We generally do not guard source in attempts to reduce
the final object sizes, but rely on garbage collection.
Most of the __unused attributes inserted here will be
removed when remaining __SIMPLE_DEVICE__ guards can
be removed.
Change-Id: I2440931fab4f41d7e8249c082e6c9b5a9cd0ef13
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/36641
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The x86 timers are a bit of a mess. Cases where different stages use
different counters and timestamps use different counters from udelays.
The original intention was to only flip TSC_CONSTANT_RATE Kconfig
to NOT_CONSTANT_TSC_RATE. The name would be incorrect though, those
counters do run with a constant rate but we just lack tsc_freq_mhz()
implementation for three platforms.
Note that for boards with UNKNOWN_TSC_RATE=y, each stage will have a
slow run of calibrate_tsc_with_pit(). This is easy enough to fix with
followup implementation of tsc_freq_mhz() for the platforms.
Implementations with LAPIC_MONOTONIC_TIMER typically will not have
tsc_freq_mhz() implemented and default to UNKNOWN_TSC_RATE. However,
as they don't use TSC for udelay() the slow calibrate_tsc_with_pit()
is avoided.
Because x86/tsc_delay.tsc was using two different guards and nb/via/vx900
claimed UDELAY_TSC, but pulled UDELAY_IO implementation, we also switch
that romstage to use UDELAY_TSC.
Change-Id: I1690cb80295d6b006b75ed69edea28899b674b68
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/33928
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This patch creates a common instance of sleepstates.asl inside intel common
code (southbridge/intel/common/acpi) and asks all IA CPU/SOC code to
refer sleepstates.asl from common code block.
TEST=Able to build and boot Hatch and ICL DE system. Dump DSDT.asl to verify
S0/S3/S4/S5 entries after booting to OS.
Change-Id: Ie2132189f91211df74f8b5546da63ded4fdf687a
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/36463
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This adds a common cbmem_top implementation to all coreboot target.
In romstage a static variable will be used to cache the result of
cbmem_top_romstage.
In ramstage if CONFIG_RAMSTAGE_CBMEM_TOP_ARG is set a global variable
needs to be populated by the stage entry with the value passed via the
calling arguments. if CONFIG_RAMSTAGE_CBMEM_TOP_ARG is not set the
same implementation as will be used as in romstage.
Change-Id: Ie767542ee25483acc9a56785ce20a885e9a63098
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/36273
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
This provides stack guards with checking and common
entry into postcar.
The code in cpu/intel/car/romstage.c is candidate
for becoming architectural so function prototype
is moved to <arch/romstage.h>.
Change-Id: I4c5a9789e7cf3f7f49a4a33e21dac894320a9639
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/34893
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Variable length arrays were a feature added in C99 that allows the
length of an array to be determined at runtime. Eg.
int sum(size_t n) {
int arr[n];
...
}
This adds a small amount of runtime overhead, but is also very
dangerous, since it allows use of an unlimited amount of stack memory,
potentially leading to stack overflow. This is only worsened in
coreboot, which often has very little stack space to begin with. Citing
concerns like this, all instances of VLA's were recently removed from the
Linux kernel. In the immortal words of Linus Torvalds [0],
AND USING VLA'S IS ACTIVELY STUPID! It generates much more code, and
much _slower_ code (and more fragile code), than just using a fixed
key size would have done. [...] Anyway, some of these are definitely
easy to just fix, and using VLA's is actively bad not just for
security worries, but simply because VLA's are a really horribly bad
idea in general in the kernel.
This patch follows suit and zaps all VLA's in coreboot. Some of the
existing VLA's are accidental ones, and all but one can be replaced with
small fixed-size buffers. The single tricky exception is in the SPI
controller interface, which will require a rewrite of old drivers
to remove [1].
[0] https://lkml.org/lkml/2018/3/7/621
[1] https://ticket.coreboot.org/issues/217
Change-Id: I7d9d1ddadbf1cee5f695165bbe3f0effb7bd32b9
Signed-off-by: Jacob Garber <jgarber1@ualberta.ca>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/33821
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
As most platforms will share the subset of enabling
both low RAM WB and high ROM WP MTRRs, provide them
with a single function.
Add possibility for the platform to skip these if
required.
Change-Id: Id1f8b7682035e654231f6133a42909a36e3e15a1
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/34809
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
It is easier to track CAR_GLOBAL_MIGRATION which is
the approach to be deprecated with the next release.
This change enforces new policy; POSTCAR_STAGE=y is
not allowed together with CAR_GLOBAL_MIGRATION=y.
Change-Id: I0dbad6a14e68bf566ac0f151dc8ea259e5ae2250
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/34804
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Let garbage-collection take care of stage_cache_external_region()
when it is not needed and move implementation to a suitable file
already building for needed stages.
Change-Id: Ia6adcc0c8bf6d4abc095ac669aaae876b33ed0f3
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/34669
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Consider the following assignment:
u64 = s32
For positive values this is fine, but if the s32 is negative, it will be
sign-extended in the conversion to a very large unsigned integer. This
manifests itself in two ways in the following code:
First, gpu_pipe{a,b}_port_select are defined as int, and can have the
values 1 or 2. In the case when they have the value 2, the shift 2 << 30
will be a negative number, making it susceptible to the sign-extension
problem above. Change these variables to something more reasonable like
a uint8_t, which is unsigned.
Second, in any bit shift, any variable with width less than an int will
be implicitly promoted to an int before performing the bit shift.
For example, the variable gpu_pipea_power_on_delay is a uint16_t, and if its
highest bit is set, the shift gpu_pipea_power_on_delay << 16 will become
negative, again introducing the above problem. To prevent this, cast all
smaller variables to a u32 before the shift, which will prevent the
implicit promotions and sign extensions.
Change-Id: Ic5db6001504cefb501dee199590a0e961a15771b
Signed-off-by: Jacob Garber <jgarber1@ualberta.ca>
Found-by: Coverity CID 1229699, 1229700, 1229701, 1229702
Reviewed-on: https://review.coreboot.org/c/coreboot/+/34487
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Alexander Couzens <lynxis@fe80.eu>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
With VBOOT=y && VBOOT_MEASURED_BOOT=y message
digest will be allocated from the stack and
1 KiB reserve used with the recent platforms
was no longer sufficient.
The comment of LZMA scratchpad consuming stack
was obsolete for postcar, so these can be reduced
to same 4 KiB.
Change-Id: Iba1fb5bfad6946f316feac2d8c998a782142a56a
Signed-off-by: Mario Scheithauer <mario.scheithauer@siemens.com>
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/33775
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Werner Zeh <werner.zeh@siemens.com>
All Baytrail boards have the same GPU PCI ID, so set it
here to avoid having to set it in each board's config.
Move the VGA_BIOS_FILE config from google/rambi into soc/baytrail
since it likewise applies to all Baytrail boards.
Change-Id: Id1e0580b55e3590d868cb839987f06c49bb07cf5
Signed-off-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/33026
Reviewed-by: HAOUAS Elyes <ehaouas@noos.fr>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
