RFA: patch to solve PR37535

[Andrew Pinski]

On Fri, Sep 26, 2008 at 11:30 AM, Vladimir Makarov <vmakarov@redhat.com> wrote:
> powerpc64-unknow-linux-gnu.  It was G5 under Fedora core8.  I'll check the
> target again.  I tried to test it well (it took 2 days) because of the patch
> nature (clobbering affects conflict building and as the consequence the
> generated code rigthness).

This also caused the following failures on x86-darwin:
+FAIL: gcc.c-torture/execute/20000808-1.c execution,  -Os
+FAIL: gcc.c-torture/execute/strct-varg-1.c execution,  -Os
+FAIL: gcc.c-torture/execute/va-arg-22.c execution,  -O1
+FAIL: gcc.c-torture/execute/va-arg-22.c execution,  -O2
+FAIL: gcc.c-torture/execute/va-arg-22.c execution,  -O3 -fomit-frame-pointer
+FAIL: gcc.c-torture/execute/va-arg-22.c execution,  -O3
-fomit-frame-pointer -funroll-loops
+FAIL: gcc.c-torture/execute/va-arg-22.c execution,  -O3
-fomit-frame-pointer -funroll-all-loops -finline-functions
+FAIL: gcc.c-torture/execute/va-arg-22.c execution,  -O3 -g
+FAIL: gcc.c-torture/execute/va-arg-22.c execution,  -Os
+FAIL: gcc.target/i386/20060125-1.c execution test

Compiling with a stage1 compiler allows those to pass.  Also the wrong
code comes in via combine so either combine.c or simplify-rtx.c is
being miscompiled.

You might be able to get this failure on x86-linux-gnu via
bootstrapping with PIC turned on since darwin has PIC turned on by
default.

Thanks,
Andrew Pinski