[cfe-dev] [ARM] Removing v7s/v7f/v7k and cortex-a9-mp
Bob Wilson
bob.wilson at apple.com
Wed Jan 15 15:04:18 PST 2014
On Nov 14, 2013, at 8:59 AM, Bob Wilson <bob.wilson at apple.com> wrote:
>
> On Nov 14, 2013, at 2:55 AM, Bradley Smith <bradley.smith at arm.com> wrote:
>
>>>> Ok, if your main concern is to get rid of the cortex-a9-mp cpu, then
>>> I would follow Renato's suggestion of just adding a -mattr=+mp option.
>>> I don't see a bigger change like you suggested below is worth it.
>>>>
>>>> Evan
>>>
>>> That would work, but it would be good to clean up this stuff a bit and
>>> I think Bradley is on the right track.
>>>
>>> First off, the issue here is specifically v7f. It has nothing to do
>>> with either v7s or v7k. I will need to double check, but I'm not aware
>>> of any fundamental reason why we need to keep v7f around. If we remove
>>> v7f, then I'm not aware of any reason to keep the cortex-a9-mp cpu.
>>
>> Our immediate concern is indeed just v7f, consistency says that whatever we
>> do to v7f we do to the others, however that depends entirely on others
>> requirements, we don't have a requirement to do anything to v7s and v7k so
>> can ignore those if needed.
>>
>> For v7f, are you suggesting that we take the approach of adding a CPUSubType
>> option to pass information through to LLVM? Or that we may be able to get
>> rid of it entirely and forget about CPUSubType generation for v7f
>> specifically?
>
> I am suggesting that we completely remove the armv7f architecture along with the cortex-a9-mp cpu. I need to do some checking first to make sure no one is using that. It will take a little while to do that checking, but I think that would be the right solution. I will get back to you.
As expected, it took a while, but I have now removed armv7f. I kept the code to recognize the cortex-a9-mp cpu name, but changed to match to the base armv7 architecture.
clang svn r199336
llvm svn r199337
compiler-rt svn r199333
Hopefully that will unblock the other changes that you wanted to make.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20140115/8312d50f/attachment.html>
More information about the cfe-dev
mailing list