]> git.ipfire.org Git - thirdparty/systemd.git/log
thirdparty/systemd.git
5 years agoMerge pull request #11352 from yuwata/rfe-11348
Lennart Poettering [Tue, 8 Jan 2019 10:43:35 +0000 (11:43 +0100)] 
Merge pull request #11352 from yuwata/rfe-11348

core/socket: logs address or path which fails to be initialized

5 years agoUnchain not related mice in 70-mouse.hwdb
Claudius Ellsel [Tue, 1 Jan 2019 20:42:26 +0000 (21:42 +0100)] 
Unchain not related mice in 70-mouse.hwdb

5 years agocat: add --stderr-priority option
Michael Sloan [Sat, 5 Jan 2019 07:21:19 +0000 (23:21 -0800)] 
cat: add --stderr-priority option

5 years agoMerge pull request #11344 from poettering/various-fixes
Zbigniew Jędrzejewski-Szmek [Tue, 8 Jan 2019 09:33:09 +0000 (10:33 +0100)] 
Merge pull request #11344 from poettering/various-fixes

multiple trivial fixlets, across the tree

5 years agoudevadm: refuse to run trigger, control, settle and monitor commands in chroot
Yu Watanabe [Mon, 7 Jan 2019 05:30:55 +0000 (14:30 +0900)] 
udevadm: refuse to run trigger, control, settle and monitor commands in chroot

Closes #11333.

5 years agoNEWS: also mention that clock file for timesyncd may need to move
Yu Watanabe [Tue, 8 Jan 2019 04:15:14 +0000 (13:15 +0900)] 
NEWS: also mention that clock file for timesyncd may need to move

Follow-up for aa2437e2aee003a361debbb34b341931caf49b3b.

Closes #11329.

5 years agoMerge pull request #11351 from yuwata/ethtool-port-table
Yu Watanabe [Tue, 8 Jan 2019 03:23:11 +0000 (12:23 +0900)] 
Merge pull request #11351 from yuwata/ethtool-port-table

ethtool: cleanups for port_table

5 years agoMerge pull request #11232 from yuwata/fix-9130-alternative
Yu Watanabe [Tue, 8 Jan 2019 03:22:21 +0000 (12:22 +0900)] 
Merge pull request #11232 from yuwata/fix-9130-alternative

network: always check link is ready when address is updated

5 years agocore/socket: logs address or path which fails to be initialized 11352/head
Yu Watanabe [Mon, 7 Jan 2019 19:38:29 +0000 (04:38 +0900)] 
core/socket: logs address or path which fails to be initialized

Closes #11348.

5 years agocore/socket: use _cleanup_ attribute for closing fds on error
Yu Watanabe [Mon, 7 Jan 2019 19:08:32 +0000 (04:08 +0900)] 
core/socket: use _cleanup_ attribute for closing fds on error

5 years agocore/socket: use log_unit_*()
Yu Watanabe [Mon, 7 Jan 2019 19:01:32 +0000 (04:01 +0900)] 
core/socket: use log_unit_*()

5 years agoethtool: minimize the size of port_table 11351/head
Yu Watanabe [Mon, 7 Jan 2019 18:29:21 +0000 (03:29 +0900)] 
ethtool: minimize the size of port_table

Note that `_NET_DEV_PORT_MAX` is 0x100.

5 years agoethtool: use kernel defined values for NET_DEV_PORT_*
Yu Watanabe [Mon, 7 Jan 2019 18:28:47 +0000 (03:28 +0900)] 
ethtool: use kernel defined values for NET_DEV_PORT_*

5 years agoudev: rework how we handle the return value from spawned programs
Zbigniew Jędrzejewski-Szmek [Sun, 6 Jan 2019 21:17:00 +0000 (22:17 +0100)] 
udev: rework how we handle the return value from spawned programs

When running PROGRAM="...", we would log
systemd-udevd[447]: Failed to wait spawned command '...': Input/output error
no matter why the program actually failed, at error level.

The code wouldn't distinguish between an internal failure and a failure in the
program being called and run sd_event_exit(..., -EIO) on any kind of error. EIO
is rather misleading here, becuase it suggests a serious error.

on_spawn_sigchld is updated to set the return code to distinguish failure to
spawn, including the program being killed by a signal (a negative return value),
and the program failing (positive return value).

The logging levels are adjusted, so that for PROGRAM= calls, which are
essentially "if" statements, we only log at debug level (unless we get a
timeout or segfault or another unexpected error).

5 years agotree-wide: always declare bitflag enums the same way 11344/head
Lennart Poettering [Mon, 7 Jan 2019 16:48:28 +0000 (17:48 +0100)] 
tree-wide: always declare bitflag enums the same way

let's always use the 1 << x syntax. No change of behaviour or even of
the compiled binary.

5 years agojson: handle NULL explicitly in json_variant_has_type()
Lennart Poettering [Fri, 4 Jan 2019 12:24:18 +0000 (13:24 +0100)] 
json: handle NULL explicitly in json_variant_has_type()

5 years agojson: fix typo
Lennart Poettering [Fri, 4 Jan 2019 12:24:40 +0000 (13:24 +0100)] 
json: fix typo

5 years agojson: teach json_log() the new SYNTHETIC_ERRNO() logic
Lennart Poettering [Thu, 3 Jan 2019 11:19:10 +0000 (12:19 +0100)] 
json: teach json_log() the new SYNTHETIC_ERRNO() logic

5 years agoloop-util: tweak codepath when a loopback file is "created" from an existing block...
Lennart Poettering [Mon, 31 Dec 2018 16:00:10 +0000 (17:00 +0100)] 
loop-util: tweak codepath when a loopback file is "created" from an existing block device

Be more careful with initialized of the 'relinquished' boolean field,
and let's return the fd, like we do for the regular codepath, too.

5 years agoloop-util: minor simplification
Lennart Poettering [Mon, 31 Dec 2018 15:56:06 +0000 (16:56 +0100)] 
loop-util: minor simplification

5 years agofsck: use symbolic names where we have them
Lennart Poettering [Sun, 23 Dec 2018 18:24:40 +0000 (19:24 +0100)] 
fsck: use symbolic names where we have them

5 years agostat-util: remove spurious newline
Lennart Poettering [Fri, 28 Dec 2018 18:02:05 +0000 (19:02 +0100)] 
stat-util: remove spurious newline

5 years agoenv-util: normalize one more flags enum
Lennart Poettering [Thu, 3 Jan 2019 11:18:47 +0000 (12:18 +0100)] 
env-util: normalize one more flags enum

5 years agoMerge pull request #11327 from keszybz/revert-dbus-address
Lennart Poettering [Mon, 7 Jan 2019 15:14:07 +0000 (16:14 +0100)] 
Merge pull request #11327 from keszybz/revert-dbus-address

Revert "pam_systemd: drop setting DBUS_SESSION_BUS_ADDRESS"

5 years agoMerge pull request #11252 from evverx/use-asan-wrapper-on-travis-ci
Frantisek Sumsal [Sun, 6 Jan 2019 17:48:38 +0000 (18:48 +0100)] 
Merge pull request #11252 from evverx/use-asan-wrapper-on-travis-ci

travis: run PID1, journald and everything else under ASan+UBsan

5 years agoMerge pull request #11241 from bengal/lldp-802-3-subtypes
Lennart Poettering [Sun, 6 Jan 2019 17:15:51 +0000 (18:15 +0100)] 
Merge pull request #11241 from bengal/lldp-802-3-subtypes

lldp: add 802.3 OUI subtype definitions

5 years agopam_systemd: set $DBUS_SESSION_BUS_ADDRESS unconditionally 11327/head
Zbigniew Jędrzejewski-Szmek [Sun, 6 Jan 2019 16:37:00 +0000 (17:37 +0100)] 
pam_systemd: set $DBUS_SESSION_BUS_ADDRESS unconditionally

There's very little lost if the variable is set for a socket that isn't
connectible, but a lot lost (races, ...) if it's not set but the socket exists.

Also, drop the FIXME note, since we don't plan to revert this revert any time
soon.

5 years agoMerge pull request #11328 from keszybz/static-array-size-declarations
Lennart Poettering [Sun, 6 Jan 2019 13:06:59 +0000 (14:06 +0100)] 
Merge pull request #11328 from keszybz/static-array-size-declarations

Use a c99 feature for additional code analysis

5 years agoman: systemctl: document effect of --all on journal output
Lucas Werkmeister [Sat, 5 Jan 2019 14:41:31 +0000 (15:41 +0100)] 
man: systemctl: document effect of --all on journal output

Based on the journalctl documentation of this option added in 23ad99b519
(#10527), but with the first reference to “fields” replaced by “journal
messages”, since I think it’s less common to show other fields with
`systemctl status` (though it’s possible with the `-o` option).

5 years agoPrint the systemd version in a format that dracut likes
Zbigniew Jędrzejewski-Szmek [Fri, 4 Jan 2019 14:07:43 +0000 (15:07 +0100)] 
Print the systemd version in a format that dracut likes

- systemd 240-63-g4199f68+
+ systemd 240 (240-63-g4199f68+)

Sad, but easy.

Fixes #11330.

5 years agoMerge pull request #11337 from keszybz/build-version-h
Lennart Poettering [Sun, 6 Jan 2019 12:55:10 +0000 (13:55 +0100)] 
Merge pull request #11337 from keszybz/build-version-h

meson: declare version.h as dep for libbasic

5 years agoDisable tools/choose-default-locale.sh when cross compiling
James Hilliard [Sat, 5 Jan 2019 03:28:59 +0000 (11:28 +0800)] 
Disable tools/choose-default-locale.sh when cross compiling

5 years agoRevert "meson: declare version.h as dep for fuzzers" 11337/head
Zbigniew Jędrzejewski-Szmek [Sat, 5 Jan 2019 12:20:52 +0000 (13:20 +0100)] 
Revert "meson: declare version.h as dep for fuzzers"

This reverts commit 0c2e93b863f8ec164e7e9e119ab127a752cb1331.

This should not be necessary anymore after previous commit.
I don't quite remember what sequence of steps was failing, but right now
"meson build -Dslow-tests=true && ninja -C build fuzzers" work fine.

5 years agomeson: declare version.h as dep for libbasic
Zbigniew Jędrzejewski-Szmek [Sat, 5 Jan 2019 12:20:37 +0000 (13:20 +0100)] 
meson: declare version.h as dep for libbasic

"meson build && ninja -C build systemd-cat" would fail on missing version.h.

Fixes #11335.

5 years agoUse c99 static array size declarations in exported functions too 11328/head
Zbigniew Jędrzejewski-Szmek [Fri, 4 Jan 2019 11:37:10 +0000 (12:37 +0100)] 
Use c99 static array size declarations in exported functions too

It seems quite useful to provide this additional information in public exported
functions.

This is a c99 feature, not supported in C++. Without the check in _sd-common.h:
FAILED: test-bus-vtable-cc@exe/src_libsystemd_sd-bus_test-bus-vtable-cc.cc.o
...
In file included from ../src/libsystemd/sd-bus/test-bus-vtable-cc.cc:9:
In file included from ../src/systemd/sd-bus-vtable.h:26:
In file included from ../src/systemd/sd-bus.h:26:
../src/systemd/sd-id128.h:38:47: error: static array size is a C99 feature, not permitted in C++
char *sd_id128_to_string(sd_id128_t id, char s[static SD_ID128_STRING_MAX]);
                                              ^

In .c files, I opted to use the define for consistency, even though we don't support
compilation with a C++ compiler, so the unconditional keyword would work too.

5 years agotree-wide: use c99 static for array size declarations
Zbigniew Jędrzejewski-Szmek [Fri, 4 Jan 2019 11:30:45 +0000 (12:30 +0100)] 
tree-wide: use c99 static for array size declarations

https://hamberg.no/erlend/posts/2013-02-18-static-array-indices.html

This only works with clang, unfortunately gcc doesn't seem to implement the check
(tested with gcc-8.2.1-5.fc29.x86_64).

Simulated error:
[2/3] Compiling C object 'systemd-nspawn@exe/src_nspawn_nspawn.c.o'.
../src/nspawn/nspawn.c:3179:45: warning: array argument is too small; contains 15 elements, callee requires at least 16 [-Warray-bounds]
                        candidate = (uid_t) siphash24(arg_machine, strlen(arg_machine), hash_key);
                                            ^                                           ~~~~~~~~
../src/basic/siphash24.h:24:64: note: callee declares array parameter as static here
uint64_t siphash24(const void *in, size_t inlen, const uint8_t k[static 16]);
                                                               ^~~~~~~~~~~~

5 years agoRevert "pam_systemd: drop setting DBUS_SESSION_BUS_ADDRESS"
Zbigniew Jędrzejewski-Szmek [Fri, 4 Jan 2019 10:19:10 +0000 (11:19 +0100)] 
Revert "pam_systemd: drop setting DBUS_SESSION_BUS_ADDRESS"

This reverts commit 2b2b7228bffef626fe8e9f131095995f3d50ee3b.

Fixes #11293.

Removing the environment variable causes problems, e.g. Xfce and Chromium and
... don't communicate with the running dbus instance. If they attempt to start their
own instance, things become even more confusing. Those packages could be fixed
one by one, but removing the variable right now is causing too many problems.

5 years agoMerge pull request #11317 from filbranden/docs1
Zbigniew Jędrzejewski-Szmek [Thu, 3 Jan 2019 17:38:57 +0000 (18:38 +0100)] 
Merge pull request #11317 from filbranden/docs1

Improvements to systemd.io generation

5 years agoMerge pull request #11244 from yuwata/revert-udev-changes
Zbigniew Jędrzejewski-Szmek [Thu, 3 Jan 2019 15:28:30 +0000 (16:28 +0100)] 
Merge pull request #11244 from yuwata/revert-udev-changes

udev: revert bind/unbind patch and one more

5 years agoMerge pull request #11250 from yuwata/dhcp-coding-style-fixes
Zbigniew Jędrzejewski-Szmek [Thu, 3 Jan 2019 14:12:12 +0000 (15:12 +0100)] 
Merge pull request #11250 from yuwata/dhcp-coding-style-fixes

dhcp,network: trivial coding style fixes

5 years agolibudev-util: make util_replace_whitespace() read only len characters
Yu Watanabe [Tue, 25 Dec 2018 03:56:48 +0000 (12:56 +0900)] 
libudev-util: make util_replace_whitespace() read only len characters

This effectively reverts df8ba4fa0e8be1ff7899d08a4b6be0196c8405a0.

Fixes #11264.

5 years agoPass separate dev_t var to device_path_parse_major_minor
YunQiang Su [Tue, 25 Dec 2018 11:01:17 +0000 (19:01 +0800)] 
Pass separate dev_t var to device_path_parse_major_minor

MIPS/O32's st_rdev member of struct stat is unsigned long, which
is 32bit, while dev_t is defined as 64bit, which make some problems
in device_path_parse_major_minor.

Don't pass st.st_rdev, st_mode to device_path_parse_major_minor,
while pass 2 seperate variables. The result of stat is alos copied
out into these 2 variables. Fixes: #11247

5 years agoMerge pull request #11273 from pseyfert/zsh-debugger-completion
Zbigniew Jędrzejewski-Szmek [Thu, 3 Jan 2019 14:00:26 +0000 (15:00 +0100)] 
Merge pull request #11273 from pseyfert/zsh-debugger-completion

zsh/coredumpctl: additional completions

5 years agohwdb: Add Lenovo IdeaPad D330 accel mount matrix
howl [Tue, 25 Dec 2018 22:09:52 +0000 (23:09 +0100)] 
hwdb: Add Lenovo IdeaPad D330 accel mount matrix

I have the fhd screen with pentium without lte model. Don't know if other models could differ in the accelerometer.

5 years agopo: update Italian translation
Daniele Medri [Wed, 2 Jan 2019 12:52:26 +0000 (13:52 +0100)] 
po: update Italian translation

5 years agosleep: install default sleep.conf
Yu Watanabe [Wed, 2 Jan 2019 17:32:57 +0000 (02:32 +0900)] 
sleep: install default sleep.conf

Closes #11310.

5 years agotest-json: check absolute and relative difference in floating point test
Zbigniew Jędrzejewski-Szmek [Fri, 21 Dec 2018 21:49:53 +0000 (22:49 +0100)] 
test-json: check absolute and relative difference in floating point test

The test fails under valgrind, so there was an exception for valgrind.
Unfortunately that check only works when valgrind-devel headers are
available during build. But it is possible to have just valgrind installed,
or simply install it after the build, and then "valgrind test-json" would
fail.

It also seems that even without valgrind, this fails on some arm32 CPUs.
Let's do the usual-style test for absolute and relative differences.

5 years agoMerge pull request #11230 from keszybz/version-string-alt
Yu Watanabe [Thu, 3 Jan 2019 12:33:55 +0000 (21:33 +0900)] 
Merge pull request #11230 from keszybz/version-string-alt

Generate version string from git describe (alternative approach)

5 years agoman: update color of journal logs in DEBUG level
Yu Watanabe [Wed, 2 Jan 2019 21:21:17 +0000 (06:21 +0900)] 
man: update color of journal logs in DEBUG level

Fixes #11303.

5 years agocgroup: s/cgroups? ?v?([0-9])/cgroup v\1/gI
Chris Down [Wed, 2 Jan 2019 20:15:15 +0000 (20:15 +0000)] 
cgroup: s/cgroups? ?v?([0-9])/cgroup v\1/gI

Nitpicky, but we've used a lot of random spacings and names in the past,
but we're trying to be completely consistent on "cgroup vN" now.

Generated by `fd -0 | xargs -0 -n1 sed -ri --follow-symlinks 's/cgroups?  ?v?([0-9])/cgroup v\1/gI'`.

I manually ignored places where it's not appropriate to replace (eg.
"cgroup2" fstype and in src/shared/linux).

5 years agoMerge pull request #11274 from yuwata/fix-11272
Yu Watanabe [Wed, 2 Jan 2019 23:29:26 +0000 (08:29 +0900)] 
Merge pull request #11274 from yuwata/fix-11272

network: set link state configuring before setting addresses

5 years agodocs: generate index.md in Jekyll 11317/head
Filipe Brandenburger [Wed, 2 Jan 2019 22:20:39 +0000 (14:20 -0800)] 
docs: generate index.md in Jekyll

This uses a {% for %} loop in Jekyll to render the page, from the "title"
information in the Front Matter of the actual page files.

This also makes `make-index-md` build rule unnecessary, since generation is
done by the template engine itself.

Tested this by running Jekyll locally.

5 years agodocs: add a "front matter" snippet to our markdown pages
Filipe Brandenburger [Wed, 2 Jan 2019 22:16:34 +0000 (14:16 -0800)] 
docs: add a "front matter" snippet to our markdown pages

It turns out Jekyll (the engine behind GitHub Pages) requires that pages
include a "Front Matter" snippet of YAML at the top for proper rendering.

Omitting it will still render the pages, but including it opens up new
possibilities, such as using a {% for %} loop to generate index.md instead of
requiring a separate script.

I'm hoping this will also fix the issue with some of the pages (notably
CODE_OF_CONDUCT.html) not being available under systemd.io

Tested locally by rendering the website with Jekyll. Before this change, the
*.md files were kept unchanged (so not sure how that even works?!), after this
commit, proper *.html files were generated from it.

5 years agoDocs: Add Missing Space Between Words
Alex Mayer [Wed, 2 Jan 2019 17:53:47 +0000 (12:53 -0500)] 
Docs: Add Missing Space Between Words

5 years agotest-network: disable DNS function of dnsmasq 11274/head
Yu Watanabe [Wed, 26 Dec 2018 22:31:26 +0000 (23:31 +0100)] 
test-network: disable DNS function of dnsmasq

5 years agotest-network: drop unused variable
Yu Watanabe [Wed, 26 Dec 2018 22:30:35 +0000 (23:30 +0100)] 
test-network: drop unused variable

5 years agotest: fix neighbor address
Yu Watanabe [Wed, 26 Dec 2018 22:16:20 +0000 (23:16 +0100)] 
test: fix neighbor address

5 years agonetwork: set *_configured flags to false before requesting addresses or freinds
Yu Watanabe [Sun, 30 Dec 2018 13:08:10 +0000 (22:08 +0900)] 
network: set *_configured flags to false before requesting addresses or freinds

Fixes #11272.

5 years agonetwork: rename link_set_routing_policy_rule() to link_request_set_routing_policy_rule()
Yu Watanabe [Sun, 30 Dec 2018 13:10:32 +0000 (22:10 +0900)] 
network: rename link_set_routing_policy_rule() to link_request_set_routing_policy_rule()

For consistency to other functions.

5 years agonetwork: do not ignore errors on link_request_set_neighbors() and link_set_routing_po...
Yu Watanabe [Sun, 30 Dec 2018 13:07:23 +0000 (22:07 +0900)] 
network: do not ignore errors on link_request_set_neighbors() and link_set_routing_policy()

5 years agoudevadm: add a workaround for dracut 11230/head
Zbigniew Jędrzejewski-Szmek [Wed, 2 Jan 2019 09:26:16 +0000 (10:26 +0100)] 
udevadm: add a workaround for dracut

Also add a comment, so we don't fall into the same trap again.

5 years agoMerge pull request #11291 from yuwata/fix-11259
Zbigniew Jędrzejewski-Szmek [Wed, 2 Jan 2019 08:36:01 +0000 (09:36 +0100)] 
Merge pull request #11291 from yuwata/fix-11259

meson: check whether C.UTF-8 exists or not and use it if exists

5 years agohwdb/sdio.ids: add BCM43438 + minor rewording of class name
Burt P [Sun, 30 Dec 2018 03:49:49 +0000 (21:49 -0600)] 
hwdb/sdio.ids: add BCM43438 + minor rewording of class name

* BCM43438 is used in the Raspberry Pi 3 B and B+
  https://www.raspberrypi.org/products/raspberry-pi-3-model-b/
  "BCM43438 wireless LAN and Bluetooth Low Energy (BLE) on board"
* With this minor change to "C 00", each class name could be prefixed
  with "Is a" and still make sense.

Signed-off-by: Burt P <pburt0@gmail.com>
5 years agoMerge pull request #11297 from yuwata/fix-11295
Zbigniew Jędrzejewski-Szmek [Wed, 2 Jan 2019 07:19:27 +0000 (08:19 +0100)] 
Merge pull request #11297 from yuwata/fix-11295

ask-password: fix assertion

5 years agotest: add test for sending/receiving an invalid device 11244/head
Yu Watanabe [Sat, 22 Dec 2018 18:28:28 +0000 (03:28 +0900)] 
test: add test for sending/receiving an invalid device

5 years agoRevert "udevd: configure a child process name for worker processes"
Yu Watanabe [Sat, 22 Dec 2018 10:49:47 +0000 (19:49 +0900)] 
Revert "udevd: configure a child process name for worker processes"

This reverts commit 49f3ee7e74c714f55aab395c080b1099fc17f7fd.

5 years agoRevert "sd-device: ignore bind/unbind events for now"
Yu Watanabe [Tue, 1 Jan 2019 19:00:14 +0000 (04:00 +0900)] 
Revert "sd-device: ignore bind/unbind events for now"

This reverts commit 56c886dc7ed5b2bb0882ba85136f4070545bfc1b.

Fixes #11277 and #11299.

5 years agomeson: check whether C.UTF-8 exists or not and use it if exists 11291/head
Yu Watanabe [Sat, 29 Dec 2018 13:00:07 +0000 (22:00 +0900)] 
meson: check whether C.UTF-8 exists or not and use it if exists

If C.UTF-8 does not exist, then fallback to en_US.UTF-8 or C.

5 years agozsh/coredumpctl: don't print warnings from completion function 11273/head
Paul Seyfert [Mon, 31 Dec 2018 20:27:09 +0000 (21:27 +0100)] 
zsh/coredumpctl: don't print warnings from completion function

5 years agosd-device: fix segfault when error occurs in device_new_from_{nulstr,strv}()
Yu Watanabe [Sat, 22 Dec 2018 18:06:47 +0000 (03:06 +0900)] 
sd-device: fix segfault when error occurs in device_new_from_{nulstr,strv}()

As devpath may not be set yet.

When debug logging is enabled, log_device_*() calls
sd_device_get_sysname(). So, we should not assume that devpath is always
set.

Fixes #11258.

5 years agoask-password-api: do not call ask_password_keyring() if keyname == NULL 11297/head
Yu Watanabe [Sun, 30 Dec 2018 12:13:43 +0000 (21:13 +0900)] 
ask-password-api: do not call ask_password_keyring() if keyname == NULL

Fixes #11295.

5 years agoask-password: make ask_password_keyring() static
Yu Watanabe [Sun, 30 Dec 2018 12:09:41 +0000 (21:09 +0900)] 
ask-password: make ask_password_keyring() static

5 years agoImprove kernel-install support for initrd files. (#11281)
ikelos [Sat, 29 Dec 2018 19:01:10 +0000 (19:01 +0000)] 
Improve kernel-install support for initrd files. (#11281)

The current support in kernel-install for initrd images doesn't copy
over the initrd file or allow a means for it to be specified (it
requires a specific filename in a particular directory).

This patchset adds support for (optionally) providing the name of
initial ramdisk file to copied over and used by kernel-install.

5 years agotests: introduce check_asan_reports and use it in check_result_{qemu|nspawn} 11252/head
Evgeny Vereshchagin [Tue, 25 Dec 2018 07:31:50 +0000 (08:31 +0100)] 
tests: introduce check_asan_reports and use it in check_result_{qemu|nspawn}

5 years agotests: look for ASan+UBSan reports in the journal
Evgeny Vereshchagin [Sun, 23 Dec 2018 14:48:18 +0000 (15:48 +0100)] 
tests: look for ASan+UBSan reports in the journal

5 years agotests: double DefaultTimeoutStartSec when systemd is run under ASan+UBSan
Evgeny Vereshchagin [Tue, 25 Dec 2018 05:13:45 +0000 (06:13 +0100)] 
tests: double DefaultTimeoutStartSec when systemd is run under ASan+UBSan

5 years agotravis: also run TEST-01-BASIC with QEMU to cover udevd and so on
Evgeny Vereshchagin [Tue, 25 Dec 2018 04:56:46 +0000 (05:56 +0100)] 
travis: also run TEST-01-BASIC with QEMU to cover udevd and so on

5 years agotravis: pass $TRAVIS when running the tests so that they can be skipped properly
Evgeny Vereshchagin [Mon, 24 Dec 2018 03:44:47 +0000 (04:44 +0100)] 
travis: pass $TRAVIS when running the tests so that they can be skipped properly

5 years agotravis: merge RUN_CLANG into RUN
Evgeny Vereshchagin [Mon, 24 Dec 2018 03:03:40 +0000 (04:03 +0100)] 
travis: merge RUN_CLANG into RUN

They are basically the same except that a couple of environment variables
have to be passed for building systemd with clang.

5 years agotests: overwrite a hard-coded timeout in systemd-hwdb-update.service
Evgeny Vereshchagin [Mon, 24 Dec 2018 00:15:43 +0000 (01:15 +0100)] 
tests: overwrite a hard-coded timeout in systemd-hwdb-update.service

5 years agojournal: rely on _cleanup_free_ to free a temporary string used in client_context_rea...
Evgeny Vereshchagin [Sun, 23 Dec 2018 23:29:56 +0000 (00:29 +0100)] 
journal: rely on _cleanup_free_ to free a temporary string used in client_context_read_cgroup

Closes https://github.com/systemd/systemd/issues/11253.

5 years agotests: fail if asan has found issues in journald
Evgeny Vereshchagin [Sun, 23 Dec 2018 14:30:41 +0000 (15:30 +0100)] 
tests: fail if asan has found issues in journald

5 years agotests: reproduce https://github.com/systemd/systemd/issues/11251
Evgeny Vereshchagin [Sun, 23 Dec 2018 13:46:00 +0000 (14:46 +0100)] 
tests: reproduce https://github.com/systemd/systemd/issues/11251

5 years agoMake default locale a compile time option
Dave Reisner [Fri, 28 Dec 2018 12:38:36 +0000 (07:38 -0500)] 
Make default locale a compile time option

Default to a locale that's guaranteed to exist everywhere, but let
distros override this with something more exotic if they choose to.

Closes #11259.

5 years agobuild: don't include shared's "serialize.h" in basic's "time-util.c"
Thomas Haller [Fri, 28 Dec 2018 16:14:01 +0000 (17:14 +0100)] 
build: don't include shared's "serialize.h" in basic's "time-util.c"

"src/basic/time-util.c" is part of src/basic and should not include
"src/shared/serialize.h". It is one of the few cases where this
is done. Also, it's not even required.

5 years agoReplace 'older then' with 'older than'
Jani Uusitalo [Fri, 28 Dec 2018 18:35:46 +0000 (20:35 +0200)] 
Replace 'older then' with 'older than'

5 years agoudev-event: do not read stdout or stderr if the pipefd is not created
Yu Watanabe [Tue, 25 Dec 2018 15:36:55 +0000 (00:36 +0900)] 
udev-event: do not read stdout or stderr if the pipefd is not created

Fixes #11255.

5 years agoswitch-root: fix error message
Yu Watanabe [Tue, 25 Dec 2018 04:10:18 +0000 (13:10 +0900)] 
switch-root: fix error message

Fixes #11261.

5 years agocore: free lines after reading them
Evgeny Vereshchagin [Sun, 23 Dec 2018 14:01:03 +0000 (15:01 +0100)] 
core: free lines after reading them

Closes https://github.com/systemd/systemd/issues/11251.

5 years agoDo not start server if it is already runnning (#11245)
rogerjames99 [Fri, 28 Dec 2018 06:34:43 +0000 (06:34 +0000)] 
Do not start server if it is already runnning (#11245)

5 years agomeson: use cross compilation compatible c++ check
James Hilliard [Thu, 27 Dec 2018 01:16:20 +0000 (09:16 +0800)] 
meson: use cross compilation compatible c++ check

5 years agozsh/coredumpctl: offer --quiet
Paul Seyfert [Wed, 26 Dec 2018 13:02:06 +0000 (14:02 +0100)] 
zsh/coredumpctl: offer --quiet

5 years agozsh/coredumpctl: offer --directory option
Paul Seyfert [Tue, 25 Dec 2018 21:57:23 +0000 (22:57 +0100)] 
zsh/coredumpctl: offer --directory option

After `-D <TAB>` and `--directory=<TAB>` _directories is called to
suggest directories.

5 years agozsh/coredumpctl: complete argument for --debugger=
Paul Seyfert [Tue, 25 Dec 2018 18:30:50 +0000 (19:30 +0100)] 
zsh/coredumpctl: complete argument for --debugger=

This way external commands (i.e. executables in the PATH) are suggested
at --debugger=<TAB>. Upon starting a path (e.g. --debugger=/<TAB> or
--debugger=./<TAB>), filenames of executables are completed.

5 years agotest: don't run TEST-01-BASIC in unprivileged containers on Travis CI
Evgeny Vereshchagin [Sun, 23 Dec 2018 13:28:01 +0000 (14:28 +0100)] 
test: don't run TEST-01-BASIC in unprivileged containers on Travis CI

5 years agotests: use systemd-journald to check whether everything has been built with ASan
Evgeny Vereshchagin [Sun, 23 Dec 2018 13:12:20 +0000 (14:12 +0100)] 
tests: use systemd-journald to check whether everything has been built with ASan

5 years agotravis: add another stage to really run everything under ASan+UBsan
Evgeny Vereshchagin [Sun, 23 Dec 2018 12:31:14 +0000 (13:31 +0100)] 
travis: add another stage to really run everything under ASan+UBsan

This should help to catch issues like https://github.com/systemd/systemd/issues/11253
and https://github.com/systemd/systemd/issues/11251.

5 years agotravis: switch to Debian Testing
Evgeny Vereshchagin [Sun, 23 Dec 2018 12:01:41 +0000 (13:01 +0100)] 
travis: switch to Debian Testing

Turns out Fedora is the least stable distribution in terms of
downloading packages. Debian is much more stable so let's use it.

5 years agotravis: skip test-bpf on Travis CI
Evgeny Vereshchagin [Sun, 23 Dec 2018 11:04:20 +0000 (12:04 +0100)] 
travis: skip test-bpf on Travis CI

More generally IPAddress[Allow|Deny] are broken in docker and lxc containers.

5 years agotravis: use xenial instead of trusty
Evgeny Vereshchagin [Sun, 23 Dec 2018 10:10:06 +0000 (11:10 +0100)] 
travis: use xenial instead of trusty

This should also fix an issue mentioned in
https://github.com/systemd/systemd/pull/11196#issuecomment-448123946.