[net 0/2][pull request] Intel Wired LAN Driver Updates 2016-06-29

[net 0/2][pull request] Intel Wired LAN Driver Updates 2016-06-29

[Jeff Kirsher]

This series contains fixes to e1000e and ixgbevf.

Jarod Wilson's fix for e1000e was a follow-on patch to his previous fix to
keep the hardware VLAN CTAG's for receive and transmit in sync, which
in turn resolves the original issue, so revert a portion of the original
fix.

Xin Long noticed that the ret_val needed to be initialized to IXGBE_ERR_MBX,
instead of -IXGBE_ERR_MBX.

The following are changes since commit 751ad819b0bf443ad8963eb7bfbd533e6a463973:
  Merge tag 'mac80211-for-davem-2016-06-29-v2' of git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211
and are available in the git repository at:
  git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-queue master

Jarod Wilson (1):
  e1000e: keep Rx/Tx HW_VLAN_CTAG in sync

Xin Long (1):
  ixgbevf: ixgbevf_write/read_posted_mbx should use IXGBE_ERR_MBX to
    initialize ret_val

 drivers/net/ethernet/intel/e1000e/netdev.c | 21 +++++++++------------
 drivers/net/ethernet/intel/ixgbevf/mbx.c   |  4 ++--
 2 files changed, 11 insertions(+), 14 deletions(-)

-- 
2.5.5

[net 1/2] e1000e: keep Rx/Tx HW_VLAN_CTAG in sync

[Jeff Kirsher]

From: Jarod Wilson <jarod@redhat.com>

The bit in the e1000 driver that mentions explicitly that the hardware
has no support for separate RX/TX VLAN accel toggling rings true for
e1000e as well, and thus both NETIF_F_HW_VLAN_CTAG_RX and
NETIF_F_HW_VLAN_CTAG_TX need to be kept in sync.

Revert a portion of commit 889ad456660461 ("e1000e: keep VLAN interfaces
functional after rxvlan off") since keeping the bits in sync resolves
the original issue.

Signed-off-by: Jarod Wilson <jarod@redhat.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/e1000e/netdev.c | 21 +++++++++------------
 1 file changed, 9 insertions(+), 12 deletions(-)

diff --git a/drivers/net/ethernet/intel/e1000e/netdev.c b/drivers/net/ethernet/intel/e1000e/netdev.c
index 73f7452..2b2e2f8 100644
--- a/drivers/net/ethernet/intel/e1000e/netdev.c
+++ b/drivers/net/ethernet/intel/e1000e/netdev.c
@@ -154,16 +154,6 @@ void __ew32(struct e1000_hw *hw, unsigned long reg, u32 val)
 	writel(val, hw->hw_addr + reg);
 }
 
-static bool e1000e_vlan_used(struct e1000_adapter *adapter)
-{
-	u16 vid;
-
-	for_each_set_bit(vid, adapter->active_vlans, VLAN_N_VID)
-		return true;
-
-	return false;
-}
-
 /**
  * e1000_regdump - register printout routine
  * @hw: pointer to the HW structure
@@ -3453,8 +3443,7 @@ static void e1000e_set_rx_mode(struct net_device *netdev)
 
 	ew32(RCTL, rctl);
 
-	if (netdev->features & NETIF_F_HW_VLAN_CTAG_RX ||
-	    e1000e_vlan_used(adapter))
+	if (netdev->features & NETIF_F_HW_VLAN_CTAG_RX)
 		e1000e_vlan_strip_enable(adapter);
 	else
 		e1000e_vlan_strip_disable(adapter);
@@ -6926,6 +6915,14 @@ static netdev_features_t e1000_fix_features(struct net_device *netdev,
 	if ((hw->mac.type >= e1000_pch2lan) && (netdev->mtu > ETH_DATA_LEN))
 		features &= ~NETIF_F_RXFCS;
 
+	/* Since there is no support for separate Rx/Tx vlan accel
+	 * enable/disable make sure Tx flag is always in same state as Rx.
+	 */
+	if (features & NETIF_F_HW_VLAN_CTAG_RX)
+		features |= NETIF_F_HW_VLAN_CTAG_TX;
+	else
+		features &= ~NETIF_F_HW_VLAN_CTAG_TX;
+
 	return features;
 }
 
-- 
2.5.5

[net 2/2] ixgbevf: ixgbevf_write/read_posted_mbx should use IXGBE_ERR_MBX to initialize ret_val

[Jeff Kirsher]

From: Xin Long <lucien.xin@gmail.com>

Now ixgbevf_write/read_posted_mbx use -IXGBE_ERR_MBX as the initiative
return value, but it's incorrect, cause in ixgbevf_vlan_rx_add_vid(),
it use err == IXGBE_ERR_MBX, the err returned from mac.ops.set_vfta,
and in ixgbevf_set_vfta_vf, it return from write/read_posted. so we
should initialize err with IXGBE_ERR_MBX, instead of -IXGBE_ERR_MBX.

With this fix, the other functions that called it also can work well,
cause they only care about if err is 0 or not.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbevf/mbx.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbevf/mbx.c b/drivers/net/ethernet/intel/ixgbevf/mbx.c
index 61a80da..2819abc 100644
--- a/drivers/net/ethernet/intel/ixgbevf/mbx.c
+++ b/drivers/net/ethernet/intel/ixgbevf/mbx.c
@@ -85,7 +85,7 @@ static s32 ixgbevf_poll_for_ack(struct ixgbe_hw *hw)
 static s32 ixgbevf_read_posted_mbx(struct ixgbe_hw *hw, u32 *msg, u16 size)
 {
 	struct ixgbe_mbx_info *mbx = &hw->mbx;
-	s32 ret_val = -IXGBE_ERR_MBX;
+	s32 ret_val = IXGBE_ERR_MBX;
 
 	if (!mbx->ops.read)
 		goto out;
@@ -111,7 +111,7 @@ out:
 static s32 ixgbevf_write_posted_mbx(struct ixgbe_hw *hw, u32 *msg, u16 size)
 {
 	struct ixgbe_mbx_info *mbx = &hw->mbx;
-	s32 ret_val = -IXGBE_ERR_MBX;
+	s32 ret_val = IXGBE_ERR_MBX;
 
 	/* exit if either we can't write or there isn't a defined timeout */
 	if (!mbx->ops.write || !mbx->timeout)
-- 
2.5.5

Re: [net 2/2] ixgbevf: ixgbevf_write/read_posted_mbx should use IXGBE_ERR_MBX to initialize ret_val

[Greg]

On Wed, 2016-06-29 at 09:30 -0700, Jeff Kirsher wrote:
> From: Xin Long <lucien.xin@gmail.com>
> 
> Now ixgbevf_write/read_posted_mbx use -IXGBE_ERR_MBX as the initiative
> return value, but it's incorrect, cause in ixgbevf_vlan_rx_add_vid(),
> it use err == IXGBE_ERR_MBX, the err returned from mac.ops.set_vfta,
> and in ixgbevf_set_vfta_vf, it return from write/read_posted. so we
> should initialize err with IXGBE_ERR_MBX, instead of -IXGBE_ERR_MBX.

I believe "initiative" should be "initial", "cause" should be "because"
and "it use err" should be "it uses err".  And I won't be a grammar nazi
about all the commas.

- Greg

> 
> With this fix, the other functions that called it also can work well,
> cause they only care about if err is 0 or not.
> 
> Signed-off-by: Xin Long <lucien.xin@gmail.com>
> Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
> ---
>  drivers/net/ethernet/intel/ixgbevf/mbx.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/ethernet/intel/ixgbevf/mbx.c b/drivers/net/ethernet/intel/ixgbevf/mbx.c
> index 61a80da..2819abc 100644
> --- a/drivers/net/ethernet/intel/ixgbevf/mbx.c
> +++ b/drivers/net/ethernet/intel/ixgbevf/mbx.c
> @@ -85,7 +85,7 @@ static s32 ixgbevf_poll_for_ack(struct ixgbe_hw *hw)
>  static s32 ixgbevf_read_posted_mbx(struct ixgbe_hw *hw, u32 *msg, u16 size)
>  {
>  	struct ixgbe_mbx_info *mbx = &hw->mbx;
> -	s32 ret_val = -IXGBE_ERR_MBX;
> +	s32 ret_val = IXGBE_ERR_MBX;
>  
>  	if (!mbx->ops.read)
>  		goto out;
> @@ -111,7 +111,7 @@ out:
>  static s32 ixgbevf_write_posted_mbx(struct ixgbe_hw *hw, u32 *msg, u16 size)
>  {
>  	struct ixgbe_mbx_info *mbx = &hw->mbx;
> -	s32 ret_val = -IXGBE_ERR_MBX;
> +	s32 ret_val = IXGBE_ERR_MBX;
>  
>  	/* exit if either we can't write or there isn't a defined timeout */
>  	if (!mbx->ops.write || !mbx->timeout)

Re: [net 0/2][pull request] Intel Wired LAN Driver Updates 2016-06-29

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Wed, 29 Jun 2016 09:30:02 -0700

> This series contains fixes to e1000e and ixgbevf.

Pulled, thanks Jeff.