|
|
|
/*
|
|
|
|
* MinIO Cloud Storage, (C) 2019 MinIO, Inc.
|
|
|
|
*
|
|
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
|
|
* you may not use this file except in compliance with the License.
|
|
|
|
* You may obtain a copy of the License at
|
|
|
|
*
|
|
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
|
|
*
|
|
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
|
|
* See the License for the specific language governing permissions and
|
|
|
|
* limitations under the License.
|
|
|
|
*/
|
|
|
|
|
|
|
|
package cmd
|
|
|
|
|
|
|
|
import (
|
|
|
|
"context"
|
|
|
|
"hash"
|
|
|
|
"io"
|
|
|
|
|
|
|
|
"github.com/minio/minio/cmd/logger"
|
|
|
|
)
|
|
|
|
|
|
|
|
// Implementation to calculate bitrot for the whole file.
|
|
|
|
type wholeBitrotWriter struct {
|
|
|
|
disk StorageAPI
|
|
|
|
volume string
|
|
|
|
filePath string
|
|
|
|
shardSize int64 // This is the shard size of the erasure logic
|
|
|
|
hash.Hash // For bitrot hash
|
|
|
|
|
|
|
|
// Following two fields are used only to make sure that Write(p) is called such that
|
|
|
|
// len(p) is always the block size except the last block and prevent programmer errors.
|
|
|
|
currentBlockIdx int
|
|
|
|
lastBlockIdx int
|
|
|
|
}
|
|
|
|
|
|
|
|
func (b *wholeBitrotWriter) Write(p []byte) (int, error) {
|
|
|
|
if b.currentBlockIdx < b.lastBlockIdx && int64(len(p)) != b.shardSize {
|
|
|
|
// All blocks except last should be of the length b.shardSize
|
|
|
|
logger.LogIf(context.Background(), errUnexpected)
|
|
|
|
return 0, errUnexpected
|
|
|
|
}
|
|
|
|
err := b.disk.AppendFile(b.volume, b.filePath, p)
|
|
|
|
if err != nil {
|
|
|
|
logger.LogIf(context.Background(), err)
|
|
|
|
return 0, err
|
|
|
|
}
|
|
|
|
_, err = b.Hash.Write(p)
|
|
|
|
if err != nil {
|
|
|
|
logger.LogIf(context.Background(), err)
|
|
|
|
return 0, err
|
|
|
|
}
|
|
|
|
b.currentBlockIdx++
|
|
|
|
return len(p), nil
|
|
|
|
}
|
|
|
|
|
|
|
|
func (b *wholeBitrotWriter) Close() error {
|
|
|
|
return nil
|
|
|
|
}
|
|
|
|
|
|
|
|
// Returns whole-file bitrot writer.
|
|
|
|
func newWholeBitrotWriter(disk StorageAPI, volume, filePath string, length int64, algo BitrotAlgorithm, shardSize int64) io.WriteCloser {
|
|
|
|
return &wholeBitrotWriter{disk, volume, filePath, shardSize, algo.New(), 0, int(length / shardSize)}
|
|
|
|
}
|
|
|
|
|
|
|
|
// Implementation to verify bitrot for the whole file.
|
|
|
|
type wholeBitrotReader struct {
|
|
|
|
disk StorageAPI
|
|
|
|
volume string
|
|
|
|
filePath string
|
|
|
|
verifier *BitrotVerifier // Holds the bit-rot info
|
|
|
|
tillOffset int64 // Affects the length of data requested in disk.ReadFile depending on Read()'s offset
|
|
|
|
buf []byte // Holds bit-rot verified data
|
|
|
|
}
|
|
|
|
|
|
|
|
func (b *wholeBitrotReader) ReadAt(buf []byte, offset int64) (n int, err error) {
|
|
|
|
if b.buf == nil {
|
|
|
|
b.buf = make([]byte, b.tillOffset-offset)
|
|
|
|
if _, err := b.disk.ReadFile(b.volume, b.filePath, offset, b.buf, b.verifier); err != nil {
|
|
|
|
ctx := context.Background()
|
|
|
|
logger.GetReqInfo(ctx).AppendTags("disk", b.disk.String())
|
|
|
|
logger.LogIf(ctx, err)
|
|
|
|
return 0, err
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if len(b.buf) < len(buf) {
|
|
|
|
logger.LogIf(context.Background(), errLessData)
|
|
|
|
return 0, errLessData
|
|
|
|
}
|
|
|
|
n = copy(buf, b.buf)
|
|
|
|
b.buf = b.buf[n:]
|
|
|
|
return n, nil
|
|
|
|
}
|
|
|
|
|
|
|
|
// Returns whole-file bitrot reader.
|
|
|
|
func newWholeBitrotReader(disk StorageAPI, volume, filePath string, algo BitrotAlgorithm, tillOffset int64, sum []byte) *wholeBitrotReader {
|
|
|
|
return &wholeBitrotReader{
|
|
|
|
disk: disk,
|
|
|
|
volume: volume,
|
|
|
|
filePath: filePath,
|
|
|
|
verifier: &BitrotVerifier{algo, sum},
|
|
|
|
tillOffset: tillOffset,
|
|
|
|
buf: nil,
|
|
|
|
}
|
|
|
|
}
|