platform/upstream/systemd.git
9 years agoMerge pull request #209 from crrodriguez/master
Kay Sievers [Mon, 15 Jun 2015 17:56:23 +0000 (19:56 +0200)]
Merge pull request #209 from crrodriguez/master

buildsys: missing SECCOMP_CFLAGS in various places

9 years agobuildsys: missing SECCOMP_CFLAGS in various places
Cristian Rodríguez [Mon, 15 Jun 2015 16:36:51 +0000 (13:36 -0300)]
buildsys: missing SECCOMP_CFLAGS in various places

libcore, systemd and nspawn fail to build when seccomp headers
are not in the include path.

9 years agoMerge pull request #208 from poettering/btrfs-rec-snapshot
Lennart Poettering [Mon, 15 Jun 2015 16:11:48 +0000 (18:11 +0200)]
Merge pull request #208 from poettering/btrfs-rec-snapshot

btrfs-util: when snapshotting make sure we don't descent into subvolu…

9 years agobtrfs-util: when snapshotting make sure we don't descent into subvolumes we just...
Lennart Poettering [Mon, 15 Jun 2015 15:53:50 +0000 (17:53 +0200)]
btrfs-util: when snapshotting make sure we don't descent into subvolumes we just created

We already had a safety check in place that we don't end up descending
to the original subvolume again, but we also should avoid descending in
the newly created one.

This is particularly important if we make a snapshot below its source,
like we do in "systemd-nspawn --ephemeral -D /".

Closes https://bugs.freedesktop.org/show_bug.cgi?id=90803

9 years agoMerge pull request #154 from dmedri/master
Daniel Mack [Mon, 15 Jun 2015 12:43:15 +0000 (14:43 +0200)]
Merge pull request #154 from dmedri/master

Italian .po updates

9 years agoMerge pull request #202 from victorenator/l10n-be
Daniel Mack [Mon, 15 Jun 2015 12:17:38 +0000 (14:17 +0200)]
Merge pull request #202 from victorenator/l10n-be

l10n: Add Belarusian translation

9 years agoMerge pull request #206 from zonque/firewall-rename
Daniel Mack [Mon, 15 Jun 2015 12:15:57 +0000 (14:15 +0200)]
Merge pull request #206 from zonque/firewall-rename

firewall: rename fw-util.[ch] → firewall-util.[ch]

9 years agofirewall: rename fw-util.[ch] → firewall-util.[ch]
Daniel Mack [Mon, 15 Jun 2015 11:50:43 +0000 (13:50 +0200)]
firewall: rename fw-util.[ch] → firewall-util.[ch]

The names fw-util.[ch] are too ambiguous, better rename the files to
firewall-util.[ch]. Also rename the test accordingly.

9 years agoMerge pull request #180 from ronnychevalier/rc/coverity_cid_1304686
Lennart Poettering [Mon, 15 Jun 2015 10:22:19 +0000 (12:22 +0200)]
Merge pull request #180 from ronnychevalier/rc/coverity_cid_1304686

login: fix potential null pointer dereference

9 years agoman: document that ExecStop= needs a synchronous tool
Lennart Poettering [Mon, 15 Jun 2015 10:05:26 +0000 (12:05 +0200)]
man: document that ExecStop= needs a synchronous tool

As requested in #199.

9 years agoman: document that SIGCONT always follows SIGTERM
Lennart Poettering [Mon, 15 Jun 2015 10:05:11 +0000 (12:05 +0200)]
man: document that SIGCONT always follows SIGTERM

As requested in #199.

9 years agoman: clarify overriding semantics of systemd-gpt-auto-generator
Lennart Poettering [Mon, 15 Jun 2015 09:48:24 +0000 (11:48 +0200)]
man: clarify overriding semantics of systemd-gpt-auto-generator

Specifically: /etc/fstab overrides the units itself, but not the deps.

See #168.

9 years agoMerge pull request #205 from endocode/iaguis/seccomp-v2
Lennart Poettering [Mon, 15 Jun 2015 09:45:48 +0000 (11:45 +0200)]
Merge pull request #205 from endocode/iaguis/seccomp-v2

nspawn: make seccomp loading errors non-fatal

9 years agohwdb: Update database of Bluetooth company identifiers
Marcel Holtmann [Mon, 15 Jun 2015 09:28:15 +0000 (11:28 +0200)]
hwdb: Update database of Bluetooth company identifiers

9 years agonspawn: make seccomp loading errors non-fatal
Iago López Galeiras [Fri, 12 Jun 2015 14:22:40 +0000 (16:22 +0200)]
nspawn: make seccomp loading errors non-fatal

seccomp_load returns -EINVAL when seccomp support is not enabled in the
kernel [1]. This should be a debug log, not an error that interrupts nspawn.
If the seccomp filter can't be set and audit is enabled, the user will
get an error message anyway.

[1]: http://man7.org/linux/man-pages/man2/prctl.2.html

9 years agologin: fix potential null pointer dereference
Ronny Chevalier [Fri, 12 Jun 2015 09:37:11 +0000 (11:37 +0200)]
login: fix potential null pointer dereference

Fix CID 1304686: Dereference after null check (FORWARD_NULL)

However, this commit does not fix any bug in logind. It helps to keep
the elect_display_compare() function generic.

9 years agosysv-generator test: always log to console
Martin Pitt [Mon, 15 Jun 2015 06:59:44 +0000 (08:59 +0200)]
sysv-generator test: always log to console

Set $SYSTEMD_LOG_TARGET so that the output always goes to stdout/stderr. This
fixes running the test as root, as that logged to the journal previously.

https://github.com/systemd/systemd/issues/195

9 years agoupdate TODO
Lennart Poettering [Sun, 14 Jun 2015 22:41:10 +0000 (00:41 +0200)]
update TODO

9 years agoupdate TODO
Lennart Poettering [Sun, 14 Jun 2015 22:15:12 +0000 (00:15 +0200)]
update TODO

9 years agol10n: Add Belarusian translation
Viktar Vauchkevich [Sun, 14 Jun 2015 21:13:43 +0000 (00:13 +0300)]
l10n: Add Belarusian translation

9 years agoMerge pull request #201 from mbiebl/drop-include_prefix
Kay Sievers [Sun, 14 Jun 2015 18:58:04 +0000 (20:58 +0200)]
Merge pull request #201 from mbiebl/drop-include_prefix

build-sys: Drop include_prefix

9 years agobuild-sys: Drop include_prefix
Michael Biebl [Sun, 14 Jun 2015 18:48:54 +0000 (20:48 +0200)]
build-sys: Drop include_prefix

Appears to be unused and a leftover from the udev merge.

9 years agoMerge pull request #144 from teg/udev-spawn-log-less-2
Kay Sievers [Sun, 14 Jun 2015 18:19:54 +0000 (20:19 +0200)]
Merge pull request #144 from teg/udev-spawn-log-less-2

udevd: event - don't log about failures of spawn processes when this …

9 years agoMerge pull request #200 from kaysievers/wip
Kay Sievers [Sun, 14 Jun 2015 18:17:19 +0000 (20:17 +0200)]
Merge pull request #200 from kaysievers/wip

build-sys: include libsystemd-journal and libudev in libshared

9 years agobuild-sys: include libsystemd-journal and libudev in libshared
Kay Sievers [Sun, 14 Jun 2015 18:02:40 +0000 (20:02 +0200)]
build-sys: include libsystemd-journal and libudev in libshared

9 years agoMerge pull request #196 from dvdhrm/bus-map-props
Tom Gundersen [Sun, 14 Jun 2015 17:35:30 +0000 (19:35 +0200)]
Merge pull request #196 from dvdhrm/bus-map-props

tree-wide: fix memory leaks in users of bus_map_all_properties()

9 years agoMerge pull request #198 from ivuk/fix_typo_timesyncd_conf
Tom Gundersen [Sun, 14 Jun 2015 16:47:41 +0000 (18:47 +0200)]
Merge pull request #198 from ivuk/fix_typo_timesyncd_conf

Fix typos in man/timesyncd.conf.xml

9 years agoMerge pull request #192 from phomes/master
Tom Gundersen [Sun, 14 Jun 2015 16:46:54 +0000 (18:46 +0200)]
Merge pull request #192 from phomes/master

test-netlink-manual: typo fix

9 years agoFix typos in man/timesyncd.conf.xml
Igor Vuk [Sun, 14 Jun 2015 16:28:55 +0000 (18:28 +0200)]
Fix typos in man/timesyncd.conf.xml

9 years agotree-wide: fix memory leaks in users of bus_map_all_properties()
David Herrmann [Sun, 14 Jun 2015 13:08:52 +0000 (15:08 +0200)]
tree-wide: fix memory leaks in users of bus_map_all_properties()

If you use bus_map_all_properties(), you must be aware that it might
touch output variables even though it may fail. This is, because we parse
many different bus-properties and cannot tell how to clean them up, in
case we fail deep down in the parser.

Fix all callers of bus_map_all_properties() to correctly cleanup any
context structures at all times.

9 years agohwdb: add support for Alienware graphics amplifier
Mario Limonciello [Thu, 11 Jun 2015 03:01:51 +0000 (22:01 -0500)]
hwdb: add support for Alienware graphics amplifier

Unplugging and plugging in the cable will create various scancodes
on the keyboard controller.

Userspace within X should be able to interact with these to show
interesting messages. Assign them to generic prog1/prog2.

(David: add comment to hwdb explaining that these keycodes are reserved)

9 years agoman: don't mention '/run' in hwdb.man
David Herrmann [Sun, 14 Jun 2015 12:24:37 +0000 (14:24 +0200)]
man: don't mention '/run' in hwdb.man

We do not support '/run' for hwdb files. Drop it from the man-pages so
people don't accidentally use it.

This was reported by: Peter Hutterer <peter.hutterer@who-t.net>

9 years agotest-netlink-manual: typo fix
Thomas Hindoe Paaboel Andersen [Sun, 14 Jun 2015 11:55:05 +0000 (13:55 +0200)]
test-netlink-manual: typo fix

No functional change, but looked weird.

9 years agoMerge pull request #178 from utezduyar/man-sd_bus_message_get_creds
David Herrmann [Sun, 14 Jun 2015 11:22:44 +0000 (13:22 +0200)]
Merge pull request #178 from utezduyar/man-sd_bus_message_get_creds

Improve the documentation of bus credentials by mentioning send-time metadata. This needs more love, we should really clarify metadata details here. However, this is still better than nothing, so it's fine.

9 years agoMerge pull request #183 from ssahani/net
David Herrmann [Sun, 14 Jun 2015 11:16:47 +0000 (13:16 +0200)]
Merge pull request #183 from ssahani/net

Improve tun/tap logging by using the new log_*errno*() functions that set 'errno' explicitly. Also fix a bunch of incorrect errno/r confusions.

9 years agoMerge pull request #191 from kaysievers/resolv
David Herrmann [Sun, 14 Jun 2015 11:09:06 +0000 (13:09 +0200)]
Merge pull request #191 from kaysievers/resolv

build-sys: merge convenience library libresolve

9 years agoMerge pull request #189 from teg/rtnl-rename
David Herrmann [Sun, 14 Jun 2015 11:07:20 +0000 (13:07 +0200)]
Merge pull request #189 from teg/rtnl-rename

Rename sd_rtnl to sd_netlink to prepare for further netlink-protocol support. Anything rtnl specific still uses the sd_rtnl prefix, but the generic parts (including the bus and message objects) are now called sd_netlink.

9 years agobuild-sys: merge convenience library libresolve
Kay Sievers [Sun, 14 Jun 2015 10:46:07 +0000 (12:46 +0200)]
build-sys: merge convenience library libresolve

9 years agoMerge pull request #187 from kaysievers/libcleanup
Kay Sievers [Sun, 14 Jun 2015 10:34:13 +0000 (12:34 +0200)]
Merge pull request #187 from kaysievers/libcleanup

build-sys: simplify lib dependencies

9 years agobuild-sys: simplify lib dependencies
Kay Sievers [Sat, 13 Jun 2015 18:35:37 +0000 (20:35 +0200)]
build-sys: simplify lib dependencies

9 years agonetworkd: tuntap improve logging
Susant Sahani [Fri, 12 Jun 2015 08:31:51 +0000 (14:01 +0530)]
networkd: tuntap improve logging

Replaces  strerror() usage with log_netdev_error_errno()

9 years agosd-netlink: socket - move some functions from main source file
Tom Gundersen [Sat, 13 Jun 2015 19:25:05 +0000 (21:25 +0200)]
sd-netlink: socket - move some functions from main source file

9 years agosd-netlink: message - split up source file
Tom Gundersen [Sat, 13 Jun 2015 19:10:39 +0000 (21:10 +0200)]
sd-netlink: message - split up source file

Split netlink-socket.c and rtnl-message.c from netlink-message.c.

9 years agosd-netlink: drop the write-queue
Tom Gundersen [Sat, 13 Jun 2015 18:51:56 +0000 (20:51 +0200)]
sd-netlink: drop the write-queue

AF_NETLINK is not write-buffered, so this was actually never used.

9 years agosd-netlink: rename from sd-rtnl
Tom Gundersen [Fri, 12 Jun 2015 14:31:33 +0000 (16:31 +0200)]
sd-netlink: rename from sd-rtnl

9 years agoMerge pull request #184 from l10n-tw/master
Daniel Mack [Sat, 13 Jun 2015 08:38:22 +0000 (10:38 +0200)]
Merge pull request #184 from l10n-tw/master

po: fix zh_TW mailing list again.

9 years agoMerge pull request #167 from keszybz/line-oriented-ima-setup
keszybz [Sat, 13 Jun 2015 03:52:13 +0000 (23:52 -0400)]
Merge pull request #167 from keszybz/line-oriented-ima-setup

ima-setup: write policy one line at a time

9 years agopo: fix zh_TW mailing list again.
Jeff Huang [Sat, 13 Jun 2015 02:45:12 +0000 (10:45 +0800)]
po: fix zh_TW mailing list again.

9 years agoMerge pull request #179 from l10n-tw/master
Ronny Chevalier [Fri, 12 Jun 2015 13:30:54 +0000 (15:30 +0200)]
Merge pull request #179 from l10n-tw/master

po: fix zh_TW mailing list.

9 years agopo: fix zh_TW mailing list.
Jeff Huang [Fri, 12 Jun 2015 13:14:04 +0000 (21:14 +0800)]
po: fix zh_TW mailing list.

9 years agoMerge pull request #164 from l10n-tw/master
Ronny Chevalier [Fri, 12 Jun 2015 13:02:00 +0000 (15:02 +0200)]
Merge pull request #164 from l10n-tw/master

Add zh_TW translation.

9 years agopo,catalog: add zh_TW translation.
Jeff Huang [Wed, 10 Jun 2015 14:39:44 +0000 (22:39 +0800)]
po,catalog: add zh_TW translation.

9 years agoman: mention sd_bus_message_get_creds as another way
Umut Tezduyar Lindskog [Fri, 12 Jun 2015 12:33:48 +0000 (14:33 +0200)]
man: mention sd_bus_message_get_creds as another way

of retrieving sd_bus_creds even though sd_bus_creds itself
contains very limited information.

9 years agoselinux: whitespace fixes
Lennart Poettering [Fri, 12 Jun 2015 11:35:59 +0000 (13:35 +0200)]
selinux: whitespace fixes

9 years agoMerge pull request #153 from crawford/cciss
Kay Sievers [Fri, 12 Jun 2015 11:38:07 +0000 (13:38 +0200)]
Merge pull request #153 from crawford/cciss

rules: re-add cciss rules

9 years agoMerge pull request #173 from mischief/ipforwarding-3
Lennart Poettering [Fri, 12 Jun 2015 10:28:16 +0000 (12:28 +0200)]
Merge pull request #173 from mischief/ipforwarding-3

IPForwarding=kernel v3

9 years agoMerge pull request #177 from ssahani/net
Daniel Mack [Fri, 12 Jun 2015 06:11:22 +0000 (08:11 +0200)]
Merge pull request #177 from ssahani/net

core: fix CID 996302

9 years agocore: fix CID 996302
Susant Sahani [Fri, 12 Jun 2015 04:54:26 +0000 (10:24 +0530)]
core: fix CID 996302

CID 996302:  Error handling issues  (CHECKED_RETURN)

9 years agoman: document IPForward=kernel option
Nick Owens [Tue, 2 Jun 2015 22:42:21 +0000 (15:42 -0700)]
man: document IPForward=kernel option

9 years agonetworkd: create "kernel" setting for IPForwarding
Nick Owens [Fri, 1 May 2015 18:48:08 +0000 (11:48 -0700)]
networkd: create "kernel" setting for IPForwarding

In 5a8bcb674f71a20e95df55319b34c556638378ce, IPForwarding was introduced
to set forwarding flags on interfaces in .network files. networkd sets
forwarding options regardless of the previous setting, even if it was
set by e.g. sysctl. This commit creates a new option for IPForwarding,
"kernel", that preserves the sysctl settings rather than always setting
them.

See https://bugs.freedesktop.org/show_bug.cgi?id=89509 for the initial
bug report.

9 years agoMerge pull request #171 from teg/rtnl-broadcast-2
David Herrmann [Thu, 11 Jun 2015 16:05:18 +0000 (18:05 +0200)]
Merge pull request #171 from teg/rtnl-broadcast-2

sd-rtnl: make joining broadcast groups implicit

9 years agosd-rtnl: make joining broadcast groups implicit
Tom Gundersen [Thu, 11 Jun 2015 13:55:37 +0000 (15:55 +0200)]
sd-rtnl: make joining broadcast groups implicit

9 years agoMerge pull request #143 from teg/networkd-packets-per-slave-mode
Lennart Poettering [Thu, 11 Jun 2015 15:42:46 +0000 (17:42 +0200)]
Merge pull request #143 from teg/networkd-packets-per-slave-mode

networkd: bond - only set packets_per_slave on balance-rr mode

9 years agoMerge pull request #156 from filbranden/journal_leading_whitespace
Lennart Poettering [Thu, 11 Jun 2015 15:42:06 +0000 (17:42 +0200)]
Merge pull request #156 from filbranden/journal_leading_whitespace

journald: do not strip leading whitespace from messages

9 years agoMerge pull request #166 from zonque/kmod
David Herrmann [Thu, 11 Jun 2015 15:01:39 +0000 (17:01 +0200)]
Merge pull request #166 from zonque/kmod

kmod-setup: don't print warning on -ENOSYS

9 years agokmod-setup: don't print warning on -ENOSYS
Daniel Mack [Thu, 11 Jun 2015 11:10:39 +0000 (13:10 +0200)]
kmod-setup: don't print warning on -ENOSYS

-ENOSYS is returned from kmod_module_probe_insert_module() if a module isn't
available, not -ENOENT. Don't spit out a warning in that case unless the
warn_if_unavailable flag is set.

Also factor out the condition into an own variable for better readability.

9 years agoMerge pull request #170 from teg/rtnl-recv
David Herrmann [Thu, 11 Jun 2015 14:27:32 +0000 (16:27 +0200)]
Merge pull request #170 from teg/rtnl-recv

sd-rtnl: use netlink header rather than ucred in socket_recv_message()

9 years agosd-rtnl: use netlink header rather than ucred in socket_recv_message()
Tom Gundersen [Thu, 11 Jun 2015 09:34:54 +0000 (11:34 +0200)]
sd-rtnl: use netlink header rather than ucred in socket_recv_message()

Lets us simplify the function and drop SO_PASSCRED.

Thanks to Alexander Larsson and David Herrmann.

9 years agoima-setup: write policy one line at a time
Zbigniew Jędrzejewski-Szmek [Wed, 10 Jun 2015 19:19:03 +0000 (15:19 -0400)]
ima-setup: write policy one line at a time

ima_write_policy() expects data to be written as one or more
rules, no more than PAGE_SIZE at a time. Easiest way to ensure
that we are not splitting rules is to read and write one line at
a time.

https://bugzilla.redhat.com/show_bug.cgi?id=1226948

9 years agoMerge pull request #84 from blueyed/zsh-optimize-filter_units_by_property
Daniel Mack [Thu, 11 Jun 2015 10:51:27 +0000 (12:51 +0200)]
Merge pull request #84 from blueyed/zsh-optimize-filter_units_by_property

zsh-completion: optimize _filter_units_by_property

9 years agokmod-setup: when we fail to load a kmod, log the error cause
Lennart Poettering [Thu, 11 Jun 2015 10:13:03 +0000 (12:13 +0200)]
kmod-setup: when we fail to load a kmod, log the error cause

(Also, downgrade message from LOG_ERROR to LOG_WARNING, after all we
don't care much and just proceed)

9 years agocore: Let two more booleans survive a daemon-reload
Werner Fink [Wed, 10 Jun 2015 12:36:50 +0000 (14:36 +0200)]
core: Let two more booleans survive a daemon-reload

Without the boolean bus_name_good services as well as cgroup_realized
for units a unit of Type=dbus and ExecReload sending SIGHUP to $MAINPID
will be terminated if systemd will be daemon reloaded.

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=746151
https://bugs.freedesktop.org/show_bug.cgi?id=78311
https://bugzilla.opensuse.org/show_bug.cgi?id=934077

9 years agoMerge pull request #158 from ssahani/net
Lennart Poettering [Thu, 11 Jun 2015 09:52:58 +0000 (11:52 +0200)]
Merge pull request #158 from ssahani/net

Net

9 years agoMerge pull request #157 from filbranden/revert_unused_result
Lennart Poettering [Thu, 11 Jun 2015 09:50:26 +0000 (11:50 +0200)]
Merge pull request #157 from filbranden/revert_unused_result

build-sys: suppress warnings of unused results

9 years agoMerge pull request #160 from kaysievers/libsplit
Daniel Mack [Thu, 11 Jun 2015 09:50:18 +0000 (11:50 +0200)]
Merge pull request #160 from kaysievers/libsplit

build-sys: split internal basic/ library from shared/

9 years agoMerge pull request #150 from poettering/journal-no-recv-loop
Lennart Poettering [Thu, 11 Jun 2015 09:34:38 +0000 (11:34 +0200)]
Merge pull request #150 from poettering/journal-no-recv-loop

journald: don't employ inner loop for reading from incoming sockets

9 years agobuild-sys: split internal basic/ library from shared/
Kay Sievers [Tue, 9 Jun 2015 12:27:33 +0000 (14:27 +0200)]
build-sys: split internal basic/ library from shared/

basic/      can be used by everything
            cannot use anything outside of basic/

libsystemd/ can use basic/
            cannot use shared/

shared/     can use libsystemd/

9 years agosd-network: fix CID 1306198
Susant Sahani [Thu, 11 Jun 2015 08:37:15 +0000 (14:07 +0530)]
sd-network: fix CID 1306198

9 years agonetworkd: improve logging
Susant Sahani [Thu, 11 Jun 2015 07:41:33 +0000 (13:11 +0530)]
networkd: improve logging

Replace strerror() usage with log_netdev_error_errno()

9 years agonetworkd: macvlan improve logging
Susant Sahani [Thu, 11 Jun 2015 07:38:19 +0000 (13:08 +0530)]
networkd: macvlan improve logging

 Replace strerror() usage with log_netdev_error_errno()

9 years agobuild-sys: suppress warnings of unused results
Filipe Brandenburger [Thu, 11 Jun 2015 05:40:31 +0000 (22:40 -0700)]
build-sys: suppress warnings of unused results

Unfortunately, gcc keeps warning about those even when we use an
explicit (void) cast to indicate we are not interested in the result.

LLVM's clang does not have that issue and works fine with the casts.

GCC bug being tracked at:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66425

Until that GCC bug is fixed (and the version with the fix is in
many/most distributions) or we switch to LLVM as the default compiler,
it looks like we'll have to disable this warning by default...

Tested by building files known to present warnings about unused results
without the suppression, confirmed that the warnings were no longer
present with this patch applied.

This partially reverts commit 00c11bc53a1d37 ("build-sys: don't suppress irrelevant warnings").

9 years agojournald: do not strip leading whitespace from messages
Filipe Brandenburger [Thu, 11 Jun 2015 05:33:44 +0000 (22:33 -0700)]
journald: do not strip leading whitespace from messages

Keep leading whitespace for compatibility with older syslog
implementations.  Also useful when piping formatted output to the
`logger` command.  Keep removing trailing whitespace.

Tested with `pstree | logger` and checking that the output of
`journalctl | tail` included aligned and formatted output.

Confirmed that all test cases still pass as expected.

9 years agoMerge pull request #3 from threatgrid/more_cgtop_enhancements
Lennart Poettering [Wed, 10 Jun 2015 23:52:25 +0000 (01:52 +0200)]
Merge pull request #3 from threatgrid/more_cgtop_enhancements

cgtop enhancements for easier machine-readable output

9 years agocgtop: IO readings are valid if any data is available, even if unchanged since last...
Charles Duffy [Tue, 2 Jun 2015 22:38:49 +0000 (17:38 -0500)]
cgtop: IO readings are valid if any data is available, even if unchanged since last tick

Emit "0" rather than "-" if no change in IO values are seen for a process since
last tick, so long as accounting has registered content at all.

9 years agoMerge pull request #6 from xnox/drop-name
Lennart Poettering [Wed, 10 Jun 2015 22:50:04 +0000 (00:50 +0200)]
Merge pull request #6 from xnox/drop-name

shared: Drop 'name=' prefix from SYSTEMD_CGROUP_CONTROLLER define.

9 years agorules: re-add cciss rules
Alex Crawford [Wed, 10 Jun 2015 20:12:02 +0000 (13:12 -0700)]
rules: re-add cciss rules

The original commit (1aff206) doesn't explain why these were removed.
This adds them back since they are in fact needed.

9 years agoAll languages .po files - general update
Daniele Medri [Wed, 10 Jun 2015 21:27:59 +0000 (23:27 +0200)]
All languages .po files - general update

9 years agoItalian .po updates
Daniele Medri [Wed, 10 Jun 2015 21:19:26 +0000 (23:19 +0200)]
Italian .po updates

9 years agoMerge pull request #152 from zonque/buildsys
Lennart Poettering [Wed, 10 Jun 2015 21:02:23 +0000 (23:02 +0200)]
Merge pull request #152 from zonque/buildsys

build-sys: make man/systemd.directives.xml depend on man/custom-entit…

9 years agoMerge pull request #148 from teg/sd-network-race
Lennart Poettering [Wed, 10 Jun 2015 18:43:38 +0000 (20:43 +0200)]
Merge pull request #148 from teg/sd-network-race

sd-network: allow the state dir to be created after the monitor

9 years agoMerge pull request #151 from mischief/dns-shared-3
Lennart Poettering [Wed, 10 Jun 2015 18:43:06 +0000 (20:43 +0200)]
Merge pull request #151 from mischief/dns-shared-3

move dns code from resolve to shared v3

9 years agobuild-sys: make man/systemd.directives.xml depend on man/custom-entities.ent
Daniel Mack [Wed, 10 Jun 2015 18:42:56 +0000 (20:42 +0200)]
build-sys: make man/systemd.directives.xml depend on man/custom-entities.ent

Currently, the following command sequence fails:

  make distclean
  ./autogen.sh c
  make distcheck

That's because the command invoked to build man/systemd.directives.xml needs
man/custom-entities.ent to function, which itself isn't a dependency.

The $(filter-out $<,$^) logic used to filter out everything from the
prerequisites except for the first word, which doesn't work anymore
now. Use $(SOURCE_XML_FILES) instead.

9 years agoInclude tools/compile-unifont.py in the source tarball
Mike Gilbert [Wed, 10 Jun 2015 03:13:36 +0000 (23:13 -0400)]
Include tools/compile-unifont.py in the source tarball

This prevents a build failure when /usr/share/fonts/unifont/unifont.hex
is newer than unifont-glyph-array.bin.

9 years agobus: we now support path_namespace=/
David Herrmann [Wed, 10 Jun 2015 18:14:56 +0000 (20:14 +0200)]
bus: we now support path_namespace=/

Our bloom-filters support root-path matching. Make sure we properly add
the path_namespace= tag.

9 years agobus: fix pattern matching
David Herrmann [Wed, 10 Jun 2015 17:34:05 +0000 (19:34 +0200)]
bus: fix pattern matching

DBus-spec defines two different pattern matchings:

  1) Path and namespace prefix matching. In this case, A matches B either
     if both are equal, or if B is fully included in the namespace of A.
     In other words, A has to be a prefix of B, but end with a separator
     character (or the following character in B must be one).

     This is used for path_namespace= and arg0namespace=

  2) The other pattern matching is used for arg0path= which does a two-way
     matching. That is, A must be a prefix of B, or B a prefix of A.
     Furthermore, the prefix must end with a separator.

Fix the sd-bus helpers to reflect that. The 'simple_' and 'complex_'
prefixes don't make any sense now, but.. eh..

9 years agobus: fix test-bus-kerne-bloom.c to match properly
David Herrmann [Wed, 10 Jun 2015 16:47:31 +0000 (18:47 +0200)]
bus: fix test-bus-kerne-bloom.c to match properly

Make sure we actually verify our match-rules are executed properly. Right
now all we test is the bloom-matches, which are non-reliable as they leave
through false-positives.

9 years agobus: fix arg0path= two-way matching
David Herrmann [Wed, 10 Jun 2015 18:04:53 +0000 (20:04 +0200)]
bus: fix arg0path= two-way matching

DBus spec clearly defines arg0path= to be a two-way matching. That is,
either the matcher or the matchee can be a prefix of the other to match.
This is not possible to implement with bloom-filters. Instead, we'd have
to add a separate filter for each prefix. This is non-trivial, though.
Hence, just skip the match for now and match locally.

9 years agobus: fix bloom_add_prefixes() to add all required data
David Herrmann [Wed, 10 Jun 2015 16:26:16 +0000 (18:26 +0200)]
bus: fix bloom_add_prefixes() to add all required data

Lets look at an example where we add arg0="/foo/bar/waldo" to a
bloom-filter. The following strings are added:
  "arg0:/foo/bar/waldo"
  "arg0-slash-prefix:/foo/bar"
  "arg0-slash-prefix:/foo"

Two problems arise:

1) If we match on "arg0path=/foo/bar/waldo", the dbus-spec explicitly
states that equal strings are also considered prefixes. However, in the
bloom-match, we can only provide a single match-filter. Therefore, we have
to add "arg0-slash-prefix:/foo/bar/waldo" there, but this never occured in
the bloom-mask of the message.
Hence, this patch makes sure bloom_add_prefixes() adds the full path as
prefix, too.

2) If we match on "arg0path=/foo/", the dbus-spec states that arg0path
does prefix-matching with the trailing slash _included_, unlike
path_namespace= matches, which does *not* include them. This is
inconsistent, but we have to support the specs. Therefore, we must add
prefixes with _and_ without trailing separators.
Hence, this patch makes sure bloom_add_prefixes() adds all prefixes with
the trailing slash included.

The final set of strings added therefore is:
  "arg0:/foo/bar/waldo"
  "arg0-slash-prefix:/foo/bar/waldo"
  "arg0-slash-prefix:/foo/bar/"
  "arg0-slash-prefix:/foo/bar"
  "arg0-slash-prefix:/foo/"
  "arg0-slash-prefix:/foo"
  "arg0-slash-prefix:/"

9 years agolibsystemd-network: use domain validation instead of hostname validation for dhcp...
Nick Owens [Tue, 2 Jun 2015 23:30:42 +0000 (16:30 -0700)]
libsystemd-network: use domain validation instead of hostname validation for dhcp domain option

previously hostname_is_valid was used to validate domain names, which
would silently drop perfectly valid dns names that were longer than a
single dns label.

9 years agoshared: add convenience function for validating dns names
Nick Owens [Tue, 2 Jun 2015 23:29:39 +0000 (16:29 -0700)]
shared: add convenience function for validating dns names