handle the locks properly for multi-pool callers (#20495)

- PutObjectMetadata()
- PutObjectTags()
- DeleteObjectTags()
- TransitionObject()
- RestoreTransitionObject()

Also improve the behavior of multipart code across
pool locks, hold locks only once per upload ID for

- CompleteMultipartUpload()
- AbortMultipartUpload()
- ListObjectParts() (read-lock)
- GetMultipartInfo() (read-lock)
- PutObjectPart() (read-lock)

This avoids lock attempts across pools for no
reason, this increases O(n) when there are n-pools.

cmd/namespace-lock.go

@@ -229,6 +229,7 @@ type localLockInstance struct {
 // path. The returned lockInstance object encapsulates the nsLockMap,
 // volume, path and operation ID.
 func (n *nsLockMap) NewNSLock(lockers func() ([]dsync.NetLocker, string), volume string, paths ...string) RWLocker {
+	sort.Strings(paths)
 	opsID := mustGetUUID()
 	if n.isDistErasure {
 		drwmutex := dsync.NewDRWMutex(&dsync.Dsync{
@@ -237,7 +238,6 @@ func (n *nsLockMap) NewNSLock(lockers func() ([]dsync.NetLocker, string), volume
 		}, pathsJoinPrefix(volume, paths...)...)
 		return &distLockInstance{drwmutex, opsID}
 	}
-	sort.Strings(paths)
 	return &localLockInstance{n, volume, paths, opsID}
 }