summaryrefslogtreecommitdiff
path: root/src/shared
AgeCommit message (Collapse)Author
2016-10-16tree-wide: introduce free_and_replace helperZbigniew Jędrzejewski-Szmek
It's a common pattern, so add a helper for it. A macro is necessary because a function that takes a pointer to a pointer would be type specific, similarly to cleanup functions. Seems better to use a macro.
2016-10-16tree-wide: use mfree moreZbigniew Jędrzejewski-Szmek
2016-10-16shared/install: use _cleanup_free_Zbigniew Jędrzejewski-Szmek
Also rewrap some comments so that they don't have a very long line and a very short line.
2016-10-12core:sandbox: Add ProtectKernelModules= optionDjalal Harouni
This is useful to turn off explicit module load and unload operations on modular kernels. This option removes CAP_SYS_MODULE from the capability bounding set for the unit, and installs a system call filter to block module system calls. This option will not prevent the kernel from loading modules using the module auto-load feature which is a system wide operation.
2016-10-12Allow block and char classes in DeviceAllow bus properties (#4353)Zbigniew Jędrzejewski-Szmek
Allowed paths are unified betwen the configuration file parses and the bus property checker. The biggest change is that the bus code now allows "block-" and "char-" classes. In addition, path_startswith("/dev") was used in the bus code, and startswith("/dev") was used in the config file code. It seems reasonable to use path_startswith() which allows a slightly broader class of strings. Fixes #3935.
2016-10-11Merge pull request #4348 from poettering/docfixesZbigniew Jędrzejewski-Szmek
Various smaller documentation fixes.
2016-10-11Merge pull request #4067 from poettering/invocation-idZbigniew Jędrzejewski-Szmek
Add an "invocation ID" concept to the service manager
2016-10-11pager: tiny beautificationLennart Poettering
2016-10-10install: let's always refer to the actual setting in errorsLennart Poettering
2016-10-08systemd-resolve: use sha256 for local-part of openpgp key (#4193)Stefan Schweter
2016-10-07core: add "invocation ID" concept to service managerLennart Poettering
This adds a new invocation ID concept to the service manager. The invocation ID identifies each runtime cycle of a unit uniquely. A new randomized 128bit ID is generated each time a unit moves from and inactive to an activating or active state. The primary usecase for this concept is to connect the runtime data PID 1 maintains about a service with the offline data the journal stores about it. Previously we'd use the unit name plus start/stop times, which however is highly racy since the journal will generally process log data after the service already ended. The "invocation ID" kinda matches the "boot ID" concept of the Linux kernel, except that it applies to an individual unit instead of the whole system. The invocation ID is passed to the activated processes as environment variable. It is additionally stored as extended attribute on the cgroup of the unit. The latter is used by journald to automatically retrieve it for each log logged message and attach it to the log entry. The environment variable is very easily accessible, even for unprivileged services. OTOH the extended attribute is only accessible to privileged processes (this is because cgroupfs only supports the "trusted." xattr namespace, not "user."). The environment variable may be altered by services, the extended attribute may not be, hence is the better choice for the journal. Note that reading the invocation ID off the extended attribute from journald is racy, similar to the way reading the unit name for a logging process is. This patch adds APIs to read the invocation ID to sd-id128: sd_id128_get_invocation() may be used in a similar fashion to sd_id128_get_boot(). PID1's own logging is updated to always include the invocation ID when it logs information about a unit. A new bus call GetUnitByInvocationID() is added that allows retrieving a bus path to a unit by its invocation ID. The bus path is built using the invocation ID, thus providing a path for referring to a unit that is valid only for the current runtime cycleof it. Outlook for the future: should the kernel eventually allow passing of cgroup information along AF_UNIX/SOCK_DGRAM messages via a unique cgroup id, then we can alter the invocation ID to be generated as hash from that rather than entirely randomly. This way we can derive the invocation race-freely from the messages.
2016-10-07bus-util: generalize helper for ID128 prpoertiesLennart Poettering
This way, we can make use of this in other code, too.
2016-10-05seccomp: add support for the s390 architecture (#4287)hbrueckner
Add seccomp support for the s390 architecture (31-bit and 64-bit) to systemd. This requires libseccomp >= 2.3.1.
2016-09-25core: add two new service settings ProtectKernelTunables= and ↵Lennart Poettering
ProtectControlGroups= If enabled, these will block write access to /sys, /proc/sys and /proc/sys/fs/cgroup.
2016-09-17Merge pull request #4123 from keszybz/network-file-dropinsMartin Pitt
Network file dropins
2016-09-16shared/conf-parser: add config_parse_many which takes strv with dirsZbigniew Jędrzejewski-Szmek
This way we don't have to create a nulstr just to unpack it in a moment.
2016-09-16tree-wide: rename config_parse_many to …_nulstrZbigniew Jędrzejewski-Szmek
In preparation for adding a version which takes a strv.
2016-09-15Merge pull request #4131 from intelfx/update-done-timestamps-precisionZbigniew Jędrzejewski-Szmek
condition: ignore nanoseconds in timestamps for ConditionNeedsUpdate= Fixes #4130.
2016-09-15update-done, condition: write the timestamp to the file as well and use it ↵Ivan Shapovalov
to prevent false-positives This fixes https://bugs.freedesktop.org/show_bug.cgi?id=90192 and #4130 for real. Also, remove timestamp check in update-done.c altogether since the whole operation is idempotent.
2016-09-13fileio: simplify mkostemp_safe() (#4090)Topi Miettinen
According to its manual page, flags given to mkostemp(3) shouldn't include O_RDWR, O_CREAT or O_EXCL flags as these are always included. Beyond those, the only flag that all callers (except a few tests where it probably doesn't matter) use is O_CLOEXEC, so set that unconditionally.
2016-09-10shared/install: fix set-default with empty root (#4118)Zbigniew Jędrzejewski-Szmek
https://bugzilla.redhat.com/show_bug.cgi?id=1374371 When root was empty or equal to "/", chroot_symlinks_same was called with root==NULL, and strjoina returned "", so the code thought both paths are equal even if they were not. Fix that by always providing a non-null first argument to strjoina.
2016-09-09shared: recognize DNS names with more than one trailing dot as invalid (#4111)Martin Pitt
One trailing dot is valid, but more than one isn't. This also fixes glibc's posix/tst-getaddrinfo5 test. Fixes #3978.
2016-09-06seccomp: also detect if seccomp filtering is enabledFelipe Sateler
In https://github.com/systemd/systemd/pull/4004 , a runtime detection method for seccomp was added. However, it does not detect the case where CONFIG_SECCOMP=y but CONFIG_SECCOMP_FILTER=n. This is possible if the architecture does not support filtering yet. Add a check for that case too. While at it, change get_proc_field usage to use PR_GET_SECCOMP prctl, as that should save a few system calls and (unnecessary) allocations. Previously, reading of /proc/self/stat was done as recommended by prctl(2) as safer. However, given that we need to do the prctl call anyway, lets skip opening, reading and parsing the file. Code for checking inspired by https://outflux.net/teach-seccomp/autodetect.html
2016-08-31networkd: add options to bridge (#4051)Tobias Jungel
This patch allows to configure AgeingTimeSec, Priority and DefaultPVID for bridge interfaces.
2016-08-30install: fix disable when /etc/systemd/system is a symlinkLukas Nykryn
2016-08-26Merge pull request #3984 from poettering/refcntEvgeny Vereshchagin
permit bus clients to pin units to avoid automatic GC
2016-08-22core: do not fail at step SECCOMP if there is no kernel support (#4004)Felipe Sateler
Fixes #3882
2016-08-22bus-util: turn on exit-on-disconnect for all command line toolsLennart Poettering
bus_connect_transport() is exclusively used from our command line tools, hence let's set exit-on-disconnect for all of them, making behaviour a bit nicer in case dbus-daemon goes down.
2016-08-22run: optionally, wait for the service to finish and show its resultLennart Poettering
2016-08-22bus-util: make sure map_basic() returns EOPNOTSUPP if called for an unknown typeLennart Poettering
Make sure we return proper errors for types not understood yet.
2016-08-22bus-util: treat an empty string as a NULLLennart Poettering
Instead of ignoring empty strings retrieved via the bus, treat them as NULL, as it's customary in systemd.
2016-08-22bus-util: support mapping signed integers with bus_map_properties()Lennart Poettering
Let's make sure we can read the exit code/status properties exposed by PID 1 properly. Let's reuse the existing code for unsigned fields, as we just use it to copy words around, and don't calculate it.
2016-08-22seccomp: make sure getrlimit() is among the default permitted syscallsLennart Poettering
A lot of basic code wants to know the stack size, and it is safe if they do, hence let's permit getrlimit() (but not setrlimit()) by default. See: #3970
2016-08-21shared/install: do not enable masked instances (#4005)Zbigniew Jędrzejewski-Szmek
When told to enable a template unit, and the DefaultInstance specified in that unit was masked, we would do this. Such a unit cannot be started or loaded, so reporting successful enabling is misleading and unexpected. $ systemctl mask getty@tty1 Created symlink /etc/systemd/system/getty@tty1.service → /dev/null. $ systemctl --root=/ enable getty@tty1 (unchanged) Failed to enable unit, unit /etc/systemd/system/getty@tty1.service is masked. $ systemctl --root=/ enable getty@ (before) Created symlink /etc/systemd/system/getty.target.wants/getty@tty1.service → /usr/lib/systemd/system/getty@.service. (now) Failed to enable unit, unit /etc/systemd/system/getty@tty1.service is masked. The same error is emitted for enable and preset. And an error is emmited, not a warning, so the failure to enable DefaultInstance is treated the same as if the instance was specified on the command line. I think that this makes most sense, for most template units. Fixes #2513.
2016-08-19Merge pull request #3909 from poettering/mount-toolEvgeny Vereshchagin
add a new tool for creating transient mount and automount units
2016-08-19Merge pull request #3955 from keszybz/fix-preset-allLennart Poettering
Fix preset-all
2016-08-19shared/install: properly report masked units listed in Also=Zbigniew Jędrzejewski-Szmek
A masked unit is listed in Also=: $ systemctl cat test1 test2 →# /etc/systemd/system/test1.service [Unit] Description=test service 1 [Service] Type=oneshot ExecStart=/usr/bin/true [Install] WantedBy=multi-user.target Also=test2.service Alias=alias1.service →# /dev/null $ systemctl --root=/ enable test1 (before) Created symlink /etc/systemd/system/alias1.service → /etc/systemd/system/test1.service. Created symlink /etc/systemd/system/multi-user.target.wants/test1.service → /etc/systemd/system/test1.service. The unit files have no installation config (WantedBy, RequiredBy, Also, Alias settings in the [Install] section, and DefaultInstance for template units). This means they are not meant to be enabled using systemctl. Possible reasons for having this kind of units are: 1) A unit may be statically enabled by being symlinked from another unit's .wants/ or .requires/ directory. 2) A unit's purpose may be to act as a helper for some other unit which has a requirement dependency on it. 3) A unit may be started when needed via activation (socket, path, timer, D-Bus, udev, scripted systemctl call, ...). 4) In case of template units, the unit is meant to be enabled with some instance name specified. (after) Created symlink /etc/systemd/system/alias1.service → /etc/systemd/system/test1.service. Created symlink /etc/systemd/system/multi-user.target.wants/test1.service → /etc/systemd/system/test1.service. Unit /etc/systemd/system/test2.service is masked, ignoring.
2016-08-19shared/install: when creating symlinks, keep existing relative symlinksZbigniew Jędrzejewski-Szmek
Running preset-all on a system installed from rpms or even created using make install would remove and recreate a lot of symlinks, changing relative to absolute symlinks. In general relative symlinks are nicer, so there is no reason to change them, and those spurious changes were obscuring more interesting stuff. $ make install DESTDIR=/var/tmp/inst1 $ systemctl preset-all --root=/var/tmp/inst1 (before) Removed /var/tmp/inst1/etc/systemd/system/network-online.target.wants/systemd-networkd-wait-online.service. Created symlink /var/tmp/inst1/etc/systemd/system/ctrl-alt-del.target → /usr/lib/systemd/system/exit.target. Removed /var/tmp/inst1/etc/systemd/system/multi-user.target.wants/remote-fs.target. Created symlink /var/tmp/inst1/etc/systemd/system/multi-user.target.wants/remote-fs.target → /usr/lib/systemd/system/remote-fs.target. Created symlink /var/tmp/inst1/etc/systemd/system/multi-user.target.wants/machines.target → /usr/lib/systemd/system/machines.target. Created symlink /var/tmp/inst1/etc/systemd/system/sockets.target.wants/systemd-journal-remote.socket → /usr/lib/systemd/system/systemd-journal-remote.socket. Removed /var/tmp/inst1/etc/systemd/system/sockets.target.wants/systemd-networkd.socket. Created symlink /var/tmp/inst1/etc/systemd/system/sockets.target.wants/systemd-networkd.socket → /usr/lib/systemd/system/systemd-networkd.socket. Removed /var/tmp/inst1/etc/systemd/system/getty.target.wants/getty@tty1.service. Created symlink /var/tmp/inst1/etc/systemd/system/getty.target.wants/getty@tty1.service → /usr/lib/systemd/system/getty@.service. Created symlink /var/tmp/inst1/etc/systemd/system/multi-user.target.wants/systemd-journal-upload.service → /usr/lib/systemd/system/systemd-journal-upload.service. Removed /var/tmp/inst1/etc/systemd/system/sysinit.target.wants/systemd-timesyncd.service. Created symlink /var/tmp/inst1/etc/systemd/system/sysinit.target.wants/systemd-timesyncd.service → /usr/lib/systemd/system/systemd-timesyncd.service. Removed /var/tmp/inst1/etc/systemd/system/multi-user.target.wants/systemd-resolved.service. Created symlink /var/tmp/inst1/etc/systemd/system/multi-user.target.wants/systemd-resolved.service → /usr/lib/systemd/system/systemd-resolved.service. Removed /var/tmp/inst1/etc/systemd/system/multi-user.target.wants/systemd-networkd.service. Created symlink /var/tmp/inst1/etc/systemd/system/multi-user.target.wants/systemd-networkd.service → /usr/lib/systemd/system/systemd-networkd.service. (after) Removed /var/tmp/inst1/etc/systemd/system/network-online.target.wants/systemd-networkd-wait-online.service. Created symlink /var/tmp/inst1/etc/systemd/system/ctrl-alt-del.target → /usr/lib/systemd/system/exit.target. Created symlink /var/tmp/inst1/etc/systemd/system/multi-user.target.wants/machines.target → /usr/lib/systemd/system/machines.target. Created symlink /var/tmp/inst1/etc/systemd/system/sockets.target.wants/systemd-journal-remote.socket → /usr/lib/systemd/system/systemd-journal-remote.socket. Created symlink /var/tmp/inst1/etc/systemd/system/multi-user.target.wants/systemd-journal-upload.service → /usr/lib/systemd/system/systemd-journal-upload.service.
2016-08-19shared/install: move root skipping into create_symlink()Zbigniew Jędrzejewski-Szmek
No functional change intended.
2016-08-19shared/install: ignore unit symlinks when doing preset-allZbigniew Jędrzejewski-Szmek
Before, when interating over unit files during preset-all, behaviour was the following: - if we hit the real unit name first, presets were queried for that name, and that unit was enabled or disabled accordingly, - if we hit an alias first (one of the symlinks chaining to the real unit), we checked the presets using the symlink name, and then proceeded to enable or disable the real unit. E.g. for systemd-networkd.service we have the alias dbus-org.freedesktop.network1.service (/usr/lib/systemd/system/dbus-org.freedesktop.network1.service), but the preset is only for the systemd-networkd.service name. The service would be enabled or disabled pseudorandomly depending on the order of iteration. For "preset", behaviour was analogous: preset on the alias name disabled the service (following the default disable policy), preset on the "real" name applied the presets. With the patch, for "preset" and "preset-all" we silently skip symlinks. This gives mostly the right behaviour, with the limitation that presets on aliases are ignored. I think that presets on aliases are not that common (at least my preset files on Fedora don't exhibit any such usage), and should not be necessary, since whoever installs the preset can just refer to the real unit file. It would be possible to overcome this limitation by gathering a list of names of a unit first, and then checking whether *any* of the names matches the presets list. That would require a significant redesign of the code, and be a lot slower (since we would have to fully read all unit directories to preset one unit) to so I'm not doing that for now. With this patch, two properties are satisfied: - preset-all and preset are idempotent, and the second and subsequent invocations do not produce any changes, - preset-all and preset for a specific name produce the same state for that unit. Fixes #3616.
2016-08-19shared/install: remove unused paramater and add more commentsZbigniew Jędrzejewski-Szmek
2016-08-19systemctl: fix preset-all with missing /etc/systemd/systemZbigniew Jędrzejewski-Szmek
If the directory is missing, we can assume that those pesky symlinks are gone too.
2016-08-19clean-ipc: debug log about all remove IPC objectsLennart Poettering
2016-08-19core: add RemoveIPC= settingLennart Poettering
This adds the boolean RemoveIPC= setting to service, socket, mount and swap units (i.e. all unit types that may invoke processes). if turned on, and the unit's user/group is not root, all IPC objects of the user/group are removed when the service is shut down. The life-cycle of the IPC objects is hence bound to the unit life-cycle. This is particularly relevant for units with dynamic users, as it is essential that no objects owned by the dynamic users survive the service exiting. In fact, this patch adds code to imply RemoveIPC= if DynamicUser= is set. In order to communicate the UID/GID of an executed process back to PID 1 this adds a new "user lookup" socket pair, that is inherited into the forked processes, and closed before the exec(). This is needed since we cannot do NSS from PID 1 due to deadlock risks, However need to know the used UID/GID in order to clean up IPC owned by it if the unit shuts down.
2016-08-18clean-ipc: shorten code a bitLennart Poettering
2016-08-18clean-ipc: don't filter out '.' and '..' twiceLennart Poettering
2016-08-18bus-util: unify loop around bus_append_unit_property_assignment()Lennart Poettering
This is done exactly the same way a couple of times at various places, let's unify this into one version.
2016-08-14Merge pull request #3905 from htejun/cgroup-v2-cpuZbigniew Jędrzejewski-Szmek
core: add cgroup CPU controller support on the unified hierarchy (zj: merging not squashing to make it clear against which upstream this patch was developed.)
2016-08-09install: follow config_path symlink (#3362)Rhys
Under NixOS, the config_path /etc/systemd/system is a symlink to /etc/static/systemd/system. Commands such as `systemctl list-unit-files` and `systemctl is-enabled` did not work as the symlink was not followed. This does not affect how symlinks are treated within the config_path directory.
2016-08-07core: add cgroup CPU controller support on the unified hierarchyTejun Heo
Unfortunately, due to the disagreements in the kernel development community, CPU controller cgroup v2 support has not been merged and enabling it requires applying two small out-of-tree kernel patches. The situation is explained in the following documentation. https://git.kernel.org/cgit/linux/kernel/git/tj/cgroup.git/tree/Documentation/cgroup-v2-cpu.txt?h=cgroup-v2-cpu While it isn't clear what will happen with CPU controller cgroup v2 support, there are critical features which are possible only on cgroup v2 such as buffered write control making cgroup v2 essential for a lot of workloads. This commit implements systemd CPU controller support on the unified hierarchy so that users who choose to deploy CPU controller cgroup v2 support can easily take advantage of it. On the unified hierarchy, "cpu.weight" knob replaces "cpu.shares" and "cpu.max" replaces "cpu.cfs_period_us" and "cpu.cfs_quota_us". [Startup]CPUWeight config options are added with the usual compat translation. CPU quota settings remain unchanged and apply to both legacy and unified hierarchies. v2: - Error in man page corrected. - CPU config application in cgroup_context_apply() refactored. - CPU accounting now works on unified hierarchy.