Re: [PATCH 5/5] mm: cleanup pfn_t usage in track_pfn_insert()

[Dan Williams]

On Tue, Sep 6, 2016 at 1:20 PM, Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> wrote:
> On Tue, 06 Sep 2016 09:49:47 -0700 Dan Williams <dan.j.williams@xxxxxxxxx> wrote:
>
>> Now that track_pfn_insert() is no longer used in the DAX path, it no
>> longer needs to comprehend pfn_t values.
>
> What's the benefit in this?  A pfn *should* have type pfn_t, shouldn't
> it?   Confused.

It should when there's extra information to consider.  I don't mind
leaving it as is, but all the other usages of pfn_t are considering or
passing through the PFN_DEV and PFN_MAP flags.  So, it's a courtesy to
the reader saying "you don't need to worry about pfn_t defined
behavior here, this is just a plain old physical address >>
PAGE_SHIFT"