Guys,

I am facing some strange error with the  open_file_ntcreate api in 3.0.28a version and symantec backup 1.12.5. We upgraded the samba from 3.0.14 to 3.0.28a for our NAS product. Everything was smooth until one of our customers found this issue with Symantec backup application. It doesn't work with NAS saying that file creation is not allowed. 

Here is the related log:

Mar  4 19:40:45 Star-6K smbd[26735]: [2003/03/04 19:40:45, 0] smbd/open.c:open_file_ntcreate(1724)
Mar  4 19:40:45 Star-6K smbd[26735]:   Could not get share mode lock
Mar  4 19:40:45 Star-6K smbd[26735]: [2003/03/04 19:40:45, 0] smbd/open.c:open_file_ntcreate(1724)
Mar  4 19:40:45 Star-6K smbd[26735]:   Could not get share mode lock
Mar  4 19:40:45 Star-6K smbd[26735]: [2003/03/04 19:40:45, 0] smbd/open.c:open_file_ntcreate(1724)
Mar  4 19:40:45 Star-6K smbd[26735]:   Could not get share mode lock
Mar  4 19:40:57 Star-6K smbd[26735]: [2003/03/04 19:40:57, 1] smbd/service.c:close_cnum(1231)
Mar  4 19:40:57 Star-6K smbd[26735]:   nas-pc25 (192.168.200.79) closed connection to service share1


Actually we had modified the smbd/open.c to have common file locking working. Here is the patch:

diff -Naur source/smbd/open.c source/smbd/open.c
--- source/smbd/open.c  2008-07-24 01:56:13.948660848 -0700
+++ source/smbd/open.c  2008-07-29 00:04:08.676849728 -0700
@@ -1718,15 +1718,17 @@
            the kernel refuses the operations then the kernel is wrong.
           note that GPFS supports it as well - jmcd */

+       if ( fsp->fh->fd > 0 ) {
        ret_flock = SMB_VFS_KERNEL_FLOCK(fsp, fsp->fh->fd, share_access);
-       if(ret_flock == -1 ){
-
+       if( file_existed && (ret_flock == 0) ){
+               DEBUG(0, ("Could not get share mode lock\n"));
                TALLOC_FREE(lck);
                fd_close(conn, fsp);
                file_free(fsp);

                return NT_STATUS_SHARING_VIOLATION;
        }
+       }

And the line number log corresponds to is locking code. Can anyone review this patch and let me know what's going wrong?

Thanks
Sagar