All of lore.kernel.org
 help / color / mirror / Atom feed
From: John Crispin <john@phrozen.org>
To: Andrew Lunn <andrew@lunn.ch>,
	Vivien Didelot <vivien.didelot@savoirfairelinux.com>,
	Florian Fainelli <f.fainelli@gmail.com>,
	"David S . Miller" <davem@davemloft.net>,
	Sean Wang <sean.wang@mediatek.com>
Cc: netdev@vger.kernel.org, linux-mediatek@lists.infradead.org,
	linux-kernel@vger.kernel.org, Muciri Gatimu <muciri@openmesh.com>,
	Shashidhar Lakkavalli <shashidhar.lakkavalli@openmesh.com>,
	John Crispin <john@phrozen.org>
Subject: [PATCH 4/4] net-next: dsa: fix flow dissection
Date: Wed,  9 Aug 2017 14:41:19 +0200	[thread overview]
Message-ID: <20170809124119.24320-5-john@phrozen.org> (raw)
In-Reply-To: <20170809124119.24320-1-john@phrozen.org>

RPS and probably other kernel features are currently broken on some if not
all DSA devices. The root cause of this is that skb_hash will call the
flow_dissector. At this point the skb still contains the magic switch
header and the skb->protocol field is not set up to the correct 802.3
value yet. By the time the tag specific code is called, removing the header
and properly setting the protocol an invalid hash is already set. In the
case of the mt7530 this will result in all flows always having the same
hash.

Signed-off-by: Muciri Gatimu <muciri@openmesh.com>
Signed-off-by: Shashidhar Lakkavalli <shashidhar.lakkavalli@openmesh.com>
Signed-off-by: John Crispin <john@phrozen.org>
---
 net/core/flow_dissector.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c
index 0cc672aba1f0..5b5be9577257 100644
--- a/net/core/flow_dissector.c
+++ b/net/core/flow_dissector.c
@@ -4,6 +4,7 @@
 #include <linux/ip.h>
 #include <linux/ipv6.h>
 #include <linux/if_vlan.h>
+#include <net/dsa.h>
 #include <net/ip.h>
 #include <net/ipv6.h>
 #include <net/gre.h>
@@ -440,6 +441,17 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 			 skb->vlan_proto : skb->protocol;
 		nhoff = skb_network_offset(skb);
 		hlen = skb_headlen(skb);
+		if (unlikely(netdev_uses_dsa(skb->dev))) {
+			const struct dsa_device_ops *ops;
+			int offset;
+
+			ops = skb->dev->dsa_ptr->tag_ops;
+			if (ops->flow_dissect &&
+			    !ops->flow_dissect(skb, &proto, &offset)) {
+				hlen -= offset;
+				nhoff += offset;
+			}
+		}
 	}
 
 	/* It is ensured by skb_flow_dissector_init() that control key will
-- 
2.11.0

WARNING: multiple messages have this Message-ID (diff)
From: John Crispin <john-Pj+rj9U5foFAfugRpC6u6w@public.gmane.org>
To: Andrew Lunn <andrew-g2DYL2Zd6BY@public.gmane.org>,
	Vivien Didelot
	<vivien.didelot-4ysUXcep3aM1wj+D4I0NRVaTQe2KTcn/@public.gmane.org>,
	Florian Fainelli
	<f.fainelli-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>,
	"David S . Miller"
	<davem-fT/PcQaiUtIeIZ0/mPfg9Q@public.gmane.org>,
	Sean Wang <sean.wang-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
Cc: Muciri Gatimu <muciri-lv6y7wLVQPlWk0Htik3J/w@public.gmane.org>,
	netdev-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	Shashidhar Lakkavalli
	<shashidhar.lakkavalli-lv6y7wLVQPlWk0Htik3J/w@public.gmane.org>,
	linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	linux-mediatek-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org,
	John Crispin <john-Pj+rj9U5foFAfugRpC6u6w@public.gmane.org>
Subject: [PATCH 4/4] net-next: dsa: fix flow dissection
Date: Wed,  9 Aug 2017 14:41:19 +0200	[thread overview]
Message-ID: <20170809124119.24320-5-john@phrozen.org> (raw)
In-Reply-To: <20170809124119.24320-1-john-Pj+rj9U5foFAfugRpC6u6w@public.gmane.org>

RPS and probably other kernel features are currently broken on some if not
all DSA devices. The root cause of this is that skb_hash will call the
flow_dissector. At this point the skb still contains the magic switch
header and the skb->protocol field is not set up to the correct 802.3
value yet. By the time the tag specific code is called, removing the header
and properly setting the protocol an invalid hash is already set. In the
case of the mt7530 this will result in all flows always having the same
hash.

Signed-off-by: Muciri Gatimu <muciri-lv6y7wLVQPlWk0Htik3J/w@public.gmane.org>
Signed-off-by: Shashidhar Lakkavalli <shashidhar.lakkavalli-lv6y7wLVQPlWk0Htik3J/w@public.gmane.org>
Signed-off-by: John Crispin <john-Pj+rj9U5foFAfugRpC6u6w@public.gmane.org>
---
 net/core/flow_dissector.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c
index 0cc672aba1f0..5b5be9577257 100644
--- a/net/core/flow_dissector.c
+++ b/net/core/flow_dissector.c
@@ -4,6 +4,7 @@
 #include <linux/ip.h>
 #include <linux/ipv6.h>
 #include <linux/if_vlan.h>
+#include <net/dsa.h>
 #include <net/ip.h>
 #include <net/ipv6.h>
 #include <net/gre.h>
@@ -440,6 +441,17 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
 			 skb->vlan_proto : skb->protocol;
 		nhoff = skb_network_offset(skb);
 		hlen = skb_headlen(skb);
+		if (unlikely(netdev_uses_dsa(skb->dev))) {
+			const struct dsa_device_ops *ops;
+			int offset;
+
+			ops = skb->dev->dsa_ptr->tag_ops;
+			if (ops->flow_dissect &&
+			    !ops->flow_dissect(skb, &proto, &offset)) {
+				hlen -= offset;
+				nhoff += offset;
+			}
+		}
 	}
 
 	/* It is ensured by skb_flow_dissector_init() that control key will
-- 
2.11.0

  parent reply	other threads:[~2017-08-09 12:42 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-08-09 12:41 [PATCH 0/4] net-next: dsa: fix flow dissection John Crispin
2017-08-09 12:41 ` John Crispin
2017-08-09 12:41 ` [PATCH 1/4] net-next: dsa: move struct dsa_device_ops to the global header file John Crispin
2017-08-09 13:45   ` Andrew Lunn
2017-08-09 13:45     ` Andrew Lunn
2017-08-09 12:41 ` [PATCH 2/4] net-next: dsa: add flow_dissect callback to struct dsa_device_ops John Crispin
2017-08-09 12:41   ` John Crispin
2017-08-09 13:51   ` Andrew Lunn
2017-08-09 13:51     ` Andrew Lunn
2017-08-09 12:41 ` [PATCH 3/4] net-next: tag_mtk: add flow_dissect callback to the ops struct John Crispin
2017-08-09 13:54   ` Andrew Lunn
2017-08-09 12:41 ` John Crispin [this message]
2017-08-09 12:41   ` [PATCH 4/4] net-next: dsa: fix flow dissection John Crispin
2017-08-09 13:55   ` Andrew Lunn
2017-08-09 13:55     ` Andrew Lunn
2017-08-09 15:07 ` [PATCH 0/4] " Vivien Didelot
2017-08-09 15:07   ` Vivien Didelot
2017-08-10  5:52 ` David Miller
2017-08-10  6:42   ` Eric Dumazet
2017-08-10  7:40     ` John Crispin
2017-08-10  8:10     ` John Crispin

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=20170809124119.24320-5-john@phrozen.org \
    --to=john@phrozen.org \
    --cc=andrew@lunn.ch \
    --cc=davem@davemloft.net \
    --cc=f.fainelli@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mediatek@lists.infradead.org \
    --cc=muciri@openmesh.com \
    --cc=netdev@vger.kernel.org \
    --cc=sean.wang@mediatek.com \
    --cc=shashidhar.lakkavalli@openmesh.com \
    --cc=vivien.didelot@savoirfairelinux.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.