netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
To: netdev@vger.kernel.org, Horatiu Vultur <horatiu.vultur@microchip.com>
Cc: linux-kernel@vger.kernel.org, Jakub Kicinski <kuba@kernel.org>,
	"David S. Miller" <davem@davemloft.net>,
	Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Subject: [PATCH net 2/2] net: switchdev: don't set port_obj_info->handled true when -EOPNOTSUPP
Date: Wed, 23 Dec 2020 15:45:33 +0100	[thread overview]
Message-ID: <20201223144533.4145-3-rasmus.villemoes@prevas.dk> (raw)
In-Reply-To: <20201223144533.4145-1-rasmus.villemoes@prevas.dk>

It's not true that switchdev_port_obj_notify() only inspects the
->handled field of "struct switchdev_notifier_port_obj_info" if
call_switchdev_blocking_notifiers() returns 0 - there's a WARN_ON()
triggering for a non-zero return combined with ->handled not being
true. But the real problem here is that -EOPNOTSUPP is not being
properly handled.

The wrapper functions switchdev_handle_port_obj_add() et al change a
return value of -EOPNOTSUPP to 0, and the treatment of ->handled in
switchdev_port_obj_notify() seems to be designed to change that back
to -EOPNOTSUPP in case nobody actually acted on the notifier (i.e.,
everybody returned -EOPNOTSUPP).

Currently, as soon as some device down the stack passes the check_cb()
check, ->handled gets set to true, which means that
switchdev_port_obj_notify() cannot actually ever return -EOPNOTSUPP.

This, for example, means that the detection of hardware offload
support in the MRP code is broken - br_mrp_set_ring_role() always ends
up setting mrp->ring_role_offloaded to 1, despite not a single
mainline driver implementing any of the SWITCHDEV_OBJ_ID*_MRP. So
since the MRP code thinks the generation of MRP test frames has been
offloaded, no such frames are actually put on the wire.

So, continue to set ->handled true if any callback returns success or
any error distinct from -EOPNOTSUPP. But if all the callbacks return
-EOPNOTSUPP, make sure that ->handled stays false, so the logic in
switchdev_port_obj_notify() can propagate that information.

Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
---
 net/switchdev/switchdev.c | 23 +++++++++++++----------
 1 file changed, 13 insertions(+), 10 deletions(-)

diff --git a/net/switchdev/switchdev.c b/net/switchdev/switchdev.c
index 23d868545362..2c1ffc9ba2eb 100644
--- a/net/switchdev/switchdev.c
+++ b/net/switchdev/switchdev.c
@@ -460,10 +460,11 @@ static int __switchdev_handle_port_obj_add(struct net_device *dev,
 	extack = switchdev_notifier_info_to_extack(&port_obj_info->info);
 
 	if (check_cb(dev)) {
-		/* This flag is only checked if the return value is success. */
-		port_obj_info->handled = true;
-		return add_cb(dev, port_obj_info->obj, port_obj_info->trans,
-			      extack);
+		err = add_cb(dev, port_obj_info->obj, port_obj_info->trans,
+			     extack);
+		if (err != -EOPNOTSUPP)
+			port_obj_info->handled = true;
+		return err;
 	}
 
 	/* Switch ports might be stacked under e.g. a LAG. Ignore the
@@ -515,9 +516,10 @@ static int __switchdev_handle_port_obj_del(struct net_device *dev,
 	int err = -EOPNOTSUPP;
 
 	if (check_cb(dev)) {
-		/* This flag is only checked if the return value is success. */
-		port_obj_info->handled = true;
-		return del_cb(dev, port_obj_info->obj);
+		err = del_cb(dev, port_obj_info->obj);
+		if (err != -EOPNOTSUPP)
+			port_obj_info->handled = true;
+		return err;
 	}
 
 	/* Switch ports might be stacked under e.g. a LAG. Ignore the
@@ -568,9 +570,10 @@ static int __switchdev_handle_port_attr_set(struct net_device *dev,
 	int err = -EOPNOTSUPP;
 
 	if (check_cb(dev)) {
-		port_attr_info->handled = true;
-		return set_cb(dev, port_attr_info->attr,
-			      port_attr_info->trans);
+		err = set_cb(dev, port_attr_info->attr, port_attr_info->trans);
+		if (err != -EOPNOTSUPP)
+			port_attr_info->handled = true;
+		return err;
 	}
 
 	/* Switch ports might be stacked under e.g. a LAG. Ignore the
-- 
2.23.0


  parent reply	other threads:[~2020-12-23 14:47 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-12-23 14:45 [PATCH net 0/2] MRP without hardware offload? Rasmus Villemoes
2020-12-23 14:45 ` [PATCH net 1/2] net: mrp: fix definitions of MRP test packets Rasmus Villemoes
2020-12-23 17:59   ` Horatiu Vultur
2020-12-23 18:41     ` Andrew Lunn
2020-12-23 19:22       ` Horatiu Vultur
2020-12-28 22:24   ` Jakub Kicinski
2021-01-03 13:29     ` Horatiu Vultur
2021-01-19 15:42     ` Rasmus Villemoes
2021-01-19 17:45       ` Jakub Kicinski
2020-12-23 14:45 ` Rasmus Villemoes [this message]
2020-12-28 22:26   ` [PATCH net 2/2] net: switchdev: don't set port_obj_info->handled true when -EOPNOTSUPP Jakub Kicinski
2020-12-23 18:14 ` [PATCH net 0/2] MRP without hardware offload? Horatiu Vultur

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=20201223144533.4145-3-rasmus.villemoes@prevas.dk \
    --to=rasmus.villemoes@prevas.dk \
    --cc=davem@davemloft.net \
    --cc=horatiu.vultur@microchip.com \
    --cc=kuba@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=netdev@vger.kernel.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;
as well as URLs for NNTP newsgroup(s).