Re: [RFC PATCH, x86_64] Use -mno-sse[,2] to fall back to x87 FP argument passing convention

[Mark Mitchell <mark at codesourcery dot com>]

Ian Lance Taylor wrote:

> Side-stepping the whole following discussion, I want to agree with a
> couple of other people that we should not change the ABI with an
> option like -mno-sse2.  We should use a different option, probably
> -mabi= along the lines of ARM, MIPS, PPC, etc.

I know this discussion is pretty well closed, but I'd just like to chime 
in here to agree with Ian: both that -mno-sse2 should affect 
code-generation, but not the ABI, and that -mabi=foo is the right way to 
spell any ABI-changing option.

In general, we should strive for greater consistency across platforms 
with respect to options.  As GCC moves from being primarily a drop-in 
replacement for system compilers to being a cross-platform standard, we 
want consistency as much as possible.  For example, -march= and -mtune= 
options should be supported on all platforms, with the same meanings. 
The options to enable threading should be the same.  The *-{elf,eabi} 
targets should be made consistent with respect to whether or not you 
have to explicitly specify a linker script to get the simulator target. 
 Much of this could be done without making backwards-incompatible 
changes.  Where that's impossible, of course, we'd have to take a bit 
more of a case-by-case approach.

--
Mark Mitchell
CodeSourcery
mark@codesourcery.com
(650) 331-3385 x713