From: Leon Romanovsky <leon@kernel.org>
To: Yonatan Maman <ymaman@nvidia.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>,
"Andrew Morton" <akpm@linux-foundation.org>,
"Jason Gunthorpe" <jgg@ziepe.ca>, "Lyude Paul" <lyude@redhat.com>,
"Danilo Krummrich" <dakr@kernel.org>,
"David Airlie" <airlied@gmail.com>,
"Simona Vetter" <simona@ffwll.ch>,
"Alistair Popple" <apopple@nvidia.com>,
"Ben Skeggs" <bskeggs@nvidia.com>,
"Michael Guralnik" <michaelgur@nvidia.com>,
"Or Har-Toov" <ohartoov@nvidia.com>,
"Daisuke Matsuda" <dskmtsd@gmail.com>,
"Shay Drory" <shayd@nvidia.com>,
linux-mm@kvack.org, linux-rdma@vger.kernel.org,
dri-devel@lists.freedesktop.org, nouveau@lists.freedesktop.org,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 0/5] *** GPU Direct RDMA (P2P DMA) for Device Private Pages ***
Date: Sun, 20 Jul 2025 13:30:03 +0300 [thread overview]
Message-ID: <20250720103003.GH402218@unreal> (raw)
In-Reply-To: <20250718115112.3881129-1-ymaman@nvidia.com>
On Fri, Jul 18, 2025 at 02:51:07PM +0300, Yonatan Maman wrote:
> From: Yonatan Maman <Ymaman@Nvidia.com>
>
> This patch series aims to enable Peer-to-Peer (P2P) DMA access in
> GPU-centric applications that utilize RDMA and private device pages. This
> enhancement reduces data transfer overhead by allowing the GPU to directly
> expose device private page data to devices such as NICs, eliminating the
> need to traverse system RAM, which is the native method for exposing
> device private page data.
>
> To fully support Peer-to-Peer for device private pages, the following
> changes are proposed:
>
> `Memory Management (MM)`
> * Leverage struct pagemap_ops to support P2P page operations: This
> modification ensures that the GPU can directly map device private pages
> for P2P DMA.
> * Utilize hmm_range_fault to support P2P connections for device private
> pages (instead of Page fault)
>
> `IB Drivers`
> Add TRY_P2P_REQ flag for the hmm_range_fault call: This flag indicates the
> need for P2P mapping, enabling IB drivers to efficiently handle P2P DMA
> requests.
>
> `Nouveau driver`
> Add support for the Nouveau p2p_page callback function: This update
> integrates P2P DMA support into the Nouveau driver, allowing it to handle
> P2P page operations seamlessly.
>
> `MLX5 Driver`
> Utilize NIC Address Translation Service (ATS) for ODP memory, to optimize
> DMA P2P for private device pages. Also, when P2P DMA mapping fails due to
> inaccessible bridges, the system falls back to standard DMA, which uses host
> memory, for the affected PFNs
I'm probably missing something very important, but why can't you always
perform p2p if two devices support it? It is strange that IB and not HMM
has a fallback mode.
Thanks
>
> Previous version:
> https://lore.kernel.org/linux-mm/20241201103659.420677-1-ymaman@nvidia.com/
> https://lore.kernel.org/linux-mm/20241015152348.3055360-1-ymaman@nvidia.com/
>
> Yonatan Maman (5):
> mm/hmm: HMM API to enable P2P DMA for device private pages
> nouveau/dmem: HMM P2P DMA for private dev pages
> IB/core: P2P DMA for device private pages
> RDMA/mlx5: Enable P2P DMA with fallback mechanism
> RDMA/mlx5: Enabling ATS for ODP memory
>
> drivers/gpu/drm/nouveau/nouveau_dmem.c | 110 +++++++++++++++++++++++++
> drivers/infiniband/core/umem_odp.c | 4 +
> drivers/infiniband/hw/mlx5/mlx5_ib.h | 6 +-
> drivers/infiniband/hw/mlx5/odp.c | 24 +++++-
> include/linux/hmm.h | 3 +-
> include/linux/memremap.h | 8 ++
> mm/hmm.c | 57 ++++++++++---
> 7 files changed, 195 insertions(+), 17 deletions(-)
>
> --
> 2.34.1
>
next prev parent reply other threads:[~2025-07-20 10:30 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-07-18 11:51 [PATCH v2 0/5] *** GPU Direct RDMA (P2P DMA) for Device Private Pages *** Yonatan Maman
2025-07-18 11:51 ` [PATCH v2 1/5] mm/hmm: HMM API to enable P2P DMA for device private pages Yonatan Maman
2025-07-18 14:17 ` Matthew Wilcox
2025-07-18 14:44 ` Jason Gunthorpe
2025-07-21 0:11 ` Alistair Popple
2025-07-21 13:23 ` Matthew Wilcox
2025-07-22 0:49 ` Alistair Popple
2025-07-23 3:51 ` Jason Gunthorpe
2025-07-23 4:10 ` Alistair Popple
2025-07-24 8:52 ` David Hildenbrand
2025-07-25 0:31 ` Alistair Popple
2025-07-25 9:51 ` David Hildenbrand
2025-08-01 16:40 ` Jason Gunthorpe
2025-08-01 16:50 ` David Hildenbrand
2025-08-01 16:57 ` Jason Gunthorpe
2025-08-04 1:51 ` Alistair Popple
2025-08-05 14:09 ` Jason Gunthorpe
2025-08-04 7:48 ` David Hildenbrand
2025-07-21 6:59 ` Christoph Hellwig
2025-07-22 5:42 ` Yonatan Maman
2025-08-01 16:52 ` Jason Gunthorpe
2025-07-18 11:51 ` [PATCH v2 2/5] nouveau/dmem: HMM P2P DMA for private dev pages Yonatan Maman
2025-07-21 7:00 ` Christoph Hellwig
2025-07-22 5:23 ` Yonatan Maman
2025-07-18 11:51 ` [PATCH v2 3/5] IB/core: P2P DMA for device private pages Yonatan Maman
2025-07-18 11:51 ` [PATCH v2 4/5] RDMA/mlx5: Enable P2P DMA with fallback mechanism Yonatan Maman
2025-07-21 7:03 ` Christoph Hellwig
2025-07-23 3:55 ` Jason Gunthorpe
2025-07-24 7:30 ` Christoph Hellwig
2025-08-01 16:46 ` Jason Gunthorpe
2025-07-18 11:51 ` [PATCH v2 5/5] RDMA/mlx5: Enabling ATS for ODP memory Yonatan Maman
2025-07-20 10:30 ` Leon Romanovsky [this message]
2025-07-20 21:03 ` [PATCH v2 0/5] *** GPU Direct RDMA (P2P DMA) for Device Private Pages *** Yonatan Maman
2025-07-21 6:49 ` Leon Romanovsky
2025-07-23 4:03 ` Jason Gunthorpe
2025-07-23 8:44 ` Leon Romanovsky
2025-07-21 6:54 ` Christoph Hellwig
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=20250720103003.GH402218@unreal \
--to=leon@kernel.org \
--cc=airlied@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=apopple@nvidia.com \
--cc=bskeggs@nvidia.com \
--cc=dakr@kernel.org \
--cc=dri-devel@lists.freedesktop.org \
--cc=dskmtsd@gmail.com \
--cc=jgg@ziepe.ca \
--cc=jglisse@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-rdma@vger.kernel.org \
--cc=lyude@redhat.com \
--cc=michaelgur@nvidia.com \
--cc=nouveau@lists.freedesktop.org \
--cc=ohartoov@nvidia.com \
--cc=shayd@nvidia.com \
--cc=simona@ffwll.ch \
--cc=ymaman@nvidia.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).