[resend,net-next] sch_tbf: add TBF_BURST/TBF_PBURST attribute

From: Yang Yingliang <yangyingliang@huawei.com>

When we set burst to 1514 with low rate in userspace,
the kernel get a value of burst that less than 1514,
which doesn't work.

Because it may make some loss when transform burst
to buffer in userspace. This makes burst lose some
bytes, when the kernel transform the buffer back to
burst.

This patch adds two new attributes to support sending
burst/mtu to kernel directly to avoid the loss.

Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
---
 include/uapi/linux/pkt_sched.h |  2 ++
 net/sched/sch_tbf.c            | 25 +++++++++++++++++++++----
 2 files changed, 23 insertions(+), 4 deletions(-)

-- 1.8.0

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

From: Yang Yingliang <yangyingliang@huawei.com>
Date: Tue, 17 Dec 2013 11:34:39 +0800

> From: Yang Yingliang <yangyingliang@huawei.com>
> 
> When we set burst to 1514 with low rate in userspace,
> the kernel get a value of burst that less than 1514,
> which doesn't work.
> 
> Because it may make some loss when transform burst
> to buffer in userspace. This makes burst lose some
> bytes, when the kernel transform the buffer back to
> burst.
> 
> This patch adds two new attributes to support sending
> burst/mtu to kernel directly to avoid the loss.
> 
> Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>

I really want to see some review of this change, thanks.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

On Tue, 2013-12-17 at 11:34 +0800, Yang Yingliang wrote:
> From: Yang Yingliang <yangyingliang@huawei.com>
> 
> When we set burst to 1514 with low rate in userspace,
> the kernel get a value of burst that less than 1514,
> which doesn't work.
> 
> Because it may make some loss when transform burst
> to buffer in userspace. This makes burst lose some
> bytes, when the kernel transform the buffer back to
> burst.
> 
> This patch adds two new attributes to support sending
> burst/mtu to kernel directly to avoid the loss.
> 
> Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
> ---
>  include/uapi/linux/pkt_sched.h |  2 ++
>  net/sched/sch_tbf.c            | 25 +++++++++++++++++++++----
>  2 files changed, 23 insertions(+), 4 deletions(-)
> 
> diff --git a/include/uapi/linux/pkt_sched.h b/include/uapi/linux/pkt_sched.h
> index a806687..fd8e17e 100644
> --- a/include/uapi/linux/pkt_sched.h
> +++ b/include/uapi/linux/pkt_sched.h
> @@ -173,6 +173,8 @@ enum {
>  	TCA_TBF_PTAB,
>  	TCA_TBF_RATE64,
>  	TCA_TBF_PRATE64,
> +	TCA_TBF_BURST,
> +	TCA_TBF_PBURST,
>  	__TCA_TBF_MAX,
>  };
>  
> diff --git a/net/sched/sch_tbf.c b/net/sched/sch_tbf.c
> index 887e672..ee84409 100644
> --- a/net/sched/sch_tbf.c
> +++ b/net/sched/sch_tbf.c
> @@ -307,6 +307,8 @@ static const struct nla_policy tbf_policy[TCA_TBF_MAX + 1] = {
>  	[TCA_TBF_PTAB]	= { .type = NLA_BINARY, .len = TC_RTAB_SIZE },
>  	[TCA_TBF_RATE64]	= { .type = NLA_U64 },
>  	[TCA_TBF_PRATE64]	= { .type = NLA_U64 },
> +	[TCA_TBF_BURST] = { .type = NLA_U32 },
> +	[TCA_TBF_PBURST] = { .type = NLA_U32 },
>  };
>  
>  static int tbf_change(struct Qdisc *sch, struct nlattr *opt)
> @@ -358,7 +360,11 @@ static int tbf_change(struct Qdisc *sch, struct nlattr *opt)
>  		rate64 = nla_get_u64(tb[TCA_TBF_RATE64]);
>  	psched_ratecfg_precompute(&rate, &qopt->rate, rate64);
>  
> -	max_size = min_t(u64, psched_ns_t2l(&rate, buffer), ~0U);
> +	if (tb[TCA_TBF_BURST]) {
> +		max_size = nla_get_u32(tb[TCA_TBF_BURST]);
> +		buffer = psched_l2t_ns(&rate, max_size);
> +	} else
> +		max_size = min_t(u64, psched_ns_t2l(&rate, buffer), ~0U);
>  
         } else {
                ...
         }

>  	if (qopt->peakrate.rate) {
>  		if (tb[TCA_TBF_PRATE64])
> @@ -371,7 +377,12 @@ static int tbf_change(struct Qdisc *sch, struct nlattr *opt)
>  			goto done;
>  		}
>  
> -		max_size = min_t(u64, max_size, psched_ns_t2l(&peak, mtu));
> +		if (tb[TCA_TBF_PBURST]) {
> +			u32 pburst = nla_get_u32(tb[TCA_TBF_PBURST]);
> +			max_size = min_t(u32, max_size, pburst);
> +			mtu = psched_l2t_ns(&peak, pburst);
> +		} else
> +			max_size = min_t(u64, max_size, psched_ns_t2l(&peak, mtu));
                 } else {
                 }

>  	}
>  
>  	if (max_size < psched_mtu(qdisc_dev(sch)))
> @@ -391,9 +402,15 @@ static int tbf_change(struct Qdisc *sch, struct nlattr *opt)
>  		q->qdisc = child;
>  	}
>  	q->limit = qopt->limit;
> -	q->mtu = PSCHED_TICKS2NS(qopt->mtu);
> +	if (tb[TCA_TBF_PBURST])
> +		q->mtu = mtu;
> +	else
> +		q->mtu = PSCHED_TICKS2NS(qopt->mtu);
>  	q->max_size = max_size;
> -	q->buffer = PSCHED_TICKS2NS(qopt->buffer);
> +	if (tb[TCA_TBF_BURST])
> +		q->buffer = buffer;
> +	else
> +		q->buffer = PSCHED_TICKS2NS(qopt->buffer);
>  	q->tokens = q->buffer;
>  	q->ptokens = q->mtu;
>  

Otherwise, looks good.


--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html