| Age | Commit message (Collapse) | Author |
|---|
|
When (for example) switching from X11 to a new VT and logging in there,
creating a new session, the user state file (/run/systemd/users/$uid) is
not updated after the session becomes active. The latest time it is
saved is when the session is in SESSION_OPENING.
This results in a /run/systemd/users/$uid file which contains
STATE=online for the current user on the current active VT, which is
obviously wrong.
As functions like sd_uid_get_state() use this file to get the user’s
state, this could result in things like PolicyKit making incorrect
decisions about the user’s state. (See
https://bugs.freedesktop.org/show_bug.cgi?id=76358.)
Fix this by re-saving the state for a session’s user after completing
the state_job for that session.
https://bugs.freedesktop.org/show_bug.cgi?id=90818
|
|
Add a regression test for the recent breakage of handling improperly
escaped exec strings in unit files.
Code contributed by Martin Pitt:
https://bugs.freedesktop.org/show_bug.cgi?id=90794
|
|
The cunescape() helper function used to handle unknown escaping sequences
gracefully by copying them over verbatim.
Commit 527b7a42 ("util: rework cunescape(), improve error handling") added
a flag to make that behavior optional, and changed to default to error out
with -EINVAL otherwise.
However, config_parse_exec(), which is used to parse the
Exec{Start,Stop}{Post,Pre,} directives of unit files, was not changed along
with that commit, which means that directives with improperly escaped
command line strings are no longer parsed.
Relevant bugreports include:
https://bugs.freedesktop.org/show_bug.cgi?id=90794
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=787256
Fix this by passing UNESCAPE_RELAX to config_parse_exec() in order to
restore the original behavior.
|
|
Don't clobber the sd_device struct, and don't leak memory when memory allocation fails.
|
|
The command is 'install', not 'instal'. Fix that typo.
|
|
In device_update_properties_bufs(), the strv is built from pointers into the
single nul-terminated buf_nulstr string, to avoid allocating the key=value
strings twice. However, we must not do that while building and
GREEDY_REALLOC0()'ing buf_nulstr, as each time when this actually reallocates
memory the pointers we wrote into buf_strv so far become invalid.
So change the logic to first completely build the new buf_nulstr, and then
iterate over it to pick out the pointers to the individual key=value strings
for properties_strv.
This fixes invalid environment for udev callouts.
|
|
https://bugs.debian.org/787367
|
|
|
|
The communication channels must all be opened before forknig in daemon mode,
or we cannot guarantee that udevadm will work correctly as soon as udevd is
started.
|
|
In daemon mode we would break sd-event as it cannot work accross different processes.
Simply delay the allocation to after the fork.
|
|
Kay said: 'it is from ancient times, when we started udevd from the
kernel's usermodhelper which had no fd 0,1,2'.
|
|
A small typo in ee818b8 caused /home to be put in read-only instead of
/usr when ProtectSystem was enabled (ie: not set to "no").
|
|
Instead of this:
[filename:1] Failed to parse nsec_t value, ignoring: garbage
we show this:
[filename:1] Failed to parse nsec value, ignoring: garbage
|
|
We ignore the return value of sd_device_get_devtype, then devtype could
be uninitialized when used with streq_ptr. So we need to initialize it
first.
|
|
|
|
|
|
add tests for the following functions:
- fdset_new_array
- fdset_steal_first
- fdset_isempty
|
|
|
|
When processing an event, the watch is disabled, make sure it is restorted after
a CHANGE event has been processed.
|
|
The intention was to turn this rule from using a blacklist to a whitelist, but
there was a stray '!'.
|
|
This reverts commit 6096d9cc. As discussed on the mailing list, we
should accept some formal incorrectness in the dependency here, and
not rebuild the man pages every time Makefile.am changes - xsltproc
is simply too expensive.
Instead, let's move man/custom-entities.ent from DISTCLEANFILES to
CLEANFILES, so a 'make clean' is sufficient to actually make changes
in Makefile.am efficient for the contents of the man pages.
|
|
|
|
Missed this one from the previous commit
|
|
|
|
No functional changes.
|
|
|
|
We are already sending watchdog notification, this tells PID1 to actually listen for
them and restart udevd in case it gets stuck.
|
|
This notifies PID1 about config being flushed, about shutdown starting and shutdown finalizing.
|
|
Only unset the env var in the workers, but otherwise keep it around in the main daemon.
|
|
Only log about starting in daemon mode, rely on PID1 to log this in notify mode. Also
explicitly set the STATUS variable, as is done in notify mode as is done for other
serivecs.
|
|
This allows us to drop the special sigterm handling in spawn_wait()
as this will now be passed directly to the worker event loop.
We now log failing spawend processes at 'warning' level, and timeouts
are in terms of CLOCK_BOOTTIME when available, otherwise the behavior
is unchanged.
|
|
|
|
Rather than trying to schedule new events on every main-loop iteration, do it explicitly when
processing an event finishes, a worker is killed, a new uevent is received, or the event queue
is explicitly restarted.
|
|
Also move builtin and rules initialization from main loop to
event_queue_start().
No functional change.
|
|
The behavior is mostly unchanged, but rather than only ever calling these functions at
fixed points in the event loop, they are called directly whenever they are invoked.
|
|
This makes path_is_mount_point() consistent with fd_is_mount_point() wrt.
flags.
|
|
Commits 27cc6f166 and f25afeb broke path_is_mount_point() for files (such as
/etc/machine-id → /run/machine-id bind mounts) as with the factorization of
fd_is_mount_point() we lost the parent directory. We cannot determine that from
an fd only as openat(fd, "..") only works for directory fds.
Change fd_is_mount_point() to behave like openat(): It now takes a file
descriptor of the containing directory, a file name in it, and flags (which can
be 0 or AT_SYMLINK_FOLLOW). Unlike name_to_handle_at() or openat(), fstatat()
only accepts the inverse flag AT_SYMLINK_NOFOLLOW and complains with EINVAL
about AT_SYMLINK_FOLLOW; so we need to transform the flags for that fallback.
Adjust rm_rf_children() accordingly (only other caller of fd_is_mount_point()
aside from path_is_mount_point()).
Add test cases for files, links, and file bind mounts (the latter will only
work when running as root). Split out a new test_path_is_mount_point() test
case function as it got significantly larger now.
|
|
Previously we always ran distcheck with --disable-split-usr. This caused
test-path-util to fail with
Assertion 'fsck_exists("minix") == 0' failed at ../src/test/test-path-util.c:224, function test_fsck_exists(). Aborting.
as looking up fsck.minix would only look into DEFAULT_PATH_NORMAL, but on these
systems fsck is in /sbin/.
|
|
There is nothing like systemd_verify_* in Makefile.am. The bug has
been invisible because automake uses the default CFLAGS when component
CFLAGS are undefined.
|
|
Consistently move EXTRA_DIST out of conditional blocks. This would have
produced incomplete dist tarballs when being run in a built tree with not
every feature enabled, which can cause broken dist tarballs.
|
|
When Makefile.am is modified, make sure custom-entities.ent is rebuilt.
After all, $(substitutions) is defined there, so changes of that variable
must be reflected in the resulting file.
|
|
If there is both a SysV init.d script and a systemd unit for a given name, we
want to do the same enable/disable operation for both, instead of just on the
systemd unit. This keeps the enablement status in sync so that switching init
systems behaves as expected.
|
|
Introduce /usr/lib/systemd/systemd-sysv-install [--root=] <action> <name>
abstraction, replacing the direct calling of chkconfig. This allows
distributions to call their specific tools like update-rc.d without patching
systemd.
Ship systemd-sysv-install.SKELETON as an example for packagers how to implement
this.
Drop the --enable-chkconfig configure option.
Document this in README and point to it in NEWS.
|
|
Export the MOUNT_PATH and UMOUNT_PATH variables as XML entities and use them in
the systemctl.1 manpage instead of hardcoding the path in /usr/bin.
Tested:
- Ran ./configure ac_cv_path_MOUNT_PATH=/bin/mount (same for umount) and
rebuilt the manpages, confirmed that the correct path was in man/systemctl.1
- Rebuilt man/systemd.directives.xml and the man pages derived from it,
confirmed that the correct paths were there as well.
|
|
In particular, use /lib/systemd instead of /usr/lib/systemd in distributions
like Debian which still have not adopted a /usr merge setup.
Use XML entities from man/custom-entities.ent to replace configured paths while
doing XSLT processing of the original XML files. There was precedent of some
files (such as systemd.generator.xml) which were already using this approach.
This addresses most of the (manual) fixes from this patch:
http://anonscm.debian.org/cgit/pkg-systemd/systemd.git/tree/debian/patches/Fix-paths-in-man-pages.patch?h=experimental-220
The idea of using generic XML entities was presented here:
http://lists.freedesktop.org/archives/systemd-devel/2015-May/032240.html
This patch solves almost all the issues, with the exception of:
- Path to /bin/mount and /bin/umount.
- Generic statements about preference of /lib over /etc.
These will be handled separately by follow up patches.
Tested:
- With default configure settings, ran "make install" to two separate
directories and compared the output to confirm they matched exactly.
- Used a set of configure flags including $CONFFLAGS from Debian:
http://anonscm.debian.org/cgit/pkg-systemd/systemd.git/tree/debian/rules
Installed the tree and confirmed the paths use /lib/systemd instead of
/usr/lib/systemd and that no other unexpected differences exist.
- Confirmed that `make distcheck` still passes.
|
|
This was a regression that broke
$ udevadm trigger -nv --property-match=DEVNAME=/dev/sda1 --attr-match=size=409600
Reported by David Reisner.
|
|
|
|
|
|
A NULL pointer was inserted as the first element of the strv.
This had the effect of always passing the empty environment to processes
spawned by udev.
Reported by Michał Bartoszkiewicz.
|
|
We were listening for SIGCHLD in the wrong process.
|
