• [00:06:24] * nemequ (n=nemequ@ip68-111-215-155.sd.sd.cox.net) Quit (Remote closed the connection)
  • [00:07:05] * guillaum1 (n=Guillaum@AMontsouris-153-1-82-57.w90-2.abo.wanadoo.fr) Quit ("Leaving.")
  • [00:07:22] * guillaum1 (n=gl@AMontsouris-153-1-82-57.w90-2.abo.wanadoo.fr) has joined #beagle
  • [00:12:50] * notaz_ (n=notaz@78.59.134.74) has left #beagle
  • [00:13:38] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
  • [00:13:42] * abitos (n=nixgibts@dslb-084-057-154-061.pools.arcor-ip.net) Quit (Read error: 113 (No route to host))
  • [00:27:04] * RobotGuy (n=robotguy@pool-96-225-210-142.ptldor.fios.verizon.net) has joined #Beagle
  • [01:04:39] * Olipro (n=Olipro@uncyclopedia/Olipro) has joined #beagle
  • [01:05:27] <Olipro> anyone know how to get the GNU assembler to evaluate the address for an LDR
  • [01:05:32] <Olipro> like
  • [01:05:50] <Olipro> LDR R0, =0x100+(SomeLabel-0x12224)
  • [01:08:16] <mru> ldr r0, label
  • [01:10:24] <mru> maybe "ldr r0, (expression)" will work
  • [01:11:27] <mru> otherwise you probably have to do ldr r0, [pc, #(expression-.-8)]
  • [01:19:48] <Olipro> mru: LDR R0, =(0x100000-(RelativeAdd+0x15151))
  • [01:19:55] <Olipro> it shits when I do that
  • [01:20:05] <Olipro> any suggestions?
  • [01:20:07] <mru> try without the =
  • [01:20:16] <Olipro> tried it
  • [01:20:28] <Olipro> Error: cannot represent THUMB_OFFSET relocation in this object file format
  • [01:20:47] <mru> what object file format?
  • [01:21:00] <Olipro> I want it to convert that expression into a constant that becomes a value in the binary
  • [01:21:02] <mru> and you didn't mention writing thumb code
  • [01:21:08] <Olipro> there's no file format
  • [01:21:17] <mru> there's always a file format
  • [01:21:18] <Olipro> it doesn't make a difference
  • [01:21:22] <Olipro> I want the compiler to make it a constant
  • [01:21:31] <mru> the error message says it makes a difference
  • [01:21:45] <Olipro> yeah, so not using = is no good
  • [01:21:50] <Olipro> I want it to become a constant
  • [01:22:04] <mru> apparently the location is not constant, so a relocation entry needs to be generated, and the type of relocation required is not available in your selected object file format
  • [01:22:28] <mru> where is the label?
  • [01:22:41] <Olipro> a couple of lines below that operation
  • [01:22:51] <Olipro> and I .org the whole thing to a specific address
  • [01:22:52] <mru> yes, but text or data section?
  • [01:22:57] <Olipro> there is no section
  • [01:23:03] <mru> there always is a section
  • [01:23:53] <mru> you have to somehow convince the assembler that the value is constant
  • [01:24:34] <Olipro> well
  • [01:24:42] <Olipro> if I do LDR R0, =RelativeAdd
  • [01:24:47] <Olipro> then it has no problem creating a constant
  • [01:25:09] <mru> exactly what machine instruction does that turn into?
  • [01:27:11] <Olipro> it writes the address into 4 bytes in the binary
  • [01:27:33] <Olipro> so the LDR references those 4 bytes, containing the address of RelativeAdd
  • [01:27:41] <Olipro> you know, in the .ltorg section
  • [01:28:03] * cmonex (n=xy6091@k9dury4k65.adsl.datanet.hu) has joined #beagle
  • [01:36:02] * Olipro (n=Olipro@uncyclopedia/Olipro) Quit (Nick collision from services.)
  • [01:39:26] * Olipro (n=Olipro@uncyclopedia/Olipro) has joined #beagle
  • [01:39:31] <Olipro> mru are you there
  • [01:39:37] <Olipro> my computer got unplugged
  • [01:39:39] <mru> yes
  • [01:39:40] <Olipro> thanks for that cmonex
  • [01:39:46] <Olipro> :P
  • [01:39:53] <Olipro> anyway, what did you say
  • [01:40:00] <Olipro> last thing I saw was what I wrote
  • [01:40:03] <Olipro> <Olipro> you know, in the .ltorg section
  • [01:40:12] <mru> I've said nothing since that
  • [01:41:09] <mru> so what errors do you get with various ways or writing the expression?
  • [01:42:14] <Olipro> well
  • [01:42:16] <Olipro> interestingly
  • [01:42:20] <Olipro> if I have the label at the start
  • [01:42:28] <Olipro> I can perform any math I like on it
  • [01:42:34] <Olipro> otherwise it craps itself
  • [01:42:57] <Olipro> LDR R0, =(RelativeAdd+0x55454)-0x100)
  • [01:42:59] <Olipro> that'd be fine
  • [01:43:04] <Olipro> it's not what I want to do though
  • [01:44:40] <Olipro> er, 2 (( at the start obviously
  • [01:48:20] * nemequ (n=nemequ@ip68-111-215-155.sd.sd.cox.net) has joined #beagle
  • [01:56:36] * FlipFlop (n=flipflop@146.244.179.23) Quit (Read error: 110 (Connection timed out))
  • [02:02:52] * SpaceAceX (i=Rakso@CPE001310d5071d-CM0011ae01368a.cpe.net.cable.rogers.com) has joined #beagle
  • [02:03:26] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit (Remote closed the connection)
  • [02:03:35] * jo2 (n=jon@host86-138-174-192.range86-138.btcentralplus.com) has left #Beagle
  • [02:04:05] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
  • [02:05:38] <SpaceAceX> .
  • [02:16:55] * geckosenator (n=sean@c-24-8-193-190.hsd1.co.comcast.net) has joined #beagle
  • [02:43:59] * Leon_Nardella (n=chatzill@200-161-14-111.dsl.telesp.net.br) Quit (Remote closed the connection)
  • [03:16:28] * TehUni (n=cory@c-24-30-35-55.hsd1.ga.comcast.net) Quit (Read error: 104 (Connection reset by peer))
  • [03:16:48] * TehUni (n=cory@c-24-30-35-55.hsd1.ga.comcast.net) has joined #beagle
  • [03:24:53] * Beagle6 (n=Beagle1@129.108.41.175) Quit (Read error: 110 (Connection timed out))
  • [03:26:05] * dcordes (n=dcordes_@unaffiliated/dcordes) Quit ("Verlassend")
  • [03:45:07] <gregoiregentil> Has anyone build/run DMAI (leveraging codec-engine) on Beagleboard? Any help/pointer how to do that?
  • [04:01:12] * nmq (n=chatzill@129.108.41.159) Quit ("ChatZilla 0.9.84 [Firefox 2.0.0.14/2008040413]")
  • [04:02:38] * SpaceAceX (i=Rakso@CPE001310d5071d-CM0011ae01368a.cpe.net.cable.rogers.com) Quit (Read error: 110 (Connection timed out))
  • [04:02:43] * SpaceAceX (i=Rakso@CPE001310d5071d-CM0011ae01368a.cpe.net.cable.rogers.com) has joined #beagle
  • [04:15:19] * robclark (n=robclark@166.128.140.36) Quit (Read error: 110 (Connection timed out))
  • [04:31:13] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
  • [04:47:53] * DaQatz (n=db@c-66-30-48-54.hsd1.nh.comcast.net) Quit (Remote closed the connection)
  • [05:11:20] * BThompson (n=BThompso@cpe-76-185-93-11.tx.res.rr.com) Quit (Read error: 104 (Connection reset by peer))
  • [05:12:36] <cmonex> if anyone is interested the solution to olipro's problem was: .set ALbl, AddMePlease - RelativeAdd LDR R0, =ALbl
  • [05:14:25] * SpaceAceX (i=Rakso@CPE001310d5071d-CM0011ae01368a.cpe.net.cable.rogers.com) Quit ()
  • [05:14:41] * BThompson (n=BThompso@cpe-76-185-93-11.tx.res.rr.com) has joined #beagle
  • [05:29:07] * DaQatz (n=db@c-66-30-48-54.hsd1.nh.comcast.net) has joined #beagle
  • [05:31:37] * BThompson (n=BThompso@cpe-76-185-93-11.tx.res.rr.com) Quit ("Trillian (http://www.ceruleanstudios.com")
  • [05:35:25] <Olipro> well, that's a partial description of the solution
  • [05:35:30] <Olipro> I added a label called AddMePlease
  • [05:35:37] <Olipro> that was .org'd to the address I wanted
  • [05:35:44] <Olipro> pretty lame assembler
  • [05:38:42] * TAK2004 (n=Administ@dslb-088-074-045-236.pools.arcor-ip.net) Quit (Read error: 104 (Connection reset by peer))
  • [06:18:09] * Olipro is now known as sergey_brin
  • [06:18:34] * sergey_brin is now known as Olipro
  • [06:25:01] * RobotGuy (n=robotguy@pool-96-225-210-142.ptldor.fios.verizon.net) Quit (Remote closed the connection)
  • [06:27:12] <TehUni> Olipro from xda-devs too?
  • [06:27:16] <TehUni> lots of you guys here, heh
  • [06:27:26] <Olipro> lol
  • [06:28:21] <TehUni> already thanked dcordes, thanks to you for hardspl :)
  • [06:29:30] <Olipro> we have a channel on here too you know
  • [06:29:33] <Olipro> #xda-devs
  • [06:29:55] <TehUni> i've kinda abandoned the wince platform, but thanks
  • [06:30:16] <TehUni> just hoping for android on the kaiser now
  • [06:37:30] * mrc3 (n=ddiaz@189.157.115.100) Quit (Read error: 104 (Connection reset by peer))
  • [06:37:31] * mrc3_ (n=ddiaz@189.157.115.100) has joined #beagle
  • [06:37:32] <geckosenator> me neither
  • [06:52:40] * vijay (n=vijay@203.199.213.3) has joined #beagle
  • [07:56:52] * cmonexaway (n=xy6091@k9dury4k65.adsl.datanet.hu) has joined #beagle
  • [08:01:44] * cmonex (n=xy6091@k9dury4k65.adsl.datanet.hu) Quit (Nick collision from services.)
  • [08:01:48] * cmonexaway is now known as cmonex
  • [08:44:41] <cmonex> TehUni: which device do you have?
  • [08:45:06] <cmonex> ah kaiser
  • [08:48:03] * Fl1pFl0p (n=flipflop@ip68-8-230-62.sd.sd.cox.net) Quit ("server is back up")
  • [08:52:08] * dirk2 (n=dirk@92.117.92.202) has joined #beagle
  • [08:58:36] * koen wonders why people either don't read errors messages or just the top 5 lines
  • [09:12:23] <zuh> koen: Because they get the answer easier by asking people like you?-)
  • [09:13:18] <koen> if "snide reply" is "answer", then yes :)
  • [09:14:00] <koen> normally I have the rule that if the same question gets asked over and over again something is wrong and needs better docs
  • [09:14:25] <koen> in this case people keep ignoring the actual part of the error message
  • [09:15:09] * dirk2 (n=dirk@92.117.92.202) has left #beagle
  • [09:22:20] <zuh> snide replies are usually the correct answer to those people ;)
  • [09:33:41] <koen> I appreciate that people make an effort to guess the error, but I'd appreciate it more if they did that *after* reading the error message
  • [09:42:25] <koen> zuh: did you say someone was working on some armv6 asm for xf86-video-omapfb?
  • [09:44:44] <zuh> LoneMeow was at least, but I never got his code for integration ;)
  • [09:48:24] <koen> zuh: gregoiregentil adapted the omapfbplay code for mplayer: http://dominion.thruhere.net/git/?p=openembedded.git;a=commitdiff;h=59f5a9c741bba79e6ebf8e6527142aeeb8ea9236
  • [09:50:10] <gregoiregentil> Koen: did you go further regarding DSP? Meaning: is there any example to play with, once you get the dsplink and codec-engine packages?
  • [09:50:43] <koen> gregoiregentil: OE should have built the dsplink and CE examples
  • [09:50:53] <koen> gregoiregentil: ringio*, etc
  • [09:52:05] <gregoiregentil> yes, I managed to build dsplink and CE. what is ringio?
  • [09:52:29] <koen> it's a test people seem to like running to test dsplink
  • [09:52:58] <koen> gregoiregentil: I suspect you'd get better answers on the beagle mailinglist :)
  • [09:53:08] <gregoiregentil> ringiogpp I guess
  • [09:53:14] <gregoiregentil> I will try
  • [09:53:39] <gregoiregentil> I think that DMAI would be an interesting package to create. I tried to write a recipe but I had some troubles
  • [09:55:17] * valhalla (n=valhalla@81-174-21-31.dynamic.ngi.it) has joined #beagle
  • [09:55:53] <koen> I'm currently blocking on TI to make clear what pieces of CE are of a certain license and how they want to have it distributed as .ipk
  • [09:57:00] <koen> since TI does seem to want to have installable packages, but seem to be caught in some theoretical discussion about licenses and not getting email-addresses from people downloding the .ipks
  • [09:57:14] <gregoiregentil> I saw the discussion with sjam and rogermon who may have a full blown CE
  • [09:57:50] <gregoiregentil> I didn't see the email of sjam on the mailing list as he wanted to do
  • [09:58:13] <koen> this reminds me of when I added a build recipe for sun cvm, but wasn't allowed to distribute the resulting binary
  • [09:59:30] <gregoiregentil> oops!
  • [10:00:04] <gregoiregentil> but the current situation is not too bad. you have the recipes, you put the link inside the recipes
  • [10:00:15] <koen> it ended up in setting every student in front of an ssh session and saying "press enter to build it"
  • [10:00:17] <gregoiregentil> and people downloads and compile (after accepting the license on TI website)
  • [10:00:43] <koen> right
  • [10:01:26] <koen> but some people in TI want to give users ready made .ipk packages
  • [10:01:37] <koen> for e.g. CE powered vlc
  • [10:03:01] <koen> gregoiregentil: 'opkg install dsplink' will get you binaries of the gpl and bsd bits
  • [10:13:25] <koen> gregoiregentil: both omapfbplay and vo_omapfb give a green picture with dss2
  • [10:13:53] <gregoiregentil> Cool! ;-)
  • [10:13:59] <gregoiregentil> I'm kidding...
  • [10:17:10] <gregoiregentil> I will try later DSS2 and play with the plane initialization of vo_omapfb. But I doubt that I will manage to fix it. Mru will be the man of the situation...
  • [10:19:12] <koen> gregoiregentil: doing 'fbset -fb /dev/fb1 -nonstd 1' during playback solved the problem on my omap3evm board
  • [10:19:47] <gregoiregentil> I'm very confused by the plane color type...
  • [10:20:55] <gregoiregentil> I don't understand if we are really outputting in YUV422 or YUY422
  • [10:21:37] <gregoiregentil> Mru's code mentioned to yuv422 but with DSS1, the plane is used with YUY422
  • [10:22:08] <gregoiregentil> when I developed my vo_omapfb on DSS1, I used plane YUV422 and it was not working. It was fine with plane YUY422
  • [10:22:24] <gregoiregentil> now you are saying that DSS2 works with plane YUV422
  • [10:22:42] <gregoiregentil> I'm just wondering if there is not a bug in DSS1 about the color format? Am I making sense?
  • [10:22:53] <koen> or a bug in dss2
  • [10:23:04] <koen> could be that someone switched nonstd 1 and 8
  • [10:23:21] <koen> I guess tomba, mru or zuh might know :)
  • [10:23:35] <mru> when my code says yuv422 it is meant in the generic sense
  • [10:23:47] <mru> i.e. it's a yuv format with 4:2:2 subsampling
  • [10:23:57] <mru> the driver may call it something else
  • [10:24:09] <mru> there are many possible permutations of such formats
  • [10:26:05] <gregoiregentil> I'm confused between yuy422 and yuv422. I don't understand well the difference. It's very confusing. Anyway, one can try the different color formats and find the one that works
  • [10:26:26] <zuh> YUV format names *are* confusing :)
  • [10:26:32] <gregoiregentil> but for sure DSS1 and DSS2 doesn't seem to have the same idea of YUV
  • [10:26:50] <gregoiregentil> err: don't
  • [10:27:42] <gregoiregentil> when you read the OMAP35xx spec, YUV wikipedia, mplayer header files, you really need the box of advil pills...
  • [10:28:07] <zuh> fourcc.org is a good source of information too
  • [10:29:09] <koen> http://fourcc.org/yuv.php
  • [10:29:19] * chelli (n=chelli@p54B87B82.dip.t-dialin.net) has joined #beagle
  • [10:30:09] <koen> could it be that DSS1 calls YUV422 'nonstd 1' and YUY422 'nonstd 8'?
  • [10:30:32] <koen> and DSS@ calls YUV422 'nonstd 8' and YUY422 'nonstd1'?
  • [10:30:32] <gregoiregentil> yes. reading it... there is also something that I have not understood. Why do codecs output yuv420 and hardware like OMAP only support yuv422. Just the pleasure of adding unnecessary CPU-intensive color conversion? and giving us headache?
  • [10:31:14] <koen> AIUI 422 was a thinko on TIs side
  • [10:31:27] <koen> or maybe TI has 1337 dsp codeds that work better with 422
  • [10:34:05] <koen> having 12bpp is nice, but kinda moot if it costs you memory bandwidth to convert it from 420
  • [10:35:29] <gregoiregentil> I read the discussion about 12bpp in the mail thread. We are far from the end...
  • [10:35:59] <gregoiregentil> Mru: what did you stop of trying to leverage DSP in omapfbplay? Source code (like DMAI) is available and shows how to decode with DSPlink and CE? Am I missing something?
  • [10:38:34] <koen> ok
  • [10:38:39] <koen> major weirdness
  • [10:38:58] <koen> for fixing the colours in xf86-video-omapfb I need: "fbset -fb /dev/fb1 -nonstd 8"
  • [10:39:59] * koen is confused now
  • [10:40:05] * RogerMonk (n=a0740758@nat/ti/x-96bc58e8004aa0c3) has joined #beagle
  • [10:40:27] <koen> gregoiregentil: your dsp guru has arrived :)
  • [10:40:38] <gregoiregentil> Yes. I saw it!
  • [10:40:45] <gregoiregentil> Ping RogerMonk?
  • [10:40:46] <koen> RogerMonk: good mornin!
  • [10:40:56] <koen> RogerMonk: you are in high demand lately :)
  • [10:41:17] <zuh> koen: fb.nonstd = rand(1,8);
  • [10:41:19] <zuh> ;)
  • [10:41:25] <RogerMonk> morning guys!
  • [10:41:46] <gregoiregentil> Hello
  • [10:41:58] <gregoiregentil> RogerMonk: Do you have a minute to discuss?
  • [10:42:17] <gregoiregentil> we were chatting about DSP and the word is that you are the real expert :-)
  • [10:42:25] <RogerMonk> can we sync up a bit later today - got to rush our right now - happy to help though
  • [10:43:05] <gregoiregentil> Sure. No problem! As long as you can help us later ;-)
  • [10:43:19] <RogerMonk> I'll try :)
  • [10:44:01] * ldesnogu_ (n=ldesnogu@ven06-2-82-247-86-183.fbx.proxad.net) has joined #beagle
  • [10:54:25] * vijay (n=vijay@203.199.213.3) Quit ("leaving")
  • [10:56:20] <koen> zuh, gregoiregentil I suspec that DSS2 has a different view of formats as DSS1
  • [10:56:37] <koen> zuh, gregoiregentil: but I have no clue which is the correct one
  • [10:57:17] * likewise (n=chatzill@82-171-51-231.ip.telfort.nl) has joined #beagle
  • [10:59:18] * vu3rdd (n=user@61.8.208.80) has joined #beagle
  • [11:06:53] <koen> hmmm
  • [11:07:04] <koen> DSS2 uses the same enum as DSS1
  • [11:07:13] <koen> http://dominion.thruhere.net/git/?p=linux-omap.git;a=blob;f=arch/arm/plat-omap/include/mach/omapfb.h#l83
  • [11:07:20] <koen> except that it adds a few 32bit modes
  • [11:18:08] * koen is puzzled
  • [11:18:26] <koen> omapfbplay, mplayer and xf86-video omapfb all use the same enum
  • [11:19:07] <koen> which for the YUV formats is the same in DSS1 and DSS2
  • [11:23:16] * geckosenator (n=sean@c-24-8-193-190.hsd1.co.comcast.net) Quit (Read error: 110 (Connection timed out))
  • [11:24:28] <koen> aaaah
  • [11:24:33] <koen> DSS2 has a fb_mode_to_dss_mode method
  • [11:28:01] * abitos (n=nixgibts@dslb-084-057-154-061.pools.arcor-ip.net) has joined #beagle
  • [11:54:05] * vu3rdd (n=user@61.8.208.80) Quit (Read error: 113 (No route to host))
  • [12:02:21] * Olipro (n=Olipro@uncyclopedia/Olipro) Quit ("Leaving")
  • [12:02:46] <tomba> do I have the modes wrong way?
  • [12:03:05] <koen> tomba: no idea on that
  • [12:03:18] <koen> tomba: it could be that you have the modes right and dss1 has them wrong
  • [12:04:44] <tomba> I don't know why dss1 didn't name the modes as they were in omap TRM, but I didn't want to change the naming, of course. thats why I convert them from fb modes to dss modes
  • [12:11:37] <RogerMonk> ping gregoiregentil
  • [12:14:06] <koen> tomba: it seems that dss2 reverses nonstd 1 and 8 wrt to DSS1
  • [12:15:01] <koen> tomba: and I say "seems" since I haven't looked into it deeply
  • [12:17:22] * DJWillis (i=djwillis@82-46-19-72.cable.ubr02.bath.blueyonder.co.uk) Quit ("Manny: It's my scythe. I like to keep it next to where my heart used to be.")
  • [12:20:19] <mru> koen, gregoiregentil: video codecs output planar yuv420 because that's what is encoded in the bitstream
  • [12:20:32] <mru> it is more efficient to encode the planes separately
  • [12:21:46] <mru> and chroma subsampling is a simple way to reduce the amount of data without too much visual impact
  • [12:23:45] <koen> if we re-encode a yuv420 h264 movie into yuv422 h264 we would get more fps, right?
  • [12:23:54] <koen> fps as in decode speed
  • [12:23:58] <mru> I'd expect less
  • [12:24:12] <mru> insofar h264 even can encode 422
  • [12:24:40] <koen> because h264 decoding would take more cpu than decoding+ colour conversion
  • [12:24:50] <mru> that's what I'd expect
  • [12:25:07] <koen> what does bluray use?
  • [12:25:10] <koen> mpeg4?
  • [12:25:35] * koen googles it
  • [12:25:39] <mru> bluray uses mpeg2, h264 or vc1
  • [12:27:55] * koen reads up on mpeg4
  • [12:28:51] <koen> what's I've been calling mpeg4 is actually mpeg4 asp
  • [12:28:58] <mru> exactly
  • [12:29:06] <mru> mpeg4 avc, aka mpeg4 part 10 is h264
  • [12:29:16] <koen> and asp is part 2
  • [12:29:26] <koen> and sort of h263
  • [12:29:50] <koen> wikipedia quotes ffmpeg on the asp page :)
  • [12:29:56] <koen> http://en.wikipedia.org/wiki/MPEG-4_Part_2
  • [12:32:53] <mru> mpeg4 asp is largely based on h263
  • [12:37:36] <mru> h264 allows 422 format only in high 4:2:2 profile
  • [12:37:41] <mru> not in base or main
  • [12:37:48] * notaz_ (n=notaz@78.59.134.74) has joined #beagle
  • [12:40:21] <notaz_> does anyone know if audio recording works with current l-o kernel?
  • [12:40:57] * Viral_Sachde (n=viral@122.167.215.107) has joined #beagle
  • [12:47:06] * hli (i=chaton@vig91-2-82-232-97-149.fbx.proxad.net) has joined #beagle
  • [12:51:07] * dirk2 (n=dirk@92.116.38.17) has joined #beagle
  • [12:52:22] <jkridner> good morning all
  • [13:02:22] <likewise> gm
  • [13:02:23] * Viral_Sachde (n=viral@122.167.215.107) Quit (Read error: 104 (Connection reset by peer))
  • [13:09:51] * jccu (n=johanes@schwarzenegger.paivola.fi) has joined #beagle
  • [13:11:20] * dirk2 (n=dirk@92.116.38.17) Quit (Remote closed the connection)
  • [13:18:45] <koen> hey jkridner & likewise
  • [13:18:59] <likewise> hey koen et all
  • [13:19:02] <likewise> -l
  • [13:19:11] <jkridner> -l?
  • [13:19:36] <likewise> jkridner: http://www.thefreedictionary.com/et%20al.
  • [13:20:05] <jkridner> ah, one extra 'l'.
  • [13:20:06] <likewise> either "and all" or "et al" - but who cares.
  • [13:20:19] <jkridner> I knew "et al", but was confused by the -l.
  • [13:20:19] <koen> y'all :)
  • [13:20:23] <koen> or all y'all
  • [13:20:41] * koen doesn't remember if y'all is singular or plural
  • [13:20:41] <jkridner> all y'all both ;)
  • [13:20:42] <likewise> howya do'ng y'all?
  • [13:21:10] <likewise> remember we are dutch. we can break havoc on anything language-related.
  • [13:21:16] <koen> yeah
  • [13:21:21] * jccu (n=johanes@schwarzenegger.paivola.fi) has left #beagle
  • [13:21:27] <koen> the dutch translation misspells 'failed'
  • [13:21:32] <jkridner> ist gut
  • [13:21:40] <koen> 'gevaald' instead of 'gefaald'
  • [13:21:54] * Viral_Sachde (n=viral@122.167.215.107) has joined #beagle
  • [13:22:07] <jkridner> what translation software was used?
  • [13:22:18] * koen has no idea
  • [13:22:38] <likewise> jkridner: "ist gut" is german, not dutch. ("is goed").
  • [13:23:05] <jkridner> I used to make use of babelfish, but I've started using Google a bit more now.
  • [13:23:29] <jkridner> then, is goed.
  • [13:24:01] <jkridner> 2 year old german is as close as I can get to dutch.
  • [13:26:39] * flo_lap (n=fuchs@g228133152.adsl.alicedsl.de) has joined #beagle
  • [13:27:40] <jkridner> anyway, the goal was to encourage you to continue out of my own curiousity, but I guess that gefaald.
  • [13:28:06] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) has joined #beagle
  • [13:30:45] <jkridner> in the next couple of weeks or so, we hope to update the u-boot used on the Beagle boards to include the L1NEON patch and generally be close to what is getting released from the mainline.
  • [13:31:14] <jkridner> would be a lot more fun to play with OpenCV and Pico-based vision apps, or even just ALSA.
  • [13:31:23] <jkridner> er, ASoC.
  • [13:31:42] <koen> jkridner: what's the holdup of using sakomans tree + splash patch?
  • [13:32:26] <jkridner> I think that's where khasim is headed, but I think the mailing list or this channel is probably more up-to-date than I am.
  • [13:32:27] <koen> or is that the "close to what is getting released from the mainline" bit :)
  • [13:32:44] <jkridner> well, the u-boot-arm is closer.
  • [13:33:02] * koen is using sakomans tree + mrus 600MHz patch
  • [13:33:35] <jkridner> so, I think it would be j24's u-boot-arm tree with patches coming from sakoman's tree (so that the patches are more isolated).
  • [13:34:13] <jkridner> I really want the script stuff from sakoman's tree, but that I think that is lower priority than being close to the mainline.
  • [13:35:08] <jkridner> I managed to get an hour or two in on struct clean-up to handle w-denx requirements, but was then quickly derailed.
  • [13:35:22] <jkridner> working on an article now.
  • [13:38:23] <jkridner> doesn't seem like a lot of traffic here, but another thing I need to talk about is the idea of having some contests.
  • [13:39:01] <jkridner> I think I can get the TI marketing folks to give away some boards as part of a demo design contest.
  • [13:39:18] <jkridner> I'm a bit concerned that getting a give-away board might not be so great, since you'd have to have one already to create the demo in the first place.
  • [13:40:35] <koen> maybe coordinate with the pandora people
  • [13:41:04] <koen> but most demos don't need a beagle
  • [13:41:17] <koen> unless you want do use 1337 neon asm
  • [13:41:39] <jkridner> well, using the DSP would likely be bonus points for TI.
  • [13:41:44] <koen> there's a simulator for the sgx and for other stuff people could use SDL or X
  • [13:42:14] <koen> yeah, using the dsp would require a real board
  • [13:42:32] <jkridner> getting Link, Codec Engine, etc. out there is a bit of my motivation for the timing.
  • [13:42:58] <jkridner> I'm also expecting to have something better for the 3D "very soon" (TM)
  • [13:43:58] <jkridner> by better, I mean easier to work with newer kernels.
  • [13:44:13] <jkridner> I'm not anxious to promote any 3D drivers bound to an old kernel.
  • [13:44:25] <koen> if someone would come up with an actual proposal to handle link and CE packaging in OE I can easily include them in the demo images
  • [13:44:56] <jkridner> *I* (me personally, not as a representative of TI) think the current packaging is fine.
  • [13:45:51] <jkridner> I don't really know what it would take to nail us down to an official position. You have the legal info.
  • [13:46:11] <jkridner> you cannot get TI to interpret the contract info stuff.
  • [13:50:53] <jkridner> I don't know if you are using code that isn't under the BSD or GPLv2 licenses, but for the code that is under those licenses, I think they have been publicly interpreted many times.
  • [13:52:15] <jkridner> as I recall, there is a big click-wrap license on the software, and at the bottom there is an exception for certain chunks of code to be under the BSD/GPL licenses...
  • [13:52:32] <jkridner> I think those exceptions cover the majority of the code.
  • [13:56:06] <koen> I'm not sure how to handle the dsp sample
  • [13:56:15] <koen> since the code is bsd but gets linked with bios
  • [13:56:19] <koen> AIUI, I could be wrong
  • [13:58:02] * dcordes (n=dcordes_@unaffiliated/dcordes) has joined #beagle
  • [14:00:43] <dcordes> koen, do you get problems like 4886 with armv6-novfp ?
  • [14:02:16] <koen> let me try, I only built gcc-cross-initial
  • [14:02:58] <dcordes> ok thanks
  • [14:04:03] <koen> dcordes: it could be that we'd need to remove the hyphen
  • [14:04:15] <koen> I think I did that for armv7-a as well
  • [14:04:22] * koen can't rememeber
  • [14:05:49] <dcordes> koen, which one?
  • [14:07:49] <koen> the one in armv7-novfp
  • [14:07:51] <koen> ehm
  • [14:07:53] <koen> v6
  • [14:08:57] * jrmuizel (n=jrmuizel@CPE001f5be79d0f-CM0017ee62f8b0.cpe.net.cable.rogers.com) Quit ()
  • [14:09:03] <Crofton|work> jkridner, at one point TI had an academic design competition for the DSP products
  • [14:09:06] <dcordes> koen, where do I need to change this
  • [14:11:50] <jkridner> Crofton|work: how did that go?
  • [14:12:31] <Crofton|work> very well, when I started grad school at Virginia Tech, they had just won :)
  • [14:12:37] * Viral_Sachde (n=viral@122.167.215.107) Quit (Read error: 104 (Connection reset by peer))
  • [14:12:38] <Crofton|work> so I am biased :)
  • [14:12:54] <Crofton|work> this would have been circa 1999 I think
  • [14:13:11] * Crofton|work was in grad school for a long time
  • [14:13:30] <Crofton|work> I can ask who at TI was involved if you like
  • [14:13:39] * dcordes_ (n=dcordes_@unaffiliated/dcordes) has joined #beagle
  • [14:14:26] * shriram (i=3d5fc3cd@gateway/web/ajax/mibbit.com/x-0e0bcbec4949c3b1) has joined #beagle
  • [14:17:07] <dcordes_> koen, ping
  • [14:19:08] <koen> dcordes_: pong
  • [14:19:21] * likewise needs some gaming activities
  • [14:19:22] <dcordes_> koen, where would I change that to try?
  • [14:19:24] * likewise (n=chatzill@82-171-51-231.ip.telfort.nl) Quit ("ChatZilla 0.9.84 [Firefox 3.0/2008061015]")
  • [14:20:19] <Crofton|work> hmm, should I go play with the PS3
  • [14:21:10] <koen> dcordes_: in the tunefile, but my glibc build seems to have succeeded
  • [14:22:17] <dcordes_> koen, that's odd. after your commits I cleaned tmp and tried the htcraphael builds.
  • [14:22:28] * likewise (n=likewise@82-171-51-231.ip.telfort.nl) has joined #beagle
  • [14:24:10] <koen> NOTE: package glibc-2.6.1-r15: task do_package: started
  • [14:25:02] * shriram (i=3d5fc3cd@gateway/web/ajax/mibbit.com/x-0e0bcbec4949c3b1) Quit ("http://www.mibbit.com ajax IRC Client")
  • [14:27:24] <dcordes_> hmm
  • [14:27:25] <dcordes_> checking for arm-angstrom-linux-gnueabi-gcc... arm-angstrom-linux-gnueabi-gcc -march=armv6j -mtune=arm1136jf-s
  • [14:27:25] <dcordes_> checking for suffix of object files... configure: error: cannot compute suffix of object files: cannot compile
  • [14:29:44] * dcordes (n=dcordes_@unaffiliated/dcordes) Quit (Read error: 110 (Connection timed out))
  • [14:29:53] * Beagle2 (n=Beagle2@e50109.upc-e.chello.nl) has joined #beagle
  • [14:30:47] * Beagle2 (n=Beagle2@e50109.upc-e.chello.nl) Quit (Client Quit)
  • [14:30:57] * Beagle8 (n=Beagle8@e50109.upc-e.chello.nl) has joined #beagle
  • [14:31:28] * Beagle8 (n=Beagle8@e50109.upc-e.chello.nl) Quit (Client Quit)
  • [14:33:02] * Viral_Sachde (n=viral@122.167.215.107) has joined #beagle
  • [14:34:32] <koen> dcordes_: try a clean rebuild with the commits I just pushed
  • [14:35:56] <dcordes_> ok
  • [14:39:38] * flo_lap (n=fuchs@g228133152.adsl.alicedsl.de) Quit ("Verlassend")
  • [14:41:48] * Viral_Sachde (n=viral@122.167.215.107) Quit (Read error: 104 (Connection reset by peer))
  • [14:43:25] * Viral_Sachde (n=viral@122.167.215.107) has joined #beagle
  • [14:49:42] * vijay_ (n=vijay@203.199.213.3) has joined #beagle
  • [14:58:20] <dcordes_> koen, looks good
  • [14:58:34] <koen> tomba, zuh: can you think of any pitfalls in teaching X to like the 32bit mode?
  • [14:59:47] * zuh doesn't know about the implications enough to comment really
  • [15:00:07] <koen> or rather 24 bit mode
  • [15:00:52] <koen> the fb console doesn't seem to like the 32 bit mode and gives funky colours
  • [15:00:52] * Viral_Sachde (n=viral@122.167.215.107) Quit (Read error: 104 (Connection reset by peer))
  • [15:00:58] <zuh> I guess it should be okay, provided that the setup is done properly
  • [15:01:12] <zuh> all the color masks and such
  • [15:01:13] <koen> I'm doing an fbset -16 early at boot
  • [15:02:10] <zuh> Since x-v-omapfb not really touching the data anyway, it's just letting the libfb do it's thing
  • [15:05:39] <dcordes_> koen, closed
  • [15:12:50] <tomba> koen, I think there's a bpp == 16 check in one or two places in omapfb. just remove the check and it worked
  • [15:14:44] <tomba> I mean Xomap, not omapfb
  • [15:15:53] <koen> I'm not using Xomap
  • [15:16:51] <koen> I used Xfbdev from kdrive, Xorg with xf86-video-fbdev and xf86-video-omapfb
  • [15:20:32] * Viral_Sachde (n=viral@122.167.221.240) has joined #beagle
  • [15:21:04] <zuh> koen: I suppose the only place x-v-omapfb cares is the call to xf86SetDepthBpp()
  • [15:21:07] <tomba> well, X should work fine. I don't know about xf86-video-omapfb
  • [15:23:02] <tomba> at least Xomap worked fine with rgb16, rgb24 packed and rgb24 unpacked modes, after I removed the check
  • [15:24:59] <zuh> http://www.xfree86.org/current/DESIGN18.html <- xf86SetDepthBpp() seems to even have some conversion flags...
  • [15:25:15] * ClaudeQC (n=claude@bas1-quebec03-1279637127.dsl.bell.ca) has joined #beagle
  • [15:35:30] * Openfree (n=df@222.65.143.118) Quit (Remote closed the connection)
  • [15:43:55] * _AV500_ (n=av500@p50996ded.dip0.t-ipconnect.de) has joined #beagle
  • [15:44:48] * valhalla (n=valhalla@81-174-21-31.dynamic.ngi.it) Quit (kubrick.freenode.net irc.freenode.net)
  • [15:44:48] * mrc3_ (n=ddiaz@189.157.115.100) Quit (kubrick.freenode.net irc.freenode.net)
  • [15:44:48] * DaQatz (n=db@c-66-30-48-54.hsd1.nh.comcast.net) Quit (kubrick.freenode.net irc.freenode.net)
  • [15:44:48] * Crofton (n=balister@66-207-66-26.black.dmt.ntelos.net) Quit (kubrick.freenode.net irc.freenode.net)
  • [15:44:48] * AV500 (n=av500@p50996ded.dip0.t-ipconnect.de) Quit (kubrick.freenode.net irc.freenode.net)
  • [15:51:05] * dcordes_ (n=dcordes_@unaffiliated/dcordes) Quit ("Verlassend")
  • [15:51:12] * likewise (n=likewise@82-171-51-231.ip.telfort.nl) Quit ()
  • [15:57:48] * DaQatz (n=db@c-66-30-48-54.hsd1.nh.comcast.net) has joined #beagle
  • [15:58:15] * Crofton (n=balister@66-207-66-26.black.dmt.ntelos.net) has joined #beagle
  • [16:02:05] * valhalla (n=valhalla@81-174-21-31.dynamic.ngi.it) has joined #beagle
  • [16:02:13] * JuanG (n=Juan@nat/ti/x-5a6fe9eb7f633850) has joined #beagle
  • [16:02:17] * JuanG (n=Juan@nat/ti/x-5a6fe9eb7f633850) has left #beagle
  • [16:13:23] * mrc3_ (n=ddiaz@189.157.115.100) has joined #beagle
  • [16:21:18] * dirk2 (n=dirk@92.116.232.232) has joined #beagle
  • [16:26:29] <dirk2> jkridner: actim registers: have a look to pdf page 1397 (table 11-156) of spruf98b.
  • [16:27:27] <dirk2> jkridner: with p = 1, all registers have offset of 0x30, but actim registers are 0x28.
  • [16:28:46] <dirk2> jkridner: I have no idea what hw guys think doing something like this ;(
  • [16:29:15] <jkridner> oh, freakish.
  • [16:29:28] <jkridner> I didn't notice the switch from 0x30 to 0x28.
  • [16:29:51] <jkridner> makes no sense whatsoever.
  • [16:30:32] <dirk2> Some minutes I thought the TRM might be wrong, but original U-Boot has this, too
  • [16:33:16] <dirk2> jkridner: Thanks for helping! Will send both patches to U-Boot list, soon.
  • [16:59:47] * FuL|OUT is now known as fulgas
  • [17:00:35] * dirk2 (n=dirk@92.116.232.232) has left #beagle
  • [17:02:22] * soman (n=somnath@122.169.162.164) has joined #beagle
  • [17:19:47] * dcordes (n=dcordes_@unaffiliated/dcordes) has joined #beagle
  • [17:21:27] <dcordes> koen, the msm kernels don't work with the defconfigs. instead of packages/linux/linux-msm7xxxx/<machine>/defconfig bitbake tries to copy packages/linux/linux-msm7xxxx/defconfig
  • [17:23:26] <dcordes> koen, the subfolders with the machine names are not correct. they all have htc prepended
  • [17:29:01] * uwe__ (n=uwe@dslb-088-064-102-078.pools.arcor-ip.net) has joined #beagle
  • [17:38:13] * macneib (n=macneib@70.29.45.22) has joined #beagle
  • [17:40:33] * uwe2 (n=uwe@dslb-088-064-072-207.pools.arcor-ip.net) Quit (Read error: 110 (Connection timed out))
  • [17:45:20] * Viral_Sachde (n=viral@122.167.221.240) Quit (Read error: 131 (Connection reset by peer))
  • [17:46:43] * Viral_Sachde (n=viral@122.167.221.240) has joined #beagle
  • [17:51:33] * Viral_Sachde (n=viral@122.167.221.240) Quit (Read error: 104 (Connection reset by peer))
  • [17:58:16] * BThompson (n=BThompso@cpe-76-185-93-11.tx.res.rr.com) has joined #beagle
  • [18:01:04] * nemequ (n=nemequ@ip68-111-215-155.sd.sd.cox.net) Quit (Read error: 145 (Connection timed out))
  • [18:02:22] * sjam (n=sojan_ja@121.247.121.94) has joined #beagle
  • [18:03:10] <sjam> Hi koen
  • [18:04:02] <sjam> I see that you've done the .bb script for the CodecEngine build on OE, right?
  • [18:09:21] <koen> with a lot of help from RogerMonk
  • [18:09:42] <sjam> ah
  • [18:10:12] <sjam> Well, going to post a question on the group.
  • [18:11:13] * Viral_Sachde (n=viral@122.167.221.240) has joined #beagle
  • [18:17:43] <gregoiregentil> sjam: you should read the logs of the IRC for the past 24 hours
  • [18:18:19] <sjam> i will do that right away
  • [18:19:46] <sjam> hmm, gregoir.. , you seem to have managed to build the CE examples.
  • [18:20:00] <sjam> did you use the OE recipes or did you manually make the changes?
  • [18:22:46] <sjam> I use XDC ver 3_00_06, CE 2.21 and I get an error from XDC, cannot find package 'gnu.targets.arm".
  • [18:23:09] <sjam> I also don't see this in /usr/local/xdc_3_00_06/packages/gnu/targets
  • [18:24:56] <gregoiregentil> yes, I managed to build the recipes though it has been a little bit painful
  • [18:25:26] <gregoiregentil> there was a small bug in CE. Koen posted the fix in the chat
  • [18:28:10] <sjam> still going through the logs, Oh BTW, I just have the CE downloaded from the TI website, from what I understand, all the code is in the package.
  • [18:30:32] <koen> yes, CE has everything it needs inside, link, lpm, cmem, etc
  • [18:30:44] <koen> but not codegen, xdc and bios
  • [18:32:53] <sjam> koen: sorry, I cannot seem to find the fix you posted for the CE bug
  • [18:33:32] <gregoiregentil> http://gitweb.openembedded.net/?p=openembedded.git;a=commitdiff;h=bfa38e42af250182a019f702379a459a92006fa1
  • [18:35:40] <sjam> thanks
  • [18:36:47] * nemequ (n=nemequ@ip68-111-215-155.sd.sd.cox.net) has joined #beagle
  • [18:40:41] * abitos (n=nixgibts@dslb-084-057-154-061.pools.arcor-ip.net) Quit (Nick collision from services.)
  • [18:40:59] * abitos (n=nixgibts@dslb-084-057-158-020.pools.arcor-ip.net) has joined #beagle
  • [18:43:35] <sjam> gregoire: which version of XDC are you using?
  • [18:44:28] <sjam> i get the error - "can't find package 'gnu.targets.arm"
  • [18:45:05] <sjam> any changes required to the XDCPATH? It seems to be set correctly though
  • [18:46:29] * cmonex (n=xy6091@k9dury4k65.adsl.datanet.hu) Quit (Read error: 60 (Operation timed out))
  • [18:46:33] <gregoiregentil> bios_setuplinux_5_33_02.bin TI-C6x-CGT-v6.0.16-eval.bin codec_engine_2_21.tar.gz dmai_1_10_00_06.tar.gz xdctools_setuplinux_3_10_03.bin dsplink_1_60.tar.gz
  • [18:48:52] * mrc3_ (n=ddiaz@189.157.115.100) Quit (Read error: 104 (Connection reset by peer))
  • [18:49:11] <sjam> Mine is XDCTools ver 3_00_06, BIOS - 5_33_03, CodeGen - 6_0_15. I think I should try downgrading my XDCTools
  • [18:49:14] * mrc3_ (n=ddiaz@189.157.115.100) has joined #beagle
  • [18:49:46] * gus2u (n=user@67-40-46-247.hlrn.qwest.net) has joined #beagle
  • [18:51:08] <sjam> This is something for tomorrow, thanks gregoire and koen, I'm sure I'll be back with some more questions.
  • [18:51:34] <gregoiregentil> OK. IN the mean time, I will try to ping RogerMon to continue improving the recipes if possible
  • [18:51:50] <gregoiregentil> Ping RogerMon?
  • [18:52:14] <gregoiregentil> and ping RogerMonk?!
  • [18:53:08] <sjam> TI says that with the new versions of Codec Engine, it should be possible to to run two processes concurrently, using the same DSP Server. I'm trying to understand how that works, and maybe also test it.
  • [18:55:41] * dcordes (n=dcordes_@unaffiliated/dcordes) Quit ("Verlassend")
  • [18:56:15] * sjam (n=sojan_ja@121.247.121.94) Quit ()
  • [19:03:32] * Viral_Sachde (n=viral@122.167.221.240) Quit (Read error: 104 (Connection reset by peer))
  • [19:05:33] * Viral_Sachde (n=viral@122.167.221.240) has joined #beagle
  • [19:08:27] * Viral_Sachde (n=viral@122.167.221.240) has left #beagle
  • [19:10:49] * koen (n=koen@s55917625.adsl.wanadoo.nl) Quit (kubrick.freenode.net irc.freenode.net)
  • [19:10:49] * surki (n=surki@gek7.kyla.fi) Quit (kubrick.freenode.net irc.freenode.net)
  • [19:10:49] * jkridner (n=jason@c-76-31-18-64.hsd1.tx.comcast.net) Quit (kubrick.freenode.net irc.freenode.net)
  • [19:10:49] * dcramer (n=davec@dcdsl.ebox.com) Quit (kubrick.freenode.net irc.freenode.net)
  • [19:10:49] * kmwang (n=angelic4@mail.CDPA.cc) Quit (kubrick.freenode.net irc.freenode.net)
  • [19:10:53] * geckosenator (n=sean@c-24-8-193-190.hsd1.co.comcast.net) has joined #beagle
  • [19:11:21] <geckosenator> anyone working on expansion boards?
  • [19:11:36] * kmwang (n=angelic4@mail.CDPA.cc) has joined #beagle
  • [19:12:16] * koen (n=koen@s55917625.adsl.wanadoo.nl) has joined #beagle
  • [19:13:28] * soman (n=somnath@122.169.162.164) Quit ("Leaving")
  • [19:15:37] * dcramer (n=davec@dcdsl.ebox.com) has joined #beagle
  • [19:15:48] * jkridner (n=jason@c-76-31-18-64.hsd1.tx.comcast.net) has joined #beagle
  • [19:20:14] * surki (n=surki@gek7.kyla.fi) has joined #beagle
  • [19:22:30] <adj_> geckosenator: yeah, I just received my first expansion board proto
  • [19:24:15] <adj_> and as soon as i received the pwb i noticed that dc/dc converter pin order was mirrored :(
  • [19:26:31] <geckosenator> heh
  • [19:26:37] <geckosenator> adj_: what expansion?
  • [19:26:43] <geckosenator> adj_: website? can I get protos?s
  • [19:28:06] <adj_> i'm building a robot of some kind, just to play with embedded linux, and the board serves as expansion/level shifter/servo controller board
  • [19:28:31] <geckosenator> so you made the board yourself?
  • [19:28:38] <adj_> yep
  • [19:28:40] <geckosenator> cool
  • [19:28:47] <geckosenator> what levels are you shifting? gpio?
  • [19:29:17] <adj_> i have schematics somewhere on the web, just a moment
  • [19:29:23] <geckosenator> great!
  • [19:29:49] <geckosenator> adj_: maybe you can glue the dc/dc converter on its back
  • [19:30:02] <adj_> http://www.hervanta.com/~antti/temp/add-on_board.pdf
  • [19:30:02] <geckosenator> adj_: then run small wires from the legs to the pads
  • [19:30:23] <adj_> that's how it is currently mounted
  • [19:30:54] <geckosenator> cool
  • [19:31:31] <geckosenator> I'm looking to something more generic.. sd card slot, serial port, and maybe soem other stuff
  • [19:32:06] <geckosenator> you have atmega on it?
  • [19:32:40] <geckosenator> maybe you can wire the beagle spi straight to the atmega spi, and be able to program it that way (as well as communication)
  • [19:32:41] <adj_> yes, atmega is mainly for providing 50 Hz PWM for servos
  • [19:32:59] <geckosenator> cool, I use avrs too
  • [19:33:21] <adj_> hmm, yeah, that programming option would be doable and quite usable
  • [19:33:33] <geckosenator> yeah
  • [19:33:41] <geckosenator> otherwise I guess you can use usb
  • [19:34:01] <geckosenator> I don't know if avrdude has support for spi like that.. but it could
  • [19:34:35] <geckosenator> wait, your regulator is linear not switching?
  • [19:34:43] <geckosenator> I'm making a board with all switching regulators
  • [19:35:12] <adj_> it's switching, tracopower's 3A converter
  • [19:36:12] <geckosenator> oh
  • [19:36:22] <geckosenator> is it a module with an inductor in it?
  • [19:37:04] * RobotGuy (n=robotguy@pool-96-225-210-142.ptldor.fios.verizon.net) has joined #Beagle
  • [19:37:14] <geckosenator> this is perfect, I wouldn't have known about that converter otherwise :-P
  • [19:37:45] <adj_> it's a damn expensive with "everything you need is inside" module :)
  • [19:37:50] <geckosenator> heh
  • [19:37:56] <adj_> http://fi.farnell.com/tracopower/tel-15-1211/converter-dc-dc-15w-5v-3a/dp/1242681?_requestid=20847
  • [19:38:03] <geckosenator> I found a part made by ti that is 96% efficient
  • [19:38:13] <geckosenator> it only needs 1 inductor and 2 caps
  • [19:38:19] <geckosenator> the rest is all inside
  • [19:38:28] <geckosenator> oh, you need 2 resistors if you want adjustable output voltage
  • [19:38:57] <adj_> how much current can it handle?
  • [19:38:59] <geckosenator> TPS63030
  • [19:39:03] <geckosenator> probably not as much
  • [19:39:12] <geckosenator> like 1.5 amps?
  • [19:39:37] <adj_> .8 it seems
  • [19:40:08] <adj_> still more than enough for powering beagle
  • [19:40:40] <geckosenator> yes
  • [19:40:46] <geckosenator> and it's buck/boost
  • [19:41:00] <geckosenator> so I can run the beagle off say a single cell lithium battery
  • [19:41:15] <geckosenator> and get longer run time when the battery voltage drops too low
  • [19:41:31] <geckosenator> it's not.. buck/boost.. it's buck or boost :-P
  • [19:43:31] <geckosenator> adj_: are you using it to power the beagle?
  • [19:43:34] <geckosenator> adj_: or for the servos?
  • [19:44:29] <adj_> currently it seems that i'll be using it to power only beagle and usb hub with peripherals such as wlan adapter and camera
  • [19:45:00] <geckosenator> oh
  • [19:45:02] <geckosenator> well here is my plan
  • [19:45:06] <adj_> if i use all 5 servos they could take too much current for that module
  • [19:45:08] <geckosenator> I'm going to have 2 of them
  • [19:45:15] <geckosenator> one will power only the beagle
  • [19:45:24] <geckosenator> at whatever the lowest voltage I can get away with
  • [19:45:32] <geckosenator> the other one powers the usb at 5v
  • [19:45:55] <geckosenator> adj_: there are reports of people powering the beagle at 2.5v instead of 5v (save some power)
  • [19:46:21] <geckosenator> adj_: the camera might use a lot of power.. mine uses like 300mA
  • [19:46:57] <adj_> well, i guess beagle's power consumption will not be problem in my case
  • [19:47:12] <geckosenator> it's snowing _again_ I need some sunshine so I can figure out how many solar panels I want to put in series (and what voltage my batteries will be)
  • [19:48:40] <geckosenator> adj_: do you have a realtime clock? apparently there are i2c parts that you can plug straight into the beagle's i2c bus and there is already a kernel driver
  • [19:49:31] <adj_> sadly no rtc
  • [19:49:38] <geckosenator> http://www.nxp.com/acrobat/datasheets/PCF8563_6.pdf
  • [19:50:13] <adj_> perhaps i should have put at least a place reservation for RTC on my board, but it didn't cross my mind as it seems i don't have use for it
  • [19:50:35] <geckosenator> oh ok
  • [19:50:57] <adj_> some pictures of my current setup:
  • [19:50:58] <adj_> http://www.hervanta.com/~antti/temp/beagle/
  • [19:51:34] <adj_> grrr, i'm still waiting for those level shifters from farnell
  • [19:53:03] <geckosenator> I am planning on using ones from ti
  • [19:53:20] <geckosenator> I have a level translator board for video (once revC is out)
  • [19:53:23] * vijay_ (n=vijay@203.199.213.3) Quit (Read error: 110 (Connection timed out))
  • [19:53:39] <geckosenator> revC is supposed to have parallel as well as hdmi video.. but the parallel is 1.8v, many lcds require 3.3v
  • [19:53:54] <geckosenator> adj_: nice board.. is it 2 layers/
  • [19:55:17] <adj_> yep, 2 layers
  • [19:55:50] <geckosenator> I can't seem to route my stuff with only 2
  • [19:56:02] <geckosenator> well, if I made the board bigger it would
  • [19:59:03] <ds2> morning
  • [19:59:47] <geckosenator> its not morning here
  • [19:59:50] <adj_> i have sometimes thought about ordering 4-layer protos from pcbcart.com
  • [20:00:45] <geckosenator> I contacted batchpcb and they said 4 layer boards have a 4 month wait
  • [20:00:54] <adj_> ouch
  • [20:01:07] <geckosenator> but it is cheap
  • [20:01:14] <ds2> heh where is that?
  • [20:01:15] <geckosenator> I'm going to check out pcbcart.com
  • [20:04:34] <geckosenator> adj_: pcbcart wants like $275 for my intended 4 layer boards
  • [20:04:40] <geckosenator> I don't know if I can pay that much :-P
  • [20:07:53] * kelledin (n=kelledin@cpe-70-123-124-210.tx.res.rr.com) has joined #beagle
  • [20:09:04] <ds2> geckosenator: tried pcbpool yet?
  • [20:10:12] * bazbell (n=a0192809@nat/ti/x-31ed735ffb560cd5) Quit (Remote closed the connection)
  • [20:10:16] * cmonex (n=xy6091@p0xvxcs9fl.adsl.datanet.hu) has joined #beagle
  • [20:10:27] <kelledin> engineering question: how much of the USB 2.0 power budget does the base beagleboard consume?
  • [20:11:10] <geckosenator> kelledin: you mean when powered off usb?
  • [20:11:18] <kelledin> and how much would it consume with, say, a Micrel KSZ8851 attached?
  • [20:11:23] <geckosenator> ds2: is it german?
  • [20:11:24] <kelledin> geckosenator: yes
  • [20:12:07] <geckosenator> ds2: pcbpool is like $220.. cheaper, but I want it more like $150 :-P
  • [20:12:24] <geckosenator> kelledin: well I have heard reports that in the worst cases it might use 500mA
  • [20:12:40] <geckosenator> well, more in really bad cases..
  • [20:12:46] <geckosenator> kelledin: I would just assume it uses all of it :-P
  • [20:13:10] <kelledin> geckosenator: true, I'm just contemplating what I could add to it without having to resort to a power brick
  • [20:13:30] <adj_> for hobbyist it's very hard to find affordable 4-layer pcb... just for one piece at least :(
  • [20:14:17] <ds2> geckosenator: yes but they claim to ship to US
  • [20:14:17] <kelledin> hmmm...i wonder what texas prototypes charges
  • [20:14:32] <geckosenator> adj_: well I'm doing like 5 designs on 1 board hopefully
  • [20:14:45] <kelledin> i know they do small runs, but i'm not familiar with their pricing
  • [20:15:04] <geckosenator> I would like to print my own pcbs with a printer
  • [20:15:06] <adj_> that's why i'm keeping my eyes closely on our work projects if our dedicated layout guy could find a suitable area in some of his work projects. That would mean 8 to 10-layer pwbs (and dozens of protos at once) :)
  • [20:15:15] <kelledin> geckosenator: you and me both ;)
  • [20:15:29] <geckosenator> instead of color cartriges, use conductive, non-conductive ones
  • [20:15:35] <kelledin> just getting a one-layer is difficult...
  • [20:15:44] <geckosenator> maybe even have a few different dielectrics so you never need any caps.. you just print htem
  • [20:15:51] <kelledin> i seem to remember you could get laser-printable acid masks
  • [20:16:12] <geckosenator> kelledin: I want to print transistors as well, but I don't think the dpi of most printers is high enough for a lot of stuff
  • [20:16:24] <kelledin> and do a little with a RepRap or...what's the other 3D fabber...
  • [20:16:33] <geckosenator> adj_: I would like 8 layer pcbs
  • [20:17:32] <geckosenator> kelledin: cool, it can build itself
  • [20:17:45] <kelledin> geckosenator: supposedly, yes
  • [20:17:57] <kelledin> i can almost remember what the other one was called
  • [20:17:58] <kelledin> but not quite
  • [20:18:05] <kelledin> but they're VERY low DPI
  • [20:18:31] <kelledin> like maybe good for cake-decoration resolution, but that's about it
  • [20:18:51] <geckosenator> that's ok for some stuff
  • [20:19:09] <geckosenator> but if I want to be able to print a die (like what's in a microprocessor)
  • [20:19:17] <geckosenator> I don't even think laser jet has sufficient dpi
  • [20:20:08] <kelledin> heh true
  • [20:20:38] <kelledin> i think even the 8086 was 12um gate width
  • [20:21:21] <geckosenator> I might make a realtime clock board for the beagle :-P
  • [20:21:28] <geckosenator> use a 3v watch battery, and i2c interface
  • [20:21:37] <kelledin> yah, that could work
  • [20:21:53] <kelledin> fab@home...that's the other one
  • [20:22:03] <kelledin> still pretty poor DPI
  • [20:22:25] <geckosenator> it's still very useful for stuff
  • [20:22:34] <geckosenator> is it only for fab? or can it assemble too?
  • [20:22:44] <kelledin> just for fabbing i think
  • [20:22:52] <kelledin> same for the RepRap as I recall
  • [20:23:01] <kelledin> it can spit out parts, but not actually put them together
  • [20:23:34] <geckosenator> ah, too bad
  • [20:23:50] <geckosenator> my roommate insists that computers are not "life" until they can completely build themselves
  • [20:24:27] <geckosenator> seems like a ridiculous require ment for me though
  • [20:25:00] <geckosenator> well.. I thought about making 2, 2 layer boards, and soldering pins in all the vias
  • [20:26:54] <kelledin> geckosenator: i assume your roommate thinks mules aren't alive then? ;-)
  • [20:27:14] <geckosenator> he thinks I'm not alive either since I don't reproduce
  • [20:27:33] <geckosenator> heh
  • [20:28:25] <kelledin> let him get mule-kicked in the gnads by something that isn't technically "alive" by his definition
  • [20:28:31] * DaQatz (n=db@c-66-30-48-54.hsd1.nh.comcast.net) Quit (Remote closed the connection)
  • [20:28:48] <kelledin> if the not-alive animal kicks hard enough, maybe your roommate will be similarly not-alive?
  • [20:29:36] <kelledin> and that's probably enough painful mental imagery
  • [20:29:43] <geckosenator> he
  • [20:30:11] <geckosenator> I don't know, I would like to see a computer that fit his definition
  • [20:30:30] <kelledin> how about software writing software?
  • [20:30:33] <geckosenator> I maintained that computers are alive.. my definition of life is anything that evolves
  • [20:30:42] <geckosenator> kelledin: that was one of my examples :-P
  • [20:31:57] <kelledin> anyways...can the beagleboard split power load across two USB jacks?
  • [20:32:06] <geckosenator> well
  • [20:32:14] <geckosenator> there is only 1 USB jack right?
  • [20:32:39] <kelledin> it is possible to make a cable that's USB on one end, and pure dc-power tap on the other
  • [20:33:07] <geckosenator> I did that
  • [20:33:21] * likewise (n=chatzill@82-171-51-231.ip.telfort.nl) has joined #beagle
  • [20:35:03] <kelledin> of course, that's still limited by jack power budget, so you'd need that in addition to the power from the USB data cable
  • [20:36:17] <geckosenator> well
  • [20:36:27] <geckosenator> kelledin: have you seen usb Y cables?
  • [20:36:34] <kelledin> geckosenator: yup
  • [20:36:34] <geckosenator> kelledin: also, a lot of hubs take external power
  • [20:36:45] <geckosenator> kelledin: what I"m planning is plug the hub into beagle, then power the hub
  • [20:37:18] <kelledin> PoE would also be interesting
  • [20:37:42] <geckosenator> eww
  • [20:37:43] <geckosenator> heh
  • [20:38:10] <geckosenator> my only regret with USB power is that it's 5v instead of 3.3
  • [20:38:17] * TAK2004 (n=Administ@dslb-088-074-060-023.pools.arcor-ip.net) has joined #beagle
  • [20:38:22] <kelledin> heh true
  • [20:38:28] <kelledin> of course, that means more watts supported
  • [20:38:32] <kelledin> so you can't complain too much
  • [20:39:13] <geckosenator> yeah
  • [20:39:38] <geckosenator> I'm boycotting usb 3 though
  • [20:39:54] <kelledin> due to intel's handling of the reference standard?
  • [20:40:37] <geckosenator> due to the fact that the same cables won't work
  • [20:41:02] <kelledin> ah yeah
  • [20:41:44] <geckosenator> I don't understand why they don't keep the same cable
  • [20:41:55] <geckosenator> and run it at 4.8gbps instead of 480mpbs
  • [20:42:43] <kelledin> probably the same reason 10GigE went to CX4 for copper
  • [20:43:00] <kelledin> too much bandwidth to try to shove down the pairs they had
  • [20:43:35] <kelledin> that reminds me of shit i discovered at the physics labs at my college
  • [20:43:58] <kelledin> in many cases they cheaped out on Cat5E and ran two 10/100 lines over a single cable
  • [20:44:15] <geckosenator> they should just use fiber then if they are going to change the cable (for future upgrades)
  • [20:44:45] <kelledin> fiber's not so good for mobility tho
  • [20:44:51] <kelledin> too easy to damage
  • [20:45:03] <kelledin> and a bit risky for eyesight as well
  • [20:45:04] <geckosenator> hmm
  • [20:45:30] <kelledin> unless you're using plain LED light, but good luck getting much length out of that
  • [20:45:35] <geckosenator> well all I'm saying is when usb 4.0 comes out, they are going to have like 16 twisted pairs
  • [20:47:03] <kelledin> true
  • [20:47:22] <kelledin> unless they get a bigger push for SerDes development
  • [20:47:30] <kelledin> which might happen now that it's catching on properly
  • [20:48:07] <geckosenator> yeah
  • [20:48:27] <kelledin> they're already hitting walls with PCIe 3.0 tho
  • [20:48:28] <geckosenator> or if there are more intelligent ways to send data so that some of it gets corrupted, and resent
  • [20:53:10] * nmq (n=chatzill@129.108.41.159) has joined #beagle
  • [20:53:51] <Ikarus> kelledin: errr, 10GigE over cat6e or cat7 cabling exists
  • [20:55:08] <Ikarus> CX4 cabling is awesome for low cost switch interconnects though
  • [21:05:01] * guillaum1 (n=gl@AMontsouris-153-1-82-57.w90-2.abo.wanadoo.fr) Quit ("Leaving.")
  • [21:05:26] * guillaum1 (n=Guillaum@AMontsouris-153-1-82-57.w90-2.abo.wanadoo.fr) has joined #beagle
  • [21:09:31] <geckosenator> they make realtime clocks that trickle charge the battery when powered
  • [21:11:01] <geckosenator> you can use super capacitors :-P
  • [21:13:39] <ds2> just build the RTC with a Li cell and a MSP430
  • [21:13:53] <geckosenator> heh
  • [21:13:58] <ds2> IIRC, that is either an appnote or a posted project
  • [21:14:13] <geckosenator> does it have an internal oscillator?
  • [21:14:15] <ds2> they mentioned it during the 430 day promo
  • [21:14:26] <ds2> yep. claim they got lifetimes of 2+ years on a single Li cell
  • [21:14:30] <geckosenator> I'm sure you can do it, but I don't know how to program 430s
  • [21:14:38] <ds2> look up ez430 =)
  • [21:14:50] <geckosenator> that would be cool though
  • [21:14:57] <ds2> uses standard C, and that is a USB stick programmer
  • [21:15:19] <geckosenator> the msp430 would be the realtime clock, and it could also perform all kinds of other functions on the expansion board
  • [21:15:22] <ds2> $20 cost...used to have coupons for $10 off
  • [21:15:26] * Xenion (n=robert@p579FCDF1.dip.t-dialin.net) has joined #beagle
  • [21:15:39] <ds2> *nod* and it comes in SOIC-16 packages
  • [21:15:47] <geckosenator> it's kind of ridiculous though :-P
  • [21:16:08] <geckosenator> they make realtime parts that have i2c, integrated oscillator, and can charge an external supercap
  • [21:16:25] <geckosenator> that seems the best to me, then it never needs replacement
  • [21:16:27] <ds2> I was tempted to use it as a TSC
  • [21:16:43] <ds2> No thanks, I prefer a Li cell
  • [21:16:48] <geckosenator> I suppose the msp430 would do all of that too
  • [21:17:04] <geckosenator> well the Li cell lasts years
  • [21:17:17] <geckosenator> the other way is to use a rechargable li cell
  • [21:17:43] <ds2> a supercap still has a lifetime and I suspect leakage on there gives it a shorter life time then a Li cell
  • [21:18:00] <ds2> besides, CR2032's are cheap... like $0.20 each
  • [21:18:07] <ds2> in packs of 5
  • [21:18:53] <geckosenator> yeah
  • [21:19:03] <geckosenator> leakage is pretty low for super caps actually
  • [21:19:19] <geckosenator> the lifetime would be longer since everytime you powerup, it charges it
  • [21:19:22] <geckosenator> heh
  • [21:19:27] * valhalla (n=valhalla@81-174-21-31.dynamic.ngi.it) Quit ("Leaving")
  • [21:19:41] <ds2> low enough to say... charge it up to full, and let it sit for 2 years and expect it to have enough charge left to power an RTC?
  • [21:20:14] <geckosenator> I would have to research that
  • [21:20:28] <ds2> then there is the whole die electric breakdown issue
  • [21:20:35] <ds2> KISS
  • [21:21:42] <geckosenator> well I woudln't overvolt it
  • [21:23:54] <geckosenator> I think the lithium cell is more reasonable :-P
  • [21:27:32] <ds2> temperature might also be a problem
  • [21:27:47] <geckosenator> for the capacitor?
  • [21:28:01] <ds2> yep
  • [21:28:10] <geckosenator> well
  • [21:29:25] <geckosenator> I might rather use a rechargable lithum cell
  • [21:29:32] <ds2> i rather not
  • [21:29:33] <nmq> I am having trouble getting an I2C device work on the expansion. It's fine if I just reduce the clock speed on the kernel config right? no risk of hardware damage right?
  • [21:31:54] <nmq> right?
  • [21:32:16] <geckosenator> I can't find any that support 1.8v and have a built in oscillator :(
  • [21:35:11] <geckosenator> ds2: the ds1340 auto-detects if the battery isn't rechargable, and doesn't recharge it in that case :-P
  • [21:37:18] * florian (n=fuchs@g228133152.adsl.alicedsl.de) has joined #beagle
  • [21:40:12] * ldesnogu_ (n=ldesnogu@ven06-2-82-247-86-183.fbx.proxad.net) Quit ()
  • [21:41:14] * RogerMon1 (n=a0740758@nat/ti/x-2d7993b66c327df5) has joined #beagle
  • [21:45:20] * RogerMonk (n=a0740758@nat/ti/x-96bc58e8004aa0c3) Quit (Remote closed the connection)
  • [21:46:23] <RogerMon1> ping gregoiregentil, sjam
  • [21:49:59] * RogerMon1 (n=a0740758@nat/ti/x-2d7993b66c327df5) has left #beagle
  • [21:53:22] * Leon_Nardella (n=Leon@200-161-14-111.dsl.telesp.net.br) has joined #beagle
  • [21:53:52] <gregoiregentil> Arghh! Ping RogerMonk?
  • [21:54:19] <gregoiregentil> Very unlucky I am to discuss with RogerMonk :-( sniff
  • [22:00:48] * RogerMonk (n=a0740758@nat/ti/x-111376351f4fe69e) has joined #beagle
  • [22:01:00] <gregoiregentil> Ping RogerMonk!?
  • [22:03:04] <RogerMonk> Hey gregoiregentil!
  • [22:03:10] <gregoiregentil> Hey!
  • [22:03:11] <RogerMonk> how's things?
  • [22:03:27] <gregoiregentil> Good. Thanks. I'm playing with CE and I think that you have good knowledge
  • [22:03:37] <RogerMonk> Anything I can help with?
  • [22:03:39] <gregoiregentil> I have a few questions to understand a few things
  • [22:03:40] <gregoiregentil> yes
  • [22:03:49] <RogerMonk> ok - let's try
  • [22:04:04] <gregoiregentil> so first, I was told that the CE repice in OE doesn't pull everything
  • [22:04:21] <gregoiregentil> is that correct? Koen mentioned that you may have a "full blown" CE recipe
  • [22:04:24] <RogerMonk> you mean doesn't autodownload?
  • [22:04:42] <RogerMonk> pull?
  • [22:04:44] <gregoiregentil> Let me ask differently: hwo can I know which codec I have when I build
  • [22:04:50] <gregoiregentil> the package?
  • [22:05:00] <RogerMonk> ah-ha, ok
  • [22:05:50] <RogerMonk> the ce recipe today, builds the CE examples, and all the necessary sub-components (dsplink, lpm, cmem, etc). The codecs are 'dummy' codecs which just do a 'copy' operation
  • [22:06:16] <gregoiregentil> so is there a place where it's possible to get the evluation codec?
  • [22:06:17] <RogerMonk> We now need to add recipes for building a 'real' dspserver, with real codecs
  • [22:06:44] <gregoiregentil> OK. will it be a free version for some codec or will one have to buy the
  • [22:06:52] <RogerMonk> yep, hoping to do that this week - now that we've got the base CE infrastructure up and running (thanks to Koen's help), we can go wild
  • [22:06:57] <RogerMonk> what codecs u looking for?
  • [22:06:57] <gregoiregentil> TI codec that jkridner referenced a few days ago
  • [22:07:16] <gregoiregentil> mpeg2, mpeg4 (including H264)
  • [22:08:21] <RogerMonk> on which platform? - beagle?
  • [22:08:23] <gregoiregentil> RogerMonk: just to understand, you work for TI. correct?
  • [22:08:26] <gregoiregentil> yes, beagle
  • [22:08:28] <RogerMonk> right
  • [22:08:51] <gregoiregentil> I would like to understand what is given for free and what should be bought eventually: something like
  • [22:08:51] <gregoiregentil> http://www.ti-estore.com/Merchant2/merchant.mvc?Screen=PROD&Product_Code=C64XPLUSBAUD1P
  • [22:10:05] <RogerMonk> ah-ok, this is audio codecs, they should work ok on omap
  • [22:10:18] <gregoiregentil> this one is audio. I wanted to show video.
  • [22:10:44] <gregoiregentil> just to understand, you are working for TI. Am I correct or Have I miread?
  • [22:10:47] <RogerMonk> yep, video codecs are still in progress I think - should be coming soon in a similar way - we already have this for davinci
  • [22:10:58] <RogerMonk> yes, I work at TI
  • [22:11:33] <RogerMonk> are you working on omap or davinci?
  • [22:11:44] <gregoiregentil> omap / beagleboard
  • [22:11:58] * ClaudeQC (n=claude@bas1-quebec03-1279637127.dsl.bell.ca) has left #beagle
  • [22:12:03] <RogerMonk> ok, and do you have an omap evm as well?
  • [22:12:20] <gregoiregentil> no
  • [22:12:27] <RogerMonk> ah, ok
  • [22:13:13] <RogerMonk> have you been able to build the CE recipes ok in OE and run the examples?
  • [22:13:27] <gregoiregentil> yes. I built them
  • [22:13:36] <gregoiregentil> the second part of my questions were about the examples
  • [22:13:49] <RogerMonk> ok
  • [22:13:50] <gregoiregentil> let me just summarize what we said - one second
  • [22:21:56] * likewise (n=chatzill@82-171-51-231.ip.telfort.nl) Quit ("ChatZilla 0.9.84 [Firefox 3.0/2008061015]")
  • [22:30:22] <gregoiregentil> RogerMonk: when you said "building a 'real' dspserver, with real codecs". Can you be more precise?
  • [22:30:28] <gregoiregentil> on this page: http://www.ti-estore.com/Merchant2/merchant.mvc?Screen=CTGY&Category_Code=dEmbSoft
  • [22:30:42] <gregoiregentil> there is only audio for C64x
  • [22:31:00] <gregoiregentil> Am I understanding correctly that video codec are not yet ready for OMAP DSP but will be soon?
  • [22:31:51] <RogerMonk> right, omap codecs are still in beta - they'll be coming soon
  • [22:32:48] <RogerMonk> by 'real' dspserver, I mean building a dsp image that contains real codecs, rather than dummy 'copy' codecs
  • [22:32:53] <gregoiregentil> so what are you intending to add in the codec-engine.bb this week?
  • [22:32:59] <RogerMonk> nothing
  • [22:33:10] <RogerMonk> that will stay the same
  • [22:33:33] <RogerMonk> I'm working on another recipe that will build some real codecs into a server
  • [22:33:51] <gregoiregentil> OK. I understand a little bit better now
  • [22:34:00] <RogerMonk> then we'll need an arm side test app that uses the dsp server
  • [22:34:02] <gregoiregentil> and those real codecs will be audio or video?
  • [22:34:16] <RogerMonk> (probably DMAI based (if you are familiar with DMAI))
  • [22:34:21] <gregoiregentil> exactly!!!
  • [22:34:23] <RogerMonk> both hopefully
  • [22:34:28] <gregoiregentil> it's what I tried to do this week-end.
  • [22:34:38] <gregoiregentil> to create a recipe for DMAI... I think that it will be the simplest
  • [22:34:41] <RogerMonk> ok, how did you get on with DMIA?
  • [22:34:53] <gregoiregentil> way to test DSP multimedia. just a proof of concept
  • [22:35:13] <RogerMonk> did you port the drivers to DSS2 yet?
  • [22:35:26] <gregoiregentil> so I downloaded DMAI from TI website and started to write the recipe but I had some trouble
  • [22:35:36] <RogerMonk> ok - what sort of trouble?
  • [22:35:42] <gregoiregentil> why DSS2? What are we talking about now can't work with DSS1?
  • [22:35:55] <gregoiregentil> I have nothing against DSS2 but it's less mature and we are still sorting out
  • [22:36:00] <gregoiregentil> some problem with Koen
  • [22:37:00] <RogerMonk> sorry, I should be more clear. DMAI currently only works with TI kernel, not mainline and will therefore need to be ported to either DSS1 or DSS2. I presumed you would be starting with DSS2
  • [22:37:30] <gregoiregentil> OK. I didn't know that. Let me think one second
  • [22:38:15] <RogerMonk> changes should be fairly minimal, we just haven't had time to look at this yet
  • [22:38:59] <gregoiregentil> Sorry again but I'm a little bit confused. I thought that DMAI was working on top of codec-engine. For me, ti doesn't touch the kernel: http://wiki.davincidsp.com/index.php?title=Davinci_Multimedia_Application_Interface
  • [22:39:32] <gregoiregentil> err, I mean: for me, DMAI doesn't touch the kernel
  • [22:39:58] <gregoiregentil> http://wiki.davincidsp.com/index.php?title=Image:Dmai_block.png
  • [22:40:17] <gregoiregentil> we DMAI touches a little bit the kernel on the diagram ;-)
  • [22:41:21] <RogerMonk> You are right, it does sit on top of CE, in particular, on top of XDM VISA classes. It also abstracts drivers as well to hide differences between OMAP/DM355/DM6446/DM6467, which all use different driver models.
  • [22:41:41] <RogerMonk> (and different kernel versions)
  • [22:42:30] <gregoiregentil> so bottom-line: you are saying that kernel l-o should be patched for DMAI even if the objective is to use only VISA?
  • [22:43:00] <RogerMonk> no, I'm saying DMAI should be patched to add support for mainline l-o
  • [22:43:47] <gregoiregentil> OK. the reverse. Got you.
  • [22:44:32] <gregoiregentil> but has this work bee done yet? I was thinking about gst-dmai plugin...
  • [22:44:40] <RogerMonk> (however, if you don't care about the driver piece of DMAI, then you don't need to build those bits - however, the dmai test apps will be less useful
  • [22:45:37] <gregoiregentil> so is there any pointer to understand what would need to be done on DMAI to support kernel l-o?
  • [22:46:41] <gregoiregentil> Also, I'm still unsure to understand what is lost if there is no DMAI-linux connection?
  • [22:47:17] <RogerMonk> hmm - not sure. Think you could look at the Display, Capture and Sound modules to see how they are currently using TI kernel - there's probably not a lot needed to be changed
  • [22:47:46] <RogerMonk> The Cpu module, will also need 'beagle' machine identifier added.
  • [22:48:58] <gregoiregentil> so I guess that it's this folder that probably needs to be patched: dmai-1.10.00.06
  • [22:48:59] <gregoiregentil> -r0/dmai_1_10_00_06/packages/ti/sdo/dmai/linux. Am I correct?
  • [22:49:10] <RogerMonk> I'll try to look at this sometime this week and will update you on my progress. Ping me during the week and let me know how u r getting on your side.
  • [22:49:24] <RogerMonk> yep, sounds about right
  • [22:49:32] <RogerMonk> (don't have source in front of me tonight)
  • [22:50:43] <gregoiregentil> do you intend to write a dmai recipe? I think that it would be useful for testing (Gregoire hoping that somebody with the expertise like RogerMonk can do it much faster than him...)
  • [22:56:34] <RogerMonk> I'll see what I can do - I've got quite a bit on this week, but will try :)
  • [22:56:53] * chelli (n=chelli@debian/developer/tschmidt) Quit ("IRC is just multiplayer notepad")
  • [22:56:58] <gregoiregentil> Cool. Thanks. Talk with you later this week then.
  • [22:57:17] <RogerMonk> ok sir - I'll be online most of the week - just ping me.
  • [22:57:30] <gregoiregentil> Great! Thanks
  • [23:15:38] * abitos (n=nixgibts@dslb-084-057-158-020.pools.arcor-ip.net) Quit (Read error: 113 (No route to host))
  • [23:22:17] * RogerMonk (n=a0740758@nat/ti/x-111376351f4fe69e) has left #beagle
  • [23:33:59] * Leon_Nardella (n=Leon@200-161-14-111.dsl.telesp.net.br) Quit (Remote closed the connection)
  • [23:35:41] * vijay (n=vijay@203.199.213.3) has joined #beagle
  • [23:36:36] * guillaum1 (n=Guillaum@AMontsouris-153-1-82-57.w90-2.abo.wanadoo.fr) Quit ("Leaving.")
  • [23:37:03] * fulgas is now known as FuL|OUT
  • [23:37:09] * guillaum1 (n=gl@AMontsouris-153-1-82-57.w90-2.abo.wanadoo.fr) has joined #beagle
  • [23:40:52] * Xenion (n=robert@p579FCDF1.dip.t-dialin.net) Quit ("Verlassend")
  • [23:41:23] * florian (n=fuchs@g228133152.adsl.alicedsl.de) Quit (Read error: 113 (No route to host))