When booting with debug logs, we print:
Setting '/proc/sys/fs/file-max' to '9223372036854775807
'
Setting '/proc/sys/fs/nr_open' to '2147483640
'
Couldn't write fs.nr_open as 2147483640, halving it.
Setting '/proc/sys/fs/nr_open' to '1073741816
'
Successfully bumped fs.nr_open to 1073741816
The strange formatting is because we explicitly appended a newline in those two
places. It seems that the kernel doesn't care. In fact, we have a few dozen other
writes to sysctl where we don't append a newline. So let's just drop those here
too, to make the code a bit simpler and avoid strange output in the logs.
Some distributions have started phasing out SHA1, which breaks
the systemd-measure test case in its current form. Let's make sure we
can use SHA1 for signing beforehand to mitigate this.
Spotted on RHEL 9, where SHA1 signatures are disallowed by [0]:
```
openssl genpkey -algorithm RSA -pkeyopt rsa_keygen_bits:2048 -out "/tmp/pcrsign-private.pem"
...
openssl rsa -pubout -in "/tmp/pcrsign-private.pem" -out "/tmp/pcrsign-public.pem"
writing RSA key
/usr/lib/systemd/systemd-measure sign --current --bank=sha1 --private-key="/tmp/pcrsign-private.pem" --public-key="/tmp/pcrsign-public.pem"
Failed to initialize signature context.
```
[0] https://gitlab.com/redhat/centos-stream/rpms/openssl/-/blob/c9s/0049-Selectively-disallow-SHA1-signatures.patch
Once mkosi migrates to systemd-repart, only "disk" will be supported
for making disk images with mkosi and the filesystem will have to be
specified in repart partition definition files. To accomodate this
change, let's remove the explicit Format= assignment which means we'll
default to a disk image with ext4 until we add our own mkosi.repart/
directory.
This fixes an issue introduced by af2aea8bb6.
When an outdated address or route is passed to link_request_address()/route(),
then they return 0 and the address or route will not be assigned. Such
situation can happen when we receive RA with zero lifetime. In that
case, we should not unset Link.ndisc_configured flag, otherwise even
no new address nor route will assigned, the interface will enter to the
configuring state, and unnecessary DBus property change is emit and the state
file will be updated. That makes resolved or timesyncd triggered to
reconfigure the interface.
Fixes#25456.
let's make sure we can probe file systems also when unprivileged:
instead of probing the partition block devices for file system
signatures, let's go via the original "whole" fd.
libblkid makes this easy actually, as it allows us to specify the
offset/size of the area to probe. And we have the partition
offsets/sizes anyway, so it's trivial for us to make use of.
This thus enables fs probing also when lacking privs and operating on
naked regular files without loopback devices or anything like this.
Let's explicitly flush the kernel's buffer cache on the whole block
device once we ran "mkfs". This is necessary, because partition and
whole block devices maintain separate buffer caches, and thus writing
to one will not be visible on the other if cached there already, until
the latter's cache is explicitly flushed.
This is preparation for later adding support for probing file sytems
also if we have no open partition block devices, and hence want to use
the whole block device instead.
Let's extend the test further, and try the codepaths where we do not
pin/add the partition block devices (i.e. which is the codepaths we use
when running without privs)
libblkid really should define an enum for this on its own, but it
currently doesn't and returns literal numeric values. Lets make this
more readable by adding our own symbolic names via an enum.
e.g. vfat doesn't support symlinks, sockets, fifos, etc so let's ignore
any copy failures related to unsupported file types when populating
filesystems.
Curently, these two flags were implied by dissect_loop_device(), but
that's not right, because this means systemd-gpt-auto-generator will
dissect the root block device with these flags set and that's not
desirable: the generator should not cause the partition devices to be
created (we don't intend to use them right-away after all, but expect
udev to find/probe them first, and then mount them though .mount units).
And there's no point in opening the partition devices, since we do not
intend to mount them via fds either.
Hence, rework this: instead of implying the flags, specify them
explicitly.
While we are at it, let's also rename the flags to make them more
descriptive:
DISSECT_IMAGE_MANAGE_PARTITION_DEVICES becomes
DISSECT_IMAGE_ADD_PARTITION_DEVICES, since that's really all this does:
add the partition devices via BLKPG.
DISSECT_IMAGE_OPEN_PARTITION_DEVICES becomes
DISSECT_IMAGE_PIN_PARTITION_DEVICES, since we not only open the devices,
but keep the devices open continously (i.e. we "pin" them).
Also, drop the DISSECT_IMAGE_BLOCK_DEVICE combination flag, since it is
misleading, i.e. it suggests it was appropriate to specify on all
dissected blocking devices, but that's precisely not the case, see the
systemd-gpt-auto-generator case. My guess is that the confusion around
this was actually the cause for this bug we are addressing here.
Fixes: #25528
reset_terminal_fd sets certain minimum required terminal attributes
that systemd relies on.
One of those attributes is `ONLCR` which ensures that when a new line
is sent to the terminal, that the cursor not only moves to the next
line, but also moves to the very beginning of that line.
In order for `ONLCR` to work, the terminal needs to perform output
post-processing. That requires an additional attribute, `OPOST`,
which reset_terminal_fd currently fails to ensure is set.
In most cases `OPOST` (and `ONLCR` actually) are both set anyway, so
it's not an issue, but it could be a problem if, e.g., the terminal was
put in raw mode by a program and the program unexpectedly died before
restoring settings.
This commit ensures when `ONLCR` is set `OPOST` is set too, which is
the only thing that really makes sense to do.
When copying between filesystems, sometimes the target filesystem
might not support symlinks/fifos/sockets/... and we want to log and
ignore any failures to copy such files when copying. Let's introduce
a new flag to enable this behavior.
Otherwise the `terminate()` method sends SIGKILL rather quickly (~0.3s),
which then leaves a dangling scope on the host system, breaking further
test executions.
