Skip to contentSkip to conversation

Xilinx LL-TEMAC driver

U-Boot New monstr at monstr.eu 4 participants 18 messages Jan 5, 2009 – Apr 6, 2009

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

Hi Ben and Wolfgang,

first patch just sort labels in drivers/net/Makefile I think this patch could be applied directly to Wolfgang tree.

The second patch is Xilinx LL Temac driver.
This driver was sent to mailing list some month ago by Yoshio Kashiwagi. I clean the driver and add support for Fifo mode.

This driver is tested on some Microblaze boards.

There is one part which should be rework in future. (FIXME comment) It is about PHY lib which we discussed in past.
I believe that this part won't be a problem.

Thanks for your comments,
Michal

#1 lore raw

PATCH 1/2 net: Sort Makefile labels

From: Michal Simek <[email protected]>

---
 drivers/net/Makefile |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)
drivers/net/Makefile +2 -2
index 631336a..128dc11 100644
@@ -26,6 +26,7 @@ include $(TOPDIR)/config.mk
 LIB	:= $(obj)libnet.a
 
 COBJS-$(CONFIG_DRIVER_3C589) += 3c589.o
+COBJS-$(CONFIG_PPC4xx_EMAC) += 4xx_enet.o
 COBJS-$(CONFIG_DRIVER_AX88180) += ax88180.o
 COBJS-$(CONFIG_BCM570x) += bcm570x.o bcm570x_autoneg.o 5701rls.o
 COBJS-$(CONFIG_BFIN_MAC) += bfin_mac.o
@@ -54,11 +55,11 @@ COBJS-$(CONFIG_NS8382X) += ns8382x.o
 COBJS-$(CONFIG_DRIVER_NS9750_ETHERNET) += ns9750_eth.o
 COBJS-$(CONFIG_PCNET) += pcnet.o
 COBJS-$(CONFIG_PLB2800_ETHER) += plb2800_eth.o
-COBJS-$(CONFIG_PPC4xx_EMAC) += 4xx_enet.o
 COBJS-$(CONFIG_DRIVER_RTL8019) += rtl8019.o
 COBJS-$(CONFIG_RTL8139) += rtl8139.o
 COBJS-$(CONFIG_RTL8169) += rtl8169.o
 COBJS-$(CONFIG_DRIVER_S3C4510_ETH) += s3c4510b_eth.o
+COBJS-$(CONFIG_SH_ETHER) += sh_eth.o
 COBJS-$(CONFIG_DRIVER_SMC91111) += smc91111.o
 COBJS-$(CONFIG_DRIVER_SMC911X) += smc911x.o
 COBJS-$(CONFIG_TIGON3) += tigon3.o bcm570x_autoneg.o 5701rls.o
@@ -68,7 +69,6 @@ COBJS-$(CONFIG_ULI526X) += uli526x.o
 COBJS-$(CONFIG_VSC7385_ENET) += vsc7385.o
 COBJS-$(CONFIG_XILINX_EMAC) += xilinx_emac.o
 COBJS-$(CONFIG_XILINX_EMACLITE) += xilinx_emaclite.o
-COBJS-$(CONFIG_SH_ETHER) += sh_eth.o
 
 COBJS	:= $(COBJS-y)
 SRCS	:= $(COBJS:.o=.c)
Signature · 1 line
1.5.5.1
#2 lore raw

PATCH 2/2 net: Add Xilinx LL Temac driver

From: Michal Simek <[email protected]>

---
 drivers/net/Makefile          |    1 +
 drivers/net/xilinx_ll_temac.c |  543 +++++++++++++++++++++++++++++++++++++++++
 2 files changed, 544 insertions(+), 0 deletions(-)
 create mode 100644 drivers/net/xilinx_ll_temac.c
drivers/net/Makefile +1 -0
index 128dc11..ed64787 100644
@@ -69,6 +69,7 @@ COBJS-$(CONFIG_ULI526X) += uli526x.o
 COBJS-$(CONFIG_VSC7385_ENET) += vsc7385.o
 COBJS-$(CONFIG_XILINX_EMAC) += xilinx_emac.o
 COBJS-$(CONFIG_XILINX_EMACLITE) += xilinx_emaclite.o
+COBJS-$(CONFIG_XILINX_LL_TEMAC) += xilinx_ll_temac.o
 
 COBJS	:= $(COBJS-y)
 SRCS	:= $(COBJS:.o=.c)
drivers/net/xilinx_ll_temac.c new file +543 -0

543 lines not on this page · raw message

Signature · 1 line
1.5.5.1
#3 lore raw

PATCH 2/2 net: Add Xilinx LL Temac driver

Hi Ben,

Could you check and pull ll temac driver to your net repo.
I prepared net branch in my microblaze repo. (It is based on my master branch which Wolfgang merged but it hasn't in his branch yet).

Thanks,
Michal

The following changes since commit e7f325be9edeb84bb457301776bbac1f7257dafc:
Michal Simek (1):
microblaze: Use cache functions (especially cache status)

are available in the git repository at:

git://git.denx.de/u-boot-microblaze.git net

Michal Simek (1):
net: Add Xilinx LL Temac driver

 drivers/net/Makefile          |    1 +
 drivers/net/xilinx_ll_temac.c |  549 +++++++++++++++++++++++++++++++++++++++++
 2 files changed, 550 insertions(+), 0 deletions(-)
 create mode 100644 drivers/net/xilinx_ll_temac.c
Quoted 571 lines

From: Michal Simek <[email protected]>

---
 drivers/net/Makefile          |    1 +
 drivers/net/xilinx_ll_temac.c |  543 +++++++++++++++++++++++++++++++++++++++++
 2 files changed, 544 insertions(+), 0 deletions(-)
 create mode 100644 drivers/net/xilinx_ll_temac.c
drivers/net/Makefile +1 -0
index 128dc11..ed64787 100644
@@ -69,6 +69,7 @@ COBJS-$(CONFIG_ULI526X) += uli526x.o
 COBJS-$(CONFIG_VSC7385_ENET) += vsc7385.o
 COBJS-$(CONFIG_XILINX_EMAC) += xilinx_emac.o
 COBJS-$(CONFIG_XILINX_EMACLITE) += xilinx_emaclite.o
+COBJS-$(CONFIG_XILINX_LL_TEMAC) += xilinx_ll_temac.o
 
 COBJS	:= $(COBJS-y)
 SRCS	:= $(COBJS:.o=.c)
drivers/net/xilinx_ll_temac.c new file +543 -0
new file mode 100644
index 0000000..d1e63e2
@@ -0,0 +1,543 @@
+/*
+ *
+ * Xilinx xps_ll_temac ethernet driver for u-boot
+ *
+ * Author: Yoshio Kashiwagi kashiwagi at co-nss.co.jp
+ *
+ * Copyright (C) 2008 Nissin Systems Co.,Ltd.
+ * March 2008 created
+ *
+ * Copyright (C) 2008 - 2009 Michal Simek <[email protected]>
+ * June 2008 Microblaze optimalization, FIFO mode support
+ *
+ * This program is free software; you can redistribute it 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.
+ *
+ */
+
+#include <config.h>
+#include <common.h>
+#include <net.h>
+#include <malloc.h>
+#include <asm/processor.h>
+#include <asm/io.h>
+
+#ifdef XILINX_LLTEMAC_FIFO_BASEADDR
+# define FIFO_MODE	1
+#elif XILINX_LLTEMAC_SDMA_CTRL_BASEADDR
+# define SDMA_MODE	1
+#else
+# error Unsupported mode
+#endif
+
+#undef ETH_HALTING
+
+#ifdef SDMA_MODE
+/* XPS_LL_TEMAC SDMA registers definition */
+# define TX_NXTDESC_PTR		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x00)
+# define TX_CURBUF_ADDR		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x04)
+# define TX_CURBUF_LENGTH	(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x08)
+# define TX_CURDESC_PTR		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x0c)
+# define TX_TAILDESC_PTR	(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x10)
+# define TX_CHNL_CTRL		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x14)
+# define TX_IRQ_REG		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x18)
+# define TX_CHNL_STS		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x1c)
+
+# define RX_NXTDESC_PTR		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x20)
+# define RX_CURBUF_ADDR		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x24)
+# define RX_CURBUF_LENGTH	(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x28)
+# define RX_CURDESC_PTR		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x2c)
+# define RX_TAILDESC_PTR	(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x30)
+# define RX_CHNL_CTRL		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x34)
+# define RX_IRQ_REG		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x38)
+# define RX_CHNL_STS		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x3c)
+
+# define DMA_CONTROL_REG	(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x40)
+#endif
+
+/* XPS_LL_TEMAC direct registers definition */
+#define TEMAC_RAF0		(XILINX_LLTEMAC_BASEADDR + 0x00)
+#define TEMAC_TPF0		(XILINX_LLTEMAC_BASEADDR + 0x04)
+#define TEMAC_IFGP0		(XILINX_LLTEMAC_BASEADDR + 0x08)
+#define TEMAC_IS0		(XILINX_LLTEMAC_BASEADDR + 0x0c)
+#define TEMAC_IP0		(XILINX_LLTEMAC_BASEADDR + 0x10)
+#define TEMAC_IE0		(XILINX_LLTEMAC_BASEADDR + 0x14)
+
+#define TEMAC_MSW0		(XILINX_LLTEMAC_BASEADDR + 0x20)
+#define TEMAC_LSW0		(XILINX_LLTEMAC_BASEADDR + 0x24)
+#define TEMAC_CTL0		(XILINX_LLTEMAC_BASEADDR + 0x28)
+#define TEMAC_RDY0		(XILINX_LLTEMAC_BASEADDR + 0x2c)
+
+#define XTE_RSE_MIIM_RR_MASK	0x0002
+#define XTE_RSE_MIIM_WR_MASK	0x0004
+#define XTE_RSE_CFG_RR_MASK	0x0020
+#define XTE_RSE_CFG_WR_MASK	0x0040
+
+/* XPS_LL_TEMAC indirect registers offset definition */
+
+#define RCW0	0x200
+#define RCW1	0x240
+#define TC	0x280
+#define FCC	0x2c0
+#define EMMC	0x300
+#define PHYC	0x320
+#define MC	0x340
+#define UAW0	0x380
+#define UAW1	0x384
+#define MAW0	0x388
+#define MAW1	0x38c
+#define AFM	0x390
+#define TIS	0x3a0
+#define TIE	0x3a4
+#define MIIMWD	0x3b0
+#define MIIMAI	0x3b4
+
+#define CNTLREG_WRITE_ENABLE_MASK	0x8000
+#define CNTLREG_EMAC1SEL_MASK		0x0400
+#define CNTLREG_ADDRESSCODE_MASK	0x03ff
+
+#define MDIO_ENABLE_MASK	0x40
+#define MDIO_CLOCK_DIV_MASK	0x3F
+#define MDIO_CLOCK_DIV_100MHz	0x28
+
+#define ETHER_MTU		1520
+
+#ifdef SDMA_MODE
+/* CDMAC descriptor status bit definitions */
+# define BDSTAT_ERROR_MASK		0x80
+# define BDSTAT_INT_ON_END_MASK		0x40
+# define BDSTAT_STOP_ON_END_MASK	0x20
+# define BDSTAT_COMPLETED_MASK		0x10
+# define BDSTAT_SOP_MASK		0x08
+# define BDSTAT_EOP_MASK		0x04
+# define BDSTAT_CHANBUSY_MASK		0x02
+# define BDSTAT_CHANRESET_MASK		0x01
+
+/* SDMA Buffer Descriptor */
+
+typedef struct cdmac_bd_t {
+	struct cdmac_bd_t *next_p;
+	unsigned char *phys_buf_p;
+	unsigned long buf_len;
+	unsigned char stat;
+	unsigned char app1_1;
+	unsigned short app1_2;
+	unsigned long app2;
+	unsigned long app3;
+	unsigned long app4;
+	unsigned long app5;
+} cdmac_bd __attribute((aligned(32))) ;
+
+static cdmac_bd	tx_bd;
+static cdmac_bd	rx_bd;
+#endif
+
+#ifdef FIFO_MODE
+typedef struct ll_fifo_s {
+	int isr; /* Interrupt Status Register 0x0 */
+	int ier; /* Interrupt Enable Register 0x4 */
+	int tdfr; /* Transmit data FIFO reset 0x8 */
+	int tdfv; /* Transmit data FIFO Vacancy 0xC */
+	int tdfd; /* Transmit data FIFO 32bit wide data write port 0x10 */
+	int tlf; /* Write Transmit Length FIFO 0x14 */
+	int rdfr; /* Read Receive data FIFO reset 0x18 */
+	int rdfo; /* Receive data FIFO Occupancy 0x1C */
+	int rdfd; /* Read Receive data FIFO 32bit wide data read port 0x20 */
+	int rlf; /* Read Receive Length FIFO 0x24 */
+	int llr; /* Read LocalLink reset 0x28 */
+} ll_fifo_s;
+
+ll_fifo_s *ll_fifo = (ll_fifo_s *) (XILINX_LLTEMAC_FIFO_BASEADDR);
+#endif
+
+static unsigned char tx_buffer[ETHER_MTU] __attribute((aligned(32)));
+static unsigned char rx_buffer[ETHER_MTU] __attribute((aligned(32)));
+
+struct xps_ll_temac_private {
+	int idx;
+	unsigned char dev_addr[6];
+};
+
+#ifdef DEBUG
+/* undirect hostif write to ll_temac */
+static void xps_ll_temac_hostif_set(int emac, int phy_addr,
+					int reg_addr, int phy_data)
+{
+	out_be32((u32 *)TEMAC_LSW0, phy_data);
+	out_be32((u32 *)TEMAC_CTL0, CNTLREG_WRITE_ENABLE_MASK | MIIMWD);
+	out_be32((u32 *)TEMAC_LSW0, (phy_addr << 5) | (reg_addr));
+	out_be32((u32 *)TEMAC_CTL0, \
+			CNTLREG_WRITE_ENABLE_MASK | MIIMAI | (emac << 10));
+	while(! (in_be32((u32 *)TEMAC_RDY0) & XTE_RSE_MIIM_WR_MASK));
+}
+#endif
+
+/* undirect hostif read from ll_temac */
+static unsigned int xps_ll_temac_hostif_get(int emac, int phy_addr, int reg_addr)
+{
+	out_be32((u32 *)TEMAC_LSW0, (phy_addr << 5) | (reg_addr));
+	out_be32((u32 *)TEMAC_CTL0, MIIMAI | (emac << 10));
+	while(! (in_be32((u32 *)TEMAC_RDY0) & XTE_RSE_MIIM_RR_MASK));
+	return in_be32((u32 *)TEMAC_LSW0);
+}
+
+/* undirect write to ll_temac */
+static void xps_ll_temac_indirect_set(int emac, int reg_offset, int reg_data)
+{
+	out_be32((u32 *)TEMAC_LSW0, reg_data);
+	out_be32((u32 *)TEMAC_CTL0, \
+			CNTLREG_WRITE_ENABLE_MASK | (emac << 10) | reg_offset);
+	while(! (in_be32((u32 *)TEMAC_RDY0) & XTE_RSE_CFG_WR_MASK));
+}
+
+/* undirect read from ll_temac */
+int xps_ll_temac_indirect_get(int emac, int reg_offset)
+{
+	out_be32((u32 *)TEMAC_CTL0, (emac << 10) | reg_offset);
+	while(! (in_be32((u32 *)TEMAC_RDY0) & XTE_RSE_CFG_RR_MASK));
+	return in_be32((u32 *)TEMAC_LSW0);
+}
+
+#ifdef DEBUG
+/* read from phy */
+static void read_phy_reg (int phy_addr)
+{
+	int j, result;
+	printf("phy%d ",phy_addr);
+	for ( j = 0; j < 32; j++) {
+		result = xps_ll_temac_hostif_get(0, phy_addr, j);
+		printf("%d: 0x%x ", j, result);
+	}
+	puts("\n");
+}
+#endif
+
+static int phy_addr = -1;
+static int link = 0;
+
+/* setting ll_temac and phy to proper setting */
+static int xps_ll_temac_phy_ctrl(void)
+{
+	int i;
+	unsigned int result;
+	unsigned retries = 10;
+
+	if(link == 1)
+		return 1; /* link is setup */
+
+	/* wait for link up */
+	while (retries-- &&
+		((xps_ll_temac_hostif_get(0, phy_addr, 1) & 0x24) == 0x24))
+		;
+
+	if(phy_addr == -1) {
+		for(i = 31; i >= 0; i--) {
+			result = xps_ll_temac_hostif_get(0, i, 1);
+			if((result & 0x0ffff) != 0x0ffff) {
+#ifdef DEBUG
+				printf ("phy %x result %x\n", i, result);
+#endif
+				phy_addr = i;
+				break;
+			}
+		}
+	}
+
+	/* get PHY id */
+	i = (xps_ll_temac_hostif_get(0, phy_addr, 2) << 16) | \
+		xps_ll_temac_hostif_get(0, phy_addr, 3);
+#ifdef DEBUG
+	printf("LL_TEMAC: Phy ID 0x%x\n", i);
+#endif
+
+#ifdef DEBUG
+	xps_ll_temac_hostif_set(0, 0, 0, 0x8000); /* phy reset */
+#endif
+	/* FIXME this part will be replaced by PHY lib */
+	/* s3e boards */
+	if (i == 0x7c0a3) {
+		xps_ll_temac_indirect_set(0, EMMC, 0x40000000); /* 100BASE-T/FD */
+		link = 1;
+		return 1;
+	}
+
+	/* Marwell 88e1111 id - ml50x */
+	if (i == 0x1410cc2) {
+		result = xps_ll_temac_hostif_get(0, phy_addr, 5);
+		if((result & 0x8000) == 0x8000) {
+			xps_ll_temac_indirect_set(0, EMMC, 0x80000000);
+			printf("1000BASE-T/FD\n");
+			link = 1;
+		} else if((result & 0x4000) == 0x4000) {
+			xps_ll_temac_indirect_set(0, EMMC, 0x40000000);
+			printf("100BASE-T/FD\n");
+			link = 1;
+		} else {
+			printf("Unsupported mode\n");
+			link = 0;
+		}
+		return 1;
+	}
+	return 0;
+}
+
+#ifdef SDMA_MODE
+/* bd init */
+static void xps_ll_temac_bd_init(void)
+{
+	memset((void *)&tx_bd, 0, sizeof(cdmac_bd));
+	memset((void *)&rx_bd, 0, sizeof(cdmac_bd));
+
+	rx_bd.phys_buf_p = &rx_buffer[0];
+
+	rx_bd.next_p = &rx_bd;
+	rx_bd.buf_len = ETHER_MTU;
+	flush_cache((u32)&rx_bd, sizeof(cdmac_bd));
+
+	out_be32((u32 *)RX_CURDESC_PTR, (u32)&rx_bd);
+	out_be32((u32 *)RX_TAILDESC_PTR, (u32)&rx_bd);
+	out_be32((u32 *)RX_NXTDESC_PTR, (u32)&rx_bd); /* setup first fd */
+
+	tx_bd.phys_buf_p = &tx_buffer[0];
+	tx_bd.next_p = &tx_bd;
+
+	flush_cache((u32)&tx_bd, sizeof(cdmac_bd));
+	out_be32((u32 *)TX_CURDESC_PTR, (u32)&tx_bd);
+}
+#endif
+
+#ifdef SDMA_MODE
+static int xps_ll_temac_send_sdma(unsigned char *buffer, int length)
+{
+	if( xps_ll_temac_phy_ctrl() == 0)
+		return 0;
+
+	memcpy (tx_buffer, buffer, length);
+	flush_cache ((u32)tx_buffer, length);
+
+	tx_bd.stat = BDSTAT_SOP_MASK | BDSTAT_EOP_MASK | BDSTAT_STOP_ON_END_MASK;
+	tx_bd.buf_len = length;
+	flush_cache ((u32)&tx_bd, sizeof(cdmac_bd));
+
+	out_be32((u32 *)TX_CURDESC_PTR, (u32)&tx_bd);
+	out_be32((u32 *)TX_TAILDESC_PTR, (u32)&tx_bd); /* DMA start */
+
+	do {
+		flush_cache ((u32)&tx_bd, sizeof(cdmac_bd));
+	} while(!(((volatile int)tx_bd.stat) & BDSTAT_COMPLETED_MASK));
+
+	return length;
+}
+
+
+static int xps_ll_temac_recv_sdma(void)
+{
+	int length;
+
+	flush_cache ((u32)&rx_bd, sizeof(cdmac_bd));
+
+	if(!(rx_bd.stat & BDSTAT_COMPLETED_MASK)) {
+		return 0;
+	}
+
+	length = rx_bd.app5;
+	flush_cache ((u32)rx_bd.phys_buf_p, length);
+
+	rx_bd.buf_len = ETHER_MTU;
+	rx_bd.stat = 0;
+	rx_bd.app5 = 0;
+
+	flush_cache ((u32)&rx_bd, sizeof(cdmac_bd));
+	out_be32((u32 *)RX_TAILDESC_PTR, (u32)&rx_bd);
+
+	if(length > 0) {
+		NetReceive(rx_bd.phys_buf_p, length);
+	}
+
+	return length;
+}
+#endif
+
+
+#ifdef FIFO_MODE
+void debugll(int count)
+{
+	printf ("%d fifo isr 0x%08x, fifo_ier 0x%08x, fifo_rdfr 0x%08x, fifo_rdfo 0x%08x fifo_rlr 0x%08x\n",count, ll_fifo->isr, \
+	ll_fifo->ier, ll_fifo->rdfr, ll_fifo->rdfo, ll_fifo->rlf);
+
+}
+
+static int xps_ll_temac_send_fifo(unsigned char *buffer, int length)
+{
+	u32 *buf = buffer;
+	u32 len, i, val;
+
+	len = (length / 4) + 1;
+
+	for (i = 0; i < len; i++) {
+		val = *buf++;
+		ll_fifo->tdfd = val;
+	}
+
+	ll_fifo->tlf = length;
+
+	return length;
+}
+
+static int xps_ll_temac_recv_fifo()
+{
+	int len, len2, i, val;
+	int *buf;
+	buf = &rx_buffer;
+
+	if (ll_fifo->isr & 0x04000000 ) {
+		ll_fifo->isr = 0xffffffff; /* reset isr */
+	
+		/* while (ll_fifo->isr); */
+		len = ll_fifo->rlf & 0x7FF;
+		len2 = (len / 4) + 1;
+	
+		for (i = 0; i < len2; i++) {
+			val = ll_fifo->rdfd;
+			*buf++ = val ;
+		}
+	
+		/* debugll(1); */
+		NetReceive (&rx_buffer, len);
+	}
+	return 0;
+}
+#endif
+
+/* setup mac addr */
+static int xps_ll_temac_addr_setup(struct xps_ll_temac_private * lp)
+{
+	char * env_p;
+	char * end;
+	int i, val;
+
+	env_p = getenv("ethaddr");
+	if (env_p == NULL) {
+		printf("cannot get enviroment for \"ethaddr\".\n");
+		return -1;
+	}
+
+	for (i = 0; i < 6; i++) {
+		lp->dev_addr[i] = env_p ? simple_strtoul(env_p, &end, 16) : 0;
+		if (env_p) env_p = (*end) ? end + 1 : end;
+	}
+
+	/* set up unicast MAC address filter */
+	val = ((lp->dev_addr[3] << 24) | (lp->dev_addr[2] << 16) |
+		(lp->dev_addr[1] << 8) | (lp->dev_addr[0] ));
+	xps_ll_temac_indirect_set(0, UAW0, val);
+	val = (lp->dev_addr[5] << 8) | lp->dev_addr[4] ;
+	xps_ll_temac_indirect_set(0, UAW1, val);
+
+	return 0;
+}
+
+static int xps_ll_temac_init(struct eth_device *dev, bd_t *bis)
+{
+	struct xps_ll_temac_private *lp = (struct xps_ll_temac_private *)dev->priv;
+
+#ifdef SDMA_MODE
+	xps_ll_temac_bd_init();
+#endif
+#ifdef FIFO_MODE
+	ll_fifo->tdfr = 0x000000a5; /* set fifo lenght */
+	ll_fifo->rdfr = 0x000000a5;
+
+	/* ll_fifo->isr = 0x0; */
+	/* ll_fifo->ier = 0x0; */
+#endif
+	xps_ll_temac_indirect_set(0, MC, MDIO_ENABLE_MASK | MDIO_CLOCK_DIV_100MHz);
+
+	xps_ll_temac_addr_setup(lp);
+	xps_ll_temac_indirect_set(0, AFM, 0x00000000); /* Promiscuos mode disable */
+	xps_ll_temac_indirect_set(0, RCW1, 0x10000000); /* Enable Receiver */
+	xps_ll_temac_indirect_set(0, TC, 0x10000000); /* Enable Transmitter */
+
+}
+
+#ifdef ETH_HALTING
+static int xps_ll_temac_halt(void)
+{
+	xps_ll_temac_indirect_set(0, RCW1, 0x00000000); /* Disable Receiver */
+	xps_ll_temac_indirect_set(0, TC, 0x00000000); /* Disable Transmitter */
+
+#ifdef SDMA_MODE
+	out_be32((u32 *)DMA_CONTROL_REG, 0x00000001);
+	while(in_be32((u32 *)DMA_CONTROL_REG) & 1);
+#endif
+#ifdef FIFO_MODE
+	/* reset fifos */
+#endif
+}
+#endif
+
+/* halt device */
+void eth_halt(void){
+	link = 0;
+#ifdef ETH_HALTING
+	xps_ll_temac_halt();
+#endif
+}
+
+int eth_init(bd_t *bis)
+{
+	static int first = 1;
+	struct eth_device *dev;
+	struct xps_ll_temac_private *lp;
+#if DEBUG
+	int i;
+#endif
+
+	if(!first)
+		return 0;
+	first = 0;
+	dev = (struct eth_device *) calloc(1, sizeof(struct eth_device));
+	if (NULL == dev)
+		return 0;
+
+	lp = (struct xps_ll_temac_private *) calloc(1, sizeof(struct xps_ll_temac_private));
+	if (lp == NULL)
+		return 0;
+	dev->priv = lp;
+	sprintf(dev->name, "eth0");
+
+	xps_ll_temac_init(dev, bis);
+
+	printf("%s: Xilinx XPS LocalLink Tri-Mode Ether MAC #%d at 0x%08X.\n",
+		dev->name, 0, XILINX_LLTEMAC_BASEADDR);
+
+#if DEBUG
+	for(i = 0;i < 32;i++) {
+		read_phy_reg(i);
+	}
+#endif
+	xps_ll_temac_phy_ctrl();
+	return 1;
+}
+
+int eth_send(volatile void *packet, int length)
+{
+#ifdef SDMA_MODE
+	return xps_ll_temac_send_sdma((unsigned char *)packet, length);
+#endif
+#ifdef FIFO_MODE
+	return xps_ll_temac_send_fifo((unsigned char *)packet, length);
+#endif
+}
+
+int eth_rx(void)
+{
+#ifdef SDMA_MODE
+	return xps_ll_temac_recv_sdma();
+#endif
+#ifdef FIFO_MODE
+	return xps_ll_temac_recv_fifo();
+#endif
+}
Signature · 2 lines
Michal Simek, Ing. (M.Eng)
w: www.monstr.eu p: +42-0-721842854
#4 lore raw

PATCH 2/2 net: Add Xilinx LL Temac driver

Hi Ben,

Quoted 3 lines

Hi Michal,

I'm really sorry for taking so long getting to this. Mostly little things:

it's ok you are busy guy. I removed uninteresting part of code. I added all your reported things to u-boot-microblaze.git net branch. Thanks for pull.

Quoted 9 lines
Quoted 9 lines
drivers/net/xilinx_ll_temac.c new file +9 -0
+#ifdef DEBUG
+/* read from phy */
+static void read_phy_reg (int phy_addr)
+{
+	int j, result;
+	printf("phy%d ",phy_addr);
+	for ( j = 0; j < 32; j++) {
+		result = xps_ll_temac_hostif_get(0, phy_addr, j);
+		printf("%d: 0x%x ", j, result);
Quoted 8 lines

drivers/net/xilinx_ll_temac.c line 216

Quoted 6 lines
drivers/net/xilinx_ll_temac.c new file +5 -0
+	}
+	puts("\n");
+}
+#endif
+

You only use this function in one place. Why not move this code to that place and use debug()

You are right that currently is this function use only once but that's for this code. When will be new board with different phy, it is easy to use this function in different place too. The second reason is that eth_init is long and added this function to it just cause that function will be longer than now. It is atomic function.

Quoted 17 lines
Quoted 17 lines
drivers/net/xilinx_ll_temac.c new file +17 -0
+static int phy_addr = -1;
+static int link = 0;
+
+/* setting ll_temac and phy to proper setting */
+static int xps_ll_temac_phy_ctrl(void)
+{
+	int i;
+	unsigned int result;
+	unsigned retries = 10;
+
+	if(link == 1)
+		return 1; /* link is setup */
+
+	/* wait for link up */
+	while (retries-- &&
+		((xps_ll_temac_hostif_get(0, phy_addr, 1) & 0x24) == 0x24))
+		;
Quoted 8 lines

drivers/net/xilinx_ll_temac.c line 241

Quoted 8 lines
drivers/net/xilinx_ll_temac.c new file +8 -0
+
+	if(phy_addr == -1) {
+		for(i = 31; i >= 0; i--) {
+			result = xps_ll_temac_hostif_get(0, i, 1);
+			if((result & 0x0ffff) != 0x0ffff) {
+#ifdef DEBUG
+				printf ("phy %x result %x\n", i, result);
+#endif

fixed debug here too.

Quoted 5 lines
Quoted 5 lines
drivers/net/xilinx_ll_temac.c new file +5 -0
+				phy_addr = i;
+				break;
+			}
+		}
+	}
Quoted 8 lines

drivers/net/xilinx_ll_temac.c line 252

Quoted 7 lines
drivers/net/xilinx_ll_temac.c new file +6 -0
+
+	/* get PHY id */
+	i = (xps_ll_temac_hostif_get(0, phy_addr, 2) << 16) | \
+		xps_ll_temac_hostif_get(0, phy_addr, 3);
+#ifdef DEBUG
+	printf("LL_TEMAC: Phy ID 0x%x\n", i);

Please use debug() instead

fixed.

Quoted 34 lines
Quoted 33 lines
drivers/net/xilinx_ll_temac.c new file +33 -0
+#endif
+
+#ifdef DEBUG
+	xps_ll_temac_hostif_set(0, 0, 0, 0x8000); /* phy reset */
+#endif
+	/* FIXME this part will be replaced by PHY lib */
+	/* s3e boards */
+	if (i == 0x7c0a3) {
+		xps_ll_temac_indirect_set(0, EMMC, 0x40000000); /* 100BASE-T/FD */
+		link = 1;
+		return 1;
+	}
+
+	/* Marwell 88e1111 id - ml50x */
+	if (i == 0x1410cc2) {
+		result = xps_ll_temac_hostif_get(0, phy_addr, 5);
+		if((result & 0x8000) == 0x8000) {
+			xps_ll_temac_indirect_set(0, EMMC, 0x80000000);
+			printf("1000BASE-T/FD\n");
+			link = 1;
+		} else if((result & 0x4000) == 0x4000) {
+			xps_ll_temac_indirect_set(0, EMMC, 0x40000000);
+			printf("100BASE-T/FD\n");
+			link = 1;
+		} else {
+			printf("Unsupported mode\n");
+			link = 0;
+		}
+		return 1;
+	}
+	return 0;
+}
+
Quoted 5 lines

drivers/net/xilinx_ll_temac.c

Quoted 5 lines
drivers/net/xilinx_ll_temac.c new file +5 -0
+
+#ifdef FIFO_MODE
+void debugll(int count)
+{
+	printf ("%d fifo isr 0x%08x, fifo_ier 0x%08x, fifo_rdfr 0x%08x, fifo_rdfo

0x%08x fifo_rlr 0x%08x\n",count, ll_fifo->isr, \

Quoted 2 lines
Quoted 1 line

This line's way too long.

Yes I know about. It was fixed in net branch.

Quoted 34 lines
Quoted 34 lines
drivers/net/xilinx_ll_temac.c new file +33 -0
+	ll_fifo->ier, ll_fifo->rdfr, ll_fifo->rdfo, ll_fifo->rlf);
+
+}
+
 
+
+#ifdef ETH_HALTING
+static int xps_ll_temac_halt(void)
+{
+	xps_ll_temac_indirect_set(0, RCW1, 0x00000000); /* Disable Receiver */
+	xps_ll_temac_indirect_set(0, TC, 0x00000000); /* Disable Transmitter */
+
+#ifdef SDMA_MODE
+	out_be32((u32 *)DMA_CONTROL_REG, 0x00000001);
+	while(in_be32((u32 *)DMA_CONTROL_REG) & 1);
+#endif
+#ifdef FIFO_MODE
+	/* reset fifos */
+#endif
+}
+#endif
+
+/* halt device */
+void eth_halt(void){
+	link = 0;
+#ifdef ETH_HALTING
+	xps_ll_temac_halt();
+#endif
+}
+
+int eth_init(bd_t *bis)
+{
+	static int first = 1;
+	struct eth_device *dev;
Quoted 8 lines

drivers/net/xilinx_ll_temac.c line 497

Quoted 6 lines
drivers/net/xilinx_ll_temac.c new file +5 -0
+	struct xps_ll_temac_private *lp;
+#if DEBUG
+	int i;
+#endif
+

I normally don't like this, but move the variable declaration to where it's used to avoid 2 #ifs

I don't like it too but ok.

Quoted 8 lines

drivers/net/xilinx_ll_temac.c

Quoted 8 lines
drivers/net/xilinx_ll_temac.c new file +8 -0
+	if(!first)
+		return 0;
+	first = 0;
+	dev = (struct eth_device *) calloc(1, sizeof(struct eth_device));
+	if (NULL == dev)
+		return 0;
+
+	lp = (struct xps_ll_temac_private *) calloc(1, sizeof(struct

xps_ll_temac_private));

Quoted 14 lines
Quoted 14 lines
drivers/net/xilinx_ll_temac.c new file +14 -0
+	if (lp == NULL)
+		return 0;
+	dev->priv = lp;
+	sprintf(dev->name, "eth0");
+
+	xps_ll_temac_init(dev, bis);
+
+	printf("%s: Xilinx XPS LocalLink Tri-Mode Ether MAC #%d at 0x%08X.\n",
+		dev->name, 0, XILINX_LLTEMAC_BASEADDR);
+
+#if DEBUG
+	for(i = 0;i < 32;i++) {
+		read_phy_reg(i);
+	}
Quoted 8 lines

drivers/net/xilinx_ll_temac.c line 524

Quoted 6 lines
drivers/net/xilinx_ll_temac.c new file +5 -0
+#endif
+	xps_ll_temac_phy_ctrl();
+	return 1;
+}
+

regards,
B en

Thanks,
Michal

Quoted 5 lines

_______________________________________________
U-Boot mailing list
U-Boot at lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot

#5 lore raw

PATCH 2/2 net: Add Xilinx LL Temac driver

Dear Michal,

Quoted 6 lines · In message <[email protected]> you wrote:
Quoted 2 lines

I'm really sorry for taking so long getting to this. Mostly little things:

it's ok you are busy guy. I removed uninteresting part of code. I added all your reported things to u-boot-microblaze.git net branch. Thanks for pull.

What exactly is the state of this today? It seems this did not hit mainline yet?

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
       There is enough for the need of everyone in this world,
       but not for the greed of everyone.     - Mahatma Gandhi
#6 lore raw

PATCH 2/2 net: Add Xilinx LL Temac driver

Hi Wolfgang and Ben,

Quoted 10 lines

Dear Michal,

Quoted 4 lines · In message <[email protected]> you wrote:
Quoted 1 line

I'm really sorry for taking so long getting to this. Mostly little things:

it's ok you are busy guy. I removed uninteresting part of code. I added all your reported things to u-boot-microblaze.git net branch. Thanks for pull.

What exactly is the state of this today? It seems this did not hit mainline yet?

Yes, this driver is not in mainline. Ben wanted to changed this driver to new network driver style. I tried to do it but I my toolchain had a problem with weak function.
I see two solution.
1. Add old style driver without new net multi interface
2. Add new style driver and microblaze community just comment one line in eth.c (this line int board_eth_init(bd_t *bis) __attribute((weak, alias("__def_eth_init")));)

Both solution are possible. If you agree with me, I'll send new version driver for comments.

And of course the best solution is update my toolchain but there we have some problems.

Thanks,
Michal

Quoted 5 lines

Best regards,

Wolfgang Denk

Signature · 2 lines
Michal Simek, Ing. (M.Eng)
w: www.monstr.eu p: +42-0-721842854
#7 lore raw

PATCH 2/2 net: Add Xilinx LL Temac driver

In reply to monstr at monstr.eu

Hi Michal,

I'm really sorry for taking so long getting to this. Mostly little things:

Quoted 235 lines · monstr at monstr.eu

From: Michal Simek <[email protected]>

---
 drivers/net/Makefile          |    1 +
 drivers/net/xilinx_ll_temac.c |  543 +++++++++++++++++++++++++++++++++++++++++
 2 files changed, 544 insertions(+), 0 deletions(-)
 create mode 100644 drivers/net/xilinx_ll_temac.c
drivers/net/Makefile +1 -0
index 128dc11..ed64787 100644
@@ -69,6 +69,7 @@ COBJS-$(CONFIG_ULI526X) += uli526x.o
 COBJS-$(CONFIG_VSC7385_ENET) += vsc7385.o
 COBJS-$(CONFIG_XILINX_EMAC) += xilinx_emac.o
 COBJS-$(CONFIG_XILINX_EMACLITE) += xilinx_emaclite.o
+COBJS-$(CONFIG_XILINX_LL_TEMAC) += xilinx_ll_temac.o
 
 COBJS	:= $(COBJS-y)
 SRCS	:= $(COBJS:.o=.c)
drivers/net/xilinx_ll_temac.c new file +206 -0
new file mode 100644
index 0000000..d1e63e2
@@ -0,0 +1,543 @@
+/*
+ *
+ * Xilinx xps_ll_temac ethernet driver for u-boot
+ *
+ * Author: Yoshio Kashiwagi kashiwagi at co-nss.co.jp
+ *
+ * Copyright (C) 2008 Nissin Systems Co.,Ltd.
+ * March 2008 created
+ *
+ * Copyright (C) 2008 - 2009 Michal Simek <[email protected]>
+ * June 2008 Microblaze optimalization, FIFO mode support
+ *
+ * This program is free software; you can redistribute it 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.
+ *
+ */
+
+#include <config.h>
+#include <common.h>
+#include <net.h>
+#include <malloc.h>
+#include <asm/processor.h>
+#include <asm/io.h>
+
+#ifdef XILINX_LLTEMAC_FIFO_BASEADDR
+# define FIFO_MODE	1
+#elif XILINX_LLTEMAC_SDMA_CTRL_BASEADDR
+# define SDMA_MODE	1
+#else
+# error Unsupported mode
+#endif
+
+#undef ETH_HALTING
+
+#ifdef SDMA_MODE
+/* XPS_LL_TEMAC SDMA registers definition */
+# define TX_NXTDESC_PTR		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x00)
+# define TX_CURBUF_ADDR		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x04)
+# define TX_CURBUF_LENGTH	(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x08)
+# define TX_CURDESC_PTR		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x0c)
+# define TX_TAILDESC_PTR	(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x10)
+# define TX_CHNL_CTRL		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x14)
+# define TX_IRQ_REG		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x18)
+# define TX_CHNL_STS		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x1c)
+
+# define RX_NXTDESC_PTR		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x20)
+# define RX_CURBUF_ADDR		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x24)
+# define RX_CURBUF_LENGTH	(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x28)
+# define RX_CURDESC_PTR		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x2c)
+# define RX_TAILDESC_PTR	(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x30)
+# define RX_CHNL_CTRL		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x34)
+# define RX_IRQ_REG		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x38)
+# define RX_CHNL_STS		(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x3c)
+
+# define DMA_CONTROL_REG	(XILINX_LLTEMAC_SDMA_CTRL_BASEADDR + 0x40)
+#endif
+
+/* XPS_LL_TEMAC direct registers definition */
+#define TEMAC_RAF0		(XILINX_LLTEMAC_BASEADDR + 0x00)
+#define TEMAC_TPF0		(XILINX_LLTEMAC_BASEADDR + 0x04)
+#define TEMAC_IFGP0		(XILINX_LLTEMAC_BASEADDR + 0x08)
+#define TEMAC_IS0		(XILINX_LLTEMAC_BASEADDR + 0x0c)
+#define TEMAC_IP0		(XILINX_LLTEMAC_BASEADDR + 0x10)
+#define TEMAC_IE0		(XILINX_LLTEMAC_BASEADDR + 0x14)
+
+#define TEMAC_MSW0		(XILINX_LLTEMAC_BASEADDR + 0x20)
+#define TEMAC_LSW0		(XILINX_LLTEMAC_BASEADDR + 0x24)
+#define TEMAC_CTL0		(XILINX_LLTEMAC_BASEADDR + 0x28)
+#define TEMAC_RDY0		(XILINX_LLTEMAC_BASEADDR + 0x2c)
+
+#define XTE_RSE_MIIM_RR_MASK	0x0002
+#define XTE_RSE_MIIM_WR_MASK	0x0004
+#define XTE_RSE_CFG_RR_MASK	0x0020
+#define XTE_RSE_CFG_WR_MASK	0x0040
+
+/* XPS_LL_TEMAC indirect registers offset definition */
+
+#define RCW0	0x200
+#define RCW1	0x240
+#define TC	0x280
+#define FCC	0x2c0
+#define EMMC	0x300
+#define PHYC	0x320
+#define MC	0x340
+#define UAW0	0x380
+#define UAW1	0x384
+#define MAW0	0x388
+#define MAW1	0x38c
+#define AFM	0x390
+#define TIS	0x3a0
+#define TIE	0x3a4
+#define MIIMWD	0x3b0
+#define MIIMAI	0x3b4
+
+#define CNTLREG_WRITE_ENABLE_MASK	0x8000
+#define CNTLREG_EMAC1SEL_MASK		0x0400
+#define CNTLREG_ADDRESSCODE_MASK	0x03ff
+
+#define MDIO_ENABLE_MASK	0x40
+#define MDIO_CLOCK_DIV_MASK	0x3F
+#define MDIO_CLOCK_DIV_100MHz	0x28
+
+#define ETHER_MTU		1520
+
+#ifdef SDMA_MODE
+/* CDMAC descriptor status bit definitions */
+# define BDSTAT_ERROR_MASK		0x80
+# define BDSTAT_INT_ON_END_MASK		0x40
+# define BDSTAT_STOP_ON_END_MASK	0x20
+# define BDSTAT_COMPLETED_MASK		0x10
+# define BDSTAT_SOP_MASK		0x08
+# define BDSTAT_EOP_MASK		0x04
+# define BDSTAT_CHANBUSY_MASK		0x02
+# define BDSTAT_CHANRESET_MASK		0x01
+
+/* SDMA Buffer Descriptor */
+
+typedef struct cdmac_bd_t {
+	struct cdmac_bd_t *next_p;
+	unsigned char *phys_buf_p;
+	unsigned long buf_len;
+	unsigned char stat;
+	unsigned char app1_1;
+	unsigned short app1_2;
+	unsigned long app2;
+	unsigned long app3;
+	unsigned long app4;
+	unsigned long app5;
+} cdmac_bd __attribute((aligned(32))) ;
+
+static cdmac_bd	tx_bd;
+static cdmac_bd	rx_bd;
+#endif
+
+#ifdef FIFO_MODE
+typedef struct ll_fifo_s {
+	int isr; /* Interrupt Status Register 0x0 */
+	int ier; /* Interrupt Enable Register 0x4 */
+	int tdfr; /* Transmit data FIFO reset 0x8 */
+	int tdfv; /* Transmit data FIFO Vacancy 0xC */
+	int tdfd; /* Transmit data FIFO 32bit wide data write port 0x10 */
+	int tlf; /* Write Transmit Length FIFO 0x14 */
+	int rdfr; /* Read Receive data FIFO reset 0x18 */
+	int rdfo; /* Receive data FIFO Occupancy 0x1C */
+	int rdfd; /* Read Receive data FIFO 32bit wide data read port 0x20 */
+	int rlf; /* Read Receive Length FIFO 0x24 */
+	int llr; /* Read LocalLink reset 0x28 */
+} ll_fifo_s;
+
+ll_fifo_s *ll_fifo = (ll_fifo_s *) (XILINX_LLTEMAC_FIFO_BASEADDR);
+#endif
+
+static unsigned char tx_buffer[ETHER_MTU] __attribute((aligned(32)));
+static unsigned char rx_buffer[ETHER_MTU] __attribute((aligned(32)));
+
+struct xps_ll_temac_private {
+	int idx;
+	unsigned char dev_addr[6];
+};
+
+#ifdef DEBUG
+/* undirect hostif write to ll_temac */
+static void xps_ll_temac_hostif_set(int emac, int phy_addr,
+					int reg_addr, int phy_data)
+{
+	out_be32((u32 *)TEMAC_LSW0, phy_data);
+	out_be32((u32 *)TEMAC_CTL0, CNTLREG_WRITE_ENABLE_MASK | MIIMWD);
+	out_be32((u32 *)TEMAC_LSW0, (phy_addr << 5) | (reg_addr));
+	out_be32((u32 *)TEMAC_CTL0, \
+			CNTLREG_WRITE_ENABLE_MASK | MIIMAI | (emac << 10));
+	while(! (in_be32((u32 *)TEMAC_RDY0) & XTE_RSE_MIIM_WR_MASK));
+}
+#endif
+
+/* undirect hostif read from ll_temac */
+static unsigned int xps_ll_temac_hostif_get(int emac, int phy_addr, int reg_addr)
+{
+	out_be32((u32 *)TEMAC_LSW0, (phy_addr << 5) | (reg_addr));
+	out_be32((u32 *)TEMAC_CTL0, MIIMAI | (emac << 10));
+	while(! (in_be32((u32 *)TEMAC_RDY0) & XTE_RSE_MIIM_RR_MASK));
+	return in_be32((u32 *)TEMAC_LSW0);
+}
+
+/* undirect write to ll_temac */
+static void xps_ll_temac_indirect_set(int emac, int reg_offset, int reg_data)
+{
+	out_be32((u32 *)TEMAC_LSW0, reg_data);
+	out_be32((u32 *)TEMAC_CTL0, \
+			CNTLREG_WRITE_ENABLE_MASK | (emac << 10) | reg_offset);
+	while(! (in_be32((u32 *)TEMAC_RDY0) & XTE_RSE_CFG_WR_MASK));
+}
+
+/* undirect read from ll_temac */
+int xps_ll_temac_indirect_get(int emac, int reg_offset)
+{
+	out_be32((u32 *)TEMAC_CTL0, (emac << 10) | reg_offset);
+	while(! (in_be32((u32 *)TEMAC_RDY0) & XTE_RSE_CFG_RR_MASK));
+	return in_be32((u32 *)TEMAC_LSW0);
+}
+
+#ifdef DEBUG
+/* read from phy */
+static void read_phy_reg (int phy_addr)
+{

drivers/net/xilinx_ll_temac.c line 216

Quoted 10 lines
drivers/net/xilinx_ll_temac.c new file +10 -0
+	int j, result;
+	printf("phy%d ",phy_addr);
+	for ( j = 0; j < 32; j++) {
+		result = xps_ll_temac_hostif_get(0, phy_addr, j);
+		printf("%d: 0x%x ", j, result);
+	}
+	puts("\n");
+}
+#endif
+

You only use this function in one place. Why not move this code to that place and use debug()

Quoted 27 lines
drivers/net/xilinx_ll_temac.c new file +26 -0
+static int phy_addr = -1;
+static int link = 0;
+
+/* setting ll_temac and phy to proper setting */
+static int xps_ll_temac_phy_ctrl(void)
+{
+	int i;
+	unsigned int result;
+	unsigned retries = 10;
+
+	if(link == 1)
+		return 1; /* link is setup */
+
+	/* wait for link up */
+	while (retries-- &&
+		((xps_ll_temac_hostif_get(0, phy_addr, 1) & 0x24) == 0x24))
+		;
+
+	if(phy_addr == -1) {
+		for(i = 31; i >= 0; i--) {
+			result = xps_ll_temac_hostif_get(0, i, 1);
+			if((result & 0x0ffff) != 0x0ffff) {
+#ifdef DEBUG
+				printf ("phy %x result %x\n", i, result);
+#endif
+				phy_addr = i;

drivers/net/xilinx_ll_temac.c line 252

Quoted 10 lines
drivers/net/xilinx_ll_temac.c new file +10 -0
+				break;
+			}
+		}
+	}
+
+	/* get PHY id */
+	i = (xps_ll_temac_hostif_get(0, phy_addr, 2) << 16) | \
+		xps_ll_temac_hostif_get(0, phy_addr, 3);
+#ifdef DEBUG
+	printf("LL_TEMAC: Phy ID 0x%x\n", i);

Please use debug() instead

Quoted 106 lines
drivers/net/xilinx_ll_temac.c new file +105 -0
+#endif
+
+#ifdef DEBUG
+	xps_ll_temac_hostif_set(0, 0, 0, 0x8000); /* phy reset */
+#endif
+	/* FIXME this part will be replaced by PHY lib */
+	/* s3e boards */
+	if (i == 0x7c0a3) {
+		xps_ll_temac_indirect_set(0, EMMC, 0x40000000); /* 100BASE-T/FD */
+		link = 1;
+		return 1;
+	}
+
+	/* Marwell 88e1111 id - ml50x */
+	if (i == 0x1410cc2) {
+		result = xps_ll_temac_hostif_get(0, phy_addr, 5);
+		if((result & 0x8000) == 0x8000) {
+			xps_ll_temac_indirect_set(0, EMMC, 0x80000000);
+			printf("1000BASE-T/FD\n");
+			link = 1;
+		} else if((result & 0x4000) == 0x4000) {
+			xps_ll_temac_indirect_set(0, EMMC, 0x40000000);
+			printf("100BASE-T/FD\n");
+			link = 1;
+		} else {
+			printf("Unsupported mode\n");
+			link = 0;
+		}
+		return 1;
+	}
+	return 0;
+}
+
+#ifdef SDMA_MODE
+/* bd init */
+static void xps_ll_temac_bd_init(void)
+{
+	memset((void *)&tx_bd, 0, sizeof(cdmac_bd));
+	memset((void *)&rx_bd, 0, sizeof(cdmac_bd));
+
+	rx_bd.phys_buf_p = &rx_buffer[0];
+
+	rx_bd.next_p = &rx_bd;
+	rx_bd.buf_len = ETHER_MTU;
+	flush_cache((u32)&rx_bd, sizeof(cdmac_bd));
+
+	out_be32((u32 *)RX_CURDESC_PTR, (u32)&rx_bd);
+	out_be32((u32 *)RX_TAILDESC_PTR, (u32)&rx_bd);
+	out_be32((u32 *)RX_NXTDESC_PTR, (u32)&rx_bd); /* setup first fd */
+
+	tx_bd.phys_buf_p = &tx_buffer[0];
+	tx_bd.next_p = &tx_bd;
+
+	flush_cache((u32)&tx_bd, sizeof(cdmac_bd));
+	out_be32((u32 *)TX_CURDESC_PTR, (u32)&tx_bd);
+}
+#endif
+
+#ifdef SDMA_MODE
+static int xps_ll_temac_send_sdma(unsigned char *buffer, int length)
+{
+	if( xps_ll_temac_phy_ctrl() == 0)
+		return 0;
+
+	memcpy (tx_buffer, buffer, length);
+	flush_cache ((u32)tx_buffer, length);
+
+	tx_bd.stat = BDSTAT_SOP_MASK | BDSTAT_EOP_MASK | BDSTAT_STOP_ON_END_MASK;
+	tx_bd.buf_len = length;
+	flush_cache ((u32)&tx_bd, sizeof(cdmac_bd));
+
+	out_be32((u32 *)TX_CURDESC_PTR, (u32)&tx_bd);
+	out_be32((u32 *)TX_TAILDESC_PTR, (u32)&tx_bd); /* DMA start */
+
+	do {
+		flush_cache ((u32)&tx_bd, sizeof(cdmac_bd));
+	} while(!(((volatile int)tx_bd.stat) & BDSTAT_COMPLETED_MASK));
+
+	return length;
+}
+
+
+static int xps_ll_temac_recv_sdma(void)
+{
+	int length;
+
+	flush_cache ((u32)&rx_bd, sizeof(cdmac_bd));
+
+	if(!(rx_bd.stat & BDSTAT_COMPLETED_MASK)) {
+		return 0;
+	}
+
+	length = rx_bd.app5;
+	flush_cache ((u32)rx_bd.phys_buf_p, length);
+
+	rx_bd.buf_len = ETHER_MTU;
+	rx_bd.stat = 0;
+	rx_bd.app5 = 0;
+
+	flush_cache ((u32)&rx_bd, sizeof(cdmac_bd));
+	out_be32((u32 *)RX_TAILDESC_PTR, (u32)&rx_bd);
+
+	if(length > 0) {
+		NetReceive(rx_bd.phys_buf_p, length);
+	}

drivers/net/xilinx_ll_temac.c line 367

Quoted 10 lines
drivers/net/xilinx_ll_temac.c new file +10 -0
+
+	return length;
+}
+#endif
+
+
+#ifdef FIFO_MODE
+void debugll(int count)
+{
+	printf ("%d fifo isr 0x%08x, fifo_ier 0x%08x, fifo_rdfr 0x%08x, fifo_rdfo 0x%08x fifo_rlr 0x%08x\n",count, ll_fifo->isr, \

This line's way too long.

Quoted 121 lines
drivers/net/xilinx_ll_temac.c new file +120 -0
+	ll_fifo->ier, ll_fifo->rdfr, ll_fifo->rdfo, ll_fifo->rlf);
+
+}
+
+static int xps_ll_temac_send_fifo(unsigned char *buffer, int length)
+{
+	u32 *buf = buffer;
+	u32 len, i, val;
+
+	len = (length / 4) + 1;
+
+	for (i = 0; i < len; i++) {
+		val = *buf++;
+		ll_fifo->tdfd = val;
+	}
+
+	ll_fifo->tlf = length;
+
+	return length;
+}
+
+static int xps_ll_temac_recv_fifo()
+{
+	int len, len2, i, val;
+	int *buf;
+	buf = &rx_buffer;
+
+	if (ll_fifo->isr & 0x04000000 ) {
+		ll_fifo->isr = 0xffffffff; /* reset isr */
+	
+		/* while (ll_fifo->isr); */
+		len = ll_fifo->rlf & 0x7FF;
+		len2 = (len / 4) + 1;
+	
+		for (i = 0; i < len2; i++) {
+			val = ll_fifo->rdfd;
+			*buf++ = val ;
+		}
+	
+		/* debugll(1); */
+		NetReceive (&rx_buffer, len);
+	}
+	return 0;
+}
+#endif
+
+/* setup mac addr */
+static int xps_ll_temac_addr_setup(struct xps_ll_temac_private * lp)
+{
+	char * env_p;
+	char * end;
+	int i, val;
+
+	env_p = getenv("ethaddr");
+	if (env_p == NULL) {
+		printf("cannot get enviroment for \"ethaddr\".\n");
+		return -1;
+	}
+
+	for (i = 0; i < 6; i++) {
+		lp->dev_addr[i] = env_p ? simple_strtoul(env_p, &end, 16) : 0;
+		if (env_p) env_p = (*end) ? end + 1 : end;
+	}
+
+	/* set up unicast MAC address filter */
+	val = ((lp->dev_addr[3] << 24) | (lp->dev_addr[2] << 16) |
+		(lp->dev_addr[1] << 8) | (lp->dev_addr[0] ));
+	xps_ll_temac_indirect_set(0, UAW0, val);
+	val = (lp->dev_addr[5] << 8) | lp->dev_addr[4] ;
+	xps_ll_temac_indirect_set(0, UAW1, val);
+
+	return 0;
+}
+
+static int xps_ll_temac_init(struct eth_device *dev, bd_t *bis)
+{
+	struct xps_ll_temac_private *lp = (struct xps_ll_temac_private *)dev->priv;
+
+#ifdef SDMA_MODE
+	xps_ll_temac_bd_init();
+#endif
+#ifdef FIFO_MODE
+	ll_fifo->tdfr = 0x000000a5; /* set fifo lenght */
+	ll_fifo->rdfr = 0x000000a5;
+
+	/* ll_fifo->isr = 0x0; */
+	/* ll_fifo->ier = 0x0; */
+#endif
+	xps_ll_temac_indirect_set(0, MC, MDIO_ENABLE_MASK | MDIO_CLOCK_DIV_100MHz);
+
+	xps_ll_temac_addr_setup(lp);
+	xps_ll_temac_indirect_set(0, AFM, 0x00000000); /* Promiscuos mode disable */
+	xps_ll_temac_indirect_set(0, RCW1, 0x10000000); /* Enable Receiver */
+	xps_ll_temac_indirect_set(0, TC, 0x10000000); /* Enable Transmitter */
+
+}
+
+#ifdef ETH_HALTING
+static int xps_ll_temac_halt(void)
+{
+	xps_ll_temac_indirect_set(0, RCW1, 0x00000000); /* Disable Receiver */
+	xps_ll_temac_indirect_set(0, TC, 0x00000000); /* Disable Transmitter */
+
+#ifdef SDMA_MODE
+	out_be32((u32 *)DMA_CONTROL_REG, 0x00000001);
+	while(in_be32((u32 *)DMA_CONTROL_REG) & 1);
+#endif
+#ifdef FIFO_MODE
+	/* reset fifos */
+#endif
+}
+#endif
+
+/* halt device */
+void eth_halt(void){
+	link = 0;
+#ifdef ETH_HALTING
+	xps_ll_temac_halt();
+#endif
+}

drivers/net/xilinx_ll_temac.c line 497

Quoted 10 lines
drivers/net/xilinx_ll_temac.c new file +10 -0
+
+int eth_init(bd_t *bis)
+{
+	static int first = 1;
+	struct eth_device *dev;
+	struct xps_ll_temac_private *lp;
+#if DEBUG
+	int i;
+#endif
+

I normally don't like this, but move the variable declaration to where it's used to avoid 2 #ifs

Quoted 37 lines
drivers/net/xilinx_ll_temac.c new file +36 -0
+	if(!first)
+		return 0;
+	first = 0;
+	dev = (struct eth_device *) calloc(1, sizeof(struct eth_device));
+	if (NULL == dev)
+		return 0;
+
+	lp = (struct xps_ll_temac_private *) calloc(1, sizeof(struct xps_ll_temac_private));
+	if (lp == NULL)
+		return 0;
+	dev->priv = lp;
+	sprintf(dev->name, "eth0");
+
+	xps_ll_temac_init(dev, bis);
+
+	printf("%s: Xilinx XPS LocalLink Tri-Mode Ether MAC #%d at 0x%08X.\n",
+		dev->name, 0, XILINX_LLTEMAC_BASEADDR);
+
+#if DEBUG
+	for(i = 0;i < 32;i++) {
+		read_phy_reg(i);
+	}
+#endif
+	xps_ll_temac_phy_ctrl();
+	return 1;
+}
+
+int eth_send(volatile void *packet, int length)
+{
+#ifdef SDMA_MODE
+	return xps_ll_temac_send_sdma((unsigned char *)packet, length);
+#endif
+#ifdef FIFO_MODE
+	return xps_ll_temac_send_fifo((unsigned char *)packet, length);
+#endif
+}

drivers/net/xilinx_ll_temac.c line 543

Quoted 10 lines
drivers/net/xilinx_ll_temac.c new file +10 -0
+
+int eth_rx(void)
+{
+#ifdef SDMA_MODE
+	return xps_ll_temac_recv_sdma();
+#endif
+#ifdef FIFO_MODE
+	return xps_ll_temac_recv_fifo();
+#endif
+}

regards,
B en

#8 lore raw

PATCH 1/2 net: Sort Makefile labels

In reply to monstr at monstr.eu

Dear monstr at monstr.eu,

In message <d4fd46d3b2d27747b728da37fad3ebca3732d837.1231154122.git.monstr@monstr.eu> you wrote:
...

Quoted 4 lines

COBJS-$(CONFIG_DRIVER_3C589) += 3c589.o
+COBJS-$(CONFIG_PPC4xx_EMAC) += 4xx_enet.o
COBJS-$(CONFIG_DRIVER_AX88180) += ax88180.o

I have to admit that I rather sort this by the name of the CONFIG_ variable. Seeing "P" here between two "D"s looks strange to me.

Please don't do this.

Quoted 4 lines

COBJS-$(CONFIG_VSC7385_ENET) += vsc7385.o COBJS-$(CONFIG_XILINX_EMAC) += xilinx_emac.o COBJS-$(CONFIG_XILINX_EMACLITE) += xilinx_emaclite.o
-COBJS-$(CONFIG_SH_ETHER) += sh_eth.o

Of course this is still wrong, like some others.

Best regards,

Wolfgang Denk

Signature · 6 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
"There are three principal ways to lose money: wine, women,  and  en-
gineers.  While  the first two are more pleasant, the third is by far
the more certain."                      -- Baron Rothschild, ca. 1800
#9 lore raw

PATCH 1/2 net: Sort Makefile labels

In reply to monstr at monstr.eu

Dear monstr at monstr.eu,

Quoted 7 lines · In message <d4fd46d3b2d27747b728da37fad3ebca3732d837.1231154122.git.monstr@mons…

From: Michal Simek <[email protected]>

---
 drivers/net/Makefile |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

Sorry, this doesn't apply either:

Applying: part_efi: Fix partition size calculation due to inclusive ending LBA.
error: patch failed: disk/part_efi.c:163
error: disk/part_efi.c: patch does not apply
Using index info to reconstruct a base tree...
error: patch failed: disk/part_efi.c:163
error: disk/part_efi.c: patch does not apply
Did you hand edit your patch?
It does not apply to blobs recorded in its index.
Cannot fall back to three-way merge.
Patch failed at 0001.

Please rebase and repost. Thanks.

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
There are two ways to write error-free programs. Only the  third  one
works.
Applied
#10 lore raw

PATCH 1/2 net: Sort Makefile labels

In reply to monstr at monstr.eu
Quoted 35 lines · monstr at monstr.eu

From: Michal Simek <[email protected]>

---
 drivers/net/Makefile |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)
drivers/net/Makefile +2 -1
index 631336a..128dc11 100644
@@ -26,6 +26,7 @@ include $(TOPDIR)/config.mk
 LIB	:= $(obj)libnet.a
 
 COBJS-$(CONFIG_DRIVER_3C589) += 3c589.o
+COBJS-$(CONFIG_PPC4xx_EMAC) += 4xx_enet.o
 COBJS-$(CONFIG_DRIVER_AX88180) += ax88180.o
 COBJS-$(CONFIG_BCM570x) += bcm570x.o bcm570x_autoneg.o 5701rls.o
 COBJS-$(CONFIG_BFIN_MAC) += bfin_mac.o
@@ -54,11 +55,11 @@ COBJS-$(CONFIG_NS8382X) += ns8382x.o
 COBJS-$(CONFIG_DRIVER_NS9750_ETHERNET) += ns9750_eth.o
 COBJS-$(CONFIG_PCNET) += pcnet.o
 COBJS-$(CONFIG_PLB2800_ETHER) += plb2800_eth.o
-COBJS-$(CONFIG_PPC4xx_EMAC) += 4xx_enet.o
 COBJS-$(CONFIG_DRIVER_RTL8019) += rtl8019.o
 COBJS-$(CONFIG_RTL8139) += rtl8139.o
 COBJS-$(CONFIG_RTL8169) += rtl8169.o
 COBJS-$(CONFIG_DRIVER_S3C4510_ETH) += s3c4510b_eth.o
+COBJS-$(CONFIG_SH_ETHER) += sh_eth.o
 COBJS-$(CONFIG_DRIVER_SMC91111) += smc91111.o
 COBJS-$(CONFIG_DRIVER_SMC911X) += smc911x.o
 COBJS-$(CONFIG_TIGON3) += tigon3.o bcm570x_autoneg.o 5701rls.o

drivers/net/Makefile COBJS-$(CONFIG_ULI526X) += uli526x.o · line 74

Quoted 7 lines
drivers/net/Makefile +0 -1
@@ -68,7 +69,6 @@ COBJS-$(CONFIG_ULI526X) += uli526x.o
 COBJS-$(CONFIG_VSC7385_ENET) += vsc7385.o
 COBJS-$(CONFIG_XILINX_EMAC) += xilinx_emac.o
 COBJS-$(CONFIG_XILINX_EMACLITE) += xilinx_emaclite.o
-COBJS-$(CONFIG_SH_ETHER) += sh_eth.o
 
 COBJS	:= $(COBJS-y)
 SRCS	:= $(COBJS:.o=.c)

Applied to net repo.

thanks,
Ben

#11 lore raw

PATCH 1/2 net: Sort Makefile labels

In reply to monstr at monstr.eu

Dear monstr at monstr.eu,

Quoted 13 lines · In message <d4fd46d3b2d27747b728da37fad3ebca3732d837.1231154122.git.monstr@mons…

From: Michal Simek <[email protected]>

---
 drivers/net/Makefile |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)
drivers/net/Makefile +0 -0
index 631336a..128dc11 100644

drivers/net/Makefile include $(TOPDIR)/config.mk · line 30

Quoted 5 lines
drivers/net/Makefile +1 -0
@@ -26,6 +26,7 @@ include $(TOPDIR)/config.mk
 LIB	:= $(obj)libnet.a
 
 COBJS-$(CONFIG_DRIVER_3C589) += 3c589.o
+COBJS-$(CONFIG_PPC4xx_EMAC) += 4xx_enet.o
 COBJS-$(CONFIG_DRIVER_AX88180) += ax88180.o

What makes you think that CONFIG_PPC4xx_EMAC has a good place between CONFIG_DRIVER_3C589 and CONFIG_DRIVER_AX88180?

drivers/net/Makefile COBJS-$(CONFIG_NS8382X) += ns8382x.o · line 57

Quoted 6 lines
drivers/net/Makefile +0 -1
 COBJS-$(CONFIG_BFIN_MAC) += bfin_mac.o
@@ -54,11 +55,11 @@ COBJS-$(CONFIG_NS8382X) += ns8382x.o
 COBJS-$(CONFIG_DRIVER_NS9750_ETHERNET) += ns9750_eth.o
 COBJS-$(CONFIG_PCNET) += pcnet.o
 COBJS-$(CONFIG_PLB2800_ETHER) += plb2800_eth.o
-COBJS-$(CONFIG_PPC4xx_EMAC) += 4xx_enet.o

that was ok here.

drivers/net/Makefile COBJS-$(CONFIG_NS8382X) += ns8382x.o · line 58

Quoted 1 line
drivers/net/Makefile +0 -0
 COBJS-$(CONFIG_DRIVER_RTL8019) += rtl8019.o

...while this isn't.

drivers/net/Makefile COBJS-$(CONFIG_ULI526X) += uli526x.o · line 71

Quoted 12 lines
drivers/net/Makefile +1 -1
 COBJS-$(CONFIG_RTL8139) += rtl8139.o
 COBJS-$(CONFIG_RTL8169) += rtl8169.o
 COBJS-$(CONFIG_DRIVER_S3C4510_ETH) += s3c4510b_eth.o
+COBJS-$(CONFIG_SH_ETHER) += sh_eth.o
 COBJS-$(CONFIG_DRIVER_SMC91111) += smc91111.o
 COBJS-$(CONFIG_DRIVER_SMC911X) += smc911x.o
 COBJS-$(CONFIG_TIGON3) += tigon3.o bcm570x_autoneg.o 5701rls.o
@@ -68,7 +69,6 @@ COBJS-$(CONFIG_ULI526X) += uli526x.o
 COBJS-$(CONFIG_VSC7385_ENET) += vsc7385.o
 COBJS-$(CONFIG_XILINX_EMAC) += xilinx_emac.o
 COBJS-$(CONFIG_XILINX_EMACLITE) += xilinx_emaclite.o
-COBJS-$(CONFIG_SH_ETHER) += sh_eth.o

Actually we should perform a global s/CONFIG_DRIVER_/CONFIG_/, me thinks.

Best regards,

Wolfgang Denk

Signature · 6 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Worlds may change, galaxies disintegrate, but a woman always  remains
a woman.
	-- Kirk, "The Conscience of the King", stardate 2818.9
#12 lore raw

PATCH 1/2 net: Sort Makefile labels

Quoted 45 lines · Wolfgang Denk

Dear monstr at monstr.eu,

Quoted 13 lines · In message <d4fd46d3b2d27747b728da37fad3ebca3732d837.1231154122.git.monstr@mons…

From: Michal Simek <[email protected]>

---
 drivers/net/Makefile |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)
drivers/net/Makefile +0 -0
index 631336a..128dc11 100644

drivers/net/Makefile include $(TOPDIR)/config.mk · line 31

Quoted 6 lines
drivers/net/Makefile +1 -0
@@ -26,6 +26,7 @@ include $(TOPDIR)/config.mk
 LIB	:= $(obj)libnet.a
 
 COBJS-$(CONFIG_DRIVER_3C589) += 3c589.o
+COBJS-$(CONFIG_PPC4xx_EMAC) += 4xx_enet.o
 COBJS-$(CONFIG_DRIVER_AX88180) += ax88180.o
 

What makes you think that CONFIG_PPC4xx_EMAC has a good place between CONFIG_DRIVER_3C589 and CONFIG_DRIVER_AX88180?

drivers/net/Makefile COBJS-$(CONFIG_NS8382X) += ns8382x.o · line 58

Quoted 7 lines
drivers/net/Makefile +0 -1
 COBJS-$(CONFIG_BFIN_MAC) += bfin_mac.o
@@ -54,11 +55,11 @@ COBJS-$(CONFIG_NS8382X) += ns8382x.o
 COBJS-$(CONFIG_DRIVER_NS9750_ETHERNET) += ns9750_eth.o
 COBJS-$(CONFIG_PCNET) += pcnet.o
 COBJS-$(CONFIG_PLB2800_ETHER) += plb2800_eth.o
-COBJS-$(CONFIG_PPC4xx_EMAC) += 4xx_enet.o
 

that was ok here.

drivers/net/Makefile COBJS-$(CONFIG_NS8382X) += ns8382x.o · line 60

Quoted 2 lines
drivers/net/Makefile +0 -0
 COBJS-$(CONFIG_DRIVER_RTL8019) += rtl8019.o
 

...while this isn't.

Quoted 18 lines

drivers/net/Makefile COBJS-$(CONFIG_ULI526X) += uli526x.o · line 72

Quoted 13 lines
drivers/net/Makefile +1 -0
 COBJS-$(CONFIG_RTL8139) += rtl8139.o
 COBJS-$(CONFIG_RTL8169) += rtl8169.o
 COBJS-$(CONFIG_DRIVER_S3C4510_ETH) += s3c4510b_eth.o
+COBJS-$(CONFIG_SH_ETHER) += sh_eth.o
 COBJS-$(CONFIG_DRIVER_SMC91111) += smc91111.o

COBJS-$(CONFIG_DRIVER_SMC911X) += smc911x.o COBJS-$(CONFIG_TIGON3) += tigon3.o bcm570x_autoneg.o 5701rls.o

drivers/net/Makefile +0 -1
@@ -68,7 +69,6 @@ COBJS-$(CONFIG_ULI526X) += uli526x.o
 COBJS-$(CONFIG_VSC7385_ENET) += vsc7385.o
 COBJS-$(CONFIG_XILINX_EMAC) += xilinx_emac.o
 COBJS-$(CONFIG_XILINX_EMACLITE) += xilinx_emaclite.o
-COBJS-$(CONFIG_SH_ETHER) += sh_eth.o
 

Actually we should perform a global s/CONFIG_DRIVER_/CONFIG_/, me thinks.

Agreed. Can you do it Michal, or should I (for net, anyway)?

Quoted 5 lines

Best regards,

Wolfgang Denk

regards,
Ben

#13 lore raw

PATCH 1/2 net: Sort Makefile labels

Hi Ben and Wolfgang,

Quoted 27 lines
Quoted 27 lines

In message <d4fd46d3b2d27747b728da37fad3ebca3732d837.1231154122.git.monstr@monstr.eu>

Quoted 13 lines · you

From: Michal Simek <[email protected]>

---
 drivers/net/Makefile |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)
drivers/net/Makefile +0 -0
index 631336a..128dc11 100644

drivers/net/Makefile include $(TOPDIR)/config.mk · line 31

Quoted 6 lines
drivers/net/Makefile +1 -0
@@ -26,6 +26,7 @@ include $(TOPDIR)/config.mk
 LIB    := $(obj)libnet.a
 
 COBJS-$(CONFIG_DRIVER_3C589) += 3c589.o
+COBJS-$(CONFIG_PPC4xx_EMAC) += 4xx_enet.o
 COBJS-$(CONFIG_DRIVER_AX88180) += ax88180.o
 

What makes you think that CONFIG_PPC4xx_EMAC has a good place between CONFIG_DRIVER_3C589 and CONFIG_DRIVER_AX88180?

#14 lore raw

PATCH 1/2 net: Sort Makefile labels

Quoted 11 lines · Michal Simek
Quoted 10 lines
Quoted 10 lines

drivers/net/Makefile include $(TOPDIR)/config.mk · line 30

Quoted 7 lines

+++ b/drivers/net/Makefile

drivers/net/Makefile +1 -0
@@ -26,6 +26,7 @@ include $(TOPDIR)/config.mk
 LIB    := $(obj)libnet.a
 
 COBJS-$(CONFIG_DRIVER_3C589) += 3c589.o
+COBJS-$(CONFIG_PPC4xx_EMAC) += 4xx_enet.o
 COBJS-$(CONFIG_DRIVER_AX88180) += ax88180.o

What makes you think that CONFIG_PPC4xx_EMAC has a good place between CONFIG_DRIVER_3C589 and CONFIG_DRIVER_AX88180?

Quoted 4 lines

From my point of view make me more sense to rename 4xx_enec.c file to correspond with driver name. I mean if the config is CONFIG_PPC4xx_EMAC the filename should be ppc4xx_emac.c or in second case CONFIG_4XX_ENET for current 4xx_enet.c file. From these two choices I like the first one.

Yes, I also think it makes sense to rename this driver to "ppc4xx_emac.c". If nobody objects I'll send a patch doing this shortly.

Best regards,
Stefan

=====================================================================
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office at denx.de
=====================================================================

#15 lore raw

PATCH 1/2 net: Sort Makefile labels

In reply to Michal Simek

Dear Michal Simek,

Quoted 6 lines · In message <[email protected]> you wrote:
Quoted 6 lines
Quoted 6 lines
Quoted 5 lines

COBJS-$(CONFIG_DRIVER_3C589) += 3c589.o
+COBJS-$(CONFIG_PPC4xx_EMAC) += 4xx_enet.o
COBJS-$(CONFIG_DRIVER_AX88180) += ax88180.o

Quoted 7 lines
Quoted 2 lines
Quoted 2 lines

What makes you think that CONFIG_PPC4xx_EMAC has a good place between CONFIG_DRIVER_3C589 and CONFIG_DRIVER_AX88180?

From my point of view make me more sense to rename 4xx_enec.c file to correspond with driver name. I mean if the config is CONFIG_PPC4xx_EMAC the filename should be ppc4xx_emac.c or in second case CONFIG_4XX_ENET for current 4xx_enet.c file. From these two choices I like the first one.

Well, that's IMO for the PPC4xx custodian to decide.

Stefan, what do you think?

Quoted 5 lines

I would like to describe what happen. I sent to mailing list two patches. One with Makefile sort and second with LL_TEMAC. First patch just sort some labels in drivers/net/Makefile. Wolfgang sent that he don't like it and he reject this patch. I haven't wanted to sort any Makefile labels I just wanted to add LL_TEMAC driver. Makefile sort was not my point.

I'm aware of this. But this is how contributing to U-Boot works : you just want to add a tiny piece of code here, and you will find yourself involved in some bigger cleanup all over the place. That's what happens to me all day, and to many others, so please help to carry that burdon. Thanks.

Best regards,

Wolfgang Denk

Signature · 7 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Yes, it's a technical challenge, and  you  have  to  kind  of  admire
people  who go to the lengths of actually implementing it, but at the
same time you wonder about their IQ...
         --  Linus Torvalds in <[email protected]>
#16 lore raw

PATCH 1/2 net: Sort Makefile labels

Dear Wolfgang,

Quoted 24 lines

Dear Michal Simek,

Quoted 10 lines · In message <[email protected]> you wrote:
Quoted 6 lines
Quoted 6 lines
Quoted 4 lines

COBJS-$(CONFIG_DRIVER_3C589) += 3c589.o
+COBJS-$(CONFIG_PPC4xx_EMAC) += 4xx_enet.o
COBJS-$(CONFIG_DRIVER_AX88180) += ax88180.o

What makes you think that CONFIG_PPC4xx_EMAC has a good place between CONFIG_DRIVER_3C589 and CONFIG_DRIVER_AX88180?

From my point of view make me more sense to rename 4xx_enec.c file to correspond with driver name. I mean if the config is CONFIG_PPC4xx_EMAC the filename should be ppc4xx_emac.c or in second case CONFIG_4XX_ENET for current 4xx_enet.c file. From these two choices I like the first one.

Well, that's IMO for the PPC4xx custodian to decide.

Stefan, what do you think?

Quoted 5 lines

I would like to describe what happen. I sent to mailing list two patches. One with Makefile sort and second with LL_TEMAC. First patch just sort some labels in drivers/net/Makefile. Wolfgang sent that he don't like it and he reject this patch. I haven't wanted to sort any Makefile labels I just wanted to add LL_TEMAC driver. Makefile sort was not my point.

Quoted 5 lines

I'm aware of this. But this is how contributing to U-Boot works : you just want to add a tiny piece of code here, and you will find yourself involved in some bigger cleanup all over the place. That's what happens to me all day, and to many others, so please help to carry that burdon. Thanks.

Yes I know that's why I sent sort Makefile patch in that pack.

Regards,
Michal

Quoted 4 lines

Best regards,

Wolfgang Denk

Signature · 2 lines
Michal Simek, Ing. (M.Eng)
w: www.monstr.eu p: +42-0-721842854
#17 lore raw

PATCH 1/2 net: Sort Makefile labels

Quoted 46 lines · Michal Simek

Dear Wolfgang,

Quoted 10 lines

Dear Michal Simek,

Quoted 7 lines · In message <[email protected]> you wrote:
Quoted 6 lines
Quoted 6 lines
Quoted 4 lines

COBJS-$(CONFIG_DRIVER_3C589) += 3c589.o
+COBJS-$(CONFIG_PPC4xx_EMAC) += 4xx_enet.o
COBJS-$(CONFIG_DRIVER_AX88180) += ax88180.o

What makes you think that CONFIG_PPC4xx_EMAC has a good place between CONFIG_DRIVER_3C589 and CONFIG_DRIVER_AX88180?

From my point of view make me more sense to rename 4xx_enec.c file to

correspond

Quoted 1 line
Quoted 1 line

with driver name. I mean if the config is CONFIG_PPC4xx_EMAC the

filename should

Quoted 1 line
Quoted 1 line

be ppc4xx_emac.c or in second case CONFIG_4XX_ENET for current

4xx_enet.c file.

Quoted 7 lines
Quoted 1 line

From these two choices I like the first one.

Well, that's IMO for the PPC4xx custodian to decide.

Stefan, what do you think?

Quoted 1 line

I would like to describe what happen. I sent to mailing list two

patches. One

Quoted 1 line
Quoted 1 line

with Makefile sort and second with LL_TEMAC. First patch just sort some

labels

Quoted 1 line
Quoted 1 line

in drivers/net/Makefile. Wolfgang sent that he don't like it and he

reject this

Quoted 8 lines
Quoted 2 lines

patch. I haven't wanted to sort any Makefile labels I just wanted to add LL_TEMAC driver. Makefile sort was not my point.

I'm aware of this. But this is how contributing to U-Boot works : you just want to add a tiny piece of code here, and you will find yourself involved in some bigger cleanup all over the place. That's what happens to me all day, and to many others, so please help to carry that burdon. Thanks.

Yes I know that's why I sent sort Makefile patch in that pack.

Regards,
Michal

Quoted 2 lines

Best regards,

Quoted 8 lines
Quoted 2 lines

Wolfgang Denk

--
Michal Simek, Ing. (M.Eng)
w: www.monstr.eu p: +42-0-721842854

OK, I'll send a clean up patch later today or tonight.

regards,
Ben