[yocto] [OE-core] bring mipsel support (little endian mode)

[Bruce Ashfield]

On Wed, Jun 27, 2012 at 2:40 AM, Khem Raj <raj.khem at gmail.com> wrote:
> On Tue, Jun 26, 2012 at 11:25 PM, Dennis.Yxun <dennis.yxun at gmail.com> wrote:
>> HI ALL:
>>   Sorry for crossing post, but I'm quite new here, so if I did
>> something wrong, please point me the right direction. thanks
>>   I'm using oe-core, and notice that mipsel support(o32, little
>> endian mode) is missing (current available choose is: qemumips,
>> qemumips64, qemumips64el).
>>   So, here I'm trying to bring up qemumipsel support . My first
>> attempt would just make qemumipsel work,
>> further would make it running on real board, thus maybe mips4kec,
>> mips24k, mips74k chips.
>>   What I achieved here current:
>>  1) toolchain/basic libs, utilities should works ,only one changes,
>> see my patch [a]
>>   for mipsel we also have to filter out -march=mips32, previously we
>> only handle mips (the big endian?)
>>   don't have the error message now, but if you request, I can provide.
>>  2) attempt to  compile kernel to support qemumipsel, unfortunately it fail.
>> The point here is linux-yocto doesn't support qemumipsel, but merely
>> support those other three mips arches,
>> So here is my attempt patch [b], and the fail log [c], complied out
>> binary still big endian.
>>
>>   Could my patch [a] be accepted? or should I send with another mail
>> for better review?
>>   Is it better to request a ticket in yocto's bugzilla? or just
>> report here, I'm not quite sure.
>>
>> Dennis
>>
>> [a] 0001-eglibc-support-mipsel-little-endian-filter-out-march.patch
>
> This patch is appropriate for OE-Core, I will add it to my tree.
>
>> [b] 0002-qemu-attempt-to-add-mipsel-little-endian-support-but.patch
>
> This looks good to me but I will leave it to Bruce to decide.

I've always had latent support for mipsel in the kernel tree, since it is
just a minor switch to the configs for the mti.

We'd also sent the KMACHINE to mti-malta32-le to pickup the le
configuration switches from the kernel .. or alternatively, I'll just update
the description to match on qemumipsel.

I'd want to do some build and boot testing myself first though, Phil seemed
to indicate that patch a) wasn't required. I'll start my own build and see if
my results match.

So if there's a desire to get this into oe-core, I can take care of the kernel
tweaks (and then carry it), but I'd expect Richard has an opinion on this, since
it really should be something that we then build and minimally test
on a regular basis.

Cheers,

Bruce

>
>> [c] build_log.txt
>> [d] log.do_package
>>
>> _______________________________________________
>> Openembedded-core mailing list
>> Openembedded-core at lists.openembedded.org
>> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
>>
> _______________________________________________
> yocto mailing list
> yocto at yoctoproject.org
> https://lists.yoctoproject.org/listinfo/yocto



-- 
"Thou shalt not follow the NULL pointer, for chaos and madness await
thee at its end"