#touchbook IRC Log


IRC Log for 2009-09-03

Timestamps are in GMT/BST.

[0:00] <hyc_tb> it appears to have more PM features than what I was looking at before, so I'm going to plow on with it
[0:01] <hyc_tb> btw, I maded some updates to the DSP Video page on the wiki with my latest findings for DSP Link
[0:01] <hyc_tb> the newest code built ok, the codecs worked, it was kinda cool
[0:01] <hyc_tb> now I'm going back to try to get dspbridge working
[0:02] <hyc_tb> eh... I generally like newest, yes, but this time I'd like newer+more stable ;)
[0:02] <hyc_tb> lol
[0:03] <hyc_tb> as long as you all send me a bottle of Scotch..... :P
[0:04] <hyc_tb> ok, back to figuring out what files I'm missing...
[0:18] * derken73 (n=quassel@ Quit ("http://quassel-irc.org - Chat comfortably. Anywhere.")
[0:18] * derken73 (n=quassel@ has joined #Touchbook
[0:21] * blunderer (n=tristan@LPuteaux-156-15-47-90.w82-127.abo.wanadoo.fr) has joined #touchbook
[1:07] * Anges (n=agnes@lns-bzn-49f-62-147-173-3.adsl.proxad.net) has joined #touchbook
[1:32] * Geep (n=JoeBelow@bas21-toronto12-1279470488.dsl.bell.ca) Quit (Remote closed the connection)
[1:55] * Meiz_n810 (n=Meizirkk@padedu-62-165-142-173.phnet.fi) Quit (Read error: 110 (Connection timed out))
[2:00] * spaetz (n=spaetz@ has joined #touchbook
[2:39] * tshirtman (n=gaby@ks22672.kimsufi.com) Quit (Remote closed the connection)
[2:39] * hyc_tb (n=hyc@ Quit ("Powered by OE: www.openembedded.org")
[3:18] * spaetz (n=spaetz@ Quit (Read error: 60 (Operation timed out))
[3:19] * thirdhandinfo (n=thirdhan@ Quit (farmer.freenode.net irc.freenode.net)
[3:19] * Pebby (n=pebby@c-98-248-129-187.hsd1.ca.comcast.net) Quit (farmer.freenode.net irc.freenode.net)
[3:19] * bstag (n=bstag@ Quit (farmer.freenode.net irc.freenode.net)
[3:19] * blunderer (n=tristan@LPuteaux-156-15-47-90.w82-127.abo.wanadoo.fr) Quit (farmer.freenode.net irc.freenode.net)
[3:19] * andrewgodwin (n=andrew@hydrae.aeracode.org) Quit (farmer.freenode.net irc.freenode.net)
[3:19] * honk (i=honk@xehonk.com) Quit (farmer.freenode.net irc.freenode.net)
[3:19] * MaceRep (i=macer@lasziv.reprehensible.net) Quit (farmer.freenode.net irc.freenode.net)
[3:19] * derken73 (n=quassel@ Quit (farmer.freenode.net irc.freenode.net)
[3:19] * Corsac (i=corsac@ Quit (farmer.freenode.net irc.freenode.net)
[3:19] * dpb (n=dpb@unaffiliated/dpb) Quit (farmer.freenode.net irc.freenode.net)
[3:19] * Pebby (n=pebby@c-98-248-129-187.hsd1.ca.comcast.net) has joined #touchbook
[3:23] * thirdhandinfo (n=thirdhan@ has joined #touchbook
[3:25] * blunderer (n=tristan@LPuteaux-156-15-47-90.w82-127.abo.wanadoo.fr) has joined #touchbook
[3:25] * honk (i=honk@xehonk.com) has joined #touchbook
[3:25] * andrewgodwin (n=andrew@hydrae.aeracode.org) has joined #touchbook
[3:25] * bstag (n=bstag@ has joined #touchbook
[3:25] * MaceRep (i=macer@lasziv.reprehensible.net) has joined #touchbook
[3:26] * Corsac (i=corsac@ has joined #touchbook
[3:26] * dpb (n=dpb@unaffiliated/dpb) has joined #touchbook
[3:49] * hyc_tb (n=hyc@ has joined #touchbook
[4:04] * spaetz (n=spaetz@ has joined #touchbook
[4:37] * GUido-- is now known as GUido-
[5:20] * robclark (n=robclark@nat/ti/x-uwvguhrsmncrnohs) has joined #touchbook
[5:48] * Meizirkki (n=Meizirkk@bbwirelessgw2-feeedc00-64.dhcp.inet.fi) has joined #touchbook
[6:01] <martinh> hmmmm. i was hoping to have my tb in time for the ohio linux fest (ohiolinux.org), but I've not hear anything so it looks like I won't.
[6:07] <Corsac> when is it and when did you pre-order?
[6:07] <dpb> (or did he even pre-order..)
[6:09] * spaetz (n=spaetz@ Quit (Read error: 110 (Connection timed out))
[6:17] <martinh> pre ordered first week of april. the OLF is the last weekend this month.
[6:18] <Corsac> did you qualified your order?
[6:19] * shiznebit (n=chatzill@ has joined #touchbook
[6:29] <martinh> sure did.
[6:30] <Corsac> at which level?
[6:30] <martinh> i can't remember exactly. but I think it was 8 or 9.
[6:30] <dpb> the pre-september shipments have been quite small
[6:30] <dpb> you'll probably get one from the september shipment
[6:30] <Corsac> yep
[6:30] <martinh> i haven't heard anything about it. . .
[6:31] <Corsac> which *should* be mid-september
[6:31] <Corsac> martinh: look at the news page
[6:48] * dec0de (n=theo@h189n1fls34o881.telia.com) has joined #touchbook
[6:52] <martinh> yea. . .that all says i may. . .but, i still haven't heard that I will.
[6:53] <martinh> I'm figuring on a week turn around from charge/email to receiving it. so, i've got about week and a half.
[6:55] <martinh> maybe up to 2 weeks.
[6:56] <Corsac> I'd say 2 weeks between charge/mail and receive
[6:56] <Corsac> but I may be a little pessimistic :)
[6:56] <dec0de> does anyone have any contact with Collins at Always Innovation?
[6:57] <Corsac> It's Always Innovating
[6:57] <dec0de> ohh.. bugger.. ;-)
[6:57] <Corsac> and I don't know who's Collins :)
[6:58] <dec0de> the person who is answering the e-mails to contact@alwaysinnovation.com
[6:58] <dpb> You sure it's with an s? An "Collin L." answered my billing address change.
[6:58] <dec0de> sorry.. innovating..
[6:58] <Corsac> ok
[6:59] <dec0de> well.. maybe Collin L it is..
[6:59] <dpb> dec0de: what about him then?
[7:00] <martinh> gregoire was hanging out here last week. but he's apparently been really busy again.
[7:02] <Corsac> and it's sleep time atm :)
[7:02] <dec0de> dpb: To be able to get the touch book quicker, I asked Mr Collin if they could send it to another address. Then he wanted to mix the new address with the old one... and now I haven't gotten any reply on my last two mails..
[7:02] <dpb> huh
[7:02] <dec0de> dpb:Thus I am just very curious what on earth is happening, and where might the touch book end up?
[7:03] <Corsac> in between
[7:03] <dec0de> or.. has did he maybe cancel the order....
[7:03] <dec0de> Corsac: well.. thats what I suspect. Mr Collins want to mix the two addresses given, which is quite unwise in my oppinion..
[7:04] <dec0de> dpb, so i thought maybe someone at Always Innovating, might hang around here.. :-)
[7:05] <shiznebit> he used too....
[7:06] <shiznebit> before he ran for the boarder
[7:06] * hyc_tb (n=hyc@ Quit (Read error: 110 (Connection timed out))
[7:07] <dec0de> It's a bit annoying to have a pre-order laying around, not knowing when, where and if the touch book will ever find its destination... :-(
[7:08] <dec0de> maybe I should have asked Mr Collin to take the touch book with him when running for the border. He could have delivered it in person. ;-)
[7:27] <dpb> soo, atleast 3 persons seem to work in AI
[7:31] <Corsac> Gregoire, M. Collin and?
[7:32] <dpb> Tim Yamin, seems to own the copyright of the ai-daemon for this year
[7:32] <dpb> http://www.alwaysinnovating.com/wiki/index.php/Ai-daemon
[7:34] <Corsac> yeah but he doesn't have an ai address
[7:34] <Corsac> so I'd say it's a contractor
[7:35] <dpb> and some Davids has admin rights to the wiki: http://www.alwaysinnovating.com/wiki/index.php?title=Special:ListUsers&limit=100
[7:39] * killring (n=killring@ has joined #touchbook
[7:40] * spaetz (n=spaetz@ has joined #touchbook
[7:43] <martinh> ha! on pcworld: Netbook Fail: Nokia Booklet Sports a Big-Laptop Price
[8:06] * jvs (n=jvs@ has joined #touchbook
[8:09] <shiznebit> lol because the OEM's need to cover their losses
[8:16] <ratonk> anyone has an idea to have rss feed with the information webpage http://www.alwaysinnovating.com/company/news.htm ??
[8:17] * ratonk_ (i=ratonk@after-work.chapeaux-noirs.org) has joined #touchbook
[8:23] * dec0de (n=theo@h189n1fls34o881.telia.com) Quit (Read error: 110 (Connection timed out))
[8:34] <spaetz> ratonk: google helps: http://www.montastic.com/
[8:34] <spaetz> never used it myself
[8:43] <ratonk_> spaetz: i did try a bit but found nothing interesting
[8:43] <ratonk_> Jay-o: i mean it would be good idea na me: thx to give me such advice la
[8:43] <ratonk_> but the pint is i'm not developper so there are nto so much jobs for me in this area
[8:43] <ratonk_> but i would consider for future ;)oups
[8:43] <ratonk_> sorry my mistake, damn right clic
[9:10] * jvs (n=jvs@ Quit (Remote closed the connection)
[9:11] * jvs (n=jvs@ has joined #touchbook
[9:22] * spaetz (n=spaetz@ Quit (Read error: 110 (Connection timed out))
[10:11] * jkridner|tb (n=ai@ Quit (Read error: 110 (Connection timed out))
[10:20] * Geep (n=JoeBelow@bas21-toronto12-1279470488.dsl.bell.ca) has joined #touchbook
[10:46] * blunderer (n=tristan@LPuteaux-156-15-47-90.w82-127.abo.wanadoo.fr) has left #touchbook
[11:43] * dirk2 (n=dirk@p5B040250.dip0.t-ipconnect.de) Quit ("Konversation terminated!")
[12:05] * Meizirkki (n=Meizirkk@bbwirelessgw2-feeedc00-64.dhcp.inet.fi) Quit (Read error: 110 (Connection timed out))
[12:11] * [AFX] (n=afx@S0106001b116229f2.tb.shawcable.net) Quit (Read error: 110 (Connection timed out))
[13:13] * jkridner|tb (n=ai@ has joined #touchbook
[13:37] * hyc_tb (n=hyc@ has joined #touchbook
[13:48] * dec0de (n=theo@c83-251-53-17.bredband.comhem.se) has joined #touchbook
[13:56] * Anges (n=agnes@lns-bzn-49f-62-147-173-3.adsl.proxad.net) Quit ("Leaving.")
[13:58] * jjwere (n=d2dadmin@ Quit ("Hey! Where'd my controlling terminal go?")
[14:46] * shiznebit (n=chatzill@ Quit (Remote closed the connection)
[14:53] <Eruquen> is there some way to attach a vga or dvi display to the touchbook?
[14:54] <koen> displaylink device would be your best bet atm
[14:55] <honk> Eruquen: get an usb to vga converter
[14:55] <Eruquen> they don't exactly come cheap
[14:56] <honk> they're not exactly expensive either
[14:56] <martinh> yea.
[14:56] <koen> have a look at http://www.displaylink.com/shop/adapters/
[14:56] <honk> 40$ or so
[14:59] <Eruquen> well, here in Germany they are sold for 60 EUR (~ 80 USD)
[15:00] <Eruquen> horror slide show? :D
[15:00] <koen> the sis ones have enough accell to be acceptable
[15:00] <koen> but still, vga, blech
[15:00] <Eruquen> I guess I'll stick to vnc or something similar
[15:01] <Eruquen> and a hell of a lot cheaper, too
[15:03] <Eruquen> I'm not a fan of usb either but afaik it's the only interface available on the touchbook (not counting wlan ofc)
[15:04] <honk> and more expensive
[15:05] <Eruquen> there's no ethernet port, right?
[15:07] <hyc> right
[15:07] <hyc> just usb ports
[15:07] <hyc> and the sd card slot
[15:07] <hyc> yeah
[15:08] <hyc> so it doesn't really count, since you have to boot off it
[15:09] <hyc> I wouldn't both, external flash usb are so common
[15:09] <martinh> you can always yank the BT one if you don't want that.
[15:09] <martinh> for 2 internals
[15:10] <martinh> but, the beagle board it's based on has also has lots of other i/o ports. I suspect that it might be possible to add more stuff if you're adventerous.
[15:10] <hyc> if it fits, why wouldn't it work?
[15:10] <honk> it's just linux
[15:10] <martinh> if there's linux drivers. . .why shouldn't it?
[15:11] <hyc> if the driver is open source, it should work
[15:11] <honk> if nothing else works, you could build 'em yourself :}
[15:11] <hyc> so what?
[15:11] <hyc> compile, load, run.
[15:11] <honk> most drivers are available for all platforms :}
[15:12] <hyc> the wifi device is a ralink RT3070, the driver is just built from the source code Ralink provides on their web site
[15:12] <hyc> it works for any Linux platform
[15:12] <hyc> if you're unfortunate enough to pick a device that only ships with binary x86 drivers, well, you bought the wrong hardware
[15:13] <hyc> but most USB modems all follow ITU standards, they respond to Hayes-style AT commands
[15:13] <hyc> as long as they support the USB CDC-modem spec, the generic Linux driver will work
[15:14] <hyc> I'm still debating whether to buy a new USB HSDPA modem or just tether with my G1 phone
[15:14] <hyc> tethering is probably simple enough
[15:17] <martinh> i tether my linux box to my centro. it's dead simple.
[15:18] <martinh> I suspect I'll be doing the same thing with my tb.
[15:46] * dec0de (n=theo@c83-251-53-17.bredband.comhem.se) Quit (Read error: 110 (Connection timed out))
[16:33] * robclark (n=robclark@nat/ti/x-uwvguhrsmncrnohs) Quit ("be seeing you")
[16:44] * edt (n=Ed@dsl-216-221-38-250.aei.ca) has joined #touchbook
[16:46] <edt> does the touchbook have a mic?
[16:48] <GUido-> no
[16:50] <edt> I can think of one change I would love to see for v2 hardware - a bluetooth keyboard.
[16:51] <edt> standard linux bluetooth support should work with a bluetooth headphone (done on my n800)
[16:52] <edt> I have one. The idea would be that when one detaches the keyboard all that detaches is the extra battery - you can still (remote) control the display
[16:54] <edt> think this should free another usb port inside. I can see two devices I would like to plug in a touch book. A 3G dongle and a gps dongle. A second free internal usb would be very nice.
[16:56] <edt> is there a picture of the touchbook's back open on the site?
[17:00] * robclark (n=robclark@nat/ti/x-yvczesmvvbyfswae) has joined #touchbook
[17:03] <edt> might be room for a small hub...
[17:09] <hyc> it's pretty packed in there. a hub would fit, but not many devices
[17:22] * robclark (n=robclark@nat/ti/x-yvczesmvvbyfswae) Quit (Remote closed the connection)
[17:24] * robclark (n=robclark@nat/ti/x-ddpbbvskhuhvqkpw) has joined #touchbook
[17:27] <edt> hyc I can see adding two devices gps & 3g doogle
[17:27] * shiznebit (n=chatzill@ool-18b99610.dyn.optonline.net) has joined #touchbook
[17:29] <hyc> edt: that would be cool, yes
[17:29] <hyc> but I have both of those in my G1 already
[17:30] <hyc> it would be simpler just to write a G1 app that exports GPS data over the network
[17:30] <hyc> and then if you tether the TB to the G1, you get it all
[17:31] <robclark> hyc: is there no GPS device that would connect over USB as a virtual UART? NMEA is pretty easy to parse..
[17:32] <hyc> robclark: I'm sure they exist; we were talking about what can be crammed inside the TB case so there are no external dongles hanging around
[17:33] <hyc> there's only 1 free internal USB port though
[17:33] <robclark> ahh, ic..
[17:34] <hyc> but yes, I'd like 3G and GPS in my TB too. but I could settle for siphoning that off the G1 if I can't cram it all into the TB case.
[17:34] <edt> robclark problem is that I would like to use a 3G doogle inside along with a gps... There seems to only be one usb free inside. Mind you the bluetooth doogle looks to be quite small - maybe it could live in an external usb without danger.
[17:34] <robclark> I was kinda interested to attache one externally, like on top of a small airplane.. moving maps would be nice for navigation
[17:34] <edt> there are many bluetooth gps units too
[17:34] <robclark> (in my case, when I'd want to be using GPS, I'd detatch BT and WLAN)
[17:35] <hyc> hmmm... I guess that'd work, as long as you have a decent map database in local storage
[17:35] <edt> opensource gps apps love having a internet connection...
[17:35] <robclark> yeah, I'd have to come up with some way to slurp down all the map data pre-flight
[17:36] <robclark> (not really sure how I'd do all that yet... but it'd be an interesting project)
[17:36] <hyc> definitely would be cool
[17:36] <edt> if someone ports maemo maps to the touchbook that will not be much of a problem
[17:37] <robclark> ok, I'll have to checkout maemo maps.. probably wouldn't be hard to port
[17:37] <edt> there are quite a few n8xx apps that would be nice on the touchbook
[17:37] <shiznebit> ovi maps
[17:37] <shiznebit> i have no clue how to get em on linux
[17:37] <robclark> maemo is very much a little linux desktop.. X11, gtk, etc..
[17:37] <shiznebit> N900 ftw
[17:37] <edt> and arm based too
[17:38] <shiznebit> but $200 says Tmobile f it up
[17:38] <robclark> yeah, I'm really interested to see what different carriers will have it in US.. I hope it isn't tmobile exclusive
[17:38] <shiznebit> if its w/e
[17:39] <shiznebit> atleast it will work in europe guaranteed
[17:39] <shiznebit> but i will bet my life on it
[17:39] <shiznebit> that we wont have root
[17:39] <shiznebit> and that some part, particularly the antenna and 3g radio will be fully binary
[17:39] <hyc> it didn't take long for folks to gain root on android
[17:40] <shiznebit> it wasn't officially supported
[17:40] <hyc> but yes, no doubt the radio section will be sealed off. it's dedicated hardware and a separate blob on the G1
[17:40] <shiznebit> they just left it open, by mistake
[17:41] <shiznebit> why do they protect that part
[17:41] <shiznebit> though
[17:41] <shiznebit> security through obscurity ?
[17:42] <hyc> something like that. trade secrets and all
[17:42] <hyc> Qualcomm is highly protective of their radio technology, and charges big $$ to license it out
[17:43] * AndroidIRC (n=shiznebi@m3f5736d0.tmodns.net) has joined #touchbook
[17:44] <AndroidIRC> what about openmoko ?
[17:45] * robclark (n=robclark@nat/ti/x-ddpbbvskhuhvqkpw) Quit (Remote closed the connection)
[17:45] <AndroidIRC> how does it consider itself opensource ?
[17:45] <edt> I would be susprised if nokia does not make root available - there is a large opensource maemo comunity that will be VERY, motivated to hack, if root is locked.
[17:46] <AndroidIRC> does openmoko use an opensource radio ?
[17:46] <edt> you are probably write about the 3g stuff being a binary blog though
[17:46] <edt> s/blog/blob/
[17:47] <AndroidIRC> edt: they'll probably have to find another user exclation vulnerbility
[17:47] <edt> I doubt it
[17:47] * AndroidIRC is now known as shiznebit-
[17:47] <edt> nokia would get VERY VERY VERY bad feedback if they try to lock root
[17:48] <shiznebit-> yeah
[17:48] <shiznebit-> but if its open then that would allow voip ?
[17:48] <edt> given that nokia wants to gain market share it would be a bad move.
[17:48] <shiznebit-> through wifi...
[17:49] <shiznebit-> z
[17:49] * robclark (n=robclark@nat/ti/x-ggvqkdusbrzqtthn) has joined #touchbook
[17:50] <shiznebit-> meh i cant make a text undo
[17:50] <edt> viop will be available on the n900
[17:50] <shiznebit-> so openmoko used an opensource radio or a blob ?
[17:51] * shtylman (n=shtylman@cpe-68-173-99-55.nyc.res.rr.com) Quit (Remote closed the connection)
[17:51] <robclark> root is avail on n900: http://flors.wordpress.com/2009/08/27/software-freedom-lovers-here-comes-maemo-5/
[17:54] <shiznebit> sweet
[17:55] <edt> exactly - I've been following the n700/n800/n810/n900 saga and from the start nokia has been a 'good guy' and have improved with time (and been open with why when they cannot open some aspect of the stack or hw)
[17:56] <shiznebit> the stack ?
[17:56] <edt> I like the touchbook with its larger screen - my 50+ eyes do not like small screens...
[17:56] <shiznebit> i understand some hardware just can't be open
[17:56] <edt> software stack
[17:56] <hyc> heh, philosophically no, everything *can* be open
[17:56] <shiznebit> what software stack can't be open ?
[17:56] <robclark> I assume the radio (if that is what you mean by stack) will be well sealed off...
[17:56] <hyc> and *should* IMO
[17:57] <hyc> the GnuRadio guys have implemented a lot of these air protocols already too
[17:57] <shiznebit> hyc, was openmoko fully open ?
[17:57] <hyc> don't recall, didn't read enough about it
[17:58] <robclark> I don't really know if n900 is using a "thin" modem, sharing memory/flash with the omap.. but I would have to think they partition it somehow if they give out root so easily
[18:00] <shiznebit> GSM cell network and the unfortunately slow GPRS data network. It has built-in GPS, Wi-Fi, and Bluetooth, and supports microSD memory cards
[18:00] <shiznebit> hmm i guess they couldn't find a manufacturer who would release soource for EDGE/3G antennas
[18:01] <shiznebit> http://www.pcworld.com/article/134987/insider_info_on_the_first_fully_opensource_cell_phone.html
[18:02] <hyc> gahhh... even the bridge from git fails to build, using the entire kernel source
[18:03] <shiznebit> o_O
[18:03] <hyc> (sorry, wrong channel, but robclark knows what I'm talking about)
[18:03] <shiznebit> what chya talking about ?
[18:04] <hyc> trying to build a new dspbridge driver for the TB
[18:05] <hyc> the source code is not exactly "clean" ...
[18:05] <shiznebit> isn't that TI's DSP
[18:05] <hyc> in this case, the driver references a variable that was ifdef'd out. the reference needs to be ifdef'd and isn't.
[18:06] <shiznebit> yea more ifdEFFFFF's
[18:06] <shiznebit> care to share ?
[18:08] <robclark> hyc: this is with the dspbridge kernel tree? What is the #ifdef (I guess it is some kernel config that the bridge team doesn't use.. but I'll tell 'em about it tomorrow)
[18:09] <hyc> robclark: yes, in rmgr/drv_interface.c there is a reference to clk_handle that should be protected by #ifdef CONFIG_BRIDGE_DVFS
[18:09] <hyc> line 485
[18:09] <hyc> I guess they must only build with BRIDGE_DVFS enabled
[18:09] <hyc> I thought I had it enabled too, but I guess not
[18:10] <robclark> ahh, probably true
[18:10] <hyc> doublechecking my config now, DVFS defaults to off...
[18:10] <robclark> I guess most of 'em are building with all the PM stuff enabled
[18:11] <hyc> yeah. I went thru the menuconfig and I thought I had turned it all on
[18:20] <hyc> I didn't have the OMAP power management option configured
[18:20] <hyc> so that was disabling the DVFS option...
[18:21] <robclark> oh, yeah.. you probably want PM ;-)
[18:22] <hyc> yeah, slight detail. ;) trying again
[18:22] <hyc> do you know anything about SmartReflex? should I have it enabled in the kernel or not?
[18:22] <hyc> I think it is disabled in our previous builds
[18:23] <robclark> keep it disabled for now.. there were some recent patches, but I don't know if they are in the tree you are using or not..
[18:24] <robclark> actually, I'm not completely sure if 35xx supports SmartReflex, or if that is only on 34xx
[18:24] <shiznebit> it should
[18:25] <robclark> ok, if it does, then enable it in the near future.. it helps cut down power consumption somewhat
[18:29] <hyc> ok
[18:32] <hyc> I actually toggled it on in my config already, so hopefully it doesn't break anything when I boot it up ;)
[18:33] <hyc> and hopefully I don't have to rebuild any other modules
[18:33] <hyc> it would suck if my framebuffer driver decided the kernel was the wrong version and refused to load :P
[18:33] <robclark> hmm, omapfb?
[18:34] <robclark> there was recent work to integrate DSS2 and omapfb.. I'm not sure if it is in the kernel you have or not..
[18:40] <edt> how much of the touchbook kernel stuff is supplied as binary blobs?
[18:41] <hyc> edt: ? it's all built from open git repos
[18:41] <edt> ahh very good news
[18:42] <hyc> I think the only thing that's not in git is the wifi
[18:42] <hyc> and that's still open source /gpl
[18:43] <edt> as longs as its easy enough to build... having it out of tree is a little bit of a pita (much much better than binary crud though).
[18:45] <edt> are you cross compiling or building on the touchbook?
[18:45] <hyc> right now cross
[18:45] <hyc> I've built a full kernel on the tb too
[18:45] <hyc> it just takes a long time
[18:45] <edt> how long to build?
[18:45] <hyc> mm, a few hours
[18:45] <edt> it takes two hours or so on my wife's acer
[18:45] <hyc> I didn't stick around to wait
[18:48] <edt> sleep time here
[18:48] <edt> nite
[18:49] <hyc> night
[18:50] <shiznebit> night
[18:51] <GUido-> any new news?
[18:51] <GUido-> I'm hoping Gregoire hasn't been on irc this last week because he's been too busy boxing touchbooks
[18:54] <hyc> I got an email reply from him last night about one of my early bug reports
[18:55] <hyc> robclark: oops, I forgot about that. this is using TI's driver: omaplfb and pvrsrvkm. I don't have the source to rebuild those. I just filled out the request form from the TI web site to get the SDK.
[18:58] * hyc_tb (n=hyc@ Quit ("Powered by OE: www.openembedded.org")
[19:00] <hyc> ooo, lots of pretty colors...
[19:00] <hyc> the new kernel turned my screen to hash
[19:00] <hyc> suffice to say, it doesn't work...
[19:00] <robclark> umm, it's a "feature"
[19:00] <shiznebit> powervr core is botched for all kernels other than that specific one
[19:01] <hyc> ok, back to square 1. original kernel, and try to figure out what's missing that the new bridge driver needs.
[19:01] <hyc> PITA
[19:01] <shiznebit> thats because its not OSS
[19:03] * robclark is hearing thunder outside.. so hopefully my network stays up..
[19:04] <shiznebit> what ?
[19:04] <shiznebit> o_O
[19:04] <shiznebit> you using satellite ?
[19:04] <robclark> hyc, ir you can find the kernel that felipec or mrc uses on beagle, that should work for you
[19:05] <robclark> no, but my power and DSL seem to go out at the drop of a hat today
[19:05] <hyc> buy a UPS, they're cheap insurance for power at least
[19:05] <hyc> I always get surges and brownouts here...
[19:05] <robclark> power, I'm not worried about.. I'm on a laptop with a good battery.. but DSL seems a bit finicky
[19:06] <robclark> it's already dropped out a couple times tonight
[19:06] <robclark> hyc: http://elinux.org/BeagleBoard/gst-openmax
[19:06] <shiznebit> i assumed DSL would be through the underground...not polls
[19:07] <robclark> I think it goes though duct tape and chicken wire
[19:07] <hyc> lol
[19:07] <shiznebit> lol
[19:13] <robclark> well, if you want a variety of kernel's to choose from: http://dev.omapzoom.org/ ;-)
[19:13] <robclark> I think the kernel-display.git tree should have what you need for omapfb display for TB
[19:14] <robclark> but I'd really need to spend some time digging through and trying different trees to see the right combo
[19:14] <robclark> unfortunately a lot of merging of stuff in the linux-omap tree and omapzoom tree is active work-in-progress..
[19:15] <robclark> (give me a weekend or two with a TB and I'd have something working ;-))
[19:15] <robclark> the omap3 integration tree might be a good bet
[19:42] <hyc> I think I'll just back and let folks finish their merges. after the dust settles I'll see what's left
[19:42] <hyc> this all started with me just trying to get a video decoder working...
[19:44] <robclark> hyc: you should be able to get stuff working with the older tiopenmax (or possible even newer, with RM disabled) using the kernel, etc, that felipec used (see http://felipec.wordpress.com/2009/03/26/omap3-public-dsp-binaries-now-work/)
[19:44] <robclark> (actually, I suspect the TB kernel should contain everything you need to use Felipe's tiopenmax, gst-openmax, etc)
[19:46] <robclark> (except that .flv you sent me yesterday... to use .flv, looks like I need to add support for parsing codec-data to get the video dimensions and framerate)
[19:51] <hyc> mmm... I would certainly like to be able to use those .flv's directly
[19:51] <robclark> hyc: me or someone on my team will add the codec_data parsing.. just give us a bit of time ;-)
[19:52] <hyc> cool
[19:52] <hyc> no rush
[19:54] <hyc> I still need to get my hands on the H.264MP decoder anyway. will see how that 24-48 hour response time works out ;)
[19:56] <shiznebit> so could we play quacklive on the touchbook ?
[19:56] <shiznebit> or is that tooo much ?
[19:56] <hyc> quacklive?
[19:57] <hyc> I'm not a gamer, couldn't tell ya
[19:57] * robclark is not a gamer either
[19:57] <hyc> but I recall a friend of mine at Google playing Quake3 on his G1
[19:57] <shiznebit> i heard that the BB could run Quack3 though
[19:57] <shiznebit> yea i recall that too
[19:58] <shiznebit> but that was with everything stripped to the bone
[19:58] <shiznebit> or else it wouldn't run
[19:58] <hyc> it looked pretty good to me, image wise
[19:58] <hyc> and the OpenGL support on the TB is at least that good
[20:01] <hyc> ok, I've cherry picked a bunch of headers from the git tree into my original tree and it looks like I may get a bridgedriver module finally. sheesh
[20:13] <hyc> but it won't have any power management features, since the base kernel is missing those patches
[20:14] <shiznebit-> we can add pm manually
[20:14] <shiznebit-> ?
[20:14] <hyc> probably, but that'll be another day
[20:15] <shiznebit-> as long as we get the dsp to work
[20:15] <hyc> I wonder why the gst_ti guys require realtime priority for their decode threads
[20:15] <hyc> that seems to be a pretty severe limitation. if you can't doublebuffer everything in advance, there's something wrong with your design.
[20:16] <shiznebit-> double buffering helps alot on an oe ?,
[20:17] <hyc> double buffering is pretty much always a good idea for audio/video processing
[20:18] <hyc> it also gives you a little cushion in your processing latency, so you don't need to jack your threads up to maximum realtime priority just to get smooth playback
[20:18] <hyc> I think the gstreamer_ti guys need to rethink their code...
[20:19] * robclark is not involved in gstreamer_ti
[20:19] <killring> are you assuming that because it needs realtime priority that it's not double buffering?
[20:19] <hyc> killring: yes, there's not many other reasons it would need it.
[20:19] <robclark> there should be no need for anything drastic.. the threads on the ARM side just need to run a couple times every ~30ms for video and ~20-25ms for audio
[20:19] <shiznebit-> well it seems pointless to have it rt and doublebuffering
[20:19] <hyc> but you're right, I'm making that assumption without having taken a deep look into the code.
[20:20] <robclark> with a queue between decoder and sink, that should absorb any transient high latency issue if you have other processes running
[20:20] <killring> iirc, apple had to hack the os x kernel years ago to give quicktime realtime priority and I'm pretty sure they were db
[20:20] <hyc> apple is notorious for incompetent hardware design though, so that doesn't surprise me
[20:20] <killring> had more to do with ensuring that you could maintain a consistent playback stream than merely getting it to work
[20:21] <killring> just sayin'...
[20:21] <hyc> all you need is two buffer addresses, and prefill two buffers. start the first one going; fire an interrupt when it's done, start on the 2nd buffer.
[20:22] <hyc> interrupt handler fills a new address into buffer 1, no problem.
[20:22] <hyc> unless your system is so busy that the interrupt handler can't get control in however many ms you allotted...
[20:22] <killring> and if you're on a loaded system where everything else is getting pretty much the same priority as your media process...
[20:22] <killring> exactly
[20:22] <robclark> hyc: in practice, you usually have a queue with ~3 buffers of video between the decoder and video display to maintain good A/V sync
[20:22] <hyc> sure - and if you're worried about latency then just put more frames into a buffer
[20:23] <robclark> it depends a bit on what video sink you use... gstomapfb doesn't support mult-buffering
[20:23] <hyc> then that needs to be fixed ;)
[20:24] <hyc> ok, my new kernel build just finished. wonder if I'll get a new light show...
[20:24] <robclark> I implemented a v4l2sink a little while back, which would let you have multiple buffers, and the video decoder could render directly into the memory that the video overlay would be programmed to display.. so zero copy from decoder to display while still allowing a queue of some small number (>2) buffers to handle jitter/latency issues
[20:27] * shiznebit (n=chatzill@ool-18b99610.dyn.optonline.net) Quit (Read error: 110 (Connection timed out))
[20:28] <hyc> that sounds very nice, what have you done with it?
[20:29] <robclark> v4l2sink is merged in gst-plugins-good.. should be in current or next release..
[20:29] <robclark> I'm actually wanting to try a 2.6.31 kernel w/ DSS2 + v4l2 display driver all merged into one, so I can verify that it is working well on the new display driver..
[20:31] <robclark> (there is also a bit of work to get the video overlay to integrate well with the x-server.. so it does the right thing when you move/resize the video player app window.. another reason why I'm waiting for my TB with it's relatively (compared to phones) large screen)
[20:31] <hyc> heh
[20:31] <hyc> yeah, working with this OS in a phone would be a pain
[20:36] <robclark> ahh, latest gst-plugins-good includes v4l2sink: http://gstreamer.freedesktop.org/releases/gst-plugins-good/0.10.16.html :-)
[20:43] * hyc_tb (n=hyc@ has joined #touchbook
[20:45] * [AFX] (n=afx@S0106001b116229f2.tb.shawcable.net) has joined #touchbook
[20:50] <hyc> bugger. the bridgedriver module complains that the kernel is the wrong version. I seem to have bunged something up in my .config mangling
[20:51] <hyc> backing up another step. good thing I have 4 SD cards here with different images...
[21:16] * hyc_tb (n=hyc@ Quit ("Powered by OE: www.openembedded.org")
[21:55] * robclark (n=robclark@nat/ti/x-ggvqkdusbrzqtthn) Quit ()
[22:23] * bstag_ (n=bstag@ has joined #touchbook
[22:29] * Meiz_n810 (n=Meizirkk@padedu-62-165-142-173.phnet.fi) has joined #Touchbook
[22:30] * dirk2 (n=dirk@p5B04059C.dip0.t-ipconnect.de) has joined #touchbook
[22:30] * bstag (n=bstag@ Quit (Read error: 60 (Operation timed out))
[22:35] * killring (n=killring@ Quit ()
[22:43] * killring (n=killring@adsl-76-226-135-57.dsl.sfldmi.sbcglobal.net) has joined #touchbook
[22:45] * Meiz_n810 is now known as Meiz_n900
[22:47] * Meiz_n900 is now known as Meiz_n810
[22:49] * killring (n=killring@adsl-76-226-135-57.dsl.sfldmi.sbcglobal.net) Quit ()
[22:52] * hyc_tb (n=hyc@ has joined #touchbook
[22:53] * jvs (n=jvs@ Quit (Remote closed the connection)
[22:56] * hyc_tb (n=hyc@ Quit (Client Quit)

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