This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Inline assembly syntax
- From: Zack Weinberg <zack at codesourcery dot com>
- To: Stan Shebs <shebs at apple dot com>
- Cc: Ian Lance Taylor <ian at wasabisystems dot com>, gcc at gcc dot gnu dot org
- Date: Fri, 07 May 2004 18:33:08 -0700
- Subject: Re: Inline assembly syntax
- References: <m3y8o3u7eg.fsf@gossamer.airs.com> <409C3575.5070907@apple.com>
Stan Shebs <shebs@apple.com> writes:
> Incidentally, as I was thinking about the assembly grokking problem,
> it occurred that we already have a rich mine of info about operand
> constraints - the machine descriptions themselves. Not always
> expressed as per-instruction, but for at least the one-instruction
> output strings, one could make up a table of instructions/constraint
> letters/types that would be pretty accurate, and not need special
> per-target hackery.
That's a good idea. I don't see multi-instruction output strings as a
huge problem, because I've thought for awhile that we ought to be
pushing people to use strictly one-instruction output strings in their
machine descriptions anyway (break up complicated operations with
splitters or expanders).
A more serious problem is the ability to execute arbitrary C code at
output time (* patterns). I don't know if that can all be shoved into
splitters and peephole2 patterns, but it would be nice.
zw