Skip to contentSkip to conversation

net/hsr: Use list_head+rcu, better frame dispatch, etc.

PATCH In review Arvid Brodin 2 participants 3 messages Jul 4, 2014 – Jul 11, 2014

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover net/hsr: Use list_head+rcu, better frame dispatch, etc.

This patch series is meant to improve the HSR code in several ways:

* Better code readability.
* In general, make the code structure more like the net/bridge code (HSR operates similarly to a bridge, but uses the HSR-specific frame headers to break up rings, instead of the STP protocol).
* Better handling of HSR ports' net_device features.
* Use list_head and the _rcu list traversing routines instead of array of slave devices.
* Make it easy to support HSR Interlink devices (for future Redbox/Quadbox support).
* Somewhat better throughput on non-HAVE_EFFICIENT_UNALIGNED_ACCESS archs, due to lesser copying of skb data.

The code has been tested in a ring together with other HSR nodes running unchanged code, on both avr32 and x86_64. There should only be one minor change in behaviour from a user perspective:

* Anyone using the Netlink HSR_C_GET_NODE_LIST message to dump the internal node database will notice that the database now also contains the self node.

All patches pass 'checkpatch.pl --ignore CAMELCASE --max-line-length=83 --strict' with only CHECKs, each of which have been deliberately left in place.

The final code passes sparse checks with no output.

Arvid Brodin (10):
net/hsr: Better variable names and update of contact info. net/hsr: Switch from dev_add_pack() to netdev_rx_handler_register() net/hsr: Move to per-hsr device prune timer.
net/hsr: Operstate handling cleanup.
net/hsr: Move slave init to hsr_slave.c.
net/hsr: Use list_head (and rcu) instead of array for slave devices. net/hsr: Implemented .ndo_fix_features (better device features
handling).
net/hsr: Added SET_NETDEV_DEVTYPE and features |= NETIF_F_NETNS_LOCAL
to dev_setup.
net/hsr: Better frame dispatch
net/hsr: Fix NULL pointer dereference on incomplete hsr_newlink()
params.

 net/hsr/Makefile       |   3 +-
 net/hsr/hsr_device.c   | 580 +++++++++++++++++++++----------------------------
 net/hsr/hsr_device.h   |  12 +-
 net/hsr/hsr_forward.c  | 368 +++++++++++++++++++++++++++++++
 net/hsr/hsr_forward.h  |  20 ++
 net/hsr/hsr_framereg.c | 481 ++++++++++++++++++++--------------------
 net/hsr/hsr_framereg.h |  45 ++--
 net/hsr/hsr_main.c     | 425 ++++--------------------------------
 net/hsr/hsr_main.h     |  61 ++++--
 net/hsr/hsr_netlink.c  | 102 +++++----
 net/hsr/hsr_netlink.h  |  11 +-
 net/hsr/hsr_slave.c    | 196 +++++++++++++++++
 net/hsr/hsr_slave.h    |  38 ++++
 13 files changed, 1285 insertions(+), 1057 deletions(-)
 create mode 100644 net/hsr/hsr_forward.c
 create mode 100644 net/hsr/hsr_forward.h
 create mode 100644 net/hsr/hsr_slave.c
 create mode 100644 net/hsr/hsr_slave.h
Signature · 6 lines
1.8.3.2

-- 
Arvid Brodin | Consultant (Linux)
ALTEN | Knarrarnäsgatan 7 | SE-164 40 Kista | Sweden
[email protected] | www.alten.se/en/
maintainer
#1 lore raw

From: Arvid Brodin <[email protected]>
Date: Fri, 4 Jul 2014 23:33:44 +0200

Quoted 18 lines

This patch series is meant to improve the HSR code in several ways:

* Better code readability.
* In general, make the code structure more like the net/bridge code (HSR operates similarly to a bridge, but uses the HSR-specific frame headers to break up rings, instead of the STP protocol).
* Better handling of HSR ports' net_device features.
* Use list_head and the _rcu list traversing routines instead of array of slave devices.
* Make it easy to support HSR Interlink devices (for future Redbox/Quadbox support).
* Somewhat better throughput on non-HAVE_EFFICIENT_UNALIGNED_ACCESS archs, due to lesser copying of skb data.

The code has been tested in a ring together with other HSR nodes running unchanged code, on both avr32 and x86_64. There should only be one minor change in behaviour from a user perspective:

Quoted 7 lines

* Anyone using the Netlink HSR_C_GET_NODE_LIST message to dump the internal node database will notice that the database now also contains the self node.

All patches pass 'checkpatch.pl --ignore CAMELCASE --max-line-length=83 --strict' with only CHECKs, each of which have been deliberately left in place.

The final code passes sparse checks with no output.

Series applied, thanks.

Just one question:

@@ -575,6 +579,13 @@ int hsr_dev_finalize(struct net_device *hsr_dev, struct net_device *slave[2],
}
}

+ for (i = 0; i < HSR_MAX_SLAVE; i++) {
+ res = netdev_rx_handler_register(slave[i], hsr_handle_frame,
+ hsr);
+ if (res)
+ goto fail;
+ }
+
/* Make sure we recognize frames from ourselves in hsr_rcv() */ res = hsr_create_self_node(&hsr->self_node_db, hsr_dev->dev_addr,
hsr->slave[1]->dev_addr);

If one of the slaves fails to register it's RX handler here, what unregisters the slaves which already succeeded?

Thanks.

#2 lore raw
Quoted 18 lines · David Miller

Just one question:

(quoted diff) +7 -0
@@ -575,6 +579,13 @@ int hsr_dev_finalize(struct net_device *hsr_dev, struct net_device *slave[2],
 		}
 	}
 
+	for (i = 0; i < HSR_MAX_SLAVE; i++) {
+		res = netdev_rx_handler_register(slave[i], hsr_handle_frame,
+						 hsr);
+		if (res)
+			goto fail;
+	}
+
 	/* Make sure we recognize frames from ourselves in hsr_rcv() */
 	res = hsr_create_self_node(&hsr->self_node_db, hsr_dev->dev_addr,
 				   hsr->slave[1]->dev_addr);
Quoted 2 lines

If one of the slaves fails to register it's RX handler here, what unregisters the slaves which already succeeded?

This looks like it's from "[PATCH net-next 02/10] net/hsr: Switch from dev_add_pack() to netdev_rx_handler_register()". If so, the code under the fail: label calls restore_slaves(), which in turn iterates over the slaves. From the same patch:

@@ -402,8 +402,12 @@ static void restore_slaves(struct net_device *hsr_dev)
netdev_info(hsr_dev,
"Cannot restore slave promiscuity (%s, %d)\n", hsr->slave[i]->name, res);
+
+ if (hsr->slave[i]->rx_handler == hsr_handle_frame)
+ netdev_rx_handler_unregister(hsr->slave[i]);
}

+
rtnl_unlock();
}

Signature · 3 lines
Arvid Brodin | Consultant (Linux)
ALTEN | Knarrarnäsgatan 7 | SE-164 40 Kista | Sweden
[email protected] | www.alten.se/en/