netdev - Re: [RFC PATCH] Regression in linux 2.6.32 virtio

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

Message-ID: <OF6BB64FFB.075C585D-ON6525768F.0047CB9A-6525768F.00487BE4@in.ibm.com>
Date:	Thu, 17 Dec 2009 18:52:55 +0530
From:	Krishna Kumar2 <krkumar2@...ibm.com>
To:	Herbert Xu <herbert@...dor.apana.org.au>
Cc:	"David S. Miller" <davem@...emloft.net>,
	Jarek Poplawski <jarkao2@...il.com>, mst@...hat.com,
	netdev@...r.kernel.org, Rusty Russell <rusty@...tcorp.com.au>,
	Sridhar Samudrala <sri@...ibm.com>
Subject: Re: [RFC PATCH] Regression in linux 2.6.32 virtio_net seen with	vhost-net

Herbert Xu <herbert@...dor.apana.org.au> wrote on 12/17/2009 06:26:54 PM:

>
> > On Thu, Dec 17, 2009 at 05:57:15PM +0530, Krishna Kumar2 wrote:
> > > the first resched. I also think the resched in requeue is probably
> > > not required, since the device will call netif_tx_wake_queue anyway.
>
> Actually, the requeue is there for the case where we hit lock
> contention on the device TX lock so it is needed for that at
> least.
>
> However, for the case where the device queue is stopped you're
> right we don't need to reschedule.

Yes, my patch had a handle for both cases by adding another arg
to dev_requeue_skb, something like:

static inline int dev_requeue_skb(struct sk_buff *skb, struct Qdisc *q,
                                  int reason)
{
      q->gso_skb = skb;
      q->qstats.requeues++;
      q->q.qlen++;    /* it's still part of the queue */

      if (reason != NETDEV_TX_BUSY)
            __netif_schedule(q);

      return 0;
}

and pass the reason in the two places that call requeue. I didn't
submit it since it felt a small optimization - avoiding one resched
per stopped txq (not for every skb), plus I was not sure if every
driver would properly call netif_tx_wake_queue.

thanks,

- KK

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