From: Jason Gunthorpe <jgg@nvidia.com>
To: "Christian König" <christian.koenig@amd.com>
Cc: "Kasireddy, Vivek" <vivek.kasireddy@intel.com>,
Simona Vetter <simona.vetter@ffwll.ch>,
"dri-devel@lists.freedesktop.org"
<dri-devel@lists.freedesktop.org>,
"intel-xe@lists.freedesktop.org" <intel-xe@lists.freedesktop.org>,
Bjorn Helgaas <bhelgaas@google.com>,
Logan Gunthorpe <logang@deltatee.com>,
"linux-pci@vger.kernel.org" <linux-pci@vger.kernel.org>
Subject: Re: [PATCH v4 1/5] PCI/P2PDMA: Don't enforce ACS check for device functions of Intel GPUs
Date: Mon, 22 Sep 2025 09:20:18 -0300 [thread overview]
Message-ID: <20250922122018.GU1391379@nvidia.com> (raw)
In-Reply-To: <045c6892-9b15-4f31-aa6a-1f45528500f1@amd.com>
On Mon, Sep 22, 2025 at 01:22:49PM +0200, Christian König wrote:
> Well what exactly is happening here? You have a PF assigned to the
> host and a VF passed through to a guest, correct?
>
> And now the PF (from the host side) wants to access a BAR of the VF?
Not quite.
It is a GPU so it has a pool of VRAM. The PF can access all VRAM and
the VF can access some VRAM.
They want to get a DMABUF handle for a bit of the VF's reachable VRAM
that the PF can import and use through it's own funciton.
The use of the VF's BAR in this series is an ugly hack. The PF never
actually uses the VF BAR, it just hackily converts the dma_addr_t back
to CPU physical and figures out where it is in the VRAM pool and then
uses a PF centric address for it.
All they want is either the actual VRAM address or the CPU physical.
Jason
next prev parent reply other threads:[~2025-09-22 12:20 UTC|newest]
Thread overview: 62+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-09-15 7:21 [PATCH v4 0/5] drm/xe/sriov: Don't migrate dmabuf BO to System RAM if P2P check succeeds Vivek Kasireddy
2025-09-15 7:21 ` [PATCH v4 1/5] PCI/P2PDMA: Don't enforce ACS check for device functions of Intel GPUs Vivek Kasireddy
2025-09-15 15:33 ` Logan Gunthorpe
2025-09-16 17:34 ` Bjorn Helgaas
2025-09-16 17:59 ` Jason Gunthorpe
2025-09-16 17:57 ` Jason Gunthorpe
2025-09-18 6:16 ` Kasireddy, Vivek
2025-09-18 12:04 ` Jason Gunthorpe
2025-09-19 6:22 ` Kasireddy, Vivek
2025-09-19 12:29 ` Jason Gunthorpe
2025-09-22 6:59 ` Kasireddy, Vivek
2025-09-22 11:22 ` Christian König
2025-09-22 12:20 ` Jason Gunthorpe [this message]
2025-09-22 12:25 ` Christian König
2025-09-22 12:29 ` Jason Gunthorpe
2025-09-22 13:20 ` Christian König
2025-09-22 13:27 ` Jason Gunthorpe
2025-09-22 13:57 ` Christian König
2025-09-22 14:00 ` Jason Gunthorpe
2025-09-23 5:53 ` Kasireddy, Vivek
2025-09-23 6:25 ` Matthew Brost
2025-09-23 6:44 ` Matthew Brost
2025-09-23 7:52 ` Christian König
2025-09-23 12:15 ` Jason Gunthorpe
2025-09-23 12:45 ` Christian König
2025-09-23 13:12 ` Jason Gunthorpe
2025-09-23 13:28 ` Christian König
2025-09-23 13:38 ` Jason Gunthorpe
2025-09-23 13:48 ` Christian König
2025-09-23 23:02 ` Matthew Brost
2025-09-24 8:29 ` Christian König
2025-09-24 6:50 ` Kasireddy, Vivek
2025-09-24 7:21 ` Christian König
2025-09-25 3:56 ` Kasireddy, Vivek
2025-09-25 10:51 ` Thomas Hellström
2025-09-25 11:28 ` Christian König
2025-09-25 13:11 ` Thomas Hellström
2025-09-25 13:33 ` Jason Gunthorpe
2025-09-25 15:40 ` Thomas Hellström
2025-09-25 15:55 ` Jason Gunthorpe
2025-09-26 6:12 ` Kasireddy, Vivek
2025-09-23 13:36 ` Christoph Hellwig
2025-09-23 6:01 ` Kasireddy, Vivek
2025-09-22 12:12 ` Jason Gunthorpe
2025-09-24 16:13 ` Simon Richter
2025-09-24 17:12 ` Jason Gunthorpe
2025-09-25 4:06 ` Kasireddy, Vivek
2025-09-15 7:21 ` [PATCH v4 2/5] drm/xe/dmabuf: Don't migrate BO to System RAM if P2P check succeeds Vivek Kasireddy
2025-09-16 20:01 ` Thomas Hellström
2025-09-15 7:21 ` [PATCH v4 3/5] drm/xe/pf: Add a helper function to get a VF's backing object in LMEM Vivek Kasireddy
2025-09-16 19:58 ` Matthew Brost
2025-09-16 20:06 ` Thomas Hellström
2025-09-17 7:10 ` Thomas Hellström
2025-09-15 7:21 ` [PATCH v4 4/5] drm/xe/bo: Create new dma_addr array for dmabuf BOs associated with VFs Vivek Kasireddy
2025-09-16 19:08 ` Matthew Brost
2025-09-16 19:44 ` Matthew Brost
2025-09-15 7:21 ` [PATCH v4 5/5] drm/xe/pt: Add an additional check for dmabuf BOs while doing bind Vivek Kasireddy
2025-09-16 19:13 ` Matthew Brost
2025-09-15 7:32 ` ✗ CI.checkpatch: warning for drm/xe/sriov: Don't migrate dmabuf BO to System RAM if P2P check succeeds Patchwork
2025-09-15 7:33 ` ✓ CI.KUnit: success " Patchwork
2025-09-15 8:16 ` ✓ Xe.CI.BAT: " Patchwork
2025-09-15 10:19 ` ✗ Xe.CI.Full: failure " 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=20250922122018.GU1391379@nvidia.com \
--to=jgg@nvidia.com \
--cc=bhelgaas@google.com \
--cc=christian.koenig@amd.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=intel-xe@lists.freedesktop.org \
--cc=linux-pci@vger.kernel.org \
--cc=logang@deltatee.com \
--cc=simona.vetter@ffwll.ch \
--cc=vivek.kasireddy@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 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).