[dpdk-dev] [PATCH 0/2] fix mbufs leakage during Rx queue release on ixgbe/i40e

[dpdk-dev] [PATCH 0/2] fix mbufs leakage during Rx queue release on ixgbe/i40e

[Yury Kylulin]

This patch set fixes problem when mbufs are not released back to the
original memory pool during Rx queue relese call leading to mbufs
"leakage".

This issue happens when device stop (rte_eth_dev_stop()) will be called
after device start (rte_eth_dev_start()) without any incoming packets
received since device start. In this case both rxq->rx_tail and
rxq->rxrearm_start equal to 0 and release cycle will not be executed.
On the next device start (rte_eth_dev_start()) new set of mbufs for the
RX queue will be allocated.

Yury Kylulin (2):
  net/ixgbe: fix mbufs leakage during Rx queue release
  net/i40e: fix mbufs leakage during Rx queue release

 drivers/net/i40e/i40e_rxtx_vec.c          | 16 ++++++++++++++--
 drivers/net/ixgbe/ixgbe_rxtx_vec_common.h | 16 ++++++++++++++--
 2 files changed, 28 insertions(+), 4 deletions(-)

-- 
2.7.4

[dpdk-dev] [PATCH 1/2] net/ixgbe: fix mbufs leakage during Rx queue release

[Yury Kylulin]

For the vector PMD release all mbufs from the Rx queue if no packets
received after device start.

Fixes: 11b220c6498d ("ixgbe: fix release queue mbufs")

Signed-off-by: Yury Kylulin <yury.kylulin@intel.com>
---
 drivers/net/ixgbe/ixgbe_rxtx_vec_common.h | 16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ixgbe/ixgbe_rxtx_vec_common.h b/drivers/net/ixgbe/ixgbe_rxtx_vec_common.h
index 62b8201..3c3c009 100644
--- a/drivers/net/ixgbe/ixgbe_rxtx_vec_common.h
+++ b/drivers/net/ixgbe/ixgbe_rxtx_vec_common.h
@@ -204,8 +204,20 @@ _ixgbe_rx_queue_release_mbufs_vec(struct ixgbe_rx_queue *rxq)
 		return;
 
 	/* free all mbufs that are valid in the ring */
-	for (i = rxq->rx_tail; i != rxq->rxrearm_start; i = (i + 1) & mask)
-		rte_pktmbuf_free_seg(rxq->sw_ring[i].mbuf);
+	if (rxq->rxrearm_nb == 0) {
+		for (i = 0; i < rxq->nb_rx_desc; i++) {
+			if (rxq->sw_ring[i].mbuf != NULL)
+				rte_pktmbuf_free_seg(rxq->sw_ring[i].mbuf);
+		}
+	} else {
+		for (i = rxq->rx_tail;
+		     i != rxq->rxrearm_start;
+		     i = (i + 1) & mask) {
+			if (rxq->sw_ring[i].mbuf != NULL)
+				rte_pktmbuf_free_seg(rxq->sw_ring[i].mbuf);
+		}
+	}
+
 	rxq->rxrearm_nb = rxq->nb_rx_desc;
 
 	/* set all entries to NULL */
-- 
2.7.4

Re: [dpdk-dev] [PATCH 1/2] net/ixgbe: fix mbufs leakage during Rx queue release

[Lu, Wenzhuo]

Hi,

> -----Original Message-----
> From: Kylulin, Yury
> Sent: Tuesday, August 30, 2016 12:51 AM
> To: Zhang, Helin; Ananyev, Konstantin; Wu, Jingjing
> Cc: Lu, Wenzhuo; dev@dpdk.org; Kylulin, Yury
> Subject: [PATCH 1/2] net/ixgbe: fix mbufs leakage during Rx queue release
> 
> For the vector PMD release all mbufs from the Rx queue if no packets received
> after device start.
> 
> Fixes: 11b220c6498d ("ixgbe: fix release queue mbufs")
> 
> Signed-off-by: Yury Kylulin <yury.kylulin@intel.com>
Acked-by: Wenzhuo Lu <Wenzhuo.lu@intel.com>

[dpdk-dev] [PATCH 2/2] net/i40e: fix mbufs leakage during Rx queue release

[Yury Kylulin]

For the vector PMD release all mbufs from the Rx queue if no packets
received after device start.

Fixes: 9ed94e5bb04e ("i40e: add vector Rx")

Signed-off-by: Yury Kylulin <yury.kylulin@intel.com>
---
 drivers/net/i40e/i40e_rxtx_vec.c | 16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)

diff --git a/drivers/net/i40e/i40e_rxtx_vec.c b/drivers/net/i40e/i40e_rxtx_vec.c
index 51fb282..82571ee 100644
--- a/drivers/net/i40e/i40e_rxtx_vec.c
+++ b/drivers/net/i40e/i40e_rxtx_vec.c
@@ -692,8 +692,20 @@ i40e_rx_queue_release_mbufs_vec(struct i40e_rx_queue *rxq)
 		return;
 
 	/* free all mbufs that are valid in the ring */
-	for (i = rxq->rx_tail; i != rxq->rxrearm_start; i = (i + 1) & mask)
-		rte_pktmbuf_free_seg(rxq->sw_ring[i].mbuf);
+	if (rxq->rxrearm_nb == 0) {
+		for (i = 0; i < rxq->nb_rx_desc; i++) {
+			if (rxq->sw_ring[i].mbuf != NULL)
+				rte_pktmbuf_free_seg(rxq->sw_ring[i].mbuf);
+		}
+	} else {
+		for (i = rxq->rx_tail;
+		     i != rxq->rxrearm_start;
+		     i = (i + 1) & mask) {
+			if (rxq->sw_ring[i].mbuf != NULL)
+				rte_pktmbuf_free_seg(rxq->sw_ring[i].mbuf);
+		}
+	}
+
 	rxq->rxrearm_nb = rxq->nb_rx_desc;
 
 	/* set all entries to NULL */
-- 
2.7.4

Re: [dpdk-dev] [PATCH 2/2] net/i40e: fix mbufs leakage during Rx queue release

[Xing, Beilei]

Hi,

> -----Original Message-----
> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Yury Kylulin
> Sent: Tuesday, August 30, 2016 12:51 AM
> To: Zhang, Helin <helin.zhang@intel.com>; Ananyev, Konstantin
> <konstantin.ananyev@intel.com>; Wu, Jingjing <jingjing.wu@intel.com>
> Cc: Lu, Wenzhuo <wenzhuo.lu@intel.com>; dev@dpdk.org; Kylulin, Yury
> <yury.kylulin@intel.com>
> Subject: [dpdk-dev] [PATCH 2/2] net/i40e: fix mbufs leakage during Rx
> queue release
> 
> For the vector PMD release all mbufs from the Rx queue if no packets
> received after device start.
> 
> Fixes: 9ed94e5bb04e ("i40e: add vector Rx")
> 
> Signed-off-by: Yury Kylulin <yury.kylulin@intel.com>
Acked-by: Beilei Xing <beilei.xing@intel.com>

Re: [dpdk-dev] [PATCH 0/2] fix mbufs leakage during Rx queue release on ixgbe/i40e

[Bruce Richardson]

On Mon, Aug 29, 2016 at 07:50:46PM +0300, Yury Kylulin wrote:
> This patch set fixes problem when mbufs are not released back to the
> original memory pool during Rx queue relese call leading to mbufs
> "leakage".
> 
> This issue happens when device stop (rte_eth_dev_stop()) will be called
> after device start (rte_eth_dev_start()) without any incoming packets
> received since device start. In this case both rxq->rx_tail and
> rxq->rxrearm_start equal to 0 and release cycle will not be executed.
> On the next device start (rte_eth_dev_start()) new set of mbufs for the
> RX queue will be allocated.
> 
> Yury Kylulin (2):
>   net/ixgbe: fix mbufs leakage during Rx queue release
>   net/i40e: fix mbufs leakage during Rx queue release
> 
Applied to dpdk-next-net/rel_16_11

Thanks,
/Bruce

Re: [dpdk-dev] [PATCH 0/2] fix mbufs leakage during Rx queue release on ixgbe/i40e

[Christian Ehrhardt]

On Mon, Sep 19, 2016 at 3:29 PM, Bruce Richardson <
bruce.richardson@intel.com> wrote:

> > This patch set fixes problem when mbufs are not released back to the
> > original memory pool during Rx queue relese call leading to mbufs
> > "leakage".
> >
> > This issue happens when device stop (rte_eth_dev_stop()) will be called
> > after device start (rte_eth_dev_start()) without any incoming packets
> > received since device start. In this case both rxq->rx_tail and
> > rxq->rxrearm_start equal to 0 and release cycle will not be executed.
> > On the next device start (rte_eth_dev_start()) new set of mbufs for the
> > RX queue will be allocated.
> >
> > Yury Kylulin (2):
> >   net/ixgbe: fix mbufs leakage during Rx queue release
> >   net/i40e: fix mbufs leakage during Rx queue release
> >
> Applied to dpdk-next-net/rel_16_11
>

I don't know the underlying code well enough, but given the description I'd
expect that qualifies for the stable tree?
Set stable@dpdk on cc.

Re: [dpdk-dev] [dpdk-stable] [PATCH 0/2] fix mbufs leakage during Rx queue release on ixgbe/i40e

[Yuanhan Liu]

On Mon, Sep 19, 2016 at 03:52:25PM +0200, Christian Ehrhardt wrote:
> On Mon, Sep 19, 2016 at 3:29 PM, Bruce Richardson <
> bruce.richardson@intel.com> wrote:
> 
> > > This patch set fixes problem when mbufs are not released back to the
> > > original memory pool during Rx queue relese call leading to mbufs
> > > "leakage".
> > >
> > > This issue happens when device stop (rte_eth_dev_stop()) will be called
> > > after device start (rte_eth_dev_start()) without any incoming packets
> > > received since device start. In this case both rxq->rx_tail and
> > > rxq->rxrearm_start equal to 0 and release cycle will not be executed.
> > > On the next device start (rte_eth_dev_start()) new set of mbufs for the
> > > RX queue will be allocated.
> > >
> > > Yury Kylulin (2):
> > >   net/ixgbe: fix mbufs leakage during Rx queue release
> > >   net/i40e: fix mbufs leakage during Rx queue release
> > >
> > Applied to dpdk-next-net/rel_16_11
> >
> 
> I don't know the underlying code well enough,

Me, neither.

> but given the description I'd
> expect that qualifies for the stable tree?

But, yes, I agree with Christian.

> Set stable@dpdk on cc.

Thanks for doing that.

	--yliu