Given that this is also the place to store raw disk images which are
very much bootable with qemu/kvm it sounds like a misnomer to call the
directory "container". Hence, let's change this sooner rather than
later, and use the generic name, in particular since we otherwise try to
use the generic "machine" preferably over the more specific "container"
or "vm".
After all, nspawn can now dissect MBR partition levels, too, hence
".gpt" appears a misnomer. Moreover, the the .raw suffix for these files
is already pretty popular (the Fedora disk images use it for example),
hence sounds like an OK scheme to adopt.
This adds three kinds of file system locks for container images:
a) a file system lock next to the actual image, in a .lck file in the
same directory the image is located. This lock has the benefit of
usually being located on the same NFS share as the image itself, and
thus allows locking container images across NFS shares.
b) a file system lock in /run, named after st_dev and st_ino of the
root of the image. This lock has the advantage that it is unique even
if the same image is bind mounted to two different places at the same
time, as the ino/dev stays constant for them.
c) a file system lock that is only taken when a new disk image is about
to be created, that ensures that checking whether the name is already
used across the search path, and actually placing the image is not
interrupted by other code taking the name.
a + b are read-write locks. When a container is booted in read-only mode
a read lock is taken, otherwise a write lock.
Lock b is always taken after a, to avoid ABBA problems.
Lock c is mostly relevant when renaming or cloning images.
This adds two new settings to networkd's .network files:
IPForwarding=yes and IPMasquerade=yes. The former controls the
"forwarding" sysctl setting of the interface, thus controlling whether
IP forwarding shall be enabled on the specific interface. The latter
controls whether a firewall rule shall be installed that exposes traffic
coming from the interface as coming from the local host to all other
interfaces.
This also enables both options by default for container network
interfaces, thus making "systemd-nspawn --network-veth" have network
connectivity out of the box.
cunescape_length_with_prefix() is called with the length as an
argument, so it cannot rely on the buffer being NUL terminated.
Move the length check before accessing the memory.
When an incomplete escape sequence was given at the end of the
buffer, c_l_w_p() would read past the end of the buffer. Fix this
and add a test.
We would ignore options like "fail" and "auto", and for any option
which takes a value the first assignment would win. Repeated and
options equivalent to the default are rarely used, but they have been
documented forever, and people might use them. Especially on the
kernel command line it is easier to append a repeated or negated
option at the end.
This fixes parsing of options in shared/generator.c. Existing code
had some issues:
- it would treate whitespace and semicolons as seperators. fstab(5)
is pretty clear that only commas matter. And the syntax does
not allow for spaces to be inserted in the field in fstab.
Whitespace might be escaped, but then it should not seperate
options. Treat whitespace and semicolons as any other character.
- it assumed that x-systemd.device-timeout would always be followed
by "=". But this is not guaranteed, hasmntopt will return this
option even if there's no value. Uninitialized memory could be read.
- some error paths would log, and inconsistently, some would just
return an error code.
Filtering is split out to a separate function and tests are added.
Similar code paths in other places are adjusted to use the new function.
Generators are different than unit files: they are never automatically
generated, so there's no point in allowing /etc to override /run. On
the other hand, overriding /etc might be useful in some cases.
Sometimes it is necessary to stop a generator from running. Either
because of a bug, or for testing, or some other reason. The only way
to do that would be to rename or chmod the generator binary, which is
inconvenient and does not survive upgrades. Allow masking and
overriding generators similarly to units and other configuration
files.
For the systemd instance, masking would be more common, rather than
overriding generators. For the user instances, it may also be useful
for users to have generators in $XDG_CONFIG_HOME to augment or
override system-wide generators.
Directories are searched according to the usual scheme (/usr/lib,
/usr/local/lib, /run, /etc), and files with the same name in higher
priority directories override files with the same name in lower
priority directories. Empty files and links to /dev/null mask a given
name.
https://bugs.freedesktop.org/show_bug.cgi?id=87230
Remove the optional sepearate opening of the directory,
it would be just too complicated with the change to
multiple directories.
Move the middle of execute_directory() to a seperate
function to make it easier to grok.
Make sure to extract the log-priority when comparing against
max-log-level, otherwise, we will always drop those messages.
This fixes bus-proxyd to properly send warnings on policy blocks.
When there are a lot of split out journal files, we might run out of fds
quicker then we want. Hence: bump RLIMIT_NOFILE to 16K if possible.
Do these even for journalctl. On Fedora the soft RLIMIT_NOFILE is at 1K,
the hard at 4K by default for normal user processes, this code hence
bumps this up for users to 4K.
https://bugzilla.redhat.com/show_bug.cgi?id=1179980
btrfs' COW logic results in heavily fragment journal files, which is
detrimental for perfomance. Hence, turn off COW for journal files as we
create them.
Turning off COW comes at the cost of data integrity guarantees, but this
should be acceptable, given that we do our own checksumming, and
generally have a pretty conservative write pattern.
Also see discussion on linux-btrfs:
http://www.spinics.net/lists/linux-btrfs/msg41001.html
So far, if we had no knowledge about the correct $TERM we defaulted to
v102, as a safe, conservative choice. However, the terminfo data for
vt102 is not aware of pageup/pagedown, which makes "less" much harder
work with than necessary. Setting vt220 allows them to work correctly.
"vt220" should be a sufficiently safe choice too, given that xterm,
gnome-terminal and the linux console all strive to implement vt220 as
baseline, already to pass pageup/pagedown correctly to apps.
Effectively, with this change "journalctl -e" run inside a
"systemd-nspawn" terminal will now run a pager where pageup/pagedown
works, which is quite an improvement of usability for containers.
Our write pattern is quite awful for CoW file systems (btrfs...), as we
keep updating file parts in the beginning of the file. This results in
fragmented journal files. Hence: when rotating files, defragment them,
since at that point we know that no further write accesses will be made.
With this change it is possible to send file descriptors to PID 1, via
sd_pid_notify_with_fds() which PID 1 will store individually for each
service, and pass via the usual fd passing logic on next invocation.
This is useful for enable daemon reload schemes where daemons serialize
their state to /run, push their fds into PID 1 and terminate, restoring
their state on next start from the data in /run and passed in from PID
1.
The fds are kept by PID 1 as long as no POLLHUP or POLLERR is seen on
them, and the service they belong to are either not dead or failed, or
have a job queued.
When setting up a namespace, mount flags like noexec, nosuid and
nodev are cleared, so the mounts always have exec, suid and dev
flags enabled.
Copy source directory mount flags to target mount when remounting
the bind mounts.
Regression introduced by ed757c0cb0
Mirror the implementation of columns(), since the fd_columns()
functions returns a negative integer for errors.
Also fix columns() to return the unsigned variable instead of the
signed intermediary (they're the same, but better to be explicit).
Even though we use fallocate() it appears that file systems like btrfs
will trigger SIGBUS on certain low-disk-space situation. We should
handle that, hence catch the signal, add it to a list of invalidated
pages, and replace the page with an empty memory area. After each write
check if SIGBUS was triggered, and consider the write invalid if it was.
This should make journald a lot more robust with file systems where
fallocate() is not reliable, for example all CoW file systems
(btrfs...), where changing written data can fail with disk full errors.
https://bugzilla.redhat.com/show_bug.cgi?id=1045810
