[linux-cifs-client] Re:CaseInsensitivePassword

[Steve French]

Does the attached patch fix your problem?  It seems harmless (if someone
disagrees let me know - I would rather be safe, or at least make it
optional, if there is a problem that it could cause).  It dups the 24
byte NTLM hash in the lanman password field (I am very curious what
windows client does when its registry is changed to only send ntlm too).

I have pushed this patch to the bk tree, but can back it out if someone
objects.
-------------- next part --------------
--- fs/cifs/connect.c	2004-12-01 01:12:41.525332272 -0600
+++ /home/stevef/connect.c	2004-12-01 01:02:14.170704672 -0600
@@ -1574,15 +1574,14 @@
 		capabilities |= CAP_DFS;
 	}
 	pSMB->req_no_secext.Capabilities = cpu_to_le32(capabilities);
-
-	pSMB->req_no_secext.CaseInsensitivePasswordLength = 
-		cpu_to_le16(CIFS_SESSION_KEY_SIZE);
-
+	/* pSMB->req_no_secext.CaseInsensitivePasswordLength =
+	   CIFS_SESSION_KEY_SIZE; */
+	pSMB->req_no_secext.CaseInsensitivePasswordLength = 0;
 	pSMB->req_no_secext.CaseSensitivePasswordLength =
 	    cpu_to_le16(CIFS_SESSION_KEY_SIZE);
 	bcc_ptr = pByteArea(smb_buffer);
-	memcpy(bcc_ptr, (char *) session_key, CIFS_SESSION_KEY_SIZE);
-	bcc_ptr += CIFS_SESSION_KEY_SIZE;
+	/* memcpy(bcc_ptr, (char *) lm_session_key, CIFS_SESSION_KEY_SIZE);
+	   bcc_ptr += CIFS_SESSION_KEY_SIZE; */
 	memcpy(bcc_ptr, (char *) session_key, CIFS_SESSION_KEY_SIZE);
 	bcc_ptr += CIFS_SESSION_KEY_SIZE;