Daniel Mack [Fri, 20 Nov 2015 15:00:32 +0000 (16:00 +0100)]
Merge pull request #1966 from evverx/fix-initrd-searching-on-debian
tests: fix initrd searching on Debian/Ubuntu
Evgeny Vereshchagin [Fri, 20 Nov 2015 05:21:06 +0000 (05:21 +0000)]
tests: fix initrd searching on Debian/Ubuntu
Lennart Poettering [Fri, 20 Nov 2015 11:24:43 +0000 (12:24 +0100)]
Merge pull request #1950 from evverx/man-systemd-exec-env
man: systemd.exec: add missing variables
Martin Pitt [Fri, 20 Nov 2015 04:16:03 +0000 (05:16 +0100)]
Merge pull request #1964 from poettering/sd-event-man
Finish the documentation for sd-event.h
Tom Gundersen [Fri, 20 Nov 2015 03:04:24 +0000 (04:04 +0100)]
Merge pull request #1962 from mbiebl/install-completion-networkctl
build: install bash completion for networkctl
Lennart Poettering [Thu, 19 Nov 2015 22:38:54 +0000 (23:38 +0100)]
man: fully document sd-event interfaces
This completes the set of man pages for sd-event and contains some minor
other fixes for other man pages too.
The sd_event_set_name(3) man page is renamed to
sd_event_source_set_description(3), which is the correct name of the
concept today.
Lennart Poettering [Thu, 19 Nov 2015 22:33:55 +0000 (23:33 +0100)]
sd-event: define a new PREPARING state
We already have a state RUNNING and EXITING when we dispatch regular and
exit callbacks. Let's introduce a new state called PREPARING that is
active while we invoke preparation callbacks. This way we have a state
each for all three kinds of event handlers.
The states are currently not documented, hence let's add a new state to
the end, before we start documenting this.
Lennart Poettering [Thu, 19 Nov 2015 22:32:52 +0000 (23:32 +0100)]
sd-event: make _ref() calls NOPs when NULL is passed to them
Let's make _ref() calls happy when NULL is passed to them, and simply
return NULL without any assertion logic. This makes them nicely
symmetric to the _unref() calls which also are happy to take NULL and
become NOPs then.
Lennart Poettering [Thu, 19 Nov 2015 22:32:30 +0000 (23:32 +0100)]
Merge pull request #1947 from phomes/sort-includes2
tree-wide: sort includes in *.h
Michael Biebl [Thu, 19 Nov 2015 16:41:41 +0000 (17:41 +0100)]
build: install bash completion for networkctl
This is a follow-up for commit
e4f246b1b512c68260581c4976d2eb01cf7febdc.
Filipe Brandenburger [Thu, 19 Nov 2015 16:10:24 +0000 (08:10 -0800)]
Merge pull request #1953 from martinpitt/fix-gitignore
gitignore: only ignore .html files in man/
Martin Pitt [Thu, 19 Nov 2015 09:51:19 +0000 (10:51 +0100)]
gitignore: only ignore .html files in man/
src/journal-remote/browse.html is git-tracked source and should not be ignored.
Avoid accidentally ignoring similar ones in the future.
Daniel Mack [Thu, 19 Nov 2015 15:00:21 +0000 (16:00 +0100)]
Merge pull request #1960 from evverx/fix-ls-can-not-access
tests: fix ls: cannot access /var/tmp/systemd-test.*/journal/*/*.journal
Evgeny Vereshchagin [Thu, 19 Nov 2015 14:48:21 +0000 (14:48 +0000)]
tests: fix ls: cannot access /var/tmp/systemd-test.*/journal/*/*.journal
don't lose logs from the test machine too
Evgeny Vereshchagin [Thu, 19 Nov 2015 06:28:03 +0000 (06:28 +0000)]
man: systemd.exec: add missing variables
Daniel Mack [Thu, 19 Nov 2015 13:18:01 +0000 (14:18 +0100)]
Merge pull request #1958 from teg/networkd-fixes
nd-ndisc: don't fail if src address is unset
Tom Gundersen [Thu, 19 Nov 2015 11:18:04 +0000 (12:18 +0100)]
nd-ndisc: don't fail if src address is unset
This happens when running our test-suite over a socketpair,
so don't fall over in that case.
Fixes issue #1952.
Filipe Brandenburger [Thu, 19 Nov 2015 08:15:01 +0000 (00:15 -0800)]
Merge pull request #1951 from martinpitt/siphash-test-endianess
test: remove wrong endianess conversion in test-siphash24
Martin Pitt [Thu, 19 Nov 2015 06:50:56 +0000 (07:50 +0100)]
test: remove wrong endianess conversion in test-siphash24
Commit
933f9caee changed the returned result of siphash24_finalize() from
little-endian to native. Follow suit in test-siphash24 and drop the endianess
conversion there as well, so that this succeeds on big-endian machines again.
Fixes #1946.
Tom Gundersen [Thu, 19 Nov 2015 01:27:10 +0000 (02:27 +0100)]
Merge pull request #1931 from bengal/dhcp-fqdn-v2
libsystemd-network: add support for "Client FQDN" DHCP option (v2)
Tom Gundersen [Thu, 19 Nov 2015 01:22:12 +0000 (02:22 +0100)]
Merge pull request #1944 from poettering/randoms-ec
add RandomSec= setting to timer units, and more
David Herrmann [Thu, 19 Nov 2015 00:45:19 +0000 (01:45 +0100)]
Merge pull request #1948 from teg/networkd-fixes
sd-ndisc: drop packets from invalid source addresses
Tom Gundersen [Wed, 18 Nov 2015 20:32:43 +0000 (21:32 +0100)]
Revert "networkd: ndisc - revert to letting the kernel handle NDisc"
This reverts commit
23f186494cf041107896fc4bd7fc5b8f921fd79f.
The bug has now been fixed, so we can go back to managing this ourselves.
Tom Gundersen [Thu, 19 Nov 2015 00:23:24 +0000 (01:23 +0100)]
sd-ndisc: drop RA packets from non-link-local addresses
See https://tools.ietf.org/html/rfc4861#section-4.2. Some routers (dnsmasq) will send packets
from global addresses, which would break the default route setup, so ignore those.
This is also what the kernel does.
Thomas Hindoe Paaboel Andersen [Wed, 18 Nov 2015 21:46:33 +0000 (22:46 +0100)]
tree-wide: sort includes in *.h
This is a continuation of the previous include sort patch, which
only sorted for .c files.
Daniel Mack [Wed, 18 Nov 2015 18:31:08 +0000 (19:31 +0100)]
Merge pull request #1945 from phomes/indentation-fix
network: fix indentation
Thomas Hindoe Paaboel Andersen [Wed, 18 Nov 2015 18:08:44 +0000 (19:08 +0100)]
network: fix indentation
Lennart Poettering [Wed, 18 Nov 2015 12:37:10 +0000 (13:37 +0100)]
update TODO
Lennart Poettering [Wed, 18 Nov 2015 14:33:37 +0000 (15:33 +0100)]
resolved: shortcut lookups names in the local zone
Previously, we'd always generate a packet on the wire, even for names
that are within our local zone. Shortcut this, and always check the
local zone first. This should minimize generated traffic and improve
security.
Lennart Poettering [Wed, 18 Nov 2015 14:30:54 +0000 (15:30 +0100)]
resolved: simplify dns zone logic: take a single key when looking up entries
Instead of taking a DnsQuestion object (i.e. an array of keys) only take
a single key. This simplifies things a bit, and as DNS/LLMNR require a
single question per query message was unnecessary anyway.
This mimics a similar change that was done a while ago for the dns cache
logic.
Lennart Poettering [Wed, 18 Nov 2015 12:37:30 +0000 (13:37 +0100)]
core: add new RandomSec= setting for time units
This allows configuration of a random time on top of the elapse events,
in order to spread time events in a network evenly across a range.
Lennart Poettering [Wed, 18 Nov 2015 11:20:21 +0000 (12:20 +0100)]
systemctl: open pager for 'status', since process tree can get very long
Lennart Poettering [Wed, 18 Nov 2015 11:20:04 +0000 (12:20 +0100)]
systemctl: make sure 'status' shows whether units are transient
Lennart Poettering [Wed, 18 Nov 2015 11:19:42 +0000 (12:19 +0100)]
bus-util: make sure --property=RemainAfterElapse=1 works
Lennart Poettering [Wed, 18 Nov 2015 11:19:15 +0000 (12:19 +0100)]
bus-util: add in forgotten error messages
Lennart Poettering [Wed, 18 Nov 2015 11:18:27 +0000 (12:18 +0100)]
man: improve and fix documentation for RemainAfterElapse=
Lennart Poettering [Wed, 18 Nov 2015 11:17:57 +0000 (12:17 +0100)]
NEWS: add in missing NEWS entry for 228 feature RemainAfterElapse=
Martin Pitt [Wed, 18 Nov 2015 14:20:35 +0000 (15:20 +0100)]
Merge pull request #1943 from teg/networkd-fixes
networkd: fixes
Tom Gundersen [Wed, 18 Nov 2015 13:21:06 +0000 (14:21 +0100)]
networkd: dhcp4 - dirty link when the lease changes
Without this commit, we risk not picking up DNS addresses
as soon as they are available.
Tom Gundersen [Wed, 18 Nov 2015 12:43:49 +0000 (13:43 +0100)]
networkd: ndisc - revert to letting the kernel handle NDisc
There is still a bug in the userspace version, so temporarily revert this.
This is a work-around for issue #1866.
Lennart Poettering [Wed, 18 Nov 2015 10:32:45 +0000 (11:32 +0100)]
Merge pull request #1940 from evverx/man-journald-syslog-forwarding
man: remove section about syslog forwarding
Lennart Poettering [Wed, 18 Nov 2015 10:31:42 +0000 (11:31 +0100)]
Merge pull request #1912 from mustrumr/calendarspec-sub-second-v3
Calendarspec sub second v3
Lennart Poettering [Wed, 18 Nov 2015 10:30:06 +0000 (11:30 +0100)]
Merge pull request #1889 from ssahani/socket-proto
socket: Add support for socket protcol
David Herrmann [Wed, 18 Nov 2015 07:59:06 +0000 (08:59 +0100)]
build: bump version numbers
Prepare for v228 release and bump version numbers.
David Herrmann [Wed, 18 Nov 2015 07:05:12 +0000 (08:05 +0100)]
Merge pull request #1936 from poettering/transient-timers
Fix transient timers and other fixes
Susant Sahani [Wed, 18 Nov 2015 04:03:51 +0000 (09:33 +0530)]
Socket: socket protocol add to dbus properties
Susant Sahani [Mon, 16 Nov 2015 06:45:47 +0000 (12:15 +0530)]
socket: Add support for socket protcol
Now we don't support the socket protocol like
sctp and udplite .
This patch add a new config param
SocketProtocol: udplite/sctp
With this now we can configure the protocol as
udplite = IPPROTO_UDPLITE
sctp = IPPROTO_SCTP
Tested with nspawn:
Evgeny Vereshchagin [Wed, 18 Nov 2015 02:02:29 +0000 (02:02 +0000)]
man: remove section about syslog forwarding
journald: turn ForwardToSyslog= off by default
After all, rsyslog and friends nowadays read their data directly from
the journal, hence the forwarding is unnecessary in most cases.
see
46b131574fdd7d77c15a0919ca9010c
Hristo Venev [Mon, 16 Nov 2015 08:15:17 +0000 (10:15 +0200)]
man: calendarspec sub-second support
Hristo Venev [Mon, 16 Nov 2015 08:15:16 +0000 (10:15 +0200)]
test: calendarspec sub-second tests
Hristo Venev [Mon, 16 Nov 2015 08:15:05 +0000 (10:15 +0200)]
calendarspec: sub-second support, v3
Lennart Poettering [Tue, 17 Nov 2015 19:38:12 +0000 (20:38 +0100)]
run: automatically clean up transient timers created by systemd-run
Lennart Poettering [Tue, 17 Nov 2015 19:13:36 +0000 (20:13 +0100)]
core: add RemainAfterElapse= setting to timer units
Previously, after a timer unit elapsed we'd leave it around for good,
which has the nice benefit that starting a timer that shall trigger at a
specific point in time multiple times will only result in one trigger
instead of possibly many. With this change a new option
RemainAfterElapse= is added. It defaults to "true", to mimic the old
behaviour. If set to "false" timer units will be unloaded after they
elapsed. This is specifically useful for transient timer units.
Lennart Poettering [Tue, 17 Nov 2015 16:11:44 +0000 (17:11 +0100)]
core: Minor cleaning up of unit/log status and log logic
We only reorder a few things and modernize some constructs. No
functional changes.
- Move some if checks from the caller to the callee of a few functions.
- Use IN_SE() where we can
- Move status printing functions together
Lennart Poettering [Tue, 17 Nov 2015 13:11:12 +0000 (14:11 +0100)]
run: when automatically generating names for transient units, use unique bus ID, fallback to random
Previously we used the process ID to generate transient unit names.
However, that is problematic as PIDs get reused easily, and applying
them to remote systems makes little sense.
Fortunately, each bus peer gets a unique, non-reusable ID assigned when
attaching to a bus, hence let's use that, if we can. In some cases we
cannot however, because we connect directly to PID's private socket, and
thus are not a proper bus peer with a unique ID. In that case generate a
random UUID to name the unit after.
Lennart Poettering [Tue, 17 Nov 2015 13:09:16 +0000 (14:09 +0100)]
core: make unit_make_transient() more thorough
Let's reset more stuff that does not apply to transient units. Also,
let's readd the unito to all queues, because it's identity now changed.
Lennart Poettering [Tue, 17 Nov 2015 13:07:38 +0000 (14:07 +0100)]
core: dispatch load queue each time we set up a transient units
manager_load_unit() will dispatch the load queue anyway, but let's make
sure we also dispatch it immediately, after truning a unit into a
transient one and loading the properties from the message. That way the
know about the validity of the unit before we begin processing the next
auxiliary unit.
Lennart Poettering [Tue, 17 Nov 2015 13:04:40 +0000 (14:04 +0100)]
core: move check whether a unit is suitable to become transient into unit.c
Lets introduce unit_is_pristine() that verifies whether a unit is
suitable to become a transient unit, by checking that it is no
referenced yet and has no data on disk assigned.
Lennart Poettering [Tue, 17 Nov 2015 13:03:13 +0000 (14:03 +0100)]
core: generate nice error messages for auxiliary transient units, too
Let's move the validation checks into the loop that sets up the main and
auxiliary transient units, so that we can generate pretty error messages
for all units a transient unit transaction generates, not just for the
main unit.
Lennart Poettering [Tue, 17 Nov 2015 13:01:20 +0000 (14:01 +0100)]
sysctl: use %P instead of %p in core pattern
That way we'll get the PID on the host, rather than the one in a PID
namespace. Which should make the coredump handler less confusing.
Fixes #1930.
Martin Pitt [Tue, 17 Nov 2015 16:01:32 +0000 (17:01 +0100)]
Merge pull request #1920 from teg/networkd-fixes
networkd fixes
Martin Pitt [Tue, 17 Nov 2015 14:59:46 +0000 (15:59 +0100)]
Merge pull request #1932 from teg/networkd-ndisc-timeout
networkd: ndisc timeout
Tom Gundersen [Tue, 17 Nov 2015 14:32:39 +0000 (15:32 +0100)]
networkd: ndisc - consider configured on timeout
Don't block indefinitely, when control has been passed on from NDisc to DHCPv6.
In this case there is likely no IPv6 support on the local link, so otherwise
this would block indefinitely.
Beniamino Galvani [Tue, 17 Nov 2015 13:08:37 +0000 (14:08 +0100)]
hostname-util: fix code comment
Beniamino Galvani [Mon, 16 Nov 2015 09:17:48 +0000 (10:17 +0100)]
libsystemd-network: add support for "Client FQDN" DHCP option
This adds support for the Client Fully Qualified Domain Name (FQDN)
option [RFC 4702] to libsystemd-network. The option can be used to
exchange information about a DHCPv4 client's fully qualified domain
name and about responsibility for updating the DNS RR related to the
client's address assignment.
Other popular DHCP clients (dhclient, dhcpcd) support this option and
it would be useful to have it in networkd too.
Beniamino Galvani [Mon, 16 Nov 2015 16:35:50 +0000 (17:35 +0100)]
dns-domain: add dns_name_to_wire_format()
The function converts a domain name string to the wire format
described in RFC 1035 Section 3.1.
Tom Gundersen [Mon, 16 Nov 2015 16:43:08 +0000 (17:43 +0100)]
sd-dhcp6-client: bind to link-local address
This ensures that several DHCPv6 clients can run on separate interfaces
simultaneously.
Tom Gundersen [Tue, 17 Nov 2015 12:14:36 +0000 (13:14 +0100)]
networkd: ndisc - fix token support
Fixes CID#1338680. Thanks to Thomas Andersen.
Tom Gundersen [Mon, 16 Nov 2015 18:05:27 +0000 (19:05 +0100)]
sd-dhcp6-client: allow multiple clients on host
We need to enable SO_REUSEADDR in order for several sockets to be allowed
to bind to the same port (even on different links).
Tom Gundersen [Mon, 16 Nov 2015 15:47:18 +0000 (16:47 +0100)]
networkd: ndisc - always configure dhcp6 client
The ndisc client may trigger the dhcpv6 client to be started (this is the common case),
so we should allocate the dhcpv6 client whenever we allocate the ndisc one.
Lennart Poettering [Tue, 17 Nov 2015 11:30:44 +0000 (12:30 +0100)]
Merge pull request #1909 from keszybz/filetriggers-v2
Move daemon-reload from package %post scripts to file triggers
Lennart Poettering [Tue, 17 Nov 2015 11:16:57 +0000 (12:16 +0100)]
Merge pull request #1925 from evverx/fix-pam-systemd-user
tests: copy /etc/pam.d/systemd-user from the host on Debian, Ubuntu
Daniel Mack [Tue, 17 Nov 2015 08:36:25 +0000 (09:36 +0100)]
Merge pull request #1926 from phomes/include-order-libudev
tree-wide: group include of libudev.h with sd-*
Daniel Mack [Tue, 17 Nov 2015 08:34:58 +0000 (09:34 +0100)]
Merge pull request #1927 from evverx/no-entries-in-quiet-mode
journalctl: don't print -- No entries -- in quiet mode
Daniel Mack [Tue, 17 Nov 2015 08:33:59 +0000 (09:33 +0100)]
Merge pull request #1924 from poettering/some-fixes-2
casting fixes, another siphash24 alignment fix, and more
Evgeny Vereshchagin [Tue, 17 Nov 2015 06:06:52 +0000 (06:06 +0000)]
journalctl: don't print -- No entries -- in quiet mode
Thomas Hindoe Paaboel Andersen [Tue, 17 Nov 2015 05:52:45 +0000 (06:52 +0100)]
tree-wide: group include of libudev.h with sd-*
Lennart Poettering [Tue, 17 Nov 2015 00:31:42 +0000 (01:31 +0100)]
Some additions to NEWS
Evgeny Vereshchagin [Mon, 16 Nov 2015 23:58:20 +0000 (23:58 +0000)]
tests: copy /etc/pam.d/systemd-user from the host on Debian, Ubuntu
Fixes:
systemd-testsuite systemd[34]: PAM _pam_load_conf_file: unable to open /etc/pam.d/system-auth
systemd-testsuite systemd[34]: PAM _pam_load_conf_file: unable to open /etc/pam.d/system-auth
systemd-testsuite systemd[34]: user@0.service: Failed at step PAM spawning /lib/systemd/systemd: Operation not permitted
...
on Debian, Ubuntu
Lennart Poettering [Tue, 17 Nov 2015 00:08:44 +0000 (01:08 +0100)]
siphash: minor coding style fixes and modernizations
Only cosmetics really, doesn't change any actual logic.
Lennart Poettering [Mon, 16 Nov 2015 23:56:05 +0000 (00:56 +0100)]
siphash: fix another alignment issue
Lennart Poettering [Mon, 16 Nov 2015 23:51:24 +0000 (00:51 +0100)]
tree-wide: make macros for converting fds to pointers and back generic and use them everywhere
Lennart Poettering [Mon, 16 Nov 2015 23:27:18 +0000 (00:27 +0100)]
remount-fs: modernize coding style a bit
a) Use _cleanup_ where it makes sense
b) Uniformly use negative errno-style errors internally, convert to
EXIT_FAILURE/EXIT_SUCCESS only when actually exiting.
c) Use log_oom() where appropriate
d) Fix minor memory leak in hashmap addition error path.
e) Don't pretend we could continue sensibly on OOM or fork() failure
f) Use PR_SET_PDEATHSIG to make sure clients we don't kill on error are
cleaned up.
g) Make use of STRV_MAKE() where it's pretty to do so.
h) Simplify error paths.
Lennart Poettering [Mon, 16 Nov 2015 23:16:21 +0000 (00:16 +0100)]
sd-ipv4ll: fix error path if sd-ipv4acd allocation fails
Let's make sure the destructor cannot hit the n_ref == 0 case.
Lennart Poettering [Mon, 16 Nov 2015 23:02:43 +0000 (00:02 +0100)]
journald: trivial simplification
Lennart Poettering [Mon, 16 Nov 2015 23:02:27 +0000 (00:02 +0100)]
coredump: modernize error logging a bit
Lennart Poettering [Mon, 16 Nov 2015 23:01:00 +0000 (00:01 +0100)]
core: don't generate warnings when write access to the cgroup fs fails in --user due to EACCES
After all, in the classic hierarchy that's pretty much the default case.
Lennart Poettering [Mon, 16 Nov 2015 23:00:32 +0000 (00:00 +0100)]
tree-wide: use right cast macros for UIDs, GIDs and PIDs
Lennart Poettering [Mon, 16 Nov 2015 23:32:06 +0000 (00:32 +0100)]
Merge pull request #1923 from zonque/siphash
siphash24: let siphash24_finalize() and siphash24() return the result…
Daniel Mack [Mon, 16 Nov 2015 22:41:22 +0000 (23:41 +0100)]
Merge pull request #1922 from phomes/sort-includes
Sort includes
Daniel Mack [Mon, 16 Nov 2015 22:17:52 +0000 (23:17 +0100)]
siphash24: let siphash24_finalize() and siphash24() return the result directly
Rather than passing a pointer to return the result, return it directly
from the function calls.
Also, return the result in native endianess, and let the callers care
about the conversion. For hash tables and bloom filters, we don't care,
but in order to keep MAC addresses and DHCP client IDs stable, we
explicitly convert to LE.
Thomas Hindoe Paaboel Andersen [Mon, 16 Nov 2015 21:09:36 +0000 (22:09 +0100)]
tree-wide: sort includes
Sort the includes accoding to the new coding style.
Thomas Hindoe Paaboel Andersen [Mon, 16 Nov 2015 19:13:09 +0000 (20:13 +0100)]
tree-wide: add missing includes
Add a few includes that we rely on to be include already.
Tom Gundersen [Mon, 16 Nov 2015 15:46:14 +0000 (16:46 +0100)]
networkd: ndisc/dhcpv6 - handle starting running clients
The clients may be triggered to be started repeatedly without being stopped first,
simply swallow the error rather than failing the link.
David Herrmann [Mon, 16 Nov 2015 15:47:07 +0000 (16:47 +0100)]
Merge pull request #1919 from martinpitt/master
siphash42: add tests with unaligned input pointers
David Herrmann [Mon, 16 Nov 2015 15:28:25 +0000 (16:28 +0100)]
Merge pull request #1918 from dvdhrm/user3
login: allow re-using users (v3)
Martin Pitt [Mon, 16 Nov 2015 15:24:56 +0000 (16:24 +0100)]
siphash42: add tests with unaligned input pointers
Add test case for calling siphash24 with unaligned input pointers, as we
commonly get with calling it on the result on basename() or similar.
This provides a test for PR #1916, rescued from the superseded PR #1911.
Thanks to Steve Langasek for the test!
David Herrmann [Mon, 16 Nov 2015 14:45:02 +0000 (15:45 +0100)]
login: ignore JobRemoved of old jobs
If we requeue jobs, we are no longer interested in old jobs. Hence, we
better ignore any JobRemoved signals for old jobs and concentrate on our
replacements.
Tom Gundersen [Mon, 16 Nov 2015 14:50:13 +0000 (15:50 +0100)]
Merge pull request #1916 from zonque/align
siphash: alignment
David Herrmann [Mon, 16 Nov 2015 14:48:21 +0000 (15:48 +0100)]
Merge pull request #1915 from poettering/btrfs-root-subvol
tmpfiles: create subvolumes for "v", "q", and "Q" only if / is a subv…