Fix issue where zero-sized BIOS region could cause bitshift
for '-1' which is an unspecified behavior.
Change-Id: Icb62bf413a1a0d293657503ef21fe97b5f9a5484
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/15727
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Since the Integrated Sensor Hub can be disabled through devicetree.cb
as a PCI device, there is no need for a separate register variable.
Remove handling the register and update mainboards' devicetrees. Also
keep ISH disabled on both Reef and Amenia.
Change-Id: I90dbf57b353ae1b80295ecf39877b10ed21de146
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/15710
Tested-by: build bot (Jenkins)
Reviewed-by: Furquan Shaikh <furquan@google.com>
1. The hotplug feature needs to be disabled
so that pcie root ports will be disabled by fsp
2. Correct PcieRootPortEn mapping.
The correct mapping should be like below
PcieRootPortEn[0] ==> 00:14.0
PcieRootPortEn[1] ==> 00:14.1
PcieRootPortEn[2] ==> 00:13.0
PcieRootPortEn[3] ==> 00:13.1
PcieRootPortEn[4] ==> 00:13.2
PcieRootPortEn[5] ==> 00:13.3
BUG=chrome-os-partner:54288
BRANCH=None
TEST=Checked pcie root port is disabled properly
and make sure pcie ports are coalesced.
Also make sure the device will still be enabled after coalescence
when pcie on function 0 is disabled devicetree
Change-Id: I39c482a0c068ddc2cc573499480c3fe6a52dd5eb
Signed-off-by: Kane Chen <kane.chen@intel.com>
Reviewed-on: https://review.coreboot.org/15595
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
The ramstage main() in lib/hardwaremain.c has the logic
to set the ACPI sleep state based on romstage_handoff. Thus,
there's no need to do it a second time.
Change-Id: I75172083587c8d4457c1466edb88d400f7ef2dd0
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/15662
Tested-by: build bot (Jenkins)
Reviewed-by: Furquan Shaikh <furquan@google.com>
The ramstage main() in lib/hardwaremain.c has the logic
to set the ACPI sleep state based on romstage_handoff. Thus,
there's no need to do it a second time.
Change-Id: I88af301024fd6f868f494a737d2cce14d85f8241
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/15661
Tested-by: build bot (Jenkins)
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Lee Leahy <leroy.p.leahy@intel.com>
It is not used in this file.
Change-Id: I59bb41370b97b79073c0fd82b1dbcae9fd8a62d0
Reported-by: GCC 6.1.0
Signed-off-by: Jonathan Neuschäfer <j.neuschaefer@gmx.net>
Reviewed-on: https://review.coreboot.org/15552
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
There is currently a SMBus driver implemented for soc/intel/broadwell
which nearly matches Broadwell-DE as well. Use this driver as template
and add minor modifications to make it work for Broadwell-DE. Support in
romstage is not available and can be added with a different patch.
Change-Id: I64649ceaa298994ee36018f5b2b0f5d49cf7ffd0
Signed-off-by: Werner Zeh <werner.zeh@siemens.com>
Reviewed-on: https://review.coreboot.org/15617
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This patch adds the support for gpio_tier1_sci_en bit which
needs to be set before going to sleep so that when
gpio_tier1_sci_sts bit gets set platform can wake
from S3.
BUG = chrome-os-partner:53992
TEST = Platform wakes from S3 on lidopen,key press.
Tested on Amenia and Reef boards.
Change-Id: I3ba79fa53ca8817149d585fa795a8f427c128dcb
Signed-off-by: Shaunak Saha <shaunak.saha@intel.com>
Reviewed-on: https://review.coreboot.org/15612
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
FSP is currently setting a hard-coded policy for the interrupt
polarity settings. When the mainboard has already set the GPIO
settings up prior to SiliconInit being called that results
in the previous settings being dropped. Work around FSP's
default policy until FSP is fixed.
BUG=chrome-os-partner:54955
Change-Id: Ibbd8c4894d8fbce479aeb73aa775b67df15dae85
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/15649
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
For APIC routed gpios, set the corresponding interrupt polarity
for the associated IRQ based on the gpio pad's invert setting.
This allows for the APIC redirection entries to match the hardware
active polarity once the double inversion takes place to meet
apollolake interrupt triggering constraints.
BUG=chrome-os-partner:54955
Change-Id: I69c395b6f861946d4774a4206cf8f5f721c6f5f4
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/15648
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Furquan Shaikh <furquan@google.com>
The interrupt and timer subsystem (ITSS) sits between the APIC
and the other logic blocks. It only supports positive polarity
events, but there's a polarity inversion setting for each IRQ such
that it can pass the signal on to the APIC according to the
expected APIC redirection entry values. This support is needed
in order for the platform/board to set the expected interrupt
polarity into the APIC for gpio signals.
BUG=chrome-os-partner:54955
Change-Id: I50ea1b7c4a7601e760878af515518cc0e808c0d1
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/15647
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Internally, apollolake routes its interrupts as active high.
This includes SCI, SMI, and ACPI. Therefore, provide helper
macros such that the user can describe an interrupt's active
high/low polarity more easily. It helps for readability when
one is comparing gpio configuration next to APIC configuration
in different files. Additionally, the gpio APIC macros always
use a LEVEL trigger in order to let the APIC handle the
filtering of the IRQ on its own end.
BUG=chrome-os-partner:54977
Change-Id: Id8fdcd98f0920936cd2b1a687fd8fa07bce9a614
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/15644
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Andrey Petrov <andrey.petrov@intel.com>
Change-Id: I195fd3a9c7fc07c35913342d2041e1ffef110466
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/15549
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
This patch adds the handler to enable bit for gpio_tier1_sci_en.
gpio_tier1_sci_en enables the setting of the GPIO_TIER1_SCI_STS
bit to generate a wake event and/or an SCI or SMI#. We are setting
the bit for gpio_tier1_sci_en from the ASL code as OS clears this bit
if set from BIOS. As per ACPI spec _GPE is defined as the Named
Object that evaluates to either an integer or a package. If _GPE
evaluates to an integer, the value is the bit assignment of the SCI
interrupt within the GPEx_STS register of a GPE block described in
the FADT that the embedded controller will trigger. FADT right now
has no mechanism to acheive the same.
Change-Id: I1e1bd3f5c89a5e6bea2d1858569a9d30e6da78fe
Signed-off-by: Shaunak Saha <shaunak.saha@intel.com>
Reviewed-on: https://review.coreboot.org/15578
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Properly obtain the top of memory address from the hardware registers
set by FSP.
TEST=Build and run on Galileo Gen2
Change-Id: I7681d32112408b8358b4dad67f8d69581c7dde2e
Signed-off-by: Lee Leahy <leroy.p.leahy@intel.com>
Reviewed-on: https://review.coreboot.org/15594
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Add host bridge register access routines and macros.
TEST=Build and run on Galileo Gen2
Change-Id: I52eb6a68e99533fbb69c0ae1e6d581e4c4fab9d2
Signed-off-by: Lee Leahy <leroy.p.leahy@intel.com>
Reviewed-on: https://review.coreboot.org/15593
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Specify the memory initialization parameters in
mainboard/intel/galileo/devicetree.cb. Pass these values into FSP to
initialize memory.
TEST=Build and run on Galileo Gen2
Change-Id: I83ee196f5fb825118a3a74b61f73f3728a1a1dc6
Signed-off-by: Lee Leahy <leroy.p.leahy@intel.com>
Reviewed-on: https://review.coreboot.org/15260
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Remove the unused Kconfig values which specify the PDAT file, its
location and inclusion into the coreboot file system. Remove the code
in romstage which locates the pdat.bin file.
TEST=Build and run on Galileo Gen2
Change-Id: I397aa22ada6c073c60485a735d6e2cb42bfd40ab
Signed-off-by: Lee Leahy <leroy.p.leahy@intel.com>
Reviewed-on: https://review.coreboot.org/15205
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Add the gpio_defs.h reference in chip.h to enable
reef and amenia devicetree.cb to use the definitions from gpio_defs.h.
Change-Id: I333d4e810e42309ac76dd90c19f05cf3e3a517b1
Signed-off-by: Sathyanarayana Nujella <sathyanarayana.nujella@intel.com>
Reviewed-on: https://review.coreboot.org/15550
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
There is a second ACPI _DSD document from the UEFI Forum that details
how _DSD style tables can be nested, creating a tree of similarly
formatted tables. This document is linked from acpi_device.h.
In order to support this the device property interface needs to be
more flexible and build up a tree of properties to write all entries
at once instead of writing each entry as it is generated.
In the end this is a more flexible solution that can support drivers
that need child tables like the DA7219 codec, while only requiring
minor changes to the existing drivers that use the device property
interface.
This was tested on reef (apollolake) and chell (skylake) boards to
ensure that there was no change in the generated SSDT AML.
Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://review.coreboot.org/15537
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Instead of having all the mainboards put similar logic
into their own code provide common mechanism for memory
SKU selection. A function, meminit_lpddr4_by_sku(), is
added that selects the proper configuration based on the
SKU id and configuration passed in. LPDDR4 speed as well
as DRAM device density configuration is associated for
each logical channel per SKU id.
BUG=chrome-os-partner:54949
BRANCH=None
TEST=Built and used on reef for memory config.
Change-Id: Ifc6a734040bb61a58bc3d4c128a6420a71245c6c
Signed-off-by: Aaron Durbin <adurbin@chromuim.org>
Reviewed-on: https://review.coreboot.org/15559
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Furquan Shaikh <furquan@google.com>
The internal pulls for gpio_input_pullup() and gpio_input_pulldown()
were using fairly strong pulls. Weaken them so that external pulls
can override the internal ones. This matches the current assumptions
of lib/gpio.c.
BUG=chrome-os-partner:54949
BRANCH=None
TEST=Built and used on reef for memory config.
Change-Id: Ifda1d04d40141325f78db277eb0bd55574994abf
Signed-off-by: Aaron Durbin <adurbin@chromuim.org>
Reviewed-on: https://review.coreboot.org/15558
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
The SCI interrupt can be routed to different IRQs using ACPI control
register. Instead of using hard coded IRQ9 for ACPI table generation
read back the register and return the used IRQ number. This way SCI IRQ
can be modified (e.g. for a given mainboard) and ACPI tables will
remain consistent.
Change-Id: I534fc69eb1df28cd8d733d1ac6b2081d2dcf7511
Signed-off-by: Werner Zeh <werner.zeh@siemens.com>
Reviewed-on: https://review.coreboot.org/15548
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: York Yang <york.yang@intel.com>
There are several different macros available to convert a PCI device and
function to a single 8 bit value. One is PCI_DEVFN and is defined in
device/pci_def.h. The other is DEV_FUNC and is defined in several intel
fsp based chipset implementations. In fsp_broadwell_de DEV_FUNC is even
used without being defined at all. This patch unifies the situation so
that only PCI_DEVFN is used.
Change-Id: Ia1c6d7f3683badc66d15053846936d88aa836632
Signed-off-by: Werner Zeh <werner.zeh@siemens.com>
Reviewed-on: https://review.coreboot.org/15546
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
This patch adds the basic framework for SCI to GPE routing code.
BUG = chrome-os-partner:53438
TEST = Toogle pch_sci_l from ec console using gpioset command and
see that the sci counter increases in /sys/firmware/acpi/interrupt
and also 9 in /proc/interrupts.
Change-Id: I3b3198276530bf6513d94e9bea02ab9751212adf
Signed-off-by: Shaunak Saha <shaunak.saha@intel.com>
Reviewed-on: https://review.coreboot.org/15324
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
On Apollolake CSE can be used to fetch firmware from boot media. However,
when this feature is not used, CSE needs to be explicitly notified of it
before memory training is complete. This way it can transition to next
state.
BUG=chrome-os-partner:53876
TEST=CSE can be power-gated during S0iX. Confirmed with LTB.
Change-Id: I5141bff350b6c0bb662424b7b709f0787ec5fd28
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/15494
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Add the Audio DSP device for apollolake as a PCI driver with a static
scan_bus handler so generic devices can be declared under it.
This is for devices like the Maxim 98357A which is connected on the
I2S bus for data but has no control channel bus and instead just has
a GPIO for channel selection and power down control and needs to
describe that GPIO connection to the OS via ACPI.
Change-Id: Icb97ccf7d6a9034877614d49166bc9e4fe659b12
Signed-off-by: Harsha Priya <harshapriya.n@intel.com>
Reviewed-on: https://review.coreboot.org/15528
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
The P2SB device is device 0xd and function 0. If hidden that
causes the latter pci devices on function >= 1 to not be probed
in the kernel. This is also a problem for coreboot if the P2SB
device is hidden by FSP. That means the coreboot driver won't
be ran. Therefore, provide hide and unhide functions for the
P2SB device.
The other quirk is to allow the GPIO devices to work correctly.
Those devices are ACPI devices. However, their resources are
sub-regions within the P2SB BAR. Sadly, linux doesn't handle
ACPI devices being children of PCI devices. This leads to resource
conflict errors when the P2SB device is visible. For the
time being keep the P2SB device hidden, but also ensure the
resources it is using are accounted for and reserved. The fallout
of that is the PMC and SPI device are no longer probed by the
kernel.
BUG=chrome-os-partner:53017
TEST=Ensured P2SB device is visible and pci resources are allocated
correctly for the devices.
Change-Id: I24e59bbde74310e1ce8425b344a3ad0b88702153
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/15530
Tested-by: build bot (Jenkins)
Reviewed-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Add the function defined in gpio.h to translate a gpio_t into a
value for use in an ACPI GPIO pin table.
For skylake this just returns the gpio_t value as the pins are
translated directly and they are all in the same ACPI device.
Change-Id: I00fad1cafec2f2d63dce9f7779063be0532649c7
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://review.coreboot.org/15520
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
There are four GPIO communities in this SOC and they are implemented
as separate ACPI devices. This means the pin number that is used in
an ACPI GPIO declaration needs to be relative to the community that
the pin resides in. Also select GENERIC_GPIO_LIB in the SOC Kconfig
so this function actually gets used.
This was tested on the reef mainboard by verifying the output of the
SSDT for the Maxim 98357A codec that the assigned GPIO_76 is listed
as pin 0x24 which is the value relative to the Northwest community.
Change-Id: Iad2ab8eccf4c91185a075ffce8d41c81f06c1113
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://review.coreboot.org/15513
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
Support the I2C interfaces on this SOC using the Intel common lpss_i2c
driver. The controllers are supported in pre-ram environments by
setting a temporary base address in bootblock and in ramstage using
the naturally enumerated base address.
The base speed of this controller is 133MHz and the SCL/SDA timing
values that are reported to the OS are calculated using that clock.
This was tested on a google/reef board doing I2C transactions to the
trackpad both in verstage and in ramstage.
Change-Id: I0a9d62cd1007caa95cdf4754f30c30aaff9f78f9
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://review.coreboot.org/15480
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Add support for the soc_acpi_name() handler in the device operations
structure to translate a device path into ACPI name.
In order to make this more complete add some missing devices in
include/soc/pci_devs.h.
Change-Id: I517bc86d8d9fe70bfa0fc4eb3828681887239587
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://review.coreboot.org/15479
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Have the Skylake SOC generate ACPI timing values for the enabled I2C
controllers instead of passing it in the DSDT with static timings.
The timing values are generated from the controller clock speed and
are more accurate than the hardcoded values that were in the ASL which
were originally copied from Broadwell where the controller is running
at a different clock speed...
Additionally it is now possible for a board to override the values
using devicetree.cb. If zero is passed in for SCL HCNT or LCNT then
the kernel will generate its own timing using the same forumla, but if
the SDA hold time value is zero the kernel will NOT generate a correct
value and the SDA hold time may be incorrect.
This was tested on the Chell platform to ensure all the I2C devices on
the board are still operational with these new timing values.
Change-Id: I4feb3df9e083592792f8fadd7105e081a984a906
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://review.coreboot.org/15291
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
The nvs.h header is the one which defines global_nvs_t proper.
Don't rely on an indirect inclusion.
Change-Id: I89d6a73f65e408c73f068b4a35b5efd361a6e5d3
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/15503
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins)
Every other platform has global_nvs_t as a typedef. For some
reason apollolake didn't bother following current conventions.
Fix this omission to allow for better code sharing and consistency.
Change-Id: Id596eed517737759a64ce803c89ea2a05cbe2cce
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/15502
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins)
This patch exposes a function to read pmc bar.
PMC bar is read in function read_pmc_mmio_bar which
is defined static in file pmutil.c. This patch exposes
that functionality to call it from other files.
BUG=chrome-os-partner:53438
TEST= Read the PMC bar value properly from outside
pmutil file.
Change-Id: I26ee13e6ab95d3a8991c7f8ea4b3856ceb015d10
Signed-off-by: Shaunak Saha <shaunak.saha@intel.com>
Reviewed-on: https://review.coreboot.org/15460
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
