Hi Beilei, PSB
Regards, Rory > > Sorry, NACK the patch since there's compilation issue with clang and coding > style issue. > Please check > http://mails.dpdk.org/archives/test-report/2020-January/112422.html and > http://mails.dpdk.org/archives/test-report/2020-January/112412.html > BTW, Doc update is needed in release notes, please also describe what's the > supported input set for the new PCTYPE. > > Beilei > The compilation issue is because for some reason patch 1/2 of this patch-set which affects rte_flow api was not applied before compiling this patch with DPDK test framework. http://mails.dpdk.org/archives/test-report/2020-January/112422.html The errors in above would have been avoided if the first patch of this patch-set had been applied before this 2/2 patch. Not sure why this happened. I've confirmed manually that there are no issues with clang compilation when first patch of set is applied correctly. As regards coding style issue (http://mails.dpdk.org/archives/test-report/2020-January/112412.html). CHECK:ASSIGNMENT_CONTINUATIONS: Assignment operator '=' should be on the previous line #343: FILE: drivers/net/i40e/i40e_flow.c:3059: + filter->input.flow.ip4_l2tpv3oip_flow.session_id + = l2tpv3oip_spec->session_id; If the '=' operator is moved to previous line then the line will exceed 80 characters. There are similar cases earlier in this same file which I followed to keep code consistent throughout. Please advise which style you prefer to be used, line exceeding 80 chars or '=' operator on 2nd line. See an earlier example for this case in same file, drivers/net/i40e/i40e_flow.c:2900, which I used as example for how to base coding style for such case: filter->input.flow.sctp4_flow.verify_tag = sctp_spec->hdr.tag; I will do DOC update for release notes separately and describe the supported input set for new PCTYPE there.
