linux-toradex.git/drivers/acpi/sysfs.c, branch v4.12

Merge branches 'acpi-button', 'acpica' and 'acpi-sysfs'

2017-06-02T22:03:29+00:00

* acpi-button:
  Revert "ACPI / button: Change default behavior to lid_init_state=open"

* acpica:
  ACPICA: Tables: Fix regression introduced by a too early mechanism enabling

* acpi-sysfs:
  ACPI / sysfs: fix acpi_get_table() leak / acpi-sysfs denial of service

ACPI / sysfs: fix acpi_get_table() leak / acpi-sysfs denial of service

2017-04-26T22:27:49+00:00

Reading an ACPI table through the /sys/firmware/acpi/tables interface
more than 65,536 times leads to the following log message:

 ACPI Error: Table ffff88033595eaa8, Validation count is zero after increment
  (20170119/tbutils-423)

...and the table being unavailable until the next reboot. Add the
missing acpi_put_table() so the table ->validation_count is decremented
after each read.

Reported-by: Anush Seetharaman 
Fixes: 174cc7187e6f "ACPICA: Tables: Back port acpi_get_table_with_size() ..."
Signed-off-by: Dan Williams 
Cc: 4.10+  # 4.10+
Signed-off-by: Rafael J. Wysocki

ACPI / scan: Drop support for force_remove

2017-04-13T01:51:47+00:00

/sys/firmware/acpi/hotplug/force_remove was presumably added to support
auto offlining in the past. This is, however, inherently dangerous for
some hotplugable resources like memory. The memory offlining fails when
the memory is still in use and cannot be dropped or migrated. If we
ignore the failure we are basically allowing for subtle memory
corruption or a crash.

We have actually noticed the later while hitting BUG() during the memory
hotremove (remove_memory):
	ret = walk_memory_range(PFN_DOWN(start), PFN_UP(start + size - 1), NULL,
			check_memblock_offlined_cb);
	if (ret)
		BUG();

it took us quite non-trivial time realize that the customer had
force_remove enabled. Even if the BUG was removed here and we could
propagate the error up the call chain it wouldn't help at all because
then we would hit a crash or a memory corruption later and harder to
debug. So force_remove is unfixable for the memory hotremove. We haven't
checked other hotplugable resources to be prone to a similar problems.

Remove the force_remove functionality because it is not fixable currently.
Keep the sysfs file and report an error if somebody tries to enable it.
Encourage users to report about the missing functionality and work with
them with an alternative solution.

Reviewed-by: Lee, Chun-Yi 
Signed-off-by: Michal Hocko 
Signed-off-by: Rafael J. Wysocki

ACPI / sysfs: Provide quirk mechanism to prevent GPE flooding

2016-12-26T22:16:39+00:00

Sometimes, the users may require a quirk to be provided from ACPI subsystem
core to prevent a GPE from flooding.
Normally, if a GPE cannot be dispatched, ACPICA core automatically prevents
the GPE from firing. But there are cases the GPE is dispatched by _Lxx/_Exx
provided via AML table, and OSPM is lacking of the knowledge to get
_Lxx/_Exx correctly executed to handle the GPE, thus the GPE flooding may
still occur.

The existing quirk mechanism can be enabled/disabled using the following
commands to prevent such kind of GPE flooding during runtime:
 # echo mask > /sys/firmware/acpi/interrupts/gpe00
 # echo unmask > /sys/firmware/acpi/interrupts/gpe00
To avoid GPE flooding during boot, we need a boot stage mechanism.

This patch provides such a boot stage quirk mechanism to stop this kind of
GPE flooding. This patch doesn't fix any feature gap but since the new
feature gaps could be found in the future endlessly, and can disappear if
the feature gaps are filled, providing a boot parameter rather than a DMI
table should suffice.

Link: https://bugzilla.kernel.org/show_bug.cgi?id=53071
Link: https://bugzilla.kernel.org/show_bug.cgi?id=117481
Link: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/887793
Signed-off-by: Lv Zheng 
Signed-off-by: Rafael J. Wysocki

ACPI / sysfs: Update sysfs signature handling code

2016-09-16T23:12:34+00:00

This patch cleans up sysfs table signature handling code:

 1. Convert the signature handling code to use the ACPICA APIs to
    benefit from the future improvements of the APIs.

 2. Add 'filename' attribute in order to handle both BE/LE name tags.

 3. Add instance check in order to avoid the possible buffer overflow
    related to the table file name.

Signed-off-by: Lv Zheng 
[ rjw: Changelog ]
Signed-off-by: Rafael J. Wysocki

ACPI / sysfs: Fix an issue for LoadTable opcode

2016-09-16T23:10:43+00:00

OEM tables can be installed via RSDT/XSDT, in this case, they have already
been created under /sys/firmware/acpi/tables.

For this kind of tables, normally LoadTable opcode will be executed to load
them. If LoadTable opcode is executed after acpi_sysfs_init(),
acpi_sysfs_table_handler() will be invoked, thus a redundant table file
will be created under /sys/firmware/acpi/tables/dynamic. Then running
"acpidump" on such platform results in an error, complaining blank empty
table (see Link 1 below).

The bug can be reproduced by customizing an OEM1 table, allowing it to be
overridden via 'table_sigs' (drivers/acpi/tables.c), adding the following
code to the customized DSDT to load it:

  Name (OEMH, Zero)
  Name (OEMF, One)
  If (LEqual (OEMF, One)) {
      Store (LoadTable ("OEM1", "Intel", "Test"), OEMH)
      Store (Zero, OEMF)
  }

In order to make sure that the OEM1 table is installed after
acpi_sysfs_init(), acpi_sysfs_init() can be moved before invoking
acpi_load_tables(). Then the following command execution result can be
seen:
 # acpidump > acpidump.txt
 Could not read table header: /sysfs/firmware/acpi/tables/dynamic/OEM12
 Could not get ACPI table at index 17, AE_BAD_HEADER

Link: https://bugzilla.kernel.org/show_bug.cgi?id=150841 # [1]
Link: https://github.com/acpica/acpica/commit/ed6a5fbc
Reported-by: Jason Voelz 
Reported-by: Francisco Leoner 
Signed-off-by: Lv Zheng 
Signed-off-by: Rafael J. Wysocki

ACPI / sysfs: Use new GPE masking mechanism in GPE interface

2016-08-17T00:10:44+00:00

Now GPE can be masked via the new acpi_mask_gpe() API and this patch
modifies /sys/firmware/acpi/interrupts/gpexx to use this new facility.

Writes "mask/unmask" to this file now invokes acpi_mask_gpe().

Reads from this file now returns new "EN/STS" when the corresponding GPE
hardware register's EN/STS bits are flagged, and new "masked/unmasked"
attribute to indicate the status of the masking mechanism.

Signed-off-by: Lv Zheng 
[ rjw: Subject ]
Signed-off-by: Rafael J. Wysocki

ACPI: add support for ACPI reconfiguration notifiers

2016-07-08T19:52:35+00:00

Add support for ACPI reconfiguration notifiers to allow subsystems
to react to changes in the ACPI tables that happen after the initial
enumeration. This is similar with the way dynamic device tree
notifications work.

The reconfigure notifications supported for now are device add and
device remove.

Since ACPICA allows only one table notification handler, this patch
makes the table notifier function generic and moves it out of the
sysfs specific code.

Signed-off-by: Octavian Purdila 
Reviewed-by: Mika Westerberg 
Signed-off-by: Rafael J. Wysocki

ACPI / sysfs: fix error code in get_status()

2016-05-05T13:28:58+00:00

The problem with ornamental, do-nothing gotos is that they lead to
"forgot to set the error code" bugs.  We should be returning -EINVAL
here but we don't.  It leads to an uninitalized variable in
counter_show():

    drivers/acpi/sysfs.c:603 counter_show()
    error: uninitialized symbol 'status'.

Fixes: 1c8fce27e275 (ACPI: introduce drivers/acpi/sysfs.c)
Signed-off-by: Dan Carpenter 
Signed-off-by: Rafael J. Wysocki

ACPI / sysfs: correctly check failing memory allocation

2015-10-26T03:57:27+00:00

Since kobject_create_and_add() can fail under memory pressure,
its return value needs to be checked against NULL before passing
it to sysfs_create_file().

Signed-off-by: Insu Yun 
[ rjw: Subject & changelog ]
Signed-off-by: Rafael J. Wysocki