Commit d142e3e6 authored by Glauber Costa's avatar Glauber Costa Committed by Linus Torvalds

memcg: split part of memcg creation to css_online

This patch is a preparatory work for later locking rework to get rid of
big cgroup lock from memory controller code.

The memory controller uses some tunables to adjust its operation.  Those
tunables are inherited from parent to children upon children
intialization.  For most of them, the value cannot be changed after the
parent has a new children.

cgroup core splits initialization in two phases: css_alloc and css_online.
After css_alloc, the memory allocation and basic initialization are done.
But the new group is not yet visible anywhere, not even for cgroup core
code.  It is only somewhere between css_alloc and css_online that it is
inserted into the internal children lists.  Copying tunable values in
css_alloc will lead to inconsistent values: the children will copy the old
parent values, that can change between the copy and the moment in which
the groups is linked to any data structure that can indicate the presence
of children.
Signed-off-by: default avatarGlauber Costa <glommer@parallels.com>
Acked-by: default avatarMichal Hocko <mhocko@suse.cz>
Cc: Tejun Heo <tj@kernel.org>
Cc: Hiroyuki Kamezawa <kamezawa.hiroyuki@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
parent ee5e8472
...@@ -6093,7 +6093,7 @@ static int mem_cgroup_soft_limit_tree_init(void) ...@@ -6093,7 +6093,7 @@ static int mem_cgroup_soft_limit_tree_init(void)
static struct cgroup_subsys_state * __ref static struct cgroup_subsys_state * __ref
mem_cgroup_css_alloc(struct cgroup *cont) mem_cgroup_css_alloc(struct cgroup *cont)
{ {
struct mem_cgroup *memcg, *parent; struct mem_cgroup *memcg;
long error = -ENOMEM; long error = -ENOMEM;
int node; int node;
...@@ -6108,7 +6108,7 @@ mem_cgroup_css_alloc(struct cgroup *cont) ...@@ -6108,7 +6108,7 @@ mem_cgroup_css_alloc(struct cgroup *cont)
/* root ? */ /* root ? */
if (cont->parent == NULL) { if (cont->parent == NULL) {
int cpu; int cpu;
parent = NULL;
if (mem_cgroup_soft_limit_tree_init()) if (mem_cgroup_soft_limit_tree_init())
goto free_out; goto free_out;
root_mem_cgroup = memcg; root_mem_cgroup = memcg;
...@@ -6117,13 +6117,43 @@ mem_cgroup_css_alloc(struct cgroup *cont) ...@@ -6117,13 +6117,43 @@ mem_cgroup_css_alloc(struct cgroup *cont)
&per_cpu(memcg_stock, cpu); &per_cpu(memcg_stock, cpu);
INIT_WORK(&stock->work, drain_local_stock); INIT_WORK(&stock->work, drain_local_stock);
} }
} else {
parent = mem_cgroup_from_cont(cont->parent); res_counter_init(&memcg->res, NULL);
memcg->use_hierarchy = parent->use_hierarchy; res_counter_init(&memcg->memsw, NULL);
memcg->oom_kill_disable = parent->oom_kill_disable; res_counter_init(&memcg->kmem, NULL);
} }
if (parent && parent->use_hierarchy) { memcg->last_scanned_node = MAX_NUMNODES;
INIT_LIST_HEAD(&memcg->oom_notify);
atomic_set(&memcg->refcnt, 1);
memcg->move_charge_at_immigrate = 0;
mutex_init(&memcg->thresholds_lock);
spin_lock_init(&memcg->move_lock);
return &memcg->css;
free_out:
__mem_cgroup_free(memcg);
return ERR_PTR(error);
}
static int
mem_cgroup_css_online(struct cgroup *cont)
{
struct mem_cgroup *memcg, *parent;
int error = 0;
if (!cont->parent)
return 0;
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;
memcg->swappiness = mem_cgroup_swappiness(parent);
if (parent->use_hierarchy) {
res_counter_init(&memcg->res, &parent->res); res_counter_init(&memcg->res, &parent->res);
res_counter_init(&memcg->memsw, &parent->memsw); res_counter_init(&memcg->memsw, &parent->memsw);
res_counter_init(&memcg->kmem, &parent->kmem); res_counter_init(&memcg->kmem, &parent->kmem);
...@@ -6144,18 +6174,9 @@ mem_cgroup_css_alloc(struct cgroup *cont) ...@@ -6144,18 +6174,9 @@ mem_cgroup_css_alloc(struct cgroup *cont)
* much sense so let cgroup subsystem know about this * much sense so let cgroup subsystem know about this
* unfortunate state in our controller. * unfortunate state in our controller.
*/ */
if (parent && parent != root_mem_cgroup) if (parent != root_mem_cgroup)
mem_cgroup_subsys.broken_hierarchy = true; mem_cgroup_subsys.broken_hierarchy = true;
} }
memcg->last_scanned_node = MAX_NUMNODES;
INIT_LIST_HEAD(&memcg->oom_notify);
if (parent)
memcg->swappiness = mem_cgroup_swappiness(parent);
atomic_set(&memcg->refcnt, 1);
memcg->move_charge_at_immigrate = 0;
mutex_init(&memcg->thresholds_lock);
spin_lock_init(&memcg->move_lock);
error = memcg_init_kmem(memcg, &mem_cgroup_subsys); error = memcg_init_kmem(memcg, &mem_cgroup_subsys);
if (error) { if (error) {
...@@ -6165,12 +6186,8 @@ mem_cgroup_css_alloc(struct cgroup *cont) ...@@ -6165,12 +6186,8 @@ mem_cgroup_css_alloc(struct cgroup *cont)
* call __mem_cgroup_free, so return directly * call __mem_cgroup_free, so return directly
*/ */
mem_cgroup_put(memcg); mem_cgroup_put(memcg);
return ERR_PTR(error);
} }
return &memcg->css; return error;
free_out:
__mem_cgroup_free(memcg);
return ERR_PTR(error);
} }
static void mem_cgroup_css_offline(struct cgroup *cont) static void mem_cgroup_css_offline(struct cgroup *cont)
...@@ -6780,6 +6797,7 @@ struct cgroup_subsys mem_cgroup_subsys = { ...@@ -6780,6 +6797,7 @@ struct cgroup_subsys mem_cgroup_subsys = {
.name = "memory", .name = "memory",
.subsys_id = mem_cgroup_subsys_id, .subsys_id = mem_cgroup_subsys_id,
.css_alloc = mem_cgroup_css_alloc, .css_alloc = mem_cgroup_css_alloc,
.css_online = mem_cgroup_css_online,
.css_offline = mem_cgroup_css_offline, .css_offline = mem_cgroup_css_offline,
.css_free = mem_cgroup_css_free, .css_free = mem_cgroup_css_free,
.can_attach = mem_cgroup_can_attach, .can_attach = mem_cgroup_can_attach,
......
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