lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Mon,  7 Dec 2020 01:59:15 +0200
From:   Vladimir Oltean <vladimir.oltean@....com>
To:     "David S . Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>
Cc:     netdev@...r.kernel.org, Andrew Lunn <andrew@...n.ch>,
        Florian Fainelli <f.fainelli@...il.com>,
        Paul Gortmaker <paul.gortmaker@...driver.com>,
        Pablo Neira Ayuso <pablo@...filter.org>,
        Jiri Benc <jbenc@...hat.com>,
        Cong Wang <xiyou.wangcong@...il.com>,
        Jamal Hadi Salim <jhs@...atatu.com>,
        Stephen Hemminger <stephen@...workplumber.org>,
        Eric Dumazet <edumazet@...gle.com>,
        George McCollister <george.mccollister@...il.com>,
        Oleksij Rempel <o.rempel@...gutronix.de>,
        "James E.J. Bottomley" <James.Bottomley@...senPartnership.com>,
        Helge Deller <deller@....de>, linux-parisc@...r.kernel.org
Subject: [RFC PATCH net-next 09/13] parisc/led: hold the netdev lists lock when retrieving device statistics

In the effort of making .ndo_get_stats64 be able to sleep, we need to
ensure the callers of dev_get_stats do not use atomic context.

The LED driver for HP-PARISC workstations uses a workqueue to
periodically check for updates in network interface statistics, and
flicker when those have changed (i.e. there has been activity on the
line). Ignoring the fact that this driver is completely duplicating
drivers/leds/trigger/ledtrig-netdev.c, there is an even bigger problem.
Now, the dev_get_stats call can sleep, and iterating through the list of
network interfaces still needs to ensure the integrity of list of
network interfaces. So that leaves us only one locking option given the
current design of the network stack, and that is the netns mutex.

Cc: "James E.J. Bottomley" <James.Bottomley@...senPartnership.com>
Cc: Helge Deller <deller@....de>
Cc: linux-parisc@...r.kernel.org
Signed-off-by: Vladimir Oltean <vladimir.oltean@....com>
---
 drivers/parisc/led.c | 21 ++++++++++++---------
 1 file changed, 12 insertions(+), 9 deletions(-)

diff --git a/drivers/parisc/led.c b/drivers/parisc/led.c
index b7005aaa782b..1289dd3ea6e4 100644
--- a/drivers/parisc/led.c
+++ b/drivers/parisc/led.c
@@ -38,7 +38,6 @@
 #include <linux/ctype.h>
 #include <linux/blkdev.h>
 #include <linux/workqueue.h>
-#include <linux/rcupdate.h>
 #include <asm/io.h>
 #include <asm/processor.h>
 #include <asm/hardware.h>
@@ -356,24 +355,28 @@ static __inline__ int led_get_net_activity(void)
 
 	rx_total = tx_total = 0;
 
-	/* we are running as a workqueue task, so we can use an RCU lookup */
-	rcu_read_lock();
-	for_each_netdev_rcu(&init_net, dev) {
+	/* we are running as a workqueue task, so we can sleep */
+	mutex_lock(&init_net->netdev_lists_lock);
+
+	for_each_netdev(&init_net, dev) {
+		struct in_device *in_dev = in_dev_get(dev);
 		const struct rtnl_link_stats64 *stats;
 		struct rtnl_link_stats64 temp;
-		struct in_device *in_dev = __in_dev_get_rcu(dev);
 
-		if (!in_dev || !in_dev->ifa_list)
+		if (!in_dev || !in_dev->ifa_list ||
+		    ipv4_is_loopback(in_dev->ifa_list->ifa_local)) {
+			in_dev_put(in_dev);
 			continue;
+		}
 
-		if (ipv4_is_loopback(in_dev->ifa_list->ifa_local))
-			continue;
+		in_dev_put(in_dev);
 
 		stats = dev_get_stats(dev, &temp);
 		rx_total += stats->rx_packets;
 		tx_total += stats->tx_packets;
 	}
-	rcu_read_unlock();
+
+	mutex_unlock(&init_net->netdev_lists_lock);
 
 	retval = 0;
 
-- 
2.25.1

Powered by blists - more mailing lists