We have a timestamp from before cache-as-ram setup saved in the MMX
registers. Recover that timestamp, and use it as the base timestamp
rather than letting lib/bootblock.c use a late timestamp.
This allows more accurate profiling of the boot flow, since CAR setup
time is no longer excluded from the timing information.
Change-Id: I055092c600438c5260ab67509434a38f1eb77fe4
Signed-off-by: Alexandru Gagniuc <alexandrux.gagniuc@intel.com>
Reviewed-on: https://review.coreboot.org/14863
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
The timer emulation works by deriving a frequency based off the
Common Timer Copy with a frequency of 19.2MHz.
The desired frequency = (19.2MHz * multiplier) >> 32;
With that knowledge update the code to let the compiler perform
the necessary math based on target frequency.
Change-Id: I716c7980f0456a7c6072bbaaddd6b7fcd8cd5b37
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/14889
Tested-by: build bot (Jenkins)
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
This reverts commit e976bd4469.
The LPC resource allocation will be completely reworked in subsequent
patches. The most straightforward approach is to start by reverting
the existing code.
Change-Id: I2475542b79817020d4c956f22ed5856f05046b16
Reviewed-on: https://review.coreboot.org/14583
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
The pad for CS2 of the Fast SPI interface needs to be configured for
automatic MMIO translation when a SPI TPM is utilized. Instead of
unconditionally configuring that pad under LPC_TPM provide a explicit
Kconfig for a mainboard to select.
Change-Id: Ia94b90e12d71a4b849359188a853f7e036cc583b
Signed-off-by: Aaron Durbin <adurbin@chormium.org>
Reviewed-on: https://review.coreboot.org/14531
Reviewed-by: Andrey Petrov <andrey.petrov@intel.com>
Tested-by: build bot (Jenkins)
Enable caching of BIOS region with variable MTRR. This is most
useful if enabled early such as in bootblock.
Change-Id: I39f33ca43f06fce26d1d48e706c97f097e3c10f1
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/14480
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
One of devices connected to FAST SPI bus is TPM. SoC uses dedicated
line for chip select for TPM function. If TPM is used, that line needs
to be configured to a specific native funciton.
Change-Id: Ib5bf4c759adf9656f7b34540d4fc924945d27a97
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/14467
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Set the default memory type in MTRRCap register to 0. This ensures
that even if the MTRR Enable bit is set in MTRRCap register, the
default memory type is still uncacheable.
Change-Id: I63e7993f8b65dabbab60e7c1bb8d6d89ef4da9ee
Signed-off-by: Furquan Shaikh <furquan@google.com>
Reviewed-on: https://review.coreboot.org/14428
Tested-by: build bot (Jenkins)
Reviewed-by: Andrey Petrov <andrey.petrov@intel.com>
BIST result is already stored by arch/x86/bootblock_ctr0.S in
mm0. Also, eax does not contain BIST result by the time control
reaches bootblock_pre_c_entry. bootblock_crt0.S saves timestamp in mm2
which was being overwritten here. Thus, remove the saving of BIST
result from SoC code.
Change-Id: I65444689cf104c59c84574019f5daf82aab10bc7
Signed-off-by: Furquan Shaikh <furquan@google.com>
Reviewed-on: https://review.coreboot.org/14381
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Andrey Petrov <andrey.petrov@intel.com>
Update all of the license headers to make sure they are compliant
with coreboot's license header policy.
Change-Id: I151d058615290e528d9d1738c17804f6b9cc8dce
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: https://review.coreboot.org/14321
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
The flush L1D to L2 operation was only being used when loading
romstage from bootblock. However, when the FSP-M component is
loaded no code coherency actions are taken. I suspect this is
because the FSP-M component is larger than the 24KiB L1D and
the entry point is early in the image. Thus, when loading
the FSP-M component the earlier part of the image is flushed
out to L2 in the process of loading the latter part of the
component. Also, once verstage is introduced the same
code coherency actions need to be taken as well. Therefore,
position the apollolake code to handle all these cases.
Change-Id: Ie71764f1b420a6072c4f149ad3e37278b6cb70e1
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/14210
Tested-by: build bot (Jenkins)
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Andrey Petrov <andrey.petrov@intel.com>
The current Apollolake flow has its code executing out of
cache-as-ram for the pre-DRAM stages. This is different from
past platforms where they were just executing-in-place against
the memory-mapped SPI flash boot media. The implication is
that when cache-as-ram needs to be torn down one needs to be
executing out of DRAM since the act of cache-as-ram going
away means the code disappears out from under the processor.
Therefore load and use the postcar infrastructure to bootstrap
this process for tearing down cache-as-ram and subsequently
loading ramstage.
Change-Id: I856f4b992dd2609b95375767bfa4fe64a267d89e
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/14141
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Provide a bootblock_soc_early_init() to that takes care of
initializing the UART on behalf of the mainboard when serial
console is enabled.
Change-Id: I2d3875110b6f58a9e0b4c113084b85817aa05a87
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/13793
Tested-by: build bot (Jenkins)
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Andrey Petrov <andrey.petrov@intel.com>
Once bootblock copied romstage into CAR it may not jump into it right
away. This is because we are in NEM mode, there is no backing store
and a miss in L1 may cause L1D line snoop that gets written back. The
solution is to flush L1D to L2 so snoop guaranteed to hit L2.
Change-Id: I2ffe46dbfdfe7f0ccd38b34ff203ff76b6d5755b
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/13703
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Early UART driver is for bootblock and romstage. It is supposed to be used
when BOOTBLOCK_CONSOLE is enabled. This also adds few configuration bits
in bootblock requiered for serial to be set up.
Change-Id: I15520d566f107797e68d618885d4379e73d0fa45
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/13677
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This is the minimum setup needed to both get cache-as-ram setup and a
C environment working. On apollolake, we only get 32 KiB of data
loaded into an SRAM that is readonly to the main CPU. Due to this
restriction we have to set CAR and a C environment very early on.
Change-Id: I65c51f972580609d2c1f03dfe2a86bc5d45d1e46
Signed-off-by: Alexandru Gagniuc <alexandrux.gagniuc@intel.com>
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/13301
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
