From mboxrd@z Thu Jan  1 00:00:00 1970
From: Anthony Liguori <anthony@codemonkey.ws>
Subject: Re: [Qemu-devel] Re: [PATCH] Make page_find() return 0 for too-large
 addresses
Date: Mon, 15 Sep 2008 10:29:14 -0500
Message-ID: <48CE7F4A.3060000@codemonkey.ws>
References: <20080912185856.GM3982@blackpad> <48CAC809.5000901@codemonkey.ws> <20080912201406.GA10147@blackpad> <20080912204404.GB10147@blackpad>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1;
	format=flowed
Content-Transfer-Encoding: QUOTED-PRINTABLE
Cc: qemu-devel@nongnu.org, kvm@vger.kernel.org, gcosta@redhat.com
To: Eduardo Habkost <ehabkost@redhat.com>
Return-path: <kvm-owner@vger.kernel.org>
Received: from yx-out-2324.google.com ([74.125.44.29]:59905 "EHLO
	yx-out-2324.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1754156AbYIOPaJ (ORCPT <rfc822;kvm@vger.kernel.org>);
	Mon, 15 Sep 2008 11:30:09 -0400
Received: by yx-out-2324.google.com with SMTP id 8so650290yxm.1
        for <kvm@vger.kernel.org>; Mon, 15 Sep 2008 08:30:08 -0700 (PDT)
In-Reply-To: <20080912204404.GB10147@blackpad>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

Eduardo Habkost wrote:
> On Fri, Sep 12, 2008 at 05:14:06PM -0300, Eduardo Habkost wrote:
>  =20
>> On Fri, Sep 12, 2008 at 02:50:33PM -0500, Anthony Liguori wrote:
>>    =20
>>> Eduardo Habkost wrote:
>>>      =20
>>>> On some cases, such as under KVM, tb_invalidate_phys_page_range()
>>>> may be called for large addresses, when qemu is configured to more=
 than
>>>> 4GB of RAM.
>>>>
>>>> On these cases, qemu was crashing because it was using an index to=
o
>>>> large for l1_map[], that supports only 32-bit addresses when compi=
ling
>>>> without CONFIG_USER_ONLY.
>>>>  =20
>>>>        =20
> <snip>
>  =20
>> BTW, I've just noticed page_find_alloc() has this:
>>
>> #if TARGET_LONG_BITS > 32
>>     /* Host memory outside guest VM.  For 32-bit targets we have alr=
eady
>>        excluded high addresses.  */
>>     if (index > ((target_ulong)L2_SIZE * L1_SIZE))
>>         return NULL;
>> #endif
>>
>> So, we can just use a similar check on page_find().
>>    =20
>
> New patch, reusing the range check from page_find_alloc() on
> page_find(). Untested.
>  =20

Have you tested this patch yet?  I like to avoid being the first one to=
=20
test something when it's not my code :-)

Regards,

Anthony Liguori

> Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
> ---
> Index: qemu/exec.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- qemu/exec.c	(revis=E3o 5200)
> +++ qemu/exec.c	(c=F3pia de trabalho)
> @@ -279,17 +279,24 @@ static void page_init(void)
>  #endif
>  }
> =20
> -static inline PageDesc *page_find_alloc(target_ulong index)
> +static inline PageDesc **page_l1_map(target_ulong index)
>  {
> -    PageDesc **lp, *p;
> -
>  #if TARGET_LONG_BITS > 32
>      /* Host memory outside guest VM.  For 32-bit targets we have alr=
eady
>         excluded high addresses.  */
>      if (index > ((target_ulong)L2_SIZE * L1_SIZE))
>          return NULL;
>  #endif
> -    lp =3D &l1_map[index >> L2_BITS];
> +    return &l1_map[index >> L2_BITS];
> +}
> +
> +static inline PageDesc *page_find_alloc(target_ulong index)
> +{
> +    PageDesc **lp, *p;
> +    lp =3D page_l1_map(index);
> +    if (!lp)
> +        return NULL;
> +
>      p =3D *lp;
>      if (!p) {
>          /* allocate if not found */
> @@ -316,9 +323,12 @@ static inline PageDesc *page_find_alloc(
> =20
>  static inline PageDesc *page_find(target_ulong index)
>  {
> -    PageDesc *p;
> +    PageDesc **lp, *p;
> +    lp =3D page_l1_map(index);
> +    if (!lp)
> +        return NULL;
> =20
> -    p =3D l1_map[index >> L2_BITS];
> +    p =3D *lp;
>      if (!p)
>          return 0;
>      return p + (index & (L2_SIZE - 1));
>
>
>  =20