Jason> I'm sure this must have been answered, but given a PCI
Jason> domain:bus:device:function tuple and a DMA address,
Jason> shouldn't any effects of an IOMMU be easially duplicated in
Jason> software to result in a cpu-bus physical address? Ie on
Jason> AMD64 it is just a matter of following the GART tables in
Jason> software - assuming the address in question hits the GART
Jason> region (which for ipath, I expect, it never would)
Yes, you could do this. However there's no exported interface for
reversing a DMA mapping. And it seems like a lot of unneeded
complexity to add -- why not just avoid the DMA mapping in the first place?
- R.
_______________________________________________
openib-general mailing list
[email protected]
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
