Re: Severe performance regression in "net: macsec: preserve ingress frame ordering"

2020-08-07 Thread Ryan Cox
On 8/6/20 9:48 PM, Scott Dial wrote: The aes-aesni driver is smart enough to use the FPU if it's not busy and fallback to the CPU otherwise. Unfortunately, the ghash-clmulni driver does not have that kind of logic in it and only provides an async version, so we are forced to use the ghash-generic

Severe performance regression in "net: macsec: preserve ingress frame ordering"

2020-08-06 Thread Ryan Cox
Hello, I have found two performance issues with MACsec on 10 Gb/s links (tested on Intel and Broadcom NICs): 1)  MACsec with encryption is much faster than MACsec without encryption (9.8 vs 7.4 Gb/s) until 5.7, where both have poor performance 2)  5.7 introduced a severe performance impact for