Daan De Meyer [Tue, 27 Aug 2024 11:17:38 +0000 (13:17 +0200)]
mkosi: Don't apply distribution specific patches
rpm upstream is going to imply --noprep when running with --build-in-place so let's do the same on older
versions of rpm (https://github.com/rpm-software-management/rpm/commit/e0925ad6e3185030b004343f73a5917fdc562d6c)
Also, to keep things consistent between distros, run with --noprepare
on Arch Linux as well (we already skip patches on Debian/Ubuntu).
To keep things working on Arch, we apply the one downstream patch
manually ourselves.
Currently, they unconditionally returns EPOLLIN and USEC_INFINITY, respectively.
Just for consistency with sd-bus, sd-journal, sd-varlink, and so on. All
they have _get_fd(), _get_events(), and _get_timeout().
Yu Watanabe [Sun, 25 Aug 2024 21:24:24 +0000 (06:24 +0900)]
sd-device: make sd_device_new_from_subsystem_sysname() stricter
As workarounded by fc0cbed2db860d163d59d04c32fa6ec30bd0606f, the pair of
subsystem and sysname is not unique. For examples,
- /sys/bus/gpio and /sys/class/gpio, both have gpiochip%N. However, these point to different devpaths.
- /sys/bus/mdio_bus and /sys/class/mdio_bus,
- /sys/bus/mei and /sys/class/mei,
- /sys/bus/typec and /sys/class/typec, and so on.
Let's refuse to provide sd_device object in such cases.
To create the sd_device object of a driver, the function
sd_device_new_from_subsystem_sysname() requires "drivers" for subsystem
and e.g. "pci:iwlwifi" for sysname. Similarly, sd_device_new_from_device_id()
also requires driver subsystem. However, we have never provided a
way to get the driver subsystem ("pci" for the previous example) from
an existing sd_device object.
Before the commit, the function returned 0 on success, but the commit
made the function always return 1, as if device->devtype is NULL, the
function returns -ENOENT in the above.
Fortunately, udev_device_get_devtype() does not propagate any
non-negative value from sd_device_get_devtype(). Hence, hopefully we can
safely revert the change.
Vitaly Kuznetsov [Tue, 27 Aug 2024 10:51:45 +0000 (12:51 +0200)]
stub: restore random seed update logic
Commit 201e0d53bdd43 ("stub: split out random seed part out of run()")
looks like refactoring but apparently it changed the logic when random
seed is refreshed in the ESP completely. Previously, process_random_seed()
was called when either:
- sd-stub was not present (LoaderFeatures var is unset) OR
- sd-stub was present but EFI_LOADER_FEATURE_RANDOM_SEED flag was unset.
Post-change, refresh_random_seed() bails under the exact same conditions (no
sd-stub or EFI_LOADER_FEATURE_RANDOM_SEED is unset) and thus
process_random_seed() is NOT called.
Restore the original logic. efivar_get_uint64_le()'s return value doesn't
require checking: loader_features is initialized to 0 and in case of failure it
stays untouched.
One of the major pait points of managing fleets of headless nodes is
that when something fails at startup, unless debug level was already
enabled (which usually isn't, as it's a firehose), one needs to manually
enable it and pray the issue can be reproduced, which often is really
hard and time consuming, just to get extra info. Usually the extra log
messages are enough to triage an issue.
This new option makes it so that when a service fails and is restarted
due to Restart=, log level for that unit is set to debug, so that all
setup code in pid1 and sd-executor logs at debug level, and also a new
DEBUG_INVOCATION=1 env var is passed to the service itself, so that it
knows it should start with a higher log level. Once the unit succeeds
or reaches the rate limit the original level is restored.
Ronan Pigott [Mon, 26 Aug 2024 08:39:20 +0000 (01:39 -0700)]
resolved: clear the AD bit for bypass packets
When the bypass logic is invoked, such as for queries to the stub with
the DO bit set, be certain to clear the AD bit in the reply before
forwarding it if the answer is not known to be authentic.
Mike Yuan [Sun, 25 Aug 2024 21:21:47 +0000 (23:21 +0200)]
core: honor FileDescriptorName= too for Accept=yes sockets
So far we manually hardcoded $LISTEN_FDNAMES to "varlink" in various
varlink service units we ship, even though FileDescriptorName=varlink
is specified in associated socket units already, because
FileDescriptorName= is currently silently ignored when combined with
Accept=yes. Let's step away from this, which seems saner.
Note that this is technically a compat break, but a mostly negligible
one as there shall be few users setting FileDescriptorName= but
still expecting LISTEN_FDNAMES=connection in the actual executable.
Daan De Meyer [Mon, 26 Aug 2024 11:08:25 +0000 (13:08 +0200)]
mkosi: Stop using git commit timestamps for package releases
This prevents bisecting to figure out which commit broke something
as when going backwards the git commit timestamp will be older meaning
package managers will refuse to upgrade to the "older" version. Let's
make sure the release is always newer by using the current date unless
$SOURCE_DATE_EPOCH is set.
Yu Watanabe [Tue, 20 Aug 2024 18:24:06 +0000 (03:24 +0900)]
man: reword about default route for DNS traffic
DefaultRoute is a D-Bus property, not a valid setting name in .network
files nor resolved.conf.
Whether a link is the default route or not is configured with
DNSDefaultRoute= setting in .network files.
Yu Watanabe [Sat, 24 Aug 2024 20:50:27 +0000 (05:50 +0900)]
network/routing-policy-rule: introduce a generic conf-parser for [RoutingPolicyRule] sectin
This introduce config_parse_routing_policy_rule(), which wraps existing
conf parsers. With this, we can drop many custom conf parsers for
[RoutingPolicyRule], and can reuse generic conf parsers in conf-parser.[ch].
Yu Watanabe [Sat, 24 Aug 2024 19:53:10 +0000 (04:53 +0900)]
conf-parser: return 1 on success
Typically, conf parsers will ignore most errors during parsing strings
and return 0. Let's return 1 on success. Otherwise it is hard to reused
these function in another conf parser.
Sam James [Sat, 24 Aug 2024 12:09:47 +0000 (13:09 +0100)]
meson: search for 'bpf-unknown-none' too
We currently search for 'bpf-gcc' and 'bpf-none-gcc'. Gentoo's
sys-devel/bpf-toolchain package uses 'bpf-unknown-none-gcc', as does Fedora's
cross-binutils. Search for this name too.
varlinkctl: output an expressive error message in case invalid method/interface names are specified
Inspired by #34098 → let's make it easier for users to understand and
correct the mistakes they made: let's early refuse invalid
interface/method names.
Yu Watanabe [Fri, 23 Aug 2024 00:54:00 +0000 (09:54 +0900)]
udev/net_id: move naming scheme check
We usually do not set r = -1 when a functionality is disabled or not
supported. Even though the error code is not used, let's set a negative
errno in such case.
Since ID_NET_NAME_SLOT was introduced we ignore slot == 0
https://github.com/systemd/systemd/blob/0035597a30d120f70df2dd7da3d6128fb8ba6051/src/udev/udev-builtin-net_id.c#L139
Qemu sets _SUN to PCI_SLOT() for all NICs, so _SUN is not unique.
https://gitlab.com/qemu-project/qemu/-/issues/2530
In my tests with libvirt I can only set 'slot="0x00"' in interface definition,
so all NICs end up with _SUN == 0, and this commit is enough to avoid the issue.
Adrian Vovk [Thu, 20 Jun 2024 01:05:04 +0000 (21:05 -0400)]
sysupdate: Repair incomplete versions in-place
A previous commit made sysupdate recognize installed versions where some
transfers are missing. This commit teaches sysupdate how to correctly
repair these incomplete versions.
Previously, if you had a incomplete installation of the OS booted, and
ran sysupdate in an attempt to repair it, sysupdate would make things
worse by creating copies of the currently-booted partitions in the
inactive slots. Then at boot you have two identical partitions, with
identical labels an UUIDs, and end up with a mess.
With this commit, sysupdate is able to recognize situations where it can
simply download the missing transfers and leave the rest of the system
undistrubed.
Partial fix for https://github.com/systemd/systemd/issues/33339
Adrian Vovk [Tue, 2 Jul 2024 16:13:45 +0000 (12:13 -0400)]
sysupdate: Track incompletely-installed versions
When enumerating what versions exist for a given target, sysupdate would
completely throw out any version that's incomplete (where some of the
transfers in the target have that version installed or available, and
other transfers do not).
If we're trying to find what versions we can offer for download, this is
great behavior. If the server side is advertising a partial update to
download, we shouldn't present it to the user.
On the other hand, if we're enumerating what versions we have currently
installed, this is a bad behavior. It makes sysupdate fragile. For
example, if a sysext introduces a new .conf file into
/usr/lib/sysupdate.d, suddenly the currently-installed OS stops being a
version that we've enumerated. Since it's not enumerated, it's not
protected, and so sysupdate will wipe the booted OS.
So if we're looking for installed versions, we now loosen the
restrictions and enumerate incomplete installations.
Partial fix for https://github.com/systemd/systemd/issues/33339
Adrian Vovk [Sat, 22 Jun 2024 00:03:12 +0000 (20:03 -0400)]
sysupdate: Fix resource_find_instance
The current implementation will never find a match, because in the event
of a match instance_cmp falls through to comparing paths and the key
we're matching against will always have a path of NULL.
So let's just use a separate compare function, just to make sure future
updates to instance_cmp don't break resource_find_instance again.
projects / thirdparty / systemd.git / log
