Original question:
Hi, Keerthy
I applied the patch, but the symptoms seem to be the same.
I tried to leave a log, but it seems like the log is not left in that part.
Did I do something wrong?
Regars,
Kim
Keerthy,
The schematic for AVS is the same as the Devkit.
Is this the correct part?
The actual i2c address is also confirmed on uboot.
Regards,
Kim
Keerthy,
// SPDX-License-Identifier: GPL-2.0
/*
* (C) Copyright 2019 Texas Instruments Incorporated, <www.ti.com>
* Keerthy <j-keerthy@ti.com>
*/
#include <common.h>
#include <fdtdec.h>
#include <errno.h>
#include <dm.h>
#include <i2c.h>
#include <log.h>
#include <linux/printk.h>
#include <power/pmic.h>
#include <power/regulator.h>
#include <power/tps65941.h>
#include <dm/device.h>
#define TPS65941_WD_THR_CFG 0x409
#define TPS65941_WD_EN_MASK BIT(6)
static const struct pmic_child_info pmic_children_info[] = {
{ .prefix = "ldo", .driver = TPS65941_LDO_DRIVER },
{ .prefix = "buck", .driver = TPS65941_BUCK_DRIVER },
{ },
};
static int tps65941_write(struct udevice *dev, uint reg, const uint8_t *buff,
int len)
{
if (dm_i2c_write(dev, reg, buff, len)) {
pr_err("write error to device: %p register: %#x!\n", dev, reg);
return -EIO;
}
return 0;
}
static int tps65941_read(struct udevice *dev, uint reg, uint8_t *buff, int len)
{
if (dm_i2c_read(dev, reg, buff, len)) {
pr_err("read error from device: %p register: %#x!\n", dev, reg);
return -EIO;
}
return 0;
}
static int tps65941_bind(struct udevice *dev)
{
ofnode regulators_node;
int children;
if (dev->driver_data == TPS65941_WD)
return 0;
regulators_node = dev_read_subnode(dev, "regulators");
if (!ofnode_valid(regulators_node)) {
debug("%s: %s regulators subnode not found!\n", __func__,
dev->name);
return -ENXIO;
}
debug("%s: '%s' - found regulators subnode\n", __func__, dev->name);
children = pmic_bind_children(dev, regulators_node, pmic_children_info);
if (!children)
printf("%s: %s - no child found\n", __func__, dev->name);
/* Probe all the child devices */
return dm_scan_fdt_dev(dev);
}
static int stop_watchdog(struct udevice *dev)
{
int ret;
/* Disable WD */
ret = dm_i2c_reg_read(dev, TPS65941_WD_THR_CFG);
if (1) {
pr_err("failed to read i2c reg (%d)\n", ret);
//return ret;
}
ret &= ~TPS65941_WD_EN_MASK;
ret = dm_i2c_reg_write(dev, TPS65941_WD_THR_CFG, ret);
if (1) {
pr_err("%s: %s write WD_EN fail!\n", __func__, dev->name);
//return ret;
}
ret = dm_i2c_reg_read(dev, TPS65941_WD_THR_CFG);
if (1) {
pr_err("failed to read back i2c reg (%d)\n", ret);
//return ret;
}
return 0;
}
static int tps65941_probe(struct udevice *dev)
{
pr_err("%s: %s ------------------------!\n", __func__, dev->name);
if (dev->driver_data == TPS65941_WD)
return stop_watchdog(dev);
return 0;
}
static struct dm_pmic_ops tps65941_ops = {
.read = tps65941_read,
.write = tps65941_write,
};
static const struct udevice_id tps65941_ids[] = {
{ .compatible = "ti,tps659411", .data = TPS659411 },
{ .compatible = "ti,tps659412", .data = TPS659411 },
{ .compatible = "ti,tps659413", .data = TPS659413 },
{ .compatible = "ti,lp876441", .data = LP876441 },
{ .compatible = "ti,tps65224", .data = TPS65224 },
{ .compatible = "ti,tps6594-q1", .data = TPS659411 },
{ .compatible = "ti,tps6593-q1", .data = TPS659413 },
{ .compatible = "ti,lp8764-q1", .data = LP876441 },
{ .compatible = "ti,tps659413_wd", .data = TPS65941_WD },
{ }
};
U_BOOT_DRIVER(pmic_tps65941) = {
.name = "tps65941_pmic",
.id = UCLASS_PMIC,
.of_match = tps65941_ids,
.bind = tps65941_bind,
.probe = tps65941_probe,
.ops = &tps65941_ops,
};
I'm trying to proceed by adding tps65941 entry to mcu_i2c0, but it doesn't work properly.
It seems that 0x12 device driver is not running.
Can't there be two drivers running?
Regards,
Kim
The schematic for AVS is the same as the Devkit.
Is this the correct part?
The actual i2c address is also confirmed on uboot
So this is a standalone regulator and this does not come under TPS6594x. The slave ID is 0x48 on the EVMs not sure if its 0x12 on your custom board.
In the PMIC probe function can you add the register write directly?
- Keerthy
Hi, Keerthy
The function is called, but it still outputs only the value 0x2c at address 0x09. This is the same value as the value at address 0x09 in the nvm.bin file.
The value I want to change is address 0x409.
Regards,
Kim
Hi Keerthy
When I enable mcu_i2c0, I get an error like below.
I found a forum with a similar topic, is there a separate patch file for mcu_i2c0?
Regards,
Kim
Hi Keerthy
I checked the trm document because it was confirmed to be a power issue. I checked the table below.
ti-processor-sdk-linux-adas-j784s4-evm-10_00_00_08/board-support/ti-u-boot-2024.04+git/arch/arm/mach-k3/r5/j784s4/dev-data.c
Then, after modifying it as shown below, it seems that the 13-minute watchdog issue was resolved.
I think I need to test it more.
Will there be any problems with the board if I apply the above patch?
Regards,
Kim