Skip to content

Commit

Permalink
buffer: re-enable Fast API for Buffer.write
Browse files Browse the repository at this point in the history 
Re-enables fast Fast API for Buffer.write after fixing
UTF8 handling.

Fixes: #54521
  • Loading branch information
@ronag
ronag committed Sep 2, 2024
1 parent 981c701 commit 0ae2ae1
Show file tree
Hide file tree
Showing 2 changed files with 165 additions and 14 deletions.
134 changes: 120 additions & 14 deletions src/node_buffer.cc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -22,6 +22,7 @@
#include "node_buffer.h"
#include "node.h"
#include "node_blob.h"
#include "node_debug.h"
#include "node_errors.h"
#include "node_external_reference.h"
#include "node_i18n.h"
Expand Down Expand Up @@ -1442,6 +1443,79 @@ void CopyArrayBuffer(const FunctionCallbackInfo<Value>& args) {
memcpy(dest, src, bytes_to_copy);
}

size_t convert_latin1_to_utf8_s(const char* src,
size_t src_len,
char* dst,
size_t dst_len) noexcept {
size_t src_pos = 0;
size_t dst_pos = 0;

const auto safe_len = std::min(src_len, dst_len >> 1);
if (safe_len > 16) {
// convert_latin1_to_utf8 will never write more than input length * 2.
dst_pos += simdutf::convert_latin1_to_utf8(src, safe_len, dst);
src_pos += safe_len;
}

// Based on:
// https://github.com/simdutf/simdutf/blob/master/src/scalar/latin1_to_utf8/latin1_to_utf8.h
// with an upper limit on the number of bytes to write.

const auto src_ptr = reinterpret_cast<const uint8_t*>(src);
const auto dst_ptr = reinterpret_cast<uint8_t*>(dst);

size_t skip_pos = src_pos;
while (src_pos < src_len && dst_pos < dst_len) {
if (skip_pos <= src_pos && src_pos + 16 <= src_len &&
dst_pos + 16 <= dst_len) {
uint64_t v1;
memcpy(&v1, src_ptr + src_pos + 0, 8);
uint64_t v2;
memcpy(&v2, src_ptr + src_pos + 8, 8);
if (((v1 | v2) & UINT64_C(0x8080808080808080)) == 0) {
memcpy(dst_ptr + dst_pos, src_ptr + src_pos, 16);
dst_pos += 16;
src_pos += 16;
} else {
skip_pos = src_pos + 16;
}
} else {
const auto byte = src_ptr[src_pos++];
if ((byte & 0x80) == 0) {
dst_ptr[dst_pos++] = byte;
} else if (dst_pos + 2 <= dst_len) {
dst_ptr[dst_pos++] = (byte >> 6) | 0b11000000;
dst_ptr[dst_pos++] = (byte & 0b111111) | 0b10000000;
} else {
break;
}
}
}

return dst_pos;
}

template <encoding encoding>
uint32_t WriteOneByteString(const char* src,
uint32_t src_len,
char* dst,
uint32_t dst_len) {
if (dst_len == 0) {
return 0;
}

if (encoding == UTF8) {
return convert_latin1_to_utf8_s(src, src_len, dst, dst_len);
} else if (encoding == LATIN1 || encoding == ASCII) {
const auto size = std::min(src_len, dst_len);
memcpy(dst, src, size);
return size;
} else {
// TODO(ronag): Add support for more encoding.
UNREACHABLE();
}
}

template <encoding encoding>
void SlowWriteString(const FunctionCallbackInfo<Value>& args) {
Environment* env = Environment::GetCurrent(args);
Expand All @@ -1464,11 +1538,22 @@ void SlowWriteString(const FunctionCallbackInfo<Value>& args) {

if (max_length == 0) return args.GetReturnValue().Set(0);

uint32_t written = StringBytes::Write(
env->isolate(), ts_obj_data + offset, max_length, str, encoding);
uint32_t written = 0;

if ((encoding == UTF8 || encoding == LATIN1 || encoding == ASCII) &&
str->IsExternalOneByte()) {
const auto src = str->GetExternalOneByteStringResource();
written = WriteOneByteString<encoding>(
src->data(), src->length(), ts_obj_data + offset, max_length);
} else {
written = StringBytes::Write(
env->isolate(), ts_obj_data + offset, max_length, str, encoding);
}

args.GetReturnValue().Set(written);
}

template <encoding encoding>
uint32_t FastWriteString(Local<Value> receiver,
const v8::FastApiTypedArray<uint8_t>& dst,
const v8::FastOneByteString& src,
Expand All @@ -1478,16 +1563,21 @@ uint32_t FastWriteString(Local<Value> receiver,
CHECK(dst.getStorageIfAligned(&dst_data));
CHECK(offset <= dst.length());
CHECK(dst.length() - offset <= std::numeric_limits<uint32_t>::max());
TRACK_V8_FAST_API_CALL("buffer.writeString");

const auto size = std::min(
{static_cast<uint32_t>(dst.length() - offset), max_length, src.length});

memcpy(dst_data + offset, src.data, size);

return size;
return WriteOneByteString<encoding>(
src.data,
src.length,
reinterpret_cast<char*>(dst_data + offset),
std::min<uint32_t>(dst.length() - offset, max_length));
}

static v8::CFunction fast_write_string(v8::CFunction::Make(FastWriteString));
static v8::CFunction fast_write_string_ascii(
v8::CFunction::Make(FastWriteString<ASCII>));
static v8::CFunction fast_write_string_latin1(
v8::CFunction::Make(FastWriteString<LATIN1>));
static v8::CFunction fast_write_string_utf8(
v8::CFunction::Make(FastWriteString<UTF8>));

void Initialize(Local<Object> target,
Local<Value> unused,
Expand Down Expand Up @@ -1554,9 +1644,21 @@ void Initialize(Local<Object> target,
SetMethod(context, target, "hexWrite", StringWrite<HEX>);
SetMethod(context, target, "ucs2Write", StringWrite<UCS2>);

SetMethod(context, target, "asciiWriteStatic", SlowWriteString<ASCII>);
SetMethod(context, target, "latin1WriteStatic", SlowWriteString<LATIN1>);
SetMethod(context, target, "utf8WriteStatic", SlowWriteString<UTF8>);
SetFastMethod(context,
target,
"asciiWriteStatic",
SlowWriteString<ASCII>,
&fast_write_string_ascii);
SetFastMethod(context,
target,
"latin1WriteStatic",
SlowWriteString<LATIN1>,
&fast_write_string_latin1);
SetFastMethod(context,
target,
"utf8WriteStatic",
SlowWriteString<UTF8>,
&fast_write_string_utf8);

SetMethod(context, target, "getZeroFillToggle", GetZeroFillToggle);
}
Expand Down Expand Up @@ -1601,8 +1703,12 @@ void RegisterExternalReferences(ExternalReferenceRegistry* registry) {
registry->Register(SlowWriteString<ASCII>);
registry->Register(SlowWriteString<LATIN1>);
registry->Register(SlowWriteString<UTF8>);
registry->Register(fast_write_string.GetTypeInfo());
registry->Register(FastWriteString);
registry->Register(FastWriteString<ASCII>);
registry->Register(fast_write_string_ascii.GetTypeInfo());
registry->Register(FastWriteString<LATIN1>);
registry->Register(fast_write_string_latin1.GetTypeInfo());
registry->Register(FastWriteString<UTF8>);
registry->Register(fast_write_string_utf8.GetTypeInfo());
registry->Register(StringWrite<ASCII>);
registry->Register(StringWrite<BASE64>);
registry->Register(StringWrite<BASE64URL>);
Expand Down
45 changes: 45 additions & 0 deletions test/parallel/test-buffer-write-fast.js
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,45 @@
// Flags: --expose-internals --no-warnings --allow-natives-syntax
'use strict';

const common = require('../common');
const assert = require('assert');

const { internalBinding } = require('internal/test/binding');

function testFastUtf8Write() {
{
const buf = Buffer.from('\x80');

assert.strictEqual(buf[0], 194);
assert.strictEqual(buf[1], 128);
}

{
const buf = Buffer.alloc(64);
const newBuf = buf.subarray(0, buf.write('éñüçßÆ'));
assert.deepStrictEqual(newBuf, Buffer.from([195, 169, 195, 177, 195, 188, 195, 167, 195, 159, 195, 134]));
}

{
const buf = Buffer.alloc(64);
const newBuf = buf.subarray(0, buf.write('¿'));
assert.deepStrictEqual(newBuf, Buffer.from([194, 191]));
}

{
const buf = Buffer.from(new ArrayBuffer(34), 0, 16);
const str = Buffer.from([50, 83, 127, 39, 104, 8, 74, 65, 108, 123, 5, 4, 82, 10, 7, 53]).toString();
const newBuf = buf.subarray(0, buf.write(str));
assert.deepStrictEqual(newBuf, Buffer.from([ 50, 83, 127, 39, 104, 8, 74, 65, 108, 123, 5, 4, 82, 10, 7, 53]));
}
}

eval('%PrepareFunctionForOptimization(Buffer.prototype.utf8Write)');
testFastUtf8Write();
eval('%OptimizeFunctionOnNextCall(Buffer.prototype.utf8Write)');
testFastUtf8Write();

if (common.isDebug) {
const { getV8FastApiCallCount } = internalBinding('debug');
assert(getV8FastApiCallCount('buffer.writeString'), 4);
}

0 comments on commit 0ae2ae1

Please sign in to comment.