syncthing/lib/scanner/blocks.go
Jakob Borg 2a6f164923 lib/scanner: When scanning a file, stick to the size given by Lstat (fixes #3440)
Otherwise if the file grows during scanning the block list will be out
of sync with the stated size and things get confused. We could fixup the
size afterwards based on the block list, but then we might see other
inconsistencies as the mtime should have changed to reflect the new size
etc. Better stick to the original state and let the next scan pick up
the change.

GitHub-Pull-Request: https://github.com/syncthing/syncthing/pull/3442
2016-07-25 19:16:49 +00:00

228 lines
5.2 KiB
Go

// Copyright (C) 2014 The Syncthing Authors.
//
// This Source Code Form is subject to the terms of the Mozilla Public
// License, v. 2.0. If a copy of the MPL was not distributed with this file,
// You can obtain one at http://mozilla.org/MPL/2.0/.
package scanner
import (
"bytes"
"crypto/sha256"
"fmt"
"io"
"github.com/syncthing/syncthing/lib/protocol"
)
var SHA256OfNothing = []uint8{0xe3, 0xb0, 0xc4, 0x42, 0x98, 0xfc, 0x1c, 0x14, 0x9a, 0xfb, 0xf4, 0xc8, 0x99, 0x6f, 0xb9, 0x24, 0x27, 0xae, 0x41, 0xe4, 0x64, 0x9b, 0x93, 0x4c, 0xa4, 0x95, 0x99, 0x1b, 0x78, 0x52, 0xb8, 0x55}
type Counter interface {
Update(bytes int64)
}
// Blocks returns the blockwise hash of the reader.
func Blocks(r io.Reader, blocksize int, sizehint int64, counter Counter) ([]protocol.BlockInfo, error) {
hf := sha256.New()
hashLength := hf.Size()
var blocks []protocol.BlockInfo
var hashes, thisHash []byte
if sizehint > 0 {
// Allocate contiguous blocks for the BlockInfo structures and their
// hashes once and for all, and stick to the specified size.
r = io.LimitReader(r, sizehint)
numBlocks := int(sizehint / int64(blocksize))
blocks = make([]protocol.BlockInfo, 0, numBlocks)
hashes = make([]byte, 0, hashLength*numBlocks)
}
// A 32k buffer is used for copying into the hash function.
buf := make([]byte, 32<<10)
var offset int64
for {
lr := io.LimitReader(r, int64(blocksize))
n, err := copyBuffer(hf, lr, buf)
if err != nil {
return nil, err
}
if n == 0 {
break
}
if counter != nil {
counter.Update(int64(n))
}
// Carve out a hash-sized chunk of "hashes" to store the hash for this
// block.
hashes = hf.Sum(hashes)
thisHash, hashes = hashes[:hashLength], hashes[hashLength:]
b := protocol.BlockInfo{
Size: int32(n),
Offset: offset,
Hash: thisHash,
}
blocks = append(blocks, b)
offset += int64(n)
hf.Reset()
}
if len(blocks) == 0 {
// Empty file
blocks = append(blocks, protocol.BlockInfo{
Offset: 0,
Size: 0,
Hash: SHA256OfNothing,
})
}
return blocks, nil
}
// PopulateOffsets sets the Offset field on each block
func PopulateOffsets(blocks []protocol.BlockInfo) {
var offset int64
for i := range blocks {
blocks[i].Offset = offset
offset += int64(blocks[i].Size)
}
}
// BlockDiff returns lists of common and missing (to transform src into tgt)
// blocks. Both block lists must have been created with the same block size.
func BlockDiff(src, tgt []protocol.BlockInfo) (have, need []protocol.BlockInfo) {
if len(tgt) == 0 && len(src) != 0 {
return nil, nil
}
if len(tgt) != 0 && len(src) == 0 {
// Copy the entire file
return nil, tgt
}
for i := range tgt {
if i >= len(src) || !bytes.Equal(tgt[i].Hash, src[i].Hash) {
// Copy differing block
need = append(need, tgt[i])
} else {
have = append(have, tgt[i])
}
}
return have, need
}
// Verify returns nil or an error describing the mismatch between the block
// list and actual reader contents
func Verify(r io.Reader, blocksize int, blocks []protocol.BlockInfo) error {
hf := sha256.New()
for i, block := range blocks {
lr := &io.LimitedReader{R: r, N: int64(blocksize)}
_, err := io.Copy(hf, lr)
if err != nil {
return err
}
hash := hf.Sum(nil)
hf.Reset()
if !bytes.Equal(hash, block.Hash) {
return fmt.Errorf("hash mismatch %x != %x for block %d", hash, block.Hash, i)
}
}
// We should have reached the end now
bs := make([]byte, 1)
n, err := r.Read(bs)
if n != 0 || err != io.EOF {
return fmt.Errorf("file continues past end of blocks")
}
return nil
}
func VerifyBuffer(buf []byte, block protocol.BlockInfo) ([]byte, error) {
if len(buf) != int(block.Size) {
return nil, fmt.Errorf("length mismatch %d != %d", len(buf), block.Size)
}
hf := sha256.New()
_, err := hf.Write(buf)
if err != nil {
return nil, err
}
hash := hf.Sum(nil)
if !bytes.Equal(hash, block.Hash) {
return hash, fmt.Errorf("hash mismatch %x != %x", hash, block.Hash)
}
return hash, nil
}
// BlocksEqual returns whether two slices of blocks are exactly the same hash
// and index pair wise.
func BlocksEqual(src, tgt []protocol.BlockInfo) bool {
if len(tgt) != len(src) {
return false
}
for i, sblk := range src {
if !bytes.Equal(sblk.Hash, tgt[i].Hash) {
return false
}
}
return true
}
// This is a copy & paste of io.copyBuffer from the Go 1.5 standard library,
// as we want this but also want to build with Go 1.3+.
// copyBuffer is the actual implementation of Copy and CopyBuffer.
// if buf is nil, one is allocated.
func copyBuffer(dst io.Writer, src io.Reader, buf []byte) (written int64, err error) {
// If the reader has a WriteTo method, use it to do the copy.
// Avoids an allocation and a copy.
if wt, ok := src.(io.WriterTo); ok {
return wt.WriteTo(dst)
}
// Similarly, if the writer has a ReadFrom method, use it to do the copy.
if rt, ok := dst.(io.ReaderFrom); ok {
return rt.ReadFrom(src)
}
if buf == nil {
buf = make([]byte, 32*1024)
}
for {
nr, er := src.Read(buf)
if nr > 0 {
nw, ew := dst.Write(buf[0:nr])
if nw > 0 {
written += int64(nw)
}
if ew != nil {
err = ew
break
}
if nr != nw {
err = io.ErrShortWrite
break
}
}
if er == io.EOF {
break
}
if er != nil {
err = er
break
}
}
return written, err
}