summaryrefslogtreecommitdiff
path: root/src
AgeCommit message (Collapse)Author
2011-04-28dbus: make daemon reexecution synchronousLennart Poettering
We simply keep open copies of the dbus connections across the reexecution and close them as last step of it. A client can thus simply wait until its connection is dropped to know when the reexecution is finished. https://bugzilla.redhat.com/show_bug.cgi?id=698198
2011-04-28label: don't print warning when we try ro relabel an non-existing fileLennart Poettering
https://bugzilla.redhat.com/show_bug.cgi?id=698674
2011-04-28service: properly notice when services with a main process that isn't a ↵Lennart Poettering
child of init die https://bugzilla.redhat.com/show_bug.cgi?id=699114
2011-04-28lookup: drop empty directories from search pathsLennart Poettering
2011-04-28path: simplify recheck logicLennart Poettering
2011-04-28unit: fix assert when trying to load unit instances for uninstanciable typesLennart Poettering
http://lists.freedesktop.org/archives/systemd-devel/2011-April/001910.html
2011-04-27systemctl: whitelist /etc instead of blacklisting /lib to figure out if ↵Lennart Poettering
[Install] is needed
2011-04-27def: lower default timeout to 90sLennart Poettering
Almost everybody found 3min too long, so lower it again
2011-04-27manager: fix serialization counterLennart Poettering
2011-04-27dbus: don't hit assert when dumping propertiesLennart Poettering
https://bugs.freedesktop.org/show_bug.cgi?id=36472
2011-04-26cryptsetup: fix keyfile size option processingMilan Broz
/etc/crypttab uses "size" argument only for plain crypt mapping (and input is limited to volume key size) For LUKS it always read whole keyfile (no arguments except "luks" should be in crypttab). Because of bug in cryptsetup 1.2.0 (fixed in 1.3.0) library did not enforced keyfile size argument, so this bug was hidden for some time. Patch also converts size in bits to bytes like excepted by libcryptsetup api. Fixes https://bugs.freedesktop.org/show_bug.cgi?id=36264
2011-04-26socket: improve warning message when we get POLLHUPLennart Poettering
2011-04-26strv: Fix gcc unitialized variable warningZbigniew Jędrzejewski-Szmek
Since strv_* functions handle null arguments, this warning is actually valid. src/strv.c: In function ‘strv_copy’: src/strv.c:68:21: warning: ‘k’ may be used uninitialized in this function [-Wuninitialized]
2011-04-26mount: failure to mount cgroup hierarchies should not be fatalLennart Poettering
If we cannot open /etc/cgroup, print an error message, but go on, to support cgroup-less builds. https://bugzilla.redhat.com/show_bug.cgi?id=628004
2011-04-25tmpfiles.d: switch to stacked config dirs in /lib, /etc, /runKay Sievers
2011-04-25sysctl.d, binfmt.d, modules-load.d: switch to stacked config dirs in /lib, ↵Kay Sievers
/etc, /run
2011-04-23manager: mkdir /run/systemd/system when starting upTollef Fog Heen
2011-04-21missing: Fix broken syscall(__NR_fanotify_mark... on o32 mips.cee1
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=%23618562 MIPS o32 ABI enforces that 64bit arguments should be 64bit-aligned, and the third argument of syscall(__NR_fanotify_mark, ...) is 64bit and not 64bit-aligned on o32 mips platform, thus extra padding is inserted before it. The syscall function doesn't know the prototype of fanotify_mark, so the padding will be passed into kernel as valid argument.
2011-04-21Revert "cryptsetup: pass 0 as key size to crypt_activate_by_keyfile()"Lennart Poettering
This reverts commit 9402841ef197e485061ab328edcaa0e6d83051bd. This freezes boot-up when /dev/urandom is used as key source
2011-04-20mount: Allow creating mount units for /var/lib/nfs/rpc_pipefs and /proc/fs/nfsd.Wulf C. Krueger
Mounting /var/lib/nfs/rpc_pipefs and /proc/fs/nfsd is required by rpcbind and the NFS server respectively. To depend on them being mounted, systemd needs to be allowed to create mount units for both in order to avoid having to implement hacks (mounting in ExecStartPre or using shell scripts).
2011-04-20cryptsetup: pass 0 as key size to crypt_activate_by_keyfile()Lennart Poettering
https://bugs.freedesktop.org/show_bug.cgi?id=36264
2011-04-20socket: support ListeSpecial= socketsLennart Poettering
2011-04-20vconsole: don't set console font/keymap if settings are emptyLennart Poettering
https://bugs.freedesktop.org/show_bug.cgi?id=36052 If KEYMAP= is set to the empty string in /etc/vconsole leave the kernel keymap loaded. Similar for FONT=.
2011-04-20nspawn: don't fail when we receive SIGCHLDLennart Poettering
https://bugs.freedesktop.org/show_bug.cgi?id=36148
2011-04-20cgroup: don't accidentaly trim on reloadLennart Poettering
https://bugzilla.redhat.com/show_bug.cgi?id=678555
2011-04-20socket: log more information about invalid poll eventsLennart Poettering
2011-04-20MeeGo supportChris E Ferron
This commit consists of the initial work to include MeeGo as a ported distribution for systemd. The majority of the changes are small configuration additions to auto tools, so that MeeGo is identified as a valid distribution option. Some small deviations will be noticed between the configuration of MeeGo and other distributions. As MeeGo is a distribution striving for compliancy to support its near embedded attributes and target users, there is less user configuration options available by default. Most services will be enabled by systemd as part of the distribution requirements, and as such most links and service files will be pre-setup for the MeeGo distribution. As much of this is going to be done within the MeeGo distribution packaging this is still noteworthy to mention, as it explains why in systemd you will observe configuration differences where the MeeGo distribution removes all links in the pkgsysconfdir for instance. MeeGo will be user configurable if there is desire, but most services will be enabled by the distribution as designated by the MeeGo compliancy standards. Other changes are in source to add such areas as meego-release defined in utils, and hostname in hostname-setup, defining vconsole-setup, localizations and rescue additions as needed. As this is all ground work, MeeGo will continue to strive for complete compatibility.
2011-04-20mount,crypto: rework meaning of noauto/nofailLennart Poettering
2011-04-19manager: when running in test mode, do not write generated unit files to ↵Lennart Poettering
/run/systemd/generator
2011-04-19mount: properly parse timeouts options in the middle of the stringLennart Poettering
2011-04-19hostnamed: improve error loggingLennart Poettering
2011-04-19execute: when we run as PID 1 the kernel doesn't give us CAP_SETPCAP by ↵Lennart Poettering
default. Get that temporarily when dropping capabilities for good
2011-04-16mount: make device timeout configurableLennart Poettering
2011-04-16cryptsetup: do not order crypto DM devices against the cryptsetup serviceLennart Poettering
2011-04-16socket: reuse existing FIFOsLennart Poettering
2011-04-16socket: guarantee order in which sockets are passed to be the one of the ↵Lennart Poettering
configuration file
2011-04-16systemctl: always consider unit files with no [Install] section but stored ↵Lennart Poettering
in /lib enabled
2011-04-16job: also print status messages when we successfully started a unitLennart Poettering
2011-04-16hostnamed: add reference to SMBIOS specsLennart Poettering
2011-04-16polkit: follow the usual syntax for polkit actionsLennart Poettering
2011-04-16hostnamed: introduce systemd-hostnamedLennart Poettering
http://www.freedesktop.org/wiki/Software/systemd/hostnamed
2011-04-16hostname: split out hostname validation into util.cLennart Poettering
2011-04-16cryptsetup: simplificationsLennart Poettering
2011-04-16dbus: split out object management code into dbus-common, and simplify itLennart Poettering
2011-04-16strv: properly override settings in env_append()Lennart Poettering
2011-04-16strv: detect non-assignments in env blocks properly in env_append()Lennart Poettering
2011-04-16strv: handle empty lists in strv_copy() properlyLennart Poettering
2011-04-16util: truncate newline inside of read_one_line_file() alreadyLennart Poettering
2011-04-16util: modernize get_parent_of_pid() a bitLennart Poettering
2011-04-16crypto: let the cryptsetup binary handles its own configurable timeoutsLennart Poettering