Re: [patch 2/8] x86 PAT: set VM_PFNMAP flag in vm_insert_pfn

[Nick Piggin <npiggin@suse.de>]

On Thu, Nov 13, 2008 at 10:47:23AM -0800, Pallipadi, Venkatesh wrote:
> >> >-----Original Message-----
> >> >From: Nick Piggin [mailto:npiggin@suse.de]
> >> >Sent: Wednesday, November 12, 2008 3:23 PM
> >> >To: Pallipadi, Venkatesh
> >> >Cc: Ingo Molnar; Thomas Gleixner; H.Peter Anvin; Hugh Dickins;
> >> >Roland Dreier; Jesse Barnes; Jeremy Fitzhardinge; Arjan van de
> >> >Ven; linux-kernel@vger.kernel.org; Siddha, Suresh B
> >> >Subject: Re: [patch 2/8] x86 PAT: set VM_PFNMAP flag in 
> >vm_insert_pfn
> >> >
> >> >You have to be careful of this, because it can be called 
> >with mmap_sem
> >> >held for read only. Hmm, I guess vm_insert_page is doing the
> >> >same thing.
> >> >Probably mostly works because all other modifiers of vm_flags
> >> >are holding
> >> >mmap_sem.
> >>
> >> Yes. I did the patch looking at vm_insert_page doing similar thing.
> >>
> >> >
> >> >However, in some cases, code can do vm_insert_pfn and vm_insert_page
> >> >(actually hmm, no vm_insert_mixed actually should cover 
> >most of those
> >> >cases).
> >> >
> >> >Still, I'd be much happier if we could make these into 
> >BUG_ON, and then
> >> >teach callers to set it in their .mmap routines.
> >>
> >> Actually, vm_insert_pfn() already has a BUG_ON() at the 
> >start for cases
> >> where neither (or both) MIXEDMAP and PFNMAP is not set. So, 
> >that should
> >> cover the case we are worried about it here and we can eliminate this
> >> patch altogether. Only part I am not sure about is why we are looking
> >> for MIXEDMAP here. Shouldn't they be using vm_insert_mixed instead?
> >
> >They should, but it will do an inesrt_pfn in some cases, won't it?
> >
> 
> Yes. It does. But, it calls a lower level insert_pfn() function. The lower
> level insert_pfn() does not have any bug checks. But the higher level
> vm_insert_pfn() checks for PFNMAP or MIXEDMAP.

Yes, but is there anything extra you need to check for cache aliases in
MIXEDMAP mappings?