Change Details

Combining non-LRO mbufs results in a 12% throughput increase, this does a similar change for LROed mbuf chains. This should mostly result in fewer calls.

Combining non-LRO mbufs results in a 12% throughput increase, Chain LROed mbufs into a chain before passing onto if_input() to reduce number of if_input() calls. No API changes.

Combining non-LRO mbufs results in a 12% throughput increase, this does a similarChain LROed mbufs into a change for LROed mbuf chains. This should mostlyin before passing onto if_input() to reduce number of if_input() calls. result in fewer callsNo API changes.