Stoney Ridge ASL uses global NVS for sharing data between ACPI and SMM.
Change-Id: I8f497870844da1f56ff5c3b126d5613be36a9bfb
Signed-off-by: Marc Jones <marcj303@gmail.com>
Reviewed-on: https://review.coreboot.org/22425
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Explicitly add #include files to romstage.c to ensure sizes of the
devicetree structures are correct. The AMD support headers have an
open #pragma pack(1) which causes structure sizes to change based on
include ordering in different compilation units. More concretely, this
fixes a bug where dev->chip_info is incorrectly detected as 0.
Also shorten a printk string to bring the source line within 80 columns.
Change-Id: I1ed51cdbb8df387a453de6cb944b90538dac4431
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/22424
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Several SPD hex files for chell were missing from upstream coreboot
(as compared to the Chromium tree/branch), which resulted in the
incorrect type and amount of RAM being reported on chell boards
with > 4GB RAM. Add these missing files and their Makefile entries.
TEST: boot google/chell m7/16GB config and observe correct RAM
type and amount reported via dmidecode and cbmem console log.
Change-Id: I37d708c96e754b438e40fc413420aa64bf234c29
Signed-off-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-on: https://review.coreboot.org/22402
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
ModuleIdentifier must be 8 bytes. Every other location else that uses
this value explicityly defines it as 8 bytes. If it's initialized here
to less than 8 bytes, it gets passed to those other locations with
garbage at the end and fails to load the AGESA binary.
TEST=Build & boot Kahlee
BUG=B:69165234
Change-Id: I11fc90748f49782e2b16ee5326aee17cfe92d0bc
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/22430
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Justin TerAvest <teravest@chromium.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Set the UMA memory size to 128 MiB. This value was empirically tested
by AMD as the lowest value one could use.
BUG=b:64927639
TEST=default, and 64, 128, 256, 384MB non-legacy configurations.
Change-Id: I2bc808d8b402c3eb16a1a5962f3fa9d6b224cf52
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Signed-off-by: Richard Spiegel <richard.spiegel@silverbackltd.com>
Reviewed-on: https://review.coreboot.org/21335
Reviewed-by: Martin Roth <martinroth@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Add three settings for the UMA configuration to correspond with
definitions in AGESA.h.
* UMA off, Auto, or size specified
* Size (if specified above)
* Legacy vs. non-legacy (if Auto)
BUG=b:64927639
Change-Id: I38b6603f365fdc1f1f615794365476f749e58be7
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Signed-off-by: Richard Spiegel <richard.spiegel@silverbackltd.com>
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/21334
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
The AcpiPm1EvtBlk base I/O address is configured in PMx60. Add a
helper function to read this. The register is not lockable so it
shouldn't be assumed to be at its original address.
Change-Id: I91ebfb454c2d2ae561e658d903f33bfb34e1ad6f
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/22413
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marc Jones <marc@marcjonesconsulting.com>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
The revision level is not checked. This was probably left over from
trying to determine hudson variants.
Remove the unused SMI command port values. This was missed in:
e9b862e amd/stoneyridge: Use generic SMM command port values
Change-Id: I91d8051372f4e238d390dd445d0bf06d06683a66
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/22412
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marc Jones <marc@marcjonesconsulting.com>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Unhardcode the acpi_get_sleep_type() function and rely on the new
function in arch/acpi.h.
Change-Id: Icd49c44fae43effb9f082db354abd327cad9f1ad
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/22410
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Marc Jones <marc@marcjonesconsulting.com>
SPD address is currenty int. It should be uint8_t.
BUG=b:62200225
Change-Id: Ia11c5994c41849ba01ecae3cee6fa97c527134d0
Signed-off-by: Richard Spiegel <richard.spiegel@silverbackltd.com>
Reviewed-on: https://review.coreboot.org/22300
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
Solve issues left from Change-Id Ib88a868e654ad127be70ecc506f6b90b784f8d1b
Unify code: smbus.c to have the actual execution code, sm.c and smbus_spd.c
call functions within smbus.c.
Fix some functions that wrongly use SMBHSTCTRL as the register for the
data being transfered. The correct register is SMBHSTDAT0.
Include file smbus.h should only be used by sm.c, smbus.c and smbus_spd.c.
BUG=b:62200225
Change-Id: Ibd55560c95b6752652a4f255b04198e7a4e77d05
Signed-off-by: Richard Spiegel <richard.spiegel@silverbackltd.com>
Reviewed-on: https://review.coreboot.org/21887
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Marc Jones <marc@marcjonesconsulting.com>
Add a #define for MB_DIMM_SLOTS and verify it doesn't exceed the max
supported for the device. AGESA's DRAM procedures follow the BKDG and
may vary depending on the number of slots on the motherboard. DIMM
numbering and ordering is also affected by this value.
Replace hardcoded integers with defined values for DIMM slots and
number of channels.
Change-Id: I4f7336da80b4e3d7f351502a63de0652e9ff5395
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/21853
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Kahlee needs to keep its DRAM contents after a reset. Move this
override out of the OemCustomize.c file to a devicetree register
setting.
Change-Id: I3196cb8b94bec64e8ce59e4285cf8d97f442bd3d
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Signed-off-by: Richard Spiegel <richard.spiegel@silverbackltd.com>
Reviewed-on: https://review.coreboot.org/21858
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
AmdInitPost() can be instructed to clear DRAM after a reset or to
preserve it. Use SetMemParams() to tell AGESA which action to take.
Note that any overrides from OemPostParams (OemCustomize.c) are not
affected by this change.
Change-Id: Ie18e7a265b6e0a00c0cc8912db6361087f772d2d
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Signed-off-by: Richard Spiegel <richard.spiegel@silverbackltd.com>
Reviewed-on: https://review.coreboot.org/21856
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Variable without an initializer will default to 0 hence no
need of an explicit initialization.
Change-Id: I208d5e475600b102cd3d972919b170c10c790b32
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/22417
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
This patch ensures Skylake/Kabylake soc can make use of common
CSE code in order to perform global reset using HECI interface.
TEST=Build and boot on soraka/eve/reef/cnl-rvp
Change-Id: I49b89be8106a19cde1eb9b488ac660637537ad71
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/22394
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
This ensures HECI1_BASE_ADDRESS macro is coming from respective
SoC dirctory and not hardcoded inside common cse code. As per
firmware specification HECI1_BASE_ADDRESS might be different
between different socs.
Change-Id: I502b5b41b449bb07f14f07435bf311bbd4f943b6
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/22393
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Some prototypes types don't match the actual function type, though there's
no error message due to the types being alias. For clarity, types should
match between prototypes and actual functions.
BUG=b:68007655
Change-Id: I9573a68f7153dbbad2fc6551d5dab000760c871e
Signed-off-by: Richard Spiegel <richard.spiegel@silverbackltd.com>
Reviewed-on: https://review.coreboot.org/22288
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marc Jones <marc@marcjonesconsulting.com>
Reviewed-by: Martin Roth <martinroth@google.com>
On Astronaunt, after the system enters the S5 power state, there is a
10-second timeout before the system transitions the power state from S5
to G3. The EN_PP3300_DX_LTE_SOC signal, which is controlled by GPIO_78
on the APL platform, remains on during that period. If the system is
powered back on before going to G3, the built-in modem won't go through
a power cycle as EN_PP3300_DX_LTE_SOC is never de-asserted.
Keeping the modem, and indirectly the SIM, powered during a quick system
power cycle may sometimes be undesirable. For instance, we would like a
SIM with PIN lock enabled to require unlocking each time the system is
powered on. After the SIM receives a PIN, it may remain unlocked until
its next power cycle.
Also, it is often desirable to power cycle the modem when the system
goes through a power cycle. For instance, a user may power cycle the
system to recover a wedged modem.
BUG=b:68365029
TEST=Tested the following on an Astronaunt device:
1. Verify that the modem is powered on after the system boots from cold.
2. Suspend the system to S0ix. Verify that the modem remains powered on
when the system is in S0ix. After the system goes back to S0, verify
that the SIM with PIN lock enabled doesn't request unlocking, and the
modem can quickly reconnect to a network.
3. Configure the system to suspend to S3 instead of S0ix, and then
repeat (2).
4. Perform a quick system power cycle, verify that the modem is powered
cycle and the SIM with PIN lock enabled requests unlocking.
Change-Id: Ie60776d5d9ebc6a69aa9e360bd882f455265dfa2
Signed-off-by: Ben Chan <benchan@chromium.org>
Reviewed-on: https://review.coreboot.org/22415
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
This further allows compilation units to be re-used without
having to add macro guards because of declarations not being
around in the __SIMPLE_DEVICE__ case. These declarations are for
functions that operate on struct device. struct device is a known
type so just expose the functions using the correct type. Also,
DEVTREE_CONST is empty while in ramstage so there's no reason
to separate the declarations. They compile regardless of stage.
Change-Id: Idd4180437d30e7dfaa9f735416c108841e43129f
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/22397
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Subrata Banik <subrata.banik@intel.com>
In case the TPM is doing a long crypto operation the initial probe
could be very delayed. Rather than end up in recovery make the delay
long enough to accommodate the (current) long crypto times.
BUG=b:65867313, b:68729265
TEST=Verified that Soraka no longer hangs during EC reboot test.
Change-Id: I3bccff70e001dfc065c24be8ad34ef239a144db1
Signed-off-by: Duncan Laurie <dlaurie@google.com>
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/22379
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This patch enables customized NIC leds as follows:
Green Orange (Amber)
100M off blinking
1000M on blinking
BUG=b:65437780, b:68284778
TEST=Make sure the registers are programmed as expected and observe the
LEDs are behaving as expected. Perform suspend/resume test and the
LEDs are still working as expected.
Change-Id: I9bb1367a4c742c2755d620e14ee6dfe70ee7f34b
Signed-off-by: Gaggery Tsai <gaggery.tsai@intel.com>
Reviewed-on: https://review.coreboot.org/22293
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Furquan Shaikh <furquan@google.com>
This compares to the acpi directory in other variants.
BUG=b:68293392
TEST=Build and boot kahlee
Change-Id: I05d402995b280d6f020bc2575063dbffefa30670
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/22372
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Even though this GPIO isn't used for Kahlee, it needs to be defined
so that the weak version of the variant_board_id() function can
compile.
BUG=b:68293392
TEST=Build and boot kahlee
Change-Id: Ia8daf70fbafe02ec37c6b5eb8421cdb11de3be8b
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/22371
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Stoney Ridge supports two different sets of SMU firmware, one for
either fanless OPNs and one for fanned.
Add a Kconfig mechanism to select the proper set and add the blobs
into cbfs.
BUG=b:66339938
Change-Id: I8510823e2232b74ec6fe001cc28953f53b2aa520
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/22057
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
An upcoming PSP firmware change will allow coreboot to load the two
SMU firmware blobs (one runs in SRAM and the other in DRAM). The
traditional method is for the PSP to control most of the process,
e.g. loading the SRAM version prior to releasing the x86 reset.
Add a new command that can instruct the PSP to load a firmware blob
from a location in the flash.
The definition for commands 19 and 1a differ from others in that they
do not use a command/response buffer. Instead, the PSP will look in
the command/response pointer registers directly for the blob's
address.
BUG=b:66339938
Change-Id: I8431af341930f45ac74f471628b4dc4ede7735f4
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/22056
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Move the two functions in fixme.c to places where they make more sense.
Coincidentally fix the todo in amd_initcpuio() and use bsp_topmem()
instead of explicitely reading the MSR.
BUG=b:62241048
Change-Id: Ica80b92f48788314ad290ccf72e6847fb6d039c3
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/22297
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Convert functionality to use coreboot-centric functions and defined
values. This change should have no functional effect.
BUG=b:62241048
Change-Id: I87b258f3187db4247b291c848b5f0366d3303c75
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/22248
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Now that recovery button has been enabled through cr50,
we can add VBOOT_PHYSICAL_REC_SWITCH config and treat
the recovery button like on any other chromebox.
BUG=b:37751915, b:63893483
BRANCH=None
TEST=With DUT in normal mode, boot into recovery, press ctrl+d
followed by pressing the recovery button. Should successfully
boot into dev mode.
CQ-DEPEND=CL:737477
Change-Id: I72fb42508083295e317dd06900796dc0cda753f6
Signed-off-by: Shelley Chen <shchen@chromium.org>
Reviewed-on: https://review.coreboot.org/22368
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
SeaBIOS 1.10.3 was tagged on October 12th, 2017 with the following
changes.
```
$ git log --oneline rel-1.10.2..rel-1.10.3
b7661dd tcgbios: Fix use of unitialized variable
6055583 boot: Increase description size in boot menu
3551613 resume: Don't attempt to use generic reboot mechanisms on QEMU
```
Change-Id: I3a9ebf10a55118fc35aed688ea7ec794333c8227
Signed-off-by: Martin Kepplinger <martink@posteo.de>
Reviewed-on: https://review.coreboot.org/22358
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Marc Jones