tsan: speed up race deduplication

Race deduplication code proved to be a performance bottleneck in the past if suppressions/annotations are used, or just some races left unaddressed. And we still get user complaints about this:
https://groups.google.com/forum/#!topic/thread-sanitizer/hB0WyiTI4e4
ReportRace already has several layers of caching for racy pcs/addresses to make deduplication faster. However, ReportRace still takes a global mutex (ThreadRegistry and ReportMutex) during deduplication and also calls mmap/munmap (which take process-wide semaphore in kernel), this makes deduplication non-scalable.

This patch moves race deduplication outside of global mutexes and also removes all mmap/munmap calls.
As the result, race_stress.cc with 100 threads and 10000 iterations become 30x faster:

before:
real	0m21.673s
user	0m5.932s
sys	0m34.885s

after:
real	0m0.720s
user	0m23.646s
sys	0m1.254s

http://reviews.llvm.org/D12554




git-svn-id: https://llvm.org/svn/llvm-project/compiler-rt/trunk@246758 91177308-0d34-0410-b5e6-96231b3b80d8

1 files changed, 2 insertions, 4 deletions

diff --git a/lib/sanitizer_common/sanitizer_suppressions.cc b/lib/sanitizer_common/sanitizer_suppressions.cc
index 2b91fbfe9..46d463813 100644
--- a/lib/sanitizer_common/sanitizer_suppressions.cc
+++ b/lib/sanitizer_common/sanitizer_suppressions.cc
@@ -127,13 +127,11 @@ void SuppressionContext::Parse(const char *str) {
         Printf("%s: failed to parse suppressions\n", SanitizerToolName);
         Die();
       }
-      Suppression s;
+      Suppression s = {};
       s.type = suppression_types_[type];
       s.templ = (char*)InternalAlloc(end2 - line + 1);
       internal_memcpy(s.templ, line, end2 - line);
       s.templ[end2 - line] = 0;
-      s.hit_count = 0;
-      s.weight = 0;
       suppressions_.push_back(s);
       has_suppression_type_[type] = true;
     }
@@ -163,7 +161,7 @@ const Suppression *SuppressionContext::SuppressionAt(uptr i) const {
 void SuppressionContext::GetMatched(
     InternalMmapVector<Suppression *> *matched) {
   for (uptr i = 0; i < suppressions_.size(); i++)
-    if (suppressions_[i].hit_count)
+    if (atomic_load_relaxed(&suppressions_[i].hit_count))
       matched->push_back(&suppressions_[i]);
 }