[PATCH 07/23] cgroup: add css_parent()

From: Tejun Heo
Date: Thu Aug 01 2013 - 17:57:52 EST

Currently, controllers have to explicitly follow the cgroup hierarchy
to find the parent of a given css. cgroup is moving towards using
cgroup_subsys_state as the main controller interface construct, so
let's provide a way to climb the hierarchy using just csses.

This patch implements css_parent() which, given a css, returns its
parent. The function is guarnateed to valid non-NULL parent css as
long as the target css is not at the top of the hierarchy.

freezer, cpuset, cpu, cpuacct, hugetlb, memory, net_cls and devices
are converted to use css_parent() instead of accessing cgroup->parent
directly.

* __parent_ca() is dropped from cpuacct and its usage is replaced with
parent_ca(). The only difference between the two was NULL test on
cgroup->parent which is now embedded in css_parent() making the
distinction moot. Note that eventually a css->parent field will be
added to css and the NULL check in css_parent() will go away.

This patch shouldn't cause any behavior differences.

Signed-off-by: Tejun Heo <tj@xxxxxxxxxx>
---
block/blk-cgroup.h | 4 +---
include/linux/cgroup.h | 15 +++++++++++++++
kernel/cgroup_freezer.c | 8 ++------
kernel/cpuset.c | 6 +-----
kernel/sched/core.c | 9 +++------
kernel/sched/cpuacct.c | 11 ++---------
mm/hugetlb_cgroup.c | 6 +-----
mm/memcontrol.c | 39 +++++++++++----------------------------
net/sched/cls_cgroup.c | 8 +++++---
security/device_cgroup.c | 18 +++++-------------
10 files changed, 46 insertions(+), 78 deletions(-)

diff --git a/block/blk-cgroup.h b/block/blk-cgroup.h
index 8e5863e..b6802c4 100644
--- a/block/blk-cgroup.h
+++ b/block/blk-cgroup.h
@@ -209,9 +209,7 @@ static inline struct blkcg *bio_blkcg(struct bio *bio)
*/
static inline struct blkcg *blkcg_parent(struct blkcg *blkcg)
{
- struct cgroup *pcg = blkcg->css.cgroup->parent;
-
- return pcg ? cgroup_to_blkcg(pcg) : NULL;
+ return css_to_blkcg(css_parent(&blkcg->css));
}

/**
diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
index 9c2f9d7..b65f6b5 100644
--- a/include/linux/cgroup.h
+++ b/include/linux/cgroup.h
@@ -648,6 +648,21 @@ struct cgroup_subsys {
#undef SUBSYS

/**
+ * css_parent - find the parent css
+ * @css: the target cgroup_subsys_state
+ *
+ * Return the parent css of @css. This function is guaranteed to return
+ * non-NULL parent as long as @css isn't the root.
+ */
+static inline
+struct cgroup_subsys_state *css_parent(struct cgroup_subsys_state *css)
+{
+ struct cgroup *parent_cgrp = css->cgroup->parent;
+
+ return parent_cgrp ? parent_cgrp->subsys[css->ss->subsys_id] : NULL;
+}
+
+/**
* cgroup_css - obtain a cgroup's css for the specified subsystem
* @cgrp: the cgroup of interest
* @subsys_id: the subsystem of interest
diff --git a/kernel/cgroup_freezer.c b/kernel/cgroup_freezer.c
index 1db686e..657a73c 100644
--- a/kernel/cgroup_freezer.c
+++ b/kernel/cgroup_freezer.c
@@ -62,11 +62,7 @@ static inline struct freezer *task_freezer(struct task_struct *task)

static struct freezer *parent_freezer(struct freezer *freezer)
{
- struct cgroup *pcg = freezer->css.cgroup->parent;
-
- if (pcg)
- return cgroup_freezer(pcg);
- return NULL;
+ return css_freezer(css_parent(&freezer->css));
}

bool cgroup_freezing(struct task_struct *task)
@@ -234,7 +230,7 @@ static void freezer_fork(struct task_struct *task)
* The root cgroup is non-freezable, so we can skip the
* following check.
*/
- if (!freezer->css.cgroup->parent)
+ if (!parent_freezer(freezer))
goto out;

spin_lock_irq(&freezer->lock);
diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index 6e9cbdd..259a4af 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -133,11 +133,7 @@ static inline struct cpuset *task_cs(struct task_struct *task)

static inline struct cpuset *parent_cs(struct cpuset *cs)
{
- struct cgroup *pcgrp = cs->css.cgroup->parent;
-
- if (pcgrp)
- return cgroup_cs(pcgrp);
- return NULL;
+ return css_cs(css_parent(&cs->css));
}

#ifdef CONFIG_NUMA
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 5bccb02..7a10742 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -7114,13 +7114,10 @@ static struct cgroup_subsys_state *cpu_cgroup_css_alloc(struct cgroup *cgrp)
static int cpu_cgroup_css_online(struct cgroup *cgrp)
{
struct task_group *tg = cgroup_tg(cgrp);
- struct task_group *parent;
+ struct task_group *parent = css_tg(css_parent(&tg->css));

- if (!cgrp->parent)
- return 0;
-
- parent = cgroup_tg(cgrp->parent);
- sched_online_group(tg, parent);
+ if (parent)
+ sched_online_group(tg, parent);
return 0;
}

diff --git a/kernel/sched/cpuacct.c b/kernel/sched/cpuacct.c
index 8ccfa10..f6926a1 100644
--- a/kernel/sched/cpuacct.c
+++ b/kernel/sched/cpuacct.c
@@ -50,16 +50,9 @@ static inline struct cpuacct *task_ca(struct task_struct *tsk)
return css_ca(task_css(tsk, cpuacct_subsys_id));
}

-static inline struct cpuacct *__parent_ca(struct cpuacct *ca)
-{
- return cgroup_ca(ca->css.cgroup->parent);
-}
-
static inline struct cpuacct *parent_ca(struct cpuacct *ca)
{
- if (!ca->css.cgroup->parent)
- return NULL;
- return cgroup_ca(ca->css.cgroup->parent);
+ return css_ca(css_parent(&ca->css));
}

static DEFINE_PER_CPU(u64, root_cpuacct_cpuusage);
@@ -284,7 +277,7 @@ void cpuacct_account_field(struct task_struct *p, int index, u64 val)
while (ca != &root_cpuacct) {
kcpustat = this_cpu_ptr(ca->cpustat);
kcpustat->cpustat[index] += val;
- ca = __parent_ca(ca);
+ ca = parent_ca(ca);
}
rcu_read_unlock();
}
diff --git a/mm/hugetlb_cgroup.c b/mm/hugetlb_cgroup.c
index 95585a0..57ecb5d 100644
--- a/mm/hugetlb_cgroup.c
+++ b/mm/hugetlb_cgroup.c
@@ -59,11 +59,7 @@ static inline bool hugetlb_cgroup_is_root(struct hugetlb_cgroup *h_cg)
static inline struct hugetlb_cgroup *
parent_hugetlb_cgroup(struct hugetlb_cgroup *h_cg)
{
- struct cgroup *parent = h_cg->css.cgroup->parent;
-
- if (!parent)
- return NULL;
- return hugetlb_cgroup_from_cgroup(parent);
+ return hugetlb_cgroup_from_css(css_parent(&h_cg->css));
}

static inline bool hugetlb_cgroup_have_usage(struct hugetlb_cgroup *h_cg)
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 11d659e..69b3e52 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -1524,10 +1524,8 @@ static unsigned long mem_cgroup_margin(struct mem_cgroup *memcg)

int mem_cgroup_swappiness(struct mem_cgroup *memcg)
{
- struct cgroup *cgrp = memcg->css.cgroup;
-
/* root ? */
- if (cgrp->parent == NULL)
+ if (!css_parent(&memcg->css))
return vm_swappiness;

return memcg->swappiness;
@@ -5026,11 +5024,7 @@ static int mem_cgroup_hierarchy_write(struct cgroup *cont, struct cftype *cft,
{
int retval = 0;
struct mem_cgroup *memcg = mem_cgroup_from_cont(cont);
- struct cgroup *parent = cont->parent;
- struct mem_cgroup *parent_memcg = NULL;
-
- if (parent)
- parent_memcg = mem_cgroup_from_cont(parent);
+ struct mem_cgroup *parent_memcg = mem_cgroup_from_css(css_parent(&memcg->css));

mutex_lock(&memcg_create_mutex);

@@ -5282,18 +5276,15 @@ static int mem_cgroup_write(struct cgroup *cont, struct cftype *cft,
static void memcg_get_hierarchical_limit(struct mem_cgroup *memcg,
unsigned long long *mem_limit, unsigned long long *memsw_limit)
{
- struct cgroup *cgroup;
unsigned long long min_limit, min_memsw_limit, tmp;

min_limit = res_counter_read_u64(&memcg->res, RES_LIMIT);
min_memsw_limit = res_counter_read_u64(&memcg->memsw, RES_LIMIT);
- cgroup = memcg->css.cgroup;
if (!memcg->use_hierarchy)
goto out;

- while (cgroup->parent) {
- cgroup = cgroup->parent;
- memcg = mem_cgroup_from_cont(cgroup);
+ while (css_parent(&memcg->css)) {
+ memcg = mem_cgroup_from_css(css_parent(&memcg->css));
if (!memcg->use_hierarchy)
break;
tmp = res_counter_read_u64(&memcg->res, RES_LIMIT);
@@ -5523,16 +5514,11 @@ static int mem_cgroup_swappiness_write(struct cgroup *cgrp, struct cftype *cft,
u64 val)
{
struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp);
- struct mem_cgroup *parent;
-
- if (val > 100)
- return -EINVAL;
+ struct mem_cgroup *parent = mem_cgroup_from_css(css_parent(&memcg->css));

- if (cgrp->parent == NULL)
+ if (val > 100 || !parent)
return -EINVAL;

- parent = mem_cgroup_from_cont(cgrp->parent);
-
mutex_lock(&memcg_create_mutex);

/* If under hierarchy, only empty-root can set this value */
@@ -5861,14 +5847,12 @@ static int mem_cgroup_oom_control_write(struct cgroup *cgrp,
struct cftype *cft, u64 val)
{
struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp);
- struct mem_cgroup *parent;
+ struct mem_cgroup *parent = mem_cgroup_from_css(css_parent(&memcg->css));

/* cannot set to root cgroup and only 0 and 1 are allowed */
- if (!cgrp->parent || !((val == 0) || (val == 1)))
+ if (!parent || !((val == 0) || (val == 1)))
return -EINVAL;

- parent = mem_cgroup_from_cont(cgrp->parent);
-
mutex_lock(&memcg_create_mutex);
/* oom-kill-disable is a flag for subhierarchy. */
if ((parent->use_hierarchy) || memcg_has_children(memcg)) {
@@ -6266,15 +6250,14 @@ free_out:
static int
mem_cgroup_css_online(struct cgroup *cont)
{
- struct mem_cgroup *memcg, *parent;
+ struct mem_cgroup *memcg = mem_cgroup_from_cont(cont);
+ struct mem_cgroup *parent = mem_cgroup_from_css(css_parent(&memcg->css));
int error = 0;

- if (!cont->parent)
+ if (!parent)
return 0;

mutex_lock(&memcg_create_mutex);
- memcg = mem_cgroup_from_cont(cont);
- parent = mem_cgroup_from_cont(cont->parent);

memcg->use_hierarchy = parent->use_hierarchy;
memcg->oom_kill_disable = parent->oom_kill_disable;
diff --git a/net/sched/cls_cgroup.c b/net/sched/cls_cgroup.c
index af412ab..9e6b75e 100644
--- a/net/sched/cls_cgroup.c
+++ b/net/sched/cls_cgroup.c
@@ -50,9 +50,11 @@ static struct cgroup_subsys_state *cgrp_css_alloc(struct cgroup *cgrp)

static int cgrp_css_online(struct cgroup *cgrp)
{
- if (cgrp->parent)
- cgrp_cls_state(cgrp)->classid =
- cgrp_cls_state(cgrp->parent)->classid;
+ struct cgroup_cls_state *cs = cgrp_cls_state(cgrp);
+ struct cgroup_cls_state *parent = css_cls_state(css_parent(&cs->css));
+
+ if (parent)
+ cs->classid = parent->classid;
return 0;
}

diff --git a/security/device_cgroup.c b/security/device_cgroup.c
index 9095364..635a49d 100644
--- a/security/device_cgroup.c
+++ b/security/device_cgroup.c
@@ -198,13 +198,11 @@ static inline bool is_devcg_online(const struct dev_cgroup *devcg)
*/
static int devcgroup_online(struct cgroup *cgroup)
{
- struct dev_cgroup *dev_cgroup, *parent_dev_cgroup = NULL;
+ struct dev_cgroup *dev_cgroup = cgroup_to_devcgroup(cgroup);
+ struct dev_cgroup *parent_dev_cgroup = css_to_devcgroup(css_parent(&dev_cgroup->css));
int ret = 0;

mutex_lock(&devcgroup_mutex);
- dev_cgroup = cgroup_to_devcgroup(cgroup);
- if (cgroup->parent)
- parent_dev_cgroup = cgroup_to_devcgroup(cgroup->parent);

if (parent_dev_cgroup == NULL)
dev_cgroup->behavior = DEVCG_DEFAULT_ALLOW;
@@ -394,12 +392,10 @@ static bool may_access(struct dev_cgroup *dev_cgroup,
static int parent_has_perm(struct dev_cgroup *childcg,
struct dev_exception_item *ex)
{
- struct cgroup *pcg = childcg->css.cgroup->parent;
- struct dev_cgroup *parent;
+ struct dev_cgroup *parent = css_to_devcgroup(css_parent(&childcg->css));

- if (!pcg)
+ if (!parent)
return 1;
- parent = cgroup_to_devcgroup(pcg);
return may_access(parent, ex, childcg->behavior);
}

@@ -524,15 +520,11 @@ static int devcgroup_update_access(struct dev_cgroup *devcgroup,
char temp[12]; /* 11 + 1 characters needed for a u32 */
int count, rc = 0;
struct dev_exception_item ex;
- struct cgroup *p = devcgroup->css.cgroup;
- struct dev_cgroup *parent = NULL;
+ struct dev_cgroup *parent = css_to_devcgroup(css_parent(&devcgroup->css));

if (!capable(CAP_SYS_ADMIN))
return -EPERM;

- if (p->parent)
- parent = cgroup_to_devcgroup(p->parent);
-
memset(&ex, 0, sizeof(ex));
b = buffer;

--
1.8.3.1

--
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/