Support unaligned load/store on more ARM targets

[JF Bastien]

> How should I name it? GCC has -mstrict-align and -mno-strict-align, and
>> LLVM currently has -arm-strict-align. I could call it -arm-no-strict-align
>> but that's a mouthful. Maybe -arm-allow-unaligned?
>>
>
> Why do we have the "arm" in the name? Can we not just call it like GCC?
> I'm not very found of calling two opposite flags completely different
> names... Between "arm-allow-unaligned" and "arm-no-strict-align", I prefer
> the latter. ;)
>

Agreed that this isn't strictly an ARM thing (as evidenced by GCC's
options), but I'd rather not create a command line option called
-m(no)strict-align and not have it work on other non-ARM targets. Keeping
the "arm" in seems like a good enough compromise without changing the other
backends. IIUC x86 instead qualifies whether the target CPU has fast
unaligned access or not, and that dictates which code gets generated, a
generic flag like -m(no)strict-align could override that behavior.

How about I go with keeping -arm-strict-align and add -arm-no-strict-align
for now? It seems like a clean path to later add GCC compatibility with the
-m(no)strict-align, but requires changing the other backends.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20130515/1d76047b/attachment.html>