Re: [PATCH v2 3/5] userfaultfd: non-cooperative: add event for exit() notification

From: Andrew Morton <akpm@linux-foundation.org>
To: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>,
	"Dr. David Alan Gilbert" <dgilbert@redhat.com>,
	Hillf Danton <hillf.zj@alibaba-inc.com>,
	Mike Kravetz <mike.kravetz@oracle.com>,
	Pavel Emelyanov <xemul@virtuozzo.com>,
	Linux-MM <linux-mm@kvack.org>,
	LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v2 3/5] userfaultfd: non-cooperative: add event for exit() notification
Date: Tue, 31 Jan 2017 16:41:32 -0800	[thread overview]
Message-ID: <20170131164132.439f9d30e3a9b3c79bcada3a@linux-foundation.org> (raw)
In-Reply-To: <1485542673-24387-4-git-send-email-rppt@linux.vnet.ibm.com>

On Fri, 27 Jan 2017 20:44:31 +0200 Mike Rapoport <rppt@linux.vnet.ibm.com> wrote:

> Allow userfaultfd monitor track termination of the processes that have
> memory backed by the uffd.
> 
> --- a/fs/userfaultfd.c
> +++ b/fs/userfaultfd.c
> @@ -774,6 +774,30 @@ void userfaultfd_unmap_complete(struct mm_struct *mm, struct list_head *uf)
>  	}
>  }
>  
> +void userfaultfd_exit(struct mm_struct *mm)
> +{
> +	struct vm_area_struct *vma = mm->mmap;
> +
> +	while (vma) {
> +		struct userfaultfd_ctx *ctx = vma->vm_userfaultfd_ctx.ctx;
> +
> +		if (ctx && (ctx->features & UFFD_FEATURE_EVENT_EXIT)) {
> +			struct userfaultfd_wait_queue ewq;
> +
> +			userfaultfd_ctx_get(ctx);
> +
> +			msg_init(&ewq.msg);
> +			ewq.msg.event = UFFD_EVENT_EXIT;
> +
> +			userfaultfd_event_wait_completion(ctx, &ewq);
> +
> +			ctx->features &= ~UFFD_FEATURE_EVENT_EXIT;
> +		}
> +
> +		vma = vma->vm_next;
> +	}
> +}

And we can do the vma walk without locking because the caller (exit_mm)
knows it now has exclusive access.  Worth a comment?