[0/2] Fix problems with with IFLA_VF_PORTS

[0/2] Fix problems with with IFLA_VF_PORTS

[David Gibson]

I've had a customer encounter a problem with getifaddrs(3) freezing up
on a system with a Cisco enic device.

I've discovered that the problem is caused by an enic device with a
large number of SR-IOV virtual functions overflowing the normal sized
packet buffer for netlink, leading to interfaces not being reported
from an RTM_GETLINK request.

The first patch here just makes the problem easier to locate if it
occurs again in a different way, by adding a WARN_ON() when we run out
of room in a netlink packet in this manner.

The second patch actually fixes the problem, by only reporting
IFLA_VF_PORTS information when the RTEXT_FILTER_VF flag is specified.

[PATCH 1/2] rtnetlink: Warn when interface's information won't fit in our packet

[David Gibson]

Without IFLA_EXT_MASK specified, the information reported for a single
interface in response to RTM_GETLINK is expected to fit within a netlink
packet of NLMSG_GOODSIZE.

If it doesn't, however, things will go badly wrong,  When listing all
interfaces, netlink_dump() will incorrectly treat -EMSGSIZE on the first
message in a packet as the end of the listing and omit information for
that interface and all subsequent ones.  This can cause getifaddrs(3) to
enter an infinite loop.

This patch won't fix the problem, but it will WARN_ON() making it easier to
track down what's going wrong.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
---
 net/core/rtnetlink.c | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)

diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
index d4ff417..5331db2 100644
--- a/net/core/rtnetlink.c
+++ b/net/core/rtnetlink.c
@@ -1198,6 +1198,7 @@ static int rtnl_dump_ifinfo(struct sk_buff *skb, struct netlink_callback *cb)
 	struct hlist_head *head;
 	struct nlattr *tb[IFLA_MAX+1];
 	u32 ext_filter_mask = 0;
+	int err;
 
 	s_h = cb->args[0];
 	s_idx = cb->args[1];
@@ -1218,11 +1219,16 @@ static int rtnl_dump_ifinfo(struct sk_buff *skb, struct netlink_callback *cb)
 		hlist_for_each_entry_rcu(dev, head, index_hlist) {
 			if (idx < s_idx)
 				goto cont;
-			if (rtnl_fill_ifinfo(skb, dev, RTM_NEWLINK,
-					     NETLINK_CB(cb->skb).portid,
-					     cb->nlh->nlmsg_seq, 0,
-					     NLM_F_MULTI,
-					     ext_filter_mask) <= 0)
+			err = rtnl_fill_ifinfo(skb, dev, RTM_NEWLINK,
+					       NETLINK_CB(cb->skb).portid,
+					       cb->nlh->nlmsg_seq, 0,
+					       NLM_F_MULTI,
+					       ext_filter_mask);
+			/* If we ran out of room on the first message,
+			 * we're in trouble */
+			WARN_ON((err == -EMSGSIZE) && (skb->len == 0));
+
+			if (err <= 0)
 				goto out;
 
 			nl_dump_check_consistent(cb, nlmsg_hdr(skb));
-- 
1.9.0

Re: [PATCH 1/2] rtnetlink: Warn when interface's information won't fit in our packet

[Jiri Pirko]

Wed, Apr 23, 2014 at 09:21:04AM CEST, david@gibson.dropbear.id.au wrote:
>Without IFLA_EXT_MASK specified, the information reported for a single
>interface in response to RTM_GETLINK is expected to fit within a netlink
>packet of NLMSG_GOODSIZE.
>
>If it doesn't, however, things will go badly wrong,  When listing all
>interfaces, netlink_dump() will incorrectly treat -EMSGSIZE on the first
>message in a packet as the end of the listing and omit information for
>that interface and all subsequent ones.  This can cause getifaddrs(3) to
>enter an infinite loop.
>
>This patch won't fix the problem, but it will WARN_ON() making it easier to
>track down what's going wrong.
>
>Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
>---
> net/core/rtnetlink.c | 16 +++++++++++-----
> 1 file changed, 11 insertions(+), 5 deletions(-)
>
>diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
>index d4ff417..5331db2 100644
>--- a/net/core/rtnetlink.c
>+++ b/net/core/rtnetlink.c
>@@ -1198,6 +1198,7 @@ static int rtnl_dump_ifinfo(struct sk_buff *skb, struct netlink_callback *cb)
> 	struct hlist_head *head;
> 	struct nlattr *tb[IFLA_MAX+1];
> 	u32 ext_filter_mask = 0;
>+	int err;
> 
> 	s_h = cb->args[0];
> 	s_idx = cb->args[1];
>@@ -1218,11 +1219,16 @@ static int rtnl_dump_ifinfo(struct sk_buff *skb, struct netlink_callback *cb)
> 		hlist_for_each_entry_rcu(dev, head, index_hlist) {
> 			if (idx < s_idx)
> 				goto cont;
>-			if (rtnl_fill_ifinfo(skb, dev, RTM_NEWLINK,
>-					     NETLINK_CB(cb->skb).portid,
>-					     cb->nlh->nlmsg_seq, 0,
>-					     NLM_F_MULTI,
>-					     ext_filter_mask) <= 0)
>+			err = rtnl_fill_ifinfo(skb, dev, RTM_NEWLINK,
>+					       NETLINK_CB(cb->skb).portid,
>+					       cb->nlh->nlmsg_seq, 0,
>+					       NLM_F_MULTI,
>+					       ext_filter_mask);
>+			/* If we ran out of room on the first message,
>+			 * we're in trouble */
>+			WARN_ON((err == -EMSGSIZE) && (skb->len == 0));
>+
>+			if (err <= 0)
			if (err)

> 				goto out;
> 
> 			nl_dump_check_consistent(cb, nlmsg_hdr(skb));
>-- 
>1.9.0
>
>--
>To unsubscribe from this list: send the line "unsubscribe netdev" in
>the body of a message to majordomo@vger.kernel.org
>More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/2] rtnetlink: Warn when interface's information won't fit in our packet

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 2475 bytes --]

On Wed, Apr 23, 2014 at 11:09:05AM +0200, Jiri Pirko wrote:
> Wed, Apr 23, 2014 at 09:21:04AM CEST, david@gibson.dropbear.id.au wrote:
> >Without IFLA_EXT_MASK specified, the information reported for a single
> >interface in response to RTM_GETLINK is expected to fit within a netlink
> >packet of NLMSG_GOODSIZE.
> >
> >If it doesn't, however, things will go badly wrong,  When listing all
> >interfaces, netlink_dump() will incorrectly treat -EMSGSIZE on the first
> >message in a packet as the end of the listing and omit information for
> >that interface and all subsequent ones.  This can cause getifaddrs(3) to
> >enter an infinite loop.
> >
> >This patch won't fix the problem, but it will WARN_ON() making it easier to
> >track down what's going wrong.
> >
> >Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
> >---
> > net/core/rtnetlink.c | 16 +++++++++++-----
> > 1 file changed, 11 insertions(+), 5 deletions(-)
> >
> >diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
> >index d4ff417..5331db2 100644
> >--- a/net/core/rtnetlink.c
> >+++ b/net/core/rtnetlink.c
> >@@ -1198,6 +1198,7 @@ static int rtnl_dump_ifinfo(struct sk_buff *skb, struct netlink_callback *cb)
> > 	struct hlist_head *head;
> > 	struct nlattr *tb[IFLA_MAX+1];
> > 	u32 ext_filter_mask = 0;
> >+	int err;
> > 
> > 	s_h = cb->args[0];
> > 	s_idx = cb->args[1];
> >@@ -1218,11 +1219,16 @@ static int rtnl_dump_ifinfo(struct sk_buff *skb, struct netlink_callback *cb)
> > 		hlist_for_each_entry_rcu(dev, head, index_hlist) {
> > 			if (idx < s_idx)
> > 				goto cont;
> >-			if (rtnl_fill_ifinfo(skb, dev, RTM_NEWLINK,
> >-					     NETLINK_CB(cb->skb).portid,
> >-					     cb->nlh->nlmsg_seq, 0,
> >-					     NLM_F_MULTI,
> >-					     ext_filter_mask) <= 0)
> >+			err = rtnl_fill_ifinfo(skb, dev, RTM_NEWLINK,
> >+					       NETLINK_CB(cb->skb).portid,
> >+					       cb->nlh->nlmsg_seq, 0,
> >+					       NLM_F_MULTI,
> >+					       ext_filter_mask);
> >+			/* If we ran out of room on the first message,
> >+			 * we're in trouble */
> >+			WARN_ON((err == -EMSGSIZE) && (skb->len == 0));
> >+
> >+			if (err <= 0)
> 			if (err)

<= 0 was what the original version had.  Why do you think it should be
changed to != 0?

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: Type: application/pgp-signature, Size: 819 bytes --]

Re: [PATCH 1/2] rtnetlink: Warn when interface's information won't fit in our packet

[Jiri Pirko]

Wed, Apr 23, 2014 at 01:41:20PM CEST, david@gibson.dropbear.id.au wrote:
>On Wed, Apr 23, 2014 at 11:09:05AM +0200, Jiri Pirko wrote:
>> Wed, Apr 23, 2014 at 09:21:04AM CEST, david@gibson.dropbear.id.au wrote:
>> >Without IFLA_EXT_MASK specified, the information reported for a single
>> >interface in response to RTM_GETLINK is expected to fit within a netlink
>> >packet of NLMSG_GOODSIZE.
>> >
>> >If it doesn't, however, things will go badly wrong,  When listing all
>> >interfaces, netlink_dump() will incorrectly treat -EMSGSIZE on the first
>> >message in a packet as the end of the listing and omit information for
>> >that interface and all subsequent ones.  This can cause getifaddrs(3) to
>> >enter an infinite loop.
>> >
>> >This patch won't fix the problem, but it will WARN_ON() making it easier to
>> >track down what's going wrong.
>> >
>> >Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
>> >---
>> > net/core/rtnetlink.c | 16 +++++++++++-----
>> > 1 file changed, 11 insertions(+), 5 deletions(-)
>> >
>> >diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
>> >index d4ff417..5331db2 100644
>> >--- a/net/core/rtnetlink.c
>> >+++ b/net/core/rtnetlink.c
>> >@@ -1198,6 +1198,7 @@ static int rtnl_dump_ifinfo(struct sk_buff *skb, struct netlink_callback *cb)
>> > 	struct hlist_head *head;
>> > 	struct nlattr *tb[IFLA_MAX+1];
>> > 	u32 ext_filter_mask = 0;
>> >+	int err;
>> > 
>> > 	s_h = cb->args[0];
>> > 	s_idx = cb->args[1];
>> >@@ -1218,11 +1219,16 @@ static int rtnl_dump_ifinfo(struct sk_buff *skb, struct netlink_callback *cb)
>> > 		hlist_for_each_entry_rcu(dev, head, index_hlist) {
>> > 			if (idx < s_idx)
>> > 				goto cont;
>> >-			if (rtnl_fill_ifinfo(skb, dev, RTM_NEWLINK,
>> >-					     NETLINK_CB(cb->skb).portid,
>> >-					     cb->nlh->nlmsg_seq, 0,
>> >-					     NLM_F_MULTI,
>> >-					     ext_filter_mask) <= 0)
>> >+			err = rtnl_fill_ifinfo(skb, dev, RTM_NEWLINK,
>> >+					       NETLINK_CB(cb->skb).portid,
>> >+					       cb->nlh->nlmsg_seq, 0,
>> >+					       NLM_F_MULTI,
>> >+					       ext_filter_mask);
>> >+			/* If we ran out of room on the first message,
>> >+			 * we're in trouble */
>> >+			WARN_ON((err == -EMSGSIZE) && (skb->len == 0));
>> >+
>> >+			if (err <= 0)
>> 			if (err)
>
><= 0 was what the original version had.  Why do you think it should be
>changed to != 0?

You are right. I misread rtnl_fill_ifinfo.

>
>-- 
>David Gibson			| I'll have my music baroque, and my code
>david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
>				| _way_ _around_!
>http://www.ozlabs.org/~dgibson

Re: [PATCH 1/2] rtnetlink: Warn when interface's information won't fit in our packet

[Sergei Shtylyov]

Hello.

On 23-04-2014 11:21, David Gibson wrote:

> Without IFLA_EXT_MASK specified, the information reported for a single
> interface in response to RTM_GETLINK is expected to fit within a netlink
> packet of NLMSG_GOODSIZE.

> If it doesn't, however, things will go badly wrong,  When listing all
> interfaces, netlink_dump() will incorrectly treat -EMSGSIZE on the first
> message in a packet as the end of the listing and omit information for
> that interface and all subsequent ones.  This can cause getifaddrs(3) to
> enter an infinite loop.

> This patch won't fix the problem, but it will WARN_ON() making it easier to
> track down what's going wrong.

> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
> ---
>   net/core/rtnetlink.c | 16 +++++++++++-----
>   1 file changed, 11 insertions(+), 5 deletions(-)

> diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
> index d4ff417..5331db2 100644
> --- a/net/core/rtnetlink.c
> +++ b/net/core/rtnetlink.c
[...]
> @@ -1218,11 +1219,16 @@ static int rtnl_dump_ifinfo(struct sk_buff *skb, struct netlink_callback *cb)
>   		hlist_for_each_entry_rcu(dev, head, index_hlist) {
>   			if (idx < s_idx)
>   				goto cont;
> -			if (rtnl_fill_ifinfo(skb, dev, RTM_NEWLINK,
> -					     NETLINK_CB(cb->skb).portid,
> -					     cb->nlh->nlmsg_seq, 0,
> -					     NLM_F_MULTI,
> -					     ext_filter_mask) <= 0)
> +			err = rtnl_fill_ifinfo(skb, dev, RTM_NEWLINK,
> +					       NETLINK_CB(cb->skb).portid,
> +					       cb->nlh->nlmsg_seq, 0,
> +					       NLM_F_MULTI,
> +					       ext_filter_mask);
> +			/* If we ran out of room on the first message,
> +			 * we're in trouble */

    Hm, comment style... multi-line comments in the networking code should 
look like this:

/* bla
  * bla
  */

> +			WARN_ON((err == -EMSGSIZE) && (skb->len == 0));
> +
> +			if (err <= 0)

WBR, Sergei

[PATCH 2/2] rtnetlink: Only supply IFLA_VF_PORTS information when RTEXT_FILTER_VF is set

[David Gibson]

Since 115c9b81928360d769a76c632bae62d15206a94a (rtnetlink: Fix problem with
buffer allocation), RTM_NEWLINK messages only contain the IFLA_VFINFO_LIST
attribute if they were solicited by a GETLINK message containing an
IFLA_EXT_MASK attribute with the RTEXT_FILTER_VF flag.

That was done because some user programs broke when they received more data
than expected - because IFLA_VFINFO_LIST contains information for each VF
it can become large if there are many VFs.

However, the IFLA_VF_PORTS attribute, supplied for devices which implement
ndo_get_vf_port (currently the 'enic' driver only), has the same problem.
It supplies per-VF information and can therefore become large, but it is
not currently conditional on the IFLA_EXT_MASK value.

Worse, it interacts badly with the existing EXT_MASK handling.  When
IFLA_EXT_MASK is not supplied, the buffer for netlink replies is fixed at
NLMSG_GOODSIZE.  If the information for IFLA_VF_PORTS exceeds this, then
rtnl_fill_ifinfo() returns -EMSGSIZE on the first message in a packet.
netlink_dump() will misinterpret this as having finished the listing and
omit data for this interface and all subsequent ones.  That can cause
getifaddrs(3) to enter an infinite loop.

This patch addresses the problem by only supplying IFLA_VF_PORTS when
IFLA_EXT_MASK is supplied with the RTEXT_FILTER_VF flag set.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
---
 net/core/rtnetlink.c | 16 ++++++++++------
 1 file changed, 10 insertions(+), 6 deletions(-)

diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
index 5331db2..32a1287 100644
--- a/net/core/rtnetlink.c
+++ b/net/core/rtnetlink.c
@@ -774,7 +774,8 @@ static inline int rtnl_vfinfo_size(const struct net_device *dev,
 		return 0;
 }
 
-static size_t rtnl_port_size(const struct net_device *dev)
+static size_t rtnl_port_size(const struct net_device *dev,
+			     u32 ext_filter_mask)
 {
 	size_t port_size = nla_total_size(4)		/* PORT_VF */
 		+ nla_total_size(PORT_PROFILE_MAX)	/* PORT_PROFILE */
@@ -790,7 +791,8 @@ static size_t rtnl_port_size(const struct net_device *dev)
 	size_t port_self_size = nla_total_size(sizeof(struct nlattr))
 		+ port_size;
 
-	if (!dev->netdev_ops->ndo_get_vf_port || !dev->dev.parent)
+	if (!dev->netdev_ops->ndo_get_vf_port || !dev->dev.parent
+	    || !(ext_filter_mask & RTEXT_FILTER_VF))
 		return 0;
 	if (dev_num_vf(dev->dev.parent))
 		return port_self_size + vf_ports_size +
@@ -826,7 +828,7 @@ static noinline size_t if_nlmsg_size(const struct net_device *dev,
 	       + nla_total_size(ext_filter_mask
 			        & RTEXT_FILTER_VF ? 4 : 0) /* IFLA_NUM_VF */
 	       + rtnl_vfinfo_size(dev, ext_filter_mask) /* IFLA_VFINFO_LIST */
-	       + rtnl_port_size(dev) /* IFLA_VF_PORTS + IFLA_PORT_SELF */
+	       + rtnl_port_size(dev, ext_filter_mask) /* IFLA_VF_PORTS + IFLA_PORT_SELF */
 	       + rtnl_link_get_size(dev) /* IFLA_LINKINFO */
 	       + rtnl_link_get_af_size(dev) /* IFLA_AF_SPEC */
 	       + nla_total_size(MAX_PHYS_PORT_ID_LEN); /* IFLA_PHYS_PORT_ID */
@@ -888,11 +890,13 @@ static int rtnl_port_self_fill(struct sk_buff *skb, struct net_device *dev)
 	return 0;
 }
 
-static int rtnl_port_fill(struct sk_buff *skb, struct net_device *dev)
+static int rtnl_port_fill(struct sk_buff *skb, struct net_device *dev,
+			  u32 ext_filter_mask)
 {
 	int err;
 
-	if (!dev->netdev_ops->ndo_get_vf_port || !dev->dev.parent)
+	if (!dev->netdev_ops->ndo_get_vf_port || !dev->dev.parent
+	    || !(ext_filter_mask & RTEXT_FILTER_VF))
 		return 0;
 
 	err = rtnl_port_self_fill(skb, dev);
@@ -1079,7 +1083,7 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev,
 		nla_nest_end(skb, vfinfo);
 	}
 
-	if (rtnl_port_fill(skb, dev))
+	if (rtnl_port_fill(skb, dev, ext_filter_mask))
 		goto nla_put_failure;
 
 	if (dev->rtnl_link_ops || rtnl_have_link_slave_info(dev)) {
-- 
1.9.0

Re: [PATCH 2/2] rtnetlink: Only supply IFLA_VF_PORTS information when RTEXT_FILTER_VF is set

[Jiri Pirko]

Wed, Apr 23, 2014 at 09:21:05AM CEST, david@gibson.dropbear.id.au wrote:
>Since 115c9b81928360d769a76c632bae62d15206a94a (rtnetlink: Fix problem with
>buffer allocation), RTM_NEWLINK messages only contain the IFLA_VFINFO_LIST
>attribute if they were solicited by a GETLINK message containing an
>IFLA_EXT_MASK attribute with the RTEXT_FILTER_VF flag.
>
>That was done because some user programs broke when they received more data
>than expected - because IFLA_VFINFO_LIST contains information for each VF
>it can become large if there are many VFs.
>
>However, the IFLA_VF_PORTS attribute, supplied for devices which implement
>ndo_get_vf_port (currently the 'enic' driver only), has the same problem.
>It supplies per-VF information and can therefore become large, but it is
>not currently conditional on the IFLA_EXT_MASK value.
>
>Worse, it interacts badly with the existing EXT_MASK handling.  When
>IFLA_EXT_MASK is not supplied, the buffer for netlink replies is fixed at
>NLMSG_GOODSIZE.  If the information for IFLA_VF_PORTS exceeds this, then
>rtnl_fill_ifinfo() returns -EMSGSIZE on the first message in a packet.
>netlink_dump() will misinterpret this as having finished the listing and
>omit data for this interface and all subsequent ones.  That can cause
>getifaddrs(3) to enter an infinite loop.
>
>This patch addresses the problem by only supplying IFLA_VF_PORTS when
>IFLA_EXT_MASK is supplied with the RTEXT_FILTER_VF flag set.
>
>Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
>---
> net/core/rtnetlink.c | 16 ++++++++++------
> 1 file changed, 10 insertions(+), 6 deletions(-)
>
>diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
>index 5331db2..32a1287 100644
>--- a/net/core/rtnetlink.c
>+++ b/net/core/rtnetlink.c
>@@ -774,7 +774,8 @@ static inline int rtnl_vfinfo_size(const struct net_device *dev,
> 		return 0;
> }
> 
>-static size_t rtnl_port_size(const struct net_device *dev)
>+static size_t rtnl_port_size(const struct net_device *dev,
>+			     u32 ext_filter_mask)
> {
> 	size_t port_size = nla_total_size(4)		/* PORT_VF */
> 		+ nla_total_size(PORT_PROFILE_MAX)	/* PORT_PROFILE */
>@@ -790,7 +791,8 @@ static size_t rtnl_port_size(const struct net_device *dev)
> 	size_t port_self_size = nla_total_size(sizeof(struct nlattr))
> 		+ port_size;
> 
>-	if (!dev->netdev_ops->ndo_get_vf_port || !dev->dev.parent)
>+	if (!dev->netdev_ops->ndo_get_vf_port || !dev->dev.parent
>+	    || !(ext_filter_mask & RTEXT_FILTER_VF))

	Do not start line with ||


> 		return 0;
> 	if (dev_num_vf(dev->dev.parent))
> 		return port_self_size + vf_ports_size +
>@@ -826,7 +828,7 @@ static noinline size_t if_nlmsg_size(const struct net_device *dev,
> 	       + nla_total_size(ext_filter_mask
> 			        & RTEXT_FILTER_VF ? 4 : 0) /* IFLA_NUM_VF */
> 	       + rtnl_vfinfo_size(dev, ext_filter_mask) /* IFLA_VFINFO_LIST */
>-	       + rtnl_port_size(dev) /* IFLA_VF_PORTS + IFLA_PORT_SELF */
>+	       + rtnl_port_size(dev, ext_filter_mask) /* IFLA_VF_PORTS + IFLA_PORT_SELF */
> 	       + rtnl_link_get_size(dev) /* IFLA_LINKINFO */
> 	       + rtnl_link_get_af_size(dev) /* IFLA_AF_SPEC */
> 	       + nla_total_size(MAX_PHYS_PORT_ID_LEN); /* IFLA_PHYS_PORT_ID */
>@@ -888,11 +890,13 @@ static int rtnl_port_self_fill(struct sk_buff *skb, struct net_device *dev)
> 	return 0;
> }
> 
>-static int rtnl_port_fill(struct sk_buff *skb, struct net_device *dev)
>+static int rtnl_port_fill(struct sk_buff *skb, struct net_device *dev,
>+			  u32 ext_filter_mask)
> {
> 	int err;
> 
>-	if (!dev->netdev_ops->ndo_get_vf_port || !dev->dev.parent)
>+	if (!dev->netdev_ops->ndo_get_vf_port || !dev->dev.parent
>+	    || !(ext_filter_mask & RTEXT_FILTER_VF))

	Same here.


> 		return 0;
> 
> 	err = rtnl_port_self_fill(skb, dev);
>@@ -1079,7 +1083,7 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev,
> 		nla_nest_end(skb, vfinfo);
> 	}
> 
>-	if (rtnl_port_fill(skb, dev))
>+	if (rtnl_port_fill(skb, dev, ext_filter_mask))
> 		goto nla_put_failure;
> 
> 	if (dev->rtnl_link_ops || rtnl_have_link_slave_info(dev)) {


Other than this, I think the patch is fine.



>-- 
>1.9.0
>
>--
>To unsubscribe from this list: send the line "unsubscribe netdev" in
>the body of a message to majordomo@vger.kernel.org
>More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 2/2] rtnetlink: Only supply IFLA_VF_PORTS information when RTEXT_FILTER_VF is set

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 2986 bytes --]

On Wed, Apr 23, 2014 at 11:17:15AM +0200, Jiri Pirko wrote:
> Wed, Apr 23, 2014 at 09:21:05AM CEST, david@gibson.dropbear.id.au wrote:
> >Since 115c9b81928360d769a76c632bae62d15206a94a (rtnetlink: Fix problem with
> >buffer allocation), RTM_NEWLINK messages only contain the IFLA_VFINFO_LIST
> >attribute if they were solicited by a GETLINK message containing an
> >IFLA_EXT_MASK attribute with the RTEXT_FILTER_VF flag.
> >
> >That was done because some user programs broke when they received more data
> >than expected - because IFLA_VFINFO_LIST contains information for each VF
> >it can become large if there are many VFs.
> >
> >However, the IFLA_VF_PORTS attribute, supplied for devices which implement
> >ndo_get_vf_port (currently the 'enic' driver only), has the same problem.
> >It supplies per-VF information and can therefore become large, but it is
> >not currently conditional on the IFLA_EXT_MASK value.
> >
> >Worse, it interacts badly with the existing EXT_MASK handling.  When
> >IFLA_EXT_MASK is not supplied, the buffer for netlink replies is fixed at
> >NLMSG_GOODSIZE.  If the information for IFLA_VF_PORTS exceeds this, then
> >rtnl_fill_ifinfo() returns -EMSGSIZE on the first message in a packet.
> >netlink_dump() will misinterpret this as having finished the listing and
> >omit data for this interface and all subsequent ones.  That can cause
> >getifaddrs(3) to enter an infinite loop.
> >
> >This patch addresses the problem by only supplying IFLA_VF_PORTS when
> >IFLA_EXT_MASK is supplied with the RTEXT_FILTER_VF flag set.
> >
> >Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
> >---
> > net/core/rtnetlink.c | 16 ++++++++++------
> > 1 file changed, 10 insertions(+), 6 deletions(-)
> >
> >diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
> >index 5331db2..32a1287 100644
> >--- a/net/core/rtnetlink.c
> >+++ b/net/core/rtnetlink.c
> >@@ -774,7 +774,8 @@ static inline int rtnl_vfinfo_size(const struct net_device *dev,
> > 		return 0;
> > }
> > 
> >-static size_t rtnl_port_size(const struct net_device *dev)
> >+static size_t rtnl_port_size(const struct net_device *dev,
> >+			     u32 ext_filter_mask)
> > {
> > 	size_t port_size = nla_total_size(4)		/* PORT_VF */
> > 		+ nla_total_size(PORT_PROFILE_MAX)	/* PORT_PROFILE */
> >@@ -790,7 +791,8 @@ static size_t rtnl_port_size(const struct net_device *dev)
> > 	size_t port_self_size = nla_total_size(sizeof(struct nlattr))
> > 		+ port_size;
> > 
> >-	if (!dev->netdev_ops->ndo_get_vf_port || !dev->dev.parent)
> >+	if (!dev->netdev_ops->ndo_get_vf_port || !dev->dev.parent
> >+	    || !(ext_filter_mask & RTEXT_FILTER_VF))
> 
> 	Do not start line with ||

Ah, sorry, I'm getting my project codingstyles mixed up.  Will fix.

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: Type: application/pgp-signature, Size: 819 bytes --]

Re: [0/2] Fix problems with with IFLA_VF_PORTS

[Greg Rose]

On Wed, 23 Apr 2014 17:21:03 +1000
David Gibson <david@gibson.dropbear.id.au> wrote:

> I've had a customer encounter a problem with getifaddrs(3) freezing up
> on a system with a Cisco enic device.
> 
> I've discovered that the problem is caused by an enic device with a
> large number of SR-IOV virtual functions overflowing the normal sized
> packet buffer for netlink, leading to interfaces not being reported
> from an RTM_GETLINK request.
> 
> The first patch here just makes the problem easier to locate if it
> occurs again in a different way, by adding a WARN_ON() when we run out
> of room in a netlink packet in this manner.
> 
> The second patch actually fixes the problem, by only reporting
> IFLA_VF_PORTS information when the RTEXT_FILTER_VF flag is specified.
> --
> To unsubscribe from this list: send the line "unsubscribe netdev" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Once you fix a couple of coding style errors mentioned by Jiri then

Acked-by: Greg Rose <gregory.v.rose@intel.com>