[linux-cifs-client] mount -t cifs : when try to cp, system hangs up

[Steven French]

> I am troubleshooting an issue with smbfs and special characters 
> (ã and à (atilde)) , to be more specific and was told to try CIFS.

cifs vfs (2.4 and 2.6) and smbfs (in 2.6 or with some patches) can use 
Unicode when contacting servers such as later versions of Samba, and 
Windows 2000 and later whic support Unicode.   Also using the iocharset 
mount option (for either cifs vfs or smbfs) may be helpful in forcing a 
particular character mapping.

> when I try to copy *any* file (small files also), I get nothing on 
> the console after my [ENTER]

Not something I remember as being a problem in 0.9.9, and certainly not 
something i have run into on 2.4.22 with my current cifs vfs testing. 
Sounds like it oopsed in the kernel, probably before it got a chance to 
write any information to /var/log/messages or equivalent.   Enabling 
/proc/fs/cifs/cifsFYI may write a little bit of data to the log before the 
oops (you could view the messages on reboot presumably) or if you can 
build with a kdb or kgdb so we could see the data on a remote debug 
console... 

What could be done to narrow it down farther would be to trace the network 
traffic from that client's ip address to the server - using a network 
analyzer such as ethereal (or even the simple tcpdump utility) running on 
the server.   Since mount causes at least half a dozen different requests 
it would be helpful in narrowing this down to known the last response that 
the server sent to client.    Since this sounds like an interesting test 
of NLS codepage mapping, might also be interesting to check if you have 
compiled in the right NLS tables in the kernel, but that seems like a long 
shot.

Steve French
Senior Software Engineer
Linux Technology Center - IBM Austin
phone: 512-838-2294
email: sfrench at-sign us dot ibm dot com
-------------- next part --------------
HTML attachment scrubbed and removed