[ofa-general][PATCH v2 1/2]mlx4: Multiple completion vectors support

Roland Dreier rdreier at cisco.com
Wed Jun 18 10:48:01 PDT 2008


 > > ?? Nothing ties the CQ event vector number to a CPU number that I know of.

 > I am not sure to follow you. Do you want to say that the suggested
 > implementation in this patch is such that the EQ is not tied to a
 > specific CPU or that there is no proper mapping from the CPU this
 > process is shielded on to the vector (=EQ) number this CQ would be
 > associated with?

Well, both of your statements seem to be true: nothing sets the affinity
for the interrupts created for multiple EQs, and there is not any simply
way to guarantee that CQ vector 5 is sent to CPU 5 that I see.



More information about the general mailing list