linux-kernel - Re: [PATCH 16 of 55] xen64: add extra pv_mmu

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <4874709A.1080801@goop.org>
Date:	Wed, 09 Jul 2008 01:02:34 -0700
From:	Jeremy Fitzhardinge <jeremy@...p.org>
To:	Mark McLoughlin <markmc@...hat.com>
CC:	Ingo Molnar <mingo@...e.hu>, LKML <linux-kernel@...r.kernel.org>,
	x86@...nel.org, Stephen Tweedie <sct@...hat.com>,
	Eduardo Habkost <ehabkost@...hat.com>
Subject: Re: [PATCH 16 of 55] xen64: add extra pv_mmu_ops

Mark McLoughlin wrote:
> Hey Jeremy,
>
> On Tue, 2008-07-08 at 15:06 -0700, Jeremy Fitzhardinge wrote:
>   
>> We need extra pv_mmu_ops for 64-bit, to deal with the extra level of
>> pagetable.
>>
>> Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@...rix.com>
>> ---
>>  arch/x86/xen/enlighten.c   |   33 +++++++++++++++++++++++++++-
>>  arch/x86/xen/mmu.c         |   51 +++++++++++++++++++++++++++++++++++++++++++-
>>  arch/x86/xen/mmu.h         |   15 +++++++++++-
>>  include/asm-x86/xen/page.h |    4 +++
>>  4 files changed, 99 insertions(+), 4 deletions(-)
>>     
> ...
>   
>> diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
>> --- a/arch/x86/xen/mmu.c
>> +++ b/arch/x86/xen/mmu.c
>> @@ -438,14 +438,19 @@
>>  
>>  void xen_set_pte(pte_t *ptep, pte_t pte)
>>  {
>> +#ifdef CONFIG_X86_PAE
>>  	ptep->pte_high = pte.pte_high;
>>  	smp_wmb();
>>  	ptep->pte_low = pte.pte_low;
>> +#else
>> +	*ptep = pte;
>> +#endif
>>  }
>>     
>
> You've dropped non-PAE support already, right? Any reason to use the
> X86_PAE macro instead of X86_32?
>   

It's a fine difference, but the specific thing I'm testing for is 
actually PAE vs 64-bit (or even more specifically "can I update a whole 
pte atomically?").  Testing for X86_32 is equivalent in the absence of 
non-PAE 32-bit, but not quite as correct.  The other way I could do it 
is as I have elsewhere:

	if (sizeof(pte_t) > sizeof(long)) {
		ptep->pte_high = pte.pte_high;
		smp_wmb();
		ptep->pte_low = pte.pte_low;
	} else
		*ptep = pte;

But there are other places where it tests for this using PAE (and even 
whole pvops which are only defined in the PAE case).

You'll notice other places where I test for PAGETABLE_LEVELS == 3 where 
the code  is generally handling 3-level pagetables, and others where I 
test for PAE where it's something specific to PAE.

    J
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/