Skip to contentSkip to conversation

ACPI / LPSS: Only call pwm_add_table() for Bay Trail PWM if PMIC HRV is 2

PATCH Applied Hans de Goede 3 participants 3 messages Apr 13, 2018 – May 13, 2018

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ACPI / LPSS: Only call pwm_add_table() for Bay Trail PWM if PMIC HRV is 2

The Point of View mobii wintab p800w Bay Trail tablet comes with a Crystal Cove PMIC, yet uses the LPSS PWM for backlight control, rather then the Crystal Cove's PWM, so we need to call pwm_add_table() to add a pwm_backlight mapping for the LPSS pwm despite there being an INT33FD ACPI device present.

On all Bay Trail devices the _HRV object of the INT33FD ACPI device will normally return 2, to indicate the Bay Trail variant of the CRC PMIC is present, except on this tablet where _HRV is 0xffff. I guess this is a hack to make the windows Crystal Cove PWM driver not bind.

Out of the 44 DSTDs with an INT33FD device in there which I have (from different model devices) only the pov mobii wintab p800w uses 0xffff for the HRV.

The byt_pwm_setup code calls acpi_dev_present to check for the presence of a INT33FD ACPI device which indicates that a CRC PMIC is present and if the INT33FD ACPI device is present then byt_pwm_setup will not add a pwm_backlight mapping for the LPSS pwm, so that the CRC PWM will get used instead.

acpi_dev_present has a hrv parameter, this commit make us pass 2 instead of -1, so that things still match on normal tablets, but on this special case with its _HRV of 0xffff, the check will now fail so that the pwm_backlight mapping for the LPSS pwm gets added fixing backlight brightness control on this device.

---
 drivers/acpi/acpi_lpss.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/acpi/acpi_lpss.c +5 -1
index c4ba9164e582..38a286975c31 100644
@@ -69,6 +69,10 @@ ACPI_MODULE_NAME("acpi_lpss");
 #define LPSS_SAVE_CTX			BIT(4)
 #define LPSS_NO_D3_DELAY		BIT(5)
 
+/* Crystal Cove PMIC shares same ACPI ID between different platforms */
+#define BYT_CRC_HRV			2
+#define CHT_CRC_HRV			3
+
 struct lpss_private_data;
 
 struct lpss_device_desc {
@@ -162,7 +166,7 @@ static void byt_pwm_setup(struct lpss_private_data *pdata)
 	if (!adev->pnp.unique_id || strcmp(adev->pnp.unique_id, "1"))
 		return;
 
-	if (!acpi_dev_present("INT33FD", NULL, -1))
+	if (!acpi_dev_present("INT33FD", NULL, BYT_CRC_HRV))
 		pwm_add_table(byt_pwm_lookup, ARRAY_SIZE(byt_pwm_lookup));
 }
 
Signature · 1 line
2.17.0
Reviewed-by on v1 1/1
#1 lore raw
Quoted 27 lines · Hans de Goede

The Point of View mobii wintab p800w Bay Trail tablet comes with a Crystal
Cove PMIC, yet uses the LPSS PWM for backlight control, rather then the
Crystal Cove's PWM, so we need to call pwm_add_table() to add a pwm_backlight mapping for the LPSS pwm despite there being an INT33FD ACPI device present.

On all Bay Trail devices the _HRV object of the INT33FD ACPI device will normally return 2, to indicate the Bay Trail variant of the CRC PMIC is present, except on this tablet where _HRV is 0xffff. I guess this
is a hack to make the windows Crystal Cove PWM driver not bind.

Out of the 44 DSTDs with an INT33FD device in there which I have (from different model devices) only the pov mobii wintab p800w uses 0xffff for
the HRV.

The byt_pwm_setup code calls acpi_dev_present to check for the presence
of a INT33FD ACPI device which indicates that a CRC PMIC is present and
if the INT33FD ACPI device is present then byt_pwm_setup will not add a pwm_backlight mapping for the LPSS pwm, so that the CRC PWM will get used instead.

Quoted 8 lines

acpi_dev_present has a hrv parameter, this commit make us pass 2 instead
of -1, so that things still match on normal tablets, but on this special
case with its _HRV of 0xffff, the check will now fail so that the pwm_backlight mapping for the LPSS pwm gets added fixing backlight brightness control on this device.

Quoted 32 lines
---
 drivers/acpi/acpi_lpss.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/acpi/acpi_lpss.c +4 -0
index c4ba9164e582..38a286975c31 100644
@@ -69,6 +69,10 @@ ACPI_MODULE_NAME("acpi_lpss");
 #define LPSS_SAVE_CTX			BIT(4)
 #define LPSS_NO_D3_DELAY		BIT(5)
 
+/* Crystal Cove PMIC shares same ACPI ID between different platforms
 */
+#define BYT_CRC_HRV			2
+#define CHT_CRC_HRV			3
+
 struct lpss_private_data;
 

struct lpss_device_desc {

drivers/acpi/acpi_lpss.c +1 -1
@@ -162,7 +166,7 @@ static void byt_pwm_setup(struct lpss_private_data
 *pdata)
 	if (!adev->pnp.unique_id || strcmp(adev->pnp.unique_id, "1"))
 		return;
 
-	if (!acpi_dev_present("INT33FD", NULL, -1))
+	if (!acpi_dev_present("INT33FD", NULL, BYT_CRC_HRV))
 		pwm_add_table(byt_pwm_lookup,
 ARRAY_SIZE(byt_pwm_lookup));

}

Signature · 2 lines
Andy Shevchenko <[email protected]>
Intel Finland Oy
#2 lore raw
Quoted 37 lines · CEST Andy Shevchenko
Quoted 27 lines · Hans de Goede

The Point of View mobii wintab p800w Bay Trail tablet comes with a Crystal
Cove PMIC, yet uses the LPSS PWM for backlight control, rather then the
Crystal Cove's PWM, so we need to call pwm_add_table() to add a pwm_backlight mapping for the LPSS pwm despite there being an INT33FD ACPI device present.

On all Bay Trail devices the _HRV object of the INT33FD ACPI device will normally return 2, to indicate the Bay Trail variant of the CRC PMIC is present, except on this tablet where _HRV is 0xffff. I guess this
is a hack to make the windows Crystal Cove PWM driver not bind.

Out of the 44 DSTDs with an INT33FD device in there which I have (from different model devices) only the pov mobii wintab p800w uses 0xffff for
the HRV.

The byt_pwm_setup code calls acpi_dev_present to check for the presence
of a INT33FD ACPI device which indicates that a CRC PMIC is present and
if the INT33FD ACPI device is present then byt_pwm_setup will not add a pwm_backlight mapping for the LPSS pwm, so that the CRC PWM will get used instead.

Quoted 8 lines

acpi_dev_present has a hrv parameter, this commit make us pass 2 instead
of -1, so that things still match on normal tablets, but on this special
case with its _HRV of 0xffff, the check will now fail so that the pwm_backlight mapping for the LPSS pwm gets added fixing backlight brightness control on this device.

Quoted 1 line

Applied, thanks!