[net 00/13][pull request] Intel Wired LAN Driver Fixes 2017-11-21

[net 00/13][pull request] Intel Wired LAN Driver Fixes 2017-11-21

[Jeff Kirsher]

This series contains fixes for igb/vf, ixgbe/vf, i40e/vf and fm10k.

Jake fixes a regression issue with older firmware, where we were using
the NVM lock to synchronize NVM reads for all devices and firmware
versions, yet this caused issues with older firmware prior to version
1.5.  Fixed this by only grabbing the lock for newer devices and firmware
version 1.5 or newer.

Zijie Pan fixes the calculation of the i40e VF MAC addresses, where it was
possible to increment to the next MAC entry without calling
i40e_add_mac_filter().

Amritha removes the upper limit of 64 queues on a channel VSI since the
upper bound is determined by the VSI's num_queue_pairs.

Filip fixes an issue during FLR resets, where should have been checking
for upcoming core reset and if so, just return with I40E_ERR_NOT_READY.

Alan fixes the notifying clients of l2 parameters by copying the
parameters to the client instance struct and re-organizes the priority
in which the client tasks fire so that if the flag for notifying l2
params is set, it will trigger before the client open task.  Also fixed
the promiscuous settings after reset for all the VSI's.

Brian King from IBM fixes an issue seen on Power systems which would
result in skb list corruption and eventual kernel oops.  Brian
provides the same fix for nearly all our drivers, to replace the
read_barrier_depends with smp_rmb() to ensure loads are ordered with
respect to the load of tx_buffer->next_to_watch.

The following are changes since commit 0c86a6bd85ff0629cd2c5141027fc1c8bb6cde9c:
  Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
and are available in the git repository at:
  git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-queue 40GbE

Alan Brady (2):
  i40evf: fix client notify of l2 params
  i40e: restore promiscuous after reset

Amritha Nambiar (1):
  i40e: Remove limit of 64 max queues per channel

Brian King (7):
  ixgbe: Fix skb list corruption on Power systems
  i40e: Use smp_rmb rather than read_barrier_depends
  ixgbevf: Use smp_rmb rather than read_barrier_depends
  igbvf: Use smp_rmb rather than read_barrier_depends
  igb: Use smp_rmb rather than read_barrier_depends
  fm10k: Use smp_rmb rather than read_barrier_depends
  i40evf: Use smp_rmb rather than read_barrier_depends

Filip Sadowski (1):
  i40e: Fix FLR reset timeout issue

Jacob Keller (1):
  i40e: Fix for NUP NVM image downgrade failure

Zijie Pan (1):
  i40e: fix the calculation of VFs mac addresses

 drivers/net/ethernet/intel/fm10k/fm10k_main.c      |   2 +-
 drivers/net/ethernet/intel/i40e/i40e.h             |   1 -
 drivers/net/ethernet/intel/i40e/i40e_adminq.c      |   6 +
 drivers/net/ethernet/intel/i40e/i40e_common.c      |  10 +-
 drivers/net/ethernet/intel/i40e/i40e_main.c        | 165 +++++++++++----------
 drivers/net/ethernet/intel/i40e/i40e_nvm.c         |   8 +-
 drivers/net/ethernet/intel/i40e/i40e_txrx.c        |   2 +-
 drivers/net/ethernet/intel/i40e/i40e_type.h        |   1 +
 drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c |  21 +--
 drivers/net/ethernet/intel/i40evf/i40e_txrx.c      |   2 +-
 drivers/net/ethernet/intel/i40evf/i40evf_client.c  |  38 +++--
 drivers/net/ethernet/intel/i40evf/i40evf_main.c    |  10 +-
 drivers/net/ethernet/intel/igb/igb_main.c          |   2 +-
 drivers/net/ethernet/intel/igbvf/netdev.c          |   2 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c      |   2 +-
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c  |   2 +-
 16 files changed, 154 insertions(+), 120 deletions(-)

-- 
2.15.0

[net 01/13] i40e: Fix for NUP NVM image downgrade failure

[Jeff Kirsher]

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

Since commit 96a39aed25e6 ("i40e: Acquire NVM lock before
reads on all devices") we've used the NVM lock
to synchronize NVM reads even on devices which don't strictly
need the lock.

Doing so can cause a regression on older firmware prior to 1.5,
especially when downgrading the firmware.

Fix this by only grabbing the lock if we're running on an X722
device (which requires the lock as it uses the AdminQ to read
the NVM), or if we're currently running 1.5 or newer firmware.

Signed-off-by: Jacob Keller <jacob.e.keller@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_adminq.c | 6 ++++++
 drivers/net/ethernet/intel/i40e/i40e_common.c | 3 ++-
 drivers/net/ethernet/intel/i40e/i40e_nvm.c    | 8 +++++---
 drivers/net/ethernet/intel/i40e/i40e_type.h   | 1 +
 4 files changed, 14 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40e/i40e_adminq.c b/drivers/net/ethernet/intel/i40e/i40e_adminq.c
index 9dcb2a961197..9af74253c3f7 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_adminq.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_adminq.c
@@ -613,6 +613,12 @@ i40e_status i40e_init_adminq(struct i40e_hw *hw)
 		hw->flags |= I40E_HW_FLAG_AQ_PHY_ACCESS_CAPABLE;
 	}
 
+	/* Newer versions of firmware require lock when reading the NVM */
+	if (hw->aq.api_maj_ver > 1 ||
+	    (hw->aq.api_maj_ver == 1 &&
+	     hw->aq.api_min_ver >= 5))
+		hw->flags |= I40E_HW_FLAG_NVM_READ_REQUIRES_LOCK;
+
 	/* The ability to RX (not drop) 802.1ad frames was added in API 1.7 */
 	if (hw->aq.api_maj_ver > 1 ||
 	    (hw->aq.api_maj_ver == 1 &&
diff --git a/drivers/net/ethernet/intel/i40e/i40e_common.c b/drivers/net/ethernet/intel/i40e/i40e_common.c
index 0203665cb53c..13c79468a6da 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_common.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_common.c
@@ -948,7 +948,8 @@ i40e_status i40e_init_shared_code(struct i40e_hw *hw)
 		hw->pf_id = (u8)(func_rid & 0x7);
 
 	if (hw->mac.type == I40E_MAC_X722)
-		hw->flags |= I40E_HW_FLAG_AQ_SRCTL_ACCESS_ENABLE;
+		hw->flags |= I40E_HW_FLAG_AQ_SRCTL_ACCESS_ENABLE |
+			     I40E_HW_FLAG_NVM_READ_REQUIRES_LOCK;
 
 	status = i40e_init_nvm(hw);
 	return status;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_nvm.c b/drivers/net/ethernet/intel/i40e/i40e_nvm.c
index 0ccab0a5d717..7689c2ee0d46 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_nvm.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_nvm.c
@@ -328,15 +328,17 @@ static i40e_status __i40e_read_nvm_word(struct i40e_hw *hw,
 i40e_status i40e_read_nvm_word(struct i40e_hw *hw, u16 offset,
 			       u16 *data)
 {
-	i40e_status ret_code;
+	i40e_status ret_code = 0;
 
-	ret_code = i40e_acquire_nvm(hw, I40E_RESOURCE_READ);
+	if (hw->flags & I40E_HW_FLAG_NVM_READ_REQUIRES_LOCK)
+		ret_code = i40e_acquire_nvm(hw, I40E_RESOURCE_READ);
 	if (ret_code)
 		return ret_code;
 
 	ret_code = __i40e_read_nvm_word(hw, offset, data);
 
-	i40e_release_nvm(hw);
+	if (hw->flags & I40E_HW_FLAG_NVM_READ_REQUIRES_LOCK)
+		i40e_release_nvm(hw);
 
 	return ret_code;
 }
diff --git a/drivers/net/ethernet/intel/i40e/i40e_type.h b/drivers/net/ethernet/intel/i40e/i40e_type.h
index 00d4833e9925..0e8568719b4e 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_type.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_type.h
@@ -629,6 +629,7 @@ struct i40e_hw {
 #define I40E_HW_FLAG_AQ_SRCTL_ACCESS_ENABLE BIT_ULL(0)
 #define I40E_HW_FLAG_802_1AD_CAPABLE        BIT_ULL(1)
 #define I40E_HW_FLAG_AQ_PHY_ACCESS_CAPABLE  BIT_ULL(2)
+#define I40E_HW_FLAG_NVM_READ_REQUIRES_LOCK BIT_ULL(3)
 	u64 flags;
 
 	/* Used in set switch config AQ command */
-- 
2.15.0

[net 02/13] i40e: fix the calculation of VFs mac addresses

[Jeff Kirsher]

From: Zijie Pan <zijie.pan@6wind.com>

num_mac should be increased only after the call to i40e_add_mac_filter().

Fixes: 5f527ba962e2 ("i40e: Limit the number of MAC and VLAN addresses that can be added for VFs")
Signed-off-by: Zijie Pan <zijie.pan@6wind.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Reviewed-by: Tushar Dave <tushar.n.dave@oracle.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 | 21 +++++++++++----------
 1 file changed, 11 insertions(+), 10 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
index f8a794b72462..a3dc9b932946 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
@@ -2218,18 +2218,19 @@ static int i40e_vc_add_mac_addr_msg(struct i40e_vf *vf, u8 *msg, u16 msglen)
 		struct i40e_mac_filter *f;
 
 		f = i40e_find_mac(vsi, al->list[i].addr);
-		if (!f)
+		if (!f) {
 			f = i40e_add_mac_filter(vsi, al->list[i].addr);
 
-		if (!f) {
-			dev_err(&pf->pdev->dev,
-				"Unable to add MAC filter %pM for VF %d\n",
-				 al->list[i].addr, vf->vf_id);
-			ret = I40E_ERR_PARAM;
-			spin_unlock_bh(&vsi->mac_filter_hash_lock);
-			goto error_param;
-		} else {
-			vf->num_mac++;
+			if (!f) {
+				dev_err(&pf->pdev->dev,
+					"Unable to add MAC filter %pM for VF %d\n",
+					al->list[i].addr, vf->vf_id);
+				ret = I40E_ERR_PARAM;
+				spin_unlock_bh(&vsi->mac_filter_hash_lock);
+				goto error_param;
+			} else {
+				vf->num_mac++;
+			}
 		}
 	}
 	spin_unlock_bh(&vsi->mac_filter_hash_lock);
-- 
2.15.0

[net 03/13] i40e: Remove limit of 64 max queues per channel

[Jeff Kirsher]

From: Amritha Nambiar <amritha.nambiar@intel.com>

It is safe to remove the upper limit of 64 queues on a channel
VSI. The upper bound is determined by the VSI's num_queue_pairs
and gets validated when the queue mapping info through mqprio
interface is subject to bound checking in the driver.

Signed-off-by: Amritha Nambiar <amritha.nambiar@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      | 1 -
 drivers/net/ethernet/intel/i40e/i40e_main.c | 8 --------
 2 files changed, 9 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
index 5829715fa342..e019baa905c5 100644
--- a/drivers/net/ethernet/intel/i40e/i40e.h
+++ b/drivers/net/ethernet/intel/i40e/i40e.h
@@ -90,7 +90,6 @@
 #define I40E_AQ_LEN			256
 #define I40E_AQ_WORK_LIMIT		66 /* max number of VFs + a little */
 #define I40E_MAX_USER_PRIORITY		8
-#define I40E_MAX_QUEUES_PER_CH		64
 #define I40E_DEFAULT_TRAFFIC_CLASS	BIT(0)
 #define I40E_DEFAULT_MSG_ENABLE		4
 #define I40E_QUEUE_WAIT_RETRY_LIMIT	10
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 4a964d6e4a9e..173e924d4dae 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -5629,14 +5629,6 @@ static int i40e_validate_num_queues(struct i40e_pf *pf, int num_queues,
 		return -EINVAL;
 
 	*reconfig_rss = false;
-
-	if (num_queues > I40E_MAX_QUEUES_PER_CH) {
-		dev_err(&pf->pdev->dev,
-			"Failed to create VMDq VSI. User requested num_queues (%d) > I40E_MAX_QUEUES_PER_VSI (%u)\n",
-			num_queues, I40E_MAX_QUEUES_PER_CH);
-		return -EINVAL;
-	}
-
 	if (vsi->current_rss_size) {
 		if (num_queues > vsi->current_rss_size) {
 			dev_dbg(&pf->pdev->dev,
-- 
2.15.0

[net 04/13] i40e: Fix FLR reset timeout issue

[Jeff Kirsher]

From: Filip Sadowski <filip.sadowski@intel.com>

This patch allows detection of upcoming core reset in case NIC gets
stuck while performing FLR reset. The i40e_pf_reset() function returns
I40E_ERR_NOT_READY when global reset was detected.

Signed-off-by: Filip Sadowski <filip.sadowski@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_common.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/net/ethernet/intel/i40e/i40e_common.c b/drivers/net/ethernet/intel/i40e/i40e_common.c
index 13c79468a6da..095965f268bd 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_common.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_common.c
@@ -1269,6 +1269,7 @@ i40e_status i40e_pf_reset(struct i40e_hw *hw)
 	 * we don't need to do the PF Reset
 	 */
 	if (!cnt) {
+		u32 reg2 = 0;
 		if (hw->revision_id == 0)
 			cnt = I40E_PF_RESET_WAIT_COUNT_A0;
 		else
@@ -1280,6 +1281,12 @@ i40e_status i40e_pf_reset(struct i40e_hw *hw)
 			reg = rd32(hw, I40E_PFGEN_CTRL);
 			if (!(reg & I40E_PFGEN_CTRL_PFSWR_MASK))
 				break;
+			reg2 = rd32(hw, I40E_GLGEN_RSTAT);
+			if (reg2 & I40E_GLGEN_RSTAT_DEVSTATE_MASK) {
+				hw_dbg(hw, "Core reset upcoming. Skipping PF reset request.\n");
+				hw_dbg(hw, "I40E_GLGEN_RSTAT = 0x%x\n", reg2);
+				return I40E_ERR_NOT_READY;
+			}
 			usleep_range(1000, 2000);
 		}
 		if (reg & I40E_PFGEN_CTRL_PFSWR_MASK) {
-- 
2.15.0

[net 05/13] i40evf: fix client notify of l2 params

[Jeff Kirsher]

From: Alan Brady <alan.brady@intel.com>

The current method for notifying clients of l2 parameters is broken
because we fail to copy the new parameters to the client instance
struct, we need to do the notification before the client 'open' function
pointer gets called, and lastly we should set the l2 parameters when
first adding a client instance.

This patch first introduces the i40evf_client_get_params function to
prevent code duplication in the i40evf_client_add_instance and the
i40evf_notify_client_l2_params functions.  We then fix the notify l2
params function to actually copy the parameters to client instance
struct and do the same in the *_add_instance' function.  Lastly this
patch reorganizes the priority in which client tasks fire so that if the
flag for notifying l2 params is set, it will trigger before the open
because the client needs these new parameters as part of a client open
task.

Signed-off-by: Alan Brady <alan.brady@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_client.c | 38 ++++++++++++++++-------
 drivers/net/ethernet/intel/i40evf/i40evf_main.c   | 10 +++---
 2 files changed, 31 insertions(+), 17 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_client.c b/drivers/net/ethernet/intel/i40evf/i40evf_client.c
index d8131139565e..da60ce12b33d 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_client.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_client.c
@@ -25,6 +25,26 @@ static struct i40e_ops i40evf_lan_ops = {
 	.setup_qvlist = i40evf_client_setup_qvlist,
 };
 
+/**
+ * i40evf_client_get_params - retrieve relevant client parameters
+ * @vsi: VSI with parameters
+ * @params: client param struct
+ **/
+static
+void i40evf_client_get_params(struct i40e_vsi *vsi, struct i40e_params *params)
+{
+	int i;
+
+	memset(params, 0, sizeof(struct i40e_params));
+	params->mtu = vsi->netdev->mtu;
+	params->link_up = vsi->back->link_up;
+
+	for (i = 0; i < I40E_MAX_USER_PRIORITY; i++) {
+		params->qos.prio_qos[i].tc = 0;
+		params->qos.prio_qos[i].qs_handle = vsi->qs_handle;
+	}
+}
+
 /**
  * i40evf_notify_client_message - call the client message receive callback
  * @vsi: the VSI associated with this client
@@ -66,10 +86,6 @@ void i40evf_notify_client_l2_params(struct i40e_vsi *vsi)
 		return;
 
 	cinst = vsi->back->cinst;
-	memset(&params, 0, sizeof(params));
-	params.mtu = vsi->netdev->mtu;
-	params.link_up = vsi->back->link_up;
-	params.qos.prio_qos[0].qs_handle = vsi->qs_handle;
 
 	if (!cinst || !cinst->client || !cinst->client->ops ||
 	    !cinst->client->ops->l2_param_change) {
@@ -77,6 +93,8 @@ void i40evf_notify_client_l2_params(struct i40e_vsi *vsi)
 			"Cannot locate client instance l2_param_change function\n");
 		return;
 	}
+	i40evf_client_get_params(vsi, &params);
+	cinst->lan_info.params = params;
 	cinst->client->ops->l2_param_change(&cinst->lan_info, cinst->client,
 					    &params);
 }
@@ -166,9 +184,9 @@ static struct i40e_client_instance *
 i40evf_client_add_instance(struct i40evf_adapter *adapter)
 {
 	struct i40e_client_instance *cinst = NULL;
-	struct netdev_hw_addr *mac = NULL;
 	struct i40e_vsi *vsi = &adapter->vsi;
-	int i;
+	struct netdev_hw_addr *mac = NULL;
+	struct i40e_params params;
 
 	if (!vf_registered_client)
 		goto out;
@@ -192,18 +210,14 @@ i40evf_client_add_instance(struct i40evf_adapter *adapter)
 	cinst->lan_info.version.major = I40EVF_CLIENT_VERSION_MAJOR;
 	cinst->lan_info.version.minor = I40EVF_CLIENT_VERSION_MINOR;
 	cinst->lan_info.version.build = I40EVF_CLIENT_VERSION_BUILD;
+	i40evf_client_get_params(vsi, &params);
+	cinst->lan_info.params = params;
 	set_bit(__I40E_CLIENT_INSTANCE_NONE, &cinst->state);
 
 	cinst->lan_info.msix_count = adapter->num_iwarp_msix;
 	cinst->lan_info.msix_entries =
 			&adapter->msix_entries[adapter->iwarp_base_vector];
 
-	for (i = 0; i < I40E_MAX_USER_PRIORITY; i++) {
-		cinst->lan_info.params.qos.prio_qos[i].tc = 0;
-		cinst->lan_info.params.qos.prio_qos[i].qs_handle =
-								vsi->qs_handle;
-	}
-
 	mac = list_first_entry(&cinst->lan_info.netdev->dev_addrs.list,
 			       struct netdev_hw_addr, list);
 	if (mac)
diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_main.c b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
index ca2ebdbd24d7..7b2a4eba92e2 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_main.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
@@ -2110,6 +2110,11 @@ static void i40evf_client_task(struct work_struct *work)
 		adapter->flags &= ~I40EVF_FLAG_SERVICE_CLIENT_REQUESTED;
 		goto out;
 	}
+	if (adapter->flags & I40EVF_FLAG_CLIENT_NEEDS_L2_PARAMS) {
+		i40evf_notify_client_l2_params(&adapter->vsi);
+		adapter->flags &= ~I40EVF_FLAG_CLIENT_NEEDS_L2_PARAMS;
+		goto out;
+	}
 	if (adapter->flags & I40EVF_FLAG_CLIENT_NEEDS_CLOSE) {
 		i40evf_notify_client_close(&adapter->vsi, false);
 		adapter->flags &= ~I40EVF_FLAG_CLIENT_NEEDS_CLOSE;
@@ -2118,11 +2123,6 @@ static void i40evf_client_task(struct work_struct *work)
 	if (adapter->flags & I40EVF_FLAG_CLIENT_NEEDS_OPEN) {
 		i40evf_notify_client_open(&adapter->vsi);
 		adapter->flags &= ~I40EVF_FLAG_CLIENT_NEEDS_OPEN;
-		goto out;
-	}
-	if (adapter->flags & I40EVF_FLAG_CLIENT_NEEDS_L2_PARAMS) {
-		i40evf_notify_client_l2_params(&adapter->vsi);
-		adapter->flags &= ~I40EVF_FLAG_CLIENT_NEEDS_L2_PARAMS;
 	}
 out:
 	clear_bit(__I40EVF_IN_CLIENT_TASK, &adapter->crit_section);
-- 
2.15.0

[net 06/13] i40e: restore promiscuous after reset

[Jeff Kirsher]

From: Alan Brady <alan.brady@intel.com>

After a reset we rebuild the VSIs which is going to clobber any
promiscuous settings we had before reset.  This makes it so that we
restore the promiscuous settings we had before reset.

Signed-off-by: Alan Brady <alan.brady@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 | 155 +++++++++++++++-------------
 1 file changed, 83 insertions(+), 72 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 173e924d4dae..775d5a125887 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -2166,6 +2166,73 @@ i40e_aqc_broadcast_filter(struct i40e_vsi *vsi, const char *vsi_name,
 	return aq_ret;
 }
 
+/**
+ * i40e_set_promiscuous - set promiscuous mode
+ * @pf: board private structure
+ * @promisc: promisc on or off
+ *
+ * There are different ways of setting promiscuous mode on a PF depending on
+ * what state/environment we're in.  This identifies and sets it appropriately.
+ * Returns 0 on success.
+ **/
+static int i40e_set_promiscuous(struct i40e_pf *pf, bool promisc)
+{
+	struct i40e_vsi *vsi = pf->vsi[pf->lan_vsi];
+	struct i40e_hw *hw = &pf->hw;
+	i40e_status aq_ret;
+
+	if (vsi->type == I40E_VSI_MAIN &&
+	    pf->lan_veb != I40E_NO_VEB &&
+	    !(pf->flags & I40E_FLAG_MFP_ENABLED)) {
+		/* set defport ON for Main VSI instead of true promisc
+		 * this way we will get all unicast/multicast and VLAN
+		 * promisc behavior but will not get VF or VMDq traffic
+		 * replicated on the Main VSI.
+		 */
+		if (promisc)
+			aq_ret = i40e_aq_set_default_vsi(hw,
+							 vsi->seid,
+							 NULL);
+		else
+			aq_ret = i40e_aq_clear_default_vsi(hw,
+							   vsi->seid,
+							   NULL);
+		if (aq_ret) {
+			dev_info(&pf->pdev->dev,
+				 "Set default VSI failed, err %s, aq_err %s\n",
+				 i40e_stat_str(hw, aq_ret),
+				 i40e_aq_str(hw, hw->aq.asq_last_status));
+		}
+	} else {
+		aq_ret = i40e_aq_set_vsi_unicast_promiscuous(
+						  hw,
+						  vsi->seid,
+						  promisc, NULL,
+						  true);
+		if (aq_ret) {
+			dev_info(&pf->pdev->dev,
+				 "set unicast promisc failed, err %s, aq_err %s\n",
+				 i40e_stat_str(hw, aq_ret),
+				 i40e_aq_str(hw, hw->aq.asq_last_status));
+		}
+		aq_ret = i40e_aq_set_vsi_multicast_promiscuous(
+						  hw,
+						  vsi->seid,
+						  promisc, NULL);
+		if (aq_ret) {
+			dev_info(&pf->pdev->dev,
+				 "set multicast promisc failed, err %s, aq_err %s\n",
+				 i40e_stat_str(hw, aq_ret),
+				 i40e_aq_str(hw, hw->aq.asq_last_status));
+		}
+	}
+
+	if (!aq_ret)
+		pf->cur_promisc = promisc;
+
+	return aq_ret;
+}
+
 /**
  * i40e_sync_vsi_filters - Update the VSI filter list to the HW
  * @vsi: ptr to the VSI
@@ -2467,81 +2534,16 @@ int i40e_sync_vsi_filters(struct i40e_vsi *vsi)
 		cur_promisc = (!!(vsi->current_netdev_flags & IFF_PROMISC) ||
 			       test_bit(__I40E_VSI_OVERFLOW_PROMISC,
 					vsi->state));
-		if ((vsi->type == I40E_VSI_MAIN) &&
-		    (pf->lan_veb != I40E_NO_VEB) &&
-		    !(pf->flags & I40E_FLAG_MFP_ENABLED)) {
-			/* set defport ON for Main VSI instead of true promisc
-			 * this way we will get all unicast/multicast and VLAN
-			 * promisc behavior but will not get VF or VMDq traffic
-			 * replicated on the Main VSI.
-			 */
-			if (pf->cur_promisc != cur_promisc) {
-				pf->cur_promisc = cur_promisc;
-				if (cur_promisc)
-					aq_ret =
-					      i40e_aq_set_default_vsi(hw,
-								      vsi->seid,
-								      NULL);
-				else
-					aq_ret =
-					    i40e_aq_clear_default_vsi(hw,
-								      vsi->seid,
-								      NULL);
-				if (aq_ret) {
-					retval = i40e_aq_rc_to_posix(aq_ret,
-							hw->aq.asq_last_status);
-					dev_info(&pf->pdev->dev,
-						 "Set default VSI failed on %s, err %s, aq_err %s\n",
-						 vsi_name,
-						 i40e_stat_str(hw, aq_ret),
-						 i40e_aq_str(hw,
-						     hw->aq.asq_last_status));
-				}
-			}
-		} else {
-			aq_ret = i40e_aq_set_vsi_unicast_promiscuous(
-							  hw,
-							  vsi->seid,
-							  cur_promisc, NULL,
-							  true);
-			if (aq_ret) {
-				retval =
-				i40e_aq_rc_to_posix(aq_ret,
-						    hw->aq.asq_last_status);
-				dev_info(&pf->pdev->dev,
-					 "set unicast promisc failed on %s, err %s, aq_err %s\n",
-					 vsi_name,
-					 i40e_stat_str(hw, aq_ret),
-					 i40e_aq_str(hw,
-						     hw->aq.asq_last_status));
-			}
-			aq_ret = i40e_aq_set_vsi_multicast_promiscuous(
-							  hw,
-							  vsi->seid,
-							  cur_promisc, NULL);
-			if (aq_ret) {
-				retval =
-				i40e_aq_rc_to_posix(aq_ret,
-						    hw->aq.asq_last_status);
-				dev_info(&pf->pdev->dev,
-					 "set multicast promisc failed on %s, err %s, aq_err %s\n",
-					 vsi_name,
-					 i40e_stat_str(hw, aq_ret),
-					 i40e_aq_str(hw,
-						     hw->aq.asq_last_status));
-			}
-		}
-		aq_ret = i40e_aq_set_vsi_broadcast(&vsi->back->hw,
-						   vsi->seid,
-						   cur_promisc, NULL);
+		aq_ret = i40e_set_promiscuous(pf, cur_promisc);
 		if (aq_ret) {
 			retval = i40e_aq_rc_to_posix(aq_ret,
-						     pf->hw.aq.asq_last_status);
+						     hw->aq.asq_last_status);
 			dev_info(&pf->pdev->dev,
-				 "set brdcast promisc failed, err %s, aq_err %s\n",
-					 i40e_stat_str(hw, aq_ret),
-					 i40e_aq_str(hw,
-						     hw->aq.asq_last_status));
+				 "Setting promiscuous %s failed on %s, err %s aq_err %s\n",
+				 cur_promisc ? "on" : "off",
+				 vsi_name,
+				 i40e_stat_str(hw, aq_ret),
+				 i40e_aq_str(hw, hw->aq.asq_last_status));
 		}
 	}
 out:
@@ -9421,6 +9423,15 @@ static void i40e_rebuild(struct i40e_pf *pf, bool reinit, bool lock_acquired)
 	if (!lock_acquired)
 		rtnl_unlock();
 
+	/* Restore promiscuous settings */
+	ret = i40e_set_promiscuous(pf, pf->cur_promisc);
+	if (ret)
+		dev_warn(&pf->pdev->dev,
+			 "Failed to restore promiscuous setting: %s, err %s aq_err %s\n",
+			 pf->cur_promisc ? "on" : "off",
+			 i40e_stat_str(&pf->hw, ret),
+			 i40e_aq_str(&pf->hw, pf->hw.aq.asq_last_status));
+
 	i40e_reset_all_vfs(pf, true);
 
 	/* tell the firmware that we're starting */
-- 
2.15.0

[net 07/13] ixgbe: Fix skb list corruption on Power systems

[Jeff Kirsher]

From: Brian King <brking@linux.vnet.ibm.com>

This patch fixes an issue seen on Power systems with ixgbe which results
in skb list corruption and an eventual kernel oops. The following is what
was observed:

CPU 1                                   CPU2
============================            ============================
1: ixgbe_xmit_frame_ring                ixgbe_clean_tx_irq
2:  first->skb = skb                     eop_desc = tx_buffer->next_to_watch
3:  ixgbe_tx_map                         read_barrier_depends()
4:   wmb                                 check adapter written status bit
5:   first->next_to_watch = tx_desc      napi_consume_skb(tx_buffer->skb ..);
6:   writel(i, tx_ring->tail);

The read_barrier_depends is insufficient to ensure that tx_buffer->skb does not
get loaded prior to tx_buffer->next_to_watch, which then results in loading
a stale skb pointer. This patch replaces the read_barrier_depends with
smp_rmb to ensure loads are ordered with respect to the load of
tx_buffer->next_to_watch.

Cc: stable <stable@vger.kernel.org>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Acked-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/ixgbe/ixgbe_main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index ca06c3cc2ca8..62a18914f00f 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -1192,7 +1192,7 @@ static bool ixgbe_clean_tx_irq(struct ixgbe_q_vector *q_vector,
 			break;
 
 		/* prevent any other reads prior to eop_desc */
-		read_barrier_depends();
+		smp_rmb();
 
 		/* if DD is not set pending work has not been completed */
 		if (!(eop_desc->wb.status & cpu_to_le32(IXGBE_TXD_STAT_DD)))
-- 
2.15.0

[net 08/13] i40e: Use smp_rmb rather than read_barrier_depends

[Jeff Kirsher]

From: Brian King <brking@linux.vnet.ibm.com>

The original issue being fixed in this patch was seen with the ixgbe
driver, but the same issue exists with i40e as well, as the code is
very similar. read_barrier_depends is not sufficient to ensure
loads following it are not speculatively loaded out of order
by the CPU, which can result in stale data being loaded, causing
potential system crashes.

Cc: stable <stable@vger.kernel.org>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Acked-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 | 2 +-
 drivers/net/ethernet/intel/i40e/i40e_txrx.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 775d5a125887..4c08cc86463e 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -3966,7 +3966,7 @@ static bool i40e_clean_fdir_tx_irq(struct i40e_ring *tx_ring, int budget)
 			break;
 
 		/* prevent any other reads prior to eop_desc */
-		read_barrier_depends();
+		smp_rmb();
 
 		/* if the descriptor isn't done, no work yet to do */
 		if (!(eop_desc->cmd_type_offset_bsz &
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index d6d352a6e6ea..4566d66ffc7c 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -759,7 +759,7 @@ static bool i40e_clean_tx_irq(struct i40e_vsi *vsi,
 			break;
 
 		/* prevent any other reads prior to eop_desc */
-		read_barrier_depends();
+		smp_rmb();
 
 		i40e_trace(clean_tx_irq, tx_ring, tx_desc, tx_buf);
 		/* we have caught up to head, no work left to do */
-- 
2.15.0

[net 09/13] ixgbevf: Use smp_rmb rather than read_barrier_depends

[Jeff Kirsher]

From: Brian King <brking@linux.vnet.ibm.com>

The original issue being fixed in this patch was seen with the ixgbe
driver, but the same issue exists with ixgbevf as well, as the code is
very similar. read_barrier_depends is not sufficient to ensure
loads following it are not speculatively loaded out of order
by the CPU, which can result in stale data being loaded, causing
potential system crashes.

Cc: stable <stable@vger.kernel.org>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Acked-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/ixgbevf/ixgbevf_main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index feed11bc9ddf..1f4a69134ade 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -326,7 +326,7 @@ static bool ixgbevf_clean_tx_irq(struct ixgbevf_q_vector *q_vector,
 			break;
 
 		/* prevent any other reads prior to eop_desc */
-		read_barrier_depends();
+		smp_rmb();
 
 		/* if DD is not set pending work has not been completed */
 		if (!(eop_desc->wb.status & cpu_to_le32(IXGBE_TXD_STAT_DD)))
-- 
2.15.0

[net 10/13] igbvf: Use smp_rmb rather than read_barrier_depends

[Jeff Kirsher]

From: Brian King <brking@linux.vnet.ibm.com>

The original issue being fixed in this patch was seen with the ixgbe
driver, but the same issue exists with igbvf as well, as the code is
very similar. read_barrier_depends is not sufficient to ensure
loads following it are not speculatively loaded out of order
by the CPU, which can result in stale data being loaded, causing
potential system crashes.

Cc: stable <stable@vger.kernel.org>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Acked-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/igbvf/netdev.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/igbvf/netdev.c b/drivers/net/ethernet/intel/igbvf/netdev.c
index 713e8df23744..4214c1519a87 100644
--- a/drivers/net/ethernet/intel/igbvf/netdev.c
+++ b/drivers/net/ethernet/intel/igbvf/netdev.c
@@ -810,7 +810,7 @@ static bool igbvf_clean_tx_irq(struct igbvf_ring *tx_ring)
 			break;
 
 		/* prevent any other reads prior to eop_desc */
-		read_barrier_depends();
+		smp_rmb();
 
 		/* if DD is not set pending work has not been completed */
 		if (!(eop_desc->wb.status & cpu_to_le32(E1000_TXD_STAT_DD)))
-- 
2.15.0

[net 11/13] igb: Use smp_rmb rather than read_barrier_depends

[Jeff Kirsher]

From: Brian King <brking@linux.vnet.ibm.com>

The original issue being fixed in this patch was seen with the ixgbe
driver, but the same issue exists with igb as well, as the code is
very similar. read_barrier_depends is not sufficient to ensure
loads following it are not speculatively loaded out of order
by the CPU, which can result in stale data being loaded, causing
potential system crashes.

Cc: stable <stable@vger.kernel.org>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Acked-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/igb/igb_main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c
index e94d3c256667..c208753ff5b7 100644
--- a/drivers/net/ethernet/intel/igb/igb_main.c
+++ b/drivers/net/ethernet/intel/igb/igb_main.c
@@ -7317,7 +7317,7 @@ static bool igb_clean_tx_irq(struct igb_q_vector *q_vector, int napi_budget)
 			break;
 
 		/* prevent any other reads prior to eop_desc */
-		read_barrier_depends();
+		smp_rmb();
 
 		/* if DD is not set pending work has not been completed */
 		if (!(eop_desc->wb.status & cpu_to_le32(E1000_TXD_STAT_DD)))
-- 
2.15.0

[net 12/13] fm10k: Use smp_rmb rather than read_barrier_depends

[Jeff Kirsher]

From: Brian King <brking@linux.vnet.ibm.com>

The original issue being fixed in this patch was seen with the ixgbe
driver, but the same issue exists with fm10k as well, as the code is
very similar. read_barrier_depends is not sufficient to ensure
loads following it are not speculatively loaded out of order
by the CPU, which can result in stale data being loaded, causing
potential system crashes.

Cc: stable <stable@vger.kernel.org>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Acked-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/fm10k/fm10k_main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_main.c b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
index dbd69310f263..538b42d5c187 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_main.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
@@ -1231,7 +1231,7 @@ static bool fm10k_clean_tx_irq(struct fm10k_q_vector *q_vector,
 			break;
 
 		/* prevent any other reads prior to eop_desc */
-		read_barrier_depends();
+		smp_rmb();
 
 		/* if DD is not set pending work has not been completed */
 		if (!(eop_desc->flags & FM10K_TXD_FLAG_DONE))
-- 
2.15.0

[net 13/13] i40evf: Use smp_rmb rather than read_barrier_depends

[Jeff Kirsher]

From: Brian King <brking@linux.vnet.ibm.com>

The original issue being fixed in this patch was seen with the ixgbe
driver, but the same issue exists with i40evf as well, as the code is
very similar. read_barrier_depends is not sufficient to ensure
loads following it are not speculatively loaded out of order
by the CPU, which can result in stale data being loaded, causing
potential system crashes.

Cc: stable <stable@vger.kernel.org>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Acked-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/i40evf/i40e_txrx.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
index fe817e2b6fef..50864f99446d 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
@@ -179,7 +179,7 @@ static bool i40e_clean_tx_irq(struct i40e_vsi *vsi,
 			break;
 
 		/* prevent any other reads prior to eop_desc */
-		read_barrier_depends();
+		smp_rmb();
 
 		i40e_trace(clean_tx_irq, tx_ring, tx_desc, tx_buf);
 		/* if the descriptor isn't done, no work yet to do */
-- 
2.15.0

Re: [net 00/13][pull request] Intel Wired LAN Driver Fixes 2017-11-21

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Wed, 22 Nov 2017 00:31:07 -0800

> This series contains fixes for igb/vf, ixgbe/vf, i40e/vf and fm10k.

Pulled, thanks a lot Jeff!