Re: [PATCH, v5 2/2] cgroups: introduce timer slack subsystem

[Balbir Singh]

* Kirill A. Shutemov <kirill@xxxxxxxxxxxxx> [2011-02-07 12:57:30]:

> On Mon, Feb 07, 2011 at 03:36:24PM +0530, Balbir Singh wrote:
> > * Kirill A. Shutsemov <kirill@xxxxxxxxxxxxx> [2011-02-07 11:46:02]:
> > 
> > > From: Kirill A. Shutemov <kirill@xxxxxxxxxxxxx>
> > > 
> > > Provides a way of tasks grouping by timer slack value. Introduces per
> > > cgroup max and min timer slack value. When a task attaches to a cgroup,
> > > its timer slack value adjusts (if needed) to fit min-max range.
> > > 
> > > It also provides a way to set timer slack value for all tasks in the
> > > cgroup at once.
> > > 
> > > This functionality is useful in mobile devices where certain background
> > > apps are attached to a cgroup and minimum wakeups are desired.
> > > 
> > > Idea-by: Jacob Pan <jacob.jun.pan@xxxxxxxxxxxxxxx>
> > > Signed-off-by: Kirill A. Shutemov <kirill@xxxxxxxxxxxxx>
> > > ---
> > >  include/linux/cgroup_subsys.h |    6 +
> > >  include/linux/init_task.h     |    4 +-
> > >  init/Kconfig                  |   10 ++
> > >  kernel/Makefile               |    1 +
> > >  kernel/cgroup_timer_slack.c   |  281 +++++++++++++++++++++++++++++++++++++++++
> > >  kernel/sys.c                  |   19 ++-
> > >  6 files changed, 314 insertions(+), 7 deletions(-)
> > >  create mode 100644 kernel/cgroup_timer_slack.c
> > > 
> > > diff --git a/include/linux/cgroup_subsys.h b/include/linux/cgroup_subsys.h
> > > index ccefff0..e399228 100644
> > > --- a/include/linux/cgroup_subsys.h
> > > +++ b/include/linux/cgroup_subsys.h
> > > @@ -66,3 +66,9 @@ SUBSYS(blkio)
> > >  #endif
> > > 
> > >  /* */
> > > +
> > > +#ifdef CONFIG_CGROUP_TIMER_SLACK
> > > +SUBSYS(timer_slack)
> > > +#endif
> > > +
> > > +/* */
> > > diff --git a/include/linux/init_task.h b/include/linux/init_task.h
> > > index caa151f..48eca8f 100644
> > > --- a/include/linux/init_task.h
> > > +++ b/include/linux/init_task.h
> > > @@ -124,6 +124,8 @@ extern struct cred init_cred;
> > >  # define INIT_PERF_EVENTS(tsk)
> > >  #endif
> > > 
> > > +#define TIMER_SLACK_NS_DEFAULT 50000
> > > +
> > >  /*
> > >   *  INIT_TASK is used to set up the first task table, touch at
> > >   * your own risk!. Base=0, limit=0x1fffff (=2MB)
> > > @@ -177,7 +179,7 @@ extern struct cred init_cred;
> > >  	.cpu_timers	= INIT_CPU_TIMERS(tsk.cpu_timers),		\
> > >  	.fs_excl	= ATOMIC_INIT(0),				\
> > >  	.pi_lock	= __RAW_SPIN_LOCK_UNLOCKED(tsk.pi_lock),	\
> > > -	.timer_slack_ns = 50000, /* 50 usec default slack */		\
> > > +	.timer_slack_ns = TIMER_SLACK_NS_DEFAULT,			\
> > >  	.pids = {							\
> > >  		[PIDTYPE_PID]  = INIT_PID_LINK(PIDTYPE_PID),		\
> > >  		[PIDTYPE_PGID] = INIT_PID_LINK(PIDTYPE_PGID),		\
> > > diff --git a/init/Kconfig b/init/Kconfig
> > > index be788c0..6cf465f 100644
> > > --- a/init/Kconfig
> > > +++ b/init/Kconfig
> > > @@ -596,6 +596,16 @@ config CGROUP_FREEZER
> > >  	  Provides a way to freeze and unfreeze all tasks in a
> > >  	  cgroup.
> > > 
> > > +config CGROUP_TIMER_SLACK
> > > +	tristate "Timer slack cgroup subsystem"
> > > +	help
> > > +	  Provides a way of tasks grouping by timer slack value.
> > > +	  Introduces per cgroup timer slack value which will override
> > > +	  the default timer slack value once a task is attached to a
> > > +	  cgroup.
> > > +	  It's useful in mobile devices where certain background apps
> > > +	  are attached to a cgroup and combined wakeups are desired.
> > > +
> > >  config CGROUP_DEVICE
> > >  	bool "Device controller for cgroups"
> > >  	help
> > > diff --git a/kernel/Makefile b/kernel/Makefile
> > > index 353d3fe..0b60239 100644
> > > --- a/kernel/Makefile
> > > +++ b/kernel/Makefile
> > > @@ -61,6 +61,7 @@ obj-$(CONFIG_BACKTRACE_SELF_TEST) += backtracetest.o
> > >  obj-$(CONFIG_COMPAT) += compat.o
> > >  obj-$(CONFIG_CGROUPS) += cgroup.o
> > >  obj-$(CONFIG_CGROUP_FREEZER) += cgroup_freezer.o
> > > +obj-$(CONFIG_CGROUP_TIMER_SLACK) += cgroup_timer_slack.o
> > >  obj-$(CONFIG_CPUSETS) += cpuset.o
> > >  obj-$(CONFIG_CGROUP_NS) += ns_cgroup.o
> > >  obj-$(CONFIG_UTS_NS) += utsname.o
> > > diff --git a/kernel/cgroup_timer_slack.c b/kernel/cgroup_timer_slack.c
> > > new file mode 100644
> > > index 0000000..609e1f1
> > > --- /dev/null
> > > +++ b/kernel/cgroup_timer_slack.c
> > > @@ -0,0 +1,281 @@
> > > +/*
> > > + * cgroup_timer_slack.c - control group timer slack subsystem
> > > + *
> > > + * Copyright Nokia Corparation, 2011
> > > + * Author: Kirill A. Shutemov
> > > + *
> > > + * This program is free software; you can redistribute it and/or modify
> > > + * it under the terms of the GNU General Public License as published by
> > > + * the Free Software Foundation; either version 2 of the License, or
> > > + * (at your option) any later version.
> > > + *
> > > + * This program is distributed in the hope that it will be useful,
> > > + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> > > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> > > + * GNU General Public License for more details.
> > > + */
> > > +#include <linux/cgroup.h>
> > > +#include <linux/init_task.h>
> > > +#include <linux/module.h>
> > > +#include <linux/slab.h>
> > > +#include <linux/rcupdate.h>
> > > +
> > > +struct cgroup_subsys timer_slack_subsys;
> > > +struct timer_slack_cgroup {
> > > +	struct cgroup_subsys_state css;
> > > +	unsigned long min_slack_ns;
> > > +	unsigned long max_slack_ns;
> > > +};
> > > +
> > > +enum {
> > > +	TIMER_SLACK_MIN,
> > > +	TIMER_SLACK_MAX,
> > > +};
> > > +
> > > +extern int (*timer_slack_check)(struct task_struct *task,
> > > +		unsigned long slack_ns);
> > > +
> > > +static struct timer_slack_cgroup *cgroup_to_tslack_cgroup(struct cgroup *cgroup)
> > > +{
> > > +	struct cgroup_subsys_state *css;
> > > +
> > > +	css = cgroup_subsys_state(cgroup, timer_slack_subsys.subsys_id);
> > > +	return container_of(css, struct timer_slack_cgroup, css);
> > > +}
> > > +
> > > +static int is_timer_slack_allowed(struct timer_slack_cgroup *tslack_cgroup,
> > > +		unsigned long slack_ns)
> > > +{
> > > +	if (slack_ns < tslack_cgroup->min_slack_ns ||
> > > +			slack_ns > tslack_cgroup->max_slack_ns)
> > > +		return false;
> > > +	return true;
> > > +}
> > > +
> > > +static int cgroup_timer_slack_check(struct task_struct *task,
> > > +		unsigned long slack_ns)
> > > +{
> > > +	struct cgroup_subsys_state *css;
> > > +	struct timer_slack_cgroup *tslack_cgroup;
> > > +
> > > +	/* XXX: lockdep false positive? */
> > > +	rcu_read_lock();
> > > +	css = task_subsys_state(task, timer_slack_subsys.subsys_id);
> > > +	tslack_cgroup = container_of(css, struct timer_slack_cgroup, css);
> > 
> > Shouldn't the rcu_read_unlock() be after the check of is_timer_slack_allowed()?
> 
> I'm not sure if this rcu_read_lock() is needed. See freezer_fork().
> 
> > > +	rcu_read_unlock();
> > > +
> > > +	if (!is_timer_slack_allowed(tslack_cgroup, slack_ns))
> > > +		return -EPERM;
> > > +	return 0;
> > > +}
> > > +
> > > +static struct cgroup_subsys_state *
> > > +tslack_cgroup_create(struct cgroup_subsys *subsys, struct cgroup *cgroup)
> > > +{
> > > +	struct timer_slack_cgroup *tslack_cgroup;
> > > +
> > > +	tslack_cgroup = kmalloc(sizeof(*tslack_cgroup), GFP_KERNEL);
> > > +	if (!tslack_cgroup)
> > > +		return ERR_PTR(-ENOMEM);
> > > +
> > > +	if (cgroup->parent) {
> > > +		struct timer_slack_cgroup *parent;
> > > +		parent = cgroup_to_tslack_cgroup(cgroup->parent);
> > > +		tslack_cgroup->min_slack_ns = parent->min_slack_ns;
> > > +		tslack_cgroup->max_slack_ns = parent->max_slack_ns;
> > > +	} else {
> > > +		tslack_cgroup->min_slack_ns = 0UL;
> > > +		tslack_cgroup->max_slack_ns = ULONG_MAX;
> > > +	}
> > > +
> > > +	return &tslack_cgroup->css;
> > > +}
> > > +
> > > +static void tslack_cgroup_destroy(struct cgroup_subsys *subsys,
> > > +		struct cgroup *cgroup)
> > > +{
> > > +	kfree(cgroup_to_tslack_cgroup(cgroup));
> > > +}
> > > +
> > > +/*
> > > + * Adjust ->timer_slack_ns and ->default_max_slack_ns of the task to fit
> > > + * limits of the cgroup.
> > > + */
> > > +static void tslack_adjust_task(struct timer_slack_cgroup *tslack_cgroup,
> > > +		struct task_struct *tsk)
> > > +{
> > > +	if (tslack_cgroup->min_slack_ns > tsk->timer_slack_ns)
> > > +		tsk->timer_slack_ns = tslack_cgroup->min_slack_ns;
> > > +	else if (tslack_cgroup->max_slack_ns < tsk->timer_slack_ns)
> > > +		tsk->timer_slack_ns = tslack_cgroup->max_slack_ns;
> > > +
> > > +	if (tslack_cgroup->min_slack_ns > tsk->default_timer_slack_ns)
> > > +		tsk->default_timer_slack_ns = tslack_cgroup->min_slack_ns;
> > > +	else if (tslack_cgroup->max_slack_ns < tsk->default_timer_slack_ns)
> > > +		tsk->default_timer_slack_ns = tslack_cgroup->max_slack_ns;
> > > +}
> > > +
> > > +static void tslack_cgroup_attach(struct cgroup_subsys *subsys,
> > > +		struct cgroup *cgroup, struct cgroup *prev,
> > > +		struct task_struct *tsk, bool threadgroup)
> > > +{
> > > +	tslack_adjust_task(cgroup_to_tslack_cgroup(cgroup), tsk);
> > > +}
> > > +
> > > +static int tslack_write_set_slack_ns(struct cgroup *cgroup, struct cftype *cft,
> > > +		u64 val)
> > > +{
> > > +	struct timer_slack_cgroup *tslack_cgroup;
> > > +	struct cgroup_iter it;
> > > +	struct task_struct *task;
> > > +
> > > +	tslack_cgroup = cgroup_to_tslack_cgroup(cgroup);
> > > +	if (!is_timer_slack_allowed(cgroup_to_tslack_cgroup(cgroup), val))
> > > +		return -EPERM;
> > > +
> > > +	/* Change timer slack value for all tasks in the cgroup */
> > > +	cgroup_iter_start(cgroup, &it);
> > > +	while ((task = cgroup_iter_next(cgroup, &it)))
> > > +		task->timer_slack_ns = val;
> > > +	cgroup_iter_end(cgroup, &it);
> > 
> > OK, I've got questions, are the following possible
> > 
> > 1. Assign a task to a cgroup based on timer slack, it is accepted
> > based on min and max values
> 
> Hm.. Currently, there is no such functionality. Do we need it? What's use
> case?
> 
> > 2. We update the slack
> > 
> > Can't this be bad for some tasks, is there a way for a task to prevent
> > the max slack from being updated?
> 
> If a non-RT task can't tolerate not being on time it's totally broken.

Yes, I agree, except that the slack might be overwritten after the
application set an acceptable slack.

> RT-tasks are not affected by timer slack.
> 
> And we believe that user knows what he does.
> 
> > 2. Can there be an overlap in the ranges?
> 
> What overlap do you mean?
> There is only one rule: parent's range should be wider then child's.

Between the children, can we have something like

        10-200
        /    \
       /      \
      V        V
   10-110    100-200

-- 
	Three Cheers,
	Balbir
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/