Commit message (Collapse) | Author | Age | Files | Lines | ||
---|---|---|---|---|---|---|
... | ||||||
* | separate "touch /.autorelabel" to a new unit | Michal Schmidt | 2011-04-04 | 3 | -4/+16 | |
| | | | | | | | | | | | | | | | | | | | Writing /.autorelabel from fedora-autorelabel does not work because the script is executed only if relabel was already requested. Create a new unit fedora-autorelabel-mark.service which will be responsible for creating /.autorelabel if SELinux is disabled. The unit takes advantage of the new "ConditionSecurity=" support in systemd. The old script checked for a read-only filesystem first. The new unit does not do that. If / is read-only, touch will simply fail. This should not be considered as a failure of the unit, so "-" is used in ExecStart. There have been arguments on systemd-devel that the root directory should not be abused for flag files like /.autorelabel. It has a long tradition in Fedora though (since 2005). Maybe we can change it eventually, but let's keep it where it is for now. | |||||
* | simplify the check for plymouth in fedora-autorelabel | Michal Schmidt | 2011-04-04 | 1 | -4/+1 | |
| | | | | | The result of the check is needed only once, so there's no advantage in caching it in a variable. | |||||
* | let systemd check the conditions for fedora-autorelabel | Michal Schmidt | 2011-04-04 | 2 | -18/+4 | |
| | | | | | | | The service unit file already checks whether /.autorelabel or the "autorelabel" parameter are present. No need to recheck in the script. Using the new "ConditionSecurity=" feature systemd can also check whether SELinux is enabled. | |||||
* | crypttab should be 0600. | Bill Nottingham | 2011-03-30 | 1 | -0/+1 | |
| | ||||||
* | Don't explictly disallow IPv6 aliases (#583409, #612877.) | Bill Nottingham | 2011-03-30 | 2 | -4/+2 | |
| | ||||||
* | Document NM_CONTROLLED. (#692123) | Bill Nottingham | 2011-03-30 | 1 | -0/+3 | |
| | ||||||
* | Don't print errors if mdadm isn't installed. (#692187) | Bill Nottingham | 2011-03-30 | 1 | -2/+2 | |
| | ||||||
* | 9.28-1 | Bill Nottingham | 2011-03-30 | 1 | -1/+5 | |
| | ||||||
* | Add an error when setting gateway fails. (#672202) | Bill Nottingham | 2011-03-30 | 1 | -2/+4 | |
| | ||||||
* | ifup-eth: use /run/initramfs rather then /dev/.run/initramfs | Harald Hoyer | 2011-03-30 | 1 | -1/+1 | |
| | ||||||
* | Convert ipv6_log users to net_log. | Bill Nottingham | 2011-03-25 | 1 | -204/+51 | |
| | ||||||
* | Use net_log where appropriate. | Bill Nottingham | 2011-03-25 | 12 | -52/+52 | |
| | ||||||
* | Switch the default logging value to error. | Bill Nottingham | 2011-03-25 | 1 | -2/+2 | |
| | | | | It's what's used the most. | |||||
* | Make net_log()'s output similar to that of the ipv6 logger. | Bill Nottingham | 2011-03-25 | 1 | -1/+16 | |
| | | | | This will cause output strings to change, of course, so it's not really appropriate for earlier releases. | |||||
* | Add a net_log() function. (#507515, #689525) | Bill Nottingham | 2011-03-25 | 1 | -0/+23 | |
| | ||||||
* | Add a comment explaining the oft-confused -t nox,y,z mount syntax. | Bill Nottingham | 2011-03-22 | 2 | -0/+2 | |
| | | | | | The 'no' applies to all listed filesystem types, not just the first. Since this comes up in bugzilla once every month or two, explicitly comment it. | |||||
* | 9.27-1initscripts-9.27-1 | Bill Nottingham | 2011-03-17 | 1 | -1/+6 | |
| | ||||||
* | Fix mishandled argument to fstab-decode. (#685137) | Bill Nottingham | 2011-03-15 | 1 | -1/+1 | |
| | ||||||
* | Y kant eye spell? (#684465) | Bill Nottingham | 2011-03-14 | 1 | -2/+2 | |
| | ||||||
* | Support ipv6 routing rules by merging IPv4/IPv6, and route/rule code. ↵ | Bill Nottingham | 2011-03-11 | 3 | -26/+30 | |
| | | | | (#680872, <tobiasoed@hotmail.com>) | |||||
* | Fix conditions so that they operate as OR, not AND. (#684125) | Bill Nottingham | 2011-03-11 | 2 | -4/+4 | |
| | ||||||
* | 9.26-1initscripts-9.26-1 | Bill Nottingham | 2011-03-10 | 1 | -1/+14 | |
| | ||||||
* | Revert "Call sync after nfs unmount, otherwise we'll hang when the kernel ↵ | Bill Nottingham | 2011-03-10 | 1 | -1/+0 | |
| | | | | | | | | | syncs later. (#637500)" This reverts commit decf19bb9dc7b70ad89f9154899e73df069f5e62. This is better handled by not doing a force/lazy umount the first time. (a9b0d6b5c655da96783851d5304c4d800d4e4553) | |||||
* | Fix check for unmanaged devices so it does the right thing with devices NM ↵ | Bill Nottingham | 2011-03-10 | 1 | -1/+1 | |
| | | | | | | doesn't know about at all. (#670154) It was returning nonexistent devices, or devies NM knew nothing about (bridges), as being managed. | |||||
* | network-scripts/ifup-eth: fix cec3312 and close braces | Harald Hoyer | 2011-03-10 | 1 | -1/+1 | |
| | ||||||
* | network-scripts/ifup-eth: also check /dev/.run/initramfs | Harald Hoyer | 2011-03-10 | 1 | -1/+1 | |
| | ||||||
* | initscripts: execute fedora-sysinit-unhack after sysinit.target | Andrey Borzenkov | 2011-03-10 | 1 | -1/+2 | |
| | | | | | | | | | Currently both are started early during boot and in the best case are no-op, in the worst - hack is executed after unhack and leaves /dev/.in_sysinit hanging around forever. Make sure unhack is executed after hack and after sysinit.target as name suggests. Signed-off-by: Andrey Borzenkov <arvidjaar@gmail.com> | |||||
* | Don't do force/lazy umount for the first nfs umount. (#676851, ↵ | Bill Nottingham | 2011-03-07 | 1 | -2/+2 | |
| | | | | | | | <jlayton@redhat.com>) This can cause hung tasks that aren't cleaned up or killed properly, as their dangling references won't get caught. | |||||
* | Initial transifex.net support, updated translations. | Bill Nottingham | 2011-03-03 | 12 | -976/+1077 | |
| | ||||||
* | Update and improve IPv6 disabling instructions. | Ville Skyttä | 2011-03-01 | 1 | -3/+6 | |
| | ||||||
* | Fix path to sysconfig.txt in comments. | Ville Skyttä | 2011-03-01 | 1 | -1/+1 | |
| | ||||||
* | Avoid some unnecessary stat calls. | Ville Skyttä | 2011-03-01 | 2 | -7/+5 | |
| | ||||||
* | Move sysctl.conf/sysctl.d handling to a function, call it where appropriate. ↵ | Bill Nottingham | 2011-02-28 | 3 | -7/+12 | |
| | | | | (#593211, continued) | |||||
* | Tweak waiting times slightly. (#596451) | Bill Nottingham | 2011-02-28 | 1 | -2/+2 | |
| | ||||||
* | Make killproc more granular when delay is passed. (#428029, <xjakub@fi.muni.cz>) | Bill Nottingham | 2011-02-28 | 1 | -6/+12 | |
| | ||||||
* | Add GVRP support (#597598, <tomek@jot23.org>) | Bill Nottingham | 2011-02-28 | 2 | -1/+8 | |
| | ||||||
* | Add support for noauto crypt devices, to mirror systemd. | Bill Nottingham | 2011-02-28 | 1 | -0/+3 | |
| | ||||||
* | Add some more docs (#613722) | Bill Nottingham | 2011-02-28 | 1 | -1/+8 | |
| | ||||||
* | Drop some unnecessary command invocations. | Ville Skyttä | 2011-02-28 | 12 | -33/+32 | |
| | ||||||
* | Allow for plymouth to enter non-LUKS passphrases (#621158, ↵ | Bill Nottingham | 2011-02-28 | 1 | -4/+10 | |
| | | | | <redhat-bugzilla@camperquake.de>) | |||||
* | Make more accurate. | Bill Nottingham | 2011-02-25 | 1 | -4/+8 | |
| | ||||||
* | Call sync after nfs unmount, otherwise we'll hang when the kernel syncs ↵ | Bill Nottingham | 2011-02-25 | 1 | -0/+1 | |
| | | | | later. (#637500) | |||||
* | remove network device naming requirement from VLAN devices | Matt Domsch | 2011-02-25 | 1 | -2/+2 | |
| | ||||||
* | Update to 9.25-1initscripts-9.25-1 | Bill Nottingham | 2011-02-25 | 1 | -1/+9 | |
| | ||||||
* | Remove outdated "Red Hat Linux" comments from sysctl.conf. | Ville Skyttä | 2011-02-25 | 3 | -3/+3 | |
| | ||||||
* | Add support for sysctl.d (#59311) | Martin Langhoff | 2011-02-25 | 1 | -0/+4 | |
| | ||||||
* | Fix OMAP patch (#678875) | Bill Nottingham | 2011-02-23 | 1 | -1/+6 | |
| | ||||||
* | Support OMAP serial consoles (#678875, <ndevos@redhat.com>) | Bill Nottingham | 2011-02-21 | 1 | -0/+2 | |
| | ||||||
* | Accept --ignore-dependencies, --skip-redirect in /sbin/service. | Bill Nottingham | 2011-02-17 | 1 | -1/+9 | |
| | ||||||
* | systemctl: honour $SYSTEMCTL_IGNORE_DEPENDENCIES | Lennart Poettering | 2011-02-17 | 1 | -1/+6 | |
| | | | | | | | | | | | | | | | | Here's a patch for initscripts: If SYSTEMCTL_IGNORE_DEPENDENCIES is set, pass --ignore-dependencies to systemctl. This allows programs which invoke /sbin/service to get more SysV-like behaviour, i.e. where starting a service does not pull in any dependencies and honours no ordering. By default we continue to honour the dependencies since this is usually safe and allows other software to hook into all services regardless by who they are started. This is the result of a discussion with the IPA people, Simo Sorce in particular. |