I removed my patch mapping ERRDOS 50 to errno 0 and applied your patch. Voila! it fixed the problem I had been experiencing.
I did not get any syslog messages though, I was a little confused about your comment about printing the fileid.
There is another problem that I discovered it has to do with have the READ-ONLY attribute set on the Win95 or WFW3.11 file.
On my linux system ls shows rwxrwxrwx for permissions. But if I try to update the file I get an errno-13 file access error.
Is it possible for the ls command to show the file as read-only?
Thanks for all your help,
Steve
Bill Hawes wrote:
> Steve Clark wrote:
> > If I copy a text file from my linux system to either of the above
> > platforms and the target filename does not already exist, the file is
> > copied but I receive and EIO error from the cp command. If I immediately
> > try to do the same thing again I get the EIO error but now the length of
> > the destination file is ZERO.
>
> OK, I have a theory on what might be going wrong. Newer SMB servers
> (e.g. Win NT) apparently create files already open and return a
> filehandle. The current smbfs isn't ready to work with this, so it
> closes the file immediately.
>
> But if an older server -- Win 95 or 3.11? -- doesn't open the file, then
> the attempted close might be generating the spurious error code. (The
> fileid is probably bogus.)
>
> To test this out, I've added code to print out the fileid after a
> create, and to close it only if the protocol level is NT1 _and_ the
> fileid is non-zero. (Win 95 unfortunately identifies itself as level
> NT1, so just the protocol test isn't enough.)
>
> Please give the attached patch a try. I'd like to track down the source
> of the problem rather than just squashing a return error code.
>
> Regards,
> Bill
>