Fix an issue where, if a task is pending on a nano timeout, the duration
it wants to wait is not taken into account by the tickless idle code.
This could cause a system to wait forever, or to the limit of the timer
hardware (which is forever, for all intents and purposes).
This fix is to add one field in the nanokernel data structure for one
task to record the amount of ticks it will wait on a nano timeout. Only
one task has to be able to record this information, since, these waits
being looping busy waits, the task of highest priority is the only task
that can be actively waiting with a nano timeout. If a task of lower
priority was previously waiting, and a new task is now waiting, it means
that the wait of the original task has been interrupted, which will
cause said task to run the busy loop on the object again when it gets
scheduled, and the number of ticks it wants to wait has to be recomputed
and recorded again.
Change-Id: Ibcf0f288fc42d96897642cfee00ab7359716703f
Signed-off-by: Benjamin Walsh <benjamin.walsh@windriver.com>
Changes the nanokernel stack API so that the timeout parameter must be
specified when invoking nano_isr_stack_pop(), nano_fiber_stack_pop(),
nano_task_stack_pop() and nano_stack_pop().
This obsoletes the following APIs:
nano_fiber_stack_pop_wait()
nano_task_stack_pop_wait()
nano_stack_pop_wait()
Note that even though the new API requires that the timeout parameter
be specified, there are currently only two acceptable values:
TICKS_NONE and TICKS_UNLIMITED
This nanokernel option does not support CONFIG_NANO_TIMEOUTS.
Change-Id: Ic7f16ee30c3534115ceffa19ef8591ecc5a79080
Signed-off-by: Peter Mitsis <peter.mitsis@windriver.com>
Changes the nanokernel LIFO API so that the timeout parameter must be
specified when invoking nano_isr_lifo_get(), nano_fiber_lifo_get(),
nano_task_lifo_get() and nano_lifo_get().
This obsoletes the following APIs:
nano_fiber_lifo_get_wait()
nano_fiber_lifo_get_wait_timeout()
nano_task_lifo_get_wait()
nano_task_lifo_get_wait_timeout()
nano_lifo_get_wait()
nano_lifo_get_wait_timeout()
Change-Id: Ie9f93e46da42ea33c32544c02ab1d70b893cc198
Signed-off-by: Peter Mitsis <peter.mitsis@windriver.com>
Adds the following generic lifo routines:
nano_lifo_put()
nano_lifo_get()
nano_lifo_get_wait()
nano_lifo_get_wait_timeout()
Those routines are convenience wrappers for invoking the task, fiber and
ISR (if applicable) specific implementations.
Change-Id: I5252e4643fe4772f1309b26c1b3e4319f5035956
Signed-off-by: Peter Mitsis <peter.mitsis@windriver.com>
The \NOMANUAL tag is a remnant from days of yore and is no longer
needed or useful. Cleaning up the code references to this.
Change-Id: I1b8cc9c9560d1dbb711f05fa63fd23386789875c
Signed-off-by: Dan Kalowsky <daniel.kalowsky@intel.com>
Add a list for each kind of nanokernel object. And link the object
to the list on the initialization function.
Change-Id: I7a51902a8d1b306ec50d9486e55af286127fcb1a
Signed-off-by: Yonattan Louise <yonattan.a.louise.mendoza@intel.com>
Change all the Intel and Wind River code license from BSD-3 to Apache 2.
Change-Id: Id8be2c1c161a06ea8a0b9f38e17660e11dbb384b
Signed-off-by: Javier B Perez Hernandez <javier.b.perez.hernandez@linux.intel.com>
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Signed-off-by: Allan Stephens <allan.stephens@windriver.com>
Signed-off-by: Benjamin Walsh <benjamin.walsh@windriver.com>
The term 'context' is vague and overloaded. Its usage for 'an execution
context' is now referred as such, in both comments and some APIs' names.
When the execution context can only be a fiber or a task (i.e. not an
ISR), it is referred to as a 'thread', again in comments and everywhere
in the code.
APIs that had their names changed:
- nano_context_id_t is now nano_thread_id_t
- context_self_get() is now sys_thread_self_get()
- context_type_get() is now sys_execution_context_type_get()
- context_custom_data_set/get() are now
sys_thread_custom_data_set/get()
The 'context' prefix namespace does not have to be reserved by the
kernel anymore.
The Context Control Structure (CCS) data structure is now the Thread
Control Structure (TCS):
- struct ccs is now struct tcs
- tCCS is now tTCS
Change-Id: I7526a76c5b01e7c86333078e2d2e77c9feef5364
Signed-off-by: Benjamin Walsh <benjamin.walsh@windriver.com>
Standardize on using the irq_lock/irq_unlock (non-inline) symbols
everywhere.
The non-inline versions provide absolutely no benefits, so they will be
removed in a subsequent commit, and the inline versions will have their
_inline suffix removed.
Change-Id: Ib0b55f450447366468723e065a60adbadf7067a9
Signed-off-by: Benjamin Walsh <benjamin.walsh@windriver.com>
Moves the nanokernel LIFO in-code comments to the .h file. The comments
were changed to the Javadoc Doxygen format. A comment per alias was
created based on the original comment.
Change-Id: Iacd6bf1136ceac439d05669ef51cf28b858b306b
Signed-off-by: Rodrigo Caballero <rodrigo.caballero.abraham@intel.com>
Remove function name from comment and add @brief instead.
Also capitilize first letter.
Change-Id: Ib708b49bf02e5bc89b0066637a55874e659637e0
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Previous comment style used RETRURNS:, use @return to comply
with javadoc style.
Change-Id: Ib1dffd92da1d97d60063ec5309b08049828f6661
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
The change replaces multiple asterisks to ** at
the beginning of comments and adds a space before
the asterisks at the beginning of lines.
Change-Id: I7656bde3bf4d9a31e38941e43b580520432dabc1
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
The new name better reflects that this file contains all private
nanokernel APIs that are used by various kernel subsystems.
Change-Id: I4c258d582e93753eec9e575fdb5f9f2109417a0f
Signed-off-by: Allan Stephens <allan.stephens@windriver.com>
Add the _timeout() variants to the _get_wait() routines. Those allow
waiting with a timeout. The return values are the same as for the
regular _get() routines: a valid pointer on success, NULL on failure.
Note that the task_ variant is still polling, and the task is not put on
a timeout queue, instead polling for the ticker change in addition to
data arriving on the LIFO.
Change-Id: Iad8fa00b3fa0656fee9a67b2dd8b316e1e8aac74
Signed-off-by: Benjamin Walsh <benjamin.walsh@windriver.com>
Eliminates references to the obsolete OS name. In most cases the
name is simply removed, as it isn't necessary.
Change-Id: I32f9e7390e436aec008a9454b72657e129d65152
Signed-off-by: Allan Stephens <allan.stephens@windriver.com>
It is a fatal error if the microkernel is configured with either too few
command packets or too few timer packets.
NOTE: During this refactoring of _Cget(), not only is it renamed to
_nano_fiber_lifo_get_panic(), but it is moved into "nano_lifo.c".
Change-Id: I1d866cda1b444da04877f7eda03762b6e83c9a6f
Signed-off-by: Peter Mitsis <peter.mitsis@windriver.com>
These APIs no longer serve a useful purpose; there removal helps
reduce microkernel-related "pollution" of nanokernel files.
Change-Id: I3fb1df1feb47abf1b3e57656175ba6c80e191932
Signed-off-by: Allan Stephens <allan.stephens@windriver.com>
This notation is more concise and carry the same information.
Change-Id: Iba2023d37c2f3168fa44750441a29a6e024c0d7e
Signed-off-by: Benjamin Walsh <benjamin.walsh@windriver.com>
Append '_non_preemptible' to _fifo_put, _lifo_put, _sem_give and
_stack_push.
The original names do not convey the fact that these routines are
usable only by fibers and ISRs because they do not invoke a context
switch if they cause a fiber to become ready.
Change-Id: I8ddc24cbc7ebcbed42479ef84b04a74ef21d1647
Signed-off-by: Benjamin Walsh <benjamin.walsh@windriver.com>
State that a task 'giving' an object to a waiting fiber will be
preempted immediately. Remove useless _Swap() comments.
Change-Id: I78e9871acc046dc9b003f926d45b0b52ba48d2b4
Signed-off-by: Benjamin Walsh <benjamin.walsh@windriver.com>
The nanokernel LIFOs could only have one fiber pending on one of
them concurrently. If a second fiber wanted to take a contested
LIFO, if would kick the pending fiber out of the wait 'container',
but would not put it in back in the ready queue with an error. Instead,
it would, for all intents and purposes, remove it from the scheduling
mechanism of the nanokernel: the fiber would be 'lost' at that point.
The nanokernel LIFOs now make use of the fiber pend queue, and thus
allow multiple fibers to pend concurrently.
sysgen is updated since the microkernel initializes statically two
LIFOs, one for the command packets and one for the timer packets, and
the LIFO data structure has changed.
The nano_timers use the LIFOs and look at their internals, so they are
updated as well.
Change-Id: I250a610fcdee9d32172c88d38e6c3cfd5b437d15
Signed-off-by: Benjamin Walsh <benjamin.walsh@windriver.com>
The 'channel' wording was misleading.
Change-Id: I0f6f620e728d24988159cb80ee98edcaf7c52ec0
Signed-off-by: Benjamin Walsh <benjamin.walsh@windriver.com>
Moves all files in this sub-directory up one level, to the main
nanokernel source directory, then deletes the sub-directory itself.
This change means that all nanokernel source files (with the
excecption of its include files) are now in the main nanokernel
source directory, which is a more intuitive arrangement.
Change-Id: I6bd4b1bb6fdd9be13d57b320947f520a5bc73914
Signed-off-by: Allan Stephens <allan.stephens@windriver.com>
