From: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
To: Chris Wilson <chris@chris-wilson.co.uk>,
Matthew Auld <matthew.william.auld@gmail.com>
Cc: Intel Graphics Development <intel-gfx@lists.freedesktop.org>,
Matthew Auld <matthew.auld@intel.com>
Subject: Re: [Intel-gfx] [PATCH] drm/i915/gem: Allow importing of shmemfs objects into any device
Date: Thu, 21 Jan 2021 11:00:25 +0000 [thread overview]
Message-ID: <5c5f16c9-9885-3eba-ee1b-21e491094acc@linux.intel.com> (raw)
In-Reply-To: <161117638753.7444.16304169600967757321@build.alporthouse.com>
On 20/01/2021 20:59, Chris Wilson wrote:
> Quoting Chris Wilson (2021-01-20 18:06:08)
>> Quoting Matthew Auld (2021-01-20 17:46:10)
>>> On Wed, 20 Jan 2021 at 15:40, Chris Wilson <chris@chris-wilson.co.uk> wrote:
>>>>
>>>> If we import a shmemfs object between devices, for example from
>>>> Tigerlake to DG1, we can simply reuse the native object and its backing
>>>> store.
>>>
>>> Hmmm interesting, so does that include re-using the actual sg mapping
>>> for the backing pages? Does that work out-of-the-box between different
>>> devices assuming we have iommu enabled?
>>
>> Indeed interesting; the dma_addr_t are supposed to be local to a device.
>
> On reflection, we are expected to use cross-device dma_addr_t with
> dma-buf. It's the exporter who assigns the dma_addr_t for the importer
> to use, and they are always given from the original device.
>
> Maybe not so bad. Definitely needs testing to see what happens in
> practice.
What about object migration? I did not spot anything preventing it once
object was exported like this so owning device could move it to device
memory afterwards which would probably be bad.
Regards,
Tvrtko
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
next prev parent reply other threads:[~2021-01-21 11:00 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-01-20 15:40 [Intel-gfx] [PATCH] drm/i915/gem: Allow importing of shmemfs objects into any device Chris Wilson
2021-01-20 17:46 ` Matthew Auld
[not found] ` <161116596844.3420.4138630766326614766@build.alporthouse.com>
2021-01-20 20:59 ` Chris Wilson
2021-01-21 11:00 ` Tvrtko Ursulin [this message]
2021-01-21 11:03 ` Chris Wilson
2021-01-21 11:14 ` Tvrtko Ursulin
2021-01-22 20:09 ` Ruhl, Michael J
2021-01-20 20:53 ` kernel test robot
2021-01-20 21:31 ` [Intel-gfx] ✗ Fi.CI.BUILD: failure for " Patchwork
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=5c5f16c9-9885-3eba-ee1b-21e491094acc@linux.intel.com \
--to=tvrtko.ursulin@linux.intel.com \
--cc=chris@chris-wilson.co.uk \
--cc=intel-gfx@lists.freedesktop.org \
--cc=matthew.auld@intel.com \
--cc=matthew.william.auld@gmail.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 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).