[patch] LTO support for Mach-O (apple-darwin)
Jack Howarth
howarth@bromo.med.uc.edu
Fri May 14 13:38:00 GMT 2010
On Fri, May 14, 2010 at 09:25:48AM -0400, Diego Novillo wrote:
> OK folks. I'm now even *more* confused. Is the patch ready for review or not?
>
>
> Diego.
Diego,
The operative patch for review should be...
http://gcc.gnu.org/ml/gcc-patches/2010-05/msg01051.html
which was originally proposed by Steven in...
http://gcc.gnu.org/ml/gcc-patches/2010-05/msg00577.html
I have verified that the gcc/config/darwin.c changes
produce no regressions on x86_64-apple-darwin10...
http://gcc.gnu.org/ml/gcc-testresults/2010-05/msg01025.html
and Dominique Dhumieres has verified that it brings powerpc-apple-darwin's
lto testsuite performance on par with i386-apple-darwin.
http://gcc.gnu.org/ml/gcc-testresults/2010-05/msg00939.html
Jack
ps Both i386-apple-darwin* and powerpc-apple-darwin* has unrelated
issues with lto apparently optimizing away symbols. Applying the
requested patch, http://gcc.gnu.org/ml/gcc-patches/2010-05/msg01051.html,
will expand the universe of testers for these remaining lto issues
by adding powerpc darwin to i386 darwin as supported lto targets.
More information about the Gcc-patches
mailing list