Re: [PATCH 03/10] mm: Convert p[te|md]_numa users to p[te|md]_protnone_numa

["Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>]

Benjamin Herrenschmidt <benh@kernel.crashing.org> writes:

> On Fri, 2014-11-21 at 13:57 +0000, Mel Gorman wrote:
>> Convert existing users of pte_numa and friends to the new helper. Note
>> that the kernel is broken after this patch is applied until the other
>> page table modifiers are also altered. This patch layout is to make
>> review easier.
>
> Aneesh, the removal of the DSISR_PROTFAULT checks, I wonder if we might
> break something here ... (I know, I asked for them to be removed :-)
>

That is the reason I converted that to a WARN_ON in later patch. 

> IE, we basically bounce all protection checks to the "normal" VMA
> protection checking, so far so good...
>
> But what about the subpage protection stuff ? Will that still work ?
>

I did look at that before. So if we had subpage access limitted, when we
take a fault for that subpage, we bail out early in hash_page_mm. (with
rc = 2). low_hash_fault handle that case directly. We will not end up
calling do_page_fault.

Now, hash_preload can possibly insert an hpte in hash page table even if
the access is not allowed by the pte permissions. But i guess even that
is ok. because we will fault again, end-up calling hash_page_mm where we
handle that part correctly.

-aneesh