OpenVZ Forum


Home » Mailing lists » Devel » [PATCH v3 00/28] kmem limitation for memcg
Re: [PATCH v3 16/28] memcg: kmem controller charge/uncharge infrastructure [message #46639 is a reply to message #46529] Wed, 30 May 2012 13:04 Go to previous messageGo to previous message
Frederic Weisbecker is currently offline  Frederic Weisbecker
Messages: 25
Registered: April 2012
Junior Member
On Fri, May 25, 2012 at 05:03:36PM +0400, Glauber Costa wrote:
> +struct kmem_cache *__mem_cgroup_get_kmem_cache(struct kmem_cache *cachep,
> + gfp_t gfp)
> +{
> + struct mem_cgroup *memcg;
> + int idx;
> + struct task_struct *p;
> +
> + gfp |= cachep->allocflags;
> +
> + if (cachep->memcg_params.memcg)
> + return cachep;
> +
> + idx = cachep->memcg_params.id;
> + VM_BUG_ON(idx == -1);
> +
> + p = rcu_dereference(current->mm->owner);
> + memcg = mem_cgroup_from_task(p);
> +
> + if (!mem_cgroup_kmem_enabled(memcg))
> + return cachep;
> +
> + if (memcg->slabs[idx] == NULL) {
> + memcg_create_cache_enqueue(memcg, cachep);
> + return cachep;
> + }
> +
> + return memcg->slabs[idx];
> +}
> +EXPORT_SYMBOL(__mem_cgroup_get_kmem_cache);
> +
> +bool __mem_cgroup_new_kmem_page(struct page *page, gfp_t gfp)
> +{
> + struct mem_cgroup *memcg;
> + struct page_cgroup *pc;
> + bool ret = true;
> + size_t size;
> + struct task_struct *p;
> +
> + if (!current->mm || in_interrupt())
> + return true;
> +
> + rcu_read_lock();
> + p = rcu_dereference(current->mm->owner);
> + memcg = mem_cgroup_from_task(p);
> +
> + if (!mem_cgroup_kmem_enabled(memcg))
> + goto out;

Do you think it's possible that this memcg can be destroyed (like ss->destroy())
concurrently?

Probably not because there is a synchronize_rcu() in cgroup_diput() so as long
as we are in rcu_read_lock() we are fine.

OTOH current->mm->owner can exit() right after we fetched its memcg and thus the css_set
can be freed concurrently? And then the cgroup itself after we call rcu_read_unlock()
due to cgroup_diput().
And yet we are doing the mem_cgroup_get() below unconditionally assuming it's
always fine to get a reference to it.

May be I'm missing something?


> + mem_cgroup_get(memcg);
> +
> + size = (1 << compound_order(page)) << PAGE_SHIFT;
> +
> + ret = memcg_charge_kmem(memcg, gfp, size) == 0;
> + if (!ret) {
> + mem_cgroup_put(memcg);
> + goto out;
> + }
> +
> + pc = lookup_page_cgroup(page);
> + lock_page_cgroup(pc);
> + pc->mem_cgroup = memcg;
> + SetPageCgroupUsed(pc);
> + unlock_page_cgroup(pc);
> +
> +out:
> + rcu_read_unlock();
> + return ret;
> +}
> +EXPORT_SYMBOL(__mem_cgroup_new_kmem_page);

Report message to a moderator

[Message index]
 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic: [PATCH] NFS: hard-code init_net for NFS callback transports
Next Topic: [PATCH v4 0/4] per cgroup cpu statistics
Goto Forum:
  

[ Syndicate this forum (XML) ] [ RSS ] [ PDF ]

Current Time: Sun Aug 31 18:26:25 GMT 2025

Total time taken to generate the page: 0.07331 seconds
Powered by FUDforum Powered by Virtuozzo Containers