public inbox for netdev@vger.kernel.org
 help / color / mirror / Atom feed
From: Eric Dumazet <eric.dumazet-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
To: Johannes Berg <johannes-cdvu00un1VgdHxzADdlk8Q@public.gmane.org>
Cc: Marc MERLIN <marc-xnduUnryOU1AfugRpC6u6w@public.gmane.org>,
	David Miller <davem-fT/PcQaiUtIeIZ0/mPfg9Q@public.gmane.org>,
	Larry.Finger-tQ5ms3gMjBLk1uMJSBkQmQ@public.gmane.org,
	bhutchings-s/n/eUQHGBpZroRs9YW3xA@public.gmane.org,
	linux-wireless-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	netdev-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
Subject: Re: 3.7.8/amd64 full interrupt hangs due to iwlwifi under big nfs copies out
Date: Wed, 20 Feb 2013 08:55:31 -0800	[thread overview]
Message-ID: <1361379331.19353.189.camel@edumazet-glaptop> (raw)
In-Reply-To: <1361377243.8629.34.camel-8Nb76shvtaUJvtFkdXX2HixXY32XiHfO@public.gmane.org>

On Wed, 2013-02-20 at 17:20 +0100, Johannes Berg wrote:
> On Wed, 2013-02-20 at 07:11 -0800, Eric Dumazet wrote:
> > On Wed, 2013-02-20 at 10:15 +0100, Johannes Berg wrote:
> > 
> > > OTOH, this affects the protocol, and when you really can't allocate any
> > > order-1 pages you pointed out yourself that many other things also won't
> > > work, so I'm not really sure it makes a big difference if we change the
> > > driver?
> > 
> > It will make a huge difference, even on non pressure mode, as TCP
> > receive window will grow twice faster.
> 
> Hmm, why does that depend on the allocation size?

I guess you missed all the patches about skb->truesize on netdev

> 
> > Unless wifi speed reaches 10Gbps, following patch should do the trick
> > 
> > 
> > diff --git a/drivers/net/wireless/iwlwifi/dvm/rx.c b/drivers/net/wireless/iwlwifi/dvm/rx.c
> > index a4eed20..77a3ee3 100644
> > --- a/drivers/net/wireless/iwlwifi/dvm/rx.c
> > +++ b/drivers/net/wireless/iwlwifi/dvm/rx.c
> > @@ -750,7 +750,12 @@ static void iwlagn_pass_packet_to_mac80211(struct iwl_priv *priv,
> >  	/* Dont use dev_alloc_skb(), we'll have enough headroom once
> >  	 * ieee80211_hdr pulled.
> >  	 */
> > -	skb = alloc_skb(128, GFP_ATOMIC);
> > +	fraglen = 128;
> > +	/* if we use order-1 pages, copy to get better TCP performance */
> > +	if (rxb->truesize > PAGE_SIZE)
> > +		fraglen = max_t(unsigned, fraglen, len);
> > +
> > +	skb = alloc_skb(fraglen, GFP_ATOMIC);
> 
> Hmm, I don't quite understand -- that's not doing any copy?
> 
> FWIW if you do the copy you should not "steal" the pages, then they'd be
> recycled in the RX ring right away.

Code should just works, please read the following lines in the same
function....

        /* If frame is small enough to fit in skb->head, pull it completely.
         * If not, only pull ieee80211_hdr so that splice() or TCP coalesce
         * are more efficient.
         */
        hdrlen = (len <= skb_tailroom(skb)) ? len : sizeof(*hdr);

        memcpy(skb_put(skb, hdrlen), hdr, hdrlen);
        fraglen = len - hdrlen;

        if (fraglen) {
                int offset = (void *)hdr + hdrlen -
                             rxb_addr(rxb) + rxb_offset(rxb);

                skb_add_rx_frag(skb, 0, rxb_steal_page(rxb), offset,
                                fraglen, rxb->truesize);
        }



--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

  parent reply	other threads:[~2013-02-20 16:55 UTC|newest]

Thread overview: 55+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-03-29 16:38 3.2.8/amd64 full interrupt hangs and deadlocks under big network copies (page allocation failure) Marc MERLIN
     [not found] ` <20120311183244.GA14001@merlins.org>
     [not found]   ` <20120329053111.GD24933@merlins.org>
     [not found]     ` <20120329163800.GH24933-xnduUnryOU1AfugRpC6u6w@public.gmane.org>
     [not found]       ` <20120329053111.GD24933-xnduUnryOU1AfugRpC6u6w@public.gmane.org>
2012-03-29 16:41         ` Marc MERLIN
2012-03-29 18:09 ` Ben Hutchings
     [not found]   ` <1333044575.2656.1.camel-/LGg1Z1CJKReKY3V0RtoKmatzQS1i7+A3tAM5lWOD0I@public.gmane.org>
2012-03-29 21:19     ` Marc MERLIN
2012-04-09 17:20   ` Marc MERLIN
2012-04-09 18:12     ` David Miller
     [not found]       ` <20120409.141241.1216091936509309354.davem-fT/PcQaiUtIeIZ0/mPfg9Q@public.gmane.org>
2012-04-09 18:36         ` Marc MERLIN
2012-04-09 18:37           ` David Miller
2012-04-09 18:58             ` Larry Finger
2012-04-09 19:11               ` Eric Dumazet
2012-04-09 19:34                 ` David Miller
     [not found]                   ` <20120409.153452.1284163346306246866.davem-fT/PcQaiUtIeIZ0/mPfg9Q@public.gmane.org>
2012-04-09 19:46                     ` Marc MERLIN
2012-04-10  3:56                   ` Eric Dumazet
2012-04-10  5:11                     ` Marc MERLIN
2012-04-10  6:11                       ` Eric Dumazet
2012-04-11  5:27                         ` Marc MERLIN
     [not found]                           ` <20120411052733.GA17352-xnduUnryOU1AfugRpC6u6w@public.gmane.org>
2012-04-11  5:43                             ` Eric Dumazet
2012-04-11  6:30                               ` [PATCH] tcp: avoid order-1 allocations on wifi and tx path Eric Dumazet
2012-04-11  7:38                                 ` Eric Dumazet
2012-04-11 14:12                                   ` David Miller
2012-04-11 14:11                                 ` David Miller
2012-04-11  6:08                           ` [PATCH] net: allow pskb_expand_head() to get maximum tailroom Eric Dumazet
2012-04-11 14:11                             ` David Miller
2012-07-15 21:59                           ` 3.4.4/amd64 full interrupt hangs under big nfs copies Marc MERLIN
2012-07-16  6:18                             ` Eric Dumazet
2012-07-16 15:18                               ` Marc MERLIN
2012-07-16 16:21                                 ` Eric Dumazet
2012-07-16 17:17                                   ` Marc MERLIN
2013-02-19  4:05                                   ` 3.7.8/amd64 full interrupt hangs due to iwlwifi under big nfs copies out Marc MERLIN
2013-02-19  5:17                                     ` Eric Dumazet
2013-02-19  5:26                                       ` Marc MERLIN
2013-02-19 10:03                                       ` Johannes Berg
2013-02-19 16:18                                         ` Marc MERLIN
2013-02-19 16:36                                           ` Eric Dumazet
2013-02-19 16:21                                         ` Eric Dumazet
2013-02-20  9:12                                           ` Johannes Berg
2013-02-20  9:15                                             ` Johannes Berg
2013-02-20 15:11                                               ` Eric Dumazet
2013-02-20 16:20                                                 ` Johannes Berg
     [not found]                                                   ` <1361377243.8629.34.camel-8Nb76shvtaUJvtFkdXX2HixXY32XiHfO@public.gmane.org>
2013-02-20 16:55                                                     ` Eric Dumazet [this message]
2013-02-20 16:59                                                       ` Johannes Berg
2013-02-20 17:39                                                         ` Eric Dumazet
2013-02-20 17:01                                                       ` Johannes Berg
2013-02-20 17:24                                                         ` Eric Dumazet
2013-02-20 18:16                                                           ` Johannes Berg
2013-02-20 19:17                                                             ` Eric Dumazet
2013-02-20 19:58                                                               ` Johannes Berg
2013-02-20 20:14                                                                 ` Eric Dumazet
2013-02-20 20:27                                                                   ` Johannes Berg
2013-02-20 20:09                                                 ` Johannes Berg
2013-02-23  6:14                                               ` Marc MERLIN
2013-02-26 20:54                                                 ` Johannes Berg
     [not found]                                                   ` <1361912099.8440.21.camel-8Nb76shvtaUJvtFkdXX2HixXY32XiHfO@public.gmane.org>
2013-06-18 16:52                                                     ` Eric Dumazet
2013-06-18 17:04                                                       ` Johannes Berg
2013-06-19 13:09                                                         ` Stanislaw Gruszka

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=1361379331.19353.189.camel@edumazet-glaptop \
    --to=eric.dumazet-re5jqeeqqe8avxtiumwx3w@public.gmane.org \
    --cc=Larry.Finger-tQ5ms3gMjBLk1uMJSBkQmQ@public.gmane.org \
    --cc=bhutchings-s/n/eUQHGBpZroRs9YW3xA@public.gmane.org \
    --cc=davem-fT/PcQaiUtIeIZ0/mPfg9Q@public.gmane.org \
    --cc=johannes-cdvu00un1VgdHxzADdlk8Q@public.gmane.org \
    --cc=linux-wireless-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
    --cc=marc-xnduUnryOU1AfugRpC6u6w@public.gmane.org \
    --cc=netdev-u79uwXL29TY76Z2rM5mHXA@public.gmane.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