The memcpy(), memmove() and memcmp() functions use word by word
operations regardless of the pointer alignment. Depending on the
platform, this could lead to a crash.
This patch makes the memcpy(), memmove() or memcmp() operate byte per
byte if they are supplied with unaligned pointers.
Change-Id: I0b668739b7b58d47266f10f2dff2dc9cbf38577e
Signed-off-by: Jeremy Compostella <jeremy.compostella@intel.com>
Reviewed-on: https://review.coreboot.org/20535
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Reviewed-by: Julius Werner <jwerner@chromium.org>
The FSP 2.0 path uses postcar to decompress ramstage. Since postcar
is entirely RAM based there's no need to have an excessively large
stack for the lzma decompression buffer. Therefore, reduce the stack
required to 1 KiB like apollolake.
Change-Id: I45e5c283f8ae87e701c94d6a123463dddde3f221
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: https://review.coreboot.org/20536
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-by: Lijian Zhao <lijian.zhao@intel.com>
Reviewed-by: Furquan Shaikh <furquan@google.com>
At process _start, the stack is expected to be aligned to a
16-byte boundary. Upon entry to any function the stack frame
must have the end of any arguments also aligned. In other words
the value of %esp+4 or %rsp+8 is always a multiple of 16 (1).
Align the stack down and change the method for executing
car_stage_entry from jmp to call which should preserve proper
alignment regardless of a 32- or 64-bit build.
Although 4-byte alignment is the minimum requirement for i386,
some AMD platforms use SSE instructions which expect 16-byte.
1) http://wiki.osdev.org/System_V_ABI
See "Initial Stack and Register State" and "The Stack Frame"
in the supplements.
BUG=chrome-os-partner:62841664
Change-Id: I8a15514f551a8e17e9fe77b8402fe0d2b106972e
Signed-off-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-on: https://review.coreboot.org/20528
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Update my email address for Apollolake and FSP 2.0 driver. Also
downgrade support from "supported" to "maintained" as currently
no other Intel persons are assigned for the role.
Change-Id: I3033fc5ec8b0882ce79eeb15ee3eb13a228611a4
Signed-off-by: Andrey Petrov <andrey.petrov@intel.com>
Reviewed-on: https://review.coreboot.org/20550
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Add capability and location data for USB ports/devices via
_PLD and _UPC ACPI methods, which is utilized by Windows and
required by macOS.
Move inclusion of mainboard.asl after southbridge asl files
so scopes referenced in usb.asl are valid.
Change-Id: Id98aa5f6dbdcbb8da4616d4fce6e7388f3ba4656
Signed-off-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-on: https://review.coreboot.org/20515
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Needs the ramstage configuration enabling of SuperIO GPIO pnp devices
for BSEL straps.
Also needs VSBGATE# to be on for ram to be powered during S3.
TESTED with 800MHz and 1067MHz FSB CPUs at the correct straps when
resuming from S3.
Change-Id: I6ac927ee9dcce15fc7621aad57969fae8f5805ca
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/19602
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Enable Azalia controller, HD Audio DSP and select
the HDAudio IoBuffer Ownership for rvp3.
Check if device is enabled in HDA codec init function
to avoid failure when Azalia controller is disabled in
the devicetree.cb.
BUG=None
TEST=Build for kblrvp3, Make sure booting is fine irrespective of HDA
enable/disable.
Change-Id: I87212fe16ecc6053d6d00372904a5fd5d6f6b209
Signed-off-by: V Sowmya <v.sowmya@intel.com>
Reviewed-on: https://review.coreboot.org/20530
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Martin Roth <martinroth@google.com>
Add and use new interface to set and get GNVS' ASLB register.
To be used by Intel's gma driver to set ASLB at ACPI table
creation and to get ASLB on S3 resume.
Change-Id: If30c6b2270069783b0892774802f47406404da5f
Signed-off-by: Patrick Rudolph <siro@das-labor.org>
Reviewed-on: https://review.coreboot.org/20435
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
The optimization of the memset() function introduced by commit
dbadb1dd63 (libpayload: Reorder default
memcpy, speed up memset and memcmp) is provoking an issue on x86
platform when compiling without the CONFIG_GPL option.
GCC is making use of the movdqa instruction to copy words. This
instruction can raise a "General Protection Fault Exception" when it
is called on a non-aligned address argument.
Change-Id: I73382a76a4399d8e78244867f2ebb1dca176a6bf
Signed-off-by: Jeremy Compostella <jeremy.compostella@intel.com>
Reviewed-on: https://review.coreboot.org/20524
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
1. Due to reset signal, PMIC loses its internal register state. This
causes PMIC to be in improper state after sleep.
2. The intent of reset signal is to reset internal state of PMIC (which
happens once during power on), hence avoid asserting reset signal
when not needed.
3. As per PMIC (TPS68470) datasheet, device can be kept in SLEEP mode
when not in use to save max possible power.
To fix the same, do not reset PMIC while entering sleep.
By keeping PMIC in SLEEP mode, Power consumption is < 1uW (Typ) upto
3.63uW (Max). Refs: TPS68470 datasheet.
Measured value: 0.66uW
TEST= Build the firmware for Soraka & boot to OS. Do S3 resume & check
whether PMIC internal registers state are preserved.
Change-Id: I93ce4d76b0376b64ae6d1067aca0fd7467af3582
Signed-off-by: Naresh G Solanki <naresh.solanki@intel.com>
Reviewed-on: https://review.coreboot.org/20264
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Furquan Shaikh <furquan@google.com>
This reverts commit 5535cead (intel/skylake: Disable SaGv in
recovery mode).
Commit 5535cead disables SaGv in recovery mode to save few seconds
booting time as we were doing memory training on every recovery flow.
Now we don't need to perform MRC training on every recovery boot
due to RECOVERY_MRC_CACHE implementation in place. Hence we don't
need to define different SaGv policy between Normal (developer) mode
and recovery mode to save few seconds.
Using different SaGv parameters between recovery and all other mode
has some significent drawbacks over warm reboot cycle. We are seeing
a MRC traning hang in eve/soraka/poppy devices with below use case.
Step 1: Boot system in developer mode (first time RW_MRC training)
Step 2: Set recovery_request=1 (using crossystem) and issue “reboot”
from OS
Step 3: System will perform recovery mode MRC training and boot to
OS (first time RECOVERY_MRC training)
Step 4: Issue “reboot” from OS console.
Step 5: System wil boot in developer mode (using RW_MRC cache)
Step 6: Set recovery_request=1 (using crossystem) and issue “reboot”
from OS
Step 7: System will pick RECOVERY_MRC_CACHE and will hang during
MRC training.
This patch fixes issue mentioned above and ensures system boot to
OS without any hang if we change mode (dev<->recovery) over warm
reset.
BUG=b:63515071
BRANCH=none
TEST=manual stress testing of dev<->recovery mode over warm boot.
No MRC hang with this fix on eve/soraka/poppy devices.
Change-Id: I8d094a8b6d78ea3bf8f929870a4a179495c29c78
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/20516
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
The macro FPGA_SET_PARAM was introduced to make the setting of different
FPGA registers with the appropriate values from hwinfo more
transparent. The hwilib takes care about the size of the provided buffer
where the requested value should be stored in. The fields in hwinfo have
not always the same size as the matching registers in the FPGA. So to
avoid errors resulting in a too small buffer when calling hwilib_get_field()
the buffer is now fixed to 32 bit and will be casted to the destination
type when the value is written into the FPGA register.
Changing the field size in hwilib would be the wrong way as the defined
lengths are specified this way to be expandable in the future.
In addition the number of maximum supported temperature sensors is
increased to 8 as the FPGA now supports more.
Change-Id: I0c697106783158420708a973c3cff2be90fa4fce
Signed-off-by: Werner Zeh <werner.zeh@siemens.com>
Reviewed-on: https://review.coreboot.org/20471
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Mario Scheithauer <mario.scheithauer@siemens.com>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
For distros that package and version gnat independently from gcc (such
as Ubuntu), try to build with gnatgcc first.
This fixes the issue of gcc -print-prog-name=gnat1 failing because gcc
is of a different version.
Change-Id: Icec6d1fba8855e88ac91d47842dcb7f6b9d35461
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://review.coreboot.org/20517
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Fix undefined behavior found by clang's -Wshift-sign-overflow, find,
and source inspection. Left shifting an int where the right operand is
>= the width of the type is undefined. Add UL suffix since it's safe
for unsigned types.
Change-Id: I5240a19647c8ad59f64925f3e1c199446a886d2d
Signed-off-by: Ryan Salsamendi <rsalsamendi@hotmail.com>
Reviewed-on: https://review.coreboot.org/20466
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Fix undefined behavior found by clang's -Wshift-sign-overflow, grep,
and source inspection. Left shifting an int where the right operand is
>= the width of the type is undefined. Add UL suffix since it's safe
for unsigned types.
Change-Id: I10db2566199200ceb3068721cfb35eadb2be1f68
Signed-off-by: Ryan Salsamendi <rsalsamendi@hotmail.com>
Reviewed-on: https://review.coreboot.org/20464
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
