12:20 <pEDrOs> hi! I want to ask if this procedure is suitable for the A13 Olinuxino wifi in order to enable wifiDirect

12:20 <pEDrOs> http://rtl8192cu.googlecode.com/hg-history/bdd3a2265bdd6a92f24cef3d52fa594b2844c9c1/android_reference_codes_ICS_nl80211/Realtek_Wi-Fi_SDK_for_Android_ICS.pdf

13:36 <bfree> debian-kernel has accepted the EMAC addition for 3.10 and applied it to their rcs, so their next experimental kernel (probably rc6) should have EMAC support

13:39 <bfree> no mention of sata at http://linux-sunxi.org/Linux_mainlining_effort (and I don't know where to put it)

13:47 <Turl> bfree: nice! :)

14:48 <hramrach_> oliv3r: for me the cubieboard jb android builds after hasking the build environment into shape

15:19 <hramrach_> debian-live for a10 anyone? ;-)

15:28 <bfree> hramrach_: does debian-live support netboot? or are you just thinking of 3.4? I've already had another debian based live system running on my cubie ;) I've yet to decide/figure out though what I think the sanest approach is for an A10&|armmp (an live system and "full" installer? an image and thin installer?). and of course u-boot+dtb/fex needs per-device love :-/ and kernel+u-boot->boot needs Debian love

16:16 <oliv3r> hramrach_: on 64bit?

16:17 <oliv3r> or did you fall back to 32bit

19:41 <oliv3r> bfree: setenv initrd_high 0xffffffff

19:41 <oliv3r> why is that needed

19:43 <bfree> oliv3r: ask Turl cause I'm damned if I know :-p but it reproducably (for me and Turl) ooops without it on the cubie

20:14 <hno> oliv3r, it does smell like there is an u-boot bug to me. Setting initrd_high makes things much less confusing. See description of initrd_high in u-boot/README

20:15 <hno> setting initrd_high=0xffffffff I meant. Can have many other values as well but that one is special.

20:16 <hno> but could also be a kernel bug, haven't studied it in detail yet.

20:17 <hramrach_> oliv3r: I built on 64bit with 32bit libraries installed

20:18 <hramrach_> installing 32bit development libraries breaks 64bit build environment on Debian but not an issue with dedicated chroot

20:19 <hramrach_> also your error does not look 64bit specific

20:20 <bfree> hramrach_: appears debian-live will netboot using nfs ... knock yourself out ;) http://live.debian.net/manual/2.x/html/live-manual#210

20:24 <hramrach_> bfree: or from http server

20:25 <bfree> ok, didn't spot that in their manual there

20:25 <hramrach_> what does u-boot netboot from?

20:26 <hramrach_> maybe it's not in the manual. don't read it much

20:28 <hramrach_> hmm, tftp

20:28 <hramrach_> even dhcp

20:28 <oliv3r> hno: well so far, i havent' even heard of that line, nor had to use it

20:28 <oliv3r> so was supprised to see it there

20:29 <oliv3r> hramrach_: so you did do a dedicated chroot

20:29 <hramrach_> yes, i did that form the start

20:29 <hramrach_> it's not like I want android crawling all over my system ;-)

20:30 <oliv3r> heh, it stays in its own library :p

20:31 <hramrach_> it turns out it requires special version of everything

20:35 <hramrach_> hmm, u-boot can load file specified by dhcp but how do you load initramfs in this case?

20:35 <hramrach_> initramfs is another file

20:35 <bfree> hramrach_: I just remembered debian live is i386/amd64 only so who knows what fun you might run into trying it on armhf ;)

20:36 <hramrach_> it's a shell script

20:36 <hramrach_> it will run just fine if you can load a kernel and initramfs

20:37 <hramrach_> there are just no arm images because nobody built them

20:37 <bfree> isn't tftpboot just a substitute for e.g. fatload/ext2load? i.e. it loads a file and doesn't boot it (you still need to bootm)? sorry I've not played with it at all (yet anyway)

20:37 <hramrach_> because nobody knows how to make a bootable arm image :p

20:37 <hramrach_> the guide says to tftp and elfboot

20:38 <hramrach_> you can technically use two files if you script tftp in boot/scr or so

20:38 <hramrach_> probably

20:39 <hramrach_> but dhcp specifies only one file so if you use that you don't get te einitramfs

20:39 <hramrach_> and if you don't you need always the same file loacation on local network

20:39 <hramrach_> because tftp is non-routable

20:42 <hramrach_> hmm

20:43 <hramrach_> I guess we need support for loading boot.scr from tftp

20:43 <hramrach_> then you can tftp both kernel and initramfs, hopefully

20:45 <hramrach_> who understands u-boot?

20:46 <hno> hramrach_, u-boot can tftboot any number of files. It's just files. But I do not thing DHCP can give more than one image name so you need to use explicit tftpboot commands (and disable the autoboot on dhcp).

20:47 <hramrach_> yes, that's the problem

20:47 <hramrach_> you need two files at least and you can't expect people to edit boot.scr

20:47 <hramrach_> expecially since it needs to be converted to some image

20:48 <hramrach_> so the boot.scr needs to be loaded from tftp

20:58 <hramrach_> well, since boot.scr is some sort of image it should have some magic and booting it from dhcp should be doable

21:01 <hno> hramrach_, there is also the uEnv.txt method..

21:02 <hno> that's a plain text file. But not so friendly if doing a longer boot script.

21:02 <oliv3r> what location is env stored again?

21:02 <oliv3r> raw partition?

21:02 <hno> uenv.txt is a kind of stored environment.

21:03 <hramrach_> raw data on card iirc

21:03 <hramrach_> so works nicely when you have serial console and can cut&paste settings and saveenv

21:03 <hno> but the normal stored environment is a binary blob, stored at raw offsets, or in a partition, or in a file.

21:07 <oliv3r> http://lxr.missinglinkelectronics.com/#uboot+v2012.04/README#L3791 is just to speed things up, hno

21:07 <oliv3r> it's an optimization thing sort of

21:08 <hramrach_> and the non-optimizeda case is fudged ;-)

21:29 <oliv3r> i never used that flag; and never had a problem

21:29 <oliv3r> though i have built-in initramfs

21:48 <bfree> wtf, I try to edit http://linux-sunxi.org/index.php?title=How_to_boot_the_A10_over_the_network just to say the filename is not random but the ip address in hex .img, and it tells me I'm adding new external links!

21:51 <hno> oliv3r, not only an optimization, also less translatoin, and for some reason requires kernel support.

21:52 <hno> built-in initramfs works without it.

22:01 <oliv3r> ahh, so that's why i might not have had issues

22:02 <oliv3r> Turl: i like lxc; i don't fully get certain things, but i like it

22:05 <Turl> oliv3r: 0xffffffff is "do not relocate"

22:05 <Turl> fdt_high also supports the special value

22:06 <Turl> I believe the region used automatically by uboot is highmem, maybe that is causing the issue?

22:07 <Turl> hno: I suppose 'kernel support' means 'make sure kernel does not uncompress itself/work over that region'

22:08 <Turl> bfree: is your account verified?

22:13 <bfree> Turl: yep, and there were no new urls ... whatever, answered the captcha and moved on ;)

22:15 <Turl> bfree: that's the beauty of our captcha also ;) you can easily complete it :p

23:32 <leowt> just received an rk8168 tablet

23:32 <leowt> do you know of any irc channel that most work around with rockchip?

23:41 <rz2k> rockchip doesnt have big community iirc

23:41 <rz2k> check forums

23:42 <leowt> this chip is brand new

23:42 <leowt> so there isnt anything for it

23:42 <leowt> as far as i can see rk30xx sdk was used to build

23:42 <leowt> and kernel seems the only difference

23:42 <leowt> i cant get usb debugging working

23:42 <leowt> :/

23:43 <leowt> it is activated but i cant get no adb connection

23:43 <leowt> i have almost no experience with android and i dont know those kind of problems