[Intel-wired-lan] [PATCH 0/5] ixgbe: fixes for configuring link on x550a

[Intel-wired-lan] [PATCH 0/5] ixgbe: fixes for configuring link on x550a

[Emil Tantilov]

The following series introduces small fixes needed to address issues
with setting up link on SFP+ based x550a parts.

---

Emil Tantilov (5):
      ixgbe: correct CS4223/7 PHY identification
      ixgbe: add write flush when configuring CS4223/7
      ixgbe: always call setup_mac_link for multispeed fiber
      ixgbe: add missing configuration for rate select 1
      ixgbe: fix incorrect status check


 drivers/net/ethernet/intel/ixgbe/ixgbe_common.c |   37 ++++++++++++-----------
 drivers/net/ethernet/intel/ixgbe/ixgbe_phy.h    |    5 ++-
 drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c   |   30 ++++++++++++++-----
 3 files changed, 45 insertions(+), 27 deletions(-)

--

[Intel-wired-lan] [PATCH 1/5] ixgbe: correct CS4223/7 PHY identification

[Emil Tantilov]

Previous method was unreliable. Use a different register to
diferentiate between the SKUs.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_phy.h  |    5 +++--
 drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c |    8 ++++----
 2 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_phy.h b/drivers/net/ethernet/intel/ixgbe/ixgbe_phy.h
index 5aa2c3c..b0cac96 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_phy.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_phy.h
@@ -84,8 +84,9 @@
 #define IXGBE_CS4227_GLOBAL_ID_LSB		0
 #define IXGBE_CS4227_GLOBAL_ID_MSB		1
 #define IXGBE_CS4227_SCRATCH			2
-#define IXGBE_CS4223_PHY_ID			0x7003	/* Quad port */
-#define IXGBE_CS4227_PHY_ID			0x3003	/* Dual port */
+#define IXGBE_CS4227_EFUSE_PDF_SKU		0x19F
+#define IXGBE_CS4223_SKU_ID			0x0010  /* Quad port */
+#define IXGBE_CS4227_SKU_ID			0x0014  /* Dual port */
 #define IXGBE_CS4227_RESET_PENDING		0x1357
 #define IXGBE_CS4227_RESET_COMPLETE		0x5AA5
 #define IXGBE_CS4227_RETRIES			15
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c
index b8d9097..870f9e1 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c
@@ -1807,16 +1807,16 @@ static s32 ixgbe_setup_sfi_x550a(struct ixgbe_hw *hw, ixgbe_link_speed *speed)
 	if (hw->phy.mdio.prtad == MDIO_PRTAD_NONE)
 		return IXGBE_ERR_PHY_ADDR_INVALID;
 
-	/* Get external PHY device id */
-	ret_val = hw->phy.ops.read_reg(hw, IXGBE_CS4227_GLOBAL_ID_MSB,
-				  IXGBE_MDIO_ZERO_DEV_TYPE, &reg_phy_ext);
+	/* Get external PHY SKU id */
+	ret_val = hw->phy.ops.read_reg(hw, IXGBE_CS4227_EFUSE_PDF_SKU,
+				       IXGBE_MDIO_ZERO_DEV_TYPE, &reg_phy_ext);
 	if (ret_val)
 		return ret_val;
 
 	/* When configuring quad port CS4223, the MAC instance is part
 	 * of the slice offset.
 	 */
-	if (reg_phy_ext == IXGBE_CS4223_PHY_ID)
+	if (reg_phy_ext == IXGBE_CS4223_SKU_ID)
 		slice_offset = (hw->bus.lan_id +
 				(hw->bus.instance_id << 1)) << 12;
 	else

[Intel-wired-lan] [PATCH 2/5] ixgbe: add write flush when configuring CS4223/7

[Emil Tantilov]

Make sure the writes are processed immediately. Without the flush it
is possible for operations on one port to spill over the other as the
resource is shared.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c |   20 ++++++++++++++++++--
 1 file changed, 18 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c
index 870f9e1..cb5d363 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c
@@ -1824,12 +1824,28 @@ static s32 ixgbe_setup_sfi_x550a(struct ixgbe_hw *hw, ixgbe_link_speed *speed)
 
 	/* Configure CS4227/CS4223 LINE side to proper mode. */
 	reg_slice = IXGBE_CS4227_LINE_SPARE24_LSB + slice_offset;
+
+	ret_val = hw->phy.ops.read_reg(hw, reg_slice,
+				       IXGBE_MDIO_ZERO_DEV_TYPE, &reg_phy_ext);
+	if (ret_val)
+		return ret_val;
+
+	reg_phy_ext &= ~((IXGBE_CS4227_EDC_MODE_CX1 << 1) |
+			 (IXGBE_CS4227_EDC_MODE_SR << 1));
+
 	if (setup_linear)
 		reg_phy_ext = (IXGBE_CS4227_EDC_MODE_CX1 << 1) | 1;
 	else
 		reg_phy_ext = (IXGBE_CS4227_EDC_MODE_SR << 1) | 1;
-	return hw->phy.ops.write_reg(hw, reg_slice, IXGBE_MDIO_ZERO_DEV_TYPE,
-				     reg_phy_ext);
+
+	ret_val = hw->phy.ops.write_reg(hw, reg_slice,
+					IXGBE_MDIO_ZERO_DEV_TYPE, reg_phy_ext);
+	if (ret_val)
+		return ret_val;
+
+	/* Flush previous write with a read */
+	return hw->phy.ops.read_reg(hw, reg_slice,
+				    IXGBE_MDIO_ZERO_DEV_TYPE, &reg_phy_ext);
 }
 
 /**

[Intel-wired-lan] [PATCH 3/5] ixgbe: always call setup_mac_link for multispeed fiber

[Emil Tantilov]

Remove the logic which would previously skip the link configuration
in the case where we are already at the requested speed in
ixgbe_setup_mac_link_multispeed_fiber().

By exiting early we are skipping the link configuration and as such
the driver may not always configure the PHY correctly for SFP+.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_common.c |   18 ------------------
 1 file changed, 18 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_common.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_common.c
index 3af6127..8a2e8bf8 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_common.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_common.c
@@ -4121,15 +4121,6 @@ s32 ixgbe_setup_mac_link_multispeed_fiber(struct ixgbe_hw *hw,
 		speedcnt++;
 		highest_link_speed = IXGBE_LINK_SPEED_10GB_FULL;
 
-		/* If we already have link at this speed, just jump out */
-		status = hw->mac.ops.check_link(hw, &link_speed, &link_up,
-						false);
-		if (status)
-			return status;
-
-		if (link_speed == IXGBE_LINK_SPEED_10GB_FULL && link_up)
-			goto out;
-
 		/* Set the module link speed */
 		switch (hw->phy.media_type) {
 		case ixgbe_media_type_fiber:
@@ -4181,15 +4172,6 @@ s32 ixgbe_setup_mac_link_multispeed_fiber(struct ixgbe_hw *hw,
 		if (highest_link_speed == IXGBE_LINK_SPEED_UNKNOWN)
 			highest_link_speed = IXGBE_LINK_SPEED_1GB_FULL;
 
-		/* If we already have link at this speed, just jump out */
-		status = hw->mac.ops.check_link(hw, &link_speed, &link_up,
-						false);
-		if (status)
-			return status;
-
-		if (link_speed == IXGBE_LINK_SPEED_1GB_FULL && link_up)
-			goto out;
-
 		/* Set the module link speed */
 		switch (hw->phy.media_type) {
 		case ixgbe_media_type_fiber:

[Intel-wired-lan] [PATCH 4/5] ixgbe: add missing configuration for rate select 1

[Emil Tantilov]

Add RS1 configuration to ixgbe_set_soft_rate_select_speed()

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_common.c |   19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_common.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_common.c
index 8a2e8bf8..9b8e594 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_common.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_common.c
@@ -4278,4 +4278,23 @@ void ixgbe_set_soft_rate_select_speed(struct ixgbe_hw *hw,
 		hw_dbg(hw, "Failed to write Rx Rate Select RS0\n");
 		return;
 	}
+
+	/* Set RS1 */
+	status = hw->phy.ops.read_i2c_byte(hw, IXGBE_SFF_SFF_8472_ESCB,
+					   IXGBE_I2C_EEPROM_DEV_ADDR2,
+					   &eeprom_data);
+	if (status) {
+		hw_dbg(hw, "Failed to read Rx Rate Select RS1\n");
+		return;
+	}
+
+	eeprom_data = (eeprom_data & ~IXGBE_SFF_SOFT_RS_SELECT_MASK) | rs;
+
+	status = hw->phy.ops.write_i2c_byte(hw, IXGBE_SFF_SFF_8472_ESCB,
+					    IXGBE_I2C_EEPROM_DEV_ADDR2,
+					    eeprom_data);
+	if (status) {
+		hw_dbg(hw, "Failed to write Rx Rate Select RS1\n");
+		return;
+	}
 }

[Intel-wired-lan] [PATCH 5/5] ixgbe: fix incorrect status check

[Emil Tantilov]

Check for ret_val instead of !ret_val to allow the rest of
the code to execute and configure the speed properly.

Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c
index cb5d363..47c8f24 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c
@@ -1798,7 +1798,7 @@ static s32 ixgbe_setup_sfi_x550a(struct ixgbe_hw *hw, ixgbe_link_speed *speed)
 	if (ret_val == IXGBE_ERR_SFP_NOT_PRESENT)
 		return 0;
 
-	if (!ret_val)
+	if (ret_val)
 		return ret_val;
 
 	/* Configure internal PHY for KR/KX. */

[Intel-wired-lan] [PATCH 3/5] ixgbe: always call setup_mac_link for multispeed fiber

[Bowers, AndrewX]

> -----Original Message-----
> From: Intel-wired-lan [mailto:intel-wired-lan-bounces at osuosl.org] On
> Behalf Of Emil Tantilov
> Sent: Wednesday, May 17, 2017 3:18 PM
> To: intel-wired-lan at lists.osuosl.org
> Subject: [Intel-wired-lan] [PATCH 3/5] ixgbe: always call setup_mac_link for
> multispeed fiber
> 
> Remove the logic which would previously skip the link configuration in the
> case where we are already at the requested speed in
> ixgbe_setup_mac_link_multispeed_fiber().
> 
> By exiting early we are skipping the link configuration and as such the driver
> may not always configure the PHY correctly for SFP+.
> 
> Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
> ---
>  drivers/net/ethernet/intel/ixgbe/ixgbe_common.c |   18 ------------------
>  1 file changed, 18 deletions(-)

Tested-by: Andrew Bowers <andrewx.bowers@intel.com>

[Intel-wired-lan] [PATCH 4/5] ixgbe: add missing configuration for rate select 1

[Bowers, AndrewX]

> -----Original Message-----
> From: Intel-wired-lan [mailto:intel-wired-lan-bounces at osuosl.org] On
> Behalf Of Emil Tantilov
> Sent: Wednesday, May 17, 2017 3:18 PM
> To: intel-wired-lan at lists.osuosl.org
> Subject: [Intel-wired-lan] [PATCH 4/5] ixgbe: add missing configuration for
> rate select 1
> 
> Add RS1 configuration to ixgbe_set_soft_rate_select_speed()
> 
> Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
> ---
>  drivers/net/ethernet/intel/ixgbe/ixgbe_common.c |   19
> +++++++++++++++++++
>  1 file changed, 19 insertions(+)

Tested-by: Andrew Bowers <andrewx.bowers@intel.com>

[Intel-wired-lan] [PATCH 5/5] ixgbe: fix incorrect status check

[Bowers, AndrewX]

> -----Original Message-----
> From: Intel-wired-lan [mailto:intel-wired-lan-bounces at osuosl.org] On
> Behalf Of Emil Tantilov
> Sent: Wednesday, May 17, 2017 3:18 PM
> To: intel-wired-lan at lists.osuosl.org
> Subject: [Intel-wired-lan] [PATCH 5/5] ixgbe: fix incorrect status check
> 
> Check for ret_val instead of !ret_val to allow the rest of the code to execute
> and configure the speed properly.
> 
> Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
> ---
>  drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)

Tested-by: Andrew Bowers <andrewx.bowers@intel.com>

[Intel-wired-lan] [PATCH 1/5] ixgbe: correct CS4223/7 PHY identification

[Bowers, AndrewX]

> -----Original Message-----
> From: Intel-wired-lan [mailto:intel-wired-lan-bounces at osuosl.org] On
> Behalf Of Emil Tantilov
> Sent: Wednesday, May 17, 2017 3:18 PM
> To: intel-wired-lan at lists.osuosl.org
> Subject: [Intel-wired-lan] [PATCH 1/5] ixgbe: correct CS4223/7 PHY
> identification
> 
> Previous method was unreliable. Use a different register to diferentiate
> between the SKUs.
> 
> Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
> ---
>  drivers/net/ethernet/intel/ixgbe/ixgbe_phy.h  |    5 +++--
>  drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c |    8 ++++----
>  2 files changed, 7 insertions(+), 6 deletions(-)

Tested-by: Andrew Bowers <andrewx.bowers@intel.com>

[Intel-wired-lan] [PATCH 2/5] ixgbe: add write flush when configuring CS4223/7

[Bowers, AndrewX]

> -----Original Message-----
> From: Intel-wired-lan [mailto:intel-wired-lan-bounces at osuosl.org] On
> Behalf Of Emil Tantilov
> Sent: Wednesday, May 17, 2017 3:18 PM
> To: intel-wired-lan at lists.osuosl.org
> Subject: [Intel-wired-lan] [PATCH 2/5] ixgbe: add write flush when
> configuring CS4223/7
> 
> Make sure the writes are processed immediately. Without the flush it is
> possible for operations on one port to spill over the other as the resource is
> shared.
> 
> Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
> ---
>  drivers/net/ethernet/intel/ixgbe/ixgbe_x550.c |   20
> ++++++++++++++++++--
>  1 file changed, 18 insertions(+), 2 deletions(-)

Tested-by: Andrew Bowers <andrewx.bowers@intel.com>