On tgl, we noticed system hang if a shutdown is triggered before fsps.
The dut is unable to shutdown properly due to tcss is stuck before
tcss_init in fsps.
This change enable power button smi on jsl, tgl, adl after fsps.
it can also prevent a shutdown failure due to lack of fsps init on
certain ip.
BUG=b:186194102, b:186815114
TEST=Power on the system and pressing power button repeatedly doesn't
cause the system hang during shutdown.
Change-Id: I70b871f2676a89bc782116e02beba5c20ec51eef
Signed-off-by: Kane Chen <kane.chen@intel.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/52874
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
If a power button SMI is triggered between where it is currently
enabled and before FSP-S exits, when the SMI handler disables
bus mastering for all devices, it inadvertently also disables
the PMC's I/O decoding, so the register write to actually go into
S5 does not succeed, and the system hangs.
This can be solved by skipping the PMC when disabling bus
mastering in the SMI handler, for which a callback,
smihandler_soc_disable_busmaster is provided.
BUG=b:186194102, b:186815114
TEST=Power on the system and pressing power button repeatedly doesn't
cause the system hang during shutdown.
Change-Id: I1cf5cf91ebad4a49df6679e01fc88ff60c81526c
Signed-off-by: Kane Chen <kane.chen@intel.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/52873
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Adding GPIO definition for community 3 which is CPU reserved GPIO used
by CPU side PCIe root ports. We did not have this definition since
FSP used to program this GPIOs. Now, instead of FSP, coreboot programs
CPU PCIe GPIOs for CLKSRC and lanes to put GPIOs in native mode.
Thus adding definition of this virtual GPIOs in this CL.
BUG=None
BRANCH=None
TEST=Check if correct registers are being programmed
Change-Id: I481ea7e3ba948bf6d37b97d08c675a18ee68125d
Signed-off-by: Maulik V Vaghela <maulik.v.vaghela@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/52783
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
The usage of `pci_devfn_t` here is misleading, as these intentionally
store the `PCH_DEVFN_*` macros so they can be used across `smm` and
`ramstage` without requiring the device model. Update to `unsigned int`
instead, as `pci_devfn_t` implies the data is an MMCONF-compatible PCI
devfn offset.
Change-Id: Ic8880de984e6eceda4cbe141e118f3a5fdd672a2
Signed-off-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/52808
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
With the recent switch to SMM module loader v2, the size of the SMM for
module google/volteer increased to above 64K in size, and thus failed to
install the permanent SMM handler. Turns out, the devicetree is all
pulled into the SMM build because of elog, which calls
`pci_dev_is_wake_source`, and is the only user of `struct device` in
SMM. Changing this function to take a pci_devfn_t instead allows the
linker to remove almost the entire devicetree from SMM (only usage left
is when disabling HECI via SMM).
BUG=b:186661594
TEST=Verify loaded program size of `smm.elf` for google/volteer is
almost ~50% smaller.
Signed-off-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Change-Id: I4c39e5188321c8711d6479b15065e5aaedad8f38
Reviewed-on: https://review.coreboot.org/c/coreboot/+/52765
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Kconfig change which enables the hwp cppc acpi support is to get the
maximum performance of each CPU to check and enable Intel Turbo Boost
Max Technology.
BUG=none
BRANCH=none
TEST=check GCPC and CPC generated in acpi tables for each CPU
Change-Id: I5d93774e8025466f1911cf77459910fe872bfcc8
Signed-off-by: ravindr1 <ravindra@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/51795
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The VT-d specification states that device scope for remapping hardware
unit which has DRHD_INCLUDE_PCI_ALL flags must be the last in the list
of hardware unit definition structure. This change fixes the devices
list in the DMAR DRHD structure.
BUG=None
TEST=Built image successfully.
Signed-off-by: John Zhao <john.zhao@intel.com>
Change-Id: I14c34ad66a5ee8c30acabd8fe5a05c22087f9120
Reviewed-on: https://review.coreboot.org/c/coreboot/+/52477
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
FSP uses PchHdaAudioLink{Hda|Dmic|Ssp|Sndw}Enable UPDs to configure
GPIO pads for audio. However, mainboard is expected to perform all
GPIO configration in coreboot and hence these UPDs must be set to
0. There is no need to expose these UPDs in chip.h and provide
mainboard an option to set these in devicetree.
This change drops PchHdaAudioLink{Hda|Dmic|Ssp|Sndw}Enable UPDs from
chip.h and the corresponding devicetree in mainboards. Currently,
shadowmountain already set these UPDs to 0, whereas adlrvp set these
to 1. But all the ADL boards are correctly configuring the GPIO pads
for audio, so this change should not impact audio for any of these
boards.
BUG=b:183482000
TEST=adlrvp and shadowmountain build successfully.
Change-Id: I90e4eb5cc242a789800f4c9f8c71e9d8c8a2becf
Signed-off-by: Furquan Shaikh <furquan@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/52559
Reviewed-by: EricR Lai <ericr_lai@compal.corp-partner.google.com>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
We need to change OC pin for type C USB3 ports and it depends
on the board design. Allowing it to be filled by devicetree will
make it easier to change the mapping based on the board design
BUG=b:184653645
BRANCH=None
TEST=compilation works fine and value of UPD is getting reflected.
Change-Id: I61faa661c12dced27c6cdd7005a61ae8de8621e1
Signed-off-by: Maulik V Vaghela <maulik.v.vaghela@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/52320
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Reviewed-by: EricR Lai <ericr_lai@compal.corp-partner.google.com>
Each TCSS DMA is grouped together with two PCIe RPs in terms of PM flow.
This change ensures that SD3C is updated for the TCSS DMA devices
corresponding to the TBT RP ports. If TBT port is 0 or 1, SD3C for DMA0
is updated, else for DMA1.
BUG=None
TEST=Built Alderlake image successfully.
Signed-off-by: John Zhao <john.zhao@intel.com>
Change-Id: Ia3462dfbb287a374960a57bb4c3541db2a435611
Reviewed-on: https://review.coreboot.org/c/coreboot/+/51965
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Lists of changes:
1. Rename MISCCFG_ENABLE_GPIO_PM_CONFIG -> MISCCFG_GPIO_PM_CONFIG_BITS
2. Move MISCCFG_GPIO_PM_CONFIG_BITS definition from intelblock/gpio.h to
soc/gpio.h. Refer to detailed description below to understand the
motivation behind this change.
An advanced GPIO PM capabilities has been introduced since CNP PCH,
refer to 'include/intelblock/gpio.h' for detailed GPIO PM bit definitions.
Now with TGP PCH, additional bits are defined in the MISCCFG register
for GPIO PM control. This results in different SoCs supporting
different number of bits. The bits defined in earlier platforms
(CNL, CML, ICL) are present on TGL, JSL and ADL too. Hence, refactor the
common GPIO code to keep the bit definitions in intelblock/gpio.h, but
the definition of MISCCFG_GPIO_PM_CONFIG_BITS is moved to soc/gpio.h so
that each SoC can provide this as per hardware support.
TEST=On ADL, TGL and JSL platform.
Without this CL :
GPIO COMM 0 MISCCFG:0xC0 (Bit 6 and 7 enable)
With this CL :
GPIO COMM 0 MISCCFG: 0x00 (Bit 6 and 7 disable)
Change-Id: Ie027cbd7b99b39752941384339a34f8995c10c94
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/51767
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
List of changes:
1. Update GPIO Group to GPE DWx assignment encoding as per MISCCFG
register per GPIO Community.
2. PMC_GPP_* macros are also updated as per GPIO_CFG register
in PMC space.
BUG=b:183464235
TEST=Able to fix the TPM IRQ issue on SM.
Change-Id: Id9f57b0b5726315f5ebba013f11d52ed3ee34484
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/51789
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Add function to allow overriding the RcompResistor and
RcompTarget UPDs from mainboard if required.
Mainboard users can pass required rcomp from memory.c file.
Refactor ddr_config structure to take out rcomp related variable
outside for all memory type to override if required.
BUG=b:182772421
TEST=Able to override the default RcompResistor and RcompTarget
values.
Change-Id: Ie8528bbf0517728534d47f9adaabfc9a2c469609
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/51683
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
List of changes:
1. Alder Lake MRC is expecting a RcompResistor value of word width.
Reference RCOMP resistors on motherboard are ~ 100 Ohms but coreboot
is passing an array of RcompResistor which is not completely in use.
Note: Rcomp resistor value represents rcomp resistor attached to
the DDR_COMP pins on the SoC.
2. Also, remove usage of '&' with memcpy the required value into
RcompTarget array.
3. Also, update RcompResistor value for ADLRVP.
BUG=b:183341229
TEST=Enable FSP debug log to verify the override value for
RcompResistor is reflecting correctly.
Change-Id: I69c7cec55b65036fc039c33374a3fd363ef7004e
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/51704
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Drop the 100ms delay in the _PS0 method because kernel already adds this
100ms. This change also drops polling TBT PCIe root ports Link Active
State because this scheme is not applicable for SW CM.
BUG=None
TEST=Built Alderlake coreboot image successfully.
Signed-off-by: John Zhao <john.zhao@intel.com>
Change-Id: I792d3c8ca4249ed74d4090ec1efba5a180429c75
Reviewed-on: https://review.coreboot.org/c/coreboot/+/51191
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The Intel ADL BIOS specification #627270 recommends reserving the
following resources for each PCIe TBT root port:
- 42 buses
- 192 MiB Non-prefetchable memory
- 448 MiB Prefetchable memory
Add a mainboard Kconfig which will auto-select these recommended values,
in addition to PCIEXP_HOTPLUG.
Signed-off-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Change-Id: Icdfa2688d69c2db0f98d0523d5aba42eec1824db
Reviewed-on: https://review.coreboot.org/c/coreboot/+/51460
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: EricR Lai <ericr_lai@compal.corp-partner.google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
UART pad configuration should not be done in common code, because that
may cause short circuits, when the user sets a wrong UART index.
Since all boards do pad setup on their own now, finally drop the pad
configuration from SoC common code.
Change-Id: Id03719eb8bd0414083148471ed05dea62a895126
Signed-off-by: Michael Niewöhner <foss@mniewoehner.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/48829
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Lance Zhao
The default value for the LidStatus is "LidClosed" mean 0
Because of this GOP skips graphics initialization assuming
lid is closed even though lid is open. This Patch is to set
LidStatus UPD to 1 whenever RUN_FSP_GOP config is selected.
BUG=b:178461282
BRANCH=None
TEST=Build and boot ADLRVP and verify eDP is coming up in
depthcharge
Change-Id: I1648ae0f06e414b2a686e325acf803deb702b7a5
Signed-off-by: Ronak Kanabar <ronak.kanabar@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/51131
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Reviewed-by: EricR Lai <ericr_lai@compal.corp-partner.google.com>
The `FORCE_ENABLE` and `FORCE_DISABLE` names do not match what FSP UPDs
say, and can be confused with the `PchHdaTestPowerClockGating` UPD.
Replace the enum with a bool, and drop the confusing names. Note that
the enum for Ice Lake was incorrect, but no mainboards used the option.
Change-Id: I2c9b4c6a2f210ffca946ca196299fa672a06ccc7
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/51154
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
