When we send a batch of nfnl messages, but e.g. without sufficient
privilege, the kernel may only return an error message for
NFNL_MSG_BATCH_BEGIN and ignore all later messages.
So, we need to read the response for the NFNL_MSG_BATCH_BEGIN,
and if it is an error ignore the replies for the rest.
When we send a message with NLM_F_ACK, but if later we are not
interested in the reply and do not want to call sd_netlink_read(),
the reply will be stored in the rqueue forever.
Let's introduce a way to ignore received message without waiting reply.
When sd-event is attached to the sd-netlink object, use the timestamp of
the current event, otherwise call now().
This mostly change nothing. Preparation for later change.
Every other log in this function is debug level, so downgrade this too.
Otherwise compiling without libcryptsetup means sd-executor logs at
error level when starting any service, regardless of whether it is
needed or not.
Fixes https://github.com/systemd/systemd/issues/39968
Follow-up for efaf5a763d
This is a follow-up for 119d332d9c and
ensures the check only is applied to the system instance of machined. It
doesn't really apply to the per-user instance, because we never want to
permit differently privileged clients access anyway.
(The process_is_owned_by_uid() call might fail if invoked unpriv, hence
there's value in not calling it if machined runs in --user mode, it's
what makes machined actually work)
Previously, we looked for scripts in the tools/ directory ad hoc,
wherever they were needed. Let's do those checks in one place.
The main meson.build file is shrunk somewhat, which is always nice.
Quite often we need to specify include_directories('.'). Normally, meson
does this automatically, but we specify implicit_include_directories : false,
see 30d20907bd for an explanation.
Passing the include_directories argument through the template layers was rather
cumbersome. Let's simplify things by always including the directory of the
first source file. This makes the definitions simpler, at the cost of having an
unnecessary include directory in the list in some cases.
(Tests are excluded from this change, because they happen to not need this, but
also because some tests have source[0] which is a custom target, and we'd need
newer meson to support that.)
Always abbreviate import/pull/export structs with their first letter
instead of only doing it in some places and using the wrong letter in
other places.
gcry_mpi_t is defined as "typedef struct gcry_mpi *gcry_mpi_t;".
When const is applied to this type, it resolves to
"struct gcry_mpi *const" instead of what we expect ("const struct gcry_mpi *").
So we end up with a const pointer to a mutable object instead of a mutable
pointer to a const object. Since the pointer passed to the function
is copied regardless, making it const has zero benefit.
You'd think we could instead stop using gcry_mpi_t and replace it with
"const struct gcry_mpi *", except that gcrypt leaked this mess into its
api, so it expects const pointers to mutable objects as well, which means
we can't take pointers to const objects as arguments, as we'd discard the qualifier
when calling a gcrypt function.
To avoid confusion, let's drop the const qualifiers from the gcry_mpi_t arguments.
Follow-up for 7336f2c748
This alignes with some other optional modules in shraed/,
and it allows dlopen_libmount() to be optimized out entirely.
Let's avoid emitting pointless symbols.
Follow-up for b3243f4bee
and 5df44d0f6a
Since we now consider this a supported senario, let's hook up
libmount loading with the high-level unit_type_supported() machinery
and gracefully skip the whole unit accordingly.
This got split in 5cabeed80b
to accommodate --empower, and later --empower received
dedicated handling again (c36942916b).
I think the new naming makes more sense - --empower is privileged
after all, just with uid left unchanged. Hence merge
privileged_execution back into it.
Follow-up for 4f6ef13f43
Special casing --area= rather than --empower makes the code
self-explanatory, as --area= is about alternative home dir
after all. On top of that this ensures when --area= and
--empower are specified in combination we honor the home dir
switch, too.
Also, unless --same-root-dir was specified, don't make the executable
absolute if we're running in a chroot. Situations like this are still iffy,
but we might as well handle them a little more gracefully.
Latest glibc uses _Generic to have strstr() and other functions return
const char* or char* based on whether the input is a const char* or a
char*. This causes build failures as we previously always expected a
char*.
Let's fix the compilation failures and add our own macros similar to
glibc's to have string functions that return a mutable or const pointer
depending on the input.
Latest glibc uses _Generic to have strstr() and other functions return
const char* or char* based on whether the input is a const char* or a
char*. This causes build failures as we previously always expected a char*.
Let's fix the compilation failures and add our own macros similar to glibc's
to have string functions that return a mutable or const pointer depending on
the input.
When in --empower mode, all created files will be owned by the current
user, which could be problematic when creating files outside of the
current user's home directory, as other processes running as the same
user would be able to edit those files.
While this is a bit of an edge case since users already have to go through
the effort of writing --empower to indicate they want a privileged session
as the current user, it's not unphatomable to think they could start an
empowered session which they later return to and continue using. Currently,
it's not easy to differentiate a regular run0 session and an empowered session
at a glance, so users might think they're using a regular run0 session when
they're actually using an empowered session.
To address this problem, let's give empowered session their own identify, by
making the background orange, using the pumpkin emoji as the shell prompt
prefix and giving them an orange circle as the PTY title.
Some of the functions were ignoring failure in cleanup, others weren't. If we
got a reply, it's better to use it, so ignore failures in cleanup everywhere.
This moves the open call earlier, so that we do any state-changing operations
if we actually managed to open the nonblocking fd. The code is easier to follow
this way and might be more robust.
Suprisingly, this fixes https://github.com/systemd/systemd/issues/39055: it
seems that run0 chowns /dev/stdin (in my case /dev/pts/0) to root:root, and the
second run0 can read and write stdin/stdout throught the already-open fds,
but fd_reopen fails.
Fixes https://github.com/systemd/systemd/issues/39055.
Yu Watanabe
Daan De Meyer
Antonio Alvarez Feijoo
Luca Boccassi
Lennart Poettering
Zbigniew Jędrzejewski-Szmek
Mike Gilbert
Mike Yuan
Luca Boccassi
Frantisek Sumsal
Jelle van der Waa