Files
versitygw/backend/common.go
niksis02 7a098b925f feat: implement conditional writes
Closes #821

**Implements conditional operations across object APIs:**

* **PutObject** and **CompleteMultipartUpload**:
  Supports conditional writes with `If-Match` and `If-None-Match` headers (ETag comparisons).
  Evaluation is based on an existing object with the same key in the bucket. The operation is allowed only if the preconditions are satisfied. If no object exists for the key, these headers are ignored.

* **CopyObject** and **UploadPartCopy**:
  Adds conditional reads on the copy source object with the following headers:

  * `x-amz-copy-source-if-match`
  * `x-amz-copy-source-if-none-match`
  * `x-amz-copy-source-if-modified-since`
  * `x-amz-copy-source-if-unmodified-since`
    The first two are ETag comparisons, while the latter two compare against the copy source’s `LastModified` timestamp.

* **AbortMultipartUpload**:
  Supports the `x-amz-if-match-initiated-time` header, which is true only if the multipart upload’s initialization time matches.

* **DeleteObject**:
  Adds support for:

  * `If-Match` (ETag comparison)
  * `x-amz-if-match-last-modified-time` (LastModified comparison)
  * `x-amz-if-match-size` (object size comparison)

Additionally, this PR updates precondition date parsing logic to support both **RFC1123** and **RFC3339** formats. Dates set in the future are ignored, matching AWS S3 behavior.
2025-09-09 01:55:38 +04:00

573 lines
16 KiB
Go

// Copyright 2023 Versity Software
// This file is licensed under the Apache License, Version 2.0
// (the "License"); you may not use this file except in compliance
// with the License. You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing,
// software distributed under the License is distributed on an
// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
// KIND, either express or implied. See the License for the
// specific language governing permissions and limitations
// under the License.
package backend
import (
"crypto/md5"
"encoding/hex"
"errors"
"fmt"
"hash"
"io"
"io/fs"
"math"
"net/url"
"os"
"regexp"
"strconv"
"strings"
"syscall"
"time"
"github.com/aws/aws-sdk-go-v2/service/s3/types"
"github.com/versity/versitygw/s3err"
"github.com/versity/versitygw/s3response"
)
const (
// this is the media type for directories in AWS and Nextcloud
DirContentType = "application/x-directory"
DefaultContentType = "binary/octet-stream"
// this is the minimum allowed size for mp parts
MinPartSize = 5 * 1024 * 1024
)
func IsValidBucketName(name string) bool { return true }
type ByBucketName []s3response.ListAllMyBucketsEntry
func (d ByBucketName) Len() int { return len(d) }
func (d ByBucketName) Swap(i, j int) { d[i], d[j] = d[j], d[i] }
func (d ByBucketName) Less(i, j int) bool { return d[i].Name < d[j].Name }
type ByObjectName []types.Object
func (d ByObjectName) Len() int { return len(d) }
func (d ByObjectName) Swap(i, j int) { d[i], d[j] = d[j], d[i] }
func (d ByObjectName) Less(i, j int) bool { return *d[i].Key < *d[j].Key }
func GetPtrFromString(str string) *string {
if str == "" {
return nil
}
return &str
}
func GetStringFromPtr(str *string) string {
if str == nil {
return ""
}
return *str
}
func GetTimePtr(t time.Time) *time.Time {
return &t
}
func TrimEtag(etag *string) *string {
if etag == nil {
return nil
}
return GetPtrFromString(strings.Trim(*etag, "\""))
}
var (
errInvalidRange = s3err.GetAPIError(s3err.ErrInvalidRange)
errInvalidCopySourceRange = s3err.GetAPIError(s3err.ErrInvalidCopySourceRange)
errPreconditionFailed = s3err.GetAPIError(s3err.ErrPreconditionFailed)
errNotModified = s3err.GetAPIError(s3err.ErrNotModified)
)
// ParseObjectRange parses input range header and returns startoffset, length, isValid
// and error. If no endoffset specified, then length is set to the object size
// for invalid inputs, it returns no error, but isValid=false
// `InvalidRange` error is returnd, only if startoffset is greater than the object size
func ParseObjectRange(size int64, acceptRange string) (int64, int64, bool, error) {
// Return full object (invalid range, no error) if header empty
if acceptRange == "" {
return 0, size, false, nil
}
rangeKv := strings.Split(acceptRange, "=")
if len(rangeKv) != 2 {
return 0, size, false, nil
}
if rangeKv[0] != "bytes" { // unsupported unit -> ignore
return 0, size, false, nil
}
bRange := strings.Split(rangeKv[1], "-")
if len(bRange) != 2 { // malformed / multi-range
return 0, size, false, nil
}
// Parse start; empty start indicates a suffix-byte-range-spec (e.g. bytes=-100)
startOffset, err := strconv.ParseInt(bRange[0], 10, strconv.IntSize)
if startOffset > int64(math.MaxInt) || startOffset < int64(math.MinInt) {
return 0, size, false, errInvalidRange
}
if err != nil && bRange[0] != "" { // invalid numeric start (non-empty) -> ignore range
return 0, size, false, nil
}
// If end part missing (e.g. bytes=100-)
if bRange[1] == "" {
if bRange[0] == "" { // bytes=- (meaningless) -> ignore
return 0, size, false, nil
}
// start beyond or at size is unsatisfiable -> error (RequestedRangeNotSatisfiable)
if startOffset >= size {
return 0, 0, false, errInvalidRange
}
// bytes=100- => from start to end
return startOffset, size - startOffset, true, nil
}
endOffset, err := strconv.ParseInt(bRange[1], 10, strconv.IntSize)
if endOffset > int64(math.MaxInt) {
return 0, size, false, errInvalidRange
}
if err != nil { // invalid numeric end -> ignore range
return 0, size, false, nil
}
// Suffix range handling (bRange[0] == "")
if bRange[0] == "" {
// Disallow -0 (always unsatisfiable)
if endOffset == 0 {
return 0, 0, false, errInvalidRange
}
// For zero-sized objects any positive suffix is treated as invalid (ignored, no error)
if size == 0 {
return 0, size, false, nil
}
// Clamp to object size (request more bytes than exist -> entire object)
endOffset = min(endOffset, size)
return size - endOffset, endOffset, true, nil
}
// Normal range (start-end)
if startOffset > endOffset { // start > end -> ignore
return 0, size, false, nil
}
// Start beyond or at end of object -> error
if startOffset >= size {
return 0, 0, false, errInvalidRange
}
// Adjust end beyond object size (trim)
if endOffset >= size {
endOffset = size - 1
}
return startOffset, endOffset - startOffset + 1, true, nil
}
// ParseCopySourceRange parses input range header and returns startoffset, length
// and error. If no endoffset specified, then length is set to the object size
func ParseCopySourceRange(size int64, acceptRange string) (int64, int64, error) {
if acceptRange == "" {
return 0, size, nil
}
rangeKv := strings.Split(acceptRange, "=")
if len(rangeKv) != 2 {
return 0, 0, errInvalidCopySourceRange
}
if rangeKv[0] != "bytes" {
return 0, 0, errInvalidCopySourceRange
}
bRange := strings.Split(rangeKv[1], "-")
if len(bRange) != 2 {
return 0, 0, errInvalidCopySourceRange
}
startOffset, err := strconv.ParseInt(bRange[0], 10, 64)
if err != nil {
return 0, 0, errInvalidCopySourceRange
}
if startOffset >= size {
return 0, 0, s3err.CreateExceedingRangeErr(size)
}
if bRange[1] == "" {
return startOffset, size - startOffset + 1, nil
}
endOffset, err := strconv.ParseInt(bRange[1], 10, 64)
if err != nil {
return 0, 0, errInvalidCopySourceRange
}
if endOffset < startOffset {
return 0, 0, errInvalidCopySourceRange
}
if endOffset >= size {
return 0, 0, s3err.CreateExceedingRangeErr(size)
}
return startOffset, endOffset - startOffset + 1, nil
}
// ParseCopySource parses x-amz-copy-source header and returns source bucket,
// source object, versionId, error respectively
func ParseCopySource(copySourceHeader string) (string, string, string, error) {
if copySourceHeader[0] == '/' {
copySourceHeader = copySourceHeader[1:]
}
var copySource, versionId string
i := strings.LastIndex(copySourceHeader, "?versionId=")
if i == -1 {
copySource = copySourceHeader
} else {
copySource = copySourceHeader[:i]
versionId = copySourceHeader[i+11:]
}
srcBucket, srcObject, ok := strings.Cut(copySource, "/")
if !ok {
return "", "", "", s3err.GetAPIError(s3err.ErrInvalidCopySourceBucket)
}
return srcBucket, srcObject, versionId, nil
}
// ParseObjectTags parses the url encoded input string into
// map[string]string with unescaped key/value pair
func ParseObjectTags(tagging string) (map[string]string, error) {
if tagging == "" {
return nil, nil
}
tagSet := make(map[string]string)
for tagging != "" {
var tag string
tag, tagging, _ = strings.Cut(tagging, "&")
// if 'tag' before the first appearance of '&' is empty continue
if tag == "" {
continue
}
key, value, found := strings.Cut(tag, "=")
// if key is empty, but "=" is present, return invalid url ecnoding err
if found && key == "" {
return nil, s3err.GetAPIError(s3err.ErrInvalidURLEncodedTagging)
}
// return invalid tag key, if the key is longer than 128
if len(key) > 128 {
return nil, s3err.GetAPIError(s3err.ErrInvalidTagKey)
}
// return invalid tag value, if tag value is longer than 256
if len(value) > 256 {
return nil, s3err.GetAPIError(s3err.ErrInvalidTagValue)
}
// query unescape tag key
key, err := url.QueryUnescape(key)
if err != nil {
return nil, s3err.GetAPIError(s3err.ErrInvalidURLEncodedTagging)
}
// query unescape tag value
value, err = url.QueryUnescape(value)
if err != nil {
return nil, s3err.GetAPIError(s3err.ErrInvalidURLEncodedTagging)
}
// check tag key to be valid
if !isValidTagComponent(key) {
return nil, s3err.GetAPIError(s3err.ErrInvalidTagKey)
}
// check tag value to be valid
if !isValidTagComponent(value) {
return nil, s3err.GetAPIError(s3err.ErrInvalidTagValue)
}
// duplicate keys are not allowed: return invalid url encoding err
_, ok := tagSet[key]
if ok {
return nil, s3err.GetAPIError(s3err.ErrInvalidURLEncodedTagging)
}
tagSet[key] = value
}
return tagSet, nil
}
var validTagComponent = regexp.MustCompile(`^[a-zA-Z0-9:/_.\-+ ]+$`)
// isValidTagComponent matches strings which contain letters, decimal digits,
// and special chars: '/', '_', '-', '+', '.', ' ' (space)
func isValidTagComponent(str string) bool {
if str == "" {
return true
}
return validTagComponent.Match([]byte(str))
}
func GetMultipartMD5(parts []types.CompletedPart) string {
var partsEtagBytes []byte
for _, part := range parts {
partsEtagBytes = append(partsEtagBytes, getEtagBytes(*part.ETag)...)
}
return fmt.Sprintf("\"%s-%d\"", md5String(partsEtagBytes), len(parts))
}
func getEtagBytes(etag string) []byte {
decode, err := hex.DecodeString(strings.ReplaceAll(etag, string('"'), ""))
if err != nil {
return []byte(etag)
}
return decode
}
func md5String(data []byte) string {
sum := md5.Sum(data)
return hex.EncodeToString(sum[:])
}
type FileSectionReadCloser struct {
R io.Reader
F *os.File
}
func (f *FileSectionReadCloser) Read(p []byte) (int, error) {
return f.R.Read(p)
}
func (f *FileSectionReadCloser) Close() error {
return f.F.Close()
}
// MoveFile moves a file from source to destination.
func MoveFile(source, destination string, perm os.FileMode) error {
// We use Rename as the atomic operation for object puts. The upload is
// written to a temp file to not conflict with any other simultaneous
// uploads. The final operation is to move the temp file into place for
// the object. This ensures the object semantics of last upload completed
// wins and is not some combination of writes from simultaneous uploads.
err := os.Rename(source, destination)
if err == nil || !errors.Is(err, syscall.EXDEV) {
return err
}
// Rename can fail if the source and destination are not on the same
// filesystem. The fallback is to copy the file and then remove the source.
// We need to be careful that the desination does not exist before copying
// to prevent any other simultaneous writes to the file.
sourceFile, err := os.Open(source)
if err != nil {
return fmt.Errorf("open source: %w", err)
}
defer sourceFile.Close()
var destFile *os.File
for {
destFile, err = os.OpenFile(destination, os.O_CREATE|os.O_EXCL|os.O_WRONLY, perm)
if err != nil {
if errors.Is(err, fs.ErrExist) {
if removeErr := os.Remove(destination); removeErr != nil {
return fmt.Errorf("remove existing destination: %w", removeErr)
}
continue
}
return fmt.Errorf("create destination: %w", err)
}
break
}
defer destFile.Close()
_, err = io.Copy(destFile, sourceFile)
if err != nil {
return fmt.Errorf("copy data: %w", err)
}
err = os.Remove(source)
if err != nil {
return fmt.Errorf("remove source: %w", err)
}
return nil
}
// GenerateEtag generates a new quoted etag from the provided hash.Hash
func GenerateEtag(h hash.Hash) string {
dataSum := h.Sum(nil)
return fmt.Sprintf("\"%s\"", hex.EncodeToString(dataSum[:]))
}
// AreEtagsSame compares 2 etags by ignoring quotes
func AreEtagsSame(e1, e2 string) bool {
return strings.Trim(e1, `"`) == strings.Trim(e2, `"`)
}
func getBoolPtr(b bool) *bool {
return &b
}
type PreConditions struct {
IfMatch *string
IfNoneMatch *string
IfModSince *time.Time
IfUnmodeSince *time.Time
}
// EvaluatePreconditions takes the object ETag, the last modified time and
// evaluates the read preconditions:
// - if-match,
// - if-none-match
// - if-modified-since
// - if-unmodified-since
// if-match and if-none-match are ETag comparisions
// if-modified-since and if-unmodified-since are last modifed time comparisons
func EvaluatePreconditions(etag string, modTime time.Time, preconditions PreConditions) error {
if preconditions.IfMatch == nil && preconditions.IfNoneMatch == nil && preconditions.IfModSince == nil && preconditions.IfUnmodeSince == nil {
return nil
}
// convert all conditions to *bool to evaluate the conditions
var ifMatch, ifNoneMatch, ifModSince, ifUnmodeSince *bool
if preconditions.IfMatch != nil {
ifMatch = getBoolPtr(*preconditions.IfMatch == etag)
}
if preconditions.IfNoneMatch != nil {
ifNoneMatch = getBoolPtr(*preconditions.IfNoneMatch != etag)
}
if preconditions.IfModSince != nil {
ifModSince = getBoolPtr(preconditions.IfModSince.UTC().Before(modTime.UTC()))
}
if preconditions.IfUnmodeSince != nil {
ifUnmodeSince = getBoolPtr(preconditions.IfUnmodeSince.UTC().After(modTime.UTC()))
}
if ifMatch != nil {
// if `if-match` doesn't matches, return PreconditionFailed
if !*ifMatch {
return errPreconditionFailed
}
// if-match matches
if *ifMatch {
if ifNoneMatch != nil {
// if `if-none-match` doesn't match return NotModified
if !*ifNoneMatch {
return errNotModified
}
// if both `if-match` and `if-none-match` match, return no error
return nil
}
// if `if-match` matches but `if-modified-since` is false return NotModified
if ifModSince != nil && !*ifModSince {
return errNotModified
}
// ignore `if-unmodified-since` as `if-match` is true
return nil
}
}
if ifNoneMatch != nil {
if *ifNoneMatch {
// if `if-none-match` is true, but `if-unmodified-since` is false
// return PreconditionFailed
if ifUnmodeSince != nil && !*ifUnmodeSince {
return errPreconditionFailed
}
// ignore `if-modified-since` as `if-none-match` is true
return nil
} else {
// if `if-none-match` is false and `if-unmodified-since` is false
// return PreconditionFailed
if ifUnmodeSince != nil && !*ifUnmodeSince {
return errPreconditionFailed
}
// in all other cases when `if-none-match` is false return NotModified
return errNotModified
}
}
if ifModSince != nil && !*ifModSince {
// if both `if-modified-since` and `if-unmodified-since` are false
// return PreconditionFailed
if ifUnmodeSince != nil && !*ifUnmodeSince {
return errPreconditionFailed
}
// if only `if-modified-since` is false, return NotModified
return errNotModified
}
// if `if-unmodified-since` is false return PreconditionFailed
if ifUnmodeSince != nil && !*ifUnmodeSince {
return errPreconditionFailed
}
return nil
}
// EvaluateMatchPreconditions evaluates if-match and if-none-match preconditions
func EvaluateMatchPreconditions(etag string, ifMatch, ifNoneMatch *string) error {
if ifMatch != nil && *ifMatch != etag {
return errPreconditionFailed
}
if ifNoneMatch != nil && *ifNoneMatch == etag {
return errPreconditionFailed
}
return nil
}
type ObjectDeletePreconditions struct {
IfMatch *string
IfMatchLastModTime *time.Time
IfMatchSize *int64
}
// EvaluateObjectDeletePreconditions evaluates preconditions for DeleteObject
func EvaluateObjectDeletePreconditions(etag string, modTime time.Time, size int64, preconditions ObjectDeletePreconditions) error {
ifMatch := preconditions.IfMatch
if ifMatch != nil && *ifMatch != etag {
return errPreconditionFailed
}
ifMatchTime := preconditions.IfMatchLastModTime
if ifMatchTime != nil && ifMatchTime.Unix() != modTime.Unix() {
return errPreconditionFailed
}
ifMatchSize := preconditions.IfMatchSize
if ifMatchSize != nil && *ifMatchSize != size {
return errPreconditionFailed
}
return nil
}