[llvm-branch-commits] [polly] [Polly] Introduce PhaseManager and remove LPM support (PR #125442)
https://github.com/Meinersbur edited https://github.com/llvm/llvm-project/pull/125442 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [polly] [Polly] Introduce PhaseManager and remove LPM support (PR #125442)
@@ -0,0 +1,419 @@
+//===-- PhaseManager.cpp *- C++
-*-===//
+//
+// 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
+//
+//===--===//
+
+#include "polly/Pass/PhaseManager.h"
+#include "polly/CodeGen/CodeGeneration.h"
+#include "polly/CodeGen/IslAst.h"
+#include "polly/CodePreparation.h"
+#include "polly/DeLICM.h"
+#include "polly/DeadCodeElimination.h"
+#include "polly/DependenceInfo.h"
+#include "polly/FlattenSchedule.h"
+#include "polly/ForwardOpTree.h"
+#include "polly/JSONExporter.h"
+#include "polly/MaximalStaticExpansion.h"
+#include "polly/PruneUnprofitable.h"
+#include "polly/ScheduleOptimizer.h"
+#include "polly/ScopDetection.h"
+#include "polly/ScopDetectionDiagnostic.h"
+#include "polly/ScopGraphPrinter.h"
+#include "polly/ScopInfo.h"
+#include "polly/Simplify.h"
+#include "llvm/Analysis/AssumptionCache.h"
+#include "llvm/Analysis/OptimizationRemarkEmitter.h"
+#include "llvm/IR/Module.h"
+
+#define DEBUG_TYPE "polly-pass"
+
+using namespace polly;
+using namespace llvm;
+
+namespace {
+
+/// Recurse through all subregions and all regions and add them to RQ.
+static void addRegionIntoQueue(Region &R, SmallVector &RQ) {
+ RQ.push_back(&R);
+ for (const auto &E : R)
+addRegionIntoQueue(*E, RQ);
+}
+
+/// The phase pipeline of Polly to be embedded into another pass manager than
+/// runs passes on functions.
+///
+/// Polly holds state besides LLVM-IR (RegionInfo and ScopInfo) between phases
+/// that LLVM pass managers do not consider when scheduling analyses and
passes.
+/// That is, the ScopInfo must persist between phases that a pass manager must
+/// not invalidate to recompute later.
+class PhaseManager {
+private:
+ Function &F;
+ FunctionAnalysisManager &FAM;
+ PollyPassOptions Opts;
+
+public:
+ PhaseManager(Function &F, FunctionAnalysisManager &FAM, PollyPassOptions
Opts)
+ : F(F), FAM(FAM), Opts(std::move(Opts)) {}
+
+ /// Execute Polly's phases as indicated by the options.
+ bool run() {
+// Get analyses from the function pass manager.
+// These must be preserved during all phases so that if processing one SCoP
+// has finished, the next SCoP can still use them. Recomputing is not an
+// option because ScopDetection stores references to the old results.
+// TODO: CodePreparation doesn't actually need these analysis, it just
keeps
+// them up-to-date. If they are not computed yet, can also compute after
the
+// prepare phase.
+auto &LI = FAM.getResult(F);
+auto &DT = FAM.getResult(F);
+bool ModifiedIR = false;
+
+// Phase: prepare
+// TODO: Setting ModifiedIR will invalidate any anlysis, even if DT, LI are
+// preserved.
+if (Opts.isPhaseEnabled(PassPhase::Prepare))
+ ModifiedIR |= runCodePreparation(F, &DT, &LI, nullptr);
+
+// Can't do anything without detection
+if (!Opts.isPhaseEnabled(PassPhase::Detection))
+ return false;
+
+auto &AA = FAM.getResult(F);
+auto &SE = FAM.getResult(F);
+auto &ORE = FAM.getResult(F);
+
+// ScopDetection is modifying RegionInfo, do not cache it, nor use a cached
+// version.
+RegionInfo RI = RegionInfoAnalysis().run(F, FAM);
+
+// Phase: detection
+ScopDetection SD(DT, SE, LI, RI, AA, ORE);
+SD.detect(F);
+if (Opts.isPhaseEnabled(PassPhase::PrintDetect)) {
+ outs() << "Detected Scops in Function " << F.getName() << "\n";
+ for (const Region *R : SD.ValidRegions)
+outs() << "Valid Region for Scop: " << R->getNameStr() << '\n';
+ outs() << "\n";
+}
+
+if (Opts.isPhaseEnabled(PassPhase::DotScops))
+ printGraphForFunction(F, &SD, "scops", false);
+if (Opts.isPhaseEnabled(PassPhase::DotScopsOnly))
+ printGraphForFunction(F, &SD, "scopsonly", true);
+
+auto ViewScops = [&](const char *Name, bool IsSimply) {
+ if (Opts.ViewFilter.empty() && !F.getName().count(Opts.ViewFilter))
+return;
+
+ if (Opts.ViewAll || std::distance(SD.begin(), SD.end()) > 0)
+viewGraphForFunction(F, &SD, Name, IsSimply);
+};
+if (Opts.isPhaseEnabled(PassPhase::ViewScops))
+ ViewScops("scops", false);
+if (Opts.isPhaseEnabled(PassPhase::ViewScopsOnly))
+ ViewScops("scopsonly", true);
+
+// Phase: scops
+auto &AC = FAM.getResult(F);
+const DataLayout &DL = F.getParent()->getDataLayout();
+ScopInfo Info(DL, SD, SE, LI, AA, DT, AC, ORE);
Meinersbur wrote:
Can you clarify? ScopS there iterated over twice:
1. For `-print-scops`: This PR's code reuses the LPM did:
https://github.com/llvm/llvm-project/blob/5a91ecf5f004d9defce3ba5f7b08015a1f2073f9/polly/lib/Pass/PhaseManager.cpp#L131
vs
https://github.com/llvm/llvm-project/blob/4504e775509483ec20912bc1d8057
[llvm-branch-commits] [polly] [Polly] Introduce PhaseManager and remove LPM support (PR #125442)
https://github.com/Meinersbur edited https://github.com/llvm/llvm-project/pull/125442 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [clang-tools-extra] [clang-doc] Extract Info into JSON values (PR #138063)
@@ -162,15 +162,264 @@ Error MustacheHTMLGenerator::generateDocs(
return Error::success();
}
+static json::Value
+extractValue(const Location &L,
+ std::optional RepositoryUrl = std::nullopt) {
+ Object Obj = Object();
+ // Should there be Start/End line numbers?
evelez7 wrote:
outdated comment?
```suggestion
```
https://github.com/llvm/llvm-project/pull/138063
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [clang-tools-extra] [clang-doc] Extract Info into JSON values (PR #138063)
@@ -162,15 +162,264 @@ Error MustacheHTMLGenerator::generateDocs(
return Error::success();
}
+static json::Value
+extractValue(const Location &L,
+ std::optional RepositoryUrl = std::nullopt) {
+ Object Obj = Object();
+ // Should there be Start/End line numbers?
+ Obj.insert({"LineNumber", L.StartLineNumber});
+ Obj.insert({"Filename", L.Filename});
+
+ if (!L.IsFileInRootDir || !RepositoryUrl)
+return Obj;
+ SmallString<128> FileURL(*RepositoryUrl);
+ sys::path::append(FileURL, sys::path::Style::posix, L.Filename);
+ FileURL += "#" + std::to_string(L.StartLineNumber);
+ Obj.insert({"FileURL", FileURL});
+
+ return Obj;
+}
+
+static json::Value extractValue(const Reference &I,
+StringRef CurrentDirectory) {
+ SmallString<64> Path = I.getRelativeFilePath(CurrentDirectory);
+ sys::path::append(Path, I.getFileBaseName() + ".html");
+ sys::path::native(Path, sys::path::Style::posix);
+ Object Obj = Object();
+ Obj.insert({"Link", Path});
+ Obj.insert({"Name", I.Name});
+ Obj.insert({"QualName", I.QualName});
+ Obj.insert({"ID", toHex(toStringRef(I.USR))});
+ return Obj;
+}
+
+static json::Value extractValue(const TypedefInfo &I) {
+ // Not Supported
+ return nullptr;
+}
+
+static json::Value extractValue(const CommentInfo &I) {
+ assert((I.Kind == "BlockCommandComment" || I.Kind == "FullComment" ||
+ I.Kind == "ParagraphComment" || I.Kind == "TextComment") &&
+ "Unknown Comment type in CommentInfo.");
+
+ Object Obj = Object();
+ json::Value Child = Object();
+
+ // TextComment has no children, so return it.
+ if (I.Kind == "TextComment") {
+Obj.insert({"TextComment", I.Text});
+return Obj;
+ }
+
+ // BlockCommandComment needs to generate a Command key.
+ if (I.Kind == "BlockCommandComment")
+Child.getAsObject()->insert({"Command", I.Name});
+
+ // Use the same handling for everything else.
+ // Only valid for:
+ // - BlockCommandComment
+ // - FullComment
+ // - ParagraphComment
+ json::Value ChildArr = Array();
+ auto &CARef = *ChildArr.getAsArray();
+ CARef.reserve(I.Children.size());
+ for (const auto &C : I.Children)
+CARef.emplace_back(extractValue(*C));
+ Child.getAsObject()->insert({"Children", ChildArr});
+ Obj.insert({I.Kind, Child});
+
+ return Obj;
+}
+
+static void maybeInsertLocation(std::optional Loc,
+const ClangDocContext &CDCtx, Object &Obj) {
+ if (!Loc)
+return;
+ Location L = *Loc;
+ Obj.insert({"Location", extractValue(L, CDCtx.RepositoryUrl)});
+}
+
+static void extractDescriptionFromInfo(ArrayRef Descriptions,
+ json::Object &EnumValObj) {
+ if (Descriptions.empty())
+return;
+ json::Value ArrDesc = Array();
+ json::Array &ADescRef = *ArrDesc.getAsArray();
+ for (const CommentInfo &Child : Descriptions)
+ADescRef.emplace_back(extractValue(Child));
+ EnumValObj.insert({"EnumValueComments", ArrDesc});
+}
+
+static json::Value extractValue(const FunctionInfo &I, StringRef ParentInfoDir,
+const ClangDocContext &CDCtx) {
+ Object Obj = Object();
+ Obj.insert({"Name", I.Name});
+ Obj.insert({"ID", toHex(toStringRef(I.USR))});
+ Obj.insert({"Access", getAccessSpelling(I.Access).str()});
+ Obj.insert({"ReturnType", extractValue(I.ReturnType.Type, ParentInfoDir)});
+
+ json::Value ParamArr = Array();
+ for (const auto Val : enumerate(I.Params)) {
+json::Value V = Object();
+auto &VRef = *V.getAsObject();
+VRef.insert({"Name", Val.value().Name});
+VRef.insert({"Type", Val.value().Type.Name});
+VRef.insert({"End", Val.index() + 1 == I.Params.size()});
+ParamArr.getAsArray()->emplace_back(V);
evelez7 wrote:
nit?: Not using the same pattern here as elsewhere where an Array reference is
created to `emplace_back` values. I'm assuming the pattern is to not use
`getAsArray` so much.
https://github.com/llvm/llvm-project/pull/138063
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [clang-tools-extra] [clang-doc] Extract Info into JSON values (PR #138063)
@@ -162,15 +162,264 @@ Error MustacheHTMLGenerator::generateDocs(
return Error::success();
}
+static json::Value
+extractValue(const Location &L,
+ std::optional RepositoryUrl = std::nullopt) {
+ Object Obj = Object();
+ // Should there be Start/End line numbers?
+ Obj.insert({"LineNumber", L.StartLineNumber});
+ Obj.insert({"Filename", L.Filename});
+
+ if (!L.IsFileInRootDir || !RepositoryUrl)
+return Obj;
+ SmallString<128> FileURL(*RepositoryUrl);
+ sys::path::append(FileURL, sys::path::Style::posix, L.Filename);
+ FileURL += "#" + std::to_string(L.StartLineNumber);
+ Obj.insert({"FileURL", FileURL});
+
+ return Obj;
+}
+
+static json::Value extractValue(const Reference &I,
+StringRef CurrentDirectory) {
+ SmallString<64> Path = I.getRelativeFilePath(CurrentDirectory);
+ sys::path::append(Path, I.getFileBaseName() + ".html");
+ sys::path::native(Path, sys::path::Style::posix);
+ Object Obj = Object();
+ Obj.insert({"Link", Path});
+ Obj.insert({"Name", I.Name});
+ Obj.insert({"QualName", I.QualName});
+ Obj.insert({"ID", toHex(toStringRef(I.USR))});
+ return Obj;
+}
+
+static json::Value extractValue(const TypedefInfo &I) {
+ // Not Supported
+ return nullptr;
+}
+
+static json::Value extractValue(const CommentInfo &I) {
+ assert((I.Kind == "BlockCommandComment" || I.Kind == "FullComment" ||
+ I.Kind == "ParagraphComment" || I.Kind == "TextComment") &&
+ "Unknown Comment type in CommentInfo.");
+
+ Object Obj = Object();
+ json::Value Child = Object();
+
+ // TextComment has no children, so return it.
+ if (I.Kind == "TextComment") {
+Obj.insert({"TextComment", I.Text});
+return Obj;
+ }
+
+ // BlockCommandComment needs to generate a Command key.
+ if (I.Kind == "BlockCommandComment")
+Child.getAsObject()->insert({"Command", I.Name});
+
+ // Use the same handling for everything else.
+ // Only valid for:
+ // - BlockCommandComment
+ // - FullComment
+ // - ParagraphComment
+ json::Value ChildArr = Array();
+ auto &CARef = *ChildArr.getAsArray();
+ CARef.reserve(I.Children.size());
+ for (const auto &C : I.Children)
+CARef.emplace_back(extractValue(*C));
+ Child.getAsObject()->insert({"Children", ChildArr});
+ Obj.insert({I.Kind, Child});
+
+ return Obj;
+}
+
+static void maybeInsertLocation(std::optional Loc,
+const ClangDocContext &CDCtx, Object &Obj) {
+ if (!Loc)
+return;
+ Location L = *Loc;
+ Obj.insert({"Location", extractValue(L, CDCtx.RepositoryUrl)});
+}
+
+static void extractDescriptionFromInfo(ArrayRef Descriptions,
+ json::Object &EnumValObj) {
+ if (Descriptions.empty())
+return;
+ json::Value ArrDesc = Array();
+ json::Array &ADescRef = *ArrDesc.getAsArray();
+ for (const CommentInfo &Child : Descriptions)
+ADescRef.emplace_back(extractValue(Child));
+ EnumValObj.insert({"EnumValueComments", ArrDesc});
+}
+
+static json::Value extractValue(const FunctionInfo &I, StringRef ParentInfoDir,
+const ClangDocContext &CDCtx) {
+ Object Obj = Object();
+ Obj.insert({"Name", I.Name});
+ Obj.insert({"ID", toHex(toStringRef(I.USR))});
+ Obj.insert({"Access", getAccessSpelling(I.Access).str()});
+ Obj.insert({"ReturnType", extractValue(I.ReturnType.Type, ParentInfoDir)});
+
+ json::Value ParamArr = Array();
+ for (const auto Val : enumerate(I.Params)) {
+json::Value V = Object();
+auto &VRef = *V.getAsObject();
+VRef.insert({"Name", Val.value().Name});
+VRef.insert({"Type", Val.value().Type.Name});
+VRef.insert({"End", Val.index() + 1 == I.Params.size()});
+ParamArr.getAsArray()->emplace_back(V);
+ }
+ Obj.insert({"Params", ParamArr});
+
+ maybeInsertLocation(I.DefLoc, CDCtx, Obj);
+ return Obj;
+}
+
+static json::Value extractValue(const EnumInfo &I,
+const ClangDocContext &CDCtx) {
+ Object Obj = Object();
+ std::string EnumType = I.Scoped ? "enum class " : "enum ";
+ EnumType += I.Name;
+ bool HasComment = std::any_of(
+ I.Members.begin(), I.Members.end(),
+ [](const EnumValueInfo &M) { return !M.Description.empty(); });
+ Obj.insert({"EnumName", EnumType});
+ Obj.insert({"HasComment", HasComment});
+ Obj.insert({"ID", toHex(toStringRef(I.USR))});
+ json::Value Arr = Array();
evelez7 wrote:
```suggestion
json::Value EnumArr = Array();
```
nit: Arrays in the other `extractValue` functions have type+"Arr" names. Same
for the reference.
https://github.com/llvm/llvm-project/pull/138063
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [clang-tools-extra] [clang-doc] Extract Info into JSON values (PR #138063)
https://github.com/evelez7 edited https://github.com/llvm/llvm-project/pull/138063 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [clang-tools-extra] [clang-doc] Extract Info into JSON values (PR #138063)
@@ -162,15 +162,264 @@ Error MustacheHTMLGenerator::generateDocs(
return Error::success();
}
+static json::Value
+extractValue(const Location &L,
+ std::optional RepositoryUrl = std::nullopt) {
+ Object Obj = Object();
+ // Should there be Start/End line numbers?
+ Obj.insert({"LineNumber", L.StartLineNumber});
+ Obj.insert({"Filename", L.Filename});
+
+ if (!L.IsFileInRootDir || !RepositoryUrl)
+return Obj;
+ SmallString<128> FileURL(*RepositoryUrl);
+ sys::path::append(FileURL, sys::path::Style::posix, L.Filename);
+ FileURL += "#" + std::to_string(L.StartLineNumber);
+ Obj.insert({"FileURL", FileURL});
+
+ return Obj;
+}
+
+static json::Value extractValue(const Reference &I,
+StringRef CurrentDirectory) {
+ SmallString<64> Path = I.getRelativeFilePath(CurrentDirectory);
+ sys::path::append(Path, I.getFileBaseName() + ".html");
+ sys::path::native(Path, sys::path::Style::posix);
+ Object Obj = Object();
+ Obj.insert({"Link", Path});
+ Obj.insert({"Name", I.Name});
+ Obj.insert({"QualName", I.QualName});
+ Obj.insert({"ID", toHex(toStringRef(I.USR))});
+ return Obj;
+}
+
+static json::Value extractValue(const TypedefInfo &I) {
+ // Not Supported
+ return nullptr;
+}
+
+static json::Value extractValue(const CommentInfo &I) {
+ assert((I.Kind == "BlockCommandComment" || I.Kind == "FullComment" ||
+ I.Kind == "ParagraphComment" || I.Kind == "TextComment") &&
+ "Unknown Comment type in CommentInfo.");
+
+ Object Obj = Object();
+ json::Value Child = Object();
+
+ // TextComment has no children, so return it.
+ if (I.Kind == "TextComment") {
+Obj.insert({"TextComment", I.Text});
+return Obj;
+ }
+
+ // BlockCommandComment needs to generate a Command key.
+ if (I.Kind == "BlockCommandComment")
+Child.getAsObject()->insert({"Command", I.Name});
+
+ // Use the same handling for everything else.
+ // Only valid for:
+ // - BlockCommandComment
+ // - FullComment
+ // - ParagraphComment
+ json::Value ChildArr = Array();
+ auto &CARef = *ChildArr.getAsArray();
+ CARef.reserve(I.Children.size());
+ for (const auto &C : I.Children)
+CARef.emplace_back(extractValue(*C));
+ Child.getAsObject()->insert({"Children", ChildArr});
+ Obj.insert({I.Kind, Child});
+
+ return Obj;
+}
+
+static void maybeInsertLocation(std::optional Loc,
+const ClangDocContext &CDCtx, Object &Obj) {
+ if (!Loc)
+return;
+ Location L = *Loc;
+ Obj.insert({"Location", extractValue(L, CDCtx.RepositoryUrl)});
+}
+
+static void extractDescriptionFromInfo(ArrayRef Descriptions,
+ json::Object &EnumValObj) {
+ if (Descriptions.empty())
+return;
+ json::Value ArrDesc = Array();
+ json::Array &ADescRef = *ArrDesc.getAsArray();
+ for (const CommentInfo &Child : Descriptions)
+ADescRef.emplace_back(extractValue(Child));
+ EnumValObj.insert({"EnumValueComments", ArrDesc});
+}
+
+static json::Value extractValue(const FunctionInfo &I, StringRef ParentInfoDir,
+const ClangDocContext &CDCtx) {
+ Object Obj = Object();
+ Obj.insert({"Name", I.Name});
+ Obj.insert({"ID", toHex(toStringRef(I.USR))});
+ Obj.insert({"Access", getAccessSpelling(I.Access).str()});
+ Obj.insert({"ReturnType", extractValue(I.ReturnType.Type, ParentInfoDir)});
+
+ json::Value ParamArr = Array();
+ for (const auto Val : enumerate(I.Params)) {
+json::Value V = Object();
+auto &VRef = *V.getAsObject();
+VRef.insert({"Name", Val.value().Name});
+VRef.insert({"Type", Val.value().Type.Name});
+VRef.insert({"End", Val.index() + 1 == I.Params.size()});
+ParamArr.getAsArray()->emplace_back(V);
+ }
+ Obj.insert({"Params", ParamArr});
+
+ maybeInsertLocation(I.DefLoc, CDCtx, Obj);
+ return Obj;
+}
+
+static json::Value extractValue(const EnumInfo &I,
+const ClangDocContext &CDCtx) {
+ Object Obj = Object();
+ std::string EnumType = I.Scoped ? "enum class " : "enum ";
+ EnumType += I.Name;
+ bool HasComment = std::any_of(
+ I.Members.begin(), I.Members.end(),
+ [](const EnumValueInfo &M) { return !M.Description.empty(); });
+ Obj.insert({"EnumName", EnumType});
+ Obj.insert({"HasComment", HasComment});
+ Obj.insert({"ID", toHex(toStringRef(I.USR))});
+ json::Value Arr = Array();
+ json::Array &ARef = *Arr.getAsArray();
+ for (const EnumValueInfo &M : I.Members) {
+json::Value EnumValue = Object();
+auto &EnumValObj = *EnumValue.getAsObject();
+EnumValObj.insert({"Name", M.Name});
+if (!M.ValueExpr.empty())
+ EnumValObj.insert({"ValueExpr", M.ValueExpr});
+else
+ EnumValObj.insert({"Value", M.Value});
+
+extractDescriptionFromInfo(M.Description, EnumValObj);
+ARef.emplace_back(EnumValue);
+ }
+ Obj.insert({"EnumValues", Arr});
+
+ extractDescriptionFromInfo(I.Description, Obj);
+ maybeInsertLocation(I.DefLoc,
[llvm-branch-commits] [clang-tools-extra] [clang-doc] Extract Info into JSON values (PR #138063)
https://github.com/evelez7 edited https://github.com/llvm/llvm-project/pull/138063 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [clang-tools-extra] [clang-doc] Extract Info into JSON values (PR #138063)
@@ -162,15 +162,264 @@ Error MustacheHTMLGenerator::generateDocs(
return Error::success();
}
+static json::Value
+extractValue(const Location &L,
+ std::optional RepositoryUrl = std::nullopt) {
+ Object Obj = Object();
+ // Should there be Start/End line numbers?
+ Obj.insert({"LineNumber", L.StartLineNumber});
+ Obj.insert({"Filename", L.Filename});
+
+ if (!L.IsFileInRootDir || !RepositoryUrl)
+return Obj;
+ SmallString<128> FileURL(*RepositoryUrl);
+ sys::path::append(FileURL, sys::path::Style::posix, L.Filename);
+ FileURL += "#" + std::to_string(L.StartLineNumber);
+ Obj.insert({"FileURL", FileURL});
+
+ return Obj;
+}
+
+static json::Value extractValue(const Reference &I,
+StringRef CurrentDirectory) {
+ SmallString<64> Path = I.getRelativeFilePath(CurrentDirectory);
+ sys::path::append(Path, I.getFileBaseName() + ".html");
+ sys::path::native(Path, sys::path::Style::posix);
+ Object Obj = Object();
+ Obj.insert({"Link", Path});
+ Obj.insert({"Name", I.Name});
+ Obj.insert({"QualName", I.QualName});
+ Obj.insert({"ID", toHex(toStringRef(I.USR))});
+ return Obj;
+}
+
+static json::Value extractValue(const TypedefInfo &I) {
+ // Not Supported
+ return nullptr;
+}
+
+static json::Value extractValue(const CommentInfo &I) {
+ assert((I.Kind == "BlockCommandComment" || I.Kind == "FullComment" ||
+ I.Kind == "ParagraphComment" || I.Kind == "TextComment") &&
+ "Unknown Comment type in CommentInfo.");
+
+ Object Obj = Object();
+ json::Value Child = Object();
+
+ // TextComment has no children, so return it.
+ if (I.Kind == "TextComment") {
+Obj.insert({"TextComment", I.Text});
+return Obj;
+ }
+
+ // BlockCommandComment needs to generate a Command key.
+ if (I.Kind == "BlockCommandComment")
+Child.getAsObject()->insert({"Command", I.Name});
+
+ // Use the same handling for everything else.
+ // Only valid for:
+ // - BlockCommandComment
+ // - FullComment
+ // - ParagraphComment
+ json::Value ChildArr = Array();
+ auto &CARef = *ChildArr.getAsArray();
+ CARef.reserve(I.Children.size());
+ for (const auto &C : I.Children)
+CARef.emplace_back(extractValue(*C));
+ Child.getAsObject()->insert({"Children", ChildArr});
+ Obj.insert({I.Kind, Child});
+
+ return Obj;
+}
+
+static void maybeInsertLocation(std::optional Loc,
+const ClangDocContext &CDCtx, Object &Obj) {
+ if (!Loc)
+return;
+ Location L = *Loc;
+ Obj.insert({"Location", extractValue(L, CDCtx.RepositoryUrl)});
+}
+
+static void extractDescriptionFromInfo(ArrayRef Descriptions,
+ json::Object &EnumValObj) {
+ if (Descriptions.empty())
+return;
+ json::Value ArrDesc = Array();
+ json::Array &ADescRef = *ArrDesc.getAsArray();
+ for (const CommentInfo &Child : Descriptions)
+ADescRef.emplace_back(extractValue(Child));
+ EnumValObj.insert({"EnumValueComments", ArrDesc});
+}
+
+static json::Value extractValue(const FunctionInfo &I, StringRef ParentInfoDir,
+const ClangDocContext &CDCtx) {
+ Object Obj = Object();
+ Obj.insert({"Name", I.Name});
+ Obj.insert({"ID", toHex(toStringRef(I.USR))});
+ Obj.insert({"Access", getAccessSpelling(I.Access).str()});
+ Obj.insert({"ReturnType", extractValue(I.ReturnType.Type, ParentInfoDir)});
+
+ json::Value ParamArr = Array();
+ for (const auto Val : enumerate(I.Params)) {
+json::Value V = Object();
+auto &VRef = *V.getAsObject();
+VRef.insert({"Name", Val.value().Name});
+VRef.insert({"Type", Val.value().Type.Name});
+VRef.insert({"End", Val.index() + 1 == I.Params.size()});
+ParamArr.getAsArray()->emplace_back(V);
+ }
+ Obj.insert({"Params", ParamArr});
+
+ maybeInsertLocation(I.DefLoc, CDCtx, Obj);
+ return Obj;
+}
+
+static json::Value extractValue(const EnumInfo &I,
+const ClangDocContext &CDCtx) {
+ Object Obj = Object();
+ std::string EnumType = I.Scoped ? "enum class " : "enum ";
+ EnumType += I.Name;
+ bool HasComment = std::any_of(
+ I.Members.begin(), I.Members.end(),
+ [](const EnumValueInfo &M) { return !M.Description.empty(); });
+ Obj.insert({"EnumName", EnumType});
+ Obj.insert({"HasComment", HasComment});
+ Obj.insert({"ID", toHex(toStringRef(I.USR))});
+ json::Value Arr = Array();
+ json::Array &ARef = *Arr.getAsArray();
+ for (const EnumValueInfo &M : I.Members) {
+json::Value EnumValue = Object();
+auto &EnumValObj = *EnumValue.getAsObject();
+EnumValObj.insert({"Name", M.Name});
+if (!M.ValueExpr.empty())
+ EnumValObj.insert({"ValueExpr", M.ValueExpr});
+else
+ EnumValObj.insert({"Value", M.Value});
+
+extractDescriptionFromInfo(M.Description, EnumValObj);
+ARef.emplace_back(EnumValue);
+ }
+ Obj.insert({"EnumValues", Arr});
+
+ extractDescriptionFromInfo(I.Description, Obj);
+ maybeInsertLocation(I.DefLoc,
[llvm-branch-commits] Implement src:*=sanitize for UBSan. (PR #140489)
https://github.com/qinkunbao created https://github.com/llvm/llvm-project/pull/140489 None ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] Implement src:*=sanitize for UBSan. (PR #140489)
github-actions[bot] wrote: :warning: C/C++ code formatter, clang-format found issues in your code. :warning: You can test this locally with the following command: ``bash git-clang-format --diff HEAD~1 HEAD --extensions h,cpp -- clang/include/clang/Basic/SanitizerSpecialCaseList.h clang/lib/Basic/NoSanitizeList.cpp clang/lib/Basic/SanitizerSpecialCaseList.cpp `` View the diff from clang-format here. ``diff diff --git a/clang/include/clang/Basic/SanitizerSpecialCaseList.h b/clang/include/clang/Basic/SanitizerSpecialCaseList.h index dd01a786d..a77f63d7e 100644 --- a/clang/include/clang/Basic/SanitizerSpecialCaseList.h +++ b/clang/include/clang/Basic/SanitizerSpecialCaseList.h @@ -44,9 +44,9 @@ public: StringRef Category = StringRef()) const; // Query ignorelisted entries if any bit in Mask matches the entry's section. - // Return 0 if not found. If found, return the line number (starts with 1). + // Return 0 if not found. If found, return the line number (starts with 1). unsigned inSectionBlame(SanitizerMask Mask, StringRef Prefix, StringRef Query, - StringRef Category = StringRef()) const; + StringRef Category = StringRef()) const; protected: // Initialize SanitizerSections. `` https://github.com/llvm/llvm-project/pull/140489 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [clang] Implement src:*=sanitize for UBSan. (PR #140489)
https://github.com/qinkunbao updated
https://github.com/llvm/llvm-project/pull/140489
>From d383fc3d23c0c302d134a76d39491c87547526a1 Mon Sep 17 00:00:00 2001
From: Qinkun Bao
Date: Mon, 19 May 2025 02:45:30 +
Subject: [PATCH] fix format
Created using spr 1.3.6
---
clang/include/clang/Basic/SanitizerSpecialCaseList.h | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/clang/include/clang/Basic/SanitizerSpecialCaseList.h
b/clang/include/clang/Basic/SanitizerSpecialCaseList.h
index dd01a786dee01..25d518e7128cf 100644
--- a/clang/include/clang/Basic/SanitizerSpecialCaseList.h
+++ b/clang/include/clang/Basic/SanitizerSpecialCaseList.h
@@ -44,9 +44,9 @@ class SanitizerSpecialCaseList : public llvm::SpecialCaseList
{
StringRef Category = StringRef()) const;
// Query ignorelisted entries if any bit in Mask matches the entry's section.
- // Return 0 if not found. If found, return the line number (starts with 1).
+ // Return 0 if not found. If found, return the line number (starts with 1).
unsigned inSectionBlame(SanitizerMask Mask, StringRef Prefix, StringRef
Query,
- StringRef Category = StringRef()) const;
+ StringRef Category = StringRef()) const;
protected:
// Initialize SanitizerSections.
@@ -54,7 +54,7 @@ class SanitizerSpecialCaseList : public llvm::SpecialCaseList
{
struct SanitizerSection {
SanitizerSection(SanitizerMask SM, SectionEntries &E)
-: Mask(SM), Entries(E){};
+: Mask(SM), Entries(E) {};
SanitizerMask Mask;
SectionEntries &Entries;
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] Implement src:*=sanitize for UBSan. (PR #140489)
llvmbot wrote:
@llvm/pr-subscribers-clang
Author: Qinkun Bao (qinkunbao)
Changes
---
Full diff: https://github.com/llvm/llvm-project/pull/140489.diff
4 Files Affected:
- (modified) clang/include/clang/Basic/SanitizerSpecialCaseList.h (+5)
- (modified) clang/lib/Basic/NoSanitizeList.cpp (+7)
- (modified) clang/lib/Basic/SanitizerSpecialCaseList.cpp (+16-5)
- (added) clang/test/CodeGen/ubsan-src-ignorelist-category.test (+37)
``diff
diff --git a/clang/include/clang/Basic/SanitizerSpecialCaseList.h
b/clang/include/clang/Basic/SanitizerSpecialCaseList.h
index d024b7dfc2e85..dd01a786dee01 100644
--- a/clang/include/clang/Basic/SanitizerSpecialCaseList.h
+++ b/clang/include/clang/Basic/SanitizerSpecialCaseList.h
@@ -43,6 +43,11 @@ class SanitizerSpecialCaseList : public
llvm::SpecialCaseList {
bool inSection(SanitizerMask Mask, StringRef Prefix, StringRef Query,
StringRef Category = StringRef()) const;
+ // Query ignorelisted entries if any bit in Mask matches the entry's section.
+ // Return 0 if not found. If found, return the line number (starts with 1).
+ unsigned inSectionBlame(SanitizerMask Mask, StringRef Prefix, StringRef
Query,
+ StringRef Category = StringRef()) const;
+
protected:
// Initialize SanitizerSections.
void createSanitizerSections();
diff --git a/clang/lib/Basic/NoSanitizeList.cpp
b/clang/lib/Basic/NoSanitizeList.cpp
index e7e63c1f419e6..811480f914ec5 100644
--- a/clang/lib/Basic/NoSanitizeList.cpp
+++ b/clang/lib/Basic/NoSanitizeList.cpp
@@ -44,6 +44,13 @@ bool NoSanitizeList::containsFunction(SanitizerMask Mask,
bool NoSanitizeList::containsFile(SanitizerMask Mask, StringRef FileName,
StringRef Category) const {
+ unsigned nosanline = SSCL->inSectionBlame(Mask, "src", FileName, Category);
+ unsigned sanline = SSCL->inSectionBlame(Mask, "src", FileName, "sanitize");
+ // If we have two cases such as `src:a.cpp=sanitize` and `src:a.cpp`, the
+ // current entry override the previous entry.
+ if (nosanline > 0 && sanline > 0) {
+return nosanline > sanline;
+ }
return SSCL->inSection(Mask, "src", FileName, Category);
}
diff --git a/clang/lib/Basic/SanitizerSpecialCaseList.cpp
b/clang/lib/Basic/SanitizerSpecialCaseList.cpp
index 2dbf04c6ede97..3bf79876235db 100644
--- a/clang/lib/Basic/SanitizerSpecialCaseList.cpp
+++ b/clang/lib/Basic/SanitizerSpecialCaseList.cpp
@@ -56,10 +56,21 @@ void SanitizerSpecialCaseList::createSanitizerSections() {
bool SanitizerSpecialCaseList::inSection(SanitizerMask Mask, StringRef Prefix,
StringRef Query,
StringRef Category) const {
- for (auto &S : SanitizerSections)
-if ((S.Mask & Mask) &&
-SpecialCaseList::inSectionBlame(S.Entries, Prefix, Query, Category))
- return true;
+ return inSectionBlame(Mask, Prefix, Query, Category) > 0;
+}
- return false;
+unsigned SanitizerSpecialCaseList::inSectionBlame(SanitizerMask Mask,
+ StringRef Prefix,
+ StringRef Query,
+ StringRef Category) const {
+ for (auto &S : SanitizerSections) {
+if (S.Mask & Mask) {
+ unsigned lineNum =
+ SpecialCaseList::inSectionBlame(S.Entries, Prefix, Query, Category);
+ if (lineNum > 0) {
+return lineNum;
+ }
+}
+ }
+ return 0;
}
diff --git a/clang/test/CodeGen/ubsan-src-ignorelist-category.test
b/clang/test/CodeGen/ubsan-src-ignorelist-category.test
new file mode 100644
index 0..f32dc5cbb9e13
--- /dev/null
+++ b/clang/test/CodeGen/ubsan-src-ignorelist-category.test
@@ -0,0 +1,37 @@
+// RUN: rm -rf %t
+// RUN: split-file %s %t
+// RUN: %clang_cc1 -triple x86_64-linux-gnu -fsanitize=signed-integer-overflow
-fsanitize-ignorelist=%t/src.ignorelist -emit-llvm %t/test1.c -o - | FileCheck
%s -check-prefix=CHECK-ALLOWLIST
+// RUN: %clang_cc1 -triple x86_64-linux-gnu -fsanitize=signed-integer-overflow
-fsanitize-ignorelist=%t/src.ignorelist -emit-llvm %t/test2.c -o - | FileCheck
%s -check-prefix=CHECK-IGNORELIST
+// RUN: %clang_cc1 -triple x86_64-linux-gnu -fsanitize=signed-integer-overflow
-fsanitize-ignorelist=%t/src.ignorelist.contradict1 -emit-llvm %t/test1.c -o -
| FileCheck %s -check-prefix=CHECK-ALLOWLISTOVERIDE1
+// RUN: %clang_cc1 -triple x86_64-linux-gnu -fsanitize=signed-integer-overflow
-fsanitize-ignorelist=%t/src.ignorelist.contradict2 -emit-llvm %t/test1.c -o -
| FileCheck %s -check-prefix=CHECK-ALLOWLISTOVERIDE2
+
+
+// Verify ubsan only emits checks for files in the allowlist
+
+//--- src.ignorelist
+src:*
+src:*/test1.c=sanitize
+
+//--- src.ignorelist.contradict1
+src:*
+src:*/test1.c=sanitize
+src:*/test1.c
+
+//--- src.ignorelist.contradict1
+src:*
+src:*/test1.c
+src:*/test1.c=sanitize
+
+//--- test1.c
+int add1(int a, int b) {
+//
[llvm-branch-commits] [clang] Implement src:*=sanitize for UBSan. (PR #140489)
https://github.com/qinkunbao edited https://github.com/llvm/llvm-project/pull/140489 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] Implement src:*=sanitize for UBSan. (PR #140489)
https://github.com/qinkunbao edited https://github.com/llvm/llvm-project/pull/140489 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [clang] Implement src:*=sanitize for UBSan. (PR #140489)
https://github.com/qinkunbao converted_to_draft https://github.com/llvm/llvm-project/pull/140489 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] [polly] [Polly] Introduce PhaseManager and remove LPM support (PR #125442)
@@ -0,0 +1,419 @@
+//===-- PhaseManager.cpp *- C++
-*-===//
+//
+// 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
+//
+//===--===//
+
+#include "polly/Pass/PhaseManager.h"
+#include "polly/CodeGen/CodeGeneration.h"
+#include "polly/CodeGen/IslAst.h"
+#include "polly/CodePreparation.h"
+#include "polly/DeLICM.h"
+#include "polly/DeadCodeElimination.h"
+#include "polly/DependenceInfo.h"
+#include "polly/FlattenSchedule.h"
+#include "polly/ForwardOpTree.h"
+#include "polly/JSONExporter.h"
+#include "polly/MaximalStaticExpansion.h"
+#include "polly/PruneUnprofitable.h"
+#include "polly/ScheduleOptimizer.h"
+#include "polly/ScopDetection.h"
+#include "polly/ScopDetectionDiagnostic.h"
+#include "polly/ScopGraphPrinter.h"
+#include "polly/ScopInfo.h"
+#include "polly/Simplify.h"
+#include "llvm/Analysis/AssumptionCache.h"
+#include "llvm/Analysis/OptimizationRemarkEmitter.h"
+#include "llvm/IR/Module.h"
+
+#define DEBUG_TYPE "polly-pass"
+
+using namespace polly;
+using namespace llvm;
+
+namespace {
+
+/// Recurse through all subregions and all regions and add them to RQ.
+static void addRegionIntoQueue(Region &R, SmallVector &RQ) {
+ RQ.push_back(&R);
+ for (const auto &E : R)
+addRegionIntoQueue(*E, RQ);
+}
+
+/// The phase pipeline of Polly to be embedded into another pass manager than
+/// runs passes on functions.
+///
+/// Polly holds state besides LLVM-IR (RegionInfo and ScopInfo) between phases
+/// that LLVM pass managers do not consider when scheduling analyses and
passes.
+/// That is, the ScopInfo must persist between phases that a pass manager must
+/// not invalidate to recompute later.
+class PhaseManager {
+private:
+ Function &F;
+ FunctionAnalysisManager &FAM;
+ PollyPassOptions Opts;
+
+public:
+ PhaseManager(Function &F, FunctionAnalysisManager &FAM, PollyPassOptions
Opts)
+ : F(F), FAM(FAM), Opts(std::move(Opts)) {}
+
+ /// Execute Polly's phases as indicated by the options.
+ bool run() {
+// Get analyses from the function pass manager.
+// These must be preserved during all phases so that if processing one SCoP
+// has finished, the next SCoP can still use them. Recomputing is not an
+// option because ScopDetection stores references to the old results.
+// TODO: CodePreparation doesn't actually need these analysis, it just
keeps
+// them up-to-date. If they are not computed yet, can also compute after
the
+// prepare phase.
+auto &LI = FAM.getResult(F);
+auto &DT = FAM.getResult(F);
+bool ModifiedIR = false;
+
+// Phase: prepare
+// TODO: Setting ModifiedIR will invalidate any anlysis, even if DT, LI are
+// preserved.
+if (Opts.isPhaseEnabled(PassPhase::Prepare))
+ ModifiedIR |= runCodePreparation(F, &DT, &LI, nullptr);
+
Meinersbur wrote:
I agree, analyses that CodePreparation does not preserve should be barred from
being used in later phases.
https://github.com/llvm/llvm-project/pull/125442
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] fix format in SpecialCaseList.cpp (PR #140479)
https://github.com/qinkunbao created https://github.com/llvm/llvm-project/pull/140479 None ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] fix format in SpecialCaseList.cpp (PR #140479)
https://github.com/qinkunbao closed https://github.com/llvm/llvm-project/pull/140479 ___ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
[llvm-branch-commits] fix format in SpecialCaseList.cpp (PR #140479)
llvmbot wrote:
@llvm/pr-subscribers-llvm-support
Author: Qinkun Bao (qinkunbao)
Changes
---
Full diff: https://github.com/llvm/llvm-project/pull/140479.diff
1 Files Affected:
- (modified) llvm/unittests/Support/SpecialCaseListTest.cpp (-1)
``diff
diff --git a/llvm/unittests/Support/SpecialCaseListTest.cpp
b/llvm/unittests/Support/SpecialCaseListTest.cpp
index 15dc0222a57c3..f9185ea7483ca 100644
--- a/llvm/unittests/Support/SpecialCaseListTest.cpp
+++ b/llvm/unittests/Support/SpecialCaseListTest.cpp
@@ -325,5 +325,4 @@ TEST_F(SpecialCaseListTest, Version3) {
EXPECT_EQ(4u, SCL->inSectionBlame("sect1", "fun", "barz"));
EXPECT_EQ(6u, SCL->inSectionBlame("sect2", "fun", "def"));
}
-
}
``
https://github.com/llvm/llvm-project/pull/140479
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
