Files
syft/syft/pkg/cataloger/binary/elf_package_cataloger.go
T
Alex Goodman 58a033f924 Prevent unnecessary allocations when parsing compressed ELF sections (#5187)
* fix(elf): bound compressed ELF section reads

`debug/elf` takes a section's decompressed size from that section's own
compression header, and a highly compressible stream really does deliver the
bytes that header promises, so `internal/saferio` does not help: it faithfully
allocates every one of them. A 2MB input file drives `elf.NewFile` to allocate
over 10GB and return no error, which is a fatal OOM rather than a recoverable
panic.

Which sections get read is not up to the caller. `elf.NewFile` always reads the
section-name string table, and `File.Symbols` reads `.symtab` plus whatever
section its `Link` field points at, so being selective about sections is not
enough to avoid it.

New `elfutil.NewFile` is a drop-in for `elf.NewFile` that rejects a declared
decompressed size over 128MB. Every production call site goes through it, and a
ruleguard rule keeps the next one from going direct.

The check runs in two parts, since `debug/elf` expands sections at two different
times. The section-name string table is the only one `elf.NewFile` expands
itself, so it is checked against the raw bytes before the call; everything else
is expanded lazily by `(*Section).Open` and is checked after the parse, where
names, types and decompressed sizes are already resolved.

Only the sections syft can actually reach are bounded, which keeps the guard
from costing real binaries. DWARF is excluded since nothing calls `File.DWARF`,
so a large compressed `.debug_info` no longer skips the whole file, and sections
`debug/elf` will not decompress anyway (`SHF_ALLOC`, `SHT_NOBITS`) are left
alone. The legacy `.zdebug` form is matched on the section name the way
`debug/elf` gates it rather than on the `ZLIB` magic, so an ordinary section
starting with those four bytes is not mistaken for a compressed one.

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* fix(elf): gate debug/buildinfo behind the compressed-section check

`debug/buildinfo.Read` opens ELF files with `debug/elf` itself, and `elf.NewFile`
expands the section-name string table as it parses, so the golang cataloger was
still reachable by the same bomb `elfutil` exists to stop. A 261KB fixture drove
1.4GB of allocation through `buildinfo.Read` and returned no error.

`elfutil.CheckSectionNameTable` is now exported for that case: callers that cannot
use `NewFile` because the `debug/elf` call is made for them inside another package.
Both `buildinfo.Read` call sites go through it, including the UPX-decompressed one.

Also corrects claims that did not hold up:

- the package doc's 2MB-to-10GB figure is not reachable with zlib (~1000:1), so it
  now carries the measured 510KB-to-2.6GB, and names zstd's 32767:1 since that is
  what makes the small inputs possible

- `.go.buildinfo` was listed as a hot-path section elfutil covers, but it is read
  through `debug/buildinfo` and never touches `Section.Data`

- the graalvm comment claimed routing size rejections away from `*elf.FormatError`
  improved reporting; both branches are skipped by the caller and only the
  FormatError branch logs, so it did the opposite

- `sharedLibraries` logged short and truncated files as real ELF failures, since
  `debug/elf` returns a bare `io.EOF` rather than an `*elf.FormatError` for those

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* added test comments around the negative cases

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* additional tests

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* better decomposition and comments

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

* use a less brittle constant for error detection

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>

---------

Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
2026-08-14 20:08:58 +00:00

219 lines
6.8 KiB
Go

package binary
import (
"bytes"
"context"
"encoding/binary"
"encoding/json"
"fmt"
"github.com/anchore/syft/internal"
"github.com/anchore/syft/internal/log"
"github.com/anchore/syft/internal/mimetype"
"github.com/anchore/syft/internal/unknown"
"github.com/anchore/syft/syft/artifact"
"github.com/anchore/syft/syft/file"
"github.com/anchore/syft/syft/internal/elfutil"
"github.com/anchore/syft/syft/internal/unionreader"
"github.com/anchore/syft/syft/pkg"
)
var _ pkg.Cataloger = (*elfPackageCataloger)(nil)
type elfPackageCataloger struct {
}
// TODO: for now this accounts for a single data shape from the .note.package section of an ELF binary.
// In the future, this should be generalized to support multiple data shapes, including non-json data.
// For example, fedora includes an ELF section header as a prefix to the JSON payload: https://github.com/anchore/syft/issues/2713
type elfBinaryPackageNotes struct {
Name string `json:"name"`
Version string `json:"version"`
PURL string `json:"purl"`
CPE string `json:"cpe"`
License string `json:"license"`
pkg.ELFBinaryPackageNoteJSONPayload `json:",inline"`
// CorrectOSCPE has the corrected casing for the osCPE field relative to the systemd ELF package metadata "spec" https://systemd.io/ELF_PACKAGE_METADATA/ .
// Ideally in syft 2.0 this field should be replaced with the pkg.ELFBinaryPackageNoteJSONPayload.OSCPE field directly (with the struct tag corrected).
CorrectOSCPE string `json:"osCpe,omitempty"`
Location file.Location `json:"-"`
}
type elfPackageKey struct {
Name string
Version string
PURL string
CPE string
}
func NewELFPackageCataloger() pkg.Cataloger {
return &elfPackageCataloger{}
}
func (c *elfPackageCataloger) Name() string {
return "elf-binary-package-cataloger"
}
func (c *elfPackageCataloger) Catalog(ctx context.Context, resolver file.Resolver) ([]pkg.Package, []artifact.Relationship, error) {
var errs error
locations, err := resolver.FilesByMIMEType(mimetype.ExecutableMIMETypeSet.List()...)
if err != nil {
return nil, nil, fmt.Errorf("unable to get binary files by mime type: %w", err)
}
// first find all ELF binaries that have notes
var notesByLocation = make(map[elfPackageKey][]elfBinaryPackageNotes)
for _, location := range locations {
notes, key, err := parseElfPackageNotes(resolver, location, c)
if err != nil {
errs = unknown.Append(errs, location, err)
continue
}
if notes == nil {
continue
}
notesByLocation[key] = append(notesByLocation[key], *notes)
}
// now we have all ELF binaries that have notes, let's create packages for them.
// we do this in a second pass since it is possible that we have multiple ELF binaries with the same name and version
// which means the set of binaries collectively represent a single logical package.
var pkgs []pkg.Package
for _, notes := range notesByLocation {
noteLocations := file.NewLocationSet()
for _, note := range notes {
noteLocations.Add(note.Location.WithAnnotation(pkg.EvidenceAnnotationKey, pkg.PrimaryEvidenceAnnotation))
}
// create a package for each unique name/version pair (based on the first note found)
pkgs = append(pkgs, newELFPackage(ctx, notes[0], noteLocations))
}
// why not return relationships? We have an executable cataloger that will note the dynamic libraries imported by
// each binary. After all files and packages are processed there is a final task that creates package-to-package
// and package-to-file relationships based on the dynamic libraries imported by each binary.
return pkgs, nil, errs
}
func parseElfPackageNotes(resolver file.Resolver, location file.Location, c *elfPackageCataloger) (*elfBinaryPackageNotes, elfPackageKey, error) {
reader, err := resolver.FileContentsByLocation(location)
if err != nil {
return nil, elfPackageKey{}, fmt.Errorf("unable to get binary contents %q: %w", location.Path(), err)
}
defer internal.CloseAndLogError(reader, location.AccessPath)
notes, err := c.parseElfNotes(file.LocationReadCloser{
Location: location,
ReadCloser: reader,
})
if err != nil {
log.WithFields("file", location.Path(), "error", err).Trace("unable to parse ELF notes")
return nil, elfPackageKey{}, err
}
if notes == nil {
return nil, elfPackageKey{}, nil
}
notes.Location = location
key := elfPackageKey{
Name: notes.Name,
Version: notes.Version,
PURL: notes.PURL,
CPE: notes.CPE,
}
return notes, key, nil
}
func (c *elfPackageCataloger) parseElfNotes(reader file.LocationReadCloser) (*elfBinaryPackageNotes, error) {
metadata, err := getELFNotes(reader)
if err != nil {
return nil, fmt.Errorf("unable to process ELF binary: %w", err)
}
if metadata == nil || metadata.Name == "" || metadata.Version == "" {
return nil, nil
}
return metadata, nil
}
func getELFNotes(r file.LocationReadCloser) (*elfBinaryPackageNotes, error) {
unionReader, err := unionreader.GetUnionReader(r)
if err != nil {
return nil, fmt.Errorf("unable to get union reader for binary: %w", err)
}
f, err := elfutil.NewFile(unionReader)
if f == nil || err != nil {
log.WithFields("file", r.Location.Path(), "error", err).Trace("unable to parse binary as ELF")
return nil, nil
}
noteSection := f.Section(".note.package")
if noteSection == nil {
return nil, nil
}
notes, err := noteSection.Data()
if err != nil {
return nil, err
}
if len(notes) == 0 {
return nil, nil
}
{
var metadata *elfBinaryPackageNotes
if metadata, err = unmarshalELFPackageNotesPayload(notes); err == nil {
return metadata, nil
}
}
{
var header elf64SectionHeader
headerSize := binary.Size(header) / 4
if len(notes) > headerSize {
var metadata *elfBinaryPackageNotes
newPayload := bytes.TrimRight(notes[headerSize:], "\x00")
if metadata, err = unmarshalELFPackageNotesPayload(newPayload); err == nil {
return metadata, nil
}
log.WithFields("file", r.Location.Path(), "error", err).Trace("unable to unmarshal ELF package notes as JSON")
}
}
return nil, err
}
func unmarshalELFPackageNotesPayload(data []byte) (*elfBinaryPackageNotes, error) {
var metadata elfBinaryPackageNotes
if err := json.Unmarshal(data, &metadata); err != nil {
return nil, fmt.Errorf("unable to unmarshal ELF package notes payload: %w", err)
}
// normalize the os CPE field
if metadata.OSCPE == "" { //nolint:staticcheck
// ensure the public field is populated for backwards compatibility
metadata.OSCPE = metadata.CorrectOSCPE //nolint:staticcheck
}
return &metadata, nil
}
type elf64SectionHeader struct {
ShName uint32
ShType uint32
ShFlags uint64
ShAddr uint64
ShOffset uint64
ShSize uint64
ShLink uint32
ShInfo uint32
ShAddralign uint64
ShEntsize uint64
}