Add CFR option "touchpad_wake" (default: false) to allow users to
enable touchpad wake from sleep. Wake is disabled by default to
prevent random wakeups when systems are moved while sleeping.
Implementation:
- Add touchpad_wake CFR option to System form
- Add device alias "touchpad" to variant devicetree files
(add touchpad2 where a 2nd touchpad option is present)
- Update variant GPIO configs to support wake capability
(this was set inconsistently)
- Conditionally disable wake in ramstage based on CFR option
When the option is disabled (default), config->wake and
config->irq.wake are set to 0, preventing the touchpad ACPI device
from defining wake methods and capability.
TEST_1: Build for ELDRID, boot system into mainlinue Linux (Fedora).
Close the lid to suspend the system, firmly grab it with both hands and
jump 10 times. Make sure that system doesn't wake up from sleep state.
TEST_2: Build for DROBIT, boot system into mainline Linux (NixOS).
Close the lid to suspend the system, firmly grab it with both hands and
jump few times. Make sure that system doesn't wake up from sleep state.
Tested-by: Ingo Reitz <9l@9lo.re>
Change-Id: Ie3b5013bcf2d5ea45388bcdce987dd9ae5870597
Co-authored-by: Alicja Michalska <alicja.michalska@9elements.com>
Signed-off-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90500
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Alicja Michalska <ahplka19@gmail.com>
Add CFR option "touchpad_wake" (default: false) to allow users to
enable touchpad wake from sleep. Wake is disabled by default to
prevent random wakeups when systems are moved while sleeping.
Implementation:
- Add touchpad_wake CFR option to System form
- Add device alias "touchpad" to variant devicetree files
- Conditionally disable wake in ramstage based on CFR option
When the option is disabled (default), config->wake and
config->irq.wake are set to 0, preventing the touchpad ACPI device
from defining wake methods and capability.
TEST=Build for KOHAKU, boot into mainline Linux (Fedora 43, kernel
6.18). Suspend the machine, close the lid, apply pressure to trackpad
area. Ensure that system doesn't wake up.
Change-Id: I620d7c0f0ebec8ca72ec018c93747f58735fd3e5
Co-authored-by: Alicja Michalska <alicja.michalska@9elements.com>
Signed-off-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90491
Reviewed-by: Alicja Michalska <ahplka19@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The STATESTS register is a very simple read/write-1-clear status
register. OS drivers have to read and clear it all the time with-
out any quirk handling. So it seems unlikely that this sequence
was ever necessary for any coreboot-supported chip.
More likely, that sequence was copied from the dance around the
reset bit when Poulsbo support was added in commit be61a17351
("Support Intel SCH (Poulsbo) and add iwave/iWRainbowG6 board
which uses it.").
TEST= Verbs were loaded correctly on off-tree HP ProBook 450 G3.
Change-Id: I1fbea8ffb71a2fcb4ce5f42b3cb8f816ec336c5b
Signed-off-by: Nico Huber <nico.h@gmx.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/89653
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The azalia_set_bits() function does some sequencing specific to the
CRST# bit. Name it accordingly and adapt its signature.
The while loop was also changed to match the other functions (e.g.
wait_for_ready()).
TEST= Verbs were loaded correctly on off-tree HP ProBook 450 G3.
Change-Id: I5ac766f2ddf8b48b436a54469815bf799ae31d52
Signed-off-by: Nico Huber <nico.h@gmx.de>
Signed-off-by: Nicholas Sudsgaard <devel+coreboot@nsudsgaard.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/89652
Reviewed-by: Alicja Michalska <ahplka19@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The spec[1] says a codec is allowed to take up to 521us before sig-
naling an initialization request. Our original SB600 implementation
had a 1ms delay here since commit 4505948fae ("Use the correct device
for switching on HDA.")
Most codecs are a lot faster, which is probably why nobody noticed
the missing delay. For instance, the Realtek ALC272 datasheet spe-
cifies a 1 frame (1/48kHz) maximum[2]. It doesn't hurt, though, to
be correct here. We have a lot longer delays around.
[1] High Definition Audio Specification 1.0a: "4.3 Codec Discovery"
[2] ALC272 datasheet: "9.2.1. Link Reset and Initialization Timing"
TEST= Verbs were loaded correctly on off-tree HP ProBook 450 G3.
Change-Id: Ifd3357758fb3678e60b4c6edcfbdb60b3bda9746
Signed-off-by: Nico Huber <nico.h@gmx.de>
Signed-off-by: Nicholas Sudsgaard <devel+coreboot@nsudsgaard.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/89651
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Alicja Michalska <ahplka19@gmail.com>
This adds the release notes template for the upcoming March 2026
release of coreboot.
Change-Id: I7702ee6b5dffce067503ff6e4ccb6508d30bf48f
Signed-off-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90498
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Alicja Michalska <ahplka19@gmail.com>
These are the preliminary release notes. They'll need to be updated
with any changes done this week. We'll need another patch after the tag
to capture the final statistics. The notes will be changed from
"Upcoming release" after the tag is done.
Change-Id: Ic8c5a4f374adc6560eff8383d45d05d16ba84759
Signed-off-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90497
Reviewed-by: Felix Singer <service+coreboot-gerrit@felixsinger.de>
Reviewed-by: Alicja Michalska <ahplka19@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
non-cacheable
Add support to split the dram_aop memory region into three in order to
map dram_aop_cmd_db as non-cacheable. The purpose of dram_aop_config
is memory region where the aop_devcfg.mbn image is loaded.
Test=1. Build and boot on X1P42100.
2. Dump the MMU table in coreboot ramstage and verify whether the region
is mapped as non-cacheable.
Change-Id: Id73d878b7d343f248a845bd5727c43e22c4c348a
Signed-off-by: Swathi Tamilselvan <tswathi@qualcomm.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90521
Reviewed-by: Subrata Banik <subratabanik@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Add support to map AOP CMD-DB region as uncached region in MMU. The
reason for this change is that the CMD-DB region is configured as
read-only region and any write to this region will be treated as
fatal. Mapping it as cacheable can lead to cache-line writebacks,
causing invalid accesses and device crashes.
Test=1. Create an image.serial.bin and ensure it boots on X1P42100.
MMU Table dump from Trace32:
'M:0000000081C60000--0000000081C7FFFF| AM:0000000081C60000--0000000081C7FFFF| s | | 00001000| read/write access exec | yes| inn| non-cacheable | 1| no | \\ramstage\Global\_dram_aop_cmd_db'
Change-Id: I296b505f670f3be28eb998fdac8164a85bf757b0
Signed-off-by: Swathi Tamilselvan <tswathi@qualcomm.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90464
Reviewed-by: Subrata Banik <subratabanik@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
On WCL RVP, ISH (Intel Sensor Hub) shares UART with FPS (Finger Print
Sensor), we can enable either ISH UART or FPS UART, or disable both
UART by changing the DIP switch settings. When DIP switch is not set
for ISH, ISH RX signal is disconnected, causing ISH low power mode
failure. Therefore, NC ISH RX pin mux to minimize the impact on ISH PM.
As a result, ISH console won't accept input since this pin is not
connected.
BUG=b:428084925
TEST=DIP SW1317 3-6, 4-5 on WCL RVP DT card ON to enable FPS UART,
ISH main firmware boots up and runs successfully.
SW1317 all switches OFF to disable both FPS and ISH UART, ISH main
firmware boots up and runs successfully.
Put system into suspend state and resume back. System enters into
low power mode as substate_residencies are updated. Verified using
console command "cat /sys/kernel/debug/pmc_core/substate_residencies".
Change-Id: I1165e0151c0a6d0e82038215703626e863739b39
Signed-off-by: Megha Verma <megha.verma@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90022
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Bhat D, Krishna P <krishna.p.bhat.d@intel.com>
Reviewed-by: P, Usha <usha.p@intel.com>
The mipi_cmd_func_t callback for mipi_panel_parse_init_commands()
currently doesn't support passing additional data for storing context.
Therefore user code would need to store any extra data in global
variables. For example, in the upcoming DSI dual channel support for
MediaTek platforms, the callback needs to know whether the MIPI panel
supports dual channel or not. To support that use case, pass an extra
`user_data` argument to mipi_cmd_func_t.
BUG=b:424782827
TEST=util/abuild/abuild -x -t GOOGLE_SAPPHIRE -a
BRANCH=none
Change-Id: Id5d7b168cdcadfe8d8435c29d7e855a535815057
Signed-off-by: Yu-Ping Wu <yupingso@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90519
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Subrata Banik <subratabanik@google.com>
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
Reviewed-by: Yidi Lin <yidilin@google.com>
Update off-mode detection logic to use ChromeEC host events. If
the system is powered on by AC insertion (without power button
or lid open events), identify the boot as off-mode charging.
BUG=b:457566143
TEST=Verify off-mode battery charging on Google/Quenbi.
Change-Id: Ibfbbf9cbeabd229595f625104f94eb814012e2f8
Signed-off-by: Kapil Porwal <kapilporwal@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90511
Reviewed-by: Subrata Banik <subratabanik@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
pcie_rp5 and pcie_rp6 belong to the same PCIe controller group
(Port C1–C2). Within this controller, pcie_rp5 represents function 0
and pcie_rp6 represents function 1. For multi-function PCIe root
ports, function 0 must be enabled for the controller to initialize
fully.
If only pcie_rp6 (function 1) is enabled, the controller does not
complete initialization and the downstream LAN device fails to
enumerate. Enabling pcie_rp5 ensures the PCIe controller group is
brought up correctly and allows the RTL8111H LAN device behind
pcie_rp6 to enumerate as expected.
BUG=b:466908212
TEST=Build and boot to OS in kodkod.
kodkod:/ # pclspci -v
01:00.0 Class 0200: Device 10ec:8168 (rev 15)
DeviceName: Ethernet controller
Kernel driver in use: r8169
Change-Id: I4332f3d612f1f66cd30dda7da723c47bcfce35a3
Signed-off-by: Ian Feng <ian_feng@compal.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90422
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Pranava Y N <pranavayn@google.com>
Define the QCLIB_GA_ENABLE_PD_NEGOTIATION bit in the QcLib global
attributes. This flag signals to QcLib that Power Delivery
negotiation should be performed.
BUG=b:457566143
TEST=Verify different boot modes on Google/Quenbi.
TEST=Verify that PD negotiation is skipped in normal mode.
Change-Id: Ia046f68ebeacaa1c1d9a73c4b957315c9f7b68b6
Signed-off-by: Kapil Porwal <kapilporwal@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90512
Reviewed-by: Subrata Banik <subratabanik@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
If the section guid is CPER_SEC_PROC_IA32X64_GUID we want the x86
processor specific section instead of the generic one.
This was probably some kind copy error from the line above.
Change-Id: I6a6a885bf8ab97cb5d256513cf8134078b707d3c
Signed-off-by: Maximilian Brune <maximilian.brune@9elements.com>
Signed-off-by: Patrick Rudolph <patrick.rudolph@amd.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90476
Reviewed-by: Matt DeVillier <matt.devillier@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-by: Jérémy Compostella <jeremy.compostella@intel.com>
Introduce a weak function qclib_mainboard_override to allow
mainboards to customize QcLib policies or global attributes.
This hook is called from the SoC-specific QcLib initialization
path.
BUG=b:457566143
TEST=Configure QcLib policy for Google/Quenbi.
Change-Id: I0397b7138db260973ea86852cfa9f408e14d195d
Signed-off-by: Kapil Porwal <kapilporwal@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90510
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Subrata Banik <subratabanik@google.com>
Ensure that existing flags in global_attributes are not overwritten
when enabling UART logging. Using a bitwise OR preserves any
previously set attributes.
BUG=b:457566143
TEST=Verify the QcLib global attributes.
With this CL -
```
[DEBUG] Global Attributes[0x3]..Table Entries Count[8]
```
Change-Id: Iffceb06cb800ba4c9e5c07381745cbed1fb7d550
Signed-off-by: Kapil Porwal <kapilporwal@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90509
Reviewed-by: Subrata Banik <subratabanik@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This commit updates the SSRAM memory layout in memlayout.ld to reserve
space for several new regions.
The following regions are added to the SSRAM section:
- qsee: 100K at 0x14680000
- qdss_usb_trace: 8K at 0x146A6000
- aop_imem: 8K at 0x146A8000
The memory map diagram in the comments is also updated to reflect the
reservation of aop_imem and qdss_usb_trace.
BUG=b:456953373
TEST=Able to build and boot google/quenbih.
Change-Id: I17c2a97d31cdcb81ffdd0c83d8c6d19b9a03a91b
Signed-off-by: Subrata Banik <subratabanik@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90443
Reviewed-by: Kapil Porwal <kapilporwal@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Jayvik Desai <jayvik@google.com>
Relocate the PRERAM_STACK region to a new address to resolve a memory
conflict with the QSEE Trust Zone environment.
Details:
- The previous location of the PRERAM_STACK starting at 0x14680000 in
SSRAM is now reserved for QC QSEE.
- This change moves the 16KB PRERAM_STACK from 0x14680000 (SSRAM) to
the available memory region starting at 0x14850000 in the BSRAM.
BUG=b:456953373
BRANCH=None
TEST=Able to build google/bluey.
Change-Id: Ifa9bc457e28b9ec21aa59c6ab9114993f23b2bc8
Signed-off-by: Subrata Banik <subratabanik@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90406
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Jayvik Desai <jayvik@google.com>
Reviewed-by: Kapil Porwal <kapilporwal@google.com>
Implement the newly introduced PRERAM_STACK and POSTRAM_STACK macros in
the x1p42100 memory layout, addressing the memory constraints on this
SoC.
Changes
- Pre-RAM Stack: The temporary stack used before DRAM is ready is
defined using PRERAM_STACK(0x14680000, 16K) in the SSRAM region,
replacing the old generic STACK definition.
- Post-RAM Stack: The final stack is defined using
POSTRAM_STACK(0x80000000, 16K) at the start of DRAM.
- The POSTRAM_DMA_COHERENT region is shifted up to 0x80004000 to
accommodate the new 16KB post-RAM stack, avoiding memory overlap.
This guarantees a distinct and properly sized stack region for each
stage of the boot process, resolving conflicts with Trust Zone and
ensuring a clean stack transition during the RAM stage.
BUG=b:456953373
BRANCH=None
TEST=Able to build google/bluey.
w/o this patch
```
[SPEW ] stack: top_of_stack address is 0x14684000
```
w/ this patch
```
[SPEW ] stack: top_of_stack address is 0x80004000
```
Change-Id: Iccf3f99aff31a8e44386ea52b2196b49797caa79
Signed-off-by: Subrata Banik <subratabanik@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90405
Reviewed-by: Jayvik Desai <jayvik@google.com>
Reviewed-by: Kapil Porwal <kapilporwal@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This change prepares an alternative entry point for the ARM64
ramstage. It is written in assembly language, avoids the usage of the
stack, and overrides the program stack pointer (SP register) if the
`preram_stack` and `postram_stack` point to different addresses.
Previous Boot Flow:
- header.ld -> jump into `stage_entry` C code for ROMSTAGE onwards ->
`stage_entry` being called and followed by `main` function
Updated Boot Flow:
- header.ld -> jump into `_start` (assembly entry point) for
ramstage specifically -> Update the existing SP (stack pointer)
register if the `preram_` or `postram_` stack address is not
same -> call into `stage_entry` and follow the `main` function.
BUG=b:456953373
BRANCH=None
TEST=Able to build google/bluey.
Change-Id: I4eec24aff1c9d01180c3452a3631dd344656c771
Signed-off-by: Subrata Banik <subratabanik@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90403
Reviewed-by: Julius Werner <jwerner@chromium.org>
Reviewed-by: Jayvik Desai <jayvik@google.com>
Reviewed-by: Kapil Porwal <kapilporwal@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Refactor the stack definition macros to explicitly define separate
memory regions for the stack, addressing resource conflicts on
certain SoCs like Qualcomm x1p42100.
The original STACK macro is split into PRERAM_STACK and
POSTRAM_STACK.
Motivation:
On the Qualcomm x1p42100 SoC, the boot flow presents two
constraints for the initial stack location:
- Boot IMEM is unavailable after the ADSP is loaded.
- The existing SSRAM stack address is reserved for QC QSEE by the
Trust Zone.
Solution:
- PRERAM_STACK: Used by coreboot (e.g., till romstage) for static
stack allocation (from an alternative SSRAM or BOOT IMEM region).
- POSTRAM_STACK: Used starting from ramstage, leveraging the
DRAM-mapped memory.
This conditional split allows coreboot to manage stack memory
independently for the limited environment before DRAM is fully
initialized (ENV_ROMSTAGE_OR_BEFORE), resolving the hardware memory
conflicts while maintaining compatibility with existing code via
aliasing.
BUG=b:456953373
BRANCH=None
TEST=Able to build google/bluey.
Change-Id: I6356adc63d595f59050e6dc5961404be4a9534c0
Signed-off-by: Subrata Banik <subratabanik@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90402
Reviewed-by: Julius Werner <jwerner@chromium.org>
Reviewed-by: Kapil Porwal <kapilporwal@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Jayvik Desai <jayvik@google.com>
On the X1P42100 SoC, the System Debug Image (SDI) flow is handled by
the Always-On Processor (AOP), unlike previous architectures (e.g.,
Kodiak) which utilized a dedicated 'QcSDI' image.
Rename the memory region at 0x14699000 from 'qcsdi' to 'aop_sdi' to
accurately reflect ownership by the AOP and to align with the new
chipset architecture. This change clarifies the memory map and serves
as a prerequisite for removing legacy QcSDI artifacts once the
modern SDI flow is fully enabled.
BUG=b:456953373
TEST=Build and boot on google/quenbih target.
Change-Id: Ic5626c06decaadfd459aa21bde4efcfec92e1e47
Signed-off-by: Subrata Banik <subratabanik@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90505
Reviewed-by: Jayvik Desai <jayvik@google.com>
Reviewed-by: Kapil Porwal <kapilporwal@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The script was failing when origin/main doesn't exist. Instead of
trying to detect or use a main branch, simply save the current HEAD
hash and restore it when done. This works regardless of branch names
or remote configuration.
Also improve the clean check to use git diff-files instead of
comparing to a specific branch.
Change-Id: I237de4b1e8a06fd4e1e3ef08286208c130e7a6bd
Signed-off-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90502
Reviewed-by: Martin L Roth <gaumless@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Replace the incorrect Intel confidential license header with the
proper BSD license header in the MemInfoHob.h file to align with
the standard licensing used for Intel FSP vendorcode.
BUG=None
TEST=None
Change-Id: I242d9abedc2910f819c70be597c5d1cbca996a2a
Signed-off-by: Appukuttan V K <appukuttan.vk@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90494
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
This fixes the build failure for commit 0b4d41004 (mb/lenovo/sklkbl: Add
Lenovo Thinkpad X280 as a variant) caused by a discrepancy between
memory/Makefile.mk and the actual content of spd/ddr4.
Change-Id: I92a4446e7bd457a7f09a107a0cb0fe1d7a6e1de4
Signed-off-by: Johann C. Rode <jcrode@gmx.net>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90503
Reviewed-by: Matt DeVillier <matt.devillier@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This proliferates the fixes from commit e5d10e5d23, CB:90023 ("mb/
lenovo/t480: Fix headphone jack") to the other SKL/KBL Thinkpad
variants T580, T470s, and T480s. This has been only validated on the
former two machines, but since the hardware changes between the
different models appear to be minimal, it should be safe to deploy
this fix to the T480s variant blindly.
Change-Id: I1edf8dc33231b9d1e1cf2eaf3f4f296736b7be32
Signed-off-by: Johann C. Rode <jcrode@gmx.net>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90482
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Matt DeVillier <matt.devillier@gmail.com>
The hardware of this machine is mostly identical to the already
supported Thinkpad Tx80 machines. Minor differences are the absence of
SODIMM slots (i.e. all RAM is soldered on), two fewer eDP lane pairs,
and different PCIe lane assignment.
All the hardware configuration settings (VBT, GPIO register dumps,
hda verbs, and so on) have been obtained running the latest BIOS/EC
firmware (1.59/N20ET74W, 1.15/N20HT28W). PCIe port assignments as well
as USB, RAM SPD and GPIO assignments have been cross-checked against
publicly available schematics (Finn-1).
Basic functionality has been validated on a Thinkpad X280 part number
20KF-002QUS with 16GB onboard RAM, i7-8650U CPU. The laptop has been
tested with SeaBIOS 1.17 as a payload booting either debian 13 or
Windows 10. A cursory check of the hardware (video, wifi, audio, wired
ethernet, reboot, sleep) shows no issues. This patch also includes a
fix for the headphone jack detection that's been already validated
on the Thinkpad T480 with commit e5d10e5d23 ("mb/lenovo/t480: Fix
headphone jack").
Change-Id: Ia8e6c40b200dee240d08d79253fbbe0842882a80
Signed-off-by: Johann C. Rode <jcrode@gmx.net>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90254
Reviewed-by: Matt DeVillier <matt.devillier@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Logs such as "DRAM-K: calibration failed: status = 1" give little
information about the failure reason. Add get_status_string() and use it
to print the return status as a string.
Change-Id: If20282f0de7ba8ce884d0016fe8da1dc93a33ea4
Signed-off-by: Yu-Ping Wu <yupingso@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90484
Reviewed-by: Yidi Lin <yidilin@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Chen-Tsung Hsieh <chentsung@google.com>
Optimize the `sdram_size` function by caching the calculated SDRAM
size in a static variable. This prevents redundant calls to
`mtk_dram_size()` or `mem_chip_info_total_density_bytes()` if the size
has already been determined, improving performance in scenarios where
`sdram_size` is called multiple times.
BUG=none
TEST=emerege-tanjiro coreboot
Change-Id: I0ca0df80ee9cb781a5bb6d55ee28a2c1153be0ad
Signed-off-by: Yidi Lin <yidilin@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90485
Reviewed-by: Yu-Ping Wu <yupingso@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
This patch enables MTE (Memory Tagging Extension) for the MediaTek
MT8196 SoC.
During `soc_init`, it calculates the required size and start address for
the MTE tag storage based on the physical DRAM size. It then calls
`booker_mte_init` to initialize the MTE hardware with the calculated
start address.
Later, during memory initialization, `bootmem_platform_add_ranges` uses
`bootmem_add_range_from` to reserve the calculated memory region for
MTE tag storage, preventing it from being used for other purposes.
BUG=b:438666196
TEST=Check cbmem log.
[DEBUG] booker_mte_init: MTE tag addr 0x460f70000
...
[DEBUG] 17. 0000000460f70000-000000047ffeffff: TAG STORAGE
[DEBUG] 18. 000000047fff0000-000000047fffffff: RESERVED
Change-Id: I7caa4fde4f314261383a68e942b0e3fb06c6184b
Signed-off-by: Yidi Lin <yidilin@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90144
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Chen-Tsung Hsieh <chentsung@google.com>
Reviewed-by: Yu-Ping Wu <yupingso@google.com>
The Armv9-A architecture introduces the Memory Tagging Extension (MTE),
which uses a dedicated memory region for tag storage.
This patch adds a new memory type, BM_MEM_TAG, to allow for the proper
accounting and reservation of this memory region. This ensures that the
payload, e.g. depthcharge, can correctly identify and utilize the tag
storage area.
BUG=b:438666196
Change-Id: I2f6d2b3c2c1a8e1f0e9b2c3d4e5f6a7b8c9d0e1f
Signed-off-by: Yidi Lin <yidilin@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90470
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Yu-Ping Wu <yupingso@google.com>
This patch introduces a new function, bootmem_add_range_from, which
allows adding a memory range of a specific type only if it is carved
out from a range of another specific type. This is useful for cases
where memory needs to be allocated from a pre-defined region.
The function checks if the target range is fully contained within a
range of `from_tag` before marking it as `new_tag`. Error reporting
is included to log cases where the allocation is not possible.
BUG=b:438666196
Change-Id: Icfdb5ef9114572c075be6ef4e57d00151300a17a
Signed-off-by: Yidi Lin <yidilin@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90469
Reviewed-by: Yu-Ping Wu <yupingso@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
GCC generates correct code for __builtin_bswapXX() on all architectures,
including ArmV4. It seems that whatever bug caused this to not work back
in commit 879ea7fce8 ("endian: Replace explicit byte swapping with
compiler builtin") has been fixed now. We can eliminate the swabXX()
functions and simplify the code.
All instances that had been calling these functions directly should have
been using real endianness conversions anyway.
Change-Id: I19713fd009aa5c0e01c4a42e0cf012364d6bed60
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90438
Reviewed-by: Maximilian Brune <maximilian.brune@9elements.com>
Reviewed-by: Yu-Ping Wu <yupingso@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Frans Hendriks <fhendriks@eltan.com>
In preparation for the upcoming DSI dual channel support, pass dsi_regs
and mipi_tx_regs to DSI functions that need to access these registers.
BUG=b:424782827
TEST=util/abuild/abuild -x -t GOOGLE_SAPPHIRE -a
BRANCH=none
Change-Id: Ia0c9051148e38a7703119f800d417f2f8b52f78a
Signed-off-by: Payne Lin <payne.lin@mediatek.corp-partner.google.com>
Signed-off-by: Yu-Ping Wu <yupingso@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90446
Reviewed-by: Yidi Lin <yidilin@google.com>
Reviewed-by: Chen-Tsung Hsieh <chentsung@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Add dsi_register_v3.h to define dsi_regs struct for mt8196, and add
dsi_reg.h to define mipi_tx_regs struct. Unlike other SoCs, mt8196
has dsi1 and mipi_tx1 registers in order to support DSI dual channel.
BUG=b:424782827
TEST=util/abuild/abuild -x -t GOOGLE_SAPPHIRE -a
BRANCH=none
Change-Id: I2b541c89a007f380de482bffa86aca60c351b526
Signed-off-by: Payne Lin <payne.lin@mediatek.corp-partner.google.com>
Signed-off-by: Yu-Ping Wu <yupingso@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90451
Reviewed-by: Chen-Tsung Hsieh <chentsung@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Yidi Lin <yidilin@google.com>
Move dsi0 variable definition to dsi_register_v*.h to be closer to the
register struct definition. This also allows us to define dsi1 for
mt8196 at the same place as dsi0 in a future patch. The dsi1 variable
cannot be defined in dsi_common.h because not all MediaTek SoCs using
the header have the DSI1_BASE register.
BUG=b:424782827
TEST=util/abuild/abuild -x -t GOOGLE_SAPPHIRE -a
BRANCH=none
Change-Id: I643f8d0317d0fb8bd8ea7bf8870ec99051c35c1b
Signed-off-by: Yu-Ping Wu <yupingso@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90452
Reviewed-by: Yidi Lin <yidilin@google.com>
Reviewed-by: Chen-Tsung Hsieh <chentsung@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The soc/dsi.h header contains SoC-specific dsi and mipi_tx register
definitions, which are not needed for SoCs not supporting MIPI DSI
panels (such as mt8195). To decouple the generic display.c (used for
both eDP and MIPI panels) from those register definitions, move the
mtk_dsi_init() declaration and MIPI_DSI_* enums to display_dsi.h.
This allows us to remove the unused soc/dsi.h for mt8195.
BUG=b:424782827
TEST=util/abuild/abuild -x -t GOOGLE_ASURADA -a
TEST=emerge-skywalker coreboot
BRANCH=none
Change-Id: I56e458ec8077ed48929637b9b5c70f08653cc73f
Signed-off-by: Yu-Ping Wu <yupingso@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/90449
Reviewed-by: Yidi Lin <yidilin@google.com>
Reviewed-by: Chen-Tsung Hsieh <chentsung@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
