linux-kernel - Re: [PATCH v2 1/1] smb: client: Fix hang in smb2

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <258e2dba-da30-795f-6408-85c06e137b61@axis.com>
Date: Mon, 15 Apr 2024 16:35:02 +0200
From: Rickard x Andersson <rickaran@...s.com>
To: Steve French <smfrench@...il.com>
Cc: linux-cifs@...r.kernel.org, samba-technical@...ts.samba.org,
 sfrench@...ba.org, pc@...guebit.com, ronniesahlberg@...il.com,
 sprasad@...rosoft.com, tom@...pey.com, linux-kernel@...r.kernel.org,
 rickard314.andersson@...il.com, kernel@...s.com
Subject: Re: [PATCH v2 1/1] smb: client: Fix hang in smb2_reconnect

On 4/13/24 11:02, Steve French wrote:
> Shyam and I tried some experiments and there are two cases where the
> patch breaks:
> 1) ChangeNotify will time out
> 2) Certainly byte range lock calls (they can be allowed to block) will timeout
> 
> An obvious alternative would be to not make this change for the
> commands like ChangeNotify and blocking locks but allow it for the
> others.
> 

Would it make sense to make the patch less intrusive by doing something 
like the following?:

@@ -1209,7 +1216,12 @@ compound_send_recv(const unsigned int xid, struct 
cifs_ses *ses,
         spin_unlock(&ses->ses_lock);

         for (i = 0; i < num_rqst; i++) {
-               rc = wait_for_response(server, midQ[i]);
+               if (flags == CIFS_NEG_OP) {

+                       rc = wait_for_response(server, midQ[i], 50*1000);
+               }
+               else
+                       rc = wait_for_response(server, midQ[i], -1);
                 if (rc != 0)
                         break;



So, we are just waiting with timeout in the case it is a "CIFS_NEG_OP".

Note that I am not familiar at all with this code base.

Best regards,
Rickard