s3: return IncompleteBody instead of 500 for truncated PUT bodies (#10186)

* s3: add IncompleteBody error code

* operation: tag a truncated source read as ErrTruncatedBody

The chunked uploader wraps both source-read failures and volume-upload failures,
and a mid-write volume server drop also carries io.ErrUnexpectedEOF. Tag only the
source read so callers can tell a truncated input apart from a server-side fault.

* s3: return IncompleteBody for a truncated PUT body

A client abort or reverse-proxy timeout truncates the request body mid-upload.
putToFiler mapped every streaming-upload failure to InternalError (500), which a
reverse proxy relays as a 502. Classify a source-read truncation as IncompleteBody
(400) so the response matches AWS and passes through. All S3 write paths share
putToFiler, so they all benefit.

weed/operation/upload_chunked.go

@@ -5,6 +5,7 @@ import (
 	"context"
 	"crypto/md5"
 	"encoding/base64"
+	"errors"
 	"fmt"
 	"hash"
 	"io"
@@ -18,6 +19,11 @@ import (
 	util_http "github.com/seaweedfs/seaweedfs/weed/util/http"
 )
 
+// ErrTruncatedBody tags a source read that ended before the expected bytes
+// arrived, so callers can tell a truncated input (client abort, reverse-proxy
+// timeout) apart from a volume-server upload fault.
+var ErrTruncatedBody = errors.New("truncated request body")
+
 // ChunkedUploadResult contains the result of a chunked upload
 type ChunkedUploadResult struct {
 	FileChunks   []*filer_pb.FileChunk
@@ -104,6 +110,9 @@ uploadLoop:
 			// before any data (offset=0,got=0) from mid-stream truncation.
 			// A bare io.ErrUnexpectedEOF is not actionable on its own (see #9149).
 			wrapped := fmt.Errorf("read chunk at offset %d (got %d bytes): %w", chunkOffset, dataSize, err)
+			if errors.Is(err, io.ErrUnexpectedEOF) {
+				wrapped = fmt.Errorf("%w: %w", ErrTruncatedBody, wrapped)
+			}
 			glog.V(2).Infof("UploadReaderInChunks: %v", wrapped)
 			chunkBufferPool.Put(bytesBuffer)
 			<-bytesBufferLimitChan