Messages in this thread | | | From | "Luis R. Rodriguez" <> | Date | Mon, 6 Jul 2015 23:53:20 -0700 | Subject | Re: [PATCH v2 2/2] x86/mm/pat, drivers/media/ivtv: move pat warn and replace WARN() with pr_warn() |
| |
On Mon, Jul 6, 2015 at 5:44 PM, Luis R. Rodriguez <mcgrof@suse.com> wrote: > If we really wanted to we could consider arch_phys_wc_add()
I mean adding a __arch_phys_wc_add() which does not check for pat_enabled().
> and > deal with that this will not check for pat_enabled() and forces MTRR... > I think Andy Luto won't like that very much though ? I at least don't > like it since we did all this work to finally leave only 1 piece of > code with direct MTRR access... Seems a bit sad. Since ipath will > be removed we'd have only ivtv driver using this API, I am not sure if > its worth it.
Since ipath is going away soon we'd just have one driver with the icky #ifdef code. I'd rather see that and then require semantics / grammer rules to require ioremap_wc() when used with arch_phys_wc_add(). I don't think ivtv is worth to consider breaking the semantics and requirements.
> Thoughts?
Luis
| |