[dpdk-dev] [PATCH] ring: fix sc dequeue performance issue

[Thomas Monjalon]

2016-07-24 22:57, Ananyev, Konstantin:
> From: Jerin Jacob [mailto:jerin.jacob at caviumnetworks.com]
> > Use of rte_smb_wmb() instead of rte_smb_rmb() in sc dequeue function 
> > creates the additional overhead of waiting for all the STOREs to be
> > completed to local buffer from ring buffer memory. The sc dequeue function 
> > demands only LOAD-STORE barrier where LOADs from ring
> > buffer memory needs to be completed before tail pointer update. Changing to 
> > rte_smb_rmb() to enable the required LOAD-STORE barrier.
> > 
> > Fixes: ecc7d10e448e ("ring: guarantee dequeue ordering before tail update")
> > 
> > Signed-off-by: Jerin Jacob 
> 
> Acked-by: Konstantin Ananyev 

Applied, thanks for the quick reaction

[dpdk-dev] [PATCH] ring: fix sc dequeue performance issue

[Ananyev, Konstantin]

> -Original Message-
> From: Jerin Jacob [mailto:jerin.jacob at caviumnetworks.com]
> Sent: Sunday, July 24, 2016 6:08 PM
> To: dev at dpdk.org
> Cc: thomas.monjalon at 6wind.com; Ananyev, Konstantin  intel.com>; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH] ring: fix sc dequeue performance issue
> 
> Use of rte_smb_wmb() instead of rte_smb_rmb() in sc dequeue function creates 
> the additional overhead of waiting for all the STOREs to be
> completed to local buffer from ring buffer memory. The sc dequeue function 
> demands only LOAD-STORE barrier where LOADs from ring
> buffer memory needs to be completed before tail pointer update. Changing to 
> rte_smb_rmb() to enable the required LOAD-STORE barrier.
> 
> Fixes: ecc7d10e448e ("ring: guarantee dequeue ordering before tail update")
> 
> Signed-off-by: Jerin Jacob 
> ---
>  lib/librte_ring/rte_ring.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/lib/librte_ring/rte_ring.h b/lib/librte_ring/rte_ring.h index 
> f928324..0e22e69 100644
> --- a/lib/librte_ring/rte_ring.h
> +++ b/lib/librte_ring/rte_ring.h
> @@ -756,7 +756,7 @@ __rte_ring_sc_do_dequeue(struct rte_ring *r, void 
> **obj_table,
> 
>   /* copy in table */
>   DEQUEUE_PTRS();
> - rte_smp_wmb();
> + rte_smp_rmb();
> 
>   __RING_STAT_ADD(r, deq_success, n);
>   r->cons.tail = cons_next;
> --

Acked-by: Konstantin Ananyev 

> 2.5.5

[dpdk-dev] [PATCH] ring: fix sc dequeue performance issue

[Jerin Jacob]

Use of rte_smb_wmb() instead of rte_smb_rmb() in sc dequeue
function creates the additional overhead of waiting for
all the STOREs to be completed to local buffer from ring buffer
memory. The sc dequeue function demands only LOAD-STORE barrier
where LOADs from ring buffer memory needs to be
completed before tail pointer update. Changing to rte_smb_rmb()
to enable the required LOAD-STORE barrier.

Fixes: ecc7d10e448e ("ring: guarantee dequeue ordering before tail update")

Signed-off-by: Jerin Jacob 
---
 lib/librte_ring/rte_ring.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/librte_ring/rte_ring.h b/lib/librte_ring/rte_ring.h
index f928324..0e22e69 100644
--- a/lib/librte_ring/rte_ring.h
+++ b/lib/librte_ring/rte_ring.h
@@ -756,7 +756,7 @@ __rte_ring_sc_do_dequeue(struct rte_ring *r, void 
**obj_table,

/* copy in table */
DEQUEUE_PTRS();
-   rte_smp_wmb();
+   rte_smp_rmb();

__RING_STAT_ADD(r, deq_success, n);
r->cons.tail = cons_next;
-- 
2.5.5