This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: optimization/8599: [3.2 regression] loop unroll bug with -march=k6-3
At 11:16 29.11.2002, Eric Botcazou wrote:
> Reopen, backported fix From Dale causes a regression on
> powerpc-linux-gnu. The testcase gcc.c-torture/execute/961213-1.c now fails
> with -O2 -funroll-loops. Presumably the backported patch relies on some
> other changes in the mainline.
Are you sure ?
http://gcc.gnu.org/ml/gcc-testresults/2002-11/msg01035.html
http://gcc.gnu.org/ml/gcc-testresults/2002-11/msg01038.html
Yes, I'm sure, that's why I mentioned -O2 -funroll-loops, which is untested
by a normal testsuite run, but I routinely run my testsuite with
{-O2,-fpic,-fPIC,} as I've been bitten too often by the laxness of the
testsuite in this respect.
Now, the problematic part is the doloop.c hunk as I just verified. So we
have to find out how to fix it or we have to revert the patch on the branch
:-(. I have an unrelated glibc build running, after that is finished I'll
build the mainline and verify what's going on there as well.
Franz.