diff options
Diffstat (limited to 'llvm/test')
| -rw-r--r-- | llvm/test/CodeGen/PowerPC/ppc64-anyregcc-crash.ll | 19 | ||||
| -rw-r--r-- | llvm/test/CodeGen/PowerPC/ppc64-anyregcc.ll | 367 | ||||
| -rw-r--r-- | llvm/test/CodeGen/PowerPC/ppc64-patchpoint.ll | 93 | ||||
| -rw-r--r-- | llvm/test/CodeGen/PowerPC/ppc64-stackmap-nops.ll | 24 | ||||
| -rw-r--r-- | llvm/test/CodeGen/PowerPC/ppc64-stackmap.ll | 289 |
5 files changed, 792 insertions, 0 deletions
diff --git a/llvm/test/CodeGen/PowerPC/ppc64-anyregcc-crash.ll b/llvm/test/CodeGen/PowerPC/ppc64-anyregcc-crash.ll new file mode 100644 index 00000000000..479c7a7af25 --- /dev/null +++ b/llvm/test/CodeGen/PowerPC/ppc64-anyregcc-crash.ll @@ -0,0 +1,19 @@ +; RUN: not llc < %s -mtriple=powerpc64-unknown-linux-gnu 2>&1 | FileCheck %s +; +; Check that misuse of anyregcc results in a compile time error. + +; CHECK: LLVM ERROR: ran out of registers during register allocation +define i64 @anyreglimit(i64 %v1, i64 %v2, i64 %v3, i64 %v4, i64 %v5, i64 %v6, i64 %v7, i64 %v8, + i64 %v9, i64 %v10, i64 %v11, i64 %v12, i64 %v13, i64 %v14, i64 %v15, i64 %v16, + i64 %v17, i64 %v18, i64 %v19, i64 %v20, i64 %v21, i64 %v22, i64 %v23, i64 %v24, + i64 %v25, i64 %v26, i64 %v27, i64 %v28, i64 %v29, i64 %v30, i64 %v31, i64 %v32) { +entry: + %result = tail call anyregcc i64 (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.i64(i64 12, i32 15, i8* inttoptr (i64 0 to i8*), i32 32, + i64 %v1, i64 %v2, i64 %v3, i64 %v4, i64 %v5, i64 %v6, i64 %v7, i64 %v8, + i64 %v9, i64 %v10, i64 %v11, i64 %v12, i64 %v13, i64 %v14, i64 %v15, i64 %v16, + i64 %v17, i64 %v18, i64 %v19, i64 %v20, i64 %v21, i64 %v22, i64 %v23, i64 %v24, + i64 %v25, i64 %v26, i64 %v27, i64 %v28, i64 %v29, i64 %v30, i64 %v31, i64 %v32) + ret i64 %result +} + +declare i64 @llvm.experimental.patchpoint.i64(i64, i32, i8*, i32, ...) diff --git a/llvm/test/CodeGen/PowerPC/ppc64-anyregcc.ll b/llvm/test/CodeGen/PowerPC/ppc64-anyregcc.ll new file mode 100644 index 00000000000..7cd3c4b3820 --- /dev/null +++ b/llvm/test/CodeGen/PowerPC/ppc64-anyregcc.ll @@ -0,0 +1,367 @@ +; RUN: llc < %s | FileCheck %s +target datalayout = "E-m:e-i64:64-n32:64" +target triple = "powerpc64-unknown-linux-gnu" + +; Stackmap Header: no constants - 6 callsites +; CHECK-LABEL: .section .llvm_stackmaps +; CHECK-NEXT: __LLVM_StackMaps: +; Header +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 0 +; CHECK-NEXT: .short 0 +; Num Functions +; CHECK-NEXT: .long 8 +; Num LargeConstants +; CHECK-NEXT: .long 0 +; Num Callsites +; CHECK-NEXT: .long 8 + +; Functions and stack size +; CHECK-NEXT: .quad test +; CHECK-NEXT: .quad 128 +; CHECK-NEXT: .quad property_access1 +; CHECK-NEXT: .quad 128 +; CHECK-NEXT: .quad property_access2 +; CHECK-NEXT: .quad 128 +; CHECK-NEXT: .quad property_access3 +; CHECK-NEXT: .quad 128 +; CHECK-NEXT: .quad anyreg_test1 +; CHECK-NEXT: .quad 160 +; CHECK-NEXT: .quad anyreg_test2 +; CHECK-NEXT: .quad 160 +; CHECK-NEXT: .quad patchpoint_spilldef +; CHECK-NEXT: .quad 256 +; CHECK-NEXT: .quad patchpoint_spillargs +; CHECK-NEXT: .quad 288 + + +; test +; CHECK-LABEL: .long .L{{.*}}-.L.test +; CHECK-NEXT: .short 0 +; 3 locations +; CHECK-NEXT: .short 3 +; Loc 0: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 4 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 1: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 4 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 2: Constant 3 +; CHECK-NEXT: .byte 4 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .long 3 +define i64 @test() nounwind ssp uwtable { +entry: + call anyregcc void (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.void(i64 0, i32 24, i8* null, i32 2, i32 1, i32 2, i64 3) + ret i64 0 +} + +; property access 1 - %obj is an anyreg call argument and should therefore be in a register +; CHECK-LABEL: .long .L{{.*}}-.L.property_access1 +; CHECK-NEXT: .short 0 +; 2 locations +; CHECK-NEXT: .short 2 +; Loc 0: Register <-- this is the return register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 1: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +define i64 @property_access1(i8* %obj) nounwind ssp uwtable { +entry: + %f = inttoptr i64 281474417671919 to i8* + %ret = call anyregcc i64 (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.i64(i64 1, i32 24, i8* %f, i32 1, i8* %obj) + ret i64 %ret +} + +; property access 2 - %obj is an anyreg call argument and should therefore be in a register +; CHECK-LABEL: .long .L{{.*}}-.L.property_access2 +; CHECK-NEXT: .short 0 +; 2 locations +; CHECK-NEXT: .short 2 +; Loc 0: Register <-- this is the return register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 1: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +define i64 @property_access2() nounwind ssp uwtable { +entry: + %obj = alloca i64, align 8 + %f = inttoptr i64 281474417671919 to i8* + %ret = call anyregcc i64 (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.i64(i64 2, i32 24, i8* %f, i32 1, i64* %obj) + ret i64 %ret +} + +; property access 3 - %obj is a frame index +; CHECK-LABEL: .long .L{{.*}}-.L.property_access3 +; CHECK-NEXT: .short 0 +; 2 locations +; CHECK-NEXT: .short 2 +; Loc 0: Register <-- this is the return register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 1: Direct FP - 8 +; CHECK-NEXT: .byte 2 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short 31 +; CHECK-NEXT: .long 112 +define i64 @property_access3() nounwind ssp uwtable { +entry: + %obj = alloca i64, align 8 + %f = inttoptr i64 281474417671919 to i8* + %ret = call anyregcc i64 (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.i64(i64 3, i32 24, i8* %f, i32 0, i64* %obj) + ret i64 %ret +} + +; anyreg_test1 +; CHECK-LABEL: .long .L{{.*}}-.L.anyreg_test1 +; CHECK-NEXT: .short 0 +; 14 locations +; CHECK-NEXT: .short 14 +; Loc 0: Register <-- this is the return register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 1: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 2: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 3: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 4: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 5: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 6: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 7: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 8: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 9: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 10: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 11: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 12: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 13: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +define i64 @anyreg_test1(i8* %a1, i8* %a2, i8* %a3, i8* %a4, i8* %a5, i8* %a6, i8* %a7, i8* %a8, i8* %a9, i8* %a10, i8* %a11, i8* %a12, i8* %a13) nounwind ssp uwtable { +entry: + %f = inttoptr i64 281474417671919 to i8* + %ret = call anyregcc i64 (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.i64(i64 4, i32 24, i8* %f, i32 13, i8* %a1, i8* %a2, i8* %a3, i8* %a4, i8* %a5, i8* %a6, i8* %a7, i8* %a8, i8* %a9, i8* %a10, i8* %a11, i8* %a12, i8* %a13) + ret i64 %ret +} + +; anyreg_test2 +; CHECK-LABEL: .long .L{{.*}}-.L.anyreg_test2 +; CHECK-NEXT: .short 0 +; 14 locations +; CHECK-NEXT: .short 14 +; Loc 0: Register <-- this is the return register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 1: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 2: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 3: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 4: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 5: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 6: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 7: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 8: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 9: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 10: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 11: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 12: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 13: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +define i64 @anyreg_test2(i8* %a1, i8* %a2, i8* %a3, i8* %a4, i8* %a5, i8* %a6, i8* %a7, i8* %a8, i8* %a9, i8* %a10, i8* %a11, i8* %a12, i8* %a13) nounwind ssp uwtable { +entry: + %f = inttoptr i64 281474417671919 to i8* + %ret = call anyregcc i64 (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.i64(i64 5, i32 24, i8* %f, i32 8, i8* %a1, i8* %a2, i8* %a3, i8* %a4, i8* %a5, i8* %a6, i8* %a7, i8* %a8, i8* %a9, i8* %a10, i8* %a11, i8* %a12, i8* %a13) + ret i64 %ret +} + +; Test spilling the return value of an anyregcc call. +; +; <rdar://problem/15432754> [JS] Assertion: "Folded a def to a non-store!" +; +; CHECK-LABEL: .long .L{{.*}}-.L.patchpoint_spilldef +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .short 3 +; Loc 0: Register (some register that will be spilled to the stack) +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 1: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 1: Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +define i64 @patchpoint_spilldef(i64 %p1, i64 %p2, i64 %p3, i64 %p4) { +entry: + %result = tail call anyregcc i64 (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.i64(i64 12, i32 24, i8* inttoptr (i64 0 to i8*), i32 2, i64 %p1, i64 %p2) + tail call void asm sideeffect "nop", "~{r0},~{r3},~{r4},~{r5},~{r6},~{r7},~{r8},~{r9},~{r10},~{r11},~{r12},~{r14},~{r15},~{r16},~{r17 +},~{r18},~{r19},~{r20},~{r21},~{r22},~{r23},~{r24},~{r25},~{r26},~{r27},~{r28},~{r29},~{r30},~{r31}"() nounwind + ret i64 %result +} + +; Test spilling the arguments of an anyregcc call. +; +; <rdar://problem/15487687> [JS] AnyRegCC argument ends up being spilled +; +; CHECK-LABEL: .long .L{{.*}}-.L.patchpoint_spillargs +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .short 5 +; Loc 0: Return a register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 1: Arg0 in a Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 2: Arg1 in a Register +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; Loc 3: Arg2 spilled to FP -96 +; CHECK-NEXT: .byte 3 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short 31 +; CHECK-NEXT: .long 128 +; Loc 4: Arg3 spilled to FP - 88 +; CHECK-NEXT: .byte 3 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short 31 +; CHECK-NEXT: .long 136 +define i64 @patchpoint_spillargs(i64 %p1, i64 %p2, i64 %p3, i64 %p4) { +entry: + tail call void asm sideeffect "nop", "~{r0},~{r3},~{r4},~{r5},~{r6},~{r7},~{r8},~{r9},~{r10},~{r11},~{r12},~{r14},~{r15},~{r16},~{r17 +},~{r18},~{r19},~{r20},~{r21},~{r22},~{r23},~{r24},~{r25},~{r26},~{r27},~{r28},~{r29},~{r30},~{r31}"() nounwind + %result = tail call anyregcc i64 (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.i64(i64 13, i32 24, i8* inttoptr (i64 0 to i8*), i32 2, i64 %p1, i64 %p2, i64 %p3, i64 %p4) + ret i64 %result +} + +declare void @llvm.experimental.patchpoint.void(i64, i32, i8*, i32, ...) +declare i64 @llvm.experimental.patchpoint.i64(i64, i32, i8*, i32, ...) diff --git a/llvm/test/CodeGen/PowerPC/ppc64-patchpoint.ll b/llvm/test/CodeGen/PowerPC/ppc64-patchpoint.ll new file mode 100644 index 00000000000..5e58fdab216 --- /dev/null +++ b/llvm/test/CodeGen/PowerPC/ppc64-patchpoint.ll @@ -0,0 +1,93 @@ +; RUN: llc < %s | FileCheck %s +; RUN: llc -fast-isel -fast-isel-abort < %s | FileCheck %s +target datalayout = "E-m:e-i64:64-n32:64" +target triple = "powerpc64-unknown-linux-gnu" + +; Trivial patchpoint codegen +; +define i64 @trivial_patchpoint_codegen(i64 %p1, i64 %p2, i64 %p3, i64 %p4) { +entry: +; CHECK-LABEL: trivial_patchpoint_codegen: + +; CHECK: li 11, -8531 +; CHECK-NEXT: rldic 11, 11, 32, 16 +; CHECK-NEXT: oris 11, 11, 48879 +; CHECK-NEXT: ori 11, 11, 51966 +; CHECK-NEXT: mtctr 11 +; CHECK-NEXT: bctrl + +; CHECK: li 11, -8531 +; CHECK-NEXT: rldic 11, 11, 32, 16 +; CHECK-NEXT: oris 11, 11, 48879 +; CHECK-NEXT: ori 11, 11, 51967 +; CHECK-NEXT: mtctr 11 +; CHECK-NEXT: bctrl + +; CHECK: blr + + %resolveCall2 = inttoptr i64 244837814094590 to i8* + %result = tail call i64 (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.i64(i64 2, i32 24, i8* %resolveCall2, i32 4, i64 %p1, i64 %p2, i64 %p3, i64 %p4) + %resolveCall3 = inttoptr i64 244837814094591 to i8* + tail call void (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.void(i64 3, i32 24, i8* %resolveCall3, i32 2, i64 %p1, i64 %result) + ret i64 %result +} + +; Caller frame metadata with stackmaps. This should not be optimized +; as a leaf function. +; +; CHECK-LABEL: caller_meta_leaf +; CHECK: stdu 1, -80(1) +; CHECK: Ltmp +; CHECK: addi 1, 1, 80 +; CHECK: blr + +define void @caller_meta_leaf() { +entry: + %metadata = alloca i64, i32 3, align 8 + store i64 11, i64* %metadata + store i64 12, i64* %metadata + store i64 13, i64* %metadata + call void (i64, i32, ...)* @llvm.experimental.stackmap(i64 4, i32 0, i64* %metadata) + ret void +} + +; Test patchpoints reusing the same TargetConstant. +; <rdar:15390785> Assertion failed: (CI.getNumArgOperands() >= NumArgs + 4) +; There is no way to verify this, since it depends on memory allocation. +; But I think it's useful to include as a working example. +define i64 @testLowerConstant(i64 %arg, i64 %tmp2, i64 %tmp10, i64* %tmp33, i64 %tmp79) { +entry: + %tmp80 = add i64 %tmp79, -16 + %tmp81 = inttoptr i64 %tmp80 to i64* + %tmp82 = load i64* %tmp81, align 8 + tail call void (i64, i32, ...)* @llvm.experimental.stackmap(i64 14, i32 8, i64 %arg, i64 %tmp2, i64 %tmp10, i64 %tmp82) + tail call void (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.void(i64 15, i32 32, i8* null, i32 3, i64 %arg, i64 %tmp10, i64 %tmp82) + %tmp83 = load i64* %tmp33, align 8 + %tmp84 = add i64 %tmp83, -24 + %tmp85 = inttoptr i64 %tmp84 to i64* + %tmp86 = load i64* %tmp85, align 8 + tail call void (i64, i32, ...)* @llvm.experimental.stackmap(i64 17, i32 8, i64 %arg, i64 %tmp10, i64 %tmp86) + tail call void (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.void(i64 18, i32 32, i8* null, i32 3, i64 %arg, i64 %tmp10, i64 %tmp86) + ret i64 10 +} + +; Test small patchpoints that don't emit calls. +define void @small_patchpoint_codegen(i64 %p1, i64 %p2, i64 %p3, i64 %p4) { +entry: +; CHECK-LABEL: small_patchpoint_codegen: +; CHECK: Ltmp +; CHECK: nop +; CHECK-NEXT: nop +; CHECK-NEXT: nop +; CHECK-NEXT: nop +; CHECK-NEXT: nop +; CHECK-NOT: nop +; CHECK: blr + %result = tail call i64 (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.i64(i64 5, i32 20, i8* null, i32 2, i64 %p1, i64 %p2) + ret void +} + +declare void @llvm.experimental.stackmap(i64, i32, ...) +declare void @llvm.experimental.patchpoint.void(i64, i32, i8*, i32, ...) +declare i64 @llvm.experimental.patchpoint.i64(i64, i32, i8*, i32, ...) + diff --git a/llvm/test/CodeGen/PowerPC/ppc64-stackmap-nops.ll b/llvm/test/CodeGen/PowerPC/ppc64-stackmap-nops.ll new file mode 100644 index 00000000000..368ddc5c833 --- /dev/null +++ b/llvm/test/CodeGen/PowerPC/ppc64-stackmap-nops.ll @@ -0,0 +1,24 @@ +; RUN: llc < %s -mtriple=powerpc64-unknown-gnu-linux | FileCheck %s + +define void @test_shadow_optimization() { +entry: +; Expect 12 bytes worth of nops here rather than 32: With the shadow optimization +; in place, 20 bytes will be consumed by the frame teardown and return instr. +; CHECK-LABEL: test_shadow_optimization: + +; CHECK: nop +; CHECK-NEXT: nop +; CHECK-NEXT: nop +; CHECK-NOT: nop +; CHECK: addi 1, 1, 64 +; CHECK: ld [[REG1:[0-9]+]], 16(1) +; CHECK: ld 31, -8(1) +; CHECK: mtlr [[REG1]] +; CHECK: blr + + tail call void (i64, i32, ...)* @llvm.experimental.stackmap(i64 0, i32 32) + ret void +} + +declare void @llvm.experimental.stackmap(i64, i32, ...) + diff --git a/llvm/test/CodeGen/PowerPC/ppc64-stackmap.ll b/llvm/test/CodeGen/PowerPC/ppc64-stackmap.ll new file mode 100644 index 00000000000..9be8d0c8ad4 --- /dev/null +++ b/llvm/test/CodeGen/PowerPC/ppc64-stackmap.ll @@ -0,0 +1,289 @@ +; RUN: llc < %s | FileCheck %s +; +; Note: Print verbose stackmaps using -debug-only=stackmaps. + +; We are not getting the correct stack alignment when cross compiling for arm64. +; So specify a datalayout here. +target datalayout = "E-m:e-i64:64-n32:64" +target triple = "powerpc64-unknown-linux-gnu" + +; CHECK-LABEL: .section .llvm_stackmaps +; CHECK-NEXT: __LLVM_StackMaps: +; Header +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 0 +; CHECK-NEXT: .short 0 +; Num Functions +; CHECK-NEXT: .long 11 +; Num LargeConstants +; CHECK-NEXT: .long 2 +; Num Callsites +; CHECK-NEXT: .long 11 + +; Functions and stack size +; CHECK-NEXT: .quad constantargs +; CHECK-NEXT: .quad 128 +; CHECK-NEXT: .quad osrinline +; CHECK-NEXT: .quad 144 +; CHECK-NEXT: .quad osrcold +; CHECK-NEXT: .quad 128 +; CHECK-NEXT: .quad propertyRead +; CHECK-NEXT: .quad 128 +; CHECK-NEXT: .quad propertyWrite +; CHECK-NEXT: .quad 128 +; CHECK-NEXT: .quad jsVoidCall +; CHECK-NEXT: .quad 128 +; CHECK-NEXT: .quad jsIntCall +; CHECK-NEXT: .quad 128 +; CHECK-NEXT: .quad spilledValue +; CHECK-NEXT: .quad 320 +; CHECK-NEXT: .quad spilledStackMapValue +; CHECK-NEXT: .quad 224 +; CHECK-NEXT: .quad liveConstant +; CHECK-NEXT: .quad 64 +; CHECK-NEXT: .quad clobberLR +; CHECK-NEXT: .quad 208 + +; Num LargeConstants +; CHECK-NEXT: .quad 4294967295 +; CHECK-NEXT: .quad 4294967296 + +; Constant arguments +; +; CHECK-NEXT: .quad 1 +; CHECK-NEXT: .long .L{{.*}}-.L.constantargs +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .short 4 +; SmallConstant +; CHECK-NEXT: .byte 4 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .long 65535 +; SmallConstant +; CHECK-NEXT: .byte 4 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .long 65536 +; SmallConstant +; CHECK-NEXT: .byte 5 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .long 0 +; LargeConstant at index 0 +; CHECK-NEXT: .byte 5 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .long 1 + +define void @constantargs() { +entry: + %0 = inttoptr i64 244837814094590 to i8* + tail call void (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.void(i64 1, i32 24, i8* %0, i32 0, i64 65535, i64 65536, i64 4294967295, i64 4294967296) + ret void +} + +; Inline OSR Exit +; +; CHECK-LABEL: .long .L{{.*}}-.L.osrinline +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .short 2 +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +define void @osrinline(i64 %a, i64 %b) { +entry: + ; Runtime void->void call. + call void inttoptr (i64 244837814094590 to void ()*)() + ; Followed by inline OSR patchpoint with 12-byte shadow and 2 live vars. + call void (i64, i32, ...)* @llvm.experimental.stackmap(i64 3, i32 12, i64 %a, i64 %b) + ret void +} + +; Cold OSR Exit +; +; 2 live variables in register. +; +; CHECK-LABEL: .long .L{{.*}}-.L.osrcold +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .short 2 +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +define void @osrcold(i64 %a, i64 %b) { +entry: + %test = icmp slt i64 %a, %b + br i1 %test, label %ret, label %cold +cold: + ; OSR patchpoint with 12-byte nop-slide and 2 live vars. + %thunk = inttoptr i64 244837814094590 to i8* + call void (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.void(i64 4, i32 24, i8* %thunk, i32 0, i64 %a, i64 %b) + unreachable +ret: + ret void +} + +; Property Read +; CHECK-LABEL: .long .L{{.*}}-.L.propertyRead +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .short 0 +; +; FIXME: There are currently no stackmap entries. After moving to +; AnyRegCC, we will have entries for the object and return value. +define i64 @propertyRead(i64* %obj) { +entry: + %resolveRead = inttoptr i64 244837814094590 to i8* + %result = call i64 (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.i64(i64 5, i32 24, i8* %resolveRead, i32 1, i64* %obj) + %add = add i64 %result, 3 + ret i64 %add +} + +; Property Write +; CHECK-LABEL: .long .L{{.*}}-.L.propertyWrite +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .short 2 +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +define void @propertyWrite(i64 %dummy1, i64* %obj, i64 %dummy2, i64 %a) { +entry: + %resolveWrite = inttoptr i64 244837814094590 to i8* + call anyregcc void (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.void(i64 6, i32 24, i8* %resolveWrite, i32 2, i64* %obj, i64 %a) + ret void +} + +; Void JS Call +; +; 2 live variables in registers. +; +; CHECK-LABEL: .long .L{{.*}}-.L.jsVoidCall +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .short 2 +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +define void @jsVoidCall(i64 %dummy1, i64* %obj, i64 %arg, i64 %l1, i64 %l2) { +entry: + %resolveCall = inttoptr i64 244837814094590 to i8* + call void (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.void(i64 7, i32 24, i8* %resolveCall, i32 2, i64* %obj, i64 %arg, i64 %l1, i64 %l2) + ret void +} + +; i64 JS Call +; +; 2 live variables in registers. +; +; CHECK-LABEL: .long .L{{.*}}-.L.jsIntCall +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .short 2 +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +; CHECK-NEXT: .byte 1 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short {{[0-9]+}} +; CHECK-NEXT: .long 0 +define i64 @jsIntCall(i64 %dummy1, i64* %obj, i64 %arg, i64 %l1, i64 %l2) { +entry: + %resolveCall = inttoptr i64 244837814094590 to i8* + %result = call i64 (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.i64(i64 8, i32 24, i8* %resolveCall, i32 2, i64* %obj, i64 %arg, i64 %l1, i64 %l2) + %add = add i64 %result, 3 + ret i64 %add +} + +; Spilled stack map values. +; +; Verify 28 stack map entries. +; +; CHECK-LABEL: .long .L{{.*}}-.L.spilledValue +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .short 28 +; +; Check that at least one is a spilled entry from r31. +; Location: Indirect FP + ... +; CHECK: .byte 3 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short 31 +define void @spilledValue(i64 %arg0, i64 %arg1, i64 %arg2, i64 %arg3, i64 %arg4, i64 %l0, i64 %l1, i64 %l2, i64 %l3, i64 %l4, i64 %l5, i64 %l6, i64 %l7, i64 %l8, i64 %l9, i64 %l10, i64 %l11, i64 %l12, i64 %l13, i64 %l14, i64 %l15, i64 %l16, i64 %l17, i64 %l18, i64 %l19, i64 %l20, i64 %l21, i64 %l22, i64 %l23, i64 %l24, i64 %l25, i64 %l26, i64 %l27) { +entry: + call void (i64, i32, i8*, i32, ...)* @llvm.experimental.patchpoint.void(i64 11, i32 24, i8* null, i32 5, i64 %arg0, i64 %arg1, i64 %arg2, i64 %arg3, i64 %arg4, i64 %l0, i64 %l1, i64 %l2, i64 %l3, i64 %l4, i64 %l5, i64 %l6, i64 %l7, i64 %l8, i64 %l9, i64 %l10, i64 %l11, i64 %l12, i64 %l13, i64 %l14, i64 %l15, i64 %l16, i64 %l17, i64 %l18, i64 %l19, i64 %l20, i64 %l21, i64 %l22, i64 %l23, i64 %l24, i64 %l25, i64 %l26, i64 %l27) + ret void +} + +; Spilled stack map values. +; +; Verify 30 stack map entries. +; +; CHECK-LABEL: .long .L{{.*}}-.L.spilledStackMapValue +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .short 30 +; +; Check that at least one is a spilled entry from r31. +; Location: Indirect FP + ... +; CHECK: .byte 3 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short 31 +define webkit_jscc void @spilledStackMapValue(i64 %l0, i64 %l1, i64 %l2, i64 %l3, i64 %l4, i64 %l5, i64 %l6, i64 %l7, i64 %l8, i64 %l9, i64 %l10, i64 %l11, i64 %l12, i64 %l13, i64 %l14, i64 %l15, i64 %l16, i64 %l17, i64 %l18, i64 %l19, i64 %l20, i64 %l21, i64 %l22, i64 %l23, i64 %l24, i64 %l25, i64 %l26, i64 %l27, i64 %l28, i64 %l29) { +entry: + call void (i64, i32, ...)* @llvm.experimental.stackmap(i64 12, i32 16, i64 %l0, i64 %l1, i64 %l2, i64 %l3, i64 %l4, i64 %l5, i64 %l6, i64 %l7, i64 %l8, i64 %l9, i64 %l10, i64 %l11, i64 %l12, i64 %l13, i64 %l14, i64 %l15, i64 %l16, i64 %l17, i64 %l18, i64 %l19, i64 %l20, i64 %l21, i64 %l22, i64 %l23, i64 %l24, i64 %l25, i64 %l26, i64 %l27, i64 %l28, i64 %l29) + ret void +} + + +; Map a constant value. +; +; CHECK-LABEL: .long .L{{.*}}-.L.liveConstant +; CHECK-NEXT: .short 0 +; 1 location +; CHECK-NEXT: .short 1 +; Loc 0: SmallConstant +; CHECK-NEXT: .byte 4 +; CHECK-NEXT: .byte 8 +; CHECK-NEXT: .short 0 +; CHECK-NEXT: .long 33 + +define void @liveConstant() { + tail call void (i64, i32, ...)* @llvm.experimental.stackmap(i64 15, i32 8, i32 33) + ret void +} + +; Map a value when LR is the only free register. +; +; CHECK-LABEL: .long .L{{.*}}-.L.clobberLR +; CHECK-NEXT: .short 0 +; 1 location +; CHECK-NEXT: .short 1 +; Loc 0: Indirect FP (r31) - offset +; CHECK-NEXT: .byte 3 +; CHECK-NEXT: .byte 4 +; CHECK-NEXT: .short 31 +; CHECK-NEXT: .long {{[0-9]+}} +define void @clobberLR(i32 %a) { + tail call void asm sideeffect "nop", "~{r0},~{r3},~{r4},~{r5},~{r6},~{r7},~{r8},~{r9},~{r10},~{r11},~{r12},~{r14},~{r15},~{r16},~{r17},~{r18},~{r19},~{r20},~{r21},~{r22},~{r23},~{r24},~{r25},~{r26},~{r27},~{r28},~{r29},~{r30},~{r31}"() nounwind + tail call void (i64, i32, ...)* @llvm.experimental.stackmap(i64 16, i32 8, i32 %a) + ret void +} + +declare void @llvm.experimental.stackmap(i64, i32, ...) +declare void @llvm.experimental.patchpoint.void(i64, i32, i8*, i32, ...) +declare i64 @llvm.experimental.patchpoint.i64(i64, i32, i8*, i32, ...) |

