On 12/16/23 20:59, Abdel Alkuor wrote:
On Sat, Dec 16, 2023 at 05:40:35PM -0800, Guenter Roeck wrote:
On 12/16/23 14:07, Abdel Alkuor wrote:
On Sat, Dec 16, 2023 at 10:46:53AM -0800, Guenter Roeck wrote:
On 12/16/23 08:39, Abdel Alkuor wrote:
Should I use tmp112 params for as6200?
Sure, or just add a separate entry for as6200.
I think some modifications need to be done regarding setting the default
configuration for chips with config reg of 16 bits.
Currently, tmp112 set_mask and clr_mask look like this
[tmp112] = {
.set_mask = 3 << 5, /* 8 samples / second */
.clr_mask = 1 << 7, /* no one-shot mode*/
...
}
and in probe function, we are using i2c_smbus_read_byte_data which
basically reads byte 1 of tmp112 config reg and in lm75_write_config
it writes byte 1 of tmp112 config reg. Now based on tmp112 set_mask,
we want to set the sample rate but we actually setting R0 and R1 instead.
According to tmp112 datasheet on pg. 16, byte 1 is written first then
byte 2, where byte 2 has the conversion rate at bit 6 and 7 (CR0/CR1).
tmp112 datasheet: https://www.ti.com/lit/ds/symlink/tmp112.pdf?ts=1702713491401&ref_url=https%253A%252F%252Fwww.google.com%252F
Now, to accommodate 16 bit config register read/write, something along these lines can
be done:
- In struct lm75_params,
- change set_mask and clr_mask from u8 to u16
- Add config reg two bytes size flag
- Use the proper function to read the config reg based on config reg size i.e
For one byte config reg, use i2c_smbus_read_byte_data, and for 2 bytes
config reg, use regmap_read.
static int lm75_probe(struct i2c_client *client)
{
...
if (data->params->config_reg_16bits)
status = regmap_read(client, LM75_REG_CONF, ®val);
if (status < 0) {
dev_dbg(dev, "Can't read config? %d\n", status);
return status;
}
data->orig_conf = regval;
data->current_conf = regval;
} else {
status = i2c_smbus_read_byte_data(client, LM75_REG_CONF);
if (status < 0) {
dev_dbg(dev, "Can't read config? %d\n", status);
return status;
}
data->orig_conf = status;
data->current_conf = status;
}
...
}
static int lm75_write_config(struct lm75_data *data, u16 set_mask,
u16 clr_mask)
{
if (data->params->config_reg_16bits)
clr_mask |= LM75_SHUTDOWN << 8;
else
clr_mask |= LM75_SHUTDOWN;
...
if (data->params->config_reg_16bits)
err = regmap_write(data->regmap, LM75_REG_CONF, value);
else
err = i2c_smbus_write_byte_data(data->client,
LM75_REG_CONF,
value);
...
}
Based on that, the new tmp112 set_mask and clr_mask would look like this instead,
[tmp112] = {
.set_mask = 3 << 6, /* 8 samples / second */
.clr_mask = 1 << 15, /* no one-shot mode*/
.config_reg_16bits = 1,
...
}
Yes, you are correct, we'll need something like that. lm75_update_interval()
tries to solve the problem for tmp112, but that doesn't work with
set_mask/clear_mask. We should have a separate function lm75_read_config(),
though, to hide the complexity.
Thanks,
Guenter
