On 08/10/2017 08:13 PM, Yuiko.Oshino@microchip.com wrote:
Hi Marek,
Hi,
-----Original Message----- From: Marek Vasut [mailto:marex@denx.de] Sent: Wednesday, August 9, 2017 1:12 PM To: Yuiko Oshino - C18177; u-boot@lists.denx.de Cc: Joe Hershberger Subject: Re: [PATCH v3 1/2] usb: net: Add support for Microchip LAN75xx and LAN78xx
On 08/09/2017 06:25 PM, yuiko.oshino@microchip.com wrote:
From: Yuiko Oshino yuiko.oshino@microchip.com
Series-Changes: 3
FYI, this will end in the commit message when applied, remove it or move it below the --- . Also commit message is missing.
I did my best to follow the patman instructions and I added "commit-notes:" tag, but I guess it wasn't good enough. Should I always manually edit the patch before sending email in the patman? Also, when I am ready to update this patch again, should I do a series or just this patch? How can I update the [PATCH v] number? If just his patch, then will it be [PATCH v4]?
TBH, I dunno, I don't use patman :)
- All #ifdef CONFIG_DM_ETH and #endif are removed.
- The lan7x_eth_recv() is modifed to correctly support the Driver Model and returns packet_en.
- Add mii_resolve_flowctrl_fdx() patch in the series.
Series-Changes: 2
- The wait_for_bit functions copy the real one.
- Uses phylib
- Unnecessary variables are removed
- All return values are checked
- Uses mii_resolve_flowctrl_fdx() from linux/mii.h
Signed-off-by: Yuiko Oshino yuiko.oshino@microchip.com
Add support for Microchip LAN7500, LAN7800 and LAN7850, USB to 10/100/1000 Ethernet Controllers.
drivers/usb/Kconfig | 2 + drivers/usb/eth/Kconfig | 17 ++ drivers/usb/eth/Makefile | 2 + drivers/usb/eth/lan75xx.c | 318 +++++++++++++++++++++++++++++ drivers/usb/eth/lan78xx.c | 477
++++++++++++++++++++++++++++++++++++++++++++
drivers/usb/eth/lan7x.c | 495
++++++++++++++++++++++++++++++++++++++++++++++
drivers/usb/eth/lan7x.h | 230 +++++++++++++++++++++ 7 files changed, 1541 insertions(+) create mode 100644 drivers/usb/eth/Kconfig create mode 100644 drivers/usb/eth/lan75xx.c create mode 100644 drivers/usb/eth/lan78xx.c create mode 100644 drivers/usb/eth/lan7x.c create mode 100644 drivers/usb/eth/lan7x.h
diff --git a/drivers/usb/Kconfig b/drivers/usb/Kconfig index da3ec2f..62126aa 100644 --- a/drivers/usb/Kconfig +++ b/drivers/usb/Kconfig @@ -94,4 +94,6 @@ endif
source "drivers/usb/gadget/Kconfig"
+source "drivers/usb/eth/Kconfig"
endif diff --git a/drivers/usb/eth/Kconfig b/drivers/usb/eth/Kconfig new file mode 100644 index 0000000..14cfa26 --- /dev/null +++ b/drivers/usb/eth/Kconfig @@ -0,0 +1,17 @@ +comment "USB to Ethernet Controller Drivers"
+config USB_ETHER_LAN75XX
- bool "Microchip LAN75XX support"
- ---help---
Say Y here if you would like to support Microchip LAN75XX Hi-SpeedUSB 2.0 to 10/100/1000 Gigabit Ethernet controller.Supports 10Base-T/ 100Base-TX/1000Base-T.This driver supports the internal PHY.+config USB_ETHER_LAN78XX
- bool "Microchip LAN78XX support"
- ---help---
Say Y here if you would like to support Microchip LAN78XX USB 3.1Gen 1 to 10/100/1000 Gigabit Ethernet controller.Supports 10Base-T/ 100Base-TX/1000Base-T.This driver supports the internal PHY.diff --git a/drivers/usb/eth/Makefile b/drivers/usb/eth/Makefile index 4c44efc..4b935a3 100644 --- a/drivers/usb/eth/Makefile +++ b/drivers/usb/eth/Makefile @@ -9,4 +9,6 @@ obj-$(CONFIG_USB_ETHER_ASIX) += asix.o obj-$(CONFIG_USB_ETHER_ASIX88179) += asix88179.o obj-$(CONFIG_USB_ETHER_MCS7830) += mcs7830.o obj-$(CONFIG_USB_ETHER_SMSC95XX) += smsc95xx.o +obj-$(CONFIG_USB_ETHER_LAN75XX) += lan7x.o lan75xx.o +obj-$(CONFIG_USB_ETHER_LAN78XX) += lan7x.o lan78xx.o obj-$(CONFIG_USB_ETHER_RTL8152) += r8152.o r8152_fw.o diff --git a/drivers/usb/eth/lan75xx.c b/drivers/usb/eth/lan75xx.c new file mode 100644 index 0000000..a3c1411 --- /dev/null +++ b/drivers/usb/eth/lan75xx.c @@ -0,0 +1,318 @@ +/*
- Copyright (c) 2017 Microchip Technology Inc. All rights reserved.
- SPDX-License-Identifier: GPL-2.0+
- */
+#include <dm.h> +#include <usb.h> +#include <linux/mii.h> +#include "usb_ether.h" +#include "lan7x.h"
+/* LAN75xx specific register/bit defines */ +#define LAN75XX_HW_CFG_BIR BIT(7)
+#define LAN75XX_BURST_CAP 0x034
+#define LAN75XX_BULK_IN_DLY 0x03C
+#define LAN75XX_RFE_CTL 0x060
+#define LAN75XX_FCT_RX_CTL 0x090
+#define LAN75XX_FCT_TX_CTL 0x094
+#define LAN75XX_FCT_RX_FIFO_END 0x098
+#define LAN75XX_FCT_TX_FIFO_END 0x09C
+#define LAN75XX_FCT_FLOW 0x0A0
+/* MAC ADDRESS PERFECT FILTER For LAN75xx */ +#define LAN75XX_ADDR_FILTX 0x300 +#define LAN75XX_ADDR_FILTX_FB_VALID BIT(31)
+/*
- Lan75xx infrastructure commands
- */
+static int lan75xx_phy_gig_workaround(struct usb_device *udev,
struct ueth_data *dev)+{
- int ret = 0;
- /* Only internal phy */
- /* Set the phy in Gig loopback */
- lan7x_mdio_write(udev, dev->phy_id, MII_BMCR,
(BMCR_LOOPBACK | BMCR_SPEED1000));- /* Wait for the link up */
- ret = lan7x_mdio_wait_for_bit(udev, "BMSR_LSTATUS",
dev->phy_id, MII_BMSR, BMSR_LSTATUS,true, PHY_CONNECT_TIMEOUT_MS, 1);- if (ret)
return ret;- /* phy reset */
- return lan7x_pmt_phy_reset(udev, dev); }
+static int lan75xx_update_flowcontrol(struct usb_device *udev,
struct ueth_data *dev)+{
- uint32_t flow = 0, fct_flow = 0;
- int ret;
- ret = lan7x_update_flowcontrol(udev, dev, &flow, &fct_flow);
- if (ret)
return ret;- ret = lan7x_write_reg(udev, LAN75XX_FCT_FLOW, fct_flow);
- if (ret)
return ret;- return lan7x_write_reg(udev, FLOW, flow); }
+static int lan75xx_read_mac(unsigned char *enetaddr,
struct usb_device *udev)+{
- /*
* Refer to the doc/README.enetaddr and doc/README.usb for* the U-Boot MAC address policy*/- return lan7x_read_eeprom_mac(enetaddr, udev);
Is this function needed ?
I wanted to maintain the same function name convention among our drivers. Therefore the wrapper for this device. But I can remove it.
Please do.
+}
+static int lan75xx_set_receive_filter(struct usb_device *udev) {
- /* No multicast in u-boot */
- return lan7x_write_reg(udev, LAN75XX_RFE_CTL,
RFE_CTL_BCAST_EN | RFE_CTL_DA_PERFECT); }+/* starts the TX path */ +static void lan75xx_start_tx_path(struct usb_device *udev) {
- /* Enable Tx at MAC */
- lan7x_write_reg(udev, MAC_TX, MAC_TX_TXEN);
- /* Enable Tx at SCSRs */
- lan7x_write_reg(udev, LAN75XX_FCT_TX_CTL, FCT_TX_CTL_EN); }
+/* Starts the Receive path */ +static void lan75xx_start_rx_path(struct usb_device *udev) {
- /* Enable Rx at MAC */
- lan7x_write_reg(udev, MAC_RX,
LAN7X_MAC_RX_MAX_SIZE_DEFAULT |MAC_RX_FCS_STRIP | MAC_RX_RXEN);- /* Enable Rx at SCSRs */
- lan7x_write_reg(udev, LAN75XX_FCT_RX_CTL, FCT_RX_CTL_EN); }
+static int lan75xx_basic_reset(struct usb_device *udev,
struct ueth_data *dev,struct lan7x_private *priv) {- int ret;
- u32 val;
- ret = lan7x_basic_reset(udev, dev);
- if (ret)
return ret;- /* Keep the chip ID */
- ret = lan7x_read_reg(udev, ID_REV, &val);
- if (ret)
return ret;- debug("LAN75xx ID_REV = 0x%08x\n", val);
Some sort of USB bus ID would be useful in the debug message to identify the chip in case multiple are present.
I can add this to the lan7x_basic_reset() debug("USB devnum %d portnr %d\n", udev->devnum, udev->portnr); This shows device number and port number. Is this ok?
Yes, thanks.