Daan De Meyer [Fri, 25 Oct 2024 15:03:37 +0000 (17:03 +0200)]
import: Draw progress bars
Currently every progress update results in a new progress message
which is extremely verbose. Instead, let's use the progress bar infra
to draw a proper progress bar similar to what we do in systemd-repart
now.
progress-bar: issue Windows Terminal progress indicating ANSI sequences (#34929)
This generates the Windows Terminal OSC sequences indicating progress.
This let's the terminal know that we are doing a slow operation, and how
we are progressing.
Windows Terminal uses this in two ways: it shows a circle in the tab
that completes, and it highlights the progress in the task bar.
I found no Linux terminal that currently supports it, but also none that
didn't like it. Thankfully most terminals correctly ignore unrecognized
OSC sequences.
I think we should just merge this, and see if this trips up too many
people, but I have reason to believe this shouldn't be too bad.
And yes, I do work from Windows Terminal sometimes, ssh into my Linux
build systems, and it is really cute seeing the progress animation
there.
Let's ramp up security for system user accounts, at least where
possible, by creating them fully locked (instead of just with an invalid
password). This matters when taking non-password (i.e. SSH) logins into
account.
* e42eed4afd test_sysusers_defined: support new ! line flag for creating fully locked accounts
* 2c6a4e2f90 Version 256.7
* bedc0270e7 Move yum/dnf protection removal config file under /usr
* 5a82129a41 Reword some descriptions
* ce99022f7b Version 256.6
We should avoid unnecessary abbreviations for such messages, and this
puts a maximum limit on things, hence it should indicate this in the
name.
Moreover, matches is a bit confusing, since most people will probably
call "busctl monitor" without any match specification, i.e. zero
matches, but that's not what was meant here at all.
Also, add a brief switch for this (-N) since I figure in particular
"-N1" might be a frequent operation people might want to use.
busctl: fix timeout calculation for "busctl monitor"
The --timeout= logic was implemented incorrectly, as it would not put a
a limit on the runtime of the operation, but only on the IO sleep.
However, spurious wakeups are possible, hence the timer would be reset
too often.
Fix that, by determining the absolute timestamp early, and checking
against that.
progress-bar: issue Windows Terminal progress indicating ANSI sequences
This generates the Windows Terminal OSC sequences indicating progress.
This let's the terminal know that we are doing a slow operation, and how
we are progressing.
Windows Terminal uses this in two ways: it shows a circle in the tab
that completes, and it highlights the progress in the task bar.
I found no Linux terminal that currently supports it, but also none that
didn't like it. Thankfully most terminals correctly ignore unrecognized
OSC sequences.
I think we should just merge this, and see if this trips up too many
people, but I have reason to believe this shouldn't be too bad.
And yes, I do work from Windows Terminal sometimes, ssh into my Linux
build systems, and it is really cute seeing the progress animation
there.
Daan De Meyer [Mon, 28 Oct 2024 20:12:19 +0000 (21:12 +0100)]
mkosi: Move copying packages to the output directory to the postinst script
Now that we have the mkosi.clangd script to run clangd from the mkosi
build script, it becomes clear that doing cleanup with mkosi.clean has
a big gap in that we always run the mkosi.clean script and thus we also
run it when we run the mkosi.clangd script, causing the previously built
packages to be removed when we run clangd without producing new ones.
In mkosi we're improving the situation by only running clean scripts when we
clean up the output directory and disallowing writing to the output directory
from build scripts.
Let's adapt systemd to these changes by moving the copying of packages to the
output directory to the postinst script.
Yu Watanabe [Sat, 26 Oct 2024 14:28:32 +0000 (23:28 +0900)]
update-utmp: wait slightly longer for the private bus socket being active
Before a339495b1d67f69f49ffffdd96002164a28f1c93, update-utmp typically
connects the public DBus socket when disconnected from the private DBus
socket, as dbus service should be active even during PID1 is being reexecuted.
However, after a339495b1d67f69f49ffffdd96002164a28f1c93, update-utmp
tries to connect only the private DBus socket, but reexecution of PID1
may be slow, hence all trials may fail when the reexecution is slow.
With this change, now it waits for 100ms to 2000ms, so in total it waits
about 37 seconds in average, previously about 4 seconds.
David Michael [Fri, 25 Oct 2024 14:45:44 +0000 (10:45 -0400)]
socket: support setting ownership of message queues
This applies the existing SocketUser=/SocketGroup= options to units
defining a POSIX message queue, bringing them in line with UNIX
sockets and FIFOs. They are set on the file descriptor rather than
a file system path because the /dev/mqueue path interface is an
optional mount unit.
Ryan Wilson [Fri, 18 Oct 2024 18:41:09 +0000 (11:41 -0700)]
cgroup: Add support for ProtectControlGroups= private and strict
This commit adds two settings private and strict to
the ProtectControlGroups= property. Private will unshare the cgroup
namespace and mount a read-write private cgroup2 filesystem at /sys/fs/cgroup.
Strict does the same except the mount is read-only. Since the unit is
running in a cgroup namespace, the new root of /sys/fs/cgroup is the unit's
own cgroup.
We also add a new dbus property ProtectControlGroupsEx which accepts strings
instead of boolean. This will allow users to use private/strict via dbus
and systemd-run in addition to service files.
Note private and strict fall back to no and yes respectively if the kernel
doesn't support cgroup2 or system is not using unified hierarchy.
Ryan Wilson [Wed, 16 Oct 2024 23:51:17 +0000 (16:51 -0700)]
core: Refactor ProtectControlGroups= to use enum vs bool
This commit refactors ProtectControlGroups= from using a boolean
in the dbus/execute backend to using an enum. There is no functional
change but this will allow adding new non-boolean values (e.g. strict,
private) a la PrivateHome.
sd-json,tree-wide: add sd_json_format_enabled() and use it everwhere
We often used a pattern like if (!FLAGS_SET(flags, SD_JSON_FORMAT_OFF)),
which is rather verbose and also contains a double negative, which we try
to avoid. Add a little helper to avoid an explicit bit check.
This change clarifies an aditional thing: in some cases we treated
SD_JSON_FORMAT_OFF as a flag (flags & SD_JSON_FORMAT_OFF), while in other cases
we treated it as an independent enum value (flags == SD_JSON_FORMAT_OFF).
In the first form, flags like SD_JSON_FORMAT_SSE do _not_ turn the json
output on, while in the second form they do. Let's use the first form
everywhere.
No functional change intended.
Initially I wasn't sure if this helper should be made public or just internal,
but it seems such a common pattern that if we expose the flags, we might just
as well expose it too, to make life easier for any consumers.
We would need to use pure if the funtion was getting pointers and
dereferencing them. But sd128_t is a structure and those functions
only access the parameters of the call.
const is stronger than pure, see
https://gcc.gnu.org/onlinedocs/gcc/Common-Function-Attributes.html#index-pure-function-attribute
and
https://gcc.gnu.org/onlinedocs/gcc/Common-Function-Attributes.html#index-const-function-attribute.
Zhou Qiankang [Mon, 28 Oct 2024 04:47:20 +0000 (12:47 +0800)]
meson: add loongarch64's definition to cpu_arch_defines
The default definition to add is `-D__loongarch64__`, which is not searched in [bpf_tracing.h](https://github.com/libbpf/libbpf/blob/09b9e83102eb8ab9e540d36b4559c55f3bcdb95d/src/bpf_tracing.h#L68)
This may avoid `error: Must specify a BPF target arch via __TARGET_ARCH_xxx` in loongarch64
Ryan Wilson [Sat, 26 Oct 2024 22:33:30 +0000 (15:33 -0700)]
core: Add RootDirectory= path to error message if directory does not exist
A colleague reported when RootDirectory= does not exist, systemd reports an error like:
```
Failed to set up mount namespacing: No such file or directory
```
Unfortunately, with large spec files, it can be hard to diagnose which path systemd is talking
about. Thus, to make the error message more helpful and similar to mount error messages, we add
the root directory/image path into the error message like:
```
Failed to set up mount namespacing: /tmp/thisdoesnotexist: No such file or directory
```
run: reconnect if our dbus connection is terminated
We must be prepared that systemd temporarily drops off the bus or
disconnects our direct connections (due to systemctl daemon-reexec or
so). Hence automatically reconnect when we watch the unit status, and
handle this case gracefully.
run: drop "-" prefix from command line when generating unit description
Let's not confuse users with the login shell indicator and drop it from
the description. This means a run0 session will now usually show up with
a description of "[run0] /bin/bash" rather than "[run0] -/bin/bash".
run: prefix unit description with our own process name
I think we should try to communicate clearly if something is a run0
session, or a systemd-run invocation. Hence, let's initialize the
description so that the command is prefixed by
program_invocation_short_name.
Effectively this means that our run0 sessions now appear as services
with a description of "[run0] -/bin/bash"
The current logic is a bit complex how systemd-run units are called. It
used to be just the unique ID of the dbus connection. Which was nice,
since its system-widely, uniquely assigned to us. But this didn't work
out well, due to direct connections to PID 1 and due to soft reboots.
We nowadays have a better ID to use though, with nicer properties: the
kernel manages a pidfd ID for every process after all, and it's globally
unique, for any process, and regardless of soft reboots. Hence use that
for naming preferably, and just keep one branch with a randomized name
as fallback.
run0: optionally show superhero emoji on each shell prompt
This makes use of the infra introduced in 229d4a980607e9478cf1935793652ddd9a14618b to indicate visually on each prompt that we are in superuser mode temporarily.
pick ad5de3222f userdbctl: add some basic client-side filtering
core: make sure that if PAMName= is set we always do the full user changing even if no user is specified explicitly
When PAMName= is set this should be enough to go through our entire user
changing story, so that PAM is definitely run, and environment variables
definitely pulled in and so on.
Previously, it would happen that under some circumstances we might no do
this when transitioning from root to root itself even though PAM was
enabled.
This adds some basic client-side user/group filtering to "userdbctl":
1. by uid/gid min/max
2. by user "disposition" (i.e. show only regular users with "userdbctl
user -R")
3. by fuzzy name (i.e. search by substring/levenshtein of user name,
real name, and other identifiers of the user/group record).
In the long run we also want to support this server side, but let's
start out with doing this client-side, since many backends won't support
server-side filtering anytime soon anyway, so we need it in either case.
The kernel regression has been hopefully fixed by
https://github.com/torvalds/linux/commit/c6508124193d42bbc3224571eb75bfa4c1821fbb
which is included in 6.12-rc4.
Let's drop the workaround.
Daan De Meyer [Tue, 22 Oct 2024 12:22:47 +0000 (14:22 +0200)]
docs: Mention that a local build might be required to use mkosi
Currently we need ukify with support for --profile and --join-profile
which isn't in an official release yet so mention that a local build
from source might be required.
dns-domain: tweak hash table comparison function for DNS names
Currently, when comparing two DNS names when storing them in a
hashtable, and the DNS names are not actually valid we'll compare the
error codes.
This is not very smart however, since this means two invalid DNS names
that happen to be equally "invalid" will be considered identical, even
if their strings are entirely different.
Let's find a better solution for this niche case: let's simple compare
the domains as strings.
This matters in case of DNS label compression: if we already added added
an invalid DNS name into the label compression hash table, and lookup
any other invalid DNS name, this lookup will likely return what the
earlier one already returned, and that's confusing.
projects / thirdparty / systemd.git / log
