Merge lp:~jamesodhunt/upstart/disable-system-bus into lp:upstart
Status: | Merged |
---|---|
Merged at revision: | 1555 |
Proposed branch: | lp:~jamesodhunt/upstart/disable-system-bus |
Merge into: | lp:upstart |
Diff against target: |
216 lines (+119/-10) (has conflicts) 4 files modified
ChangeLog (+11/-0) init/main.c (+37/-10) init/man/init.8 (+4/-0) util/tests/test_initctl.c (+67/-0) Text conflict in ChangeLog Text conflict in init/main.c Text conflict in util/tests/test_initctl.c |
To merge this branch: | bzr merge lp:~jamesodhunt/upstart/disable-system-bus |
Related bugs: |
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
James Hunt | Pending | ||
Review via email: mp+193764@code.launchpad.net |
This proposal supersedes a proposal from 2013-09-19.
Description of the change
* init/control.c: Typo.
* init/main.c: Add '--no-dbus' command-line option.
init/man/init.8: Added '--no-dbus' option.
This option has the effect of stopping Session Inits having access to system-level events for those systems which require such behaviour; in such environments, even if the upstart-
A side-effect of booting with '--no-dbus' is that a non-priv user will be unable to query system jobs using initctl (since such users will not have access to the private socket). However, for those making use of '--no-dbus', such behaviour would be deemed a security advantage rather than a limitation.
The code looks good to me, we should just add one test to confirm that upstart with --disable-dbus won't do anything on SIGUSR1.