Merge branch 'extend_sg_map' into xegpu_simt_2d_to_1d

Author: chencha3

.ci/compute_projects.py

@@ -200,6 +200,11 @@ def _get_modified_projects(modified_files: list[str]) -> Set[str]:
         # documentation builds.
         if len(path_parts) > 2 and path_parts[1] == "docs":
             continue
+        # Exclude files for the gn build. We do not test it within premerge
+        # and changes occur often enough that they otherwise take up
+        # capacity.
+        if len(path_parts) > 3 and path_parts[:3] == ("llvm", "utils", "gn"):
+            continue
         modified_projects.add(pathlib.Path(modified_file).parts[0])
     return modified_projects
 

.ci/compute_projects_test.py

@@ -179,6 +179,15 @@ def test_exclude_docs(self):
         self.assertEqual(env_variables["runtimes_to_build"], "")
         self.assertEqual(env_variables["runtimes_check_targets"], "")
 
+    def test_exclude_gn(self):
+        env_variables = compute_projects.get_env_variables(
+            ["llvm/utils/gn/build/BUILD.gn"], "Linux"
+        )
+        self.assertEqual(env_variables["projects_to_build"], "")
+        self.assertEqual(env_variables["project_check_targets"], "")
+        self.assertEqual(env_variables["runtimes_to_build"], "")
+        self.assertEqual(env_variables["runtimes_check_targets"], "")
+
 
 if __name__ == "__main__":
     unittest.main()

bolt/include/bolt/Core/MCPlusBuilder.h

@@ -577,6 +577,16 @@ class MCPlusBuilder {
     return getNoRegister();
   }
 
+  /// Returns the register used as call destination, or no-register, if not
+  /// an indirect call. Sets IsAuthenticatedInternally if the instruction
+  /// accepts a signed pointer as its operand and authenticates it internally.
+  virtual MCPhysReg
+  getRegUsedAsCallDest(const MCInst &Inst,
+                       bool &IsAuthenticatedInternally) const {
+    llvm_unreachable("not implemented");
+    return getNoRegister();
+  }
+
   virtual bool isTerminator(const MCInst &Inst) const;
 
   virtual bool isNoop(const MCInst &Inst) const {
@@ -1741,6 +1751,15 @@ class MCPlusBuilder {
     return {};
   }
 
+  /// Create a sequence of instructions to compare contents of a register
+  /// \p RegNo to immediate \Imm and jump to \p Target if they are different.
+  virtual InstructionListType createCmpJNE(MCPhysReg RegNo, int64_t Imm,
+                                           const MCSymbol *Target,
+                                           MCContext *Ctx) const {
+    llvm_unreachable("not implemented");
+    return {};
+  }
+
   /// Creates inline memcpy instruction. If \p ReturnEnd is true, then return
   /// (dest + n) instead of dest.
   virtual InstructionListType createInlineMemcpy(bool ReturnEnd) const {

bolt/include/bolt/Core/Relocation.h

@@ -86,6 +86,9 @@ class Relocation {
   /// Adjust value depending on relocation type (make it PC relative or not).
   static uint64_t encodeValue(uint32_t Type, uint64_t Value, uint64_t PC);
 
+  /// Return true if there are enough bits to encode the relocation value.
+  static bool canEncodeValue(uint32_t Type, uint64_t Value, uint64_t PC);
+
   /// Extract current relocated value from binary contents. This is used for
   /// RISC architectures where values are encoded in specific bits depending
   /// on the relocation value. For X86, we limit to sign extending the value

bolt/include/bolt/Passes/PAuthGadgetScanner.h

@@ -248,6 +248,9 @@ struct FunctionAnalysisResult {
 };
 
 class Analysis : public BinaryFunctionPass {
+  /// Only search for pac-ret violations.
+  bool PacRetGadgetsOnly;
+
   void runOnFunction(BinaryFunction &Function,
                      MCPlusBuilder::AllocatorIdTy AllocatorId);
   FunctionAnalysisResult findGadgets(BinaryFunction &BF,
@@ -261,7 +264,8 @@ class Analysis : public BinaryFunctionPass {
   std::mutex AnalysisResultsMutex;
 
 public:
-  explicit Analysis() : BinaryFunctionPass(false) {}
+  explicit Analysis(bool PacRetGadgetsOnly)
+      : BinaryFunctionPass(false), PacRetGadgetsOnly(PacRetGadgetsOnly) {}
 
   const char *getName() const override { return "pauth-gadget-scanner"; }
 

bolt/include/bolt/Utils/CommandLineOpts.h

@@ -81,9 +81,9 @@ extern llvm::cl::opt<unsigned> Verbosity;
 /// Return true if we should process all functions in the binary.
 bool processAllFunctions();
 
-enum GadgetScannerKind { GS_PACRET, GS_ALL };
+enum GadgetScannerKind { GS_PACRET, GS_PAUTH, GS_ALL };
 
-extern llvm::cl::list<GadgetScannerKind> GadgetScannersToRun;
+extern llvm::cl::bits<GadgetScannerKind> GadgetScannersToRun;
 
 } // namespace opts
 

bolt/lib/Core/BinaryFunction.cpp

@@ -1795,6 +1795,8 @@ bool BinaryFunction::scanExternalRefs() {
     // Create relocation for every fixup.
     for (const MCFixup &Fixup : Fixups) {
       std::optional<Relocation> Rel = BC.MIB->createRelocation(Fixup, *BC.MAB);
+      // Can be skipped in case of overlow during relocation value encoding.
+      Rel->setOptional();
       if (!Rel) {
         Success = false;
         continue;

bolt/lib/Core/BinarySection.cpp

@@ -12,6 +12,7 @@
 
 #include "bolt/Core/BinarySection.h"
 #include "bolt/Core/BinaryContext.h"
+#include "bolt/Utils/CommandLineOpts.h"
 #include "bolt/Utils/Utils.h"
 #include "llvm/MC/MCStreamer.h"
 #include "llvm/Support/CommandLine.h"
@@ -22,8 +23,8 @@ using namespace llvm;
 using namespace bolt;
 
 namespace opts {
-extern cl::opt<bool> PrintRelocations;
 extern cl::opt<bool> HotData;
+extern cl::opt<bool> PrintRelocations;
 } // namespace opts
 
 uint64_t BinarySection::Count = 0;
@@ -174,11 +175,30 @@ void BinarySection::flushPendingRelocations(raw_pwrite_stream &OS,
     OS.pwrite(Patch.Bytes.data(), Patch.Bytes.size(),
               SectionFileOffset + Patch.Offset);
 
+  uint64_t SkippedPendingRelocations = 0;
   for (Relocation &Reloc : PendingRelocations) {
     uint64_t Value = Reloc.Addend;
     if (Reloc.Symbol)
       Value += Resolver(Reloc.Symbol);
 
+    // Safely skip any optional pending relocation that cannot be encoded.
+    if (Reloc.isOptional() &&
+        !Relocation::canEncodeValue(Reloc.Type, Value,
+                                    SectionAddress + Reloc.Offset)) {
+
+      // A successful run of 'scanExternalRefs' means that all pending
+      // relocations are flushed. Otherwise, PatchEntries should run.
+      if (!opts::ForcePatch) {
+        BC.errs()
+            << "BOLT-ERROR: cannot encode relocation for symbol "
+            << Reloc.Symbol->getName()
+            << " as it is out-of-range. To proceed must use -force-patch\n";
+        exit(1);
+      }
+
+      ++SkippedPendingRelocations;
+      continue;
+    }
     Value = Relocation::encodeValue(Reloc.Type, Value,
                                     SectionAddress + Reloc.Offset);
 
@@ -197,6 +217,11 @@ void BinarySection::flushPendingRelocations(raw_pwrite_stream &OS,
   }
 
   clearList(PendingRelocations);
+
+  if (SkippedPendingRelocations > 0 && opts::Verbosity >= 1) {
+    BC.outs() << "BOLT-INFO: skipped " << SkippedPendingRelocations
+              << " out-of-range optional relocations\n";
+  }
 }
 
 BinarySection::~BinarySection() { updateContents(nullptr, 0); }

bolt/lib/Core/Relocation.cpp

@@ -271,6 +271,16 @@ static uint64_t encodeValueX86(uint32_t Type, uint64_t Value, uint64_t PC) {
   return Value;
 }
 
+static bool canEncodeValueAArch64(uint32_t Type, uint64_t Value, uint64_t PC) {
+  switch (Type) {
+  default:
+    llvm_unreachable("unsupported relocation");
+  case ELF::R_AARCH64_CALL26:
+  case ELF::R_AARCH64_JUMP26:
+    return isInt<28>(Value - PC);
+  }
+}
+
 static uint64_t encodeValueAArch64(uint32_t Type, uint64_t Value, uint64_t PC) {
   switch (Type) {
   default:
@@ -303,6 +313,16 @@ static uint64_t encodeValueAArch64(uint32_t Type, uint64_t Value, uint64_t PC) {
   return Value;
 }
 
+static uint64_t canEncodeValueRISCV(uint32_t Type, uint64_t Value,
+                                    uint64_t PC) {
+  switch (Type) {
+  default:
+    llvm_unreachable("unsupported relocation");
+  case ELF::R_RISCV_64:
+    return true;
+  }
+}
+
 static uint64_t encodeValueRISCV(uint32_t Type, uint64_t Value, uint64_t PC) {
   switch (Type) {
   default:
@@ -739,6 +759,19 @@ uint64_t Relocation::encodeValue(uint32_t Type, uint64_t Value, uint64_t PC) {
   }
 }
 
+bool Relocation::canEncodeValue(uint32_t Type, uint64_t Value, uint64_t PC) {
+  switch (Arch) {
+  default:
+    llvm_unreachable("Unsupported architecture");
+  case Triple::aarch64:
+    return canEncodeValueAArch64(Type, Value, PC);
+  case Triple::riscv64:
+    return canEncodeValueRISCV(Type, Value, PC);
+  case Triple::x86_64:
+    return true;
+  }
+}
+
 uint64_t Relocation::extractValue(uint32_t Type, uint64_t Contents,
                                   uint64_t PC) {
   switch (Arch) {

bolt/lib/Passes/PAuthGadgetScanner.cpp

@@ -401,11 +401,11 @@ class PacRetAnalysis
 
 public:
   std::vector<MCInstReference>
-  getLastClobberingInsts(const MCInst Ret, BinaryFunction &BF,
-                         const ArrayRef<MCPhysReg> UsedDirtyRegs) const {
+  getLastClobberingInsts(const MCInst &Inst, BinaryFunction &BF,
+                         const ArrayRef<MCPhysReg> UsedDirtyRegs) {
     if (RegsToTrackInstsFor.empty())
       return {};
-    auto MaybeState = getStateAt(Ret);
+    auto MaybeState = getStateBefore(Inst);
     if (!MaybeState)
       llvm_unreachable("Expected State to be present");
     const State &S = *MaybeState;
@@ -453,6 +453,29 @@ shouldReportReturnGadget(const BinaryContext &BC, const MCInstReference &Inst,
   return std::make_shared<GadgetReport>(RetKind, Inst, RetReg);
 }
 
+static std::shared_ptr<Report>
+shouldReportCallGadget(const BinaryContext &BC, const MCInstReference &Inst,
+                       const State &S) {
+  static const GadgetKind CallKind("non-protected call found");
+  if (!BC.MIB->isCall(Inst) && !BC.MIB->isBranch(Inst))
+    return nullptr;
+
+  bool IsAuthenticated = false;
+  MCPhysReg DestReg = BC.MIB->getRegUsedAsCallDest(Inst, IsAuthenticated);
+  if (IsAuthenticated || DestReg == BC.MIB->getNoRegister())
+    return nullptr;
+
+  LLVM_DEBUG({
+    traceInst(BC, "Found call inst", Inst);
+    traceReg(BC, "Call destination reg", DestReg);
+    traceRegMask(BC, "SafeToDerefRegs", S.SafeToDerefRegs);
+  });
+  if (S.SafeToDerefRegs[DestReg])
+    return nullptr;
+
+  return std::make_shared<GadgetReport>(CallKind, Inst, DestReg);
+}
+
 FunctionAnalysisResult
 Analysis::findGadgets(BinaryFunction &BF,
                       MCPlusBuilder::AllocatorIdTy AllocatorId) {
@@ -469,7 +492,7 @@ Analysis::findGadgets(BinaryFunction &BF,
   for (BinaryBasicBlock &BB : BF) {
     for (int64_t I = 0, E = BB.size(); I < E; ++I) {
       MCInstReference Inst(&BB, I);
-      const State &S = *PRA.getStateAt(Inst);
+      const State &S = *PRA.getStateBefore(Inst);
 
       // If non-empty state was never propagated from the entry basic block
       // to Inst, assume it to be unreachable and report a warning.
@@ -481,6 +504,12 @@ Analysis::findGadgets(BinaryFunction &BF,
 
       if (auto Report = shouldReportReturnGadget(BC, Inst, S))
         Result.Diagnostics.push_back(Report);
+
+      if (PacRetGadgetsOnly)
+        continue;
+
+      if (auto Report = shouldReportCallGadget(BC, Inst, S))
+        Result.Diagnostics.push_back(Report);
     }
   }
   return Result;

bolt/lib/Rewrite/RewriteInstance.cpp

@@ -247,12 +247,14 @@ static cl::opt<bool> WriteBoltInfoSection(
     "bolt-info", cl::desc("write bolt info section in the output binary"),
     cl::init(true), cl::Hidden, cl::cat(BoltOutputCategory));
 
-cl::list<GadgetScannerKind>
-    GadgetScannersToRun("scanners", cl::desc("which gadget scanners to run"),
-                        cl::values(clEnumValN(GS_PACRET, "pacret", "pac-ret"),
-                                   clEnumValN(GS_ALL, "all", "all")),
-                        cl::ZeroOrMore, cl::CommaSeparated,
-                        cl::cat(BinaryAnalysisCategory));
+cl::bits<GadgetScannerKind> GadgetScannersToRun(
+    "scanners", cl::desc("which gadget scanners to run"),
+    cl::values(
+        clEnumValN(GS_PACRET, "pacret",
+                   "pac-ret: return address protection (subset of \"pauth\")"),
+        clEnumValN(GS_PAUTH, "pauth", "All Pointer Authentication scanners"),
+        clEnumValN(GS_ALL, "all", "All implemented scanners")),
+    cl::ZeroOrMore, cl::CommaSeparated, cl::cat(BinaryAnalysisCategory));
 
 } // namespace opts
 
@@ -3539,12 +3541,18 @@ void RewriteInstance::runBinaryAnalyses() {
   // FIXME: add a pass that warns about which functions do not have CFG,
   // and therefore, analysis is most likely to be less accurate.
   using GSK = opts::GadgetScannerKind;
-  // if no command line option was given, act as if "all" was specified.
-  if (opts::GadgetScannersToRun.empty())
-    opts::GadgetScannersToRun.addValue(GSK::GS_ALL);
-  for (GSK ScannerToRun : opts::GadgetScannersToRun) {
-    if (ScannerToRun == GSK::GS_PACRET || ScannerToRun == GSK::GS_ALL)
-      Manager.registerPass(std::make_unique<PAuthGadgetScanner::Analysis>());
+  using PAuthScanner = PAuthGadgetScanner::Analysis;
+
+  // If no command line option was given, act as if "all" was specified.
+  bool RunAll = !opts::GadgetScannersToRun.getBits() ||
+                opts::GadgetScannersToRun.isSet(GSK::GS_ALL);
+
+  if (RunAll || opts::GadgetScannersToRun.isSet(GSK::GS_PAUTH)) {
+    Manager.registerPass(
+        std::make_unique<PAuthScanner>(/*OnlyPacRetChecks=*/false));
+  } else if (RunAll || opts::GadgetScannersToRun.isSet(GSK::GS_PACRET)) {
+    Manager.registerPass(
+        std::make_unique<PAuthScanner>(/*OnlyPacRetChecks=*/true));
   }
 
   BC->logBOLTErrorsAndQuitOnFatal(Manager.runPasses());