]> git.ipfire.org Git - thirdparty/systemd.git/log
thirdparty/systemd.git
4 years agolog: drop unused LogRealm 18281/head
Yu Watanabe [Sun, 17 Jan 2021 04:59:33 +0000 (13:59 +0900)] 
log: drop unused LogRealm

Already no binary is built with LOG_REALM= argument.
Hence, we can safely drop LogRealm now.

4 years agoudev: stop to use LOG_REALM_UDEV
Yu Watanabe [Sun, 17 Jan 2021 04:32:29 +0000 (13:32 +0900)] 
udev: stop to use LOG_REALM_UDEV

Before this commit, udevd is built with LOG_REALM=LOG_REALM_UDEV.
However, log level specified by e.g. environment variable or kernel
command line option are also passed to LOG_REALM_SYSTEMD. So, the
maximum log level for the two realms are always equivalent, and it is
not necessary to specify the build option. Hence drop it.

4 years agoselinux-util: do not use log_internal_realm() without setting log realm
Yu Watanabe [Sun, 17 Jan 2021 04:02:54 +0000 (13:02 +0900)] 
selinux-util: do not use log_internal_realm() without setting log realm

4 years agolog: drop unused log_full_errno_realm() macro
Yu Watanabe [Sun, 17 Jan 2021 03:57:07 +0000 (12:57 +0900)] 
log: drop unused log_full_errno_realm() macro

4 years agoMerge pull request #18363 from bluca/test_image_reuse
Zbigniew Jędrzejewski-Szmek [Mon, 25 Jan 2021 08:25:56 +0000 (09:25 +0100)] 
Merge pull request #18363 from bluca/test_image_reuse

tests: build the image once and then copy/extend it

4 years agoTranslated using Weblate (Punjabi)
A S Alam [Sun, 24 Jan 2021 16:38:24 +0000 (17:38 +0100)] 
Translated using Weblate (Punjabi)

Currently translated at 3.7% (7 of 189 strings)

Co-authored-by: A S Alam <amanpreet.alam@gmail.com>
Translate-URL: https://translate.fedoraproject.org/projects/systemd/master/pa/
Translation: systemd/main

4 years agosd-bus: Add sd_bus_reply()
Daan De Meyer [Sat, 23 Jan 2021 15:16:57 +0000 (15:16 +0000)] 
sd-bus: Add sd_bus_reply()

While sd-bus already provides sd_bus_call() for calling a method
from a complete bus message object, We don't have an equivalent
function for replying from a method with a complete bus message
object.

Currently, we use sd_bus_send(call->bus, m, NULL) instead. Let's
add a shorthand for this pattern and name it sd_bus_reply().

4 years agoAvoid error when bind mount source already exists
Stefan Agner [Mon, 28 Dec 2020 20:46:40 +0000 (21:46 +0100)] 
Avoid error when bind mount source already exists

4 years agoMerge pull request #18361 from anitazha/oomdconfigtime
Zbigniew Jędrzejewski-Szmek [Sun, 24 Jan 2021 19:12:28 +0000 (20:12 +0100)] 
Merge pull request #18361 from anitazha/oomdconfigtime

oom: some improvements and fixes

4 years agohwdb.d/60-keyboard.hwdb: Fixes for Samsung 900X laptops
John Slade [Sun, 24 Jan 2021 14:04:03 +0000 (14:04 +0000)] 
hwdb.d/60-keyboard.hwdb: Fixes for Samsung 900X laptops

It is only the Samsung 900X3A which has the alternative function keys.

The 900X3B, 900X4B have the standard Series 9 layout - as defined above.

Note: the 900X4A is not a valid laptop model.

4 years agotests: build the image once and then copy/extend it 18363/head
Luca Boccassi [Wed, 6 Jan 2021 21:42:28 +0000 (21:42 +0000)] 
tests: build the image once and then copy/extend it

Building custom images for each test takes a lot of time.
Build the default one, and if the test needs incompatible changes
just copy it and extend it instead.

4 years agotest: note where the bionic-* CI tool/integration lives
Luca Boccassi [Tue, 5 Jan 2021 14:05:15 +0000 (14:05 +0000)] 
test: note where the bionic-* CI tool/integration lives

4 years agoMerge pull request #18358 from jwrdegoede/hwdb-2-accel-quirks
Luca Boccassi [Sun, 24 Jan 2021 11:35:59 +0000 (11:35 +0000)] 
Merge pull request #18358 from jwrdegoede/hwdb-2-accel-quirks

Hwdb 2 accel quirks

4 years agodocs: Update HACKING.md with the mkosi boot/qemu commands + options
Daan De Meyer [Sat, 23 Jan 2021 12:16:18 +0000 (12:16 +0000)] 
docs: Update HACKING.md with the mkosi boot/qemu commands + options

Let's use the mkosi commands in HACKING.md and recommend some options
that speed up mkosi builds. Also includes some other small improvements.

4 years agodocs: Add a section to HACKING.md on using mkosi and clangd together
Daan De Meyer [Sat, 23 Jan 2021 12:17:36 +0000 (12:17 +0000)] 
docs: Add a section to HACKING.md on using mkosi and clangd together

While it's perfectly possible today to completely rely on mkosi for
building and testing systemd, to get code completion and other IDE
niceties to work properly, it's still necessary to build systemd
locally.

Recently, mkosi gained the ability to allow external programs to
communicate with the build script. We can use this feature to run
the clangd language server in the mkosi build image via a custom
build script to provide IDE features in editors without requiring
developers to build systemd on the host or install any of systemd's
build dependencies locally.

This commit adds the necessary information on how to set this up
to HACKING.md.

4 years agooom: update extended test to remove swap gating 18361/head
Anita Zhang [Sun, 24 Jan 2021 09:34:23 +0000 (01:34 -0800)] 
oom: update extended test to remove swap gating

4 years agooom: fix reclaim activity detection
Anita Zhang [Sun, 24 Jan 2021 09:22:51 +0000 (01:22 -0800)] 
oom: fix reclaim activity detection

This should have been checking for any reclaim activity within a larger interval
of time rather than within the past second. On systems with swap this
doesn't seem to have mattered too much as reclaim would always increase when
memory pressure was elevated. But testing in the no swap case having
this larger interval made a difference between oomd killing or not.

4 years agooom: make swap a soft requirement
Anita Zhang [Sun, 24 Jan 2021 06:10:42 +0000 (22:10 -0800)] 
oom: make swap a soft requirement

4 years agooom: make memory pressure duration configurable through oomd.conf
Anita Zhang [Sun, 24 Jan 2021 08:16:19 +0000 (00:16 -0800)] 
oom: make memory pressure duration configurable through oomd.conf

4 years agohwdb: Add accel orientation quirk for the Jumper Ezpad 7 tablet 18358/head
Hans de Goede [Sat, 23 Jan 2021 21:36:58 +0000 (22:36 +0100)] 
hwdb: Add accel orientation quirk for the Jumper Ezpad 7 tablet

Add a quirk to fix the accelerometer orientation on
the Jumper Ezpad 7 tablet.

4 years agohwdb: Add accel orientation quirk for the Estar Beauty HD tablet
Hans de Goede [Wed, 23 Dec 2020 20:45:18 +0000 (21:45 +0100)] 
hwdb: Add accel orientation quirk for the Estar Beauty HD tablet

Add a quirk to fix the accelerometer orientation on the Estar Beauty HD
(marked as model-number MID 7316R on the back) tablet.

4 years agoMerge pull request #18349 from poettering/import-fixlets2
Luca Boccassi [Sat, 23 Jan 2021 00:21:48 +0000 (00:21 +0000)] 
Merge pull request #18349 from poettering/import-fixlets2

minor importd fixlets

4 years agorepart: improve help text
Lennart Poettering [Fri, 22 Jan 2021 19:29:58 +0000 (20:29 +0100)] 
repart: improve help text

Finding "partitions" in a "directory" is a bit weird. Let's find
"partition definitions" there, after all the option is called
--definitions=

4 years agoimport: downgrade error messages we ignore to LOG_WARNING 18349/head
Lennart Poettering [Fri, 22 Jan 2021 19:39:29 +0000 (20:39 +0100)] 
import: downgrade error messages we ignore to LOG_WARNING

4 years agoimport: set up btrfs qgroups on correct hierarchy
Lennart Poettering [Fri, 22 Jan 2021 17:49:44 +0000 (18:49 +0100)] 
import: set up btrfs qgroups on correct hierarchy

Also, simplify import_assign_pool_quota_and_warn(), don't do the same
thing twice. Let's just allow the caller call this twice.

4 years agoimport: drop redundant {}, as per coding style
Lennart Poettering [Fri, 22 Jan 2021 16:35:47 +0000 (17:35 +0100)] 
import: drop redundant {}, as per coding style

4 years agoimport: introduce ImportFlags flags field
Lennart Poettering [Fri, 22 Jan 2021 16:10:50 +0000 (17:10 +0100)] 
import: introduce ImportFlags flags field

This merges the two flags that are passed to the ImportTar/ImportRaw
objects into a single flags parameter, which we then can extend more
easily later on.

No change in behaviour.

This is inspired by 133b34f69a72dc90d4e336837d699245390c9f50 which does
the same for PullTar/PullRaw.

4 years agoimport: don't apply empty_or_dash_to_null() to stuff we know is NULL anyway
Lennart Poettering [Fri, 22 Jan 2021 15:47:22 +0000 (16:47 +0100)] 
import: don't apply empty_or_dash_to_null() to stuff we know is NULL anyway

4 years agofeature: display status with a different shape depending on the status (#17728)
Jiehong [Wed, 20 Jan 2021 14:52:35 +0000 (15:52 +0100)] 
feature: display status with a different shape depending on the status (#17728)

4 years agoMerge pull request #18302 from bluca/mount_image_runtime
Luca Boccassi [Fri, 22 Jan 2021 17:49:11 +0000 (17:49 +0000)] 
Merge pull request #18302 from bluca/mount_image_runtime

systemctl/core: add DBUS method to mount images without service restart

4 years agocore: update setings on the unit and job as the result of ExecCondition=
Anita Zhang [Tue, 12 Jan 2021 04:04:20 +0000 (20:04 -0800)] 
core: update setings on the unit and job as the result of ExecCondition=

Update ExecCondition= to set Unit->condition_result and return JOB_DONE
in the Job results  if the check fails so as to match the current behavior
of ConditionXYZ= w.r.t units/jobs dependency checks.

Fixes: #18207
4 years agohostname: accept an empty hostname to unset transient hostname
Yu Watanabe [Fri, 22 Jan 2021 01:13:01 +0000 (10:13 +0900)] 
hostname: accept an empty hostname to unset transient hostname

Follow-up for 60e4fb4240b24bdd2d4299d8d844f48093df8807.

Before 60e4fb4240b24bdd2d4299d8d844f48093df8807,
`hostnamectl --transient set-hostname ""` unsets the transient hostname.
But after the commit, it is refused. This fixes the issue.

4 years agoMerge pull request #16228 from ddstreet/administrative_state
Yu Watanabe [Fri, 22 Jan 2021 03:11:29 +0000 (12:11 +0900)] 
Merge pull request #16228 from ddstreet/administrative_state

Add ActivationPolicy= [Link] parameter

4 years agoMerge pull request #18338 from poettering/print-table-with-bells-and-whistles
Luca Boccassi [Thu, 21 Jan 2021 22:41:18 +0000 (22:41 +0000)] 
Merge pull request #18338 from poettering/print-table-with-bells-and-whistles

json/table printing unification work

4 years agoportabled: add BindPaths=/run back into "trusted" policy
Lennart Poettering [Thu, 21 Jan 2021 14:06:13 +0000 (15:06 +0100)] 
portabled: add BindPaths=/run back into "trusted" policy

This partly reverts fe239c7d7d6227209234100f4e0a36dd952a5587, where I
accidentally dropped only line for /run that was unlike the others.
Oops.

@bluca, thanks for noticing:

https://github.com/systemd/systemd/pull/18329#pullrequestreview-573343549

4 years agoci: point the Fossies badge to main
Evgeny Vereshchagin [Thu, 21 Jan 2021 19:02:22 +0000 (19:02 +0000)] 
ci: point the Fossies badge to main

Another follow-up to https://github.com/systemd/systemd/issues/16834

4 years agosystemctl: add new option to mount image inside a running service namespace 18302/head
Luca Boccassi [Thu, 21 Jan 2021 18:37:40 +0000 (18:37 +0000)] 
systemctl: add new option to mount image inside a running service namespace

Use the new DBUS method and follow the same pattern as the
systemctl bind command.

4 years agopid1: add DBUS method to mount image inside a running service namespace
Luca Boccassi [Thu, 21 Jan 2021 18:36:33 +0000 (18:36 +0000)] 
pid1: add DBUS method to mount image inside a running service namespace

Follow the same pattern as the BindMount method.

4 years agomount-util: add helper to mount image inside live namespace
Luca Boccassi [Thu, 21 Jan 2021 18:35:31 +0000 (18:35 +0000)] 
mount-util: add helper to mount image inside live namespace

4 years agodissect: split verity_dissect_and_mount helper out for reuse
Luca Boccassi [Thu, 21 Jan 2021 18:32:44 +0000 (18:32 +0000)] 
dissect: split verity_dissect_and_mount helper out for reuse

4 years agocore: split read_mount_options helper out for reuse
Luca Boccassi [Thu, 21 Jan 2021 18:31:45 +0000 (18:31 +0000)] 
core: split read_mount_options helper out for reuse

4 years agoci: change the Semaphore CI badge branch to 'main'
Frantisek Sumsal [Thu, 21 Jan 2021 17:51:15 +0000 (18:51 +0100)] 
ci: change the Semaphore CI badge branch to 'main'

Follow-up to #16834

4 years agoman: share a common explanation for --json= 18338/head
Lennart Poettering [Thu, 21 Jan 2021 17:21:41 +0000 (18:21 +0100)] 
man: share a common explanation for --json=

4 years agobusctl: show --help in pager
Lennart Poettering [Thu, 21 Jan 2021 16:59:01 +0000 (17:59 +0100)] 
busctl: show --help in pager

We do this for other larger --help texts, do so here too.

4 years agodissect: almost port systemd-dissect to table_print_with_pager()
Lennart Poettering [Thu, 21 Jan 2021 17:12:23 +0000 (18:12 +0100)] 
dissect: almost port systemd-dissect to table_print_with_pager()

This doesn't actually port systemd-dissect to table_print_with_pager()
but at least rearranges things so that similar behaviour is exposed. The
reason it's not ported over 1:1 is that systemd-dissect actually adjusts
the JSON output of the table with additional info, and doesn't print the
table 1:1 as JSON.

4 years agosysext: port to table_print_with_pager()
Lennart Poettering [Thu, 21 Jan 2021 16:57:13 +0000 (17:57 +0100)] 
sysext: port to table_print_with_pager()

This adds --no-legend as side-effect.

4 years agorepart: port to table_print_with_pager()
Lennart Poettering [Thu, 21 Jan 2021 16:40:02 +0000 (17:40 +0100)] 
repart: port to table_print_with_pager()

This adds support for showing a pager, turning of the legend, like in
the other tools.

4 years agobusctl/homectl: port the obvious cases to table_print_with_pager()
Lennart Poettering [Thu, 21 Jan 2021 16:39:01 +0000 (17:39 +0100)] 
busctl/homectl: port the obvious cases to table_print_with_pager()

4 years agoformat-table: add new helper table_print_with_pager()
Lennart Poettering [Thu, 21 Jan 2021 16:36:53 +0000 (17:36 +0100)] 
format-table: add new helper table_print_with_pager()

This adds a new function table_print_with_pager() which is a wrapper
around table_print_json()/table_print() but spawns a pager first, if
that's enabled, and optionally turns off the header line of the table.

This addresses the fact that many of our tools actually keep doing very
this very similar stuff, over and over again. Let's unify this in one
place.

4 years agosave link activation policy to state file and display in networkctl 16228/head
Dan Streetman [Thu, 13 Aug 2020 15:52:53 +0000 (11:52 -0400)] 
save link activation policy to state file and display in networkctl

4 years agotest: add ActivationPolicy= unit tests
Dan Streetman [Thu, 18 Jun 2020 22:31:18 +0000 (18:31 -0400)] 
test: add ActivationPolicy= unit tests

4 years agonetwork: add ActivationPolicy= configuration parameter
Dan Streetman [Thu, 18 Jun 2020 20:09:40 +0000 (16:09 -0400)] 
network: add ActivationPolicy= configuration parameter

This parameter allows configuring the activation policy for an interface,
meaning how it manages the interface's administrative state (IFF_UP flag).
The policy can be configured to bring the interface either up or down when
the interface is (re)configured, to always force the interface either up or
down, or to never change the interface administrative state.

If the interface is bound with BindCarrier=, its administrative state is
controlled by the interface(s) it's bound to, and this parameter is forced
to 'bound'.

This changes the default behavior of how systemd-networkd sets the IFF_UP
flag; previously, it was set up (if not already up) every time the
link_joined() function was called. Now, with the default ActivationPolicy=
setting of 'up', it will only set the IFF_UP flag once, the first time
link_joined() is called, during an interface's configuration; and on
the first link_joined() call each time the interface is reconfigured.

Fixes: #3031
Fixes: #17437
4 years agoGH Actions: switch to main
Evgeny Vereshchagin [Thu, 21 Jan 2021 13:52:18 +0000 (13:52 +0000)] 
GH Actions: switch to main

It's just a follow-up to https://github.com/systemd/systemd/issues/16834

4 years agohwdb: add Medion Akoya E2228T MD61900 (#18317)
corvusnix [Thu, 21 Jan 2021 08:21:50 +0000 (09:21 +0100)] 
hwdb: add Medion Akoya E2228T MD61900 (#18317)

Add a model-specific entry for Medion Akoya E2228T MD61900 accelerometer to 60-sensor.hwdb so the orientation is detected correctly.

4 years agoMerge pull request #18327 from benzea/benzea/libfprint-autosuspend-hwdb
Zbigniew Jędrzejewski-Szmek [Thu, 21 Jan 2021 08:09:34 +0000 (09:09 +0100)] 
Merge pull request #18327 from benzea/benzea/libfprint-autosuspend-hwdb

Pull libfprint autosuspend hwdb

4 years agoMerge pull request #18329 from poettering/notify-chroot
Yu Watanabe [Thu, 21 Jan 2021 04:16:59 +0000 (13:16 +0900)] 
Merge pull request #18329 from poettering/notify-chroot

chroot/sd_notify() fixes

4 years agovarlink: make 'userdata' pointer inheritance from varlink server to connection optional
Lennart Poettering [Wed, 20 Jan 2021 18:15:55 +0000 (19:15 +0100)] 
varlink: make 'userdata' pointer inheritance from varlink server to connection optional

@keszybz's right on
https://github.com/systemd/systemd/pull/18248#issuecomment-760798473:
swapping out the userdata pointer of a live varlink connection is iffy.

Let's fix this by making the userdata inheritance from VarlinkServer
object to the Varlink connection object optional: we want it for most
cases, but not all, i.e. all those cases where the calls implemented as
varlink methods are stateless and can be answered synchronously. For the
other cases (i.e. where we want per-connection objects that wrap the
asynchronous operation as it goes on) let's not do such inheritance but
initialize the userdata pointer only once we have it. THis means the
original manager object must be manually retrieved from the
VarlinkServer object, which in turn needs to be requested from the
Varlink connection object.

The userdata inheritance is now controlled by the
VARLINK_INHERIT_USERDATA flag passed at VarlinkServer construction.

Alternative-to: #18248
4 years agoMerge pull request #18311 from poettering/sysext-fixups
Yu Watanabe [Wed, 20 Jan 2021 22:20:04 +0000 (07:20 +0900)] 
Merge pull request #18311 from poettering/sysext-fixups

sysext: post-merge fixups

4 years agoportabled: update profiles to current semantics 18329/head
Lennart Poettering [Wed, 20 Jan 2021 21:28:30 +0000 (22:28 +0100)] 
portabled: update profiles to current semantics

MountAPIVFS= implicitly mounts /run as tmpfs now, no need to do this
explicitly.

The notification socket is now implicitly mounted too, if NotifyAccess=
and RootImage=/RootDirectory= are used together.

4 years agoman: document how get logging to work in a RootDirectory=/RootImage= environment
Lennart Poettering [Wed, 20 Jan 2021 21:27:41 +0000 (22:27 +0100)] 
man: document how get logging to work in a RootDirectory=/RootImage= environment

Fixes: #18051
4 years agocore: make NotifyAccess= in combination with RootDirectory=/RootImage= work
Lennart Poettering [Wed, 20 Jan 2021 20:57:21 +0000 (21:57 +0100)] 
core: make NotifyAccess= in combination with RootDirectory=/RootImage= work

Previously if people enabled RootDirectory=/RootImage= and NotifyAccess=
together, things wouldn't work, they'd have to explicitly add
BindReadOnlyPaths=/run/systemd/notify too.

Let's make this implicit. Since both options are opt-in, if people use
them together it would be pointless not also defining the
BindReadOnlyPaths= entry, in which case we can just do it automatically.

See: #18051

4 years agohwdb: Pull autosuspend rules from upstream libfprint 18327/head
Benjamin Berg [Tue, 19 Jan 2021 13:59:42 +0000 (14:59 +0100)] 
hwdb: Pull autosuspend rules from upstream libfprint

libfprint includes a list of known fingerprint readers that can be
autosuspended. Upstream libfprint generates this file from the USB IDs
registered to drivers and a list of well-known readers that are
currently unsupported.

Closes: #17663
4 years agoMerge pull request #18325 from ssahani/more-cleanup
Luca Boccassi [Wed, 20 Jan 2021 20:48:18 +0000 (20:48 +0000)] 
Merge pull request #18325 from ssahani/more-cleanup

Tree wide various cleanup

4 years agoudev/cdrom_id: re-enable logging related functions
Yu Watanabe [Wed, 20 Jan 2021 09:48:42 +0000 (18:48 +0900)] 
udev/cdrom_id: re-enable logging related functions

The logging related functions are mistakenly located in main() by
a084b3878955bc1b93adcaa7a41acb83b18eff9b, and dropped by
23afa884d4f3bcd97160a893816f9ba170f62ad4.

4 years agoman: document recent systemd-sysext interface changes 18311/head
Lennart Poettering [Tue, 19 Jan 2021 18:53:20 +0000 (19:53 +0100)] 
man: document recent systemd-sysext interface changes

4 years agosysext: add --force swich for forcibly ignoring version incompatibilities
Lennart Poettering [Tue, 19 Jan 2021 18:59:53 +0000 (19:59 +0100)] 
sysext: add --force swich for forcibly ignoring version incompatibilities

4 years agosysext: rework command line interface to be verb-based
Lennart Poettering [Tue, 19 Jan 2021 18:57:56 +0000 (19:57 +0100)] 
sysext: rework command line interface to be verb-based

As suggested by @yuwata:

https://github.com/systemd/systemd/pull/18181#pullrequestreview-570826113

4 years agosysext: split version validation logic into function of its own
Lennart Poettering [Tue, 19 Jan 2021 18:57:13 +0000 (19:57 +0100)] 
sysext: split version validation logic into function of its own

Just some simple refactoring to simplify the logic.

4 years agosysext: use log_setup_cli()
Lennart Poettering [Tue, 19 Jan 2021 18:54:59 +0000 (19:54 +0100)] 
sysext: use log_setup_cli()

4 years agopath-util: simplify check_x_access()
Yu Watanabe [Tue, 19 Jan 2021 19:43:32 +0000 (04:43 +0900)] 
path-util: simplify check_x_access()

Follow-up for ece852c84592220c3d6bb5a055fd8b84ea097290.

This addresses the following comments by the Lennart:
---
hmm, so this now does two access() calls for the case where the fd is
not requested, and opens things up for races (theoretically, …). now,
the access() code path was in place for optimization, but if an optimization
is less sexy than the original (and i think it is less sexy, since more
than one syscall, and non-atomic), i think we shouldn't do the optimization.

maybe we should just always use open(O_PATH) now, and then fstat() it to
check if regular file, and then access_fd() it for checking if its executable.

4 years agosysext: install in rootbindir, remove template from unit
Luca Boccassi [Tue, 19 Jan 2021 17:23:57 +0000 (17:23 +0000)] 
sysext: install in rootbindir, remove template from unit

This reverts commit 71ad75f30641b90f9ca0088869f164d9d085430f.

4 years agobtrfs-util: tighten variable scope used in loop 18325/head
Susant Sahani [Wed, 20 Jan 2021 14:14:30 +0000 (15:14 +0100)] 
btrfs-util: tighten variable scope used in loop

4 years agoanalyze: tighten variable scope used in loop
Susant Sahani [Wed, 20 Jan 2021 13:21:37 +0000 (14:21 +0100)] 
analyze: tighten variable scope used in loop

4 years agosd-event: Use hashmap_ensure_put
Susant Sahani [Wed, 20 Jan 2021 13:13:08 +0000 (14:13 +0100)] 
sd-event: Use hashmap_ensure_put

4 years agosd-event: Use hashmap_ensure_put
Susant Sahani [Wed, 20 Jan 2021 12:58:00 +0000 (13:58 +0100)] 
sd-event: Use hashmap_ensure_put

4 years agosd-device: Use TAKE_PTR
Susant Sahani [Wed, 20 Jan 2021 12:54:59 +0000 (13:54 +0100)] 
sd-device: Use TAKE_PTR

4 years agosd-device: Use hashmap_ensure_put
Susant Sahani [Wed, 20 Jan 2021 12:54:01 +0000 (13:54 +0100)] 
sd-device: Use hashmap_ensure_put

4 years agoudev-rules: use ordered_hashmap_ensure_put
Susant Sahani [Wed, 20 Jan 2021 14:13:02 +0000 (15:13 +0100)] 
udev-rules: use ordered_hashmap_ensure_put

4 years agonetwork: networkd-network use TAKE_PTR
Susant Sahani [Wed, 20 Jan 2021 12:00:15 +0000 (13:00 +0100)] 
network: networkd-network use TAKE_PTR

4 years agonetwork: Use hashmap_ensure_put
Susant Sahani [Wed, 20 Jan 2021 11:59:03 +0000 (12:59 +0100)] 
network: Use hashmap_ensure_put

4 years agonetwork: ndisc - Use ordered_set_ensure_put
Susant Sahani [Wed, 20 Jan 2021 11:56:44 +0000 (12:56 +0100)] 
network: ndisc - Use ordered_set_ensure_put

4 years agomachine: Use hashmap_ensure_put
Susant Sahani [Wed, 20 Jan 2021 11:53:15 +0000 (12:53 +0100)] 
machine: Use hashmap_ensure_put

4 years agologind: Use hashmap_ensure_put
Susant Sahani [Wed, 20 Jan 2021 11:51:08 +0000 (12:51 +0100)] 
logind: Use hashmap_ensure_put

4 years agojournal: Use cleanup_free
Susant Sahani [Wed, 20 Jan 2021 11:46:42 +0000 (12:46 +0100)] 
journal: Use cleanup_free

4 years agoJournal: Use hashmap_ensure_put
Susant Sahani [Wed, 20 Jan 2021 11:29:08 +0000 (12:29 +0100)] 
Journal: Use hashmap_ensure_put

4 years agoudev: use DEFINE_MAIN_FUNCTION in cdrom_id
Dan Streetman [Sun, 17 Jan 2021 22:46:12 +0000 (17:46 -0500)] 
udev: use DEFINE_MAIN_FUNCTION in cdrom_id

This was failing s390x ubuntu ci due to an old version of binutils;
the binutils package in use for ubuntu ci tests has been updated and
the build no longer fails, so this can use the macro again.

Fixes: #18165
4 years agotree-wide: ignore messages with too long control data
Lennart Poettering [Tue, 19 Jan 2021 20:34:20 +0000 (21:34 +0100)] 
tree-wide: ignore messages with too long control data

Apparently SELinux inserts control data into AF_UNIX datagrams where we
don't expect it, thus miscalculating the control data. This looks like
something to fix in SELinux, but we still should handle this gracefully
and just drop the offending datagram and continue.

recvmsg_safe() actually already drops the datagram, it's just a matter
of actually ignoring EXFULL (which it generates if control data is too
large) in the right places.

This does this wherever an AF_UNIX/SOCK_DGRAM socket is used with
recvmsg_safe() that is not just internal communication.

Fixes: #17795
Follow-up for: 3691bcf3c5eebdcca5b4f1c51c745441c57a6cd1

4 years agoRevert "test-functions: make sure we test our own libudev instead of the host libudev"
Zbigniew Jędrzejewski-Szmek [Tue, 19 Jan 2021 17:39:50 +0000 (18:39 +0100)] 
Revert "test-functions: make sure we test our own libudev instead of the host libudev"

This reverts commit 73484ecff90f2cc235d827c0e955999bffe64dd0.

3976f372ae91e1cbe5ff9873aadfc2c43400452e moved libudev.so to be built in the
main directory, so this addition to $LD_LIBRARY_PATH is now obsolete.

After that commit, we build the following shared libraries:

build/libnss_myhostname.so.2
build/libnss_mymachines.so.2
build/libnss_resolve.so.2
build/libnss_systemd.so.2
build/libsystemd.so.0.30.0
build/libudev.so.1.7.0
build/pam_systemd.so
build/pam_systemd_home.so
build/src/boot/efi/stub.so
build/src/boot/efi/systemd_boot.so
build/src/shared/libsystemd-shared-247.so

EFI stubs don't matter, and libsystemd-shared-nnn.so is loaded through rpath,
and is doesn't need to and shouldn't be in $LD_LIBRARY_PATH. In effect, we only
ever need to add the main build directory to the search path.

4 years agomarkdown: fix comment
Lennart Poettering [Tue, 19 Jan 2021 21:24:05 +0000 (22:24 +0100)] 
markdown: fix comment

Follow-up for 2d816c9804c1d565797faff85c0fdbfb15f1e9d5

4 years agomarkdown: suggest backticks around uname -a output
Lennart Poettering [Tue, 19 Jan 2021 21:24:05 +0000 (22:24 +0100)] 
markdown: suggest backticks around uname -a output

Fixes: #18238
4 years agoexecute: for processes where creds logic is turned off, don't pass creds path to...
Lennart Poettering [Tue, 19 Jan 2021 16:31:21 +0000 (17:31 +0100)] 
execute: for processes where creds logic is turned off, don't pass creds path to namespace logic

Otherwise, the namespace logic will try to mount a dir that doesn't
actually exist.

Fixes: #18116
4 years agoMerge pull request #18307 from poettering/import-verity-download
Lennart Poettering [Tue, 19 Jan 2021 19:02:34 +0000 (20:02 +0100)] 
Merge pull request #18307 from poettering/import-verity-download

importd: when downloading raw image, also download .roothash.p7s and .verity along with it

4 years agoimport: refactor how we do gpg validation 18307/head
Lennart Poettering [Fri, 15 Jan 2021 23:06:13 +0000 (00:06 +0100)] 
import: refactor how we do gpg validation

Let's split out the actual gpg logic into a helper function, so that we
can add alternative validations later on.

4 years agoimport: optionally pull .verity + .roothash.p7s data when downloading
Lennart Poettering [Fri, 15 Jan 2021 22:18:54 +0000 (23:18 +0100)] 
import: optionally pull .verity + .roothash.p7s data when downloading

We already had support for downlading a .nspawn and a .roothash file,
let's make the set complete, and also download .verity + roothash.p7s if
it exists, as nspawn consumes that.

Since there are now four kinds of additional resources to acquire, let's
introduce a PullFlags flags value for this instead of separate 'bool'
variables, it's just too many to always pass those around on the
function parameter list.

4 years agoimport: ignore non-successful HTTP codes for collecing image metadata
Lennart Poettering [Fri, 15 Jan 2021 21:55:15 +0000 (22:55 +0100)] 
import: ignore non-successful HTTP codes for collecing image metadata

Previously we'd collect the data from redirects too, which wasn't
particularly terrible, since these typically don't carry the data we
were interested in, but it's still incorrect to do so.

4 years agoimport: turn on HTTP logging in debug mode
Lennart Poettering [Fri, 15 Jan 2021 21:54:42 +0000 (22:54 +0100)] 
import: turn on HTTP logging in debug mode

4 years agoimport: small memory management simplification
Lennart Poettering [Fri, 15 Jan 2021 20:23:20 +0000 (21:23 +0100)] 
import: small memory management simplification

4 years agoimport: reset PullJob properly
Lennart Poettering [Fri, 15 Jan 2021 20:22:27 +0000 (21:22 +0100)] 
import: reset PullJob properly

Properly reset all fields that have to do with the current GET job when
we restart things. Previously we freed/reset only some stuff, leaking
some memory even.

4 years agoimport: rework how verification works
Lennart Poettering [Fri, 15 Jan 2021 15:45:29 +0000 (16:45 +0100)] 
import: rework how verification works

Previously the PullJob object took internal care of rerequested the
SHA256SUMS file, if requesting <image>.sha256 didn't work. This was a
weird a non-abstraction only used when actually getting the checksum
files.

Let's move this out of the PullJob, so that it is generic again, and
does roughly the same stuff for all resources it is used for: let's
define a generic .on_not_found() handler that can be set on a PullJob
object, and is called whenever with see HTTP 404, and may be used to
provide a new URL to try if the first didn't work.

This is also preparation for later work to support PKCS#7 signatures
instead of gpg signatures, where a similar logic is needed, and we thus
should have a generic infrastructure place.

This gets rid of the VerificationStyle field in the PullJob object:
instead of storing this non-generic field we just derive the same
information from the URL itself, which is safe, since we generated it
ourselves earlier.

4 years agoimport: use TAKE_PTR() where available
Lennart Poettering [Fri, 15 Jan 2021 15:42:58 +0000 (16:42 +0100)] 
import: use TAKE_PTR() where available