On Sun, Aug 11, 2019 at 02:08:52PM -0400, Joel Fernandes wrote:
> Hi Paul, everyone,
>
> I noticed on reading code that the need_heavy_qs check and
> rcu_momentary_dyntick_idle() is only called for !PREEMPT kernels. Don't we
> need to call this for PREEMPT kernels for the benefit of nohz_full CPUs?
>
> Consider the following events:
> 1. Kernel is PREEMPT=y configuration.
> 2. CPU 2 is a nohz_full CPU running only a single task and the tick is off.
> 3. CPU 2 is running only in kernel mode and does not enter user mode or idle.
> 4. Grace period thread running on CPU 3 enter the fqs loop.
> 5. Enough time passes and it sets the need_heavy_qs for CPU2.
> 6. CPU 2 is still in kernel mode but does cond_resched().
> 7. cond_resched() does not call rcu_momentary_dyntick_idle() because PREEMPT=y.
>
> Is 7. not calling rcu_momentary_dyntick_idle() a lost opportunity for the FQS
> loop to detect that the CPU has crossed a quiescent point?
>
> Is this done so that cond_resched() is fast for PREEMPT=y kernels?
The problem is that the definiton of cond_resched() in PREEMPT=y
kernels is as follows:
static inline int _cond_resched(void) { return 0; }
If (but only if!) someone shows a problem in a PREEMPT=y kernel, the
code could be updated to do something like a resched_cpu() earlier
rather than later.
The reason that I do not expect a problem in NO_HZ_FULL=n&&PREEMPT=y
kernels is that the scheduling-clock tick will with high probability
happen when the CPU is not in an RCU read-side critical section, and
this quiescent state will be reported reasonably quickly.
This leaves NO_HZ_FULL=y&&PREEMPT=y kernels. In that case, RCU is
more aggressive about using resched_cpu() on CPUs that have not yet
reported a quiescent state for the current grace period.
So we should be good as is.
Or am I missing a key corner case here?
Thanx, Paul
