Since both functions are only called from one function each, inline them
into those functions. Also get_gpio_mux just returned the return value
of iomux_read8, so there were two functions with identical functionality
which shouldn't be the case.
Suggested-by: Martin Roth <martinroth@chromium.org>
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I5662d0226edb25a9954fa47b42e208729a79e5a9
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56830
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Jason Glenesk <jason.glenesk@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This aligns the GPIO MUX access more with the GPIO control register
access and will facilitate adding support for the remote GPIO bank. Also
change the GPIO number argument type to gpio_t.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I4054656c5cc23ea942e8dd370fbbffca304755d6
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56787
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
Introduce amdblocks/acpimmio_legacy_gpio100.h so that the old pre-SoC
chipsets can still access the raw GPIO100 and IOMUX ACPIMMIO registers
while only allowing GPIO accesses through the GPIO API on the SoCs.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I18872dfa40d53ba8b0d7802eec52ede5e2ae617a
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56786
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
Move those two functions near the top of the file to have all functions
that do the hardware accesses in one place.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: If787e6e1d124a932beafd73e5ce7d0ce4869e800
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56782
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Use unsigned integers for variables that aren't supposed to become
negative.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I5ee037221b9818b0474fe0376323e522c1b3b516
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56701
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Since the raw GPIO MMIO register access is now only used inside the
gpio_banks block, the gpio_read32 and gpio_write32 functions can be
moved to that block to reduce the visibility and enforce the usage of
the functions provided by the gpio_banks block.
The iomux_read8 and iomux_write8 functions can't be easily moved to the
gpio_banks block, since it's also used in the pre-SOC AMD chipsets that
use the ACPIMMIO access functions directly.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ia0d6dea72c6bebbbe6ce545bedfc74f91e0042c4
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56781
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Jason Glenesk <jason.glenesk@amd.corp-partner.google.com>
The I2C code should use some GPIO API to access the GPIO registers
instead of accessing the GPIO MMIO regions itself.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I84dff381ad86e0c7f879f0f079186aec9cafc604
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56779
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
No need to do raw GPIO MMIO accesses when basically the same
functionality can be achieved by using existing APIs. Using the existing
GPIO API instead of raw GPIO MMIO register accesses allows containing
all direct GPIO MMIO accesses inside the common AMD GPIO code which will
be done in subsequent patches. Since the value parameter of gpio_set is
int, change the type of the val parameter of drive_scl to int as well
even though I'm not sure why a signed integer was used for this in the
common GPIO API. Since program_gpios already configures the SCL GPIOs as
outputs, gpio_set can be used in drive_scl which only sets the output
value, but doesn't configure the direction.
TEST=The waveform on the SCL pin of I2C3 on a barla/careena Chromebook
looks similar to the same as before during the reset_i2c_peripherals
call, but due to the additional overhead of the read-modify-write to the
GPIO register instead of just a write, the pulse width gets about 50%
longer. Since the udelay call in drive_scl still has an open TODO to
make this configurable and the pulses being longer is in the safe side,
this side-effect can be addressed in a follow-up patch.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ic323cebc1c83ecd6f0e1fbab419c69489d77face
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56777
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
southbridge/amd/pi/hudson uses the common GPIO bank access code from
soc/amd, but doesn't provide all functionality that would be needed to
use the full functionality. Add a Kconfig option that switches off some
functionality in the common SoC GPIO access code, so that more of the
functionality proviced by the common SoC GPIO code can be used in the
AMD binaryPI chipset and board code.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ib10d5d5580aab30a359aa001bb6fc7e9fdb8fc41
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56783
Reviewed-by: Martin Roth <martinroth@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
In CB:56884 we discussed changing the default fast_read speed from
66MHz, which some platforms may not be capable of running, to 33MHz,
which should be generally suitable for all platforms. This same
change has been applied to the default for all SPI speeds.
Signed-off-by: Martin Roth <martinroth@chromium.org>
Change-Id: Ibf926df6829ffdcbae947aaa245356f219615ce8
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57148
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This patch adds code to print the current SPI speeds for each of the 4
different speeds, Normal, Fast-read, Alt-mode, & TPM. It also displays
the SPI mode and whether or not SPI100 mode is enabled.
BUG=b:194919326
TEST: Display the speed, change speeds, show that new speeds are the
expected values.
Signed-off-by: Martin Roth <martinroth@chromium.org>
Change-Id: I7825a9337474c147b803c85c9af7f9dc24670459
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56960
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
Get the settings for fast-read and mode from EFS, and reprogram those.
Program Normal reads, Alt-mode, and TPM speeds from Kconfig settings.
BUG=b:195943311
TEST=Boot and see that SPI was set to the correct speed & mode
Signed-off-by: Martin Roth <martinroth@chromium.org>
Change-Id: I8a24f637b2a0061f60a8f736121d224d4c4ba69b
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56959
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
The common_i2c_save struct isn't specific to the I2C code and since it
contains the state of the GPIO control & status register and the state
of the GPIO MUX register, move it to include/amdblocks/gpio_banks.h and
rename it to soc_amd_gpio_register_save.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: If7cd47e5a32427d856948e319de8dfad8c928e96
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56778
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
No need to strip the quotes of the FSP-M file path in the size check and
it's always a good idea to not remove the quotes around file paths that
will get passed as parameters to shell programs so that spaces in the
path can't cause malfunction.
TEST=All cases still behave as expected for Mandolin.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Suggested-by: Nico Huber <nico.h@gmx.de>
Change-Id: Ieeea84b5861f9d15b2472208432169dc8e3f0049
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57218
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
When CONFIG_FSP_M_FILE isn't defined, the parameter of the file-size
call evaluates to an empty string, so the file-size call will run
"cat | wc -c" which will cause make to get stuck in there. Also print a
message when no FSP-M file is specified that the resulting image won't
boot successfully.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I6b02774e2c79d12554fd076aa01bbe972176f372
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57189
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
Reviewed-by: Martin Roth <martinroth@google.com>
Currently, The SPI speed/mode configuration is split between Kconfig
and devicetree. We'd like to have everything in one place. Since we
need the fast-read speed and the mode available in the Makefile to build
the AMD EFS table, we currently need it in Kconfig. Move all of the
settings to Kconfig and remove them from Devicetree in a later commit.
BUG=b:195943311
TEST=boot majolica & guybrush, verify spi settings
Signed-off-by: Martin Roth <martinroth@chromium.org>
Change-Id: I8f29e49e886bd99b39172905e21bfd392c6c10e2
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56884
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
When modules are added to the FSP and they won't fit into the FSP binary
any more, the size can be increased in the FSP build. Especially in the
case of debug builds the increased size might not fit into the memory
region it gets decompressed into which starts at FSP_M_ADDR and has a
size of FSP_M_SIZE. SoCs can implement the soc_validate_fspm_header
function that ends up being called by the FSP driver in romstage to do
some additional checks on the FSP binary's header that includes the
version number and the image size. We can use the image size field to
check if it fits into the reserved region. Since the FSP-M memory region
is located after romstage loading it won't clobber the romstage code
where we do the check.
This runtime check is added in addition to the build-time check to also
cover the case when the FSP binaries in CBFS get replaced with ones that
don't fit into the reserved memory region after the coreboot build.
BUG=b:186149011
TEST=Mandolin still boots fine with the patch applied. When as a test
the FSP_M_SIZE Kconfig option in soc/amd/picasso is decreased to 0x10000
which is by far not enough for the decompressed FSP-M binary to fit into
it prints the newly added error message on the console and then stops.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I9b74a2d03993ba50b166eb6e87d4e57b93afc069
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57068
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The FSP-M binary needs to fit into the memory region that starts at
FSP_M_ADDR and is FSP_M_SIZE bytes large, so error out during build time
if the uncompressed FSP-M file is larger than the size of the region it
will be copied into.
BUG=b:186149011
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Suggested-by: Furquan Shaikh <furquan@google.com>
Change-Id: Ice4a59e5a723c3c0a40b1f3f3227aee6b9dcb39a
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57129
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
PSP_Verstage will take almost the entire time to run that
is allotted to S0i3 resume. Since coreboot isn't running,
the PSP needs to handle any security requirements. The long-
term plan is that the PSP won't even load psp_verstage on S0i3
resume, but when it is loaded, this makes sure we exit
immediately
BUG=b:177064859
TEST=Verify that PSP_verstage doesn't run on S0i3 resume
Signed-off-by: Martin Roth <martinroth@chromium.org>
Change-Id: Ia7b2560ff3d7621922ec4bc0e8793961f5d7550f
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56919
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Use correct format for constraint list as expected by kernel driver.
With this change, kernel is able to correctly list dummy device in
constraint list.
BUG=b:194687976
TEST=Build and boot to OS in Guybrush.
Change-Id: I7af1941ffd21cd5864c7285f44cb2d063d2f225f
Signed-off-by: Pratik Vishwakarma <Pratik.Vishwakarma@amd.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57012
Reviewed-by: Jason Glenesk <jason.glenesk@amd.corp-partner.google.com>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-by: Rob Barnes <robbarnes@google.com>
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Disabling the 4 DWORD bursts causes SPI DMA operations to stall, so
leave it enabled when SPI DMA is used.
BUG=b:194919326
TEST=Build and boot to OS in Guybrush.
Change-Id: I363acdcdb4178a10e4f7eb2bbcbd6d0ca7924f2d
Signed-off-by: Karthikeyan Ramasubramanian <kramasub@google.com>
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56683
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Jason Glenesk <jason.glenesk@gmail.com>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
Add a new Kconfig option to enable or disable the 4 DWORD burst support
of the SPI controller and use this setting to determine if the
corresponding feature bit in SPI100_HOST_PREF_CONFIG will be set or
cleared. Since fch_spi_disable_4dw_burst can now enable or disable the
feature, rename it to fch_spi_configure_4dw_burst. On Stoneyridge the
SPI_RD4DW_EN_HOST bit needs to be cleared (see the Rd4dw_en_host bit
definition in the SPIx2C SPI100 Host Prefetch Config register in the
public BKDG #55072 Rev 3.09), so add a SoC dependency to the Kconfig
option.
Signed-off-by: Karthikeyan Ramasubramanian <kramasub@google.com>
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Id754fa8d5f9554ed25cf9f3341bfdd1968693788
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56814
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Jason Glenesk <jason.glenesk@gmail.com>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
Move IVRS acpi table generation code to common, so that it can be shared
by other programs.
BUG=b:190515051
TEST=Build picasso coreboot image. Compare IVRS tables before/after
change.
Change-Id: Icd5fec3a9d66e8301e267312020e726d9bc1aa70
Signed-off-by: Jason Glenesk <jason.glenesk@amd.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56450
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Add new IVRS kconfig option to control IVRS generation.
BUG=b:190515051
Change-Id: Iad0c6401dbccd2f3f75464a69e4c27f64d3507a5
Signed-off-by: Jason Glenesk <jason.glenesk@amd.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56694
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
A valid GEVENT number is never negative. The local variable in
set_single_gpio still needs to be a signed integer, since the return
value of get_gpio_gevent being -1 indicates that the GPIO can't generate
a GEVENT. The check for that makes the function return before calling
program_smi of program_sci, so the parameter of those functions can be
changed to unsigned.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I6ce23ceed1585589932824b8cab2a138328672a9
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56705
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Each bit in the GPIO wake status index registers is set to 1 when at
least one of 4 corresponding GPIO pins has its wake status register set.
Added the comment since the gpio_base + i * 4 in the next line looked as
if it calculates some absolute register value which is not what the code
does or should be doing.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I2fc8e9c5bd7c1b011f364b05d0cfdeb0df88ada6
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56703
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Since the parameter the variable gets compared with is size_t type, use
size_t as type for that variable too.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: If82a948bf71079d456616f4438f4b754e0d7262d
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56702
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
With the addition of the remote GPIO support, the GPIO number won't fit
into 8 bit any more, so use the gpio_t type instead which is an uint32_t
typedef.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I3de93fd3a2f2af3c1e3b335fef84019c56482051
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56693
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The offsets of all GPIOs in the up to four regular banks are all
calculated relatively to ACPIMMIO_GPIO0_BANK, so we can just drop the
unused defines for ACPIMMIO_GPIO1_BANK and ACPIMMIO_GPIO2_BANK.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I832ffdca479c1f07219a23b4a7f9be69322dfe03
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56675
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
GPIO bank 3 isn't used in coreboot, but the existence is documented in
both the Picasso PPR #55570 Rev 3.16 and Cezanne PPR #56569 Rev 3.01 and
for those two SoCs all 4 banks are covered by the corresponding
Memory32Fixed region in the DSDT.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Id444a97a398d7e3abfd1f5c4a32e762ee6ff68f1
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56674
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The biosram accessor support in soc/amd/common/block/acpimmio/biosram.c
is only used on Stoneyridge and the old amd/southbridge code and not on
Picasso or Cezanne. It also only builds as a 32 bit binary and breaks
when trying to build as a 64 bit binary, since the size of an uintptr_t
is different on those two. There is no support for using the 32 bit
binaryPI with a 64 bit coreboot while there is code to use a 32 bit FSP
with 64 bit coreboot, so not building this for FSP-based SoC support
moves us one step closer to be able to build coreboot as 64 bit binary
for Picasso and Cezanne.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I2d87ec2fa1b217eaf55d865e4390308812502e56
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56570
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Move the event logging helpers defined in acpi into a separate library.
This will allow logging power management and GPE events for both S3 and
Modern Standby. Introduce a single helper acpi_log_events function to
log both PM and GPE events.
BUG=None
TEST=Build and boot to OS in Guybrush.
Change-Id: I96df66edfc824eb3db108098a560d33d758f55ba
Signed-off-by: Karthikeyan Ramasubramanian <kramasub@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56360
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
TEST=Checked on amd/mandolin with PCO APU and google/guybrush with CZN
APU that the McaXEnable bit is set in the CONFIG registers of all used
MCAX banks.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ia4515ba529e758f910d1d135cdce819f83ea0b5c
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56535
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Enable RSA and SHA for cezanne since support has been added to the PSP.
Also picasso and cezanne have different enums definitions for
hash algorithm, so split that out into chipset.c.
BUG=b:187906425
TEST=boot guybrush, check cbmem -t and the logs
Signed-off-by: Kangheui Won <khwon@chromium.org>
Change-Id: I725b0cac801ac0429f362a83aa58a8b9de158550
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55833
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
This allows keeping track of how long it takes to load the microcode.
BUG=b:179699789
TEST=Boot guybrush
112:started reading uCode 990,448 (10,615)
113:finished reading uCode 991,722 (1,274)
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I86b67cf9d17786a380e90130a8fe424734e64657
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56391
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Jason Glenesk <jason.glenesk@gmail.com>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
This change allows preloading the payload.
BUG=b:179699789
TEST=Boot guybrush and see payload read/decompress drop by 20 ms. We
now spend 7ms decompression from RAM. By switching to LZ4 we drop that
to 500us.
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I3ec78e628f24f2ba0c9fcf2a9e3bde64687eec44
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56053
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Martin Roth <martinroth@google.com>
This CL adds a method that can start the processes of reading the APOB
from SPI. It does require more RAM in ramstage since we no longer mmap
the buffer in the happy path. This will allow us to reduce our
boot time by ~10ms. The SoC code will need to be updated to call
start_apob_cache_read at a point where it makes sense.
BUG=b:179699789
TEST=With this and the patches above I can see a 10 ms reduction in
boot time on guybrush.
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I930d58b76eb4558bc4f48ed928c4d6538fefb1e5
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56232
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
There is no telling when the next udelay will be, so explicitly call
`thread_yield()` after completing a transaction. This will allow any
pending transactions to immediately start.
BUG=b:179699789
TEST=Verify new transaction is enqueued right after another.
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I9c1272bde46c3e0c15305b76c2ea7a6dde5ed0b0
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56321
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Once we enable COOP_MULTITASKING, we need to guarantee that we don't
have multiple threads trying to access the DMA hardware.
BUG=b:179699789
TEST=Boot guybrush with APOB patches.
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: Ibb8e31c95d6722521425772f4210af45626c8e09
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56231
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This change will make it so the standard rdev readat call will use the
SPI DMA controller if the alignment is correct, and the transfer size is
larger than 64 bytes.
There is a magic bit that needs to be set for the SPI DMA controller to
function correctly. This is only available in RN/CZN+.
BUG=b:179699789
TEST=Boot guybrush to OS. This reduces loading verstage by 40ms,
verifying RW by 500us and loading romstage by 500 us.
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I0be555956581fd82bbe1482d8afa8828c61aaa01
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56228
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
Also rename the existing PCI_DEVICE_ID_ATI_FAM19H_MODEL51H_GPU
definition to PCI_DEVICE_ID_ATI_FAM19H_MODEL51H_GPU_CEZANNE to clarify
that that is the one for Cezanne.
BUG=b:193888172
Change-Id: I1c5446c1517f2e0cd708d3275b08d2bce4be0ea8
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56396
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Matt Papageorge <matthewpapa07@gmail.com>
Reviewed-by: Nikolai Vyssotski <nikolai.vyssotski@amd.corp-partner.google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Also move the registers in the order they are in the hardware.
Change-Id: If018e746e58c14475caeda76feb8b5281d7732f1
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56315
Reviewed-by: Martin Roth <martinroth@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
