Freerunner android wifi

[michael]

Hi,

Andy Green wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Somebody in the thread at some point said:
> | Andy Green wrote:
> | Somebody in the thread at some point said:
> | | Hi all,
> | |
> | | I have installed android on a freerunner and I don't understand 
> why the
> | | wpa_supplicant doesn't work when it try
> | | to scan network. I see at the code and the SIOCSIWSCAN is 
> implemented.
> | |
> | | I take a look at the code, what is the open issue?
> |
> | Where did this kernel come from and what does /proc/version say?
> |
> |> Linux version 2.6.28-GTA02_koolu_a68394f09592d151-mokodev
> |> (michael at panicking) (gcc version 4.1.2) #4 PREEMPT Wed Dec 31 11:10:10
> |> CET 2008
> |
> |> The version is take from the koolu repository and I think that is
> |> aligned with the stable version (look at
> |> the git).
> |> I can put some debug stuff in the wireless driver and see if the ioctl
> |> is regulary called.
>
> I doubt it's much to do with our stable, which is still 2.6.24.  It
> looks like andy-tracking from about 20 days ago.
>
> http://git.koolu.org/?p=kernel/openmoko.git;a=summary
>
> Marcelo, we should have Android stuff enabled and almost completely
> working in our gta02 configs down ./arch/arm/configs/, the only
> difference is that we leave up the framebuffer console.
>
Where is the git repository? I can clone and use for my test?
> I'd really like to avoid having YAKB (Yet Another Kernel Binary) because
> when we consider the number of targets we will hopefully end up
> supporting, it's going to double the number of kernel binaries
> singlehanded.  And as you can see I dunno what to do if someone turns up
> asking why something is busted.
I agree with you. I use koolo repository for a starting point.
>
> I don't particularly like having YAFFS (and Sean told it wasn't actually
> necessary) but if it is necessary I am willing to carry it as a local
> patch directly.
>
> The other two open issues are:
>
> ~ - Android should be able to coexist with framebuffer console, it seems
> we need some way to ask it not to switch away to last VC written to
>
> ~ - The menu patch can go in our tree directly if someone just adds some
> code to make it defeated by default but able to be enabled by the rootfs
> presumably down /sys, only Android would do that
What is the menu patch?
>
> If we can solve those we can really reduce your patchset down to nothing
> so stock kernels from here will do what you need.
>
> - -Andy
Michael