Lennart Poettering [Thu, 23 Nov 2017 16:39:53 +0000 (17:39 +0100)]
core: don't reference rescue/emergency targets in --user mode
They are only defined for system mode, hence let's not check for them in
--user mode.
Follow-up for #7433
Lennart Poettering [Thu, 23 Nov 2017 16:39:25 +0000 (17:39 +0100)]
core: minor simplification
Lennart Poettering [Thu, 23 Nov 2017 16:37:49 +0000 (17:37 +0100)]
path-lookup: LookupPathsFlags are a flags type, hence define it like one
We shouldn't rely on C's incremental assignment of values of enums for
bit fields. That'll work only between the first two flags, but for
everything following will break horrible. Hence, let's avoid any
ambiguity here, and let's clearly define the flags as shifts of 1.
Lennart Poettering [Thu, 23 Nov 2017 16:36:30 +0000 (17:36 +0100)]
systemctl: reword "needs-reload" message a bit
This message is displayed either when the unit file itself is newer than
what is loaded, but also when any of the drop-ins is newer. Say so in
the message, in order not to confuse the user unnecessarily.
Lennart Poettering [Wed, 22 Nov 2017 15:34:56 +0000 (16:34 +0100)]
core: allow changing the description for units dynamically at runtime
This makes "systemctl set-property foobar.service Description=waldo"
work.
See: #7245
Lennart Poettering [Wed, 22 Nov 2017 14:15:11 +0000 (15:15 +0100)]
tree-wide: generate SD_BUS_ERROR_INVALID_ARGS when we get invalid arguments on bus calls
Let's make sure that when we return a D-Bus error, we return a native
one, if we generate it ourselves, and use errno-based error
synthetization only if we received an errno ourselves. Yes, this makes
things slightly longer, but is highly misleading as we propagate D-Bus
errors, and not errnos to the client.
Lennart Poettering [Wed, 22 Nov 2017 14:03:51 +0000 (15:03 +0100)]
core: add proper escaping to writing of drop-ins/transient unit files
This majorly refactors the transient unit file and drop-in writing
logic, so that we properly C-escape and specifier-escape (% → %%)
everything we write out, so that when we read it back again, specifiers
are parsed that aren't supposed to be parsed.
This renames unit_write_drop_in() and friends by unit_write_setting().
The name change is supposed to clarify that the functions are not only
used to write drop-in files, but also transient unit files.
The previous "mode" parameter to this function is replaced by a more
generic "flags", which knows additional flags for implicit C-style and
specifier escaping before writing things out. This can cover most
properties where either form of escaping is defined. For the cases where
this isn't sufficient, we add helpers unit_escape_setting() and
unit_concat_strv() for escaping individual strings or strvs properly.
While we are at it, we also prettify generation of transient unit files:
we try to reduce the number of section headers written out: previously
we'd write the right section header our for each setting. With this
change we do so only if the setting lives in a different section than
the one before.
(This should also be considered preparation for when we add proper APIs
to systemd to write normal, persistant unit files through the bus API)
Lennart Poettering [Wed, 22 Nov 2017 11:10:39 +0000 (12:10 +0100)]
logind: fix sysfs change trigger code
We can't create files in sysfs, hence don't bother. Also if we ignore
the return value, do so explicitly by casting to void.
Lennart Poettering [Tue, 21 Nov 2017 19:09:31 +0000 (20:09 +0100)]
generators: be more careful when writing unit settings that support specifier expansion
Let's always escape strings we receive from the user before writing them
out to unit file settings that suppor specifier expansion, so that user
strings are transported as-is.
Lennart Poettering [Tue, 21 Nov 2017 19:03:51 +0000 (20:03 +0100)]
core: enable specifier expansion for What=/Where=/Type=/SourcePath= too
Using specifiers in these settings isn't particularly useful by itself,
but it unifies behaviour a bit. It's kinda surprising that What= in
mount units resolves specifies, but Where= does not. Hence let's add
that too. Also, it's surprising Where=/What= in mount units behaves
differently than in automount and swap units, hence resolve specifiers
there too. Then, Type= in mount units is nowadays an arbitrary,
sometimes non-trivial string (think fuse!), hence let's also expand
specifiers there, to match the rest of the mount settings.
This has the benefit that when writing code that generates unit files,
less care has to be taken to check whether escaping of specifiers is
necessary or not: broadly everything that takes arbitrary user strings
now does specifier expansion, while enums/numerics/booleans do not.
Lennart Poettering [Tue, 21 Nov 2017 18:50:52 +0000 (19:50 +0100)]
core: process C-style escapes in UnsetEnvironment=
We process C-style escapes in Environment=, hence we should process it
in UnsetEnvironment= too, as the latter accepts assignments much like
the former, including arbitrary values specified by the user.
Lennart Poettering [Tue, 21 Nov 2017 18:23:31 +0000 (19:23 +0100)]
core: use config_parse_unit_string_printf() for decoding RebootArgument=
All other cases where we accept a reboot argument are decoded with
config_parse_unit_string_printf() rather than
config_parse_unit_path_printf(), and that's really the only thing what
makes sense here, hence adjust this here, too.
Lennart Poettering [Tue, 21 Nov 2017 16:58:51 +0000 (17:58 +0100)]
CODING_STYLE: document the order in which unescaping + specifier expansion should be done
Lennart Poettering [Tue, 21 Nov 2017 16:57:56 +0000 (17:57 +0100)]
specifier: unify specifier implementations for user-related specifiers
The code in install-printf.c and unit-printf.c for these is pretty much
the same and very generic. Let's move this all over to the generic
specifier.c, and share the implementations.
Lennart Poettering [Tue, 21 Nov 2017 16:52:31 +0000 (17:52 +0100)]
specifier: add helper for escaping '%' characters to avoid making them subject for expansion
This is ultimately just a wrapper around strreplace(), but it makes
things a bit more self-descriptive.
Lennart Poettering [Tue, 21 Nov 2017 18:38:49 +0000 (19:38 +0100)]
string-util: update strreplace() a bit, use GREEDY_REALLOC()
Lennart Poettering [Tue, 21 Nov 2017 16:32:01 +0000 (17:32 +0100)]
core: drop specifier expansion when parsing transient dbus properties
Specifier expansion (much like C escape handling) should be a helper for
writing unit files, but should be nothing we do on programatic APIs. For
those, the client can do the necessary replacements anyway, and we
really should be careful with doing such string processing of data we
get via lower level programmatic APIs.
We currently do specifier expansion only for the env var transient unit
APIs, no other properties do this. Let's remove it here too, to be fully
systematic.
Yes, in a way this is API breakage, but then again this API isn't
documented yet, and an outlier, hence let's clear this up now, before it
is too late.
Yu Watanabe [Wed, 29 Nov 2017 03:50:49 +0000 (12:50 +0900)]
Merge pull request #7492 from keszybz/coverity-fixes
A few more cleanups based on coverity and valgrind
Zbigniew Jędrzejewski-Szmek [Tue, 28 Nov 2017 12:26:15 +0000 (13:26 +0100)]
udev/collect: use _cleanup_ in one more place
This doesn't make the code pretty, but it at least makes it a tiny bit shorter.
Zbigniew Jędrzejewski-Szmek [Tue, 28 Nov 2017 12:18:09 +0000 (13:18 +0100)]
udev/collect: remove dead conditional
Coverity rightly says that 'word' is assigned to 'ptr', and 'word'
was assigned from 'ptr' (yikes) right after a non-null check.
Zbigniew Jędrzejewski-Szmek [Tue, 28 Nov 2017 12:11:35 +0000 (13:11 +0100)]
journal: avoid undefined behaviour in float division by 0.0
Coverity says that's undefined. I'm pretty sure we always would get a nan, but
let's avoid (formally) undefined behaviour since that can cause compilers to do
strange things.
Zbigniew Jędrzejewski-Szmek [Tue, 28 Nov 2017 12:02:33 +0000 (13:02 +0100)]
networkd: fix memleak of Rules
This was the issue mentioned in
93f9da6e0a.
Zbigniew Jędrzejewski-Szmek [Tue, 28 Nov 2017 11:40:14 +0000 (12:40 +0100)]
journal: use new helpers with journal_file_close
journal_file_close_set() is not necessary anymore.
Zbigniew Jędrzejewski-Szmek [Tue, 28 Nov 2017 11:38:15 +0000 (12:38 +0100)]
journal-remote: add helper function and use hashmap_free_with_destructor
Zbigniew Jędrzejewski-Szmek [Tue, 28 Nov 2017 11:37:51 +0000 (12:37 +0100)]
cryptsetup-generator: add helper function and use hashmap_free_with_destructor
Zbigniew Jędrzejewski-Szmek [Tue, 28 Nov 2017 11:37:14 +0000 (12:37 +0100)]
coredump: fix typo and use hashmap_free_with_destructor
Zbigniew Jędrzejewski-Szmek [Tue, 28 Nov 2017 11:36:35 +0000 (12:36 +0100)]
machine-image: convert image_hashmap_free() to an static inline helper
Zbigniew Jędrzejewski-Szmek [Tue, 28 Nov 2017 11:35:49 +0000 (12:35 +0100)]
Add set/hashmap helpers for non-trivial freeing and use where straighforward
A macro is needed because otherwise we couldn't ensure type safety.
Some simple tests are included.
No functional change intended.
Zbigniew Jędrzejewski-Szmek [Tue, 28 Nov 2017 11:31:27 +0000 (12:31 +0100)]
Use hashmap_free_free where appropriate
Zbigniew Jędrzejewski-Szmek [Tue, 28 Nov 2017 17:15:30 +0000 (18:15 +0100)]
Merge pull request #7490 from yuwata/test-basic
add more tests for basic functions
Susant Sahani [Tue, 28 Nov 2017 16:24:52 +0000 (21:54 +0530)]
networkd: Add comment for setting_mtu #7460 (#7489)
Follow up #7460
Lennart Poettering [Tue, 28 Nov 2017 14:04:11 +0000 (15:04 +0100)]
Merge pull request #7486 from keszybz/coverity-fixes
Fixes for issues spotted by coverity (and some cleanups)
Yu Watanabe [Tue, 28 Nov 2017 13:17:00 +0000 (22:17 +0900)]
test-cap-list: add more tests about capability_set_{from,to}_string()
Yu Watanabe [Tue, 28 Nov 2017 13:14:54 +0000 (22:14 +0900)]
test: move tests about cap_list_cap() from test-cap-list to test-capability
Yu Watanabe [Tue, 28 Nov 2017 13:06:34 +0000 (22:06 +0900)]
cap-list: check range of numeric value
Yu Watanabe [Tue, 28 Nov 2017 09:19:04 +0000 (18:19 +0900)]
test-bitmap: add tests for bitmap_copy()
Yu Watanabe [Tue, 28 Nov 2017 09:14:55 +0000 (18:14 +0900)]
test-architecture: add more tests
Yu Watanabe [Tue, 28 Nov 2017 11:33:58 +0000 (20:33 +0900)]
test-alloc-util: add tests for memdup_suffix0() and greedy_realloc()
Yu Watanabe [Tue, 28 Nov 2017 09:13:05 +0000 (18:13 +0900)]
tests: add tests for empty string
Yu Watanabe [Tue, 28 Nov 2017 09:11:58 +0000 (18:11 +0900)]
alloc-util: coding style fix
Lennart Poettering [Tue, 28 Nov 2017 10:35:02 +0000 (11:35 +0100)]
Merge pull request #7398 from keszybz/udev-list
udev: use list.h instead of udev_list_node
Lennart Poettering [Tue, 28 Nov 2017 10:33:38 +0000 (11:33 +0100)]
Merge pull request #7484 from poettering/sd-resolve-timeout
timeout sd-resolve tests
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 22:05:19 +0000 (22:05 +0000)]
pid1: improve syntax in error message
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 22:01:48 +0000 (22:01 +0000)]
pid1: clarify that min_max is always initialized
read_one_line_file() always returns <= 0, so the code was OK, but let's write
the check a bit differently to make it obvious that min_max is always set.
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 21:54:38 +0000 (21:54 +0000)]
test-hostname-util: add assert_se's to make coverity happy
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 21:52:32 +0000 (21:52 +0000)]
test-process-util: add (void) cast to make coverity happy
The intent is for the call to succeed only when privileged, so make
that clear.
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 21:51:11 +0000 (21:51 +0000)]
test-fd-util: add assert_se to make coverity happy
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 21:48:46 +0000 (21:48 +0000)]
udevadm-monitor: add some assert_se's to make coverity happy
We do that elsewhere and it's nice to be consistent.
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 16:13:04 +0000 (16:13 +0000)]
core/umount: use _cleanup_ in one more place
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 13:54:55 +0000 (13:54 +0000)]
sd-resolve: check that name fits in the specified packet length
Coverity complained that we didn't check if the data is long enough.
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 13:31:46 +0000 (13:31 +0000)]
shim: fix logically dead code
The alternate verification path could never be taken.
Found by coverity.
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 13:25:08 +0000 (13:25 +0000)]
networkd: remove dead code
Found by coverity.
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 12:47:52 +0000 (12:47 +0000)]
test-ndisc-ra: add assert_se() to make coverity happy
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 12:36:21 +0000 (12:36 +0000)]
boot/shim: fix char ** vs char * mismatch
This code cannot have ever worked ;(
Found by coverity, obvious when the type is not obfuscated to void*.
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 12:22:56 +0000 (12:22 +0000)]
boot/efi: fix NULL dereference
The comment above makes the intent of the code pretty clear:
"use security2_protocol == NULL as indicator".
So revert the condition in the check and fix the logic in the comment while
at it.
The question is how this could have ever worked: if BS->LocateProtocol
(which is supposedly optional) ever failed, we'd crash here. Strange.
Found by coverity.
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 12:16:09 +0000 (12:16 +0000)]
dissect-image: simplify check for passphrase presence
m->encrypted is set when fstype=="crypto_LUKS", but this is not obvious when
reading decrypt_partition(). Just check if passphrase is set before using
it.
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 12:08:37 +0000 (12:08 +0000)]
networkd: fix reversed arguments
Found by coverity.
Also drop unused return value while at it.
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 12:01:41 +0000 (12:01 +0000)]
networkd: plug some memleaks in rule serialization/deserialization
This fixes the (mostly theoretical, since we're only parsing data that we write
ourselves) memleak when iif or oif is deserialized multiple times. Unfortunately
it does not fix the memleak when rule is freed, but that'll require a bigger
effort.
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 11:51:08 +0000 (11:51 +0000)]
networkd: fix serialization of {Incoming,Outgoing}Interface
Let's just say that the code wasn't fully functional ;(
Since we only had the parser for serialization, and not the writer, we are
free to change the format. So while at it, let's use shorter names in the
serialization format that match the surrounding style.
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 11:38:24 +0000 (11:38 +0000)]
Add test-routing-policy-rule
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 11:23:31 +0000 (11:23 +0000)]
test-network: use automatic cleanup
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 11:21:19 +0000 (11:21 +0000)]
networkd: parse the rvalue, not the option name
:(
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 10:22:51 +0000 (10:22 +0000)]
networkd: move rule loading to a separate function
No functional change.
Zbigniew Jędrzejewski-Szmek [Tue, 28 Nov 2017 08:15:00 +0000 (09:15 +0100)]
Merge pull request #7316 from poettering/fd-store-remove
add new FDSTOREREMOVE=1 sd_notify() message
Ari [Tue, 28 Nov 2017 02:30:31 +0000 (21:30 -0500)]
Added Logitech MX Master 2s (#7487)
Lennart Poettering [Mon, 27 Nov 2017 15:29:41 +0000 (16:29 +0100)]
test-resolve: make sure we don't hang forever
If test-resolve is running in some CI environment that doesn't have
reliably working DNS, it's a good idea not to hang forever, let's
time-out after 20s
Lennart Poettering [Mon, 27 Nov 2017 15:28:53 +0000 (16:28 +0100)]
sd-resolve: propagate timeouts in sd_resolve_wait() the same way as in sd_bus_wait(): ETIMEDOUT
Thankfully this is an internal API still, so we can mkae changes like
this.
Lennart Poettering [Mon, 27 Nov 2017 15:58:46 +0000 (16:58 +0100)]
service: shortcut operations if the MAINPID= doesn't actually cause a change
Lennart Poettering [Mon, 27 Nov 2017 15:53:03 +0000 (16:53 +0100)]
service: use parse_errno() for parsing error numbers
Let's always use the same logic when parsing error numbers, i.e. use
parse_errno() here too, to unify some code, and tighten the checks a
bit.
This also allows clients to pass errors as symbolic names. Probably
nothing we want to advertise too eagerly (since new daemons generating
this on old service managers won't understand), but still pretty
useful I think, in particular in scripting languages and such, where the
numeric error numbers might not be readily available.
Lennart Poettering [Mon, 13 Nov 2017 18:16:56 +0000 (19:16 +0100)]
update TODO
Lennart Poettering [Mon, 13 Nov 2017 18:06:11 +0000 (19:06 +0100)]
logind: use the new FDSTOREREMOVE=1 sd_notify() message
Let's explicitly tell PID 1 that we don't need an fd anymore, instead of
relying exclusively on POLLERR/POLLHUP for it to be removed.
Fixes: #6908
Lennart Poettering [Mon, 13 Nov 2017 17:26:04 +0000 (18:26 +0100)]
core: add a new sd_notify() message for removing fds from the FD store again
Currenly the only way to remove fds from the fdstore is to fully
stop the service, or to somehow trigger POLLERR/POLLHUP on the fd, in
which case systemd will remove the fd automatically.
Let's add another way: a new message that can be sent to remove fds
explicitly, given their name.
Lennart Poettering [Mon, 13 Nov 2017 17:19:53 +0000 (18:19 +0100)]
man: slightly improve the sd_notify() documentation regarding READY=1
READY=1 may be used to signal when a service finished startup, but also
when it finished reloading. Say so.
Lennart Poettering [Mon, 13 Nov 2017 17:14:20 +0000 (18:14 +0100)]
core: only process one of READY=1, STOPPING=1 or RELOADING=1 in sd_notify() handling
Of course, it's not really a valid sd_notify() message if multiple of
these fields are used in one, but let's handle this somewhat gracefully,
by only processing one of them, and ignoring the rest.
Lennart Poettering [Mon, 13 Nov 2017 17:06:09 +0000 (18:06 +0100)]
service: reorder sd_notify() handling a bit
Let's keep handling of WATCHDOG= and WATCHDOG_USEC= together. No
functional changes.
Lennart Poettering [Mon, 13 Nov 2017 16:48:50 +0000 (17:48 +0100)]
service: split out sd_notify() message authorization code into a function of its own
Let's shorten service_notify_message() a bit, and do the authentication
outside of the main function body.
No functional changes.
Lennart Poettering [Mon, 13 Nov 2017 16:44:00 +0000 (17:44 +0100)]
core: only log about sd_notify() message contents, when debug logging is on
Let's optimize things a bit for the non-debug case. No change in
behaviour.
Main reason to do this is not so much the speed benefit though, but
merely to isolate the code from its surroundings more.
Vladislav Vishnyakov [Mon, 27 Nov 2017 15:32:07 +0000 (22:32 +0700)]
Corrected Chuwi HiBook entry (#7483)
Patrik Flykt [Mon, 27 Nov 2017 12:07:22 +0000 (14:07 +0200)]
networkd: Free Prefix in case allocation fails (#7480)
Set the code to call _cleanup_prefix_free_ in case the next allocation
fails so that Prefix does not leak memory.
Reported by evverx.
Zbigniew Jędrzejewski-Szmek [Mon, 27 Nov 2017 10:06:55 +0000 (10:06 +0000)]
networkd: unbork rule serialization
We'd write something like
RULE=from=250.250.240.0/24 to=(null)/0 tos=0 fwmark=0/0 table=12
RULE=from=250.250.250.0/24 to=(null)/0 tos=0 fwmark=0/0 table=11
Lennart Poettering [Mon, 27 Nov 2017 07:15:07 +0000 (08:15 +0100)]
logind: make sure we don't acces m->action_what if it's not initialized (#7475)
Fixes: #7466
Lennart Poettering [Sun, 26 Nov 2017 17:37:00 +0000 (18:37 +0100)]
Merge pull request #7472 from keszybz/tests-on-a-slow-disk
Tests on a slow disk
Susant Sahani [Sun, 26 Nov 2017 14:21:45 +0000 (19:51 +0530)]
networkd: DHCP client do not get into a loop while setting MTU (#7460)
Some devices get reset itself while setting the MTU. we get in to a LOOP .
Once the MTU changed then the DHCP client talking with DHCP server never stops.
networkd gets into a loop and generates endless DHCP requests.
fixes #6593
fixes #7380
Zbigniew Jędrzejewski-Szmek [Sun, 26 Nov 2017 12:20:08 +0000 (13:20 +0100)]
test: increase the timeout for hwdb-test
This test does a lot of mmap/madvise/unmmap, which is slow under address sanitizer.
Just increase the timeout to avoid spurious failure.
Zbigniew Jędrzejewski-Szmek [Sun, 26 Nov 2017 12:18:45 +0000 (13:18 +0100)]
test: increase timeout for test-journal-enum
This reads the journal from the host system, and in case of a slow
disk, the 30s default can be easily exceeded.
Evgeny Vereshchagin [Sun, 26 Nov 2017 11:46:56 +0000 (14:46 +0300)]
exec-util: initialize `new` before using it (#7471)
CID #1383004
Zbigniew Jędrzejewski-Szmek [Sun, 26 Nov 2017 10:34:24 +0000 (11:34 +0100)]
Merge pull request #7464 from poettering/cgroup-control-fix
various cgroup fixes
Yu Watanabe [Sun, 26 Nov 2017 03:35:07 +0000 (12:35 +0900)]
Merge pull request #7468 from yuwata/fix-7362
test and meson cleanups
Zbigniew Jędrzejewski-Szmek [Sat, 25 Nov 2017 22:13:12 +0000 (23:13 +0100)]
Merge pull request #7467 from yuwata/fix-7302
nspawn: adjust path to static resolv.conf to support split usr
Yu Watanabe [Sat, 25 Nov 2017 17:40:23 +0000 (02:40 +0900)]
fileio: include sys/mman.h
Yu Watanabe [Sat, 25 Nov 2017 17:17:06 +0000 (02:17 +0900)]
meson: update header file to detect memfd_create()
Lennart Poettering [Fri, 24 Nov 2017 17:29:29 +0000 (18:29 +0100)]
update TODO
Lennart Poettering [Fri, 24 Nov 2017 21:02:22 +0000 (22:02 +0100)]
core: warn about left-over processes in cgroup on unit start
Now that we don't kill control processes anymore, let's at least warn
about any processes left-over in the unit cgroup at the moment of
starting the unit.
Lennart Poettering [Fri, 24 Nov 2017 19:27:01 +0000 (20:27 +0100)]
core: generalize the cgroup empty check on GC
Let's move the cgroup empty check for all unit types into the generic
unit_check_gc() call, out of the per-unit-type _check_gc() type. This
not only allows us to share some code, but also hooks up mount and
socket units with this kind of check, for free, as it was missing there
previously.
Lennart Poettering [Fri, 24 Nov 2017 19:11:27 +0000 (20:11 +0100)]
cgroup: remove logic for maintaining /control subcgroup for the service unit type
Previously, in the service unit type we ran all control processes in a
special subcgroup /control of the unit's main cgroup. Remove that, and
run the control program in the main cgroup instead.
The concept conflicts with cgroupv2's logic of "no processes in inner
nodes": if a unit has a main daemon process running in the main cgroup,
and a reload control process would be started in the /control subcgroup,
then this would necessarily fail, as the main daemon process would
become an inner node process that way.
We could in theory continue to support this in cgroupv1, but in the
interest in keeping behaviour similar in both hierarchies, let's drop
this altogether.
Philosophically maybe it wasn't the greatest idea anyway to just go
berserk and SIGKILL all those processes — loud warning logging might
have sufficed, too.
Lennart Poettering [Fri, 24 Nov 2017 18:51:36 +0000 (19:51 +0100)]
unit: initialize bpf cgroup realization state properly
Before this patch, the bpf cgroup realization state was implicitly set
to "NO", meaning that the bpf configuration was realized but was turned
off. That means invalidation requests for the bpf stuff (which we issue
in blanket fashion when doing a daemon reload) would actually later
result in a us re-realizing the unit, under the assumption it was
already realized once, even though in reality it never was realized
before.
This had the effect that after each daemon-reload we'd end up realizing
*all* defined units, even the unloaded ones, populating cgroupfs with
lots of unneeded empty cgroups.
With this fix we properly set the realiazation state to "INVALIDATED",
i.e. indicating the bpf stuff was never set up for the unit, and hence
when we try to invalidate it later we won't do anything.
Lennart Poettering [Fri, 24 Nov 2017 18:48:38 +0000 (19:48 +0100)]
cgroup: when dispatching the cgroup realization queue, check again if we shall actually realize
We add units to the cgroup realization queue when propagating realizing
requests to sibling units, and when invalidating cgroup settings because
some cgroup setting changed. In the time between where we add the unit
to the queue until the cgroup is actually dispatched the unit's state
might have changed however, so that the unit doesn't actually need to be
realized anymore, for example because the unit went down. To handle
that, check the unit state again, if realization makes sense.
Redundant realization is usually not a problem, except when the unit is
not actually running, hence check exactly for that.
Lennart Poettering [Fri, 24 Nov 2017 18:37:01 +0000 (19:37 +0100)]
cgroup: drop unused parameter from function
Lennart Poettering [Fri, 24 Nov 2017 17:30:23 +0000 (18:30 +0100)]
cgroup-util: merge cg_set_tasks_access() and cg-set_group_access() into one
We never use these functions seperately, hence don't bother splitting
them into to.
Also, simplify things a bit, and maintain tables for the attribute files
to chown. Let's also update those tables a bit, and include thenew
"cgroup.threads" file in it, that needs to be delegated too, according
to the documentation.
Lennart Poettering [Fri, 24 Nov 2017 17:29:35 +0000 (18:29 +0100)]
mkosi: let's switch to the unified mode for mkosi
This is a toolconfiguration for developers, and hence most likely should
be the first thing to be switched over. Do so.