Daan De Meyer [Wed, 14 May 2025 19:41:43 +0000 (21:41 +0200)]
networkd-util: Avoid call to endswith()
Instead of recalculating the length of the string again after
reading it, let's make sd_netlink_message_read_string() return the
length of the string that we then use to check if the last character
is a dot or not.
This allows us to get rid of the string-util.h include in #37344.
Luca Boccassi [Wed, 14 May 2025 19:02:34 +0000 (20:02 +0100)]
test: fix assertion failure with CONFIG_UNIX_DIAG disabled
On OBS the build VM is heavily locked down, with network
disabled in various ways in the custom kernel, to isolate the
build, including disabling CONFIG_UNIX_DIAG.
[ 456s] /* test_af_unix_get_qlen */
[ 456s] src/test/test-socket-netlink.c:393: Assertion failed: Expected "af_unix_get_qlen(unix_fd, &q)" to succeed, but got error: No such file or directory
[ 454s] /* test_sock_diag_unix */
[ 454s] src/libsystemd/sd-netlink/test-netlink.c:727: Assertion failed: Expected "sd_netlink_call(nl, message, 0, &reply)" to succeed, but got error: No such file or directory
Daan De Meyer [Wed, 14 May 2025 16:46:58 +0000 (18:46 +0200)]
capability-util: Ignore unknown capabilities instead of aborting
capability_quintet_mangle() can be called with capability sets
containing unknown capabilities. Let's not crash when this is the
case but instead ignore the unknown capabilities.
With 'or', we ignore the empty string (but not '0'), and we only call
time.time() lazily. So this works the same as the code that is replaced,
but avoids the ugly repetition.
We recommend that users create overriddes. This creates the problem that there
is no syntax to unset a property. Thus, the user needs to just set the property
to "something else" in the override file. But then the blurb saying that
"VAR=1" (or "VAR=0" in some cases) is the only allowed value can be confusing.
Say that both 0 and 1 can be set, since this documentation is also intended
for end users.
In our files, we generally don't want the override values anywhere. But we
have a test which checks the rvalue, which should be enough.
Daan De Meyer [Wed, 14 May 2025 13:33:36 +0000 (15:33 +0200)]
blockdev-util: Remove dependency on string-util.h
Let's insist on a string literal in SYS_BLOCK_PATH_MAX() so that
we don't accidentally allocate VLAs and let's inline strempty() in
xsprintf_sys_block_path() so we don't need to include string-util.h
in blockdev-util.h
We'll remove the actual string-util.h include as part of #37344.
Daan De Meyer [Wed, 14 May 2025 14:51:10 +0000 (16:51 +0200)]
Introduce forward.h header with forward declarations (#37428)
In preparation for adopting forward declarations to reduce unnecessary
transitive includes across the tree, let's introduce a forward.h header
with forward declarations for all libc, libsystemd, basic and shared
types.
Additionally, this header exports all basic integer types and errno
constants, as well as all macros including assertions macros. These
header files contain types often used in headers and are always included
in every source file one way or another anyway.
To avoid having to include memory-util.h and alloc-util.h in forward.h,
we split off the parts we need from both into cleanup-util.h and only
include cleanup-util.h in forward.h.
To keep this commit self-contained, we include cleanup-fundamental.h and
cleanup-util.h from the headers that originally contained the same
macros. We'll remove these again in a later commit that optimizes the
includes in src/basic and src/fundamental.
validatefs: properly authenticate all subordinate devices of DM devices
Previously, we'd only authenticate "one" of the subordinate devices of a
DM device, and which one was somewhat undefined, it would be what we
find in slaves/ first. This is in particular a problem with dm-verity
which generally has two subordinate devices: the data device and the
hash device.
Let's fix this properly. This means two things:
1. iterate through *all* subordinate devices of a DM device (i.e.
iterate through the sysfs slaves/ subdir), not just
one
2. permit configuring a list of gpt labels and gpt type uuids in the
xattrs of mount points, so that all valid combinations can be listed.
This only updates the validation like this. The generation of xattrs
that carry multiple type uuids/labels in systemd-repart will follow in a
later commit.
This extends the syntax of the two gpt-related xattrs, to allow lists of
things. This is a true extension, without breaking compat (but even if
it was, it wouldn't matter given that validatefs was added post v257,
i.e. is not included in a stable release.
Daan De Meyer [Tue, 13 May 2025 09:50:05 +0000 (11:50 +0200)]
Introduce forward.h header with forward declarations
In preparation for adopting forward declarations to reduce unnecessary
transitive includes across the tree, let's introduce a forward.h header
with forward declarations for all libc, libsystemd, basic and shared types.
Additionally, this header exports all basic integer types and errno constants,
as well as all macros including assertions macros. These header files contain
types often used in headers and are always included in every source file one
way or another anyway.
To avoid having to include memory-util.h and alloc-util.h in forward.h, we
split off the parts we need from both into cleanup-util.h and only include
cleanup-util.h in forward.h.
To keep this commit self-contained, we include cleanup-fundamental.h and
cleanup-util.h from the headers that originally contained the same macros.
We'll remove these again in a later commit that optimizes the includes in
src/basic and src/fundamental.
journald: rename primary object from "Server" to "Manager"
In all our daemons the primary entrypoint object is called "Manager".
But so far there was one exception: in journald it was called "Server".
Let's normalize that, and stick to the same nomenclature everywhere, to
make journald less special.
Mike Yuan [Tue, 13 May 2025 20:58:02 +0000 (22:58 +0200)]
fork-journal: use char* const* for strv input param
This is compatible with char** and is what I originally
asked for in
https://github.com/systemd/systemd/pull/36858#discussion_r2086792739
Someone needs to read better ;-)
Yu Watanabe [Fri, 9 May 2025 07:56:48 +0000 (16:56 +0900)]
integration-tests: adjust priorities
When running with sanitizers:
```
26/95 systemd:integration-tests / TEST-21-DFUZZER OK 1517.75s
40/95 systemd:integration-tests / TEST-85-NETWORK-NetworkdDHCPClientTests OK 779.18s
42/95 systemd:integration-tests / TEST-04-JOURNAL OK 716.17s
```
and without sanitizers:
```
44/95 systemd:integration-tests / TEST-85-NETWORK-NetworkdDHCPClientTests OK 730.33s
29/95 systemd:integration-tests / TEST-64-UDEV-STORAGE-simultaneous_events OK 701.49s
40/95 systemd:integration-tests / TEST-04-JOURNAL OK 348.05s
```
So, let's set higher priorities only on these tests.
Yu Watanabe [Tue, 13 May 2025 14:02:13 +0000 (23:02 +0900)]
login,udev: avoid race between systemd-logind and systemd-udevd in setting ACLs
Previously, both udevd and logind modifies ACLs of a device node. Hence,
there exists a race something like the following:
1. udevd reads an old state file,
2. logind updates the state file, and apply new ACLs,
3. udevd applies ACLs based on the old state file.
This makes logind not update ACLs but trigger uevents for relevant
devices to make ACLs updated by udevd.
Yu Watanabe [Tue, 13 May 2025 14:50:22 +0000 (23:50 +0900)]
login: do not call manager_process_seat_device() more than once per event
When udevd broadcasts an event for e.g. a graphics device with master-of-seat
tag, then previously manager_process_seat_device() was called twice for
the event.
With this commit, the function is called only once even for an event for
such device.
This returns to the original approach proposed in
https://github.com/systemd/systemd/pull/17270. After review, the approach was
changed to use sd_pid_get_owner_uid() instead. Back then, when running in a
typical graphical session, sd_pid_get_owner_uid() would usually return the user
UID, and when running under sudo, geteuid() would return 0, so we'd trigger the
secure path.
sudo may allocate a new session if is invoked outside of a session (depending
on the PAM config). Since nowadays desktop environments usually start the user
shell through user units, the typical shell in a terminal emulator is not part
of a session, and when sudo is invoked, a new session is allocated, and
sd_pid_get_owner_uid() returns 0 too. Technically, the code still works as
documented in the man page, but in the common case, it doesn't do the expected
thing.
$ build/test-sd-login |& rg 'get_(owner_uid|cgroup|session)'
sd_pid_get_session(0) → No data available
sd_pid_get_owner_uid(0) → 1000
sd_pid_get_cgroup(0) → /user.slice/user-1000.slice/user@1000.service/app.slice/app-ghostty-transient-5088.scope/surfaces/556FAF50BA40.scope
I think it's worth checking for sudo because it is a common case used by users.
There obviously are other mechanims, so the man page is extended to say that
only some common mechanisms are supported, and to (again) recommend setting
SYSTEMD_LESSSECURE explicitly. The other option would be to set "secure mode"
by default. But this would create an inconvenience for users doing the right
thing, running systemctl and other tools directly, because then they can't run
privileged commands from the pager, e.g. to save the output to a file. (Or the
user would need to explicitly set SYSTEMD_LESSSECURE. One option would be to
set it always in the environment and to rely on sudo and other tools stripping
it from the environment before running privileged code. But that is also fairly
fragile and it obviously relies on the user doing a complicated setup to
support a fairly common use case. I think this decreases usability of the
system quite a bit. I don't think we should build solutions that work in
priniciple, but are painfully inconvenient in common cases.)
compress: deal with zstd decoder issues gracefully
If zstd frames are corrupted the initial size returned for the current
frame might be wrong. Don#t assert() on that, but handle it gracefully,
as EBADMSG
logs-show: use memory_startswith() rather than startswith()
Let's be strict here: this data is conceptually not NUL terminated,
hence use memory_startswith() rather than startswith() (which implies
NUL termination). All other similar cases in logs-show.c got this right.
Fix the remaining three, too.
projects / thirdparty / systemd.git / log
