v8/test/cctest/test-sync-primitives-arm64.cc
Jakob Gruber c7cb9beca1 Reland "Reland "[deoptimizer] Change deopt entries into builtins""
This is a reland of fbfa9bf4ec

The arm64 was missing proper codegen for CFI, thus sizes were off.

Original change's description:
> Reland "[deoptimizer] Change deopt entries into builtins"
>
> This is a reland of 7f58ced72e
>
> It fixes the different exit size emitted on x64/Atom CPUs due to
> performance tuning in TurboAssembler::Call. Additionally, add
> cctests to verify the fixed size exits.
>
> Original change's description:
> > [deoptimizer] Change deopt entries into builtins
> >
> > While the overall goal of this commit is to change deoptimization
> > entries into builtins, there are multiple related things happening:
> >
> > - Deoptimization entries, formerly stubs (i.e. Code objects generated
> >   at runtime, guaranteed to be immovable), have been converted into
> >   builtins. The major restriction is that we now need to preserve the
> >   kRootRegister, which was formerly used on most architectures to pass
> >   the deoptimization id. The solution differs based on platform.
> > - Renamed DEOPT_ENTRIES_OR_FOR_TESTING code kind to FOR_TESTING.
> > - Removed heap/ support for immovable Code generation.
> > - Removed the DeserializerData class (no longer needed).
> > - arm64: to preserve 4-byte deopt exits, introduced a new optimization
> >   in which the final jump to the deoptimization entry is generated
> >   once per Code object, and deopt exits can continue to emit a
> >   near-call.
> > - arm,ia32,x64: change to fixed-size deopt exits. This reduces exit
> >   sizes by 4/8, 5, and 5 bytes, respectively.
> >
> > On arm the deopt exit size is reduced from 12 (or 16) bytes to 8 bytes
> > by using the same strategy as on arm64 (recalc deopt id from return
> > address). Before:
> >
> >  e300a002       movw r10, <id>
> >  e59fc024       ldr ip, [pc, <entry offset>]
> >  e12fff3c       blx ip
> >
> > After:
> >
> >  e59acb35       ldr ip, [r10, <entry offset>]
> >  e12fff3c       blx ip
> >
> > On arm64 the deopt exit size remains 4 bytes (or 8 bytes in same cases
> > with CFI). Additionally, up to 4 builtin jumps are emitted per Code
> > object (max 32 bytes added overhead per Code object). Before:
> >
> >  9401cdae       bl <entry offset>
> >
> > After:
> >
> >  # eager deoptimization entry jump.
> >  f95b1f50       ldr x16, [x26, <eager entry offset>]
> >  d61f0200       br x16
> >  # lazy deoptimization entry jump.
> >  f95b2b50       ldr x16, [x26, <lazy entry offset>]
> >  d61f0200       br x16
> >  # the deopt exit.
> >  97fffffc       bl <eager deoptimization entry jump offset>
> >
> > On ia32 the deopt exit size is reduced from 10 to 5 bytes. Before:
> >
> >  bb00000000     mov ebx,<id>
> >  e825f5372b     call <entry>
> >
> > After:
> >
> >  e8ea2256ba     call <entry>
> >
> > On x64 the deopt exit size is reduced from 12 to 7 bytes. Before:
> >
> >  49c7c511000000 REX.W movq r13,<id>
> >  e8ea2f0700     call <entry>
> >
> > After:
> >
> >  41ff9560360000 call [r13+<entry offset>]
> >
> > Bug: v8:8661,v8:8768
> > Change-Id: I13e30aedc360474dc818fecc528ce87c3bfeed42
> > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2465834
> > Commit-Queue: Jakob Gruber <jgruber@chromium.org>
> > Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
> > Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#70597}
>
> Tbr: ulan@chromium.org, tebbi@chromium.org, rmcilroy@chromium.org
> Bug: v8:8661,v8:8768,chromium:1140165
> Change-Id: Ibcd5c39c58a70bf2b2ac221aa375fc68d495e144
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2485506
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
> Commit-Queue: Jakob Gruber <jgruber@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#70655}

Tbr: ulan@chromium.org, tebbi@chromium.org, rmcilroy@chromium.org
Bug: v8:8661
Bug: v8:8768
Bug: chromium:1140165
Change-Id: I471cc94fc085e527dc9bfb5a84b96bd907c2333f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2488682
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70672}
2020-10-21 06:01:38 +00:00

408 lines
13 KiB
C++

// Copyright 2017 the V8 project authors. All rights reserved.
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are
// met:
//
// * Redistributions of source code must retain the above copyright
// notice, this list of conditions and the following disclaimer.
// * Redistributions in binary form must reproduce the above
// copyright notice, this list of conditions and the following
// disclaimer in the documentation and/or other materials provided
// with the distribution.
// * Neither the name of Google Inc. nor the names of its
// contributors may be used to endorse or promote products derived
// from this software without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "src/init/v8.h"
#include "test/cctest/cctest.h"
#include "src/codegen/macro-assembler-inl.h"
#include "src/execution/arm64/simulator-arm64.h"
#include "src/heap/factory.h"
#include "src/objects/objects-inl.h"
namespace v8 {
namespace internal {
// These tests rely on the behaviour specific to the simulator so we cannot
// expect the same results on real hardware. The reason for this is that our
// simulation of synchronisation primitives is more conservative than the
// reality.
// For example:
// ldxr x1, [x2] ; Load acquire at address x2; x2 is now marked as exclusive.
// ldr x0, [x4] ; This is a normal load, and at a different address.
// ; However, any memory accesses can potentially clear the
// ; exclusivity (See ARM DDI 0487B.a B2.9.5). This is unlikely
// ; on real hardware but to be conservative, the simulator
// ; always does it.
// stxr w3, x1, [x2] ; As a result, this will always fail in the simulator but
// ; will likely succeed on hardware.
#if defined(USE_SIMULATOR)
#ifndef V8_TARGET_LITTLE_ENDIAN
#error Expected ARM to be little-endian
#endif
#define __ masm.
struct MemoryAccess {
enum class Kind {
None,
Load,
LoadExcl,
Store,
StoreExcl,
};
enum class Size {
Byte,
HalfWord,
Word,
};
MemoryAccess() : kind(Kind::None) {}
MemoryAccess(Kind kind, Size size, size_t offset, int value = 0)
: kind(kind), size(size), offset(offset), value(value) {}
Kind kind = Kind::None;
Size size = Size::Byte;
size_t offset = 0;
int value = 0;
};
struct TestData {
explicit TestData(int w) : w(w) {}
union {
int32_t w;
int16_t h;
int8_t b;
};
int dummy;
};
namespace {
void AssembleMemoryAccess(MacroAssembler* assembler, MemoryAccess access,
Register dest_reg, Register value_reg,
Register addr_reg) {
MacroAssembler& masm = *assembler;
__ Add(addr_reg, x0, Operand(access.offset));
switch (access.kind) {
case MemoryAccess::Kind::None:
break;
case MemoryAccess::Kind::Load:
switch (access.size) {
case MemoryAccess::Size::Byte:
__ ldrb(value_reg, MemOperand(addr_reg));
break;
case MemoryAccess::Size::HalfWord:
__ ldrh(value_reg, MemOperand(addr_reg));
break;
case MemoryAccess::Size::Word:
__ ldr(value_reg, MemOperand(addr_reg));
break;
}
break;
case MemoryAccess::Kind::LoadExcl:
switch (access.size) {
case MemoryAccess::Size::Byte:
__ ldaxrb(value_reg, addr_reg);
break;
case MemoryAccess::Size::HalfWord:
__ ldaxrh(value_reg, addr_reg);
break;
case MemoryAccess::Size::Word:
__ ldaxr(value_reg, addr_reg);
break;
}
break;
case MemoryAccess::Kind::Store:
switch (access.size) {
case MemoryAccess::Size::Byte:
__ Mov(value_reg, Operand(access.value));
__ strb(value_reg, MemOperand(addr_reg));
break;
case MemoryAccess::Size::HalfWord:
__ Mov(value_reg, Operand(access.value));
__ strh(value_reg, MemOperand(addr_reg));
break;
case MemoryAccess::Size::Word:
__ Mov(value_reg, Operand(access.value));
__ str(value_reg, MemOperand(addr_reg));
break;
}
break;
case MemoryAccess::Kind::StoreExcl:
switch (access.size) {
case MemoryAccess::Size::Byte:
__ Mov(value_reg, Operand(access.value));
__ stlxrb(dest_reg, value_reg, addr_reg);
break;
case MemoryAccess::Size::HalfWord:
__ Mov(value_reg, Operand(access.value));
__ stlxrh(dest_reg, value_reg, addr_reg);
break;
case MemoryAccess::Size::Word:
__ Mov(value_reg, Operand(access.value));
__ stlxr(dest_reg, value_reg, addr_reg);
break;
}
break;
}
}
void AssembleLoadExcl(MacroAssembler* assembler, MemoryAccess access,
Register value_reg, Register addr_reg) {
DCHECK(access.kind == MemoryAccess::Kind::LoadExcl);
AssembleMemoryAccess(assembler, access, no_reg, value_reg, addr_reg);
}
void AssembleStoreExcl(MacroAssembler* assembler, MemoryAccess access,
Register dest_reg, Register value_reg,
Register addr_reg) {
DCHECK(access.kind == MemoryAccess::Kind::StoreExcl);
AssembleMemoryAccess(assembler, access, dest_reg, value_reg, addr_reg);
}
void TestInvalidateExclusiveAccess(TestData initial_data, MemoryAccess access1,
MemoryAccess access2, MemoryAccess access3,
int expected_res, TestData expected_data) {
Isolate* isolate = CcTest::i_isolate();
HandleScope scope(isolate);
MacroAssembler masm(isolate, v8::internal::CodeObjectRequired::kYes);
AssembleLoadExcl(&masm, access1, w1, x1);
AssembleMemoryAccess(&masm, access2, w3, w2, x1);
AssembleStoreExcl(&masm, access3, w0, w3, x1);
__ Ret();
CodeDesc desc;
masm.GetCode(isolate, &desc);
Handle<Code> code =
Factory::CodeBuilder(isolate, desc, CodeKind::FOR_TESTING).Build();
TestData t = initial_data;
Simulator::current(isolate)->Call<void>(code->entry(), &t);
int res = Simulator::current(isolate)->wreg(0);
CHECK_EQ(expected_res, res);
switch (access3.size) {
case MemoryAccess::Size::Byte:
CHECK_EQ(expected_data.b, t.b);
break;
case MemoryAccess::Size::HalfWord:
CHECK_EQ(expected_data.h, t.h);
break;
case MemoryAccess::Size::Word:
CHECK_EQ(expected_data.w, t.w);
break;
}
}
} // namespace
TEST(simulator_invalidate_exclusive_access) {
using Kind = MemoryAccess::Kind;
using Size = MemoryAccess::Size;
MemoryAccess ldaxr_w(Kind::LoadExcl, Size::Word, offsetof(TestData, w));
MemoryAccess stlxr_w(Kind::StoreExcl, Size::Word, offsetof(TestData, w), 7);
// Address mismatch.
TestInvalidateExclusiveAccess(
TestData(1), ldaxr_w,
MemoryAccess(Kind::LoadExcl, Size::Word, offsetof(TestData, dummy)),
stlxr_w, 1, TestData(1));
// Size mismatch.
TestInvalidateExclusiveAccess(
TestData(1), ldaxr_w, MemoryAccess(),
MemoryAccess(Kind::StoreExcl, Size::HalfWord, offsetof(TestData, w), 7),
1, TestData(1));
// Load between ldaxr/stlxr.
TestInvalidateExclusiveAccess(
TestData(1), ldaxr_w,
MemoryAccess(Kind::Load, Size::Word, offsetof(TestData, dummy)), stlxr_w,
1, TestData(1));
// Store between ldaxr/stlxr.
TestInvalidateExclusiveAccess(
TestData(1), ldaxr_w,
MemoryAccess(Kind::Store, Size::Word, offsetof(TestData, dummy)), stlxr_w,
1, TestData(1));
// Match
TestInvalidateExclusiveAccess(TestData(1), ldaxr_w, MemoryAccess(), stlxr_w,
0, TestData(7));
}
namespace {
int ExecuteMemoryAccess(Isolate* isolate, TestData* test_data,
MemoryAccess access) {
HandleScope scope(isolate);
MacroAssembler masm(isolate, v8::internal::CodeObjectRequired::kYes);
AssembleMemoryAccess(&masm, access, w0, w2, x1);
__ Ret();
CodeDesc desc;
masm.GetCode(isolate, &desc);
Handle<Code> code =
Factory::CodeBuilder(isolate, desc, CodeKind::FOR_TESTING).Build();
Simulator::current(isolate)->Call<void>(code->entry(), test_data);
return Simulator::current(isolate)->wreg(0);
}
} // namespace
class MemoryAccessThread : public v8::base::Thread {
public:
MemoryAccessThread()
: Thread(Options("MemoryAccessThread")),
test_data_(nullptr),
is_finished_(false),
has_request_(false),
did_request_(false),
isolate_(nullptr) {}
virtual void Run() {
v8::Isolate::CreateParams create_params;
create_params.array_buffer_allocator = CcTest::array_buffer_allocator();
isolate_ = v8::Isolate::New(create_params);
Isolate* i_isolate = reinterpret_cast<Isolate*>(isolate_);
{
v8::Isolate::Scope scope(isolate_);
v8::base::MutexGuard lock_guard(&mutex_);
while (!is_finished_) {
while (!(has_request_ || is_finished_)) {
has_request_cv_.Wait(&mutex_);
}
if (is_finished_) {
break;
}
ExecuteMemoryAccess(i_isolate, test_data_, access_);
has_request_ = false;
did_request_ = true;
did_request_cv_.NotifyOne();
}
}
isolate_->Dispose();
}
void NextAndWait(TestData* test_data, MemoryAccess access) {
DCHECK(!has_request_);
v8::base::MutexGuard lock_guard(&mutex_);
test_data_ = test_data;
access_ = access;
has_request_ = true;
has_request_cv_.NotifyOne();
while (!did_request_) {
did_request_cv_.Wait(&mutex_);
}
did_request_ = false;
}
void Finish() {
v8::base::MutexGuard lock_guard(&mutex_);
is_finished_ = true;
has_request_cv_.NotifyOne();
}
private:
TestData* test_data_;
MemoryAccess access_;
bool is_finished_;
bool has_request_;
bool did_request_;
v8::base::Mutex mutex_;
v8::base::ConditionVariable has_request_cv_;
v8::base::ConditionVariable did_request_cv_;
v8::Isolate* isolate_;
};
TEST(simulator_invalidate_exclusive_access_threaded) {
using Kind = MemoryAccess::Kind;
using Size = MemoryAccess::Size;
Isolate* isolate = CcTest::i_isolate();
HandleScope scope(isolate);
TestData test_data(1);
MemoryAccessThread thread;
CHECK(thread.Start());
MemoryAccess ldaxr_w(Kind::LoadExcl, Size::Word, offsetof(TestData, w));
MemoryAccess stlxr_w(Kind::StoreExcl, Size::Word, offsetof(TestData, w), 7);
// Exclusive store completed by another thread first.
test_data = TestData(1);
thread.NextAndWait(&test_data, MemoryAccess(Kind::LoadExcl, Size::Word,
offsetof(TestData, w)));
ExecuteMemoryAccess(isolate, &test_data, ldaxr_w);
thread.NextAndWait(&test_data, MemoryAccess(Kind::StoreExcl, Size::Word,
offsetof(TestData, w), 5));
CHECK_EQ(1, ExecuteMemoryAccess(isolate, &test_data, stlxr_w));
CHECK_EQ(5, test_data.w);
// Exclusive store completed by another thread; different address, but masked
// to same
test_data = TestData(1);
ExecuteMemoryAccess(isolate, &test_data, ldaxr_w);
thread.NextAndWait(&test_data, MemoryAccess(Kind::LoadExcl, Size::Word,
offsetof(TestData, dummy)));
thread.NextAndWait(&test_data, MemoryAccess(Kind::StoreExcl, Size::Word,
offsetof(TestData, dummy), 5));
CHECK_EQ(1, ExecuteMemoryAccess(isolate, &test_data, stlxr_w));
CHECK_EQ(1, test_data.w);
// Test failure when store between ldaxr/stlxr.
test_data = TestData(1);
ExecuteMemoryAccess(isolate, &test_data, ldaxr_w);
thread.NextAndWait(&test_data, MemoryAccess(Kind::Store, Size::Word,
offsetof(TestData, dummy)));
CHECK_EQ(1, ExecuteMemoryAccess(isolate, &test_data, stlxr_w));
CHECK_EQ(1, test_data.w);
thread.Finish();
thread.Join();
}
#undef __
#endif // USE_SIMULATOR
} // namespace internal
} // namespace v8