From: Jason Gunthorpe <jgg@nvidia.com>
To: Daniel Vetter <daniel@ffwll.ch>
Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com>,
"David Airlie" <airlied@linux.ie>,
"Christian Koenig" <christian.koenig@amd.com>,
dri-devel@lists.freedesktop.org, "Huang Rui" <ray.huang@amd.com>,
"Dan Williams" <dan.j.williams@intel.com>,
"Ralph Campbell" <rcampbell@nvidia.com>,
"Roland Scheidegger" <sroland@vmware.com>
Subject: Re: [PATCH v2] drm/ttm: Do not put non-struct page memory into PUD/PMDs
Date: Fri, 22 Oct 2021 16:00:32 -0300 [thread overview]
Message-ID: <20211022190032.GP2744544@nvidia.com> (raw)
In-Reply-To: <YXFRrvQCho36OTHE@phenom.ffwll.local>
On Thu, Oct 21, 2021 at 01:40:30PM +0200, Daniel Vetter wrote:
> On Wed, Oct 20, 2021 at 11:09:58AM -0300, Jason Gunthorpe wrote:
> > On Wed, Oct 20, 2021 at 08:34:33AM +0200, Thomas Hellström wrote:
> >
> > > Follow up question: If we resurrect this in the proper way (and in that case
> > > only for x86_64) is there something we need to pay particular attention to
> > > WRT the ZONE_DEVICE refcounting fixing you mention above?
> >
> > Similar to PTE it should be completely separated from ZONE_DEVICE.
> >
> > Seeing the special bit set at any level should trigger all page table
> > walkers to never try to get a struct page.
> >
> > Today some of the page table walkers are trying to do this with
> > vma_is_special(), all of those should end up being the Pxx_SPECIAL
> > test instead.
>
> My understanding is that vma_is_special is for platforms which don't have
> pte_special, and hence can't do gup_fast. At least I was assuming this is
> why vma_is_special is a thing, and why some architectures cannot do
> gup_fast.
AFAIK that is what vm_normal_page() is for as the rules for MIXED are
complicated
Once the semi-bogus usages in the PUD/PMD page table walkers are
removed the only remaining users are in arch code around vdso, and I
haven't tried to figure that out..
Jason
prev parent reply other threads:[~2021-10-22 19:00 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-10-19 23:27 [PATCH v2] drm/ttm: Do not put non-struct page memory into PUD/PMDs Jason Gunthorpe
2021-10-20 6:34 ` Thomas Hellström
2021-10-20 6:41 ` Christian König
2021-10-20 6:52 ` Thomas Hellström
2021-10-20 19:37 ` Jason Gunthorpe
2021-10-21 7:04 ` Christian König
2021-10-21 7:05 ` Thomas Hellström
2021-10-21 11:41 ` Daniel Vetter
2021-10-22 18:57 ` Jason Gunthorpe
2021-10-28 15:14 ` Daniel Vetter
2021-10-28 19:42 ` Jason Gunthorpe
2021-11-05 10:12 ` Daniel Vetter
2021-10-20 14:09 ` Jason Gunthorpe
2021-10-21 11:40 ` Daniel Vetter
2021-10-22 19:00 ` Jason Gunthorpe [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20211022190032.GP2744544@nvidia.com \
--to=jgg@nvidia.com \
--cc=airlied@linux.ie \
--cc=christian.koenig@amd.com \
--cc=dan.j.williams@intel.com \
--cc=daniel@ffwll.ch \
--cc=dri-devel@lists.freedesktop.org \
--cc=ray.huang@amd.com \
--cc=rcampbell@nvidia.com \
--cc=sroland@vmware.com \
--cc=thomas.hellstrom@linux.intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.