[Bug target/51753] Many gcc.dg/simultate-thread tests fail on Solaris 10+/x86

[ro at CeBiTec dot Uni-Bielefeld.DE]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51753

--- Comment #4 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot Uni-Bielefeld.DE> 2012-02-07 18:09:45 UTC ---
> Please look at gcc.log files in your testsuite directory. There are extensive
> single-instruction traces to analyse what went wrong. Comparing your traces to
> linux execution should give you a hint.

The Solaris results were completely nonsensical at first, but ultimately
could be explained by a (long-standing, I suppose) and stupid gdb bug:

    http://sourceware.org/ml/gdb-patches/2012-02/msg00075.html

With this fixed, all but the

FAIL: gcc.dg/simulate-thread/atomic-load-int128.c  -O0 -g  thread simulation
test

failures (at all optimization levels) were gone.  I didn't have the
Linux/x86_64 gcc.log available for comparison, so I still have to
investigate that one.

On Solaris 10/x86, I'm sometimes still seeing FAILs, but they may just
be timeouts.  I've have to check that, too.

I'm still seeing a large number of failures both on IRIX 6.5 and Tru64
UNIX V5.1B (both with gdb 7.4), but I suppose those are better filed as
separate PRs.  I certainly cannot compare gcc.log/g++.log output from
other OSes since have no machines running e.g. Linux for comparison.

    Rainer