[llvm] r256655 - [ValueTracking] fix bug computing isKnownToBeAPowerOfTwo() with arithmetic shift right (PR25900)
Sanjay Patel via llvm-commits
llvm-commits at lists.llvm.org
Wed Dec 30 14:40:52 PST 2015
Author: spatel
Date: Wed Dec 30 16:40:52 2015
New Revision: 256655
URL: http://llvm.org/viewvc/llvm-project?rev=256655&view=rev
Log:
[ValueTracking] fix bug computing isKnownToBeAPowerOfTwo() with arithmetic shift right (PR25900)
This is a fix for:
https://llvm.org/bugs/show_bug.cgi?id=25900
If we think that an arithmetic right shift of a power of two is always a power of two,
an sdiv gets wrongly converted to udiv.
Differential Revision: http://reviews.llvm.org/D15827
Added:
llvm/trunk/test/Analysis/ValueTracking/known-power-of-two.ll
Modified:
llvm/trunk/lib/Analysis/ValueTracking.cpp
Modified: llvm/trunk/lib/Analysis/ValueTracking.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Analysis/ValueTracking.cpp?rev=256655&r1=256654&r2=256655&view=diff
==============================================================================
--- llvm/trunk/lib/Analysis/ValueTracking.cpp (original)
+++ llvm/trunk/lib/Analysis/ValueTracking.cpp Wed Dec 30 16:40:52 2015
@@ -1743,9 +1743,10 @@ bool isKnownToBeAPowerOfTwo(Value *V, bo
return false;
Value *X = nullptr, *Y = nullptr;
- // A shift of a power of two is a power of two or zero.
+ // A shift left or a logical shift right of a power of two is a power of two
+ // or zero.
if (OrZero && (match(V, m_Shl(m_Value(X), m_Value())) ||
- match(V, m_Shr(m_Value(X), m_Value()))))
+ match(V, m_LShr(m_Value(X), m_Value()))))
return isKnownToBeAPowerOfTwo(X, /*OrZero*/ true, Depth, Q, DL);
if (ZExtInst *ZI = dyn_cast<ZExtInst>(V))
Added: llvm/trunk/test/Analysis/ValueTracking/known-power-of-two.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Analysis/ValueTracking/known-power-of-two.ll?rev=256655&view=auto
==============================================================================
--- llvm/trunk/test/Analysis/ValueTracking/known-power-of-two.ll (added)
+++ llvm/trunk/test/Analysis/ValueTracking/known-power-of-two.ll Wed Dec 30 16:40:52 2015
@@ -0,0 +1,20 @@
+; RUN: opt -S -instcombine < %s | FileCheck %s
+
+; https://llvm.org/bugs/show_bug.cgi?id=25900
+; An arithmetic shift right of a power of two is not a power
+; of two if the original value is the sign bit. Therefore,
+; we can't transform the sdiv into a udiv.
+
+define i32 @pr25900(i32 %d) {
+ %and = and i32 %d, -2147483648
+; The next 3 lines prevent another fold from masking the bug.
+ %ext = zext i32 %and to i64
+ %or = or i64 %ext, 4294967296
+ %trunc = trunc i64 %or to i32
+ %ashr = ashr exact i32 %trunc, 31
+ %div = sdiv i32 4, %ashr
+ ret i32 %div
+
+; CHECK: sdiv
+}
+
More information about the llvm-commits
mailing list