Description of problem: The init script that launches mailman qrunners uses the form "interpreter script" instead of "script" to launch the mailmanctl script. Because of this proper SELinux domain transitions cannot occur and the Mailman qrunner ends up in initrc_t domain. Version-Release number of selected component (if applicable): mailman-2.1.9-2 How reproducible: Always Steps to Reproduce: 1. Start mailman with SELinux in permissive mode. 2. Execute "ps xuwa -Z | grep qrunner" Actual results: The processes run in initrc_t domain. Expected results: The processes should run in mailman_mail_t domain. Additional info: A patch against the init script is attached.
Created attachment 161280 [details] A patch to get rid of direct interpreter invocation in mailman init script
I got some AVC messages after applying the patch. This means that we might need to update also the policy and need to test this for some time. Maybe we should start with this in Fedora first.
Created attachment 234841 [details] AVC messages
I've filled the bug #350511 -- the selinux-policy might need to be changed as well.
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on the solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHBA-2008-0416.html