u-boot/board/freescale/ls1012aqds/eth.c

310 lines
7.3 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0+
/*
* Copyright 2015-2016 Freescale Semiconductor, Inc.
* Copyright 2017 NXP
*/
#include <common.h>
#include <dm.h>
#include <asm/io.h>
#include <netdev.h>
#include <fm_eth.h>
#include <fsl_mdio.h>
#include <malloc.h>
#include <asm/types.h>
#include <fsl_dtsec.h>
#include <asm/arch/soc.h>
#include <asm/arch-fsl-layerscape/config.h>
#include <asm/arch-fsl-layerscape/immap_lsch2.h>
#include <asm/arch/fsl_serdes.h>
#include <linux/delay.h>
#include "../common/qixis.h"
#include <net/pfe_eth/pfe_eth.h>
#include <dm/platform_data/pfe_dm_eth.h>
#include "ls1012aqds_qixis.h"
#define EMI_NONE 0xFF
#define EMI1_RGMII 1
#define EMI1_SLOT1 2
#define EMI1_SLOT2 3
#define DEFAULT_PFE_MDIO_NAME "PFE_MDIO"
#define DEFAULT_PFE_MDIO1_NAME "PFE_MDIO1"
static const char * const mdio_names[] = {
"NULL",
"LS1012AQDS_MDIO_RGMII",
"LS1012AQDS_MDIO_SLOT1",
"LS1012AQDS_MDIO_SLOT2",
"NULL",
};
static const char *ls1012aqds_mdio_name_for_muxval(u8 muxval)
{
return mdio_names[muxval];
}
struct ls1012aqds_mdio {
u8 muxval;
struct mii_dev *realbus;
};
static void ls1012aqds_mux_mdio(u8 muxval)
{
u8 brdcfg4;
if (muxval < 7) {
brdcfg4 = QIXIS_READ(brdcfg[4]);
brdcfg4 &= ~BRDCFG4_EMISEL_MASK;
brdcfg4 |= (muxval << BRDCFG4_EMISEL_SHIFT);
QIXIS_WRITE(brdcfg[4], brdcfg4);
}
}
static int ls1012aqds_mdio_read(struct mii_dev *bus, int addr, int devad,
int regnum)
{
struct ls1012aqds_mdio *priv = bus->priv;
ls1012aqds_mux_mdio(priv->muxval);
return priv->realbus->read(priv->realbus, addr, devad, regnum);
}
static int ls1012aqds_mdio_write(struct mii_dev *bus, int addr, int devad,
int regnum, u16 value)
{
struct ls1012aqds_mdio *priv = bus->priv;
ls1012aqds_mux_mdio(priv->muxval);
return priv->realbus->write(priv->realbus, addr, devad, regnum, value);
}
static int ls1012aqds_mdio_reset(struct mii_dev *bus)
{
struct ls1012aqds_mdio *priv = bus->priv;
if (priv->realbus->reset)
return priv->realbus->reset(priv->realbus);
else
return -1;
}
static int ls1012aqds_mdio_init(char *realbusname, u8 muxval)
{
struct ls1012aqds_mdio *pmdio;
struct mii_dev *bus = mdio_alloc();
if (!bus) {
printf("Failed to allocate ls1012aqds MDIO bus\n");
return -1;
}
pmdio = malloc(sizeof(*pmdio));
if (!pmdio) {
printf("Failed to allocate ls1012aqds private data\n");
free(bus);
return -1;
}
bus->read = ls1012aqds_mdio_read;
bus->write = ls1012aqds_mdio_write;
bus->reset = ls1012aqds_mdio_reset;
sprintf(bus->name, ls1012aqds_mdio_name_for_muxval(muxval));
pmdio->realbus = miiphy_get_dev_by_name(realbusname);
if (!pmdio->realbus) {
printf("No bus with name %s\n", realbusname);
free(bus);
free(pmdio);
return -1;
}
pmdio->muxval = muxval;
bus->priv = pmdio;
return mdio_register(bus);
}
int pfe_eth_board_init(struct udevice *dev)
{
static int init_done;
struct mii_dev *bus;
static const char *mdio_name;
struct pfe_mdio_info mac_mdio_info;
struct ccsr_gur __iomem *gur = (void *)CFG_SYS_FSL_GUTS_ADDR;
u8 data8;
struct pfe_eth_dev *priv = dev_get_priv(dev);
int srds_s1 = in_be32(&gur->rcwsr[4]) &
FSL_CHASSIS2_RCWSR4_SRDS1_PRTCL_MASK;
srds_s1 >>= FSL_CHASSIS2_RCWSR4_SRDS1_PRTCL_SHIFT;
ls1012aqds_mux_mdio(EMI1_SLOT1);
if (!init_done) {
mac_mdio_info.reg_base = (void *)EMAC1_BASE_ADDR;
mac_mdio_info.name = DEFAULT_PFE_MDIO_NAME;
bus = pfe_mdio_init(&mac_mdio_info);
if (!bus) {
printf("Failed to register mdio\n");
return -1;
}
init_done = 1;
}
if (priv->gemac_port) {
mac_mdio_info.reg_base = (void *)EMAC2_BASE_ADDR;
mac_mdio_info.name = DEFAULT_PFE_MDIO1_NAME;
bus = pfe_mdio_init(&mac_mdio_info);
if (!bus) {
printf("Failed to register mdio\n");
return -1;
}
}
switch (srds_s1) {
case 0x3508:
printf("ls1012aqds:supported SerDes PRCTL= %d\n", srds_s1);
#ifdef CONFIG_PFE_RGMII_RESET_WA
/*
* Work around for FPGA registers initialization
* This is needed for RGMII to work.
*/
printf("Reset RGMII WA....\n");
data8 = QIXIS_READ(rst_frc[0]);
data8 |= 0x2;
QIXIS_WRITE(rst_frc[0], data8);
data8 = QIXIS_READ(rst_frc[0]);
data8 = QIXIS_READ(res8[6]);
data8 |= 0xff;
QIXIS_WRITE(res8[6], data8);
data8 = QIXIS_READ(res8[6]);
#endif
if (priv->gemac_port) {
mdio_name = ls1012aqds_mdio_name_for_muxval(EMI1_RGMII);
if (ls1012aqds_mdio_init(DEFAULT_PFE_MDIO_NAME, EMI1_RGMII)
< 0) {
printf("Failed to register mdio for %s\n", mdio_name);
}
/* MAC2 */
mdio_name = ls1012aqds_mdio_name_for_muxval(EMI1_RGMII);
bus = miiphy_get_dev_by_name(mdio_name);
pfe_set_mdio(priv->gemac_port, bus);
pfe_set_phy_address_mode(priv->gemac_port,
CONFIG_PFE_EMAC2_PHY_ADDR,
PHY_INTERFACE_MODE_RGMII);
} else {
mdio_name = ls1012aqds_mdio_name_for_muxval(EMI1_SLOT1);
if (ls1012aqds_mdio_init(DEFAULT_PFE_MDIO_NAME, EMI1_SLOT1)
< 0) {
printf("Failed to register mdio for %s\n", mdio_name);
}
/* MAC1 */
mdio_name = ls1012aqds_mdio_name_for_muxval(EMI1_SLOT1);
bus = miiphy_get_dev_by_name(mdio_name);
pfe_set_mdio(priv->gemac_port, bus);
pfe_set_phy_address_mode(priv->gemac_port,
CONFIG_PFE_EMAC1_PHY_ADDR,
PHY_INTERFACE_MODE_SGMII);
}
break;
case 0x2205:
printf("ls1012aqds:supported SerDes PRCTL= %d\n", srds_s1);
/*
* Work around for FPGA registers initialization
* This is needed for RGMII to work.
*/
printf("Reset SLOT1 SLOT2....\n");
data8 = QIXIS_READ(rst_frc[2]);
data8 |= 0xc0;
QIXIS_WRITE(rst_frc[2], data8);
mdelay(100);
data8 = QIXIS_READ(rst_frc[2]);
data8 &= 0x3f;
QIXIS_WRITE(rst_frc[2], data8);
if (priv->gemac_port) {
mdio_name = ls1012aqds_mdio_name_for_muxval(EMI1_SLOT2);
if (ls1012aqds_mdio_init(DEFAULT_PFE_MDIO_NAME, EMI1_SLOT2)
< 0) {
printf("Failed to register mdio for %s\n", mdio_name);
}
/* MAC2 */
mdio_name = ls1012aqds_mdio_name_for_muxval(EMI1_SLOT2);
bus = miiphy_get_dev_by_name(mdio_name);
pfe_set_mdio(1, bus);
pfe_set_phy_address_mode(1, CONFIG_PFE_SGMII_2500_PHY2_ADDR,
net: freescale: replace usage of phy-mode = "sgmii-2500" with "2500base-x" After the discussion here: https://lore.kernel.org/netdev/20210603143453.if7hgifupx5k433b@pali/ which resulted in this patch: https://patchwork.kernel.org/project/netdevbpf/patch/20210704134325.24842-1-pali@kernel.org/ and many other discussions before it, notably: https://patchwork.kernel.org/project/linux-arm-kernel/patch/1512016235-15909-1-git-send-email-Bhaskar.Upadhaya@nxp.com/ it became apparent that nobody really knows what "SGMII 2500" is. Certainly, Freescale/NXP hardware engineers name this protocol "SGMII 2500" in the reference manuals, but the PCS devices do not support any "SGMII" specific features when operating at the speed of 2500 Mbps, no in-band autoneg and no speed change via symbol replication . So that leaves a fixed speed of 2500 Mbps using a coding of 8b/10b with a SERDES lane frequency of 3.125 GHz. In fact, "SGMII 2500 without in-band autoneg and at a fixed speed" is indistinguishable from "2500base-x without in-band autoneg", which is precisely what these NXP devices support. So it just appears that "SGMII 2500" is an unclear name with no clear definition that stuck. As such, in the Linux kernel, the drivers which use this SERDES protocol use the 2500base-x phy-mode. This patch converts U-Boot to use 2500base-x too, or at least, as much as it can. Note that I would have really liked to delete PHY_INTERFACE_MODE_SGMII_2500 completely, but the mvpp2 driver seems to even distinguish between SGMII 2500 and 2500base-X. Namely, it enables in-band autoneg for one but not the other, and forces flow control for one but not the other. This goes back to the idea that maybe 2500base-X is a fiber protocol and SGMII-2500 is an MII protocol (connects a MAC to a PHY such as Aquantia), but the two are practically indistinguishable through everything except use case. NXP devices can support both use cases through an identical configuration, for example RX flow control can be unconditionally enabled in order to support rate adaptation performed by an Aquantia PHY. At least I can find no indication in online documents published by Cisco which would point towards "SGMII-2500" being an actual standard with an actual definition, so I cannot say "yes, NXP devices support it". Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
2021-09-18 12:32:35 +00:00
PHY_INTERFACE_MODE_2500BASEX);
data8 = QIXIS_READ(brdcfg[12]);
data8 |= 0x20;
QIXIS_WRITE(brdcfg[12], data8);
} else {
mdio_name = ls1012aqds_mdio_name_for_muxval(EMI1_SLOT1);
if (ls1012aqds_mdio_init(DEFAULT_PFE_MDIO_NAME, EMI1_SLOT1)
< 0) {
printf("Failed to register mdio for %s\n", mdio_name);
}
/* MAC1 */
mdio_name = ls1012aqds_mdio_name_for_muxval(EMI1_SLOT1);
bus = miiphy_get_dev_by_name(mdio_name);
pfe_set_mdio(0, bus);
pfe_set_phy_address_mode(0,
CONFIG_PFE_SGMII_2500_PHY1_ADDR,
net: freescale: replace usage of phy-mode = "sgmii-2500" with "2500base-x" After the discussion here: https://lore.kernel.org/netdev/20210603143453.if7hgifupx5k433b@pali/ which resulted in this patch: https://patchwork.kernel.org/project/netdevbpf/patch/20210704134325.24842-1-pali@kernel.org/ and many other discussions before it, notably: https://patchwork.kernel.org/project/linux-arm-kernel/patch/1512016235-15909-1-git-send-email-Bhaskar.Upadhaya@nxp.com/ it became apparent that nobody really knows what "SGMII 2500" is. Certainly, Freescale/NXP hardware engineers name this protocol "SGMII 2500" in the reference manuals, but the PCS devices do not support any "SGMII" specific features when operating at the speed of 2500 Mbps, no in-band autoneg and no speed change via symbol replication . So that leaves a fixed speed of 2500 Mbps using a coding of 8b/10b with a SERDES lane frequency of 3.125 GHz. In fact, "SGMII 2500 without in-band autoneg and at a fixed speed" is indistinguishable from "2500base-x without in-band autoneg", which is precisely what these NXP devices support. So it just appears that "SGMII 2500" is an unclear name with no clear definition that stuck. As such, in the Linux kernel, the drivers which use this SERDES protocol use the 2500base-x phy-mode. This patch converts U-Boot to use 2500base-x too, or at least, as much as it can. Note that I would have really liked to delete PHY_INTERFACE_MODE_SGMII_2500 completely, but the mvpp2 driver seems to even distinguish between SGMII 2500 and 2500base-X. Namely, it enables in-band autoneg for one but not the other, and forces flow control for one but not the other. This goes back to the idea that maybe 2500base-X is a fiber protocol and SGMII-2500 is an MII protocol (connects a MAC to a PHY such as Aquantia), but the two are practically indistinguishable through everything except use case. NXP devices can support both use cases through an identical configuration, for example RX flow control can be unconditionally enabled in order to support rate adaptation performed by an Aquantia PHY. At least I can find no indication in online documents published by Cisco which would point towards "SGMII-2500" being an actual standard with an actual definition, so I cannot say "yes, NXP devices support it". Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
2021-09-18 12:32:35 +00:00
PHY_INTERFACE_MODE_2500BASEX);
}
break;
default:
printf("ls1012aqds:unsupported SerDes PRCTL= %d\n", srds_s1);
break;
}
return 0;
}
static struct pfe_eth_pdata pfe_pdata0 = {
.pfe_eth_pdata_mac = {
.iobase = (phys_addr_t)EMAC1_BASE_ADDR,
.phy_interface = 0,
},
.pfe_ddr_addr = {
.ddr_pfe_baseaddr = (void *)CONFIG_DDR_PFE_BASEADDR,
.ddr_pfe_phys_baseaddr = CONFIG_DDR_PFE_PHYS_BASEADDR,
},
};
static struct pfe_eth_pdata pfe_pdata1 = {
.pfe_eth_pdata_mac = {
.iobase = (phys_addr_t)EMAC2_BASE_ADDR,
.phy_interface = 1,
},
.pfe_ddr_addr = {
.ddr_pfe_baseaddr = (void *)CONFIG_DDR_PFE_BASEADDR,
.ddr_pfe_phys_baseaddr = CONFIG_DDR_PFE_PHYS_BASEADDR,
},
};
U_BOOT_DRVINFO(ls1012a_pfe0) = {
.name = "pfe_eth",
.plat = &pfe_pdata0,
};
U_BOOT_DRVINFO(ls1012a_pfe1) = {
.name = "pfe_eth",
.plat = &pfe_pdata1,
};