On Mon, Mar 31, 2025 at 12:54 AM Miroslav Lichvar <[email protected]> wrote: > On Thu, Mar 27, 2025 at 06:32:27PM +0100, Thomas Gleixner wrote: > > On Thu, Mar 27 2025 at 16:42, Miroslav Lichvar wrote: > > > On Thu, Mar 27, 2025 at 10:22:31AM +0100, Thomas Gleixner wrote: > > > To clearly see the difference with the new code, I made an attempt > > > to update the old linux-tktest simulation that was used back when the > > > multiplier adjustment was reworked, but there are too many missing > > > things now and I gave up. > > > > Can you point me to that code? > > It's this thing: https://github.com/mlichvar/linux-tktest > > > It would be probably useful to create a test mechanism which allows to > > exercise all of this in a simulated way so we actually don't have to > > wonder every time we change a bit what the consequences are. > > Yes, that would be very nice if we could run the timekeeping code in a > deterministic simulated environment with a configurable clocksource, > timing of kernel updates, timing and values of injected adjtimex() > calls, etc. The question is how to isolate it.
Miroslav, Have you looked at KUNIT? https://docs.kernel.org/dev-tools/kunit/index.html I've not yet done much with it, but it seems like it might be a good match for moving some of this simulation logic (which has always impressed me) into the kernel tree. thanks -john
