After the obsoletion of Processor() it is necessary to provide
_CST package to define P_LVLx IO addresses for C2/C3 transitions.
The latency values from _CST will always replace those in FADT.
Change-Id: I3230be719659fe9cdf9ed6ae73bc91b05093ab97
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/74430
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Boards with SOC_INTEL_COMMON_BLOCK_ACPI_CPU_HYBRID have
special handling for the time being.
Change of aopen/dxplplusu is coupled with sb/intel/i82801dx.
Change of emulation/qemu-i440fx is coupled with intel/i82371eb.
For asus/p2b, this adds MADT LAPIC entries, even though platform
has ACPI_NO_MADT selected. Even previously ACPI_NO_MADT creates
the MADT, including an entry for LAPIC address.
Change-Id: I1f8d7ee9891553742d73a92b55a87c04fa95a132
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/74316
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Factor out the get_pstate_core_freq function from the SoC's acpi.c files
to both avoid duplication and to also be able to use the same function
in the TSC frequency calculation in a follow-up patch. The family 17h
and 19h SoCs use the same frequency encoding in the P state MSRs while
the family 1Ah SoCs use a different encoding. The family 15h and 16h
SoCs use another encoding, but since this isn't implemented in
Stoneyridge's acpi.c, this will be added in a follow-up patch.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I8619822c2c61e06ae5db86896d5323c9b105b25b
Reviewed-on: https://review.coreboot.org/c/coreboot/+/74010
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Eric Lai <eric_lai@quanta.corp-partner.google.com>
Reviewed-by: Martin Roth <martin.roth@amd.corp-partner.google.com>
Instead of implementing the conversion from the raw serial voltage ID
value to the voltage in microvolts in every SoC, introduce the
SOC_AMD_COMMON_BLOCK_SVI[2,3] Kconfig options for the SoC to select the
correct version, implement get_uvolts_from_vid for both cases and only
include the selected implementation in the build.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I344641217e6e4654fd281d434b88e346e0482f57
Reviewed-on: https://review.coreboot.org/c/coreboot/+/73995
Reviewed-by: Fred Reitberger <reitbergerfred@gmail.com>
Reviewed-by: Matt DeVillier <matt.devillier@amd.corp-partner.google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Picasso and Cezanne use the serial voltage ID 2 standard to communicate
the CPU voltage to the voltage regulator module on the mainboard, while
Mendocino, Phoenix and Glinda use the serial voltage ID 3 standard for
this. Both standards encode the voltage in a different way, so add the
serial VID version number to the defines to clarify for which version
the define is.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I8ddab8df27c86dc2c70a6dfb47908d9405d86240
Reviewed-on: https://review.coreboot.org/c/coreboot/+/73994
Reviewed-by: Fred Reitberger <reitbergerfred@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Matt DeVillier <matt.devillier@amd.corp-partner.google.com>
Since we already have and use the pstate_msr union in get_pstate_info,
also pass it directly to the get_pstate_core_freq and
get_pstate_core_power function calls avoids having to sort-of convert
the msr_t type parameter in the implementations of those two functions.
In amdblocks/cpu.h a forward declaration of the pstate_msr union is used
since soc/msr.h doesn't exist in the two pre-Zen SoCs that also include
amdblocks/cpu.h.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I112030a15211587ccdc949807d1a1d552fe662b4
Reviewed-on: https://review.coreboot.org/c/coreboot/+/73926
Reviewed-by: Fred Reitberger <reitbergerfred@gmail.com>
Reviewed-by: Matt DeVillier <matt.devillier@amd.corp-partner.google.com>
Reviewed-by: Eric Lai <eric_lai@quanta.corp-partner.google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Add the pstate_msr union of a bitfield struct and a raw uint64_t to
allow easier access of the bitfields of the P state MSRs and use this
bitfield struct in get_pstate_core_freq and get_pstate_core_power. The
signature of those two function will be changed in a follow-up commit.
TEST=The coreboot-generated SSDT containing the P state packages stays
identical on Mandolin.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I8dc293351f9941cfb8a9c84d9fb9a4fd76361d5d
Reviewed-on: https://review.coreboot.org/c/coreboot/+/73643
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Fred Reitberger <reitbergerfred@gmail.com>
The implementations of get_pstate_info of Picasso, Cezanne, Mendocino,
Phoenix and Glinda are identical, so factor it out and move it to the
common AMD SoC code. The SoC-specific get_pstate_core_freq and
get_pstate_core_power functions remain in the SoC-specific code.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ibe0494f1747f381a75b3dd71a8cc38fdc6dce042
Reviewed-on: https://review.coreboot.org/c/coreboot/+/73505
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Fred Reitberger <reitbergerfred@gmail.com>
With the exception of the generate_cppc_entries call, the
implementations of generate_cpu_entries of Picasso, Cezanne, Mendocino,
Phoenix and Glinda are identical, so factor it out and move it to the
common AMD SoC code. Since all SoCs that support CPPC already select the
SOC_AMD_COMMON_BLOCK_ACPI_CPPC Kconfig option, this can be used to only
call generate_cppc_entries for platforms where it is available.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I71323d9d071b6f9d82852479b60dc56c24f2b9ec
Reviewed-on: https://review.coreboot.org/c/coreboot/+/73504
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Fred Reitberger <reitbergerfred@gmail.com>
Rework the way the C state info is generated before it gets passed to
acpigen_write_CST_package in generate_cpu_entries by separating the data
from the code. For this, the newly introduced common get_cstate_info
function is used. Separating the data from the code will eventually
allow moving generate_cpu_entries to the common AMD code.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Id6bd8879ce5968b24893b43041be98db55a4c3c6
Reviewed-on: https://review.coreboot.org/c/coreboot/+/73499
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Eric Lai <eric_lai@quanta.corp-partner.google.com>
Reviewed-by: Fred Reitberger <reitbergerfred@gmail.com>
The FADT data structure is zero-initialized in acpi_create_fadt which
then calls the SoC-specific acpi_fill_fadt function, therefore it's not
needed to assign 0 to the duty_offset and duty_width FADT field in
acpi_fill_fadt for all SoC except Stoneyridge.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ib63b24891d44298841153dfc500b030619e1a5ea
Reviewed-on: https://review.coreboot.org/c/coreboot/+/73421
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Fred Reitberger <reitbergerfred@gmail.com>
Picasso neither has the corresponding P_CNT register implemented nor
writes a _PTC ACPI object that would specify the P_CNT register. The
Picasso UEFI reference code also sets the duty_width FADT entry to 0.
This also aligns the Picasso code with the Cezanne code in this regard.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I74645e5c4e54a2ad6bc7f9e72f5f656027a79860
Reviewed-on: https://review.coreboot.org/c/coreboot/+/73420
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Fred Reitberger <reitbergerfred@gmail.com>
Instead of adding the P-state number to the PSTATE_0_MSR number to get
the P-state MSR number for the rdmsr call, provide a macro that directly
calculates the MSR number for a given power state. Also drop the unused
PSTATE_[1..4]_MSR definitions which also didn't cover all P-state MSRs
available in the hardware.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: If85acf556efe82c209e1608e56c05f7a2a748403
Reviewed-on: https://review.coreboot.org/c/coreboot/+/73323
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
Reviewed-by: Fred Reitberger <reitbergerfred@gmail.com>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
The latency values in the _CST package override the values in the
p_lvl2_lat and p_lvl3_lat FADT fields. In Picasso, Cezanne, Mendocino,
Phoenix and Glinda generate_cpu_entries generates the _CST packages for
each CPU device. The coreboot code for Stoneyridge doesn't generate _CST
packages for the CPU objects, but those are provided via the PSTATE SSDT
binaryPI generates and agesa_write_acpi_tables gets and adds to the ACPI
tables. The AGESA reference code also sets those two FADT entries to the
equivalents of ACPI_FADT_C2_NOT_SUPPORTED and ACPI_FADT_C3_NOT_SUPPORTED
so this also matches the AGESA behavior.
From the ACPI 6.4 spec: "Values provided by the _CST object override
P_LVLx values in P_BLK and P_LVLx_LAT values in the FADT."
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I1116a3013576b18b6f521604d6b0a9d75b971e0b
Reviewed-on: https://review.coreboot.org/c/coreboot/+/73231
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martin.roth@amd.corp-partner.google.com>
The coreboot-common acpi_create_fadt writes a pointer to the FACS table
into both firmware_ctrl and x_firmware_ctl_l FADT fields and sets
x_firmware_ctl_h to zero. When x_firmware_ctl_[l,h] is non-zero, the
pointer in firmware_ctrl will be ignored, but that's what is already
done on Cezanne and newer.
TEST=Linux doesn't complain about any new ACPI problem on Mandolin.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ib9eab4dcf828f28a60c6312ec96872aac4cfb266
Reviewed-on: https://review.coreboot.org/c/coreboot/+/73190
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
Reviewed-by: Fred Reitberger <reitbergerfred@gmail.com>
The FADT data structure is zero-initialized in acpi_create_fadt which
then calls the SoC-specific acpi_fill_fadt function, therefore it's not
needed to assign 0 to the ARM_boot_arch FADT field in acpi_fill_fadt.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ica968db1228a2d63e83f2b6c4ea57c5f02bf1504
Reviewed-on: https://review.coreboot.org/c/coreboot/+/73187
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Fred Reitberger <reitbergerfred@gmail.com>
In CB:71614 Kyösti pointed out that ACPI_GPE0_BLK is the wrong address
to assign to proc_blk_addr; the correct one would be ACPI_CPU_CONTROL.
When looking a bit closer into this, it turned out that
acpigen_write_processor is generating deprecated AML opcodes, so replace
the acpigen_write_processor call with a call to the newly added
acpigen_write_processor_device function that also doesn't have the
proc_blk_addr and proc_blk_len parameters. The information about the IO
port for entering C-states is already written into an SSDT by
acpigen_write_CST_package which is likely also the reason why the wrong
proc_blk_addr value wasn't noticed for a very long time.
TEST=Mandolin still boots Ubuntu 22.04 LTS and Windows 10 and no
possibly related errors show up. Linux gets the expected C-state
information from the _CST package inside the processor device scope.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ie67416e19e431029dd12da66ad44ddfa8586df03
Reviewed-on: https://review.coreboot.org/c/coreboot/+/72490
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Elyes Haouas <ehaouas@noos.fr>
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Generate the PPKG object in the generate_cpu_entries function instead of
generating the PCNT object that is the used in the PPKG method in
cpu.asl to provide the PPKG object. This both simplifies the code and
aligns Picasso with Cezanne and Sabrina. This will also make the code
behave correctly in a case where the number of CPU cores/threads isn't a
power of two.
TEST=Mandolin still boots successfully to Linux desktop and dmesg
doesn't show any any possibly related problems.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ifb84435345c6d8c5d11a8b42e5538cfb86432780
Reviewed-on: https://review.coreboot.org/c/coreboot/+/62517
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
When "fadt->FADT_MinorVersion" is not explicitly set to the right value, gcc sets it up to "0".
So set it correctly for treewide.
Signed-off-by: Elyes Haouas <ehaouas@noos.fr>
Change-Id: Ic9a8e097f78622cd78ba432e3b1141b142485b9a
Reviewed-on: https://review.coreboot.org/c/coreboot/+/62221
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Jason Glenesk <jason.glenesk@gmail.com>
Reviewed-by: Lance Zhao
The `acpigen_write_CST_package` and `acpigen_write_CST_package_entry`
functions don't modify the provided C-state information. So, make the
pointer parameters read-only to enforce this. Also constify arguments
where possible.
Change-Id: I9e18d82ee6c16e4435b8fad6d467e58c33194cf4
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58391
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
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>
