On Thu, Dec 5, 2024 at 7:24 AM Rafael J. Wysocki <rjw@xxxxxxxxxxxxx> wrote: > > Add at least 50 us on top of the requested sleep time in case the > timer can be subject to coalescing, which is consistent with what's > done in user space in this context [2], but for sleeps longer than 5 ms > use 1% of the requested sleep time for this purpose. > > The rationale here is that longer sleeps don't need that much of a timer > precision as a rule and making the timer a more likely candidate for > coalescing in these cases is generally desirable. It starts at 5 ms so > that the delta between the requested sleep time and the effective > deadline is a contiuous function of the former. timerslack_ns defaults to 50,000 ns. So when a user invokes nanosleep(50ms), they get slacked out to 50.050 ms With this patch, if the AML BIOS programmer invokes Sleep(50ms), it gets slacked out to 50.500 ms -- a 10x longer slack period. I have not seen an explanation for why these cases should be treated differently. Len Brown, Intel Open Source Technology Center
