These functions can do all kinds of things, such as converting an ELF
image into SELF, or (in the future) compress or checksum entire files.
This may require changing or adding fields to the header, so they
need to have access to it.
The header_size parameter that was provided (but never used) is
equivalent to cbfs_file's offset field.
Change-Id: I7c10ab15f3dff4412461103e9763a1d78b7be7bb
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/11325
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
It's sole use was comparing it to the header's "len" field.
Change-Id: Ic3657a709dee0d2b9288373757345a1a56124f37
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/11324
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
->len used to be set to the file data length plus the size of the
padding used for the cbfs_file header. This isn't the case anymore,
so no patching of this field is necessary anymore.
->offset still needs to be patched in that case because its final
value can only be determined when the file's actual location is known.
Change-Id: I1037885f81b4ed3b68898dd7d0e515cf7a9c90a8
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/11322
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Up to now cbfstool creates the cbfs_file header at the latest possible
time, which is unsuitable when the idea is to add further fields to it
that need to be configured earlier.
Thus, have it ripple up the call chain.
Change-Id: I7c160681c31818bc550ed2098008146043d0ee01
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/11320
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
"target", for what? It's the offset where the file header of the currently
added file will be located, name it as such.
Change-Id: I382f08f81991faf660e217566849773d9a7ec227
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/11319
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Refactor the code to be better understandable.
Change-Id: Ia815a27f7cc83c226a32e87485d712a5fbf4168e
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: Zheng Bao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/11318
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
The Kconfig symbol CACHE_MRC_BIN was getting forced enabled everywhere
it existed.
Remove the Kconfig symbol and get rid of the #if statements
surrounding the code.
This fixes the Kconfig warning for Haswell & Broadwell chips:
warning: (NORTHBRIDGE_INTEL_HASWELL &&
NORTHBRIDGE_INTEL_SANDYBRIDGE &&
NORTHBRIDGE_INTEL_SANDYBRIDGE_NATIVE &&
NORTHBRIDGE_INTEL_IVYBRIDGE &&
NORTHBRIDGE_INTEL_IVYBRIDGE_NATIVE &&
CPU_SPECIFIC_OPTIONS) selects CACHE_MRC_BIN
which has unmet direct dependencies
(CPU_INTEL_SOCKET_RPGA988B || CPU_INTEL_SOCKET_RPGA989)
Change-Id: Ie0f0726e3d6f217e2cb3be73034405081ce0735a
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: http://review.coreboot.org/11270
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
In testing other localization changes, I found that I couldn't build
anymore because xcompile wasn't picking up my toolchain. I traced it
to the regex comparison of '.*format \(.[a-z0-9-]*\)' to the string
'formato del fichero elf32-i386'. Forcing the localization of
objdump to C before doing the comparison fixes the issue.
Change-Id: I6bed5a9824807dd5bc5a38b711ab47e2af4b0c29
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: http://review.coreboot.org/11304
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-by: Zheng Bao <zheng.bao@amd.com>
There doesn't seem to have been an olddefconfig target in the coreboot
version of the Kconfig makefile. It's listed in the .PHONY, but it
doesn't seem like it's ever been there. This is useful for expanding
a miniconfig saved with 'make savedefconfig'.
Change-Id: I3798f8469135b58d32da68d4b0e434ab5351b501
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: http://review.coreboot.org/11273
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
When the check for global symbols in romstage happens, if everything is
good, a warning appears, telling us that the segment is empty. While the
empty segment is good, the warning is distracting:
"BFD: build/cbfs/fallback/romstage_null.debug: warning: Empty loadable
segment detected, is this intentional ?"
This change hides that particular warning, but shouldn't hide any other
output from objcopy.
Change-Id: If22489280712d02a61c3ee5e0cb2a53db87d6082
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: http://review.coreboot.org/11302
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Tested-by: build bot (Jenkins)
Some patch implementations (eg. BSD) create new files by taking the "---" file
name instead of the "+++" one, so set both to the file name that is to be
created.
Change-Id: I6f37748b4cf0852d292f8f5156fc27ab8fd481b6
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reported-by: Idwer Vollering <vidwer@gmail.com>
Reported-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
Reviewed-on: http://review.coreboot.org/11303
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
The AMD K8 northbridge uses the Kconfig symbol QRANK_DIMM_SUPPORT,
but the symbol was used on a number of Family 10 boards as well.
AMD Family 10 doesn't use this Kconfig symbol for anything.
I verified that the symbol wasn't used actually getting used in any
of these platforms.
Fixes Kconfig warnings for these 19 mainboards:
warning: (BOARD_SPECIFIC_OPTIONS...) selects QRANK_DIMM_SUPPORT which
has unmet direct dependencies (NORTHBRIDGE_AMD_AMDK8)
Change-Id: I454992a4975566fd6439a21f5a800d0cfa1b4d3b
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: http://review.coreboot.org/11300
Tested-by: build bot (Jenkins)
Reviewed-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
Add CHROMEOS dependencies to selects for the following Kconfig
symbols:
CHROMEOS_RAMOOPS_DYNAMIC
CHROMEOS_RAMOOPS_NON_ACPI
CHROMEOS_VBNV_CMOS
CHROMEOS_VBNV_EC
CHROMEOS_VBNV_FLASH
EC_SOFTWARE_SYNC
LID_SWITCH
RETURN_FROM_VERSTAGE
SEPARATE_VERSTAGE
VBOOT_DISABLE_DEV_ON_RECOVERY
VBOOT_EC_SLOW_UPDATE
VBOOT_OPROM_MATTERS
VBOOT_STARTS_IN_BOOTBLOCK
WIPEOUT_SUPPORTED
This gets rid of these sorts of Kconfig errors:
warning: BOARD_SPECIFIC_OPTIONS selects CHROMEOS_VBNV_EC which has
unmet direct dependencies (MAINBOARD_HAS_CHROMEOS && CHROMEOS)
Note: These two boards would never actually have CHROMEOS enabled:
intel/emeraldlake2 has MAINBOARD_HAS_CHROMEOS commented out
google/peach_pit doesn't have MAINBOARD_HAS_CHROMEOS
Change-Id: I51b4ee326f082c6a656a813ee5772e9c34f5c343
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: http://review.coreboot.org/11272
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
It still needs to pass --32 (yes, 32) to the assembler.
x86_64-linux does this (through some other config file),
x86_64-elf did not.
This fixes building SeaBIOS with our x86_64-elf multilib compiler.
Change-Id: Ibe2a70e46e64e71c947482be5ec0eaf7f7bf300d
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/11289
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
CHROMEOS is a user-visible bool. It must not be 'select'ed in Kconfig.
That's why we have MAINBOARD_HAS_CHROMEOS. This is the fifth time I
find this being used wrong.
Why is this confusing/so hard to get right?
Change-Id: Icb4629355c63508f5a044b46842524b3d203c2da
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-on: http://review.coreboot.org/11290
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
All compilers built, incl. x86_64-elf as multilib and riscv-elf.
Change-Id: Iafa61b1d2ffc9c737ab67a417c62417593b69372
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/10975
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
If I wanted to fill the whole memory address space with one byte, I
wouldn't try it that subtle.
With size_t beeing unsigned the loop condition >= 0 was always true.
Change-Id: Idee6a4901f6697093c88bda354b5e43066c0d948
Signed-off-by: Nico Huber <nico.huber@secunet.com>
Reviewed-on: http://review.coreboot.org/11286
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
Store the payload config and version files in CBFS if using a
SeaBIOS or filo payload if INCLUDE_CONFIG_FILE is enabled.
Change-Id: I0c1b4da8f6179b9cee06cecfa76bc631b43196e0
Signed-off-by: Martin Roth <gaumless@gmail.com>
Reviewed-on: http://review.coreboot.org/10607
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Tested-by: build bot (Jenkins)
cbmem_top was using CHIPSET_RESERVED_MEM_BYTES to w/a unknown memory
regions reserved by fsp for chipset use. With that being removed, the
function needs to properly walk though the memory map resulted from fsp
memory init to find out the usable address for cbmem root.
Refer the FSP 1.3.0 Integartion guide for more details on the Memory
Map.
systemagent should also use the same mechanism to create the reserved
RAM resource.
BRANCH=None
BUG=None
TEST=Build and Boot kunimitsu (FAB3)
CQ-DEPEND=CL:*226035,CL:*226045,CL:291573
Original-Change-Id: Id0954cf8e6388e549c7d4df67b468572b5bea539
Original-Signed-off-by: Rizwan Qureshi <rizwan.qureshi@intel.com>
Original-Reviewed-on: https://chromium-review.googlesource.com/291611
Original-Tested-by: Wenkai Du <wenkai.du@intel.com>
Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Original-Commit-Queue: Robbie Zhang <robbie.zhang@intel.com>
Change-Id: I4e716170f40936081ce9d4878bf74c75f469f78d
Signed-off-by: Rizwan Qureshi <rizwan.qureshi@intel.com>
Reviewed-on: http://review.coreboot.org/11239
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
(1) Wifi is connected on RP1 which is 1c.0 , so enabling
1c.0 and disabling 1d.0
(2) kepler is on RP5 which is 1c.4, so enabling it
(3) enabling ClkReqSupport for RP1 and RP5 so that L1 substates can
get enabled.
BRANCH=None
BUG=chrome-os-partner:43738
TEST=Built and boot for Kunimitsu. checked all PCIe powersaving
states (LTR, L1, L1S) are enabled
Original-Change-Id: I525661399d1a4d939b53d5ed5f7991598b84ddcd
Original-Signed-off-by: Pratik Prajapati <pratikkumar.v.prajapati@intel.com>
Original-Reviewed-on: https://chromium-review.googlesource.com/293482
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: Ib9a771a6ec137217668fb0385efc13b1824772b4
Signed-off-by: Pratik Prajapati <pratikkumar.v.prajapati@intel.com>
Reviewed-on: http://review.coreboot.org/11237
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Check device connect status while waiting for usb transfer complete
Avoid coreboot get stuck when usb device unplugged
BUG=chrome-os-partner:35525
TEST=None
BRANCH=None
Original-Change-Id: Id103501aa0d8b31b0b81bef773679c0fad79f689
Original-Signed-off-by: Yunzhi Li <lyz@rock-chips.com>
Original-Reviewed-on: https://chromium-review.googlesource.com/292630
Original-Reviewed-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Reviewed-by: Julius Werner <jwerner@chromium.org>
Original-Commit-Queue: Lin Huang <hl@rock-chips.com>
Original-Tested-by: Lin Huang <hl@rock-chips.com>
Original-Reviewed-on: https://chromium-review.googlesource.com/292966
Original-Reviewed-by: David Hendricks <dhendrix@chromium.org>
Original-Tested-by: David Hendricks <dhendrix@chromium.org>
Original-Commit-Queue: David Hendricks <dhendrix@chromium.org>
Change-Id: I49396b74131dbfda505d9d3de5adbdc87eb92ce1
Signed-off-by: Yunzhi Li <lyz@rock-chips.com>
Reviewed-on: http://review.coreboot.org/11236
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
The skylake IO-APIC supports up to 120 redirection entries.
In practice it seems FSP has already written to this write-once
register. However, it doesn't hurt to actually be correct within
the source.
BUG=chrome-os-partner:43522
BRANCH=None
TEST=Built and booted glados.
Original-Change-Id: I666b1b6034f0d37a37ea918f802317f9d5f15718
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/293251
Original-Reviewed-by: Robbie Zhang <robbie.zhang@intel.com>
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: I6ddbc89c98c262e2dd0f9f0b76adb092d3043602
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11235
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
The skylake SoC code now has macros for the previously
hard-code numbers for IRQs and GPEs. Switch over to using
those as they bring a little more clarity.
BUG=chrome-os-partner:43522
BRANCH=None
TEST=Built and booted glados.
Original-Change-Id: Ic8fcc59d680cdddec9dfbc3bf679731f6d786793
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/293411
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: I594907005372100a3c9d17dda9d17769844ad272
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11234
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
One thing that is brittle is lining up GPE0 bits in ASL
and with a board's design proper. This results in open
calculated magic numbers. To help alleviate this provide
just #defines that C preprocessor can use before handing
the source off to the ASL compiler.
BUG=chrome-os-partner:43522
BRANCH=None
TEST=Built and booted glados. Everything's intact.
Original-Change-Id: I359616ebe4bfc83c05bafe0ca36b766efd16dcca
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/293410
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: I32513c324b923fa0adbd6a0ee920c27e9b97dd1b
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11233
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
The get_region() function was using fixed masks for
the base and limit. However, newer descriptors (on
skylake, e.g.) use a 15-bit mask -- not a 12-bit one.
Choose the right mask based on ifd_version.
BUG=chrome-os-partner:43461
BRANCH=None
TEST=Built glados bootimage.
Original-Change-Id: Ibcbfd649a561d36b17ea2cc8fbeb30ffdbbb2c96
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/293250
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: I7f2ef9fb8e5b6c7114225fecc2798668d6507ac3
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11229
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
The descriptor format has changed with Skylake and some fields have
moved or been expanded.
This includes new SPI frequencies and chip densities, though unfortunately
30MHz in the new format conflicts with 50MHz in the old format...
There are also new regions with a few reserved regions inserted before
a new embedded controller region.
Unfortunately there does not seem to be a documented version field
so there does not seem to be an official way to determine if a
specific descriptor is new or old. To work around this ifdtool
checks the hardcoded "SPI Read Frequency" to see if it set for
20MHz (old descriptor) or 17MHz (new descriptor).
BUG=chrome-os-partner:40635
BUG=chrome-os-partner:43461
BRANCH=none
TEST=run ifdtool on skylake and broadwell images
Original-Change-Id: I0561b3c65fcb3e77c0a24be58b01db9b3a36e5a9
Original-Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/281001
Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Original-Commit-Queue: Aaron Durbin <adurbin@chromium.org>
Change-Id: I9a08c26432e13c4000afc50de9d8473e6f911805
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/293240
Reviewed-on: http://review.coreboot.org/11228
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
The location of the AMD ROMSIG binary was being checked and warnings
were being printed even when the ROMSIG file wasn't being used.
These false warnings are avoided by moving the warnings into the
block where the CBFS file for the ROMSIG is generated.
Change-Id: Ie44a2ad97ff3b15df6dc9b8166992de6ed837997
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: http://review.coreboot.org/11161
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Commit 27baa32 (cpu/amd/model_10xxx: Do not initialize SMM memory if
SMM is disabled) deactivated TSeg SMRAM, which had the side effect
of routing legacy VGA memory access to DRAM. Restore the correct
MMIO mapping via the MMIO configuration registers.
TEST: Booted KGPE-D16 with nVidia 7300LE card and verified proper VGA
functionality.
Change-Id: Ie4b7c0b2d6f9a02af9a022565fe514119513190a
Signed-off-by: Timothy Pearson <tpearson@raptorengineeringinc.com>
Reviewed-on: http://review.coreboot.org/11240
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Broadwell and Skylake chipsets, along with a few mainboards were
selecting ALWAYS_LOAD_OPROM without making sure that the dependency
for that symbol was met as well.
Looking at the dependencies for VGA_RUN_ROM, we see:
PCI && !PAYLOAD_SEABIOS && !MAINBOARD_DO_NATIVE_VGA_INIT
Since ARCH_X86 selects PCI, that's always met here.
Since Broadwell and Skylake don't have native VGA init yet, that's
not needed.
- Make sure that VGA_RUN_ROM is selected as well.
- Add dependency on !PAYLOAD_SEABIOS for both ALWAYS_LOAD_OPROM and
VGA_RUN_ROM symbols where they're selected.
Fixes Kconfig warning for these boards and chipsets:
warning: (BOARD_SPECIFIC_OPTIONS && BOARD_SPECIFIC_OPTIONS &&
BOARD_SPECIFIC_OPTIONS && CPU_SPECIFIC_OPTIONS && CPU_SPECIFIC_OPTIONS)
selects ALWAYS_LOAD_OPROM which has unmet direct dependencies
(VGA_ROM_RUN)
Change-Id: I787a87e9467e1fc7afe8b04864b2a89b54824b9f
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: http://review.coreboot.org/11246
Tested-by: build bot (Jenkins)
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Change the dependency on CONSOLE_SERIAL to select CONSOLE_SERIAL based
on this question.
The dependency was causing multiple warnings on every platform tested.
src/console/Kconfig:21:error: recursive dependency detected!
src/console/Kconfig:21: symbol CONSOLE_SERIAL depends on
DRIVERS_UART_8250MEM
src/drivers/uart/Kconfig:16: symbol DRIVERS_UART_8250MEM is selected by
UART_DEBUG
src/soc/intel/skylake/Kconfig:198: symbol UART_DEBUG depends on
CONSOLE_SERIAL
Change-Id: Ia0426cd150561694081b5ea7c6797d36022c1f57
Signed-off-by: Martin Roth <martinroth@google.com>
Reviewed-on: http://review.coreboot.org/11243
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins)
When the user's primary group contains a space ls -l and awk get the
wrong value for the file size. This results in padding the
coreboot_psp_directory_combine_pubkey.bin file too much which ultimately
means RtmPubSigned.key can not be placed at the necessary offset.
Changing from ls -l to ls -ln seemed like the most minimal,
POSIX-friendly way to effect this change.
Change-Id: Icbeaad476753924626adb6de53dc9a30052d91a6
Signed-off-by: Dan Christensen <opello@opello.org>
Reviewed-on: http://review.coreboot.org/11242
Tested-by: build bot (Jenkins)
Reviewed-by: Zheng Bao <zheng.bao@amd.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
seabios integration interprets the CC variable with a special case when
ccache is prepended to the compiler.
Since the integration also tries to extract compiler flags (which I'm
not sure we still add to CC _ever_), that also needs to look at only
the part of the string that contains compiler and (maybe) flags, so
skip the first word if it was determined to be the path to the ccache
binary.
Change-Id: I717863f456bf4fd6f08427d86633079ecda039df
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Reviewed-on: http://review.coreboot.org/11227
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
In order for the EC_SCI_L to work the GPE0 route needs
to be set along w/ the GPE event for the EC. As the GPE0
route is dynamic the EC_SCI_GPI needs to be set along
with the route so everything lines up. In this case, the
GPE0 route is set to the defaults such that GPP_C, GPP_D,
and GPP_E are routed to GPE0 block 0, 1, and 2, respectively.
This works out for glados because the EC_SCI_L is connected
to GPP_E16.
BUG=chrome-os-partner:43778
BRANCH=None
TEST=Built and booted glados. The 'acpi' interrupt in /proc/interrupts
is incrementing as well as /sys/firmware/acpi/interrupts/gpe50.
Original-Change-Id: I71fc4bec124f3ac87453a099412154e67aba6280
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/292011
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: Idbb6d29364655537abc9ae6f012b3abb38edf138
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11210
Tested-by: build bot (Jenkins)
Set the EC_SMI_GPI define to be GPP_E15 and route that
GPIO for SMI generation. Also, the mainboard_smi_gpi_handler()
was introduced on skylake in order to process any GPI that could
generate an SMI. Switch to this handler so one can process the
appropriate events.
BUG=chrome-os-partner:43778
BRANCH=None
TEST=Used 'lidclose' on EC command line during depthcharge
to confirm EC_SMI_L generates SMI and shutdown happens.
Original-Change-Id: Ia365b86161670a809e3fa99dde38fccc612d5e77
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/291934
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: Ic16ea8e8d6ff564977ed2081d2353c82af71adea
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11209
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
The current construction for processing SMI GPI events
didn't allow for the mainboard to query the state of a
particular GPI for the snapshotted SMI event. The
skylake part can route GPIs from any (there are design
limitations) GPIO group. Those status and enable registers
are within the GPIO community so one needs to gather
all the possibilities in order to query the state.
The call chain did this:
southbridge_smi_gpi(
clear_alt_smi_status() -> reset_alt_smi_status() ->
print_all_smi_status() -> return 0)
As a replacement the following functions and types are
introduced:
struct gpi_status - represent gpi status.
gpi_status_get() - per gpi query on struct gpi_status
gpi_clear_get_smi_status() - clear and retrieve SMI GPI status
mainboard_smi_gpi_handler() - mainboard handler using gpi_status
Also remove gpio_enable_all_smi() as that construct was never
used, but it also is quite heavy handed in that it would
enable SMI generation for all GPIs.
BUG=chrome-os-partner:43778
BRANCH=None
TEST=Built.
Original-Change-Id: Ief977e60de65d9964b8ee58f2433cae5c93872ca
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/291933
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: Ida009393c6af88ffe910195dc79a4c0d2a4c029e
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11208
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
The first pass of the GPIO configuration patch didn't
enable the SMI# generation for GPIs marked as SMI
routed. Now when a pad is configured as SMI routed
the bit for the SMI enablement is set accordingly.
BUG=chrome-os-partner:43778
BRANCH=None
TEST=Built and booted glados. Confirmed SMI_EN being set
for SMI routed GPIOs.
Original-Change-Id: I796b68accb7a49b03ef18539861e72fa9d169c26
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/292010
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: I3be770234d3f605ae630ecd5cd4cfe4867243999
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11207
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
The gpio pad configuration currently defaults to ACPI
owned GPIs. A '0' was used which wasn't so clear. Add
a comment and explicitly set it to ACPI. Also,
PAD_CFG_GPI_ACPI_SMI wasn't using the _PAD_CFG_ATTRS
macro which causes compliation errors if attempted
to be instantiated. No piece of code tried to use
it so the error was overlooked.
Lastly, allow for soc/gpio.h to be included during
ASL compilation. That allows for gpio_defs.h to be
included and those macros utilized without needing
to know the file name and where it lives; just use
the generic gpio.h.
BUG=chrome-os-partner:43778
BRANCH=None
TEST=Built and booted glados.
Original-Change-Id: I9dbadb0b494683ab38babfc1ac5e13093ee37730
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/291935
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: Id4fa8b65ec1e1537dbf09824c2155119a768807e
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11206
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Instead of using a hard-coded value leverage the existing
definitions to perform GPE0 block length calculations. There
are 4 pairs of 32-bit status/enable registers.
BUG=chrome-os-partner:43522
BRANCH=None
TEST=Built and booted glados.
Original-Change-Id: I14d08298b5750c91ce0ac3fa33569813396f7089
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/291932
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: I127f026f15180fa79625d4cad96d5e35f85e5090
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11205
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
The ec_smi_gpio and alt_gp_smi_en devicetree options are
goign to be removed. The plan for skylake is to set the
settings by the mainboard through either gpio pad
configuration or through helper functions.
Moreover, these values only allow *1* SMI GPIO configuration
in that the following has to be true:
alt_gp_smi_en = 1 << (ec_smi_gpio % 24)
If not, then another gpio(s) from the same group has the
SMI_EN bit set for it.
Lastly, remove all the subsequent dependencies as they are
no longer used: enable_alt_smi() and gpio_enable_group().
BUG=chrome-os-partner:43778
BRANCH=None
TEST=None
Original-Change-Id: I749a499c810d83de522a2ccce1dd9efb0ad2e20a
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/291931
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: I2e1cd6879b76923157268a1449c617ef2aada9c4
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11204
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
On skylake the GPE0 routing can be dynamically changed to
a particular GPIO group. Provide the ability for the mainboard
to set the route accordingly. If any of the values in the
devicetree are the same the current setting in the PMC register
is used. The GPIO communities need to have matching configuration
for the plumbing to work properly.
BUG=chrome-os-partner:43778
BRANCH=None
TEST=Built and booted glados w/ and w/o devicetree changes. Fields
are set accordingly.
Original-Change-Id: I263d648c8ea8a70b21570f01b333d05a5fa2a4e3
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/291930
Original-Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Change-Id: I966d38bc197dbb52a2ba50927c06e243e169afbe
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11203
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Patrick Georgi