The code in cpu/x86/lapic/apic_timer.c for timer_monotonic_get()
is not SMP safe as LAPIC timers do not run as synchronised as TSCs.
The times reported for console for boot_states does not accumulate
from APs now. Also remove console time tracking from ENV_SMM.
Change-Id: I1ea2c1e7172f8ab3692b42dee3f669c5942d864a
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37398
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
TEST=Set MAX_CPUS=2 and run qemu with -smp 2
Signed-off-by: Philipp Hug <philipp@hug.cx>
Change-Id: I94fb25fad103e3cb5db676eb4caead11d54ae0ae
Reviewed-on: https://review.coreboot.org/c/coreboot/+/35246
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Xiang Wang <merle@hardenedlinux.org>
gcc seems to have some stupid problem with deciding when to inline byte
swapping functions (https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92716).
Using the compiler builtin instead seems to solve the problem.
(This doesn't yet solve the issue for the read_be32()-family of
functions, which we should maybe just get rid of at some point?)
Change-Id: Ia2a6d8ea98987266ccc32ffaa0a7f78965fca1cd
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37343
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
factory_initialize_tpm() calls secdata_xxx_create() (for both firmware
and kernel space) and then immediately writes those spaces out to the
TPM. The create() functions make vboot think it just changed the secdata
(because it reinitialized the byte arrays in the context), so we also
need to clear the VB2_CONTEXT_SECDATA_xxx_CHANGED flags again, otherwise
vboot thinks it still needs to flush the spaces out to the TPM even
though we already did that.
Also clean up some minor related stuff (VB2_CONTEXT_SECDATA_CHANGED
notation is deprecated, and secdata space intialization should use the
same write-and-readback function we use for updates).
Change-Id: I231fadcf7b35a1aec3b39254e7e41c3d456d4911
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37471
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Going forwards, vb2ex_commit_data will be used to flush both
nvdata and secdata.
The patch that is circularly dependent on this lies between a patch that
makes vboot no longer build and the patch that fixes that, so we have to
pull the whole thing in at once to sort out the mess.
Updating from commit id 1c4dbaa0:
2019-11-18 Julius Werner Makefile: Fix typo for MOCK_TPM
to commit id 695c56dc:
2019-12-04 Julius Werner Makefile: Make loop unrolling fully
controllable by the caller
BUG=b:124141368, chromium:1006689
TEST=make clean && make test-abuild
BRANCH=none
Change-Id: Ia2612da0df101cd3c46151dbce728633a39fada1
Signed-off-by: Joel Kitching <kitching@google.com>
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37315
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
1. No gpio control in bootblock
2. Disable power and assert reset in ramstage gpio
3. Power on and then deassert reset at the end of ramstage gpio
4. Disable power and assert reset when entering S5
On "reboot", the amount of time the power is disabled for is
equivalent to the amount of time between triggering #4 and wrapping
around to #3, which is about 400ms on Kohaku.
Since #2 forces power off for FPMCU, S3 resume will still
not work properly.
Additionally, we must ensure that GPP_A12 is reconfigured as an output
before going to any sleep state, since user space could have configured
it to use its native3 function.
See https://review.coreboot.org/c/coreboot/+/32111 for more detail.
The control signals have been validated on a Kohaku in
the following scenarios:
1. Cold startup
2. Issuing a "reboot" command
3. Issuing a "halt -p" and powering back on within 10 seconds
4. Issuing a "halt -p" and powering back on after 10 seconds
5. Entering and leaving S3 (does not work properly)
6. Entering and leaving S0iX
BRANCH=hatch
BUG=b/142751685
TEST=Verify all signals as mentioned above
TEST=reboot
flash_fp_mcu /opt/google/biod/fw/dartmonkey_v2.0.2417-af88cc91a.bin
TEST=halt -p
# power back on within 10 seconds
flash_fp_mcu /opt/google/biod/fw/dartmonkey_v2.0.2417-af88cc91a.bin
TEST=halt -p
# power back on after 10 seconds
flash_fp_mcu /opt/google/biod/fw/dartmonkey_v2.0.2417-af88cc91a.bin
Change-Id: I2e3ff42715611d519677a4256bdd172ec98687f9
Signed-off-by: Craig Hesling <hesling@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37459
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
This patch moves the common devicetree settings into baseboard and
creates overridetree.cb for each variant. For PCIe root port settings,
SATA, eMMC, I2Cs and GBe, they are in overridetree.
TEST=build an image for each variant
Change-Id: I067bdb3fcf1218b93e52801f6db093e24d7d2b62
Signed-off-by: Gaggery Tsai <gaggery.tsai@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/36794
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Definitions were moved so that now device/mmio.h needs to be included
instead of arch/mmio.h. Also, don't use le32 conversion.
This follows the activities of commit 55009af42 (Change all
clrsetbits_leXX() to clrsetbitsXX()) and commit 1c37157218 (mmio: Add
clrsetbitsXX() API in place of updateX()).
Change-Id: Ie3af0d4f0b3331fe5572fc56915952547b512db7
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37534
Reviewed-by: Julius Werner <jwerner@chromium.org>
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Over time our printk() seems to acquire more and more features... which
is nice, but it also makes it a little less robust when something goes
wrong. If the wrong global is trampled by some buffer overflow, it
suddenly doesn't print anymore. It would be nice to have at least some
way to tell that we triggered a real exception in that case.
With this patch, arm64 exceptions will print a '!' straight to the UART
before trying any of the more fancy printk() stuff. It's not much but it
should tell the difference between an exception and a hang and hopefully
help someone dig in the right direction sooner. This violates loglevels
(which is part of the point), but presumably when you have a fatal
exception you shouldn't care about that anymore.
Change-Id: I3b08ab86beaee55263786011caa5588d93bbc720
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37465
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
To avoid trampling over interesting exception artifacts on the real
stack, our arm64 systems switch to a separate exception stack when
entering an exception handler. We don't want that to use up too much
SRAM so we just set it to 512 bytes. I mean it just prints a bunch of
registers, how much stack could it need, right?
Quite a bit it turns out. The whole vtxprintf() call stack goes pretty
deep, and aarch64 generally seems to be very generous with stack space.
Just the varargs handling seems to require 128 bytes for some reason,
and the other stuff adds up too. In the end the current implementation
takes 1008 bytes, so bump the exception stack size to 2K to make sure it
fits.
Change-Id: I910be4c5f6b29fae35eb53929c733a1bd4585377
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37464
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
Arm CPUs have always had an odd feature that allows you to mask not only
true interrupts, but also "external aborts" (memory bus errors from
outside the CPU). CPUs usually have all of these masked after reset,
which we quickly learned was a bad idea back when bringing up the first
arm32 systems in coreboot. Masking external aborts means that if any of
your firmware code does an illegal memory access, you will only see it
once the kernel comes up and unmasks the abort (not when it happens).
Therefore, we always unmask everything in early bootblock assembly code.
When arm64 came around, it had very similar masking bits and we did the
same there, thinking the issue resolved. Unfortunately Arm, in their
ceaseless struggle for more complexity, decided that having a single bit
to control this masking behavior is no longer enough: on AArch64, in
addition to the PSTATE.DAIF bits that are analogous to arm32's CPSR,
there are additional bits in SCR_EL3 that can override the PSTATE
setting for some but not all cases (makes perfect sense, I know...).
When aborts are unmasked in PSTATE, but SCR.EA is not set, then
synchronous external aborts will cause an exception while asynchronous
external aborts will not. It turns out we never intialize SCR in
coreboot and on RK3399 it comes up with all zeroes (even the reserved-1
bits, which is super weird). If you get an asynchronous external abort
in coreboot it will silently hide in the CPU until BL31 enables SCR.EA
before it has its own console handlers registered and silently hangs.
This patch resolves the issue by also initializing SCR to a known good
state early in the bootblock. It also cleans up some bit defintions and
slightly reworks the DAIF unmasking... it doesn't actually make that
much sense to unmask anything before our console and exception handlers
are up. The new code will mask everything until the exception handler is
installed and then unmask it, so that if there was a super early
external abort we could still see it. (Of course there are still dozens
of other processor exceptions that could happen which we have no way to
mask.)
Change-Id: I5266481a7aaf0b72aca8988accb671d92739af6f
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37463
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
This includes USB QUSB2,QMP Phy and Controller support
And libpayload support for USB
Change-Id: I0651fc28dc227efbeb23eeefe9b96a3b940ae995
Signed-off-by: Sandeep Maheswaram <sanm@codeaurora.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/35503
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
This implements the SPI-NOR driver for the Qualcomm QSPI core.
Developer/Reviewer, be aware of this patch from Napali:
https://review.coreboot.org/c/coreboot/+/27483/58
Change-Id: I2eb8cf90aa4559541ba293b3fd2870896bed20b7
Signed-off-by: Akash Asthana <akashast@codeaurora.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/35501
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-by: Julius Werner <jwerner@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
The PCH uses the SRCCLKREQ# pin to detect PCIe device in the slot in
order to send clock signal to it. However, this logic is not required
for the Realtek LAN device, since this chip is soldered to the board
and always uses clocking. The chipset can't receive the clock request
signal (most likely this pin isn't connected) and doesn't enable the
CLK. For this reason, the device is broken during the initialization
phase. The patch disables clock request logic for the PCH PCIe port 6
to initialize the onboard LAN device correctly.
Change-Id: I5cbce6177c89052eb50959f43903b6f8a607e77f
Signed-off-by: Maxim Polyakov <max.senia.poliak@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/36377
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Add support for clock driver for SC7180
Developer/Reviewer, be aware of this patch from Napali:
https://review.coreboot.org/c/coreboot/+/31083/6
Change-Id: I3f39252c887c36e8af43bc49289795000e4638d8
Signed-off-by: Taniya Das <tdas@codeaurora.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/35496
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
TEST=boot PC Engines apu1 with C bootblock patch and launch
Debian with Linux kernel 4.14.50
Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Change-Id: Ie81198f5034a84d319ee7143aa032433f82be254
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37329
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This patch changes all existing instances of clrsetbits_leXX() to the
new endian-independent clrsetbitsXX(), after double-checking that
they're all in SoC-specific code operating on CPU registers and not
actually trying to make an endian conversion.
This patch was created by running
sed -i -e 's/\([cs][le][rt]bits\)_le\([136][624]\)/\1\2/g'
across the codebase and cleaning up formatting a bit.
Change-Id: I7fc3e736e5fe927da8960fdcd2aae607b62b5ff4
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37433
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
This patch removes the recently added update8/16/32/64() API and
replaces it with clrsetbits8/16/32/64(). This is more in line with the
existing endian-specific clrsetbits_le16/32/64() functions that have
been used for this task on some platforms already. Rename clrsetbits_8()
to clrsetbits8() to be in line with the new naming.
Keep this stuff in <device/mmio.h> and get rid of <mmio.h> again because
having both is confusing and we seem to have been standardizing on
<device/mmio.h> as the standard arch-independent header that all
platforms should include already.
Also sync libpayload back up with what we have in coreboot. (I'm the
original author of the clrsetbits_le32-definitions so I'm relicensing
them to BSD here.)
Change-Id: Ie4f7b9fdbdf9e8c0174427b4288f79006d56978b
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37432
Reviewed-by: Nico Huber <nico.h@gmx.de>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Create new variant for Lick that is copied from phaser variant.
Remove unnecessary code, due to not support touchscreen and stylus.
Set to default_override_table.
Remove variant.c.
BUG=b:145181137
BRANCH=octopus
TEST=./util/abuild/abuild -p none -t google/octopus -x -a
Change-Id: If732d94194defb9f5ee9c847ee93dd58aef01174
Signed-off-by: Hash.Hung <hash1.hung@lcfc.corp-partner.google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37247
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-by: Henry Sun <henrysun@google.com>
Reviewed-by: Marco Chen <marcochen@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
All serial I2C bus frequencies should not be over 400KHz in kohaku,
but the measurement showed frequencies of I2C1 and I2C4 were over
400KHz. (b:144885961)
This change adjusts I2C speed settings to limit that frequencies to
400KHz.
The new setting values have been from other projects using same I2C
components, and verified I2C1 and I2C4 frequencies < 400MHz internally.
BUG=b:144885961
BRANCH=firmware-hatch-12672.B
TEST=Verified I2C1 and I2C4 frequency not over 400KHz
Change-Id: I9614fb39b6e55cb2ce1b0879a9f5204e55002f8d
Signed-off-by: Seunghwan Kim <sh_.kim@samsung.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37313
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Make some room for C environment bootblock. The S3 resume
feature needs less than 2 KiB.
Change-Id: Ic49c313d492f1d18f59d61e84f81f106e3b41fb1
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37439
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Move the stoneyridge implementation of get/set AP entry to the common
block.
Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Change-Id: I9c73940ffe5f735dcd844911361355c384f617b1
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37416
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Richard Spiegel <richard.spiegel@silverbackltd.com>
All AMD CPU families supported in coreboot have BIOSRAM space. Looking at
the source code, every family could have the same API to save and restore
cbmem top or UMA base and size.
Unify BIOSRAM layout and add implementation for cbmem top and UMA storing.
Also replace the existing implementation of cbmem top and UMA with the
BIOSRAM access.
TEST=boot PC Engines apu1 and apu2
Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Change-Id: I69a03e4f01d7fb2ffc9f8b5af73d7e4e7ec027da
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37402
Reviewed-by: Richard Spiegel <richard.spiegel@silverbackltd.com>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
According to BKDGs for families 15h 60-6fh or newer and families 16h the
ACPI MMIO decode enable bit is the second LSB, not the first LSB.
Additionally create another enable function for older families where
the register and bit is different.
It does not seem to impact any current board, but may be crucial for
incoming C bootblock implementations when this bit will need to be set
very early. Most likely this bit is set by AGESA right now.
Signed-off-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Change-Id: Iaa31abc3dbdf77d8513fa83c7415b9a1b7fd266f
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37178
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Fix regression with commit 5639736
binaryPI: Drop CAR teardown without POSTCAR_STAGE
Occassionally (maybe 1 boot in 10) SMP lapic_cpu_init() fails
with following errors in the logs of pcengines/apu2:
CPU 0x03 would not start!
CPU 0x03 did not initialize!
The CPU number is sometimes 0x02, never seen 0x01. Work-around also
suggests something to do with cache coherency and MTRRs that is really
at fault.
As a work-around return the BSP CAR teardown to use wbinvd instead
of invd. These platforms do not support S3 resume so this is the
easy work-around for the time being.
Change-Id: I3dac8785aaf4af5c7c105ec9dd0b95156b7cca21
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37438
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
&imd_cbmem is never NULL, so remove that path
Change-Id: Ib9a9c88d6cd4842df447f046bc0abaa7ef5032c7
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37361
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
BUG=b:144002424
TEST=Ensured no TPM time out issue and system can boot to OS
Change-Id: I7282e6c2d9627846039638bdc0db3ee7ebba5f12
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/37320
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Kyösti Mälkki