The push/pop of %ebx was only added because smm_stub saves the canary
value in it. Now that we no longer use cpu_info in smm, we no longer
need to save the register.
BUG=b:179699789
TEST=Boot guybrush to the OS
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I554dbe016db8b1c61246c8ffc7fa252b2542ba92
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58205
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
Now that cpu_info() is no longer used by COOP_MULTITASKING, we no
longer need to set up cpu_info in SMM. When using CPU_INFO_V2, if
something does manage to call cpu_info() while executing in SMM mode,
the %gs segment is disabled, so it will generate an exception.
BUG=b:179699789
TEST=Boot guybrush to OS with threads enabled
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: Id64f32cc63082880a92dab6deb473431b2238cd0
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58204
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
Reviewed-by: Paul Menzel <paulepanter@mailbox.org>
We only ever start and execute threads on the BSP. By explicitly
checking to see if the CPU is the BSP we can remove the dependency on
cpu_info. With this change we can in theory enable threads in all
stages.
BUG=b:194391185, b:179699789
TEST=Boot guybrush to OS and verify coop multithreading still works
Suggested-by: Julius Werner <jwerner@chromium.org>
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: Iea4622d52c36d529e100b7ea55f32c334acfdf3e
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58199
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
For older CPU models where CPUID leaf 0xb is not supported, use
initial LAPIC ID from CPUID instead of LAPIC register space to
to detect if logical CPU is a hyperthreading sibling. The one
in LAPIC space is more complex to read, and might not reflect
CPU topology as it can be modified in XAPIC mode.
Change-Id: I8c458824db1ea66948126622a3e0d0604e391e4b
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/58385
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
Remove code, which was only needed for B and C2 stepping
of P54C. The linux kernel source has commentary on X86_BUG_11AP:
* See if we have a good local APIC by checking for buggy Pentia,
* i.e. all B steppings and the C2 stepping of P54C when using their
* integrated APIC (see 11AP erratum in "Pentium Processor
* Specification Update")
Change-Id: Iec10335f603674bcef2e7494831cf11200795d38
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55199
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
Make use of the newly introduced ACPI macros for CPPC table generation
that currently exists of a bunch of confusing assignments of structs
that only get partially filled.
Test: dumped SSDT before and after do not differ.
Change-Id: I844d191b1134b98e409240ede71e2751e51e2159
Signed-off-by: Michael Niewöhner <foss@mniewoehner.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57888
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Reviewed-by: Lance Zhao
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
There is currently a fundamental flaw in the current cpu_info()
implementation. It assumes that current stack is CONFIG_STACK_SIZE
aligned. This assumption breaks down when performing SMM relocation.
The first step in performing SMM relocation is changing the SMBASE. This
is accomplished by installing the smmstub at 0x00038000, which is the
default SMM entry point. The stub is configured to set up a new stack
with the size of 1 KiB (CONFIG_SMM_STUB_STACK_SIZE), and an entry point
of smm_do_relocation located in RAMSTAGE RAM.
This means that when smm_do_relocation is executed, it is running in SMM
with a different sized stack. When cpu_info() gets called it will be
using CONFIG_STACK_SIZE to calculate the location of the cpu_info
struct. This results in reading random memory. Since cpu_info() has to
run in multiple environments, we can't use a compile time constant to
locate the cpu_info struct.
This CL introduces a new way of locating cpu_info. It uses a per-cpu
segment descriptor that points to a per-cpu segment that is allocated on
the stack. By using a segment descriptor to point to the per-cpu data,
we no longer need to calculate the location of the cpu_info struct. This
has the following advantages:
* Stacks no longer need to be CONFIG_STACK_SIZE aligned.
* Accessing an unconfigured segment will result in an exception. This
ensures no one can call cpu_info() from an unsupported environment.
* Segment selectors are cleared when entering SMM and restored when
leaving SMM.
* There is a 1:1 mapping between cpu and cpu_info. When using
COOP_MULTITASKING, a new cpu_info is currently allocated at the top of
each thread's stack. This no longer needs to happen.
This CL guards most of the code with CONFIG(CPU_INFO_V2). I did this so
reviewers can feel more comfortable knowing most of the CL is a no-op. I
would eventually like to remove most of the guards though.
This CL does not touch the LEGACY_SMP_INIT code path. I don't have any
way of testing it.
The %gs segment was chosen over the %fs segment because it's what the
linux kernel uses for per-cpu data in x86_64 mode.
BUG=b:194391185, b:179699789
TEST=Boot guybrush with CPU_INFO_V2 and verify BSP and APs have correct
%gs segment. Verify cpu_info looks sane. Verify booting to the OS
works correctly with COOP_MULTITASKING enabled.
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I79dce9597cb784acb39a96897fb3c2f2973bfd98
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57627
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Eric Peers <epeers@google.com>
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
The %fs and %gs segment are typically used to implement thread local
storage or cpu local storage. We don't currently use these in coreboot,
so there is no reason to map them. By setting the segment index to 0,
it disables the segment. If an instruction tries to read from one of
these segments an exception will be raised.
The end goal is to make cpu_info() use the %gs segment. This will remove
the stack alignment requirements and fix smm_do_relocation.
BUG=b:194391185, b:179699789
TEST=Boot guybrush to OS
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: Iaa376e562acc6bd1dfffb7a23bdec82aa474c1d5
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57860
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Eric Peers <epeers@google.com>
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
This will help reduce duplication and make it easier to add new members
to the cpu_info struct.
BUG=b:194391185, b:179699789
TEST=Compare assembly of romstage and ramstage before and after
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I31f264f4bb8b605fa3cb3bfff0d9bf79224072aa
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57859
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Karthik Ramasubramanian <kramasub@google.com>
The code under `cpu/x86/tsc` is only compiled in when its `Makefile.inc`
is included from platform (CPU/SoC) code and the `UDELAY_TSC` Kconfig
option is enabled.
Include `cpu/x86/tsc/Makefile.inc` once from `cpu/x86/Makefile.inc` and
drop the now-redundant inclusions from platform code. Also, deduplicate
the `UDELAY_TSC` guards.
Change-Id: I41e96026f37f19de954fd5985b92a08cb97876c1
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/57456
Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-by: Felix Held <felix-coreboot@felixheld.de>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
IDS (Integrated Debug Services) options are meant to be enabled when one
wants to debug AGESA. Since they are compile-time options, using Kconfig
is the logical choice. Currently, none of the options builds.
Tested with BUILD_TIMELESS=1 without adding the configuration options
into the binary, and Asus A88XM-E does not change.
Change-Id: I465627c19c9856e58ca94aa0efedbddb6baaf3f6
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/53985
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: David Hendricks <david.hendricks@gmail.com>
Since current AMD SoCs don't need some wait time between INIT and SIPI,
we can skip the 10ms wait there, which improves the boot time a bit.
before: CPU_CLUSTER: 0 init finished in 632 msecs
after: CPU_CLUSTER: 0 init finished in 619 msecs
mpinit still works on Mandolin and all CPU cores show up and are usable.
This also doesn't change the binary in a timeless build for boards/SoCs
that don't select X86_AMD_INIT_SIPI which I verified for lenovo/x230.
BUG=b:193885336
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Change-Id: I1e044776f45021742a88a5e369a74383c1baaab6
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56533
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Marshall Dawson <marshalldawson3rd@gmail.com>
The alignment for `struct cpu_info` is wrong on x86_64. c_start.S uses
the `push` instruction when setting up the cpu_info struct. This
instruction will push 8 bytes but `unsigned int` is 4 bytes. By making
it a `size_t` we get the correct size for both x86_32 and x86_64.
BUG=b:179699789
TEST=Boot guybrush to the OS
Suggested-by: Julius Werner <jwerner@chromium.org>
Signed-off-by: Raul E Rangel <rrangel@chromium.org>
Change-Id: I8ef311aaa8333ccf8a5b3f1f0e852bb26777671c
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56573
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
Use the common mca_get_bank_count function instead of open-coding the
functionality to get the MCA bank number. Also re-type the num_banks
variable from signed in to unsigned int, since the number of MCA bank is
always positive, and make it constant.
In the case of Intel model 2065x the mca_get_bank_count() call replaces
a magic number.
Change-Id: I245b15f57e77edca179e9e28965383a227617174
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56244
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
When accessing the MCA MSRs, the MCA bank number gets multiplied by 4
and added to the IA32_MC0_* define to get the MSR number. Add a macro
that already does this calculation to avoid open coding this repeatedly.
Change-Id: I2de753b8c8ac8dcff5a94d5bba43aa13bbf94b99
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56243
Reviewed-by: Raul Rangel <rrangel@chromium.org>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Use the common mca_get_bank_count function instead of open-coding the
functionality to get the MCA bank number. Also re-type the num_banks
variable from signed in to unsigned int, since the number of MCA bank is
always positive.
Change-Id: I70ad423aab484cf4ec8f51b43624cd434647aad4
Signed-off-by: Felix Held <felix-coreboot@felixheld.de>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56184
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Michał Żygowski <michal.zygowski@3mdeb.com>
Commit 1aa60a95bd broke microcode loading for chipsets that have a
microcode blob with a total_size field set to 0. This appears to be
support for older chipsets, where the size was set to 0 and assumed to
be 2048 bytes. The fix is to change the result of the subtraction to a
signed type, and ensure the following comparison is done without
promoting the signed type to an unsigned one.
Resolves: https://ticket.coreboot.org/issues/313
Change-Id: I62def8014fd3f3bbf607b4d58ddc4dca4c695622
Signed-off-by: Tim Wawrzynczak <twawrzynczak@chromium.org>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/56153
Reviewed-by: Furquan Shaikh <furquan@google.com>
Reviewed-by: Nico Huber <nico.h@gmx.de>
Reviewed-by: Stefan Ott <coreboot@desire.ch>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Introduce `USE_EXP_X86_64_SUPPORT` in `src/arch/x86/Kconfig` and guard
it with `HAVE_EXP_X86_64_SUPPORT`. Replace the per-CPU implementations
of the same functionality with the newly-added Kconfig options. Update
documentation and the config file for QEMU accordingly.
Change-Id: I550216fd2a8323342d6b605306b0b95ffd5dcd1c
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55760
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
Introduce the `ARCH_ALL_STAGES_X86` Kconfig symbol to automatically
select the per-stage arch options. Subsequent commits will leverage
this to allow choosing between 32-bit and 64-bit coreboot where all
stages are x86. AMD Picasso and AMD Cezanne are the only exceptions
to this rule: they disable `ARCH_ALL_STAGES_X86` and explicitly set
the per-stage arch options accordingly.
Change-Id: Ia2ddbae8c0dfb5301352d725032f6ebd370428c9
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55759
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
To generalise the choice of 32-bit or 64-bit coreboot on x86 hardware,
have platforms select `ARCH_X86` directly instead of through per-stage
Kconfig options, effectively reversing the dependency order.
Change-Id: If15436817ba664398055e9efc6c7c656de3bf3e4
Signed-off-by: Angel Pons <th3fanbus@gmail.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55758
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Patrick Rudolph <siro@das-labor.org>
This commit adds a method called `mainboard_smi_finalize` which provides
a mechanism for a mainboard to execute some code as part of the finalize
method in the SMM stage before SoC does its finalization.
BUG=b:191189275
BRANCH=None
TEST=Implement `mainboard_smi_finalize` on lalala and verify that the
code executes in SMM.
Signed-off-by: Aseda Aboagye <aaboagye@google.com>
Change-Id: If1ee63431e3c2a5831a4656c3a361229acff3f42
Reviewed-on: https://review.coreboot.org/c/coreboot/+/55649
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Karthik Ramasubramanian <kramasub@google.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>
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>
