]> git.ipfire.org Git - thirdparty/systemd.git/log
thirdparty/systemd.git
2 years agoescape: add missing non-NULL parameter assertions 26628/head
msizanoen1 [Thu, 2 Mar 2023 11:37:02 +0000 (18:37 +0700)] 
escape: add missing non-NULL parameter assertions

2 years agotest-escape: Add tests for escaping bogus UTF-8 sequences
msizanoen1 [Wed, 1 Mar 2023 14:48:08 +0000 (21:48 +0700)] 
test-escape: Add tests for escaping bogus UTF-8 sequences

2 years agoescape: Ensure that output is always valid UTF-8
msizanoen1 [Wed, 1 Mar 2023 10:35:17 +0000 (17:35 +0700)] 
escape: Ensure that output is always valid UTF-8

This ensures that shell string escape operations will not produce output
with invalid UTF-8 from the input by escaping invalid UTF-8 data as if
they were single byte characters.

2 years agoMerge pull request #26548 from poettering/user-record-ambient
Luca Boccassi [Tue, 28 Feb 2023 23:40:57 +0000 (23:40 +0000)] 
Merge pull request #26548 from poettering/user-record-ambient

add ambient/bounding set fields to user records, and set CAP_WAKE_ALARM by default for local regular user sessions

2 years agoupdate NEWS 26548/head
Lennart Poettering [Wed, 22 Feb 2023 17:43:45 +0000 (18:43 +0100)] 
update NEWS

2 years agotest: test setting ambient caps via pam_systemd.so
Lennart Poettering [Wed, 22 Feb 2023 17:05:18 +0000 (18:05 +0100)] 
test: test setting ambient caps via pam_systemd.so

2 years agopam_systemd: process the two new capabilities user records fields in pam_systemd
Lennart Poettering [Fri, 17 Feb 2023 21:49:16 +0000 (22:49 +0100)] 
pam_systemd: process the two new capabilities user records fields in pam_systemd

And also: by default, for the systemd-user service and for local
sessions (i.e. those assigned to a seat): let's imply CAP_WAKE_SYSTEM
for them by default. Yes, let's pass one specific capability by default to local
unprivileged users.

The capability services exactly once purpose: to allow system wake-up
from suspend via alarm clocks, hence is relatively limited in focus. By
adding this tools such as GNOME's Alarm Clock app can simply allocate a
CLOCK_REALTIME_ALARM (or ask systemd --user to do this) timer and it
will wake up the system as necessary.

Note that systemd --user will not pass the ambient caps on by default,
so even with this change, individual services need to use
AmbientCapabilities= to pass this on to the individual programs.

Fixes: #17564 #21382
2 years agohomectl: make the new caps field configurable via homectl
Lennart Poettering [Fri, 17 Feb 2023 21:25:47 +0000 (22:25 +0100)] 
homectl: make the new caps field configurable via homectl

2 years agouser-record: extend user records with an ambient and bounding caps set field
Lennart Poettering [Fri, 17 Feb 2023 21:22:16 +0000 (22:22 +0100)] 
user-record: extend user records with an ambient and bounding caps set field

In particular the ambieht caps field is useful: we can use it later to
pass caps such as CAP_WAKE_ALARM to regular users on login.

2 years agomkosi: add some really basic tools to default mkosi image
Lennart Poettering [Wed, 22 Feb 2023 12:04:46 +0000 (13:04 +0100)] 
mkosi: add some really basic tools to default mkosi image

"passwd" and "pscap" are extremely useful to debug basic OS behaviour,
and tiny. So let's add them to our default development images, just to
save us some headaches.

2 years agoMerge pull request #26608 from medhefgo/boot-arm
Luca Boccassi [Tue, 28 Feb 2023 18:48:45 +0000 (18:48 +0000)] 
Merge pull request #26608 from medhefgo/boot-arm

boot: Some ARM fixes

2 years agoman: fix typo in systemd.generator
Antonio Alvarez Feijoo [Tue, 28 Feb 2023 15:04:03 +0000 (16:04 +0100)] 
man: fix typo in systemd.generator

2 years agoboot: Provide div0 handlers for ARM 26608/head
Jan Janssen [Mon, 27 Feb 2023 17:43:07 +0000 (18:43 +0100)] 
boot: Provide div0 handlers for ARM

This is a cleaner approach to 59833064742310bfccf028b0278811ba5cff8dcf.

2 years agoboot: Fix data model detection for ARM
Jan Janssen [Mon, 27 Feb 2023 14:58:42 +0000 (15:58 +0100)] 
boot: Fix data model detection for ARM

For whatever reason, ARM does not define __ILP32__.

2 years agopsi-util: fix error handling
Lennart Poettering [Mon, 27 Feb 2023 18:02:41 +0000 (19:02 +0100)] 
psi-util: fix error handling

We checked ERRNO_IS_NOT_SUPPORTED on a possible positive non-error code,
which isn't right.

Fix that. Also add caching, since we are about to call this more often.

2 years agosd-event: handle kernels that set CONFIG_PSI_DEFAULT_DISABLED more gracefully
Lennart Poettering [Mon, 27 Feb 2023 18:05:19 +0000 (19:05 +0100)] 
sd-event: handle kernels that set CONFIG_PSI_DEFAULT_DISABLED more gracefully

If CONFIG_PSI_DEFAULT_DISABLED is set in the kernel, then the PSI files
will be there, and you can open them, but read()/write() will fail.
Which is terrible, since that happens so late. But anyway, handle this
gracefully.

2 years agoMerge pull request #26605 from poettering/noxenix
Luca Boccassi [Tue, 28 Feb 2023 12:14:16 +0000 (12:14 +0000)] 
Merge pull request #26605 from poettering/noxenix

journal error message fixes

2 years agojournald: remove triplicate logging about failure to write log lines 26605/head
Lennart Poettering [Mon, 27 Feb 2023 14:48:45 +0000 (15:48 +0100)] 
journald: remove triplicate logging about failure to write log lines

Let's log exactly at one place about failed writing of log lines to
journal file: in shall_try_append_again().

Then, if we decide to suppress a retry-after-vacuum because we already
vacuumed anyway then say this explicitly as "supressed rotation",
because that's what we do here.

This removes triplicate logging about the same error, and logs exactly
once, plus optional one "suppressed rotation" message. (plus more debug
output). The triplicate logging was bad in particular because it had no
understanding of the actual error codes and just showed generic UNIX
error strings ("Not a XENIX named type file"). By relying on
shall_try_append_again() to do all logging we now get very clean error
strings for all conditions.

Fixes: #26496
2 years agojournald: always pass error code to logging function, even if we don't use it with %m
Lennart Poettering [Mon, 27 Feb 2023 14:47:11 +0000 (15:47 +0100)] 
journald: always pass error code to logging function, even if we don't use it with %m

We always want to pass the error code along with the log call, so that
it can add it to structured logging, even if the format string does not
contain %m.

2 years agojournald: upgrade log message about missing mmap() support to LOG_ERR
Lennart Poettering [Mon, 27 Feb 2023 14:42:59 +0000 (15:42 +0100)] 
journald: upgrade log message about missing mmap() support to LOG_ERR

in this case we really cannot proceed and fail. Let's log about this
loudly, since we simply fail the operation to write anything.

2 years agojournald: downgrade various log messages from LOG_WARNING to LOG_INFO
Lennart Poettering [Mon, 27 Feb 2023 14:40:48 +0000 (15:40 +0100)] 
journald: downgrade various log messages from LOG_WARNING to LOG_INFO

None of these conditions are real issues, but they can simply happen
because we just swtched from /run to /var as backend for logging and
there are old files from different boots with different systemd versions
and so on.

Let's not make more noise than necessary: still log, but not consider it
a warning, but just some normal thing.

We are handling these issues safely after all: by rotating and starting
anew, i.e. there's no reason to be concerned.

2 years agojournald: start log message uppercase, like every other one around
Lennart Poettering [Mon, 27 Feb 2023 14:39:24 +0000 (15:39 +0100)] 
journald: start log message uppercase, like every other one around

2 years agojournald: make sure shall_try_append_again() logs about all return codes passed in...
Lennart Poettering [Mon, 27 Feb 2023 14:35:01 +0000 (15:35 +0100)] 
journald: make sure shall_try_append_again() logs about all return codes passed in, not just some

2 years agosystemctl: print better message if default target is masked
David Tardon [Wed, 2 Mar 2022 20:02:21 +0000 (21:02 +0100)] 
systemctl: print better message if default target is masked

If the default target is masked, `systemctl get-default` prints

  Failed to get default target: Operation not possible due to RF-kill

That's a bit too cryptic, so let's make it clear what's actually
happening.

Fixes #26589.

2 years agotests: temporarily disable time-util failing tests
Luca Boccassi [Mon, 27 Feb 2023 18:42:04 +0000 (18:42 +0000)] 
tests: temporarily disable time-util failing tests

These are failing since https://github.com/systemd/systemd/pull/26409
disable for now

2 years agoRevert "dissect-image: don't probe swap partitions needlessly"
Lennart Poettering [Mon, 27 Feb 2023 12:45:17 +0000 (13:45 +0100)] 
Revert "dissect-image: don't probe swap partitions needlessly"

This reverts commit df4524cb3796529b435af309f5e6a2d92bcdaaa0.

This commit is just wrong. The thing is that we *don't* know
automatically that the partition contains a swap image, because it could
be encrypted.

Hence revert.

Fixes: #26595
2 years agoupdate TODO
Lennart Poettering [Mon, 27 Feb 2023 18:02:35 +0000 (19:02 +0100)] 
update TODO

2 years agotest: add coverage test for udevadm
David Tardon [Fri, 11 Nov 2022 14:28:25 +0000 (15:28 +0100)] 
test: add coverage test for udevadm

2 years agoMerge pull request #26578 from thkukuk/main
Yu Watanabe [Sat, 25 Feb 2023 22:54:11 +0000 (07:54 +0900)] 
Merge pull request #26578 from thkukuk/main

sd-login: add sd_uid_get_login_time interface #26574

2 years agorules: remove redundant duplicate comparisons
Dmitry V. Levin [Sat, 25 Feb 2023 20:00:00 +0000 (20:00 +0000)] 
rules: remove redundant duplicate comparisons

$ grep -F -n -o 'ENV{DISKSEQ}=="?*", ' rules.d/60-persistent-storage.rules
139:ENV{DISKSEQ}=="?*",
139:ENV{DISKSEQ}=="?*",
140:ENV{DISKSEQ}=="?*",
140:ENV{DISKSEQ}=="?*",

Reported-by: Alexey Gladkov <legion@kernel.org>
Fixes: 17d97d4c90f8 ("udev: create disk/by-diskseq symlink only when the device has diskseq")
Fixes: 583dc6d933d8 ("udev: also create partition /dev/disk/by-diskseq/ symlinks")
2 years agosd-login: document that result is in microseconds 26578/head
Thorsten Kukuk [Fri, 24 Feb 2023 13:51:00 +0000 (14:51 +0100)] 
sd-login: document that result is in microseconds

2 years agosd-login: add sd_uid_get_login_time interface #26574
Thorsten Kukuk [Fri, 24 Feb 2023 08:47:52 +0000 (09:47 +0100)] 
sd-login: add sd_uid_get_login_time interface #26574

2 years agotpm2-util: use compound initialization when allocating tpm2 objects
Lennart Poettering [Fri, 24 Feb 2023 17:20:50 +0000 (18:20 +0100)] 
tpm2-util: use compound initialization when allocating tpm2 objects

2 years agotree-wide: Use correct SPDX license identifier
Jan Janssen [Fri, 24 Feb 2023 18:23:01 +0000 (19:23 +0100)] 
tree-wide: Use correct SPDX license identifier

2 years agoman: add two missing commands to synopsys
Lennart Poettering [Fri, 24 Feb 2023 13:24:18 +0000 (14:24 +0100)] 
man: add two missing commands to synopsys

2 years agoupdate TODO
Lennart Poettering [Fri, 24 Feb 2023 16:08:56 +0000 (17:08 +0100)] 
update TODO

2 years agoMerge pull request #26579 from poettering/manager-env-clean-up
Lennart Poettering [Fri, 24 Feb 2023 13:11:57 +0000 (14:11 +0100)] 
Merge pull request #26579 from poettering/manager-env-clean-up

service manager env var clean-up fixes

2 years agomempool: fix tile alignment check
Lennart Poettering [Fri, 24 Feb 2023 09:41:47 +0000 (10:41 +0100)] 
mempool: fix tile alignment check

We should check alignment *after* determining the pointer points into
our pool, not before. Otherwise might might end up checking alignment of
the pointer relative to our base, even though it is taken relative to
some other base.

Follow-up for: a2b052b29f8bc141e94a4af95d1653a38a57eaeb
See: https://github.com/systemd/systemd/pull/26393#issuecomment-1442295012

2 years agolibsystemd: sd_journal_get_seqnum() must be tagged with 254 symver, not 253
Lennart Poettering [Fri, 24 Feb 2023 10:15:46 +0000 (11:15 +0100)] 
libsystemd: sd_journal_get_seqnum() must be tagged with 254 symver, not 253

This is a follow-up for b1712fabd1702640b04b0acdbba2d78294313a4d which
was prepped for 253, but merged into early 254 development cycle. It
thus had the symbol it adds at the wrong symver. Fix thta.

2 years agoMerge pull request #26409 from yuwata/parse-timestamp
Luca Boccassi [Fri, 24 Feb 2023 11:45:03 +0000 (11:45 +0000)] 
Merge pull request #26409 from yuwata/parse-timestamp

fix parse_timestamp()

2 years agoMerge pull request #26217 from bluca/dump_mem
Luca Boccassi [Fri, 24 Feb 2023 10:31:21 +0000 (10:31 +0000)] 
Merge pull request #26217 from bluca/dump_mem

Add D-Bus method and analyze command to dump malloc_info()

2 years agoman: document REMOTE_ADDR/REMOTE_PORT along with the other env vars 26579/head
Lennart Poettering [Fri, 24 Feb 2023 09:21:49 +0000 (10:21 +0100)] 
man: document REMOTE_ADDR/REMOTE_PORT along with the other env vars

2 years agomanager: be more careful with env var clean-up
Lennart Poettering [Fri, 24 Feb 2023 09:15:35 +0000 (10:15 +0100)] 
manager: be more careful with env var clean-up

Let's add some more env vars we nowadays set for activated services to
the list of env vars we strip from the environment block we pass on.

2 years agoMerge pull request #26574 from YHNdnzj/sd-login-new-interface
Yu Watanabe [Fri, 24 Feb 2023 00:20:52 +0000 (09:20 +0900)] 
Merge pull request #26574 from YHNdnzj/sd-login-new-interface

sd-login: add two interfaces for retriving session info

2 years agotest: add tests for format_timestamp() and parse_timestamp() with various timezone 26409/head
Yu Watanabe [Sat, 11 Feb 2023 20:30:49 +0000 (05:30 +0900)] 
test: add tests for format_timestamp() and parse_timestamp() with various timezone

2 years agotime-util: make parse_timestamp() use the RFC-822/ISO 8601 standard timezone spec
Yu Watanabe [Mon, 13 Feb 2023 18:39:15 +0000 (03:39 +0900)] 
time-util: make parse_timestamp() use the RFC-822/ISO 8601 standard timezone spec

If the timezone is specified with a number e.g. +0900 or so, then
let's parse the time as a UTC, and adjust it with the specified time
shift.

Otherwise, if an area has timezone change, e.g.
---
Africa/Casablanca  Sun Jun 17 01:59:59 2018 UT = Sun Jun 17 01:59:59 2018 +00 isdst=0 gmtoff=0
Africa/Casablanca  Sun Jun 17 02:00:00 2018 UT = Sun Jun 17 03:00:00 2018 +01 isdst=1 gmtoff=3600
Africa/Casablanca  Sun Oct 28 01:59:59 2018 UT = Sun Oct 28 02:59:59 2018 +01 isdst=1 gmtoff=3600
Africa/Casablanca  Sun Oct 28 02:00:00 2018 UT = Sun Oct 28 03:00:00 2018 +01 isdst=0 gmtoff=3600
---
then we could not determine isdst from the timezone (+01 in the above)
and mktime() will provide wrong results.

Fixes #26370.

2 years agotime-util: drop unnecessary assignment of timezone name
Yu Watanabe [Wed, 15 Feb 2023 04:51:15 +0000 (13:51 +0900)] 
time-util: drop unnecessary assignment of timezone name

As mktime() does not use timezone neme.

2 years agotime-util: rename variables
Yu Watanabe [Wed, 15 Feb 2023 04:46:50 +0000 (13:46 +0900)] 
time-util: rename variables

No functional changes, just refactoring.

2 years agotime-util: shorten code a bit
Yu Watanabe [Mon, 13 Feb 2023 19:27:52 +0000 (04:27 +0900)] 
time-util: shorten code a bit

No functional change, just refactoring.

2 years agotime-util: use usec_add() and usec_sub_unsigned()
Yu Watanabe [Mon, 13 Feb 2023 19:27:27 +0000 (04:27 +0900)] 
time-util: use usec_add() and usec_sub_unsigned()

And move the check with USEC_TIMESTAMP_FORMATTABLE_MAX at the end,
as usec_add() can handle overflow correctly.

2 years agotime-util: use result from startswith_no_case()
Yu Watanabe [Mon, 13 Feb 2023 19:14:24 +0000 (04:14 +0900)] 
time-util: use result from startswith_no_case()

No functional change, just refactoring.

2 years agotime-util: do not use strdupa()
Yu Watanabe [Mon, 13 Feb 2023 18:41:26 +0000 (03:41 +0900)] 
time-util: do not use strdupa()

The input string may come from command line, config files.

2 years agotime-util: drop redundant else
Yu Watanabe [Mon, 13 Feb 2023 18:43:58 +0000 (03:43 +0900)] 
time-util: drop redundant else

2 years agotime-util: add assertions
Yu Watanabe [Mon, 13 Feb 2023 17:06:13 +0000 (02:06 +0900)] 
time-util: add assertions

2 years agotime-util: rename variables
Yu Watanabe [Mon, 13 Feb 2023 17:04:31 +0000 (02:04 +0900)] 
time-util: rename variables

2 years agotime-util: align string table
Yu Watanabe [Mon, 13 Feb 2023 16:41:38 +0000 (01:41 +0900)] 
time-util: align string table

2 years agotime-util: use DEFINE_STRING_TABLE_LOOKUP_TO_STRING() macro
Yu Watanabe [Mon, 13 Feb 2023 16:40:56 +0000 (01:40 +0900)] 
time-util: use DEFINE_STRING_TABLE_LOOKUP_TO_STRING() macro

2 years agotime-util: drop redundant space
Yu Watanabe [Mon, 13 Feb 2023 16:38:57 +0000 (01:38 +0900)] 
time-util: drop redundant space

2 years agotime-util: drop unused definition of FORMAT_TIMESTAMP_WIDTH
Yu Watanabe [Sat, 11 Feb 2023 20:30:34 +0000 (05:30 +0900)] 
time-util: drop unused definition of FORMAT_TIMESTAMP_WIDTH

2 years agoman,catalog: fix typo
Yu Watanabe [Thu, 23 Feb 2023 23:52:39 +0000 (08:52 +0900)] 
man,catalog: fix typo

Follow-up for #26448.

2 years agoshared: move cg_set_access() declaration to right header file
Lennart Poettering [Thu, 23 Feb 2023 17:27:43 +0000 (18:27 +0100)] 
shared: move cg_set_access() declaration to right header file

This function was moved from cgroup-util.c to cgroup-setup.c a while
back, but the prototype in the matching header files wasn't migrated.
Let's fix that.

2 years agojournal-file: Improve debug messages
Daan De Meyer [Thu, 23 Feb 2023 14:55:36 +0000 (15:55 +0100)] 
journal-file: Improve debug messages

2 years agojournal-file: Fix return value in bump_entry_array()
Daan De Meyer [Thu, 23 Feb 2023 14:40:38 +0000 (15:40 +0100)] 
journal-file: Fix return value in bump_entry_array()

2 years agosystemctl: use bus_message_hash_ops
Yu Watanabe [Thu, 23 Feb 2023 12:49:29 +0000 (21:49 +0900)] 
systemctl: use bus_message_hash_ops

2 years agoman: link up sd_event_add_memory_pressure() more
Lennart Poettering [Thu, 23 Feb 2023 16:22:57 +0000 (17:22 +0100)] 
man: link up sd_event_add_memory_pressure() more

2 years agoanalyze: add 'malloc' verb to dump malloc_info() 26217/head
Luca Boccassi [Thu, 26 Jan 2023 13:28:15 +0000 (13:28 +0000)] 
analyze: add 'malloc' verb to dump malloc_info()

Gets the memory state of the manager:

root@image:~# systemd-analyze malloc
<malloc version=1>
<heap nr=0>
<sizes>
  <size from=33 to=33 total=396 count=12/>
  <unsorted from=20385 to=20385 total=20385 count=1/>
</sizes>
<total type=fast count=0 size=0/>
<total type=rest count=14 size=36589/>
<system type=current size=1691648/>
<system type=max size=1839104/>
<aspace type=total size=1691648/>
<aspace type=mprotect size=1691648/>
</heap>
<total type=fast count=0 size=0/>
<total type=rest count=14 size=36589/>
<total type=mmap count=0 size=0/>
<system type=current size=1691648/>
<system type=max size=1839104/>
<aspace type=total size=1691648/>
<aspace type=mprotect size=1691648/>
</malloc>

2 years agosd-login: add sd_session_get_start_time 26574/head
Mike Yuan [Thu, 23 Feb 2023 17:45:50 +0000 (01:45 +0800)] 
sd-login: add sd_session_get_start_time

2 years agosd-login: add sd_session_get_username
Mike Yuan [Thu, 23 Feb 2023 16:48:10 +0000 (00:48 +0800)] 
sd-login: add sd_session_get_username

2 years agoMerge pull request #26258 from DaanDeMeyer/boot-smbios
Daan De Meyer [Thu, 23 Feb 2023 14:59:18 +0000 (15:59 +0100)] 
Merge pull request #26258 from DaanDeMeyer/boot-smbios

stub: Read extra kernel command line items from SMBIOS

2 years agomanager: add GetMallocInfo() hidden/debug method
Luca Boccassi [Thu, 26 Jan 2023 12:51:55 +0000 (12:51 +0000)] 
manager: add GetMallocInfo() hidden/debug method

Return the output of malloc_info() via a file descriptor (in case it
gets large on a busy system). Useful to get live data about memory
usage when it is not possible to run under a profiler from the get-go.
Do not formally register the method, but add a 'hidden' interface
so that it cannot be seen by introspection or by looking at the
object.

2 years agoMerge pull request #26535 from yuwata/systemctl-list-cleanups
Yu Watanabe [Thu, 23 Feb 2023 12:38:13 +0000 (21:38 +0900)] 
Merge pull request #26535 from yuwata/systemctl-list-cleanups

systemctl-list: several cleanups

2 years agoMerge pull request #26560 from yuwata/journal-send-fix-memleak-and-cleanups
Yu Watanabe [Thu, 23 Feb 2023 12:37:20 +0000 (21:37 +0900)] 
Merge pull request #26560 from yuwata/journal-send-fix-memleak-and-cleanups

sd-journal: several fixlets and cleanups for sd_journal_send()

2 years agocore/dbus-socket: check the socket path is absolute
Yu Watanabe [Thu, 23 Feb 2023 09:03:50 +0000 (18:03 +0900)] 
core/dbus-socket: check the socket path is absolute

In config_parse_socket_listen(), we have checked the path is absolute,
however we have not in the dbus method.

2 years agoexecute: pass ambient caps from PAM through to invoked service
Lennart Poettering [Wed, 22 Feb 2023 12:06:29 +0000 (13:06 +0100)] 
execute: pass ambient caps from PAM through to invoked service

If a PAM service sets some ambient caps, we should honour that, hence
query it, and merge it with our own ambient settings.

This needs to be done manually since otherwise dropping privs via
setresuid() will undo all such caps, and we need to manually tweak
things to keep them.

2 years agoMerge pull request #26563 from dtardon/fd-init
Zbigniew Jędrzejewski-Szmek [Thu, 23 Feb 2023 10:43:09 +0000 (11:43 +0100)] 
Merge pull request #26563 from dtardon/fd-init

Initialize fds with -EBADF + drop one unneeded initialization

2 years agoMerge pull request #26556 from keszybz/list-timers-optimize-left
Zbigniew Jędrzejewski-Szmek [Thu, 23 Feb 2023 10:37:48 +0000 (11:37 +0100)] 
Merge pull request #26556 from keszybz/list-timers-optimize-left

Do not say " left" on every timer in LEFT column

2 years agosystemctl: cleanups for list-paths 26535/head
Yu Watanabe [Wed, 22 Feb 2023 07:28:43 +0000 (16:28 +0900)] 
systemctl: cleanups for list-paths

- Rename struct path_info -> PathInfo,
- Drop struct path_infos,
- Use CLEANUP_ARRAY(),

No functional change, just refactoring.

2 years agosystemctl: cleanups for list-automounts
Yu Watanabe [Wed, 22 Feb 2023 07:12:29 +0000 (16:12 +0900)] 
systemctl: cleanups for list-automounts

- Rename struct automount_info -> AutomountInfo,
- use CLEANUP_ARRAY(), and etc.

No functional change, just refactoring.

2 years agosystemctl: cleanups for list-timers
Yu Watanabe [Wed, 22 Feb 2023 06:57:33 +0000 (15:57 +0900)] 
systemctl: cleanups for list-timers

- Rename struct timer_info -> TimerInfo,
- use CLEANUP_ARRAY().

No functional change, just refactoring.

2 years agosystemctl: cleanups for list-sockets
Yu Watanabe [Wed, 22 Feb 2023 06:20:20 +0000 (15:20 +0900)] 
systemctl: cleanups for list-sockets

- Rename struct socket_info -> SocketInfo.
- Drop own_triggered field, and always copy the triggered field.
- Use CLEANUP_ARRAY().

No functional changes, just refactoring.

2 years agovmm: Modernize get_smbios_table() 26258/head
Daan De Meyer [Wed, 22 Feb 2023 16:04:58 +0000 (17:04 +0100)] 
vmm: Modernize get_smbios_table()

2 years agostub: Read extra kernel command line items from SMBIOS
Daan De Meyer [Mon, 30 Jan 2023 15:26:50 +0000 (16:26 +0100)] 
stub: Read extra kernel command line items from SMBIOS

Let's read more kernel command line arguments from SMBIOS OEM string
io.systemd.stub.kernel-cmdline-extra. This allows adding debug kernel
command line arguments when booting in qemy without having to modify
the UKI.

2 years agobootctl: drop unneeded initialization 26563/head
David Tardon [Thu, 23 Feb 2023 08:50:26 +0000 (09:50 +0100)] 
bootctl: drop unneeded initialization

2 years agovmm: Add smbios_find_oem_string()
Daan De Meyer [Mon, 30 Jan 2023 15:26:14 +0000 (16:26 +0100)] 
vmm: Add smbios_find_oem_string()

This function can be used to find SMBIOS strings in the SMBIOS Type 11
table.

2 years agovmm: Add more const
Daan De Meyer [Mon, 30 Jan 2023 20:15:12 +0000 (21:15 +0100)] 
vmm: Add more const

SMBIOS tables are immutable, so let's access it via const pointers
where possible.

2 years agoefi-string: Add efi_memchr()
Daan De Meyer [Tue, 31 Jan 2023 14:39:40 +0000 (15:39 +0100)] 
efi-string: Add efi_memchr()

2 years agotree-wide: initialize fds with -EBADF
David Tardon [Thu, 23 Feb 2023 08:48:54 +0000 (09:48 +0100)] 
tree-wide: initialize fds with -EBADF

2 years agoMerge pull request #26458 from yuwata/core-network-namespace-remount-sysfs
Lennart Poettering [Thu, 23 Feb 2023 08:44:06 +0000 (09:44 +0100)] 
Merge pull request #26458 from yuwata/core-network-namespace-remount-sysfs

core: remount sysfs when network and mount namespace are enabled

2 years agosd-event: fix error handling
Yu Watanabe [Wed, 22 Feb 2023 22:31:01 +0000 (07:31 +0900)] 
sd-event: fix error handling

Follow-up for 6d2326e036ceed30f9ccdb0266713c10a44dcf6c.

2 years agoMerge pull request #26559 from yuwata/sd-event-memory-pressure-follow-ups
Lennart Poettering [Thu, 23 Feb 2023 08:27:20 +0000 (09:27 +0100)] 
Merge pull request #26559 from yuwata/sd-event-memory-pressure-follow-ups

sd-event: follow-ups for memory pressure event source

2 years agosystemctl-list-units: drop unused return value
Yu Watanabe [Wed, 22 Feb 2023 07:38:36 +0000 (16:38 +0900)] 
systemctl-list-units: drop unused return value

2 years agosystemctl-list-units: fix memleak on error
Yu Watanabe [Wed, 22 Feb 2023 07:38:18 +0000 (16:38 +0900)] 
systemctl-list-units: fix memleak on error

2 years agosd-journal: fix memleak and freeing invalid pointers 26560/head
Yu Watanabe [Thu, 23 Feb 2023 02:39:42 +0000 (11:39 +0900)] 
sd-journal: fix memleak and freeing invalid pointers

This also makes
- use GREEDY_REALLOC() or GREEDY_REALLOC0(),
- use CLEANUP_ARRAY() macro.

2 years agoio-util: drop double evaluation in IOVEC_INIT_STRING()
Yu Watanabe [Thu, 23 Feb 2023 04:10:31 +0000 (13:10 +0900)] 
io-util: drop double evaluation in IOVEC_INIT_STRING()

This also makes IOVEC_INIT_STRING() and IOVEC_MAKE_STRING() identical.

2 years agoio-util: introduce IOVEC_NULL
Yu Watanabe [Thu, 23 Feb 2023 04:09:46 +0000 (13:09 +0900)] 
io-util: introduce IOVEC_NULL

2 years agoio-util: introduce iovec_array_free()
Lennart Poettering [Thu, 23 Feb 2023 02:42:11 +0000 (11:42 +0900)] 
io-util: introduce iovec_array_free()

It will be used in later commits.

2 years agotest-execute: add test for PrivateNetwork= with/without mount namespacing 26458/head
Yu Watanabe [Fri, 17 Feb 2023 05:26:29 +0000 (14:26 +0900)] 
test-execute: add test for PrivateNetwork= with/without mount namespacing

2 years agocore/namespace: mount new sysfs when new network namespace is requested
Yu Watanabe [Fri, 17 Feb 2023 04:50:17 +0000 (13:50 +0900)] 
core/namespace: mount new sysfs when new network namespace is requested

Even when a mount namespace is created, previously host's sysfs is used,
especially with RootDirectory= or RootImage=, thus service processes can
still access the properties of the network interfaces in the main network
namespace through sysfs.

This makes, sysfs is remounted with the new network namespace tag, except
when PrivateMounts= is explicitly disabled. Hence, the properties of the
network interfaces in the main network namespace cannot be accessed by
service processes through sysfs.

Fixes #26422.

2 years agocore/namespace: introduce a new namespace mount mode PRIVATE_SYSFS
Yu Watanabe [Sat, 18 Feb 2023 00:44:43 +0000 (09:44 +0900)] 
core/namespace: introduce a new namespace mount mode PRIVATE_SYSFS

This is useful when a service running with a new network namespace.

The mount mode is not used yet, but will be used in a later commit.

2 years agocore/namespace: rename SYSFS -> BIND_SYSFS
Yu Watanabe [Fri, 17 Feb 2023 22:12:37 +0000 (07:12 +0900)] 
core/namespace: rename SYSFS -> BIND_SYSFS

No functional change, just preparation for later commits.