rpm/systemd
SHA256
1
0
Fork 0
forked from pool/systemd
Code Pull Requests Activity
00a36bdb9a
systemd/0001-core-introduce-new-KillMode-mixed-which-sends-SIGTER.patch
Stephan Kulow 8f6010ffdf Accepting request 215598 from Base:System 
- Change patch 0001-add-hdflush-for-reboot-or-hddown-for-poweroff.patch
  to skip already by the kernel managed devices 

- fix timeout stopping user@.service (bnc#841544)
  * 0001-core-introduce-new-KillMode-mixed-which-sends-SIGTER.patch
  * 0002-service-allow-KillMode-mixed-in-conjunction-with-PAM.patch
  * 0003-core-make-sure-to-always-go-through-both-SIGTERM-and.patch

- Add patch 0001-upstream-systemctl-halt-reboot-error-handling.patch
  to be able to detect if the sysctl reboot() returns.
- Add patch 0001-add-hdflush-for-reboot-or-hddown-for-poweroff.patch
  A check for unmaintained disk like devices is added to be able to
  flush and maybe shut them down.  Also the missing sync() system
  call is added for the direct halt/reboot systemctl command. Then
  the system halt is used as fallback if poweroff fails for both
  the direct poweroff systemctl command as well as for the
  systemd-shutdown utility.

- Make systemd-mini build 

- Make requires bash-completion a recommends

- Add patch 1017-skip-native-unit-handling-if-sysv-already-handled.patch
  to avoid that enabled boot scripts will be handled as unit files
  by systemctl status command (bnc#818044)

- Drop patch 1017-enforce-sufficient-shutdown-warnings.patch
  as the original code behaves exactly as the shutdown code of
  the old SysVinit (bnc#750845)
- Rename support-powerfail-with-powerstatus.patch to

OBS-URL: https://build.opensuse.org/request/show/215598
OBS-URL: https://build.opensuse.org/package/show/openSUSE:Factory/systemd?expand=0&rev=173
2014-01-31 12:36:06 +00:00

226 lines
11 KiB
Diff
Raw Blame History

From 6fa7e1a944a2dbb89e794ad0f9da5d0fda5dc4a9 Mon Sep 17 00:00:00 2001
From: Lennart Poettering <lennart@poettering.net>
Date: Wed, 29 Jan 2014 13:38:55 +0100
Subject: [PATCH 1/3] core: introduce new KillMode=mixed which sends SIGTERM
only to the main process, but SIGKILL to all daemon processes
This should fix some race with terminating systemd --user, where the
system systemd instance might race against the user systemd instance
when sending SIGTERM.
---
man/systemd.kill.xml | 77 +++++++++++++++++++++++++++++++++-----------------
src/core/kill.c | 1 +
src/core/kill.h | 1 +
src/core/unit.c | 3 +-
units/user@.service.in | 1 +
5 files changed, 56 insertions(+), 27 deletions(-)
diff --git a/man/systemd.kill.xml b/man/systemd.kill.xml
index 1b10fba..a4009aa 100644
--- a/man/systemd.kill.xml
+++ b/man/systemd.kill.xml
@@ -44,39 +44,44 @@
<refnamediv>
<refname>systemd.kill</refname>
- <refpurpose>Kill environment configuration</refpurpose>
+ <refpurpose>Process killing procedure
+ configuration</refpurpose>
</refnamediv>
<refsynopsisdiv>
<para><filename><replaceable>service</replaceable>.service</filename>,
<filename><replaceable>socket</replaceable>.socket</filename>,
<filename><replaceable>mount</replaceable>.mount</filename>,
- <filename><replaceable>swap</replaceable>.swap</filename></para>
+ <filename><replaceable>swap</replaceable>.swap</filename>,
+ <filename><replaceable>scope</replaceable>.scope</filename></para>
</refsynopsisdiv>
<refsect1>
<title>Description</title>
<para>Unit configuration files for services, sockets,
- mount points and swap devices share a subset of
- configuration options which define the process killing
- parameters of spawned processes.</para>
+ mount points, swap devices and scopes share a subset
+ of configuration options which define the
+ killing procedure of processes belonging to the unit.</para>
<para>This man page lists the configuration options
- shared by these four unit types. See
+ shared by these five unit types. See
<citerefentry><refentrytitle>systemd.unit</refentrytitle><manvolnum>5</manvolnum></citerefentry>
- for the common options of all unit configuration
- files, and
+ for the common options shared by all unit
+ configuration files, and
<citerefentry><refentrytitle>systemd.service</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
<citerefentry><refentrytitle>systemd.socket</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
- <citerefentry><refentrytitle>systemd.swap</refentrytitle><manvolnum>5</manvolnum></citerefentry>
- and
+ <citerefentry><refentrytitle>systemd.swap</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
<citerefentry><refentrytitle>systemd.mount</refentrytitle><manvolnum>5</manvolnum></citerefentry>
- for more information on the specific unit
- configuration files. The execution specific
+ and
+ <citerefentry><refentrytitle>systemd.scope</refentrytitle><manvolnum>5</manvolnum></citerefentry>
+ for more information on the configuration file options
+ specific to each unit type.</para>
+
+ <para>The kill procedure
configuration options are configured in the [Service],
- [Socket], [Mount], or [Swap] section, depending on the unit
- type.</para>
+ [Socket], [Mount] or [Swap] section, depending on the
+ unit type.</para>
</refsect1>
<refsect1>
@@ -87,32 +92,40 @@
<varlistentry>
<term><varname>KillMode=</varname></term>
<listitem><para>Specifies how
- processes of this service shall be
+ processes of this unit shall be
killed. One of
<option>control-group</option>,
<option>process</option>,
+ <option>mixed</option>,
<option>none</option>.</para>
<para>If set to
<option>control-group</option>, all
remaining processes in the control
- group of this unit will be terminated
- on unit stop (for services: after the
+ group of this unit will be killed on
+ unit stop (for services: after the
stop command is executed, as
configured with
<varname>ExecStop=</varname>). If set
to <option>process</option>, only the
main process itself is killed. If set
- to <option>none</option>, no process is
+ to <option>mixed</option> the
+ <constant>SIGTERM</constant> signal
+ (see below) is sent to the main
+ process while the subsequent
+ <constant>SIGKILL</constant> signal
+ (see below) is sent to all remaining
+ processes of the unit's control
+ group. If set to
+ <option>none</option>, no process is
killed. In this case only the stop
- command will be executed on unit
- stop, but no process be killed
+ command will be executed on unit stop,
+ but no process be killed
otherwise. Processes remaining alive
after stop are left in their control
group and the control group continues
to exist after stop unless it is
- empty. Defaults to
- <option>control-group</option>.</para>
+ empty.</para>
<para>Processes will first be
terminated via
@@ -133,14 +146,24 @@
option). See
<citerefentry><refentrytitle>kill</refentrytitle><manvolnum>2</manvolnum></citerefentry>
for more
- information.</para></listitem>
+ information.</para>
+
+ <para>Defaults to
+ <option>control-group</option>.</para></listitem>
</varlistentry>
<varlistentry>
<term><varname>KillSignal=</varname></term>
<listitem><para>Specifies which signal
- to use when killing a
- service. Defaults to <constant>SIGTERM</constant>.
+ to use when killing a service. This
+ controls the signal that is sent as
+ first step of shutting down a unit
+ (see above), and is usually followed
+ by <constant>SIGKILL</constant> (see
+ above and below). For a list of valid
+ signals, see
+ <citerefentry><refentrytitle>signal</refentrytitle><manvolnum>7</manvolnum></citerefentry>. Defaults
+ to <constant>SIGTERM</constant>.
</para></listitem>
</varlistentry>
@@ -184,7 +207,9 @@
<citerefentry><refentrytitle>systemd.swap</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
<citerefentry><refentrytitle>systemd.mount</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
<citerefentry><refentrytitle>systemd.exec</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
- <citerefentry><refentrytitle>systemd.directives</refentrytitle><manvolnum>7</manvolnum></citerefentry>
+ <citerefentry><refentrytitle>systemd.directives</refentrytitle><manvolnum>7</manvolnum></citerefentry>,
+ <citerefentry><refentrytitle>kill</refentrytitle><manvolnum>2</manvolnum></citerefentry>,
+ <citerefentry><refentrytitle>signal</refentrytitle><manvolnum>7</manvolnum></citerefentry>
</para>
</refsect1>
diff --git a/src/core/kill.c b/src/core/kill.c
index ea947c2..4271346 100644
--- a/src/core/kill.c
+++ b/src/core/kill.c
@@ -52,6 +52,7 @@ void kill_context_dump(KillContext *c, FILE *f, const char *prefix) {
static const char* const kill_mode_table[_KILL_MODE_MAX] = {
[KILL_CONTROL_GROUP] = "control-group",
[KILL_PROCESS] = "process",
+ [KILL_MIXED] = "mixed",
[KILL_NONE] = "none"
};
diff --git a/src/core/kill.h b/src/core/kill.h
index 41773f0..d5f125f 100644
--- a/src/core/kill.h
+++ b/src/core/kill.h
@@ -32,6 +32,7 @@ typedef enum KillMode {
/* The kill mode is a property of a unit. */
KILL_CONTROL_GROUP = 0,
KILL_PROCESS,
+ KILL_MIXED,
KILL_NONE,
_KILL_MODE_MAX,
_KILL_MODE_INVALID = -1
diff --git a/src/core/unit.c b/src/core/unit.c
index 4b97710..0b10e57 100644
--- a/src/core/unit.c
+++ b/src/core/unit.c
@@ -3007,7 +3007,7 @@ int unit_kill_context(
}
}
- if (c->kill_mode == KILL_CONTROL_GROUP && u->cgroup_path) {
+ if ((c->kill_mode == KILL_CONTROL_GROUP || (c->kill_mode == KILL_MIXED && sigkill)) && u->cgroup_path) {
_cleanup_set_free_ Set *pid_set = NULL;
/* Exclude the main/control pids from being killed via the cgroup */
@@ -3021,6 +3021,7 @@ int unit_kill_context(
log_warning_unit(u->id, "Failed to kill control group: %s", strerror(-r));
} else if (r > 0) {
wait_for_exit = true;
+
if (c->send_sighup) {
set_free(pid_set);
diff --git a/units/user@.service.in b/units/user@.service.in
index 3718a57..3bb8696 100644
--- a/units/user@.service.in
+++ b/units/user@.service.in
@@ -17,3 +17,4 @@ Environment=SHELL=%s
ExecStart=-@rootlibexecdir@/systemd --user
Environment=DBUS_SESSION_BUS_ADDRESS=unix:path=/run/user/%I/dbus/user_bus_socket
Slice=user-%i.slice
+KillMode=mixed
--
1.8.4
View Git Blame Copy Permalink