On 09/18/2018 01:48 PM, Stephen Rothwell wrote: > Hi all, > > On Tue, 18 Sep 2018 10:17:03 +0000 Vakul Garg <vakul.garg@xxxxxxx> wrote: >> >> Got it. >> Thanks for the guidance. > > So, what should I remove? ;-) My (very own personal) preference in general would be that we test / assert the kernel behavior that exists /today/, meaning once we implement support for multi-record peek we add the corresponding test case along with that code. Fwiw, this practice would be consistent with the rest of the kernel selftests development model we have under net (& bpf). Thanks, Daniel
