Add function to have CSME firmware enter Soft Temporary Disable mode,
and a corresponding function to put it back into Normal mode. A global
reset is required for the CSME to change modes.
Control changing modes by a new option "ime_mode". Possible values are
- Enable (0): Set the current operation mode to Normal
- Disable (1): Set the current operation mode to Soft Temporary Disable
Reference:
- Intel doc #612229 (CSME 15.0 BIOS Specification)
Change-Id: I38d320fbb157a628c5decc90e6ced78efbf85e0d
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Add a driver for systems with NVIDIA Optimus (hybrid) graphics using
GC6 3.0. The driver provides ACPI support for dynamically powering on
and off the GPU, and a function for enabling the GPU power in romstage.
Tested on system76/gaze15.
Change-Id: I2dec7aa2c8db7994f78a7cc1220502676e248465
Signed-off-by: Jeremy Soller <jeremy@system76.com>
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Setting this causes boot to fail when upgrading from a version that did
not have it already set to Enabled.
Change-Id: I3d04cd659d5d53745de618703ec1590ca499f70a
Signed-off-by: Tim Crawford <tcrawford@system76.com>
The Oryx Pro 6 has the same board layout as the next model in series,
Oryx Pro 7. The primary difference between the two is the dGPU (20
series to 30 series). Convert oryp6 to a variant setup in preparation
for adding the oryp7.
Change-Id: I976750c7724d23b303d0012f2d83c21a459e5eed
Signed-off-by: Tim Crawford <tcrawford@system76.com>
Some older devices, like the x230 Thinkpad, do not boot with the
newer Tianocore UefiPayloadPkg build target, and cannot easily be
debugged without serial UART output. As a stopgap solution, re-add
the older (now deprecated/removed) CorebootPayloadPkg build target.
This partially reverts commit d3b49b4c,
"payloads/Tianocore: Update default build target, simplify build options"
Change-Id: I81490c277626fc69d95920868d80cb24c0763de4
Signed-off-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58710
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
Some brya0 units have HPS fitted and connected to PCH I2C2, rather than
a user-facing camera.
Because HPS uses I2C address 0x51, which may conflict with the
user-facing camera EEPROM, introduce a new fw_config bit to indicate
whether HPS is present.
BUG=b:202784200
TEST=FW_NAME=brya0 emerge-brya coreboot chromeos-bootimage
TEST=ectool cbi set 6 0x28191 4 # set bit 17 for HPS
TEST=flashrom -p internal -w image-brya0.serial.bin
Signed-off-by: Dan Callaghan <dcallagh@google.com>
Change-Id: I322548bcfccf16ba571396bc88fd6fc03c036a4e
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58646
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Make use of `smbios_bus_width_to_spd_width()` for filling DIMM info.
Additionally, ensures dimm_info_util.c file is getting compiled for
romstage.
TEST=dmidecode -t 17 output Total Width and Data Width as expected.
Change-Id: I7fdc19fadc576dec43e12f182fe088707e6654d9
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58655
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
The reason cbfs_cache was disabled on x86 was due to the lack of
.data sections in the pre-RAM stages. By using
ENV_STAGE_HAS_DATA_SECTION we enable x86 to start using the cbfs_cache.
We still need to add a cbfs_cache region into the memlayout for it to
be enabled.
BUG=b:179699789
TEST=Build guybrush and verify cbfs_cache.size == 0.
Suggested-by: Julius Werner <jwerner@chromium.org>
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I74434ef9250ff059e7587147b1456aeabbee33aa
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56577
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
The PCI-SIG engineering change requirement provides the ACPI additions
for firmware latency optimization. This change adds additional ACPI DSM
function with both of FW_RESET_TIME and FW_D3HOT_TO_D0_TIME to the
USB4/TBT topology which has the same implementation on Tiger Lake in
commit I5a19118b75ed0a78b7436f2f90295c03928300d7.
BUG=b:199757442
TEST= It was validated that the first connected device waits only 50ms
instead of 100ms and all functions work on Alder Lake platform boards.
Signed-off-by: John Zhao <john.zhao@intel.com>
Change-Id: I0c8977c96de27ab0e554469eba658660975b8493
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58098
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Used H1 firmware where the last version number is 0.0.22, 0.3.22 or
less to production that will need to disable autonomous GPIO power
management and then can get H1 version by gsctool -a -f -M
BUG=b:201266532
TEST=FW_NAME=kano emerge-brya coreboot and verify it builds
without error.
Signed-off-by: David Wu <david_wu@quanta.corp-partner.google.com>
Change-Id: If6783e0df1404c9a353061fb564210aa0d12896e
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58682
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Set the `PcieRpLtrEnable` option to enable the LTR capability on all PCH
PCIe root ports.
TEST=Verify LTR capability enabled in `DevCap2` using `lspci -vv`
Change-Id: I07ea37d178ea61d904c4f131fdea31479e899ef3
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58326
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Map each PCIe clock source to the corresponding root port. Also, correct
the CLKREQ# mapping for clock sources not associated to any CLKREQ# pin.
The default `PcieClkSrcClkReq` value of 0 corresponds to CLKREQ# 0.
TEST=Check that Linux sees the same PCIe devices with this commit:
- All 5 onboard Ethernet NICs
- BMC
- Two random graphics cards in PEG0 and PEG1 slots
- M.2 M NVMe SSD
Change-Id: I0515877a36d42fb8858a0f0b3c0af1199a18d9af
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58368
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
The current check for XDCI enabled uses a static device path to an
internal PCI device at a very late point in the boot flow. At this
time the devicetree has been processed and disabled devices have been
already removed. If this device (00:15.1, XDCI) is disabled in
devicetree this will trigger the message
'BUG: check_xdci_enable requests hidden 00:15.1' in the log.
This looks weird and is wrong since it is not a bug to disable this
device when it is not needed.
To avoid this look up the devicetree by a tree walk instead of using
a static value for the devicetree.
Change-Id: If193be724299c4017e7e10142fac8db9fac44383
Signed-off-by: Werner Zeh <werner.zeh@siemens.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58524
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Correct the PCIe clock source configuration as per the schematics.
Apparently, FSP does not turn off unused PCIe clock sources when using
SPS (Server Platform Services) firmware, but it does when using CSME
firmware.
TEST=BMC and Ethernet NICs get detected when using CSME firmware.
Change-Id: Id25a34816f512510640db95251a7a792c1eebe62
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58065
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
As per MP service specification, EDK2 is allowed to specify the mode
in which a 'func' routine should be executed on APs.
`SingleThread` sets to 'true' meaning to execute the function one by
one (serially) or sets to 'false' meaning to execute the function
simultaneously.
MP service API `StartupAllAPs` was designed to pass such options as
part of function argument.
But another MP service API `StartupAllCPUs` doesn't specify any such
requirement. Running the `func` simultaneously on APs results in
a coherency issue (hang while executing `func`) due to lack of
acquiring a spin lock while accessing common data structure in
multiprocessor environment.
BUG=b:199246420
Change-Id: Ia95d11408f663212fd40daa9fd9b0881a07f1ce7
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Signed-off-by: Ronak Kanabar <ronak.kanabar@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57343
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Maulik V Vaghela <maulik.v.vaghela@intel.com>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
