ctr-remote: add wildcard support for prefetch-list option

Signed-off-by: abushwang <[email protected]>

Author: abushwang

analyzer/recorder/recorder.go

@@ -56,6 +56,8 @@ type ImageRecorder struct {
 	recordW        content.Writer
 	recordWMu      sync.Mutex
 	recorded       map[string]struct{}
+	allPaths       map[string]struct{}
+	allPathsOnce   sync.Once
 }
 
 func NewImageRecorder(ctx context.Context, cs content.Store, img images.Image, platformMC platforms.MatchComparer) (*ImageRecorder, error) {
@@ -173,6 +175,46 @@ func (r *ImageRecorder) Record(name string) error {
 	})
 }
 
+// RecordGlob records all files matching the given glob pattern.
+// Pattern semantics follow path.Match and paths are normalized via cleanEntryName.
+func (r *ImageRecorder) RecordGlob(pattern string, matcher func(string, string) (bool, error)) (int, error) {
+	if pattern == "" {
+		return 0, nil
+	}
+
+	pattern = cleanEntryName(pattern)
+	r.ensureAllPaths()
+
+	var matchedCount int
+	for name := range r.allPaths {
+		matched, err := matcher(pattern, name)
+		if err != nil {
+			return matchedCount, err
+		}
+		if !matched {
+			continue
+		}
+		if err := r.Record(name); err != nil {
+			continue
+		}
+		matchedCount++
+	}
+
+	return matchedCount, nil
+}
+
+func (r *ImageRecorder) ensureAllPaths() {
+	r.allPathsOnce.Do(func() {
+		paths := make(map[string]struct{})
+		for i := range r.index {
+			for name := range r.index[i] {
+				paths[name] = struct{}{}
+			}
+		}
+		r.allPaths = paths
+	})
+}
+
 func (r *ImageRecorder) Commit(ctx context.Context) (digest.Digest, error) {
 	r.recordWMu.Lock()
 	defer r.recordWMu.Unlock()

cmd/ctr-remote/commands/optimize.go

@@ -29,6 +29,7 @@ import (
 	"strings"
 	"time"
 
+	"github.com/bmatcuk/doublestar/v4"
 	containerd "github.com/containerd/containerd/v2/client"
 	"github.com/containerd/containerd/v2/cmd/ctr/commands"
 	"github.com/containerd/containerd/v2/core/content"
@@ -129,7 +130,7 @@ var OptimizeCommand = &cli.Command{
 		},
 		&cli.StringFlag{
 			Name:  "prefetch-list",
-			Usage: "path to a text file containing list of files to prefetch (one file path per line)",
+			Usage: "path to a text file listing files/patterns to prefetch (one per line; supports glob *, ?, [], and **)",
 		},
 	}, samplerFlags...),
 	Action: func(clicontext *cli.Context) error {
@@ -384,6 +385,17 @@ func analyzePrefetchList(
 	defer rc.Close()
 
 	for _, filePath := range prefetchPaths {
+		if strings.ContainsAny(filePath, "*?[") {
+			n, err := rc.RecordGlob(filePath, doublestar.Match)
+			if err != nil {
+				log.G(ctx).WithError(err).Debugf("failed to record by glob %q", filePath)
+				continue
+			}
+			if n == 0 {
+				log.G(ctx).Warnf("glob pattern %q matched no files", filePath)
+			}
+			continue
+		}
 		if err := rc.Record(filePath); err != nil {
 			log.G(ctx).WithError(err).Debugf("failed to record %q", filePath)
 		}

cmd/go.mod

@@ -5,6 +5,7 @@ go 1.24.0
 toolchain go1.24.2
 
 require (
+	github.com/bmatcuk/doublestar/v4 v4.9.1
 	github.com/containerd/containerd/api v1.9.0
 	github.com/containerd/containerd/v2 v2.1.4
 	github.com/containerd/go-cni v1.1.13

cmd/go.sum

@@ -14,6 +14,8 @@ github.com/beorn7/perks v1.0.1 h1:VlbKKnNfV8bJzeqoa4cOKqO6bYr3WgKZxO8Z16+hsOM=
 github.com/beorn7/perks v1.0.1/go.mod h1:G2ZrVWU2WbWT9wwq4/hrbKbnv/1ERSJQ0ibhJ6rlkpw=
 github.com/blang/semver/v4 v4.0.0 h1:1PFHFE6yCCTv8C1TeyNNarDzntLi7wMI5i/pzqYIsAM=
 github.com/blang/semver/v4 v4.0.0/go.mod h1:IbckMUScFkM3pff0VJDNKRiT6TG/YpiHIM2yvyW5YoQ=
+github.com/bmatcuk/doublestar/v4 v4.9.1 h1:X8jg9rRZmJd4yRy7ZeNDRnM+T3ZfHv15JiBJ/avrEXE=
+github.com/bmatcuk/doublestar/v4 v4.9.1/go.mod h1:xBQ8jztBU6kakFMg+8WGxn0c6z1fTSPVIjEY1Wr7jzc=
 github.com/census-instrumentation/opencensus-proto v0.2.1/go.mod h1:f6KPmirojxKA12rnyqOA5BBL4O983OfeGPqjHWSTneU=
 github.com/cespare/xxhash/v2 v2.3.0 h1:UL815xU9SqsFlibzuggzjXhog7bL6oX9BbNZnL2UFvs=
 github.com/cespare/xxhash/v2 v2.3.0/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=

docs/ctr-remote.md

@@ -132,6 +132,28 @@ Other options are also available for configuring the workload.
 |`-t`or`--terminal`|Attach terminal to the container. This flag must be specified with `-i`|
 |`-i`|Attach stdin to the container|
 
+### Prefetch list glob patterns
+
+`ctr-remote image optimize` supports specifying a prefetch list file via `--prefetch-list`. Each line can be a path or a glob pattern. Supported patterns:
+
+- `*` matches any sequence except `/` (single directory level)
+- `?` matches a single character except `/`
+- `[]` character class, e.g. `[ab]`, `[a-z]`
+- `**` recursive wildcard that can match across `/` (multiple directory levels)
+
+Examples:
+
+```
+/usr/bin/*
+/usr/lib/**/*.so
+/var/log/**/app.*.log
+/opt/data/img_[ab].png
+```
+
+Notes:
+- Empty lines and lines starting with `#` are ignored.
+- Paths are normalized internally; both `/foo/bar` and `foo/bar` are accepted.
+
 ## Mounting files from the host
 
 There are several cases where sharing files from host to the container during optimization is useful.