systemd/.git
5 years agotty-ask-pw-agent: properly propagate error
Lennart Poettering [Wed, 26 Aug 2020 21:10:50 +0000 (23:10 +0200)]
tty-ask-pw-agent: properly propagate error

(cherry picked from commit a4fd6cd3f575eb7e1452a74c2e35548739e3b252)

5 years agotty-ask-pw-agent: the message string might not be set
Lennart Poettering [Wed, 26 Aug 2020 21:05:46 +0000 (23:05 +0200)]
tty-ask-pw-agent: the message string might not be set

(cherry picked from commit 66bff73b4f91f8d2fdd385f9f1e2b6339055c9e4)

5 years agotty-ask-pw-agent: make sure "--list" works correctly
Lennart Poettering [Wed, 26 Aug 2020 21:02:13 +0000 (23:02 +0200)]
tty-ask-pw-agent: make sure "--list" works correctly

Fixes: #16836
(cherry picked from commit 4c4a018caba30a58c3549924b8521074bbe5adad)

5 years agoadd "list" verb to autocompleted commands
Olivier Le Moal [Thu, 27 Aug 2020 08:01:36 +0000 (10:01 +0200)]
add "list" verb to autocompleted commands

(cherry picked from commit 50574ed1ac2e27d8f25894c9db75bb0b4c28978f)

5 years agoshell-completion/zsh: add missing verbs for networkctl
Olivier Le Moal [Wed, 26 Aug 2020 14:03:35 +0000 (16:03 +0200)]
shell-completion/zsh: add missing verbs for networkctl

(cherry picked from commit 6ff45bc1a18855cb9b0c43ba13843c11dd10728f)

5 years agopath: Improve $PATH search directory case
Chris Down [Wed, 26 Aug 2020 17:49:27 +0000 (18:49 +0100)]
path: Improve $PATH search directory case

Previously:

1. last_error wouldn't be updated with errors from is_dir;
2. We'd always issue a stat(), even for binaries without execute;
3. We used stat() instead of access(), which is cheaper.

This change avoids all of those, by only checking inside X_OK-positive
case whether access() works on the path with an extra slash appended.
Thanks to Lennart for the suggestion.

(cherry picked from commit 33e1a5d8d3f792e1d98377fe439e123231032ec7)

5 years agopath: Skip directories when finalising $PATH search v246.3
Chris Down [Tue, 25 Aug 2020 20:59:11 +0000 (21:59 +0100)]
path: Skip directories when finalising $PATH search

Imagine $PATH /a:/b. There is an echo command at /b/echo. Under this
configuration, this works fine:

    % systemd-run --user --scope echo .
    Running scope as unit: run-rfe98e0574b424d63a641644af511ff30.scope
    .

However, if I do `mkdir /a/echo`, this happens:

    % systemd-run --user --scope echo .
    Running scope as unit: run-rcbe9369537ed47f282ee12ce9f692046.scope
    Failed to execute: Permission denied

We check whether the resulting file is executable for the performing
user, but of course, most directories are anyway, since that's needed to
list within it. As such, another is_dir() check is needed prior to
considering the search result final.

Another approach might be to check S_ISREG, but there may be more gnarly
edge cases there than just eliminating this obviously pathological
example, so let's just do this for now.

(cherry picked from commit 8b5cb69bc8b70d1dcc39ed2165907723099bd9d8)

5 years agorules: don't install 80-drivers.rules when kmod is disabled
Alec Moskvin [Sat, 1 Aug 2020 13:25:05 +0000 (09:25 -0400)]
rules: don't install 80-drivers.rules when kmod is disabled

(cherry picked from commit dd47b25220f69f869679089da5cc848cf9cd0c78)

5 years agozsh: correct journalctl command completion parsing
Ronan Pigott [Tue, 25 Aug 2020 02:33:37 +0000 (19:33 -0700)]
zsh: correct journalctl command completion parsing

(cherry picked from commit 45b156c1559da468f1c12aa5170858574c9b5831)

5 years agobasic/missing_syscall: fix syscall numbers for arm64 :(
Zbigniew Jędrzejewski-Szmek [Sat, 22 Aug 2020 14:55:56 +0000 (16:55 +0200)]
basic/missing_syscall: fix syscall numbers for arm64 :(

(cherry picked from commit b6ce3d2c0152a17210bb7fd31bb92a289f181a57)

5 years agoshared/install: fix preset operations for non-service instantiated units
Zbigniew Jędrzejewski-Szmek [Sat, 22 Aug 2020 09:58:15 +0000 (11:58 +0200)]
shared/install: fix preset operations for non-service instantiated units

Fixes https://github.com/coreos/ignition/issues/1064.

(cherry picked from commit 47ab95fe4315b3f7ee5a3694460a744bb88c52fd)

5 years agonss-resolve: treat BUS_ERROR_NO_SUCH_UNIT the same as SD_BUS_ERROR_SERVICE_UNKNOWN too
Zbigniew Jędrzejewski-Szmek [Sat, 22 Aug 2020 16:48:43 +0000 (18:48 +0200)]
nss-resolve: treat BUS_ERROR_NO_SUCH_UNIT the same as SD_BUS_ERROR_SERVICE_UNKNOWN too

Seems safer to do so.

(cherry picked from commit 8e34f4cc62aaeaa8881a1d6cd9136de5a22777a1)

5 years agovarious: treat BUS_ERROR_NO_SUCH_UNIT the same as SD_BUS_ERROR_SERVICE_UNKNOWN
Zbigniew Jędrzejewski-Szmek [Sat, 22 Aug 2020 16:48:03 +0000 (18:48 +0200)]
various: treat BUS_ERROR_NO_SUCH_UNIT the same as SD_BUS_ERROR_SERVICE_UNKNOWN

We return BUS_ERROR_NO_SUCH_UNIT a.k.a. org.freedesktop.systemd1.NoSuchUnit
in various places. In #16813:
Aug 22 06:14:48 core sudo[2769199]: pam_systemd_home(sudo:account): Failed to query user record: Unit dbus-org.freedesktop.home1.service not found.
Aug 22 06:14:48 core dbus-daemon[5311]: [system] Activation via systemd failed for unit 'dbus-org.freedesktop.home1.service': Unit dbus-org.freedesktop.home1.service not found.
Aug 22 06:14:48 core dbus-daemon[5311]: [system] Activating via systemd: service name='org.freedesktop.home1' unit='dbus-org.freedesktop.home1.service' requested by ':1.6564' (uid=0 pid=2769199 comm="sudo su ")

This particular error comes from bus_unit_validate_load_state() in pid1:
  case UNIT_NOT_FOUND:
       return sd_bus_error_setf(error, BUS_ERROR_NO_SUCH_UNIT, "Unit %s not found.", u->id);

It seems possible that we should return a different error, but it doesn't really
matter: if we change pid1 to return a different error, we still need to handle
BUS_ERROR_NO_SUCH_UNIT as in this patch to handle pid1 with current code.

(cherry picked from commit 73d3ac8e2440cda3b7f2310f329f0798de6c041c)

5 years agoman: drop reference to long gone .busname unit type
Lennart Poettering [Mon, 24 Aug 2020 17:41:09 +0000 (19:41 +0200)]
man: drop reference to long gone .busname unit type

Seems we missed one occurence.

(cherry picked from commit 68dd195c1b8fdba3fccf166a1c3c2d7ca5a912fe)

5 years agoman: fix a fix of a typo in systemd.service example
Jan Chren [Mon, 24 Aug 2020 14:40:11 +0000 (16:40 +0200)]
man: fix a fix of a typo in systemd.service example

The fix from cb263973acf83de22a86f08fe502a9cbd6c01d2b was made the other way around,
i.e. `SIGKILL` was changed to `SIGUSR1`, but the sentence is about a "termination signal", i.e. `SIGKILL`, not `SIGUSR1`.

(cherry picked from commit be3f62faf5d498aaab4cf6ceb3ca56e0d994ddf5)

5 years agonetwork: can: Fix CAN initialization
Clemens Gruber [Fri, 21 Aug 2020 14:03:23 +0000 (16:03 +0200)]
network: can: Fix CAN initialization

When introducing CAN-FD support, the .can_fd_mode was not initalized
with -1 and due to cm.mask containing the CAN_CTRLMODE_FD bit, it was
not ignored when FDMode was not configured but instead disabled.
The same thing happened when listen-only mode support was introduced.

On chips that do not support these features, this lead to an error:
can0: Failed to configure CAN link: Operation not supported

Fix it by intializing all the CAN related tristate variables
(.can_listen_only, .can_fd_mode and .can_non_iso) to -1.

(cherry picked from commit f594b5feabee38dded4b7ae288fba6c3b6c79a45)

5 years agoman: update autogenerated dbus api lists
Zbigniew Jędrzejewski-Szmek [Sat, 22 Aug 2020 10:25:44 +0000 (12:25 +0200)]
man: update autogenerated dbus api lists

We forgot to do this before the release :(
Relavant commits are:
4e11ddfdd3c1f93721b8ca534e33e16ced32ff06,
0bb007f7a23c41e23481373ded47ee3ddcf8f26b,
a3d19f5d99c44940831a33df8b5bece4aaf749f7,
bf760801804e55b045aed54bf9b1d0b0131be3f2,
4793c31083031e729e6eb17b87b540a3944bba3b.

Suitable for backporting.

(cherry picked from commit 9653108f1135ad5105fbe71ccf6f818aeb7aa5f8)

5 years agoman: fix invalid tag place
Yu Watanabe [Fri, 21 Aug 2020 06:30:05 +0000 (15:30 +0900)]
man: fix invalid tag place

(cherry picked from commit d91b9bbce881c7e0d2ee80208325be59924b0e8f)

5 years agoman: add conditionals to more man pages
Zbigniew Jędrzejewski-Szmek [Thu, 20 Aug 2020 15:41:19 +0000 (17:41 +0200)]
man: add conditionals to more man pages

Fixes #16701.

(cherry picked from commit bb5a34fb3eeef824afe02b679d391853244161b1)

5 years agomeson: add ENABLE_ANALYZE conditional
Zbigniew Jędrzejewski-Szmek [Thu, 20 Aug 2020 15:35:50 +0000 (17:35 +0200)]
meson: add ENABLE_ANALYZE conditional

(cherry picked from commit b3259a6e5f05ebaf5282e624f80ecf04b77a0adb)

5 years agocore: add missing conditions/asserts to unit file parsing
Lennart Poettering [Thu, 20 Aug 2020 11:44:12 +0000 (13:44 +0200)]
core: add missing conditions/asserts to unit file parsing

(cherry picked from commit 4f55a5b0bf1e68e4595120d8ac4b518654355fc3)

5 years agoanalyze: rework condition testing
Lennart Poettering [Thu, 20 Aug 2020 11:43:00 +0000 (13:43 +0200)]
analyze: rework condition testing

Let's drop the private table and just use the generic concepts we have
in place already that make the same information available.

Fixes: #16781
(cherry picked from commit 625a164069aff9efb61dcc5916c572f53c2a7ab0)

5 years agosd-bus: fix error handling on readv()
Lennart Poettering [Thu, 20 Aug 2020 10:59:23 +0000 (12:59 +0200)]
sd-bus: fix error handling on readv()

let's make sure we collect the right error code from errno, otherwise
we'll see EPERM (i.e. error 1) for all errors readv() returns (since it
returns -1 on error), including EAGAIN.

This is definitely backport material.

A fix-up for 3691bcf3c5eebdcca5b4f1c51c745441c57a6cd1.

Fixes: #16699
(cherry picked from commit c1093c34d7d81b5b13cc72d4d1941000813001da)

5 years agouser-runtime-dir: deal gracefully with missing logind properties
Lennart Poettering [Wed, 19 Aug 2020 15:05:44 +0000 (17:05 +0200)]
user-runtime-dir: deal gracefully with missing logind properties

Fixes: #16685
(cherry picked from commit 5d1e68b49489574ef57947c5f1adfc761547eea9)

5 years agoshared/seccomp: do not use ifdef guards around textual syscall names
Zbigniew Jędrzejewski-Szmek [Tue, 18 Aug 2020 14:10:47 +0000 (16:10 +0200)]
shared/seccomp: do not use ifdef guards around textual syscall names

It is possible that we will be running with an upgraded libseccomp, in which
case libseccomp might know the syscall name, even if the number is not known at
the time when systemd is being compiled. The guard only serves to break such
upgrades, by requiring that we also recompile systemd.

For s390-specific syscalls, use a define to exclude them, so that that we don't
try to filter them on other arches.

(cherry picked from commit 6da432fd542af5553742b905a0f87a825a28a399)

5 years agomachine-id-setup: don't use KVM or container manager supplied uuid if in chroot env
Lennart Poettering [Wed, 19 Aug 2020 15:07:41 +0000 (17:07 +0200)]
machine-id-setup: don't use KVM or container manager supplied uuid if in chroot env

Fixes: #16758
(cherry picked from commit 1f894e682cb34d8d713378d01dc0565b7b5a245f)

5 years agoman: Fix typo in systemd-tmpfiles
Phaedrus Leeds [Wed, 19 Aug 2020 16:36:32 +0000 (09:36 -0700)]
man: Fix typo in systemd-tmpfiles

(cherry picked from commit c2892a244c82ee142089f8fb7848547dada4ed32)

5 years agohomework: downgrade chattr failure log message
Lennart Poettering [Mon, 17 Aug 2020 20:07:55 +0000 (22:07 +0200)]
homework: downgrade chattr failure log message

NOCOW is a btrfs-only thing hence don't log louder than necessary if we
don't have it.

(cherry picked from commit 64dc138d1ebbe11eda7872522263a35237916183)

5 years agohomework: explicitly close cryptsetup context, to not keep loopback device busy
Lennart Poettering [Mon, 17 Aug 2020 20:09:04 +0000 (22:09 +0200)]
homework: explicitly close cryptsetup context, to not keep loopback device busy

The cryptsetup context pins the loop device even after deactivation.
Let's explicitly release the context to make sure the subsequent
loopback device detaching works cleanly.

(cherry picked from commit df14bda2b5de8ea022a915774c5a4e4660e421b6)

5 years agohomework: correct error passed into log message
Lennart Poettering [Mon, 17 Aug 2020 18:20:47 +0000 (20:20 +0200)]
homework: correct error passed into log message

(cherry picked from commit 6d7b47eb3eb71b37c768e588173f2c8169e953e7)

5 years agohomework: sync everything to disk before we rename LUKS loopback file into place
Lennart Poettering [Mon, 17 Aug 2020 17:54:38 +0000 (19:54 +0200)]
homework: sync everything to disk before we rename LUKS loopback file into place

This how this works on Linux: when atomically creating a file we need to
fully populate it under a temporary name and then when we are fully
done, sync it and the directory it is contained in, before renaming it
to the final name.

(cherry picked from commit a2a8a509cd56955c0ce42ba0650d163e0d3eeb82)

5 years agohomed: downgrade quota message in containers
Lennart Poettering [Mon, 17 Aug 2020 16:19:30 +0000 (18:19 +0200)]
homed: downgrade quota message in containers

quota syscalls and operations are typically prohibited in containers.
Let's not make noise about that, needlessly.

(cherry picked from commit 5e5e11b8744fbe9dcd87a2d165f342dc9bcd4008)

5 years agoanalyze-security: do not assign badness to filtered-out syscalls
Zbigniew Jędrzejewski-Szmek [Sat, 1 Aug 2020 09:28:09 +0000 (11:28 +0200)]
analyze-security: do not assign badness to filtered-out syscalls

Fixes #16451, https://bugzilla.redhat.com/show_bug.cgi?id=1856273.

(cherry picked from commit 01ecb3674ad3650bcbb14155b2dcbd4b9f4ed57e)

5 years agoNEWS: clarify two points
Zbigniew Jędrzejewski-Szmek [Fri, 31 Jul 2020 13:36:07 +0000 (15:36 +0200)]
NEWS: clarify two points

I was reading a summary of changes on Phoronix, and (while not incorrect)
those two points were rather misleading.

(cherry picked from commit c2cfb12641eb6d47c66ae14d916f0df86de5d8f8)

5 years agomeson: add min version for libfdisk
Anita Zhang [Tue, 18 Aug 2020 06:09:38 +0000 (23:09 -0700)]
meson: add min version for libfdisk

Was trying to run src/partition/test-repart.sh on CentOS 8 and the first
resize call kept failing with ERANGE. Turned out that CentOS 8 comes
with libfdisk-devel-2.32.1 which is missing
https://github.com/karelzak/util-linux/commit/2f35c1ead621f42f32f7777232568cb03185b473
(in libfdisk 2.33 and up).

(cherry picked from commit e71f5585b9b0580428f9530d0a485265c9c25165)

5 years agoload-fragment: fix grammar in error messages
Alyssa Ross [Tue, 18 Aug 2020 20:56:59 +0000 (20:56 +0000)]
load-fragment: fix grammar in error messages

(cherry picked from commit 556a7bbed607ec8cbbf4affc5d862ef92403418e)

5 years agoFix function description in logind man page
Nicolas Fella [Tue, 18 Aug 2020 20:36:47 +0000 (22:36 +0200)]
Fix function description in logind man page

Fix a presumed copy&paste error. SetLockedHint corresponds to the "locked hint", not the "idle hint".

(cherry picked from commit 07b4f4496868578e56932a292de0f0b1881ecc84)

5 years agonetwork: do not fail if UseMTU=yes on DHCP lease lost
Yu Watanabe [Tue, 18 Aug 2020 10:21:25 +0000 (19:21 +0900)]
network: do not fail if UseMTU=yes on DHCP lease lost

This fixes a bug introduced by 6906794dd1698954b349ec3ec84f1494ecd63cd3.

Fixes #16768.

(cherry picked from commit 46b875fb803ce8ee9bcb6a23ad6339db4f735d3a)

5 years agomissing_syscall: do not use function name that may conflict with glibc
Zbigniew Jędrzejewski-Szmek [Sun, 16 Aug 2020 14:19:29 +0000 (16:19 +0200)]
missing_syscall: do not use function name that may conflict with glibc

This was done for all replacements back in 5187dd2c403caf92d09f3491e41f1ceb3f10491f,
but some newer stuff didn't do this.

(cherry picked from commit faeae444ea452139519718ceb681e8b5831b6890)

5 years agomissing_syscall: fix pidfd_{send_signal,open} numbers for alpha
Zbigniew Jędrzejewski-Szmek [Sat, 15 Aug 2020 17:01:23 +0000 (19:01 +0200)]
missing_syscall: fix pidfd_{send_signal,open} numbers for alpha

Also order the syscalls by syscall number for easier comparisons with the
kernel headers.

Fixup for 5f152f43d04e5aad6a3f98f45f020a66e3aac717.

(cherry picked from commit 23654cee136862996d92e7f1e9887786ddb3dfe6)

5 years agonetwork: wait for previous address removal before configuring static addresses
Yu Watanabe [Sat, 8 Aug 2020 03:36:18 +0000 (12:36 +0900)]
network: wait for previous address removal before configuring static addresses

Fixes #16696.

(cherry picked from commit 9b966cee435fecfdee79f3e0aff0dec0a4e68f14)

5 years agonetwork: only process non-error message
Yu Watanabe [Sat, 8 Aug 2020 03:31:12 +0000 (12:31 +0900)]
network: only process non-error message

(cherry picked from commit fb282d4e256a79b19d3299999024e8fafd2ab0a0)

5 years agotest: accept that char device 0/0 can now be created witout privileges
Lennart Poettering [Fri, 14 Aug 2020 19:50:55 +0000 (21:50 +0200)]
test: accept that char device 0/0 can now be created witout privileges

Fixes: #16721
(cherry picked from commit 5b5ce6298e5a1c09beacd5c963e2350979cbf94a)

5 years agotools/make-man-index: fix purpose text that contains tags v246.2
Haochen Tong [Sun, 16 Aug 2020 01:28:46 +0000 (03:28 +0200)]
tools/make-man-index: fix purpose text that contains tags

(cherry picked from commit f3317336450e1145b97ae9e38bd626f3d4c88eb8)

5 years agotest-fs-util: skip encrypted path test if we get EACCES
Topi Miettinen [Mon, 17 Aug 2020 09:08:57 +0000 (12:08 +0300)]
test-fs-util: skip encrypted path test if we get EACCES

Unprivileged test-fs-util fails on my system since /sys/dev/block is
inaccessible for unprivileged users, so let's skip encrypted path test if we
get EACCES or similar.

(cherry picked from commit 209650b7200115d2cad9081cb97e22608fce41f8)

5 years agoNewer Glibc use faccessat2 to implement faccessat
Michael Scherer [Sat, 15 Aug 2020 16:12:02 +0000 (18:12 +0200)]
Newer Glibc use faccessat2 to implement faccessat

cf https://repo.or.cz/glibc.git/commit/3d3ab573a5f3071992cbc4f57d50d1d29d55bde2

This cause breakage on Fedora Rawhide: https://bugzilla.redhat.com/show_bug.cgi?id=1869030

(cherry picked from commit bcf08acbffdee0d6360d3c31d268e73d0623e5dc)

5 years agonamespace: fix minor memory leak
Lennart Poettering [Fri, 14 Aug 2020 09:07:18 +0000 (11:07 +0200)]
namespace: fix minor memory leak

(cherry picked from commit 3f181262f4aa746177ea311827c268652bdb1c33)

5 years agoman: fix incorrectly placed full stop
Lennart Poettering [Tue, 11 Aug 2020 12:50:32 +0000 (14:50 +0200)]
man: fix incorrectly placed full stop

(cherry picked from commit 6b49257f6b92c1bcdf02ca0e896009da36ed9bb0)

5 years agoman: fix typo
Zbigniew Jędrzejewski-Szmek [Wed, 5 Aug 2020 17:44:53 +0000 (19:44 +0200)]
man: fix typo

(cherry picked from commit d08a529518c37734e4e77540844373249589fb78)

5 years agobless-boot: add missing verb to --help
Zbigniew Jędrzejewski-Szmek [Wed, 5 Aug 2020 10:16:03 +0000 (12:16 +0200)]
bless-boot: add missing verb to --help

There is also "help" verb which duplicates the --help option. I don't
think we should advertise it.

(cherry picked from commit ddd8e23dc9d6ab33b3b3c243369042a325ba618d)

5 years agouser-record: deal with invalid GECOS fields gracefully
Lennart Poettering [Thu, 6 Aug 2020 15:00:07 +0000 (17:00 +0200)]
user-record: deal with invalid GECOS fields gracefully

Let's fix up invalid GECOS fields both when we convert from NSS to JSON
and the other way round.

Kinda sucks we have to do that, but NSS does it when writing data to
/etc/passwd, so let's do the same.

Fixes: #16668
(cherry picked from commit 5cd12abaa0c0f3a06c9ff2048941fbe6e8b3577e)

5 years agouser-util: add mangle_gecos() call for turning strings into fields suitable as GECOS...
Lennart Poettering [Thu, 6 Aug 2020 14:46:18 +0000 (16:46 +0200)]
user-util: add mangle_gecos() call for turning strings into fields suitable as GECOS fields

(cherry picked from commit b10fd796f56e4f16f7430cd22f59f544766d3bef)

5 years agofix typo in systemctl help v246.1
Matt Fenwick [Fri, 7 Aug 2020 14:43:11 +0000 (10:43 -0400)]
fix typo in systemctl help

(cherry picked from commit c1c28fe2f70573270b0544670dba38b3a2f06c13)

5 years agoman: clarify that LogNamespace= is for system services only
Lennart Poettering [Thu, 6 Aug 2020 13:27:00 +0000 (15:27 +0200)]
man: clarify that LogNamespace= is for system services only

Fixes: #16638
(cherry picked from commit a6991726f80c299ac7275f4570e310e1dd5bce96)

5 years ago_sd-common.h: avoid parsing errors with Coverity
Kamil Dudka [Wed, 5 Aug 2020 21:53:40 +0000 (23:53 +0200)]
_sd-common.h: avoid parsing errors with Coverity

The commit 1070d271fa8fa553d57dd5f74dd1e3f60732d0b9 which was supposed
too fix this does not seem to take effect any more.  We get again 34%
compilation success rate while scanning systemd itself.  Moreover, the
installed header file breaks compilation of programs that include it:

"/usr/include/systemd/_sd-common.h", line 23: error #35: #error directive: "Do
          not include _sd-common.h directly; it is a private header."
  #  error "Do not include _sd-common.h directly; it is a private header."
     ^

(cherry picked from commit 4191b3282afbca9f1ef333f91bb6566c374da1fe)

5 years agoanalyze: fix 'cat-config systemd/zram-generator.conf'
Zbigniew Jędrzejewski-Szmek [Sat, 1 Aug 2020 10:25:56 +0000 (12:25 +0200)]
analyze: fix 'cat-config systemd/zram-generator.conf'

Also makes this work for various systemd config files that support .d.

(cherry picked from commit 28365e88d0dfc6332b26120e23b2c39b7958f907)

5 years agoman: describe that changing Storage= does not move existing data
Zbigniew Jędrzejewski-Szmek [Mon, 3 Aug 2020 11:22:01 +0000 (13:22 +0200)]
man: describe that changing Storage= does not move existing data

Fixes #16384.

(cherry picked from commit f254abcd72b6ca6c4567c245a7dbaf739d02dfc2)

5 years agocore: reset bus error before reuse
Zbigniew Jędrzejewski-Szmek [Mon, 3 Aug 2020 09:39:25 +0000 (11:39 +0200)]
core: reset bus error before reuse

From a report in https://bugzilla.redhat.com/show_bug.cgi?id=1861463:
usb-gadget.target: Failed to load configuration: No such file or directory
usb-gadget.target: Failed to load configuration: No such file or directory
usb-gadget.target: Trying to enqueue job usb-gadget.target/start/fail
usb-gadget.target: Failed to load configuration: No such file or directory
Assertion '!bus_error_is_dirty(e)' failed at src/libsystemd/sd-bus/bus-error.c:239, function bus_error_setfv(). Ignoring.
sys-devices-platform-soc-2100000.bus-2184000.usb-ci_hdrc.0-udc-ci_hdrc.0.device: Failed to enqueue SYSTEMD_WANTS= job, ignoring: Unit usb-gadget.target not found.

I *think* this is the place where the reuse occurs: we call
bus_unit_validate_load_state(unit, e) twice in a row.

(cherry picked from commit 94efaa31812f664f0ad659f474db3da0f9c9bacd)

5 years agonspawn: Fix incorrect usage of putenv
Daan De Meyer [Sat, 1 Aug 2020 14:05:01 +0000 (15:05 +0100)]
nspawn: Fix incorrect usage of putenv

strv_env_get only returns the environment variable value. putenv expects
KEY=VALUE format strings. Use setenv instead to fix the use.

(cherry picked from commit 6f646e01755df587bb33dae4ca78cdaad5721f5e)

5 years agoman: fix typo in systemd.service
Etienne Doms [Sat, 1 Aug 2020 10:47:17 +0000 (12:47 +0200)]
man: fix typo in systemd.service

5 years agonetwork: fix DHCPv6 Prefix Delegation example after option rename
Florian Klink [Fri, 31 Jul 2020 09:35:15 +0000 (11:35 +0200)]
network: fix DHCPv6 Prefix Delegation example after option rename

99e015e28c8322bf714f45cd1defcf20ac2103c5 missed to update the example
below - DHCPv6.AssignAcquiredDelegatedPrefixAddress was moved to
DHCPv6PrefixDelegation.Assign.

As it already defaulted to true since it's introduction in
9efa8a3cff9948d3a78597b74dca75c805716de4, there's no need to explicitly
list it at all.

5 years agoNEWS: last rites v246
Zbigniew Jędrzejewski-Szmek [Thu, 30 Jul 2020 18:57:50 +0000 (20:57 +0200)]
NEWS: last rites

5 years agoloop-device: implicitly sync device on detach
Lennart Poettering [Thu, 30 Jul 2020 16:42:13 +0000 (18:42 +0200)]
loop-device: implicitly sync device on detach

Apparently, if IO is still in flight at the moment we invoke LOOP_CLR_FD
it is likely simply dropped (probably because yanking physical storage,
such as a USB stick would drop it too). Let's protect ourselves against
that and always sync explicitly before we invoke it.

5 years agoMerge pull request #16624 from keszybz/timesync-retry-interval
Zbigniew Jędrzejewski-Szmek [Thu, 30 Jul 2020 18:47:41 +0000 (20:47 +0200)]
Merge pull request #16624 from keszybz/timesync-retry-interval

Make timesyncd retry interval grow more slowly

5 years agoudev: fix codesonar warnings
Amitanand.Chikorde [Thu, 30 Jul 2020 13:18:48 +0000 (18:48 +0530)]
udev: fix codesonar warnings

Fixed below systemd codesonar warning.
isprint() is invoked here with an argument of signed
type char, but only has defined behavior for int arguments that are
either representable as unsigned char or equal to the value
of macro EOF(-1).

As per codesonar report, in a number of libc implementations, isprint()
function implemented using lookup tables (arrays): passing in a
negative value can result in a read underrun.

5 years agopid1: stop limiting size of /dev/shm
Zbigniew Jędrzejewski-Szmek [Thu, 30 Jul 2020 11:36:10 +0000 (13:36 +0200)]
pid1: stop limiting size of /dev/shm

The explicit limit is dropped, which means that we return to the kernel default
of 50% of RAM. See 362a55fc14 for a discussion why that is not as much as it
seems. It turns out various applications need more space in /dev/shm and we
would break them by imposing a low limit.

While at it, rename the define and use a single macro for various tmpfs mounts.
We don't really care what the purpose of the given tmpfs is, so it seems
reasonable to use a single macro.

This effectively reverts part of 7d85383edbab7. Fixes #16617.

5 years agopath-util: make use of TAKE_PTR() where we can
Lennart Poettering [Wed, 29 Jul 2020 17:09:48 +0000 (19:09 +0200)]
path-util: make use of TAKE_PTR() where we can

5 years agohwdb: Add EliteBook to use micmute hotkey
Kai-Heng Feng [Tue, 16 Jun 2020 05:24:27 +0000 (13:24 +0800)]
hwdb: Add EliteBook to use micmute hotkey

Like HP ZBooks, all EliteBooks use the same micmute scancode.

5 years agorepart: log fixes
Lennart Poettering [Wed, 29 Jul 2020 17:20:39 +0000 (19:20 +0200)]
repart: log fixes

5 years agotimesync: increase retry interval more slowly
Zbigniew Jędrzejewski-Szmek [Thu, 30 Jul 2020 13:15:11 +0000 (15:15 +0200)]
timesync: increase retry interval more slowly

The new retry intervals are [15, 20, 26, 34, 45, 60, 80, 106, 141, 188, 250,
333, 360, ...]. This should allow graceful response if a transient network
failure is encountered. Growth is exponential, but with a small power and
capped to a non-too-large value so that we resynchronize within a few minutes
after network is restored. I made the minimum 15 s to make sure that we never
send packets more often than that.

Fixes #16492.

5 years agoadded alternative version of SurfTab Twin
Benjamin Dahlhoff [Wed, 29 Jul 2020 18:02:55 +0000 (20:02 +0200)]
added alternative version of SurfTab Twin

5 years agonetwork: fix use of uninitialized value
Yu Watanabe [Thu, 30 Jul 2020 03:45:36 +0000 (12:45 +0900)]
network: fix use of uninitialized value

Follow-up for 6e537f62d71edea5bfcd2e2675c35ee840ffdc1f.

5 years agonetwork: Fix "Unknown section 'DHCPv6PrefixDelegation'." message
Michael Marley [Wed, 29 Jul 2020 19:13:35 +0000 (15:13 -0400)]
network: Fix "Unknown section 'DHCPv6PrefixDelegation'." message

Follow-up for https://github.com/systemd/systemd/commit/99e015e28c8322bf714f45cd1defcf20ac2103c5

5 years agoNEWS: mention about [DHCPv6PrefixDelegation] section
Yu Watanabe [Wed, 29 Jul 2020 16:02:52 +0000 (01:02 +0900)]
NEWS: mention about [DHCPv6PrefixDelegation] section

Follow-up for 99e015e28c8322bf714f45cd1defcf20ac2103c5.

5 years agomeson: run 'ninja man/update-man-rules' again
Lennart Poettering [Wed, 29 Jul 2020 13:50:22 +0000 (15:50 +0200)]
meson: run 'ninja man/update-man-rules' again

5 years agoman: fix link markup
Lennart Poettering [Wed, 29 Jul 2020 15:37:33 +0000 (17:37 +0200)]
man: fix link markup

5 years agoMerge pull request #16585 from yuwata/network-dhcp6-fixes
Lennart Poettering [Wed, 29 Jul 2020 13:59:11 +0000 (15:59 +0200)]
Merge pull request #16585 from yuwata/network-dhcp6-fixes

network: fix several DHCP6 PD issues, and cleanup DHCP4 and NDISC

5 years agoMerge pull request #16590 from keszybz/test-fs-util-relax
Lennart Poettering [Wed, 29 Jul 2020 13:54:51 +0000 (15:54 +0200)]
Merge pull request #16590 from keszybz/test-fs-util-relax

Small adjustements to let various tests pass in overloaded Fedora infra and related changes

5 years agoMerge pull request #16576 from keszybz/bump-tmp-size
Lennart Poettering [Wed, 29 Jul 2020 13:53:30 +0000 (15:53 +0200)]
Merge pull request #16576 from keszybz/bump-tmp-size

Increase /tmp size to 50% of RAM

5 years agotest-cgroup-util: Handle result=NULL as empty string
Gaurav Singh [Tue, 28 Jul 2020 02:28:56 +0000 (22:28 -0400)]
test-cgroup-util: Handle result=NULL as empty string

Signed-off-by: Gaurav Singh <gaurav1086@gmail.com>
5 years agofirstboot: Update help string with --root-shell options
Daan De Meyer [Wed, 29 Jul 2020 07:21:01 +0000 (08:21 +0100)]
firstboot: Update help string with --root-shell options

5 years agotest-sleep: add more logging, show secure boot mode
Zbigniew Jędrzejewski-Szmek [Mon, 27 Jul 2020 15:36:56 +0000 (17:36 +0200)]
test-sleep: add more logging, show secure boot mode

In https://bugzilla.redhat.com/show_bug.cgi?id=1858219 a user
reported that hibernation is not supported. This is probably caused
by secure boot, but our current logging does not make this
obvious.

5 years agoNEWS: mention new tmpfs limits
Zbigniew Jędrzejewski-Szmek [Fri, 24 Jul 2020 20:24:59 +0000 (22:24 +0200)]
NEWS: mention new tmpfs limits

5 years agoBump /tmp size back to 50% of RAM
Zbigniew Jędrzejewski-Szmek [Fri, 24 Jul 2020 20:05:21 +0000 (22:05 +0200)]
Bump /tmp size back to 50% of RAM

This should be enough to fix https://bugzilla.redhat.com/show_bug.cgi?id=1856514.
But the limit should be significantly higher than 10% anyway. By setting a
limit on /tmp at 10% we'll break many reasonable use cases, even though the
machine would deal fine with a much larger fraction devoted to /tmp.
(In the first version of this patch I made it 25% with the comment that
"Even 25% might be too low.". The kernel default is 50%, and we have been using
that seemingly without trouble since https://fedoraproject.org/wiki/Features/tmp-on-tmpfs.
So let's just make it 50% again.)

See 7d85383edbab73274dc81cc888d884bb01070bc2.

(Another consideration is that we learned from from the whole initiative with
zram in Fedora that a reasonable size for zram is 0.5-1.5 of RAM, and that pretty
much all systems benefit from having zram or zswap enabled. Thus it is reasonable
to assume that it'll become widely used. Taking the usual compression effectiveness
of 0.2 into account, machines have effective memory available of between
1.0 - 0.2*0.5 + 0.5 = 1.4 (for zram sized to 0.5 of RAM) and
1.0 - 0.2*1.5 + 1.5 = 2.2 (for zram 1.5 sized to 1.5 of RAM) times RAM size.
This means that the 10% was really like 7-4% of effective memory.)

A comment is added to mount-util.h to clarify that tmp.mount is separate.

5 years agonetwork: rename settings about DHCPv6 Prefix Delegation
Yu Watanabe [Tue, 28 Jul 2020 20:50:04 +0000 (05:50 +0900)]
network: rename settings about DHCPv6 Prefix Delegation

Closes #16602.

5 years agonetwork: make RADVPrefixDelegation enum bitfield
Yu Watanabe [Tue, 28 Jul 2020 20:09:38 +0000 (05:09 +0900)]
network: make RADVPrefixDelegation enum bitfield

This should not change any behavior.

5 years agonetwork: update address infomation even if link is in failed or linger state
Yu Watanabe [Thu, 23 Jul 2020 22:18:39 +0000 (07:18 +0900)]
network: update address infomation even if link is in failed or linger state

As the link may be reconfigured later. If we do not update the address
information, then its setup state or operstate may not be updated
correctly.

5 years agonetwork: do not assume static addresses are configured
Yu Watanabe [Thu, 23 Jul 2020 15:52:32 +0000 (00:52 +0900)]
network: do not assume static addresses are configured

link_request_set_routes() is also called when a dynamic address is
configured. At that time, static addresses may not be configured yet.

Fixes #16546.

5 years agonetwork: check at least one dynamic address is assigned when DHCP is enabled
Yu Watanabe [Thu, 23 Jul 2020 12:18:24 +0000 (21:18 +0900)]
network: check at least one dynamic address is assigned when DHCP is enabled

5 years agonetwork: dhcp6: drop addresses and delegated prefixes on client stop
Yu Watanabe [Wed, 22 Jul 2020 18:13:42 +0000 (03:13 +0900)]
network: dhcp6: drop addresses and delegated prefixes on client stop

Previously, we did not drop addresses and delegated prefixes when
DHCP6 client is stopped.

Fixes #15455.
Fixes #13564.

5 years agonetwork: rename ipv4ll_address -> ipv4ll_address_configured
Yu Watanabe [Wed, 22 Jul 2020 16:21:40 +0000 (01:21 +0900)]
network: rename ipv4ll_address -> ipv4ll_address_configured

5 years agonetwork: ndisc: also remove old DNSSL or RDNSS records after an SLAAC address is...
Yu Watanabe [Wed, 22 Jul 2020 12:44:44 +0000 (21:44 +0900)]
network: ndisc: also remove old DNSSL or RDNSS records after an SLAAC address is ready

5 years agonetwork: ndisc: remove old addresses and routes after at least one SLAAC address...
Yu Watanabe [Wed, 22 Jul 2020 02:55:07 +0000 (11:55 +0900)]
network: ndisc: remove old addresses and routes after at least one SLAAC address becomes ready

Otherwise, the old addresses will exist in deperecated state.

5 years agonetwork: dhcp4: release old lease after the new address become ready
Yu Watanabe [Tue, 21 Jul 2020 14:06:51 +0000 (23:06 +0900)]
network: dhcp4: release old lease after the new address become ready

Previously, on DHCPv4 address renewal, the old address may be removed
while the new address is not ready yet.

This also simplifies the logic of removing address and routes.

5 years agonetwork: make address/route_configure optionally return created Address/Route object
Yu Watanabe [Wed, 22 Jul 2020 00:41:30 +0000 (09:41 +0900)]
network: make address/route_configure optionally return created Address/Route object

5 years agonetwork: set key destructor in several hash_ops
Yu Watanabe [Tue, 21 Jul 2020 23:22:55 +0000 (08:22 +0900)]
network: set key destructor in several hash_ops

5 years agonetwork: make link_check_ready() return earlier if the link is not in 'configuring...
Yu Watanabe [Tue, 21 Jul 2020 21:07:42 +0000 (06:07 +0900)]
network: make link_check_ready() return earlier if the link is not in 'configuring' state

5 years agonetwork: ndisc: do not set configured flags when addresses or routes are not assigned yet
Yu Watanabe [Tue, 21 Jul 2020 21:03:10 +0000 (06:03 +0900)]
network: ndisc: do not set configured flags when addresses or routes are not assigned yet

Just for safety.

5 years agonetwork: make link enter failed state if address_update() failed
Yu Watanabe [Thu, 23 Jul 2020 18:16:42 +0000 (03:16 +0900)]
network: make link enter failed state if address_update() failed

5 years agonetwork: introduce callback called when an address becomes ready
Yu Watanabe [Tue, 21 Jul 2020 14:02:35 +0000 (23:02 +0900)]
network: introduce callback called when an address becomes ready