[llvm] da41932 - [NFC][IndVarSimplify] Test: replacing IV user with unsafe loop-invariant (PR45360)

Thu Jun 18 09:36:00 PDT 2020

Author: Roman Lebedev
Date: 2020-06-18T19:35:35+03:00
New Revision: da419320ef22c68b8178e1e349398877448b163f

URL: https://github.com/llvm/llvm-project/commit/da419320ef22c68b8178e1e349398877448b163f
DIFF: https://github.com/llvm/llvm-project/commit/da419320ef22c68b8178e1e349398877448b163f.diff

LOG: [NFC][IndVarSimplify] Test: replacing IV user with unsafe loop-invariant (PR45360)

https://bugs.llvm.org/show_bug.cgi?id=45360

This is reduced from the (runnable) test provided in the bug report.
The remainder operation is originally guarded, it never divides by zero.
Indvars should not make it execute unconditionally.

This is not a great test, running whole -O2 is fragile,
but i really don't understand why running -indvars on the IR before
that tranform happens doesn't work.

Added: 
    llvm/test/Transforms/IndVarSimplify/X86/pr45360.ll

Modified: 
    

Removed: 
    


################################################################################
diff  --git a/llvm/test/Transforms/IndVarSimplify/X86/pr45360.ll b/llvm/test/Transforms/IndVarSimplify/X86/pr45360.ll
new file mode 100644
index 000000000000..c4562d811d09

--- /dev/null
+++ b/llvm/test/Transforms/IndVarSimplify/X86/pr45360.ll
@@ -0,0 +1,136 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt -S -O2 -scev-cheap-expansion-budget=1024 %s | FileCheck %s
+
+; See https://bugs.llvm.org/show_bug.cgi?id=45360
+; This is reduced from that (runnable) test.
+; The remainder operation is originally guarded, it never divides by zero.
+; Indvars should not make it execute unconditionally.
+
+target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-pc-linux-gnu"
+
+ at f = dso_local global i32 0, align 4
+ at a = dso_local global i32 0, align 4
+ at d = dso_local global i32 0, align 4
+ at c = dso_local global i32 0, align 4
+ at b = dso_local global i32 0, align 4
+ at e = dso_local global i32 0, align 4
+
+define dso_local i32 @main() {
+; CHECK-LABEL: @main(
+; CHECK-NEXT:  bb:
+; CHECK-NEXT:    [[I6:%.*]] = load i32, i32* @a, align 4
+; CHECK-NEXT:    [[I24:%.*]] = load i32, i32* @b, align 4
+; CHECK-NEXT:    [[D_PROMOTED8:%.*]] = load i32, i32* @d, align 4
+; CHECK-NEXT:    br label [[BB19:%.*]]
+; CHECK:       bb19:
+; CHECK-NEXT:    [[I8_LCSSA9:%.*]] = phi i32 [ [[D_PROMOTED8]], [[BB:%.*]] ], [ [[I8:%.*]], [[BB27:%.*]] ]
+; CHECK-NEXT:    [[I8]] = and i32 [[I8_LCSSA9]], [[I6]]
+; CHECK-NEXT:    [[TMP0:%.*]] = urem i32 [[I24]], [[I8]]
+; CHECK-NEXT:    [[I21:%.*]] = icmp eq i32 [[I8]], 0
+; CHECK-NEXT:    br i1 [[I21]], label [[BB27_THREAD:%.*]], label [[BB27]]
+; CHECK:       bb27.thread:
+; CHECK-NEXT:    store i32 -1, i32* @f, align 4
+; CHECK-NEXT:    store i32 0, i32* @d, align 4
+; CHECK-NEXT:    store i32 0, i32* @c, align 4
+; CHECK-NEXT:    br label [[BB32:%.*]]
+; CHECK:       bb27:
+; CHECK-NEXT:    store i32 [[TMP0]], i32* @e, align 4
+; CHECK-NEXT:    [[I30:%.*]] = icmp eq i32 [[TMP0]], 0
+; CHECK-NEXT:    br i1 [[I30]], label [[BB32_LOOPEXIT:%.*]], label [[BB19]]
+; CHECK:       bb32.loopexit:
+; CHECK-NEXT:    store i32 -1, i32* @f, align 4
+; CHECK-NEXT:    store i32 [[I8]], i32* @d, align 4
+; CHECK-NEXT:    br label [[BB32]]
+; CHECK:       bb32:
+; CHECK-NEXT:    [[C_SINK:%.*]] = phi i32* [ @c, [[BB32_LOOPEXIT]] ], [ @e, [[BB27_THREAD]] ]
+; CHECK-NEXT:    store i32 0, i32* [[C_SINK]], align 4
+; CHECK-NEXT:    ret i32 0
+;
+bb:
+  %i = alloca i32, align 4
+  store i32 0, i32* %i, align 4
+  br label %bb1
+
+bb1:
+  store i32 0, i32* @f, align 4
+  br label %bb2
+
+bb2:
+  %i3 = load i32, i32* @f, align 4
+  %i4 = icmp sge i32 %i3, 0
+  br i1 %i4, label %bb5, label %bb12
+
+bb5:
+  %i6 = load i32, i32* @a, align 4
+  %i7 = load i32, i32* @d, align 4
+  %i8 = and i32 %i7, %i6
+  store i32 %i8, i32* @d, align 4
+  br label %bb9
+
+bb9:
+  %i10 = load i32, i32* @f, align 4
+  %i11 = add nsw i32 %i10, -1
+  store i32 %i11, i32* @f, align 4
+  br label %bb2
+
+bb12:
+  store i32 0, i32* @c, align 4
+  br label %bb13
+
+bb13:
+  %i14 = load i32, i32* @c, align 4
+  %i15 = icmp sle i32 %i14, 0
+  br i1 %i15, label %bb16, label %bb39
+
+bb16:
+  %i17 = load i32, i32* @f, align 4
+  %i18 = icmp ne i32 %i17, 0
+  br i1 %i18, label %bb19, label %bb34
+
+bb19:
+  %i20 = load i32, i32* @d, align 4
+  %i21 = icmp eq i32 %i20, 0
+  br i1 %i21, label %bb22, label %bb23
+
+bb22:
+  br label %bb27
+
+bb23:
+  %i24 = load i32, i32* @b, align 4
+  %i25 = load i32, i32* @d, align 4
+  %i26 = urem i32 %i24, %i25
+  br label %bb27
+
+bb27:
+  %i28 = phi i32 [ 0, %bb22 ], [ %i26, %bb23 ]
+  store i32 %i28, i32* @e, align 4
+  %i29 = load i32, i32* @e, align 4
+  %i30 = icmp ne i32 %i29, 0
+  br i1 %i30, label %bb31, label %bb32
+
+bb31:
+  br label %bb33
+
+bb32:
+  ret i32 0
+
+bb33:
+  br label %bb35
+
+bb34:
+  store i32 0, i32* @d, align 4
+  br label %bb35
+
+bb35:
+  br label %bb36
+
+bb36:
+  %i37 = load i32, i32* @c, align 4
+  %i38 = add nsw i32 %i37, 1
+  store i32 %i38, i32* @c, align 4
+  br label %bb13
+
+bb39:
+  br label %bb1
+}


        
