With CSE-lite enabled, we were going through the lengthy memory
training procedure twice on the first power-on boot or after full BIOS
SPI flash update. This moves the global reset performed to achieve the
CSE-lite RO to RW reboot to a later boot phase so that it happens
after the memory training data has been written to the MRC cache. Now,
the 2nd (and subsequent) reboot can utilize the memory training data
established during the 1st boot.
This reduces the first boot time by about 20s on a 16GB system.
Looking at the timing stats form cbmem, the normal boot penalty is
about 300ms - mostly attributed to running FspSiliconInit a 2nd
time. We will get this time back when the mrc_cache refactoring effort
lands (cb:44196, et. al).
BUG=b:162021048
TEST=Booted on volteer, confirmed 20s faster boot time.
Change-Id: Ia42d72fdec41f9792ab8f04205b20a55758a4235
Signed-off-by: Caveh Jalali <caveh@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44430
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Reviewed-by: Nick Vaccaro <nvaccaro@google.com>
Add code for IVRS generation to coreboot. Publish coreboot generated
structure rather than IVRS generated by FSP binary.
Reference Doc: 48882_IOMMU_3.05_PUB.pdf
BUG=b:155307433
TEST=Boot trembyle to shell and extract and compare IVRS tables and make
sure they cover the same devices.
Change-Id: I693f4399766c71c3ad53539634c65ba59afd0fe1
Signed-off-by: Jason Glenesk <jason.glenesk@amd.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/43804
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
For Volteer (and future Tiger Lake boards) we can enable mode S0i3.4
only if we know that the Cr50 is generating 100us interrupt pulses.
We have to do so, because the SoC is not guaranteed to detect pulses
shorter than 100us in S0i3.4 substate.
A new Kconfig setting CR50_USE_LONG_INTERRUPT_PULSES controls new code
running in verstage, which will program a new Cr50 register, provided that
Cr50 firmware is new enough to support the register.
BUG=b:154333137
TEST=util/abuild/abuild -t GOOGLE_VOLTEER -c max -x
Signed-off-by: Jes Bodi Klinke <jbk@chromium.org>
Change-Id: If83188fd09fe69c2cda4ce1a8bf5b2efe1ca86da
Reviewed-on: https://review.coreboot.org/c/coreboot/+/43741
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Julius Werner <jwerner@chromium.org>
Kconfig 5.8 interprets $(...) itself using environment variables, which
generally means that they expand to the empty string. \$(...) works
with both our current and new Kconfig with the desired behavior
(to pass it through unmodified).
Change-Id: I726567eeb61d2035560152677d2b4548c1472be9
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44584
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: HAOUAS Elyes <ehaouas@noos.fr>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Not selecting `ME_MBP_CLEAR_LATE` results in a build failure. Since both
traditional and ULT platforms are known to be working, drop the option.
Change-Id: I09ce27f812966800e36f6c0624c93759089faf45
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44547
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Together with the "AMD_XMP" changes, now this board with Crucial
BLT8G3D1869DT1TX0 sticks could run at 1600 MT/s CL8 (8-8-9-23) speeds.
Earlier only 1333 MT/s CL9 (9-9-10-27) has been possible with coreboot.
1866 MT/s CL9 is impossible on f16kb without northbridge overclocking.
tRP in "CL-tRCD-tRP-tRAS" gets set 1 point higher by AGESA because of
Errata 638. See more info in a BKDG for AMD Family 16h Models 00h-0Fh.
Signed-off-by: Mike Banon <mikebdp2@gmail.com>
Change-Id: I7e9f5120421221043f9f9dfe143b51bfa61936be
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44462
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
AMD f16kb boards are perfectly capable of working at 1600MT/s RAM speeds
even with two DDR3 UDIMM modules per channel. AM1I-A only supports a
single-channel operation, with at most two DIMMs per channel, so raising
these limit values is required to let it and similar boards run faster.
Successfully tested on AM1I-A and two Crucial BLT8G3D1869DT1TX0 UDIMMs,
together with related AMD_XMP changes - also required to get a 1600MT/s
with this set of modules which have only 1333MT/s at JEDEC part of SPD.
Signed-off-by: Mike Banon <mikebdp2@gmail.com>
Change-Id: I2a9da4e594ab3dc38b5ba87520633cbd01c9ce01
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44461
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Since Volteer also uses the CSE Lite SKU and the cr50, it is subject
to a problem where old cr50 FW will not be able to properly detect an
SoC reset, so the reset on cold boots caused by the CSE Lite RO->RW
jump should instead get an assist from the EC, which can perform a
full cold reset.
BUG=b:162977697
TEST=Verify EC performs the cold reset
Signed-off-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Change-Id: Ie8ae21c203da218459d5fd30a23be23520ed0598
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44536
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Caveh Jalali <caveh@chromium.org>
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
* Enable optional x86_64 romstage, postcar and ramstage
* Add Kconfig for x86_64 compilation
* Add documentation for x86 qemu mainboards
* Increase CAR stack as x86_64 uses more than 0x4000 bytes
Working:
* Boots to Linux
* Boots to SeaBIOS
* Drops to protected mode at end of ramstage
* Enumerates PCI devices
* Relocateable ramstage
* SMM
Change-Id: If2f02a95b2f91ab51043d4e81054354f4a6eb5d5
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/29667
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
* On ARCH_RAMSTAGE_X86_64 jump to the payload in protected mode.
* Add a helper function to jump to arbitrary code in protected mode,
similar to the real mode call handler.
* Doesn't affect existing x86_32 code.
* Add a macro to cast pointer to uint32_t that dies if it would overflow
on conversion
Tested on QEMU Q35 using SeaBIOS as payload.
Tested on Lenovo T410 with additional x86_64 patches.
Change-Id: I6552ac30f1b6205e08e16d251328e01ce3fbfd14
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/30118
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
The AM335X is a SoC, so should be in the soc tree.
This moves all the existing am335x code to soc/ and updates any
references. It also adds a soc.c file as required for the ramstage.
Change-Id: Ic1ccb0e9b9c24a8b211b723b5f4cc26cdd0eaaab
Signed-off-by: Sam Lewis <sam.vr.lewis@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44378
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
Implements fit_payload_arch for the arm (aarch32) architecture, so that
FIT images can be used. The implementation is very similar to the
existing implementations for arm64 and riscv, and has mostly been
lifted from these other ports.
TEST: Booted Beaglebone Black (in progress port, to be submitted soon!)
with a FIT image containing a 5.4 kernel, dtb and initramfs.
Change-Id: I6b50c6f06b83c00a5b3622b5bbafe67130b6d233
Signed-off-by: Sam Lewis <sam.vr.lewis@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44377
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
In order to log gpio events for wake purposes the state
of the gpio subsystem should be snapshotted. Add the ability
to capture state of gpio subystem as well as saving up to 16
gpios that indicate their wake status.
Likewise, provide the eventlog additions based on state.
BUG=b:159947207
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Change-Id: I49fca56c87543aa8aad0eb7da5c5cb570c4349d5
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44534
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Switch from locating the AMD firmware in the RW_A &
RW_B regions with their hardcoded locations to using CBFS to find
them. They still need to be at the hardcoded locations so that we
can set the location inside the binary, but instead of just setting
the pointer directly to them, we now search for them with cbfs.
BUG=b:154441227
TEST=Boot & verify that binaries are located in both RW-A & RW-B
Signed-off-by: Martin Roth <martin@coreboot.org>
Change-Id: I27b0593e0db7a9e6ba9b0633ac93b4d93954f002
Reviewed-on: https://review.coreboot.org/c/coreboot/+/42831
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Eric Peers <epeers@google.com>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
In CB:44488 the cbmem addition was re-filling the object
when it should be memcpy()ing from static object. Correct
that oversight. The side effect from the previous implementation
would be if FSP-M modified the GPE state.
BUG=b:159947207
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Change-Id: I158a89ae28431896fa9b5789292000fcbf0b066d
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44533
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
GPP_A19(DP_HPD1) and GPP_A20(DP_HPD2) were configured native function
(NF1) without internal pull-down which wrongly presents HPD interrupts.
DP_HPD had been removed for EVT design as those events are through eSPI.
This change configures GPP_A19 and GPP_A20 to be no connection and
disables DdiPort1Hpd and DdiPort2Hpd.
BUG=b:162566436
TEST=Booted to kernel and verified no kernel HPD pins assertion message
on Volteer EVT board.
Signed-off-by: John Zhao <john.zhao@intel.com>
Change-Id: Ia3245741b776b75073d2b43d36c8ea40b476b3ed
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44501
Reviewed-by: Caveh Jalali <caveh@chromium.org>
Reviewed-by: Wonkyu Kim <wonkyu.kim@intel.com>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The wake source macro for GPE events was using 'GPIO'. However,
current usage is really all GPEs. Therefore, provide clarity
in the naming in order to allow for proper GPIO wake events
that are separate from the ACPI GPE block.
BUG=b:159947207
Change-Id: I27d0ab439c58b1658ed39158eddb1213c24d328f
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44527
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
