From 28c5869f2bc4e319027bc22a66a1aa9eefc69a16 Mon Sep 17 00:00:00 2001
From: Helin Zhang <helin.zhang@intel.com>
Date: Mon, 26 Oct 2015 19:44:27 -0400
Subject: [PATCH 01/17] i40e: add new fields to store user configuration

This patch adds new fields to i40e_vsi to store user configured
RSS config data and code to use it.

Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Change-ID: I73886469dca9e9f6b16d842182a87f3f4009f95d
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40e/i40e.h        |  2 +
 .../net/ethernet/intel/i40e/i40e_ethtool.c    | 33 +++++++-----
 drivers/net/ethernet/intel/i40e/i40e_main.c   | 52 +++++++++++++++++--
 3 files changed, 69 insertions(+), 18 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
index 89f5323362ed..6efab66a05ae 100644
--- a/drivers/net/ethernet/intel/i40e/i40e.h
+++ b/drivers/net/ethernet/intel/i40e/i40e.h
@@ -508,6 +508,8 @@ struct i40e_vsi {
 
 	u16 rss_table_size;
 	u16 rss_size;
+	u8  *rss_hkey_user; /* User configured hash keys */
+	u8  *rss_lut_user;  /* User configured lookup table entries */
 
 	u16 max_frame;
 	u16 rx_hdr_len;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
index 6cb2b340cdc2..b52c50924274 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_ethtool.c
@@ -2651,10 +2651,8 @@ static int i40e_set_rxfh(struct net_device *netdev, const u32 *indir,
 {
 	struct i40e_netdev_priv *np = netdev_priv(netdev);
 	struct i40e_vsi *vsi = np->vsi;
-	u8 seed_def[I40E_HKEY_ARRAY_SIZE];
-	u8 *lut, *seed = NULL;
+	u8 *seed = NULL;
 	u16 i;
-	int ret;
 
 	if (hfunc != ETH_RSS_HASH_NO_CHANGE && hfunc != ETH_RSS_HASH_TOP)
 		return -EOPNOTSUPP;
@@ -2663,18 +2661,27 @@ static int i40e_set_rxfh(struct net_device *netdev, const u32 *indir,
 		return 0;
 
 	if (key) {
-		memcpy(seed_def, key, I40E_HKEY_ARRAY_SIZE);
-		seed = seed_def;
+		if (!vsi->rss_hkey_user) {
+			vsi->rss_hkey_user = kzalloc(I40E_HKEY_ARRAY_SIZE,
+						     GFP_KERNEL);
+			if (!vsi->rss_hkey_user)
+				return -ENOMEM;
+		}
+		memcpy(vsi->rss_hkey_user, key, I40E_HKEY_ARRAY_SIZE);
+		seed = vsi->rss_hkey_user;
+	}
+	if (!vsi->rss_lut_user) {
+		vsi->rss_lut_user = kzalloc(I40E_HLUT_ARRAY_SIZE, GFP_KERNEL);
+		if (!vsi->rss_lut_user)
+			return -ENOMEM;
 	}
-	lut = kzalloc(I40E_HLUT_ARRAY_SIZE, GFP_KERNEL);
-	if (!lut)
-		return -ENOMEM;
-	for (i = 0; i < I40E_HLUT_ARRAY_SIZE; i++)
-		lut[i] = (u8)(indir[i]);
-	ret = i40e_config_rss(vsi, seed, lut, I40E_HLUT_ARRAY_SIZE);
-	kfree(lut);
 
-	return ret;
+	/* Each 32 bits pointed by 'indir' is stored with a lut entry */
+	for (i = 0; i < I40E_HLUT_ARRAY_SIZE; i++)
+		vsi->rss_lut_user[i] = (u8)(indir[i]);
+
+	return i40e_config_rss(vsi, seed, vsi->rss_lut_user,
+			       I40E_HLUT_ARRAY_SIZE);
 }
 
 /**
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 4b7d87447fe1..c0a784fadeb0 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -7300,6 +7300,23 @@ static void i40e_vsi_free_arrays(struct i40e_vsi *vsi, bool free_qvectors)
 	vsi->rx_rings = NULL;
 }
 
+/**
+ * i40e_clear_rss_config_user - clear the user configured RSS hash keys
+ * and lookup table
+ * @vsi: Pointer to VSI structure
+ */
+static void i40e_clear_rss_config_user(struct i40e_vsi *vsi)
+{
+	if (!vsi)
+		return;
+
+	kfree(vsi->rss_hkey_user);
+	vsi->rss_hkey_user = NULL;
+
+	kfree(vsi->rss_lut_user);
+	vsi->rss_lut_user = NULL;
+}
+
 /**
  * i40e_vsi_clear - Deallocate the VSI provided
  * @vsi: the VSI being un-configured
@@ -7337,6 +7354,7 @@ static int i40e_vsi_clear(struct i40e_vsi *vsi)
 	i40e_put_lump(pf->irq_pile, vsi->base_vector, vsi->idx);
 
 	i40e_vsi_free_arrays(vsi, true);
+	i40e_clear_rss_config_user(vsi);
 
 	pf->vsi[vsi->idx] = NULL;
 	if (vsi->idx < pf->next_vsi)
@@ -8015,8 +8033,6 @@ static int i40e_pf_config_rss(struct i40e_pf *pf)
 	wr32(hw, I40E_PFQF_HENA(0), (u32)hena);
 	wr32(hw, I40E_PFQF_HENA(1), (u32)(hena >> 32));
 
-	vsi->rss_size = min_t(int, pf->rss_size, vsi->num_queue_pairs);
-
 	/* Determine the RSS table size based on the hardware capabilities */
 	reg_val = rd32(hw, I40E_PFQF_CTL_0);
 	reg_val = (pf->rss_table_size == 512) ?
@@ -8024,15 +8040,28 @@ static int i40e_pf_config_rss(struct i40e_pf *pf)
 			(reg_val & ~I40E_PFQF_CTL_0_HASHLUTSIZE_512);
 	wr32(hw, I40E_PFQF_CTL_0, reg_val);
 
+	/* Determine the RSS size of the VSI */
+	if (!vsi->rss_size)
+		vsi->rss_size = min_t(int, pf->rss_size, vsi->num_queue_pairs);
+
 	lut = kzalloc(vsi->rss_table_size, GFP_KERNEL);
 	if (!lut)
 		return -ENOMEM;
 
-	i40e_fill_rss_lut(pf, lut, vsi->rss_table_size, vsi->rss_size);
+	/* Use user configured lut if there is one, otherwise use default */
+	if (vsi->rss_lut_user)
+		memcpy(lut, vsi->rss_lut_user, vsi->rss_table_size);
+	else
+		i40e_fill_rss_lut(pf, lut, vsi->rss_table_size, vsi->rss_size);
 
-	netdev_rss_key_fill((void *)seed, I40E_HKEY_ARRAY_SIZE);
+	/* Use user configured hash key if there is one, otherwise
+	 * use default.
+	 */
+	if (vsi->rss_hkey_user)
+		memcpy(seed, vsi->rss_hkey_user, I40E_HKEY_ARRAY_SIZE);
+	else
+		netdev_rss_key_fill((void *)seed, I40E_HKEY_ARRAY_SIZE);
 	ret = i40e_config_rss(vsi, seed, lut, vsi->rss_table_size);
-
 	kfree(lut);
 
 	return ret;
@@ -8063,6 +8092,19 @@ int i40e_reconfig_rss_queues(struct i40e_pf *pf, int queue_count)
 		pf->rss_size = new_rss_size;
 
 		i40e_reset_and_rebuild(pf, true);
+
+		/* Discard the user configured hash keys and lut, if less
+		 * queues are enabled.
+		 */
+		if (queue_count < vsi->rss_size) {
+			i40e_clear_rss_config_user(vsi);
+			dev_dbg(&pf->pdev->dev,
+				"discard user configured hash keys and lut\n");
+		}
+
+		/* Reset vsi->rss_size, as number of enabled queues changed */
+		vsi->rss_size = min_t(int, pf->rss_size, vsi->num_queue_pairs);
+
 		i40e_pf_config_rss(pf);
 	}
 	dev_info(&pf->pdev->dev, "RSS count:  %d\n", pf->rss_size);

From acd65448f2e6b3407b8bb28b57d711ca417ea679 Mon Sep 17 00:00:00 2001
From: Helin Zhang <helin.zhang@intel.com>
Date: Mon, 26 Oct 2015 19:44:28 -0400
Subject: [PATCH 02/17] i40e: rename rss_size to alloc_rss_size in i40e_pf

This patch renames rss_size to alloc_rss_size in i40e_pf, which is
clearer and avoids confusion. It also adds comments to the other
related structure members to help clarify usage.

Change-ID: Ia90090609d006ab589cb639975bb8a0af795d16f
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40e/i40e.h      |  8 +++---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 31 ++++++++++++---------
 2 files changed, 22 insertions(+), 17 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
index 6efab66a05ae..d854a46bba1e 100644
--- a/drivers/net/ethernet/intel/i40e/i40e.h
+++ b/drivers/net/ethernet/intel/i40e/i40e.h
@@ -266,7 +266,7 @@ struct i40e_pf {
 	u16 num_lan_qps;           /* num lan queues this PF has set up */
 	u16 num_lan_msix;          /* num queue vectors for the base PF vsi */
 	int queues_left;           /* queues left unclaimed */
-	u16 rss_size;              /* num queues in the RSS array */
+	u16 alloc_rss_size;        /* allocated RSS queues */
 	u16 rss_size_max;          /* HW defined max RSS queues */
 	u16 fdir_pf_filter_count;  /* num of guaranteed filters for this PF */
 	u16 num_alloc_vsi;         /* num VSIs this driver supports */
@@ -413,7 +413,7 @@ struct i40e_pf {
 	u32 rx_hwtstamp_cleared;
 	bool ptp_tx;
 	bool ptp_rx;
-	u16 rss_table_size;
+	u16 rss_table_size; /* HW RSS table size */
 	/* These are only valid in NPAR modes */
 	u32 npar_max_bw;
 	u32 npar_min_bw;
@@ -506,8 +506,8 @@ struct i40e_vsi {
 	u16 tx_itr_setting;
 	u16 int_rate_limit;  /* value in usecs */
 
-	u16 rss_table_size;
-	u16 rss_size;
+	u16 rss_table_size; /* HW RSS table size */
+	u16 rss_size;       /* Allocated RSS queues */
 	u8  *rss_hkey_user; /* User configured hash keys */
 	u8  *rss_lut_user;  /* User configured lookup table entries */
 
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index c0a784fadeb0..491c82fe6467 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -1630,7 +1630,8 @@ static void i40e_vsi_setup_queue_map(struct i40e_vsi *vsi,
 
 			switch (vsi->type) {
 			case I40E_VSI_MAIN:
-				qcount = min_t(int, pf->rss_size, num_tc_qps);
+				qcount = min_t(int, pf->alloc_rss_size,
+					       num_tc_qps);
 				break;
 #ifdef I40E_FCOE
 			case I40E_VSI_FCOE:
@@ -7883,7 +7884,7 @@ static int i40e_vsi_config_rss(struct i40e_vsi *vsi)
 
 	i40e_fill_rss_lut(pf, lut, vsi->rss_table_size, vsi->rss_size);
 	netdev_rss_key_fill((void *)seed, I40E_HKEY_ARRAY_SIZE);
-	vsi->rss_size = min_t(int, pf->rss_size, vsi->num_queue_pairs);
+	vsi->rss_size = min_t(int, pf->alloc_rss_size, vsi->num_queue_pairs);
 	ret = i40e_config_rss_aq(vsi, seed, lut, vsi->rss_table_size);
 	kfree(lut);
 
@@ -8042,7 +8043,8 @@ static int i40e_pf_config_rss(struct i40e_pf *pf)
 
 	/* Determine the RSS size of the VSI */
 	if (!vsi->rss_size)
-		vsi->rss_size = min_t(int, pf->rss_size, vsi->num_queue_pairs);
+		vsi->rss_size = min_t(int, pf->alloc_rss_size,
+				      vsi->num_queue_pairs);
 
 	lut = kzalloc(vsi->rss_table_size, GFP_KERNEL);
 	if (!lut)
@@ -8089,7 +8091,7 @@ int i40e_reconfig_rss_queues(struct i40e_pf *pf, int queue_count)
 		vsi->req_queue_pairs = queue_count;
 		i40e_prep_for_reset(pf);
 
-		pf->rss_size = new_rss_size;
+		pf->alloc_rss_size = new_rss_size;
 
 		i40e_reset_and_rebuild(pf, true);
 
@@ -8103,12 +8105,13 @@ int i40e_reconfig_rss_queues(struct i40e_pf *pf, int queue_count)
 		}
 
 		/* Reset vsi->rss_size, as number of enabled queues changed */
-		vsi->rss_size = min_t(int, pf->rss_size, vsi->num_queue_pairs);
+		vsi->rss_size = min_t(int, pf->alloc_rss_size,
+				      vsi->num_queue_pairs);
 
 		i40e_pf_config_rss(pf);
 	}
-	dev_info(&pf->pdev->dev, "RSS count:  %d\n", pf->rss_size);
-	return pf->rss_size;
+	dev_info(&pf->pdev->dev, "RSS count:  %d\n", pf->alloc_rss_size);
+	return pf->alloc_rss_size;
 }
 
 /**
@@ -8279,13 +8282,14 @@ static int i40e_sw_init(struct i40e_pf *pf)
 	 * maximum might end up larger than the available queues
 	 */
 	pf->rss_size_max = BIT(pf->hw.func_caps.rss_table_entry_width);
-	pf->rss_size = 1;
+	pf->alloc_rss_size = 1;
 	pf->rss_table_size = pf->hw.func_caps.rss_table_size;
 	pf->rss_size_max = min_t(int, pf->rss_size_max,
 				 pf->hw.func_caps.num_tx_qp);
 	if (pf->hw.func_caps.rss) {
 		pf->flags |= I40E_FLAG_RSS_ENABLED;
-		pf->rss_size = min_t(int, pf->rss_size_max, num_online_cpus());
+		pf->alloc_rss_size = min_t(int, pf->rss_size_max,
+					   num_online_cpus());
 	}
 
 	/* MFP mode enabled */
@@ -10152,7 +10156,7 @@ static void i40e_determine_queue_usage(struct i40e_pf *pf)
 	    !(pf->flags & I40E_FLAG_MSIX_ENABLED)) {
 		/* one qp for PF, no queues for anything else */
 		queues_left = 0;
-		pf->rss_size = pf->num_lan_qps = 1;
+		pf->alloc_rss_size = pf->num_lan_qps = 1;
 
 		/* make sure all the fancies are disabled */
 		pf->flags &= ~(I40E_FLAG_RSS_ENABLED	|
@@ -10169,7 +10173,7 @@ static void i40e_determine_queue_usage(struct i40e_pf *pf)
 				  I40E_FLAG_FD_ATR_ENABLED |
 				  I40E_FLAG_DCB_CAPABLE))) {
 		/* one qp for PF */
-		pf->rss_size = pf->num_lan_qps = 1;
+		pf->alloc_rss_size = pf->num_lan_qps = 1;
 		queues_left -= pf->num_lan_qps;
 
 		pf->flags &= ~(I40E_FLAG_RSS_ENABLED	|
@@ -10239,8 +10243,9 @@ static void i40e_determine_queue_usage(struct i40e_pf *pf)
 		"qs_avail=%d FD SB=%d lan_qs=%d lan_tc0=%d vf=%d*%d vmdq=%d*%d, remaining=%d\n",
 		pf->hw.func_caps.num_tx_qp,
 		!!(pf->flags & I40E_FLAG_FD_SB_ENABLED),
-		pf->num_lan_qps, pf->rss_size, pf->num_req_vfs, pf->num_vf_qps,
-		pf->num_vmdq_vsis, pf->num_vmdq_qps, queues_left);
+		pf->num_lan_qps, pf->alloc_rss_size, pf->num_req_vfs,
+		pf->num_vf_qps, pf->num_vmdq_vsis, pf->num_vmdq_qps,
+		queues_left);
 #ifdef I40E_FCOE
 	dev_dbg(&pf->pdev->dev, "fcoe queues = %d\n", pf->num_fcoe_qps);
 #endif

From 6a7fded776a778f728b13d83a2c9fc893580c080 Mon Sep 17 00:00:00 2001
From: Anjali Singhai Jain <anjali.singhai@intel.com>
Date: Mon, 26 Oct 2015 19:44:29 -0400
Subject: [PATCH 03/17] i40e/i40evf: Fix RS bit update in Tx path and disable
 force WB workaround

This patch fixes the issue of forcing WB too often causing us to not
benefit from NAPI.

Without this patch we were forcing WB/arming interrupt too often taking
away the benefits of NAPI and causing a performance impact.

With this patch we disable force WB in the clean routine for X710
and XL710 adapters. X722 adapters do not enable interrupt to force
a WB and benefit from WB_ON_ITR and hence force WB is left enabled
for those adapters.
For XL710 and X710 adapters if we have less than 4 packets pending
a software Interrupt triggered from service task will force a WB.

This patch also changes the conditions for setting RS bit as described
in code comments. This optimizes when the HW does a tail bump amd when
it does a WB. It also optimizes when we do a wmb.

Change-ID: Id831e1ae7d3e2ec3f52cd0917b41ce1d22d75d9d
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40evf/i40e_txrx.c | 118 +++++++++++-------
 drivers/net/ethernet/intel/i40evf/i40e_txrx.h |   2 +
 2 files changed, 77 insertions(+), 43 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
index 8629a9f77e42..23fd11243583 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
@@ -245,16 +245,6 @@ static bool i40e_clean_tx_irq(struct i40e_ring *tx_ring, int budget)
 	tx_ring->q_vector->tx.total_bytes += total_bytes;
 	tx_ring->q_vector->tx.total_packets += total_packets;
 
-	/* check to see if there are any non-cache aligned descriptors
-	 * waiting to be written back, and kick the hardware to force
-	 * them to be written back in case of napi polling
-	 */
-	if (budget &&
-	    !((i & WB_STRIDE) == WB_STRIDE) &&
-	    !test_bit(__I40E_DOWN, &tx_ring->vsi->state) &&
-	    (I40E_DESC_UNUSED(tx_ring) != tx_ring->count))
-		tx_ring->arm_wb = true;
-
 	netdev_tx_completed_queue(netdev_get_tx_queue(tx_ring->netdev,
 						      tx_ring->queue_index),
 				  total_packets, total_bytes);
@@ -1770,6 +1760,9 @@ static inline void i40evf_tx_map(struct i40e_ring *tx_ring, struct sk_buff *skb,
 	u32 td_tag = 0;
 	dma_addr_t dma;
 	u16 gso_segs;
+	u16 desc_count = 0;
+	bool tail_bump = true;
+	bool do_rs = false;
 
 	if (tx_flags & I40E_TX_FLAGS_HW_VLAN) {
 		td_cmd |= I40E_TX_DESC_CMD_IL2TAG1;
@@ -1810,6 +1803,8 @@ static inline void i40evf_tx_map(struct i40e_ring *tx_ring, struct sk_buff *skb,
 
 			tx_desc++;
 			i++;
+			desc_count++;
+
 			if (i == tx_ring->count) {
 				tx_desc = I40E_TX_DESC(tx_ring, 0);
 				i = 0;
@@ -1829,6 +1824,8 @@ static inline void i40evf_tx_map(struct i40e_ring *tx_ring, struct sk_buff *skb,
 
 		tx_desc++;
 		i++;
+		desc_count++;
+
 		if (i == tx_ring->count) {
 			tx_desc = I40E_TX_DESC(tx_ring, 0);
 			i = 0;
@@ -1843,35 +1840,7 @@ static inline void i40evf_tx_map(struct i40e_ring *tx_ring, struct sk_buff *skb,
 		tx_bi = &tx_ring->tx_bi[i];
 	}
 
-	/* Place RS bit on last descriptor of any packet that spans across the
-	 * 4th descriptor (WB_STRIDE aka 0x3) in a 64B cacheline.
-	 */
 #define WB_STRIDE 0x3
-	if (((i & WB_STRIDE) != WB_STRIDE) &&
-	    (first <= &tx_ring->tx_bi[i]) &&
-	    (first >= &tx_ring->tx_bi[i & ~WB_STRIDE])) {
-		tx_desc->cmd_type_offset_bsz =
-			build_ctob(td_cmd, td_offset, size, td_tag) |
-			cpu_to_le64((u64)I40E_TX_DESC_CMD_EOP <<
-					 I40E_TXD_QW1_CMD_SHIFT);
-	} else {
-		tx_desc->cmd_type_offset_bsz =
-			build_ctob(td_cmd, td_offset, size, td_tag) |
-			cpu_to_le64((u64)I40E_TXD_CMD <<
-					 I40E_TXD_QW1_CMD_SHIFT);
-	}
-
-	netdev_tx_sent_queue(netdev_get_tx_queue(tx_ring->netdev,
-						 tx_ring->queue_index),
-			     first->bytecount);
-
-	/* Force memory writes to complete before letting h/w
-	 * know there are new descriptors to fetch.  (Only
-	 * applicable for weak-ordered memory model archs,
-	 * such as IA-64).
-	 */
-	wmb();
-
 	/* set next_to_watch value indicating a packet is present */
 	first->next_to_watch = tx_desc;
 
@@ -1881,15 +1850,78 @@ static inline void i40evf_tx_map(struct i40e_ring *tx_ring, struct sk_buff *skb,
 
 	tx_ring->next_to_use = i;
 
+	netdev_tx_sent_queue(netdev_get_tx_queue(tx_ring->netdev,
+						 tx_ring->queue_index),
+						 first->bytecount);
 	i40evf_maybe_stop_tx(tx_ring, DESC_NEEDED);
+
+	/* Algorithm to optimize tail and RS bit setting:
+	 * if xmit_more is supported
+	 *	if xmit_more is true
+	 *		do not update tail and do not mark RS bit.
+	 *	if xmit_more is false and last xmit_more was false
+	 *		if every packet spanned less than 4 desc
+	 *			then set RS bit on 4th packet and update tail
+	 *			on every packet
+	 *		else
+	 *			update tail and set RS bit on every packet.
+	 *	if xmit_more is false and last_xmit_more was true
+	 *		update tail and set RS bit.
+	 * else (kernel < 3.18)
+	 *	if every packet spanned less than 4 desc
+	 *		then set RS bit on 4th packet and update tail
+	 *		on every packet
+	 *	else
+	 *		set RS bit on EOP for every packet and update tail
+	 *
+	 * Optimization: wmb to be issued only in case of tail update.
+	 * Also optimize the Descriptor WB path for RS bit with the same
+	 * algorithm.
+	 *
+	 * Note: If there are less than 4 packets
+	 * pending and interrupts were disabled the service task will
+	 * trigger a force WB.
+	 */
+	if (skb->xmit_more  &&
+	    !netif_xmit_stopped(netdev_get_tx_queue(tx_ring->netdev,
+						    tx_ring->queue_index))) {
+		tx_ring->flags |= I40E_TXR_FLAGS_LAST_XMIT_MORE_SET;
+		tail_bump = false;
+	} else if (!skb->xmit_more &&
+		   !netif_xmit_stopped(netdev_get_tx_queue(tx_ring->netdev,
+						       tx_ring->queue_index)) &&
+		   (!(tx_ring->flags & I40E_TXR_FLAGS_LAST_XMIT_MORE_SET)) &&
+		   (tx_ring->packet_stride < WB_STRIDE) &&
+		   (desc_count < WB_STRIDE)) {
+		tx_ring->packet_stride++;
+	} else {
+		tx_ring->packet_stride = 0;
+		tx_ring->flags &= ~I40E_TXR_FLAGS_LAST_XMIT_MORE_SET;
+		do_rs = true;
+	}
+	if (do_rs)
+		tx_ring->packet_stride = 0;
+
+	tx_desc->cmd_type_offset_bsz =
+			build_ctob(td_cmd, td_offset, size, td_tag) |
+			cpu_to_le64((u64)(do_rs ? I40E_TXD_CMD :
+						  I40E_TX_DESC_CMD_EOP) <<
+						  I40E_TXD_QW1_CMD_SHIFT);
+
 	/* notify HW of packet */
-	if (!skb->xmit_more ||
-	    netif_xmit_stopped(netdev_get_tx_queue(tx_ring->netdev,
-						   tx_ring->queue_index)))
-		writel(i, tx_ring->tail);
-	else
+	if (!tail_bump)
 		prefetchw(tx_desc + 1);
 
+	if (tail_bump) {
+		/* Force memory writes to complete before letting h/w
+		 * know there are new descriptors to fetch.  (Only
+		 * applicable for weak-ordered memory model archs,
+		 * such as IA-64).
+		 */
+		wmb();
+		writel(i, tx_ring->tail);
+	}
+
 	return;
 
 dma_error:
diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.h b/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
index 997b374e86af..929ddd9f8215 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
@@ -268,6 +268,8 @@ struct i40e_ring {
 
 	bool ring_active;		/* is ring online or not */
 	bool arm_wb;		/* do something to arm write back */
+	u8 packet_stride;
+#define I40E_TXR_FLAGS_LAST_XMIT_MORE_SET BIT(2)
 
 	u16 flags;
 #define I40E_TXR_FLAGS_WB_ON_ITR	BIT(0)

From b74118f08356fd8ab6cb5d1a15705a2760c9afdd Mon Sep 17 00:00:00 2001
From: Jesse Brandeburg <jesse.brandeburg@intel.com>
Date: Mon, 26 Oct 2015 19:44:30 -0400
Subject: [PATCH 04/17] i40e/i40evf: prefetch skb data on transmit

Issue a prefetch for data early in the transmit path.
This should not be generally needed for Tx traffic, but
it helps immensely for pktgen workloads and should help
for forwarding workloads as well.

Change-ID: Iefee870c20599e0c4240e1d8637e4f16b625f83a
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40e/i40e_txrx.c   | 3 +++
 drivers/net/ethernet/intel/i40evf/i40e_txrx.c | 3 +++
 2 files changed, 6 insertions(+)

diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index dbd2bcae763d..7371df7b6fb0 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -2806,6 +2806,9 @@ static netdev_tx_t i40e_xmit_frame_ring(struct sk_buff *skb,
 	int tsyn;
 	int tso;
 
+	/* prefetch the data, we'll need it later */
+	prefetch(skb->data);
+
 	if (0 == i40e_xmit_descriptor_count(skb, tx_ring))
 		return NETDEV_TX_BUSY;
 
diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
index 23fd11243583..0f1a4ef5e3be 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
@@ -1993,6 +1993,9 @@ static netdev_tx_t i40e_xmit_frame_ring(struct sk_buff *skb,
 	u8 hdr_len = 0;
 	int tso;
 
+	/* prefetch the data, we'll need it later */
+	prefetch(skb->data);
+
 	if (0 == i40evf_xmit_descriptor_count(skb, tx_ring))
 		return NETDEV_TX_BUSY;
 

From 96a8198652e5b80a2cfc87397a3512bf8e45cd63 Mon Sep 17 00:00:00 2001
From: Helin Zhang <helin.zhang@intel.com>
Date: Mon, 26 Oct 2015 19:44:31 -0400
Subject: [PATCH 05/17] i40evf: rename VF adapter specific RSS function

This patch renames old VF adapter specific RSS function to clarify
its scope.

Change-ID: Ie5253083a44c677ebb7709a8a3a18402ad2dc6a6
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40evf/i40evf_main.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_main.c b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
index 6ad62656c75e..1b38c8883952 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_main.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
@@ -1295,10 +1295,10 @@ static void i40evf_configure_rss_reg(struct i40evf_adapter *adapter,
 }
 
 /**
- * i40evf_configure_rss - Prepare for RSS
+ * i40evf_init_rss - Prepare for RSS
  * @adapter: board private structure
  **/
-static void i40evf_configure_rss(struct i40evf_adapter *adapter)
+static void i40evf_init_rss(struct i40evf_adapter *adapter)
 {
 	struct i40e_hw *hw = &adapter->hw;
 	u8 seed[I40EVF_HKEY_ARRAY_SIZE];
@@ -1564,7 +1564,7 @@ static void i40evf_watchdog_task(struct work_struct *work)
 		 * PF, so we don't have to set current_op as we will
 		 * not get a response through the ARQ.
 		 */
-		i40evf_configure_rss(adapter);
+		i40evf_init_rss(adapter);
 		adapter->aq_required &= ~I40EVF_FLAG_AQ_CONFIGURE_RSS;
 		goto watchdog_done;
 	}
@@ -2312,7 +2312,7 @@ static void i40evf_init_task(struct work_struct *work)
 		    I40E_VIRTCHNL_VF_OFFLOAD_WB_ON_ITR)
 		adapter->flags |= I40EVF_FLAG_WB_ON_ITR_CAPABLE;
 	if (!RSS_AQ(adapter))
-		i40evf_configure_rss(adapter);
+		i40evf_init_rss(adapter);
 	err = i40evf_request_misc_irq(adapter);
 	if (err)
 		goto err_sw_init;
@@ -2342,7 +2342,7 @@ static void i40evf_init_task(struct work_struct *work)
 		adapter->aq_required |= I40EVF_FLAG_AQ_CONFIGURE_RSS;
 		mod_timer_pending(&adapter->watchdog_timer, jiffies + 1);
 	} else {
-		i40evf_configure_rss(adapter);
+		i40evf_init_rss(adapter);
 	}
 	return;
 restart:

From 2c86ac3c70794f0ecc3684b0ae0fd75b0cf0c1f6 Mon Sep 17 00:00:00 2001
From: Helin Zhang <helin.zhang@intel.com>
Date: Tue, 27 Oct 2015 16:15:06 -0400
Subject: [PATCH 06/17] i40evf: create a generic config RSS function

There are two ways to configure RSS, this patch adjusts those two
functions with the same input parameters, and creates a more
generic function for configuring RSS.

Change-ID: Iace73bdeba4831909979bef221011060ab327f71
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40evf/i40evf.h    |   3 +
 .../ethernet/intel/i40evf/i40evf_ethtool.c    |  29 +--
 .../net/ethernet/intel/i40evf/i40evf_main.c   | 168 +++++++++++-------
 3 files changed, 130 insertions(+), 70 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40evf/i40evf.h b/drivers/net/ethernet/intel/i40evf/i40evf.h
index 22fc3d49c4b9..b5aa377fb8ae 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf.h
+++ b/drivers/net/ethernet/intel/i40evf/i40evf.h
@@ -99,6 +99,7 @@ struct i40e_vsi {
 #define MAX_TX_QUEUES MAX_RX_QUEUES
 
 #define I40EVF_HKEY_ARRAY_SIZE ((I40E_VFQF_HKEY_MAX_INDEX + 1) * 4)
+#define I40EVF_HLUT_ARRAY_SIZE ((I40E_VFQF_HLUT_MAX_INDEX + 1) * 4)
 
 /* MAX_MSIX_Q_VECTORS of these are allocated,
  * but we only use one per queue-specific vector.
@@ -313,4 +314,6 @@ void i40evf_request_reset(struct i40evf_adapter *adapter);
 void i40evf_virtchnl_completion(struct i40evf_adapter *adapter,
 				enum i40e_virtchnl_ops v_opcode,
 				i40e_status v_retval, u8 *msg, u16 msglen);
+int i40evf_config_rss(struct i40e_vsi *vsi, const u8 *seed, u8 *lut,
+		      u16 lut_size);
 #endif /* _I40EVF_H_ */
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c b/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
index 4790437a50ac..cf253df7f3b2 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
@@ -668,9 +668,11 @@ static int i40evf_set_rxfh(struct net_device *netdev, const u32 *indir,
 			   const u8 *key, const u8 hfunc)
 {
 	struct i40evf_adapter *adapter = netdev_priv(netdev);
-	struct i40e_hw *hw = &adapter->hw;
-	u32 hlut_val;
-	int i, j;
+	struct i40e_vsi *vsi = &adapter->vsi;
+	u8 seed_def[I40EVF_HKEY_ARRAY_SIZE];
+	u8 *seed = NULL, *lut;
+	int ret;
+	u16 i;
 
 	/* We do not allow change in unsupported parameters */
 	if (key ||
@@ -679,15 +681,22 @@ static int i40evf_set_rxfh(struct net_device *netdev, const u32 *indir,
 	if (!indir)
 		return 0;
 
-	for (i = 0, j = 0; i <= I40E_VFQF_HLUT_MAX_INDEX; i++) {
-		hlut_val = indir[j++];
-		hlut_val |= indir[j++] << 8;
-		hlut_val |= indir[j++] << 16;
-		hlut_val |= indir[j++] << 24;
-		wr32(hw, I40E_VFQF_HLUT(i), hlut_val);
+	if (key) {
+		memcpy(seed_def, key, I40EVF_HKEY_ARRAY_SIZE);
+		seed = seed_def;
 	}
+	lut = kzalloc(I40EVF_HLUT_ARRAY_SIZE, GFP_KERNEL);
+	if (!lut)
+		return -ENOMEM;
 
-	return 0;
+	/* Each 32 bits pointed by 'indir' is stored with a lut entry */
+	for (i = 0; i < I40EVF_HLUT_ARRAY_SIZE; i++)
+		lut[i] = (u8)(indir[i]);
+
+	ret = i40evf_config_rss(vsi, seed, lut, I40EVF_HLUT_ARRAY_SIZE);
+	kfree(lut);
+
+	return ret;
 }
 
 static const struct ethtool_ops i40evf_ethtool_ops = {
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_main.c b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
index 1b38c8883952..63d7e470d17c 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_main.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
@@ -1211,110 +1211,158 @@ out:
 }
 
 /**
- * i40e_configure_rss_aq - Prepare for RSS using AQ commands
+ * i40e_config_rss_aq - Prepare for RSS using AQ commands
  * @vsi: vsi structure
  * @seed: RSS hash seed
+ * @lut: Lookup table
+ * @lut_size: Lookup table size
+ *
+ * Return 0 on success, negative on failure
  **/
-static void i40evf_configure_rss_aq(struct i40e_vsi *vsi, const u8 *seed)
+static int i40evf_config_rss_aq(struct i40e_vsi *vsi, const u8 *seed,
+				u8 *lut, u16 lut_size)
 {
-	struct i40e_aqc_get_set_rss_key_data rss_key;
 	struct i40evf_adapter *adapter = vsi->back;
 	struct i40e_hw *hw = &adapter->hw;
-	int ret = 0, i;
-	u8 *rss_lut;
+	int ret = 0;
 
 	if (!vsi->id)
-		return;
+		return -EINVAL;
 
 	if (adapter->current_op != I40E_VIRTCHNL_OP_UNKNOWN) {
 		/* bail because we already have a command pending */
 		dev_err(&adapter->pdev->dev, "Cannot confiure RSS, command %d pending\n",
 			adapter->current_op);
-		return;
+		return -EBUSY;
 	}
 
-	memset(&rss_key, 0, sizeof(rss_key));
-	memcpy(&rss_key, seed, sizeof(rss_key));
-
-	rss_lut = kzalloc(((I40E_VFQF_HLUT_MAX_INDEX + 1) * 4), GFP_KERNEL);
-	if (!rss_lut)
-		return;
-
-	/* Populate the LUT with max no. PF queues in round robin fashion */
-	for (i = 0; i <= (I40E_VFQF_HLUT_MAX_INDEX * 4); i++)
-		rss_lut[i] = i % adapter->num_active_queues;
-
-	ret = i40evf_aq_set_rss_key(hw, vsi->id, &rss_key);
-	if (ret) {
-		dev_err(&adapter->pdev->dev,
-			"Cannot set RSS key, err %s aq_err %s\n",
-			i40evf_stat_str(hw, ret),
-			i40evf_aq_str(hw, hw->aq.asq_last_status));
-		return;
+	if (seed) {
+		struct i40e_aqc_get_set_rss_key_data *rss_key =
+			(struct i40e_aqc_get_set_rss_key_data *)seed;
+		ret = i40evf_aq_set_rss_key(hw, vsi->id, rss_key);
+		if (ret) {
+			dev_err(&adapter->pdev->dev, "Cannot set RSS key, err %s aq_err %s\n",
+				i40evf_stat_str(hw, ret),
+				i40evf_aq_str(hw, hw->aq.asq_last_status));
+			return ret;
+		}
 	}
 
-	ret = i40evf_aq_set_rss_lut(hw, vsi->id, false, rss_lut,
-				    (I40E_VFQF_HLUT_MAX_INDEX + 1) * 4);
-	if (ret)
-		dev_err(&adapter->pdev->dev,
-			"Cannot set RSS lut, err %s aq_err %s\n",
-			i40evf_stat_str(hw, ret),
-			i40evf_aq_str(hw, hw->aq.asq_last_status));
+	if (lut) {
+		ret = i40evf_aq_set_rss_lut(hw, vsi->id, false, lut, lut_size);
+		if (ret) {
+			dev_err(&adapter->pdev->dev,
+				"Cannot set RSS lut, err %s aq_err %s\n",
+				i40evf_stat_str(hw, ret),
+				i40evf_aq_str(hw, hw->aq.asq_last_status));
+			return ret;
+		}
+	}
+
+	return ret;
 }
 
 /**
- * i40e_configure_rss_reg - Prepare for RSS if used
- * @adapter: board private structure
+ * i40evf_config_rss_reg - Configure RSS keys and lut by writing registers
+ * @vsi: Pointer to vsi structure
  * @seed: RSS hash seed
+ * @lut: Lookup table
+ * @lut_size: Lookup table size
+ *
+ * Returns 0 on success, negative on failure
  **/
-static void i40evf_configure_rss_reg(struct i40evf_adapter *adapter,
-				     const u8 *seed)
+static int i40evf_config_rss_reg(struct i40e_vsi *vsi, const u8 *seed,
+				 const u8 *lut, u16 lut_size)
 {
+	struct i40evf_adapter *adapter = vsi->back;
 	struct i40e_hw *hw = &adapter->hw;
-	u32 *seed_dw = (u32 *)seed;
-	u32 cqueue = 0;
-	u32 lut = 0;
-	int i, j;
+	u16 i;
 
-	/* Fill out hash function seed */
-	for (i = 0; i <= I40E_VFQF_HKEY_MAX_INDEX; i++)
-		wr32(hw, I40E_VFQF_HKEY(i), seed_dw[i]);
+	if (seed) {
+		u32 *seed_dw = (u32 *)seed;
 
-	/* Populate the LUT with max no. PF queues in round robin fashion */
-	for (i = 0; i <= I40E_VFQF_HLUT_MAX_INDEX; i++) {
-		lut = 0;
-		for (j = 0; j < 4; j++) {
-			if (cqueue == adapter->num_active_queues)
-				cqueue = 0;
-			lut |= ((cqueue) << (8 * j));
-			cqueue++;
-		}
-		wr32(hw, I40E_VFQF_HLUT(i), lut);
+		for (i = 0; i <= I40E_VFQF_HKEY_MAX_INDEX; i++)
+			wr32(hw, I40E_VFQF_HKEY(i), seed_dw[i]);
+	}
+
+	if (lut) {
+		u32 *lut_dw = (u32 *)lut;
+
+		if (lut_size != I40EVF_HLUT_ARRAY_SIZE)
+			return -EINVAL;
+
+		for (i = 0; i <= I40E_VFQF_HLUT_MAX_INDEX; i++)
+			wr32(hw, I40E_VFQF_HLUT(i), lut_dw[i]);
 	}
 	i40e_flush(hw);
+
+	return 0;
+}
+
+/**
+ * i40evf_config_rss - Configure RSS keys and lut
+ * @vsi: Pointer to vsi structure
+ * @seed: RSS hash seed
+ * @lut: Lookup table
+ * @lut_size: Lookup table size
+ *
+ * Returns 0 on success, negative on failure
+ **/
+int i40evf_config_rss(struct i40e_vsi *vsi, const u8 *seed,
+		      u8 *lut, u16 lut_size)
+{
+	struct i40evf_adapter *adapter = vsi->back;
+
+	if (RSS_AQ(adapter))
+		return i40evf_config_rss_aq(vsi, seed, lut, lut_size);
+	else
+		return i40evf_config_rss_reg(vsi, seed, lut, lut_size);
+}
+
+/**
+ * i40evf_fill_rss_lut - Fill the lut with default values
+ * @lut: Lookup table to be filled with
+ * @rss_table_size: Lookup table size
+ * @rss_size: Range of queue number for hashing
+ **/
+static void i40evf_fill_rss_lut(u8 *lut, u16 rss_table_size, u16 rss_size)
+{
+	u16 i;
+
+	for (i = 0; i < rss_table_size; i++)
+		lut[i] = i % rss_size;
 }
 
 /**
  * i40evf_init_rss - Prepare for RSS
  * @adapter: board private structure
+ *
+ * Return 0 on success, negative on failure
  **/
-static void i40evf_init_rss(struct i40evf_adapter *adapter)
+static int i40evf_init_rss(struct i40evf_adapter *adapter)
 {
+	struct i40e_vsi *vsi = &adapter->vsi;
 	struct i40e_hw *hw = &adapter->hw;
 	u8 seed[I40EVF_HKEY_ARRAY_SIZE];
 	u64 hena;
-
-	netdev_rss_key_fill((void *)seed, I40EVF_HKEY_ARRAY_SIZE);
+	u8 *lut;
+	int ret;
 
 	/* Enable PCTYPES for RSS, TCP/UDP with IPv4/IPv6 */
 	hena = I40E_DEFAULT_RSS_HENA;
 	wr32(hw, I40E_VFQF_HENA(0), (u32)hena);
 	wr32(hw, I40E_VFQF_HENA(1), (u32)(hena >> 32));
 
-	if (RSS_AQ(adapter))
-		i40evf_configure_rss_aq(&adapter->vsi, seed);
-	else
-		i40evf_configure_rss_reg(adapter, seed);
+	lut = kzalloc(I40EVF_HLUT_ARRAY_SIZE, GFP_KERNEL);
+	if (!lut)
+		return -ENOMEM;
+	i40evf_fill_rss_lut(lut, I40EVF_HLUT_ARRAY_SIZE,
+			    adapter->num_active_queues);
+	netdev_rss_key_fill((void *)seed, I40EVF_HKEY_ARRAY_SIZE);
+	ret = i40evf_config_rss(vsi, seed, lut, I40EVF_HLUT_ARRAY_SIZE);
+	kfree(lut);
+
+	return ret;
 }
 
 /**

From 90b02b4392eb9778265ec6b4fef034010a937910 Mon Sep 17 00:00:00 2001
From: Helin Zhang <helin.zhang@intel.com>
Date: Mon, 26 Oct 2015 19:44:33 -0400
Subject: [PATCH 07/17] i40evf: create a generic get RSS function

There are two ways to get RSS, this patch implements two functions
with the same input parameters, and creates a more generic function
for getting RSS configuration.

Change-ID: I12d3b712c21455d47dd0a5aae58fc9b7c680db59
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40evf/i40evf.h    |  2 +
 .../ethernet/intel/i40evf/i40evf_ethtool.c    | 35 ++++---
 .../net/ethernet/intel/i40evf/i40evf_main.c   | 97 +++++++++++++++++++
 3 files changed, 121 insertions(+), 13 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40evf/i40evf.h b/drivers/net/ethernet/intel/i40evf/i40evf.h
index b5aa377fb8ae..5f2e61dc20ec 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf.h
+++ b/drivers/net/ethernet/intel/i40evf/i40evf.h
@@ -316,4 +316,6 @@ void i40evf_virtchnl_completion(struct i40evf_adapter *adapter,
 				i40e_status v_retval, u8 *msg, u16 msglen);
 int i40evf_config_rss(struct i40e_vsi *vsi, const u8 *seed, u8 *lut,
 		      u16 lut_size);
+int i40evf_get_rss(struct i40e_vsi *vsi, const u8 *seed, u8 *lut,
+		   u16 lut_size);
 #endif /* _I40EVF_H_ */
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c b/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
index cf253df7f3b2..a5209707bcf3 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
@@ -634,25 +634,34 @@ static int i40evf_get_rxfh(struct net_device *netdev, u32 *indir, u8 *key,
 			   u8 *hfunc)
 {
 	struct i40evf_adapter *adapter = netdev_priv(netdev);
-	struct i40e_hw *hw = &adapter->hw;
-	u32 hlut_val;
-	int i, j;
+	struct i40e_vsi *vsi = &adapter->vsi;
+	u8 *seed = NULL, *lut;
+	int ret;
+	u16 i;
 
 	if (hfunc)
 		*hfunc = ETH_RSS_HASH_TOP;
 	if (!indir)
 		return 0;
 
-	if (indir) {
-		for (i = 0, j = 0; i <= I40E_VFQF_HLUT_MAX_INDEX; i++) {
-			hlut_val = rd32(hw, I40E_VFQF_HLUT(i));
-			indir[j++] = hlut_val & 0xff;
-			indir[j++] = (hlut_val >> 8) & 0xff;
-			indir[j++] = (hlut_val >> 16) & 0xff;
-			indir[j++] = (hlut_val >> 24) & 0xff;
-		}
-	}
-	return 0;
+	seed = key;
+
+	lut = kzalloc(I40EVF_HLUT_ARRAY_SIZE, GFP_KERNEL);
+	if (!lut)
+		return -ENOMEM;
+
+	ret = i40evf_get_rss(vsi, seed, lut, I40EVF_HLUT_ARRAY_SIZE);
+	if (ret)
+		goto out;
+
+	/* Each 32 bits pointed by 'indir' is stored with a lut entry */
+	for (i = 0; i < I40EVF_HLUT_ARRAY_SIZE; i++)
+		indir[i] = (u32)lut[i];
+
+out:
+	kfree(lut);
+
+	return ret;
 }
 
 /**
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_main.c b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
index 63d7e470d17c..68251fa081ec 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_main.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
@@ -1299,6 +1299,84 @@ static int i40evf_config_rss_reg(struct i40e_vsi *vsi, const u8 *seed,
 	return 0;
 }
 
+/**
+ *  * i40evf_get_rss_aq - Get RSS keys and lut by using AQ commands
+ *  @vsi: Pointer to vsi structure
+ *  @seed: RSS hash seed
+ *  @lut: Lookup table
+ *  @lut_size: Lookup table size
+ *
+ *  Return 0 on success, negative on failure
+ **/
+static int i40evf_get_rss_aq(struct i40e_vsi *vsi, const u8 *seed,
+			     u8 *lut, u16 lut_size)
+{
+	struct i40evf_adapter *adapter = vsi->back;
+	struct i40e_hw *hw = &adapter->hw;
+	int ret = 0;
+
+	if (seed) {
+		ret = i40evf_aq_get_rss_key(hw, vsi->id,
+			(struct i40e_aqc_get_set_rss_key_data *)seed);
+		if (ret) {
+			dev_err(&adapter->pdev->dev,
+				"Cannot get RSS key, err %s aq_err %s\n",
+				i40evf_stat_str(hw, ret),
+				i40evf_aq_str(hw, hw->aq.asq_last_status));
+			return ret;
+		}
+	}
+
+	if (lut) {
+		ret = i40evf_aq_get_rss_lut(hw, vsi->id, seed, lut, lut_size);
+		if (ret) {
+			dev_err(&adapter->pdev->dev,
+				"Cannot get RSS lut, err %s aq_err %s\n",
+				i40evf_stat_str(hw, ret),
+				i40evf_aq_str(hw, hw->aq.asq_last_status));
+			return ret;
+		}
+	}
+
+	return ret;
+}
+
+/**
+ *  * i40evf_get_rss_reg - Get RSS keys and lut by reading registers
+ *  @vsi: Pointer to vsi structure
+ *  @seed: RSS hash seed
+ *  @lut: Lookup table
+ *  @lut_size: Lookup table size
+ *
+ *  Returns 0 on success, negative on failure
+ **/
+static int i40evf_get_rss_reg(struct i40e_vsi *vsi, const u8 *seed,
+			      const u8 *lut, u16 lut_size)
+{
+	struct i40evf_adapter *adapter = vsi->back;
+	struct i40e_hw *hw = &adapter->hw;
+	u16 i;
+
+	if (seed) {
+		u32 *seed_dw = (u32 *)seed;
+
+		for (i = 0; i <= I40E_VFQF_HKEY_MAX_INDEX; i++)
+			seed_dw[i] = rd32(hw, I40E_VFQF_HKEY(i));
+	}
+
+	if (lut) {
+		u32 *lut_dw = (u32 *)lut;
+
+		if (lut_size != I40EVF_HLUT_ARRAY_SIZE)
+			return -EINVAL;
+
+		for (i = 0; i <= I40E_VFQF_HLUT_MAX_INDEX; i++)
+			lut_dw[i] = rd32(hw, I40E_VFQF_HLUT(i));
+	}
+
+	return 0;
+}
+
 /**
  * i40evf_config_rss - Configure RSS keys and lut
  * @vsi: Pointer to vsi structure
@@ -1319,6 +1397,25 @@ int i40evf_config_rss(struct i40e_vsi *vsi, const u8 *seed,
 		return i40evf_config_rss_reg(vsi, seed, lut, lut_size);
 }
 
+/**
+ * i40evf_get_rss - Get RSS keys and lut
+ * @vsi: Pointer to vsi structure
+ * @seed: RSS hash seed
+ * @lut: Lookup table
+ * @lut_size: Lookup table size
+ *
+ * Returns 0 on success, negative on failure
+ **/
+int i40evf_get_rss(struct i40e_vsi *vsi, const u8 *seed, u8 *lut, u16 lut_size)
+{
+	struct i40evf_adapter *adapter = vsi->back;
+
+	if (RSS_AQ(adapter))
+		return i40evf_get_rss_aq(vsi, seed, lut, lut_size);
+	else
+		return i40evf_get_rss_reg(vsi, seed, lut, lut_size);
+}
+
 /**
  * i40evf_fill_rss_lut - Fill the lut with default values
  * @lut: Lookup table to be filled with

From 66f9af855a4df146713866f6076207edea44a4f5 Mon Sep 17 00:00:00 2001
From: Helin Zhang <helin.zhang@intel.com>
Date: Mon, 26 Oct 2015 19:44:34 -0400
Subject: [PATCH 08/17] i40evf: add new fields to store user configuration of
 RSS

This patch adds new fields to i40e_vsi to store user configured
RSS config data and code to use it.

Change-ID: Ic5d3db8d9df52182b560248f8cdca9c5c7546879
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40evf/i40evf.h    |  2 +
 .../ethernet/intel/i40evf/i40evf_ethtool.c    | 31 +++++++++-------
 .../net/ethernet/intel/i40evf/i40evf_main.c   | 37 +++++++++++++++++--
 3 files changed, 54 insertions(+), 16 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40evf/i40evf.h b/drivers/net/ethernet/intel/i40evf/i40evf.h
index 5f2e61dc20ec..090c6048f0df 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf.h
+++ b/drivers/net/ethernet/intel/i40evf/i40evf.h
@@ -67,6 +67,8 @@ struct i40e_vsi {
 	u16 rx_itr_setting;
 	u16 tx_itr_setting;
 	u16 qs_handle;
+	u8 *rss_hkey_user; /* User configured hash keys */
+	u8 *rss_lut_user;  /* User configured lookup table entries */
 };
 
 /* How many Rx Buffers do we bundle into one write to the hardware ? */
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c b/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
index a5209707bcf3..966863c93dc4 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
@@ -678,9 +678,7 @@ static int i40evf_set_rxfh(struct net_device *netdev, const u32 *indir,
 {
 	struct i40evf_adapter *adapter = netdev_priv(netdev);
 	struct i40e_vsi *vsi = &adapter->vsi;
-	u8 seed_def[I40EVF_HKEY_ARRAY_SIZE];
-	u8 *seed = NULL, *lut;
-	int ret;
+	u8 *seed = NULL;
 	u16 i;
 
 	/* We do not allow change in unsupported parameters */
@@ -691,21 +689,28 @@ static int i40evf_set_rxfh(struct net_device *netdev, const u32 *indir,
 		return 0;
 
 	if (key) {
-		memcpy(seed_def, key, I40EVF_HKEY_ARRAY_SIZE);
-		seed = seed_def;
+		if (!vsi->rss_hkey_user) {
+			vsi->rss_hkey_user = kzalloc(I40EVF_HKEY_ARRAY_SIZE,
+						     GFP_KERNEL);
+			if (!vsi->rss_hkey_user)
+				return -ENOMEM;
+		}
+		memcpy(vsi->rss_hkey_user, key, I40EVF_HKEY_ARRAY_SIZE);
+		seed = vsi->rss_hkey_user;
+	}
+	if (!vsi->rss_lut_user) {
+		vsi->rss_lut_user = kzalloc(I40EVF_HLUT_ARRAY_SIZE,
+					    GFP_KERNEL);
+		if (!vsi->rss_lut_user)
+			return -ENOMEM;
 	}
-	lut = kzalloc(I40EVF_HLUT_ARRAY_SIZE, GFP_KERNEL);
-	if (!lut)
-		return -ENOMEM;
 
 	/* Each 32 bits pointed by 'indir' is stored with a lut entry */
 	for (i = 0; i < I40EVF_HLUT_ARRAY_SIZE; i++)
-		lut[i] = (u8)(indir[i]);
+		vsi->rss_lut_user[i] = (u8)(indir[i]);
 
-	ret = i40evf_config_rss(vsi, seed, lut, I40EVF_HLUT_ARRAY_SIZE);
-	kfree(lut);
-
-	return ret;
+	return i40evf_config_rss(vsi, seed, vsi->rss_lut_user,
+				 I40EVF_HLUT_ARRAY_SIZE);
 }
 
 static const struct ethtool_ops i40evf_ethtool_ops = {
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_main.c b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
index 68251fa081ec..1cd5d5d1b660 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_main.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
@@ -1453,9 +1453,21 @@ static int i40evf_init_rss(struct i40evf_adapter *adapter)
 	lut = kzalloc(I40EVF_HLUT_ARRAY_SIZE, GFP_KERNEL);
 	if (!lut)
 		return -ENOMEM;
-	i40evf_fill_rss_lut(lut, I40EVF_HLUT_ARRAY_SIZE,
-			    adapter->num_active_queues);
-	netdev_rss_key_fill((void *)seed, I40EVF_HKEY_ARRAY_SIZE);
+
+	/* Use user configured lut if there is one, otherwise use default */
+	if (vsi->rss_lut_user)
+		memcpy(lut, vsi->rss_lut_user, I40EVF_HLUT_ARRAY_SIZE);
+	else
+		i40evf_fill_rss_lut(lut, I40EVF_HLUT_ARRAY_SIZE,
+				    adapter->num_active_queues);
+
+	/* Use user configured hash key if there is one, otherwise
+	 * user default.
+	 */
+	if (vsi->rss_hkey_user)
+		memcpy(seed, vsi->rss_hkey_user, I40EVF_HKEY_ARRAY_SIZE);
+	else
+		netdev_rss_key_fill((void *)seed, I40EVF_HKEY_ARRAY_SIZE);
 	ret = i40evf_config_rss(vsi, seed, lut, I40EVF_HLUT_ARRAY_SIZE);
 	kfree(lut);
 
@@ -1582,6 +1594,22 @@ err_set_interrupt:
 	return err;
 }
 
+/**
+ * i40evf_clear_rss_config_user - Clear user configurations of RSS
+ * @vsi: Pointer to VSI structure
+ **/
+static void i40evf_clear_rss_config_user(struct i40e_vsi *vsi)
+{
+	if (!vsi)
+		return;
+
+	kfree(vsi->rss_hkey_user);
+	vsi->rss_hkey_user = NULL;
+
+	kfree(vsi->rss_lut_user);
+	vsi->rss_lut_user = NULL;
+}
+
 /**
  * i40evf_watchdog_timer - Periodic call-back timer
  * @data: pointer to adapter disguised as unsigned long
@@ -2770,6 +2798,9 @@ static void i40evf_remove(struct pci_dev *pdev)
 
 	flush_scheduled_work();
 
+	/* Clear user configurations for RSS */
+	i40evf_clear_rss_config_user(&adapter->vsi);
+
 	if (hw->aq.asq.count)
 		i40evf_shutdown_adminq(hw);
 

From 2b2426a760764017cb53ffae128c87b7808dcf11 Mon Sep 17 00:00:00 2001
From: Carolyn Wyborny <carolyn.wyborny@intel.com>
Date: Mon, 26 Oct 2015 19:44:35 -0400
Subject: [PATCH 09/17] i40e: Update error messaging

This patch fixes an issue where adminq init failures always provided
a message that NVM was newer than expected.  This is not always the
case for init_adminq failures. Without this patch, if adminq init
fails for any reason, newer NVM message would be given.  This
problem is fixed by adding  a check for that specific error
condition and a different hopefully helpful message otherwise.

Change-ID: Iaeaebee4e398989eae40bb70f943ab66a3a521a5
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 16 ++++++++++------
 1 file changed, 10 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 491c82fe6467..508cf9a52522 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -10471,6 +10471,16 @@ static int i40e_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	pf->hw.fc.requested_mode = I40E_FC_NONE;
 
 	err = i40e_init_adminq(hw);
+	if (err) {
+		if (err == I40E_ERR_FIRMWARE_API_VERSION)
+			dev_info(&pdev->dev,
+				 "The driver for the device stopped because the NVM image is newer than expected. You must install the most recent version of the network driver.\n");
+		else
+			dev_info(&pdev->dev,
+				 "The driver for the device stopped because the device firmware failed to init. Try updating your NVM image.\n");
+
+		goto err_pf_reset;
+	}
 
 	/* provide nvm, fw, api versions */
 	dev_info(&pdev->dev, "fw %d.%d.%05d api %d.%d nvm %s\n",
@@ -10478,12 +10488,6 @@ static int i40e_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 		 hw->aq.api_maj_ver, hw->aq.api_min_ver,
 		 i40e_nvm_version_str(hw));
 
-	if (err) {
-		dev_info(&pdev->dev,
-			 "The driver for the device stopped because the NVM image is newer than expected. You must install the most recent version of the network driver.\n");
-		goto err_pf_reset;
-	}
-
 	if (hw->aq.api_maj_ver == I40E_FW_API_VERSION_MAJOR &&
 	    hw->aq.api_min_ver > I40E_FW_API_VERSION_MINOR)
 		dev_info(&pdev->dev,

From e36b0b111bcc651ce030e88e042f83b015dd04b8 Mon Sep 17 00:00:00 2001
From: Helin Zhang <helin.zhang@intel.com>
Date: Mon, 26 Oct 2015 19:44:36 -0400
Subject: [PATCH 10/17] i40e: fix confusing message

This patch fixes the confusing kernel message of enabled RSS size,
by reporting it together with the hardware maximum RSS size.

Change-ID: I64864dbfbc13beccc180a7871680def1f3d5a339
Signed-off-by: Helin Zhang <helin.zhang@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 508cf9a52522..4e9d6e564861 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -8110,7 +8110,8 @@ int i40e_reconfig_rss_queues(struct i40e_pf *pf, int queue_count)
 
 		i40e_pf_config_rss(pf);
 	}
-	dev_info(&pf->pdev->dev, "RSS count:  %d\n", pf->alloc_rss_size);
+	dev_info(&pf->pdev->dev, "RSS count/HW max RSS count:  %d/%d\n",
+		 pf->alloc_rss_size, pf->rss_size_max);
 	return pf->alloc_rss_size;
 }
 

From e7ffb72d65c784ba445ac849bd46d2533365fae7 Mon Sep 17 00:00:00 2001
From: Mitch Williams <mitch.a.williams@intel.com>
Date: Mon, 26 Oct 2015 19:44:37 -0400
Subject: [PATCH 11/17] i40e: make error message more useful

If we get an invalid message from a VF, we should tell the user which VF
is being naughty, rather than making them guess.

Change-ID: I9252cef7baea3d8584043ed6ff12619a94e2f99c
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
index 9c54ca2bd04d..26f247d5cad3 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
@@ -1094,8 +1094,8 @@ static int i40e_vc_send_msg_to_vf(struct i40e_vf *vf, u32 v_opcode,
 	/* single place to detect unsuccessful return values */
 	if (v_retval) {
 		vf->num_invalid_msgs++;
-		dev_err(&pf->pdev->dev, "Failed opcode %d Error: %d\n",
-			v_opcode, v_retval);
+		dev_err(&pf->pdev->dev, "VF %d failed opcode %d, error: %d\n",
+			vf->vf_id, v_opcode, v_retval);
 		if (vf->num_invalid_msgs >
 		    I40E_DEFAULT_NUM_INVALID_MSGS_ALLOWED) {
 			dev_err(&pf->pdev->dev,

From e743072fd16e6e772d5b5e4aef38873946f9a3e0 Mon Sep 17 00:00:00 2001
From: Mitch Williams <mitch.a.williams@intel.com>
Date: Mon, 26 Oct 2015 19:44:38 -0400
Subject: [PATCH 12/17] i40evf: quoth the VF driver, Nevermore

If, upon a midnight dreary, the PF returns ERR_PARAM when the VF is
requesting resources, that's fatal. Either the firmware or NVM is badly,
badly misconfigured, or this VF has been disabled due to a previous VF
driver sending a bunch of bogus messages.

Either way, there is no recovery from this. Don't ponder weak and weary,
just quit.

Change-ID: I09d9f16cc4ee7fec3b57646a289d33838c1c5bf5
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40evf/i40evf_main.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_main.c b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
index 1cd5d5d1b660..8476fc548f0f 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_main.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
@@ -2435,6 +2435,14 @@ static void i40evf_init_task(struct work_struct *work)
 		if (err == I40E_ERR_ADMIN_QUEUE_NO_WORK) {
 			err = i40evf_send_vf_config_msg(adapter);
 			goto err;
+		} else if (err == I40E_ERR_PARAM) {
+			/* We only get ERR_PARAM if the device is in a very bad
+			 * state or if we've been disabled for previous bad
+			 * behavior. Either way, we're done now.
+			 */
+			i40evf_shutdown_adminq(hw);
+			dev_err(&pdev->dev, "Unable to get VF config due to PF error condition, not retrying\n");
+			return;
 		}
 		if (err) {
 			dev_err(&pdev->dev, "Unable to get VF config (%d)\n",

From 7d96ba1a8b7c36d1f4a04dc40e031b6f8c677496 Mon Sep 17 00:00:00 2001
From: Mitch Williams <mitch.a.williams@intel.com>
Date: Mon, 26 Oct 2015 19:44:39 -0400
Subject: [PATCH 13/17] i40evf: allocate queue vectors dynamically

Change the queue_vector array from a statically-sized member of the
adapter structure to a dynamically-allocated and -sized array.

This reduces the size of the adapter structure, and allows us to support
any number of queue vectors in the future without changing the code.

Change-ID: I08dc622cb2f2ad01e832e51c1ad9b86524730693
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40evf/i40evf.h    |  5 +--
 .../ethernet/intel/i40evf/i40evf_ethtool.c    |  2 +-
 .../net/ethernet/intel/i40evf/i40evf_main.c   | 38 +++++++++----------
 .../ethernet/intel/i40evf/i40evf_virtchnl.c   |  2 +-
 4 files changed, 21 insertions(+), 26 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40evf/i40evf.h b/drivers/net/ethernet/intel/i40evf/i40evf.h
index 090c6048f0df..bd185add7948 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf.h
+++ b/drivers/net/ethernet/intel/i40evf/i40evf.h
@@ -145,9 +145,6 @@ struct i40e_q_vector {
 #define OTHER_VECTOR 1
 #define NONQ_VECS (OTHER_VECTOR)
 
-#define MAX_MSIX_Q_VECTORS 4
-#define MAX_MSIX_COUNT 5
-
 #define MIN_MSIX_Q_VECTORS 1
 #define MIN_MSIX_COUNT (MIN_MSIX_Q_VECTORS + NONQ_VECS)
 
@@ -193,7 +190,7 @@ struct i40evf_adapter {
 	struct work_struct reset_task;
 	struct work_struct adminq_task;
 	struct delayed_work init_task;
-	struct i40e_q_vector *q_vector[MAX_MSIX_Q_VECTORS];
+	struct i40e_q_vector *q_vectors;
 	struct list_head vlan_filter_list;
 	char misc_vector_name[IFNAMSIZ + 9];
 	int num_active_queues;
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c b/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
index 966863c93dc4..90c5110bb6de 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
@@ -351,7 +351,7 @@ static int i40evf_set_coalesce(struct net_device *netdev,
 		vsi->tx_itr_setting &= ~I40E_ITR_DYNAMIC;
 
 	for (i = 0; i < adapter->num_msix_vectors - NONQ_VECS; i++) {
-		q_vector = adapter->q_vector[i];
+		q_vector = &adapter->q_vectors[i];
 		q_vector->rx.itr = ITR_TO_REG(vsi->rx_itr_setting);
 		wr32(hw, I40E_VFINT_ITRN1(0, i), q_vector->rx.itr);
 		q_vector->tx.itr = ITR_TO_REG(vsi->tx_itr_setting);
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_main.c b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
index 8476fc548f0f..129c274c3673 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_main.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
@@ -347,7 +347,7 @@ static irqreturn_t i40evf_msix_clean_rings(int irq, void *data)
 static void
 i40evf_map_vector_to_rxq(struct i40evf_adapter *adapter, int v_idx, int r_idx)
 {
-	struct i40e_q_vector *q_vector = adapter->q_vector[v_idx];
+	struct i40e_q_vector *q_vector = &adapter->q_vectors[v_idx];
 	struct i40e_ring *rx_ring = adapter->rx_rings[r_idx];
 
 	rx_ring->q_vector = q_vector;
@@ -368,7 +368,7 @@ i40evf_map_vector_to_rxq(struct i40evf_adapter *adapter, int v_idx, int r_idx)
 static void
 i40evf_map_vector_to_txq(struct i40evf_adapter *adapter, int v_idx, int t_idx)
 {
-	struct i40e_q_vector *q_vector = adapter->q_vector[v_idx];
+	struct i40e_q_vector *q_vector = &adapter->q_vectors[v_idx];
 	struct i40e_ring *tx_ring = adapter->tx_rings[t_idx];
 
 	tx_ring->q_vector = q_vector;
@@ -464,7 +464,7 @@ static void i40evf_netpoll(struct net_device *netdev)
 		return;
 
 	for (i = 0; i < q_vectors; i++)
-		i40evf_msix_clean_rings(0, adapter->q_vector[i]);
+		i40evf_msix_clean_rings(0, &adapter->q_vectors[i]);
 }
 
 #endif
@@ -486,7 +486,7 @@ i40evf_request_traffic_irqs(struct i40evf_adapter *adapter, char *basename)
 	q_vectors = adapter->num_msix_vectors - NONQ_VECS;
 
 	for (vector = 0; vector < q_vectors; vector++) {
-		struct i40e_q_vector *q_vector = adapter->q_vector[vector];
+		struct i40e_q_vector *q_vector = &adapter->q_vectors[vector];
 
 		if (q_vector->tx.ring && q_vector->rx.ring) {
 			snprintf(q_vector->name, sizeof(q_vector->name) - 1,
@@ -531,7 +531,7 @@ free_queue_irqs:
 			adapter->msix_entries[vector + NONQ_VECS].vector,
 			NULL);
 		free_irq(adapter->msix_entries[vector + NONQ_VECS].vector,
-			 adapter->q_vector[vector]);
+			 &adapter->q_vectors[vector]);
 	}
 	return err;
 }
@@ -581,7 +581,7 @@ static void i40evf_free_traffic_irqs(struct i40evf_adapter *adapter)
 		irq_set_affinity_hint(adapter->msix_entries[i+1].vector,
 				      NULL);
 		free_irq(adapter->msix_entries[i+1].vector,
-			 adapter->q_vector[i]);
+			 &adapter->q_vectors[i]);
 	}
 }
 
@@ -953,7 +953,7 @@ static void i40evf_napi_enable_all(struct i40evf_adapter *adapter)
 	for (q_idx = 0; q_idx < q_vectors; q_idx++) {
 		struct napi_struct *napi;
 
-		q_vector = adapter->q_vector[q_idx];
+		q_vector = &adapter->q_vectors[q_idx];
 		napi = &q_vector->napi;
 		napi_enable(napi);
 	}
@@ -970,7 +970,7 @@ static void i40evf_napi_disable_all(struct i40evf_adapter *adapter)
 	int q_vectors = adapter->num_msix_vectors - NONQ_VECS;
 
 	for (q_idx = 0; q_idx < q_vectors; q_idx++) {
-		q_vector = adapter->q_vector[q_idx];
+		q_vector = &adapter->q_vectors[q_idx];
 		napi_disable(&q_vector->napi);
 	}
 }
@@ -1483,21 +1483,22 @@ static int i40evf_init_rss(struct i40evf_adapter *adapter)
  **/
 static int i40evf_alloc_q_vectors(struct i40evf_adapter *adapter)
 {
-	int q_idx, num_q_vectors;
+	int q_idx = 0, num_q_vectors;
 	struct i40e_q_vector *q_vector;
 
 	num_q_vectors = adapter->num_msix_vectors - NONQ_VECS;
+	adapter->q_vectors = kzalloc(sizeof(*q_vector) * num_q_vectors,
+				     GFP_KERNEL);
+	if (!adapter->q_vectors)
+		goto err_out;
 
 	for (q_idx = 0; q_idx < num_q_vectors; q_idx++) {
-		q_vector = kzalloc(sizeof(*q_vector), GFP_KERNEL);
-		if (!q_vector)
-			goto err_out;
+		q_vector = &adapter->q_vectors[q_idx];
 		q_vector->adapter = adapter;
 		q_vector->vsi = &adapter->vsi;
 		q_vector->v_idx = q_idx;
 		netif_napi_add(adapter->netdev, &q_vector->napi,
 			       i40evf_napi_poll, NAPI_POLL_WEIGHT);
-		adapter->q_vector[q_idx] = q_vector;
 	}
 
 	return 0;
@@ -1505,11 +1506,10 @@ static int i40evf_alloc_q_vectors(struct i40evf_adapter *adapter)
 err_out:
 	while (q_idx) {
 		q_idx--;
-		q_vector = adapter->q_vector[q_idx];
+		q_vector = &adapter->q_vectors[q_idx];
 		netif_napi_del(&q_vector->napi);
-		kfree(q_vector);
-		adapter->q_vector[q_idx] = NULL;
 	}
+	kfree(adapter->q_vectors);
 	return -ENOMEM;
 }
 
@@ -1530,13 +1530,11 @@ static void i40evf_free_q_vectors(struct i40evf_adapter *adapter)
 	napi_vectors = adapter->num_active_queues;
 
 	for (q_idx = 0; q_idx < num_q_vectors; q_idx++) {
-		struct i40e_q_vector *q_vector = adapter->q_vector[q_idx];
-
-		adapter->q_vector[q_idx] = NULL;
+		struct i40e_q_vector *q_vector = &adapter->q_vectors[q_idx];
 		if (q_idx < napi_vectors)
 			netif_napi_del(&q_vector->napi);
-		kfree(q_vector);
 	}
+	kfree(adapter->q_vectors);
 }
 
 /**
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_virtchnl.c b/drivers/net/ethernet/intel/i40evf/i40evf_virtchnl.c
index 46b051684908..24b3af31b5a2 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_virtchnl.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_virtchnl.c
@@ -360,7 +360,7 @@ void i40evf_map_queues(struct i40evf_adapter *adapter)
 	vimi->num_vectors = adapter->num_msix_vectors;
 	/* Queue vectors first */
 	for (v_idx = 0; v_idx < q_vectors; v_idx++) {
-		q_vector = adapter->q_vector[v_idx];
+		q_vector = adapter->q_vectors + v_idx;
 		vimi->vecmap[v_idx].vsi_id = adapter->vsi_res->vsi_id;
 		vimi->vecmap[v_idx].vector_id = v_idx + NONQ_VECS;
 		vimi->vecmap[v_idx].txq_map = q_vector->ring_mask;

From 0dd438d8ad5db85c1a805ce2244252e37d5586a0 Mon Sep 17 00:00:00 2001
From: Mitch Williams <mitch.a.williams@intel.com>
Date: Mon, 26 Oct 2015 19:44:40 -0400
Subject: [PATCH 14/17] i40evf: allocate ring structs dynamically

Instead of awkwardly keeping a fixed array of pointers in the adapter
struct and then allocating ring structs individually, just keep a single
pointer and allocate a single blob for the arrays. This simplifies code,
shrinks the adapter structure, and future-proofs the driver by not
limiting the number of rings we can handle.

Change-ID: I31334ff911a6474954232cfe4bc98ccca3c769ff
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40evf/i40e_txrx.c |  2 +-
 drivers/net/ethernet/intel/i40evf/i40evf.h    |  4 +-
 .../ethernet/intel/i40evf/i40evf_ethtool.c    |  8 +--
 .../net/ethernet/intel/i40evf/i40evf_main.c   | 57 +++++++++----------
 .../ethernet/intel/i40evf/i40evf_virtchnl.c   | 10 ++--
 5 files changed, 40 insertions(+), 41 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
index 0f1a4ef5e3be..11561b648417 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
@@ -2063,7 +2063,7 @@ out_drop:
 netdev_tx_t i40evf_xmit_frame(struct sk_buff *skb, struct net_device *netdev)
 {
 	struct i40evf_adapter *adapter = netdev_priv(netdev);
-	struct i40e_ring *tx_ring = adapter->tx_rings[skb->queue_mapping];
+	struct i40e_ring *tx_ring = &adapter->tx_rings[skb->queue_mapping];
 
 	/* hardware can't handle really short frames, hardware padding works
 	 * beyond this point
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf.h b/drivers/net/ethernet/intel/i40evf/i40evf.h
index bd185add7948..a6318c421695 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf.h
+++ b/drivers/net/ethernet/intel/i40evf/i40evf.h
@@ -196,13 +196,13 @@ struct i40evf_adapter {
 	int num_active_queues;
 
 	/* TX */
-	struct i40e_ring *tx_rings[I40E_MAX_VSI_QP];
+	struct i40e_ring *tx_rings;
 	u32 tx_timeout_count;
 	struct list_head mac_filter_list;
 	u32 tx_desc_count;
 
 	/* RX */
-	struct i40e_ring *rx_rings[I40E_MAX_VSI_QP];
+	struct i40e_ring *rx_rings;
 	u64 hw_csum_rx_error;
 	u32 rx_desc_count;
 	int num_msix_vectors;
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c b/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
index 90c5110bb6de..a4c9feb589e7 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_ethtool.c
@@ -121,12 +121,12 @@ static void i40evf_get_ethtool_stats(struct net_device *netdev,
 		data[i] =  *(u64 *)p;
 	}
 	for (j = 0; j < adapter->num_active_queues; j++) {
-		data[i++] = adapter->tx_rings[j]->stats.packets;
-		data[i++] = adapter->tx_rings[j]->stats.bytes;
+		data[i++] = adapter->tx_rings[j].stats.packets;
+		data[i++] = adapter->tx_rings[j].stats.bytes;
 	}
 	for (j = 0; j < adapter->num_active_queues; j++) {
-		data[i++] = adapter->rx_rings[j]->stats.packets;
-		data[i++] = adapter->rx_rings[j]->stats.bytes;
+		data[i++] = adapter->rx_rings[j].stats.packets;
+		data[i++] = adapter->rx_rings[j].stats.bytes;
 	}
 }
 
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_main.c b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
index 129c274c3673..af14160524c5 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_main.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
@@ -348,7 +348,7 @@ static void
 i40evf_map_vector_to_rxq(struct i40evf_adapter *adapter, int v_idx, int r_idx)
 {
 	struct i40e_q_vector *q_vector = &adapter->q_vectors[v_idx];
-	struct i40e_ring *rx_ring = adapter->rx_rings[r_idx];
+	struct i40e_ring *rx_ring = &adapter->rx_rings[r_idx];
 
 	rx_ring->q_vector = q_vector;
 	rx_ring->next = q_vector->rx.ring;
@@ -369,7 +369,7 @@ static void
 i40evf_map_vector_to_txq(struct i40evf_adapter *adapter, int v_idx, int t_idx)
 {
 	struct i40e_q_vector *q_vector = &adapter->q_vectors[v_idx];
-	struct i40e_ring *tx_ring = adapter->tx_rings[t_idx];
+	struct i40e_ring *tx_ring = &adapter->tx_rings[t_idx];
 
 	tx_ring->q_vector = q_vector;
 	tx_ring->next = q_vector->tx.ring;
@@ -610,7 +610,7 @@ static void i40evf_configure_tx(struct i40evf_adapter *adapter)
 	int i;
 
 	for (i = 0; i < adapter->num_active_queues; i++)
-		adapter->tx_rings[i]->tail = hw->hw_addr + I40E_QTX_TAIL1(i);
+		adapter->tx_rings[i].tail = hw->hw_addr + I40E_QTX_TAIL1(i);
 }
 
 /**
@@ -655,8 +655,8 @@ static void i40evf_configure_rx(struct i40evf_adapter *adapter)
 	}
 
 	for (i = 0; i < adapter->num_active_queues; i++) {
-		adapter->rx_rings[i]->tail = hw->hw_addr + I40E_QRX_TAIL1(i);
-		adapter->rx_rings[i]->rx_buf_len = rx_buf_len;
+		adapter->rx_rings[i].tail = hw->hw_addr + I40E_QRX_TAIL1(i);
+		adapter->rx_rings[i].rx_buf_len = rx_buf_len;
 	}
 }
 
@@ -991,7 +991,7 @@ static void i40evf_configure(struct i40evf_adapter *adapter)
 	adapter->aq_required |= I40EVF_FLAG_AQ_CONFIGURE_QUEUES;
 
 	for (i = 0; i < adapter->num_active_queues; i++) {
-		struct i40e_ring *ring = adapter->rx_rings[i];
+		struct i40e_ring *ring = &adapter->rx_rings[i];
 
 		i40evf_alloc_rx_buffers_1buf(ring, ring->count);
 		ring->next_to_use = ring->count - 1;
@@ -1111,16 +1111,10 @@ i40evf_acquire_msix_vectors(struct i40evf_adapter *adapter, int vectors)
  **/
 static void i40evf_free_queues(struct i40evf_adapter *adapter)
 {
-	int i;
-
 	if (!adapter->vsi_res)
 		return;
-	for (i = 0; i < adapter->num_active_queues; i++) {
-		if (adapter->tx_rings[i])
-			kfree_rcu(adapter->tx_rings[i], rcu);
-		adapter->tx_rings[i] = NULL;
-		adapter->rx_rings[i] = NULL;
-	}
+	kfree(adapter->tx_rings);
+	kfree(adapter->rx_rings);
 }
 
 /**
@@ -1135,13 +1129,20 @@ static int i40evf_alloc_queues(struct i40evf_adapter *adapter)
 {
 	int i;
 
+	adapter->tx_rings = kcalloc(adapter->num_active_queues,
+				    sizeof(struct i40e_ring), GFP_KERNEL);
+	if (!adapter->tx_rings)
+		goto err_out;
+	adapter->rx_rings = kcalloc(adapter->num_active_queues,
+				    sizeof(struct i40e_ring), GFP_KERNEL);
+	if (!adapter->rx_rings)
+		goto err_out;
+
 	for (i = 0; i < adapter->num_active_queues; i++) {
 		struct i40e_ring *tx_ring;
 		struct i40e_ring *rx_ring;
 
-		tx_ring = kzalloc(sizeof(*tx_ring) * 2, GFP_KERNEL);
-		if (!tx_ring)
-			goto err_out;
+		tx_ring = &adapter->tx_rings[i];
 
 		tx_ring->queue_index = i;
 		tx_ring->netdev = adapter->netdev;
@@ -1149,14 +1150,12 @@ static int i40evf_alloc_queues(struct i40evf_adapter *adapter)
 		tx_ring->count = adapter->tx_desc_count;
 		if (adapter->flags & I40E_FLAG_WB_ON_ITR_CAPABLE)
 			tx_ring->flags |= I40E_TXR_FLAGS_WB_ON_ITR;
-		adapter->tx_rings[i] = tx_ring;
 
-		rx_ring = &tx_ring[1];
+		rx_ring = &adapter->rx_rings[i];
 		rx_ring->queue_index = i;
 		rx_ring->netdev = adapter->netdev;
 		rx_ring->dev = &adapter->pdev->dev;
 		rx_ring->count = adapter->rx_desc_count;
-		adapter->rx_rings[i] = rx_ring;
 	}
 
 	return 0;
@@ -1487,7 +1486,7 @@ static int i40evf_alloc_q_vectors(struct i40evf_adapter *adapter)
 	struct i40e_q_vector *q_vector;
 
 	num_q_vectors = adapter->num_msix_vectors - NONQ_VECS;
-	adapter->q_vectors = kzalloc(sizeof(*q_vector) * num_q_vectors,
+	adapter->q_vectors = kcalloc(num_q_vectors, sizeof(*q_vector),
 				     GFP_KERNEL);
 	if (!adapter->q_vectors)
 		goto err_out;
@@ -2035,8 +2034,8 @@ void i40evf_free_all_tx_resources(struct i40evf_adapter *adapter)
 	int i;
 
 	for (i = 0; i < adapter->num_active_queues; i++)
-		if (adapter->tx_rings[i]->desc)
-			i40evf_free_tx_resources(adapter->tx_rings[i]);
+		if (adapter->tx_rings[i].desc)
+			i40evf_free_tx_resources(&adapter->tx_rings[i]);
 }
 
 /**
@@ -2054,8 +2053,8 @@ static int i40evf_setup_all_tx_resources(struct i40evf_adapter *adapter)
 	int i, err = 0;
 
 	for (i = 0; i < adapter->num_active_queues; i++) {
-		adapter->tx_rings[i]->count = adapter->tx_desc_count;
-		err = i40evf_setup_tx_descriptors(adapter->tx_rings[i]);
+		adapter->tx_rings[i].count = adapter->tx_desc_count;
+		err = i40evf_setup_tx_descriptors(&adapter->tx_rings[i]);
 		if (!err)
 			continue;
 		dev_err(&adapter->pdev->dev,
@@ -2081,8 +2080,8 @@ static int i40evf_setup_all_rx_resources(struct i40evf_adapter *adapter)
 	int i, err = 0;
 
 	for (i = 0; i < adapter->num_active_queues; i++) {
-		adapter->rx_rings[i]->count = adapter->rx_desc_count;
-		err = i40evf_setup_rx_descriptors(adapter->rx_rings[i]);
+		adapter->rx_rings[i].count = adapter->rx_desc_count;
+		err = i40evf_setup_rx_descriptors(&adapter->rx_rings[i]);
 		if (!err)
 			continue;
 		dev_err(&adapter->pdev->dev,
@@ -2103,8 +2102,8 @@ void i40evf_free_all_rx_resources(struct i40evf_adapter *adapter)
 	int i;
 
 	for (i = 0; i < adapter->num_active_queues; i++)
-		if (adapter->rx_rings[i]->desc)
-			i40evf_free_rx_resources(adapter->rx_rings[i]);
+		if (adapter->rx_rings[i].desc)
+			i40evf_free_rx_resources(&adapter->rx_rings[i]);
 }
 
 /**
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_virtchnl.c b/drivers/net/ethernet/intel/i40evf/i40evf_virtchnl.c
index 24b3af31b5a2..9b5557622657 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_virtchnl.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_virtchnl.c
@@ -255,19 +255,19 @@ void i40evf_configure_queues(struct i40evf_adapter *adapter)
 	for (i = 0; i < pairs; i++) {
 		vqpi->txq.vsi_id = vqci->vsi_id;
 		vqpi->txq.queue_id = i;
-		vqpi->txq.ring_len = adapter->tx_rings[i]->count;
-		vqpi->txq.dma_ring_addr = adapter->tx_rings[i]->dma;
+		vqpi->txq.ring_len = adapter->tx_rings[i].count;
+		vqpi->txq.dma_ring_addr = adapter->tx_rings[i].dma;
 		vqpi->txq.headwb_enabled = 1;
 		vqpi->txq.dma_headwb_addr = vqpi->txq.dma_ring_addr +
 		    (vqpi->txq.ring_len * sizeof(struct i40e_tx_desc));
 
 		vqpi->rxq.vsi_id = vqci->vsi_id;
 		vqpi->rxq.queue_id = i;
-		vqpi->rxq.ring_len = adapter->rx_rings[i]->count;
-		vqpi->rxq.dma_ring_addr = adapter->rx_rings[i]->dma;
+		vqpi->rxq.ring_len = adapter->rx_rings[i].count;
+		vqpi->rxq.dma_ring_addr = adapter->rx_rings[i].dma;
 		vqpi->rxq.max_pkt_size = adapter->netdev->mtu
 					+ ETH_HLEN + VLAN_HLEN + ETH_FCS_LEN;
-		vqpi->rxq.databuffer_size = adapter->rx_rings[i]->rx_buf_len;
+		vqpi->rxq.databuffer_size = adapter->rx_rings[i].rx_buf_len;
 		vqpi++;
 	}
 

From e1c227919548fb9d8ee771e715d2764f1d41e8e0 Mon Sep 17 00:00:00 2001
From: Catherine Sullivan <catherine.sullivan@intel.com>
Date: Mon, 26 Oct 2015 19:44:41 -0400
Subject: [PATCH 15/17] i40e/i40evf: Bump i40e version to 1.4.4 and i40evf to
 1.4.1

Bump.

Change-ID: I00ebbb2e5e5572f947502b8f6db4d94f666d6b14
Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40e/i40e_main.c     | 2 +-
 drivers/net/ethernet/intel/i40evf/i40evf_main.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 4e9d6e564861..7759703a40bc 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -39,7 +39,7 @@ static const char i40e_driver_string[] =
 
 #define DRV_VERSION_MAJOR 1
 #define DRV_VERSION_MINOR 4
-#define DRV_VERSION_BUILD 2
+#define DRV_VERSION_BUILD 4
 #define DRV_VERSION __stringify(DRV_VERSION_MAJOR) "." \
 	     __stringify(DRV_VERSION_MINOR) "." \
 	     __stringify(DRV_VERSION_BUILD)    DRV_KERN
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_main.c b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
index af14160524c5..7ef10424a67f 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_main.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
@@ -34,7 +34,7 @@ char i40evf_driver_name[] = "i40evf";
 static const char i40evf_driver_string[] =
 	"Intel(R) XL710/X710 Virtual Function Network Driver";
 
-#define DRV_VERSION "1.3.33"
+#define DRV_VERSION "1.4.1"
 const char i40evf_driver_version[] = DRV_VERSION;
 static const char i40evf_copyright[] =
 	"Copyright (c) 2013 - 2015 Intel Corporation.";

From 0e4425ed641f3eef67c892bc541949cd745a9ba9 Mon Sep 17 00:00:00 2001
From: Jesse Brandeburg <jesse.brandeburg@intel.com>
Date: Thu, 5 Nov 2015 17:01:01 -0800
Subject: [PATCH 16/17] i40e: fix: do not sleep in netdev_ops

The driver was being called by VLAN, bonding, teaming operations
that expected to be able to hold locks like rcu_read_lock().

This causes the driver to be held to the requirement to not sleep,
and was found by the kernel debug options for checking sleep
inside critical section, and the locking validator.

Change-ID: Ibc68c835f5ffa8ffe0638ffe910a66fc5649a7f7
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 44 ++++++++++-----------
 1 file changed, 20 insertions(+), 24 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 7759703a40bc..3119ee44b0bf 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -1552,9 +1552,11 @@ static int i40e_set_mac(struct net_device *netdev, void *p)
 		spin_unlock_bh(&vsi->mac_filter_list_lock);
 	}
 
-	i40e_sync_vsi_filters(vsi, false);
 	ether_addr_copy(netdev->dev_addr, addr->sa_data);
-
+	/* schedule our worker thread which will take care of
+	 * applying the new filter changes
+	 */
+	i40e_service_event_schedule(vsi->back);
 	return 0;
 }
 
@@ -2118,12 +2120,7 @@ int i40e_sync_vsi_filters(struct i40e_vsi *vsi, bool grab_rtnl)
 			 */
 			if (pf->cur_promisc != cur_promisc) {
 				pf->cur_promisc = cur_promisc;
-				if (grab_rtnl)
-					i40e_do_reset_safe(pf,
-						BIT(__I40E_PF_RESET_REQUESTED));
-				else
-					i40e_do_reset(pf,
-						BIT(__I40E_PF_RESET_REQUESTED));
+				set_bit(__I40E_PF_RESET_REQUESTED, &pf->state);
 			}
 		} else {
 			ret = i40e_aq_set_vsi_unicast_promiscuous(
@@ -2383,16 +2380,13 @@ int i40e_vsi_add_vlan(struct i40e_vsi *vsi, s16 vid)
 		}
 	}
 
-	/* Make sure to release before sync_vsi_filter because that
-	 * function will lock/unlock as necessary
-	 */
 	spin_unlock_bh(&vsi->mac_filter_list_lock);
 
-	if (test_bit(__I40E_DOWN, &vsi->back->state) ||
-	    test_bit(__I40E_RESET_RECOVERY_PENDING, &vsi->back->state))
-		return 0;
-
-	return i40e_sync_vsi_filters(vsi, false);
+	/* schedule our worker thread which will take care of
+	 * applying the new filter changes
+	 */
+	i40e_service_event_schedule(vsi->back);
+	return 0;
 }
 
 /**
@@ -2465,16 +2459,13 @@ int i40e_vsi_kill_vlan(struct i40e_vsi *vsi, s16 vid)
 		}
 	}
 
-	/* Make sure to release before sync_vsi_filter because that
-	 * function with lock/unlock as necessary
-	 */
 	spin_unlock_bh(&vsi->mac_filter_list_lock);
 
-	if (test_bit(__I40E_DOWN, &vsi->back->state) ||
-	    test_bit(__I40E_RESET_RECOVERY_PENDING, &vsi->back->state))
-		return 0;
-
-	return i40e_sync_vsi_filters(vsi, false);
+	/* schedule our worker thread which will take care of
+	 * applying the new filter changes
+	 */
+	i40e_service_event_schedule(vsi->back);
+	return 0;
 }
 
 /**
@@ -2717,6 +2708,11 @@ static void i40e_config_xps_tx_ring(struct i40e_ring *ring)
 		netif_set_xps_queue(ring->netdev, mask, ring->queue_index);
 		free_cpumask_var(mask);
 	}
+
+	/* schedule our worker thread which will take care of
+	 * applying the new filter changes
+	 */
+	i40e_service_event_schedule(vsi->back);
 }
 
 /**

From 17652c6336fdca0f48906f0a1d9b5ac77b29e5a7 Mon Sep 17 00:00:00 2001
From: Jesse Brandeburg <jesse.brandeburg@intel.com>
Date: Thu, 5 Nov 2015 17:01:02 -0800
Subject: [PATCH 17/17] i40e: remove unused argument

With the final edition of the patches to remove sleeps from
the driver's entry points, the grab_rtnl argument is no
longer needed, so partially revert the commit that added it.

Change-ID: Ib9778476242586cc9e58b670f5f48d415cb59003
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40e/i40e.h           |  2 +-
 drivers/net/ethernet/intel/i40e/i40e_debugfs.c   |  4 ++--
 drivers/net/ethernet/intel/i40e/i40e_main.c      | 16 +++++++++++-----
 .../net/ethernet/intel/i40e/i40e_virtchnl_pf.c   |  8 ++++----
 4 files changed, 18 insertions(+), 12 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
index d854a46bba1e..0b9537b37412 100644
--- a/drivers/net/ethernet/intel/i40e/i40e.h
+++ b/drivers/net/ethernet/intel/i40e/i40e.h
@@ -697,7 +697,7 @@ struct i40e_mac_filter *i40e_add_filter(struct i40e_vsi *vsi,
 					bool is_vf, bool is_netdev);
 void i40e_del_filter(struct i40e_vsi *vsi, u8 *macaddr, s16 vlan,
 		     bool is_vf, bool is_netdev);
-int i40e_sync_vsi_filters(struct i40e_vsi *vsi, bool grab_rtnl);
+int i40e_sync_vsi_filters(struct i40e_vsi *vsi);
 struct i40e_vsi *i40e_vsi_setup(struct i40e_pf *pf, u8 type,
 				u16 uplink, u32 param1);
 int i40e_vsi_release(struct i40e_vsi *vsi);
diff --git a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
index d1a91c8178d6..10744a698d6f 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
@@ -1138,7 +1138,7 @@ static ssize_t i40e_dbg_command_write(struct file *filp,
 		spin_lock_bh(&vsi->mac_filter_list_lock);
 		f = i40e_add_filter(vsi, ma, vlan, false, false);
 		spin_unlock_bh(&vsi->mac_filter_list_lock);
-		ret = i40e_sync_vsi_filters(vsi, true);
+		ret = i40e_sync_vsi_filters(vsi);
 		if (f && !ret)
 			dev_info(&pf->pdev->dev,
 				 "add macaddr: %pM vlan=%d added to VSI %d\n",
@@ -1177,7 +1177,7 @@ static ssize_t i40e_dbg_command_write(struct file *filp,
 		spin_lock_bh(&vsi->mac_filter_list_lock);
 		i40e_del_filter(vsi, ma, vlan, false, false);
 		spin_unlock_bh(&vsi->mac_filter_list_lock);
-		ret = i40e_sync_vsi_filters(vsi, true);
+		ret = i40e_sync_vsi_filters(vsi);
 		if (!ret)
 			dev_info(&pf->pdev->dev,
 				 "del macaddr: %pM vlan=%d removed from VSI %d\n",
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 3119ee44b0bf..9e6268b4295a 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -1859,13 +1859,12 @@ static void i40e_cleanup_add_list(struct list_head *add_list)
 /**
  * i40e_sync_vsi_filters - Update the VSI filter list to the HW
  * @vsi: ptr to the VSI
- * @grab_rtnl: whether RTNL needs to be grabbed
  *
  * Push any outstanding VSI filter changes through the AdminQ.
  *
  * Returns 0 or error value
  **/
-int i40e_sync_vsi_filters(struct i40e_vsi *vsi, bool grab_rtnl)
+int i40e_sync_vsi_filters(struct i40e_vsi *vsi)
 {
 	struct list_head tmp_del_list, tmp_add_list;
 	struct i40e_mac_filter *f, *ftmp, *fclone;
@@ -2169,8 +2168,15 @@ static void i40e_sync_filters_subtask(struct i40e_pf *pf)
 
 	for (v = 0; v < pf->num_alloc_vsi; v++) {
 		if (pf->vsi[v] &&
-		    (pf->vsi[v]->flags & I40E_VSI_FLAG_FILTER_CHANGED))
-			i40e_sync_vsi_filters(pf->vsi[v], true);
+		    (pf->vsi[v]->flags & I40E_VSI_FLAG_FILTER_CHANGED)) {
+			int ret = i40e_sync_vsi_filters(pf->vsi[v]);
+
+			if (ret) {
+				/* come back and try again later */
+				pf->flags |= I40E_FLAG_FILTER_SYNC;
+				break;
+			}
+		}
 	}
 }
 
@@ -9219,7 +9225,7 @@ int i40e_vsi_release(struct i40e_vsi *vsi)
 				f->is_vf, f->is_netdev);
 	spin_unlock_bh(&vsi->mac_filter_list_lock);
 
-	i40e_sync_vsi_filters(vsi, false);
+	i40e_sync_vsi_filters(vsi);
 
 	i40e_vsi_delete(vsi);
 	i40e_vsi_free_q_vectors(vsi);
diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
index 26f247d5cad3..819803c8e461 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
@@ -565,7 +565,7 @@ static int i40e_alloc_vsi_res(struct i40e_vf *vf, enum i40e_vsi_type type)
 	}
 
 	/* program mac filter */
-	ret = i40e_sync_vsi_filters(vsi, false);
+	ret = i40e_sync_vsi_filters(vsi);
 	if (ret)
 		dev_err(&pf->pdev->dev, "Unable to program ucast filters\n");
 
@@ -1633,7 +1633,7 @@ static int i40e_vc_add_mac_addr_msg(struct i40e_vf *vf, u8 *msg, u16 msglen)
 	spin_unlock_bh(&vsi->mac_filter_list_lock);
 
 	/* program the updated filter list */
-	if (i40e_sync_vsi_filters(vsi, false))
+	if (i40e_sync_vsi_filters(vsi))
 		dev_err(&pf->pdev->dev, "Unable to program VF %d MAC filters\n",
 			vf->vf_id);
 
@@ -1687,7 +1687,7 @@ static int i40e_vc_del_mac_addr_msg(struct i40e_vf *vf, u8 *msg, u16 msglen)
 	spin_unlock_bh(&vsi->mac_filter_list_lock);
 
 	/* program the updated filter list */
-	if (i40e_sync_vsi_filters(vsi, false))
+	if (i40e_sync_vsi_filters(vsi))
 		dev_err(&pf->pdev->dev, "Unable to program VF %d MAC filters\n",
 			vf->vf_id);
 
@@ -2102,7 +2102,7 @@ int i40e_ndo_set_vf_mac(struct net_device *netdev, int vf_id, u8 *mac)
 
 	dev_info(&pf->pdev->dev, "Setting MAC %pM on VF %d\n", mac, vf_id);
 	/* program mac filter */
-	if (i40e_sync_vsi_filters(vsi, false)) {
+	if (i40e_sync_vsi_filters(vsi)) {
 		dev_err(&pf->pdev->dev, "Unable to program ucast filters\n");
 		ret = -EIO;
 		goto error_param;