Red Hat Bugzilla – Bug 379451
Cifs unmount on shutdown takes more than a minute
Last modified: 2014-06-18 03:37:05 EDT
Description of problem:
If i shutdown my pc the cifs unmount takes very long. and if i take a look at
the dmesg log file i get the following error:
eth0: no IPv6 routers present
CIFS VFS: RFC1001 size 35 bigger than SMB for Mid=10124
Bad SMB: : dump of 48 bytes of data at 0xe7998a00
00000023 424d53ff 00000074 00018800 # . . . � S M B t . . . . . . .
00000000 00000000 00000000 69230000 . . . . . . . . . . . . . . # i
278c0000 0000ff00 12a50640 b400020a . . . ' . � . . @ . � . . . . �
This seem to be a cifs.ko issue, re-assigning to Jeff.
What sort of server is this?
the clients are:
redhat es 3 incl. latest updates
redhat es 4 incl. latest updates
fedora 8 incl. latest updates
the server is a netapp fas2050
here is the mount entry from fstab:
//foo/INSTALL /mnt/net/foo/install cifs
auto,credentials=/root/.smbcredentials,uid=500,gid=500 1 1
//foo/INSTALL on /mnt/net/foo/install type cifs (rw,mand)
I suspect that this is a duplicate of:
...you might want to try looking at captures with wireshark and see if the
server is sending a bogus BCC value in the LogoffANDX response. I'm going to
close this as a duplicate of that bug and suggest opening a case with netapp.
They may already have a filer patch that fixes this (I'm not sure).
Note that a patch recently went upstream to fix a potential buffer overflow in
CIFS. That patch has the side effect of making the client not expect a response
from the server for Logoff requests. That may also work around this problem, but
it's not in Fedora yet.
*** This bug has been marked as a duplicate of 191112 ***