Input: goodix - try to reset the controller if the i2c-test fails

On some ACPI/x86 devices (where we use one of the ACPI IRQ pin access
methods) the firmware is buggy, it does not properly reset the controller
at boot, and we cannot communicate with it.

Normally on ACPI/x86 devices we do not want to reset the controller at
probe time since in some cases this causes the controller to loose its
configuration and this is loaded into it by the system's firmware.
So on these systems we leave the reset_controller_at_probe flag unset,
even though we have a access to both the IRQ and reset pins and thus
could reset it.

In the case of the buggy firmware we have to reset the controller to
actually be able to talk to it.

This commit adds a special case for this, if the goodix_i2c_test() fails,
and we have not reset the controller yet; and we do have a way to reset
the controller then retry the i2c-test after resetting the controller.

This fixes the driver failing at probe on ACPI/x86 systems with this
firmware bug.
Reported-and-tested-by: Dmitry Mastykin <dmastykin@astralinux.ru>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20200311191013.10826-2-hdegoede@redhat.comSigned-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>

drivers/input/touchscreen/goodix.c

@@ -1159,6 +1159,7 @@ static int goodix_ts_probe(struct i2c_client *client,
 	if (error)
 		return error;
 
+reset:
 	if (ts->reset_controller_at_probe) {
 		/* reset the controller */
 		error = goodix_reset(ts);
@@ -1170,6 +1171,12 @@ static int goodix_ts_probe(struct i2c_client *client,
 
 	error = goodix_i2c_test(client);
 	if (error) {
+		if (!ts->reset_controller_at_probe &&
+		    ts->irq_pin_access_method != IRQ_PIN_ACCESS_NONE) {
+			/* Retry after a controller reset */
+			ts->reset_controller_at_probe = true;
+			goto reset;
+		}
 		dev_err(&client->dev, "I2C communication failure: %d\n", error);
 		return error;
 	}