b4d7209428
Summary: The first key is used to defer reading the data block until this file gets to the top of merging iterator's heap. For short range scans, most files never make it to the top of the heap, so this change can reduce read amplification by a lot sometimes. Consider the following workload. There are a few data streams (we'll be calling them "logs"), each stream consisting of a sequence of blobs (we'll be calling them "records"). Each record is identified by log ID and a sequence number within the log. RocksDB key is concatenation of log ID and sequence number (big endian). Reads are mostly relatively short range scans, each within a single log. Writes are mostly sequential for each log, but writes to different logs are randomly interleaved. Compactions are disabled; instead, when we accumulate a few tens of sst files, we create a new column family and start writing to it. So, a typical sst file consists of a few ranges of blocks, each range corresponding to one log ID (we use FlushBlockPolicy to cut blocks at log boundaries). A typical read would go like this. First, iterator Seek() reads one block from each sst file. Then a series of Next()s move through one sst file (since writes to each log are mostly sequential) until the subiterator reaches the end of this log in this sst file; then Next() switches to the next sst file and reads sequentially from that, and so on. Often a range scan will only return records from a small number of blocks in small number of sst files; in this case, the cost of initial Seek() reading one block from each file may be bigger than the cost of reading the actually useful blocks. Neither iterate_upper_bound nor bloom filters can prevent reading one block from each file in Seek(). But this PR can: if the index contains first key from each block, we don't have to read the block until this block actually makes it to the top of merging iterator's heap, so for short range scans we won't read any blocks from most of the sst files. This PR does the deferred block loading inside value() call. This is not ideal: there's no good way to report an IO error from inside value(). As discussed with siying offline, it would probably be better to change InternalIterator's interface to explicitly fetch deferred value and get status. I'll do it in a separate PR. Pull Request resolved: https://github.com/facebook/rocksdb/pull/5289 Differential Revision: D15256423 Pulled By: al13n321 fbshipit-source-id: 750e4c39ce88e8d41662f701cf6275d9388ba46a
469 lines
15 KiB
C++
469 lines
15 KiB
C++
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
|
// This source code is licensed under both the GPLv2 (found in the
|
|
// COPYING file in the root directory) and Apache 2.0 License
|
|
// (found in the LICENSE.Apache file in the root directory).
|
|
//
|
|
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
|
|
// Use of this source code is governed by a BSD-style license that can be
|
|
// found in the LICENSE file. See the AUTHORS file for names of contributors.
|
|
//
|
|
// Endian-neutral encoding:
|
|
// * Fixed-length numbers are encoded with least-significant byte first
|
|
// * In addition we support variable length "varint" encoding
|
|
// * Strings are encoded prefixed by their length in varint format
|
|
|
|
#pragma once
|
|
#include <algorithm>
|
|
#include <stdint.h>
|
|
#include <string.h>
|
|
#include <string>
|
|
|
|
#include "rocksdb/write_batch.h"
|
|
#include "port/port.h"
|
|
|
|
// Some processors does not allow unaligned access to memory
|
|
#if defined(__sparc)
|
|
#define PLATFORM_UNALIGNED_ACCESS_NOT_ALLOWED
|
|
#endif
|
|
|
|
namespace rocksdb {
|
|
|
|
// The maximum length of a varint in bytes for 64-bit.
|
|
const unsigned int kMaxVarint64Length = 10;
|
|
|
|
// Standard Put... routines append to a string
|
|
extern void PutFixed16(std::string* dst, uint16_t value);
|
|
extern void PutFixed32(std::string* dst, uint32_t value);
|
|
extern void PutFixed64(std::string* dst, uint64_t value);
|
|
extern void PutVarint32(std::string* dst, uint32_t value);
|
|
extern void PutVarint32Varint32(std::string* dst, uint32_t value1,
|
|
uint32_t value2);
|
|
extern void PutVarint32Varint32Varint32(std::string* dst, uint32_t value1,
|
|
uint32_t value2, uint32_t value3);
|
|
extern void PutVarint64(std::string* dst, uint64_t value);
|
|
extern void PutVarint64Varint64(std::string* dst, uint64_t value1,
|
|
uint64_t value2);
|
|
extern void PutVarint32Varint64(std::string* dst, uint32_t value1,
|
|
uint64_t value2);
|
|
extern void PutVarint32Varint32Varint64(std::string* dst, uint32_t value1,
|
|
uint32_t value2, uint64_t value3);
|
|
extern void PutLengthPrefixedSlice(std::string* dst, const Slice& value);
|
|
extern void PutLengthPrefixedSliceParts(std::string* dst,
|
|
const SliceParts& slice_parts);
|
|
|
|
// Standard Get... routines parse a value from the beginning of a Slice
|
|
// and advance the slice past the parsed value.
|
|
extern bool GetFixed64(Slice* input, uint64_t* value);
|
|
extern bool GetFixed32(Slice* input, uint32_t* value);
|
|
extern bool GetFixed16(Slice* input, uint16_t* value);
|
|
extern bool GetVarint32(Slice* input, uint32_t* value);
|
|
extern bool GetVarint64(Slice* input, uint64_t* value);
|
|
extern bool GetVarsignedint64(Slice* input, int64_t* value);
|
|
extern bool GetLengthPrefixedSlice(Slice* input, Slice* result);
|
|
// This function assumes data is well-formed.
|
|
extern Slice GetLengthPrefixedSlice(const char* data);
|
|
|
|
extern Slice GetSliceUntil(Slice* slice, char delimiter);
|
|
|
|
// Borrowed from
|
|
// https://github.com/facebook/fbthrift/blob/449a5f77f9f9bae72c9eb5e78093247eef185c04/thrift/lib/cpp/util/VarintUtils-inl.h#L202-L208
|
|
constexpr inline uint64_t i64ToZigzag(const int64_t l) {
|
|
return (static_cast<uint64_t>(l) << 1) ^ static_cast<uint64_t>(l >> 63);
|
|
}
|
|
inline int64_t zigzagToI64(uint64_t n) {
|
|
return (n >> 1) ^ -static_cast<int64_t>(n & 1);
|
|
}
|
|
|
|
// Pointer-based variants of GetVarint... These either store a value
|
|
// in *v and return a pointer just past the parsed value, or return
|
|
// nullptr on error. These routines only look at bytes in the range
|
|
// [p..limit-1]
|
|
extern const char* GetVarint32Ptr(const char* p,const char* limit, uint32_t* v);
|
|
extern const char* GetVarint64Ptr(const char* p,const char* limit, uint64_t* v);
|
|
inline const char* GetVarsignedint64Ptr(const char* p, const char* limit,
|
|
int64_t* value) {
|
|
uint64_t u = 0;
|
|
const char* ret = GetVarint64Ptr(p, limit, &u);
|
|
*value = zigzagToI64(u);
|
|
return ret;
|
|
}
|
|
|
|
// Returns the length of the varint32 or varint64 encoding of "v"
|
|
extern int VarintLength(uint64_t v);
|
|
|
|
// Lower-level versions of Put... that write directly into a character buffer
|
|
// REQUIRES: dst has enough space for the value being written
|
|
extern void EncodeFixed16(char* dst, uint16_t value);
|
|
extern void EncodeFixed32(char* dst, uint32_t value);
|
|
extern void EncodeFixed64(char* dst, uint64_t value);
|
|
|
|
// Lower-level versions of Put... that write directly into a character buffer
|
|
// and return a pointer just past the last byte written.
|
|
// REQUIRES: dst has enough space for the value being written
|
|
extern char* EncodeVarint32(char* dst, uint32_t value);
|
|
extern char* EncodeVarint64(char* dst, uint64_t value);
|
|
|
|
// Lower-level versions of Get... that read directly from a character buffer
|
|
// without any bounds checking.
|
|
|
|
inline uint16_t DecodeFixed16(const char* ptr) {
|
|
if (port::kLittleEndian) {
|
|
// Load the raw bytes
|
|
uint16_t result;
|
|
memcpy(&result, ptr, sizeof(result)); // gcc optimizes this to a plain load
|
|
return result;
|
|
} else {
|
|
return ((static_cast<uint16_t>(static_cast<unsigned char>(ptr[0]))) |
|
|
(static_cast<uint16_t>(static_cast<unsigned char>(ptr[1])) << 8));
|
|
}
|
|
}
|
|
|
|
inline uint32_t DecodeFixed32(const char* ptr) {
|
|
if (port::kLittleEndian) {
|
|
// Load the raw bytes
|
|
uint32_t result;
|
|
memcpy(&result, ptr, sizeof(result)); // gcc optimizes this to a plain load
|
|
return result;
|
|
} else {
|
|
return ((static_cast<uint32_t>(static_cast<unsigned char>(ptr[0])))
|
|
| (static_cast<uint32_t>(static_cast<unsigned char>(ptr[1])) << 8)
|
|
| (static_cast<uint32_t>(static_cast<unsigned char>(ptr[2])) << 16)
|
|
| (static_cast<uint32_t>(static_cast<unsigned char>(ptr[3])) << 24));
|
|
}
|
|
}
|
|
|
|
inline uint64_t DecodeFixed64(const char* ptr) {
|
|
if (port::kLittleEndian) {
|
|
// Load the raw bytes
|
|
uint64_t result;
|
|
memcpy(&result, ptr, sizeof(result)); // gcc optimizes this to a plain load
|
|
return result;
|
|
} else {
|
|
uint64_t lo = DecodeFixed32(ptr);
|
|
uint64_t hi = DecodeFixed32(ptr + 4);
|
|
return (hi << 32) | lo;
|
|
}
|
|
}
|
|
|
|
// Internal routine for use by fallback path of GetVarint32Ptr
|
|
extern const char* GetVarint32PtrFallback(const char* p,
|
|
const char* limit,
|
|
uint32_t* value);
|
|
inline const char* GetVarint32Ptr(const char* p,
|
|
const char* limit,
|
|
uint32_t* value) {
|
|
if (p < limit) {
|
|
uint32_t result = *(reinterpret_cast<const unsigned char*>(p));
|
|
if ((result & 128) == 0) {
|
|
*value = result;
|
|
return p + 1;
|
|
}
|
|
}
|
|
return GetVarint32PtrFallback(p, limit, value);
|
|
}
|
|
|
|
// -- Implementation of the functions declared above
|
|
inline void EncodeFixed16(char* buf, uint16_t value) {
|
|
if (port::kLittleEndian) {
|
|
memcpy(buf, &value, sizeof(value));
|
|
} else {
|
|
buf[0] = value & 0xff;
|
|
buf[1] = (value >> 8) & 0xff;
|
|
}
|
|
}
|
|
|
|
inline void EncodeFixed32(char* buf, uint32_t value) {
|
|
if (port::kLittleEndian) {
|
|
memcpy(buf, &value, sizeof(value));
|
|
} else {
|
|
buf[0] = value & 0xff;
|
|
buf[1] = (value >> 8) & 0xff;
|
|
buf[2] = (value >> 16) & 0xff;
|
|
buf[3] = (value >> 24) & 0xff;
|
|
}
|
|
}
|
|
|
|
inline void EncodeFixed64(char* buf, uint64_t value) {
|
|
if (port::kLittleEndian) {
|
|
memcpy(buf, &value, sizeof(value));
|
|
} else {
|
|
buf[0] = value & 0xff;
|
|
buf[1] = (value >> 8) & 0xff;
|
|
buf[2] = (value >> 16) & 0xff;
|
|
buf[3] = (value >> 24) & 0xff;
|
|
buf[4] = (value >> 32) & 0xff;
|
|
buf[5] = (value >> 40) & 0xff;
|
|
buf[6] = (value >> 48) & 0xff;
|
|
buf[7] = (value >> 56) & 0xff;
|
|
}
|
|
}
|
|
|
|
// Pull the last 8 bits and cast it to a character
|
|
inline void PutFixed16(std::string* dst, uint16_t value) {
|
|
if (port::kLittleEndian) {
|
|
dst->append(const_cast<const char*>(reinterpret_cast<char*>(&value)),
|
|
sizeof(value));
|
|
} else {
|
|
char buf[sizeof(value)];
|
|
EncodeFixed16(buf, value);
|
|
dst->append(buf, sizeof(buf));
|
|
}
|
|
}
|
|
|
|
inline void PutFixed32(std::string* dst, uint32_t value) {
|
|
if (port::kLittleEndian) {
|
|
dst->append(const_cast<const char*>(reinterpret_cast<char*>(&value)),
|
|
sizeof(value));
|
|
} else {
|
|
char buf[sizeof(value)];
|
|
EncodeFixed32(buf, value);
|
|
dst->append(buf, sizeof(buf));
|
|
}
|
|
}
|
|
|
|
inline void PutFixed64(std::string* dst, uint64_t value) {
|
|
if (port::kLittleEndian) {
|
|
dst->append(const_cast<const char*>(reinterpret_cast<char*>(&value)),
|
|
sizeof(value));
|
|
} else {
|
|
char buf[sizeof(value)];
|
|
EncodeFixed64(buf, value);
|
|
dst->append(buf, sizeof(buf));
|
|
}
|
|
}
|
|
|
|
inline void PutVarint32(std::string* dst, uint32_t v) {
|
|
char buf[5];
|
|
char* ptr = EncodeVarint32(buf, v);
|
|
dst->append(buf, static_cast<size_t>(ptr - buf));
|
|
}
|
|
|
|
inline void PutVarint32Varint32(std::string* dst, uint32_t v1, uint32_t v2) {
|
|
char buf[10];
|
|
char* ptr = EncodeVarint32(buf, v1);
|
|
ptr = EncodeVarint32(ptr, v2);
|
|
dst->append(buf, static_cast<size_t>(ptr - buf));
|
|
}
|
|
|
|
inline void PutVarint32Varint32Varint32(std::string* dst, uint32_t v1,
|
|
uint32_t v2, uint32_t v3) {
|
|
char buf[15];
|
|
char* ptr = EncodeVarint32(buf, v1);
|
|
ptr = EncodeVarint32(ptr, v2);
|
|
ptr = EncodeVarint32(ptr, v3);
|
|
dst->append(buf, static_cast<size_t>(ptr - buf));
|
|
}
|
|
|
|
inline char* EncodeVarint64(char* dst, uint64_t v) {
|
|
static const unsigned int B = 128;
|
|
unsigned char* ptr = reinterpret_cast<unsigned char*>(dst);
|
|
while (v >= B) {
|
|
*(ptr++) = (v & (B - 1)) | B;
|
|
v >>= 7;
|
|
}
|
|
*(ptr++) = static_cast<unsigned char>(v);
|
|
return reinterpret_cast<char*>(ptr);
|
|
}
|
|
|
|
inline void PutVarint64(std::string* dst, uint64_t v) {
|
|
char buf[kMaxVarint64Length];
|
|
char* ptr = EncodeVarint64(buf, v);
|
|
dst->append(buf, static_cast<size_t>(ptr - buf));
|
|
}
|
|
|
|
inline void PutVarsignedint64(std::string* dst, int64_t v) {
|
|
char buf[kMaxVarint64Length];
|
|
// Using Zigzag format to convert signed to unsigned
|
|
char* ptr = EncodeVarint64(buf, i64ToZigzag(v));
|
|
dst->append(buf, static_cast<size_t>(ptr - buf));
|
|
}
|
|
|
|
inline void PutVarint64Varint64(std::string* dst, uint64_t v1, uint64_t v2) {
|
|
char buf[20];
|
|
char* ptr = EncodeVarint64(buf, v1);
|
|
ptr = EncodeVarint64(ptr, v2);
|
|
dst->append(buf, static_cast<size_t>(ptr - buf));
|
|
}
|
|
|
|
inline void PutVarint32Varint64(std::string* dst, uint32_t v1, uint64_t v2) {
|
|
char buf[15];
|
|
char* ptr = EncodeVarint32(buf, v1);
|
|
ptr = EncodeVarint64(ptr, v2);
|
|
dst->append(buf, static_cast<size_t>(ptr - buf));
|
|
}
|
|
|
|
inline void PutVarint32Varint32Varint64(std::string* dst, uint32_t v1,
|
|
uint32_t v2, uint64_t v3) {
|
|
char buf[20];
|
|
char* ptr = EncodeVarint32(buf, v1);
|
|
ptr = EncodeVarint32(ptr, v2);
|
|
ptr = EncodeVarint64(ptr, v3);
|
|
dst->append(buf, static_cast<size_t>(ptr - buf));
|
|
}
|
|
|
|
inline void PutLengthPrefixedSlice(std::string* dst, const Slice& value) {
|
|
PutVarint32(dst, static_cast<uint32_t>(value.size()));
|
|
dst->append(value.data(), value.size());
|
|
}
|
|
|
|
inline void PutLengthPrefixedSliceParts(std::string* dst,
|
|
const SliceParts& slice_parts) {
|
|
size_t total_bytes = 0;
|
|
for (int i = 0; i < slice_parts.num_parts; ++i) {
|
|
total_bytes += slice_parts.parts[i].size();
|
|
}
|
|
PutVarint32(dst, static_cast<uint32_t>(total_bytes));
|
|
for (int i = 0; i < slice_parts.num_parts; ++i) {
|
|
dst->append(slice_parts.parts[i].data(), slice_parts.parts[i].size());
|
|
}
|
|
}
|
|
|
|
inline int VarintLength(uint64_t v) {
|
|
int len = 1;
|
|
while (v >= 128) {
|
|
v >>= 7;
|
|
len++;
|
|
}
|
|
return len;
|
|
}
|
|
|
|
inline bool GetFixed64(Slice* input, uint64_t* value) {
|
|
if (input->size() < sizeof(uint64_t)) {
|
|
return false;
|
|
}
|
|
*value = DecodeFixed64(input->data());
|
|
input->remove_prefix(sizeof(uint64_t));
|
|
return true;
|
|
}
|
|
|
|
inline bool GetFixed32(Slice* input, uint32_t* value) {
|
|
if (input->size() < sizeof(uint32_t)) {
|
|
return false;
|
|
}
|
|
*value = DecodeFixed32(input->data());
|
|
input->remove_prefix(sizeof(uint32_t));
|
|
return true;
|
|
}
|
|
|
|
inline bool GetFixed16(Slice* input, uint16_t* value) {
|
|
if (input->size() < sizeof(uint16_t)) {
|
|
return false;
|
|
}
|
|
*value = DecodeFixed16(input->data());
|
|
input->remove_prefix(sizeof(uint16_t));
|
|
return true;
|
|
}
|
|
|
|
inline bool GetVarint32(Slice* input, uint32_t* value) {
|
|
const char* p = input->data();
|
|
const char* limit = p + input->size();
|
|
const char* q = GetVarint32Ptr(p, limit, value);
|
|
if (q == nullptr) {
|
|
return false;
|
|
} else {
|
|
*input = Slice(q, static_cast<size_t>(limit - q));
|
|
return true;
|
|
}
|
|
}
|
|
|
|
inline bool GetVarint64(Slice* input, uint64_t* value) {
|
|
const char* p = input->data();
|
|
const char* limit = p + input->size();
|
|
const char* q = GetVarint64Ptr(p, limit, value);
|
|
if (q == nullptr) {
|
|
return false;
|
|
} else {
|
|
*input = Slice(q, static_cast<size_t>(limit - q));
|
|
return true;
|
|
}
|
|
}
|
|
|
|
inline bool GetVarsignedint64(Slice* input, int64_t* value) {
|
|
const char* p = input->data();
|
|
const char* limit = p + input->size();
|
|
const char* q = GetVarsignedint64Ptr(p, limit, value);
|
|
if (q == nullptr) {
|
|
return false;
|
|
} else {
|
|
*input = Slice(q, static_cast<size_t>(limit - q));
|
|
return true;
|
|
}
|
|
}
|
|
|
|
// Provide an interface for platform independent endianness transformation
|
|
inline uint64_t EndianTransform(uint64_t input, size_t size) {
|
|
char* pos = reinterpret_cast<char*>(&input);
|
|
uint64_t ret_val = 0;
|
|
for (size_t i = 0; i < size; ++i) {
|
|
ret_val |= (static_cast<uint64_t>(static_cast<unsigned char>(pos[i]))
|
|
<< ((size - i - 1) << 3));
|
|
}
|
|
return ret_val;
|
|
}
|
|
|
|
inline bool GetLengthPrefixedSlice(Slice* input, Slice* result) {
|
|
uint32_t len = 0;
|
|
if (GetVarint32(input, &len) && input->size() >= len) {
|
|
*result = Slice(input->data(), len);
|
|
input->remove_prefix(len);
|
|
return true;
|
|
} else {
|
|
return false;
|
|
}
|
|
}
|
|
|
|
inline Slice GetLengthPrefixedSlice(const char* data) {
|
|
uint32_t len = 0;
|
|
// +5: we assume "data" is not corrupted
|
|
// unsigned char is 7 bits, uint32_t is 32 bits, need 5 unsigned char
|
|
auto p = GetVarint32Ptr(data, data + 5 /* limit */, &len);
|
|
return Slice(p, len);
|
|
}
|
|
|
|
inline Slice GetSliceUntil(Slice* slice, char delimiter) {
|
|
uint32_t len = 0;
|
|
for (len = 0; len < slice->size() && slice->data()[len] != delimiter; ++len) {
|
|
// nothing
|
|
}
|
|
|
|
Slice ret(slice->data(), len);
|
|
slice->remove_prefix(len + ((len < slice->size()) ? 1 : 0));
|
|
return ret;
|
|
}
|
|
|
|
template<class T>
|
|
#ifdef ROCKSDB_UBSAN_RUN
|
|
#if defined(__clang__)
|
|
__attribute__((__no_sanitize__("alignment")))
|
|
#elif defined(__GNUC__)
|
|
__attribute__((__no_sanitize_undefined__))
|
|
#endif
|
|
#endif
|
|
inline void PutUnaligned(T *memory, const T &value) {
|
|
#if defined(PLATFORM_UNALIGNED_ACCESS_NOT_ALLOWED)
|
|
char *nonAlignedMemory = reinterpret_cast<char*>(memory);
|
|
memcpy(nonAlignedMemory, reinterpret_cast<const char*>(&value), sizeof(T));
|
|
#else
|
|
*memory = value;
|
|
#endif
|
|
}
|
|
|
|
template<class T>
|
|
#ifdef ROCKSDB_UBSAN_RUN
|
|
#if defined(__clang__)
|
|
__attribute__((__no_sanitize__("alignment")))
|
|
#elif defined(__GNUC__)
|
|
__attribute__((__no_sanitize_undefined__))
|
|
#endif
|
|
#endif
|
|
inline void GetUnaligned(const T *memory, T *value) {
|
|
#if defined(PLATFORM_UNALIGNED_ACCESS_NOT_ALLOWED)
|
|
char *nonAlignedMemory = reinterpret_cast<char*>(value);
|
|
memcpy(nonAlignedMemory, reinterpret_cast<const char*>(memory), sizeof(T));
|
|
#else
|
|
*value = *memory;
|
|
#endif
|
|
}
|
|
|
|
} // namespace rocksdb
|