Lennart Poettering [Fri, 15 Nov 2024 08:38:38 +0000 (09:38 +0100)]
boot: explain the 4G quirks we apply to initrd memory allocations
Given how long it took to come to a conclusion of the discussions around
https://github.com/systemd/systemd/issues/35026, let's add a comment
that makes this easier to grok for the next time this comes up.
Follow-up for:
6e207b370e91e681efb08c497a6c8ad78e3c8d83
Yu Watanabe [Fri, 15 Nov 2024 00:07:00 +0000 (09:07 +0900)]
network/nexthop: fix copy-and-paste error
Follow-up for
688f166972916b5cfc9287055582ac1aeef3d486.
Luca Boccassi [Thu, 14 Nov 2024 16:19:25 +0000 (16:19 +0000)]
test: skip TEST-84-STORAGETM if running with bugged libnvme
libnvme 1.11 appears to require a kernel built with NVME TLS
kconfigs, and fails hard if it is not, as the expected
privileged keyring '.nvme' is not present. We cannot just
create it from userspace, as privileged keyrings can only
be created by the kernel itself (those starting with '.').
Skip the test if the library exactly matches this version.
https://github.com/linux-nvme/nvme-cli/issues/2573
Fixes https://github.com/systemd/systemd/issues/35130
Luca Boccassi [Thu, 14 Nov 2024 16:26:01 +0000 (16:26 +0000)]
ukify: Support building UKIs with .dtbauto and .hwids sections (#34158)
Stub behavior will be as following:
1. If there are no `.dtbauto` sections then is used `.dtb` if present
2. If there are `.dtbauto` sections and there is at least one matching
(either with the firmware-provided DT or via `.hwids`) then it'll be
used instead of the `.dtb`.
Based on #28959 and [dtbloader](https://github.com/TravMurav/dtbloader)
Closes #28959
Fixes #31946
Yu Watanabe [Thu, 14 Nov 2024 15:12:40 +0000 (00:12 +0900)]
nspawn: several follow-ups for recent changes (#35146)
Closes #35116.
Yu Watanabe [Thu, 14 Nov 2024 15:10:22 +0000 (00:10 +0900)]
network/ndisc: fix coalescing of ndisc routes when multiple router exists (#35119)
Fixes #33470.
anonymix007 [Sat, 31 Aug 2024 12:51:42 +0000 (15:51 +0300)]
ukify: Support building UKIs with a .hwids section
This section contains a predefined set of HWIDs and the corresponding compatibles to be used in dtb matching
anonymix007 [Tue, 22 Oct 2024 12:27:27 +0000 (15:27 +0300)]
ukify: Support building UKIs with .dtbauto sections
Mike Yuan [Mon, 11 Nov 2024 18:38:36 +0000 (19:38 +0100)]
process-util: refuse FORK_DETACH + FORK_DEATHSIG_*
There's no synchoronization between the intermediate process
and the double-forked child, and the semantics are not useful.
Refuse such combination.
Mike Yuan [Mon, 11 Nov 2024 18:11:38 +0000 (19:11 +0100)]
async: block SIGTERM in asynchronous_rm_rf()
See justifications at https://github.com/systemd/systemd/pull/32235#issuecomment-
2062327783
Luca Boccassi [Thu, 14 Nov 2024 11:36:26 +0000 (11:36 +0000)]
network/netdev: follow-ups for reloading .netdev files (#34979)
Follow-ups for #34909.
Yu Watanabe [Thu, 14 Nov 2024 09:08:35 +0000 (18:08 +0900)]
TEST-13-NSPAWN: add test case for /dev/net/tun
For issue #35116.
Yu Watanabe [Wed, 13 Nov 2024 04:36:11 +0000 (13:36 +0900)]
nspawn: ignore failure in creating /dev/net/tun when --private-network is unspecified
Follow-up for
efedb6b0f3cff37950112fd37cb750c16d599bc7.
Closes #35116.
Yu Watanabe [Wed, 13 Nov 2024 04:17:42 +0000 (13:17 +0900)]
nspawn: split out copy_devnode_one() and bind_mount_devnode() from copy_devnodes()
While doing that, even if mknod() failed, we anyway try to fall back to
use bind mount if arg_uid_shift == 0.
Mostly no functional change, just refactoring and preparation for later commit.
Yu Watanabe [Wed, 13 Nov 2024 04:42:03 +0000 (13:42 +0900)]
nspawn: silence warning about failure in getting fuse version
Follow-up for
dc3223919f663b7c8b8d8d1d6072b4487df7709b.
If nspawn is invoked with DevicePolicy= but DeviceAllow= does not
contain /dev/fuse, nspawn will fail to get fuse version with -EPERM.
Let's silence the warning in that case.
andre4ik3 [Thu, 14 Nov 2024 04:20:09 +0000 (08:20 +0400)]
boot/stub: allocate pages for combined initrds below 4GiB only on x86 (#35149)
Outside of x86, some machines (e.g. Apple silicon, AMD Opteron A1100)
have physical memory mapped above 4GiB, meaning this allocation will
fail, causing the entire boot process to fail on these machines.
This commit makes it so that the below-4GB address space allocation
requirement is only set on x86 platforms, and not on other platforms
(that don't have the specific Linux x86 boot protocol), thereby fixing
boot on those that have no memory mapped below 4GiB in their address
space.
Tested on an Apple silicon M1 laptop and an AMD x86_64 desktop tower.
Fixes: #35026
Yu Watanabe [Thu, 14 Nov 2024 03:15:38 +0000 (12:15 +0900)]
fetch-distro: use git log --first-parent and update debian commit (#35151)
Yu Watanabe [Thu, 14 Nov 2024 03:04:30 +0000 (12:04 +0900)]
logind-session: be more specific about session_kill() errors, plus minor fixes for sd_bus_error handling (#35150)
Yu Watanabe [Thu, 14 Nov 2024 03:02:34 +0000 (12:02 +0900)]
tmpfiles.d/meson.build: two minor tweaks (#35153)
Yu Watanabe [Mon, 11 Nov 2024 17:13:04 +0000 (02:13 +0900)]
network/ndisc: dynamically configure nexthops when routes with gateway are requested
Previously, when multiple routers send RAs with the same preference,
then the kernel merges routes with the same gateway address:
===
default proto ra metric 1024 expires 595sec pref medium
nexthop via fe80::200:10ff:fe10:1060 dev enp0s9 weight 1
nexthop via fe80::200:10ff:fe10:1061 dev enp0s9 weight 1
===
This causes IPv6 Conformance Test v6LC.2.2.11 failure, as reported in #33470.
To avoid the coalescing issue, we can use nexthop, as suggested by Ido Schimmel:
https://lore.kernel.org/netdev/ZytjEINNRmtpadr_@shredder/
> BTW, you can avoid the coalescing problem by using the nexthop API.
> # ip nexthop add id 1 via fe80::200:10ff:fe10:1060 dev enp0s9
> # ip -6 route add default nhid 1 expires 600 proto ra
> # ip nexthop add id 2 via fe80::200:10ff:fe10:1061 dev enp0s9
> # ip -6 route append default nhid 2 expires 600 proto ra
> # ip -6 route
> fe80::/64 dev enp0s9 proto kernel metric 256 pref medium
> default nhid 1 via fe80::200:10ff:fe10:1060 dev enp0s9 proto ra metric 1024 expires 563sec pref medium
> default nhid 2 via fe80::200:10ff:fe10:1061 dev enp0s9 proto ra metric 1024 expires 594sec pref medium
Fixes #33470.
Suggested-by: Ido Schimmel <idosch@idosch.org>
Yu Watanabe [Mon, 11 Nov 2024 16:52:07 +0000 (01:52 +0900)]
network/nexthop: serialize/deserialize nexthops
Yu Watanabe [Mon, 11 Nov 2024 07:19:26 +0000 (16:19 +0900)]
network/nexthop: preparation for dynamically configuring nexthops
Preparation for later commits.
Yu Watanabe [Thu, 14 Nov 2024 02:59:44 +0000 (11:59 +0900)]
network: serialize and deserialize current configuration (#34989)
Replaces #34963.
Fixes #26602.
Fixes #32569.
Yu Watanabe [Tue, 5 Nov 2024 03:17:51 +0000 (12:17 +0900)]
test-network: update KeepConfiguration=dhcp -> dynamic
Yu Watanabe [Tue, 5 Nov 2024 03:15:04 +0000 (12:15 +0900)]
man/network: update documentation for KeepConfiguration=
Yu Watanabe [Tue, 5 Nov 2024 03:14:01 +0000 (12:14 +0900)]
network: rename KeepConfiguration=dhcp -> dynamic
KeepConfiguration=dhcp keeps not only DHCP configurations but
also SLAAC or IPV4LL. Let's rename the value to 'dynamic'.
Yu Watanabe [Wed, 6 Nov 2024 18:40:04 +0000 (03:40 +0900)]
network/ipv4ll: use a foreign IPv4LL address when KeepConfiguration=dhcp
This is similar to what we do for DHCPv4 address, but for IPv4LL
address.
Yu Watanabe [Wed, 6 Nov 2024 18:43:50 +0000 (03:43 +0900)]
network: keep all dynamically acquired configurations when KeepConfiguration=dhcp-on-stop
By the previous commit, configuration source of addresses and routes are
saved on stop and restored on start. Hence, we can keep dynamic
configurations on stop.
Co-authored-by: Jian Zhang <zhangjian.3032@bytedance.com>
Yu Watanabe [Fri, 1 Nov 2024 21:03:09 +0000 (06:03 +0900)]
network: introduce manager_serialize()/deserialize()
Currently, only configuration sources and providers of addresses and
routes are serialized/deserialized.
This should mostly not change behavior, as dynamic (except for DHCPv4)
configurations will be dropped before stopping networkd, and for DHCPv4
protocol, we have already had another logic to handle DHCPv4
configurations.
Preparation for later commits.
Yu Watanabe [Tue, 12 Nov 2024 10:43:57 +0000 (19:43 +0900)]
networkd-test.py: show current status when wait-online failed
For easier debugging on failure.
Yu Watanabe [Fri, 8 Nov 2024 16:26:29 +0000 (01:26 +0900)]
test-network: add tests for reloading .netdev files for independent netdevs
Yu Watanabe [Fri, 8 Nov 2024 17:48:17 +0000 (02:48 +0900)]
network/tuntap: manage tun/tap fds by manager
Otherwise, when a .netdev file for tun or tap netdev is updated,
reloading the file leaks the previous file descriptor.
Yu Watanabe [Fri, 8 Nov 2024 16:46:05 +0000 (01:46 +0900)]
network/bond: do not update several parameters if already up or has slaves
Some bonding parameters cannot be updated when the netdev is already up
or already has at least one slave interface.
Yu Watanabe [Sun, 10 Nov 2024 00:51:32 +0000 (09:51 +0900)]
network/netdev: do not try to update if not supported
Some netdevs cannot update there properties after created.
Let's skip requests in that case.
Yu Watanabe [Mon, 11 Nov 2024 02:13:39 +0000 (11:13 +0900)]
network/netdev: fix counter handling if request is cancelled
Follow-up for
1003093604661bd984574889167f2ff4dfd6209c.
If a netdev is detached for some reasons, then previously the request
was simply cancelled, and the underlying interface never enter the
configured state, as the 'stacked_netdevs_created' flag never set.
This makes the counter decremented manually by the function, and set the
flag. So, the underlying interface can eter the configured state.
Yu Watanabe [Thu, 31 Oct 2024 19:56:12 +0000 (04:56 +0900)]
network/netdev: always queue request of creating netdev then process it later
After PR #34909, networkd tries to update an existing netdev interface if
possible. But, when .netdev files are loaded on start, we have not
enumerate interfaces, so we do not know if the corresponding interface
exists or not. Let's delay processing request a bit.
Yu Watanabe [Fri, 8 Nov 2024 15:58:29 +0000 (00:58 +0900)]
network/netdev: enter ready state only when it is created by us
Follow-up for PR #34909.
This fixes an issue that network interfaces cannot join a master netdev,
like bond or bridge, when the corresponding .netdev is reloaded.
With PR #34909, networkd supports reloading .netdev files. However,
When a .netdev file is modified and reloaded, ifindex is copied from
the old NetDev object to the new one. Thus, even if the interface is
successfully updated, netdev_set_ifindex_impl() will return 0 and
netdev_enter_ready() will never called. If the netdev is a kind of
master netdev, then port interfaces cannot join the master netdev,
as REQUEST_TYPE_SET_LINK_MASTER requires that the master netdev is
in the ready state.
Yu Watanabe [Wed, 13 Nov 2024 02:44:46 +0000 (11:44 +0900)]
network/netdev: do not update MAC address if netdev is already running
Follow-up for
17c5337f7b2993619d84acc2088b2ba1789e6477.
Older kernels (older than v6.5) refuse RTM_NEWLINK messages with IFLA_ADDRESS
attribute when the netdev already exists and is running, even if the MAC
address is unchanged.
So, let's not set IFLA_ADDRESS or IFLA_MTU if they are unchanged, and
set the attributes only when we can update them.
Yu Watanabe [Wed, 13 Nov 2024 05:13:01 +0000 (14:13 +0900)]
network/netdev: set interface name only when creating a new netdev
Otherwise, the kernel older than v6.2 will refuse the netlink message.
Mike Yuan [Wed, 13 Nov 2024 21:30:40 +0000 (22:30 +0100)]
tmpfiles.d/meson: remove the need of specifying empty condition
Mike Yuan [Wed, 13 Nov 2024 21:23:40 +0000 (22:23 +0100)]
tmpfiles.d/meson: call subdir_done() early if tmpfiles is disabled
Luca Boccassi [Wed, 13 Nov 2024 17:03:45 +0000 (17:03 +0000)]
mkosi: update debian commit reference
*
48fabbd5d2 Install new sd-keyutil binary in sd-repart package
*
6dd9ab10fe Update changelog for 257~rc1-4 release
*
6dd325f04b Backport patch to fix TEST-07-PID1 integration test
*
5988cc60ee Update changelog for 257~rc1-3 release
*
cf3a2f7ccc Backport another patch to fix test failure on buildd
*
5d6a226dbb Update changelog for 257~rc1-2 release
*
ebe97c52c8 Backport patch to fix unit test failure on buildd
*
21f63b20bb Update changelog for 257~rc1-1 release
*
0dfec51bbb d/copyright: remove pattern for directory that is no longer present
*
337b3bb2dd Ignore Lintian warning dh-exec-script-without-dh-exec-features
*
b680e6b448 List new libsystemd0 symbols
*
3c00aa000c gbp.conf: use --first-parent for dch to avoid upstream commits
*
d53ecc7769 Install new files
*
546e8c9137 Drop all patches, merged upstream
*
6757597480 Update upstream source from tag 'upstream/257_rc1'
*
4b82805020 gbp.conf: switch upstream branch to full upstream history
*
e60c637a95 gbp.conf: enable signing tags by default
*
2ad27b63c4 Update changelog for 256.7-3 release
*
a212c36c54 systemd-boot: provide integration with shim
Luca Boccassi [Wed, 13 Nov 2024 17:03:35 +0000 (17:03 +0000)]
fetch-distro: use git log --first-parent
We now import the upstream tag in the debian repository, so
this explodes as it tries to walk all upstream commits. Use
--first-parent so that merges only get added via the merge
commit.
Mike Yuan [Wed, 13 Nov 2024 16:41:14 +0000 (17:41 +0100)]
logind-session: be more specific about session_kill() errors
When kill_whom == _ALL, there can be two cases that lead to
ESRCH: the session expects no scope at all or the scope is
not active. Let's distinguish the two cases.
Mike Yuan [Wed, 13 Nov 2024 16:45:53 +0000 (17:45 +0100)]
portable: do not use SYNTHETIC_ERRNO for sd_bus_error_set_errno()
The concept of synthetic errnos is about logging, which
is irrelevant irt bus error and we don't do any special
treatment in sd-bus for them, meaning the value propagated
would be spurious.
Mike Yuan [Wed, 13 Nov 2024 16:26:07 +0000 (17:26 +0100)]
sd-bus/bus-common-errors: reorder one pid1 error to group with others
Lennart Poettering [Fri, 8 Nov 2024 22:06:44 +0000 (23:06 +0100)]
namespace-util: pin pid via pidfd during namespace_open()
Luca Boccassi [Wed, 13 Nov 2024 14:06:21 +0000 (14:06 +0000)]
network/ndisc: fix removal of unnecessary routes (#35128)
Follow-up for
972f1d17ab461a51142a142609dd3ec50bae8440.
This fixes the logic of removing unnecessary routes configured by the
previously received RAs. Previously, we wrongly handled existing routes
could be updated, and unexpected routes would be kept.
Yu Watanabe [Wed, 13 Nov 2024 01:09:05 +0000 (10:09 +0900)]
sd-boot/sd-stub: two log message fixes (#35143)
Fixes: #35033
Fixes: #35100
Yu Watanabe [Wed, 13 Nov 2024 01:08:29 +0000 (10:08 +0900)]
audit-util: return -ENODATA from audit_{session|loginuid}_from_pid() if invoked in a container (#35072)
The auditing subsystem is still not virtualized for containers, hence
the two values don't really make sense inside them, they will just leak
information from outside into the container. Hence don't make use of the
data if we detect we are run inside of a container.
This has visible effects: logind will no longer try to reuse the
auditing session ids as its own session ids when run inside a container.
While are at it, modernize the calls in more ways:
1. switch to pidref behaviour, all but one of our uses are using pidref
anyway already.
2. use read_virtual_file() + proc_mounted()
3. reasonably distinguish ENOENT errors when reading the process proc
files: distinguish the case where /proc is not mounted, from the case
where the process is already gone, from where auditing is not enabled in
the kernel build.
Lennart Poettering [Tue, 12 Nov 2024 21:18:05 +0000 (22:18 +0100)]
ptyfwd: ellipsize overly long window titles
Apparently some terminal emulators have problems with overly long
titles, hence truncate them at some safe length (128).
Also, when parsing ANSI sequences ourselves accept longer sequences
(192), after all we should be fine when parsing our own title sequences.
Fixes: #35104
Mike Yuan [Tue, 12 Nov 2024 18:30:38 +0000 (19:30 +0100)]
various: check meson feature flag early
Prompted by https://github.com/systemd/systemd/pull/35110#discussion_r1835885340
Lennart Poettering [Tue, 12 Nov 2024 22:17:41 +0000 (23:17 +0100)]
pe: use PE_SECTION_VECTOR_IS_SET() macro where appropriate
Lennart Poettering [Tue, 12 Nov 2024 22:15:17 +0000 (23:15 +0100)]
pe: remove unnecessary log message about DT/HWID
Fixes: #35100
Lennart Poettering [Tue, 12 Nov 2024 22:22:39 +0000 (23:22 +0100)]
efi: don't log if EFI RNG isn't ready
Apparently this happens IRL on some systems, let's handle this
gracefully and don't log.
Fixes: #35033
Lennart Poettering [Tue, 12 Nov 2024 13:18:31 +0000 (14:18 +0100)]
run0: when changing privileges to non-root, do not show superhero emoji
Let's show an idcard logo instead, to indicate that we changed ids.
Lennart Poettering [Tue, 12 Nov 2024 14:22:31 +0000 (15:22 +0100)]
dbus-manager: add missing word 'unit' to PK message
Lennart Poettering [Tue, 12 Nov 2024 22:08:45 +0000 (23:08 +0100)]
tree-wide: remove some dead code (#35137)
Lennart Poettering [Thu, 7 Nov 2024 13:54:21 +0000 (14:54 +0100)]
process-util: more gracefully handle oom adjust parsing/setting
Who knows what kind of mount shenanigans people employ, let's gracefully
handle parse failures of proc files, like we alway do otherwsie.
Lennart Poettering [Thu, 7 Nov 2024 16:49:34 +0000 (17:49 +0100)]
audit-util: modernize use_audit() a bit
Use ERRNO_IS_xyz() macros where appropriate.
Also, reduce indentation a bit by inverted early check.
And log in more error codepaths.
Lennart Poettering [Thu, 7 Nov 2024 13:48:57 +0000 (14:48 +0100)]
audit-util: return -ENODATA from audit_{session|loginuid}_from_pid() if invoked in a container
The auditing subsystem is still not virtualized for containers, hence the two
values don't really make sense inside them, they will just leak
information from outside into the container. Hence don't make use of the
data if we detect we are run inside of a container.
This has visible effects: logind will no longer try to reuse the
auditing session ids as its own session ids when run inside a container.
While are at it, modernize the calls in more ways:
1. switch to pidref behaviour, all but one of our uses are using pidref
anyway already.
2. use read_virtual_file() + proc_mounted()
3. reasonable distinguish ENOENT errors when reading the process proc
files: distinguish the case where /proc is not mounted, from the case
where the process is already gone, from where auditing is not enabled
in the kernel build.
Davide Cavalca [Tue, 12 Nov 2024 04:49:36 +0000 (20:49 -0800)]
mkosi: ruff is not available on all distros
Refactor to only install ruff where it is available
Maanya Goenka [Tue, 12 Nov 2024 18:00:23 +0000 (10:00 -0800)]
TODO: Fix typo (#35138)
Replace confex with confext
Lennart Poettering [Mon, 11 Nov 2024 16:49:32 +0000 (17:49 +0100)]
nspawn: fix indentation of run_container() parameter list
Lennart Poettering [Fri, 8 Nov 2024 22:07:00 +0000 (23:07 +0100)]
mntwork: shorten code
Lennart Poettering [Mon, 11 Nov 2024 16:19:26 +0000 (17:19 +0100)]
dissect-image: remove dead code
Lennart Poettering [Fri, 8 Nov 2024 11:14:52 +0000 (12:14 +0100)]
mountfsd: drop unused variable
Antonio Alvarez Feijoo [Tue, 12 Nov 2024 15:18:30 +0000 (16:18 +0100)]
sbsign: remove unused --no-pager option
Davide Cavalca [Tue, 12 Nov 2024 04:44:50 +0000 (20:44 -0800)]
mkosi: Install tpm2-tss-devel to tools for CentOS and Fedora instead of tss2-devel
tss2-devel is the IBM TPM stack, we want the Intel TPM stack, so let's
use the correct package.
Yu Watanabe [Tue, 12 Nov 2024 09:23:47 +0000 (18:23 +0900)]
Fix man page links broken due to incorrect volume numbers (#35122)
Yu Watanabe [Tue, 12 Nov 2024 07:04:16 +0000 (16:04 +0900)]
test-network: several cleanups
- fix verifiers in test_router_preference() to make them actually check
if unnecessary routes are removed,
- stop radv in test_ndisc_vs_static_route() before checking if the static
route is preserved even when the router sends a RA with zero lifetime,
- make verifiers in NetworkdIPv6PrefixTests stricter.
Yu Watanabe [Mon, 11 Nov 2024 17:11:12 +0000 (02:11 +0900)]
network/ndisc: restore the original preference and priority before checking if existing route can be updated
Follow-up for
972f1d17ab461a51142a142609dd3ec50bae8440.
This fixes the logic of removing unnecessary routes configured by the
previously received RAs. Previously, we wrongly handled existing routes
could be updated, and unexpected routes would be kept.
Yu Watanabe [Mon, 11 Nov 2024 19:31:55 +0000 (04:31 +0900)]
network/ndisc: introduce ndisc_route_prepare() and ndisc_router_route_prepare()
These applies common parameters to the route to be requested or removed.
No functional change, just refactoring and preparation for later
commits.
Yu Watanabe [Mon, 11 Nov 2024 17:06:21 +0000 (02:06 +0900)]
network/ndisc: several cleanups for ndisc_remove_route()
- drop unnecessary call of ndisc_set_route_priority() at the beginning,
as it is called later in the loop below,
- use RET_GATHER() and remove all possible routes even if failed.
Yu Watanabe [Mon, 11 Nov 2024 17:03:44 +0000 (02:03 +0900)]
network/ndisc: introduce route_is_bound_to_link() helper function and use it where applicable
No functional change, and preparation for later commits.
Antonio Alvarez Feijoo [Tue, 12 Nov 2024 07:29:28 +0000 (08:29 +0100)]
man/systemd-keyutil: fix rendering typo
Štěpán Němec [Mon, 11 Nov 2024 19:31:43 +0000 (20:31 +0100)]
man/systemd.special: fix a typo
Štěpán Němec [Mon, 11 Nov 2024 19:10:00 +0000 (20:10 +0100)]
man: fix incorrect volume numbers in internal man page references
Some ambiguity (e.g., same-named man pages in multiple volumes)
makes it impossible to fully automate this, but the following
Python snippet (run inside the man/ directory of the systemd repo)
helped to generate the sed command lines (which were subsequently
manually reviewed, run and the false positives reverted):
from pathlib import Path
import lxml
from lxml import etree as ET
man2vol: dict[str, str] = {}
man2citerefs: dict[str, list] = {}
for file in Path(".").glob("*.xml"):
tree = ET.parse(file, lxml.etree.XMLParser(recover=True))
meta = tree.find("refmeta")
if meta is not None:
title = meta.findtext("refentrytitle")
if title is not None:
vol = meta.findtext("manvolnum")
if vol is not None:
man2vol[title] = vol
citerefs = list(tree.iter("citerefentry"))
if citerefs:
man2citerefs[title] = citerefs
for man, refs in man2citerefs.items():
for ref in refs:
title = ref.findtext("refentrytitle")
if title is not None:
has = ref.findtext("manvolnum")
try:
should_have = man2vol[title]
except KeyError: # Non-systemd man page reference? Ignore.
continue
if has != should_have:
print(
f"sed -i '\\|<citerefentry><refentrytitle>{title}"
f"</refentrytitle><manvolnum>{has}</manvolnum>"
f"</citerefentry>|s|<manvolnum>{has}</manvolnum>|"
f"<manvolnum>{should_have}</manvolnum>|' {man}.xml"
)
Yu Watanabe [Mon, 11 Nov 2024 16:07:43 +0000 (01:07 +0900)]
network: forget IPv4 non-local routes when an interface went down (#35099)
Fixes #35047.
Lennart Poettering [Mon, 11 Nov 2024 15:09:07 +0000 (16:09 +0100)]
Introduce systemd-keyutil to do various key/certificate operations (#35095)
Let's gather generic key/certificate operations in a new tool
systemd-keyutil instead of spreading them across various special purpose
tools.
Fixes #35087
Yu Watanabe [Fri, 8 Nov 2024 13:55:20 +0000 (22:55 +0900)]
test-network: add test case for issue #35047
Yu Watanabe [Mon, 11 Nov 2024 04:00:10 +0000 (13:00 +0900)]
network/nexthop: also forget IPv4 nexthops when an interface went down
Similar to the previous commit, but for nexthop.
Yu Watanabe [Fri, 8 Nov 2024 13:07:53 +0000 (22:07 +0900)]
network/route: forget IPv4 non-local routes when an interface went down
When an interface went down, IPv4 non-local routes are removed by the
kernel without any notifications. Let's forget the routes in that case.
Fixes #35047.
Yu Watanabe [Mon, 11 Nov 2024 03:26:48 +0000 (12:26 +0900)]
network/nexthop: forget dependent routes without trying to remove
When a nexthop is removed, routes depend on the removed nexthop are
already removed. It is not necessary to remove them, as already
commented. Let's forget them without trying to remove.
Yu Watanabe [Mon, 11 Nov 2024 04:51:49 +0000 (13:51 +0900)]
network/nexthop: do not remove depending nexthops when a nexthop is removed
Previously, when a nexthop is removed, depending nexthops were removed, but
that's not necessary, as the kernel keeps them, at least with v6.11.
Yu Watanabe [Mon, 11 Nov 2024 07:05:54 +0000 (16:05 +0900)]
network/route: update reference of the route from nexthop
Follow-up for
6f09031e4d04727cc72164fefcbc763e37556493.
The function has been introduced by the commit, but it has never been used...
Luca Boccassi [Mon, 11 Nov 2024 12:59:31 +0000 (12:59 +0000)]
network: further rework for reconfiguring interfaces (#35059)
Follow-ups for #35035.
Split-out of #34989.
Fixes #35092.
Yu Watanabe [Mon, 11 Nov 2024 09:36:26 +0000 (18:36 +0900)]
Split src/partition (#35110)
Gabriel Elyas [Mon, 11 Nov 2024 08:36:51 +0000 (09:36 +0100)]
po: Translated using Weblate (Portuguese (Brazil))
Currently translated at 96.1% (247 of 257 strings)
Co-authored-by: Gabriel Elyas <gabrielelyas@protonmail.com>
Translate-URL: https://translate.fedoraproject.org/projects/systemd/main/pt_BR/
Translation: systemd/main
Zbigniew Jędrzejewski-Szmek [Sun, 10 Nov 2024 13:05:08 +0000 (14:05 +0100)]
Rename src/partition to src/repart
Yu Watanabe [Wed, 6 Nov 2024 18:26:39 +0000 (03:26 +0900)]
network: make 'networkctl reconfigure' work safely even when KeepConfiguration=dhcp or yes
Previously, even if KeepConfiguration=dhcp or yes is specified in the
new .network file, dynamic configurations like DHCP address and routes
were dropped when 'networkctl reconfigure INTERFACE' is invoked.
If the setting is specified, let's gracefully handle the dynamic
configurations. Then, 'networkctl reconfigure' can be also used for
an interface that has critical connections.
Yu Watanabe [Wed, 6 Nov 2024 18:02:54 +0000 (03:02 +0900)]
network: drop static configs later
Follow-up for
dd6d53a8dc58c5e6e310b09ba7f7a22600a87ba9.
Unnecessary static configs will be anyway dropped later in
link_configure() -> link_drop_unmanaged_config(). Hence, even if we are
reconfiguring an interface cleanly, it is not necessary to drop static
configs here.
Yu Watanabe [Thu, 7 Nov 2024 00:16:06 +0000 (09:16 +0900)]
network/dhcp-pd: do not remove unreachable route when reconfiguring non-upstream interface
Unreachable routes are not owned by any interfaces, and its ifindex is
zero. Previously, if a non-upstream interface is reconfigured, all routes
including unreachable routes configured by the upstream interface are
removed.
This makes unreachable routes are always handled by the upstream interface,
and only removed when the delegated prefixes are changed or lost.
Yu Watanabe [Thu, 7 Nov 2024 01:13:37 +0000 (10:13 +0900)]
network: reorder dropping dynamic configuration
Follow-up for
451c2baf30f50b95d73e648058c7c2348dbf0c31.
Yu Watanabe [Thu, 7 Nov 2024 00:52:03 +0000 (09:52 +0900)]
test-network: reconfigure interface cleanly to drop previous DHCP lease and friends
Follow-up for
451c2baf30f50b95d73e648058c7c2348dbf0c31.
With the commits, reloading .network files does not release previously
acquired DHCP lease and friends if possible.
On graceful reconfigure triggered by the reload, the interface may
acquire a new DHCPv4 lease earlier than DHCPv6 lease. In that case,
the check will fail as it is done with the new DHCPv4 lease and old
DHCPv6 lease, which does not contain any IPv6 DNS servers or so.
So, when switching from no -> yes, we need to wait a new lease with DNS
servers or so. To achieve that, we need to clean reconfigure the interface.
Yu Watanabe [Thu, 7 Nov 2024 00:52:03 +0000 (09:52 +0900)]
network: reset 'configured' flags even if we keep DHCP lease and friends on reconfigure
Follow-up for
451c2baf30f50b95d73e648058c7c2348dbf0c31.
With the commits, reloading .network files does not release previously
acquired DHCP lease and friends if possible. If previously a DHCP client
was configured as not requesting DNS servers or so, then the previously
acquired lease might not contain any DNS servers. In that case, if the
new .network file enables UseDNS=, then the interface should enter the
configured state after a new lease is acquired. To achieve that, we need
to reset the flags.
With this change, the workaround applied to the test by the commit
451c2baf30f50b95d73e648058c7c2348dbf0c31 can be dropped.
Yu Watanabe [Thu, 7 Nov 2024 03:00:38 +0000 (12:00 +0900)]
network: drop unnecessary size specifier
It does not save any memory usage but increase code complexity.
Yu Watanabe [Fri, 8 Nov 2024 11:38:42 +0000 (20:38 +0900)]
netwrok: call link_drop_unmanaged_config() earlier in link_configure()
Otherwise, even if a link enters the configuring state at the beginning
of link_configure(), link_check_ready() may be called before
link_drop_unmanaged_config() is called, and the link may enter the
configured state.
Fixes #35092.
Zbigniew Jędrzejewski-Szmek [Sun, 10 Nov 2024 13:01:34 +0000 (14:01 +0100)]
Move growfs+makefs to src/growfs/
Those two programs are used together and it makes sense to keep them
together. makefs is smaller, so name the directory after growfs.
Yu Watanabe [Sun, 10 Nov 2024 02:19:10 +0000 (11:19 +0900)]
Various multi-dt fixes and CHID test (#35056)
Part of #34158