We generally operate on the assumption that a source is "gone" as soon
as we unref it. This is generally true because we have the only reference.
But if something else holds the reference, our unref doesn't really stop
the source and it could fire again.
In particular, on_query_timeout() is called with DnsQuery* as userdata, and
it calls dns_query_stop() which invalidates that pointer. If it was ever
called again, we'd be accessing already-freed memory.
I don't see what would hold the reference. sd-event takes a temporary reference,
but on the sd_event object, not on the individual sources. And our sources
are non-floating, so there is no reference from the sd_event object to the
sources.
For #18427.
sockaddr_pretty() is the function that is ultimately called to print
the address, and afaics, it'll always use 0.0.0.0 if family==AF_INET.
So the test must have always been broken on systems with IPv6 support.
Fixes a failure in semaphore debian autopkgtest ci:
"65535" → ✓ 0 → "0.0.0.0:65535" (expect 0 / "[::]:65535") 15:15
Assertion 'streq(out, expected ?: in)' failed at src/test/test-socket-netlink.c:32, function test_socket_address_parse_one(). Aborting. 15:15
It causes a regression in certain running environments (networkd under LXC),
so avoid enabling for now.
Fixes#18795
Suggested-by: Topi Miettinen <toiwoton@gmail.com>
This addresses the issue described in https://lwn.net/Articles/846403/
and makes sure we will be able to stream bytes from procfs/sysfs via
copy_bytes() if people ask us to.
Let's add a special tweak to the btrfs builtin: if /dev/btrfs-control is
not there, let's consider all btrfs file systems as SYSTEMD_READY=0.
This is useful in initrds, where btrfs.ko might be missing. After the
initrd → host transition we can then retigger the device and undo the
SYSTEMD_READY=0 marking.
Prompted by https://bugzilla.redhat.com/show_bug.cgi?id=1930875 in which
I had previously used json_dispatch_unsigned and passed a return variable of
type unsigned when json_dispatch_unsigned writes a uintmax_t.
When we parse properties set via "systemctl set-property" we should
validate to the point where the data we pass over the bus makes rough
sense, but we shouldn't needlessly check whether specified enum values
are among the known enum values. The server side checks that anyway
again, and it's kinda nice if an older systemctl can be used to talk to
a newer systemd.
Our coding style dictates we should not clobber return parameters on
failure, and always initialize them on success. Do so here.
This changes getenv_path_list() to return ENXIO if the env var is not
set, which is similar to how we handle this in getenv_bool().
This drops debug logging from parse_env_extension_hierarchies(), since
it's done anyway in getenv_path_list()
Follow-up for: #18018
Clean up ignore_signals() + default_signals() + sigaction_many() a bit:
make it unnecessary to explicitly terminate the signal list with -1.
Merge all three calls into a single function that is just called with
slightly different parameters. And eliminate an unnecessary extra
iteration in its inner for() loop.
No change in behaviour.
The firmware spec doesn't really say whether FilePath of the LoadedImage
protocol may be NULL or not. So far we assumed it to be non-NULL, but
apparently the FreeBSD UEFI chainloader sets it to NULL. Handle this
gracefully.
(Noticed and tracked down by Alexander Schreiber)
Fixes: #18733
Yu Watanabe
Lennart Poettering
Zbigniew Jędrzejewski-Szmek
Einsler Lee
Luca Boccassi
Anita Zhang
Dan Streetman
Joerg Behrmann
Annika Wickert
Sergey Bugaev