Re: [PATCH 2/2] x86: replace BIO_VMERGE_BOUNDARY with BIOVEC_PHYS_MERGEABLE

[Jeremy Fitzhardinge <jeremy@goop.org>]

Ingo Molnar wrote:
> * FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp> wrote:
>
>   
>> On Mon, 27 Oct 2008 12:57:00 +0100
>> Jens Axboe <jens.axboe@oracle.com> wrote:
>>
>>     
>>> On Mon, Oct 27 2008, Jeremy Fitzhardinge wrote:
>>>       
>>>> Jens Axboe wrote:
>>>>         
>>>>> Pretty much baffles me as well, xen should just need to do
>>>>>
>>>>> #define BIOVEC_PHYS_MERGEABLE(vec1, vec2)       0
>>>>>  
>>>>>           
>>>> It needs to be a runtime switch, since we only want to do this when 
>>>> actually running under Xen.  Also, its possible that the two pages might 
>>>> actually be physically contiguous, so they could be merged anyway.
>>>>         
>>> Alright, then add a xen_biovec_phys_mergeable(vec1, vec2) in the xen
>>> code that actually checks this for real. You can add your switch there
>>> as well. Then put the BIOVEC_PHYS_MERGEABLE() in the xen arch includes,
>>> done.
>>>
>>> What Tomo is saying is that this has nothing to do with virtual merging,
>>> and he's right.
>>>       
>> Yeah, overriding BIOVEC_PHYS_MERGEABLE perfectly works for Xen. And it
>> is not related with BIO_VMERGE_BOUNDARY at all.
>>
>> Ingo, please put this patch into your tree:
>>
>> http://marc.info/?l=linux-kernel&m=122482703716620&w=2
>>     
>
> does it have any dependency on:
>
>   Subject: [PATCH 1/2] bio: define __BIOVEC_PHYS_MERGEABLE
>   

No, they're independent.  Defining __BIOVEC_PHYS_MERGEABLE will be 
useful for the Xen changes I need to make in the wake of Tomo's patch.

    J