diff options
author | gdkchan <gab.dark.100@gmail.com> | 2019-08-08 15:56:22 -0300 |
---|---|---|
committer | emmauss <emmausssss@gmail.com> | 2019-08-08 21:56:22 +0300 |
commit | a731ab3a2aad56e6ceb8b4e2444a61353246295c (patch) | |
tree | c7f13f51bfec6b19431e62167811ae31e9d2fea9 /ARMeilleure/Translation/SsaConstruction.cs | |
parent | 1ba58e9942e54175e3f3a0e1d57a48537f4888b1 (diff) |
Add a new JIT compiler for CPU code (#693)
* Start of the ARMeilleure project
* Refactoring around the old IRAdapter, now renamed to PreAllocator
* Optimize the LowestBitSet method
* Add CLZ support and fix CLS implementation
* Add missing Equals and GetHashCode overrides on some structs, misc small tweaks
* Implement the ByteSwap IR instruction, and some refactoring on the assembler
* Implement the DivideUI IR instruction and fix 64-bits IDIV
* Correct constant operand type on CSINC
* Move division instructions implementation to InstEmitDiv
* Fix destination type for the ConditionalSelect IR instruction
* Implement UMULH and SMULH, with new IR instructions
* Fix some issues with shift instructions
* Fix constant types for BFM instructions
* Fix up new tests using the new V128 struct
* Update tests
* Move DIV tests to a separate file
* Add support for calls, and some instructions that depends on them
* Start adding support for SIMD & FP types, along with some of the related ARM instructions
* Fix some typos and the divide instruction with FP operands
* Fix wrong method call on Clz_V
* Implement ARM FP & SIMD move instructions, Saddlv_V, and misc. fixes
* Implement SIMD logical instructions and more misc. fixes
* Fix PSRAD x86 instruction encoding, TRN, UABD and UABDL implementations
* Implement float conversion instruction, merge in LDj3SNuD fixes, and some other misc. fixes
* Implement SIMD shift instruction and fix Dup_V
* Add SCVTF and UCVTF (vector, fixed-point) variants to the opcode table
* Fix check with tolerance on tester
* Implement FP & SIMD comparison instructions, and some fixes
* Update FCVT (Scalar) encoding on the table to support the Half-float variants
* Support passing V128 structs, some cleanup on the register allocator, merge LDj3SNuD fixes
* Use old memory access methods, made a start on SIMD memory insts support, some fixes
* Fix float constant passed to functions, save and restore non-volatile XMM registers, other fixes
* Fix arguments count with struct return values, other fixes
* More instructions
* Misc. fixes and integrate LDj3SNuD fixes
* Update tests
* Add a faster linear scan allocator, unwinding support on windows, and other changes
* Update Ryujinx.HLE
* Update Ryujinx.Graphics
* Fix V128 return pointer passing, RCX is clobbered
* Update Ryujinx.Tests
* Update ITimeZoneService
* Stop using GetFunctionPointer as that can't be called from native code, misc. fixes and tweaks
* Use generic GetFunctionPointerForDelegate method and other tweaks
* Some refactoring on the code generator, assert on invalid operations and use a separate enum for intrinsics
* Remove some unused code on the assembler
* Fix REX.W prefix regression on float conversion instructions, add some sort of profiler
* Add hardware capability detection
* Fix regression on Sha1h and revert Fcm** changes
* Add SSE2-only paths on vector extract and insert, some refactoring on the pre-allocator
* Fix silly mistake introduced on last commit on CpuId
* Generate inline stack probes when the stack allocation is too large
* Initial support for the System-V ABI
* Support multiple destination operands
* Fix SSE2 VectorInsert8 path, and other fixes
* Change placement of XMM callee save and restore code to match other compilers
* Rename Dest to Destination and Inst to Instruction
* Fix a regression related to calls and the V128 type
* Add an extra space on comments to match code style
* Some refactoring
* Fix vector insert FP32 SSE2 path
* Port over the ARM32 instructions
* Avoid memory protection races on JIT Cache
* Another fix on VectorInsert FP32 (thanks to LDj3SNuD
* Float operands don't need to use the same register when VEX is supported
* Add a new register allocator, higher quality code for hot code (tier up), and other tweaks
* Some nits, small improvements on the pre allocator
* CpuThreadState is gone
* Allow changing CPU emulators with a config entry
* Add runtime identifiers on the ARMeilleure project
* Allow switching between CPUs through a config entry (pt. 2)
* Change win10-x64 to win-x64 on projects
* Update the Ryujinx project to use ARMeilleure
* Ensure that the selected register is valid on the hybrid allocator
* Allow exiting on returns to 0 (should fix test regression)
* Remove register assignments for most used variables on the hybrid allocator
* Do not use fixed registers as spill temp
* Add missing namespace and remove unneeded using
* Address PR feedback
* Fix types, etc
* Enable AssumeStrictAbiCompliance by default
* Ensure that Spill and Fill don't load or store any more than necessary
Diffstat (limited to 'ARMeilleure/Translation/SsaConstruction.cs')
-rw-r--r-- | ARMeilleure/Translation/SsaConstruction.cs | 293 |
1 files changed, 293 insertions, 0 deletions
diff --git a/ARMeilleure/Translation/SsaConstruction.cs b/ARMeilleure/Translation/SsaConstruction.cs new file mode 100644 index 00000000..ccf52591 --- /dev/null +++ b/ARMeilleure/Translation/SsaConstruction.cs @@ -0,0 +1,293 @@ +using ARMeilleure.Common; +using ARMeilleure.IntermediateRepresentation; +using ARMeilleure.State; +using System.Collections.Generic; + +using static ARMeilleure.IntermediateRepresentation.OperandHelper; + +namespace ARMeilleure.Translation +{ + static partial class Ssa + { + private class DefMap + { + private Dictionary<Register, Operand> _map; + + private BitMap _phiMasks; + + public DefMap() + { + _map = new Dictionary<Register, Operand>(); + + _phiMasks = new BitMap(RegisterConsts.TotalCount); + } + + public bool TryAddOperand(Register reg, Operand operand) + { + return _map.TryAdd(reg, operand); + } + + public bool TryGetOperand(Register reg, out Operand operand) + { + return _map.TryGetValue(reg, out operand); + } + + public bool AddPhi(Register reg) + { + return _phiMasks.Set(GetIdFromRegister(reg)); + } + + public bool HasPhi(Register reg) + { + return _phiMasks.IsSet(GetIdFromRegister(reg)); + } + } + + public static void Construct(ControlFlowGraph cfg) + { + DefMap[] globalDefs = new DefMap[cfg.Blocks.Count]; + + foreach (BasicBlock block in cfg.Blocks) + { + globalDefs[block.Index] = new DefMap(); + } + + Queue<BasicBlock> dfPhiBlocks = new Queue<BasicBlock>(); + + // First pass, get all defs and locals uses. + foreach (BasicBlock block in cfg.Blocks) + { + Operand[] localDefs = new Operand[RegisterConsts.TotalCount]; + + LinkedListNode<Node> node = block.Operations.First; + + Operand RenameLocal(Operand operand) + { + if (operand != null && operand.Kind == OperandKind.Register) + { + Operand local = localDefs[GetIdFromRegister(operand.GetRegister())]; + + operand = local ?? operand; + } + + return operand; + } + + while (node != null) + { + if (node.Value is Operation operation) + { + for (int index = 0; index < operation.SourcesCount; index++) + { + operation.SetSource(index, RenameLocal(operation.GetSource(index))); + } + + Operand dest = operation.Destination; + + if (dest != null && dest.Kind == OperandKind.Register) + { + Operand local = Local(dest.Type); + + localDefs[GetIdFromRegister(dest.GetRegister())] = local; + + operation.Destination = local; + } + } + + node = node.Next; + } + + for (int index = 0; index < RegisterConsts.TotalCount; index++) + { + Operand local = localDefs[index]; + + if (local == null) + { + continue; + } + + Register reg = GetRegisterFromId(index); + + globalDefs[block.Index].TryAddOperand(reg, local); + + dfPhiBlocks.Enqueue(block); + + while (dfPhiBlocks.TryDequeue(out BasicBlock dfPhiBlock)) + { + foreach (BasicBlock domFrontier in dfPhiBlock.DominanceFrontiers) + { + if (globalDefs[domFrontier.Index].AddPhi(reg)) + { + dfPhiBlocks.Enqueue(domFrontier); + } + } + } + } + } + + // Second pass, rename variables with definitions on different blocks. + foreach (BasicBlock block in cfg.Blocks) + { + Operand[] localDefs = new Operand[RegisterConsts.TotalCount]; + + LinkedListNode<Node> node = block.Operations.First; + + Operand RenameGlobal(Operand operand) + { + if (operand != null && operand.Kind == OperandKind.Register) + { + int key = GetIdFromRegister(operand.GetRegister()); + + Operand local = localDefs[key]; + + if (local == null) + { + local = FindDef(globalDefs, block, operand); + + localDefs[key] = local; + } + + operand = local; + } + + return operand; + } + + while (node != null) + { + if (node.Value is Operation operation) + { + for (int index = 0; index < operation.SourcesCount; index++) + { + operation.SetSource(index, RenameGlobal(operation.GetSource(index))); + } + } + + node = node.Next; + } + } + } + + private static Operand FindDef(DefMap[] globalDefs, BasicBlock current, Operand operand) + { + if (globalDefs[current.Index].HasPhi(operand.GetRegister())) + { + return InsertPhi(globalDefs, current, operand); + } + + if (current != current.ImmediateDominator) + { + return FindDefOnPred(globalDefs, current.ImmediateDominator, operand); + } + + return Undef(); + } + + private static Operand FindDefOnPred(DefMap[] globalDefs, BasicBlock current, Operand operand) + { + BasicBlock previous; + + do + { + DefMap defMap = globalDefs[current.Index]; + + Register reg = operand.GetRegister(); + + if (defMap.TryGetOperand(reg, out Operand lastDef)) + { + return lastDef; + } + + if (defMap.HasPhi(reg)) + { + return InsertPhi(globalDefs, current, operand); + } + + previous = current; + current = current.ImmediateDominator; + } + while (previous != current); + + return Undef(); + } + + private static Operand InsertPhi(DefMap[] globalDefs, BasicBlock block, Operand operand) + { + // This block has a Phi that has not been materialized yet, but that + // would define a new version of the variable we're looking for. We need + // to materialize the Phi, add all the block/operand pairs into the Phi, and + // then use the definition from that Phi. + Operand local = Local(operand.Type); + + PhiNode phi = new PhiNode(local, block.Predecessors.Count); + + AddPhi(block, phi); + + globalDefs[block.Index].TryAddOperand(operand.GetRegister(), local); + + for (int index = 0; index < block.Predecessors.Count; index++) + { + BasicBlock predecessor = block.Predecessors[index]; + + phi.SetBlock(index, predecessor); + phi.SetSource(index, FindDefOnPred(globalDefs, predecessor, operand)); + } + + return local; + } + + private static void AddPhi(BasicBlock block, PhiNode phi) + { + LinkedListNode<Node> node = block.Operations.First; + + if (node != null) + { + while (node.Next?.Value is PhiNode) + { + node = node.Next; + } + } + + if (node?.Value is PhiNode) + { + block.Operations.AddAfter(node, phi); + } + else + { + block.Operations.AddFirst(phi); + } + } + + private static int GetIdFromRegister(Register reg) + { + if (reg.Type == RegisterType.Integer) + { + return reg.Index; + } + else if (reg.Type == RegisterType.Vector) + { + return RegisterConsts.IntRegsCount + reg.Index; + } + else /* if (reg.Type == RegisterType.Flag) */ + { + return RegisterConsts.IntAndVecRegsCount + reg.Index; + } + } + + private static Register GetRegisterFromId(int id) + { + if (id < RegisterConsts.IntRegsCount) + { + return new Register(id, RegisterType.Integer); + } + else if (id < RegisterConsts.IntAndVecRegsCount) + { + return new Register(id - RegisterConsts.IntRegsCount, RegisterType.Vector); + } + else /* if (id < RegisterConsts.TotalCount) */ + { + return new Register(id - RegisterConsts.IntAndVecRegsCount, RegisterType.Flag); + } + } + } +}
\ No newline at end of file |