test: add a test case that validates cgroup delegation
This test runs on the unified hierarchy, and ensures that cgroup
delegation works properly, i.e. writ access is granted and the requested
controllers are enabled.
Make sure to add the delegation mask to the mask of controllers we have
to enable on our own unit. Do not claim it was a members mask, as such
a logic would mean we'd collide with cgroupv2's "no processes on inner
nodes policy".
This change does the right thing: it means any controller enabled
through Controllers= will be made available to subcrgoups of our unit,
but the unit itself has to still enable it through
cgroup.subtree_control (which it can since that file is delegated too)
to be inherited further down.
Or to say this differently: we only should manipulate
cgroup.subtree_control ourselves for inner nodes (i.e. slices), and
for leaves we need to provide a way to enable controllers in the slices
above, but stay away from the cgroup's own cgroup.subtree_control —
which is what this patch ensures.
cgroup: properly determine cgroups zombie processes belong to
When a process becomes a zombie its cgroup might be deleted. Let's add
some minimal code to detect cases like this, so that we can still
attribute this back to the original cgroup.
core: unify common code for preparing for forking off unit processes
This introduces a new function unit_prepare_exec() that encapsulates a
number of calls we do in preparation for spawning off some processes in
all our unit types that do so.
This allows us to neatly unify a bit of code between unit types and
shorten our code.
Let's rename get_controllers() → get_process_controllers(), in order to
underline the difference to cg_kernel_controllers(). After all, one
returns the controllers available to the process, the other the
controllers enabled in the kernel at all).
Let's also update the code to use read_line() and set_put_strdup() to
shorten the code a bit, and make it more robust.
nspawn: rework mount_systemd_cgroup_writable() a bit
We shouldn't call alloca() as part of function calls, that's not really
defined in C. Hence, let's first do our stack allocations, and then
invoke functions.
Also, some coding style fixes, and minor shuffling around.
Shawn Landden [Tue, 21 Nov 2017 07:24:12 +0000 (23:24 -0800)]
shared: silence gcc warning (#7402)
[346/1860] Compiling C object 'src/shared/systemd-shared-235@sha/firewall-util.c.o'.
../src/shared/firewall-util.c: In function ‘entry_fill_basics’:
../src/shared/firewall-util.c:81:79: warning: logical ‘and’ of equal expressions [-Wlogical-op]
[543/1860] Compiling C object 'src/shared/systemd-shared-235@sta/firewall-util.c.o'.
../src/shared/firewall-util.c: In function ‘entry_fill_basics’:
../src/shared/firewall-util.c:81:79: warning: logical ‘and’ of equal expressions [-Wlogical-op]
Susant Sahani [Mon, 20 Nov 2017 16:50:48 +0000 (22:20 +0530)]
networkd: Set RoutingPolicyRule in link_configure (#7235)
The RoutingPolicyRules are not added when we are calling from set_address
the link->message++ and link->message-- never reaches to zero in the callback function
resulting routes are never gets added.
meson: add -Wimplicit-fallthrough=3 to compilation options (#7393)
At some point before gcc-7 was released, -Wimplicit-fallthrough=3 was included
in -Wextra. The documentation for gcc-7.2.1-2.fc27.x86_64 still says that, but
empirical testing shows that it's not. The documentation also misstates that
-Wimplicit-fallthrough is equivalent to -Wimplicit-fallthrough=3.
Let's add -Wimplicit-fallthrough=3 explicitly to get the warnings if we regress.
This little new command can parse, validate, normalize calendar events,
and calculate when they will elapse next. This should be useful for
anyone writing calendar events and who'd like to validate the expression
before running them as timer units.
Yu Watanabe [Sun, 19 Nov 2017 16:00:34 +0000 (01:00 +0900)]
core/swap: load() should fail when neither of corresponding unit file nor /proc/swap entry does not exist
It is not necessary to label as loaded to a swap unit when neither of
corresponding unit file nor entry in /proc/swap does not exist.
This makes swap_load() to fail such a case.
Yu Watanabe [Sun, 19 Nov 2017 07:45:35 +0000 (16:45 +0900)]
core/automount: load() should fail when the unit file does not exist
It is not necessary to label as loaded to automount unit when its unit
file does not exist. So, let's make automount_load() to fail when the
unit file does not exist.
Add license headers and SPDX identifiers to meson.build files
So far I avoided adding license headers to meson files, but they are pretty
big and important and should carry license headers like everything else.
I added my own copyright, even though other people modified those files too.
But this is mostly symbolic, so I hope that's OK.
Add SPDX license identifiers to source files under the LGPL
This follows what the kernel is doing, c.f.
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=5fd54ace4721fc5ce2bb5aef6318fcf17f421460.
In principle we shouldn't merge this until after 4.15 is released, but the
chances of a revert upstream are low, and in that unlikely scenario we can just
revert this patch, it's a trivial documentation update after all.
systemctl: make sure the kernel is loaded before kexec'ing
We just load the same kernel that would be loaded by default by sd-boot, with
the same options. Changing the kernel or initramfs or options is left for
later.
Now we will refuse to continue if loading fails. This makes 'systemctl kexec'
more predictable: it will not fall back to normal reboot if the kernel is
not loaded.
core: never apply first boot presets in the initrd
Presets are useful to initialize uninitialized /etc, but that doesn't
apply to the initrd.
Also, let's rename etc_empty → first_boot. After all, the variable
doesn't actually reflect whether /etc is really empty, it just reflects
whether /etc/machine-id existed originally or not. Moreover, we later on
directly initialize manager_set_first_boot() from it, hence let's just
name it the same way all through the codepath, to make this all less
confusing.
This function is really not a method of the Manager object (implemented
in manager.c), but just a helper in main.c. Hence let's not confusingly
name it the way methods are called.
Already, path_is_safe() refused paths container the "." dir. Doing that
isn't strictly necessary to be "safe" by most definitions of the word.
But it is necessary in order to consider a path "normalized". Hence,
"path_is_safe()" is slightly misleading a name, but
"path_is_normalize()" is more descriptive, hence let's rename things
accordingly.
core: don't allow DefaultStandardOutput= be set to socket/fd:/file:
These three settings only make sense within the context of actual unit
files, hence filter this out when applied to the per-manager default,
and generate a log message about it.
core: add support for StandardInputFile= and friends
These new settings permit specifiying arbitrary paths as
stdin/stdout/stderr locations. We try to open/create them as necessary.
Some special magic is applied:
1) if the same path is specified for both input and output/stderr, we'll
open it only once O_RDWR, and duplicate them fd instead.
2) If we an AF_UNIX socket path is specified, we'll connect() to it,
rather than open() it. This allows invoking systemd services with
stdin/stdout/stderr connected to arbitrary foreign service sockets.
core: fold property_get_input_fdname() and property_get_output_fdname() into one
property_get_output_fdname() already had two different control flows for
stdout and stderr, it might as well handle stdin too, thus shortening
our code a bit.
projects / thirdparty / systemd.git / log
