While using '3' is fine for the standard gpe0 for skylake, I want
to make sure anyone that copies this code doesn't tweak GPE0_REG_MAX
without the hard coded index. If that does happen now things will
still work, but it may just not match the hardware proper.
BUG=chrome-os-partner:58666
Change-Id: I434b9a765a0a2f263490bb2b4ecb3635292d46c9
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/17160
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Add a GPIO macro that allows a pin to be routed to the APIC with
the input inverted. This allows a normal interrupt to get used as
a GPE during firmware and still be used as a perhiperal interrupt
in the kernel.
BUG=chrome-os-partner:58666
TEST=boot en eve and use TPM IRQ in firmware and OS
Change-Id: I77f727f749fdd5281ff595a9237fe1e634daba96
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://review.coreboot.org/17176
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Add the necessary infrastructure to support eSPI SMI events,
and a mainboard handler to pass control to the EC.
BUG=chrome-os-partner:58666
TEST=tested on eve board with eSPI enabled, verified that lid
close event from the EC during firmware will result in an SMI
and shut down the system.
Change-Id: I6367e233e070a8fca053a7bdd2534c0578d15d12
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://review.coreboot.org/17134
Tested-by: build bot (Jenkins)
Reviewed-by: Furquan Shaikh <furquan@google.com>
Export the pmc_gpe_init() function from pmc.c to pmutil.c
so it can be used in bootblock, and then call it from there
to initialize any GPEs for use in firmware.
BUG=chrome-os-partner:58666
TEST=test working GPE as TPM interrupt on skylake board
Change-Id: I6b4f7d0aa689db42dc455075f84ab5694e8c9661
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://review.coreboot.org/17135
Tested-by: build bot (Jenkins)
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Add the SOC definition for acpi_get_gpe() so it can be used
by the I2C TPM driver. Also add the I2C support code to
verstage so it can get used by vboot.
BUG=chrome-os-partner:58666
TEST=boot with I2C TPM on skylake board
Change-Id: I553f00a6ec25955ecc18a7616d9c3e1e7cbbb8ca
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://review.coreboot.org/17136
Tested-by: build bot (Jenkins)
Reviewed-by: Furquan Shaikh <furquan@google.com>
The Deep S3 state will lose a lot of register contents that we
used to rely on for determining wake source.
In order to make use of this override the enable bit for wake
sources that are enabled for Deep S3 in devicetree.cb.
BUG=chrome-os-partner:58666
TEST=check for _SWS reporting wake source on S3 resume on skylake
Change-Id: If5113d6890f6cbecc32f92af67a29952266fe0ac
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://review.coreboot.org/17137
Tested-by: build bot (Jenkins)
Reviewed-by: Furquan Shaikh <furquan@google.com>
Remove the FADT from the individual mainboards and select and
use COMMON_FADT in the SOC instead. Set the ACPI revision to 5.
Change-Id: Ieb87c467c71bc125f80c7d941486c2fbc9cd4020
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://review.coreboot.org/17138
Tested-by: build bot (Jenkins)
Reviewed-by: Furquan Shaikh <furquan@google.com>
Three things are required to enable wake-on-usb:
1. 5V to USB ports should be enabled in S3.
2. ASL file needs to have appropriate wake bit set.
3. XHCI controller should have the wake on attach/detach bit set for the
corresponding port in PORTSCN register.
Only part missing was #3.
This CL adds support to allow mainboard to define a bitmap in
devicetree corresponding to the ports that it wants to enable
wake-on-usb feature. Based on the bitmap, wake on attach/detach bits in
PORTSCN would be set by xhci.asl for the appropriate ports.
BUG=chrome-os-partner:58734
BRANCH=None
TEST=Verified that with port 5 enabled, chell wakes up from S3 on usb
attach/detach.
Change-Id: I40a22a450e52f74a0ab93ebb8170555d834ebdaf
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/17056
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Use intel common infrastructure to enable support for write-protecting
SPI flash range. Also, enable this protection for RW_MRC_CACHE.
BUG=chrome-os-partner:58896
TEST=Verified that write to RW_MRC_CACHE fails in OS using
"flashrom -p host -i RW_MRC_CACHE -w /tmp/test.bin"
Change-Id: I35df12bc295d141e314ec2cb092d904842432394
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/17117
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
Write-protect SPI flash range provided by caller by using a free Flash
Protected Range (FPR) register. This expects SoC to define a callback
for providing information about the first FPR register address and
maximum number of FPRs supported.
BUG=chrome-os-partner:58896
Change-Id: I4e34ede8784e5587a5e08ffa10e20d2d14e20add
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/17115
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
We found sdram may fail in pctl_cfg(), so we check the status in this
function. If it exceeds 100ms still in this function, we will restart
the system. We also found there are rare chances DDR training fails,
so also restart system in that case.
BUG=chrome-os-partner:57988
BRANCH=None
TEST=coreboot resets on failure and eventually the system comes up
Change-Id: Icc0688da028a8f4f81eafe36bbaa79fdf2bcea74
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: 89e45f8352
Original-Change-Id: If4e78983abcfdfe1e0e26847448d86169e598700
Original-Signed-off-by: Lin Huang <hl@rock-chips.com>
Original-Reviewed-on: https://chromium-review.googlesource.com/397439
Original-Reviewed-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/17045
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Implement GPIO ACPI AML generating functions that can be called by
coreboot drivers to generate GPIO manipulation code in AML. Following
functions are implemented:
1. acpigen_soc_read_rx_gpio
2. acpigen_soc_get_tx_gpio
3. acpigen_soc_set_tx_gpio
4. acpigen_soc_clear_tx_gpio
BUG=chrome-os-partner:55988
Change-Id: I3d8695d73a1c43555032de90f14ee47ccee45559
Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: https://review.coreboot.org/17082
Tested-by: build bot (Jenkins)
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
This patch adds a Makefile rule for mvmap2315 to install a BDB and
bootblock code in the BOOTBLOCK region. The resulting BDB has a
header and data both signed by a RSA-4096 key.
BUG=chrome-os-partner:57889
BRANCH=none
TEST=emerge-rotor coreboot and examined the output binary. Booted
coreboot.rom.
Change-Id: I1e20a09b12f8f8ed4d095aa588e3eb930f359fc5
Signed-off-by: Daisuke Nojiri <dnojiri@chromium.org>
Reviewed-on: https://review.coreboot.org/16747
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
While we stub out free(), tools like coverity scan have no idea, and it
might change in the future. So free it.
Change-Id: I1d93a6f45b64445662daa95b51128140ad0a87e2
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Found-by: Coverity Scan #1260716
Reviewed-on: https://review.coreboot.org/17055
Tested-by: build bot (Jenkins)
Reviewed-by: Jonathan Neuschäfer <j.neuschaefer@gmx.net>
This refactoring was already carried into RK3288 with commit 6911219
(edid: Add helper function to calculate bits-per-pixel dependent values)
but it seems that the code for RK3399 was copy&pasted from it too early
to pick this up. Fix that so that future Rockchip SoCs can copy&paste
the right thing.
Change-Id: I5050c58d18db38fffabc7666e67a622d4a828590
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/17050
Tested-by: build bot (Jenkins)
Reviewed-by: David Hendricks <dhendrix@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Skip FSP initiated core/MP init as it is implemented and initiated
in coreboot.
Add soc core init to set up the following feature MSRs:
1. C-states
2. IO/Mwait redirection
BUG=chrome-os-partner:56922
BRANCH=None
TEST= Check C-state functioning using 'powertop'. Check 0xE2 and
0xE4 MSR to verify IO/Mwait redirection.
Signed-off-by: Ravi Sarawadi <ravishankar.sarawadi@intel.com>
Change-Id: I97c3d82f654be30a0d2d88cb68c8212af3d6f767
Reviewed-on: https://review.coreboot.org/16587
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This patch sets the package power limit (PL1) value in RAPL MSR
and disables MMIO register. Added configurable PL1 override
parameter to leverage full TDP capacity.
BUG=chrome-os-partner:56922
TEST=webGL performance(fps) not impacted before and after S3.
Change-Id: I34208048a6d4a127e9b1267d2df043cb2c46cf77
Signed-off-by: Sumeet Pawnikar <sumeet.r.pawnikar@intel.com>
Signed-off-by: Venkateswarlu Vinjamuri <venkateswarlu.v.vinjamuri@intel.com>
Reviewed-on: https://review.coreboot.org/16884
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
In FSP1.1 all the platform resets including global was handled
on its own without any intervention from coreboot.
In FSP2.0, any reset required will be notified to coreboot
and it is expected that coreboot will perform platform reset.
Hence, implement platform global reset hooks in coreboot. If Intel
ME is in non ERROR state then MEI message will able to perform
global reset else force global reset by writing 0x6 or 0xE to
0xCF9 port with PCH ETR3 register bit [20] set.
BUG=none
BRANCH=none
TEST=Verified platform global reset is working with MEI
message or writing to PCH ETR3.
Change-Id: I57e55caa6d20b15644bac686be8734d9652f21e5
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Signed-off-by: Rizwan Qureshi <rizwan.qureshi@intel.com>
Reviewed-on: https://review.coreboot.org/16903
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
As per ME BWG, there are two mechanism to generate a Global
Reset (resets both host and Intel ME), one is through CF9h
IO write of 6h or Eh with "CF9h Global Reset" (CF9GR) bit set,
PMC PCI offset ACh[20]. Another is to issue the Global Reset
MEI message. Because any attempts to cause global reset without
synchronizing the two sides might cause unwanted side effects,
such as unwritten flash data that will get destroyed if the
host were to cause a global reset without informing Intel ME
firmware, the recommended method is to send a Global Reset MEI
message when the following conditions are met:
The PCH chipset firmware just needs to complete the Intel ME
Interface #1 initialization and check the Intel ME HFSTS state
if Intel ME is not in ERROR state and is accepting MEI commands
then firmware should be able to use Global Reset MEI message to
trigger global reset.
Furthermore, if Intel ME is in ERROR state, BIOS can use I/O 0xCF9
write of 0x06 or 0x0E command with PCH ETR3 register bit [20]
to perform the global reset.
BUG=none
BRANCH=none
TEST=Verified Global Reset MEI message is able to perform platform
global issue in ME good state.
Change-Id: If326a137eeadaa695668b76b84c510e12c546024
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Signed-off-by: Rizwan Qureshi <rizwan.qureshi@intel.com>
Reviewed-on: https://review.coreboot.org/16902
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This patch programs and enables BAR for ME (bus:0/
device:0x16/function:0) device to have early ME communication.
BUG=none
BRANCH=none
TEST=Verified Global Reset MEI message can able to perform platform
global reset during romstage.
Change-Id: I99ce0ccd42610112a361a48ba31168c9feaa0332
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/17016
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
VBOOT_EC_SLOW_UPDATE should be selected if EC_GOOGLE_CHROMEEC is used as
building coreboot with Chrome OS support & without Chrome EC gives a
build error in coreboot.
Change-Id: I77eed0e1bdc1ba49381b72e21b0e18f573cadff0
Signed-off-by: Naresh G Solanki <naresh.solanki@intel.com>
Reviewed-on: https://review.coreboot.org/17020
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
The clearing of the PMC registers was not being called resulting
in state persisting across reboots. This state is queried and
events are added to the eventlog like 'RTC reset' events. However,
the RTC reset event is a one time thing so it should only be logged
once. Without the clearing of the state the event was logged on
every boot.
BUG=chrome-os-partner:58496
Change-Id: I60aa7102977c2b1775ab8c54d1c147737d2af5e2
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/17027
Reviewed-by: Furquan Shaikh <furquan@google.com>
Tested-by: build bot (Jenkins)
Reviewed-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
When timestamp is enabled, the system hangs because the timestamp data
is not yet available. Add a temporary work around that starts the
timestamp after the FspInit() making this data available.
Verified on Intel Camelback Mountain CRB and ensured that system can
boot to payload with timpstamp feature enabled.
Change-Id: I59c4bb83ae7e166cceca34988d5a392e5a831afa
Signed-off-by: York Yang <york.yang@intel.com>
Reviewed-on: https://review.coreboot.org/16894
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
The enforced FSP 1.0 APIs call was used to work around an fsp1.0 driver
issue. As the issue has been addressed in fsp1.0 driver (Change 9780),
remove the enforced workaround. Otherwise will see error message
'FSP API NotifyPhase failed' in serial log.
Verified on Intel Camelback Mountain CRB and confirmed that the serial
log error message regarding the 'FSP API NotifyPhase failed' is gone.
Change-Id: Iafa1d22e2476769fd841a3ebaa1ab4f9713c6c39
Signed-off-by: York Yang <york.yang@intel.com>
Reviewed-on: https://review.coreboot.org/16892
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Though we don't use Type-C PHY to support USB3 in firmware,
we still need to initialize the Type-C PHY, and make sure
the power state of pipe is always fixed to U2/P2. After
this, we can force USB3 controller to work in USB2 only
mode.
BRANCH=none
BUG=chrome-os-partner:56425
TEST=Go to recovery mode, plug a Type-C USB drive containing
chrome OS image into both ports in all orientations, check if
system can boot from USB.
Change-Id: I95bb96ff27d4fecafb7b2b9e9dc2839b5c132654
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: 8ec9850784
Original-Change-Id: Ie3654cd1c1cb76b62aa9b247879b60cbecee0155
Original-Signed-off-by: William wu <wulf@rock-chips.com>
Original-Reviewed-on: https://chromium-review.googlesource.com/391412
Original-Commit-Ready: Julius Werner <jwerner@chromium.org>
Original-Tested-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/16910
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Converged Security Engine (CSE) has a secure variable storage feature.
However, this storage is expected to be reset during S3 resume flow.
Since coreboot does not use secure storage feature, disable HECI2 reset
request. This saves appr. 130ms of resume time.
BUG=chrome-os-partner:56941
BRANCH=none
TEST=powerd_dbus_suspend; resume; check time with cbmem -t. Note
FspMemoryInit time is not significantly different from normal boot
time case.
Change-Id: I485a980369c6bd97c43b9e554d65ee89e84d8233
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/16870
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
This patch enables stage cache to save ~40ms during S3 resume.
It saves ramstage in the stage cache and restores it on resume
so that ramstage does not have to reinitialize during the
resume flow. Stage cache functionality is added to postcar stage
since ramstage is called from postcar.
BUG=chrome-os-partner:56941
BRANCH=none
TEST=built for Reef and tested ramstage being cached
Change-Id: I1551fd0faca536bd8c8656f0a8ec7f900aae1f72
Signed-off-by: Brandon Breitenstein <brandon.breitenstein@intel.com>
Reviewed-on: https://review.coreboot.org/16833
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
This reverts commit 28821dbb22.
(https://review.coreboot.org/16649)
This change causes the kernel to boot really slow. Maybe there is an
interrupt storm that prevents the kernel from making any
progress. Reverting until the proper kernel dependency is met.
BUG=chrome-os-partner:57364
BRANCH=None
TEST=Kernels boots to prompt fine on DVT.
Change-Id: I1c9913b4476a08303f9dd887b8631601c847dcf7
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: d7014ee1bb
Original-Change-Id: I061c0b03b43b516a190b370c04888e73a410fcf1
Original-Signed-off-by: Furquan Shaikh <furquan@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/391233
Original-Reviewed-by: Duncan Laurie <dlaurie@google.com>
Reviewed-on: https://review.coreboot.org/16881
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Martin Roth <martinroth@google.com>
CL:377541 was supposed to remove the big CPU cluster initialization from
rkclk_init() in the bootblock and move it to a more suitable place in
ramstage. Except that next to all the code cleanup I did in that patch,
I seem to have forgotten to actually remove that old code.
Big thanks to Nico for spotting that in the upstream coreboot review.
BRANCH=gru
BUG=chrome-os-partner:54906
TEST=Booted Kevin.
Change-Id: I09fe948b4587536802b42329b813177439e0804f
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: 77f9eaf044
Original-Change-Id: I13dab208225b7e43ad864f2f3cf51b3c104acd4b
Original-Reported-by: Nico Huber <nico.h@gmx.de>
Original-Signed-off-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/389236
Original-Reviewed-by: Douglas Anderson <dianders@chromium.org>
Reviewed-on: https://review.coreboot.org/16769
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
This selects the rank to train before training is triggered. This is
to prevent any race conditions with the hardware.
BRANCH=none
BUG=chrome-os-partner:56940
TEST=stressapptest -M 1536 -s 1000
Change-Id: I892bace414cf4495619d41bdaea0c4e91c1e29b3
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: 8f2dd6f529
Original-Change-Id: I4e7118d8509b59e391d0a254477b5390dfdd43a5
Original-Signed-off-by: Derek Basehore <dbasehore@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/387907
Original-Commit-Ready: Douglas Anderson <dianders@chromium.org>
Original-Tested-by: Douglas Anderson <dianders@chromium.org>
Original-Reviewed-by: Douglas Anderson <dianders@chromium.org>
Original-Reviewed-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-by: 云平 汤 <typ@rock-chips.com>
Reviewed-on: https://review.coreboot.org/16768
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Martin Roth <martinroth@google.com>
There are two modifications in the driver:
1. Correctly set speeds based on DDR frequency.
Control the speeds in the predriver circuits to reduce power.
SPEED[1:0]
2'b00:less than 800Mbps(400MHz)
2b01 : 800Mbps(400MHz) to 1600Mbps(800MHz)
2b10 : 1600Mbsp(800MHz) to 2400Mbps(1200MHz)
2b11 : 3200Mbps and greater
2. Configure the number of cycles for the phy clock pll wait time after
locking, based on the DDR config file.
BRANCH=none
BUG=chrome-os-partner:56940
TEST=do memtester on kevin board, and pass
Change-Id: Iaf6da59c6c5c290867e0922a2a99de272f4c7bde
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: 125cf8afac
Original-Change-Id: Iabc17df37a701c4f052540c3c259f209a1db3c59
Original-Signed-off-by: Lin Huang <hl@rock-chips.com>
Original-Reviewed-on: https://chromium-review.googlesource.com/387428
Original-Reviewed-by: Douglas Anderson <dianders@chromium.org>
Original-Reviewed-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/16722
Reviewed-by: Martin Roth <martinroth@google.com>
Tested-by: build bot (Jenkins)
In USB2 only mode, the Type-C PHY will be held in reset and
only the USB2 logic of the USB3 OTG controller and PHY will be
used over the USB2 pins on the Type-C connector to support Low,
Full and High-speed USB operation.
BRANCH=none
BUG=chrome-os-partner:56425
TEST=Go to recovery mode, plug a Type-C USB drive containing
chrome OS image into both ports in all orientations, check if
system can boot from USB.
Change-Id: Ic265c0c91c24f63b2f9c3106eb2bb277a589233b
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: a37ccc5b60
Original-Change-Id: I582f04f84eef447ff0ba691ce60e9461ed31cfad
Original-Signed-off-by: Liangfeng Wu <wulf@rock-chips.com>
Original-Reviewed-on: https://chromium-review.googlesource.com/385837
Original-Commit-Ready: Julius Werner <jwerner@chromium.org>
Original-Tested-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/16717
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
To improve sdram 800MHz and 933MHz stability, we
need to modify write leveling flow to get the
proper write leveling value.
BUG=chrome-os-partner:56940
BRANCH=none
TEST=Boot from kevin on 933MHz, and do stressapptest
Change-Id: I5b24c93d4a57917fb9af7e5e2a95d8423ccbaa7e
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: d84bf25b3e
Original-Change-Id: I87efddf628c3683fcb85d6875e029cf3cbc482be
Original-Signed-off-by: Jianqun Xu <jay.xu@rock-chips.com>
Original-Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>
Original-Signed-off-by: Lin Huang <hl@rock-chips.com>
Original-Reviewed-on: https://chromium-review.googlesource.com/384292
Original-Commit-Ready: Julius Werner <jwerner@chromium.org>
Original-Tested-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/16716
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
We found that we may want to load some components of BL31 on the RK3399
into SRAM. As usual, these components may not overlap any coreboot
regions still in use at that time, as is already statically checked by
the check-ramstage-overlaps rule in Makefile.inc.
On RK3399, the only such regions are TTB and STACK. This patch moves the
TTB region back to the end of SRAM (right before STACK), so that a large
contiguous region of SRAM before that remains usable for BL31.
BRANCH=gru
BUG=None
TEST=Booted Kevin.
Change-Id: I1689d0280d79bad805fea5fc3759c2ae3ba24915
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: 1d4c6c6f6c
Original-Change-Id: I37c94f2460ef63aec4526caabe58f35ae851bab0
Original-Signed-off-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/384635
Original-Reviewed-by: Simon Glass <sjg@google.com>
Reviewed-on: https://review.coreboot.org/16714
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
With a SPI clock above about 24MHz the APB cannot keep up when doing
individual byte transfers. Adjust the driver to use 16-bit reads when
it can, to remove this bottleneck.
Any transaction which involves writing bytes still uses 8-bit transfers,
to simplify the code. These are the transfers that are not time-critical
since they tend to be small. The case that really matters is reading from
SPI flash.
In general we can use 16-bit reads anytime we are transferring an even
number of bytes. If the code detects an odd number of bytes, it tries to
perform the operation in two steps: once in 16-bit mode with an even
number of bytes, and once in 8-bit mode for the final byte. This allow
us to use 16-bit reads even if asked to transfer (for example) 0xf423
bytes.
The limit on in_now and out_now is adjusted to 0xfffe to avoid an extra
transfer when transferring ~>=64KB.
CQ-DEPEND=CL:383232
BUG=chrome-os-partner:56556
BRANCH=none
TEST=boot on gru and see that things still work correctly. I tested (with
extra debugging) that the 16-bit case is being picked when it should be.
Change-Id: If5effae9a84e4de06537fd594bedf7f01d6a9c88
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: ec250b4931
Original-Change-Id: Idc5b7e5d82cdbdc1e8fe8b2d6da819edf2d5570c
Original-Signed-off-by: Simon Glass <sjg@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/381312
Original-Commit-Ready: Julius Werner <jwerner@chromium.org>
Original-Tested-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/16712
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Martin Roth <martinroth@google.com>
Some of the asserts for valid clock divisor ranges were off by one. This
patch corrects them and writes them all in a consistent way.
BRANCH=None
BUG=None
TEST=Booted Kevin.
Change-Id: I81749408a40822100797f1734f3b88987d12d8d5
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: e09cdfde26
Original-Change-Id: I429edb99e2d5ff2302d9750e6569b3d21f5686fa
Original-Signed-off-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/381574
Original-Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-on: https://review.coreboot.org/16704
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
This patch moves the big CPU cluster initialization on the RK3399 from
the clock init bootblock function into ramstage. We're only really doing
this to put the cluster into a sane state for the OS, we're never
actually taking it out of reset ourselves... so there's no reason to do
this so early.
Also cleaned up the interface for rkclk_configure_cpu() a bit to make it
more readable.
BRANCH=None
BUG=chrome-os-partner:54906
TEST=Booted Kevin.
Change-Id: I568b891da0abb404760d120cef847737c1f9e3ec
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: bd7aa7ec3e
Original-Change-Id: Ic3d01a51531683b53e17addf1942441663a8ea40
Original-Signed-off-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/377541
Original-Reviewed-by: Douglas Anderson <dianders@chromium.org>
Reviewed-on: https://review.coreboot.org/16698
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
At higher SPI bus speeds the SPI RX value is not available in time for
sampling at the normal time. Add a delay to ensure that we read the
correct data.
The value of 40ns is chosen arbitrarily. In my testing I can use a sample
delay of 1 even at 24MHz. But since it is not necessary, I have left that
case alone. It kicks in at 25MHz and up.
BUG=chrome-os-partner:56556
BRANCH=none
TEST=boot on gru and see no change at current speed
Change-Id: I3ef335d9a532eaef1e76034bd02e185acf11176a
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: e9b620c47f
Original-Change-Id: I65d66d752cbbbee4d02f475de23a52069a0e9782
Original-Signed-off-by: Simon Glass <sjg@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/381311
Original-Commit-Ready: Julius Werner <jwerner@chromium.org>
Original-Tested-by: Simon Glass <sjg@google.com>
Original-Reviewed-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/16707
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
This patch fixes a typo in the clock initialization code that caused the
PERILP1_PCLK_HZ constant to be ignored and the clock to always run at
the same speed as its parent (PERILP1_HCLK_HZ). Since we've done all our
previous tests and validation with this bug, we should probably increase
the value of the constant (that had not actually been used) to the value
that we had been incorrectly using instead (which also makes effective
SPI read times faster).
BRANCH=None
BUG=chrome-os-partner:56556
TEST=Booted Kevin.
Change-Id: Ibeb08f5fe5e984a74e3f57e60c62d4bfb644b6ca
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: 06e605a5fc
Original-Change-Id: Icb5e079f53eb22b0dbf0ea4d1c2ff08688e3fa8e
Original-Signed-off-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/381031
Original-Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-on: https://review.coreboot.org/16703
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Output GPIOs should never have a pull-up or pull-down resistor attached
since they're actively driven. Since some GPIOs get initialized with a
pull at power-on reset, we should explicitly overwrite that setting.
Most other platforms do this on gpio_output, but Rockchip hadn't yet.
Also, shuffle some code around to make things cleaner and allow for
easier code reuse.
BRANCH=None
BUG=chrome-os-partner:52526
TEST=Booted Kevin.
Change-Id: I1425d074ea1e90f4484e1e84a8002b057192c5f7
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: df5b236bfd
Original-Change-Id: I044266d71ef8bd0518316ff72d829d1ca1e30f35
Original-Signed-off-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/382531
Original-Reviewed-by: Simon Glass <sjg@google.com>
Reviewed-on: https://review.coreboot.org/16710
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
As far as I know, the Cortex-A53 cores in RK3399 are of a newer revision
that is not affected by ARM erratum 843419. If it was, the workaround
would also need to be enabled in libpayload and Chrome OS userspace,
which it currently isn't. I assume this was just incorrectly copied over
from another SoC and we can safely remove it.
BRANCH=None
BUG=chrome-os-partner:56700
TEST=Booted Kevin.
Change-Id: I5b1534c954a6d985499b481738723cabbdc07253
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: 4891cc8665
Original-Change-Id: Iadb57428f8727ce0e563204723644e2c79e3007c
Original-Signed-off-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/376363
Original-Commit-Queue: Douglas Anderson <dianders@chromium.org>
Original-Reviewed-by: Douglas Anderson <dianders@chromium.org>
Reviewed-on: https://review.coreboot.org/16702
Tested-by: build bot (Jenkins)
Reviewed-by: Martin Roth <martinroth@google.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Kconfig hex values don't need to be in quotes, and should start with
'0x'. If the default value isn't set this way, Kconfig will add the
0x to the start, and the entry can be added unnecessarily to the
defconfig since it's "different" than what was set by the default.
A check for this has been added to the Kconfig lint tool.
Change-Id: I86f37340682771700011b6285e4b4af41b7e9968
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/16834
Tested-by: build bot (Jenkins)
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
