summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2015-10-14Merge pull request #1562 from michich/overlinkingLennart Poettering
build: fix overlinking
2015-10-14Merge pull request #1563 from poettering/exciting-fixesZbigniew Jędrzejewski-Szmek
A variety of fixes
2015-10-14unit: remove [Install] section from the user exit.target unitLennart Poettering
There's no concept of ctrl-alt-del for user systemd instances, hence don't suggest it woud make sense to symlink the unit to it. Fixes #1525.
2015-10-14tty-ask-password-agent: don't warn if we cannot disable wall messages on the ↵Lennart Poettering
calling tty For example, due to perm issues. THis simply downgrades the message about it, since this is purely cosmetical anyway. Fixes #1543.
2015-10-14core: log if the service to activate vanished before we activate itLennart Poettering
Normally this shouldn't happen unless the daemon is reloaded. A similar check is already in place in socket.c for the socket activation case. This hopefully makes #1526 non-fatal, though it will not fix this, and there's something else to fix.
2015-10-14build: libdl is needed only for libnss_resolveMichal Schmidt
Not everything needs to link to libdl. dlopen+dlsym are used only by libnss_resolve.
2015-10-14build: don't link everything to libcapMichal Schmidt
The intent of the assignments around the check for libcap was obviously to avoid polluting LIBS. To work properly, LIBS must be restored from the save_LIBS variable. The practical effect is small though, because pretty much everything links with libbasic, which links with CAP_LIBS.
2015-10-14build: fix overlinking to libdwMichal Schmidt
We have two AC_CHECK_LIB checks for libdw. The first one has an empty action-if-found, so it defaults to adding "-ldw" to LIBS. LIBS are applied to everything we build. But only systemd-coredump needs libdw. It already links to ELFUTILS_LIBS correctly in Makefile.am. Drop the first AC_CHECK_LIB check. The second check is sufficient. Q: Don't we already use "-Wl,--as-needed" to eliminate overlinking? A: We do, but it is effective only for executables, not for shared libraries. This is due to a libtool bug: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=347650
2015-10-14Merge pull request #1530 from teg/network-fixes-2Lennart Poettering
networkd/libsystemd-network: collection of trivial patches v2
2015-10-14Merge pull request #1529 from evverx/dbus-props-syslog-fac-priLennart Poettering
core: expose `SyslogFacility` and `SyslogLevel` as properties on dbus
2015-10-14core: expose SyslogLevel and SyslogFacility as properties on dbusEvgeny Vereshchagin
2015-10-14Merge pull request #1560 from evverx/fix-default-choppingLennart Poettering
bus-util: change `Default`-chopping to `Limit`-searching
2015-10-14Merge pull request #1558 from martinpitt/masterLennart Poettering
Fix keymap aliases and add support for Lenovo Thinkpad Yoga S1
2015-10-14Merge pull request #1561 from phomes/unused-functionLennart Poettering
manager: remove unused function
2015-10-14keymap: Drop keyboard-keys-to-name.hMartin Pitt
We don't use that anywhere any more. With the introduction of alias names it also is not a proper mapping any more as several keys (e. g. KEY_COFFEE and KEY_SCREENLOCK) have the same numerical mapping.
2015-10-14keymap: Drop ThinkPad YogaMartin Pitt
The Yoga's firmware sends key events whenever it's being folded or unfolded. These are thus *not* a button for requesting a screen orientation change, just an indication that this already happened. Thus they should not be assigned to "direction", but be ignored. Assigning them to "reserved" does not silence the "unknown key pressed" kernel warning, so there's no point in maintaining a mapping here. Fixes #1440
2015-10-14keymap: Recognize KEY_* aliasesMartin Pitt
linux/input.h contains alias definitions like #define KEY_COFFEE 152 #define KEY_SCREENLOCK KEY_COFFEE #define KEY_ROTATE_DISPLAY 153 #define KEY_DIRECTION KEY_ROTATE_DISPLAY But we ignored these when building keyboard-keys-list.txt. Also allow the value to start with "K" now (for KEY_*), and drop the hardcoded COFFEE → SCREENLOCK aliasing. This fixes assignments to key "direction". Fixes #1151
2015-10-13manager: remove unused functionThomas Hindoe Paaboel Andersen
2015-10-13bus-util: change `Default`-chopping to `Limit`-searchingEvgeny Vereshchagin
See https://github.com/systemd/systemd/pull/1534#commitcomment-13744013 Actually, thinking about this, maybe it would be nicer to actually look for "Limit" in the string rather than chopping off a "Default".... Sounds more generic...
2015-10-13man: /dev/kmsg is not a socketZbigniew Jędrzejewski-Szmek
Fixup for #1542.
2015-10-13Merge pull request #1548 from evverx/journalctl-catalog-ops-fixesLennart Poettering
Fix journalctl --dump-catalog, journalctl --list-catalog
2015-10-13Merge pull request #1534 from evverx/expose-rlimits-on-dbusLennart Poettering
Expose `DefaultLimit*` as properties on dbus
2015-10-13Merge pull request #1542 from keszybz/journal-audit-optionalLennart Poettering
Make journald audit socket maskable
2015-10-13Merge pull request #1544 from keszybz/man-IPv6AcceptRouterAdvertisementsDaniel Mack
man: describe IPv6AcceptRouterAdvertisements= better
2015-10-13Merge pull request #1545 from keszybz/check-for-xsltprocDaniel Mack
build-sys: check for xsltproc when building manpages
2015-10-13Merge pull request #1552 from jsynacek/short-flags-since-untilLennart Poettering
journalctl: introduce short options for --since and --until
2015-10-13journalctl: introduce short options for --since and --untilJan Synacek
Fixes #1514.
2015-10-13journalctl: enable args for --dump-catalog and --list-catalogEvgeny Vereshchagin
`journalctl --dump-catalog ID1 ID2 ...` works fine.
2015-10-13journalctl: open pager for --dump-catalog and --list-catalogEvgeny Vereshchagin
2015-10-12Document journald sockets paths a bit moreZbigniew Jędrzejewski-Szmek
Adding them to the documentation makes it easier to find the right man page for people who are trying to understand where some socket in the filesystem is coming from.
2015-10-12build-sys: check for xsltproc when building manpagesZbigniew Jędrzejewski-Szmek
Only check for xsltproc if it will be used. If not found, complain. https://github.com/systemd/systemd/issues/1521
2015-10-12man: describe IPv6AcceptRouterAdvertisements= betterZbigniew Jędrzejewski-Szmek
With the previous description it wasn't clear that the kernel default is being described. Add link to kernel docs.
2015-10-12journald: make audit socket optionalZbigniew Jędrzejewski-Szmek
If we were given some sockets through socket activation, and audit socket is not among them, do not try to open it. This way, if the socket unit is disabled, we will not receive audit events. https://bugzilla.redhat.com/show_bug.cgi?id=1227379
2015-10-12Merge pull request #1536 from jsynacek/masterDaniel Mack
po/da: use unix line endings
2015-10-12Merge pull request #1540 from zonque/cmsgDavid Herrmann
sd-daemon: wipe out memory before using CMSG_NXTHDR()
2015-10-12sd-daemon: wipe out memory before using CMSG_NXTHDR()Daniel Mack
CMSG_NXTHDR() checks for cmsg->cmsg_len *after* it increased the pointer. While this makes sense for parsing received messages, that's a pitfall for code crafting messages with this macro. Wipe out the allocated memory to fix this.
2015-10-12po/da: use unix line endingsJan Synacek
2015-10-12Merge pull request #1532 from phomes/trivial-fixesDaniel Mack
Trivial fixes
2015-10-12core: manager: expose DefaultLimit* as properties on dbusEvgeny Vereshchagin
2015-10-12bus-util: add bus_property_get_rlimit (move from core/dbus_execute)Evgeny Vereshchagin
2015-10-11shared: remove an unused variableThomas Hindoe Paaboel Andersen
2015-10-11rfkill: consistently use = in designatorThomas Hindoe Paaboel Andersen
2015-10-11networkd: address - introduce address_is_ready()Tom Gundersen
Checks that a given address is not tentative nor deprecated.
2015-10-11networkd: make address_{establish,release}() staticTom Gundersen
No need to expose these functions, but rather call them from address_{add,drop}.
2015-10-11networkd: manager - avoid unnecessary memory allocationTom Gundersen
Don't allocate Address objects only to free them again when processing rtnl events.
2015-10-11networkd: address - store active addresses in a Set rather than a ListTom Gundersen
We need to be able to look these things up quickly as we will be updating them continuously and there can in principle be many of them.
2015-10-11networkd: address - process in manager.c rather than link.cTom Gundersen
2015-10-11networkd: link - refactor (un)refTom Gundersen
2015-10-11networkd: link - fix error handling in link_add()Tom Gundersen
2015-10-11sd-netlink: add support for subscribing to ROUTE messagesTom Gundersen
Also, explicitly don't support subscribing to GET or SET messages, as these will never be emitted by the kernel.