Skip to contentSkip to conversation

platform/x86: int3472: discrete: Fix getting active_value

PATCH In review Hans de Goede 4 participants 11 messages Jun 12, 2023 – Jun 13, 2023

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/2 platform/x86: int3472: discrete: Fix getting active_value

acpi_object.integer.value is 64 bit, so to get bits 31-24 the value not only needs to be shifted but also masked.

---
 drivers/platform/x86/intel/int3472/discrete.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/platform/x86/intel/int3472/discrete.c +1 -1
index fc839a73e411..4ef60883154d 100644
@@ -179,7 +179,7 @@ static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares,
 	int3472_get_func_and_polarity(type, &func, &polarity);
 
 	/* If bits 31-24 of the _DSM entry are all 0 then the signal is inverted */
-	active_value = obj->integer.value >> 24;
+	active_value = (obj->integer.value >> 24) & 0xff;
 	if (!active_value)
 		polarity ^= GPIO_ACTIVE_LOW;
 
Signature · 1 line
2.40.1
#1 lore raw
Quoted 3 lines · Hans de Goede

acpi_object.integer.value is 64 bit, so to get bits 31-24 the value not only needs to be shifted but also masked.

Why? The type of active_value is u8. How does this make any difference?

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#2 lore raw

Hi,

Quoted 6 lines · Andy Shevchenko
Quoted 3 lines · Hans de Goede

acpi_object.integer.value is 64 bit, so to get bits 31-24 the value not only needs to be shifted but also masked.

Why? The type of active_value is u8. How does this make any difference?

Ah right I forgot about that, I'll drop this patch and drop the & 0xff for the second patch.

Regards,

Hans

#3 lore raw

PATCH 2/2 platform/x86: int3472: discrete: Log a warning if the pin-numbers don't match

In reply to Hans de Goede

The INT3472 discrete code assumes that the ACPI GPIO resources are in the same order as the pin-info _DSM entries.

The returned pin-info includes the pin-number in bits 15-8. Add a check that this matches with the ACPI GPIO resource pin-number in case the assumption is not true with some ACPI tables.

---
 drivers/platform/x86/intel/int3472/discrete.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
drivers/platform/x86/intel/int3472/discrete.c +9 -1
index 4ef60883154d..c1132bbbff41 100644
@@ -149,8 +149,8 @@ static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares,
 {
 	struct int3472_discrete_device *int3472 = data;
 	struct acpi_resource_gpio *agpio;
+	u8 active_value, pin, type;
 	union acpi_object *obj;
-	u8 active_value, type;
 	const char *err_msg;
 	const char *func;
 	u32 polarity;
@@ -174,10 +174,18 @@ static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares,
 		return 1;
 	}
 
+	/* Bits 7-0 contain the type/function of the pin */
 	type = obj->integer.value & 0xff;
 
 	int3472_get_func_and_polarity(type, &func, &polarity);
 
+	/* Bits 15-8 contain the pin-number on the GPIO chip */
+	pin = (obj->integer.value >> 8) & 0xff;
+	if (pin != agpio->pin_table[0])
+		dev_warn(int3472->dev, "%s %s pin number mismatch _DSM %d resource %d\n",
+			 func, agpio->resource_source.string_ptr, pin,
+			 agpio->pin_table[0]);
+
 	/* If bits 31-24 of the _DSM entry are all 0 then the signal is inverted */
 	active_value = (obj->integer.value >> 24) & 0xff;
 	if (!active_value)
Signature · 1 line
2.40.1
maintainer Reviewed-by on v1 2/2
#4 lore raw

Re: [PATCH 2/2] platform/x86: int3472: discrete: Log a warning if the pin-numbers don't match

Hi Hans

Quoted 8 lines · Hans de Goede

The INT3472 discrete code assumes that the ACPI GPIO resources are in the same order as the pin-info _DSM entries.

The returned pin-info includes the pin-number in bits 15-8. Add a check that this matches with the ACPI GPIO resource pin-number in case the assumption is not true with some ACPI tables.

That seems like a good idea to me:

Quoted 37 lines
---
  drivers/platform/x86/intel/int3472/discrete.c | 10 +++++++++-
  1 file changed, 9 insertions(+), 1 deletion(-)
drivers/platform/x86/intel/int3472/discrete.c +9 -1
index 4ef60883154d..c1132bbbff41 100644
@@ -149,8 +149,8 @@ static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares,
  {
  	struct int3472_discrete_device *int3472 = data;
  	struct acpi_resource_gpio *agpio;
+	u8 active_value, pin, type;
  	union acpi_object *obj;
-	u8 active_value, type;
  	const char *err_msg;
  	const char *func;
  	u32 polarity;
@@ -174,10 +174,18 @@ static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares,
  		return 1;
  	}
 
+	/* Bits 7-0 contain the type/function of the pin */
  	type = obj->integer.value & 0xff;
 
  	int3472_get_func_and_polarity(type, &func, &polarity);
 
+	/* Bits 15-8 contain the pin-number on the GPIO chip */
+	pin = (obj->integer.value >> 8) & 0xff;
+	if (pin != agpio->pin_table[0])
+		dev_warn(int3472->dev, "%s %s pin number mismatch _DSM %d resource %d\n",
+			 func, agpio->resource_source.string_ptr, pin,
+			 agpio->pin_table[0]);
+
  	/* If bits 31-24 of the _DSM entry are all 0 then the signal is inverted */
  	active_value = (obj->integer.value >> 24) & 0xff;
  	if (!active_value)
#5 lore raw

Re: [PATCH 2/2] platform/x86: int3472: discrete: Log a warning if the pin-numbers don't match

Hi,

Quoted 10 lines · Dan Scally

Hi Hans

Quoted 7 lines · Hans de Goede

The INT3472 discrete code assumes that the ACPI GPIO resources are in the same order as the pin-info _DSM entries.

The returned pin-info includes the pin-number in bits 15-8. Add a check that this matches with the ACPI GPIO resource pin-number in case the assumption is not true with some ACPI tables.

Quoted 7 lines
Quoted 1 line

That seems like a good idea to me:

Thanks. Did you also see my:
"[PATCH 0/4] platform/x86: int3472: discrete: Regulator rework / Lenovo Miix 510 support"

series ? It would be great if I can get your input on that.

Personally I'm not entirely happy with patch 3/4 there, but I don't really see any other way of solving the issue on the mix 510 that 3/4 fixes.

Regards,

Hans

Quoted 36 lines
Quoted 35 lines

---
  drivers/platform/x86/intel/int3472/discrete.c | 10 +++++++++-

  1 file changed, 9 insertions(+), 1 deletion(-)
drivers/platform/x86/intel/int3472/discrete.c +1 -1
index 4ef60883154d..c1132bbbff41 100644
@@ -149,8 +149,8 @@ static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares,
   {
       struct int3472_discrete_device *int3472 = data;
       struct acpi_resource_gpio *agpio;
+    u8 active_value, pin, type;
       union acpi_object *obj;
-    u8 active_value, type;
       const char *err_msg;
       const char *func;
       u32 polarity;
@@ -174,10 +174,18 @@ static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares,
           return 1;
       }

  +    /* Bits 7-0 contain the type/function of the pin */
      type = obj->integer.value & 0xff; int3472_get_func_and_polarity(type, &func, &polarity);
  +    /* Bits 15-8 contain the pin-number on the GPIO chip */
+    pin = (obj->integer.value >> 8) & 0xff;
+    if (pin != agpio->pin_table[0])
+        dev_warn(int3472->dev, "%s %s pin number mismatch _DSM %d resource %d\n", +             func, agpio->resource_source.string_ptr, pin,
+             agpio->pin_table[0]);
+
      /* If bits 31-24 of the _DSM entry are all 0 then the signal is inverted */ active_value = (obj->integer.value >> 24) & 0xff;
      if (!active_value)

maintainer
#6 lore raw

Re: [PATCH 2/2] platform/x86: int3472: discrete: Log a warning if the pin-numbers don't match

Hi Hans

Quoted 19 lines · Hans de Goede

Hi,

Quoted 10 lines · Dan Scally

Hi Hans

Quoted 7 lines · Hans de Goede

The INT3472 discrete code assumes that the ACPI GPIO resources are in the same order as the pin-info _DSM entries.

The returned pin-info includes the pin-number in bits 15-8. Add a check that this matches with the ACPI GPIO resource pin-number in case the assumption is not true with some ACPI tables.

Quoted 6 lines
Quoted 1 line

That seems like a good idea to me:

Quoted 8 lines

Thanks. Did you also see my:
"[PATCH 0/4] platform/x86: int3472: discrete: Regulator rework / Lenovo Miix 510 support"

series ? It would be great if I can get your input on that.

Personally I'm not entirely happy with patch 3/4 there, but I don't really see any other way of solving the issue on the mix 510 that 3/4 fixes.

I did; I was hoping to go through the ov2680 set and that set today but I've been a bit busy; I'll try to look at and test both asap.

Quoted 40 lines

Regards,

Hans

Quoted 34 lines
Quoted 34 lines
---
   drivers/platform/x86/intel/int3472/discrete.c | 10 +++++++++-
   1 file changed, 9 insertions(+), 1 deletion(-)
drivers/platform/x86/intel/int3472/discrete.c +7 -1
index 4ef60883154d..c1132bbbff41 100644
@@ -149,8 +149,8 @@ static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares,
   {
       struct int3472_discrete_device *int3472 = data;
       struct acpi_resource_gpio *agpio;
+    u8 active_value, pin, type;
       union acpi_object *obj;
-    u8 active_value, type;
       const char *err_msg;
       const char *func;
       u32 polarity;
@@ -174,10 +174,18 @@ static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares,
           return 1;
       }
   +    /* Bits 7-0 contain the type/function of the pin */
       type = obj->integer.value & 0xff;
         int3472_get_func_and_polarity(type, &func, &polarity);
   +    /* Bits 15-8 contain the pin-number on the GPIO chip */
+    pin = (obj->integer.value >> 8) & 0xff;
+    if (pin != agpio->pin_table[0])
+        dev_warn(int3472->dev, "%s %s pin number mismatch _DSM %d resource %d\n",
+             func, agpio->resource_source.string_ptr, pin,
+             agpio->pin_table[0]);
+
       /* If bits 31-24 of the _DSM entry are all 0 then the signal is inverted */
       active_value = (obj->integer.value >> 24) & 0xff;
       if (!active_value)
#7 lore raw

Re: [PATCH 2/2] platform/x86: int3472: discrete: Log a warning if the pin-numbers don't match

Hi,

Quoted 25 lines · Dan Scally

Hi Hans

Quoted 22 lines · Hans de Goede

Hi,

Quoted 10 lines · Dan Scally

Hi Hans

Quoted 7 lines · Hans de Goede

The INT3472 discrete code assumes that the ACPI GPIO resources are in the same order as the pin-info _DSM entries.

The returned pin-info includes the pin-number in bits 15-8. Add a check that this matches with the ACPI GPIO resource pin-number in case the assumption is not true with some ACPI tables.

Quoted 6 lines
Quoted 1 line

That seems like a good idea to me:

Thanks. Did you also see my:
"[PATCH 0/4] platform/x86: int3472: discrete: Regulator rework / Lenovo Miix 510 support"

Quoted 8 lines
Quoted 5 lines

series ? It would be great if I can get your input on that.

Personally I'm not entirely happy with patch 3/4 there, but I don't really see any other way of solving the issue on the mix 510 that 3/4 fixes.

I did; I was hoping to go through the ov2680 set and that set today but I've been a bit busy; I'll try to look at and test both asap.

Ok, thank you.

Note the ov2680 set does not work with the IPU3 yet. I've tried with the ipu3-capture.sh script after first testing the script with the ov5693 and my current ov2680 set is missing the link-rate / pixel-rate v4l2 ctrls.

I plan to continue working on this, adding those ctrls tomorrow.

Regards,

Hans

Quoted 35 lines
Quoted 34 lines
Quoted 34 lines
Quoted 34 lines

---
   drivers/platform/x86/intel/int3472/discrete.c | 10 +++++++++-

   1 file changed, 9 insertions(+), 1 deletion(-)
drivers/platform/x86/intel/int3472/discrete.c +1 -1
index 4ef60883154d..c1132bbbff41 100644
@@ -149,8 +149,8 @@ static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares,
    {
        struct int3472_discrete_device *int3472 = data;
        struct acpi_resource_gpio *agpio;
+    u8 active_value, pin, type;
        union acpi_object *obj;
-    u8 active_value, type;
        const char *err_msg;
        const char *func;
        u32 polarity;
@@ -174,10 +174,18 @@ static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares,
            return 1;
        }

   +    /* Bits 7-0 contain the type/function of the pin */
       type = obj->integer.value & 0xff; int3472_get_func_and_polarity(type, &func, &polarity);
   +    /* Bits 15-8 contain the pin-number on the GPIO chip */
+    pin = (obj->integer.value >> 8) & 0xff;
+    if (pin != agpio->pin_table[0])
+        dev_warn(int3472->dev, "%s %s pin number mismatch _DSM %d resource %d\n", +             func, agpio->resource_source.string_ptr, pin,
+             agpio->pin_table[0]);
+
       /* If bits 31-24 of the _DSM entry are all 0 then the signal is inverted */ active_value = (obj->integer.value >> 24) & 0xff;
       if (!active_value)

#8 lore raw

Re: [PATCH 2/2] platform/x86: int3472: discrete: Log a warning if the pin-numbers don't match

In reply to Hans de Goede
Quoted 7 lines · Hans de Goede

The INT3472 discrete code assumes that the ACPI GPIO resources are in the same order as the pin-info _DSM entries.

The returned pin-info includes the pin-number in bits 15-8. Add a check that this matches with the ACPI GPIO resource pin-number in case the assumption is not true with some ACPI tables.

...

drivers/platform/x86/intel/int3472/discrete.c static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares, · line 183

Quoted 4 lines
drivers/platform/x86/intel/int3472/discrete.c +3 -0
+       u8 active_value, pin, type;
 
+       /* Bits 15-8 contain the pin-number on the GPIO chip */
+       pin = (obj->integer.value >> 8) & 0xff;

All the same here, as long as the pin is u8, the ' & 0xff' is no-op as it's already implied.

drivers/platform/x86/intel/int3472/discrete.c static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares, · line 185

Quoted 2 lines
drivers/platform/x86/intel/int3472/discrete.c +2 -0
+       if (pin != agpio->pin_table[0])
+               dev_warn(int3472->dev, "%s %s pin number mismatch _DSM %d resource %d\n",

%u
%u

Quoted 2 lines
drivers/platform/x86/intel/int3472/discrete.c +2 -0
+                        func, agpio->resource_source.string_ptr, pin,
+                        agpio->pin_table[0]);
Signature · 2 lines
With Best Regards,
Andy Shevchenko
maintainer
#9 lore raw

Re: [PATCH 2/2] platform/x86: int3472: discrete: Log a warning if the pin-numbers don't match

In reply to Hans de Goede
Quoted 36 lines · Hans de Goede

The INT3472 discrete code assumes that the ACPI GPIO resources are in the same order as the pin-info _DSM entries.

The returned pin-info includes the pin-number in bits 15-8. Add a check that this matches with the ACPI GPIO resource pin-number in case the assumption is not true with some ACPI tables.

---
 drivers/platform/x86/intel/int3472/discrete.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
drivers/platform/x86/intel/int3472/discrete.c +2 -1
index 4ef60883154d..c1132bbbff41 100644
@@ -149,8 +149,8 @@ static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares,
 {
 	struct int3472_discrete_device *int3472 = data;
 	struct acpi_resource_gpio *agpio;
+	u8 active_value, pin, type;
 	union acpi_object *obj;
-	u8 active_value, type;
 	const char *err_msg;
 	const char *func;
 	u32 polarity;
@@ -174,10 +174,18 @@ static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares,
 		return 1;
 	}
 
+	/* Bits 7-0 contain the type/function of the pin */
 	type = obj->integer.value & 0xff;
 
 	int3472_get_func_and_polarity(type, &func, &polarity);
 

drivers/platform/x86/intel/int3472/discrete.c static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares, · line 191

Quoted 10 lines
drivers/platform/x86/intel/int3472/discrete.c +7 -0
+	/* Bits 15-8 contain the pin-number on the GPIO chip */
+	pin = (obj->integer.value >> 8) & 0xff;
+	if (pin != agpio->pin_table[0])
+		dev_warn(int3472->dev, "%s %s pin number mismatch _DSM %d resource %d\n",
+			 func, agpio->resource_source.string_ptr, pin,
+			 agpio->pin_table[0]);
+
 	/* If bits 31-24 of the _DSM entry are all 0 then the signal is inverted */
 	active_value = (obj->integer.value >> 24) & 0xff;
 	if (!active_value)

These changes made me wonder why there aren't defines for the fields? And then FIELD_GET() used to read the field. Most of those comments would be documented by the define name itself.

Signature · 1 line
 i.
#10 lore raw

Re: [PATCH 2/2] platform/x86: int3472: discrete: Log a warning if the pin-numbers don't match

Hi,

Quoted 37 lines · Ilpo Järvinen
Quoted 35 lines · Hans de Goede

The INT3472 discrete code assumes that the ACPI GPIO resources are in the same order as the pin-info _DSM entries.

The returned pin-info includes the pin-number in bits 15-8. Add a check that this matches with the ACPI GPIO resource pin-number in case the assumption is not true with some ACPI tables.

---
 drivers/platform/x86/intel/int3472/discrete.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
drivers/platform/x86/intel/int3472/discrete.c +2 -1
index 4ef60883154d..c1132bbbff41 100644
@@ -149,8 +149,8 @@ static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares,
 {
 	struct int3472_discrete_device *int3472 = data;
 	struct acpi_resource_gpio *agpio;
+	u8 active_value, pin, type;
 	union acpi_object *obj;
-	u8 active_value, type;
 	const char *err_msg;
 	const char *func;
 	u32 polarity;
@@ -174,10 +174,18 @@ static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares,
 		return 1;
 	}
 
+	/* Bits 7-0 contain the type/function of the pin */
 	type = obj->integer.value & 0xff;
 
 	int3472_get_func_and_polarity(type, &func, &polarity);
 
Quoted 15 lines

drivers/platform/x86/intel/int3472/discrete.c static int skl_int3472_handle_gpio_resources(struct acpi_resource *ares, · line 191

Quoted 11 lines
drivers/platform/x86/intel/int3472/discrete.c +7 -0
+	/* Bits 15-8 contain the pin-number on the GPIO chip */
+	pin = (obj->integer.value >> 8) & 0xff;
+	if (pin != agpio->pin_table[0])
+		dev_warn(int3472->dev, "%s %s pin number mismatch _DSM %d resource %d\n",
+			 func, agpio->resource_source.string_ptr, pin,
+			 agpio->pin_table[0]);
+
 	/* If bits 31-24 of the _DSM entry are all 0 then the signal is inverted */
 	active_value = (obj->integer.value >> 24) & 0xff;
 	if (!active_value)

These changes made me wonder why there aren't defines for the fields? And then FIELD_GET() used to read the field. Most of those comments would be documented by the define name itself.

That is a good idea for v2 I'll add a new 1/2 adding defines + switching the existing cases to FIELD_GET() and I'll also switch to FIELD_GET() here for v2.

Regards,

Hans