lei updated this revision to Diff 359050.
lei marked 3 inline comments as done.
lei added a comment.
Address review comments to add/upate:
- builtin encoding for params that need to be folded into constant expr
- llvm intrinsic property for immediates
- test line for `-mattr=-vsx`
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D105930/new/
https://reviews.llvm.org/D105930
Files:
clang/include/clang/Basic/BuiltinsPPC.def
clang/lib/Basic/Targets/PPC.cpp
clang/lib/Sema/SemaChecking.cpp
clang/test/CodeGen/builtins-ppc-xlcompat-error.c
clang/test/CodeGen/builtins-ppc-xlcompat-math.c
clang/test/CodeGen/builtins-ppc-xlcompat-pwr9-64bit.c
clang/test/CodeGen/builtins-ppc-xlcompat-pwr9.c
llvm/include/llvm/IR/IntrinsicsPowerPC.td
llvm/lib/Target/PowerPC/PPCInstrInfo.td
llvm/lib/Target/PowerPC/PPCInstrVSX.td
llvm/test/CodeGen/PowerPC/builtins-ppc-xlcompat-math.ll
llvm/test/CodeGen/PowerPC/builtins-ppc-xlcompat-pwr9-64bit.ll
llvm/test/CodeGen/PowerPC/builtins-ppc-xlcompat-pwr9.ll
Index: llvm/test/CodeGen/PowerPC/builtins-ppc-xlcompat-pwr9.ll
===================================================================
--- /dev/null
+++ llvm/test/CodeGen/PowerPC/builtins-ppc-xlcompat-pwr9.ll
@@ -0,0 +1,27 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
+; RUN: -mcpu=pwr9 < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
+; RUN: -mcpu=pwr9 < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-aix \
+; RUN: -mcpu=pwr9 < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc-unknown-aix \
+; RUN: -mcpu=pwr9 < %s | FileCheck %s --check-prefix=CHECK-32BIT
+
+define dso_local zeroext i32 @extract_exp(double %d) {
+; CHECK-LABEL: extract_exp:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xsxexpdp 3, 1
+; CHECK-NEXT: clrldi 3, 3, 32
+; CHECK-NEXT: blr
+;
+; CHECK-32BIT-LABEL: extract_exp:
+; CHECK-32BIT: # %bb.0: # %entry
+; CHECK-32BIT-NEXT: xsxexpdp 3, 1
+; CHECK-32BIT-NEXT: # kill: def $r3 killed $r3 killed $x3
+; CHECK-32BIT-NEXT: blr
+entry:
+ %0 = tail call i32 @llvm.ppc.extract.exp(double %d)
+ ret i32 %0
+}
+declare i32 @llvm.ppc.extract.exp(double)
Index: llvm/test/CodeGen/PowerPC/builtins-ppc-xlcompat-pwr9-64bit.ll
===================================================================
--- /dev/null
+++ llvm/test/CodeGen/PowerPC/builtins-ppc-xlcompat-pwr9-64bit.ll
@@ -0,0 +1,31 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
+; RUN: -mcpu=pwr9 < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
+; RUN: -mcpu=pwr9 < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-aix \
+; RUN: -mcpu=pwr9 < %s | FileCheck %s
+
+define dso_local i64 @extract_sig(double %d) {
+; CHECK-LABEL: extract_sig:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: xsxsigdp 3, 1
+; CHECK-NEXT: blr
+entry:
+ %0 = tail call i64 @llvm.ppc.extract.sig(double %d)
+ ret i64 %0
+}
+declare i64 @llvm.ppc.extract.sig(double)
+
+define dso_local double @insert_exp(double %d, i64 %ull) {
+; CHECK-LABEL: insert_exp:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: mffprd 3, 1
+; CHECK-NEXT: xsiexpdp 1, 3, 4
+; CHECK-NEXT: # kill: def $f1 killed $f1 killed $vsl1
+; CHECK-NEXT: blr
+entry:
+ %0 = tail call double @llvm.ppc.insert.exp(double %d, i64 %ull)
+ ret double %0
+}
+declare double @llvm.ppc.insert.exp(double, i64)
Index: llvm/test/CodeGen/PowerPC/builtins-ppc-xlcompat-math.ll
===================================================================
--- /dev/null
+++ llvm/test/CodeGen/PowerPC/builtins-ppc-xlcompat-math.ll
@@ -0,0 +1,231 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu < %s |\
+; RUN: FileCheck %s --check-prefix=CHECK-PWR8
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
+; RUN: -mattr=-vsx < %s | FileCheck %s --check-prefix=CHECK-NOVSX
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
+; RUN: -mcpu=pwr7 < %s | FileCheck %s --check-prefix=CHECK-PWR7
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-aix -mcpu=pwr7 \
+; RUN: < %s | FileCheck %s --check-prefix=CHECK-PWR7
+; RUN: llc -verify-machineinstrs -mtriple=powerpc-unknown-aix -mcpu=pwr8 < %s |\
+; RUN: FileCheck %s --check-prefix=CHECK-PWR8
+
+define dso_local void @mtfsb0() {
+; CHECK-PWR8-LABEL: mtfsb0:
+; CHECK-PWR8: # %bb.0: # %entry
+; CHECK-PWR8-NEXT: mtfsb0 10
+; CHECK-PWR8-NEXT: blr
+;
+; CHECK-NOVSX-LABEL: mtfsb0:
+; CHECK-NOVSX: # %bb.0: # %entry
+; CHECK-NOVSX-NEXT: mtfsb0 10
+; CHECK-NOVSX-NEXT: blr
+;
+; CHECK-PWR7-LABEL: mtfsb0:
+; CHECK-PWR7: # %bb.0: # %entry
+; CHECK-PWR7-NEXT: mtfsb0 10
+; CHECK-PWR7-NEXT: blr
+entry:
+ tail call void @llvm.ppc.mtfsb0(i32 10)
+ ret void
+}
+
+declare void @llvm.ppc.mtfsb0(i32 immarg) #1
+
+define dso_local void @mtfsb1() {
+; CHECK-PWR8-LABEL: mtfsb1:
+; CHECK-PWR8: # %bb.0: # %entry
+; CHECK-PWR8-NEXT: mtfsb1 0
+; CHECK-PWR8-NEXT: blr
+;
+; CHECK-NOVSX-LABEL: mtfsb1:
+; CHECK-NOVSX: # %bb.0: # %entry
+; CHECK-NOVSX-NEXT: mtfsb1 0
+; CHECK-NOVSX-NEXT: blr
+;
+; CHECK-PWR7-LABEL: mtfsb1:
+; CHECK-PWR7: # %bb.0: # %entry
+; CHECK-PWR7-NEXT: mtfsb1 0
+; CHECK-PWR7-NEXT: blr
+entry:
+ tail call void @llvm.ppc.mtfsb1(i32 0)
+ ret void
+}
+
+declare void @llvm.ppc.mtfsb1(i32 immarg) #1
+
+define dso_local double @fmsub_t0(double %d, double %d2, double %d3) {
+; CHECK-PWR8-LABEL: fmsub_t0:
+; CHECK-PWR8: # %bb.0: # %entry
+; CHECK-PWR8-NEXT: xsmsubmdp 1, 2, 3
+; CHECK-PWR8-NEXT: blr
+;
+; CHECK-NOVSX-LABEL: fmsub_t0:
+; CHECK-NOVSX: # %bb.0: # %entry
+; CHECK-NOVSX-NEXT: fmsub 1, 1, 2, 3
+; CHECK-NOVSX-NEXT: blr
+;
+; CHECK-PWR7-LABEL: fmsub_t0:
+; CHECK-PWR7: # %bb.0: # %entry
+; CHECK-PWR7-NEXT: xsmsubmdp 1, 2, 3
+; CHECK-PWR7-NEXT: blr
+entry:
+ %0 = tail call double @llvm.ppc.fmsub(double %d, double %d2, double %d3)
+ ret double %0
+}
+
+declare double @llvm.ppc.fmsub(double, double, double)
+
+define dso_local float @fmsubs_t0(float %f, float %f2, float %f3) {
+; CHECK-PWR8-LABEL: fmsubs_t0:
+; CHECK-PWR8: # %bb.0: # %entry
+; CHECK-PWR8-NEXT: xsmsubmsp 1, 2, 3
+; CHECK-PWR8-NEXT: blr
+;
+; CHECK-NOVSX-LABEL: fmsubs_t0:
+; CHECK-NOVSX: # %bb.0: # %entry
+; CHECK-NOVSX-NEXT: fmsubs 1, 1, 2, 3
+; CHECK-NOVSX-NEXT: blr
+;
+; CHECK-PWR7-LABEL: fmsubs_t0:
+; CHECK-PWR7: # %bb.0: # %entry
+; CHECK-PWR7-NEXT: fmsubs 1, 1, 2, 3
+; CHECK-PWR7-NEXT: blr
+entry:
+ %0 = tail call float @llvm.ppc.fmsubs(float %f, float %f2, float %f3)
+ ret float %0
+}
+
+declare float @llvm.ppc.fmsubs(float, float, float)
+
+define dso_local double @fnmadd_t0(double %d, double %d2, double %d3) {
+; CHECK-PWR8-LABEL: fnmadd_t0:
+; CHECK-PWR8: # %bb.0: # %entry
+; CHECK-PWR8-NEXT: xsnmaddmdp 1, 2, 3
+; CHECK-PWR8-NEXT: blr
+;
+; CHECK-NOVSX-LABEL: fnmadd_t0:
+; CHECK-NOVSX: # %bb.0: # %entry
+; CHECK-NOVSX-NEXT: fnmadd 1, 1, 2, 3
+; CHECK-NOVSX-NEXT: blr
+;
+; CHECK-PWR7-LABEL: fnmadd_t0:
+; CHECK-PWR7: # %bb.0: # %entry
+; CHECK-PWR7-NEXT: xsnmaddmdp 1, 2, 3
+; CHECK-PWR7-NEXT: blr
+entry:
+ %0 = tail call double @llvm.ppc.fnmadd(double %d, double %d2, double %d3)
+ ret double %0
+}
+
+declare double @llvm.ppc.fnmadd(double, double, double)
+
+define dso_local float @fnmadds_t0(float %f, float %f2, float %f3) {
+; CHECK-PWR8-LABEL: fnmadds_t0:
+; CHECK-PWR8: # %bb.0: # %entry
+; CHECK-PWR8-NEXT: xsnmaddmsp 1, 2, 3
+; CHECK-PWR8-NEXT: blr
+;
+; CHECK-NOVSX-LABEL: fnmadds_t0:
+; CHECK-NOVSX: # %bb.0: # %entry
+; CHECK-NOVSX-NEXT: fnmadds 1, 1, 2, 3
+; CHECK-NOVSX-NEXT: blr
+;
+; CHECK-PWR7-LABEL: fnmadds_t0:
+; CHECK-PWR7: # %bb.0: # %entry
+; CHECK-PWR7-NEXT: fnmadds 1, 1, 2, 3
+; CHECK-PWR7-NEXT: blr
+entry:
+ %0 = tail call float @llvm.ppc.fnmadds(float %f, float %f2, float %f3)
+ ret float %0
+}
+
+declare float @llvm.ppc.fnmadds(float, float, float)
+
+define dso_local double @fnmsub_t0(double %d, double %d2, double %d3) {
+; CHECK-PWR8-LABEL: fnmsub_t0:
+; CHECK-PWR8: # %bb.0: # %entry
+; CHECK-PWR8-NEXT: xsnmsubmdp 1, 2, 3
+; CHECK-PWR8-NEXT: blr
+;
+; CHECK-NOVSX-LABEL: fnmsub_t0:
+; CHECK-NOVSX: # %bb.0: # %entry
+; CHECK-NOVSX-NEXT: fnmsub 1, 1, 2, 3
+; CHECK-NOVSX-NEXT: blr
+;
+; CHECK-PWR7-LABEL: fnmsub_t0:
+; CHECK-PWR7: # %bb.0: # %entry
+; CHECK-PWR7-NEXT: xsnmsubmdp 1, 2, 3
+; CHECK-PWR7-NEXT: blr
+entry:
+ %0 = tail call double @llvm.ppc.fnmsub(double %d, double %d2, double %d3)
+ ret double %0
+}
+
+declare double @llvm.ppc.fnmsub(double, double, double)
+
+define dso_local float @fnmsubs_t0(float %f, float %f2, float %f3) {
+; CHECK-PWR8-LABEL: fnmsubs_t0:
+; CHECK-PWR8: # %bb.0: # %entry
+; CHECK-PWR8-NEXT: xsnmsubmsp 1, 2, 3
+; CHECK-PWR8-NEXT: blr
+;
+; CHECK-NOVSX-LABEL: fnmsubs_t0:
+; CHECK-NOVSX: # %bb.0: # %entry
+; CHECK-NOVSX-NEXT: fnmsubs 1, 1, 2, 3
+; CHECK-NOVSX-NEXT: blr
+;
+; CHECK-PWR7-LABEL: fnmsubs_t0:
+; CHECK-PWR7: # %bb.0: # %entry
+; CHECK-PWR7-NEXT: fnmsubs 1, 1, 2, 3
+; CHECK-PWR7-NEXT: blr
+entry:
+ %0 = tail call float @llvm.ppc.fnmsubs(float %f, float %f2, float %f3)
+ ret float %0
+}
+
+declare float @llvm.ppc.fnmsubs(float, float, float)
+
+define dso_local double @fre(double %d) {
+; CHECK-PWR8-LABEL: fre:
+; CHECK-PWR8: # %bb.0: # %entry
+; CHECK-PWR8-NEXT: xsredp 1, 1
+; CHECK-PWR8-NEXT: blr
+;
+; CHECK-NOVSX-LABEL: fre:
+; CHECK-NOVSX: # %bb.0: # %entry
+; CHECK-NOVSX-NEXT: fre 1, 1
+; CHECK-NOVSX-NEXT: blr
+;
+; CHECK-PWR7-LABEL: fre:
+; CHECK-PWR7: # %bb.0: # %entry
+; CHECK-PWR7-NEXT: xsredp 1, 1
+; CHECK-PWR7-NEXT: blr
+entry:
+ %0 = tail call double @llvm.ppc.fre(double %d)
+ ret double %0
+}
+
+declare double @llvm.ppc.fre(double)
+
+define dso_local float @fres(float %f) {
+; CHECK-PWR8-LABEL: fres:
+; CHECK-PWR8: # %bb.0: # %entry
+; CHECK-PWR8-NEXT: xsresp 1, 1
+; CHECK-PWR8-NEXT: blr
+;
+; CHECK-NOVSX-LABEL: fres:
+; CHECK-NOVSX: # %bb.0: # %entry
+; CHECK-NOVSX-NEXT: fres 1, 1
+; CHECK-NOVSX-NEXT: blr
+;
+; CHECK-PWR7-LABEL: fres:
+; CHECK-PWR7: # %bb.0: # %entry
+; CHECK-PWR7-NEXT: fres 1, 1
+; CHECK-PWR7-NEXT: blr
+entry:
+ %0 = tail call float @llvm.ppc.fres(float %f)
+ ret float %0
+}
+
+declare float @llvm.ppc.fres(float)
Index: llvm/lib/Target/PowerPC/PPCInstrVSX.td
===================================================================
--- llvm/lib/Target/PowerPC/PPCInstrVSX.td
+++ llvm/lib/Target/PowerPC/PPCInstrVSX.td
@@ -1539,10 +1539,10 @@
// Insert Exponent DP/QP
// XT NOTE: XT.dword[1] = 0xUUUU_UUUU_UUUU_UUUU
+ def XSIEXPDP : XX1Form <60, 918, (outs vsrc:$XT), (ins g8rc:$rA, g8rc:$rB),
+ "xsiexpdp $XT, $rA, $rB", IIC_VecFP, []>;
// FIXME: Setting the hasSideEffects flag here to match current behaviour.
let hasSideEffects = 1 in {
- def XSIEXPDP : XX1Form <60, 918, (outs vsrc:$XT), (ins g8rc:$rA, g8rc:$rB),
- "xsiexpdp $XT, $rA, $rB", IIC_VecFP, []>;
// vB NOTE: only vB.dword[0] is used, that's why we don't use
// X_VT5_VA5_VB5 form
def XSIEXPQP : XForm_18<63, 868, (outs vrrc:$vT), (ins vrrc:$vA, vsfrc:$vB),
@@ -1550,11 +1550,11 @@
}
// Extract Exponent/Significand DP/QP
+ def XSXEXPDP : XX2_RT5_XO5_XB6<60, 0, 347, "xsxexpdp", []>;
+ def XSXSIGDP : XX2_RT5_XO5_XB6<60, 1, 347, "xsxsigdp", []>;
+
// FIXME: Setting the hasSideEffects flag here to match current behaviour.
let hasSideEffects = 1 in {
- def XSXEXPDP : XX2_RT5_XO5_XB6<60, 0, 347, "xsxexpdp", []>;
- def XSXSIGDP : XX2_RT5_XO5_XB6<60, 1, 347, "xsxsigdp", []>;
-
def XSXEXPQP : X_VT5_XO5_VB5 <63, 2, 804, "xsxexpqp", []>;
def XSXSIGQP : X_VT5_XO5_VB5 <63, 18, 804, "xsxsigqp", []>;
}
@@ -2850,6 +2850,12 @@
def : Pat<(v2i64 (PPCvcmp_rec v2i64:$vA, v2i64:$vB, 199)),
(VCMPGTUB_rec DblwdCmp.MRGEQ, (v2i64 (XXLXORz)))>;
} // AddedComplexity = 0
+
+// XL Compat builtins.
+def : Pat<(int_ppc_fmsub f64:$A, f64:$B, f64:$C), (XSMSUBMDP $A, $B, $C)>;
+def : Pat<(int_ppc_fnmsub f64:$A, f64:$B, f64:$C), (XSNMSUBMDP $A, $B, $C)>;
+def : Pat<(int_ppc_fnmadd f64:$A, f64:$B, f64:$C), (XSNMADDMDP $A, $B, $C)>;
+def : Pat<(int_ppc_fre f64:$A), (XSREDP $A)>;
} // HasVSX
// Any big endian VSX subtarget.
@@ -3240,6 +3246,18 @@
(v8i16 (COPY_TO_REGCLASS(XXLEQVOnes), VSRC))>;
def : Pat<(v16i8 (bitconvert (v16i8 immAllOnesV))),
(v16i8 (COPY_TO_REGCLASS(XXLEQVOnes), VSRC))>;
+
+// XL Compat builtins.
+def : Pat<(int_ppc_fmsubs f32:$A, f32:$B, f32:$C), (XSMSUBMSP $A, $B, $C)>;
+def : Pat<(int_ppc_fnmsubs f32:$A, f32:$B, f32:$C), (XSNMSUBMSP $A, $B, $C)>;
+def : Pat<(int_ppc_fnmadds f32:$A, f32:$B, f32:$C), (XSNMADDMSP $A, $B, $C)>;
+def : Pat<(int_ppc_fres f32:$A), (XSRESP $A)>;
+def : Pat<(i32 (int_ppc_extract_exp f64:$A)),
+ (EXTRACT_SUBREG (XSXEXPDP (COPY_TO_REGCLASS $A, VSFRC)), sub_32)>;
+def : Pat<(int_ppc_extract_sig f64:$A),
+ (XSXSIGDP (COPY_TO_REGCLASS $A, VSFRC))>;
+def : Pat<(f64 (int_ppc_insert_exp f64:$A, i64:$B)),
+ (COPY_TO_REGCLASS (XSIEXPDP (COPY_TO_REGCLASS $A, G8RC), $B), F8RC)>;
} // HasVSX, HasP8Vector
// Any big endian Power8 VSX subtarget.
Index: llvm/lib/Target/PowerPC/PPCInstrInfo.td
===================================================================
--- llvm/lib/Target/PowerPC/PPCInstrInfo.td
+++ llvm/lib/Target/PowerPC/PPCInstrInfo.td
@@ -3084,12 +3084,16 @@
// When FM is 30/31, we are setting the 62/63 bit of FPSCR, the implicit-def
// RM should be set.
+let hasSideEffects = 1 in {
def MTFSB0 : XForm_43<63, 70, (outs), (ins u5imm:$FM),
- "mtfsb0 $FM", IIC_IntMTFSB0, []>,
+ "mtfsb0 $FM", IIC_IntMTFSB0,
+ [(int_ppc_mtfsb0 timm:$FM)]>,
PPC970_DGroup_Single, PPC970_Unit_FPU;
def MTFSB1 : XForm_43<63, 38, (outs), (ins u5imm:$FM),
- "mtfsb1 $FM", IIC_IntMTFSB0, []>,
+ "mtfsb1 $FM", IIC_IntMTFSB0,
+ [(int_ppc_mtfsb1 timm:$FM)]>,
PPC970_DGroup_Single, PPC970_Unit_FPU;
+}
let Defs = [RM] in {
let isCodeGenOnly = 1 in
@@ -3642,6 +3646,16 @@
(FCPSGNS (COPY_TO_REGCLASS $frA, F4RC), $frB)>;
}
+// XL Compat intrinsics.
+def : Pat<(int_ppc_fmsub f64:$A, f64:$B, f64:$C), (FMSUB $A, $B, $C)>;
+def : Pat<(int_ppc_fmsubs f32:$A, f32:$B, f32:$C), (FMSUBS $A, $B, $C)>;
+def : Pat<(int_ppc_fnmsub f64:$A, f64:$B, f64:$C), (FNMSUB $A, $B, $C)>;
+def : Pat<(int_ppc_fnmsubs f32:$A, f32:$B, f32:$C), (FNMSUBS $A, $B, $C)>;
+def : Pat<(int_ppc_fnmadd f64:$A, f64:$B, f64:$C), (FNMADD $A, $B, $C)>;
+def : Pat<(int_ppc_fnmadds f32:$A, f32:$B, f32:$C), (FNMADDS $A, $B, $C)>;
+def : Pat<(int_ppc_fre f64:$A), (FRE $A)>;
+def : Pat<(int_ppc_fres f32:$A), (FRES $A)>;
+
include "PPCInstrAltivec.td"
include "PPCInstrSPE.td"
include "PPCInstr64Bit.td"
Index: llvm/include/llvm/IR/IntrinsicsPowerPC.td
===================================================================
--- llvm/include/llvm/IR/IntrinsicsPowerPC.td
+++ llvm/include/llvm/IR/IntrinsicsPowerPC.td
@@ -1565,5 +1565,67 @@
def int_ppc_stwcx : GCCBuiltin<"__builtin_ppc_stwcx">,
Intrinsic<[llvm_i32_ty], [llvm_ptr_ty, llvm_i32_ty],
[IntrWriteMem]>;
+ def int_ppc_insert_exp
+ : GCCBuiltin<"__builtin_ppc_insert_exp">,
+ Intrinsic <[llvm_double_ty], [llvm_double_ty, llvm_i64_ty],
+ [IntrNoMem]>;
+ def int_ppc_extract_exp
+ : GCCBuiltin<"__builtin_ppc_extract_exp">,
+ Intrinsic <[llvm_i32_ty], [llvm_double_ty], [IntrNoMem]>;
+ def int_ppc_extract_sig
+ : GCCBuiltin<"__builtin_ppc_extract_sig">,
+ Intrinsic <[llvm_i64_ty], [llvm_double_ty], [IntrNoMem]>;
+ def int_ppc_mtfsb0
+ : GCCBuiltin<"__builtin_ppc_mtfsb0">,
+ Intrinsic <[], [llvm_i32_ty],
+ [IntrNoMem, IntrHasSideEffects, ImmArg<ArgIndex<0>>]>;
+ def int_ppc_mtfsb1
+ : GCCBuiltin<"__builtin_ppc_mtfsb1">,
+ Intrinsic <[], [llvm_i32_ty],
+ [IntrNoMem, IntrHasSideEffects, ImmArg<ArgIndex<0>>]>;
+ def int_ppc_mtfsf
+ : GCCBuiltin<"__builtin_ppc_mtfsf">,
+ Intrinsic <[], [llvm_i32_ty, llvm_i32_ty],
+ [IntrNoMem, IntrHasSideEffects, ImmArg<ArgIndex<0>>]>;
+ def int_ppc_mtfsfi
+ : GCCBuiltin<"__builtin_ppc_mtfsfi">,
+ Intrinsic <[], [llvm_i32_ty, llvm_i32_ty],
+ [IntrNoMem, IntrHasSideEffects,
+ ImmArg<ArgIndex<0>>,ImmArg<ArgIndex<1>>]>;
+ def int_ppc_fmsub
+ : GCCBuiltin<"__builtin_ppc_fmsub">,
+ Intrinsic <[llvm_double_ty],
+ [llvm_double_ty, llvm_double_ty, llvm_double_ty],
+ [IntrNoMem]>;
+ def int_ppc_fmsubs
+ : GCCBuiltin<"__builtin_ppc_fmsubs">,
+ Intrinsic <[llvm_float_ty],
+ [llvm_float_ty, llvm_float_ty, llvm_float_ty],
+ [IntrNoMem]>;
+ def int_ppc_fnmadd
+ : GCCBuiltin<"__builtin_ppc_fnmadd">,
+ Intrinsic <[llvm_double_ty],
+ [llvm_double_ty, llvm_double_ty, llvm_double_ty],
+ [IntrNoMem]>;
+ def int_ppc_fnmadds
+ : GCCBuiltin<"__builtin_ppc_fnmadds">,
+ Intrinsic <[llvm_float_ty],
+ [llvm_float_ty, llvm_float_ty, llvm_float_ty],
+ [IntrNoMem]>;
+ def int_ppc_fnmsub
+ : GCCBuiltin<"__builtin_ppc_fnmsub">,
+ Intrinsic <[llvm_double_ty],
+ [llvm_double_ty, llvm_double_ty, llvm_double_ty],
+ [IntrNoMem]>;
+ def int_ppc_fnmsubs
+ : GCCBuiltin<"__builtin_ppc_fnmsubs">,
+ Intrinsic <[llvm_float_ty],
+ [llvm_float_ty, llvm_float_ty, llvm_float_ty],
+ [IntrNoMem]>;
+ def int_ppc_fre
+ : GCCBuiltin<"__builtin_ppc_fre">,
+ Intrinsic <[llvm_double_ty], [llvm_double_ty], [IntrNoMem]>;
+ def int_ppc_fres
+ : GCCBuiltin<"__builtin_ppc_fres">,
+ Intrinsic <[llvm_float_ty], [llvm_float_ty], [IntrNoMem]>;
}
-
Index: clang/test/CodeGen/builtins-ppc-xlcompat-pwr9.c
===================================================================
--- /dev/null
+++ clang/test/CodeGen/builtins-ppc-xlcompat-pwr9.c
@@ -0,0 +1,18 @@
+// RUN: %clang_cc1 -triple powerpc64-unknown-unknown -emit-llvm %s \
+// RUN: -target-cpu pwr9 -o - | FileCheck %s
+// RUN: %clang_cc1 -triple powerpc64le-unknown-unknown -emit-llvm %s \
+// RUN: -target-cpu pwr9 -o - | FileCheck %s
+// RUN: %clang_cc1 -triple powerpc64-unknown-aix -emit-llvm %s \
+// RUN: -target-cpu pwr9 -o - | FileCheck %s
+// RUN: %clang_cc1 -triple powerpc-unknown-aix %s -emit-llvm %s \
+// RUN: -target-cpu pwr9 -o - | FileCheck %s
+// RUN: not %clang_cc1 -triple powerpc64-unknown-aix -emit-llvm %s \
+// RUN: -target-cpu pwr8 2>&1 | FileCheck %s --check-prefix=CHECK-NONPWR9-ERR
+
+unsigned int extract_exp (double d) {
+// CHECK-LABEL: @extract_exp
+// CHECK: [[TMP1:%.*]] = call i32 @llvm.ppc.extract.exp(double %0)
+// CHECK-NEXT: ret i32 [[TMP1]]
+// CHECK-NONPWR9-ERR: error: this builtin is only valid on POWER9 or later CPUs
+ return __extract_exp (d);
+}
Index: clang/test/CodeGen/builtins-ppc-xlcompat-pwr9-64bit.c
===================================================================
--- /dev/null
+++ clang/test/CodeGen/builtins-ppc-xlcompat-pwr9-64bit.c
@@ -0,0 +1,30 @@
+// RUN: %clang_cc1 -triple powerpc64-unknown-unknown -emit-llvm %s \
+// RUN: -target-cpu pwr9 -o - | FileCheck %s
+// RUN: %clang_cc1 -triple powerpc64le-unknown-unknown -emit-llvm %s \
+// RUN: -target-cpu pwr9 -o - | FileCheck %s
+// RUN: %clang_cc1 -triple powerpc64-unknown-aix -emit-llvm %s \
+// RUN: -target-cpu pwr9 -o - | FileCheck %s
+// RUN: not %clang_cc1 -triple powerpc-unknown-aix %s \
+// RUN: -target-cpu pwr9 2>&1 | FileCheck %s --check-prefix=CHECK-32-ERROR
+// RUN: not %clang_cc1 -triple powerpc64-unknown-aix -emit-llvm %s \
+// RUN: -target-cpu pwr8 2>&1 | FileCheck %s --check-prefix=CHECK-NONPWR9-ERR
+
+unsigned long long extract_sig (double d) {
+// CHECK-LABEL: @extract_sig(
+// CHECK: [[TMP1:%.*]] = call i64 @llvm.ppc.extract.sig(double %0)
+// CHECK-NEXT: ret i64 [[TMP1]]
+//
+// CHECK-32-ERROR: error: this builtin is only available on 64-bit targets
+// CHECK-NONPWR9-ERR: error: this builtin is only valid on POWER9 or later CPUs
+ return __extract_sig (d);
+}
+
+double insert_exp (double d, unsigned long long ull) {
+// CHECK-LABEL: @insert_exp(
+// CHECK: [[TMP2:%.*]] = call double @llvm.ppc.insert.exp(double %0, i64 %1)
+// CHECK-NEXT: ret double [[TMP2]]
+//
+// CHECK-32-ERROR: error: this builtin is only available on 64-bit targets
+// CHECK-NONPWR9-ERR: error: this builtin is only valid on POWER9 or later CPUs
+ return __insert_exp (d, ull);
+}
Index: clang/test/CodeGen/builtins-ppc-xlcompat-math.c
===================================================================
--- /dev/null
+++ clang/test/CodeGen/builtins-ppc-xlcompat-math.c
@@ -0,0 +1,143 @@
+// RUN: %clang_cc1 -triple powerpc64-unknown-unknown -emit-llvm %s \
+// RUN: -target-cpu pwr7 -o - | FileCheck %s
+// RUN: %clang_cc1 -triple powerpc64le-unknown-unknown -emit-llvm %s \
+// RUN: -target-cpu pwr8 -o - | FileCheck %s
+// RUN: %clang_cc1 -triple powerpc64-unknown-aix -emit-llvm %s \
+// RUN: -target-cpu pwr7 -o - | FileCheck %s
+// RUN: %clang_cc1 -triple powerpc-unknown-aix -emit-llvm %s \
+// RUN: -target-cpu pwr7 -o - | FileCheck %s
+
+// CHECK-LABEL: @mtfsb0(
+// CHECK: call void @llvm.ppc.mtfsb0(i32 10)
+// CHECK-NEXT: ret void
+//
+void mtfsb0 () {
+ __mtfsb0 (10);
+}
+
+// CHECK-LABEL: @mtfsb1(
+// CHECK: call void @llvm.ppc.mtfsb1(i32 0)
+// CHECK-NEXT: ret void
+//
+void mtfsb1 () {
+ __mtfsb1 (0);
+}
+
+// CHECK-LABEL: @mtfsf(
+// CHECK: [[UI_ADDR:%.*]] = alloca i32, align 4
+// CHECK-NEXT: store i32 [[UI:%.*]], i32* [[UI_ADDR]], align 4
+// CHECK-NEXT: [[TMP0:%.*]] = load i32, i32* [[UI_ADDR]], align 4
+// CHECK-NEXT: call void @llvm.ppc.mtfsf(i32 8, i32 [[TMP0]])
+// CHECK-NEXT: ret void
+//
+void mtfsf (unsigned int ui) {
+ __mtfsf (8, ui);
+}
+
+// CHECK-LABEL: @mtfsfi(
+// CHECK: call void @llvm.ppc.mtfsfi(i32 7, i32 15)
+// CHECK-NEXT: ret void
+//
+void mtfsfi () {
+ __mtfsfi (7, 15);
+}
+
+// CHECK-LABEL: @fmsub(
+// CHECK: [[D_ADDR:%.*]] = alloca double, align 8
+// CHECK-NEXT: store double [[D:%.*]], double* [[D_ADDR]], align 8
+// CHECK-NEXT: [[TMP0:%.*]] = load double, double* [[D_ADDR]], align 8
+// CHECK-NEXT: [[TMP1:%.*]] = load double, double* [[D_ADDR]], align 8
+// CHECK-NEXT: [[TMP2:%.*]] = load double, double* [[D_ADDR]], align 8
+// CHECK-NEXT: [[TMP3:%.*]] = call double @llvm.ppc.fmsub(double [[TMP0]], double [[TMP1]], double [[TMP2]])
+// CHECK-NEXT: ret double [[TMP3]]
+//
+double fmsub (double d) {
+ return __fmsub (d, d, d);
+}
+
+// CHECK-LABEL: @fmsubs(
+// CHECK: [[F_ADDR:%.*]] = alloca float, align 4
+// CHECK-NEXT: store float [[F:%.*]], float* [[F_ADDR]], align 4
+// CHECK-NEXT: [[TMP0:%.*]] = load float, float* [[F_ADDR]], align 4
+// CHECK-NEXT: [[TMP1:%.*]] = load float, float* [[F_ADDR]], align 4
+// CHECK-NEXT: [[TMP2:%.*]] = load float, float* [[F_ADDR]], align 4
+// CHECK-NEXT: [[TMP3:%.*]] = call float @llvm.ppc.fmsubs(float [[TMP0]], float [[TMP1]], float [[TMP2]])
+// CHECK-NEXT: ret float [[TMP3]]
+//
+float fmsubs (float f) {
+ return __fmsubs (f, f, f);
+}
+
+// CHECK-LABEL: @fnmadd(
+// CHECK: [[D_ADDR:%.*]] = alloca double, align 8
+// CHECK-NEXT: store double [[D:%.*]], double* [[D_ADDR]], align 8
+// CHECK-NEXT: [[TMP0:%.*]] = load double, double* [[D_ADDR]], align 8
+// CHECK-NEXT: [[TMP1:%.*]] = load double, double* [[D_ADDR]], align 8
+// CHECK-NEXT: [[TMP2:%.*]] = load double, double* [[D_ADDR]], align 8
+// CHECK-NEXT: [[TMP3:%.*]] = call double @llvm.ppc.fnmadd(double [[TMP0]], double [[TMP1]], double [[TMP2]])
+// CHECK-NEXT: ret double [[TMP3]]
+//
+double fnmadd (double d) {
+ return __fnmadd (d, d, d);
+}
+
+// CHECK-LABEL: @fnmadds(
+// CHECK: [[F_ADDR:%.*]] = alloca float, align 4
+// CHECK-NEXT: store float [[F:%.*]], float* [[F_ADDR]], align 4
+// CHECK-NEXT: [[TMP0:%.*]] = load float, float* [[F_ADDR]], align 4
+// CHECK-NEXT: [[TMP1:%.*]] = load float, float* [[F_ADDR]], align 4
+// CHECK-NEXT: [[TMP2:%.*]] = load float, float* [[F_ADDR]], align 4
+// CHECK-NEXT: [[TMP3:%.*]] = call float @llvm.ppc.fnmadds(float [[TMP0]], float [[TMP1]], float [[TMP2]])
+// CHECK-NEXT: ret float [[TMP3]]
+//
+float fnmadds (float f) {
+ return __fnmadds (f, f, f);
+}
+
+// CHECK-LABEL: @fnmsub(
+// CHECK: [[D_ADDR:%.*]] = alloca double, align 8
+// CHECK-NEXT: store double [[D:%.*]], double* [[D_ADDR]], align 8
+// CHECK-NEXT: [[TMP0:%.*]] = load double, double* [[D_ADDR]], align 8
+// CHECK-NEXT: [[TMP1:%.*]] = load double, double* [[D_ADDR]], align 8
+// CHECK-NEXT: [[TMP2:%.*]] = load double, double* [[D_ADDR]], align 8
+// CHECK-NEXT: [[TMP3:%.*]] = call double @llvm.ppc.fnmsub(double [[TMP0]], double [[TMP1]], double [[TMP2]])
+// CHECK-NEXT: ret double [[TMP3]]
+//
+double fnmsub (double d) {
+ return __fnmsub (d, d, d);
+}
+
+// CHECK-LABEL: @fnmsubs(
+// CHECK: [[F_ADDR:%.*]] = alloca float, align 4
+// CHECK-NEXT: store float [[F:%.*]], float* [[F_ADDR]], align 4
+// CHECK-NEXT: [[TMP0:%.*]] = load float, float* [[F_ADDR]], align 4
+// CHECK-NEXT: [[TMP1:%.*]] = load float, float* [[F_ADDR]], align 4
+// CHECK-NEXT: [[TMP2:%.*]] = load float, float* [[F_ADDR]], align 4
+// CHECK-NEXT: [[TMP3:%.*]] = call float @llvm.ppc.fnmsubs(float [[TMP0]], float [[TMP1]], float [[TMP2]])
+// CHECK-NEXT: ret float [[TMP3]]
+//
+float fnmsubs (float f) {
+ return __fnmsubs (f, f, f);
+}
+
+// CHECK-LABEL: @fre(
+// CHECK: [[D_ADDR:%.*]] = alloca double, align 8
+// CHECK-NEXT: store double [[D:%.*]], double* [[D_ADDR]], align 8
+// CHECK-NEXT: [[TMP0:%.*]] = load double, double* [[D_ADDR]], align 8
+// CHECK-NEXT: [[TMP1:%.*]] = call double @llvm.ppc.fre(double [[TMP0]])
+// CHECK-NEXT: ret double [[TMP1]]
+//
+double fre (double d) {
+ return __fre (d);
+}
+
+// CHECK-LABEL: @fres(
+// CHECK: [[F_ADDR:%.*]] = alloca float, align 4
+// CHECK-NEXT: store float [[F:%.*]], float* [[F_ADDR]], align 4
+// CHECK-NEXT: [[TMP0:%.*]] = load float, float* [[F_ADDR]], align 4
+// CHECK-NEXT: [[TMP1:%.*]] = call float @llvm.ppc.fres(float [[TMP0]])
+// CHECK-NEXT: ret float [[TMP1]]
+//
+float fres (float f) {
+ return __fres (f);
+}
Index: clang/test/CodeGen/builtins-ppc-xlcompat-error.c
===================================================================
--- clang/test/CodeGen/builtins-ppc-xlcompat-error.c
+++ clang/test/CodeGen/builtins-ppc-xlcompat-error.c
@@ -17,3 +17,20 @@
#endif
__tw(ia, ib, 50); //expected-error {{argument value 50 is outside the valid range [0, 31]}}
}
+
+extern unsigned int usi;
+extern double d;
+extern float f;
+
+void testMathBuiltin(void) {
+ __mtfsb0(usi); //expected-error {{argument to '__builtin_ppc_mtfsb0' must be a constant integer}}
+ __mtfsb0(32); //expected-error {{argument value 32 is outside the valid range [0, 31]}}
+ __mtfsb1(usi); //expected-error {{argument to '__builtin_ppc_mtfsb1' must be a constant integer}}
+ __mtfsb1(45); //expected-error {{argument value 45 is outside the valid range [0, 31]}}
+ __mtfsf(usi, usi); //expected-error {{argument to '__builtin_ppc_mtfsf' must be a constant integer}}
+ __mtfsf(350, usi); //expected-error {{argument value 350 is outside the valid range [0, 255]}}
+ __mtfsfi(usi, 0); //expected-error {{argument to '__builtin_ppc_mtfsfi' must be a constant integer}}
+ __mtfsfi(0, usi); //expected-error {{argument to '__builtin_ppc_mtfsfi' must be a constant integer}}
+ __mtfsfi(8, 0); //expected-error {{argument value 8 is outside the valid range [0, 7]}}
+ __mtfsfi(5, 24); //expected-error {{argument value 24 is outside the valid range [0, 15]}}
+}
Index: clang/lib/Sema/SemaChecking.cpp
===================================================================
--- clang/lib/Sema/SemaChecking.cpp
+++ clang/lib/Sema/SemaChecking.cpp
@@ -3269,6 +3269,8 @@
case PPC::BI__builtin_ppc_stdcx:
case PPC::BI__builtin_ppc_tdw:
case PPC::BI__builtin_ppc_trapd:
+ case PPC::BI__builtin_ppc_insert_exp:
+ case PPC::BI__builtin_ppc_extract_sig:
return true;
}
return false;
@@ -3360,6 +3362,19 @@
case PPC::BI__builtin_ppc_tw:
case PPC::BI__builtin_ppc_tdw:
return SemaBuiltinConstantArgRange(TheCall, 2, 0, 31);
+ case PPC::BI__builtin_ppc_extract_exp:
+ case PPC::BI__builtin_ppc_extract_sig:
+ case PPC::BI__builtin_ppc_insert_exp:
+ return SemaFeatureCheck(*this, TheCall, "power9-vector",
+ diag::err_ppc_builtin_only_on_arch, "9");
+ case PPC::BI__builtin_ppc_mtfsb0:
+ case PPC::BI__builtin_ppc_mtfsb1:
+ return SemaBuiltinConstantArgRange(TheCall, 0, 0, 31);
+ case PPC::BI__builtin_ppc_mtfsf:
+ return SemaBuiltinConstantArgRange(TheCall, 0, 0, 255);
+ case PPC::BI__builtin_ppc_mtfsfi:
+ return SemaBuiltinConstantArgRange(TheCall, 0, 0, 7) ||
+ SemaBuiltinConstantArgRange(TheCall, 1, 0, 15);
#define CUSTOM_BUILTIN(Name, Intr, Types, Acc) \
case PPC::BI__builtin_##Name: \
return SemaBuiltinPPCMMACall(TheCall, Types);
Index: clang/lib/Basic/Targets/PPC.cpp
===================================================================
--- clang/lib/Basic/Targets/PPC.cpp
+++ clang/lib/Basic/Targets/PPC.cpp
@@ -130,6 +130,21 @@
Builder.defineMacro("__fctiwz", "__builtin_ppc_fctiwz");
Builder.defineMacro("__fctudz", "__builtin_ppc_fctudz");
Builder.defineMacro("__fctuwz", "__builtin_ppc_fctuwz");
+ Builder.defineMacro("__extract_exp", "__builtin_ppc_extract_exp");
+ Builder.defineMacro("__extract_sig", "__builtin_ppc_extract_sig");
+ Builder.defineMacro("__mtfsb0", "__builtin_ppc_mtfsb0");
+ Builder.defineMacro("__mtfsb1", "__builtin_ppc_mtfsb1");
+ Builder.defineMacro("__mtfsf", "__builtin_ppc_mtfsf");
+ Builder.defineMacro("__mtfsfi", "__builtin_ppc_mtfsfi");
+ Builder.defineMacro("__insert_exp", "__builtin_ppc_insert_exp");
+ Builder.defineMacro("__fmsub", "__builtin_ppc_fmsub");
+ Builder.defineMacro("__fmsubs", "__builtin_ppc_fmsubs");
+ Builder.defineMacro("__fnmadd", "__builtin_ppc_fnmadd");
+ Builder.defineMacro("__fnmadds", "__builtin_ppc_fnmadds");
+ Builder.defineMacro("__fnmsub", "__builtin_ppc_fnmsub");
+ Builder.defineMacro("__fnmsubs", "__builtin_ppc_fnmsubs");
+ Builder.defineMacro("__fre", "__builtin_ppc_fre");
+ Builder.defineMacro("__fres", "__builtin_ppc_fres");
}
/// PPCTargetInfo::getTargetDefines - Return a set of the PowerPC-specific
Index: clang/include/clang/Basic/BuiltinsPPC.def
===================================================================
--- clang/include/clang/Basic/BuiltinsPPC.def
+++ clang/include/clang/Basic/BuiltinsPPC.def
@@ -29,7 +29,7 @@
#define UNALIASED_CUSTOM_BUILTIN(ID, TYPES, ACCUMULATE) \
CUSTOM_BUILTIN(ID, ID, TYPES, ACCUMULATE)
-// builtins for compatibility with the XL compiler
+// XL Compatibility built-ins
BUILTIN(__builtin_ppc_popcntb, "ULiULi", "")
BUILTIN(__builtin_ppc_eieio, "v", "")
BUILTIN(__builtin_ppc_iospace_eieio, "v", "")
@@ -71,6 +71,21 @@
BUILTIN(__builtin_ppc_fctiwz, "dd", "")
BUILTIN(__builtin_ppc_fctudz, "dd", "")
BUILTIN(__builtin_ppc_fctuwz, "dd", "")
+BUILTIN(__builtin_ppc_extract_exp, "Uid", "")
+BUILTIN(__builtin_ppc_extract_sig, "ULLid", "")
+BUILTIN(__builtin_ppc_mtfsb0, "vUIi", "")
+BUILTIN(__builtin_ppc_mtfsb1, "vUIi", "")
+BUILTIN(__builtin_ppc_mtfsf, "vUIiUi", "")
+BUILTIN(__builtin_ppc_mtfsfi, "vUIiUIi", "")
+BUILTIN(__builtin_ppc_insert_exp, "ddULLi", "")
+BUILTIN(__builtin_ppc_fmsub, "dddd", "")
+BUILTIN(__builtin_ppc_fmsubs, "ffff", "")
+BUILTIN(__builtin_ppc_fnmadd, "dddd", "")
+BUILTIN(__builtin_ppc_fnmadds, "ffff", "")
+BUILTIN(__builtin_ppc_fnmsub, "dddd", "")
+BUILTIN(__builtin_ppc_fnmsubs, "ffff", "")
+BUILTIN(__builtin_ppc_fre, "dd", "")
+BUILTIN(__builtin_ppc_fres, "ff", "")
BUILTIN(__builtin_ppc_get_timebase, "ULLi", "n")
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
