It was reported that the resume stats received from the firmware are
always zero. This happens because the SMU expects the driver to send the
command to dump the log data after clearing the OS_HINT.

Adjust the order of the commands sent to SMU.
Signed-off-by: NShyam Sundar S K <Shyam-sundar.S-k@amd.com>
Signed-off-by: NSanket Goswami <Sanket.Goswami@amd.com>
Link: https://lore.kernel.org/r/20210921115910.19401-1-Sanket.Goswami@amd.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

The amd_pmc_get_smu_version() and amd_pmc_idlemask_read() functions are
used in the probe / suspend/resume code, so they are also used when
CONFIG_DEBUGFS is disabled, move them outside of the #ifdef CONFIG_DEBUGFS
block.

Note this purely moves the code to above the #ifdef CONFIG_DEBUGFS,
the code is completely unchanged.

Fixes: f6045de1 ("platform/x86: amd-pmc: Export Idlemask values based on the APU")
Cc: Shyam Sundar S K <Shyam-sundar.S-k@amd.com>
Cc: Sanket Goswami <Sanket.Goswami@amd.com>
Reported-by: NNathan Chancellor <nathan@kernel.org>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Adds support for:

- Dell Inspiron 2in1 tablet mode switch notifications. These are
  delivered by a type 0x0011 message with code 0xe070, followed by a
  flag (1 for laptop mode, 0 for tablet mode).

- Recognising (but not otherwise processing) the Dell Ultra Performance
  mode request switch. This is delivered by a type 0x0012 message with
  code 0x000d, followed by a parameter that is either 1 or 2. It is
  not clear what (if anything) should be done with this notification, so
  it is ignored.
Signed-off-by: NTroy Rollo <linux2021@troy.rollo.name>
Link: https://lore.kernel.org/r/20210918073131.2966942-1-linux2021@troy.rollo.nameReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Tested with a AMD Ryzen 7 5800X.
Signed-off-by: NTobias Jakobi <tjakobi@math.uni-bielefeld.de>
Acked-by: NThomas Weißschuh <thomas@weissschuh.net>
Link: https://lore.kernel.org/r/20210921100702.3838-1-tjakobi@math.uni-bielefeld.deSigned-off-by: NHans de Goede <hdegoede@redhat.com>

Some devices, even non convertible ones, can send incorrect
SW_TABLET_MODE reports.

Add an allow list and accept such reports only from devices in it.

Bug reported for Dell XPS 17 9710 on:
https://gitlab.freedesktop.org/libinput/libinput/-/issues/662Reported-by: NTobias Gurtzick <magic@wizardtales.com>
Suggested-by: NHans de Goede <hdegoede@redhat.com>
Tested-by: NTobias Gurtzick <magic@wizardtales.com>
Signed-off-by: NJosé Expósito <jose.exposito89@gmail.com>
Link: https://lore.kernel.org/r/20210920160312.9787-1-jose.exposito89@gmail.com
[hdegoede@redhat.com: Check dmi_switches_auto_add_allow_list only once]
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

When DELL_WMI=y, DELL_WMI_PRIVACY=y, and LEDS_TRIGGER_AUDIO=m, there
is a linker error since the LEDS trigger code is built as a loadable
module. This happens because DELL_WMI_PRIVACY is a bool that depends
on a tristate (LEDS_TRIGGER_AUDIO=m), which can be dangerous.

ld: drivers/platform/x86/dell/dell-wmi-privacy.o: in function `dell_privacy_wmi_probe':
dell-wmi-privacy.c:(.text+0x3df): undefined reference to `ledtrig_audio_get'

Fixes: 8af9fa37 ("platform/x86: dell-privacy: Add support for Dell hardware privacy")
Signed-off-by: NRandy Dunlap <rdunlap@infradead.org>
Cc: Perry Yuan <Perry.Yuan@dell.com>
Cc: Dell.Client.Kernel@dell.com
Cc: platform-driver-x86@vger.kernel.org
Cc: Hans de Goede <hdegoede@redhat.com>
Cc: Mark Gross <mgross@linux.intel.com>
Link: https://lore.kernel.org/r/20210918044829.19222-1-rdunlap@infradead.orgReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

The MODULE_DEVICE_TABLE already creates proper alias for ACPI driver.
Having another MODULE_ALIAS causes the alias to be duplicated.
Signed-off-by: NKrzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20210916170054.136790-1-krzysztof.kozlowski@canonical.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

IdleMask is the metric used by the PM firmware to know the status of each
of the Hardware IP blocks monitored by the PM firmware.

Knowing this value is key to get the information of s2idle suspend/resume
status. This value is mapped to PMC scratch registers, retrieve them
accordingly based on the CPU family and the underlying firmware support.
Co-developed-by: NShyam Sundar S K <Shyam-sundar.S-k@amd.com>
Signed-off-by: NShyam Sundar S K <Shyam-sundar.S-k@amd.com>
Signed-off-by: NSanket Goswami <Sanket.Goswami@amd.com>
Reviewed-by: NMario Limonciello <mario.limonciello@amd.com>
Link: https://lore.kernel.org/r/20210916124002.2529-1-Sanket.Goswami@amd.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

As the PM firmware returns the status of the last s0i3 in the smu_metrics
structure, the existing name "s0i3_cyclecount" seems to be a misnomer.
Change it accordingly to "s0i3_last_entry_status".
Signed-off-by: NSanket Goswami <Sanket.Goswami@amd.com>
Acked-by: NShyam Sundar S K <Shyam-sundar.S-k@amd.com>
Link: https://lore.kernel.org/r/20210916124130.2581-1-Sanket.Goswami@amd.comSigned-off-by: NHans de Goede <hdegoede@redhat.com>

This driver is for accessing the PSE (Programmable Service Engine) - an
Embedded Controller like IP - using ISHTP (Integratd Sensor Hub Transport
Protocol) to get battery, thermal and UCSI (USB Type-C Connector System
Software Interface) related data from the platform.
Signed-off-by: NK Naduvalath, Sumesh <sumesh.k.naduvalath@intel.com>
Reviewed-by: NMark Gross <mgross@linux.intel.com>
Link: https://lore.kernel.org/r/20210913051056.28736-1-sumesh.k.naduvalath@intel.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

checkpatch.pl tool warns about using __attribute__((packed))
"WARNING: __packed is preferred over __attribute__((packed))"
To fix this __attribute__((packed)) is replaced by __packed
Signed-off-by: NJules Irenge <jbi.octave@gmail.com>
Link: https://lore.kernel.org/r/20210912011741.30495-1-jbi.octave@gmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Make `find_guid()` return an acpi_status, and make it handle NULL
pointer GUID strings; and adapt users accordingly.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-31-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Previously, `acpi_wmi_notify_handler()` and `wmi_get_event_data()`
shared more or less the exact same code to query the data for
a particular event.

Introduce a function to get rid of the duplication, and use it
from `acpi_wmi_notify_handler()` and `wmi_get_event_data()`.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-30-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Introduce helper function to determine the appropriate
ACPI type for the input parameter.

This also fixes the following checkpatch warning:
"braces {} are not necessary for any arm of this statement".
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-29-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Instead of "manually" constructing the ACPI method name and
hard-coding sizes in WMI functions, introduce a helper method
which generates the method name for an arbitrary WMI block.

Furthermore, save the appropriate buffer size into a macro.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-28-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Introduce a helper function which wraps the appropriate
`container_of()` macro invocation to convert
a `struct device_driver` to `struct wmi_driver`.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-27-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

The current code carries out the following ACPI status
mapping:

  AE_NOT_FOUND -> AE_OK
  AE_OK        -> AE_OK
  AE_$X        -> AE_$X

That is, everything is mapped to itself, except AE_NOT_FOUND.
The current code does not do it in the most straighforward way.
Simplify the logic.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-26-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Previously, `__query_block()` would fail if the
second WCxx method call failed. However, the
WQxx method might have succeeded, and potentially
allocated memory for the result. Instead of
throwing away the result and potentially
leaking memory, ignore the result of
the second WCxx call.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-25-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Print the event identifier number in addition to
the already printed information, and use %u for
printing unsigned values in `wmi_notify_debug()`.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-24-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Align the arguments of
  * wmi_evaluate_method()
  * wmi_install_notify_handler()
  * wmidev_evaluate_method()
  * find_guid_context()
  * acpi_wmi_ec_space_handler()
  * wmi_char_read()
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-23-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Move some variables in order to keep them
in the narrowest possible scope.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-22-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

The `block` variable is assigned and only used once, the code
shorter and probably clearer without it; so remove it.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-21-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

As per the coding style guide, the preferred way
to pass the size of objects to allocator functions
is `sizeof(*p)`. Use that.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-20-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Other parts of the code use the `!p` idiom to check
for NULL pointers, convert `find_guid_context()` to
do the same.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-19-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Instead of `sprintf()` use the new `sysfs_emit()` function.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-17-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

The `guid_block` struct is overlaid onto a buffer
coming from the _WDG ACPI object of the device.
For this reason mark the struct packed and add
assertions about sizes.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-16-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Instead of hard-coding a 16 long byte array,
use the available `guid_t` type and related methods.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-15-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

The `bool` type is more expressive for a yes/no
kind of value, so use that as the type of the
`enable` parameter of `wmi_method_enable()`.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-13-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Instead of manually creating the bit masks,
use the `BIT()` macro to do it.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-12-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

The `find_guid_context()` is only called from one place,
and `wblock` and `wdriver` cannot be NULL there.
So remove the currently redundant checks.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-11-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Remove an empty line after the last statement
in `acpi_wmi_notify_handler()` which serves
no purpose.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-10-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Void pointers are implictly cast to arbitrary pointer types,
so remove superfluous casts.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-9-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

The GUID block is available for `wmi_create_device()`
through `wblock->gblock`. Use that consistently in
the function instead of using a mix of `gblock` and
`wblock->gblock`.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-8-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

The `status` variable was assigned at the end, and then
immediately returned. Remove it altogether, and return
the previously assigned value directly.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-7-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Some pointers are initialized when they are defined,
but they are almost immediately reassigned in the
following lines. Remove these superfluous assignments.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-6-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

The empty initializer `{ }` is enough to properly initialize
the terminating acpi_device_id entry in the device table,
so use that.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-5-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Remove commas that are after terminating entries in arrays.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-4-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

Fix the following two checkpatch warnings:

 * "space required before the open parenthesis '('"
 * "that open brace { should be on the previous line"
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-3-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

The kernel doc erroneously specified `wmi_uninstall_notify_handler()`
for the `wmi_remove_notify_handler()` function. Fix that.
Signed-off-by: NBarnabás Pőcze <pobrn@protonmail.com>
Link: https://lore.kernel.org/r/20210904175450.156801-2-pobrn@protonmail.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>

A number of upcoming notebook computer designs drive the internal
display panel's backlight PWM through the Embedded Controller (EC).
This EC-based backlight control can be plumbed through to an ACPI
"WMAA" method interface, which in turn can be wrapped by WMI with
the GUID handle 603E9613-EF25-4338-A3D0-C46177516DB7.

Add a new driver, aliased to the WMAA WMI GUID, to expose a sysfs
backlight class driver to control backlight levels on systems with
EC-driven backlights.
Signed-off-by: NDaniel Dadap <ddadap@nvidia.com>
Reviewed-by: NThomas Weißschuh <linux@weissschuh.net>
Link: https://lore.kernel.org/r/20210903003838.15797-1-ddadap@nvidia.comReviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>