filer: routed PosixLock RPC over the in-memory authority (#9664)

* filer: in-memory POSIX lock authority (Manager)

Concurrent multi-inode authority over the per-inode Set: a Set per opaque
inode key (path, or hl:<HardLinkId>) plus a session->keys index so a dead
mount's locks reap in O(locks held). Lock state stays in memory like the
distributed lock manager's, off the replicated meta-log. TryLock/Unlock/
GetLk/ReleasePosixOwner/ReleaseFlockOwner/ReleaseSession; empty sets and
stale index entries are pruned on release.

* filer: routed PosixLock RPC over the in-memory authority

Adds the PosixLock RPC (try/unlock/get_lk + the flush/release owner
drops) that the owner filer answers from its in-memory Manager. The
request key is the inode identity ring key; a non-owner filer forwards
one hop (is_moved-bounded), mirroring ObjectTransaction, so the owner's
table stays the single authority under a stale ring view. Strictly
non-blocking; SetLkw polling lives in the mount.

other/java/client/src/main/proto/filer.proto

@@ -37,6 +37,9 @@ service SeaweedFiler {
     rpc ObjectTransactionBatch (ObjectTransactionBatchRequest) returns (ObjectTransactionBatchResponse) {
     }
 
+    rpc PosixLock (PosixLockRequest) returns (PosixLockResponse) {
+    }
+
     rpc AtomicRenameEntry (AtomicRenameEntryRequest) returns (AtomicRenameEntryResponse) {
     }
     rpc StreamRenameEntry (StreamRenameEntryRequest) returns (stream StreamRenameEntryResponse) {
@@ -351,7 +354,7 @@ message ObjectTransactionRequest {
     repeated ObjectMutation mutations = 3;
     bool is_from_other_cluster = 4;
     repeated int32 signatures = 5;
-    string condition_key = 6;
+    string condition_key = 6;          // if set, evaluate the condition against this entry instead of lock_key (still locking lock_key)
     string route_key = 7;              // ring key identifying the owner filer; a non-owner forwards the whole transaction to it
     bool is_moved = 8;                 // set on a forwarded transaction so the receiver applies it locally instead of forwarding again
 }
@@ -361,6 +364,46 @@ message ObjectTransactionResponse {
     FilerError error_code = 2;
 }
 
+// PosixLockRange is one advisory byte-range lock. Owner identity is (sid, owner):
+// sid is the mount session, owner the FUSE lock owner within it, so owners from
+// different mounts never alias. end is inclusive (max uint64 = to EOF); is_flock
+// separates the flock and fcntl namespaces, which never conflict.
+message PosixLockRange {
+    uint64 start = 1;
+    uint64 end = 2;
+    uint32 type = 3;       // 1=read, 2=write, 3=unlock
+    uint64 sid = 4;
+    uint64 owner = 5;
+    uint32 pid = 6;        // holder pid, for get_lk reporting only
+    bool is_flock = 7;
+}
+
+// PosixLock routes an advisory lock operation to the inode's owner filer, which
+// holds the authoritative in-memory lock table. key is the inode identity ring
+// key (the file path, or hl:<HardLinkId> for a hardlink) used both to resolve the
+// owner and to index the table. A non-owner filer forwards the request one hop;
+// is_moved bounds it so a stale ring view cannot loop.
+message PosixLockRequest {
+    string key = 1;
+    bool is_moved = 2;
+    PosixLockOp op = 3;
+    PosixLockRange lock = 4;
+}
+
+enum PosixLockOp {
+    TRY_LOCK = 0;             // grant lock or report conflict (non-blocking)
+    UNLOCK = 1;              // release lock's owner's locks over its range
+    GET_LK = 2;             // report a conflicting lock, if any
+    RELEASE_POSIX_OWNER = 3; // drop the owner's fcntl locks (flush-time)
+    RELEASE_FLOCK_OWNER = 4; // drop the owner's flock locks (release-time)
+}
+
+message PosixLockResponse {
+    bool granted = 1;            // for TRY_LOCK: whether the lock was granted
+    bool has_conflict = 2;       // whether conflict is populated
+    PosixLockRange conflict = 3; // the blocking lock (TRY_LOCK conflict / GET_LK result)
+}
+
 // ObjectTransactionBatch applies several object transactions in one round trip,
 // each under its own per-path lock and independent of the others (no cross-key
 // atomicity). A caller groups keys that route to the same owner filer and sends