From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?ISO-8859-2?Q?Micha=B3_Miros=B3aw?= Subject: Re: [PATCH] pktgen: Clone skb to avoid corruption of skbs in ndo_start_xmit methods Date: Wed, 20 Jul 2011 18:08:39 +0200 Message-ID: References: <1311105738.3113.11.camel@edumazet-laptop> <20110719202922.GA2352@minipsycho> <1311108107.3113.22.camel@edumazet-laptop> <20110720001904.GA1992@neilslaptop.think-freely.org> <1311122593.3113.46.camel@edumazet-laptop> <4E2626E1.6030005@candelatech.com> <20110720020737.GB2692@neilslaptop.think-freely.org> <1311135855.3113.66.camel@edumazet-laptop> <20110720151827.GD12349@hmsreliant.think-freely.org> <20110720154057.GF12349@hmsreliant.think-freely.org> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Eric Dumazet , Ben Greear , Jiri Pirko , netdev@vger.kernel.org, Alexey Dobriyan , "David S. Miller" , robert.olsson@its.uu.se To: Neil Horman Return-path: Received: from mail-qw0-f46.google.com ([209.85.216.46]:37004 "EHLO mail-qw0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751302Ab1GTQI7 convert rfc822-to-8bit (ORCPT ); Wed, 20 Jul 2011 12:08:59 -0400 Received: by qwk3 with SMTP id 3so197489qwk.19 for ; Wed, 20 Jul 2011 09:08:59 -0700 (PDT) In-Reply-To: <20110720154057.GF12349@hmsreliant.think-freely.org> Sender: netdev-owner@vger.kernel.org List-ID: W dniu 20 lipca 2011 17:40 u=C5=BCytkownik Neil Horman napisa=C5=82: > On Wed, Jul 20, 2011 at 05:35:47PM +0200, Micha=C5=82 Miros=C5=82aw w= rote: >> 2011/7/20 Neil Horman : >> > On Wed, Jul 20, 2011 at 06:24:15AM +0200, Eric Dumazet wrote: >> >> Le mardi 19 juillet 2011 =C3=A0 22:07 -0400, Neil Horman a =C3=A9= crit : >> >> > > >> >> > I think this is a good idea. =C2=A0It lets pktgen dynamically m= ake the clone/share >> >> > decision dynamically and only impacts performance for those sys= tems. >> >> > >> >> >> >> Just let pktgen refuse to use clone_skb command for these devices= =2E >> >> >> > copy that, This is by no means final, but what do you think of thi= s? =C2=A0If its >> > agreeable to you, Ben, et al. I can add this to my local tree and = start auditing >> > all the drivers that may need to have the flag set. >> > >> > Regards >> > Neil >> > >> > >> > diff --git a/include/linux/if.h b/include/linux/if.h >> > index 3bc63e6..ae904fe 100644 >> > --- a/include/linux/if.h >> > +++ b/include/linux/if.h >> > @@ -76,6 +76,7 @@ >> > =C2=A0#define IFF_BRIDGE_PORT =C2=A0 =C2=A0 =C2=A0 =C2=A00x4000 =C2= =A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0/* device used as bridge port */ >> > =C2=A0#define IFF_OVS_DATAPATH =C2=A0 =C2=A0 =C2=A0 0x8000 =C2=A0/= * device used as Open vSwitch >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2= =A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0= * datapath port */ >> > +#define IFF_CANT_SHARE_SKB =C2=A0 =C2=A0 0x10000 /* Device can't = share skbs in tx path */ >> > >> > =C2=A0#define IF_GET_IFACE =C2=A0 0x0001 =C2=A0 =C2=A0 =C2=A0 =C2=A0= =C2=A0/* for querying only */ >> > =C2=A0#define IF_GET_PROTO =C2=A0 0x0002 >> > diff --git a/net/core/pktgen.c b/net/core/pktgen.c >> > index f76079c..bf6d88d 100644 >> > --- a/net/core/pktgen.c >> > +++ b/net/core/pktgen.c >> > @@ -1071,6 +1071,9 @@ static ssize_t pktgen_if_write(struct file *= file, >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0if (len < 0= ) >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2= =A0 =C2=A0 =C2=A0return len; >> > >> > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 if (pkt_dev->pr= iv_flags & IFF_CANT_SHARE_SKB) >> > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2= =A0 =C2=A0 return -EOPNOTSUPP; >> > + >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0i +=3D len; >> > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0pkt_dev->cl= one_skb =3D value; >> > >> >> I would prefer that the flag be inclusive, i.e. it should mark drive= rs >> which can use shared skbs. And it might be better to clone the skb i= n >> case the flag is disabled to keep functionality working. > Ok, I can agree with that. =C2=A0But you're ok with the general appro= ach? I assumed you wanted to use netdev->features and I noticed priv_flags just now. If the flag is supposed to be permanent then its fine by me. Actually this makes me wonder if NETIF_F_LLTX and similar should get moved to netdev->priv_flags. Best Regards, Micha=C5=82 Miros=C5=82aw