I applied this V7 to my github tree: https://github.com/netoptimizer/prototype-kernel/commits/master
And adjusted my benchmark and unit-tests of this skb_array API. Added an unit test of re-sizing to small queue size. https://github.com/netoptimizer/prototype-kernel/blob/81ca09c2a9/kernel/lib/skb_array_test01.c#L130 On Thu, 2 Jun 2016 19:08:14 +0300 "Michael S. Tsirkin" <m...@redhat.com> wrote: [...] > My testing seems to show that at least for the common usecase > in networking, which isn't lockless, circular buffer > with indices does not perform that well, because > each index access causes a cache line to bounce between > CPUs, and index access causes stalls due to the dependency. > > By comparison, an array of pointers where NULL means invalid > and !NULL means valid I confirm and validate that this is a good approach. In module skb_array_parallel01: https://github.com/netoptimizer/prototype-kernel/blob/master/kernel/lib/skb_array_parallel01.c > As this might be useful outside of networking, I implemented > a generic array of void pointers, with a type-safe wrapper for skbs. Nice [...] > The only issue is with calls within a loop using the __ptr_ring_XXX > accessors - in theory compiler could hoist accesses out of the loop. > > Following volatile-considered-harmful.txt I merely > documented that callers that busy-poll should invoke cpu_relax(). > Most people will use the external skb_array_XXX APIs with a spinlock, > so this should not be an issue for them. I would like to see some bulking support... As my experiments[1] show that alf_queue (primarily) can beat skb_array due to bulking support. It seems like an obvious optimization for the virt tun use-case to bulk dequeue SKBs. -- Best regards, Jesper Dangaard Brouer MSc.CS, Principal Kernel Engineer at Red Hat Author of http://www.iptv-analyzer.org LinkedIn: http://www.linkedin.com/in/brouer [1] https://github.com/netoptimizer/prototype-kernel/blob/master/kernel/lib/alf_queue_parallel01.c
