archive/rpcs3
1
0
Fork 0
mirror of https://github.com/RPCS3/rpcs3.git synced 2025-07-14 10:48:36 +12:00
Code Issues Projects Releases Packages Wiki Activity Actions 2

PPU LLVM arm64+macOS port (#12115)

Browse source 
* BufferUtils: use naive function pointer on Apple arm64

Use naive function pointer on Apple arm64 because ASLR breaks asmjit.
See BufferUtils.cpp comment for explanation on why this happens and how
to fix if you want to use asmjit.

* build-macos: fix source maps for Mac

Tell Qt not to strip debug symbols when we're in debug or relwithdebinfo
modes.

* LLVM PPU: fix aarch64 on macOS

Force MachO on macOS to fix LLVM being unable to patch relocations
during codegen. Adds Aarch64 NEON intrinsics for x86 intrinsics used by
PPUTranslator/Recompiler.

* virtual memory: use 16k pages on aarch64 macOS

Temporary hack to get things working by using 16k pages instead of 4k
pages in VM emulation.

* PPU/SPU: fix NEON intrinsics and compilation for arm64 macOS

Fixes some intrinsics usage and patches usages of asmjit to properly
emit absolute jmps so ASLR doesn't cause out of bounds rel jumps. Also
patches the SPU recompiler to properly work on arm64 by telling LLVM to
target arm64.

* virtual memory: fix W^X toggles on macOS aarch64

Fixes W^X on macOS aarch64 by setting all JIT mmap'd regions to default
to RW mode. For both SPU and PPU execution threads, when initialization
finishes we toggle to RX mode. This exploits Apple's per-thread setting
for RW/RX to let us be technically compliant with the OS's W^X
    enforcement while not needing to actually separate the memory
    allocated for code/data.

* PPU: implement aarch64 specific functions

Implements ppu_gateway for arm64 and patches LLVM initialization to use
the correct triple. Adds some fixes for macOS W^X JIT restrictions when
entering/exiting JITed code.

* PPU: Mark rpcs3 calls as non-tail

Strictly speaking, rpcs3 JIT -> C++ calls are not tail calls. If you
call a function inside e.g. an L2 syscall, it will clobber LR on arm64
and subtly break returns in emulated code. Only JIT -> JIT "calls"
should be tail.

* macOS/arm64: compatibility fixes

* vm: patch virtual memory for arm64 macOS

Tag mmap calls with MAP_JIT to allow W^X on macOS. Fix mmap calls to
existing mmap'd addresses that were tagged with MAP_JIT on macOS. Fix
memory unmapping on 16K page machines with a hack to mark "unmapped"
pages as RW.

* PPU: remove wrong comment

* PPU: fix a merge regression

* vm: remove 16k page hacks

* PPU: formatting fixes

* PPU: fix arm64 null function assembly

* ppu: clean up arch-specific instructions
This commit is contained in:
Jeff Guo 2022-06-14 05:28:38 -07:00 • committed by GitHub
parent 264253757c
commit cefc37a553
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
14 changed files with 306 additions and 16 deletions
Download patch file Download diff file Expand all files Collapse all files

33
rpcs3/Emu/Cell/SPURecompiler.cpp
View file

@ -34,6 +34,15 @@ const extern spu_decoder<spu_iflag> g_spu_iflag;
// Move 4 args for calling native function from a GHC calling convention function
static u8* move_args_ghc_to_native(u8* raw)
{
#ifdef ARCH_ARM64
// Note: this is a placeholder to get rpcs3 working for now
// mov x0, x22
// mov x1, x23
// mov x2, x24
// mov x3, x25
std::memcpy(raw, "\xE0\x03\x16\xAA\xE1\x03\x17\xAA\xE2\x03\x18\xAA\xE3\x03\x19\xAA", 16);
return raw + 16;
#else
#ifdef _WIN32
// mov rcx, r13
// mov rdx, rbp
@ -49,10 +58,14 @@ static u8* move_args_ghc_to_native(u8* raw)
#endif
return raw + 12;
#endif
}
DECLARE(spu_runtime::tr_dispatch) = []
{
#ifdef __APPLE__
pthread_jit_write_protect_np(false);
#endif
// Generate a special trampoline to spu_recompiler_base::dispatch with pause instruction
u8* const trptr = jit_runtime::alloc(32, 16);
u8* raw = move_args_ghc_to_native(trptr);
@ -439,6 +452,9 @@ void spu_cache::initialize()
named_thread_group workers("SPU Worker ", worker_count, [&]() -> uint
{
#ifdef __APPLE__
pthread_jit_write_protect_np(false);
#endif
// Set low priority
thread_ctrl::scoped_priority low_prio(-1);
@ -4412,7 +4428,7 @@ public:
// Create LLVM module
std::unique_ptr<Module> _module = std::make_unique<Module>(m_hash + ".obj", m_context);
_module->setTargetTriple(Triple::normalize("x86_64-unknown-linux-gnu"));
_module->setTargetTriple(utils::c_llvm_default_triple);
_module->setDataLayout(m_jit.get_engine().getTargetMachine()->createDataLayout());
m_module = _module.get();
@ -4672,7 +4688,12 @@ public:
// Function that executes check_state and escapes if necessary
m_test_state = llvm::cast<llvm::Function>(m_module->getOrInsertFunction("spu_test_state", get_ftype<void, u8*>()).getCallee());
m_test_state->setLinkage(GlobalValue::InternalLinkage);
#ifdef ARCH_ARM64
// LLVM doesn't support PreserveAll on arm64.
m_test_state->setCallingConv(CallingConv::GHC);
#else
m_test_state->setCallingConv(CallingConv::PreserveAll);
#endif
m_ir->SetInsertPoint(BasicBlock::Create(m_context, "", m_test_state));
const auto escape_yes = BasicBlock::Create(m_context, "", m_test_state);
const auto escape_no = BasicBlock::Create(m_context, "", m_test_state);
@ -5069,7 +5090,7 @@ public:
// Create LLVM module
std::unique_ptr<Module> _module = std::make_unique<Module>("spu_interpreter.obj", m_context);
_module->setTargetTriple(Triple::normalize("x86_64-unknown-linux-gnu"));
_module->setTargetTriple(utils::c_llvm_default_triple);
_module->setDataLayout(m_jit.get_engine().getTargetMachine()->createDataLayout());
m_module = _module.get();
@ -5114,7 +5135,11 @@ public:
// Save host thread's stack pointer
const auto native_sp = spu_ptr<u64>(&spu_thread::saved_native_sp);
#if defined(ARCH_X64)
const auto rsp_name = MetadataAsValue::get(m_context, MDNode::get(m_context, {MDString::get(m_context, "rsp")}));
#elif defined(ARCH_ARM64)
const auto rsp_name = MetadataAsValue::get(m_context, MDNode::get(m_context, {MDString::get(m_context, "sp")}));
#endif
m_ir->CreateStore(m_ir->CreateCall(get_intrinsic<u64>(Intrinsic::read_register), {rsp_name}), native_sp);
// Decode (shift) and load function pointer
@ -5328,7 +5353,11 @@ public:
else if (!(itype & spu_itype::branch))
{
// Hack: inline ret instruction before final jmp; this is not reliable.
#ifdef ARCH_X64
m_ir->CreateCall(InlineAsm::get(get_ftype<void>(), "ret", "", true, false, InlineAsm::AD_Intel));
#else
m_ir->CreateCall(InlineAsm::get(get_ftype<void>(), "ret", "", true, false));
#endif
fret = ret_func;
}