Bug 5791 - mount.cifs stuck in uninterruptable sleep - have to reboot
Summary: mount.cifs stuck in uninterruptable sleep - have to reboot
Status: RESOLVED FIXED
Alias: None
Product: CifsVFS
Classification: Unclassified
Component: kernel fs (show other bugs)
Version: 2.6
Hardware: x64 Linux
: P3 regression
Target Milestone: ---
Assignee: Steve French
QA Contact:
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2008-09-24 08:33 UTC by Moritz
Modified: 2009-03-07 11:05 UTC (History)
1 user (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Moritz 2008-09-24 08:33:14 UTC
client: debian, linux 2.6.23-1-amd64, cifs 1.50
server1: windows 2003 server, latest sp
server2: windows 2003 server, clean install (without sp)
mount opts: rw,user=user,dom=domain,pass=pass

On many of our machines mount.cifs hangs with state "D" in ps.

Doing a echo 7 > cifsFYI i see in demsg:

 fs/cifs/connect.c: Socket created
 fs/cifs/connect.c: Error -512 connecting to server via ipv4
 fs/cifs/connect.c: reconnect error -512

-512 is ERESTARTSYS. Makeing the server available / unavailable, plugging network cables and so on did not help, the only way to solve it is rebooting the machine, then it works for a while.

cat DebugData shows:
Active VFS Requests: 1718
a lot of:
8) Name: 10.32.1.83  Domain: DOMAIN Mounts: -1 OS: Windows Server 2003 R2 3790 Service Pack 2
        NOS: Windows Server 2003 R2 5.2 Capability: 0x1f3fd
        SMB session status: 2
Comment 1 Jeff Layton 2008-11-23 09:53:52 UTC
-ERESTARTSYS usually means that the task (which is probably cifsd in this case) caught a signal. That shouldn't happen -- cifsd should be insulated from signals originating in userspace. The kernel should only send it a signal when it's taking down the thread.

The Mounts -1 here means that the refcounting is broken (that should never dip below 0).

Are you able to reproduce this on recent kernels? In particular, the current state of Linus tree or later? We've recently corrected a number of mount/umount refcounting bugs there.
Comment 2 Steve French 2009-02-18 23:30:31 UTC
As Jeff said, we fixed similar sounding problems in cifs mount in 2.6.27 stable series, and later kernels.
Comment 3 Steve French 2009-03-07 11:05:43 UTC
Please reopen if the changes to mount code in recent 2.6.27 and 2.6.28 do not fix it.