2014-06-05 21:15:12 +07:00
|
|
|
/*
|
|
|
|
* AMD 10Gb Ethernet PHY driver
|
|
|
|
*
|
|
|
|
* This file is available to you under your choice of the following two
|
|
|
|
* licenses:
|
|
|
|
*
|
|
|
|
* License 1: GPLv2
|
|
|
|
*
|
|
|
|
* Copyright (c) 2014 Advanced Micro Devices, Inc.
|
|
|
|
*
|
|
|
|
* This file is free software; you may copy, redistribute and/or modify
|
|
|
|
* it under the terms of the GNU General Public License as published by
|
|
|
|
* the Free Software Foundation, either version 2 of the License, or (at
|
|
|
|
* your option) any later version.
|
|
|
|
*
|
|
|
|
* This file is distributed in the hope that it will be useful, but
|
|
|
|
* WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
|
|
* General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU General Public License
|
|
|
|
* along with this program. If not, see <http://www.gnu.org/licenses/>.
|
|
|
|
*
|
|
|
|
*
|
|
|
|
* License 2: Modified BSD
|
|
|
|
*
|
|
|
|
* Copyright (c) 2014 Advanced Micro Devices, Inc.
|
|
|
|
* All rights reserved.
|
|
|
|
*
|
|
|
|
* Redistribution and use in source and binary forms, with or without
|
|
|
|
* modification, are permitted provided that the following conditions are met:
|
|
|
|
* * Redistributions of source code must retain the above copyright
|
|
|
|
* notice, this list of conditions and the following disclaimer.
|
|
|
|
* * Redistributions in binary form must reproduce the above copyright
|
|
|
|
* notice, this list of conditions and the following disclaimer in the
|
|
|
|
* documentation and/or other materials provided with the distribution.
|
|
|
|
* * Neither the name of Advanced Micro Devices, Inc. nor the
|
|
|
|
* names of its contributors may be used to endorse or promote products
|
|
|
|
* derived from this software without specific prior written permission.
|
|
|
|
*
|
|
|
|
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
|
|
|
|
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
|
|
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
|
|
* ARE DISCLAIMED. IN NO EVENT SHALL <COPYRIGHT HOLDER> BE LIABLE FOR ANY
|
|
|
|
* DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
|
|
|
|
* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
|
|
|
|
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
|
|
|
|
* ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
|
|
|
|
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
|
|
|
|
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/kernel.h>
|
|
|
|
#include <linux/device.h>
|
|
|
|
#include <linux/platform_device.h>
|
|
|
|
#include <linux/string.h>
|
|
|
|
#include <linux/errno.h>
|
|
|
|
#include <linux/unistd.h>
|
|
|
|
#include <linux/slab.h>
|
|
|
|
#include <linux/interrupt.h>
|
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/delay.h>
|
|
|
|
#include <linux/netdevice.h>
|
|
|
|
#include <linux/etherdevice.h>
|
|
|
|
#include <linux/skbuff.h>
|
|
|
|
#include <linux/mm.h>
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/mii.h>
|
|
|
|
#include <linux/ethtool.h>
|
|
|
|
#include <linux/phy.h>
|
|
|
|
#include <linux/mdio.h>
|
|
|
|
#include <linux/io.h>
|
|
|
|
#include <linux/of.h>
|
|
|
|
#include <linux/of_platform.h>
|
|
|
|
#include <linux/of_device.h>
|
|
|
|
#include <linux/uaccess.h>
|
|
|
|
#include <asm/irq.h>
|
|
|
|
|
|
|
|
|
|
|
|
MODULE_AUTHOR("Tom Lendacky <thomas.lendacky@amd.com>");
|
|
|
|
MODULE_LICENSE("Dual BSD/GPL");
|
|
|
|
MODULE_VERSION("1.0.0-a");
|
|
|
|
MODULE_DESCRIPTION("AMD 10GbE (amd-xgbe) PHY driver");
|
|
|
|
|
|
|
|
#define XGBE_PHY_ID 0x000162d0
|
|
|
|
#define XGBE_PHY_MASK 0xfffffff0
|
|
|
|
|
|
|
|
#define XGBE_AN_INT_CMPLT 0x01
|
|
|
|
#define XGBE_AN_INC_LINK 0x02
|
|
|
|
#define XGBE_AN_PG_RCV 0x04
|
|
|
|
|
|
|
|
#define XNP_MCF_NULL_MESSAGE 0x001
|
|
|
|
#define XNP_ACK_PROCESSED (1 << 12)
|
|
|
|
#define XNP_MP_FORMATTED (1 << 13)
|
|
|
|
#define XNP_NP_EXCHANGE (1 << 15)
|
|
|
|
|
|
|
|
#ifndef MDIO_PMA_10GBR_PMD_CTRL
|
|
|
|
#define MDIO_PMA_10GBR_PMD_CTRL 0x0096
|
|
|
|
#endif
|
|
|
|
#ifndef MDIO_PMA_10GBR_FEC_CTRL
|
|
|
|
#define MDIO_PMA_10GBR_FEC_CTRL 0x00ab
|
|
|
|
#endif
|
|
|
|
#ifndef MDIO_AN_XNP
|
|
|
|
#define MDIO_AN_XNP 0x0016
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#ifndef MDIO_AN_INTMASK
|
|
|
|
#define MDIO_AN_INTMASK 0x8001
|
|
|
|
#endif
|
|
|
|
#ifndef MDIO_AN_INT
|
|
|
|
#define MDIO_AN_INT 0x8002
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#ifndef MDIO_CTRL1_SPEED1G
|
|
|
|
#define MDIO_CTRL1_SPEED1G (MDIO_CTRL1_SPEED10G & ~BMCR_SPEED100)
|
|
|
|
#endif
|
|
|
|
|
|
|
|
/* SerDes integration register offsets */
|
|
|
|
#define SIR0_STATUS 0x0040
|
|
|
|
#define SIR1_SPEED 0x0000
|
|
|
|
|
|
|
|
/* SerDes integration register entry bit positions and sizes */
|
|
|
|
#define SIR0_STATUS_RX_READY_INDEX 0
|
|
|
|
#define SIR0_STATUS_RX_READY_WIDTH 1
|
|
|
|
#define SIR0_STATUS_TX_READY_INDEX 8
|
|
|
|
#define SIR0_STATUS_TX_READY_WIDTH 1
|
|
|
|
#define SIR1_SPEED_DATARATE_INDEX 4
|
|
|
|
#define SIR1_SPEED_DATARATE_WIDTH 2
|
|
|
|
#define SIR1_SPEED_PI_SPD_SEL_INDEX 12
|
|
|
|
#define SIR1_SPEED_PI_SPD_SEL_WIDTH 4
|
|
|
|
#define SIR1_SPEED_PLLSEL_INDEX 3
|
|
|
|
#define SIR1_SPEED_PLLSEL_WIDTH 1
|
|
|
|
#define SIR1_SPEED_RATECHANGE_INDEX 6
|
|
|
|
#define SIR1_SPEED_RATECHANGE_WIDTH 1
|
|
|
|
#define SIR1_SPEED_TXAMP_INDEX 8
|
|
|
|
#define SIR1_SPEED_TXAMP_WIDTH 4
|
|
|
|
#define SIR1_SPEED_WORDMODE_INDEX 0
|
|
|
|
#define SIR1_SPEED_WORDMODE_WIDTH 3
|
|
|
|
|
|
|
|
#define SPEED_10000_CDR 0x7
|
|
|
|
#define SPEED_10000_PLL 0x1
|
|
|
|
#define SPEED_10000_RATE 0x0
|
|
|
|
#define SPEED_10000_TXAMP 0xa
|
|
|
|
#define SPEED_10000_WORD 0x7
|
|
|
|
|
|
|
|
#define SPEED_2500_CDR 0x2
|
|
|
|
#define SPEED_2500_PLL 0x0
|
|
|
|
#define SPEED_2500_RATE 0x2
|
|
|
|
#define SPEED_2500_TXAMP 0xf
|
|
|
|
#define SPEED_2500_WORD 0x1
|
|
|
|
|
|
|
|
#define SPEED_1000_CDR 0x2
|
|
|
|
#define SPEED_1000_PLL 0x0
|
|
|
|
#define SPEED_1000_RATE 0x3
|
|
|
|
#define SPEED_1000_TXAMP 0xf
|
|
|
|
#define SPEED_1000_WORD 0x1
|
|
|
|
|
|
|
|
|
|
|
|
/* SerDes RxTx register offsets */
|
|
|
|
#define RXTX_REG20 0x0050
|
|
|
|
#define RXTX_REG114 0x01c8
|
|
|
|
|
|
|
|
/* SerDes RxTx register entry bit positions and sizes */
|
|
|
|
#define RXTX_REG20_BLWC_ENA_INDEX 2
|
|
|
|
#define RXTX_REG20_BLWC_ENA_WIDTH 1
|
|
|
|
#define RXTX_REG114_PQ_REG_INDEX 9
|
|
|
|
#define RXTX_REG114_PQ_REG_WIDTH 7
|
|
|
|
|
|
|
|
#define RXTX_10000_BLWC 0
|
|
|
|
#define RXTX_10000_PQ 0x1e
|
|
|
|
|
|
|
|
#define RXTX_2500_BLWC 1
|
|
|
|
#define RXTX_2500_PQ 0xa
|
|
|
|
|
|
|
|
#define RXTX_1000_BLWC 1
|
|
|
|
#define RXTX_1000_PQ 0xa
|
|
|
|
|
|
|
|
/* Bit setting and getting macros
|
|
|
|
* The get macro will extract the current bit field value from within
|
|
|
|
* the variable
|
|
|
|
*
|
|
|
|
* The set macro will clear the current bit field value within the
|
|
|
|
* variable and then set the bit field of the variable to the
|
|
|
|
* specified value
|
|
|
|
*/
|
|
|
|
#define GET_BITS(_var, _index, _width) \
|
|
|
|
(((_var) >> (_index)) & ((0x1 << (_width)) - 1))
|
|
|
|
|
|
|
|
#define SET_BITS(_var, _index, _width, _val) \
|
|
|
|
do { \
|
|
|
|
(_var) &= ~(((0x1 << (_width)) - 1) << (_index)); \
|
|
|
|
(_var) |= (((_val) & ((0x1 << (_width)) - 1)) << (_index)); \
|
|
|
|
} while (0)
|
|
|
|
|
|
|
|
/* Macros for reading or writing SerDes integration registers
|
|
|
|
* The ioread macros will get bit fields or full values using the
|
|
|
|
* register definitions formed using the input names
|
|
|
|
*
|
|
|
|
* The iowrite macros will set bit fields or full values using the
|
|
|
|
* register definitions formed using the input names
|
|
|
|
*/
|
|
|
|
#define XSIR0_IOREAD(_priv, _reg) \
|
|
|
|
ioread16((_priv)->sir0_regs + _reg)
|
|
|
|
|
|
|
|
#define XSIR0_IOREAD_BITS(_priv, _reg, _field) \
|
|
|
|
GET_BITS(XSIR0_IOREAD((_priv), _reg), \
|
|
|
|
_reg##_##_field##_INDEX, \
|
|
|
|
_reg##_##_field##_WIDTH)
|
|
|
|
|
|
|
|
#define XSIR0_IOWRITE(_priv, _reg, _val) \
|
|
|
|
iowrite16((_val), (_priv)->sir0_regs + _reg)
|
|
|
|
|
|
|
|
#define XSIR0_IOWRITE_BITS(_priv, _reg, _field, _val) \
|
|
|
|
do { \
|
|
|
|
u16 reg_val = XSIR0_IOREAD((_priv), _reg); \
|
|
|
|
SET_BITS(reg_val, \
|
|
|
|
_reg##_##_field##_INDEX, \
|
|
|
|
_reg##_##_field##_WIDTH, (_val)); \
|
|
|
|
XSIR0_IOWRITE((_priv), _reg, reg_val); \
|
|
|
|
} while (0)
|
|
|
|
|
|
|
|
#define XSIR1_IOREAD(_priv, _reg) \
|
|
|
|
ioread16((_priv)->sir1_regs + _reg)
|
|
|
|
|
|
|
|
#define XSIR1_IOREAD_BITS(_priv, _reg, _field) \
|
|
|
|
GET_BITS(XSIR1_IOREAD((_priv), _reg), \
|
|
|
|
_reg##_##_field##_INDEX, \
|
|
|
|
_reg##_##_field##_WIDTH)
|
|
|
|
|
|
|
|
#define XSIR1_IOWRITE(_priv, _reg, _val) \
|
|
|
|
iowrite16((_val), (_priv)->sir1_regs + _reg)
|
|
|
|
|
|
|
|
#define XSIR1_IOWRITE_BITS(_priv, _reg, _field, _val) \
|
|
|
|
do { \
|
|
|
|
u16 reg_val = XSIR1_IOREAD((_priv), _reg); \
|
|
|
|
SET_BITS(reg_val, \
|
|
|
|
_reg##_##_field##_INDEX, \
|
|
|
|
_reg##_##_field##_WIDTH, (_val)); \
|
|
|
|
XSIR1_IOWRITE((_priv), _reg, reg_val); \
|
|
|
|
} while (0)
|
|
|
|
|
|
|
|
|
|
|
|
/* Macros for reading or writing SerDes RxTx registers
|
|
|
|
* The ioread macros will get bit fields or full values using the
|
|
|
|
* register definitions formed using the input names
|
|
|
|
*
|
|
|
|
* The iowrite macros will set bit fields or full values using the
|
|
|
|
* register definitions formed using the input names
|
|
|
|
*/
|
|
|
|
#define XRXTX_IOREAD(_priv, _reg) \
|
|
|
|
ioread16((_priv)->rxtx_regs + _reg)
|
|
|
|
|
|
|
|
#define XRXTX_IOREAD_BITS(_priv, _reg, _field) \
|
|
|
|
GET_BITS(XRXTX_IOREAD((_priv), _reg), \
|
|
|
|
_reg##_##_field##_INDEX, \
|
|
|
|
_reg##_##_field##_WIDTH)
|
|
|
|
|
|
|
|
#define XRXTX_IOWRITE(_priv, _reg, _val) \
|
|
|
|
iowrite16((_val), (_priv)->rxtx_regs + _reg)
|
|
|
|
|
|
|
|
#define XRXTX_IOWRITE_BITS(_priv, _reg, _field, _val) \
|
|
|
|
do { \
|
|
|
|
u16 reg_val = XRXTX_IOREAD((_priv), _reg); \
|
|
|
|
SET_BITS(reg_val, \
|
|
|
|
_reg##_##_field##_INDEX, \
|
|
|
|
_reg##_##_field##_WIDTH, (_val)); \
|
|
|
|
XRXTX_IOWRITE((_priv), _reg, reg_val); \
|
|
|
|
} while (0)
|
|
|
|
|
|
|
|
|
|
|
|
enum amd_xgbe_phy_an {
|
|
|
|
AMD_XGBE_AN_READY = 0,
|
|
|
|
AMD_XGBE_AN_START,
|
|
|
|
AMD_XGBE_AN_EVENT,
|
|
|
|
AMD_XGBE_AN_PAGE_RECEIVED,
|
|
|
|
AMD_XGBE_AN_INCOMPAT_LINK,
|
|
|
|
AMD_XGBE_AN_COMPLETE,
|
|
|
|
AMD_XGBE_AN_NO_LINK,
|
|
|
|
AMD_XGBE_AN_EXIT,
|
|
|
|
AMD_XGBE_AN_ERROR,
|
|
|
|
};
|
|
|
|
|
|
|
|
enum amd_xgbe_phy_rx {
|
|
|
|
AMD_XGBE_RX_READY = 0,
|
|
|
|
AMD_XGBE_RX_BPA,
|
|
|
|
AMD_XGBE_RX_XNP,
|
|
|
|
AMD_XGBE_RX_COMPLETE,
|
|
|
|
};
|
|
|
|
|
|
|
|
enum amd_xgbe_phy_mode {
|
|
|
|
AMD_XGBE_MODE_KR,
|
|
|
|
AMD_XGBE_MODE_KX,
|
|
|
|
};
|
|
|
|
|
|
|
|
struct amd_xgbe_phy_priv {
|
|
|
|
struct platform_device *pdev;
|
|
|
|
struct device *dev;
|
|
|
|
|
|
|
|
struct phy_device *phydev;
|
|
|
|
|
|
|
|
/* SerDes related mmio resources */
|
|
|
|
struct resource *rxtx_res;
|
|
|
|
struct resource *sir0_res;
|
|
|
|
struct resource *sir1_res;
|
|
|
|
|
|
|
|
/* SerDes related mmio registers */
|
|
|
|
void __iomem *rxtx_regs; /* SerDes Rx/Tx CSRs */
|
|
|
|
void __iomem *sir0_regs; /* SerDes integration registers (1/2) */
|
|
|
|
void __iomem *sir1_regs; /* SerDes integration registers (2/2) */
|
|
|
|
|
|
|
|
/* Maintain link status for re-starting auto-negotiation */
|
|
|
|
unsigned int link;
|
|
|
|
enum amd_xgbe_phy_mode mode;
|
|
|
|
|
|
|
|
/* Auto-negotiation state machine support */
|
|
|
|
struct mutex an_mutex;
|
|
|
|
enum amd_xgbe_phy_an an_result;
|
|
|
|
enum amd_xgbe_phy_an an_state;
|
|
|
|
enum amd_xgbe_phy_rx kr_state;
|
|
|
|
enum amd_xgbe_phy_rx kx_state;
|
|
|
|
struct work_struct an_work;
|
|
|
|
struct workqueue_struct *an_workqueue;
|
|
|
|
};
|
|
|
|
|
|
|
|
static int amd_xgbe_an_enable_kr_training(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PMAPMD, MDIO_PMA_10GBR_PMD_CTRL);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
ret |= 0x02;
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_PMAPMD, MDIO_PMA_10GBR_PMD_CTRL, ret);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_an_disable_kr_training(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PMAPMD, MDIO_PMA_10GBR_PMD_CTRL);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
ret &= ~0x02;
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_PMAPMD, MDIO_PMA_10GBR_PMD_CTRL, ret);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_phy_pcs_power_cycle(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL1);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
ret |= MDIO_CTRL1_LPOWER;
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL1, ret);
|
|
|
|
|
|
|
|
usleep_range(75, 100);
|
|
|
|
|
|
|
|
ret &= ~MDIO_CTRL1_LPOWER;
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL1, ret);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void amd_xgbe_phy_serdes_start_ratechange(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
struct amd_xgbe_phy_priv *priv = phydev->priv;
|
|
|
|
|
|
|
|
/* Assert Rx and Tx ratechange */
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, RATECHANGE, 1);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void amd_xgbe_phy_serdes_complete_ratechange(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
struct amd_xgbe_phy_priv *priv = phydev->priv;
|
|
|
|
|
|
|
|
/* Release Rx and Tx ratechange */
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, RATECHANGE, 0);
|
|
|
|
|
|
|
|
/* Wait for Rx and Tx ready */
|
|
|
|
while (!XSIR0_IOREAD_BITS(priv, SIR0_STATUS, RX_READY) &&
|
|
|
|
!XSIR0_IOREAD_BITS(priv, SIR0_STATUS, TX_READY))
|
|
|
|
usleep_range(10, 20);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_phy_xgmii_mode(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
struct amd_xgbe_phy_priv *priv = phydev->priv;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
/* Enable KR training */
|
|
|
|
ret = amd_xgbe_an_enable_kr_training(phydev);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
/* Set PCS to KR/10G speed */
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL2);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
ret &= ~MDIO_PCS_CTRL2_TYPE;
|
|
|
|
ret |= MDIO_PCS_CTRL2_10GBR;
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL2, ret);
|
|
|
|
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL1);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
ret &= ~MDIO_CTRL1_SPEEDSEL;
|
|
|
|
ret |= MDIO_CTRL1_SPEED10G;
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL1, ret);
|
|
|
|
|
|
|
|
ret = amd_xgbe_phy_pcs_power_cycle(phydev);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
/* Set SerDes to 10G speed */
|
|
|
|
amd_xgbe_phy_serdes_start_ratechange(phydev);
|
|
|
|
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, DATARATE, SPEED_10000_RATE);
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, WORDMODE, SPEED_10000_WORD);
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, TXAMP, SPEED_10000_TXAMP);
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, PLLSEL, SPEED_10000_PLL);
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, PI_SPD_SEL, SPEED_10000_CDR);
|
|
|
|
|
|
|
|
XRXTX_IOWRITE_BITS(priv, RXTX_REG20, BLWC_ENA, RXTX_10000_BLWC);
|
|
|
|
XRXTX_IOWRITE_BITS(priv, RXTX_REG114, PQ_REG, RXTX_10000_PQ);
|
|
|
|
|
|
|
|
amd_xgbe_phy_serdes_complete_ratechange(phydev);
|
|
|
|
|
|
|
|
priv->mode = AMD_XGBE_MODE_KR;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_phy_gmii_2500_mode(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
struct amd_xgbe_phy_priv *priv = phydev->priv;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
/* Disable KR training */
|
|
|
|
ret = amd_xgbe_an_disable_kr_training(phydev);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
/* Set PCS to KX/1G speed */
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL2);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
ret &= ~MDIO_PCS_CTRL2_TYPE;
|
|
|
|
ret |= MDIO_PCS_CTRL2_10GBX;
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL2, ret);
|
|
|
|
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL1);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
ret &= ~MDIO_CTRL1_SPEEDSEL;
|
|
|
|
ret |= MDIO_CTRL1_SPEED1G;
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL1, ret);
|
|
|
|
|
|
|
|
ret = amd_xgbe_phy_pcs_power_cycle(phydev);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
/* Set SerDes to 2.5G speed */
|
|
|
|
amd_xgbe_phy_serdes_start_ratechange(phydev);
|
|
|
|
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, DATARATE, SPEED_2500_RATE);
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, WORDMODE, SPEED_2500_WORD);
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, TXAMP, SPEED_2500_TXAMP);
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, PLLSEL, SPEED_2500_PLL);
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, PI_SPD_SEL, SPEED_2500_CDR);
|
|
|
|
|
|
|
|
XRXTX_IOWRITE_BITS(priv, RXTX_REG20, BLWC_ENA, RXTX_2500_BLWC);
|
|
|
|
XRXTX_IOWRITE_BITS(priv, RXTX_REG114, PQ_REG, RXTX_2500_PQ);
|
|
|
|
|
|
|
|
amd_xgbe_phy_serdes_complete_ratechange(phydev);
|
|
|
|
|
|
|
|
priv->mode = AMD_XGBE_MODE_KX;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_phy_gmii_mode(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
struct amd_xgbe_phy_priv *priv = phydev->priv;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
/* Disable KR training */
|
|
|
|
ret = amd_xgbe_an_disable_kr_training(phydev);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
/* Set PCS to KX/1G speed */
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL2);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
ret &= ~MDIO_PCS_CTRL2_TYPE;
|
|
|
|
ret |= MDIO_PCS_CTRL2_10GBX;
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL2, ret);
|
|
|
|
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL1);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
ret &= ~MDIO_CTRL1_SPEEDSEL;
|
|
|
|
ret |= MDIO_CTRL1_SPEED1G;
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL1, ret);
|
|
|
|
|
|
|
|
ret = amd_xgbe_phy_pcs_power_cycle(phydev);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
/* Set SerDes to 1G speed */
|
|
|
|
amd_xgbe_phy_serdes_start_ratechange(phydev);
|
|
|
|
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, DATARATE, SPEED_1000_RATE);
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, WORDMODE, SPEED_1000_WORD);
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, TXAMP, SPEED_1000_TXAMP);
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, PLLSEL, SPEED_1000_PLL);
|
|
|
|
XSIR1_IOWRITE_BITS(priv, SIR1_SPEED, PI_SPD_SEL, SPEED_1000_CDR);
|
|
|
|
|
|
|
|
XRXTX_IOWRITE_BITS(priv, RXTX_REG20, BLWC_ENA, RXTX_1000_BLWC);
|
|
|
|
XRXTX_IOWRITE_BITS(priv, RXTX_REG114, PQ_REG, RXTX_1000_PQ);
|
|
|
|
|
|
|
|
amd_xgbe_phy_serdes_complete_ratechange(phydev);
|
|
|
|
|
|
|
|
priv->mode = AMD_XGBE_MODE_KX;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_phy_switch_mode(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
struct amd_xgbe_phy_priv *priv = phydev->priv;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
/* If we are in KR switch to KX, and vice-versa */
|
|
|
|
if (priv->mode == AMD_XGBE_MODE_KR)
|
|
|
|
ret = amd_xgbe_phy_gmii_mode(phydev);
|
|
|
|
else
|
|
|
|
ret = amd_xgbe_phy_xgmii_mode(phydev);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static enum amd_xgbe_phy_an amd_xgbe_an_switch_mode(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = amd_xgbe_phy_switch_mode(phydev);
|
|
|
|
if (ret < 0)
|
|
|
|
return AMD_XGBE_AN_ERROR;
|
|
|
|
|
|
|
|
return AMD_XGBE_AN_START;
|
|
|
|
}
|
|
|
|
|
|
|
|
static enum amd_xgbe_phy_an amd_xgbe_an_tx_training(struct phy_device *phydev,
|
|
|
|
enum amd_xgbe_phy_rx *state)
|
|
|
|
{
|
|
|
|
struct amd_xgbe_phy_priv *priv = phydev->priv;
|
|
|
|
int ad_reg, lp_reg, ret;
|
|
|
|
|
|
|
|
*state = AMD_XGBE_RX_COMPLETE;
|
|
|
|
|
|
|
|
/* If we're in KX mode then we're done */
|
|
|
|
if (priv->mode == AMD_XGBE_MODE_KX)
|
|
|
|
return AMD_XGBE_AN_EVENT;
|
|
|
|
|
|
|
|
/* Enable/Disable FEC */
|
|
|
|
ad_reg = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_AN_ADVERTISE + 2);
|
|
|
|
if (ad_reg < 0)
|
|
|
|
return AMD_XGBE_AN_ERROR;
|
|
|
|
|
|
|
|
lp_reg = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_AN_LPA + 2);
|
|
|
|
if (lp_reg < 0)
|
|
|
|
return AMD_XGBE_AN_ERROR;
|
|
|
|
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PMAPMD, MDIO_PMA_10GBR_FEC_CTRL);
|
|
|
|
if (ret < 0)
|
|
|
|
return AMD_XGBE_AN_ERROR;
|
|
|
|
|
|
|
|
if ((ad_reg & 0xc000) && (lp_reg & 0xc000))
|
|
|
|
ret |= 0x01;
|
|
|
|
else
|
|
|
|
ret &= ~0x01;
|
|
|
|
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_PMAPMD, MDIO_PMA_10GBR_FEC_CTRL, ret);
|
|
|
|
|
|
|
|
/* Start KR training */
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PMAPMD, MDIO_PMA_10GBR_PMD_CTRL);
|
|
|
|
if (ret < 0)
|
|
|
|
return AMD_XGBE_AN_ERROR;
|
|
|
|
|
|
|
|
ret |= 0x01;
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_PMAPMD, MDIO_PMA_10GBR_PMD_CTRL, ret);
|
|
|
|
|
|
|
|
return AMD_XGBE_AN_EVENT;
|
|
|
|
}
|
|
|
|
|
|
|
|
static enum amd_xgbe_phy_an amd_xgbe_an_tx_xnp(struct phy_device *phydev,
|
|
|
|
enum amd_xgbe_phy_rx *state)
|
|
|
|
{
|
|
|
|
u16 msg;
|
|
|
|
|
|
|
|
*state = AMD_XGBE_RX_XNP;
|
|
|
|
|
|
|
|
msg = XNP_MCF_NULL_MESSAGE;
|
|
|
|
msg |= XNP_MP_FORMATTED;
|
|
|
|
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_AN, MDIO_AN_XNP + 2, 0);
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_AN, MDIO_AN_XNP + 1, 0);
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_AN, MDIO_AN_XNP, msg);
|
|
|
|
|
|
|
|
return AMD_XGBE_AN_EVENT;
|
|
|
|
}
|
|
|
|
|
|
|
|
static enum amd_xgbe_phy_an amd_xgbe_an_rx_bpa(struct phy_device *phydev,
|
|
|
|
enum amd_xgbe_phy_rx *state)
|
|
|
|
{
|
|
|
|
struct amd_xgbe_phy_priv *priv = phydev->priv;
|
|
|
|
unsigned int link_support;
|
|
|
|
int ret, ad_reg, lp_reg;
|
|
|
|
|
|
|
|
/* Read Base Ability register 2 first */
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_AN_LPA + 1);
|
|
|
|
if (ret < 0)
|
|
|
|
return AMD_XGBE_AN_ERROR;
|
|
|
|
|
|
|
|
/* Check for a supported mode, otherwise restart in a different one */
|
|
|
|
link_support = (priv->mode == AMD_XGBE_MODE_KR) ? 0x80 : 0x20;
|
|
|
|
if (!(ret & link_support))
|
|
|
|
return amd_xgbe_an_switch_mode(phydev);
|
|
|
|
|
|
|
|
/* Check Extended Next Page support */
|
|
|
|
ad_reg = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_AN_ADVERTISE);
|
|
|
|
if (ad_reg < 0)
|
|
|
|
return AMD_XGBE_AN_ERROR;
|
|
|
|
|
|
|
|
lp_reg = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_AN_LPA);
|
|
|
|
if (lp_reg < 0)
|
|
|
|
return AMD_XGBE_AN_ERROR;
|
|
|
|
|
|
|
|
return ((ad_reg & XNP_NP_EXCHANGE) || (lp_reg & XNP_NP_EXCHANGE)) ?
|
|
|
|
amd_xgbe_an_tx_xnp(phydev, state) :
|
|
|
|
amd_xgbe_an_tx_training(phydev, state);
|
|
|
|
}
|
|
|
|
|
|
|
|
static enum amd_xgbe_phy_an amd_xgbe_an_rx_xnp(struct phy_device *phydev,
|
|
|
|
enum amd_xgbe_phy_rx *state)
|
|
|
|
{
|
|
|
|
int ad_reg, lp_reg;
|
|
|
|
|
|
|
|
/* Check Extended Next Page support */
|
|
|
|
ad_reg = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_AN_ADVERTISE);
|
|
|
|
if (ad_reg < 0)
|
|
|
|
return AMD_XGBE_AN_ERROR;
|
|
|
|
|
|
|
|
lp_reg = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_AN_LPA);
|
|
|
|
if (lp_reg < 0)
|
|
|
|
return AMD_XGBE_AN_ERROR;
|
|
|
|
|
|
|
|
return ((ad_reg & XNP_NP_EXCHANGE) || (lp_reg & XNP_NP_EXCHANGE)) ?
|
|
|
|
amd_xgbe_an_tx_xnp(phydev, state) :
|
|
|
|
amd_xgbe_an_tx_training(phydev, state);
|
|
|
|
}
|
|
|
|
|
|
|
|
static enum amd_xgbe_phy_an amd_xgbe_an_start(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
struct amd_xgbe_phy_priv *priv = phydev->priv;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
/* Be sure we aren't looping trying to negotiate */
|
|
|
|
if (priv->mode == AMD_XGBE_MODE_KR) {
|
|
|
|
if (priv->kr_state != AMD_XGBE_RX_READY)
|
|
|
|
return AMD_XGBE_AN_NO_LINK;
|
|
|
|
priv->kr_state = AMD_XGBE_RX_BPA;
|
|
|
|
} else {
|
|
|
|
if (priv->kx_state != AMD_XGBE_RX_READY)
|
|
|
|
return AMD_XGBE_AN_NO_LINK;
|
|
|
|
priv->kx_state = AMD_XGBE_RX_BPA;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Set up Advertisement register 3 first */
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_AN_ADVERTISE + 2);
|
|
|
|
if (ret < 0)
|
|
|
|
return AMD_XGBE_AN_ERROR;
|
|
|
|
|
|
|
|
if (phydev->supported & SUPPORTED_10000baseR_FEC)
|
|
|
|
ret |= 0xc000;
|
|
|
|
else
|
|
|
|
ret &= ~0xc000;
|
|
|
|
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_AN, MDIO_AN_ADVERTISE + 2, ret);
|
|
|
|
|
|
|
|
/* Set up Advertisement register 2 next */
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_AN_ADVERTISE + 1);
|
|
|
|
if (ret < 0)
|
|
|
|
return AMD_XGBE_AN_ERROR;
|
|
|
|
|
|
|
|
if (phydev->supported & SUPPORTED_10000baseKR_Full)
|
|
|
|
ret |= 0x80;
|
|
|
|
else
|
|
|
|
ret &= ~0x80;
|
|
|
|
|
|
|
|
if (phydev->supported & SUPPORTED_1000baseKX_Full)
|
|
|
|
ret |= 0x20;
|
|
|
|
else
|
|
|
|
ret &= ~0x20;
|
|
|
|
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_AN, MDIO_AN_ADVERTISE + 1, ret);
|
|
|
|
|
|
|
|
/* Set up Advertisement register 1 last */
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_AN_ADVERTISE);
|
|
|
|
if (ret < 0)
|
|
|
|
return AMD_XGBE_AN_ERROR;
|
|
|
|
|
|
|
|
if (phydev->supported & SUPPORTED_Pause)
|
|
|
|
ret |= 0x400;
|
|
|
|
else
|
|
|
|
ret &= ~0x400;
|
|
|
|
|
|
|
|
if (phydev->supported & SUPPORTED_Asym_Pause)
|
|
|
|
ret |= 0x800;
|
|
|
|
else
|
|
|
|
ret &= ~0x800;
|
|
|
|
|
|
|
|
/* We don't intend to perform XNP */
|
|
|
|
ret &= ~XNP_NP_EXCHANGE;
|
|
|
|
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_AN, MDIO_AN_ADVERTISE, ret);
|
|
|
|
|
|
|
|
/* Enable and start auto-negotiation */
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_AN, MDIO_AN_INT, 0);
|
|
|
|
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_CTRL1);
|
|
|
|
if (ret < 0)
|
|
|
|
return AMD_XGBE_AN_ERROR;
|
|
|
|
|
|
|
|
ret |= MDIO_AN_CTRL1_ENABLE;
|
|
|
|
ret |= MDIO_AN_CTRL1_RESTART;
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_AN, MDIO_CTRL1, ret);
|
|
|
|
|
|
|
|
return AMD_XGBE_AN_EVENT;
|
|
|
|
}
|
|
|
|
|
|
|
|
static enum amd_xgbe_phy_an amd_xgbe_an_event(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
enum amd_xgbe_phy_an new_state;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_AN_INT);
|
|
|
|
if (ret < 0)
|
|
|
|
return AMD_XGBE_AN_ERROR;
|
|
|
|
|
|
|
|
new_state = AMD_XGBE_AN_EVENT;
|
|
|
|
if (ret & XGBE_AN_PG_RCV)
|
|
|
|
new_state = AMD_XGBE_AN_PAGE_RECEIVED;
|
|
|
|
else if (ret & XGBE_AN_INC_LINK)
|
|
|
|
new_state = AMD_XGBE_AN_INCOMPAT_LINK;
|
|
|
|
else if (ret & XGBE_AN_INT_CMPLT)
|
|
|
|
new_state = AMD_XGBE_AN_COMPLETE;
|
|
|
|
|
|
|
|
if (new_state != AMD_XGBE_AN_EVENT)
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_AN, MDIO_AN_INT, 0);
|
|
|
|
|
|
|
|
return new_state;
|
|
|
|
}
|
|
|
|
|
|
|
|
static enum amd_xgbe_phy_an amd_xgbe_an_page_received(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
struct amd_xgbe_phy_priv *priv = phydev->priv;
|
|
|
|
enum amd_xgbe_phy_rx *state;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
state = (priv->mode == AMD_XGBE_MODE_KR) ? &priv->kr_state
|
|
|
|
: &priv->kx_state;
|
|
|
|
|
|
|
|
switch (*state) {
|
|
|
|
case AMD_XGBE_RX_BPA:
|
|
|
|
ret = amd_xgbe_an_rx_bpa(phydev, state);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case AMD_XGBE_RX_XNP:
|
|
|
|
ret = amd_xgbe_an_rx_xnp(phydev, state);
|
|
|
|
break;
|
|
|
|
|
|
|
|
default:
|
|
|
|
ret = AMD_XGBE_AN_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static enum amd_xgbe_phy_an amd_xgbe_an_incompat_link(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
return amd_xgbe_an_switch_mode(phydev);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void amd_xgbe_an_state_machine(struct work_struct *work)
|
|
|
|
{
|
|
|
|
struct amd_xgbe_phy_priv *priv = container_of(work,
|
|
|
|
struct amd_xgbe_phy_priv,
|
|
|
|
an_work);
|
|
|
|
struct phy_device *phydev = priv->phydev;
|
|
|
|
enum amd_xgbe_phy_an cur_state;
|
|
|
|
int sleep;
|
|
|
|
|
|
|
|
while (1) {
|
|
|
|
mutex_lock(&priv->an_mutex);
|
|
|
|
|
|
|
|
cur_state = priv->an_state;
|
|
|
|
|
|
|
|
switch (priv->an_state) {
|
|
|
|
case AMD_XGBE_AN_START:
|
|
|
|
priv->an_state = amd_xgbe_an_start(phydev);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case AMD_XGBE_AN_EVENT:
|
|
|
|
priv->an_state = amd_xgbe_an_event(phydev);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case AMD_XGBE_AN_PAGE_RECEIVED:
|
|
|
|
priv->an_state = amd_xgbe_an_page_received(phydev);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case AMD_XGBE_AN_INCOMPAT_LINK:
|
|
|
|
priv->an_state = amd_xgbe_an_incompat_link(phydev);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case AMD_XGBE_AN_COMPLETE:
|
|
|
|
case AMD_XGBE_AN_NO_LINK:
|
|
|
|
case AMD_XGBE_AN_EXIT:
|
|
|
|
goto exit_unlock;
|
|
|
|
|
|
|
|
default:
|
|
|
|
priv->an_state = AMD_XGBE_AN_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (priv->an_state == AMD_XGBE_AN_ERROR) {
|
|
|
|
netdev_err(phydev->attached_dev,
|
|
|
|
"error during auto-negotiation, state=%u\n",
|
|
|
|
cur_state);
|
|
|
|
goto exit_unlock;
|
|
|
|
}
|
|
|
|
|
|
|
|
sleep = (priv->an_state == AMD_XGBE_AN_EVENT) ? 1 : 0;
|
|
|
|
|
|
|
|
mutex_unlock(&priv->an_mutex);
|
|
|
|
|
|
|
|
if (sleep)
|
|
|
|
usleep_range(20, 50);
|
|
|
|
}
|
|
|
|
|
|
|
|
exit_unlock:
|
|
|
|
priv->an_result = priv->an_state;
|
|
|
|
priv->an_state = AMD_XGBE_AN_READY;
|
|
|
|
|
|
|
|
mutex_unlock(&priv->an_mutex);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_phy_soft_reset(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
int count, ret;
|
|
|
|
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL1);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
ret |= MDIO_CTRL1_RESET;
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL1, ret);
|
|
|
|
|
|
|
|
count = 50;
|
|
|
|
do {
|
|
|
|
msleep(20);
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL1);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
} while ((ret & MDIO_CTRL1_RESET) && --count);
|
|
|
|
|
|
|
|
if (ret & MDIO_CTRL1_RESET)
|
|
|
|
return -ETIMEDOUT;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_phy_config_init(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
/* Initialize supported features */
|
|
|
|
phydev->supported = SUPPORTED_Autoneg;
|
|
|
|
phydev->supported |= SUPPORTED_Pause | SUPPORTED_Asym_Pause;
|
|
|
|
phydev->supported |= SUPPORTED_Backplane;
|
|
|
|
phydev->supported |= SUPPORTED_1000baseKX_Full |
|
|
|
|
SUPPORTED_2500baseX_Full;
|
|
|
|
phydev->supported |= SUPPORTED_10000baseKR_Full |
|
|
|
|
SUPPORTED_10000baseR_FEC;
|
|
|
|
phydev->advertising = phydev->supported;
|
|
|
|
|
|
|
|
/* Turn off and clear interrupts */
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_AN, MDIO_AN_INTMASK, 0);
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_AN, MDIO_AN_INT, 0);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_phy_setup_forced(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
/* Disable auto-negotiation */
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_CTRL1);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
ret &= ~MDIO_AN_CTRL1_ENABLE;
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_AN, MDIO_CTRL1, ret);
|
|
|
|
|
|
|
|
/* Validate/Set specified speed */
|
|
|
|
switch (phydev->speed) {
|
|
|
|
case SPEED_10000:
|
|
|
|
ret = amd_xgbe_phy_xgmii_mode(phydev);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case SPEED_2500:
|
|
|
|
ret = amd_xgbe_phy_gmii_2500_mode(phydev);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case SPEED_1000:
|
|
|
|
ret = amd_xgbe_phy_gmii_mode(phydev);
|
|
|
|
break;
|
|
|
|
|
|
|
|
default:
|
|
|
|
ret = -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
/* Validate duplex mode */
|
|
|
|
if (phydev->duplex != DUPLEX_FULL)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
phydev->pause = 0;
|
|
|
|
phydev->asym_pause = 0;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_phy_config_aneg(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
struct amd_xgbe_phy_priv *priv = phydev->priv;
|
|
|
|
u32 mmd_mask = phydev->c45_ids.devices_in_package;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (phydev->autoneg != AUTONEG_ENABLE)
|
|
|
|
return amd_xgbe_phy_setup_forced(phydev);
|
|
|
|
|
|
|
|
/* Make sure we have the AN MMD present */
|
|
|
|
if (!(mmd_mask & MDIO_DEVS_AN))
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
/* Get the current speed mode */
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL2);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
/* Start/Restart the auto-negotiation state machine */
|
|
|
|
mutex_lock(&priv->an_mutex);
|
|
|
|
priv->an_result = AMD_XGBE_AN_READY;
|
|
|
|
priv->an_state = AMD_XGBE_AN_START;
|
|
|
|
priv->kr_state = AMD_XGBE_RX_READY;
|
|
|
|
priv->kx_state = AMD_XGBE_RX_READY;
|
|
|
|
mutex_unlock(&priv->an_mutex);
|
|
|
|
|
|
|
|
queue_work(priv->an_workqueue, &priv->an_work);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_phy_aneg_done(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
struct amd_xgbe_phy_priv *priv = phydev->priv;
|
|
|
|
enum amd_xgbe_phy_an state;
|
|
|
|
|
|
|
|
mutex_lock(&priv->an_mutex);
|
|
|
|
state = priv->an_result;
|
|
|
|
mutex_unlock(&priv->an_mutex);
|
|
|
|
|
|
|
|
return (state == AMD_XGBE_AN_COMPLETE);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_phy_update_link(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
struct amd_xgbe_phy_priv *priv = phydev->priv;
|
|
|
|
enum amd_xgbe_phy_an state;
|
|
|
|
unsigned int check_again, autoneg;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
/* If we're doing auto-negotiation don't report link down */
|
|
|
|
mutex_lock(&priv->an_mutex);
|
|
|
|
state = priv->an_state;
|
|
|
|
mutex_unlock(&priv->an_mutex);
|
|
|
|
|
|
|
|
if (state != AMD_XGBE_AN_READY) {
|
|
|
|
phydev->link = 1;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Since the device can be in the wrong mode when a link is
|
|
|
|
* (re-)established (cable connected after the interface is
|
|
|
|
* up, etc.), the link status may report no link. If there
|
|
|
|
* is no link, try switching modes and checking the status
|
|
|
|
* again.
|
|
|
|
*/
|
|
|
|
check_again = 1;
|
|
|
|
again:
|
|
|
|
/* Link status is latched low, so read once to clear
|
|
|
|
* and then read again to get current state
|
|
|
|
*/
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PCS, MDIO_STAT1);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PCS, MDIO_STAT1);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
phydev->link = (ret & MDIO_STAT1_LSTATUS) ? 1 : 0;
|
|
|
|
|
|
|
|
if (!phydev->link) {
|
|
|
|
ret = amd_xgbe_phy_switch_mode(phydev);
|
|
|
|
if (check_again) {
|
|
|
|
check_again = 0;
|
|
|
|
goto again;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
autoneg = (phydev->link && !priv->link) ? 1 : 0;
|
|
|
|
priv->link = phydev->link;
|
|
|
|
if (autoneg) {
|
|
|
|
/* Link is (back) up, re-start auto-negotiation */
|
|
|
|
ret = amd_xgbe_phy_config_aneg(phydev);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_phy_read_status(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
u32 mmd_mask = phydev->c45_ids.devices_in_package;
|
|
|
|
int ret, mode, ad_ret, lp_ret;
|
|
|
|
|
|
|
|
ret = amd_xgbe_phy_update_link(phydev);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
mode = phy_read_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL2);
|
|
|
|
if (mode < 0)
|
|
|
|
return mode;
|
|
|
|
mode &= MDIO_PCS_CTRL2_TYPE;
|
|
|
|
|
|
|
|
if (phydev->autoneg == AUTONEG_ENABLE) {
|
|
|
|
if (!(mmd_mask & MDIO_DEVS_AN))
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
if (!amd_xgbe_phy_aneg_done(phydev))
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
/* Compare Advertisement and Link Partner register 1 */
|
|
|
|
ad_ret = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_AN_ADVERTISE);
|
|
|
|
if (ad_ret < 0)
|
|
|
|
return ad_ret;
|
|
|
|
lp_ret = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_AN_LPA);
|
|
|
|
if (lp_ret < 0)
|
|
|
|
return lp_ret;
|
|
|
|
|
|
|
|
ad_ret &= lp_ret;
|
|
|
|
phydev->pause = (ad_ret & 0x400) ? 1 : 0;
|
|
|
|
phydev->asym_pause = (ad_ret & 0x800) ? 1 : 0;
|
|
|
|
|
|
|
|
/* Compare Advertisement and Link Partner register 2 */
|
|
|
|
ad_ret = phy_read_mmd(phydev, MDIO_MMD_AN,
|
|
|
|
MDIO_AN_ADVERTISE + 1);
|
|
|
|
if (ad_ret < 0)
|
|
|
|
return ad_ret;
|
|
|
|
lp_ret = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_AN_LPA + 1);
|
|
|
|
if (lp_ret < 0)
|
|
|
|
return lp_ret;
|
|
|
|
|
|
|
|
ad_ret &= lp_ret;
|
|
|
|
if (ad_ret & 0x80) {
|
|
|
|
phydev->speed = SPEED_10000;
|
|
|
|
if (mode != MDIO_PCS_CTRL2_10GBR) {
|
|
|
|
ret = amd_xgbe_phy_xgmii_mode(phydev);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
phydev->speed = SPEED_1000;
|
|
|
|
if (mode == MDIO_PCS_CTRL2_10GBR) {
|
|
|
|
ret = amd_xgbe_phy_gmii_mode(phydev);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
phydev->duplex = DUPLEX_FULL;
|
|
|
|
} else {
|
|
|
|
phydev->speed = (mode == MDIO_PCS_CTRL2_10GBR) ? SPEED_10000
|
|
|
|
: SPEED_1000;
|
|
|
|
phydev->duplex = DUPLEX_FULL;
|
|
|
|
phydev->pause = 0;
|
|
|
|
phydev->asym_pause = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_phy_suspend(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
mutex_lock(&phydev->lock);
|
|
|
|
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL1);
|
|
|
|
if (ret < 0)
|
|
|
|
goto unlock;
|
|
|
|
|
|
|
|
ret |= MDIO_CTRL1_LPOWER;
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL1, ret);
|
|
|
|
|
|
|
|
ret = 0;
|
|
|
|
|
|
|
|
unlock:
|
|
|
|
mutex_unlock(&phydev->lock);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_phy_resume(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
mutex_lock(&phydev->lock);
|
|
|
|
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL1);
|
|
|
|
if (ret < 0)
|
|
|
|
goto unlock;
|
|
|
|
|
|
|
|
ret &= ~MDIO_CTRL1_LPOWER;
|
|
|
|
phy_write_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL1, ret);
|
|
|
|
|
|
|
|
ret = 0;
|
|
|
|
|
|
|
|
unlock:
|
|
|
|
mutex_unlock(&phydev->lock);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_phy_probe(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
struct amd_xgbe_phy_priv *priv;
|
|
|
|
struct platform_device *pdev;
|
|
|
|
struct device *dev;
|
|
|
|
char *wq_name;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (!phydev->dev.of_node)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
pdev = of_find_device_by_node(phydev->dev.of_node);
|
|
|
|
if (!pdev)
|
|
|
|
return -EINVAL;
|
|
|
|
dev = &pdev->dev;
|
|
|
|
|
|
|
|
wq_name = kasprintf(GFP_KERNEL, "%s-amd-xgbe-phy", phydev->bus->name);
|
|
|
|
if (!wq_name) {
|
|
|
|
ret = -ENOMEM;
|
|
|
|
goto err_pdev;
|
|
|
|
}
|
|
|
|
|
|
|
|
priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
|
|
|
|
if (!priv) {
|
|
|
|
ret = -ENOMEM;
|
|
|
|
goto err_name;
|
|
|
|
}
|
|
|
|
|
|
|
|
priv->pdev = pdev;
|
|
|
|
priv->dev = dev;
|
|
|
|
priv->phydev = phydev;
|
|
|
|
|
|
|
|
/* Get the device mmio areas */
|
|
|
|
priv->rxtx_res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
|
|
|
|
priv->rxtx_regs = devm_ioremap_resource(dev, priv->rxtx_res);
|
|
|
|
if (IS_ERR(priv->rxtx_regs)) {
|
|
|
|
dev_err(dev, "rxtx ioremap failed\n");
|
|
|
|
ret = PTR_ERR(priv->rxtx_regs);
|
|
|
|
goto err_priv;
|
|
|
|
}
|
|
|
|
|
|
|
|
priv->sir0_res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
|
|
|
|
priv->sir0_regs = devm_ioremap_resource(dev, priv->sir0_res);
|
|
|
|
if (IS_ERR(priv->sir0_regs)) {
|
|
|
|
dev_err(dev, "sir0 ioremap failed\n");
|
|
|
|
ret = PTR_ERR(priv->sir0_regs);
|
|
|
|
goto err_rxtx;
|
|
|
|
}
|
|
|
|
|
|
|
|
priv->sir1_res = platform_get_resource(pdev, IORESOURCE_MEM, 2);
|
|
|
|
priv->sir1_regs = devm_ioremap_resource(dev, priv->sir1_res);
|
|
|
|
if (IS_ERR(priv->sir1_regs)) {
|
|
|
|
dev_err(dev, "sir1 ioremap failed\n");
|
|
|
|
ret = PTR_ERR(priv->sir1_regs);
|
|
|
|
goto err_sir0;
|
|
|
|
}
|
|
|
|
|
|
|
|
priv->link = 1;
|
|
|
|
|
|
|
|
ret = phy_read_mmd(phydev, MDIO_MMD_PCS, MDIO_CTRL2);
|
|
|
|
if (ret < 0)
|
|
|
|
goto err_sir1;
|
|
|
|
if ((ret & MDIO_PCS_CTRL2_TYPE) == MDIO_PCS_CTRL2_10GBR)
|
|
|
|
priv->mode = AMD_XGBE_MODE_KR;
|
|
|
|
else
|
|
|
|
priv->mode = AMD_XGBE_MODE_KX;
|
|
|
|
|
|
|
|
mutex_init(&priv->an_mutex);
|
|
|
|
INIT_WORK(&priv->an_work, amd_xgbe_an_state_machine);
|
|
|
|
priv->an_workqueue = create_singlethread_workqueue(wq_name);
|
|
|
|
if (!priv->an_workqueue) {
|
|
|
|
ret = -ENOMEM;
|
|
|
|
goto err_sir1;
|
|
|
|
}
|
|
|
|
|
|
|
|
phydev->priv = priv;
|
|
|
|
|
|
|
|
kfree(wq_name);
|
|
|
|
of_dev_put(pdev);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
err_sir1:
|
|
|
|
devm_iounmap(dev, priv->sir1_regs);
|
|
|
|
devm_release_mem_region(dev, priv->sir1_res->start,
|
|
|
|
resource_size(priv->sir1_res));
|
|
|
|
|
|
|
|
err_sir0:
|
|
|
|
devm_iounmap(dev, priv->sir0_regs);
|
|
|
|
devm_release_mem_region(dev, priv->sir0_res->start,
|
|
|
|
resource_size(priv->sir0_res));
|
|
|
|
|
|
|
|
err_rxtx:
|
|
|
|
devm_iounmap(dev, priv->rxtx_regs);
|
|
|
|
devm_release_mem_region(dev, priv->rxtx_res->start,
|
|
|
|
resource_size(priv->rxtx_res));
|
|
|
|
|
|
|
|
err_priv:
|
|
|
|
devm_kfree(dev, priv);
|
|
|
|
|
|
|
|
err_name:
|
|
|
|
kfree(wq_name);
|
|
|
|
|
|
|
|
err_pdev:
|
|
|
|
of_dev_put(pdev);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void amd_xgbe_phy_remove(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
struct amd_xgbe_phy_priv *priv = phydev->priv;
|
|
|
|
struct device *dev = priv->dev;
|
|
|
|
|
|
|
|
/* Stop any in process auto-negotiation */
|
|
|
|
mutex_lock(&priv->an_mutex);
|
|
|
|
priv->an_state = AMD_XGBE_AN_EXIT;
|
|
|
|
mutex_unlock(&priv->an_mutex);
|
|
|
|
|
|
|
|
flush_workqueue(priv->an_workqueue);
|
|
|
|
destroy_workqueue(priv->an_workqueue);
|
|
|
|
|
|
|
|
/* Release resources */
|
|
|
|
devm_iounmap(dev, priv->sir1_regs);
|
|
|
|
devm_release_mem_region(dev, priv->sir1_res->start,
|
|
|
|
resource_size(priv->sir1_res));
|
|
|
|
|
|
|
|
devm_iounmap(dev, priv->sir0_regs);
|
|
|
|
devm_release_mem_region(dev, priv->sir0_res->start,
|
|
|
|
resource_size(priv->sir0_res));
|
|
|
|
|
|
|
|
devm_iounmap(dev, priv->rxtx_regs);
|
|
|
|
devm_release_mem_region(dev, priv->rxtx_res->start,
|
|
|
|
resource_size(priv->rxtx_res));
|
|
|
|
|
|
|
|
devm_kfree(dev, priv);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int amd_xgbe_match_phy_device(struct phy_device *phydev)
|
|
|
|
{
|
|
|
|
return phydev->c45_ids.device_ids[MDIO_MMD_PCS] == XGBE_PHY_ID;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct phy_driver amd_xgbe_phy_driver[] = {
|
|
|
|
{
|
|
|
|
.phy_id = XGBE_PHY_ID,
|
|
|
|
.phy_id_mask = XGBE_PHY_MASK,
|
|
|
|
.name = "AMD XGBE PHY",
|
|
|
|
.features = 0,
|
|
|
|
.probe = amd_xgbe_phy_probe,
|
|
|
|
.remove = amd_xgbe_phy_remove,
|
|
|
|
.soft_reset = amd_xgbe_phy_soft_reset,
|
|
|
|
.config_init = amd_xgbe_phy_config_init,
|
|
|
|
.suspend = amd_xgbe_phy_suspend,
|
|
|
|
.resume = amd_xgbe_phy_resume,
|
|
|
|
.config_aneg = amd_xgbe_phy_config_aneg,
|
|
|
|
.aneg_done = amd_xgbe_phy_aneg_done,
|
|
|
|
.read_status = amd_xgbe_phy_read_status,
|
|
|
|
.match_phy_device = amd_xgbe_match_phy_device,
|
|
|
|
.driver = {
|
|
|
|
.owner = THIS_MODULE,
|
|
|
|
},
|
|
|
|
},
|
|
|
|
};
|
|
|
|
|
|
|
|
static int __init amd_xgbe_phy_init(void)
|
|
|
|
{
|
|
|
|
return phy_drivers_register(amd_xgbe_phy_driver,
|
|
|
|
ARRAY_SIZE(amd_xgbe_phy_driver));
|
|
|
|
}
|
|
|
|
|
|
|
|
static void __exit amd_xgbe_phy_exit(void)
|
|
|
|
{
|
|
|
|
phy_drivers_unregister(amd_xgbe_phy_driver,
|
|
|
|
ARRAY_SIZE(amd_xgbe_phy_driver));
|
|
|
|
}
|
|
|
|
|
|
|
|
module_init(amd_xgbe_phy_init);
|
|
|
|
module_exit(amd_xgbe_phy_exit);
|
|
|
|
|
2014-06-07 16:07:48 +07:00
|
|
|
static struct mdio_device_id __maybe_unused amd_xgbe_phy_ids[] = {
|
2014-06-05 21:15:12 +07:00
|
|
|
{ XGBE_PHY_ID, XGBE_PHY_MASK },
|
|
|
|
{ }
|
|
|
|
};
|
|
|
|
MODULE_DEVICE_TABLE(mdio, amd_xgbe_phy_ids);
|