Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1453278 > unrolled thread

Re: [PATCH] radix-tree: account nodes to memcg only if explicitly requested

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [PATCH] radix-tree: account nodes to memcg only if explicitly  requested Johannes Weiner <hannes@cmpxchg.org> - 2016-08-01 17:30 +0200
    Re: [PATCH] radix-tree: account nodes to memcg only if explicitly  requested Johannes Weiner <hannes@cmpxchg.org> - 2016-08-01 19:20 +0200

#1453278 — Re: [PATCH] radix-tree: account nodes to memcg only if explicitly requested

On Mon, Aug 01, 2016 at 04:13:08PM +0300, Vladimir Davydov wrote:
> Radix trees may be used not only for storing page cache pages, so
> unconditionally accounting radix tree nodes to the current memory cgroup
> is bad: if a radix tree node is used for storing data shared among
> different cgroups we risk pinning dead memory cgroups forever. So let's
> only account radix tree nodes if it was explicitly requested by passing
> __GFP_ACCOUNT to INIT_RADIX_TREE. Currently, we only want to account
> page cache entries, so mark mapping->page_tree so.

Is this a theoretical fix, or did you actually run into problems? I
wouldn't expect any other radix tree node consumer in the kernel to
come anywhere close to the page cache, so I wonder why it matters.

> @@ -351,6 +351,12 @@ static int __radix_tree_preload(gfp_t gfp_mask, int nr)
>  	struct radix_tree_node *node;
>  	int ret = -ENOMEM;
>  
> +	/*
> +	 * Nodes preloaded by one cgroup can be be used by another cgroup, so
> +	 * they should never be accounted to any particular memory cgroup.
> +	 */
> +	gfp_mask &= ~__GFP_ACCOUNT;

But *all* page cache radix tree nodes are allocated from inside the
preload code, since the tree insertions need mapping->tree_lock. So
this would effectively disable accounting of the biggest radix tree
consumer in the kernel, no?

[toc] | [next] | [standalone]

#1453342

On Mon, Aug 01, 2016 at 07:06:05PM +0300, Vladimir Davydov wrote:
> On Mon, Aug 01, 2016 at 11:24:09AM -0400, Johannes Weiner wrote:
> > On Mon, Aug 01, 2016 at 04:13:08PM +0300, Vladimir Davydov wrote:
> > > @@ -351,6 +351,12 @@ static int __radix_tree_preload(gfp_t gfp_mask, int nr)
> > >  	struct radix_tree_node *node;
> > >  	int ret = -ENOMEM;
> > >  
> > > +	/*
> > > +	 * Nodes preloaded by one cgroup can be be used by another cgroup, so
> > > +	 * they should never be accounted to any particular memory cgroup.
> > > +	 */
> > > +	gfp_mask &= ~__GFP_ACCOUNT;
> > 
> > But *all* page cache radix tree nodes are allocated from inside the
> > preload code, since the tree insertions need mapping->tree_lock. So
> > this would effectively disable accounting of the biggest radix tree
> > consumer in the kernel, no?
> 
> No, that's not how accounting of radix tree nodes works. We never
> account preloaded nodes, because this could result in a node accounted
> to one cgroup used by an unrelated cgroup. Instead we always try to
> kmalloc a node on insertion falling back on preloads only if kmalloc
> fails - see commit 58e698af4c634 ("radix-tree: account radix_tree_node
> to memory cgroup").

You are right, I forgot we are doing this. The patch makes sense then.

Acked-by: Johannes Weiner <hannes@cmpxchg.org>

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web