Re: [dpdk-dev] [PATCH 2/2] lib/timer: relax barrier for status update

["Carrillo, Erik G" <erik.g.carrillo@intel.com>]

> -----Original Message-----
> From: Phil Yang <phil.yang@arm.com>
> Sent: Monday, February 24, 2020 12:42 AM
> To: rsanford@akamai.com; Carrillo, Erik G <erik.g.carrillo@intel.com>;
> dev@dpdk.org
> Cc: david.marchand@redhat.com; Burakov, Anatoly
> <anatoly.burakov@intel.com>; thomas@monjalon.net; jerinj@marvell.com;
> hemant.agrawal@nxp.com; Honnappa.Nagarahalli@arm.com;
> gavin.hu@arm.com; phil.yang@arm.com; nd@arm.com
> Subject: [PATCH 2/2] lib/timer: relax barrier for status update
> 
> Volatile has no ordering semantics. The rte_timer structure defines timer
> status as a volatile variable and uses the rte_r/wmb barrier to guarantee
> inter-thread visibility.
> 
> This patch optimized the volatile operation with c11 atomic operations and
> one-way barrier to save the performance penalty. According to the
> timer_perf_autotest benchmarking results, this patch can uplift 10%~16%
> timer appending performance, 3%~20% timer resetting performance and
> 45% timer callbacks scheduling performance on aarch64 and no loss in
> performance for x86.
> 
> Suggested-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
> Signed-off-by: Phil Yang <phil.yang@arm.com>
> Reviewed-by: Gavin Hu <gavin.hu@arm.com>

Hi Phil,

It seems like the consensus is to generally avoid replacing rte_atomic_* interfaces with the GCC builtins directly.   In other areas of DPDK that are being patched, are the <std_atomic.h> C11 APIs going to be investigated?   It seems like that decision will apply here as well.

Thanks,
Erik

> ---
>  lib/librte_timer/rte_timer.c | 90 +++++++++++++++++++++++++++++++----
> ---------
>  lib/librte_timer/rte_timer.h |  2 +-
>  2 files changed, 65 insertions(+), 27 deletions(-)
> 
> diff --git a/lib/librte_timer/rte_timer.c b/lib/librte_timer/rte_timer.c index
> 269e921..be0262d 100644
> --- a/lib/librte_timer/rte_timer.c
> +++ b/lib/librte_timer/rte_timer.c
> @@ -10,7 +10,6 @@
>  #include <assert.h>
>  #include <sys/queue.h>
> 
> -#include <rte_atomic.h>
>  #include <rte_common.h>
>  #include <rte_cycles.h>
>  #include <rte_eal_memconfig.h>
> @@ -218,7 +217,7 @@ rte_timer_init(struct rte_timer *tim)
> 
>  	status.state = RTE_TIMER_STOP;
>  	status.owner = RTE_TIMER_NO_OWNER;
> -	tim->status.u32 = status.u32;
> +	__atomic_store_n(&tim->status.u32, status.u32,
> __ATOMIC_RELAXED);
>  }
> 
>  /*
> @@ -239,9 +238,9 @@ timer_set_config_state(struct rte_timer *tim,
> 
>  	/* wait that the timer is in correct status before update,
>  	 * and mark it as being configured */
> -	while (success == 0) {
> -		prev_status.u32 = tim->status.u32;
> +	prev_status.u32 = __atomic_load_n(&tim->status.u32,
> __ATOMIC_RELAXED);
> 
> +	while (success == 0) {
>  		/* timer is running on another core
>  		 * or ready to run on local core, exit
>  		 */
> @@ -258,9 +257,20 @@ timer_set_config_state(struct rte_timer *tim,
>  		 * mark it atomically as being configured */
>  		status.state = RTE_TIMER_CONFIG;
>  		status.owner = (int16_t)lcore_id;
> -		success = rte_atomic32_cmpset(&tim->status.u32,
> -					      prev_status.u32,
> -					      status.u32);
> +		/* If status is observed as RTE_TIMER_CONFIG earlier,
> +		 * that's not going to cause any issues because the
> +		 * pattern is read for status then read the other members.
> +		 * In one of the callers to timer_set_config_state
> +		 * (the __rte_timer_reset) we set other members to the
> +		 * structure (period, expire, f, arg) we want these
> +		 * changes to be observed after our change to status.
> +		 * So we need __ATOMIC_ACQUIRE here.
> +		 */
> +		success = __atomic_compare_exchange_n(&tim-
> >status.u32,
> +					      &prev_status.u32,
> +					      status.u32, 0,
> +					      __ATOMIC_ACQUIRE,
> +					      __ATOMIC_RELAXED);
>  	}
> 
>  	ret_prev_status->u32 = prev_status.u32; @@ -279,20 +289,27 @@
> timer_set_running_state(struct rte_timer *tim)
> 
>  	/* wait that the timer is in correct status before update,
>  	 * and mark it as running */
> -	while (success == 0) {
> -		prev_status.u32 = tim->status.u32;
> +	prev_status.u32 = __atomic_load_n(&tim->status.u32,
> __ATOMIC_RELAXED);
> 
> +	while (success == 0) {
>  		/* timer is not pending anymore */
>  		if (prev_status.state != RTE_TIMER_PENDING)
>  			return -1;
> 
>  		/* here, we know that timer is stopped or pending,
> -		 * mark it atomically as being configured */
> +		 * mark it atomically as being running
> +		 */
>  		status.state = RTE_TIMER_RUNNING;
>  		status.owner = (int16_t)lcore_id;
> -		success = rte_atomic32_cmpset(&tim->status.u32,
> -					      prev_status.u32,
> -					      status.u32);
> +		/* RUNNING states are acting as locked states. If the
> +		 * timer is in RUNNING state, the state cannot be changed
> +		 * by other threads. So, we should use ACQUIRE here.
> +		 */
> +		success = __atomic_compare_exchange_n(&tim-
> >status.u32,
> +					      &prev_status.u32,
> +					      status.u32, 0,
> +					      __ATOMIC_ACQUIRE,
> +					      __ATOMIC_RELAXED);
>  	}
> 
>  	return 0;
> @@ -520,10 +537,12 @@ __rte_timer_reset(struct rte_timer *tim, uint64_t
> expire,
> 
>  	/* update state: as we are in CONFIG state, only us can modify
>  	 * the state so we don't need to use cmpset() here */
> -	rte_wmb();
>  	status.state = RTE_TIMER_PENDING;
>  	status.owner = (int16_t)tim_lcore;
> -	tim->status.u32 = status.u32;
> +	/* The "RELEASE" ordering guarantees the memory operations above
> +	 * the status update are observed before the update by all threads
> +	 */
> +	__atomic_store_n(&tim->status.u32, status.u32,
> __ATOMIC_RELEASE);
> 
>  	if (tim_lcore != lcore_id || !local_is_locked)
>  		rte_spinlock_unlock(&priv_timer[tim_lcore].list_lock);
> @@ -600,10 +619,12 @@ __rte_timer_stop(struct rte_timer *tim, int
> local_is_locked,
>  	}
> 
>  	/* mark timer as stopped */
> -	rte_wmb();
>  	status.state = RTE_TIMER_STOP;
>  	status.owner = RTE_TIMER_NO_OWNER;
> -	tim->status.u32 = status.u32;
> +	/* The "RELEASE" ordering guarantees the memory operations above
> +	 * the status update are observed before the update by all threads
> +	 */
> +	__atomic_store_n(&tim->status.u32, status.u32,
> __ATOMIC_RELEASE);
> 
>  	return 0;
>  }
> @@ -637,7 +658,8 @@ rte_timer_stop_sync(struct rte_timer *tim)  int
> rte_timer_pending(struct rte_timer *tim)  {
> -	return tim->status.state == RTE_TIMER_PENDING;
> +	return __atomic_load_n(&tim->status.state,
> +				__ATOMIC_RELAXED) ==
> RTE_TIMER_PENDING;
>  }
> 
>  /* must be called periodically, run all timer that expired */ @@ -739,8
> +761,12 @@ __rte_timer_manage(struct rte_timer_data *timer_data)
>  			/* remove from done list and mark timer as stopped
> */
>  			status.state = RTE_TIMER_STOP;
>  			status.owner = RTE_TIMER_NO_OWNER;
> -			rte_wmb();
> -			tim->status.u32 = status.u32;
> +			/* The "RELEASE" ordering guarantees the memory
> +			 * operations above the status update are observed
> +			 * before the update by all threads
> +			 */
> +			__atomic_store_n(&tim->status.u32, status.u32,
> +				__ATOMIC_RELEASE);
>  		}
>  		else {
>  			/* keep it in list and mark timer as pending */ @@ -
> 748,8 +774,12 @@ __rte_timer_manage(struct rte_timer_data *timer_data)
>  			status.state = RTE_TIMER_PENDING;
>  			__TIMER_STAT_ADD(priv_timer, pending, 1);
>  			status.owner = (int16_t)lcore_id;
> -			rte_wmb();
> -			tim->status.u32 = status.u32;
> +			/* The "RELEASE" ordering guarantees the memory
> +			 * operations above the status update are observed
> +			 * before the update by all threads
> +			 */
> +			__atomic_store_n(&tim->status.u32, status.u32,
> +				__ATOMIC_RELEASE);
>  			__rte_timer_reset(tim, tim->expire + tim->period,
>  				tim->period, lcore_id, tim->f, tim->arg, 1,
>  				timer_data);
> @@ -919,8 +949,12 @@ rte_timer_alt_manage(uint32_t timer_data_id,
>  			/* remove from done list and mark timer as stopped
> */
>  			status.state = RTE_TIMER_STOP;
>  			status.owner = RTE_TIMER_NO_OWNER;
> -			rte_wmb();
> -			tim->status.u32 = status.u32;
> +			/* The "RELEASE" ordering guarantees the memory
> +			 * operations above the status update are observed
> +			 * before the update by all threads
> +			 */
> +			__atomic_store_n(&tim->status.u32, status.u32,
> +				__ATOMIC_RELEASE);
>  		} else {
>  			/* keep it in list and mark timer as pending */
>  			rte_spinlock_lock(
> @@ -928,8 +962,12 @@ rte_timer_alt_manage(uint32_t timer_data_id,
>  			status.state = RTE_TIMER_PENDING;
>  			__TIMER_STAT_ADD(data->priv_timer, pending, 1);
>  			status.owner = (int16_t)this_lcore;
> -			rte_wmb();
> -			tim->status.u32 = status.u32;
> +			/* The "RELEASE" ordering guarantees the memory
> +			 * operations above the status update are observed
> +			 * before the update by all threads
> +			 */
> +			__atomic_store_n(&tim->status.u32, status.u32,
> +				__ATOMIC_RELEASE);
>  			__rte_timer_reset(tim, tim->expire + tim->period,
>  				tim->period, this_lcore, tim->f, tim->arg, 1,
>  				data);
> diff --git a/lib/librte_timer/rte_timer.h b/lib/librte_timer/rte_timer.h index
> c6b3d45..df533fa 100644
> --- a/lib/librte_timer/rte_timer.h
> +++ b/lib/librte_timer/rte_timer.h
> @@ -101,7 +101,7 @@ struct rte_timer
>  {
>  	uint64_t expire;       /**< Time when timer expire. */
>  	struct rte_timer *sl_next[MAX_SKIPLIST_DEPTH];
> -	volatile union rte_timer_status status; /**< Status of timer. */
> +	union rte_timer_status status; /**< Status of timer. */
>  	uint64_t period;       /**< Period of timer (0 if not periodic). */
>  	rte_timer_cb_t f;      /**< Callback function. */
>  	void *arg;             /**< Argument to callback function. */
> --
> 2.7.4