u-boot/board/freescale/ls1046aqds/eth.c

432 lines
11 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0+
/*
* Copyright 2016 Freescale Semiconductor, Inc.
* Copyright 2018-2020 NXP
*/
#include <common.h>
#include <log.h>
#include <net.h>
#include <asm/io.h>
#include <netdev.h>
#include <fdt_support.h>
#include <fm_eth.h>
#include <fsl_mdio.h>
#include <fsl_dtsec.h>
#include <malloc.h>
#include <asm/arch/fsl_serdes.h>
#include "../common/qixis.h"
#include "../common/fman.h"
#include "ls1046aqds_qixis.h"
#define EMI_NONE 0xFF
#define EMI1_RGMII1 0
#define EMI1_RGMII2 1
#define EMI1_SLOT1 2
#define EMI1_SLOT2 3
#define EMI1_SLOT4 4
static const char * const mdio_names[] = {
"LS1046AQDS_MDIO_RGMII1",
"LS1046AQDS_MDIO_RGMII2",
"LS1046AQDS_MDIO_SLOT1",
"LS1046AQDS_MDIO_SLOT2",
"LS1046AQDS_MDIO_SLOT4",
"NULL",
};
/* Map SerDes 1 & 2 lanes to default slot. */
#ifdef CONFIG_FMAN_ENET
static int mdio_mux[NUM_FM_PORTS];
static u8 lane_to_slot[] = {1, 1, 1, 1, 0, 4, 0 , 0};
#endif
static const char *ls1046aqds_mdio_name_for_muxval(u8 muxval)
{
return mdio_names[muxval];
}
struct mii_dev *mii_dev_for_muxval(u8 muxval)
{
struct mii_dev *bus;
const char *name;
if (muxval > EMI1_SLOT4)
return NULL;
name = ls1046aqds_mdio_name_for_muxval(muxval);
if (!name) {
printf("No bus for muxval %x\n", muxval);
return NULL;
}
bus = miiphy_get_dev_by_name(name);
if (!bus) {
printf("No bus by name %s\n", name);
return NULL;
}
return bus;
}
#ifdef CONFIG_FMAN_ENET
struct ls1046aqds_mdio {
u8 muxval;
struct mii_dev *realbus;
};
static void ls1046aqds_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 ls1046aqds_mdio_read(struct mii_dev *bus, int addr, int devad,
int regnum)
{
struct ls1046aqds_mdio *priv = bus->priv;
ls1046aqds_mux_mdio(priv->muxval);
return priv->realbus->read(priv->realbus, addr, devad, regnum);
}
static int ls1046aqds_mdio_write(struct mii_dev *bus, int addr, int devad,
int regnum, u16 value)
{
struct ls1046aqds_mdio *priv = bus->priv;
ls1046aqds_mux_mdio(priv->muxval);
return priv->realbus->write(priv->realbus, addr, devad,
regnum, value);
}
static int ls1046aqds_mdio_reset(struct mii_dev *bus)
{
struct ls1046aqds_mdio *priv = bus->priv;
return priv->realbus->reset(priv->realbus);
}
static int ls1046aqds_mdio_init(char *realbusname, u8 muxval)
{
struct ls1046aqds_mdio *pmdio;
struct mii_dev *bus = mdio_alloc();
if (!bus) {
printf("Failed to allocate ls1046aqds MDIO bus\n");
return -1;
}
pmdio = malloc(sizeof(*pmdio));
if (!pmdio) {
printf("Failed to allocate ls1046aqds private data\n");
free(bus);
return -1;
}
bus->read = ls1046aqds_mdio_read;
bus->write = ls1046aqds_mdio_write;
bus->reset = ls1046aqds_mdio_reset;
sprintf(bus->name, ls1046aqds_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);
}
void board_ft_fman_fixup_port(void *fdt, char *compat, phys_addr_t addr,
enum fm_port port, int offset)
{
struct fixed_link f_link;
const char *phyconn;
if (fm_info_get_enet_if(port) == PHY_INTERFACE_MODE_SGMII) {
switch (port) {
case FM1_DTSEC9:
fdt_set_phy_handle(fdt, compat, addr, "sgmii-s1-p1");
break;
case FM1_DTSEC10:
fdt_set_phy_handle(fdt, compat, addr, "sgmii-s1-p2");
break;
case FM1_DTSEC5:
fdt_set_phy_handle(fdt, compat, addr, "sgmii-s1-p3");
break;
case FM1_DTSEC6:
fdt_set_phy_handle(fdt, compat, addr, "sgmii-s1-p4");
break;
case FM1_DTSEC2:
fdt_set_phy_handle(fdt, compat, addr, "sgmii-s4-p1");
break;
default:
break;
}
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
} else if (fm_info_get_enet_if(port) == PHY_INTERFACE_MODE_2500BASEX) {
/* 2.5G SGMII interface */
f_link.phy_id = cpu_to_fdt32(port);
f_link.duplex = cpu_to_fdt32(1);
f_link.link_speed = cpu_to_fdt32(1000);
f_link.pause = 0;
f_link.asym_pause = 0;
/* no PHY for 2.5G SGMII on QDS */
fdt_delprop(fdt, offset, "phy-handle");
fdt_setprop(fdt, offset, "fixed-link", &f_link, sizeof(f_link));
fdt_setprop_string(fdt, offset, "phy-connection-type",
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
"2500base-x");
} else if (fm_info_get_enet_if(port) == PHY_INTERFACE_MODE_QSGMII) {
switch (port) {
case FM1_DTSEC1:
fdt_set_phy_handle(fdt, compat, addr, "qsgmii-s2-p4");
break;
case FM1_DTSEC5:
fdt_set_phy_handle(fdt, compat, addr, "qsgmii-s2-p2");
break;
case FM1_DTSEC6:
fdt_set_phy_handle(fdt, compat, addr, "qsgmii-s2-p1");
break;
case FM1_DTSEC10:
fdt_set_phy_handle(fdt, compat, addr, "qsgmii-s2-p3");
break;
default:
break;
}
fdt_delprop(fdt, offset, "phy-connection-type");
fdt_setprop_string(fdt, offset, "phy-connection-type",
"qsgmii");
} else if (fm_info_get_enet_if(port) == PHY_INTERFACE_MODE_XGMII &&
(port == FM1_10GEC1 || port == FM1_10GEC2)) {
phyconn = fdt_getprop(fdt, offset, "phy-connection-type", NULL);
if (is_backplane_mode(phyconn)) {
/* Backplane KR mode: skip fixups */
printf("Interface %d in backplane KR mode\n", port);
} else {
/* 10GBase-R interface */
f_link.phy_id = cpu_to_fdt32(port);
f_link.duplex = cpu_to_fdt32(1);
f_link.link_speed = cpu_to_fdt32(10000);
f_link.pause = 0;
f_link.asym_pause = 0;
/* no PHY for 10GBase-R */
fdt_delprop(fdt, offset, "phy-handle");
fdt_setprop(fdt, offset, "fixed-link", &f_link,
sizeof(f_link));
fdt_setprop_string(fdt, offset, "phy-connection-type",
"xgmii");
}
}
}
void fdt_fixup_board_enet(void *fdt)
{
int i;
for (i = FM1_DTSEC1; i < NUM_FM_PORTS; i++) {
switch (fm_info_get_enet_if(i)) {
case PHY_INTERFACE_MODE_SGMII:
case PHY_INTERFACE_MODE_QSGMII:
switch (mdio_mux[i]) {
case EMI1_SLOT1:
fdt_status_okay_by_alias(fdt, "emi1-slot1");
break;
case EMI1_SLOT2:
fdt_status_okay_by_alias(fdt, "emi1-slot2");
break;
case EMI1_SLOT4:
fdt_status_okay_by_alias(fdt, "emi1-slot4");
break;
default:
break;
}
break;
default:
break;
}
}
}
int board_eth_init(struct bd_info *bis)
{
int i, idx, lane, slot, interface;
struct memac_mdio_info dtsec_mdio_info;
struct ccsr_gur *gur = (void *)(CFG_SYS_FSL_GUTS_ADDR);
u32 srds_s1, srds_s2;
u8 brdcfg12;
srds_s1 = in_be32(&gur->rcwsr[4]) &
FSL_CHASSIS2_RCWSR4_SRDS1_PRTCL_MASK;
srds_s1 >>= FSL_CHASSIS2_RCWSR4_SRDS1_PRTCL_SHIFT;
srds_s2 = in_be32(&gur->rcwsr[4]) &
FSL_CHASSIS2_RCWSR4_SRDS2_PRTCL_MASK;
srds_s2 >>= FSL_CHASSIS2_RCWSR4_SRDS2_PRTCL_SHIFT;
/* Initialize the mdio_mux array so we can recognize empty elements */
for (i = 0; i < NUM_FM_PORTS; i++)
mdio_mux[i] = EMI_NONE;
dtsec_mdio_info.regs =
(struct memac_mdio_controller *)CONFIG_SYS_FM1_DTSEC_MDIO_ADDR;
dtsec_mdio_info.name = DEFAULT_FM_MDIO_NAME;
/* Register the 1G MDIO bus */
fm_memac_mdio_init(bis, &dtsec_mdio_info);
/* Register the muxing front-ends to the MDIO buses */
ls1046aqds_mdio_init(DEFAULT_FM_MDIO_NAME, EMI1_RGMII1);
ls1046aqds_mdio_init(DEFAULT_FM_MDIO_NAME, EMI1_RGMII2);
ls1046aqds_mdio_init(DEFAULT_FM_MDIO_NAME, EMI1_SLOT1);
ls1046aqds_mdio_init(DEFAULT_FM_MDIO_NAME, EMI1_SLOT2);
ls1046aqds_mdio_init(DEFAULT_FM_MDIO_NAME, EMI1_SLOT4);
/* Set the two on-board RGMII PHY address */
fm_info_set_phy_address(FM1_DTSEC3, RGMII_PHY1_ADDR);
fm_info_set_phy_address(FM1_DTSEC4, RGMII_PHY2_ADDR);
switch (srds_s1) {
case 0x3333:
/* SGMII on slot 1, MAC 9 */
fm_info_set_phy_address(FM1_DTSEC9, SGMII_CARD_PORT1_PHY_ADDR);
case 0x1333:
case 0x2333:
/* SGMII on slot 1, MAC 10 */
fm_info_set_phy_address(FM1_DTSEC10, SGMII_CARD_PORT2_PHY_ADDR);
case 0x1133:
case 0x2233:
/* SGMII on slot 1, MAC 5/6 */
fm_info_set_phy_address(FM1_DTSEC5, SGMII_CARD_PORT3_PHY_ADDR);
fm_info_set_phy_address(FM1_DTSEC6, SGMII_CARD_PORT4_PHY_ADDR);
break;
case 0x1040:
case 0x2040:
/* QSGMII on lane B, MAC 6/5/10/1 */
fm_info_set_phy_address(FM1_DTSEC6,
QSGMII_CARD_PORT1_PHY_ADDR_S2);
fm_info_set_phy_address(FM1_DTSEC5,
QSGMII_CARD_PORT2_PHY_ADDR_S2);
fm_info_set_phy_address(FM1_DTSEC10,
QSGMII_CARD_PORT3_PHY_ADDR_S2);
fm_info_set_phy_address(FM1_DTSEC1,
QSGMII_CARD_PORT4_PHY_ADDR_S2);
break;
case 0x3363:
/* SGMII on slot 1, MAC 9/10 */
fm_info_set_phy_address(FM1_DTSEC9, SGMII_CARD_PORT1_PHY_ADDR);
fm_info_set_phy_address(FM1_DTSEC10, SGMII_CARD_PORT2_PHY_ADDR);
case 0x1163:
case 0x2263:
case 0x2223:
/* SGMII on slot 1, MAC 6 */
fm_info_set_phy_address(FM1_DTSEC6, SGMII_CARD_PORT4_PHY_ADDR);
break;
default:
printf("Invalid SerDes protocol 0x%x for LS1046AQDS\n",
srds_s1);
break;
}
if (srds_s2 == 0x5a59 || srds_s2 == 0x5a06)
/* SGMII on slot 4, MAC 2 */
fm_info_set_phy_address(FM1_DTSEC2, SGMII_CARD_PORT1_PHY_ADDR);
for (i = FM1_DTSEC1; i < FM1_DTSEC1 + CONFIG_SYS_NUM_FM1_DTSEC; i++) {
idx = i - FM1_DTSEC1;
interface = fm_info_get_enet_if(i);
switch (interface) {
case PHY_INTERFACE_MODE_SGMII:
case PHY_INTERFACE_MODE_QSGMII:
if (interface == PHY_INTERFACE_MODE_SGMII) {
if (i == FM1_DTSEC5) {
/* route lane 2 to slot1 so to have
* one sgmii riser card supports
* MAC5 and MAC6.
*/
brdcfg12 = QIXIS_READ(brdcfg[12]);
QIXIS_WRITE(brdcfg[12],
brdcfg12 | 0x80);
}
lane = serdes_get_first_lane(FSL_SRDS_1,
SGMII_FM1_DTSEC1 + idx);
} else {
/* clear the bit 7 to route lane B on slot2. */
brdcfg12 = QIXIS_READ(brdcfg[12]);
QIXIS_WRITE(brdcfg[12], brdcfg12 & 0x7f);
lane = serdes_get_first_lane(FSL_SRDS_1,
QSGMII_FM1_A);
lane_to_slot[lane] = 2;
}
if (i == FM1_DTSEC2)
lane = 5;
if (lane < 0)
break;
slot = lane_to_slot[lane];
debug("FM1@DTSEC%u expects SGMII in slot %u\n",
idx + 1, slot);
if (QIXIS_READ(present2) & (1 << (slot - 1)))
fm_disable_port(i);
switch (slot) {
case 1:
mdio_mux[i] = EMI1_SLOT1;
fm_info_set_mdio(i, mii_dev_for_muxval(
mdio_mux[i]));
break;
case 2:
mdio_mux[i] = EMI1_SLOT2;
fm_info_set_mdio(i, mii_dev_for_muxval(
mdio_mux[i]));
break;
case 4:
mdio_mux[i] = EMI1_SLOT4;
fm_info_set_mdio(i, mii_dev_for_muxval(
mdio_mux[i]));
break;
default:
break;
}
break;
case PHY_INTERFACE_MODE_RGMII:
case PHY_INTERFACE_MODE_RGMII_TXID:
case PHY_INTERFACE_MODE_RGMII_RXID:
case PHY_INTERFACE_MODE_RGMII_ID:
if (i == FM1_DTSEC3)
mdio_mux[i] = EMI1_RGMII1;
else if (i == FM1_DTSEC4)
mdio_mux[i] = EMI1_RGMII2;
fm_info_set_mdio(i, mii_dev_for_muxval(mdio_mux[i]));
break;
default:
break;
}
}
cpu_eth_init(bis);
return pci_eth_init(bis);
}
#endif /* CONFIG_FMAN_ENET */