feat: report unknowns in sbom (#2998)

Signed-off-by: Keith Zantow <kzantow@gmail.com>
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
Co-authored-by: Alex Goodman <wagoodman@users.noreply.github.com>
This commit is contained in:
Keith Zantow
2024-10-07 16:11:37 -04:00
committed by GitHub
co-authored by Alex Goodman
parent 4d7ed9f749
commit ccbee94b87
145 changed files with 4420 additions and 233 deletions
+31 -3
View File
@@ -11,8 +11,10 @@ import (
"github.com/anchore/syft/internal/log"
"github.com/anchore/syft/internal/sbomsync"
"github.com/anchore/syft/internal/unknown"
"github.com/anchore/syft/syft/event/monitor"
"github.com/anchore/syft/syft/file"
"github.com/anchore/syft/syft/sbom"
)
type Executor struct {
@@ -35,6 +37,12 @@ func NewTaskExecutor(tasks []Task, numWorkers int) *Executor {
}
func (p *Executor) Execute(ctx context.Context, resolver file.Resolver, s sbomsync.Builder, prog *monitor.CatalogerTaskProgress) error {
var lock sync.Mutex
withLock := func(fn func()) {
lock.Lock()
defer lock.Unlock()
fn()
}
var errs error
wg := &sync.WaitGroup{}
for i := 0; i < p.numWorkers; i++ {
@@ -48,9 +56,16 @@ func (p *Executor) Execute(ctx context.Context, resolver file.Resolver, s sbomsy
return
}
if err := runTaskSafely(ctx, tsk, resolver, s); err != nil {
errs = multierror.Append(errs, fmt.Errorf("failed to run task: %w", err))
prog.SetError(err)
err := runTaskSafely(ctx, tsk, resolver, s)
unknowns, err := unknown.ExtractCoordinateErrors(err)
if len(unknowns) > 0 {
appendUnknowns(s, tsk.Name(), unknowns)
}
if err != nil {
withLock(func() {
errs = multierror.Append(errs, fmt.Errorf("failed to run task: %w", err))
prog.SetError(err)
})
}
prog.Increment()
}
@@ -62,6 +77,19 @@ func (p *Executor) Execute(ctx context.Context, resolver file.Resolver, s sbomsy
return errs
}
func appendUnknowns(builder sbomsync.Builder, taskName string, unknowns []unknown.CoordinateError) {
if accessor, ok := builder.(sbomsync.Accessor); ok {
accessor.WriteToSBOM(func(sb *sbom.SBOM) {
for _, u := range unknowns {
if sb.Artifacts.Unknowns == nil {
sb.Artifacts.Unknowns = map[file.Coordinates][]string{}
}
sb.Artifacts.Unknowns[u.Coordinates] = append(sb.Artifacts.Unknowns[u.Coordinates], formatUnknown(u.Reason.Error(), taskName))
}
})
}
}
func runTaskSafely(ctx context.Context, t Task, resolver file.Resolver, s sbomsync.Builder) (err error) {
// handle individual cataloger panics
defer func() {
+4 -17
View File
@@ -3,7 +3,6 @@ package task
import (
"context"
"crypto"
"fmt"
"github.com/anchore/syft/internal/sbomsync"
"github.com/anchore/syft/syft/artifact"
@@ -32,15 +31,12 @@ func NewFileDigestCatalogerTask(selection file.Selection, hashers ...crypto.Hash
}
result, err := digestsCataloger.Catalog(ctx, resolver, coordinates...)
if err != nil {
return fmt.Errorf("unable to catalog file digests: %w", err)
}
accessor.WriteToSBOM(func(sbom *sbom.SBOM) {
sbom.Artifacts.FileDigests = result
})
return nil
return err
}
return NewTask("file-digest-cataloger", fn)
@@ -62,15 +58,12 @@ func NewFileMetadataCatalogerTask(selection file.Selection) Task {
}
result, err := metadataCataloger.Catalog(ctx, resolver, coordinates...)
if err != nil {
return err
}
accessor.WriteToSBOM(func(sbom *sbom.SBOM) {
sbom.Artifacts.FileMetadata = result
})
return nil
return err
}
return NewTask("file-metadata-cataloger", fn)
@@ -87,15 +80,12 @@ func NewFileContentCatalogerTask(cfg filecontent.Config) Task {
accessor := builder.(sbomsync.Accessor)
result, err := cat.Catalog(ctx, resolver)
if err != nil {
return err
}
accessor.WriteToSBOM(func(sbom *sbom.SBOM) {
sbom.Artifacts.FileContents = result
})
return nil
return err
}
return NewTask("file-content-cataloger", fn)
@@ -112,15 +102,12 @@ func NewExecutableCatalogerTask(selection file.Selection, cfg executable.Config)
accessor := builder.(sbomsync.Accessor)
result, err := cat.Catalog(resolver)
if err != nil {
return err
}
accessor.WriteToSBOM(func(sbom *sbom.SBOM) {
sbom.Artifacts.Executables = result
})
return nil
return err
}
return NewTask("file-executable-cataloger", fn)
+1 -4
View File
@@ -103,9 +103,6 @@ func NewPackageTask(cfg CatalogingFactoryConfig, c pkg.Cataloger, tags ...string
t := bus.StartCatalogerTask(info, -1, "")
pkgs, relationships, err := c.Catalog(ctx, resolver)
if err != nil {
return fmt.Errorf("unable to catalog packages with %q: %w", catalogerName, err)
}
log.WithFields("cataloger", catalogerName).Debugf("discovered %d packages", len(pkgs))
@@ -120,7 +117,7 @@ func NewPackageTask(cfg CatalogingFactoryConfig, c pkg.Cataloger, tags ...string
t.SetCompleted()
log.WithFields("name", catalogerName).Trace("package cataloger completed")
return nil
return err
}
tags = append(tags, pkgcataloging.PackageTag)
+114
View File
@@ -0,0 +1,114 @@
package task
import (
"context"
"strings"
"github.com/mholt/archiver/v3"
"github.com/anchore/syft/internal/log"
"github.com/anchore/syft/internal/sbomsync"
"github.com/anchore/syft/syft/cataloging"
"github.com/anchore/syft/syft/file"
"github.com/anchore/syft/syft/pkg"
"github.com/anchore/syft/syft/sbom"
)
const unknownsLabelerTaskName = "unknowns-labeler"
func NewUnknownsLabelerTask(cfg cataloging.UnknownsConfig) Task {
return NewTask(unknownsLabelerTaskName, unknownsLabelerTask{cfg}.processUnknowns)
}
type unknownsLabelerTask struct {
cataloging.UnknownsConfig
}
// processUnknowns removes unknown entries that have valid packages reported for the locations
func (c unknownsLabelerTask) processUnknowns(_ context.Context, resolver file.Resolver, builder sbomsync.Builder) error {
accessor := builder.(sbomsync.Accessor)
accessor.WriteToSBOM(func(s *sbom.SBOM) {
c.finalize(resolver, s)
})
return nil
}
func (c unknownsLabelerTask) finalize(resolver file.Resolver, s *sbom.SBOM) {
hasPackageReference := coordinateReferenceLookup(resolver, s)
if c.RemoveWhenPackagesDefined {
for coords := range s.Artifacts.Unknowns {
if !hasPackageReference(coords) {
continue
}
delete(s.Artifacts.Unknowns, coords)
}
}
if s.Artifacts.Unknowns == nil {
s.Artifacts.Unknowns = map[file.Coordinates][]string{}
}
if c.IncludeExecutablesWithoutPackages {
for coords := range s.Artifacts.Executables {
if !hasPackageReference(coords) {
s.Artifacts.Unknowns[coords] = append(s.Artifacts.Unknowns[coords], formatUnknown("no package identified in executable file", unknownsLabelerTaskName))
}
}
}
if c.IncludeUnexpandedArchives {
for coords := range s.Artifacts.FileMetadata {
unarchiver, notArchiveErr := archiver.ByExtension(coords.RealPath)
if unarchiver != nil && notArchiveErr == nil && !hasPackageReference(coords) {
s.Artifacts.Unknowns[coords] = append(s.Artifacts.Unknowns[coords], "archive not cataloged")
}
}
}
}
func formatUnknown(err string, task ...string) string {
return strings.Join(task, "/") + ": " + err
}
func coordinateReferenceLookup(resolver file.Resolver, s *sbom.SBOM) func(coords file.Coordinates) bool {
allPackageCoords := file.NewCoordinateSet()
// include all directly included locations that result in packages
for p := range s.Artifacts.Packages.Enumerate() {
allPackageCoords.Add(p.Locations.CoordinateSet().ToSlice()...)
}
// include owned files, for example specified by package managers.
// relationships for these owned files may be disabled, but we always want to include them
for p := range s.Artifacts.Packages.Enumerate() {
if f, ok := p.Metadata.(pkg.FileOwner); ok {
for _, ownedFilePath := range f.OwnedFiles() {
// resolve these owned files, as they may have symlinks
// but coordinates we will test against are always absolute paths
locations, err := resolver.FilesByPath(ownedFilePath)
if err != nil {
log.Debugf("unable to resolve owned file '%s': %v", ownedFilePath, err)
}
for _, loc := range locations {
allPackageCoords.Add(loc.Coordinates)
}
}
}
}
// include relationships
for _, r := range s.Relationships {
_, fromPkgOk := r.From.(pkg.Package)
fromFile, fromFileOk := r.From.(file.Coordinates)
_, toPkgOk := r.To.(pkg.Package)
toFile, toFileOk := r.To.(file.Coordinates)
if fromPkgOk && toFileOk {
allPackageCoords.Add(toFile)
} else if fromFileOk && toPkgOk {
allPackageCoords.Add(fromFile)
}
}
return allPackageCoords.Contains
}