In acpi_soc_get_bert_region after the bert_errors_region call is was
checked if the region parameter is NULL after the call; since region is
a parameter of acpi_soc_get_bert_region, it's non-NULL. What we should
be checking here is if region points to a non-NULL pointer.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reported-by: Coverity (CID:1457506)
Change-Id: I0523504d65725ab2d2df4db28a5dedd90697b917
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55541
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Even though the code is currently commented out, replace the magic
numbers with the existing defines.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Id0dbbadf71f2e5a4d23ee998e2aa0a8b67205845
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55534
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Factoring out those defines allows using them easily in the ACPI code
without having to use preprocessor macros.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: Ib9dfddb0d4f32a542fa652ff8c14e932c224f247
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55533
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Factoring out those defines allows using them easily in the ACPI code
without having to use preprocessor macros.
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I99cb03de8782a0eeeb505f567b982099b0e8a18d
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55531
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Lynx Point PCH ACPI reference code version 1.9.1 has two additional
magic steps, which were already present in Broadwell. Add them.
Change-Id: Ia8ca6dcfcfb4ed6b0d957d249b93640ef74670d7
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55494
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
The printed values are unsigned, and should be printed accordingly.
Change-Id: Ie5edce914c389c70460b1ed3390731e3568340dd
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55493
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
The typedef needlessly hides the actual type of the variables.
Tested with BUILD_TIMELESS=1, Purism Librem 13 v1 remains identical.
Change-Id: I58a58cd402ec679960f460e80b37ff2afb8e3974
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55492
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
This option is valid for Broadwell as well as Haswell.
Change-Id: I4f1e9663806bae279f6aca36f09a0c989c12e507
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55491
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Split up PCH Kconfig into a separate file. While we're at it, also sort
selected options alphabetically.
Tested with BUILD_TIMELESS=1, coreboot.rom for the Purism Librem 13 v1
remains identical when not adding the .config file in it.
Change-Id: Ic3ff982e7108bf2d25a22e56ac2fbb93070df164
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55490
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
The BMC EEPROM layout has been updated to contain system and mainboard
serial numbers. Use these values in SMBIOS Type 1 and Type 2 tables.
Change-Id: I55b51a856b4ad28fd56b02015b2e1d49cd629735
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55275
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
These macros were used to generate ACPI P-state entries, but Broadwell
now uses Haswell CPU code. These macros are unused and can be removed.
Tested with BUILD_TIMELESS=1, Purism Librem 13 v1 remains identical.
Change-Id: Ib2baca2964d9177e7ab6630d4ced22c5d332fb6e
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55487
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Keep deduplicating code. Have Broadwell PCH ASL borrow some equivalent
Lynx Point ASL files, and drop the now-unused files from Broadwell PCH.
Tested with BUILD_TIMELESS=1, Purism Librem 13 v1 remains identical.
Change-Id: If5a8712a846bbf7c42db92167763935dee74c26f
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55485
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Use hostbridge.asl from Haswell instead of Broadwell. Both files are
equivalent. Then, drop the now-unused hostbridge.asl from Broadwell.
Tested with BUILD_TIMELESS=1, Purism Librem 13 v1 remains identical.
Change-Id: I87d51727b75a9c59e2f5f3ba8d48c575ce93c78c
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55484
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Include Haswell memmap.h from Broadwell iomap.h to deduplicate identical
definitions. This also prevents the definitions from falling out of sync
while the unification process is ongoing.
Tested with BUILD_TIMELESS=1, Purism Librem 13 v1 remains identical.
Change-Id: I850e5521effba3818f4e2a13b94281bf07857d50
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55483
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Move the inclusion of `pci_irqs.asl` into PCH scope in order to allow
deduplicating northbridge ACPI code.
Change-Id: I541913226b26662f3798ae9c25ab1ac33cf2ed45
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55482
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Taken from Haswell code. These resources also exist on Broadwell and
should be reported to the OS.
Change-Id: I45f2a6a9140d72c1cc2ee8b72621dc16c815b621
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55481
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
GDXCBAR and EDRAMBAR are accounted for when reporting resources to the
allocator, but they are not present in the DSDT. In addition, coreboot
does not enable either range, but MRC.bin sets up GDXCBAR and does not
disable it afterwards. Not reporting GDXCBAR in the DSDT can result in
resource conflicts, and not enabling EDRAMBAR can cause issues on CPUs
with eDRAM.
Enable both GDXCBAR and EDRAMBAR in coreboot code, and report these
ranges in the DSDT. This matches what Broadwell does. The value for
the `GDXC_BASE_ADDRESS` macro matches what MRC.bin programs as well.
Tested on Asrock B85M Pro4 with an i7-4770S (no eDRAM):
- Still boots
- EDRAMBAR is now enabled with base address of 0xfed80000
- GDXCBAR is still mapped with base address of 0xfed84000
Change-Id: I5538873b30e3d02053e4ba125528d32453ef6572
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55480
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Add defines for the sizes of northbridge MMIO windows and use them where
applicable. The macro names have been taken from Broadwell.
Tested with BUILD_TIMELESS=1, Asrock B85M Pro4 remains identical.
Change-Id: I845cba8acbd478cd325d2e364138336d985f9c34
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55479
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Mainboard information can be found in the included documentation.
Signed-off-by: Bill XIE <persmule@hardenedlinux.org>
Change-Id: Ic811e24bd72da84e5ca8f5b09f2eb65872153b72
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55111
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
This is only called locally.
Change-Id: Ie3eaf659a2868eee1d4688885495c413f94f42e2
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55469
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Johnny Lin <Johnny_Lin@wiwynn.com>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Christian Walter <christian.walter@9elements.com>
1) FSP-S should not run XIP
2) Overriding the FSP-T location conflicts with the location set in
drivers/intel/fsp2_0
This fixes a regression caused by commit
0f068a600e (drivers/intel/fsp2_0: Fix the FSP-T position)
Link: https://mail.coreboot.org/hyperkitty/list/coreboot@coreboot.org/thread/G6WRFITANOS2JEYG3GKB2ZNVCLUZ6W7P/
Change-Id: I381781c1de7c6dad32d66b295c927419dea7d8be
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55389
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Mariusz Szafrański <mariuszx.szafranski@intel.com>
Reviewed-by: King Sumo <kingsumos@gmail.com>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Avoid using deprecated macros, where possible.
"GpioResetPwrGood" represents multiple valid updated values, depending
on the GPIO community and will be more difficult to update.
While Kabylake supports both sets of macros, it will cause build errors
on Coffeelake. In the GPD group, replace with "GpioDswReset."
Replace with "GpioResumeReset" in any GPP group.
Change-Id: Iab0bb09adad997bef3a2133c443471d4c634f423
Signed-off-by: Benjamin Doron <benjamin.doron00@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/54974
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Maxim Polyakov <max.senia.poliak@gmail.com>
One of the Memory32Fixed entries covers the TXT private and public
spaces, and another covers the TPM registers. Update the comments.
Change-Id: I261d74c113fabf1d152964efd8c91de85eba4179
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55462
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
Move locking CPU MSRs during CPU init instead of using
CONFIG_PARALLEL_MP_AP_WORK functions.
The AES Lock enable bit caused CPU exception errors as this should not
run on HT siblings. The set_aesni_lock() function takes care of that.
Change-Id: I21598c3e9a153dce25a09b187ddf9cf6363039d3
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55098
Reviewed-by: Sumeet R Pawnikar <sumeet.r.pawnikar@intel.com>
Reviewed-by: Christian Walter <christian.walter@9elements.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Disable all display related UPDs if IGD is not enabled as FSP
don't need to perform display port initialization while IGD itself
is disabled else assign UPDs based on devicetree config.
TEST=Dump FSP-M display related UPDs with IGD enable and disable
to ensure patch integrity.
Change-Id: I0479904141dfc5e707679109aa18b7ef4264cf96
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55273
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Rizwan Qureshi <rizwan.qureshi@intel.com>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
1. Replace all pcidev_path_on_root() and is_dev_enabled() functions
combination with is_devfn_enabled().
2. Remove unused local variable of device structure type
(struct device *).
3. Replace pcidev_path_on_root() and dev->enabled check with
is_devfn_enabled() call.
TEST=Able to build and boot without any regression seen on ADL.
Change-Id: I92671992ec14fd2adca1635b0791ac8b456332e9
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55292
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
1. Replace all pcidev_path_on_root() and is_dev_enabled() functions
combination with is_devfn_enabled().
2. Remove unused local variable of device structure type (struct device *).
3. Replace pcidev_path_on_root() and dev->enabled check with is_devfn_enabled()
call.
TEST=Able to build and boot without any regression seen on EHL.
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Change-Id: Iadf9145a11f27ff0e182f146b6fe5a01e6cf3ed8
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55331
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Lean Sheng Tan <lean.sheng.tan@intel.com>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
1. Replace all pcidev_path_on_root() and is_dev_enabled() functions
combination with is_devfn_enabled().
2. Remove unused local variable of device structure type (struct device *).
3. Replace pcidev_path_on_root() and dev->enabled check with is_devfn_enabled()
call.
TEST=Able to build and boot without any regression seen on TGLRVP.
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Change-Id: Ic9d91b711bab83de1911e0b7ea876f2ad018c937
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55330
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
1. Replace all pcidev_path_on_root() and is_dev_enabled() functions
combination with is_devfn_enabled().
2. Remove unused local variable of device structure type (struct device *).
3. Replace pcidev_path_on_root() and dev->enabled check with is_devfn_enabled()
call.
TEST=Able to build and boot without any regression seen on dedede
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Change-Id: I4919a1ec02df50bc41fd66d5f3a352108a7aa04c
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55329
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
1. Replace all pcidev_path_on_root() and is_dev_enabled() functions
combination with is_devfn_enabled().
2. Remove unused local variable of device structure type (struct device *).
3. Replace pcidev_path_on_root() and dev->enabled check with is_devfn_enabled()
call.
TEST=Able to build and boot without any regression seen on CML.
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Change-Id: Ib5df5fd32e2e2742d349754a942bf81ca505dd39
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55328
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
1. Replace all pcidev_path_on_root() and is_dev_enabled() functions
combination with is_devfn_enabled().
2. Remove unused local variable of device structure type
(struct device *).
3. Replace pcidev_path_on_root() and dev->enabled check with
is_devfn_enabled() call.
TEST=Able to build and boot without any regression seen on Reef.
Signed-off-by: Subrata Banik <subrata.banik@intel.com>
Change-Id: I900038dd4b2e2d89b1236bbd26bec5f34483b9f3
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55327
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Trigger mode LAPIC_INT_LEVELTRIG was only used with LAPIC_DM_INIT,
specifically for (obsolete) Init Level De-assert.
Level LAPIC_INT_ASSERT is required to be set for all other delivery
modes other than LAPIC_DM_INIT.
This reverts the two above changes that X2APIC mode support introduced
to the IPI for LAPIC_DM_SMI.
Change-Id: I7264f39143cc6edb7a9687d0bd763cb2703a8265
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55197
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Depending on how the "middle-end" (yes, the gcc developers are
serious about that) optimizer ends up mangling the code, there may
or may not be a complaint about x being used uninitialized when it's
clearly not used at all.
So instead, why keep x in the first place? memcpy(foo, NULL, 0) is
the same as memcpy(foo, some_uninitialized_variable, 0) in that it
does nothing.
Change-Id: Ib0a97c3e3fd1a2a6aff37da63376373c88ac595d
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55499
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
This update adds a commit to fix building libgfxinit with gcc 11
Change-Id: I5c0e3823ab7219667f9430bce74e4f2fba0c0c3a
Signed-off-by: Patrick Georgi <pgeorgi@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55497
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
CB:55356 removed static inline declarations from get_log_level(). This
commit puts them back. It also changes the method of accessing static
symbols in tests/console/routing-test to source file inclusion like
in CB:46458 to avoid changing tested source file.
Signed-off-by: Jakub Czapiga <jacz@semihalf.com>
Change-Id: Iaa5dcbccb327f819374967be51ef642b1fb25e7b
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55473
Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Add support for the ThinkPad W541 based on Peter Lemenkov's initial W541
port. Compiled and tested with SeaBIOS and Tianocore booting into Arch
Linux 5.10.32-lts. The Haswell mrc.bin blob is required.
Tested working:
- SATA SSD
- SATA DVD drive
- M.2 SATA
- All USB ports
- SD card reader
- Speakers/headphone jack
- Keyboard/touchpad
- libgfxinit
- VGA
- mini DisplayPort (Thunderbolt untested)
- eDP laptop screen
- NVIDIA GPU in Linux
- Camera/Mic
- Smartcard reader
- Internal flashing when IFD is unlocked
- ThinkPad basic dock (VGA, USB, Ethernet)
- CMOS options
- WLAN
- Bluetooth
- Ethernet
- Using me_cleaner
- All DDR3 slots
Not working:
- Keyboard backlight
- First boot can take up to 20s (MRC.bin is slow)
Untested:
- Thunderbolt
- Internal flashing when IFD is locked
- Other ThinkPad docks (DisplayPort, DVI, Audio)
- ExpressCard slot
- Battery thresholds
- WWAN card
- Fingerprint reader
- USB Debug console
Signed-off-by: Justin Wu <amersel@runbox.me>
Change-Id: Ia43070f51bba3cf59ba9b7d9e29e4e778efbeb08
Reviewed-on: https://review.coreboot.org/c/coreboot/+/52659
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Peter Lemenkov <lemenkov@gmail.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Fix compilation on x86_64 by using compatible types.
The MRC blob isn't supported yet as there's no x86_32 wrapper.
Tested on HP8200:
* Still boots on x86_32.
* Boots to payload in x86_64
Change-Id: Iab29a87d52ad3f6c480f21a3b8389a7f49cb5dd8
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/44677
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
In x86_64 code every function call consumes 32byte of stack with
no stack local variables being used. That limits the function call depth
in SMM to 32 or less.
Double the stack size to prevent overwriting the stack canary as seen
on HP8200 and x86_64 enabled.
Change-Id: Iee202ba2ae609a474d0eb3b06f49690f33f4eda8
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55449
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
This fixes a hard to debug hang that could occur in any stage, but in
the end it follows simple rules and is easy to fix.
In long mode the 32bit displacement addressing used on 'mov' and 'lea'
instructions is sign-extended. Those instructions can be found using
readelf on the stage and searching for relocation type R_X86_64_32S.
The sign extension is no issue when either running in protected mode or
the code module and thus the address is below 2GiB. If the address is
greater than 2GiB, as usually the case for code in TSEG, the higher
address bits [64:32] are all set to 1 and the effective address is
pointing to memory not paged. Accessing this memory will cause a page
fault, which isn't handled either.
To prevent such problems
- disable R_AMD64_32S relocations in rmodtool
- add comment explaining why it's not allowed
- use the pseudo op movabs, which doesn't use 32bit displacement addressing
- Print a useful error message if such a reloc is present in the code
Fixes a crash in TSEG and when in long mode seen on Intel Sandybridge.
Change-Id: Ia5f5a9cde7c325f67b12e3a8e9a76283cc3870a3
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55448
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Felix Held