[PATCH v2] net: moxa: fix TX overrun memory leak
Jonas Jensen
jonas.jensen at gmail.com
Mon Mar 27 05:31:19 PDT 2017
moxart_mac_start_xmit() doesn't care where tx_tail is, tx_head can
catch and pass tx_tail, which is bad because moxart_tx_finished()
isn't guaranteed to catch up on freeing resources from tx_tail.
Add a check in moxart_mac_start_xmit() stopping the queue at
the end of the circular buffer. Wake it on completion.
Addresses https://bugzilla.kernel.org/show_bug.cgi?id=99451
Signed-off-by: Jonas Jensen <jonas.jensen at gmail.com>
---
Notes:
ChangeLog v1->v2:
- stop queue instead of dropping frames
The following trick was used to trigger the leak.
On the host (where this driver runs):
1. iptables-restore /etc/iptables.conf && echo 1 > /proc/sys/net/ipv4/ip_forward && ifconfig eth0:0 192.168.5.1
2. cat /dev/zero | nc -l -p 3334
On a client configured with 192.168.5.1 as a gateway:
1. nc -v 192.168.5.1 3334 > /dev/null &
repeat the following multiple times, interrup after a few seconds with CTRL+C:
2. wget http://cdimage.debian.org/debian-cd/current/amd64/iso-cd/debian-8.7.1-amd64-netinst.iso
Result (especially note columns <active_objs> and <num_objs> of kmalloc-192 and kmalloc-2048):
date && ifconfig && echo 1 > /proc/sys/vm/drop_caches && cat /proc/slabinfo
Fri Mar 24 16:22:54 CET 2017
..
RX bytes:57737 (56.3 KiB) TX bytes:6638 (6.4 KiB)
..
# name <active_objs> <num_objs> <objsize> <objperslab> <pagesperslab> : tunables <limit> ..
..
kmalloc-8192 8 8 8192 4 8 : tunables 0 0 0 : slabdata 2 2 0
kmalloc-4096 13 16 4096 8 8 : tunables 0 0 0 : slabdata 2 2 0
kmalloc-2048 40 40 2048 8 4 : tunables 0 0 0 : slabdata 5 5 0
kmalloc-1024 94 96 1024 8 2 : tunables 0 0 0 : slabdata 12 12 0
kmalloc-512 179 184 512 8 1 : tunables 0 0 0 : slabdata 23 23 0
kmalloc-256 76 80 256 16 1 : tunables 0 0 0 : slabdata 5 5 0
kmalloc-192 126 126 192 21 1 : tunables 0 0 0 : slabdata 6 6 0
kmalloc-128 340 416 128 32 1 : tunables 0 0 0 : slabdata 13 13 0
kmalloc-96 8353 8358 96 42 1 : tunables 0 0 0 : slabdata 199 199 0
kmalloc-64 313 320 64 64 1 : tunables 0 0 0 : slabdata 5 5 0
kmalloc-32 1460 1536 32 128 1 : tunables 0 0 0 : slabdata 12 12 0
date && ifconfig && echo 1 > /proc/sys/vm/drop_caches && cat /proc/slabinfo
Fri Mar 24 16:26:36 CET 2017
..
RX bytes:70381213 (67.1 MiB) TX bytes:86208719 (82.2 MiB)
..
# name <active_objs> <num_objs> <objsize> <objperslab> <pagesperslab> : tunables <limit> ..
..
kmalloc-8192 8 8 8192 4 8 : tunables 0 0 0 : slabdata 2 2 0
kmalloc-4096 13 16 4096 8 8 : tunables 0 0 0 : slabdata 2 2 0
kmalloc-2048 2159 2194 2048 8 4 : tunables 0 0 0 : slabdata 275 275 0
kmalloc-1024 100 104 1024 8 2 : tunables 0 0 0 : slabdata 13 13 0
kmalloc-512 182 184 512 8 1 : tunables 0 0 0 : slabdata 23 23 0
kmalloc-256 76 80 256 16 1 : tunables 0 0 0 : slabdata 5 5 0
kmalloc-192 2638 2667 192 21 1 : tunables 0 0 0 : slabdata 127 127 0
kmalloc-128 344 416 128 32 1 : tunables 0 0 0 : slabdata 13 13 0
kmalloc-96 8353 8358 96 42 1 : tunables 0 0 0 : slabdata 199 199 0
kmalloc-64 313 320 64 64 1 : tunables 0 0 0 : slabdata 5 5 0
kmalloc-32 1625 1664 32 128 1 : tunables 0 0 0 : slabdata 13 13 0
Applies to next-20170310
drivers/net/ethernet/moxa/moxart_ether.c | 12 ++++++++++--
1 file changed, 10 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/moxa/moxart_ether.c b/drivers/net/ethernet/moxa/moxart_ether.c
index 06c9f41..fa571d5 100644
--- a/drivers/net/ethernet/moxa/moxart_ether.c
+++ b/drivers/net/ethernet/moxa/moxart_ether.c
@@ -25,6 +25,7 @@
#include <linux/of_irq.h>
#include <linux/crc32.h>
#include <linux/crc32c.h>
+#include <linux/circ_buf.h>
#include "moxart_ether.h"
@@ -297,6 +298,7 @@ static void moxart_tx_finished(struct net_device *ndev)
tx_tail = TX_NEXT(tx_tail);
}
priv->tx_tail = tx_tail;
+ netif_wake_queue(ndev);
}
static irqreturn_t moxart_mac_interrupt(int irq, void *dev_id)
@@ -324,13 +326,19 @@ static int moxart_mac_start_xmit(struct sk_buff *skb, struct net_device *ndev)
struct moxart_mac_priv_t *priv = netdev_priv(ndev);
void *desc;
unsigned int len;
- unsigned int tx_head = priv->tx_head;
+ unsigned int tx_head, tx_tail;
u32 txdes1;
int ret = NETDEV_TX_BUSY;
+ spin_lock_irq(&priv->txlock);
+
+ tx_head = priv->tx_head;
+ tx_tail = priv->tx_tail;
desc = priv->tx_desc_base + (TX_REG_DESC_SIZE * tx_head);
- spin_lock_irq(&priv->txlock);
+ if (CIRC_SPACE(tx_head, tx_tail, TX_DESC_NUM) == 1)
+ netif_stop_queue(ndev);
+
if (moxart_desc_read(desc + TX_REG_OFFSET_DESC0) & TX_DESC0_DMA_OWN) {
net_dbg_ratelimited("no TX space for packet\n");
priv->stats.tx_dropped++;
--
1.8.2.1
More information about the linux-arm-kernel
mailing list