]> git.ipfire.org Git - thirdparty/systemd.git/log
thirdparty/systemd.git
4 years agoudevadm: introduce -a|--action option for test-builtin command 20460/head
Yu Watanabe [Tue, 17 Aug 2021 14:14:29 +0000 (23:14 +0900)] 
udevadm: introduce -a|--action option for test-builtin command

As net_setup_link builtin requires that a device action is set for the
sd_device object.

4 years agoshell-completion: add missing uevent actions for udevadm
Yu Watanabe [Tue, 17 Aug 2021 14:13:07 +0000 (23:13 +0900)] 
shell-completion: add missing uevent actions for udevadm

4 years agoudevadm: introduce parse_device_action() helper function
Yu Watanabe [Tue, 17 Aug 2021 13:57:04 +0000 (22:57 +0900)] 
udevadm: introduce parse_device_action() helper function

4 years agoudevadm: introduce find_device_with_action() helper function
Yu Watanabe [Tue, 17 Aug 2021 13:46:32 +0000 (22:46 +0900)] 
udevadm: introduce find_device_with_action() helper function

4 years agoMerge pull request #20281 from bluca/ext_release_naming
Lennart Poettering [Tue, 17 Aug 2021 14:06:26 +0000 (16:06 +0200)] 
Merge pull request #20281 from bluca/ext_release_naming

extension-release: allow fallback when image name is mangled after build, improve docs

4 years agodocs: document layered images in PORTABLE_SERVICES.md 20281/head
Luca Boccassi [Fri, 23 Jul 2021 15:07:05 +0000 (16:07 +0100)] 
docs: document layered images in PORTABLE_SERVICES.md

4 years agoman: further document extension-release
Luca Boccassi [Fri, 23 Jul 2021 13:45:38 +0000 (14:45 +0100)] 
man: further document extension-release

4 years agoextension-release: search for other files if expected name not found
Luca Boccassi [Thu, 22 Jul 2021 19:41:34 +0000 (20:41 +0100)] 
extension-release: search for other files if expected name not found

In some cases image names are unpredictable - some orchestrators/deployment
tools like to mangle names to suit their internal formats. In these cases,
the requirement that the extension-release file matches exactly the image
name where it's contained cannot work.

Allow falling back to loading the first regular file which name starts with
'extension-release' located in /usr/lib/extension-release.d/ and tagged with
a user.extension-release.strict extended attribute with a true value, if the
one with the expected name cannot be found.

4 years agoxattr-util: add fgetxattrat_fake_malloc variant
Luca Boccassi [Mon, 16 Aug 2021 14:58:09 +0000 (15:58 +0100)] 
xattr-util: add fgetxattrat_fake_malloc variant

4 years agoenv-util: add unsetenv_erase() helper
Lennart Poettering [Mon, 16 Aug 2021 14:41:34 +0000 (16:41 +0200)] 
env-util: add unsetenv_erase() helper

Let's unify how we remove secrets from the env block.

4 years agoMerge pull request #20420 from poettering/import-beef-up
Lennart Poettering [Tue, 17 Aug 2021 09:53:18 +0000 (11:53 +0200)] 
Merge pull request #20420 from poettering/import-beef-up

import: modernizations, and various additions

4 years agoimport: drop some now unused functions from import-common.c 20420/head
Lennart Poettering [Fri, 26 Feb 2021 21:49:38 +0000 (22:49 +0100)] 
import: drop some now unused functions from import-common.c

4 years agodocs: document how to turn off btrfs quota support in importd
Lennart Poettering [Wed, 3 Feb 2021 20:50:27 +0000 (21:50 +0100)] 
docs: document how to turn off btrfs quota support in importd

Fixes: #18421 #15903
4 years agoimport-fs: make various options controllable via cmdline/env var
Lennart Poettering [Thu, 25 Feb 2021 12:09:30 +0000 (13:09 +0100)] 
import-fs: make various options controllable via cmdline/env var

This basically does what the previous two commits did for systemd-import
+ systemd-pull but for systemd-import-fs.

This commit is a bit simpler though, as a --direct mode doesn't change
that much. It's mostly about not searching for existing, conflicting
images and not much else.

4 years agopull: add --direct mode + make various eatures optional + explicit checksum verification
Lennart Poettering [Wed, 9 Jun 2021 15:19:15 +0000 (17:19 +0200)] 
pull: add --direct mode + make various eatures optional + explicit checksum verification

This does what the previous commit did for systemd-import the same way
for systemd-pull.

It also adds one more thing: the checksum validation is extended, in
addition of doing SHA256SUMS/gpg verification it is now possible to
immediately specify a hash value on the command line that the download
needs to match. This is particularly useful in --direct mode as we can
download/decompress/unpack arbitrary files and check the hash of the
downloaded file on-the-fly.

4 years agoimport: add new "--direct" mode + add controls for turning certain features on/off
Lennart Poettering [Fri, 22 Jan 2021 16:40:51 +0000 (17:40 +0100)] 
import: add new "--direct" mode + add controls for turning certain features on/off

This reworks/modernizes the tar/raw import logic and adds the following
new features:

- Adds the ability to control btrfs subvol and quota behaviour which was
  previously always on via an env var and cmdline arg

- Adds control whether to sync() stuff after writing it, similar via env
  var + cmdline arg

- Similar, the QCOW2 unpacking logic that was previously the implied
  default may now be controlled via env var + cmdline arg.

- adds a "direct" mode. In this mode, the systemd-import tool can be
  used as a simple tool for decompressing/unpacking/installing arbitrary
  files, without all the additional meta data and auxiliary resources,
  i.e.  outside of the immediate disk image context. Via the new
  --offset= and --size-max= switches the downloaded data can be written
  to specific locations of a file (which is particularly useful to use
  the tool to download fs images and write them to a partition location
  before actually creating the partition).

We'll later use the latter feature for "sysupdate" concept, where images
can be directly be written to partitions. That way the systemd-import
binary will be used as backend for both "systemd-importd" and
"systemd-sysupdate" and share most of the same code.

4 years agoshared: add generic helper tools for installing files/dir trees
Lennart Poettering [Fri, 26 Feb 2021 16:41:30 +0000 (17:41 +0100)] 
shared: add generic helper tools for installing files/dir trees

This adds a bit of generic helper tools for installing files/dir trees.
"installing" is supposed to mean the final step when preparing a disk
image or directory tree, where the result is renamed to its final name.
It has some bells and whistles, as it is able to replace existing files
sanely, can fsync() things carefully and can mark things read-only in a
nice way.

This is supposed to be generic, unified code that can be used eventually
for any of our tools that prepare disk images/directory trees, including
importd, nspawn's --template= mechanism, the discover-image.c logic,
and more.

4 years agopo: Translated using Weblate (Portuguese (Brazil))
Gustavo Costa [Tue, 17 Aug 2021 07:04:48 +0000 (09:04 +0200)] 
po: Translated using Weblate (Portuguese (Brazil))

Currently translated at 100.0% (189 of 189 strings)

Co-authored-by: Gustavo Costa <xfgusta@gmail.com>
Translate-URL: https://translate.fedoraproject.org/projects/systemd/master/pt_BR/
Translation: systemd/main

4 years agoMerge pull request #20450 from yuwata/ethtool-cleanups
Yu Watanabe [Mon, 16 Aug 2021 20:22:40 +0000 (05:22 +0900)] 
Merge pull request #20450 from yuwata/ethtool-cleanups

ethtool: trivial cleanups

4 years agoMerge pull request #20443 from yuwata/network-conf-parser-cleanups
Yu Watanabe [Mon, 16 Aug 2021 17:42:27 +0000 (02:42 +0900)] 
Merge pull request #20443 from yuwata/network-conf-parser-cleanups

network: conf parser cleanups

4 years agounit: coldplug both job and nop_job if possible
Geass-LL [Fri, 2 Apr 2021 03:27:59 +0000 (11:27 +0800)] 
unit: coldplug both job and nop_job if possible

Sometimes, both job and nop_job are deserialized. In this case,
if we only cold plug the job, the nop_job will also stuck in the
job list.

4 years agoMerge pull request #20442 from yuwata/network-can-introduce-many-settings
Yu Watanabe [Mon, 16 Aug 2021 17:40:32 +0000 (02:40 +0900)] 
Merge pull request #20442 from yuwata/network-can-introduce-many-settings

network: introduce several CAN interface related settings

4 years agonetwork: add address label on dhcpv4
Maxime de Roucy [Sat, 14 Aug 2021 20:50:14 +0000 (22:50 +0200)] 
network: add address label on dhcpv4

Fixes: #13967
4 years agonetwork: allow users to forbid passthru MACVLAN from putting its link into promiscuou...
Tom Yan [Mon, 16 Aug 2021 10:00:42 +0000 (18:00 +0800)] 
network: allow users to forbid passthru MACVLAN from putting its link into promiscuous mode

While we haven't implemented a key for users to set MACVLAN/MACVTAP flags,
we can at least allow them to make use of the Promiscuous= key of
the corresponding link to set the nopromisc flag.

4 years agoethtool: make ethtool_set_features() return earlier when nothing is requested 20450/head
Yu Watanabe [Mon, 16 Aug 2021 15:44:00 +0000 (00:44 +0900)] 
ethtool: make ethtool_set_features() return earlier when nothing is requested

4 years agoethtool: make the size of 'features' array static
Yu Watanabe [Mon, 16 Aug 2021 14:47:40 +0000 (23:47 +0900)] 
ethtool: make the size of 'features' array static

4 years agoman: address label can be set only for IPv4 addresses 20443/head
Yu Watanabe [Mon, 16 Aug 2021 03:57:18 +0000 (12:57 +0900)] 
man: address label can be set only for IPv4 addresses

4 years agoMerge pull request #20438 from medhefgo/boot
Lennart Poettering [Mon, 16 Aug 2021 13:35:08 +0000 (15:35 +0200)] 
Merge pull request #20438 from medhefgo/boot

sd-boot: Better self-detection and windows loader title

4 years agonetwork: can: allow to specify bit-timing with TimeQuantaNSec= and friends 20442/head
Yu Watanabe [Sun, 15 Aug 2021 18:36:58 +0000 (03:36 +0900)] 
network: can: allow to specify bit-timing with TimeQuantaNSec= and friends

Closes #19424 and #20435.

4 years agonetwork: SamplePoint= should be specified only when BitRate= is specified
Yu Watanabe [Sun, 15 Aug 2021 17:41:15 +0000 (02:41 +0900)] 
network: SamplePoint= should be specified only when BitRate= is specified

See can_get_bittiming() in drivers/net/can/dev/bittiming.c of kernel.

4 years agonetwork: can: make Termination= optionally take a raw resistor value
Yu Watanabe [Sun, 15 Aug 2021 17:16:53 +0000 (02:16 +0900)] 
network: can: make Termination= optionally take a raw resistor value

Note that this slightly breaks backward compatibility when
Termination=1. Previously, this is handled as boolean true, then 120 ohm
was used. But now with this commit, it is handled as 1 ohm.

4 years agonetwork: can: add missing control modes
Yu Watanabe [Sun, 15 Aug 2021 16:53:26 +0000 (01:53 +0900)] 
network: can: add missing control modes

4 years agonetwork: can: introduce config_parse_can_control_mode()
Yu Watanabe [Sun, 15 Aug 2021 16:03:25 +0000 (01:03 +0900)] 
network: can: introduce config_parse_can_control_mode()

4 years agonetwork: can: refuse too large restart sec earlier
Yu Watanabe [Sun, 15 Aug 2021 15:45:23 +0000 (00:45 +0900)] 
network: can: refuse too large restart sec earlier

4 years agosd-boot: Detect windows boot loader title from BCD 20438/head
Jan Janssen [Fri, 13 Aug 2021 17:03:35 +0000 (19:03 +0200)] 
sd-boot: Detect windows boot loader title from BCD

4 years agosd-boot: Add memmem_safe and memory_startswith
Jan Janssen [Fri, 13 Aug 2021 17:03:35 +0000 (19:03 +0200)] 
sd-boot: Add memmem_safe and memory_startswith

4 years agosd-boot: Try harder to detect ourselves
Jan Janssen [Fri, 13 Aug 2021 17:03:35 +0000 (19:03 +0200)] 
sd-boot: Try harder to detect ourselves

By moving our magic string into its own PE section, we can forego
grepping for it.

4 years agosd-boot: Fix PE section parsing
Jan Janssen [Fri, 13 Aug 2021 16:55:32 +0000 (18:55 +0200)] 
sd-boot: Fix PE section parsing

We only need the PE header offset from the DOS header, not
its size. Previously, the section table could be cut off in the middle.

While we are at it, also modernize the remaining code.

4 years agonetwork: check validity before copying the input string
Yu Watanabe [Mon, 16 Aug 2021 03:48:04 +0000 (12:48 +0900)] 
network: check validity before copying the input string

Prompted by https://github.com/systemd/systemd/pull/20440#discussion_r689136337.

4 years agonetwork: rebreak conf parser arguments
Yu Watanabe [Mon, 16 Aug 2021 03:41:34 +0000 (12:41 +0900)] 
network: rebreak conf parser arguments

4 years agonetwork: can: move function
Yu Watanabe [Sun, 15 Aug 2021 15:32:10 +0000 (00:32 +0900)] 
network: can: move function

4 years agobasic/linux: update linux uapi headers
Yu Watanabe [Sun, 15 Aug 2021 16:06:11 +0000 (01:06 +0900)] 
basic/linux: update linux uapi headers

4 years agoDon't open /var journals in volatile mode when runtime_journal==NULL
Milo Turner [Fri, 13 Aug 2021 14:28:58 +0000 (10:28 -0400)] 
Don't open /var journals in volatile mode when runtime_journal==NULL

4 years agoMerge pull request #20432 from yuwata/network-recreate-stacked-netdevs
Yu Watanabe [Fri, 13 Aug 2021 08:31:28 +0000 (17:31 +0900)] 
Merge pull request #20432 from yuwata/network-recreate-stacked-netdevs

network: recreate stacked netdevs when underlying device is re-added

4 years agomeson: Make unused-function an error
Jan Janssen [Thu, 12 Aug 2021 14:18:10 +0000 (16:18 +0200)] 
meson: Make unused-function an error

4 years agoresolved: retry on SERVFAIL before downgrading feature level
Steven Siloti [Sun, 11 Jul 2021 18:05:26 +0000 (11:05 -0700)] 
resolved: retry on SERVFAIL before downgrading feature level

The SERVFAIL RCODE can be generated for many reasons which may not be related
to lack of feature support. For example, the Stubby resolver generates
SERVFAIL when a request times out. Such transient failures can cause
unnecessary downgrades to both the transaction and the server's feature level.
The consequences of this are especially severe if the server is in DNSSEC
strict mode. In this case repeated downgrades eventually cause the server to
stop resolving entirely with the error "incompatible-server".

To avoid unnecessary downgrades the request should be retried once with the
current level before the transaction's feature level is downgraded.

4 years agoMerge pull request #20233 from maanyagoenka/log-error
Lennart Poettering [Fri, 13 Aug 2021 07:22:48 +0000 (09:22 +0200)] 
Merge pull request #20233 from maanyagoenka/log-error

systemd-analyze: add option to return an error value when unit verification fails

4 years agoMerge pull request #20350 from medhefgo/boot
Lennart Poettering [Fri, 13 Aug 2021 07:22:12 +0000 (09:22 +0200)] 
Merge pull request #20350 from medhefgo/boot

Grab bag of sd-boot improvements

4 years agosystemd-analyze: option to exit with an error when 'verify' fails 20233/head
Maanya Goenka [Mon, 26 Jul 2021 20:02:17 +0000 (13:02 -0700)] 
systemd-analyze: option to exit with an error when 'verify' fails

The commit introduces a callback invoked from log_syntax_internal.
Use it from systemd-analyze to gather a list of units that contain
syntax warnings. A new command line option is added to make use of this.

The new option --recursive-errors takes in three possible modes:

1. yes - which is the default. systemd-analyze exits with an error when syntax warnings arise during verification of the
 specified units or any of their dependencies.
3. no - systemd-analyze exits with an error when syntax warnings arise during verification of only the selected unit.
Analyzing and loading any dependencies will be skipped.
4. one - systemd-analyze exits with an error when syntax warnings arise during verification
 of only the selected units and their direct dependencies.

Below are two service unit files that I created for the purposes of testing:

1. First, we run the commands on a unit that does not have dependencies but has a non-existing key-value setting (i.e. foo = bar).

> cat <<EOF>testcase.service

[Unit]
foo = bar

[Service]
ExecStart = echo hello
EOF

OUTPUT:

maanya-goenka@debian:~/systemd (log-error)$ sudo build/systemd-analyze verify testcase.service
/home/maanya-goenka/systemd/testcase.service:2: Unknown key name 'foo' in section 'Unit', ignoring.
/usr/lib/systemd/system/plymouth-start.service:15: Unit configured to use KillMode=none. This is unsafe, as it disables systemd's process lifecycle management for the service. Please update your service to use a safer KillMode=, such as 'mixed' or 'control-group'. Support for KillMode=none is deprecated and will eventually be removed.
/usr/lib/systemd/system/dbus.socket:5: ListenStream= references a path below legacy directory /var/run/, updating /var/run/dbus/system_bus_socket → /run/dbus/system_bus_socket; please update the unit file accordingly.
/usr/lib/systemd/system/gdm.service:30: Standard output type syslog is obsolete, automatically updating to journal. Please update your unit file, and consider removing the setting altogether.
maanya-goenka@debian:~/systemd (log-error)$ echo $?
1

maanya-goenka@debian:~/systemd (log-error)$ sudo build/systemd-analyze verify --recursive-errors=yes testcase.service
/home/maanya-goenka/systemd/testcase.service:2: Unknown key name 'foo' in section 'Unit', ignoring.
/usr/lib/systemd/system/plymouth-start.service:15: Unit configured to use KillMode=none. This is unsafe, as it disables systemd's process lifecycle management for the service. Please update your service to use a safer KillMode=, such as 'mixed' or 'control-group'. Support for KillMode=none is deprecated and will eventually be removed.
/usr/lib/systemd/system/dbus.socket:5: ListenStream= references a path below legacy directory /var/run/, updating /var/run/dbus/system_bus_socket → /run/dbus/system_bus_socket; please update the unit file accordingly.
/usr/lib/systemd/system/gdm.service:30: Standard output type syslog is obsolete, automatically updating to journal. Please update your unit file, and consider removing the setting altogether.
maanya-goenka@debian:~/systemd (log-error)$ echo $?
1

maanya-goenka@debian:~/systemd (log-error)$ sudo build/systemd-analyze verify --recursive-errors=no testcase.service
/home/maanya-goenka/systemd/testcase.service:2: Unknown key name 'foo' in section 'Unit', ignoring.
maanya-goenka@debian:~/systemd (log-error)$ echo $?
1

maanya-goenka@debian:~/systemd (log-error)$ sudo build/systemd-analyze verify --recursive-errors=one testcase.service
/home/maanya-goenka/systemd/testcase.service:2: Unknown key name 'foo' in section 'Unit', ignoring.
/usr/lib/systemd/system/plymouth-start.service:15: Unit configured to use KillMode=none. This is unsafe, as it disables systemd's process lifecycle management for the service. Please update your service to use a safer KillMode=, such as 'mixed' or 'control-group'. Support for KillMode=none is deprecated and will eventually be removed.
/usr/lib/systemd/system/dbus.socket:5: ListenStream= references a path below legacy directory /var/run/, updating /var/run/dbus/system_bus_socket → /run/dbus/system_bus_socket; please update the unit file accordingly.
/usr/lib/systemd/system/gdm.service:30: Standard output type syslog is obsolete, automatically updating to journal. Please update your unit file, and consider removing the setting altogether.
maanya-goenka@debian:~/systemd (log-error)$ echo $?
1

2. Next, we run the commands on a unit that is syntactically valid but has a non-existing dependency (i.e. foo2.service)

> cat <<EOF>foobar.service

[Unit]
Requires = foo2.service

[Service]
ExecStart = echo hello
EOF

OUTPUT:

maanya-goenka@debian:~/systemd (log-error)$ sudo build/systemd-analyze verify foobar.service
/usr/lib/systemd/system/plymouth-start.service:15: Unit configured to use KillMode=none. This is unsafe, as it disables systemd's process lifecycle management for the service. Please update your service to use a safer KillMode=, such as 'mixed' or 'control-group'. Support for KillMode=none is deprecated and will eventually be removed.
/usr/lib/systemd/system/dbus.socket:5: ListenStream= references a path below legacy directory /var/run/, updating /var/run/dbus/system_bus_socket → /run/dbus/system_bus_socket; please update the unit file accordingly.
/usr/lib/systemd/system/gdm.service:30: Standard output type syslog is obsolete, automatically updating to journal. Please update your unit file, and consider removing the setting altogether.
foobar.service: Failed to create foobar.service/start: Unit foo2.service not found.
maanya-goenka@debian:~/systemd (log-error)$ echo $?
1

maanya-goenka@debian:~/systemd (log-error)$ sudo build/systemd-analyze verify --recursive-errors=yes foobar.service
/usr/lib/systemd/system/plymouth-start.service:15: Unit configured to use KillMode=none. This is unsafe, as it disables systemd's process lifecycle management for the service. Please update your service to use a safer KillMode=, such as 'mixed' or 'control-group'. Support for KillMode=none is deprecated and will eventually be removed.
/usr/lib/systemd/system/dbus.socket:5: ListenStream= references a path below legacy directory /var/run/, updating /var/run/dbus/system_bus_socket → /run/dbus/system_bus_socket; please update the unit file accordingly.
/usr/lib/systemd/system/gdm.service:30: Standard output type syslog is obsolete, automatically updating to journal. Please update your unit file, and consider removing the setting altogether.
foobar.service: Failed to create foobar.service/start: Unit foo2.service not found.
maanya-goenka@debian:~/systemd (log-error)$ echo $?
1

maanya-goenka@debian:~/systemd (log-error)$ sudo build/systemd-analyze verify --recursive-errors=no foobar.service
maanya-goenka@debian:~/systemd (log-error)$ echo $?
0

maanya-goenka@debian:~/systemd (log-error)$ sudo build/systemd-analyze verify --recursive-errors=one foobar.service
/usr/lib/systemd/system/plymouth-start.service:15: Unit configured to use KillMode=none. This is unsafe, as it disables systemd's process lifecycle management for the service. Please update your service to use a safer KillMode=, such as 'mixed' or 'control-group'. Support for KillMode=none is deprecated and will eventually be removed.
/usr/lib/systemd/system/dbus.socket:5: ListenStream= references a path below legacy directory /var/run/, updating /var/run/dbus/system_bus_socket → /run/dbus/system_bus_socket; please update the unit file accordingly.
/usr/lib/systemd/system/gdm.service:30: Standard output type syslog is obsolete, automatically updating to journal. Please update your unit file, and consider removing the setting altogether.
foobar.service: Failed to create foobar.service/start: Unit foo2.service not found.
maanya-goenka@debian:~/systemd (log-error)$ echo $?
1

4 years agomanager: add a test flag to ignore dependencies
Maanya Goenka [Thu, 8 Jul 2021 01:28:20 +0000 (18:28 -0700)] 
manager: add a test flag to ignore dependencies

The MANAGER_TEST_RUN_IGNORE_DEPENDENCIES flag was added in order to allow the caller
to skip the recursive loading of dependency units when loading specific
unit files. This includes the default dependencies, the specified dependencies, the slice.
This will be used by systemd-analyze to allow checking individual unit files in isolation.

4 years agomanager: use FLAGS_SET when checking for MANAGER_TEST_RUN_MINIMAL
Maanya Goenka [Thu, 29 Jul 2021 20:01:08 +0000 (13:01 -0700)] 
manager: use FLAGS_SET when checking for MANAGER_TEST_RUN_MINIMAL

Allows multiple flags to be set, for example, in systemd-analyze.

4 years agosd-boot: Allow on/off and t/f for booleans too 20350/head
Jan Janssen [Wed, 11 Aug 2021 12:59:46 +0000 (14:59 +0200)] 
sd-boot: Allow on/off and t/f for booleans too

4 years agosd-boot: Provide error messages when parsing a config option fails
Jan Janssen [Wed, 11 Aug 2021 12:59:46 +0000 (14:59 +0200)] 
sd-boot: Provide error messages when parsing a config option fails

4 years agosd-boot: Rework console input handling
Jan Janssen [Wed, 11 Aug 2021 12:59:46 +0000 (14:59 +0200)] 
sd-boot: Rework console input handling

Fixes: #15847
Probably fixes: #19191

4 years agoMerge pull request #20199 from ddstreet/unit_cgroup_catchup
Lennart Poettering [Thu, 12 Aug 2021 14:04:40 +0000 (16:04 +0200)] 
Merge pull request #20199 from ddstreet/unit_cgroup_catchup

cgroup: do 'catchup' for unit cgroup inotify watch files

4 years agosystemd-analyze: parse ip_filters_custom_egress correctly
Maanya Goenka [Tue, 10 Aug 2021 21:30:46 +0000 (14:30 -0700)] 
systemd-analyze: parse ip_filters_custom_egress correctly

Fixed bug in original assignment of security_info variable: ip_filters_custom_egress.

4 years agonetwork: fix configuring of CAN devices
Yu Watanabe [Thu, 12 Aug 2021 04:39:53 +0000 (13:39 +0900)] 
network: fix configuring of CAN devices

Fix a bug introduced by 7558f9e717381eef0ddc8ddfb5a754ea4b0f3e6c.

Fixes #20428.

4 years agosd-boot: Use StrSize where it makes sense
Jan Janssen [Wed, 11 Aug 2021 12:59:46 +0000 (14:59 +0200)] 
sd-boot: Use StrSize where it makes sense

4 years agosd-boot: Assert all the things!
Jan Janssen [Wed, 11 Aug 2021 12:59:46 +0000 (14:59 +0200)] 
sd-boot: Assert all the things!

4 years agosd-boot: Add assert implementation
Jan Janssen [Wed, 11 Aug 2021 12:59:46 +0000 (14:59 +0200)] 
sd-boot: Add assert implementation

There is a ASSERT() macro from gnu-efi, but that does not show any
output to ConOut. Having to do some additional setup just to get
some debug output is tedious and outright difficult on real hardware.

4 years agosd-boot: Fix possible null pointer dereference
Jan Janssen [Wed, 11 Aug 2021 12:59:46 +0000 (14:59 +0200)] 
sd-boot: Fix possible null pointer dereference

Auto entries are showing garbage for the version in print_status()
because StrDuplicate does not expect null pointers.

4 years agosd-boot: Don't use magic integer constants
Jan Janssen [Wed, 11 Aug 2021 12:59:46 +0000 (14:59 +0200)] 
sd-boot: Don't use magic integer constants

4 years agosd-boot: Unify error handling
Jan Janssen [Wed, 11 Aug 2021 12:59:46 +0000 (14:59 +0200)] 
sd-boot: Unify error handling

log_error_stall() and log_error_status_stall() will ensure the user has
a chance to catch an error message by stalling and also forcing a
lightred/black color on it. Also, convert several Print() calls to it
since they are actually error messages.

4 years agoMerge pull request #20418 from yuwata/hostnamed-fix-20417
Lennart Poettering [Thu, 12 Aug 2021 07:34:21 +0000 (09:34 +0200)] 
Merge pull request #20418 from yuwata/hostnamed-fix-20417

hostname: fix off-by-one issue in gethostname()

4 years agotest-network: add a testcase for recreating stacked netdevs 20432/head
Yu Watanabe [Thu, 12 Aug 2021 05:45:40 +0000 (14:45 +0900)] 
test-network: add a testcase for recreating stacked netdevs

4 years agonetwork: recreate stacked netdevs when underlying device is re-added
Yu Watanabe [Thu, 12 Aug 2021 05:31:27 +0000 (14:31 +0900)] 
network: recreate stacked netdevs when underlying device is re-added

Closes #20430.

4 years agonetwork: adjust log messages, function names, etc.
Yu Watanabe [Thu, 12 Aug 2021 05:38:34 +0000 (14:38 +0900)] 
network: adjust log messages, function names, etc.

4 years agonetwork: use netdev_enter_failed() instead of netdev_drop() on error
Yu Watanabe [Thu, 12 Aug 2021 05:23:54 +0000 (14:23 +0900)] 
network: use netdev_enter_failed() instead of netdev_drop() on error

Preparation for later commits to support reconfiguring netdevs.

4 years agonetwork: introduce a helper function netdev_is_stacked_and_independent()
Yu Watanabe [Thu, 12 Aug 2021 05:10:45 +0000 (14:10 +0900)] 
network: introduce a helper function netdev_is_stacked_and_independent()

4 years agohostname: introduce gethostname_full() and use it in various gethostname() variants 20418/head
Yu Watanabe [Tue, 10 Aug 2021 18:44:24 +0000 (03:44 +0900)] 
hostname: introduce gethostname_full() and use it in various gethostname() variants

4 years agohostname: fix off-by-one issue in gethostname()
Yu Watanabe [Tue, 10 Aug 2021 15:12:40 +0000 (00:12 +0900)] 
hostname: fix off-by-one issue in gethostname()

gethostname() returns null-terminated hostname.

Fixes #20309 and #20417.

4 years agoMerge pull request #20423 from yuwata/dhcp-server-static-lease-outside-pool
Luca Boccassi [Wed, 11 Aug 2021 17:40:00 +0000 (18:40 +0100)] 
Merge pull request #20423 from yuwata/dhcp-server-static-lease-outside-pool

sd-dhcp-server: support static lease outside of pool

4 years agoshared/copy: add a new flag COPY_ALL_XATTRS
Andrej Lajovic [Sun, 8 Aug 2021 23:43:54 +0000 (01:43 +0200)] 
shared/copy: add a new flag COPY_ALL_XATTRS

When the flag COPY_ALL_XATTRS is set, it causes the complete set of xattrs
to be copied. If the flag is unset, only xattrs from the "user" namespace
are copied.

Fixes #17178.

4 years agoMerge pull request #20419 from keszybz/setenv-no-value
Lennart Poettering [Wed, 11 Aug 2021 15:47:45 +0000 (17:47 +0200)] 
Merge pull request #20419 from keszybz/setenv-no-value

Allow --setenv=FOO in various programs

4 years agomacro: Move some macros to macro-fundamental.h
Jan Janssen [Wed, 11 Aug 2021 12:59:46 +0000 (14:59 +0200)] 
macro: Move some macros to macro-fundamental.h

Also, make sure STRLEN works with wide strings too.

4 years agoactivate: simplify/rework implementation of --setenv 20419/head
Zbigniew Jędrzejewski-Szmek [Tue, 10 Aug 2021 15:58:17 +0000 (17:58 +0200)] 
activate: simplify/rework implementation of --setenv

Previous implementation is simplified by using the new helper. The new code
does more looping, but considering that it's unlikely that people set more
than a handful of variables through commandline options, this should be OK.

If a variable is specified on the command line, it overrides any automatically
set variable. Effective behaviour was already were like this, because we would
specify two variables, both would be set, and since glibc will return
the first matching entry.
('systemd-socket-activate -E TERM=FOO -l 2000 --inetd -a env' would give
'TERM=FOO TERM=xterm-256color PATH=...', and getenv("TERM") returns "FOO".)
But it's nicer to filter out any duplicate entries and only pass the intended
variable to the child process.

4 years agoMerge pull request #20079 from maanyagoenka/img-support
Lennart Poettering [Wed, 11 Aug 2021 08:02:46 +0000 (10:02 +0200)] 
Merge pull request #20079 from maanyagoenka/img-support

systemd-analyze: root validation and verify support for root arg and discrete images

4 years agotest-network: test static lease outside of pool 20423/head
Yu Watanabe [Wed, 11 Aug 2021 07:42:06 +0000 (16:42 +0900)] 
test-network: test static lease outside of pool

4 years agosd-dhcp-server: support static lease outside of address pool
Yu Watanabe [Wed, 11 Aug 2021 07:18:45 +0000 (16:18 +0900)] 
sd-dhcp-server: support static lease outside of address pool

Closes #20341.

4 years agosd-dhcp-server: fix possible double-free or use-after-free
Yu Watanabe [Wed, 11 Aug 2021 07:20:35 +0000 (16:20 +0900)] 
sd-dhcp-server: fix possible double-free or use-after-free

4 years agosd-dhcp-server: use hashmap_ensure_put()
Yu Watanabe [Wed, 11 Aug 2021 06:25:20 +0000 (15:25 +0900)] 
sd-dhcp-server: use hashmap_ensure_put()

4 years agoGet rid of dangling setutxent()
MertsA [Wed, 11 Aug 2021 03:54:50 +0000 (20:54 -0700)] 
Get rid of dangling setutxent()

utmp_wall() and utmp_put_dead_process() called setutxent() directly instead of the stub in utmp-wtmp.h and never called endutxent(). This would leave /run/utmp left open by PID 1 or journald. This can be reproduced by e.g. lsof /run/utmp and systemd-cat -p 0 echo test. For utmp_put_dead_process() it would only leave it open if it returned early before calling write_utmp_wtmp()

4 years agoactivate: use global variable instead of passing char **envp around
Zbigniew Jędrzejewski-Szmek [Tue, 10 Aug 2021 15:34:07 +0000 (17:34 +0200)] 
activate: use global variable instead of passing char **envp around

The effect should be the same, but the code is less verbose.
In particular, the variable was called envp in parts of the code,
but in other parts, we had a local envp variable, and envp was called
env.

4 years agomachinectl: allow --setenv=FOO
Zbigniew Jędrzejewski-Szmek [Tue, 10 Aug 2021 15:28:47 +0000 (17:28 +0200)] 
machinectl: allow --setenv=FOO

4 years agorun: allow --setenv=FOO
Zbigniew Jędrzejewski-Szmek [Tue, 10 Aug 2021 15:25:47 +0000 (17:25 +0200)] 
run: allow --setenv=FOO

4 years agohomectl: allow --setenv=FOO
Zbigniew Jędrzejewski-Szmek [Tue, 10 Aug 2021 15:19:31 +0000 (17:19 +0200)] 
homectl: allow --setenv=FOO

4 years agonspawn: allow --setenv=FOO as equivalent to --setenv=FOO=$FOO
Zbigniew Jędrzejewski-Szmek [Tue, 10 Aug 2021 15:11:56 +0000 (17:11 +0200)] 
nspawn: allow --setenv=FOO as equivalent to --setenv=FOO=$FOO

systemd-socket-activate has supported such a mode since
5e65c93a433447b15180249166f7b3944c3e6156. '--setenv=FOO=$FOO' is a fairly
common use in scripts, and it's nicer to do this automatically without worrying
about quoting and whatnot.

https://github.com/systemd/mkosi/pull/765 added the same to 'mkosi --environment='.

4 years agobasic/env-util: add a mode where we pull in the variable value from environment
Zbigniew Jędrzejewski-Szmek [Tue, 10 Aug 2021 15:03:12 +0000 (17:03 +0200)] 
basic/env-util: add a mode where we pull in the variable value from environment

4 years agotest-env-util: extend the test for strv_env_merge() a bit
Zbigniew Jędrzejewski-Szmek [Tue, 10 Aug 2021 14:53:00 +0000 (16:53 +0200)] 
test-env-util: extend the test for strv_env_merge() a bit

4 years agoAdd implicit sentinel to strv_env_merge()
Zbigniew Jędrzejewski-Szmek [Sat, 7 Aug 2021 08:16:19 +0000 (10:16 +0200)] 
Add implicit sentinel to strv_env_merge()

Just to make it a tiny bit nicer to use.

4 years agoalloc-util: make mfree() typesafe
Lennart Poettering [Tue, 10 Aug 2021 12:36:00 +0000 (14:36 +0200)] 
alloc-util: make mfree() typesafe

Make sure we return the same type as we accept.

One incorrect use of mfree() is discovered and fixed this way.

4 years agoMerge pull request #20414 from yuwata/sd-dhcp6-client-trivial-fixes
Luca Boccassi [Tue, 10 Aug 2021 19:32:29 +0000 (20:32 +0100)] 
Merge pull request #20414 from yuwata/sd-dhcp6-client-trivial-fixes

sd-dhcp6-client: several fixes and cleanups

4 years agoMerge pull request #20413 from poettering/mempmen
Luca Boccassi [Tue, 10 Aug 2021 19:31:36 +0000 (20:31 +0100)] 
Merge pull request #20413 from poettering/mempmen

add and use mempmem_safe()

4 years agocore: don't go to disk for masked unit state if we know it anyway
Lennart Poettering [Tue, 10 Aug 2021 12:50:54 +0000 (14:50 +0200)] 
core: don't go to disk for masked unit state if we know it anyway

This should speed up our selinux checks a bit, since in most cases we
already know if a unit is masked, so use that info.

4 years agosystemd-analyze: add root to find and verify executable 20079/head
Maanya Goenka [Wed, 4 Aug 2021 19:00:31 +0000 (12:00 -0700)] 
systemd-analyze: add root to find and verify executable

4 years agopath-util: teach find_executable_full how to look into the root directory
Maanya Goenka [Wed, 4 Aug 2021 18:59:45 +0000 (11:59 -0700)] 
path-util: teach find_executable_full how to look into the root directory

When the root parameter in find_executable_full is set, chase_symlinks prefixes this root
to every check of the path name to find the complete path of the execuatble in case the
path provided is not absolute. This is only done for the non NULL root because otherwise
the chase_symlinks function would alter the behavior of some of the callers which would
in turn alter the outputs in a way that is undesirable. The find_execuatble_full function is
invoked by the verify_executable function in analyze-verify.

4 years agounits: added factory-reset.target
Yegor Alexeyev [Sun, 25 Jul 2021 02:20:27 +0000 (05:20 +0300)] 
units: added factory-reset.target

4 years agoresolved: Don't omit AD bit in reply if DO is set in the query
Jack Dähn [Tue, 3 Aug 2021 06:07:05 +0000 (08:07 +0200)] 
resolved: Don't omit AD bit in reply if DO is set in the query

Set the AD bit in the response if the data is authenticated and AD or DO is set in the query, as suggested by section 5.8 of RFC6840.

Fixes #20332

4 years agotest: fix TEST-10-ISSUE-2467
Michael Biebl [Mon, 9 Aug 2021 17:45:48 +0000 (19:45 +0200)] 
test: fix TEST-10-ISSUE-2467

Depending on the timing, socat will either get ECONNREFUSED oder EPIPE
from systemd. The latter will cause it to exit(1) and subsequently the
test to fail.
We are not actually interested in the return code of socat though. The
test is supposed to check, whether rate limiting of a socket unit works
properly.

So ignore any failures from the socat invocation and instead check, if
test10.socket is in state "failed" with result "trigger-limit-hit" after
it has been triggered.

TriggerLimitIntervalSec= by default is set to 2s. A "sleep 10" should
give systemd enough time even on slower machines, to reach the trigger
limit.

For better readability, break the test into separate ExecStart lines.

Fixes #19154.

4 years agoupdate TODO
Lennart Poettering [Tue, 10 Aug 2021 12:35:42 +0000 (14:35 +0200)] 
update TODO