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>
There are several subsystems and boards which require a relatively large
system heap (used by k_malloc()) to function properly. This became even
more notable with the recent introduction of the ACPICA library, which
causes ACPI-using boards to require a system heap of up to several
megabytes in size.
Until now, subsystems and boards have tried to solve this by having
Kconfig overlays which modify the default value of HEAP_MEM_POOL_SIZE.
This works ok, except when applications start explicitly setting values
in their prj.conf files:
$ git grep CONFIG_HEAP_MEM_POOL_SIZE= tests samples|wc -l
157
The vast majority of values set by current sample or test applications
is much too small for subsystems like ACPI, which results in the
application not being able to run on such boards.
To solve this situation, we introduce support for subsystems to specify
their own custom system heap size requirement. Subsystems do
this by defining Kconfig options with the prefix HEAP_MEM_POOL_ADD_SIZE_.
The final value of the system heap is the sum of the custom
minimum requirements, or the value existing HEAP_MEM_POOL_SIZE option,
whichever is greater.
We also introduce a new HEAP_MEM_POOL_IGNORE_MIN Kconfig option which
applications can use to force a lower value than what subsystems have
specficied, however this behavior is disabled by default.
Whenever the minimum is greater than the requested value a CMake warning
will be issued in the build output.
This patch ends up modifying several places outside of kernel code,
since the presence of the system heap is no longer detected using a
non-zero CONFIG_HEAP_MEM_POOL_SIZE value, rather it's now detected using
a new K_HEAP_MEM_POOL_SIZE value that's evaluated at build.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The switch_handle field in the thread struct is used as an atomic flag
between CPUs in SMP, and has been known for a long time to technically
require memory barriers for correct operation. We have an API for
that now, so put them in:
* The code immediately before arch_switch() needs a write barrier to
ensure that thread state written by the scheduler is seen to happen
before the outgoing thread is flagged with a valid switch handle.
* The loop in z_sched_switch_spin() needs a read barrier at the end,
to make sure the calling context doesn't load state from before the
other CPU stored the switch handle.
Also, that same spot in switch_spin was spinning with interrupts held,
which means it needs a call to arch_spin_relax() to avoid a FPU state
deadlock on some architectures.
Signed-off-by: Andy Ross <andyross@google.com>
This trick turns out also to be needed by the abort/join code.
Promote it to a more formal-looking internal API and clean up the
documentation to (hopefully) clarify the exact behavior and better
explain the need.
This is one of the more... enchanted bits of the scheduler, and while
the trick is IMHO pretty clean, it remains a big SMP footgun.
Signed-off-by: Andy Ross <andyross@google.com>
Fixes#46324
Set dummy_thread->base.slice_ticks to 0 when
CONFIG_TIMESLICE_PER_THREAD is set. To avoid
_current_cpu->slice_ticks be a big number.
Signed-off-by: Hu Zhenyu <zhenyu.hu@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:
5d02614e34
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>
Zephyr's timeslice implementation has always been somewhat primitive.
You get a global timeslice that applies broadly to the whole bottom of
the priority space, with no ability (beyond that one priority
threshold) to tune it to work on certain threads, etc...
This adds an (optionally configurable) API that allows timeslicing to
be controlled on a per-thread basis: any thread at any priority can be
set to timeslice, for a configurable per-thread slice time, and at the
end of its slice a callback can be provided that can take action.
This allows the application to implement things like responsiveness
heuristics, "fair" scheduling algorithms, etc... without requiring
that facility in the core kernel.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The resource pool of the short-lived dummy thread "stub" may be
inherited by other threads created during system initialization. This
commit initializes this resource pool to NULL or the system pool to
ensure that a well-defined resource pool propagates to other threads
that inherit it from the dummy thread.
Fixes#41482.
Signed-off-by: Berend Ozceri <berend@recogni.com>
This is an alternate backend that does what THREAD_RUNTIME_STATS is
doing currently, but with a few advantages:
* Correctly synchronized: you can't race against a running thread
(potentially on another CPU!) while querying its usage.
* Realtime results: you get the right answer always, up to timer
precision, even if a thread has been running for a while
uninterrupted and hasn't updated its total.
* Portable, no need for per-architecture code at all for the simple
case. (It leverages the USE_SWITCH layer to do this, so won't work
on older architectures)
* Faster/smaller: minimizes use of 64 bit math; lower overhead in
thread struct (keeps the scratch "started" time in the CPU struct
instead). One 64 bit counter per thread and a 32 bit scratch
register in the CPU struct.
* Standalone. It's a core (but optional) scheduler feature, no
dependence on para-kernel configuration like the tracing
infrastructure.
* More precise: allows architectures to optionally call a trivial
zero-argument/no-result cdecl function out of interrupt entry to
avoid accounting for ISR runtime in thread totals. No configuration
needed here, if it's called then you get proper ISR accounting, and
if not you don't.
For right now, pending unification, it's added side-by-side with the
older API and left as a z_*() internal symbol.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
There was a brief (but seen in practice on real apps on real
hardware!) race with the switch-based z_swap() implementation. The
thread return value was being initialized to -EAGAIN after the
enclosing lock had been released. But that lock is supposed to be
atomic with the thread suspend.
This opened a window for another racing thread to come by and "wake
up" our pending thread (which is fine on its own), set its return
value (e.g. to 0 for success) and then have that value clobbered by
the thread continuing to suspend itself outside the lock.
Melodramatic aside: I continue to hate this
arch_thread_return_value_set() API; it needs to die. At best it's a
mild optimization on a handful of architectures (e.g. x86 implements
it by writing to the EAX register save slot in the context block).
Asynchronous APIs are almost always worse than synchronous ones, and
in this case it's an async operation that races against literal
context switch code that can't use traditional locking strategies.
Fixes#39575
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Our z_swap() API takes a key returned from arch_irq_lock() and
releases it atomically with the context switch. Make sure that the
action of the unlocking is to unmask interrupts globally. If
interrupts would still be masked then that means there is an OUTER
interrupt lock still held, and the code that locked it surely doesn't
expect the thread to be suspended and interrupts unmasked while it's
held!
Unfortunately, this kind of mistake is very easy to make. We should
catch that with a simple assertion. This is essentially a crude
Zephyr equivalent of the extremely common "BUG: scheduling while
atomic" error in Linux drivers (just google it).
The one exception made is the circumstance where a thread has already
aborted itself. At that stage, whatever upthread lock state might
have existed will have already been messed up, so there's no value in
our asserting here. We can't catch all bugs, and this can actually
happen in error handling and/or test frameworks.
Fixes#33319
Signed-off-by: Andy Ross <andrew.j.ross@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>
Commit 6b84ab3830 ("kernel/sched: Adjust locking in z_swap()") moved
the call to arch_cohere_stacks() out of the scheduler lock while doing
some reorgnizing. On further reflection, this is incorrect. When
done outside the lock, the two arch_cohere_stacks() calls will race
against each other.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Swap was originally written to use the scheduler lock just to select a
new thread, but it would be nice to be able to rely on scheduler
atomicity later in the process (in particular it would be nice if the
assignment to cpu.current could be seen atomically). Rework the code
a bit so that swap takes the lock itself and holds it until just
before the call to arch_switch().
Note that the local interrupt mask has always been required to be held
across the swap, so extending the lock here has no effect on latency
at all on uniprocessor setups, and even on SMP only affects average
latency and not worst case.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
It was possible with pathological timing (see below) for the scheduler
to pick a cycle of threads on each CPU and enter the context switch
path on all of them simultaneously.
Example:
* CPU0 is idle, CPU1 is running thread A
* CPU1 makes high priority thread B runnable
* CPU1 reaches a schedule point (or returns from an interrupt) and
decides to run thread B instead
* CPU0 simultaneously takes its IPI and returns, selecting thread A
Now both CPUs enter wait_for_switch() to spin, waiting for the context
switch code on the other thread to finish and mark the thread
runnable. So we have a deadlock, each CPU is spinning waiting for the
other!
Actually, in practice this seems not to happen on existing hardware
platforms, it's only exercisable in emulation. The reason is that the
hardware IPI time is much faster than the software paths required to
reach a schedule point or interrupt exit, so CPU1 always selects the
newly scheduled thread and no deadlock appears. I tried for a bit to
make this happen with a cycle of three threads, but it's complicated
to get right and I still couldn't get the timing to hit correctly. In
qemu, though, the IPI is implemented as a Unix signal sent to the
thread running the other CPU, which is far slower and opens the window
to see this happen.
The solution is simple enough: don't store the _current thread in the
run queue until we are on the tail end of the context switch path,
after wait_for_switch() and going to reach the end in guaranteed time.
Note that this requires changing a little logic to handle the yield
case: because we can no longer rely on _current's position in the run
queue to suppress it, we need to do the priority comparison directly
based on the existing "swap_ok" flag (which has always meant
"yielded", and maybe should be renamed).
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The "null out the switch handle and put it back" code in the swap
implementation is a holdover from some defensive coding (not wanting
to break the case where we picked our current thread), but it hides a
subtle SMP race: when that field goes NULL, another CPU that may have
selected that thread (which is to say, our current thread) as its next
to run will be spinning on that to detect when the field goes
non-NULL. So it will get the signal to move on when we revert the
value, when clearly we are still running on the stack!
In practice this was found on x86 which poisons the switch context
such that it crashes instantly.
Instead, be firm about state and always set the switch handle of a
currently running thread to NULL immediately before it starts running:
right before entering arch_switch() and symmetrically on the interrupt
exit path.
Fixes#28105
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The z_swap_unlocked() function used a dummy spinlock for simplicity.
But this runs afouls of checking for stack-resident spinlocks
(forbidden when KERNEL_COHERENCE is set). And it's executing needless
code to release the lock anyway. Replace with a compile time NULL,
which will improve performance, correctness and code size.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Since the tracing of thread being switched in/out has the same
instrumentation points, we can roll the tracing function calls
into the one for thread stats gathering functions.
This avoids duplicating code to call another function.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This adds the bits to gather the first thread runtime statictic:
thread execution time. It provides a rough idea of how much time
a thread is spent in active execution. Currently it is not being
used, pending following commits where it combines with the trace
points on context switch as they instrument the same locations.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Zephyr SMP kernels need to be able to run on architectures with
incoherent caches. Naive implementation of synchronization on such
architectures requires extensive cache flushing (e.g. flush+invalidate
everything on every spin lock operation, flush on every unlock!) and
is a performance problem.
Instead, many of these systems will have access to separate "coherent"
(usually uncached) and "incoherent" regions of memory. Where this is
available, place all writable data sections by default into the
coherent region. An "__incoherent" attribute flag is defined for data
regions that are known to be CPU-local and which should use the cache.
By default, this is used for stack memory.
Stack memory will be incoherent by default, as by definition it is
local to its current thread. This requires special cache management
on context switch, so an arch API has been added for that.
Also, when enabled, add assertions to strategic places to ensure that
shared kernel data is indeed coherent. We check thread objects, the
_kernel struct, waitq's, timeouts and spinlocks. In practice almost
all kernel synchronization is built on top of these structures, and
any shared data structs will contain at least one of them.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
This code had one purpose only, feed timing information into a test and
was not used by anything else. The custom trace points unfortunatly were
not accurate and this test was delivering informatin that conflicted
with other tests we have due to placement of such trace points in the
architecture and kernel code.
For such measurements we are planning to use the tracing functionality
in a special mode that would be used for metrics without polluting the
architecture and kernel code with additional tracing and timing code.
Furthermore, much of the assembly code used had issues.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
* Move switched_in into the arch context switch assembly code,
which will correctly record the switched_in information.
* Add switched_in/switched_out for context switch in irq exit.
Signed-off-by: Watson Zeng <zhiwei@synopsys.com>
We make a policy change here: all threads are members of a
memory domain, never NULL. We introduce a default memory domain
for threads that haven't been assigned to or inherited another one.
Primary motivation for this change is better MMU support, as
one common configuration will be to maintain page tables at
the memory domain level.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Move tracing switched_in and switched_out to the architecture code and
remove duplications. This changes swap tracing for x86, xtensa.
Signed-off-by: Anas Nashif <anas.nashif@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>
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>
On SMP, there is an inherent race when swapping: the old thread adds
itself back to the run queue before calling into the arch layer to do
the context switch. The former is properly synchronized under the
scheduler lock, and the later operates with interrupts locally
disabled. But until somewhere in the middle of arch_switch(), the old
thread (that is in the run queue!) does not have complete saved state
that can be restored.
So it's possible for another CPU to grab a thread before it is saved
and try to restore its unsaved register contents (which are garbage --
typically whatever state it had at the last interrupt).
Fix this by leveraging the "swapped_from" pointer already passed to
arch_switch() as a synchronization primitive. When the switch
implementation writes the new handle value, we know the switch is
complete. Then we can wait for that in z_swap() and at interrupt
exit.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Improve positioning of tracing calls. Avoid multiple calls and missing
events because of complex logix. Trace the event where things happen
really.
Signed-off-by: Anas Nashif <anas.nashif@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>
In uniprocessor mode, the kernel knows when a context switch "is
coming" because of the cache optimization and can use that to do
things like update time slice state. But on SMP the scheduler state
may be updated on the other CPU at any time, so we don't know that a
switch is going to happen until the last minute.
Expose reset_time_slice() as a public function and call it when needed
out of z_swap().
Signed-off-by: Andy Ross <andrew.j.ross@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>
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>
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>
The call to _arch_switch is a giant screaming sign inviting optimizer
bugs. The code that appears before is what happened long ago when we
were switched out, but the version that EXECUTED just now is actually
in a different thread. So the assignment to _current before the
switch actually assigned OUR thread (the "new_thread" of the old
context!) to _current.
But obviously the optimizer looks at that code and assumes that the
_current which got assigned to the thread we were switching to long
ago is still correct, and used it when retrieving the swap return
value.
Obviously the real bug here is that the _arch_switch() in question
lacked a memory clobber (and it's getting one).
But we can remove two lines, remove code from inside the interrupt
lock and make the implementation more robust by moving the read to
after the irq_unlock() (which generally also has a memory clobber).
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Instead of checking every time we hit the low-level context switch
path to see if the new thread has a "partner" with which it needs to
share time, just run the slice timer always and reset it from the
scheduler at the points where it has already decided a switch needs to
happen. In TICKLESS_KERNEL situations, we pay the cost of extra timer
interrupts at ~10Hz or whatever, which is low (note also that this
kind of regular wakeup architecture is required on SMP anyway so the
scheduler can "notice" threads scheduled by other CPUs). Advantages:
1. Much simpler logic. Significantly smaller code. No variance or
dependence on tickless modes or timer driver (beyond setting a
simple timeout).
2. No arch-specific assembly integration with _Swap() needed
3. Better performance on many workloads, as the accounting now happens
at most once per timer interrupt (~5 Hz) and true rescheduling and
not on every unrelated context switch and interrupt return.
4. It's SMP-safe. The previous scheme kept the slice ticks as a
global variable, which was an unnoticed bug.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Any word started with underscore followed by and uppercase letter or a
second underscore is a reserved word according with C99.
With have *many* violations on Zephyr's code, this commit is tackling
only the violations caused by headers guards. It also takes the
opportunity to normalize them using the filename in uppercase and
replacing dot with underscore. e.g file.h -> FILE_H
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
__swap function was returning -EAGAIN in some case, though its return
value was declared as unsigned int.
This commit changes this function to return int since it can return a
negative value and its return was already been propagate as int.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
Move to more generic tracing hooks that can be implemented in different
ways and do not interfere with the kernel.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Define generic interface and hooks for tracing to replace
kernel_event_logger and existing tracing facilities with something more
common.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
This patch provides support needed to get timing related
information from xtensa based SOC.
Signed-off-by: Adithya Baglody <adithya.nagaraj.baglody@intel.com>
Recent changes post-scheduler-rewrite broke scheduling on SMP:
The "preempt_ok" feature added to isolate preemption points wasn't
honored in SMP mode. Fix this by adding a "swap_ok" field to the CPU
record (not the thread) which is set at the same time out of
update_cache().
The "queued" flag wasn't being maintained correctly when swapping away
from _current (it was added back to the queue, but the flag wasn't
set).
Abstract out a "should_preempt()" predicate so SMP and uniprocessor
paths share the same logic, which is distressingly subtle.
There were two places where _Swap() was predicated on
_get_next_ready_thread() != _current. That's no longer a benign
optimization in SMP, where the former function REMOVES the next thread
from the queue. Just call _Swap() directly in SMP, which has a
unified C implementation that does this test already. Don't change
other architectures in case it exposes bugs with _Swap() switching
back to the same thread (it should work, I just don't want to break
anything).
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
This replaces the existing scheduler (but not priority handling)
implementation with a somewhat simpler one. Behavior as to thread
selection does not change. New features:
+ Unifies SMP and uniprocessing selection code (with the sole
exception of the "cache" trick not being possible in SMP).
+ The old static multi-queue implementation is gone and has been
replaced with a build-time choice of either a "dumb" list
implementation (faster and significantly smaller for apps with only
a few threads) or a balanced tree queue which scales well to
arbitrary numbers of threads and priority levels. This is
controlled via the CONFIG_SCHED_DUMB kconfig variable.
+ The balanced tree implementation is usable symmetrically for the
wait_q abstraction, fixing a scalability glitch Zephyr had when many
threads were waiting on a single object. This can be selected via
CONFIG_WAITQ_FAST.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The SMP testing missed the case where _Swap() decides to return back
into the _current. Obviously there is no valid switch handle for the
running thread into which we can restore, and everything blows up.
(What happened is that the new scheduler code opened up a spot where
k_thread_priority_set() does a _reschedule() unconditionally and
doens't check to see whether or not it's needed like the old code).
But that isn't incorrect! It's entirely possible that _Swap() may
find that no thread is runnable except _current (due, for example, to
another CPU racing the other thread you expected off to sleep or
something). Don't blow up, check and return a noop.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Simon Hein