From: Jason Gunthorpe <jgg@ziepe.ca>
To: "Christian König" <christian.koenig@amd.com>
Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com>,
"Leon Romanovsky" <leon@kernel.org>,
"Bjorn Helgaas" <bhelgaas@google.com>,
"Logan Gunthorpe" <logang@deltatee.com>,
"Chaitanya Kulkarni" <kch@nvidia.com>,
"Greg Kroah-Hartman" <gregkh@linuxfoundation.org>,
"Jens Axboe" <axboe@kernel.dk>,
"Alex Williamson" <alex@shazbot.org>,
"Ankit Agrawal" <ankita@nvidia.com>,
"Jonathan Corbet" <corbet@lwn.net>,
"Shuah Khan" <skhan@linuxfoundation.org>,
"Joerg Roedel (AMD)" <joro@8bytes.org>,
"Will Deacon" <will@kernel.org>,
"Robin Murphy" <robin.murphy@arm.com>,
"Randy Dunlap" <rdunlap@infradead.org>,
"Sumit Semwal" <sumit.semwal@linaro.org>,
linux-pci@vger.kernel.org, linux-kernel@vger.kernel.org,
linux-doc@vger.kernel.org, iommu@lists.linux.dev,
"Tushar Dave" <tdave@nvidia.com>,
linux-media@vger.kernel.org, dri-devel@lists.freedesktop.org,
linaro-mm-sig@lists.linaro.org, linux-rdma@vger.kernel.org,
kvm@vger.kernel.org
Subject: Re: [PATCH v6 18/18] RDMA/mlx5: Ask P2PDMA whether ATS takes a direct peer-to-peer route
Date: Thu, 17 Sep 2026 11:28:08 -0300 [thread overview]
Message-ID: <20260917142808.GH3196566@ziepe.ca> (raw)
In-Reply-To: <f3708893-4760-4157-83a3-313867a94e78@amd.com>
On Thu, Sep 17, 2026 at 04:00:33PM +0200, Christian König wrote:
> > However, at least for the Intel GPUs, ATS enablement is based on the
> > PCIe-side enable bit.
>
> Yeah, I can confirm as well that this is something really exotic.
In most datacenter systems it is only the NIC that initiates the P2P
transfers, so it should not be surprising that NIC HW has more
optimizations for this. In real systems ATS carries such a perf cost
that you do need to be selective about it. I would expect every
serious NIC to have this capability or it will not be perf
competitive.
NIC and GPU are using ATS for completely different reasons.
> Additional to that the setup that ACS seems to be disabled for some
> bridges bit not all also sounds extremely uncommon.
I'm not sure where you saw ACS disabled, you can look rdma-core now
has a tool to help people get their RDMA NICs working in GPU systems:
https://github.com/linux-rdma/rdma-core/blob/master/kernel-boot/rdma_topo
I would view most of that stuff as completely normal at this point,
people have been doing versions of this by hand for almost 10 years
now. Especially the asymmetric translated/untranslated routing.
Jason
next prev parent reply other threads:[~2026-09-17 14:28 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-14 11:22 [PATCH v6 00/18] PCI/P2PDMA: Route peer-to-peer DMA by TLP class Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 01/18] PCI/P2PDMA: Document pdev->p2pdma lifetime rules Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 02/18] PCI/P2PDMA: Document the TLP attribute assumptions Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 03/18] PCI/P2PDMA: Derive routing from directional ACS controls Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 04/18] PCI: Reject unreadable ACS controls in isolation checks Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 05/18] PCI/P2PDMA: Evaluate ACS controls at the path divergence Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 06/18] PCI/P2PDMA: Document directional ACS routing Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 07/18] PCI/P2PDMA: Collect the path's ACS controls before deciding Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 08/18] PCI/P2PDMA: Answer routing per TLP class Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 09/18] PCI/P2PDMA: Route Relaxed Ordering Completions directly Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 10/18] PCI/P2PDMA: Reject Translated Requests blocked by Translation Blocking Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 11/18] PCI/P2PDMA: Route Translated Requests under Direct Translated P2P Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 12/18] PCI/P2PDMA: Log detailed ACS routing diagnostics Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 13/18] PCI/P2PDMA: Add KUnit tests for the ACS routing decisions Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 14/18] PCI/P2PDMA: Test the ACS P2P routing walk Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 15/18] PCI: Add KUnit coverage for ACS isolation checks Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 16/18] PCI/P2PDMA: Document TLP-class routing Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 17/18] dma-buf: Let importers ask how peer-to-peer traffic is routed Leon Romanovsky
2026-09-14 14:27 ` Christian König
2026-09-14 22:10 ` Jason Gunthorpe
2026-09-15 6:28 ` Leon Romanovsky
2026-09-15 6:23 ` Leon Romanovsky
2026-09-14 11:22 ` [PATCH v6 18/18] RDMA/mlx5: Ask P2PDMA whether ATS takes a direct peer-to-peer route Leon Romanovsky
2026-09-17 13:56 ` Thomas Hellström
2026-09-17 14:00 ` Christian König
2026-09-17 14:28 ` Jason Gunthorpe [this message]
2026-09-17 14:02 ` Jason Gunthorpe
2026-09-17 14:16 ` Thomas Hellström
2026-09-17 15:15 ` Jason Gunthorpe
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=20260917142808.GH3196566@ziepe.ca \
--to=jgg@ziepe.ca \
--cc=alex@shazbot.org \
--cc=ankita@nvidia.com \
--cc=axboe@kernel.dk \
--cc=bhelgaas@google.com \
--cc=christian.koenig@amd.com \
--cc=corbet@lwn.net \
--cc=dri-devel@lists.freedesktop.org \
--cc=gregkh@linuxfoundation.org \
--cc=iommu@lists.linux.dev \
--cc=joro@8bytes.org \
--cc=kch@nvidia.com \
--cc=kvm@vger.kernel.org \
--cc=leon@kernel.org \
--cc=linaro-mm-sig@lists.linaro.org \
--cc=linux-doc@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-media@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
--cc=linux-rdma@vger.kernel.org \
--cc=logang@deltatee.com \
--cc=rdunlap@infradead.org \
--cc=robin.murphy@arm.com \
--cc=skhan@linuxfoundation.org \
--cc=sumit.semwal@linaro.org \
--cc=tdave@nvidia.com \
--cc=thomas.hellstrom@linux.intel.com \
--cc=will@kernel.org \
/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
all inboxes | Powered by JetHome®