[net-next 00/16][pull request] 100GbE Intel Wired LAN Driver Updates 2019-02-22

[net-next 00/16][pull request] 100GbE Intel Wired LAN Driver Updates 2019-02-22

[Jeff Kirsher]

This series contains updates to the ice driver only.

Bruce adds the __always_unused attribute to a parameter to avoid
compiler warnings when using -Wunused-parameter.  Fixed unnecessary
type-casting and the use of sizeof().  Fix the allocation of structs
that have become memory hogs, so allocate them in heaps and fix all the
associated references.  Fixed the "possible" numeric overflow issues
that were caught with static analysis.

Maciej fixes the maximum MTU calculation by taking into account double
VLAN tagging amd ensure that the operations are done in the correct
order.

Victor fixes the supported node calculation, where we were not taking
into account if there is space to add the new VSI or intermediate node
above that layer, then it is not required to continue the calculation.
Added a check for a leaf node presence for a given VSI, which is needed
before removing a VSI.

Jake fixes an issue where the VSI list is shared, so simply removing a
VSI from the list will cause issues for the other users who reference
the list.  Since we also free the memory, this could lead to
segmentation faults.

Brett fixes an issue where driver unload could cause a system reboot
when intel_iommu=on parameter is set.  The issue is that we are not
clearing the CAUSE_ENA bit for the appropriate control queues register
when freeing the miscellaneous interrupt vector.

Mitch is so kind, he prevented spamming the VF with link messages when
the link status really has not changed.  Updates the driver to use the
absolute vector ID and not the per-PF vector ID for the VF MSIx vector
allocation.

Lukasz fixes the ethtool pause parameter for the ice driver, which was
originally based off the link status but is now based off the PHY
configuration.  This is to resolve an issue where pause parameters could
be set while link was down.

Jesse updates the string that reports statistics so the string does not
get modified at runtime and cause reports of string truncation.

The following are changes since commit e59d790959b48a42874fa2cb16475a621663f085:
  Merge branch 'net-phy-at803x-Update-delays-for-RGMII-modes'
and are available in the git repository at:
  git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queue 100GbE

Brett Creeley (2):
  ice: fix issue where host reboots on unload when iommu=on
  ice: only use the VF for ICE_VSI_VF in ice_vsi_release

Bruce Allan (4):
  ice: Mark extack argument as __always_unused
  ice: sizeof(<type>) should be avoided
  ice: fix stack hogs from struct ice_vsi_ctx structures
  ice: fix numeric overflow warning

Jacob Keller (1):
  ice: fix ice_remove_rule_internal vsi_list handling

Jesse Brandeburg (1):
  ice: fix overlong string, update stats output

Lukasz Czapnik (1):
  ice: Fix for FC get rx/tx pause params

Maciej Fijalkowski (1):
  ice: Fix the calculation of ICE_MAX_MTU

Mitch Williams (3):
  ice: don't spam VFs with link messages
  ice: clear VF ARQLEN register on reset
  ice: use absolute vector ID for VFs

Victor Raj (3):
  ice: Fix added in VSI supported nodes calc
  ice: flush Tx pipe on disable queue timeout
  ice: check for a leaf node presence

 drivers/net/ethernet/intel/ice/ice.h          |   2 +-
 drivers/net/ethernet/intel/ice/ice_common.c   |  21 ++-
 drivers/net/ethernet/intel/ice/ice_ethtool.c  | 132 +++++++++--------
 .../net/ethernet/intel/ice/ice_hw_autogen.h   |   1 +
 drivers/net/ethernet/intel/ice/ice_lib.c      | 112 +++++++++------
 drivers/net/ethernet/intel/ice/ice_main.c     | 133 ++++++++++++------
 drivers/net/ethernet/intel/ice/ice_nvm.c      |   7 +-
 drivers/net/ethernet/intel/ice/ice_sched.c    |  41 +++++-
 drivers/net/ethernet/intel/ice/ice_status.h   |   1 +
 drivers/net/ethernet/intel/ice/ice_switch.c   |  17 ++-
 drivers/net/ethernet/intel/ice/ice_txrx.c     |  26 ++--
 .../net/ethernet/intel/ice/ice_virtchnl_pf.c  |  70 +++++----
 12 files changed, 364 insertions(+), 199 deletions(-)

-- 
2.20.1

[net-next 01/16] ice: Mark extack argument as __always_unused

[Jeff Kirsher]

From: Bruce Allan <bruce.w.allan@intel.com>

Commit 87b0984ebfab ("net: Add extack argument to ndo_fdb_add()") in
net-next added an extended parameter to the .ndo_fdb_add op and changed
ice_fdb_add() accordingly. Update the function header and add the
__always_unused attribute to the new parameter to avoid -Wunused-parameter
warnings.

Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_main.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 48f033928aa2..9d266d754445 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -2435,11 +2435,12 @@ static void ice_set_rx_mode(struct net_device *netdev)
  * @addr: the MAC address entry being added
  * @vid: VLAN id
  * @flags: instructions from stack about fdb operation
+ * @extack: netlink extended ack
  */
-static int ice_fdb_add(struct ndmsg *ndm, struct nlattr __always_unused *tb[],
-		       struct net_device *dev, const unsigned char *addr,
-		       u16 vid, u16 flags,
-		       struct netlink_ext_ack *extack)
+static int
+ice_fdb_add(struct ndmsg *ndm, struct nlattr __always_unused *tb[],
+	    struct net_device *dev, const unsigned char *addr, u16 vid,
+	    u16 flags, struct netlink_ext_ack __always_unused *extack)
 {
 	int err;
 
-- 
2.20.1

[net-next 02/16] ice: Fix the calculation of ICE_MAX_MTU

[Jeff Kirsher]

From: Maciej Fijalkowski <maciej.fijalkowski@intel.com>

Currently ICE_MAX_MTU subtracts only ETH_HLEN from max frame size and
adds ETH_FCS_LEN and VLAN_HLEN, which is not what was intended.
The ETH_HLEN + ETH_FCS_LEN + VLAN_HLEN expression should be surrounded
with parentheses.

Wrap mentioned expression and take into account VLAN double tagging.

Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ice/ice.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/ice/ice.h b/drivers/net/ethernet/intel/ice/ice.h
index 55944e089558..89440775aea1 100644
--- a/drivers/net/ethernet/intel/ice/ice.h
+++ b/drivers/net/ethernet/intel/ice/ice.h
@@ -83,7 +83,7 @@ extern const char ice_drv_ver[];
 #define ICE_DFLT_NETIF_M (NETIF_MSG_DRV | NETIF_MSG_PROBE | NETIF_MSG_LINK)
 
 #define ICE_MAX_MTU	(ICE_AQ_SET_MAC_FRAME_SIZE_MAX - \
-			 ETH_HLEN + ETH_FCS_LEN + VLAN_HLEN)
+			(ETH_HLEN + ETH_FCS_LEN + (VLAN_HLEN * 2)))
 
 #define ICE_UP_TABLE_TRANSLATE(val, i) \
 		(((val) << ICE_AQ_VSI_UP_TABLE_UP##i##_S) & \
-- 
2.20.1

[net-next 03/16] ice: Fix added in VSI supported nodes calc

[Jeff Kirsher]

From: Victor Raj <victor.raj@intel.com>

VSI supported nodes are calculated in order to add the VSI parent or
intermediate nodes to the scheduler tree. If one of the node in below
layers (from VSI layer) has space to add the new VSI or intermediate node
above that layer then it's not required to continue the calculation further
for below layers.

Signed-off-by: Victor Raj <victor.raj@intel.com>
Reviewed-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_sched.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_sched.c b/drivers/net/ethernet/intel/ice/ice_sched.c
index fb38e8be1e2e..81fb7d19c0fd 100644
--- a/drivers/net/ethernet/intel/ice/ice_sched.c
+++ b/drivers/net/ethernet/intel/ice/ice_sched.c
@@ -1344,9 +1344,14 @@ ice_sched_calc_vsi_support_nodes(struct ice_hw *hw,
 				node = node->sibling;
 			}
 
+			/* tree has one intermediate node to add this new VSI.
+			 * So no need to calculate supported nodes for below
+			 * layers.
+			 */
+			if (node)
+				break;
 			/* all the nodes are full, allocate a new one */
-			if (!node)
-				num_nodes[i]++;
+			num_nodes[i]++;
 		}
 }
 
-- 
2.20.1

[net-next 04/16] ice: sizeof(<type>) should be avoided

[Jeff Kirsher]

From: Bruce Allan <bruce.w.allan@intel.com>

With sizeof(), it is preferable to use the variable of type <type> instead
of sizeof(<type>).

There are multiple places where a temporary variable is used to hold a
'size' value which is then used for a subsequent alloc/memset. Get rid
of the temporary variable by calculating size as part of the alloc/memset
statement.

Also remove unnecessary type-cast.

Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_ethtool.c | 15 ++++++------
 drivers/net/ethernet/intel/ice/ice_lib.c     | 10 ++++----
 drivers/net/ethernet/intel/ice/ice_main.c    | 23 +++++++++----------
 drivers/net/ethernet/intel/ice/ice_sched.c   |  9 ++++----
 drivers/net/ethernet/intel/ice/ice_switch.c  |  2 +-
 drivers/net/ethernet/intel/ice/ice_txrx.c    | 24 ++++++++------------
 6 files changed, 38 insertions(+), 45 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_ethtool.c b/drivers/net/ethernet/intel/ice/ice_ethtool.c
index a82f0202652d..a93daf660171 100644
--- a/drivers/net/ethernet/intel/ice/ice_ethtool.c
+++ b/drivers/net/ethernet/intel/ice/ice_ethtool.c
@@ -1084,7 +1084,7 @@ ice_get_settings_link_up(struct ethtool_link_ksettings *ks,
 	 * current PHY type, get what is supported by the NVM and intersect
 	 * them to get what is truly supported
 	 */
-	memset(&cap_ksettings, 0, sizeof(struct ethtool_link_ksettings));
+	memset(&cap_ksettings, 0, sizeof(cap_ksettings));
 	ice_phy_type_to_ethtool(netdev, &cap_ksettings);
 	ethtool_intersect_link_masks(ks, &cap_ksettings);
 
@@ -1416,7 +1416,7 @@ ice_set_link_ksettings(struct net_device *netdev,
 		return -EOPNOTSUPP;
 
 	/* copy the ksettings to copy_ks to avoid modifying the original */
-	memcpy(&copy_ks, ks, sizeof(struct ethtool_link_ksettings));
+	memcpy(&copy_ks, ks, sizeof(copy_ks));
 
 	/* save autoneg out of ksettings */
 	autoneg = copy_ks.base.autoneg;
@@ -1435,7 +1435,7 @@ ice_set_link_ksettings(struct net_device *netdev,
 		return -EINVAL;
 
 	/* get our own copy of the bits to check against */
-	memset(&safe_ks, 0, sizeof(struct ethtool_link_ksettings));
+	memset(&safe_ks, 0, sizeof(safe_ks));
 	safe_ks.base.cmd = copy_ks.base.cmd;
 	safe_ks.base.link_mode_masks_nwords =
 		copy_ks.base.link_mode_masks_nwords;
@@ -1449,8 +1449,7 @@ ice_set_link_ksettings(struct net_device *netdev,
 	/* If copy_ks.base and safe_ks.base are not the same now, then they are
 	 * trying to set something that we do not support.
 	 */
-	if (memcmp(&copy_ks.base, &safe_ks.base,
-		   sizeof(struct ethtool_link_settings)))
+	if (memcmp(&copy_ks.base, &safe_ks.base, sizeof(copy_ks.base)))
 		return -EOPNOTSUPP;
 
 	while (test_and_set_bit(__ICE_CFG_BUSY, pf->state)) {
@@ -1474,7 +1473,7 @@ ice_set_link_ksettings(struct net_device *netdev,
 	}
 
 	/* Copy abilities to config in case autoneg is not set below */
-	memset(&config, 0, sizeof(struct ice_aqc_set_phy_cfg_data));
+	memset(&config, 0, sizeof(config));
 	config.caps = abilities->caps & ~ICE_AQC_PHY_AN_MODE;
 	if (abilities->caps & ICE_AQC_PHY_AN_MODE)
 		config.caps |= ICE_AQ_PHY_ENA_AUTO_LINK_UPDT;
@@ -1668,7 +1667,7 @@ ice_set_ringparam(struct net_device *netdev, struct ethtool_ringparam *ring)
 		    vsi->tx_rings[0]->count, new_tx_cnt);
 
 	tx_rings = devm_kcalloc(&pf->pdev->dev, vsi->alloc_txq,
-				sizeof(struct ice_ring), GFP_KERNEL);
+				sizeof(*tx_rings), GFP_KERNEL);
 	if (!tx_rings) {
 		err = -ENOMEM;
 		goto done;
@@ -1700,7 +1699,7 @@ ice_set_ringparam(struct net_device *netdev, struct ethtool_ringparam *ring)
 		    vsi->rx_rings[0]->count, new_rx_cnt);
 
 	rx_rings = devm_kcalloc(&pf->pdev->dev, vsi->alloc_rxq,
-				sizeof(struct ice_ring), GFP_KERNEL);
+				sizeof(*rx_rings), GFP_KERNEL);
 	if (!rx_rings) {
 		err = -ENOMEM;
 		goto done;
diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c
index 27c3760ae5cb..82aae530d6c5 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.c
+++ b/drivers/net/ethernet/intel/ice/ice_lib.c
@@ -249,12 +249,12 @@ static int ice_vsi_alloc_arrays(struct ice_vsi *vsi, bool alloc_qvectors)
 
 	/* allocate memory for both Tx and Rx ring pointers */
 	vsi->tx_rings = devm_kcalloc(&pf->pdev->dev, vsi->alloc_txq,
-				     sizeof(struct ice_ring *), GFP_KERNEL);
+				     sizeof(*vsi->tx_rings), GFP_KERNEL);
 	if (!vsi->tx_rings)
 		goto err_txrings;
 
 	vsi->rx_rings = devm_kcalloc(&pf->pdev->dev, vsi->alloc_rxq,
-				     sizeof(struct ice_ring *), GFP_KERNEL);
+				     sizeof(*vsi->rx_rings), GFP_KERNEL);
 	if (!vsi->rx_rings)
 		goto err_rxrings;
 
@@ -262,7 +262,7 @@ static int ice_vsi_alloc_arrays(struct ice_vsi *vsi, bool alloc_qvectors)
 		/* allocate memory for q_vector pointers */
 		vsi->q_vectors = devm_kcalloc(&pf->pdev->dev,
 					      vsi->num_q_vectors,
-					      sizeof(struct ice_q_vector *),
+					      sizeof(*vsi->q_vectors),
 					      GFP_KERNEL);
 		if (!vsi->q_vectors)
 			goto err_vectors;
@@ -355,7 +355,7 @@ void ice_vsi_delete(struct ice_vsi *vsi)
 		ctxt.vf_num = vsi->vf_id;
 	ctxt.vsi_num = vsi->vsi_num;
 
-	memcpy(&ctxt.info, &vsi->info, sizeof(struct ice_aqc_vsi_props));
+	memcpy(&ctxt.info, &vsi->info, sizeof(ctxt.info));
 
 	status = ice_free_vsi(&pf->hw, vsi->idx, &ctxt, false, NULL);
 	if (status)
@@ -1620,7 +1620,7 @@ ice_vsi_cfg_txqs(struct ice_vsi *vsi, struct ice_ring **rings, int offset)
 	u16 buf_len, i, pf_q;
 	int err = 0, tc;
 
-	buf_len = sizeof(struct ice_aqc_add_tx_qgrp);
+	buf_len = sizeof(*qg_buf);
 	qg_buf = devm_kzalloc(&pf->pdev->dev, buf_len, GFP_KERNEL);
 	if (!qg_buf)
 		return -ENOMEM;
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 9d266d754445..0731b8994958 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -1513,8 +1513,8 @@ static int ice_cfg_netdev(struct ice_vsi *vsi)
 	u8 mac_addr[ETH_ALEN];
 	int err;
 
-	netdev = alloc_etherdev_mqs(sizeof(struct ice_netdev_priv),
-				    vsi->alloc_txq, vsi->alloc_rxq);
+	netdev = alloc_etherdev_mqs(sizeof(*np), vsi->alloc_txq,
+				    vsi->alloc_rxq);
 	if (!netdev)
 		return -ENOMEM;
 
@@ -1867,7 +1867,7 @@ static int ice_ena_msix_range(struct ice_pf *pf)
 	v_left -= pf->num_lan_msix;
 
 	pf->msix_entries = devm_kcalloc(&pf->pdev->dev, v_budget,
-					sizeof(struct msix_entry), GFP_KERNEL);
+					sizeof(*pf->msix_entries), GFP_KERNEL);
 
 	if (!pf->msix_entries) {
 		err = -ENOMEM;
@@ -1955,7 +1955,6 @@ static void ice_clear_interrupt_scheme(struct ice_pf *pf)
 static int ice_init_interrupt_scheme(struct ice_pf *pf)
 {
 	int vectors = 0, hw_vectors = 0;
-	ssize_t size;
 
 	if (test_bit(ICE_FLAG_MSIX_ENA, pf->flags))
 		vectors = ice_ena_msix_range(pf);
@@ -1966,9 +1965,9 @@ static int ice_init_interrupt_scheme(struct ice_pf *pf)
 		return vectors;
 
 	/* set up vector assignment tracking */
-	size = sizeof(struct ice_res_tracker) + (sizeof(u16) * vectors);
-
-	pf->sw_irq_tracker = devm_kzalloc(&pf->pdev->dev, size, GFP_KERNEL);
+	pf->sw_irq_tracker =
+		devm_kzalloc(&pf->pdev->dev, sizeof(*pf->sw_irq_tracker) +
+			     (sizeof(u16) * vectors), GFP_KERNEL);
 	if (!pf->sw_irq_tracker) {
 		ice_dis_msix(pf);
 		return -ENOMEM;
@@ -1980,9 +1979,9 @@ static int ice_init_interrupt_scheme(struct ice_pf *pf)
 
 	/* set up HW vector assignment tracking */
 	hw_vectors = pf->hw.func_caps.common_cap.num_msix_vectors;
-	size = sizeof(struct ice_res_tracker) + (sizeof(u16) * hw_vectors);
-
-	pf->hw_irq_tracker = devm_kzalloc(&pf->pdev->dev, size, GFP_KERNEL);
+	pf->hw_irq_tracker =
+		devm_kzalloc(&pf->pdev->dev, sizeof(*pf->hw_irq_tracker) +
+			     (sizeof(u16) * hw_vectors), GFP_KERNEL);
 	if (!pf->hw_irq_tracker) {
 		ice_clear_interrupt_scheme(pf);
 		return -ENOMEM;
@@ -2116,7 +2115,7 @@ static int ice_probe(struct pci_dev *pdev,
 	}
 
 	pf->vsi = devm_kcalloc(&pdev->dev, pf->num_alloc_vsi,
-			       sizeof(struct ice_vsi *), GFP_KERNEL);
+			       sizeof(*pf->vsi), GFP_KERNEL);
 	if (!pf->vsi) {
 		err = -ENOMEM;
 		goto err_init_pf_unroll;
@@ -2148,7 +2147,7 @@ static int ice_probe(struct pci_dev *pdev,
 	}
 
 	/* create switch struct for the switch element created by FW on boot */
-	pf->first_sw = devm_kzalloc(&pdev->dev, sizeof(struct ice_sw),
+	pf->first_sw = devm_kzalloc(&pdev->dev, sizeof(*pf->first_sw),
 				    GFP_KERNEL);
 	if (!pf->first_sw) {
 		err = -ENOMEM;
diff --git a/drivers/net/ethernet/intel/ice/ice_sched.c b/drivers/net/ethernet/intel/ice/ice_sched.c
index 81fb7d19c0fd..f355f59d8b64 100644
--- a/drivers/net/ethernet/intel/ice/ice_sched.c
+++ b/drivers/net/ethernet/intel/ice/ice_sched.c
@@ -1066,11 +1066,10 @@ enum ice_status ice_sched_query_res_alloc(struct ice_hw *hw)
 		hw->max_children[i] = le16_to_cpu(max_sibl);
 	}
 
-	hw->layer_info = (struct ice_aqc_layer_props *)
-			  devm_kmemdup(ice_hw_to_dev(hw), buf->layer_props,
-				       (hw->num_tx_sched_layers *
-					sizeof(*hw->layer_info)),
-				       GFP_KERNEL);
+	hw->layer_info = devm_kmemdup(ice_hw_to_dev(hw), buf->layer_props,
+				      (hw->num_tx_sched_layers *
+				       sizeof(*hw->layer_info)),
+				      GFP_KERNEL);
 	if (!hw->layer_info) {
 		status = ICE_ERR_NO_MEMORY;
 		goto sched_query_out;
diff --git a/drivers/net/ethernet/intel/ice/ice_switch.c b/drivers/net/ethernet/intel/ice/ice_switch.c
index 2e5693107fa4..d560a8aa5015 100644
--- a/drivers/net/ethernet/intel/ice/ice_switch.c
+++ b/drivers/net/ethernet/intel/ice/ice_switch.c
@@ -98,7 +98,7 @@ enum ice_status ice_init_def_sw_recp(struct ice_hw *hw)
 	u8 i;
 
 	recps = devm_kcalloc(ice_hw_to_dev(hw), ICE_MAX_NUM_RECIPES,
-			     sizeof(struct ice_sw_recipe), GFP_KERNEL);
+			     sizeof(*recps), GFP_KERNEL);
 	if (!recps)
 		return ICE_ERR_NO_MEMORY;
 
diff --git a/drivers/net/ethernet/intel/ice/ice_txrx.c b/drivers/net/ethernet/intel/ice/ice_txrx.c
index 2357fcac996b..67d129bfe489 100644
--- a/drivers/net/ethernet/intel/ice/ice_txrx.c
+++ b/drivers/net/ethernet/intel/ice/ice_txrx.c
@@ -48,7 +48,6 @@ static struct netdev_queue *txring_txq(const struct ice_ring *ring)
  */
 void ice_clean_tx_ring(struct ice_ring *tx_ring)
 {
-	unsigned long size;
 	u16 i;
 
 	/* ring already cleared, nothing to do */
@@ -59,8 +58,7 @@ void ice_clean_tx_ring(struct ice_ring *tx_ring)
 	for (i = 0; i < tx_ring->count; i++)
 		ice_unmap_and_free_tx_buf(tx_ring, &tx_ring->tx_buf[i]);
 
-	size = sizeof(struct ice_tx_buf) * tx_ring->count;
-	memset(tx_ring->tx_buf, 0, size);
+	memset(tx_ring->tx_buf, 0, sizeof(*tx_ring->tx_buf) * tx_ring->count);
 
 	/* Zero out the descriptor ring */
 	memset(tx_ring->desc, 0, tx_ring->size);
@@ -226,21 +224,21 @@ static bool ice_clean_tx_irq(struct ice_vsi *vsi, struct ice_ring *tx_ring,
 int ice_setup_tx_ring(struct ice_ring *tx_ring)
 {
 	struct device *dev = tx_ring->dev;
-	int bi_size;
 
 	if (!dev)
 		return -ENOMEM;
 
 	/* warn if we are about to overwrite the pointer */
 	WARN_ON(tx_ring->tx_buf);
-	bi_size = sizeof(struct ice_tx_buf) * tx_ring->count;
-	tx_ring->tx_buf = devm_kzalloc(dev, bi_size, GFP_KERNEL);
+	tx_ring->tx_buf =
+		devm_kzalloc(dev, sizeof(*tx_ring->tx_buf) * tx_ring->count,
+			     GFP_KERNEL);
 	if (!tx_ring->tx_buf)
 		return -ENOMEM;
 
 	/* round up to nearest 4K */
-	tx_ring->size = tx_ring->count * sizeof(struct ice_tx_desc);
-	tx_ring->size = ALIGN(tx_ring->size, 4096);
+	tx_ring->size = ALIGN(tx_ring->count * sizeof(struct ice_tx_desc),
+			      4096);
 	tx_ring->desc = dmam_alloc_coherent(dev, tx_ring->size, &tx_ring->dma,
 					    GFP_KERNEL);
 	if (!tx_ring->desc) {
@@ -267,7 +265,6 @@ int ice_setup_tx_ring(struct ice_ring *tx_ring)
 void ice_clean_rx_ring(struct ice_ring *rx_ring)
 {
 	struct device *dev = rx_ring->dev;
-	unsigned long size;
 	u16 i;
 
 	/* ring already cleared, nothing to do */
@@ -292,8 +289,7 @@ void ice_clean_rx_ring(struct ice_ring *rx_ring)
 		rx_buf->page_offset = 0;
 	}
 
-	size = sizeof(struct ice_rx_buf) * rx_ring->count;
-	memset(rx_ring->rx_buf, 0, size);
+	memset(rx_ring->rx_buf, 0, sizeof(*rx_ring->rx_buf) * rx_ring->count);
 
 	/* Zero out the descriptor ring */
 	memset(rx_ring->desc, 0, rx_ring->size);
@@ -331,15 +327,15 @@ void ice_free_rx_ring(struct ice_ring *rx_ring)
 int ice_setup_rx_ring(struct ice_ring *rx_ring)
 {
 	struct device *dev = rx_ring->dev;
-	int bi_size;
 
 	if (!dev)
 		return -ENOMEM;
 
 	/* warn if we are about to overwrite the pointer */
 	WARN_ON(rx_ring->rx_buf);
-	bi_size = sizeof(struct ice_rx_buf) * rx_ring->count;
-	rx_ring->rx_buf = devm_kzalloc(dev, bi_size, GFP_KERNEL);
+	rx_ring->rx_buf =
+		devm_kzalloc(dev, sizeof(*rx_ring->rx_buf) * rx_ring->count,
+			     GFP_KERNEL);
 	if (!rx_ring->rx_buf)
 		return -ENOMEM;
 
-- 
2.20.1

[net-next 05/16] ice: fix stack hogs from struct ice_vsi_ctx structures

[Jeff Kirsher]

From: Bruce Allan <bruce.w.allan@intel.com>

struct ice_vsi_ctx has gotten large enough that function local declarations
of it on the stack are causing stack hogs.  Fix that by allocating the
structs on heap.  Cleanup some formatting issues in the code around these
changes and fix incorrect data type uses of returned functions in a couple
places.

Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_lib.c      | 98 ++++++++++++-------
 drivers/net/ethernet/intel/ice/ice_main.c     | 27 +++--
 .../net/ethernet/intel/ice/ice_virtchnl_pf.c  | 59 ++++++-----
 3 files changed, 117 insertions(+), 67 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c
index 82aae530d6c5..d93a0d5a783c 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.c
+++ b/drivers/net/ethernet/intel/ice/ice_lib.c
@@ -348,19 +348,25 @@ static int ice_get_free_slot(void *array, int size, int curr)
 void ice_vsi_delete(struct ice_vsi *vsi)
 {
 	struct ice_pf *pf = vsi->back;
-	struct ice_vsi_ctx ctxt;
+	struct ice_vsi_ctx *ctxt;
 	enum ice_status status;
 
+	ctxt = devm_kzalloc(&pf->pdev->dev, sizeof(*ctxt), GFP_KERNEL);
+	if (!ctxt)
+		return;
+
 	if (vsi->type == ICE_VSI_VF)
-		ctxt.vf_num = vsi->vf_id;
-	ctxt.vsi_num = vsi->vsi_num;
+		ctxt->vf_num = vsi->vf_id;
+	ctxt->vsi_num = vsi->vsi_num;
 
-	memcpy(&ctxt.info, &vsi->info, sizeof(ctxt.info));
+	memcpy(&ctxt->info, &vsi->info, sizeof(ctxt->info));
 
-	status = ice_free_vsi(&pf->hw, vsi->idx, &ctxt, false, NULL);
+	status = ice_free_vsi(&pf->hw, vsi->idx, ctxt, false, NULL);
 	if (status)
 		dev_err(&pf->pdev->dev, "Failed to delete VSI %i in FW\n",
 			vsi->vsi_num);
+
+	devm_kfree(&pf->pdev->dev, ctxt);
 }
 
 /**
@@ -908,37 +914,41 @@ static void ice_set_rss_vsi_ctx(struct ice_vsi_ctx *ctxt, struct ice_vsi *vsi)
  */
 static int ice_vsi_init(struct ice_vsi *vsi)
 {
-	struct ice_vsi_ctx ctxt = { 0 };
 	struct ice_pf *pf = vsi->back;
 	struct ice_hw *hw = &pf->hw;
+	struct ice_vsi_ctx *ctxt;
 	int ret = 0;
 
+	ctxt = devm_kzalloc(&pf->pdev->dev, sizeof(*ctxt), GFP_KERNEL);
+	if (!ctxt)
+		return -ENOMEM;
+
 	switch (vsi->type) {
 	case ICE_VSI_PF:
-		ctxt.flags = ICE_AQ_VSI_TYPE_PF;
+		ctxt->flags = ICE_AQ_VSI_TYPE_PF;
 		break;
 	case ICE_VSI_VF:
-		ctxt.flags = ICE_AQ_VSI_TYPE_VF;
+		ctxt->flags = ICE_AQ_VSI_TYPE_VF;
 		/* VF number here is the absolute VF number (0-255) */
-		ctxt.vf_num = vsi->vf_id + hw->func_caps.vf_base_id;
+		ctxt->vf_num = vsi->vf_id + hw->func_caps.vf_base_id;
 		break;
 	default:
 		return -ENODEV;
 	}
 
-	ice_set_dflt_vsi_ctx(&ctxt);
+	ice_set_dflt_vsi_ctx(ctxt);
 	/* if the switch is in VEB mode, allow VSI loopback */
 	if (vsi->vsw->bridge_mode == BRIDGE_MODE_VEB)
-		ctxt.info.sw_flags |= ICE_AQ_VSI_SW_FLAG_ALLOW_LB;
+		ctxt->info.sw_flags |= ICE_AQ_VSI_SW_FLAG_ALLOW_LB;
 
 	/* Set LUT type and HASH type if RSS is enabled */
 	if (test_bit(ICE_FLAG_RSS_ENA, pf->flags))
-		ice_set_rss_vsi_ctx(&ctxt, vsi);
+		ice_set_rss_vsi_ctx(ctxt, vsi);
 
-	ctxt.info.sw_id = vsi->port_info->sw_id;
-	ice_vsi_setup_q_map(vsi, &ctxt);
+	ctxt->info.sw_id = vsi->port_info->sw_id;
+	ice_vsi_setup_q_map(vsi, ctxt);
 
-	ret = ice_add_vsi(hw, vsi->idx, &ctxt, NULL);
+	ret = ice_add_vsi(hw, vsi->idx, ctxt, NULL);
 	if (ret) {
 		dev_err(&pf->pdev->dev,
 			"Add VSI failed, err %d\n", ret);
@@ -946,11 +956,12 @@ static int ice_vsi_init(struct ice_vsi *vsi)
 	}
 
 	/* keep context for update VSI operations */
-	vsi->info = ctxt.info;
+	vsi->info = ctxt->info;
 
 	/* record VSI number returned */
-	vsi->vsi_num = ctxt.vsi_num;
+	vsi->vsi_num = ctxt->vsi_num;
 
+	devm_kfree(&pf->pdev->dev, ctxt);
 	return ret;
 }
 
@@ -1823,26 +1834,34 @@ int ice_vsi_manage_vlan_insertion(struct ice_vsi *vsi)
 {
 	struct device *dev = &vsi->back->pdev->dev;
 	struct ice_hw *hw = &vsi->back->hw;
-	struct ice_vsi_ctx ctxt = { 0 };
+	struct ice_vsi_ctx *ctxt;
 	enum ice_status status;
+	int ret = 0;
+
+	ctxt = devm_kzalloc(dev, sizeof(*ctxt), GFP_KERNEL);
+	if (!ctxt)
+		return -ENOMEM;
 
 	/* Here we are configuring the VSI to let the driver add VLAN tags by
 	 * setting vlan_flags to ICE_AQ_VSI_VLAN_MODE_ALL. The actual VLAN tag
 	 * insertion happens in the Tx hot path, in ice_tx_map.
 	 */
-	ctxt.info.vlan_flags = ICE_AQ_VSI_VLAN_MODE_ALL;
+	ctxt->info.vlan_flags = ICE_AQ_VSI_VLAN_MODE_ALL;
 
-	ctxt.info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_VLAN_VALID);
+	ctxt->info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_VLAN_VALID);
 
-	status = ice_update_vsi(hw, vsi->idx, &ctxt, NULL);
+	status = ice_update_vsi(hw, vsi->idx, ctxt, NULL);
 	if (status) {
 		dev_err(dev, "update VSI for VLAN insert failed, err %d aq_err %d\n",
 			status, hw->adminq.sq_last_status);
-		return -EIO;
+		ret = -EIO;
+		goto out;
 	}
 
-	vsi->info.vlan_flags = ctxt.info.vlan_flags;
-	return 0;
+	vsi->info.vlan_flags = ctxt->info.vlan_flags;
+out:
+	devm_kfree(dev, ctxt);
+	return ret;
 }
 
 /**
@@ -1854,35 +1873,42 @@ int ice_vsi_manage_vlan_stripping(struct ice_vsi *vsi, bool ena)
 {
 	struct device *dev = &vsi->back->pdev->dev;
 	struct ice_hw *hw = &vsi->back->hw;
-	struct ice_vsi_ctx ctxt = { 0 };
+	struct ice_vsi_ctx *ctxt;
 	enum ice_status status;
+	int ret = 0;
+
+	ctxt = devm_kzalloc(dev, sizeof(*ctxt), GFP_KERNEL);
+	if (!ctxt)
+		return -ENOMEM;
 
 	/* Here we are configuring what the VSI should do with the VLAN tag in
 	 * the Rx packet. We can either leave the tag in the packet or put it in
 	 * the Rx descriptor.
 	 */
-	if (ena) {
+	if (ena)
 		/* Strip VLAN tag from Rx packet and put it in the desc */
-		ctxt.info.vlan_flags = ICE_AQ_VSI_VLAN_EMOD_STR_BOTH;
-	} else {
+		ctxt->info.vlan_flags = ICE_AQ_VSI_VLAN_EMOD_STR_BOTH;
+	else
 		/* Disable stripping. Leave tag in packet */
-		ctxt.info.vlan_flags = ICE_AQ_VSI_VLAN_EMOD_NOTHING;
-	}
+		ctxt->info.vlan_flags = ICE_AQ_VSI_VLAN_EMOD_NOTHING;
 
 	/* Allow all packets untagged/tagged */
-	ctxt.info.vlan_flags |= ICE_AQ_VSI_VLAN_MODE_ALL;
+	ctxt->info.vlan_flags |= ICE_AQ_VSI_VLAN_MODE_ALL;
 
-	ctxt.info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_VLAN_VALID);
+	ctxt->info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_VLAN_VALID);
 
-	status = ice_update_vsi(hw, vsi->idx, &ctxt, NULL);
+	status = ice_update_vsi(hw, vsi->idx, ctxt, NULL);
 	if (status) {
 		dev_err(dev, "update VSI for VLAN strip failed, ena = %d err %d aq_err %d\n",
 			ena, status, hw->adminq.sq_last_status);
-		return -EIO;
+		ret = -EIO;
+		goto out;
 	}
 
-	vsi->info.vlan_flags = ctxt.info.vlan_flags;
-	return 0;
+	vsi->info.vlan_flags = ctxt->info.vlan_flags;
+out:
+	devm_kfree(dev, ctxt);
+	return ret;
 }
 
 /**
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 0731b8994958..aff348e42562 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -3707,30 +3707,39 @@ static int ice_vsi_update_bridge_mode(struct ice_vsi *vsi, u16 bmode)
 	struct device *dev = &vsi->back->pdev->dev;
 	struct ice_aqc_vsi_props *vsi_props;
 	struct ice_hw *hw = &vsi->back->hw;
-	struct ice_vsi_ctx ctxt = { 0 };
+	struct ice_vsi_ctx *ctxt;
 	enum ice_status status;
+	int ret = 0;
 
 	vsi_props = &vsi->info;
-	ctxt.info = vsi->info;
+
+	ctxt = devm_kzalloc(dev, sizeof(*ctxt), GFP_KERNEL);
+	if (!ctxt)
+		return -ENOMEM;
+
+	ctxt->info = vsi->info;
 
 	if (bmode == BRIDGE_MODE_VEB)
 		/* change from VEPA to VEB mode */
-		ctxt.info.sw_flags |= ICE_AQ_VSI_SW_FLAG_ALLOW_LB;
+		ctxt->info.sw_flags |= ICE_AQ_VSI_SW_FLAG_ALLOW_LB;
 	else
 		/* change from VEB to VEPA mode */
-		ctxt.info.sw_flags &= ~ICE_AQ_VSI_SW_FLAG_ALLOW_LB;
-	ctxt.info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_SW_VALID);
+		ctxt->info.sw_flags &= ~ICE_AQ_VSI_SW_FLAG_ALLOW_LB;
+	ctxt->info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_SW_VALID);
 
-	status = ice_update_vsi(hw, vsi->idx, &ctxt, NULL);
+	status = ice_update_vsi(hw, vsi->idx, ctxt, NULL);
 	if (status) {
 		dev_err(dev, "update VSI for bridge mode failed, bmode = %d err %d aq_err %d\n",
 			bmode, status, hw->adminq.sq_last_status);
-		return -EIO;
+		ret = -EIO;
+		goto out;
 	}
 	/* Update sw flags for book keeping */
-	vsi_props->sw_flags = ctxt.info.sw_flags;
+	vsi_props->sw_flags = ctxt->info.sw_flags;
 
-	return 0;
+out:
+	devm_kfree(dev, ctxt);
+	return ret;
 }
 
 /**
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
index 80b50e67cbef..79d793a76042 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
@@ -345,25 +345,33 @@ static int ice_vsi_set_pvid(struct ice_vsi *vsi, u16 vid)
 {
 	struct device *dev = &vsi->back->pdev->dev;
 	struct ice_hw *hw = &vsi->back->hw;
-	struct ice_vsi_ctx ctxt = { 0 };
+	struct ice_vsi_ctx *ctxt;
 	enum ice_status status;
+	int ret = 0;
+
+	ctxt = devm_kzalloc(dev, sizeof(*ctxt), GFP_KERNEL);
+	if (!ctxt)
+		return -ENOMEM;
 
-	ctxt.info.vlan_flags = ICE_AQ_VSI_VLAN_MODE_UNTAGGED |
-			       ICE_AQ_VSI_PVLAN_INSERT_PVID |
-			       ICE_AQ_VSI_VLAN_EMOD_STR;
-	ctxt.info.pvid = cpu_to_le16(vid);
-	ctxt.info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_VLAN_VALID);
+	ctxt->info.vlan_flags = (ICE_AQ_VSI_VLAN_MODE_UNTAGGED |
+				 ICE_AQ_VSI_PVLAN_INSERT_PVID |
+				 ICE_AQ_VSI_VLAN_EMOD_STR);
+	ctxt->info.pvid = cpu_to_le16(vid);
+	ctxt->info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_VLAN_VALID);
 
-	status = ice_update_vsi(hw, vsi->idx, &ctxt, NULL);
+	status = ice_update_vsi(hw, vsi->idx, ctxt, NULL);
 	if (status) {
 		dev_info(dev, "update VSI for VLAN insert failed, err %d aq_err %d\n",
 			 status, hw->adminq.sq_last_status);
-		return -EIO;
+		ret = -EIO;
+		goto out;
 	}
 
-	vsi->info.pvid = ctxt.info.pvid;
-	vsi->info.vlan_flags = ctxt.info.vlan_flags;
-	return 0;
+	vsi->info.pvid = ctxt->info.pvid;
+	vsi->info.vlan_flags = ctxt->info.vlan_flags;
+out:
+	devm_kfree(dev, ctxt);
+	return ret;
 }
 
 /**
@@ -2479,11 +2487,12 @@ int ice_get_vf_cfg(struct net_device *netdev, int vf_id,
 int ice_set_vf_spoofchk(struct net_device *netdev, int vf_id, bool ena)
 {
 	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_vsi_ctx ctx = { 0 };
 	struct ice_vsi *vsi = np->vsi;
 	struct ice_pf *pf = vsi->back;
+	struct ice_vsi_ctx *ctx;
+	enum ice_status status;
 	struct ice_vf *vf;
-	int status;
+	int ret = 0;
 
 	/* validate the request */
 	if (vf_id >= pf->num_alloc_vfs) {
@@ -2503,25 +2512,31 @@ int ice_set_vf_spoofchk(struct net_device *netdev, int vf_id, bool ena)
 		return 0;
 	}
 
-	ctx.info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_SECURITY_VALID);
+	ctx = devm_kzalloc(&pf->pdev->dev, sizeof(*ctx), GFP_KERNEL);
+	if (!ctx)
+		return -ENOMEM;
+
+	ctx->info.valid_sections = cpu_to_le16(ICE_AQ_VSI_PROP_SECURITY_VALID);
 
 	if (ena) {
-		ctx.info.sec_flags |= ICE_AQ_VSI_SEC_FLAG_ENA_MAC_ANTI_SPOOF;
-		ctx.info.sw_flags2 |= ICE_AQ_VSI_SW_FLAG_RX_PRUNE_EN_M;
+		ctx->info.sec_flags |= ICE_AQ_VSI_SEC_FLAG_ENA_MAC_ANTI_SPOOF;
+		ctx->info.sw_flags2 |= ICE_AQ_VSI_SW_FLAG_RX_PRUNE_EN_M;
 	}
 
-	status = ice_update_vsi(&pf->hw, vsi->idx, &ctx, NULL);
+	status = ice_update_vsi(&pf->hw, vsi->idx, ctx, NULL);
 	if (status) {
 		dev_dbg(&pf->pdev->dev,
 			"Error %d, failed to update VSI* parameters\n", status);
-		return -EIO;
+		ret = -EIO;
+		goto out;
 	}
 
 	vf->spoofchk = ena;
-	vsi->info.sec_flags = ctx.info.sec_flags;
-	vsi->info.sw_flags2 = ctx.info.sw_flags2;
-
-	return status;
+	vsi->info.sec_flags = ctx->info.sec_flags;
+	vsi->info.sw_flags2 = ctx->info.sw_flags2;
+out:
+	devm_kfree(&pf->pdev->dev, ctx);
+	return ret;
 }
 
 /**
-- 
2.20.1

[net-next 06/16] ice: fix ice_remove_rule_internal vsi_list handling

[Jeff Kirsher]

From: Jacob Keller <jacob.e.keller@intel.com>

When adding multiple VLANs to the same VSI, the ice_add_vlan code will
share the VSI list, so as not to create multiple unnecessary VSI lists.

Consider the following flow

  ice_add_vlan(hw, <VSI 0 VID 7, VSI 0 VID 8, VSI 0 VID 9>)

Where we add three VLAN filters for VIDs 7, 8, and 9, all for VSI 0.

The ice_add_vlan will create a single vsi_list and share it among all
the filters.

Later, if we try to remove a VLAN,

  ice_remove_vlan(hw, <VSI 0 VID 7>)

Then the removal code will update the vsi_list and remove VSI 0 from it.
But, since the vsi_list is shared, this breaks the list for the other
users who reference it. We actually even free the VSI list memory, and
may result in segmentation faults.

This is due to the way that VLAN rule share VSI lists with reference
counts, and is caused because we call ice_rem_update_vsi_list even when
the ref_cnt is greater than one.

To fix this, handle the case where ref_cnt is greater than one
separately. In this case, we need to remove the associated rule without
modifying the vsi_list, since it is currently being referenced by
another rule. Instead, we just need to decrement the VSI list ref_cnt.

The case for handling sharing of VSI lists with multiple VSIs is not
currently supported by this code. No such rules will be created today,
and this code will require changes if/when such code is added.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_switch.c | 15 +++++++++++++--
 1 file changed, 13 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_switch.c b/drivers/net/ethernet/intel/ice/ice_switch.c
index d560a8aa5015..09d1c314b68f 100644
--- a/drivers/net/ethernet/intel/ice/ice_switch.c
+++ b/drivers/net/ethernet/intel/ice/ice_switch.c
@@ -1538,9 +1538,20 @@ ice_remove_rule_internal(struct ice_hw *hw, u8 recp_id,
 	} else if (!list_elem->vsi_list_info) {
 		status = ICE_ERR_DOES_NOT_EXIST;
 		goto exit;
+	} else if (list_elem->vsi_list_info->ref_cnt > 1) {
+		/* a ref_cnt > 1 indicates that the vsi_list is being
+		 * shared by multiple rules. Decrement the ref_cnt and
+		 * remove this rule, but do not modify the list, as it
+		 * is in-use by other rules.
+		 */
+		list_elem->vsi_list_info->ref_cnt--;
+		remove_rule = true;
 	} else {
-		if (list_elem->vsi_list_info->ref_cnt > 1)
-			list_elem->vsi_list_info->ref_cnt--;
+		/* a ref_cnt of 1 indicates the vsi_list is only used
+		 * by one rule. However, the original removal request is only
+		 * for a single VSI. Update the vsi_list first, and only
+		 * remove the rule if there are no further VSIs in this list.
+		 */
 		vsi_handle = f_entry->fltr_info.vsi_handle;
 		status = ice_rem_update_vsi_list(hw, vsi_handle, list_elem);
 		if (status)
-- 
2.20.1

[net-next 07/16] ice: fix issue where host reboots on unload when iommu=on

[Jeff Kirsher]

From: Brett Creeley <brett.creeley@intel.com>

Currently if the kernel has the intel_iommu=on parameter set, on some
platforms removing the driver causes a system reboot. In initialization
we associate the control queue interrupts with the pf->hw_oicr_idx and
enable the interrupts by setting the CAUSE_ENA bit. The problem comes
on teardown because we are not clearing the CAUSE_ENA bit for the
control queues, but the vector at pf->hw_oicr_idx (miscellaneous
interrupt vector) gets disabled.

Fix this by clearing the CAUSE_ENA bit in the appropriate control queue
registers on when freeing the miscellaneous interrupt vector. Also,
move the call to ice_free_irq_msix_misc() to after ice_deinit_sw() in
ice_remove() because ice_deinit_sw() makes an AQ call, but
ice_free_irq_msix_misc() disables the miscellaneous vector and it's
associated interrupts.

Also, create two small helper functions to enable and disable the
control queue interrupts respectively.

Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_main.c | 71 +++++++++++++++++------
 1 file changed, 54 insertions(+), 17 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index aff348e42562..fb04a5ebdc0b 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -1355,15 +1355,40 @@ static irqreturn_t ice_misc_intr(int __always_unused irq, void *data)
 	return ret;
 }
 
+/**
+ * ice_dis_ctrlq_interrupts - disable control queue interrupts
+ * @hw: pointer to HW structure
+ */
+static void ice_dis_ctrlq_interrupts(struct ice_hw *hw)
+{
+	/* disable Admin queue Interrupt causes */
+	wr32(hw, PFINT_FW_CTL,
+	     rd32(hw, PFINT_FW_CTL) & ~PFINT_FW_CTL_CAUSE_ENA_M);
+
+	/* disable Mailbox queue Interrupt causes */
+	wr32(hw, PFINT_MBX_CTL,
+	     rd32(hw, PFINT_MBX_CTL) & ~PFINT_MBX_CTL_CAUSE_ENA_M);
+
+	/* disable Control queue Interrupt causes */
+	wr32(hw, PFINT_OICR_CTL,
+	     rd32(hw, PFINT_OICR_CTL) & ~PFINT_OICR_CTL_CAUSE_ENA_M);
+
+	ice_flush(hw);
+}
+
 /**
  * ice_free_irq_msix_misc - Unroll misc vector setup
  * @pf: board private structure
  */
 static void ice_free_irq_msix_misc(struct ice_pf *pf)
 {
+	struct ice_hw *hw = &pf->hw;
+
+	ice_dis_ctrlq_interrupts(hw);
+
 	/* disable OICR interrupt */
-	wr32(&pf->hw, PFINT_OICR_ENA, 0);
-	ice_flush(&pf->hw);
+	wr32(hw, PFINT_OICR_ENA, 0);
+	ice_flush(hw);
 
 	if (test_bit(ICE_FLAG_MSIX_ENA, pf->flags) && pf->msix_entries) {
 		synchronize_irq(pf->msix_entries[pf->sw_oicr_idx].vector);
@@ -1377,6 +1402,32 @@ static void ice_free_irq_msix_misc(struct ice_pf *pf)
 	ice_free_res(pf->hw_irq_tracker, pf->hw_oicr_idx, ICE_RES_MISC_VEC_ID);
 }
 
+/**
+ * ice_ena_ctrlq_interrupts - enable control queue interrupts
+ * @hw: pointer to HW structure
+ * @v_idx: HW vector index to associate the control queue interrupts with
+ */
+static void ice_ena_ctrlq_interrupts(struct ice_hw *hw, u16 v_idx)
+{
+	u32 val;
+
+	val = ((v_idx & PFINT_OICR_CTL_MSIX_INDX_M) |
+	       PFINT_OICR_CTL_CAUSE_ENA_M);
+	wr32(hw, PFINT_OICR_CTL, val);
+
+	/* enable Admin queue Interrupt causes */
+	val = ((v_idx & PFINT_FW_CTL_MSIX_INDX_M) |
+	       PFINT_FW_CTL_CAUSE_ENA_M);
+	wr32(hw, PFINT_FW_CTL, val);
+
+	/* enable Mailbox queue Interrupt causes */
+	val = ((v_idx & PFINT_MBX_CTL_MSIX_INDX_M) |
+	       PFINT_MBX_CTL_CAUSE_ENA_M);
+	wr32(hw, PFINT_MBX_CTL, val);
+
+	ice_flush(hw);
+}
+
 /**
  * ice_req_irq_msix_misc - Setup the misc vector to handle non queue events
  * @pf: board private structure
@@ -1389,7 +1440,6 @@ static int ice_req_irq_msix_misc(struct ice_pf *pf)
 {
 	struct ice_hw *hw = &pf->hw;
 	int oicr_idx, err = 0;
-	u32 val;
 
 	if (!pf->int_name[0])
 		snprintf(pf->int_name, sizeof(pf->int_name) - 1, "%s-%s:misc",
@@ -1438,20 +1488,7 @@ static int ice_req_irq_msix_misc(struct ice_pf *pf)
 skip_req_irq:
 	ice_ena_misc_vector(pf);
 
-	val = ((pf->hw_oicr_idx & PFINT_OICR_CTL_MSIX_INDX_M) |
-	       PFINT_OICR_CTL_CAUSE_ENA_M);
-	wr32(hw, PFINT_OICR_CTL, val);
-
-	/* This enables Admin queue Interrupt causes */
-	val = ((pf->hw_oicr_idx & PFINT_FW_CTL_MSIX_INDX_M) |
-	       PFINT_FW_CTL_CAUSE_ENA_M);
-	wr32(hw, PFINT_FW_CTL, val);
-
-	/* This enables Mailbox queue Interrupt causes */
-	val = ((pf->hw_oicr_idx & PFINT_MBX_CTL_MSIX_INDX_M) |
-	       PFINT_MBX_CTL_CAUSE_ENA_M);
-	wr32(hw, PFINT_MBX_CTL, val);
-
+	ice_ena_ctrlq_interrupts(hw, pf->hw_oicr_idx);
 	wr32(hw, GLINT_ITR(ICE_RX_ITR, pf->hw_oicr_idx),
 	     ITR_REG_ALIGN(ICE_ITR_8K) >> ICE_ITR_GRAN_S);
 
-- 
2.20.1

[net-next 08/16] ice: fix numeric overflow warning

[Jeff Kirsher]

From: Bruce Allan <bruce.w.allan@intel.com>

When compiling and analyzing the driver on newer kernels, a static
analyzer warns about the following "numeric overflow" issues:

  "The result of expression: 'budget-1' generates 4-byte type while casting
   to a bigger size of 8-byte".

  "The result of expression: '*words-words_read' generates 4-byte type
   while casting to a bigger size of 8-byte".

Fix them both.

Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_nvm.c  | 7 ++++---
 drivers/net/ethernet/intel/ice/ice_txrx.c | 2 +-
 2 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_nvm.c b/drivers/net/ethernet/intel/ice/ice_nvm.c
index ce64cecdae9c..413fdbbcc4d0 100644
--- a/drivers/net/ethernet/intel/ice/ice_nvm.c
+++ b/drivers/net/ethernet/intel/ice/ice_nvm.c
@@ -152,9 +152,10 @@ ice_read_sr_buf_aq(struct ice_hw *hw, u16 offset, u16 *words, u16 *data)
 		 */
 		off_w = offset % ICE_SR_SECTOR_SIZE_IN_WORDS;
 		read_size = off_w ?
-			min(*words,
-			    (u16)(ICE_SR_SECTOR_SIZE_IN_WORDS - off_w)) :
-			min((*words - words_read), ICE_SR_SECTOR_SIZE_IN_WORDS);
+			min_t(u16, *words,
+			      (ICE_SR_SECTOR_SIZE_IN_WORDS - off_w)) :
+			min_t(u16, (*words - words_read),
+			      ICE_SR_SECTOR_SIZE_IN_WORDS);
 
 		/* Check if this is last command, if so set proper flag */
 		if ((words_read + read_size) >= *words)
diff --git a/drivers/net/ethernet/intel/ice/ice_txrx.c b/drivers/net/ethernet/intel/ice/ice_txrx.c
index 67d129bfe489..c289d97f477d 100644
--- a/drivers/net/ethernet/intel/ice/ice_txrx.c
+++ b/drivers/net/ethernet/intel/ice/ice_txrx.c
@@ -1169,7 +1169,7 @@ int ice_napi_poll(struct napi_struct *napi, int budget)
 		if (test_bit(ICE_FLAG_MSIX_ENA, pf->flags))
 			ice_update_ena_itr(vsi, q_vector);
 
-	return min(work_done, budget - 1);
+	return min_t(int, work_done, budget - 1);
 }
 
 /* helper function for building cmd/type/offset */
-- 
2.20.1

[net-next 09/16] ice: only use the VF for ICE_VSI_VF in ice_vsi_release

[Jeff Kirsher]

From: Brett Creeley <brett.creeley@intel.com>

In ice_vsi_release we are always assigning a value to the local VF
variable. Change this to only be assigned if the VSI is a VF VSI.

Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_lib.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_lib.c b/drivers/net/ethernet/intel/ice/ice_lib.c
index d93a0d5a783c..fa61203bee26 100644
--- a/drivers/net/ethernet/intel/ice/ice_lib.c
+++ b/drivers/net/ethernet/intel/ice/ice_lib.c
@@ -2518,13 +2518,15 @@ void ice_vsi_dis_irq(struct ice_vsi *vsi)
  */
 int ice_vsi_release(struct ice_vsi *vsi)
 {
+	struct ice_vf *vf = NULL;
 	struct ice_pf *pf;
-	struct ice_vf *vf;
 
 	if (!vsi->back)
 		return -ENODEV;
 	pf = vsi->back;
-	vf = &pf->vf[vsi->vf_id];
+
+	if (vsi->type == ICE_VSI_VF)
+		vf = &pf->vf[vsi->vf_id];
 	/* do not unregister and free netdevs while driver is in the reset
 	 * recovery pending state. Since reset/rebuild happens through PF
 	 * service task workqueue, its not a good idea to unregister netdev
-- 
2.20.1

[net-next 10/16] ice: don't spam VFs with link messages

[Jeff Kirsher]

From: Mitch Williams <mitch.a.williams@intel.com>

Don't send a link message to the VFs unless link actually changes state.
This avoids a small timing hole in some VF drivers that can cause an
apparent TX hang if they receive a link status message at the wrong time.

Although we have fixed the timing hole in the current VF driver, there
are still lots of drivers in the field that have this timing hole. Let's
not fall into it if we can avoid it.

Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_main.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index fb04a5ebdc0b..47cc3f905b7f 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -609,7 +609,8 @@ ice_link_event(struct ice_pf *pf, struct ice_port_info *pi)
 		}
 	}
 
-	ice_vc_notify_link_state(pf);
+	if (!new_link_same_as_old && pf->num_alloc_vfs)
+		ice_vc_notify_link_state(pf);
 
 	return 0;
 }
-- 
2.20.1

[net-next 11/16] ice: clear VF ARQLEN register on reset

[Jeff Kirsher]

From: Mitch Williams <mitch.a.williams@intel.com>

On older devices like X710 and X722, the VF's ARQLEN register is cleared
on reset, so the VF driver uses that register to detect an unannounced
reset. Unfortunately, on devices controlled by ice, this register is NOT
cleared on reset. This causes the VF to miss resets, and even on
properly-announced resets, the VF driver complains that it didn't see
the reset.

To fix this, we'll do it in software. When we handle a VF reset (whether
triggered by software or VFLR), clear this register after the HW reset
is complete.

Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_hw_autogen.h  | 1 +
 drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c | 5 +++++
 2 files changed, 6 insertions(+)

diff --git a/drivers/net/ethernet/intel/ice/ice_hw_autogen.h b/drivers/net/ethernet/intel/ice/ice_hw_autogen.h
index f9a38f2cd470..6bf5cc064270 100644
--- a/drivers/net/ethernet/intel/ice/ice_hw_autogen.h
+++ b/drivers/net/ethernet/intel/ice/ice_hw_autogen.h
@@ -30,6 +30,7 @@
 #define PF_FW_ATQLEN_ATQVFE_M			BIT(28)
 #define PF_FW_ATQLEN_ATQOVFL_M			BIT(29)
 #define PF_FW_ATQLEN_ATQCRIT_M			BIT(30)
+#define VF_MBX_ARQLEN(_VF)			(0x0022BC00 + ((_VF) * 4))
 #define PF_FW_ATQLEN_ATQENABLE_M		BIT(31)
 #define PF_FW_ATQT				0x00080400
 #define PF_MBX_ARQBAH				0x0022E400
diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
index 79d793a76042..458e179ea863 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
@@ -310,6 +310,11 @@ static void ice_trigger_vf_reset(struct ice_vf *vf, bool is_vflr)
 	 */
 	clear_bit(ICE_VF_STATE_INIT, vf->vf_states);
 
+	/* Clear the VF's ARQLEN register. This is how the VF detects reset,
+	 * since the VFGEN_RSTAT register doesn't stick at 0 after reset.
+	 */
+	wr32(hw, VF_MBX_ARQLEN(vf_abs_id), 0);
+
 	/* In the case of a VFLR, the HW has already reset the VF and we
 	 * just need to clean up, so don't hit the VFRTRIG register.
 	 */
-- 
2.20.1

[net-next 12/16] ice: flush Tx pipe on disable queue timeout

[Jeff Kirsher]

From: Victor Raj <victor.raj@intel.com>

Set the flush Tx pipe flag instead of getting an EAGAIN error when FW
times out in processing the disable Tx queue command.

Signed-off-by: Victor Raj <victor.raj@intel.com>
Reviewed-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_common.c | 21 +++++++++++++++++++--
 1 file changed, 19 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_common.c b/drivers/net/ethernet/intel/ice/ice_common.c
index b17ade424423..63f003441300 100644
--- a/drivers/net/ethernet/intel/ice/ice_common.c
+++ b/drivers/net/ethernet/intel/ice/ice_common.c
@@ -2450,6 +2450,7 @@ ice_aq_dis_lan_txq(struct ice_hw *hw, u8 num_qgrps,
 {
 	struct ice_aqc_dis_txqs *cmd;
 	struct ice_aq_desc desc;
+	enum ice_status status;
 	u16 i, sz = 0;
 
 	cmd = &desc.params.dis_txqs;
@@ -2485,6 +2486,8 @@ ice_aq_dis_lan_txq(struct ice_hw *hw, u8 num_qgrps,
 		break;
 	}
 
+	/* flush pipe on time out */
+	cmd->cmd_type |= ICE_AQC_Q_DIS_CMD_FLUSH_PIPE;
 	/* If no queue group info, we are in a reset flow. Issue the AQ */
 	if (!qg_list)
 		goto do_aq;
@@ -2510,7 +2513,17 @@ ice_aq_dis_lan_txq(struct ice_hw *hw, u8 num_qgrps,
 		return ICE_ERR_PARAM;
 
 do_aq:
-	return ice_aq_send_cmd(hw, &desc, qg_list, buf_size, cd);
+	status = ice_aq_send_cmd(hw, &desc, qg_list, buf_size, cd);
+	if (status) {
+		if (!qg_list)
+			ice_debug(hw, ICE_DBG_SCHED, "VM%d disable failed %d\n",
+				  vmvf_num, hw->adminq.sq_last_status);
+		else
+			ice_debug(hw, ICE_DBG_SCHED, "disable Q %d failed %d\n",
+				  le16_to_cpu(qg_list[0].q_id[0]),
+				  hw->adminq.sq_last_status);
+	}
+	return status;
 }
 
 /* End of FW Admin Queue command wrappers */
@@ -2796,8 +2809,12 @@ ice_ena_vsi_txq(struct ice_port_info *pi, u16 vsi_handle, u8 tc, u8 num_qgrps,
 
 	/* add the lan q */
 	status = ice_aq_add_lan_txq(hw, num_qgrps, buf, buf_size, cd);
-	if (status)
+	if (status) {
+		ice_debug(hw, ICE_DBG_SCHED, "enable Q %d failed %d\n",
+			  le16_to_cpu(buf->txqs[0].txq_id),
+			  hw->adminq.sq_last_status);
 		goto ena_txq_exit;
+	}
 
 	node.node_teid = buf->txqs[0].q_teid;
 	node.data.elem_type = ICE_AQC_ELEM_TYPE_LEAF;
-- 
2.20.1

[net-next 13/16] ice: check for a leaf node presence

[Jeff Kirsher]

From: Victor Raj <victor.raj@intel.com>

Check for a leaf node presence for a given VSI. This check is required
before removing a VSI since VSIs can't be removed with enabled queues
(with leaf nodes) from the FW scheduler tree unless its a reset.

Signed-off-by: Victor Raj <victor.raj@intel.com>
Reviewed-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_sched.c  | 23 +++++++++++++++++++++
 drivers/net/ethernet/intel/ice/ice_status.h |  1 +
 2 files changed, 24 insertions(+)

diff --git a/drivers/net/ethernet/intel/ice/ice_sched.c b/drivers/net/ethernet/intel/ice/ice_sched.c
index f355f59d8b64..56049739a250 100644
--- a/drivers/net/ethernet/intel/ice/ice_sched.c
+++ b/drivers/net/ethernet/intel/ice/ice_sched.c
@@ -1615,6 +1615,23 @@ ice_sched_rm_agg_vsi_info(struct ice_port_info *pi, u16 vsi_handle)
 	}
 }
 
+/**
+ * ice_sched_is_leaf_node_present - check for a leaf node in the sub-tree
+ * @node: pointer to the sub-tree node
+ *
+ * This function checks for a leaf node presence in a given sub-tree node.
+ */
+static bool ice_sched_is_leaf_node_present(struct ice_sched_node *node)
+{
+	u8 i;
+
+	for (i = 0; i < node->num_children; i++)
+		if (ice_sched_is_leaf_node_present(node->children[i]))
+			return true;
+	/* check for a leaf node */
+	return (node->info.data.elem_type == ICE_AQC_ELEM_TYPE_LEAF);
+}
+
 /**
  * ice_sched_rm_vsi_cfg - remove the VSI and its children nodes
  * @pi: port information structure
@@ -1649,6 +1666,12 @@ ice_sched_rm_vsi_cfg(struct ice_port_info *pi, u16 vsi_handle, u8 owner)
 		if (!vsi_node)
 			continue;
 
+		if (ice_sched_is_leaf_node_present(vsi_node)) {
+			ice_debug(pi->hw, ICE_DBG_SCHED,
+				  "VSI has leaf nodes in TC %d\n", i);
+			status = ICE_ERR_IN_USE;
+			goto exit_sched_rm_vsi_cfg;
+		}
 		while (j < vsi_node->num_children) {
 			if (vsi_node->children[j]->owner == owner) {
 				ice_free_sched_node(pi, vsi_node->children[j]);
diff --git a/drivers/net/ethernet/intel/ice/ice_status.h b/drivers/net/ethernet/intel/ice/ice_status.h
index f49f299ddf2c..683f48824a29 100644
--- a/drivers/net/ethernet/intel/ice/ice_status.h
+++ b/drivers/net/ethernet/intel/ice/ice_status.h
@@ -22,6 +22,7 @@ enum ice_status {
 	ICE_ERR_OUT_OF_RANGE			= -13,
 	ICE_ERR_ALREADY_EXISTS			= -14,
 	ICE_ERR_DOES_NOT_EXIST			= -15,
+	ICE_ERR_IN_USE				= -16,
 	ICE_ERR_MAX_LIMIT			= -17,
 	ICE_ERR_RESET_ONGOING			= -18,
 	ICE_ERR_BUF_TOO_SHORT			= -52,
-- 
2.20.1

[net-next 14/16] ice: use absolute vector ID for VFs

[Jeff Kirsher]

From: Mitch Williams <mitch.a.williams@intel.com>

When the PF driver sets up the VF MSI-X vector allocation, it needs to
use the hardware absolute vector ID, not the per-PF vector ID. Without
this change we see (apparent) TX hangs when using VFs on multiple PFs.

Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
index 458e179ea863..57155b4a59dc 100644
--- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c
@@ -173,7 +173,8 @@ static void ice_dis_vf_mappings(struct ice_vf *vf)
 	wr32(hw, VPINT_ALLOC(vf->vf_id), 0);
 	wr32(hw, VPINT_ALLOC_PCI(vf->vf_id), 0);
 
-	first = vf->first_vector_idx;
+	first = vf->first_vector_idx +
+		hw->func_caps.common_cap.msix_vector_first_id;
 	last = first + pf->num_vf_msix - 1;
 	for (v = first; v <= last; v++) {
 		u32 reg;
@@ -523,7 +524,8 @@ static void ice_ena_vf_mappings(struct ice_vf *vf)
 
 	hw = &pf->hw;
 	vsi = pf->vsi[vf->lan_vsi_idx];
-	first = vf->first_vector_idx;
+	first = vf->first_vector_idx +
+		hw->func_caps.common_cap.msix_vector_first_id;
 	last = (first + pf->num_vf_msix) - 1;
 	abs_vf_id = vf->vf_id + hw->func_caps.vf_base_id;
 
-- 
2.20.1

[net-next 15/16] ice: Fix for FC get rx/tx pause params

[Jeff Kirsher]

From: Lukasz Czapnik <lukasz.czapnik@intel.com>

Ethtool reported pause params based on the currently negotiated
link settings instead of current PHY config. User was not able
to turn off pause params because ethtool was incorrectly reporting
parameters as off when link was down even though PHY was configured
to support pause frames. Now pause params are taken from PHY config
instead of link status.

Signed-off-by: Lukasz Czapnik <lukasz.czapnik@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_ethtool.c | 37 ++++++++++++++------
 1 file changed, 26 insertions(+), 11 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_ethtool.c b/drivers/net/ethernet/intel/ice/ice_ethtool.c
index a93daf660171..295571b1c4e8 100644
--- a/drivers/net/ethernet/intel/ice/ice_ethtool.c
+++ b/drivers/net/ethernet/intel/ice/ice_ethtool.c
@@ -1818,21 +1818,36 @@ static void
 ice_get_pauseparam(struct net_device *netdev, struct ethtool_pauseparam *pause)
 {
 	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_port_info *pi;
+	struct ice_port_info *pi = np->vsi->port_info;
+	struct ice_aqc_get_phy_caps_data *pcaps;
+	struct ice_vsi *vsi = np->vsi;
+	enum ice_status status;
 
-	pi = np->vsi->port_info;
-	pause->autoneg =
-		((pi->phy.link_info.an_info & ICE_AQ_AN_COMPLETED) ?
-		 AUTONEG_ENABLE : AUTONEG_DISABLE);
+	/* Initialize pause params */
+	pause->rx_pause = 0;
+	pause->tx_pause = 0;
 
-	if (pi->fc.current_mode == ICE_FC_RX_PAUSE) {
-		pause->rx_pause = 1;
-	} else if (pi->fc.current_mode == ICE_FC_TX_PAUSE) {
+	pcaps = devm_kzalloc(&vsi->back->pdev->dev, sizeof(*pcaps),
+			     GFP_KERNEL);
+	if (!pcaps)
+		return;
+
+	/* Get current phy config */
+	status = ice_aq_get_phy_caps(pi, false, ICE_AQC_REPORT_SW_CFG, pcaps,
+				     NULL);
+	if (status)
+		goto out;
+
+	pause->autoneg = ((pcaps->caps & ICE_AQC_PHY_AN_MODE) ?
+			AUTONEG_ENABLE : AUTONEG_DISABLE);
+
+	if (pcaps->caps & ICE_AQC_PHY_EN_TX_LINK_PAUSE)
 		pause->tx_pause = 1;
-	} else if (pi->fc.current_mode == ICE_FC_FULL) {
+	if (pcaps->caps & ICE_AQC_PHY_EN_RX_LINK_PAUSE)
 		pause->rx_pause = 1;
-		pause->tx_pause = 1;
-	}
+
+out:
+	devm_kfree(&vsi->back->pdev->dev, pcaps);
 }
 
 /**
-- 
2.20.1

[net-next 16/16] ice: fix overlong string, update stats output

[Jeff Kirsher]

From: Jesse Brandeburg <jesse.brandeburg@intel.com>

A test started warning on a string truncation. This led to an unfortunate
realization that we are likely not accounting for the stats length
correctly before this patch, so fix the issue by putting "port." in front
of all the PF stats, instead of magically prepending it at runtime.

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ice/ice_ethtool.c | 80 ++++++++++----------
 1 file changed, 40 insertions(+), 40 deletions(-)

diff --git a/drivers/net/ethernet/intel/ice/ice_ethtool.c b/drivers/net/ethernet/intel/ice/ice_ethtool.c
index 295571b1c4e8..eb8d149e317c 100644
--- a/drivers/net/ethernet/intel/ice/ice_ethtool.c
+++ b/drivers/net/ethernet/intel/ice/ice_ethtool.c
@@ -63,45 +63,45 @@ static const struct ice_stats ice_gstrings_vsi_stats[] = {
  * is queried on the base PF netdev.
  */
 static const struct ice_stats ice_gstrings_pf_stats[] = {
-	ICE_PF_STAT("tx_bytes", stats.eth.tx_bytes),
-	ICE_PF_STAT("rx_bytes", stats.eth.rx_bytes),
-	ICE_PF_STAT("tx_unicast", stats.eth.tx_unicast),
-	ICE_PF_STAT("rx_unicast", stats.eth.rx_unicast),
-	ICE_PF_STAT("tx_multicast", stats.eth.tx_multicast),
-	ICE_PF_STAT("rx_multicast", stats.eth.rx_multicast),
-	ICE_PF_STAT("tx_broadcast", stats.eth.tx_broadcast),
-	ICE_PF_STAT("rx_broadcast", stats.eth.rx_broadcast),
-	ICE_PF_STAT("tx_errors", stats.eth.tx_errors),
-	ICE_PF_STAT("tx_size_64", stats.tx_size_64),
-	ICE_PF_STAT("rx_size_64", stats.rx_size_64),
-	ICE_PF_STAT("tx_size_127", stats.tx_size_127),
-	ICE_PF_STAT("rx_size_127", stats.rx_size_127),
-	ICE_PF_STAT("tx_size_255", stats.tx_size_255),
-	ICE_PF_STAT("rx_size_255", stats.rx_size_255),
-	ICE_PF_STAT("tx_size_511", stats.tx_size_511),
-	ICE_PF_STAT("rx_size_511", stats.rx_size_511),
-	ICE_PF_STAT("tx_size_1023", stats.tx_size_1023),
-	ICE_PF_STAT("rx_size_1023", stats.rx_size_1023),
-	ICE_PF_STAT("tx_size_1522", stats.tx_size_1522),
-	ICE_PF_STAT("rx_size_1522", stats.rx_size_1522),
-	ICE_PF_STAT("tx_size_big", stats.tx_size_big),
-	ICE_PF_STAT("rx_size_big", stats.rx_size_big),
-	ICE_PF_STAT("link_xon_tx", stats.link_xon_tx),
-	ICE_PF_STAT("link_xon_rx", stats.link_xon_rx),
-	ICE_PF_STAT("link_xoff_tx", stats.link_xoff_tx),
-	ICE_PF_STAT("link_xoff_rx", stats.link_xoff_rx),
-	ICE_PF_STAT("tx_dropped_link_down", stats.tx_dropped_link_down),
-	ICE_PF_STAT("rx_undersize", stats.rx_undersize),
-	ICE_PF_STAT("rx_fragments", stats.rx_fragments),
-	ICE_PF_STAT("rx_oversize", stats.rx_oversize),
-	ICE_PF_STAT("rx_jabber", stats.rx_jabber),
-	ICE_PF_STAT("rx_csum_bad", hw_csum_rx_error),
-	ICE_PF_STAT("rx_length_errors", stats.rx_len_errors),
-	ICE_PF_STAT("rx_dropped", stats.eth.rx_discards),
-	ICE_PF_STAT("rx_crc_errors", stats.crc_errors),
-	ICE_PF_STAT("illegal_bytes", stats.illegal_bytes),
-	ICE_PF_STAT("mac_local_faults", stats.mac_local_faults),
-	ICE_PF_STAT("mac_remote_faults", stats.mac_remote_faults),
+	ICE_PF_STAT("port.tx_bytes", stats.eth.tx_bytes),
+	ICE_PF_STAT("port.rx_bytes", stats.eth.rx_bytes),
+	ICE_PF_STAT("port.tx_unicast", stats.eth.tx_unicast),
+	ICE_PF_STAT("port.rx_unicast", stats.eth.rx_unicast),
+	ICE_PF_STAT("port.tx_multicast", stats.eth.tx_multicast),
+	ICE_PF_STAT("port.rx_multicast", stats.eth.rx_multicast),
+	ICE_PF_STAT("port.tx_broadcast", stats.eth.tx_broadcast),
+	ICE_PF_STAT("port.rx_broadcast", stats.eth.rx_broadcast),
+	ICE_PF_STAT("port.tx_errors", stats.eth.tx_errors),
+	ICE_PF_STAT("port.tx_size_64", stats.tx_size_64),
+	ICE_PF_STAT("port.rx_size_64", stats.rx_size_64),
+	ICE_PF_STAT("port.tx_size_127", stats.tx_size_127),
+	ICE_PF_STAT("port.rx_size_127", stats.rx_size_127),
+	ICE_PF_STAT("port.tx_size_255", stats.tx_size_255),
+	ICE_PF_STAT("port.rx_size_255", stats.rx_size_255),
+	ICE_PF_STAT("port.tx_size_511", stats.tx_size_511),
+	ICE_PF_STAT("port.rx_size_511", stats.rx_size_511),
+	ICE_PF_STAT("port.tx_size_1023", stats.tx_size_1023),
+	ICE_PF_STAT("port.rx_size_1023", stats.rx_size_1023),
+	ICE_PF_STAT("port.tx_size_1522", stats.tx_size_1522),
+	ICE_PF_STAT("port.rx_size_1522", stats.rx_size_1522),
+	ICE_PF_STAT("port.tx_size_big", stats.tx_size_big),
+	ICE_PF_STAT("port.rx_size_big", stats.rx_size_big),
+	ICE_PF_STAT("port.link_xon_tx", stats.link_xon_tx),
+	ICE_PF_STAT("port.link_xon_rx", stats.link_xon_rx),
+	ICE_PF_STAT("port.link_xoff_tx", stats.link_xoff_tx),
+	ICE_PF_STAT("port.link_xoff_rx", stats.link_xoff_rx),
+	ICE_PF_STAT("port.tx_dropped_link_down", stats.tx_dropped_link_down),
+	ICE_PF_STAT("port.rx_undersize", stats.rx_undersize),
+	ICE_PF_STAT("port.rx_fragments", stats.rx_fragments),
+	ICE_PF_STAT("port.rx_oversize", stats.rx_oversize),
+	ICE_PF_STAT("port.rx_jabber", stats.rx_jabber),
+	ICE_PF_STAT("port.rx_csum_bad", hw_csum_rx_error),
+	ICE_PF_STAT("port.rx_length_errors", stats.rx_len_errors),
+	ICE_PF_STAT("port.rx_dropped", stats.eth.rx_discards),
+	ICE_PF_STAT("port.rx_crc_errors", stats.crc_errors),
+	ICE_PF_STAT("port.illegal_bytes", stats.illegal_bytes),
+	ICE_PF_STAT("port.mac_local_faults", stats.mac_local_faults),
+	ICE_PF_STAT("port.mac_remote_faults", stats.mac_remote_faults),
 };
 
 static const u32 ice_regs_dump_list[] = {
@@ -304,7 +304,7 @@ static void ice_get_strings(struct net_device *netdev, u32 stringset, u8 *data)
 			return;
 
 		for (i = 0; i < ICE_PF_STATS_LEN; i++) {
-			snprintf(p, ETH_GSTRING_LEN, "port.%s",
+			snprintf(p, ETH_GSTRING_LEN, "%s",
 				 ice_gstrings_pf_stats[i].stat_string);
 			p += ETH_GSTRING_LEN;
 		}
-- 
2.20.1

Re: [net-next 00/16][pull request] 100GbE Intel Wired LAN Driver Updates 2019-02-22

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Mon, 25 Feb 2019 10:42:50 -0800

> This series contains updates to the ice driver only.

Pulled, thanks Jeff.

Re: [net-next 01/16] ice: Mark extack argument as __always_unused

[Stephen Hemminger]

On Mon, 25 Feb 2019 10:42:51 -0800
Jeff Kirsher <jeffrey.t.kirsher@intel.com> wrote:

> From: Bruce Allan <bruce.w.allan@intel.com>
> 
> Commit 87b0984ebfab ("net: Add extack argument to ndo_fdb_add()") in
> net-next added an extended parameter to the .ndo_fdb_add op and changed
> ice_fdb_add() accordingly. Update the function header and add the
> __always_unused attribute to the new parameter to avoid -Wunused-parameter
> warnings.
> 
> Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
> Signed-off-by: Anirudh Venkataramanan <anirudh.venkataramanan@intel.com>
> Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
> ---
>  drivers/net/ethernet/intel/ice/ice_main.c | 9 +++++----
>  1 file changed, 5 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
> index 48f033928aa2..9d266d754445 100644
> --- a/drivers/net/ethernet/intel/ice/ice_main.c
> +++ b/drivers/net/ethernet/intel/ice/ice_main.c
> @@ -2435,11 +2435,12 @@ static void ice_set_rx_mode(struct net_device *netdev)
>   * @addr: the MAC address entry being added
>   * @vid: VLAN id
>   * @flags: instructions from stack about fdb operation
> + * @extack: netlink extended ack
>   */
> -static int ice_fdb_add(struct ndmsg *ndm, struct nlattr __always_unused *tb[],
> -		       struct net_device *dev, const unsigned char *addr,
> -		       u16 vid, u16 flags,
> -		       struct netlink_ext_ack *extack)
> +static int
> +ice_fdb_add(struct ndmsg *ndm, struct nlattr __always_unused *tb[],
> +	    struct net_device *dev, const unsigned char *addr, u16 vid,
> +	    u16 flags, struct netlink_ext_ack __always_unused *extack)
>  {
>  	int err;
>  

There are several log messages in ice_fdb_add, why don't you convert those
to be user friendly and use the extended ack mechanism?