summaryrefslogtreecommitdiff
path: root/src/login/logind-user.c
Commit message (Collapse)AuthorAge
* Prep v238: Do not kill user processes if not forced and the user is lingering.Sven Eden2018-06-29
|
* logind-user.c: As we need special.h elsewhere now, use it here, too.Sven Eden2018-06-08
| | | | (cherry picked from commit a96cb75c17b008312fca8f81bc2a126f8a87850c)
* logind: change check_gc to may_gc everywhereZbigniew Jędrzejewski-Szmek2018-05-30
|
* logind: use free_and_replace in one spotZbigniew Jędrzejewski-Szmek2018-05-30
| | | | No functional change.
* logind: fix misleading messageZbigniew Jędrzejewski-Szmek2018-05-30
| | | | | This message would also be emitted at boot for any user with linger enabled, so "logged in" is the wrong term to use.
* logind: simplify one conditionalZbigniew Jędrzejewski-Szmek2018-05-30
| | | | Don't bother with removing the directory if we didn't create it.
* Move /var/lib/systemd/linger to /var/lib/elogind/.Sven Eden2018-04-23
|
* Reverted accidential renaming of /run/systemd to /run/elogind. Applications ↵Sven Eden2018-04-23
| | | | using elogind as a drop-in replacement expect the first.
* Prep v236 : Add missing SPDX-License-Identifier (5/9) src/loginSven Eden2018-03-26
|
* user-util: add new uid_is_system() helperLennart Poettering2017-12-02
| | | | | | | This adds uid_is_system() and gid_is_system(), similar in style to uid_is_dynamic(). That a helper like this is useful is illustrated by the fact that test-condition.c didn't get the check right so far, which this patch fixes.
* logind: allow two very close error messages to be distinguishedZbigniew Jędrzejewski-Szmek2017-10-14
| | | | | | | | | | In https://bugzilla.redhat.com/show_bug.cgi?id=1486859 error messages appera: Sep 06 19:09:07 ld92.e.math.uh.edu audit[21482]: AVC avc: denied { read } for pid=21482 comm="elogind-logind" name="dbus-1" dev="tmpfs" ino=5548194 scontext=system_u:system_r:elogind_logind_t:s0 tcontext=unconfined_u:object_r:session_dbusd_tmp_t:s0 tclass=dir permissive=0 Sep 06 19:09:07 ld92.e.math.uh.edu elogind-logind[21482]: Failed to remove runtime directory /run/user/8664: Permission denied But it's not clear which of the two rm_rf's is the source. Let's make them different.
* tree-wide: use IN_SET macro (#6977)Yu Watanabe2017-12-08
|
* tree-wide: use `!IN_SET(..)` for `a != b && a != c && …`Andreas Rammhold2017-09-29
| | | | | | The included cocci was used to generate the changes. Thanks to @flo-wer for pointing this case out.
* Prep v235: Apply upstream fixes (6/10) [src/login]Sven Eden2017-08-14
|
* Prep v234: Re-add user/session to the gc_queue when stopping.Sven Eden2017-07-26
| | | | | | If manager_gc() does not find the user/session, their respective finalize_*() method is never called, leaving stale state files in /run/systemd/sessions and run/systemd/users.
* tree-wide: stop using canonicalize_file_name(), use chase_symlinks() insteadLennart Poettering2017-07-17
| | | | | | | | Let's use chase_symlinks() everywhere, and stop using GNU canonicalize_file_name() everywhere. For most cases this should not change behaviour, however increase exposure of our function to get better tested. Most importantly in a few cases (most notably nspawn) it can take the correct root directory into account when chasing symlinks.
* Rename formats-util.h to format-util.hZbigniew Jędrzejewski-Szmek2017-07-17
| | | | | | We don't have plural in the name of any other -util files and this inconsistency trips me up every time I try to type this file name from memory. "formats-util" is even hard to pronounce.
* logind: fix /run/user/$UID creation in apparmor-confined containers (#4154)Tomáš Janoušek2017-07-05
| | | | | | | | | | | | | | | | When a docker container is confined with AppArmor [1] and happens to run on top of a kernel that supports mount mediation [2], e.g. any Ubuntu kernel, mount(2) returns EACCES instead of EPERM. This then leads to: elogind-logind[33]: Failed to mount per-user tmpfs directory /run/user/1000: Permission denied login[42]: pam_elogind(login:session): Failed to create session: Access denied and user sessions don't start. This also applies to selinux that too returns EACCES on mount denial. [1] https://github.com/docker/docker/blob/master/docs/security/apparmor.md#understand-the-policies [2] http://bazaar.launchpad.net/~apparmor-dev/apparmor/master/view/head:/kernel-patches/4.7/0025-UBUNTU-SAUCE-apparmor-Add-the-ability-to-mediate-mou.patch
* logind: update empty and "infinity" handling for [User]TasksMax (#3835)Tejun Heo2017-07-05
| | | | | | | | | | | | | | | | | The parsing functions for [User]TasksMax were inconsistent. Empty string and "infinity" were interpreted as no limit for TasksMax but not accepted for UserTasksMax. Update them so that they're consistent with other knobs. * Empty string indicates the default value. * "infinity" indicates no limit. While at it, replace opencoded (uint64_t) -1 with CGROUP_LIMIT_MAX in TasksMax handling. v2: Update empty string to indicate the default value as suggested by Zbigniew Jędrzejewski-Szmek. v3: Fixed empty UserTasksMax handling.
* core: add RemoveIPC= settingLennart Poettering2017-07-05
| | | | | | | | | | | | | | | | | | 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.
* Prep v231.2: Minor changes to make future patching easier.Sven Eden2017-06-28
|
* logind: 0% and 100% should be valid for UserTasksMax (#3836)Tejun Heo2017-06-19
| | | | | | | config_parse_user_tasks_max() was incorrectly accepting percentage value between 1 and 99. Update it to accept 0% and 100%. This brings it in line with TasksMax handling in elogind. (cherry picked from commit cb3e4417590196bd30e1b8097348dca6ba34bd15)
* Use "return log_error_errno" in more places"Zbigniew Jędrzejewski-Szmek2017-06-16
|
* logind: change TasksMax= value for user logins to 33%Lennart Poettering2017-06-16
| | | | | | | | | | | | | | | | | | | | Let's change from a fixed value of 12288 tasks per user to a relative value of 33%, which with the kernel's default of 32768 translates to 10813. This is a slight decrease of the limit, for no other reason than "33%" sounding like a nice round number that is close enough to 12288 (which would translate to 37.5%). (Well, it also has the nice effect of still leaving a bit of room in the PID space if there are 3 cooperating evil users that try to consume all PIDs... Also, I like my bikesheds blue). Since the new value is taken relative, and machined's TasksMax= setting defaults to 16384, 33% inside of containers is usually equivalent to 5406, which should still be ample space. To summarize: | on the host | in the container old default | 12288 | 12288 new default | 10813 | 5406
* util: introduce physical_memory_scale() to unify how we scale by physical memoryLennart Poettering2017-06-16
| | | | | The various bits of code did the scaling all different, let's unify this, given that the code is not trivial.
* util-lib: introduce parse_percent() for parsing percent specificationsLennart Poettering2017-06-16
| | | | And port a couple of users over to it.
* Prep v230: Apply missing upstream fixes and updates (6/8) src/login.Sven Eden2017-06-16
|
* time-util: Rename and fix call of deserialize_timestamp_value()Benjamin Robin2017-06-16
| | | | | | The deserialize_timestamp_value() is renamed timestamp_deserialize() to be more consistent with dual_timestamp_deserialize() And add the NULL check back on realtime and monotonic
* logind: use deserialize_timestamp_value()Alexander Kuleshov2017-06-16
| | | | which is introduced in the ebf30a086d commit.
* Prep v229: Add missing fixes from upstream [4/6] src/loginSven Eden2017-05-17
|
* tree-wide: check if errno is greater than zero (2)Zbigniew Jędrzejewski-Szmek2017-05-17
| | | | | Compare errno with zero in a way that tells gcc that (if the condition is true) errno is positive.
* Prep v228: Silence uninitialized usage warnings.Sven Eden2017-04-26
|
* Prep v228: Condense elogind source masks (5/5)Sven Eden2017-04-26
|
* Prep v228: Add remaining updates from upstream (3/3)Sven Eden2017-04-26
| | | | | Apply remaining fixes and the performed move of utility functions into their own foo-util.[hc] files on the rest of elogind.
* login: fix re-use of usersDavid Herrmann2017-04-26
| | | | | | | | | | | | If the last reference to a user is released, we queue stop-jobs for the user-service and slice. Only once those are finished, we drop the user-object. However, if a new session is opened before the user object is fully dropped, we currently incorrectly re-use the object. This has the effect, that we get stale sessions without a valid "elogind --user" instance. Fix this by properly allowing user_start() to be called, even if user->stopping is true.
* [4/5] Apply missing fixes from upstreamSven Eden2017-03-29
|
* Unifiy free() usageSven Eden2017-03-14
| | | | | | | | This commit substitutes all occurrences of free(foo); foo = NULL; with foo = mfree(foo);
* user_start_slice() : The slice name must end with .slice:Sven Eden2017-03-14
|
* Classify processes from sessions into cgroupsSven Eden2017-03-14
| | | | | | | Create a private cgroup tree associated with no controllers, and use it to map PIDs to sessions. Since we use our own path structure, remove internal cgroup-related helpers that interpret the cgroup path structure to pull out users, slices, and scopes.
* Remove src/basic/special.h, as all defines in there are systemd-only.Sven Eden2017-03-14
|
* Remove dependency of systemd units, services and slices for new sessions.Sven Eden2017-03-14
| | | | | | | | | | | | | | | | | | | | * elogind does not support systemd services and units. But at least the units are needed to support the systemd cgroup slice/scope system. * Remove systemd subscription to scope, service and slice jobs. These can not be supported in any way, as they depend on systemd running the machine. * The functions session_start_scope(), user_start_service() and user_start_slice() no longer try to call systemd via dbus for assistance. This way they generate their proper scope, service and slice names, and store them in the Managers HashMaps for session and user units. This should enable us to reverse track pids to users and such stuff, as that is what systemd-logind does, not knowing whether any unit *really* has been started or not. However, this will not work out of the box until we find a way to integrate cg_create_everywhere() into elogind without becoming dependent of systemd unit, service and job knowledge again.
* Prep v225: Applying various fixes and changes to src/login that got lost ↵Sven Eden2017-03-14
| | | | during git am transfer.
* Prep v220: Use new cgroups functionsSven Eden2017-03-14
| | | | | Prep v220: Update logind and loginctl to upstream version. Prep v220: src/shared/rm-rf.c does not need to be able to handle btrfs subvolumes for elogind.
* Classify processes from sessions into cgroupsAndy Wingo2017-03-14
| | | | | | | Create a private cgroup tree associated with no controllers, and use it to map PIDs to sessions. Since we use our own path structure, remove internal cgroup-related helpers that interpret the cgroup path structure to pull out users, slices, and scopes.
* Remove systemd subscription and user/session unitsAndy Wingo2017-03-14
| | | | | | This removes attempts by logind to listen to systemd messages over the bus, and to start and/or manage units associated with sessions and users.
* Actually working with pamAndy Wingo2017-03-14
| | | | | | | | | | * src/login/logind-user.c (user_start): Don't start slices and systemd-user services. * src/login/logind-dbus.c (method_create_session): Send a reply directly instead of waiting on systemd that isn't there. * configure.ac: Bump version.
* logind: save /run/systemd/users/UID before starting user@.serviceSimon McVittie2017-03-14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously, this had a race condition during a user's first login. Some component calls CreateSession (most likely by a PAM service other than 'systemd-user' running pam_systemd), with the following results: - logind: * create the user's XDG_RUNTIME_DIR * tell pid 1 to create user-UID.slice * tell pid 1 to start user@UID.service Then these two processes race: - logind: * save information including XDG_RUNTIME_DIR to /run/systemd/users/UID - the subprocess of pid 1 responsible for user@service: * start a 'systemd-user' PAM session, which reads XDG_RUNTIME_DIR and puts it in the environment * run systemd --user, which requires XDG_RUNTIME_DIR in the environment If logind wins the race, which usually happens, everything is fine; but if the subprocesses of pid 1 win the race, which can happen under load, then systemd --user exits unsuccessfully. To avoid this race, we have to write out /run/systemd/users/UID even though the service has not "officially" started yet; previously this did an early-return without saving anything. Record its state as OPENING in this case. Bug: https://github.com/systemd/systemd/issues/232 Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
* logind: apply selinux label to XDG_RUNTIME_DIRLennart Poettering2017-03-14
| | | | | As discussed in #257: we should ensure the selinux label is correctly applied to each user's XDG_RUNTIME_DIR.
* everywhere: actually make use of DUAL_TIMESTAMP_NULL macroLennart Poettering2017-03-14
| | | | Let's use it as initializer where appropriate.
* login: fix potential null pointer dereferenceRonny Chevalier2017-03-14
| | | | | | | Fix CID 1304686: Dereference after null check (FORWARD_NULL) However, this commit does not fix any bug in logind. It helps to keep the elect_display_compare() function generic.
generated by cgit v1.2.3 (git 2.25.1) at 2024-06-20 16:06:54 +0000