Ryujinx-uplift/ARMeilleure/Translation/Cache/JitCache.cs
FICTURE7 9d7627af64
Add multi-level function table (#2228)
* Add AddressTable<T>

* Use AddressTable<T> for dispatch

* Remove JumpTable & co.

* Add fallback for out of range addresses

* Add PPTC support

* Add documentation to `AddressTable<T>`

* Make AddressTable<T> configurable

* Fix table walk

* Fix IsMapped check

* Remove CountTableCapacity

* Add PPTC support for fast path

* Rename IsMapped to IsValid

* Remove stale comment

* Change format of address in exception message

* Add TranslatorStubs

* Split DispatchStub

Avoids recompilation of stubs during tests.

* Add hint for 64bit or 32bit

* Add documentation to `Symbol`

* Add documentation to `TranslatorStubs`

Make `TranslatorStubs` disposable as well.

* Add documentation to `SymbolType`

* Add `AddressTableEventSource` to monitor function table size

Add an EventSource which measures the amount of unmanaged bytes
allocated by AddressTable<T> instances.

 dotnet-counters monitor -n Ryujinx --counters ARMeilleure

* Add `AllowLcqInFunctionTable` optimization toggle

This is to reduce the impact this change has on the test duration.
Before everytime a test was ran, the FunctionTable would be initialized
and populated so that the newly compiled test would get registered to
it.

* Implement unmanaged dispatcher

Uses the DispatchStub to dispatch into the next translation, which
allows execution to stay in unmanaged for longer and skips a
ConcurrentDictionary look up when the target translation has been
registered to the FunctionTable.

* Remove redundant null check

* Tune levels of FunctionTable

Uses 5 levels instead of 4 and change unit of AddressTableEventSource
from KB to MB.

* Use 64-bit function table

Improves codegen for direct branches:

    mov qword [rax+0x408],0x10603560
 -  mov rcx,sub_10603560_OFFSET
 -  mov ecx,[rcx]
 -  mov ecx,ecx
 -  mov rdx,JIT_CACHE_BASE
 -  add rdx,rcx
 +  mov rcx,sub_10603560
 +  mov rdx,[rcx]
    mov rcx,rax

Improves codegen for dispatch stub:

    and rax,byte +0x1f
 -  mov eax,[rcx+rax*4]
 -  mov eax,eax
 -  mov rcx,JIT_CACHE_BASE
 -  lea rax,[rcx+rax]
 +  mov rax,[rcx+rax*8]
    mov rcx,rbx

* Remove `JitCacheSymbol` & `JitCache.Offset`

* Turn `Translator.Translate` into an instance method

We do not have to add more parameter to this method and related ones as
new structures are added & needed for translation.

* Add symbol only when PTC is enabled

Address LDj3SNuD's feedback

* Change `NativeContext.Running` to a 32-bit integer

* Fix PageTable symbol for host mapped
2021-05-29 18:06:28 -03:00

184 lines
5.0 KiB
C#

using ARMeilleure.CodeGen;
using ARMeilleure.CodeGen.Unwinding;
using ARMeilleure.Memory;
using System;
using System.Collections.Generic;
using System.Diagnostics;
using System.Runtime.InteropServices;
namespace ARMeilleure.Translation.Cache
{
static class JitCache
{
private const int PageSize = 4 * 1024;
private const int PageMask = PageSize - 1;
private const int CodeAlignment = 4; // Bytes.
private const int CacheSize = 2047 * 1024 * 1024;
private static ReservedRegion _jitRegion;
private static CacheMemoryAllocator _cacheAllocator;
private static readonly List<CacheEntry> _cacheEntries = new List<CacheEntry>();
private static readonly object _lock = new object();
private static bool _initialized;
public static IntPtr Base => _jitRegion.Pointer;
public static void Initialize(IJitMemoryAllocator allocator)
{
if (_initialized) return;
lock (_lock)
{
if (_initialized) return;
_jitRegion = new ReservedRegion(allocator, CacheSize);
_cacheAllocator = new CacheMemoryAllocator(CacheSize);
if (RuntimeInformation.IsOSPlatform(OSPlatform.Windows))
{
JitUnwindWindows.InstallFunctionTableHandler(_jitRegion.Pointer, CacheSize, _jitRegion.Pointer + Allocate(PageSize));
}
_initialized = true;
}
}
public static IntPtr Map(CompiledFunction func)
{
byte[] code = func.Code;
lock (_lock)
{
Debug.Assert(_initialized);
int funcOffset = Allocate(code.Length);
IntPtr funcPtr = _jitRegion.Pointer + funcOffset;
ReprotectAsWritable(funcOffset, code.Length);
Marshal.Copy(code, 0, funcPtr, code.Length);
ReprotectAsExecutable(funcOffset, code.Length);
Add(funcOffset, code.Length, func.UnwindInfo);
return funcPtr;
}
}
public static void Unmap(IntPtr pointer)
{
lock (_lock)
{
Debug.Assert(_initialized);
int funcOffset = (int)(pointer.ToInt64() - _jitRegion.Pointer.ToInt64());
bool result = TryFind(funcOffset, out CacheEntry entry);
Debug.Assert(result);
_cacheAllocator.Free(funcOffset, AlignCodeSize(entry.Size));
Remove(funcOffset);
}
}
private static void ReprotectAsWritable(int offset, int size)
{
int endOffs = offset + size;
int regionStart = offset & ~PageMask;
int regionEnd = (endOffs + PageMask) & ~PageMask;
_jitRegion.Block.MapAsRwx((ulong)regionStart, (ulong)(regionEnd - regionStart));
}
private static void ReprotectAsExecutable(int offset, int size)
{
int endOffs = offset + size;
int regionStart = offset & ~PageMask;
int regionEnd = (endOffs + PageMask) & ~PageMask;
_jitRegion.Block.MapAsRx((ulong)regionStart, (ulong)(regionEnd - regionStart));
}
private static int Allocate(int codeSize)
{
codeSize = AlignCodeSize(codeSize);
int allocOffset = _cacheAllocator.Allocate(codeSize);
if (allocOffset < 0)
{
throw new OutOfMemoryException("JIT Cache exhausted.");
}
_jitRegion.ExpandIfNeeded((ulong)allocOffset + (ulong)codeSize);
return allocOffset;
}
private static int AlignCodeSize(int codeSize)
{
return checked(codeSize + (CodeAlignment - 1)) & ~(CodeAlignment - 1);
}
private static void Add(int offset, int size, UnwindInfo unwindInfo)
{
CacheEntry entry = new CacheEntry(offset, size, unwindInfo);
int index = _cacheEntries.BinarySearch(entry);
if (index < 0)
{
index = ~index;
}
_cacheEntries.Insert(index, entry);
}
private static void Remove(int offset)
{
int index = _cacheEntries.BinarySearch(new CacheEntry(offset, 0, default));
if (index < 0)
{
index = ~index - 1;
}
if (index >= 0)
{
_cacheEntries.RemoveAt(index);
}
}
public static bool TryFind(int offset, out CacheEntry entry)
{
lock (_lock)
{
int index = _cacheEntries.BinarySearch(new CacheEntry(offset, 0, default));
if (index < 0)
{
index = ~index - 1;
}
if (index >= 0)
{
entry = _cacheEntries[index];
return true;
}
}
entry = default;
return false;
}
}
}