restic/internal/archiver/blob_saver.go
Michael Eischer bdf7ba20cb archiver: Fix race condition triggered by TestArchiverAbortEarlyOnError
The Save methods of the BlobSaver, FileSaver and TreeSaver return early
on when the archiver is stopped due to an error. For that they select on
both the tomb.Dying() and context.Done() channels, which can lead to a
race condition when the tomb is killed due to an error: The tomb first
closes its Dying channel before canceling all child contexts.
Archiver.SaveDir only aborts its execution once the context was
canceled. When the tomb killing is paused between closing its Dying
channel and canceling the child contexts, this lets the
FileSaver/TreeSaver.Save methods return immediately, however, ScanDir
still reads further files causing the test case to fail.

As a killed tomb always cancels all child contexts and as the Savers
always use a context bound to the tomb, it is sufficient to just use
context.Done() as escape hatch in the Save functions. This fixes the
mismatch between SaveDir and Save.

Adjust the tests to use contexts bound to the tomb for all interactions
with the Savers.
2020-04-13 18:23:17 +02:00

171 lines
3.6 KiB
Go

package archiver
import (
"context"
"sync"
"github.com/restic/restic/internal/debug"
"github.com/restic/restic/internal/restic"
tomb "gopkg.in/tomb.v2"
)
// Saver allows saving a blob.
type Saver interface {
SaveBlob(ctx context.Context, t restic.BlobType, data []byte, id restic.ID) (restic.ID, error)
Index() restic.Index
}
// BlobSaver concurrently saves incoming blobs to the repo.
type BlobSaver struct {
repo Saver
m sync.Mutex
knownBlobs restic.BlobSet
ch chan<- saveBlobJob
}
// NewBlobSaver returns a new blob. A worker pool is started, it is stopped
// when ctx is cancelled.
func NewBlobSaver(ctx context.Context, t *tomb.Tomb, repo Saver, workers uint) *BlobSaver {
ch := make(chan saveBlobJob)
s := &BlobSaver{
repo: repo,
knownBlobs: restic.NewBlobSet(),
ch: ch,
}
for i := uint(0); i < workers; i++ {
t.Go(func() error {
return s.worker(t.Context(ctx), ch)
})
}
return s
}
// Save stores a blob in the repo. It checks the index and the known blobs
// before saving anything. The second return parameter is true if the blob was
// previously unknown.
func (s *BlobSaver) Save(ctx context.Context, t restic.BlobType, buf *Buffer) FutureBlob {
ch := make(chan saveBlobResponse, 1)
select {
case s.ch <- saveBlobJob{BlobType: t, buf: buf, ch: ch}:
case <-ctx.Done():
debug.Log("not sending job, context is cancelled")
close(ch)
return FutureBlob{ch: ch}
}
return FutureBlob{ch: ch, length: len(buf.Data)}
}
// FutureBlob is returned by SaveBlob and will return the data once it has been processed.
type FutureBlob struct {
ch <-chan saveBlobResponse
length int
res saveBlobResponse
}
// Wait blocks until the result is available or the context is cancelled.
func (s *FutureBlob) Wait(ctx context.Context) {
select {
case <-ctx.Done():
return
case res, ok := <-s.ch:
if ok {
s.res = res
}
}
}
// ID returns the ID of the blob after it has been saved.
func (s *FutureBlob) ID() restic.ID {
return s.res.id
}
// Known returns whether or not the blob was already known.
func (s *FutureBlob) Known() bool {
return s.res.known
}
// Length returns the length of the blob.
func (s *FutureBlob) Length() int {
return s.length
}
type saveBlobJob struct {
restic.BlobType
buf *Buffer
ch chan<- saveBlobResponse
}
type saveBlobResponse struct {
id restic.ID
known bool
}
func (s *BlobSaver) saveBlob(ctx context.Context, t restic.BlobType, buf []byte) (saveBlobResponse, error) {
id := restic.Hash(buf)
h := restic.BlobHandle{ID: id, Type: t}
// check if another goroutine has already saved this blob
known := false
s.m.Lock()
if s.knownBlobs.Has(h) {
known = true
} else {
s.knownBlobs.Insert(h)
known = false
}
s.m.Unlock()
// blob is already known, nothing to do
if known {
return saveBlobResponse{
id: id,
known: true,
}, nil
}
// check if the repo knows this blob
if s.repo.Index().Has(id, t) {
return saveBlobResponse{
id: id,
known: true,
}, nil
}
// otherwise we're responsible for saving it
_, err := s.repo.SaveBlob(ctx, t, buf, id)
if err != nil {
return saveBlobResponse{}, err
}
return saveBlobResponse{
id: id,
known: false,
}, nil
}
func (s *BlobSaver) worker(ctx context.Context, jobs <-chan saveBlobJob) error {
for {
var job saveBlobJob
select {
case <-ctx.Done():
return nil
case job = <-jobs:
}
res, err := s.saveBlob(ctx, job.BlobType, job.buf.Data)
if err != nil {
debug.Log("saveBlob returned error, exiting: %v", err)
close(job.ch)
return err
}
job.ch <- res
close(job.ch)
job.buf.Release()
}
}