Commit 3b1b3f6e authored by Li Zefan's avatar Li Zefan Committed by Linus Torvalds

freezer_cg: disable writing freezer.state of root cgroup

With this change, control file 'freezer.state' doesn't exist in root
cgroup, making root cgroup unfreezable.

I think it's reasonable to disallow freeze tasks in the root cgroup.  And
then we can avoid fork overhead when freezer subsystem is compiled but not
used.

Also make writing invalid value to freezer.state returns EINVAL rather
than EIO.  This is more consistent with other cgroup subsystem.
Signed-off-by: default avatarLi Zefan <lizf@cn.fujitsu.com>
Acked-by: default avatarPaul Menage <menage@google.com>
Cc: Cedric Le Goater <clg@fr.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: "Serge E. Hallyn" <serue@us.ibm.com>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
parent 68744676
The cgroup freezer is useful to batch job management system which start The cgroup freezer is useful to batch job management system which start
and stop sets of tasks in order to schedule the resources of a machine and stop sets of tasks in order to schedule the resources of a machine
according to the desires of a system administrator. This sort of program according to the desires of a system administrator. This sort of program
is often used on HPC clusters to schedule access to the cluster as a is often used on HPC clusters to schedule access to the cluster as a
...@@ -6,7 +6,7 @@ whole. The cgroup freezer uses cgroups to describe the set of tasks to ...@@ -6,7 +6,7 @@ whole. The cgroup freezer uses cgroups to describe the set of tasks to
be started/stopped by the batch job management system. It also provides be started/stopped by the batch job management system. It also provides
a means to start and stop the tasks composing the job. a means to start and stop the tasks composing the job.
The cgroup freezer will also be useful for checkpointing running groups The cgroup freezer will also be useful for checkpointing running groups
of tasks. The freezer allows the checkpoint code to obtain a consistent of tasks. The freezer allows the checkpoint code to obtain a consistent
image of the tasks by attempting to force the tasks in a cgroup into a image of the tasks by attempting to force the tasks in a cgroup into a
quiescent state. Once the tasks are quiescent another task can quiescent state. Once the tasks are quiescent another task can
...@@ -16,7 +16,7 @@ recoverable error occur. This also allows the checkpointed tasks to be ...@@ -16,7 +16,7 @@ recoverable error occur. This also allows the checkpointed tasks to be
migrated between nodes in a cluster by copying the gathered information migrated between nodes in a cluster by copying the gathered information
to another node and restarting the tasks there. to another node and restarting the tasks there.
Sequences of SIGSTOP and SIGCONT are not always sufficient for stopping Sequences of SIGSTOP and SIGCONT are not always sufficient for stopping
and resuming tasks in userspace. Both of these signals are observable and resuming tasks in userspace. Both of these signals are observable
from within the tasks we wish to freeze. While SIGSTOP cannot be caught, from within the tasks we wish to freeze. While SIGSTOP cannot be caught,
blocked, or ignored it can be seen by waiting or ptracing parent tasks. blocked, or ignored it can be seen by waiting or ptracing parent tasks.
...@@ -37,26 +37,29 @@ demonstrate this problem using nested bash shells: ...@@ -37,26 +37,29 @@ demonstrate this problem using nested bash shells:
<at this point 16990 exits and causes 16644 to exit too> <at this point 16990 exits and causes 16644 to exit too>
This happens because bash can observe both signals and choose how it This happens because bash can observe both signals and choose how it
responds to them. responds to them.
Another example of a program which catches and responds to these Another example of a program which catches and responds to these
signals is gdb. In fact any program designed to use ptrace is likely to signals is gdb. In fact any program designed to use ptrace is likely to
have a problem with this method of stopping and resuming tasks. have a problem with this method of stopping and resuming tasks.
In contrast, the cgroup freezer uses the kernel freezer code to In contrast, the cgroup freezer uses the kernel freezer code to
prevent the freeze/unfreeze cycle from becoming visible to the tasks prevent the freeze/unfreeze cycle from becoming visible to the tasks
being frozen. This allows the bash example above and gdb to run as being frozen. This allows the bash example above and gdb to run as
expected. expected.
The freezer subsystem in the container filesystem defines a file named The freezer subsystem in the container filesystem defines a file named
freezer.state. Writing "FROZEN" to the state file will freeze all tasks in the freezer.state. Writing "FROZEN" to the state file will freeze all tasks in the
cgroup. Subsequently writing "THAWED" will unfreeze the tasks in the cgroup. cgroup. Subsequently writing "THAWED" will unfreeze the tasks in the cgroup.
Reading will return the current state. Reading will return the current state.
Note freezer.state doesn't exist in root cgroup, which means root cgroup
is non-freezable.
* Examples of usage : * Examples of usage :
# mkdir /containers/freezer # mkdir /containers
# mount -t cgroup -ofreezer freezer /containers # mount -t cgroup -ofreezer freezer /containers
# mkdir /containers/0 # mkdir /containers/0
# echo $some_pid > /containers/0/tasks # echo $some_pid > /containers/0/tasks
...@@ -94,6 +97,6 @@ things happens: ...@@ -94,6 +97,6 @@ things happens:
the freezer.state file the freezer.state file
2) Userspace retries the freezing operation by writing "FROZEN" to 2) Userspace retries the freezing operation by writing "FROZEN" to
the freezer.state file (writing "FREEZING" is not legal the freezer.state file (writing "FREEZING" is not legal
and returns EIO) and returns EINVAL)
3) The tasks that blocked the cgroup from entering the "FROZEN" 3) The tasks that blocked the cgroup from entering the "FROZEN"
state disappear from the cgroup's set of tasks. state disappear from the cgroup's set of tasks.
...@@ -192,6 +192,13 @@ static void freezer_fork(struct cgroup_subsys *ss, struct task_struct *task) ...@@ -192,6 +192,13 @@ static void freezer_fork(struct cgroup_subsys *ss, struct task_struct *task)
*/ */
freezer = task_freezer(task); freezer = task_freezer(task);
/*
* The root cgroup is non-freezable, so we can skip the
* following check.
*/
if (!freezer->css.cgroup->parent)
return;
spin_lock_irq(&freezer->lock); spin_lock_irq(&freezer->lock);
BUG_ON(freezer->state == CGROUP_FROZEN); BUG_ON(freezer->state == CGROUP_FROZEN);
...@@ -335,7 +342,7 @@ static int freezer_write(struct cgroup *cgroup, ...@@ -335,7 +342,7 @@ static int freezer_write(struct cgroup *cgroup,
else if (strcmp(buffer, freezer_state_strs[CGROUP_FROZEN]) == 0) else if (strcmp(buffer, freezer_state_strs[CGROUP_FROZEN]) == 0)
goal_state = CGROUP_FROZEN; goal_state = CGROUP_FROZEN;
else else
return -EIO; return -EINVAL;
if (!cgroup_lock_live_group(cgroup)) if (!cgroup_lock_live_group(cgroup))
return -ENODEV; return -ENODEV;
...@@ -354,6 +361,8 @@ static struct cftype files[] = { ...@@ -354,6 +361,8 @@ static struct cftype files[] = {
static int freezer_populate(struct cgroup_subsys *ss, struct cgroup *cgroup) static int freezer_populate(struct cgroup_subsys *ss, struct cgroup *cgroup)
{ {
if (!cgroup->parent)
return 0;
return cgroup_add_files(cgroup, ss, files, ARRAY_SIZE(files)); return cgroup_add_files(cgroup, ss, files, ARRAY_SIZE(files));
} }
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment