Follow-up for 179dcf924f
After the mentioned commit, event_add_child_pidref() duplicates
the pidfd internally, hence the original pidfd would be leaked
with TAKE_PIDREF.
Now that sd-event doesn't insist on blocking
SIGCHLD anymore when watching for process exit,
let's stop blocking SIGCHLD tree-wide where not
required anymore.
We use symbols provided by unistd.h without including it. E.g.
open(), close(), read(), write(), access(), symlink(), unlink(), rmdir(),
fsync(), syncfs(), lseek(), ftruncate(), fchown(), dup2(), pipe2(),
getuid(), getgid(), gettid(), getppid(), pipe2(), execv(), _exit(),
environ, STDIN_FILENO, STDOUT_FILENO, STDERR_FILENO, F_OK, and their
friends and variants, so on.
Currently, unistd.h is indirectly included mainly in the following two paths:
- through missing_syscall.h, which is planned to covert to .c file.
- through signal.h -> bits/sigstksz.h, which is new since glibc-2.34.
Note, signal.h is included by sd-eevent.h. So, many source files
indirectly include unistd.h if newer glibc is used.
Currently, our baseline on glibc is 2.31. We need to support glibc older
than 2.34, but unfortunately, we do not have any CI environments with
such old glibc. CIFuzz uses glibc-2.31, but it builds only fuzzers, and
many files are even not compiled.
Let's move some more implementation logic into functions. We keep
the logic that requires the macro in the macro and move the rest into
functions.
While we're at it, let's also make the parameter declarations of
all the string table macros less clausthrophobic.
Follow-up for 7f6af95dab
- Allocate internal buf on the stack, memdup() only at the end.
This ensures we're able to handle OOM gracefully, i.e.
return -EAGAIN on OOM while still emptying socket buffer.
- Do not treat empty notify message as error.
- Raise log level since all callers log loudly anyway.
We nowadays support unprivileged invocation of systemd-nspawn +
systemd-vmspawn, but there was no support for discovering suitable disk
images (i.e. no per-user counterpart of /var/lib/machines). Add this
now, and hook it up everywhere.
Instead of hardcoding machined's, importd's, portabled's, sysupdated's
image discovery to RUNTIME_SCOPE_SYSTEM I introduced a field that make
the scope variable, even if this field is always initialized to
RUNTIME_SCOPE_SYSTEM for now. I think these four services should
eventually be updated to support a per-user concept too, this is
preparation for that, even though it doesn't outright add support for
this.
This is for the largest part not user visible, except for in nspawn,
vmspawn and the dissect tool. For the latter I added a pair of
--user/--system switches to select the discovery scope.
We'd log that we're skipping the target, but it would never actually get
removed from the manager's list. Thus, we'd advertise targets that don't
actually exist to clients.
In the original version of the sysupdated PR, this was handled by
removing the target from the manager's list in target_free, and using a
_cleanup_ attribute to free the target when skipping. However, this
changed at some point during review. So, this commit takes the
alternative approach
First, this fixes a case where an error is logged twice at the LOG_ERR
level. Second, this goes through and raises the log level where
appropriate so that sysupdated is easier to debug. Finally, it replaces
EINVAL with EPROTO where appropriate, since EINVAL implies that the
caller passed some incorrect arguments (which is incorrect; sysupdate
passed some invalid JSON back to us)
If someone runs `updatectl update`, sysupdate will be running multiple
update jobs at the same time, which can make reasoning about the output
in the journal quite difficult. Especially if things go wrong: the error
messages didn't mention which job failed. Nor was there any link between
job ID and the PID of the worker process logging to the journal. This
is all fixed here!
Cuts out some `strdup`s, and also avoids a rather weird case of donating
memory to a function. Basically just duplicates the solution I just
implemented for sysupdate's callout handler.
This fixes a bug introduced during review of sysupdated. Originally,
we just returned EALREADY verbatim to signify that the target is
already up-to-date. Then we switched this to a proper error
(org.freedesktop.sysupdate1.NoCandidate) during review. But that now
maps to EIO, not EALREADY. Thus, whenever there's nothing to update,
updatectl would report I/O errors to the user, even though nothing
actually went wrong.
Mike Yuan
Daan De Meyer
Yu Watanabe
Lennart Poettering
Adrian Vovk
