From: Zach Pfeffer <zpfeffer@codeaurora.org>
To: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Russell King - ARM Linux <linux@arm.linux.org.uk>,
FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>,
linux-arch@vger.kernel.org, dwalker@codeaurora.org,
mel@csn.ul.ie, linux-arm-msm@vger.kernel.org,
linux-kernel@vger.kernel.org, linux-mm@kvack.org,
andi@firstfloor.org, linux-omap@vger.kernel.org,
linux-arm-kernel@lists.infradead.org
Subject: Re: [RFC 1/3 v3] mm: iommu: An API to unify IOMMU, CPU and device memory management
Date: Wed, 14 Jul 2010 22:40:51 -0700 [thread overview]
Message-ID: <20100715054050.GB3623@codeaurora.org> (raw)
In-Reply-To: <m1mxttiki1.fsf@fess.ebiederm.org>
On Wed, Jul 14, 2010 at 06:47:34PM -0700, Eric W. Biederman wrote:
> Zach Pfeffer <zpfeffer@codeaurora.org> writes:
>
> > On Wed, Jul 14, 2010 at 11:05:36PM +0100, Russell King - ARM Linux wrote:
> >> On Wed, Jul 14, 2010 at 01:11:49PM -0700, Zach Pfeffer wrote:
> >> > If the DMA-API contained functions to allocate virtual space separate
> >> > from physical space and reworked how chained buffers functioned it
> >> > would probably work - but then things start to look like the VCM API
> >> > which does graph based map management.
> >>
> >> Every additional virtual mapping of a physical buffer results in
> >> additional cache aliases on aliasing caches, and more workload for
> >> developers to sort out the cache aliasing issues.
> >>
> >> What does VCM to do mitigate that?
> >
> > The VCM ensures that all mappings that map a given physical buffer:
> > IOMMU mappings, CPU mappings and one-to-one device mappings all map
> > that buffer using the same (or compatible) attributes. At this point
> > the only attribute that users can pass is CACHED. In the absence of
> > CACHED all accesses go straight through to the physical memory.
> >
> > The architecture of the VCM allows these sorts of consistency checks
> > to be made since all mappers of a given physical resource are
> > tracked. This is feasible because the physical resources we're
> > tracking are typically large.
>
> On x86 this is implemented in the pat code, and could reasonably be
> generalized to be cross platform.
>
> This is controlled by HAVE_PFNMAP_TRACKING and with entry points
> like track_pfn_vma_new.
>
> Given that we already have an implementation that tracks the cached
> vs non-cached attribute using the dma api. I don't see that the
> API has to change. An implementation of the cached vs non-cached
> status for arm and other architectures is probably appropriate.
>
> It is definitely true that getting your mapping caching attributes
> out of sync can be a problem.
Sure, but we're still stuck with needing lots of scatterlist list
elements and needing to copy them to share physical buffers.
next prev parent reply other threads:[~2010-07-15 5:40 UTC|newest]
Thread overview: 53+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-07-03 5:38 [RFC 1/3 v3] mm: iommu: An API to unify IOMMU, CPU and device memory management Zach Pfeffer
2010-07-03 19:06 ` Eric W. Biederman
2010-07-07 22:44 ` Zach Pfeffer
2010-07-07 23:07 ` Russell King - ARM Linux
2010-07-08 23:59 ` Zach Pfeffer
2010-07-12 1:25 ` FUJITA Tomonori
2010-07-13 5:57 ` Zach Pfeffer
2010-07-13 6:03 ` FUJITA Tomonori
2010-07-13 12:14 ` Zach Pfeffer
2010-07-14 1:59 ` FUJITA Tomonori
2010-07-14 20:11 ` Zach Pfeffer
2010-07-14 22:05 ` Russell King - ARM Linux
2010-07-15 1:29 ` Zach Pfeffer
2010-07-15 1:47 ` Eric W. Biederman
2010-07-15 5:40 ` Zach Pfeffer [this message]
2010-07-15 5:35 ` Zach Pfeffer
2010-07-15 8:55 ` Russell King - ARM Linux
2010-07-16 0:48 ` Tim HRM
2010-07-16 7:58 ` Russell King - ARM Linux
2010-07-17 0:01 ` Larry Bassel
2010-07-19 9:21 ` Tim HRM
2010-07-21 0:44 ` Zach Pfeffer
2010-07-21 1:44 ` Timothy Meade
2010-07-22 4:06 ` Zach Pfeffer
2010-07-19 17:55 ` Michael Bohan
2010-07-19 18:40 ` Russell King - ARM Linux
2010-07-20 22:02 ` stepanm
2010-07-20 22:29 ` Russell King - ARM Linux
2010-07-21 5:49 ` Shilimkar, Santosh
2010-07-21 7:28 ` Russell King - ARM Linux
2010-07-21 7:45 ` Shilimkar, Santosh
2010-07-21 18:04 ` stepanm
2010-07-20 20:45 ` Zach Pfeffer
2010-07-20 20:54 ` Russell King - ARM Linux
2010-07-20 21:56 ` Zach Pfeffer
2010-07-19 6:52 ` Zach Pfeffer
2010-07-19 7:44 ` Eric W. Biederman
2010-07-22 4:25 ` Zach Pfeffer
2010-07-22 7:34 ` Russell King - ARM Linux
2010-07-22 16:25 ` Zach Pfeffer
2010-07-14 23:07 ` FUJITA Tomonori
2010-07-15 1:41 ` Zach Pfeffer
2010-07-19 8:22 ` Russell King - ARM Linux
2010-07-20 10:09 ` FUJITA Tomonori
2010-07-20 22:20 ` Zach Pfeffer
2010-07-21 1:44 ` FUJITA Tomonori
2010-07-22 4:30 ` Zach Pfeffer
2010-07-22 4:43 ` FUJITA Tomonori
2010-07-22 16:44 ` Zach Pfeffer
2010-07-22 7:39 ` Russell King - ARM Linux
2010-07-22 16:28 ` Zach Pfeffer
2010-07-06 15:42 Zach Pfeffer
2010-07-21 5:18 stepanm
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=20100715054050.GB3623@codeaurora.org \
--to=zpfeffer@codeaurora.org \
--cc=andi@firstfloor.org \
--cc=dwalker@codeaurora.org \
--cc=ebiederm@xmission.com \
--cc=fujita.tomonori@lab.ntt.co.jp \
--cc=linux-arch@vger.kernel.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-arm-msm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-omap@vger.kernel.org \
--cc=linux@arm.linux.org.uk \
--cc=mel@csn.ul.ie \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).