Dan Streetman [Fri, 14 May 2021 12:08:33 +0000 (08:08 -0400)]
oom: log one-time warning if kernel doesn't provide memory.swap.current
The kernel can be compiled without support for any memory.swap.* files, or
it can be disabled at boot time with the 'swapaccount=0' boot parameter,
so if the file doesn't exist log warning indicating the kernel doesn't
support the file and the user may need to try using the 'swapaccount=1'
boot param.
Note that the actual error from the call to fopen() is ENOENT, but
that is translated into ENODATA in cg_get_attribute_as_uint64()
Dan Streetman [Wed, 19 May 2021 18:22:28 +0000 (14:22 -0400)]
psi: update is_pressure_supported to read file
The kernel still provides the /proc and cgroup pressure files even
if its psi support is disabled, so we need to actually read the files
to verify they don't return -EOPNOTSUPP
Dan Streetman [Wed, 19 May 2021 14:22:21 +0000 (10:22 -0400)]
log: add log_once() and log_once_errno() macros
These macros will log a message at the specified level only the first time
they are called. On all later calls, if the specified level is debug, the
logs will be suppressed; otherwise the message will be logged at debug.
This ensures that the fuzz test code is also built by default.
It also increases the test coverage a bit. Compiling the tests
*with* sanitizers is painfully slow, so this is not enabled. But
just compiling them sauté is hardly noticable. Running the tests
increases the test count and runtime:
622 tests, 26 s
to
922 tests, 35 s
I think this is acceptable.
Frantisek Sumsal [Wed, 19 May 2021 18:15:53 +0000 (20:15 +0200)]
ci: work around #19442 to make CI happy again
Let's introduce a somewhat ugly workaround for #19442 and retry
the systemd-nspawn image boot test up to three times in case it dies
with the dissect timeout. Since this issue occurs only in the Arch job,
limit the workaround to this job only.
Hardcoding major numbers sucks. And we generally don't do it, except
when determining whether something is a PTY. Thing though is that we
don't actually need to do that here either, hence don#t.
nspawn: add new --bind-user= option for binding a host user into the container
This new option does three things for a host user specified via
--bind-user=:
1. Bind mount the home directory from the host directory into
/run/host/home/<username>
2. Install an additional user namepace UID/GID mapping mapping the host
UID/GID of the host user to an unused one from the container in the range
60514…60577.
3. Synthesize a user/group record for the user/group under the same name
as on the host, with minimized information, and the UID/GID set to
the mapped UID/GID. This data is written to /run/host/userdb/ where
nss-system will pick it up.
This should make sharing users and home directories from host into the
container pretty seamless, under some conditions:
1. User namespacing must be used.
2. The host UID/GID of the user/group cannot be in the range assigned to
the container (kernel already refuses this, as this would mean two
host UIDs/GIDs might end up being mapped to the same continer
UID/GID.
3. There's a free UID/GID in the aforementioned range in the container,
and the name of the user/group is not used in the container.
4. Container payload is new enough to include an nss-systemd version
that picks up records from /run/host/userdb/
alloc-util: simplify GREEDY_REALLOC() logic by relying on malloc_usable_size()
We recently started making more use of malloc_usable_size() and rely on
it (see the string_erase() story). Given that we don't really support
sytems where malloc_usable_size() cannot be trusted beyond statistics
anyway, let's go fully in and rework GREEDY_REALLOC() on top of it:
instead of passing around and maintaining the currenly allocated size
everywhere, let's just derive it automatically from
malloc_usable_size().
I am mostly after this for the simplicity this brings. It also brings
minor efficiency improvements I guess, but things become so much nicer
to look at if we can avoid these allocation size variables everywhere.
Note that the malloc_usable_size() man page says relying on it wasn't
"good programming practice", but I think it does this for reasons that
don't apply here: the greedy realloc logic specifically doesn't rely on
the returned extra size, beyond the fact that it is equal or larger than
what was requested.
(This commit was supposed to be a quick patch btw, but apparently we use
the greedy realloc stuff quite a bit across the codebase, so this ends
up touching *a*lot* of code.)
It's a wrapper around malloc_usable_size() that is supposed to be
compatible with _FORTIFY_SOURCES=1, by taking the
__builtin_object_size() data into account, the same way as the
_FORTIFY_SOURCES=1 logic does.
docs: use {% raw %} to wrap jinja2 tags in documentation
As reported by @mrc0mmand:
> Since https://github.com/systemd/systemd/commit/89f52a780e54b2eb0905a6e613f6d4afcb22256b#diff-b842e6ab4a95a695d9449d106f091e6a134d9eac8d2aee1cd8b169fcb6b3a98bR109
> the GH pages fail to build, since they use the Liquid templating language,
> which coincidentally uses a very similar tags as jinja:
> https://shopify.github.io/liquid/tags/control-flow/
>
>> The tag elif on line 112 in HACKING.md is not a recognized Liquid tag.
mkosi/fedora: use pkgconfig virtual provides to refer to packages
... and /usr/bin/ path for a library package which provides an executable we
care about (libxslt).
This way the mkosi dependency list corresponds directly to the names which are
used in the dependency() and find_program() lines in meson.build. It also makes
the thing more resilient to package splits and renames.
network: show online state in networkctl status output
In case the link online state is invalid, networkctl will print
"unknown", which is sufficiently neutral. The same goes for the overall
manager online state if there are no managed links, or if
RequiredForOnline=no for all managed links.
Example output:
$ networkctl status
● State: routable
Online state: partial
Address: 172.22.0.130 on wlan0
...
man: clarify RequiredFamilyForOnline= behaviour for online state
With new "online state" semantics in networkd, make the description of
RequiredFamilyForOnline= a little more broad. Some rewording has been
done to make the passage easier to understand.
Alvin Šipraga [Tue, 30 Mar 2021 21:18:08 +0000 (23:18 +0200)]
network: use the overall online state in network_is_online()
Since networkd advertises a reliable online state, use it in
network_is_online(). If for some reason networkd does not know the
online state (e.g. it does not manage any of the network interfaces),
fall back to the original best-guess logic.
Alvin Šipraga [Tue, 30 Mar 2021 19:39:18 +0000 (21:39 +0200)]
network: add an online state for links and manager
Add a new state of type LinkOnlineState which indicates whether a link
is online or not. The state is also used by networkd's manager to expose
the overall online state of the system.
The possible states are:
offline the link (or system) is offline
partial at least one required link is online (see below)
online all required links are online
For links, a link is defined to be "online" if:
- it is managed; and
- its operational state is within the range defined by
RequiredForOnline=; and
- it has an IPv4 address if RequiredFamilyForOnline=ipv4 or =both; and
- it has an IPv6 address if RequiredFamilyForOnline=ipv6 or =both.
A link is defined to be "offline" if:
- it is managed; and
- it is not online, i.e. its operational state is not within the range
defined by RequiredForOnline=, and/or it is missing an IP address in
a required address family.
Otherwise, the link online state is undefined (represented internally as
_LINK_ONLINE_STATUS_INVALID or -EINVAL). Put another way, networkd will
only offer a meaningful online state for managed links where
RequiredForOnline=yes.
For the manager, the online state is a function of the online state of
all links which are requried for online, i.e. RequiredForOnline=yes. If
all required links are online, then the manager online state is defined
to be "online". If at least one of the required links is online, then
the manager online state is defined to be "partial". If none of
the required links are online, then the manager online state is defined
to be "offline". If there are no managed links, or RequiredForOnline=no
for all managed links, then the manager online state is undefined as
above.
The purpose of the "partial" state is analogous to the --any switch in
systemd-networkd-wait-online.service(8). For example, a required link
which lacks a carrier on boot will not force the overall (manager)
online state to "offline" if there is an alternative link available.
Recent meson versions include the directory name in the target name,
so there is no conflict for files with the same name in different
directories. But at least with meson-0.49.2 in buster we have conflict
with sysusers.d/systemd.conf.
This doesn't matter too much, but makes things a bit more consistent.
A minor advantage is that the file is not a configuration file for meson
anymore, so:
a) It is not built unless pulled in by another target. Since
we don't usually build man pages by default, this saves a tiny
amount of work.
b) When the .in file is updated, meson does not reconfigure everything,
but just rebuilds the dependent targets.
Now that the conversion is finished, time for benchmarking:
a full build with default settings (and -Dstandalonebinaries=true), yields
before this pull request: 1687 targets, 148.13s user 35.17s system 317% cpu 57.697 total
with the full pull request: 1714 targets, 143.07s user 27.87s system 314% cpu 54.369 total
The difference doesn't seem significant. Partial rebuilds might be faster as
mentioned before.
We had two big 'configuration_data' objects in meson config. (There are in fact
more. On is added in this series, and there's one for efi… But those others
have a handful variables only for specific purposes and don't matter). The two
sets are 'conf' and 'substs', and were inherited from the original autotools
system. In the past there was even a third set ('m4_defines'), but @yuwata
removed it in 348b44372f36010d48d9a7dda14ef67155753a71. And those two/three
systems had very similar data, but with different variable names, because of
historical reasons. They also used subtly different quoting (.set()
vs. .set10() vs. .set_quoted()), which was required because the templating
engines were not flexible enough. This meants we had more work when changing
things, and we needed to search for different variable names, etc.
With a more flexible templating engine we can do with just one
configuration_data object.
One stanza had "if install_sysconfdir_samples", while the other
"if install_sysconfdir", which looks like a mistake.
install_sysconfdir_samples is now used for both.
The naming of variables is very inconsistent. I tried to use more
modern style naming (UNDERSCORED_TITLE_CASE), but I didn't change existing
names too much. Only SYSTEM_DATA_UNIT_PATH is renamed to SYSTEM_DATA_UNIT_DIR
to match SYSTEM_CONFIG_UNIT_DIR.
meson: use conf configuration_data object to generate test-sysusers.sh
I wanted to use jinja2 templating here too, but it's hard to get right:
custom_target() strips the executable bit by default (unlike configure_file
apparently). custom_target() has install_mode setting, but it was only added
in meson-0.47, so it can't be used while we support 0.46. And without the
executable bit the test is not invoked properly. For example, "root-unittests"
in the debian package calls test-* after installation, so the executable bit
there is necessary. It would be possible to adjust the file mode after the
fact, but it would make things more complicated.
So let's use the native meson substitutions here. We don't need anything more
fancy.
meson: use custom configuration_data() object for man/man and man/html helpers
I want to stop using 'substs'. But in this case, configure_file() is nicer
than custom_target(), because it causes meson to immediately generate the
helpers after configuration, so it's possible to do
'meson build && build/man/man ...', without building anything first.
We only substitute one variable here, so let's use a custom configuration_data()
object.
m4 was hugely popular in the past, because autotools, automake, flex, bison and
many other things used it. But nowadays it much less popular, and might not even
be installed in the buildroot. (m4 is small, so it doesn't make a big difference.)
(FWIW, Fedora dropped make from the buildroot now,
https://fedoraproject.org/wiki/Changes/Remove_make_from_BuildRoot. I think it's
reasonable to assume that m4 will be dropped at some point too.)
The main reason to drop m4 is that the syntax is not very nice, and we should
minimize the number of different syntaxes that we use. We still have two
(configure_file() with @FOO@ and jinja2 templates with {{foo}} and the
pythonesque conditional expressions), but at least we don't need m4 (with
m4_dnl and `quotes').
Jinja2 inserts an empty line after the first macro body, which I don't know how
to get rid of. Only the first macro causes problems: the other ones don't have
conditional statements at the end and the issue does not occur. As a work-around
I moved ProtectHostname to the end of the first macro.
Output is identical, except for horizontal whitespace and change in position of
ProtectHostname.
HAVE_SMACK_RUN_LABEL was dropped back in 348b44372f36010d48d9a7dda14ef67155753a71,
so one line in etc.conf was not rendered as expected ;(
Checking if names are defined is paying for itself!
projects / thirdparty / systemd.git / log
