[Bug 1853] New: smbfs reconnect does not quite work properly

[Martin J. Bligh]

http://bugme.osdl.org/show_bug.cgi?id=1853

           Summary: smbfs reconnect does not quite work properly
    Kernel Version: 2.6.1-rc3
            Status: NEW
          Severity: normal
             Owner: fs_samba-smb@xxxxxxxxxxxxxxxxxxxx
         Submitter: kernel@xxxxxxxxxxxxxxxxxxxxxx


Distribution: Debian testing/unstable 
Hardware Environment: PIII/450 
Software Environment:  
Problem Description: 
smbfs mount reconnection due to inactivity timeout should not pass failure 
back to userspace 
 
Steps to reproduce: 
* Mount remote smbfs drive (eg. from a Win98 or W2K box). 
* Wait for it to time out (eg. by not accessing it for a couple of hours). 
* Access the drive. 
    Under 2.4.x I would have seen message such as: 
      Jan 10 06:25:48 wibble kernel: smb_get_length: recv error = 5 
      Jan 10 06:25:48 wibble kernel: smb_request: result -5, setting invalid 
      Jan 10 06:25:48 wibble kernel: smb_retry: successful, new pid=31576, 
generation=43 
    The connection would immediately get reestablished and userspace would be 
none the wiser that anything went wrong. 
    Under 2.6.1-rc3 however: 
      Jan 12 02:01:28 wibble kernel: SMB connection re-established (-5) 
      Jan 12 02:01:28 wibble kernel: smb_lookup: find //Program Files failed, 
error=-57 
      And userspace would see an error: 
      find: /mnt/blue: Invalid slot 
    Subsequent accesses to the drive work fine. It would be infinitely 
preferable for the error not reach userspace.


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/