Files
syft/syft/internal/unionreader/union_reader.go
T
Alex Goodman 04a6fa1b41 fix(unionreader): honor the io.ReaderAt contract in readerAtAdapter (#5186)
`readerAtAdapter.ReadAt` seeks and then issues a single `Read`, which breaks the
`io.ReaderAt` contract in both directions against the squashfs reader it exists
to wrap:

- `squashfs.File.Read` copies against the decompressed block length but advances
  its block cursor by the nominal block size, so a block that decompresses short
  silently stops copying and returns fewer bytes with a nil error. `ReadAt`
  forbids that, and callers rely on it: anything decoding a fixed-size structure
  off the result gets zero padding it has no way to detect and parses it as real
  data. The GraalVM PE export table and the UPX block reader both size a buffer
  from a header field and then ignore `n` entirely, so a crafted image drives
  them straight through the padding.

- a read landing exactly on the end of the file returns a *full* buffer paired
  with `io.EOF`. `bytes.Reader.ReadAt` returns nil there, and the callers that
  treat any error as fatal were written against that, so squashfs-resident
  binaries sized near a read boundary were being skipped outright.

`io.ReadFull` normalizes both: it fills the buffer across short reads, and it
clears the error once the buffer is full. A genuinely short tail is reported as
`io.EOF`, which is what `ReadAt` implementations return at the end of a file, and
what the buffering branch of `GetUnionReader` already returns.

Affects squashfs-backed sources (snaps), so in practice the binary catalogers
reading structure out of executables.

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
2026-08-14 14:09:10 -04:00

149 lines
4.0 KiB
Go

package unionreader
import (
"bytes"
"fmt"
"io"
"sync"
"github.com/diskfs/go-diskfs/filesystem/squashfs"
macho "github.com/anchore/go-macholibre"
"github.com/anchore/syft/internal/log"
"github.com/anchore/syft/syft/file"
)
// UnionReader is a single interface with all reading functions needed by multi-arch binary catalogers
// cataloger.
type UnionReader interface {
io.Reader
io.ReaderAt
io.Seeker
io.Closer
}
// GetReaders extracts one or more io.ReaderAt objects representing binaries that can be processed (multiple binaries in the case for multi-architecture binaries).
func GetReaders(f UnionReader) ([]io.ReaderAt, error) {
if macho.IsUniversalMachoBinary(f) {
machoReaders, err := macho.ExtractReaders(f)
if err != nil {
log.Debugf("extracting readers: %v", err)
return nil, err
}
var readers []io.ReaderAt
for _, e := range machoReaders {
readers = append(readers, e.Reader)
}
return readers, nil
}
return []io.ReaderAt{f}, nil
}
func GetUnionReader(readerCloser io.ReadCloser) (UnionReader, error) {
reader, ok := readerCloser.(UnionReader)
if ok {
return reader, nil
}
// file.LocationReadCloser embeds a ReadCloser, which is likely
// to implement UnionReader. Check whether the embedded read closer
// implements UnionReader, and just return that if so.
if r, ok := readerCloser.(file.LocationReadCloser); ok {
return GetUnionReader(r.ReadCloser)
}
if r, ok := readerCloser.(*squashfs.File); ok {
// seeking is implemented, but not io.ReaderAt. Lets wrap it to prevent from degrading performance
// by copying all data.
return newReaderAtAdapter(r), nil
}
b, err := io.ReadAll(readerCloser) //nolint:gocritic // buffering non-seekable to ReaderAt
if err != nil {
return nil, fmt.Errorf("unable to read contents from binary: %w", err)
}
bytesReader := bytes.NewReader(b)
reader = struct {
io.ReadCloser
io.ReaderAt
io.Seeker
}{
ReadCloser: io.NopCloser(bytesReader),
ReaderAt: bytesReader,
Seeker: bytesReader,
}
return reader, nil
}
type readerAtAdapter struct {
io.ReadSeekCloser
mu *sync.Mutex
}
func newReaderAtAdapter(rs io.ReadSeekCloser) UnionReader {
return &readerAtAdapter{
ReadSeekCloser: rs,
mu: &sync.Mutex{},
}
}
func (r *readerAtAdapter) Read(p []byte) (n int, err error) {
r.mu.Lock()
defer r.mu.Unlock()
return r.ReadSeekCloser.Read(p)
}
func (r *readerAtAdapter) Seek(offset int64, whence int) (int64, error) {
r.mu.Lock()
defer r.mu.Unlock()
return r.ReadSeekCloser.Seek(offset, whence)
}
func (r *readerAtAdapter) ReadAt(p []byte, off int64) (n int, err error) {
r.mu.Lock()
defer r.mu.Unlock()
currentPos, err := r.ReadSeekCloser.Seek(0, io.SeekCurrent) // save current pos
if err != nil {
return 0, err
}
_, err = r.ReadSeekCloser.Seek(off, io.SeekStart) // seek to absolute position `off`
if err != nil {
return 0, err
}
// io.ReaderAt requires len(p) bytes or a non-nil error, and the underlying reader honors neither half of
// that: a squashfs block that decompresses short returns fewer bytes with err == nil, and a read landing
// exactly on the end of the file returns a full buffer alongside io.EOF. io.ReadFull normalizes both.
n, err = io.ReadFull(r.ReadSeekCloser, p)
// ReadAt reports a short read at the end of the file as io.EOF. compared by identity on purpose:
// io.ReadFull returns this sentinel bare, and errors.Is would also downgrade a wrapped one from the
// underlying reader, turning genuine stream corruption into a benign end of file
if err == io.ErrUnexpectedEOF {
err = io.EOF
}
// restore the position for the stateful read/seek operations. a read error wins over a restore failure:
// callers compare against io.EOF, and masking it would strand them without the data they did get
if restoreErr := r.restorePosition(currentPos); restoreErr != nil {
if err == nil {
err = restoreErr
}
}
return n, err
}
func (r *readerAtAdapter) restorePosition(pos int64) error {
_, err := r.ReadSeekCloser.Seek(pos, io.SeekStart)
return err
}