maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/Transforms/IndVarSimplify/lftr.ll

633 lines
25 KiB
LLVM
Raw Normal View History

[Tests] Start consolidating lftr tests into a single file llvm-svn: 360934
2019-05-17 04:33:41 +08:00
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt < %s -indvars -dce -S | FileCheck %s
; Provide legal integer types.
target datalayout = "n8:16:32:64"
@A = external global i32
[Tests] Expand basic lftr coverage Newly written tests to cover the simple cases. We don't appear to have broad coverage of this transform anywhere. llvm-svn: 360957
2019-05-17 07:41:28 +08:00
;; Convert a pre-increment check on the latch into a post increment check
define i32 @pre_to_post_add() {
; CHECK-LABEL: @pre_to_post_add(
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[LOOP:%.*]]
; CHECK: loop:
; CHECK-NEXT: [[I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[I_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[I_NEXT]] = add nuw nsw i32 [[I]], 1
; CHECK-NEXT: store i32 [[I]], i32* @A
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[I_NEXT]], 1001
; CHECK-NEXT: br i1 [[EXITCOND]], label [[LOOP]], label [[LOOPEXIT:%.*]]
; CHECK: loopexit:
; CHECK-NEXT: ret i32 1000
;
entry:
br label %loop
loop:
%i = phi i32 [ 0, %entry ], [ %i.next, %loop ]
%i.next = add i32 %i, 1
store i32 %i, i32* @A
%c = icmp slt i32 %i, 1000
br i1 %c, label %loop, label %loopexit
loopexit:
ret i32 %i
}
; TODO: we should be able to convert the subtract into a post-decrement check
define i32 @pre_to_post_sub() {
; CHECK-LABEL: @pre_to_post_sub(
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[LOOP:%.*]]
; CHECK: loop:
; CHECK-NEXT: [[I:%.*]] = phi i32 [ 1000, [[ENTRY:%.*]] ], [ [[I_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[I_NEXT]] = sub nsw i32 [[I]], 1
; CHECK-NEXT: store i32 [[I]], i32* @A
; CHECK-NEXT: [[C:%.*]] = icmp ugt i32 [[I]], 0
; CHECK-NEXT: br i1 [[C]], label [[LOOP]], label [[LOOPEXIT:%.*]]
; CHECK: loopexit:
; CHECK-NEXT: ret i32 0
;
entry:
br label %loop
loop:
%i = phi i32 [ 1000, %entry ], [ %i.next, %loop ]
%i.next = sub i32 %i, 1
store i32 %i, i32* @A
%c = icmp sgt i32 %i, 0
br i1 %c, label %loop, label %loopexit
loopexit:
ret i32 %i
}
; LFTR should eliminate the need for the computation of i*i completely. It
; is only used to compute the exit value.
define i32 @quadratic_slt() {
; CHECK-LABEL: @quadratic_slt(
[Tests] Start consolidating lftr tests into a single file llvm-svn: 360934
2019-05-17 04:33:41 +08:00
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[LOOP:%.*]]
; CHECK: loop:
; CHECK-NEXT: [[I:%.*]] = phi i32 [ 7, [[ENTRY:%.*]] ], [ [[I_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[I_NEXT]] = add nuw nsw i32 [[I]], 1
; CHECK-NEXT: store i32 [[I]], i32* @A
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[I_NEXT]], 33
; CHECK-NEXT: br i1 [[EXITCOND]], label [[LOOP]], label [[LOOPEXIT:%.*]]
; CHECK: loopexit:
; CHECK-NEXT: ret i32 32
;
entry:
br label %loop
loop:
%i = phi i32 [ 7, %entry ], [ %i.next, %loop ]
%i.next = add i32 %i, 1
store i32 %i, i32* @A
%i2 = mul i32 %i, %i
%c = icmp slt i32 %i2, 1000
br i1 %c, label %loop, label %loopexit
loopexit:
ret i32 %i
}
[Tests] Expand basic lftr coverage Newly written tests to cover the simple cases. We don't appear to have broad coverage of this transform anywhere. llvm-svn: 360957
2019-05-17 07:41:28 +08:00
; Same as previous but with sle test
define i32 @quadratic_sle() {
; CHECK-LABEL: @quadratic_sle(
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[LOOP:%.*]]
; CHECK: loop:
; CHECK-NEXT: [[I:%.*]] = phi i32 [ 7, [[ENTRY:%.*]] ], [ [[I_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[I_NEXT]] = add nuw nsw i32 [[I]], 1
; CHECK-NEXT: store i32 [[I]], i32* @A
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[I_NEXT]], 33
; CHECK-NEXT: br i1 [[EXITCOND]], label [[LOOP]], label [[LOOPEXIT:%.*]]
; CHECK: loopexit:
; CHECK-NEXT: ret i32 32
;
entry:
br label %loop
loop:
%i = phi i32 [ 7, %entry ], [ %i.next, %loop ]
%i.next = add i32 %i, 1
store i32 %i, i32* @A
%i2 = mul i32 %i, %i
%c = icmp sle i32 %i2, 1000
br i1 %c, label %loop, label %loopexit
loopexit:
ret i32 %i
}
; Same as previous but with ule test
define i32 @quadratic_ule() {
; CHECK-LABEL: @quadratic_ule(
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[LOOP:%.*]]
; CHECK: loop:
; CHECK-NEXT: [[I:%.*]] = phi i32 [ 7, [[ENTRY:%.*]] ], [ [[I_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[I_NEXT]] = add nuw nsw i32 [[I]], 1
; CHECK-NEXT: store i32 [[I]], i32* @A
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[I_NEXT]], 33
; CHECK-NEXT: br i1 [[EXITCOND]], label [[LOOP]], label [[LOOPEXIT:%.*]]
; CHECK: loopexit:
; CHECK-NEXT: ret i32 32
;
entry:
br label %loop
loop:
%i = phi i32 [ 7, %entry ], [ %i.next, %loop ]
%i.next = add i32 %i, 1
store i32 %i, i32* @A
%i2 = mul i32 %i, %i
%c = icmp ule i32 %i2, 1000
br i1 %c, label %loop, label %loopexit
loopexit:
ret i32 %i
}
[Tests] Start consolidating lftr tests into a single file llvm-svn: 360934
2019-05-17 04:33:41 +08:00
@data = common global [240 x i8] zeroinitializer, align 16
define void @test_zext(i8* %a) #0 {
; CHECK-LABEL: @test_zext(
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[LOOP:%.*]]
; CHECK: loop:
; CHECK-NEXT: [[P_0:%.*]] = phi i8* [ getelementptr inbounds ([240 x i8], [240 x i8]* @data, i64 0, i64 0), [[ENTRY:%.*]] ], [ [[TMP3:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[DOT0:%.*]] = phi i8* [ [[A:%.*]], [[ENTRY]] ], [ [[TMP:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[TMP]] = getelementptr inbounds i8, i8* [[DOT0]], i64 1
; CHECK-NEXT: [[TMP2:%.*]] = load i8, i8* [[DOT0]], align 1
; CHECK-NEXT: [[TMP3]] = getelementptr inbounds i8, i8* [[P_0]], i64 1
; CHECK-NEXT: store i8 [[TMP2]], i8* [[P_0]], align 1
Fix a bug w/inbounds invalidation in LFTR (recommit) Recommit r363289 with a bug fix for crash identified in pr42279. Issue was that a loop exit test does not have to be an icmp, leading to a null dereference crash when new logic was exercised for that case. Test case previously committed in r363601. Original commit comment follows: This contains fixes for two cases where we might invalidate inbounds and leave it stale in the IR (a miscompile). Case 1 is when switching to an IV with no dynamically live uses, and case 2 is when doing pre-to-post conversion on the same pointer type IV. The basic scheme used is to prove that using the given IV (pre or post increment forms) would have to already trigger UB on the path to the test we're modifying. As such, our potential UB triggering use does not change the semantics of the original program. As was pointed out in the review thread by Nikita, this is defending against a separate issue from the hasConcreteDef case. This is about poison, that's about undef. Unfortunately, the two are different, see Nikita's comment for a fuller explanation, he explains it well. (Note: I'm going to address Nikita's last style comment in a separate commit just to minimize chance of subtle bugs being introduced due to typos.) Differential Revision: https://reviews.llvm.org/D62939 llvm-svn: 363613
2019-06-18 04:32:22 +08:00
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i8* [[P_0]], getelementptr inbounds ([240 x i8], [240 x i8]* @data, i64 0, i64 239)
[Tests] Start consolidating lftr tests into a single file llvm-svn: 360934
2019-05-17 04:33:41 +08:00
; CHECK-NEXT: br i1 [[EXITCOND]], label [[LOOP]], label [[EXIT:%.*]]
; CHECK: exit:
; CHECK-NEXT: ret void
;
entry:
br label %loop
loop:
%i.0 = phi i8 [ 0, %entry ], [ %tmp4, %loop ]
%p.0 = phi i8* [ getelementptr inbounds ([240 x i8], [240 x i8]* @data, i64 0, i64 0), %entry ], [ %tmp3, %loop ]
%.0 = phi i8* [ %a, %entry ], [ %tmp, %loop ]
%tmp = getelementptr inbounds i8, i8* %.0, i64 1
%tmp2 = load i8, i8* %.0, align 1
%tmp3 = getelementptr inbounds i8, i8* %p.0, i64 1
store i8 %tmp2, i8* %p.0, align 1
%tmp4 = add i8 %i.0, 1
%tmp5 = icmp ult i8 %tmp4, -16
br i1 %tmp5, label %loop, label %exit
exit:
ret void
}
; It is okay to do LFTR on this loop even though the trip count is a
; division because in this case the division can be optimized to a
; shift.
define void @test_udiv_as_shift(i8* %a, i8 %n) nounwind uwtable ssp {
; CHECK-LABEL: @test_udiv_as_shift(
; CHECK-NEXT: entry:
; CHECK-NEXT: [[E:%.*]] = icmp sgt i8 [[N:%.*]], 3
; CHECK-NEXT: br i1 [[E]], label [[LOOP_PREHEADER:%.*]], label [[EXIT:%.*]]
; CHECK: loop.preheader:
; CHECK-NEXT: [[TMP0:%.*]] = add i8 [[N]], 3
; CHECK-NEXT: [[TMP1:%.*]] = lshr i8 [[TMP0]], 2
[SCEV] Pass NoWrapFlags when expanding an AddExpr InsertBinop now accepts NoWrapFlags, so pass them through when expanding a simple add expression. This is the first re-commit of the functional changes from rL362687, which was previously reverted. Differential Revision: https://reviews.llvm.org/D61934 llvm-svn: 363364
2019-06-14 17:19:41 +08:00
; CHECK-NEXT: [[TMP2:%.*]] = add nuw nsw i8 [[TMP1]], 1
[Tests] Start consolidating lftr tests into a single file llvm-svn: 360934
2019-05-17 04:33:41 +08:00
; CHECK-NEXT: br label [[LOOP:%.*]]
; CHECK: loop:
; CHECK-NEXT: [[I1:%.*]] = phi i8 [ [[I1_INC:%.*]], [[LOOP]] ], [ 0, [[LOOP_PREHEADER]] ]
; CHECK-NEXT: [[I1_INC]] = add nuw nsw i8 [[I1]], 1
; CHECK-NEXT: store volatile i8 0, i8* [[A:%.*]]
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i8 [[I1_INC]], [[TMP2]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[LOOP]], label [[EXIT_LOOPEXIT:%.*]]
; CHECK: exit.loopexit:
; CHECK-NEXT: br label [[EXIT]]
; CHECK: exit:
; CHECK-NEXT: ret void
;
entry:
%e = icmp sgt i8 %n, 3
br i1 %e, label %loop, label %exit
loop:
%i = phi i8 [ 0, %entry ], [ %i.inc, %loop ]
%i1 = phi i8 [ 0, %entry ], [ %i1.inc, %loop ]
%i.inc = add nsw i8 %i, 4
%i1.inc = add i8 %i1, 1
store volatile i8 0, i8* %a
%c = icmp slt i8 %i, %n
br i1 %c, label %loop, label %exit
exit:
ret void
}
[Tests] More consolidation of lftr tests llvm-svn: 360936
2019-05-17 04:42:00 +08:00
; Don't RAUW the loop's original comparison instruction if it has other uses
; which aren't dominated by the new comparison instruction (which we insert
; at the branch user).
define void @use_before_branch() {
; CHECK-LABEL: @use_before_branch(
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[LOOPENTRY_0:%.*]]
; CHECK: loopentry.0:
; CHECK-NEXT: [[MB_Y_0:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[TMP_152:%.*]], [[LOOPENTRY_1:%.*]] ]
; CHECK-NEXT: [[TMP_14:%.*]] = icmp ule i32 [[MB_Y_0]], 3
; CHECK-NEXT: br i1 [[TMP_14]], label [[LOOPENTRY_1]], label [[LOOPEXIT_0:%.*]]
; CHECK: loopentry.1:
; CHECK-NEXT: [[TMP_152]] = add nuw nsw i32 [[MB_Y_0]], 2
; CHECK-NEXT: br label [[LOOPENTRY_0]]
; CHECK: loopexit.0:
; CHECK-NEXT: unreachable
;
entry:
br label %loopentry.0
loopentry.0:
%mb_y.0 = phi i32 [ 0, %entry ], [ %tmp.152, %loopentry.1 ]
%tmp.14 = icmp sle i32 %mb_y.0, 3
%tmp.15 = zext i1 %tmp.14 to i32
br i1 %tmp.14, label %loopentry.1, label %loopexit.0
loopentry.1:
%tmp.152 = add i32 %mb_y.0, 2
br label %loopentry.0
loopexit.0: ; preds = %loopentry.0
unreachable
}
@.str3 = private constant [6 x i8] c"%lld\0A\00", align 1
declare i32 @printf(i8* noalias nocapture, ...) nounwind
; PR13371: indvars pass incorrectly substitutes 'undef' values
;
; LFTR should not user %undef as the loop counter.
define i64 @no_undef_counter() nounwind {
; CHECK-LABEL: @no_undef_counter(
; CHECK-NEXT: func_start:
; CHECK-NEXT: br label [[BLOCK9:%.*]]
; CHECK: block9:
; CHECK-NEXT: [[UNDEF:%.*]] = phi i64 [ [[NEXT_UNDEF:%.*]], [[BLOCK9]] ], [ undef, [[FUNC_START:%.*]] ]
; CHECK-NEXT: [[ITER:%.*]] = phi i64 [ [[NEXT_ITER:%.*]], [[BLOCK9]] ], [ 1, [[FUNC_START]] ]
; CHECK-NEXT: [[NEXT_ITER]] = add nuw nsw i64 [[ITER]], 1
; CHECK-NEXT: [[TMP0:%.*]] = tail call i32 (i8*, ...) @printf(i8* noalias nocapture getelementptr inbounds ([6 x i8], [6 x i8]* @.str3, i64 0, i64 0), i64 [[NEXT_ITER]], i64 [[UNDEF]])
; CHECK-NEXT: [[NEXT_UNDEF]] = add nsw i64 [[UNDEF]], 1
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i64 [[NEXT_ITER]], 100
; CHECK-NEXT: br i1 [[EXITCOND]], label [[BLOCK9]], label [[EXIT:%.*]]
; CHECK: exit:
; CHECK-NEXT: ret i64 0
;
func_start:
br label %block9
block9: ; preds = %block9,%func_start
%undef = phi i64 [ %next_undef, %block9 ], [ undef, %func_start ]
%iter = phi i64 [ %next_iter, %block9 ], [ 1, %func_start ]
%next_iter = add nsw i64 %iter, 1
%0 = tail call i32 (i8*, ...) @printf(i8* noalias nocapture getelementptr inbounds ([6 x i8], [6 x i8]* @.str3, i64 0, i64 0), i64 %next_iter, i64 %undef)
%next_undef = add nsw i64 %undef, 1
%_tmp_3 = icmp slt i64 %next_iter, 100
br i1 %_tmp_3, label %block9, label %exit
exit: ; preds = %block9
ret i64 0
}
[Tests] Consolidate more lftr tests These are all of the ones involving the same data layout string. Remainder take a bit more consideration, but at least everything can be auto-updated now. llvm-svn: 360961
2019-05-17 08:19:28 +08:00
define void @extend_const() #0 {
; CHECK-LABEL: @extend_const(
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i32 [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ], [ 0, [[ENTRY:%.*]] ]
; CHECK-NEXT: call void @bar(i32 [[INDVARS_IV]]) #2
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i32 [[INDVARS_IV]], 1
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[INDVARS_IV_NEXT]], 512
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_BODY]], label [[FOR_END:%.*]]
; CHECK: for.end:
; CHECK-NEXT: ret void
;
entry:
br label %for.body
for.body: ; preds = %entry, %for.body
%i.01 = phi i16 [ 0, %entry ], [ %inc, %for.body ]
%conv2 = sext i16 %i.01 to i32
call void @bar(i32 %conv2) #1
%inc = add i16 %i.01, 1
%cmp = icmp slt i16 %inc, 512
br i1 %cmp, label %for.body, label %for.end
for.end: ; preds = %for.body
ret void
}
; Check that post-incrementing the backedge taken count does not overflow.
define i32 @extend_const_postinc() #0 {
; CHECK-LABEL: @extend_const_postinc(
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[DO_BODY:%.*]]
; CHECK: do.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i32 [ [[INDVARS_IV_NEXT:%.*]], [[DO_BODY]] ], [ 0, [[ENTRY:%.*]] ]
; CHECK-NEXT: call void @bar(i32 [[INDVARS_IV]]) #2
; CHECK-NEXT: [[CMP:%.*]] = icmp eq i32 [[INDVARS_IV]], 255
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i32 [[INDVARS_IV]], 1
; CHECK-NEXT: br i1 [[CMP]], label [[DO_END:%.*]], label [[DO_BODY]]
; CHECK: do.end:
; CHECK-NEXT: ret i32 0
;
entry:
br label %do.body
do.body: ; preds = %do.body, %entry
%first.0 = phi i8 [ 0, %entry ], [ %inc, %do.body ]
%conv = zext i8 %first.0 to i32
call void @bar(i32 %conv) #1
%inc = add i8 %first.0, 1
%cmp = icmp eq i8 %first.0, -1
br i1 %cmp, label %do.end, label %do.body
do.end: ; preds = %do.body
ret i32 0
}
declare void @bar(i32)
attributes #0 = { nounwind uwtable }
attributes #1 = { nounwind }
; With the given initial value for IV, it is not legal to widen
; trip count to IV size
define void @wide_trip_count_test1(float* %autoc,
; CHECK-LABEL: @wide_trip_count_test1(
; CHECK-NEXT: entry:
; CHECK-NEXT: [[SUB:%.*]] = sub i32 [[DATA_LEN:%.*]], [[SAMPLE:%.*]]
; CHECK-NEXT: [[CMP4:%.*]] = icmp eq i32 [[DATA_LEN]], [[SAMPLE]]
; CHECK-NEXT: br i1 [[CMP4]], label [[FOR_END:%.*]], label [[FOR_BODY_PREHEADER:%.*]]
; CHECK: for.body.preheader:
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ], [ 68719476736, [[FOR_BODY_PREHEADER]] ]
; CHECK-NEXT: [[TEMP:%.*]] = trunc i64 [[INDVARS_IV]] to i32
; CHECK-NEXT: [[ADD:%.*]] = add i32 [[TEMP]], [[SAMPLE]]
; CHECK-NEXT: [[IDXPROM:%.*]] = zext i32 [[ADD]] to i64
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds float, float* [[DATA:%.*]], i64 [[IDXPROM]]
; CHECK-NEXT: [[TEMP1:%.*]] = load float, float* [[ARRAYIDX]], align 4
; CHECK-NEXT: [[MUL:%.*]] = fmul float [[TEMP1]], [[D:%.*]]
; CHECK-NEXT: [[ARRAYIDX2:%.*]] = getelementptr inbounds float, float* [[AUTOC:%.*]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[TEMP2:%.*]] = load float, float* [[ARRAYIDX2]], align 4
; CHECK-NEXT: [[ADD3:%.*]] = fadd float [[TEMP2]], [[MUL]]
; CHECK-NEXT: store float [[ADD3]], float* [[ARRAYIDX2]], align 4
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[LFTR_WIDEIV]], [[SUB]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT:%.*]]
; CHECK: for.end.loopexit:
; CHECK-NEXT: br label [[FOR_END]]
; CHECK: for.end:
; CHECK-NEXT: ret void
;
float* %data,
float %d, i32 %data_len, i32 %sample) nounwind {
entry:
%sub = sub i32 %data_len, %sample
%cmp4 = icmp eq i32 %data_len, %sample
br i1 %cmp4, label %for.end, label %for.body
for.body: ; preds = %entry, %for.body
%indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 68719476736, %entry ]
%temp = trunc i64 %indvars.iv to i32
%add = add i32 %temp, %sample
%idxprom = zext i32 %add to i64
%arrayidx = getelementptr inbounds float, float* %data, i64 %idxprom
%temp1 = load float, float* %arrayidx, align 4
%mul = fmul float %temp1, %d
%arrayidx2 = getelementptr inbounds float, float* %autoc, i64 %indvars.iv
%temp2 = load float, float* %arrayidx2, align 4
%add3 = fadd float %temp2, %mul
store float %add3, float* %arrayidx2, align 4
%indvars.iv.next = add i64 %indvars.iv, 1
%temp3 = trunc i64 %indvars.iv.next to i32
%cmp = icmp ult i32 %temp3, %sub
br i1 %cmp, label %for.body, label %for.end
for.end: ; preds = %for.body, %entry
ret void
}
; Trip count should be widened and LFTR should canonicalize the condition
define float @wide_trip_count_test2(float* %a,
; CHECK-LABEL: @wide_trip_count_test2(
; CHECK-NEXT: entry:
; CHECK-NEXT: [[CMP5:%.*]] = icmp ugt i32 [[M:%.*]], 500
; CHECK-NEXT: br i1 [[CMP5]], label [[FOR_BODY_PREHEADER:%.*]], label [[FOR_END:%.*]]
; CHECK: for.body.preheader:
[LFTR] Hoist extend expressions outside of loops w/o waiting for LICM The motivation for this is two fold: 1) Make the output (and thus tests) a bit more readable to a human trying to understand the result of the transform 2) Reduce spurious diffs in a potential future change to restructure all of this logic to use SCEVExpander (which hoists by default) llvm-svn: 365066
2019-07-04 02:18:36 +08:00
; CHECK-NEXT: [[WIDE_TRIP_COUNT:%.*]] = zext i32 [[M]] to i64
[Tests] Consolidate more lftr tests These are all of the ones involving the same data layout string. Remainder take a bit more consideration, but at least everything can be auto-updated now. llvm-svn: 360961
2019-05-17 08:19:28 +08:00
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ], [ 500, [[FOR_BODY_PREHEADER]] ]
; CHECK-NEXT: [[SUM_07:%.*]] = phi float [ [[ADD:%.*]], [[FOR_BODY]] ], [ 0.000000e+00, [[FOR_BODY_PREHEADER]] ]
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds float, float* [[B:%.*]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[TEMP:%.*]] = load float, float* [[ARRAYIDX]], align 4
; CHECK-NEXT: [[ARRAYIDX2:%.*]] = getelementptr inbounds float, float* [[A:%.*]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[TEMP1:%.*]] = load float, float* [[ARRAYIDX2]], align 4
; CHECK-NEXT: [[MUL:%.*]] = fmul float [[TEMP]], [[TEMP1]]
; CHECK-NEXT: [[ADD]] = fadd float [[SUM_07]], [[MUL]]
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT]], [[WIDE_TRIP_COUNT]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT:%.*]]
; CHECK: for.end.loopexit:
; CHECK-NEXT: [[ADD_LCSSA:%.*]] = phi float [ [[ADD]], [[FOR_BODY]] ]
; CHECK-NEXT: br label [[FOR_END]]
; CHECK: for.end:
; CHECK-NEXT: [[SUM_0_LCSSA:%.*]] = phi float [ 0.000000e+00, [[ENTRY:%.*]] ], [ [[ADD_LCSSA]], [[FOR_END_LOOPEXIT]] ]
; CHECK-NEXT: ret float [[SUM_0_LCSSA]]
;
float* %b,
i32 zeroext %m) local_unnamed_addr #0 {
entry:
%cmp5 = icmp ugt i32 %m, 500
br i1 %cmp5, label %for.body.preheader, label %for.end
for.body.preheader: ; preds = %entry
br label %for.body
for.body: ; preds = %for.body.preheader, %for.body
%sum.07 = phi float [ %add, %for.body ], [ 0.000000e+00, %for.body.preheader ]
%i.06 = phi i32 [ %inc, %for.body ], [ 500, %for.body.preheader ]
%idxprom = zext i32 %i.06 to i64
%arrayidx = getelementptr inbounds float, float* %b, i64 %idxprom
%temp = load float, float* %arrayidx, align 4
%arrayidx2 = getelementptr inbounds float, float* %a, i64 %idxprom
%temp1 = load float, float* %arrayidx2, align 4
%mul = fmul float %temp, %temp1
%add = fadd float %sum.07, %mul
%inc = add i32 %i.06, 1
%cmp = icmp ult i32 %inc, %m
br i1 %cmp, label %for.body, label %for.end.loopexit
for.end.loopexit: ; preds = %for.body
br label %for.end
for.end: ; preds = %for.end.loopexit, %entry
%sum.0.lcssa = phi float [ 0.000000e+00, %entry ], [ %add, %for.end.loopexit ]
ret float %sum.0.lcssa
}
; Trip count should be widened and LFTR should canonicalize the condition
define float @wide_trip_count_test3(float* %b,
; CHECK-LABEL: @wide_trip_count_test3(
; CHECK-NEXT: entry:
; CHECK-NEXT: [[CMP5:%.*]] = icmp sgt i32 [[M:%.*]], -10
; CHECK-NEXT: br i1 [[CMP5]], label [[FOR_BODY_PREHEADER:%.*]], label [[FOR_END:%.*]]
; CHECK: for.body.preheader:
[LFTR] Hoist extend expressions outside of loops w/o waiting for LICM The motivation for this is two fold: 1) Make the output (and thus tests) a bit more readable to a human trying to understand the result of the transform 2) Reduce spurious diffs in a potential future change to restructure all of this logic to use SCEVExpander (which hoists by default) llvm-svn: 365066
2019-07-04 02:18:36 +08:00
; CHECK-NEXT: [[WIDE_TRIP_COUNT:%.*]] = sext i32 [[M]] to i64
[Tests] Consolidate more lftr tests These are all of the ones involving the same data layout string. Remainder take a bit more consideration, but at least everything can be auto-updated now. llvm-svn: 360961
2019-05-17 08:19:28 +08:00
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ], [ -10, [[FOR_BODY_PREHEADER]] ]
; CHECK-NEXT: [[SUM_07:%.*]] = phi float [ [[ADD1:%.*]], [[FOR_BODY]] ], [ 0.000000e+00, [[FOR_BODY_PREHEADER]] ]
; CHECK-NEXT: [[TMP0:%.*]] = add nsw i64 [[INDVARS_IV]], 20
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds float, float* [[B:%.*]], i64 [[TMP0]]
; CHECK-NEXT: [[TEMP:%.*]] = load float, float* [[ARRAYIDX]], align 4
; CHECK-NEXT: [[TMP1:%.*]] = trunc i64 [[INDVARS_IV]] to i32
; CHECK-NEXT: [[CONV:%.*]] = sitofp i32 [[TMP1]] to float
; CHECK-NEXT: [[MUL:%.*]] = fmul float [[CONV]], [[TEMP]]
; CHECK-NEXT: [[ADD1]] = fadd float [[SUM_07]], [[MUL]]
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nsw i64 [[INDVARS_IV]], 1
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT]], [[WIDE_TRIP_COUNT]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT:%.*]]
; CHECK: for.end.loopexit:
; CHECK-NEXT: [[ADD1_LCSSA:%.*]] = phi float [ [[ADD1]], [[FOR_BODY]] ]
; CHECK-NEXT: br label [[FOR_END]]
; CHECK: for.end:
; CHECK-NEXT: [[SUM_0_LCSSA:%.*]] = phi float [ 0.000000e+00, [[ENTRY:%.*]] ], [ [[ADD1_LCSSA]], [[FOR_END_LOOPEXIT]] ]
; CHECK-NEXT: ret float [[SUM_0_LCSSA]]
;
i32 signext %m) local_unnamed_addr #0 {
entry:
%cmp5 = icmp sgt i32 %m, -10
br i1 %cmp5, label %for.body.preheader, label %for.end
for.body.preheader: ; preds = %entry
br label %for.body
for.body: ; preds = %for.body.preheader, %for.body
%sum.07 = phi float [ %add1, %for.body ], [ 0.000000e+00, %for.body.preheader ]
%i.06 = phi i32 [ %inc, %for.body ], [ -10, %for.body.preheader ]
%add = add nsw i32 %i.06, 20
%idxprom = sext i32 %add to i64
%arrayidx = getelementptr inbounds float, float* %b, i64 %idxprom
%temp = load float, float* %arrayidx, align 4
%conv = sitofp i32 %i.06 to float
%mul = fmul float %conv, %temp
%add1 = fadd float %sum.07, %mul
%inc = add nsw i32 %i.06, 1
%cmp = icmp slt i32 %inc, %m
br i1 %cmp, label %for.body, label %for.end.loopexit
for.end.loopexit: ; preds = %for.body
br label %for.end
for.end: ; preds = %for.end.loopexit, %entry
%sum.0.lcssa = phi float [ 0.000000e+00, %entry ], [ %add1, %for.end.loopexit ]
ret float %sum.0.lcssa
}
; Trip count should be widened and LFTR should canonicalize the condition
define float @wide_trip_count_test4(float* %b,
; CHECK-LABEL: @wide_trip_count_test4(
; CHECK-NEXT: entry:
; CHECK-NEXT: [[CMP5:%.*]] = icmp sgt i32 [[M:%.*]], 10
; CHECK-NEXT: br i1 [[CMP5]], label [[FOR_BODY_PREHEADER:%.*]], label [[FOR_END:%.*]]
; CHECK: for.body.preheader:
[LFTR] Hoist extend expressions outside of loops w/o waiting for LICM The motivation for this is two fold: 1) Make the output (and thus tests) a bit more readable to a human trying to understand the result of the transform 2) Reduce spurious diffs in a potential future change to restructure all of this logic to use SCEVExpander (which hoists by default) llvm-svn: 365066
2019-07-04 02:18:36 +08:00
; CHECK-NEXT: [[WIDE_TRIP_COUNT:%.*]] = zext i32 [[M]] to i64
[Tests] Consolidate more lftr tests These are all of the ones involving the same data layout string. Remainder take a bit more consideration, but at least everything can be auto-updated now. llvm-svn: 360961
2019-05-17 08:19:28 +08:00
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ], [ 10, [[FOR_BODY_PREHEADER]] ]
; CHECK-NEXT: [[SUM_07:%.*]] = phi float [ [[ADD1:%.*]], [[FOR_BODY]] ], [ 0.000000e+00, [[FOR_BODY_PREHEADER]] ]
; CHECK-NEXT: [[TMP0:%.*]] = add nuw nsw i64 [[INDVARS_IV]], 20
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds float, float* [[B:%.*]], i64 [[TMP0]]
; CHECK-NEXT: [[TEMP:%.*]] = load float, float* [[ARRAYIDX]], align 4
; CHECK-NEXT: [[TMP1:%.*]] = trunc i64 [[INDVARS_IV]] to i32
; CHECK-NEXT: [[CONV:%.*]] = sitofp i32 [[TMP1]] to float
; CHECK-NEXT: [[MUL:%.*]] = fmul float [[CONV]], [[TEMP]]
; CHECK-NEXT: [[ADD1]] = fadd float [[SUM_07]], [[MUL]]
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT]], [[WIDE_TRIP_COUNT]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT:%.*]]
; CHECK: for.end.loopexit:
; CHECK-NEXT: [[ADD1_LCSSA:%.*]] = phi float [ [[ADD1]], [[FOR_BODY]] ]
; CHECK-NEXT: br label [[FOR_END]]
; CHECK: for.end:
; CHECK-NEXT: [[SUM_0_LCSSA:%.*]] = phi float [ 0.000000e+00, [[ENTRY:%.*]] ], [ [[ADD1_LCSSA]], [[FOR_END_LOOPEXIT]] ]
; CHECK-NEXT: ret float [[SUM_0_LCSSA]]
;
i32 signext %m) local_unnamed_addr #0 {
entry:
%cmp5 = icmp sgt i32 %m, 10
br i1 %cmp5, label %for.body.preheader, label %for.end
for.body.preheader: ; preds = %entry
br label %for.body
for.body: ; preds = %for.body.preheader, %for.body
%sum.07 = phi float [ %add1, %for.body ], [ 0.000000e+00, %for.body.preheader ]
%i.06 = phi i32 [ %inc, %for.body ], [ 10, %for.body.preheader ]
%add = add nsw i32 %i.06, 20
%idxprom = sext i32 %add to i64
%arrayidx = getelementptr inbounds float, float* %b, i64 %idxprom
%temp = load float, float* %arrayidx, align 4
%conv = sitofp i32 %i.06 to float
%mul = fmul float %conv, %temp
%add1 = fadd float %sum.07, %mul
%inc = add nsw i32 %i.06, 1
%cmp = icmp slt i32 %inc, %m
br i1 %cmp, label %for.body, label %for.end.loopexit
for.end.loopexit: ; preds = %for.body
%add1.lcssa = phi float [ %add1, %for.body ]
br label %for.end
for.end: ; preds = %for.end.loopexit, %entry
%sum.0.lcssa = phi float [ 0.000000e+00, %entry ], [ %add1.lcssa, %for.end.loopexit ]
ret float %sum.0.lcssa
}
Reduced test case for pr42279 in advance of the relevant re-commit + fix llvm-svn: 363601
2019-06-18 03:27:45 +08:00
define void @ptr_non_cmp_exit_test() {
; CHECK-LABEL: @ptr_non_cmp_exit_test(
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[FOR_BODY29:%.*]]
; CHECK: for.body29:
; CHECK-NEXT: [[IV:%.*]] = phi i8* [ null, [[ENTRY:%.*]] ], [ [[IV_NEXT:%.*]], [[FOR_BODY29]] ]
; CHECK-NEXT: [[TMP0:%.*]] = load volatile i8, i8* [[IV]], align 1
; CHECK-NEXT: [[IV_NEXT]] = getelementptr inbounds i8, i8* [[IV]], i64 1
Fix a bug w/inbounds invalidation in LFTR (recommit) Recommit r363289 with a bug fix for crash identified in pr42279. Issue was that a loop exit test does not have to be an icmp, leading to a null dereference crash when new logic was exercised for that case. Test case previously committed in r363601. Original commit comment follows: This contains fixes for two cases where we might invalidate inbounds and leave it stale in the IR (a miscompile). Case 1 is when switching to an IV with no dynamically live uses, and case 2 is when doing pre-to-post conversion on the same pointer type IV. The basic scheme used is to prove that using the given IV (pre or post increment forms) would have to already trigger UB on the path to the test we're modifying. As such, our potential UB triggering use does not change the semantics of the original program. As was pointed out in the review thread by Nikita, this is defending against a separate issue from the hasConcreteDef case. This is about poison, that's about undef. Unfortunately, the two are different, see Nikita's comment for a fuller explanation, he explains it well. (Note: I'm going to address Nikita's last style comment in a separate commit just to minimize chance of subtle bugs being introduced due to typos.) Differential Revision: https://reviews.llvm.org/D62939 llvm-svn: 363613
2019-06-18 04:32:22 +08:00
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i8* [[IV]], inttoptr (i64 10 to i8*)
Reduced test case for pr42279 in advance of the relevant re-commit + fix llvm-svn: 363601
2019-06-18 03:27:45 +08:00
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_BODY29]], label [[EXIT:%.*]]
; CHECK: exit:
; CHECK-NEXT: ret void
;
entry:
br label %for.body29
for.body29:
%iv = phi i8* [ null, %entry ], [ %iv.next, %for.body29 ]
load volatile i8, i8* %iv, align 1
%iv.next = getelementptr inbounds i8, i8* %iv, i64 1
%cmp = icmp ne i8* %iv.next, inttoptr (i64 11 to i8*)
%and = and i1 %cmp, %cmp
br i1 %and, label %for.body29, label %exit
exit:
ret void
}