The SOC code can set whether it's known that the CPU
is immune, don't default to turning this off.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Retpolines were never completely implemented, even on x86.
Move this particular Kconfig to only concern itself with
the assembly code, and don't default it on ever since we
prefer SSBD instead.
We can restore the common kernel-wide CONFIG_RETPOLINE once
we have an end-to-end implementation.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
PAE page tables (the only kind we support) have 512
entries per page directory, not 1024.
Fixes: #13838
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
We don't need the build system to pull out the mmu
region specifiers from the kernel binary when the
script can just as easily do this itself.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
This is an integral part of userspace and cannot be used
on its own. Fold into the main userspace configuration.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Incremental builds have been broken in x86 due to a misconfigured
dependency. mmu_tables.bin is always generated, even for "nothing to
do" builds.
We fix this by removing the stray dependency on user_mmu_tables.bin
when not CONFIG_X86_KPTI.
Steps to reproduce:
Build any sample twice with qemu_x86 and observe that the second build
regenerates mmu_tables.bin.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
The same pattern is used five times In the x86 build scripts and the
same code has been copied and modified the same amount of times. This
has resulted in a system that is difficult to make changes to.
To enforce consistency and improve maintainability we refactor the
code into a function.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
In general, to have correct dependencies, one must not only depend on
files, but also a wrapper target for the file. This is done for some
of the files in arch/x86/CMakeLists.txt, but not all.
To be consistent with how dependency management is done we add wrapper
targets and add dependencies to them.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
The arch/x86/CMakeLists.txt build scripts names five sections that are
generated from .bin files. Two of them are named the same as the .bin
file, and the other three are named inconsistently.
To be consistent, we will rename the three that are named inconistenly
to align with the two that are named as the .bin file.
Being consistent simplifies the system and fosters code-reuse.
This patch renames irq_vectors_alloc_data.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
The arch/x86/CMakeLists.txt build scripts names five sections that are
generated from .bin files. Two of them are named the same as the .bin
file, and the other three are named inconsistently.
To be consistent, we will rename the three that are named inconistenly
to align with the two that are named as the .bin file.
Being consistent simplifies the system and fosters code-reuse.
This patch renames gdt.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
The arch/x86/CMakeLists.txt build scripts names five sections that are
generated from .bin files. Two of them are named the same as the .bin
file, and the other three are named inconsistently.
To be consistent, we will rename the three that are named inconistenly
to align with the two that are named as the .bin file.
Being consistent simplifies the system and fosters code-reuse.
This patch renames user_mmu_tables.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
The arch/x86/CMakeLists.txt build scripts names five sections that are
generated from .bin files. Two of them are named the same as the .bin
file, and the other three are named inconsistently.
To be consistent, we will rename the three that are named inconistenly
to align with the two that are named as the .bin file.
Being consistent simplifies the system and fosters code-reuse.
This patch renames mmu_tables.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
If the faulting context is in user mode, then we are
not on the same stack due to HW-level stack switching
on privilege elevation, and the faulting ESP is on
the stack itself.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
The code did not consider privilege level stack switches.
We have the original stack pointer in the NANO_ESF,
just use that.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
We now have a dedicated function to test whether
a memory region is withing the boundary of the
faulting context's stack buffer.
We use this to determine whether a page or double fault
was due to ESP being outside the bounds of the stack,
as well as when unwinding stack frames to print debug
output.
Fixes two issues:
- Stack overflows in user mode being incorrectly reported
as just page fault exceptions
- Exceptions that occur when unwinding corrupted stacks
The type of fault which triggered the stack overflow
logic (double or page fault) is now always shown.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
The code wasn't checking if the memory address to check
corresponded to a non-present page directory pointer
table entry.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Upon hard/soft irq or exception entry/exit, handle transitions
off or onto the trampoline stack, which is the only stack that
can be used on the kernel side when the shadow page table
is active. We swap page tables when on this stack.
Adjustments to page tables are now as follows:
- Any adjustments for stack memory access now are always done
to the user page tables
- Any adjustments for memory domains are now always done to
the user page tables
- With KPTI, resetting a page now clears the present bit
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
If kernel page table isolation is enabled, we generate a second
set of page tables. These tables, except for the shared page, have
all non-user pages marked as non-present.
The MMU generation script has been refactored:
- Debugging output has been make significantly simpler and less
verbose
- Useless globals removed or adjusted
- MMU region list is validated as it is read
- Some tuples unpacked into individual variables to make the
code easier to read.
- Useless command line option for output binary endian-ness
remobved
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
In the event of a double fault, we do a HW task switch to
a special _df_tss hardware task which resets the stack
pointer to the interrupt stack and otherwise restores
the main hardware task to a runnable state so that
_df_handler_bottom() can run.
However, we need to make sure that _df_handler_bottom()
runs with interrupts locked, otherwise another IRQ could
corrupt the interrupt stack resulting in undefined
behavior.
We have very little stack space to work with in this
context, just zero it. It's a fatal error for the thread
in any event.
Fixes: #7291
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
This commit changes the names of SYS_POWER_DEEP_SLEEP* Kconfig
options in order to match SYS_POWER_LOW_POWER_STATE* naming
scheme.
Signed-off-by: Piotr Zięcik <piotr.ziecik@nordicsemi.no>
At boot, user threads were being granted access to the entire
app shared memory section. This is incorrect; user threads should
have no access until they are added to a memory domain, which
may contain partitions defined within it.
Change from MMU_ENTRY_USER (which grants permission at boot)
to MMU_ENTRY_RUNTIME_USER (which indicates that the pages may
be granted to user mode at runtime, but not at boot).
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
We want a _Swap() variant that can atomically release/restore a
spinlock state in addition to the legacy irqlock. The function as it
was is now named "_Swap_irqlock()", while _Swap() now refers to a
spinlock and takes two arguments. The former will be going away once
existing users (not that many! Swap() is an internal API, and the
long port away from legacy irqlocking is going to be happening mostly
in drivers) are ported to spinlocks.
Obviously on uniprocessor setups, these produce identical code. But
SMP requires that the correct API be used to maintain the global lock.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
This was never a long-term solution, more of a gross hack
to get test cases working until we could figure out a good
end-to-end solution for memory domains that generated
appropriate linker sections. Now that we have this with
the app shared memory feature, and have converted all tests
to remove it, delete this feature.
To date all userspace APIs have been tagged as 'experimental'
which sidesteps deprecation policies.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
This diverges from policy for all of our other arches
and C libraries. Global access to the malloc arena
may not be desirable.
Forthcoming patch will expose, for all C libraries, a
k_mem_partition with the malloc arena which can be
added to domains as desired.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
This is a separate data section which needs to be copied into
RAM.
Most arches just use the kernel's _data_copy(), but x86 has its
own optimized copying code.
Signed-off-by: Adithya Baglody <adithya.nagaraj.baglody@intel.com>
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
PAE tables introduce the NX bit which is very desirable
from a security perspetive, back in 1995.
PAE tables are larger, but we are not targeting x86 memory
protection for RAM constrained devices.
Remove the old style 32-bit tables to make the x86 port
easier to maintain.
Renamed some verbosely named data structures, and fixed
incorrect number of entries for the page directory
pointer table.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
There is an effort underway to make most of the Zephyr build script's
reentrant. Meaning, the build scripts can be executed multiple times
during the same CMake invocation.
Reentrancy enables several use-cases, the motivating one is the
ability to build several Zephyr executables, or images, for instance a
bootloader and an application.
For build scripts to be reentrant they cannot be directly referencing
global variables, like target names, but must instead reference
variables, which can vary from entry to entry.
Therefore, in this patch, we replace global targets with variables.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
This patch adds all the required hooks needed in the kernel to
get the coverage reports from x86 SoCs.
Signed-off-by: Adithya Baglody <adithya.nagaraj.baglody@intel.com>
LTO is not supported yet, but there are a handful of references to the
flag '-flto' and the non-existent Kconfig option 'LTO'. To not confuse
users about whether LTO is supported or not, we should remove this
dead code.
As an aside, prototyping has shown that supporting LTO will give
signicant (10%) code size improvments, but will not be trivial to
support due to how we process object files with python.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
The operation was shifiting bit using a signed constant in the left
operand. Use BIT macro to do it properly.
MISRA-C rule 12.2
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
In C90 was introduced function prototype, that allows argument types
to be checked against parameter types, though it is not necessary
specify names for the parameters. MISRA-C requires names for function
prototype parameters, it claims that names can provide useful
information regarding the function interface.
MISRA-C rule 8.2
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
When __ASSERT is not enabled there is an attribution to the variable
total_partitions and it is never used.
MISRA-C rule 2.2
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
There is a function called _thread_entry defined in
lib/thread_entry.c. Just changing name to fix MISRA-C violation.
MISRA-C rule 5.8
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
Previously, this was only built if CONFIG_EXCEPTION_DEBUG
was enabled, but CONFIG_USERSPACE needs it too for validating
strings sent in from user mode.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
This commit exposes k_mem_partition_attr_t outside User Mode, so
we can use struct k_mem_partition for defining memory partitions
outside the scope of user space (for example, to describe thread
stack guards or no-cacheable MPU regions). A requirement is that
the Zephyr build supports Memory protection. To signify this, a
new hidden, all-architecture Kconfig symbol is defined (MPU). In
the wake of exposing k_mem_partition_attr_t, the commit exposes
the MPU architecture-specific access permission attribute macros
outside the User space context (for all ARCHs), so they can be
used in a more generic way.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
