rls2606: aes128nullgmac regression in March

[vrpolakatcisco]

Trending [0] shows something decreased the performance. Bisect in progress, might take some time due to noisy results.

[0] 3n-icx e810 dpdk

[vrpolakatcisco]

Bisect run [1] quite confidently concluded this [2] refactor is the cause. Comparing the telemetry before [3] and after [4], I see the crypto nodes taking longer due to more cache misses in general.

[1] https://github.com/FDio/csit/actions/runs/30263180414/job/89967644720
[2] 45131: crypto: refactor crypto ctx, dispatch and engine integration | https://gerrit.fd.io/r/c/vpp/+/45131
[3] https://logs.fd.io/vex-yul-rot-jenkins-1/csit-vpp-bisect-3n-icx/30263180414/csit_earliest/log.html.gz#s1-s1-s1-s1-s1-t1-k2-k10-k10-k10-k1-k1-k1-k14
[4] https://logs.fd.io/vex-yul-rot-jenkins-1/csit-vpp-bisect-3n-icx/30263180414/csit_middle/3/log.html.gz#s1-s1-s1-s1-s1-t1-k2-k10-k10-k10-k1-k1-k1-k14