This is part of a series to move memory management functions
away from the z_ namespace and into its own namespace. Also
make documentation available via doxygen.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Move this to a call in the init process. arch_* calls are no services
and should be called consistently during initialization.
Place it between PRE_KERNEL_1 and PRE_KERNEL_2 as some drivers
initialized in PRE_KERNEL_2 might depend on SMP being setup.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Platforms that support IPIs allow them to be broadcast via the
new arch_sched_broadcast_ipi() routine (replacing arch_sched_ipi()).
Those that also allow IPIs to be directed to specific CPUs may
use arch_sched_directed_ipi() to do so.
As the kernel has the capability to track which CPUs may need an IPI
(see CONFIG_IPI_OPTIMIZE), this commit updates the signalling of
tracked IPIs to use the directed version if supported; otherwise
they continue to use the broadcast version.
Platforms that allow directed IPIs may see a significant reduction
in the number of IPI related ISRs when CONFIG_IPI_OPTIMIZE is
enabled and the number of CPUs increases. These platforms can be
identified by the Kconfig option CONFIG_ARCH_HAS_DIRECTED_IPIS.
Signed-off-by: Peter Mitsis <peter.mitsis@intel.com>
Make `struct arch_esf` compulsory for all architectures by
declaring it in the `arch_interface.h` header.
After this commit, the named struct `z_arch_esf_t` is only used
internally to generate offsets, and is slated to be removed
from the `arch_interface.h` header in the future.
Signed-off-by: Yong Cong Sin <ycsin@meta.com>
The irq_offload mechanism was using the same entry of the IDT vector for
all CPUs on SMP systems. This caused race conditions when two CPUs were
doing irq_offload() calls.
This patch addresses that by adding one indirection layer: the
irq_offload() now sets a per CPU entry with the routine and parameter to
be run. Then a software interrupt is generated, and a default handler
will do the appropriate dispatching.
Finally, test "kernel/smp_abort" is enabled for x86 as it should work
now.
Fixes#72172.
Signed-off-by: Ederson de Souza <ederson.desouza@intel.com>
Remove the limitation of number of cpu support in x86 arch.
Also add support for retrieve cpu informations such as for
hybird cores.
Signed-off-by: Najumon B.A <najumon.ba@intel.com>
This adds the necessary bits to enable memory mapping thread
stacks on both x86 and x86_64. Note that currently these do
not support multi level mappings (e.g. demand paging and
running in virtual address space: qemu_x86/atom/virt board)
as the mapped stacks require actual physical addresses.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
When kernel stack is defined as an array, K_KERNEL_STACK_LEN()
is used to calculate the size for each stack in the array.
However, standalone kernel stack has its size calculated by
Z_KERNEL_STACK_SIZE_ADJUST() instead. Depending on the arch
alignment requirement, they may not be the same... which
could cause some confusions. So align them both to use
K_KERNEL_STACK_LEN().
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Simple rename to align the kernel naming scheme. This is being
used throughout the tree, especially in the architecture code.
As this is not a private API internal to kernel, prefix it
appropriately with K_.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
During early boot in assembly, the function parameter to
z_x86_early_tls_update_gdt() should be the pointer to
the interrupt stack. However, what was passed instead
was the pointer to the x86_cpuboot struct. So fix it to
actually pass the stack pointer (which is stashed inside
the x86_cpuboot struct).
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
For some reason, unrelated code change triggered compiler
warning about this function returns even though it is
marked nonreturn. So add CODE_UNREACHABLE to silence
the warning, possibly to catch any errors.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
When building without optimizations and with only one core the linker
does not throw away arch_start_cpu and we get an undefined reference to
x86_ap_start
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
This moves the k_* memory management functions from sys/ into
kernel/ includes, as there are kernel public APIs. The z_*
functions are further separated into the kernel internal
header directory.
Also made a quick change to doxygen to group sys_mem_* into
the OS Memory Management group so they will appear in doc.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Use ACPI_MADT_LOCAL_APIC instead of struct acpi_madt_local_apic. In the
same go, switch to IF_ENABLED from ifdef - slightly more readable, and
this keeps some static analyzers happy (e.g. upstream Compliance check).
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
After 79d0bf39b8 was merged, the inclusion
of <zephyr/acpi/acpi.h> with CONFIG_ACPI=n caused a build failure because
<acpica/source/include/acpi.h> could no longer be included due to the
inlcude path not being injected anymore.
Fix this by guarding the header inclusion when CONFIG_ACPI
is not set.
Fixes#62679.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
The old acpi implimentation is replaced with acpica interface
and updated x86 arch porting with the new interface.
Signed-off-by: Najumon B.A <najumon.ba@intel.com>
As per #26393, Local APIC is using Kconfig based option for
the base address. This patch adds DTS binding support in the driver,
just like its conunter part I/O APIC.
Signed-off-by: Umar Nisar <umar.nisar@intel.com>
With all the stacks and TSS (etc), the x86_64 arch code can only
support maximum of 4 CPUs at the moment. So add a build assert
if more CPUs are specified via CONFIG_MP_MAX_NUM_CPUS, also
overwrite the range value for CONFIG_MP_MAX_NUM_CPUS.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Until now iterable sections APIs have been part of the toolchain
(common) headers. They are not strictly related to a toolchain, they
just rely on linker providing support for sections. Most files relied on
indirect includes to access the API, now, it is included as needed.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Move the FS_BASE MSR code to to the top of __resume to ensure
that %fs relative addressing run in the thread switching hook
works.
Signed-off-by: Keith Packard <keithp@keithp.com>
It is frequent to find variable definitions like this:
```c
static const struct device *dev = DEVICE_DT_GET(...)
```
That is, module level variables that are statically initialized with a
device reference. Such value is, in most cases, never changed meaning
the variable can also be declared as const (immutable). This patch
constifies all such cases.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
MISRA C:2012 Rule 8.2 (Function types shall be in prototype form with
named parameters.)
Added missing parameter names.
Signed-off-by: Abramo Bagnara <abramo.bagnara@bugseng.com>
Assembler files were not migrated with the new <zephyr/...> prefix.
Note that the conversion has been scripted, refer to #45388 for more
details.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
In order to bring consistency in-tree, migrate all arch code to the new
prefix <zephyr/...>. Note that the conversion has been scripted, refer
to zephyrproject-rtos#45388 for more details.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
As for Multiboot, let prep_c be aware of EFI boot.
In the futur, EFI will pass an argument to it.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
In order to mitigate at runtime whether it booted on multiboot or EFI,
let's introduce a dedicated x86 cpu argument structure which holds the
type and the actual pointer delivered by the method (multiboot_info, or
efi_system_table)
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Improve code by using DEVICE_DT_GET_ONE instead of device_get_binding,
since the intel_vt_d device instance can be obtained at compile time.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
When XIP is not enabled, z_data_copy() already falls back to an empty
function. No need to ifdef it.
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
Allocate an IRTE for all irq being connected through
arch_irq_connect_dynamic(). This will be mandatory since VT-D expects to
filter all interruptions (but the one it generates, as we will see
later).
Taking into account CONFIG_INTEL_VTD_ICTL_XAPIC_PASSTHROUGH, which could
help for debugging.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Call into z_thread_usage_stop() before ISR entry to avoid including
interrupt handling totals in thread usage stats.
This has to go into the assembly immediately before the callback-based
dispatch. Note that the dispatch code was putting the vector number
in RCX, which was unfortunate as that's a caller-saved register.
Would be nice to clean this up in the future so it lives in a
preserved register but it's mildly complicated to make work with the
way we do the stack layout right now.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
This adds arch_float_enable() and arch_float_disable() to x86-64.
As x86-64 always has FP/SSE enabled, these operations are basically
no-ops. These are added just for the completeness of arch interface.
Fixes#38022
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
The 16 bit bootstrap code for SMP CPUs was using the 286-era "lmsw"
instruction (load machine status word) to set the protected bit in CR0
(which is the modern evolution of the same register), presumably
because this is 16 bit code and we can't move a dword into CR0.
But that's wrong, because the full instruction set *is* available in
real mode on a 386, you just have to use a operand size prefix to get
to it, which the assembler emits for you automatically when you use
the .code16 directive.
Write this conventionally and use modern (e.g. 1986-era) instructions.
It also has the advantage of not confusing much more modern
hypervisors like ACRN by issuing instructions they (and I!) never knew
existed.
Fixes#35076
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Because of a historical misunderstanding, by default the ACRN
hypervisor wants to load Zephyr at address 0x1000 and enter the binary
at that same address. This entry point corresponds to the __start
symbol of the build they were given, which is a 1-cpu non-SMP
configuration. Unfortunately, when we build with
CONFIG_MP_NUM_CPUS=1, the code in locore.S #if's out the 16 bit entry
point for the auxiliary CPUs at the start of the section. So in the
build ACRN received, the start address happened to be 0x7000, the same
address we need to launch the AP processors from.
That's right: under ACRN, the SAME ADDRESS used to enter the OS in 32
bit mode needs to be used later to boot CPUs running in 16 bit real
mode!
The solution, such as it is, is to put a 32 bit jump at the entry
address which hops to the 32 bit OS entry code, and then scribble NOP
instructions over that jump once we get there so that the next time we
reach that address (in real mode) we fall through to the correct
entry.
This patch should be considered a temporary workaround. While it
works on all x86 hardware, it's not really needed. A much better
solution would be to eliminate the locore linker region entirely
(which causes other headaches) and enter the Zephyr binary in a 32 bit
address somewhere in the contiguous high memory area. All that locore
is needed for is the 16 bit bootstrap code for SMP processors, which
is ~6 instructions and can be copied in from the kernel at runtime.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Both operands of an operator in which the usual arithmetic
conversions are performed shall have the same essential
type category.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
This reverts commit 7d32e9f9a5.
We now allow the kernel to be linked virtually. This patch:
- Properly converts between virtual/physical addresses
- Handles early boot instruction pointer transition
- Double-maps SRAM to both virtual and physical locations
in boot page tables to facilitate instruction pointer
transition, with logic to clean this up after completed.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Instead of doing these in assembly, use the common z_bss_zero()
and z_data_copy() C functions instead. This simplifies code
a bit and we won't miss any additions to these two functions
(if any) under x86 in the future (as x86_64 was actually not
clearing gcov bss area).
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This moves calling z_loapic_enable() from crt0.S into
z_x86_prep_c(). This is done so we can move BSS clearing
and data section copying inside z_x86_prep_c() as
these are needed before calling z_loapic_enable().
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
The x86_64 SysV ABI requires 16 byte alignment for the stack pointer
during execution of normal code. That means that on entry to an
ABI-compatible C function (which is reached via a CALL instruction
that pushes the return address) the RSP register must be MISaligned by
exactly 8 bytes. The kernel mode thread setup got this right, but we
missed the equivalent condition in userspace entry.
The end result was a misaligned stack, which is surprisingly robust
for most use. But recent toolchains have starting doing some more
elaborate vectorization, and the resulting SSE instructions started
failing in userspace on the misaliged loads.
Note that there's a comment about optimization: we're doing the stack
alignment in the "wrong place" and are needlessly wasting bytes in
some cases. We should see the raw stack boundaries where we are
setting up RSP values. Add a FIXME to this effect, but don't touch
anything as this patch is a targeted bugfix.
Also fix a somewhat embarassing 32-bit-ism that would have truncated
the address of a userspace stack that we tried to put above 4G.
Fixes#31018
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
