleboard/Angstrom-Beagleboard-demo-image-glibc-ipk-2009.X-test-20090104-beagleboard.rootfs.tar.bz2 has XV support
[12:43:50] <koen> you do need to follow the readme.txt, though
[12:44:24] <tomba> koen: yes. I keep doing that to hide all the crap commits, and to make it easier to review the patch set. also because otherwise I could get "secret" lines to public patches
[12:45:17] <koen> tomba: it makes extracting patches easier for me :)
[12:45:22] <jkridner> good morning all.
[12:45:28] <koen> hey jkridner
[12:45:29] <florian> hi jkridner
[12:45:45] <koen> jkridner: 13:33 < tomba> but, I'm just in the process of publishing new version of DSS. it supports command line parameters, so you can define the primary display
[12:45:59] <ddompe> morning jkridner
[12:46:14] <jkridner> ah, good.
[12:46:29] <CarlMle> koen: Yeah thats the demo I'm using, i'll re-enter the setenv commands to see if that helps
[12:46:47] <CarlMle> koen: from the readme.txt
[12:47:31] * garren|work (n=chatzill@mail.dm.co.za) has joined #beagle
[12:52:09] <jkridner> ah, looks like some musb fixes went out on linux-usb today.
[12:52:51] <tomba> with the new DSS you have boot args like "omap-dss.def_disp=dvi omapfb.video_mode=800x600MR-24@60"
[12:53:52] <tomba> and also changing the display mode dynamically works, at least in some cases.
[12:54:42] <CarlMle> koen: cool, i get video now, but still as jerky, im playing it off a usb stick, i guess it might be better to play it off the SD card?
[12:55:07] <dcordes> koen: ping
[12:55:21] <tomba> git fetch over http:// is unbelievably slow. I've been updating my linux-omap remote for 15min now. or perhaps it's the nokia web proxy.
[12:56:11] <jkridner> What does "MR" mean?
[12:57:18] <dcordes> koen: could you? http://linuxtogo.org/~lgorris/diffs/msm-update-010709.diff
[12:57:20] <tomba> R was reduced timing. M... hmm. well, they are from drivers/video/modedb.c, you can check from there
[12:57:23] <kulve> tomba: hmm.. I got the DVI working now as the primary, but the colors are quite red..?
[12:58:17] <tomba> kulve: that's most likely a misconfiguration problem. unfortunately I don't have evm board.
[12:58:35] <kulve> (==) FBDEV(0): Depth 24, (==) framebuffer bpp 32
[12:58:36] <kulve> (==) FBDEV(0): RGB weight 888
[12:58:42] <kulve> probably because of that..
[12:58:49] <koen> kulve: probably the same DAC problem with ES revisions we had with the LCD
[12:59:24] <koen> kulve: now you have a 32bpp dislpay you can fix xf86-video-omapfb to work with 32bpp and 16bpp displays :)
[12:59:29] <tomba> kulve: always try without X first =). otherwise there are too many components that can be wrong
[12:59:57] <raster> oooh 32bpp is working now?
[13:01:03] <recalcati> usbnet: with 2.6.28 is up, ping from beagleboard to pc works, now I try ssh and after nfsroot
[13:01:05] <koen> raster: it is with DSS2, you need a small patch to the X driver
[13:01:09] <tomba> if somebody has EVM schematics or TRM, it'd be interesting to see what all those gpios mean that are related to the display
[13:01:40] <kulve> koen: which x driver..?
[13:01:50] <koen> xf86-video-omapfb
[13:02:03] <recalcati> using bitbake linux-omap-2.6.28
[13:02:47] <kulve> I'm now running fbdev, I guess I need to build a new rootfs with xf86-video-omapfb
[13:02:49] * soman (n=somnath@61.16.248.242) Quit (Remote closed the connection)
[13:03:02] <raster> koen: werd. it will soon be time to.. update my oe tree
[13:03:20] <raster> i know there is a need for 16bpp
[13:03:31] <raster> but downconveritn gfrom 32 to 16bpp either generated ugly output
[13:03:38] <raster> or means u sink cycles into dithering it
[13:03:51] <raster> and thats just all not happy-lucky-fun-times
[13:03:53] * __alanc__ (n=a-campbe@nat/ti/x-a269046b809dd1c9) has joined #beagle
[13:04:31] * gcohler (n=gcohler@nat/ti/x-bfe7edc5bc610b8f) has joined #beagle
[13:04:45] <koen> raster: 32bpp takes a lot of memory bw
[13:04:55] <raster> 2x as much
[13:05:04] * chomb (i=4dc716a9@gateway/web/ajax/mibbit.com/x-8ce1635f27eaac10) has joined #beagle
[13:05:08] <raster> it depends on your target
[13:05:11] <raster> if u want high-res
[13:05:27] <raster> like xga+ - sure. its quite significant an increase in total mb
[13:05:34] <raster> if u are down at sub-vga land or so
[13:05:42] <raster> its more of a "meh" amount
[13:05:51] <tomba> you can use 24bpp if you really want to conserve memory
[13:06:02] <raster> nah
[13:06:07] * garren|work (n=chatzill@mail.dm.co.za) Quit ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
[13:06:14] <raster> that still means i need to munge 32->24
[13:06:21] <raster> as all the rendering is in 32
[13:06:22] <koen> tomba: are you going to ask tony to add your dss2 tree to l-o git for testing?
[13:06:35] <raster> as there is always an alpha channel - used or not
[13:06:36] <tomba> koen: at some point, yes
[13:08:34] <chomb> hi there, I am trying to play video with the tiopenmax stuff and getting a DSP error (8000800c) - can s/o help?
[13:10:20] * garren|work (n=chatzill@mail.dm.co.za) has joined #beagle
[13:12:13] * chombourger (n=mvista-f@169.22.199-77.rev.gaoland.net) has joined #beagle
[13:12:29] * mib_2hyom6 (i=5198b1fe@gateway/web/ajax/mibbit.com/x-57dffa0038d8c831) has joined #beagle
[13:13:08] <__alanc__> can u print the whole error - sounds like a dsplink err - can then check http://tiexpressdsp.com/wiki/index.php?title=Troubleshooting_DSPLink_configuration_issues
[13:13:18] * guillaum1 (n=GZBA4143@AMontsouris-153-1-25-222.w86-212.abo.wanadoo.fr) Quit (Remote closed the connection)
[13:14:19] * guillaum1 (n=GZBA4143@AMontsouris-153-1-25-222.w86-212.abo.wanadoo.fr) has joined #beagle
[13:14:30] <chomb> __alanc__: gst-launch-0.10 filesrc location="foobar.avi" ! avidemux ! omx_mpeg4dec ! omap fbsink Setting pipeline to PAUSED ... Pipeline is PREROLLING ... #### src/lcml/LCML_DspCodec.c:621 DSPNode_Create=8000800c ****************LCML ERROR : DSP ************************ Error: Create the Node : Err Num = 8000800c ****************LCML ERROR : DSP ************************
[13:14:48] <chomb> __alanc__: will now check your link
[13:14:52] * mrc3 (n=ddiaz@189.157.115.100) Quit (Read error: 104 (Connection reset by peer))
[13:15:04] <chomb> __alanc__: note that the ping app works
[13:15:05] * mrc3 (n=ddiaz@189.157.115.100) has joined #beagle
[13:18:28] <CarlMle> ahh okay use the -nosound in mplayer and i get smoooth playback
[13:19:13] * atin_OUT (n=atin@c-71-232-7-71.hsd1.ma.comcast.net) has joined #beagle
[13:19:18] * atin_OUT is now known as atin_
[13:20:46] <recalcati> CarlMle: developers told me that mplayer is not optimized, instead omapfbplay it is
[13:22:12] <__alanc__> chomb : DSPLink errbase.h shows for that err.../* A memory allocation failure occurred. */
[13:22:12] <__alanc__> #define DSP_EMEMORY (DSP_EBASE + 0xCl)
[13:23:00] <CarlMle> recalcati: yeah i tried omapbfplay i cant get it to play, im doing 'omapfbplay /myvideo.mov'
[13:23:05] <__alanc__> chomb : ignore the confusing-to-the-eye "l" at end of the def
[13:23:47] <chomb> __alanc__: that's what I found indeed - don't know how I can work around this error
[13:23:58] <CarlMle> it says mmap invalid argument
[13:24:14] <chomb> __alanc__: I have followed steps described at http://elinux.org/BeagleBoard/gst-openmax
[13:25:05] * Openfree (n=df@218.82.191.188) Quit (Connection timed out)
[13:25:50] <tomba> koen, kulve: I updated my git, but only the test branch for now. so if you have time to test that, please do =)
[13:27:11] <tomba> the tree is based on today's linux-omap
[13:28:14] * mckoan|away is now known as mckoan
[13:28:52] * jsync (n=jess@59.160.172.220) has left #beagle
[13:29:04] <tomba> koen: debugging prints have also changed a bit. so I suggest you enable the debug support kconfig option for both DSS and omapfb by default. that way the code is there, and you can enable the actual printing with "debug" module parameter
[13:29:38] <tomba> so no more "could you enable debugging in the kernel config" =)
[13:30:01] * docelic_ (n=docelic@78.134.194.50) Quit (Read error: 60 (Operation timed out))
[13:31:13] * MT (i=86630288@gateway/web/ajax/mibbit.com/x-ef5c7fa867aabb69) has joined #beagle
[13:31:50] * ceyusa (n=ceyusa@189.163.35.95) has joined #beagle
[13:35:10] <recalcati> CarlMle: first I try nfs over usbnet and than omapfbplay
[13:35:34] * mib_35n8z9 (i=5118136a@gateway/web/ajax/mibbit.com/x-952830644b23b773) Quit ("http://www.mibbit.com ajax IRC Client")
[13:35:40] * docelic_ (n=docelic@78.134.197.242) has joined #beagle
[13:35:41] * mib_nsj92o (i=5056380b@gateway/web/ajax/mibbit.com/x-fcd916b01b838d61) has joined #beagle
[13:36:26] * CarlMle2 (i=d5d20d1e@gateway/web/ajax/mibbit.com/x-5d4486affc4e00fb) has joined #beagle
[13:40:53] * abeard (i=c6f6abc8@gateway/web/ajax/mibbit.com/x-2dd2fddef1171987) has joined #beagle
[13:42:22] <ScriptRipper> jkridner: are you around already?
[13:42:39] <jkridner> yeah. on a teleconference.
[13:45:52] * kevinsc_ (n=kevinsc@nat/ti/x-6c10fb91b150ba79) has left #beagle
[13:46:25] * mib_nsj92o (i=5056380b@gateway/web/ajax/mibbit.com/x-fcd916b01b838d61) Quit ("http://www.mibbit.com ajax IRC Client")
[13:48:57] <Crofton|work> urg
[13:50:38] * mib_2hyom6 (i=5198b1fe@gateway/web/ajax/mibbit.com/x-57dffa0038d8c831) Quit ("http://www.mibbit.com ajax IRC Client")
[13:51:59] * mib_2hyom6 (i=5198b1fe@gateway/web/ajax/mibbit.com/x-ec1bf8337d6d56d7) has joined #beagle
[13:53:56] * bazbell (n=a0192809@nat/ti/x-41783e8432673b0f) has joined #beagle
[13:54:53] <koen> man, this cold is getting annoying
[13:54:59] * koen fixes a fix for a fix
[13:55:35] <Crofton|work> does dsplink 1.60 still assume 256M of ram?
[13:56:12] <koen> Crofton|work: depends on the platform
[13:56:22] <Crofton|work> davinci ....
[13:56:22] <koen> Crofton|work: __alanc__ is the link guru :)
[13:56:23] * cbrake_away is now known as cbrake
[13:56:32] <koen> Crofton|work: he wrote the book on it
[13:56:33] <Crofton|work> maybe he is awake
[13:56:44] <Crofton|work> the one I have on my desk?
[13:57:18] <abeard> anybody know where the 2.6.22 TI kernel source lives?
[13:57:26] <abeard> The ones that has S-Video control?
[13:57:38] * mib_2hyom6 (i=5198b1fe@gateway/web/ajax/mibbit.com/x-ec1bf8337d6d56d7) Quit ("http://www.mibbit.com ajax IRC Client")
[13:58:09] <koen> tomba: booting a kernel with your latest stuff now
[13:58:13] <abeard> I've got the omap git kernel source, but I'm having issues finding the TI source
[14:01:25] <tomba> koen: you have to modify your scripts also, the display sysfs file has changed
[14:01:49] <koen> lcd e:1 u:1 t:0 h:480/1/28/2 v:640/1/1/1 p:19200
[14:02:17] <tomba> Documentation/arm/OMAP/DSS has an example script again
[14:02:54] <tomba> the sysfs files are getting more and more crowded, soon it's a must to fix them somehow =)
[14:03:27] <koen> this latest stuff has the overlay colour modes fixed?
[14:03:35] <tomba> yes
[14:04:00] <tomba> the boot arguments are also described in the readme, check them also
[14:04:51] <dcordes> koen: thanks
[14:07:27] * bazbell (n=a0192809@nat/ti/x-41783e8432673b0f) Quit (Remote closed the connection)
[14:08:20] <RobertK> abeard: they are on the bb google page.
[14:09:07] <RobertK> abeard: http://code.google.com/p/beagleboard/wiki/BeagleSourceCode
[14:09:29] <RobertK> abeard: but what do you what with such an old kernel?
[14:12:02] <abeard> RobertK: I need S-Video out support, and the from what I can tell the git kernel doesn't have it
[14:13:05] <koen> kulve, tomba: http://cgit.openembedded.net/cgit.cgi?url=openembedded/commit/&id=d45651ebb1a1ac66b850a35f16b2a58d14b56ec2
[14:13:06] <abeard> RobertK: I've found a couple google groups posts asking the same question, and they all seem to point at using the 2.6.22-18 TI kernel source
[14:13:32] <abeard> RobertK: Thanks for the link, I'll give that tarball a try
[14:13:52] <koen> abeard: 2.6.28 in OE has s-video support for beagle
[14:15:37] <abeard> koen OE == OpenEmbedded?
[14:15:48] <dcordes> koen: btw, how's the projector situation?
[14:16:04] <koen> dcordes: I have a new one which is working
[14:16:08] <koen> abeard: yes
[14:16:24] <dcordes> koen: they sent your a broken one first?
[14:16:34] <Crofton|work> koen broke the first one!
[14:16:36] <abeard> koen: Thanks, I'll give that a shot first
[14:16:40] <dcordes> :)
[14:16:44] <Crofton|work> I should get one for FOSDEM
[14:17:35] <koen> I bricked my with i2ctools
[14:17:35] <dcordes> I should get one for showing off and playing
[14:17:36] <Crofton|work> is there a way to the beagle video hw to flip the image for rear projection systems?
[14:17:37] <florian> koen: we need some kernel/feature support matrix for BB and OE
[14:17:50] <koen> florian: glad you volunteered!
[14:17:53] <Crofton|work> :)
[14:17:54] * philippe (n=fille@a83-245-252-47.elisa-laajakaista.fi) has joined #beagle
[14:18:07] * MT (i=86630288@gateway/web/ajax/mibbit.com/x-ef5c7fa867aabb69) Quit ("http://www.mibbit.com ajax IRC Client")
[14:18:12] <Crofton|work> we need to figure out how to get more money for OE work :)
[14:18:31] <florian> koen: hmm right, I could work on this. but you will have to provide data to fill it
[14:18:43] <dcordes> is it possible to brick vice versa, i.e. break the beagle with some device on the other i2c end?
[14:18:54] * sh1 (i=3d5fc3cd@gateway/web/ajax/mibbit.com/x-d3d649b7c55acb38) has joined #beagle
[14:18:59] <koen> no, you can't brick the beagle
[14:19:07] <Crofton|work> thank god
[14:19:09] <koen> unless you use a hammer
[14:20:39] <tomba> a brick works also
[14:21:05] <__alanc__> Crofton|work - this might help : http://tiexpressdsp.com/wiki/index.php?title=Changing_DSPLink_Memory_Map
[14:21:15] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
[14:21:31] <Crofton|work> yes, but you still need to adjust it manually
[14:22:43] <dcordes> koen: good. I want to add some controller for servo motors on i2c
[14:23:48] * abeard (i=c6f6abc8@gateway/web/ajax/mibbit.com/x-2dd2fddef1171987) Quit ("http://www.mibbit.com ajax IRC Client")
[14:26:13] <__alanc__> Crofton|work - ...__alanc__ will talk to Link crew...
[14:26:17] <kulve> koen: hmm.. I looking at the linux-omap_2.6.28.bb recipe. How do I know from where the mentioned patches are taken from? I.e. under linux-omap_2.6.28 directory? Or under linux-omap directory..?
[14:26:47] <Crofton|work> __alanc__, a guy on the linux-davinci list built link for this board
[14:26:59] <Crofton|work> I'm not sure if they realize you need to adjust the map
[14:27:13] <Crofton|work> one day, I need to make sure what is in OE work son this board
[14:27:23] <Crofton|work> my last attempt failed :)
[14:36:46] <tomba> koen: note also that you don't need huge amount of dma coherent memory anymore, as the dss uses it's own vram allocator
[14:40:15] * sh1 (i=3d5fc3cd@gateway/web/ajax/mibbit.com/x-d3d649b7c55acb38) Quit ("http://www.mibbit.com ajax IRC Client")
[14:41:10] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
[14:41:50] <koen> kulve: OE checks PN-PV/, then PN/ then files/ and . after that for patches
[14:42:16] <koen> kulve: and each of those dirs can have subdirs like beagleboard/ which would override patches in the dir below
[14:42:25] * Leon_Nardella (n=leon@200-161-14-111.dsl.telesp.net.br) has joined #beagle
[14:42:52] * Shivdas (n=root@59.160.172.220) has left #beagle
[14:43:00] <koen> Crofton|work: you do need all the TI toolchains for CE to build
[14:43:09] <koen> Crofton|work: the recipes should have the URLs
[14:44:24] <jkridner> koen: did you specify which C6x codegen to use?
[14:45:14] * prpplague (n=dave@mail.americanmicrosystems.com) has joined #beagle
[14:45:15] <Crofton|work> I have to bug out for an hour or so ...
[14:45:27] * Crofton|work bolts
[14:45:31] <koen> # To build the DSP applications you need to have DSP/BIOS, xdctools and codegen installed. Get these from:
[14:45:35] <koen> # https://www-a.ti.com/downloads/sds_support/targetcontent/bios/bios_5_33/sabios-j05/bios_setuplinux_5_33_02.bin
[14:45:37] <koen> # https://www-a.ti.com/downloads/sds_support/targetcontent/rtsc/xdctools_3_10_03/exports/xdctools_setuplinux_3_10_03.bin
[14:45:40] <koen> # https://www-a.ti.com/downloads/sds_support/targetcontent/LinuxDspTools/download.html
[14:46:11] <jkridner> koen: I mentioned the C6x codegen, because I think it is the only one you didn't provide a version in the link.
[14:46:40] <koen> ah, right
[14:47:04] <jkridner> I thought the include file was reasonably easy to find, but you can see that it can still stump people.
[14:47:48] * jkridner makes mental note to check the wiki page on installing Angstrom development tools for the BeagleBoard.
[14:50:47] <tomba> very strange. with DSS2, using mplayer with -vo xv, the video plays fine, but if window manager is running the video window is not close on exit. but if window manager is not running, it closes fine
[14:50:56] * JuanG (n=Juan@nat/ti/x-36d8b4655319c6b2) has joined #beagle
[14:50:59] * JuanG (n=Juan@nat/ti/x-36d8b4655319c6b2) has left #beagle
[14:51:07] <tomba> and looking at dss debug, nobody even tries to turn it off
[14:56:11] * garren|work (n=chatzill@mail.dm.co.za) Quit ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
[14:57:00] * garren|work (n=chatzill@mail.dm.co.za) has joined #beagle
[14:59:55] <koen> tomba: zuh and kulve are the xv masters :)
[15:01:38] * maelcum (n=quassel@e178133113.adsl.alicedsl.de) Quit (Read error: 110 (Connection timed out))
[15:02:43] <tomba> hmm, it has something to do with being the only window on the screen when ran without window manager
[15:03:10] <tomba> koen: did you get far with your testing? I have two quick fixes for the DSS already =)
[15:03:26] * garren|work (n=chatzill@mail.dm.co.za) Quit (Read error: 113 (No route to host))
[15:03:32] <kulve> tomba: X restarts when last client exists? It does that without -noreset
[15:04:31] <tomba> kulve: yes. with -noreset the video again stays on the screen. so now the question is: why nobody tries to close the overlay
[15:04:48] <kulve> I think koen pointed that out earlier
[15:04:56] <kulve> I might have time to debug that later this week..
[15:05:09] <kulve> i.e. bother zuh with it ;)
[15:05:54] <koen> tomba: X works on my evm, that's about all I tested
[15:06:02] * emeb (n=ericb@ip72-223-90-212.ph.ph.cox.net) has joined #beagle
[15:06:55] <zuh> xf86-video-omapfb never closes the handle if that's required. It just sets the plane_info.enabled-flag to zero and calls SETUP_PLANE
[15:07:09] <tomba> koen: ok. well, I pushed the two quick fixes to the test branch also
[15:07:11] <zuh> and unmaps the memory
[15:07:35] <tomba> zuh: for some reason that setup plane never comes to omapfb driver
[15:08:32] <tomba> zuh: are there tests or anything that could cause the cleanup code to be skipped?
[15:08:35] <koen> tomba: http://scap.linuxtogo.org/files/a9d9d199e936e45ef675f663cd3280e2.png
[15:08:41] <koen> the output on the lcd looks the same :)
[15:08:44] * koen pulls
[15:09:15] <zuh> if you run Xorg with -verbose you should see OMAPFBXVStopVideoGeneric entered if the driver tries to do it.
[15:09:22] <tomba> hmm I do get: (EE) OMAPFBXVStopVideoGeneric: Graphics sync failed
[15:09:41] <zuh> http://cgit.pingu.fi/xf86-video-omapfb/tree/src/omapfb-xv-generic.c#n292 <- then it doesn't reach that
[15:09:49] * matt_c (n=mcroydon@gateway.sunflowerbroadband.com) has joined #beagle
[15:09:54] <zuh> err, doesn't reach the SETUP_PLANE
[15:10:35] <tomba> hmm it shouldn't call sync if the display is not in manual update mode
[15:11:04] * jrmuizel (n=jrmuizel@mozca02.ca.mozilla.com) has joined #beagle
[15:11:05] * mib_2hyom6 (i=5198b1fe@gateway/web/ajax/mibbit.com/x-8918e0334e683c6d) has joined #beagle
[15:11:25] <tomba> I guess DSS1 did not return an error if using sync with non-syncable display. let's see
[15:12:02] <tomba> yes, DSS1 never returns an error with SYNC call. perhaps I'll make it so also
[15:12:58] * rupeshgujare (n=rupesh@59.160.172.220) has left #beagle
[15:13:25] * garren|work (n=chatzill@mail.dm.co.za) has joined #beagle
[15:15:52] <koen> I have fake vsync enabled, would that help?
[15:16:17] <tomba> no, you don't need fake vsync for anything. disable it
[15:16:59] <tomba> but now mplayer overlay closing works also. X didn't like GFX_SYNC returning an error. pushing fix to test branch in a minute
[15:18:01] <tomba> there
[15:18:24] <koen> drat
[15:18:38] <koen> that means I need to get the powervr stuff working with DSS2 :)
[15:19:08] <tomba> it's quite trivial to do that
[15:19:25] <tomba> if you know how things work =). I have it running here
[15:20:23] * NishanthMenon (n=nmenon@nat/ti/x-c841b2684472d39c) has joined #beagle
[15:21:53] <koen> tomba: you patched omaplfb or did you add stubs to dss2?
[15:21:58] <zuh> tomba: Hmm, so if not using manual updates, the sync is doing nothing?
[15:22:27] <tomba> zuh: yes. I thought it's better to return an error in that case, but perhaps not =)
[15:23:53] <tomba> koen: patched omaplfb
[15:24:56] <tomba> koen: altough I only fixed it for X output. well, fb output works also, but not the multi-buffering. so just a quick hack. but the omaplfb is a horrible hack anyway =)
[15:27:48] * sh1 (i=3d5fc3cd@gateway/web/ajax/mibbit.com/x-7398107e39290d59) has joined #beagle
[15:28:47] * gabba (i=d87245c2@gateway/web/ajax/mibbit.com/x-ae5c9da4e0df8555) Quit ("http://www.mibbit.com ajax IRC Client")
[15:29:46] <zuh> tomba: Well, I'll fix the driver to not call it unless in manual mode anyway, but it is very likely that someone would be tempted to use it like I did (ie. with the intent of making sure there's no pending work going on for teardown)
[15:30:33] <koen> zuh: so that pretty much only leaves the 32bit init :)
[15:32:42] <tomba> zuh: perhaps, but isn't it better to return an error then, so that people won't use it the wrong way =)
[15:33:35] <tomba> somebody should write documentation about the dss
[15:34:25] <zuh> tomba: Assuming there would be a way to know that it failed because you shouldn't call it :)
[15:34:45] <zuh> and not because some hw state error or whatever
[15:37:00] <tomba> =)
[15:38:36] <tomba> hmm does the xv driver support overlay scaling? I tried with mplayer's -fs switch, and I think it scaled with cpu
[15:39:00] <zuh> It should.
[15:39:47] <tomba> ah sorry, it does.
[15:39:48] <tomba> nice
[15:42:30] <koen> tomba: it does 420->422 in software (using C, not NEON)
[15:43:06] <garren|work> anyone got a openpandora board?
[15:43:25] <koen> no (t yet)
[15:43:57] <muriani> heh
[15:44:04] <DJWillis> garren|work: yep, a few
[15:44:20] <garren|work> it looks pretty awesome.. is it the same omap as the beagleboard?
[15:44:23] <muriani> I'm actually looking at the beagle just because it's like a stripped down openpandora
[15:44:29] <muriani> AFAIK yes it is
[15:44:42] <muriani> .hmm.. or does it not have the DSP? I thought it did.
[15:45:02] * muriani checks
[15:45:07] <DJWillis> Same SoC as the Beagle, DSP and PowerVR
[15:45:18] * RobertK (i=3edc0622@gateway/web/ajax/mibbit.com/x-3a68f6ef515ed911) Quit ("http://www.mibbit.com ajax IRC Client")
[15:45:34] <muriani> ah, yeah
[15:45:35] <koen> <ldesnogu mode>and NEON!</ldesnogu mode>
[15:45:48] <DJWillis> koen: ;-)
[15:46:47] <garren|work> be keen to get one when more are available... can you use oe to build linux for it
[15:46:56] * mib_8mgznu (i=43451e7a@gateway/web/ajax/mibbit.com/x-eaf7a1ebc38601a6) has joined #beagle
[15:46:57] <muriani> I don't see why now.
[15:47:05] <muriani> They're using Angstrom already, no?
[15:47:16] <muriani> s/now/not
[15:47:41] <DJWillis> garren|work: basic support is already in mainline OE and the whole base is Angstrom
[15:50:19] <garren|work> awesome... thats really cool.
[15:50:24] <RogerMonk> koen - clutter demo working at my end - nice work
[15:50:58] * mib_8mgznu (i=43451e7a@gateway/web/ajax/mibbit.com/x-eaf7a1ebc38601a6) Quit (Client Quit)
[15:51:45] <DJWillis> garren|work: there is a Pandora chan #openpandora but it is fairly quiet.
[15:53:00] <garren|work> ok... will keep my eye out on the project
[15:54:29] * garren|work (n=chatzill@mail.dm.co.za) Quit ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
[15:57:11] * recalcati (i=5d904010@gateway/web/ajax/mibbit.com/x-e20dbe312c99fa53) Quit ("http://www.mibbit.com ajax IRC Client")
[16:07:54] * zedstar (n=john@fsf/member/zedstar) Quit (Remote closed the connection)
[16:08:57] * jkridner|work (n=a0321898@nat/ti/x-c017bcbf176b3506) has joined #beagle
[16:10:38] * aleij (n=ad@204-50.78-83.cust.bluewin.ch) has joined #beagle
[16:11:56] * aleij (n=ad@204-50.78-83.cust.bluewin.ch) Quit (Client Quit)
[16:13:14] * recalcati (i=5d904010@gateway/web/ajax/mibbit.com/x-72379aed3706b4a8) has joined #beagle
[16:19:49] * mib_2hyom6 (i=5198b1fe@gateway/web/ajax/mibbit.com/x-8918e0334e683c6d) Quit ("http://www.mibbit.com ajax IRC Client")
[16:20:26] <chombourger> dumb question? how do I get traces out of the dspbridge kernel module? I have compiled it with CONFIG_BRIDGE_DEBUG but I am not getting any of the GT_[0-x]trace out
[16:21:35] * garren (n=garren@dsl-244-208-148.telkomadsl.co.za) has joined #beagle
[16:22:37] * mib_2hyom6 (i=5198b1fe@gateway/web/ajax/mibbit.com/x-aa07b7f101570745) has joined #beagle
[16:24:57] * aleij (n=ad@204-50.78-83.cust.bluewin.ch) has joined #beagle
[16:27:49] * bmxr_laptop (n=chatzill@142.232.129.24) has joined #beagle
[16:29:38] * bmxr_laptop (n=chatzill@142.232.129.24) Quit (Remote closed the connection)
[16:30:02] * _don_ (n=don@nat/ti/x-782894d2a237cd64) has joined #beagle
[16:35:05] * CarlMle2 (i=d5d20d1e@gateway/web/ajax/mibbit.com/x-5d4486affc4e00fb) Quit ("http://www.mibbit.com ajax IRC Client")
[16:35:56] <gregoiregentil> koen: yesterday, you removed the pvr patches for 2.6.27 but I think that we should keep the first one (pvr-dispc). Otherwise, I get an error during the insert of omaplfb because omap_..._plane is not there. do you agree?
[16:37:50] * mib_2hyom6 (i=5198b1fe@gateway/web/ajax/mibbit.com/x-aa07b7f101570745) Quit ("http://www.mibbit.com ajax IRC Client")
[16:37:59] * sh1 (i=3d5fc3cd@gateway/web/ajax/mibbit.com/x-7398107e39290d59) Quit ("http://www.mibbit.com ajax IRC Client")
[16:38:55] <gregoiregentil> koen: also, I'm still getting a better result with Pratheesh's binary module than yours coming from new omap3-sgx-modules. I get this error with yours: http://pastebin.com/d7e7454e2 Any idea?
[16:41:50] <koen> gregoiregentil: rogermonk also noticed I was a bit too sleepy yesterday evening to be committing stuff
[16:44:05] <gregoiregentil> No problem. You were too enthusiastic yesterday ;-) but I still get this error which is not linked to this missing patch. Something is wrong in pvrsrvkm. Don't know if it's just me
[16:45:01] * Leon_Nardella (n=leon@200-161-14-111.dsl.telesp.net.br) Quit (Remote closed the connection)
[16:46:20] * mib_maxiiq (i=4bd0bff0@gateway/web/ajax/mibbit.com/x-70159089ef2ffcf4) has joined #beagle
[16:47:40] <gregoiregentil> I got this too from omap3-sgx-modules: http://pastebin.com/d309c2d8b
[16:48:38] <koen> gregoiregentil: yeah, it reports the cpuclock being twice what it is
[16:49:09] <koen> sakoman: e-wm-config-angstrom now ships and enables the places module
[16:49:27] * TAK2004 (n=Administ@dslb-088-074-047-011.pools.arcor-ip.net) has joined #beagle
[16:49:32] <koen> sakoman: it also disabled the battery and temp module now
[16:49:47] <RogerMonk> koen - have you changes something for places module today?
[16:50:14] <RogerMonk> I'm getting module load error when desktop starts - is this related?
[16:50:17] <koen> yes
[16:50:23] <koen> I forgot to ship it in the image
[16:50:28] <RogerMonk> ah-ha
[16:50:31] <RogerMonk> I'll rebase
[16:50:33] <koen> I fixed that 5 minutes ago
[16:50:38] <RogerMonk> thanks
[16:50:54] <koen> the default desktop looks like http://scap.linuxtogo.org/files/c62b20dec1e5ae2250ffeaf7f5ca3cca.png now
[16:51:42] * koen foods
[16:51:56] <gregoiregentil> koen: so to confirm, you are getting it too? Still strange that I don't get it with Pratheesh, doesn't it?
[16:54:22] * Leon_Nardella (n=leon@200-161-14-111.dsl.telesp.net.br) has joined #beagle
[17:00:02] <sakoman_> koen: heh, I was just making the same changes :-)
[17:02:06] * cringthis (n=cringthi@nat/ti/x-6cb70990ce6b5b20) has joined #beagle
[17:02:52] * mib_maxiiq (i=4bd0bff0@gateway/web/ajax/mibbit.com/x-70159089ef2ffcf4) Quit ("http://www.mibbit.com ajax IRC Client")
[17:04:35] * mckoan is now known as mckoan|away
[17:10:00] * erbo (n=Erik@c-7b7de455.115-16-64736c14.cust.bredbandsbolaget.se) has joined #beagle
[17:12:19] * orifice_work (n=aho@67.71.223.35) has joined #beagle
[17:14:54] * n6pfk (n=mike@c-76-104-54-202.hsd1.va.comcast.net) has joined #beagle
[17:17:55] * CarlMle (i=d5d20d1e@gateway/web/ajax/mibbit.com/x-3e67edfc76e4b379) Quit ("http://www.mibbit.com ajax IRC Client")
[17:23:50] * dirk2 (n=dirk@p5B04320C.dip0.t-ipconnect.de) has joined #beagle
[17:25:52] <tharvey> is there an issue with NAND/jffs2 and 2.6.27 kernel from oe? - when trying to write to jffs2/NAND to populate my filesystem I get oops in jffs2_write_end - http://www.pastebin.ca/1301361
[17:27:23] * gregoiregentil (n=zonbu@adsl-71-135-114-242.dsl.pltn13.pacbell.net) has left #beagle
[17:27:56] * rsalveti (n=salveti@189.70.60.88) Quit (Read error: 104 (Connection reset by peer))
[17:28:55] <tharvey> this is with updated MLO/u-boot/kernel/filesystem built from oe pulled yesterday
[17:30:52] <dirk2> tharvey: http://www.beagleboard.org/irclogs/index.php?date=2008-10-23#T07:02:09 Maybe you like to ask jkridner if he can remember or/and search archives (at google enter: site:beagleboard.org jffs2)
[17:33:58] * Xenion (n=robert@p579FCDE3.dip.t-dialin.net) has joined #beagle
[17:34:22] <Xenion> hello
[17:34:27] <Xenion> jkridner, are you there ?
[17:34:50] <tharvey> hmmm... so not many people running rfs from NAND/jffs2? or if so - with what kernels?
[17:35:29] <Xenion> <- just wondering if by chance Thomas Leyrer from TI is using this channel :)
[17:35:51] <RogerMonk> Xenion - he's on sometimes
[17:36:03] <Xenion> RogerMonk, uhh, and with what nickname ?
[17:36:12] * ceyusa (n=ceyusa@189.163.35.95) Quit (Read error: 60 (Operation timed out))
[17:36:31] * gmansi (n=gmansi@190.173.76.176) Quit (Read error: 104 (Connection reset by peer))
[17:36:35] <RogerMonk> can't remember
[17:36:58] <Xenion> :/ too bad
[17:37:28] <RogerMonk> do you have something specific for him - if so I'll ping
[17:38:39] <Xenion> RogerMonk, it's just that i would ask him something about an email he send me some time ago.. andin this regards irc is always the most practical thing
[17:38:52] <RogerMonk> k
[17:42:37] <Xenion> i'll just mail him .. RogerMonk but if you remember his nickname, tell me :-)
[17:42:56] <RogerMonk> sure
[17:43:43] <garren> what is a .so file? a shared library file?
[17:44:30] * docelic_ is now known as docelic
[17:44:58] <koen> yay!
[17:45:06] * koen just got confirmed as speaker at bossa 2009
[17:45:15] <muriani> sweet
[17:45:29] <DJWillis> koen: awesome, flight included?
[17:45:49] <koen> it was last time
[17:46:08] * rsalveti (n=salveti@200.184.118.130) has joined #beagle
[17:47:59] <muriani> that's pretty badass
[17:52:02] <dirk2> Beagle contest: I updated the final schedule as it seems to be agreed at mailing list. Please change/fix if I missed something. http://elinux.org/BeagleBoard/contest#Contest_.231
[17:52:35] * aleij is now known as a|a
[17:53:54] <dirk2> additionally, I added (empty) contact templates for the projects. Seems we missed it the first time. I think we should have a chance to contact the winner ;) Please add your contact info: http://elinux.org/BeagleBoard/contest#Contest_projects
[17:55:42] * dixson3 (n=dixson3@cpe-24-28-79-164.austin.res.rr.com) Quit ()
[17:56:09] * dixson3 (n=dixson3@cpe-24-28-79-164.austin.res.rr.com) has joined #beagle
[17:57:30] * batteryman2 (i=4b34c99e@gateway/web/ajax/mibbit.com/x-3df243b5066d78aa) has joined #beagle
[17:59:04] * mib_2hyom6 (i=5198b1fe@gateway/web/ajax/mibbit.com/x-19bdab8ae88113f1) has joined #beagle
[18:00:05] * batteryman2 (i=4b34c99e@gateway/web/ajax/mibbit.com/x-3df243b5066d78aa) Quit (Client Quit)
[18:04:00] <koen> dirk2: looks good to me
[18:06:20] * felipec (n=felipec@189.153.88.243) has joined #beagle
[18:06:46] * RogerMonk (n=a0740758@nat/ti/x-9fe3daa891b2e395) has left #beagle
[18:14:14] * ZeZu (n=null@c-98-227-56-17.hsd1.in.comcast.net) Quit ()
[18:16:16] <sakoman_> dirk2: looks good to me too
[18:18:29] * gmansi (n=gmansi@190.173.80.23) has joined #beagle
[18:21:56] <Crofton|work> koen, congratulations
[18:25:34] * guillaum1 (n=GZBA4143@AMontsouris-153-1-25-222.w86-212.abo.wanadoo.fr) Quit ("Leaving.")
[18:26:01] * mib_2hyom6 (i=5198b1fe@gateway/web/ajax/mibbit.com/x-19bdab8ae88113f1) Quit ("http://www.mibbit.com ajax IRC Client")
[18:42:42] * mib_dts4n9 (i=4eb012f4@gateway/web/ajax/mibbit.com/x-8ce9186079f23313) has joined #beagle
[18:43:02] * recalcati (i=5d904010@gateway/web/ajax/mibbit.com/x-72379aed3706b4a8) Quit ("http://www.mibbit.com ajax IRC Client")
[18:43:02] <keesj> koen great !
[18:46:29] * ceyusa (n=ceyusa@189.163.35.95) has joined #beagle
[18:46:51] <mru> hi guys
[18:47:13] <Crofton|work> gm
[18:48:38] * eFfeM (n=frans@195-241-226-180.ip.telfort.nl) has joined #beagle
[18:49:02] <koen> hey mru
[18:49:16] <mru> raster: ping
[18:54:46] * mib_dts4n9 (i=4eb012f4@gateway/web/ajax/mibbit.com/x-8ce9186079f23313) Quit ("http://www.mibbit.com ajax IRC Client")
[18:55:16] * florian (n=fuchs@217.146.132.69) Quit (Remote closed the connection)
[18:57:40] * mib_maxiiq (i=803713d8@gateway/web/ajax/mibbit.com/x-9042c0dbac194724) has joined #beagle
[18:58:06] * koen discovers you need to be root for to GLES
[19:02:34] <DJWillis> koen: urgg :(. Well that could screw up one of my plans then ;-)
[19:02:51] <koen> maybe I just need to chmod /dev/pvr*
[19:03:44] <eFfeM> koen, probably, otherwise use a setuid-ed app
[19:03:59] <eFfeM> chmod 4711 whatever
[19:04:06] <eFfeM> does not smell :-)
[19:04:53] * suihkulokki (i=nchip@naru.obs2.net) has joined #beagle
[19:05:31] * mib_2hyom6 (i=5198b1fe@gateway/web/ajax/mibbit.com/x-80b6ae7ce64605df) has joined #beagle
[19:10:59] <adj_> damn, streaming video from usb webcam to browser seems to crash musb functionality very often
[19:11:31] * koen hates writing bios
[19:11:54] <koen> adj_: at least your cam works :)
[19:13:06] * guillaum1 (n=Guillaum@AMontsouris-153-1-25-222.w86-212.abo.wanadoo.fr) has joined #beagle
[19:13:15] <adj_> at least a few minutes at a time, until usb dies :)
[19:13:34] <koen> tried making musb PIO only?
[19:13:40] <koen> it's somewhere int he kernel defconfig
[19:13:58] <koen> CONFIG_USB_INVENTRA_DMA=y
[19:15:10] <adj_> haven't tried that yet. I'll have try it, thanks for the tip
[19:16:40] <mru> do you have that workaround for some bug with musb full-duplex dma?
[19:16:42] <tharvey> I'm having some issues with musb as well, booting with a kernel from NAND and mounting rootfs from USB etx2 partition - I can't write to the rootfilesystem
[19:16:56] <tharvey> usb over otg is not stable currently?
[19:18:35] * odesus (n=OMAR@148.223.106.43) has joined #beagle
[19:24:54] <tharvey> my USB rootfs issues only appear when going through a USB 2.0 HUB
[19:29:47] <garren> is there a way to set the resolution on boot up, my screen is saying out of range
[19:30:27] * odesus (n=OMAR@148.223.106.43) has left #beagle
[19:31:09] * bazbell (n=a0192809@nat/ti/x-223a5acb35d9dfed) has joined #beagle
[19:35:03] * dirk2 (n=dirk@p5B04320C.dip0.t-ipconnect.de) Quit ("Konversation terminated!")
[19:40:51] * jrmuizel (n=jrmuizel@mozca02.ca.mozilla.com) Quit ()
[19:43:50] <tharvey> jkridner|work, you mention back in previous IRC logs about having a kernel that was stable with JFFS2?
[19:46:19] * jrmuizel (n=jrmuizel@bas19-toronto12-1177849598.dsl.bell.ca) has joined #beagle
[19:49:22] <garren> will this command set the display to 720: video=omapfb:mode:720p60
[19:50:41] <adj_> mru: propably not, i'll check that, too
[19:53:04] * jrmuizel_ (n=jrmuizel@bas19-toronto12-1177849598.dsl.bell.ca) has joined #beagle
[19:53:56] * jrmuizel (n=jrmuizel@bas19-toronto12-1177849598.dsl.bell.ca) Quit (Read error: 104 (Connection reset by peer))
[19:56:34] * maelcum (n=quassel@e178185150.adsl.alicedsl.de) has joined #beagle
[20:02:01] <tharvey> are there known/common issues with OTG going through a hub? I've seen flakey issues going through both a 1.0 and 2.0 hub (both self powered)
[20:03:10] * mib_2hyom6 (i=5198b1fe@gateway/web/ajax/mibbit.com/x-80b6ae7ce64605df) Quit ("http://www.mibbit.com ajax IRC Client")
[20:09:37] <sakoman_> tharvey: depends on which kernel you are using. IIRC 26.27 had some issues with low speed devices on second level hubs
[20:11:09] <tharvey> sakoman_, I'm using 2.6.27-omap1 built from OE (linux-omap-2.6.27-r8)
[20:12:21] <sakoman_> my guess is that kernel suffers from the bug I mentioned
[20:12:26] * jrmuizel_ (n=jrmuizel@bas19-toronto12-1177849598.dsl.bell.ca) Quit ()
[20:12:33] <tharvey> what kernel 'should' I be using? I figured the 2.6.27-r8 kernel wouldn't be default in OE unless it was fairly stable?
[20:12:51] <tharvey> I've had NAND/jffs2 issues as well as USB/hub issues
[20:13:00] <sakoman_> but it only showed up with things like USB KVM switches, keyboards with built in hubs, ec.
[20:13:31] <tharvey> I'm having issues with usb mass storage through hubs
[20:13:36] <sakoman_> it is stable - just not bug free :-)
[20:13:52] <tharvey> using a USB rootfs (because I can't seem to flash a JFFS2 onto NAND)
[20:14:05] <sakoman_> haven't heard about any bugs w/ usb mass storage
[20:14:11] <chombourger> Anyone got a clue why tiopenmax (through gst-openmax) is trying to create a DSP node with a 10MB heap size (Node_Create therefore failing with DSP_EMEMORY)?
[20:14:29] <tharvey> well... can anyone recommend a kernel that works with NAND/jffs2 and fairly well with USB?
[20:14:40] <sakoman_> tharvey: are you using u-boot or linux mtd-tools to write nand?
[20:15:35] <tharvey> I'm using the same 2.6.27-omap1 kernel with mtd-tools to write nand - I'm open to trying any other kernel/rootfs
[20:15:53] <tharvey> the Angstrom demo I tried didn't have mtd-tools
[20:16:15] <sakoman_> I work mostly with Overo and never have had any issues with mtd-tools and 2.6.27
[20:16:18] * jrmuizel (n=jrmuizel@mozca02.ca.mozilla.com) has joined #beagle
[20:16:26] <tharvey> sakoman_, but I did try uboot as well as a 2nd choice and it failed but I believe it was because of differences in nand.jffs2 vs kernel's jffs2
[20:16:32] <koen> es3.0 beagles out there, interesting
[20:16:38] <sakoman_> In fact Gumstix uses mtd-ools to fflash the rootfs for every uni they ship
[20:16:58] <tharvey> sakoman_, what kernel are you using?
[20:17:17] <sakoman_> I'm always at the bleeding edge
[20:17:23] <tharvey> I mean to say, where did you build it... you using OE?
[20:17:31] <sakoman_> yes, OE
[20:17:40] * hugsig (n=hsi@2001:16d8:ff1a:0:216:3eff:fe5a:59cd) has joined #beagle
[20:17:47] <tharvey> oe-git? what recipe?
[20:18:11] <sakoman_> Remember I am working on Overo 99% of the time
[20:18:25] <tharvey> I wonder if I should be using linux-omap_git.bb
[20:18:31] * a|a is now known as aleij
[20:18:51] <tharvey> koen, you do all the OE recipes correct? what kernel would you say I should be using that you think should have stable jffs2/nand?
[20:18:53] <sakoman_> I use a recipe that is extremely close to koen's linux-omap_git
[20:19:12] * mib_maxiiq (i=803713d8@gateway/web/ajax/mibbit.com/x-9042c0dbac194724) Quit ("http://www.mibbit.com ajax IRC Client")
[20:19:19] * mib_maxiiq (i=803713d8@gateway/web/ajax/mibbit.com/x-43d7a362dc1115f1) has joined #beagle
[20:19:45] * jrmuizel (n=jrmuizel@mozca02.ca.mozilla.com) Quit (Client Quit)
[20:20:01] <sakoman_> Not sure why you are having issues with nand -- not something I have seen
[20:20:13] <koen> tharvey: I'd advise using the kernel OE builds by default
[20:20:37] <sakoman_> sounds like good advice to me :-)
[20:20:51] <garren> is alsa configure in kernel 2.6.27 ?
[20:21:18] <tharvey> koen, thats 2.6.27-r8 - which is not allowing me to flash jffs2 onto nand - does that surprise you? I'm wondering if its due to something in my uboot? I'm using uboot/mlo built from oe yesterday as well
[20:21:29] <koen> no idea on that
[20:22:29] * ceyusa (n=ceyusa@189.163.35.95) Quit (Read error: 60 (Operation timed out))
[20:23:41] <tharvey> perhaps I'll try linux-omap_git.bb
[20:24:30] * jrmuizel (n=jrmuizel@mozca02.ca.mozilla.com) has joined #beagle
[20:32:02] * mib_2hyom6 (i=5198b1fe@gateway/web/ajax/mibbit.com/x-6915b3a50e899766) has joined #beagle
[20:36:06] <adj_> tharvey: what issues are you having with nand and jffs2?
[20:36:13] * mib_2hyom6 (i=5198b1fe@gateway/web/ajax/mibbit.com/x-6915b3a50e899766) Quit (Client Quit)
[20:36:15] <adj_> i use those all the time
[20:37:17] * florian (n=fuchs@f049004246.adsl.alicedsl.de) has joined #beagle
[20:42:07] <adj_> hmm, seems that kernels 2.6.27-r5 and -r7 don't have that patch which corrects nand functionality on beagle
[20:42:36] <adj_> i mean those kernels in OE
[20:44:39] <mru> random thought: how come "nor" is a commonly used word, whereas "nand" is not?
[20:46:45] <raster> mru: pong
[20:47:09] <mru> raster: you were talking about arm/neon communication earlier today
[20:47:26] <mru> your idea to rotate the registers doesn't work
[20:47:43] <mru> sending data from arm to neon is fast
[20:47:51] <mru> retrieving data is slow
[20:48:16] <Crofton|work> so save intermediate results in NEON ....
[20:48:40] <raster> mru: no no.. what i meant was
[20:48:43] <mru> it's impossible to move a value from neon to arm in less than 20 cycles
[20:48:46] <raster> WITHIN neon - you rotate
[20:48:50] <adj_> tharvey: http://marc.info/?l=linux-omap&m=122609645332227&w=4
[20:49:03] <raster> otherwise u keep neon/arm pretty much separate
[20:49:06] <mru> raster: no need for that for that reason
[20:49:21] <Crofton|work> that is ok for many to one result calcs, bad for many to man y ...
[20:49:29] <mru> obviously unrolling loops with a bit of pipelining is a good idea
[20:49:51] <raster> so data stays within arm or within neon
[20:50:06] <mru> also avoid accessing the same memory areas from both arm and neon
[20:50:07] <raster> and goes in from memory and back to memory when finished processing
[20:50:12] * mib_2hyom6 (i=5198b1fe@gateway/web/ajax/mibbit.com/x-67fdf9391bc541e2) has joined #beagle
[20:50:21] <raster> but you avoid sending data to/from arm/neon
[20:50:32] <raster> so they just work on different things entirely
[20:50:48] <raster> yup
[20:50:53] <raster> for example
[20:51:11] <raster> neon goes and process top-half of image, arm does bottom-half
[20:51:16] <raster> (silly example - but you get the idea)
[20:51:17] <mru> ideally you should not access the same cache lines from arm and neon within 20 cycles
[20:51:28] <Crofton|work> what if you move something between two neon regs ...
[20:51:31] * cringthis (n=cringthi@nat/ti/x-6cb70990ce6b5b20) has left #beagle
[20:51:47] <raster> and you then interleave your arm and neon instructions
[20:51:52] <mru> moving between neon regs is fast
[20:52:03] <mru> I understand what you're getting at
[20:52:21] <mru> in practise it's very hard to make it any faster
[20:52:22] <raster> it'll be hell to code
[20:52:23] <raster> :)
[20:52:32] <Crofton|work> you guys are nuts :)
[20:53:59] <mru> what tends to happen is that the arm pipeline stalls, starving the neon unit of instructions
[20:54:42] <raster> mmm good point
[20:54:45] <raster> didn't think of that
[20:55:01] <koen> would the a9 be better at that?
[20:55:06] <mru> you can do a small amount of arm work when the neon instruction queue has something in it
[20:55:06] * ceyusa (n=ceyusa@189.163.35.95) has joined #beagle
[20:55:44] <mru> I don't know how a9 will be in that regard
[20:55:52] <mru> I only know they've made some changes
[20:56:48] * Leon_Nardella1 (n=leon@200-161-14-111.dsl.telesp.net.br) has joined #beagle
[20:57:15] * mib_tcr87v (i=5a0f152d@gateway/web/ajax/mibbit.com/x-f3c7b40b73d6f2bd) Quit ("http://www.mibbit.com ajax IRC Client")
[20:57:18] <mru> ldesnogu could probably tell us more be doubt he's allowed to
[20:57:32] <raster> koen: it's going to be mighty-tricky
[20:57:56] <raster> as then they need to allow out-of-order processing between arm and neon instrs
[20:58:16] <mru> a9 is out of order
[20:58:20] <raster> and then have to synchronise when they hit a known common resource again
[20:58:23] * gregoiregentil (n=zonbu@adsl-71-135-114-242.dsl.pltn13.pacbell.net) has joined #beagle
[20:58:42] <mru> on a8 the arm an neon pipelines are already asynchronous
[20:58:52] <mru> that's why it takes so many cycles to get data back
[20:59:13] <raster> mru: yeah - but the stall is the problem
[20:59:26] <raster> if u can avoid the stall...
[21:00:14] <mru> the really annoying bit on a8 is that *any* register access will stall the arm pipeline if it's waiting for neon data
[21:01:01] <mru> you can issue more neon instructions during the delay though
[21:01:18] <raster> as long as u queued them up before...
[21:01:18] <raster> :)
[21:01:49] <mru> so vmov r0, r1, d0; lots of neon; add r2, r0, r1 won't stall
[21:02:58] <mru> if you don't have any neon work to do while waiting, you're in trouble
[21:03:06] <mru> since any normal arm instruction will stall
[21:03:54] <mru> one workaround is to write the neon data to a memory location, do 20 cycles worth of arm stuff, then read the memory
[21:04:10] * bazbell (n=a0192809@nat/ti/x-223a5acb35d9dfed) Quit ("Leaving.")
[21:04:11] <mru> but don't touch the same cacheline before the 20 cycles are up
[21:04:42] <mru> actually its only a 16-byte block that is locked for 20 cycles
[21:04:56] <mru> the rest of the cacheline becomes accessible sooner
[21:05:06] * ssvb (n=ssvb___@a88-114-221-132.elisa-laajakaista.fi) has joined #beagle
[21:05:38] <mru> http://hardwarebug.org/2008/12/31/arm-neon-memory-hazards/
[21:06:03] <Pavlov> a a9 beagle would rule
[21:06:06] * Leon_Nardella (n=leon@200-161-14-111.dsl.telesp.net.br) Quit (Connection timed out)
[21:06:59] <mru> they're probably working on an a9 based chip as we speak
[21:07:05] <Pavlov> no doubt
[21:07:43] * mib_tcr87v (i=5a1d5fb2@gateway/web/ajax/mibbit.com/x-1847b82da6cff9a8) has joined #beagle
[21:13:51] <raster> mru: i'll keep it in mind
[21:14:09] <raster> mru: i just like the idea of getting as manu transistors as i can doing evil deeds for me... at once
[21:14:19] <raster> idle transistors are like idle children
[21:14:21] <muriani> hahaha
[21:14:25] <raster> sooner or later they cause trouble
[21:14:37] <raster> and are a waste - they could have been set to work in the mines!
[21:14:38] <raster> :)
[21:14:48] <mru> I totally agree
[21:15:13] <mru> I'm just telling you what difficulties I've experienced
[21:15:19] <raster> thats cool
[21:15:27] <raster> what were you trying to do?
[21:17:49] <mru> some pixel shuffling
[21:18:18] <mru> doing some of it with arm instructions only made it slower
[21:18:30] * garren (n=garren@dsl-244-208-148.telkomadsl.co.za) Quit ("Ex-Chat")
[21:21:00] <raster> pixels for what
[21:21:01] * mib_c6f8of (i=558823a4@gateway/web/ajax/mibbit.com/x-0c918ff62b444e53) has joined #beagle
[21:21:11] <raster> video codec stuff?
[21:21:21] <raster> or more raw rendering?
[21:22:34] <mru> unpacking 4-bit data into 8 bits
[21:22:55] <mru> that was one thing I tried such tricks on
[21:24:21] * philippe (n=fille@a83-245-252-47.elisa-laajakaista.fi) Quit (Read error: 113 (No route to host))
[21:26:02] <jkridner|work> tharvey: I found that the Angstrom 2.6.26-r64 kernel was stable with JFFS2.
[21:27:03] <raster> oooh
[21:27:19] * chomb (i=4dc716a9@gateway/web/ajax/mibbit.com/x-8ce1635f27eaac10) Quit ("http://www.mibbit.com ajax IRC Client")
[21:27:22] <raster> so it seems you get cacheline thrashing in the neon
[21:27:45] <raster> it doesnt keep the cacheline split into segments ans rotate them etc.
[21:27:53] <eFfeM> mru: Pavlov: http://www.arm.com/products/CPUs/ARMCortex-A9_MPCore.html
[21:28:11] * emeb (n=ericb@ip72-223-90-212.ph.ph.cox.net) has left #beagle
[21:28:31] <eFfeM> mru and for unpacking 4 bit data into 8 bits use table lookup (if I understand the problem properly)
[21:29:36] <florian> koen: http://wiki.openembedded.net/index.php/BeagleBoard
[21:29:46] <florian> koen: Something like this....
[21:29:49] * robtow (n=rtow@nat/montavista/x-a4f9823c293e64b1) has joined #beagle
[21:29:59] <mru> eFfeM: repeating the 4 bits gets close enough
[21:39:32] <eFfeM> ok fine
[21:41:16] * dcordes_ (n=dcordes@unaffiliated/dcordes) has joined #beagle
[21:43:18] * mib_c6f8of (i=558823a4@gateway/web/ajax/mibbit.com/x-0c918ff62b444e53) Quit ("http://www.mibbit.com ajax IRC Client")
[21:43:20] * chombourger (n=mvista-f@169.22.199-77.rev.gaoland.net) Quit (Read error: 104 (Connection reset by peer))
[21:43:54] * mib_c6f8of (i=558823a4@gateway/web/ajax/mibbit.com/x-d2dc8e46691443ea) has joined #beagle
[21:43:57] * eFfeM (n=frans@195-241-226-180.ip.telfort.nl) Quit ("Leaving.")
[21:44:12] * felipec (n=felipec@189.153.88.243) Quit (Read error: 110 (Connection timed out))
[21:49:03] * ceyusa (n=ceyusa@189.163.35.95) Quit (Read error: 110 (Connection timed out))
[21:49:03] <tharvey> adj_, here's what I had with nand/jffs2 and 2.6.27-omap http://www.pastebin.ca/1301361
[21:49:49] * dcordes (n=dcordes@unaffiliated/dcordes) Quit (Read error: 145 (Connection timed out))
[21:50:07] * felipec (n=felipec@189.153.88.243) has joined #beagle
[21:50:23] <tharvey> adj_, thx for the pointer on hte NAND patch - the oe default kernel definately does not implement that patch - I'll test and provide feedback
[21:51:11] <tharvey> jkridner|work, thx is the Angstrom 2.6.26-r64 kernel bin available for download anywhere?
[21:55:40] * mib_wvj1fp (i=c72ec8e7@gateway/web/ajax/mibbit.com/x-ac27739b26cc234c) has joined #beagle
[21:56:01] * mib_wvj1fp (i=c72ec8e7@gateway/web/ajax/mibbit.com/x-ac27739b26cc234c) Quit (Client Quit)
[22:06:43] * danrens (i=42b90382@gateway/web/ajax/mibbit.com/x-9be729a612d115f9) has joined #beagle
[22:07:50] * zedstar (n=john@fsf/member/zedstar) has joined #beagle
[22:08:22] <danrens> Could someone help me out. I have a beagle board all set up. I'm getting ready to port some code the the board that uses the serial port for I/O using XBee. can I use the existing DB9 or is it only used for the debug console?
[22:10:36] <tharvey> danrens, I assume you mean the 10pin header (no db9 on the beagle) - you should be able to use this serial port but not as a console at the same time
[22:10:56] * gcohler (n=gcohler@nat/ti/x-bfe7edc5bc610b8f) Quit (Remote closed the connection)
[22:11:01] <tharvey> the omap has several uarts... not sure what the state of the uart driver is or if they are made avail on the expansion headers
[22:11:16] <tharvey> or you can use a usb-to-serial device
[22:11:19] * assem (n=assem@terranpro.org) has joined #beagle
[22:12:17] <danrens> tharvey, thanks... do you know how to switch the existing 10 pin header to and from console mode?
[22:12:45] <danrens> the last thing I want to do is lock myself out.
[22:13:21] * mib_tcr87v (i=5a1d5fb2@gateway/web/ajax/mibbit.com/x-1847b82da6cff9a8) Quit ("http://www.mibbit.com ajax IRC Client")
[22:13:25] <mru> add/remove the corresponding getty line in /etc/inittab and the console=ttyS2... kernel parameter
[22:14:01] <mru> if your root filesystem is on an sd card you can always edit it in a pc if you make a mistake
[22:14:36] <danrens> cool, thanks i do appreciate your help
[22:15:21] * mib_c6f8of (i=558823a4@gateway/web/ajax/mibbit.com/x-d2dc8e46691443ea) Quit ("http://www.mibbit.com ajax IRC Client")
[22:15:33] * mib_c6f8of (i=558823a4@gateway/web/ajax/mibbit.com/x-72668355bfa164e5) has joined #beagle
[22:17:16] * geckosenator (n=sean@ppp-69-218-241-177.dsl.dytnoh.ameritech.net) has joined #beagle
[22:17:38] * alecrim (n=alecrim@189.2.128.130) has joined #beagle
[22:19:57] * valhalla (n=valhalla@81-174-21-57.dynamic.ngi.it) Quit ("Leaving")
[22:20:17] * robtow (n=rtow@nat/montavista/x-a4f9823c293e64b1) Quit (Excess Flood)
[22:20:44] * robtow (n=rtow@nat/montavista/x-9b921fc1dad185ea) has joined #beagle
[22:21:00] <ldesnogu_> koen, you should not quote me as NEON supporter, I'm not a big fan of it to be honest :)
[22:21:22] * mib_2hyom6 (i=5198b1fe@gateway/web/ajax/mibbit.com/x-67fdf9391bc541e2) Quit ("http://www.mibbit.com ajax IRC Client")
[22:23:17] <koen> ldesnogu_: no, but you always point out that cortex doesn't need to have neon :)
[22:23:50] <ldesnogu_> oh yes but that's for A9 :)
[22:28:04] * guillaum1 (n=Guillaum@AMontsouris-153-1-25-222.w86-212.abo.wanadoo.fr) Quit ("Leaving.")
[22:28:34] * guillaum1 (n=gl@AMontsouris-153-1-25-222.w86-212.abo.wanadoo.fr) has joined #beagle
[22:28:49] <BeagleBot> http://beagleboard.org/pico/ was edited by jmalina.myopenid.com
[22:29:52] * abitos (n=nixgibts@dslb-084-057-136-099.pools.arcor-ip.net) has joined #beagle
[22:37:44] * danrens (i=42b90382@gateway/web/ajax/mibbit.com/x-9be729a612d115f9) Quit ("http://www.mibbit.com ajax IRC Client")
[22:40:02] * ceyusa (n=ceyusa@189.163.35.95) has joined #beagle
[22:47:24] * mib_c6f8of (i=558823a4@gateway/web/ajax/mibbit.com/x-72668355bfa164e5) Quit ("http://www.mibbit.com ajax IRC Client")
[22:47:33] <tharvey> adj_, thx again the one-liner patch here http://marc.info/?l=linux-omap&m=122609645332227&w=4 does solve the NAND issues for 2.6.27
[22:58:41] * vpdn (n=vpdn@dslb-084-056-128-078.pools.arcor-ip.net) has joined #beagle
[22:59:45] <vpdn> Hi, I'm looking for a way to run flash movies on a small embedded device and heared about the beagleboard.
[23:00:38] <vpdn> anyone knows whether the processor on the beagleboard is fast enough to run swf's?
[23:01:53] <geckosenator> swf?
[23:01:59] * uwe2 (n=uwe@dslb-084-056-008-063.pools.arcor-ip.net) Quit (Read error: 60 (Operation timed out))
[23:02:22] <tharvey> swf = shockwave flash
[23:02:51] <jkridner|work> vpdn: try swf-dec and see, but the answer is yes.
[23:03:11] <jkridner|work> it can do 720p video even.
[23:05:55] * uwe2 (n=uwe@dslb-084-056-054-220.pools.arcor-ip.net) has joined #beagle
[23:05:55] <vpdn> is there a reason why you recommend swf-dec over the original player?
[23:07:03] <raster> theres no source for the original
[23:07:08] <raster> thus good luck compiling it
[23:07:14] * florian just managed to make his usb camera work with the beagle
[23:08:03] <vpdn> raster: you got a point here =D
[23:08:20] <koen> florian: I suspect moving the keymap from 12keymap to 98keymap would also fix the kdrive 1.4 race
[23:11:45] * cbrake is now known as cbrake_away
[23:12:47] * matt_c (n=mcroydon@gateway.sunflowerbroadband.com) Quit ()
[23:13:21] * ldesnogu_ (n=ldesnogu@ven06-2-82-247-86-183.fbx.proxad.net) Quit ()
[23:13:34] <florian> koen: that would be good, could you verify this?
[23:15:01] <koen> I'll try to have xora verify his theory :)
[23:16:02] * koen should get his bike fixed first
[23:17:19] <koen> one of the sprockets is trying real hard to be a circle
[23:18:01] * uwe_ (n=uwe_@dslb-084-056-008-063.pools.arcor-ip.net) Quit (Read error: 110 (Connection timed out))
[23:18:28] <jkridner|work> fyi, 2.6.28 EHCI works with Beagle Rev C1D.
[23:18:39] <koen> yay!
[23:18:48] <tharvey> nice... when do they ship?
[23:18:48] <koen> with a patched uboot I presume?
[23:20:19] <jkridner|work> March.
[23:20:27] <jkridner|work> oh, right. forgot about that.
[23:20:52] <jkridner|work> I have the kernel patch here that I was prepping. skipisz did the kernel patch.
[23:21:17] <koen> I should probably apply the uboot patch to the uboot recipe in OE soon
[23:21:39] <jkridner|work> I got the kernel source from steve and built a patch off of that.
[23:22:07] <jkridner|work> is there a u-boot patch already?
[23:22:17] <koen> jkridner|work: FWIW, tony applied a bunch of patches today
[23:22:30] <koen> he's going to tag 2.6.28-omap1 this week
[23:22:56] <jkridner|work> hmmm.... I might be out of date then.
[23:23:03] <jkridner|work> I last pulled 9 hours ago.
[23:23:12] <koen> I say that because he's going to remove things like this:
[23:23:13] <koen> - gpio_request(gpio + 1, "EHCI_nOC");
[23:23:14] <koen> - gpio_direction_input(gpio + 1);
[23:23:31] <koen> if I read RMKs email right
[23:24:14] <koen> ah, no
[23:24:18] <koen> I read it wrong
[23:24:19] * geckosenator (n=sean@ppp-69-218-241-177.dsl.dytnoh.ameritech.net) Quit (Read error: 104 (Connection reset by peer))
[23:25:30] <koen> jkridner|work: I am fairly sure that omap_writel is deprecated
[23:26:12] * geckosenator (n=sean@ppp-69-218-241-177.dsl.dytnoh.ameritech.net) has joined #beagle
[23:26:53] * abitos (n=nixgibts@dslb-084-057-136-099.pools.arcor-ip.net) Quit (Read error: 113 (No route to host))
[23:27:28] * koen double checks
[23:27:30] <koen> NOTE: - Use omap_read/write[bwl] for physical register addresses
[23:27:37] <koen> - Use __raw_read/write[bwl]() for virtual register addresses
[23:27:51] <koen> - Use IO_ADDRESS(phys_addr) to convert registers to virtual addresses
[23:31:30] <tharvey> koen, theres a patch needed to 2.6.27 OE recipe for NAND to work - http://thread.gmane.org/gmane.linux.ports.arm.omap/14945
[23:32:22] <koen> you're behind the times :)
[23:32:24] <koen> tharvey: http://cgit.openembedded.net/cgit.cgi?url=openembedded/commit/&id=bbbf56d7c111a4a8f6162803675be1feb74b0084
[23:33:41] <tharvey> damn... that just went in today?
[23:33:55] * aleij_ (n=ad@204-50.78-83.cust.bluewin.ch) has joined #beagle
[23:34:57] * aleij (n=ad@204-50.78-83.cust.bluewin.ch) Quit (Read error: 104 (Connection reset by peer))
[23:35:24] <tharvey> koen, dang... nice work - difficult to keep up with you - I was up to date yesterday :P
[23:39:36] * ddompe (n=ddompe@200.122.155.113) Quit ()
[23:39:55] * emeb_mac (n=ericb@ip72-223-90-212.ph.ph.cox.net) has joined #beagle
[23:44:47] * mib_c6f8of (i=558823a4@gateway/web/ajax/mibbit.com/x-0b7e6e1d4acff532) has joined #beagle
[23:45:30] <sakoman_> koen: heh, I forgot all about that :-)
[23:45:42] <sakoman_> fix the bugs & move on . . .
[23:46:00] * rsalveti (n=salveti@200.184.118.130) Quit (Remote closed the connection)
[23:47:40] * gregoiregentil (n=zonbu@adsl-71-135-114-242.dsl.pltn13.pacbell.net) has left #beagle
[23:48:00] * ceyusa (n=ceyusa@189.163.35.95) Quit (Read error: 60 (Operation timed out))
[23:48:07] <jkridner|work> koen: are you looking at my try-steve3 branch or something?
[23:48:12] <raster> there are no bugs
[23:48:14] <raster> just features
[23:48:26] * Leon_Nardella1 is now known as Leon_Nardella
[23:49:24] * florian (n=fuchs@f049004246.adsl.alicedsl.de) Quit ("Verlassend")
[23:51:14] * prpplague (n=dave@mail.americanmicrosystems.com) Quit ("Leaving")
[23:52:52] * uwe_ (n=uwe_@dslb-084-056-054-220.pools.arcor-ip.net) has joined #beagle
[23:53:35] * felipec (n=felipec@189.153.88.243) Quit (Read error: 145 (Connection timed out))
[23:56:20] * vpdn (n=vpdn@dslb-084-056-128-078.pools.arcor-ip.net) Quit ()
[23:58:43] <Crofton|work> finally, made some real progress

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.