[PowerPC] Add support for -mcpu=pwr10 in both clang and llvm

[lei]

Summary

This patch simply adds support for the new CPU in anticipation of
Power10. There isn't really any functionality added so there are no
associated test cases at this time.

[lei]

Add support in llvm.

[lei]

missed a file

[amyk]

I believe we're also missing IsISA3_1 = false; in PPCSubtarget.cpp.

llvm/lib/Target/PowerPC/PPC.td

338

Are we missing FeatureISA3_1 in P10AdditionalFeatures?

[steven.zhang]

clang/lib/Basic/Targets/PPC.cpp

272

Do we miss to define this macro ?

__POWER10_VECTOR__

llvm/lib/Target/PowerPC/PPC.td

211

Nit: no ending dot.

[steven.zhang]

clang/lib/Basic/Targets/PPC.cpp

272

Well, pls ignore this comments as that macro should be added when p10 vector feature is added.

[stefanp]

Most of my comments are related to the fact that we are now inserting P10 between P9 and Future and so a few things need to change for the Future code to sit on top of P10 now.

clang/test/Preprocessor/init-ppc64.c

654

Since we are adding P10 between P9 and Future we should add another line here:

PPCFUTURE:#define _ARCH_PWR10 1

llvm/lib/Target/PowerPC/PPC.td

340

I think these can be moved up to P10AdditionalFeatures. That way everything on P10 is now inheritable by future and we don't have to specify anything for FutureSpecificFeatures.

351

These features are now no longer FutureSpecificFeatures I would think that they would now be part of Power10 and should be inherited by future CPU.

llvm/test/CodeGen/PowerPC/check-cpu.ll

11

nit:
We should probably update this comment too.

[NeHuang]

clang/lib/Basic/Targets/PPC.cpp

338

I think we also need to check for ArchDefinePwr10 and ArchDefineFuture based on the comment "// We have __float128 on PPC but not power 9 and above."

!(ArchDefs & ArchDefinePwr9) -> !(ArchDefs & (ArchDefinePwr9 | ArchDefinePwr10 | ArchDefineFuture))

[lei]

address review comments

clang/lib/Basic/Targets/PPC.cpp

338

I would think this would be redundant since -mcpu=pwr10 also defines ArchDefinePwr9 ...

llvm/lib/Target/PowerPC/PPC.td

211

That's needed to indicate the end of the sentence. Similar to line 208.

[NeHuang]

clang/lib/Basic/Targets/PPC.cpp

338

Yeah. That makes sense. Thanks for the explanation!

[amyk]

llvm/lib/Target/PowerPC/PPCTargetTransformInfo.cpp

654

Should this comment be updated to include P10 as well?

[nemanjai]

LGTM aside from a couple of minor nits.

clang/lib/Basic/Targets/PPC.cpp

319

Please remove this since HTM was removed in P10.

clang/test/Preprocessor/init-ppc64.c

644

I am not sure what the story is with not checking for _ARCH_PWR8 for the P9 test, but I don't think we need to continue that precedent.

[stefanp]

LGTM

[amyk]

I think this looks good aside from the comments I had.

llvm/lib/Target/PowerPC/PPCSubtarget.h

142

Missing IsISA3_1 = false; in PPCSubtarget.cpp.

[lei]

update as per reviewers comments

[lei]

change how we generate p10 feature list.

[stefanp]

LGTM

[amyk]

This LGTM.

[lei]

rebased

[lei]

fix up rebase issue after revert

clang/lib/Basic/Targets/PPC.h

typedef enum {
     ArchDefinePwr4 = 1 << 6,
     ArchDefinePwr5 = 1 << 7,
     ArchDefinePwr5x = 1 << 8,
     ArchDefinePwr6 = 1 << 9,
     ArchDefinePwr6x = 1 << 10,
     ArchDefinePwr7 = 1 << 11,
     ArchDefinePwr8 = 1 << 12,
     ArchDefinePwr9 = 1 << 13,
-    ArchDefineFuture = 1 << 14,
-    ArchDefineA2 = 1 << 15,
-    ArchDefineA2q = 1 << 16,
-    ArchDefineE500 = 1 << 17
+    ArchDefinePwr10 = 1 << 14,
+    ArchDefineFuture = 1 << 15,
+    ArchDefineA2 = 1 << 16,
+    ArchDefineA2q = 1 << 17,
+    ArchDefineE500 = 1 << 18
   } ArchDefineTypes;
 
-
   ArchDefineTypes ArchDefs = ArchDefineNone;
   static const Builtin::Info BuiltinInfo[];
   static const char *const GCCRegNames[];
   static const TargetInfo::GCCRegAlias GCCRegAliases[];
   std::string CPU;
   enum PPCFloatABI { HardFloat, SoftFloat } FloatABI;
 
   // Target cpu features.
if (CPUKnown) {
               .Case("7450", ArchDefineName | ArchDefinePpcgr)
               .Case("750", ArchDefineName | ArchDefinePpcgr)
               .Case("970", ArchDefineName | ArchDefinePwr4 | ArchDefinePpcgr |
                                ArchDefinePpcsq)
               .Case("a2", ArchDefineA2)
               .Case("a2q", ArchDefineName | ArchDefineA2 | ArchDefineA2q)
               .Cases("power3", "pwr3", ArchDefinePpcgr)
               .Cases("power4", "pwr4",
                      ArchDefinePwr4 | ArchDefinePpcgr | ArchDefinePpcsq)
               .Cases("power5", "pwr5",
                      ArchDefinePwr5 | ArchDefinePwr4 | ArchDefinePpcgr |
                          ArchDefinePpcsq)
               .Cases("power5x", "pwr5x",
                      ArchDefinePwr5x | ArchDefinePwr5 | ArchDefinePwr4 |
                          ArchDefinePpcgr | ArchDefinePpcsq)
               .Cases("power6", "pwr6",
                      ArchDefinePwr6 | ArchDefinePwr5x | ArchDefinePwr5 |
                          ArchDefinePwr4 | ArchDefinePpcgr | ArchDefinePpcsq)
               .Cases("power6x", "pwr6x",
                      ArchDefinePwr6x | ArchDefinePwr6 | ArchDefinePwr5x |
                          ArchDefinePwr5 | ArchDefinePwr4 | ArchDefinePpcgr |
                          ArchDefinePpcsq)
               .Cases("power7", "pwr7",
                      ArchDefinePwr7 | ArchDefinePwr6 | ArchDefinePwr5x |
                          ArchDefinePwr5 | ArchDefinePwr4 | ArchDefinePpcgr |
                          ArchDefinePpcsq)
               // powerpc64le automatically defaults to at least power8.
               .Cases("power8", "pwr8", "ppc64le",
                      ArchDefinePwr8 | ArchDefinePwr7 | ArchDefinePwr6 |
                          ArchDefinePwr5x | ArchDefinePwr5 | ArchDefinePwr4 |
                          ArchDefinePpcgr | ArchDefinePpcsq)
               .Cases("power9", "pwr9",
                      ArchDefinePwr9 | ArchDefinePwr8 | ArchDefinePwr7 |
                          ArchDefinePwr6 | ArchDefinePwr5x | ArchDefinePwr5 |
                          ArchDefinePwr4 | ArchDefinePpcgr | ArchDefinePpcsq)
-              .Case("future",
-                    ArchDefineFuture | ArchDefinePwr9 | ArchDefinePwr8 |
+              .Cases("power10", "pwr10",
+                     ArchDefinePwr10 | ArchDefinePwr9 | ArchDefinePwr8 |
                          ArchDefinePwr7 | ArchDefinePwr6 | ArchDefinePwr5x |
                          ArchDefinePwr5 | ArchDefinePwr4 | ArchDefinePpcgr |
                          ArchDefinePpcsq)
+              .Case("future",
+                    ArchDefineFuture | ArchDefinePwr10 | ArchDefinePwr9 |
+                        ArchDefinePwr8 | ArchDefinePwr7 | ArchDefinePwr6 |
+                        ArchDefinePwr5x | ArchDefinePwr5 | ArchDefinePwr4 |
+                        ArchDefinePpcgr | ArchDefinePpcsq)
               .Cases("8548", "e500", ArchDefineE500)
               .Default(ArchDefineNone);
     }
     return CPUKnown;
   }
 
   StringRef getABI() const override { return ABI; }
 

clang/lib/Basic/Targets/PPC.cpp

void PPCTargetInfo::getTargetDefines(const LangOptions &Opts,
   if (ArchDefs & ArchDefinePwr6x)
     Builder.defineMacro("_ARCH_PWR6X");
   if (ArchDefs & ArchDefinePwr7)
     Builder.defineMacro("_ARCH_PWR7");
   if (ArchDefs & ArchDefinePwr8)
     Builder.defineMacro("_ARCH_PWR8");
   if (ArchDefs & ArchDefinePwr9)
     Builder.defineMacro("_ARCH_PWR9");
+  if (ArchDefs & ArchDefinePwr10)
+    Builder.defineMacro("_ARCH_PWR10");
   if (ArchDefs & ArchDefineA2)
     Builder.defineMacro("_ARCH_A2");
   if (ArchDefs & ArchDefineA2q) {
     Builder.defineMacro("_ARCH_A2Q");
     Builder.defineMacro("_ARCH_QP");
   }
   if (ArchDefs & ArchDefineE500)
     Builder.defineMacro("__NO_LWSYNC__");
Features["altivec"] = llvm::StringSwitch<bool>(CPU)
                             .Case("7450", true)
                             .Case("g4+", true)
                             .Case("970", true)
                             .Case("g5", true)
                             .Case("pwr6", true)
                             .Case("pwr7", true)
                             .Case("pwr8", true)
                             .Case("pwr9", true)
+                            .Case("pwr10", true)
                             .Case("ppc64", true)
                             .Case("ppc64le", true)
                             .Default(false);
 
   Features["qpx"] = (CPU == "a2q");
-  Features["power9-vector"] = (CPU == "pwr9");
+  Features["power9-vector"] = llvm::StringSwitch<bool>(CPU)
+                                  .Case("pwr10", true)
+                                  .Case("pwr9", true)
+                                  .Default(false);
   Features["crypto"] = llvm::StringSwitch<bool>(CPU)
                            .Case("ppc64le", true)
+                           .Case("pwr10", true)
                            .Case("pwr9", true)
                            .Case("pwr8", true)
                            .Default(false);
   Features["power8-vector"] = llvm::StringSwitch<bool>(CPU)
                                   .Case("ppc64le", true)
+                                  .Case("pwr10", true)
                                   .Case("pwr9", true)
                                   .Case("pwr8", true)
                                   .Default(false);
   Features["bpermd"] = llvm::StringSwitch<bool>(CPU)
                            .Case("ppc64le", true)
+                           .Case("pwr10", true)
                            .Case("pwr9", true)
                            .Case("pwr8", true)
                            .Case("pwr7", true)
                            .Default(false);
   Features["extdiv"] = llvm::StringSwitch<bool>(CPU)
                            .Case("ppc64le", true)
+                           .Case("pwr10", true)
                            .Case("pwr9", true)
                            .Case("pwr8", true)
                            .Case("pwr7", true)
                            .Default(false);
   Features["direct-move"] = llvm::StringSwitch<bool>(CPU)
                                 .Case("ppc64le", true)
+                                .Case("pwr10", true)
                                 .Case("pwr9", true)
                                 .Case("pwr8", true)
                                 .Default(false);
   Features["vsx"] = llvm::StringSwitch<bool>(CPU)
                         .Case("ppc64le", true)
+                        .Case("pwr10", true)
                         .Case("pwr9", true)
                         .Case("pwr8", true)
                         .Case("pwr7", true)
                         .Default(false);
   Features["htm"] = llvm::StringSwitch<bool>(CPU)
                         .Case("ppc64le", true)
                         .Case("pwr9", true)
                         .Case("pwr8", true)
                         .Default(false);
 
   Features["spe"] = llvm::StringSwitch<bool>(CPU)
                         .Case("8548", true)
                         .Case("e500", true)
                         .Default(false);
 
-  // Future CPU should include all of the features of Power 9 as well as any
+  // Future CPU should include all of the features of Power 10 as well as any
   // additional features (yet to be determined) specific to it.
   if (CPU == "future") {
-    initFeatureMap(Features, Diags, "pwr9", FeaturesVec);
+    initFeatureMap(Features, Diags, "pwr10", FeaturesVec);
     addFutureSpecificFeatures(Features);
   }
 
   if (!ppcUserFeaturesCheck(Diags, FeaturesVec))
     return false;
 
   if (!(ArchDefs & ArchDefinePwr9) && (ArchDefs & ArchDefinePpcgr) &&
       llvm::find(FeaturesVec, "+float128") != FeaturesVec.end()) {
     // We have __float128 on PPC but not power 9 and above.
     Diags.Report(diag::err_opt_not_valid_with_opt) << "-mfloat128" << CPU;
     return false;
   }
 
   return TargetInfo::initFeatureMap(Features, Diags, CPU, FeaturesVec);
 }

 ArrayRef<TargetInfo::AddlRegName> PPCTargetInfo::getGCCAddlRegNames() const {
   if (ABI == "elfv2")
     return llvm::makeArrayRef(GCCAddlRegNames);
   else 
     return TargetInfo::getGCCAddlRegNames(); 
 }
 
 static constexpr llvm::StringLiteral ValidCPUNames[] = {
-    {"generic"},   {"440"},       {"450"},         {"601"},         {"602"},
-    {"603"},       {"603e"},      {"603ev"},       {"604"},         {"604e"},
-    {"620"},       {"630"},       {"g3"},          {"7400"},        {"g4"},
-    {"7450"},      {"g4+"},       {"750"},         {"8548"},        {"970"},
-    {"g5"},        {"a2"},        {"a2q"},         {"e500"},        {"e500mc"},
-    {"e5500"},     {"power3"},    {"pwr3"},        {"power4"},      {"pwr4"},
-    {"power5"},    {"pwr5"},      {"power5x"},     {"pwr5x"},       {"power6"},
-    {"pwr6"},      {"power6x"},   {"pwr6x"},       {"power7"},      {"pwr7"},
-    {"power8"},    {"pwr8"},      {"power9"},      {"pwr9"},        {"powerpc"},
-    {"ppc"},       {"powerpc64"}, {"ppc64"},       {"powerpc64le"}, {"ppc64le"},
-    {"future"}
-};
+    {"generic"},     {"440"},     {"450"},     {"601"},       {"602"},
+    {"603"},         {"603e"},    {"603ev"},   {"604"},       {"604e"},
+    {"620"},         {"630"},     {"g3"},      {"7400"},      {"g4"},
+    {"7450"},        {"g4+"},     {"750"},     {"8548"},      {"970"},
+    {"g5"},          {"a2"},      {"a2q"},     {"e500"},      {"e500mc"},
+    {"e5500"},       {"power3"},  {"pwr3"},    {"power4"},    {"pwr4"},
+    {"power5"},      {"pwr5"},    {"power5x"}, {"pwr5x"},     {"power6"},
+    {"pwr6"},        {"power6x"}, {"pwr6x"},   {"power7"},    {"pwr7"},
+    {"power8"},      {"pwr8"},    {"power9"},  {"pwr9"},      {"power10"},
+    {"pwr10"},       {"powerpc"}, {"ppc"},     {"powerpc64"}, {"ppc64"},
+    {"powerpc64le"}, {"ppc64le"}, {"future"}};
 
 bool PPCTargetInfo::isValidCPUName(StringRef Name) const {
   return llvm::find(ValidCPUNames, Name) != std::end(ValidCPUNames);
 }
 
 void PPCTargetInfo::fillValidCPUList(SmallVectorImpl<StringRef> &Values) const {
   Values.append(std::begin(ValidCPUNames), std::end(ValidCPUNames));
 }

clang/lib/Driver/ToolChains/Arch/PPC.cpp

return llvm::StringSwitch<const char *>(CPUName)
         .Case("power4", "pwr4")
         .Case("power5", "pwr5")
         .Case("power5x", "pwr5x")
         .Case("power6", "pwr6")
         .Case("power6x", "pwr6x")
         .Case("power7", "pwr7")
         .Case("power8", "pwr8")
         .Case("power9", "pwr9")
+        .Case("power10", "pwr10")
         .Case("future", "future")
         .Case("pwr3", "pwr3")
         .Case("pwr4", "pwr4")
         .Case("pwr5", "pwr5")
         .Case("pwr5x", "pwr5x")
         .Case("pwr6", "pwr6")
         .Case("pwr6x", "pwr6x")
         .Case("pwr7", "pwr7")
         .Case("pwr8", "pwr8")
         .Case("pwr9", "pwr9")
+        .Case("pwr10", "pwr10")
         .Case("powerpc", "ppc")
         .Case("powerpc64", "ppc64")
         .Case("powerpc64le", "ppc64le")
         .Default("");
   }
 
   return "";
 }
 
 const char *ppc::getPPCAsmModeForCPU(StringRef Name) {
   return llvm::StringSwitch<const char *>(Name)
-        .Case("pwr7", "-mpower7")
-        .Case("power7", "-mpower7")
-        .Case("pwr8", "-mpower8")
-        .Case("power8", "-mpower8")
-        .Case("ppc64le", "-mpower8")
-        .Case("pwr9", "-mpower9")
-        .Case("power9", "-mpower9")
-        .Default("-many");
+      .Case("pwr10", "-mpower10")
+      .Case("power10", "-mpower10")
 }
 
 void ppc::getPPCTargetFeatures(const Driver &D, const llvm::Triple &Triple,
                                const ArgList &Args,
                                std::vector<StringRef> &Features) {
   if (Triple.getSubArch() == llvm::Triple::PPCSubArch_spe)
     Features.push_back("+spe");
 

clang/test/Misc/target-invalid-cpu-note.c

 // SPARCV9: note: valid target CPU values are: v9, ultrasparc, ultrasparc3, niagara, niagara2, niagara3, niagara4
 
 // RUN: not %clang_cc1 -triple powerpc--- -target-cpu not-a-cpu -fsyntax-only %s 2>&1 | FileCheck %s --check-prefix PPC
 // PPC: error: unknown target CPU 'not-a-cpu'
 // PPC: note: valid target CPU values are: generic, 440, 450, 601, 602, 603,
 // PPC-SAME: 603e, 603ev, 604, 604e, 620, 630, g3, 7400, g4, 7450, g4+, 750,
 // PPC-SAME: 8548, 970, g5, a2, a2q, e500, e500mc, e5500, power3, pwr3, power4,
 // PPC-SAME: pwr4, power5, pwr5, power5x, pwr5x, power6, pwr6, power6x, pwr6x,
-// PPC-SAME: power7, pwr7, power8, pwr8, power9, pwr9, powerpc, ppc, powerpc64,
+// PPC-SAME: power7, pwr7, power8, pwr8, power9, pwr9, power10, pwr10, powerpc, ppc, powerpc64,
 // PPC-SAME: ppc64, powerpc64le, ppc64le, future
 
 // RUN: not %clang_cc1 -triple mips--- -target-cpu not-a-cpu -fsyntax-only %s 2>&1 | FileCheck %s --check-prefix MIPS
 // MIPS: error: unknown target CPU 'not-a-cpu'
 // MIPS: note: valid target CPU values are: mips1, mips2, mips3, mips4, mips5,
 // MIPS-SAME: mips32, mips32r2, mips32r3, mips32r5, mips32r6, mips64, mips64r2,
 // MIPS-SAME: mips64r3, mips64r5, mips64r6, octeon, octeon+, p5600
 

clang/test/Preprocessor/init-ppc64.c

 // PPCPOWER9:#define _ARCH_PWR4 1
 // PPCPOWER9:#define _ARCH_PWR5 1
 // PPCPOWER9:#define _ARCH_PWR5X 1
 // PPCPOWER9:#define _ARCH_PWR6 1
 // PPCPOWER9-NOT:#define _ARCH_PWR6X 1
 // PPCPOWER9:#define _ARCH_PWR7 1
 // PPCPOWER9:#define _ARCH_PWR9 1
 //
+// RUN: %clang_cc1 -E -dM -ffreestanding -triple=powerpc64-none-none -target-cpu pwr10 -fno-signed-char < /dev/null | FileCheck -match-full-lines -check-prefix PPCPOWER10 %s
+// RUN: %clang_cc1 -E -dM -ffreestanding -triple=powerpc64-none-none -target-cpu power10 -fno-signed-char < /dev/null | FileCheck -match-full-lines -check-prefix PPCPOWER10 %s
+//
+// PPCPOWER10:#define _ARCH_PPC 1
+// PPCPOWER10:#define _ARCH_PPC64 1
+// PPCPOWER10:#define _ARCH_PPCGR 1
+// PPCPOWER10:#define _ARCH_PPCSQ 1
+// PPCPOWER10:#define _ARCH_PWR10 1
+// PPCPOWER10:#define _ARCH_PWR4 1
+// PPCPOWER10:#define _ARCH_PWR5 1
+// PPCPOWER10:#define _ARCH_PWR5X 1
+// PPCPOWER10:#define _ARCH_PWR6 1
+// PPCPOWER10-NOT:#define _ARCH_PWR6X 1
+// PPCPOWER10:#define _ARCH_PWR7 1
+// PPCPOWER10:#define _ARCH_PWR8 1
+// PPCPOWER10:#define _ARCH_PWR9 1
+//
 // RUN: %clang_cc1 -E -dM -ffreestanding -triple=powerpc64-none-none -target-cpu future -fno-signed-char < /dev/null | FileCheck -match-full-lines -check-prefix PPCFUTURE %s
 //
 // PPCFUTURE:#define _ARCH_PPC 1
 // PPCFUTURE:#define _ARCH_PPC64 1
 // PPCFUTURE:#define _ARCH_PPCGR 1
 // PPCFUTURE:#define _ARCH_PPCSQ 1
+// PPCFUTURE:#define _ARCH_PWR10 1
 // PPCFUTURE:#define _ARCH_PWR4 1
 // PPCFUTURE:#define _ARCH_PWR5 1
 // PPCFUTURE:#define _ARCH_PWR5X 1
 // PPCFUTURE:#define _ARCH_PWR6 1
 // PPCFUTURE-NOT:#define _ARCH_PWR6X 1
 // PPCFUTURE:#define _ARCH_PWR7 1
 // PPCFUTURE:#define _ARCH_PWR8 1
 // PPCFUTURE:#define _ARCH_PWR9 1
 // PPCFUTURE:#define _ARCH_PWR_FUTURE 1

llvm/lib/Support/Host.cpp

return StringSwitch<const char *>(StringRef(CPUStart, CPULen))
       .Case("POWER5", "g5")
       .Case("A2", "a2")
       .Case("POWER6", "pwr6")
       .Case("POWER7", "pwr7")
       .Case("POWER8", "pwr8")
       .Case("POWER8E", "pwr8")
       .Case("POWER8NVL", "pwr8")
       .Case("POWER9", "pwr9")
+      .Case("POWER10", "pwr10")
       // FIXME: If we get a simulator or machine with the capabilities of
       // mcpu=future, we should revisit this and add the name reported by the
       // simulator/machine.
       .Default(generic);
 }
 
 StringRef sys::detail::getHostCPUNameForARM(StringRef ProcCpuinfoContent) {
   // The cpuid register on arm is not accessible from user space. On Linux,

llvm/lib/Target/PowerPC/PPC.td

 def DirectivePwr5x
     : SubtargetFeature<"", "CPUDirective", "PPC::DIR_PWR5X", "">;
 def DirectivePwr6: SubtargetFeature<"", "CPUDirective", "PPC::DIR_PWR6", "">;
 def DirectivePwr6x
     : SubtargetFeature<"", "CPUDirective", "PPC::DIR_PWR6X", "">;
 def DirectivePwr7: SubtargetFeature<"", "CPUDirective", "PPC::DIR_PWR7", "">;
 def DirectivePwr8: SubtargetFeature<"", "CPUDirective", "PPC::DIR_PWR8", "">;
 def DirectivePwr9: SubtargetFeature<"", "CPUDirective", "PPC::DIR_PWR9", "">;
+def DirectivePwr10: SubtargetFeature<"", "CPUDirective", "PPC::DIR_PWR10", "">;
 def DirectivePwrFuture
     : SubtargetFeature<"", "CPUDirective", "PPC::DIR_PWR_FUTURE", "">;
 
 def Feature64Bit     : SubtargetFeature<"64bit","Has64BitSupport", "true",
                                         "Enable 64-bit instructions">;
 def FeatureHardFloat : SubtargetFeature<"hard-float", "HasHardFloat", "true",
                               "Enable floating-point instructions">;
 def Feature64BitRegs : SubtargetFeature<"64bitregs","Use64BitRegs", "true",
def FeatureSlowPOPCNTD : SubtargetFeature<"slow-popcntd","HasPOPCNTD",
                                           "Has slow popcnt[dw] instructions">;
 
 def DeprecatedDST    : SubtargetFeature<"", "DeprecatedDST", "true",
   "Treat vector data stream cache control instructions as deprecated">;
 
 def FeatureISA3_0 : SubtargetFeature<"isa-v30-instructions", "IsISA3_0",
                                      "true",
                                      "Enable instructions added in ISA 3.0.">;
+def FeatureISA3_1 : SubtargetFeature<"isa-v31-instructions", "IsISA3_1",
+                                     "true",
+                                     "Enable instructions added in ISA 3.1.">;
 def FeatureP9Altivec : SubtargetFeature<"power9-altivec", "HasP9Altivec", "true",
                                         "Enable POWER9 Altivec instructions",
                                         [FeatureISA3_0, FeatureP8Altivec]>;
 def FeatureP9Vector  : SubtargetFeature<"power9-vector", "HasP9Vector", "true",
                                         "Enable POWER9 vector instructions",
                                         [FeatureISA3_0, FeatureP8Vector,
                                          FeatureP9Altivec]>;
 // A separate feature for this even though it is equivalent to P9Vector
def ProcessorFeatures {
   list<SubtargetFeature> P9SpecificFeatures = [FeatureVectorsUseTwoUnits,
                                                FeaturePPCPreRASched,
                                                FeaturePPCPostRASched];
   list<SubtargetFeature> P9InheritableFeatures =
     !listconcat(P8InheritableFeatures, P9AdditionalFeatures);
   list<SubtargetFeature> P9Features =
     !listconcat(P9InheritableFeatures, P9SpecificFeatures);
 
-  // Future
-  // For future CPU we assume that all of the existing features from Power 9
-  // still exist with the exception of those we know are Power 9 specific.
+  // Power10
+  // For P10 CPU we assume that all of the existing features from Power9
+  // still exist with the exception of those we know are Power9 specific.
+  list<SubtargetFeature> P10AdditionalFeatures =
+    [DirectivePwr10, FeatureISA3_1, FeaturePrefixInstrs,
+     FeaturePCRelativeMemops];
+  list<SubtargetFeature> P10SpecificFeatures = [];
+  list<SubtargetFeature> P10InheritableFeatures =
+    !listconcat(P9InheritableFeatures, P10AdditionalFeatures);
+  list<SubtargetFeature> P10Features =
+    !listconcat(P10InheritableFeatures, P10SpecificFeatures);
+
+  // Future
+  // For future CPU we assume that all of the existing features from Power10
+  // still exist with the exception of those we know are Power10 specific.
   list<SubtargetFeature> FutureAdditionalFeatures = [];
-  list<SubtargetFeature> FutureSpecificFeatures =
-    [FeaturePrefixInstrs, FeaturePCRelativeMemops];
+  list<SubtargetFeature> FutureSpecificFeatures = [];
   list<SubtargetFeature> FutureInheritableFeatures =
-    !listconcat(P9InheritableFeatures, FutureAdditionalFeatures);
+    !listconcat(P10InheritableFeatures, FutureAdditionalFeatures);
   list<SubtargetFeature> FutureFeatures =
     !listconcat(FutureInheritableFeatures, FutureSpecificFeatures);
 }
 
 // Note: Future features to add when support is extended to more
 // recent ISA levels:
 //
 // DFP          p6, p6x, p7        decimal floating-point instructions
def : ProcessorModel<"pwr6x", G5Model,
                    FeatureFCPSGN, FeatureFSqrt, FeatureFRE, FeatureFRES,
                    FeatureFRSQRTE, FeatureFRSQRTES, FeatureRecipPrec,
                    FeatureSTFIWX, FeatureLFIWAX, FeatureCMPB,
                    FeatureFPRND, Feature64Bit,
                    FeatureMFTB, DeprecatedDST]>;
 def : ProcessorModel<"pwr7", P7Model, ProcessorFeatures.P7Features>;
 def : ProcessorModel<"pwr8", P8Model, ProcessorFeatures.P8Features>;
 def : ProcessorModel<"pwr9", P9Model, ProcessorFeatures.P9Features>;
+// No scheduler model yet.
+def : ProcessorModel<"pwr10", NoSchedModel, ProcessorFeatures.P10Features>;
 // No scheduler model for future CPU.
 def : ProcessorModel<"future", NoSchedModel,
                   ProcessorFeatures.FutureFeatures>;
 def : Processor<"ppc", G3Itineraries, [Directive32, FeatureHardFloat,
                                        FeatureMFTB]>;
 def : Processor<"ppc32", G3Itineraries, [Directive32, FeatureHardFloat,
                                          FeatureMFTB]>;
 def : ProcessorModel<"ppc64", G5Model,

llvm/lib/Target/PowerPC/PPCISelLowering.cpp

   case PPC::DIR_PWR4:
   case PPC::DIR_PWR5:
   case PPC::DIR_PWR5X:
   case PPC::DIR_PWR6:
   case PPC::DIR_PWR6X:
   case PPC::DIR_PWR7:
   case PPC::DIR_PWR8:
   case PPC::DIR_PWR9:
+  case PPC::DIR_PWR10:
   case PPC::DIR_PWR_FUTURE:
     setPrefLoopAlignment(Align(16));
     setPrefFunctionAlignment(Align(16));
     break;
   }
 
   if (Subtarget.enableMachineScheduler())
     setSchedulingPreference(Sched::Source);
Align PPCTargetLowering::getPrefLoopAlignment(MachineLoop *ML) const {
   case PPC::DIR_PWR4:
   case PPC::DIR_PWR5:
   case PPC::DIR_PWR5X:
   case PPC::DIR_PWR6:
   case PPC::DIR_PWR6X:
   case PPC::DIR_PWR7:
   case PPC::DIR_PWR8:
   case PPC::DIR_PWR9:
+  case PPC::DIR_PWR10:
   case PPC::DIR_PWR_FUTURE: {
     if (!ML)
       break;
 
     if (!DisableInnermostLoopAlign32) {
       // If the nested loop is an innermost loop, prefer to a 32-byte alignment,
       // so that we can decrease cache misses and branch-prediction misses.
       // Actual alignment of the loop will depend on the hotness check and other
default:
       // TODO: enhance the condition for subtarget before pwr8
       return false;
     case PPC::DIR_PWR8:
       //  type        mul     add    shl
       // scalar        4       1      1
       // vector        7       2      2
       return true;
     case PPC::DIR_PWR9:
+    case PPC::DIR_PWR10:
     case PPC::DIR_PWR_FUTURE:
       //  type        mul     add    shl
       // scalar        5       2      2
       // vector        7       2      2
 
       // The cycle RATIO of related operations are showed as a table above.
       // Because mul is 5(scalar)/7(vector), add/sub/shl are all 2 for both
       // scalar and vector type. For 2 instrs patterns, add/sub + shl

llvm/lib/Target/PowerPC/PPCSubtarget.h

 // GCC #defines PPC on Linux but we use it as our namespace name
 #undef PPC
 
 namespace llvm {
 class StringRef;
 
 namespace PPC {
   // -m directive values.
-  enum {
-    DIR_NONE,
-    DIR_32,
-    DIR_440,
-    DIR_601,
-    DIR_602,
-    DIR_603,
-    DIR_7400,
-    DIR_750,
-    DIR_970,
-    DIR_A2,
-    DIR_E500,
-    DIR_E500mc,
-    DIR_E5500,
-    DIR_PWR3,
-    DIR_PWR4,
-    DIR_PWR5,
-    DIR_PWR5X,
-    DIR_PWR6,
-    DIR_PWR6X,
-    DIR_PWR7,
-    DIR_PWR8,
-    DIR_PWR9,
-    DIR_PWR_FUTURE,
-    DIR_64
-  };
+  DIR_PWR10,
 }
 
 class GlobalValue;
 
 class PPCSubtarget : public PPCGenSubtargetInfo {
 public:
   enum POPCNTDKind {
     POPCNTD_Unavailable,
protected:
   bool HasPartwordAtomics;
   bool HasDirectMove;
   bool HasHTM;
   bool HasFloat128;
   bool HasFusion;
   bool HasAddiLoadFusion;
   bool HasAddisLoadFusion;
   bool IsISA3_0;
+  bool IsISA3_1;
   bool UseLongCalls;
   bool SecurePlt;
   bool VectorsUseTwoUnits;
   bool UsePPCPreRASchedStrategy;
   bool UsePPCPostRASchedStrategy;
   bool PredictableSelectIsExpensive;
 
   POPCNTDKind HasPOPCNTD;
unsigned  getRedZoneSize() const {
     // AIX PPC32: 220 bytes = 18*8 (FPRs) + 19*4 (GPRs);
     // PPC32 SVR4ABI has no redzone.
     return isAIXABI() ? 220 : 0;
   }
 
   bool hasHTM() const { return HasHTM; }
   bool hasFloat128() const { return HasFloat128; }
   bool isISA3_0() const { return IsISA3_0; }
+  bool isISA3_1() const { return IsISA3_1; }
   bool useLongCalls() const { return UseLongCalls; }
   bool hasFusion() const { return HasFusion; }
   bool hasAddiLoadFusion() const { return HasAddiLoadFusion; }
   bool hasAddisLoadFusion() const { return HasAddisLoadFusion; }
   bool needsSwapsForVSXMemOps() const {
     return hasVSX() && isLittleEndian() && !hasP9Vector();
   }
 

llvm/lib/Target/PowerPC/PPCSubtarget.cpp

void PPCSubtarget::initializeEnvironment() {
   HasDirectMove = false;
   IsQPXStackUnaligned = false;
   HasHTM = false;
   HasFloat128 = false;
   HasFusion = false;
   HasAddiLoadFusion = false;
   HasAddisLoadFusion = false;
   IsISA3_0 = false;
+  IsISA3_1 = false;
   UseLongCalls = false;
   SecurePlt = false;
   VectorsUseTwoUnits = false;
   UsePPCPreRASchedStrategy = false;
   UsePPCPostRASchedStrategy = false;
   PredictableSelectIsExpensive = false;
 
   HasPOPCNTD = POPCNTD_Unavailable;

llvm/lib/Target/PowerPC/PPCTargetTransformInfo.cpp

 
 }
 
 unsigned PPCTTIImpl::getCacheLineSize() const {
   // Check first if the user specified a custom line size.
   if (CacheLineSize.getNumOccurrences() > 0)
     return CacheLineSize;
 
-  // On P7, P8 or P9 we have a cache line size of 128.
+  // Starting with P7 we have a cache line size of 128.
   unsigned Directive = ST->getCPUDirective();
   // Assume that Future CPU has the same cache line size as the others.
   if (Directive == PPC::DIR_PWR7 || Directive == PPC::DIR_PWR8 ||
-      Directive == PPC::DIR_PWR9 || Directive == PPC::DIR_PWR_FUTURE)
+      Directive == PPC::DIR_PWR9 || Directive == PPC::DIR_PWR10 ||
+      Directive == PPC::DIR_PWR_FUTURE)
     return 128;
 
   // On other processors return a default of 64 bytes.
   return 64;
 }
 
 unsigned PPCTTIImpl::getPrefetchDistance() const {
   // This seems like a reasonable default for the BG/Q (this pass is enabled, by
unsigned PPCTTIImpl::getMaxInterleaveFactor(unsigned VF) {
 
   // FIXME: For lack of any better information, do no harm...
   if (Directive == PPC::DIR_E500mc || Directive == PPC::DIR_E5500)
     return 1;
 
   // For P7 and P8, floating-point instructions have a 6-cycle latency and
   // there are two execution units, so unroll by 12x for latency hiding.
   // FIXME: the same for P9 as previous gen until POWER9 scheduling is ready
+  // FIXME: the same for P10 as previous gen until POWER10 scheduling is ready
   // Assume that future is the same as the others.
   if (Directive == PPC::DIR_PWR7 || Directive == PPC::DIR_PWR8 ||
-      Directive == PPC::DIR_PWR9 || Directive == PPC::DIR_PWR_FUTURE)
+      Directive == PPC::DIR_PWR9 || Directive == PPC::DIR_PWR10 ||
+      Directive == PPC::DIR_PWR_FUTURE)
     return 12;
 
   // For most things, modern systems have two execution units (and
   // out-of-order execution).
   return 2;
 }
 
 // Adjust the cost of vector instructions on targets which there is overlap

llvm/test/CodeGen/PowerPC/check-cpu.ll

 ; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
 ; RUN:     -mcpu=future < %s | FileCheck %s
 ; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
 ; RUN:     -mcpu=future < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
+; RUN:     -mcpu=power10 < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
+; RUN:     -mcpu=pwr10 < %s | FileCheck %s
 
 
-; Test mcpu=future that should be recognized on PowerPC.
+; Test -mcpu=[pwr10|future] is recognized on PowerPC.
 
 ; CHECK-NOT: is not a recognized processor for this target
 ; CHECK:     .text