Re: [PATCH v0.8 4/6] sched/umcg, lib/umcg: implement libumcg

From: Tao Zhou
Date: Mon Nov 08 2021 - 02:11:37 EST


On Sun, Nov 07, 2021 at 10:27:02AM -0800, Peter Oskolkov wrote:
> On Sun, Nov 7, 2021 at 8:33 AM Tao Zhou <tao.zhou@xxxxxxxxx> wrote:
> >
> > On Thu, Nov 04, 2021 at 12:58:02PM -0700, Peter Oskolkov wrote:
> >
> > > +/* Update the state variable, set new timestamp. */
> > > +static bool umcg_update_state(uint64_t *state, uint64_t *prev, uint64_t next)
> > > +{
> > > + uint64_t prev_ts = (*prev) >> (64 - UMCG_STATE_TIMESTAMP_BITS);
> > > + struct timespec now;
> > > + uint64_t next_ts;
> > > + int res;
> > > +
> > > + /*
> > > + * clock_gettime(CLOCK_MONOTONIC, ...) takes less than 20ns on a
> > > + * typical Intel processor on average, even when run concurrently,
> > > + * so the overhead is low enough for most applications.
> > > + *
> > > + * If this is still too high, `next_ts = prev_ts + 1` should work
> > > + * as well. The only real requirement is that the "timestamps" are
> > > + * uniqueue per thread within a reasonable time frame.
> > > + */
> > > + res = clock_gettime(CLOCK_MONOTONIC, &now);
> > > + assert(!res);
> > > + next_ts = (now.tv_sec * NSEC_PER_SEC + now.tv_nsec) >>
> > > + UMCG_STATE_TIMESTAMP_GRANULARITY;
> > > +
> > > + /* Cut higher order bits. */
> > > + next_ts &= ((1ULL << UMCG_STATE_TIMESTAMP_BITS) - 1);
> >
> > This is the right cut.. The same to the kernel side.
>
> Yes, thanks!
>
> >
> > > +
> > > + if (next_ts == prev_ts)
> > > + ++next_ts;
> > > +
> > > +#ifndef NDEBUG
> > > + if (prev_ts > next_ts) {
> > > + fprintf(stderr, "%s: time goes back: prev_ts: %lu "
> > > + "next_ts: %lu diff: %lu\n", __func__,
> > > + prev_ts, next_ts, prev_ts - next_ts);
> > > + }
> > > +#endif
> > > +
> > > + /* Remove old timestamp, if any. */
> > > + next &= ((1ULL << (64 - UMCG_STATE_TIMESTAMP_BITS)) - 1);
> > > +
> > > + /* Set the new timestamp. */
> > > + next |= (next_ts << (64 - UMCG_STATE_TIMESTAMP_BITS));
> > > +
> > > + /*
> > > + * TODO: review whether memory order below can be weakened to
> > > + * memory_order_acq_rel for success and memory_order_acquire for
> > > + * failure.
> > > + */
> > > + return atomic_compare_exchange_strong_explicit(state, prev, next,
> > > + memory_order_seq_cst, memory_order_seq_cst);
> > > +}
> > > +
> >
> > > +static void task_unlock(struct umcg_task_tls *task, uint64_t expected_state,
> > > + uint64_t new_state)
> > > +{
> > > + bool ok;
> > > + uint64_t next;
> > > + uint64_t prev = atomic_load_explicit(&task->umcg_task.state_ts,
> > > + memory_order_acquire);
> > > +
> > > + next = ((prev & ~UMCG_TASK_STATE_MASK_FULL) | new_state) & ~UMCG_TF_LOCKED;
> >
> > Use UMCG_TASK_STATE_MASK instead and the other state flag can be checked.
>
> Why? We want to clear the TF_LOCKED flag and keep every other bit of
> state, including other state flags (but excluding timestamp).

This one is different because the task state flag is also masked by
UMCG_TASK_STATE_MASK_FULL. But no effect I think. Document says that

'
PREEMPTED: the userspace indicates it wants the worker to be preempted;
there are no situations when both LOCKED and PREEMPTED flags are set at
the same time.
'

The task state flag this time only have UMCG_TF_LOCKED be set even if we
cleared the task state flag.
>
>
> >
> > All others places that use UMCG_TASK_STATE_MASK_FULL to mask to check
> > the task state may seems reasonable if the state flag not allowed to
> > be set when we check that task state, otherwise use UMCG_TASK_STATE_MASK
> > will be enough.
> >
> > Not sure.
> >
> >
> > Thanks,
> > Tao
> > > + assert(next != prev);
> > > + assert((prev & UMCG_TASK_STATE_MASK_FULL & ~UMCG_TF_LOCKED) == expected_state);
> > > +
> > > + ok = umcg_update_state(&task->umcg_task.state_ts, &prev, next);
> > > + assert(ok);
> > > +}