core: split system/user job timeouts and make them configurable
Config options are -Ddefault-timeout-sec= and -Ddefault-user-timeout-sec=.
Existing -Dupdate-helper-user-timeout= is renamed to -Dupdate-helper-user-timeout-sec=
for consistency. All three options take an integer value in seconds. The
renaming and type-change of the option is a small compat break, but it's just
at compile time and result in a clear error message. I also doubt that anyone was
actually using the option.
This commit separates the user manager timeouts, but keeps them unchanged at 90 s.
The timeout for the user manager is set to 4/3*user-timeout, which means that it
is still 120 s.
Fedora wants to experiment with lower timeouts, but doing this via a patch would
be annoying and more work than necessary. Let's make this easy to configure.
journal-file: refuse writing to journal files where the header size is different then expected
We keep adding fields to the header, and it's fine reading files with
different header sizes, as we check via the size if the fields we need
are included. However, let's be stricter when writing journal files than
when reading, and insist that the header structure in the file actually
matches our expectations. Refuse otherwise, so that a new file is
created after rotation that then matches our expectations.
This makes sure that mismatch in header size is treated exactly as
unknown "compatible" flags, which is our other mechanism to allow
extending the journal file format in a non-breaking way.
In https://bugzilla.redhat.com/show_bug.cgi?id=2156900 sysusers was reporting a
conflict between the following lines:
u root 0:0 "Super User" /root /bin/bash
u root 0 "Super User" /root
The problem is that those configurations are indeed not equivalent. If group 0
exists with a different name, the first line would just create the user, but the
second line would create a 'root' group with a different GID. The second
behaviour seems definitely wrong. (Or at least more confusing in practice than
the first one. The system is in a strange shape, but the second approach takes
an additional step than is worse than doing nothing.)
When this line was initially added, we didn't have the uid:gid functionality for
'u', so we didn't think about this too much. But now we do, so we should use it.
$ build/systemd-sysusers --root=/var/tmp/inst7 --inline 'g foobar 0'
Creating group 'foobar' with GID 0.
$ build/systemd-sysusers --root=/var/tmp/inst7 --inline 'u root 0 "Zuper zuper"'
src/sysusers/sysusers.c:1365: Creating group 'root' with GID 999.
src/sysusers/sysusers.c:1115: Suggested user ID 0 for root already used.
src/sysusers/sysusers.c:1183: Creating user 'root' (Zuper zuper) with UID 999 and GID 999.
vs.
$ build/systemd-sysusers --root=/var/tmp/inst7 --inline 'u root 0:0 "Zuper zuper"'
src/sysusers/sysusers.c:1183: Creating user 'root' (Zuper zuper) with UID 0 and GID 0.
Tuetuopay [Fri, 27 Jan 2023 14:10:49 +0000 (15:10 +0100)]
network/dhcp4: accept local subnet routes from DHCP
RFC3442 specifies option 121 (Classless Static Routes) that allow a DHCP
server to push arbitrary routes to a client. It has a Local Subnet
Routes section expliciting the behavior of routes with a null (0.0.0.0)
gateway.
Such routes are to be installed on the interface with a Link scope, to
mark them as directly available on the link without any gateway.
Networkd currently drops those routes, which is against the RFC, as
Linux has proper support for such routes.
Fixes: 7f20627 ("network: dhcp4: ignore gateway in static routes if destination is link-local or in the same network")
Ilya Leoshkevich [Mon, 30 Jan 2023 20:21:48 +0000 (21:21 +0100)]
bpf: fix restrict_fs on s390x
Linux kernel's bpf-next contains BPF LSM support for s390x. systemd's
test-bpf-lsm currently fails with this kernel.
This is an endianness issue: in the restrict_fs bpf program,
magic_number has type unsigned long (64 bits on s390x), but magic_map
keys are uint32_t (32 bits). Accessing magic_map using 64-bit keys may
work by accident on little-endian systems, but fails hard on big-endian
ones.
Arsen Arsenović [Sat, 28 Jan 2023 21:32:41 +0000 (22:32 +0100)]
importd: Always specify file unpacked by tar
Despite popular belief, the default file extracted by GNU tar is not stdin. It
is the value of the TAPE environment variable, falling back on a compile-time
constant. On my system, the default value is /dev/full, which causes tar to
just spin forever due to --ignore-zeros. Always specifying this flag is the
safe thing to do.
~$ tar --show-defaults
--format=gnu -f/dev/full -b20 --quoting-style=escape
--rmt-command=/usr/sbin/grmt
See also: ``(tar)defaults'', available via Info viewers, and in HTML form at:
https://www.gnu.org/s/tar/manual/html_node/defaults.html
Daan De Meyer [Sun, 29 Jan 2023 14:17:06 +0000 (15:17 +0100)]
mkosi: Add back CentOS Stream 8 to CI
It's still useful to test the EFI handover logic in systemd-boot.
We use a mkosi.prepare script to install a newer python and update
the system to use it.
Daan De Meyer [Sun, 29 Jan 2023 14:04:13 +0000 (15:04 +0100)]
mkosi: Don't modify rootfs in build script
When unprivileged mkosi becomes available, builds will be executed
as an unprivileged user, so we won't be able to modify the rootfs
anymore. Let's update the build script to account for this.
Daan De Meyer [Sat, 28 Jan 2023 14:12:08 +0000 (15:12 +0100)]
nspawn: Make sure we create bind mount points as the correct UID/GID
When using --private-users, we have to create bind mount points as
the user that will become root in the user namespace, so let's take
that into account.
Yu Watanabe [Sun, 13 Nov 2022 17:49:19 +0000 (02:49 +0900)]
sleep: introduce siphash24_compress_id128()
Also, rename get_battery_identifier() to siphash24_compress_device_sysattr().
This also makes any errors in sd_id128_get_machine() or id128_get_product()
ignored. For the machine ID, the failure should not be significant unless
the file stored in the discharge level is reused by another system, which
is quite unusual. For the product ID, if the firmware provides useless
ID (all zero or all 0xFF), then loading/storing the discharge rate
becomes completely broken, that should be avoided.
Note, now sysattrs are used instead of properties in uevent files, but
both provide the same information, hence no functionality should be
changed.
Yu Watanabe [Sun, 13 Nov 2022 17:08:05 +0000 (02:08 +0900)]
sleep: introduce SuspendEstimationSec=
Before v252, HibernateDelaySec= specifies the maximum timespan that the
system in suspend state, and the system hibernate after the timespan.
However, after 96d662fa4c8cab24da57523c5e49e6ef3967fc13, the setting is
repurposed as the default interval to measure battery charge level and
estimate the battery discharging late. And if the system has enough
battery capacity, then the system will stay in suspend state and not
hibernate even if the time passed. See issue #25269.
To keep the backward compatibility, let's introduce another setting
SuspendEstimationSec= for controlling the interval to measure
battery charge level, and make HibernateDelaySec= work as of v251.
This also drops implementation details from the man page.
tmpfiles: automatically create /etc/credstore/ and friends
This adds a tmpfiles.d/ snippet for LoadCredential= style credentials
directories in /etc/ and /run/.
This is done primarily to ensure that the access modes for the dirs are
set up properly, in the most restrictive ways. Specifically these are
set to 0000, so that CAP_DAC_OVERRIDE is necessary to enumerate and read
the credentials, and being UID=0 is not sufficient to do so.
This creates /etc/credstore/, but leaves /run/credstore/ absent if
missing, for now. Thinking is: the latter being non-persistent is
created by software usually, not manually by users, and hence more
likely right. But dunno, we might want to revisit this sooner or later.
This is ultimately an exercise to advertise the LoadCredential= concept
a bit, and do so in a reasonably secure way, underlining the safety of
the concept.
Daan De Meyer [Thu, 26 Jan 2023 21:18:47 +0000 (22:18 +0100)]
nspawn: Drop CAP_NET_BIND_SERVICE when in userns but not in netns
If we're in a user namespace but not unsharing the network namespace,
we won't be able to bind any privileged ports even with
CAP_NET_BIND_SERVICE, so let's drop it from the retained capabilities
so services can condition themselves on that.
journal: automatically pick up boot ID in journal_file_append_entry()
Let's pick up the boot ID early if unspecified, in
journal_file_append_entry(). This is symmetric to the fact that we
already pick up the monotonic timestamp in journal_file_append_entry()
if unspecified, and given that the monotonic clock is not too useful
without its boot ID it makes a lot of sense to pick them up at the same
time.
There are two relevant callers of journal_file_append_entry() right now:
journald (which leaves the boot ID unspecified) and journal-remote
(there are also some tests, but those don't matter too much). The former
calls it to store new entries in the journal file, the latter for
converting/processing/merging existing ones (where it passes along the
original boot ID). This new code hence only is relevant on the former,
and using the boot ID of the current system is the right choice for live
generated entries.
Note that this effectively changes little, since the lower-level
function journal_file_append_entry_internal() will copy boot ID stored
in the file header into all records if unspecified, and typically that's
the one of the local system. But strictly speaking this is not the right
thing to do, since we actually might end up appending to journal files
from previous boots. (The lower level function is indirectly used by
various tests, where the copying-from-header logic kinda makes sense
since they are detached from any live messages streaming in from the
host after all).
This is a follow-up for 1d8d483f59ffa62974772fb58a8ef4abe88550ec and
makes the strict ordering by realtime clock within each journal file
optional, not mandatory. It then enables it for all journal files
written by journald, but leaves it off on others (for example those
written by journald-remote).
This relaxes the logic behind writing journal files to the status quo
ante for all cases where the journal files are not generated, but are
merged/processed/propagated. Typically when processing journal records
from many files ordering by realtime clock and monotonic clock are
contradictory, and cannot be universally guaranteed as the records are
interleaved. By enforcing strict rules we would thus end up generating
myriads of separate journal files, each with just a few records in them.
Hence, let's losen restrictions again, but continue to enforce them in
journald, i.e. when we original create the journal files locally.
Note that generally there's nothing really wring with having journal
files with non-monotonically ordered entries by realtime clock. Looking
for records will not be deterministic anymore, but that's inherent to a
realtime clock that jumps up and down. So you won't get the "only"
answer, but still *a* answer that is correct if you seek for a realtime
clock.
This also adds similar logic on the monotonic clock, which is also only
enabled when generating journal files locally. This should be harder to
trigger (as journald will generate the messages, and should run with a
stable boot id and monotonic clock), but let's better be safe than
sorry, and refuse on the lower layer what makes no sense, even if it's
unlikely the higher layer will ever generate records that aren't ordered
by their monotonic clock.
Frantisek Sumsal [Wed, 25 Jan 2023 12:21:09 +0000 (13:21 +0100)]
partition: fix build with newer linux/btrfs.h uapi header
linux/btrfs.h needs to be included after sys/mount.h, as since [0]
linux/btrfs.h includes linux/fs.h causing build errors:
```
In file included from /usr/include/linux/fs.h:19,
from ../src/basic/linux/btrfs.h:29,
from ../src/partition/growfs.c:6:
/usr/include/sys/mount.h:35:3: error: expected identifier before numeric constant
35 | MS_RDONLY = 1, /* Mount read-only. */
| ^~~~~~~~~
[1222/2169] Compiling C object systemd-creds.p/src_creds_creds.c.o
ninja: build stopped: subcommand failed.
```
Frantisek Sumsal [Wed, 25 Jan 2023 11:37:49 +0000 (12:37 +0100)]
basic/linux: update linux uapi headers
IPPROTO_L2TP was moved from linux/l2tp.h to linux/in.h [0], so let's
reflect that change to fix build with newer kernels:
```
In file included from ../src/libsystemd/sd-netlink/netlink-types-genl.c:10:
../src/basic/linux/l2tp.h:16: error: "IPPROTO_L2TP" redefined [-Werror]
16 | #define IPPROTO_L2TP 115
|
In file included from ../src/libsystemd/sd-netlink/netlink-types-genl.c:3:
/usr/include/netinet/in.h:85: note: this is the location of the previous definition
85 | #define IPPROTO_L2TP IPPROTO_L2TP
|
cc1: all warnings being treated as errors
```
When at it, update the rest of the headers we ship as well.
projects / thirdparty / systemd.git / log
