[PATCH] bonding: fix race condition in bonding_store_slaves_active

Subsystems: bonding driver, networking drivers, the rest

STALE5011d

5 messages, 3 authors, 2012-11-29 · open the first message on its own page

[PATCH] bonding: fix race condition in bonding_store_slaves_active

From: Nikolay Aleksandrov <hidden>
Date: 2012-11-24 12:22:00

 Race between bonding_store_slaves_active() and slave manipulation functions
 The bond_for_each_slave use in bonding_store_slaves_active() is not protected
 by any synchronization mechanism. NULL pointer dereference is easy to reach.
 Fixed by acquiring the bond->lock for the slave walk.

Signed-off-by: Nikolay Aleksandrov <redacted>
---
 drivers/net/bonding/bond_sysfs.c | 2 ++
 1 file changed, 2 insertions(+)
diff --git a/drivers/net/bonding/bond_sysfs.c b/drivers/net/bonding/bond_sysfs.c
index ef8d2a0..ba4f95b 100644
--- a/drivers/net/bonding/bond_sysfs.c
+++ b/drivers/net/bonding/bond_sysfs.c
@@ -1582,6 +1582,7 @@ static ssize_t bonding_store_slaves_active(struct device *d,
 		goto out;
 	}
 
+	read_lock(&bond->lock);
 	bond_for_each_slave(bond, slave, i) {
 		if (!bond_is_active_slave(slave)) {
 			if (new_value)
@@ -1590,6 +1591,7 @@ static ssize_t bonding_store_slaves_active(struct device *d,
 				slave->inactive = 1;
 		}
 	}
+	read_unlock(&bond->lock);
 out:
 	return ret;
 }
-- 
1.7.11.7

Re: [PATCH] bonding: fix race condition in bonding_store_slaves_active

From: David Miller <davem@davemloft.net>
Date: 2012-11-28 16:21:37

Some bonding developer reviews would be appreciated.

[PATCH v2] bonding: fix race condition in bonding_store_slaves_active

From: Nikolay Aleksandrov <hidden>
Date: 2012-11-29 12:18:02

 Race between bonding_store_slaves_active() and slave manipulation 
 functions. The bond_for_each_slave use in bonding_store_slaves_active()
 is not protected by any synchronization mechanism.
 NULL pointer dereference is easy to reach.
 Fixed by acquiring the bond->lock for the slave walk.

 v2: Make description text < 75 columns

Signed-off-by: Nikolay Aleksandrov <redacted>
---
 drivers/net/bonding/bond_sysfs.c | 2 ++
 1 file changed, 2 insertions(+)
diff --git a/drivers/net/bonding/bond_sysfs.c b/drivers/net/bonding/bond_sysfs.c
index ef8d2a0..ba4f95b 100644
--- a/drivers/net/bonding/bond_sysfs.c
+++ b/drivers/net/bonding/bond_sysfs.c
@@ -1582,6 +1582,7 @@ static ssize_t bonding_store_slaves_active(struct device *d,
 		goto out;
 	}
 
+	read_lock(&bond->lock);
 	bond_for_each_slave(bond, slave, i) {
 		if (!bond_is_active_slave(slave)) {
 			if (new_value)
@@ -1590,6 +1591,7 @@ static ssize_t bonding_store_slaves_active(struct device *d,
 				slave->inactive = 1;
 		}
 	}
+	read_unlock(&bond->lock);
 out:
 	return ret;
 }
-- 
1.7.11.7

Re: [PATCH v2] bonding: fix race condition in bonding_store_slaves_active

From: Jay Vosburgh <hidden>
Date: 2012-11-29 17:38:22

Nikolay Aleksandrov [off-list ref] wrote:
Race between bonding_store_slaves_active() and slave manipulation 
functions. The bond_for_each_slave use in bonding_store_slaves_active()
is not protected by any synchronization mechanism.
NULL pointer dereference is easy to reach.
Fixed by acquiring the bond->lock for the slave walk.

v2: Make description text < 75 columns

Signed-off-by: Nikolay Aleksandrov <redacted>
Signed-off-by: Jay Vosburgh <redacted>
quoted hunk
---
drivers/net/bonding/bond_sysfs.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/bonding/bond_sysfs.c b/drivers/net/bonding/bond_sysfs.c
index ef8d2a0..ba4f95b 100644
--- a/drivers/net/bonding/bond_sysfs.c
+++ b/drivers/net/bonding/bond_sysfs.c
@@ -1582,6 +1582,7 @@ static ssize_t bonding_store_slaves_active(struct device *d,
		goto out;
	}

+	read_lock(&bond->lock);
	bond_for_each_slave(bond, slave, i) {
		if (!bond_is_active_slave(slave)) {
			if (new_value)
@@ -1590,6 +1591,7 @@ static ssize_t bonding_store_slaves_active(struct device *d,
				slave->inactive = 1;
		}
	}
+	read_unlock(&bond->lock);
out:
	return ret;
}
-- 
1.7.11.7

Re: [PATCH v2] bonding: fix race condition in bonding_store_slaves_active

From: David Miller <davem@davemloft.net>
Date: 2012-11-29 18:14:11

From: Jay Vosburgh <redacted>
Date: Thu, 29 Nov 2012 09:37:56 -0800
Nikolay Aleksandrov [off-list ref] wrote:
quoted
Race between bonding_store_slaves_active() and slave manipulation 
functions. The bond_for_each_slave use in bonding_store_slaves_active()
is not protected by any synchronization mechanism.
NULL pointer dereference is easy to reach.
Fixed by acquiring the bond->lock for the slave walk.

v2: Make description text < 75 columns

Signed-off-by: Nikolay Aleksandrov <redacted>
Signed-off-by: Jay Vosburgh <redacted>
Applied.
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help