diff options
Diffstat (limited to 'test/CodeGen/ARM')
29 files changed, 907 insertions, 117 deletions
diff --git a/test/CodeGen/ARM/2009-09-09-fpcmp-ole.ll b/test/CodeGen/ARM/2009-09-09-fpcmp-ole.ll index 3909c6a..0a157c9 100644 --- a/test/CodeGen/ARM/2009-09-09-fpcmp-ole.ll +++ b/test/CodeGen/ARM/2009-09-09-fpcmp-ole.ll @@ -1,16 +1,16 @@ -; RUN: llc -O1 -march=arm -mattr=+vfp2 < %s | FileCheck %s +; RUN: llc -O1 -march=arm -mattr=+vfp2 -mtriple=arm-linux-gnueabi < %s | FileCheck %s ; pr4939 define void @test(double* %x, double* %y) nounwind { - %1 = load double* %x, align 4 - %2 = load double* %y, align 4 + %1 = load double* %x + %2 = load double* %y %3 = fsub double -0.000000e+00, %1 %4 = fcmp ugt double %2, %3 br i1 %4, label %bb1, label %bb2 bb1: ;CHECK: vstrhi.64 - store double %1, double* %y, align 4 + store double %1, double* %y br label %bb2 bb2: diff --git a/test/CodeGen/ARM/2011-04-07-schediv.ll b/test/CodeGen/ARM/2011-04-07-schediv.ll index a61908f..19f756f 100644 --- a/test/CodeGen/ARM/2011-04-07-schediv.ll +++ b/test/CodeGen/ARM/2011-04-07-schediv.ll @@ -13,6 +13,7 @@ entry: ; Make sure the scheduler schedules all uses of the preincrement ; induction variable before defining the postincrement value. ; CHECK: t: +; CHECK: %bb ; CHECK-NOT: mov bb: ; preds = %entry, %bb %j.05 = phi i32 [ %2, %bb ], [ 0, %entry ] diff --git a/test/CodeGen/ARM/2011-04-11-MachineLICMBug.ll b/test/CodeGen/ARM/2011-04-11-MachineLICMBug.ll index a9dd971..568718c 100644 --- a/test/CodeGen/ARM/2011-04-11-MachineLICMBug.ll +++ b/test/CodeGen/ARM/2011-04-11-MachineLICMBug.ll @@ -14,15 +14,15 @@ for.cond: br i1 %cmp, label %for.body, label %return for.body: -; CHECK: %for.body -; CHECK: movs r{{[0-9]+}}, #1 +; CHECK: %for. +; CHECK: movs r{{[0-9]+}}, #{{[01]}} %arrayidx = getelementptr i32* %A, i32 %0 %tmp4 = load i32* %arrayidx, align 4 %cmp6 = icmp eq i32 %tmp4, %value br i1 %cmp6, label %return, label %for.inc -; CHECK: %for.cond -; CHECK: movs r{{[0-9]+}}, #0 +; CHECK: %for. +; CHECK: movs r{{[0-9]+}}, #{{[01]}} for.inc: %inc = add i32 %0, 1 diff --git a/test/CodeGen/ARM/2011-05-04-MultipleLandingPadSuccs.ll b/test/CodeGen/ARM/2011-05-04-MultipleLandingPadSuccs.ll new file mode 100644 index 0000000..0b5f962 --- /dev/null +++ b/test/CodeGen/ARM/2011-05-04-MultipleLandingPadSuccs.ll @@ -0,0 +1,93 @@ +; RUN: llc < %s -verify-machineinstrs +; <rdar://problem/9187612> +target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:64:64-v128:128:128-a0:0:32-n32" +target triple = "thumbv7-apple-darwin" + +define void @func() unnamed_addr align 2 { +entry: + br label %for.cond + +for.cond: + %tmp2 = phi i32 [ 0, %entry ], [ %add, %for.cond.backedge ] + %cmp = icmp ult i32 %tmp2, 14 + br i1 %cmp, label %for.body, label %for.end + +for.body: + %add = add i32 %tmp2, 1 + switch i32 %tmp2, label %sw.default [ + i32 0, label %sw.bb + i32 1, label %sw.bb + i32 2, label %sw.bb + i32 4, label %sw.bb + i32 5, label %sw.bb + i32 10, label %sw.bb + ] + +sw.bb: + invoke void @foo() + to label %invoke.cont17 unwind label %lpad + +invoke.cont17: + invoke void @foo() + to label %for.cond.backedge unwind label %lpad26 + +for.cond.backedge: + br label %for.cond + +lpad: + %exn = tail call i8* @llvm.eh.exception() nounwind + %eh.selector = tail call i32 (i8*, i8*, ...)* @llvm.eh.selector(i8* %exn, i8* bitcast (i32 (...)* @__gxx_personality_sj0 to i8*), i8* null) nounwind + invoke void @foo() + to label %eh.resume unwind label %terminate.lpad + +lpad26: + %exn27 = tail call i8* @llvm.eh.exception() nounwind + %eh.selector28 = tail call i32 (i8*, i8*, ...)* @llvm.eh.selector(i8* %exn27, i8* bitcast (i32 (...)* @__gxx_personality_sj0 to i8*), i8* null) nounwind + invoke void @foo() + to label %eh.resume unwind label %terminate.lpad + +sw.default: + br label %for.cond.backedge + +for.end: + invoke void @foo() + to label %call8.i.i.i.noexc unwind label %lpad44 + +call8.i.i.i.noexc: + ret void + +lpad44: + %exn45 = tail call i8* @llvm.eh.exception() nounwind + %eh.selector46 = tail call i32 (i8*, i8*, ...)* @llvm.eh.selector(i8* %exn45, i8* bitcast (i32 (...)* @__gxx_personality_sj0 to i8*), i8* null) nounwind + invoke void @foo() + to label %eh.resume unwind label %terminate.lpad + +eh.resume: + %exn.slot.0 = phi i8* [ %exn27, %lpad26 ], [ %exn, %lpad ], [ %exn45, %lpad44 ] + tail call void @_Unwind_SjLj_Resume_or_Rethrow(i8* %exn.slot.0) noreturn + unreachable + +terminate.lpad: + %exn51 = tail call i8* @llvm.eh.exception() nounwind + %eh.selector52 = tail call i32 (i8*, i8*, ...)* @llvm.eh.selector(i8* %exn51, i8* bitcast (i32 (...)* @__gxx_personality_sj0 to i8*), i8* null) nounwind + tail call void @_ZSt9terminatev() noreturn nounwind + unreachable +} + +declare void @foo() + +declare i8* @llvm.eh.exception() nounwind readonly + +declare i32 @__gxx_personality_sj0(...) + +declare i32 @llvm.eh.selector(i8*, i8*, ...) nounwind + +declare void @_Unwind_SjLj_Resume_or_Rethrow(i8*) + +declare void @_ZSt9terminatev() + +!0 = metadata !{metadata !"any pointer", metadata !1} +!1 = metadata !{metadata !"omnipotent char", metadata !2} +!2 = metadata !{metadata !"Simple C/C++ TBAA", null} +!3 = metadata !{metadata !"bool", metadata !1} +!4 = metadata !{metadata !"int", metadata !1} diff --git a/test/CodeGen/ARM/2011-06-09-TailCallByVal.ll b/test/CodeGen/ARM/2011-06-09-TailCallByVal.ll new file mode 100644 index 0000000..4db3acf --- /dev/null +++ b/test/CodeGen/ARM/2011-06-09-TailCallByVal.ll @@ -0,0 +1,39 @@ +; RUN: llc < %s -relocation-model=pic -mcpu=cortex-a8 -arm-tail-calls=1 | FileCheck %s +target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:32:64-v128:32:128-a0:0:32-n32" +target triple = "thumbv7-apple-darwin10" + +%struct._RuneCharClass = type { [14 x i8], i32 } +%struct._RuneEntry = type { i32, i32, i32, i32* } +%struct._RuneLocale = type { [8 x i8], [32 x i8], i32 (i8*, i32, i8**)*, i32 (i32, i8*, i32, i8**)*, i32, [256 x i32], [256 x i32], [256 x i32], %struct._RuneRange, %struct._RuneRange, %struct._RuneRange, i8*, i32, i32, %struct._RuneCharClass* } +%struct._RuneRange = type { i32, %struct._RuneEntry* } +%struct.__collate_st_chain_pri = type { [10 x i32], [2 x i32] } +%struct.__collate_st_char_pri = type { [2 x i32] } +%struct.__collate_st_info = type { [2 x i8], i8, i8, [2 x i32], [2 x i32], i32, i32 } +%struct.__collate_st_large_char_pri = type { i32, %struct.__collate_st_char_pri } +%struct.__collate_st_subst = type { i32, [10 x i32] } +%struct.__xlocale_st_collate = type { i32, void (i8*)*, [32 x i8], %struct.__collate_st_info, [2 x %struct.__collate_st_subst*], %struct.__collate_st_chain_pri*, %struct.__collate_st_large_char_pri*, [256 x %struct.__collate_st_char_pri] } +%struct.__xlocale_st_messages = type { i32, void (i8*)*, i8*, %struct.lc_messages_T } +%struct.__xlocale_st_monetary = type { i32, void (i8*)*, i8*, %struct.lc_monetary_T } +%struct.__xlocale_st_numeric = type { i32, void (i8*)*, i8*, %struct.lc_numeric_T } +%struct.__xlocale_st_runelocale = type { i32, void (i8*)*, [32 x i8], i32, i32, i32 (i32*, i8*, i32, %union.__mbstate_t*, %struct._xlocale*)*, i32 (%union.__mbstate_t*, %struct._xlocale*)*, i32 (i32*, i8**, i32, i32, %union.__mbstate_t*, %struct._xlocale*)*, i32 (i8*, i32, %union.__mbstate_t*, %struct._xlocale*)*, i32 (i8*, i32**, i32, i32, %union.__mbstate_t*, %struct._xlocale*)*, i32, %struct._RuneLocale } +%struct.__xlocale_st_time = type { i32, void (i8*)*, i8*, %struct.lc_time_T } +%struct._xlocale = type { i32, void (i8*)*, %union.__mbstate_t, %union.__mbstate_t, %union.__mbstate_t, %union.__mbstate_t, %union.__mbstate_t, %union.__mbstate_t, %union.__mbstate_t, %union.__mbstate_t, %union.__mbstate_t, %union.__mbstate_t, i32, i64, i8, i8, i8, i8, i8, i8, i8, i8, i8, %struct.__xlocale_st_collate*, %struct.__xlocale_st_runelocale*, %struct.__xlocale_st_messages*, %struct.__xlocale_st_monetary*, %struct.__xlocale_st_numeric*, %struct._xlocale*, %struct.__xlocale_st_time*, %struct.lconv } +%struct.lc_messages_T = type { i8*, i8*, i8*, i8* } +%struct.lc_monetary_T = type { i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8* } +%struct.lc_numeric_T = type { i8*, i8*, i8* } +%struct.lc_time_T = type { [12 x i8*], [12 x i8*], [7 x i8*], [7 x i8*], i8*, i8*, i8*, i8*, i8*, i8*, [12 x i8*], i8*, i8* } +%struct.lconv = type { i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8, i8, i8, i8, i8, i8, i8, i8, i8, i8, i8, i8, i8, i8 } +%union.__mbstate_t = type { i64, [120 x i8] } + +@"\01_fnmatch.initial" = external constant %union.__mbstate_t, align 4 + +; CHECK: _fnmatch +; CHECK: blx _fnmatch1 + +define i32 @"\01_fnmatch"(i8* %pattern, i8* %string, i32 %flags) nounwind optsize { +entry: + %call4 = tail call i32 @fnmatch1(i8* %pattern, i8* %string, i8* %string, i32 %flags, %union.__mbstate_t* byval @"\01_fnmatch.initial", %union.__mbstate_t* byval @"\01_fnmatch.initial", %struct._xlocale* undef, i32 64) optsize + ret i32 %call4 +} + +declare i32 @fnmatch1(i8*, i8*, i8*, i32, %union.__mbstate_t* byval, %union.__mbstate_t* byval, %struct._xlocale*, i32) nounwind optsize diff --git a/test/CodeGen/ARM/arm-modifier.ll b/test/CodeGen/ARM/arm-modifier.ll new file mode 100644 index 0000000..0a7bb6c --- /dev/null +++ b/test/CodeGen/ARM/arm-modifier.ll @@ -0,0 +1,59 @@ +; RUN: llc < %s -march=arm -mattr=+vfp2 | FileCheck %s + +define i32 @foo(float %scale, float %scale2) nounwind { +entry: + %scale.addr = alloca float, align 4 + %scale2.addr = alloca float, align 4 + store float %scale, float* %scale.addr, align 4 + store float %scale2, float* %scale2.addr, align 4 + %tmp = load float* %scale.addr, align 4 + %tmp1 = load float* %scale2.addr, align 4 + call void asm sideeffect "vmul.f32 q0, q0, ${0:y} \0A\09vmul.f32 q1, q1, ${0:y} \0A\09vmul.f32 q1, q0, ${1:y} \0A\09", "w,w,~{q0},~{q1}"(float %tmp, float %tmp1) nounwind + ret i32 0 +} + +define void @f0() nounwind { +entry: +; CHECK: f0 +; CHECK: .word -1 +call void asm sideeffect ".word ${0:B} \0A\09", "i"(i32 0) nounwind +ret void +} + +define void @f1() nounwind { +entry: +; CHECK: f1 +; CHECK: .word 65535 +call void asm sideeffect ".word ${0:L} \0A\09", "i"(i32 -1) nounwind +ret void +} + +@f2_ptr = internal global i32* @f2_var, align 4 +@f2_var = external global i32 + +define void @f2() nounwind { +entry: +; CHECK: f2 +; CHECK: ldr r0, [r{{[0-9]+}}] +call void asm sideeffect "ldr r0, [${0:m}]\0A\09", "*m,~{r0}"(i32** @f2_ptr) nounwind +ret void +} + +@f3_ptr = internal global i64* @f3_var, align 4 +@f3_var = external global i64 +@f3_var2 = external global i64 + +define void @f3() nounwind { +entry: +; CHECK: f3 +; CHECK: stm r{{[0-9]+}}, {[[REG1:(r[0-9]+)]], r{{[0-9]+}}} +; CHECK: adds lr, [[REG1]] +; CHECK: ldm r{{[0-9]+}}, {r{{[0-9]+}}, r{{[0-9]+}}} +%tmp = load i64* @f3_var, align 4 +%tmp1 = load i64* @f3_var2, align 4 +%0 = call i64 asm sideeffect "stm ${0:m}, ${1:M}\0A\09adds $3, $1\0A\09", "=*m,=r,1,r"(i64** @f3_ptr, i64 %tmp, i64 %tmp1) nounwind +store i64 %0, i64* @f3_var, align 4 +%1 = call i64 asm sideeffect "ldm ${1:m}, ${0:M}\0A\09", "=r,*m"(i64** @f3_ptr) nounwind +store i64 %1, i64* @f3_var, align 4 +ret void +} diff --git a/test/CodeGen/ARM/atomic-op.ll b/test/CodeGen/ARM/atomic-op.ll new file mode 100644 index 0000000..03940e3 --- /dev/null +++ b/test/CodeGen/ARM/atomic-op.ll @@ -0,0 +1,103 @@ +; RUN: llc < %s -mtriple=armv7-apple-darwin10 | FileCheck %s +; RUN: llc < %s -mtriple=thumbv7-apple-darwin10 | FileCheck %s + +define void @func(i32 %argc, i8** %argv) nounwind { +entry: + %argc.addr = alloca i32 ; <i32*> [#uses=1] + %argv.addr = alloca i8** ; <i8***> [#uses=1] + %val1 = alloca i32 ; <i32*> [#uses=2] + %val2 = alloca i32 ; <i32*> [#uses=15] + %andt = alloca i32 ; <i32*> [#uses=2] + %ort = alloca i32 ; <i32*> [#uses=2] + %xort = alloca i32 ; <i32*> [#uses=2] + %old = alloca i32 ; <i32*> [#uses=18] + %temp = alloca i32 ; <i32*> [#uses=2] + store i32 %argc, i32* %argc.addr + store i8** %argv, i8*** %argv.addr + store i32 0, i32* %val1 + store i32 31, i32* %val2 + store i32 3855, i32* %andt + store i32 3855, i32* %ort + store i32 3855, i32* %xort + store i32 4, i32* %temp + %tmp = load i32* %temp + ; CHECK: ldrex + ; CHECK: add + ; CHECK: strex + call i32 @llvm.atomic.load.add.i32.p0i32( i32* %val1, i32 %tmp ) ; <i32>:0 [#uses=1] + store i32 %0, i32* %old + ; CHECK: ldrex + ; CHECK: sub + ; CHECK: strex + call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %val2, i32 30 ) ; <i32>:1 [#uses=1] + store i32 %1, i32* %old + ; CHECK: ldrex + ; CHECK: add + ; CHECK: strex + call i32 @llvm.atomic.load.add.i32.p0i32( i32* %val2, i32 1 ) ; <i32>:2 [#uses=1] + store i32 %2, i32* %old + ; CHECK: ldrex + ; CHECK: sub + ; CHECK: strex + call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %val2, i32 1 ) ; <i32>:3 [#uses=1] + store i32 %3, i32* %old + ; CHECK: ldrex + ; CHECK: and + ; CHECK: strex + call i32 @llvm.atomic.load.and.i32.p0i32( i32* %andt, i32 4080 ) ; <i32>:4 [#uses=1] + store i32 %4, i32* %old + ; CHECK: ldrex + ; CHECK: or + ; CHECK: strex + call i32 @llvm.atomic.load.or.i32.p0i32( i32* %ort, i32 4080 ) ; <i32>:5 [#uses=1] + store i32 %5, i32* %old + ; CHECK: ldrex + ; CHECK: eor + ; CHECK: strex + call i32 @llvm.atomic.load.xor.i32.p0i32( i32* %xort, i32 4080 ) ; <i32>:6 [#uses=1] + store i32 %6, i32* %old + ; CHECK: ldrex + ; CHECK: cmp + ; CHECK: strex + call i32 @llvm.atomic.load.min.i32.p0i32( i32* %val2, i32 16 ) ; <i32>:7 [#uses=1] + store i32 %7, i32* %old + %neg = sub i32 0, 1 ; <i32> [#uses=1] + ; CHECK: ldrex + ; CHECK: cmp + ; CHECK: strex + call i32 @llvm.atomic.load.min.i32.p0i32( i32* %val2, i32 %neg ) ; <i32>:8 [#uses=1] + store i32 %8, i32* %old + ; CHECK: ldrex + ; CHECK: cmp + ; CHECK: strex + call i32 @llvm.atomic.load.max.i32.p0i32( i32* %val2, i32 1 ) ; <i32>:9 [#uses=1] + store i32 %9, i32* %old + ; CHECK: ldrex + ; CHECK: cmp + ; CHECK: strex + call i32 @llvm.atomic.load.max.i32.p0i32( i32* %val2, i32 0 ) ; <i32>:10 [#uses=1] + store i32 %10, i32* %old + ret void +} + +declare i32 @llvm.atomic.load.add.i32.p0i32(i32*, i32) nounwind + +declare i32 @llvm.atomic.load.sub.i32.p0i32(i32*, i32) nounwind + +declare i32 @llvm.atomic.load.and.i32.p0i32(i32*, i32) nounwind + +declare i32 @llvm.atomic.load.or.i32.p0i32(i32*, i32) nounwind + +declare i32 @llvm.atomic.load.xor.i32.p0i32(i32*, i32) nounwind + +declare i32 @llvm.atomic.load.min.i32.p0i32(i32*, i32) nounwind + +declare i32 @llvm.atomic.load.max.i32.p0i32(i32*, i32) nounwind + +declare i32 @llvm.atomic.load.umax.i32.p0i32(i32*, i32) nounwind + +declare i32 @llvm.atomic.load.umin.i32.p0i32(i32*, i32) nounwind + +declare i32 @llvm.atomic.swap.i32.p0i32(i32*, i32) nounwind + +declare i32 @llvm.atomic.cmp.swap.i32.p0i32(i32*, i32, i32) nounwind diff --git a/test/CodeGen/ARM/bfi.ll b/test/CodeGen/ARM/bfi.ll index 946db19..c94b096 100644 --- a/test/CodeGen/ARM/bfi.ll +++ b/test/CodeGen/ARM/bfi.ll @@ -31,8 +31,7 @@ define i32 @f3(i32 %A, i32 %B) nounwind { entry: ; CHECK: f3 ; CHECK: lsr{{.*}} #7 -; CHECK: mov r0, r1 -; CHECK: bfi r0, r2, #7, #16 +; CHECK: bfi {{.*}}, #7, #16 %and = and i32 %A, 8388480 ; <i32> [#uses=1] %and2 = and i32 %B, -8388481 ; <i32> [#uses=1] %or = or i32 %and2, %and ; <i32> [#uses=1] @@ -42,8 +41,8 @@ entry: ; rdar://8752056 define i32 @f4(i32 %a) nounwind { ; CHECK: f4 -; CHECK: movw r1, #3137 -; CHECK: bfi r1, r0, #15, #5 +; CHECK: movw [[R1:r[0-9]+]], #3137 +; CHECK: bfi [[R1]], {{r[0-9]+}}, #15, #5 %1 = shl i32 %a, 15 %ins7 = and i32 %1, 1015808 %ins12 = or i32 %ins7, 3137 diff --git a/test/CodeGen/ARM/call-tc.ll b/test/CodeGen/ARM/call-tc.ll index 4dc37aa..c460f7a 100644 --- a/test/CodeGen/ARM/call-tc.ll +++ b/test/CodeGen/ARM/call-tc.ll @@ -1,6 +1,6 @@ ; RUN: llc < %s -mtriple=armv6-apple-darwin -mattr=+vfp2 -arm-tail-calls | FileCheck %s -check-prefix=CHECKV6 ; RUN: llc < %s -mtriple=armv6-linux-gnueabi -relocation-model=pic -mattr=+vfp2 -arm-tail-calls | FileCheck %s -check-prefix=CHECKELF -; RUN: llc < %s -mtriple=thumbv7-apple-darwin -arm-tail-calls | FileCheck %s -check-prefix=CHECKT2 +; RUN: llc < %s -mtriple=thumbv7-apple-darwin -arm-tail-calls | FileCheck %s -check-prefix=CHECKT2D @t = weak global i32 ()* null ; <i32 ()**> [#uses=1] @@ -16,6 +16,10 @@ define void @t1() { define void @t2() { ; CHECKV6: t2: ; CHECKV6: bx r0 @ TAILCALL +; CHECKT2D: t2: +; CHECKT2D: ldr +; CHECKT2D-NEXT: ldr +; CHECKT2D-NEXT: bx r0 @ TAILCALL %tmp = load i32 ()** @t ; <i32 ()*> [#uses=1] %tmp.upgrd.2 = tail call i32 %tmp( ) ; <i32> [#uses=0] ret void @@ -26,6 +30,9 @@ define void @t3() { ; CHECKV6: b _t2 @ TAILCALL ; CHECKELF: t3: ; CHECKELF: b t2(PLT) @ TAILCALL +; CHECKT2D: t3: +; CHECKT2D: b.w _t2 @ TAILCALL + tail call void @t2( ) ; <i32> [#uses=0] ret void } @@ -71,10 +78,10 @@ declare void @foo() nounwind define void @t7() nounwind { entry: -; CHECKT2: t7: -; CHECKT2: blxeq _foo -; CHECKT2-NEXT: pop.w -; CHECKT2-NEXT: b _foo +; CHECKT2D: t7: +; CHECKT2D: blxeq _foo +; CHECKT2D-NEXT: pop.w +; CHECKT2D-NEXT: b.w _foo br i1 undef, label %bb, label %bb1.lr.ph bb1.lr.ph: diff --git a/test/CodeGen/ARM/debug-info-branch-folding.ll b/test/CodeGen/ARM/debug-info-branch-folding.ll new file mode 100644 index 0000000..9bdae43 --- /dev/null +++ b/test/CodeGen/ARM/debug-info-branch-folding.ll @@ -0,0 +1,94 @@ +; RUN: llc < %s - | FileCheck %s +target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:32:64-v128:32:128-a0:0:32-n32" +target triple = "thumbv7-apple-macosx10.6.7" + +;CHECK: vadd.f32 q4, q8, q8 +;CHECK-NEXT: Ltmp +;CHECK-NEXT: @DEBUG_VALUE: y <- Q4+0 +;CHECK-NEXT: @DEBUG_VALUE: x <- Q4+0 + + +@.str = external constant [13 x i8] + +declare <4 x float> @test0001(float) nounwind readnone ssp + +define i32 @main(i32 %argc, i8** nocapture %argv) nounwind ssp { +entry: + br label %for.body9 + +for.body9: ; preds = %for.body9, %entry + %add19 = fadd <4 x float> undef, <float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 1.000000e+00>, !dbg !39 + tail call void @llvm.dbg.value(metadata !{<4 x float> %add19}, i64 0, metadata !27), !dbg !39 + %add20 = fadd <4 x float> undef, <float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 1.000000e+00>, !dbg !39 + tail call void @llvm.dbg.value(metadata !{<4 x float> %add20}, i64 0, metadata !28), !dbg !39 + br i1 undef, label %for.end54, label %for.body9, !dbg !44 + +for.end54: ; preds = %for.body9 + %tmp115 = extractelement <4 x float> %add19, i32 1 + %conv6.i75 = fpext float %tmp115 to double, !dbg !45 + %call.i82 = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([13 x i8]* @.str, i32 0, i32 0), double undef, double %conv6.i75, double undef, double undef) nounwind, !dbg !45 + %tmp116 = extractelement <4 x float> %add20, i32 1 + %conv6.i76 = fpext float %tmp116 to double, !dbg !45 + %call.i83 = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([13 x i8]* @.str, i32 0, i32 0), double undef, double %conv6.i76, double undef, double undef) nounwind, !dbg !45 + ret i32 0, !dbg !49 +} + +declare i32 @printf(i8* nocapture, ...) nounwind + +declare void @llvm.dbg.value(metadata, i64, metadata) nounwind readnone + +!llvm.dbg.sp = !{!0, !10, !14} +!llvm.dbg.lv.test0001 = !{!18} +!llvm.dbg.lv.main = !{!19, !20, !24, !26, !27, !28, !29} +!llvm.dbg.lv.printFV = !{!30} + +!0 = metadata !{i32 589870, i32 0, metadata !1, metadata !"test0001", metadata !"test0001", metadata !"", metadata !1, i32 3, metadata !3, i1 false, i1 true, i32 0, i32 0, i32 0, i32 256, i1 true, <4 x float> (float)* @test0001, null} ; [ DW_TAG_subprogram ] +!1 = metadata !{i32 589865, metadata !"build2.c", metadata !"/private/tmp", metadata !2} ; [ DW_TAG_file_type ] +!2 = metadata !{i32 589841, i32 0, i32 12, metadata !"build2.c", metadata !"/private/tmp", metadata !"clang version 3.0 (trunk 129915)", i1 true, i1 true, metadata !"", i32 0} ; [ DW_TAG_compile_unit ] +!3 = metadata !{i32 589845, metadata !1, metadata !"", metadata !1, i32 0, i64 0, i64 0, i32 0, i32 0, i32 0, metadata !4, i32 0, i32 0} ; [ DW_TAG_subroutine_type ] +!4 = metadata !{metadata !5} +!5 = metadata !{i32 589846, metadata !2, metadata !"v4f32", metadata !1, i32 14, i64 0, i64 0, i64 0, i32 0, metadata !6} ; [ DW_TAG_typedef ] +!6 = metadata !{i32 590083, metadata !2, metadata !"", metadata !2, i32 0, i64 128, i64 128, i32 0, i32 0, metadata !7, metadata !8, i32 0, i32 0} ; [ DW_TAG_vector_type ] +!7 = metadata !{i32 589860, metadata !2, metadata !"float", null, i32 0, i64 32, i64 32, i64 0, i32 0, i32 4} ; [ DW_TAG_base_type ] +!8 = metadata !{metadata !9} +!9 = metadata !{i32 589857, i64 0, i64 3} ; [ DW_TAG_subrange_type ] +!10 = metadata !{i32 589870, i32 0, metadata !1, metadata !"main", metadata !"main", metadata !"", metadata !1, i32 59, metadata !11, i1 false, i1 true, i32 0, i32 0, i32 0, i32 256, i1 true, i32 (i32, i8**)* @main, null} ; [ DW_TAG_subprogram ] +!11 = metadata !{i32 589845, metadata !1, metadata !"", metadata !1, i32 0, i64 0, i64 0, i32 0, i32 0, i32 0, metadata !12, i32 0, i32 0} ; [ DW_TAG_subroutine_type ] +!12 = metadata !{metadata !13} +!13 = metadata !{i32 589860, metadata !2, metadata !"int", null, i32 0, i64 32, i64 32, i64 0, i32 0, i32 5} ; [ DW_TAG_base_type ] +!14 = metadata !{i32 589870, i32 0, metadata !15, metadata !"printFV", metadata !"printFV", metadata !"", metadata !15, i32 41, metadata !16, i1 true, i1 true, i32 0, i32 0, i32 0, i32 256, i1 true, null, null} ; [ DW_TAG_subprogram ] +!15 = metadata !{i32 589865, metadata !"/Volumes/Lalgate/work/llvm/projects/llvm-test/SingleSource/UnitTests/Vector/helpers.h", metadata !"/private/tmp", metadata !2} ; [ DW_TAG_file_type ] +!16 = metadata !{i32 589845, metadata !15, metadata !"", metadata !15, i32 0, i64 0, i64 0, i32 0, i32 0, i32 0, metadata !17, i32 0, i32 0} ; [ DW_TAG_subroutine_type ] +!17 = metadata !{null} +!18 = metadata !{i32 590081, metadata !0, metadata !"a", metadata !1, i32 16777219, metadata !7, i32 0} ; [ DW_TAG_arg_variable ] +!19 = metadata !{i32 590081, metadata !10, metadata !"argc", metadata !1, i32 16777275, metadata !13, i32 0} ; [ DW_TAG_arg_variable ] +!20 = metadata !{i32 590081, metadata !10, metadata !"argv", metadata !1, i32 33554491, metadata !21, i32 0} ; [ DW_TAG_arg_variable ] +!21 = metadata !{i32 589839, metadata !2, metadata !"", null, i32 0, i64 32, i64 32, i64 0, i32 0, metadata !22} ; [ DW_TAG_pointer_type ] +!22 = metadata !{i32 589839, metadata !2, metadata !"", null, i32 0, i64 32, i64 32, i64 0, i32 0, metadata !23} ; [ DW_TAG_pointer_type ] +!23 = metadata !{i32 589860, metadata !2, metadata !"char", null, i32 0, i64 8, i64 8, i64 0, i32 0, i32 6} ; [ DW_TAG_base_type ] +!24 = metadata !{i32 590080, metadata !25, metadata !"i", metadata !1, i32 60, metadata !13, i32 0} ; [ DW_TAG_auto_variable ] +!25 = metadata !{i32 589835, metadata !10, i32 59, i32 33, metadata !1, i32 14} ; [ DW_TAG_lexical_block ] +!26 = metadata !{i32 590080, metadata !25, metadata !"j", metadata !1, i32 60, metadata !13, i32 0} ; [ DW_TAG_auto_variable ] +!27 = metadata !{i32 590080, metadata !25, metadata !"x", metadata !1, i32 61, metadata !5, i32 0} ; [ DW_TAG_auto_variable ] +!28 = metadata !{i32 590080, metadata !25, metadata !"y", metadata !1, i32 62, metadata !5, i32 0} ; [ DW_TAG_auto_variable ] +!29 = metadata !{i32 590080, metadata !25, metadata !"z", metadata !1, i32 63, metadata !5, i32 0} ; [ DW_TAG_auto_variable ] +!30 = metadata !{i32 590081, metadata !14, metadata !"F", metadata !15, i32 16777257, metadata !31, i32 0} ; [ DW_TAG_arg_variable ] +!31 = metadata !{i32 589839, metadata !2, metadata !"", null, i32 0, i64 32, i64 32, i64 0, i32 0, metadata !32} ; [ DW_TAG_pointer_type ] +!32 = metadata !{i32 589846, metadata !2, metadata !"FV", metadata !15, i32 25, i64 0, i64 0, i64 0, i32 0, metadata !33} ; [ DW_TAG_typedef ] +!33 = metadata !{i32 589847, metadata !2, metadata !"", metadata !15, i32 22, i64 128, i64 128, i64 0, i32 0, i32 0, metadata !34, i32 0, i32 0} ; [ DW_TAG_union_type ] +!34 = metadata !{metadata !35, metadata !37} +!35 = metadata !{i32 589837, metadata !15, metadata !"V", metadata !15, i32 23, i64 128, i64 128, i64 0, i32 0, metadata !36} ; [ DW_TAG_member ] +!36 = metadata !{i32 589846, metadata !2, metadata !"v4sf", metadata !15, i32 3, i64 0, i64 0, i64 0, i32 0, metadata !6} ; [ DW_TAG_typedef ] +!37 = metadata !{i32 589837, metadata !15, metadata !"A", metadata !15, i32 24, i64 128, i64 32, i64 0, i32 0, metadata !38} ; [ DW_TAG_member ] +!38 = metadata !{i32 589825, metadata !2, metadata !"", metadata !2, i32 0, i64 128, i64 32, i32 0, i32 0, metadata !7, metadata !8, i32 0, i32 0} ; [ DW_TAG_array_type ] +!39 = metadata !{i32 79, i32 7, metadata !40, null} +!40 = metadata !{i32 589835, metadata !41, i32 75, i32 35, metadata !1, i32 18} ; [ DW_TAG_lexical_block ] +!41 = metadata !{i32 589835, metadata !42, i32 75, i32 5, metadata !1, i32 17} ; [ DW_TAG_lexical_block ] +!42 = metadata !{i32 589835, metadata !43, i32 71, i32 32, metadata !1, i32 16} ; [ DW_TAG_lexical_block ] +!43 = metadata !{i32 589835, metadata !25, i32 71, i32 3, metadata !1, i32 15} ; [ DW_TAG_lexical_block ] +!44 = metadata !{i32 75, i32 5, metadata !42, null} +!45 = metadata !{i32 42, i32 2, metadata !46, metadata !48} +!46 = metadata !{i32 589835, metadata !47, i32 42, i32 2, metadata !15, i32 20} ; [ DW_TAG_lexical_block ] +!47 = metadata !{i32 589835, metadata !14, i32 41, i32 28, metadata !15, i32 19} ; [ DW_TAG_lexical_block ] +!48 = metadata !{i32 95, i32 3, metadata !25, null} +!49 = metadata !{i32 99, i32 3, metadata !25, null} diff --git a/test/CodeGen/ARM/debug-info-sreg2.ll b/test/CodeGen/ARM/debug-info-sreg2.ll new file mode 100644 index 0000000..16aeab3 --- /dev/null +++ b/test/CodeGen/ARM/debug-info-sreg2.ll @@ -0,0 +1,61 @@ +; RUN: llc < %s - | FileCheck %s +; Radar 9376013 +target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:32:64-v128:32:128-a0:0:32-n32" +target triple = "thumbv7-apple-macosx10.6.7" + +;CHECK: Ldebug_loc0: +;CHECK-NEXT: .long Ltmp1 +;CHECK-NEXT: .long Ltmp3 +;CHECK-NEXT: Lset9 = Ltmp10-Ltmp9 @ Loc expr size +;CHECK-NEXT: .short Lset9 +;CHECK-NEXT: Ltmp9: +;CHECK-NEXT: .byte 144 @ DW_OP_regx for S register + +define void @_Z3foov() optsize ssp { +entry: + %call = tail call float @_Z3barv() optsize, !dbg !11 + tail call void @llvm.dbg.value(metadata !{float %call}, i64 0, metadata !5), !dbg !11 + %call16 = tail call float @_Z2f2v() optsize, !dbg !12 + %cmp7 = fcmp olt float %call, %call16, !dbg !12 + br i1 %cmp7, label %for.body, label %for.end, !dbg !12 + +for.body: ; preds = %entry, %for.body + %k.08 = phi float [ %inc, %for.body ], [ %call, %entry ] + %call4 = tail call float @_Z2f3f(float %k.08) optsize, !dbg !13 + %inc = fadd float %k.08, 1.000000e+00, !dbg !14 + %call1 = tail call float @_Z2f2v() optsize, !dbg !12 + %cmp = fcmp olt float %inc, %call1, !dbg !12 + br i1 %cmp, label %for.body, label %for.end, !dbg !12 + +for.end: ; preds = %for.body, %entry + ret void, !dbg !15 +} + +declare float @_Z3barv() optsize + +declare float @_Z2f2v() optsize + +declare float @_Z2f3f(float) optsize + +declare void @llvm.dbg.value(metadata, i64, metadata) nounwind readnone + +!llvm.dbg.cu = !{!0} +!llvm.dbg.sp = !{!1} +!llvm.dbg.lv._Z3foov = !{!5, !8} + +!0 = metadata !{i32 589841, i32 0, i32 4, metadata !"k.cc", metadata !"/private/tmp", metadata !"clang version 3.0 (trunk 130845)", i1 true, i1 true, metadata !"", i32 0} ; [ DW_TAG_compile_unit ] +!1 = metadata !{i32 589870, i32 0, metadata !2, metadata !"foo", metadata !"foo", metadata !"_Z3foov", metadata !2, i32 5, metadata !3, i1 false, i1 true, i32 0, i32 0, i32 0, i32 256, i1 true, void ()* @_Z3foov, null, null} ; [ DW_TAG_subprogram ] +!2 = metadata !{i32 589865, metadata !"k.cc", metadata !"/private/tmp", metadata !0} ; [ DW_TAG_file_type ] +!3 = metadata !{i32 589845, metadata !2, metadata !"", metadata !2, i32 0, i64 0, i64 0, i32 0, i32 0, i32 0, metadata !4, i32 0, i32 0} ; [ DW_TAG_subroutine_type ] +!4 = metadata !{null} +!5 = metadata !{i32 590080, metadata !6, metadata !"k", metadata !2, i32 6, metadata !7, i32 0} ; [ DW_TAG_auto_variable ] +!6 = metadata !{i32 589835, metadata !1, i32 5, i32 12, metadata !2, i32 0} ; [ DW_TAG_lexical_block ] +!7 = metadata !{i32 589860, metadata !0, metadata !"float", null, i32 0, i64 32, i64 32, i64 0, i32 0, i32 4} ; [ DW_TAG_base_type ] +!8 = metadata !{i32 590080, metadata !9, metadata !"y", metadata !2, i32 8, metadata !7, i32 0} ; [ DW_TAG_auto_variable ] +!9 = metadata !{i32 589835, metadata !10, i32 7, i32 25, metadata !2, i32 2} ; [ DW_TAG_lexical_block ] +!10 = metadata !{i32 589835, metadata !6, i32 7, i32 3, metadata !2, i32 1} ; [ DW_TAG_lexical_block ] +!11 = metadata !{i32 6, i32 18, metadata !6, null} +!12 = metadata !{i32 7, i32 3, metadata !6, null} +!13 = metadata !{i32 8, i32 20, metadata !9, null} +!14 = metadata !{i32 7, i32 20, metadata !10, null} +!15 = metadata !{i32 10, i32 1, metadata !6, null} diff --git a/test/CodeGen/ARM/divmod.ll b/test/CodeGen/ARM/divmod.ll deleted file mode 100644 index 34313aa..0000000 --- a/test/CodeGen/ARM/divmod.ll +++ /dev/null @@ -1,58 +0,0 @@ -; RUN: llc < %s -mtriple=arm-apple-ios | FileCheck %s - -define void @foo(i32 %x, i32 %y, i32* nocapture %P) nounwind ssp { -entry: -; CHECK: foo: -; CHECK: bl ___divmodsi4 -; CHECK-NOT: bl ___divmodsi4 - %div = sdiv i32 %x, %y - store i32 %div, i32* %P, align 4 - %rem = srem i32 %x, %y - %arrayidx6 = getelementptr inbounds i32* %P, i32 1 - store i32 %rem, i32* %arrayidx6, align 4 - ret void -} - -define void @bar(i32 %x, i32 %y, i32* nocapture %P) nounwind ssp { -entry: -; CHECK: bar: -; CHECK: bl ___udivmodsi4 -; CHECK-NOT: bl ___udivmodsi4 - %div = udiv i32 %x, %y - store i32 %div, i32* %P, align 4 - %rem = urem i32 %x, %y - %arrayidx6 = getelementptr inbounds i32* %P, i32 1 - store i32 %rem, i32* %arrayidx6, align 4 - ret void -} - -; rdar://9280991 -@flags = external unnamed_addr global i32 -@tabsize = external unnamed_addr global i32 - -define void @do_indent(i32 %cols) nounwind { -entry: -; CHECK: do_indent: - %0 = load i32* @flags, align 4 - %1 = and i32 %0, 67108864 - %2 = icmp eq i32 %1, 0 - br i1 %2, label %bb1, label %bb - -bb: -; CHECK: bl ___divmodsi4 - %3 = load i32* @tabsize, align 4 - %4 = srem i32 %cols, %3 - %5 = sdiv i32 %cols, %3 - %6 = tail call i32 @llvm.objectsize.i32(i8* null, i1 false) - %7 = tail call i8* @__memset_chk(i8* null, i32 9, i32 %5, i32 %6) nounwind - br label %bb1 - -bb1: - %line_indent_len.0 = phi i32 [ %4, %bb ], [ 0, %entry ] - %8 = getelementptr inbounds i8* null, i32 %line_indent_len.0 - store i8 0, i8* %8, align 1 - ret void -} - -declare i32 @llvm.objectsize.i32(i8*, i1) nounwind readnone -declare i8* @__memset_chk(i8*, i32, i32, i32) nounwind diff --git a/test/CodeGen/ARM/eh-resume-darwin.ll b/test/CodeGen/ARM/eh-resume-darwin.ll new file mode 100644 index 0000000..e475508 --- /dev/null +++ b/test/CodeGen/ARM/eh-resume-darwin.ll @@ -0,0 +1,29 @@ +; RUN: llc < %s -march=arm | FileCheck %s +target triple = "armv6-apple-macosx10.6" + +declare void @func() + +declare i8* @llvm.eh.exception() nounwind readonly + +declare i32 @llvm.eh.selector(i8*, i8*, ...) nounwind + +declare void @llvm.eh.resume(i8*, i32) + +declare i32 @__gxx_personality_sj0(...) + +define void @test0() { +entry: + invoke void @func() + to label %cont unwind label %lpad + +cont: + ret void + +lpad: + %exn = call i8* @llvm.eh.exception() + %sel = call i32 (i8*, i8*, ...)* @llvm.eh.selector(i8* %exn, i8* bitcast (i32 (...)* @__gxx_personality_sj0 to i8*), i32 0) + call void @llvm.eh.resume(i8* %exn, i32 %sel) noreturn + unreachable +} + +; CHECK: __Unwind_SjLj_Resume diff --git a/test/CodeGen/ARM/fast-isel-crash2.ll b/test/CodeGen/ARM/fast-isel-crash2.ll new file mode 100644 index 0000000..aa06299 --- /dev/null +++ b/test/CodeGen/ARM/fast-isel-crash2.ll @@ -0,0 +1,9 @@ +; RUN: llc < %s -O0 -mtriple=thumbv7-apple-darwin +; rdar://9515076 +; (Make sure this doesn't crash.) + +define i32 @test(i32 %i) { + %t = trunc i32 %i to i4 + %r = sext i4 %t to i32 + ret i32 %r +} diff --git a/test/CodeGen/ARM/fast-isel.ll b/test/CodeGen/ARM/fast-isel.ll index 6aad92f..499c97f 100644 --- a/test/CodeGen/ARM/fast-isel.ll +++ b/test/CodeGen/ARM/fast-isel.ll @@ -61,3 +61,100 @@ b3: ; THUMB: orr {{.*}} #4 ; ARM: orr {{.*}} #4 } + +define void @test3(i32 %tmp, i32* %ptr1, i16* %ptr2, i8* %ptr3) nounwind { +; THUMB: test3: +; ARM: test3: + +bb1: + %a1 = trunc i32 %tmp to i16 + %a2 = trunc i16 %a1 to i8 + %a3 = trunc i8 %a2 to i1 + %a4 = zext i1 %a3 to i8 + store i8 %a4, i8* %ptr3 + %a5 = zext i8 %a4 to i16 + store i16 %a5, i16* %ptr2 + %a6 = zext i16 %a5 to i32 + store i32 %a6, i32* %ptr1 + br label %bb2 + +; THUMB: and +; THUMB: strb +; THUMB: uxtb +; THUMB: strh +; THUMB: uxth +; ARM: and +; ARM: strb +; ARM: uxtb +; ARM: strh +; ARM: uxth + +bb2: + %b1 = trunc i32 %tmp to i16 + %b2 = trunc i16 %b1 to i8 + store i8 %b2, i8* %ptr3 + %b3 = sext i8 %b2 to i16 + store i16 %b3, i16* %ptr2 + %b4 = sext i16 %b3 to i32 + store i32 %b4, i32* %ptr1 + br label %bb3 + +; THUMB: strb +; THUMB: sxtb +; THUMB: strh +; THUMB: sxth +; ARM: strb +; ARM: sxtb +; ARM: strh +; ARM: sxth + +bb3: + %c1 = load i8* %ptr3 + %c2 = load i16* %ptr2 + %c3 = load i32* %ptr1 + %c4 = zext i8 %c1 to i32 + %c5 = sext i16 %c2 to i32 + %c6 = add i32 %c4, %c5 + %c7 = sub i32 %c3, %c6 + store i32 %c7, i32* %ptr1 + ret void + +; THUMB: ldrb +; THUMB: ldrh +; THUMB: uxtb +; THUMB: sxth +; THUMB: add +; THUMB: sub +; ARM: ldrb +; ARM: ldrh +; ARM: uxtb +; ARM: sxth +; ARM: add +; ARM: sub +} + +; Check loads/stores with globals +@test4g = external global i32 + +define void @test4() { + %a = load i32* @test4g + %b = add i32 %a, 1 + store i32 %b, i32* @test4g + ret void + +; THUMB: ldr.n r0, LCPI4_1 +; THUMB: ldr r0, [r0] +; THUMB: ldr r0, [r0] +; THUMB: adds r0, #1 +; THUMB: ldr.n r1, LCPI4_0 +; THUMB: ldr r1, [r1] +; THUMB: str r0, [r1] + +; ARM: ldr r0, LCPI4_1 +; ARM: ldr r0, [r0] +; ARM: ldr r0, [r0] +; ARM: add r0, r0, #1 +; ARM: ldr r1, LCPI4_0 +; ARM: ldr r1, [r1] +; ARM: str r0, [r1] +} diff --git a/test/CodeGen/ARM/fcopysign.ll b/test/CodeGen/ARM/fcopysign.ll index f241c26..c4dbeb9 100644 --- a/test/CodeGen/ARM/fcopysign.ll +++ b/test/CodeGen/ARM/fcopysign.ll @@ -10,7 +10,7 @@ entry: ; HARD: test1: ; HARD: vmov.i32 [[REG1:(d[0-9]+)]], #0x80000000 -; HARD: vbsl [[REG1]], d2, d0 +; HARD: vbsl [[REG1]], d %0 = tail call float @copysignf(float %x, float %y) nounwind ret float %0 } diff --git a/test/CodeGen/ARM/fnmscs.ll b/test/CodeGen/ARM/fnmscs.ll index 9facf20..6081712 100644 --- a/test/CodeGen/ARM/fnmscs.ll +++ b/test/CodeGen/ARM/fnmscs.ll @@ -29,7 +29,7 @@ entry: ; NEON: vnmla.f32 ; A8: t2: -; A8: vnmul.f32 s{{[0123]}}, s{{[0123]}}, s{{[0123]}} +; A8: vnmul.f32 s{{[01234]}}, s{{[01234]}}, s{{[01234]}} ; A8: vsub.f32 d{{[0-9]}}, d{{[0-9]}}, d{{[0-9]}} %0 = fmul float %a, %b %1 = fmul float -1.0, %0 diff --git a/test/CodeGen/ARM/inlineasm3.ll b/test/CodeGen/ARM/inlineasm3.ll index 9d6eba8..58687b9 100644 --- a/test/CodeGen/ARM/inlineasm3.ll +++ b/test/CodeGen/ARM/inlineasm3.ll @@ -23,3 +23,38 @@ entry: %asmtmp2 = tail call i32 asm sideeffect "vmov d30, $1\0Avmov.32 $0, d30[0]\0A", "=r,w,~{d30}"(<2 x i32> undef) nounwind ret void } + +; Radar 9306086 + +%0 = type { <8 x i8>, <16 x i8>* } + +define hidden void @conv4_8_E() nounwind { +entry: +%asmtmp31 = call %0 asm "vld1.u8 {$0}, [$1, :128]!\0A", "=w,=r,1"(<16 x i8>* undef) nounwind +unreachable +} + +; Radar 9037836 & 9119939 + +define i32 @t3() nounwind { +entry: +tail call void asm sideeffect "flds s15, $0 \0A", "^Uv|m,~{s15}"(float 1.000000e+00) nounwind +ret i32 0 +} + +; Radar 9037836 & 9119939 + +@k.2126 = internal unnamed_addr global float 1.000000e+00 +define i32 @t4() nounwind { +entry: +call void asm sideeffect "flds s15, $0 \0A", "*^Uv,~{s15}"(float* @k.2126) nounwind +ret i32 0 +} + +; Radar 9037836 & 9119939 + +define i32 @t5() nounwind { +entry: +call void asm sideeffect "flds s15, $0 \0A", "*^Uvm,~{s15}"(float* @k.2126) nounwind +ret i32 0 +} diff --git a/test/CodeGen/ARM/intrinsics.ll b/test/CodeGen/ARM/intrinsics.ll new file mode 100644 index 0000000..54cc3e0 --- /dev/null +++ b/test/CodeGen/ARM/intrinsics.ll @@ -0,0 +1,39 @@ +; RUN: llc < %s -mtriple=armv7-eabi -mcpu=cortex-a8 | FileCheck %s +; RUN: llc < %s -march=thumb -mtriple=thumbv7-eabi -mcpu=cortex-a8 | FileCheck %s + +define void @coproc() nounwind { +entry: + ; CHECK: mrc + %0 = tail call i32 @llvm.arm.mrc(i32 7, i32 1, i32 1, i32 1, i32 4) nounwind + ; CHECK: mcr + tail call void @llvm.arm.mcr(i32 7, i32 1, i32 %0, i32 1, i32 1, i32 4) nounwind + ; CHECK: mrc2 + %1 = tail call i32 @llvm.arm.mrc2(i32 7, i32 1, i32 1, i32 1, i32 4) nounwind + ; CHECK: mcr2 + tail call void @llvm.arm.mcr2(i32 7, i32 1, i32 %1, i32 1, i32 1, i32 4) nounwind + ; CHECK: mcrr + tail call void @llvm.arm.mcrr(i32 7, i32 1, i32 %0, i32 %1, i32 1) nounwind + ; CHECK: mcrr2 + tail call void @llvm.arm.mcrr2(i32 7, i32 1, i32 %0, i32 %1, i32 1) nounwind + ; CHECK: cdp + tail call void @llvm.arm.cdp(i32 7, i32 3, i32 1, i32 1, i32 1, i32 5) nounwind + ; CHECK: cdp2 + tail call void @llvm.arm.cdp2(i32 7, i32 3, i32 1, i32 1, i32 1, i32 5) nounwind + ret void +} + +declare void @llvm.arm.cdp2(i32, i32, i32, i32, i32, i32) nounwind + +declare void @llvm.arm.cdp(i32, i32, i32, i32, i32, i32) nounwind + +declare void @llvm.arm.mcrr2(i32, i32, i32, i32, i32) nounwind + +declare void @llvm.arm.mcrr(i32, i32, i32, i32, i32) nounwind + +declare void @llvm.arm.mcr2(i32, i32, i32, i32, i32, i32) nounwind + +declare i32 @llvm.arm.mrc2(i32, i32, i32, i32, i32) nounwind + +declare void @llvm.arm.mcr(i32, i32, i32, i32, i32, i32) nounwind + +declare i32 @llvm.arm.mrc(i32, i32, i32, i32, i32) nounwind diff --git a/test/CodeGen/ARM/ldst-f32-2-i32.ll b/test/CodeGen/ARM/ldst-f32-2-i32.ll index 2d016f6..1c69e15 100644 --- a/test/CodeGen/ARM/ldst-f32-2-i32.ll +++ b/test/CodeGen/ARM/ldst-f32-2-i32.ll @@ -10,8 +10,8 @@ entry: br i1 %0, label %return, label %bb bb: -; CHECK: ldr [[REGISTER:(r[0-9]+)]], [r1], r3 -; CHECK: str [[REGISTER]], [r2], #4 +; CHECK: ldr [[REGISTER:(r[0-9]+)]], [{{r[0-9]+}}], {{r[0-9]+}} +; CHECK: str [[REGISTER]], [{{r[0-9]+}}], #4 %j.05 = phi i32 [ %2, %bb ], [ 0, %entry ] %tmp = mul i32 %j.05, %index %uglygep = getelementptr i8* %src6, i32 %tmp diff --git a/test/CodeGen/ARM/ldstrexd.ll b/test/CodeGen/ARM/ldstrexd.ll new file mode 100644 index 0000000..0c0911a --- /dev/null +++ b/test/CodeGen/ARM/ldstrexd.ll @@ -0,0 +1,33 @@ +; RUN: llc < %s -mtriple=armv7-apple-darwin | FileCheck %s +; RUN: llc < %s -mtriple=thumbv7-apple-darwin | FileCheck %s + +%0 = type { i32, i32 } + +; CHECK: f0: +; CHECK: ldrexd +define i64 @f0(i8* %p) nounwind readonly { +entry: + %ldrexd = tail call %0 @llvm.arm.ldrexd(i8* %p) + %0 = extractvalue %0 %ldrexd, 1 + %1 = extractvalue %0 %ldrexd, 0 + %2 = zext i32 %0 to i64 + %3 = zext i32 %1 to i64 + %shl = shl nuw i64 %2, 32 + %4 = or i64 %shl, %3 + ret i64 %4 +} + +; CHECK: f1: +; CHECK: strexd +define i32 @f1(i8* %ptr, i64 %val) nounwind { +entry: + %tmp4 = trunc i64 %val to i32 + %tmp6 = lshr i64 %val, 32 + %tmp7 = trunc i64 %tmp6 to i32 + %strexd = tail call i32 @llvm.arm.strexd(i32 %tmp4, i32 %tmp7, i8* %ptr) + ret i32 %strexd +} + +declare %0 @llvm.arm.ldrexd(i8*) nounwind readonly +declare i32 @llvm.arm.strexd(i32, i32, i8*) nounwind + diff --git a/test/CodeGen/ARM/lsr-unfolded-offset.ll b/test/CodeGen/ARM/lsr-unfolded-offset.ll new file mode 100644 index 0000000..e3e6eae --- /dev/null +++ b/test/CodeGen/ARM/lsr-unfolded-offset.ll @@ -0,0 +1,80 @@ +; RUN: llc -regalloc=greedy < %s | FileCheck %s + +; LSR shouldn't introduce more induction variables than needed, increasing +; register pressure and therefore spilling. There is more room for improvement +; here. + +; CHECK: sub sp, #{{32|24}} + +; CHECK: ldr r{{.*}}, [sp, #4] +; CHECK-NEXT: ldr r{{.*}}, [sp, #16] +; CHECK-NEXT: ldr r{{.*}}, [sp, #12] +; CHECK-NEXT: adds + +target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:32:64-v128:32:128-a0:0:32-n32" +target triple = "thumbv7-apple-macosx10.7.0" + +%struct.partition_entry = type { i32, i32, i64, i64 } + +define i32 @partition_overlap_check(%struct.partition_entry* nocapture %part, i32 %num_entries) nounwind readonly optsize ssp { +entry: + %cmp79 = icmp sgt i32 %num_entries, 0 + br i1 %cmp79, label %outer.loop, label %for.end72 + +outer.loop: ; preds = %for.inc69, %entry + %overlap.081 = phi i32 [ %overlap.4, %for.inc69 ], [ 0, %entry ] + %0 = phi i32 [ %inc71, %for.inc69 ], [ 0, %entry ] + %offset = getelementptr %struct.partition_entry* %part, i32 %0, i32 2 + %len = getelementptr %struct.partition_entry* %part, i32 %0, i32 3 + %tmp5 = load i64* %offset, align 4, !tbaa !0 + %tmp15 = load i64* %len, align 4, !tbaa !0 + %add = add nsw i64 %tmp15, %tmp5 + br label %inner.loop + +inner.loop: ; preds = %for.inc, %outer.loop + %overlap.178 = phi i32 [ %overlap.081, %outer.loop ], [ %overlap.4, %for.inc ] + %1 = phi i32 [ 0, %outer.loop ], [ %inc, %for.inc ] + %cmp23 = icmp eq i32 %0, %1 + br i1 %cmp23, label %for.inc, label %if.end + +if.end: ; preds = %inner.loop + %len39 = getelementptr %struct.partition_entry* %part, i32 %1, i32 3 + %offset28 = getelementptr %struct.partition_entry* %part, i32 %1, i32 2 + %tmp29 = load i64* %offset28, align 4, !tbaa !0 + %tmp40 = load i64* %len39, align 4, !tbaa !0 + %add41 = add nsw i64 %tmp40, %tmp29 + %cmp44 = icmp sge i64 %tmp29, %tmp5 + %cmp47 = icmp slt i64 %tmp29, %add + %or.cond = and i1 %cmp44, %cmp47 + %overlap.2 = select i1 %or.cond, i32 1, i32 %overlap.178 + %cmp52 = icmp sle i64 %add41, %add + %cmp56 = icmp sgt i64 %add41, %tmp5 + %or.cond74 = and i1 %cmp52, %cmp56 + %overlap.3 = select i1 %or.cond74, i32 1, i32 %overlap.2 + %cmp61 = icmp sgt i64 %tmp29, %tmp5 + %cmp65 = icmp slt i64 %add41, %add + %or.cond75 = or i1 %cmp61, %cmp65 + br i1 %or.cond75, label %for.inc, label %if.then66 + +if.then66: ; preds = %if.end + br label %for.inc + +for.inc: ; preds = %if.end, %if.then66, %inner.loop + %overlap.4 = phi i32 [ %overlap.178, %inner.loop ], [ 1, %if.then66 ], [ %overlap.3, %if.end ] + %inc = add nsw i32 %1, 1 + %exitcond = icmp eq i32 %inc, %num_entries + br i1 %exitcond, label %for.inc69, label %inner.loop + +for.inc69: ; preds = %for.inc + %inc71 = add nsw i32 %0, 1 + %exitcond83 = icmp eq i32 %inc71, %num_entries + br i1 %exitcond83, label %for.end72, label %outer.loop + +for.end72: ; preds = %for.inc69, %entry + %overlap.0.lcssa = phi i32 [ 0, %entry ], [ %overlap.4, %for.inc69 ] + ret i32 %overlap.0.lcssa +} + +!0 = metadata !{metadata !"long long", metadata !1} +!1 = metadata !{metadata !"omnipotent char", metadata !2} +!2 = metadata !{metadata !"Simple C/C++ TBAA", null} diff --git a/test/CodeGen/ARM/memfunc.ll b/test/CodeGen/ARM/memfunc.ll index 41d5944..032129d 100644 --- a/test/CodeGen/ARM/memfunc.ll +++ b/test/CodeGen/ARM/memfunc.ll @@ -1,10 +1,26 @@ -; RUN: llc < %s -march=arm +; RUN: llc < %s -march=arm -o - | FileCheck %s +; RUN: llc < %s -mtriple=arm-none-eabi -o - | FileCheck --check-prefix=EABI %s + +@from = common global [500 x i32] zeroinitializer, align 4 +@to = common global [500 x i32] zeroinitializer, align 4 define void @f() { entry: - call void @llvm.memmove.i32( i8* null, i8* null, i32 64, i32 0 ) - call void @llvm.memcpy.i32( i8* null, i8* null, i32 64, i32 0 ) - call void @llvm.memset.i32( i8* null, i8 64, i32 0, i32 0 ) + + ; CHECK: memmove + ; EABI: __aeabi_memmove + call void @llvm.memmove.i32( i8* bitcast ([500 x i32]* @from to i8*), i8* bitcast ([500 x i32]* @to to i8*), i32 500, i32 0 ) + + ; CHECK: memcpy + ; EABI: __aeabi_memcpy + call void @llvm.memcpy.i32( i8* bitcast ([500 x i32]* @from to i8*), i8* bitcast ([500 x i32]* @to to i8*), i32 500, i32 0 ) + + ; EABI memset swaps arguments + ; CHECK: mov r1, #0 + ; CHECK: memset + ; EABI: mov r2, #0 + ; EABI: __aeabi_memset + call void @llvm.memset.i32( i8* bitcast ([500 x i32]* @from to i8*), i8 0, i32 500, i32 0 ) unreachable } diff --git a/test/CodeGen/ARM/movt-movw-global.ll b/test/CodeGen/ARM/movt-movw-global.ll index 886ff3f..991d728 100644 --- a/test/CodeGen/ARM/movt-movw-global.ll +++ b/test/CodeGen/ARM/movt-movw-global.ll @@ -1,20 +1,39 @@ -; RUN: llc < %s | FileCheck %s -target datalayout = "e-p:32:32:32-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" -target triple = "armv7-eabi" +; RUN: llc < %s -mtriple=armv7-eabi | FileCheck %s -check-prefix=EABI +; RUN: llc < %s -mtriple=armv7-apple-ios -relocation-model=dynamic-no-pic | FileCheck %s -check-prefix=IOS +; RUN: llc < %s -mtriple=armv7-apple-ios -relocation-model=pic | FileCheck %s -check-prefix=IOS-PIC +; RUN: llc < %s -mtriple=armv7-apple-ios -relocation-model=static | FileCheck %s -check-prefix=IOS-STATIC -@foo = common global i32 0 ; <i32*> [#uses=1] +@foo = common global i32 0 -define arm_aapcs_vfpcc i32* @bar1() nounwind readnone { +define i32* @bar1() nounwind readnone { entry: -; CHECK: movw r0, :lower16:foo -; CHECK-NEXT: movt r0, :upper16:foo +; EABI: movw r0, :lower16:foo +; EABI-NEXT: movt r0, :upper16:foo + +; IOS: movw r0, :lower16:L_foo$non_lazy_ptr +; IOS-NEXT: movt r0, :upper16:L_foo$non_lazy_ptr + +; IOS-PIC: movw r0, :lower16:(L_foo$non_lazy_ptr-(LPC0_0+8)) +; IOS-PIC-NEXT: movt r0, :upper16:(L_foo$non_lazy_ptr-(LPC0_0+8)) + +; IOS-STATIC-NOT: movw r0, :lower16:_foo +; IOS-STATIC-NOT: movt r0, :upper16:_foo ret i32* @foo } -define arm_aapcs_vfpcc void @bar2(i32 %baz) nounwind { +define void @bar2(i32 %baz) nounwind { entry: -; CHECK: movw r1, :lower16:foo -; CHECK-NEXT: movt r1, :upper16:foo +; EABI: movw r1, :lower16:foo +; EABI-NEXT: movt r1, :upper16:foo + +; IOS: movw r1, :lower16:L_foo$non_lazy_ptr +; IOS-NEXT: movt r1, :upper16:L_foo$non_lazy_ptr + +; IOS-PIC: movw r1, :lower16:(L_foo$non_lazy_ptr-(LPC1_0+8)) +; IOS-PIC-NEXT: movt r1, :upper16:(L_foo$non_lazy_ptr-(LPC1_0+8)) + +; IOS-STATIC-NOT: movw r1, :lower16:_foo +; IOS-STATIC-NOT: movt r1, :upper16:_foo store i32 %baz, i32* @foo, align 4 ret void } diff --git a/test/CodeGen/ARM/select-imm.ll b/test/CodeGen/ARM/select-imm.ll index 82ed018..43f8a66 100644 --- a/test/CodeGen/ARM/select-imm.ll +++ b/test/CodeGen/ARM/select-imm.ll @@ -5,8 +5,8 @@ define i32 @t1(i32 %c) nounwind readnone { entry: ; ARM: t1: -; ARM: mov r1, #101 -; ARM: orr r1, r1, #1, #24 +; ARM: mov [[R1:r[0-9]+]], #101 +; ARM: orr [[R1b:r[0-9]+]], [[R1]], #1, #24 ; ARM: movgt r0, #123 ; ARMT2: t1: @@ -34,7 +34,7 @@ entry: ; ARMT2: movwgt r0, #357 ; THUMB2: t2: -; THUMB2: mov.w r0, #123 +; THUMB2: mov{{(s|\.w)}} r0, #123 ; THUMB2: movwgt r0, #357 %0 = icmp sgt i32 %c, 1 @@ -53,7 +53,7 @@ entry: ; ARMT2: moveq r0, #1 ; THUMB2: t3: -; THUMB2: mov.w r0, #0 +; THUMB2: mov{{(s|\.w)}} r0, #0 ; THUMB2: moveq r0, #1 %0 = icmp eq i32 %a, 160 %1 = zext i1 %0 to i32 @@ -67,11 +67,11 @@ entry: ; ARM: movlt ; ARMT2: t4: -; ARMT2: movwlt r0, #65365 -; ARMT2: movtlt r0, #65365 +; ARMT2: movwlt [[R0:r[0-9]+]], #65365 +; ARMT2: movtlt [[R0]], #65365 ; THUMB2: t4: -; THUMB2: mvnlt.w r0, #11141290 +; THUMB2: mvnlt.w [[R0:r[0-9]+]], #11141290 %0 = icmp slt i32 %a, %b %1 = select i1 %0, i32 4283826005, i32 %x ret i32 %1 diff --git a/test/CodeGen/ARM/stm.ll b/test/CodeGen/ARM/stm.ll index 2f5fadb..82dc14d 100644 --- a/test/CodeGen/ARM/stm.ll +++ b/test/CodeGen/ARM/stm.ll @@ -9,7 +9,7 @@ define i32 @main() nounwind { entry: ; CHECK: main ; CHECK: push -; CHECK: stmib +; CHECK: stm %0 = tail call i32 (i8*, ...)* @printf(i8* getelementptr ([26 x i8]* @"\01LC1", i32 0, i32 0), i32 -2, i32 -3, i32 2, i32 -6) nounwind ; <i32> [#uses=0] %1 = tail call i32 (i8*, ...)* @printf(i8* getelementptr ([32 x i8]* @"\01LC", i32 0, i32 0), i32 0, i32 1, i32 0, i32 1, i32 0, i32 1) nounwind ; <i32> [#uses=0] ret i32 0 diff --git a/test/CodeGen/ARM/vldlane.ll b/test/CodeGen/ARM/vldlane.ll index 805aad5..0d7d4ec 100644 --- a/test/CodeGen/ARM/vldlane.ll +++ b/test/CodeGen/ARM/vldlane.ll @@ -125,7 +125,7 @@ define <2 x i32> @vld2lanei32(i32* %A, <2 x i32>* %B) nounwind { ;Check for a post-increment updating load. define <2 x i32> @vld2lanei32_update(i32** %ptr, <2 x i32>* %B) nounwind { ;CHECK: vld2lanei32_update: -;CHECK: vld2.32 {d16[1], d17[1]}, [r1]! +;CHECK: vld2.32 {d16[1], d17[1]}, [{{r[0-9]+}}]! %A = load i32** %ptr %tmp0 = bitcast i32* %A to i8* %tmp1 = load <2 x i32>* %B @@ -153,7 +153,7 @@ define <2 x float> @vld2lanef(float* %A, <2 x float>* %B) nounwind { define <8 x i16> @vld2laneQi16(i16* %A, <8 x i16>* %B) nounwind { ;CHECK: vld2laneQi16: ;Check the (default) alignment. -;CHECK: vld2.16 {d17[1], d19[1]}, [r0] +;CHECK: vld2.16 {d17[1], d19[1]}, [{{r[0-9]+}}] %tmp0 = bitcast i16* %A to i8* %tmp1 = load <8 x i16>* %B %tmp2 = call %struct.__neon_int16x8x2_t @llvm.arm.neon.vld2lane.v8i16(i8* %tmp0, <8 x i16> %tmp1, <8 x i16> %tmp1, i32 5, i32 1) @@ -166,7 +166,7 @@ define <8 x i16> @vld2laneQi16(i16* %A, <8 x i16>* %B) nounwind { define <4 x i32> @vld2laneQi32(i32* %A, <4 x i32>* %B) nounwind { ;CHECK: vld2laneQi32: ;Check the alignment value. Max for this instruction is 64 bits: -;CHECK: vld2.32 {d17[0], d19[0]}, [r0, :64] +;CHECK: vld2.32 {d17[0], d19[0]}, [{{r[0-9]+}}, :64] %tmp0 = bitcast i32* %A to i8* %tmp1 = load <4 x i32>* %B %tmp2 = call %struct.__neon_int32x4x2_t @llvm.arm.neon.vld2lane.v4i32(i8* %tmp0, <4 x i32> %tmp1, <4 x i32> %tmp1, i32 2, i32 16) @@ -222,7 +222,7 @@ define <8 x i8> @vld3lanei8(i8* %A, <8 x i8>* %B) nounwind { define <4 x i16> @vld3lanei16(i16* %A, <4 x i16>* %B) nounwind { ;CHECK: vld3lanei16: ;Check the (default) alignment value. VLD3 does not support alignment. -;CHECK: vld3.16 {d16[1], d17[1], d18[1]}, [r0] +;CHECK: vld3.16 {d{{.*}}[1], d{{.*}}[1], d{{.*}}[1]}, [{{r[0-9]+}}] %tmp0 = bitcast i16* %A to i8* %tmp1 = load <4 x i16>* %B %tmp2 = call %struct.__neon_int16x4x3_t @llvm.arm.neon.vld3lane.v4i16(i8* %tmp0, <4 x i16> %tmp1, <4 x i16> %tmp1, <4 x i16> %tmp1, i32 1, i32 8) @@ -265,7 +265,7 @@ define <2 x float> @vld3lanef(float* %A, <2 x float>* %B) nounwind { define <8 x i16> @vld3laneQi16(i16* %A, <8 x i16>* %B) nounwind { ;CHECK: vld3laneQi16: ;Check the (default) alignment value. VLD3 does not support alignment. -;CHECK: vld3.16 {d16[1], d18[1], d20[1]}, [r0] +;CHECK: vld3.16 {d{{.*}}[1], d{{.*}}[1], d{{.*}}[1]}, [{{r[0-9]+}}] %tmp0 = bitcast i16* %A to i8* %tmp1 = load <8 x i16>* %B %tmp2 = call %struct.__neon_int16x8x3_t @llvm.arm.neon.vld3lane.v8i16(i8* %tmp0, <8 x i16> %tmp1, <8 x i16> %tmp1, <8 x i16> %tmp1, i32 1, i32 8) @@ -280,7 +280,7 @@ define <8 x i16> @vld3laneQi16(i16* %A, <8 x i16>* %B) nounwind { ;Check for a post-increment updating load with register increment. define <8 x i16> @vld3laneQi16_update(i16** %ptr, <8 x i16>* %B, i32 %inc) nounwind { ;CHECK: vld3laneQi16_update: -;CHECK: vld3.16 {d16[1], d18[1], d20[1]}, [{{r[0-9]+}}], {{r[0-9]+}} +;CHECK: vld3.16 {d{{.*}}[1], d{{.*}}[1], d{{.*}}[1]}, [{{r[0-9]+}}], {{r[0-9]+}} %A = load i16** %ptr %tmp0 = bitcast i16* %A to i8* %tmp1 = load <8 x i16>* %B @@ -344,7 +344,7 @@ declare %struct.__neon_float32x4x3_t @llvm.arm.neon.vld3lane.v4f32(i8*, <4 x flo define <8 x i8> @vld4lanei8(i8* %A, <8 x i8>* %B) nounwind { ;CHECK: vld4lanei8: ;Check the alignment value. Max for this instruction is 32 bits: -;CHECK: vld4.8 {d16[1], d17[1], d18[1], d19[1]}, [r0, :32] +;CHECK: vld4.8 {d{{.*}}[1], d{{.*}}[1], d{{.*}}[1], d{{.*}}[1]}, [{{r[0-9]+}}, :32] %tmp1 = load <8 x i8>* %B %tmp2 = call %struct.__neon_int8x8x4_t @llvm.arm.neon.vld4lane.v8i8(i8* %A, <8 x i8> %tmp1, <8 x i8> %tmp1, <8 x i8> %tmp1, <8 x i8> %tmp1, i32 1, i32 8) %tmp3 = extractvalue %struct.__neon_int8x8x4_t %tmp2, 0 @@ -360,7 +360,7 @@ define <8 x i8> @vld4lanei8(i8* %A, <8 x i8>* %B) nounwind { ;Check for a post-increment updating load. define <8 x i8> @vld4lanei8_update(i8** %ptr, <8 x i8>* %B) nounwind { ;CHECK: vld4lanei8_update: -;CHECK: vld4.8 {d16[1], d17[1], d18[1], d19[1]}, [r1, :32]! +;CHECK: vld4.8 {d16[1], d17[1], d18[1], d19[1]}, [{{r[0-9]+}}, :32]! %A = load i8** %ptr %tmp1 = load <8 x i8>* %B %tmp2 = call %struct.__neon_int8x8x4_t @llvm.arm.neon.vld4lane.v8i8(i8* %A, <8 x i8> %tmp1, <8 x i8> %tmp1, <8 x i8> %tmp1, <8 x i8> %tmp1, i32 1, i32 8) @@ -380,7 +380,7 @@ define <4 x i16> @vld4lanei16(i16* %A, <4 x i16>* %B) nounwind { ;CHECK: vld4lanei16: ;Check that a power-of-two alignment smaller than the total size of the memory ;being loaded is ignored. -;CHECK: vld4.16 {d16[1], d17[1], d18[1], d19[1]}, [r0] +;CHECK: vld4.16 {d16[1], d17[1], d18[1], d19[1]}, [{{r[0-9]+}}] %tmp0 = bitcast i16* %A to i8* %tmp1 = load <4 x i16>* %B %tmp2 = call %struct.__neon_int16x4x4_t @llvm.arm.neon.vld4lane.v4i16(i8* %tmp0, <4 x i16> %tmp1, <4 x i16> %tmp1, <4 x i16> %tmp1, <4 x i16> %tmp1, i32 1, i32 4) @@ -398,7 +398,7 @@ define <2 x i32> @vld4lanei32(i32* %A, <2 x i32>* %B) nounwind { ;CHECK: vld4lanei32: ;Check the alignment value. An 8-byte alignment is allowed here even though ;it is smaller than the total size of the memory being loaded. -;CHECK: vld4.32 {d16[1], d17[1], d18[1], d19[1]}, [r0, :64] +;CHECK: vld4.32 {d16[1], d17[1], d18[1], d19[1]}, [{{r[0-9]+}}, :64] %tmp0 = bitcast i32* %A to i8* %tmp1 = load <2 x i32>* %B %tmp2 = call %struct.__neon_int32x2x4_t @llvm.arm.neon.vld4lane.v2i32(i8* %tmp0, <2 x i32> %tmp1, <2 x i32> %tmp1, <2 x i32> %tmp1, <2 x i32> %tmp1, i32 1, i32 8) @@ -431,7 +431,7 @@ define <2 x float> @vld4lanef(float* %A, <2 x float>* %B) nounwind { define <8 x i16> @vld4laneQi16(i16* %A, <8 x i16>* %B) nounwind { ;CHECK: vld4laneQi16: ;Check the alignment value. Max for this instruction is 64 bits: -;CHECK: vld4.16 {d16[1], d18[1], d20[1], d22[1]}, [r0, :64] +;CHECK: vld4.16 {d16[1], d18[1], d20[1], d22[1]}, [{{r[0-9]+}}, :64] %tmp0 = bitcast i16* %A to i8* %tmp1 = load <8 x i16>* %B %tmp2 = call %struct.__neon_int16x8x4_t @llvm.arm.neon.vld4lane.v8i16(i8* %tmp0, <8 x i16> %tmp1, <8 x i16> %tmp1, <8 x i16> %tmp1, <8 x i16> %tmp1, i32 1, i32 16) @@ -448,7 +448,7 @@ define <8 x i16> @vld4laneQi16(i16* %A, <8 x i16>* %B) nounwind { define <4 x i32> @vld4laneQi32(i32* %A, <4 x i32>* %B) nounwind { ;CHECK: vld4laneQi32: ;Check the (default) alignment. -;CHECK: vld4.32 {d17[0], d19[0], d21[0], d23[0]}, [r0] +;CHECK: vld4.32 {d17[0], d19[0], d21[0], d23[0]}, [{{r[0-9]+}}] %tmp0 = bitcast i32* %A to i8* %tmp1 = load <4 x i32>* %B %tmp2 = call %struct.__neon_int32x4x4_t @llvm.arm.neon.vld4lane.v4i32(i8* %tmp0, <4 x i32> %tmp1, <4 x i32> %tmp1, <4 x i32> %tmp1, <4 x i32> %tmp1, i32 2, i32 1) diff --git a/test/CodeGen/ARM/vrev.ll b/test/CodeGen/ARM/vrev.ll index f0f9e4e..34acd16 100644 --- a/test/CodeGen/ARM/vrev.ll +++ b/test/CodeGen/ARM/vrev.ll @@ -147,3 +147,34 @@ define void @test_with_vcombine(<4 x float>* %v) nounwind { store <4 x float> %tmp8, <4 x float>* %v, align 16 ret void } + +; vrev <4 x i16> should use VREV32 and not VREV64 +define void @test_vrev64(<4 x i16>* nocapture %source, <2 x i16>* nocapture %dst) nounwind ssp { +; CHECK: test_vrev64: +; CHECK: vext.16 +; CHECK: vrev32.16 +entry: + %0 = bitcast <4 x i16>* %source to <8 x i16>* + %tmp2 = load <8 x i16>* %0, align 4 + %tmp3 = extractelement <8 x i16> %tmp2, i32 6 + %tmp5 = insertelement <2 x i16> undef, i16 %tmp3, i32 0 + %tmp9 = extractelement <8 x i16> %tmp2, i32 5 + %tmp11 = insertelement <2 x i16> %tmp5, i16 %tmp9, i32 1 + store <2 x i16> %tmp11, <2 x i16>* %dst, align 4 + ret void +} + +; Test vrev of float4 +define void @float_vrev64(float* nocapture %source, <4 x float>* nocapture %dest) nounwind noinline ssp { +; CHECK: float_vrev64 +; CHECK: vext.32 +; CHECK: vrev64.32 +entry: + %0 = bitcast float* %source to <4 x float>* + %tmp2 = load <4 x float>* %0, align 4 + %tmp5 = shufflevector <4 x float> <float 0.000000e+00, float undef, float undef, float undef>, <4 x float> %tmp2, <4 x i32> <i32 0, i32 7, i32 0, i32 0> + %arrayidx8 = getelementptr inbounds <4 x float>* %dest, i32 11 + store <4 x float> %tmp5, <4 x float>* %arrayidx8, align 4 + ret void +} + diff --git a/test/CodeGen/ARM/vstlane.ll b/test/CodeGen/ARM/vstlane.ll index d1bc15a..08b7232 100644 --- a/test/CodeGen/ARM/vstlane.ll +++ b/test/CodeGen/ARM/vstlane.ll @@ -54,7 +54,8 @@ define void @vst1lanef(float* %A, <2 x float>* %B) nounwind { define void @vst1laneQi8(i8* %A, <16 x i8>* %B) nounwind { ;CHECK: vst1laneQi8: -;CHECK: vst1.8 {d17[1]}, [r0] +; // Can use scalar load. No need to use vectors. +; // CHE-CK: vst1.8 {d17[1]}, [r0] %tmp1 = load <16 x i8>* %B %tmp2 = extractelement <16 x i8> %tmp1, i32 9 store i8 %tmp2, i8* %A, align 8 @@ -72,7 +73,8 @@ define void @vst1laneQi16(i16* %A, <8 x i16>* %B) nounwind { define void @vst1laneQi32(i32* %A, <4 x i32>* %B) nounwind { ;CHECK: vst1laneQi32: -;CHECK: vst1.32 {d17[1]}, [r0, :32] +; // Can use scalar load. No need to use vectors. +; // CHE-CK: vst1.32 {d17[1]}, [r0, :32] %tmp1 = load <4 x i32>* %B %tmp2 = extractelement <4 x i32> %tmp1, i32 3 store i32 %tmp2, i32* %A, align 8 @@ -82,7 +84,8 @@ define void @vst1laneQi32(i32* %A, <4 x i32>* %B) nounwind { ;Check for a post-increment updating store. define void @vst1laneQi32_update(i32** %ptr, <4 x i32>* %B) nounwind { ;CHECK: vst1laneQi32_update: -;CHECK: vst1.32 {d17[1]}, [r1, :32]! +; // Can use scalar load. No need to use vectors. +; // CHE-CK: vst1.32 {d17[1]}, [r1, :32]! %A = load i32** %ptr %tmp1 = load <4 x i32>* %B %tmp2 = extractelement <4 x i32> %tmp1, i32 3 @@ -94,7 +97,8 @@ define void @vst1laneQi32_update(i32** %ptr, <4 x i32>* %B) nounwind { define void @vst1laneQf(float* %A, <4 x float>* %B) nounwind { ;CHECK: vst1laneQf: -;CHECK: vst1.32 {d17[1]}, [r0] +; // Can use scalar load. No need to use vectors. +; // CHE-CK: vst1.32 {d17[1]}, [r0] %tmp1 = load <4 x float>* %B %tmp2 = extractelement <4 x float> %tmp1, i32 3 store float %tmp2, float* %A |
