On 11/1/2017 1:53 AM, Tan, Jianfeng wrote: > > >> -----Original Message----- > ... >>> >>>> Regarding the KNI crash, it can be avoid by first checking the exiting >>>> mode(rte_eal_iova_mode()). i.e since legacy driver like KNI need real >>>> physical address to work "now", it can grace full exit on the init time if >>>> mode == IOVA_VA; >>> >>> For the KNI crash, do we really need to correct the case: make those PMDs >>> (requires IOVA_VA) also work well with KNI? >>> - If no, we just need to report RTE_PCI_DRV_IOVA_AS_VA smartly. >>> - If yes, we shall make KNI work with VA. >> >> I would say "yes". But for this immediate release, I would say "no". > > Ferruh, if you also agree on "no" for now, we might fix it like this:
I agree with Jerin, for this release "no", for long term "yes". > > For PMDs like ixgbe/i40e/..., we detect if phys_addrs_available to decide > RTE_PCI_DRV_IOVA_AS_VA in a RTE_INIT function. > > How do you think of that? Instead Santosh's suggestion looks simpler. Detect if kni module inserted, if so set IOVA PA mode. I guess rte_bus_get_iommu_class() can updated for this check as workaround. > > Thanks, > Jianfeng >
