shylie/llvm-project
1
0
Fork 0
Code Issues Pull Requests Actions 6 Packages Projects Releases Wiki Activity

Temporarily promote values to i64 again

Browse Source 
Operands of binary operations that might overflow will be temporarily
  promoted to i64 again, though that is not a sound solution for the problem.

llvm-svn: 271538
This commit is contained in:
Johannes Doerfert 2016-06-02 17:09:22 +00:00
parent 5c3ff137f1
commit 6393ef135c
3 changed files with 55 additions and 46 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
polly/lib/CodeGen/IslExprBuilder.cpp
View File

@ -74,7 +74,10 @@ Value *IslExprBuilder::getOverflowState() const {
Value *IslExprBuilder::createBinOp(BinaryOperator::BinaryOps Opc, Value *LHS,
Value *RHS, const Twine &Name) {
unifyTypes(LHS, RHS);
// @TODO Temporarily promote types of potentially overflowing binary
// operations to at least i64.
Value *I64C = Builder.getInt64(0);
unifyTypes(LHS, RHS, I64C);
// Handle the plain operation (without overflow tracking) first.
if (!OverflowState) {

72
polly/test/Isl/CodeGen/invariant_load_base_pointer_conditional_2.ll
View File

@ -14,32 +14,34 @@
; CHECK-NEXT: Execution Context: [N, p, q] -> { : N > 0 }
; CHECK-NEXT: }
;
; IR: polly.preload.merge:
; IR: polly.preload.merge:
; IR-NEXT: %polly.preload.tmp1.merge = phi i32* [ %polly.access.I.load, %polly.preload.exec ], [ null, %polly.preload.cond ]
; IR-NEXT: store i32* %polly.preload.tmp1.merge, i32** %tmp1.preload.s2a
; IR-NEXT: %10 = sext i32 %N to i64
; IR-NEXT: %11 = icmp sge i64 %10, 1
; IR-NEXT: %12 = call { i32, i1 } @llvm.sadd.with.overflow.i32(i32 %p, i32 %q)
; IR-NEXT: %.obit4 = extractvalue { i32, i1 } %12, 1
; IR-NEXT: %12 = sext i32 %N to i64
; IR-NEXT: %13 = icmp sge i64 %12, 1
; IR-NEXT: %14 = sext i32 %p to i64
; IR-NEXT: %15 = sext i32 %q to i64
; IR-NEXT: %16 = call { i64, i1 } @llvm.sadd.with.overflow.i64(i64 %14, i64 %15)
; IR-NEXT: %.obit4 = extractvalue { i64, i1 } %16, 1
; IR-NEXT: %polly.overflow.state5 = or i1 false, %.obit4
; IR-NEXT: %.res6 = extractvalue { i32, i1 } %12, 0
; IR-NEXT: %13 = sext i32 %.res6 to i64
; IR-NEXT: %14 = icmp sle i64 %13, 2147483647
; IR-NEXT: %15 = and i1 %11, %14
; IR-NEXT: %16 = call { i32, i1 } @llvm.sadd.with.overflow.i32(i32 %p, i32 %q)
; IR-NEXT: %.obit7 = extractvalue { i32, i1 } %16, 1
; IR-NEXT: %.res6 = extractvalue { i64, i1 } %16, 0
; IR-NEXT: %17 = icmp sle i64 %.res6, 2147483647
; IR-NEXT: %18 = and i1 %13, %17
; IR-NEXT: %19 = sext i32 %p to i64
; IR-NEXT: %20 = sext i32 %q to i64
; IR-NEXT: %21 = call { i64, i1 } @llvm.sadd.with.overflow.i64(i64 %19, i64 %20)
; IR-NEXT: %.obit7 = extractvalue { i64, i1 } %21, 1
; IR-NEXT: %polly.overflow.state8 = or i1 %polly.overflow.state5, %.obit7
; IR-NEXT: %.res9 = extractvalue { i32, i1 } %16, 0
; IR-NEXT: %17 = sext i32 %.res9 to i64
; IR-NEXT: %18 = icmp sge i64 %17, -2147483648
; IR-NEXT: %19 = and i1 %15, %18
; IR-NEXT: %.res9 = extractvalue { i64, i1 } %21, 0
; IR-NEXT: %22 = icmp sge i64 %.res9, -2147483648
; IR-NEXT: %23 = and i1 %18, %22
; IR-NEXT: %polly.preload.cond.overflown10 = xor i1 %polly.overflow.state8, true
; IR-NEXT: %polly.preload.cond.result11 = and i1 %19, %polly.preload.cond.overflown10
; IR-NEXT: %polly.preload.cond.result11 = and i1 %23, %polly.preload.cond.overflown10
; IR-NEXT: br label %polly.preload.cond12
;
; IR: polly.preload.cond12:
; IR-NEXT: br i1 %polly.preload.cond.result11, label %polly.preload.exec14, label %polly.preload.merge13
; IR: polly.preload.cond12:
; IR-NEXT: br i1 %polly.preload.cond.result11
;
; IR: polly.preload.exec14:
; IR-NEXT: %polly.access.polly.preload.tmp1.merge = getelementptr i32, i32* %polly.preload.tmp1.merge, i64 0
; IR-NEXT: %polly.access.polly.preload.tmp1.merge.load = load i32, i32* %polly.access.polly.preload.tmp1.merge, align 4
@ -47,22 +49,24 @@
; IRA: polly.preload.merge:
; IRA-NEXT: %polly.preload.tmp1.merge = phi i32* [ %polly.access.I.load, %polly.preload.exec ], [ null, %polly.preload.cond ]
; IRA-NEXT: store i32* %polly.preload.tmp1.merge, i32** %tmp1.preload.s2a
; IRA-NEXT: %10 = sext i32 %N to i64
; IRA-NEXT: %11 = icmp sge i64 %10, 1
; IRA-NEXT: %12 = call { i32, i1 } @llvm.sadd.with.overflow.i32(i32 %p, i32 %q)
; IRA-NEXT: %.obit5 = extractvalue { i32, i1 } %12, 1
; IRA-NEXT: %.res6 = extractvalue { i32, i1 } %12, 0
; IRA-NEXT: %13 = sext i32 %.res6 to i64
; IRA-NEXT: %14 = icmp sle i64 %13, 2147483647
; IRA-NEXT: %15 = and i1 %11, %14
; IRA-NEXT: %16 = call { i32, i1 } @llvm.sadd.with.overflow.i32(i32 %p, i32 %q)
; IRA-NEXT: %.obit7 = extractvalue { i32, i1 } %16, 1
; IRA-NEXT: %.res8 = extractvalue { i32, i1 } %16, 0
; IRA-NEXT: %17 = sext i32 %.res8 to i64
; IRA-NEXT: %18 = icmp sge i64 %17, -2147483648
; IRA-NEXT: %19 = and i1 %15, %18
; IRA-NEXT: %12 = sext i32 %N to i64
; IRA-NEXT: %13 = icmp sge i64 %12, 1
; IRA-NEXT: %14 = sext i32 %p to i64
; IRA-NEXT: %15 = sext i32 %q to i64
; IRA-NEXT: %16 = call { i64, i1 } @llvm.sadd.with.overflow.i64(i64 %14, i64 %15)
; IRA-NEXT: %.obit5 = extractvalue { i64, i1 } %16, 1
; IRA-NEXT: %.res6 = extractvalue { i64, i1 } %16, 0
; IRA-NEXT: %17 = icmp sle i64 %.res6, 2147483647
; IRA-NEXT: %18 = and i1 %13, %17
; IRA-NEXT: %19 = sext i32 %p to i64
; IRA-NEXT: %20 = sext i32 %q to i64
; IRA-NEXT: %21 = call { i64, i1 } @llvm.sadd.with.overflow.i64(i64 %19, i64 %20)
; IRA-NEXT: %.obit7 = extractvalue { i64, i1 } %21, 1
; IRA-NEXT: %.res8 = extractvalue { i64, i1 } %21, 0
; IRA-NEXT: %22 = icmp sge i64 %.res8, -2147483648
; IRA-NEXT: %23 = and i1 %18, %22
; IRA-NEXT: %polly.preload.cond.overflown9 = xor i1 %.obit7, true
; IRA-NEXT: %polly.preload.cond.result10 = and i1 %19, %polly.preload.cond.overflown9
; IRA-NEXT: %polly.preload.cond.result10 = and i1 %23, %polly.preload.cond.overflown9
; IRA-NEXT: br label %polly.preload.cond11
;
; IRA: polly.preload.cond11:

24
polly/test/Isl/CodeGen/no-overflow-tracking.ll
View File

@ -16,20 +16,22 @@
; IR: polly.preload.merge:
; IR-NEXT: %polly.preload.tmp1.merge = phi i32* [ %polly.access.I.load, %polly.preload.exec ], [ null, %polly.preload.cond ]
; IR-NEXT: store i32* %polly.preload.tmp1.merge, i32** %tmp1.preload.s2a
; IR-NEXT: %10 = sext i32 %N to i64
; IR-NEXT: %11 = icmp sge i64 %10, 1
; IR-NEXT: %12 = add nsw i32 %p, %q
; IR-NEXT: %13 = sext i32 %12 to i64
; IR-NEXT: %14 = icmp sle i64 %13, 2147483647
; IR-NEXT: %15 = and i1 %11, %14
; IR-NEXT: %16 = add nsw i32 %p, %q
; IR-NEXT: %17 = sext i32 %16 to i64
; IR-NEXT: %18 = icmp sge i64 %17, -2147483648
; IR-NEXT: %19 = and i1 %15, %18
; IR-NEXT: %12 = sext i32 %N to i64
; IR-NEXT: %13 = icmp sge i64 %12, 1
; IR-NEXT: %14 = sext i32 %p to i64
; IR-NEXT: %15 = sext i32 %q to i64
; IR-NEXT: %16 = add nsw i64 %14, %15
; IR-NEXT: %17 = icmp sle i64 %16, 2147483647
; IR-NEXT: %18 = and i1 %13, %17
; IR-NEXT: %19 = sext i32 %p to i64
; IR-NEXT: %20 = sext i32 %q to i64
; IR-NEXT: %21 = add nsw i64 %19, %20
; IR-NEXT: %22 = icmp sge i64 %21, -2147483648
; IR-NEXT: %23 = and i1 %18, %22
; IR-NEXT: br label %polly.preload.cond1
;
; IR: polly.preload.cond1:
; IR-NEXT: br i1 %19
; IR-NEXT: br i1 %23
;
; IR: polly.preload.exec3:
; IR-NEXT: %polly.access.polly.preload.tmp1.merge = getelementptr i32, i32* %polly.preload.tmp1.merge, i64 0