Skip to contentSkip to conversation

ACPI: Do not call _STA on battery devices with unmet dependencies

PATCH Superseded Hans de Goede 4 participants 14 messages Jan 22, 2018 – Feb 26, 2018

Subsystems (3)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

v2 cover ACPI: Do not call _STA on battery devices with unmet dependencies

Hi All,

Here is v2 of this set, new in v2 is the addressing of Bjorn's comments to the "PCI: acpiphp_ibm: prepare for acpi_get_object_info no longer returning status" patch and a new patch (the first patch in v2) which exports acpi_bus_get_status_handle() for use in modules as acpiphp_ibm will now depend on it.

Here is the old coverletter from v1 (adjusted for the new patch):

The ACPI code already contains quite a bit of code to not bind the ACPI-battery until all deps for an ACPI battery device have been met, but on some devices calling _STA before all deps are met is a problem too because the _STA method uses an i2c OpRegion there.

Here is the DSDT of the device I'm seeing this on:
https://fedorapeople.org/~jwrdegoede/toshiba-click-mini-dsdt.dsl

This series modifies the kernel to not call _STA until all deps are met, mirroring the binding behavior of the battery driver.

Without this series a total of 32 ACPI errors get printend to the console on boot, there are 4 errors per _STA call, 2 battery devices on this system and 4 _STA calls per battery device.

The first 2 commits are preparation commits for making the ACPICA changes in the 4th commit, these commits are necessary to not break things after the ACPICA changes.

The 3th commit modifies acpi_bus_get_status to not call _STA on battery devices until all deps are met. This fixes 2 of the 4 too early _STA calls triggering these errors.

The 4th commit makes the device instantiation code use acpi_bus_get_status instead of acpi_bus_get_status_handle so that the code to get the initial status also does not makes 1 too early _STA call.

The 5th commit changes the ACPICA acpi_get_object_info function to not call _STA. Only 1 user (which is fixed in the first commit) cares about acpi_device_info.current_status. And the ACPICA code has this comment:

* Note: This interface is intended to be used during the initial device
* discovery namespace traversal. Therefore, no complex methods can be
* executed, especially those that access operation regions. Therefore, do
* not add any additional methods that could cause problems in this area.
* Because of this reason support for the following methods has been removed:
* this was the fate of the _SUB method which was found to cause such
* problems and was removed (11/2015).

The described problems with the _SUB method clearly also apply to the _STA method, so removing it from acpi_get_object_info seems like it is the right thing to do here. This too fixes 1 too early _STA call, so that with all 5 patches in place we've fixed all 4 too early _STA calls.

Regards,

Hans

#1 lore raw

PATCH v2 1/5 ACPI: export acpi_bus_get_status_handle()

Some modular drivers need this, export it.

---
Changes in v2:
-This is a new patch in v2 of this patch-set

---
 drivers/acpi/bus.c | 1 +
 1 file changed, 1 insertion(+)
drivers/acpi/bus.c +1 -0
index 4d0979e02a28..db0c05997cc2 100644
@@ -108,6 +108,7 @@ acpi_status acpi_bus_get_status_handle(acpi_handle handle,
 	}
 	return status;
 }
+EXPORT_SYMBOL(acpi_bus_get_status_handle);
 
 int acpi_bus_get_status(struct acpi_device *device)
 {
Signature · 1 line
2.14.3
#2 lore raw

Re: [Devel] [PATCH v2 1/5] ACPI: export acpi_bus_get_status_handle()

Quoted 15 lines · Hans de Goede

Some modular drivers need this, export it.

Signed-off-by Hans de Goede <hdegoede(a)redhat.com>

---
Changes in v2:

(quoted diff) +0 -2
-This is a new patch in v2 of this patch-set
---
 drivers/acpi/bus.c | 1 +
 1 file changed, 1 insertion(+)
drivers/acpi/bus.c +0 -0
index 4d0979e02a28..db0c05997cc2 100644

drivers/acpi/bus.c acpi_status acpi_bus_get_status_handle(acpi_handle handle, · line 111

Quoted 4 lines
drivers/acpi/bus.c +1 -0
@@ -108,6 +108,7 @@ acpi_status acpi_bus_get_status_handle(acpi_handle handle,
        }
        return status;
 }
+EXPORT_SYMBOL(acpi_bus_get_status_handle);

EXPORT_SYMBOL_GPL(), please.

Quoted 4 lines
drivers/acpi/bus.c +0 -1
 int acpi_bus_get_status(struct acpi_device *device)
 {
--
#3 lore raw

Re: [PATCH v2 1/5] ACPI: export acpi_bus_get_status_handle()

Hi,

Quoted 11 lines · Rafael J. Wysocki
Quoted 10 lines · Hans de Goede

Some modular drivers need this, export it.

---
Changes in v2:
-This is a new patch in v2 of this patch-set

---
  drivers/acpi/bus.c | 1 +
  1 file changed, 1 insertion(+)
Quoted 11 lines

drivers/acpi/bus.c acpi_status acpi_bus_get_status_handle(acpi_handle handle, · line 111

Quoted 9 lines
drivers/acpi/bus.c +1 -0
index 4d0979e02a28..db0c05997cc2 100644
@@ -108,6 +108,7 @@ acpi_status acpi_bus_get_status_handle(acpi_handle handle,
         }
         return status;
  }
+EXPORT_SYMBOL(acpi_bus_get_status_handle);

EXPORT_SYMBOL_GPL(), please.

Ok, v3 with this fixed coming up.

Regards,

Hans

#4 lore raw

PATCH v2 2/5 PCI: acpiphp_ibm: prepare for acpi_get_object_info() no longer returning status

In reply to Hans de Goede

acpi_get_object_info() is intended for early probe usage and as such should not call any methods which may rely on OpRegions, but it used to also call _STA to get the status, which on some systems does rely on OpRegions, this behavior and the acpi_device_info.current_status member are being removed.

This commit prepares the acpiphp_ibm code for this by having it get the status itself using acpi_bus_get_status_handle(). Note no error handling is necessary on any errors acpi_bus_get_status_handle() leaves the value of the passed in current_status at its 0 initialization value.

---
Changes in v2:
-Use () after function names in commit message
-Add Bjorn's Acked-by

---
 drivers/pci/hotplug/acpiphp_ibm.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/pci/hotplug/acpiphp_ibm.c +4 -1
index 984c7e8cec5a..8472c4a27f70 100644
@@ -399,6 +399,7 @@ static acpi_status __init ibm_find_acpi_device(acpi_handle handle,
 		u32 lvl, void *context, void **rv)
 {
 	acpi_handle *phandle = (acpi_handle *)context;
+	unsigned long long current_status = 0;
 	acpi_status status;
 	struct acpi_device_info *info;
 	int retval = 0;
@@ -410,7 +411,9 @@ static acpi_status __init ibm_find_acpi_device(acpi_handle handle,
 		return retval;
 	}
 
-	if (info->current_status && (info->valid & ACPI_VALID_HID) &&
+	acpi_bus_get_status_handle(handle, &current_status);
+
+	if (current_status && (info->valid & ACPI_VALID_HID) &&
 			(!strcmp(info->hardware_id.string, IBM_HARDWARE_ID1) ||
 			 !strcmp(info->hardware_id.string, IBM_HARDWARE_ID2))) {
 		pr_debug("found hardware: %s, handle: %p\n",
Signature · 1 line
2.14.3
#5 lore raw

PATCH v2 3/5 ACPI / bus: Do not call _STA on battery devices with unmet dependencies

In reply to Hans de Goede

The battery code uses acpi_device->dep_unmet to check for unmet deps and if there are unmet deps it does not bind to the device to avoid errors about missing OpRegions when calling ACPI methods on the device.

The missing OpRegions when there are unmet deps problem also applies to the _STA method of some battery devices and calling it too early results in errors like these:

[    0.123579] ACPI Error: No handler for Region [ECRM] (00000000ba9edc4c)
               [GenericSerialBus] (20170831/evregion-166)
[    0.123601] ACPI Error: Region GenericSerialBus (ID=9) has no handler
               (20170831/exfldio-299)
[    0.123618] ACPI Error: Method parse/execution failed
               \_SB.I2C1.BAT1._STA, AE_NOT_EXIST (20170831/psparse-550)

This commit fixes these errors happening when acpi_get_bus_status gets called by checking dep_unmet for battery devices and reporting a status of 0 until all dependencies are met.

---
 drivers/acpi/bus.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/acpi/bus.c +6 -0
index db0c05997cc2..d0d2c2543395 100644
@@ -120,6 +120,12 @@ int acpi_bus_get_status(struct acpi_device *device)
 		return 0;
 	}
 
+	/* Battery devices must have their deps met before calling _STA */
+	if (acpi_device_is_battery(device) && device->dep_unmet) {
+		acpi_set_device_status(device, 0);
+		return 0;
+	}
+
 	status = acpi_bus_get_status_handle(device->handle, &sta);
 	if (ACPI_FAILURE(status))
 		return -ENODEV;
Signature · 1 line
2.14.3
#6 lore raw

PATCH v2 4/5 ACPI / scan: Use acpi_bus_get_status for initial status of ACPI_TYPE_DEVICE devs

In reply to Hans de Goede

The acpi_get_bus_status wrapper for acpi_bus_get_status_handle has some code to handle certain device quirks, in some cases we also need this quirk handling for the initial _STA call.

Specifically on some devices calling _STA before all _DEP dependencies are met results in errors like these:

[    0.123579] ACPI Error: No handler for Region [ECRM] (00000000ba9edc4c)
               [GenericSerialBus] (20170831/evregion-166)
[    0.123601] ACPI Error: Region GenericSerialBus (ID=9) has no handler
               (20170831/exfldio-299)
[    0.123618] ACPI Error: Method parse/execution failed
               \_SB.I2C1.BAT1._STA, AE_NOT_EXIST (20170831/psparse-550)

acpi_get_bus_status already has code to avoid this, so by using it we also silence these errors from the initial _STA call.

Note that in order for the acpi_get_bus_status handling for this to work, we initialize dep_unmet to 1 until acpi_device_dep_initialize gets called, this means that battery devices will be instantiated with an initial status of 0. This is not a problem, acpi_bus_attach will get called soon after the instantiation anyways and it will update the status as first point of order.

---
Note if we were to:
1) move the acpi_bus_init_power call from acpi_bus_get_power_flags to acpi_bus_attach, which already calls acpi_bus_init_power for uninitialized devices; and
2) move the acpi_bus_get_wakeup_device_flags call from acpi_add_single_object to acpi_bus_attach. This one is tricky, because acpi_device_add checks wakeup.flags.valid which gets set by this. And the acpi_wakeup_gpe_init call actually is the troublesome one because it uses acpi_match_device_ids which relies on the status...

Then we can entirely rely on the acpi_bus_get_status call in acpi_bus_attach and remove the acpi_bus_get_status call from the device-instantiation path.

---
 drivers/acpi/scan.c | 20 +++++++++++++++++---
 1 file changed, 17 insertions(+), 3 deletions(-)
drivers/acpi/scan.c +17 -3
index e0bd49b4c3f6..ffb64424771b 100644
@@ -1572,6 +1572,8 @@ void acpi_init_device_object(struct acpi_device *device, acpi_handle handle,
 	device_initialize(&device->dev);
 	dev_set_uevent_suppress(&device->dev, true);
 	acpi_init_coherency(device);
+	/* Assume there are unmet deps until acpi_device_dep_initialize runs */
+	device->dep_unmet = 1;
 }
 
 void acpi_device_add_finalize(struct acpi_device *device)
@@ -1595,6 +1597,14 @@ static int acpi_add_single_object(struct acpi_device **child,
 	}
 
 	acpi_init_device_object(device, handle, type, sta);
+	/*
+	 * For ACPI_BUS_TYPE_DEVICE getting the status is delayed till here so
+	 * that we can call acpi_bus_get_status and use its quirk handling.
+	 * Note this must be done before the get power-/wakeup_dev-flags calls.
+	 */
+	if (type == ACPI_BUS_TYPE_DEVICE)
+		acpi_bus_get_status(device);
+
 	acpi_bus_get_power_flags(device);
 	acpi_bus_get_wakeup_device_flags(device);
 
@@ -1667,9 +1677,11 @@ static int acpi_bus_type_and_status(acpi_handle handle, int *type,
 			return -ENODEV;
 
 		*type = ACPI_BUS_TYPE_DEVICE;
-		status = acpi_bus_get_status_handle(handle, sta);
-		if (ACPI_FAILURE(status))
-			*sta = 0;
+		/*
+		 * acpi_add_single_object updates this once we've an acpi_device
+		 * so that acpi_bus_get_status' quirk handling can be used.
+		 */
+		*sta = 0;
 		break;
 	case ACPI_TYPE_PROCESSOR:
 		*type = ACPI_BUS_TYPE_PROCESSOR;
@@ -1767,6 +1779,8 @@ static void acpi_device_dep_initialize(struct acpi_device *adev)
 	acpi_status status;
 	int i;
 
+	adev->dep_unmet = 0;
+
 	if (!acpi_has_method(adev->handle, "_DEP"))
 		return;
 
Signature · 1 line
2.14.3
#7 lore raw

PATCH v2 5/5 ACPICA: Remove calling of _STA from acpi_get_object_info

In reply to Hans de Goede

As the comment above it indicates, acpi_get_object_info is intended for early probe usage and as such should not call any methods which may rely on OpRegions, before this commit it was also calling _STA, which on some systems does rely on OpRegions.

Calling _STA before things are ready leads to errors such as these:

[    0.123579] ACPI Error: No handler for Region [ECRM] (00000000ba9edc4c)
               [GenericSerialBus] (20170831/evregion-166)
[    0.123601] ACPI Error: Region GenericSerialBus (ID=9) has no handler
               (20170831/exfldio-299)
[    0.123618] ACPI Error: Method parse/execution failed
               \_SB.I2C1.BAT1._STA, AE_NOT_EXIST (20170831/psparse-550)

End 2015 support for the _SUB method was removed for exactly the same reason. Removing current_status from struct acpi_device_info only has a limit impact. Within ACPICA it is only used by 2 debug messages, both of which are modified to no longer print it with this commit.

Outside of ACPICA, for Linux there is only one user. For which a patch to remove the dependency on the current_status field is available.

---
 drivers/acpi/acpica/dbdisply.c |  5 ++---
 drivers/acpi/acpica/nsdumpdv.c |  5 ++---
 drivers/acpi/acpica/nsxfname.c | 19 ++++---------------
 include/acpi/actypes.h         |  2 --
 4 files changed, 8 insertions(+), 23 deletions(-)
drivers/acpi/acpica/dbdisply.c +2 -3
index 5a606eac0c22..7b5eb33fe962 100644
@@ -642,9 +642,8 @@ void acpi_db_display_object_type(char *object_arg)
 		return;
 	}
 
-	acpi_os_printf("ADR: %8.8X%8.8X, STA: %8.8X, Flags: %X\n",
-		       ACPI_FORMAT_UINT64(info->address),
-		       info->current_status, info->flags);
+	acpi_os_printf("ADR: %8.8X%8.8X, Flags: %X\n",
+		       ACPI_FORMAT_UINT64(info->address), info->flags);
 
 	acpi_os_printf("S1D-%2.2X S2D-%2.2X S3D-%2.2X S4D-%2.2X\n",
 		       info->highest_dstates[0], info->highest_dstates[1],
drivers/acpi/acpica/nsdumpdv.c +2 -3
index 5026594763ea..573a5f36f01a 100644
@@ -88,10 +88,9 @@ acpi_ns_dump_one_device(acpi_handle obj_handle,
 		}
 
 		ACPI_DEBUG_PRINT_RAW((ACPI_DB_TABLES,
-				      "    HID: %s, ADR: %8.8X%8.8X, Status: %X\n",
+				      "    HID: %s, ADR: %8.8X%8.8X\n",
 				      info->hardware_id.value,
-				      ACPI_FORMAT_UINT64(info->address),
-				      info->current_status));
+				      ACPI_FORMAT_UINT64(info->address));
 		ACPI_FREE(info);
 	}
 
drivers/acpi/acpica/nsxfname.c +4 -15
index 106966235805..0a9c600c2599 100644
@@ -241,7 +241,7 @@ static char *acpi_ns_copy_device_id(struct acpi_pnp_device_id *dest,
  *              namespace node and possibly by running several standard
  *              control methods (Such as in the case of a device.)
  *
- * For Device and Processor objects, run the Device _HID, _UID, _CID, _STA,
+ * For Device and Processor objects, run the Device _HID, _UID, _CID,
  * _CLS, _ADR, _sx_w, and _sx_d methods.
  *
  * Note: Allocates the return buffer, must be freed by the caller.
@@ -250,8 +250,9 @@ static char *acpi_ns_copy_device_id(struct acpi_pnp_device_id *dest,
  * discovery namespace traversal. Therefore, no complex methods can be
  * executed, especially those that access operation regions. Therefore, do
  * not add any additional methods that could cause problems in this area.
- * this was the fate of the _SUB method which was found to cause such
- * problems and was removed (11/2015).
+ * Because of this reason support for the following methods has been removed:
+ * 1) _SUB method was removed (11/2015)
+ * 2) _STA method was removed (01/2018)
  *
  ******************************************************************************/
 
@@ -374,20 +375,8 @@ acpi_get_object_info(acpi_handle handle,
 		 * Notes: none of these methods are required, so they may or may
 		 * not be present for this device. The Info->Valid bitfield is used
 		 * to indicate which methods were found and run successfully.
-		 *
-		 * For _STA, if the method does not exist, then (as per the ACPI
-		 * specification), the returned current_status flags will indicate
-		 * that the device is present/functional/enabled. Otherwise, the
-		 * current_status flags reflect the value returned from _STA.
 		 */
 
-		/* Execute the Device._STA method */
-
-		status = acpi_ut_execute_STA(node, &info->current_status);
-		if (ACPI_SUCCESS(status)) {
-			valid |= ACPI_VALID_STA;
-		}
-
 		/* Execute the Device._ADR method */
 
 		status = acpi_ut_evaluate_numeric_object(METHOD_NAME__ADR, node,
include/acpi/actypes.h +0 -2
index 4f077edb9b81..220ef8674763 100644
@@ -1191,7 +1191,6 @@ struct acpi_device_info {
 	u8 flags;		/* Miscellaneous info */
 	u8 highest_dstates[4];	/* _sx_d values: 0xFF indicates not valid */
 	u8 lowest_dstates[5];	/* _sx_w values: 0xFF indicates not valid */
-	u32 current_status;	/* _STA value */
 	u64 address;	/* _ADR value */
 	struct acpi_pnp_device_id hardware_id;	/* _HID value */
 	struct acpi_pnp_device_id unique_id;	/* _UID value */
@@ -1205,7 +1204,6 @@ struct acpi_device_info {
 
 /* Flags for Valid field above (acpi_get_object_info) */
 
-#define ACPI_VALID_STA                  0x0001
 #define ACPI_VALID_ADR                  0x0002
 #define ACPI_VALID_HID                  0x0004
 #define ACPI_VALID_UID                  0x0008
Signature · 1 line
2.14.3
#8 lore raw
In reply to Hans de Goede
Quoted 49 lines · CET Hans de Goede

Hi All,

Here is v2 of this set, new in v2 is the addressing of Bjorn's comments to the "PCI: acpiphp_ibm: prepare for acpi_get_object_info no longer returning status" patch and a new patch (the first patch in v2) which exports acpi_bus_get_status_handle() for use in modules as acpiphp_ibm will now depend on it.

Here is the old coverletter from v1 (adjusted for the new patch):

The ACPI code already contains quite a bit of code to not bind the ACPI-battery until all deps for an ACPI battery device have been met, but on some devices calling _STA before all deps are met is a problem too because the _STA method uses an i2c OpRegion there.

Here is the DSDT of the device I'm seeing this on:
https://fedorapeople.org/~jwrdegoede/toshiba-click-mini-dsdt.dsl

This series modifies the kernel to not call _STA until all deps are met, mirroring the binding behavior of the battery driver.

Without this series a total of 32 ACPI errors get printend to the console on boot, there are 4 errors per _STA call, 2 battery devices on this system and 4 _STA calls per battery device.

The first 2 commits are preparation commits for making the ACPICA changes in the 4th commit, these commits are necessary to not break things after the ACPICA changes.

The 3th commit modifies acpi_bus_get_status to not call _STA on battery devices until all deps are met. This fixes 2 of the 4 too early _STA calls triggering these errors.

The 4th commit makes the device instantiation code use acpi_bus_get_status instead of acpi_bus_get_status_handle so that the code to get the initial status also does not makes 1 too early _STA call.

The 5th commit changes the ACPICA acpi_get_object_info function to not call _STA. Only 1 user (which is fixed in the first commit) cares about acpi_device_info.current_status. And the ACPICA code has this comment:

* Note: This interface is intended to be used during the initial device
* discovery namespace traversal. Therefore, no complex methods can be
* executed, especially those that access operation regions. Therefore, do
* not add any additional methods that could cause problems in this area.
* Because of this reason support for the following methods has been removed:
* this was the fate of the _SUB method which was found to cause such
* problems and was removed (11/2015).

Quoted 4 lines

The described problems with the _SUB method clearly also apply to the _STA method, so removing it from acpi_get_object_info seems like it is the right thing to do here. This too fixes 1 too early _STA call, so that with all 5 patches in place we've fixed all 4 too early _STA calls.

I've applied [1-4/5] from this series.

The last one needs to go in via ACPICA anyway or we'll need to carry a Linux-specific replacement for acpi_get_object_info().

#9 lore raw

Hi,

Quoted 50 lines · Rafael J. Wysocki
Quoted 49 lines · CET Hans de Goede

Hi All,

Here is v2 of this set, new in v2 is the addressing of Bjorn's comments to the "PCI: acpiphp_ibm: prepare for acpi_get_object_info no longer returning status" patch and a new patch (the first patch in v2) which exports acpi_bus_get_status_handle() for use in modules as acpiphp_ibm will now depend on it.

Here is the old coverletter from v1 (adjusted for the new patch):

The ACPI code already contains quite a bit of code to not bind the ACPI-battery until all deps for an ACPI battery device have been met, but on some devices calling _STA before all deps are met is a problem too because the _STA method uses an i2c OpRegion there.

Here is the DSDT of the device I'm seeing this on:
https://fedorapeople.org/~jwrdegoede/toshiba-click-mini-dsdt.dsl

This series modifies the kernel to not call _STA until all deps are met, mirroring the binding behavior of the battery driver.

Without this series a total of 32 ACPI errors get printend to the console on boot, there are 4 errors per _STA call, 2 battery devices on this system and 4 _STA calls per battery device.

The first 2 commits are preparation commits for making the ACPICA changes in the 4th commit, these commits are necessary to not break things after the ACPICA changes.

The 3th commit modifies acpi_bus_get_status to not call _STA on battery devices until all deps are met. This fixes 2 of the 4 too early _STA calls triggering these errors.

The 4th commit makes the device instantiation code use acpi_bus_get_status instead of acpi_bus_get_status_handle so that the code to get the initial status also does not makes 1 too early _STA call.

The 5th commit changes the ACPICA acpi_get_object_info function to not call _STA. Only 1 user (which is fixed in the first commit) cares about acpi_device_info.current_status. And the ACPICA code has this comment:

* Note: This interface is intended to be used during the initial device
* discovery namespace traversal. Therefore, no complex methods can be
* executed, especially those that access operation regions. Therefore, do
* not add any additional methods that could cause problems in this area.
* Because of this reason support for the following methods has been removed:
* this was the fate of the _SUB method which was found to cause such
* problems and was removed (11/2015).

Quoted 6 lines
Quoted 4 lines

The described problems with the _SUB method clearly also apply to the _STA method, so removing it from acpi_get_object_info seems like it is the right thing to do here. This too fixes 1 too early _STA call, so that with all 5 patches in place we've fixed all 4 too early _STA calls.

I've applied [1-4/5] from this series.

Thank you.

Quoted 2 lines

The last one needs to go in via ACPICA anyway or we'll need to carry a Linux-specific replacement for acpi_get_object_info().

Ack, I'm discussing how to proceed with this with the ACPICA people.

Regards,

Hans

#10 lore raw
Quoted 10 lines
Quoted 4 lines

The last one needs to go in via ACPICA anyway or we'll need to carry a Linux-specific replacement for acpi_get_object_info().

Ack, I'm discussing how to proceed with this with the ACPICA people.

Regards,

Hans

I normally don't like to change external interfaces (the biggest reason being the sheer number of different OS vendors that use ACPICA), but in this case, I feel it may be OK. This interface is probably not heavily used, and there is precedent for this when the invocation of _SUB was removed.

On the other hand, I think I saw it mentioned that Linux calls this interface only once, is this correct? If so, perhaps it would be easiest to simply remove this call and use direct control method invocations to obtain the desired data.

If we go ahead and make this change to ACPICA: I have not seen the patch, but please make sure that all vestiges of this are removed. This includes all invocations within ACPICA that examine the result of the _STA invocation, the CurrentStatus field of ACPI_DEVICE_INFO, the ACPI_VALID_STA flag, etc.

Bob

#11 lore raw

Hi,

Quoted 7 lines · Robert
Quoted 7 lines
Quoted 4 lines

The last one needs to go in via ACPICA anyway or we'll need to carry a Linux-specific replacement for acpi_get_object_info().

Ack, I'm discussing how to proceed with this with the ACPICA people.

Quoted 8 lines
Quoted 3 lines

Regards,

Hans

I normally don't like to change external interfaces (the biggest reason being the sheer number of different OS vendors that use ACPICA), but in this case, I feel it may be OK. This interface is probably not heavily used, and there is precedent for this when the invocation of _SUB was removed.

On the other hand, I think I saw it mentioned that Linux calls this interface only once, is this correct?

More or less, Linux calls it once when enumerating all the devices and a few ACPI drivers call it again to get some extra info they need.

Quoted 1 line

If so, perhaps it would be easiest to simply remove this call and use direct control method invocations to obtain the desired data.

If we go this route we would probably end up copying the function, since as said several drivers use it too.

Quoted 1 line

If we go ahead and make this change to ACPICA: I have not seen the patch, but please make sure that all vestiges of this are removed. This includes all invocations within ACPICA that examine the result of the _STA invocation, the CurrentStatus field of ACPI_DEVICE_INFO, the ACPI_VALID_STA flag, etc.

The patch already does that:

https://patchwork.kernel.org/patch/10186195/

Regards,

Hans

#12 lore raw
Quoted 47 lines

-----Original Message-----
From: Hans de Goede [mailto:hdegoede(a)redhat.com]
Sent: Wednesday, February 14, 2018 2:33 AM
To: Moore, Robert <robert.moore(a)intel.com>; Rafael J. Wysocki <rjw(a)rjwysocki.net>; linux-acpi(a)vger.kernel.org; Schmauss, Erik <erik.schmauss(a)intel.com>; Box, David E <david.e.box(a)intel.com>
Cc: Len Brown <lenb(a)kernel.org>; Bjorn Helgaas <bhelgaas(a)google.com>; linux-pci(a)vger.kernel.org; devel(a)acpica.org
Subject: Re: [PATCH v2 0/5] ACPI: Do not call _STA on battery devices with unmet dependencies

Hi,

Quoted 5 lines · Robert
Quoted 5 lines
Quoted 4 lines

The last one needs to go in via ACPICA anyway or we'll need to carry a Linux-specific replacement for acpi_get_object_info().

Quoted 8 lines
Quoted 5 lines

Ack, I'm discussing how to proceed with this with the ACPICA people.

Regards,

Hans

I normally don't like to change external interfaces (the biggest

reason being the sheer number of different OS vendors that use ACPICA), but in this case, I feel it may be OK. This interface is probably not heavily used, and there is precedent for this when the invocation of _SUB was removed.

Quoted 2 lines

On the other hand, I think I saw it mentioned that Linux calls this

interface only once, is this correct?

More or less, Linux calls it once when enumerating all the devices and a few ACPI drivers call it again to get some extra info they need.

Quoted 1 line

If so, perhaps it would be easiest to simply remove this call and use

direct control method invocations to obtain the desired data.

If we go this route we would probably end up copying the function, since as said several drivers use it too.

Quoted 1 line

If we go ahead and make this change to ACPICA: I have not seen the

Quoted 9 lines

patch, but please make sure that all vestiges of this are removed. This includes all invocations within ACPICA that examine the result of the _STA invocation, the CurrentStatus field of ACPI_DEVICE_INFO, the ACPI_VALID_STA flag, etc.

The patch already does that:

https://patchwork.kernel.org/patch/10186195/

Of course, we will need the ACPICA version of the patch, not the "linuxized" version. Bob

Quoted 4 lines

Regards,

Hans

#13 lore raw

Hi,

Quoted 52 lines · Robert
Quoted 52 lines

-----Original Message-----
From: Hans de Goede [mailto:[email protected]]
Sent: Wednesday, February 14, 2018 2:33 AM
To: Moore, Robert <[email protected]>; Rafael J. Wysocki <[email protected]>; [email protected]; Schmauss, Erik <[email protected]>; Box, David E <[email protected]>
Cc: Len Brown <[email protected]>; Bjorn Helgaas <[email protected]>; [email protected]; [email protected]
Subject: Re: [PATCH v2 0/5] ACPI: Do not call _STA on battery devices with unmet dependencies

Hi,

Quoted 5 lines · Robert
Quoted 5 lines
Quoted 4 lines

The last one needs to go in via ACPICA anyway or we'll need to carry a Linux-specific replacement for acpi_get_object_info().

Quoted 8 lines
Quoted 5 lines

Ack, I'm discussing how to proceed with this with the ACPICA people.

Regards,

Hans

I normally don't like to change external interfaces (the biggest

reason being the sheer number of different OS vendors that use ACPICA), but in this case, I feel it may be OK. This interface is probably not heavily used, and there is precedent for this when the invocation of _SUB was removed.

Quoted 2 lines

On the other hand, I think I saw it mentioned that Linux calls this

interface only once, is this correct?

More or less, Linux calls it once when enumerating all the devices and a few ACPI drivers call it again to get some extra info they need.

Quoted 1 line

If so, perhaps it would be easiest to simply remove this call and use

direct control method invocations to obtain the desired data.

If we go this route we would probably end up copying the function, since as said several drivers use it too.

Quoted 1 line

If we go ahead and make this change to ACPICA: I have not seen the

patch, but please make sure that all vestiges of this are removed. This includes all invocations within ACPICA that examine the result of the _STA invocation, the CurrentStatus field of ACPI_DEVICE_INFO, the ACPI_VALID_STA flag, etc.

Quoted 8 lines
Quoted 4 lines

The patch already does that:

https://patchwork.kernel.org/patch/10186195/

Of course, we will need the ACPICA version of the patch, not the "linuxized" version. Bob

Ok, I've submitted a pull-request with an ACPICA version of the patch here:

https://github.com/acpica/acpica/pull/364

Regards,

Hans