platform/upstream/systemd.git
7 years agoHACKING: update qemu-kvm instructions (#6138)
AsciiWolf [Tue, 20 Jun 2017 21:16:16 +0000 (23:16 +0200)]
HACKING: update qemu-kvm instructions (#6138)

7 years agogithub: how to report bugs for older distribution versions (#6146)
Lukáš Nykrýn [Tue, 20 Jun 2017 18:20:25 +0000 (20:20 +0200)]
github: how to report bugs for older distribution versions (#6146)

7 years agoMerge pull request #6113 from keszybz/shell-quoting
Lennart Poettering [Tue, 20 Jun 2017 18:17:03 +0000 (20:17 +0200)]
Merge pull request #6113 from keszybz/shell-quoting

Use "dollar-single-quotes" to escape shell-sensitive strings

7 years agoUse "dollar-single-quotes" to escape shell-sensitive strings
Zbigniew Jędrzejewski-Szmek [Sun, 11 Jun 2017 19:24:07 +0000 (15:24 -0400)]
Use "dollar-single-quotes" to escape shell-sensitive strings

Also called "ANSI-C Quoting" in info:(bash) ANSI-C Quoting.

The escaping rules are a POSIX proposal, and are described in
http://austingroupbugs.net/view.php?id=249. There's a lot of back-and-forth on
the details of escaping of control characters, but we'll be only using a small
subset of the syntax that is common to all proposals and is widely supported.
Unfortunately dash and fish and maybe some other shells do not support it (see
the man page patch for a list).

This allows environment variables to be safely exported using show-environment
and imported into the shell. Shells which do not support this syntax will have
to do something like
    export $(systemctl show-environment|grep -v '=\$')
or whatever is appropriate in their case. I think csh and fish do not support
the A=B syntax anyway, so the change is moot for them.

Fixes #5536.

v2:
- also escape newlines (which currently disallowed in shell values, so this
  doesn't really matter), and tabs (as $'\t'), and ! (as $'!'). This way quoted
  output can be included directly in both interactive and noninteractive bash.

7 years agogithub: tweak the issue template a bit more
Lennart Poettering [Mon, 19 Jun 2017 10:15:44 +0000 (12:15 +0200)]
github: tweak the issue template a bit more

7 years agogithub: tweak the issue template a bit, comment more
Lennart Poettering [Mon, 19 Jun 2017 10:13:28 +0000 (12:13 +0200)]
github: tweak the issue template a bit, comment more

7 years agojob: Ensure JobRunningTimeoutSec= survives serialization (#6128)
Michal Koutný [Sun, 18 Jun 2017 15:51:17 +0000 (17:51 +0200)]
job: Ensure JobRunningTimeoutSec= survives serialization (#6128)

This is a fixup of commit a2df3ea4ae058693bc7bf203d144e8af3c9493d2.
When there is a running job with JobRunningTimeoutSec= and systemd serializes
its state (e.g. during daemon-reload), the timer event source won't be properly
restored in job_coldplug().
Thus save and serialize begin_running_usec too and reinitialize the timer based
on that value.

7 years agoudev: use interface before the string that interface points to is freed by device_add...
Evgeny Vereshchagin [Sun, 18 Jun 2017 09:31:30 +0000 (12:31 +0300)]
udev: use interface before the string that interface points to is freed by device_add_property_internal (#6105)

This prevents udev from reading the data after freeing it.

See https://github.com/systemd/systemd/issues/6040#issuecomment-306589836
==264== Invalid read of size 1
==264==    at 0x4C2E112: strlen (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==264==    by 0x5943EBD: strdup (in /usr/lib/libc-2.25.so)
==264==    by 0x13E263: device_add_property_aux (sd-device.c:122)
==264==    by 0x14788C: device_add_property_internal (sd-device.c:150)
==264==    by 0x14788C: device_rename (device-private.c:786)
==264==    by 0x120DB6: udev_device_rename (libudev-device-private.c:213)
==264==    by 0x120DB6: udev_event_execute_rules (udev-event.c:895)
==264==    by 0x120DB6: worker_spawn (udevd.c:456)
==264==    by 0x1216E5: event_run (udevd.c:584)
==264==    by 0x1216E5: event_queue_start (udevd.c:823)
==264==    by 0x122213: on_uevent (udevd.c:927)
==264==    by 0x141F2F: source_dispatch (sd-event.c:2272)
==264==    by 0x142D52: sd_event_dispatch (sd-event.c:2631)
==264==    by 0x142D52: sd_event_run (sd-event.c:2690)
==264==    by 0x142D52: sd_event_loop (sd-event.c:2710)
==264==    by 0x1159CB: run (udevd.c:1643)
==264==    by 0x1159CB: main (udevd.c:1772)
==264==  Address 0x7b251a0 is 0 bytes inside a block of size 5 free'd
==264==    at 0x4C2C14B: free (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==264==    by 0x13E2A2: freep (alloc-util.h:57)
==264==    by 0x13E2A2: device_add_property_aux (sd-device.c:111)
==264==    by 0x147873: device_add_property_internal (sd-device.c:150)
==264==    by 0x147873: device_rename (device-private.c:781)
==264==    by 0x120DB6: udev_device_rename (libudev-device-private.c:213)
==264==    by 0x120DB6: udev_event_execute_rules (udev-event.c:895)
==264==    by 0x120DB6: worker_spawn (udevd.c:456)
==264==    by 0x1216E5: event_run (udevd.c:584)
==264==    by 0x1216E5: event_queue_start (udevd.c:823)
==264==    by 0x122213: on_uevent (udevd.c:927)
==264==    by 0x141F2F: source_dispatch (sd-event.c:2272)
==264==    by 0x142D52: sd_event_dispatch (sd-event.c:2631)
==264==    by 0x142D52: sd_event_run (sd-event.c:2690)
==264==    by 0x142D52: sd_event_loop (sd-event.c:2710)
==264==    by 0x1159CB: run (udevd.c:1643)
==264==    by 0x1159CB: main (udevd.c:1772)
==264==  Block was alloc'd at
==264==    at 0x4C2AF1F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==264==    by 0x5943EC9: strdup (in /usr/lib/libc-2.25.so)
==264==    by 0x13E263: device_add_property_aux (sd-device.c:122)
==264==    by 0x143B45: device_add_property_internal (sd-device.c:150)
==264==    by 0x143B45: device_amend.lto_priv.235 (device-private.c:454)
==264==    by 0x1387B7: device_append (device-private.c:516)
==264==    by 0x1387B7: device_new_from_nulstr (device-private.c:620)
==264==    by 0x1387B7: udev_device_new_from_nulstr (libudev-device-private.c:268)
==264==    by 0x1387B7: udev_monitor_receive_device (libudev-monitor.c:682)
==264==    by 0x11FC69: worker_spawn (udevd.c:509)
==264==    by 0x1216E5: event_run (udevd.c:584)
==264==    by 0x1216E5: event_queue_start (udevd.c:823)
==264==    by 0x122213: on_uevent (udevd.c:927)
==264==    by 0x141F2F: source_dispatch (sd-event.c:2272)
==264==    by 0x142D52: sd_event_dispatch (sd-event.c:2631)
==264==    by 0x142D52: sd_event_run (sd-event.c:2690)
==264==    by 0x142D52: sd_event_loop (sd-event.c:2710)
==264==    by 0x1159CB: run (udevd.c:1643)
==264==    by 0x1159CB: main (udevd.c:1772)
==264==

7 years agopid1: properly encode infinity when writing CPUQuota snippet (#6141)
Zbigniew Jędrzejewski-Szmek [Sun, 18 Jun 2017 09:18:41 +0000 (05:18 -0400)]
pid1: properly encode infinity when writing CPUQuota snippet (#6141)

We would write
  [Slice]
  CPUQuota=1844674407370955%
which is (numerically) correct, but it seems better to just write
  [Slice]
  CPUQuota=
which is interpreted as USEC_INFINITY by the parser in config_parse_cpu_quota().

Fixes #5965.

7 years agoMerge pull request #6064 from keszybz/lock-session
Lennart Poettering [Sun, 18 Jun 2017 09:16:53 +0000 (11:16 +0200)]
Merge pull request #6064 from keszybz/lock-session

Improvements to loginctl operating on current session

7 years agoMerge pull request #6140 from keszybz/rm-rf-symlink
Lennart Poettering [Sun, 18 Jun 2017 09:15:13 +0000 (11:15 +0200)]
Merge pull request #6140 from keszybz/rm-rf-symlink

Allow symlinks to root to be removed

7 years agoMerge pull request #6142 from keszybz/man-runtimedirectorymode
Djalal Harouni [Sun, 18 Jun 2017 07:36:50 +0000 (09:36 +0200)]
Merge pull request #6142 from keszybz/man-runtimedirectorymode

man: describe RuntimeDirectoryMode=

7 years agoresolved: ignore DNSSEC= option when resolved is built without gcrypt (#6084)
Yu Watanabe [Sat, 17 Jun 2017 20:22:32 +0000 (05:22 +0900)]
resolved: ignore DNSSEC= option when resolved is built without gcrypt (#6084)

Fixes #5583.

7 years agoman: describe RuntimeDirectoryMode=
Zbigniew Jędrzejewski-Szmek [Sat, 17 Jun 2017 19:23:02 +0000 (15:23 -0400)]
man: describe RuntimeDirectoryMode=

Fixes #5509.

7 years agobasic/rm-rf: allow a symlink to / to be removed
Zbigniew Jędrzejewski-Szmek [Sat, 17 Jun 2017 16:41:08 +0000 (12:41 -0400)]
basic/rm-rf: allow a symlink to / to be removed

We open the target path with O_DIRECTORY|O_NOFOLLOW, and if that doesn't work,
we call unlink() on the path. In neither case we will follow the symlink, so
we can relax our check to also not follow symlinks.

Fixes #5864.

7 years agobasic/path-util: allow flags for path_equal_or_files_same
Zbigniew Jędrzejewski-Szmek [Sat, 17 Jun 2017 16:37:16 +0000 (12:37 -0400)]
basic/path-util: allow flags for path_equal_or_files_same

No functional change, just a new parameters and the tests that
AT_SYMLINK_NOFOLLOW works as expected.

7 years agoudev-builtin-blkid: Use _cleanup_blkid_free_probe_ to free probe (#6108)
Nate Clark [Fri, 16 Jun 2017 20:44:57 +0000 (16:44 -0400)]
udev-builtin-blkid: Use _cleanup_blkid_free_probe_ to free probe (#6108)

Use the _cleanup_blkid_free_probe_ to guarantee that the
blkid_probe will be freed when builtin_blkid exits.

7 years agoMerge pull request #6134 from pfl/radv_fixes
Lennart Poettering [Fri, 16 Jun 2017 20:43:36 +0000 (22:43 +0200)]
Merge pull request #6134 from pfl/radv_fixes

Router Advertisement fixes

7 years agoMerge pull request #6133 from wbx-github/idn_config
Lennart Poettering [Fri, 16 Jun 2017 20:06:48 +0000 (22:06 +0200)]
Merge pull request #6133 from wbx-github/idn_config

make IDN support conditional

7 years agomeson: add -Didn=yes|no option
Zbigniew Jędrzejewski-Szmek [Fri, 16 Jun 2017 13:16:28 +0000 (09:16 -0400)]
meson: add -Didn=yes|no option

7 years agoMake IDN support conditional
Waldemar Brodkorb [Thu, 15 Jun 2017 15:44:59 +0000 (17:44 +0200)]
Make IDN support conditional

[zj: rename HAVE_IDN to ENABLE_IDN]

7 years agonetworkd-address: Use DIV_ROUND_UP for consistency
Patrik Flykt [Fri, 16 Jun 2017 06:09:35 +0000 (09:09 +0300)]
networkd-address: Use DIV_ROUND_UP for consistency

Use DIV_ROUND_UP for consistency instead of the previous construct.

7 years agonetworkd-radv: Fix router lifetime assignment
Patrik Flykt [Fri, 16 Jun 2017 06:35:12 +0000 (09:35 +0300)]
networkd-radv: Fix router lifetime assignment

Router lifetime is stored as usec_t, use value accordingly.

7 years agonetworkd: Fix comment
Patrik Flykt [Fri, 16 Jun 2017 06:07:13 +0000 (09:07 +0300)]
networkd: Fix comment

Comment about not being a router is in the wrong place. Router lifetime
is assigned with config_parse_sec().

7 years agomeson: only run c++ tests when c++ compiler is available (#6123)
Davide Cavalca [Thu, 15 Jun 2017 08:55:13 +0000 (09:55 +0100)]
meson: only run c++ tests when c++ compiler is available (#6123)

7 years agoman: systemd-timesyncd.service(8) (#6109)
Pat Riehecky [Fri, 9 Jun 2017 22:48:25 +0000 (17:48 -0500)]
man: systemd-timesyncd.service(8) (#6109)

Updates the documentation to note use of SNTP (resolves #5735)

7 years agocore/mount: pass "-c" flag to /bin/umount (#6093)
NeilBrown [Wed, 7 Jun 2017 12:28:23 +0000 (22:28 +1000)]
core/mount: pass "-c" flag to /bin/umount (#6093)

"-c", which is short for "--no-canonicalize", tells /bin/umount
that the path name is canonical (no .. or symlinks etc).

systemd always uses a canonical name, so this flag is appropriate
for systemd to use.
Knowing that the path is canonical allows umount to avoid
some calls to lstat() on the path.

From v2.30 "-c" goes further and causes umount to avoid all
attempts to 'lstat()' (or similar) the path.  This is important
when automatically unmounting a filesystem, as lstat() can
hang indefinitely in some cases such as when an NFS server
is not accessible.

"-c" has been supported since util-linux 2.17 which is before the
earliest version supported by systemd.
So "-c" is safe to use now, and once util-linux v2.30 is in use,
it will allow mounts from non-responsive NFS servers to be
unmounted.

7 years agoudev: stop freeing value after using it for setting sysattr (#6094)
Evgeny Vereshchagin [Wed, 7 Jun 2017 01:47:47 +0000 (04:47 +0300)]
udev: stop freeing value after using it for setting sysattr (#6094)

This prevents udev from double-freeing and crashing.

See https://github.com/systemd/systemd/issues/6040#issuecomment-306589836
==351== Invalid free() / delete / delete[] / realloc()
==351==    at 0x4C2C14B: free (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==351==    by 0x13CBE8: hashmap_clear_free_free (hashmap.c:900)
==351==    by 0x13CBE8: hashmap_free_free_free (hashmap.c:852)
==351==    by 0x147F4F: sd_device_unref (sd-device.c:88)
==351==    by 0x130CCC: udev_device_unref (libudev-device.c:552)
==351==    by 0x130CD5: udev_device_unref (libudev-device.c:553)
==351==    by 0x11FBBB: worker_spawn (udevd.c:488)
==351==    by 0x1216E5: event_run (udevd.c:584)
==351==    by 0x1216E5: event_queue_start (udevd.c:823)
==351==    by 0x122213: on_uevent (udevd.c:927)
==351==    by 0x141F2F: source_dispatch (sd-event.c:2272)
==351==    by 0x142D52: sd_event_dispatch (sd-event.c:2631)
==351==    by 0x142D52: sd_event_run (sd-event.c:2690)
==351==    by 0x142D52: sd_event_loop (sd-event.c:2710)
==351==    by 0x1159CB: run (udevd.c:1643)
==351==    by 0x1159CB: main (udevd.c:1772)
==351==  Address 0x81745b0 is 0 bytes inside a block of size 1 free'd
==351==    at 0x4C2C14B: free (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==351==    by 0x1447F0: freep (alloc-util.h:57)
==351==    by 0x1447F0: sd_device_set_sysattr_value (sd-device.c:1859)
==351==    by 0x132081: udev_device_set_sysattr_value (libudev-device.c:849)
==351==    by 0x12E777: set_trackpoint_sensitivity (udev-builtin-keyboard.c:180)
==351==    by 0x12E777: builtin_keyboard.lto_priv.170 (udev-builtin-keyboard.c:263)
==351==    by 0x14D03F: udev_builtin_run.constprop.75 (udev-builtin.c:133)
==351==    by 0x11FAEB: udev_event_execute_run (udev-event.c:957)
==351==    by 0x11FAEB: worker_spawn (udevd.c:461)
==351==    by 0x1216E5: event_run (udevd.c:584)
==351==    by 0x1216E5: event_queue_start (udevd.c:823)
==351==    by 0x122213: on_uevent (udevd.c:927)
==351==    by 0x141F2F: source_dispatch (sd-event.c:2272)
==351==    by 0x142D52: sd_event_dispatch (sd-event.c:2631)
==351==    by 0x142D52: sd_event_run (sd-event.c:2690)
==351==    by 0x142D52: sd_event_loop (sd-event.c:2710)
==351==    by 0x1159CB: run (udevd.c:1643)
==351==    by 0x1159CB: main (udevd.c:1772)
==351==  Block was alloc'd at
==351==    at 0x4C2CF35: calloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==351==    by 0x144853: sd_device_set_sysattr_value (sd-device.c:1888)
==351==    by 0x132081: udev_device_set_sysattr_value (libudev-device.c:849)
==351==    by 0x12E777: set_trackpoint_sensitivity (udev-builtin-keyboard.c:180)
==351==    by 0x12E777: builtin_keyboard.lto_priv.170 (udev-builtin-keyboard.c:263)
==351==    by 0x14D03F: udev_builtin_run.constprop.75 (udev-builtin.c:133)
==351==    by 0x11FAEB: udev_event_execute_run (udev-event.c:957)
==351==    by 0x11FAEB: worker_spawn (udevd.c:461)
==351==    by 0x1216E5: event_run (udevd.c:584)
==351==    by 0x1216E5: event_queue_start (udevd.c:823)
==351==    by 0x122213: on_uevent (udevd.c:927)
==351==    by 0x141F2F: source_dispatch (sd-event.c:2272)
==351==    by 0x142D52: sd_event_dispatch (sd-event.c:2631)
==351==    by 0x142D52: sd_event_run (sd-event.c:2690)
==351==    by 0x142D52: sd_event_loop (sd-event.c:2710)
==351==    by 0x1159CB: run (udevd.c:1643)
==351==    by 0x1159CB: main (udevd.c:1772)

7 years agozsh: add completion for add-wants and add-requires (#6082)
Felipe Sateler [Wed, 7 Jun 2017 01:32:15 +0000 (21:32 -0400)]
zsh: add completion for add-wants and add-requires (#6082)

7 years agonetworkd: fix route table from unsigned char to uint32_t (#6083)
Susant Sahani [Tue, 6 Jun 2017 16:02:31 +0000 (16:02 +0000)]
networkd: fix route table from unsigned char to uint32_t (#6083)

[zj: struct Route is defined with uint32_t route, so this makes the type of the function
parameter match the field it is assigned to.]

7 years agol10n: update Czech Translation (#6090)
AsciiWolf [Tue, 6 Jun 2017 15:52:25 +0000 (17:52 +0200)]
l10n: update Czech Translation (#6090)

add missing Report-Msgid-Bugs-To

7 years agosd-device: Try /sys/firmware for sysname to allow device-tree (#5837)
Paul Kocialkowski [Tue, 6 Jun 2017 15:27:30 +0000 (18:27 +0300)]
sd-device: Try /sys/firmware for sysname to allow device-tree (#5837)

This adds /sys/firmware lookup for sysname when creating a new device,
which allows device-tree properties lookup. This look-up can then be
used in udev rules, allowing device-tree-based model detection.

7 years agosystemd-nspawn@.service: start after /var/lib/machines is mounted (#6079)
Josef Gajdusek [Tue, 6 Jun 2017 15:18:22 +0000 (17:18 +0200)]
systemd-nspawn@.service: start after /var/lib/machines is mounted (#6079)

This fixes a race condition during boot, where an nspawn container would start
before /var/lib/machines got mounted resulting in a failure.

7 years agomkosi: update Debian config for mkosi (#6089)
AsciiWolf [Tue, 6 Jun 2017 14:52:52 +0000 (16:52 +0200)]
mkosi: update Debian config for mkosi (#6089)

Drop g++ package — it's only used for a test and we want to keep the mkosi installation footprint small-ish.

7 years agosd-bus: silence format warnings in kdbus code (#6072)
Zbigniew Jędrzejewski-Szmek [Sat, 3 Jun 2017 09:41:17 +0000 (05:41 -0400)]
sd-bus: silence format warnings in kdbus code (#6072)

The code is mostly correct, but gcc is trying to outsmart us, and emits a
warning for a "llu vs lu" mismatch, even though they are the same size (on alpha):

src/libsystemd/sd-bus/bus-control.c: In function ‘kernel_get_list’:
src/libsystemd/sd-bus/bus-control.c:267:42: error: format ‘%llu’ expects argument of type ‘long long unsigned int’, but argument 3 has type ‘__u64 {aka long unsigned int}’ [-Werror=format=]
                         if (asprintf(&n, ":1.%llu", name->id) < 0) {
                                          ^
src/libsystemd/sd-bus/bus-control.c: In function ‘bus_get_name_creds_kdbus’:
src/libsystemd/sd-bus/bus-control.c:714:47: error: format ‘%llu’ expects argument of type ‘long long unsigned int’, but argument 3 has type ‘__u64 {aka long unsigned int}’ [-Werror=format=]
                 if (asprintf(&c->unique_name, ":1.%llu", conn_info->id) < 0) {
                                               ^
This is hard to work around properly, because kdbus.h uses __u64 which is
defined-differently-despite-being-the-same-size then uint64_t. Thus the simple
solution of using %PRIu64 fails on amd64:

src/libsystemd/sd-bus/bus-control.c:714:47: error: format ‘%lu’ expects argument of type ‘long unsigned int’, but argument 3 has type ‘__u64 {aka long long unsigned int}’ [-Werror=format=]
                 if (asprintf(&c->unique_name, ":1.%"PRIu64, conn_info->id) < 0) {
                                               ^~~~~~

Let's just avoid the whole issue for now by silencing the warning.
After the next release, we should just get rid of the kdbus code.

Fixes #5561.

7 years agotests: skip test_exec_inaccessiblepaths_proc when inaccessible dir is unavailable...
Evgeny Vereshchagin [Thu, 1 Jun 2017 17:58:37 +0000 (20:58 +0300)]
tests: skip test_exec_inaccessiblepaths_proc when inaccessible dir is unavailable (#6068)

test_exec_inaccessiblepaths_proc depends on
/run/systemd/inaccessible/dir, which may be unavailable.

See https://github.com/systemd/systemd/issues/5987#issuecomment-305448958.

7 years agohwdb: add multimedia keys for Medion Akoya S3409 (#6062)
Lluís Gili [Thu, 1 Jun 2017 10:16:32 +0000 (12:16 +0200)]
hwdb: add multimedia keys for Medion Akoya S3409 (#6062)

7 years agoMerge pull request #6058 from keszybz/chrooted-test-fixes
Martin Pitt [Thu, 1 Jun 2017 07:37:13 +0000 (09:37 +0200)]
Merge pull request #6058 from keszybz/chrooted-test-fixes

Chrooted test fixes

7 years agologinctl: also use $XDG_SESSION_ID for session-status
Zbigniew Jędrzejewski-Szmek [Thu, 1 Jun 2017 02:42:14 +0000 (22:42 -0400)]
loginctl: also use $XDG_SESSION_ID for session-status

7 years agologinctl: use $XDG_SESSION_ID for "our" session
Zbigniew Jędrzejewski-Szmek [Thu, 1 Jun 2017 02:23:30 +0000 (22:23 -0400)]
loginctl: use $XDG_SESSION_ID for "our" session

Instead of always letting logind guess what the caller's session is, let's
give it the value from $XDG_SESSION_ID when it is present in the caller's
environment.

Nowadays terminal emulators are often running as services under systemd --user,
and not as part of an actual session, so all loginctl calls which depend on
logind guessing the session will fail. I don't see a reason not to honour
$XDG_SESSION_ID.

This applies to LockSession, UnlockSession, TerminateSession, ActivateSession,
SetUserLinger.

Fixes #6032.

7 years agologind: nicer error message when we cannot guess the caller's session
Zbigniew Jędrzejewski-Szmek [Thu, 1 Jun 2017 02:10:15 +0000 (22:10 -0400)]
logind: nicer error message when we cannot guess the caller's session

Partial fix for #6032.

7 years agohwdb: use path_join() to generate the hwdb_bin path (#6063)
Michael Biebl [Thu, 1 Jun 2017 01:21:11 +0000 (03:21 +0200)]
hwdb: use path_join() to generate the hwdb_bin path (#6063)

This avoids having double slashes which can confuse selinux.

7 years agoman: update sd_get_seats(3)
Yu Watanabe [Wed, 31 May 2017 15:27:47 +0000 (00:27 +0900)]
man: update sd_get_seats(3)

7 years agosd-login: sd_get_machine_names(): do not return -EINVAL when output parameter is...
Yu Watanabe [Wed, 31 May 2017 15:12:32 +0000 (00:12 +0900)]
sd-login: sd_get_machine_names(): do not return -EINVAL when output parameter is NULL

Other functions in sd-login generally allow the output parameter to be NULL, in
which case only the number of items that would be stored in the array is returned.
Be nice and do the same here.

7 years agosd-login: treat missing /run/systemd/{seats,sessions,users} the same as empty
Yu Watanabe [Wed, 31 May 2017 14:48:47 +0000 (23:48 +0900)]
sd-login: treat missing /run/systemd/{seats,sessions,users} the same as empty

C.f. 0543105b0fb13e4243b71a78f62f81fb9dde5d51.
This makes if /run/systemd/{seats,sessions,users} are missing, then
sd_get_seats(), sd_get_sessions() and sd_get_uids() return 0, that is,
an empty list, instead of -ENOENT.

7 years agoMerge pull request #6053 from keszybz/doc-tweaks
Martin Pitt [Wed, 31 May 2017 21:24:57 +0000 (23:24 +0200)]
Merge pull request #6053 from keszybz/doc-tweaks

Man page and unit file and mailmap and meson tweaks

7 years agoMerge pull request #6060 from keszybz/reboot-f-warning
Djalal Harouni [Wed, 31 May 2017 14:57:55 +0000 (16:57 +0200)]
Merge pull request #6060 from keszybz/reboot-f-warning

systemctl: avoid spurious warning about missing reboot-param file

7 years agobuild: *really* install tmpfiles.d/systemd-remote.conf when necessary (#6061)
Franck Bui [Wed, 31 May 2017 13:44:08 +0000 (15:44 +0200)]
build: *really* install tmpfiles.d/systemd-remote.conf when necessary (#6061)

This fixes commit 5e354b22520bbb02e which was an attempt to avoid installation
of tmpfiles.d/systemd-remote.conf when it was not needed (ie HAVE_REMOTE=false).

Before this fix and with the autotool build, systemd-remote.conf was
distributed and also installed (although it was empty) even though
HAVE_REMOTE=false.

That's what happens when doing last second changes without retesting...

While at it, update tmpfiles.d/.gitignore (var.conf was missing as well
since commit a083537e5d11b).

7 years agosystemctl: avoid spurious warning about missing reboot-param file
Zbigniew Jędrzejewski-Szmek [Wed, 31 May 2017 03:14:31 +0000 (23:14 -0400)]
systemctl: avoid spurious warning about missing reboot-param file

$ reboot -f
Failed to read reboot parameter file: No such file or directory

It seems that the warning on ENOENT was inadvertently introduced in
27c06cb516c3b87c34f2a1c2c227152997d05c8c.

The warning reported in #5646 comes from systemctl, but let's fix the other
call site in the same way too.

Fixes #5646.

7 years agomeson: output TAGS file in source not build directory
Zbigniew Jędrzejewski-Szmek [Tue, 30 May 2017 23:56:12 +0000 (19:56 -0400)]
meson: output TAGS file in source not build directory

The TAGS file was generated in build/ following what autotools did.
Nevertheless, it's more convenient to put in the source dir.
(It doesn't get deleted by mistake, and it's easier to find for the
editor.)

7 years agomailmap: add entries for recent commits from github
Zbigniew Jędrzejewski-Szmek [Tue, 23 May 2017 23:46:10 +0000 (19:46 -0400)]
mailmap: add entries for recent commits from github

7 years agotest-bus-vtable: it's OK if dbus is not running
Zbigniew Jędrzejewski-Szmek [Wed, 31 May 2017 00:36:18 +0000 (20:36 -0400)]
test-bus-vtable: it's OK if dbus is not running

Fixes #5989.

7 years agosd-login: translate -ENOMEDIUM to -ENODATA
Zbigniew Jędrzejewski-Szmek [Wed, 31 May 2017 00:21:45 +0000 (20:21 -0400)]
sd-login: translate -ENOMEDIUM to -ENODATA

The -ENOMEDIUM return value was introduced in v232-1001-g2977724b09,
('core: make hybrid cgroup unified mode keep compat /sys/fs/cgroup/systemd hierarchy'),
and would be returned by cg_pid_get_path_shifted(), but the documented and
expected return value is -ENODATA. Let's just catch ENXIO/ENOMEDIUM and translate
it to ENODATA in all cases.

Complements 171f8f591ff27ebb5ff475b7a9d1f13a846c9331, fixes #6012.

7 years agosulogin-shell: avoid heredocs
Zbigniew Jędrzejewski-Szmek [Tue, 30 May 2017 21:03:25 +0000 (17:03 -0400)]
sulogin-shell: avoid heredocs

They require a writable /tmp dir (in the bash implementation).
Let's use echo, and not 'echo -e' since that doesn't seem to be completely
portable.

Fixes #6052.

7 years agoman: update MemoryDenyWriteExecute description for executable stacks
Zbigniew Jędrzejewski-Szmek [Tue, 30 May 2017 20:43:48 +0000 (16:43 -0400)]
man: update MemoryDenyWriteExecute description for executable stacks

Without going into details, mention that libraries are also covered by the
filters, and that executable stacks are a no no.

Closes #5970.

7 years agoman: don't say that ExecStart syntax is "very shell"
Zbigniew Jędrzejewski-Szmek [Tue, 30 May 2017 20:31:51 +0000 (16:31 -0400)]
man: don't say that ExecStart syntax is "very shell"

Fixes #6035.

7 years agobuild: only install tmpfiles.d/systemd-remote.conf when necessary (#6051)
Franck Bui [Tue, 30 May 2017 15:03:25 +0000 (17:03 +0200)]
build: only install tmpfiles.d/systemd-remote.conf when necessary (#6051)

Also only include the relevant parts for systemd-journal-remote and
systemd-journal-upload when needed.

7 years agokernel-install: support the case /etc/machine-id is missing or empty (#5975)
Yu Watanabe [Tue, 30 May 2017 13:45:10 +0000 (22:45 +0900)]
kernel-install: support the case /etc/machine-id is missing or empty (#5975)

Some .install plugins does not require that machine ID is set such as
20-grubby.install for Fedora and 50-depmod.install.
To support such plugins to run without valid machine-id, this commit
makes the following change:
* if /etc/machine-id is missing or empty, create temporary directory
  and set its path to BOOT_DIR_ABS,
* run the .install helpers with KERNEL_INSTALL_MACHINE_ID environment
  variable that'd be empty if /etc/machine-id is missing or empty.
This may be useful for installing kernel for e.g. stateless systems
which initialize machine-id while booting the systems.

7 years agorules: watch metadata changes in mmcblk devices (#6050)
kjackiewicz [Tue, 30 May 2017 13:19:03 +0000 (15:19 +0200)]
rules: watch metadata changes in mmcblk devices (#6050)

Formatting sd-cards does not trigger "change" uevents. As a result clients
using udev API don't get any updates afterwards and get outdated information
about the device.

Include mmcblk* in a match for watch option assignment.

7 years agotimesyncd: properly handle OOM errors when parsing fallback servers (#6047)
Lennart Poettering [Tue, 30 May 2017 12:49:35 +0000 (14:49 +0200)]
timesyncd: properly handle OOM errors when parsing fallback servers (#6047)

7 years agoMerge pull request #4512 from pfl/ndisc_exponential_backoff
Lennart Poettering [Tue, 30 May 2017 12:49:07 +0000 (14:49 +0200)]
Merge pull request #4512 from pfl/ndisc_exponential_backoff

Ndisc exponential backoff

7 years agotest-ndisc-rs.c: Test Router Solicitation timer computation
Patrik Flykt [Fri, 19 May 2017 13:22:47 +0000 (16:22 +0300)]
test-ndisc-rs.c: Test Router Solicitation timer computation

Test ICMPv6 Router Solicitation exponential backoff timer by
computing the minimum and maximum values according to RFC 7559,
Section 2 and the algorithm itself described in RFC 3315, Section
14. Reset the sd_ndisc timer to trigger after a zero second delay,
which causes the ndisc timeout to be triggered immediately once
the caller of the "sending" function returns to the main loop.

7 years agosd-ndisc.c: Move Router Solicitation sending after timer computaion
Patrik Flykt [Fri, 19 May 2017 13:22:46 +0000 (16:22 +0300)]
sd-ndisc.c: Move Router Solicitation sending after timer computaion

Move ICMPv6 Router Solicitation sending after timer computation so
that timers are already set up when the packet is being sent. This
makes it possible to create a test that inspects Router
Solicitation timer values when the Router Solicitation is sent out
on the network.

7 years agosd-ndisc: Implement Router Solicitation backoff method
Patrik Flykt [Fri, 19 May 2017 13:22:45 +0000 (16:22 +0300)]
sd-ndisc: Implement Router Solicitation backoff method

Instead of sending a fixed amount of Router Solicitiations, implement
the backoff algorithm proposed in RFC 7559. The backoff algorithm is
the same as used by DHCPv6.

Time out after 12s as specified in RFC 4861 in order not to delay
setting up a link for too long while sending Router Solicitations
in the background. Notice that after this change the callback will
receive a SD_NDISC_EVENT_TIMEOUT timeout event, and at a later point
when a router appears, a received Router Advertisment will cause the
callback to be called again with the SD_NDISC_EVENT_ROUTER event.

7 years agonetwork: bridge vlan without PVID (#5899)
Tobias Jungel [Mon, 29 May 2017 15:20:01 +0000 (17:20 +0200)]
network: bridge vlan without PVID (#5899)

this patch makes it possible to configure a vlan aware bridge without the
PVID. To configure no PVID set DefaultPVID=none in the [BridgeVLAN] section.

fixes #5716

7 years agoMerge pull request #5926 from fsateler/condition-uid
Lennart Poettering [Mon, 29 May 2017 13:18:38 +0000 (15:18 +0200)]
Merge pull request #5926 from fsateler/condition-uid

core: add ConditionUID and ConditionGID

7 years agoMerge pull request #6031 from teg/monitor
Lennart Poettering [Mon, 29 May 2017 13:08:20 +0000 (15:08 +0200)]
Merge pull request #6031 from teg/monitor

busctl: fix up the monitor

7 years agoshell-completion: add systemctl revert (#6042)
Lucas Werkmeister [Mon, 29 May 2017 13:01:01 +0000 (15:01 +0200)]
shell-completion: add systemctl revert (#6042)

The `systemctl revert` command was added in v230 (commit 344ca7556b),
but was missing from the shell completion specifications.

Fixes #5978.

7 years agoMerge pull request #6029 from keszybz/vconsole-no-vga
Lennart Poettering [Fri, 26 May 2017 16:09:57 +0000 (18:09 +0200)]
Merge pull request #6029 from keszybz/vconsole-no-vga

Avoid systemd-vconsole-setup failure on machines w/o VGA

7 years agocore: add @system special value to ConditionUser=
Felipe Sateler [Fri, 19 May 2017 02:12:14 +0000 (22:12 -0400)]
core: add @system special value to ConditionUser=

It allows checking if the user is a system user or a normal user

7 years agocore: add ConditionUser and ConditionGroup
Felipe Sateler [Sat, 8 Apr 2017 23:32:13 +0000 (20:32 -0300)]
core: add ConditionUser and ConditionGroup

This adds two options that are useful for user units. In particular, it
is useful to check ConditionUser=!0 to not start for the root user.

Closes: #5187

7 years agorules: Handle MMC boot partitions by-path correctly (#6026)
George McCollister [Fri, 26 May 2017 02:13:50 +0000 (21:13 -0500)]
rules: Handle MMC boot partitions by-path correctly (#6026)

Many eMMC devices have separate boot partitions that aren't part of the
normal partition table that show up as /dev/mmcblk[0-9]boot[0-9]. These
partitions are generally small (128KB to 16MB) and typically hold a boot
loader, boot loader data or a recovery image. Match these and create
-boot%n by-path symlinks.

Prior to this change by-path symlinks for the main device would be
incorrectly linked to one of the boot partitions.

For instance before:
/dev/disk/by-path/platform-219c000.usdhc linked to /dev/mmcblk1boot1

Now:
/dev/disk/by-path/platform-219c000.usdhc links to /dev/mmcblk1
/dev/disk/by-path/platform-219c000.usdhc-boot0 links to /dev/mmcblk1boot0
/dev/disk/by-path/platform-219c000.usdhc-boot1 links to /dev/mmcblk1boot1

On systems that support multiple SD/MMC devices it can be essential to
have by-path links to these devices since device names vary depending on
which other devices are connected.

7 years agoMerge pull request #5999 from mbiebl/timesyncd-fallback-server
Zbigniew Jędrzejewski-Szmek [Thu, 25 May 2017 22:51:08 +0000 (18:51 -0400)]
Merge pull request #5999 from mbiebl/timesyncd-fallback-server

timesyncd: don't use compiled-in list if FallbackNTP has been configured

7 years agobusctl: monitor - ignore the final NameLost message
Tom Gundersen [Thu, 25 May 2017 14:10:37 +0000 (16:10 +0200)]
busctl: monitor - ignore the final NameLost message

Commit f5938e8ff3cf5b6fadd6b440b3b10fc0e5a64733 started dropping all
messages before we become a monitor, but the last one was getting
through.

This drops also the last NameLost message, which indicatse the switch
from a regular peer to a monitor.

7 years agologn: tests - don't compare signed with unsigned
Tom Gundersen [Thu, 25 May 2017 14:11:04 +0000 (16:11 +0200)]
logn: tests - don't compare signed with unsigned

7 years agotest-timesync: add first test for timesyncd conf parsing
Zbigniew Jędrzejewski-Szmek [Thu, 25 May 2017 14:26:29 +0000 (10:26 -0400)]
test-timesync: add first test for timesyncd conf parsing

We parse the string supplied in NTP_SERVERS during configuration under an
assert_se(). Right now we will accept pretty much anything there, but in case
we are more picky in the future, add a simple test which checks that we can
actually parse whatever is in NTP_SERVERS so that we don't fail the assertion
at runtime.

7 years agotest: ensure 'InaccessiblePaths=/proc' option works (#6017)
Timothée Ravier [Thu, 25 May 2017 04:47:08 +0000 (06:47 +0200)]
test: ensure 'InaccessiblePaths=/proc' option works (#6017)

Test case for PR #5985.

7 years agovconsole-setup: skip setting fonts when setfont returns EX_OSERR
Zbigniew Jędrzejewski-Szmek [Thu, 25 May 2017 03:25:44 +0000 (23:25 -0400)]
vconsole-setup: skip setting fonts when setfont returns EX_OSERR

On a machine without a VGA console, /dev/tty{0,1,…} exist, so
systemd-vconsole-setup is started, but all setfont operations fail.

setfont has a bunch of return codes for different failure modes. It uses
EX_OSERR when the communication with the kernel using ioctls fails. This isn't
too specific, but at least it's only used this general class of errors. Let's
swallow the error in this case to avoid systemd-vconsole-setup.service failing
on cloud vms.

On a machine from https://bugzilla.redhat.com/show_bug.cgi?id=1272686#c4:
$ build/systemd-vconsole-setup
setfont: putfont: 512,8x16:  failed: -1
putfont: PIO_FONT: Invalid argument
/usr/bin/setfont failed with error code 71.
Setting fonts failed with a "system error", ignoring.

$ SYSTEMD_LOG_LEVEL=debug build/systemd-vconsole-setup
Found container virtualization none.
Sysfs UTF-8 flag enabled
UTF-8 kbdmode enabled on /dev/tty0
Executing "/usr/bin/setfont -C /dev/tty0 eurlatgr"...
setfont: putfont: 512,8x16:  failed: -1
putfont: PIO_FONT: Invalid argument
/usr/bin/setfont failed with error code 71.
Executing "/usr/bin/loadkeys -q -C /dev/tty0 -u us"...
/usr/bin/loadkeys succeeded.
Setting fonts failed with a "system error", ignoring.

$ lspci | grep -i vga

$ ls /dev/tty?
/dev/tty0  /dev/tty2  /dev/tty4  /dev/tty6  /dev/tty8
/dev/tty1  /dev/tty3  /dev/tty5  /dev/tty7  /dev/tty9

If we have a better test for /dev/tty? being connected to something that has a
font, we could avoid running setfont at all… ATM, I'm not aware of a simple
test like that.

7 years agovconsole-setup: add more log messages
Zbigniew Jędrzejewski-Szmek [Thu, 25 May 2017 03:25:10 +0000 (23:25 -0400)]
vconsole-setup: add more log messages

This makes it quite a bit easier to see what failed.

strv_join is called inline in log_debug so that it is under the conditional
that kills the whole thing if debugging is disabled.

7 years agotimesyncd: don't use compiled-in list if FallbackNTP has been configured explicitly
Michael Biebl [Sun, 21 May 2017 01:22:43 +0000 (03:22 +0200)]
timesyncd: don't use compiled-in list if FallbackNTP has been configured explicitly

Parse the config files first and only apply the compiled-in list of
fallback servers if no NTP server was configured via FallbackNTP.

Closes: #5091

7 years agoDHCP: Fail link_dhcp_set_routes promotely if no address is assigned from lease (...
Daniel Wang [Wed, 24 May 2017 12:05:49 +0000 (05:05 -0700)]
DHCP: Fail link_dhcp_set_routes promotely if no address is assigned from lease (#6009)

Currently the local variable `address` is unintialized if the DHCP lease
doesn't provide a router address (when r == -ENODATA). Thus the
subsequent call to route_scope_from_address will result in accessing an
unintialized variable.

As a matter of fact, sd-dhcp-client ignores DHCP leases without an
address so link_dhcp_set_routes probably will never be called without a
valid address.

7 years agotmpfiles: Remove unnecessary utmp file creation (#6006)
codekipper [Wed, 24 May 2017 09:10:59 +0000 (11:10 +0200)]
tmpfiles: Remove unnecessary utmp file creation (#6006)

If utmp is disabled (--disable-utmp) then there is no need to create
the wtmp and btmp files.

7 years agoresolved: bugfix of null pointer p->question dereferencing (#6020)
Evgeny Vereshchagin [Wed, 24 May 2017 05:56:48 +0000 (08:56 +0300)]
resolved: bugfix of null pointer p->question dereferencing (#6020)

See https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1621396

7 years agomailmap: add entry for John Paul Adrian Glaubitz (#6015)
John Paul Adrian Glaubitz [Tue, 23 May 2017 13:22:44 +0000 (15:22 +0200)]
mailmap: add entry for John Paul Adrian Glaubitz (#6015)

7 years agoAllow TimeoutSec=0 to work as documented in mount units and elsewhere (#6013)
NeilBrown [Tue, 23 May 2017 07:42:26 +0000 (17:42 +1000)]
Allow TimeoutSec=0 to work as documented in mount units and elsewhere (#6013)

Since commit 36c16a7cdd6c ("core: rework unit timeout handling, and add
new setting RuntimeMaxSec=") TimeoutSec=0 in mount units has
cause the mount to timeout immediately instead of never as documented.

There is a similar problem with Socket.TimeoutSec and Swap.TimeoutSec.

These are easily fixed using config_parse_sec_fix_0().

Automount.TimeoutIdleSec looks like it could have the same problem,
but doesn't because the kernel treats '0' as 'no timeout'.
It handle USEC_INFINITY correctly only because that constant has
the value '-1', and when round up, it becomes zero.
To avoid possible confusion, use config_parse_sec_fix_0() as well, and
explicitly handle USEC_INFINITY.

7 years agonspawn: add nosuid and nodev to /tmp mount (#6004)
tomty89 [Tue, 23 May 2017 07:41:36 +0000 (15:41 +0800)]
nspawn: add nosuid and nodev to /tmp mount (#6004)

When automatic /tmp mount was introduced to nspawn in v219, it was done without having the nosuid and nodev mount options, which was the same case as systemd's default tmp.mount unit back then.

nosuid and nodev was added to tmp.mount(.m4) in v231 for security reasons. matching the nspawn /tmp mount entry against that.

Ref.:
https://github.com/systemd/systemd/commit/2f9df7c96a25adb42093ee3ee201577f3e01da42
https://github.com/systemd/systemd/commit/bbb99c30d01a8bcdc27fb151cc6376a7877a6b07

7 years agonetworkd-link: Receive LLDP on Bridge slaves not master (#5995)
sjoerd-ccu [Tue, 23 May 2017 07:10:59 +0000 (09:10 +0200)]
networkd-link: Receive LLDP on Bridge slaves not master (#5995)

LLDP should be received on bridge slaves as they're the entities
directly connected to a peer. Receiving LLDP on the bridge device makes
little sense, Linux by default even filters out LLDP going onto the
bridge device.

Flip the current logic, receive LLDP on bridge slaves don't listen for
them on the bridge itself.

7 years agoMerge pull request #6005 from grawity/github
Lennart Poettering [Mon, 22 May 2017 13:47:50 +0000 (15:47 +0200)]
Merge pull request #6005 from grawity/github

doc: update GitHub ISSUE_TEMPLATE

7 years agoload-dropin: propagate errors properly from unit_name_compatible() (#6002)
Lennart Poettering [Mon, 22 May 2017 13:18:00 +0000 (15:18 +0200)]
load-dropin: propagate errors properly from unit_name_compatible() (#6002)

Let's log about this in the caller.

Doesn't really matter, but let's do something about my OCD, and
propagate errors properly, so that the caller can log about them.

7 years agogithub: comment out submitter note in ISSUE_TEMPLATE
Mantas Mikulėnas [Mon, 22 May 2017 13:07:28 +0000 (16:07 +0300)]
github: comment out submitter note in ISSUE_TEMPLATE

It's for the person filling in the form, not for people reading it later.

7 years agogithub: remove checkboxes from ISSUE_TEMPLATE
Mantas Mikulėnas [Mon, 22 May 2017 13:06:50 +0000 (16:06 +0300)]
github: remove checkboxes from ISSUE_TEMPLATE

The issue list page thinks those are in fact todo items.

7 years agosd-dhcp: library code shouldn't log above LOG_DEBUG (#6001)
Lennart Poettering [Mon, 22 May 2017 12:38:01 +0000 (14:38 +0200)]
sd-dhcp: library code shouldn't log above LOG_DEBUG (#6001)

Let's downgrade the warning introduced by
955d99edc7991386a36e3d33924cc584931fde91 to debug, as we really
shouldn't log at more than debug level from library code.

(And while we are at it, print the MTU as the right (unsigned) type in
the format string.)

7 years agoMerge pull request #5958 from keszybz/explicit-log-errno
Lennart Poettering [Mon, 22 May 2017 08:12:18 +0000 (10:12 +0200)]
Merge pull request #5958 from keszybz/explicit-log-errno

Use explicit errno in log calls

7 years agomkosi: update Debian config for mkosi (#5997)
AsciiWolf [Mon, 22 May 2017 02:24:30 +0000 (04:24 +0200)]
mkosi: update Debian config for mkosi (#5997)

* mkosi.debian: fix libidn2 package name
* mkosi.debian: drop duplicate diffutils package
* mkosi.debian: add missing g++ package

7 years agoAllow bad MTU values with warning to be able to connect to the machine. (#5954)
Matthias Greiner [Mon, 22 May 2017 01:11:25 +0000 (03:11 +0200)]
Allow bad MTU values with warning to be able to connect to the machine. (#5954)

Ensure the MTU value is valid. Emit a warning and ignore otherwise.

7 years agoMerge pull request #5990 from keszybz/logind
Djalal Harouni [Sun, 21 May 2017 05:14:21 +0000 (07:14 +0200)]
Merge pull request #5990 from keszybz/logind

A bunch of sd-login improvements

7 years agoMerge pull request #6000 from keszybz/fix-oom-warning
Djalal Harouni [Sun, 21 May 2017 05:11:33 +0000 (07:11 +0200)]
Merge pull request #6000 from keszybz/fix-oom-warning

core/load-droping: avoid oom warning when the unit symlink is not a template

7 years agocore/load-droping: avoid oom warning when the unit symlink is not a template
Zbigniew Jędrzejewski-Szmek [Sat, 20 May 2017 23:34:50 +0000 (19:34 -0400)]
core/load-droping: avoid oom warning when the unit symlink is not a template

unit_name_template returns -EINVAL if the unit name is not a template, but
the code assumed that OOM is the only failure mode. Fix that to emit the warning
if a non-template unit is encountered (because in this case we expect the name
to match exactly), and just skip the warning on other errors (presumably oom).

Fixes #5543.

7 years agoMerge pull request #5960 from keszybz/journald-memleak
Evgeny Vereshchagin [Sat, 20 May 2017 22:41:48 +0000 (01:41 +0300)]
Merge pull request #5960 from keszybz/journald-memleak

Journald and journal-remote memleak fixes

7 years agoDHCP: when adding static routes set scopes properly (#5982)
Daniel Wang [Sat, 20 May 2017 11:05:18 +0000 (04:05 -0700)]
DHCP: when adding static routes set scopes properly (#5982)

DHCP responses could include static routes, but unfortunately not an
option to tell what scope to use. So it's important that the client sets
it properly.

This mimics what the `ip route add` command does when adding a static
route without an explicit scope:

* If the destination IP is on the local host, use scope `host`
* Otherwise if the gateway IP is null (direct route), use scope `link`
* If anything else, use the current default `global`.

Fixes #5979.