linux-cifs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH][SMB3 client] cifs: Don't need state locking in smb2_get_mid_entry()
@ 2025-08-06 23:22 Steve French
  0 siblings, 0 replies; only message in thread
From: Steve French @ 2025-08-06 23:22 UTC (permalink / raw)
  To: CIFS; +Cc: David Howells

[-- Attachment #1: Type: text/plain, Size: 489 bytes --]

An attempt to rebase one of David's patches to current for-next.
There might be value in this cleanup.   Any opinions?  See attached.

There's no need to get ->srv_lock or ->ses_lock in smb2_get_mid_entry() as
all that happens of relevance (to the lock) inside the locked sections is
the reading of one status value in each.

Replace the locking with READ_ONCE() and use a switch instead of a chain of
if-statements.

Signed-off-by: David Howells <dhowells@redhat.com>

-- 
Thanks,

Steve

[-- Attachment #2: 0001-cifs-Don-t-need-state-locking-in-smb2_get_mid_entry.patch --]
[-- Type: text/x-patch, Size: 2699 bytes --]

From 469464ba8e225aa715677eb632fae92eba0bd288 Mon Sep 17 00:00:00 2001
From: Steve French <stfrench@microsoft.com>
Date: Wed, 6 Aug 2025 18:17:07 -0500
Subject: [PATCH] cifs: Don't need state locking in smb2_get_mid_entry()

There's no need to get ->srv_lock or ->ses_lock in smb2_get_mid_entry() as
all that happens of relevance (to the lock) inside the locked sections is
the reading of one status value in each.

Replace the locking with READ_ONCE() and use a switch instead of a chain of
if-statements.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Paulo Alcantara <pc@manguebit.org>
cc: Shyam Prasad N <sprasad@microsoft.com>
cc: Tom Talpey <tom@talpey.com>
cc: linux-cifs@vger.kernel.org
Signed-off-by: Steve French <stfrench@microsoft.com>
---
 fs/smb/client/smb2transport.c | 44 ++++++++++++++---------------------
 1 file changed, 18 insertions(+), 26 deletions(-)

diff --git a/fs/smb/client/smb2transport.c b/fs/smb/client/smb2transport.c
index ff9ef7fcd010..956173a930b7 100644
--- a/fs/smb/client/smb2transport.c
+++ b/fs/smb/client/smb2transport.c
@@ -799,43 +799,35 @@ static int
 smb2_get_mid_entry(struct cifs_ses *ses, struct TCP_Server_Info *server,
 		   struct smb2_hdr *shdr, struct mid_q_entry **mid)
 {
-	spin_lock(&server->srv_lock);
-	if (server->tcpStatus == CifsExiting) {
-		spin_unlock(&server->srv_lock);
+	switch (READ_ONCE(server->tcpStatus)) {
+	case CifsExiting:
 		return -ENOENT;
-	}
-
-	if (server->tcpStatus == CifsNeedReconnect) {
-		spin_unlock(&server->srv_lock);
+	case CifsNeedReconnect:
 		cifs_dbg(FYI, "tcp session dead - return to caller to retry\n");
 		return -EAGAIN;
+	case CifsNeedNegotiate:
+		if (shdr->Command != SMB2_NEGOTIATE)
+			return -EAGAIN;
+		break;
+	default:
+		break;
 	}
 
-	if (server->tcpStatus == CifsNeedNegotiate &&
-	   shdr->Command != SMB2_NEGOTIATE) {
-		spin_unlock(&server->srv_lock);
-		return -EAGAIN;
-	}
-	spin_unlock(&server->srv_lock);
-
-	spin_lock(&ses->ses_lock);
-	if (ses->ses_status == SES_NEW) {
+	switch (READ_ONCE(ses->ses_status)) {
+	case SES_NEW:
 		if ((shdr->Command != SMB2_SESSION_SETUP) &&
-		    (shdr->Command != SMB2_NEGOTIATE)) {
-			spin_unlock(&ses->ses_lock);
+		    (shdr->Command != SMB2_NEGOTIATE))
 			return -EAGAIN;
-		}
 		/* else ok - we are setting up session */
-	}
-
-	if (ses->ses_status == SES_EXITING) {
-		if (shdr->Command != SMB2_LOGOFF) {
-			spin_unlock(&ses->ses_lock);
+		break;
+	case SES_EXITING:
+		if (shdr->Command != SMB2_LOGOFF)
 			return -EAGAIN;
-		}
 		/* else ok - we are shutting down the session */
+		break;
+	default:
+		break;
 	}
-	spin_unlock(&ses->ses_lock);
 
 	*mid = smb2_mid_entry_alloc(shdr, server);
 	if (*mid == NULL)
-- 
2.43.0


^ permalink raw reply related	[flat|nested] only message in thread

only message in thread, other threads:[~2025-08-06 23:22 UTC | newest]

Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-08-06 23:22 [PATCH][SMB3 client] cifs: Don't need state locking in smb2_get_mid_entry() Steve French

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).