The turbo-freq enable with auto mode, prints result for the last possible
CPU, which is not correct when either CPU is not present or user wants
command to be limited to a single die/package. For example, in the
below command user wants to limit to die/package 0, but the
"turbo-freq --auto" result is displayed using the other package.

$ sudo intel-speed-select -c 0 turbo-freq enable -a
Intel(R) Speed Select Technology
 package-0
  die-0
    cpu-0
      turbo-freq
        enable:success
 package--1
  die-0
    cpu-31
      turbo-freq --auto
        enable:success

Since we do have to traverse all CPUs, don't display CPU info for
"turbo-freq --auto", as we already displayed the result for
turbo-freq enable with the CPU information.

With the fix, the same command results in:

$ sudo intel-speed-select -c 0 turbo-freq enable -a
Intel(R) Speed Select Technology
package-0
  die-0
    cpu-0
      turbo-freq
        enable:success
      turbo-freq --auto
        enable:success

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

It is possible that BIOS may not enable core-power feature. In this case
this additional interface will allow to enable from this utility. Also
the information dump, includes the current status of core-power.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

To discover core-power capability, some new mailbox commands are added. Allow
those commands to execute.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Add support for new Mellanox system types of basic class VMOD0010,
containing new Mellanox systems equipped with new switch device
Spectrum 3 (32x400GbE/64x200G/128x100G Ethernet switch).
These are the Top of the Rack 1U/2U/4U systems, equipped with
Mellanox Comex card and with the switch board with Mellanox Spectrum-3
device.
This class of devices can be equipped with two PS units for 1U/2U or
with four PS units for 4U systems.

Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Add support for capability register, which is used for detection of the
actual number of interrupt capable components within the particular
group, supported by the specific system.
Such components could be for example the number of power units and
interrupts related to these units.
The motivation is to avoid adding a new code in the future in order to
distinct between the systems type supported different number of the
components like power supplies, FANs, ASICs, line cards.

Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Add support for capability register, which contains information about
the number of PS units equipped on the system and about minimum I2C
frequency supported by the all system's I2C devices.
Utilization of this register allows to avoid necessity of providing new
system description, in case it differs in number of PS units or in
minimal I2C frequency.

Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Add support for new Mellanox system types of basic class VMOD0009,
containing Mellanox systems equipped with the switch devices
Spectrum 1 (32x100GbE Ethernet switch) and Switch-IB/Switch-IB2
(36x100Gbe InfiniBand switch).
These are the Top of the Rack system, equipped with Mellanox Comex
card.

Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Separate assignment for systems mux configuration based on system type,
instead of setting the same configuration for the all.
The motivation is to allow introduction of new systems types with the
different mux topology.

Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Add documentation for the new attributes for:
- Exposing reset causes types asserted by: platform reset, SoC reset,
  AC power failure, software power off request.
- Setting and removing system VPD (EEPROM) hardware write protection.
- Voltage regulator devices configuration update status and firmware
  version.
- Setting PCIe ASIC reset to disable or enable state during PCIe root
  complex reset.
- System static topology identification, like system's static I2C
  topology, number and type of FPGA devices within the system and so on.

Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Add new attributes for "next-generation" type systems:
- Reset cause indication, when system reset has been caused by the
  platform reset request through CPLD, by AC power failure,
  by software power off request through CPLD. by signal asserted by SOC
  through ACPI register. It introduces more reset causes, which can
  be monitored after the reboots.
- Setting and removing system VPD (EEPROM) hardware write protection.
  It allows to access VPD during production cycle and prevents VPD
  corruption on system in field.
- Voltage regulator devices configuration update status and version. It
  allows to monitor configuration update status and current active
  version for such sort of devices.
- PCIe ASIC reset disable - when set ASIC will go down upon PCIe
  root complex reset, otherwise ASIC will stay up during PCIe root
  complex reset.
- System configuration Ids to provide system static topology
  identification, like system's static I2C topology, number and type of
  FPGA devices within the system and so on.

Add the existing attribute for "iio" bank selection for system type
"msn21xx".

All the above attributes are exposed through "sysfs" by "mlxreg-io"
driver.

Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Remove blank lines between "What" and "Date" keywords.
Start each section with "What" keyword.

Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Add missed "cpld4_version" attribute.

Fixes: 52675da1

 ("Documentation/ABI: Add new attribute for mlxreg-io sysfs interfaces")
Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Fix attribute name from "jtag_enable", which described twice to
"cpld3_version", which is expected to be instead of second appearance
of "jtag_enable".

Fixes: 2752e344

 ("Documentation/ABI: Add new attribute for mlxreg-io sysfs interfaces")
Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Remove redundant semicolons at the end of few functions.

Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Add an entry for drivers/platform/x86/intel-uncore-frequency.c.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Some server users set limits on the uncore frequency using MSR 620H, while
running latency sensitive workloads. Here uncore frequency controls
RING/LLC(last-level cache) clocks.

But MSR control is not always possible from the user space, so this driver
provides a sysfs interface to set max and min frequency limits. This MSR
620H is a die scoped in multi-die system or package scoped in non multi-die
systems.

When this driver is loaded, a new directory is created under
 /sys/devices/system/cpu.

For example on a two package Skylake server:
$cd /sys/devices/system/cpu/intel_uncore_frequency

$ls
package_00_die_00 package_01_die_00

$ls package_00_die_00
max_freq_khz  min_freq_khz  initial_max_freq_khz
initial_min_freq_khz

$grep . *
    max_freq_khz:2400000
    min_freq_khz:1200000
    initial_max_freq_khz:2400000
    initial_min_freq_khz:1200000

Here, initial_max_freq_khz and initial_min_freq_khz are read only
attributes to show power up or initial values of max and min frequencies
respectively. Other attributes are read-write, so that users can modify.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Some of ASUS laptops like UX431FL keyboard backlight cannot be set to
brightness 0. According to ASUS' information, the brightness should be
0x80 ~ 0x83. This patch fixes it by following the logic.

Fixes: e9809c0b

 ("asus-wmi: add keyboard backlight support")
Signed-off-by: Jian-Hong Pan <jian-hong@endlessm.com>
Reviewed-by: Daniel Drake <drake@endlessm.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Fix ambiguous spelling in the comments.

While here, update two lines to fit one.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

The timeout loop look more naturally when done like

	unsigned long timeout = ...;
	...
	do {
		...
		if (cond)
			return %OK;

		sleep(...);
	} while (time_before(timeout));

	...print timeout error...
	return %ERROR;

It also saves LOCs. Convert the driver to this format of timeout loop.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

ASUS TUF FX705DY/FX505DY starts in silent mode and under heavy
CPU load it overheats and drops CPU frequency to 399MHz and stays
at it until reboot [1]. Set throttle thermal policy to default
to avoid overheating and throttlig.

[1] Link: https://bugzilla.kernel.org/show_bug.cgi?id=203733



Signed-off-by: Leonid Maksymchuk <leonmaxx@gmail.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Throttle thermal policy ACPI device is used to control CPU cooling and
throttling. This patch adds sysfs entry for setting current mode and
Fn+F5 hotkey that switches to next.

Policy modes:
* 0x00 - default
* 0x01 - overboost
* 0x02 - silent

Signed-off-by: Leonid Maksymchuk <leonmaxx@gmail.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

1. Put on one line PROPERTY_ENTRY_STRING("firmware-name", ...) lines. In most
   cases this doesn't over the limit, the rest is pretty much obvious from
   reading point of view.
2. Put on one line DMI_MATCH() parameters, we already have over the limit lines.
3. Reduce indentation of function parameters.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Add touchscreen info for the PiPO W11 tablet.

Signed-off-by: Tim Josten <timjosten@gmail.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Adding new CML CPU model ID into platform driver support list.

Signed-off-by: Harry Pan <harry.pan@intel.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Allow the user to configure the fan to turn on / speed-up at lower
thresholds then before (20 degrees Celcius as minimum instead of 40) and
likewise also allow the user to delay the fan speeding-up till the
temperature hits 90 degrees Celcius (was 70).

Cc: Jason Anderson <jasona.594@gmail.com>
Reported-by: Jason Anderson <jasona.594@gmail.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Use our default values when wrong module-parameters are given, instead of
refusing to load. Refusing to load leaves the fan at the BIOS default
setting, which is "Off". The CPU's thermal throttling should protect the
system from damage, but not-loading is really not the best fallback in this
case.

This commit fixes this by re-setting module-parameter values to their
defaults if they are out of range, instead of failing the probe with
-EINVAL.

Cc: stable@vger.kernel.org
Cc: Jason Anderson <jasona.594@gmail.com>
Reported-by: Jason Anderson <jasona.594@gmail.com>
Fixes: 594ce6db

 ("platform/x86: GPD pocket fan: Use a min-speed of 2 while charging")
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

This commit fixes the potential deadlock caused by the console Rx
and Tx processing at the same time. Rx and Tx both take the console
and tmfifo spinlock but in different order which causes potential
deadlock. The fix is to use different tmfifo spinlock for Rx and
Tx since they protect different resources and it's safe to split
the lock.

Below is the reported call trace when copying/pasting large string
in the console.

Rx:
    _raw_spin_lock_irqsave (hvc lock)
    __hvc_poll
    hvc_poll
    in_intr
    vring_interrupt
    mlxbf_tmfifo_rxtx_one_desc (tmfifo lock)
    mlxbf_tmfifo_rxtx
    mlxbf_tmfifo_work_rxtx
Tx:
    _raw_spin_lock_irqsave (tmfifo lock)
    mlxbf_tmfifo_virtio_notify
    virtqueue_notify
    virtqueue_kick
    put_chars
    hvc_push
    hvc_write (hvc lock)
    ...
    do_tty_write
    tty_write

Fixes: 1357dfd7

 ("platform/mellanox: Add TmFifo driver for Mellanox BlueField Soc")
Cc: <stable@vger.kernel.org> # 5.4+
Reviewed-by: David Woods <dwoods@mellanox.com>
Signed-off-by: Liming Sun <lsun@mellanox.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

This patch corrects the SPDX License Identifier style in
header files related to x86 Platform Specific Drivers.
For C header files Documentation/process/license-rules.rst
mandates C-like comments (opposed to C source files where
C++ style should be used).

Changes made by using a script provided by Joe Perches here:
https://lkml.org/lkml/2019/2/7/46

.

Suggested-by: Joe Perches <joe@perches.com>
Signed-off-by: Nishad Kamdar <nishadkamdar@gmail.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

kstrtou32_from_user() may return different error codes on certain
circumstances. Respect all possible values.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Fix spelling of MHz unit in a comment.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

There are two problematic places where indentation is not good to a reader
and maintainer. Fix them here.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

There is a function which is solely used for DebugFS interface,
do not build it otherwise.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

kstrtou32_from_user() may return different error codes on certain
circumstances. Respect all possible values.

While here, move it out of the lock: there is no data access that lock guards.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Use devm_platform_ioremap_resource() to simplify the code a bit.

While here, drop initialized but unused ssram_base_addr and ssram_size members.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Add Intel Elkhart Lake to the list of the platforms that driver supports
for the PMC device.

Just like Ice Lake and Tiger Lake, Elkhart Lake can also reuse all the
Cannon Lake PCH IPs. Also, it uses the same PCH IPs of Tiger Lake, no
additional effort is needed to enable but to simply reuse them.

Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Srinivas Pandruvada <srinivas.pandruvada@intel.com>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: David E. Box <david.e.box@intel.com>
Cc: Rajneesh Bhardwaj <irenic.rajneesh@gmail.com>
Cc: Tony Luck <tony.luck@intel.com>
Reviewed-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Gayatri Kammela <gayatri.kammela@intel.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Add Intel Tiger Lake to the list of the platforms that driver supports
for the PMC device.

Just like Ice Lake, Tiger Lake can also reuse all the Cannon Lake PCH
IPs. Since Tiger Lake has almost the same number of PCH IPs as Ice Lake,
reuse Ice Lake's PPFEAR_NUM_ENTRIES instead of defining a new macro.

Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Srinivas Pandruvada <srinivas.pandruvada@intel.com>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: David E. Box <david.e.box@intel.com>
Cc: Rajneesh Bhardwaj <irenic.rajneesh@gmail.com>
Cc: Tony Luck <tony.luck@intel.com>
Reviewed-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Gayatri Kammela <gayatri.kammela@intel.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Check if the platform supports and only then add a debugfs entry for PCH
IP power gating status.

Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Srinivas Pandruvada <srinivas.pandruvada@intel.com>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: David E. Box <david.e.box@intel.com>
Cc: Rajneesh Bhardwaj <irenic.rajneesh@gmail.com>
Cc: Tony Luck <tony.luck@intel.com>
Reviewed-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Gayatri Kammela <gayatri.kammela@intel.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

The current implementation of intel_pmc_core driver allows to reuse, but does
not reflect the exact number and names of IPs for a newer platform which
does not necessarily support all the IPs in the entries. The names and
number of these IPs might differ from its previous platforms. The number
of PCH IPs per platform is calculated based on PPFEAR_NUM_ENTRIES
defined, where each entry represents a bucket (8 bits). The platform can
support 'n' entries, but not necessarily all 'n * 8' IPs.

Create platform dependent bitmap structures to specify the exact number,
names of IPs while reusing the existing IPs.

The changes in this patch are preparatory to accommodate future SoCs
that might reuse the CNL/ICL PCH IPs, and to reflect the exact number of
IPs with its names.

Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Srinivas Pandruvada <srinivas.pandruvada@intel.com>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: David E. Box <david.e.box@intel.com>
Cc: Rajneesh Bhardwaj <irenic.rajneesh@gmail.com>
Cc: Tony Luck <tony.luck@intel.com>
Reviewed-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Gayatri Kammela <gayatri.kammela@intel.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

There is no need to assign NULL or 0 in the static structures explicitly.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

It is common practice to place a comma after the last entry in an
initialized array to avoid the need to add one in future patches that
extend the array. But when the last element is a termination marker, the
comma could be harmful. Any new entries must be added before the
terminator (and the comma would prevent the compiler from issuing a
warning about an entry after the terminator).

Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Srinivas Pandruvada <srinivas.pandruvada@intel.com>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: David E. Box <david.e.box@intel.com>
Cc: Rajneesh Bhardwaj <irenic.rajneesh@gmail.com>
Cc: Tony Luck <tony.luck@intel.com>
Suggested-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Gayatri Kammela <gayatri.kammela@intel.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>