* [dpdk-dev] [PATCH] net/bonding:fix balance-xor link down stack overflow
@ 2021-08-10 6:43 jilei
2021-08-10 7:50 ` Min Hu (Connor)
2022-07-18 13:08 ` [dpdk-dev] [PATCH] net/bonding: fix array overflow in Rx burst Yunjian Wang
0 siblings, 2 replies; 6+ messages in thread
From: jilei @ 2021-08-10 6:43 UTC (permalink / raw)
To: dev
Cc: chas3, humin29, zhaohui8, yu.yutao, hejiajun, liucheng11,
yinpeijun, jilei, stable
When slave link down, deactivate_slave will internals->active_slaves
and internals->active_slave_count.Active_slave in bond_ethdev_rx_burst
may out of range in internals->active_slaves.It will get bond's port_id
cause stack overflow
Cc: stable@dpdk.org
Signed-off-by: jilei <jilei8@huawei.com>
---
drivers/net/bonding/rte_eth_bond_pmd.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c b/drivers/net/bonding/rte_eth_bond_pmd.c
index a6755661c4..46f2c42d60 100644
--- a/drivers/net/bonding/rte_eth_bond_pmd.c
+++ b/drivers/net/bonding/rte_eth_bond_pmd.c
@@ -82,7 +82,7 @@ bond_ethdev_rx_burst(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
bufs + num_rx_total, nb_pkts);
num_rx_total += num_rx_slave;
nb_pkts -= num_rx_slave;
- if (++active_slave == slave_count)
+ if (++active_slave >= slave_count)
active_slave = 0;
}
--
2.23.0
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [dpdk-dev] [PATCH] net/bonding:fix balance-xor link down stack overflow
2021-08-10 6:43 [dpdk-dev] [PATCH] net/bonding:fix balance-xor link down stack overflow jilei
@ 2021-08-10 7:50 ` Min Hu (Connor)
2022-05-20 15:33 ` Ferruh Yigit
2022-07-18 13:08 ` [dpdk-dev] [PATCH] net/bonding: fix array overflow in Rx burst Yunjian Wang
1 sibling, 1 reply; 6+ messages in thread
From: Min Hu (Connor) @ 2021-08-10 7:50 UTC (permalink / raw)
To: jilei, dev
Cc: chas3, zhaohui8, yu.yutao, hejiajun, liucheng11, yinpeijun, stable
Hi,
Your patch is OK, but the description is misleading and has
syntax errors. Please fix it ,thanks.
在 2021/8/10 14:43, jilei 写道:
> When slave link down, deactivate_slave will internals->active_slaves
> and internals->active_slave_count.Active_slave in bond_ethdev_rx_burst
> may out of range in internals->active_slaves.It will get bond's port_id
> cause stack overflow
>
> Cc: stable@dpdk.org
> Signed-off-by: jilei <jilei8@huawei.com>
> ---
> drivers/net/bonding/rte_eth_bond_pmd.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c b/drivers/net/bonding/rte_eth_bond_pmd.c
> index a6755661c4..46f2c42d60 100644
> --- a/drivers/net/bonding/rte_eth_bond_pmd.c
> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c
> @@ -82,7 +82,7 @@ bond_ethdev_rx_burst(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
> bufs + num_rx_total, nb_pkts);
> num_rx_total += num_rx_slave;
> nb_pkts -= num_rx_slave;
> - if (++active_slave == slave_count)
> + if (++active_slave >= slave_count)
> active_slave = 0;
> }
>
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [dpdk-dev] [PATCH] net/bonding:fix balance-xor link down stack overflow
2021-08-10 7:50 ` Min Hu (Connor)
@ 2022-05-20 15:33 ` Ferruh Yigit
0 siblings, 0 replies; 6+ messages in thread
From: Ferruh Yigit @ 2022-05-20 15:33 UTC (permalink / raw)
To: Min Hu (Connor), jilei, dev
Cc: chas3, zhaohui8, yu.yutao, hejiajun, liucheng11, yinpeijun, stable
On 8/10/2021 8:50 AM, Min Hu (Connor) wrote:
> 在 2021/8/10 14:43, jilei 写道:
>> When slave link down, deactivate_slave will internals->active_slaves
>> and internals->active_slave_count.Active_slave in bond_ethdev_rx_burst
>> may out of range in internals->active_slaves.It will get bond's port_id
>> cause stack overflow
>>
>> Cc: stable@dpdk.org
>> Signed-off-by: jilei <jilei8@huawei.com>
>> ---
>> drivers/net/bonding/rte_eth_bond_pmd.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c
>> b/drivers/net/bonding/rte_eth_bond_pmd.c
>> index a6755661c4..46f2c42d60 100644
>> --- a/drivers/net/bonding/rte_eth_bond_pmd.c
>> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c
>> @@ -82,7 +82,7 @@ bond_ethdev_rx_burst(void *queue, struct rte_mbuf
>> **bufs, uint16_t nb_pkts)
>> bufs + num_rx_total, nb_pkts);
>> num_rx_total += num_rx_slave;
>> nb_pkts -= num_rx_slave;
>> - if (++active_slave == slave_count)
>> + if (++active_slave >= slave_count)
>> active_slave = 0;
>> }
>>
>
> Hi,
> Your patch is OK, but the description is misleading and has
> syntax errors. Please fix it ,thanks.
>
Hi jilei,
The patch is waiting for a minor update from you, can you please send a
new version with the requested change?
Btw, please add your sign off in following format:
Signed-off-by: Name Surname <email@address>
Thanks,
ferruh
^ permalink raw reply [flat|nested] 6+ messages in thread
* [dpdk-dev] [PATCH] net/bonding: fix array overflow in Rx burst
2021-08-10 6:43 [dpdk-dev] [PATCH] net/bonding:fix balance-xor link down stack overflow jilei
2021-08-10 7:50 ` Min Hu (Connor)
@ 2022-07-18 13:08 ` Yunjian Wang
2022-07-20 1:28 ` 答复: " humin (Q)
1 sibling, 1 reply; 6+ messages in thread
From: Yunjian Wang @ 2022-07-18 13:08 UTC (permalink / raw)
To: dev; +Cc: chas3, humin29, huangshaozhang, jilei8, Yunjian Wang, stable
In bond_ethdev_rx_burst() function, we check the validity of the
'active_slave' as this code:
if (++active_slave == slave_count)
active_slave = 0;
However, the value of 'active_slave' maybe equal to 'slave_count',
when a slave is down. This is wrong and it can cause buffer overflow.
This patch fixes the issue by using '>=' instead of '=='.
Fixes: e1110e977648 ("net/bonding: fix Rx slave fairness")
Cc: stable@dpdk.org
Signed-off-by: Lei Ji <jilei8@huawei.com>
Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
---
drivers/net/bonding/rte_eth_bond_pmd.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c b/drivers/net/bonding/rte_eth_bond_pmd.c
index 73e6972035..6f8a6da108 100644
--- a/drivers/net/bonding/rte_eth_bond_pmd.c
+++ b/drivers/net/bonding/rte_eth_bond_pmd.c
@@ -82,7 +82,7 @@ bond_ethdev_rx_burst(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
bufs + num_rx_total, nb_pkts);
num_rx_total += num_rx_slave;
nb_pkts -= num_rx_slave;
- if (++active_slave == slave_count)
+ if (++active_slave >= slave_count)
active_slave = 0;
}
--
2.27.0
^ permalink raw reply [flat|nested] 6+ messages in thread* 答复: [dpdk-dev] [PATCH] net/bonding: fix array overflow in Rx burst
2022-07-18 13:08 ` [dpdk-dev] [PATCH] net/bonding: fix array overflow in Rx burst Yunjian Wang
@ 2022-07-20 1:28 ` humin (Q)
2022-08-25 16:39 ` Ferruh Yigit
0 siblings, 1 reply; 6+ messages in thread
From: humin (Q) @ 2022-07-20 1:28 UTC (permalink / raw)
To: wangyunjian, dev; +Cc: chas3, Huangshaozhang, jilei (F), stable
Acked-by: Min Hu (Connor) <humin29@huawei.com>
-----邮件原件-----
发件人: wangyunjian <wangyunjian@huawei.com>
发送时间: 2022年7月18日 21:09
收件人: dev@dpdk.org
抄送: chas3@att.com; humin (Q) <humin29@huawei.com>; Huangshaozhang <huangshaozhang@huawei.com>; jilei (F) <jilei8@huawei.com>; wangyunjian <wangyunjian@huawei.com>; stable@dpdk.org
主题: [dpdk-dev] [PATCH] net/bonding: fix array overflow in Rx burst
In bond_ethdev_rx_burst() function, we check the validity of the 'active_slave' as this code:
if (++active_slave == slave_count)
active_slave = 0;
However, the value of 'active_slave' maybe equal to 'slave_count', when a slave is down. This is wrong and it can cause buffer overflow.
This patch fixes the issue by using '>=' instead of '=='.
Fixes: e1110e977648 ("net/bonding: fix Rx slave fairness")
Cc: stable@dpdk.org
Signed-off-by: Lei Ji <jilei8@huawei.com>
Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
---
drivers/net/bonding/rte_eth_bond_pmd.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c b/drivers/net/bonding/rte_eth_bond_pmd.c
index 73e6972035..6f8a6da108 100644
--- a/drivers/net/bonding/rte_eth_bond_pmd.c
+++ b/drivers/net/bonding/rte_eth_bond_pmd.c
@@ -82,7 +82,7 @@ bond_ethdev_rx_burst(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
bufs + num_rx_total, nb_pkts);
num_rx_total += num_rx_slave;
nb_pkts -= num_rx_slave;
- if (++active_slave == slave_count)
+ if (++active_slave >= slave_count)
active_slave = 0;
}
--
2.27.0
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: 答复: [dpdk-dev] [PATCH] net/bonding: fix array overflow in Rx burst
2022-07-20 1:28 ` 答复: " humin (Q)
@ 2022-08-25 16:39 ` Ferruh Yigit
0 siblings, 0 replies; 6+ messages in thread
From: Ferruh Yigit @ 2022-08-25 16:39 UTC (permalink / raw)
To: humin (Q), wangyunjian, dev; +Cc: chas3, Huangshaozhang, jilei (F), stable
On 7/20/2022 2:28 AM, humin (Q) wrote:
> -----邮件原件-----
> 发件人: wangyunjian <wangyunjian@huawei.com>
> 发送时间: 2022年7月18日 21:09
> 收件人: dev@dpdk.org
> 抄送: chas3@att.com; humin (Q) <humin29@huawei.com>; Huangshaozhang <huangshaozhang@huawei.com>; jilei (F) <jilei8@huawei.com>; wangyunjian <wangyunjian@huawei.com>; stable@dpdk.org
> 主题: [dpdk-dev] [PATCH] net/bonding: fix array overflow in Rx burst
>
> In bond_ethdev_rx_burst() function, we check the validity of the 'active_slave' as this code:
> if (++active_slave == slave_count)
> active_slave = 0;
> However, the value of 'active_slave' maybe equal to 'slave_count', when a slave is down. This is wrong and it can cause buffer overflow.
> This patch fixes the issue by using '>=' instead of '=='.
>
> Fixes: e1110e977648 ("net/bonding: fix Rx slave fairness")
> Cc: stable@dpdk.org
>
> Signed-off-by: Lei Ji <jilei8@huawei.com>
> Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
>
> Acked-by: Min Hu (Connor) <humin29@huawei.com>
>
Applied to dpdk-next-net/main, thanks.
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2022-08-25 16:39 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-08-10 6:43 [dpdk-dev] [PATCH] net/bonding:fix balance-xor link down stack overflow jilei
2021-08-10 7:50 ` Min Hu (Connor)
2022-05-20 15:33 ` Ferruh Yigit
2022-07-18 13:08 ` [dpdk-dev] [PATCH] net/bonding: fix array overflow in Rx burst Yunjian Wang
2022-07-20 1:28 ` 答复: " humin (Q)
2022-08-25 16:39 ` Ferruh Yigit
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).