| Commit message (Collapse) | Author | Files | Lines |
|---|
|
|
|
|
|
Follow-up for 7ea4392f1e444388caa706d6bd64fb7b30dc2903
Let's not hide errors needlessly aggressively.
|
|
Before:
```
write(2, "Device /dev/loop1p1 is too small.\n", 34) = -1 ENOTCONN (Transport
endpoint is not connected)
```
After:
```
$ journalctl -b -e | grep 'too small'
Apr 02 16:53:30 loora systemd[343579]: Device /dev/loop1p1 is too small.
```
|
|
On certain distributions such as NixOS the mtime of `/etc/hosts` is
locked to a fixed value. In such cases, only checking the last mtime of
`/etc/hosts` is not enough - we also need to check if the st_ino/st_dev
match up. Thus, let's make sure make sure that systemd-resolved also
rereads `/etc/hosts` if the inode or the device containing `/etc/hosts` changes.
Test script:
```bash
hosts="/etc/hosts"
echo "127.0.0.1 testpr" > "hosts_new"
mv "hosts_new" "$hosts"
resolvectl query testpr || exit 1
mtime="$(stat -c %y "$hosts")"
echo "127.0.0.1 newhost" > "hosts_tmp"
touch -d "$mtime" "hosts_tmp"
install -p "hosts_tmp" "$hosts"
sleep 10
resolvectl query newhost || exit 1
rm -f "hosts_tmp"
```
Closes #14456.
|
|
|
|
|
|
|
|
This reverts commit 7e1ed1f3b29162df25064b33dc55ac8cf432bb0b.
systemd-repart is not a user service that should be something people
enable/disable, instead it should just work if there's configuration for
it. It's like systemd-tmpfiles, systemd-sysusers, systemd-load-modules,
systemd-binfmt, systemd-systemd-sysctl which are NOPs if they have no
configuration, and thus don't hurt, but cannot be disabled since they
are too deep part of the OS.
This doesn't mean people couldn't disable the service if they really
want to, there's after all "systemctl mask" and build-time disabling,
but those are OS developer facing instead of admin facing, that's how it
should be.
Note that systemd-repart is in particular an initrd service, and so far
enable/disable state of those is not managed anyway via "systemctl
enable/disable" but more what dracut decides to package up and what not.
|
|
|
|
|
|
split() and FOREACH_WORD really should die, and everything be moved to
extract_first_word() and friends, but let's at least make sure that for
the remaining code using it we can't deadlock by not progressing in the
word iteration.
Fixes: #15305
|
|
|
|
|
|
Consumers of the sd-bus convenience API can't make convenience
helpers of their own without va_list variants.
This commit is a mechanical change splitting out the existing function
bodies into bare va_list variants having a 'v' suffixed to the names.
The original functions now simply create the va_list before forwarding
the call on to the va_list variant, and the va_list variants dispense
with those steps.
|
|
The function sd_device_get_property_value has some paths where it exits without
touching the n pointer. In those cases, n remained uninitialized until it was
eventually read inside isempty where it caused the segmentation fault.
Fixes #15078
|
|
|
|
For now, this function is nearly equivalent to the si_uint64 parser, except for
an additional range check as Linux only takes 32-bit values as bitrates. In
future, this may also be used to introduce fancier bitrate config formats.
|
|
|
|
|
|
Apparently people wondered about that:
https://lists.freedesktop.org/archives/systemd-devel/2020-March/044091.html
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Also convert these to use take_fdopendir().
|
|
Also convert these to use take_fdopen().
|
|
Some fdopendir() calls remain where safe_close() is manually
performed, those could be simplified as well by converting to
use the _cleanup_close_ machinery, but makes things less trivial
to review so left for a future cleanup.
|
|
fdopendir() wrapper analogous to take_fdopen()
|
|
Mechanical change to eliminate some cruft by using the
new take_fdopen{_unlocked}() wrappers where trivial.
|
|
With the addition of _cleanup_close_ there's a repetitious
pattern of assigning -1 to the fd after a successful fdopen
to prevent its close on cleanup now that the FILE * owns the
fd.
This introduces a wrapper that instead takes a pointer to the
fd being opened, and always overwrites the fd with -1 on success.
A future commit will cleanup all the fdopen call sites to use the
wrapper and elide the manual -1 fd assignment.
|
|
/home is posibly a remote file system. it makes sense to order homed
after it, so that we can properly enumerate users in it, but we probably
shouldn't pull it in ourselves, and leave that to users to configure
otherwise.
Fixes: #15102
|
|
|
|
This way it's easier to see when a wrong $BUILD_DIR was passed.
|
|
It's lightweight and generally useful, so it should be enabled by default. But
users might want to disable it for whatever reason, and things should be fine
without it, so let's make it installable so it can be disabled if wanted.
Fixes #15175.
|
|
Fixes #15083. Users might want to disable homed if not used to save resources.
|
|
This essentially adds another layer of configurability:
build disable, this, presence of configuration. The default is
set to enabled, because the service does nothing w/o config.
|
|
When we are supposed to accept numeric UIDs formatted as string, then
let's check that first, before passing things on to
valid_user_group_name_full(), since that might log about, and not the
other way round.
See: #15201
Follow-up for: 93c23c9297e48e594785e0bb9c51504aae5fbe3e
|
|
trivial cosmetic cleanup
|
|
stroage -> storage
|
|
|
|
|
|
That's what you get for changing these signals around, after checking
everything works, not before.
Bad, Lennart, bad!
Fixes: #15085
|
|
I don't think anyone uses it.
|
|
https://mesonbuild.com/Release-notes-for-0-54-0.html#ninja-version-requirement-bumped-to-17
https://github.com/mesonbuild/meson/issues/6867
|
|
|
|
|
|
|
