Files
chatlog/internal/wechat/key/darwin/v4.go
2025-04-17 18:42:45 +08:00

282 lines
6.4 KiB
Go

package darwin
import (
"bytes"
"context"
"encoding/hex"
"runtime"
"sync"
"github.com/rs/zerolog/log"
"github.com/sjzar/chatlog/internal/errors"
"github.com/sjzar/chatlog/internal/wechat/decrypt"
"github.com/sjzar/chatlog/internal/wechat/key/darwin/glance"
"github.com/sjzar/chatlog/internal/wechat/model"
)
const (
MaxWorkers = 8
MinChunkSize = 1 * 1024 * 1024 // 1MB
ChunkOverlapBytes = 1024 // Greater than all offsets
ChunkMultiplier = 2 // Number of chunks = MaxWorkers * ChunkMultiplier
)
var V4KeyPatterns = []KeyPatternInfo{
{
Pattern: []byte{0x20, 0x66, 0x74, 0x73, 0x35, 0x28, 0x25, 0x00},
Offsets: []int{16, -80, 64},
},
}
type V4Extractor struct {
validator *decrypt.Validator
keyPatterns []KeyPatternInfo
}
func NewV4Extractor() *V4Extractor {
return &V4Extractor{
keyPatterns: V4KeyPatterns,
}
}
func (e *V4Extractor) Extract(ctx context.Context, proc *model.Process) (string, error) {
if proc.Status == model.StatusOffline {
return "", errors.ErrWeChatOffline
}
// Check if SIP is disabled, as it's required for memory reading on macOS
if !glance.IsSIPDisabled() {
return "", errors.ErrSIPEnabled
}
if e.validator == nil {
return "", errors.ErrValidatorNotSet
}
// Create context to control all goroutines
searchCtx, cancel := context.WithCancel(ctx)
defer cancel()
// Create channels for memory data and results
memoryChannel := make(chan []byte, 100)
resultChannel := make(chan string, 1)
// Determine number of worker goroutines
workerCount := runtime.NumCPU()
if workerCount < 2 {
workerCount = 2
}
if workerCount > MaxWorkers {
workerCount = MaxWorkers
}
log.Debug().Msgf("Starting %d workers for V4 key search", workerCount)
// Start consumer goroutines
var workerWaitGroup sync.WaitGroup
workerWaitGroup.Add(workerCount)
for index := 0; index < workerCount; index++ {
go func() {
defer workerWaitGroup.Done()
e.worker(searchCtx, memoryChannel, resultChannel)
}()
}
// Start producer goroutine
var producerWaitGroup sync.WaitGroup
producerWaitGroup.Add(1)
go func() {
defer producerWaitGroup.Done()
defer close(memoryChannel) // Close channel when producer is done
err := e.findMemory(searchCtx, uint32(proc.PID), memoryChannel)
if err != nil {
log.Err(err).Msg("Failed to read memory")
}
}()
// Wait for producer and consumers to complete
go func() {
producerWaitGroup.Wait()
workerWaitGroup.Wait()
close(resultChannel)
}()
// Wait for result
select {
case <-ctx.Done():
return "", ctx.Err()
case result, ok := <-resultChannel:
if ok && result != "" {
return result, nil
}
}
return "", errors.ErrNoValidKey
}
// findMemory searches for memory regions using Glance
func (e *V4Extractor) findMemory(ctx context.Context, pid uint32, memoryChannel chan<- []byte) error {
// Initialize a Glance instance to read process memory
g := glance.NewGlance(pid)
// Read memory data
memory, err := g.Read()
if err != nil {
return err
}
totalSize := len(memory)
log.Debug().Msgf("Read memory region, size: %d bytes", totalSize)
// If memory is small enough, process it as a single chunk
if totalSize <= MinChunkSize {
select {
case memoryChannel <- memory:
log.Debug().Msg("Memory sent as a single chunk for analysis")
case <-ctx.Done():
return ctx.Err()
}
return nil
}
chunkCount := MaxWorkers * ChunkMultiplier
// Calculate chunk size based on fixed chunk count
chunkSize := totalSize / chunkCount
if chunkSize < MinChunkSize {
// Reduce number of chunks if each would be too small
chunkCount = totalSize / MinChunkSize
if chunkCount == 0 {
chunkCount = 1
}
chunkSize = totalSize / chunkCount
}
// Process memory in chunks from end to beginning
for i := chunkCount - 1; i >= 0; i-- {
select {
case <-ctx.Done():
return ctx.Err()
default:
// Calculate start and end positions for this chunk
start := i * chunkSize
end := (i + 1) * chunkSize
// Ensure the last chunk includes all remaining memory
if i == chunkCount-1 {
end = totalSize
}
// Add overlap area to catch patterns at chunk boundaries
if i > 0 {
start -= ChunkOverlapBytes
if start < 0 {
start = 0
}
}
chunk := memory[start:end]
log.Debug().
Int("chunk_index", i+1).
Int("total_chunks", chunkCount).
Int("chunk_size", len(chunk)).
Int("start_offset", start).
Int("end_offset", end).
Msg("Processing memory chunk")
select {
case memoryChannel <- chunk:
case <-ctx.Done():
return ctx.Err()
}
}
}
return nil
}
// worker processes memory regions to find V4 version key
func (e *V4Extractor) worker(ctx context.Context, memoryChannel <-chan []byte, resultChannel chan<- string) {
for {
select {
case <-ctx.Done():
return
case memory, ok := <-memoryChannel:
if !ok {
return
}
if key, ok := e.SearchKey(ctx, memory); ok {
select {
case resultChannel <- key:
default:
}
}
}
}
}
func (e *V4Extractor) SearchKey(ctx context.Context, memory []byte) (string, bool) {
for _, keyPattern := range e.keyPatterns {
index := len(memory)
for {
select {
case <-ctx.Done():
return "", false
default:
}
// Find pattern from end to beginning
index = bytes.LastIndex(memory[:index], keyPattern.Pattern)
if index == -1 {
break // No more matches found
}
// Try each offset for this pattern
for _, offset := range keyPattern.Offsets {
// Check if we have enough space for the key
keyOffset := index + offset
if keyOffset < 0 || keyOffset+32 > len(memory) {
continue
}
// Extract the key data, which is at the offset position and 32 bytes long
keyData := memory[keyOffset : keyOffset+32]
// Validate key against database header
if keyData, ok := e.validate(ctx, keyData); ok {
log.Debug().
Str("pattern", hex.EncodeToString(keyPattern.Pattern)).
Int("offset", offset).
Str("key", hex.EncodeToString(keyData)).
Msg("Key found")
return hex.EncodeToString(keyData), true
}
}
index -= 1
}
}
return "", false
}
func (e *V4Extractor) validate(ctx context.Context, keyDate []byte) ([]byte, bool) {
if e.validator.Validate(keyDate) {
return keyDate, true
}
// Try to find a valid key by ***
return nil, false
}
func (e *V4Extractor) SetValidate(validator *decrypt.Validator) {
e.validator = validator
}
type KeyPatternInfo struct {
Pattern []byte
Offsets []int
}