[ofa-general] vendor_id in struct ib_device_attr

Jason Gunthorpe jgunthorpe at obsidianresearch.com
Mon Jun 23 12:44:21 PDT 2008


On Mon, Jun 23, 2008 at 03:31:40PM -0400, Jeff Squyres wrote:

>> One question is whether we want to change (fix?) the ipath driver to use
>> OUI.  Strict IB spec compliances says we should.
>
>
> FWIW, OMPI checks these values for QL HCAs:
>
> [QLogic InfiniPath]
> vendor_id = 0x1fc1
> vendor_part_id = 13
> use_eager_rdma = 1
> mtu = 2048
>
> [QLogic InfiniPath]
> vendor_id = 0x1fc1,0x1077
> vendor_part_id = 16,29216
> use_eager_rdma = 1
> mtu = 4096
>
> Translation: vendor_id values 0x1fc1 and 0x1077.

If you are using the vendor_id to key sillicon specific optimizations
then I'd think it is better to look at the PCI ID rather than the
OUI. The OUI belongs to whoever sticks the chip on a card, and has no
relation to the sillicon capabilities..

Jason



More information about the general mailing list