Skip to contentSkip to conversation

Better I2C support for OMAP

U-Boot New Tom Rix 3 participants 10 messages Nov 1, 2009 – Nov 5, 2009 I2C

lore
Share & tools
lore

Tom Rix

author
#0 lore raw
Quoted 19 lines
Quoted 16 lines · Hugo Vincent

Hi everyone,

This patch improves I2C support on OMAP3 platforms, including supporting the second and third I2C controllers (mainline only supports the first controller), and supporting writes where alen=0 (i.e. no register address - this was needed for some I2C devices I needed to drive from u-boot).

Tested on Gumstix Overo with several I2C devices, on I2C1 and 3.

The multi-bus support is mostly from this patch:
http://www.beagleboard.org/gitweb/?p=u-boot-arm.git;a=commit;h=52eddcd07c2e7ad61d15bab2cf2d0d21466eaca2

Best regards,
Hugo Vincent

Quoted 8 lines
Quoted 1 line

Does anybody have an up to date version of this patch? Seems we need it for Beagle Zippy, too.

Thanks

Dirk

I have cherry picked this change from the above git repo.

There was some cleanup.

The omap24xx version changes were done.

The structure elements were changed from unsigned int to unsigned shorts + pads to make the elements consistent with the writew/readw 's used in the driver.

The cmd_i2c changes that did not seem to fit were removed.

I have run tested this on beagle and zoom1.

Since I don't have a zippy board handy I did not get anything meaningfull when switching to i2c dev 2. The default i2c dev worked as expected.

Tom

author
#1 lore raw

PATCH 1/1 Add I2C multibus support for OMAP2/3 boards

From: Syed Mohammed Khasim <[email protected]>

This was cherry-picked from

repo: http://www.beagleboard.org/u-boot-arm.git
commit: 52eddcd07c2e7ad61d15bab2cf2d0d21466eaca2

In addition to adding multibus support, this patch also cleans up the register access. The register access has been changed from #defines to a structure.

---
 drivers/i2c/omap24xx_i2c.c          |  220 ++++++++++++++++++++++-------------
 include/asm-arm/arch-omap24xx/i2c.h |   52 ++++++---
 include/asm-arm/arch-omap3/i2c.h    |   48 +++++---
 include/configs/omap3_beagle.h      |    1 +
 4 files changed, 209 insertions(+), 112 deletions(-)
drivers/i2c/omap24xx_i2c.c +139 -81
index 1a4c8c9..763d2f8 100644
@@ -1,7 +1,7 @@
 /*
  * Basic I2C functions
  *
- * Copyright (c) 2004 Texas Instruments
+ * Copyright (c) 2004, 2009 Texas Instruments
  *
  * This package is free software;  you can redistribute it and/or
  * modify it under the terms of the license found in the file
@@ -25,10 +25,18 @@
 #include <asm/arch/i2c.h>
 #include <asm/io.h>
 
+#ifdef CONFIG_OMAP34XX
+#define I2C_NUM_IF 3
+#else
+#define I2C_NUM_IF 2
+#endif
+
 static void wait_for_bb (void);
 static u16 wait_for_pin (void);
 static void flush_fifo(void);
 
+static i2c_t *i2c = (i2c_t *)I2C_DEFAULT_BASE;
+
 void i2c_init (int speed, int slaveadd)
 {
 	int psc, fsscll, fssclh;
@@ -95,30 +103,30 @@ void i2c_init (int speed, int slaveadd)
 		sclh = (unsigned int)fssclh;
 	}
 
-	writew(0x2, I2C_SYSC); /* for ES2 after soft reset */
+	writew(0x2, &i2c->sysc);		/* for ES2 after soft reset */
 	udelay(1000);
-	writew(0x0, I2C_SYSC); /* will probably self clear but */
+	writew(0x0, &i2c->sysc);		/* will probably self clear but */
 
-	if (readw (I2C_CON) & I2C_CON_EN) {
-		writew (0, I2C_CON);
-		udelay (50000);
+	if (readw(&i2c->con) & I2C_CON_EN) {
+		writew(0, &i2c->con);
+		udelay(50000);
 	}
 
-	writew(psc, I2C_PSC);
-	writew(scll, I2C_SCLL);
-	writew(sclh, I2C_SCLH);
+	writew(psc, &i2c->psc);
+	writew(scll, &i2c->scll);
+	writew(sclh, &i2c->sclh);
 
 	/* own address */
-	writew (slaveadd, I2C_OA);
-	writew (I2C_CON_EN, I2C_CON);
+	writew(slaveadd, &i2c->oa);
+	writew(I2C_CON_EN, &i2c->con);
 
 	/* have to enable intrrupts or OMAP i2c module doesn't work */
-	writew (I2C_IE_XRDY_IE | I2C_IE_RRDY_IE | I2C_IE_ARDY_IE |
-		I2C_IE_NACK_IE | I2C_IE_AL_IE, I2C_IE);
-	udelay (1000);
+	writew(I2C_IE_XRDY_IE | I2C_IE_RRDY_IE | I2C_IE_ARDY_IE |
+		I2C_IE_NACK_IE | I2C_IE_AL_IE, &i2c->ie);
+	udelay(1000);
 	flush_fifo();
-	writew (0xFFFF, I2C_STAT);
-	writew (0, I2C_CNT);
+	writew(0xFFFF, &i2c->stat);
+	writew(0, &i2c->cnt);
 }
 
 static int i2c_read_byte (u8 devaddr, u8 regoffset, u8 * value)
@@ -130,19 +138,19 @@ static int i2c_read_byte (u8 devaddr, u8 regoffset, u8 * value)
 	wait_for_bb ();
 
 	/* one byte only */
-	writew (1, I2C_CNT);
+	writew(1, &i2c->cnt);
 	/* set slave address */
-	writew (devaddr, I2C_SA);
+	writew(devaddr, &i2c->sa);
 	/* no stop bit needed here */
-	writew (I2C_CON_EN | I2C_CON_MST | I2C_CON_STT | I2C_CON_TRX, I2C_CON);
+	writew(I2C_CON_EN | I2C_CON_MST | I2C_CON_STT | I2C_CON_TRX, &i2c->con);
 
 	status = wait_for_pin ();
 
 	if (status & I2C_STAT_XRDY) {
 		/* Important: have to use byte access */
-		writeb (regoffset, I2C_DATA);
-		udelay (20000);
-		if (readw (I2C_STAT) & I2C_STAT_NACK) {
+		writeb(regoffset, &i2c->data);
+		udelay(20000);
+		if (readw(&i2c->stat) & I2C_STAT_NACK) {
 			i2c_error = 1;
 		}
 	} else {
@@ -151,28 +159,28 @@ static int i2c_read_byte (u8 devaddr, u8 regoffset, u8 * value)
 
 	if (!i2c_error) {
 		/* free bus, otherwise we can't use a combined transction */
-		writew (0, I2C_CON);
-		while (readw (I2C_STAT) || (readw (I2C_CON) & I2C_CON_MST)) {
+		writew(0, &i2c->con);
+		while (readw(&i2c->stat) || (readw(&i2c->con) & I2C_CON_MST)) {
 			udelay (10000);
 			/* Have to clear pending interrupt to clear I2C_STAT */
-			writew (0xFFFF, I2C_STAT);
+			writew(0xFFFF, &i2c->stat);
 		}
 
 		wait_for_bb ();
 		/* set slave address */
-		writew (devaddr, I2C_SA);
+		writew(devaddr, &i2c->sa);
 		/* read one byte from slave */
-		writew (1, I2C_CNT);
+		writew(1, &i2c->cnt);
 		/* need stop bit here */
-		writew (I2C_CON_EN | I2C_CON_MST | I2C_CON_STT | I2C_CON_STP,
-			I2C_CON);
+		writew(I2C_CON_EN | I2C_CON_MST | I2C_CON_STT | I2C_CON_STP,
+			&i2c->con);
 
 		status = wait_for_pin ();
 		if (status & I2C_STAT_RRDY) {
 #if defined(CONFIG_OMAP243X) || defined(CONFIG_OMAP34XX)
-			*value = readb (I2C_DATA);
+			*value = readb(&i2c->data);
 #else
-			*value = readw (I2C_DATA);
+			*value = readw(&i2c->data);
 #endif
 			udelay (20000);
 		} else {
@@ -180,17 +188,17 @@ static int i2c_read_byte (u8 devaddr, u8 regoffset, u8 * value)
 		}
 
 		if (!i2c_error) {
-			writew (I2C_CON_EN, I2C_CON);
-			while (readw (I2C_STAT)
-			       || (readw (I2C_CON) & I2C_CON_MST)) {
+			writew(I2C_CON_EN, &i2c->con);
+			while (readw(&i2c->stat)
+			       || (readw(&i2c->con) & I2C_CON_MST)) {
 				udelay (10000);
-				writew (0xFFFF, I2C_STAT);
+				writew(0xFFFF, &i2c->stat);
 			}
 		}
 	}
 	flush_fifo();
-	writew (0xFFFF, I2C_STAT);
-	writew (0, I2C_CNT);
+	writew(0xFFFF, &i2c->stat);
+	writew(0, &i2c->cnt);
 	return i2c_error;
 }
 
@@ -203,12 +211,12 @@ static int i2c_write_byte (u8 devaddr, u8 regoffset, u8 value)
 	wait_for_bb ();
 
 	/* two bytes */
-	writew (2, I2C_CNT);
+	writew(2, &i2c->cnt);
 	/* set slave address */
-	writew (devaddr, I2C_SA);
+	writew(devaddr, &i2c->sa);
 	/* stop bit needed here */
-	writew (I2C_CON_EN | I2C_CON_MST | I2C_CON_STT | I2C_CON_TRX |
-		I2C_CON_STP, I2C_CON);
+	writew(I2C_CON_EN | I2C_CON_MST | I2C_CON_STT | I2C_CON_TRX |
+		I2C_CON_STP, &i2c->con);
 
 	/* wait until state change */
 	status = wait_for_pin ();
@@ -216,24 +224,24 @@ static int i2c_write_byte (u8 devaddr, u8 regoffset, u8 value)
 	if (status & I2C_STAT_XRDY) {
 #if defined(CONFIG_OMAP243X) || defined(CONFIG_OMAP34XX)
 		/* send out 1 byte */
-		writeb (regoffset, I2C_DATA);
-		writew (I2C_STAT_XRDY, I2C_STAT);
+		writeb(regoffset, &i2c->data);
+		writew(I2C_STAT_XRDY, &i2c->stat);
 
 		status = wait_for_pin ();
 		if ((status & I2C_STAT_XRDY)) {
 			/* send out next 1 byte */
-			writeb (value, I2C_DATA);
-			writew (I2C_STAT_XRDY, I2C_STAT);
+			writeb(value, &i2c->data);
+			writew(I2C_STAT_XRDY, &i2c->stat);
 		} else {
 			i2c_error = 1;
 		}
 #else
 		/* send out two bytes */
-		writew ((value << 8) + regoffset, I2C_DATA);
+		writew((value << 8) + regoffset, &i2c->data);
 #endif
 		/* must have enough delay to allow BB bit to go low */
-		udelay (50000);
-		if (readw (I2C_STAT) & I2C_STAT_NACK) {
+		udelay(50000);
+		if (readw(&i2c->stat) & I2C_STAT_NACK) {
 			i2c_error = 1;
 		}
 	} else {
@@ -243,18 +251,18 @@ static int i2c_write_byte (u8 devaddr, u8 regoffset, u8 value)
 	if (!i2c_error) {
 		int eout = 200;
 
-		writew (I2C_CON_EN, I2C_CON);
-		while ((stat = readw (I2C_STAT)) || (readw (I2C_CON) & I2C_CON_MST)) {
-			udelay (1000);
+		writew(I2C_CON_EN, &i2c->con);
+		while ((stat = readw(&i2c->stat)) || (readw(&i2c->con) & I2C_CON_MST)) {
+			udelay(1000);
 			/* have to read to clear intrrupt */
-			writew (0xFFFF, I2C_STAT);
+			writew(0xFFFF, &i2c->stat);
 			if(--eout == 0) /* better leave with error than hang */
 				break;
 		}
 	}
 	flush_fifo();
-	writew (0xFFFF, I2C_STAT);
-	writew (0, I2C_CNT);
+	writew(0xFFFF, &i2c->stat);
+	writew(0, &i2c->cnt);
 	return i2c_error;
 }
 
@@ -265,14 +273,14 @@ static void flush_fifo(void)
 	 * you get a bus error
 	 */
 	while(1){
-		stat = readw(I2C_STAT);
+		stat = readw(&i2c->stat);
 		if(stat == I2C_STAT_RRDY){
 #if defined(CONFIG_OMAP243X) || defined(CONFIG_OMAP34XX)
-			readb(I2C_DATA);
+			readb(&i2c->data);
 #else
-			readw(I2C_DATA);
+			readw(&i2c->data);
 #endif
-			writew(I2C_STAT_RRDY,I2C_STAT);
+			writew(I2C_STAT_RRDY, &i2c->stat);
 			udelay(1000);
 		}else
 			break;
@@ -283,7 +291,7 @@ int i2c_probe (uchar chip)
 {
 	int res = 1; /* default = fail */
 
-	if (chip == readw (I2C_OA)) {
+	if (chip == readw(&i2c->oa)) {
 		return res;
 	}
 
@@ -291,27 +299,27 @@ int i2c_probe (uchar chip)
 	wait_for_bb ();
 
 	/* try to read one byte */
-	writew (1, I2C_CNT);
+	writew(1, &i2c->cnt);
 	/* set slave address */
-	writew (chip, I2C_SA);
+	writew(chip, &i2c->sa);
 	/* stop bit needed here */
-	writew (I2C_CON_EN | I2C_CON_MST | I2C_CON_STT | I2C_CON_STP, I2C_CON);
+	writew(I2C_CON_EN | I2C_CON_MST | I2C_CON_STT | I2C_CON_STP, &i2c->con);
 	/* enough delay for the NACK bit set */
-	udelay (50000);
+	udelay(50000);
 
-	if (!(readw (I2C_STAT) & I2C_STAT_NACK)) {
+	if (!(readw(&i2c->stat) & I2C_STAT_NACK)) {
 		res = 0;      /* success case */
 		flush_fifo();
-		writew(0xFFFF, I2C_STAT);
+		writew(0xFFFF, &i2c->stat);
 	} else {
-		writew(0xFFFF, I2C_STAT);	 /* failue, clear sources*/
-		writew (readw (I2C_CON) | I2C_CON_STP, I2C_CON); /* finish up xfer */
+		writew(0xFFFF, &i2c->stat);	 /* failue, clear sources*/
+		writew(readw(&i2c->con) | I2C_CON_STP, &i2c->con); /* finish up xfer */
 		udelay(20000);
 		wait_for_bb ();
 	}
 	flush_fifo();
-	writew (0, I2C_CNT); /* don't allow any more data in...we don't want it.*/
-	writew(0xFFFF, I2C_STAT);
+	writew(0, &i2c->cnt); /* don't allow any more data in...we don't want it.*/
+	writew(0xFFFF, &i2c->stat);
 	return res;
 }
 
@@ -345,18 +353,18 @@ int i2c_write (uchar chip, uint addr, int alen, uchar * buffer, int len)
 	int i;
 
 	if (alen > 1) {
-		printf ("I2C read: addr len %d not supported\n", alen);
+		printf ("I2C write: addr len %d not supported\n", alen);
 		return 1;
 	}
 
 	if (addr + len > 256) {
-		printf ("I2C read: address out of range\n");
+		printf ("I2C write: address out of range\n");
 		return 1;
 	}
 
 	for (i = 0; i < len; i++) {
 		if (i2c_write_byte (chip, addr + i, buffer[i])) {
-			printf ("I2C read: I/O error\n");
+			printf ("I2C write: I/O error\n");
 			i2c_init (CONFIG_SYS_I2C_SPEED, CONFIG_SYS_I2C_SLAVE);
 			return 1;
 		}
@@ -370,17 +378,17 @@ static void wait_for_bb (void)
 	int timeout = 10;
 	u16 stat;
 
-	writew(0xFFFF, I2C_STAT);	 /* clear current interruts...*/
-	while ((stat = readw (I2C_STAT) & I2C_STAT_BB) && timeout--) {
-		writew (stat, I2C_STAT);
-		udelay (50000);
+	writew(0xFFFF, &i2c->stat);	 /* clear current interruts...*/
+	while ((stat = readw(&i2c->stat) & I2C_STAT_BB) && timeout--) {
+		writew(stat, &i2c->stat);
+		udelay(50000);
 	}
 
 	if (timeout <= 0) {
 		printf ("timed out in wait_for_bb: I2C_STAT=%x\n",
-			readw (I2C_STAT));
+			readw(&i2c->stat));
 	}
-	writew(0xFFFF, I2C_STAT);	 /* clear delayed stuff*/
+	writew(0xFFFF, &i2c->stat);	 /* clear delayed stuff*/
 }
 
 static u16 wait_for_pin (void)
@@ -390,7 +398,7 @@ static u16 wait_for_pin (void)
 
 	do {
 		udelay (1000);
-		status = readw (I2C_STAT);
+		status = readw(&i2c->stat);
 	} while (  !(status &
 		   (I2C_STAT_ROVR | I2C_STAT_XUDF | I2C_STAT_XRDY |
 		    I2C_STAT_RRDY | I2C_STAT_ARDY | I2C_STAT_NACK |
@@ -398,8 +406,58 @@ static u16 wait_for_pin (void)
 
 	if (timeout <= 0) {
 		printf ("timed out in wait_for_pin: I2C_STAT=%x\n",
-			readw (I2C_STAT));
-			writew(0xFFFF, I2C_STAT);
+			readw(&i2c->stat));
+			writew(0xFFFF, &i2c->stat);
 }
 	return status;
 }
+
+int i2c_set_bus_num(unsigned int bus)
+{
+	if ((bus < 0) || (bus >= I2C_NUM_IF)) {
+		printf("Bad bus ID-%d\n", bus);
+		return -1;
+	}
+
+#if defined(CONFIG_OMAP34XX)
+	if (bus == 2)
+		i2c = (i2c_t *)I2C_BASE3;
+	else
+#endif
+	if (bus == 1)
+		i2c = (i2c_t *)I2C_BASE2;
+	else
+		i2c = (i2c_t *)I2C_BASE1;
+
+	return 0;
+}
+
+unsigned int i2c_get_bus_num(void)
+{
+	if (i2c == (i2c_t *)I2C_BASE1)
+		return 0;
+	else
+	if (i2c == (i2c_t *)I2C_BASE2)
+		return 1;
+#if defined(CONFIG_OMAP34XX)
+	else
+	if (i2c == (i2c_t *)I2C_BASE3)
+		return 2;
+#endif
+	else
+		return 0xFFFFFFFF;
+}
+
+/*
+ * To be Implemented
+ */
+int i2c_set_bus_speed(unsigned int speed)
+{
+	return 0;
+}
+
+unsigned int i2c_get_bus_speed(void)
+{
+	return 0;
+}
+
include/asm-arm/arch-omap24xx/i2c.h +36 -16

60 lines not on this page · raw message

include/asm-arm/arch-omap3/i2c.h +33 -15

54 lines not on this page · raw message

include/configs/omap3_beagle.h +1 -0
index 19a5ec9..d5a0d49 100644
@@ -128,6 +128,7 @@
 #define CONFIG_SYS_I2C_BUS		0
 #define CONFIG_SYS_I2C_BUS_SELECT	1
 #define CONFIG_DRIVER_OMAP34XX_I2C	1
+#define CONFIG_I2C_MULTI_BUS		1
 
 /*
  * TWL4030
Signature · 1 line
1.6.0.6
#2 lore raw

PATCH 1/1 Add I2C multibus support for OMAP2/3 boards

Quoted 10 lines · Tom Rix

From: Syed Mohammed Khasim <[email protected]>

This was cherry-picked from

repo: http://www.beagleboard.org/u-boot-arm.git
commit: 52eddcd07c2e7ad61d15bab2cf2d0d21466eaca2

In addition to adding multibus support, this patch also cleans up the register access. The register access has been changed from #defines to a structure.

Have you looked at my proposal I sent some hours before your patch?

http://lists.denx.de/pipermail/u-boot/2009-October/063556.html

Quoted 24 lines
---
 drivers/i2c/omap24xx_i2c.c          |  220 ++++++++++++++++++++++-------------
 include/asm-arm/arch-omap24xx/i2c.h |   52 ++++++---
 include/asm-arm/arch-omap3/i2c.h    |   48 +++++---
 include/configs/omap3_beagle.h      |    1 +
 4 files changed, 209 insertions(+), 112 deletions(-)
drivers/i2c/omap24xx_i2c.c +1 -1
index 1a4c8c9..763d2f8 100644
@@ -1,7 +1,7 @@
 /*
  * Basic I2C functions
  *
- * Copyright (c) 2004 Texas Instruments
+ * Copyright (c) 2004, 2009 Texas Instruments
  *
  * This package is free software;  you can redistribute it and/or
  * modify it under the terms of the license found in the file

drivers/i2c/omap24xx_i2c.c line 32

Quoted 8 lines
drivers/i2c/omap24xx_i2c.c +5 -0
@@ -25,10 +25,18 @@
 #include <asm/arch/i2c.h>
 #include <asm/io.h>
 
+#ifdef CONFIG_OMAP34XX
+#define I2C_NUM_IF 3
+#else
+#define I2C_NUM_IF 2
+#endif

I prefer something like I2C_BUS_MAX for this. And move it to header file. Moving it OMAP2 and OMAP3 i2c.h headers will remove the #ifdef, too.

drivers/i2c/omap24xx_i2c.c line 38

Quoted 5 lines
drivers/i2c/omap24xx_i2c.c +1 -0
 static void wait_for_bb (void);
 static u16 wait_for_pin (void);
 static void flush_fifo(void);
 
+static i2c_t *i2c = (i2c_t *)I2C_DEFAULT_BASE;

Looking at the other OMAP3 files, the recent syntax seems to be

static struct i2c *i2c_base = (struct i2c *)I2C_DEFAULT_BASE;

Just to be consistent.

Quoted 9 lines
drivers/i2c/omap24xx_i2c.c +1 -1
 void i2c_init (int speed, int slaveadd)
 {
 	int psc, fsscll, fssclh;
@@ -95,30 +103,30 @@ void i2c_init (int speed, int slaveadd)
 		sclh = (unsigned int)fssclh;
 	}
 
-	writew(0x2, I2C_SYSC); /* for ES2 after soft reset */
+	writew(0x2, &i2c->sysc);		/* for ES2 after soft reset */

drivers/i2c/omap24xx_i2c.c void i2c_init (int speed, int slaveadd) · line 112

Quoted 10 lines
drivers/i2c/omap24xx_i2c.c +4 -4
 	udelay(1000);
-	writew(0x0, I2C_SYSC); /* will probably self clear but */
+	writew(0x0, &i2c->sysc);		/* will probably self clear but */
 
-	if (readw (I2C_CON) & I2C_CON_EN) {
-		writew (0, I2C_CON);
-		udelay (50000);
+	if (readw(&i2c->con) & I2C_CON_EN) {
+		writew(0, &i2c->con);
+		udelay(50000);

udelay: Are this formatting changes? You are correct, a lot of formatting improvements are needed here. But they should go to an other patch.

Quoted 10 lines
drivers/i2c/omap24xx_i2c.c +3 -5
-	writew(psc, I2C_PSC);
-	writew(scll, I2C_SCLL);
-	writew(sclh, I2C_SCLH);
+	writew(psc, &i2c->psc);
+	writew(scll, &i2c->scll);
+	writew(sclh, &i2c->sclh);
 
 	/* own address */
-	writew (slaveadd, I2C_OA);
-	writew (I2C_CON_EN, I2C_CON);

drivers/i2c/omap24xx_i2c.c void i2c_init (int speed, int slaveadd) · line 126

Quoted 10 lines
drivers/i2c/omap24xx_i2c.c +5 -3
+	writew(slaveadd, &i2c->oa);
+	writew(I2C_CON_EN, &i2c->con);
 
 	/* have to enable intrrupts or OMAP i2c module doesn't work */
-	writew (I2C_IE_XRDY_IE | I2C_IE_RRDY_IE | I2C_IE_ARDY_IE |
-		I2C_IE_NACK_IE | I2C_IE_AL_IE, I2C_IE);
-	udelay (1000);
+	writew(I2C_IE_XRDY_IE | I2C_IE_RRDY_IE | I2C_IE_ARDY_IE |
+		I2C_IE_NACK_IE | I2C_IE_AL_IE, &i2c->ie);
+	udelay(1000);

Formatting change?

Quoted 20 lines
drivers/i2c/omap24xx_i2c.c +5 -5
 	flush_fifo();
-	writew (0xFFFF, I2C_STAT);
-	writew (0, I2C_CNT);
+	writew(0xFFFF, &i2c->stat);
+	writew(0, &i2c->cnt);
 }
 
 static int i2c_read_byte (u8 devaddr, u8 regoffset, u8 * value)
@@ -130,19 +138,19 @@ static int i2c_read_byte (u8 devaddr, u8 regoffset, u8 * value)
 	wait_for_bb ();
 
 	/* one byte only */
-	writew (1, I2C_CNT);
+	writew(1, &i2c->cnt);
 	/* set slave address */
-	writew (devaddr, I2C_SA);
+	writew(devaddr, &i2c->sa);
 	/* no stop bit needed here */
-	writew (I2C_CON_EN | I2C_CON_MST | I2C_CON_STT | I2C_CON_TRX, I2C_CON);
+	writew(I2C_CON_EN | I2C_CON_MST | I2C_CON_STT | I2C_CON_TRX, &i2c->con);

drivers/i2c/omap24xx_i2c.c static int i2c_read_byte (u8 devaddr, u8 regoffset, u8 * value) · line 152

Quoted 10 lines
drivers/i2c/omap24xx_i2c.c +2 -3
 
 	status = wait_for_pin ();
 
 	if (status & I2C_STAT_XRDY) {
 		/* Important: have to use byte access */
-		writeb (regoffset, I2C_DATA);
-		udelay (20000);
-		if (readw (I2C_STAT) & I2C_STAT_NACK) {
+		writeb(regoffset, &i2c->data);
+		udelay(20000);

Formatting change?

Quoted 116 lines
drivers/i2c/omap24xx_i2c.c +27 -26
+		if (readw(&i2c->stat) & I2C_STAT_NACK) {
 			i2c_error = 1;
 		}
 	} else {
@@ -151,28 +159,28 @@ static int i2c_read_byte (u8 devaddr, u8 regoffset, u8 * value)
 
 	if (!i2c_error) {
 		/* free bus, otherwise we can't use a combined transction */
-		writew (0, I2C_CON);
-		while (readw (I2C_STAT) || (readw (I2C_CON) & I2C_CON_MST)) {
+		writew(0, &i2c->con);
+		while (readw(&i2c->stat) || (readw(&i2c->con) & I2C_CON_MST)) {
 			udelay (10000);
 			/* Have to clear pending interrupt to clear I2C_STAT */
-			writew (0xFFFF, I2C_STAT);
+			writew(0xFFFF, &i2c->stat);
 		}
 
 		wait_for_bb ();
 		/* set slave address */
-		writew (devaddr, I2C_SA);
+		writew(devaddr, &i2c->sa);
 		/* read one byte from slave */
-		writew (1, I2C_CNT);
+		writew(1, &i2c->cnt);
 		/* need stop bit here */
-		writew (I2C_CON_EN | I2C_CON_MST | I2C_CON_STT | I2C_CON_STP,
-			I2C_CON);
+		writew(I2C_CON_EN | I2C_CON_MST | I2C_CON_STT | I2C_CON_STP,
+			&i2c->con);
 
 		status = wait_for_pin ();
 		if (status & I2C_STAT_RRDY) {
 #if defined(CONFIG_OMAP243X) || defined(CONFIG_OMAP34XX)
-			*value = readb (I2C_DATA);
+			*value = readb(&i2c->data);
 #else
-			*value = readw (I2C_DATA);
+			*value = readw(&i2c->data);
 #endif
 			udelay (20000);
 		} else {
@@ -180,17 +188,17 @@ static int i2c_read_byte (u8 devaddr, u8 regoffset, u8 * value)
 		}
 
 		if (!i2c_error) {
-			writew (I2C_CON_EN, I2C_CON);
-			while (readw (I2C_STAT)
-			       || (readw (I2C_CON) & I2C_CON_MST)) {
+			writew(I2C_CON_EN, &i2c->con);
+			while (readw(&i2c->stat)
+			       || (readw(&i2c->con) & I2C_CON_MST)) {
 				udelay (10000);
-				writew (0xFFFF, I2C_STAT);
+				writew(0xFFFF, &i2c->stat);
 			}
 		}
 	}
 	flush_fifo();
-	writew (0xFFFF, I2C_STAT);
-	writew (0, I2C_CNT);
+	writew(0xFFFF, &i2c->stat);
+	writew(0, &i2c->cnt);
 	return i2c_error;
 }
 
@@ -203,12 +211,12 @@ static int i2c_write_byte (u8 devaddr, u8 regoffset, u8 value)
 	wait_for_bb ();
 
 	/* two bytes */
-	writew (2, I2C_CNT);
+	writew(2, &i2c->cnt);
 	/* set slave address */
-	writew (devaddr, I2C_SA);
+	writew(devaddr, &i2c->sa);
 	/* stop bit needed here */
-	writew (I2C_CON_EN | I2C_CON_MST | I2C_CON_STT | I2C_CON_TRX |
-		I2C_CON_STP, I2C_CON);
+	writew(I2C_CON_EN | I2C_CON_MST | I2C_CON_STT | I2C_CON_TRX |
+		I2C_CON_STP, &i2c->con);
 
 	/* wait until state change */
 	status = wait_for_pin ();
@@ -216,24 +224,24 @@ static int i2c_write_byte (u8 devaddr, u8 regoffset, u8 value)
 	if (status & I2C_STAT_XRDY) {
 #if defined(CONFIG_OMAP243X) || defined(CONFIG_OMAP34XX)
 		/* send out 1 byte */
-		writeb (regoffset, I2C_DATA);
-		writew (I2C_STAT_XRDY, I2C_STAT);
+		writeb(regoffset, &i2c->data);
+		writew(I2C_STAT_XRDY, &i2c->stat);
 
 		status = wait_for_pin ();
 		if ((status & I2C_STAT_XRDY)) {
 			/* send out next 1 byte */
-			writeb (value, I2C_DATA);
-			writew (I2C_STAT_XRDY, I2C_STAT);
+			writeb(value, &i2c->data);
+			writew(I2C_STAT_XRDY, &i2c->stat);
 		} else {
 			i2c_error = 1;
 		}
 #else
 		/* send out two bytes */
-		writew ((value << 8) + regoffset, I2C_DATA);
+		writew((value << 8) + regoffset, &i2c->data);
 #endif
 		/* must have enough delay to allow BB bit to go low */
-		udelay (50000);
-		if (readw (I2C_STAT) & I2C_STAT_NACK) {
+		udelay(50000);
+		if (readw(&i2c->stat) & I2C_STAT_NACK) {
 			i2c_error = 1;
 		}
 	} else {

drivers/i2c/omap24xx_i2c.c static int i2c_write_byte (u8 devaddr, u8 regoffset, u8 value) · line 256

Quoted 9 lines
drivers/i2c/omap24xx_i2c.c +3 -3
@@ -243,18 +251,18 @@ static int i2c_write_byte (u8 devaddr, u8 regoffset, u8 value)
 	if (!i2c_error) {
 		int eout = 200;
 
-		writew (I2C_CON_EN, I2C_CON);
-		while ((stat = readw (I2C_STAT)) || (readw (I2C_CON) & I2C_CON_MST)) {
-			udelay (1000);
+		writew(I2C_CON_EN, &i2c->con);
+		while ((stat = readw(&i2c->stat)) || (readw(&i2c->con) & I2C_CON_MST)) {
+			udelay(1000);

Formatting change?

Quoted 48 lines
drivers/i2c/omap24xx_i2c.c +8 -9
 			/* have to read to clear intrrupt */
-			writew (0xFFFF, I2C_STAT);
+			writew(0xFFFF, &i2c->stat);
 			if(--eout == 0) /* better leave with error than hang */
 				break;
 		}
 	}
 	flush_fifo();
-	writew (0xFFFF, I2C_STAT);
-	writew (0, I2C_CNT);
+	writew(0xFFFF, &i2c->stat);
+	writew(0, &i2c->cnt);
 	return i2c_error;
 }
 
@@ -265,14 +273,14 @@ static void flush_fifo(void)
 	 * you get a bus error
 	 */
 	while(1){
-		stat = readw(I2C_STAT);
+		stat = readw(&i2c->stat);
 		if(stat == I2C_STAT_RRDY){
 #if defined(CONFIG_OMAP243X) || defined(CONFIG_OMAP34XX)
-			readb(I2C_DATA);
+			readb(&i2c->data);
 #else
-			readw(I2C_DATA);
+			readw(&i2c->data);
 #endif
-			writew(I2C_STAT_RRDY,I2C_STAT);
+			writew(I2C_STAT_RRDY, &i2c->stat);
 			udelay(1000);
 		}else
 			break;
@@ -283,7 +291,7 @@ int i2c_probe (uchar chip)
 {
 	int res = 1; /* default = fail */
 
-	if (chip == readw (I2C_OA)) {
+	if (chip == readw(&i2c->oa)) {
 		return res;
 	}
 
@@ -291,27 +299,27 @@ int i2c_probe (uchar chip)
 	wait_for_bb ();
 
 	/* try to read one byte */
-	writew (1, I2C_CNT);

drivers/i2c/omap24xx_i2c.c int i2c_probe (uchar chip) · line 308

Quoted 10 lines
drivers/i2c/omap24xx_i2c.c +4 -3
+	writew(1, &i2c->cnt);
 	/* set slave address */
-	writew (chip, I2C_SA);
+	writew(chip, &i2c->sa);
 	/* stop bit needed here */
-	writew (I2C_CON_EN | I2C_CON_MST | I2C_CON_STT | I2C_CON_STP, I2C_CON);
+	writew(I2C_CON_EN | I2C_CON_MST | I2C_CON_STT | I2C_CON_STP, &i2c->con);
 	/* enough delay for the NACK bit set */
-	udelay (50000);
+	udelay(50000);

Formatting change?

Quoted 24 lines
drivers/i2c/omap24xx_i2c.c +6 -6
-	if (!(readw (I2C_STAT) & I2C_STAT_NACK)) {
+	if (!(readw(&i2c->stat) & I2C_STAT_NACK)) {
 		res = 0;      /* success case */
 		flush_fifo();
-		writew(0xFFFF, I2C_STAT);
+		writew(0xFFFF, &i2c->stat);
 	} else {
-		writew(0xFFFF, I2C_STAT);	 /* failue, clear sources*/
-		writew (readw (I2C_CON) | I2C_CON_STP, I2C_CON); /* finish up xfer */
+		writew(0xFFFF, &i2c->stat);	 /* failue, clear sources*/
+		writew(readw(&i2c->con) | I2C_CON_STP, &i2c->con); /* finish up xfer */
 		udelay(20000);
 		wait_for_bb ();
 	}
 	flush_fifo();
-	writew (0, I2C_CNT); /* don't allow any more data in...we don't want it.*/
-	writew(0xFFFF, I2C_STAT);
+	writew(0, &i2c->cnt); /* don't allow any more data in...we don't want it.*/
+	writew(0xFFFF, &i2c->stat);
 	return res;
 }
 
@@ -345,18 +353,18 @@ int i2c_write (uchar chip, uint addr, int alen, uchar * buffer, int len)
 	int i;

drivers/i2c/omap24xx_i2c.c int i2c_write (uchar chip, uint addr, int alen, uchar * buffer, int len) · line 361

Quoted 10 lines
drivers/i2c/omap24xx_i2c.c +2 -2
 
 	if (alen > 1) {
-		printf ("I2C read: addr len %d not supported\n", alen);
+		printf ("I2C write: addr len %d not supported\n", alen);
 		return 1;
 	}
 
 	if (addr + len > 256) {
-		printf ("I2C read: address out of range\n");
+		printf ("I2C write: address out of range\n");

Here and above, typo change?

drivers/i2c/omap24xx_i2c.c int i2c_write (uchar chip, uint addr, int alen, uchar * buffer, int len) · line 367

Quoted 7 lines
drivers/i2c/omap24xx_i2c.c +1 -1
 		return 1;
 	}
 
 	for (i = 0; i < len; i++) {
 		if (i2c_write_byte (chip, addr + i, buffer[i])) {
-			printf ("I2C read: I/O error\n");
+			printf ("I2C write: I/O error\n");

Typo change?

drivers/i2c/omap24xx_i2c.c static void wait_for_bb (void) · line 376

Quoted 11 lines
drivers/i2c/omap24xx_i2c.c +0 -4
 			i2c_init (CONFIG_SYS_I2C_SPEED, CONFIG_SYS_I2C_SLAVE);
 			return 1;
 		}
@@ -370,17 +378,17 @@ static void wait_for_bb (void)
 	int timeout = 10;
 	u16 stat;
 
-	writew(0xFFFF, I2C_STAT);	 /* clear current interruts...*/
-	while ((stat = readw (I2C_STAT) & I2C_STAT_BB) && timeout--) {
-		writew (stat, I2C_STAT);
-		udelay (50000);

Formatting change?

Quoted 28 lines
drivers/i2c/omap24xx_i2c.c +7 -3
+	writew(0xFFFF, &i2c->stat);	 /* clear current interruts...*/
+	while ((stat = readw(&i2c->stat) & I2C_STAT_BB) && timeout--) {
+		writew(stat, &i2c->stat);
+		udelay(50000);
 	}
 
 	if (timeout <= 0) {
 		printf ("timed out in wait_for_bb: I2C_STAT=%x\n",
-			readw (I2C_STAT));
+			readw(&i2c->stat));
 	}
-	writew(0xFFFF, I2C_STAT);	 /* clear delayed stuff*/
+	writew(0xFFFF, &i2c->stat);	 /* clear delayed stuff*/
 }
 
 static u16 wait_for_pin (void)
@@ -390,7 +398,7 @@ static u16 wait_for_pin (void)
 
 	do {
 		udelay (1000);
-		status = readw (I2C_STAT);
+		status = readw(&i2c->stat);
 	} while (  !(status &
 		   (I2C_STAT_ROVR | I2C_STAT_XUDF | I2C_STAT_XRDY |
 		    I2C_STAT_RRDY | I2C_STAT_ARDY | I2C_STAT_NACK |
@@ -398,8 +406,58 @@ static u16 wait_for_pin (void)
 
 	if (timeout <= 0) {

drivers/i2c/omap24xx_i2c.c static u16 wait_for_pin (void) · line 415

Quoted 10 lines
drivers/i2c/omap24xx_i2c.c +4 -2
 		printf ("timed out in wait_for_pin: I2C_STAT=%x\n",
-			readw (I2C_STAT));
-			writew(0xFFFF, I2C_STAT);
+			readw(&i2c->stat));
+			writew(0xFFFF, &i2c->stat);
 }
 	return status;
 }
+
+int i2c_set_bus_num(unsigned int bus)

Do we need an extern declaration in i2c.h for this? To be able to call it from somewhere else without warning?

drivers/i2c/omap24xx_i2c.c static u16 wait_for_pin (void) · line 417

Quoted 2 lines
drivers/i2c/omap24xx_i2c.c +2 -0
+{
+	if ((bus < 0) || (bus >= I2C_NUM_IF)) {

As mentioned above, I'd like something like bus max here.

Quoted 6 lines
drivers/i2c/omap24xx_i2c.c +6 -0
+		printf("Bad bus ID-%d\n", bus);
+		return -1;
+	}
+
+#if defined(CONFIG_OMAP34XX)
+	if (bus == 2)

drivers/i2c/omap24xx_i2c.c static u16 wait_for_pin (void) · line 433

Quoted 10 lines
drivers/i2c/omap24xx_i2c.c +10 -0
+		i2c = (i2c_t *)I2C_BASE3;
+	else
+#endif
+	if (bus == 1)
+		i2c = (i2c_t *)I2C_BASE2;
+	else
+		i2c = (i2c_t *)I2C_BASE1;
+
+	return 0;
+}

I would remove the following three functions as they are not needed at the moment (i.e. without command line interface).

include/asm-arm/arch-omap24xx/i2c.h

Quoted 33 lines
drivers/i2c/omap24xx_i2c.c +29 -0
+unsigned int i2c_get_bus_num(void)
+{
+	if (i2c == (i2c_t *)I2C_BASE1)
+		return 0;
+	else
+	if (i2c == (i2c_t *)I2C_BASE2)
+		return 1;
+#if defined(CONFIG_OMAP34XX)
+	else
+	if (i2c == (i2c_t *)I2C_BASE3)
+		return 2;
+#endif
+	else
+		return 0xFFFFFFFF;
+}
+
+/*
+ * To be Implemented
+ */
+int i2c_set_bus_speed(unsigned int speed)
+{
+	return 0;
+}
+
+unsigned int i2c_get_bus_speed(void)
+{
+	return 0;
+}
+
include/asm-arm/arch-omap24xx/i2c.h +0 -0
index 44db7a2..832a0e1 100644

Ah, thanks for pointing to this. I missed this.

include/asm-arm/arch-omap24xx/i2c.h line 27

Quoted 7 lines
include/asm-arm/arch-omap24xx/i2c.h +1 -1
@@ -23,24 +23,44 @@
 #ifndef _OMAP24XX_I2C_H_
 #define _OMAP24XX_I2C_H_
 
-#define I2C_BASE                0x48070000
+#define I2C_BASE1               0x48070000
 #define I2C_BASE2               0x48072000 /* nothing hooked up on h4 */

I wonder why arch-omap24xx/i2c.h has no I2C_DEFAULT_BASE (used above now).

Quoted 8 lines
include/asm-arm/arch-omap24xx/i2c.h +0 -8
-#define I2C_REV                 (I2C_BASE + 0x00)
-#define I2C_IE                  (I2C_BASE + 0x04)
-#define I2C_STAT                (I2C_BASE + 0x08)
-#define I2C_IV                  (I2C_BASE + 0x0c)
-#define I2C_BUF                 (I2C_BASE + 0x14)
-#define I2C_CNT                 (I2C_BASE + 0x18)
-#define I2C_DATA                (I2C_BASE + 0x1c)
-#define I2C_SYSC                (I2C_BASE + 0x20)

include/asm-arm/arch-omap24xx/i2c.h line 31

Quoted 10 lines
include/asm-arm/arch-omap24xx/i2c.h +3 -7
-#define I2C_CON                 (I2C_BASE + 0x24)
-#define I2C_OA                  (I2C_BASE + 0x28)
-#define I2C_SA                  (I2C_BASE + 0x2c)
-#define I2C_PSC                 (I2C_BASE + 0x30)
-#define I2C_SCLL                (I2C_BASE + 0x34)
-#define I2C_SCLH                (I2C_BASE + 0x38)
-#define I2C_SYSTEST             (I2C_BASE + 0x3c)
+#define I2C_DEFAULT_BASE	I2C_BASE1
+
+typedef struct i2c {

Seems the recent syntax is without typedef.

include/asm-arm/arch-omap24xx/i2c.h line 32

Quoted 2 lines
include/asm-arm/arch-omap24xx/i2c.h +2 -0
+	unsigned short rev;		/* 0x00 */
+	unsigned short pad_rev;		/* 0x02 */

Looking at

include/asm-arm/arch-omap3/cpu.h

the current style is to use resX for padding. Like done below with res1 already. We should stay consistent here.

include/asm-arm/arch-omap24xx/i2c.h line 40

Quoted 7 lines
include/asm-arm/arch-omap24xx/i2c.h +7 -0
+	unsigned short ie;		/* 0x04 */
+	unsigned short pad_ie;		/* 0x06 */
+	unsigned short stat;		/* 0x08 */
+	unsigned short pad_stat;	/* 0x0a */
+	unsigned short iv;		/* 0x0c */
+	unsigned short pad_iv;		/* 0x0e */
+	unsigned int res1;		/* 0x10 */

unsigned short resX[3] instead?

Quoted 40 lines
include/asm-arm/arch-omap24xx/i2c.h +23 -0
+	unsigned short buf;		/* 0x14 */
+	unsigned short pad_buf;		/* 0x16 */
+	unsigned short cnt;		/* 0x18 */
+	unsigned short pad_cnt;		/* 0x1a */
+	unsigned short data;		/* 0x1c */
+	unsigned short pad_data;	/* 0x1e */
+	unsigned short sysc;		/* 0x20 */
+	unsigned short pad_sysc;	/* 0x22 */
+	unsigned short con;		/* 0x24 */
+	unsigned short pad_con;		/* 0x26 */
+	unsigned short oa;		/* 0x28 */
+	unsigned short pad_oa;		/* 0x2a */
+	unsigned short sa;		/* 0x2c */
+	unsigned short pad_sa;		/* 0x2e */
+	unsigned short psc;		/* 0x30 */
+	unsigned short pad_psc;		/* 0x32 */
+	unsigned short scll;		/* 0x34 */
+	unsigned short pad_scll;	/* 0x36 */
+	unsigned short sclh;		/* 0x38 */
+	unsigned short pad_sclh;	/* 0x3a */
+	unsigned short systest;		/* 0x3c */
+	unsigned short pad_systest;	/* 0x3e */
+} i2c_t;
 
 /* I2C masks */
 
include/asm-arm/arch-omap3/i2c.h +0 -6
index 8b339cc..3da1fcf 100644
@@ -25,21 +25,39 @@
 
 #define I2C_DEFAULT_BASE	I2C_BASE1
 
-#define I2C_REV		(I2C_DEFAULT_BASE + 0x00)
-#define I2C_IE 		(I2C_DEFAULT_BASE + 0x04)
-#define I2C_STAT	(I2C_DEFAULT_BASE + 0x08)
-#define I2C_IV 		(I2C_DEFAULT_BASE + 0x0c)
-#define I2C_BUF		(I2C_DEFAULT_BASE + 0x14)
-#define I2C_CNT		(I2C_DEFAULT_BASE + 0x18)

include/asm-arm/arch-omap3/i2c.h line 28

Quoted 10 lines
include/asm-arm/arch-omap3/i2c.h +1 -9
-#define I2C_DATA	(I2C_DEFAULT_BASE + 0x1c)
-#define I2C_SYSC	(I2C_DEFAULT_BASE + 0x20)
-#define I2C_CON		(I2C_DEFAULT_BASE + 0x24)
-#define I2C_OA 		(I2C_DEFAULT_BASE + 0x28)
-#define I2C_SA 		(I2C_DEFAULT_BASE + 0x2c)
-#define I2C_PSC		(I2C_DEFAULT_BASE + 0x30)
-#define I2C_SCLL	(I2C_DEFAULT_BASE + 0x34)
-#define I2C_SCLH	(I2C_DEFAULT_BASE + 0x38)
-#define I2C_SYSTEST	(I2C_DEFAULT_BASE + 0x3c)
+typedef struct i2c {

Comments from above apply here, too.

Quoted 40 lines
include/asm-arm/arch-omap3/i2c.h +32 -0
+	unsigned short rev;		/* 0x00 */
+	unsigned short pad_rev;		/* 0x02 */
+	unsigned short ie;		/* 0x04 */
+	unsigned short pad_ie;		/* 0x06 */
+	unsigned short stat;		/* 0x08 */
+	unsigned short pad_stat;	/* 0x0a */
+	unsigned short iv;		/* 0x0c */
+	unsigned short pad_iv;		/* 0x0e */
+	unsigned int res1;		/* 0x10 */
+	unsigned short buf;		/* 0x14 */
+	unsigned short pad_buf;		/* 0x16 */
+	unsigned short cnt;		/* 0x18 */
+	unsigned short pad_cnt;		/* 0x1a */
+	unsigned short data;		/* 0x1c */
+	unsigned short pad_data;	/* 0x1e */
+	unsigned short sysc;		/* 0x20 */
+	unsigned short pad_sysc;	/* 0x22 */
+	unsigned short con;		/* 0x24 */
+	unsigned short pad_con;		/* 0x26 */
+	unsigned short oa;		/* 0x28 */
+	unsigned short pad_oa;		/* 0x2a */
+	unsigned short sa;		/* 0x2c */
+	unsigned short pad_sa;		/* 0x2e */
+	unsigned short psc;		/* 0x30 */
+	unsigned short pad_psc;		/* 0x32 */
+	unsigned short scll;		/* 0x34 */
+	unsigned short pad_scll;	/* 0x36 */
+	unsigned short sclh;		/* 0x38 */
+	unsigned short pad_sclh;	/* 0x3a */
+	unsigned short systest;		/* 0x3c */
+	unsigned short pad_systest;	/* 0x3e */
+} i2c_t;
 
 /* I2C masks */
 
include/configs/omap3_beagle.h +0 -0
index 19a5ec9..d5a0d49 100644

include/configs/omap3_beagle.h line 131

Quoted 4 lines
include/configs/omap3_beagle.h +1 -0
@@ -128,6 +128,7 @@
 #define CONFIG_SYS_I2C_BUS		0
 #define CONFIG_SYS_I2C_BUS_SELECT	1
 #define CONFIG_DRIVER_OMAP34XX_I2C	1
+#define CONFIG_I2C_MULTI_BUS		1

Not needed.

While all above are only style questions, what's about the main functionality topic:

Do we have to call i2c_init() for bus 1 and 2 if switching to it? If yes, who does it? For bus 0 it's already in the code. I'd like that the user doesn't have to care about it. Therefore, I added

static unsigned int bus_initialized[3] = {0, 0, 0}; static unsigned int current_bus = 0;

void i2c_init (int speed, int slaveadd) {
...
bus_initialized[current_bus] = 1;
}

int i2c_set_bus_num(unsigned int bus) {
...
current_bus = bus;

if(!bus_initialized[current_bus])

	i2c_init(CONFIG_SYS_I2C_SPEED, CONFIG_SYS_I2C_SLAVE);

...
}

Do you like to test my patch in attachment?

It is an extension of

http://lists.denx.de/pipermail/u-boot/2009-October/063556.html

to cover arch-omap24xx/i2c.h, too.

Compile tested for omap3_beagle and omap2420h4.

Best regards

Dirk
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: omap3_i2c_multibus_patch.txt
Url: http://lists.denx.de/pipermail/u-boot/attachments/20091101/62cb74ab/attachment.txt

author
#3 lore raw

PATCH 1/1 Add I2C multibus support for OMAP2/3 boards

Quoted 5 lines · Dirk Behme
Quoted 4 lines · Tom Rix

From: Syed Mohammed Khasim <[email protected]>

This was cherry-picked from

Quoted 8 lines
Quoted 6 lines

repo: http://www.beagleboard.org/u-boot-arm.git
commit: 52eddcd07c2e7ad61d15bab2cf2d0d21466eaca2

In addition to adding multibus support, this patch also cleans up the register access. The register access has been changed from #defines to a structure.

Have you looked at my proposal I sent some hours before your patch?

Sorry. I did not.
Not surprisingly it looks similar.

The formatting changes you mentioned were part of the original cherry-picked commit. I tried to leave as much of the original commit intact as possible. I ran but ignored the output of checkpatch.. The biggest complaint it has is with 80 char lines.

I looked over your patch.
It looks like an improvement to mine.

I'm all for dropping mine and using yours.
I will run test yours.

Do you have a zippy board that you can use to test the 2nd bus ?

Tom

author
#4 lore raw

PATCH 1/1 Add I2C multibus support for OMAP2/3 boards

Quoted 22 lines · Tom
Quoted 5 lines · Dirk Behme
Quoted 4 lines · Tom Rix

From: Syed Mohammed Khasim <[email protected]>

This was cherry-picked from

Quoted 7 lines
Quoted 6 lines

repo: http://www.beagleboard.org/u-boot-arm.git
commit: 52eddcd07c2e7ad61d15bab2cf2d0d21466eaca2

In addition to adding multibus support, this patch also cleans up the register access. The register access has been changed from #defines to a structure.

Have you looked at my proposal I sent some hours before your patch?

Sorry. I did not.
Not surprisingly it looks similar.

The formatting changes you mentioned were part of the original cherry-picked commit. I tried to leave as much of the original commit intact as possible. I ran but ignored the output of checkpatch.. The biggest complaint it has is with 80 char lines.

Quoted 8 lines

I looked over your patch.
It looks like an improvement to mine.

I'm all for dropping mine and using yours.
I will run test yours.

Do you have a zippy board that you can use to test the 2nd bus ?

Dirk,

Some feedback.

No regressions on MAKEALL arm

I ran tested your patch Zoom1 and Beagle.
Default i2c looks good.
I did not test the multibus support.

Looks good, please officially submit your patch.

Tom

Quoted 5 lines

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

#5 lore raw

PATCH 1/1 Add I2C multibus support for OMAP2/3 boards

In reply to Dirk Behme

Hello Dirk,

Quoted 44 lines · Dirk Behme
Quoted 10 lines · Tom Rix

From: Syed Mohammed Khasim <[email protected]>

This was cherry-picked from

repo: http://www.beagleboard.org/u-boot-arm.git
commit: 52eddcd07c2e7ad61d15bab2cf2d0d21466eaca2

In addition to adding multibus support, this patch also cleans up the register access. The register access has been changed from #defines to a structure.

Have you looked at my proposal I sent some hours before your patch?

http://lists.denx.de/pipermail/u-boot/2009-October/063556.html

Quoted 28 lines
---
 drivers/i2c/omap24xx_i2c.c          |  220
++++++++++++++++++++++-------------
 include/asm-arm/arch-omap24xx/i2c.h |   52 ++++++---
 include/asm-arm/arch-omap3/i2c.h    |   48 +++++---
 include/configs/omap3_beagle.h      |    1 +
 4 files changed, 209 insertions(+), 112 deletions(-)
drivers/i2c/omap24xx_i2c.c +1 -1
index 1a4c8c9..763d2f8 100644
@@ -1,7 +1,7 @@
 /*
  * Basic I2C functions
  *
- * Copyright (c) 2004 Texas Instruments
+ * Copyright (c) 2004, 2009 Texas Instruments
  *
  * This package is free software;  you can redistribute it and/or
  * modify it under the terms of the license found in the file
@@ -25,10 +25,18 @@
 #include <asm/arch/i2c.h>
 #include <asm/io.h>
 
Quoted 8 lines

drivers/i2c/omap24xx_i2c.c line 32

Quoted 5 lines
drivers/i2c/omap24xx_i2c.c +5 -0
+#ifdef CONFIG_OMAP34XX
+#define I2C_NUM_IF 3
+#else
+#define I2C_NUM_IF 2
+#endif

I prefer something like I2C_BUS_MAX for this. And move it to header file. Moving it OMAP2 and OMAP3 i2c.h headers will remove the #ifdef, too.

Yep, I2C_BUS_MAX would be better.

[...]

Quoted 16 lines

drivers/i2c/omap24xx_i2c.c static u16 wait_for_pin (void) · line 415

Quoted 13 lines
drivers/i2c/omap24xx_i2c.c +4 -2
@@ -398,8 +406,58 @@ static u16 wait_for_pin (void)
 
     if (timeout <= 0) {
         printf ("timed out in wait_for_pin: I2C_STAT=%x\n",
-            readw (I2C_STAT));
-            writew(0xFFFF, I2C_STAT);
+            readw(&i2c->stat));
+            writew(0xFFFF, &i2c->stat);
 }
     return status;
 }
+
+int i2c_set_bus_num(unsigned int bus)

Do we need an extern declaration in i2c.h for this? To be able to call it from somewhere else without warning?

Hmm.. isn;t it declared in i2c.h?

Quoted 16 lines

drivers/i2c/omap24xx_i2c.c static u16 wait_for_pin (void) · line 417

Quoted 2 lines
drivers/i2c/omap24xx_i2c.c +2 -0
+{
+    if ((bus < 0) || (bus >= I2C_NUM_IF)) {

As mentioned above, I'd like something like bus max here.

Quoted 11 lines
drivers/i2c/omap24xx_i2c.c +11 -0
+        printf("Bad bus ID-%d\n", bus);
+        return -1;
+    }
+
+#if defined(CONFIG_OMAP34XX)
+    if (bus == 2)
+        i2c = (i2c_t *)I2C_BASE3;
+    else
+#endif
+    if (bus == 1)
+        i2c = (i2c_t *)I2C_BASE2;
Quoted 8 lines

drivers/i2c/omap24xx_i2c.c static u16 wait_for_pin (void) · line 433

Quoted 5 lines
drivers/i2c/omap24xx_i2c.c +5 -0
+    else
+        i2c = (i2c_t *)I2C_BASE1;
+
+    return 0;
+}

I would remove the following three functions as they are not needed at the moment (i.e. without command line interface).

Hmm... really? If someone uses multibus support,
this functions are needed, or?

Quoted 29 lines
Quoted 29 lines
drivers/i2c/omap24xx_i2c.c +29 -0
+unsigned int i2c_get_bus_num(void)
+{
+    if (i2c == (i2c_t *)I2C_BASE1)
+        return 0;
+    else
+    if (i2c == (i2c_t *)I2C_BASE2)
+        return 1;
+#if defined(CONFIG_OMAP34XX)
+    else
+    if (i2c == (i2c_t *)I2C_BASE3)
+        return 2;
+#endif
+    else
+        return 0xFFFFFFFF;
+}
+
+/*
+ * To be Implemented
+ */
+int i2c_set_bus_speed(unsigned int speed)
+{
+    return 0;
+}
+
+unsigned int i2c_get_bus_speed(void)
+{
+    return 0;
+}
+

[...]

Quoted 14 lines

include/configs/omap3_beagle.h line 131

Quoted 10 lines
include/configs/omap3_beagle.h +1 -0
index 19a5ec9..d5a0d49 100644
@@ -128,6 +128,7 @@
 #define CONFIG_SYS_I2C_BUS        0
 #define CONFIG_SYS_I2C_BUS_SELECT    1
 #define CONFIG_DRIVER_OMAP34XX_I2C    1
+#define CONFIG_I2C_MULTI_BUS        1

Not needed.

Quoted 6 lines

While all above are only style questions, what's about the main functionality topic:

Do we have to call i2c_init() for bus 1 and 2 if switching to it? If yes, who does it? For bus 0 it's already in the code. I'd like that the user doesn't have to care about it. Therefore, I added

I solved this in the multibus_v2 branch, see:

http://git.denx.de/?p=u-boot/u-boot-i2c.git;a=shortlog;h=refs/heads/multibus_v2

i2c_init() gets just called if switching to an I2C bus. This is done in i2c_set_bus_num()

http://git.denx.de/?p=u-boot/u-boot-i2c.git;a=blob;f=drivers/i2c/i2c_core.c;h=e1bb67f1c136f5904c195d31d2916b17a30c9632;hb=6ef1bb6e9a0c03d5cbc8ef296b5022b5698207a6

pro:

a I2C adapter gets only initialized, if used, and there is the option to extend this functionality to:
- when switching to another i2c adapter call (a not yet existing) i2c adapter deinit function ...

Quoted 8 lines

static unsigned int bus_initialized[3] = {0, 0, 0}; static unsigned int current_bus = 0;

void i2c_init (int speed, int slaveadd) {
...
bus_initialized[current_bus] = 1;
}

Quoted 8 lines

int i2c_set_bus_num(unsigned int bus) {
...
current_bus = bus;

if(!bus_initialized[current_bus])

    i2c_init(CONFIG_SYS_I2C_SPEED, CONFIG_SYS_I2C_SLAVE);

...
}

Ah, you also tend to this direction ;-) Great!

bye
Heiko

Signature · 2 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
#6 lore raw

PATCH 1/1 Add I2C multibus support for OMAP2/3 boards

Hi Heiko,

Quoted 6 lines · Heiko Schocher
Quoted 2 lines · Dirk Behme

I would remove the following three functions as they are not needed at the moment (i.e. without command line interface).

Hmm... really? If someone uses multibus support,
this functions are needed, or?

Well, most probably yes. But I was not really looking for multibus support ;)

The functionality needed for Zippy I'm currently working on is just to be able to do

i2c_get_bus_num(1); /* Switch to bus 1 */
... do some stuff on bus 1 ... i2c_get_bus_num(0); /* Switch back to default bus 0 */

And for this having only i2c_get_bus_num() is sufficient. And the only tested ;)

Quoted 29 lines
Quoted 29 lines
Quoted 29 lines
drivers/i2c/omap24xx_i2c.c +29 -0
+unsigned int i2c_get_bus_num(void)
+{
+    if (i2c == (i2c_t *)I2C_BASE1)
+        return 0;
+    else
+    if (i2c == (i2c_t *)I2C_BASE2)
+        return 1;
+#if defined(CONFIG_OMAP34XX)
+    else
+    if (i2c == (i2c_t *)I2C_BASE3)
+        return 2;
+#endif
+    else
+        return 0xFFFFFFFF;
+}
+
+/*
+ * To be Implemented
+ */
+int i2c_set_bus_speed(unsigned int speed)
+{
+    return 0;
+}
+
+unsigned int i2c_get_bus_speed(void)
+{
+    return 0;
+}
+

...

Quoted 10 lines
Quoted 6 lines

While all above are only style questions, what's about the main functionality topic:

Do we have to call i2c_init() for bus 1 and 2 if switching to it? If yes, who does it? For bus 0 it's already in the code. I'd like that the user doesn't have to care about it. Therefore, I added

I solved this in the multibus_v2 branch, see:

http://git.denx.de/?p=u-boot/u-boot-i2c.git;a=shortlog;h=refs/heads/multibus_v2

Looks like we have to merge that with what is in this patch somehow. What are your plans with that branch? Is it intended to go into mainline, soon?

It seems to me that multibus_v2 and this patch are somehow orthogonal (but yes, they will need some merging).
Regarding omap24xx_i2c.c, multibus_v2 seems to deal mainly with the API stuff

http://git.denx.de/?p=u-boot/u-boot-i2c.git;a=blobdiff;f=drivers/i2c/omap24xx_i2c.c;h=993b995d34a298d07fbd11bce23f2d0cfdebb40f;hp=1a4c8c9ad2701b550fc1ad0ab2a9fde741902080;hb=899dbf9fbd089a9e8133026be3817e0116f44390;hpb=24d3d6307e1e5d5cddddee399a15e70ef9da8818

while the main part of this patch is to replace all hard coded registers by the ability to switch to an other bus.

I will send an updated patch with fixing at least the review comments, soon.

Best regards

Dirk

#7 lore raw

PATCH 1/1 Add I2C multibus support for OMAP2/3 boards

Hello Dirk,

Quoted 2 lines · Dirk Behme
Quoted 1 line · Heiko Schocher

Dirk Behme wrote:

[...]

Quoted 7 lines
Quoted 7 lines
Quoted 6 lines

While all above are only style questions, what's about the main functionality topic:

Do we have to call i2c_init() for bus 1 and 2 if switching to it? If yes, who does it? For bus 0 it's already in the code. I'd like that the user doesn't have to care about it. Therefore, I added

Quoted 8 lines
Quoted 4 lines

I solved this in the multibus_v2 branch, see:

http://git.denx.de/?p=u-boot/u-boot-i2c.git;a=shortlog;h=refs/heads/multibus_v2

Looks like we have to merge that with what is in this patch somehow. What are your plans with that branch? Is it intended to go into mainline, soon?

I have to do some more work, before this is ready to go in mainline...

Quoted 2 lines

It seems to me that multibus_v2 and this patch are somehow orthogonal (but yes, they will need some merging).

Yep, I hold the multibus_v2 branch in sync with mainline.

Quoted 4 lines

Regarding omap24xx_i2c.c, multibus_v2 seems to deal mainly with the API stuff

http://git.denx.de/?p=u-boot/u-boot-i2c.git;a=blobdiff;f=drivers/i2c/omap24xx_i2c.c;h=993b995d34a298d07fbd11bce23f2d0cfdebb40f;hp=1a4c8c9ad2701b550fc1ad0ab2a9fde741902080;hb=899dbf9fbd089a9e8133026be3817e0116f44390;hpb=24d3d6307e1e5d5cddddee399a15e70ef9da8818

Yep, I adapt this driver only for working with this new API.

Quoted 5 lines

while the main part of this patch is to replace all hard coded registers by the ability to switch to an other bus.

I will send an updated patch with fixing at least the review comments, soon.

OK, thanks.

bye
Heiko

Signature · 2 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
#8 lore raw

PATCH 1/1 Add I2C multibus support for OMAP2/3 boards

Hi Heiko,

Quoted 25 lines · Heiko Schocher

Hello Dirk,

Quoted 2 lines · Dirk Behme
Quoted 1 line · Heiko Schocher

Dirk Behme wrote:

[...]

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

While all above are only style questions, what's about the main functionality topic:

Do we have to call i2c_init() for bus 1 and 2 if switching to it? If yes, who does it? For bus 0 it's already in the code. I'd like that the user doesn't have to care about it. Therefore, I added

Quoted 7 lines
Quoted 4 lines

I solved this in the multibus_v2 branch, see:

http://git.denx.de/?p=u-boot/u-boot-i2c.git;a=shortlog;h=refs/heads/multibus_v2

Looks like we have to merge that with what is in this patch somehow. What are your plans with that branch? Is it intended to go into mainline, soon?

I have to do some more work, before this is ready to go in mainline...

Quoted 2 lines

It seems to me that multibus_v2 and this patch are somehow orthogonal (but yes, they will need some merging).

Quoted 14 lines

Yep, I hold the multibus_v2 branch in sync with mainline.

Quoted 4 lines

Regarding omap24xx_i2c.c, multibus_v2 seems to deal mainly with the API stuff

http://git.denx.de/?p=u-boot/u-boot-i2c.git;a=blobdiff;f=drivers/i2c/omap24xx_i2c.c;h=993b995d34a298d07fbd11bce23f2d0cfdebb40f;hp=1a4c8c9ad2701b550fc1ad0ab2a9fde741902080;hb=899dbf9fbd089a9e8133026be3817e0116f44390;hpb=24d3d6307e1e5d5cddddee399a15e70ef9da8818

Yep, I adapt this driver only for working with this new API.

Quoted 5 lines

while the main part of this patch is to replace all hard coded registers by the ability to switch to an other bus.

I will send an updated patch with fixing at least the review comments, soon.

Hmm, do you like to push

http://lists.denx.de/pipermail/u-boot/2009-November/063658.html

to I2C git main branch then?

Or are you fine that we push it to u-boot-ti?

Or?

Thanks

Dirk

#9 lore raw

PATCH 1/1 Add I2C multibus support for OMAP2/3 boards

Hello Dirk,

Quoted 9 lines · Dirk Behme

Hmm, do you like to push

http://lists.denx.de/pipermail/u-boot/2009-November/063658.html

to I2C git main branch then?

Or are you fine that we push it to u-boot-ti?

Or?

It;s in u-boot-i2c next, see

http://git.denx.de/?p=u-boot/u-boot-i2c.git;a=shortlog;h=refs/heads/next

because I think it is for the next merge window ...

bye
Heiko

Signature · 2 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany