Using existing defines instead of magic values improves readability of
the code. Also add comments to the MADT IRQ overrides to make it clearer
what those actually do.
TEST=Timeless build results in identical binary for amd/gardenia
(Stoneyridge), amd/mandolin (Picasso) and amd/majolica (Cezanne)
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I224ffbe8eb65bcdd5fc70c0ff8b15d55b3f6be01
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55613
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Factoring out those defines brings the Stoneyridge SoC code a bit more
in line with the Cezanne and Picasso SoC code.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ifba7f13cc926ac28376233aa0bf317164ca9bbd6
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55588
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Since the default state of the MMIO UART devices in the chipset
devicetree is off, the mainboard devicetree entries that disable MMIO
UART devices are removed.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I913a587802020ce4e182b48632cdde1104c2a6e6
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55545
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Updating the APOB takes a considerable amount of time. I decided to be
granular and split out the operations so we know when we read vs read +
erase + write.
BUG=b:179092979
TEST=Boot guybrush and dump timestamps
3:after RAM initialization 3,025,425 (44)
920:starting APOB read 3,025,430 (5)
921:starting APOB erase 3,025,478 (48)
922:starting APOB write 3,027,727 (2,249)
923:finished APOB 3,210,965 (183,238)
4:end of romstage 3,210,971 (6)
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I08e371873112e38f623f452af0eb946f5471c399
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55401
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
In acpi_soc_get_bert_region after the bert_errors_region call is was
checked if the region parameter is NULL after the call; since region is
a parameter of acpi_soc_get_bert_region, it's non-NULL. What we should
be checking here is if region points to a non-NULL pointer.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reported-by: Coverity (CID:1457506)
Change-Id: I0523504d65725ab2d2df4db28a5dedd90697b917
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55541
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Even though the code is currently commented out, replace the magic
numbers with the existing defines.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Id0dbbadf71f2e5a4d23ee998e2aa0a8b67205845
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55534
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Factoring out those defines allows using them easily in the ACPI code
without having to use preprocessor macros.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ib9dfddb0d4f32a542fa652ff8c14e932c224f247
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55533
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Factoring out those defines allows using them easily in the ACPI code
without having to use preprocessor macros.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I99cb03de8782a0eeeb505f567b982099b0e8a18d
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55531
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Part of the soc/amd/stoneyridge code already uses the FCH_IOAPIC_ID and
GNB_IOAPIC_ID defines. Use those defines in the remaining location to
make sure that the IOAPIC IDs are always consistent between the hardware
register, the MADT and the IVRS ACPI tables.
TEST=Timeless build of amd/gardenia results in identical binary.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I410a6560de66889b153c8a66b8dc5474ac114ba7
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55429
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Enable generation of DMI Type 17 data on Cezanne.
BUG=b:184124605
TEST="dmidecode --type 17" in OS on Majolica
Change-Id: Iaa89ee1ce6efa0280f17a443e07571a1190873a6
Signed-off-by: Nikolai Vyssotski <nikolai.vyssotski@amd.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/52230
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Move dmi.c code to common/fsp to be shared among different SOCs.
BUG=b:184124605
Change-Id: I46071556bbbbf6435d9e3724bba19e102bd02535
Signed-off-by: Nikolai Vyssotski <nikolai.vyssotski@amd.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/52746
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
if VBOOT_STARTS_BEFORE_BOOTBLOCK is set, call boot_with_psp_timestamp to
migrate PSP timestamps into x86 timestamp table.
Signed-off-by: Kangheui Won <khwon@chromium.org>
Change-Id: I4d51802145263145d40908889de29147af54f50f
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55405
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
The warm reset bit in the NCP_ERR register doesn't behave as the PPR [1]
suggested; no matter if something was written to the register, the
NCP_WARM_BOOT bit never got set and the NCP_ERR register in I/O-space
always reads back as 0x7f.
[1] checked with PPR for AMD Family 19h Model 51h A1 (CZN) #56569 Rev
3.01
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I569372db9f36ec7bbc741f4d7312ade312daa70b
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55101
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
On Linux, in order to set wake alarms >24 hours, the RTC Date Alarm
field must be set to a valid non-zero value. If not, there are two
consequences:
1. Alarms >24 hours don't work
2. The kernel will refuse to enter suspend because it can't resume as
expected to service the alarm.
Since the RTC Date Alarm and RTC AltCentury fields are supported on
Stoneyridge, set them.
This is a mirror of commit 041fcf5902
("soc/amd/picasso/acpi: Set missing RTC offsets") for picasso.
BUG=b:187516317
TEST=On a Chrome OS 'grunt' device, run
`time powerd_dbus_suspend --suspend_for_sec=172800`
and verify the system suspended and woke up after 48 hours
BRANCH=grunt
Signed-off-by: Anand K Mistry <amistry@google.com>
Change-Id: I10831b982662e680fa71aa81d02935e1b7e7a7a1
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55001
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Note that there are assumptions about LAPIC MMIO location
in both AMD and Intel sources in coreboot proper.
Change-Id: I2c668f5f9b93d170351c00d77d003c230900e0b4
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55194
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This also adds BERT table gerenation support for Cezanne, but since the
functionality to populate the BERT memory region isn't implemented yet,
this won't result in a BERT table being generated on Cezanne, since
bert_generate_ssdt will always return false there.
TEST=BERT ACPI table generation still works on AMD/Mandolin
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I69b4a9a7432041e1f4902436fa4e6dee5332dbd2
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55056
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Implement acpi_soc_get_bert_region so that the common ACPI code will
generate a BERT ACPI table that points to the BERT memory region instead
of generating the BERT table in the SoC=specific code.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I86d4f5ef74d4d40cb93ac4a3feaf28b99022ebd5
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55055
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Since bert_errors_present() is only available when ACPI_BERT is selected
the ACPI table generation code needs to check that before calling the
function, so add bert_should_generate_acpi_table that returns false when
ACPI_BERT isn't selected or the return value of bert_errors_present()
when ACPI_BERT is selected.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ia955f627c190ea38e05b5aaedc7cb2d030274e83
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55024
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
EDK2 mandates HOB to be in increments of qword (8). This HOB has 13
elements which causes it be padded with 4 bytes of garbage. This
results in coreboot failing intermittently with invalid data. Add
"number of entries" field to specify the number of valid entries in
the table.
BUG=b:190153208
Cq-depend: chrome-internal:3889619
TEST=verify HOB is present and correct size (13) is reported
Change-Id: Iaafae304f04a5f26d75a41a6d6fcb4ee69954d20
Signed-off-by: Nikolai Vyssotski <nikolai.vyssotski@amd.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55237
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Some of the I2C buses are required to operate at different voltage level
compared to other I2C buses eg. I2C bus to Google Security Chip (GSC)
should be at 1.8V level. By default, all the I2C buses are initialized
to operate at 3.3 V. Add support to configure I2C pad RX select through
devicetree and update the concerned devicetree.
BUG=b:188538373
TEST=Build and boot to OS in Guybrush. Ensure that the communication
with GSC is fine. Build Majolica mainboard.
Change-Id: I595a64736fdac0274abffb68c5e521302275b845
Signed-off-by: Karthikeyan Ramasubramanian <kramasub@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55149
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
The new cezanne PSP release added support for these svcs. So add those
functionality back to cezanne psp_verstage.
BUG=b:187906425
Signed-off-by: Kangheui Won <khwon@chromium.org>
Change-Id: Id542f5ed0762f582ea966466d67ed938ecb9c1f0
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55137
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
GSC is connected with AP via i2c bus so we need to enable i2c in
psp_verstage.
Signed-off-by: Kangheui Won <khwon@chromium.org>
Change-Id: I5f7b73be67a692ea7de31ae53bd111d0e4b6998c
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55136
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Since the MCA(X) registers have defined values on the cold boot path,
the is_warm_reset check can be dropped. Also the warm reset bit in the
NCP_ERR register doesn't behave as the PPR [1] suggested; no matter if
something was written to the register or the machine went through a warm
reset cycle, the NCP_WARM_BOOT bit never got set.
[1] checked with PPR for AMD Family 17h Models 11h,18h B1 (RV,PCO)
#55570 Rev 3.15
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I4e6df98ffd5d15ca204c9847a76c19c753726737
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55059
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
The chipset devicetree only has the essential PCIe devices enabled that
are needed for the SoC code to work. It also defines aliases for all
PCIe devices that can be used to reference the devices in the mainboard-
specific devicetrees and devicetree overrides. To make the change easier
to review that part will be done in a follow-up patch.
Despite missing in the PPR, device pci 18.7 exists on Picasso.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I6b7c3fd32579a23539594672593a243172c161c7
Reviewed-on: https://review.coreboot.org/c/coreboot/+/50626
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
The Guybrush platform needs to set up some GPIOs immediately before the
FSP-M runs. Add a platform specific call. This will be used in a
follow-on commit.
BUG=b:184796302, b:184598323
TEST=Build
Signed-off-by: Martin Roth <martinroth@chromium.org>
Change-Id: I37d2625ff426347852e98a9a50f15368e0213449
Reviewed-on: https://review.coreboot.org/c/coreboot/+/54638
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
PLTRST# is currently asserted and latched when eSPI_RST# gets asserted.
If eSPI_RST# isn't used on a platform or it doesn't properly assert
in all cases, then PLTRST# will never be asserted. This could result in
the AP and EC being out of sync.
BUG=b:188188172, b:188935533
TEST=Warm reset guybrush with partial #22 rework. Verify that peripheral
channel is correctly reset.
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I20d12edf3efc6100096e24aa8d1aec76bbde264f
Reviewed-on: https://review.coreboot.org/c/coreboot/+/54884
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Rob Barnes <robbarnes@google.com>
On RW boot path psp_verstage call cbfs_map which calls chain of
_cbfs_alloc, cbfs_boot_lookup and cbfs_get_boot_device. Then
cbfs_get_boot_device initializes MCACHE which is used later.
However on RO boot path psp_verstage doesn't try to find anything in the
CBFS which results RO MCACHE not to be initialized. Add
cbfs_get_boot_device(true) to explicitly initialize MCACHE on recovery
boot.
BUG=b:177091575
BRANCH=none
TEST=build and boot jelboz
Signed-off-by: Kangheui Won <khwon@chromium.org>
Change-Id: I6c4b522fef5a4affd215faa122bdf6b53190cf3d
Reviewed-on: https://review.coreboot.org/c/coreboot/+/54711
Reviewed-by: Julius Werner <jwerner@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
We were missing the stdint.h header, and the header was sorted
incorrectly in chip.h
BUG=non
TEST=build guybrush
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I209d3c9c48e5b06b2a56759af51cf2858eb99f51
Reviewed-on: https://review.coreboot.org/c/coreboot/+/54922
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Since the default for the corresponding UPD of the Picasso FSP is
DXIO_PSPP_POWERSAVE and the devicetree default is DXIO_PSPP_PERFORMANCE,
add a deviectree setting for each board that's using the Picasso SoC
code to not change the setting for the existing boards.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I0008ebb0c0f339ed3bdf24ab95a20aa83d5be2c9
Reviewed-on: https://review.coreboot.org/c/coreboot/+/54934
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
This allows boards to specify which PSPP policy (basically a dynamic
trade-off between power consumption and PCIe link speed) should be used
and also makes sure that the boards are using the expected PSPP policy
and not just the UPD default from the FSP binary that has already
changed once during the development.
BUG=b:188793754
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I1b6459b2984711e72b79f5d4d90e04cb4b78d512
Reviewed-on: https://review.coreboot.org/c/coreboot/+/54930
Reviewed-by: Matt Papageorge <matthewpapa07@gmail.com>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This patch also adds the additional 10 MCAX registers to the BERT MSR
error record.
BUG=b:186038401
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I31912d3b3e77e905f64b6143042f5e7f73db7407
Reviewed-on: https://review.coreboot.org/c/coreboot/+/52616
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
The AMD FSP is using a new structure for USB and USB C phy settings.
This patch removes old, unused structures, adds the new one and
enables the devicetree interface for it.
Signed-off-by: Julian Schroeder <julianmarcusschroeder@gmail.com>
Change-Id: I011ca40a334e4fd26778ca7f18b653298b14019b
Reviewed-on: https://review.coreboot.org/c/coreboot/+/54065
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-by: Rob Barnes <robbarnes@google.com>
ESPI_SLAVE_CHANNEL_READY is a read-only bit from the host perspective.
It is set when the eSPI peripheral has configured the channel.
We actually want to set the ESPI_SLAVE_CHANNEL_ENABLE flag. This never
caused an issue before because the peripheral channel is enabled by
default after PLTRST# is deasserted. This does fix the case where
periph_ch_en == 0. It now properly clears the enable flag.
BUG=b:188188172, b:188935533
TEST=Boot guybrush to OS, perform warm reset
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I24e0734d5652601ae9c967da528fec5e3f780991
Reviewed-on: https://review.coreboot.org/c/coreboot/+/54883
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Rob Barnes <robbarnes@google.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
The ChromeEC might take longer than 1ms for the peripheral channel to be
enabled. The PLTRST# interrupt handler takes about ~539us.
This doesn't account for the time it takes for the interrupt handler
to be scheduled. Increasing the timeout to 10ms gives ample time.
BUG=b:188188172, b:188935533
TEST=Boot guybrush and no longer see channel enable errors
Suggested-by: Rob Barnes <robbarnes@google.com>
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: Ib6db577bf06175ceb17b446af706ad8c9f891481
Reviewed-on: https://review.coreboot.org/c/coreboot/+/54788
Reviewed-by: Jason Glenesk <jason.glenesk@amd.corp-partner.google.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The default of CBFS_MCACHE_SIZE is increased to 0x4000 in CB:54146 but
we have limited space on the PSP thus cannot afford it.
BUG=b:177091575
BRANCH=none
Signed-off-by: Kangheui Won <khwon@chrmoium.org>
Change-Id: I94dd782ae00d0b18ad6dd2fc061e4318bda88579
Reviewed-on: https://review.coreboot.org/c/coreboot/+/54710
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
NCP_ERR is a 1 byte register in I/O-space, so use inb and not inw. The
variable the result gets assigned to is also a uint8_t.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I9fd8c139004111d6227c0316ba2a8b0281541654
Reviewed-on: https://review.coreboot.org/c/coreboot/+/54736
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
commit ce0e2a0140 (drivers/intel/fsp2_0:
use FSP to allocate APEI BERT memory region) adds a mechanism to reserve
the BERT region inside the coreboot code, so we can get rid of the
workaround to reserve it in the FSP and return the location in a HOB.
mcfg->bert_size defaults to 0 which makes the FSP not generate the
corresponding HOB, but that field is planned to be removed at least on
Cezanne, so don't explicitly set it to 0.
BUG=b:169934025
TEST=BERT table that gets generated in a follow-up patch for Picasso
points to expected BERT region and Linux is able to access, decode and
display it.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Iaca89b47793bf9982181560f026459a18e7db134
Reviewed-on: https://review.coreboot.org/c/coreboot/+/52584
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
When performing an in-band reset the host controller and the
peripheral can have mismatched IO configs.
i.e., The eSPI peripheral can be in IO-4 mode while, the
eSPI host will be in IO-1. This results in the peripheral
getting invalid packets and thus not responding. This causes the
NO_RESPONSE status bit to be set and cause eSPI init to fail.
If the peripheral is alerting when we perform an in-band
reset, there is a race condition in espi_send_command.
1) espi_send_command clears the interrupt status.
2) eSPI host controller hardware notices the alert and sends
a GET_STATUS.
3) espi_send_command writes the in-band reset command.
4) eSPI hardware enqueues the in-band reset until GET_STATUS
is complete.
5) GET_STATUS fails with NO_RESPONSE and sets the interrupt
status.
6) eSPI hardware performs in-band reset.
7) espi_send_command checks the status and sees a
NO_RESPONSE bit.
As a workaround we allow the NO_RESPONSE status code when
we perform an in-band reset.
BUG=b:186135022
TEST=suspend_stress_test and S5->S0 tests on guybrush and zork.
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I71271377f20eaf29032214be98794e1645d9b70a
Reviewed-on: https://review.coreboot.org/c/coreboot/+/54070
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-by: Rob Barnes <robbarnes@google.com>
I'm not 100% sure if this should rather be duplicated from Picasso or
commonized. Checked with the docs and this won't be compatible with
Stoneyridge and one future product's PPR lacked the corresponding
register. Some other chip has a compatible register layout, but a
different number of PCIe GPP clock outputs, so the common code would
need to use some SoC-dependent defines and possibly a SoC-specific
lookup table for the mapping which is also not that great.
TEST=Checked Cezanne PPR
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I6b6d0cb8d7eb0288d8a18fcb975dc377b2c6846a
Reviewed-on: https://review.coreboot.org/c/coreboot/+/54685
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
