Harden chunk plan cache storage and reuse

This commit is contained in:
2026-07-18 00:49:50 +00:00
parent 6fc6ce0adb
commit 6d0a19c94c
7 changed files with 228 additions and 5 deletions

View File

@@ -0,0 +1,62 @@
package chunkplan
import (
"go/parser"
"go/token"
"path/filepath"
"runtime"
"strconv"
"strings"
"testing"
)
const repositoryImportPrefix = "gitea.maximumdirect.net/eric/notarius/internal/"
func TestPlanStoreAndSourceImportBoundaries(t *testing.T) {
repositoryRoot := repositoryRoot(t)
for _, tc := range []struct {
name string
directory string
forbidden []string
}{
{name: "source is framework and module independent", directory: "internal/core/source", forbidden: []string{"framework/", "modules/"}},
{name: "plan store is module independent", directory: "internal/framework/chunkplan", forbidden: []string{"modules/"}},
} {
t.Run(tc.name, func(t *testing.T) {
files, err := filepath.Glob(filepath.Join(repositoryRoot, tc.directory, "*.go"))
if err != nil {
t.Fatal(err)
}
for _, filename := range files {
if strings.HasSuffix(filename, "_test.go") {
continue
}
parsed, err := parser.ParseFile(token.NewFileSet(), filename, nil, parser.ImportsOnly)
if err != nil {
t.Fatal(err)
}
for _, item := range parsed.Imports {
path, err := strconv.Unquote(item.Path.Value)
if err != nil {
t.Fatal(err)
}
path = strings.TrimPrefix(path, repositoryImportPrefix)
for _, prefix := range tc.forbidden {
if strings.HasPrefix(path, prefix) {
t.Fatalf("%s imports %q, forbidden by %s boundary", filepath.Base(filename), path, tc.name)
}
}
}
}
})
}
}
func repositoryRoot(t *testing.T) string {
t.Helper()
_, filename, _, ok := runtime.Caller(0)
if !ok {
t.Fatal("resolve test location")
}
return filepath.Clean(filepath.Join(filepath.Dir(filename), "..", "..", ".."))
}

View File

@@ -17,7 +17,8 @@ import (
const SchemaVersion = pipeline.ChunkPlanSchemaVersion
type filesystemStore struct {
root string
root string
write func(string, []byte) error
}
func NewFilesystemStore(root string) (pipeline.ChunkPlanStore, error) {
@@ -28,7 +29,7 @@ func NewFilesystemStore(root string) (pipeline.ChunkPlanStore, error) {
if strings.ContainsRune(root, '\x00') {
return nil, fmt.Errorf("chunk plan root must not contain NUL")
}
return &filesystemStore{root: filepath.Clean(root)}, nil
return &filesystemStore{root: filepath.Clean(root), write: writeAtomic}, nil
}
func (s *filesystemStore) Load(sourceDigest string) (pipeline.ChunkPlanRecord, pipeline.ChunkPlanDecision, error) {
@@ -77,7 +78,11 @@ func (s *filesystemStore) Save(record pipeline.ChunkPlanRecord) error {
return fmt.Errorf("encode chunk plan record: %w", err)
}
data = append(data, '\n')
if err := writeAtomic(target, data); err != nil {
writer := s.write
if writer == nil {
writer = writeAtomic
}
if err := writer(target, data); err != nil {
return fmt.Errorf("write chunk plan: %w", err)
}
return nil
@@ -164,7 +169,16 @@ func invalidDecision(reason string) (pipeline.ChunkPlanRecord, pipeline.ChunkPla
return pipeline.ChunkPlanRecord{}, pipeline.ChunkPlanDecision{Status: pipeline.ChunkPlanInvalid, Reason: reason}, nil
}
type atomicWriteHooks struct {
BeforeCreateTemp func() error
BeforeRename func() error
}
func writeAtomic(target string, data []byte) error {
return writeAtomicWithHooks(target, data, atomicWriteHooks{})
}
func writeAtomicWithHooks(target string, data []byte, hooks atomicWriteHooks) error {
dir := filepath.Dir(target)
if err := os.MkdirAll(dir, 0o700); err != nil {
return err
@@ -173,6 +187,11 @@ func writeAtomic(target string, data []byte) error {
return err
}
if hooks.BeforeCreateTemp != nil {
if err := hooks.BeforeCreateTemp(); err != nil {
return err
}
}
temp, err := os.CreateTemp(dir, ".plan.json.tmp-*")
if err != nil {
return err
@@ -199,6 +218,11 @@ func writeAtomic(target string, data []byte) error {
if err := temp.Close(); err != nil {
return err
}
if hooks.BeforeRename != nil {
if err := hooks.BeforeRename(); err != nil {
return err
}
}
if err := os.Rename(tempPath, target); err != nil {
return err
}

View File

@@ -3,6 +3,8 @@ package chunkplan
import (
"bytes"
"encoding/json"
"errors"
"fmt"
"os"
"path/filepath"
"reflect"
@@ -125,6 +127,7 @@ func TestFilesystemStoreReportsInvalidRecordsAsRecoverable(t *testing.T) {
{name: "unknown field", mutate: func(data []byte) []byte {
return bytes.Replace(data, []byte(`{"schema_version"`), []byte(`{"unknown":true,"schema_version"`), 1)
}, want: "unknown"},
{name: "truncated JSON", mutate: func(data []byte) []byte { return data[:len(data)/2] }, want: "decode"},
{name: "schema mismatch", mutate: replaceJSON(`notarius.chunk-plan.v1`, `notarius.chunk-plan.v2`), want: "schema_version"},
{name: "source mismatch", mutate: replaceJSON(testSourceDigest, "sha256:"+strings.Repeat("b", 64)), want: "source_digest"},
{name: "plan digest mismatch", mutate: func(data []byte) []byte {
@@ -138,6 +141,9 @@ func TestFilesystemStoreReportsInvalidRecordsAsRecoverable(t *testing.T) {
{name: "noncanonical annotation", mutate: func(data []byte) []byte {
return bytes.Replace(data, []byte(`{"value":1}`), []byte(`{ "value": 1 }`), 1)
}, want: "canonical JSON"},
{name: "bad boundary", mutate: func(data []byte) []byte {
return bytes.Replace(data, []byte(`"start_unit_id":1`), []byte(`"start_unit_id":0`), 1)
}, want: "boundaries must be positive"},
{name: "trailing JSON", mutate: func(data []byte) []byte { return append(data, []byte(` {}`)...) }, want: "trailing"},
}
for _, tc := range tests {
@@ -226,6 +232,77 @@ func TestFilesystemStoreConcurrentWritersExposeCompleteRecord(t *testing.T) {
}
}
func TestFilesystemStoreReadersObserveOnlyCompleteRecordsDuringWrites(t *testing.T) {
store := newStore(t, t.TempDir())
if err := store.Save(testRecord(t, 1)); err != nil {
t.Fatal(err)
}
const writers = 12
const readers = 12
errs := make(chan error, writers+readers)
start := make(chan struct{})
var writersDone sync.WaitGroup
for i := 0; i < writers; i++ {
writersDone.Add(1)
go func(value int) {
defer writersDone.Done()
<-start
errs <- store.Save(testRecord(t, value+2))
}(i)
}
for i := 0; i < readers; i++ {
go func() {
<-start
for attempt := 0; attempt < 50; attempt++ {
record, decision, err := store.Load(testSourceDigest)
if err != nil || decision.Status != pipeline.ChunkPlanHit {
errs <- fmt.Errorf("Load() decision=%#v error=%v", decision, err)
return
}
if err := validateRecord(record, testSourceDigest); err != nil {
errs <- fmt.Errorf("reader observed invalid record: %w", err)
return
}
}
errs <- nil
}()
}
close(start)
writersDone.Wait()
for i := 0; i < readers+writers; i++ {
if err := <-errs; err != nil {
t.Fatal(err)
}
}
}
func TestFilesystemStoreInterruptedWritesPreservePreviousRecord(t *testing.T) {
store := newStore(t, t.TempDir()).(*filesystemStore)
first := testRecord(t, 1)
if err := store.Save(first); err != nil {
t.Fatal(err)
}
for _, tc := range []struct {
name string
hooks atomicWriteHooks
}{
{name: "before temporary file", hooks: atomicWriteHooks{BeforeCreateTemp: func() error { return errors.New("interrupted before temporary file") }}},
{name: "before rename", hooks: atomicWriteHooks{BeforeRename: func() error { return errors.New("interrupted before rename") }}},
} {
t.Run(tc.name, func(t *testing.T) {
store.write = func(target string, data []byte) error { return writeAtomicWithHooks(target, data, tc.hooks) }
if err := store.Save(testRecord(t, 2)); err == nil {
t.Fatal("Save() error = nil")
}
store.write = writeAtomic
got, decision, err := store.Load(testSourceDigest)
if err != nil || decision.Status != pipeline.ChunkPlanHit || !reflect.DeepEqual(got, first) {
t.Fatalf("record after interruption=%#v decision=%#v error=%v", got, decision, err)
}
})
}
}
func newStore(t *testing.T, root string) pipeline.ChunkPlanStore {
t.Helper()
store, err := NewFilesystemStore(root)