[PATCH] net: cgroup: fix out of bounds accesses

From: Eric Dumazet <edumazet@google.com>

dev->priomap is allocated by extend_netdev_table() called from
update_netdev_tables().
And this is only called if write_priomap() is called.

But if write_priomap() is not called, it seems we can have out of bounds
accesses in cgrp_destroy(), read_priomap() & skb_update_prio()

With help from Gao Feng

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neil Horman <nhorman@tuxdriver.com>
Cc: Gao feng <redacted>
---
net/core/dev.c            |    8 ++++++--
net/core/netprio_cgroup.c |    4 ++--
2 files changed, 8 insertions(+), 4 deletions(-)

diff --git a/net/core/dev.c b/net/core/dev.c
index 84f01ba..0f28a9e 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c

@@ -2444,8 +2444,12 @@ static void skb_update_prio(struct sk_buff *skb)
 {
 	struct netprio_map *map = rcu_dereference_bh(skb->dev->priomap);
 
-	if ((!skb->priority) && (skb->sk) && map)
-		skb->priority = map->priomap[skb->sk->sk_cgrp_prioidx];
+	if (!skb->priority && skb->sk && map) {
+		unsigned int prioidx = skb->sk->sk_cgrp_prioidx;
+
+		if (prioidx < map->priomap_len)
+			skb->priority = map->priomap[prioidx];
+	}
 }
 #else
 #define skb_update_prio(skb)

diff --git a/net/core/netprio_cgroup.c b/net/core/netprio_cgroup.c
index aa907ed..3e953ea 100644
--- a/net/core/netprio_cgroup.c
+++ b/net/core/netprio_cgroup.c

@@ -142,7 +142,7 @@ static void cgrp_destroy(struct cgroup *cgrp)
 	rtnl_lock();
 	for_each_netdev(&init_net, dev) {
 		map = rtnl_dereference(dev->priomap);
-		if (map)
+		if (map && cs->prioidx < map->priomap_len)
 			map->priomap[cs->prioidx] = 0;
 	}
 	rtnl_unlock();

@@ -166,7 +166,7 @@ static int read_priomap(struct cgroup *cont, struct cftype *cft,
 	rcu_read_lock();
 	for_each_netdev_rcu(&init_net, dev) {
 		map = rcu_dereference(dev->priomap);
-		priority = map ? map->priomap[prioidx] : 0;
+		priority = (map && prioidx < map->priomap_len) ? map->priomap[prioidx] : 0;
 		cb->fill(cb, dev->name, priority);
 	}
 	rcu_read_unlock();

于 2012年07月09日 15:45, Eric Dumazet 写道:

From: Eric Dumazet <edumazet@google.com>

dev->priomap is allocated by extend_netdev_table() called from
update_netdev_tables().
And this is only called if write_priomap() is called.

But if write_priomap() is not called, it seems we can have out of bounds
accesses in cgrp_destroy(), read_priomap() & skb_update_prio()

With help from Gao Feng

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neil Horman <nhorman@tuxdriver.com>
Cc: Gao feng <redacted>
---
net/core/dev.c            |    8 ++++++--
net/core/netprio_cgroup.c |    4 ++--
2 files changed, 8 insertions(+), 4 deletions(-)

Acked-by: Gao feng <redacted>

On Mon, Jul 09, 2012 at 09:45:10AM +0200, Eric Dumazet wrote:

From: Eric Dumazet <edumazet@google.com>

dev->priomap is allocated by extend_netdev_table() called from
update_netdev_tables().
And this is only called if write_priomap() is called.

But if write_priomap() is not called, it seems we can have out of bounds
accesses in cgrp_destroy(), read_priomap() & skb_update_prio()

With help from Gao Feng

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neil Horman <nhorman@tuxdriver.com>
Cc: Gao feng <redacted>
---
net/core/dev.c            |    8 ++++++--
net/core/netprio_cgroup.c |    4 ++--
2 files changed, 8 insertions(+), 4 deletions(-)

Thank you for doing this Eric, Gao.  Just to be sure (I asked in the previous
thread), would it be better to avoid the length check in skb_update_prio, and
instead update the netdev tables to be long enough in cgrp_create and in
netprio_device_event on device registration?

Neil

On Mon, 2012-07-09 at 07:01 -0400, Neil Horman wrote:

Thank you for doing this Eric, Gao.  Just to be sure (I asked in the previous
thread), would it be better to avoid the length check in skb_update_prio, and
instead update the netdev tables to be long enough in cgrp_create and in
netprio_device_event on device registration?

Yes probably, and it is even needed because extend_netdev_table() can
acutally fail to expand the table if kzalloc() returned NULL.

Current code just ignores this allocation failure so we also can crash
in write_priomap()

On Mon, Jul 09, 2012 at 01:50:52PM +0200, Eric Dumazet wrote:

On Mon, 2012-07-09 at 07:01 -0400, Neil Horman wrote:

quoted

Thank you for doing this Eric, Gao.  Just to be sure (I asked in the previous
thread), would it be better to avoid the length check in skb_update_prio, and
instead update the netdev tables to be long enough in cgrp_create and in
netprio_device_event on device registration?

Yes probably, and it is even needed because extend_netdev_table() can
acutally fail to expand the table if kzalloc() returned NULL.

Current code just ignores this allocation failure so we also can crash
in write_priomap()

ACK, can you follow up with a patch please?

Thanks!
Neil

On Mon, 2012-07-09 at 08:13 -0400, Neil Horman wrote:

On Mon, Jul 09, 2012 at 01:50:52PM +0200, Eric Dumazet wrote:

quoted

On Mon, 2012-07-09 at 07:01 -0400, Neil Horman wrote:

quoted

Thank you for doing this Eric, Gao.  Just to be sure (I asked in the previous
thread), would it be better to avoid the length check in skb_update_prio, and
instead update the netdev tables to be long enough in cgrp_create and in
netprio_device_event on device registration?

Yes probably, and it is even needed because extend_netdev_table() can
acutally fail to expand the table if kzalloc() returned NULL.

Current code just ignores this allocation failure so we also can crash
in write_priomap()

ACK, can you follow up with a patch please?

Gao was working on this allocation problem (he privately sent me a v1 of
his patch), so I think we can wait Gao submit a v2 to combine all the
work/ideas in a single patch.

(ie make sure we dont need additional bound checkings in fast path)

On Mon, Jul 09, 2012 at 02:40:25PM +0200, Eric Dumazet wrote:

On Mon, 2012-07-09 at 08:13 -0400, Neil Horman wrote:

quoted

On Mon, Jul 09, 2012 at 01:50:52PM +0200, Eric Dumazet wrote:

quoted

On Mon, 2012-07-09 at 07:01 -0400, Neil Horman wrote:

quoted

Thank you for doing this Eric, Gao.  Just to be sure (I asked in the previous
thread), would it be better to avoid the length check in skb_update_prio, and
instead update the netdev tables to be long enough in cgrp_create and in
netprio_device_event on device registration?

Yes probably, and it is even needed because extend_netdev_table() can
acutally fail to expand the table if kzalloc() returned NULL.

Current code just ignores this allocation failure so we also can crash
in write_priomap()

ACK, can you follow up with a patch please?

Gao was working on this allocation problem (he privately sent me a v1 of
his patch), so I think we can wait Gao submit a v2 to combine all the
work/ideas in a single patch.

(ie make sure we dont need additional bound checkings in fast path)

Ok, I agree.  thanks!
Neil

From: Gao feng <redacted>
Date: Mon, 09 Jul 2012 16:15:29 +0800

于 2012年07月09日 15:45, Eric Dumazet 写道:

quoted

From: Eric Dumazet <edumazet@google.com>

dev->priomap is allocated by extend_netdev_table() called from
update_netdev_tables().
And this is only called if write_priomap() is called.

But if write_priomap() is not called, it seems we can have out of bounds
accesses in cgrp_destroy(), read_priomap() & skb_update_prio()

With help from Gao Feng

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neil Horman <nhorman@tuxdriver.com>
Cc: Gao feng <redacted>
---
net/core/dev.c            |    8 ++++++--
net/core/netprio_cgroup.c |    4 ++--
2 files changed, 8 insertions(+), 4 deletions(-)

Acked-by: Gao feng <redacted>

Applied.

there are some out of bound accesses in netprio cgroup.
when creating a new netprio cgroup,we only set a prioidx for
the new cgroup,without allocate memory for dev->priomap.

because we don't want to see additional bound checkings in
fast path, so I think the best way is to allocate memory when we
creating a new netprio cgroup.

and because netdev can be created or registered after cgroup being
created, so extend_netdev_table is also needed in write_priomap.

this patch add a return value for update_netdev_tables & extend_netdev_table,
so when new_priomap is allocated failed,write_priomap will stop to access
the priomap,and return -ENOMEM back to the userspace to tell the user
what happend.

Signed-off-by: Gao feng <redacted>
Cc: Neil Horman <nhorman@tuxdriver.com>
Cc: Eric Dumazet <edumazet@google.com>
---
 net/core/netprio_cgroup.c |   43 +++++++++++++++++++++++++++++--------------
 1 files changed, 29 insertions(+), 14 deletions(-)

diff --git a/net/core/netprio_cgroup.c b/net/core/netprio_cgroup.c
index aa907ed..3554f28 100644
--- a/net/core/netprio_cgroup.c
+++ b/net/core/netprio_cgroup.c

@@ -65,7 +65,7 @@ static void put_prioidx(u32 idx)
 	spin_unlock_irqrestore(&prioidx_map_lock, flags);
 }
 
-static void extend_netdev_table(struct net_device *dev, u32 new_len)
+static int extend_netdev_table(struct net_device *dev, u32 new_len)
 {
 	size_t new_size = sizeof(struct netprio_map) +
 			   ((sizeof(u32) * new_len));

@@ -77,7 +77,7 @@ static void extend_netdev_table(struct net_device *dev, u32 new_len)
 
 	if (!new_priomap) {
 		pr_warn("Unable to alloc new priomap!\n");
-		return;
+		return -ENOMEM;
 	}
 
 	for (i = 0;

@@ -90,10 +90,12 @@ static void extend_netdev_table(struct net_device *dev, u32 new_len)
 	rcu_assign_pointer(dev->priomap, new_priomap);
 	if (old_priomap)
 		kfree_rcu(old_priomap, rcu);
+	return 0;
 }
 
-static void update_netdev_tables(void)
+static int update_netdev_tables(void)
 {
+	int ret = 0;
 	struct net_device *dev;
 	u32 max_len = atomic_read(&max_prioidx) + 1;
 	struct netprio_map *map;

@@ -102,34 +104,44 @@ static void update_netdev_tables(void)
 	for_each_netdev(&init_net, dev) {
 		map = rtnl_dereference(dev->priomap);
 		if ((!map) ||
-		    (map->priomap_len < max_len))
-			extend_netdev_table(dev, max_len);
+		    (map->priomap_len < max_len)) {
+			ret = extend_netdev_table(dev, max_len);
+			if (ret < 0)
+				break;
+		}
 	}
 	rtnl_unlock();
+	return ret;
 }
 
 static struct cgroup_subsys_state *cgrp_create(struct cgroup *cgrp)
 {
 	struct cgroup_netprio_state *cs;
-	int ret;
+	int ret = -EINVAL;
 
 	cs = kzalloc(sizeof(*cs), GFP_KERNEL);
 	if (!cs)
 		return ERR_PTR(-ENOMEM);
 
-	if (cgrp->parent && cgrp_netprio_state(cgrp->parent)->prioidx) {
-		kfree(cs);
-		return ERR_PTR(-EINVAL);
-	}
+	if (cgrp->parent && cgrp_netprio_state(cgrp->parent)->prioidx)
+		goto out;
 
 	ret = get_prioidx(&cs->prioidx);
-	if (ret != 0) {
+	if (ret < 0) {
 		pr_warn("No space in priority index array\n");
-		kfree(cs);
-		return ERR_PTR(ret);
+		goto out;
+	}
+
+	ret = update_netdev_tables();
+	if (ret < 0) {
+		put_prioidx(cs->prioidx);
+		goto out;
 	}
 
 	return &cs->css;
+out:
+	kfree(cs);
+	return ERR_PTR(ret);
 }
 
 static void cgrp_destroy(struct cgroup *cgrp)

@@ -221,7 +233,10 @@ static int write_priomap(struct cgroup *cgrp, struct cftype *cft,
 	if (!dev)
 		goto out_free_devname;
 
-	update_netdev_tables();
+	ret = update_netdev_tables();
+	if (ret < 0)
+		goto out_free_devname;
+
 	ret = 0;
 	rcu_read_lock();
 	map = rcu_dereference(dev->priomap);

-- 
1.7.7.6

于 2012年07月10日 05:51, David Miller 写道:

From: Gao feng <redacted>
Date: Mon, 09 Jul 2012 16:15:29 +0800

quoted

于 2012年07月09日 15:45, Eric Dumazet 写道:

quoted

From: Eric Dumazet <edumazet@google.com>

dev->priomap is allocated by extend_netdev_table() called from
update_netdev_tables().
And this is only called if write_priomap() is called.

But if write_priomap() is not called, it seems we can have out of bounds
accesses in cgrp_destroy(), read_priomap() & skb_update_prio()

With help from Gao Feng

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neil Horman <nhorman@tuxdriver.com>
Cc: Gao feng <redacted>
---
net/core/dev.c            |    8 ++++++--
net/core/netprio_cgroup.c |    4 ++--
2 files changed, 8 insertions(+), 4 deletions(-)

Acked-by: Gao feng <redacted>

Applied.

Hi David

Please see my patch in this thread, I think it's a better way to fix this bug.

Thanks.

From: Gao feng <redacted>
Date: Tue, 10 Jul 2012 10:33:23 +0800

Please see my patch in this thread, I think it's a better way to fix this bug.

You'll need to work that out with Eric, fwiw I think his patch was
clean and just fine and it's staying in my tree.

On Tue, 2012-07-10 at 10:31 +0800, Gao feng wrote:

there are some out of bound accesses in netprio cgroup.
when creating a new netprio cgroup,we only set a prioidx for
the new cgroup,without allocate memory for dev->priomap.

because we don't want to see additional bound checkings in
fast path, so I think the best way is to allocate memory when we
creating a new netprio cgroup.

and because netdev can be created or registered after cgroup being
created, so extend_netdev_table is also needed in write_priomap.

this patch add a return value for update_netdev_tables & extend_netdev_table,
so when new_priomap is allocated failed,write_priomap will stop to access
the priomap,and return -ENOMEM back to the userspace to tell the user
what happend.

Signed-off-by: Gao feng <redacted>
Cc: Neil Horman <nhorman@tuxdriver.com>
Cc: Eric Dumazet <edumazet@google.com>
---

Hi Gao

Is it still needed to call update_netdev_tables() from write_priomap() ?

Hi Gao

Is it still needed to call update_netdev_tables() from write_priomap() ?

Yes, I think it's needed,because read_priomap will show all of the net devices,

But we may add the netdev after create a netprio cgroup, so the new added netdev's
priomap will not be allocated. if we don't call update_netdev_tables in write_priomap,
we may access this unallocated memory.

On Tue, 2012-07-10 at 16:53 +0800, Gao feng wrote:

quoted

Hi Gao

Is it still needed to call update_netdev_tables() from write_priomap() ?

Yes, I think it's needed,because read_priomap will show all of the net devices,

But we may add the netdev after create a netprio cgroup, so the new added netdev's
priomap will not be allocated. if we don't call update_netdev_tables in write_priomap,
we may access this unallocated memory.

I realize my question was not clear.

If we write in write_priomap() a field of a single netdevice,
why should we allocate memory for all netdevices on the machine ?

So the question was : Do we really need to call
update_netdev_tables(alldevs), instead of extend_netdev_table(dev)

于 2012年07月10日 17:15, Eric Dumazet 写道:

On Tue, 2012-07-10 at 16:53 +0800, Gao feng wrote:

quoted

quoted

Hi Gao

Is it still needed to call update_netdev_tables() from write_priomap() ?

Yes, I think it's needed,because read_priomap will show all of the net devices,

But we may add the netdev after create a netprio cgroup, so the new added netdev's
priomap will not be allocated. if we don't call update_netdev_tables in write_priomap,
we may access this unallocated memory.

I realize my question was not clear.

If we write in write_priomap() a field of a single netdevice,
why should we allocate memory for all netdevices on the machine ?

So the question was : Do we really need to call
update_netdev_tables(alldevs), instead of extend_netdev_table(dev)

I get it.

You are right,Indeed we only need to call extend_netdev_table
for the netdev witch we want to change.

and I read the commit f5c38208d32412d72b97a4f0d44af0eb39feb20b,
found why we need delay allocation.

I will send a v2 patch.

Thanks!