platform/upstream/systemd.git
11 months agotest: reenable TEST-30 on i*86
Frantisek Sumsal [Mon, 4 Dec 2023 09:57:39 +0000 (10:57 +0100)]
test: reenable TEST-30 on i*86

The original reason for deny-listing it was that it's flaky there. I'm
not sure if that's still the case, but the Ubuntu CI jobs for i*86 are
gone, so this file shouldn't be needed anymore anyway.

11 months agotest: skip TEST-36 on s390x and powerpc
Frantisek Sumsal [Mon, 4 Dec 2023 09:51:35 +0000 (10:51 +0100)]
test: skip TEST-36 on s390x and powerpc

As QEMU there doesn't support NUMA nodes. Also, drop the now unneeded
deny-list file for Ubuntu CI.

11 months agotest: check if 'btrfs filesystem' supports 'mkswapfile'
Frantisek Sumsal [Mon, 4 Dec 2023 09:23:42 +0000 (10:23 +0100)]
test: check if 'btrfs filesystem' supports 'mkswapfile'

Instead of deny-listing the test on Ubuntu CI, so it gets enabled
automagically once btrfs-progs is updated to a newer version there.

11 months agotest: reenable TEST-25-IMPORT in Ubuntu CI
Frantisek Sumsal [Mon, 4 Dec 2023 09:16:02 +0000 (10:16 +0100)]
test: reenable TEST-25-IMPORT in Ubuntu CI

Let's see if #13973 is still an issue.

11 months agoexecutor: apply LogLevelMax earlier
Luca Boccassi [Sun, 3 Dec 2023 18:30:07 +0000 (18:30 +0000)]
executor: apply LogLevelMax earlier

SELinux logs before we have a chance to apply it, move it up as it
breaks TEST-04-JOURNAL:

[  408.578624] testsuite-04.sh[11463]: ++ journalctl -b -q -u silent-success.service
[  408.578743] testsuite-04.sh[11098]: + [[ -z Dec 03 13:38:41 H systemd-executor[11459]: SELinux enabled state cached to: disabled ]]

Follow-up for: bb5232b6a3b8a

11 months agoMerge pull request #30305 from yuwata/seccomp-fix
Yu Watanabe [Mon, 4 Dec 2023 02:44:58 +0000 (11:44 +0900)]
Merge pull request #30305 from yuwata/seccomp-fix

seccomp: override the default action only when the filter is allow-list

11 months agotests: fix section mapping in test_ukify.py
Zbigniew Jędrzejewski-Szmek [Sat, 2 Dec 2023 09:26:05 +0000 (10:26 +0100)]
tests: fix section mapping in test_ukify.py

The regexp only worked if the sections were small enough for the size to
start with "0". I have an initrd that is 0x1078ec7e bytes, so the tests
would spuriously fail.

11 months agoMerge pull request #30297 from keszybz/fixups
Luca Boccassi [Sun, 3 Dec 2023 12:34:24 +0000 (12:34 +0000)]
Merge pull request #30297 from keszybz/fixups

A few unrelated fixups for recent commits

11 months agoseccomp-util: also use ENOSYS for unknown syscalls in seccomp_load_syscall_filter_set()
Yu Watanabe [Sun, 3 Dec 2023 09:01:20 +0000 (18:01 +0900)]
seccomp-util: also use ENOSYS for unknown syscalls in seccomp_load_syscall_filter_set()

Follow-up for 2331c02d06cae97b87637a0fc6bb4961b509ccf2.

Note, currently, the function is always called with SCMP_ACT_ALLOW as
the default action, except for the test. So, this should not change
anything in the runtime code.

11 months agoseccomp-util: override default action only when the filter is allow-list
Yu Watanabe [Sun, 3 Dec 2023 08:37:02 +0000 (17:37 +0900)]
seccomp-util: override default action only when the filter is allow-list

Follow-up for 2331c02d06cae97b87637a0fc6bb4961b509ccf2.

Fixes #30304.

11 months agostdio-bridge: return immediately if we can
Zbigniew Jędrzejewski-Szmek [Wed, 29 Nov 2023 13:38:06 +0000 (14:38 +0100)]
stdio-bridge: return immediately if we can

Follow-up for 0321248b79d14ceddd36140b327332f145ae68e7.

11 months agorun: fix bad escaping and memory ownership confusion
Zbigniew Jędrzejewski-Szmek [Wed, 29 Nov 2023 13:13:33 +0000 (14:13 +0100)]
run: fix bad escaping and memory ownership confusion

arg_description was either set to arg_unit (i.e. a const char*), or to
char *description, the result of allocation in run(). But description
was decorated with _cleanup_, so it would be freed when going out of the
function. Nothing bad would happen, because the program would exit after
exiting from run(), but this is just all too messy.

Also, strv_join(" ") + shell_escape() is not a good way to escape command
lines. In particular, one the join has happened, we cannot distinguish
empty arguments, or arguments with whitespace, etc. We have a helper
function to do the escaping properly, so let's use that.

Fixup for 2c29813da3421b77eca5e5cdc3b9a863cad473b9.

11 months agorun: adjust indentation
Zbigniew Jędrzejewski-Szmek [Wed, 29 Nov 2023 13:01:13 +0000 (14:01 +0100)]
run: adjust indentation

11 months agotest-macro: use capital test names for macro tests
Zbigniew Jędrzejewski-Szmek [Sun, 26 Nov 2023 14:54:11 +0000 (15:54 +0100)]
test-macro: use capital test names for macro tests

It's just easier if the test is named as the thing being tested. Also, this way
inconsistent, because lower in the file uppercase test names are used.

11 months agoNEWS: finalize for v255-rc4
Luca Boccassi [Sat, 2 Dec 2023 02:05:27 +0000 (02:05 +0000)]
NEWS: finalize for v255-rc4

11 months agoMerge pull request #30291 from keszybz/seccomp-unknown-syscall
Luca Boccassi [Sat, 2 Dec 2023 02:04:24 +0000 (02:04 +0000)]
Merge pull request #30291 from keszybz/seccomp-unknown-syscall

Backwardscompatibly handle syscalls unknown to us or libseccomp

11 months agocore: turn on higher optimization level in seccomp
Zbigniew Jędrzejewski-Szmek [Fri, 1 Dec 2023 22:13:54 +0000 (23:13 +0100)]
core: turn on higher optimization level in seccomp

This mirrors what d75615f398a0fbf986cf03924462863ca6ee2f9f did for nspawn.

It isn't really a fatal failure if we can't set that, so ignore it in libseccomp
cannot set the attribute.

 line  OP   JT   JF   K
=================================
 0000: 0x20 0x00 0x00 0x00000004   ld  $data[4]
 0001: 0x15 0x00 0xb7 0x40000003   jeq 1073741827 true:0002 false:0185
 0002: 0x20 0x00 0x00 0x00000000   ld  $data[0]
 0003: 0x15 0xb5 0x00 0x00000000   jeq 0    true:0185 false:0004
 0004: 0x15 0xb4 0x00 0x00000001   jeq 1    true:0185 false:0005
 0005: 0x15 0xb3 0x00 0x00000002   jeq 2    true:0185 false:0006
 0006: 0x15 0xb2 0x00 0x00000003   jeq 3    true:0185 false:0007
 0007: 0x15 0xb1 0x00 0x00000004   jeq 4    true:0185 false:0008
 0008: 0x15 0xb0 0x00 0x00000005   jeq 5    true:0185 false:0009
 0009: 0x15 0xaf 0x00 0x00000006   jeq 6    true:0185 false:0010
 ...
 0438: 0x15 0x03 0x00 0x000001be   jeq 446  true:0442 false:0439
 0439: 0x15 0x02 0x00 0x000001bf   jeq 447  true:0442 false:0440
 0440: 0x15 0x01 0x00 0x000001c0   jeq 448  true:0442 false:0441
 0441: 0x06 0x00 0x00 0x00050026   ret ERRNO(38)
 0442: 0x06 0x00 0x00 0x7fff0000   ret ALLOW

 line  OP   JT   JF   K
=================================
 0000: 0x20 0x00 0x00 0x00000004   ld  $data[4]
 0001: 0x15 0x00 0x27 0x40000003   jeq 1073741827 true:0002 false:0041
 0002: 0x20 0x00 0x00 0x00000000   ld  $data[0]
 0003: 0x25 0x01 0x00 0x000000b5   jgt 181  true:0005 false:0004
 0004: 0x05 0x00 0x00 0x00000143   jmp 0328
 0005: 0x25 0x00 0xa1 0x00000139   jgt 313  true:0006 false:0167
 0006: 0x25 0x00 0x51 0x00000179   jgt 377  true:0007 false:0088
 0007: 0x25 0x00 0x29 0x000001a0   jgt 416  true:0008 false:0049
 0008: 0x25 0x00 0x13 0x000001b0   jgt 432  true:0009 false:0028
 0009: 0x25 0x00 0x09 0x000001b8   jgt 440  true:0010 false:0019
 ...
 0551: 0x15 0x03 0x00 0x00000002   jeq 2    true:0555 false:0552
 0552: 0x15 0x02 0x01 0x00000001   jeq 1    true:0555 false:0554
 0553: 0x15 0x01 0x00 0x00000000   jeq 0    true:0555 false:0554
 0554: 0x06 0x00 0x00 0x00050026   ret ERRNO(38)
 0555: 0x06 0x00 0x00 0x7fff0000   ret ALLOW

The program is longer but hopefully faster because of the binary search.

11 months agoshared/seccomp-util: use the same error message for the same condition
Zbigniew Jędrzejewski-Szmek [Fri, 1 Dec 2023 18:04:27 +0000 (19:04 +0100)]
shared/seccomp-util: use the same error message for the same condition

We were calling seccomp_syscall_resolve_name three times and using a
slightly different error message in each of the cases.

11 months agocore: when applying syscall filters, use ENOSYS for unknown calls
Zbigniew Jędrzejewski-Szmek [Fri, 1 Dec 2023 18:03:23 +0000 (19:03 +0100)]
core: when applying syscall filters, use ENOSYS for unknown calls

glibc starting using fchmodat2 to implement fchmod with flags [1], but
current version of libseccomp does not support fchmodat2 [2]. This is
causing problems with programs sandboxed by systemd. libseccomp needs to know
a syscall to be able to set any kind of filter for it, so for syscalls unknown
by libseccomp we would always do the default action, i.e. either return the
errno set by SystemCallErrorNumber or send a fatal signal. For glibc to ignore
the unknown syscall and gracefully fall back to the older implementation,
we need to return ENOSYS. In particular, tar now fails with the default
SystemCallFilter="@system-service" sandbox [3].

This is of course a wider problem: any time the kernel gains new syscalls,
before libseccomp and systemd have caught up, we'd behave incorrectly. Let's
do the same as we already were doing in nspawn since
3573e032f26724949e86626eace058d006b8bf70, and do the "default action" only
for syscalls which are known by us and libseccomp, and return ENOSYS for
anything else. This means that users can start using a sandbox with the new
syscalls only after libseccomp and systemd have been updated, but before that
happens they behaviour that is backwards-compatible.

[1] https://github.com/bminor/glibc/commit/65341f7bbea824d2ff9d37db15d8be162df42bd3
[2] https://github.com/seccomp/libseccomp/issues/406
[2] https://github.com/systemd/systemd/issues/30250

Fixes https://github.com/systemd/systemd/issues/30250.

In seccomp_restrict_sxid() there's a chunk conditionalized with
'#if defined(__SNR_fchmodat2)'. We need to kep that because seccomp_restrict_sxid()
seccomp_restrict_suid_sgid() uses SCMP_ACT_ALLOW as the default action.

11 months agocore/cgroup: for non-cached attrs, don't return ENODATA blindly
Mike Yuan [Fri, 1 Dec 2023 19:06:16 +0000 (03:06 +0800)]
core/cgroup: for non-cached attrs, don't return ENODATA blindly

Follow-up for f17b07f4d72238da95312920dcc2ad076568cba3

Hope I won't break this thing again...

11 months agoMerge pull request #30294 from bluca/news
Zbigniew Jędrzejewski-Szmek [Sat, 2 Dec 2023 00:08:45 +0000 (01:08 +0100)]
Merge pull request #30294 from bluca/news

hwdb and NEWS updates

11 months agosystemd-homed.service.in: add quotactl to SystemCallFilter
Neil Wilson [Fri, 1 Dec 2023 16:29:32 +0000 (16:29 +0000)]
systemd-homed.service.in: add quotactl to SystemCallFilter

Standard directories make a call to the quotactl system call to enforce disk size limits.

Fixes #30287

11 months agohwdb: update
Luca Boccassi [Fri, 1 Dec 2023 22:07:08 +0000 (22:07 +0000)]
hwdb: update

11 months agoNEWS: update contributors list
Luca Boccassi [Fri, 1 Dec 2023 22:05:08 +0000 (22:05 +0000)]
NEWS: update contributors list

11 months agoMerge pull request #30268 from yuwata/network-fix-too-many-waiting-replies
Zbigniew Jędrzejewski-Szmek [Fri, 1 Dec 2023 18:54:18 +0000 (19:54 +0100)]
Merge pull request #30268 from yuwata/network-fix-too-many-waiting-replies

network: fix issue caused by too many waiting replies

11 months agocore: fix comment
Zbigniew Jędrzejewski-Szmek [Fri, 1 Dec 2023 14:53:06 +0000 (15:53 +0100)]
core: fix comment

11 months agofix: prefix of dmesg pstore files
Дамјан Георгиевски [Fri, 1 Dec 2023 10:46:36 +0000 (11:46 +0100)]
fix: prefix of dmesg pstore files

A change in the kernel[1] renamed the prefix of the pstore files from
`dmesg-efi-` to `dmesg-efi_pstore-`.

[1]
https://git.kernel.org/linus/893c5f1de620

11 months agohomework-quota.c: correct error message in home_update_quota_btrfs
Neil Wilson [Fri, 1 Dec 2023 16:32:18 +0000 (16:32 +0000)]
homework-quota.c: correct error message in home_update_quota_btrfs

Fixes #30286

11 months agobuild(deps): bump meson from 1.2.3 to 1.3.0 in /.github/workflows
dependabot[bot] [Fri, 1 Dec 2023 10:01:04 +0000 (10:01 +0000)]
build(deps): bump meson from 1.2.3 to 1.3.0 in /.github/workflows

Bumps [meson](https://github.com/mesonbuild/meson) from 1.2.3 to 1.3.0.
- [Release notes](https://github.com/mesonbuild/meson/releases)
- [Commits](https://github.com/mesonbuild/meson/compare/1.2.3...1.3.0)

---
updated-dependencies:
- dependency-name: meson
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>
11 months agobuild(deps): bump redhat-plumbers-in-action/differential-shellcheck
dependabot[bot] [Fri, 1 Dec 2023 09:04:01 +0000 (09:04 +0000)]
build(deps): bump redhat-plumbers-in-action/differential-shellcheck

Bumps [redhat-plumbers-in-action/differential-shellcheck](https://github.com/redhat-plumbers-in-action/differential-shellcheck) from 5.0.1 to 5.0.2.
- [Release notes](https://github.com/redhat-plumbers-in-action/differential-shellcheck/releases)
- [Changelog](https://github.com/redhat-plumbers-in-action/differential-shellcheck/blob/main/docs/CHANGELOG.md)
- [Commits](https://github.com/redhat-plumbers-in-action/differential-shellcheck/compare/aa647ec4466543e8555c2c3b648124a9813cee44...91e2582e40236f831458392d905578d680baa138)

---
updated-dependencies:
- dependency-name: redhat-plumbers-in-action/differential-shellcheck
  dependency-type: direct:production
  update-type: version-update:semver-patch
...

Signed-off-by: dependabot[bot] <support@github.com>
11 months agobuild(deps): bump redhat-plumbers-in-action/devel-freezer
dependabot[bot] [Fri, 1 Dec 2023 09:03:54 +0000 (09:03 +0000)]
build(deps): bump redhat-plumbers-in-action/devel-freezer

Bumps [redhat-plumbers-in-action/devel-freezer](https://github.com/redhat-plumbers-in-action/devel-freezer) from 1.0.7 to 1.0.8.
- [Release notes](https://github.com/redhat-plumbers-in-action/devel-freezer/releases)
- [Commits](https://github.com/redhat-plumbers-in-action/devel-freezer/compare/13b6551f19ade74ca79be4cab06b815a4ffffa64...67aec4a153bd9fca5322e1c4dd4d7c419fb36362)

---
updated-dependencies:
- dependency-name: redhat-plumbers-in-action/devel-freezer
  dependency-type: direct:production
  update-type: version-update:semver-patch
...

Signed-off-by: dependabot[bot] <support@github.com>
11 months agobuild(deps): bump actions/github-script from 6.4.1 to 7.0.1
dependabot[bot] [Fri, 1 Dec 2023 09:03:48 +0000 (09:03 +0000)]
build(deps): bump actions/github-script from 6.4.1 to 7.0.1

Bumps [actions/github-script](https://github.com/actions/github-script) from 6.4.1 to 7.0.1.
- [Release notes](https://github.com/actions/github-script/releases)
- [Commits](https://github.com/actions/github-script/compare/d7906e4ad0b1822421a7e6a35d5ca353c962f410...60a0d83039c74a4aee543508d2ffcb1c3799cdea)

---
updated-dependencies:
- dependency-name: actions/github-script
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
11 months agobuild(deps): bump github/codeql-action from 2.21.9 to 2.22.8
dependabot[bot] [Fri, 1 Dec 2023 09:03:50 +0000 (09:03 +0000)]
build(deps): bump github/codeql-action from 2.21.9 to 2.22.8

Bumps [github/codeql-action](https://github.com/github/codeql-action) from 2.21.9 to 2.22.8.
- [Release notes](https://github.com/github/codeql-action/releases)
- [Changelog](https://github.com/github/codeql-action/blob/main/CHANGELOG.md)
- [Commits](https://github.com/github/codeql-action/compare/ddccb873888234080b77e9bc2d4764d5ccaaccf9...407ffafae6a767df3e0230c3df91b6443ae8df75)

---
updated-dependencies:
- dependency-name: github/codeql-action
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>
11 months agovconsole-setup: use a consistent log level when setfont fails with EX_OSERR
Franck Bui [Fri, 1 Dec 2023 07:44:53 +0000 (08:44 +0100)]
vconsole-setup: use a consistent log level when setfont fails with EX_OSERR

Since we accept that setfont(8) can fail with EX_OSERR and we dont treat it as
an error, dont log this failure at LOG_ERR.

Before:
-------
/usr/bin/setfont failed with exit status 71.           [LOG_ERR]
Setting fonts failed with a "system error", ignoring.  [LOG_NOTICE]

After:
-----
/usr/bin/setfont failed with a "system error" (EX_OSERR), ignoring.   [LOG_NOTICE]
Setting source virtual console failed, ignoring remaining ones        [LOG_NOTICE]

Follow-up for 93c9a9d235e2304500c490b4868534385f925c76

11 months agoukify: avoid deprecated datetime call
Zbigniew Jędrzejewski-Szmek [Fri, 1 Dec 2023 09:36:04 +0000 (10:36 +0100)]
ukify: avoid deprecated datetime call

DeprecationWarning: datetime.datetime.utcnow() is deprecated and scheduled for
                    removal in a future version. Use timezone-aware objects to
                    represent datetimes in UTC: datetime.datetime.now(datetime.UTC).

The difference between the two is that .now(datetime.UTC) returns an object with
a timezone attached, "the numbers" are the same.

>>> datetime.datetime.utcnow(), datetime.datetime.now(datetime.UTC)
(datetime.datetime(2023, 12, 1, 9, 37, 53, 891669),
 datetime.datetime(2023, 12, 1, 9, 37, 53, 891688, tzinfo=datetime.timezone.utc))

This value is fed to cryptography's x509.CertificateBuilder object, so as long
as it can accept a datetime object with tzinfo, the result should be identical.

11 months agocore: do not drop CAP_SETUID if it is in AmbientCapabilities=
Luca Boccassi [Fri, 1 Dec 2023 01:44:54 +0000 (01:44 +0000)]
core: do not drop CAP_SETUID if it is in AmbientCapabilities=

Follow-up for 24832d10b604848cf46624bb439c7fac27f3ce3f

11 months agonetwork: do not send too many netlink messages in a single event
Yu Watanabe [Thu, 30 Nov 2023 10:07:36 +0000 (19:07 +0900)]
network: do not send too many netlink messages in a single event

Fixes #26743.

11 months agofirewall-util: introduce fw_ctx_get_reply_callback_count()
Yu Watanabe [Fri, 1 Dec 2023 05:36:30 +0000 (14:36 +0900)]
firewall-util: introduce fw_ctx_get_reply_callback_count()

11 months agosd-netlink: introduce netlink_get_reply_callback_count()
Yu Watanabe [Thu, 30 Nov 2023 10:04:37 +0000 (19:04 +0900)]
sd-netlink: introduce netlink_get_reply_callback_count()

11 months agoMerge pull request #30211 from yuwata/sd-journal-generic-array-bisect-fix
Daan De Meyer [Fri, 1 Dec 2023 10:43:04 +0000 (11:43 +0100)]
Merge pull request #30211 from yuwata/sd-journal-generic-array-bisect-fix

sd-journal: fix generic_array_bisect()

11 months agosd-netlink: change error code of the case that too many replies waiting
Yu Watanabe [Fri, 1 Dec 2023 05:17:58 +0000 (14:17 +0900)]
sd-netlink: change error code of the case that too many replies waiting

ERANGE should be used when setting or parsing a number.

11 months agohwdb: ieee1394-unit-function: arrangement for Sony DVMC-DA1
Takashi Sakamoto [Wed, 29 Nov 2023 13:39:50 +0000 (22:39 +0900)]
hwdb: ieee1394-unit-function: arrangement for Sony DVMC-DA1

A commit 6a42bdb37e39 ("hwdb: ieee1394-unit-function: add Sony
DVMC-DA1") is based on kernel feature unreleased yet (furthermore, not
merged yet). The original intension of new entry is to configure permission
of special file for FireWire character device, so this commit changes the
entry so that it can covers the issued case in existent version of Linux
kernel as out best effort.

When the new version of Linux kernel is released with the new feature,
then following commits would fulfill the hwdb with vendor and model names.

11 months agotest: add tests for generic_array_bisect()
Yu Watanabe [Wed, 29 Nov 2023 04:38:36 +0000 (13:38 +0900)]
test: add tests for generic_array_bisect()

11 months agotest: make append_number() optionally return offset of the entry object
Yu Watanabe [Wed, 29 Nov 2023 04:37:53 +0000 (13:37 +0900)]
test: make append_number() optionally return offset of the entry object

Preparation for later commits.

11 months agotest: split out finalization task
Yu Watanabe [Tue, 28 Nov 2023 21:08:11 +0000 (06:08 +0900)]
test: split out finalization task

11 months agosd-journal: fix corrupted journal handling of generic_array_bisect()
Yu Watanabe [Mon, 27 Nov 2023 02:55:49 +0000 (11:55 +0900)]
sd-journal: fix corrupted journal handling of generic_array_bisect()

Let's consider the following case:
- the direction is down,
- no cached entry,
- the array has 5 entry objects,
- the function test_object() reutns TEST_LEFT for the 1st object,
- the 2nd, 3rd, and 4th objects are broken, so generic_array_bisect_step()
  returns TEST_RIGHT for the object.

Then, previously, generic_array_bisect_step() updated the values like the following:
  0th: (m = 5, left = 0, right = 4, i = 4) -> (m = 4, left = 0, right = 3, RIGHT)
  1st: (m = 4, left = 0, right = 3, i = 1) -> (m = 4, left = 2, right = 3, LEFT)
  2nd: (m = 4, left = 2, right = 3, i = 2) -> (m = 2, left = 2, right = 1, RIGHT) <- ouch!!
So, assert(left < right) in generic_array_bisect() was triggered.
See issue #30210.

In such situation, there is no matching entry in the array. By returning
TEST_GOTO_PREVIOUS, generic_array_bisect() handles the result so.

Fixes a bug introduced by ab8f553d1e09088fb1f633e014299e7bf6c30c9e.

Fixes #30210.

11 months agosd-journal: ignore failure in testing cached corrupted entry
Yu Watanabe [Wed, 29 Nov 2023 21:46:21 +0000 (06:46 +0900)]
sd-journal: ignore failure in testing cached corrupted entry

Let's consider the case that the 1st entry in an array is broken, but
n-th entry is valid. Then, if generic_array_get() is called to read
n-th object, the offset of the broken entry is cached by the function.
If generic_array_bisect() is followed, even if the matching entry is
valid, it always fail with -EBADMSG or friends, as the function test the
cached entry at the beginnning. Let's ignore the failure in testing the
cached entry.

11 months agoman: document how to properly use a target as the Unit= of a timer
Joerg Behrmann [Thu, 30 Nov 2023 10:01:16 +0000 (11:01 +0100)]
man: document how to properly use a target as the Unit= of a timer

11 months agoMake sure we close bpf outer map fd in systemd-executor
Daan De Meyer [Thu, 30 Nov 2023 10:01:14 +0000 (11:01 +0100)]
Make sure we close bpf outer map fd in systemd-executor

Not doing so leaks it into the child service and causes selinux
denials.

11 months agocontrib: include co-authors to the contributor list of NEWS
Yu Watanabe [Thu, 30 Nov 2023 02:55:27 +0000 (11:55 +0900)]
contrib: include co-authors to the contributor list of NEWS

Closes #30261.

11 months agocore/executor: avoid double closing serialization fd
Mike Yuan [Wed, 29 Nov 2023 15:18:23 +0000 (23:18 +0800)]
core/executor: avoid double closing serialization fd

Before this commit, between fdopen() (in parse_argv()) and fdset_remove(),
the serialization fd is owned by both arg_serialization FILE stream and fdset.
Therefore, if something wrong happens between the two calls, or if --deserialize=
is specified more than once, we end up closing the serialization fd twice.
Normally this doesn't matter much, but I still think it's better to fix this.

Let's call fdset_new_fill() after parsing serialization fd hence.
We set the fd to CLOEXEC in parse_argv(), so it will be filtered
when the fdset is created.

While at it, also move fdset_new_fill() under the second log_open(), so
that we always log to the log target specified in arguments.

11 months agosd-journal: fix typo in function name
Yu Watanabe [Thu, 30 Nov 2023 04:05:04 +0000 (13:05 +0900)]
sd-journal: fix typo in function name

11 months agocore: Always call log_open() in systemd-executor
Daan De Meyer [Wed, 29 Nov 2023 14:57:46 +0000 (15:57 +0100)]
core: Always call log_open() in systemd-executor

log_setup() will open the console in systemd-executor because it's
not pid 1 and it's not connected to the journal. So if the log target
is later changed to kmsg, we have to reopen the log.

But since log_open() won't open the same log twice, let's just call it
unconditionally since it will be a noop if we try to reopen the same log.

This makes sure that systemd-executor will log to the log target passed
via --log-target= after parsing arguments.

11 months agotest: ignore gcov errors with RestrictFileSystems=
Frantisek Sumsal [Wed, 29 Nov 2023 12:36:54 +0000 (13:36 +0100)]
test: ignore gcov errors with RestrictFileSystems=

The errors are valid, since the file system is indeed not writable, but
we don't care about the missing coverage data in this case.

Follow-up to 4a43c2b3a1.

11 months agonetwork: Add missing comma
Daan De Meyer [Wed, 29 Nov 2023 20:18:31 +0000 (21:18 +0100)]
network: Add missing comma

11 months agoRevert "mkosi ci: enable jammy-proposed"
Luca Boccassi [Wed, 29 Nov 2023 16:01:21 +0000 (16:01 +0000)]
Revert "mkosi ci: enable jammy-proposed"

libsolv has migrated to jammy-updates, so we can disable the
proposed-updates repository again.

This reverts commit 48bfc6791dca3b130befe419fc07f15f515b291c.

11 months agomkosi: Drop building custom kernel logic
Daan De Meyer [Wed, 29 Nov 2023 13:46:20 +0000 (14:46 +0100)]
mkosi: Drop building custom kernel logic

Now that mkosi-kernel is a thing, this logic in systemd is just mostly
bitrotting since I just use mkosi-kernel these days. If I ever need to
hack on systemd and the kernel in tandem, I'll just add support for
building systemd to mkosi-kernel instead, so let's drop the support for
building a custom kernel in systemd's mkosi configuration.

11 months agomkosi: Install integritysetup on CentOS/Fedora
Daan De Meyer [Wed, 29 Nov 2023 09:29:06 +0000 (10:29 +0100)]
mkosi: Install integritysetup on CentOS/Fedora

Required for running integration tests

11 months agorules: go to the end of rules indeed when dm is suspended
janana [Wed, 29 Nov 2023 03:36:52 +0000 (11:36 +0800)]
rules: go to the end of rules indeed when dm is suspended

The previous patch 466266c does not make sense indeed, that is to say, if the SYSTEMD_READY is not recorded in the database, the GOTO="systemd_end" will not be applied.

The IMPORT{db} is actually a matching token, it returns false when there is no SYSTEMD_READY recorded in the database.

The previous patch 466266c tended to inherit the state of SYSTEMD_READY from the database and skip to the end of current rule file. But when the database does not contain SYSTEMD_READY, e.g., the dm-* is not set db_persistent during initrd and the database will be cleared after switching root, the following rules will still be applied not as expected.

11 months agostdio-bridge: properly handle org.freedesktop.DBus.Local.Disconnected signal
Lennart Poettering [Tue, 28 Nov 2023 22:29:50 +0000 (23:29 +0100)]
stdio-bridge: properly handle org.freedesktop.DBus.Local.Disconnected signal

Previously, we'd forward org.freedesktop.DBus.Local.Disconnected like
any other message to the other side. But that's not OK, as messages in
the org.freedesktop.DBus.Local.* namespace are supposed to never touch
the wire, and are synthetic messages that the library uses to
communicate with the app, but never with other apps.

dbus-daemon never cared, but dbus-broker complains about this, hence
clean this up.

See: #28514

11 months agostub: get_extra_dir() can return NULL
Lennart Poettering [Tue, 28 Nov 2023 20:21:37 +0000 (21:21 +0100)]
stub: get_extra_dir() can return NULL

If we have a device path that is not the usual file system directory
get_extra_dir() will return NULL. Handle that gracefully.

Fixes: #30228

11 months agoMerge pull request #30236 from DaanDeMeyer/mkosi
Daan De Meyer [Tue, 28 Nov 2023 22:39:41 +0000 (23:39 +0100)]
Merge pull request #30236 from DaanDeMeyer/mkosi

Update to mkosi v19

11 months agomkosi: pin CentOS8 kernel to working version
Daan De Meyer [Tue, 28 Nov 2023 22:19:44 +0000 (23:19 +0100)]
mkosi: pin CentOS8 kernel to working version

Newer kernels are affected by a regression that causes a kernel panic
on boot when using cgroupv2, so pin them for now. Can be reverted once
that problem is fixed.

11 months agohomed: properly handle operation exit statusses
Lennart Poettering [Tue, 28 Nov 2023 14:41:41 +0000 (15:41 +0100)]
homed: properly handle operation exit statusses

11 months agoshared/cryptsetup-util: build problematic code only in developer mode
Zbigniew Jędrzejewski-Szmek [Tue, 28 Nov 2023 16:28:06 +0000 (17:28 +0100)]
shared/cryptsetup-util: build problematic code only in developer mode

This code doesn't link when gcc+lld is used:

$ LDFLAGS=-fuse-ld=lld meson setup build-lld && ninja -C build-lld udevadm
...
ld.lld: error: src/shared/libsystemd-shared-255.a(libsystemd-shared-255.a.p/cryptsetup-util.c.o):
  symbol crypt_token_external_path@@ has undefined version
collect2: error: ld returned 1 exit status

As a work-around, restrict it to developer mode.

Closes https://github.com/systemd/systemd/issues/30218.

11 months agoman: typo, it's DHCP not DCHP in DHCPPrefixDelegation=
Дамјан Георгиевски [Tue, 28 Nov 2023 18:53:57 +0000 (19:53 +0100)]
man: typo, it's DHCP not DCHP in DHCPPrefixDelegation=

11 months agoUpdate to mkosi v19
Daan De Meyer [Tue, 28 Nov 2023 17:38:22 +0000 (18:38 +0100)]
Update to mkosi v19

- Use mkosi.images/ instead of mkosi.presets/
- Use the .chroot suffix to run scripts in the image
- Use BuildSources= match for the kernel build
- Move 10-systemd.conf to mkosi.conf and rely on mkosi.local.conf
  for local configuration

11 months agoboot: suffix SecureBoot message with newline
Lennart Poettering [Mon, 27 Nov 2023 11:50:09 +0000 (12:50 +0100)]
boot: suffix SecureBoot message with newline

Output is otherwise so weird, since this is the last log line seen for a
while typically, and if it doesn#t put the cursor back in the first
column it looks like something is incomplete and hanging. Hence do what
we always do: finish log messages with a newline.

11 months agopam-util: fix pam_syslog_pam_error() format string
Lennart Poettering [Tue, 28 Nov 2023 10:05:52 +0000 (11:05 +0100)]
pam-util: fix pam_syslog_pam_error() format string

To cut off a string at some position we must set the "precision" not the
field width in format strings.

This led to some assert()s being hit where they really should not be.

11 months agogpt-auto-generator: enable TPM2 unlocking in gpt-auto-generator
Lennart Poettering [Fri, 24 Nov 2023 13:35:43 +0000 (14:35 +0100)]
gpt-auto-generator: enable TPM2 unlocking in gpt-auto-generator

If we detect a TPM, let's also unlock the disk with it, if it has an
enrollment for that.

Fixes: #30176

11 months agoMerge pull request #30203 from yuwata/resolve-dump-cache
Luca Boccassi [Tue, 28 Nov 2023 15:35:11 +0000 (15:35 +0000)]
Merge pull request #30203 from yuwata/resolve-dump-cache

resolve: fix varlink message verification

11 months agoMerge pull request #30216 from poettering/loginctl-table-tweaks
Luca Boccassi [Tue, 28 Nov 2023 15:33:35 +0000 (15:33 +0000)]
Merge pull request #30216 from poettering/loginctl-table-tweaks

loginctl: minor tweaks to tabular output

11 months agoMerge pull request #30220 from yuwata/journald-vs-soft-reboot
Luca Boccassi [Tue, 28 Nov 2023 15:32:08 +0000 (15:32 +0000)]
Merge pull request #30220 from yuwata/journald-vs-soft-reboot

unit: stop journald before SIGKILL on soft-reboot

11 months agologinctl: show null fields for unset seat/tty fields of sessions
Lennart Poettering [Mon, 27 Nov 2023 14:10:42 +0000 (15:10 +0100)]
loginctl: show null fields for unset seat/tty fields of sessions

11 months agologinctl: set appropriate ersatz strings for all tables
Lennart Poettering [Mon, 27 Nov 2023 14:09:53 +0000 (15:09 +0100)]
loginctl: set appropriate ersatz strings for all tables

11 months agologind: align columns of a table
Lennart Poettering [Mon, 27 Nov 2023 14:00:12 +0000 (15:00 +0100)]
logind: align columns of a table

11 months agoutmp-wtmp: fix wrong suffix assignment on putting dead process
Yu Watanabe [Mon, 27 Nov 2023 23:45:31 +0000 (08:45 +0900)]
utmp-wtmp: fix wrong suffix assignment on putting dead process

Fixes a bug introduced by f1d553e9dfd56f95b7564dd20a0b56e6a0d6492c.

Fixes #30213.

11 months agocore/cgroup: fix compile error
Yu Watanabe [Tue, 28 Nov 2023 04:02:02 +0000 (13:02 +0900)]
core/cgroup: fix compile error

With gcc-13,
```
CFLAGS="-O3 -fno-semantic-interposition" meson setup build
```
triggers the following error:
```
../src/core/cgroup.c: In function ‘cgroup_context_dump’:
../src/core/cgroup.c:633:44: error: ‘%s’ directive argument is null [-Werror=format-overflow=]
  633 |                         "%sDeviceAllow: %s %s\n",
      |                                            ^~
cc1: some warnings being treated as errors
```

Fixes #30223.

11 months agotest: check journal files are not corrupted after soft-reboot
Yu Watanabe [Sun, 26 Nov 2023 00:25:22 +0000 (09:25 +0900)]
test: check journal files are not corrupted after soft-reboot

11 months agounit: make journald stopped on soft-reboot before broadcasting SIGKILL
Yu Watanabe [Mon, 27 Nov 2023 23:32:22 +0000 (08:32 +0900)]
unit: make journald stopped on soft-reboot before broadcasting SIGKILL

Workaround for #30195.

11 months agoman: also SIGKILL is sent on soft-reboot
Yu Watanabe [Sat, 25 Nov 2023 20:43:46 +0000 (05:43 +0900)]
man: also SIGKILL is sent on soft-reboot

11 months agovarlinkctl: add short comment
Yu Watanabe [Sat, 25 Nov 2023 20:34:10 +0000 (05:34 +0900)]
varlinkctl: add short comment

11 months agocapability-util: avoid false-positive use-of-uninitialized-value error
Yu Watanabe [Mon, 27 Nov 2023 23:25:06 +0000 (08:25 +0900)]
capability-util: avoid false-positive use-of-uninitialized-value error

fuzzers randomly fail with the following:
```
==172==WARNING: MemorySanitizer: use-of-uninitialized-value
    #0 0x7f41169cb39b in update_argv /work/build/../../src/systemd/src/basic/argv-util.c:96:13
    #1 0x7f41169cb39b in rename_process /work/build/../../src/systemd/src/basic/argv-util.c:210:16
    #2 0x7f4116b6824e in safe_fork_full /work/build/../../src/systemd/src/basic/process-util.c:1516:21
    #3 0x7f4116bffa36 in safe_fork /work/build/../../src/systemd/src/basic/process-util.h:191:16
    #4 0x7f4116bffa36 in parse_timestamp /work/build/../../src/systemd/src/basic/time-util.c:1047:13
    #5 0x4a61e6 in LLVMFuzzerTestOneInput /work/build/../../src/systemd/src/fuzz/fuzz-time-util.c:16:16
    #6 0x4c4a13 in fuzzer::Fuzzer::ExecuteCallback(unsigned char const*, unsigned long) /src/llvm-project/compiler-rt/lib/fuzzer/FuzzerLoop.cpp:611:15
    #7 0x4c41fa in fuzzer::Fuzzer::RunOne(unsigned char const*, unsigned long, bool, fuzzer::InputInfo*, bool, bool*) /src/llvm-project/compiler-rt/lib/fuzzer/FuzzerLoop.cpp:514:3
    #8 0x4c58c9 in fuzzer::Fuzzer::MutateAndTestOne() /src/llvm-project/compiler-rt/lib/fuzzer/FuzzerLoop.cpp:757:19
    #9 0x4c6595 in fuzzer::Fuzzer::Loop(std::__Fuzzer::vector<fuzzer::SizedFile, std::__Fuzzer::allocator<fuzzer::SizedFile> >&) /src/llvm-project/compiler-rt/lib/fuzzer/FuzzerLoop.cpp:895:5
    #10 0x4b58ff in fuzzer::FuzzerDriver(int*, char***, int (*)(unsigned char const*, unsigned long)) /src/llvm-project/compiler-rt/lib/fuzzer/FuzzerDriver.cpp:912:6
    #11 0x4def52 in main /src/llvm-project/compiler-rt/lib/fuzzer/FuzzerMain.cpp:20:10
    #12 0x7f4115ea3082 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x24082) (BuildId: e678fe54a5d2c2092f8e47eb0b33105e380f7340)
    #13 0x41f5ad in _start (build-out/fuzz-time-util+0x41f5ad)

DEDUP_TOKEN: update_argv--rename_process--safe_fork_full
  Uninitialized value was created by an allocation of 'fv' in the stack frame of function 'have_effective_cap'
    #0 0x7f41169d3540 in have_effective_cap /work/build/../../src/systemd/src/basic/capability-util.c:21
```

11 months agotest: verify json format passed through varlink
Yu Watanabe [Sun, 26 Nov 2023 21:26:07 +0000 (06:26 +0900)]
test: verify json format passed through varlink

This drops logs of failed unit tests in TEST-02-UNITTESTS from the journal,
as the expected logs from test-varlink-idl makes the post script fail.
Saving logs in journal is simply noisy, and we will output anyway after
qemu or nspawn finished by check_result_{qemu,nspawn}_unittests().

11 months agotest: show saved journal files after TEST-02-UNITTESTS
Yu Watanabe [Mon, 27 Nov 2023 00:42:15 +0000 (09:42 +0900)]
test: show saved journal files after TEST-02-UNITTESTS

11 months agotest: drop remaining console output
Yu Watanabe [Sun, 26 Nov 2023 20:59:56 +0000 (05:59 +0900)]
test: drop remaining console output

Follow-up for fa6f37c04391d0103c95e24813ad345c2d5c4b67.

Partially reverts d5ab23aa54229e9e89427e7d20883bd8c93fd976 and
3540ce8587cbd21ce9c2dbec72ea7fa3d1b38a5f.

11 months agotest: add simple coverity test for 'resolvectl show-cache'
Yu Watanabe [Sun, 26 Nov 2023 04:12:19 +0000 (13:12 +0900)]
test: add simple coverity test for 'resolvectl show-cache'

11 months agovarlink: fix key name in reply of org.varlink.service.GetInterfaceDescription
Yu Watanabe [Sun, 26 Nov 2023 20:41:22 +0000 (05:41 +0900)]
varlink: fix key name in reply of org.varlink.service.GetInterfaceDescription

Fixes https://github.com/systemd/systemd/pull/30203#issuecomment-1826746111.

11 months agoresolve: fix varink message verification
Yu Watanabe [Sun, 26 Nov 2023 04:10:15 +0000 (13:10 +0900)]
resolve: fix varink message verification

Fixes #30199.

11 months agotest-time-util: suppress timestamp conversion failures for Africa/Khartoum timezone
Zbigniew Jędrzejewski-Szmek [Sun, 26 Nov 2023 19:58:43 +0000 (20:58 +0100)]
test-time-util: suppress timestamp conversion failures for Africa/Khartoum timezone

Our timestamp conversion roundtrip test was failing. But I think that this
is not our bug:

$ TZ='Africa/Khartoum' date --date='@1509482094'
Tue Oct 31 23:34:54 EAT 2017
$ TZ='Africa/Khartoum' date --date='Tue Oct 31 23:34:54 EAT 2017' +%s
1509485694
$ TZ='Africa/Khartoum' date --date='@1509485694'
Tue Oct 31 23:34:54 CAT 2017
$ echo $[1509485694 - 1509482094]
3600

This is essentially the same as what happens in our test. After a round-trip, we
end up one hour ahead.

> For 1509482094632752, from the change log of tzdata:
>
> Release 2017c - 2017-10-20 14:49:34 -0700
>
>  Changes to future timestamps
>    Sudan will switch from +03 to +02 on 2017-11-01.

Fixes https://github.com/systemd/systemd/issues/28472.

11 months agoMerge pull request #30170 from bluca/exec_bpf_fd
Luca Boccassi [Mon, 27 Nov 2023 15:44:50 +0000 (15:44 +0000)]
Merge pull request #30170 from bluca/exec_bpf_fd

core: pass bpf_outer_map_fd to sd-executor only if RestrictFileSystems was set

11 months agohwdb: ieee1394-unit-function: add Sony DVMC-DA1
Adam Goldman [Sun, 26 Nov 2023 10:45:25 +0000 (02:45 -0800)]
hwdb: ieee1394-unit-function: add Sony DVMC-DA1

This commit adds a hwdb entry for the Sony DVMC-DA1. This media converter
works with video capture software such as dvgrab, but it doesn't support
the AV/C command set and doesn't match the general entry.

11 months agoMerge pull request #30200 from mrc0mmand/test-tweaks
Yu Watanabe [Sun, 26 Nov 2023 00:10:09 +0000 (09:10 +0900)]
Merge pull request #30200 from mrc0mmand/test-tweaks

A couple of assorted fixes

11 months agotest: drop 'noearly' from crypttab
Frantisek Sumsal [Sat, 25 Nov 2023 21:30:01 +0000 (22:30 +0100)]
test: drop 'noearly' from crypttab

It's a Debian thing, we don't support it.

[   30.639971] testsuite-64.sh[475]: + systemctl restart cryptsetup.target
[   30.697667] systemd-cryptsetup[3389]: Encountered unknown /etc/crypttab option 'noearly', ignoring.
[   30.700529] systemd-cryptsetup[3390]: Encountered unknown /etc/crypttab option 'noearly', ignoring.
[   30.700594] systemd-cryptsetup[3391]: Encountered unknown /etc/crypttab option 'noearly', ignoring.
[   30.704638] systemd-cryptsetup[3392]: Encountered unknown /etc/crypttab option 'noearly', ignoring

11 months agotest: drop invalid LogTarget= assignment
Frantisek Sumsal [Sat, 25 Nov 2023 20:59:57 +0000 (21:59 +0100)]
test: drop invalid LogTarget= assignment

It's been there since the test was introduced and I'm not really sure
what was the original intention behind it, but it makes systemd sad:

[    4.909056] systemd[1]: /usr/lib/systemd/tests/testdata/units/testsuite-44.service:13: Unknown key name 'LogTarget' in section 'Service', ignoring.

11 months agosystemctl: include unit ID in the property map
Frantisek Sumsal [Sat, 25 Nov 2023 20:36:10 +0000 (21:36 +0100)]
systemctl: include unit ID in the property map

Otherwise we get a not very nice message when trying to display a
non-existent man page:

~# systemctl cat test.service
[Unit]
Description=Hello
[Service]
ExecStart=true
~# systemctl help test.service
Documentation for (null) not known.

11 months agoimport: append % to X_IMPORT_PROGRESS=
Frantisek Sumsal [Sat, 25 Nov 2023 19:42:13 +0000 (20:42 +0100)]
import: append % to X_IMPORT_PROGRESS=

Since we parse it on the other side via parse_percent() which requires
that, otherwise we get an error:

[    8.133131] testsuite-13.sh[649]: + machinectl import-raw /tmp/container.raw container-raw
[    8.175035] machinectl[1143]: Enqueued transfer job 1. Press C-c to continue download in background.
[    8.182130] machinectl[1143]: Importing '/tmp/container.raw', saving as 'container-raw'.
[    8.182377] systemd-importd[1144]: Got invalid percent value '0', ignoring.
[    8.182451] machinectl[1143]: Imported 0%.
[    8.282669] systemd-importd[1144]: Got invalid percent value '40', ignoring.
[    8.282746] machinectl[1143]: Imported 40%.
[    8.366448] machinectl[1143]: Wrote 64.0M.
[    8.366519] machinectl[1143]: Operation completed successfully.
[    8.366617] machinectl[1143]: Exiting.

11 months agotest: remove the reboot counter just before spawning container/VM
Frantisek Sumsal [Sat, 25 Nov 2023 19:22:24 +0000 (20:22 +0100)]
test: remove the reboot counter just before spawning container/VM

Doing that in test_run() is not enough, since it's called only once and
spawns all QEMU/nspawn jobs.

11 months agoMerge pull request #30196 from YHNdnzj/fchmodat2-no-symlink
Yu Watanabe [Sat, 25 Nov 2023 18:30:58 +0000 (03:30 +0900)]
Merge pull request #30196 from YHNdnzj/fchmodat2-no-symlink

Revert "nspawn-patch-uid: try fchmodat2() to restore mode of symlink"

11 months agoMerge pull request #30197 from keszybz/vconsole-restart-no-limit
Yu Watanabe [Sat, 25 Nov 2023 17:28:06 +0000 (02:28 +0900)]
Merge pull request #30197 from keszybz/vconsole-restart-no-limit

Drop restart limit for systemd-vconsole-setup.service and some man page tweaks