Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1648132 > unrolled thread

Re: [GIT PULL] x86/mm changes for v4.12

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [GIT PULL] x86/mm changes for v4.12 "Kirill A. Shutemov" <kirill@shutemov.name> - 2017-05-23 16:30 +0200
    Re: [GIT PULL] x86/mm changes for v4.12 Ingo Molnar <mingo@kernel.org> - 2017-05-24 08:30 +0200
      Re: [GIT PULL] x86/mm changes for v4.12 "Kirill A. Shutemov" <kirill@shutemov.name> - 2017-05-24 10:30 +0200

#1648132 — Re: [GIT PULL] x86/mm changes for v4.12

On Mon, May 01, 2017 at 02:00:53PM +0200, Ingo Molnar wrote:
>  - Continued Intel 5-level paging enablement: in particular the conversion 
>    of x86 GUP to the generic GUP code. (Kirill A. Shutemov)
> 
...
> 
> I'd like to draw special attention to this late MM commit:
> 
>   71389703839e: mm, zone_device: Replace {get, put}_zone_device_page() with a single reference to fix pmem crash
> 
> ... which is a late fix, and which I believe is the right thing to do - but please 
> double check.

I've just noticied that the x86 GUP conversion haven't made it's way to
v4.12: the fix got applied (71389703839e), but the final commit that does
convertion keeped reverted -- 6dd29b3df975.

I guess it's too late for v4.12...

Ingo could you re-apply 2947ba054a4d ("x86/mm/gup: Switch GUP to the
generic get_user_page_fast() implementation") to be queued to v4.13?

-- 
 Kirill A. Shutemov

[toc] | [next] | [standalone]

#1649149

* Kirill A. Shutemov <kirill@shutemov.name> wrote:

> On Mon, May 01, 2017 at 02:00:53PM +0200, Ingo Molnar wrote:
> >  - Continued Intel 5-level paging enablement: in particular the conversion 
> >    of x86 GUP to the generic GUP code. (Kirill A. Shutemov)
> > 
> ...
> > 
> > I'd like to draw special attention to this late MM commit:
> > 
> >   71389703839e: mm, zone_device: Replace {get, put}_zone_device_page() with a single reference to fix pmem crash
> > 
> > ... which is a late fix, and which I believe is the right thing to do - but please 
> > double check.
> 
> I've just noticied that the x86 GUP conversion haven't made it's way to
> v4.12: the fix got applied (71389703839e), but the final commit that does
> convertion keeped reverted -- 6dd29b3df975.
> 
> I guess it's too late for v4.12...
> 
> Ingo could you re-apply 2947ba054a4d ("x86/mm/gup: Switch GUP to the
> generic get_user_page_fast() implementation") to be queued to v4.13?

There were some late problems and I wanted x86/mm to be regression-free - we don't 
have all the 5-level paging patches applied yet so the exact cut we made at v4.12 
does not matter much - v4.13 is going to be the real deal.

Could you please make the re-application part of the remaining 5-level paging 
patches?

Thanks,

	Ingo

[toc] | [prev] | [next] | [standalone]

#1649251

On Wed, May 24, 2017 at 08:25:26AM +0200, Ingo Molnar wrote:
> 
> * Kirill A. Shutemov <kirill@shutemov.name> wrote:
> 
> > On Mon, May 01, 2017 at 02:00:53PM +0200, Ingo Molnar wrote:
> > >  - Continued Intel 5-level paging enablement: in particular the conversion 
> > >    of x86 GUP to the generic GUP code. (Kirill A. Shutemov)
> > > 
> > ...
> > > 
> > > I'd like to draw special attention to this late MM commit:
> > > 
> > >   71389703839e: mm, zone_device: Replace {get, put}_zone_device_page() with a single reference to fix pmem crash
> > > 
> > > ... which is a late fix, and which I believe is the right thing to do - but please 
> > > double check.
> > 
> > I've just noticied that the x86 GUP conversion haven't made it's way to
> > v4.12: the fix got applied (71389703839e), but the final commit that does
> > convertion keeped reverted -- 6dd29b3df975.
> > 
> > I guess it's too late for v4.12...
> > 
> > Ingo could you re-apply 2947ba054a4d ("x86/mm/gup: Switch GUP to the
> > generic get_user_page_fast() implementation") to be queued to v4.13?
> 
> There were some late problems and I wanted x86/mm to be regression-free - we don't 
> have all the 5-level paging patches applied yet so the exact cut we made at v4.12 
> does not matter much - v4.13 is going to be the real deal.
> 
> Could you please make the re-application part of the remaining 5-level paging 
> patches?

Sure. Will post later today.

-- 
 Kirill A. Shutemov

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web