On 27/05/16 09:22, Iyappan Subramanian wrote:
This patch fixes MAC configuration to support 10/100GbE for SGMII and
link_state call back. It also sets pdata->mdio_driver flag based on
ethernet mdio subnode and prepare for MDIO driver support.
In summary, following are the changes,
- Added set_speed function pointer in mac_ops
- Changed link_state to call the set_speed
- Add 10/100 support for SGMII based 1G
- Fixed mac_init for 1G
- Call mac_ops rx_enable/disable and tx_enable/disable function pointers
This is just a code clean-up, right? If so, this should be a patch apart.
- Add acpi_phy_find_device to find PHY using phy-handle reference object
- Changing phy_start and phy_stop calls based on phy_dev object existence
- Calling phy_connect based on pdata->mdio_driver flag
Signed-off-by: Iyappan Subramanian <isubraman...@apm.com>
Tested-by: Fushen Chen <fc...@apm.com>
Tested-by: Toan Le <toa...@apm.com>
---
drivers/net/ethernet/apm/xgene/xgene_enet_hw.c | 189 +++++++++++++---------
drivers/net/ethernet/apm/xgene/xgene_enet_hw.h | 4 +
drivers/net/ethernet/apm/xgene/xgene_enet_main.c | 40 +++--
drivers/net/ethernet/apm/xgene/xgene_enet_main.h | 2 +
drivers/net/ethernet/apm/xgene/xgene_enet_sgmac.c | 106 +++++++++++-
drivers/net/ethernet/apm/xgene/xgene_enet_sgmac.h | 8 +
drivers/net/ethernet/apm/xgene/xgene_enet_xgmac.h | 4 +
7 files changed, 256 insertions(+), 97 deletions(-)
diff --git a/drivers/net/ethernet/apm/xgene/xgene_enet_hw.c
b/drivers/net/ethernet/apm/xgene/xgene_enet_hw.c
index 2f5638f..6bc8360 100644
--- a/drivers/net/ethernet/apm/xgene/xgene_enet_hw.c
+++ b/drivers/net/ethernet/apm/xgene/xgene_enet_hw.c
@@ -512,14 +512,11 @@ static void xgene_enet_configure_clock(struct
xgene_enet_pdata *pdata)
[...]
static void xgene_sgmac_init(struct xgene_enet_pdata *p)
{
u32 data, loop = 10;
- u32 offset = p->port_id * 4;
+ u32 offset = 0;
u32 enet_spare_cfg_reg, rsif_config_reg;
u32 cfg_bypass_reg, rx_dv_gate_reg;
xgene_sgmac_reset(p);
/* Enable auto-negotiation */
- xgene_mii_phy_write(p, INT_PHY_ADDR, SGMII_CONTROL_ADDR >> 2, 0x1000);
+ xgene_mii_phy_write(p, INT_PHY_ADDR, SGMII_TBI_CONTROL_ADDR >> 2,
+ 0x8000);
+ xgene_mii_phy_write(p, INT_PHY_ADDR, SGMII_CONTROL_ADDR >> 2, 0x9000);
xgene_mii_phy_write(p, INT_PHY_ADDR, SGMII_TBI_CONTROL_ADDR >> 2, 0);
while (loop--) {
@@ -256,16 +343,14 @@ static void xgene_sgmac_init(struct xgene_enet_pdata *p)
if (!(data & AUTO_NEG_COMPLETE) || !(data & LINK_STATUS))
netdev_err(p->ndev, "Auto-negotiation failed\n");
- data = xgene_enet_rd_mac(p, MAC_CONFIG_2_ADDR);
- ENET_INTERFACE_MODE2_SET(&data, 2);
- xgene_enet_wr_mac(p, MAC_CONFIG_2_ADDR, data | FULL_DUPLEX2);
- xgene_enet_wr_mac(p, INTERFACE_CONTROL_ADDR, ENET_GHD_MODE);
+ xgene_sgmac_set_speed(p);
if (p->enet_id == XGENE_ENET1) {
enet_spare_cfg_reg = ENET_SPARE_CFG_REG_ADDR;
rsif_config_reg = RSIF_CONFIG_REG_ADDR;
cfg_bypass_reg = CFG_BYPASS_ADDR;
rx_dv_gate_reg = SG_RX_DV_GATE_REG_0_ADDR;
+ offset = p->port_id * 4;
At least for me it is not clear where the "4" comes from. Maybe you can
add a comment or declare a constant for that.
Anyway this looks like it fixes the driver for xgene2-sgnet and therefor
should go into a separate patch, or am I wrong?
Regards,
Matthias
