acpi_is_wakeup_s3() was introduced in upstream coreboot
while the FSP support code was written. Move to using
that instead of using the romstage_handoff structure
directly.
BUG=chrome-os-partner:43636
BRANCH=None
TEST=Built, booted, suspended, and resumed on glados.
Original-Change-Id: I71601a4be3c981672e25e189c98abb6a676462bf
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/290720
Original-Reviewed-by: Leroy P Leahy <leroy.p.leahy@intel.com>
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: I2ae4d9906e0891080481fb58b941921922a989d3
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11190
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Explicitly clear all write-1-to-clear fields in the
appropriate power state registers. That way stale
state isn't left around from boot to boot. The
MMIO PMC registers are always added such that the
resource can be accessed from reg_script. It doesn't
hurt to add the resource, and it's actually more
informative by attaching the actual resources
owned by the device.
BUG=chrome-os-partner:43625
BRANCH=None
TEST=Built and boot glados. Did global reset. Noticed bits
set. Did normal reset and saw those same bits no longer set.
Original-Change-Id: Idd412bd6bf2c6c57b46c74f9411bdf8413ddd83e
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/290339
Change-Id: Ibef1aefedf6ba006f17f9f94998a10b39cc6bfec
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11186
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Leaving a sentinel 0xC0DEBABE and fixing it up is
is the old way of setting the correct base address
for GNVS. One just needs to reference NVSA which is
already filled in by the skylake ACPI code.
BUG=chrome-os-partner:43611
BUG=chrome-os-partner:43522
BRANCH=None
TEST=Built and booted glados. /sys/firmware/log shows
up as well as ramoops using the correct address.
Original-Change-Id: I1d4979b1bb65faa76316a4ec4c551a7b9b9eed32
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/290338
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: I25efea73a383215f9365ce91230f79516b0201a6
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11185
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Provide #defines for the bit fields in the SMI status register.
This allows for one to set the callback accordingly without
hard coding the index.
BUG=chrome-os-partner:43522
BRANCH=None
TEST=Built and booted glados.
Original-Change-Id: I3e61d431717c725748409ef5b543ad2eb82955c4
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/289802
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: I1a91f2c8b903de4297aaa66f5c6ff15f1b9c54f6
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11184
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
DISB (bit 23) in GEN_PMCON_A represents to MRC that DRAM
training is complete. However, as a 8-bit write was
being performed the bit was never being set.
BUG=chrome-os-partner:43516
BRANCH=None
TEST=Built and booted to kernel. Rebooted. Noted full memory
training was not being peformed.
Original-Change-Id: If2a9cc2f80bc38ea86fb0d7ff855ef95540b561b
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/290337
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: Ic7973e0ec279304797e0b3d83d7378f620f2b548
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11183
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Open coding bitfields is really annoying as no one knows
what they are unless you have a doc in front of you.
Fill in the bitfields for the GEN_PMCON_A and GEN_PMCON_B
registers.
BUG=chrome-os-partner:43522
BRANCH=None
TEST=Built and booted glados.
Original-Change-Id: Id48de68eaa3896c17d5da2ffb0bcf17062f73e5e
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/290336
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: I968be9736419e26a771e0a0c3c964d540fbb1efe
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11182
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
In order to run with the debug FSP the SMBus device needs
to be enabled. Additionally, the TCO block lives within
the SMBus device so if TCO is to be employed then the
SMBus device needs to be enabled as a prerequisite.
BUG=chrome-os-partner:42407
BRANCH=None
TEST=Buit and booted into kernel.
Original-Change-Id: I269650fa5222b4741ef495188dff1f4b8176fe89
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/290364
Original-Reviewed-by: Bernie Thompson <bhthompson@chromium.org>
Original-Reviewed-by: Robbie Zhang <robbie.zhang@intel.com>
Change-Id: Ia1f72ea7bd70728de83cdff07df9810a326266c2
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11181
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
FSP was setting up the TCO registers to be mapped at 0x400.
However, the SMBus initialization in romstage was mapping
its I/O BAR to 0x400 as well. The result seemed to cause the
TCO register to be hidden. However, the board was rebooting in
depthcharge when the SMBus device was enabled from a TCO timeout.
As the TCO timer was halted before the double resource assignment
it's not clear how the TCO was getting re-enabled. In either case,
the current behavior is wrong.
BUG=chrome-os-partner:42407
BRANCH=None
TEST=Built and booted glados w/ SMBus enabled.
Original-Change-Id: I43c0d67a76abac51ccfd5105245792981fbcd04c
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/290363
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: I3839290768c27626c3fd2d67d5de94c291c1386e
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11180
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Instead of relying on FSP to do gpio configuration in one
place use the native support in coreboot. This also removes
the open coded configuration of the memory configuration
ids.
BUG=chrome-os-partner:42982
BRANCH=None
TEST=Built and booted glados.
Original-Change-Id: I4655221d821d91a2270d774305a02d6bd5c3959c
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/289800
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: I2e66242d050c3825f6bc65d3d2c7f51d2cdfbd73
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11175
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
It's important to be able to configure the gpio pads at
various stages instead of a single place using FSP. Without
this support there is a lot of duplicated open-coded pad
configuration taking place both within the SoC code and
mainboards.
Current limitation is that all GPIOs are in ACPI mode. i.e.
The HostSW ownership register sets the pad configuration to
only update GPI_GPE_STS, GPI_NMI_STS and/or GPI_SMI_STS. The
GPI_STS update is masked within the GPIO community registers.
BUG=chrome-os-partner:42982
BRANCH=None
TEST=Built and booted glados.
Original-Change-Id: Id8a00e99c7a4c3912de2feaff9cea12b402f2c68
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/289789
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: I4c86b47ac5ab004f2bfd7cb07dd23c458f7dbb7c
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11174
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
In the wake of the recent Intel "Memoy Sinkhole" exploit a code review
of the AMD SMM code was undertaken. While native Family 10h support
does not appear to be affected by the same SMM flaw, it also does not
require SMM to function. Therefore, the SMM memory range initialization
should only be executed if SMM will be used on the target platform.
Change-Id: I6531908a7724933e4ba5a2bbefeb89356197e8fd
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/11211
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Many Kconfig options changed in coreboot.org since
skylake was first started. Fix Kconfig option name
changes, and also provide a common option, UART_DEBUG
that can be selected to select all the necessary
options.
Note: It's still a requirement to manually unset the
8250IO option because that's unconditionally set.
BUG=chrome-os-partner:43419
BUG=chrome-os-partner:43463
BRANCH=None
TEST=Built glados. Booted into kernel. Kernel reboots somewhere.
Original-Change-Id: I9e6549ea0f1d6b9ffe64a73856ec87b5bc7b7091
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/289951
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: I0e6b492d7279cc35d4fb3ac17fd727177adce39d
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11172
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Enable the Deep Sx pins to allow wake from the EC via LAN_WAKE#.
Report the EC wake pin LAN_WAKE as GPE[112].
BUG=chrome-os-partner:43079
BRANCH=none
TEST=suspend/resume on glados with wake from keyboard
Original-Change-Id: I99664e1e406d15e7460046a6168cbd3a377aaca4
Original-Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/288921
Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Change-Id: I19db144ed5db183f47af03340886a5e770af8bc8
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/11171
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Add support for enabling various pins in Deep Sx by setting
a register in the mainboard devicetree.
BUG=chrome-os-partner:43079
BRANCH=none
TEST=build and boot on glados
Original-Change-Id: I1b4fb51f72b88bdc49096268bdd781750dcd089d
Original-Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/288920
Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Change-Id: I7555a92fecc6e78b579ec0bc18da202cb0c824e2
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: http://review.coreboot.org/11170
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
There was no implementation for uart_fill_lb() in the 8250mem
driver. Rectify this so when 8250MEM and CONSOLE_SERIAL are
employed then the build doesn't fail.
BUG=chrome-os-partner:43419
BRANCH=None
TEST=Built with glados using 8250MEM
Original-Change-Id: I35d6b15e47989c1854ddcee9c6d46711edffaf3e
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/289899
Original-Reviewed-by: Patrick Georgi <pgeorgi@chromium.org>
Change-Id: I972b069a4def666f509268816de91ed6c0f655d9
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11169
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
CBFS_SIZE is living as a mainboard attribute. Because
of the Kconfig include ordering the SoC *cannot* set
the default. Remove from the soc Kconfig and add a
default Kconfig for SOC_INTEL_SKYLAKE.
BUG=chrome-os-partner:43419
BRANCH=None
TEST=built glados
Original-Change-Id: I8808177b573ce8e2158c9e598dbfea9ff84b97c7
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/289833
Original-Reviewed-by: Martin Roth <martinroth@google.com>
Change-Id: Icf52d7861eee016a35be899e5486deb0924a0f3c
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11168
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
The asl_template previously unconditionally included
dsdt.aml. However, COMPILE_IN_DSDT=y results in the
dsdt.aml being linked directly into ramstage. Thus
the information is duplicated.
The inclusion of this file unconditionally throws
some errors as certain assets need to be included
in CBFS. However, as there isn't fine-grained
ordering control in how files are added fixed
resource requirements for other assets collide
result in failure to build.
To remedy both things, provide a 2nd argument to
asl_template which defaults to 'y' for CBFS
addition. In the COMPILE_IN_DSDT=y case pass
'n' so that dsdt.aml is no longer added.
BUG=chrome-os-partner:43419
BRANCH=None
TEST=For glados:
Built with COMPILE_IN_DSDT=y. dsdt.aml not included.
Built with COMPILE_IN_DSDT=n. dsdt.aml was included.
Original-Change-Id: I4767e5be2915c1732251fe415017f30314c5efc9
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/289840
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: Id1828627ba0a034eb05b2fe23be76e19f3040444
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11166
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Tested-by: build bot (Jenkins)
The sysinfo object within the k8 ram init is used
to communicate progess/status from all the nodes in the
system. However, the code was assuming where the sysinfo
object lived in cache-as-ram. The layout of cache-as-ram
is dynamic so one needs to do the lookup of the correct
address at runtime. The way the amd code is compiled
by #include'ing .c files makes the solution a little
more complex in that some cache-as-ram support code
needed to be refactored.
Change-Id: I6500fa7b005dc082c4c0b3382ee2c3a138d9ac31
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/10961
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
One may prefer to include vboot from another directory than 3rdparty for
convenience. This is especially the case in Libreboot, where 3rdparty is not
checked out at all.
Change-Id: I13167eb604a777a2ba87c3567f134ef3ff9610e4
Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
Reviewed-on: http://review.coreboot.org/11116
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
$(obj) might be defined either as a relative or an absolute path. Thus, it has
to be filtered out before adding $(top) to it (in case of an absolute path) when
building vboot. It is then provided separately in CFLAGS (as an absolute path).
In addition, VB2_LIB inherits $(obj), so it might also already be an absolute
path, and prefixing $(top) to it doesn't apply. Thus, the absolute path to it
should be passed to the vboot make command.
Change-Id: I13e893ebdf22c4513ee40d9331a30ac7de8f9788
Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
Reviewed-on: http://review.coreboot.org/11120
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
As acpi_write_hpet() uses CONFIG_HPET_ADDRESS in the HPET table we
need to use CONFIG_HPET_ADDRESS when assigning it to the device.
Change-Id: I656f917658f1c1717bb3653fa048a6d36fca2454
Signed-off-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
Reviewed-on: http://review.coreboot.org/10925
Tested-by: build bot (Jenkins)
Reviewed-by: Idwer Vollering <vidwer@gmail.com>
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Related-to: I3175c8b29e94a27a2db6b11f8fc9e1d91bde11f9
(ACPI: Fix corrupt SSDT table on multiprocessor AMD Family 10h systems)
Change-Id: I0b5f265278d90cbaeddc6fc4432933856050f784
Signed-off-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
Reviewed-on: http://review.coreboot.org/10912
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
This should probably be moved out of lib and to arch/x86,
since it does not even apply on x86-64, and ARM has its
own copy of libgcc.
Change-Id: I4fca1323927f8d37128472ed60d059f7a459fc71
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: http://review.coreboot.org/11110
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Spike support: QEMU RISCV is broken, and the maintainers at Berkeley
are working on it, but at the moment spike is the only way to test
on riscv. Add support for spike console output for debugging.
Privileged ISA: Update to privileged ISA in RISCV (machine,
supervisor, hypervisor, user modes) broke exisitng RISCV asm, and
bootblock.S was updated to match the new spec. Clean old assembly
[pg: things build with gcc 4.9 now, but don't expect them to work.
Hardcoding register names into the assembler language may not be the smartest
idea of the RISCV folks.]
Change-Id: Ie2c109d3c26712c207512f74f28ce1a925e6e181
Signed-off-by: Thaminda Edirisooriya <thaminda@google.com>
Reviewed-on: http://review.coreboot.org/11078
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
The spec states (5.2.10): "The BIOS aligns the FACS on a 64-byte boundary
anywhere within the system's memory address space."
Change-Id: Ie9415e505525dbdd418028d4954018c829921a18
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Found-by: fwts 15.08
Reviewed-on: http://review.coreboot.org/11141
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
The AMD AGESA binaryPI sources were incorrectly committed to
3rdparty/blobs. Move them from blobs to vendorcode and fix
Kconfig and Makefile.inc to match.
Change-Id: I55a777553c1203464d7f7f4293b361fedcfa3283
Signed-off-by: Marc Jones <marc.jones@se-eng.com>
Reviewed-on: http://review.coreboot.org/10982
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
The *_SELECTED Kconfig variables are not needed with the
options contained within "if CPU_AMD_AGESA_BINARY_PI"
introduced in e4c17ce8. It also removes the need to
source and select the default prior to selecting the
AGESA source or AGESA PI option.
Change-Id: Iffa366f575f7f155bd6c7e7ece2a985f747c83be
Signed-off-by: Marc Jones <marc.jones@se-eng.com>
Reviewed-on: http://review.coreboot.org/10981
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
On x86-64 the current way of calculating the base address
of the boot device (SPI flash) gets an unwanted sign extension,
making it live somewhere at the end of 64bit address space.
Enforce rom_base to be at the upper end of the 4G address space.
Change-Id: Ia81e82094d3c51f6c10e02b4b0df2f3e1519d39e
Signed-off-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-on: http://review.coreboot.org/11121
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
