Commit Graph
78527 Commits
Author SHA1 Message Date
Daan De Meyer eaf0f79293 machine: several follow-ups for #35066 (#35868)
Addresses the post-merge review comments.
2025-01-06 11:59:55 +01:00
Lennart Poettering 336acebc77 basic: port various pidfd/pidref helpers to PIDFD_GET_INFO and PIDFD_GET_*_NAMESPACE (#35242)
Supersedes #35308 (cherry-picked one commit and replaced the rest)

(I left a few comments that's folded by GitHub. Please make sure to
check them too.)
2025-01-06 11:23:08 +01:00
Lennart Poettering 7f72184f12 more pidref'ification (#35839)
This is split out of #35264, but makes a ton of sense on its own.
2025-01-06 11:21:43 +01:00
Daan De Meyer 65b603492d tty-ask-password-agent: if we're spawning further agents, grant them notify access (#35855)
Alternative to #35853
2025-01-06 10:05:29 +01:00
Lennart Poettering dd445d6e99 cgroup-util: add remoteness checks to all cg_pidref_get_xyz() calls 2025-01-06 09:54:41 +01:00
Lennart Poettering 1740d3e04b sd-login: make use of getpeerpidref() and cg_pidref_get_*() 2025-01-06 09:54:41 +01:00
Lennart Poettering 92d78966fd cgroup-util: add pidref counterparts for cg_pid_get_session() + cg_pid_get_owner_uid() 2025-01-06 09:54:41 +01:00
Lennart Poettering b2206fe514 socket-util: introduce getpeerpidref()
This combines getpeercred() and getpeerpidfd() and returns a PidRef
2025-01-06 09:45:57 +01:00
Yu Watanabe c192d87d83 machine: comment source and destination must be absolute
Addresses the post-merge review comment:
https://github.com/systemd/systemd/pull/35066#discussion_r1901841247
2025-01-06 17:14:26 +09:00
Yu Watanabe 8551ac2092 machine: use json_dispatch_const_path() where applicable
Follow-up for #35066 and 3866e9ce9d.
2025-01-06 17:09:43 +09:00
Yu Watanabe 5eec51a6c1 packit: Move fmf metadata into upstream (#35700)
This moves https://src.fedoraproject.org/tests/systemd into upstream so
it's one less moving piece that we need to remember.
2025-01-06 10:37:56 +09:00
Daan De Meyer b0b88f18b3 packit: Move fmf metadata into upstream
This moves https://src.fedoraproject.org/tests/systemd into upstream
so it's one less moving piece that we need to remember. While we're at
it, do various cleanups as well:

- Remove non-packit logic
- Set NO_BUILD in mkosi.local.conf as it's not picked up from the
  process environment
- Don't set unused environment variables
- Remove workdir logic as we run in an ephemeral VM already
- Drop --verbose from meson test to reduce noise
  from passed tests
- Use mkosi sandbox and drop duplicated deps from metadata file
- Stop running the unit tests as they already run as part of the rpm
  build and as part of the integration tests.
- Various style fixes
2025-01-05 23:21:36 +01:00
Daan De Meyer 8a9b989cba mkosi: Make path exclude more generic
The systemd rpms we try to install in packit have /usr/bin/bash and
/usr/bin/python3 as dependencies which breaks dnf5 because mkosi
doesn't download filelists metadata and
https://bugzilla.redhat.com/show_bug.cgi?id=2263771 is still not fixed
almost a year after being reported.
2025-01-05 23:19:44 +01:00
Daan De Meyer c008b70370 mkosi: Update to latest 2025-01-05 21:18:32 +01:00
Luca Boccassi c9098abd57 mkosi: update debian commit reference
* 7dc7ee8007 d/rules: drop duplicated config flag
* 2f417d66d1 Install new files for upstream CI
2025-01-05 01:49:43 +09:00
Chris Hofstaedtler 18f3c91a04 analyze: fix URLs to DeviceAllow=, IPAddressDeny=
Signed-off-by: Chris Hofstaedtler <zeha@debian.org>
2025-01-05 01:49:05 +09:00
Yu Watanabe be68dcdc20 firewall-util: allow to override the table and map names through env var
Then, use test specific names to make the test not conflict with
networkd.

Hopefully fixes #35526.
2025-01-05 01:48:29 +09:00
Mike Yuan 9598708a12 cgroup-util: explain why cg_pidref_get_path() cannot be ported over to pidfd helpers (yet)
See also: https://github.com/systemd/systemd/pull/35242#issuecomment-2506686806
2025-01-04 17:48:23 +01:00
Mike Yuan f1ba5c900b cgroup-util: introduce generic cg_path_from_cgroupid() helper
Taken from nsresourced/userns-registry.c userns_destroy_cgroup()
2025-01-04 17:48:22 +01:00
Mike Yuan 223d455670 process-util: make pid_is_unwaited() wrapper around pidref version 2025-01-04 17:48:22 +01:00
Mike Yuan 47f64104d1 process-util: port pidref_get_uid() and pidref_is_my_child() to pidfd helpers 2025-01-04 17:48:22 +01:00
Yu Watanabe 12f77fe058 hostname: read hardware serial from device tree (#35797)
Closes #35792.
2025-01-05 01:46:58 +09:00
Mike Yuan 85e7bbfaa4 pidfd-util: introduce pidfd_get_{ppid,uid,cgroupid} which goes via PIDFD_GET_INFO too 2025-01-04 17:08:01 +01:00
Mike Yuan dcf0ef3f42 pidfd-util: try to translate pidfd -> pid through ioctl(PIDFD_GET_INFO) 2025-01-04 17:08:01 +01:00
Mike Yuan 92b8e5e72f namespace-util: introduce pidref_in_same_namespace() 2025-01-04 17:08:01 +01:00
Mike Yuan a33f691374 process-util: move namespace_get_leader() to namespace-util
This allows us to drop the hack for recursive includes.
2025-01-04 17:08:00 +01:00
Mike Yuan 0577339d7d various: add missing include for namespace-util.h
Preparation for later commits for removing inclusion
of namespace-util from process-util.h
2025-01-04 17:08:00 +01:00
Mike Yuan 4cad58788b namespace-util: hook pidref_namespace_open() up with pidfd_get_namespace()
Supersedes #35308
2025-01-04 17:08:00 +01:00
Mike Yuan 2f7cd6c70a namespace-util: correct assertion in namespace_is_init()
Follow-up for 18ead2b03d
2025-01-04 17:07:59 +01:00
Mike Yuan 07610cafcf namespace-util: modernize fd_is_namespace() and is_our_namespace()
- Make fd_is_namespace() take NamespaceType
- Drop support for kernel without NS_GET_NSTYPE (< 4.11)
- Port is_our_namespace() to namespace_open_by_type()
  (preparation for later commits, where the latter
  would go by pidfd if available, avoiding procfs)
2025-01-04 17:07:59 +01:00
Yu Watanabe 1775337a36 namespace-util: refuse remote pidref in pidref_namespace_open() 2025-01-04 17:07:59 +01:00
Mike Yuan 72661e0091 pidfd-util: introduce helper for obtaining ns fd using PIDFD_GET_*_NAMESPACE directly 2025-01-04 17:07:59 +01:00
Mike Yuan 9a7b20b6a6 errno-util: add ERRNO_IS_IOCTL_NOT_SUPPORTED that checks EINVAL additionally 2025-01-04 17:07:58 +01:00
Mike Yuan 5d873679b1 core/socket: include peer pidfd id in socket service instance if available 2025-01-04 17:07:58 +01:00
Mike Yuan 2fa42318c3 pidref: move generic pidfd_get_inode_id() to pidfd-util
Prompted by https://github.com/systemd/systemd/pull/35224/commits/221d6e54c60389f26c2c79dbfa4e83204d2775d7

Also generalize pidfd_check_pidfs() and expose have_pidfs for later use.
2025-01-04 17:07:58 +01:00
Mike Yuan ebddfb0dee basic: introduce missing_pidfd.h 2025-01-04 16:58:13 +01:00
Mike Yuan b234026d09 process-util: extract pidfd-related funcs into pidfd-util.[ch] 2025-01-04 16:58:13 +01:00
Mike Yuan 4b59f6e63d tty-ask-password-agent: if we're spawning further agents, grant them notify access
Follow-up for 254649d576

Otherwise, systemd-ask-password-console.service times out
being started since the main process never delivers READY=1
notification.

Alternative to #35853
2025-01-04 16:44:53 +01:00
Mike Yuan c56935a09d tty-ask-password-agent: coding style tweaks 2025-01-04 16:44:50 +01:00
Mike Yuan 5734893f82 tree-wide: make sigactions static const 2025-01-04 16:40:53 +01:00
Mike Yuan bed4386e29 signal-util: generalize sigaction_nop_nocldstop 2025-01-04 16:40:53 +01:00
Yu Watanabe 573e8fc217 ptyfwd: fix typo
Follow-up for 8224a48da9.
2025-01-04 23:46:57 +09:00
Yu Watanabe ffe2923303 sd-varlink: introduce sd_varlink_get_current_method() 2025-01-04 23:33:09 +09:00
Yu Watanabe 8252b74f2d varlink: add comments for io.systemd.service interface 2025-01-04 23:17:47 +09:00
Yu Watanabe 04b09924b2 udev: split on_ctrl_msg() into small pieces (#35736) 2025-01-04 23:17:17 +09:00
Yu Watanabe 2da8da32b0 varlink: several cleanups for io.systemd.service interface (#35733) 2025-01-04 23:17:03 +09:00
Yu Watanabe 3e91aed6c4 hostname: also read serial number from device tree
This makes hostnamed try to read hardware serial number from
/proc/device-tree/serial-number.

Closes #35792.
2025-01-04 22:29:01 +09:00
Yu Watanabe c1507748c6 hostname: cache sd-device object for DMI
Also, always check if the obtained string is safe for passing through
DBus.
Just refactoring and preparation for later change.
2025-01-04 22:28:31 +09:00
Luca Boccassi 993f6f4281 mkosi: update debian commit reference
* 5d185f7901 Install new files for upstream CI
* c6484e14fc Update changelog for 257.1-7 release
* 543c1e6a42 d/t/tests-in-lxd: more fixes for debci compatibility
* a7d4a7ffa3 Add missing d/copyright file to signed template packages
* ab48efa495 Install new files for upstream CI
* d0cae6d06c Update changelog for 257.1-6 release
* 2fd23ffaaa Drop build dependency on libxen-dev on armhf, no longer available
* 35fbac92a5 d/t/upstream: update mkosi setting name
* 4d9862709e d/t/tests-in-lxd: drop auto-apt-proxy.conf from container
* cf133a8d32 Update changelog for 257.1-5 release
* 77ec315ec5 d/t/upstream: switch from btrfs to ext4 and use nspawn again
* 9c57f51fb8 d/t/upstream: do not mask machined and instead use mkosi workaround
* eb8b91d98b d/t/upstream: do not use --verbose with 'meson test'
* 21c65b7e22 d/t/tests-in-lxd: auto-apt-proxy is in universe in Ubuntu
* 081a78884b d/t/tests-in-lxd: install auto-apt-proxy in the nested container
* faa014a64e Update changelog for 257.1-4 release
* 4a48235928 Backport patch to workaround issue in new mdadm
* 005390f039 d/t/upstream: mask systemd-machined
* f7c86d8ec3 d/t/upstream: disable homed/userdbd in the test runner
* a1c30e55e0 d/t/upstream: do not use nspawn for tests
* 9bad1d3021 d/t/upstream: decrease parallelism to nproc - 1
* 7d1d583d7b Update changelog for 257.1-3 release
* dfe5996032 autopkgtest: mark tests-in-lxd as flaky
* 331567b13d d/t/upstream: disable ToolsTree= in mkosi config
* 04af5d8604 Update changelog for 257.1-2 release
* 2e58f0092c d/t/tests-in-lxd: use a single autopkgtest invocation
* 4b608de1d2 d/t/tests-in-lxd: prefer $AUTOPKGTEST_TMP to manual discovery
* fb4f3e4272 d/t/tests-in-lxd: do not fail if /etc/apt/sources.list.d/autopkgtest.list does not exist
* 385bbd0147 d/t/tests-in-lxd: set -x to get more logs
* ab678ff879 d/t/tests-in-lxd: quote variables to fix shellcheck warning
* 0d0eb96f16 d/t/tests-in-lxd: workaround broken os-release in sid
* 916a518d8e d/t/tests-in-lxd: use correct lxc image for debian
* 97deb28f99 Update changelog for 257.1-1 release
* 6276567a1d Install systemd-creds bash completion
* 084248775c Drop patches, merged upstream
* c4aa4ac6bd Update upstream source from tag 'upstream/257.1'
* dffde7eae8 d/watch: restrict to v257.x series
* 3c5143991c d/t/tests-in-lxd: fix autopkgtest source dir
* 5626a78ad1 systemd-ukify: recommend python3-cryptography
* ca5fdf7bc5 d/t/control: move tests-in-lxd above upstream suite
* 71c3f1014c d/systemd-resolved.postinst: copy existing /etc/resolv.conf on new installs
* 6ca4f4c1e5 d/control: make systemd-sysv Depends: on matching version of systemd
* eae57033c2 debian/control: add Breaks: systemd (<< ${binary:Version}) for udev
* 408cc1872c debian/systemd.postinst: Skip daemon-reexec and try-restarts during shutdown
* 6b3a01118b debian/tests/boot-and-services: drop test_no_failed
* 5a333eef8c d/t/tests-in-lxd: run some tests under LXD
* 882d221400 d/t/upstream: fix mkosi syntax warnings
2025-01-04 22:02:17 +09:00
Yu Watanabe bbb0dbe6b1 udev-ctrl: refuse ENV control message with invalid environment assignment
Previously, udevd accepts an arbitrary pair of key and value.
Let's make the environment variable assignment more strict for safety.

Note, we already refuse environment variables with the same way in
net/link-config.c.
2025-01-04 21:54:40 +09:00