Skip to contentSkip to conversation

ksmbd: refactor smb2_notify() to a blocking wait

PATCH In review ChenXiaoSong 2 participants 5 messages Aug 25 – Aug 28

Subsystems (2)
lore
Share & tools
lore

ChenXiaoSong

author reviewer
#0 lore raw

v5 cover ksmbd: refactor smb2_notify() to a blocking wait

From: ChenXiaoSong <[email protected]>

v4->v5: Rebased on ksmbd-for-next branch.

v4: https://lore.kernel.org/linux-cifs/[email protected]/

ChenXiaoSong (1):
smb/server: support compound fid in notify requests

Gael Blivet (2):
ksmbd: temporarily remove interim-work change notify handling
ksmbd: refactor smb2_notify() to a blocking wait

 fs/smb/server/ksmbd_work.c |   3 -
 fs/smb/server/ksmbd_work.h |   5 +-
 fs/smb/server/smb2pdu.c    | 349 +++++++------------------------------
 fs/smb/server/vfs_cache.c  |  48 -----
 fs/smb/server/vfs_cache.h  |   6 -
 5 files changed, 65 insertions(+), 346 deletions(-)
Signature · 1 line
2.55.0
author reviewer
#1 lore raw

PATCH v5 1/3 smb/server: support compound fid in notify requests

From: ChenXiaoSong <[email protected]>

A Windows client can send a compound request containing:

Create Request, File: <share>; Notify Request

The Notify Request uses FFFF...FFFF as the compound FID.

---
 fs/smb/server/smb2pdu.c | 41 -----------------------------------------
 1 file changed, 41 deletions(-)
fs/smb/server/smb2pdu.c +0 -41
index 2fbd9010513e..86f93f81cdfd 100644
@@ -11818,47 +11818,6 @@ int smb2_notify(struct ksmbd_work *work)
 		return -EIO;
 	}
 
-	/*
-	 * macOS backupd sends CHANGE_NOTIFY with FileId=FFFF...FFFF (share-root
-	 * sentinel) to watch for changes on the share root without holding an
-	 * open handle. Respond STATUS_PENDING + STATUS_NOTIFY_CLEANUP immediately;
-	 * without this, backupd aborts Time Machine setup on STATUS_FILE_CLOSED.
-	 */
-	if (req->VolatileFileId == SMB2_NO_FID &&
-	    req->PersistentFileId == SMB2_NO_FID) {
-		in_work = ksmbd_alloc_work_struct();
-		if (!in_work || allocate_interim_rsp_buf(in_work)) {
-			if (in_work)
-				ksmbd_free_work_struct(in_work);
-			rsp->hdr.Status = STATUS_INSUFFICIENT_RESOURCES;
-			smb2_set_err_rsp(work);
-			return 0;
-		}
-		if (setup_async_work(work, NULL, NULL)) {
-			ksmbd_free_work_struct(in_work);
-			rsp->hdr.Status = STATUS_INSUFFICIENT_RESOURCES;
-			smb2_set_err_rsp(work);
-			return 0;
-		}
-		smb2_send_interim_resp(work, STATUS_PENDING);
-		in_work->conn = work->conn;
-		in_hdr = smb_get_msg(in_work->response_buf);
-		memcpy(in_hdr, ksmbd_resp_buf_next(work),
-		       __SMB2_HEADER_STRUCTURE_SIZE);
-		in_hdr->Flags |= SMB2_FLAGS_ASYNC_COMMAND;
-		in_hdr->Id.AsyncId = cpu_to_le64(work->async_id);
-		smb2_set_err_rsp(in_work);
-		in_hdr->Status = STATUS_NOTIFY_CLEANUP;
-		in_work->async_id = work->async_id;
-		work->async_id = 0;
-		release_async_work(work);
-		if (smb2_send_interim_work(in_work, work, false))
-			ksmbd_debug(SMB, "failed to send notify cleanup\n");
-		ksmbd_free_work_struct(in_work);
-		work->send_no_response = 1;
-		return 0;
-	}
-
 	/*
 	 * KSMBD does not implement a real change-notification backend.
 	 * Genuine SMB2 servers (and macOS smbfs) never complete a
Signature · 1 line
2.55.0
author reviewer
#2 lore raw

PATCH v5 2/3 ksmbd: temporarily remove interim-work change notify handling

In reply to ChenXiaoSong

From: Gael Blivet <[email protected]>

Prepare for the next patch. Split the removal of the old implementation into this separate patch to make the next patch easier to review.

The current change notify path keeps a synthetic ksmbd_work, named in_work, after the original request has returned. It also needs a per-file notify_pendings list and special close, cancel, and connection lifetime handling to complete and free that work.

Next patch replaces this design with a blocking wait in smb2_notify().

This patch reverts the following patches (or parts of them):

- Revert part of 2a0e037648da ("ksmbd: encrypt interim responses to encrypted requests")
- Revert 4818df5aaac0 ("ksmbd: wait for deferred notify cancellation")
- Revert part of a47634cd729b ("ksmbd: expire SMB sessions when Kerberos tickets expire")
- Revert f495154703cb ("ksmbd: retain connection for pending notify work")
- Revert part of 528af7cf69f3 ("smb/server: avoid registering async requests during connection close")
- Revert 152036e875e8 ("ksmbd: defer CHANGE_NOTIFY completion instead of STATUS_NOT_IMPLEMENTED")

---
 fs/smb/server/ksmbd_work.c |   3 -
 fs/smb/server/ksmbd_work.h |   4 -
 fs/smb/server/smb2pdu.c    | 271 +------------------------------------
 fs/smb/server/vfs_cache.c  |  48 -------
 fs/smb/server/vfs_cache.h  |   6 -
 5 files changed, 4 insertions(+), 328 deletions(-)
fs/smb/server/ksmbd_work.c +0 -3
index f35335307670..097de59f807c 100644
@@ -57,7 +57,6 @@ struct ksmbd_work *ksmbd_alloc_work_struct(void)
 		INIT_LIST_HEAD(&work->request_entry);
 		INIT_LIST_HEAD(&work->async_request_entry);
 		INIT_LIST_HEAD(&work->fp_entry);
-		INIT_LIST_HEAD(&work->notify_entry);
 		INIT_LIST_HEAD(&work->aux_read_list);
 		work->iov_alloc_cnt = ARRAY_SIZE(work->iov_inline);
 		work->iov = work->iov_inline;
@@ -87,8 +86,6 @@ void ksmbd_free_work_struct(struct ksmbd_work *work)
 
 	if (work->async_id)
 		ksmbd_release_id(&work->conn->async_ida, work->async_id);
-	if (work->owns_conn_ref)
-		ksmbd_conn_put(work->conn);
 	ksmbd_fd_put(work, work->request_open);
 	kmem_cache_free(work_cache, work);
 }
fs/smb/server/ksmbd_work.h +0 -4
index 5f1d3ebab4fb..0b55ad457cd3 100644
@@ -91,8 +91,6 @@ struct ksmbd_work {
 	bool                            compress_response:1;
 	/* Is this SYNC or ASYNC ksmbd_work */
 	bool                            asynchronous:1;
-	/* Work owns a reference to @conn. */
-	bool				owns_conn_ref:1;
 	bool                            need_invalidate_rkey:1;
 	bool				request_open_chseq_tracked:1;
 	bool				session_setup_reauth:1;
@@ -116,8 +114,6 @@ struct ksmbd_work {
 	/* List head at conn->async_requests */
 	struct list_head                async_request_entry;
 	struct list_head                fp_entry;
-	/* List head at ksmbd_file->notify_pendings */
-	struct list_head                notify_entry;
 };
 
 /**
fs/smb/server/smb2pdu.c +4 -267

311 lines not on this page · raw message

fs/smb/server/vfs_cache.c +0 -48

66 lines not on this page · raw message

fs/smb/server/vfs_cache.h +0 -6
index 502efb16f05f..e86e3cbe3bf2 100644
@@ -162,12 +162,6 @@ struct ksmbd_file {
 	unsigned int			outstanding_requests;
 	unsigned int			outstanding_pre_requests;
 	struct ksmbd_lock_sequence	lock_seq[KSMBD_LOCK_SEQ_ARRAY_SIZE];
-
-	/*
-	 * Pending CHANGE_NOTIFY completions for this handle, sent with
-	 * STATUS_NOTIFY_CLEANUP when the handle is closed.
-	 */
-	struct list_head		notify_pendings;
 };
 
 static inline void set_ctx_actor(struct dir_context *ctx,
Signature · 1 line
2.55.0
author reviewer
#3 lore raw

PATCH v5 3/3 ksmbd: refactor smb2_notify() to a blocking wait

In reply to ChenXiaoSong

From: Gael Blivet <[email protected]>

The previous design registered a synthetic work struct (in_work) directly on conn->async_requests and deferred the response to a workqueue -- a bespoke async/cancel implementation duplicating what setup_async_work(), release_async_work(), and smb2_send_interim_resp() already provide for smb2_lock()'s pending byte-range lock.

Replace it with that same pattern: setup_async_work() on the calling work itself, registered on fp->blocked_works, woken by cancel or by the handle closing via the existing set_close_state_blocked_works(). This removes the synthetic work struct, the notify_pendings list and its close-time drain, and the deferred workqueue send, leaving smb2_notify() sharing the same async/cancel machinery as smb2_lock() instead of its own separate copy.

The worker now blocks on ksmbd_wq for as long as the watch stays open, instead of returning immediately. This also makes the skeleton ready for a future event-delivery implementation on the same blocking wait.

---
 fs/smb/server/ksmbd_work.h |  1 +
 fs/smb/server/smb2pdu.c    | 93 ++++++++++++++++++++++++++++++++++++--
 2 files changed, 89 insertions(+), 5 deletions(-)
fs/smb/server/ksmbd_work.h +1 -0
index 0b55ad457cd3..4c0b7b2535b9 100644
@@ -113,6 +113,7 @@ struct ksmbd_work {
 	struct list_head                request_entry;
 	/* List head at conn->async_requests */
 	struct list_head                async_request_entry;
+	/* List head at ksmbd_file->blocked_works */
 	struct list_head                fp_entry;
 };
 
fs/smb/server/smb2pdu.c +88 -5

113 lines not on this page · raw message

Signature · 1 line
2.55.0
maintainer
#4 lore raw
In reply to ChenXiaoSong
Quoted 5 lines · ChenXiaoSong

From: ChenXiaoSong <[email protected]>

v4->v5: Rebased on ksmbd-for-next branch.

Quoted 8 lines

v4: https://lore.kernel.org/linux-cifs/[email protected]/

ChenXiaoSong (1):
smb/server: support compound fid in notify requests

Gael Blivet (2):
ksmbd: temporarily remove interim-work change notify handling
ksmbd: refactor smb2_notify() to a blocking wait

The smbtorture failure was caused by 0002 patch removing the signing path in smb2_send_notify_cancelled(). I combined patches 0002, 0003 and the signing path into a single patch.
And I will apply them to #ksmbd-for-next.
Thanks!