[cdac] Begin adding MethodDesc APIs to the RuntimeTypeSystem contract (…

…dotnet#104811)

This is largely a placeholder just to get the MethodDesc/MethodDescChunk infrastructure in place. To further fill out ValidateMethodDesc and to implement the GetMethodDescData DAC API, I also need to add an executable code manager contract for mapping between native code pointer values and methods. I'd like to do that as a separate PR.

Debugged WinDbg enough to verify that GetMethodDescData is implemented enough to extract the correct MethodTable pointer value from a MethodDesc.

* start GetMethodDescDataImpl

* add MethodDesc and MethodDescChunk

* checkpoint: MethodDesc validation

* update contract

* fix RuntimeTypeSystem unit tests

   mock the additional data and globals

* fix GetMethodDescChunkPointerMayThrow

* add data descriptor description to the contract

* Apply suggestions from code review

Co-authored-by: Elinor Fung <[email protected]>

* MayThrow -> Throwing

* Slot is ushort not byte

* remove unused property

* add TargetPointer 32-/64-bit max constants

* use NewArrayHolder

* spelling

* add globals to RTS contract

* remove unused usings

* constexpr cdac_offsets, not const

* Apply suggestions from code review

Co-authored-by: Elinor Fung <[email protected]>

* make GetNumVtableSlots private

---------

Co-authored-by: Elinor Fung <[email protected]>

docs/design/datacontracts/RuntimeTypeSystem.md

@@ -4,10 +4,11 @@ This contract is for exploring the properties of the runtime types of values on
 
 ## APIs of contract
 
+### TypeHandle
+
 A `TypeHandle` is the runtime representation of the type information about a value which is represented as a TypeHandle.
 Given a `TargetPointer` address, the `RuntimeTypeSystem` contract provides a `TypeHandle` for querying the details of the `TypeHandle`.
 
-
 ``` csharp
 struct TypeHandle
 {
@@ -28,6 +29,8 @@ internal enum CorElementType
 A `TypeHandle` is the runtime representation of the type information about a value.  This can be constructed from the address of a `TypeHandle` or a `MethodTable`.
 
 ``` csharp
+partial interface IRuntimeTypeSystem : IContract
+{
     #region TypeHandle inspection APIs
     public virtual TypeHandle GetTypeHandle(TargetPointer targetPointer);
 
@@ -73,10 +76,35 @@ A `TypeHandle` is the runtime representation of the type information about a val
     public virtual bool IsFunctionPointer(TypeHandle typeHandle, out ReadOnlySpan<TypeHandle> retAndArgTypes, out byte callConv);
 
     #endregion TypeHandle inspection APIs
+}
+```
+
+### MethodDesc
+
+A `MethodDesc` is the runtime representation of a managed method (either from IL, from reflection emit, or generated by the runtime).
+
+```csharp
+struct MethodDescHandle
+{
+    // no public properties or constructors
+
+    internal TargetPointer Address { get; }
+}
+```
+
+```csharp
+partial interface IRuntimeTypeSystem : IContract
+{
+    public virtual MethodDescHandle GetMethodDescHandle(TargetPointer methodDescPointer);
+
+    public virtual TargetPointer GetMethodTable(MethodDescHandle methodDesc);
+}
 ```
 
 ## Version 1
 
+### TypeHandle
+
 The `MethodTable` inspection APIs are implemented in terms of the following flags on the runtime `MethodTable` structure:
 
 ``` csharp
@@ -233,7 +261,11 @@ static class RuntimeTypeSystem_1_Helpers
 }
 ```
 
-The contract depends on the global pointer value `FreeObjectMethodTablePointer`.
+The contract depends on the following globals
+
+| Global name | Meaning |
+| --- | --- |
+| `FreeObjectMethodTablePointer` | A pointer to the address of a `MethodTable` used by the GC to indicate reclaimed memory
 
 The contract additionally depends on these data descriptors
 
@@ -251,6 +283,7 @@ The contract additionally depends on these data descriptors
 | `EEClass` | `InternalCorElementType` | An InternalCorElementType uses the enum values of a CorElementType to indicate some of the information about the type of the type which uses the EEClass In particular, all reference types are CorElementType.Class, Enums are the element type of their underlying type and ValueTypes which can exactly be represented as an element type are represented as such, all other values types are represented as CorElementType.ValueType. |
 | `EEClass` | `MethodTable` | Pointer to the canonical MethodTable of this type |
 | `EEClass` | `NumMethods` | Count of methods attached to the EEClass |
+| `EEClass` | `NumNonVirtualSlots` | Count of non-virtual slots for the EEClass |
 | `EEClass` | `CorTypeAttr` | Various flags |
 | `ArrayClass` | `Rank` | Rank of the associated array MethodTable |
 | `TypeDesc` | `TypeAndFlags` | The lower 8 bits are the CorElementType of the `TypeDesc`, the upper 24 bits are reserved for flags |
@@ -523,3 +556,28 @@ The contract additionally depends on these data descriptors
         return true;
     }
 ```
+
+### MethodDesc
+
+The version 1 `MethodDesc` APIs depend on the `MethodDescAlignment` global and the `MethodDesc` and `MethodDescChunk` data descriptors.
+
+| Global name | Meaning |
+| --- | --- |
+| `MethodDescAlignment` | `MethodDescChunk` trailing data is allocated in multiples of this constant.  The size (in bytes) of each `MethodDesc` (or subclass) instance is a multiple of this constant.
+
+
+In the runtime a `MethodDesc` implicitly belongs to a single `MethodDescChunk` and some common data is shared between method descriptors that belong to the same chunk.  A single method table
+will typically have multiple chunks.  There are subkinds of MethodDescs at runtime of varying sizes (but the sizes must be mutliples of `MethodDescAlignment`) and each chunk contains method descriptors of the same size.
+
+We depend on the following data descriptors:
+| Data Descriptor Name | Field | Meaning |
+| --- | --- | --- |
+| `MethodDesc` | `ChunkIndex` | Offset of this `MethodDesc` relative to the end of its containing `MethodDescChunk` - in multiples of `MethodDescAlignment`
+| `MethodDesc` | `Slot` | The method's slot
+| `MethodDesc` | `Flags` | The method's flags
+| `MethodDescChunk` | `MethodTable` | The method table set of methods belongs to
+| `MethodDescChunk` | `Next` | The next chunk of methods
+| `MethodDescChunk` | `Size` | The size of this `MethodDescChunk`  following this `MethodDescChunk` header, minus 1. In multiples of `MethodDescAlignment`
+| `MethodDescChunk` | `Count` | The number of `MethodDesc` entries in this chunk, minus 1.
+
+**TODO(cdac)**

src/coreclr/debug/daccess/dacimpl.h

@@ -1243,6 +1243,7 @@ class ClrDataAccess
     HRESULT GetObjectExceptionDataImpl(CLRDATA_ADDRESS objAddr, struct DacpExceptionObjectData *data);
     HRESULT GetObjectStringDataImpl(CLRDATA_ADDRESS obj, unsigned int count, _Inout_updates_z_(count) WCHAR *stringData, unsigned int *pNeeded);
     HRESULT GetUsefulGlobalsImpl(struct DacpUsefulGlobalsData *globalsData);
+    HRESULT GetMethodDescDataImpl(CLRDATA_ADDRESS methodDesc, CLRDATA_ADDRESS ip, struct DacpMethodDescData *data, ULONG cRevertedRejitVersions, DacpReJitData * rgRevertedRejitData, ULONG * pcNeededRevertedRejitData);
 
     BOOL IsExceptionFromManagedCode(EXCEPTION_RECORD * pExceptionRecord);
 #ifndef TARGET_UNIX

src/coreclr/debug/daccess/request.cpp

@@ -1041,6 +1041,70 @@ HRESULT ClrDataAccess::GetMethodDescData(
     }
 
     SOSDacEnter();
+    if (m_cdacSos != NULL)
+    {
+        // Try the cDAC first - it will return E_NOTIMPL if it doesn't support this method yet. Fall back to the DAC.
+        hr = m_cdacSos->GetMethodDescData(methodDesc, ip, methodDescData, cRevertedRejitVersions, rgRevertedRejitData, pcNeededRevertedRejitData);
+        if (FAILED(hr))
+        {
+            hr = GetMethodDescDataImpl(methodDesc, ip, methodDescData, cRevertedRejitVersions, rgRevertedRejitData, pcNeededRevertedRejitData);
+        }
+#ifdef _DEBUG
+        else
+        {
+            // Assert that the data is the same as what we get from the DAC.
+            DacpMethodDescData mdDataLocal;
+            NewArrayHolder<DacpReJitData> rgRevertedRejitDataLocal{};
+            if (rgRevertedRejitData != nullptr)
+            {
+                rgRevertedRejitDataLocal = new DacpReJitData[cRevertedRejitVersions];
+            }
+            ULONG cNeededRevertedRejitDataLocal = 0;
+            ULONG *pcNeededRevertedRejitDataLocal = NULL;
+            if (pcNeededRevertedRejitData != NULL)
+            {
+                pcNeededRevertedRejitDataLocal = &cNeededRevertedRejitDataLocal;
+            }
+            HRESULT hrLocal = GetMethodDescDataImpl(methodDesc, ip,&mdDataLocal, cRevertedRejitVersions, rgRevertedRejitDataLocal, pcNeededRevertedRejitDataLocal);
+            _ASSERTE(hr == hrLocal);
+            _ASSERTE(methodDescData->bHasNativeCode == mdDataLocal.bHasNativeCode);
+            _ASSERTE(methodDescData->bIsDynamic == mdDataLocal.bIsDynamic);
+            _ASSERTE(methodDescData->wSlotNumber == mdDataLocal.wSlotNumber);
+            _ASSERTE(methodDescData->NativeCodeAddr == mdDataLocal.NativeCodeAddr);
+            _ASSERTE(methodDescData->AddressOfNativeCodeSlot == mdDataLocal.AddressOfNativeCodeSlot);
+            //TODO[cdac]: assert the rest of mdDataLocal contains the same info as methodDescData
+            if (rgRevertedRejitData != NULL)
+            {
+                _ASSERTE (cNeededRevertedRejitDataLocal == *pcNeededRevertedRejitData);
+                for (ULONG i = 0; i < cNeededRevertedRejitDataLocal; i++)
+                {
+                    _ASSERTE(rgRevertedRejitData[i].rejitID == rgRevertedRejitDataLocal[i].rejitID);
+                    _ASSERTE(rgRevertedRejitData[i].NativeCodeAddr == rgRevertedRejitDataLocal[i].NativeCodeAddr);
+                    _ASSERTE(rgRevertedRejitData[i].flags == rgRevertedRejitDataLocal[i].flags);
+                }
+            }
+        }
+#endif
+    }
+    else
+    {
+        hr = GetMethodDescDataImpl(methodDesc, ip, methodDescData, cRevertedRejitVersions, rgRevertedRejitData, pcNeededRevertedRejitData);
+    }
+
+    SOSDacLeave();
+    return hr;
+}
+
+HRESULT ClrDataAccess::GetMethodDescDataImpl(
+    CLRDATA_ADDRESS methodDesc,
+    CLRDATA_ADDRESS ip,
+    struct DacpMethodDescData *methodDescData,
+    ULONG cRevertedRejitVersions,
+    DacpReJitData * rgRevertedRejitData,
+    ULONG * pcNeededRevertedRejitData)
+{
+
+    HRESULT hr = S_OK;
 
     PTR_MethodDesc pMD = PTR_MethodDesc(TO_TADDR(methodDesc));
 
@@ -1236,7 +1300,6 @@ HRESULT ClrDataAccess::GetMethodDescData(
         }
     }
 
-    SOSDacLeave();
     return hr;
 }
 

src/coreclr/debug/runtimeinfo/datadescriptor.h

@@ -202,7 +202,7 @@ CDAC_TYPE_FIELD(Module, /*pointer*/, TypeDefToMethodTableMap, cdac_offsets<Modul
 CDAC_TYPE_FIELD(Module, /*pointer*/, TypeRefToMethodTableMap, cdac_offsets<Module>::TypeRefToMethodTableMap)
 CDAC_TYPE_END(Module)
 
-// Metadata
+// RuntimeTypeSystem
 
 CDAC_TYPE_BEGIN(MethodTable)
 CDAC_TYPE_INDETERMINATE(MethodTable)
@@ -223,6 +223,7 @@ CDAC_TYPE_FIELD(EEClass, /*pointer*/, MethodTable, cdac_offsets<EEClass>::Method
 CDAC_TYPE_FIELD(EEClass, /*uint16*/, NumMethods, cdac_offsets<EEClass>::NumMethods)
 CDAC_TYPE_FIELD(EEClass, /*uint32*/, CorTypeAttr, cdac_offsets<EEClass>::CorTypeAttr)
 CDAC_TYPE_FIELD(EEClass, /*uint8*/, InternalCorElementType, cdac_offsets<EEClass>::InternalCorElementType)
+CDAC_TYPE_FIELD(EEClass, /*uint16*/, NumNonVirtualSlots, cdac_offsets<EEClass>::NumNonVirtualSlots)
 CDAC_TYPE_END(EEClass)
 
 CDAC_TYPE_BEGIN(ArrayClass)
@@ -263,6 +264,21 @@ CDAC_TYPE_FIELD(DynamicMetadata, /*uint32*/, Size, cdac_offsets<DynamicMetadata>
 CDAC_TYPE_FIELD(DynamicMetadata, /*inline byte array*/, Data, cdac_offsets<DynamicMetadata>::Data)
 CDAC_TYPE_END(DynamicMetadata)
 
+CDAC_TYPE_BEGIN(MethodDesc)
+CDAC_TYPE_INDETERMINATE(MethodDesc)
+CDAC_TYPE_FIELD(MethodDesc, /*uint8*/, ChunkIndex, cdac_offsets<MethodDesc>::ChunkIndex)
+CDAC_TYPE_FIELD(MethodDesc, /*uint16*/, Slot, cdac_offsets<MethodDesc>::Slot)
+CDAC_TYPE_FIELD(MethodDesc, /*uint16*/, Flags, cdac_offsets<MethodDesc>::Flags)
+CDAC_TYPE_END(MethodDesc)
+
+CDAC_TYPE_BEGIN(MethodDescChunk)
+CDAC_TYPE_SIZE(sizeof(MethodDescChunk))
+CDAC_TYPE_FIELD(MethodDescChunk, /*pointer*/, MethodTable, cdac_offsets<MethodDescChunk>::MethodTable)
+CDAC_TYPE_FIELD(MethodDescChunk, /*pointer*/, Next, cdac_offsets<MethodDescChunk>::Next)
+CDAC_TYPE_FIELD(MethodDescChunk, /*uint8*/, Size, cdac_offsets<MethodDescChunk>::Size)
+CDAC_TYPE_FIELD(MethodDescChunk, /*uint8*/, Count, cdac_offsets<MethodDescChunk>::Count)
+CDAC_TYPE_END(MethodDescChunk)
+
 CDAC_TYPES_END()
 
 CDAC_GLOBALS_BEGIN()
@@ -282,6 +298,7 @@ CDAC_GLOBAL(ObjectToMethodTableUnmask, uint8, 1 | 1 << 1 | 1 << 2)
 CDAC_GLOBAL(ObjectToMethodTableUnmask, uint8, 1 | 1 << 1)
 #endif //TARGET_64BIT
 CDAC_GLOBAL(SOSBreakingChangeVersion, uint8, SOS_BREAKING_CHANGE_VERSION)
+CDAC_GLOBAL(MethodDescAlignment, uint64, MethodDesc::ALIGNMENT)
 CDAC_GLOBAL_POINTER(ExceptionMethodTable, &::g_pExceptionClass)
 CDAC_GLOBAL_POINTER(FreeObjectMethodTable, &::g_pFreeObjectMethodTable)
 CDAC_GLOBAL_POINTER(ObjectMethodTable, &::g_pObjectClass)

src/coreclr/vm/class.h

@@ -1807,6 +1807,7 @@ template<> struct cdac_offsets<EEClass>
     static constexpr size_t MethodTable = offsetof(EEClass, m_pMethodTable);
     static constexpr size_t NumMethods = offsetof(EEClass, m_NumMethods);
     static constexpr size_t CorTypeAttr = offsetof(EEClass, m_dwAttrClass);
+    static constexpr size_t NumNonVirtualSlots = offsetof(EEClass, m_NumNonVirtualSlots);
 };
 
 // --------------------------------------------------------------------------------------------

src/coreclr/vm/method.hpp

@@ -1908,6 +1908,15 @@ class MethodDesc
 public:
     static void Init();
 #endif
+
+    template<typename T> friend struct ::cdac_offsets;
+};
+
+template<> struct cdac_offsets<MethodDesc>
+{
+    static constexpr size_t ChunkIndex = offsetof(MethodDesc, m_chunkIndex);
+    static constexpr size_t Slot = offsetof(MethodDesc, m_wSlotNumber);
+    static constexpr size_t Flags = offsetof(MethodDesc, m_wFlags);
 };
 
 #ifndef DACCESS_COMPILE
@@ -2328,6 +2337,17 @@ class MethodDescChunk
     UINT16               m_flagsAndTokenRange;
 
     // Followed by array of method descs...
+
+    template<typename T> friend struct ::cdac_offsets;
+};
+
+template<>
+struct cdac_offsets<MethodDescChunk>
+{
+    static constexpr size_t MethodTable = offsetof(MethodDescChunk, m_methodTable);
+    static constexpr size_t Next = offsetof(MethodDescChunk, m_next);
+    static constexpr size_t Size = offsetof(MethodDescChunk, m_size);
+    static constexpr size_t Count = offsetof(MethodDescChunk, m_count);
 };
 
 inline int MethodDesc::GetMethodDescChunkIndex() const

src/coreclr/vm/methodtable.h

@@ -621,7 +621,7 @@ struct DynamicStaticsInfo
                 // If it has, then we don't need to do anything
                 return false;
             }
-            
+
             if (isClassInitedByUpdatingStaticPointer)
             {
                 oldValFromInterlockedOp = InterlockedCompareExchangeT(pAddr, newVal, oldVal);

src/native/managed/cdacreader/src/Constants.cs

@@ -25,5 +25,7 @@ internal static class Globals
 
         internal const string MiniMetaDataBuffAddress = nameof(MiniMetaDataBuffAddress);
         internal const string MiniMetaDataBuffMaxSize = nameof(MiniMetaDataBuffMaxSize);
+
+        internal const string MethodDescAlignment = nameof(MethodDescAlignment);
     }
 }

src/native/managed/cdacreader/src/Contracts/RuntimeTypeSystem.cs

@@ -55,16 +55,27 @@ internal enum CorElementType
     Sentinel = 0x41,
 }
 
+internal readonly struct MethodDescHandle
+{
+    internal MethodDescHandle(TargetPointer address)
+    {
+        Address = address;
+    }
+
+    internal TargetPointer Address { get; }
+}
+
 internal interface IRuntimeTypeSystem : IContract
 {
     static string IContract.Name => nameof(RuntimeTypeSystem);
     static IContract IContract.Create(Target target, int version)
     {
         TargetPointer targetPointer = target.ReadGlobalPointer(Constants.Globals.FreeObjectMethodTable);
         TargetPointer freeObjectMethodTable = target.ReadPointer(targetPointer);
+        ulong methodDescAlignment = target.ReadGlobal<ulong>(Constants.Globals.MethodDescAlignment);
         return version switch
         {
-            1 => new RuntimeTypeSystem_1(target, freeObjectMethodTable),
+            1 => new RuntimeTypeSystem_1(target, freeObjectMethodTable, methodDescAlignment),
             _ => default(RuntimeTypeSystem),
         };
     }
@@ -113,6 +124,11 @@ static IContract IContract.Create(Target target, int version)
     public virtual bool IsFunctionPointer(TypeHandle typeHandle, out ReadOnlySpan<TypeHandle> retAndArgTypes, out byte callConv) => throw new NotImplementedException();
     // Returns null if the TypeHandle is not a class/struct/generic variable
     #endregion TypeHandle inspection APIs
+
+    #region MethodDesc inspection APIs
+    public virtual MethodDescHandle GetMethodDescHandle(TargetPointer targetPointer) => throw new NotImplementedException();
+    public virtual TargetPointer GetMethodTable(MethodDescHandle methodDesc) => throw new NotImplementedException();
+    #endregion MethodDesc inspection APIs
 }
 
 internal struct RuntimeTypeSystem : IRuntimeTypeSystem