platform/upstream/systemd.git
5 years agocore: remove unused prototypes
Zbigniew Jędrzejewski-Szmek [Tue, 1 Oct 2019 12:25:10 +0000 (14:25 +0200)]
core: remove unused prototypes

5 years agocore: minor formatting adjustment
Zbigniew Jędrzejewski-Szmek [Tue, 1 Oct 2019 12:13:35 +0000 (14:13 +0200)]
core: minor formatting adjustment

5 years agocryptsetup: bump minimum libcryptsetup version to v2.0.1
Jay Strict [Thu, 26 Sep 2019 13:54:29 +0000 (15:54 +0200)]
cryptsetup: bump minimum libcryptsetup version to v2.0.1

libcryptsetup v2.0.1 introduced new API calls, supporting 64 bit wide
integers for `keyfile_offset`. This change invokes the new function
call, gets rid of the warning that was added in #7689, and removes
redundant #ifdefery and constant definitions.
See https://gitlab.com/cryptsetup/cryptsetup/issues/359.

Fixes #7677.

5 years agoMerge pull request #13689 from cdown/default_memory_min
Chris Down [Tue, 1 Oct 2019 02:08:15 +0000 (03:08 +0100)]
Merge pull request #13689 from cdown/default_memory_min

cgroup: A bunch of protection-related fixes

5 years agocgroup: Mark memory protections as explicitly set in transient units
Chris Down [Mon, 30 Sep 2019 17:36:13 +0000 (18:36 +0100)]
cgroup: Mark memory protections as explicitly set in transient units

A later version of the DefaultMemory{Low,Min} patch changed these to
require explicitly setting memory_foo_set, but we only set that in
load-fragment, not dbus-cgroup.

Without these, we may fall back to either DefaultMemoryFoo or
CGROUP_LIMIT_MIN when we really shouldn't.

5 years agoMerge pull request #13687 from cdown/cgroupv2_docs
Anita Zhang [Mon, 30 Sep 2019 21:00:38 +0000 (14:00 -0700)]
Merge pull request #13687 from cdown/cgroupv2_docs

cgroup: minor doc improvements

5 years agocgroup: Respect DefaultMemoryMin when setting memory.min
Chris Down [Mon, 30 Sep 2019 17:25:09 +0000 (18:25 +0100)]
cgroup: Respect DefaultMemoryMin when setting memory.min

This is an oversight from https://github.com/systemd/systemd/pull/12332.

Sadly the tests didn't catch it since it requires a real cgroup
hierarchy to see, and it wasn't seen in prod since we're only currently
using DefaultMemoryLow, not DefaultMemoryMin. :-(

5 years agocgroup: Check ancestor memory min for unified memory config
Chris Down [Mon, 30 Sep 2019 17:24:26 +0000 (18:24 +0100)]
cgroup: Check ancestor memory min for unified memory config

Otherwise we might not enable it when we should, ie. DefaultMemoryMin is
set in a parent, but not MemoryMin in the current unit.

5 years agoConsider smb3 as remote filesystem
Kenneth D'souza [Mon, 30 Sep 2019 10:31:21 +0000 (16:01 +0530)]
Consider smb3 as remote filesystem

Currently systemd will treat smb3 as local filesystem and cause
can't boot failures. Add smb3 to the list of remote filesystems
to fix this issue.

Signed-off-by: Kenneth D'souza <kdsouza@redhat.com>
5 years agocgroup: docs: memory.high doc fixups
Chris Down [Mon, 30 Sep 2019 13:25:14 +0000 (14:25 +0100)]
cgroup: docs: memory.high doc fixups

The docs just tautologically call this the "high limit". Just call it
throttling as we do in cgroup-v2.rst.

5 years agocgroup: docs: Mention unbounded protection for memory.{low,min}
Chris Down [Mon, 30 Sep 2019 13:23:32 +0000 (14:23 +0100)]
cgroup: docs: Mention unbounded protection for memory.{low,min}

I got asked why Memory{Low,Min} don't allow "infinity". They do, but the
docs don't say that like they already do for Memory{High,Max}.

5 years agoHandle d_type == DT_UNKNOWN correctly
Michael Olbrich [Sun, 29 Sep 2019 11:43:00 +0000 (13:43 +0200)]
Handle d_type == DT_UNKNOWN correctly

As documented in the man-page, readdir() may return a directory entry with
d_type == DT_UNKNOWN. This must be handled for regular filesystems.

dirent_ensure_type() is available to set d_type if necessary. Use it in
some more places.

Without this systemd will fail to boot correctly with nfsroot and some
other filesystems.

Closes #13609

5 years agotest: allow use of nested KVM
Frantisek Sumsal [Thu, 26 Sep 2019 21:29:38 +0000 (23:29 +0200)]
test: allow use of nested KVM

Introduce TEST_NESTED_KVM variable, which allows use of nested KVM for
QEMU VMs (which is disabled by default by KVM autodetection).

5 years agoMerge pull request #13651 from cdown/find_binary_2
Anita Zhang [Fri, 27 Sep 2019 16:39:10 +0000 (09:39 -0700)]
Merge pull request #13651 from cdown/find_binary_2

util-lib: Don't propagate EACCES from find_binary PATH lookup to caller

5 years agoshared: fix a copy-paste error in ethtool-util
Frantisek Sumsal [Thu, 26 Sep 2019 13:13:21 +0000 (15:13 +0200)]
shared: fix a copy-paste error in ethtool-util

CID#1405818

5 years agoutil-lib: Don't propagate EACCES from find_binary PATH lookup to caller
Chris Down [Wed, 25 Sep 2019 16:09:38 +0000 (17:09 +0100)]
util-lib: Don't propagate EACCES from find_binary PATH lookup to caller

On one of my test machines, test-path-util was failing because the
find_binary("xxxx-xxxx") was returning -EACCES instead of -ENOENT. This
happens because the PATH entry on that host contains a directory which
the user in question doesn't have access to. Typically applications
ignore permission errors when searching through PATH, for example in
bash:

    $ whoami
    cdown
    $ PATH=/root:/bin type sh
    sh is /bin/sh

This behaviour is present on zsh and other shells as well, though. This
patch brings our PATH search behaviour closer to other major Unix tools.

5 years agodhcp4: make IPServiceType configurable
Siddharth Chandrasekara [Mon, 23 Sep 2019 11:25:21 +0000 (04:25 -0700)]
dhcp4: make IPServiceType configurable

IPServiceType set to CS6 (network control) causes problems on some old
network setups that continue to interpret the field as IP TOS.

Make DHCP work on such networks by allowing this field to be set to
CS4 (Realtime) instead, as this maps to IPTOS_LOWDELAY.

Signed-off-by: Siddharth Chandrasekaran <csiddharth@vmware.com>
5 years agonetwork: drop noisy log message
Yu Watanabe [Wed, 25 Sep 2019 15:21:00 +0000 (00:21 +0900)]
network: drop noisy log message

Closes #13595.

5 years agoMerge pull request #13649 from keszybz/arphrd-minimization
Yu Watanabe [Wed, 25 Sep 2019 15:12:20 +0000 (00:12 +0900)]
Merge pull request #13649 from keszybz/arphrd-minimization

Arphrd minimization

5 years agoUpdated log message when the timesync happens for the first time (#13624)
cbzxt [Wed, 25 Sep 2019 15:02:46 +0000 (20:32 +0530)]
Updated log message when the timesync happens for the first time (#13624)

5 years agounits: make systemd-binfmt.service easier to work with no autofs
Zbigniew Jędrzejewski-Szmek [Wed, 18 Sep 2019 10:01:41 +0000 (12:01 +0200)]
units: make systemd-binfmt.service easier to work with no autofs

See https://bugzilla.redhat.com/show_bug.cgi?id=1731772:
when autofs4 is disabled in the kernel,
proc-sys-fs-binfmt_misc.automount is not started, so the binfmt_misc module is
never loaded. If we added a dependency on proc-sys-fs-binfmt_misc.mount
to systemd-binfmt.service, things would work even if autofs4 was disabled, but
we would unconditionally pull in the module and mount, which we don't want to do.
(Right now we ony load the module if some binfmt is configured.)
But let's make it easier to handle this case by doing two changes:
1. order systemd-binfmt.service after the .mount unit (so that the .service
   can count on the mount if both units are pulled in, even if .automount
   is skipped)
2. add [Install] section to the service unit. This way the user can do
   'systemctl enable proc-sys-fs-binfmt_misc.mount' to get the appropriate behaviour.

5 years agobasic/arphrd: stop discriminating against NETROM and CISCO
Zbigniew Jędrzejewski-Szmek [Wed, 25 Sep 2019 10:14:07 +0000 (12:14 +0200)]
basic/arphrd: stop discriminating against NETROM and CISCO

ARPHRD_NETROM was excluded, most likely just because it is protocol No. 0,
and ARPHRD_CISCO was reported under its alias name "HDLC". Let's just
allow defined aliases under the main name.

5 years agobasic: massively reduce the size of arphdr lookup functions
Zbigniew Jędrzejewski-Szmek [Wed, 25 Sep 2019 10:06:05 +0000 (12:06 +0200)]
basic: massively reduce the size of arphdr lookup functions

Our biggest object in libsystemd was a table full of zeros, for the arphdr
names. Let's use a switch (which gcc nicely optimizes for us), instead a
table with a gap between 826 and 65534:

$ ls -l build{,2}/src/basic/a6ba3eb@@basic@sta/arphrd-list.c.o
-rw-rw-r--. 1 zbyszek zbyszek 540232 Sep 22 00:29 build/src/basic/a6ba3eb\@\@basic\@sta/arphrd-list.c.o
-rw-rw-r--. 1 zbyszek zbyszek  20512 Sep 25 11:56 build2/src/basic/a6ba3eb\@\@basic\@sta/arphrd-list.c.o

$ ls -l build{,2}/src/shared/libsystemd-shared-243.so
-rwxrwxr-x. 1 zbyszek zbyszek 6774368 Sep 22 00:29 build/src/shared/libsystemd-shared-243.so
-rwxrwxr-x. 1 zbyszek zbyszek 6254808 Sep 25 12:16 build2/src/shared/libsystemd-shared-243.so

No functional change.

5 years agoMerge pull request #13648 from evverx/unbreak-semaphore
Chris Down [Wed, 25 Sep 2019 09:41:03 +0000 (11:41 +0200)]
Merge pull request #13648 from evverx/unbreak-semaphore

semaphore: turn on systemd-networkd in lxc-containers on Semaphore

5 years agotest-arphdr-list: modernize and provide some useful output
Zbigniew Jędrzejewski-Szmek [Wed, 25 Sep 2019 09:39:10 +0000 (11:39 +0200)]
test-arphdr-list: modernize and provide some useful output

5 years agoMerge pull request #13636 from traylenator/manpage
Chris Down [Wed, 25 Sep 2019 09:32:20 +0000 (11:32 +0200)]
Merge pull request #13636 from traylenator/manpage

Corect man page reference in systemd-nologin.conf comments

5 years agoMerge pull request #13641 from pwithnall/machinectl-docs-typo
Chris Down [Wed, 25 Sep 2019 09:24:11 +0000 (11:24 +0200)]
Merge pull request #13641 from pwithnall/machinectl-docs-typo

man: Add a missing space in machinectl(1)

5 years agolog: Add missing "%" in "%m" log format strings
Filipe Brandenburger [Tue, 24 Sep 2019 22:31:27 +0000 (15:31 -0700)]
log: Add missing "%" in "%m" log format strings

These were clearly intended to be "%m" to display the human readable version
of the error stored in errno.

5 years agosemaphore: turn on systemd-networkd in lxc-containers on Semaphore
Evgeny Vereshchagin [Wed, 25 Sep 2019 05:19:12 +0000 (05:19 +0000)]
semaphore: turn on systemd-networkd in lxc-containers on Semaphore

It appears networking.service has broken down on Debian Buster.

5 years agoMerge pull request #13481 from Antique/cgroupv2
Chris Down [Tue, 24 Sep 2019 19:57:06 +0000 (21:57 +0200)]
Merge pull request #13481 from Antique/cgroupv2

cgroup: introduce support for cgroup v2 CPUSET controller

5 years agoman: Add a missing space in machinectl(1)
Philip Withnall [Tue, 24 Sep 2019 15:07:23 +0000 (16:07 +0100)]
man: Add a missing space in machinectl(1)

Signed-off-by: Philip Withnall <withnall@endlessm.com>
5 years agolink: Add support to configure NIC ring buffer size
Susant Sahani [Mon, 23 Sep 2019 14:51:02 +0000 (16:51 +0200)]
link: Add support to configure NIC ring buffer size

5 years agotest-mountpoint-util: improve output a bit
Zbigniew Jędrzejewski-Szmek [Tue, 24 Sep 2019 13:21:47 +0000 (15:21 +0200)]
test-mountpoint-util: improve output a bit

5 years agocgroup-util: fix obsolete comment about supported controllers
Pavel Hrdina [Tue, 10 Sep 2019 07:09:46 +0000 (09:09 +0200)]
cgroup-util: fix obsolete comment about supported controllers

The list might grow so make the comment more generic to not worry about
it if some controller is implemented.

5 years agocgroup: introduce support for cgroup v2 CPUSET controller
Pavel Hrdina [Mon, 29 Jul 2019 15:50:05 +0000 (17:50 +0200)]
cgroup: introduce support for cgroup v2 CPUSET controller

Introduce support for configuring cpus and mems for processes using
cgroup v2 CPUSET controller.  This allows users to limit which cpus
and memory NUMA nodes can be used by processes to better utilize
system resources.

The cgroup v2 interfaces to control it are cpuset.cpus and cpuset.mems
where the requested configuration is written.  However, it doesn't mean
that the requested configuration will be actually used as parent cgroup
may limit the cpus or mems as well.  In order to reflect the real
configuration cgroup v2 provides read-only files cpuset.cpus.effective
and cpuset.mems.effective which are exported to users as well.

5 years agoCorect man page reference in systemd-nologin.conf comments
Steve Traylen [Tue, 24 Sep 2019 07:48:03 +0000 (09:48 +0200)]
Corect man page reference in systemd-nologin.conf comments

The reference to the man page of `systemd-user-sessions.service`
in the comments of `tmpfiles.d/systemd-nologin.conf` is corrected.

5 years agoMerge pull request #13600 from keszybz/ratelimit
Lennart Poettering [Mon, 23 Sep 2019 10:59:09 +0000 (12:59 +0200)]
Merge pull request #13600 from keszybz/ratelimit

Clean up ratelimit functions and disable /dev/kmsg ratelimit

5 years agoMerge pull request #11883 from yuwata/network-dhcp-renew
Yu Watanabe [Sun, 22 Sep 2019 05:44:23 +0000 (07:44 +0200)]
Merge pull request #11883 from yuwata/network-dhcp-renew

network: dhcp renew

5 years agoMerge pull request #13610 from yuwata/network-check-one-more-section
Yu Watanabe [Sun, 22 Sep 2019 05:18:28 +0000 (07:18 +0200)]
Merge pull request #13610 from yuwata/network-check-one-more-section

network: assorted follow-ups

5 years agoman: fix typo
Michael Biebl [Sat, 21 Sep 2019 15:06:25 +0000 (17:06 +0200)]
man: fix typo

5 years agotree-wide: various code-formatting improvements
Frantisek Sumsal [Sat, 21 Sep 2019 14:01:14 +0000 (16:01 +0200)]
tree-wide: various code-formatting improvements

Reported/found by Coccinelle

5 years agonetworkctl: add renew command
Yu Watanabe [Thu, 18 Jul 2019 13:26:49 +0000 (22:26 +0900)]
networkctl: add renew command

5 years agonetwork: introduce new bus method Renew() and RenewLink()
Yu Watanabe [Mon, 4 Mar 2019 08:54:48 +0000 (17:54 +0900)]
network: introduce new bus method Renew() and RenewLink()

5 years agonetwork: add one more section validty check
Yu Watanabe [Thu, 19 Sep 2019 06:20:00 +0000 (15:20 +0900)]
network: add one more section validty check

Follow-up for 203d4df5732b1fdcf50db498ddeb74a934b21f87.

5 years agopo: update Japanese translations
Yu Watanabe [Sat, 21 Sep 2019 11:15:06 +0000 (20:15 +0900)]
po: update Japanese translations

Follow-up for ea2552639b22004c89df26f2374e3d3f1d2acace.

5 years agol10n: update Czech Translation
AsciiWolf [Thu, 19 Sep 2019 13:36:42 +0000 (15:36 +0200)]
l10n: update Czech Translation

5 years agoMerge pull request #13605 from cdown/ono
Chris Down [Fri, 20 Sep 2019 16:10:34 +0000 (18:10 +0200)]
Merge pull request #13605 from cdown/ono

Fixes for recent SIP work

5 years agobasic/log: put a ratelimit on our logging to /dev/kmsg
Zbigniew Jędrzejewski-Szmek [Thu, 19 Sep 2019 16:01:23 +0000 (18:01 +0200)]
basic/log: put a ratelimit on our logging to /dev/kmsg

See the inline comment for details.

5 years agopid1: disable printk ratelimit in early boot
Zbigniew Jędrzejewski-Szmek [Wed, 18 Sep 2019 19:02:07 +0000 (21:02 +0200)]
pid1: disable printk ratelimit in early boot

We have the problem that many early boot or late shutdown issues are harder
to solve than they could be because we have no logs. When journald is not
running, messages are redirected to /dev/kmsg. It is also the time when many
things happen in a rapid succession, so we tend to hit the kernel printk
ratelimit fairly reliably. The end result is that we get no logs from the time
where they would be most useful. Thus let's disable the kernels ratelimit.

Once the system is up and running, the ratelimit is not a problem. But during
normal runtime, things also log to journald, and not to /dev/kmsg, so the
ratelimit is not useful. Hence, there doesn't seem to be much point in trying
to restore the ratelimit after boot is finished and journald is up and running.

See kernel's commit 750afe7babd117daabebf4855da18e4418ea845e for the
description of the kenrel interface. Our setting has lower precedence than
explicit configuration on the kenrel command line.

5 years agotree-wide: say "ratelimit" not "rate_limit"
Zbigniew Jędrzejewski-Szmek [Thu, 19 Sep 2019 15:49:14 +0000 (17:49 +0200)]
tree-wide: say "ratelimit" not "rate_limit"

"ratelimit" is a real word, so we don't need to use the other form anywhere.
We had both forms in various places, let's standarize on the shorter and more
correct one.

5 years agopid1: rename start_limit to start_ratelimit
Zbigniew Jędrzejewski-Szmek [Thu, 19 Sep 2019 15:45:41 +0000 (17:45 +0200)]
pid1: rename start_limit to start_ratelimit

This way it is clearer what the type is. We also have auto_stop_ratelimit adjacent,
and it feels ugly to have a different suffix for those two.

5 years agoDrop RATELIMIT macros
Zbigniew Jędrzejewski-Szmek [Thu, 19 Sep 2019 15:41:20 +0000 (17:41 +0200)]
Drop RATELIMIT macros

Using plain structure initialization is both shorter _and_ more clearer.
We get type safety for free.

5 years agosd-radv: if lifetime < SD_RADV_DEFAULT_MAX_TIMEOUT_USEC, adjust timeout (#13491)
Georg Müller [Fri, 20 Sep 2019 08:23:45 +0000 (10:23 +0200)]
sd-radv: if lifetime < SD_RADV_DEFAULT_MAX_TIMEOUT_USEC, adjust timeout (#13491)

The RFC states that lifetime (AdvDefaultLifetime) must be at least
MaxRtrAdvInterval (which more or less corresponds to SD_RADV_DEFAULT_MAX_TIMEOUT_USEC
in systemd).

To fulfill this limit, virtually lower MaxRtrAdvInterval and MinRtrAdvInterval
accordingly.

Also check that min is not lower than 3s and max is not lower than 4s.

5 years agonetworkd: Fix SIP double-free
Chris Down [Fri, 20 Sep 2019 13:14:05 +0000 (15:14 +0200)]
networkd: Fix SIP double-free

This is already strv_freed right above

5 years agodhcp: Log SIP server information correctly
Chris Down [Fri, 20 Sep 2019 13:13:00 +0000 (15:13 +0200)]
dhcp: Log SIP server information correctly

5 years agonetwork: DHCP server Add support to transmit SIP server
Susant Sahani [Wed, 18 Sep 2019 13:22:47 +0000 (15:22 +0200)]
network: DHCP server Add support to transmit SIP server

1. DHCP server trasmit
2. Client parses and saves in leases
Implements http://www.rfc-editor.org/rfc/rfc3361.txt

```
Frame 134: 348 bytes on wire (2784 bits), 348 bytes captured (2784 bits) on interface 0
Ethernet II, Src: 42:65:85:d6:4e:32 (42:65:85:d6:4e:32), Dst: 1e:04:f8:b8:2f:d4 (1e:04:f8:b8:2f:d4)
Internet Protocol Version 4, Src: 192.168.5.1, Dst: 192.168.5.11
User Datagram Protocol, Src Port: 67, Dst Port: 68
Dynamic Host Configuration Protocol (ACK)
    Message type: Boot Reply (2)
    Hardware type: Ethernet (0x01)
    Hardware address length: 6
    Hops: 0
    Transaction ID: 0x7cc87cb4
    Seconds elapsed: 0
    Bootp flags: 0x0000 (Unicast)
    Client IP address: 0.0.0.0
    Your (client) IP address: 192.168.5.11
    Next server IP address: 0.0.0.0
    Relay agent IP address: 0.0.0.0
    Client MAC address: 1e:04:f8:b8:2f:d4 (1e:04:f8:b8:2f:d4)
    Client hardware address padding: 00000000000000000000
    Server host name not given
    Boot file name not given
    Magic cookie: DHCP
    Option: (53) DHCP Message Type (ACK)
        Length: 1
        DHCP: ACK (5)
    Option: (51) IP Address Lease Time
        Length: 4
        IP Address Lease Time: (3600s) 1 hour
    Option: (1) Subnet Mask (255.255.255.0)
        Length: 4
        Subnet Mask: 255.255.255.0
    Option: (3) Router
        Length: 4
        Router: 192.168.5.1
    Option: (6) Domain Name Server
        Length: 4
        Domain Name Server: 192.168.5.1
    Option: (42) Network Time Protocol Servers
        Length: 4
        Network Time Protocol Server: 192.168.1.1
    Option: (120) SIP Servers <=====here
        Length: 9
        SIP Server Encoding: IPv4 Address (1)
        SIP Server Address: 192.168.1.1
        SIP Server Address: 192.168.5.2
    Option: (101) TCode
        Length: 13
        TZ TCode: Europe/Berlin
    Option: (54) DHCP Server Identifier (192.168.5.1)
        Length: 4
        DHCP Server Identifier: 192.168.5.1
    Option: (255) End
        Option End: 255
```

```
cat /run/systemd/netif/state                                                                                                   ✔  ⚡  3148  16:40:51
OPER_STATE=routable
CARRIER_STATE=carrier
ADDRESS_STATE=routable
DNS=192.168.94.2 192.168.5.1
NTP=192.168.5.1
SIP=192.168.1.1 192.168.5.2

```

aa

5 years agonetwork dhcp4: Add support send request options in a generic manner
Susant Sahani [Tue, 17 Sep 2019 12:49:22 +0000 (14:49 +0200)]
network dhcp4: Add support send request options in a generic manner

5 years agopolkit: fix typo
AsciiWolf [Thu, 19 Sep 2019 12:34:31 +0000 (14:34 +0200)]
polkit: fix typo

5 years agotest-network: Drop start dnsmasq
Susant Sahani [Wed, 18 Sep 2019 12:14:15 +0000 (14:14 +0200)]
test-network: Drop start dnsmasq

5 years agodhcp6: use unaligned_read_be32()
Yu Watanabe [Wed, 18 Sep 2019 13:14:45 +0000 (22:14 +0900)]
dhcp6: use unaligned_read_be32()

Closes #13591.

5 years agopid1: do not warn if /run/systemd/relabel-extra.d/ doesn't exist
Zbigniew Jędrzejewski-Szmek [Wed, 18 Sep 2019 15:52:30 +0000 (17:52 +0200)]
pid1: do not warn if /run/systemd/relabel-extra.d/ doesn't exist

After all, that is the expected state.

5 years agoMerge pull request #13582 from yuwata/udevadm-reload
Yu Watanabe [Tue, 17 Sep 2019 23:15:01 +0000 (08:15 +0900)]
Merge pull request #13582 from yuwata/udevadm-reload

udevadm: use 'udevadm control' to reload udevd

5 years agoMerge pull request #13583 from keszybz/networkd-hash-compare-equality
Yu Watanabe [Tue, 17 Sep 2019 23:09:48 +0000 (08:09 +0900)]
Merge pull request #13583 from keszybz/networkd-hash-compare-equality

Networkd hash compare equality

5 years agodhcp6: add missing option length check
Yu Watanabe [Tue, 17 Sep 2019 13:18:49 +0000 (22:18 +0900)]
dhcp6: add missing option length check

Closes #13578.

5 years agoMerge pull request #13577 from yuwata/network-fix-ci-failure
Frantisek Sumsal [Tue, 17 Sep 2019 17:42:26 +0000 (17:42 +0000)]
Merge pull request #13577 from yuwata/network-fix-ci-failure

network: fix CI failure

5 years agonetworkd: unbreak routing_policy_rule_compare_func()
Zbigniew Jędrzejewski-Szmek [Tue, 17 Sep 2019 17:27:58 +0000 (19:27 +0200)]
networkd: unbreak routing_policy_rule_compare_func()

!r is the same r == 0, so this was short-circuiting the comparison when
streq(a->iff, b->iff) or streq(a->off, b->off). Before the parent commit which
moved those comparisons to the end, this was short-circuiting quite a bit
of the comparison function.

5 years agoMerge pull request #13508 from yuwata/network-route-fix-13506
Zbigniew Jędrzejewski-Szmek [Tue, 17 Sep 2019 17:20:21 +0000 (19:20 +0200)]
Merge pull request #13508 from yuwata/network-route-fix-13506

network: takes more route information into hash func

5 years agonetworkd: make neighbor_hash_func() handle the same data as neighbor_compare_func()
Zbigniew Jędrzejewski-Szmek [Tue, 17 Sep 2019 17:13:09 +0000 (19:13 +0200)]
networkd: make neighbor_hash_func() handle the same data as neighbor_compare_func()

In practice this probably makes little difference, because we only use AF_INET
and AF_INET6, but it's cleaner to use the same logic in both places.

5 years agonetworkd: use same order in _hash_func() and _compare_func()
Zbigniew Jędrzejewski-Szmek [Tue, 17 Sep 2019 17:11:31 +0000 (19:11 +0200)]
networkd: use same order in _hash_func() and _compare_func()

This makes it easier to see that the same data is handled in both cases.
No functional change.

5 years agonetworkd: split out helper function for address extraction
Zbigniew Jędrzejewski-Szmek [Tue, 17 Sep 2019 17:09:57 +0000 (19:09 +0200)]
networkd: split out helper function for address extraction

5 years agounit: add ExecReload= in systemd-udevd.service
Yu Watanabe [Tue, 17 Sep 2019 16:32:46 +0000 (01:32 +0900)]
unit: add ExecReload= in systemd-udevd.service

5 years agoudevadm: use usec_add()
Yu Watanabe [Tue, 17 Sep 2019 16:28:23 +0000 (01:28 +0900)]
udevadm: use usec_add()

Also fixes description of event source.

5 years agoudevadm: missing initialization of descriptor
Yu Watanabe [Tue, 17 Sep 2019 15:52:07 +0000 (00:52 +0900)]
udevadm: missing initialization of descriptor

5 years agocore: add ExecXYZEx= bus hook ups to all exec command properties
Anita Zhang [Wed, 21 Aug 2019 00:39:38 +0000 (17:39 -0700)]
core: add ExecXYZEx= bus hook ups to all exec command properties

The "Ex" variant was originally only added for ExecStartXYZ= but it makes
sense to have feature parity for the rest of the exec command properties
as well (e.g. ExecReload=, ExecStop=, etc).

5 years agotest: rename TEST-EXECRELOAD to avoid name conflict
Frantisek Sumsal [Tue, 17 Sep 2019 10:22:19 +0000 (12:22 +0200)]
test: rename TEST-EXECRELOAD to avoid name conflict

5 years agoMerge pull request #13559 from ssahani/ipv6ra-route
Yu Watanabe [Tue, 17 Sep 2019 15:13:08 +0000 (00:13 +0900)]
Merge pull request #13559 from ssahani/ipv6ra-route

network: make networkd able to advertise IPv6 routes on links

5 years agocore: coldplug possible nop_job
ypf791 [Fri, 19 Jul 2019 10:28:04 +0000 (18:28 +0800)]
core: coldplug possible nop_job

5 years agonetwork: also take Route::initcwnd and ::initrwnd into hash func
Yu Watanabe [Mon, 9 Sep 2019 15:18:05 +0000 (00:18 +0900)]
network: also take Route::initcwnd and ::initrwnd into hash func

Fixes #13506.

5 years agonetwork: take more route information into hash func
Yu Watanabe [Mon, 9 Sep 2019 15:17:28 +0000 (00:17 +0900)]
network: take more route information into hash func

5 years agonetwork: make route_get() or friends take Route object
Yu Watanabe [Mon, 9 Sep 2019 15:16:41 +0000 (00:16 +0900)]
network: make route_get() or friends take Route object

5 years agosd-netlink: fix invalid assertion
Yu Watanabe [Sun, 15 Sep 2019 13:50:36 +0000 (22:50 +0900)]
sd-netlink: fix invalid assertion

It is natural that n_attiributes is less than type. But in that case,
the message does not contain any message about the type. So, we should
not abort execution with assertion, but just return -ENODATA.

5 years agoMerge pull request #13567 from keszybz/logind-two-refactoring-patches
Yu Watanabe [Tue, 17 Sep 2019 12:41:38 +0000 (21:41 +0900)]
Merge pull request #13567 from keszybz/logind-two-refactoring-patches

Logind two refactoring patches

5 years agonetwork: do not enter failed state if device's sysfs entry does not exist yet
Yu Watanabe [Tue, 17 Sep 2019 12:34:06 +0000 (21:34 +0900)]
network: do not enter failed state if device's sysfs entry does not exist yet

5 years agonetwork: add missing link->network checks
Yu Watanabe [Tue, 17 Sep 2019 12:29:23 +0000 (21:29 +0900)]
network: add missing link->network checks

When the function is called, no network file may be assigned to the
link.

5 years agoMerge pull request #13573 from fbuihuu/ask-password-small-fixes
Frantisek Sumsal [Tue, 17 Sep 2019 12:21:05 +0000 (12:21 +0000)]
Merge pull request #13573 from fbuihuu/ask-password-small-fixes

Ask password small fixes

5 years agoAdd test for IPv6 Route advertisement
Susant Sahani [Mon, 16 Sep 2019 14:43:34 +0000 (20:13 +0530)]
Add test for IPv6 Route advertisement

5 years agonetwork: Add support to advertie ipv6 route
Susant Sahani [Sat, 14 Sep 2019 11:14:22 +0000 (16:44 +0530)]
network: Add support to advertie ipv6 route

Implements https://tools.ietf.org/html/rfc4191

cat veth99.network
```
[Match]
Name=veth99

[Network]
DHCP=no
IPv6PrefixDelegation=yes
Address=2001:db8:0:1::1/64

[IPv6Prefix]
Prefix=2001:db8:0:1::4/64

[IPv6RoutePrefix]
Route=2001:db0:fff::/48

```
Wireshark

```
Frame 481: 142 bytes on wire (1136 bits), 142 bytes captured (1136 bits) on interface 0
Ethernet II, Src: 1e:04:f8:b8:2f:d4 (1e:04:f8:b8:2f:d4), Dst: IPv6mcast_01 (33:33:00:00:00:01)
Internet Protocol Version 6, Src: fe80::1c04:f8ff:feb8:2fd4, Dst: ff02::1
Internet Control Message Protocol v6
    Type: Router Advertisement (134)
    Code: 0
    Checksum: 0xec77 [correct]
    [Checksum Status: Good]
    Cur hop limit: 0
    Flags: 0x00, Prf (Default Router Preference): Medium
    Router lifetime (s): 0
    Reachable time (ms): 0
    Retrans timer (ms): 0
    ICMPv6 Option (Source link-layer address : 1e:04:f8:b8:2f:d4)
        Type: Source link-layer address (1)
        Length: 1 (8 bytes)
        Link-layer address: 1e:04:f8:b8:2f:d4 (1e:04:f8:b8:2f:d4)
    ICMPv6 Option (MTU : 1500)
        Type: MTU (5)
        Length: 1 (8 bytes)
        Reserved
        MTU: 1500
    ICMPv6 Option (Prefix information : 2001:db8:0:1::4/64)
        Type: Prefix information (3)
        Length: 4 (32 bytes)
        Prefix Length: 64
        Flag: 0xc0, On-link flag(L), Autonomous address-configuration flag(A)
        Valid Lifetime: 2592000
        Preferred Lifetime: 604800
        Reserved
        Prefix: 2001:db8:0:1::4
    ICMPv6 Option (Route Information : Medium 2001:db0:fff::/48)
        Type: Route Information (24)
        Length: 3 (24 bytes)
        Prefix Length: 48
        Flag: 0x00, Route Preference: Medium
            ...0 0... = Route Preference: Medium (0)
            000. .000 = Reserved: 0
        Route Lifetime: 604800
        Prefix: 2001:db0:fff::
```

5 years agoMerge pull request #13300 from mrc0mmand/test-execreload
Zbigniew Jędrzejewski-Szmek [Tue, 17 Sep 2019 08:54:18 +0000 (10:54 +0200)]
Merge pull request #13300 from mrc0mmand/test-execreload

 test: introduce TEST-37-EXECRELOAD

5 years agopath: stop watching path specs once we triggered the target unit
Michal Sekletar [Mon, 9 Sep 2019 12:38:35 +0000 (14:38 +0200)]
path: stop watching path specs once we triggered the target unit

We start watching them again once we get a notification that triggered
unit entered inactive or failed state.

Fixes: #10503

5 years agoMerge pull request #13565 from yuwata/network-drop-ipv6ll-address
Zbigniew Jędrzejewski-Szmek [Tue, 17 Sep 2019 07:57:21 +0000 (09:57 +0200)]
Merge pull request #13565 from yuwata/network-drop-ipv6ll-address

network: drop IPv6LL address when LinkLocalAddressing=no|ipv4

5 years agotty-ask-pwd-agent: simplify handling of --wall a bit
Franck Bui [Tue, 17 Sep 2019 07:48:12 +0000 (09:48 +0200)]
tty-ask-pwd-agent: simplify handling of --wall a bit

"wall" was always NULL when passed to parse_password() so let's simplify this
part a bit.

The effective changes are small but wall_tty_match() needed to be moved so it's
still visible from parse_password().

No functional changes.

5 years agoMerge pull request #13494 from yuwata/dhcp6-delay
Zbigniew Jędrzejewski-Szmek [Tue, 17 Sep 2019 07:14:12 +0000 (09:14 +0200)]
Merge pull request #13494 from yuwata/dhcp6-delay

dhcp6: rfc8415 fixes

5 years agocore: Fix setting StatusUnitFormat from config files
Maciej Stanczew [Mon, 16 Sep 2019 00:09:35 +0000 (02:09 +0200)]
core: Fix setting StatusUnitFormat from config files

5 years agotest-network: add one more test case for LinkLocalAddressing=
Yu Watanabe [Sun, 15 Sep 2019 21:16:48 +0000 (06:16 +0900)]
test-network: add one more test case for LinkLocalAddressing=

5 years agonetwork: drop IPv6LL address when LinkLocalAddressing=no|ipv4
Yu Watanabe [Sun, 15 Sep 2019 21:07:38 +0000 (06:07 +0900)]
network: drop IPv6LL address when LinkLocalAddressing=no|ipv4

C.f. disscussion in #13533.

Hopefully fixes #12886.

5 years agonetwork DHCP4: Dont mislead the logs.
Susant Sahani [Mon, 16 Sep 2019 18:46:39 +0000 (20:46 +0200)]
network DHCP4: Dont mislead the logs.

The debug log making sense to the end user as an error.

5 years agotty-ask-pwd-agent: fix message forwarded to wall(1)
Franck Bui [Mon, 16 Sep 2019 20:03:43 +0000 (22:03 +0200)]
tty-ask-pwd-agent: fix message forwarded to wall(1)

Commit a1c111c2d12429d2 wrongly replaced '!' with ':' when the message is
forwarded to wall(1).

Indeed in this case we are not requesting the user for providing a password but
instead we are simply displaying a message to suggest starting
tty-ask-password-agent for providing the password.

5 years agoUpdate m4 for selective utmp support.
Donald A. Cupp Jr [Thu, 5 Sep 2019 22:03:54 +0000 (15:03 -0700)]
Update m4 for selective utmp support.
modified:   tmpfiles.d/systemd.conf.m4

5 years agohwdb: add Medion Akoya E2292 (#13498)
matthiasroos [Mon, 16 Sep 2019 19:10:15 +0000 (21:10 +0200)]
hwdb: add Medion Akoya E2292 (#13498)

Fixes #13480.

5 years agopo: update Brazilian Portuguese translation
Rafael Fontenelle [Sun, 8 Sep 2019 22:08:14 +0000 (19:08 -0300)]
po: update Brazilian Portuguese translation