smb: client: remove redundant lstrp update in negotiate protocol

Commit 34331d7 ("smb: client: fix first command failure during
re-negotiation") addressed a race condition by updating lstrp before
entering negotiate state. However, this approach may have some unintended
side effects.

The lstrp field is documented as "when we got last response from this
server", and updating it before actually receiving a server response
could potentially affect other mechanisms that rely on this timestamp.
For example, the SMB echo detection logic also uses lstrp as a reference
point. In scenarios with frequent user operations during reconnect states,
the repeated calls to cifs_negotiate_protocol() might continuously
update lstrp, which could interfere with the echo detection timing.

Additionally, commit 266b5d0 ("smb: client: fix race condition in
negotiate timeout by using more precise timing") introduced a dedicated
neg_start field specifically for tracking negotiate start time. This
provides a more precise solution for the original race condition while
preserving the intended semantics of lstrp.

Since the race condition is now properly handled by the neg_start
mechanism, the lstrp update in cifs_negotiate_protocol() is no longer
necessary and can be safely removed.

Fixes: 266b5d0 ("smb: client: fix race condition in negotiate timeout by using more precise timing")
Cc: [email protected]
Acked-by: Paulo Alcantara (Red Hat) <[email protected]>
Signed-off-by: Wang Zhaolong <[email protected]>
Signed-off-by: Steve French <[email protected]>

Author: Wang Zhaolong

fs/smb/client/connect.c

@@ -4205,7 +4205,6 @@ cifs_negotiate_protocol(const unsigned int xid, struct cifs_ses *ses,
 		return 0;
 	}
 
-	server->lstrp = jiffies;
 	server->tcpStatus = CifsInNegotiate;
 	server->neg_start = jiffies;
 	spin_unlock(&server->srv_lock);