[fedora-arm] armv7hl requirements

[Peter Robinson]

On Tue, May 10, 2011 at 4:28 PM,  <omalleys at msu.edu> wrote:
> Quoting Gordan Bobic <gordan at bobich.net>:
>
>> Dennis Gilmore wrote:
>>>> Hi Jon,
>>>>
>>>> On Wed, May 04 2011, Jon Masters wrote:
>>>>> I'd like to kick off a discussion about flags for ARMv7. My proposal
>>>>> here is that we treat v7hl as an entirely different architecture, and
>>>>> don't try any multi-arch kind of hacks (there isn't the established user
>>>>> base for Fedora ARM to justify doing any of those things at the moment).
>>>>>
>>>>> Things I think we should consider as a minimum:
>>>>>
>>>>> *). Little endian (obviously, but worth stating) (l)
>>>>> *). Cortex-A8 or higher fully compliant core(s)
>>>>> *). ARM VFP3 hardware floating point (h)
>>>>> *). ARM NEON Architecture
>>>>> *). Thumb2 interworking
>>>>> *). Your suggestion here?
>>>>>
>>>>> I think we should build for ARM (as opposed to Thumb2) but we should
>>>>> support interworking with Thumb2 code through the toolchain options. We
>>>>> should then later consider implementing some Thumb2 optimization. It's
>>>>> more armv7thl, but the (t) is implied since it's ARMv7 anyway.
>>>>>
>>>>> Several folks have begun looking at toolchain bringup based on the F-15
>>>>> toolchain applied to an F-13 userspace initially. But I'd like us to
>>>>> discuss options/requirements for toolchains before we go too far.
>>>>>
>>>>> Once I get some feedback, I'll be updating the wiki, along with some
>>>>> more F-15 goals and (hopefully) generally useful stuff.
>>>> Just for the record, this sounds great from OLPC's perspective; +1.
>>>> (I expect we'd rather build for Thumb2, even if only for the size
>>>> benefit.)
>>>
>>> ive started building some f15 rpms with hardfp
>>>
>>> i set in redhat-rpm-config
>>> -march=armv7-a -mfpu=vfpv3-d16  -mfloat-abi=hard -mthumb
>>>
>>> im using meego as a base to bootstrap, we will need to build a couple of
>>> times to get everything bootstrapped right with the full set of flags.
>>> meego dropped some things like selinux. im slowly making some progress. i
>>> want to get to having a fedora minimal buildroot by the end the week.
>>> though that might be a bit hard since gcc will take some time to compile.
>>
>> Does that mean we are skipping F14 alltogether? I'm not against the
>> idea, just curious. Anything that helps close the gap to primary distros
>> is a good thing. :)
>
> I don't think armv5 is skipping f14.

No definitely not skipping armv5 F-14. I'm working on it but having a
few issues. All help and assistance appreciated.

> It is probably a good idea to skip F14 for armv7 though and start off
> at least in the general ballpark of the mainline distro.

The armv7+hardfp would be hard for F-14 as it would need a lot of
patches to gcc etc. armv7+soft as we have now probably won't provide
massive improvements. Ultimately I think we'll end up with a arm5tel
for maximum hardware support as we have now, and then a arm7hl (plus
possibly thumb) for best performance on armv7 platforms.

> The rpm tweaks need to make it upstream though. :)

Correct, and it would be good to get this into there soon assuming the
above settings are what we all agree upon for v7 + hardfp for F-15.

Peter