[linux-cifs-client] Re: [mount.cifs] character set conversion

[Steve French (smfltc)]

>I have successfully mounted from my debian/etch Linux machine my 
>company's Windows file server. However, I still face a problem with the 
>appropriate encoding of the filenames. Filenames containing non-english 
>characters (greek chars in particular) appear as ????????? in the client.
>
>I suppose that the reason for this problem is an encoding mismatch, when 
>I mount the target directory. Probably the server dir filenames are 
>iso-8859-7 or cp-1253 encoded, while my terminal expects utf-8.
>
>I have extensively tried 'iocharset' option, but without match success. 
>Further, when I try to use cp-1253 in iocharset, I receive the following 
>error:
>"mount error 79 = Can not access a needed shared library", which 
>indicates a missing library.
>
>When I change my terminal from utf-8 to iso-8859-7 or cp-1253 the 
>problem remains exactly the same, regardless of the iocharset parameter 
>I use.
>
>  
>
Most Windows servers (since Windows NT 4 or later) default to using 
Unicode (UCS-16) on the wire.
If various filenames are appearing as ???????? on new files created and 
on existing files on the server,
then it is likely that the library (UCS-16 to UTF-8) is missing mappings 
for various characters.
Are you getting warning messages in the log about unmapped characters?

>I use mount.smbfs version 3.0.24 and linux kernel 2.6.18.4
>
>  
>
mount.smbfs or mount.cifs  (they are not related, mount.smbfs is not 
needed for cifs mounts)

>A similar question has been posted in this list (subject: charset 
>convertion issue with linux mount.cifs and samba 3.0.23 at Oct 16,2006), 
>but no appropriate answer was provided then.
>  
>
A change was made to the Linux cifs client (cifs.ko module) to better 
log characters which are
unmapped.