On Tue, Nov 12, 2013 at 10:57 AM, Jakub Jelinek <ja...@redhat.com> wrote: > On Tue, Nov 12, 2013 at 10:07:32AM -0800, Kostya Serebryany wrote: >> On Tue, Nov 12, 2013 at 10:04 AM, Michael Meissner >> <meiss...@linux.vnet.ibm.com> wrote: >> > On Tue, Nov 12, 2013 at 09:43:38AM -0800, Kostya Serebryany wrote: >> >> or, alternatively, we can disable libsanitizer on PowerPC if the >> >> PowerPC community does not care enough about it being healthy. >> > >> > I think there should be a global --enable-libsanitizer or whatever option >> > that >> > would allow people to enable it. It should only be default on x86_64 until >> > people are motivated to fix libsantizer on all instances of the platform. >> >> I don't mind that. > > Perhaps I'm totally blind, but how exactly are the > sanitizer_common_syscalls.inc interceptors used (seems most of the portability > trouble is caused by that)? I see it defines tons of functions like > __sanitizer_syscall_post_impl_* > __sanitizer_syscall_pre_impl_* > but I couldn't find anything actually referencing those symbols. Is that > just dead code for the time being?
This is all dead code in gcc repo. This is why I am asking for any quick #ifdef. in clang repo this code is used by MemorySanitizer (and will be used by asan/tsan later). > > Jakub
