Logan,
> I think this list would be pretty easy to approve if we know the two
> Atom parts work.
They are the Xeon 6 E-core parts, Sierra Forest and Clearwater Forest,
and expose the same 0x09a2 host bridge. Lukas lists both among the
Xeons that support P2PDMA in the commit message of 0ba76b19fd4c
("PCI/P2PDMA: Add Intel QAT, DSA, IAA devices to whitelist") [1].
[1] https://git.kernel.org/torvalds/c/0ba76b19fd4c
Thanks,
Pavel
> -----Original Message-----
> From: Logan Gunthorpe <[email protected]>
> Sent: Friday, October 16, 2026 1:12 PM
> To: Popov, Pavel E <[email protected]>; Bjorn Helgaas
> <[email protected]>
> Cc: Leon Romanovsky <[email protected]>; Jim Chow
> <[email protected]>; Bjorn Helgaas <[email protected]>; Radu
> Rugina <[email protected]>; Alexey Makhalov
> <[email protected]>; Wei Liu <[email protected]>; Michael
> Kelley <[email protected]>; Lukas Wunner <[email protected]>; Rafael
> J. Wysocki <[email protected]>; Nathan Ciobanu
> <[email protected]>; bcm-kernel-feedback-
> [email protected]; [email protected]; linux-
> [email protected]; [email protected]; linux-
> [email protected]
> Subject: Re: [PATCH 0/2] PCI/P2PDMA: Allow P2PDMA in VMware and Hyper-
> V guests on Intel hosts
>
>
>
> On 2026-10-16 13:45, Popov, Pavel E wrote:
> > X86_MATCH_VFM(INTEL_SKYLAKE_X, NULL),
> > X86_MATCH_VFM(INTEL_ICELAKE_X, NULL),
> > X86_MATCH_VFM(INTEL_ICELAKE_D, NULL),
> > X86_MATCH_VFM(INTEL_SAPPHIRERAPIDS_X, NULL),
> > X86_MATCH_VFM(INTEL_EMERALDRAPIDS_X, NULL),
> > X86_MATCH_VFM(INTEL_GRANITERAPIDS_X, NULL),
> > X86_MATCH_VFM(INTEL_GRANITERAPIDS_D, NULL),
> > X86_MATCH_VFM(INTEL_ATOM_CRESTMONT_X, NULL), /*
> Sierra Forest */
> > X86_MATCH_VFM(INTEL_ATOM_DARKMONT_X, NULL), /*
> Clearwater Forest */
> > X86_MATCH_VFM(INTEL_DIAMONDRAPIDS_X, NULL),
>
> I think this list would be pretty easy to approve if we know the two Atom
> parts
> work.
>
> > If you still prefer the kernel command line option you suggested, v3
> > can add it, per-endpoint or global. It works without a patched kernel,
> > but puts the burden on every VMware user to know about and set it,
> > where v1 or v2 just works.
>
> A kernel command line option probably wouldn't be a bad idea independent
> of this. But up to you if you want to try to implement it.
>
> Logan
>
