On 01/23/2017 09:23 AM, Timothy Pearson wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 01/21/2017 02:57 AM, Taiidan@gmx.com wrote:
It seems the PCI-e root ports on the KGPE-D16 have ARI which is needed for SR-IOV, however it is not reported via # lspci -vv I assume that is why the VF's are not assigned to IOMMU groups and thus can't be assigned.
I am running coreboot v4.5 (I forgot to note this as it is another late night)
Tim was the ARI feature implemented when you guys did the programming? https://support.amd.com/TechDocs/44549.pdf
We did not explicitly enable nor verify the ARI feature as we did not need it for our internal machines. As always, this is something that we could definitely work on under contract, and I am always happy to review patches on Gerrit if you choose to tackle this on your own!
Timothy Pearson Raptor Engineering +1 (415) 727-8645 (direct line) +1 (512) 690-0200 (switchboard) https://www.raptorengineering.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iQEcBAEBAgAGBQJYhhHzAAoJEK+E3vEXDOFbwP0IAIbCI5HLMEYNmDOeYyO/IYQT eVFMs+Bd2Q9bJA+x6e0D8MXL6LQzCOjxcg8qbeS4UuI6Kq6HszHqmKKH6rusmPqD Ew2DCffmoQnvsshmL/YvqPSL1SZhYWjjNCdUrBNCAvJAZLEh9Tef4eKpVp9aHjt4 WBvrttU7uFyTf5zbWAPsOFfe2aQ0TNGlCgl+EOqJQWeIaHw+8gu6MPDOIc7aorJ/ 2wC6ReI9iybAAlI0ZfYvsYFtuypqmJEOYmC/9uaU8xYtKOvrNumRPeJw1gB/m1yD Z1joJMC9WQ/1b9+L9D9wRnhFrYp00e5GVZYXWbSM31v81zjg+GKThsAZYwpTnO4= =haoQ -----END PGP SIGNATURE-----
TY for info.
I finally found some real info on ARI. https://us-east.manta.joyent.com/jmc/public/opensolaris/ARChive/FWARC/2010/0... Apparently SR-IOV doesn't actually require it, it is only needed for more than 8 total functions per device. I will continue troubleshooting and get back to you, SR-IOV is an important feature for today's modern servers so I will try to figure it out.
I would DIY but hardware engineering is currently above my paygrade, but when I get a job I am definitely going to do some bankrollin'