* [dpdk-dev] [PATCH] net/iavf: fix mbuf leak
@ 2021-09-10 8:23 Qiming Chen
2021-09-10 8:31 ` [dpdk-dev] [PATCH v2] " Qiming Chen
0 siblings, 1 reply; 7+ messages in thread
From: Qiming Chen @ 2021-09-10 8:23 UTC (permalink / raw)
To: dev; +Cc: beilei.xing, jingjing.wu, Qiming Chen, stable
In the iavf_dev_rx_queue_start function, if the iavf_switch_queue
or iavf_switch_queue_lv function fails, the previously applied mbuf
is not released, resulting in leakage. The patch fixes the problem.
Fixes: 9cf9c02bf6ee ("net/iavf: add enable/disable queues for large VF")
Cc: stable@dpdk.org
---
drivers/net/iavf/iavf_rxtx.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/drivers/net/iavf/iavf_rxtx.c b/drivers/net/iavf/iavf_rxtx.c
index e33fe4576b..55393a9400 100644
--- a/drivers/net/iavf/iavf_rxtx.c
+++ b/drivers/net/iavf/iavf_rxtx.c
@@ -848,12 +848,14 @@ iavf_dev_rx_queue_start(struct rte_eth_dev *dev, uint16_t rx_queue_id)
else
err = iavf_switch_queue_lv(adapter, rx_queue_id, true, true);
- if (err)
+ if (err) {
+ release_rxq_mbufs(rxq);
PMD_DRV_LOG(ERR, "Failed to switch RX queue %u on",
rx_queue_id);
- else
+ } else {
dev->data->rx_queue_state[rx_queue_id] =
RTE_ETH_QUEUE_STATE_STARTED;
+ }
return err;
}
--
2.30.1.windows.1
^ permalink raw reply [flat|nested] 7+ messages in thread* [dpdk-dev] [PATCH v2] net/iavf: fix mbuf leak
2021-09-10 8:23 [dpdk-dev] [PATCH] net/iavf: fix mbuf leak Qiming Chen
@ 2021-09-10 8:31 ` Qiming Chen
2021-09-15 3:27 ` Zhang, Qi Z
2021-09-16 14:04 ` Qiming Chen
0 siblings, 2 replies; 7+ messages in thread
From: Qiming Chen @ 2021-09-10 8:31 UTC (permalink / raw)
To: dev; +Cc: beilei.xing, jingjing.wu, Qiming Chen, stable
In the iavf_dev_rx_queue_start function, if the iavf_switch_queue
or iavf_switch_queue_lv function fails, the previously applied mbuf
is not released, resulting in leakage. The patch fixes the problem.
Fixes: 9cf9c02bf6ee ("net/iavf: add enable/disable queues for large VF")
Cc: stable@dpdk.org
Signed-off-by: Qiming Chen <chenqiming_huawei@163.com>
---
v2:
Modify coding style warning.
---
drivers/net/iavf/iavf_rxtx.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/drivers/net/iavf/iavf_rxtx.c b/drivers/net/iavf/iavf_rxtx.c
index e33fe4576b..55393a9400 100644
--- a/drivers/net/iavf/iavf_rxtx.c
+++ b/drivers/net/iavf/iavf_rxtx.c
@@ -848,12 +848,14 @@ iavf_dev_rx_queue_start(struct rte_eth_dev *dev, uint16_t rx_queue_id)
else
err = iavf_switch_queue_lv(adapter, rx_queue_id, true, true);
- if (err)
+ if (err) {
+ release_rxq_mbufs(rxq);
PMD_DRV_LOG(ERR, "Failed to switch RX queue %u on",
rx_queue_id);
- else
+ } else {
dev->data->rx_queue_state[rx_queue_id] =
RTE_ETH_QUEUE_STATE_STARTED;
+ }
return err;
}
--
2.30.1.windows.1
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [dpdk-dev] [PATCH v2] net/iavf: fix mbuf leak
2021-09-10 8:31 ` [dpdk-dev] [PATCH v2] " Qiming Chen
@ 2021-09-15 3:27 ` Zhang, Qi Z
2021-09-16 14:04 ` Qiming Chen
1 sibling, 0 replies; 7+ messages in thread
From: Zhang, Qi Z @ 2021-09-15 3:27 UTC (permalink / raw)
To: Qiming Chen, dev; +Cc: Xing, Beilei, Wu, Jingjing, stable
> -----Original Message-----
> From: dev <dev-bounces@dpdk.org> On Behalf Of Qiming Chen
> Sent: Friday, September 10, 2021 4:32 PM
> To: dev@dpdk.org
> Cc: Xing, Beilei <beilei.xing@intel.com>; Wu, Jingjing <jingjing.wu@intel.com>;
> Qiming Chen <chenqiming_huawei@163.com>; stable@dpdk.org
> Subject: [dpdk-dev] [PATCH v2] net/iavf: fix mbuf leak
>
> In the iavf_dev_rx_queue_start function, if the iavf_switch_queue or
> iavf_switch_queue_lv function fails, the previously applied mbuf is not
> released, resulting in leakage. The patch fixes the problem.
>
> Fixes: 9cf9c02bf6ee ("net/iavf: add enable/disable queues for large VF")
> Cc: stable@dpdk.org
>
> Signed-off-by: Qiming Chen <chenqiming_huawei@163.com>
> ---
> v2:
> Modify coding style warning.
> ---
> drivers/net/iavf/iavf_rxtx.c | 6 ++++--
> 1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/net/iavf/iavf_rxtx.c b/drivers/net/iavf/iavf_rxtx.c index
> e33fe4576b..55393a9400 100644
> --- a/drivers/net/iavf/iavf_rxtx.c
> +++ b/drivers/net/iavf/iavf_rxtx.c
> @@ -848,12 +848,14 @@ iavf_dev_rx_queue_start(struct rte_eth_dev *dev,
> uint16_t rx_queue_id)
> else
> err = iavf_switch_queue_lv(adapter, rx_queue_id, true, true);
>
> - if (err)
> + if (err) {
> + release_rxq_mbufs(rxq);
This looks good, but I saw another potential memory leak in " alloc_rxq_mbufs",
It is the case when only part of the memory are available, the already allocated mbuf not be released before return -ENOMEM
for (i = 0; i < rxq->nb_rx_desc; i++) {
mbuf = rte_mbuf_raw_alloc(rxq->mp);
if (unlikely(!mbuf)) {
PMD_DRV_LOG(ERR, "Failed to allocate mbuf for RX");
return -ENOMEM;
}
Do you mind to merge them into the same patch? Thanks.
> PMD_DRV_LOG(ERR, "Failed to switch RX queue %u on",
> rx_queue_id);
> - else
> + } else {
> dev->data->rx_queue_state[rx_queue_id] =
> RTE_ETH_QUEUE_STATE_STARTED;
> + }
>
> return err;
> }
> --
> 2.30.1.windows.1
^ permalink raw reply [flat|nested] 7+ messages in thread* [dpdk-dev] [PATCH v2] net/iavf: fix mbuf leak
2021-09-10 8:31 ` [dpdk-dev] [PATCH v2] " Qiming Chen
2021-09-15 3:27 ` Zhang, Qi Z
@ 2021-09-16 14:04 ` Qiming Chen
2021-09-22 7:04 ` Zhang, Qi Z
1 sibling, 1 reply; 7+ messages in thread
From: Qiming Chen @ 2021-09-16 14:04 UTC (permalink / raw)
To: dev; +Cc: beilei.xing, jingjing.wu, Qiming Chen, stable
In the iavf_dev_rx_queue_start function, if the iavf_switch_queue
or iavf_switch_queue_lv function fails, the previously applied mbuf
is not released, resulting in leakage. The patch fixes the problem.
Fixes: 9cf9c02bf6ee ("net/iavf: add enable/disable queues for large VF")
Cc: stable@dpdk.org
Signed-off-by: Qiming Chen <chenqiming_huawei@163.com>
---
v2:
Solve the potential mbuf leak of the alloc_rxq_mbufs function itself.
---
drivers/net/iavf/iavf_rxtx.c | 12 +++++++++---
1 file changed, 9 insertions(+), 3 deletions(-)
diff --git a/drivers/net/iavf/iavf_rxtx.c b/drivers/net/iavf/iavf_rxtx.c
index e33fe4576b..7ec4f241e2 100644
--- a/drivers/net/iavf/iavf_rxtx.c
+++ b/drivers/net/iavf/iavf_rxtx.c
@@ -274,11 +274,15 @@ alloc_rxq_mbufs(struct iavf_rx_queue *rxq)
volatile union iavf_rx_desc *rxd;
struct rte_mbuf *mbuf = NULL;
uint64_t dma_addr;
- uint16_t i;
+ uint16_t i, j;
for (i = 0; i < rxq->nb_rx_desc; i++) {
mbuf = rte_mbuf_raw_alloc(rxq->mp);
if (unlikely(!mbuf)) {
+ for (j = 0; j < i; j++) {
+ rte_pktmbuf_free_seg(rxq->sw_ring[j]);
+ rxq->sw_ring[j] = NULL;
+ }
PMD_DRV_LOG(ERR, "Failed to allocate mbuf for RX");
return -ENOMEM;
}
@@ -848,12 +852,14 @@ iavf_dev_rx_queue_start(struct rte_eth_dev *dev, uint16_t rx_queue_id)
else
err = iavf_switch_queue_lv(adapter, rx_queue_id, true, true);
- if (err)
+ if (err) {
+ release_rxq_mbufs(rxq);
PMD_DRV_LOG(ERR, "Failed to switch RX queue %u on",
rx_queue_id);
- else
+ } else {
dev->data->rx_queue_state[rx_queue_id] =
RTE_ETH_QUEUE_STATE_STARTED;
+ }
return err;
}
--
2.30.1.windows.1
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [dpdk-dev] [PATCH v2] net/iavf: fix mbuf leak
2021-09-16 14:04 ` Qiming Chen
@ 2021-09-22 7:04 ` Zhang, Qi Z
0 siblings, 0 replies; 7+ messages in thread
From: Zhang, Qi Z @ 2021-09-22 7:04 UTC (permalink / raw)
To: Qiming Chen, dev; +Cc: Xing, Beilei, Wu, Jingjing, stable
> -----Original Message-----
> From: dev <dev-bounces@dpdk.org> On Behalf Of Qiming Chen
> Sent: Thursday, September 16, 2021 10:04 PM
> To: dev@dpdk.org
> Cc: Xing, Beilei <beilei.xing@intel.com>; Wu, Jingjing <jingjing.wu@intel.com>;
> Qiming Chen <chenqiming_huawei@163.com>; stable@dpdk.org
> Subject: [dpdk-dev] [PATCH v2] net/iavf: fix mbuf leak
>
> In the iavf_dev_rx_queue_start function, if the iavf_switch_queue or
> iavf_switch_queue_lv function fails, the previously applied mbuf is not released,
> resulting in leakage. The patch fixes the problem.
>
> Fixes: 9cf9c02bf6ee ("net/iavf: add enable/disable queues for large VF")
> Cc: stable@dpdk.org
>
> Signed-off-by: Qiming Chen <chenqiming_huawei@163.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Applied to dpdk-next-net-intel.
Thanks
Qi
^ permalink raw reply [flat|nested] 7+ messages in thread
* [dpdk-dev] [PATCH] net/iavf: fix mbuf leak
@ 2021-09-11 1:47 Qiming Chen
2021-09-15 1:20 ` Zhang, Qi Z
0 siblings, 1 reply; 7+ messages in thread
From: Qiming Chen @ 2021-09-11 1:47 UTC (permalink / raw)
To: dev; +Cc: beilei.xing, jingjing.wu, Qiming Chen, stable
A local test found that repeated port start and stop operations during
the continuous SSE vector bufflist receiving process will cause the mbuf
resource to run out. The final positioning is when the port is stopped,
the mbuf of the pkt_first_seg pointer is not released. Resources leak.
The patch scheme is to judge whether the pointer is empty when the port
is stopped, and release the corresponding mbuf if it is not empty.
Fixes: 69dd4c3d0898 ("net/avf: enable queue and device")
Cc: stable@dpdk.org
Signed-off-by: Qiming Chen <chenqiming_huawei@163.com>
---
drivers/net/iavf/iavf_rxtx.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/drivers/net/iavf/iavf_rxtx.c b/drivers/net/iavf/iavf_rxtx.c
index 55393a9400..59448eae98 100644
--- a/drivers/net/iavf/iavf_rxtx.c
+++ b/drivers/net/iavf/iavf_rxtx.c
@@ -225,6 +225,10 @@ reset_rx_queue(struct iavf_rx_queue *rxq)
rxq->rx_tail = 0;
rxq->nb_rx_hold = 0;
+
+ if (rxq->pkt_first_seg != NULL)
+ rte_pktmbuf_free(rxq->pkt_first_seg);
+
rxq->pkt_first_seg = NULL;
rxq->pkt_last_seg = NULL;
rxq->rxrearm_nb = 0;
--
2.30.1.windows.1
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [dpdk-dev] [PATCH] net/iavf: fix mbuf leak
2021-09-11 1:47 [dpdk-dev] [PATCH] " Qiming Chen
@ 2021-09-15 1:20 ` Zhang, Qi Z
0 siblings, 0 replies; 7+ messages in thread
From: Zhang, Qi Z @ 2021-09-15 1:20 UTC (permalink / raw)
To: Qiming Chen, dev; +Cc: Xing, Beilei, Wu, Jingjing, stable
> -----Original Message-----
> From: dev <dev-bounces@dpdk.org> On Behalf Of Qiming Chen
> Sent: Saturday, September 11, 2021 9:47 AM
> To: dev@dpdk.org
> Cc: Xing, Beilei <beilei.xing@intel.com>; Wu, Jingjing <jingjing.wu@intel.com>;
> Qiming Chen <chenqiming_huawei@163.com>; stable@dpdk.org
> Subject: [dpdk-dev] [PATCH] net/iavf: fix mbuf leak
>
> A local test found that repeated port start and stop operations during the
> continuous SSE vector bufflist receiving process will cause the mbuf resource
> to run out. The final positioning is when the port is stopped, the mbuf of the
> pkt_first_seg pointer is not released. Resources leak.
> The patch scheme is to judge whether the pointer is empty when the port is
> stopped, and release the corresponding mbuf if it is not empty.
>
> Fixes: 69dd4c3d0898 ("net/avf: enable queue and device")
> Cc: stable@dpdk.org
>
> Signed-off-by: Qiming Chen <chenqiming_huawei@163.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Applied to dpdk-next-net-intel.
Thanks
Qi
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2021-09-22 7:05 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-09-10 8:23 [dpdk-dev] [PATCH] net/iavf: fix mbuf leak Qiming Chen
2021-09-10 8:31 ` [dpdk-dev] [PATCH v2] " Qiming Chen
2021-09-15 3:27 ` Zhang, Qi Z
2021-09-16 14:04 ` Qiming Chen
2021-09-22 7:04 ` Zhang, Qi Z
2021-09-11 1:47 [dpdk-dev] [PATCH] " Qiming Chen
2021-09-15 1:20 ` Zhang, Qi Z
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).