Persistent loongarch-unit-testing errors do not seem to be related to the patch since the first test failure happens before the bpf_autotest is even reached: https://mails.dpdk.org/archives/test-report/2026-August/1034609.html