]> git.ipfire.org Git - thirdparty/systemd.git/log
thirdparty/systemd.git
2 years agodocs: adapt forward layout to systemd.io design 23460/head
Benjamin Franzke [Sat, 21 May 2022 11:08:19 +0000 (13:08 +0200)] 
docs: adapt forward layout to systemd.io design

2 years agodocs: rename COREDUMP_PACKAGE_METADATA → ELF_PACKAGE_METADATA
Zbigniew Jędrzejewski-Szmek [Sat, 21 May 2022 16:07:38 +0000 (18:07 +0200)] 
docs: rename COREDUMP_PACKAGE_METADATA → ELF_PACKAGE_METADATA

The format described by this document is used not only for coredumps,
but also for other purposes, so we've outgrown the old title.

A redirect was added based on https://codepo8.github.io/github-redirection-demo/.
I tried to use a relative link, but it's hard to test if it works
before merging.

Co-authored-by: Benjamin Franzke <benjaminfranzke@googlemail.com>
2 years agosysext: refuse empty release ID to avoid triggering assertion
Yu Watanabe [Fri, 20 May 2022 18:03:21 +0000 (03:03 +0900)] 
sysext: refuse empty release ID to avoid triggering assertion

Otherwise, the assertion in extension_release_validate() will be
triggered.

2 years agotest: add test for bus introspection of portable1
Yu Watanabe [Fri, 20 May 2022 19:14:03 +0000 (04:14 +0900)] 
test: add test for bus introspection of portable1

Follow-up for #23454.

2 years agoMerge pull request #23439 from keszybz/kernel-install-verbose
Yu Watanabe [Fri, 20 May 2022 19:26:03 +0000 (04:26 +0900)] 
Merge pull request #23439 from keszybz/kernel-install-verbose

Make kernel-install --verbose more verbose and some autodetection logic fixups

2 years agoMerge pull request #23451 from medhefgo/boot-static-assert
Yu Watanabe [Fri, 20 May 2022 19:22:17 +0000 (04:22 +0900)] 
Merge pull request #23451 from medhefgo/boot-static-assert

boot: C11/static_assert

2 years agosd-bus: add comment and test in sd_bus_path_decode() for empty string
Luca Boccassi [Fri, 20 May 2022 13:00:39 +0000 (14:00 +0100)] 
sd-bus: add comment and test in sd_bus_path_decode() for empty string

3970         e = object_path_startswith(path, prefix);
(gdb) p path
$1 = 0x55c5a166f768 "/org/freedesktop/portable1/image"
(gdb) p prefix
$2 = 0x55c59ffc2928 "/org/freedesktop/portable1/image"
(gdb) p e
$1 = 0x5581a1675788 ""

This can be a bit confusing in certain cases, so add a comment and a
test to make the behaviour clearer and explicit.

2 years agoMerge pull request #23454 from keszybz/portable-introspect
Yu Watanabe [Fri, 20 May 2022 18:41:22 +0000 (03:41 +0900)] 
Merge pull request #23454 from keszybz/portable-introspect

Fix bus introspection of portable1

2 years agokernel-install: restore priority of check for /boot/loader/entries 23439/head
Zbigniew Jędrzejewski-Szmek [Thu, 19 May 2022 20:22:44 +0000 (22:22 +0200)] 
kernel-install: restore priority of check for /boot/loader/entries

Before 9e82a74cb0f08a288f9db228a0b5bec8a7188cdb, we had a check like the
following:

if [[ -d /efi/loader/entries ]] || [[ -d /efi/$MACHINE_ID ]]; then
    ENTRY_DIR_ABS="/efi/$MACHINE_ID/$KERNEL_VERSION"
elif [[ -d /boot/loader/entries ]] || [[ -d /boot/$MACHINE_ID ]]; then
    ENTRY_DIR_ABS="/boot/$MACHINE_ID/$KERNEL_VERSION"
elif [[ -d /boot/efi/loader/entries ]] || [[ -d /boot/efi/$MACHINE_ID ]]; then
    ENTRY_DIR_ABS="/boot/efi/$MACHINE_ID/$KERNEL_VERSION"


In stock Fedora 34-, /efi isn't used, but grub creates /boot/loader/entries and
installs kernels and initrds directly in /boot. Thus the second arm of the
check wins, and we end up with BOOT_ROOT=/boot.

After 9e82a74cb0f08a288f9db228a0b5bec8a7188cdb, we iterate over the inner
directory first and over the second directory later:

[ -d /efi/<machine-id> ]
[ -d /boot/efi/<machine-id> ]
[ -d /boot/<machine-id> ]
[ -d /efi/Default ]
[ -d /boot/efi/Default ]
[ -d /boot/Default ]
[ -d /efi/loader/entries ]
[ -d /boot/efi/loader/entries ]
[ -d /boot/loader/entries ]

This was partially reverted by 447a822f8ee47b63a4cae00423c4d407bfa5e516 which
removed Default from the list, and a5307e173bf86d695fe85b8e15e91126e8618a14,
which moved checks for /boot up, so we ended up with:

[ -d /efi/<machine-id> ]
[ -d /boot/<machine-id> ]
[ -d /boot/efi/<machine-id> ]
[ -d /efi/loader/entries ]
[ -d /boot/loader/entries ]
[ -d /boot/efi/loader/entries ]

6637cf9db67237857279262d93ee0e39023c5b85 added autodetection of an entry
token, so we end up checking the following suffixes:

<machine-id>, $IMAGE_ID, $ID, Default

But the important unchanged characteristic is that we iterate over the suffix
first. Sadly this breaks Fedora, because we find /boot/efi/<machine-id> before
we could find /boot/loader/entries. It seems that every possible aspect of
behaviour matters for somebody, so we need to keep the original order of
detection.

With the patch:

[ -d /efi/<machine-id> ]
...
[ -d /efi/loader/entries ]
[ -d /boot/<machine-id> ]
...
[ -d /boot/loader/entries ]
[ -d /boot/efi/<machine-id> ]
...
[ -d /boot/efi/loader/entries ]

Note that we need to check for "loader/entries" too, even though it is not
an entry-token candidate, so that we get the same detection priority as
before.

Fixes https://bugzilla.redhat.com/show_bug.cgi?id=2071034.

2 years agoportabled: wrap long lines and fix typo in error message 23454/head
Zbigniew Jędrzejewski-Szmek [Fri, 20 May 2022 12:53:50 +0000 (14:53 +0200)] 
portabled: wrap long lines and fix typo in error message

2 years agoportabled: refuse queries for empty image name
Zbigniew Jędrzejewski-Szmek [Fri, 20 May 2022 12:53:22 +0000 (14:53 +0200)] 
portabled: refuse queries for empty image name

I took inspiration from pid1:
bus_unit_find()
  → find_unit()
    → manager_load_unit_from_dbus_path()
      → unit_name_from_dbus_path()
        → !startswith(path, "/org/freedesktop/systemd1/unit/")
          → return -EINVAL
          ←
        ←
      ←
    ← if (r < 0) return 0
  ← 0


i.e. we return 0 when queried for "/org/freedesktop/systemd1/unit".

Fixes #23445.

2 years agoshared/extension-release: drop unnecessary strna() wrapper
Yu Watanabe [Fri, 20 May 2022 07:49:14 +0000 (16:49 +0900)] 
shared/extension-release: drop unnecessary strna() wrapper

These are not NULL.

2 years agoportable: reject root directories without an ID field in os-release 23452/head
Luca Boccassi [Fri, 20 May 2022 11:24:45 +0000 (12:24 +0100)] 
portable: reject root directories without an ID field in os-release

We always require at least ID to be set in os-release, reject
and propagate error to the caller instead of asserting later

2 years agodissect: ID from os-release should be non-empty, not just non-NULL
Luca Boccassi [Fri, 20 May 2022 11:05:34 +0000 (12:05 +0100)] 
dissect: ID from os-release should be non-empty, not just non-NULL

2 years agomacro: Use C11 static_assert 23451/head
Jan Janssen [Fri, 20 May 2022 10:49:50 +0000 (12:49 +0200)] 
macro: Use C11 static_assert

Now that we require C11 everywhere, we can always use static_assert.

2 years agoboot: Build with C11 too
Jan Janssen [Fri, 20 May 2022 10:47:43 +0000 (12:47 +0200)] 
boot: Build with C11 too

2 years agokernel-install: fix detection of entry-token if $BOOT_ROOT is configured
Zbigniew Jędrzejewski-Szmek [Thu, 19 May 2022 13:34:32 +0000 (15:34 +0200)] 
kernel-install: fix detection of entry-token if $BOOT_ROOT is configured

If $BOOT_ROOT is specified, but entry-token not, we'd skip the detection
altogether, effectively defaulting to entry-token=machine-id.
The case where $BOOT_ROOT was not specied, but entry-token was configured
was handled correctly.
This patch makes the handling of both symmetrical, i.e. will only set what
wasn't configured.

2 years agokernel-install: debug the configuration detection if --verbose
Zbigniew Jędrzejewski-Szmek [Thu, 19 May 2022 12:50:07 +0000 (14:50 +0200)] 
kernel-install: debug the configuration detection if --verbose

No changes to behaviour, but let's print everything out as we discover it.

The docs say that BOOT_ROOT can be specified by the environment. I have
it locally in /etc/kernel/install.conf, and then the override doesn't work.
It'd be nice to handle such cases more reliably.

2 years agoAdd sys/stat.h for S_IFDIR
Khem Raj [Thu, 19 May 2022 18:01:04 +0000 (11:01 -0700)] 
Add sys/stat.h for S_IFDIR

Fixes
../git/src/shared/mkdir-label.c:13:61: error: use of undeclared identifier 'S_IFDIR'
        r = mac_selinux_create_file_prepare_at(dirfd, path, S_IFDIR);

Signed-off-by: Khem Raj <raj.khem@gmail.com>
2 years agosd-bus: Fix introspect memory leak
Daan De Meyer [Thu, 19 May 2022 21:59:50 +0000 (23:59 +0200)] 
sd-bus: Fix introspect memory leak

We have to free the contents of the set on top of the set itself.

Fixes #23443.

2 years agokernel-install: bail if machine id generation fails
Zbigniew Jędrzejewski-Szmek [Thu, 19 May 2022 12:43:03 +0000 (14:43 +0200)] 
kernel-install: bail if machine id generation fails

The call is unlikely to fail, but systemd-id128 might not be installed.
We shouldn't continue with the empty string.

2 years agokernel-install: actually export KERNEL_INSTALL_VERBOSE
Zbigniew Jędrzejewski-Szmek [Thu, 19 May 2022 12:23:48 +0000 (14:23 +0200)] 
kernel-install: actually export KERNEL_INSTALL_VERBOSE

:(

2 years agoMerge pull request #23432 from bnf/docs-relative-links
Zbigniew Jędrzejewski-Szmek [Thu, 19 May 2022 06:54:49 +0000 (08:54 +0200)] 
Merge pull request #23432 from bnf/docs-relative-links

docs: use relative links

2 years agoportable: Remove unnecessary assert()
Daan De Meyer [Wed, 18 May 2022 22:44:31 +0000 (00:44 +0200)] 
portable: Remove unnecessary assert()

Fixes #23433

matches is plumbed through until it finally gets used in unit_match()
which can deal with NULL matches so the assert() is unnecessary and
can be removed.

The two call sites of extract_image_and_extensions() also don't
assert() on matches either.

2 years agotest: double timeout of TEST-50-DISSECT
Luca Boccassi [Wed, 18 May 2022 23:15:30 +0000 (00:15 +0100)] 
test: double timeout of TEST-50-DISSECT

It times out on slow CIs near the end of the test, eg:

[  553.539368] kernel: loop_reread_partitions: partition scan of loop3 () failed (rc=-5)
TEST-50-DISSECT: (timeout)

https://autopkgtest.ubuntu.com/results/autopkgtest-focal-upstream-systemd-ci-systemd-ci/focal/ppc64el/s/systemd-upstream/20220518_172659_bf20f@/log.gz

2 years agomeson: install 70-power-switch.rules
Yu Watanabe [Wed, 18 May 2022 15:15:12 +0000 (00:15 +0900)] 
meson: install 70-power-switch.rules

Fixes a bug introduced by 155078c835a00fed264a7b36b06f709d9b57cb1b.

Fixes #23425.

2 years agoudev: fixed config_parse_ifalias() logic not to skip setting IFLA_IFALIAS
Tomasz Pala [Wed, 18 May 2022 16:11:42 +0000 (18:11 +0200)] 
udev: fixed config_parse_ifalias() logic not to skip setting IFLA_IFALIAS
this flaw was introduced in 6a74900002981eacbde382f659ab706ca36155cc

2 years agodocs: streamline headline structure in JOURNAL_EXPORT_FORMATS.md 23432/head
Benjamin Franzke [Tue, 17 May 2022 22:37:53 +0000 (00:37 +0200)] 
docs: streamline headline structure in JOURNAL_EXPORT_FORMATS.md

Add one general first-level headline and transform
the existing ones into second-level headlines.

2 years agodocs: use relative links
Benjamin Franzke [Tue, 17 May 2022 22:05:38 +0000 (00:05 +0200)] 
docs: use relative links

Allows for links to work both on systemd.io (or forks) and
when viewed on https://github.com/systemd/systemd/tree/main/docs

Note that the markdown links are converted by jekyll-relative-links[1]
to html. This plugin is enabled by default on github pages[2][3].

Due to a bug in jekyll-relative-links – see
https://github.com/benbalter/jekyll-relative-links/issues/61 –
we need to avoid line-wrapped links when using relative markdown links.

[1] https://github.com/benbalter/jekyll-relative-links
[2] https://github.blog/2016-12-05-relative-links-for-github-pages/
[3] https://docs.github.com/en/pages/setting-up-a-github-pages-site-with-jekyll/about-github-pages-and-jekyll#plugins

2 years agoMerge pull request #23417 from bnf/docs-dark-mode
Zbigniew Jędrzejewski-Szmek [Wed, 18 May 2022 11:09:50 +0000 (13:09 +0200)] 
Merge pull request #23417 from bnf/docs-dark-mode

docs: add dark mode

2 years agodoc: fix typo
Yu Watanabe [Wed, 18 May 2022 08:57:05 +0000 (17:57 +0900)] 
doc: fix typo

2 years agoMerge pull request #23422 from keszybz/two-doc-tweaks
Yu Watanabe [Wed, 18 May 2022 08:36:33 +0000 (17:36 +0900)] 
Merge pull request #23422 from keszybz/two-doc-tweaks

Two doc tweaks

2 years agoREADME: fix typo 23422/head
Zbigniew Jędrzejewski-Szmek [Wed, 18 May 2022 06:28:48 +0000 (08:28 +0200)] 
README: fix typo

2 years agoNEWS: reword description of credstore
Zbigniew Jędrzejewski-Szmek [Tue, 17 May 2022 08:32:20 +0000 (10:32 +0200)] 
NEWS: reword description of credstore

The "in" at the end made the sentence hard to parse…

2 years agodocs: overhaul styling for example log messages on startpage 23417/head
Benjamin Franzke [Tue, 17 May 2022 11:48:27 +0000 (13:48 +0200)] 
docs: overhaul styling for example log messages on startpage

 * Use brand-green for demo log messages on the startpage
   instead of the default `green` color defined by the browser vendor
   (after all, the brand green color is intended to reflect the
   boot status output)
 * Add a matching blue intro color (mixed using HSL from brand-green
   by 55deg hue rotation and 4% lightness)
 * Use a defined font-size
   (to avoid blurry rendering at 13.33333px, due to font-size: smaller)
 * Add possibility for the browser to show horizontal scrollbar
   (avoid overflow in mobile view)

2 years agodocs: add dark mode
Benjamin Franzke [Tue, 17 May 2022 08:38:07 +0000 (10:38 +0200)] 
docs: add dark mode

Respect the user agent property  `prefers-color-scheme` by
rendering the documentation in dark mode, if requested.

Reorganise CSS to store the color palette into CSS variables
and reference these from another set of CSS variables that
are dynamically switched based on the prefers-color-scheme
media query.

Light mode variables stay they same as before,
but to ease color mixing, the current RGB values
have been transformed to HSL (using chrome devtools).
The current body background is now --sd-gray-extralight,
the current body color is now --sd-gray-extradark.
Other gray-variations, needed for dark-mode constrasts
are derived from these colors using HSL lightness shifting.

The systemd brand black color is used as dark mode background
and a matching extralight gray font color has been selected.

The link font-weight is reduced to 400 in dark mode,
as the green color on dark ground becomes to overwhelming
with a bold font.

The systemd logo color is dynamically swapped by using
the special fill value `currentColor` for dark/light-mode
depending parts – as per specification on brand.systemd.io

2 years agodocs: regenerate highlight style and streamline background overwrite
Benjamin Franzke [Tue, 17 May 2022 17:05:29 +0000 (19:05 +0200)] 
docs: regenerate highlight style and streamline background overwrite

 * Update to rougify 3.26.1 styling
 * Adapt background overwrite by removing default background color
   and only adding explicit overwrites for syntax error
   (error stylings appear for example in json-excerpts –
   see HACKING.md – and are therefore hidden by our css)

2 years agodocs: avoid multiple first-level headlines
Benjamin Franzke [Tue, 17 May 2022 13:54:46 +0000 (15:54 +0200)] 
docs: avoid multiple first-level headlines

One document should only contain one main
headline, both for structuring and for
visual reasons.

Relevant cases found via:

  git grep -A2 '^# ' docs/

2 years agodocs: provide a first level headline for the frontpage
Benjamin Franzke [Tue, 17 May 2022 15:05:22 +0000 (17:05 +0200)] 
docs: provide a first level headline for the frontpage

 * Provides a nice introduction to the intro paragraph
 * Ensures text of subpages start at the same position as the intro text
   on the homepage
   => less visual "jumps" while navigating

2 years agodocs: respect default browser font-size
Benjamin Franzke [Tue, 17 May 2022 12:21:13 +0000 (14:21 +0200)] 
docs: respect default browser font-size

16px is default by all browser vendors when
font-setting is defined to be 'medium'.

Scale all font sizes and margins to
based on 16px base size.

2 years agodocs: add reasonable difference between h2 and h3 font-size
Benjamin Franzke [Tue, 17 May 2022 13:57:14 +0000 (15:57 +0200)] 
docs: add reasonable difference between h2 and h3 font-size

2 years agodocs: replace <pre>, <code> and <tt> tags by backticks
Benjamin Franzke [Tue, 17 May 2022 14:59:00 +0000 (16:59 +0200)] 
docs: replace <pre>, <code> and <tt> tags by backticks

Let jekyll render a proper codeblock, instead
of using custom HTML code in markdown.
Also <tt> is deprecated.

2 years agodocs: adapt theme color meta property
Benjamin Franzke [Tue, 17 May 2022 11:11:03 +0000 (13:11 +0200)] 
docs: adapt theme color meta property

The theme color is used on android to style the chrome
browser-ui with a color that suits the webpage.

Use the dark brand color instead of a random default blue color
in order to fit to the rest of the color scheme.

2 years agodocs: remove unneeded font-related attributes from svg logo
Benjamin Franzke [Tue, 17 May 2022 08:19:51 +0000 (10:19 +0200)] 
docs: remove unneeded font-related attributes from svg logo

The font has been vectorized and therefore all
font related properties are ignored anyway.

2 years agodocs: use whole numbers for circle in page-logo
Benjamin Franzke [Tue, 17 May 2022 08:16:15 +0000 (10:16 +0200)] 
docs: use whole numbers for circle in page-logo

0.001 is usually rounded to 0 by browsers, still,
better to be explicit and save some bytes.

2 years agodocs: replace invalid h1 font-weight
Benjamin Franzke [Tue, 17 May 2022 13:05:01 +0000 (15:05 +0200)] 
docs: replace invalid h1 font-weight

h1 font weight is defined to be 100 but no font-face
definition for weight 100 is included.
The browser will use the nearest available font-weight
instead. As that is 400, we do specify it explicitly now.

2 years agodocs: remove link to missing apple-touch-icon
Benjamin Franzke [Tue, 17 May 2022 11:16:25 +0000 (13:16 +0200)] 
docs: remove link to missing apple-touch-icon

2 years agodocs: remove invalid link to feed.xml
Benjamin Franzke [Tue, 17 May 2022 11:16:09 +0000 (13:16 +0200)] 
docs: remove invalid link to feed.xml

2 years agodocs: remove trailing space
Benjamin Franzke [Tue, 17 May 2022 11:47:21 +0000 (13:47 +0200)] 
docs: remove trailing space

2 years agodocs: configure editorconfig for css and html
Benjamin Franzke [Tue, 17 May 2022 08:38:00 +0000 (10:38 +0200)] 
docs: configure editorconfig for css and html

2 years agoMerge pull request #23412 from yuwata/network-bond-param
Luca Boccassi [Tue, 17 May 2022 19:11:55 +0000 (20:11 +0100)] 
Merge pull request #23412 from yuwata/network-bond-param

networkctl: bond param

2 years agoUpdate TODO
Luca Boccassi [Tue, 17 May 2022 16:47:38 +0000 (17:47 +0100)] 
Update TODO

2 years agoMerge pull request #23114 from yuwata/resolve-dnssec
Zbigniew Jędrzejewski-Szmek [Tue, 17 May 2022 14:53:09 +0000 (16:53 +0200)] 
Merge pull request #23114 from yuwata/resolve-dnssec

resolve: always request additional record to verify negative answer

2 years agoAdd a "test" that prints the SBAT table
Zbigniew Jędrzejewski-Szmek [Thu, 12 May 2022 17:06:29 +0000 (19:06 +0200)] 
Add a "test" that prints the SBAT table

The SBAT section was included in a special section in the EFI code, but
the contents weren't directly visible in any way. Let's add a "test" that
prints them for visual inspection.

If there's some external linter for this format, we could hook it up in the
future.

2 years agotest-boot-timestamp: use ERRNO_IS_PRIVILEGE() to also mask -EPERM
Yu Watanabe [Tue, 17 May 2022 09:13:13 +0000 (18:13 +0900)] 
test-boot-timestamp: use ERRNO_IS_PRIVILEGE() to also mask -EPERM

Fixes #23395.

2 years agotest-network: add test for showing Bond parameters 23412/head
Yu Watanabe [Tue, 17 May 2022 13:35:03 +0000 (22:35 +0900)] 
test-network: add test for showing Bond parameters

2 years agonetworkctl: fix units for bond parameters
Yu Watanabe [Tue, 17 May 2022 13:20:32 +0000 (22:20 +0900)] 
networkctl: fix units for bond parameters

Fixes RHBZ#2086166 (https://bugzilla.redhat.com/show_bug.cgi?id=2086166).

2 years agoMerge pull request #23409 from keszybz/network-online-docs
Luca Boccassi [Tue, 17 May 2022 12:42:51 +0000 (13:42 +0100)] 
Merge pull request #23409 from keszybz/network-online-docs

Network online docs

2 years agoMerge pull request #23407 from keszybz/bpf-cleanup-warning
Yu Watanabe [Tue, 17 May 2022 11:31:10 +0000 (20:31 +0900)] 
Merge pull request #23407 from keszybz/bpf-cleanup-warning

Clean up bpf cleanup warning

2 years agodocs: say how to delay network-online.target 23409/head
Zbigniew Jędrzejewski-Szmek [Tue, 17 May 2022 10:25:52 +0000 (12:25 +0200)] 
docs: say how to delay network-online.target

Could be related to https://access.redhat.com/solutions/4879291.
Unfortunately I can't access this page, but the title seems
relevant.

2 years agodocs: copy the page about network-online.target from the wiki
Zbigniew Jędrzejewski-Szmek [Tue, 17 May 2022 10:01:23 +0000 (12:01 +0200)] 
docs: copy the page about network-online.target from the wiki

From https://www.freedesktop.org/wiki/Software/systemd/NetworkTarget/.

I changed the order in the page, but didn't change the text too much. Now the
discussion of the different targets is at the top, and they are ordered like
during boot (network-pre.target first, then network.target, and
network-online.target last). The parts about LSB and $network are pushed down a
bit. I think it is still useful to have them, but not as the main entry point
into the discussion. I tried to clean up the grammar and wording a bit.

One meanigful change is that we now don't say that network-online.target means
interfaces are up and IP addresses have been assigned. In other places we were
saying that the actual implementation is provided by
NetworkManager-wait-online.service, so the actual meaning is not under our
control. The text is changed to say "usually".

The last paragraph is new, I think it's good to say that
"dnf-makecache.service" is fine to use "network-online.target".

2 years agomanager: skip BPF cleanup if we never initialized 23407/head
Zbigniew Jędrzejewski-Szmek [Tue, 17 May 2022 08:13:49 +0000 (10:13 +0200)] 
manager: skip BPF cleanup if we never initialized

This fixes a spurious warning from the manager running in user mode:

systemd[1668]: Reached target sockets.target.
systemd[1669]: Failed to create BPF map: Operation not permitted
systemd[1669]: Finished systemd-tmpfiles-setup.service.
systemd[1669]: Listening on dbus.socket.
systemd[1669]: Reached target sockets.target.
systemd[1669]: Reached target basic.target.
systemd[1]: Started user@6.service.

Fixes https://bugzilla.redhat.com/show_bug.cgi?id=2084955.

2 years agotest-execute: skip one more test when PID1 is not systemd
Yu Watanabe [Mon, 16 May 2022 22:06:39 +0000 (07:06 +0900)] 
test-execute: skip one more test when PID1 is not systemd

Fixes #23332.

2 years agocore/bpf: lsm_bpf_supported() returns a boolean
Zbigniew Jędrzejewski-Szmek [Tue, 17 May 2022 08:11:05 +0000 (10:11 +0200)] 
core/bpf: lsm_bpf_supported() returns a boolean

The code was corret, but confusing, because it was treating the int
as a boolean.

2 years agoMerge pull request #23403 from keszybz/docs-update
Yu Watanabe [Mon, 16 May 2022 21:02:08 +0000 (06:02 +0900)] 
Merge pull request #23403 from keszybz/docs-update

Say that we support recent version of major distros and other doc updates

2 years agorpm: remove check if systemd is running in a few cases
Zbigniew Jędrzejewski-Szmek [Mon, 16 May 2022 18:06:59 +0000 (20:06 +0200)] 
rpm: remove check if systemd is running in a few cases

rpms can be installed in two different modes: into a chroot, where the system
is not running, and onto a live system. In the first mode, where should create
all changes that are "permanent", and in the second mode, all changes which are
"permanent" but also those which only affect the running system. Thus, changes
like new modprobe rules, tmpfiles rules, binfmt rules, udev rules, etc., are
guarded by 'test -d "/run/systemd/system"' which is the official way to check
if systemd is running, so that they are *not* executed when installed into a
chroot. But the same logic does not apply to sysusers, hwdb, and the journal
catalog: all those files can and should result in changes being performed
immediately to the system. This makes the creation of immutable images possible
(because there are no permanent changes to executed after a reboot), and allows
other packages to depend on the the effect of those changes.

Thus, the guard to check if we're not in a chroot is dropped from triggers for
sysusers, hwdb, and the journal catalog. This means that those triggers will
execute, and no subsequent work is needed. systemd-sysusers.service,
systemd-journal-catalog-update.service, and systemd-hwdb-update.service.in all
have ConditionNeedsUpdate= so they they generally won't be invoked after a
reboot. (systemd.rpm does not touch /usr to trigger the condition, because the
%transfiletriggers make that unnecessary.)

https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=2085481

2 years agohwdb: Add Google Hangouts Meet speakermic
Pablo Ceballos [Fri, 13 May 2022 23:33:07 +0000 (23:33 +0000)] 
hwdb: Add Google Hangouts Meet speakermic

This device implements the phone mute HID usage as a toggle switch,
where 1 indicates muted, and 0 indicates unmuted. However, for a key
event 1 indicates that the key has been pressed and 0 indicates it has
been released. This mismatch causes issues, so prevent key events from
being generated for this HID usage.

2 years agoREADME: say that recent releases are supported 23403/head
Zbigniew Jędrzejewski-Szmek [Mon, 16 May 2022 17:49:20 +0000 (19:49 +0200)] 
README: say that recent releases are supported

Prompted by the discussion in
https://github.com/systemd/systemd/pull/23391.

2 years agoREADME: recommend --no-rebuild for sudo meson install
Zbigniew Jędrzejewski-Szmek [Mon, 16 May 2022 17:23:37 +0000 (19:23 +0200)] 
README: recommend --no-rebuild for sudo meson install

sudo and rebuilding don't mix well.
C.f. fe2b7631ad20b0e518b9f566608ff5f9b84d4d8b.

2 years agoREADME: add missing colons and wrap to ~80 columns
Zbigniew Jędrzejewski-Szmek [Mon, 16 May 2022 17:22:11 +0000 (19:22 +0200)] 
README: add missing colons and wrap to ~80 columns

Some paragraphs were narrow for no good reason. Let's make things
a bit more uniform.

2 years agonetworkd: reword debug messages
Zbigniew Jędrzejewski-Szmek [Mon, 16 May 2022 11:14:37 +0000 (13:14 +0200)] 
networkd: reword debug messages

"left from <something>" is not correct. "left <something>" would be the
usual form, but "left master interface" is not clear at all. So reword
those messages totally.

Follow-up for 3881fd406b13746336c6c654fd2f68261161fd8a.

2 years agomeson: add 'pam' install_tag to pam modules
Mike Gilbert [Sun, 15 May 2022 21:43:48 +0000 (17:43 -0400)] 
meson: add 'pam' install_tag to pam modules

2 years agohwdb: fix accelerometer mount matrix for Aquarius NS483
Nikolai Kostrigin [Fri, 13 May 2022 14:13:00 +0000 (17:13 +0300)] 
hwdb: fix accelerometer mount matrix for Aquarius NS483

Signed-off-by: Nikolai Kostrigin <nickel@altlinux.org>
2 years agoMerge pull request #23383 from yuwata/meson-suppress-warning
Luca Boccassi [Mon, 16 May 2022 16:57:13 +0000 (17:57 +0100)] 
Merge pull request #23383 from yuwata/meson-suppress-warning

meson: suppress warning

2 years agoman: fix typo
Zbigniew Jędrzejewski-Szmek [Mon, 16 May 2022 09:49:52 +0000 (11:49 +0200)] 
man: fix typo

2 years agoMerge pull request #23377 from nabijaczleweli/shopt
Yu Watanabe [Mon, 16 May 2022 09:17:14 +0000 (18:17 +0900)] 
Merge pull request #23377 from nabijaczleweli/shopt

Fix shopt in 90-loaderentry, don't not lint kernel-install, respect `$TMPDIR`

2 years agofuzz: drop too large input
Yu Watanabe [Sun, 15 May 2022 19:44:27 +0000 (04:44 +0900)] 
fuzz: drop too large input

The original issue oss-fuzz#10734 (https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=10734)
is that just the file size is too large, and not a issue in functions
tested by the fuzzer. It is not necessary to include the testcase.

Follow-up for c4f883b78e5ffd326a82eaf18e01a9e4e243db58.

Fixes #23390.

2 years agocore: fix comment
Yu Watanabe [Sun, 15 May 2022 20:18:30 +0000 (05:18 +0900)] 
core: fix comment

2 years agomeson: always get libbpf include directory from pkgconfig 23383/head 23399/head
Yu Watanabe [Sun, 15 May 2022 18:38:45 +0000 (03:38 +0900)] 
meson: always get libbpf include directory from pkgconfig

2 years agomeson: use fs.name() and fs.parent()
Yu Watanabe [Sun, 15 May 2022 19:05:51 +0000 (04:05 +0900)] 
meson: use fs.name() and fs.parent()

Follow-up for 1bd0cc452ce1cbfb330db6b0855e178b4814e7de.

2 years agomeson: always use ExternalProgram.path()
Yu Watanabe [Sun, 15 May 2022 17:43:59 +0000 (02:43 +0900)] 
meson: always use ExternalProgram.path()

Even if we use meson >= 0.55, using path() does not produce any error or
warning if the required version is below 0.55.

Let's convert path() with full_path() when we requires meson >= 0.55.

2 years agomeson: disable bpf if skip-deps is enabled
Yu Watanabe [Sun, 15 May 2022 14:13:38 +0000 (23:13 +0900)] 
meson: disable bpf if skip-deps is enabled

2 years agomeson: use boolean for configuration_data.set10()
Yu Watanabe [Sun, 15 May 2022 09:10:25 +0000 (18:10 +0900)] 
meson: use boolean for configuration_data.set10()

Setting with number is deprecated:
```
meson.build:1008: DEPRECATION: configuration_data.set10 with number. the `set10` method should only be used with booleans
```

2 years agomeson: use better shellscript argument passing
Eli Schwartz [Sun, 15 May 2022 15:11:24 +0000 (11:11 -0400)] 
meson: use better shellscript argument passing

Passing potentially arbitrary data into a shellscript is potentially
very broken if you do not correctly quote it for use. This quoting must
be done as part of the interpretation of the data itself, e.g. python's
shlex.quote; simply formatting it into a string with double quotes is
NOT sufficient.

An alternative is to communicate the data reliably via argv to the shell
process, and allow the shell to internally handle it via `"$1"`, which
is quote-safe and will expand the data from argv as a single tokenized
word.

2 years agocatalog,po: update Polish translation
Piotr Drąg [Sun, 15 May 2022 11:21:13 +0000 (13:21 +0200)] 
catalog,po: update Polish translation

2 years agopo: update Japanese translation
Yu Watanabe [Sun, 15 May 2022 14:26:59 +0000 (23:26 +0900)] 
po: update Japanese translation

2 years agogit: ignore manifest
Yu Watanabe [Sun, 15 May 2022 14:08:53 +0000 (23:08 +0900)] 
git: ignore manifest

2 years agoMerge pull request #23378 from floppym/install-tags
Luca Boccassi [Sun, 15 May 2022 13:52:19 +0000 (14:52 +0100)] 
Merge pull request #23378 from floppym/install-tags

Add install tags for nss modules and pkgconfig files

2 years agoMerge pull request #23382 from yuwata/resolve-fix-dnssd-assertion
Luca Boccassi [Sun, 15 May 2022 13:49:24 +0000 (14:49 +0100)] 
Merge pull request #23382 from yuwata/resolve-fix-dnssd-assertion

resolve: fix dnssd assertion

2 years agoresolve: merge variable declaration with same type 23382/head
Yu Watanabe [Sun, 15 May 2022 01:26:24 +0000 (10:26 +0900)] 
resolve: merge variable declaration with same type

2 years agoresolve/dnssd: make dnssd_render_instance_name() take Manager
Yu Watanabe [Sun, 15 May 2022 01:23:25 +0000 (10:23 +0900)] 
resolve/dnssd: make dnssd_render_instance_name() take Manager

As DnssdService object passed to dnssd_render_instance_name() may
not owned by Manager.

Fixes #23381.

2 years ago hwdb: Add accel orientation for the I15-TG
Martin [Sat, 14 May 2022 22:51:59 +0000 (00:51 +0200)] 
 hwdb: Add accel orientation for the I15-TG

Add accel orientation for the Cube iWork 10 I15-TG

2 years agonetwork: do not update interface group by default
Yu Watanabe [Thu, 12 May 2022 19:43:37 +0000 (04:43 +0900)] 
network: do not update interface group by default

This fixes a minor bug introduced by 10af8bb24b39a815079f6bf31b449c6e5aaa2adf.
Before the commit, the interface group was set only when Group= is explicitly
specified, otherwise the interface group was kept. However, after the commit,
we need to specify Group= with an empty string to keep the current interface
group.

2 years agohwdb: add touchpad parameters for Lenovo T15g Gen1 (#23373)
Nikolai Grigoriev [Sat, 14 May 2022 21:39:02 +0000 (17:39 -0400)] 
hwdb: add touchpad parameters for Lenovo T15g Gen1 (#23373)

2 years agomeson: add 'devel' install_tag to libsystemd.pc and libudev.pc 23378/head
Mike Gilbert [Sat, 14 May 2022 15:23:02 +0000 (11:23 -0400)] 
meson: add 'devel' install_tag to libsystemd.pc and libudev.pc

2 years agomeson: add 'nss' install_tag to nss modules
Mike Gilbert [Sat, 14 May 2022 15:21:20 +0000 (11:21 -0400)] 
meson: add 'nss' install_tag to nss modules

2 years agoDon't not lint kernel-install 23377/head
наб [Sat, 14 May 2022 13:12:14 +0000 (15:12 +0200)] 
Don't not lint kernel-install

2 years agokernel-install: respect $TMPDIR
наб [Sat, 14 May 2022 13:10:30 +0000 (15:10 +0200)] 
kernel-install: respect $TMPDIR

2 years agokernel-install: 90-loaderentry: remove shopt
наб [Sat, 14 May 2022 13:07:04 +0000 (15:07 +0200)] 
kernel-install: 90-loaderentry: remove shopt

Fixes: #23374
Fixes: 367165a4069ac0c04882a05a8a80f6afb1e42760
2 years agoresolve: fix memleak
Yu Watanabe [Sat, 14 May 2022 06:09:11 +0000 (15:09 +0900)] 
resolve: fix memleak

Fixes a bug introduced by 71aee23dba7faeef68e7232f444626267a6c90d7.

Fixes CID#1488763.