PC Engines apu2 had many problems with PCIe cards detection. The cards
were inconsistently detected when booted from G3, S5 or after a reboot.
AGESA can reset PCIe slots using GPIO via callback. Use it to reset the
slots that support using GPIO as reset signal.
Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Change-Id: I8ff7db6ff85cce45b84729be905e6c895a24f6f2
Reviewed-on: https://review.coreboot.org/c/coreboot/+/39703
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
The `link_frequency_270_mhz` setting was originally used by the native
graphics init code for Sandy/Ivy Bridge, which is long gone.
The value of this information (which board had it set) is questionable.
The only board that had an LVDS panel and set it to 0 was the ThinkPad
L520, where native graphics init was never reported to work. Also, the
native graphics init only used it for calculations, but never confi-
gured the hardware to use a specific frequency. A look into the docu-
mentation also doesn't reveal any straps that could be used to confi-
gure it.
Change-Id: Ieceaa13e4529096a8ba9036479fd84969faebd14
Signed-off-by: Nico Huber <nico.h@gmx.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/39763
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Alexander Couzens <lynxis@fe80.eu>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The PWM-granularity chicken bit in the Wildcat Point and Lynx Point
PCHs has actually the opposite meaning of the one for Sunrise Point
and later. When the bit is set, we get a divider of 16, when it's
unset 128. Flip the bit!
Change-Id: I1dbde1915d8b269c11643a1636565a560eb07334
Signed-off-by: Nico Huber <nico.h@gmx.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/39770
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
_DCS, _DGS and _DSS are required by specification. However,
we never implemented them properly, and no OS driver com-
plained yet. So we stub them out and keep the traditional
behavior in case an OS driver checks for their existence.
The old implementations also only returned static values as
there never was any write to their GNVS variables. The TRAP()
that was called in one place is actually implemented by some
ThinkPad's SMI handler as docking event. However, as the call
precedes these SMI handlers in coreboot history, it's most
likely an accident.
Change-Id: Ib0b9fcdd58df254d3b2290900e3bc206a7abd92d
Signed-off-by: Nico Huber <nico.h@gmx.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/39726
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: HAOUAS Elyes <ehaouas@noos.fr>
According to the latest Tigerlake Platform FSP Integration Guide, the
minimum amount of stack needed for FSP-M is 256KiB. Change
DCACHE_BSP_STACK_SIZE to reflect that (plus 1KB previously determined
empirically). JSL requires 192KiB.
Signed-off-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Change-Id: Ic9be6446c4db7f62479deab06ebeba2c7326e681
Reviewed-on: https://review.coreboot.org/c/coreboot/+/39706
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Subrata Banik <subrata.banik@intel.com>
Native raminit asserts that the DIMMs haven't been replaced before
reusing the saved training data. However, it does not check if the CPU
is still the same, so it can end up happily reusing data from an Ivy
Bridge CPU onto a Sandy Bridge CPU, which runs the raminit_ivy.c code
path. This can make the CPU run in unsupported configurations, which may
result in an unstable system, or a failure to boot.
To prevent that, ensure that the stored CPUID matches the CPUID of the
installed CPU. If they differ, print a message and do not use the saved
data. As it does not pose a problem for a regular boot, but precludes
resuming from S3, use different loglevels depending on the bootpath.
Tested on Asus P8Z77-V LX2 with an i7-2600 and an i5-3330, works well.
Change-Id: Ib0691f1f849b567579f6afa845c9460e14f8fa27
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/39734
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Instead of storing an int with a single bit of information taken from
the CPUID, we might as well store the actual CPUID. And since we are
changing the definition of the saved data, bump the version number.
Tested on Asus P8Z77-V LX2, still boots fine.
Change-Id: I6ac435fb83900a52890f823e7614055061299e23
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/39720
Reviewed-by: HAOUAS Elyes <ehaouas@noos.fr>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
When changing any of the structures that are cached in non-volatile
storage, it is necessary to bump MRC_CACHE_VERSION so that the old
information is not misinterpreted.
Change-Id: Idefbc38b3a8198b1b5909e775b3c289db689fc0c
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/39756
Reviewed-by: HAOUAS Elyes <ehaouas@noos.fr>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Header files are supposed to not make allocations from .bss. Builds
fail if said file is included multiple times. To prevent this from
happening, move the definitions to a C file.
Also, rename raminit_patterns to raminit_tables. This is because more
tables that are not patterns will be added here in subsequent changes.
Tested on Asus P8Z77-V LX2, still boots fine.
Change-Id: If8e3a285ecdc4df9e978ae156be915ced6e1750b
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/39754
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The ACPI Spec 2.0 states, that Processor declarations should be made
within the ACPI namespace \_SB and not \_PR anymore. \_PR is deprecated
and is removed here for Intel CPUs only.
Tested on:
* X11SSH (Kabylake)
* CFL Platform
* Asus P8Z77-V LX2 and Windows 10
FWTS does not return FAIL anymore on ACPI tests
Tested-by: Angel Pons <th3fanbus@gmail.com>
Change-Id: Ib101ed718f90f9056d2ecbc31b13b749ed1fc438
Signed-off-by: Christian Walter <christian.walter@9elements.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37814
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
According to BKDGs HTC temperature limit field indicates the threshold
where HTC becomes active. HTC active state means that processor is
limiting its power consumption and maximum P-State. Using this threshold
as _CRT is incorrect, since HTC active is designed to prevent
overheating, not causing immediate shutdown.
Change the behavior of temperature limit to act as a passive cooling
threshold. Make the passive cooling threshold a reference value for
critical and hot temperature with 5 degrees step.
TEST=boot FreeBSD on PC Engines apu2 and check the thermal zone
temperature using sysctl
Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Change-Id: Ife64c3aab76f8e125493ecc8183a6e87fb012e3b
Reviewed-on: https://review.coreboot.org/c/coreboot/+/39697
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
This is required for PD notifications on the cros_ec driver.
BUG=b:150649744
TEST=Boot volteer with this patch and verify that PD notifier events are
being generated.
Signed-off-by: Prashant Malani <pmalani@chromium.org>
Change-Id: I2e72320b025a3dfa7412181586cb142a4503eda5
Reviewed-on: https://review.coreboot.org/c/coreboot/+/39675
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Currently cbfstool cherry-picks a few files from vboot and hopes these
files will work standalone without any dependencies. This is pretty
brittle (for example, CL:2084062 will break it), and could be improved
by building the whole vboot library and then linking against it.
Therefore, this patch creates a new target $(VBOOT_HOSTLIB) and includes
it as a dependency for cbfstool and ifittool.
To prevent building the vboot lib twice (one for cbfstool and the other
for futility) when building coreboot tools together, add the variable
'VBOOT_BUILD' in Makefile to define a shared build path among different
tools so that vboot files don't need to be recompiled.
Also ignore *.o.d and *.a for vboot library.
BRANCH=none
BUG=none
TEST=make -C util/cbfstool
TEST=make -C util/futility
TEST=Run 'make tools' and make sure common files such as 2sha1.c are
compiled only once
TEST=emerge-nami coreboot-utils
Change-Id: Ifc826896d895f53d69ea559a88f75672c2ec3146
Signed-off-by: Yu-Ping Wu <yupingso@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/39390
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
It does not make sense to disable an optimization that was not enabled
before, especially if that optimization only applies to Ivy Bridge.
Tested, still boots and can suspend correctly with:
- Asus P8Z77-V LX2 with i5-3330 and Windows 10
- Gigabyte GA-H61MA-D3V with i5-2400 and Arch Linux
Change-Id: I9f3eb545585824bbdf51e33f0592e7daa1c425af
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/39623
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-by: Nico Huber <nico.h@gmx.de>
