* [PATCH] net: fix "queues" uevent between network namespaces
@ 2014-01-16 9:24 Chen Weilong
2014-01-16 16:19 ` Greg KH
2014-01-20 4:04 ` David Miller
0 siblings, 2 replies; 4+ messages in thread
From: Chen Weilong @ 2014-01-16 9:24 UTC (permalink / raw)
To: davem, gregkh; +Cc: netdev
From: Weilong Chen <chenweilong@huawei.com>
When I create a new namespace with 'ip netns add net0', or add/remove
new links in a namespace with 'ip link add/delete type veth', rx/tx
queues events can be got in all namespaces. That is because rx/tx queue
ktypes do not have namespace support, and their kobj parents are setted to
NULL. This patch is to fix it.
Reported-by: Libo Chen <chenlibo@huawei.com>
Signed-off-by: Libo Chen <chenlibo@huawei.com>
Signed-off-by: Weilong Chen <chenweilong@huawei.com>
---
lib/kobject_uevent.c | 10 ++++++++--
net/core/net-sysfs.c | 26 ++++++++++++++++++++++++++
2 files changed, 34 insertions(+), 2 deletions(-)
diff --git a/lib/kobject_uevent.c b/lib/kobject_uevent.c
index 52e5abb..5f72767 100644
--- a/lib/kobject_uevent.c
+++ b/lib/kobject_uevent.c
@@ -88,11 +88,17 @@ out:
#ifdef CONFIG_NET
static int kobj_bcast_filter(struct sock *dsk, struct sk_buff *skb, void *data)
{
- struct kobject *kobj = data;
+ struct kobject *kobj = data, *ksobj;
const struct kobj_ns_type_operations *ops;
ops = kobj_ns_ops(kobj);
- if (ops) {
+ if (!ops && kobj->kset) {
+ ksobj = &kobj->kset->kobj;
+ if (ksobj->parent != NULL)
+ ops = kobj_ns_ops(ksobj->parent);
+ }
+
+ if (ops && ops->netlink_ns && kobj->ktype->namespace) {
const void *sock_ns, *ns;
ns = kobj->ktype->namespace(kobj);
sock_ns = ops->netlink_ns(dsk);
diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c
index 49843bf..0cabe7c 100644
--- a/net/core/net-sysfs.c
+++ b/net/core/net-sysfs.c
@@ -748,10 +748,23 @@ static void rx_queue_release(struct kobject *kobj)
dev_put(queue->dev);
}
+static const void *rx_queue_namespace(struct kobject *kobj)
+{
+ struct netdev_rx_queue *queue = to_rx_queue(kobj);
+ struct device *dev = &queue->dev->dev;
+ const void *ns = NULL;
+
+ if (dev->class && dev->class->ns_type)
+ ns = dev->class->namespace(dev);
+
+ return ns;
+}
+
static struct kobj_type rx_queue_ktype = {
.sysfs_ops = &rx_queue_sysfs_ops,
.release = rx_queue_release,
.default_attrs = rx_queue_default_attrs,
+ .namespace = rx_queue_namespace
};
static int rx_queue_add_kobject(struct net_device *net, int index)
@@ -1082,10 +1095,23 @@ static void netdev_queue_release(struct kobject *kobj)
dev_put(queue->dev);
}
+static const void *netdev_queue_namespace(struct kobject *kobj)
+{
+ struct netdev_queue *queue = to_netdev_queue(kobj);
+ struct device *dev = &queue->dev->dev;
+ const void *ns = NULL;
+
+ if (dev->class && dev->class->ns_type)
+ ns = dev->class->namespace(dev);
+
+ return ns;
+}
+
static struct kobj_type netdev_queue_ktype = {
.sysfs_ops = &netdev_queue_sysfs_ops,
.release = netdev_queue_release,
.default_attrs = netdev_queue_default_attrs,
+ .namespace = netdev_queue_namespace,
};
static int netdev_queue_add_kobject(struct net_device *net, int index)
--
1.7.12
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH] net: fix "queues" uevent between network namespaces
2014-01-16 9:24 [PATCH] net: fix "queues" uevent between network namespaces Chen Weilong
@ 2014-01-16 16:19 ` Greg KH
2014-01-17 1:47 ` chenweilong
2014-01-20 4:04 ` David Miller
1 sibling, 1 reply; 4+ messages in thread
From: Greg KH @ 2014-01-16 16:19 UTC (permalink / raw)
To: Chen Weilong; +Cc: davem, netdev
On Thu, Jan 16, 2014 at 05:24:31PM +0800, Chen Weilong wrote:
> From: Weilong Chen <chenweilong@huawei.com>
>
> When I create a new namespace with 'ip netns add net0', or add/remove
> new links in a namespace with 'ip link add/delete type veth', rx/tx
> queues events can be got in all namespaces. That is because rx/tx queue
> ktypes do not have namespace support, and their kobj parents are setted to
> NULL. This patch is to fix it.
>
> Reported-by: Libo Chen <chenlibo@huawei.com>
> Signed-off-by: Libo Chen <chenlibo@huawei.com>
> Signed-off-by: Weilong Chen <chenweilong@huawei.com>
> ---
> lib/kobject_uevent.c | 10 ++++++++--
> net/core/net-sysfs.c | 26 ++++++++++++++++++++++++++
> 2 files changed, 34 insertions(+), 2 deletions(-)
I can't test this, but it looks good to me:
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] net: fix "queues" uevent between network namespaces
2014-01-16 16:19 ` Greg KH
@ 2014-01-17 1:47 ` chenweilong
0 siblings, 0 replies; 4+ messages in thread
From: chenweilong @ 2014-01-17 1:47 UTC (permalink / raw)
To: Greg KH; +Cc: davem, netdev
On 2014/1/17 0:19, Greg KH wrote:
> On Thu, Jan 16, 2014 at 05:24:31PM +0800, Chen Weilong wrote:
>> From: Weilong Chen <chenweilong@huawei.com>
>>
>> When I create a new namespace with 'ip netns add net0', or add/remove
>> new links in a namespace with 'ip link add/delete type veth', rx/tx
>> queues events can be got in all namespaces. That is because rx/tx queue
>> ktypes do not have namespace support, and their kobj parents are setted to
>> NULL. This patch is to fix it.
>>
>> Reported-by: Libo Chen <chenlibo@huawei.com>
>> Signed-off-by: Libo Chen <chenlibo@huawei.com>
>> Signed-off-by: Weilong Chen <chenweilong@huawei.com>
>> ---
>> lib/kobject_uevent.c | 10 ++++++++--
>> net/core/net-sysfs.c | 26 ++++++++++++++++++++++++++
>> 2 files changed, 34 insertions(+), 2 deletions(-)
>
> I can't test this, but it looks good to me:
>
> Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>
>
Hi,
Here's detail about the test, I hope it can be useful:
step1: create two netns
suse11-sp3:~ # ip netns add net0
suse11-sp3:~ # ip netns add net1
suse11-sp3:~ # ip netns list
net1
net0
setp2: monitor udev events
Term1:
suse11-sp3:~ # ip netns exec net0 udevadm monitor
Term2:
suse11-sp3:~ # ip netns exec net1 udevadm monitor
setp3: add link to net0
suse11-sp3:~ # ip netns exec net0 ip link add type veth
Then you'll see the below events in net0 and net1.
KERNEL[1389972662.984988] add /devices/virtual/net/veth0/queues/rx-0 (queues)
KERNEL[1389972662.985008] add /devices/virtual/net/veth0/queues/tx-0 (queues)
KERNEL[1389972662.985234] add /devices/virtual/net/veth1/queues/rx-0 (queues)
KERNEL[1389972662.985247] add /devices/virtual/net/veth1/queues/tx-0 (queues)
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] net: fix "queues" uevent between network namespaces
2014-01-16 9:24 [PATCH] net: fix "queues" uevent between network namespaces Chen Weilong
2014-01-16 16:19 ` Greg KH
@ 2014-01-20 4:04 ` David Miller
1 sibling, 0 replies; 4+ messages in thread
From: David Miller @ 2014-01-20 4:04 UTC (permalink / raw)
To: chenweilong; +Cc: gregkh, netdev
From: Chen Weilong <chenweilong@huawei.com>
Date: Thu, 16 Jan 2014 17:24:31 +0800
> From: Weilong Chen <chenweilong@huawei.com>
>
> When I create a new namespace with 'ip netns add net0', or add/remove
> new links in a namespace with 'ip link add/delete type veth', rx/tx
> queues events can be got in all namespaces. That is because rx/tx queue
> ktypes do not have namespace support, and their kobj parents are setted to
> NULL. This patch is to fix it.
>
> Reported-by: Libo Chen <chenlibo@huawei.com>
> Signed-off-by: Libo Chen <chenlibo@huawei.com>
> Signed-off-by: Weilong Chen <chenweilong@huawei.com>
Applied to net-next, thanks.
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2014-01-20 4:04 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-01-16 9:24 [PATCH] net: fix "queues" uevent between network namespaces Chen Weilong
2014-01-16 16:19 ` Greg KH
2014-01-17 1:47 ` chenweilong
2014-01-20 4:04 ` David Miller
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).