[linux-cifs-client] trying to mount.cifs a DFS share - system freeze

[allen.orlowski at thomson.com]

Best bet is to mount the hidden share, and not the dfs root.

So do a 
Mount.cifs //192.168.10.10/<sharename>$ /mnt/1 -o username=<user>  //
then prompts you for a password.

Or what I mostly do is not use mount.cifs and do the following.

mount -o username=<user>/<domain_name> //<server>/<share>$
/mnt_dexnas02/

I use a 2.6.9-42 kernel.

I used to mount the root dfs share, until our group was taken away from
local admins, so I mount the hidden share for our department and all is
well.

Hope that helps


-----Original Message-----
From:
linux-cifs-client-bounces+allen.orlowski=thomson.com at lists.samba.org
[mailto:linux-cifs-client-bounces+allen.orlowski=thomson.com at lists.samba
.org] On Behalf Of Tomasz Chmielewski
Sent: Friday, March 23, 2007 6:59 AM
To: linux-cifs-client at lists.samba.org
Subject: [linux-cifs-client] trying to mount.cifs a DFS share - system
freeze

I just tried to mount a DFS share from a server running kernel
2.6.17.13:

mount.cifs //192.168.10.10/dfs /mnt/1 -o username=someusername
Password:


When I look into /mnt/1, I can see directories offered by DFS, let's
say:


/mnt/1
      /files
      /users
      /homes
      /soft

Unfortunately, the content isn't there - inside those directories I can 
see DFS share again, up to two levels:

/mnt/1
      /files
      /users
      /homes
      /soft
           /files
           /users
           /homes
           /soft


When I wanted to travel deeper, it returned an error, and the whole 
/mnt/1 was gone.

Then, when I tried to unmount /mnt/1, the server just died (it's a 
remote machine, so it's a pity).


I tried to reproduce locally with 2.6.21-rc3 kernel - whenever I want to

see three levels down the mounted DFS share, kernel shows some nice
info:

slab error in verify_redzone_free(): cache `size-64': memory outside 
object was overwritten
  [<c010524c>] show_trace_log_lvl+0x1a/0x2f
  [<c01057cc>] show_trace+0x12/0x14
  [<c0105850>] dump_stack+0x16/0x18
  [<c015d09d>] __slab_error+0x26/0x28
  [<c015d5fa>] cache_free_debugcheck+0xa6/0x181
  [<c015d916>] kfree+0x6d/0xb9
  [<c020d6f1>] release_one_tty+0x7f/0xdc
  [<c020d773>] release_tty+0x25/0x35
  [<c020ef1b>] release_dev+0x5d4/0x61b
  [<c020ef6c>] tty_release+0xa/0xe
  [<c01626d9>] __fput+0xa1/0x14c
  [<c016279a>] fput+0x16/0x18
  [<c0160286>] filp_close+0x54/0x5c
  [<c0119d67>] put_files_struct+0x66/0xa8
  [<c011adde>] do_exit+0x228/0x701
  [<c011b327>] sys_exit_group+0x0/0x11
  [<c0122838>] get_signal_to_deliver+0x360/0x39f
  [<c0103385>] do_notify_resume+0x93/0x66f
  [<c0103dbd>] work_notifysig+0x13/0x26
  =======================
c50dd558: redzone 1:0x0, redzone 2:0x170fc2a5.


"memory outside object was overwritten" worries me a bit, and was 
perhaps why the remote machine running 2.6.17.13 kernel died.


Isn't it a severe bug?


-- 
Tomasz Chmielewski
http://wpkg.org




_______________________________________________
linux-cifs-client mailing list
linux-cifs-client at lists.samba.org
https://lists.samba.org/mailman/listinfo/linux-cifs-client