good news ... or not
I just try to add vfs object = audit to print$ share
and the bug dispear
strange ...
Le 13/04/2010 19:34, Volker Lendecke a écrit :
On Tue, Apr 13, 2010 at 05:26:57PM +0200, Christophe Deze wrote:
here is my valgrind.log
I hope it helps!
It does -- a bit:
==32107== Invalid read of size 4
==32107== at 0x4B86AF9: (within /usr/lib/samba/vfs/scannedonly.so)
==32107== by 0x815234F: smb_vfs_call_open (in /usr/sbin/smbd)
==32107== by 0x8149F56: (within /usr/sbin/smbd)
==32107== by 0x814C059: create_file_default (in /usr/sbin/smbd)
==32107== by 0x8189161: (within /usr/sbin/smbd)
==32107== by 0x81523FE: smb_vfs_call_create_file (in /usr/sbin/smbd)
==32107== by 0x83E165A: (within /usr/sbin/smbd)
==32107== by 0x83E1B7B: clean_up_driver_struct (in /usr/sbin/smbd)
==32107== by 0x82D179D: _spoolss_AddPrinterDriver (in /usr/sbin/smbd)
==32107== by 0x82D1E07: _spoolss_AddPrinterDriverEx (in /usr/sbin/smbd)
==32107== by 0x82E362C: (within /usr/sbin/smbd)
==32107== by 0x83272EA: api_pipe_request (in /usr/sbin/smbd)
==32107== Address 0x18 is not stack'd, malloc'd or (recently) free'd
It would be very helpful if you could recompile with -g to
get line numbers. From the message we see that apparently
some pointer (very likely "fsp") is NULL while
scannedonly_open is called. You might want to contact the
author of the scannedonly module, Olivier Sessink, directly
with this analysis. Either he is listening here on this
list, or you can find his email address from the checkin
reachable under
http://git.samba.org/?p=samba.git;a=commitdiff_plain;h=31e14285
Volker
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/options/samba
