2007-01-09 15:02:31

by Thomas Hellström

Subject: Replace nopage() / nopfn() with fault()

Nick,

We're working to slowly get the new DRM memory manager into the
mainstream kernel.
This means we have a need for the page fault handler patch you wrote
some time ago.
I guess we could take the no_pfn() route, but that would need a check
for racing
unmap_mapping_range(), and other problems may arise.

What is the current status and plans for inclusion of the fault() code?

Thanks,
Thomas Hellstrom
Tungsten Graphics.

2007-01-12 02:14:30

by Nick Piggin

Subject: Re: Replace nopage() / nopfn() with fault()

On Tue, Jan 09, 2007 at 04:02:08PM +0100, Thomas Hellstr?m wrote:
> Nick,
>
> We're working to slowly get the new DRM memory manager into the
> mainstream kernel.
> This means we have a need for the page fault handler patch you wrote
> some time ago.
> I guess we could take the no_pfn() route, but that would need a check
> for racing
> unmap_mapping_range(), and other problems may arise.
>
> What is the current status and plans for inclusion of the fault() code?

Hi Thomas,

fault should have gone in already, but the ordering of my patchset was
a little bit unfortunate :P

I will submit them to Andrew later today.

Nick

2007-01-12 11:47:53

by Thomas Hellström

Subject: Re: Replace nopage() / nopfn() with fault()

Nick Piggin wrote:

>On Tue, Jan 09, 2007 at 04:02:08PM +0100, Thomas Hellstr?m wrote:
>
>
>>Nick,
>>
>>We're working to slowly get the new DRM memory manager into the
>>mainstream kernel.
>>This means we have a need for the page fault handler patch you wrote
>>some time ago.
>>I guess we could take the no_pfn() route, but that would need a check
>>for racing
>>unmap_mapping_range(), and other problems may arise.
>>
>>What is the current status and plans for inclusion of the fault() code?
>>
>>
>
>Hi Thomas,
>
>fault should have gone in already, but the ordering of my patchset was
>a little bit unfortunate :P
>
>I will submit them to Andrew later today.
>
>Nick
>
>
>
Thanks, Nick.

/Thomas.