Fixes a regression caused by d307410327.
The link_mode_masks flex array in struct ethtool_link_settings contains
three packed arrays, and the length of each array is given by
link_mode_masks_nwords field:
```
__u32 link_mode_masks[];
/* layout of link_mode_masks fields:
* __u32 map_supported[link_mode_masks_nwords];
* __u32 map_advertising[link_mode_masks_nwords];
* __u32 map_lp_advertising[link_mode_masks_nwords];
*/
```
Hence, we cannot use the received data as is through the union, but need
to shift the array to make each map accessible through the union.
Fixes the following warning:
```
[ 130.968170] H systemd-networkd[4288]: Restart operation initiated.
[ 130.968265] H systemd-networkd[4288]: Serializing...
[ 130.968274] H systemd-networkd[4288]: There is nothing to serialize.
(snip)
[ 130.978756] H systemd[1]: Stopped systemd-networkd.service.
(snip)
[ 130.991407] H systemd[1]: Starting systemd-networkd.service...
(snip)
[ 131.077440] H systemd-networkd[4318]: Enumeration completed.
[ 131.077471] H systemd-networkd[4318]: Deserializing...
[ 131.077491] H systemd-networkd[4318]: Failed to parse json (line=0, column=0): No data available
[ 131.077524] H systemd-networkd[4318]: Failed to deserialize the previous invocation, ignoring: No data available
```
Fixes the following varlink IDL validation error:
```
systemd-oomd[1687]: varlink-11-11: Parameters for method io.systemd.oom.ReportManagedOOMCGroups()
didn't pass validation on field 'cgroups': No anode
```
Follow-ups for 23ac08115a.
Since the commit, the cgroup path is set only when it is realized.
Fixes the following varlink IDL validation failure:
```
systemd-oomd[1687]: Varlink-IDL: Mandatory field 'path' is null or missing on object, refusing.
```
We might be operating with a newer systemctl on an image with older
systemd and thus without an initrd-preset directory. Before
4a8c395167, we would use the system
presets, let's make sure we keep doing that if we're operating on an
image without initrd presets.
Follow up for 4a8c395167.
Otherwise, ExecXYZ= given by drop-ins may be added to a masked unit,
then commands may be unexpectedly executed. For example, if an ExecStop=
is set in a drop-in for a running unit, then 'systemctl mask --now' for
the unit try to invoke the command specified in ExecStop=.
Fixes#38802.
As you can see, something spurious happens in the logs below.
```
initrd-switch-root.target: Trying to enqueue job initrd-switch-root.target/start/isolate
systemd-repart.service: Looking at job systemd-repart.service/stop conflicted_by=no
systemd-repart.service: Looking at job systemd-repart.service/start conflicted_by=no
systemd-repart.service: Fixing conflicting jobs systemd-repart.service/stop,systemd-repart.service/start by deleting job systemd-repart.service/stop
initrd-switch-root.target: Fixing conflicting jobs initrd-switch-root.target/stop,initrd-switch-root.target/start by deleting job initrd-switch-root.target/stop
systemd-repart.service: Deleting job systemd-repart.service/start as dependency of job initrd-switch-root.target/stop
```
The two conflicting jobs for systemd-repart.service are initially queued
as the following:
- initrd-switch-root.target has Wants=initrd-root-fs.target, and
initrd-root-fs.target has Wants=systemd-repart.service (through symlink),
hence starting initrd-switch-root.target tries to start
systemd-repart.service,
- systemd-repart.service has Conflicts=initrd-switch-root.target, hence
starting initrd-switch-root.target tries to stop
systemd-repart.service.
As similar, interestingly(?) starting initrd-switch-root.target tries to
stop initrd-switch-root.target.
So, now there are at least two pairs of conflicting jobs:
- systemd-repart.service: start vs stop,
- initrd-switch-root.target: start vs stop.
As these jobs are induced by starting initrd-switch-root.target, of course
the most important one is the start job for initrd-switch-root.target.
Previously, as you can see in the logs at the beginning, even if
the start job for initrd-switch-root.target is important, we may first
try to resolve the conflict in systemd-repart.service, and may drop the
stop job for systemd-repart.service even if it is relevant to the start
job of initrd-switch-root.target.
This makes first we solve the pair of conflicting jobs for anchor task.
So the stop job for initrd-switch-root.target is dropped first, and the
induced start job for systemd-repart.service is automatically removed,
thus it is not necessary to solve the conflict in systemd-repart.service
anymore.
This is especially important for services that are enabled both in initrd
and after switching root. If a stop job for one of the service is
unexpectedly dropped during switching root, then the service is not stopped
before switching root, and will never start after that.
Fixes#38765.
As a hardening measure it makes sense not to process potentially crafted
data in `openssl_pubkey_to_pem()` before we know that the caller is
authorized to perform the action.
On reload, resolved does not apply new DNSSEC= (or DNSOverTLS=) settings
on links, because the link unicast scopes are not re-created. However,
the servers and link states are updated correctly, so resolvectl and link
state files do show the new setting, leading users to believe the change
took effect immediately, the same way `resolvectl dnssec` does.
Fix this by freeing all of the link unicast scopes during reload, so
that they are re-created with the new settings in link_allocate_scopes().
A boot loader entry for a unified kernel image has
BootEntry.kernel : path to the image relative to ESP or XBOOTLDR,
BootEntry.path : path to the image.
Hence, these two effectively point to the same file.
Hence, by unlink command, the image is removed by
```
deref_unlink_file(&known_files, e->kernel, e->root);
```
then later tried again by
```
r = chase_and_unlink(e->path, root, ...);
```
and of course it fails with -ENOENT.
Let's ignore the failure there. We already ignore ENOENT on removal
at various places, especially in deref_unlink_file().
Fixes#38706.
Follow-ups for 8702496bfb.
If we are asked to come up with our own name for the namespace to
allocate (because client enabled "mangle"), then we so far created a
randomized name if shortening what was proposed didn't work. This broke
polkit authorization however, because when polkit is in the mix, we
process method calls twice, submitting the polkit request on the first
and then assuming a response is known on the second invocation. But if
we generate a randomized name for the two checks we'll not be ablet to
match up the requests because it's going to be different. Let's fix that
by not using a randomized name, but one hashed from the socket
connection we are processing mixed with the client provided name. This
will ensure that for the same method call we'll generate the same name,
but different calls (i.e. calls with different names on the same socket,
or with any name on any socket) we'll end up with different names,
minimizing chance of collision.
This ensures PK starts to work with nsresourced userns registration when
a bad or no name is specified, which previously would end up in a PK
query loop.
tmpfiles attempts to correct the label of a file during various actions
via the function fd_set_perms(). Currently, said function generally
respects the dry-run mode. However, it attempts to fix the label of a
given file regardless of the state of said dry-run mode.
This causes problems, because a user could attempt to run tmpfiles with
elevated permissions and dry run enabled, expecting the tool to not
modify their system. Instead, tmpfiles would falsely relabel a file,
modifying their system.
This commit explicitly checks for when dry-run is enabled and skips the
file relabelling process. Furthermore, I added logging for both cases.
I found helpful during debugging. That said, I don't think it's
necessary to use the level LOG_INFO on the dry-run path, as it would
always produce an info log.
manager_add_job() -> transaction_add_job_and_dependencies() may update
dependencies when a unit is not loaded yet. Hence, we need to restart
dependency loop in that case.
Follow-up for b7777d0884 (#37465).
Fixes#38676.
Commit 70b7e03 introduced 3 calls to log_debug() about the presence or
absence of NX_COMPAT support. Since sd-boot does not yet have the
ability to only print messages above a certain loglevel, these will
always be printed, even on top of the configured splash screen. This
commit removes the log_debug() call after a success and only prints
those for missing firmware support if the UEFI should support them in
the first place (i.e. starting with version 2.10).
This change introduced a regression that stops DNSOverTLS from working
after some time. Revert it for now, as there's no fixup available at
the moment.
Fixes https://github.com/systemd/systemd/issues/38509
This reverts commit 4b6dbd4ea5.
Since c5de7b14ae
file searching implies a new mount api syscall by default,
to trigger automounts.
But, this is not necessary in most cases, e.g. when chasing
syspath in sd-device (actually this causes regression in umockdev,
see https://github.com/martinpitt/umockdev/issues/271).
Another example is reading unit files, especially .network files,
as automount may trigger mounting network filesystems...
Also, when this is used in NSS plugins, programs that load the
plugins may fail because of spuriously configured seccomp. See #38565.
Let's not trigger automount by default, and do only when explicitly
requested.
This introduces CHASE_TRIGGER_AUTOFS, and use it in
- service manager,
- bootctl and finding ESP/xbootldr,
- sysupdate,
- mountfsd,
- systemd-mount.
There may be several more places we should trigger automount, but let's
do that later.
Follow-up for c5de7b14ae.
Fixes#38565.
Replaces #38569.
Co-authored-by: Luca Boccassi <luca.boccassi@gmail.com>
This reverts commit 490aa05ca1.
As commented https://github.com/systemd/systemd/pull/38569#discussion_r2284978273,
the commit makes autofs check bypassed. Before the commit, when
CHASE_NO_AUTOFS is set, we did not shortcut chasing paths, and refused
any autofs mount points in the path. However, with the commit, the flag
was swapped but even when CHASE_AUTOFS is unset, the autofs check may be
skipped.
To fix the issue, rather than swapping the flag, we should introduce
another flag, say CHASE_TRIGGER_AUTOFS. This revert the commit, and in a
later commit, the new flag will be introduced.
When NX_COMPAT gets enabled, firmwares will enforce that executable
memory is either writable or executable.
This needs kernel compatibility, when it will happen the kernel will
have the NX_COMPAT bit set. If it is, set the memory buffer to RO.
Note that this must be undone on failure, as EDK2 in some configurations
overwrites memory ranges that are returned with FreePages() with a
fixed pattern, so if the pages are RO it will crash.
This is only an issue with the new custom PE loader, as LoadImage()
and StartImage() will always do the right thing automatically.
https://microsoft.github.io/mu/WhatAndWhy/enhancedmemoryprotection/https://www.kraxel.org/blog/2023/12/uefi-nx-linux-boot/
Follow-up for cab9c7b5a4
Fixes https://github.com/systemd/systemd/issues/38545
Since https://github.com/systemd/systemd/commit/c5de7b14ae2e08d267d8d75bc88934ac6aa7dcd6
file searching implies a new mount api syscall by default,
to trigger automounts.
This is problematic in NSS plugins, as they are dlopen'ed inside
processes by glibc, for two reasons.
First of all, potentially searching on a networked filesystem
automount could lead to nasty surprises, such as the process
responsible for setting up the network filesystem trying to
search on that same filesystem.
More importantly, the new mount api syscall was never part of
the filesystem seccomp filter that we provide by default, and
given mounting/remounting/bind mounting is one of the possible
ways to bypass sandboxing it is very likely not allowed when
custom filters are used in sandboxed processes, if they don't
need to do these operations otherwise.
The filesystem seccomp mask we provide has been updated, however
this only takes effect on the next restart of a service. When
systemd is upgraded via a package upgrade, the new nss plugin is
installed and will be immediately dlopen'ed by glibc when needed,
without waiting for the process to restart, which means the existing
seccomp filter applies, causing the filter to trigger.
Given it's not really possible for any arbitrary program to
predict which NSS modules glibc will load, given programs do not
configure that and instead nsswitch is set up by the sysadmin,
it's impossible to handle at each process level. It's also not
possible to know when it will be triggered, given the plugin
is not linked in each binary tools like need-restart cannot
even pre-emptively restart services that may be affected.
This means in practice, upgrading from systemd << v258 to >= v258
requires a reboot to avoid either subtle or catastrophic system
failures.
By avoiding to trigger automounts in nss-systemd we can avoid
both issues.
userdb drop-ins are searched for in:
/etc/userdb/
/run/userdb/
/run/host/userdb/
/usr/local/lib/userdb/
/usr/lib/userdb/
none of which are supported as automounts anyway.
Note that this happens only when the userdbd service is not running,
as otherwise nss-systemd will go through the varlink IPC, rather than
doing the searches in-process.
So invert CHASE_NO_AUTOFS to CHASE_AUTOFS and set it in the places where
we do want to trigger automounts, like looking for the ESP.
Follow-up for https://github.com/systemd/systemd/commit/c5de7b14ae2e08d267d8d75bc88934ac6aa7dcd6
Fixes https://github.com/systemd/systemd/issues/38565
The SHA256SUMS files provided by https://images.linuxcontainers.org/
are slightly non-conforming, insted of using " *" or " " as separator
between hash and file name they use " ". Let's accept that too, in the
interest of maximizing compatibility.
Zbigniew Jędrzejewski-Szmek
Yu Watanabe
Daan De Meyer
Matthias Gerstner
Antonio Alvarez Feijoo
Luca Boccassi
Nick Rosbrook
Lennart Poettering
Jan Fooken
Felix Pehla
Luca Boccassi