10:50 <jukivili> Is one of the usb-ports on mele a1000 wired to usb-otg host?

11:00 <rm> jukivili, afaik all ports on the A10 are reconfigurable

11:00 <rm> via script.bin

11:00 <rm> you can have it however you prefer

11:06 <rm> also

11:06 <rm> isn't it either "usb otg" or "usb host"

11:06 <rm> and "usb-otg host" doesn't make sense?

11:11 <w00tc0d3> hi

11:36 <mnemoc> jukivili: the usb0 (aka otg) por of the mele in not exposed. there is a 5pin header inside

11:37 <mnemoc> all the external usb ports come from a hub

12:48 <chujalt> hello

13:16 <jukivili> mnemoc: ok

13:22 <jukivili> rm: I meant, one port being otg hardwired to host-mode

13:22 <rm> well yes if those are via hub, then you need one that isn't, to use OTG

13:24 <rm> http://linux-sunxi.org/Fex_Guide#24_USB_control_flags

13:25 <rm> usb_port_type = xx USB port to use. (Xx = 0/1/2) 0: device only 1: host only 2: OTG

13:25 <rm> nothing is "hardwired"

13:40 <jukivili> rm: I was thinking of possibility that mele a1000 had ID pin of otg port hardwired so that otg is forced to host-mode, then expose that port as normal usb port on device. Plus regular 2 host ports of A10. But as already found out, this is not the case, mele has build in USB-hub.

18:46 <lina> Hi all

18:59 <WarheadsSE> anyone ever get around to making a grand-unified sunxi mkcard type script?

19:00 <techn__> WarheadsSE: ?

19:01 <WarheadsSE> ala, 1 script that will happily take a set of pre-made .bin's and say, ./mksunxi-card.sh hackberry will do the necessary spl/uboot (maybe fex too)

19:01 <WarheadsSE> No, it isn't hard, in any way

19:01 <WarheadsSE> ust wondering

19:02 <techn__> sunxi-bsp

19:02 <WarheadsSE> see, precisely what I derp'd and missed.

19:08 <WarheadsSE> looks like that mostly works, albeit the Arch hwpack will be a bit off from the debian model

19:10 <techn__> Yeah.. patches welcome :)

19:12 <WarheadsSE> well the major difference is that the kernel & modules are part of the Arch pacman packaging system

19:13 <WarheadsSE> thus, the whole /lib/modules/ et al is unneeded, and might safely be skipped in the hwpack generation

21:37 <mnemoc> techn__: around?

21:37 <techn__> mnemoc: here

21:38 <mnemoc> /q ?

21:59 <hramrach> WarheadsSE: kernel and modules are part of Debian packaging system too

21:59 <hramrach> the sunxi kernels are just not packaged

22:02 <WarheadsSE> :) they are on Arch

22:04 <hramrach> packaging kernels on debian is somewhat convoluted

22:05 <hramrach> even Linus imported a Debian packaing hack into mainline so that Linux users on Debian don't have to go through that :>

22:19 <bfree> you can try "my" packaging if you want redistributable kernel deb's rather then just your own local make-kpkg ;) just be warned it doesn't cross-compile so you'll face a slow native build. http://dl.linux-sunxi.org/niall/debian/ (I've the roughest of rough sunxi-tools including sunxi-boards locally now that I'm still pondering whether to upload there or whether to give it some more love first)

22:20 <hramrach> I want the latest kernel

22:21 <hramrach> since debian packaging is convoluted it directly conflicts with that

22:22 <hramrach> I am sure some distributions are geared towards making kernel packages of git trees but Debian is not one fo them

22:22 <hramrach> unless you use make-kpkg but that's not 'official enough'

22:25 <bfree> you can do it and "my" package is dirty in that the orig.tar.gz is just a git archive (which is still latest sunxi-3.4) ... but for an actual official Debian kernel I think things will need to go upstream. if it's just for your own system then make-kpkg should do

22:26 <hramrach> I do the cp arch/arm/boot/uImage /boot/

22:41 <bfree> no mention of uInitrd in the wiki ... I can't be the only one with rootfs on sata :-p no idea where to add it (mkimage call and uboot config to load it ... and maybe howto tell initramfs-tools to load modules early)

23:26 <Turl> bfree: why do you want an initrd?

23:30 <bfree> Turl: I have one ... sata modular and root on sata (also been hacking up a live-cd/usb like system which needs one)

23:31 <Turl> I'd use sata as =y and then just pass root= to the kernel cmdline

23:32 <Turl> this might help if you want to use an initrd though

23:32 <Turl> http://www.armadeus.com/wiki/index.php?title=Using_Initial_RAM_disk

23:33 <Turl> just load it with your favorite means and pass the address to bootm and appropriate cmdline params

23:33 <bfree> that would work for the normal install (though ftr pretty sure defconfig leaves sata modular) ... wouldn't work for the live system ;) I have it working fine

23:34 <Turl> oh

23:35 <bfree> sorry, re-reading what I said I wasn't clear I wasn't sure where to add the info to the wiki ... not that I needed the info ;)

23:36 <Turl> well, it's not sunxi specific

23:36 <Turl> but it might be worth documenting

23:37 <Turl> maybe as an advanced section on http://linux-sunxi.org/FirstSteps

23:37 <Turl> although it wouldn't be first steps

23:38 <Turl> maybe http://linux-sunxi.org/Booting_from_SATA ?

23:39 <Turl> or Rootfs_on_SATA, booting would be not entirely accurate

23:43 <hramrach> it's Using_Initramfs_with_u-boot