Add GPIO initialization for mc_bdx1 mainboard. Call init_gpios() as
early as possible as FSP will set up things (like hiding PCI devices)
rather early. If connections on the mainboard are dependent on GPIO
settings then FSP can screw things up (e.g. disabling not yet connected
PCI root ports).
Change-Id: I003277cfb871f861900b7fcdc5ec851d4c1c1e6a
Signed-off-by: Werner Zeh <werner.zeh@siemens.com>
Reviewed-on: https://review.coreboot.org/22035
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Add functionality to initialize, set and read back GPIOs on FSP based
Broadwell-DE implementation.
Change-Id: Ibbd86e2142bbf5772eb4a91ebb9166c31d52476e
Signed-off-by: Werner Zeh <werner.zeh@siemens.com>
Reviewed-on: https://review.coreboot.org/22034
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
In amd64_smm_state_save_area_t break out fields in reserved4 to allow access.
BUG=b:65485690
Change-Id: I592fbf18c166dc1890010dde29f76900a6849016
Signed-off-by: John E. Kabat Jr <john.kabat@scarletltd.com>
Reviewed-on: https://review.coreboot.org/22092
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
XMP profiles can have a restriction on max supported DIMMs per channel,
but many configurations work with more DIMMs.
This is relevant on mainboards with 2 DIMM slots per channel (usually 4
in total). Populating both slots with DIMMs that support XMP profiles only
with 1 DIMM per channel turns off said XMP profiles.
TEST=On a system with two DIMM slots per channel populate both slots on
one channel and ensure that DIMMs run with XMP profiles enabled.
Change-Id: I1f22d981afcef0ee73785823b0a943cf3d3564e3
Signed-off-by: Vagiz Trakhanov <rakkin@autistici.org>
Reviewed-on: https://review.coreboot.org/21841
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
These were OEM configurations hidden inside a header file, notation
was already dropped for f15tb and f16kb.
Change-Id: Id64fa861fd516e9f7cae9eba9b8145e033fe9bdd
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/21650
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Marc Jones <marc@marcjonesconsulting.com>
This change adds timestamps to romstage in order to keep
PC Engines apu2 platform in active codebase.
Change-Id: Ie0286d4982623da9d035c47df6077edaf51e5110
Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Reviewed-on: https://review.coreboot.org/22071
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Collecting time stamps is useful, especially for board status uploads,
and doesn’t come with any downsides. So enable it by default on as many
boards as possible.
The boards below currently fail to build properly, so only enable it by
default on x86.
1. board.CUBIETECH_CUBIEBOARD
2. board.EMULATION_QEMU_POWER8
3. board.EMULATION_QEMU_UCB_RISCV
4. board.EMULATION_SPIKE_UCB_RISCV
5. board.LOWRISC_NEXYS4DDR
6. board.TI_BEAGLEBONE
Change-Id: Ib01176fc2a4dffe37827c136bb8214083ce61180
Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-on: https://review.coreboot.org/11864
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Earlier the EC expected the host to set appropriate masks before
reading host events. However, with recent change in EC, this is no
longer required. This change removes the setting of wake_mask before
and after reading the host events. However, in order to support older
versions of EC, a new feature flag is added on the EC side that
informs the host whether or not it is using the new way of reporting
host events without having to set wake mask.
CQ-DEPEND=CL:719578
TEST=Verified that EC wake events are correctly logged with both old
and new versions of EC.
Change-Id: Ib17e1296fb7d3bbc84fc7581fd0a9bd179ac87b9
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/22006
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
If GPE_STS indicates that the wake source is internal PME, but none of
the controllers have the PME_STS bit set, then try probing individual
XHCI ports to see if one of those was a wake source. In some cases
e.g. gsmi logging with S0ix, pci_pm_resume_noirq runs before gsmi
callback and clears PME_STS_BIT in controller register. In such cases,
xhci port status might provide a better idea about the wake source.
BUG=b:67874513
Change-Id: I841bb2abccfa9bd6553c1513e88a6306b40315e4
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/22089
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
1. Ensure that port_status read is not all 1s to ensure that read from
mmio address returned valid data.
2. If device connect/disconnect shows that it was a wake source, there
is no need to check for usb activity.
BUG=b:67874513
Change-Id: Id8b4a1fec7bfe530fe435a0f52944b273cdd89ad
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/22088
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This change adds a mask to allow SMI handlers to be run even in SCI
mode. This prevents any SMI handlers from accidentally taking
unnecessary action in SCI mode.
Add APM_STS and SMI_ON_SLP_EN_STS to this mask to allow gsmi and sleep
to work in SCI mode.
BUG=b:67874513
Change-Id: I298f8f6ce28c9746cbc1fb6fc96035b98a17a9e3
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/22087
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This change adds support for logging wake source information in gsmi
callbacks. With this change, all the elog logging infrastructure can
be used for S0ix as well as S3 on skylake.
BUG=b:67874513
Change-Id: Ie1f81e956fe0bbe2e5e4c706f27997b7bd30d5e0
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/22086
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This change moves soc_fill_power_state and soc_prev_sleep_state to
pmutil.c. It allows the functions to be used across romstage and smm.
BUG=b:67874513
Change-Id: I375ac029520c2cdd52926f3ab3c2d5559936dd8c
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/22085
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This change allows the same functions to be used across ramstage and
smm without having to add checks for what stage is using it.
BUG=b:67874513
Change-Id: I3b10c9e8975e8622d8cb0f66d90d39a914ba7e1c
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/22084
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This change creates a new function pmc_fill_pm_reg_info that fills
chipset_power_state structure with all the PM register
information. On the other hand, already existing pmc_fill_power_state
calls into pmc_fill_pm_reg_info and then checks and returns previous
sleep state information. This allows caller to get all the PM register
information when previous sleep state is not relevant.
BUG=b:67874513
Change-Id: Idc91e4aef5379549355aceb685f7afafa6a220c5
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/22080
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
1. Add support for new GSMI commands to log S0ix entry/exit
information in elog.
2. In case of resume, provide callbacks to allow platform and
mainboard to log any wake source information.
BUG=b:67874513
Change-Id: I593e8a9e31cad720ac1f77aab447a0dbdbe9a28b
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/22079
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
In commit c06a3f72 (arch/x86: initialize EBDA in S3 and S0/S5 path)
the BDA and EBDA are wiped in the resume path. However, the coreboot
table forwarding entry wasn't taken into account so that was wiped
which resulted in cbmem not working on the resume path. Fix this
by stashing the forwarding table in cbmem and restoring it on
the resume path.
Change-Id: I142503535a78635fbb1c698fc7d032c1a2921813
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/22078
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Add option for user to select what kind of probe can be used for
platform debug.
TEST=Set to XDP and boot up system with XDP hooked, able to halt.
Change-Id: Ib6add93e3f1c8a646aa625a4cea9be0acecc0487
Signed-off-by: Lijian Zhao <lijian.zhao@intel.com>
Reviewed-on: https://review.coreboot.org/21942
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Pratikkumar V Prajapati <pratikkumar.v.prajapati@intel.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Before OS boot up, the following actions need to be taken.
1. Lock down PMC/SPI/DMI/TCO register.
2. Disable Sideband Access.
3. Disable Heci interface.
4. Disable PMtimer base on config settings.
TEST=Boot up into OS properly on both cannonlake y and cannonlake u rvp
board.
Change-Id: Icfa05b50fd76fbaeb856d398918990aedac4c5e6
Signed-off-by: Lijian Zhao <lijian.zhao@intel.com>
Reviewed-on: https://review.coreboot.org/21943
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This patch skips SPD data reading when system resumes from S3 since
MRC cahce is adopted and validated in fsp_memory_init.
BUG=b:67021596
TEST=Run suspend/resume on Fizz and make sure the systems are
working well when system resumes from S3. Checked dmidecode
information and SMBIOS type 17 data is the same with cold
boot.
Change-Id: I1692fca8456290d1471973b746537b5fec504e03
Signed-off-by: Gaggery Tsai <gaggery.tsai@intel.com>
Reviewed-on: https://review.coreboot.org/21987
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This patch implements soc override function to calculate reserve memory
size (PRMRR, ME stolen, PTT etc). System memory should reserve those
memory ranges.
BRANCH=none
BUG=b:63974384
TEST=Ensures DRAM based resource allocation has taken care
of intel soc reserved ranges.
Change-Id: I3052a255c4496dc81c8dfc6882d3ad504abae9c6
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/21986
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Lijian Zhao <lijian.zhao@intel.com>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This patch uses BIOS EBDA area to store relevent details
like cbmem top during romstage after MRC init is done.
Also provide provision to use the same EBDA data across
various stages without reexecuting memory map algorithm.
BRANCH=none
BUG=b:63974384
TEST=Ensures HW based memmap algorithm is executing once in romstage
and store required data into EBDA for other stage to avoid redundant
calculation and get cbmem_top start from EBDA area.
Change-Id: I763ad8181396ea8d8c0d5cf088264791ba62dceb
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/21985
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This patch uses PMC common function to get previous sleep state
using cbmem or chipset_power_state global structure.
acpi_get_sleep_type is needed in PRE_RAM stage when soc selects
CONFIG_EARLY_EBDA_INIT kconfig option.
Change-Id: Ib9f8bdc1c682807450b6c01941b9a0927789b2d8
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/21988
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
EC maintains a FIFO of all MKBP events and sets host event whenever
a new entry is added to the FIFO. Clearing only the host event for
MKBP creates an inconsistent state where there are pending MKBP events
in the FIFO but host event for MKBP is cleared. In order to maintain a
consistent view, host should clear all MKBP events in the FIFO if host
event is being cleared.
This change drains out all the MKBP events in the FIFO when
clear_pending_events is called.
TEST=Verified by adding debug logs in EC to verify that all the MKBP
events that occur before clear_pending_events is called get cleared
from the FIFO.
Change-Id: I131722dc01608dff30230fe341e6b23ae4cc409e
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/22005
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
This change adds a new helper function google_chromeec_get_mkbp_event
that allows coreboot to query EC for the next available MKBP event.
Change-Id: Ia6d64586ca62378d08025c96c2689c00c816041f
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/22007
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Instead of disabling all GPEs during PMC init in bootblock, this
change moves it to pmc_fill_power_state which allows romstage to
correctly fill up GPE_EN registers in chipset_power_state. This is
essential for correctly identifying the wake source.
Disabling all GPEs was added recently in change 74145f76
(intel/common/pmc: Disable all GPEs during pmc_init) because keeping
GPEs enabled in coreboot while enabling SMI could lead to
side-effects as explained in the change. Moving pmc_disable_all_gpe to
pmc_fill_power_state should be safe as that happens before SMI is
enabled in coreboot.
TEST=Verified that GPE-based wake source is correctly
identified. Also, no issues observed while resuming from S3.
Change-Id: I8e992ad09ffdefba62de11fa572e783715776bf1
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/22033
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
bc is one of the standard posix utilities. I'm surprised that it's
not in the debian docker image by default.
Change-Id: I02f2d5296e7f87876b236af119965d1f4e6a0bc0
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/21889
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Jonathan Neuschäfer <j.neuschaefer@gmx.net>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
This is already selected in `src/cpu/amd/agesa/Kconfig`.
Change-Id: I691a2ade10ee461b6bc34ea24d57a911281791f3
Signed-off-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-on: https://review.coreboot.org/22011
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
DRAM Controller is always 00.0. No need to check its PCI ID.
Change-Id: I9c5f3e5658905e464491579f8da01aa6a03bd3b7
Signed-off-by: Vagiz Tarkhanov <rakkin@autistici.org>
Reviewed-on: https://review.coreboot.org/21754
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
It's more consistent to re-initialize EBDA in all boot paths.
That way, the data living in EBDA is cleared prior to be
accessed (assuming it's after setup_ebda()).
Change-Id: I05ff84f869f7b6a463e52b4cb954acc5566475cd
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/21997
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Subrata Banik <subrata.banik@intel.com>
It's possible for chipsets utilizing ebda to cache the cbmem_top()
value to be called prior to the object being entirely setup. As
such it's important to check the signature to ensure the object
has been initialized. Do that in a newly introduced function,
retrieve_ebda_object(), which will zero out the object if the
signature doesn't match.
Change-Id: I66b07c36f46ed9112bc39b44914c860ba68677ae
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/21995
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-by: Matt DeVillier <matt.devillier@gmail.com>
Add hexdump.c to Makefile.inc and change an elog_debug format to
use %z for size_t arguments. This corrects build errors when ELOG_DEBUG
is used.
Change-Id: I3d5547eed8ada7c4bdcbbb8bb9d6965ade73beda
Signed-off-by: John E. Kabat Jr <john.kabat@scarletltd.com>
Reviewed-on: https://review.coreboot.org/21769
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Werner Zeh