This is heavily based on Kevin Kuehler's work, but the logic is also
significantly changed: instead of a straighforward port to openssl, both
versions of the code are kept, and at compile time we pick one or the other.
The code is purposefully kept "dumb" — the idea is that the libgcrypt codepaths
are only temporary and will be removed after everybody upgrades to openssl 3.
Thus, a separate abstraction layer is not introduced. Instead, very simple
ifdefs are used to select one or the other. If we added an abstraction layer,
we'd have to remove it again afterwards, and it don't think it makes sense to
do that for a temporary solution.
Co-authored-by: Zbigniew Jędrzejewski-Szmek <zbyszek@in.waw.pl>
# Conflicts:
# meson.build
This reduces dependencies. The speed of the code here is uimportant, because we
hash only a tiny amount of input data.
Debian and Ubuntu currently build without repart, see
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=976959
> repart requires openssl and so far I tried to avoid linking against
> both gnutls and openssl.
Co-authored-by: Zbigniew Jędrzejewski-Szmek <zbyszek@in.waw.pl>
On distros using Debian's base-passwd, the name of the group with gid 65534 is
nogroup. Currently, systemd-sysusers creates a spurious "nobody" group
systemd-sysusers[243]: Creating group nobody with gid 996
That's both confusing and redundant, as the nobody user still has primary group
65534 aka nogroup, and the nobody group simply goes completely unused.
So explicitly specify the primary group of the nobody user, and add a line
ensuring that that group exists.
This is not a problem for Debian (or Ubuntu) itself, as they add their own
version of basic.conf in their systemd build logic. But it appears on for
example Yocto/OpenEmbedded.
Assumption in edc027b was that job we first skipped because of active
ratelimit is still in run_queue. Hence we trigger the queue and dispatch
it in the next iteration. Actually we remove jobs from run_queue in
job_run_and_invalidate() before we call unit_start(). Hence if we want
to attempt to run the job again in the future we need to add it back
to run_queue.
Fixes#21458
I've seen this in `NetworkManager-1.34.0-0.3.el8.x86_64` (latest in CentOS
Stream 8 at the time of writing this message) which does not use the latest
Systemd but probably the code base is the same (see
https://github.com/NetworkManager/NetworkManager/commit/51f93e00a23fbd09f5ad96da6290bf4ca737d46a).
Before the patch:
```
libsystemd: eth0: DHCPv6 client: T1 expires in 34y 3w 6d 45min 31s
libsystemd: eth0: DHCPv6 client: T2 expires in 54y 5month 3w 3d 23h 20min 35s
```
After the patch:
```
libsystemd: eth0: DHCPv6 client: T1 expires in 3d 7h 58min 3s
libsystemd: eth0: DHCPv6 client: T2 expires in 5d 2h 26min 50s
```
same box (x86_64 system) and same DHCPv6 server.
This regression has likely been introduced by 8a8955507a.
This was broken in a subtle way: we'd get an ELF ref, but not the right one,
so no metadata note would be found.
Change the parsing function to return 1 when it finds something, so that
we can return early only when that happens.
The code looks a bit more complicated, but the compiler generates a simpler and
more compact text.
An additional advantage is that if any of the fields were repeating or not
power-of-two, the compiler would warn about an overridden entry in the table.
The macro variants can be used in static initializers.
The same guard against calling __builtin_clz(0) is added as for
__builtin_clzll(0), since that's undefined behaviour too. Our code
wouldn't call it, but this avoids a potential pitfall with the macro.
All variants map 0→0. Otherwise we'd often have to handle 0 specially
in callers.
__builtin_clz takes unsigned as the argument, so there's no LOG2I macro.
It seems that `pam_systemd` was the only thing left that wanted to use sockets
in file system path `/tmp/.X11-unix/X*`. X11 apps actually prefer using the
abstract socket version.
This allows running Xserver with `-nolisten tcp -nolisten unix`, which makes
the server only listen to an abstract socket.
Also in my setup, Xserver is running as a separate system service instead of
starting from display manager service, and now `PrivateTmp=yes` can be used for
both. The file system of the display manager service is inherited by user apps
and now their `/tmp` will be separate from `/tmp` of PID1 namespace as well as
`/tmp` of Xserver.
Any recent shim will refuse starting an image that does not have an sbat section
and will do so with a generic "Security Violation" message. And it is very easy
to forget passing -Dsbat-distro=auto to meson when creating a fresh build dir.
Adding sbat info when shim is not used or secure boot is disabled does not hurt
anyone, so default to auto. This still ensures to not add auto-detected info in
case we are cross building.
