Hi Andreas! On Tue, 06 Oct 2015 12:53:03 +0200, Andreas Schwab <sch...@linux-m68k.org> wrote: > Thomas Schwinge <tho...@codesourcery.com> writes: > > On Fri, 02 Oct 2015 13:58:55 +0200, Andreas Schwab <sch...@linux-m68k.org> > > wrote: > >> LAST_UPDATED: Fri Oct 2 02:31:25 UTC 2015 (revision 228367) > >> > >> Native configuration is aarch64-suse-linux-gnu > > > >> === libgomp tests === > >> > >> > >> Running target unix > >> FAIL: libgomp.oacc-c/../libgomp.oacc-c-c++-common/abort-1.c > >> -DACC_DEVICE_TYPE_host=1 -DACC_MEM_SHARED=1 output pattern test > >> FAIL: libgomp.oacc-c/../libgomp.oacc-c-c++-common/abort-3.c > >> -DACC_DEVICE_TYPE_host=1 -DACC_MEM_SHARED=1 output pattern test > >> FAIL: libgomp.oacc-c++/../libgomp.oacc-c-c++-common/abort-1.c > >> -DACC_DEVICE_TYPE_host=1 -DACC_MEM_SHARED=1 output pattern test > >> FAIL: libgomp.oacc-c++/../libgomp.oacc-c-c++-common/abort-3.c > >> -DACC_DEVICE_TYPE_host=1 -DACC_MEM_SHARED=1 output pattern test > > > > Can you please show me what's in libgomp.log? > > The logfile says the output is empty. It aborts before output is > flushed.
Given that besides your aarch64-suse-linux-gnu configuration there are no other reports of this behavior, and my reasoning: | The two regressed test cases use __builtin_printf instead of fprintf to | stderr, but as far as I know, abort is to flush all open streams before | process termination? (And it doesn, in some quick testing I've just done | -- but I'm not set up to run the libgomp testsuite myself.) ..., are you going to look into that? Maybe a problem in the Linux kernel/glibc? Grüße, Thomas
signature.asc
Description: PGP signature
