http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59215
--- Comment #8 from Kostya Serebryany <kcc at gcc dot gnu.org> --- (In reply to Jonathan Wakely from comment #6) > I realise that, but the general point is still valid: for race detectors to > understand the atomic updates in the library they library needs to be > compiled with the race detector enabled. We can update the docs to cover > tsan as well, but the point is till that you need to recompile the library > to prevent false positives from race detectors. Yes, correct. Thanks! We tried to suppress all known false reports like this so that one can *try* to use tsan w/o recompiling libstdc++. Probably we may need to suppress more, so a small test will help. But the preferred way is still recompilation of libstdc++. Since we ourselves use tsan with a recompiled library, we do not observe any false positives like this.
