[PATCH] Implements -mfpu=softvfp+variants on ARM driver

Bernie Ogden bogden at arm.com
Tue Feb 18 04:20:32 PST 2014

Thanks for the updates.

If we care about cross-target consistency then I think we must continue
supporting all of the possible values of -mfloat-abi for ARM.

Just thought: what does -mfpu=<fp arch> (i.e. without the 'softvfp+' part)
mean now? It could mean 'calling convention unspecified' or it could mean
'hardfloat'. I think it should mean the latter, to be consistent with the
armcc/gcc usage, but I _think_ that means changing the default calling
convention? I don't like changing defaults - does the community (or you)
have any position on this?



> -----Original Message-----
> From: Renato Golin [mailto:renato.golin at linaro.org]
> Sent: 18 February 2014 11:56
> To: Bernard Ogden
> Cc: Clang Commits
> Subject: Re: [PATCH] Implements -mfpu=softvfp+variants on ARM driver
> On 18 February 2014 10:53, Bernie Ogden <bogden at arm.com> wrote:
> > Nitpick: line 724 of Tools.cpp is blank.
> Done.
> > You're testing that we're setting the right target features for FP
> hardware,
> > but I don't see tests that we're getting the right calling
> convention. Is
> > that something that could reasonably be added, or am I just showing
> my
> > ignorance here?
> No, I completely forgot. Thanks! ;)
> > Is float-abi an exclusively ARM thing?
> I don't think so, though it probably more important in ARM than other
> archs.
> > And do we care about cross-target CLI consistency?
> We do, unfortunately. It's a lot easier to cross-compile using the
> same build system.
> > Basically LGTM, modulo the nitpick and the tests.
> Thanks!
> --renato

More information about the cfe-commits mailing list