xref: /freebsd/contrib/llvm-project/llvm/lib/Target/AMDGPU/AMDGPUHSAMetadataStreamer.cpp (revision 77013d11e6483b970af25e13c9b892075742f7e5) 
1 //===--- AMDGPUHSAMetadataStreamer.cpp --------------------------*- C++ -*-===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 //
9 /// \file
10 /// AMDGPU HSA Metadata Streamer.
11 ///
12 //
13 //===----------------------------------------------------------------------===//
14 
15 #include "AMDGPUHSAMetadataStreamer.h"
16 #include "AMDGPU.h"
17 #include "GCNSubtarget.h"
18 #include "MCTargetDesc/AMDGPUTargetStreamer.h"
19 #include "SIMachineFunctionInfo.h"
20 #include "SIProgramInfo.h"
21 #include "llvm/IR/Module.h"
22 using namespace llvm;
23 
24 static std::pair<Type *, Align> getArgumentTypeAlign(const Argument &Arg,
25                                                      const DataLayout &DL) {
26   Type *Ty = Arg.getType();
27   MaybeAlign ArgAlign;
28   if (Arg.hasByRefAttr()) {
29     Ty = Arg.getParamByRefType();
30     ArgAlign = Arg.getParamAlign();
31   }
32 
33   if (!ArgAlign)
34     ArgAlign = DL.getABITypeAlign(Ty);
35 
36   return std::make_pair(Ty, *ArgAlign);
37 }
38 
39 namespace llvm {
40 
41 static cl::opt<bool> DumpHSAMetadata(
42     "amdgpu-dump-hsa-metadata",
43     cl::desc("Dump AMDGPU HSA Metadata"));
44 static cl::opt<bool> VerifyHSAMetadata(
45     "amdgpu-verify-hsa-metadata",
46     cl::desc("Verify AMDGPU HSA Metadata"));
47 
48 namespace AMDGPU {
49 namespace HSAMD {
50 
51 //===----------------------------------------------------------------------===//
52 // HSAMetadataStreamerV2
53 //===----------------------------------------------------------------------===//
54 void MetadataStreamerV2::dump(StringRef HSAMetadataString) const {
55   errs() << "AMDGPU HSA Metadata:\n" << HSAMetadataString << '\n';
56 }
57 
58 void MetadataStreamerV2::verify(StringRef HSAMetadataString) const {
59   errs() << "AMDGPU HSA Metadata Parser Test: ";
60 
61   HSAMD::Metadata FromHSAMetadataString;
62   if (fromString(HSAMetadataString, FromHSAMetadataString)) {
63     errs() << "FAIL\n";
64     return;
65   }
66 
67   std::string ToHSAMetadataString;
68   if (toString(FromHSAMetadataString, ToHSAMetadataString)) {
69     errs() << "FAIL\n";
70     return;
71   }
72 
73   errs() << (HSAMetadataString == ToHSAMetadataString ? "PASS" : "FAIL")
74          << '\n';
75   if (HSAMetadataString != ToHSAMetadataString) {
76     errs() << "Original input: " << HSAMetadataString << '\n'
77            << "Produced output: " << ToHSAMetadataString << '\n';
78   }
79 }
80 
81 AccessQualifier
82 MetadataStreamerV2::getAccessQualifier(StringRef AccQual) const {
83   if (AccQual.empty())
84     return AccessQualifier::Unknown;
85 
86   return StringSwitch<AccessQualifier>(AccQual)
87              .Case("read_only",  AccessQualifier::ReadOnly)
88              .Case("write_only", AccessQualifier::WriteOnly)
89              .Case("read_write", AccessQualifier::ReadWrite)
90              .Default(AccessQualifier::Default);
91 }
92 
93 AddressSpaceQualifier
94 MetadataStreamerV2::getAddressSpaceQualifier(
95     unsigned AddressSpace) const {
96   switch (AddressSpace) {
97   case AMDGPUAS::PRIVATE_ADDRESS:
98     return AddressSpaceQualifier::Private;
99   case AMDGPUAS::GLOBAL_ADDRESS:
100     return AddressSpaceQualifier::Global;
101   case AMDGPUAS::CONSTANT_ADDRESS:
102     return AddressSpaceQualifier::Constant;
103   case AMDGPUAS::LOCAL_ADDRESS:
104     return AddressSpaceQualifier::Local;
105   case AMDGPUAS::FLAT_ADDRESS:
106     return AddressSpaceQualifier::Generic;
107   case AMDGPUAS::REGION_ADDRESS:
108     return AddressSpaceQualifier::Region;
109   default:
110     return AddressSpaceQualifier::Unknown;
111   }
112 }
113 
114 ValueKind MetadataStreamerV2::getValueKind(Type *Ty, StringRef TypeQual,
115                                            StringRef BaseTypeName) const {
116   if (TypeQual.find("pipe") != StringRef::npos)
117     return ValueKind::Pipe;
118 
119   return StringSwitch<ValueKind>(BaseTypeName)
120              .Case("image1d_t", ValueKind::Image)
121              .Case("image1d_array_t", ValueKind::Image)
122              .Case("image1d_buffer_t", ValueKind::Image)
123              .Case("image2d_t", ValueKind::Image)
124              .Case("image2d_array_t", ValueKind::Image)
125              .Case("image2d_array_depth_t", ValueKind::Image)
126              .Case("image2d_array_msaa_t", ValueKind::Image)
127              .Case("image2d_array_msaa_depth_t", ValueKind::Image)
128              .Case("image2d_depth_t", ValueKind::Image)
129              .Case("image2d_msaa_t", ValueKind::Image)
130              .Case("image2d_msaa_depth_t", ValueKind::Image)
131              .Case("image3d_t", ValueKind::Image)
132              .Case("sampler_t", ValueKind::Sampler)
133              .Case("queue_t", ValueKind::Queue)
134              .Default(isa<PointerType>(Ty) ?
135                           (Ty->getPointerAddressSpace() ==
136                            AMDGPUAS::LOCAL_ADDRESS ?
137                            ValueKind::DynamicSharedPointer :
138                            ValueKind::GlobalBuffer) :
139                       ValueKind::ByValue);
140 }
141 
142 std::string MetadataStreamerV2::getTypeName(Type *Ty, bool Signed) const {
143   switch (Ty->getTypeID()) {
144   case Type::IntegerTyID: {
145     if (!Signed)
146       return (Twine('u') + getTypeName(Ty, true)).str();
147 
148     auto BitWidth = Ty->getIntegerBitWidth();
149     switch (BitWidth) {
150     case 8:
151       return "char";
152     case 16:
153       return "short";
154     case 32:
155       return "int";
156     case 64:
157       return "long";
158     default:
159       return (Twine('i') + Twine(BitWidth)).str();
160     }
161   }
162   case Type::HalfTyID:
163     return "half";
164   case Type::FloatTyID:
165     return "float";
166   case Type::DoubleTyID:
167     return "double";
168   case Type::FixedVectorTyID: {
169     auto VecTy = cast<FixedVectorType>(Ty);
170     auto ElTy = VecTy->getElementType();
171     auto NumElements = VecTy->getNumElements();
172     return (Twine(getTypeName(ElTy, Signed)) + Twine(NumElements)).str();
173   }
174   default:
175     return "unknown";
176   }
177 }
178 
179 std::vector<uint32_t>
180 MetadataStreamerV2::getWorkGroupDimensions(MDNode *Node) const {
181   std::vector<uint32_t> Dims;
182   if (Node->getNumOperands() != 3)
183     return Dims;
184 
185   for (auto &Op : Node->operands())
186     Dims.push_back(mdconst::extract<ConstantInt>(Op)->getZExtValue());
187   return Dims;
188 }
189 
190 Kernel::CodeProps::Metadata
191 MetadataStreamerV2::getHSACodeProps(const MachineFunction &MF,
192                                     const SIProgramInfo &ProgramInfo) const {
193   const GCNSubtarget &STM = MF.getSubtarget<GCNSubtarget>();
194   const SIMachineFunctionInfo &MFI = *MF.getInfo<SIMachineFunctionInfo>();
195   HSAMD::Kernel::CodeProps::Metadata HSACodeProps;
196   const Function &F = MF.getFunction();
197 
198   assert(F.getCallingConv() == CallingConv::AMDGPU_KERNEL ||
199          F.getCallingConv() == CallingConv::SPIR_KERNEL);
200 
201   Align MaxKernArgAlign;
202   HSACodeProps.mKernargSegmentSize = STM.getKernArgSegmentSize(F,
203                                                                MaxKernArgAlign);
204   HSACodeProps.mGroupSegmentFixedSize = ProgramInfo.LDSSize;
205   HSACodeProps.mPrivateSegmentFixedSize = ProgramInfo.ScratchSize;
206   HSACodeProps.mKernargSegmentAlign =
207       std::max(MaxKernArgAlign, Align(4)).value();
208   HSACodeProps.mWavefrontSize = STM.getWavefrontSize();
209   HSACodeProps.mNumSGPRs = ProgramInfo.NumSGPR;
210   HSACodeProps.mNumVGPRs = ProgramInfo.NumVGPR;
211   HSACodeProps.mMaxFlatWorkGroupSize = MFI.getMaxFlatWorkGroupSize();
212   HSACodeProps.mIsDynamicCallStack = ProgramInfo.DynamicCallStack;
213   HSACodeProps.mIsXNACKEnabled = STM.isXNACKEnabled();
214   HSACodeProps.mNumSpilledSGPRs = MFI.getNumSpilledSGPRs();
215   HSACodeProps.mNumSpilledVGPRs = MFI.getNumSpilledVGPRs();
216 
217   return HSACodeProps;
218 }
219 
220 Kernel::DebugProps::Metadata
221 MetadataStreamerV2::getHSADebugProps(const MachineFunction &MF,
222                                      const SIProgramInfo &ProgramInfo) const {
223   return HSAMD::Kernel::DebugProps::Metadata();
224 }
225 
226 void MetadataStreamerV2::emitVersion() {
227   auto &Version = HSAMetadata.mVersion;
228 
229   Version.push_back(VersionMajor);
230   Version.push_back(VersionMinor);
231 }
232 
233 void MetadataStreamerV2::emitPrintf(const Module &Mod) {
234   auto &Printf = HSAMetadata.mPrintf;
235 
236   auto Node = Mod.getNamedMetadata("llvm.printf.fmts");
237   if (!Node)
238     return;
239 
240   for (auto Op : Node->operands())
241     if (Op->getNumOperands())
242       Printf.push_back(
243           std::string(cast<MDString>(Op->getOperand(0))->getString()));
244 }
245 
246 void MetadataStreamerV2::emitKernelLanguage(const Function &Func) {
247   auto &Kernel = HSAMetadata.mKernels.back();
248 
249   // TODO: What about other languages?
250   auto Node = Func.getParent()->getNamedMetadata("opencl.ocl.version");
251   if (!Node || !Node->getNumOperands())
252     return;
253   auto Op0 = Node->getOperand(0);
254   if (Op0->getNumOperands() <= 1)
255     return;
256 
257   Kernel.mLanguage = "OpenCL C";
258   Kernel.mLanguageVersion.push_back(
259       mdconst::extract<ConstantInt>(Op0->getOperand(0))->getZExtValue());
260   Kernel.mLanguageVersion.push_back(
261       mdconst::extract<ConstantInt>(Op0->getOperand(1))->getZExtValue());
262 }
263 
264 void MetadataStreamerV2::emitKernelAttrs(const Function &Func) {
265   auto &Attrs = HSAMetadata.mKernels.back().mAttrs;
266 
267   if (auto Node = Func.getMetadata("reqd_work_group_size"))
268     Attrs.mReqdWorkGroupSize = getWorkGroupDimensions(Node);
269   if (auto Node = Func.getMetadata("work_group_size_hint"))
270     Attrs.mWorkGroupSizeHint = getWorkGroupDimensions(Node);
271   if (auto Node = Func.getMetadata("vec_type_hint")) {
272     Attrs.mVecTypeHint = getTypeName(
273         cast<ValueAsMetadata>(Node->getOperand(0))->getType(),
274         mdconst::extract<ConstantInt>(Node->getOperand(1))->getZExtValue());
275   }
276   if (Func.hasFnAttribute("runtime-handle")) {
277     Attrs.mRuntimeHandle =
278         Func.getFnAttribute("runtime-handle").getValueAsString().str();
279   }
280 }
281 
282 void MetadataStreamerV2::emitKernelArgs(const Function &Func) {
283   for (auto &Arg : Func.args())
284     emitKernelArg(Arg);
285 
286   emitHiddenKernelArgs(Func);
287 }
288 
289 void MetadataStreamerV2::emitKernelArg(const Argument &Arg) {
290   auto Func = Arg.getParent();
291   auto ArgNo = Arg.getArgNo();
292   const MDNode *Node;
293 
294   StringRef Name;
295   Node = Func->getMetadata("kernel_arg_name");
296   if (Node && ArgNo < Node->getNumOperands())
297     Name = cast<MDString>(Node->getOperand(ArgNo))->getString();
298   else if (Arg.hasName())
299     Name = Arg.getName();
300 
301   StringRef TypeName;
302   Node = Func->getMetadata("kernel_arg_type");
303   if (Node && ArgNo < Node->getNumOperands())
304     TypeName = cast<MDString>(Node->getOperand(ArgNo))->getString();
305 
306   StringRef BaseTypeName;
307   Node = Func->getMetadata("kernel_arg_base_type");
308   if (Node && ArgNo < Node->getNumOperands())
309     BaseTypeName = cast<MDString>(Node->getOperand(ArgNo))->getString();
310 
311   StringRef AccQual;
312   if (Arg.getType()->isPointerTy() && Arg.onlyReadsMemory() &&
313       Arg.hasNoAliasAttr()) {
314     AccQual = "read_only";
315   } else {
316     Node = Func->getMetadata("kernel_arg_access_qual");
317     if (Node && ArgNo < Node->getNumOperands())
318       AccQual = cast<MDString>(Node->getOperand(ArgNo))->getString();
319   }
320 
321   StringRef TypeQual;
322   Node = Func->getMetadata("kernel_arg_type_qual");
323   if (Node && ArgNo < Node->getNumOperands())
324     TypeQual = cast<MDString>(Node->getOperand(ArgNo))->getString();
325 
326   const DataLayout &DL = Func->getParent()->getDataLayout();
327 
328   MaybeAlign PointeeAlign;
329   if (auto PtrTy = dyn_cast<PointerType>(Arg.getType())) {
330     if (PtrTy->getAddressSpace() == AMDGPUAS::LOCAL_ADDRESS) {
331       // FIXME: Should report this for all address spaces
332       PointeeAlign = DL.getValueOrABITypeAlignment(Arg.getParamAlign(),
333                                                    PtrTy->getElementType());
334     }
335   }
336 
337   Type *ArgTy;
338   Align ArgAlign;
339   std::tie(ArgTy, ArgAlign) = getArgumentTypeAlign(Arg, DL);
340 
341   emitKernelArg(DL, ArgTy, ArgAlign,
342                 getValueKind(ArgTy, TypeQual, BaseTypeName), PointeeAlign, Name,
343                 TypeName, BaseTypeName, AccQual, TypeQual);
344 }
345 
346 void MetadataStreamerV2::emitKernelArg(const DataLayout &DL, Type *Ty,
347                                        Align Alignment, ValueKind ValueKind,
348                                        MaybeAlign PointeeAlign, StringRef Name,
349                                        StringRef TypeName,
350                                        StringRef BaseTypeName,
351                                        StringRef AccQual, StringRef TypeQual) {
352   HSAMetadata.mKernels.back().mArgs.push_back(Kernel::Arg::Metadata());
353   auto &Arg = HSAMetadata.mKernels.back().mArgs.back();
354 
355   Arg.mName = std::string(Name);
356   Arg.mTypeName = std::string(TypeName);
357   Arg.mSize = DL.getTypeAllocSize(Ty);
358   Arg.mAlign = Alignment.value();
359   Arg.mValueKind = ValueKind;
360   Arg.mPointeeAlign = PointeeAlign ? PointeeAlign->value() : 0;
361 
362   if (auto PtrTy = dyn_cast<PointerType>(Ty))
363     Arg.mAddrSpaceQual = getAddressSpaceQualifier(PtrTy->getAddressSpace());
364 
365   Arg.mAccQual = getAccessQualifier(AccQual);
366 
367   // TODO: Emit Arg.mActualAccQual.
368 
369   SmallVector<StringRef, 1> SplitTypeQuals;
370   TypeQual.split(SplitTypeQuals, " ", -1, false);
371   for (StringRef Key : SplitTypeQuals) {
372     auto P = StringSwitch<bool*>(Key)
373                  .Case("const",    &Arg.mIsConst)
374                  .Case("restrict", &Arg.mIsRestrict)
375                  .Case("volatile", &Arg.mIsVolatile)
376                  .Case("pipe",     &Arg.mIsPipe)
377                  .Default(nullptr);
378     if (P)
379       *P = true;
380   }
381 }
382 
383 void MetadataStreamerV2::emitHiddenKernelArgs(const Function &Func) {
384   int HiddenArgNumBytes =
385       getIntegerAttribute(Func, "amdgpu-implicitarg-num-bytes", 0);
386 
387   if (!HiddenArgNumBytes)
388     return;
389 
390   auto &DL = Func.getParent()->getDataLayout();
391   auto Int64Ty = Type::getInt64Ty(Func.getContext());
392 
393   if (HiddenArgNumBytes >= 8)
394     emitKernelArg(DL, Int64Ty, Align(8), ValueKind::HiddenGlobalOffsetX);
395   if (HiddenArgNumBytes >= 16)
396     emitKernelArg(DL, Int64Ty, Align(8), ValueKind::HiddenGlobalOffsetY);
397   if (HiddenArgNumBytes >= 24)
398     emitKernelArg(DL, Int64Ty, Align(8), ValueKind::HiddenGlobalOffsetZ);
399 
400   auto Int8PtrTy = Type::getInt8PtrTy(Func.getContext(),
401                                       AMDGPUAS::GLOBAL_ADDRESS);
402 
403   // Emit "printf buffer" argument if printf is used, otherwise emit dummy
404   // "none" argument.
405   if (HiddenArgNumBytes >= 32) {
406     if (Func.getParent()->getNamedMetadata("llvm.printf.fmts"))
407       emitKernelArg(DL, Int8PtrTy, Align(8), ValueKind::HiddenPrintfBuffer);
408     else if (Func.getParent()->getFunction("__ockl_hostcall_internal")) {
409       // The printf runtime binding pass should have ensured that hostcall and
410       // printf are not used in the same module.
411       assert(!Func.getParent()->getNamedMetadata("llvm.printf.fmts"));
412       emitKernelArg(DL, Int8PtrTy, Align(8), ValueKind::HiddenHostcallBuffer);
413     } else
414       emitKernelArg(DL, Int8PtrTy, Align(8), ValueKind::HiddenNone);
415   }
416 
417   // Emit "default queue" and "completion action" arguments if enqueue kernel is
418   // used, otherwise emit dummy "none" arguments.
419   if (HiddenArgNumBytes >= 48) {
420     if (Func.hasFnAttribute("calls-enqueue-kernel")) {
421       emitKernelArg(DL, Int8PtrTy, Align(8), ValueKind::HiddenDefaultQueue);
422       emitKernelArg(DL, Int8PtrTy, Align(8), ValueKind::HiddenCompletionAction);
423     } else {
424       emitKernelArg(DL, Int8PtrTy, Align(8), ValueKind::HiddenNone);
425       emitKernelArg(DL, Int8PtrTy, Align(8), ValueKind::HiddenNone);
426     }
427   }
428 
429   // Emit the pointer argument for multi-grid object.
430   if (HiddenArgNumBytes >= 56)
431     emitKernelArg(DL, Int8PtrTy, Align(8), ValueKind::HiddenMultiGridSyncArg);
432 }
433 
434 bool MetadataStreamerV2::emitTo(AMDGPUTargetStreamer &TargetStreamer) {
435   return TargetStreamer.EmitHSAMetadata(getHSAMetadata());
436 }
437 
438 void MetadataStreamerV2::begin(const Module &Mod) {
439   emitVersion();
440   emitPrintf(Mod);
441 }
442 
443 void MetadataStreamerV2::end() {
444   std::string HSAMetadataString;
445   if (toString(HSAMetadata, HSAMetadataString))
446     return;
447 
448   if (DumpHSAMetadata)
449     dump(HSAMetadataString);
450   if (VerifyHSAMetadata)
451     verify(HSAMetadataString);
452 }
453 
454 void MetadataStreamerV2::emitKernel(const MachineFunction &MF,
455                                     const SIProgramInfo &ProgramInfo) {
456   auto &Func = MF.getFunction();
457   if (Func.getCallingConv() != CallingConv::AMDGPU_KERNEL)
458     return;
459 
460   auto CodeProps = getHSACodeProps(MF, ProgramInfo);
461   auto DebugProps = getHSADebugProps(MF, ProgramInfo);
462 
463   HSAMetadata.mKernels.push_back(Kernel::Metadata());
464   auto &Kernel = HSAMetadata.mKernels.back();
465 
466   Kernel.mName = std::string(Func.getName());
467   Kernel.mSymbolName = (Twine(Func.getName()) + Twine("@kd")).str();
468   emitKernelLanguage(Func);
469   emitKernelAttrs(Func);
470   emitKernelArgs(Func);
471   HSAMetadata.mKernels.back().mCodeProps = CodeProps;
472   HSAMetadata.mKernels.back().mDebugProps = DebugProps;
473 }
474 
475 //===----------------------------------------------------------------------===//
476 // HSAMetadataStreamerV3
477 //===----------------------------------------------------------------------===//
478 
479 void MetadataStreamerV3::dump(StringRef HSAMetadataString) const {
480   errs() << "AMDGPU HSA Metadata:\n" << HSAMetadataString << '\n';
481 }
482 
483 void MetadataStreamerV3::verify(StringRef HSAMetadataString) const {
484   errs() << "AMDGPU HSA Metadata Parser Test: ";
485 
486   msgpack::Document FromHSAMetadataString;
487 
488   if (!FromHSAMetadataString.fromYAML(HSAMetadataString)) {
489     errs() << "FAIL\n";
490     return;
491   }
492 
493   std::string ToHSAMetadataString;
494   raw_string_ostream StrOS(ToHSAMetadataString);
495   FromHSAMetadataString.toYAML(StrOS);
496 
497   errs() << (HSAMetadataString == StrOS.str() ? "PASS" : "FAIL") << '\n';
498   if (HSAMetadataString != ToHSAMetadataString) {
499     errs() << "Original input: " << HSAMetadataString << '\n'
500            << "Produced output: " << StrOS.str() << '\n';
501   }
502 }
503 
504 Optional<StringRef>
505 MetadataStreamerV3::getAccessQualifier(StringRef AccQual) const {
506   return StringSwitch<Optional<StringRef>>(AccQual)
507       .Case("read_only", StringRef("read_only"))
508       .Case("write_only", StringRef("write_only"))
509       .Case("read_write", StringRef("read_write"))
510       .Default(None);
511 }
512 
513 Optional<StringRef>
514 MetadataStreamerV3::getAddressSpaceQualifier(unsigned AddressSpace) const {
515   switch (AddressSpace) {
516   case AMDGPUAS::PRIVATE_ADDRESS:
517     return StringRef("private");
518   case AMDGPUAS::GLOBAL_ADDRESS:
519     return StringRef("global");
520   case AMDGPUAS::CONSTANT_ADDRESS:
521     return StringRef("constant");
522   case AMDGPUAS::LOCAL_ADDRESS:
523     return StringRef("local");
524   case AMDGPUAS::FLAT_ADDRESS:
525     return StringRef("generic");
526   case AMDGPUAS::REGION_ADDRESS:
527     return StringRef("region");
528   default:
529     return None;
530   }
531 }
532 
533 StringRef MetadataStreamerV3::getValueKind(Type *Ty, StringRef TypeQual,
534                                            StringRef BaseTypeName) const {
535   if (TypeQual.find("pipe") != StringRef::npos)
536     return "pipe";
537 
538   return StringSwitch<StringRef>(BaseTypeName)
539       .Case("image1d_t", "image")
540       .Case("image1d_array_t", "image")
541       .Case("image1d_buffer_t", "image")
542       .Case("image2d_t", "image")
543       .Case("image2d_array_t", "image")
544       .Case("image2d_array_depth_t", "image")
545       .Case("image2d_array_msaa_t", "image")
546       .Case("image2d_array_msaa_depth_t", "image")
547       .Case("image2d_depth_t", "image")
548       .Case("image2d_msaa_t", "image")
549       .Case("image2d_msaa_depth_t", "image")
550       .Case("image3d_t", "image")
551       .Case("sampler_t", "sampler")
552       .Case("queue_t", "queue")
553       .Default(isa<PointerType>(Ty)
554                    ? (Ty->getPointerAddressSpace() == AMDGPUAS::LOCAL_ADDRESS
555                           ? "dynamic_shared_pointer"
556                           : "global_buffer")
557                    : "by_value");
558 }
559 
560 std::string MetadataStreamerV3::getTypeName(Type *Ty, bool Signed) const {
561   switch (Ty->getTypeID()) {
562   case Type::IntegerTyID: {
563     if (!Signed)
564       return (Twine('u') + getTypeName(Ty, true)).str();
565 
566     auto BitWidth = Ty->getIntegerBitWidth();
567     switch (BitWidth) {
568     case 8:
569       return "char";
570     case 16:
571       return "short";
572     case 32:
573       return "int";
574     case 64:
575       return "long";
576     default:
577       return (Twine('i') + Twine(BitWidth)).str();
578     }
579   }
580   case Type::HalfTyID:
581     return "half";
582   case Type::FloatTyID:
583     return "float";
584   case Type::DoubleTyID:
585     return "double";
586   case Type::FixedVectorTyID: {
587     auto VecTy = cast<FixedVectorType>(Ty);
588     auto ElTy = VecTy->getElementType();
589     auto NumElements = VecTy->getNumElements();
590     return (Twine(getTypeName(ElTy, Signed)) + Twine(NumElements)).str();
591   }
592   default:
593     return "unknown";
594   }
595 }
596 
597 msgpack::ArrayDocNode
598 MetadataStreamerV3::getWorkGroupDimensions(MDNode *Node) const {
599   auto Dims = HSAMetadataDoc->getArrayNode();
600   if (Node->getNumOperands() != 3)
601     return Dims;
602 
603   for (auto &Op : Node->operands())
604     Dims.push_back(Dims.getDocument()->getNode(
605         uint64_t(mdconst::extract<ConstantInt>(Op)->getZExtValue())));
606   return Dims;
607 }
608 
609 void MetadataStreamerV3::emitVersion() {
610   auto Version = HSAMetadataDoc->getArrayNode();
611   Version.push_back(Version.getDocument()->getNode(VersionMajor));
612   Version.push_back(Version.getDocument()->getNode(VersionMinor));
613   getRootMetadata("amdhsa.version") = Version;
614 }
615 
616 void MetadataStreamerV3::emitPrintf(const Module &Mod) {
617   auto Node = Mod.getNamedMetadata("llvm.printf.fmts");
618   if (!Node)
619     return;
620 
621   auto Printf = HSAMetadataDoc->getArrayNode();
622   for (auto Op : Node->operands())
623     if (Op->getNumOperands())
624       Printf.push_back(Printf.getDocument()->getNode(
625           cast<MDString>(Op->getOperand(0))->getString(), /*Copy=*/true));
626   getRootMetadata("amdhsa.printf") = Printf;
627 }
628 
629 void MetadataStreamerV3::emitKernelLanguage(const Function &Func,
630                                             msgpack::MapDocNode Kern) {
631   // TODO: What about other languages?
632   auto Node = Func.getParent()->getNamedMetadata("opencl.ocl.version");
633   if (!Node || !Node->getNumOperands())
634     return;
635   auto Op0 = Node->getOperand(0);
636   if (Op0->getNumOperands() <= 1)
637     return;
638 
639   Kern[".language"] = Kern.getDocument()->getNode("OpenCL C");
640   auto LanguageVersion = Kern.getDocument()->getArrayNode();
641   LanguageVersion.push_back(Kern.getDocument()->getNode(
642       mdconst::extract<ConstantInt>(Op0->getOperand(0))->getZExtValue()));
643   LanguageVersion.push_back(Kern.getDocument()->getNode(
644       mdconst::extract<ConstantInt>(Op0->getOperand(1))->getZExtValue()));
645   Kern[".language_version"] = LanguageVersion;
646 }
647 
648 void MetadataStreamerV3::emitKernelAttrs(const Function &Func,
649                                          msgpack::MapDocNode Kern) {
650 
651   if (auto Node = Func.getMetadata("reqd_work_group_size"))
652     Kern[".reqd_workgroup_size"] = getWorkGroupDimensions(Node);
653   if (auto Node = Func.getMetadata("work_group_size_hint"))
654     Kern[".workgroup_size_hint"] = getWorkGroupDimensions(Node);
655   if (auto Node = Func.getMetadata("vec_type_hint")) {
656     Kern[".vec_type_hint"] = Kern.getDocument()->getNode(
657         getTypeName(
658             cast<ValueAsMetadata>(Node->getOperand(0))->getType(),
659             mdconst::extract<ConstantInt>(Node->getOperand(1))->getZExtValue()),
660         /*Copy=*/true);
661   }
662   if (Func.hasFnAttribute("runtime-handle")) {
663     Kern[".device_enqueue_symbol"] = Kern.getDocument()->getNode(
664         Func.getFnAttribute("runtime-handle").getValueAsString().str(),
665         /*Copy=*/true);
666   }
667 }
668 
669 void MetadataStreamerV3::emitKernelArgs(const Function &Func,
670                                         msgpack::MapDocNode Kern) {
671   unsigned Offset = 0;
672   auto Args = HSAMetadataDoc->getArrayNode();
673   for (auto &Arg : Func.args())
674     emitKernelArg(Arg, Offset, Args);
675 
676   emitHiddenKernelArgs(Func, Offset, Args);
677 
678   Kern[".args"] = Args;
679 }
680 
681 void MetadataStreamerV3::emitKernelArg(const Argument &Arg, unsigned &Offset,
682                                        msgpack::ArrayDocNode Args) {
683   auto Func = Arg.getParent();
684   auto ArgNo = Arg.getArgNo();
685   const MDNode *Node;
686 
687   StringRef Name;
688   Node = Func->getMetadata("kernel_arg_name");
689   if (Node && ArgNo < Node->getNumOperands())
690     Name = cast<MDString>(Node->getOperand(ArgNo))->getString();
691   else if (Arg.hasName())
692     Name = Arg.getName();
693 
694   StringRef TypeName;
695   Node = Func->getMetadata("kernel_arg_type");
696   if (Node && ArgNo < Node->getNumOperands())
697     TypeName = cast<MDString>(Node->getOperand(ArgNo))->getString();
698 
699   StringRef BaseTypeName;
700   Node = Func->getMetadata("kernel_arg_base_type");
701   if (Node && ArgNo < Node->getNumOperands())
702     BaseTypeName = cast<MDString>(Node->getOperand(ArgNo))->getString();
703 
704   StringRef AccQual;
705   if (Arg.getType()->isPointerTy() && Arg.onlyReadsMemory() &&
706       Arg.hasNoAliasAttr()) {
707     AccQual = "read_only";
708   } else {
709     Node = Func->getMetadata("kernel_arg_access_qual");
710     if (Node && ArgNo < Node->getNumOperands())
711       AccQual = cast<MDString>(Node->getOperand(ArgNo))->getString();
712   }
713 
714   StringRef TypeQual;
715   Node = Func->getMetadata("kernel_arg_type_qual");
716   if (Node && ArgNo < Node->getNumOperands())
717     TypeQual = cast<MDString>(Node->getOperand(ArgNo))->getString();
718 
719   const DataLayout &DL = Func->getParent()->getDataLayout();
720 
721   MaybeAlign PointeeAlign;
722   Type *Ty = Arg.hasByRefAttr() ? Arg.getParamByRefType() : Arg.getType();
723 
724   // FIXME: Need to distinguish in memory alignment from pointer alignment.
725   if (auto PtrTy = dyn_cast<PointerType>(Ty)) {
726     if (PtrTy->getAddressSpace() == AMDGPUAS::LOCAL_ADDRESS) {
727       PointeeAlign = DL.getValueOrABITypeAlignment(Arg.getParamAlign(),
728                                                    PtrTy->getElementType());
729     }
730   }
731 
732   // There's no distinction between byval aggregates and raw aggregates.
733   Type *ArgTy;
734   Align ArgAlign;
735   std::tie(ArgTy, ArgAlign) = getArgumentTypeAlign(Arg, DL);
736 
737   emitKernelArg(DL, ArgTy, ArgAlign,
738                 getValueKind(ArgTy, TypeQual, BaseTypeName), Offset, Args,
739                 PointeeAlign, Name, TypeName, BaseTypeName, AccQual, TypeQual);
740 }
741 
742 void MetadataStreamerV3::emitKernelArg(
743     const DataLayout &DL, Type *Ty, Align Alignment, StringRef ValueKind,
744     unsigned &Offset, msgpack::ArrayDocNode Args, MaybeAlign PointeeAlign,
745     StringRef Name, StringRef TypeName, StringRef BaseTypeName,
746     StringRef AccQual, StringRef TypeQual) {
747   auto Arg = Args.getDocument()->getMapNode();
748 
749   if (!Name.empty())
750     Arg[".name"] = Arg.getDocument()->getNode(Name, /*Copy=*/true);
751   if (!TypeName.empty())
752     Arg[".type_name"] = Arg.getDocument()->getNode(TypeName, /*Copy=*/true);
753   auto Size = DL.getTypeAllocSize(Ty);
754   Arg[".size"] = Arg.getDocument()->getNode(Size);
755   Offset = alignTo(Offset, Alignment);
756   Arg[".offset"] = Arg.getDocument()->getNode(Offset);
757   Offset += Size;
758   Arg[".value_kind"] = Arg.getDocument()->getNode(ValueKind, /*Copy=*/true);
759   if (PointeeAlign)
760     Arg[".pointee_align"] = Arg.getDocument()->getNode(PointeeAlign->value());
761 
762   if (auto PtrTy = dyn_cast<PointerType>(Ty))
763     if (auto Qualifier = getAddressSpaceQualifier(PtrTy->getAddressSpace()))
764       Arg[".address_space"] = Arg.getDocument()->getNode(*Qualifier, /*Copy=*/true);
765 
766   if (auto AQ = getAccessQualifier(AccQual))
767     Arg[".access"] = Arg.getDocument()->getNode(*AQ, /*Copy=*/true);
768 
769   // TODO: Emit Arg[".actual_access"].
770 
771   SmallVector<StringRef, 1> SplitTypeQuals;
772   TypeQual.split(SplitTypeQuals, " ", -1, false);
773   for (StringRef Key : SplitTypeQuals) {
774     if (Key == "const")
775       Arg[".is_const"] = Arg.getDocument()->getNode(true);
776     else if (Key == "restrict")
777       Arg[".is_restrict"] = Arg.getDocument()->getNode(true);
778     else if (Key == "volatile")
779       Arg[".is_volatile"] = Arg.getDocument()->getNode(true);
780     else if (Key == "pipe")
781       Arg[".is_pipe"] = Arg.getDocument()->getNode(true);
782   }
783 
784   Args.push_back(Arg);
785 }
786 
787 void MetadataStreamerV3::emitHiddenKernelArgs(const Function &Func,
788                                               unsigned &Offset,
789                                               msgpack::ArrayDocNode Args) {
790   int HiddenArgNumBytes =
791       getIntegerAttribute(Func, "amdgpu-implicitarg-num-bytes", 0);
792 
793   if (!HiddenArgNumBytes)
794     return;
795 
796   auto &DL = Func.getParent()->getDataLayout();
797   auto Int64Ty = Type::getInt64Ty(Func.getContext());
798 
799   if (HiddenArgNumBytes >= 8)
800     emitKernelArg(DL, Int64Ty, Align(8), "hidden_global_offset_x", Offset,
801                   Args);
802   if (HiddenArgNumBytes >= 16)
803     emitKernelArg(DL, Int64Ty, Align(8), "hidden_global_offset_y", Offset,
804                   Args);
805   if (HiddenArgNumBytes >= 24)
806     emitKernelArg(DL, Int64Ty, Align(8), "hidden_global_offset_z", Offset,
807                   Args);
808 
809   auto Int8PtrTy =
810       Type::getInt8PtrTy(Func.getContext(), AMDGPUAS::GLOBAL_ADDRESS);
811 
812   // Emit "printf buffer" argument if printf is used, otherwise emit dummy
813   // "none" argument.
814   if (HiddenArgNumBytes >= 32) {
815     if (Func.getParent()->getNamedMetadata("llvm.printf.fmts"))
816       emitKernelArg(DL, Int8PtrTy, Align(8), "hidden_printf_buffer", Offset,
817                     Args);
818     else if (Func.getParent()->getFunction("__ockl_hostcall_internal")) {
819       // The printf runtime binding pass should have ensured that hostcall and
820       // printf are not used in the same module.
821       assert(!Func.getParent()->getNamedMetadata("llvm.printf.fmts"));
822       emitKernelArg(DL, Int8PtrTy, Align(8), "hidden_hostcall_buffer", Offset,
823                     Args);
824     } else
825       emitKernelArg(DL, Int8PtrTy, Align(8), "hidden_none", Offset, Args);
826   }
827 
828   // Emit "default queue" and "completion action" arguments if enqueue kernel is
829   // used, otherwise emit dummy "none" arguments.
830   if (HiddenArgNumBytes >= 48) {
831     if (Func.hasFnAttribute("calls-enqueue-kernel")) {
832       emitKernelArg(DL, Int8PtrTy, Align(8), "hidden_default_queue", Offset,
833                     Args);
834       emitKernelArg(DL, Int8PtrTy, Align(8), "hidden_completion_action", Offset,
835                     Args);
836     } else {
837       emitKernelArg(DL, Int8PtrTy, Align(8), "hidden_none", Offset, Args);
838       emitKernelArg(DL, Int8PtrTy, Align(8), "hidden_none", Offset, Args);
839     }
840   }
841 
842   // Emit the pointer argument for multi-grid object.
843   if (HiddenArgNumBytes >= 56)
844     emitKernelArg(DL, Int8PtrTy, Align(8), "hidden_multigrid_sync_arg", Offset,
845                   Args);
846 }
847 
848 msgpack::MapDocNode
849 MetadataStreamerV3::getHSAKernelProps(const MachineFunction &MF,
850                                       const SIProgramInfo &ProgramInfo) const {
851   const GCNSubtarget &STM = MF.getSubtarget<GCNSubtarget>();
852   const SIMachineFunctionInfo &MFI = *MF.getInfo<SIMachineFunctionInfo>();
853   const Function &F = MF.getFunction();
854 
855   auto Kern = HSAMetadataDoc->getMapNode();
856 
857   Align MaxKernArgAlign;
858   Kern[".kernarg_segment_size"] = Kern.getDocument()->getNode(
859       STM.getKernArgSegmentSize(F, MaxKernArgAlign));
860   Kern[".group_segment_fixed_size"] =
861       Kern.getDocument()->getNode(ProgramInfo.LDSSize);
862   Kern[".private_segment_fixed_size"] =
863       Kern.getDocument()->getNode(ProgramInfo.ScratchSize);
864   Kern[".kernarg_segment_align"] =
865       Kern.getDocument()->getNode(std::max(Align(4), MaxKernArgAlign).value());
866   Kern[".wavefront_size"] =
867       Kern.getDocument()->getNode(STM.getWavefrontSize());
868   Kern[".sgpr_count"] = Kern.getDocument()->getNode(ProgramInfo.NumSGPR);
869   Kern[".vgpr_count"] = Kern.getDocument()->getNode(ProgramInfo.NumVGPR);
870   Kern[".max_flat_workgroup_size"] =
871       Kern.getDocument()->getNode(MFI.getMaxFlatWorkGroupSize());
872   Kern[".sgpr_spill_count"] =
873       Kern.getDocument()->getNode(MFI.getNumSpilledSGPRs());
874   Kern[".vgpr_spill_count"] =
875       Kern.getDocument()->getNode(MFI.getNumSpilledVGPRs());
876 
877   return Kern;
878 }
879 
880 bool MetadataStreamerV3::emitTo(AMDGPUTargetStreamer &TargetStreamer) {
881   return TargetStreamer.EmitHSAMetadata(*HSAMetadataDoc, true);
882 }
883 
884 void MetadataStreamerV3::begin(const Module &Mod) {
885   emitVersion();
886   emitPrintf(Mod);
887   getRootMetadata("amdhsa.kernels") = HSAMetadataDoc->getArrayNode();
888 }
889 
890 void MetadataStreamerV3::end() {
891   std::string HSAMetadataString;
892   raw_string_ostream StrOS(HSAMetadataString);
893   HSAMetadataDoc->toYAML(StrOS);
894 
895   if (DumpHSAMetadata)
896     dump(StrOS.str());
897   if (VerifyHSAMetadata)
898     verify(StrOS.str());
899 }
900 
901 void MetadataStreamerV3::emitKernel(const MachineFunction &MF,
902                                     const SIProgramInfo &ProgramInfo) {
903   auto &Func = MF.getFunction();
904   auto Kern = getHSAKernelProps(MF, ProgramInfo);
905 
906   assert(Func.getCallingConv() == CallingConv::AMDGPU_KERNEL ||
907          Func.getCallingConv() == CallingConv::SPIR_KERNEL);
908 
909   auto Kernels =
910       getRootMetadata("amdhsa.kernels").getArray(/*Convert=*/true);
911 
912   {
913     Kern[".name"] = Kern.getDocument()->getNode(Func.getName());
914     Kern[".symbol"] = Kern.getDocument()->getNode(
915         (Twine(Func.getName()) + Twine(".kd")).str(), /*Copy=*/true);
916     emitKernelLanguage(Func, Kern);
917     emitKernelAttrs(Func, Kern);
918     emitKernelArgs(Func, Kern);
919   }
920 
921   Kernels.push_back(Kern);
922 }
923 
924 } // end namespace HSAMD
925 } // end namespace AMDGPU
926 } // end namespace llvm
927