DPDK patches and discussions
 help / color / mirror / Atom feed
From: "Dumitrescu, Cristian" <cristian.dumitrescu@intel.com>
To: "alanrobertsonatt@gmail.com" <alanrobertsonatt@gmail.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>, Alan Robertson <ar771e@att.com>,
	"Alan Robertson" <alan.robertson@att.com>
Subject: Re: [dpdk-dev] [PATCH v2] Allow -ve frame_overhead values
Date: Thu, 3 May 2018 17:03:30 +0000	[thread overview]
Message-ID: <3EB4FA525960D640B5BDFFD6A3D891267BB6575D@IRSMSX108.ger.corp.intel.com> (raw)
In-Reply-To: <1516116717-24297-1-git-send-email-alanrobertsonatt@gmail.com>

> -----Original Message-----
> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of
> alanrobertsonatt@gmail.com
> Sent: Tuesday, January 16, 2018 3:32 PM
> To: Dumitrescu, Cristian <cristian.dumitrescu@intel.com>
> Cc: dev@dpdk.org; Alan Robertson <ar771e@att.com>; Alan Robertson
> <alan.robertson@att.com>
> Subject: [dpdk-dev] [PATCH v2] Allow -ve frame_overhead values
> 
> From: Alan Robertson <ar771e@att.com>
> 
> When forwarding traffic across a TDM network the ethernet header will
> be replaced with a ML-PPP one thereby reducing the size of the packet.
> 
> Signed-off-by: Alan Robertson <alan.robertson@att.com>
> ---
>  lib/librte_sched/rte_sched.c | 14 +++++++++++++-
>  lib/librte_sched/rte_sched.h |  5 +++--
>  2 files changed, 16 insertions(+), 3 deletions(-)
> 
> diff --git a/lib/librte_sched/rte_sched.c b/lib/librte_sched/rte_sched.c
> index ad2f7c6d5..c971fd0d4 100644
> --- a/lib/librte_sched/rte_sched.c
> +++ b/lib/librte_sched/rte_sched.c
> @@ -187,7 +187,7 @@ struct rte_sched_port {
>  	uint32_t n_pipes_per_subport;
>  	uint32_t rate;
>  	uint32_t mtu;
> -	uint32_t frame_overhead;
> +	int32_t frame_overhead;
>  	uint16_t qsize[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
>  	uint32_t n_pipe_profiles;
>  	uint32_t pipe_tc3_rate_max;
> @@ -1591,6 +1591,10 @@ grinder_credits_check(struct rte_sched_port
> *port, uint32_t pos)
>  	uint32_t pipe_tc_credits = pipe->tc_credits[tc_index];
>  	int enough_credits;
> 
> +#ifdef RTE_SCHED_DEBUG
> +	assert((int)(pkt->pkt_len + port->frame_overhead) > 0);
> +#endif /* RTE_SCHED_DEBUG */
> +
>  	/* Check queue credits */
>  	enough_credits = (pkt_len <= subport_tb_credits) &&
>  		(pkt_len <= subport_tc_credits) &&
> @@ -1629,6 +1633,10 @@ grinder_credits_check(struct rte_sched_port
> *port, uint32_t pos)
>  	uint32_t pipe_tc_ov_credits = pipe_tc_ov_mask1[tc_index];
>  	int enough_credits;
> 
> +#ifdef RTE_SCHED_DEBUG
> +	assert((int)(pkt->pkt_len + port->frame_overhead) > 0);
> +#endif /* RTE_SCHED_DEBUG */
> +
>  	/* Check pipe and subport credits */
>  	enough_credits = (pkt_len <= subport_tb_credits) &&
>  		(pkt_len <= subport_tc_credits) &&
> @@ -1663,6 +1671,10 @@ grinder_schedule(struct rte_sched_port *port,
> uint32_t pos)
>  	if (!grinder_credits_check(port, pos))
>  		return 0;
> 
> +#ifdef RTE_SCHED_DEBUG
> +	assert((int)(pkt->pkt_len + port->frame_overhead) > 0);
> +#endif /* RTE_SCHED_DEBUG */
> +
>  	/* Advance port time */
>  	port->time += pkt_len;
> 
> diff --git a/lib/librte_sched/rte_sched.h b/lib/librte_sched/rte_sched.h
> index 5d2a688dc..3e135c1e5 100644
> --- a/lib/librte_sched/rte_sched.h
> +++ b/lib/librte_sched/rte_sched.h
> @@ -190,8 +190,9 @@ struct rte_sched_port_params {
>  	uint32_t mtu;                    /**< Maximum Ethernet frame size
>  					  * (measured in bytes).
>  					  * Should not include the framing
> overhead. */
> -	uint32_t frame_overhead;         /**< Framing overhead per packet
> -					  * (measured in bytes) */
> +	int32_t frame_overhead;
> +	/**< Framing overhead per packet (measured in bytes).
> +	 * Can have negative value. */
>  	uint32_t n_subports_per_port;    /**< Number of subports */
>  	uint32_t n_pipes_per_subport;    /**< Number of pipes per subport
> */
>  	uint16_t qsize[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
> --
> 2.11.0
> 

Hi Alan,

This patch is straightforward now, but it has build failures, do you plan to send a v3 to fix them?

Essentially, you need to replace assert() calls from your code with RTE_ASSERT/RTE_VERIFY from rte_debug.h.

Also, I don't see the point in having the assert in grinder_credits_check(), as this is purely a check on packet length, not credits.  IMO we should have the assert in grinder_schedule() as the first instruction, before the call to grinder_credits_check(), as this is a sort of validating the inputs into the scheduling decision.

Thanks,
Cristian


      reply	other threads:[~2018-05-03 17:03 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-01-08 14:21 [dpdk-dev] [PATCH] " alanrobertsonatt
2018-01-16 15:31 ` [dpdk-dev] [PATCH v2] " alanrobertsonatt
2018-05-03 17:03   ` Dumitrescu, Cristian [this message]

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=3EB4FA525960D640B5BDFFD6A3D891267BB6575D@IRSMSX108.ger.corp.intel.com \
    --to=cristian.dumitrescu@intel.com \
    --cc=alan.robertson@att.com \
    --cc=alanrobertsonatt@gmail.com \
    --cc=ar771e@att.com \
    --cc=dev@dpdk.org \
    /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).