AMD family 17h and newer don't use cache as RAM, since the RAM is
already initialized by the PSP when the x86 cores are released from
reset. Therefore they use a different linker script as the rest of the
x86 chips in coreboot do. Since there will be support for newer
generations than Picasso will be added, move those linker scripts from
soc/amd/picasso to soc/amd/common/block/cpu/noncar.
TEST=Timeless build of amd/mandolin and amd/gardenia result in identical
binaries.
Change-Id: Ie60372aa498b6e505708f97213b502c9d0b3534b
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/47828
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
By renaming the AMD SOC common Kconfig file the wildcard to source all
AMD SoC-specific Kconfig files won't match to it and it can be sourced
after all SoC-specific Kconfig files in the sub-directories are sourced.
This change allows adding new SoCs without having to edit the soc/amd
Kconfig file.
Change-Id: Iaaa5aad23eb6364d46b279101f3969db9f182607
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/47701
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
The code that uses the GPU device ID uses the correct ATI vendor ID, but
the description wrongly used AMD as vendor. In the AMD APUs the GPU PCI
device and the corresponding audio controller use the ATI PCI vendor ID
while all other PCI devices in the SoC use the AMD PCI vendor ID.
Also move the two entries in a separate section right below the one they
were in.
Change-Id: Ia0b5bd4638f5b07c487f223321872563b36337e9
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/47674
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
SOC_AMD_COMMON needs to be selected to be able to select
SOC_AMD_COMMON_BLOCK which only includes the Kconfig files from the
function block sub-folder. Removing SOC_AMD_COMMON_BLOCK and the
corresponding Kconfig file and make SOC_AMD_COMMON include all Kconfig
files from the sub-folders simplifies this a bit.
Change-Id: I9068d57a80bdc144e73d2b8c00e7b2cae730d4b6
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/47672
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Jason Glenesk <jason.glenesk@gmail.com>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Since coreboot is written in C and not C++, having the double
underscores as a prefix is not an issue, but it also doesn't add much
information, so drop them and the trailing ones as well.
Change-Id: I1028fb9097efab8ffae5ffa9fe85a97feebc78a9
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/47583
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The SMU mailbox access code from Picasso can be reused in the next
generation, so factor out the code to soc/amd/common/block/smu. Since
the mailbox register offsets in the indirect address space, the number
of arguments and the message IDs don't always match between different
devices, keep those in the soc-specific directories.
Change-Id: Ibaf5b91ab35428e4c771e7163c6e0c4fc50371e7
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/47483
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Jeremy Soller <jeremy@system76.com>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
The cpu/car code only applies to pre-family-17h CPUs that still use
cache as RAM (CAR) and the PI code only applies to the pre-FSP vendor
code blob binaryPI interface.
Change-Id: I5a13d7e202bb745255fabb46110850c36b07de7a
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/47274
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Currently, there's no check for the table being programmed. This skips
programming a table if the table size is zero, or the pointer to the
table has been set to NULL.
BUG=None
TEST=Set table pointer to NULL, table doesn't run.
BRANCH=Zork
Signed-off-by: Martin Roth <martinroth@chromium.org>
Change-Id: I7d09b47e7d619428b64cc0695f220fb64c71ef4c
Reviewed-on: https://review.coreboot.org/c/coreboot/+/47307
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Expand fam17h to family 17h to make the help text slightly easier to
understand and add family 19h to the description of
SOC_AMD_COMMON_BLOCK_PSP_GEN2.
Change-Id: I9284c9c638e1d217d4605523dde004781f4343f9
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/47275
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Two of the PSP mailbox commands are only applicable to the first
generation of PSP mailbox interface.
Change-Id: Ice940ee780c3d96ae1d9ec7ba49ea4add00e8723
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/47180
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Fix function call to invoke 16-bit read in 16-bit api instead of 8-bit read.
Signed-off-by: Igor Bagnucki <bagnucki02@gmail.com>
Change-Id: Ifd9079fc6446125e0e58402fdb64bc198bb8e381
Reviewed-on: https://review.coreboot.org/c/coreboot/+/45374
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marc Jones <marc@marcjonesconsulting.com>
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Calling espi_open_generic_io_window in espi_open_io_window depends on
the condition in the preceding if statement, so move the command into an
else block to make it more obvious that this is the case.
TEST=Timeless build results in identical image.
Change-Id: I3039817afd79c30a2df2f2f54e7848f52dc2c487
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44353
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Create two new functions to fetch mrc_cache data (replacing
mrc_cache_get_current):
- mrc_cache_load_current: fetches the mrc_cache data and drops it into
the given buffer. This is useful for ARM platforms where the mmap
operation is very expensive.
- mrc_cache_mmap_leak: fetch the mrc_cache data and puts it into a
given buffer. This is useful for platforms where the mmap operation
is a no-op (like x86 platforms). As the name mentions, we are not
freeing the memory that we allocated with the mmap, so it is the
caller's responsibility to do so.
Additionally, we are replacing mrc_cache_latest with
mrc_cache_get_latest_slot_info, which does not check the validity of
the data when retrieving the current mrc_cache slot. This allows the
caller some flexibility in deciding where they want the mrc_cache data
stored (either in an mmaped region or at a given address).
BUG=b:150502246
BRANCH=None
TEST=Testing on a nami (x86) device:
reboot from ec console. Make sure memory training happens.
reboot from ec console. Make sure that we don't do training again.
Signed-off-by: Shelley Chen <shchen@google.com>
Change-Id: I259dd4f550719d821bbafa2d445cbae6ea22e988
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44006
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Move interrupt status and wake status clearing to after GPIO config so
that configuration does not incorrectly set interrupt or wake status.
i.e. when PULL_UP is configured on a pad, it incorrectly sets in the
interrupt status bit. Thus, the interrupt status bit must be cleared
after initial pad configuration is complete.
BUG=b:164892883, b:165342107
TEST=None
BRANCH=None
Signed-off-by: Josie Nordrum <josienordrum@google.com>
Change-Id: If4a5db4bfa6a2ee9827f38e9595f487a4dcfac2c
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44640
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
The Linux kenerl driver for AMD gpu currently has a floor
value of 12 for brightness settings (AMDGPU_DM_DEFAULT_MIN_BACKLIGHT).
AMD indicates they did this because they were concerned with certain
panels flickering at lower backlight values. However, for unaffected
panels it's desirable to be able to have the panel "turn off" at
the lowest backlight setting. The only way to do that is to provide
ATIF bindings that indicate backlight range.
Option SOC_AMD_COMMON_BLOCK_GRAPHICS_ATIF is added to provide a full
range for the backlight setting. If needed, this path can be built upon
for fuller support, but for the time being this is the only thing
necessary to make the backlight be full range.
BUG=b:163583825
Change-Id: If76801a8daf6a5e56ba7d118956f3ebce74e567a
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44642
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
In order to log gpio events for wake purposes the state
of the gpio subsystem should be snapshotted. Add the ability
to capture state of gpio subystem as well as saving up to 16
gpios that indicate their wake status.
Likewise, provide the eventlog additions based on state.
BUG=b:159947207
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Change-Id: I49fca56c87543aa8aad0eb7da5c5cb570c4349d5
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44534
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The wake source macro for GPE events was using 'GPIO'. However,
current usage is really all GPEs. Therefore, provide clarity
in the naming in order to allow for proper GPIO wake events
that are separate from the ACPI GPE block.
BUG=b:159947207
Change-Id: I27d0ab439c58b1658ed39158eddb1213c24d328f
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44527
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
GPE events were not be recorded in the eventlog. Add those
to the eventlog when the status register indicates those events.
BUG=b:159947207
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Change-Id: Ifb3167fd24f2171b2baf1a65eb81a318eb3e7a86
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44489
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Now that all users of the functions manipulating global state
and using soc-specific objects are removed remove those functions.
BUG=b:159947207
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Change-Id: I18c4c8b0c7852dde8cf0b6b3f11e43e15c3ce155
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44485
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
In order to reduce code duplication provide an acpi_fill_gnvs()
helper function. Intent is to move stoneyridge and picasso over
to using this common implementation instead of duplicating it.
BUG=b:159947207
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Change-Id: I21c6e2c24eaf42f31ae57c05df7f633d7dc266d9
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44482
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
The existing code in common/block/acpi is mixing multiple operations:
saving things to cbmem in common code but then soc code uses that
information, reliant upon soc-specific struct soc_power_reg object,
and only saving/snapshotting ACPI registers very deep in ramstage.
To unwind the above provide some functions that are more targeted:
- Add struct acpi_pm_gpe_state object
- Add acpi_fill_pm_gpe_state()
- Add acpi_pm_gpe_add_events_print_events()
- Add acpi_clear_pm_gpe_status()
BUG=b:159947207
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Change-Id: Ia7afed2861343802b3c78728784f7cfaf6f53f62
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44481
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
espi_poll_status describes better what the function actually does, since
it polls the status register instead of just doing a single read to
check.
Change-Id: I0feeef5504bd911e1fb0a00d4f4c546df3548db2
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44354
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
It's only an error if bits other than ESPI_STATUS_DNCMD_COMPLETE are set
in the status register. If ESPI_STATUS_DNCMD_COMPLETE isn't set, the
command failed, so we expect that one to be set.
Change-Id: I6f1fb5a59b1ecadd6724a07212626f21fb90e7e7
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44352
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
When address and data register for the SIO control register access is
passed as one I/O region with a size of 2, the corresponding special
decode enable register should be used instead of a generic one to save
the rather limited generic ones for other decode ranges.
Change-Id: Ie54ff6afa2bd2156f7b3a3cf83091f1f932b6993
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44351
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
We can just return at all places where the ret variable was written
before its value gets returned at the end of the function.
Change-Id: Id87f41c0d9e3397879ac3d15b13179cca1a1263f
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44350
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Since this is a bit mask applied to the raw value of a 32 bit register,
this should be a 32 bit unsigned type.
Change-Id: I9d9930963d8c827a84dc1f67e2f2fa8f95ab40f2
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44349
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Th register number passed to the low level read/write functions should
never be negative.
Change-Id: I5d7e117b3badab900d030be8e69ded026d659f8a
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44348
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
I would like to make assertions evaluate at compile time where possible,
but sometimes people used a literal assert(0) to force an assertion in a
certain code path. We already have BUG() for that so let's just replace
those instances with that.
Signed-off-by: Julius Werner <jwerner@chromium.org>
Change-Id: I674e5f8ec7f5fe8b92b1c7c95d9f9202d422ce32
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44047
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
The Embedded Firmware Structure should contain SPI speed, mode
and Micron support for the PSP to program. Add Kconfig options
to specify these values to use for future platform changes.
BUG=b:158755102
TEST=Test menuconfig and platform build for Trembyle and Mandolin.
Signed-off-by: Matt Papageorge <matt.papageorge@amd.corp-partner.google.com>
Change-Id: I78558fa3fa27c70820f0f3d636544127adab6f8b
Reviewed-on: https://review.coreboot.org/c/coreboot/+/42567
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Because there was a lot of discussion about the size increase,
I also looked at the impact of calling the get_spi_bar() function
vs reading spi_base directly and just not worring about whether
or not spi_base was already set.
Using the spi_base variable directly is 77 bytes bytes for all 6
functions. it's roughly double the size to call the function at
153 bytes. This was almost entirely due to setting up a call stack.
If we add an assert into each function to make sure that the spi_base
variable is set, it doubles from the size of the function call to
333 bytes.
For my money, the function call is the best bet, because it not only
protects us from using spi_base before it's set, it also gets the
value for us (at least on x86, on the PSP, it still just dies.)
BUG=b:161366241
TEST: Build
Signed-off-by: Martin Roth <martin@coreboot.org>
Change-Id: I0b0d005426ef90f09bf090789acb9d6383f17bd2
Reviewed-on: https://review.coreboot.org/c/coreboot/+/43772
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Previously, the spi base address code was using a number of different
functions in a way that didn't work for use on the PSP.
This patch consolidates all of that to a single saved value that gets
the LPC SPI base address by default on X86, and allows the PSP to set
it to a different value.
BUG=b:159811539
TEST=Build with following patch to set the SPI speed in psp_verstage.
Signed-off-by: Martin Roth <martinroth@chromium.org>
Change-Id: I50d9de269bcb88fbf510056a6216e22a050cae6b
Reviewed-on: https://review.coreboot.org/c/coreboot/+/43307
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This change fixes the definition of `GPIO_INT_ENABLE_STATUS_DELIVERY`
to use `GPIO_INT_ENABLE_DELIVERY` instead of
`GPIO_INT_ENABLE_STATUS_DELIVERY`.
Signed-off-by: Furquan Shaikh <furquan@google.com>
Change-Id: I64d912200779875cf121cec4476fd39de74c0223
Reviewed-on: https://review.coreboot.org/c/coreboot/+/43695
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This change adds a macro `PAD_NC` for configuring no-connect pads. This
configures the pad as input with pull-down.
Change-Id: I47c41c88ccfebe2c5dd9a24f85a120af9c8f56b5
Signed-off-by: Furquan Shaikh <furquan@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/43521
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Code has evolved such that there seems to be little
use for global definition of cbmem_top_chipset().
Even for AMD we had three different implementations.
Change-Id: I44805aa49eab526b940e57bd51cd1d9ae0377b4b
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/43326
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The SMM_LOCK bit isn't in SMM_MASK_MSR, but in HWCR_MSR, so move it
there. The soc/amd/* code itself uses the bit definition when accessing
HWCR_MSR, so SMM_LOCK was just below the wrong MSR definition.
Also remove SMM_LOCK from comment about masking bits in SMM_MASK_MSR,
since that bit isn't in that MSR.
TEST=Checked the code and the corresponding BKDG/PPR.
Change-Id: I2df446f5a9e11e1e7c8d10256f3c2803b18f9088
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/43309
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
