[memprof] Add access checks to PortableMemInfoBlock::get*

  commit 4c8ec8f
  Author: Kazu Hirata <[email protected]>
  Date:   Wed Apr 24 16:25:35 2024 -0700

introduced the idea of serializing/deserializing a subset of the
fields in PortableMemInfoBlock.  While it reduces the size of the
indexed MemProf profile file, we now could inadvertently access
unavailable fields and go without noticing.

To protect ourselves from the risk, this patch adds access checks to
PortableMemInfoBlock::get* methods by embedding a bit set representing
available fields into PortableMemInfoBlock.

Author: kazutakahirata

llvm/include/llvm/ProfileData/MemProf.h

@@ -2,6 +2,7 @@
 #define LLVM_PROFILEDATA_MEMPROF_H_
 
 #include "llvm/ADT/MapVector.h"
+#include "llvm/ADT/STLForwardCompat.h"
 #include "llvm/ADT/STLFunctionalExtras.h"
 #include "llvm/ADT/SmallVector.h"
 #include "llvm/IR/GlobalValue.h"
@@ -10,6 +11,7 @@
 #include "llvm/Support/EndianStream.h"
 #include "llvm/Support/raw_ostream.h"
 
+#include <bitset>
 #include <cstdint>
 #include <optional>
 
@@ -55,7 +57,10 @@ MemProfSchema getHotColdSchema();
 // deserialize methods.
 struct PortableMemInfoBlock {
   PortableMemInfoBlock() = default;
-  explicit PortableMemInfoBlock(const MemInfoBlock &Block) {
+  explicit PortableMemInfoBlock(const MemInfoBlock &Block,
+                                const MemProfSchema &IncomingSchema) {
+    for (const Meta Id : IncomingSchema)
+      Schema.set(llvm::to_underlying(Id));
 #define MIBEntryDef(NameTag, Name, Type) Name = Block.Name;
 #include "llvm/ProfileData/MIBEntryDef.inc"
 #undef MIBEntryDef
@@ -67,10 +72,12 @@ struct PortableMemInfoBlock {
 
   // Read the contents of \p Ptr based on the \p Schema to populate the
   // MemInfoBlock member.
-  void deserialize(const MemProfSchema &Schema, const unsigned char *Ptr) {
+  void deserialize(const MemProfSchema &IncomingSchema,
+                   const unsigned char *Ptr) {
     using namespace support;
 
-    for (const Meta Id : Schema) {
+    Schema.reset();
+    for (const Meta Id : IncomingSchema) {
       switch (Id) {
 #define MIBEntryDef(NameTag, Name, Type)                                       \
   case Meta::Name: {                                                           \
@@ -82,6 +89,8 @@ struct PortableMemInfoBlock {
         llvm_unreachable("Unknown meta type id, is the profile collected from "
                          "a newer version of the runtime?");
       }
+
+      Schema.set(llvm::to_underlying(Id));
     }
   }
 
@@ -116,15 +125,22 @@ struct PortableMemInfoBlock {
 
   // Define getters for each type which can be called by analyses.
 #define MIBEntryDef(NameTag, Name, Type)                                       \
-  Type get##Name() const { return Name; }
+  Type get##Name() const {                                                     \
+    assert(Schema[llvm::to_underlying(Meta::Name)]);                           \
+    return Name;                                                               \
+  }
 #include "llvm/ProfileData/MIBEntryDef.inc"
 #undef MIBEntryDef
 
   void clear() { *this = PortableMemInfoBlock(); }
 
   bool operator==(const PortableMemInfoBlock &Other) const {
+    if (Other.Schema != Schema)
+      return false;
+
 #define MIBEntryDef(NameTag, Name, Type)                                       \
-  if (Other.get##Name() != get##Name())                                        \
+  if (Schema[llvm::to_underlying(Meta::Name)] &&                               \
+      Other.get##Name() != get##Name())                                        \
     return false;
 #include "llvm/ProfileData/MIBEntryDef.inc"
 #undef MIBEntryDef
@@ -155,6 +171,9 @@ struct PortableMemInfoBlock {
   }
 
 private:
+  // The set of available fields, indexed by Meta::Name.
+  std::bitset<llvm::to_underlying(Meta::Size)> Schema;
+
 #define MIBEntryDef(NameTag, Name, Type) Type Name = Type();
 #include "llvm/ProfileData/MIBEntryDef.inc"
 #undef MIBEntryDef
@@ -296,8 +315,9 @@ struct IndexedAllocationInfo {
 
   IndexedAllocationInfo() = default;
   IndexedAllocationInfo(ArrayRef<FrameId> CS, CallStackId CSId,
-                        const MemInfoBlock &MB)
-      : CallStack(CS.begin(), CS.end()), CSId(CSId), Info(MB) {}
+                        const MemInfoBlock &MB,
+                        const MemProfSchema &Schema = getFullSchema())
+      : CallStack(CS.begin(), CS.end()), CSId(CSId), Info(MB, Schema) {}
 
   // Returns the size in bytes when this allocation info struct is serialized.
   size_t serializedSize(const MemProfSchema &Schema,

llvm/unittests/ProfileData/InstrProfTest.cpp

@@ -407,13 +407,13 @@ IndexedMemProfRecord makeRecord(
 IndexedMemProfRecord
 makeRecordV2(std::initializer_list<::llvm::memprof::CallStackId> AllocFrames,
              std::initializer_list<::llvm::memprof::CallStackId> CallSiteFrames,
-             const MemInfoBlock &Block) {
+             const MemInfoBlock &Block, const memprof::MemProfSchema &Schema) {
   llvm::memprof::IndexedMemProfRecord MR;
   for (const auto &CSId : AllocFrames)
     // We don't populate IndexedAllocationInfo::CallStack because we use it only
     // in Version0 and Version1.
     MR.AllocSites.emplace_back(::llvm::SmallVector<memprof::FrameId>(), CSId,
-                               Block);
+                               Block, Schema);
   for (const auto &CSId : CallSiteFrames)
     MR.CallSiteIds.push_back(CSId);
   return MR;
@@ -506,7 +506,7 @@ TEST_F(InstrProfTest, test_memprof_v2_full_schema) {
 
   const IndexedMemProfRecord IndexedMR = makeRecordV2(
       /*AllocFrames=*/{0x111, 0x222},
-      /*CallSiteFrames=*/{0x333}, MIB);
+      /*CallSiteFrames=*/{0x333}, MIB, memprof::getFullSchema());
   const FrameIdMapTy IdToFrameMap = getFrameMapping();
   const auto CSIdToCallStackMap = getCallStackMapping();
   for (const auto &I : IdToFrameMap) {
@@ -548,7 +548,7 @@ TEST_F(InstrProfTest, test_memprof_v2_partial_schema) {
 
   const IndexedMemProfRecord IndexedMR = makeRecordV2(
       /*AllocFrames=*/{0x111, 0x222},
-      /*CallSiteFrames=*/{0x333}, MIB);
+      /*CallSiteFrames=*/{0x333}, MIB, memprof::getHotColdSchema());
   const FrameIdMapTy IdToFrameMap = getFrameMapping();
   const auto CSIdToCallStackMap = getCallStackMapping();
   for (const auto &I : IdToFrameMap) {

llvm/unittests/ProfileData/MemProfTest.cpp

@@ -241,7 +241,7 @@ TEST(MemProf, PortableWrapper) {
                     /*dealloc_cpu=*/4);
 
   const auto Schema = llvm::memprof::getFullSchema();
-  PortableMemInfoBlock WriteBlock(Info);
+  PortableMemInfoBlock WriteBlock(Info, Schema);
 
   std::string Buffer;
   llvm::raw_string_ostream OS(Buffer);