Lennart Poettering [Fri, 22 Oct 2021 22:30:14 +0000 (00:30 +0200)]
stat-util: optimize dir_is_empty_at() a bit, by using getdents64()
That way we have a single syscall only for it, instead of the multiple
readdir() and friends do. And we can operate entirely on the stack, no
malloc() implicit.
Lennart Poettering [Fri, 22 Oct 2021 22:28:24 +0000 (00:28 +0200)]
dirent-util: add FOREACH macro for iterating through getdents64() buffers
We already have a similar loop twice, let's make it easier to read via
an iteration macro.
(The new macro is a bit more careful even, as it verifies the full
dirent fits into the remaining buffer when returning it)
Lennart Poettering [Fri, 22 Oct 2021 22:21:20 +0000 (00:21 +0200)]
dirent-util: move getdents64() related definitions to common header
We want to reuse getdents64() wherever necessary, let's hence move
definitions useful for that into public code.
Lennart Poettering [Thu, 21 Oct 2021 16:07:06 +0000 (18:07 +0200)]
stat-util: specify O_DIRECTORY when reopening dir in dir_is_empty_at()
That way we can fail earlier if the specified fd is not actually a
directory.
(Also, it's not exactly according to standards to open things without
either O_RDONLY/O_RDWR...)
Yu Watanabe [Sun, 24 Oct 2021 11:43:34 +0000 (20:43 +0900)]
udev: do not try to rename interface if it is already up
See dev_change_name() in kernel's net/core/dev.c.
Fixes #21105.
Yu Watanabe [Mon, 25 Oct 2021 14:16:14 +0000 (23:16 +0900)]
Merge pull request #21116 from poettering/test-cleaner
make tests a bit cleaner in regards to oomd handling
Yu Watanabe [Mon, 25 Oct 2021 14:15:47 +0000 (23:15 +0900)]
Merge pull request #21117 from mrc0mmand/last-coverage-related-tweaks
Last batch of coverage-related tweaks
Yu Watanabe [Mon, 25 Oct 2021 14:15:10 +0000 (23:15 +0900)]
Merge pull request #21077 from poettering/mount-setattr
optimize remounting with mount_setattr() syscall
Yu Watanabe [Mon, 25 Oct 2021 07:42:19 +0000 (16:42 +0900)]
nspawn: fix build when SECCOMP is disabled
Follow-up for
20e458ae3c3b386e2b720bbd3c37d95781e69ae7.
Frantisek Sumsal [Mon, 25 Oct 2021 09:35:47 +0000 (11:35 +0200)]
test: merge coverage reports from previous test runs
Relevant mainly for tests which utilize both QEMU and nspawn.
Frantisek Sumsal [Mon, 25 Oct 2021 09:02:22 +0000 (11:02 +0200)]
test: tweak TriggerLimitIntervalSec= when built with coverage
Collecting coverage causes a significant slowdown in general, but since
this test requires certain timing, we need to tweak the defaults to make
it reliably pass.
Lennart Poettering [Mon, 25 Oct 2021 08:39:32 +0000 (10:39 +0200)]
varlink: don't try to talk to oomd from unit tests
Talking to external daemons we ourselves maintain is a job for the
integration tests, not the unit tests. This communication is likely to
fail hence don#t even bother.
This makes our tests a bit cleaner.
Lennart Poettering [Mon, 25 Oct 2021 08:39:12 +0000 (10:39 +0200)]
cgroup: handle gracefully if we can't read oom_kill cgroup attribute
Lennart Poettering [Wed, 20 Oct 2021 21:27:04 +0000 (23:27 +0200)]
mount-util: move opening of /proc/self/mountinfo into bind_remount_one_with_mountinfo()
Let's move things around a bit, and open /proc/self/mountinfo if needed
inside of bind_remount_one_with_mountinfo(). That way bind_remount_one()
can become a superthin inline wrapper around
bind_remount_one_with_mountinfo(). Main benefit is that we don't even
have to open /p/s/mi in case mount_setattr() actually worked for us.
Lennart Poettering [Wed, 20 Oct 2021 21:12:53 +0000 (23:12 +0200)]
mount-util: port over bind_remount_recursive_with_mountinfo() to mount_setattr()
Lennart Poettering [Wed, 20 Oct 2021 20:57:22 +0000 (22:57 +0200)]
mount-util: use modern mount_setattr() syscall for bind_remount_one_with_mountinfo()
New kernels have a nice syscall for changing bind mount flags. Let's use
it. This makes the complex libmount based iteration logic unnecessary.
Luca Boccassi [Sun, 24 Oct 2021 16:00:49 +0000 (17:00 +0100)]
TODO: add note about chase_symlink flag
Yu Watanabe [Sun, 24 Oct 2021 13:54:23 +0000 (22:54 +0900)]
Merge pull request #21108 from mrc0mmand/here-comes-the-coverage
A couple more coverage-related tweaks
Frantisek Sumsal [Sat, 23 Oct 2021 20:10:24 +0000 (22:10 +0200)]
test: rename the global service override file for coverage runs
Otherwise we break TEST-15-DROPIN, since it uses
/usr/lib/systemd/system/service.d/override.conf in some of its
sub-tests.
Frantisek Sumsal [Sat, 23 Oct 2021 19:46:53 +0000 (21:46 +0200)]
README: fix CentOS CI badge
Frantisek Sumsal [Sat, 23 Oct 2021 19:20:37 +0000 (21:20 +0200)]
test: disable ProtectHome= when collecting coverage as well
Depending on the location of the original build dir, either ProtectHome=
or ProtectSystem= may get in the way when creating the gcov metadata
files.
Follow-up to:
*
02d7e73013a92c5580023e4e548d19adbc254dbe
*
6c9efba67715cd1ced170ac46c04d47934ad276a
Yu Watanabe [Sat, 23 Oct 2021 14:58:22 +0000 (23:58 +0900)]
Merge pull request #21082 from yuwata/network-dhcp6-pd-trivial-cleanups
network: DHCPv6 PD: several trivial cleanups
Anssi Hannula [Mon, 18 Oct 2021 13:31:30 +0000 (16:31 +0300)]
efivars: skip writing if variable is already in wanted state
In order to minimize EFI variable NVRAM wear, do not rewrite variables
if they are already in the wanted state (i.e. same data and attributes).
This allows e.g. performing repeat calls of "bootctl install" (which
always rewrites the EFI boot entry) without consuming EFI NVRAM write
cycles.
Lennart Poettering [Fri, 22 Oct 2021 22:42:54 +0000 (00:42 +0200)]
homed: don't drop caches on activation
We should drop caches if we are configured to do so in all cases where
we are done with home dir operations: except if that operation is
activation, because in that case we are not destroying anything, but
leaving it on.
Hence, turn off the flag that reminds us that we should drop caches
before exiting, once activation completed fully,
Follow-up for
86019efa4416d7b548cab321c15bc22a65463786
Yu Watanabe [Fri, 22 Oct 2021 13:32:45 +0000 (22:32 +0900)]
nspawn: ignore --suppress-sync=yes when seccomp is disabled
Follow-up for
4a4654e0241fbeabecb8587fd3520b6b39264b9c.
Fixes #21090.
Lennart Poettering [Fri, 22 Oct 2021 15:34:46 +0000 (17:34 +0200)]
nspawn: bump RLIMIT_NOFILE for nspawn payload similar to how host PID 1 does it for its payload
We try to pass containers roughly the same rlimits as the host gets from
the kernel. However, this means we'd set the RLIMIT_NOFILE to 4K. Which
is quite limiting though, and is something we actually departed from in
PID1: since
52d620757817bc0fa7de3ddbe43024544ced7ea0 we raise the limit
substantially for all userspace.
Given that nspawn is quite often invoked without proper PID1, let's raise the
limits for container payloads the same way as we do from the real PID1
to its service payloads.
Lennart Poettering [Fri, 22 Oct 2021 17:19:56 +0000 (19:19 +0200)]
update TODO
Jan Janssen [Wed, 20 Oct 2021 10:15:03 +0000 (12:15 +0200)]
sd-boot: Add keys to reboot into firmware interface
This is useful if the auto-firmware setting has been disabled. The
keys used here are based on what the majority of firmware employ in
the wild.
This also ensures there's a chance for the user to discover this in
case they were too slow during POST or simply used the wrong ones.
Lennart Poettering [Fri, 22 Oct 2021 17:12:21 +0000 (19:12 +0200)]
Merge pull request #21094 from poettering/userns-split
util-lib: split out userns allocation into new helper function
Lennart Poettering [Fri, 22 Oct 2021 17:12:01 +0000 (19:12 +0200)]
Merge pull request #21093 from poettering/homework-trivial-tweaks
homed trivial refactoring
Lennart Poettering [Fri, 22 Oct 2021 15:57:13 +0000 (17:57 +0200)]
update TODO
Lennart Poettering [Fri, 22 Oct 2021 15:41:17 +0000 (17:41 +0200)]
Merge pull request #21079 from poettering/fd-reopen-directory-tweak
fd-util: when re-opening a directory with fd_reopen() go via openat(……
Lennart Poettering [Fri, 22 Oct 2021 14:37:10 +0000 (16:37 +0200)]
process-util: move sync() out of freeze()
We are using this for creating userns namespaces, and we really
shouldn't try to sync there. Moreover the use of free() in shutdown code
doesn't need it anyway, since it just sync()ed right before anyway. Only
the third user of freeze() we have actually needs the syc(), hence do it
there and nowhere else.
Lennart Poettering [Wed, 20 Oct 2021 11:15:27 +0000 (13:15 +0200)]
namespace-util: introduce userns_acquire() as helper for allocating new unbound userns
This returns a namespace fd, and takes a uidmap/gidmap as string. This
is split out out mount-util.c's remount_idmap() logic, so that we can
allocate a userns independently.
Lennart Poettering [Wed, 20 Oct 2021 11:14:11 +0000 (13:14 +0200)]
basic: move freeze() from shared/exec-util.h to basic/process-util.h
That way we can use it in other code from basic/. It fits into both
headers equally well or badly, hence let's just move this one function.
Lennart Poettering [Fri, 22 Oct 2021 14:08:26 +0000 (16:08 +0200)]
homework: split home_unshare_and_mount() in two
Previously the call did two things, and the second thing was optional
(depending on first arg being NULL). Let's simplify this and just make
it two distinct functions, where one calls the other.
This should make things a bit more readable, given that we called a
function called "…and_mount()" which didn't actually mount...
No actual code changes, just some refactoring.
Lennart Poettering [Fri, 22 Oct 2021 13:51:49 +0000 (15:51 +0200)]
homework: use bit fields where we deal with lots of separate boolean flags
No actual code changes, just making a structure a bit shorter.
Lennart Poettering [Fri, 22 Oct 2021 14:01:42 +0000 (16:01 +0200)]
homework: add macro for "/run/systemd/user-home-mount"
We use this work dir a various places, and it's easy to mistype, hence
let the compiler detect this for us, and introduce a macro for it.
No code changes, just some search/replace.
Yu Watanabe [Fri, 22 Oct 2021 06:49:26 +0000 (15:49 +0900)]
network: dhcp6pd: fix the default value of subnet ID
The type of dhcp6_pd_subnet_id is int64_t.
Lennart Poettering [Thu, 21 Oct 2021 15:29:48 +0000 (17:29 +0200)]
varlink: disconnect varlink link in one more case
Previously we'd possibly see POLLHUP on a varlink link, and continue to
run epoll on it even though we have nothing to read nor write anymore.
Let's fix that, and once we know that there's nothing to write anymore
(or we saw a write error already) we'll disconnect after POLLHUP.
Fixes: #20062
Lennart Poettering [Fri, 22 Oct 2021 12:04:28 +0000 (14:04 +0200)]
test-fd-util: add test case for fd_reopen()
Lennart Poettering [Fri, 22 Oct 2021 12:03:46 +0000 (14:03 +0200)]
fd-util: tweak error handling in fd_reopen()
If we know that /proc/ works, then ENOENT when reopening an fd means the
fd didn't exist. Let's return the correct error code for that, i.e.
EBADF.
Lennart Poettering [Thu, 21 Oct 2021 08:21:03 +0000 (10:21 +0200)]
fd-util: when re-opening a directory with fd_reopen() go via openat(…, ".", …)
This adds a tiny shortcut to fd_reopen(): if we are about to reopen the
fd via O_DIRECTORY then we know it#s a directory and we might as well
reopen it via opening "." using the fd as "at fd" in openat().
This has the benefit that we don't need /proc/self/fd/ around for this
special case: fewer sources of errors.
Lennart Poettering [Fri, 22 Oct 2021 07:39:55 +0000 (09:39 +0200)]
test-networkd-address: fix clock type
The clock to use internally is clock_boottime_or_monotonic(), but the
test used CLOCK_MONOTONIC. After one system suspend the test thus likely
starts to fail.
Jan Janssen [Wed, 20 Oct 2021 08:11:45 +0000 (10:11 +0200)]
sd-boot: Be more precise about secure boot modes
Fixes: #11559
Yu Watanabe [Thu, 21 Oct 2021 07:09:44 +0000 (16:09 +0900)]
network: dhcp6pd: check link state earlier before assigning prefixes to downstream
Also, narrow the acceptable range of the states.
Yu Watanabe [Thu, 21 Oct 2021 08:54:06 +0000 (17:54 +0900)]
network: dhcp6pd: also call dhcp6_pd_prepare() and dhcp6_pd_finalize() for upstream interface
As now the subnet prefix may be assigned to the upstream interface.
Yu Watanabe [Thu, 21 Oct 2021 19:48:40 +0000 (04:48 +0900)]
Merge pull request #21081 from mrc0mmand/even-more-coverage-tweaks
test: collect even more coverage
Jonas Dreßler [Thu, 21 Oct 2021 12:47:35 +0000 (14:47 +0200)]
hwdb: Allow USB autosuspend for MS Surface Pro (2017) Type Cover
The Surface Type Covers are known to support USB autosuspend just fine
(and it's also enabled by default on Windows), so enable it in hwdb.
Yu Watanabe [Wed, 20 Oct 2021 20:47:16 +0000 (05:47 +0900)]
network: dhcp6pd: not necessary to drop routes when Assign=yes
Yu Watanabe [Thu, 21 Oct 2021 16:46:55 +0000 (01:46 +0900)]
Merge pull request #21072 from yuwata/network-address-lifetime
network: address: use usec_t for handling lifetime
Yu Watanabe [Thu, 21 Oct 2021 16:46:15 +0000 (01:46 +0900)]
Merge pull request #21050 from yuwata/network-bpf-cleanups
libsystemd-network: cleanup BPF code
Frantisek Sumsal [Thu, 21 Oct 2021 16:34:21 +0000 (18:34 +0200)]
test: make the coverage check safer for non-compiled builds
Frantisek Sumsal [Thu, 21 Oct 2021 16:12:06 +0000 (18:12 +0200)]
test: collect the coverage _before_ unmounting the rootfs
d'oh!
Frantisek Sumsal [Thu, 21 Oct 2021 16:06:36 +0000 (18:06 +0200)]
test: loosen sandbox restrictions for integration tests as well
Otherwise we miss quite a lot of coverage (mainly from logind,
hostnamed, networkd, and possibly others), since they can't write their
reports with `ProtectSystem=strict`.
Frantisek Sumsal [Thu, 21 Oct 2021 15:34:43 +0000 (17:34 +0200)]
test: loosen certain sandbox restrictions when collecting coverage
With `ProtectSystem=strict` gcov is unable to write the *.gcda files
with collected coverage. Let's add a yet another switch to make such
restriction less strict to make gcov happy.
This addresses following errors:
```
...
systemd-networkd[272469]: profiling:/systemd-meson-build/src/shared/libsystemd-shared-249.a.p/binfmt-util.c.gcda:Cannot open
systemd-networkd[272469]: profiling:/systemd-meson-build/src/shared/libsystemd-shared-249.a.p/base-filesystem.c.gcda:Cannot open
systemd-networkd[272469]: profiling:/systemd-meson-build/src/shared/libsystemd-shared-249.a.p/barrier.c.gcda:Cannot open
systemd-networkd[272469]: profiling:/systemd-meson-build/src/shared/libsystemd-shared-249.a.p/ask-password-api.c.gcda:Cannot open
systemd-networkd[272469]: profiling:/systemd-meson-build/src/shared/libsystemd-shared-249.a.p/apparmor-util.c.gcda:Cannot open
systemd-networkd[272469]: profiling:/systemd-meson-build/src/shared/libsystemd-shared-249.a.p/acpi-fpdt.c.gcda:Cannot open
...
```
Frantisek Sumsal [Thu, 21 Oct 2021 13:59:57 +0000 (15:59 +0200)]
test: wait a bit for the given PID to die if it's still alive
When playing around with the coverage-enabled build I kept hitting
an issue where dnsmasq failed to start because the previous instance was
still shutting down. This should, hopefully, help to mitigate that.
Yu Watanabe [Tue, 19 Oct 2021 11:32:40 +0000 (20:32 +0900)]
icmp6: shorten code a bit
Yu Watanabe [Tue, 19 Oct 2021 11:28:57 +0000 (20:28 +0900)]
dhcp: shorten BPF code a bit
- replace `A = k` and `X = A` with `X = k`,
- replace `A ^= X` and `A == 0` with `A == X`.
- use UINT32_MAX when accept a packet
Yu Watanabe [Tue, 19 Oct 2021 11:24:09 +0000 (20:24 +0900)]
dhcp: rebreak function arguments
Yu Watanabe [Tue, 19 Oct 2021 11:23:47 +0000 (20:23 +0900)]
dhcp: fix assertions
Yu Watanabe [Tue, 19 Oct 2021 11:22:56 +0000 (20:22 +0900)]
dhcp: shorten code a bit
Yu Watanabe [Tue, 19 Oct 2021 11:16:34 +0000 (20:16 +0900)]
lldp: shorten code a bit
Yu Watanabe [Tue, 19 Oct 2021 11:05:11 +0000 (20:05 +0900)]
arp-util: shorten BPF code a bit
- replace `A = k` and `X = A` with `X = k`,
- replace `A ^= X` and `A == 0` with `A == X`.
- use UINT32_MAX when accept a packet
Yu Watanabe [Wed, 20 Oct 2021 17:21:59 +0000 (02:21 +0900)]
network: address: use usec_t for handling lifetime
This drops stuct ifa_cacheinfo from Address, and store lifetime with
usec_t.
Why? Now, all requests of address configurations are once stored in
the request queue, and will be processed when it is ready. So, the
timestamp value passed to the kernel must be evaluated on configure.
This also fixes the following two issues.
- Time values in struct ifa_cacheinfo are stored in uint32_t.
So, the validity check of the address configured by NDisc may fail on
long running systems.
- If a system uses DHCPv6PD, when an interface may appear or be
reconfigured later, then the lifetime value may be inappropriate.
We need to adjust the lifetime with the current time and the timestamp
of the lease.
Yu Watanabe [Thu, 21 Oct 2021 11:50:12 +0000 (20:50 +0900)]
network: ndisc: rename valid_until -> lifetime_usec
Yu Watanabe [Thu, 21 Oct 2021 11:43:57 +0000 (20:43 +0900)]
network: ndisc: add missing lifetime check
Yu Watanabe [Wed, 20 Oct 2021 15:22:49 +0000 (00:22 +0900)]
network: route: rename lifetime -> lifetime_usec
Zbigniew Jędrzejewski-Szmek [Thu, 21 Oct 2021 06:33:08 +0000 (08:33 +0200)]
Merge pull request #21065 from keszybz/efi-warnings
Clean up unused variables in efi code
Yu Watanabe [Wed, 20 Oct 2021 19:26:56 +0000 (04:26 +0900)]
network: drop unnecessary header inclusion
Luca Boccassi [Wed, 20 Oct 2021 21:10:56 +0000 (22:10 +0100)]
CI: disable code coverage in GH Action
It is now ran on the nightly CentOS build, so that it can cover
integration tests too, and not just unit tests. It's nightly as
it considerably increases the integration test runtime, so it's
not appropriate for all PRs.
Frantisek Sumsal [Wed, 20 Oct 2021 17:43:34 +0000 (19:43 +0200)]
test: don't install test-network-generator-conversion.sh w/o networkd
otherwise TEST-02 will fail:
```
=== Failed test log ===
--- test-network-generator-conversion.sh begin ---
+ [[ -n '' ]]
+ [[ -x /usr/lib/systemd/systemd-network-generator ]]
+ [[ -x /lib/systemd/systemd-network-generator ]]
+ exit 1
--- test-network-generator-conversion.sh end ---
```
Before:
```
$ meson build -Dnetworkd=false -Dinstall-tests=true
$ ninja -C build
$ DESTDIR=$PWD/test-install ninja -C build install
$ find test-install/ -name test-network-generator-conversion.sh
test-install/usr/lib/systemd/tests/test-network-generator-conversion.sh
```
After:
```
$ find test-install/ -name test-network-generator-conversion.sh
<no output>
```
Frantisek Sumsal [Wed, 20 Oct 2021 11:00:37 +0000 (13:00 +0200)]
test: collect coverage in a couple more codepaths
Otherwise we'd miss coverage from TEST-{02,24,61}, since they don't go
through the "standard" `check_result_common` hook.
Follow-up to
7bf20e48bd7d641a39a14a7feb749b7e8b0fc0f6.
Yu Watanabe [Wed, 20 Oct 2021 13:44:35 +0000 (22:44 +0900)]
zsh-completion: nspawn: add --suppress-sync option
Follow-up for
4a4654e0241fbeabecb8587fd3520b6b39264b9c.
Yu Watanabe [Wed, 20 Oct 2021 13:42:43 +0000 (22:42 +0900)]
Merge pull request #21051 from poettering/nspawn-no-sync
nspawn: add --suppress-sync=yes mode for turning sync() and friends i…
Luca Boccassi [Wed, 20 Oct 2021 13:38:03 +0000 (14:38 +0100)]
Merge pull request #21055 from yuwata/network-dhcp6-pd-route-lifetime-metric
network: dhcp6pd: set lifetime and route metric
Yu Watanabe [Wed, 20 Oct 2021 13:21:30 +0000 (22:21 +0900)]
Merge pull request #21061 from poettering/direct-io-loopback-tweaks
tweaks to O_DIRECT loop work
Dimitri Papadopoulos [Wed, 20 Oct 2021 07:38:57 +0000 (09:38 +0200)]
Typos found by codespell
Zbigniew Jędrzejewski-Szmek [Wed, 20 Oct 2021 11:57:38 +0000 (13:57 +0200)]
efi: check input parameter
clang complained about an unused variable. We cast the buffer to
struct boot_params, so let's check that it's at least this long.
Zbigniew Jędrzejewski-Szmek [Wed, 20 Oct 2021 11:53:27 +0000 (13:53 +0200)]
efi: drop two unused variables
It seems the code was added in
2553a5482c77980429565d14662f309e385bd472 and
those two were never used for anything.
Lennart Poettering [Tue, 19 Oct 2021 12:56:49 +0000 (14:56 +0200)]
nspawn: add --suppress-sync=yes mode for turning sync() and friends into NOPs via seccomp
This is supposed to be used by package/image builders such as mkosi to
speed up building, since it allows us to suppress sync() inside a
container.
This does what Debian's eatmydata tool does, but for a container, and
via seccomp (instead of LD_PRELOAD).
Lennart Poettering [Wed, 20 Oct 2021 07:57:16 +0000 (09:57 +0200)]
loop-util: call loop_device_make_internal() at the right place
The whole reason loop_device_make_internal() exists (as opposed to just
loop_device_make()) is to avoid mangling the loop flags value/call
getenv twice. Hence let's actually call it when we already mangled the
flags value.
Lennart Poettering [Wed, 20 Oct 2021 07:56:20 +0000 (09:56 +0200)]
loop-util: add debug logging about O_RDWR vs. O_RDONLY + O_DIRECT mode
Once we managed to open the file let's log what we wanted and what we
got.
Lennart Poettering [Wed, 20 Oct 2021 07:55:33 +0000 (09:55 +0200)]
loop-util: minor coding style updates
As suggested here: https://github.com/systemd/systemd/pull/21044#pullrequestreview-
783530343
Yu Watanabe [Wed, 20 Oct 2021 04:30:32 +0000 (13:30 +0900)]
Merge pull request #21041 from yuwata/network-bpf-neighbor
network: introduce BPF to reject netlink messages about non-static neighbor
Yu Watanabe [Wed, 20 Oct 2021 04:29:45 +0000 (13:29 +0900)]
Merge pull request #21056 from yuwata/test-network-cleanups
test-network: several cleanups related to dnsmasq
Yu Watanabe [Thu, 14 Oct 2021 17:00:27 +0000 (02:00 +0900)]
test-network: cleanup dnsmasq related file on setup
Just for safety.
Yu Watanabe [Thu, 14 Oct 2021 16:54:11 +0000 (01:54 +0900)]
test-network: drop pid_file argument from stop_dnsmasq()
Yu Watanabe [Thu, 14 Oct 2021 16:36:08 +0000 (01:36 +0900)]
test-network: use constant variables in dnsmasq command
Yu Watanabe [Thu, 14 Oct 2021 20:57:30 +0000 (05:57 +0900)]
network: dhcp6pd: set default metric 256 for delegated prefix
When Assign= in [DHCPv6PrefixDelegation] is enabled, then the kernel
will create the prefix route for the assigned address with metric 256.
When Assign= is disabled, then the kernel will create the route with
metric 1024.
For the default value, we should choose a smaller value (higher priority)
than 1024, as the unreachable routes for delegated prefix will be
configured with 1024.
Yu Watanabe [Thu, 14 Oct 2021 21:00:22 +0000 (06:00 +0900)]
network: dhcp6: explicitly specify metric for unreachable route
For static IPv6 routes without metric is specified, then we use 1024.
But such an adjustment is not performed to dynamic routes. So, let's
specify the metric explicitly.
Otherwise, configured routes will be handled as foreign.
Yu Watanabe [Thu, 14 Oct 2021 19:21:21 +0000 (04:21 +0900)]
network: dhcp6pd: set lifetime to routes for assigned prefixes
Note that the kernel (at least 5.14.11) seems not to support lifetime
for IPv6 unreachable routes. The lifetime for routes of the type will be
handled by sd-event's timer event source.
So, we cannot confirm the lifetime with 'ip route' command.
Yu Watanabe [Thu, 14 Oct 2021 19:19:59 +0000 (04:19 +0900)]
sd-dhcp6-client: introduce sd_dhcp6_lease_get_timestamp()
Yu Watanabe [Tue, 19 Oct 2021 16:48:44 +0000 (01:48 +0900)]
Merge pull request #21035 from yuwata/network-route-fix-lifetime
network: route: drop kernel version check
Lennart Poettering [Tue, 19 Oct 2021 16:42:25 +0000 (18:42 +0200)]
Merge pull request #21044 from poettering/loopback-direct-io
loop-util: enable LO_FLAGS_DIRECT_IO by default on loopback devices
Daan De Meyer [Tue, 19 Oct 2021 13:54:07 +0000 (14:54 +0100)]
CI: disable opensuse mkosi CI
Until https://github.com/systemd/systemd/issues/21019 is fixed,
there's no point in running the opensuse CI job so let's disable
it for now.
Lennart Poettering [Mon, 18 Oct 2021 20:34:54 +0000 (22:34 +0200)]
loop-util: work around cache invalidation bug in older kernels
Inspired by the discussions in #21003.
Inspired in particular by what Android apexd does:
https://android.googlesource.com/platform/system/apex/+/refs/heads/master/apexd/apexd_loop.cpp
Lennart Poettering [Mon, 18 Oct 2021 15:23:30 +0000 (17:23 +0200)]
loop-util: enable LO_FLAGS_DIRECT_IO by default on loopback devices
Fixes: #21003
Yu Watanabe [Mon, 18 Oct 2021 18:55:24 +0000 (03:55 +0900)]
network: introduce BPF to reject netlink messages about non-static neighbor
Fixes #21031.
Yu Watanabe [Mon, 18 Oct 2021 17:41:38 +0000 (02:41 +0900)]
sd-netlink: introduce sd_netlink_attach_filter()