All boards with DRIVERS_GENERIC_IOAPIC select it.
Presumably the related configuration of routing IRQ0 when
IOAPIC is enabled should be always done to provide i8259
legacy compatibility for payloads.
Change-Id: Ie87816271fa63bba892c8615aa5e72ee68f6ba93
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55287
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Use the common ACPI code to reduce code duplication.
After this change, `PSS_MAX_ENTRIES` is honored correctly in P-state
table generation (as of commit c2540a9) and the number reduces from 10
to 7 entries.
Also, remnants of P_BLK support missed in CB:58096 will vanish.
Tested on google/fizz: no errors in dmesg, ACPI tables remain the same
(except PSS, as mentioned above).
Change-Id: I1ec804ae4006a2d9b69c0d93a658eb3b84d60b40
Tested-by: Matt DeVillier <matt.devillier@gmail.com>
Signed-off-by: Michael Niewöhner <foss@mniewoehner.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44138
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Matt DeVillier <matt.devillier@gmail.com>
Apollo Lake and Gemini Lake do not have a hardware PM ACPI timer but
only uCode PM Timer emulation. Add a Kconfig `NO_PM_ACPI_TIMER` denoting
SoCs without PM Timer and make it mutually exclusive with the Kconfig
`USE_PM_ACPI_TIMER`.
This is partly redundant to `PM_ACPI_TIMER_OPTIONAL`, which will be
dropped in the follow-up change, though.
Change-Id: Ic323bbfb7089c53a6f22724910a0ff3df8904ebd
Signed-off-by: Michael Niewöhner <foss@mniewoehner.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57931
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
uCode PM Timer emulation is only needed when the hardware PM ACPI timer
is disabled. Also, since it redirects any register accesses to uCode,
it overrides the hardware PM Timer. Thus, only enable emulation
when required.
Change-Id: I60a775bd6eb4206750f606ce8a8777d2e2dfb579
Signed-off-by: Michael Niewöhner <foss@mniewoehner.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57930
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Set `EnableTcoTimer=1` in order to keep FSP from
1) enabling ACPI Timer emulation in uCode.
2) disabling the PM ACPI Timer.
Both actions are now done in coreboot.
`EnableTcoTimer=1` makes FSP skip these steps in any possible case
including `SkipMpInit=0`, `SkipMpInit=1`, use of the MP PPI or FSP
Multiphase Init. This way full control is left to coreboot.
Change-Id: I8005daed732c031980ccc379375ff5b09df8dac1
Signed-off-by: Michael Niewöhner <foss@mniewoehner.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57933
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Reviewed-by: Lance Zhao
Disable the PM ACPI timer during PMC init, when `USE_PM_ACPI_TIMER` is
disabled. This is done to bring SKL, CNL, DNV in line with the other
platforms, in order to transition handling of the PM timer from FSP to
coreboot in the follow-up changes.
For SKL and CNL, this temporarly redundantly disables the PM Timer,
since FSP does that, too. This redundancy is resolved in the follow-up.
Change-Id: I47280cd670a96c8fa5af107986496234f04e1f77
Signed-off-by: Michael Niewöhner <foss@mniewoehner.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57932
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Move disabling of PM Timer to SoC PMC code.
The original reason for placing that in `finalize` [1] was FSP hanging
due to use of the PM timer without enabling timer emulation first in
coreboot, which was added later [2].
[1] commit 6c1bf27dae (intel/skylake: disable ACPI PM Timer to enable
XTAL OSC shutdown)
[2] commit f004f66ca7 (soc/intel/skylake: Enable ACPI PM timer emulation
on all CPUs)
Change-Id: I354c3aea0c8c1f8ff3d698e0636932b7b76125f7
Signed-off-by: Michael Niewöhner <foss@mniewoehner.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58019
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Currently, only the PM1_STS mask gets passed to `acpi_fill_soc_wake`. To
be able to override the GPE0_STS mask as well, also pass that one. To
accomplish that, pointers to the variables are passed now.
Change-Id: If9f28cf054ae8b602c0587e4dd4a13a4aba810c7
Signed-off-by: Michael Niewöhner <foss@mniewoehner.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58071
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
`RTC_EN` is in the RTC well* so we can rely on the actual register
content instead of statically overriding it. Drop it from the static
wake bits mask.
* Tested on clevo/l140cu
Change-Id: Ia0ae71f0a472513233bc0fd5625faf15bf86beaf
Signed-off-by: Michael Niewöhner <foss@mniewoehner.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58211
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
The PM1_EN bits WAK_STS, RTC_EN, PWRBTN_EN don't need any SoC-specific
handling. Deduplicate `acpi_fill_soc_wake` by setting these bits in
common code.
Change-Id: I06628aeb5b82b30142a383b87c82a1e22a073ef5
Signed-off-by: Michael Niewöhner <foss@mniewoehner.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58043
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
The `IRQ_SLOT_COUNT` value is only meaningful when generating a PIRQ
table. None of these boards do it, so specifying this value achieves
absolutely nothing. Drop it to prevent further useless copy-pasting.
Change-Id: I2d63b850c03fc1471c0eef180e8b621311b2c336
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58362
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Reviewed-by: Michael Niewöhner <foss@mniewoehner.de>
Picasso and Cezanne define and use APU_I2C[01234]_BASE for the base
addresses of the I2C controllers, so align Stoneyridge with this. The
ACPI device names aren't changed from I2C[ABCD] to I2C[0123] for now
since this might change behavior in the OS and would also change the
resulting binary of a timeless build.
TEST=Timeless build results in identical image for Google/Treeya.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I9c400c073eba5c14bd35703b717f75df89a8719d
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58370
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Since ACPI 5.0A it is allowed to disable the ACPI Timer, when the
according FADT flag `ACPI_FADT_PLATFORM_CLOCK` is unset.
Starting with Skylake, most platforms (except Xeon-SP) support PM Timer
emulation, so even legacy OSes and payloads should work fine with the
hardware PM Timer disabled. However, when the `TMR_STS` functionality
is required, some legacy OSes might still not work (properly).
Add a note about this to the Kconfig help.
Change-Id: I53f1814113902124779ed85da030374439570688
Signed-off-by: Michael Niewöhner <foss@mniewoehner.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58018
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Reviewed-by: Lance Zhao
The FADT contains a flag `ACPI_FADT_PLATFORM_CLOCK` telling the OSPM if
a specification-compliant PM Timer is present. Currently, this flag is
set regardless of the timer being enabled or disabled.
To be specification-compliant, only set that flag, when the hardware PM
Timer is enabled. This changes behaviour of all mainboards defaulting to
USE_PM_ACPI_TIMER=n.
Note: On platforms supporting uCode PM ACPI Timer emulation, this is
required, too, because emulation does not support `TMR_STS`. Any
OS or software checking this flag and thus relying on the overflow
flag would not work (properly).
Change-Id: Id2e5d69b5515c21e6ce922dab2cb88b494c65ebe
Signed-off-by: Michael Niewöhner <foss@mniewoehner.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57934
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
The writes were originally added due to being part of the initialization
sequence in the reference code, but coreboot already has those registers
cleared by the time we reach this part of the code, so we can drop these
redundant writes.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I43344460e5355664841d77daf1df3fd386e047e9
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58341
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
If a system doesn't use eSPI or has the eSPI interface already
configured in verstage on PSP, not calling configure_espi_with_mb_hook
from fch_pre_init makes it a bit more obvious that the eSPI interface
initialization will be skipped.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ia77b83d56a5dab1bac6cfbbd92d33aa60a9e8b89
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58339
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Now that the I2C[ABCD]_BASE_ADDRESS defines aren't macros that calculate
the MMIO addresses any more, those defines can also be used in the ACPI
code.
TEST=Timeless build results in identical image for Google/Treeya.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I7de2f83dc2f8061d8f1735caf10314bcddb2d3fa
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58337
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
The I2C_BUS_ADDRESS(x) macro isn't used to iterate over the I2C
controller base addresses, so drop this and use the fixed MMIO address
for the I2C[ABCD]_BASE_ADDRESS defines instead which also allows using
those defines in the ACPI code.
TEST=Timeless build results in identical image for Google/Treeya.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Idd7484a0322dc5167cbb7fdcd9a2583f0dbed50e
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58336
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
I2C_BASE_ADDRESS is the beginning of the MMIO space that contains the
I2C controllers MMIO. I2C[ABCD]_BASE_ADDRESS are the base addresses of
the 4 I2C controllers, so use I2CA_BASE_ADDRESS instead here.
TEST=Timeless build results in identical image for Google/Treeya.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ie8d6a438f76cd33929f5070f9ec6b2f280f471a0
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58335
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Before this patch the reservation of the MMIO region of the I2C
controllers was done in the LPC controller PCI device despite the I2C
controllers already being devices in the devicetree. This patch
implements this functionality as read_resources function of the I2C
device instead. This will only reserve the memory when the I2C devices
are enabled in devicetree which is a change from the previous behavior.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I67c853df3be2f593ecfa113ae2f74e5df7cf74e0
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58307
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This change splits CSE metadata structure (added to CBFS) into two
separate CBFS files (me_rw.hash and me_rw.version). Since `struct
cse_rw_metadata` is now used, it is dropped completely.
This change is being made in order to prepare for the upcoming changes
to stitch CSE binary at build time. Since the binary might not be
available pre-built, it complicates the order of operations for the
addition of CSE metadata structure and declaring hash and version as
CPPFLAGS_common. Instead rules can be enabled for individual CBFS file
targets for hash and version that ensure proper ordering as well.
BUG=b:184892226
TEST=Ensured that update works correctly on brya by forcing version
mismatch. In case of version match, no update is triggered.
Signed-off-by: Furquan Shaikh <furquan@google.com>
Change-Id: I7c9bb165e6a64415affcd0b3331628092195fa0d
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58158
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
The SHRM region needs to be 4 byte aligned, which make enabling
compression slightly more complicated. We need to map it to cached
memory before loading it and flushing to memory (in aligned chunks)
then remapping the address space back to device memory before
beginning execution of the SHRM region.
Also, did some cleanup in this file based on comments in CB:49392.
BUG=b:182963902
BRANCH=None
TEST=Make sure we can still boot to kernel on herobrine
Change-Id: Iaad8a8a02abe40bd01766d94ef0b61aac7671936
Signed-off-by: Shelley Chen <shchen@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58191
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
Configure I2C high / low time in the device tree to ensure Touchpad
I2C CLK runs accurately between 380 kHz and 400 kHz.
Measured I2C frequency just as below after tuning:
Touch Pad CLK: 389.2 KHz
BUG=b:202787528
TEST=Build and check after tuning I2C clock is between 380 kHz and 400 kHz
Change-Id: I0f9d062fc611de0062a39849aee1174268391682
Signed-off-by: Ian Feng <ian_feng@compal.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58238
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Henry Sun <henrysun@google.com>
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
If the VGA BIOS file path for `VGA_BIOS_FILE` in a mainboard's Kconfig
does not exist in the coreboot tree (including submodules), drop it.
These files should be stored in the `site-local` subdirectory and the
paths specified for each board in `site-local/Kconfig`. For example:
config VGA_BIOS_FILE
default "site-local/x200_vbios.bin" if BOARD_LENOVO_X200
Note that this is just an example. There are better ways to structure
one's `site-local` subfolder. Using the `CONFIG_MAINBOARD_DIR` option
would be one of them, though variants may still need special handling.
Also, update autoport to not generate `VGA_BIOS_FILE` defaults.
Change-Id: I1b5dfba035a42d7943f270f95fb7d32b285584d2
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/51340
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Reviewed-by: Michael Niewöhner <foss@mniewoehner.de>
Reviewed-by: Felix Singer <felixsinger@posteo.net>
