On Thu, Aug 6, 2026 at 5:51 PM Steven Rostedt <[email protected]> wrote: > > On Thu, 6 Aug 2026 17:45:00 -0400 > Andrey Grodzovsky <[email protected]> wrote: > > > > Since the selftests are shipped with the kernel code, I think > > > we don't need to handle backward compatibility here. Instead, > > > we can just assume disabling ftrace is no longer supported. > > Incorrect. > > > > > > > Thanks, > > > Song > > > > I tend to agree but I think here[1] Steve and Miroslav asked for > > backward compatability. V1 was actually the way you prefer. > > Right. We test older kernels (think LTS) with newer testing environments. > We need them to be backward compatible.
Alternatively, we can test the LTS kernel with the selftests in the LTS tree. I guess this alternative is not preferred because the selftests in the LTS tree didn't have as good coverage as the newer selftests. However, requiring compatibility in selftests has its own issues: 1) It makes writing and/or maintaining selftests harder. For example, shall the selftest for a new feature fail or skip on older kernels. Either way, some extra work is required to handle this, e.g., someone need to remember that this kernel-test combination is expected to fail, and it is OK. 2) It delays the actual deprecation of a feature. IOW, the feature will somehow stays in the selftests for much longer time. Overall, I think requiring backward compatibility appears to be attractive when we go from "having very little selftests" to "having some selftest coverage". However, not requiring backward compatibility will help us achieve better test coverage in the long run. I hope this makes sense. Thanks, Song
