Add a function to check if the CPU placed at the specified socket was
found usable during QPI init. This is useful for multi-socket
platforms were a CPU is missing or has been disabled due to an error.
Change-Id: I135968fcc905928b9bc6511e3ddbd7d12bad0096
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/76556
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Use the FSP define to iterate over all sockets as the runtime value of
numofIIO is the detected number of sockets, not the highest working
socket.
This fixes printing the HOB on multi-socket platforms where a CPU has
been removed or has been disabled (4S system running as 3S).
Change-Id: Ieed67cd48d26c7634636c0aae6a56f3b6fbdf640
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/76492
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
On Intel Meteor Lake (MTL), PCIe CLK control register is accessed by
P2SB on IOE/SOC die.
So this patch does:
1. Enable PCIE_CLOCK_CONTROL_THROUGH_P2SB
2. Include pcie_clk.asl
3. Set the correct IOE_DIE_CLOCK_START for MTL-U/H.
BUG=b:288976547, b:289461604
TEST=Test on google/screebo and found the pcie clock is on/off properly
and sdcard PCIe port doesn't block S0ix with RTD3 cold enabled.
Change-Id: I6788ae766f36c9a0d4910fda1d6700f20ce73ea8
Signed-off-by: Kane Chen <kane.chen@intel.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/76356
Reviewed-by: Kapil Porwal <kapilporwal@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
In the older platform such as Raptor Lake (RPL), Tiger Lake (TGL), it
needs PMC IPC cmd to turn on/off the corresponding clock.
Now on Meteor Lake (MTL), it control pcie clock registers on P2SB on
IOE or SoC die.
BUG=b:288976547, b:289461604
TEST=Test on google/screebo and found the pcie clock is on/off properly
and sdcard pcie port doesn't block S0ix with RTD3 cold enabled.
Change-Id: Ia729444b561daafc2dca0ed86c797eb98ce1f165
Signed-off-by: Kane Chen <kane.chen@intel.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/76347
Reviewed-by: Kapil Porwal <kapilporwal@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This patch creates a helper library to migrate all the common P2SB
access routines. The PCH P2SB ACPI implementation will now rely on the
common library to perform PCR read/write operations. This will make the
code more modular and easier to maintain.
The helper library provides a single interface for accessing P2SB
registers. This makes it easier to port the code to different platforms,
for example: adding support for PS2B belongs to the IOE die for
Meteor Lake SoC generation.
BUG=b:290856936
TEST=Able to build and boot google/rex.
Change-Id: I0b2e7ea416ca7082f68d0b822ebb9a87025b4a8b
Signed-off-by: Subrata Banik <subratabanik@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/76408
Reviewed-by: Kapil Porwal <kapilporwal@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
In the JPEG decoder, use `bytes_per_line` instead of `width` for
address calculations, to allow for bigger framebuffers. When
calling jpeg_decode(), add an offset to the framebuffer address
so the picture gets centered.
Change-Id: I0174bdccfaad425e708a5fa50bcb28a1b98a23f7
Signed-off-by: Nico Huber <nico.h@gmx.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/76424
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin L Roth <gaumless@gmail.com>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
Reviewed-by: Subrata Banik <subratabanik@google.com>
In all SoC pm_set_power_failure_state gets called either after a call to
enable_acpimmio_decode_pm04() or the ACPIMMIO mapping is already enabled
after reset on the SoC. This allows to use pm_read8 and pm_write8 that
use the ACPIMMIO mapping of the PM registers instead of pm_io_read8 and
pm_io_write8 which won't work on Phoenix and Glinda due to the IO ports
used on older generations to access to the PM registers not being
implemented any more.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Id0d0523d2c4920da41b3fb73cf62f22a60f1643a
Reviewed-on: https://review.coreboot.org/c/coreboot/+/76463
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin L Roth <gaumless@gmail.com>
rv32iafc-ilp32 is compatible with rv32iac-ilp32 for library
implementation, so add a reuse rule allowing the default configuration
to support rv32iafc.
-IAFC is an unusual configuration (much less common than -IMAFC),
but multilib reuse has essentially no cost: this change is useful to
users of platforms that support hardware floating-point but cannot
use hardware multiply/divide for any reason. To avoid generating a
new set of libraries this is limited to the soft-float ABI.
Tested by verifying that `gcc -march=rv32iafc -mabi=ilp32
--print-search-dirs` refers to the rv32iac/ilp32 library directory
as expected, rather than just the root library directory as occurs
when an unsupported target is selected (for instance, rv32id).
Change-Id: Ie056ba6488a138fe0876eebf7cbc59477b3c3518
Signed-off-by: Peter Marheine <pmarheine@chromium.org>
Signed-off-by: Patrick Georgi <patrick@coreboot.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/76539
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
In all SoC lpc_early_init gets called either after a call to
enable_acpimmio_decode_pm04() or the ACPIMMIO mapping is already enabled
after reset on the SoC. This allows to use pm_read8 and pm_write8 that
use the ACPIMMIO mapping of the PM registers to set the PM_LPC_ENABLE
bit in the PM_LPC_GATING register instead of pm_io_read8 and
pm_io_write8 which won't work on Phoenix and Glinda due to the IO ports
used on older generations to access to the PM registers not being
implemented any more.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I8b31ec4e03a06796502c89e3c2cfaac2d41b0ed9
Reviewed-on: https://review.coreboot.org/c/coreboot/+/76461
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The enable_acpimmio_decode_pm04 function uses the IO port based indirect
access of the PM register space. The PM_INDEX and PM_DATA registers
don't exist any more on Glinda, so the code shouldn't access those.
Since the PM_04_ACPIMMIO_DECODE_EN bit in the
ACPIMMIO_DECODE_REGISTER_04 register is 1 after reset, the ACPIMMIO
space is still accessible.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ic6bc0479ea4ea2b9fe3629a6e15940b31b2864d3
Reviewed-on: https://review.coreboot.org/c/coreboot/+/76456
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
The enable_acpimmio_decode_pm04 function uses the IO port based indirect
access of the PM register space. The PM_INDEX and PM_DATA registers
don't exist any more on Phoenix, so the code shouldn't access those.
Since the PM_04_ACPIMMIO_DECODE_EN bit in the
ACPIMMIO_DECODE_REGISTER_04 register is 1 after reset, the ACPIMMIO
space is still accessible.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ia41f239b023edc094f5cbae63ed7c079649c74da
Reviewed-on: https://review.coreboot.org/c/coreboot/+/76437
Reviewed-by: Matt DeVillier <matt.devillier@amd.corp-partner.google.com>
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This patch disables early EC sync to avoid an idle delay (~3sec)
without a provision to notify the user about some critical task
in progress.
Doing EC sync at later stage allows us to notify using graphical msg
on screen to make user aware of the WIP task.
BUG=b:279944831
TEST=Able to perform EC sync from depthcharge on google/rex.
Signed-off-by: Subrata Banik <subratabanik@google.com>
Change-Id: I03ed40827c50e75ceaaf94e30d675014ebf22dac
Reviewed-on: https://review.coreboot.org/c/coreboot/+/74837
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
Reviewed-by: Yu-Ping Wu <yupingso@google.com>
Add support for AP initiated mode entry. The code flow has been
optimized as below -
Code flow when AP initiated mode entry is disabled:
+-------+
| Start |
+---+---+
|
|
+---------+---------+
|wait_for_connection|
| Is DP ALT mode |
| available? |
+---------+---------+
|
+--------------->-------+
Yes| No |
+---------+---------+ |
|Skip enter_dp_mode | |
+---------+---------+ |
| |
| |
+-----------+----------+ |
|wait_for_dp_mode_entry| |
| Is DP flag set? | |
+-----------+----------+ |
| |
+--------------->--------
Yes| No |
+-----------+----------+ |
| wait_for_hpd | |
| Is HPD_LVL flag set? | |
+-----------+----------+ |
| |
+--------------->--------
Yes| No |
+-----------+----------+ |
| Rest of the code | |
+-----------+----------+ |
| |
+---------------<-------+
|
+---+---+
| End |
+-------+
Code flow when AP initiated mode entry is enabled:
+-------+
| Start |
+---+---+
|
+------------+-----------+
|Skip wait_for_connection|
+------------+-----------+
|
+--------+-------+
| enter_dp_mode |
| Is USB device? |
+--------+-------+
|
+--------------->-------+
Yes| No |
+---------+---------+ |
| enter_dp_mode | |
| Send DP mode | |
| entry command | |
+---------+---------+ |
| |
+-----------+----------+ |
|wait_for_dp_mode_entry| |
| Is DP flag set? | |
| (If not, loop wait | |
| until timed out) | |
+-----------+----------+ |
| |
+--------------->--------
Yes| No |
+-----------+----------+ |
| wait_for_hpd | |
| Is HPD_LVL flag set? | |
| (If not, loop wait | |
| until timed out) | |
+-----------+----------+ |
| |
+--------------->--------
Yes| No |
+-----------+----------+ |
| Rest of the code | |
+-----------+----------+ |
| |
+---------------<-------+
|
+---+---+
| End |
+-------+
BUG=b:247670186
TEST=Verify display over TCSS and its impact on boot time for
google/rex
Time taken by enter_dp_mode / wait_for_dp+hpd / MultiPhaseSiInit
functions with this patch train:
1. When AP Mode entry is enabled
- With type-c display on C1 and SuzyQ on C0: 6.9ms / 420ms / 616ms
- With USB key on C1 and SuzyQ on C0 : 6.0ms / 505ms / 666ms
- Without any device on C1 and SuzyQ on C0 : 3.7ms / 0ms / 178ms
2. When AP Mode entry is disabled
- With type-c display on C1 and SuzyQ on C0: 1.7ms / 2.5ms / 213ms
- With USB key on C1 and SuzyQ on C0 : 0.9ms / 3.3ms / 177ms
- Without any device on C1 and SuzyQ on C0 : 0.8ms / 1.8ms / 165ms
Without this patch train, wait_for_hpd would cause a constant delay of
WAIT_FOR_HPD_TIMEOUT_MS (i.e. 3 seconds) per type-c port when there is
no device connected.
Signed-off-by: Kapil Porwal <kapilporwal@google.com>
Change-Id: I514ccbdbaf905c49585dc00746d047554d7c7a58
Reviewed-on: https://review.coreboot.org/c/coreboot/+/76366
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Eric Lai <eric_lai@quanta.corp-partner.google.com>
Reviewed-by: Subrata Banik <subratabanik@google.com>
Because of the way that the CBFS filename is generated from the contents
of the microcode patch, if a duplicate microcode patch is included in
the build, the makefile would create a second copy of the name, which
doesn't work. This led to "odd" results where the other attributes of
the first copy were erased, causing cbfstool to fail. The cause of the
failure is not immediately obvious, and is a little difficult to track
down.
This patch causes an immediate failure and gives a reason as to the
cause of the issue.
When a failure is seen, this is the result:
File1: 3rdparty/amd_blobs/phoenix/psp/TypeId0x66_UcodePatch_PHXn4_A0.bin
File2: 3rdparty/amd_blobs/phoenix/psp/TypeId0x66_UcodePatch_PHX4_A0.bin
src/soc/amd/common/block/cpu/Makefile.inc:25: *** Error: The cbfs
filename "cpu_microcode_a740.bin" is used for both above files. Check
your microcode patches for duplicates.. Stop.
TEST=Now checked for both positive and negative failures.
Signed-off-by: Martin Roth <gaumless@gmail.com>
Change-Id: I3d34dc5585182545bdcbfa6370ebc34aa767cae2
Reviewed-on: https://review.coreboot.org/c/coreboot/+/76423
Reviewed-by: Matt DeVillier <matt.devillier@amd.corp-partner.google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
There is a data abort in ABL when the memory training data is used from
APOB Cache. Disable APOB Cache until the cause is identified. The
downside of this change is that the memory training happens in every
boot cycle.
BUG=b:290763369
TEST=Build BIOS image and boot to OS in Myst. Trigger a reboot from AP
console and ensure that the system boots to OS.
Change-Id: I20f4f40cdaac68bca6e121e3a238d13fe80d0d3c
Signed-off-by: Karthikeyan Ramasubramanian <kramasub@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/76422
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martin.roth@amd.corp-partner.google.com>
Reviewed-by: Matt DeVillier <matt.devillier@amd.corp-partner.google.com>
