netdev - Re: [PATCH net-next 3/4] net_sched: netem: remove qdisc_is

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <1470314151.13693.19.camel@edumazet-glaptop3.roam.corp.google.com>
Date:	Thu, 04 Aug 2016 14:35:51 +0200
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Eric Dumazet <edumazet@...gle.com>
Cc:	"David S . Miller" <davem@...emloft.net>,
	netdev <netdev@...r.kernel.org>
Subject: Re: [PATCH net-next 3/4] net_sched: netem: remove
 qdisc_is_throttled() use

On Fri, 2016-06-10 at 16:41 -0700, Eric Dumazet wrote:
> Looks like it is only there as some optimization attempt.
> 
> Since __QDISC_STATE_THROTTLED set/unset is way too expensive,
> and netem is the last user, just remove this check.
> 
> Signed-off-by: Eric Dumazet <edumazet@...gle.com>
> ---
>  net/sched/sch_netem.c | 3 ---
>  1 file changed, 3 deletions(-)
> 
> diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c
> index 9ca7947ab643..2dbe732ca135 100644
> --- a/net/sched/sch_netem.c
> +++ b/net/sched/sch_netem.c
> @@ -582,9 +582,6 @@ static struct sk_buff *netem_dequeue(struct Qdisc *sch)
>  	struct sk_buff *skb;
>  	struct rb_node *p;
>  
> -	if (qdisc_is_throttled(sch))
> -		return NULL;
> -
>  tfifo_dequeue:
>  	skb = __skb_dequeue(&sch->q);
>  	if (skb) {


Note that this commit (42117927cab5a13192ecc227bea19da5059ffc6c
"net_sched: netem: remove qdisc_is_throttled() use")
is a stable candidate, as it incidentally removes one long standing bug.

Namely, we were not calling qdisc_unthrottled() from netem enqueue(), so
we would be forced to wait a timer activation to be able to drain the
queue(s).

This matters with either a jitter, or a "tc qdisc change ... delay <new
delay>"