[dpdk-stable] patch 'kni: fix crash caused by freeing mempool' has been queued to LTS release 16.11.2

[dpdk-stable] patch 'kni: fix crash caused by freeing mempool' has been queued to LTS release 16.11.2

[Yuanhan Liu]

Hi,

FYI, your patch has been queued to LTS release 16.11.2

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable
yet. It will be pushed if I get no objections before 05/13/17.
So please shout if anyone has objections.

Thanks.

	--yliu

---
>From 02b8ce63972bc7f12a73131941dfb9aacdc57f11 Mon Sep 17 00:00:00 2001
From: Ferruh Yigit <ferruh.yigit@intel.com>
Date: Wed, 3 May 2017 17:10:37 +0100
Subject: [PATCH] kni: fix crash caused by freeing mempool

[ upstream commit 57a484e08a7ef3dea2018157d518b01a891ddfca ]

To clean alloc_q, which has physical addresses of the mbufs, kni lib
free the pkt_mempool, but this leads a crash in kni unit test.

KNI library shouldn't free the pkt_mempool.

Implementation updated to find the mbufs in the alloc_q and return them
back to mempool.

Fixes: 8eba5ebd1811 ("kni: fix possible memory leak")

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
---
 lib/librte_kni/rte_kni.c | 34 ++++++++++++++++++++++------------
 1 file changed, 22 insertions(+), 12 deletions(-)

diff --git a/lib/librte_kni/rte_kni.c b/lib/librte_kni/rte_kni.c
index 52fcd4b..c3f9208 100644
--- a/lib/librte_kni/rte_kni.c
+++ b/lib/librte_kni/rte_kni.c
@@ -451,17 +451,35 @@ kni_free_fifo(struct rte_kni_fifo *fifo)
 	} while (ret);
 }
 
+static void *
+va2pa(struct rte_mbuf *m)
+{
+	return (void *)((unsigned long)m -
+			((unsigned long)m->buf_addr -
+			 (unsigned long)m->buf_physaddr));
+}
+
 static void
-kni_free_fifo_phy(struct rte_mempool *pktmbuf_pool, struct rte_kni_fifo *fifo)
+obj_free(struct rte_mempool *mp __rte_unused, void *opaque, void *obj,
+		unsigned obj_idx __rte_unused)
+{
+	struct rte_mbuf *m = obj;
+	void *mbuf_phys = opaque;
+
+	if (va2pa(m) == mbuf_phys)
+		rte_pktmbuf_free(m);
+}
+
+static void
+kni_free_fifo_phy(struct rte_mempool *mp, struct rte_kni_fifo *fifo)
 {
 	void *mbuf_phys;
 	int ret;
 
-	rte_mempool_free(pktmbuf_pool);
-
-	/* All mbufs alredy freed with rte_mempoll_free, just free the fifo */
 	do {
 		ret = kni_fifo_get(fifo, &mbuf_phys, 1);
+		if (ret)
+			rte_mempool_obj_iter(mp, obj_free, mbuf_phys);
 	} while (ret);
 }
 
@@ -557,14 +575,6 @@ rte_kni_handle_request(struct rte_kni *kni)
 	return 0;
 }
 
-static void *
-va2pa(struct rte_mbuf *m)
-{
-	return (void *)((unsigned long)m -
-			((unsigned long)m->buf_addr -
-			 (unsigned long)m->buf_physaddr));
-}
-
 unsigned
 rte_kni_tx_burst(struct rte_kni *kni, struct rte_mbuf **mbufs, unsigned num)
 {
-- 
1.9.0

[dpdk-stable] patch 'net: fix stripped VLAN flag for offload emulation' has been queued to LTS release 16.11.2

[Yuanhan Liu]

Hi,

FYI, your patch has been queued to LTS release 16.11.2

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable
yet. It will be pushed if I get no objections before 05/13/17.
So please shout if anyone has objections.

Thanks.

	--yliu

---
>From 1223e6395e15382dcc8c3c4813143f940fb87f08 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Micha=C5=82=20Miros=C5=82aw?=
 <michal.miroslaw@atendesoftware.pl>
Date: Fri, 5 May 2017 00:36:13 +0200
Subject: [PATCH] net: fix stripped VLAN flag for offload emulation
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

[ upstream commit 47aa48b969f8db88de40ddf3121a45443a9db990 ]

Apply the new flag PKT_RX_VLAN_STRIPPED to the software emulation case
(currently only for virtio and af_packet).

Fixes: b37b528d957c ("mbuf: add new Rx flags for stripped VLAN")

Signed-off-by: Michał Mirosław <michal.miroslaw@atendesoftware.pl>
Signed-off-by: Thomas Monjalon <thomas@monjalon.net>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
---
 lib/librte_net/rte_ether.h | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/lib/librte_net/rte_ether.h b/lib/librte_net/rte_ether.h
index ff3d065..5edf66c 100644
--- a/lib/librte_net/rte_ether.h
+++ b/lib/librte_net/rte_ether.h
@@ -357,7 +357,7 @@ static inline int rte_vlan_strip(struct rte_mbuf *m)
 		return -1;
 
 	struct vlan_hdr *vh = (struct vlan_hdr *)(eh + 1);
-	m->ol_flags |= PKT_RX_VLAN_PKT;
+	m->ol_flags |= PKT_RX_VLAN_PKT | PKT_RX_VLAN_STRIPPED;
 	m->vlan_tci = rte_be_to_cpu_16(vh->vlan_tci);
 
 	/* Copy ether header over rather than moving whole packet */
@@ -407,6 +407,8 @@ static inline int rte_vlan_insert(struct rte_mbuf **m)
 	vh = (struct vlan_hdr *) (nh + 1);
 	vh->vlan_tci = rte_cpu_to_be_16((*m)->vlan_tci);
 
+	(*m)->ol_flags &= ~PKT_RX_VLAN_STRIPPED;
+
 	return 0;
 }
 
-- 
1.9.0

[dpdk-stable] patch 'net/igb: fix VF MAC address setting' has been queued to LTS release 16.11.2

[Yuanhan Liu]

Hi,

FYI, your patch has been queued to LTS release 16.11.2

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable
yet. It will be pushed if I get no objections before 05/13/17.
So please shout if anyone has objections.

Thanks.

	--yliu

---
>From a1a28336fa21a459acc89cb4bb95e85645952da9 Mon Sep 17 00:00:00 2001
From: Qiming Yang <qiming.yang@intel.com>
Date: Wed, 3 May 2017 13:51:32 +0800
Subject: [PATCH] net/igb: fix VF MAC address setting

[ upstream commit 580b85f6ba4cf38268c50185a6f5becce8b69846 ]

VF default MAC address be added in PF Mac address list
instead of VF MAC address list, makes VF can't receive
packets. This patch fixes this issue.

Fixes: be2d648a2dd3 ("igb: add PF support")

Signed-off-by: Qiming Yang <qiming.yang@intel.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>
---
 drivers/net/e1000/igb_pf.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/net/e1000/igb_pf.c b/drivers/net/e1000/igb_pf.c
index 5845bc2..3f8f152 100644
--- a/drivers/net/e1000/igb_pf.c
+++ b/drivers/net/e1000/igb_pf.c
@@ -330,12 +330,16 @@ igb_vf_set_mac_addr(struct rte_eth_dev *dev, uint32_t vf, uint32_t *msgbuf)
 		*(E1000_DEV_PRIVATE_TO_P_VFDATA(dev->data->dev_private));
 	int rar_entry = hw->mac.rar_entry_count - (vf + 1);
 	uint8_t *new_mac = (uint8_t *)(&msgbuf[1]);
+	int rah;
 
 	if (is_unicast_ether_addr((struct ether_addr *)new_mac)) {
 		if (!is_zero_ether_addr((struct ether_addr *)new_mac))
 			rte_memcpy(vfinfo[vf].vf_mac_addresses, new_mac,
 				sizeof(vfinfo[vf].vf_mac_addresses));
 		hw->mac.ops.rar_set(hw, new_mac, rar_entry);
+		rah = E1000_READ_REG(hw, E1000_RAH(rar_entry));
+		rah |= (0x1 << (E1000_RAH_POOLSEL_SHIFT + vf));
+		E1000_WRITE_REG(hw, E1000_RAH(rar_entry), rah);
 		return 0;
 	}
 	return -1;
-- 
1.9.0

[dpdk-stable] patch 'net/igb: fix VF MAC address setting' has been queued to LTS release 16.11.2

[Yuanhan Liu]

Hi,

FYI, your patch has been queued to LTS release 16.11.2

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable
yet. It will be pushed if I get no objections before 05/13/17.
So please shout if anyone has objections.

Thanks.

	--yliu

---
>From bf23f10bad67b44816e3f19907cfd7daec7cf5e1 Mon Sep 17 00:00:00 2001
From: Qiming Yang <qiming.yang@intel.com>
Date: Wed, 3 May 2017 13:52:47 +0800
Subject: [PATCH] net/igb: fix VF MAC address setting

[ upstream commit 10f28c196c344aecac870bdf7c78602e6aceca37 ]

We find that VF receive address register is not set
if MAC address is assigned by PF. This patch fixes it.

Fixes: d82170d27918 ("igb: add VF support")

Signed-off-by: Qiming Yang <qiming.yang@intel.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>
---
 drivers/net/e1000/igb_ethdev.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/net/e1000/igb_ethdev.c b/drivers/net/e1000/igb_ethdev.c
index 2fddf0c..be2600d 100644
--- a/drivers/net/e1000/igb_ethdev.c
+++ b/drivers/net/e1000/igb_ethdev.c
@@ -1012,12 +1012,6 @@ eth_igbvf_dev_init(struct rte_eth_dev *eth_dev)
 	/* Generate a random MAC address, if none was assigned by PF. */
 	if (is_zero_ether_addr(perm_addr)) {
 		eth_random_addr(perm_addr->addr_bytes);
-		diag = e1000_rar_set(hw, perm_addr->addr_bytes, 0);
-		if (diag) {
-			rte_free(eth_dev->data->mac_addrs);
-			eth_dev->data->mac_addrs = NULL;
-			return diag;
-		}
 		PMD_INIT_LOG(INFO, "\tVF MAC address not assigned by Host PF");
 		PMD_INIT_LOG(INFO, "\tAssign randomly generated MAC address "
 			     "%02x:%02x:%02x:%02x:%02x:%02x",
@@ -1029,6 +1023,12 @@ eth_igbvf_dev_init(struct rte_eth_dev *eth_dev)
 			     perm_addr->addr_bytes[5]);
 	}
 
+	diag = e1000_rar_set(hw, perm_addr->addr_bytes, 0);
+	if (diag) {
+		rte_free(eth_dev->data->mac_addrs);
+		eth_dev->data->mac_addrs = NULL;
+		return diag;
+	}
 	/* Copy the permanent MAC address */
 	ether_addr_copy((struct ether_addr *) hw->mac.perm_addr,
 			&eth_dev->data->mac_addrs[0]);
-- 
1.9.0

[dpdk-stable] patch 'net/ixgbe: fix VF Rx mode for allmulticast disabled' has been queued to LTS release 16.11.2

[Yuanhan Liu]

Hi,

FYI, your patch has been queued to LTS release 16.11.2

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable
yet. It will be pushed if I get no objections before 05/13/17.
So please shout if anyone has objections.

Thanks.

	--yliu

---
>From 7a5cd8985798c158679f5b1ca5e5b87ab828c10b Mon Sep 17 00:00:00 2001
From: Wei Dai <wei.dai@intel.com>
Date: Thu, 4 May 2017 17:54:40 +0800
Subject: [PATCH] net/ixgbe: fix VF Rx mode for allmulticast disabled

[ upstream commit b604589d572ee8844afaf196e82d2d563f51f543 ]

Some customers find that 82599 NIC DPDK VF PMD can't receive any
broadcast packets when it is bound to igb_uio in the first time
to run a DPDK application like testpmd. But when the application
is quited and run again, the DPDK VF PMD can receive broadcast
packets again. The associated PF is run by kernel driver when
the VF is driven by DPDK PMD.

Fixes: 260e2e22e26f ("net/ixgbe/base: move multicast mode update")
Fixes: 72dec9e37a84 ("ixgbe: support multicast promiscuous mode on VF")

Signed-off-by: Wei Dai <wei.dai@intel.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>
---
 drivers/net/ixgbe/ixgbe_ethdev.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ixgbe/ixgbe_ethdev.c b/drivers/net/ixgbe/ixgbe_ethdev.c
index 5fe357d..0fdf010 100644
--- a/drivers/net/ixgbe/ixgbe_ethdev.c
+++ b/drivers/net/ixgbe/ixgbe_ethdev.c
@@ -7530,7 +7530,7 @@ ixgbevf_dev_allmulticast_disable(struct rte_eth_dev *dev)
 {
 	struct ixgbe_hw *hw = IXGBE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
 
-	hw->mac.ops.update_xcast_mode(hw, IXGBEVF_XCAST_MODE_NONE);
+	hw->mac.ops.update_xcast_mode(hw, IXGBEVF_XCAST_MODE_MULTI);
 }
 
 static void ixgbevf_mbx_process(struct rte_eth_dev *dev)
-- 
1.9.0

[dpdk-stable] patch 'net/ixgbe: fix setting MTU on stopped device' has been queued to LTS release 16.11.2

[Yuanhan Liu]

Hi,

FYI, your patch has been queued to LTS release 16.11.2

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable
yet. It will be pushed if I get no objections before 05/13/17.
So please shout if anyone has objections.

Thanks.

	--yliu

---
>From 774c084d7d95db35645d3ce709d28dbef78f38d9 Mon Sep 17 00:00:00 2001
From: Jia Yu <jyu@vmware.com>
Date: Thu, 27 Apr 2017 23:11:31 -0700
Subject: [PATCH] net/ixgbe: fix setting MTU on stopped device

[ upstream commit 22d0e5aed3c22bd8b08d58cfd2ed261d5e45ae9c ]

There exists case that software sets mtu (i.e jumbo frame) of
ixgbe device when it's stopped. Before the fix, scattered_rx
is cleared during device stop, and setting jumbo frame mtu
after device stop will always fail as scattered_rx is 0.

Signed-off-by: Jia Yu <jyu@vmware.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>
---
 drivers/net/ixgbe/ixgbe_ethdev.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ixgbe/ixgbe_ethdev.c b/drivers/net/ixgbe/ixgbe_ethdev.c
index 0fdf010..d6686f6 100644
--- a/drivers/net/ixgbe/ixgbe_ethdev.c
+++ b/drivers/net/ixgbe/ixgbe_ethdev.c
@@ -4090,6 +4090,7 @@ ixgbe_dev_mtu_set(struct rte_eth_dev *dev, uint16_t mtu)
 	struct ixgbe_hw *hw;
 	struct rte_eth_dev_info dev_info;
 	uint32_t frame_size = mtu + ETHER_HDR_LEN + ETHER_CRC_LEN;
+	struct rte_eth_rxmode *rx_conf = &dev->data->dev_conf.rxmode;
 
 	ixgbe_dev_info_get(dev, &dev_info);
 
@@ -4100,7 +4101,7 @@ ixgbe_dev_mtu_set(struct rte_eth_dev *dev, uint16_t mtu)
 	/* refuse mtu that requires the support of scattered packets when this
 	 * feature has not been enabled before.
 	 */
-	if (!dev->data->scattered_rx &&
+	if (!rx_conf->enable_scatter &&
 	    (frame_size + 2 * IXGBE_VLAN_TAG_SIZE >
 	     dev->data->min_rx_buf_size - RTE_PKTMBUF_HEADROOM))
 		return -EINVAL;
@@ -5723,6 +5724,7 @@ ixgbevf_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu)
 {
 	struct ixgbe_hw *hw;
 	uint32_t max_frame = mtu + ETHER_HDR_LEN + ETHER_CRC_LEN;
+	struct rte_eth_rxmode *rx_conf = &dev->data->dev_conf.rxmode;
 
 	hw = IXGBE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
 
@@ -5732,7 +5734,7 @@ ixgbevf_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu)
 	/* refuse mtu that requires the support of scattered packets when this
 	 * feature has not been enabled before.
 	 */
-	if (!dev->data->scattered_rx &&
+	if (!rx_conf->enable_scatter &&
 	    (max_frame + 2 * IXGBE_VLAN_TAG_SIZE >
 	     dev->data->min_rx_buf_size - RTE_PKTMBUF_HEADROOM))
 		return -EINVAL;
-- 
1.9.0

[dpdk-stable] patch 'net/ixgbe: fix memory overflow in 32-bit SSE Rx' has been queued to LTS release 16.11.2

[Yuanhan Liu]

Hi,

FYI, your patch has been queued to LTS release 16.11.2

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable
yet. It will be pushed if I get no objections before 05/13/17.
So please shout if anyone has objections.

Thanks.

	--yliu

---
>From ad146b8ea41aa3bcda6e34860e56485608219989 Mon Sep 17 00:00:00 2001
From: Qi Zhang <qi.z.zhang@intel.com>
Date: Sun, 30 Apr 2017 01:26:50 -0400
Subject: [PATCH] net/ixgbe: fix memory overflow in 32-bit SSE Rx

[ upstream commit 6d65eba145cc3d84d9d0de1824eacfa2392a5eee ]

Return mbuf points of _recv_raw_pkts_vec are modified out of bound.

Fixes: c95584dc2b18 ("ixgbe: new vectorized functions for Rx/Tx")

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
---
 drivers/net/ixgbe/ixgbe_rxtx_vec_sse.c | 16 ++++++++++++----
 1 file changed, 12 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ixgbe/ixgbe_rxtx_vec_sse.c b/drivers/net/ixgbe/ixgbe_rxtx_vec_sse.c
index abbf284..dd7d177 100644
--- a/drivers/net/ixgbe/ixgbe_rxtx_vec_sse.c
+++ b/drivers/net/ixgbe/ixgbe_rxtx_vec_sse.c
@@ -335,9 +335,13 @@ _recv_raw_pkts_vec(struct ixgbe_rx_queue *rxq, struct rte_mbuf **rx_pkts,
 		__m128i descs[RTE_IXGBE_DESCS_PER_LOOP];
 		__m128i pkt_mb1, pkt_mb2, pkt_mb3, pkt_mb4;
 		__m128i zero, staterr, sterr_tmp1, sterr_tmp2;
-		__m128i mbp1, mbp2; /* two mbuf pointer in one XMM reg. */
+		/* 2 64 bit or 4 32 bit mbuf pointers in one XMM reg. */
+		__m128i mbp1;
+#if defined(RTE_ARCH_X86_64)
+		__m128i mbp2;
+#endif
 
-		/* B.1 load 1 mbuf point */
+		/* B.1 load 2 (64 bit) or 4 (32 bit) mbuf points */
 		mbp1 = _mm_loadu_si128((__m128i *)&sw_ring[pos]);
 
 		/* Read desc statuses backwards to avoid race condition */
@@ -345,11 +349,13 @@ _recv_raw_pkts_vec(struct ixgbe_rx_queue *rxq, struct rte_mbuf **rx_pkts,
 		descs[3] = _mm_loadu_si128((__m128i *)(rxdp + 3));
 		rte_compiler_barrier();
 
-		/* B.2 copy 2 mbuf point into rx_pkts  */
+		/* B.2 copy 2 64 bit or 4 32 bit mbuf point into rx_pkts */
 		_mm_storeu_si128((__m128i *)&rx_pkts[pos], mbp1);
 
-		/* B.1 load 1 mbuf point */
+#if defined(RTE_ARCH_X86_64)
+		/* B.1 load 2 64 bit mbuf points */
 		mbp2 = _mm_loadu_si128((__m128i *)&sw_ring[pos+2]);
+#endif
 
 		descs[2] = _mm_loadu_si128((__m128i *)(rxdp + 2));
 		rte_compiler_barrier();
@@ -358,8 +364,10 @@ _recv_raw_pkts_vec(struct ixgbe_rx_queue *rxq, struct rte_mbuf **rx_pkts,
 		rte_compiler_barrier();
 		descs[0] = _mm_loadu_si128((__m128i *)(rxdp));
 
+#if defined(RTE_ARCH_X86_64)
 		/* B.2 copy 2 mbuf point into rx_pkts  */
 		_mm_storeu_si128((__m128i *)&rx_pkts[pos+2], mbp2);
+#endif
 
 		if (split_packet) {
 			rte_mbuf_prefetch_part2(rx_pkts[pos]);
-- 
1.9.0

[dpdk-stable] patch 'net/i40e: fix memory overflow in 32-bit SSE Rx' has been queued to LTS release 16.11.2

[Yuanhan Liu]

Hi,

FYI, your patch has been queued to LTS release 16.11.2

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable
yet. It will be pushed if I get no objections before 05/13/17.
So please shout if anyone has objections.

Thanks.

	--yliu

---
>From 230b64b61eb82f69cec2e6d6e8c08ed78bef5b67 Mon Sep 17 00:00:00 2001
From: Qi Zhang <qi.z.zhang@intel.com>
Date: Sun, 30 Apr 2017 01:26:49 -0400
Subject: [PATCH] net/i40e: fix memory overflow in 32-bit SSE Rx

[ upstream commit f4afa1a0487fc73055a6d5571a20fdc0ed99866d ]

Return mbuf points of _recv_raw_pkts_vec are modified out of bound.

Fixes: 8e109464c022 ("i40e: allow vector Rx and Tx usage")

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
---
 drivers/net/i40e/i40e_rxtx_vec_sse.c | 16 ++++++++++++----
 1 file changed, 12 insertions(+), 4 deletions(-)

diff --git a/drivers/net/i40e/i40e_rxtx_vec_sse.c b/drivers/net/i40e/i40e_rxtx_vec_sse.c
index b95cc8e..9644dd6 100644
--- a/drivers/net/i40e/i40e_rxtx_vec_sse.c
+++ b/drivers/net/i40e/i40e_rxtx_vec_sse.c
@@ -320,20 +320,26 @@ _recv_raw_pkts_vec(struct i40e_rx_queue *rxq, struct rte_mbuf **rx_pkts,
 		__m128i descs[RTE_I40E_DESCS_PER_LOOP];
 		__m128i pkt_mb1, pkt_mb2, pkt_mb3, pkt_mb4;
 		__m128i zero, staterr, sterr_tmp1, sterr_tmp2;
-		__m128i mbp1, mbp2; /* two mbuf pointer in one XMM reg. */
+		/* 2 64 bit or 4 32 bit mbuf pointers in one XMM reg. */
+		__m128i mbp1;
+#if defined(RTE_ARCH_X86_64)
+		__m128i mbp2;
+#endif
 
-		/* B.1 load 1 mbuf point */
+		/* B.1 load 2 (64 bit) or 4 (32 bit) mbuf points */
 		mbp1 = _mm_loadu_si128((__m128i *)&sw_ring[pos]);
 		/* Read desc statuses backwards to avoid race condition */
 		/* A.1 load 4 pkts desc */
 		descs[3] = _mm_loadu_si128((__m128i *)(rxdp + 3));
 		rte_compiler_barrier();
 
-		/* B.2 copy 2 mbuf point into rx_pkts  */
+		/* B.2 copy 2 64 bit or 4 32 bit mbuf point into rx_pkts */
 		_mm_storeu_si128((__m128i *)&rx_pkts[pos], mbp1);
 
-		/* B.1 load 1 mbuf point */
+#if defined(RTE_ARCH_X86_64)
+		/* B.1 load 2 64 bit mbuf points */
 		mbp2 = _mm_loadu_si128((__m128i *)&sw_ring[pos+2]);
+#endif
 
 		descs[2] = _mm_loadu_si128((__m128i *)(rxdp + 2));
 		rte_compiler_barrier();
@@ -342,8 +348,10 @@ _recv_raw_pkts_vec(struct i40e_rx_queue *rxq, struct rte_mbuf **rx_pkts,
 		rte_compiler_barrier();
 		descs[0] = _mm_loadu_si128((__m128i *)(rxdp));
 
+#if defined(RTE_ARCH_X86_64)
 		/* B.2 copy 2 mbuf point into rx_pkts  */
 		_mm_storeu_si128((__m128i *)&rx_pkts[pos+2], mbp2);
+#endif
 
 		if (split_packet) {
 			rte_mbuf_prefetch_part2(rx_pkts[pos]);
-- 
1.9.0

[dpdk-stable] patch 'net/fm10k: fix memory overflow in 32-bit SSE Rx' has been queued to LTS release 16.11.2

[Yuanhan Liu]

Hi,

FYI, your patch has been queued to LTS release 16.11.2

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable
yet. It will be pushed if I get no objections before 05/13/17.
So please shout if anyone has objections.

Thanks.

	--yliu

---
>From 81ea8e49fdcada184cbc0823ccad125740aa4c22 Mon Sep 17 00:00:00 2001
From: Qi Zhang <qi.z.zhang@intel.com>
Date: Sun, 30 Apr 2017 01:26:51 -0400
Subject: [PATCH] net/fm10k: fix memory overflow in 32-bit SSE Rx

[ upstream commit 70fca51612722ed686d5565700717384b8f52887 ]

Return mbuf points of _recv_raw_pkts_vec are modified out of bound.

Fixes: a6ce64a97520 ("fm10k: introduce vector driver")

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
---
 drivers/net/fm10k/fm10k_rxtx_vec.c | 16 ++++++++++++----
 1 file changed, 12 insertions(+), 4 deletions(-)

diff --git a/drivers/net/fm10k/fm10k_rxtx_vec.c b/drivers/net/fm10k/fm10k_rxtx_vec.c
index 27f3e43..3d21652 100644
--- a/drivers/net/fm10k/fm10k_rxtx_vec.c
+++ b/drivers/net/fm10k/fm10k_rxtx_vec.c
@@ -470,9 +470,13 @@ fm10k_recv_raw_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
 		__m128i descs0[RTE_FM10K_DESCS_PER_LOOP];
 		__m128i pkt_mb1, pkt_mb2, pkt_mb3, pkt_mb4;
 		__m128i zero, staterr, sterr_tmp1, sterr_tmp2;
-		__m128i mbp1, mbp2; /* two mbuf pointer in one XMM reg. */
+		__m128i mbp1;
+		/* 2 64 bit or 4 32 bit mbuf pointers in one XMM reg. */
+#if defined(RTE_ARCH_X86_64)
+		__m128i mbp2;
+#endif
 
-		/* B.1 load 1 mbuf point */
+		/* B.1 load 2 (64 bit) or 4 (32 bit) mbuf points */
 		mbp1 = _mm_loadu_si128((__m128i *)&mbufp[pos]);
 
 		/* Read desc statuses backwards to avoid race condition */
@@ -480,11 +484,13 @@ fm10k_recv_raw_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
 		descs0[3] = _mm_loadu_si128((__m128i *)(rxdp + 3));
 		rte_compiler_barrier();
 
-		/* B.2 copy 2 mbuf point into rx_pkts  */
+		/* B.2 copy 2 64 bit or 4 32 bit mbuf point into rx_pkts */
 		_mm_storeu_si128((__m128i *)&rx_pkts[pos], mbp1);
 
-		/* B.1 load 1 mbuf point */
+#if defined(RTE_ARCH_X86_64)
+		/* B.1 load 2 64 bit mbuf poitns */
 		mbp2 = _mm_loadu_si128((__m128i *)&mbufp[pos+2]);
+#endif
 
 		descs0[2] = _mm_loadu_si128((__m128i *)(rxdp + 2));
 		rte_compiler_barrier();
@@ -493,8 +499,10 @@ fm10k_recv_raw_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
 		rte_compiler_barrier();
 		descs0[0] = _mm_loadu_si128((__m128i *)(rxdp));
 
+#if defined(RTE_ARCH_X86_64)
 		/* B.2 copy 2 mbuf point into rx_pkts  */
 		_mm_storeu_si128((__m128i *)&rx_pkts[pos+2], mbp2);
+#endif
 
 		/* avoid compiler reorder optimization */
 		rte_compiler_barrier();
-- 
1.9.0

[dpdk-stable] patch 'net/vmxnet3: fix build with gcc 7' has been queued to LTS release 16.11.2

[Yuanhan Liu]

Hi,

FYI, your patch has been queued to LTS release 16.11.2

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable
yet. It will be pushed if I get no objections before 05/13/17.
So please shout if anyone has objections.

Thanks.

	--yliu

---
>From db47746ea14a1043178c56551711231a0a7283dc Mon Sep 17 00:00:00 2001
From: Bruce Richardson <bruce.richardson@intel.com>
Date: Thu, 4 May 2017 16:38:17 +0100
Subject: [PATCH] net/vmxnet3: fix build with gcc 7

[ upstream commit 2fff4ff7b77d6b9054b9c894d0a78fe5ebd0d746 ]

GCC 7 flags a value as uninitialized before used. While it's a false
positive, there is little harm in providing an initial value for the
variable.

Fixes: bb1d14b87fc3 ("vmxnet3: fix link state handling")

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
---
 drivers/net/vmxnet3/vmxnet3_ethdev.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/vmxnet3/vmxnet3_ethdev.c b/drivers/net/vmxnet3/vmxnet3_ethdev.c
index 8bb13e5..f123df9 100644
--- a/drivers/net/vmxnet3/vmxnet3_ethdev.c
+++ b/drivers/net/vmxnet3/vmxnet3_ethdev.c
@@ -771,7 +771,7 @@ vmxnet3_dev_link_update(struct rte_eth_dev *dev,
 			__rte_unused int wait_to_complete)
 {
 	struct vmxnet3_hw *hw = dev->data->dev_private;
-	struct rte_eth_link old, link;
+	struct rte_eth_link old = { 0 }, link;
 	uint32_t ret;
 
 	/* Link status doesn't change for stopped dev */
-- 
1.9.0

[dpdk-stable] patch 'doc: explain zlib dependency for bnx2x' has been queued to LTS release 16.11.2

[Yuanhan Liu]

Hi,

FYI, your patch has been queued to LTS release 16.11.2

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable
yet. It will be pushed if I get no objections before 05/13/17.
So please shout if anyone has objections.

Thanks.

	--yliu

---
>From f1b996f636b5ad018ba8ce87d672a3406bd6b98f Mon Sep 17 00:00:00 2001
From: Rasesh Mody <rasesh.mody@cavium.com>
Date: Tue, 2 May 2017 16:57:49 -0700
Subject: [PATCH] doc: explain zlib dependency for bnx2x

[ upstream commit 6d4e18669575ad0d036b2b7a16a15b122b482b0a ]

Correct CONFIG_RTE_LIBRTE_BNX2X_PMD config file option and add a note
about external zlib dependency for loading the firmware image.

Fixes: ce9b8bb8b998 ("config: disable bnx2x driver")

Signed-off-by: Rasesh Mody <rasesh.mody@cavium.com>
---
 doc/guides/nics/bnx2x.rst | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/doc/guides/nics/bnx2x.rst b/doc/guides/nics/bnx2x.rst
index 6d1768a..2e34942 100644
--- a/doc/guides/nics/bnx2x.rst
+++ b/doc/guides/nics/bnx2x.rst
@@ -96,9 +96,11 @@ Config File Options
 The following options can be modified in the ``.config`` file. Please note that
 enabling debugging options may affect system performance.
 
-- ``CONFIG_RTE_LIBRTE_BNX2X_PMD`` (default **y**)
+- ``CONFIG_RTE_LIBRTE_BNX2X_PMD`` (default **n**)
 
-  Toggle compilation of bnx2x driver.
+  Toggle compilation of bnx2x driver. To use bnx2x PMD set this config parameter
+  to 'y'. Also, in order for firmware binary to load user will need zlib devel
+  package installed.
 
 - ``CONFIG_RTE_LIBRTE_BNX2X_DEBUG`` (default **n**)
 
-- 
1.9.0

[dpdk-stable] patch 'test/cmdline: fix missing break in switch' has been queued to LTS release 16.11.2

[Yuanhan Liu]

Hi,

FYI, your patch has been queued to LTS release 16.11.2

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable
yet. It will be pushed if I get no objections before 05/13/17.
So please shout if anyone has objections.

Thanks.

	--yliu

---
>From c3d82296efc79c11f8f116448ce25933893a0717 Mon Sep 17 00:00:00 2001
From: Bruce Richardson <bruce.richardson@intel.com>
Date: Thu, 4 May 2017 16:38:21 +0100
Subject: [PATCH] test/cmdline: fix missing break in switch

[ upstream commit f4e6ff881ffcdd1178550fc26a6343bbc8d36530 ]

Issue flagged by GCC 7 as a switch fall-through.

Fixes: dbb860e03eb1 ("cmdline: tests")

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
---
 app/test/test_cmdline_num.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/app/test/test_cmdline_num.c b/app/test/test_cmdline_num.c
index 04263d3..e8f60cf 100644
--- a/app/test/test_cmdline_num.c
+++ b/app/test/test_cmdline_num.c
@@ -315,6 +315,7 @@ can_parse_signed(int64_t expected_result, enum cmdline_numtype type)
 	case UINT64:
 		if (expected_result < 0)
 			return 0;
+		break;
 	case INT8:
 		if (expected_result > INT8_MAX || expected_result < INT8_MIN)
 			return 0;
-- 
1.9.0