On Mon, Feb 09, 2015 at 03:54:22AM +0100, Rafael J. Wysocki wrote:
> > > > The only remaining issue might be a NMI calling into
> > > > ktime_get_mono_fast_ns() before timekeeping is resumed. Its probably a
> > > > non issue on x86/tsc, but it might be a problem on other platforms
> > > > which turn off devices, clocks, It's not rocket science to prevent
> > > > that.
> I'm not sure if this is what you meant, but here it goes.
>
> The idea is to set up a dummy readout base for the fast timekeeper during
> timekeeping_suspend() such that it will always return the same number of
> cycles.
> After the last timekeeping_update() in timekeeping_suspend() we read the
> clocksource and store the result as cycles_at_suspend. The readout base
> from the current timekeeper is copied onto the dummy and the ->read pointer
> of the dummy is set to a routine unconditionally returning cycles_at_suspend.
> Next, the dummy is passed to update_fast_timekeeper() (which has been modified
> slightly to accept the readout base instead of a whole timekeeper).
>
> Then, if I'm not mistaken, ktime_get_mono_fast_ns() should work until the
> subsequent timekeeping_resume() and then the proper readout base for the
> fast timekeeper will be restored by the timekeeping_update() called right
> after we've cleared timekeeping_suspended.
> Index: linux-pm/kernel/time/timekeeping.c
> ===================================================================
> --- linux-pm.orig/kernel/time/timekeeping.c
> +++ linux-pm/kernel/time/timekeeping.c
> @@ -230,9 +230,7 @@ static inline s64 timekeeping_get_ns_raw
>
> /**
> * update_fast_timekeeper - Update the fast and NMI safe monotonic
> timekeeper.
> - * @tk: The timekeeper from which we take the update
> - * @tkf: The fast timekeeper to update
> - * @tbase: The time base for the fast timekeeper (mono/raw)
> + * @tkr: Timekeeping readout base from which we take the update
> *
> * We want to use this from any context including NMI and tracing /
> * instrumenting the timekeeping code itself.
> @@ -244,11 +242,11 @@ static inline s64 timekeeping_get_ns_raw
> * smp_wmb(); <- Ensure that the last base[1] update is visible
> * tkf->seq++;
> * smp_wmb(); <- Ensure that the seqcount update is visible
> - * update(tkf->base[0], tk);
> + * update(tkf->base[0], tkr);
> * smp_wmb(); <- Ensure that the base[0] update is visible
> * tkf->seq++;
> * smp_wmb(); <- Ensure that the seqcount update is visible
> - * update(tkf->base[1], tk);
> + * update(tkf->base[1], tkr);
> *
> * The reader side does:
> *
> @@ -269,7 +267,7 @@ static inline s64 timekeeping_get_ns_raw
> * slightly wrong timestamp (a few nanoseconds). See
> * @ktime_get_mono_fast_ns.
> */
> -static void update_fast_timekeeper(struct timekeeper *tk)
> +static void update_fast_timekeeper(struct tk_read_base *tkr)
> {
> struct tk_read_base *base = tk_fast_mono.base;
>
> @@ -277,7 +275,7 @@ static void update_fast_timekeeper(struc
> raw_write_seqcount_latch(&tk_fast_mono.seq);
>
> /* Update base[0] */
> - memcpy(base, &tk->tkr, sizeof(*base));
> + memcpy(base, tkr, sizeof(*base));
>
> /* Force readers back to base[0] */
> raw_write_seqcount_latch(&tk_fast_mono.seq);
> @@ -462,7 +460,7 @@ static void timekeeping_update(struct ti
> memcpy(&shadow_timekeeper, &tk_core.timekeeper,
> sizeof(tk_core.timekeeper));
>
> - update_fast_timekeeper(tk);
> + update_fast_timekeeper(&tk->tkr);
> }
>
> /**
> @@ -1251,9 +1249,18 @@ static void timekeeping_resume(void)
> hrtimers_resume();
> }
>
> -static int timekeeping_suspend(void)
> +static struct tk_read_base tkr_dummy;
> +static cycle_t cycles_at_suspend;
> +
> +static cycle_t dummy_clock_read(struct clocksource *cs)
> +{
> + return cycles_at_suspend;
> +}
> +
> +int timekeeping_suspend(void)
> {
> struct timekeeper *tk = &tk_core.timekeeper;
> + struct clocksource *clock = tk->tkr.clock;
> unsigned long flags;
> struct timespec64 delta, delta_delta;
> static struct timespec64 old_delta;
> @@ -1296,6 +1303,14 @@ static int timekeeping_suspend(void)
> }
>
> timekeeping_update(tk, TK_MIRROR);
> +
> + if (!(clock->flags & CLOCK_SOURCE_SUSPEND_NONSTOP)) {
> + memcpy(&tkr_dummy, &tk->tkr, sizeof(tkr_dummy));
> + cycles_at_suspend = tk->tkr.read(clock);
> + tkr_dummy.read = dummy_clock_read;
> + update_fast_timekeeper(&tkr_dummy);
> + }
> +
> write_seqcount_end(&tk_core.seq);
> raw_spin_unlock_irqrestore(&timekeeper_lock, flags);
Yeah, I think that that should work. John any objections?
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
