[dpdk-dev] [PATCH v3 5/9] ring: introduce HTS ring mode

[Ananyev, Konstantin]

Hi Honnappa,

> 
> Hi Konstantin,
> 	Few nits/comments inline.
> 
> <snip>
> 
> > diff --git a/lib/librte_ring/rte_ring_hts.h b/lib/librte_ring/rte_ring_hts.h new
> > file mode 100644 index 000000000..062d7be6c
> > --- /dev/null
> > +++ b/lib/librte_ring/rte_ring_hts.h
> > @@ -0,0 +1,210 @@
> > +/* SPDX-License-Identifier: BSD-3-Clause
> > + *
> > + * Copyright (c) 2010-2017 Intel Corporation
> > + * Copyright (c) 2007-2009 Kip Macy kmacy at freebsd.org
> > + * All rights reserved.
> > + * Derived from FreeBSD's bufring.h
> > + * Used as BSD-3 Licensed with permission from Kip Macy.
> > + */
> > +
> > +#ifndef _RTE_RING_HTS_H_
> > +#define _RTE_RING_HTS_H_
> > +
> > +/**
> > + * @file rte_ring_hts.h
> > + * @b EXPERIMENTAL: this API may change without prior notice
> > + * It is not recommended to include this file directly.
> > + * Please include <rte_ring.h> instead.
> > + *
> > + * Contains functions for serialized, aka Head-Tail Sync (HTS) ring mode.
> > + * In that mode enqueue/dequeue operation is fully serialized:
> > + * at any given moement only one enqueue/dequeue operation can proceed.
>                                  ^^^^^^^^ moment
> > + * This is achieved by thread is allowed to proceed with changing
>                                             ^^^^^^^^^^^^^^ allowing a thread
> > +head.value
> > + * only when head.value == tail.value.
> > + * Both head and tail values are updated atomically (as one 64-bit value).
> > + * To achieve that 64-bit CAS is used by head update routine.
> > + */
> > +
> > +#ifdef __cplusplus
> > +extern "C" {
> > +#endif
> > +
> > +#include <rte_ring_hts_generic.h>
> > +
> 
> <snip>
> 
> > diff --git a/lib/librte_ring/rte_ring_hts_generic.h
> > b/lib/librte_ring/rte_ring_hts_generic.h
> > new file mode 100644
> > index 000000000..da08f1d94
> > --- /dev/null
> > +++ b/lib/librte_ring/rte_ring_hts_generic.h
> > @@ -0,0 +1,198 @@
> > +/* SPDX-License-Identifier: BSD-3-Clause
> > + *
> > + * Copyright (c) 2010-2020 Intel Corporation
> > + * Copyright (c) 2007-2009 Kip Macy kmacy at freebsd.org
> > + * All rights reserved.
> > + * Derived from FreeBSD's bufring.h
> > + * Used as BSD-3 Licensed with permission from Kip Macy.
> > + */
> > +
> > +#ifndef _RTE_RING_HTS_GENERIC_H_
> > +#define _RTE_RING_HTS_GENERIC_H_
> > +
> > +/**
> > + * @file rte_ring_hts_generic.h
> > + * It is not recommended to include this file directly,
> > + * include <rte_ring.h> instead.
> > + * Contains internal helper functions for head/tail sync (HTS) ring mode.
> > + * For more information please refer to <rte_ring_hts.h>.
> > + */
> > +
> > +static __rte_always_inline void
> > +__rte_ring_hts_update_tail(struct rte_ring_hts_headtail *ht, uint32_t num,
> > +	uint32_t enqueue)
> > +{
> > +	union rte_ring_ht_pos p;
> > +
> > +	if (enqueue)
> > +		rte_smp_wmb();
> > +	else
> > +		rte_smp_rmb();
> > +
> > +	p.raw = rte_atomic64_read((rte_atomic64_t *)(uintptr_t)&ht-
> > >ht.raw);
> This read can be avoided if the new head can be returned from '__rte_ring_hts_head_wait'.

Yes, or even cur_head and num should be enough to avoid read here.

> 
> > +
> > +	p.pos.head = p.pos.tail + num;
> > +	p.pos.tail = p.pos.head;
> > +
> > +	rte_atomic64_set((rte_atomic64_t *)(uintptr_t)&ht->ht.raw, p.raw); }
> Why not use 32b atomic operation here and update just the tail?

Agree, this code-path can do just 32-bit store (as head is not going to change).  

> 
> > +
> > +/**
> > + * @internal waits till tail will become equal to head.
> > + * Means no writer/reader is active for that ring.
> > + * Suppose to work as serialization point.
> > + */
> > +static __rte_always_inline void
> > +__rte_ring_hts_head_wait(const struct rte_ring_hts_headtail *ht,
> > +		union rte_ring_ht_pos *p)
> > +{
> > +	p->raw = rte_atomic64_read((rte_atomic64_t *)
> > +			(uintptr_t)&ht->ht.raw);
> > +
> > +	while (p->pos.head != p->pos.tail) {
> > +		rte_pause();
> > +		p->raw = rte_atomic64_read((rte_atomic64_t *)
> > +				(uintptr_t)&ht->ht.raw);
> > +	}
> > +}
> > +
> > +/**
> > + * @internal This function updates the producer head for enqueue
> > + *
> > + * @param r
> > + *   A pointer to the ring structure
> > + * @param is_sp
> > + *   Indicates whether multi-producer path is needed or not
> > + * @param n
> > + *   The number of elements we will want to enqueue, i.e. how far should the
> > + *   head be moved
> > + * @param behavior
> > + *   RTE_RING_QUEUE_FIXED:    Enqueue a fixed number of items from a ring
> > + *   RTE_RING_QUEUE_VARIABLE: Enqueue as many items as possible from
> > ring
> > + * @param old_head
> > + *   Returns head value as it was before the move, i.e. where enqueue starts
> > + * @param new_head
> > + *   Returns the current/new head value i.e. where enqueue finishes

Ups, copy/paste thing - will remove.

> Would be good to return the new_head from this function and use it in '__rte_ring_hts_update_tail'.

I think old_head + num should be enough here (see above).

> 
> > + * @param free_entries
> > + *   Returns the amount of free space in the ring BEFORE head was moved
> > + * @return
> > + *   Actual number of objects enqueued.
> > + *   If behavior == RTE_RING_QUEUE_FIXED, this will be 0 or n only.
> > + */
> Minor, suggest removing the elaborate comments, it is not required and difficult to maintain.
> I think we should do the same thing for other files too.

Sorry, didn't get you here: what exactly do you suggest to remove?

> 
> > +static __rte_always_inline unsigned int
> > +__rte_ring_hts_move_prod_head(struct rte_ring *r, unsigned int num,
> > +	enum rte_ring_queue_behavior behavior, uint32_t *old_head,
> > +	uint32_t *free_entries)
> > +{
> > +	uint32_t n;
> > +	union rte_ring_ht_pos np, op;
> > +
> > +	const uint32_t capacity = r->capacity;
> > +
> > +	do {
> > +		/* Reset n to the initial burst count */
> > +		n = num;
> > +
> > +		/* wait for tail to be equal to head */
> > +		__rte_ring_hts_head_wait(&r->hts_prod, &op);
> > +
> > +		/* add rmb barrier to avoid load/load reorder in weak
> > +		 * memory model. It is noop on x86
> > +		 */
> > +		rte_smp_rmb();
> > +
> > +		/*
> > +		 *  The subtraction is done between two unsigned 32bits
> > value
> > +		 * (the result is always modulo 32 bits even if we have
> > +		 * *old_head > cons_tail). So 'free_entries' is always between
> > 0
> > +		 * and capacity (which is < size).
> > +		 */
> > +		*free_entries = capacity + r->cons.tail - op.pos.head;
> > +
> > +		/* check that we have enough room in ring */
> > +		if (unlikely(n > *free_entries))
> > +			n = (behavior == RTE_RING_QUEUE_FIXED) ?
> > +					0 : *free_entries;
> > +
> > +		if (n == 0)
> > +			break;
> > +
> > +		np.pos.tail = op.pos.tail;
> > +		np.pos.head = op.pos.head + n;
> > +
> > +	} while (rte_atomic64_cmpset(&r->hts_prod.ht.raw,
> > +			op.raw, np.raw) == 0);
> I think we can use 32b atomic operation here and just update the head.

I think we have to do proper 64 bit CAS here, otherwise ABA race could arise:
Thread reads head/tail values, then get suspended just before CAS instruction for a while.
Thread resumes when ring head value is equal to thread's local head value,
but tail differs (some other thread enqueuing into the ring).
If we'll do CAS just for head - it would succeed, though it shouldn't.    
I understand that with 32-bit head/tail values probability of such situation
is really low, but still.

> 
> > +
> > +	*old_head = op.pos.head;
> > +	return n;
> > +}
> > +
> > +/**
> > + * @internal This function updates the consumer head for dequeue
> > + *
> > + * @param r
> > + *   A pointer to the ring structure
> > + * @param is_sc
> > + *   Indicates whether multi-consumer path is needed or not
> > + * @param n
> > + *   The number of elements we will want to enqueue, i.e. how far should the
> > + *   head be moved
> > + * @param behavior
> > + *   RTE_RING_QUEUE_FIXED:    Dequeue a fixed number of items from a
> > ring
> > + *   RTE_RING_QUEUE_VARIABLE: Dequeue as many items as possible from
> > ring
> > + * @param old_head
> > + *   Returns head value as it was before the move, i.e. where dequeue starts
> > + * @param new_head
> > + *   Returns the current/new head value i.e. where dequeue finishes
> > + * @param entries
> > + *   Returns the number of entries in the ring BEFORE head was moved
> > + * @return
> > + *   - Actual number of objects dequeued.
> > + *     If behavior == RTE_RING_QUEUE_FIXED, this will be 0 or n only.
> > + */
> > +static __rte_always_inline unsigned int
> > +__rte_ring_hts_move_cons_head(struct rte_ring *r, unsigned int num,
> > +	enum rte_ring_queue_behavior behavior, uint32_t *old_head,
> > +	uint32_t *entries)
> > +{
> > +	uint32_t n;
> > +	union rte_ring_ht_pos np, op;
> > +
> > +	/* move cons.head atomically */
> > +	do {
> > +		/* Restore n as it may change every loop */
> > +		n = num;
> > +
> > +		/* wait for tail to be equal to head */
> > +		__rte_ring_hts_head_wait(&r->hts_cons, &op);
> > +
> > +		/* add rmb barrier to avoid load/load reorder in weak
> > +		 * memory model. It is noop on x86
> > +		 */
> > +		rte_smp_rmb();
> > +
> > +		/* The subtraction is done between two unsigned 32bits value
> > +		 * (the result is always modulo 32 bits even if we have
> > +		 * cons_head > prod_tail). So 'entries' is always between 0
> > +		 * and size(ring)-1.
> > +		 */
> > +		*entries = r->prod.tail - op.pos.head;
> > +
> > +		/* Set the actual entries for dequeue */
> > +		if (n > *entries)
> > +			n = (behavior == RTE_RING_QUEUE_FIXED) ? 0 :
> > *entries;
> > +
> > +		if (unlikely(n == 0))
> > +			break;
> > +
> > +		np.pos.tail = op.pos.tail;
> > +		np.pos.head = op.pos.head + n;
> > +
> > +	} while (rte_atomic64_cmpset(&r->hts_cons.ht.raw,
> > +			op.raw, np.raw) == 0);
> > +
> > +	*old_head = op.pos.head;
> > +	return n;
> > +}
> > +
> > +#endif /* _RTE_RING_HTS_GENERIC_H_ */
> > --
> > 2.17.1