On Wed, Sep 22, 2021 at 01:05:12PM +0200, Peter Zijlstra wrote:
> Use rcu_user_{enter,exit}() calls to provide SMP ordering on context
> tracking state stores:
>
> __context_tracking_exit()
> __this_cpu_write(context_tracking.state, CONTEXT_KERNEL)
> rcu_user_exit()
> rcu_eqs_exit()
> rcu_dynticks_eqs_eit()
> rcu_dynticks_inc()
> atomic_add_return() /* smp_mb */
>
> __context_tracking_enter()
> rcu_user_enter()
> rcu_eqs_enter()
> rcu_dynticks_eqs_enter()
> rcu_dynticks_inc()
> atomic_add_return() /* smp_mb */
> __this_cpu_write(context_tracking.state, state)
>
> This separates USER/KERNEL state with an smp_mb() on each side,
> therefore, a user of context_tracking_state_cpu() can say the CPU must
> pass through an smp_mb() before changing.
>
> Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
For the transformation to negative errno return value and name change
from an RCU perspective:
Acked-by: Paul E. McKenney <paulmck@xxxxxxxxxx>
For the sampling of nohz_full userspace state:
Another approach is for the rcu_data structure's ->dynticks variable to
use the lower two bits to differentiate between idle, nohz_full userspace
and kernel. In theory, inlining should make this zero cost for idle
transition, and should allow you to safely sample nohz_full userspace
state with a load and a couple of memory barriers instead of an IPI.
To make this work nicely, the low-order bits have to be 00 for kernel,
and (say) 01 for idle and 10 for nohz_full userspace. 11 would be an
error.
The trick would be for rcu_user_enter() and rcu_user_exit() to atomically
increment ->dynticks by 2, for rcu_nmi_exit() to increment by 1 and
rcu_nmi_enter() to increment by 3. The state sampling would need to
change accordingly.
Does this make sense, or am I missing something?
Thanx, Paul
> ---
> include/linux/context_tracking_state.h | 12 ++++++++++++
> kernel/context_tracking.c | 7 ++++---
> 2 files changed, 16 insertions(+), 3 deletions(-)
>
> --- a/include/linux/context_tracking_state.h
> +++ b/include/linux/context_tracking_state.h
> @@ -45,11 +45,23 @@ static __always_inline bool context_trac
> {
> return __this_cpu_read(context_tracking.state) == CONTEXT_USER;
> }
> +
> +static __always_inline bool context_tracking_state_cpu(int cpu)
> +{
> + struct context_tracking *ct = per_cpu_ptr(&context_tracking);
> +
> + if (!context_tracking_enabled() || !ct->active)
> + return CONTEXT_DISABLED;
> +
> + return ct->state;
> +}
> +
> #else
> static inline bool context_tracking_in_user(void) { return false; }
> static inline bool context_tracking_enabled(void) { return false; }
> static inline bool context_tracking_enabled_cpu(int cpu) { return false; }
> static inline bool context_tracking_enabled_this_cpu(void) { return false; }
> +static inline bool context_tracking_state_cpu(int cpu) { return CONTEXT_DISABLED; }
> #endif /* CONFIG_CONTEXT_TRACKING */
>
> #endif
> --- a/kernel/context_tracking.c
> +++ b/kernel/context_tracking.c
> @@ -82,7 +82,7 @@ void noinstr __context_tracking_enter(en
> vtime_user_enter(current);
> instrumentation_end();
> }
> - rcu_user_enter();
> + rcu_user_enter(); /* smp_mb */
> }
> /*
> * Even if context tracking is disabled on this CPU, because it's outside
> @@ -149,12 +149,14 @@ void noinstr __context_tracking_exit(enu
> return;
>
> if (__this_cpu_read(context_tracking.state) == state) {
> + __this_cpu_write(context_tracking.state, CONTEXT_KERNEL);
> +
> if (__this_cpu_read(context_tracking.active)) {
> /*
> * We are going to run code that may use RCU. Inform
> * RCU core about that (ie: we may need the tick again).
> */
> - rcu_user_exit();
> + rcu_user_exit(); /* smp_mb */
> if (state == CONTEXT_USER) {
> instrumentation_begin();
> vtime_user_exit(current);
> @@ -162,7 +164,6 @@ void noinstr __context_tracking_exit(enu
> instrumentation_end();
> }
> }
> - __this_cpu_write(context_tracking.state, CONTEXT_KERNEL);
> }
> context_tracking_recursion_exit();
> }
>
>
