Currently it is only possible to add upstream _system_ units. The option
systemd.additionalUpstreamSystemUnits can be used for this.
However, this was not yet possible for systemd.user. In a similar
fashion this was added to systemd-user.nix.
This is intended to have other modules add upstream units.
Use a quoted heredoc to inject installBootLoader safely into the script,
and restore the previous invocation of `system` with a single argument so
that shell commands keep working.
As of systemd/systemd@e908434458,
systemd-networkd now automatically configures routes to addresses
specified in AllowedIPs unless explicitly disabled with
"RouteTable=off".
This bug is so obscure and unlikely that I was honestly not able to
properly write a test for it. What happens is that we are calling
handleModifiedUnit() with $unitsToStart=\%unitsToRestart. We do this to
make sure that the unit is stopped before it's started again which is
not possible by regular means because the stop phase is already done
when calling the activation script.
recordUnit() still gets $startListFile, however which is the wrong file.
The bug would be triggered if an activation script requests a service
restart for a service that has `stopIfChanged = true` and
switch-to-configuration is killed before the restart phase was run. If
the script is run again, but the activation script is not requesting
more restarts, the unit would be started instead of restarted.
When initializing a system (e.g. first boot / livecd) we have no good
reference source for time. systemd-timesyncd however would revert back
to its configured fallback time (in our case 01.01.1980). Since we
probably don't want to hardcode a specific date as fallback we are now
using the current system time (wherever that might have come from) to
initialize the reference clock file.
The only systems that might be remotely affected by this change are
machines that have highly unreliable RTCs or those where the battery
that backs the RTC is running empty.
Historically these systems always had a tough time with anything time
related and likely required manual intervention.
For stateless systems (those that wipe / between reboots or our
installer CDs) this has the consequence that time will always be reset
to whatever the system comes up with on boot. This is likely the correct
time coming from an RTC. No harm done here the situation is likely
unchanged for them.
For stateful systems (those that retain the / partition across reboots)
there shouldn't be a change at all. They'll provide an initial clock
value once on their lifetime (during first boot / after installation).
From then onwards systemd-timesyncd will update the file with the newer
fallback time (that will be picked up on the next boot).
This effectively fixes the majority of all VM tests which were broken
because `/dev/vda` (or any other block device) wasn't mountable:
machine # mounting /dev/vda on /...
machine # mount: mounting /dev/vda on /mnt-root/ failed: No such device[ 2.820976] Kernel panic - not syncing: Attempted to kill init! exitcode=0x00000100
machine # [ 2.821757] CPU: 0 PID: 1 Comm: init Not tainted 5.10.72 #1-NixOS
machine # [ 2.821757] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.14.0-0-g155821a1990b-prebuilt.qemu.org 04/01/2014
machine # [ 2.821757] Call Trace:
machine # [ 2.821757] dump_stack+0x6b/0x83
machine # [ 2.821757] panic+0x101/0x2c8
machine # [ 2.821757] do_exit.cold+0x14/0xb3
machine # [ 2.821757] do_group_exit+0x33/0xa0
machine # [ 2.821757] __x64_sys_exit_group+0x14/0x20
machine # [ 2.821757] do_syscall_64+0x33/0x40
machine # [ 2.821757] entry_SYSCALL_64_after_hwframe+0x44/0xa9
machine # [ 2.821757] RIP: 0033:0x7f67ec2800f6
machine # [ 2.821757] Code: 00 4c 8b 0d 2c 5d 11 00 eb 19 66 2e 0f 1f 84 00 00 00 00 00 89 d7 89 f0 0f 05 48 3d 00 f0 ff ff 77 22 f4 89 d7 44 89 c0 0f 05 <48> 3d 00 f0 ff ff 76 e2 f7 d8 64 41 89 01 eb da 66 2e 0f 1f 84 00
machine # [ 2.821757] RSP: 002b:00007fff8f5a71d8 EFLAGS: 00000202 ORIG_RAX: 00000000000000e7
machine # [ 2.821757] RAX: ffffffffffffffda RBX: 0000000000699704 RCX: 00007f67ec2800f6
machine # [ 2.821757] RDX: 0000000000000001 RSI: 000000000000003c RDI: 0000000000000001
machine # [ 2.821757] RBP: 0000000000000004 R08: 00000000000000e7 R09: ffffffffffffff80
machine # [ 2.821757] R10: 00007f67ec33f3e0 R11: 0000000000000202 R12: 000000000000000b
machine # [ 2.821757] R13: 00007fff8f5a75a8 R14: 0000000000000000 R15: 00000000004fc198
machine # [ 2.821757] Kernel Offset: 0x31e00000 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffffbfffffff)
machine # [ 2.821757] Rebooting in 1 seconds..
This happened because the kernel failed to load modules such as `ext4`
from `boot.initrd.availableKernelModules`[1] on e.g. a `mount(2)` syscall.
The problem is that `kmod` isn't linked against `libpthread.so.0`
anymore because it got merged into `libc.so.6` (however, the .so still
exists), but still needs it:
machine # newfstatat(AT_FDCWD, "/nix/store/eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee-glibc-2.34-36/lib/x86_64", 0x7ffd951114c0, 0) = -1 ENOENT (No such file or directory)
machine # openat(AT_FDCWD, "/nix/store/eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee-glibc-2.34-36/lib/x86_64/libpthread.so.0", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory)
machine # newfstatat(AT_FDCWD, "/nix/store/eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee-glibc-2.34-36/lib/x86_64", 0x7ffd951114c0, 0) = -1 ENOENT (No such file or directory)
machine # openat(AT_FDCWD, "/nix/store/eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee-glibc-2.34-36/lib/libpthread.so.0", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory)
machine # newfstatat(AT_FDCWD, "/nix/store/eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee-glibc-2.34-36/lib", 0x7ffd951114c0, 0) = -1 ENOENT (No such file or directory)
machine # openat(AT_FDCWD, "/nix/store/eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee-glibc-2.34-36/etc/ld.so.cache", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory)
machine # writev(2, [{iov_base="/nix/store/kdc9n48ksdc1a8y8w512w"..., iov_len=69}, {iov_base=": ", iov_len=2}, {iov_base="error while loading shared libra"..., iov_len=36}, {iov_base=": ", iov_len=2}, {iov_base="libpthread.so.0", iov_len=15}, {iov_base=": ", iov_len=2}, {iov_base="cy
machine # ) = 184
machine # exit_group(127) = ?
machine # +++ exited with 127 +++
machine # mount: mounting /dev/vda on /mnt-root/ failed: No such device
machine # [ 19.167180] Kernel panic - not syncing: Attempted to kill init! exitcode=0x00000100
machine # [ 19.167711] CPU: 0 PID: 1 Comm: init Not tainted 5.10.72 #1-NixOS
This is not a problem
* inside stage-1 because `LD_LIBRARY_PATH` points to `$out/lib` of
extra-utils where `libpthread.so.6` also exists.
* on a running system because `${pkgs.glibc}/lib` is part of kmod's
rpath.
However this is a problem inside the kernel which calls `modprobe` (in
our case `kmod`) to load modules and doesn't know about
`LD_LIBRARY_PATH`. Also, the rpath-reference was nuked.
To work around this, the kernel's `modprobe`
(i.e. `/proc/sys/kernel/modprobe`) now points to a wrapper which
explicitly declares `LD_LIBRARY_PATH`. We can't use `makeWrapper` here
because `modprobe` itself must not be renamed. Otherwise, `kmod` (which
is the link-target of `modprobe`) won't work because it expects
`argv[0] == "modprobe"` to perform modprobe's tasks.
[1] https://nixos.org/manual/nixos/stable/options.html#opt-boot.initrd.availableKernelModules
systemd needs this so special characters (like the ones in wireguard
units that appear because they are part of base64) can be escaped using
the \x syntax.
Root of the issue is that `glob()` handles the backslash internally
which is obviously not what we want here.
Also add a test case and fix some perlcritic issues in the subroutine.
wtmp and btmp are created by systemd, so the rules are more appropriate there.
They can be disabled explicitly with something like
services.ogrotate.paths = {
"/var/log/btmp".enable = false;
"/var/log/wtmp".enable = false;
};
if required.
This is accomplished by comparing the hashes that the unit files
contain. By filtering for a special key `X-Reload-Triggers` in the
`[Unit]` section, we can differentiate between reloads and restarts.
Since activation scripts can request reloads of units as well, more
checking of this behaviour is implemented. If a unit is to be restarted,
it's never reloaded as well which would make no sense.
Also removes a useless subroutine and perl dependencies that are
nowadays handled by the propagated build inputs feature of
`perl.withPackages`.
The mount options need to be passed as a comma-separated list of options so that they
end up one a single Options line in the resulting mount unit.
The current code passed the options as a list, resulting in several Options lines in
the mount unit, all but the first of these were ignored by systemd however.
This behaviour is not clearly defined in the systemd man page.
The `nix.*` options, apart from options for setting up the
daemon itself, currently provide a lot of setting mappings
for the Nix daemon configuration. The scope of the mapping yields
convience, but the line where an option is considered essential
is blurry. For instance, the `extra-sandbox-paths` mapping is
provided without its primary consumer, and the corresponding
`sandbox-paths` option is also not mapped.
The current system increases the maintenance burden as maintainers have to
closely follow upstream changes. In this case, there are two state versions
of Nix which have to be maintained collectively, with different options
avaliable.
This commit aims to following the standard outlined in RFC 42[1] to
implement a structural setting pattern. The Nix configuration is encoded
at its core as key-value pairs which maps nicely to attribute sets, making
it feasible to express in the Nix language itself. Some existing options are
kept such as `buildMachines` and `registry` which present a simplified interface
to managing the respective settings. The interface is exposed as `nix.settings`.
Legacy configurations are mapped to their corresponding options under `nix.settings`
for backwards compatibility.
Various options settings in other nixos modules and relevant tests have been
updated to use structural setting for consistency.
The generation and validation of the configration file has been modified to
use `writeTextFile` instead of `runCommand` for clarity. Note that validation
is now mandatory as strict checking of options has been pushed down to the
derivation level due to freeformType consuming unmatched options. Furthermore,
validation can not occur when cross-compiling due to current limitations.
A new option `publicHostKey` was added to the `buildMachines`
submodule corresponding to the base64 encoded public host key settings
exposed in the builder syntax. The build machine generation was subsequently
rewritten to use `concatStringsSep` for better performance by grouping
concatenations.
[1] - https://github.com/NixOS/rfcs/blob/master/rfcs/0042-config-option.md
This option behaves exactly like `boot.extraModprobeConfig`, except that it also includes the generated modprobe.d file in the initrd.
Many years ago, someone tried to include the normal modprobe.d/nixos.conf file generated by `boot.extraModprobeConfig` in the initrd: 0aa2c1dc46. This file contains a reference to a directory with firmware files inside. Including firmware in the initrd made it too big, so the commit was reverted again in 4a4c051a95.
The `boot.extraModprobeConfig` option not changing the initrd caused me much confusion because I tried to set the maximum cache size for ZFS and it didn't work.
Closes https://github.com/NixOS/nixpkgs/issues/25456.
Modules that do not depend on e.g. toplevel should not have to include it just to set
things in `system.build`. As a general rule, this keeps tests simple, usage flexible
and evaluation fast. While one module is insignificant, consistency and good practices
are.
If the Nix daemon has never been enabled (nix.enable has always been
set to false), the gcroots directory won't exist. If the Nix daemon
is later enabled, the GC roots for booted-system and current-system
will be missing, and they might end up being garbage collected. Since
it's cheap to add GC roots even if the daemon will never be enabled,
let's just always add them so we're okay in the case where the daemon
is enabled later.
- Fully get rid of `parseKeyValues` and use systemctl features for that
- Add some regex modifiers recommended by perlcritic
- Get rid of a postfix if
- Sort units when showing their status
- Clean the logic for showing what failed from `elif` to `next`
- Switch from `state` to `substate` for `auto-restart` because that's
actually where the value is stored
- Show status of units with one single systemctl call and get rid of
COLUMNS in favor of --full
- Add a test for failing units
This replaces the naive K=V unit parser with a proper INI parser from a
library and adds proper support for override files. Also adds a bunch of
comments about parsing, I hope this makes it easier to understand and
maintain in the future.
There are multiple reasons to do so, the first one is just general
correctness with is nice imo. But to get to more serious reasons (I
didn't put in all that effort for nothing) is that this is the first
step torwards more clever restart/reload handling. By using a library
like Data::Compare a future PR could replace the current way of
fingerprinting units (which is to compare store paths) by comparing the
hashes. This is more precise because units won't get restarted because
the order of the options change, comments are added, some dependency of
writeText changes, .... Also this allows us to add a feature like
`X-Reload-Triggers` so the unit can either be reloaded when these change
or restarted when everything else changes, giving module authors the
ability to have their services reloaded without having to fear that
updates are not applied because the service doesn't get restarted.
Another reason why this feature is nice is that now that the unit files
are parsed correctly (and values are just extracted from one section),
potential future rewrites can just rely on some INI library without
having to implement their own weird parser that is compatible with this
script.
This also comes with a new subroutine to handle systemd booleans because
I thought the current way of handling it was just ugly. This also allows
overriding values this script reads in an override file.
Apart from making this script more compatible with the world around it,
this also fixes two issues I saw bugging exactly 0 (zero) people. First
is that this script now supports multiple override files, also ones that
are not called override.conf and the second one is that `1` and `on` are
treated as bools by systemd but were previously not parsed as such by
switch-to-configuration.
This removes `/run/nixos/activation-reload-list` (which we will need in
the future when reworking the reload logic) and makes
`/run/nixos/activation-restart-list` honor `restartIfChanged` and
`reloadIfChanged`. This way activation scripts don't have to bother with
choosing between reloading and restarting.
I was confused why I couldn't find a mention of udev.log_priority in
systemd-udevd.service(8). It turns out that it was renamed[1] to
udev.log_level. The old name is still accepted, but it'll avoid
further confusion if we use the new name in our documentation.
[1]: 64a3494c3d
most modules can be evaluated for their documentation in a very
restricted environment that doesn't include all of nixpkgs. this
evaluation can then be cached and reused for subsequent builds, merging
only documentation that has changed into the cached set. since nixos
ships with a large number of modules of which only a few are used in any
given config this can save evaluation a huge percentage of nixos
options available in any given config.
in tests of this caching, despite having to copy most of nixos/, saves
about 80% of the time needed to build the system manual, or about two
second on the machine used for testing. build time for a full system
config shrank from 9.4s to 7.4s, while turning documentation off
entirely shortened the build to 7.1s.
Add `shellDryRun` to the generic stdenv and substitute it for uses of
`${stdenv.shell} -n`. The point of this layer of abstraction is to add
the flag `-O extglob`, which resolves#126344 in a more direct way.
some options have default that are best described in prose, such as
defaults that depend on the system stateVersion, defaults that are
derivations specific to the surrounding context, or those where the
expression is much longer and harder to understand than a simple text
snippet.
The first one doesn't make any sense because the directory where the
init binary resides does not contain other tools we need like
systemd-escape.
The second one doesn't make sense either because the errors are already
ignored.
Add `systemd.network.networks.*.dhcpServerStaticLeaseConfig` to allow
for configuring static DHCP leases through the `[DHCPServerStaticLease]`
section. See systemd.network(5) of systemd 249 for details.
Also adds the NixOS test `systemd-networkd-dhcpserver-static-lease` to
test the assignment of static leases.
This reverts commit 57961d2b83, reversing
changes made to b04f913afc.
(I.e. this reverts PR #141192.)
While well-intended, this change does unfortunately introduce very
serious regressions that are especially disruptive/noticeable on desktop
systems (e.g. users of Sway will loose their graphical session when
running "nixos-rebuild switch").
Therefore, this change has to be reverted ASAP instead of trying to fix
it in "production".
Note: An updated version should be extensively discussed, reviewed, and
tested before re-landing this change as an earlier version also had to
be reverted for the exact same issues [0].
Fix: #146727
[0]: https://github.com/NixOS/nixpkgs/pull/73871#issuecomment-559783752
mkDerivedConfig : Option a -> (a -> Definition b) -> Definition b
Create config definitions with the same priority as the definition of another option.
This should be used for option definitions where one option sets the value of another as a convenience.
For instance a config file could be set with a `text` or `source` option, where text translates to a `source`
value using `mkDerivedConfig options.text (pkgs.writeText "filename.conf")`.
It takes care of setting the right priority using `mkOverride`.
Before this change, one could set environment.etc.*.text and .source.
.source would always take precedence, regardless of the priorities set.
This change means that if, for instance, .text is set with mkForce but
.source is set normally, the .text content will be the one to take
effect. If they are set with the same priority they will conflict.
The regexp was only matching numbers and not the '.', so everyone using
systemd-boot would always see `could not find any previously installed
systemd-boot` on a `nixos-rebuild`.
By using the new extendModules function to produce the specialisations,
we avoid reimplementing the eval-config.nix logic in reverse and fix
cross compilation support for specialisations in the process.
Some specialisations (such as those which affect various boot-time
attributes) cannot be switched to at runtime. This allows picking the
specialisation at boot time.
/etc/crypttab can contain the _netdev option, which adds crypto devices
to the remote-cryptsetup.target.
remote-cryptsetup.target has a dependency on cryptsetup-pre.target. So
let's add both of them.
Currently, one needs to manually ssh in and invoke `systemctl start
systemd-cryptsetup@<name>.service` to unlock volumes.
After this change, systemd will properly add it to the target, and
assuming remote-cryptsetup.target is pulled in somewhere, you can simply
pass the passphrase by invoking `systemd-tty-ask-password-agent` after
ssh-ing in, without having to manually start these services.
Whether remote-cryptsetup.target should be added to multi-user.target
(as it is on other distros) is part of another discussion - right now
the following snippet will do:
```
systemd.targets.multi-user.wants = [ "remote-cryptsetup.target" ];
```
This makes the order of operations the same in dry-activate and a "true"
activate. Also fixes the indentation I messed up and drop a useless
unlink() call (we are already unlinking that file earlier).
The previous logic failed to detect that units were socket-activated
when the socket was stopped before switch-to-configuration was run. This
commit fixes that and also starts the socket in question.
The first FIXME is removed because it doesn't make sense to use
/proc/1/exe since that points to a directory that doesn't have all tools
the activation script needs (like systemd-escape).
The second one is removed because there is already no error handling
(compare with the restart logic where the return code is checked).
This commit changes a lot more that you'd expect but it also adds a lot
of new testing code so nothing breaks in the future. The main change is
that sockets are now restarted when they change. The main reason for
the large amount of changes is the ability of activation scripts to
restart/reload units. This also works for socket-activated units now,
and honors reloadIfChanged and restartIfChanged. The two changes don't
really work without each other so they are done in the one large commit.
The test should show what works now and ensure it will continue to do so
in the future.
allows configuration of foo-over-udp decapsulation endpoints. sadly networkd
seems to lack the features necessary to support local and peer address
configuration, so those are only supported when using scripted configuration.
Checks whether all spaces are inside double quotes, thus ensuring that one
string parses as no more than one kernel param.
Co-authored-by: pennae <82953136+pennae@users.noreply.github.com>
On some systems bootctl cannot write the `LoaderSystemToken` EFI variable
during installation, which results in a failure to install the boot
loader. Upstream provides a flag (--graceful) to ignore such write failures -
this change exposes it as a configuration option.
As the exact semantics of this option appear to be somewhat volatile it
should be used only if systemd-boot otherwise fails to install.
The multipath-tools package had existed in Nixpkgs for some time but
without a nixos module to configure/drive it. This module provides
attributes to drive the majority of multipath configuration options
and is being successfully used in stage-1 and stage-2 boot to mount
/nix from a multipath-serviced iSCSI volume.
Credit goes to @grahamc for early contributions to the module and
authoring the NixOS module test.
This change is strictly functionally equivalent because we're just
lifting the transformation of `source` out of `mapAttrs` to the single point of
use (in escapeShellArgs).
This is also much faster because we can skip a map over all `etc` items.
Fixes:
```
activating the configuration...
warning: user ‘systemd-coredump’ has unknown group ‘systemd-coredump’
setting up /etc...
```
Oversight of #133166
When cross-compiling, we can't run the runtime shell to check syntax
if it's e.g. for a different architecture. We have two options here.
We can disable syntax checking when cross compiling, but that risks
letting errors through. Or, we can do what I've done here, and change
the syntax check to use stdenv's shell instead of the runtime shell.
This requires the stdenv shell and runtime shell to be broadly
compatible, but I think that's so ingrained in Nixpkgs anyway that
it's fine. And this way we avoid conditionals that check for cross.
The primary use case is tools like sops-nix and agenix to restart units
when secrets change. There's probably other reasons to restart units as
well and a nice thing to have in general.
The treatment of the "source" parameter changed
with eb7120dc79, breaking stuff.
Before that commit, the source parameter was converted to a
string by implicit coercion, which would copy the file to the
store and yield an string containing the store path. Now, by
the virtue of escapeShellArg, toString is called explicitly on
that path, which will yield an string containing the absolute
path of the file.
This commit restores the old behavior.
- boost 167 removed on staging-next (7915d1e03f) × boost attributes are inherited on staging (d20aa4955d)
- linux kernels were moved to linux-kernels.nix on staging-next (c62f911507) × hardened kernels are versioned on staging (a5341beb78) + removed linux_5_12 (e55554491d)
- conflict in node-packages – I regenerated it using node2nix from nixos-unstable (does not build on staging)
The main goal of this commit is to replace the rather fragile passing of
multiple arrays which could break in cases like #130935.
While I could have just added proper shell escaping to the variables
being passed, I opted for the more painful approach of replacing the
fragile and somewhat strange construct with the 5 bash lists. While
there are currently no more problems present with the current approach
(at least none that I know of), the new approach seems more solid and
might get around problems that could arise in the future stemming from
either the multiple-lists situation or from the absence of proper shell
quoting all over the script.
systemd-coredump tries to drop privileges to a systemd-coredump user if
present (and falls back to the root user if it's not available).
Create that user, and recycle uid 151 for it. We don't really care about
the gid.
Fixes https://github.com/NixOS/nixpkgs/issues/120803.
This modifies initialRamdiskSecretAppender to stage secrets in
/.initrd-secrets/ and stage-1-init to copy them into place after mounting
special file systems. This allows secrets to be copied into ramfs mounts
like /run/keys for use after stage-1 finishes without copying them to disk
(which would not be very secure).
- The order of NSS (host) modules has been brought in line with upstream
recommendations:
- The `myhostname` module is placed before the `resolve` (optional) and `dns`
entries, but after `file` (to allow overriding via `/etc/hosts` /
`networking.extraHosts`, and prevent ISPs with catchall-DNS resolvers from
hijacking `.localhost` domains)
- The `mymachines` module, which provides hostname resolution for local
containers (registered with `systemd-machined`) is placed to the front, to
make sure its mappings are preferred over other resolvers.
- If systemd-networkd is enabled, the `resolve` module is placed before
`files` and `myhostname`, as it provides the same logic internally, with
caching.
- The `mdns(_minimal)` module has been updated to the new priorities.
If you use your own NSS host modules, make sure to update your priorities
according to these rules:
- NSS modules which should be queried before `resolved` DNS resolution should
use mkBefore.
- NSS modules which should be queried after `resolved`, `files` and
`myhostname`, but before `dns` should use the default priority
- NSS modules which should come after `dns` should use mkAfter.
binfmt activation script creates /run/binfmt before mounting /run
when system activation.
To fix it I added dependency to specialfs to binfmt activation
script.
os.readlink only resolves one layer of symlinks. This script explicitly relies on the real path of a file for deduplication, hence symlink resolution should recurse.
Previously the code took the kernelPatches of the final derivation, which
might or might not be what was passed to the derivation in the original call.
The previous behaviour caused various hacks to become neccessary to avoid duplicates in kernelPatches.
`systemd.network.networks.*.dhcpServerConfig` did not accept all of
the options which are valid for networkd's [DHCPServer] section. See
systemd.network(5) of systemd 247 for details.
Since 03eaa48 added perl.withPackages, there is a canonical way to
create a perl interpreter from a list of libraries, for use in script
shebangs or generic build inputs. This method is declarative (what we
are doing is clear), produces short shebangs[1] and needs not to wrap
existing scripts.
Unfortunately there are a few exceptions that I've found:
1. Scripts that are calling perl with the -T switch. This makes perl
ignore PERL5LIB, which is what perl.withPackages is using to inform
the interpreter of the library paths.
2. Perl packages that depends on libraries in their own path. This
is not possible because perl.withPackages works at build time. The
workaround is to add `-I $out/${perl.libPrefix}` to the shebang.
In all other cases I propose to switch to perl.withPackages.
[1]: https://lwn.net/Articles/779997/
Adding template overrides allows for custom behavior for specific
instances of a template. Previously, it was not possible to provide
bind mounts for systemd-nspawn. This change allows it.
Currently, kernel config options whose value is "yes" always override
options whose value is "no".
This is not always desired.
Generally speaking, if someone defines an option to have the value
"no", presumably they are disabling the option for a reason, so it's
not always OK to silently enable it due to another, probably unrelated
reason.
For example, a user may want to reduce the kernel attack surface and
therefore may want to disable features that are being enabled in
common-config.nix.
In fact, common-config.nix was already silently enabling options that
were intended to be disabled in hardened/config.nix for security
reasons, such as INET_DIAG.
By eliminating the custom merge function, these config options will
now use the default module option merge functions which make sure
that all options with the highest priority have the same value.
A user that wishes to override an option defined in common-config.nix
can currently use mkForce or mkOverride to do so, e.g.:
BINFMT_MISC = mkForce (option no);
That said, this is not going to be necessary in the future, because
the plan is for kernel config options defined in nixpkgs to use a
lower priority by default, like it currently happens for other module
options.
Catch and ignore errors during writing of the boot entries. These
errors could stem from profile names that are not valid filenames on
vfat filesystems.
fixes#114552
The BGRT theme is probably a close as to "FlickerFree" we can
get without https://github.com/NixOS/nixpkgs/pull/74842.
It's more agnostic than the Breeze theme.
We also install all of themes provided by the packages, as it's possible
that one theme needs the ImageDir of another, and they're small files
anyways.
Lastly, how plymouth handles logo and header files is
a total mess, so hopefully when they have an actual release
we won't need to do all this symlinking.
This allows Plymouth to show the “NixOS 21.03” label under the logo at
startup like it already does at shutdown.
Fixes#59992.
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
Renaming an interface must be done in stage-1: otherwise udev will
report the interface as ready and network daemons (networkd, dhcpcd,
etc.) will bring it up. Once up the interface can't be changed and the
renaming will fail.
Note: link files are read directly by udev, so they can be used even
without networkd enabled.
It was introduced in c10fe14 but removed in c4f910f.
It remained such that people with older generations in their boot
entries could still boot those. Given that the parameter hasn't had any
use in quite some years, it seems safe to remove now.
Fixes#60184
`unitOption` is only used inside of `attrsOf` wich is perfectly capable of
handling the attrsets from `mkIf`, though the checkUnitConfig test
forbids it. This commit weakens that restriction to allow the usage of
`mkIf` inside of `systemd.services.<name>.serviceConfig.<something>`
etc.
networkd's [IPv6PrefixDelegation] section and IPv6PrefixDelegation=
options have been renamed as [IPv6SendRA] and IPv6SendRA= in systemd
247.
Throws if the deprecated option ipv6PrefixDelegationConfig is used.
The `platform` field is pointless nesting: it's just stuff that happens
to be defined together, and that should be an implementation detail.
This instead makes `linux-kernel` and `gcc` top level fields in platform
configs. They join `rustc` there [all are optional], which was put there
and not in `platform` in anticipation of a change like this.
`linux-kernel.arch` in particular also becomes `linuxArch`, to match the
other `*Arch`es.
The next step after is this to combine the *specific* machines from
`lib.systems.platforms` with `lib.systems.examples`, keeping just the
"multiplatform" ones for defaulting.
Declaring them as lists enables the concatenation, supporting
lib.mkBefore, lib.mkOrder, etc.
This is useful when you need to extend a service with a pre-start
script that needs to run as root.
Originally, changes to the kernel don't propagate to the other
derivation within the same package set. This commit allows for the
changes in the kernel to be propagated.
A distinct example is setting `boot.kernel.randstructSeed` to a non-zero
length string which would result in building 2 kernels, one with the
correct seed and the other with the zero length seed. Then, when using
an out-of-tree kernel driver, it would be built with the zero length
seed which differs from the non-zero length seed used to boot,
contradicting the purpose of the `boot.kernel.randstructSeed`.
For in NixOS it is beneficial if both plasma5 and pam use the same Qt5
version. Because the plasma5 desktop may use a different version as the
default Qt5 version, we introduce plasma5Packages.
https://github.com/NixOS/nixpkgs/pull/107497 broke booting on many systems that
use tmpOnTmpfs due to the lack of specifying the mount type.
This commit explicitly adds the mount type, which should fix booting
such systems.
The original change may want to be revisited however too.
@poettering decided we only need a limited number of inodes in our /tmp,
so why not limit that for every systemd user? That makes medium-sized nix
builds impossible so this commit restores the old behaviour which is the
kernel default of half the number of physical RAM pages which does not
seem too unreasonable to me.
See https://www.redhat.com/sysadmin/fedora-31-control-group-v2 for
details on why this is desirable, and how it impacts containers.
Users that need to keep using the old cgroup hierarchy can re-enable it
by setting `systemd.unifiedCgroupHierarchy` to `false`.
Well-known candidates not supporting that hierarchy, like docker and
hidepid=… will disable it automatically.
Fixes#73800
This used to be done by udev, but that was removed in
systemd/systemd@6b2229c. The links are created by systemd at the end of
stage-2, but activation scripts might need them earlier.
These were broken since 2016:
f0367da7d1
since StartLimitIntervalSec got moved into [Unit] from [Service].
StartLimitBurst has also been moved accordingly, so let's fix that one
too.
NixOS systems have been producing logs such as:
/nix/store/wf98r55aszi1bkmln1lvdbp7znsfr70i-unit-caddy.service/caddy.service:31:
Unknown key name 'StartLimitIntervalSec' in section 'Service', ignoring.
I have also removed some unnecessary duplication in units disabling
rate limiting since setting either interval or burst to zero disables it
(ad16158c10/src/basic/ratelimit.c (L16))
When the stage-1 logs get imported in to the journal, they all get
loaded with the same timestamp. This makes it difficult to identify
what might be taking a long time in early boot.
It looks like the test sshd key can never be used, because of too open
permissions. My guess is that the current test script works fine once
the user defined ssh-key has been copied into initrd.
At "nixos-install" however, the user specified host key is not present
in initrd yet and validation fails.
fixes#91486
Otherwise, stage-2-init.sh will complain about not having access to
/dev/fd/62 as of systemd v246.
On IRC, flokli said:
15:14 <flokli> cole-h: hmmm... I could imagine some of the setup inside /dev has been moved into other parts of systemd
15:14 <flokli> And given we run systemd much later (outside initramfs only) it doesn't work properly here
15:17 <flokli> We probably don't invoke udev correctly
This is a temporary fix for #97433. A more proper fix has been
implemented upstream in systemd/systemd#17001, however until it gets
backported, we are stuck with ignoring the error.
After the backport lands, this commit should be reverted.
For UEFI setups, "device" will generally be the special value "nodev"
which represents not running grub-install at all. Using "nodev" for
boot mirrors should therefore be allowed.
The commit enforces buildPackages in the builder but neglects
the fact that the builder is intended to run on the target system.
Because of that, the builder will fail when remotely building a
configuration eg. with nixops or nix-copy-closure.
This reverts commit a6ac6d00f9.
The cyclic dependency of systemd → cryptsetup → lvm2 → udev=systemd
needs to be broken somewhere. The previous strategy of building
cryptsetup with an lvm2 built without udev (#66856) caused the
installer.luksroot test to fail. Instead, build lvm2 with a udev built
without cryptsetup.
Fixes#96479.
Signed-off-by: Anders Kaseorg <andersk@mit.edu>
This allows the user to configure systemd tmpfiles.d via
`environment.etc."tmpfiles.d/X.conf".text = "..."`, which after #93073
causes permission denied (with new X.conf):
```
ln: failed to create symbolic link '/nix/store/...-etc/etc/tmpfiles.d/X.conf': Permission denied
builder for '/nix/store/...-etc.drv' failed with exit code 1
```
or collision between environment.etc and systemd-default-tmpfiles
packages (with existing X.conf, such as tmp.conf):
```
duplicate entry tmpfiles.d/tmp.conf -> /nix/store/...-etc-tmp.conf
mismatched duplicate entry /nix/store/...-systemd-246/example/tmpfiles.d/tmp.conf <-> /nix/store/...-etc-tmp.conf
builder for '/nix/store/...-etc.drv' failed with exit code 1
```
Fixes#96755
declare -a is not sufficient to make the array variable actually
exist, which resulted in the script failing when the target object did
not have any DT_NEEDED entries. This in turn resulted in some
initramfs libraries not having their rpaths patched to point to
extra-utils, which in turn broke the extra-utils tests.
symlinkJoin can break (silently) when the passed paths contain symlinks
to directories. This should work now.
Down-side: when lib/tmpfiles.d doesn't exist for some passed package,
the error message is a little less explicit, because we never get
to the postBuild phase (and symlinkJoin doesn't provide a better way):
/nix/store/HASH-NAME/lib/tmpfiles.d: No such file or directory
Also, it seemed pointless to create symlinks for whole package trees
and using only a part of the result (usually very small part).
These are now only installed by systemd if HAVE_SYSV_COMPAT is true,
which only is the case if you set sysvinit-path and sysvrcnd-path (which
we explicitly unset in the systemd derivation for quite some time)
From the systemd release notes:
nss-mymachines lost support for resolution of users and groups, and now
only does resolution of hostnames. This functionality is now provided by
nss-systemd. Thus, the 'mymachines' entry should be removed from the
'passwd:' and 'group:' lines in /etc/nsswitch.conf (and 'systemd' added
if it is not already there).
Since systemd 246, these are only installed by systemd if
HAVE_SYSV_COMPAT is true, which only is the case if you set
sysvinit-path and sysvrcnd-path (which we explicitly unset in the
systemd derivation for quite some time)
There's a circular dependency to systemd via cryptsetup and lvm2
(systemd -> cryptsetup -> lvm2 -> udev=systemd).
However, cryptsetup only really needs the devmapper component shipped
with lvm2. So build `pkgs.cryptsetup` with a lvm2 that doesn't come with
udev.
- Give a more accurate description of how fileSystems.<name/>.neededForBoot
works
- Give a more detailed description of how fileSystems.<name/>.encrypted.keyFile
works
The toplevel derivations of systems that have `networking.hostName`
set to `""` (because they want their hostname to be set by DHCP) used
to be all named
`nixos-system-unnamed-${config.system.nixos.label}`.
This makes them hard to distinguish.
A similar problem existed in NixOS tests where `vmName` is used in the
`testScript` to refer to the VM. It defaulted to the
`networking.hostName` which when set to `""` won't allow you to refer
to the machine from the `testScript`.
This commit makes the `system.name` configurable. It still defaults to:
```
if config.networking.hostName == ""
then "unnamed"
else config.networking.hostName;
```
but in case `networking.hostName` needs to be to `""` the
`system.name` can be set to a distinguishable name.
This lets users do sneaky things before systemd starts, and
permanently affect the environment in which systemd runs. For example,
we could start systemd in a non-default network namespace by setting
the systemdExecutable to a wrapper script containing:
#!/bin/sh
ip netns add virtual
touch /var/run/netns/physical
mount -o bind /proc/self/ns/net /var/run/netns/physical
exec ip netns exec virtual systemd
_note: the above example does literally work, but there are unresolved
problems with udev and dhcp._
Fixes error
Can't use an undefined value as an ARRAY reference at /nix/store/...-install-grub.pl line 642, <FILE> line 5.
with `/boot/grub/state` being:
```
grub
2.04
no
/dev/sda
/boot
```
I am not sure where the trailing empty line can come from; the script does not
seem to write it. In any case, now we handle that situation as well.
Further, ensure that `extraGrubInstallArgs` defaults to the empty array
if its key is not present in the `jsonState`.
For example, turns the error
cannot copy /nix/store/g24xsmmsz46hzi6whv7qwwn17myn3jfq-grub-2.04/share/grub/unicode.pf2 to /boot
into the more useful
cannot copy /nix/store/g24xsmmsz46hzi6whv7qwwn17myn3jfq-grub-2.04/share/grub/unicode.pf2 to /boot: Read-only file system
Useful for when you need to build grub modules into your grub kernel
to get a working boot, as shown in the added example.
To store this new value, we switch to more structural JSON approach.
Using one line per value to store in `/boot/grub/state` gets really messy when
the values are arrays, or even worse, can contain newlines (escaping would be
needed). Further, removing a value from the file would get extra messy
(empty lines we'd have to keep for backwards compatibility).
Thus, from now on we use JSON to store all values we'll need in the future.
nixos/tests/initrd-openvpn: Add test for openvpn in the initramfs
The module in this commit adds new options that allows the
integration of an OpenVPN client into the initrd.
This can be used e.g. to remotely unlock LUKS devices.
This commit also adds two tests for `boot.initrd.network.openvpn`.
The first one is a basic test to validate that a failing connection
does not prevent the machine from booting.
The second test validates that this module actually creates a valid
openvpn connection.
For this, it spawns three nodes:
- The client that uses boot.initrd.network.openvpn
- An OpenVPN server that acts as gateway and forwards a port
to the client
- A node that is external to the OpenVPN network
The client connects to the OpenVPN server and spawns a netcat instance
that echos a value to every client.
Afterwards, the external node checks if it receives this value over the
forwarded port on the OpenVPN gateway.
This can be used to explicitly specify a specific dtb file, relative to
the dtb base.
Update the generic-extlinux-compatible module to make use of this option.
Some bootloaders might not properly detect the model.
If the specific model is known by configuration, provide a way to
explicitly point to a specific dtb in the extlinux.conf.
This option exposes the builder command used to populate an image,
honoring all options except the -c <path-to-default-configuration>
argument.
Useful to have for sdImage.populateRootCommands.
Special care needs to be taken w.r.t cross - the populate command runs
on the host platform, the activation script on the build platform (so
the builders differ)
Turns out, #75510 was too restrictive.
We also need to allow str here, as some modules set this to
"/run/wrappers" to bring `/run/wrappers/bin` into $PATH of a unit.
Upstream has this alias too; so that dbus activation works.
What I don't fully understand is why this would ever be useful given
this unit is already started way in early boot; even before dbus is up.
But lets just keep behaviour similar to upstream and then ask these
questions to upstream.
With this systemd buffers netlink messages in early boot from the kernel
itself; and passes them on to networkd for processing once it's started.
Makes sure no routing messages are missed.
Also makes an alias so that dbus can activate this unit. Upstream has
this too.
This will make dbus socket activation for it work
When `systemd-resolved` is restarted; this would lead to unavailability
of DNS lookups. You're supposed to use DBUS socket activation to buffer
resolved requests; such that restarts happen without downtime
The way this ends up being called with the raspberry pi 4 image builder
ends up not using the `-e` from the shebang.
In turn, the builds fails during cross-compilation. The wrong coreutils
ends up being used, but this is not made apparent.
The issue I faced is already fixed on master, but this ensures no one
ends up with a failed build "succeeding".
Also, remove the dangling systemd.services.systemd-binfmt.wants = [
"proc-sys-fs-binfmt_misc.mount" ]; in systemd.nix.
If boot.binfmt.registrations != {}, systemd will install
proc-sys-fs-binfmt_misc.automount, which will auto-mount
`/proc/sys/fs/binfmt_misc` as soon as systemd-binfmt tries to access it.
Fixes https://github.com/NixOS/nixpkgs/issues/87687
Fixes https://github.com/NixOS/nixops/issues/574
A disabled nscd breaks nss module loading on NixOS, and systemd without
its nss modules doesn't really work either - instead of silently
disabling its nss modules if nscd is disabled, let the assertion in
nsswitch handle this.
nixos/modules/config/nsswitch.nix uses `passwdArray` for both `passwd`
and `group`, but when moving this into the systemd module in
c0995d22ee, it didn't get split
appropriately.
This will make dbus socket activation for it work
When `systemd-resolved` is restarted; this would lead to unavailability
of DNS lookups. You're supposed to use DBUS socket activation to buffer
resolved requests; such that restarts happen without downtime
This reverts commit 764c8203b8.
While this is desireable in principle, some of our modules and services
fail during service startup if no network is available don't currently
properly set Wants=network-online.target.
If nothing pulls in this target anymore, systemd won't try to reach it.
We have many VM tests waiting for `network-online.target`, and after
764c8203b8 fail with the following error
message:
```
error: unit "network-online.target" is inactive and there are no pending jobs
```
Most likely, test scripts shouldn't wait for `network-online.target` in
first place (as `network-online.target` says nothing about whether a
service has been started), but instead, the script should wait for the
network ports of the corresponding service to be open.
Let's revert this for now, and re-apply in a draft PR, fixing the tests
before merging it back in.
This follows upstreams change in documentation. While the `[DHCP]`
section might still work it is undocumented and we should probably not
be using it anymore. Users can just upgrade to the new option without
much hassle.
I had to create a bit of custom module deprecation code since the usual
approach doesn't support wildcards in the path.
You can now specify option for the `[DHCPv6]` section with
`systemd.network.<name>.dhcpV6Config.…`. Previously you could only use
the combined legacy DHCP configuration.
Systemd upstream has deprecated CriticalConnection with v244 in favor of
KeepConnection as that seems to be more flexible:
The CriticalConnection= setting in .network files is now deprecated,
and replaced by a new KeepConfiguration= setting which allows more
detailed configuration of the IP configuration to keep in place.
Not all systems need to be online to boot up. So, don’t pull
network-online.target into multi-user.target. Services that need
online network can still require it.
This increases my boot time from ~9s to ~5s.
1d61efb7f1 accidentially changed the
restartTriggers of systemd-networkd.service` to point to the attribute
name (in this case, a location relative to `/etc`), instead of the
location of the network-related unit files in the nix store.
This caused systemd-networkd to not get restarted on activation of new
networking config, if the file name hasn't changed.
Fix this, by pointing this back to the location in the nix store.
Add a distinctive `unit-script` prefix to systemd unit scripts to make
them easier to find in the store directory. Do not add this prefix to
actual script file name as it clutters logs.
Current journal output from services started by `script` rather than
`ExexStart` is unreadable because the name of the file (which journalctl
records and outputs) quite literally takes 1/3 of the screen (on smaller
screens).
Make it shorter. In particular:
* Drop the `unit-script` prefix as it is not very useful.
* Use `writeShellScriptBin` to write them because:
* It has a `checkPhase` which is better than no checkPhase.
* The script itself ends up having a short name.
systemd-tmpfiles will load all files in lexicographic order and ignores rules
for the same path in later files with a warning Since we apply the default rules
provided by systemd, we should load user-defines rules first so users have a
chance to override defaults.
Commit 1d2c3279311e4f03fcf164e1366f2fda9f4bfccf in the upstream kernel
repository removed support for the scalar x86_64 and i586 AES
assembly implementations, since the generic AES implementation generated
by the compiler is faster for both platforms. Remove the modules from
the cryptoModules list. This causes a regression for kernel versions
>=5.4 which include the removal. This should have no negative impact on
AES performance on older kernels since the generic implementation should
be faster there as well since the implementation was hardly touched from
its initial submission.
Fixes#84842
In contrast to `.service`-units, it's not possible to declare an
`overrides.conf`, however this is done by `generateUnits` for `.nspawn`
units as well. This change breaks the build if you have two derivations
configuring one nspawn unit.
This will happen in a case like this:
``` nix
{ pkgs, ... }: {
systemd.packages = [
(pkgs.writeTextDir "etc/systemd/nspawn/container0.nspawn" ''
[Files]
Bind=/tmp
'')
];
systemd.nspawn.container0 = {
/* ... */
};
}
```
This option allows replacing the tmpfs mounted on / by
the live CD's init script with a physical device
Since nixOS symlinks everything there's no trouble
at all.
That enables the user to easily use a nixOS live CD
as a portable installation.
Note that due to some limitations in how the store is mounted
currently only the non-store things are persisted.
Dropbear lags behind OpenSSH significantly in both support for modern
key formats like `ssh-ed25519`, let alone the recently-introduced
U2F/FIDO2-based `sk-ssh-ed25519@openssh.com` (as I found when I switched
my `authorizedKeys` over to it and promptly locked myself out of my
server's initrd SSH, breaking reboots), as well as security features
like multiprocess isolation. Using the same SSH daemon for stage-1 and
the main system ensures key formats will always remain compatible, as
well as more conveniently allowing the sharing of configuration and
host keys.
The main reason to use Dropbear over OpenSSH would be initrd space
concerns, but NixOS initrds are already large (17 MiB currently on my
server), and the size difference between the two isn't huge (the test's
initrd goes from 9.7 MiB to 12 MiB with this change). If the size is
still a problem, then it would be easy to shrink sshd down to a few
hundred kilobytes by using an initrd-specific build that uses musl and
disables things like Kerberos support.
This passes the test and works on my server, but more rigorous testing
and review from people who use initrd SSH would be appreciated!
`$toplevel/system` of a system closure with `x86_64` kernel and `i686` userland should contain "x86_64-linux".
If `$toplevel/system` contains "i686-linux", the closure will be run using `qemu-system-i386`, which is able to run `x86_64` kernel on most Intel CPU, but fails on AMD.
So this fix is for a rare case of `x86_64` kernel + `i686` userland + AMD CPU
This mirrors the behaviour of systemd - It's udev that parses `.link`
files, not `systemd-networkd`.
This was originally applied in 36ef112a47,
but was reverted due to 1115959a8d causing
evaluation errors on hydra.
...even when networkd is disabled
This reverts commit ce78f3ac70, reversing
changes made to dc34da0755.
I'm sorry; Hydra has been unable to evaluate, always returning
> error: unexpected EOF reading a line
and I've been unable to reproduce the problem locally. Bisecting
pointed to this merge, but I still can't see what exactly was wrong.
This is to facilitate units that should _only_ be manually started and
not activated when a configuration is switched to.
More specifically this is to be used by the new Nixops deploy-*
targets created in https://github.com/NixOS/nixops/pull/1245 that are
triggered by Nixops before/after switch-to-configuration is called.
