tsan: do not instrument not captured values

I've built some tests in WebRTC with and without this change. With this change number of __tsan_read/write calls is reduced by 20-40%, binary size decreases by 5-10% and execution time drops by ~5%. For example:

$ ls -l old/modules_unittests new/modules_unittests
-rwxr-x--- 1 dvyukov 41708976 Jan 20 18:35 old/modules_unittests
-rwxr-x--- 1 dvyukov 38294008 Jan 20 18:29 new/modules_unittests
$ objdump -d old/modules_unittests | egrep "callq.*__tsan_(read|write|unaligned)" | wc -l
239871
$ objdump -d new/modules_unittests | egrep "callq.*__tsan_(read|write|unaligned)" | wc -l
148365

http://reviews.llvm.org/D7069

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@228917 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Transforms/Instrumentation/ThreadSanitizer.cpp b/lib/Transforms/Instrumentation/ThreadSanitizer.cpp
index e9999c486fad1e7774190364e30799dd6361f2f2..e4a49112144e49c04ccd48d09e17d571d44d2b2e 100644 (file)
--- a/lib/Transforms/Instrumentation/ThreadSanitizer.cpp
+++ b/lib/Transforms/Instrumentation/ThreadSanitizer.cpp
@@ -19,6 +19,8 @@
 // The rest is handled by the run-time library.
 //===----------------------------------------------------------------------===//
 
+#include "llvm/Analysis/CaptureTracking.h"
+#include "llvm/Analysis/ValueTracking.h"
 #include "llvm/Transforms/Instrumentation.h"
 #include "llvm/ADT/SmallSet.h"
 #include "llvm/ADT/SmallString.h"
@@ -68,6 +70,7 @@ STATISTIC(NumInstrumentedVtableReads, "Number of vtable ptr reads");
 STATISTIC(NumOmittedReadsFromConstantGlobals,
           "Number of reads from constant globals");
 STATISTIC(NumOmittedReadsFromVtable, "Number of vtable reads");
+STATISTIC(NumOmittedNonCaptured, "Number of accesses ignored due to capturing");
 
 namespace {
 
@@ -272,6 +275,7 @@ bool ThreadSanitizer::addrPointsToConstantData(Value *Addr) {
 // Instrumenting some of the accesses may be proven redundant.
 // Currently handled:
 //  - read-before-write (within same BB, no calls between)
+//  - not captured variables
 //
 // We do not handle some of the patterns that should not survive
 // after the classic compiler optimizations.
@@ -303,6 +307,17 @@ void ThreadSanitizer::chooseInstructionsToInstrument(
         continue;
       }
     }
+    Value *Addr = isa<StoreInst>(*I)
+        ? cast<StoreInst>(I)->getPointerOperand()
+        : cast<LoadInst>(I)->getPointerOperand();
+    if (isa<AllocaInst>(GetUnderlyingObject(Addr, nullptr)) &&
+        !PointerMayBeCaptured(Addr, true, true)) {
+      // The variable is addressable but not captured, so it cannot be
+      // referenced from a different thread and participate in a data race
+      // (see llvm/Analysis/CaptureTracking.h for details).
+      NumOmittedNonCaptured++;
+      continue;
+    }
     All.push_back(I);
   }
   Local.clear();

diff --git a/test/Instrumentation/ThreadSanitizer/capture.ll b/test/Instrumentation/ThreadSanitizer/capture.ll
new file mode 100644 (file)
index 0000000..d6c62f0
--- /dev/null
+++ b/test/Instrumentation/ThreadSanitizer/capture.ll
@@ -0,0 +1,91 @@
+; RUN: opt < %s -tsan -S | FileCheck %s
+
+target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
+
+declare void @escape(i32*)
+
+@sink = global i32* null, align 4
+
+define void @captured0() nounwind uwtable sanitize_thread {
+entry:
+  %ptr = alloca i32, align 4
+  ; escapes due to call
+  call void @escape(i32* %ptr)
+  store i32 42, i32* %ptr, align 4
+  ret void
+}
+; CHECK-LABEL: define void @captured0
+; CHECK: __tsan_write
+; CHECK: ret void
+
+define void @captured1() nounwind uwtable sanitize_thread {
+entry:
+  %ptr = alloca i32, align 4
+  ; escapes due to store into global
+  store i32* %ptr, i32** @sink, align 8
+  store i32 42, i32* %ptr, align 4
+  ret void
+}
+; CHECK-LABEL: define void @captured1
+; CHECK: __tsan_write
+; CHECK: __tsan_write
+; CHECK: ret void
+
+define void @captured2() nounwind uwtable sanitize_thread {
+entry:
+  %ptr = alloca i32, align 4
+  %tmp = alloca i32*, align 8
+  ; transitive escape
+  store i32* %ptr, i32** %tmp, align 8
+  %0 = load i32** %tmp, align 8
+  store i32* %0, i32** @sink, align 8
+  store i32 42, i32* %ptr, align 4
+  ret void
+}
+; CHECK-LABEL: define void @captured2
+; CHECK: __tsan_write
+; CHECK: __tsan_write
+; CHECK: ret void
+
+define void @notcaptured0() nounwind uwtable sanitize_thread {
+entry:
+  %ptr = alloca i32, align 4
+  store i32 42, i32* %ptr, align 4
+  ; escapes due to call
+  call void @escape(i32* %ptr)
+  ret void
+}
+; CHECK-LABEL: define void @notcaptured0
+; CHECK: __tsan_write
+; CHECK: ret void
+
+define void @notcaptured1() nounwind uwtable sanitize_thread {
+entry:
+  %ptr = alloca i32, align 4
+  store i32 42, i32* %ptr, align 4
+  ; escapes due to store into global
+  store i32* %ptr, i32** @sink, align 8
+  ret void
+}
+; CHECK-LABEL: define void @notcaptured1
+; CHECK: __tsan_write
+; CHECK: __tsan_write
+; CHECK: ret void
+
+define void @notcaptured2() nounwind uwtable sanitize_thread {
+entry:
+  %ptr = alloca i32, align 4
+  %tmp = alloca i32*, align 8
+  store i32 42, i32* %ptr, align 4
+  ; transitive escape
+  store i32* %ptr, i32** %tmp, align 8
+  %0 = load i32** %tmp, align 8
+  store i32* %0, i32** @sink, align 8
+  ret void
+}
+; CHECK-LABEL: define void @notcaptured2
+; CHECK: __tsan_write
+; CHECK: __tsan_write
+; CHECK: ret void
+
+