Hi Logan, Drive-by nits: On Thu, Mar 21, 2019 at 06:06:44PM -0600, Logan Gunthorpe wrote: > Introduce the module parameter 'use_msi' which, when set uses s/set/set,/ > MSI interrupts instead of doorbells for each queue pair (QP). T > he parameter is only available if NTB MSI support is configured into Spurious newline in the middle of "The". > the kernel. We also require there to be more than one memory window > (MW) so that an extra one is available to forward the APIC region. > > To use MSIs, we request one interrupt per QP and forward the MSI address > and data to the peer using scratch pad registers (SPADS) above the MW > spads. (If there are not enough SPADS the MSI interrupt will not be used.) /spads/SPADS/ for consistency > Once registered, we simply use ntb_msi_peer_trigger and the recieving s/recieving/receiving/ > ISR simply queues up the rxc_db_work for the queue. > > This addition can significantly improve performance of ntb_transport. > In a simple, untuned, apples-to-apples comparision using ntb_netdev > and iperf with switchtec hardware, I see 3.88Gb/s without MSI > interrupts and 14.1Gb/s which is a more than 3x improvement. s/which is/with MSI, which is/ Bjorn
