xref: /linux/Documentation/ABI/testing/sysfs-firmware-acpi (revision ca55b2fef3a9373fcfc30f82fd26bc7fccbda732)
1What:		/sys/firmware/acpi/bgrt/
2Date:		January 2012
3Contact:	Matthew Garrett <mjg@redhat.com>
4Description:
5		The BGRT is an ACPI 5.0 feature that allows the OS
6		to obtain a copy of the firmware boot splash and
7		some associated metadata. This is intended to be used
8		by boot splash applications in order to interact with
9		the firmware boot splash in order to avoid jarring
10		transitions.
11
12		image: The image bitmap. Currently a 32-bit BMP.
13		status: 1 if the image is valid, 0 if firmware invalidated it.
14		type: 0 indicates image is in BMP format.
15		version: The version of the BGRT. Currently 1.
16		xoffset: The number of pixels between the left of the screen
17			 and the left edge of the image.
18		yoffset: The number of pixels between the top of the screen
19			 and the top edge of the image.
20
21What:		/sys/firmware/acpi/hotplug/
22Date:		February 2013
23Contact:	Rafael J. Wysocki <rafael.j.wysocki@intel.com>
24Description:
25		There are separate hotplug profiles for different classes of
26		devices supported by ACPI, such as containers, memory modules,
27		processors, PCI root bridges etc.  A hotplug profile for a given
28		class of devices is a collection of settings defining the way
29		that class of devices will be handled by the ACPI core hotplug
30		code.  Those profiles are represented in sysfs as subdirectories
31		of /sys/firmware/acpi/hotplug/.
32
33		The following setting is available to user space for each
34		hotplug profile:
35
36		enabled: If set, the ACPI core will handle notifications of
37			hotplug events associated with the given class of
38			devices and will allow those devices to be ejected with
39			the help of the _EJ0 control method.  Unsetting it
40			effectively disables hotplug for the correspoinding
41			class of devices.
42
43		The value of the above attribute is an integer number: 1 (set)
44		or 0 (unset).  Attempts to write any other values to it will
45		cause -EINVAL to be returned.
46
47What:		/sys/firmware/acpi/hotplug/force_remove
48Date:		May 2013
49Contact:	Rafael J. Wysocki <rafael.j.wysocki@intel.com>
50Description:
51		The number in this file (0 or 1) determines whether (1) or not
52		(0) the ACPI subsystem will allow devices to be hot-removed even
53		if they cannot be put offline gracefully (from the kernel's
54		viewpoint).  That number can be changed by writing a boolean
55		value to this file.
56
57What:		/sys/firmware/acpi/interrupts/
58Date:		February 2008
59Contact:	Len Brown <lenb@kernel.org>
60Description:
61		All ACPI interrupts are handled via a single IRQ,
62		the System Control Interrupt (SCI), which appears
63		as "acpi" in /proc/interrupts.
64
65		However, one of the main functions of ACPI is to make
66		the platform understand random hardware without
67		special driver support.  So while the SCI handles a few
68		well known (fixed feature) interrupts sources, such
69		as the power button, it can also handle a variable
70		number of a "General Purpose Events" (GPE).
71
72		A GPE vectors to a specified handler in AML, which
73		can do a anything the BIOS writer wants from
74		OS context.  GPE 0x12, for example, would vector
75		to a level or edge handler called _L12 or _E12.
76		The handler may do its business and return.
77		Or the handler may send send a Notify event
78		to a Linux device driver registered on an ACPI device,
79		such as a battery, or a processor.
80
81		To figure out where all the SCI's are coming from,
82		/sys/firmware/acpi/interrupts contains a file listing
83		every possible source, and the count of how many
84		times it has triggered.
85
86		$ cd /sys/firmware/acpi/interrupts
87		$ grep . *
88		error:	     0
89		ff_gbl_lock:	   0   enable
90		ff_pmtimer:	  0  invalid
91		ff_pwr_btn:	  0   enable
92		ff_rt_clk:	 2  disable
93		ff_slp_btn:	  0  invalid
94		gpe00:	     0	invalid
95		gpe01:	     0	 enable
96		gpe02:	   108	 enable
97		gpe03:	     0	invalid
98		gpe04:	     0	invalid
99		gpe05:	     0	invalid
100		gpe06:	     0	 enable
101		gpe07:	     0	 enable
102		gpe08:	     0	invalid
103		gpe09:	     0	invalid
104		gpe0A:	     0	invalid
105		gpe0B:	     0	invalid
106		gpe0C:	     0	invalid
107		gpe0D:	     0	invalid
108		gpe0E:	     0	invalid
109		gpe0F:	     0	invalid
110		gpe10:	     0	invalid
111		gpe11:	     0	invalid
112		gpe12:	     0	invalid
113		gpe13:	     0	invalid
114		gpe14:	     0	invalid
115		gpe15:	     0	invalid
116		gpe16:	     0	invalid
117		gpe17:	  1084	 enable
118		gpe18:	     0	 enable
119		gpe19:	     0	invalid
120		gpe1A:	     0	invalid
121		gpe1B:	     0	invalid
122		gpe1C:	     0	invalid
123		gpe1D:	     0	invalid
124		gpe1E:	     0	invalid
125		gpe1F:	     0	invalid
126		gpe_all:    1192
127		sci:	1194
128		sci_not:     0
129
130		sci - The number of times the ACPI SCI
131		has been called and claimed an interrupt.
132
133		sci_not - The number of times the ACPI SCI
134		has been called and NOT claimed an interrupt.
135
136		gpe_all - count of SCI caused by GPEs.
137
138		gpeXX - count for individual GPE source
139
140		ff_gbl_lock - Global Lock
141
142		ff_pmtimer - PM Timer
143
144		ff_pwr_btn - Power Button
145
146		ff_rt_clk - Real Time Clock
147
148		ff_slp_btn - Sleep Button
149
150		error - an interrupt that can't be accounted for above.
151
152		invalid: it's either a GPE or a Fixed Event that
153			doesn't have an event handler.
154
155		disable: the GPE/Fixed Event is valid but disabled.
156
157		enable: the GPE/Fixed Event is valid and enabled.
158
159		Root has permission to clear any of these counters.  Eg.
160		# echo 0 > gpe11
161
162		All counters can be cleared by clearing the total "sci":
163		# echo 0 > sci
164
165		None of these counters has an effect on the function
166		of the system, they are simply statistics.
167
168		Besides this, user can also write specific strings to these files
169		to enable/disable/clear ACPI interrupts in user space, which can be
170		used to debug some ACPI interrupt storm issues.
171
172		Note that only writting to VALID GPE/Fixed Event is allowed,
173		i.e. user can only change the status of runtime GPE and
174		Fixed Event with event handler installed.
175
176		Let's take power button fixed event for example, please kill acpid
177		and other user space applications so that the machine won't shutdown
178		when pressing the power button.
179		# cat ff_pwr_btn
180		0	enabled
181		# press the power button for 3 times;
182		# cat ff_pwr_btn
183		3	enabled
184		# echo disable > ff_pwr_btn
185		# cat ff_pwr_btn
186		3	disabled
187		# press the power button for 3 times;
188		# cat ff_pwr_btn
189		3	disabled
190		# echo enable > ff_pwr_btn
191		# cat ff_pwr_btn
192		4	enabled
193		/*
194		 * this is because the status bit is set even if the enable bit is cleared,
195		 * and it triggers an ACPI fixed event when the enable bit is set again
196		 */
197		# press the power button for 3 times;
198		# cat ff_pwr_btn
199		7	enabled
200		# echo disable > ff_pwr_btn
201		# press the power button for 3 times;
202		# echo clear > ff_pwr_btn	/* clear the status bit */
203		# echo disable > ff_pwr_btn
204		# cat ff_pwr_btn
205		7	enabled
206
207