Re: sh-unknown-linux-gnu (Support of no fdiv for integer division)

[Alexandre Oliva <aoliva at redhat dot com>]

On Oct  3, 2001, Joern Rennecke <amylaar@redhat.com> wrote:

>> Hmm...  Yes, indeed, we seem to have a problem in PASS_IN_REG_P() and
>> FUNCTION_ARG() that would render structure-passing incompatible with
>> SH4 and compatible with SH3.  This is quite unfortunate.

> It is necessary in order to use the sh2 multilibs.

Huh?  What is necessary?  What does sh2 have to do with whether
-m4-nofpu follows the SH3 ABI or the SH4 ABI except for FP argument
passing?

>> I'd much rather adjust the macros above so that -m4-nofpu has the
>> obvious meaning (which would unfortunately pose a binary-compatibility
>> problem) than introducing yet another flag.  But, if we decide not to
>> break binary compatibility, I'd rather deprecate -m4-nofpu and
>> introduce say -m4-no-fpu, that uses the SH4 ABI except for FP types,
>> and use -m3 -mtune=sh4 to get what we currently get with -m4-nofpu.

> That will require yet another item in the cpu/abi multilib dimension.
> I think that can give us currently up to 32 new multilibs, all things
> counted.

Well, we could stop building a -m4-nofpu multilib, since it's
compatible with -m3 anyway, and we could refrain from building
-m4-no-fpu, assuming it would only be used for kernel build and this
kind of stuff that doesn't require target libraries and, if it does,
it could well use the -m4 multilib.  I suppose.

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                  aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist    *Please* write to mailing lists, not to me