Thanks, Krzysztof! Really appreciate your input, as I'm currently not a
regular contributor to the kernel.
On 2024-05-04 08:37, Krzysztof Kozlowski wrote:
On 04/05/2024 04:04, Felix Kaechele wrote:
Handle regulators used on this chip family, namely AVDD and VDD. These
definitions are taken from the GPLv2 licensed vendor driver.
Signed-off-by: Felix Kaechele <felix@xxxxxxxxxxx>
Link: https://github.com/HimaxSoftware/HX83112_Android_Driver
---
drivers/input/touchscreen/himax_hx83112b.c | 48 ++++++++++++++++++++--
1 file changed, 44 insertions(+), 4 deletions(-)
diff --git a/drivers/input/touchscreen/himax_hx83112b.c b/drivers/input/touchscreen/himax_hx83112b.c
index 4f6609dcdef3..0a797789e548 100644
--- a/drivers/input/touchscreen/himax_hx83112b.c
+++ b/drivers/input/touchscreen/himax_hx83112b.c
@@ -19,10 +19,12 @@
#include <linux/interrupt.h>
#include <linux/kernel.h>
#include <linux/regmap.h>
+#include <linux/regulator/consumer.h>
#define HIMAX_ID_83112B 0x83112b
#define HIMAX_MAX_POINTS 10
+#define HIMAX_MAX_SUPPLIES 2
#define HIMAX_REG_CFG_SET_ADDR 0x00
#define HIMAX_REG_CFG_INIT_READ 0x0c
@@ -50,6 +52,7 @@ struct himax_event {
static_assert(sizeof(struct himax_event) == 56);
struct himax_ts_data {
+ struct regulator_bulk_data supplies[HIMAX_MAX_SUPPLIES];
struct gpio_desc *gpiod_rst;
struct input_dev *input_dev;
struct i2c_client *client;
@@ -63,6 +66,11 @@ static const struct regmap_config himax_regmap_config = {
.val_format_endian = REGMAP_ENDIAN_LITTLE,
};
+static const char *const himax_supply_names[] = {
+ "avdd",
+ "vdd",
+};
+
That's confusing. Binding said only HX83100A family has regulators, but
you request for everyone.
You're right. Looking at the vendor driver for each of models I could
see that it defined AVDD and VDD in both drivers. So I thought it could
make sense to offer it for the entire chip family, with which I meant
all HX831xx in this case.
But it seems after some more testing (and with this touch IC family
generally being a part of the display controller) the regulators are
sufficiently handled by the panel driver / bootloader for the use case
of having the touchscreen on when the display is on.
It wouldn't allow for waking the screen by using the touchscreen, and
while I'd have to go back to the original OS on the device to verify
this again, I don't remember that working on the original OS either.
So I'm thinking I may drop the whole regulator part of the patchset to
keep it smaller.
static int himax_read_config(struct himax_ts_data *ts, u32 address, u32 *dst)
{
int error;
@@ -267,7 +275,7 @@ static irqreturn_t himax_irq_handler(int irq, void *dev_id)
static int himax_probe(struct i2c_client *client)
{
- int error;
+ int error, i;
struct device *dev = &client->dev;
struct himax_ts_data *ts;
@@ -290,11 +298,31 @@ static int himax_probe(struct i2c_client *client)
return error;
}
+ int num_supplies = ARRAY_SIZE(himax_supply_names);
+
+ for (i = 0; i < num_supplies; i++)
+ ts->supplies[i].supply = himax_supply_names[i];
+
+ error = devm_regulator_bulk_get(dev,
devm_regulator_bulk_get_enable and drop rest of the code here.
That's pretty neat. If I do decide to bring in regulator handling this
removes quite a bit of boilerplate.
+ num_supplies,
+ ts->supplies);
Wrap it properly at 80, not one argument in one line.
+ if (error) {
+ dev_err(dev, "Failed to get supplies: %d\n", error);
return dev_err_probe()
Same here. Thanks for the hint.
+ return error;
+ }
+
+ error = regulator_bulk_enable(num_supplies,
+ ts->supplies);
+ if (error) {
+ dev_err(dev, "Failed to enable supplies: %d\n", error);
+ goto error_out;
+ }
+
I'll be sending a v2 shortly.
Thanks again,
Felix
