svn commit: r344487 - in head/sys: conf gnu/gcov

[Enji Cooper]

> On Feb 26, 2019, at 8:27 AM, Rodney W. Grimes <freebsd at pdx.rh.CN85.dnsmgr.net> wrote:

…

> Didnt we just remove an inbase, compiling BSD licensed chunk of
> code called DRM and move it to ports.  So if that was possible
> this should be very rapidly applied here and this issue goes away.
> 
> I am still shaking my head over this one.  Yes, there is some
> expediance to this.  Also could it not live on a project
> branch?  Like.. um.. the ZoL project branch?

Kernel gcov provides a lot of value beyond ZoL. Knowing which branches are being hit by analysis, then producing tests (or pruning dead branches if need be), can greatly improve the quality of kernel code, instead of making a discovery like, "I just tested a thing by running a common workload, and it hit these branches by side-effect, but I omitted these sets of branches, which resulted in panics post-release".

Isilon used its own homegrown variant of this for many moons (and it was a mess). This solution is great for developers/testers, minus the concerns I (and others) have over licensing.

Thanks,
-Enji