Nicolas Pitre points out that since these thread structs are just
dummies for the context swtiching, they can be presumed to be "write
only" and thus there's no point in having one per CPU, everyone can
share the same one.
The only gotcha is that we never really documented (nor really have a
place to document) that rule, so it's not theoretically impossible for
an architecture to read back what it might have written underneath
arch_switch(). Leave this in a separate commit for bisection
purposes, but the risk seems very low.
Signed-off-by: Andy Ross <andyross@google.com>
After a k_thread_abort(), the resulting thread struct is documented as
unused/free memory that may be re-used (for example, to respawn a new
thread).
But in the special case of aborting the current thread from within an
ISR, that wasn't quite happening. The scheduler cleanup would
complete, but the architecture layer would still try to context switch
away from the aborted thread on exit, and that can include writes to
the now-reused thread struct! The specifics will depend on
architecture (some do a full context save on entry, most don't), but
in the case of USE_SWITCH=y it will at the very least write the
switch_handle field.
Fix this simply, with a per-cpu "switch dummy" thread struct for use
as a target for context switches like this. There is some non-trivial
memory cost to that; thread structs on many architectures are large.
Pleasingly, this also addresses a known deadlock on SMP: because the
"spin in ISR" step now happens as the very last stage of
k_thread_abort() handling, the existing scheduler lock works to
serialize calls such that it's impossible for a cycle of threads to
independently decide to spin on each other: at least one will see
itself as "already aborting" and break the cycle.
Fixes#64646
Signed-off-by: Andy Ross <andyross@google.com>
This reverts commit 61c70626a5.
This PR introduced 2 regressions in main CI:
71977 & 71978
Let's revert it by now to get main's CI passing again.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
This reverts commit 20611f13ca.
This PR introduced 2 regressions in main CI:
71977 & 71978
Let's revert it by now to get main's CI passing again.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Nicolas Pitre points out that since these thread structs are just
dummies for the context swtiching, they can be presumed to be "write
only" and thus there's no point in having one per CPU, everyone can
share the same one.
The only gotcha is that we never really documented (nor really have a
place to document) that rule, so it's not theoretically impossible for
an architecture to read back what it might have written underneath
arch_switch(). Leave this in a separate commit for bisection
purposes, but the risk seems very low.
Signed-off-by: Andy Ross <andyross@google.com>
After a k_thread_abort(), the resulting thread struct is documented as
unused/free memory that may be re-used (for example, to respawn a new
thread).
But in the special case of aborting the current thread from within an
ISR, that wasn't quite happening. The scheduler cleanup would
complete, but the architecture layer would still try to context switch
away from the aborted thread on exit, and that can include writes to
the now-reused thread struct! The specifics will depend on
architecture (some do a full context save on entry, most don't), but
in the case of USE_SWITCH=y it will at the very least write the
switch_handle field.
Fix this simply, with a per-cpu "switch dummy" thread struct for use
as a target for context switches like this. There is some non-trivial
memory cost to that; thread structs on many architectures are large.
Pleasingly, this also addresses a known deadlock on SMP: because the
"spin in ISR" step now happens as the very last stage of
k_thread_abort() handling, the existing scheduler lock works to
serialize calls such that it's impossible for a cycle of threads to
independently decide to spin on each other: at least one will see
itself as "already aborting" and break the cycle.
Fixes#64646
Signed-off-by: Andy Ross <andyross@google.com>
Add a closing comment to the endif with the configuration
information to which the endif belongs too.
To make the code more clearer if the configs need adaptions.
Signed-off-by: Simon Hein <Shein@baumer.com>
cpu_start_fn is global, it's used by the initiator CPU to start or
resume secondary CPUs. However it's possible, that the initiator CPU
goes ahead and starts a second secondary CPU before the first one has
finished using the object. Fix this by creating a local copy of the
global object.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
It is observed that starting up CPU may result in other CPUs
crashing due to de-referencing NULL pointers. Note that this
happened on the up_squared board, but there was no way to
attach debugger to verify. It started working again after
moving z_dummy_thread_init() before smp_timer_init(), which
was the old behavior before commit
eefaeee061 where the issue
first appeared. So mimic the old behavior to workaround
the issue.
Fixes#68115
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This removes z_smp_thread_init() and z_smp_thread_swap() as
SOF has been updated to use k_smp_cpu_custom_start() instead.
This removes the need for SOF to mirror part of the SMP
power up code, and thus these two functions are no longer
needed.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This provides a path to resume a previously suspended CPU.
This differs from k_smp_cpu_start() where per-CPU kernel
structs are not initialized such that execution context
can be saved during suspend and restored during resume.
Though the actual context saving and restoring are
platform specific.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This renames z_smp_cpu_start() to k_smp_cpu_start().
This effectively promotes z_smp_cpu_start() into a public API
which allows out of tree usage. Since this is a new API,
we can afford to change it signature, where it allows
an additional initialization steps to be done before a newly
powered up CPU starts participating in scheduling threads
to run.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Change for loops of the form:
for (i = 0; i < CONFIG_MP_NUM_CPUS; i++)
...
to
unsigned int num_cpus = arch_num_cpus();
for (i = 0; i < num_cpus; i++)
...
We do the call outside of the for loop so that it only happens once,
rather than on every iteration.
Signed-off-by: Kumar Gala <kumar.gala@intel.com>
MISRA C:2012 Rule 14.4 (The controlling expression of an if statement
and the controlling expression of an iteration-statement shall have
essentially Boolean type.)
Use `bool' instead of `int' to represent Boolean values.
Use `do { ... } while (false)' instead of `do { ... } while (0)'.
Use comparisons with zero instead of implicitly testing integers.
This commit is a subset of the original commit:
5d02614e34a86b549c7707d3d9f0984bc3a5f22a
Signed-off-by: Simon Hein <SHein@baumer.com>
In order to bring consistency in-tree, migrate all kernel code to the
new prefix <zephyr/...>. Note that the conversion has been scripted,
refer to zephyrproject-rtos#45388 for more details.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Things had gotten a little tangled in there so let's do some cleanup.
Remove the distressingly-special-purpose z_reinit_idle_thread() hook
(which existed to support secondary core bringup when
SMP_BOOT_DELAY=y), and just fold that into a generic z_init_cpu(),
which we can call in obvious and symmetric ways from main
initialization, z_smp_init(), and z_smp_start_cpu() (the now-official
programmatic hook for starting cpus).
Remove the "#if CONFIG_MP_NUM_CPUS > 1" exclusions. These weren't
saving any code size and were propagating themselves into platform
layers trying to avoid build failures.
There are some "special" APIs added for SOF which need to go away in
favor of the newer/generic z_smp_start_cpu(). Collect them in one
place and put them under a "#ifdef CONFIG_SOF" to prevent them from
being used in Zephyr apps.
Move some function declarations that didn't have homes into
<kernel/thread.h>.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Commit 3457118 changed the sequence of z_smp_thread_init() and
smp_timer_init() in smp_init_top() subroutine, which initializes
other BPs. In some boards (up_squared, acrn_ehl_crb) it will fail to
work while SMP enabling. If the timer interrupt is enabled before the
first thread was initialized. Now change back to its original order.
Fixes#41835
Signed-off-by: Enjia Mai <enjia.mai@intel.com>
This has bitrotten a bit. Early implementations had a synchronous
arch_start_cpu(), but then we started allowing that to be an async
operation. But that means that CPU start now becomes surprisingly
reentrant to the arch layer (cpu 0 can get a call to start cpu 2 while
cpu 1's initialization code is still running). That's just error
prone; we never documented the requirements cleanly (the window is
very small, but not so small to a slow simulator!).
Add an extra flag so we don't issue the next start until the last is
out of the arch layer and running in smp_init_top().
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Add a SOC API to allow for application control over deep idle power
states. Note that the hardware idle entry happens out of the WAITI
instruction, so the application has to be responsibile for ensuring
the CPU to be halted actually reaches idle deterministically. Lots of
warnings in the docs to this effect.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
smp.c only has to be built if CONFIG_SMP is enabled. Remove
preprocessor checks from the file itself and update cmake rules
instead.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Usually Zephyr boots all secondary CPUs as a part of system
boot. Some applications however need an ability to boot on
the main CPU only and enable secondary CPUs selectively at
run-time. Add a Kconfig option to support this behaviour.
When booting CPUs on demand applications also need helpers
to initialise a dummy thread and begin threaded execution
on those CPUs, add two such helpers.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
The internal function z_smp_reacquire_global_lock() has not used by
anywhere inside zephyr code, so remove it.
Fixes#33273.
Signed-off-by: Enjia Mai <enjiax.mai@intel.com>
1. Exclude the CODE UNREACHABLE line while generating coverage report.
2. Exclude the memory domain deprecated API when calculating code
coverage.
Signed-off-by: Enjia Mai <enjiax.mai@intel.com>
MISRA-C Rule 5.3 states that identifiers in inner scope should
not hide identifiers in outer scope.
In the function smp_init_top(), the variable "start_flags"
collide with global variable of the same name. So rename the one
inside the function.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
- simplify dummy thread initialization to a kswap.h
inline function
- use the same inline function for both early boot and
SMP setup
- add a note on necessity of the dummy thread even if
a custom swap to main is implemented
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
The set of interrupt stacks is now expressed as an array. We
also define the idle threads and their associated stacks this
way. This allows for iteration in cases where we have multiple
CPUs.
There is now a centralized declaration in kernel_internal.h.
On uniprocessor systems, z_interrupt_stacks has one element
and can be used in the same way as _interrupt_stack.
The IRQ stack for CPU 0 is now set in init.c instead of in
arch code.
The extern definition of the main thread stack is now removed,
this doesn't need to be in a header.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Use of the _current_cpu pointer cannot be done safely in a preemptible
context. If a thread is preempted and migrates to another CPU, the
old CPU record will be wrong.
Add a validation assert to the expression that catches incorrect
usages, and fix up the spots where it was wrong (most important being
a few uses of _current outside of locks, and the arch_is_in_isr()
implementation).
Note that the resulting _current expression now requires locking and
is going to be somewhat slower. Longer term it's going to be better
to augment the arch API to allow SMP architectures to implement a
faster "get current thread pointer" action than this default.
Note also that this change means that "_current" is no longer
expressible as an lvalue (long ago, it was just a static variable), so
the places where it gets assigned now assign to _current_cpu->current
instead.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Promote the private z_arch_* namespace, which specifies
the interface between the core kernel and the
architecture code, to a new top-level namespace named
arch_*.
This allows our documentation generation to create
online documentation for this set of interfaces,
and this set of interfaces is worth treating in a
more formal way anyway.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
This name collides with one in the bt subsystem, and wasn't named in
proper zephyrese anyway.
Fixes#16604
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>
Tickless timers mean that k_busy_wait() won't work until after the
timer driver is initialized, which is very early but not as early as
SMP. No need for it, just spin.
Also the original code used a stack variable for the start flag, which
racily presumed that _arch_start_cpu() would comes back synchronously
with the other CPU fired up and running right now. The cleaned up smp
bringup API on x86_64 isn't so perky, so it exposed the bug. The flag
just needs to be static.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
When CONFIG_SMP is enabled but CONFIG_MP_NUM_CPUS is 1 (which is a
legal configuration, though a weird one) this static function ends up
being defined but unused, producing a compiler warning.
Signed-off-by: Andy Ross <andrew.j.ross@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>
These functions, for good design reason, take a locking key to
atomically release along with the context swtich. But there's still a
common pattern in code to do a switch unconditionally by passing
irq_lock() directly. On SMP that's a little hurtful as it spams the
global lock. Provide an _unlocked() variant for
_Swap/_reschedule/_pend_curr for simplicity and efficiency.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Checking the return of some scattered functions across kernel.
MISRA-C requires that all non-void functions have their return value
checked, though, in some cases there is nothing to do. Just
acknowledging it.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
irq_lock returns an unsigned int, though, several places was using
signed int. This commit fix this behaviour.
In order to avoid this error happens again, a coccinelle script was
added and can be used to check violations.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
This was wrong in two ways, one subtle and one awful.
The subtle problem was that the IRQ lock isn't actually globally
recursive, it gets reset when you context switch (i.e. a _Swap()
implicitly releases and reacquires it). So the recursive count I was
keeping needs to be per-thread or else we risk deadlock any time we
swap away from a thread holding the lock.
And because part of my brain apparently knew this, there was an
"optimization" in the code that tested the current count vs. zero
outside the lock, on the argument that if it was non-zero we must
already hold the lock. Which would be true of a per-thread counter,
but NOT a global one: the other CPU may be holding that lock, and this
test will tell you *you* do. The upshot is that a recursive
irq_lock() would almost always SUCCEED INCORRECTLY when there was lock
contention. That this didn't break more things is amazing to me.
The rework is actually simpler than the original, thankfully. Though
there are some further subtleties:
* The lock state implied by irq_lock() allows the lock to be
implicitly released on context switch (i.e. you can _Swap() with the
lock held at a recursion level higher than 1, which needs to allow
other processes to run). So return paths into threads from _Swap()
and interrupt/exception exit need to check and restore the global
lock state, spinning as needed.
* The idle loop design specifies a k_cpu_idle() function that is on
common architectures expected to enable interrupts (for obvious
reasons), but there is no place to put non-arch code to wire it into
the global lock accounting. So on SMP, even CPU0 needs to use the
"dumb" spinning idle loop.
Finally this patch contains a simple bugfix too, found by inspection:
the interrupt return code used when CONFIG_SWITCH is enabled wasn't
correctly setting the active flag on the threads, opening up the
potential for a race that might result in a thread being scheduled on
two CPUs simultaneously.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Names that begin with an underscore are reserved by the C standard.
This patch does not change names of functions defined and implemented
in header files.
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
In SMP, the system timer is used for timeslicing on auxiliary CPUs,
but the base system timekeeping via _nano_sys_clock_tick_announce() is
still done on CPU0 only (because the framework isn't prepared for
asynchronous notification yet). Skip processing on CPU1+.
Also, due to a hardware interaction* that is difficult to work around,
timer initialization on the auxiliary CPUs is done at the very end of
the CPU bringup, just before the swap into the scheduler. A
smp_timer_init() API has been added for this purpose.
* On ESP-32, enabling the timer seems to result in a near-synchronous
interrupt being delivered despite my best attempts to keep it
masked, then blowing things up because the CPU record isn't set up
to handle it yet.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Now that all the pieces are in place, enable SMP for real:
Initialize the CPU records, launch the CPUs at the end of kernel
initialization, have them wait for a flag to release them into the
scheduler, then enter into the runnable threads via _Swap().
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
In SMP mode, the idea of a single "IRQ lock" goes away. Long term,
all usage needs to migrate to spinlocks (which become simple IRQ locks
in the uniprocessor case). For the near term, we can ease the
migration (at the expense of performance) by providing a compatibility
implementation around a single global lock.
Note that one complication is that the older lock was recursive, while
spinlocks will deadlock if you try to lock them twice. So we
implement a simple "count" semantic to handle multiple locks.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
