[PATCH 0/3] ACPI: Support for platform initiated graceful shutdown

[PATCH 0/3] ACPI: Support for platform initiated graceful shutdown

[Prashanth Prakash]

This patch-set adds support for platform initited graceful shutdown as
described in the sections 5.6.6(Table-143) and 6.3.5.1 of ACPI 6.1 spec

Jonathan (Zhixiong) Zhang (1):
  ACPI: move ACPI_SYSTEM_HID to acpi_drivers.h

Prashanth Prakash (2):
  ACPI / sybus: add a driver for LNXSYBUS device
  ACPI: update the definition of common notification values

 drivers/acpi/Kconfig        |   6 +++
 drivers/acpi/Makefile       |   1 +
 drivers/acpi/scan.c         |   3 +-
 drivers/acpi/sybus.c        | 116 ++++++++++++++++++++++++++++++++++++++++++++
 include/acpi/acpi_drivers.h |   1 +
 include/acpi/actypes.h      |   2 +-
 6 files changed, 127 insertions(+), 2 deletions(-)
 create mode 100644 drivers/acpi/sybus.c

-- 
Qualcomm Technologies, Inc. on behalf
of the Qualcomm Innovation Center, Inc.  The Qualcomm Innovation Center, Inc. 
is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project.

[PATCH 1/3] ACPI: move ACPI_SYSTEM_HID to acpi_drivers.h

[Prashanth Prakash]

From: "Jonathan (Zhixiong) Zhang" <zjzhang@codeaurora.org>

ACPI_SYSTEM_HID is defined in drivers/acpi/bus.c out-of-band.

Move the definition to include/acpi/acpi_drivers.h, together
with other Linux specific HIDs, so that it can be used by other
code.

Signed-off-by: Jonathan (Zhixiong) Zhang <zjzhang@codeaurora.org>
---
 drivers/acpi/scan.c         | 3 ++-
 include/acpi/acpi_drivers.h | 1 +
 2 files changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c
index 407a376..358e3c8 100644
--- a/drivers/acpi/scan.c
+++ b/drivers/acpi/scan.c
@@ -13,6 +13,8 @@
 #include <linux/nls.h>
 #include <linux/dma-mapping.h>
 
+#include <acpi/acpi_drivers.h>
+
 #include <asm/pgtable.h>
 
 #include "internal.h"
@@ -22,7 +24,6 @@ ACPI_MODULE_NAME("scan");
 extern struct acpi_device *acpi_root;
 
 #define ACPI_BUS_CLASS			"system_bus"
-#define ACPI_BUS_HID			"LNXSYBUS"
 #define ACPI_BUS_DEVICE_NAME		"System Bus"
 
 #define ACPI_IS_ROOT_DEVICE(device)    (!(device)->parent)
diff --git a/include/acpi/acpi_drivers.h b/include/acpi/acpi_drivers.h
index 29c6912..3638f6f 100644
--- a/include/acpi/acpi_drivers.h
+++ b/include/acpi/acpi_drivers.h
@@ -52,6 +52,7 @@
 #define ACPI_POWER_HID			"LNXPOWER"
 #define ACPI_PROCESSOR_OBJECT_HID	"LNXCPU"
 #define ACPI_SYSTEM_HID			"LNXSYSTM"
+#define ACPI_BUS_HID			"LNXSYBUS"
 #define ACPI_THERMAL_HID		"LNXTHERM"
 #define ACPI_BUTTON_HID_POWERF		"LNXPWRBN"
 #define ACPI_BUTTON_HID_SLEEPF		"LNXSLPBN"
-- 
Qualcomm Technologies, Inc. on behalf
of the Qualcomm Innovation Center, Inc.  The Qualcomm Innovation Center, Inc. 
is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project.

Re: [PATCH 1/3] ACPI: move ACPI_SYSTEM_HID to acpi_drivers.h

[Rafael J. Wysocki]

On Wed, Mar 16, 2016 at 8:37 PM, Prashanth Prakash
<pprakash@codeaurora.org> wrote:
> From: "Jonathan (Zhixiong) Zhang" <zjzhang@codeaurora.org>
>
> ACPI_SYSTEM_HID is defined in drivers/acpi/bus.c out-of-band.
>
> Move the definition to include/acpi/acpi_drivers.h, together
> with other Linux specific HIDs, so that it can be used by other
> code.
>
> Signed-off-by: Jonathan (Zhixiong) Zhang <zjzhang@codeaurora.org>
> ---
>  drivers/acpi/scan.c         | 3 ++-
>  include/acpi/acpi_drivers.h | 1 +
>  2 files changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c
> index 407a376..358e3c8 100644
> --- a/drivers/acpi/scan.c
> +++ b/drivers/acpi/scan.c
> @@ -13,6 +13,8 @@
>  #include <linux/nls.h>
>  #include <linux/dma-mapping.h>
>
> +#include <acpi/acpi_drivers.h>
> +
>  #include <asm/pgtable.h>
>
>  #include "internal.h"
> @@ -22,7 +24,6 @@ ACPI_MODULE_NAME("scan");
>  extern struct acpi_device *acpi_root;
>
>  #define ACPI_BUS_CLASS                 "system_bus"
> -#define ACPI_BUS_HID                   "LNXSYBUS"

No.

This is not a valid device ID and should never be used as such.

Thanks,
Rafael

Re: [PATCH 1/3] ACPI: move ACPI_SYSTEM_HID to acpi_drivers.h

[Prakash, Prashanth]

Hi Rafael,

On 3/16/2016 7:50 PM, Rafael J. Wysocki wrote:
>  #include "internal.h"
> @@ -22,7 +24,6 @@ ACPI_MODULE_NAME("scan");
>  extern struct acpi_device *acpi_root;
>
>  #define ACPI_BUS_CLASS                 "system_bus"
> -#define ACPI_BUS_HID                   "LNXSYBUS"
> No.
>
> This is not a valid device ID and should never be used as such.
Section 5.6.6 in ACPI 6.1 is defining the notification value for the graceful
shutdown request as a ACPI device specific value(0x81) and it is targeting
the system bus(\_SB), so we took the current approach of creating a sybus
driver to handle the notifications targeting the system bus.

Let me go through the code again and see if we can avoid using the
ACPI_BUS_HID. In the meantime, if you have any specific thoughts on
how this should be handled, please let us know.

Thanks,
Prashanth

Re: [PATCH 1/3] ACPI: move ACPI_SYSTEM_HID to acpi_drivers.h

[Rafael J. Wysocki]

On Thu, Mar 17, 2016 at 5:53 PM, Prakash, Prashanth
<pprakash@codeaurora.org> wrote:
> Hi Rafael,
>
> On 3/16/2016 7:50 PM, Rafael J. Wysocki wrote:
>>  #include "internal.h"
>> @@ -22,7 +24,6 @@ ACPI_MODULE_NAME("scan");
>>  extern struct acpi_device *acpi_root;
>>
>>  #define ACPI_BUS_CLASS                 "system_bus"
>> -#define ACPI_BUS_HID                   "LNXSYBUS"
>> No.
>>
>> This is not a valid device ID and should never be used as such.
> Section 5.6.6 in ACPI 6.1 is defining the notification value for the graceful
> shutdown request as a ACPI device specific value(0x81) and it is targeting
> the system bus(\_SB), so we took the current approach of creating a sybus
> driver to handle the notifications targeting the system bus.
>
> Let me go through the code again and see if we can avoid using the
> ACPI_BUS_HID.

You can.

> In the meantime, if you have any specific thoughts on
> how this should be handled, please let us know.

Your goal is relatively simple.  You want to register a notify handler
for the \_SB object.  You don't need a driver for that, because the
\_SB object is always present, so you don't need to enumerate it or
wait for it to show up etc.

I'd just scan the namespace 1 level below the root and install the
handler for the object with the matching name (not a fake device ID).

Thanks,
Rafael

Re: [PATCH 1/3] ACPI: move ACPI_SYSTEM_HID to acpi_drivers.h

[Prakash, Prashanth]

Than

On 3/17/2016 3:54 PM, Rafael J. Wysocki wrote:
> On Thu, Mar 17, 2016 at 5:53 PM, Prakash, Prashanth
> <pprakash@codeaurora.org> wrote:
>> Hi Rafael,
>>
>> On 3/16/2016 7:50 PM, Rafael J. Wysocki wrote:
>>>  #include "internal.h"
>>> @@ -22,7 +24,6 @@ ACPI_MODULE_NAME("scan");
>>>  extern struct acpi_device *acpi_root;
>>>
>>>  #define ACPI_BUS_CLASS                 "system_bus"
>>> -#define ACPI_BUS_HID                   "LNXSYBUS"
>>> No.
>>>
>>> This is not a valid device ID and should never be used as such.
>> Section 5.6.6 in ACPI 6.1 is defining the notification value for the graceful
>> shutdown request as a ACPI device specific value(0x81) and it is targeting
>> the system bus(\_SB), so we took the current approach of creating a sybus
>> driver to handle the notifications targeting the system bus.
>>
>> Let me go through the code again and see if we can avoid using the
>> ACPI_BUS_HID.
> You can.
>
>> In the meantime, if you have any specific thoughts on
>> how this should be handled, please let us know.
> Your goal is relatively simple.  You want to register a notify handler
> for the \_SB object.  You don't need a driver for that, because the
> \_SB object is always present, so you don't need to enumerate it or
> wait for it to show up etc.
>
> I'd just scan the namespace 1 level below the root and install the
> handler for the object with the matching name (not a fake device ID).
Thanks Rafael! I will update the patch using the above approach.

-Prashanth
> Thanks,
> Rafael
>
>

[PATCH 2/3] ACPI / sybus: add a driver for LNXSYBUS device

[Prashanth Prakash]

This patch adds a bus driver to handle the graceful shutdown
requests as described in the section 6.3.5.1, section 5.6.6
and table 5-143 of ACPI spec 6.1

The OSPM will get a graceful shutdown request via a Notify
operator on \_SB device with a value of 0x81 per section 5.6.6.
Following the shutdown request from platform the OSPM needs to
follow the processing sequence described in section 6.2.5.1 -
"Processing Sequence for Graceful Shutdown Request"

Signed-off-by: Prashanth Prakash <pprakash@codeaurora.org>
Signed-off-by: Jonathan (Zhixiong) Zhang <zjzhang@codeaurora.org>
---
 drivers/acpi/Kconfig  |   6 +++
 drivers/acpi/Makefile |   1 +
 drivers/acpi/sybus.c  | 116 ++++++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 123 insertions(+)
 create mode 100644 drivers/acpi/sybus.c

diff --git a/drivers/acpi/Kconfig b/drivers/acpi/Kconfig
index 82b96ee..78cf770 100644
--- a/drivers/acpi/Kconfig
+++ b/drivers/acpi/Kconfig
@@ -209,6 +209,12 @@ config ACPI_DOCK
 	  This driver supports ACPI-controlled docking stations and removable
 	  drive bays such as the IBM Ultrabay and the Dell Module Bay.
 
+config ACPI_SYBUS
+	bool "System Bus"
+	help
+	  This driver handles the events on system bus which includes
+	  support for a graceful shutdown request from the platform.
+
 config ACPI_CPU_FREQ_PSS
 	bool
 	select THERMAL
diff --git a/drivers/acpi/Makefile b/drivers/acpi/Makefile
index cb648a4..fe56427 100644
--- a/drivers/acpi/Makefile
+++ b/drivers/acpi/Makefile
@@ -80,6 +80,7 @@ obj-$(CONFIG_ACPI_CUSTOM_METHOD)+= custom_method.o
 obj-$(CONFIG_ACPI_BGRT)		+= bgrt.o
 obj-$(CONFIG_ACPI_CPPC_LIB)	+= cppc_acpi.o
 obj-$(CONFIG_ACPI_DEBUGGER_USER) += acpi_dbg.o
+obj-$(CONFIG_ACPI_SYBUS)	+= sybus.o
 
 # processor has its own "processor." module_param namespace
 processor-y			:= processor_driver.o
diff --git a/drivers/acpi/sybus.c b/drivers/acpi/sybus.c
new file mode 100644
index 0000000..f9fa821
--- /dev/null
+++ b/drivers/acpi/sybus.c
@@ -0,0 +1,116 @@
+/*
+ * ACPI System Bus Device (\_SB, LNXSYBUS) Driver
+ * ACPI System Bus Device Driver is used to handle events reported to
+ * the device.
+ *
+ * Copyright (c) 2016, The Linux Foundation. All rights reserved.
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/acpi.h>
+#include <linux/workqueue.h>
+#include <linux/reboot.h>
+
+#include <acpi/acpi_drivers.h>
+
+#define _COMPONENT			ACPI_BUS_COMPONENT
+#define SYBUS_PFX			"ACPI SYBUS: "
+
+/*
+ * According to section 6.3.5 of ACPI 6.0 spec, the kernel
+ * should evaluate _OST (an ACPI control method) every 10 seconds
+ * to indicate "OS shutdown in progress" to the platform.
+ */
+#define SYBUS_INDICATE_INTERVAL		10000
+
+#define SYBUS_NOTIFY_RESERVED		    (0x80)
+#define SYBUS_NOTIFY_SHUTDOWN_REQUEST	(0x81)
+
+ACPI_MODULE_NAME("sybus");
+
+static void sybus_evaluate_ost(struct work_struct *);
+
+static struct acpi_device_id acpi_sybus_ids[] = {
+	{ACPI_BUS_HID, 0},
+	{"", 0},
+};
+MODULE_DEVICE_TABLE(acpi, acpi_sybus_ids);
+
+static acpi_handle sybus_handle;
+static DECLARE_DELAYED_WORK(acpi_sybus_work, sybus_evaluate_ost);
+
+static void sybus_evaluate_ost(struct work_struct *dummy)
+{
+	pr_info(SYBUS_PFX "OS shutdown in progress.\n");
+	acpi_evaluate_ost(sybus_handle, ACPI_OST_EC_OSPM_SHUTDOWN,
+			  ACPI_OST_SC_OS_SHUTDOWN_IN_PROGRESS, NULL);
+	schedule_delayed_work(&acpi_sybus_work,
+			msecs_to_jiffies(SYBUS_INDICATE_INTERVAL));
+}
+
+static void acpi_sybus_notify(struct acpi_device *device, u32 event)
+{
+	/*
+	 * The only event that ACPI System Bus Device should handle is
+	 * SYBUS_NOTIFY_SHUTDOWN_REQUEST.
+	 */
+	if (event != SYBUS_NOTIFY_SHUTDOWN_REQUEST) {
+		ACPI_DEBUG_PRINT((ACPI_DB_INFO,
+			"event %x is not supported by ACPI system bus device.\n",
+			event));
+		return;
+	}
+
+	pr_warn(SYBUS_PFX "Shutdown request notification received.\n");
+
+	if (!delayed_work_pending(&acpi_sybus_work)) {
+		sybus_evaluate_ost(NULL);
+		schedule_delayed_work(&acpi_sybus_work,
+				msecs_to_jiffies(SYBUS_INDICATE_INTERVAL));
+
+		orderly_poweroff(true);
+	} else
+		pr_info(SYBUS_PFX "Shutdown in already progress!\n");
+}
+
+static int acpi_sybus_add(struct acpi_device *device)
+{
+	/* Only one ACPI system bus device */
+	if (sybus_handle)
+		return -EINVAL;
+	sybus_handle = device->handle;
+	return 0;
+}
+
+static int acpi_sybus_remove(struct acpi_device *device)
+{
+	cancel_delayed_work_sync(&acpi_sybus_work);
+	sybus_handle = NULL;
+	return 0;
+}
+
+static struct acpi_driver acpi_sybus_driver = {
+	.name = "system_bus_device",
+	.class = "system_bus",
+	.ids = acpi_sybus_ids,
+	.flags = ACPI_DRIVER_ALL_NOTIFY_EVENTS,
+	.ops = {
+		.notify = acpi_sybus_notify,
+		.add = acpi_sybus_add,
+		.remove = acpi_sybus_remove,
+	},
+};
+module_acpi_driver(acpi_sybus_driver);
+
+MODULE_DESCRIPTION("ACPI System Bus Device Driver");
+MODULE_LICENSE("GPL");
-- 
Qualcomm Technologies, Inc. on behalf
of the Qualcomm Innovation Center, Inc.  The Qualcomm Innovation Center, Inc. 
is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project.

[PATCH 3/3] ACPI: update the definition of common notification values

[Prashanth Prakash]

Per ACPI spec 6.1, the device object notification value of 0x0C is
now reserved, so updating the definition to reflect the same.

Signed-off-by: Prashanth Prakash <pprakash@codeaurora.org>
---
 include/acpi/actypes.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/include/acpi/actypes.h b/include/acpi/actypes.h
index db46546..4553300 100644
--- a/include/acpi/actypes.h
+++ b/include/acpi/actypes.h
@@ -627,7 +627,7 @@ typedef u64 acpi_integer;
 #define ACPI_NOTIFY_DEVICE_PLD_CHECK    (u8) 0x09
 #define ACPI_NOTIFY_RESERVED            (u8) 0x0A
 #define ACPI_NOTIFY_LOCALITY_UPDATE     (u8) 0x0B
-#define ACPI_NOTIFY_SHUTDOWN_REQUEST    (u8) 0x0C
+#define ACPI_NOTIFY_RESERVED_2          (u8) 0x0C
 #define ACPI_NOTIFY_AFFINITY_UPDATE     (u8) 0x0D
 
 #define ACPI_NOTIFY_MAX                 0x0D
-- 
Qualcomm Technologies, Inc. on behalf
of the Qualcomm Innovation Center, Inc.  The Qualcomm Innovation Center, Inc. 
is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project.

Re: [PATCH 0/3] ACPI: Support for platform initiated graceful shutdown

[Al Stone]

On 03/16/2016 01:36 PM, Prashanth Prakash wrote:
> This patch-set adds support for platform initited graceful shutdown as
> described in the sections 5.6.6(Table-143) and 6.3.5.1 of ACPI 6.1 spec
> 
> Jonathan (Zhixiong) Zhang (1):
>   ACPI: move ACPI_SYSTEM_HID to acpi_drivers.h
> 
> Prashanth Prakash (2):
>   ACPI / sybus: add a driver for LNXSYBUS device
>   ACPI: update the definition of common notification values
> 
>  drivers/acpi/Kconfig        |   6 +++
>  drivers/acpi/Makefile       |   1 +
>  drivers/acpi/scan.c         |   3 +-
>  drivers/acpi/sybus.c        | 116 ++++++++++++++++++++++++++++++++++++++++++++
>  include/acpi/acpi_drivers.h |   1 +
>  include/acpi/actypes.h      |   2 +-
>  6 files changed, 127 insertions(+), 2 deletions(-)
>  create mode 100644 drivers/acpi/sybus.c
> 

Is v2 coming along soon?  I'd like to try these out on a couple of
systems.  Thanks, Prashanth.

-- 
ciao,
al
-----------------------------------
Al Stone
Software Engineer
Red Hat, Inc.
ahs3@redhat.com
-----------------------------------

Re: [PATCH 0/3] ACPI: Support for platform initiated graceful shutdown

[Prakash, Prashanth]

Hi Al,

I should have v2 sometime next week.

Sorry about the delay. Vacation and couple of other issues kept me little busy and couldn't update this patch set in a timely manner.

Thanks,
Prashanth

On 5/10/2016 4:50 PM, Al Stone wrote:
> On 03/16/2016 01:36 PM, Prashanth Prakash wrote:
>> This patch-set adds support for platform initited graceful shutdown as
>> described in the sections 5.6.6(Table-143) and 6.3.5.1 of ACPI 6.1 spec
>>
>> Jonathan (Zhixiong) Zhang (1):
>>   ACPI: move ACPI_SYSTEM_HID to acpi_drivers.h
>>
>> Prashanth Prakash (2):
>>   ACPI / sybus: add a driver for LNXSYBUS device
>>   ACPI: update the definition of common notification values
>>
>>  drivers/acpi/Kconfig        |   6 +++
>>  drivers/acpi/Makefile       |   1 +
>>  drivers/acpi/scan.c         |   3 +-
>>  drivers/acpi/sybus.c        | 116 ++++++++++++++++++++++++++++++++++++++++++++
>>  include/acpi/acpi_drivers.h |   1 +
>>  include/acpi/actypes.h      |   2 +-
>>  6 files changed, 127 insertions(+), 2 deletions(-)
>>  create mode 100644 drivers/acpi/sybus.c
>>
> Is v2 coming along soon?  I'd like to try these out on a couple of
> systems.  Thanks, Prashanth.
>

Re: [PATCH 0/3] ACPI: Support for platform initiated graceful shutdown

[Al Stone]

On 05/10/2016 05:28 PM, Prakash, Prashanth wrote:
> Hi Al,
> 
> I should have v2 sometime next week.
> 
> Sorry about the delay. Vacation and couple of other issues kept me little busy and couldn't update this patch set in a timely manner.
> 
> Thanks,
> Prashanth

No worries, Prashanth.  Just curiosity, on my part.  Thanks for posting these.

> On 5/10/2016 4:50 PM, Al Stone wrote:
>> On 03/16/2016 01:36 PM, Prashanth Prakash wrote:
>>> This patch-set adds support for platform initited graceful shutdown as
>>> described in the sections 5.6.6(Table-143) and 6.3.5.1 of ACPI 6.1 spec
>>>
>>> Jonathan (Zhixiong) Zhang (1):
>>>   ACPI: move ACPI_SYSTEM_HID to acpi_drivers.h
>>>
>>> Prashanth Prakash (2):
>>>   ACPI / sybus: add a driver for LNXSYBUS device
>>>   ACPI: update the definition of common notification values
>>>
>>>  drivers/acpi/Kconfig        |   6 +++
>>>  drivers/acpi/Makefile       |   1 +
>>>  drivers/acpi/scan.c         |   3 +-
>>>  drivers/acpi/sybus.c        | 116 ++++++++++++++++++++++++++++++++++++++++++++
>>>  include/acpi/acpi_drivers.h |   1 +
>>>  include/acpi/actypes.h      |   2 +-
>>>  6 files changed, 127 insertions(+), 2 deletions(-)
>>>  create mode 100644 drivers/acpi/sybus.c
>>>
>> Is v2 coming along soon?  I'd like to try these out on a couple of
>> systems.  Thanks, Prashanth.
>>
> 


-- 
ciao,
al
-----------------------------------
Al Stone
Software Engineer
Red Hat, Inc.
ahs3@redhat.com
-----------------------------------