We re-wrote the xtensa arch code, but never got around
to purging the old implementation.
Removed those boards which hadn't been moved to the new
arch code. These were all xt-sim simulator targets and not
real hardware.
Fixes: #18138
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
From the Jailhouse days, this has been a function call. That's silly.
We now inline the EOI in the ISR when in x2APIC mode. Also clean up
z_irq_controller_eoi(), so it now uses the inline macros.
Also, we now enable x2APIC on up_squared by default.
Fixes: #17133
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
Use enumerate() to fix this pylint warning:
C0200: Consider using enumerate instead of iterating with range and
len (consider-using-enumerate)
enumerate() is handy when the loop body needs both the element and its
index. It returns (index, element) tuples.
Also use a tuple unpacking to extract 'handler' from the elements in
'vector'.
Piggyback a slightly simpler way to build a list of num_chars 0s.
Getting rid of warnings for a CI check.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
Accidentally passed two arguments instead of one. Fixes this pylint
error:
arch/x86/gen_idt.py:132:8: E1121: Too many positional arguments for
function call (too-many-function-args)
Fixing pylint warning for a CI check.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
Fix this warning, as a preparation for a CI check:
arch/common/gen_isr_tables.py:167:11: C0123: Using type() instead of
isinstance() for a typecheck. (unidiomatic-typecheck)
isinstance() has the advantage that it also handles inheritance, though
it doesn't really matter here. It's more common at least.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
Getting slightly subjective, but fixes this pylint warning:
arch/x86/gen_idt.py:281:11: R1714: Consider merging these
comparisons with "in" to 'handler not in (spur_code, spur_nocode)'
(consider-using-in)
Getting rid of pylint warnings for a CI check. I could disable any
controversial ones (it's already a list of warnings to enable anyway).
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
Promote a handy and often-overlooked sys.exit() feature: Passing it a
string (or any other non-int object) prints it to stderr and exits with
status 1.
See the documentation at
https://docs.python.org/3/library/sys.html#sys.exit.
This indirectly prints some errors to stderr that previously went to
stdout.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
This adds a simple infinite loop when double exception is raised.
Without this, if double exception occurs, it would execute
arbitrary code.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This follows the z_arch_irq_en-/dis-able() so that the SoC
definitions are responsible for functions related to multi-level
interrupts.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Currently, the interrupt service code manually raises the CPU task
priority to the priority level of the vector being serviced to defer
any lower-priority interrupts. This is unnecessary; the local APIC
is aware that an interrupt is in-service and accounts for its priority
when deciding whether to issue an overriding interrupt to the CPU.
Signed-off-by: Charles E. Youse <charles@gnuless.org>
Use the 'not in' operator. Fixes this pylint warning:
arch/xtensa/core/xtensa_intgen.py:77:7: C0113: Consider changing
"not lvl in ints_by_lvl" to "lvl not in ints_by_lvl" (unneeded-not)
Fixing pylint warnings for a CI check.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
Reported by pylint's 'bad-whitespace' warning.
Not gonna enable this warning in the CI check, because it flags stuff
like deliberately aligning assignments and gets too cultish. Just a
cleanup pass.
For whatever reason, the common convention in Python is to skip spaces
around '=' when passing keyword arguments and giving default arguments:
f(x=3, y=4)
def f(x, y=8):
...
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
The code in question is very non-trivial so without good explanation
it takes a lot of time to realize what's done there and why
it still works in the end.
Here I'm trying to save a couple of man-days for the next developers
who's going to touch that piece of code.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
This commit makes it possible to infer Z_ARCH_EXCEPT()
calls in SVCs that escalate to HardFault due to being
invoked from priority level equal or higher to the
interrupt priority level of the SVC Handler.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
commit 780324b8ed ("cleanup: rename fiber/task -> thread")
seems to be done by a script and in that particular case turned
menaingful sentence into nonsense. Alas, threads might be in all
four states.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Anas Nashif <anas.nashif@intel.com>
We manage IRQs in a quite a different way now since
commit f8d061faf7 ("arch: arc: add nested interrupt support")
so that comment not only makes no sense but also may fool a reader
as disabling of interrupts happens in the very beginning of
_rirq_exit() but not here.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
we should not rely on that eret has a copy of ilink in fast
irq handling. This will cause crash for hs cores.
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
For the old codes, if nest interrupts come out after _isr_wrapper
and before _check_nest_int_by_irq_act, then multi-bits in irq_act
will be set, this will result irq stack will not be switched in
correctly
As a fix, it's still need to use nest interrupt counter to do
interrupt stack switch as before
The difference is in the past exc_nest_count is used, but here
_kernel.nested/_kernel.cpus[cpu_id].nested is used.
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
* in arc secureshield interrupts can be configured
as secure or normal
* in sw design, high interrupt priorites are allocated to
secure world, low priorities are allocated to normal world.
* secure interrupt > secure thread > normal interrupt > normal
thead
So, here secure world/firmware only checks secure interrupt
priorities
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
it's not allowed to switch to thread preempted by exception as
its context is not saved.
So if a thread switch is required in exception handling, e.g.
kill a thread, the old thread cannot be switched back
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
For arc processor equiped with secureshield, SEC_STAT.IRM
bit should be recorded, it determins which mode irq should
return
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
Set the recommended thread stack size to 40 bytes in case a build is
made for a 64-bit native posix board
Signed-off-by: Jan Van Winkel <jan.van_winkel@dxplore.eu>
Update the xtensa backend to work better with the new fatal error
architecture. Move the stack frame dump (xtensa uses a variable-size
frame becuase we don't spill unused register windows, so it doesn't
strictly have an ESF struct) into z_xtensa_fatal_error(). Unify the
older exception logging with the newer one (they'd been sort of glomed
together in the recent rework), mostly using the asm2 code but with
the exception cause stringification and the PS register field
extraction from the older one.
Note that one shortcoming is that the way the dispatch code works, we
don't have access to the spilled frame from within the spurious error
handler, so this can't log the interrupted CPU state. This isn't
fixable easily without adding overhead to every interrupt entry, so it
needs to stay the way it is for now. Longer term we could exract the
caller frame from the window state and figure it out with some
elaborate assembly, I guess.
Fixes#18140
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
It was discovered that the xtensa version of
z_arch_irq_connect_dynamic() was being removed along with the old
xtensa architecture support, because it was never included in the asm2
builds.
But there's no xtensa-specific code in it at all. Architectures that
use the existing sw_isr_table mechanism and don't (or can't, in the
case of xtensa which has fixed interrupt priority) interpret the other
parameters might as well have access to a working generic
implementation.
Fixes#18272
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Consistently place C++ use of extern "C" after all include directives,
within the negative branch of _ASMLANGUAGE if used.
Remove extern "C" support from files that don't declare objects or
functions.
In arch/arc/arch.h the extern "C" in the including context is left
active during an include to avoid more complex restructuring.
Background from issue #17997:
Declarations that use C linkage should be placed within extern "C"
so the language linkage is correct when the header is included by
a C++ compiler.
Similarly #include directives should be outside the extern "C" to
ensure the language-specific default linkage is applied to any
declarations provided by the included header.
See: https://en.cppreference.com/w/cpp/language/language_linkage
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Consistently place C++ use of extern "C" after all include directives,
within the negative branch of _ASMLANGUAGE if used.
Remove extern "C" support from files that don't declare objects or
functions.
Background from issue #17997:
Declarations that use C linkage should be placed within extern "C"
so the language linkage is correct when the header is included by
a C++ compiler.
Similarly #include directives should be outside the extern "C" to
ensure the language-specific default linkage is applied to any
declarations provided by the included header.
See: https://en.cppreference.com/w/cpp/language/language_linkage
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Consistently place C++ use of extern "C" after all include directives,
within the negative branch of _ASMLANGUAGE if used.
Background from issue #17997:
Declarations that use C linkage should be placed within extern "C"
so the language linkage is correct when the header is included by
a C++ compiler.
Similarly #include directives should be outside the extern "C" to
ensure the language-specific default linkage is applied to any
declarations provided by the included header.
See: https://en.cppreference.com/w/cpp/language/language_linkage
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Consistently place C++ use of extern "C" after all include directives,
within the negative branch of _ASMLANGUAGE if used.
In arch.h the extern "C" in the including context is left active during
include of target-specific mpu headers to avoid more complex
restructuring.
Background from issue #17997:
Declarations that use C linkage should be placed within extern "C"
so the language linkage is correct when the header is included by
a C++ compiler.
Similarly #include directives should be outside the extern "C" to
ensure the language-specific default linkage is applied to any
declarations provided by the included header.
See: https://en.cppreference.com/w/cpp/language/language_linkage
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Consistently place C++ use of extern "C" after all include directives,
within the negative branch of _ASMLANGUAGE if used.
Background from issue #17997:
Declarations that use C linkage should be placed within extern "C"
so the language linkage is correct when the header is included by
a C++ compiler.
Similarly #include directives should be outside the extern "C" to
ensure the language-specific default linkage is applied to any
declarations provided by the included header.
See: https://en.cppreference.com/w/cpp/language/language_linkage
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
'recoverable' is a value passed by reference and we
should be dereferencing the pointer, to check if the
fault has been classified as recoverable.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Consistently place C++ use of extern "C" after all include directives,
within the negative branch of _ASMLANGUAGE if used.
Remove extern "C" support from files that don't declare objects or
functions.
Background from issue #17997:
Declarations that use C linkage should be placed within extern "C"
so the language linkage is correct when the header is included by
a C++ compiler.
Similarly #include directives should be outside the extern "C" to
ensure the language-specific default linkage is applied to any
declarations provided by the included header.
See: https://en.cppreference.com/w/cpp/language/language_linkage
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
SR and LR were used as global names for load and store RISC-V assembler
operations, colliding with other uses such as SR for STATUS REGISTER in
some peripherals. Renamed them to a longer more specific name to avoid
the collision.
Signed-off-by: Karsten Koenig <karsten.koenig.030@gmail.com>
When coming out of an exception, we need to mask interrupts
to avoid races when decrementing the nested count. Move
the instruction that does this earlier.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Related to #17997, for the POSIX arch:
* Remove some unnecessary extern "C" and ifdef blocks
* Move an include out of one of these blocks
* Add a missing extern "C" block
Background:
Declarations that use C linkage should be placed within extern "C"
so the language linkage is correct when the header is included by
a C++ compiler.
Similarly #include directives should be outside the extern "C" to
ensure the language-specific default linkage is applied to any
declarations provided by the included header.
See: https://en.cppreference.com/w/cpp/language/language_linkage
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
* it's based on ARC SecureShield
* add basic secure service in arch/arc/core/secureshield
* necesssary changes in arch level
* thread switch
* irq/exception handling
* initialization
* add secure time support
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
according to high-level design,in user mode software-triggered system
fatal exceptions only allow oops and stack check failure
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
exception, different with irq offload, may be raised interrupt
handling, e.g.
* z_check_stack_sentinel
* wrong code
we need to add specific handling of this case in exception handling
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
after appling the new "_get_curr_cpu_irq_stack" in _exc_entry,
the caculation of exception stack is wrong, this will
cause stack overflow, make the exception handling corrupt.
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
This commit enables the option to route the BusFault,
HardFault, and NMI exceptions in Secure state, when
building for Cortex-M CPUs with ARM_SECURE_FIRMWARE=y.
This allows the various test to utilize BusFault,
HardFault and NMI exceptions during testing.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
For now we enforce the medany code model for 64-bit builds as we get
reloc issues otherwise. The instruction set and ABI are also set to
soft-float usage.
The ilp32 ABI is explicitly specified on 32-bit build to make sure
it is not using a wrong default if the same toolchain is used for both
32- and 64-bit builds. The archittecture options are the same as the
SDK's riscv32 toolchain default in that case.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Since commit c535300539 ("drivers/timer: New ARM SysTick driver"),
_NanoIdleValGet and _NanoIdleValClear have been unused.
Signed-off-by: Bradley Bolen <bbolen@lexmark.com>
Assembly language start code will enter here, which sets up
early kernel initialization and then calls z_cstart() when
finished.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Removes very complex boot-time generation of page tables
with a much simpler runtime generation of them at bootup.
For those x86 boards that enable the MMU in the defconfig,
set the number of page pool pages appropriately.
The MMU_RUNTIME_* flags have been removed. They were an
artifact of the old page table generation and did not
correspond to any hardware state.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
* modify the reset flow for SMP
* add smp related initialization
* implement ipi related functions
* implement thread switch in isr/exception
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
* arc connect is a component to connect multiple arc cores
* it's necessary for arc smp support
* the following features are implemented
* inter-core interrupt unit
* gloabl free running counter
* inter-core debug unit
* interrupt distribute unit
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
Makes the code that defines stacks, and code referencing
areas within the stack object, much clearer.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Previously, context switching on x86 with memory protection
enabled involved walking the page tables, de-configuring all
the partitions in the outgoing thread's memory domain, and
then configuring all the partitions in the incoming thread's
domain, on a global set of page tables.
We now have a much faster design. Each thread has reserved in
its stack object a number of pages to store page directories
and page tables pertaining to the system RAM area. Each
thread also has a toplevel PDPT which is configured to use
the per-thread tables for system RAM, and the global tables
for the rest of the address space.
The result of this is on context switch, at most we just have
to update the CR3 register to the incoming thread's PDPT.
The x86_mmu_api test was making too many assumptions and has
been adjusted to work with the new design.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
The current API was assuming too much, in that it expected that
arch-specific memory domain configuration is only maintained
in some global area, and updates to domains that are not currently
active have no effect.
This was true when all memory domain state was tracked in page
tables or MPU registers, but no longer works when arch-specific
memory management information is stored in thread-specific areas.
This is needed for: #13441#13074#15135
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
These turned out to be quite useful when debugging MMU
issues, commit them to the tree. The output format is
virtually the same as gen_mmu_x86.py's verbose output.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Currently page tables have to be re-computed in
an expensive operation on context switch. Here we
reserve some room in the page tables such that
we can have per-thread page table data, which will
be much simpler to update on context switch at
the expense of memory.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Has the same effect of catching stack overflows, but
makes debugging with GDB simpler since we won't get
errors when inspecting such regions. Making these
areas non-present was more than we needed, read-only
is sufficient.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
There are two aspects to this: CPU registers are twice as big, and the
load and store instructions must use the 'd' suffix instead of the 'w'
one. To abstract register differences, we simply use a ulong_t instead
of u32_t given that RISC-V is either ILP32 or LP64. And the relevant
lw/sw instructions are replaced by LR/SR (load/store register) that get
defined as either lw/sw or ld/sd. Finally a few constants to deal with
register offsets are also provided.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
With the upcoming riscv64 support, it is best to use "riscv" as the
subdirectory name and common symbols as riscv32 and riscv64 support
code is almost identical. Then later decide whether 32-bit or 64-bit
compilation is wanted.
Redirects for the web documentation are also included.
Then zephyrbot complained about this:
"
New files added that are not covered in CODEOWNERS:
dts/riscv/microsemi-miv.dtsi
dts/riscv/riscv32-fe310.dtsi
Please add one or more entries in the CODEOWNERS file to cover
those files
"
So I assigned them to those who created them. Feel free to readjust
as necessary.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
The ARM specific _impl_k_thread_abort function only applies to Cortex-M
so move it to the cortex_m specific directory.
Signed-off-by: Bradley Bolen <bbolen@lexmark.com>
Adapted from similar code in the x86_64 port.
Useful when debugging boot problems on actual x86
hardware if a JTAG isn't handy or feasible.
Turn this on for qemu_x86.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
* when fpu is configured or mpy_option > 6,
accl regs (r58, r59) will be configured,
they are used by fpu and mac, and are caller
-saved scratch regs, so need to be saved before
jumping to interrupt handlers
* r25 and r30 are also caller-saved scratch reg.
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
for arc, floating point support cannot be enabled
automatically, so k_float_enable is requred.
z_arch_float_enable is for k_float_enable
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
The ARC HS is a family of high performance CPUs from Synopsys
capable of running wide range of applications from heavy DPS
calculation to full-scale OS.
Still as with other ARC cores ARC HS might be tailored to
a particular application.
As opposed to EM cores ARC HS cores always have support of unaligned
data access and by default GCC generates such a data layout with
so we have to always enable unaligned data access in runtime otherwise
on attempt to access such data we'd see "Unaligned memory exception".
Note we had to explicitly mention CONFIG_CPU_ARCEM=y in
all current defconfigs as CPU_ARC{EM|HS} are now parts of a
choice so we cannot simply select ether option in board's Kconfig.
And while at it change "-mmpy-option" of ARC EM to "wlh1"
which is the same as previously used "6" but matches
Programmer's Reference Manual (PRM) and is more human-friendly.
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
ARCv2 cores may access data not aligned by the data size boundary.
I.e. read entire 32-bit word from address 0x1.
This feature is configurable for ARC EM cores excluding those with
secure shield 2+2 mode. When it's available in hardware it's required
to enable that feature in run-time as well setting status32.AD bit.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
KFLAG instruction might affect multiple flags in STATUS32 register
and so when we need just AE-bit to be reset we need first read current
state of STATUS32, then change our bit and set STATUS32 again.
Otherwise critical flags including stack checking, unaligned access etc
will be dropped for good.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Up until now only ARC EM family has been supported in Zephyr
which don't support atomic operations other than
compare-and-excange, so custom atomic ops with load-locked(LLOCK)/
store-conditional(SCOND) were never used that's how we never
realised CONFIG_ATOMIC_OPERATIONS_CUSTOM points to the wrong file:
"atomic.c" while real implementation is in "atomic.S".
Fix that now.
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
It looks like, at some point in the past, initializing thread stacks
was the responsibility of the arch layer. After that was centralized,
we forgot to remove the related conditional header inclusion. Fixed.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
* use IRQ_ACT to check nest interrupt
* implement an asm macro for nest interrupt check
* no need to use exc_nest_count, remove it
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
* do not use a specific variable (saved_r0/saved_sp) to free r0
/exchange sp, but use stack to do that.
* it will make code scalable, e.g. for SMP, no need to define
variables for each core
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
* as ilink has a copy in ERET, it can be reused as a gp
* use ilink to do the job of arc_exc_saved_sp to save 4 bytes
and save some cycles because no load/store of memory
* it will make code scalable, e.g. for SMP, no need to
define variables for each core
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
User mode is only allowed to induce oopses and stack check
failures via software-triggered system fatal exceptions. This
commit forces a kernel oops if any other fatal exception reason
is enforced.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This is now called z_arch_esf_t, conforming to our naming
convention.
This needs to remain a typedef due to how our offset generation
header mechanism works.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
We had a function that did this, but it was dead code.
Move to fatal.c and call from z_arm_fatal_error().
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
We introduce a new z_fatal_print() API and replace all
occurrences of exception handling code to use it.
This routes messages to the logging subsystem if enabled.
Otherwise, messages are sent to printk().
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
* z_NanoFatalErrorHandler() is now moved to common kernel code
and renamed z_fatal_error(). Arches dump arch-specific info
before calling.
* z_SysFatalErrorHandler() is now moved to common kernel code
and renamed k_sys_fatal_error_handler(). It is now much simpler;
the default policy is simply to lock interrupts and halt the system.
If an implementation of this function returns, then the currently
running thread is aborted.
* New arch-specific APIs introduced:
- z_arch_system_halt() simply powers off or halts the system.
* We now have a standard set of fatal exception reason codes,
namespaced under K_ERR_*
* CONFIG_SIMPLE_FATAL_ERROR_HANDLER deleted
* LOG_PANIC() calls moved to k_sys_fatal_error_handler()
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
We are standardizing to a arch-independent set of exception
reason codes, don't overload it with internal state of
the ARM fault handling code.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
sw_isr_table has two entries, an argument and an ISR function. The
comment on struct _isr_table_entry in include/sw_isr_table.h says that
"This allows a table entry to be loaded [...] with one ldmia
instruction, on ARM [...]". Some arch, e.g. SPARC, also has a double
word load instruction, "ldd", but the instruct must have address align
to double word or 8 bytes.
This commit makes the table alignment configurable. It allows each
architecture to specify it, if needed. The default value is 0 for no
alignment.
Signed-off-by: Yasushi SHOJI <y-shoji@ispace-inc.com>
The SoC initialization code used system clock frequency
as a CPU clock frequency. This commit corrects that by
obtaining the needed value from DTS.
Signed-off-by: Piotr Zięcik <piotr.ziecik@nordicsemi.no>
Fix a race which seems to have been presenting itself
very sporadically on loaded systems.
The race seems to have caused tests/kernel/sched/schedule_api
to fail at random on native_posix.
The case is a bit convoluted:
When the kernel calls z_new_thread(), the POSIX arch saves
the new thread entry call in that new Zephyr thread stack
together with a bit of extra info for the POSIX arch.
And spawns a new pthread (posix_thread_starter()) which
will eventually (after the Zephyr kernel swapped to it),
call that entry function.
(Note that in principle a thread spawned by pthreads may
be arbitrarily delayed)
The POSIX arch does not try to synchronize to that new
pthread (because why should it) until the first time the
Zephyr kernel tries to swap to that thread.
But, the kernel may never try to swap to it.
And therefore that thread's posix_thread_starter() may never
have got to run before the thread was aborted, and its
Zephyr stack reused for something else by the Zephyr app.
As posix_thread_starter() was relaying on looking into that
thread stack, it may now be looking into another thread stack
or anything else.
So, this commit fixes it by having posix_thread_starter()
get the input it always needs not from the Zephyr stack,
but from its own pthread_create() parameter pointing to a
structure kept by the POSIX arch.
Note that if the thread was aborted before reaching that point
posix_thread_starter() will NOT call the Zephyr thread entry
function, but just cleanup.
With this change all "asynchronous" parts of the POSIX arch
should relay only on the POSIX arch own structures.
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
Qemu is already updated past 2.9 release, so this
workaround for QEMU_CORTEX_M3 is now obsolete and
can be removed.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Before, attempting to induce a kernel oops would instead
lead to a general protection fault as the interrupt vector
was at DPL=0.
Now we allow by setting DPL=3. We restrict the allowable
reason codes to either stack overflows or kernel oops; we
don't want user mode to be able to create a kernel panic,
or fake some other kind of exception.
Fixes an issue where the stack canary test case was triggering
a GPF instead of a stack check exception on x86.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
We need to pass -m64 instead of -m32 when CONFIG_64BIT is set.
This is pretty x86 centric. Many platforms don't have the ability
to select between 32-bits or 64-bits builds and either of those should
be dropped in that case with restriction on the available configuration
done elsewhere. But for the time being this allows for testing both.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Correct the storage type of the thread status pointer
not assuming 32bit pointer and integer size
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
For the x86 architecture the z_arch_float_disable() is only
implemented when building with CONFIG_LAZY_FP_SHARING, so we
make z_arch_float_disable() return -ENOSYS when we build with
FLOAT and FP_SHARING but on an x86 platform where
LAZY_FP_SHARING is not supported.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
The current version is 32-bit specific, so move it to ia32/
and add a layer of indirection via an arch-level header file.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
Refactoring 32- and 64-bit subarchitectures, so this file is moved
to ia32/ and a new "redirector" header file is introduced.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
This data is subarchitecture-specific, so move it to ia32/
and add a layer of indirection at the architecture level.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
Some of this is 32-bit specific, some applies to all subarchitectures.
A preliminary attempt is made to refactor and place 32-bit-specific
portions in ia32/kernel_arch_data.h.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
This file merely declares external functions referenced only
by ia32/cache.c, so the declarations are inlined instead.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
This file was used to generate offsets for host tools that are no
longer in use, so it's removed and the offsets are no longer generated.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
Over time, this has been reduced to a few functions dealing solely
with floating-point support, referenced only from core/ia32/float.c.
Thus they are moved into that file and the header is eliminated.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
* if thread switchs in interrupt, the target sp must be in
thread's kernel stack, no need to do hardware sp switch
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
The MVIC is no longer supported, and only the APIC-based interrupt
subsystem remains. Thus this layer of indirection is unnecessary.
This also corrects an oversight left over from the Jailhouse x2APIC
implementation affecting EOI delivery for direct ISRs only.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
This header is currently IA32-specific, so move it into the subarch
directory and update references to it.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
Making room for the Intel64 subarch in this tree. This header is
32-bit specific and so it's relocated, and references rewritten
to find it in its new location.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
This file is 32-bit specific, so it is moved into the ia32/ directory
and references to it are updated accordingly.
Also, SP_ARG* definitions are no longer used, so they are removed.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
Eliminate definitions for MSRs that we don't use. Centralize the
definitions for the MSRs that we do use, including their fields.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
This pattern exists in both the include/arch/x86 and arch/x86/include
trees. This indirection is historic and unnecessary, as all supported
toolchains for x86 support gas/gcc-style inline assembly.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
z_arm_do_syscall is executing in privileged mode. This implies
that we shall not be allowed to use the thread's default
unprivileged stack, (i.e push to or pop from it), to avoid any
possible stack corruptions.
Note that since we execute in PRIV mode and no MPU guard or
PSPLIM register is guarding the end of the default stack, we
won't be able to detect any stack overflows.
This commit implement the above change, by forcing
z_arm_do_syscall() to FIRST switch to privileged
stack and then do all the preparations to execute
the system call.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
We need to correct the inline comment in swap_helper.S,
which is suggesting that system call attempts with
invalid syscall IDs (i.e. above the limit) do not force
the CPU to elevate privileges. This is in fact not true,
since the execution flow moves into valid syscall ID
handling.
In other words, all we do for system calls with invalid
ID numbers is to treat them as valid syscalls with the
K_SYSCALL_BAD ID value.
We fix the inline documentation to reflect the actual
execution flow.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
System call arguments are indexed from 1 to 6, so arg0
is corrected to arg1 in two occasions. In addition, the
ARM function for system calls is now called z_arm_do_syscall,
so we update the inline comment in __svc handler.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
If the offset within the thread struct to the
ARC arch-specific 'relinquish_cause' member is too
large, ld_s instructions referencing it will not
compile. This happens easily if CONFIG_THREAD_NAME
reserves a name buffer within the thread struct, since
all the arch-specific members come last.
Use the regular 'ld' instruction instead.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
ARMv8-M architecture supports the built-in stack overflow
detection mechanisms via the SPLIM registers. However, the
user might still wish to use the traditional MPU-based stack
overflow detection mechanism (for testing or other reasons).
We now allow the user to enable HW stack protection, but
manually turn off BUILTIN_STACK_GUARD option. This will force
the MPU_STACK_GUARD option to be selected.
It is still not allowed for the user to not select any stack
guard mechanisms, if HW_STACK_PROTECTION is selected.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Work around a testcase problem, where we want to check some
logic for the bounds check bypass mitigation in the common
kernel code. By changing the ifdef to the x86-specific option
for these lfence instructions, we avoid IAMCU build errors
but still test the common code.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Fix the start of the guard to take into account the
configurable size of the guard.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
* here use new style z_arch_switch,i.e. CONFIG_USE_SWITCH
to replace old swap mechnism.
* it's also required by SMP support
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
move misc/reboot.h to power/reboot.h and
create a shim for backward-compatibility.
No functional changes to the headers.
A warning in the shim can be controlled with CONFIG_COMPAT_INCLUDES.
Related to #16539
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
move misc/util.h to sys/util.h and
create a shim for backward-compatibility.
No functional changes to the headers.
A warning in the shim can be controlled with CONFIG_COMPAT_INCLUDES.
Related to #16539
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
move misc/printk.h to sys/printk.h and
create a shim for backward-compatibility.
No functional changes to the headers.
A warning in the shim can be controlled with CONFIG_COMPAT_INCLUDES.
Related to #16539
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
move misc/dlist.h to sys/dlist.h and
create a shim for backward-compatibility.
No functional changes to the headers.
A warning in the shim can be controlled with CONFIG_COMPAT_INCLUDES.
Related to #16539
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
move misc/__assert.h to sys/__assert.h and
create a shim for backward-compatibility.
No functional changes to the headers.
A warning in the shim can be controlled with CONFIG_COMPAT_INCLUDES.
Related to #16539
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
move atomic.h to sys/atomic.h and
create a shim for backward-compatibility.
No functional changes to the headers.
A warning in the shim can be controlled with CONFIG_COMPAT_INCLUDES.
Related to #16539
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
move tracing.h to debug/tracing.h and
create a shim for backward-compatibility.
No functional changes to the headers.
A warning in the shim can be controlled with CONFIG_COMPAT_INCLUDES.
Related to #16539
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Under FP shared registers mode (CONFIG_FP_SHARING=y),
a thread's user_options flag is checked during swap and
during stack fail check. Therefore, in k_float_disable()
we want to ensure that a thread won't be swapped-out with
K_FP_REGS flag cleared but still FP-active (CONTROL.FPCA
being not zero). To ensure that we temporarily disable
interrupts.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit reworks the ARM stack fail checking, under FP
Sharing registers mode, to account for the right width of
the MPU stack guard.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
For threads that appear to be FP-capable (i.e. with K_FP_REGS
option flag set), we configure a wide MPU stack guard, if we
build with stack protection enabled (CONFIG_MPU_STACK_GUARD=y).
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
When an FP capable thread (i.e. with K_FP_REGS option)
transitions into user mode, we want to allocate a wider
MPU stack guard region, to be able to successfully detect
overflows of the privilege stack during system calls. For
that we also need to re-adjust the .priv_stack_start pointer,
which denotes the start of the writable area of the privilege
stack buffer.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
When an FP capable thread is created (i.e. with K_FP_REGS
option) we want to allocate a wider MPU stack guard region,
to be able to successfully detect stack overflows. For that
we also need to re-adjust the values that will be passed to
the thread's stack_info .start and .size parameters.
applicable) for a thread which intends to use the FP services.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This adds the necessary bits to utilize the x86_64 toolchain
built by sdk-ng for x86_64 when toolchain variant is either
zephyr or xtools. This allows decoupling the builds from
the host toolchain.
Newlib is also available with this toolchain so remove
the Kconfig restriction on CONFIG_NEWLIB_LIBC.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
The libc hooks for Newlib requires CONFIG_SRAM_SIZE and
the symbol "_end" at the end of memory. This is in preparation
for enabling Newlib for x86_64.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Compilers (at least gcc and clang) already provide definitions to
create standard types and their range. For example, __INT16_TYPE__ is
normally defined as a short to be used with the int16_t typedef, and
__INT16_MAX__ is defined as 32767. So it makes sense to rely on them
rather than hardcoding our own, especially for the fast types where
the compiler itself knows what basic type is best.
Using compiler provided definitions makes even more sense when dealing
with 64-bit targets where some types such as intptr_t and size_t must
have a different size and range. Those definitions are then adjusted
by the compiler directly.
However there are two cases for which we should override those
definitions:
* The __INT32_TYPE__ definition on 32-bit targets vary between an int
and a long int depending on the architecture and configuration.
Notably, all compilers shipped with the Zephyr SDK, except for the
i586-zephyr-elfiamcu variant, define __INT32_TYPE__ to a long int.
Whereas, all Linux configurations for gcc, both 32-bit and 64-bit,
always define __INT32_TYPE__ as an int. Having variability here is
not welcome as pointers to a long int and to an int are not deemed
compatible by the compiler, and printing an int32_t defined with a
long using %d makes the compiler to complain, even if they're the
same size on 32-bit targets. Given that an int is always 32 bits
on all targets we might care about, and given that Zephyr hardcoded
int32_t to an int before, then we just redefine __INT32_TYPE__ and
derrivatives to an int to keep the peace in the code.
* The confusion also exists with __INTPTR_TYPE__. Looking again at the
Zephyr SDK, it is defined as an int, even even when __INT32_TYPE__ is
initially a long int. One notable exception is i586-zephyr-elf where
__INTPTR_TYPE__ is a long int even when using -m32. On 64-bit targets
this is always a long int. So let's redefine __INTPTR_TYPE__ to always
be a long int on Zephyr which simplifies the code, works for both
32-bit and 64-bit targets, and mimics what the Linux kernel does.
Only a few print format strings needed adjustment.
In those two cases, there is a safeguard to ensure the type we're
enforcing has the right size and fail the build otherwise.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
This was only enabled by the MVIC, which in turn was only used
by the Quark D2000, which has been removed.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
The Quark D2000 is the only x86 with an MVIC, and since support for
it has been dropped, the interrupt controller is orphaned. Removed.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
This commit aligns the programming of the privileged stack MPU
guard with that of the default stack guard (i.e of supervisor
threads). In particular:
- the guard is programmed BELOW the address indicated in
arch.priv_stack_start; it is, therefore, similar to the
default guard that is programmed BELOW stack_info.start.
An ASSERT is added to confirm that the guard is programmed
inside the thread privilege stack area.
- the stack fail check is updated accordningly
- arch.priv_stack_start is adjusted in arch_userspace_enter(),
to make sure we account for a (possible) guard requirement,
that is, if building with CONFIG_MPU_STACK_GUARD=y.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit re-organizes the macro definitions in arch.h for
the ARM architecture. In particular, the commit:
- defines the minimum alignment requirement for thread stacks,
that is, excluding alignment requirement for (possible)
MPU stack guards.
- defines convenience macros for the MPU stack guard align and
size for threads using the FP services under Shared registers
mode (CONFIG_FP_SHARING=y). For that, a hidden Kconfig option
is defined in arch/arm/core/cortex_m/mpu/Kconfig.
- enforces stack alignment with a wide MPU stack guard (128
bytes) under CONFIG_FP_SHARING=y for the ARMv7-M architecture,
which requires start address alignment with power-of-two and
region size.
The commit does not change the amount of stack that is reserved
with K_THREAD_STACK_DEFINE; it only determines the stack buffer
alignment as explained above.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
These constants do not need global exposure, as they're only
referenced in the reboot API implementation. Also their names
are trimmed to fit into the X86-arch-specific namespace.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
This appears to date all the way back to the initial import
and is used in exactly one place if DEBUG is on. Removed.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
Previously the existing EFLAGS was used as a base which was
then manipulated accordingly. This is unnecessary as the bits
preserved contain no useful state related to the new thread.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
Found a few annoying typos and figured I better run script and
fix anything it can find, here are the results...
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
The ARCH_CHOICE symbol on the arch/Kconfig choice for Architecture
selection, allows for multiple definitions of the choice group which
makes it possible for out-of-tree architectures to add entries to the
list as needed.
For example, in $(ARCH_DIR)/$(ARCH)/Kconfig by adding something like;
choice ARCH_CHOICE
config xARCH
bool "xARCH architecture"
endchoice
No functional change expected.
Signed-off-by: Danny Oerndrup <daor@demant.com>
No-cache SRAM section is currently used for ARM-only builds
with support for no-cacheable memory sections (i.e.
CONFIG_ARCH_HAS_NOCACHE_MEMORY_SUPPORT) and it holds
uninitialized data. This commit properly defines the
corresponding linker section using SECTION_DATA_PROLOGUE
and GROUP_DATA_LINK_IN macros.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Create source directory for IA32-subarch specific files, and move
qualifying files to that subdirectory.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
The only we support cores that don't have CMOV insns are the MINUTEIAs,
so we simply check for that rather this using a layer of indirection.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
This option is set iff CONFIG_X86 is set, thus it provides no useful
information. Remove the option and replace references with CONFIG_X86.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
This commit adds the architecture-specific implementation
of k_float_disable() for ARM and x86.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Support for Floating Point (both un-sharing and sharing
registers mode) is currently implemented for ARM and X86
architectures, so reflect this in the Kconfig symbol
definition.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
When building without support for user mode (CONFIG_USERSPACE=n)
we need to correct the starting address of the MPU Guard, before
passing it to the function that evaluates whether a stack
corruption has occurred. The bug was introduced by commit
(60bae5de38) in
PR-13619, where the start address of the MPU guard was properly
corrected, but the guard start at the corresponding stack-fail
check was not adjusted accordingly.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
More clearly differentiate MVIC vs. APIC timer code, and use new APIC
accessors in include/drivers/loapic.h. Remove extraneous comments, and
other light cleanup work.
This driver is in need of a serious overhaul -- despite appearing to
have support for TICKLESS_KERNEL and DEVICE_POWER_MANAGEMENT, bitrot
has taken its toll and the driver will not build with these enabled.
These should be removed or made to work... but not in this patch.
Old x2APIC-related accessors in kernel_arch_func.h are eliminated.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
Simple renaming and Kconfig reorganization. Choice of local APIC
access method isn't specific to the Jailhouse hypervisor.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
The real-mode startup code is trivially changed to refer to MSR
definitions in include/arch/x86/msr.h, rather than its ad-hoc ones.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
Light reorganization. All MSR definitions and manipulation functions
are consolidated into one header. The names are changed to use an
X86_* prefix instead of IA32_* which is misleading/incorrect.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
drivers/interrupt_controller/i8259.c is not a driver; it exists
solely to disable the i8259s when the configuration calls for it.
The six-byte sequence to mask the controllers is moved to crt0.S
and the pseudo-driver is removed.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
This is the generic symbol to select or otherwise test for when 64-bit
compilation is desired. Two trivial usages of this symbol are also
included.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Change removes tracing hooks before threads are initialized
and thread switched out hook for ARM before first time switching
to main thread.
Signed-off-by: Marek Pieta <Marek.Pieta@nordicsemi.no>
ARC EM4 is just a baseline configuration of ARC EM family of CPU cores.
But with addition of more featuers like caches, DSP extensions etc
we're effectively getting EM6, EM5D etc templates.
So to not confuse users let's talk about families of ARC cores
as that's what makes sense together with extra features but not
templates itself.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
A basic display driver is added for a generic 32-bpp framebuffer.
Glue logic is added to the x86 arch to request the intitialization
of a linear framebuffer by the Multiboot loader (GRUB) and connect
it to this generic driver.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
When booting using GRUB, some useful information about the environment
is given to us via a boot information structure. We've not made any
use of this information so far, but the x86 framebuffer driver will.
A skeletal definition of the structure is given, and provisions are
made to preserve its contents at boot if the configuration requires it.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
Per guidelines, all statements should have braces around them. We do not
have a CI check for this, so a few went in unnoticed.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
ip register holds the stack_info.size (it is passed as argument
into z_arch_user_mode_enter(.)). We trust that the value of
stack_info.size contains the accurate size of the writable
stack buffer, above stack_info.start (as specified in kernel.h).
Therefore, we do not need to subtract any bytes for the MPU
stack guard. This allows us to clean-up one more occurrence of
CONFIG_MPU_REQUIRES_POWER_OF_TWO_ALIGNMENT in userspace.S.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
It's useful to be able to inspect the key returned from
z_arch_irq_unlock() to see if interrupts were enabled at the point
where z_arch_irq_lock() was called. Architectures tend to represent
this is a simple way that doesn't require platform assembly to
inspect.
Adds a simple test to kernel/common that validates this predicate with
a nested lock.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
We do have a multi-architecture latency benchmark now, this one was x86
only, was never used or compiled in and is out-dated.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
arm_core_mpu_enable() and arm_core_mpu_disable() functions are
effectively static functions, only used in the drivers for ARM
and NXP MPU, therefore, we do not need to expose them in the
arm_core_mpu_dev.h header.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
arm_core_mpu_dev.h is an internal API, and is not supposed to
be directly called by kernel / application functions, therefore,
we can move it inside arch/arm/core/cortex_m/mpu directory.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Add Kconfig options:
- ARM_FIRMWARE_USES_SECURE_ENTRY_FUNCS
- ARM_ENTRY_VENEERS_LIB_NAME
Use these to link the veneers lib into the Non-Secure Firmware when
needed.
Also, make the path passed to the linker absolute to make it work with
makefiles.
Signed-off-by: Øyvind Rønningstad <oyvind.ronningstad@nordicsemi.no>
Rename function _PlatformInit() to z_platform_init() to
comply with naming conventions.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
CONFIG_MPU_REQUIRES_POWER_OF_TWO_ALIGNMENT is taken into
account when allocating the area for the interrupt stack
using the K_THREAD_STACK_DEFINE macro. Therefore, we can
simplify how the top of the stack is derived during the
initialization of the system, by removing one more
occurrence of CONFIG_MPU_REQUIRES_POWER_OF_TWO_ALIGNMENT.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
For architectures with custom swap to main, currently:
- arm
- posix
we are now using K_THREAD_STACK_SIZEOF macro to pass the
main thread stack size to z_arch_switch_to_main_thread().
This does not introduce any behavioral changes for posix;
the K_THREAD_STACK_SIZEOF() simply returns the sizeof()
the stack object. For Arm, this allows us to clean-up one
more occurence of CONFIG_MPU_REQUIRES_POWER_OF_TWO_ALIGNMENT
in kernel_arch_func.h.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
When in Sharing Floating Point Services mode we want to
dynamically save and restore the FP registers in thread
context switch, depending on whether the swapped-in and
swapped-out threads are currently using the FP registers.
This commit adds this functionality to the ARM context
switch mechanism. The logic consists of inspecting the
corresponding status flag (present in thread.arch.mode)
to decide whether to save or restore the FP registers.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Only a single bit of the 4-byte internal (thread.arch) status
variable 'mode' is currently used, when we build with User mode
support (CONFIG_USERSPACE=y). In this commit we extend the usage
of 'mode' variable, adding an additional bit-flag to track the
status of the floating point context in a particular thread,
i.e. to track whether FP context is active or not. The status
bit is meant to be used in context-switch, to restore the FP
register context when required.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
The initial carved stack frame only needs to contain the basic
stack frame (state context), because:
- under No/Unshared FP Services mode no FP context is stacked.
- under FP Sharing mode FP context is cleared upon thread
creation.
This commit removes the unnecessary stacking of the FP context
during thread initialization. This results in reduced stack
usage under the Shared FP Services mode (CONFIG_FP_SHARING=y).
Also, we do not need to initialize the FPSCR, because there is
no FP stack frame carved.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
In FP Sharing mode we enable both automatic and lazy state
preservation of the FP context. This configuration improves
interrupt latency. Under this configuration the FP registers
will only be stacked when the thread is swapped out during
context-switch. Before jumping to main() we clear the FPCA
flag of the CONTROL register, so the FP context can be used
by the main thread.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
The compiler generates errors of the form
error: "CONFIG_MPU_REQUIRES_POWER_OF_TWO_ALIGNMENT" is not defined,
evaluates to 0 [-Werror=undef]
when -Wundef is used and the config option was turned off. Change
check to if defined().
Signed-off-by: Bradley Bolen <bbolen@lexmark.com>
The compiler generates errors of the form
error: "CONFIG_ZERO_LATENCY_IRQS" is not defined, evaluates to 0
[-Werror=undef]
when -Wundef is used and the config option was turned off. Change
check to if defined().
Signed-off-by: Bradley Bolen <bbolen@lexmark.com>
The compiler generates errors of the form
error: "CONFIG_IRQ_OFFLOAD" is not defined, evaluates to 0
[-Werror=undef]
when -Wundef is used and the config option was turned off. Change
check to if defined().
Signed-off-by: Bradley Bolen <bbolen@lexmark.com>
The compiler generates errors of the form
error: "CONFIG_USERSPACE" is not defined, evaluates to 0 [-Werror=undef]
when -Wundef is used and the config option was turned off. Change
check to if defined().
Signed-off-by: Bradley Bolen <bbolen@lexmark.com>
The compiler generates errors of the form
error: "CONFIG_STACK_POINTER_RANDOM" is not defined, evaluates to 0
[-Werror=undef]
when -Wundef is used and the config option was turned off. Change
check to if defined().
Signed-off-by: Bradley Bolen <bbolen@lexmark.com>
The compiler generates errors of the form
error: "CONFIG_ARC_FIRQ" is not defined, evaluates to 0 [-Werror=undef]
when -Wundef is used and the config option was turned off. Change
check to if defined().
Signed-off-by: Bradley Bolen <bbolen@lexmark.com>
The compiler generates errors of the form
error: "CONFIG_USERSPACE" is not defined, evaluates to 0 [-Werror=undef]
when -Wundef is used and the config option was turned off. Change
check to if defined().
Signed-off-by: Bradley Bolen <bbolen@lexmark.com>
Make _get_num_regions() return a constant representing the
number of HW MPU regions, defined in DTS, if such define
is available. This removes the need of evaluating the
number of regions at run-time. The ASSERT in arm_mpu_init()
is expanded, to cover that case, where the number of
regions is taken from DTS.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
In z_NanoFatalErrorHandler() we want to print the name
of the thread that has faulted, when building with
CONFIG_THREAD_NAME. This commit adds this functionality
for the Arm architecture.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
The typeof definitions in tz.h and cmse.h are not required
any more, as the ARM GCC toolchain intrinsics have been
modified to use __typeof__ directly.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
In kernel_arch_init() we initialize the ARM core (interrupt
setup, fault init, etc.) so we can also move z_clearfaults()
in the same function and skip invoking it in the SoC init
functions.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
During the conversion of .bin to .o objcopy was not setting the
section to be readonly causing the .rodata in the final image has
write permission.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
With -O0 optimizion, gcc compiler doesn't inline "static inline"
marked function. So when function call return from function
set_and_switch_to_psp which is to switch sp from MSP to PSP, the
ending "mov sp, r7" instruction will overwrite the just updated
sp value(PSP) with the beginning stack pointer(should be MSP)
stored in r7 register, so the switch doesn't happen. And it causes
unpredictable problems in the initialization process, the backward
analysis for this problem can be found on Github issue #15794.
Fixes: #15794.
Signed-off-by: Wentong Wu <wentong.wu@intel.com>
This commit fixes a build error, when building with
CONFIG_ARM_SECURE_FIRMWARE=y. The error was introduced
in #15930 (6f19d0), where we added internal structure
to the exception stack frame struct.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
When compiling the kernel with CONFIG_SYS_CLOCK_TICKS_PER_SEC=0,
the CONFIG_SYS_CLOCK_EXISTS internal variable is unset.
This completely disables timer handling in the kernel, but a couple of
spots missed the required conditional compilation.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
The only use of the BOOTLOADER_UNKNOWN config option is on x86, where
it controls whether a multiboot header is embedded in the output.
This patch renames the option to be more descriptive, and makes it
an x86-specific option, rather than a Zephyr top-level option.
This also enables X86_MULTIBOOT by default, since the header only
occupies 12-16 bytes of memory and is (almost always) harmless.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
The ARM Cortex-M Exception Stack Frame (ESF) may consist of
several stack frame contexts (basic state context, additional
state context, FP context, etc.). To reflect these structural
properties, this commit re-factors the ESF, splitting out the
basic stack frame, holding the state context, into its own
struct container. The commit does not introduce behavioral
changes.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Within the past few days, an update to the Ubuntu 18.04 toolchain has
begun emitting code sections during link that are messing with our
stub generation. They are appearing in the 32 bit stub link despite
not being defined in the single object file, and (worse) being
included in the output segment (i.e. at the start of the bootloader
entry point!) despite not being specifically included by the linker
script. I don't understand this behavior at all, and it appears to be
directly contrary to the way the linker is documented.
Marc Herbert discovered this was down to gcc being called with
--enable-default-pie, so -no-pie works to suppress this behavior and
restore the default. And it's correct: we aren't actually generating
a position independent executable, even if we don't understand why the
linker script is being disregarded (to include sections we don't
include). See discussion in the linked github issue.
Fixes#15877
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The zero-riscy core on the rv32m1 soc does not implement hardware loop
extensions and thus should not enable RISCV_SOC_CONTEXT_SAVE, however it
does still need access to the EVENTx_INTPTPENDCLEAR symbol which comes
from GEN_SOC_OFFSET_SYMS().
Split out the soc offset symbols into a separate config so we can enable
them without enabling soc context saving.
Signed-off-by: Maureen Helm <maureen.helm@nxp.com>
When some header are included into C++ source file, this kind of
compilations errors are generated:
error: invalid conversion from 'void*'
to 'u32_t*' {aka 'unsigned int*'} [-fpermissive]
Signed-off-by: Benoit Leforestier <benoit.leforestier@gmail.com>
Initialize the Floating Point Status and Control Register when in
Unshared FP Registers mode (In Shared FP Registers mode, FPSCR is
initialized at thread creation for threads that make use of the FP).
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit removes the activation of the FP context
in ARM system boot. There is no need to do this, since
the FP context will be activated in the presence of
floating point instructions. We update the reference
documentation accordingly.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Under Unshared FP register mode we are not sharing the
FP context among different threads, so we do not need to
include the FP high registers bank in the thread.arch
container.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Under Unshared FP register mode we are not stacking the
FP context in exception entries, so we do not need to
include the FP registers bank in the exception stack
frame structure.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Under unshared FP registers mode the FP register bank is
meant to be used by a single thread context. Therefore,
there is no need for automatic stacking of the FP register
bank at exception entries, or context switch, as the
registers are not expected to be shared among multiple
contexts.
Under unshared FP registers mode we only need to clear the
FPSCR register once, before jumping to main(). However, we
initialize the FPSCR already at boot in case FP operations
need to be performed during boot.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
In ARM builds with FP services (CONFIG_FLOAT=y) but without user
mode support (CONFIG_USERSPACE=n) we do not need and should not
enable full-access to the FP co-processor. Instead, we should
enabled access by privileged code only.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Update the name of mem-domain API function to add a partition
so that it complies with the 'z_' prefix convention. Correct
the function documentation.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
When performing thread context-switch it is not necessary to
have IRQs locked while saving the current thread's callee-saved
(and possibly floating point) registers. We only need to lock
the interrupts when accessing the thread ready queue cache.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit renames the symbol CPU_HAS_SYSTICK to
CPU_CORTEX_M_HAS_SYSTICK, to look similar to all
other CPU_CORTEX_M_HAS_ options, and moves the
K-config symbol definition from arm/core/Kconfig to
arm/core/cortex_m/Kconfig.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
ARM SysTick timer is implemented by default in ARMv7-M
and Mainline ARMv8-M processors, so we include the
corresponding Kconfig symbol in arch/arm/core/cortex-M/Kconfig
and remove the selections from the Cortex-M SOCs.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
It's relatively hard to figure out what thread a crash happens in
from the crash dump. E.g, it's usually not immediately possible to
find it out from linker map due to the fact that static symbols are
not there (https://sourceware.org/bugzilla/show_bug.cgi?id=16566).
So, try to do it as easy if possible, by just printing thread name
in a dump, if thread names are enabled at all.
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
The struct _caller_saved is not used. Most architectures put
automatically the registers onto stack, in others architectures the
exception code does it.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
The struct _kernel_ach exists only because ARC' s port needed it, in
all other ports this was defined as an empty struct. Turns out that
this struct is not required even for ARC anymore, this is a legacy
code from nanokernel time.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
A parallel PCI implementation ("pcie") is added with features for PCIe.
In particular, message-signaled interrupts (MSI) are supported, which
are essential to the use of any non-trivial PCIe device.
The NS16550 UART driver is modified to use pcie.
pcie is a complete replacement for the old PCI support ("pci"). It is
smaller, by an order of magnitude, and cleaner. Both pci and pcie can
(and do) coexist in the same builds, but the intent is to rework any
existing drivers that depend on pci and ultimately remove pci entirely.
This patch is large, but things in mirror are smaller than they appear.
Most of the modified files are configuration-related, and are changed
only slightly to accommodate the modified UART driver.
Deficiencies:
64-bit support is minimal. The code works fine with 64-bit capable
devices, but will not cooperate with MMIO regions (or MSI targets) that
have high bits set. This is not needed on any current boards, and is
unlikely to be needed in the future. Only superficial changes would
be required if we change our minds.
The method specifying PCI endpoints in devicetree is somewhat kludgey.
The "right" way would be to hang PCI devices off a topological tree;
while this would be more aesthetically pleasing, I don't think it's
worth the effort, given our non-standard use of devicetree.
Signed-off-by: Charles E. Youse <charles.youse@intel.com>
Right now only numerical values are printed which must
be looked up in the Designware ARCv2 ISA Programmer's
Reference, which is not public.
Add a non-default Kconfig to print more information at
the expense of footprint, and enable it for all the simulator
targets.
We only print code/parameter details for machine check and
protection violations, more may be added later as desired.
This should cover all the exceptions we commonly encounter
for memory protection.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
The nxp mpu uses the logical OR of access permissions when multiple
region descriptors apply to a given memory access. This means that we
must partition the sram into two non-overlapping regions to implement
the mpu stack guard. This partitioning gets reconfigured at every
context switch, and if an interrupt occurs during this time, it can
cause a fault because we do not have a valid mpu region descriptor for
the sram.
This scenario was observed on frdm_k64f in tests/posix/common before
commit 2895da02a4, which changed timing.
In this case, we couldn't even print fault information to the console
and the hardware would reset. It looked a lot like a watchdog reset,
unless you attached a debugger to see the fault.
A similar problem was fixed in commit
ec424b7a45, but this change temporarily
disabled the mpu.
Fix both cases by locking interrupts during the critical sections, as
this is more secure than disabling the mpu.
Signed-off-by: Maureen Helm <maureen.helm@nxp.com>
* fix the stack allocation and initialization
for mpu stack guard when USERSPACE is not configured
* fixes#15163
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
* it's not reliable to use exc_nest_count to decide nest interrupt.
A better option is to use IRQ_ACT
* ext_nest_count is just used to switch interrupt stack
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
Update the files which contain no license information with the
'Apache-2.0' SPDX license identifier. Many source files in the tree are
missing licensing information, which makes it harder for compliance
tools to determine the correct license.
By default all files without license information are under the default
license of Zephyr, which is Apache version 2.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
The results were incorrect because the timer was firing the
interrupts before the measurement was made.
Fixes: GH-14556
Signed-off-by: Adithya Baglody <adithya.nagaraj.baglody@intel.com>
This macro is slated for complete removal, as it's not possible
on arches with an MPU stack guard to know the true buffer bounds
without also knowing the runtime state of its associated thread.
As removing this completely would be invasive to where we are
in the 1.14 release, demote to a private kernel Z_ API instead.
The current way that the macro is being used internally will
not cause any undue harm, we just don't want any external code
depending on it.
The final work to remove this (and overhaul stack specification in
general) will take place in 1.15 in the context of #14269Fixes: #14766
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Rename reserved function names in arch/ subdirectory. The Python
script gen_priv_stacks.py was updated to follow the 'z_' prefix
naming.
Signed-off-by: Patrik Flykt <patrik.flykt@intel.com>
Rename reserved function names in drivers/ subdirectory. Update
function macros concatenatenating function names with '##'. As
there is a conflict between the existing gpio_sch_manage_callback()
and _gpio_sch_manage_callback() names, leave the latter unmodified.
Signed-off-by: Patrik Flykt <patrik.flykt@intel.com>
Similar issue to what was fixed earlier in the MPUv3
code. start + size should be <= r_addr_end. Fixes
a problem where the last byte of an MPU region is
incorrectly reported as out-of-bounds.
Fixes: #15131
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
The legacy struct s_coopFloatReg was never being used, though it was
an empty struct (not wasting space), some symbols were being generate
for it.
Nevertheless, neither C99 nor C11 allow empty structs, so this
was also a violation to the C standards.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
This commit initializes the thread.mode variable, right before
dropping thread privilege level to user mode. This is required,
as we need to know the privilege level of the thread, in case
we need to context-switch it -out and -in again (e.g. in case
an interrupt triggers a context-switch).
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
The arch.mode status flag is always inline with the actual
thread execution privilege leve. Therefore, we do not need
to be saving-off the thread's mode every time we intend to
context switch-out the thread.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit updates the thread.mode variable, in system
calls, whenever we modify the execution privilege level.
We need to do this, in order to be able to properly handle
context-switching triggered by ISRs, while doing the system
calls. The commit ensures that a context-switch, triggered
by an ISR during or after,a system call will preserve the
right privilege level.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
The original value of 256 was selected more or less randomly
and special cases keep proliferating. Until we have a formal
method of proving maximum syscall stack depth, set to 1024.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
* mpu stack guard exception is a kind of mpu violation exception
* use SW way to distingusih it from other mpu vioation exception
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
* HW_STACK_PROTECTION can be done by STACK_
CHECKING or MPU stack guard. ARC STACK_CHECKING is prioritized
over MPU-based stack guard
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
Discovered with pylint3.
Use the placeholder name '_' for unproblematic unused variables. It's
what I'm used to, and pylint knows not to flag it.
Python tip:
for i in range(n):
some_list.append(0)
can be replaced with
some_list += n*[0]
Similarly, 3*'\t' gives '\t\t\t'.
(Relevant here because pylint flagged the loop index as unused.)
To do integer division in Python 3, use // instead of /.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
BIT macro uses an unsigned int avoiding implementation-defined behavior
when shifting signed types.
MISRA-C rule 10.1
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
stack check exception may come out with other protection
vilation, e.g. MPU read/write. So the possible paramter
will be 0x02 | [0x4 | 0x8].
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
MISRA defines a serie of essential types, boolean, signed/unsigned
integers, float, ... and operations must respect these essential types.
MISRA-C rule 10.1
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
MISRA defines a serie of essential types, boolean, signed/unsigned
integers, float, ... and operations must respect these essential types.
MISRA-C rule 10.1
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
BIT macro uses an unsigned int avoiding implementation-defined behavior
when shifting signed types.
MISRA-C rule 10.1
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
This adds a compiler option -fno-inline for code coverage on
architectures which supports doing code coverage. This also
modifies the ALWAYS_INLINE macro to not do any inlining. This
needs to be done so code coverage can count the number of
executions to the correct lines.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Add SYS_POWER_ prefix to HAS_STATE_SLEEP_, HAS_STATE_DEEP_SLEEP_
options to align them with names of power states they control.
Following is a detailed list of string replacements used:
s/HAS_STATE_SLEEP_(\d)/HAS_SYS_POWER_STATE_SLEEP_$1/
s/HAS_STATE_DEEP_SLEEP_(\d)/HAS_SYS_POWER_STATE_DEEP_SLEEP_$1/
Signed-off-by: Piotr Mienkowski <piotr.mienkowski@gmail.com>
This commit cleans up names of system power management functions by
assuring that:
- all functions start with 'sys_pm_' prefix
- API functions which should not be exposed to the user start with '_'
- name of the function hints at its purpose
Signed-off-by: Piotr Mienkowski <piotr.mienkowski@gmail.com>
There exists SoCs, e.g. STM32L4, where one of the low power modes
reduces CPU frequency and supply voltage but does not stop the CPU. Such
power modes are currently not supported by Zephyr.
To facilitate adding support for such class of power modes in the future
and to ensure the naming convention makes it clear that the currently
supported power modes stop the CPU this commit renames Low Power States
to Slep States and updates the documentation.
Signed-off-by: Piotr Mienkowski <piotr.mienkowski@gmail.com>
Added an option to stop the execution of the posix arch based
executable on the first fault, even if the fault stemmed from a
non essential thread.
Having it fail faster, in the first fault, will ease debugging
in many cases.
The option is disabled by default to preserve the old behavior.
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
Not needed in Python. Detected by check C0325 in pylint3.
Also replace an
if len(tag):
with just
if tag:
Empty strings, byte strings, lists, etc., are falsy in Python.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
Making a clean slate for some pylint CI tests. Only enabling relatively
uncontroversial stuff.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
Field stack_info.size is supposed to hold the actual writable
thread stack area, above what is pointed to by stack_info.start.
Therefore, in all MPU architectures we can program the Thread
Stack MPU region as starting from thread->stack_obj and with
size equal to stack_info.size plus the (possible) difference
between stack_obj and stack_info.start.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Consider a stack buffer at address 0x10000 with size 1024.
If a thread is created with this stack object, the resulting
fields in thread.stack_info ended up being a base address of
0x10020 with size 1024. The guard size needed to be subtracted
from the size in order for the bounds to be correct.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
%z isn't available in Python, and makes the code raise a ValueError. Use
%d instead. Integers in Python 3 are not sized/signed (though it's
probably a typo from C).
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
FPU's control and status register (FPCSR) for a thread is not
initialized. Random values are written to this register and this leads
to failure on fp_sharing test.
FPCSR register is set to 0 according to the value of FPDSCR.
Signed-off-by: Andrei Gansari <andrei.gansari@nxp.com>
The code was probing the byte immediately after the last byte
of the buffer, instead of the last byte of the buffer.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Fix the type of struct field index to int, to avoid
loss of precision, in case the variable holds the value
-EINVAL (in the case of errors in MPU configuration).
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
In arm_core_mpu_mem_partition_config_update() we must fix the
index value of the start of the for loop, to address the case
when dynamic MPU regions are programmed in index values lower
than static_regions_num. (This might occur in ARMv8-M MPU,
depending on the order of programming the map of the
dynamic regions in run-time). We introduce function
_get_dyn_region_min_index(), and provide the implementation in
arm_mpu_v8_internal.h and arm_mpu_v7_internal.h (trivial).
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
In this commit we fix the implementation of internal function
_mpu_buffer_validate(), so it can work in the presence of a
security attribution unit (SAU, or IDAU). The fast validation
based on the CMSE address range check intrinsic is performed
first: if it fails, then a second, MPU-only check is performed
for ARMV8-M platforms that have TEE capabilities (i.e. SAU/IDAU
units).
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Enhance the documentation for internal ARMv8-M MPU
function _mpu_buffer_validate().
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit adds an explanatory comment in ARM __swap()
function, stressing when, exactly, context-switch is
expected to occur, when we pend the PendSV exception.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
The ARM Cortex-M 321 application note is stressing that if
we disable interrupts by executing CPSID i(f), or by MSR
instructions (on PRIMASK, FAULTMASK registers), there is no
requirement to add barrier instructions after disabling
interupts. However, in ARMv7-M (and ARMv8-M Mainline) we use
BASEPRI, instead. Therefore, if we need the effect of disabling
interrupts to be recongnized immediately we should add barrier
instructions. This commit adds DSB and ISB barriers when
disabling interrupt using BASEPRI in the generic
arm _irq_lock() function as well as in the PendSV handler,
where we need to access kernel globals right after the interrups
are disabled.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
It's simpler and cleaner to move read_timer_start_of_isr to
_isr_demux. But the interrupt latency may be not accurate.
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
Since #14545 was merged building with USERSPACE and NO_OPTIMIZATIONS
has been broken due to #5226.
To fix#5226 we increase the size of KOBJECT_TEXT_AREA when
NO_OPTIMIZATIONS.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
The clang ARM assembler is a bit stricter than GNU as. Change mov to
movs for ARMv6 case of z_arch_switch_to_main_thread.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
* separate the ARC MPU driver into 2 parts
* arc_mpu_v2_internal.h for ARC MPUv2
* arc_mpu_v3_internal.h for ARC MPUv3
* For ARC MPUv2, keep the main design, but update and optimize the code
* For ARC MPUv3, implement mpu region split to supprt MPU region overlap
* misc updates and bug fixes
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
This commit partially reverts the changes introduced in
(bbe1a19786), where the PSP
modification, the interrupt enabling, and the branch to main
thread were refactored in C functions. This introduced an issue
with stack usage when compiling without any optimization.
Reverting to use assembly functions to jump to main thread
ensures, now, that the pointer to main() is preserved intact
while changing the PSP and passed correctly to z_thread_entry().
Fixes#14471.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit simply moves the MPU re-programming of
dynamic regions during initialization to occur right
before switching PSP to the main thread stack. As a
result, the MPU re-programming will execute using
the interrupt stack. No functional changes are
introduced.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
The clang ARM assembler is a bit stricter than GNU as, make a few
changes so things build with both
Signed-off-by: Peter Smith <peter.smith@linaro.org>
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
Architecture defconfigs are not used anymore and are stale. Remove them
to avoid confusion.
Related to #14442
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
The shared page is inherently used in multiprocessor contexts where
the compiler optimizer can trip us up (specifically, a spin on
num_active_pus was being hoisted out of the loop on some gcc's). Put
the volatile declartion into the struct pointer itself instead of
relying on the code to get it right.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Currently thread abort doesn't work if a thread is currently scheduled
on a different CPU, because we have no way of delivering an interrupt
to the other CPU to force the issue. This patch adds a simple
framework for an architecture to provide such an IPI, implements it
for x86_64, and uses it to implement a spin loop in abort for the case
where a thread is currently scheduled elsewhere.
On SMP architectures (xtensa) where no such IPI is implemented, we
fall back to waiting on an arbitrary interrupt to occur. This "works"
for typical code (and all current tests), but of course it cannot be
guaranteed on such an architecture that k_thread_abort() will return
in finite time (e.g. the other thread on the other CPU might have
taken a spinlock and entered an infinite loop, so it will never
receive an interrupt to terminate itself)!
On non-SMP architectures this patch changes no code paths at all.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Before we're initialized and can use proper synchronization, the CPU
initialization path spins on the thread entry function to be non-null.
But the data wasn't tagged volatile, and with gcc 8.2.1 (but not
6.2.0) the optimizer was hoisting the reads to SMP init would spin
forever.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The previous scheme where the xuk layer would call out to the to
"fetch" the stack for a SMP CPU at startup was sorta weird, and an
impedance mismatch with Zephyr which has a "start this CPU" call
instead. It also got broken when x86_64 started launching CPUs
(correctly) on their interrupt stacks instead of a temporary area;
they weren't ready yet when xuk initialization was happening and the
system would deadlock waiting for code that can't run yet to provide a
stack.
Note that this preserves the somewhat quirky behavior that Zephyr's
CPU numbering is just the order in which the SMP CPUs emerge from
initialization and not a hardware ID.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
When this code was written, there was no "stack frame" struct defined.
There is now, so use that for clarity and concision. Also remove an
obvious comment (I mean, duh, we can put any segment selectors in
those fields we want).
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
There was a spot where the early boot code was logging using %p to
emit an integer, and Coverity doesn't like that. Fixes#14420 and
Fixes#14418
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Speculative execution side channel attacks can read the
entire FPU/SIMD register state on affected Intel Core
processors, see CVE-2018-3665.
We now have two options for managing floating point
context between threads on x86: CONFIG_EAGER_FP_SHARING
and CONFIG_LAZY_FP_SHARING.
The mitigation is to unconditionally save/restore these
registers on context switch, instead of the lazy sharing
algorithm used by CONFIG_LAZY_FP_SHARING.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Update reserved function names starting with one underscore, replacing
them as follows:
'_k_' with 'z_'
'_K_' with 'Z_'
'_handler_' with 'z_handl_'
'_Cstart' with 'z_cstart'
'_Swap' with 'z_swap'
This renaming is done on both global and those static function names
in kernel/include and include/. Other static function names in kernel/
are renamed by removing the leading underscore. Other function names
not starting with any prefix listed above are renamed starting with
a 'z_' or 'Z_' prefix.
Function names starting with two or three leading underscores are not
automatcally renamed since these names will collide with the variants
with two or three leading underscores.
Various generator scripts have also been updated as well as perf,
linker and usb files. These are
drivers/serial/uart_handlers.c
include/linker/kobject-text.ld
kernel/include/syscall_handler.h
scripts/gen_kobject_list.py
scripts/gen_syscall_header.py
Signed-off-by: Patrik Flykt <patrik.flykt@intel.com>
We add two points where we add lfences to disable
speculation:
* In the memory buffer validation code, which takes memory
addresses and sizes from userspace and determins whether
this memory is actually accessible.
* In the system call landing site, after the system call ID
has been validated but before it is used.
Kconfigs have been added to enable these checks if the CPU
is not known to be immune on X86.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
We introduce hidden Kconfigs for all speculative
side channel attacks that we plan to address in the
kernel and update the existing ones to indicate their
CVEs.
This list keeps growing, so introduce a new config
CONFIG_X86_NO_SPECULATIVE_VULNERABILITIES, for CPUs
which don't speculatively execute, or are otherwise
immune by design.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Commit 45a7e5d076 removed scripts/sysgen
and _k_task_list/_k_event_list sections are no longer being used.
So remove them.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
The commit adds an __ASSERT statement when configuring
the ARMv8-M build-in stack guard for supervisor thread
stack overflow protection, to catch the case when the
stack_info.start does not point to the beginning of the
thread stack allocated area, pointed by stack_obj.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit does the following:
- moves the reservation of the area for local thread data
inside the arch-specific _new_thread() function,
- aligns the end of the thread's stack with the granularity
of the MPU, already at thread's initialization, moving it
out of userspace_enter() function. In this way, the local
thread data are stored inside the thread stack area that
is, later, assigned unprivileged RW permissions (if thread
enters user mode).
- passes the right thread stack size information to function
_new_thread_init(), which stores it into stack_info.start.
As a result, the whole usable User Thread stack area (i.e.
excluding a possible guard area) falls inside the region
programmed as unprivileged RW, during context-switch.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Appears within a menu that already has 'depends on ARM_SECURE_FIRMWARE'.
'depends on FOO' on a menu will add 'depends on FOO' to each item within
it. 'if FOO' work similarly.
Tip: When adding a symbol, check its dependencies in the menuconfig
('ninja menuconfig', then / to jump to the symbol). The menuconfig also
shows how the file with the symbol got included, so if you see
duplicated dependencies, it's easy to hunt down where they come from.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
Appears within a menu that already has a 'depends on ARCH_POSIX'.
'depends on FOO' on a menu will add 'depends on FOO' to each item within
it. 'if FOO' work similarly.
Tip: When adding a symbol, check its dependencies in the menuconfig
('ninja menuconfig', then / to jump to the symbol). The menuconfig also
shows how the file with the symbol got included, so if you see
duplicated dependencies, it's easy to hunt down where they come from.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
- The ARC CPU_HAS_MPU dependencies were added within the menu
menu "ARCH MPU Options"
depends on CPU_HAS_MPU
(arch/arc/core/mpu/Kconfig is source'd within it).
- The ARM CPU_HAS_MPU dependencies were redundantly added by
if CPU_HAS_MPU
source "arch/arm/core/cortex_m/mpu/Kconfig"
endif
and by some 'depends on CPU_HAS_MPU' within that file. Remove the
'depends on' and move the 'if' into the file instead.
Tip: Jump to symbols with '/' in the menuconfig and press '?' to check
their dependencies. If there are duplicated dependencies, the
'included via ...' path can be handy to discover where they are added.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
The lowest level of the x86_64 bit port supports the full 64 bit ABI
just fine, but Zephyr does not (yet) and builds under the x32 ABI
instead. The xuk layer can be built with or without the -mx32 switch,
and it had a configurable in a header to tell it what it was. At some
point during development I swept through and turned all those tunables
into kconfigs, but this one wasn't used by zephyr and so it got the
CONFIG_* rename but never had an entry added to a Kconfig file to
match it, and eventually got picked up by Ulf's unused symbol
detector.
Rename back. It will probably become a kconfig again someday when we
need it. Fixes#14059.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Refactor internal ARM v7m and v8m MPU region re-programming
functions to take as argument a pointer to k_mem_partition
objects, instead of the object itself.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Refactor internal NXP MPU region re-programming functions
to take as argument a pointer to k_mem_partition objects,
instead of the object itself.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit refactors the MPU region re-programming functions,
to take as argument an array of pointers to memory partition
structures, instead of the whole array of the partitions. In
this way the stack usage can be minimized, if the actual
partition information is kept in statically allocated memory.
instead of the map itself.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
The "ARM Cortex-M options" menu is already within a 'if CPU_CORTEX_M',
so no need to put 'depends on CPU_CORTEX_M' on it.
Tip: Jump to symbols with '/' in the menuconfig and press '?' to check
their dependencies. If there are duplicated dependencies, the
'included via ...' path can be handy to discover where they are added.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
Based on the definition of _ARCH_THREAD_STACK_DEFINE() macro
for ARM, the MPU Stack Guard region is placed inside the
allocated stack object, only if
CONFIG_MPU_REQUIRES_POWER_OF_TWO_ALIGNMENT and CONFIG_USERSPACE
are both set. For ARM stack objects, allocated using the
_ARCH_THREAD_STACK_DEFINE() macro, such as the threads' stacks
and the interrupt stack, the above must be reflected on how we
set the initial start of the stacks.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
On x86, if a supervisor thread belonging to a memory domain
adds a new partition to that domain, subsequent context switches
to another thread in the same domain, or dropping itself to user
mode, does not have the correct setup in the page tables.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
We need a copy of the flags field for ever PTE we are
updating, we can't just keep OR-ing in the address
field.
Fixes issues seen when setting flags for memory regions
larger than a page.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
During speculative execution, non-present pages are treated
as valid, which may expose their contents through side
channels.
Any non-present PTE will now have its address bits zeroed,
such that any speculative reads to them will go to the NULL
page.
The expected hit on performance is so minor that this is
enabled at all times.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
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>
This commit introduces a generic arch/arm function to evaluate
thread stack corruption. The function shall be used upon
occurrence of MemManage and Bus Faults.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit adds an explanatory note in the places
where we manually adjust the PSP after stacking Bus
and MemManage errors, stressing that this is due to
the fact that interrupt tail-chaining cannot always
be guaranteed by the processor.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
The commit removes a redundant check for imprecise bus fault,
when a precise bus fault has occurred, now that we have
re-worked the BusFault handler to iterate through all possible
bus errors.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit reworks the logic, based on which we assess
thread stack corruptions upon BusFault exceptions triggered
by the NXP MPU module. BusFaults rely on calling external
function to assess whether a thread stack corruption due to
overflow has occurred. Upon detection of stack corruption,
the current stack pointer is manually adjusted upwards, to
prevent un-stacking errors from firing up upon returning from
the exception.
Additionally, the commit allows to be checking for all different
types of BusFault occurrences that are reported through BFSR,
without stopping after having spotted the first BusFault type.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit reworks the logic, based on which we assess
thread stack corruptions upon MemManageFault exceptions.
Furthermore, it takes out all kernel-related logic and relies
on calling external function 'evaluate_thread_stack_corruption'
to assess whether a thread stack overflow has occurred. Upon
detection of stack corruption, the current stack pointer is
manually adjusted upwards, to prevent un-stacking errors from
firing up upon returning from the exception.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit improves the debug message on UsageFaults
related to Stack overflow, stating that the context
area is not valid. An inline comment is added, with
more detailed explanation.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit fixes two Constant Expression Result errors,
in the NXP MPU driver due to incompatible integer types.
Fixes#13836.
Fixes#13865.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit fixes several essential inline comments in the
core assembly code for ARM, improving code readability.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit applies the same fixes, concering the Constant
Expression Result errors identified in the ARMv7-M and the
NXP MPU drivers, into the ARMv8-M MPU implementation.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit fixes two Constant Expression Result errors,
in the ARMv7-M MPU driver due to incompatible integer types.
Fixes#13844.
Fixes#13854.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit aligns the implementation of arm_core_mpu_disable()
function in NXP MPU with the implementation in the ARM MPU
module, by introducing a Data Memory Barrier, (DMB) instead of
a DSB instruction. This is in accordance with the ARM guidelines
of using the memory protection unit.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Some more were added since the cleanup pass in June 2018. See e.g.
commit 2d50da70a1 ("drivers: ipm: Kconfig: Remove redundant 'default n'
properties") for a motivation. It also avoids people wondering whether
or not they need to put in 'default n'.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
CPU_LPS_n name used to indicate a low power state is cryptic and
incorrect. The low power states act on the whole SoC and not exclusively
on the CPU. This patch renames CPU_LPS_n states to LOW_POWER_n. Also
HAS_ pattern for Kconfig options is used in favor of a non standard
_SUPPORTED. Naming of deep sleep states was adjusted accordingly.
Following is a detailed list of string replacements used:
s/SYS_POWER_STATE_CPU_LPS_(\d)_SUPPORTED/HAS_STATE_LOW_POWER_$1/
s/SYS_POWER_STATE_CPU_LPS_(\d)/SYS_POWER_STATE_LOW_POWER_$1/
s/SYS_POWER_STATE_DEEP_SLEEP_(\d)_SUPPORTED/HAS_STATE_DEEP_SLEEP_$1/
Signed-off-by: Piotr Mienkowski <piotr.mienkowski@gmail.com>
This commit removes dependency on SYS_POWER_LOW_POWER_STATES_SUPPORTED,
SYS_POWER_DEEP_SLEEP_STATES_SUPPORTED Kconfig options. Power management
SYS_POWER_LOW_POWER_STATES, SYS_POWER_DEEP_SLEEP_STATES options depend
now directly on specific power states supported by the given SoC. This
simplifies maintenance of SoC Kconfig files.
Signed-off-by: Piotr Mienkowski <piotr.mienkowski@gmail.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>
Instead of having to enable ramfunc support manually, just make it
transparently available to users, keeping the MPU region disabled if not
used to not waste a MPU region. This however wastes 24 bytes of code
area when the MPU is disabled and 48 bytes when it is enabled, and
probably a dozen of CPU cycles during boot. I believe it is something
acceptable.
Note that when XIP is used, code is already in RAM, so the __ramfunc
keyword does nothing, but does not generate an error.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The linker file defines the __ramfunc_ram_size symbols to get the size
of the __ramfunc_ram section. Use that instead of computing the value at
runtime from the start and end symbols. This saves 16 bytes of code with
CONFIG_RAM_FUNCTION=y.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Using __ramfunc to places a function in RAM instead of Flash.
Code that for example reprograms flash at runtime can't execute
from flash, in that case must placing code into RAM.
This commit create a new section named '.ramfunc' in link scripts,
all functions has __ramfunc keyword saved in thats sections and
will load from flash to sram after the system booted.
Fixes: #10253
Signed-off-by: qianfan Zhao <qianfanguijin@163.com>
This commit fixes a bug that was always propagating the
default error reason (_NANO_ERR_HW_EXCEPTION) to the
system fatal error handler for (strictly) Secure and
for Non-Secure FW builds on ARMv8-M.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Include the EXC_RETURN payload flag macro definitions into
compilation when building for Non-Secure firmware, too.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit fixes a bug for ARM builds with no USERSPACE support
where the MPU stack guard had been erroneously placed, i.e. not
at the beginning of the stack object. This was eating 32-bytes
off of the thread stack. In addition to that, the commit adds a
usefull ASSERT statement for builds with USERSPACE, which checks
that the start of the guard matches the beginning of the entire
stack object.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Its possible that memcpy is marked to check that a NULL value is passed
to it. The instance of memcpy in relocate_vector_table its valid that
we're are copying to 0/NULL. So supress the possible warning generated
by gcc -Wnonnull.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
With this commit we clear the BusFault Status Register error
bits, after processing the fault, for all ARM architecture
variants and implementations (not only for ARMv8-M mainline).
That is a fix, since the bits are sticky in the ARMv7-M
variants, as well. The commit fixes, additionally, a typo
in the spelling of BFSR.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
128 already gets exceeded in a lot of cases, and the binaries
with the new SDK are very slightly larger. Just kick this up
to 256 and get rid of all the exceptions.
Fixes: #13594
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>
This commit adds the logic to indicate that a Non-Secure fault
exception has been triggered due to a processor fault that has,
in fact, occurred in the Secure state.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit fixes the way we determine where the stacked
frame lies for (strictly) ARMv8-M Non-Secure firmware. It
introduces no changes to the default ARMv8-M behavior (i.e.
ARMv7-M compatible).
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Some of power states used numerical suffix while otthers not.
This commit adds proper suffix to all power state names.
Signed-off-by: Piotr Zięcik <piotr.ziecik@nordicsemi.no>
f
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>
This is needed, because an interrupt can happen after the main/static
MPU region is disabled and before it is re-enabled.
This region gets implicitly disabled inside the _region_init call, when
its configuration registers change:
SYSMPU->WORD[index][0] = region_base;
SYSMPU->WORD[index][1] = region_end;
SYSMPU->WORD[index][2] = region_attr;
SYSMPU->WORD[index][3] = SYSMPU_WORD_VLD_MASK;
The TRM says this about the WORD0, WORD1 and WORD2 registers:
Writes to this register clear the region descriptor’s valid bit
(RGDn_WORD3[VLD]).
And thus if an interrupt happens after writing to WORD0 and before
writing VLD to WORD3 again, the code executes with enabled and yet
misconfigured MPU.
Fixes#13482
Signed-off-by: Tomasz Gorochowik <tgorochowik@antmicro.com>
This reverts commit c9ace83c89 which
bypasses setting cache attributes.
The previous cache attributes actually set the text/data/etc.
sections to be inaccessible. So fix it.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
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 thread switching tracing calls are done by the kernel,
and not by the archs. So, remove the redundant trace call.
Related to #13357
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.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>
The linker file defines the __gcov_bss_size and _nocache_ram_size
symbols to get the size of the __gcov_bss and _nocache_ram section. Use
that instead of computing the value at runtime from the start and end
symbols.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
_arch_configure_dynamic_mpu_regions() may re-program the MPU
Stack Guard, to guard the privilege stack for overflows,
(if building with option CONFIG_MPU_STACK_GUARD). There is a
risk of actually overflowing the stack while doing the
re-programming. We minimize the risk by placing the
re-programming function immediately after we have switched to
the privileged stack so that the whole privilege stack area is
available for this critical operation.
Note that the risk for overflow is higher if using the normal
thread stack, since we do not control how much stack is actually
left, when user invokes _arm_userspace_enter().
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit moves the clearing of PSPLIM before determining the
privilege stack start and size. In this way the clearing can be
done without unnecessary pushing and popping r3 core register.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit provides two fixes for the ARMv8-M built-in stack
protection in userspace.S. First, it clears the PSPLIM register
before setting the PSP to the privileged stack. In the wake of
this, the privilege stack area can now be positioned enywhere,
i.e. does not have to be above the user stack area. Second, it
removes unncessary push/pop when clearing the PSPLIM register,
upon returning from the system call, since the r3 register is
not used anymore in this routine.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Fixing the documentation for ARM core __svc function, which was
outdated and present only for the ARMv7-M version.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Stacking r3 at this point is not required as the register value
is not used afterwards, therefore, it does not need to be
retained.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Due to code re-organization inside __pendSV(), the stacking
and unstacking of scratch registers around the invocation of
read_timer_end_of_swap() is no longer required.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
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>
The SYS_POWER_LOW_POWER_STATE_SUPPORTED and SYS_POWER_LOW_POWER_STATE
suggests one low power state but these options control multiple
low power state. This commit uses plural in the names to indicate
that.
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>
Just like with _Swap(), we need two variants of these utilities which
can atomically release a lock and context switch. The naming shifts
(for byte count reasons) to _reschedule/_pend_curr, and both have an
_irqlock variant which takes the traditional locking.
Just refactoring. No logic changes.
Signed-off-by: Andy Ross <andrew.j.ross@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>
Introduces the ARCH_ROOT argument, similar to BOARD_ROOT and SOC_ROOT.
This enables support for out-of-tree architectures.
The ARCH_ROOT out-of-tree layout is expected to be the following:
* ${ARCH_ROOT}/arch/${ARCH}/
* ${ARCH_ROOT}/include/arch/${ARCH}/ (Optional)
Signed-off-by: Klaus Petersen <kape@oticon.com>
Replace Cortex-M3 with Cortex-M architecture family
in the header documentation of kernel_arch_data.h and
kernel_arch_func.h, which are generic header files for
the entire familty of ARM Cortex-M CPUs. The commit
adds some more minor style fixes in functions'
documentation.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit enforces default double-word stack alignmnet
on exception entry for Cortex-M3 and Cortex-M4 MCUs. The
patch ensures that we have consistent behavior in all
Cortex-M MCUs (double-world stack alignment on exception
entry is enforced by default in ARMv6, ARMv8, and Cortex-M7
processors).
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit updates the documentation of Kconfig option:
STACK_ALIGN_DOUBLE_WORD for Cortex-M microprocessors, stating
that the option is used in ARMv7-M MCUs to enforce 8-byte
stack alignment upon exception entry.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
use cmsis functions to avoid using direct assembly for main thread
switching.
This patch uses cmsis functions to:
- set PSP(process stack pointer) to main stack
- enable interrupts
Fixes#12878
Signed-off-by: Vikas Manocha <vikas.manocha@st.com>
There is no need to rename the POSIX types, so let's not do it to
simplify things
Also remove an unnecessary guard (POSIX_ARCH) to avoid mystifying
this any more than necessary
Related to #13054
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
The commit 77cb942a97 broke the generation of sw_isr_table for
multi-level IRQs. This patch fixes it.
Fixes#13082.
Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
add the handling of APP_SHARED_MEM.
privileged threads can access all the mem
explictly defined in user mode, i.e., APP_MEM & APP_SHARED_MEM
Signed-off-by: Wayne Ren <wei.ren@synopsys.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>
The selection of the Cortex M systick driver to be used
as a system clock driver is controlled by
CONFIG_CORTEX_M_SYSTICK.
To replace it by another driver CONFIG_CORTEX_M_SYSTICK
must be set to 'n'. Unfortunately this also controls
the interrupt vector for the systick interrupt. It is
now routed to __reserved. More bad the interrupt vector
can not be set by IRQ_CONNECT as it is one of the hard
coded interrupts in the interrupt table.
Route the hard coded systick interrupt to z_clock_isr
and make z_clock_isr a weak symbol that can be overwritten
by an alternative systick system clock driver.
Signed-off-by: Bobby Noelte <b0661n0e17e@gmail.com>
This patch adjusts the way we are doing push/pop operations before
making function calls inside of assembly routines. ARM requires 8
byte aligned stack operations across public interfaces. This means
that we need to be sure to push multiples of 2 registers.
Fixes#2108
Signed-off-by: Andy Gross <andy.gross@linaro.org>
If you enable HW_STACK_PROTECTION, DEBUG_OPTIMIZATIONS, and
CPU_STATS, you can get a crash when switching to the main thread
due to the r3 getting munged during a inline ASM call. This patch
fixes that by adding r3 to the clobber list.
Fixes#12821
Signed-off-by: Andy Gross <andy.gross@linaro.org>
This commit re-works the NXP MPU driver implementation so that
it aligns with the implementation for ARMv7-M and ARMv8-M MPU
architectures.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit removes obsolete ARM CORE MPU API definitions
and related implementation from arm_mpu.c, in the wake of
the transition to the new ARM MPU design.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit moves all internal implementation, relevant for
User mode, in a single place in the arm_mpu_v7_internal.h,
arm_mpu_v8_internal, and in arm_mpu.c. Additionally, the
commit cleans up internal function _get_region_attr() that
is not used any more.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit updates the ARM Core MPU API for memory domains,
to align with the principle of de-coupling the partitioning
and the access attribution with the architecture-specific
MPU driver implementation.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit adds the implementation of internal ARM MPU
function, arm_core_mpu_get_max_available_dyn_regions(),
required by _arch_mem_domain_max_partitions_get(),
according to the new architecture. The function returns
the number of the available partitions for memory domains
based on the total amount of HW regions and the
actual number of statically configured MPU regions.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit implements the internal (ARMv7-M and ARMv8-M)
API, which allows to user to (re)configure a memory partition
at run-time.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit contains the added functionality to program the
ARM8-M MPU so it can configure the dynamic memory regions at
run-time.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit contains the added functionality to program the
ARM7-M MPU so it can configure the dynamic memory regions at
run-time.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit introduces an ARM API that allows the user to
program a set of dynamic MPU regions at run-time. The API
function is invoked every time the memory map needs to be
re-programmed (for example at thread context-switch). The
functionality is implementated in arm_core_mpu.c.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit implements the common ARMv7-M and ARMv8-M internal
shim for configuring dynamic memory regions at run-time.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit contains the added functionality to program the
ARM8-M MPU, so it can configure the static memory regions at
boot time, as well as mark the areas, inside which dynamic MPU
region configuration at run-time will be allowed.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit contains the added functionality to program the
ARM7-M MPU, so it can configure the static memory regions at
boot time.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit removes the initialization of static memory regions
from arm_mpu_init(), since they are now initialized, directly,
by the kernel. The programming of MPU regions configured in SOC
definition is kept here.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit implements the common ARMv7-M and ARMv8-M internal
shim for configuring static memory regions at boot time. It also
adds LOG error messages, and ASSERTS, which fire if the static
MPU region configuration is not performed successfully.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit introduces the generic ARM (core) API, which allows
the user to program a set of static (fixed) MPU regions at boot
time. The API function is invoked upon initialization, in the
ARM-specific call of _arch_switch_to_main_thread(). The API
implementation is provided in arm_core_mpu.c.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit introduces a new Kconfig option symbol,
MPU_REQUIRES_NON_OVERLAPPING_REGIONS, to signify the use of an
MPU architecture that requires the active MPU regions be
non-overlapping. This requirement concerns the standard
(unmodified) ARMv8-M MPU implementation. For that, we select
the option in the definition of ARM_MPU, if the CPU architecture
is ARMv8-M (Baseline or Mainline).
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Fixes incorrect PR #12480
lp_count set to zero doesn't reset zero overhead loops,
it encodes a maximum loopcount.
Signed-off-by: Ruud Derwig <Ruud.Derwig@synopsys.com>
It's incorrect behavior to call IRQ_CONNECT() on the
same IRQ line more than once, but only x86 was catching
this.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
In 8dc69e09da (#10280)
the POSIX API main kconfig option was replaced from
PTHREAD_IPC to POSIX_API.
But the posix_cheats.h header was left using the old option.
This means that some applications/tests which were not
selecting the PTHREAD_IPC API, but using some other Zephyr
POSIX compatible APIs could have trouble when compiled for the
POSIX architecture.
Fixes#13011
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
In #9717, 777407b9ea
coverage support was broken for all NATIVE_APPLICATION except
native_posix
This includes the nrf52_bsim board
Fix it.
Fixes: #13009
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
Now that all supported xtensa boards use DTS we can move the Kconfig
setting to the arch level. Remove HAS_DTS from board Kconfig files.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
dts will now generate DT_SRAM_BASE_ADDRESS, DT_SRAM_SIZE,
DT_FLASH_BASE_ADDRESS, and DT_FLASH_SIZE defines. Kconfig can utilize
these defines to set defaults for the CONFIG_ variants.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
The Kconfig sybmol referenced in the comment is called
CONFIG_RISCV_SOC_CONTEXT_SAVE not CONFIG_RISCV_SOC_CONTEXT.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
cmsis provides intrinsic functions to configure processor special
function register like control, psp(process stack pointer). Let's use
these functions which are maintained in cmsis repository.
This patch uses cmsis function to:
- set process stack pointer(PSP)
- switch SP to PSP
- disable interrupts
It also moves stack initialization from assembly to C
Signed-off-by: Vikas Manocha <vikas.manocha@st.com>
This commit removes the local state variable nxp_mpu_enabled,
defined and used in NXP MPU driver, as it is not useful.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
The script looks for CONFIG_3RD_LEVEL_INTR_xx_OFFSET while
the config is actually CONFIG_3RD_LVL_INTR_xx_OFFSET.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Borrow from an Arm Cortex-M convention where each Kconfig.soc can
define a 'config WDOG_INIT' that does watchdog initialization early in
the boot process if that SoC needs it.
Some SoCs have watchdogs that are enabled by default and need to be
turned off during reset handling (to be re-enabled if necessary by a
Zephyr watchdog driver).
Signed-off-by: Marti Bolivar <marti@foundries.io>
A couple of follow-on patches suggested after previous RISCV32 arch
changes were merged.
Tweak some help in arch/riscv32/kconfig to better work with the RST
docs.
Take out all the CONFIG_PRINTK ifdeffery in fatal.c. The cause_str()
routine should get compiled out if PRINTK=n anyway.
Signed-off-by: Marti Bolivar <marti@foundries.io>
Some extensions to the multi-level interrupt controller are required
to support SoCs with more than four level 2 interrupt "aggregators".
Extend existing support to allow at most 8 level 2 or level 3
aggregators. Use Kconfig macro templates to cut down on boilerplate.
Try to clarify some aspects of the Kconfig help while we're at it, and
change the type of options which count things or are table offsets
from "hex" to "int", so that the generated .config is easier to read.
Finally, make some improvements to gen_isr_tables.py while we are
here. In particular, move some assignments around to cut down on
duplicated work, don't check for symbols we know must exist, and
improve the debug logging output's readability.
Signed-off-by: Marti Bolivar <marti@foundries.io>
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>
in the _rirq_return_from_coop, a fake irq stack frame is created.
But the value of lp_start, lp_end, lp_count are not set correctly.
When rtie is executed, the zero-overhead loop will be enabled, if
the pc crosses the value of lp_end and lp_count is not zero. The pc
will jump to lp_start. This will break down the system.
This commit fixs this bug and set the correct value of lp_start, lp_end
and lp_count(only need to set lp_count).
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
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>
This patch adds all the required hooks needed in the kernel to
get the coverage reports from ARM SoCs.
Signed-off-by: Adithya Baglody <adithya.nagaraj.baglody@intel.com>
This patch provides support for generating Code coverage reports.
The prj.conf needs to enable CONFIG_COVERAGE. Once enabled, the
code coverage data dump now comes via UART.
This data dump on the UART is triggered once the main
thread exits.
Next step is to save this data dump on file. Then run
scripts/gen_gcov_files.py with the serial console log as argument.
The last step would be be to run the gcovr. Use the following cmd
gcovr -r . --html -o gcov_report/coverage.html --html-details
Currently supported architectures are ARM and x86.
Signed-off-by: Adithya Baglody <adithya.nagaraj.baglody@intel.com>
This patch adds a required cast for the size calculation utilizing
pointer arithmetic. The stack_obj needs a cast to u32_t. This was
caught using a newer compiler.
Signed-off-by: Andy Gross <andy.gross@linaro.org>
On ARM, _Swap() isn't atomic and a hardware interrupt can land after
the (irq_locked) caller has entered _Swap() but before the context
switch actually happens. This will require some platform-specific
workarounds in a few places in the scheduler.
This commit is just the Kconfig and selection on ARM.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
This partially reverts commit 5a47c60dbf.
The soc.h is now only included when _soc_irq_*() is being referred.
Fixes#11077.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This patch adds a x86_64 architecture and qemu_x86_64 board to Zephyr.
Only the basic architecture support needed to run 64 bit code is
added; no drivers are added, though a low-level console exists and is
wired to printk().
The support is built on top of a "X86 underkernel" layer, which can be
built in isolation as a unit test on a Linux host.
Limitations:
+ Right now the SDK lacks an x86_64 toolchain. The build will fall
back to a host toolchain if it finds no cross compiler defined,
which is tested to work on gcc 8.2.1 right now.
+ No x87/SSE/AVX usage is allowed. This is a stronger limitation than
other architectures where the instructions work from one thread even
if the context switch code doesn't support it. We are passing
-no-sse to prevent gcc from automatically generating SSE
instructions for non-floating-point purposes, which has the side
effect of changing the ABI. Future work to handle the FPU registers
will need to be combined with an "application" ABI distinct from the
kernel one (or just to require USERSPACE).
+ Paging is enabled (it has to be in long mode), but is a 1:1 mapping
of all memory. No MMU/USERSPACE support yet.
+ We are building with -mno-red-zone for stack size reasons, but this
is a valuable optimization. Enabling it requires automatic stack
switching, which requires a TSS, which means it has to happen after
MMU support.
+ The OS runs in 64 bit mode, but for compatibility reasons is
compiled to the 32 bit "X32" ABI. So while the full 64 bit
registers and instruction set are available, C pointers are 32 bits
long and Zephyr is constrained to run in the bottom 4G of memory.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
It's worth using custom timing information on a few systems to save
cycles or gain precision. But make the use of k_cycle_get_32() a
proper default instead of hardcoding all the platforms and failing to
build on new ones. On Xtensa and RISC-V (and now x86_64) the cycle
informatoin from that call is a very fast wrapper around the native
counters anyway -- all you would save would be the function call
overhead.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
This function returns an essential boolean type, just changing the
signature to return a boolean type.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
Don't present USE_SWITCH and SMP to user applications that are
configuring for platforms that do not support SMP or USE_SWITCH.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
As this was written before, a 'FLAOT' menu entry was present, and to
see what kind of features or configuration it had, one had to navigate
into the menu to determine if anything related to 'FLOAT' was enabled
at all.
Now there is an unchecked box on the 'FLOAT' menu when it is disabled,
which clearly expresses that it is disabled without the need to enter
the menu entry.
This change only changes the menu layout, and not the dependencies of
any options.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Fix misspellings in documentation (.rst, Kconfig help text, and .h
doxygen API comments), missed during regular reviews.
Signed-off-by: David B. Kinder <david.b.kinder@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>
According with MISRA-C the value returned by a non-void function has
to be used. As memcpy return is almost useless, we are explicitly
ignoring it.
MISRA-C rule 17.7
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
This commit removes the priv_stack_size field from the _thread_arch
on arc architecture as there is no code using value stored in this
variable.
Signed-off-by: Piotr Zięcik <piotr.ziecik@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>
This commit removes the priv_stack_size field from the _thread_arch
on arm architecture as there is no code using value stored in this
variable.
Signed-off-by: Piotr Zięcik <piotr.ziecik@nordicsemi.no>
This commit removes the custom_data field from _thread_arch
for xtensa platform as it is currently unused.
Signed-off-by: Piotr Zięcik <piotr.ziecik@nordicsemi.no>
This reverts commit 140863f6a7.
This was found to be causing problems with certain linkers which
generate different code depending on whether a symbol is weak or
not.
Fixes#11916
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
When code relocation feature with userspace mode is turned on we
need a bit more memory to fit the text section for these 2
generated files.
Signed-off-by: Adithya Baglody <adithya.nagaraj.baglody@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>
arm_core_mpu.h and arm_core_mpu.c defined and implement kernel
APIs for memory protection, respectively. Therefore, they do not
need to directly include ARM CMSIS headers, or arm_mpu.h (or
nxp_mpu.h) which are supposed to define MPU-related kernel types
and convenience macros for the specific MPU architecture. These
headers are indirectly included by including kernel.h.
Similarly, arm_mpu.h shall not need to include internal/external
headers of memory protection APIs.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
This commit does the following:
- it introduces additional convenience macros for representing
MPU attributions for no-cacheability, in both ARMv7-M and
ARMv8-M MPU architectures,
- it adds documentation in K_MEM_PARTITION_IS_WRITABLE/CACHEABLE
macros in all macro definitions in the different MPU variants
- it moves the type definition of k_mem_partition_attr_t inside
the architecture-specific MPU headers, so it can be defined
per-architecture. It generalizes app_mem_domain.h, to be able
to work with _any_ (struct) type of k_mem_partition_attr_t.
- it refactors the type of k_mem_partition_attr_t for ARMv8-M
to comply with the MPU register API.
- for NXP MPU, the commit moves the macros for region access
permissions' attributes inside nxp_mpu.h, to align with what
we do for ARM MPU.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
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>
RISC-V permits myriad extensions to the ISA, any of which may imply
additional context that must be saved and restored on ISR entry and
exit. The current in-tree example is the Pulpino core, which has extra
registers used by ISA extensions for running loops that shouldn't get
clobbered by an ISR.
This is currently supported by including pulpino-specific definitions
in the generic architecture code. This works, but it's a bit inelegant
and is something of a layering violation. A more generic mechanism is
required to support other RISC-V SoCs with similar requirements
without cluttering the arch code too much.
Provide that by extending the semantics of the existing
CONFIG_RISCV_SOC_CONTEXT_SAVE option to allow other SoCs to allocate
space for saving and restoring their own state, promoting the
currently pulpino-specific __soc_save_context / __soc_restore_context
routines to a RISC-V arch API.
The cost of making this generic is two more instructions in each ISR
to pass the SoC specific context to these routines in a0 rather than
just assuming the stack points to the right place. This is minimal,
and should have been done anyway to keep with the ABI.
As a first (and currently only in-tree) customer, convert the Pulpino
SoC code to this new mechanism.
Signed-off-by: Marti Bolivar <marti@foundries.io>
The way that CONFIG_EXECUTION_BENCHMARKING=y is handled on this
architecture is incorrect. The goals are:
- call read_timer_start_of_isr() as close as possible to the
beginning of the ISR
- call read_timer_end_of_isr() after all preparations have
been made to call the driver-level IRQ handler, but it hasn't
been called yet
The current implementation could cause kernel crashes, though.
The read_timer_start_of_isr() call is made before saving MEPC or any
SoC-specific context. The MEPC issue is not that big of a deal, but
doing it before saving SoC context could clobber state that hasn't
been saved yet and corrupt the kernel.
One example is a pulpino style RISC-V SoC. Some Pulpino cores have
extra registers that are used for ISA extensions used to generate code
for C loops. There's no guarantee read_timer_start_of_isr() will never
have a loop inside: in fact, the RISC-V User-Level ISA v2.2 explicitly
recommends using a loop to get the 64-bit value of the "cycle" CSR. A
Pulpino-like SoC with a cycle CSR could thus naturally have a
read_timer_start_of_isr() implementation that involves loops. Saving
the loop state before reading the timer would then be needed.
Fix this issue by moving the call to read_timer_start_of_isr to after
all context saving is done. (This is a fairer comparison to Arm
Cortex-M SoCs anyway, since register stacking is performed in hardware
on Cortex M and is done before the first ISR instruction executes.)
The call to read_timer_end_of_isr() has an issue as well: it's called
after the ISR's argument has been stored in a0 and the ISR address is
in t1, but before actually calling the ISR itself.
In the standard RV32I calling convention, both t1 and a0 are caller
saved, so read_timer_end_of_isr() is within its rights to set them to
any garbage, which we'll happily treat as a function and its argument
and execute.
Avoid that possibility by saving the register values to the stack in
this configuration.
Signed-off-by: Marti Bolivar <marti@foundries.io>