Files
clang-p2996/llvm/lib/ProfileData/SampleProf.cpp
Wenlei He 6b989a1710 [CSSPGO] Infrastructure for context-sensitive Sample PGO and Inlining
This change adds the context-senstive sample PGO infracture described in CSSPGO RFC (https://groups.google.com/g/llvm-dev/c/1p1rdYbL93s). It introduced an abstraction between input profile and profile loader that queries input profile for functions. Specifically, there's now the notion of base profile and context profile, and they are managed by the new SampleContextTracker for adjusting and merging profiles based on inline decisions. It works with top-down profiled guided inliner in profile loader (https://reviews.llvm.org/D70655) for better inlining with specialization and better post-inline profile fidelity. In the future, we can also expose this infrastructure to CGSCC inliner in order for it to take advantage of context-sensitive profile. This change is the consumption part of context-sensitive profile (The generation part is in this stack: https://reviews.llvm.org/D89707). We've seen good results internally in conjunction with Pseudo-probe (https://reviews.llvm.org/D86193). Pacthes for integration with Pseudo-probe coming up soon.

Currently the new infrastructure kick in when input profile contains the new context-sensitive profile; otherwise it's no-op and does not affect existing AutoFDO.

**Interface**

There're two sets of interfaces for query and tracking respectively exposed from SampleContextTracker. For query, now instead of simply getting a profile from input for a function, we can explicitly query base profile or context profile for given call path of a function. For tracking, there're separate APIs for marking context profile as inlined, or promoting and merging not inlined context profile.

- Query base profile (`getBaseSamplesFor`)
Base profile is the merged synthetic profile for function's CFG profile from any outstanding (not inlined) context. We can query base profile by function.

- Query context profile (`getContextSamplesFor`)
Context profile is a function's CFG profile for a given calling context. We can query context profile by context string.

- Track inlined context profile (`markContextSamplesInlined`)
When a function is inlined for given calling context, we need to mark the context profile for that context as inlined. This is to make sure we don't include inlined context profile when synthesizing base profile for that inlined function.

- Track not-inlined context profile (`promoteMergeContextSamplesTree`)
When a function is not inlined for given calling context, we need to promote the context profile tree so the not inlined context becomes top-level context. This preserve the sub-context under that function so later inline decision for that not inlined function will still have context profile for its call tree. Note that profile will be merged if needed when promoting a context profile tree if any of the node already exists at its promoted destination.

**Implementation**

Implementation-wise, `SampleContext` is created as abstraction for context. Currently it's a string for call path, and we can later optimize it to something more efficient, e.g. context id. Each `SampleContext` also has a `ContextState` indicating whether it's raw context profile from input, whether it's inlined or merged, whether it's synthetic profile created by compiler. Each `FunctionSamples` now has a `SampleContext` that tells whether it's base profile or context profile, and for context profile what is the context and state.

On top of the above context representation, a custom trie tree is implemented to track and manager context profiles. Specifically, `SampleContextTracker` is implemented that encapsulates a trie tree with `ContextTireNode` as node. Each node of the trie tree represents a frame in calling context, thus the path from root to a node represents a valid calling context. We also track `FunctionSamples` for each node, so this trie tree can serve efficient query for context profile. Accordingly, context profile tree promotion now becomes moving a subtree to be under the root of entire tree, and merge nodes for subtree if this move encounters existing nodes.

**Integration**

`SampleContextTracker` is now also integrated with AutoFDO, `SampleProfileReader` and `SampleProfileLoader`. When we detected input profile contains context-sensitive profile, `SampleContextTracker` will be used to track profiles, and all profile query will go to `SampleContextTracker` instead of `SampleProfileReader` automatically. Tracking APIs are called automatically for each inline decision from `SampleProfileLoader`.

Differential Revision: https://reviews.llvm.org/D90125
2020-12-06 11:49:18 -08:00

292 lines
9.4 KiB
C++

//=-- SampleProf.cpp - Sample profiling format support --------------------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
//
// This file contains common definitions used in the reading and writing of
// sample profile data.
//
//===----------------------------------------------------------------------===//
#include "llvm/ProfileData/SampleProf.h"
#include "llvm/Config/llvm-config.h"
#include "llvm/IR/DebugInfoMetadata.h"
#include "llvm/ProfileData/SampleProfReader.h"
#include "llvm/Support/Compiler.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/Error.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/LEB128.h"
#include "llvm/Support/ManagedStatic.h"
#include "llvm/Support/raw_ostream.h"
#include <string>
#include <system_error>
using namespace llvm;
using namespace sampleprof;
namespace llvm {
namespace sampleprof {
SampleProfileFormat FunctionSamples::Format;
bool FunctionSamples::ProfileIsCS = false;
bool FunctionSamples::UseMD5;
} // namespace sampleprof
} // namespace llvm
namespace {
// FIXME: This class is only here to support the transition to llvm::Error. It
// will be removed once this transition is complete. Clients should prefer to
// deal with the Error value directly, rather than converting to error_code.
class SampleProfErrorCategoryType : public std::error_category {
const char *name() const noexcept override { return "llvm.sampleprof"; }
std::string message(int IE) const override {
sampleprof_error E = static_cast<sampleprof_error>(IE);
switch (E) {
case sampleprof_error::success:
return "Success";
case sampleprof_error::bad_magic:
return "Invalid sample profile data (bad magic)";
case sampleprof_error::unsupported_version:
return "Unsupported sample profile format version";
case sampleprof_error::too_large:
return "Too much profile data";
case sampleprof_error::truncated:
return "Truncated profile data";
case sampleprof_error::malformed:
return "Malformed sample profile data";
case sampleprof_error::unrecognized_format:
return "Unrecognized sample profile encoding format";
case sampleprof_error::unsupported_writing_format:
return "Profile encoding format unsupported for writing operations";
case sampleprof_error::truncated_name_table:
return "Truncated function name table";
case sampleprof_error::not_implemented:
return "Unimplemented feature";
case sampleprof_error::counter_overflow:
return "Counter overflow";
case sampleprof_error::ostream_seek_unsupported:
return "Ostream does not support seek";
case sampleprof_error::compress_failed:
return "Compress failure";
case sampleprof_error::uncompress_failed:
return "Uncompress failure";
case sampleprof_error::zlib_unavailable:
return "Zlib is unavailable";
}
llvm_unreachable("A value of sampleprof_error has no message.");
}
};
} // end anonymous namespace
static ManagedStatic<SampleProfErrorCategoryType> ErrorCategory;
const std::error_category &llvm::sampleprof_category() {
return *ErrorCategory;
}
void LineLocation::print(raw_ostream &OS) const {
OS << LineOffset;
if (Discriminator > 0)
OS << "." << Discriminator;
}
raw_ostream &llvm::sampleprof::operator<<(raw_ostream &OS,
const LineLocation &Loc) {
Loc.print(OS);
return OS;
}
#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
LLVM_DUMP_METHOD void LineLocation::dump() const { print(dbgs()); }
#endif
/// Print the sample record to the stream \p OS indented by \p Indent.
void SampleRecord::print(raw_ostream &OS, unsigned Indent) const {
OS << NumSamples;
if (hasCalls()) {
OS << ", calls:";
for (const auto &I : getSortedCallTargets())
OS << " " << I.first << ":" << I.second;
}
OS << "\n";
}
#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
LLVM_DUMP_METHOD void SampleRecord::dump() const { print(dbgs(), 0); }
#endif
raw_ostream &llvm::sampleprof::operator<<(raw_ostream &OS,
const SampleRecord &Sample) {
Sample.print(OS, 0);
return OS;
}
/// Print the samples collected for a function on stream \p OS.
void FunctionSamples::print(raw_ostream &OS, unsigned Indent) const {
OS << TotalSamples << ", " << TotalHeadSamples << ", " << BodySamples.size()
<< " sampled lines\n";
OS.indent(Indent);
if (!BodySamples.empty()) {
OS << "Samples collected in the function's body {\n";
SampleSorter<LineLocation, SampleRecord> SortedBodySamples(BodySamples);
for (const auto &SI : SortedBodySamples.get()) {
OS.indent(Indent + 2);
OS << SI->first << ": " << SI->second;
}
OS.indent(Indent);
OS << "}\n";
} else {
OS << "No samples collected in the function's body\n";
}
OS.indent(Indent);
if (!CallsiteSamples.empty()) {
OS << "Samples collected in inlined callsites {\n";
SampleSorter<LineLocation, FunctionSamplesMap> SortedCallsiteSamples(
CallsiteSamples);
for (const auto &CS : SortedCallsiteSamples.get()) {
for (const auto &FS : CS->second) {
OS.indent(Indent + 2);
OS << CS->first << ": inlined callee: " << FS.second.getName() << ": ";
FS.second.print(OS, Indent + 4);
}
}
OS.indent(Indent);
OS << "}\n";
} else {
OS << "No inlined callsites in this function\n";
}
}
raw_ostream &llvm::sampleprof::operator<<(raw_ostream &OS,
const FunctionSamples &FS) {
FS.print(OS);
return OS;
}
unsigned FunctionSamples::getOffset(const DILocation *DIL) {
return (DIL->getLine() - DIL->getScope()->getSubprogram()->getLine()) &
0xffff;
}
const FunctionSamples *FunctionSamples::findFunctionSamples(
const DILocation *DIL, SampleProfileReaderItaniumRemapper *Remapper) const {
assert(DIL);
SmallVector<std::pair<LineLocation, StringRef>, 10> S;
const DILocation *PrevDIL = DIL;
for (DIL = DIL->getInlinedAt(); DIL; DIL = DIL->getInlinedAt()) {
S.push_back(std::make_pair(
LineLocation(getOffset(DIL), DIL->getBaseDiscriminator()),
PrevDIL->getScope()->getSubprogram()->getLinkageName()));
PrevDIL = DIL;
}
if (S.size() == 0)
return this;
const FunctionSamples *FS = this;
for (int i = S.size() - 1; i >= 0 && FS != nullptr; i--) {
FS = FS->findFunctionSamplesAt(S[i].first, S[i].second, Remapper);
}
return FS;
}
void FunctionSamples::findAllNames(DenseSet<StringRef> &NameSet) const {
NameSet.insert(Name);
for (const auto &BS : BodySamples)
for (const auto &TS : BS.second.getCallTargets())
NameSet.insert(TS.getKey());
for (const auto &CS : CallsiteSamples) {
for (const auto &NameFS : CS.second) {
NameSet.insert(NameFS.first);
NameFS.second.findAllNames(NameSet);
}
}
}
const FunctionSamples *FunctionSamples::findFunctionSamplesAt(
const LineLocation &Loc, StringRef CalleeName,
SampleProfileReaderItaniumRemapper *Remapper) const {
std::string CalleeGUID;
CalleeName = getRepInFormat(CalleeName, UseMD5, CalleeGUID);
auto iter = CallsiteSamples.find(Loc);
if (iter == CallsiteSamples.end())
return nullptr;
auto FS = iter->second.find(CalleeName);
if (FS != iter->second.end())
return &FS->second;
if (Remapper) {
if (auto NameInProfile = Remapper->lookUpNameInProfile(CalleeName)) {
auto FS = iter->second.find(*NameInProfile);
if (FS != iter->second.end())
return &FS->second;
}
}
// If we cannot find exact match of the callee name, return the FS with
// the max total count. Only do this when CalleeName is not provided,
// i.e., only for indirect calls.
if (!CalleeName.empty())
return nullptr;
uint64_t MaxTotalSamples = 0;
const FunctionSamples *R = nullptr;
for (const auto &NameFS : iter->second)
if (NameFS.second.getTotalSamples() >= MaxTotalSamples) {
MaxTotalSamples = NameFS.second.getTotalSamples();
R = &NameFS.second;
}
return R;
}
#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
LLVM_DUMP_METHOD void FunctionSamples::dump() const { print(dbgs(), 0); }
#endif
std::error_code ProfileSymbolList::read(const uint8_t *Data,
uint64_t ListSize) {
const char *ListStart = reinterpret_cast<const char *>(Data);
uint64_t Size = 0;
while (Size < ListSize) {
StringRef Str(ListStart + Size);
add(Str);
Size += Str.size() + 1;
}
if (Size != ListSize)
return sampleprof_error::malformed;
return sampleprof_error::success;
}
std::error_code ProfileSymbolList::write(raw_ostream &OS) {
// Sort the symbols before output. If doing compression.
// It will make the compression much more effective.
std::vector<StringRef> SortedList;
SortedList.insert(SortedList.begin(), Syms.begin(), Syms.end());
llvm::sort(SortedList);
std::string OutputString;
for (auto &Sym : SortedList) {
OutputString.append(Sym.str());
OutputString.append(1, '\0');
}
OS << OutputString;
return sampleprof_error::success;
}
void ProfileSymbolList::dump(raw_ostream &OS) const {
OS << "======== Dump profile symbol list ========\n";
std::vector<StringRef> SortedList;
SortedList.insert(SortedList.begin(), Syms.begin(), Syms.end());
llvm::sort(SortedList);
for (auto &Sym : SortedList)
OS << Sym << "\n";
}