From mboxrd@z Thu Jan  1 00:00:00 1970
From: Martin Schwidefsky <schwidefsky@de.ibm.com>
Subject: Re: [patch 10/12] [PATCH] kvm-s390: storage key interface
Date: Mon, 12 Dec 2011 11:06:23 +0100
Message-ID: <20111212110623.7535c819@de.ibm.com>
References: <20111210123529.379102332@de.ibm.com>
	<20111210123707.571900097@de.ibm.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Cc: Avi Kivity <avi@redhat.com>, Marcelo Tossati <mtosatti@redhat.com>,
	Christian Borntraeger <borntraeger@de.ibm.com>,
	Heiko Carstens <heiko.carstens@de.ibm.com>,
	Cornelia Huck <cornelia.huck@de.ibm.com>,
	KVM <kvm@vger.kernel.org>,
	Joachim von Buttlar <joachim_von_buttlar@de.ibm.com>,
	Jens Freimann <jfrei@de.ibm.com>,
	Constantin Werner <constantin.werner@de.ibm.com>, agraf@suse.de
To: Carsten Otte <cotte@de.ibm.com>
Return-path: <kvm-owner@vger.kernel.org>
Received: from e06smtp16.uk.ibm.com ([195.75.94.112]:48021 "EHLO
	e06smtp16.uk.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1750748Ab1LLKGf (ORCPT <rfc822;kvm@vger.kernel.org>);
	Mon, 12 Dec 2011 05:06:35 -0500
Received: from /spool/local
	by e06smtp16.uk.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted
	for <kvm@vger.kernel.org> from <schwidefsky@de.ibm.com>;
	Mon, 12 Dec 2011 10:06:33 -0000
Received: from d06av08.portsmouth.uk.ibm.com (d06av08.portsmouth.uk.ibm.com [9.149.37.249])
	by d06nrmr1307.portsmouth.uk.ibm.com (8.13.8/8.13.8/NCO v10.0) with ESMTP id pBCA6QGI2719994
	for <kvm@vger.kernel.org>; Mon, 12 Dec 2011 10:06:26 GMT
Received: from d06av08.portsmouth.uk.ibm.com (loopback [127.0.0.1])
	by d06av08.portsmouth.uk.ibm.com (8.14.4/8.13.1/NCO v10.0 AVout) with ESMTP id pBCA6OUv026704
	for <kvm@vger.kernel.org>; Mon, 12 Dec 2011 10:06:25 GMT
In-Reply-To: <20111210123707.571900097@de.ibm.com>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

On Sat, 10 Dec 2011 13:35:39 +0100
Carsten Otte <cotte@de.ibm.com> wrote:

> --- a/arch/s390/mm/pgtable.c
> +++ b/arch/s390/mm/pgtable.c
> @@ -393,6 +393,33 @@ out_unmap:
>  }
>  EXPORT_SYMBOL_GPL(gmap_map_segment);
> 
> +static pmd_t *__pmdp_for_addr(struct mm_struct *mm, unsigned long addr)
> +{
> +	struct vm_area_struct *vma;
> +	pgd_t *pgd;
> +	pud_t *pud;
> +	pmd_t *pmd;
> +
> +	vma = find_vma(mm, addr);
> +	if (!vma || (vma->vm_start > addr))
> +		return ERR_PTR(-EFAULT);
> +
> +	pgd = pgd_offset(mm, addr);
> +	pud = pud_alloc(mm, pgd, addr);
> +	if (!pud)
> +		return ERR_PTR(-ENOMEM);
> +
> +	pmd = pmd_alloc(mm, pud, addr);
> +	if (!pmd)
> +		return ERR_PTR(-ENOMEM);
> +
> +	if (!pmd_present(*pmd) &&
> +	    __pte_alloc(mm, vma, pmd, addr))
> +		return ERR_PTR(-ENOMEM);
> +
> +	return pmd;
> +}
> +
>  /*
>   * this function is assumed to be called with mmap_sem held
>   */

The __pmdp_for_addr function is fine for the usage in __gmap_fault.

> @@ -806,6 +820,26 @@ int s390_enable_sie(void)
>  }
>  EXPORT_SYMBOL_GPL(s390_enable_sie);
> 
> +pte_t *ptep_for_addr(unsigned long addr)
> +{
> +	pmd_t *pmd;
> +	pte_t *pte;
> +
> +	down_read(&current->mm->mmap_sem);
> +
> +	pmd = __pmdp_for_addr(current->mm, addr);
> +	if (IS_ERR(pmd)) {
> +		pte = (pte_t *)pmd;
> +		goto up_out;
> +	}
> +
> +	pte = pte_offset(pmd, addr);
> +up_out:
> +	up_read(&current->mm->mmap_sem);
> +	return pte;
> +}
> +EXPORT_SYMBOL_GPL(ptep_for_addr);
> +
>  #if defined(CONFIG_DEBUG_PAGEALLOC) && defined(CONFIG_HIBERNATION)
>  bool kernel_page_present(struct page *page)
>  {

There is a fundamental locking sanfu. The pointer the the page table
entry is only valid until the mmap_sem is released. The down_read/up_read
has to be done in the caller.

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.