Currently attributes will show things like:
`BootOrderLock,Disable`
rather than just
`Disable`.
Of course this works, but the attribute is intended to be read by
userspace tools and not require further processing. That is a userspace
tool can display a drop down of `possible_values` and `current_value` is
one of them from the list.
This also aligns `think-lmi` with how `dell-wmi-sysman` works.
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Link: https://lore.kernel.org/r/20210622200755.12379-3-mario.limonciello@amd.com
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
On an AMD based Lenovo T14, I find that the module doesn't work at
all, and instead has a traceback with messages like:
```
sysfs: cannot create duplicate filename '/devices/virtual/firmware-attributes/thinklmi/attributes/Reserved'
```
Duplicate and reserved values showing up appear to be a firmware bug,
but they shouldn't make the driver explode. So catch them and skip
them.
Fixes: a40cd7ef22 ("platform/x86: think-lmi: Add WMI interface support on Lenovo platforms")
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Link: https://lore.kernel.org/r/20210622200755.12379-2-mario.limonciello@amd.com
[hdegoede@redhat.com: Add missing kfree(tlmi_priv.setting[i])]
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Commit 0ddcf3a6b4 ("platform/x86: think-lmi: Avoid potential read before
start of the buffer") moved the length == 0 up to before stripping the '\n'
which typically gets added when users echo a value to a sysfs-attribute
from the shell.
This avoids a potential buffer-underrun, but it also causes a behavioral
change, prior to this change "echo > kbdlang", iow writing just a single
'\n' would result in an EINVAL error, but after the change this gets
accepted setting kbdlang to an empty string.
Fix this by replacing the manual '\n' check with using strchrnul() to get
the length till '\n' or terminating 0 in one go; and then do the
length != 0 check after this.
Fixes: 0ddcf3a6b4 ("platform/x86: think-lmi: Avoid potential read before start of the buffer")
Reported-by: Juha Leppänen <juha_efku@dnainternet.net>
Suggested-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20210621193648.44138-1-hdegoede@redhat.com
There is a problem in mapping CPU to a PCI device instance when the
bus numbers are reused in different packages. This was observed on
some Sapphire Rapids systems.
The current implementation reads bus number assigned to a CPU package
via MSR 0x128. This allows to establish relationship between a CPU
and a PCI device. This allows to update power related parameters to a
MMIO offset in a PCI device space which is unique to a CPU. But if
two packages uses same bus number then this mapping will not be unique.
When bus number is reused, PCI device will use different domain number
or segment number. So we need to be aware of this domain information
while matching CPU to PCI bus number. This domain information is not
available via any MSR. So need to use ACPI numa node information.
There is an interface already available in the Linux to read numa
node for a CPU and a PCI device. This change uses this interface
to check the numa node of a match PCI device with bus number.
If the bus number and numa node matches with the CPU's assigned
bus number and numa node, the matched PCI device instance will be
returned to the caller.
It is possible that before Sapphire Rapids, the numa node is not
defined for the Speed Select PCI device in some OEM systems. In this
case to restore old behavior, return the last matched PCI device
for domain 0 unlsess there are more than one matches.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Link: https://lore.kernel.org/r/20210616221329.1909276-2-srinivas.pandruvada@linux.intel.com
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
It was observed that some of the high performance benchmarks are spending
more time in kernel depending on which CPU package they are executing.
The difference is significant and benchmark scores varies more than 10%.
These benchmarks adjust class of service to improve thread performance
which run in parallel. This class of service change causes access to
MMIO region of Intel Speed Select PCI devices depending on the CPU
package they are executing.
This mapping from CPU to PCI device instance uses a standard Linux PCI
interface "pci_get_domain_bus_and_slot()". This function does a linear
search to get to a PCI device. Since these platforms have 100+ PCI
devices, this search can be expensive in fast path for benchmarks.
Since the device and function of PCI device is fixed for Intel
Speed Select PCI devices, the CPU to PCI device information can be cached
at the same time when bus number for the CPU is read. In this way during
runtime the cached information can be used. This improves performance
of these benchmarks significantly.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Link: https://lore.kernel.org/r/20210616221329.1909276-1-srinivas.pandruvada@linux.intel.com
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
ACPI devices with _HID INT3472 are currently matched to the tps68470
driver, however this does not cover all situations in which that _HID
occurs. We've encountered three possibilities:
1. On Chrome OS devices, an ACPI device with _HID INT3472 (representing
a physical TPS68470 device) that requires a GPIO and OpRegion driver
2. On devices designed for Windows, an ACPI device with _HID INT3472
(again representing a physical TPS68470 device) which requires GPIO,
Clock and Regulator drivers.
3. On other devices designed for Windows, an ACPI device with _HID
INT3472 which does **not** represent a physical TPS68470, and is instead
used as a dummy device to group some system GPIO lines which are meant
to be consumed by the sensor that is dependent on this entry.
This commit adds a new module, registering a platform driver to deal
with the 3rd scenario plus an i2c driver to deal with #1 and #2, by
querying the CLDB buffer found against INT3472 entries to determine
which is most appropriate.
Suggested-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Daniel Scally <djrscally@gmail.com>
Link: https://lore.kernel.org/r/20210603224007.120560-6-djrscally@gmail.com
[hdegoede@redhat.com Make skl_int3472_tps68470_calc_type() static]
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
tlmi_priv.pwd_admin->password is an array (not a pointer), so the correct
way to check for the password being set is to check for
tlmi_priv.pwd_admin->password[0] != 0.
For the second check, replace the check with checking that auth_str is
set instead.
Cc: Mark Pearson <markpearson@lenovo.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Reported-by: coverity-bot <keescook+coverity-bot@chromium.org>
Addresses-Coverity-ID: 1505158 ("NO_EFFECT")
Fixes: a7314b3b1d8a ("platform/x86: think-lmi: Add WMI interface support on Lenovo platforms")
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20210609151752.156902-1-hdegoede@redhat.com
The dell-wmi-sysman and think-lmi kernel modules both have a global
struct class *fw_attr_class variable, leading to the following compile
errors when both are builtin:
ld: drivers/platform/x86/think-lmi.o:(.bss+0x0): multiple definition of `fw_attr_class'; drivers/platform/x86/dell/dell-wmi-sysman/sysman.o:(.bss+0x0): first defined here
In both cases the variable is only used in the file where it is declared.
Make both declarations static to avoid the linker error.
Cc: Mark Pearson <markpearson@lenovo.com>
Cc: Dell.Client.Kernel@dell.com
Reported-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20210609145952.113393-1-hdegoede@redhat.com
The status variable in ssam_nf_refcount_disable_free() is only set when
the reference count equals zero. Otherwise, it is returned
uninitialized. Fix this by always initializing status to zero.
Reported-by: kernel test robot <lkp@intel.com>
Fixes: 640ee17199e4 ("platform/surface: aggregator: Allow enabling of events without notifiers")
Signed-off-by: Maximilian Luz <luzmaximilian@gmail.com>
Link: https://lore.kernel.org/r/20210604210907.25738-2-luzmaximilian@gmail.com
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
While events can already be enabled and disabled via the generic request
IOCTL, this bypasses the internal reference counting mechanism of the
controller. Due to that, disabling an event will turn it off regardless
of any other client having requested said event, which may break
functionality of that client.
To solve this, add IOCTLs wrapping the ssam_controller_event_enable()
and ssam_controller_event_disable() functions, which have been
previously introduced for this specific purpose.
Signed-off-by: Maximilian Luz <luzmaximilian@gmail.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20210604134755.535590-6-luzmaximilian@gmail.com
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Currently, debugging unknown events requires writing a custom driver.
This is somewhat difficult, slow to adapt, and not entirely
user-friendly for quickly trying to figure out things on devices of some
third-party user. We can do better. We already have a user-space
interface intended for debugging SAM EC requests, so let's add support
for receiving events to that.
This commit provides support for receiving events by reading from the
controller file. It additionally introduces two new IOCTLs to control
which event categories will be forwarded. Specifically, a user-space
client can specify which target categories it wants to receive events
from by registering the corresponding notifier(s) via the IOCTLs and
after that, read the received events by reading from the controller
device.
Signed-off-by: Maximilian Luz <luzmaximilian@gmail.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20210604134755.535590-5-luzmaximilian@gmail.com
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
We can already enable and disable SAM events via one of two ways: either
via a (non-observer) notifier tied to a specific event group, or a
generic event enable/disable request. In some instances, however,
neither method may be desirable.
The first method will tie the event enable request to a specific
notifier, however, when we want to receive notifications for multiple
event groups of the same target category and forward this to the same
notifier callback, we may receive duplicate events, i.e. one event per
registered notifier. The second method will bypass the internal
reference counting mechanism, meaning that a disable request will
disable the event regardless of any other client driver using it, which
may break the functionality of that driver.
To address this problem, add new functions that allow enabling and
disabling of events via the event reference counting mechanism built
into the controller, without needing to register a notifier.
This can then be used in combination with observer notifiers to process
multiple events of the same target category without duplication in the
same callback function.
Signed-off-by: Maximilian Luz <luzmaximilian@gmail.com>
Link: https://lore.kernel.org/r/20210604134755.535590-3-luzmaximilian@gmail.com
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Currently, each SSAM event notifier is directly tied to one group of
events. This makes sense as registering a notifier will automatically
take care of enabling the corresponding event group and normally drivers
only need notifications for a very limited number of events, associated
with different callbacks for each group.
However, there are rare cases, especially for debugging, when we want to
get notifications for a whole event target category instead of just a
single group of events in that category. Registering multiple notifiers,
i.e. one per group, may be infeasible due to two issues: a) we might not
know every event enable/disable specification as some events are
auto-enabled by the EC and b) forwarding this to the same callback will
lead to duplicate events as we might not know the full event
specification to perform the appropriate filtering.
This commit introduces observer-notifiers, which are notifiers that are
not tied to a specific event group and do not attempt to manage any
events. In other words, they can be registered without enabling any
event group or incrementing the corresponding reference count and just
act as silent observers, listening to all currently/previously enabled
events based on their match-specification.
Essentially, this allows us to register one single notifier for a full
event target category, meaning that we can process all events of that
target category in a single callback without duplication. Specifically,
this will be used in the cdev debug interface to forward events to
user-space via a device file from which the events can be read.
Signed-off-by: Maximilian Luz <luzmaximilian@gmail.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20210604134755.535590-2-luzmaximilian@gmail.com
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
It causes mlxreg-hotplug probing failure: request_threaded_irq()
returns -EINVAL due to true value of condition:
((irqflags & IRQF_SHARED) && (irqflags & IRQF_NO_AUTOEN))
after flag "IRQF_NO_AUTOEN" has been added to:
err = devm_request_irq(&pdev->dev, priv->irq,
mlxreg_hotplug_irq_handler, IRQF_TRIGGER_FALLING
| IRQF_SHARED | IRQF_NO_AUTOEN,
"mlxreg-hotplug", priv);
This reverts commit bee3ecfed0 ("platform/mellanox: mlxreg-hotplug: move
to use request_irq by IRQF_NO_AUTOEN flag").
Signed-off-by: Mykola Kostenok <c_mykolak@nvidia.com>
Acked-by: Vadim Pasternak <vadimp@nvidia.com>
Link: https://lore.kernel.org/r/20210603172827.2599908-1-c_mykolak@nvidia.com
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Disabling events silently fails due to the wrong command ID being used.
Instead of the command ID for the disable call, the command ID for the
enable call was being used. This causes the disable call to enable the
event instead. As the event is already enabled when we call this
function, the EC silently drops this command and does nothing.
Use the correct command ID for disabling the event to fix this.
Fixes: c167b9c7e3 ("platform/surface: Add Surface Aggregator subsystem")
Signed-off-by: Maximilian Luz <luzmaximilian@gmail.com>
Link: https://lore.kernel.org/r/20210603000636.568846-1-luzmaximilian@gmail.com
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
add support for Dell privacy driver for the Dell units equipped
hardware privacy design, which protect users privacy of audio and
camera from hardware level. Once the audio or camera privacy mode
activated, any applications will not get any audio or video stream
when user pressed ctrl+F4 hotkey, audio privacy mode will be
enabled, micmute led will be also changed accordingly
The micmute led is fully controlled by hardware & EC(embedded controller)
and camera mute hotkey is Ctrl+F9. Currently design only emits
SW_CAMERA_LENS_COVER event while the camera lens shutter will be
changed by EC & HW(hardware) control
*The flow is like this:
1) User presses key. HW does stuff with this key (timeout timer is started)
2) WMI event is emitted from BIOS to kernel
3) WMI event is received by dell-privacy
4) KEY_MICMUTE emitted from dell-privacy
5) Userland picks up key and modifies kcontrol for SW mute
6) Codec kernel driver catches and calls ledtrig_audio_set
7) dell-privacy notifies EC, the timeout is cancelled and the HW mute
is activated. If the EC is not notified then the HW mic mute will
activate when the timeout triggers, just a bit later than with the
active ack.
Signed-off-by: Perry Yuan <perry_yuan@dell.com>
Link: https://lore.kernel.org/r/20210506115605.1504-1-Perry_Yuan@Dell.com
[hdegoede@redhat.com: Rework Kconfig/Makefile bits + other small fixups]
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Rename dell-wmi.c to dell-wmi-base.c, so that we can have other
dell-wmi-foo.c files which can be added to dell-wmi.ko as "plugins"
controlled by separate boolean Kconfig options.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
VPC event bit 10 gets set on a Yoga 300-11IBR when the EC believes that the
device has changed between laptop/tent/stand/tablet mode.
The EC relies on getting angle info from 2 accelerometers through a special
windows service calling a DSM on the DUAL250E ACPI-device. Linux does not
do this, making the laptop/tent/stand/tablet mode info unreliable.
Ignore VPC event bit 10 to avoid the warnings triggered by the default case
in ideapad_acpi_notify().
Note that the plan for Linux is to have iio-sensor-proxy read the 2
accelerometers and have it provide info about which mode 360° hinges
2-in-1s to the rest of userspace:
https://gitlab.freedesktop.org/hadess/iio-sensor-proxy/-/issues/216
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20210523172331.177834-1-hdegoede@redhat.com
