Re: [PATCH v2] mm: per-thread vma caching

From: Peter Zijlstra <peterz@infradead.org>
To: Davidlohr Bueso <davidlohr@hp.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
	Ingo Molnar <mingo@kernel.org>,
	Linus Torvalds <torvalds@linux-foundation.org>,
	Michel Lespinasse <walken@google.com>,
	Mel Gorman <mgorman@suse.de>, Rik van Riel <riel@redhat.com>,
	KOSAKI Motohiro <kosaki.motohiro@gmail.com>,
	aswin@hp.com, scott.norton@hp.com, linux-mm@kvack.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2] mm: per-thread vma caching
Date: Wed, 26 Feb 2014 09:50:48 +0100	[thread overview]
Message-ID: <20140226085048.GE18404@twins.programming.kicks-ass.net> (raw)
In-Reply-To: <1393352206.2577.36.camel@buesod1.americas.hpqcorp.net>

On Tue, Feb 25, 2014 at 10:16:46AM -0800, Davidlohr Bueso wrote:
> +void vmacache_invalidate_all(void)
> +{
> +	struct task_struct *g, *p;
> +
> +	rcu_read_lock();
> +	for_each_process_thread(g, p) {
> +		/*
> +		 * Only flush the vmacache pointers as the
> +		 * mm seqnum is already set and curr's will
> +		 * be set upon invalidation when the next
> +		 * lookup is done.
> +		 */
> +		memset(p->vmacache, 0, sizeof(p->vmacache));
> +	}
> +	rcu_read_unlock();
> +}

With all the things being said on this particular piece already; I
wanted to add that the iteration there is incomplete; we can clone()
using CLONE_VM without using CLONE_THREAD.

Its not common, but it can be done. In that case the above iteration
will miss a task that shares the same mm.