added in src/lock.c to replace call to lxc_get_lock and lxc_put_lock,
when only need to check if container is active or not.
Signed-off-by: Michel Normand <normand@fr.ibm.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
This is adding a stderror log appender that is used as default one.
Signed-off-by: Michel Normand <normand@fr.ibm.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
the support of multiple appenders need to associate
one va_list per appender.
This is the purpose of this patch to copy the va_list
before to call the appender.
Signed-off-by: Michel Normand <normand@fr.ibm.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Instead of messing with the netlink messages, let's
use the abstract unix socket and assume we will have
a single receiver and multiple sender.
With this patch mcr-wait and mcr-monitor are mutually
exclusive... for the moment.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
and report error message as soon as detected error in these two functions
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Signed-off-by: Michel Normand <normand@fr.ibm.com>
this is adding -o and -l options to all cli of lxc
Signed-off-by: Michel Normand <normand@fr.ibm.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
do it in all cli, except:
* the usage functions that will be removed by another patch
* the lxc_init that need more work
* the lxc_priority that is not used anymore
Signed-off-by: Michel Normand <normand@fr.ibm.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
this is required to log error where needed
and to to avoid additionnal log later on.
Signed-off-by: Michel Normand <normand@fr.ibm.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
required to support either:
lxc-cgroup -n foo subsystem value
lxc-cgroup -nfoo subsystem value
Signed-off-by: Michel Normand <normand@fr.ibm.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
replace the specific clone options by a -s option
with the Ored list of namespaces to be cloned.
take the opportunity of this change to have
the -u <uid> option to not automatically set
the clone flag to CLONE_NEWUSER but to be dependant
of the new -s USER option.
Signed-off-by: Michel Normand <normand@fr.ibm.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Do it except for the usage functions that will be removed by another patch
This is a follow on a previous patch that treated
all src/lxc/lxc_*.c but this file.
Signed-off-by: Michel Normand <normand@fr.ibm.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
avoid segfault when invalid parameter
git-unshare -u 666666
and improve error reporting
Signed-off-by: Michel Normand <normand@fr.ibm.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
When no tty number is specified in the command line,
let the tty service to provide choose one available
tty and provide this one.
The documentation is updated wrt this modification and
I did a little fix to generate the date of the documentation.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Export the fine grain api of lxc to be usable for external
component which wants to have more control on the container.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Define a handler to store the resource informations for the container.
That allows to factor out some code and will help to encapsulate a
little more the 'start' code which is too big.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
This reverts commit 2c9bd0073f.
Pointless and useless right now and certainly wrong as several
threads in the library may want to log to the same, to be defined
later.
When auto-regen is done at the make time, eg. when the configure.in
script has been modified, the configure is not correctly generated
because the 'config' directory is missing for aclocal, and that makes
the generation to fail due to the missing macros defined in config
directory.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Add the autogen script to the tarball so people using the
dist version can patch the auto* files and regenerate them
directly for the dist tarball.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Reported-by: Michael K Johnson <johnsonm@rpath.com>
By convention the script calling the auto tools is named
autogen.sh. I rename 'bootstrap' to 'autogen.sh' in order to
stick to this convention.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Reported-by: Michael K Johnson <johnsonm@rpath.com>
this is a required patch to be able to compile lxc on a RedHat 5 update 1
which has such define.
This is also required for the F_DUPFD_CLOEXEC define
Signed-off-by: Michel Normand <normand@fr.ibm.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
some headers and macros are now redundant
Signed-off-by: Cedric Le Goater <legoater@free.fr>
Acked-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
This patch defines a new log category in each file using logging and
replaced the former logging routines with the new ones.
Signed-off-by: Cedric Le Goater <legoater@free.fr>
Acked-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
lxc_log_init() should be called in each main() of a command
to define the default log priority and log file.
Signed-off-by: Cedric Le Goater <legoater@free.fr>
Acked-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
exclude generated tarball from git managed files
Signed-off-by: Michel Normand <normand@fr.ibm.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
This is required to be able to make the build
on a machine that mount lxc source but not all
machine filesystem.
.
Signed-off-by: Michel Normand <normand@fr.ibm.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Set the close on exec flag on the pty fd so they are automatically
closed when execing the container.
Signed-off-by: Môshe van der Sterre <me@moshe.nl>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Hello,
It took me some thinking to figure out DEVPTS_MULTIPLE_INSTANCES was disabled.
Maybe checking for it in lxc-checkconfig will be helpfull to others.
Greetings,
Môshe van der Sterre
Signed-off-by: Môshe van der Sterre <moshevds@gmail.com>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
At present the 'init' file is deleted from the 'stop' function and
I don't remember why it is done in this place :)
The 'init' file is removed from the 'start' when the first process
has been deleted.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
As the state of the container is monitored with the netlink,
the file state is no longer used to watch the state changes.
The previous hack, which adds a tempo of 200ms, is removed and
that makes the container being created, started, stopped, destroyed
faster, we gain 1 second in a container lifecycle. One second is
too much if we launch thousand of containers, one by one like for
example in a tests suite.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
This patch fixes compile warnings: ignoring return value of function,
declared with attribute warn_unused_result, and adds error handling.
Signed-off-by: Ryousei Takano <takano-ryousei@aist.go.jp>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
The current code assumes that localstatedir is equal to $(prefix)/var,
thus failing for example on debian, where prefix is /usr and
localstatedir is /var. This patch fixes this by expanding LXCPATH just
once in configure.ac to $(localstatedir)/lib/lxc and expanding that
variable everywhere else.
install-exec-local is changed to just do one mkdir -p, and taking into
account of the DESTDIR variable, user for example for packaging.
Changing the permission of LXCPATH is done in lxc-setcap
Signed-off-by: Guido Trotter <ultrotter@quaqua.net>
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
