netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Petr Machata <petrm@mellanox.com>
To: Jakub Kicinski <kuba@kernel.org>
Cc: Ido Schimmel <idosch@idosch.org>,
	netdev@vger.kernel.org, davem@davemloft.net, jiri@mellanox.com,
	jhs@mojatatu.com, xiyou.wangcong@gmail.com, mlxsw@mellanox.com,
	Ido Schimmel <idosch@mellanox.com>
Subject: Re: [PATCH net-next 3/6] net: sched: RED: Introduce an ECN tail-dropping mode
Date: Tue, 10 Mar 2020 10:48:17 +0100	[thread overview]
Message-ID: <87tv2wy1zy.fsf@mellanox.com> (raw)
In-Reply-To: <20200309151255.29bb3a4b@kicinski-fedora-PC1C0HJN>


Jakub Kicinski <kuba@kernel.org> writes:

> On Mon,  9 Mar 2020 20:35:00 +0200 Ido Schimmel wrote:
>> diff --git a/net/sched/sch_red.c b/net/sched/sch_red.c
>> index f9839d68b811..d72db7643a37 100644
>> --- a/net/sched/sch_red.c
>> +++ b/net/sched/sch_red.c
>> @@ -44,7 +44,8 @@ struct red_sched_data {
>>  	struct Qdisc		*qdisc;
>>  };
>>
>> -#define RED_SUPPORTED_FLAGS (TC_RED_ECN | TC_RED_HARDDROP | TC_RED_ADAPTATIVE)
>> +#define RED_SUPPORTED_FLAGS (TC_RED_ECN | TC_RED_HARDDROP | \
>> +			     TC_RED_ADAPTATIVE | TC_RED_TAILDROP)
>>
>>  static inline int red_use_ecn(struct red_sched_data *q)
>>  {
>> @@ -56,6 +57,11 @@ static inline int red_use_harddrop(struct red_sched_data *q)
>>  	return q->flags & TC_RED_HARDDROP;
>>  }
>>
>> +static inline int red_use_taildrop(struct red_sched_data *q)
>
> Please don't do static inlines in C code, even if the neighboring code
> does.

Ok.

>> +{
>> +	return q->flags & TC_RED_TAILDROP;
>> +}
>> +
>>  static int red_enqueue(struct sk_buff *skb, struct Qdisc *sch,
>>  		       struct sk_buff **to_free)
>>  {
>> @@ -76,23 +82,36 @@ static int red_enqueue(struct sk_buff *skb, struct Qdisc *sch,
>>
>>  	case RED_PROB_MARK:
>>  		qdisc_qstats_overlimit(sch);
>> -		if (!red_use_ecn(q) || !INET_ECN_set_ce(skb)) {
>> +		if (!red_use_ecn(q)) {
>>  			q->stats.prob_drop++;
>>  			goto congestion_drop;
>>  		}
>>
>> -		q->stats.prob_mark++;
>> +		if (INET_ECN_set_ce(skb)) {
>> +			q->stats.prob_mark++;
>> +		} else if (red_use_taildrop(q)) {
>
> This condition is inverted, no?
>
> If user requested taildrop the packet should be queued.

Argh, yes! I was meddling with different ways of coding the if-else
after Ido reviewed it and screwed it up. Thanks for catching it.

My tests have not caught this, because they all hit the hard mark case
:(

>> +			q->stats.prob_drop++;
>> +			goto congestion_drop;
>> +		}
>> +
>> +		/* Non-ECT packet in ECN taildrop mode: queue it. */
>>  		break;
>>
>>  	case RED_HARD_MARK:
>>  		qdisc_qstats_overlimit(sch);
>> -		if (red_use_harddrop(q) || !red_use_ecn(q) ||
>> -		    !INET_ECN_set_ce(skb)) {
>> +		if (red_use_harddrop(q) || !red_use_ecn(q)) {
>> +			q->stats.forced_drop++;
>> +			goto congestion_drop;
>> +		}
>> +
>> +		if (INET_ECN_set_ce(skb)) {
>> +			q->stats.forced_mark++;
>> +		} else if (!red_use_taildrop(q)) {
>
> This one looks correct.
>
>>  			q->stats.forced_drop++;
>>  			goto congestion_drop;
>>  		}
>>
>> -		q->stats.forced_mark++;
>> +		/* Non-ECT packet in ECN taildrop mode: queue it. */
>>  		break;
>>  	}
>>

  reply	other threads:[~2020-03-10  9:48 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-03-09 18:34 [PATCH net-next 0/6] RED: Introduce an ECN tail-dropping mode Ido Schimmel
2020-03-09 18:34 ` [PATCH net-next 1/6] selftests: qdiscs: Add TDC test for RED Ido Schimmel
2020-03-10 15:40   ` Roman Mashak
2020-03-10 16:56     ` Petr Machata
2020-03-10 17:28       ` Roman Mashak
2020-03-09 18:34 ` [PATCH net-next 2/6] net: sched: Add centralized RED flag checking Ido Schimmel
2020-03-09 22:18   ` Jakub Kicinski
2020-03-10  9:48     ` Petr Machata
2020-03-10 19:53       ` Jakub Kicinski
2020-03-10 22:23         ` Petr Machata
2020-03-10 23:00           ` Jakub Kicinski
2020-03-10 23:53             ` Petr Machata
2020-03-09 18:35 ` [PATCH net-next 3/6] net: sched: RED: Introduce an ECN tail-dropping mode Ido Schimmel
2020-03-09 22:12   ` Jakub Kicinski
2020-03-10  9:48     ` Petr Machata [this message]
2020-03-09 18:35 ` [PATCH net-next 4/6] mlxsw: spectrum_qdisc: Offload RED " Ido Schimmel
2020-03-09 18:35 ` [PATCH net-next 5/6] selftests: qdiscs: RED: Add taildrop tests Ido Schimmel
2020-03-09 18:35 ` [PATCH net-next 6/6] selftests: mlxsw: RED: Test RED ECN taildrop offload Ido Schimmel

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=87tv2wy1zy.fsf@mellanox.com \
    --to=petrm@mellanox.com \
    --cc=davem@davemloft.net \
    --cc=idosch@idosch.org \
    --cc=idosch@mellanox.com \
    --cc=jhs@mojatatu.com \
    --cc=jiri@mellanox.com \
    --cc=kuba@kernel.org \
    --cc=mlxsw@mellanox.com \
    --cc=netdev@vger.kernel.org \
    --cc=xiyou.wangcong@gmail.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).