DPDK patches and discussions
 help / color / mirror / Atom feed
From: Bruce Richardson <bruce.richardson@intel.com>
To: Yongseok Koh <yskoh@mellanox.com>
Cc: anatoly.burakov@intel.com, dev@dpdk.org, stable@dpdk.org
Subject: Re: [dpdk-dev] [PATCH] examples/multi_process: fix buffer underrun
Date: Wed, 10 Apr 2019 10:14:57 +0100	[thread overview]
Message-ID: <20190410091457.GC700@bricha3-MOBL.ger.corp.intel.com> (raw)
Message-ID: <20190410091457.k4C14zcHdR-JINWZmYiM6eLS04BYoMi8oAzKi_8ffGY@z> (raw)
In-Reply-To: <20190409225937.9269-1-yskoh@mellanox.com>

On Tue, Apr 09, 2019 at 03:59:37PM -0700, Yongseok Koh wrote:
> For client_server_mp, the total number of buffers for the mbuf mempool
> should be correctly calculated. Otherwise, having more clients will stop
> traffic.
> 
> Fixes: af75078fece3 ("first public release")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
> ---
>  examples/multi_process/client_server_mp/mp_server/init.c | 10 ++++++----
>  1 file changed, 6 insertions(+), 4 deletions(-)
> 
> diff --git a/examples/multi_process/client_server_mp/mp_server/init.c b/examples/multi_process/client_server_mp/mp_server/init.c
> index 30c8e44bc0..0491ba8a1a 100644
> --- a/examples/multi_process/client_server_mp/mp_server/init.c
> +++ b/examples/multi_process/client_server_mp/mp_server/init.c
> @@ -37,8 +37,6 @@
>  #include "args.h"
>  #include "init.h"
>  
> -#define MBUFS_PER_CLIENT 1536
> -#define MBUFS_PER_PORT 1536
>  #define MBUF_CACHE_SIZE 512
>  
>  #define RTE_MP_RX_DESC_DEFAULT 1024
> @@ -63,8 +61,12 @@ struct port_info *ports;
>  static int
>  init_mbuf_pools(void)
>  {
> -	const unsigned num_mbufs = (num_clients * MBUFS_PER_CLIENT) \
> -			+ (ports->num_ports * MBUFS_PER_PORT);
> +	const unsigned int num_mbufs =
> +		RTE_MP_RX_DESC_DEFAULT * ports->num_ports + /* Sever */

typo: server

> +		num_clients *
> +		(CLIENT_QUEUE_RINGSIZE +
> +		 RTE_MP_TX_DESC_DEFAULT * ports->num_ports) + /* Clients */
> +		(num_clients + 1) * MBUF_CACHE_SIZE; /* Mempool cache */
>  

The calculation looks correct to me, it's just rather long with strange
wrapping. I'd suggest either splitting up the calculation to have a
separate variable for per-client mbuf counts, or else just use a longer
line for readability for the client calculation.

>  	/* don't pass single-producer/single-consumer flags to mbuf create as it
>  	 * seems faster to use a cache instead */
> -- 
> 2.11.0
> 

  parent reply	other threads:[~2019-04-10  9:15 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-04-09 22:59 Yongseok Koh
2019-04-09 22:59 ` Yongseok Koh
2019-04-10  9:14 ` Bruce Richardson [this message]
2019-04-10  9:14   ` Bruce Richardson
2019-04-10 19:41 ` Yongseok Koh
2019-04-10 19:41   ` Yongseok Koh
2019-04-11  7:18   ` [dpdk-dev] [PATCH v2] " Yongseok Koh
2019-04-11  7:18     ` Yongseok Koh
2019-04-30 18:01     ` [dpdk-dev] [dpdk-stable] " Yongseok Koh
2019-04-30 18:01       ` Yongseok Koh
2019-05-02 23:36   ` [dpdk-dev] [PATCH] " Thomas Monjalon
2019-05-02 23:36     ` Thomas Monjalon

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20190410091457.GC700@bricha3-MOBL.ger.corp.intel.com \
    --to=bruce.richardson@intel.com \
    --cc=anatoly.burakov@intel.com \
    --cc=dev@dpdk.org \
    --cc=stable@dpdk.org \
    --cc=yskoh@mellanox.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).