According to changes in the soc/xeon_sp code [1,2], server motherboards
with Lewisburg PCH can use the soc/intel/common/gpio driver to configure
GPIO controller. This patch adds pads configuration map, which has the
format required by the GPIO driver. The data for this was taken from the
inteltool register dump with vendors firmware. The gpio.h file with pad
configuration was generated automatically using the util/intelp2m [3]:
./intelp2m -raw -p lbg -file cedarisland/vendorbios/inteltool_gpio.log
[1] https: //review.coreboot.org/c/coreboot/+/39425
[2] https: //review.coreboot.org/c/coreboot/+/39428
[3] https: //review.coreboot.org/c/coreboot/+/35643
Change-Id: I90b91e6dbf8c65c747d0e0d94c61023e610f93ab
Signed-off-by: Maxim Polyakov <max.senia.poliak@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40734
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Andrey Petrov <andrey.petrov@gmail.com>
-n needs to check against a string, but if CONFIG_TIANOCORE_BOOTSPLASH_FILE
is unset, then $(CONFIG_TIANOCORE_BOOTSPLASH_FILE) evaluates to nothing
and the check fails, leading the Makefile to try and copy a non-
existant file/path.
Change-Id: Iff717dd48748cff16f485bafaa91c7a225fb5bdb
Signed-off-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40860
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Most payloads, like GRUB, SeaBIOS and Linux, are able to initialize the
PS/2 keyboard themselves, so coreboot does not need to initialize it.
Therefore, this option should not be hard-coded for the mainboard, and
be left for the user to select.
Change-Id: Ibfb7efa22c525e60399afc198af6632330faaac3
Signed-off-by: Paul Menzel <pmenzel@molgen.mpg.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40752
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Since commit 6aa8c5bc (drivers/pc80: Do not initialize PS2 keyboard by
default), the Kconfig option `DRIVERS_PS2_KEYBOARD` already defaults to
`n`. So, remove it here, as this option should be user selectable
anyway depending on the payload.
Change-Id: I3d08fb6bbb3e9d53fd2fed96f26679e8b1e73f8c
Signed-off-by: Paul Menzel <pmenzel@molgen.mpg.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40751
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Most payloads, like GRUB, SeaBIOS and Linux, are able to initialize the
PS/2 keyboard themselves, so coreboot does not need to initialize it.
Therefore, this option should not be hard-coded for the mainboard, and
be left for the user to select.
Change-Id: Iac835d2e7a3232f8e5c76f10984ae3f172f9c0ca
Signed-off-by: Paul Menzel <pmenzel@molgen.mpg.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40750
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
This change enables support for em100 for Picasso platform. Since
em100 requires lower SPI speed, this change configures speed in all
modes as 16MHz.
BUG=b:147758054,b:153675510
BRANCH=trembyle-bringup
TEST=Verified that em100 works fine on trembyle.
Signed-off-by: Furquan Shaikh <furquan@google.com>
Change-Id: Ib5ea1fe094fda9b8dba63e94b37e61791629564f
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40825
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This change adds options to allow mainboard to configure SPI speed for
different modes as well as the SPI read mode.
BUG=b:153675510,b:147758054
BRANCH=trembyle-bringup
TEST=Verified that SPI settings are configured correctly for trembyle.
Change-Id: I24c27ec39101c7c07bedc27056f690cf2cc54951
Signed-off-by: Furquan Shaikh <furquan@google.com>
Signed-off-by: Rob Barnes <robbarnes@google.com>
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40421
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This change adds a helper sb_spi_init() that makes all the required
calls for configuring SPI to ROM.
BUG=b:147758054,b:153675510
BRANCH=trembyle-bringup
TEST=Verified that SPI configuration is correct for trembyle.
Signed-off-by: Furquan Shaikh <furquan@google.com>
Change-Id: Ic5b395a8d3bdab449c24b05d1b6b8777e128b5e0
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40824
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This change adds enums for spi_read_mode and spi100_speed in
preparation for adding these to chip.h in follow-up CLs. This makes it
easier to reference what the mainboard is expected to set for these
SPI configs.
BUG=b:147758054,b:153675510
BRANCH=trembyle-bringup
TEST=Verified that SPI configuration is correct for trembyle.
Signed-off-by: Furquan Shaikh <furquan@google.com>
Change-Id: I7f9778b41bd059a50f20993415ebd8702a1ad58e
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40823
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
southbridge.h does not really need chip.h. So, this change removes the
inclusion of chip.h from it.
Signed-off-by: Furquan Shaikh <furquan@google.com>
Change-Id: I09c87b975ecd5f7798da8dd858be0c729aef42de
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40822
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This reverts commit 285975dbba.
Reason for revert: VB2_RECOVERY_TRAIN_AND_REBOOT was never meant to have
any special effect on memory training behavior. It was just supposed to
be a "reboot automatically after reaching kernel verification" recovery
reason. On x86 devices this was used to prime the separate recovery
MRC cache in the factory (make sure it is initialized before shipping).
This isn't used on Kukui anyway, but in order to make sure nobody copies
this code and keep the behavior consistent between platforms, let's
remove it.
Change-Id: I5df5e00526e90cb573131de3c8bac9f85f4e3a5f
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40623
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Enable aspeed's function that port 80h direct to GPIO for LED display,
refer to section 9.4 Port 80h Direct to GPIO Guide of aspeed's
Application Design Guide, also configure GPIO to UART for output
serial console messages.
Tested=Check if port 80h LED debug card can display POST codes at
early stage, and serial console can see the related messages.
Change-Id: I087d5a81b881533b4550c193e4e9720a134fb8e7
Signed-off-by: BryantOu <Bryant.Ou.Q@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40481
Reviewed-by: Patrick Rudolph <patrick.rudolph@9elements.com>
Reviewed-by: Andrey Petrov <andrey.petrov@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
BUG=b:152893285
BRANCH=none
TEST=Build and boot to OS volteer with Intel SATA and reboot
from OS console
Signed-off-by: Wonkyu Kim <wonkyu.kim@intel.com>
Change-Id: Ibed8f8c445bf2ac2290ffb670d8dfb83fc960438
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40584
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Create the trondo variant of the volteer reference board by copying the
template files to a new directory named for the variant.
BUG=b:154678884
BRANCH=None
TEST=util/abuild/abuild -p none -t google/volteer -x -a
make sure the build includes GOOGLE_TRONDO
Change-Id: Ie4f9bfe4798e14f91c6cb439f5c5ab2b9ea52b51
Signed-off-by: David Wu <david_wu@quanta.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40686
Reviewed-by: Caveh Jalali <caveh@chromium.org>
Reviewed-by: Paul Fagerburg <pfagerburg@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Based on USB SI report to fine tune the strength for USB2 port0.
BRANCH=none
BUG=b:153590143
TEST=build and test USB2 port0 function works fine.
Change-Id: I070c9e1c8153a680fb8f827889738a764d7ea9f4
Signed-off-by: Tim Chen <tim-chen@quanta.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40482
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Kangheui Won <khwon@chromium.org>
Reviewed-by: Edward O'Callaghan <quasisec@chromium.org>
memcpy() is meant to be used on normal memory and often implemented with
architecture-specific optimizations to make that as performant as
possible. MMIO registers often have special access restrictions that may
be incompatible with whatever memcpy() does. For example, on arm64 it
uses the LDP (load pair) to load 16 bytes at a time, which makes 4-byte
MMIO registers unhappy.
This patch removes the caching of the XHCI capreg registers and changes
it back to a pointer. The CAP_GET() macro is still accessing a full
(non-bitfield) uint32_t at the end so this should still generate a
4-byte access (which was the goal of the original change in CB:39838).
Signed-off-by: Julius Werner <jwerner@chromium.org>
Change-Id: Id058c8813087a8e8cb85f570399e07fb8a597108
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40895
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
It seems that all SC7180 boards we have can well handle 37.5MHz of SPI
flash speed, so bump that up from the current 25MHz so that we don't
leave boot speed on the table. (The next step would be 50MHz which
currently doesn't work on all boards so we're not going there yet.)
BUG=b:117440651
Signed-off-by: Julius Werner <jwerner@chromium.org>
Change-Id: Id6e98fcbc89f5f3bfa408c7e8bbc90b4c92ceeea
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40874
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Philip Chen <philipchen@google.com>
Check SPD is not NULL before print. This can prevent the system
from hanging up.
BUG=b:154445630
TEST=Check NULL SPD is not print.
Signed-off-by: Eric Lai <ericr_lai@compal.corp-partner.google.com>
Change-Id: Iccd9fce99eda7ae2b8fb1b4f3c2e635c2a428f04
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40560
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Looks like 5 is a valid system type, as Google Beltino and Slippy are
using it. According to comments on these mainboards' code, this value
corresponds to ULT systems. So, add it to the comment on the pei_data
struct, which was likely copied from Sandy Bridge and was not updated.
Change-Id: I3654bb6022839dba3e1499cf43e8beaa97d1def1
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40692
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Matt DeVillier <matt.devillier@gmail.com>
This change adds a Kconfig option PSP_UNLOCK_SECURE_DEBUG which
when enabled includes secure debug unlock blobs and sets the required
softfuses and options for amdfwtool. By default this is set to 'N'.
BUG=b:154880818
Signed-off-by: Furquan Shaikh <furquan@google.com>
Change-Id: I47d8af67989b06242d662c77b7d9db97f624edd5
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40683
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Add LPC read/write functions for access aspeed's memory, also
create config data table to config memory and SIO. These
functions are used at early stages to configure devices.
Signed-off-by: Bryant Ou <Bryant.Ou.Q@gmail.com>
Change-Id: Ib59c29a042b2c7bf196b8a5bd5218704d8075855
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40483
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
It replicates the functionality of savedefconfig because back when the
script was added, savedefconfig didn't work for us. It now does, is
the official way of doing things, is recommended in our documentation
and is also a fair bit faster.
Change-Id: Ia8e0377537ff7cd638c564037ea6a77b01a87243
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37150
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
This change updates Makefile.inc to use a helper function set-bit to
set a bit for the soft fuses. It gets rid of the different checks that
were done to set soft fuses to magic values in different places.
This is still not the best way to handle the fuses and instead this
logic should be embedded within the amdfwtool by making it aware of
specific platforms. But until that happens, we want to avoid having to
add PSP_SOFTFUSE setting in various places with different values.
BUG=b:154880818
TEST=Verified that the softfuse values are same with and without this
change.
Change-Id: I73887eb9c56ca5bb1c08d298fa818d698da1080b
Signed-off-by: Furquan Shaikh <furquan@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40700
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Some of the PSP Kconfig options that are prompted to the user should
really be selected by mainboard. This change updates such options to
not make them user-visible any more.
BUG=b:154880818
Change-Id: Iaff02fb1e720e0562b740799593322e59b022212
Signed-off-by: Furquan Shaikh <furquan@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40699
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
PSPTRUSTLETS_FILE was including a binary for fTPM which according to
BIOS architecture design guide is the firmware enabled TPM. Chrome OS
does not really use firmware enabled TPM. Also, this is an option
which is mainboard dependent.
This change drops the addition of PSPTRUSTLETS_FILE to PSP
directory. If this is something that is required by any mainboard,
there should be a separate Kconfig to include the required files.
BUG=b:154880818
TEST=Verified that trembyle still boots
Signed-off-by: Furquan Shaikh <furquan@google.com>
Change-Id: Iaa2126c879986d00c921c85fb5cb5257c7065006
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40685
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
This change drops the addition of PUBSIGNEDKEY_FILE to PSP
directory. This file is used to add OEM key for BIOS, however this is
currently unused for upcoming zork board. In the future, if any
mainboard needs this, it can be added based on some Kconfig selection.
BUG=b:154880818
TEST=Verified that trembyle still boots up fine.
Signed-off-by: Furquan Shaikh <furquan@google.com>
Change-Id: Icd97856a94a100898678702d99bbe29b82956004
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40682
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Inclusion of MP2 firmware is optional and dependent on mainboard. Set
default option for including MP2 firmware in PSP directory to 'n'.
BUG=b:154880818
Signed-off-by: Furquan Shaikh <furquan@google.com>
Change-Id: I1ff7527a409d8ac7f4d30e69eafc53975b63e49b
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40680
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
This change adds support in common block HDA driver to add a PCI
device for HDA in SSDT and removes the HDA device from DSDT for
Stoneyridge and Picasso.
_INI method is still retained in stoneyridge since I am unsure why it
was added. In order to support the _INI method, HDA driver makes a
callback hda_soc_ssdt_quirks() to allow SoCs to add any quirks
required for the HDA device. This callback is implemented by
Stoneyridge to provide the _INI method which retains the same
functionality for HDA device.
This makes it easier to ensure that we don't accidentally
make the DSDT and SSDT entries inconsistent w.r.t. ACPI name and
scope.
BUG=b:153858769,b:155132752
TEST=Verified that audio still works fine on Trembyle.
Change-Id: I89dc46b92fdcb785bd37e18f0456935c0e57eff5
Signed-off-by: Furquan Shaikh <furquan@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40785
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
acpigen_write_field() does not need to modify the fieldlist
parameter. Thus, this change makes this parameter as const.
Change-Id: I94688913cee8948f42ae5e184f2d24264876648d
Signed-off-by: Furquan Shaikh <furquan@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40784
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This change adds the following acpigen helpers:
a. acpigen_write_store_op_to_namestr: This generates ACPI code for
storing an ACPI OP to name string
b. acpigen_write_if_lequal_namestr_int: This generates ACPI code for
checking if operand1 and operand2 are equal where operand1 is
namestring and operand2 is an integer.
Change-Id: I84c158361c0725c2927f06be35391e61f627a453
Signed-off-by: Furquan Shaikh <furquan@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40783
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
_INI method for AZHD device for Picasso family was just copied from
Stoneyridge as part of initial change. There is no evidence that this
is required for Picasso. Also, removing the _INI method works
perfectly fine. Thus, this change drops the _INI method for AZHD
device on Picasso.
Since the _INI method was the only entity using the OperationRegion
fields, this change also drops the operation region.
BUG=b:155132752
TEST=Verified that audio still works on Trembyle
Change-Id: If42abf91ee5cd47a881b0a3b4ca1916ea5169261
Signed-off-by: Furquan Shaikh <furquan@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40782
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This change adds .acpi_name() callback to HDA driver that returns
"AZHD" as the ACPI device name for HDA controller. Since this is now
done by the common HDA driver, this change also removes the HDA device
name returned by stoneyridge in chip.c.
BUG=b:153858769
Change-Id: I89eaa799518572f3c46c7ce9ef8dd3f85daa12bb
Signed-off-by: Furquan Shaikh <furquan@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40781
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This is a cosmetic change to use tabs to align hda_audio_ops and
hdaaudio_driver entries.
Change-Id: I8e398706cbe7087d0178b2433606f8984651c0d6
Signed-off-by: Furquan Shaikh <furquan@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40780
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This change sets ops_pci for hda_audio_ops to default pci_dev_ops_pci
and removes the custom lops_pci since the driver does not really need
a custom ops_pci.
BUG=b:153858769
Change-Id: I4b46e22ef556c0f49152c41a07f3c54c513ae37a
Signed-off-by: Furquan Shaikh <furquan@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40779
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
PCI device PCI_DEVICE_ID_AMD_FAM17H_HDA0 does not really use the same
vendor ID as PCI_VENDOR_ID_AMD. Thus, drop this device from the list
of pci_device_ids[] that are supported by the common hda driver.
BUG=b:153858769
Change-Id: If41dc7179e1e5b476878ee24c8a355b1cde762eb
Signed-off-by: Furquan Shaikh <furquan@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40778
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Since `commit 0ee9b14c09c` the SPD array is set to NULL if no
DIMM is present. This causes failure due to an unconditional use
of `blk.spd_array[i]`, : i={0,1}.
This validates the spd_array is non-NULL before use otherwise it
sets the DIMM as not present.
Puff fails boot with the following log:
```
...
SPD: banks 16, ranks 2, rows 16, columns 10, density 8192 Mb
SPD: device width 8 bits, bus width 64 bits
SPD: module size is 16384 MB (per channel)
ASSERTION ERROR: file 'src/soc/intel/cannonlake/cnl_memcfg_init.c', line 47
```
BUG=b:155220125
BRANCH=none
TEST=none
Change-Id: I5f47c849344951d53fa8c67e779b7c46d632d124
Signed-off-by: Edward O'Callaghan <quasisec@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40820
Reviewed-by: Jamie Chen <jamie.chen@intel.com>
Reviewed-by: Sam McNally <sammc@google.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Daniel Kurtz <djkurtz@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The original plan for Picasso was to combine the features of bootblock
with romstage due to its unique way of coming out of reset. Early in
development, all bootblock support was removed from the directory.
All Picasso designs will now use a bootblock as their first stage. The
reason being that it requires less invasive changes than using a hybrid
romstage.
Add a basic bootblock back to the directory, and compatible with the
design of lib/bootblock.c. The files support RESET_VECTOR_IN_RAM
and add appropriate settings in Kconfig. Make Makefile.inc calculates
the size and base of bootblock from known parameters.
* Future work may attempt to streamline this further, in conjunction
with changes in amdfwtool. See b/154957411.
BUG=b:147042464, b:153675909
Change-Id: I1d0784025f2b39f140b16f37726d4a7f36df6c6c
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37490
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Add support for devices with the reset vector pointing into DRAM. This
is a specific implementation that assumes a paradigm of AMD Family 17h
(a.k.a. "Zen"). Until the first ljmpl for protected mode, the core's
state appears to software like other designs, and then the actual
physical addressing becomes recognizable.
These systems cannot implement cache-as-RAM as in more traditional
x86 products. Therefore instead of reusing CAR names and variables,
a substitute called "earlyram" is introduced. This change makes
adjustments to CAR-aware files accordingly.
Enable NO_XIP_EARLY_STAGES. The first stage is already in DRAM, and
running subsequent stages as XIP in the boot device would reduce
performance.
Finally, add a new early_ram.ld linker file. Because all stages run in
DRAM, they can be linked with their .data and .bss as normal, i.e. they
don't need to rely on storage available only at a fixed location like
CAR systems. The primary purpose of the early_ram.ld is to provide
consistent locations for PRERAM_CBMEM_CONSOLE, TIMESTAMP regions, etc.
across stages until cbmem is brought online.
BUG=b:147042464
TEST=Build for trembyle, and boot to ramstage.
$ objdump -h cbfs/fallback/bootblock.debug
Idx ,Name ,Size ,VMA ,LMA ,File off Algn
0 ,.text ,000074d0 ,08076000 ,08076000 ,00001000 2**12
1 ,.data ,00000038 ,0807d4d0 ,0807d4d0 ,000084d0 2**2
2 ,.bss ,00000048 ,0807d508 ,0807d508 ,00008508 2**2
3 ,.stack ,00000800 ,0807daf0 ,0807daf0 ,00000000 2**0
4 ,.persistent ,00001cfa ,0807e2f0 ,0807e2f0 ,00000000 2**0
5 ,.reset ,00000010 ,0807fff0 ,0807fff0 ,0000aff0 2**0
6 ,.debug_info ,0002659c ,00000000 ,00000000 ,0000b000 2**0
7 ,.debug_abbrev ,000074a2 ,00000000 ,00000000 ,0003159c 2**0
8 ,.debug_aranges,00000dd0 ,00000000 ,00000000 ,00038a40 2**3
9 ,.debug_line ,0000ad65 ,00000000 ,00000000 ,00039810 2**0
10 ,.debug_str ,00009655 ,00000000 ,00000000 ,00044575 2**0
11 ,.debug_loc ,0000b7ce ,00000000 ,00000000 ,0004dbca 2**0
12 ,.debug_ranges ,000029c0 ,00000000 ,00000000 ,00059398 2**3
Change-Id: I9c084ff6fdcf7e9154436f038705e8679daea780
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/35035
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Picasso does not define the state of variable MTRRs on boot. Add a
helper function to clear all MTRRs.
BUG=b:147042464
TEST=Build trembyle
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I21b887ce12849a95ddd8f1698028fb6bbfb4a7f6
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40764
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Setting the default values for GPIO community power management, causes
issues in detecting TPM interrupts. So to avoid that GPIO PM has to be
disabled in devicetree. But for S0ix it is needed. This patch implements
a workaround in ASL code to enable GPIO PM on S0ix entry and disable it
on S0ix exit.
This patch adds the following three platform specific methods.
1. MS0X to enable power management features for GPIO communities on
entry and on exit, it disables them.
2. MPTS to enable power management features for GPIO communities when
preparing to sleep.
3. MWAK to disable power management features for GPIO communities on
waking up.
BUG=b:148892882
BRANCH=none
TEST="Boot with this change on volteer proto1 and check for GPIO
community config with debugger"
Signed-off-by: Venkata Krishna Nimmagadda <venkata.krishna.nimmagadda@intel.com>
Change-Id: If522c82c0069a4bf5738beb73a2b4f11ed6f51d3
Reviewed-on: https://review.coreboot.org/c/coreboot/+/40261
Reviewed-by: Nick Vaccaro <nvaccaro@google.com>
Reviewed-by: Venkata Krishna Nimmagadda <Venkata.krishna.nimmagadda@intel.com>
Reviewed-by: Wonkyu Kim <wonkyu.kim@intel.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Maxim Polyakov