18069996051/arduino-libs
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

初始化提交

Browse Source
This commit is contained in:
王立帮
2024-07-20 22:09:06 +08:00
commit c247dd07a6
6876 changed files with 2743096 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

502
arduino-cli/libraries/CoopTask/LICENSE Normal file
View File

@@ -0,0 +1,502 @@
GNU LESSER GENERAL PUBLIC LICENSE
Version 2.1, February 1999
Copyright (C) 1991, 1999 Free Software Foundation, Inc.
51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed.
[This is the first released version of the Lesser GPL. It also counts
as the successor of the GNU Library Public License, version 2, hence
the version number 2.1.]
Preamble
The licenses for most software are designed to take away your
freedom to share and change it. By contrast, the GNU General Public
Licenses are intended to guarantee your freedom to share and change
free software--to make sure the software is free for all its users.
This license, the Lesser General Public License, applies to some
specially designated software packages--typically libraries--of the
Free Software Foundation and other authors who decide to use it. You
can use it too, but we suggest you first think carefully about whether
this license or the ordinary General Public License is the better
strategy to use in any particular case, based on the explanations below.
When we speak of free software, we are referring to freedom of use,
not price. Our General Public Licenses are designed to make sure that
you have the freedom to distribute copies of free software (and charge
for this service if you wish); that you receive source code or can get
it if you want it; that you can change the software and use pieces of
it in new free programs; and that you are informed that you can do
these things.
To protect your rights, we need to make restrictions that forbid
distributors to deny you these rights or to ask you to surrender these
rights. These restrictions translate to certain responsibilities for
you if you distribute copies of the library or if you modify it.
For example, if you distribute copies of the library, whether gratis
or for a fee, you must give the recipients all the rights that we gave
you. You must make sure that they, too, receive or can get the source
code. If you link other code with the library, you must provide
complete object files to the recipients, so that they can relink them
with the library after making changes to the library and recompiling
it. And you must show them these terms so they know their rights.
We protect your rights with a two-step method: (1) we copyright the
library, and (2) we offer you this license, which gives you legal
permission to copy, distribute and/or modify the library.
To protect each distributor, we want to make it very clear that
there is no warranty for the free library. Also, if the library is
modified by someone else and passed on, the recipients should know
that what they have is not the original version, so that the original
author's reputation will not be affected by problems that might be
introduced by others.
Finally, software patents pose a constant threat to the existence of
any free program. We wish to make sure that a company cannot
effectively restrict the users of a free program by obtaining a
restrictive license from a patent holder. Therefore, we insist that
any patent license obtained for a version of the library must be
consistent with the full freedom of use specified in this license.
Most GNU software, including some libraries, is covered by the
ordinary GNU General Public License. This license, the GNU Lesser
General Public License, applies to certain designated libraries, and
is quite different from the ordinary General Public License. We use
this license for certain libraries in order to permit linking those
libraries into non-free programs.
When a program is linked with a library, whether statically or using
a shared library, the combination of the two is legally speaking a
combined work, a derivative of the original library. The ordinary
General Public License therefore permits such linking only if the
entire combination fits its criteria of freedom. The Lesser General
Public License permits more lax criteria for linking other code with
the library.
We call this license the "Lesser" General Public License because it
does Less to protect the user's freedom than the ordinary General
Public License. It also provides other free software developers Less
of an advantage over competing non-free programs. These disadvantages
are the reason we use the ordinary General Public License for many
libraries. However, the Lesser license provides advantages in certain
special circumstances.
For example, on rare occasions, there may be a special need to
encourage the widest possible use of a certain library, so that it becomes
a de-facto standard. To achieve this, non-free programs must be
allowed to use the library. A more frequent case is that a free
library does the same job as widely used non-free libraries. In this
case, there is little to gain by limiting the free library to free
software only, so we use the Lesser General Public License.
In other cases, permission to use a particular library in non-free
programs enables a greater number of people to use a large body of
free software. For example, permission to use the GNU C Library in
non-free programs enables many more people to use the whole GNU
operating system, as well as its variant, the GNU/Linux operating
system.
Although the Lesser General Public License is Less protective of the
users' freedom, it does ensure that the user of a program that is
linked with the Library has the freedom and the wherewithal to run
that program using a modified version of the Library.
The precise terms and conditions for copying, distribution and
modification follow. Pay close attention to the difference between a
"work based on the library" and a "work that uses the library". The
former contains code derived from the library, whereas the latter must
be combined with the library in order to run.
GNU LESSER GENERAL PUBLIC LICENSE
TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
0. This License Agreement applies to any software library or other
program which contains a notice placed by the copyright holder or
other authorized party saying it may be distributed under the terms of
this Lesser General Public License (also called "this License").
Each licensee is addressed as "you".
A "library" means a collection of software functions and/or data
prepared so as to be conveniently linked with application programs
(which use some of those functions and data) to form executables.
The "Library", below, refers to any such software library or work
which has been distributed under these terms. A "work based on the
Library" means either the Library or any derivative work under
copyright law: that is to say, a work containing the Library or a
portion of it, either verbatim or with modifications and/or translated
straightforwardly into another language. (Hereinafter, translation is
included without limitation in the term "modification".)
"Source code" for a work means the preferred form of the work for
making modifications to it. For a library, complete source code means
all the source code for all modules it contains, plus any associated
interface definition files, plus the scripts used to control compilation
and installation of the library.
Activities other than copying, distribution and modification are not
covered by this License; they are outside its scope. The act of
running a program using the Library is not restricted, and output from
such a program is covered only if its contents constitute a work based
on the Library (independent of the use of the Library in a tool for
writing it). Whether that is true depends on what the Library does
and what the program that uses the Library does.
1. You may copy and distribute verbatim copies of the Library's
complete source code as you receive it, in any medium, provided that
you conspicuously and appropriately publish on each copy an
appropriate copyright notice and disclaimer of warranty; keep intact
all the notices that refer to this License and to the absence of any
warranty; and distribute a copy of this License along with the
Library.
You may charge a fee for the physical act of transferring a copy,
and you may at your option offer warranty protection in exchange for a
fee.
2. You may modify your copy or copies of the Library or any portion
of it, thus forming a work based on the Library, and copy and
distribute such modifications or work under the terms of Section 1
above, provided that you also meet all of these conditions:
a) The modified work must itself be a software library.
b) You must cause the files modified to carry prominent notices
stating that you changed the files and the date of any change.
c) You must cause the whole of the work to be licensed at no
charge to all third parties under the terms of this License.
d) If a facility in the modified Library refers to a function or a
table of data to be supplied by an application program that uses
the facility, other than as an argument passed when the facility
is invoked, then you must make a good faith effort to ensure that,
in the event an application does not supply such function or
table, the facility still operates, and performs whatever part of
its purpose remains meaningful.
(For example, a function in a library to compute square roots has
a purpose that is entirely well-defined independent of the
application. Therefore, Subsection 2d requires that any
application-supplied function or table used by this function must
be optional: if the application does not supply it, the square
root function must still compute square roots.)
These requirements apply to the modified work as a whole. If
identifiable sections of that work are not derived from the Library,
and can be reasonably considered independent and separate works in
themselves, then this License, and its terms, do not apply to those
sections when you distribute them as separate works. But when you
distribute the same sections as part of a whole which is a work based
on the Library, the distribution of the whole must be on the terms of
this License, whose permissions for other licensees extend to the
entire whole, and thus to each and every part regardless of who wrote
it.
Thus, it is not the intent of this section to claim rights or contest
your rights to work written entirely by you; rather, the intent is to
exercise the right to control the distribution of derivative or
collective works based on the Library.
In addition, mere aggregation of another work not based on the Library
with the Library (or with a work based on the Library) on a volume of
a storage or distribution medium does not bring the other work under
the scope of this License.
3. You may opt to apply the terms of the ordinary GNU General Public
License instead of this License to a given copy of the Library. To do
this, you must alter all the notices that refer to this License, so
that they refer to the ordinary GNU General Public License, version 2,
instead of to this License. (If a newer version than version 2 of the
ordinary GNU General Public License has appeared, then you can specify
that version instead if you wish.) Do not make any other change in
these notices.
Once this change is made in a given copy, it is irreversible for
that copy, so the ordinary GNU General Public License applies to all
subsequent copies and derivative works made from that copy.
This option is useful when you wish to copy part of the code of
the Library into a program that is not a library.
4. You may copy and distribute the Library (or a portion or
derivative of it, under Section 2) in object code or executable form
under the terms of Sections 1 and 2 above provided that you accompany
it with the complete corresponding machine-readable source code, which
must be distributed under the terms of Sections 1 and 2 above on a
medium customarily used for software interchange.
If distribution of object code is made by offering access to copy
from a designated place, then offering equivalent access to copy the
source code from the same place satisfies the requirement to
distribute the source code, even though third parties are not
compelled to copy the source along with the object code.
5. A program that contains no derivative of any portion of the
Library, but is designed to work with the Library by being compiled or
linked with it, is called a "work that uses the Library". Such a
work, in isolation, is not a derivative work of the Library, and
therefore falls outside the scope of this License.
However, linking a "work that uses the Library" with the Library
creates an executable that is a derivative of the Library (because it
contains portions of the Library), rather than a "work that uses the
library". The executable is therefore covered by this License.
Section 6 states terms for distribution of such executables.
When a "work that uses the Library" uses material from a header file
that is part of the Library, the object code for the work may be a
derivative work of the Library even though the source code is not.
Whether this is true is especially significant if the work can be
linked without the Library, or if the work is itself a library. The
threshold for this to be true is not precisely defined by law.
If such an object file uses only numerical parameters, data
structure layouts and accessors, and small macros and small inline
functions (ten lines or less in length), then the use of the object
file is unrestricted, regardless of whether it is legally a derivative
work. (Executables containing this object code plus portions of the
Library will still fall under Section 6.)
Otherwise, if the work is a derivative of the Library, you may
distribute the object code for the work under the terms of Section 6.
Any executables containing that work also fall under Section 6,
whether or not they are linked directly with the Library itself.
6. As an exception to the Sections above, you may also combine or
link a "work that uses the Library" with the Library to produce a
work containing portions of the Library, and distribute that work
under terms of your choice, provided that the terms permit
modification of the work for the customer's own use and reverse
engineering for debugging such modifications.
You must give prominent notice with each copy of the work that the
Library is used in it and that the Library and its use are covered by
this License. You must supply a copy of this License. If the work
during execution displays copyright notices, you must include the
copyright notice for the Library among them, as well as a reference
directing the user to the copy of this License. Also, you must do one
of these things:
a) Accompany the work with the complete corresponding
machine-readable source code for the Library including whatever
changes were used in the work (which must be distributed under
Sections 1 and 2 above); and, if the work is an executable linked
with the Library, with the complete machine-readable "work that
uses the Library", as object code and/or source code, so that the
user can modify the Library and then relink to produce a modified
executable containing the modified Library. (It is understood
that the user who changes the contents of definitions files in the
Library will not necessarily be able to recompile the application
to use the modified definitions.)
b) Use a suitable shared library mechanism for linking with the
Library. A suitable mechanism is one that (1) uses at run time a
copy of the library already present on the user's computer system,
rather than copying library functions into the executable, and (2)
will operate properly with a modified version of the library, if
the user installs one, as long as the modified version is
interface-compatible with the version that the work was made with.
c) Accompany the work with a written offer, valid for at
least three years, to give the same user the materials
specified in Subsection 6a, above, for a charge no more
than the cost of performing this distribution.
d) If distribution of the work is made by offering access to copy
from a designated place, offer equivalent access to copy the above
specified materials from the same place.
e) Verify that the user has already received a copy of these
materials or that you have already sent this user a copy.
For an executable, the required form of the "work that uses the
Library" must include any data and utility programs needed for
reproducing the executable from it. However, as a special exception,
the materials to be distributed need not include anything that is
normally distributed (in either source or binary form) with the major
components (compiler, kernel, and so on) of the operating system on
which the executable runs, unless that component itself accompanies
the executable.
It may happen that this requirement contradicts the license
restrictions of other proprietary libraries that do not normally
accompany the operating system. Such a contradiction means you cannot
use both them and the Library together in an executable that you
distribute.
7. You may place library facilities that are a work based on the
Library side-by-side in a single library together with other library
facilities not covered by this License, and distribute such a combined
library, provided that the separate distribution of the work based on
the Library and of the other library facilities is otherwise
permitted, and provided that you do these two things:
a) Accompany the combined library with a copy of the same work
based on the Library, uncombined with any other library
facilities. This must be distributed under the terms of the
Sections above.
b) Give prominent notice with the combined library of the fact
that part of it is a work based on the Library, and explaining
where to find the accompanying uncombined form of the same work.
8. You may not copy, modify, sublicense, link with, or distribute
the Library except as expressly provided under this License. Any
attempt otherwise to copy, modify, sublicense, link with, or
distribute the Library is void, and will automatically terminate your
rights under this License. However, parties who have received copies,
or rights, from you under this License will not have their licenses
terminated so long as such parties remain in full compliance.
9. You are not required to accept this License, since you have not
signed it. However, nothing else grants you permission to modify or
distribute the Library or its derivative works. These actions are
prohibited by law if you do not accept this License. Therefore, by
modifying or distributing the Library (or any work based on the
Library), you indicate your acceptance of this License to do so, and
all its terms and conditions for copying, distributing or modifying
the Library or works based on it.
10. Each time you redistribute the Library (or any work based on the
Library), the recipient automatically receives a license from the
original licensor to copy, distribute, link with or modify the Library
subject to these terms and conditions. You may not impose any further
restrictions on the recipients' exercise of the rights granted herein.
You are not responsible for enforcing compliance by third parties with
this License.
11. If, as a consequence of a court judgment or allegation of patent
infringement or for any other reason (not limited to patent issues),
conditions are imposed on you (whether by court order, agreement or
otherwise) that contradict the conditions of this License, they do not
excuse you from the conditions of this License. If you cannot
distribute so as to satisfy simultaneously your obligations under this
License and any other pertinent obligations, then as a consequence you
may not distribute the Library at all. For example, if a patent
license would not permit royalty-free redistribution of the Library by
all those who receive copies directly or indirectly through you, then
the only way you could satisfy both it and this License would be to
refrain entirely from distribution of the Library.
If any portion of this section is held invalid or unenforceable under any
particular circumstance, the balance of the section is intended to apply,
and the section as a whole is intended to apply in other circumstances.
It is not the purpose of this section to induce you to infringe any
patents or other property right claims or to contest validity of any
such claims; this section has the sole purpose of protecting the
integrity of the free software distribution system which is
implemented by public license practices. Many people have made
generous contributions to the wide range of software distributed
through that system in reliance on consistent application of that
system; it is up to the author/donor to decide if he or she is willing
to distribute software through any other system and a licensee cannot
impose that choice.
This section is intended to make thoroughly clear what is believed to
be a consequence of the rest of this License.
12. If the distribution and/or use of the Library is restricted in
certain countries either by patents or by copyrighted interfaces, the
original copyright holder who places the Library under this License may add
an explicit geographical distribution limitation excluding those countries,
so that distribution is permitted only in or among countries not thus
excluded. In such case, this License incorporates the limitation as if
written in the body of this License.
13. The Free Software Foundation may publish revised and/or new
versions of the Lesser General Public License from time to time.
Such new versions will be similar in spirit to the present version,
but may differ in detail to address new problems or concerns.
Each version is given a distinguishing version number. If the Library
specifies a version number of this License which applies to it and
"any later version", you have the option of following the terms and
conditions either of that version or of any later version published by
the Free Software Foundation. If the Library does not specify a
license version number, you may choose any version ever published by
the Free Software Foundation.
14. If you wish to incorporate parts of the Library into other free
programs whose distribution conditions are incompatible with these,
write to the author to ask for permission. For software which is
copyrighted by the Free Software Foundation, write to the Free
Software Foundation; we sometimes make exceptions for this. Our
decision will be guided by the two goals of preserving the free status
of all derivatives of our free software and of promoting the sharing
and reuse of software generally.
NO WARRANTY
15. BECAUSE THE LIBRARY IS LICENSED FREE OF CHARGE, THERE IS NO
WARRANTY FOR THE LIBRARY, TO THE EXTENT PERMITTED BY APPLICABLE LAW.
EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR
OTHER PARTIES PROVIDE THE LIBRARY "AS IS" WITHOUT WARRANTY OF ANY
KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE
LIBRARY IS WITH YOU. SHOULD THE LIBRARY PROVE DEFECTIVE, YOU ASSUME
THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
16. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN
WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY
AND/OR REDISTRIBUTE THE LIBRARY AS PERMITTED ABOVE, BE LIABLE TO YOU
FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR
CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE THE
LIBRARY (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA BEING
RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A
FAILURE OF THE LIBRARY TO OPERATE WITH ANY OTHER SOFTWARE), EVEN IF
SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
DAMAGES.
END OF TERMS AND CONDITIONS
How to Apply These Terms to Your New Libraries
If you develop a new library, and you want it to be of the greatest
possible use to the public, we recommend making it free software that
everyone can redistribute and change. You can do so by permitting
redistribution under these terms (or, alternatively, under the terms of the
ordinary General Public License).
To apply these terms, attach the following notices to the library. It is
safest to attach them to the start of each source file to most effectively
convey the exclusion of warranty; and each file should have at least the
"copyright" line and a pointer to where the full notice is found.
<one line to give the library's name and a brief idea of what it does.>
Copyright (C) <year> <name of author>
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
Also add information on how to contact you by electronic and paper mail.
You should also get your employer (if you work as a programmer) or your
school, if any, to sign a "copyright disclaimer" for the library, if
necessary. Here is a sample; alter the names:
Yoyodyne, Inc., hereby disclaims all copyright interest in the
library `Frob' (a library for tweaking knobs) written by James Random Hacker.
<signature of Ty Coon>, 1 April 1990
Ty Coon, President of Vice
That's all there is to it!

123
arduino-cli/libraries/CoopTask/README.md Normal file
View File

@@ -0,0 +1,123 @@
# Warning
In this release on ESP8266 scanning WiFi networks and WPS are both broken in CoopTask tasks.
Using these features directly from loop() is unaffected.
This warning will be silently removed once ESP8266 Git master has accepted the necessary PR.
# CoopTask
An all-C++ implementation of a cooperative multitasking layer for ESP8266/ESP32,
Arduino boards, Linux, and Windows x86 and x86_64
During regular development it's built and tested on the ESP MCUs and
Arduino Pro/Pro Mini.
Tasks in this scheduler are stackful coroutines. They act almost the same as
the main ``setup()``/``loop()`` code in Arduino sketches, but there can be many of them
simultaneously on the same device. It's even powerful enough to run the
ESP8266 and ESP32 WebServer in a CoopTask.
Use the normal global delay() function to suspend execution of a task for the
given number of milliseconds, use yield() to give up the CPU, both return after
other cooperative tasks have run awhile.
A simple blink task can be written just like this:
```
#include <CoopTask.h>
int loopBlink()
{
// like setup():
pinMode(LED_BUILTIN, OUTPUT);
// like loop():
for (;;)
{
digitalWrite(LED_BUILTIN, LOW);
delay(2000);
digitalWrite(LED_BUILTIN, HIGH);
delay(3000);
}
// tasks can return or exit() and leave an exit code
return 0;
}
CoopTask<>* taskBlink;
void setup()
{
Serial.begin(115200);
delay(500);
#if defined(ESP8266) || defined(ESP32)
taskBlink = createCoopTask(F("Blink"), loopBlink, 0x240);
#else
taskBlink = createCoopTask(F("Blink"), loopBlink, 0x40);
#endif
if (!taskBlink) Serial.println("CoopTask Blink not created");
}
void loop()
{
runCoopTasks();
}
```
The ``runCoopTasks()`` scheduling helper has two optional callback arguments.
The first, ``reaper``, gets called each time a task exits. Retrieving the
exit code or deleting the CoopTask object would typically be performed in a
task reaper function.
The second callback, ``onDelay``, is called after each scheduling rountrip with the
total minimum delay (can be zero) of all managed tasks. A use scenario for this
is to put the MCU into a power saving sleep mode for the given duration.
## Using Arduino or Linux default loop stack space for CoopTask
Given that CoopTasks are scheduled from the Arduino default ``loop()`` or the
``main()`` function on Linux, any code in these functions is non-cooperative.
This is great for incompatible sketches or libraries, but otherwise puts the
memory of that main stack to waste. It is therefore good practice to allocate
the local stack for a single, infinitely running, CoopTask on the main stack.
Reserve enough stack to remain for ``loop()`` internals. In its most simple
form, borrowing from the example above, where taskBlink meets the requirement
of never returning, a CoopTask that uses the default stack space is created
like so:
```
#if defined(ESP8266) || defined(ESP32)
taskBlink = createCoopTask<int, CoopTaskStackAllocatorFromLoop<>>(
F("Blink"), loopBlink, 0x240);
#else
taskBlink = createCoopTask<int, CoopTaskStackAllocatorFromLoop<>>(
F("Blink"), loopBlink, 0x40);
#endif
```
## ESP8266 Core For Arduino specifics
ESP8266 Core For Arduino release 2.6.0 and later include all support for this
release of CoopTask.
## Arduino-ESP32 specifics
The ESP32 runs the Arduino API on top of the FreeRTOS real-time operating system.
This OS has all the capabilities for real-time programming and offers prioritized,
preemptive multitasking. The purpose of CoopTask on the other hand is to take
the complexity out of multi-threaded/tasked programming, and offers a cooperative
multi-tasking scheme instead.
Arduino-ESP32 has the necessary support for CoopTask beginning with
commit-ish c2b3f2d, dated Oct 4 2019, in Github master branch post release 1.4.0.
For Arduino sketches, and the libraries used in these, that never use the global
Arduino ``delay()``, don't make use of FreeRTOS ``vTaskDelay()``, and implement
delays only ever using the CoopTask metaphor ``CoopTaskBase::delay()``, CoopTask
would not require anything specific for the ESP32.
If the convenient Arduino ``delay()`` does get used, or there is any chance that
the FreeRTOS ``vTaskDelay()`` gets used, though, on the ESP32 it is necessary to
prevent unsolicited preemptive concurrency and control the CPU time for the
idle task.
This is being taken care of by CoopTask when using the ``runCoopTasks()``
scheduling helper in the Sketch ``loop()`` function.

99
arduino-cli/libraries/CoopTask/examples/Callbacks/Callbacks.ino Normal file
View File

@@ -0,0 +1,99 @@
#include <CoopTaskBase.h>
#include <CoopTask.h>
#include <CoopSemaphore.h>
#include <CoopMutex.h>
#include <BasicCoopTask.h>
#if defined(ARDUINO_attiny)
#define LED_BUILTIN 1
struct DummySerial {
void print(const __FlashStringHelper* s = nullptr) {}
void println(const __FlashStringHelper* s = nullptr) {}
void println(long unsigned int) {}
void flush() {}
};
DummySerial Serial;
#endif
CoopTask<void>* blinkTask = nullptr;
CoopTask<void>* switchTask = nullptr;
void blinkFunction()
{
pinMode(LED_BUILTIN, OUTPUT);
digitalWrite(LED_BUILTIN, HIGH);
for (;;)
{
yield(); // A
digitalWrite(LED_BUILTIN, LOW);
delay(5000); // B
digitalWrite(LED_BUILTIN, HIGH);
CoopTask<void>::sleep(); // D
}
}
void switchFunction()
{
for (;;)
{
yield(); // A
Serial.println(F("Switch on"));
delay(100); // B
Serial.println(F("Switch off"));
CoopTask<void>::sleep(); // C
}
}
bool delayCb(uint32_t ms)
{
Serial.print(F("delayDb, ms = "));
Serial.println(ms);
delay(ms);
return true;
}
bool sleepCb()
{
Serial.println(F("sleepCb"));
Serial.flush();
delay(10000);
if (blinkTask) blinkTask->wakeup();
if (switchTask) switchTask->wakeup();
return true;
}
void setup()
{
#if !defined(ARDUINO_attiny)
Serial.begin(74880);
while (!Serial);
delay(100);
Serial.println();
Serial.println(F("runTasks callback test"));
#endif
runCoopTasks(nullptr, delayCb, sleepCb);
Serial.println(F("no tasks yet, sleepCb()?"));
blinkTask = new CoopTask<void>(F("blink"), blinkFunction);
switchTask = new CoopTask<void>(F("switch"), switchFunction);
blinkTask->scheduleTask();
switchTask->scheduleTask();
}
// Add the main program code into the continuous loop() function
void loop()
{
runCoopTasks(nullptr, delayCb, sleepCb);
Serial.println(F("A - both tasks yielded, no Cb?"));
yield();
runCoopTasks(nullptr, delayCb, sleepCb);
Serial.println(F("B - both tasks delayed, delayCb(100)?"));
yield();
runCoopTasks(nullptr, delayCb, sleepCb);
Serial.println(F("C - blink task delayed, switch task sleeping, delayCb(4900)?"));
yield();
runCoopTasks(nullptr, delayCb, sleepCb);
Serial.println(F("D - both tasks sleeping, sleepCb()?"));
}

143
arduino-cli/libraries/CoopTask/examples/MultipleBlinks/MultipleBlinks.ino Normal file
View File

@@ -0,0 +1,143 @@
/*
Multiple Blinks
Ported to CoopTask from the version that
demonstrates the use of the Scheduler library for the Arduino Due.
CoopTask works on Arduino AVR (including ATtiny), ESP8266, ESP32, ARM Linux and PC OSs.
created 8 Oct 2012
by Cristian Maglie
Modified by
Scott Fitzgerald 19 Oct 2012
Ported to CoopTask by
Dirk O. Kaar 22 Dec 2019
This example code is in the public domain
http://www.arduino.cc/en/Tutorial/MultipleBlinks
*/
// Include CoopTask since we want to manage multiple tasks.
#include <CoopTask.h>
#include <CoopSemaphore.h>
// ATtiny85 max. working memory utilization, AVR 1.8.2 toolchain:
// "Minimum Memory Usage: 357 bytes (70% of a 512 byte maximum)"
#if defined(ARDUINO_attiny)
#define LED_BUILTIN 1
#endif
#if defined(ARDUINO_AVR_MICRO)
#define STACKSIZE_8BIT 92
#else
#define STACKSIZE_8BIT 40
#endif
CoopSemaphore taskSema(1, 1);
int taskToken = 1;
// Task no.1: blink LED with 1 second delay.
void loop1() {
for (;;) // explicitly run forever without returning
{
taskSema.wait();
if (1 != taskToken)
{
taskSema.post();
yield();
continue;
}
for (int i = 0; i < 3; ++i)
{
digitalWrite(LED_BUILTIN, HIGH);
// IMPORTANT:
// When multiple tasks are running 'delay' passes control to
// other tasks while waiting and guarantees they get executed.
delay(1000);
digitalWrite(LED_BUILTIN, LOW);
delay(1000);
}
taskToken = 2;
taskSema.post();
}
}
// Task no.2: blink LED with 0.25 second delay.
void loop2() {
for (;;) // explicitly run forever without returning
{
taskSema.wait();
if (2 != taskToken)
{
taskSema.post();
yield();
continue;
}
for (int i = 0; i < 6; ++i)
{
digitalWrite(LED_BUILTIN, HIGH);
// IMPORTANT:
// When multiple tasks are running 'delay' passes control to
// other tasks while waiting and guarantees they get executed.
delay(250);
digitalWrite(LED_BUILTIN, LOW);
delay(250);
}
taskToken = 3;
taskSema.post();
}
}
// Task no.3: blink LED with 0.05 second delay.
void loop3() {
for (;;) // explicitly run forever without returning
{
taskSema.wait();
if (3 != taskToken)
{
taskSema.post();
yield();
continue;
}
for (int i = 0; i < 6; ++i)
{
digitalWrite(LED_BUILTIN, HIGH);
// IMPORTANT:
// When multiple tasks are running 'delay' passes control to
// other tasks while waiting and guarantees they get executed.
delay(50);
digitalWrite(LED_BUILTIN, LOW);
delay(50);
}
taskToken = 1;
taskSema.post();
}
}
BasicCoopTask<CoopTaskStackAllocatorAsMember<sizeof(unsigned) >= 4 ? 800 : STACKSIZE_8BIT>> task1("l1", loop1);
BasicCoopTask<CoopTaskStackAllocatorAsMember<sizeof(unsigned) >= 4 ? 800 : STACKSIZE_8BIT>> task2("l2", loop2);
BasicCoopTask<CoopTaskStackAllocatorFromLoop<sizeof(unsigned) >= 4 ? 800 : STACKSIZE_8BIT>> task3("l3", loop3, sizeof(unsigned) >= 4 ? 800 : STACKSIZE_8BIT);
void setup() {
//Serial.begin(9600);
// Setup the 3 pins as OUTPUT
pinMode(LED_BUILTIN, OUTPUT);
// Add "loop1", "loop2" and "loop3" to CoopTask scheduling.
// "loop" is always started by default, and is not under the control of CoopTask.
task1.scheduleTask();
task2.scheduleTask();
task3.scheduleTask();
}
void loop() {
// loops forever by default
runCoopTasks();
}

451
arduino-cli/libraries/CoopTask/examples/blinkbuttonandweb/blinkbuttonandweb.ino Normal file
View File

@@ -0,0 +1,451 @@
#include <CoopTask.h>
#include <CoopMutex.h>
#include <CoopSemaphore.h>
#if defined(ESP8266)
#include <ESP8266WiFi.h>
#include <WiFiClient.h>
#include <ESP8266WebServer.h>
#include <ESP8266mDNS.h>
ESP8266WebServer server(80);
#elif defined(ESP32)
#include <WiFi.h>
#include <WiFiClient.h>
#include <WebServer.h>
#include <ESPmDNS.h>
WebServer server(80);
#endif
#if !defined(ESP8266) && !defined(ESP32)
#define IRAM_ATTR
#endif
#if defined(ESP8266)
constexpr auto LEDON = LOW;
constexpr auto LEDOFF = HIGH;
#else
constexpr auto LEDON = HIGH;
constexpr auto LEDOFF = LOW;
#endif
#if defined(ESP32)
#define BUTTON1 17
//#define BUTTON1 GPIO_NUM_27
#elif defined(ARDUINO_ESP8266_WEMOS_D1MINI)
#define BUTTON1 D3
#else
#define BUTTON1 0
#endif
#define USE_BUILTIN_TASK_SCHEDULER
// enter your WiFi configuration below
static const char AP_SSID[] PROGMEM = "SSID"; // your router's SSID here
static const char AP_PASS[] PROGMEM = "PSK"; // your router's password here
CoopMutex serialMutex;
CoopSemaphore blinkSema(0);
#if defined(ESP8266) || defined(ESP32)
class Button {
protected:
CoopSemaphore& reportSema;
public:
Button(uint8_t reqPin, CoopSemaphore& _reportSema) : reportSema(_reportSema), pushSema(0), PIN(reqPin) {
pinMode(PIN, INPUT_PULLUP);
attachInterruptArg(PIN, Button::buttonIsr_static, this, FALLING);
};
~Button() {
detachInterrupt(PIN);
}
CoopSemaphore pushSema;
void IRAM_ATTR buttonIsr() {
numberKeyPresses += 1;
pressed = true;
pushSema.post();
reportSema.post();
}
static void IRAM_ATTR buttonIsr_static(void* const self) {
reinterpret_cast<Button*>(self)->buttonIsr();
}
unsigned testResetPressed() {
if (pressed) {
CoopMutexLock serialLock(serialMutex);
Serial.printf_P(PSTR("Button on pin %u has been pressed %u times\n"), PIN, numberKeyPresses);
pressed = false;
}
return numberKeyPresses;
}
private:
const uint8_t PIN;
volatile unsigned numberKeyPresses = 0;
volatile bool pressed = false;
};
#endif
void loopBlink() noexcept
{
for (;;)
{
digitalWrite(LED_BUILTIN, LEDOFF);
blinkSema.wait(1000);
digitalWrite(LED_BUILTIN, LEDON);
CoopTask<>::delay(4000);
}
}
#if defined(ESP8266) || defined(ESP32)
Button* button1;
void loopButton() noexcept
{
int count = 0;
for (;;)
{
if (!button1->pushSema.wait())
{
CoopMutexLock serialLock(serialMutex);
Serial.println(F("loopButton: wait failed"));
yield();
continue;
}
else
{
++count;
}
{
CoopMutexLock serialLock(serialMutex);
Serial.print(F("loopButton: count = "));
Serial.println(count);
}
if (nullptr != button1 && 8000 < button1->testResetPressed()) {
delete button1;
button1 = nullptr;
CoopTask<>::exit();
}
yield();
}
}
void handleRoot() {
server.send(200, F("text/plain"), F("hello from esp8266!"));
}
void handleNotFound() {
String message = F("File Not Found\n\n");
message += F("URI: ");
message += server.uri();
message += F("\nMethod: ");
message += (server.method() == HTTP_GET) ? F("GET") : F("POST");
message += F("\nArguments: ");
message += server.args();
message += "\n";
for (uint8_t i = 0; i < server.args(); i++) {
message += ' ' + server.argName(i) + F(": ") + server.arg(i) + '\n';
}
server.send(404, F("text/plain"), message);
}
#endif
#if defined(ESP8266) || defined(ESP32)
CoopTask<void>* taskButton;
#endif
CoopTask<void, CoopTaskStackAllocatorFromLoop<>>* taskBlink;
CoopTask<unsigned>* taskText;
CoopTask<void>* taskReport0;
CoopTask<void>* taskReport1;
CoopTask<void>* taskReport2;
#if defined(ESP8266) || defined(ESP32)
CoopTask<void>* taskReport3;
CoopTask<void>* taskReport4;
CoopTask<void>* taskWeb;
#endif
CoopSemaphore reportSema(0);
void printStackReport(CoopTaskBase* task)
{
if (!task) return;
Serial.print(task->name().c_str());
Serial.print(F(" free stack = "));
Serial.println(task->getFreeStack());
}
uint32_t iterations = 0;
uint32_t start;
// to demonstrate that yield and delay work in subroutines
void printReport()
{
//CoopTask<>::delayMicroseconds(4000000);
Serial.print(F("cycle period/us = "));
if (iterations)
{
Serial.println(1.0F * (micros() - start) / iterations);
}
else
{
Serial.println(F("N/A"));
}
#if defined(ESP8266) || defined(ESP32)
printStackReport(taskButton);
#endif
printStackReport(taskBlink);
printStackReport(taskText);
printStackReport(CoopTask<>::self());
#if defined(ESP8266) || defined(ESP32)
printStackReport(taskWeb);
#endif
iterations = 0;
};
class RAIITest
{
public:
~RAIITest()
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name());
Serial.println(F(" stack unwound, RAIITest object destructed"));
}
};
void setup()
{
#ifdef ESP8266
Serial.begin(74880);
#else
Serial.begin(115200);
#endif
while (!Serial) {}
delay(500);
Serial.println(F("Scheduler test"));
#if defined(ESP8266) || defined(ESP32)
WiFi.mode(WIFI_STA);
WiFi.begin(FPSTR(AP_SSID), FPSTR(AP_PASS));
// Wait for connection
while (WiFi.status() != WL_CONNECTED) {
delay(500);
Serial.print('.');
}
Serial.println();
Serial.print(F("IP address: "));
Serial.println(WiFi.localIP());
if (MDNS.begin(F("esp"))) {
Serial.println(F("MDNS responder started"));
}
server.on(F("/"), handleRoot);
server.on(F("/inline"), []() {
server.send(200, F("text/plain"), F("this works as well"));
});
server.onNotFound(handleNotFound);
server.begin();
Serial.println(F("HTTP server started"));
#endif
#if defined(ESP8266) && defined(USE_BUILTIN_TASK_SCHEDULER)
CoopTaskBase::useBuiltinScheduler();
#endif
pinMode(LED_BUILTIN, OUTPUT);
#if defined(ESP8266) || defined(ESP32)
button1 = new Button(BUTTON1, reportSema);
taskButton = new CoopTask<void>(F("Button"), loopButton,
#if defined(ESP8266)
0x700);
#elif defined(ESP32)
0x940);
#endif
if (!*taskButton) Serial.printf_P(PSTR("CoopTask %s out of stack\n"), taskButton->name().c_str());
#endif
taskBlink = new CoopTask<void, CoopTaskStackAllocatorFromLoop<>>(F("Blink"), loopBlink,
#if defined(ESP8266)
0x400);
#elif defined(ESP32)
0x540);
#else
0x40);
#endif
if (!*taskBlink) Serial.println(F("CoopTask Blink out of stack"));
taskText = new CoopTask<unsigned>(F("Text"), []() -> unsigned
{
RAIITest raii;
{
CoopMutexLock serialLock(serialMutex);
Serial.println(F("Task1 - A"));
}
yield();
{
CoopMutexLock serialLock(serialMutex);
Serial.println(F("Task1 - B"));
}
uint32_t start = millis();
CoopTask<>::delay(6000);
{
CoopMutexLock serialLock(serialMutex);
Serial.print(F("!!!Task1 - C - "));
Serial.println(millis() - start);
printStackReport(taskText);
}
#if !defined(ARDUINO)
throw static_cast<unsigned>(41);
#endif
CoopMutexLock serialLock(serialMutex);
Serial.print(F("exiting from task "));
Serial.println(CoopTaskBase::self()->name());
//CoopTask<unsigned>::exit(42);
return 43;
}
#if defined(ESP8266)
, 0x380);
#elif defined(ESP32)
, 0x4c0);
#else
, 0x70);
#endif
if (!*taskText) Serial.println(F("CoopTask Text out of stack"));
auto reportFunc = []() noexcept
{
uint32_t count = 0;
for (;;) {
if (!reportSema.wait(120000))
{
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name().c_str());
Serial.println(F(": wait failed"));
}
yield();
continue;
}
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTask<>::self()->name());
Serial.print(F(" ("));
Serial.print(++count);
Serial.println(F("x)"));
printReport();
}
yield();
reportSema.setval(0);
}
};
taskReport0 = new CoopTask<void>(F("Report0"), reportFunc
#if defined(ESP8266) || defined(ESP32)
, 0x600);
#else
, 0x70);
#endif
if (!*taskReport0) Serial.println(F("CoopTask Report out of stack"));
taskReport1 = new CoopTask<void>(F("Report1"), reportFunc
#if defined(ESP8266) || defined(ESP32)
, 0x600);
#else
, 0x70);
#endif
if (!*taskReport1) Serial.println(F("CoopTask Report out of stack"));
taskReport2 = new CoopTask<void>(F("Report2"), reportFunc
#if defined(ESP8266) || defined(ESP32)
, 0x600);
#else
, 0x70);
#endif
if (!*taskReport2) Serial.println(F("CoopTask Report out of stack"));
#if defined(ESP8266) || defined(ESP32)
taskReport3 = new CoopTask<void>(F("Report3"), reportFunc
, 0x600);
if (!*taskReport3) Serial.println(F("CoopTask Report out of stack"));
taskReport4 = new CoopTask<void>(F("Report4"), reportFunc
, 0x600);
if (!*taskReport4) Serial.println(F("CoopTask Report out of stack"));
taskWeb = new CoopTask<void>(F("Web"), []() noexcept
{
for (;;) {
server.handleClient();
#ifdef ESP8266
MDNS.update();
#endif
yield();
}
},
#if defined(ESP8266)
0x800);
#else
0xa00);
#endif
if (!*taskWeb) Serial.printf_P(PSTR("CoopTask %s out of stack\n"), taskWeb->name().c_str());
if (!taskButton->scheduleTask()) { Serial.printf_P(PSTR("Could not schedule task %s\n"), taskButton->name().c_str()); }
if (!taskReport3->scheduleTask()) { Serial.printf_P(PSTR("Could not schedule task %s\n"), taskReport3->name().c_str()); }
if (!taskReport4->scheduleTask()) { Serial.printf_P(PSTR("Could not schedule task %s\n"), taskReport4->name().c_str()); }
if (!taskWeb->scheduleTask()) { Serial.printf_P(PSTR("Could not schedule task %s\n"), taskWeb->name().c_str()); }
#endif
if (!taskBlink->scheduleTask()) { Serial.print(F("Could not schedule task ")); Serial.println(taskBlink->name().c_str()); }
if (!taskText->scheduleTask()) { Serial.print(F("Could not schedule task ")); Serial.println(taskText->name().c_str()); }
if (!taskReport0->scheduleTask()) { Serial.print(F("Could not schedule task ")); Serial.println(taskReport0->name().c_str()); }
if (!taskReport1->scheduleTask()) { Serial.print(F("Could not schedule task ")); Serial.println(taskReport1->name().c_str()); }
if (!taskReport2->scheduleTask()) { Serial.print(F("Could not schedule task ")); Serial.println(taskReport2->name().c_str()); }
#ifdef ESP32
Serial.print(F("Loop free stack = ")); Serial.println(uxTaskGetStackHighWaterMark(NULL));
#endif
}
void taskReaper(const CoopTaskBase* const task)
{
if (task == taskText)
{
delete task;
taskText = nullptr;
}
}
void loop()
{
#if defined(ESP8266) && defined(USE_BUILTIN_TASK_SCHEDULER)
if (taskText && !*taskText)
{
taskReaper(taskText);
}
#else
runCoopTasks(taskReaper);
#endif
// taskReport sleeps on first run(), and after each report.
// It resets iterations to 0 on each report.
if (!iterations) start = micros();
++iterations;
#ifdef ESP32_FREERTOS
if (iterations >= 50000)
#else
if (iterations >= 200000)
#endif
{
reportSema.post();
}
}

252
arduino-cli/libraries/CoopTask/examples/mutex/mutex.ino Normal file
View File

@@ -0,0 +1,252 @@
/*
Name: mutex.ino
Created: 2019-08-04 22:40:11
Author: dok@dok-net.net
*/
#include <CoopTask.h>
#include <CoopSemaphore.h>
#include <CoopMutex.h>
#define USE_BUILTIN_TASK_SCHEDULER
CoopMutex serialMutex;
CoopMutex mutex;
CoopTaskBase* hasMutex(nullptr);
void haveMutex()
{
if (hasMutex)
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name());
Serial.print(F(" called haveMutex, despite "));
Serial.print(hasMutex->name());
Serial.println(F(" is known to have mutex."));
return;
}
hasMutex = CoopTaskBase::self();
}
void yieldMutex()
{
if (hasMutex != CoopTaskBase::self())
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name());
Serial.println(F(" called yieldMutex, but no task currently has the mutex."));
return;
}
hasMutex = nullptr;
}
CoopTask<int>* firstTask;
CoopTask<int>* secondTask;
CoopTask<int>* thirdTask;
void setup() {
#ifdef ESP8266
Serial.begin(74880);
#else
Serial.begin(115200);
#endif
while (!Serial) {}
delay(500);
Serial.println(F("Mutex test"));
#if defined(ESP8266) && defined(USE_BUILTIN_TASK_SCHEDULER)
CoopTaskBase::useBuiltinScheduler();
#endif
firstTask = createCoopTask(F("first"), []()
{
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name());
Serial.println(F(" starts"));
}
for (int i = 0; i < 30; ++i)
{
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name());
Serial.println(F(" locks mutex"));
}
{
CoopMutexLock lock(mutex);
if (!lock) {
CoopMutexLock serialLock(serialMutex);
Serial.print(F("failed to lock mutex in "));
Serial.println(CoopTaskBase::self()->name());
}
haveMutex();
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name());
Serial.println(F(" has mutex"));
}
yield();
yieldMutex();
}
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name());
Serial.print(F(" runs ("));
Serial.print(i);
Serial.println(')');
}
yield();
}
CoopMutexLock serialLock(serialMutex);
Serial.print(F("exiting from task "));
Serial.println(CoopTaskBase::self()->name());
return 0;
}
#if defined(ESP8266) || defined(ESP32)
);
#else
, 0x120);
#endif
if (!firstTask) Serial.println(F("firstTask not created"));
secondTask = createCoopTask(F("second"), []()
{
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name());
Serial.println(F(" starts"));
}
for (int i = 0; i < 30; ++i)
{
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name());
Serial.println(F(" locks mutex"));
}
{
CoopMutexLock lock(mutex);
if (!lock) {
CoopMutexLock serialLock(serialMutex);
Serial.print(F("failed to lock mutex in "));
Serial.println(CoopTaskBase::self()->name());
}
haveMutex();
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name());
Serial.println(F(" has mutex"));
}
yield();
yieldMutex();
}
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name());
Serial.print(F(" runs ("));
Serial.print(i);
Serial.println(')');
}
yield();
}
CoopMutexLock serialLock(serialMutex);
Serial.print(F("exiting from task "));
Serial.println(CoopTaskBase::self()->name());
return 0;
}
#if defined(ESP8266) || defined(ESP32)
);
#else
, 0x120);
#endif
if (!secondTask) Serial.println(F("secondTask not created"));
thirdTask = createCoopTask(F("third"), []()
{
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name());
Serial.println(F(" starts"));
}
for (int i = 0; i < 10; ++i)
{
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name());
Serial.println(F(" locks mutex"));
}
{
CoopMutexLock lock(mutex);
if (!lock) {
CoopMutexLock serialLock(serialMutex);
Serial.print(F("failed to lock mutex in "));
Serial.println(CoopTaskBase::self()->name());
}
haveMutex();
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name());
Serial.println(F(" has mutex"));
}
yield();
yieldMutex();
}
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name());
Serial.print(F(" runs ("));
Serial.print(i);
Serial.println(')');
}
yield();
}
for (int i = 0; i < 10; ++i)
{
{
CoopMutexLock serialLock(serialMutex);
Serial.print(CoopTaskBase::self()->name());
Serial.print(F(" still runs ("));
Serial.print(i);
Serial.println(')');
}
yield();
}
CoopMutexLock serialLock(serialMutex);
Serial.print(F("exiting from task "));
Serial.println(CoopTaskBase::self()->name());
return 0;
}
#if defined(ESP8266) || defined(ESP32)
);
#else
, 0x120);
#endif
if (!thirdTask) Serial.println(F("thirdTask not created"));
#ifdef ESP32
Serial.print(F("Loop free stack = ")); Serial.println(uxTaskGetStackHighWaterMark(NULL));
#endif
}
void taskReaper(const CoopTaskBase* const task)
{
delete task;
}
// the loop function runs over and over again until power down or reset
void loop()
{
#if defined(ESP8266) && defined(USE_BUILTIN_TASK_SCHEDULER)
if (firstTask && !*firstTask)
{
taskReaper(firstTask); firstTask = nullptr;
}
if (secondTask && !*secondTask)
{
taskReaper(secondTask); secondTask = nullptr;
}
if (thirdTask && !*thirdTask)
{
taskReaper(thirdTask); thirdTask = nullptr;
}
#else
runCoopTasks(taskReaper);
#endif
}

96
arduino-cli/libraries/CoopTask/examples/portable/portable.cpp Normal file
View File

@@ -0,0 +1,96 @@
// portable.cpp
// This is a basic portable example, without a scheduler.
// All tasks are run round-robin inside a for loop.
// It shows CoopTask creation, synchronization, and termination.
#include <iostream>
#include "CoopTask.h"
#include "CoopSemaphore.h"
#include "CoopMutex.h"
template<typename StackAllocator> void printStackReport(BasicCoopTask<StackAllocator>& task)
{
if (!task) return;
std::cerr << task.name().c_str() << " free stack = " << task.getFreeStack() << std::endl;
}
CoopMutex blinkMutex;
int main()
{
CoopSemaphore terminatorSema(0);
CoopSemaphore helloSema(0);
auto& hello = *createCoopTask<void>(std::string("hello"), [&terminatorSema, &helloSema]() noexcept
{
std::cerr << "Hello" << std::endl;
yield();
for (int x = 0; x < 10; ++x)
{
{
CoopMutexLock lock(blinkMutex);
std::cerr << "Loop" << std::endl;
}
helloSema.wait(2000);
}
terminatorSema.post();
}, 0x2000);
if (!hello) std::cerr << hello.name() << " CoopTask not created" << std::endl;
bool keepBlinking = true;
auto& terminator = *createCoopTask<void>(std::string("terminator"), [&keepBlinking, &terminatorSema]() noexcept
{
if (!terminatorSema.wait()) std::cerr << "terminatorSema.wait() failed" << std::endl;
keepBlinking = false;
}, 0x2000);
if (!terminator) std::cerr << terminator.name() << " CoopTask not created" << std::endl;
auto& blink = *createCoopTask<std::string, CoopTaskStackAllocatorFromLoop<>>(std::string("blink"), [&keepBlinking]()
{
while (keepBlinking)
{
{
CoopMutexLock lock(blinkMutex);
std::cerr << "LED on" << std::endl;
delay(1000);
std::cerr << "LED off" << std::endl;
}
delay(1000);
}
throw std::string("sixtynine");
return "fortytwo";
}, 0x2000);
if (!blink) std::cerr << blink.name() << " CoopTask not created" << std::endl;
auto& report = *createCoopTask<void>(std::string("report"), [&hello, &blink]() noexcept
{
for (;;) {
delay(5000);
{
CoopMutexLock lock(blinkMutex);
printStackReport(hello);
printStackReport(blink);
}
}
}, 0x2000);
if (!report) std::cerr << report.name() << " CoopTask not created" << std::endl;
auto taskReaper = [&blink](const CoopTaskBase* const task)
{
// once: hello posts terminatorSema -> terminator sets keepBlinking = false -> blink exits -> break leaves for-loop -> program exits
if (task == &blink)
{
std::cerr << task->name() << " returns = " << blink.exitCode() << std::endl;
delete task;
exit(0);
}
};
for (;;)
{
runCoopTasks(taskReaper);
}
return 0;
}

21
arduino-cli/libraries/CoopTask/library.json Normal file
View File

@@ -0,0 +1,21 @@
{
"name": "CoopTask",
"version": "3.6.4",
"description": "Portable C++ library for cooperative multitasking like Arduino Scheduler on ESP8266/ESP32, AVR, Linux, Windows",
"keywords": [
"multitasking", "timing"
],
"repository":
{
"type": "git",
"url": "https://github.com/dok-net/CoopTask"
},
"authors": [
{
"name": "Dirk Kaar"
}
],
"license": "LGPL-2.1+",
"frameworks": "arduino",
"platforms": "*"
}

9
arduino-cli/libraries/CoopTask/library.properties Normal file
View File

@@ -0,0 +1,9 @@
name=CoopTask
version=3.6.4
author=Dirk O. Kaar
maintainer=Dirk O. Kaar <dok@dok-net.net>
sentence=Portable C++ library for cooperative multitasking like Arduino Scheduler on ESP8266/ESP32, AVR, Linux, Windows
paragraph=Run multiple concurrent setup()/loop() tasks in Arduino sketches. Use the normal global delay() function, use yield() to give up the CPU to other tasks and the main loop().
category=Timing
url=https://github.com/dok-net/CoopTask
architectures=*

61
arduino-cli/libraries/CoopTask/src/BasicCoopTask.cpp Normal file
View File

@@ -0,0 +1,61 @@
/*
BasicCoopTask.cpp - Implementation of cooperative scheduling tasks
Copyright (c) 2019 Dirk O. Kaar. All rights reserved.
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "BasicCoopTask.h"
#if defined(ARDUINO) && !defined(ESP32_FREERTOS)
#include <alloca.h>
#endif
#if !defined(_MSC_VER) && !defined(ESP32_FREERTOS)
char* CoopTaskStackAllocator::allocateStack(size_t stackSize)
{
char* stackTop = nullptr;
if (stackSize <= CoopTaskBase::MAXSTACKSPACE - (CoopTaskBase::FULLFEATURES ? 2 : 1) * sizeof(CoopTaskBase::STACKCOOKIE))
{
#if defined(ESP8266)
stackTop = new (std::nothrow) char[stackSize + (CoopTaskBase::FULLFEATURES ? 2 : 1) * sizeof(CoopTaskBase::STACKCOOKIE)];
#else
stackTop = new char[stackSize + (CoopTaskBase::FULLFEATURES ? 2 : 1) * sizeof(CoopTaskBase::STACKCOOKIE)];
#endif
}
return stackTop;
}
#endif // !defined(_MSC_VER) && !defined(ESP32_FREERTOS)
#if (defined(ARDUINO) && !defined(ESP32_FREERTOS)) || defined(__GNUC__)
char* CoopTaskStackAllocatorFromLoopBase::allocateStack(size_t loopReserve, size_t stackSize)
{
char* bp = static_cast<char*>(alloca(
(sizeof(unsigned) >= 4) ? ((loopReserve + sizeof(unsigned) - 1) / sizeof(unsigned)) * sizeof(unsigned) : loopReserve
));
std::atomic_thread_fence(std::memory_order_release);
char* stackTop = nullptr;
if (stackSize <= CoopTaskBase::MAXSTACKSPACE - (CoopTaskBase::FULLFEATURES ? 2 : 1) * sizeof(CoopTaskBase::STACKCOOKIE))
{
stackTop = reinterpret_cast<char*>(
reinterpret_cast<long unsigned>(bp) - stackSize + (CoopTaskBase::FULLFEATURES ? 2 : 1) * sizeof(CoopTaskBase::STACKCOOKIE));
}
return stackTop;
}
#endif // (defined(ARDUINO) && !defined(ESP32_FREERTOS)) || defined(__GNUC__)

113
arduino-cli/libraries/CoopTask/src/BasicCoopTask.h Normal file
View File

@@ -0,0 +1,113 @@
/*
BasicCoopTask.h - Implementation of cooperative scheduling tasks
Copyright (c) 2019 Dirk O. Kaar. All rights reserved.
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#ifndef __BasicCoopTask_h
#define __BasicCoopTask_h
#include "CoopTaskBase.h"
class CoopTaskStackAllocator
{
public:
static constexpr size_t DEFAULTTASKSTACKSIZE = CoopTaskBase::DEFAULTTASKSTACKSIZE;
#if !defined(_MSC_VER) && !defined(ESP32_FREERTOS)
static char* allocateStack(size_t stackSize);
static void disposeStack(char* stackTop) { delete[] stackTop; }
#endif
};
template<size_t StackSize = CoopTaskBase::DEFAULTTASKSTACKSIZE>
class CoopTaskStackAllocatorAsMember
{
public:
static constexpr size_t DEFAULTTASKSTACKSIZE =
(sizeof(unsigned) >= 4) ? ((StackSize + sizeof(unsigned) - 1) / sizeof(unsigned)) * sizeof(unsigned) : StackSize;
#if !defined(_MSC_VER) && !defined(ESP32_FREERTOS)
protected:
char _stackTop[DEFAULTTASKSTACKSIZE + (CoopTaskBase::FULLFEATURES ? 2 : 1) * sizeof(CoopTaskBase::STACKCOOKIE)];
public:
char* allocateStack(size_t stackSize)
{
return (DEFAULTTASKSTACKSIZE >= stackSize) ?
_stackTop : nullptr;
}
static void disposeStack(char* stackTop) { }
#endif
};
class CoopTaskStackAllocatorFromLoopBase
{
public:
static constexpr size_t DEFAULTTASKSTACKSIZE = CoopTaskBase::DEFAULTTASKSTACKSIZE;
#if (defined(ARDUINO) && !defined(ESP32_FREERTOS)) || defined(__GNUC__)
protected:
static char* allocateStack(size_t loopReserve, size_t stackSize);
#endif
public:
static void disposeStack(char* stackTop) { }
};
template<size_t LoopReserve = (CoopTaskBase::DEFAULTTASKSTACKSIZE / 2)>
class CoopTaskStackAllocatorFromLoop : public CoopTaskStackAllocatorFromLoopBase
{
public:
static constexpr size_t DEFAULTTASKSTACKSIZE = CoopTaskBase::DEFAULTTASKSTACKSIZE;
#if (defined(ARDUINO) && !defined(ESP32_FREERTOS)) || defined(__GNUC__)
static char* allocateStack(size_t stackSize)
{
return CoopTaskStackAllocatorFromLoopBase::allocateStack(LoopReserve, stackSize);
}
#endif
};
template<class StackAllocator = CoopTaskStackAllocator> class BasicCoopTask : public CoopTaskBase
{
public:
#ifdef ARDUINO
BasicCoopTask(const String& name, taskfunction_t _func, size_t stackSize = StackAllocator::DEFAULTTASKSTACKSIZE) :
#else
BasicCoopTask(const std::string& name, taskfunction_t _func, size_t stackSize = StackAllocator::DEFAULTTASKSTACKSIZE) :
#endif
CoopTaskBase(name, _func, stackSize)
{
#if !defined(_MSC_VER) && !defined(ESP32_FREERTOS)
taskStackTop = stackAllocator.allocateStack(taskStackSize);
#endif
}
BasicCoopTask(const BasicCoopTask&) = delete;
BasicCoopTask& operator=(const BasicCoopTask&) = delete;
~BasicCoopTask()
{
#if !defined(_MSC_VER) && !defined(ESP32_FREERTOS)
stackAllocator.disposeStack(taskStackTop);
#endif
}
/// Every task is entered into this list by scheduleTask(). It is removed when it exits
/// or gets deleted.
static const std::array< std::atomic<BasicCoopTask* >, MAXNUMBERCOOPTASKS + 1>& getRunnableTasks()
{
// this is safe to do because CoopTaskBase ctor is protected.
return reinterpret_cast<const std::array< std::atomic<BasicCoopTask* >, MAXNUMBERCOOPTASKS + 1>&>(CoopTaskBase::getRunnableTasks());
}
protected:
StackAllocator stackAllocator;
};
#endif // __BasicCoopTask_h

93
arduino-cli/libraries/CoopTask/src/CoopMutex.h Normal file
View File

@@ -0,0 +1,93 @@
/*
CoopMutex.h - Implementation of a mutex and an RAII lock for cooperative scheduling tasks
Copyright (c) 2019 Dirk O. Kaar. All rights reserved.
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#ifndef __CoopMutex_h
#define __CoopMutex_h
#include "CoopSemaphore.h"
/// A mutex that is safe to use from CoopTasks.
class CoopMutex : private CoopSemaphore
{
protected:
std::atomic<CoopTaskBase*> owner;
public:
CoopMutex(size_t maxPending = 10) : CoopSemaphore(1, maxPending), owner(nullptr) {}
CoopMutex(const CoopMutex&) = delete;
CoopMutex& operator=(const CoopMutex&) = delete;
/// @returns: true, or false, if the current task does not own the mutex.
bool unlock()
{
if (CoopTaskBase::running() && CoopTaskBase::self() == owner.load() && post())
{
owner.store(nullptr);
return true;
}
return false;
}
/// @returns: true if the mutex becomes locked. false if it is already locked by the same task, or the maximum number of pending tasks is exceeded.
bool lock()
{
if (CoopTaskBase::running() && CoopTaskBase::self() != owner.load() && wait())
{
owner.store(CoopTaskBase::self());
return true;
}
return false;
}
/// @returns: true if the mutex becomes freshly locked without waiting, otherwise false.
bool try_lock()
{
if (CoopTaskBase::running() && CoopTaskBase::self() != owner.load() && try_wait())
{
owner.store(CoopTaskBase::self());
return true;
}
return false;
}
};
/// A RAII CoopMutex lock class.
class CoopMutexLock {
protected:
CoopMutex& mutex;
bool locked;
public:
/// The constructor returns if the mutex was locked, or locking failed.
explicit CoopMutexLock(CoopMutex& _mutex) : mutex(_mutex) {
locked = mutex.lock();
}
CoopMutexLock() = delete;
CoopMutexLock(const CoopMutexLock&) = delete;
CoopMutexLock& operator=(const CoopMutexLock&) = delete;
/// @returns: true if the mutex became locked, potentially after blocking, otherwise false.
operator bool() const {
return locked;
}
/// The destructor unlocks the mutex.
~CoopMutexLock() {
if (locked) mutex.unlock();
}
};
#endif // __CoopMutex_h

257
arduino-cli/libraries/CoopTask/src/CoopSemaphore.cpp Normal file
View File

@@ -0,0 +1,257 @@
/*
CoopSemaphore.cpp - Implementation of a semaphore for cooperative scheduling tasks
Copyright (c) 2019 Dirk O. Kaar. All rights reserved.
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "CoopSemaphore.h"
#if defined(ESP8266)
#include <interrupts.h>
using esp8266::InterruptLock;
#elif defined(ESP32) || !defined(ARDUINO)
using std::min;
#else
class InterruptLock {
public:
InterruptLock() {
noInterrupts();
}
~InterruptLock() {
interrupts();
}
};
#endif
#ifndef ARDUINO
#include <chrono>
namespace
{
uint32_t millis()
{
return std::chrono::duration_cast<std::chrono::milliseconds>(std::chrono::system_clock::now().time_since_epoch()).count();
}
}
#endif
bool CoopSemaphore::_wait(const bool withDeadline, const uint32_t ms)
{
const uint32_t start = withDeadline ? millis() : 0;
uint32_t expired = 0;
bool selfFirst = false;
for (;;)
{
auto self = CoopTaskBase::self();
unsigned val;
#if !defined(ESP32) && defined(ARDUINO)
{
InterruptLock lock;
val = value.load();
if (val)
{
value.store(val - 1);
}
}
#else
val = 1;
while (val && !value.compare_exchange_weak(val, val - 1)) {}
#endif
const unsigned valOnEntry = val;
if (withDeadline) expired = millis() - start;
if (!(selfFirst && valOnEntry))
{
if (pendingTasks.push(self))
{
if (!withDeadline) self->sleep(true);
}
else
{
selfFirst = true;
}
}
bool fwd = !selfFirst && val;
bool stop = false;
CoopTaskBase* pendingTask = nullptr;
bool selfSuccess = false;
for (;;)
{
if (pendingTasks.available())
{
#if !defined(ESP32) && defined(ARDUINO)
{
InterruptLock lock;
pendingTask = pendingTask0.load();
if (fwd || !pendingTask) pendingTask0.store(pendingTasks.pop());
}
#else
pendingTask = nullptr;
bool exchd = false;
while ((fwd || !pendingTask) && !(exchd = pendingTask0.compare_exchange_weak(pendingTask, pendingTasks.peek()))) {}
if (exchd) pendingTasks.pop();
#endif
}
else
{
#if !defined(ESP32) && defined(ARDUINO)
{
InterruptLock lock;
pendingTask = pendingTask0.load();
if (fwd && pendingTask) pendingTask0.store(nullptr);
}
#else
pendingTask = nullptr;
if (fwd) pendingTask = pendingTask0.exchange(nullptr);
#endif
stop = true;
}
if (!val)
{
break;
}
if (!(pendingTask || stop))
{
continue;
}
if (selfFirst)
{
selfFirst = false;
if (!withDeadline) self->sleep(false);
selfSuccess = true;
}
else if (pendingTask == self)
{
if (!selfSuccess)
{
if (!withDeadline) self->sleep(false);
return true;
}
if (!stop) continue;
}
else if (pendingTask)
{
pendingTask->scheduleTask(true);
}
if (stop)
{
break;
}
val -= 1;
fwd = val;
}
if (selfSuccess)
{
return true;
}
if (valOnEntry)
{
#if !defined(ESP32) && defined(ARDUINO)
InterruptLock lock;
val = value.load();
value.store(val + 1);
#else
while (!value.compare_exchange_weak(val, val + 1)) {}
#endif
}
if (withDeadline)
{
if (expired >= ms)
{
pendingTasks.for_each_rev_requeue(notIsSelfTask);
#if !defined(ESP32) && defined(ARDUINO)
{
InterruptLock lock;
pendingTask = pendingTask0.load();
if (pendingTask == self) pendingTask0.store(pendingTasks.available() ? pendingTasks.pop() : nullptr);
}
#else
bool exchd = false;
pendingTask = self;
while ((pendingTask == self) && !(exchd = pendingTask0.compare_exchange_weak(pendingTask, pendingTasks.available() ? pendingTasks.peek() : nullptr))) {}
if (exchd && pendingTasks.available()) pendingTasks.pop();
#endif
return false;
}
CoopTaskBase::delay(ms - expired);
}
else
{
CoopTaskBase::yield();
}
selfFirst = true;
}
}
bool IRAM_ATTR CoopSemaphore::post()
{
CoopTaskBase* pendingTask;
#if !defined(ESP32) && defined(ARDUINO)
{
InterruptLock lock;
unsigned val = value.load();
value.store(val + 1);
pendingTask = pendingTask0.load();
if (pendingTask) pendingTask0.store(nullptr);
}
#else
unsigned val = 0;
while (!value.compare_exchange_weak(val, val + 1)) {}
pendingTask = pendingTask0.exchange(nullptr);
#endif
if (!pendingTask || !pendingTask->suspended()) return true;
return pendingTask->scheduleTask(true);
}
bool CoopSemaphore::setval(unsigned newVal)
{
CoopTaskBase* pendingTask = nullptr;
unsigned val;
#if !defined(ESP32) && defined(ARDUINO)
{
InterruptLock lock;
val = value.load();
value.store(newVal);
if (newVal > val)
{
pendingTask = pendingTask0.load();
pendingTask0.store(nullptr);
}
}
#else
val = value.exchange(newVal);
if (newVal > val) pendingTask = pendingTask0.exchange(nullptr);
#endif
if (!pendingTask || !pendingTask->suspended()) return true;
return pendingTask->scheduleTask(true);
}
bool CoopSemaphore::try_wait()
{
unsigned val;
#if !defined(ESP32) && defined(ARDUINO)
{
InterruptLock lock;
val = value.load();
if (val)
{
value.store(val - 1);
}
}
#else
val = 1;
while (val && !value.compare_exchange_weak(val, val - 1)) {}
#endif
return val;
}

93
arduino-cli/libraries/CoopTask/src/CoopSemaphore.h Normal file
View File

@@ -0,0 +1,93 @@
/*
CoopSemaphore.h - Implementation of a semaphore for cooperative scheduling tasks
Copyright (c) 2019 Dirk O. Kaar. All rights reserved.
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#ifndef __CoopSemaphore_h
#define __CoopSemaphore_h
#include "CoopTaskBase.h"
#include "circular_queue/circular_queue.h"
/// A semaphore that is safe to use from CoopTasks.
/// Only post() is safe to use from interrupt service routines,
/// or concurrent OS threads that must synchronized with the singled thread running CoopTasks.
class CoopSemaphore
{
protected:
std::atomic<unsigned> value;
std::atomic<CoopTaskBase*> pendingTask0;
circular_queue<CoopTaskBase*> pendingTasks;
// capture-less functions for iterators.
static void awakeAndSchedule(CoopTaskBase*&& task)
{
task->scheduleTask(true);
}
static bool notIsSelfTask(CoopTaskBase*& task)
{
return CoopTaskBase::self() != task;
}
/// @param withDeadline true: the ms parameter specifies the relative timeout for a successful
/// aquisition of the semaphore.
/// false: there is no deadline, the ms parameter is disregarded.
/// @param ms the relative timeout measured in milliseconds.
/// @returns: true if it sucessfully acquired the semaphore, either immediately or after sleeping.
/// false if the deadline expired, or the maximum number of pending tasks is exceeded.
bool _wait(const bool withDeadline = false, const uint32_t ms = 0);
public:
/// @param val the initial value of the semaphore.
/// @param maxPending the maximum supported number of concurrently waiting tasks.
CoopSemaphore(unsigned val, size_t maxPending = 10) : value(val), pendingTask0(nullptr), pendingTasks(maxPending) {}
CoopSemaphore(const CoopSemaphore&) = delete;
CoopSemaphore& operator=(const CoopSemaphore&) = delete;
~CoopSemaphore()
{
// wake up all queued tasks
pendingTasks.for_each(awakeAndSchedule);
}
/// post() is the only operation that is allowed from an interrupt service routine,
/// or a concurrent OS thread that is synchronized with the singled thread running CoopTasks.
bool IRAM_ATTR post();
/// @param newVal: the semaphore is immediately set to the specified value. if newVal is greater
/// than the current semaphore value, the behavior is identical to as many post operations.
bool setval(unsigned newVal);
/// @returns: true if it sucessfully acquired the semaphore, either immediately or after sleeping.
/// false if the maximum number of pending tasks is exceeded.
bool wait()
{
return _wait();
}
/// @param ms the relative timeout, measured in milliseconds, for a successful aquisition of the semaphore.
/// @returns: true if it sucessfully acquired the semaphore, either immediately or after sleeping.
/// false if the deadline expired, or the maximum number of pending tasks is exceeded.
bool wait(uint32_t ms)
{
return _wait(true, ms);
}
/// @returns: true if the semaphore was acquired immediately, otherwise false.
bool try_wait();
};
#endif // __CoopSemaphore_h

129
arduino-cli/libraries/CoopTask/src/CoopTask.h Normal file
View File

@@ -0,0 +1,129 @@
/*
CoopTask.h - Implementation of cooperative scheduling tasks
Copyright (c) 2019 Dirk O. Kaar. All rights reserved.
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#ifndef __CoopTask_h
#define __CoopTask_h
#include "BasicCoopTask.h"
template<typename Result = int, class StackAllocator = CoopTaskStackAllocator> class CoopTask : public BasicCoopTask<StackAllocator>
{
public:
using taskfunction_t = Delegate< Result() >;
#if defined(ARDUINO)
CoopTask(const String& name, CoopTask::taskfunction_t _func, size_t stackSize = BasicCoopTask<StackAllocator>::DEFAULTTASKSTACKSIZE) :
#else
CoopTask(const std::string& name, CoopTask::taskfunction_t _func, size_t stackSize = BasicCoopTask<StackAllocator>::DEFAULTTASKSTACKSIZE) :
#endif
// Wrap _func into _exit() to capture return value as exit code
BasicCoopTask<StackAllocator>(name, captureFuncReturn, stackSize), func(_func)
{
}
protected:
Result _exitCode = {};
static void captureFuncReturn() noexcept
{
#if !defined(ARDUINO)
try {
#endif
self()->_exitCode = self()->func();
#if !defined(ARDUINO)
}
catch (const Result code)
{
self()->_exitCode = code;
}
catch (...)
{
}
#endif
}
void _exit(Result&& code = Result{}) noexcept
{
_exitCode = std::move(code);
BasicCoopTask<StackAllocator>::_exit();
}
void _exit(const Result& code) noexcept
{
_exitCode = code;
BasicCoopTask<StackAllocator>::_exit();
}
private:
taskfunction_t func;
public:
/// @returns: The exit code is either the return value of of the task function, or set by using the exit() function.
Result exitCode() const noexcept { return _exitCode; }
/// @returns: a pointer to the CoopTask instance that is running. nullptr if not called from a CoopTask function (running() == false).
static CoopTask* self() noexcept { return static_cast<CoopTask*>(BasicCoopTask<StackAllocator>::self()); }
/// Use only in running CoopTask function. As stack unwinding is corrupted
/// by exit(), which among other issues breaks the RAII idiom,
/// using regular return or exceptions is to be preferred in most cases.
/// @param code default exit code is default value of CoopTask<>'s template argument, use exit() to set a different value.
static void exit(Result&& code = Result{}) noexcept { self()->_exit(std::move(code)); }
/// Use only in running CoopTask function. As stack unwinding is corrupted
/// by exit(), which among other issues breaks the RAII idiom,
/// using regular return or exceptions is to be preferred in most cases.
/// @param code default exit code is default value of CoopTask<>'s template argument, use exit() to set a different value.
static void exit(const Result& code) noexcept { self()->_exit(code); }
};
template<class StackAllocator> class CoopTask<void, StackAllocator> : public BasicCoopTask<StackAllocator>
{
public:
using CoopTaskBase::taskfunction_t;
#if defined(ARDUINO)
CoopTask(const String& name, CoopTaskBase::taskfunction_t func, size_t stackSize = BasicCoopTask<StackAllocator>::DEFAULTTASKSTACKSIZE) :
#else
CoopTask(const std::string& name, CoopTaskBase::taskfunction_t func, size_t stackSize = BasicCoopTask<StackAllocator>::DEFAULTTASKSTACKSIZE) :
#endif
BasicCoopTask<StackAllocator>(name, func, stackSize)
{
}
/// @returns: a pointer to the CoopTask instance that is running. nullptr if not called from a CoopTask function (running() == false).
static CoopTask* self() noexcept { return static_cast<CoopTask*>(BasicCoopTask<StackAllocator>::self()); }
};
/// A convenience function that creates a new CoopTask instance for the supplied task function, with the
/// given name and stack size, and schedules it.
/// @returns: the pointer to the new CoopTask instance, or nullptr if the creation or preparing for scheduling failed.
template<typename Result = int, class StackAllocator = CoopTaskStackAllocator>
CoopTask<Result, StackAllocator>* createCoopTask(
#if defined(ARDUINO)
const String& name, typename CoopTask<Result, StackAllocator>::taskfunction_t func, size_t stackSize = CoopTaskBase::DEFAULTTASKSTACKSIZE)
#else
const std::string& name, typename CoopTask<Result, StackAllocator>::taskfunction_t func, size_t stackSize = CoopTaskBase::DEFAULTTASKSTACKSIZE)
#endif
{
auto task = new CoopTask<Result, StackAllocator>(name, func, stackSize);
if (task && task->scheduleTask()) return task;
delete task;
return nullptr;
}
#endif // __CoopTask_h

965
arduino-cli/libraries/CoopTask/src/CoopTaskBase.cpp Normal file
View File

@@ -0,0 +1,965 @@
/*
CoopTaskBase.cpp - Implementation of cooperative scheduling tasks
Copyright (c) 2019 Dirk O. Kaar. All rights reserved.
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "CoopTaskBase.h"
#ifdef ARDUINO
#include <alloca.h>
#else
#include <chrono>
#endif
#if defined(ESP8266)
#include <Schedule.h>
#include <interrupts.h>
using esp8266::InterruptLock;
#elif !defined(ESP32) && defined(ARDUINO)
class InterruptLock {
public:
InterruptLock() {
noInterrupts();
}
~InterruptLock() {
interrupts();
}
};
#endif
extern "C" {
// Integration into global yield() and delay()
#if defined(ESP8266) || defined(ESP32)
void __yield();
void yield()
{
auto self = CoopTaskBase::self();
if (self) CoopTaskBase::yield(self);
else __yield();
}
#elif defined(ARDUINO)
void yield()
{
auto self = CoopTaskBase::self();
if (self) CoopTaskBase::yield(self);
}
#endif
#if defined(ESP8266)
void __delay(unsigned long ms);
void delay(unsigned long ms)
{
auto self = CoopTaskBase::self();
if (self) CoopTaskBase::delay(self, ms);
else __delay(ms);
}
#if defined(HAVE_ESP_SUSPEND)
void __esp_suspend();
// disable CONT suspend, resume by esp_schedule pattern
void esp_suspend()
{
auto self = CoopTaskBase::self();
if (self) CoopTaskBase::yield(self);
else __esp_suspend();
}
#else
void __esp_yield();
// disable CONT suspend, resume by esp_schedule pattern
void esp_yield()
{
auto self = CoopTaskBase::self();
if (self) CoopTaskBase::yield(self);
else __esp_yield();
}
#endif
void __esp_delay(unsigned long ms);
// disable CONT suspend, resume by esp_schedule pattern
void esp_delay(unsigned long ms)
{
auto self = CoopTaskBase::self();
if (self) CoopTaskBase::yield(self);
else __esp_delay(ms);
}
#elif defined(ESP32) && !defined(ESP32_FREERTOS)
void __delay(uint32_t ms);
void delay(uint32_t ms)
{
auto self = CoopTaskBase::self();
if (self) CoopTaskBase::delay(self, ms);
else __delay(ms);
}
#endif // ESP32_FREERTOS
}
std::array< std::atomic<CoopTaskBase* >, CoopTaskBase::MAXNUMBERCOOPTASKS + 1> CoopTaskBase::runnableTasks {};
std::atomic<size_t> CoopTaskBase::runnableTasksCount(0);
CoopTaskBase* CoopTaskBase::current = nullptr;
#ifndef ARDUINO
namespace
{
uint32_t millis()
{
return std::chrono::duration_cast<std::chrono::milliseconds>(std::chrono::system_clock::now().time_since_epoch()).count();
}
uint32_t micros()
{
return std::chrono::duration_cast<std::chrono::microseconds>(std::chrono::system_clock::now().time_since_epoch()).count();
}
void delayMicroseconds(uint32_t us)
{
const uint32_t start = micros();
while (micros() - start < us) {}
}
}
#elif defined(ESP8266) || defined(ESP32)
namespace
{
static const uint32_t CYCLES_PER_MS = ESP.getCpuFreqMHz() * 1000;
}
#endif
#if defined(ESP8266)
bool CoopTaskBase::usingBuiltinScheduler = false;
bool CoopTaskBase::rescheduleTask(uint32_t repeat_us)
{
auto stat = run();
if (sleeping()) return false;
switch (stat)
{
case -1: // exited.
return false;
break;
case 0: // runnable.
// rather keep scheduling at wrong delayed interval than drop altogether
if (repeat_us)
{
return !schedule_recurrent_function_us([this]() { return rescheduleTask(0); }, 0);
}
break;
default: // delayed for stat milliseconds or microseconds, check delayIsMs().
uint32_t next_repeat_us = delayIsMs() ? stat * 1000 : stat;
if (next_repeat_us > 26000000) next_repeat_us = 26000000;
if (next_repeat_us == repeat_us) break;
// rather keep scheduling at wrong interval than drop altogether
return !schedule_recurrent_function_us([this, next_repeat_us]() { return rescheduleTask(next_repeat_us); }, next_repeat_us, [this]() { return !delayed(); });
break;
}
return true;
}
#endif
bool IRAM_ATTR CoopTaskBase::enrollRunnable()
{
bool enrolled = false;
bool inserted = false;
for (size_t i = 0; i < runnableTasks.size(); ++i)
{
#if !defined(ESP32) && defined(ARDUINO)
InterruptLock lock;
auto task = runnableTasks[i].load();
if (!enrolled && nullptr == task)
{
runnableTasks[i].store(this);
enrolled = true;
inserted = true;
}
else if (this == task)
{
if (enrolled)
{
runnableTasks[i].store(nullptr);
inserted = false;
}
else
{
enrolled = true;
}
break;
}
}
if (inserted) runnableTasksCount.store(runnableTasksCount.load() + 1);
#else
CoopTaskBase* cmpTo = nullptr;
if (!enrolled && runnableTasks[i].compare_exchange_strong(cmpTo, this))
{
enrolled = true;
inserted = true;
}
else if (enrolled)
{
cmpTo = this;
if (runnableTasks[i].compare_exchange_strong(cmpTo, nullptr))
{
inserted = false;
break;
}
}
else if (this == runnableTasks[i].load())
{
enrolled = true;
break;
}
}
if (inserted) ++runnableTasksCount;
#endif
return enrolled;
}
void CoopTaskBase::delistRunnable()
{
#if !defined(ESP32) && defined(ARDUINO)
InterruptLock lock;
for (size_t i = 0; i < runnableTasks.size(); ++i)
{
if (runnableTasks[i].load() == this)
{
runnableTasks[i].store(nullptr);
runnableTasksCount.store(runnableTasksCount.load() - 1);
break;
}
}
#else
for (size_t i = 0; i < runnableTasks.size(); ++i)
{
CoopTaskBase* self = this;
if (runnableTasks[i].compare_exchange_strong(self, nullptr))
{
--runnableTasksCount;
break;
}
}
#endif
}
bool IRAM_ATTR CoopTaskBase::scheduleTask(bool wakeup)
{
if (!*this || !enrollRunnable()) return false;
#if defined(ESP8266)
bool reschedule = usingBuiltinScheduler && sleeping();
#endif
if (wakeup)
{
sleep(false);
}
#if defined(ESP8266)
return !reschedule || schedule_function([this]() { rescheduleTask(1); });
#else
return true;
#endif
}
#if defined(_MSC_VER)
CoopTaskBase::~CoopTaskBase()
{
if (taskFiber) DeleteFiber(taskFiber);
delistRunnable();
}
LPVOID CoopTaskBase::primaryFiber = nullptr;
void __stdcall CoopTaskBase::taskFiberFunc(void* self)
{
static_cast<CoopTaskBase*>(self)->func();
static_cast<CoopTaskBase*>(self)->_exit();
}
int32_t CoopTaskBase::initialize()
{
if (!cont || init) return -1;
init = true;
if (*this)
{
if (!primaryFiber) primaryFiber = ConvertThreadToFiber(nullptr);
if (primaryFiber)
{
taskFiber = CreateFiber(taskStackSize, taskFiberFunc, this);
if (taskFiber) return 0;
}
}
cont = false;
delistRunnable();
return -1;
}
int32_t CoopTaskBase::run()
{
if (!cont) return -1;
if (sleeps.load()) return 0;
if (delays.load())
{
if (delay_ms)
{
auto expired = millis() - delay_start;
if (expired < delay_duration)
{
auto delay_rem = delay_duration - expired;
return static_cast<int32_t>(delay_rem) < 0 ? DELAY_MAXINT : delay_rem;
}
}
else
{
auto expired = micros() - delay_start;
if (expired < delay_duration)
{
auto delay_rem = delay_duration - expired;
if (delay_rem >= DELAYMICROS_THRESHOLD)
{
return static_cast<int32_t>(delay_rem) < 0 ? DELAY_MAXINT : delay_rem;
}
::delayMicroseconds(delay_rem);
}
}
delays.store(false);
delay_duration = 0;
}
current = this;
if (!init && initialize() < 0) return -1;
SwitchToFiber(taskFiber);
current = nullptr;
// val = 0: init; -1: exit() task; 1: yield task; 2: sleep task; 3: delay task for delay_duration
cont = cont && (val > 0);
sleeps.store(sleeps.load() || (val == 2));
delays.store(delays.load() || (val > 2));
if (!cont) {
DeleteFiber(taskFiber);
taskFiber = NULL;
delistRunnable();
return -1;
}
switch (val)
{
case 1:
case 2:
return 0;
break;
case 3:
default:
return static_cast<int32_t>(delay_duration) < 0 ? DELAY_MAXINT : delay_duration;
break;
}
}
size_t CoopTaskBase::getFreeStack() const
{
return taskFiber ? taskStackSize : 0;
}
void CoopTaskBase::doYield(unsigned val) noexcept
{
self()->val = val;
SwitchToFiber(primaryFiber);
}
void CoopTaskBase::_delay(uint32_t ms) noexcept
{
delay_ms = true;
delay_start = millis();
delay_duration = ms;
// CoopTask::run() defers task for delay_duration milliseconds.
doYield(3);
}
void CoopTaskBase::_delayMicroseconds(uint32_t us) noexcept
{
if (us < DELAYMICROS_THRESHOLD) {
::delayMicroseconds(us);
return;
}
delay_ms = false;
delay_start = micros();
delay_duration = us;
// CoopTask::run() defers task for delay_duration microseconds.
doYield(3);
}
void CoopTaskBase::_exit() noexcept
{
self()->val = -1;
SwitchToFiber(primaryFiber);
}
void CoopTaskBase::_yield() noexcept
{
doYield(1);
}
void CoopTaskBase::_sleep() noexcept
{
doYield(2);
}
void IRAM_ATTR CoopTaskBase::sleep(const bool state) noexcept
{
sleeps.store(state);
if (!state)
{
delays.store(false);
delay_duration = 0;
}
}
#elif defined(ESP32_FREERTOS)
CoopTaskBase::~CoopTaskBase()
{
if (taskHandle) vTaskDelete(taskHandle);
taskHandle = nullptr;
delistRunnable();
}
void CoopTaskBase::taskFunc(void* _self)
{
static_cast<CoopTaskBase*>(_self)->func();
static_cast<CoopTaskBase*>(_self)->_exit();
}
int32_t CoopTaskBase::initialize()
{
if (!cont || init) return -1;
init = true;
if (*this)
{
xTaskCreateUniversal(taskFunc, name().c_str(), taskStackSize, this, tskIDLE_PRIORITY, &taskHandle, CONFIG_ARDUINO_RUNNING_CORE);
if (taskHandle)
{
vTaskSuspend(taskHandle);
return 0;
}
}
cont = false;
delistRunnable();
return -1;
}
int32_t CoopTaskBase::run()
{
if (!cont) return -1;
if (sleeps.load()) return 0;
if (delays.load())
{
if (delay_ms)
{
if (0 == delay_duration && eSuspended != eTaskGetState(taskHandle))
{
// fall through, blocked during FreeRTOS delay or asynchronously ready after is specifically handled below
}
else
{
auto expired = (ESP.getCycleCount() - delay_start) / CYCLES_PER_MS;
while (expired && delay_duration)
{
delay_start += CYCLES_PER_MS;
--delay_duration;
--expired;
}
if (expired < delay_duration)
{
auto delay_rem = delay_duration - expired;
return static_cast<int32_t>(delay_rem) < 0 ? DELAY_MAXINT : delay_rem;
}
delays.store(false);
delay_duration = 0;
}
}
else
{
auto expired = micros() - delay_start;
if (expired < delay_duration)
{
auto delay_rem = delay_duration - expired;
if (delay_rem >= DELAYMICROS_THRESHOLD)
{
return static_cast<int32_t>(delay_rem) < 0 ? DELAY_MAXINT : delay_rem;
}
::delayMicroseconds(delay_rem);
}
delays.store(false);
delay_duration = 0;
}
}
current = this;
if (!init)
{
if (initialize() < 0)
{
current = nullptr;
return -1;
}
}
bool resume = true;
for (;;)
{
auto taskState = eTaskGetState(taskHandle);
if (eSuspended == taskState)
{
if (!resume)
{
vTaskPrioritySet(taskHandle, tskIDLE_PRIORITY);
break;
}
resume = false;
vTaskPrioritySet(taskHandle, 1);
vTaskResume(taskHandle);
continue;
}
else if (eReady == taskState)
{
if (resume)
{
resume = false;
sleep(false);
delay_duration = 0;
vTaskPrioritySet(taskHandle, 1);
continue;
}
vPortYield();
continue;
}
else if (eBlocked == taskState)
{
if (resume)
{
if (!delays.load())
{
vTaskSuspend(taskHandle);
continue;
}
break;
}
vTaskPrioritySet(taskHandle, tskIDLE_PRIORITY);
if (!delays.exchange(true))
{
delay_ms = true;
delay_start = ESP.getCycleCount();
delay_duration = 0;
}
break;
}
else if (eDeleted == taskState)
{
cont = false;
break;
}
}
current = nullptr;
if (!cont) {
vTaskDelete(taskHandle);
taskHandle = nullptr;
delistRunnable();
return -1;
}
return static_cast<int32_t>(delay_duration) < 0 ? DELAY_MAXINT : delay_duration;
}
size_t CoopTaskBase::getFreeStack() const
{
return taskHandle ? uxTaskGetStackHighWaterMark(taskHandle) : 0;
}
void CoopTaskBase::_delay(uint32_t ms) noexcept
{
delays.store(true);
delay_ms = true;
delay_start = ESP.getCycleCount();
delay_duration = ms;
vTaskSuspend(taskHandle);
}
void CoopTaskBase::_delayMicroseconds(uint32_t us) noexcept
{
if (us < DELAYMICROS_THRESHOLD) {
::delayMicroseconds(us);
return;
}
delays.store(true);
delay_ms = false;
delay_start = micros();
delay_duration = us;
vTaskSuspend(taskHandle);
}
void CoopTaskBase::_sleep() noexcept
{
sleeps.store(true);
vTaskSuspend(taskHandle);
}
void CoopTaskBase::_yield() noexcept
{
delay_duration = 0;
delays.store(false);
vTaskSuspend(taskHandle);
}
void CoopTaskBase::_exit() noexcept
{
cont = false;
vTaskSuspend(taskHandle);
}
void IRAM_ATTR CoopTaskBase::sleep(const bool state) noexcept
{
sleeps.store(state);
if (!state)
{
delay_duration = 0;
delays.store(false);
}
}
CoopTaskBase* CoopTaskBase::self() noexcept
{
const auto currentTaskHandle = xTaskGetCurrentTaskHandle();
auto cur = current;
if (cur && currentTaskHandle == cur->taskHandle) return cur;
for (size_t i = 0; i < runnableTasks.size(); ++i)
{
cur = runnableTasks[i].load();
if (cur && currentTaskHandle == cur->taskHandle) return cur;
}
return nullptr;
}
#else
jmp_buf CoopTaskBase::env;
CoopTaskBase::~CoopTaskBase()
{
delistRunnable();
}
int32_t CoopTaskBase::initialize()
{
if (!cont || init) return -1;
init = true;
// fill stack with magic values to check overflow, corruption, and high water mark
for (size_t pos = 0; pos <= (taskStackSize + (FULLFEATURES ? sizeof(STACKCOOKIE) : 0)) / sizeof(STACKCOOKIE); ++pos)
{
reinterpret_cast<unsigned*>(taskStackTop)[pos] = STACKCOOKIE;
}
#if defined(__GNUC__) && (defined(__amd64__) || defined(__amd64) || defined(__x86_64__) || defined(__x86_64))
asm volatile (
"movq %0, %%rsp"
:
: "r" (((reinterpret_cast<long unsigned>(taskStackTop) + taskStackSize + (FULLFEATURES ? sizeof(STACKCOOKIE) : 0)) >> 4) << 4)
);
#elif defined(ARDUINO) || defined(__GNUC__)
char* bp = static_cast<char*>(alloca(
reinterpret_cast<long unsigned>(&bp) - reinterpret_cast<long unsigned>(taskStackTop) - (taskStackSize + (FULLFEATURES ? sizeof(STACKCOOKIE) : 0))
));
std::atomic_thread_fence(std::memory_order_release);
#else
#error Setting stack pointer is not implemented on this target
#endif
func();
self()->_exit();
cont = false;
delistRunnable();
return -1;
}
int32_t CoopTaskBase::run()
{
if (!cont) return -1;
if (sleeps.load()) return 0;
if (delays.load())
{
if (delay_ms)
{
#if defined(ESP8266) || defined(ESP32)
uint32_t expired;
#ifdef ESP8266
if (usingBuiltinScheduler)
{
expired = millis() - delay_start;
}
else
#endif
{
expired = (ESP.getCycleCount() - delay_start) / CYCLES_PER_MS;
while (expired && delay_duration)
{
delay_start += CYCLES_PER_MS;
--delay_duration;
--expired;
}
}
#else
auto expired = millis() - delay_start;
#endif
if (expired < delay_duration)
{
auto delay_rem = delay_duration - expired;
return static_cast<int32_t>(delay_rem) < 0 ? DELAY_MAXINT : delay_rem;
}
}
else
{
auto expired = micros() - delay_start;
if (expired < delay_duration)
{
auto delay_rem = delay_duration - expired;
if (delay_rem >= DELAYMICROS_THRESHOLD)
{
return static_cast<int32_t>(delay_rem) < 0 ? DELAY_MAXINT : delay_rem;
}
::delayMicroseconds(delay_rem);
}
}
delays.store(false);
delay_duration = 0;
}
auto val = setjmp(env);
// val = 0: init; -1: exit() task; 1: yield task; 2: sleep task; 3: delay task for delay_duration
if (!val) {
current = this;
if (!init) return initialize();
if (FULLFEATURES && *reinterpret_cast<unsigned*>(taskStackTop + taskStackSize + sizeof(STACKCOOKIE)) != STACKCOOKIE)
{
#ifndef ARDUINO_attiny
::printf(PSTR("FATAL ERROR: CoopTask %s stack corrupted\n"), name().c_str());
#endif
::abort();
}
longjmp(env_yield, 1);
}
else
{
current = nullptr;
if (*reinterpret_cast<unsigned*>(taskStackTop) != STACKCOOKIE)
{
#ifndef ARDUINO_attiny
::printf(PSTR("FATAL ERROR: CoopTask %s stack overflow\n"), name().c_str());
#endif
::abort();
}
cont = cont && (val > 0);
sleeps.store(sleeps.load() || (val == 2));
delays.store(delays.load() || (val > 2));
}
if (!cont) {
delistRunnable();
return -1;
}
switch (val)
{
case 1:
case 2:
return 0;
break;
case 3:
default:
return static_cast<int32_t>(delay_duration) < 0 ? DELAY_MAXINT : delay_duration;
break;
}
}
void CoopTaskBase::dumpStack() const
{
if (!taskStackTop) return;
size_t pos;
for (pos = 1; pos < (taskStackSize + (FULLFEATURES ? sizeof(STACKCOOKIE) : 0)) / sizeof(STACKCOOKIE); ++pos)
{
if (STACKCOOKIE != reinterpret_cast<unsigned*>(taskStackTop)[pos])
break;
}
#ifndef ARDUINO_attiny
::printf(PSTR(">>>stack>>>\n"));
#endif
while (pos < (taskStackSize + (FULLFEATURES ? sizeof(STACKCOOKIE) : 0)) / sizeof(STACKCOOKIE))
{
#ifndef ARDUINO_attiny
auto* sp = &reinterpret_cast<unsigned*>(taskStackTop)[pos];
// rough indicator: stack frames usually have SP saved as the second word
bool looksLikeStackFrame = (sp[2] == reinterpret_cast<size_t>(&sp[4]));
::printf(PSTR("%08x: %08x %08x %08x %08x %c\n"),
reinterpret_cast<size_t>(sp), sp[0], sp[1], sp[2], sp[3], looksLikeStackFrame ? '<' : ' ');
#endif
pos += 4;
}
#ifndef ARDUINO_attiny
::printf(PSTR("<<<stack<<<\n"));
#endif
}
size_t CoopTaskBase::getFreeStack() const
{
if (!taskStackTop) return 0;
size_t pos;
for (pos = 1; pos < (taskStackSize + (FULLFEATURES ? sizeof(STACKCOOKIE) : 0)) / sizeof(STACKCOOKIE); ++pos)
{
if (STACKCOOKIE != reinterpret_cast<unsigned*>(taskStackTop)[pos])
break;
}
return (pos - 1) * sizeof(unsigned);
}
void CoopTaskBase::doYield(unsigned val) noexcept
{
if (!setjmp(env_yield))
{
longjmp(env, val);
}
}
void CoopTaskBase::_delay(uint32_t ms) noexcept
{
delay_ms = true;
#ifdef ESP8266
delay_start = usingBuiltinScheduler ? millis() : ESP.getCycleCount();
#elif ESP32
delay_start = ESP.getCycleCount();
#else
delay_start = millis();
#endif
delay_duration = ms;
// CoopTask::run() defers task for delay_duration milliseconds.
doYield(3);
}
void CoopTaskBase::_delayMicroseconds(uint32_t us) noexcept
{
if (us < DELAYMICROS_THRESHOLD) {
::delayMicroseconds(us);
return;
}
delay_ms = false;
delay_start = micros();
delay_duration = us;
// CoopTask::run() defers task for delay_duration microseconds.
doYield(3);
}
void CoopTaskBase::_exit() noexcept
{
longjmp(env, -1);
}
void CoopTaskBase::_yield() noexcept
{
doYield(1);
}
void CoopTaskBase::_sleep() noexcept
{
doYield(2);
}
void IRAM_ATTR CoopTaskBase::sleep(const bool state) noexcept
{
sleeps.store(state);
if (!state)
{
delays.store(false);
delay_duration = 0;
}
}
#endif // _MSC_VER
void runCoopTasks(const Delegate<void(const CoopTaskBase* const task)>& reaper,
const Delegate<bool(uint32_t ms)>& onDelay, const Delegate<bool()>& onSleep)
{
#ifdef ESP32_FREERTOS
static TaskHandle_t yieldGuardHandle = nullptr;
if (!yieldGuardHandle)
{
xTaskCreateUniversal([](void*)
{
for (;;)
{
vPortYield();
}
}, "YieldGuard", 0x200, nullptr, 1, &yieldGuardHandle, CONFIG_ARDUINO_RUNNING_CORE);
}
#endif
auto taskCount = CoopTaskBase::getRunnableTasksCount();
bool allSleeping = true;
uint32_t minDelay_ms = ~(decltype(minDelay_ms))0U;
for (size_t i = 0; taskCount && i < CoopTaskBase::getRunnableTasks().size(); ++i)
{
#if defined(ESP8266) || defined(ESP32)
optimistic_yield(10000);
#endif
auto task = CoopTaskBase::getRunnableTasks()[i].load();
if (task)
{
--taskCount;
auto runResult = task->run();
if (runResult < 0 && reaper)
reaper(task);
else if (minDelay_ms)
{
if (task->delayed())
{
allSleeping = false;
uint32_t delay_ms = task->delayIsMs() ? static_cast<uint32_t>(runResult) : static_cast<uint32_t>(runResult) / 1000UL;
if (delay_ms < minDelay_ms)
minDelay_ms = delay_ms;
}
else if (!task->sleeping())
{
allSleeping = false;
minDelay_ms = 0;
}
}
}
}
bool cleanup = true;
if (allSleeping && onSleep)
{
cleanup = onSleep();
}
else if (minDelay_ms && onDelay)
{
cleanup = onDelay(minDelay_ms);
}
if (cleanup)
{
#ifdef ESP32_FREERTOS
vTaskSuspend(yieldGuardHandle);
vTaskDelay(1);
vTaskResume(yieldGuardHandle);
#endif
}
}

270
arduino-cli/libraries/CoopTask/src/CoopTaskBase.h Normal file
View File

@@ -0,0 +1,270 @@
/*
CoopTaskBase.h - Implementation of cooperative scheduling tasks
Copyright (c) 2019 Dirk O. Kaar. All rights reserved.
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#ifndef __CoopTaskBase_h
#define __CoopTaskBase_h
#ifdef ESP32
#define ESP32_FREERTOS
#endif
#include "circular_queue/Delegate.h"
#if defined(ESP8266) || defined(ESP32)
#include <array>
#include <memory>
#include <csetjmp>
#include <Arduino.h>
#elif defined(ARDUINO)
#include <setjmp.h>
#include <Arduino.h>
#elif defined(_MSC_VER)
#include <array>
#include <Windows.h>
#include <string>
#else
#include <array>
#include <csetjmp>
#include <string>
#endif
#if defined(ESP8266) || defined(ESP32) || !defined(ARDUINO)
#include <atomic>
#else
#include "circular_queue/ghostl.h"
#endif
#if !defined(ESP32) && !defined(ESP8266)
#define IRAM_ATTR
#endif
#ifdef _MSC_VER
#define __attribute__(_)
#endif
class CoopTaskBase
{
public:
static constexpr bool FULLFEATURES = sizeof(unsigned) >= 4;
protected:
using taskfunction_t = Delegate< void() >;
#ifdef ARDUINO
CoopTaskBase(const String& name, taskfunction_t _func, size_t stackSize = DEFAULTTASKSTACKSIZE) :
#else
CoopTaskBase(const std::string& name, taskfunction_t _func, size_t stackSize = DEFAULTTASKSTACKSIZE) :
#endif
taskName(name), sleeps(true), delays(false), func(_func)
{
taskStackSize = (sizeof(unsigned) >= 4) ? ((stackSize + sizeof(unsigned) - 1) / sizeof(unsigned)) * sizeof(unsigned) : stackSize;
}
CoopTaskBase(const CoopTaskBase&) = delete;
CoopTaskBase& operator=(const CoopTaskBase&) = delete;
static constexpr int32_t DELAYMICROS_THRESHOLD = 50;
static constexpr uint32_t DELAY_MAXINT = (~(uint32_t)0) >> 1;
#ifdef ARDUINO
const String taskName;
#else
const std::string taskName;
#endif
size_t taskStackSize;
#if defined(_MSC_VER)
static LPVOID primaryFiber;
LPVOID taskFiber = nullptr;
int val = 0;
static void __stdcall taskFiberFunc(void* self);
#elif defined(ESP32_FREERTOS)
TaskHandle_t taskHandle = nullptr;
static void taskFunc(void* _self);
#else
char* taskStackTop = nullptr;
static jmp_buf env;
jmp_buf env_yield;
#endif
static constexpr size_t MAXNUMBERCOOPTASKS = FULLFEATURES ? 32 : 8;
// for lock-free insertion, must be one element larger than max task count
static std::array< std::atomic<CoopTaskBase* >, MAXNUMBERCOOPTASKS + 1> runnableTasks;
static std::atomic<size_t> runnableTasksCount;
static CoopTaskBase* current;
bool init = false;
bool cont = true;
std::atomic<bool> sleeps;
// ESP32 FreeRTOS (#define ESP32_FREERTOS) handles delays, on this platfrom delays is always false
std::atomic<bool> delays;
int32_t initialize();
void doYield(unsigned val) noexcept;
#if defined(ESP8266)
static bool usingBuiltinScheduler;
bool rescheduleTask(uint32_t repeat_us);
#endif
bool IRAM_ATTR enrollRunnable();
void delistRunnable();
void _exit() noexcept;
void _yield() noexcept;
void _sleep() noexcept;
void _delay(uint32_t ms) noexcept;
void _delayMicroseconds(uint32_t us) noexcept;
private:
// true: delay_start/delay_duration are in milliseconds; false: delay_start/delay_duration are in microseconds.
bool delay_ms = false;
uint32_t delay_start = 0;
uint32_t delay_duration = 0;
taskfunction_t func;
public:
virtual ~CoopTaskBase();
#if defined(ESP32)
static constexpr size_t MAXSTACKSPACE = 0x2000;
#elif defined(ESP8266)
static constexpr size_t MAXSTACKSPACE = 0x1000;
#elif defined(ARDUINO)
static constexpr size_t MAXSTACKSPACE = FULLFEATURES ? 0x180 : 0xc0;
#else
static constexpr size_t MAXSTACKSPACE = 0x10000;
#endif
static constexpr unsigned STACKCOOKIE = FULLFEATURES ? 0xdeadbeefUL : 0xdeadU;
static constexpr size_t DEFAULTTASKSTACKSIZE = MAXSTACKSPACE - (FULLFEATURES ? 2 : 1) * sizeof(STACKCOOKIE);
#ifdef ARDUINO
const String& name() const noexcept { return taskName; }
#else
const std::string& name() const noexcept { return taskName; }
#endif
/// @returns: true if the CoopTask object is ready to run, including stack allocation.
/// false if either initialization has failed, or the task has exited().
#if !defined(_MSC_VER) && !defined(ESP32_FREERTOS)
operator bool() const noexcept { return cont && taskStackTop; }
/// Prints the task stack, decodable by the ESP exception decoder
void dumpStack() const;
#else
operator bool() const noexcept { return cont; }
#endif
/// Ready the task for scheduling, by default waking up the task from both sleep and delay.
/// @returns: true on success.
bool IRAM_ATTR scheduleTask(bool wakeup = true);
inline bool IRAM_ATTR wakeup() __attribute__((always_inline)) { return scheduleTask(true); }
#ifdef ESP8266
/// For full access to all features, cyclic task scheduling, state evaluation
/// and running are performed explicitly from user code. For convenience, the function
/// runCoopTasks() implements the pattern as best practice. See the CoopTask examples for this.
/// If only a pre-determined number of loop tasks need to run indefinitely
/// without exit codes or explicit deep sleep system states, on platforms where a
/// scheduler exists that is suffiently capable to iteratively run each of these CoopTasks,
/// calling this function switches all task creation and scheduling to using that, obviating
/// the need to call a scheduler explicitly from user code.
/// The scheduler selection should be done before the first CoopTask is created, and not
/// changed thereafter during runtime.
/// By default, builtin schedulers are not used, for well-defined behavior and portability.
/// @param state true: The parameter default value. All subsequent scheduling of tasks is
/// handed to the builtin scheduler.
static void useBuiltinScheduler(bool state = true)
{
usingBuiltinScheduler = state;
}
#endif
/// Every task is entered into this list by scheduleTask(). It is removed when it exits
/// or gets deleted.
static const decltype(runnableTasks)& getRunnableTasks()
{
return runnableTasks;
}
/// @returns: the count of runnable, non-nullptr, tasks in the return of getRunnableTasks().
static size_t getRunnableTasksCount()
{
return runnableTasksCount.load();
}
/// @returns: -1: exited. 0: runnable or sleeping. >0: delayed for milliseconds or microseconds, check delayIsMs().
int32_t run();
/// @returns: size of unused stack space. 0 if stack is not allocated yet or was deleted after task exited.
size_t getFreeStack() const;
bool delayIsMs() const noexcept { return delay_ms; }
/// Modifies the sleep flag. if called from a running task, it is not immediately suspended.
/// @param state true: a suspended task becomes sleeping, if call from the running task,
/// the next call to yield() or delay() puts it into sleeping state.
/// false: clears the sleeping and delay state of the task.
void IRAM_ATTR sleep(const bool state) noexcept;
#ifdef ESP32_FREERTOS
/// @returns: a pointer to the CoopTask instance that is running. nullptr if not called from a CoopTask function (running() == false).
static CoopTaskBase* self() noexcept;
#else
/// @returns: a pointer to the CoopTask instance that is running. nullptr if not called from a CoopTask function (running() == false).
static CoopTaskBase* self() noexcept { return current; }
#endif
/// @returns: true if called from the task function of a CoopTask, false otherwise.
static bool running() noexcept { return self(); }
/// @returns: true if the task's is set to sleep.
/// For a non-running task, this implies it is also currently not scheduled.
inline bool IRAM_ATTR sleeping() const noexcept __attribute__((always_inline)) { return sleeps.load(); }
inline bool IRAM_ATTR delayed() const noexcept __attribute__((always_inline)) { return delays.load(); }
inline bool IRAM_ATTR suspended() const noexcept __attribute__((always_inline)) { return sleeps.load() || delays.load(); }
/// use only in running CoopTask function. As stack unwinding is corrupted
/// by exit(), which among other issues breaks the RAII idiom,
/// using regular return or exceptions is to be preferred in most cases.
static void exit() noexcept { self()->_exit(); }
/// use only in running CoopTask function.
static void yield() noexcept { self()->_yield(); }
static void yield(CoopTaskBase* self) noexcept { self->_yield(); }
/// use only in running CoopTask function.
static void sleep() noexcept { self()->_sleep(); }
/// use only in running CoopTask function.
static void delay(uint32_t ms) noexcept { self()->_delay(ms); }
static void delay(CoopTaskBase* self, uint32_t ms) noexcept { self->_delay(ms); }
/// use only in running CoopTask function.
static void delayMicroseconds(uint32_t us) noexcept { self()->_delayMicroseconds(us); }
};
#ifndef ARDUINO
inline void yield() { CoopTaskBase::yield(); }
inline void delay(uint32_t ms) { CoopTaskBase::delay(ms); }
#endif
/// An optional convenience funtion that does all the work to cyclically perform CoopTask execution.
/// @param reaper An optional function that is called once when a task exits.
/// @param onDelay An optional function to handle a global delay greater or equal 1 millisecond, resulting
/// from the minimum time interval for which at this time all CoopTasks are delayed.
/// This can be used for power saving, if wake up by asynchronous events is properly considered.
/// If onSleep is not set, onDelay() is called instead with the maximum value for the ms delay parameter.
/// onDelay() must return a bool value, if true, runCoopTasks performs the default housekeeping actions,
/// otherwise it skips those.
/// @param onSleep An optional function indicating that all CoopTasks are sleeping, that is, are infinitely delayed.
/// This can be used for power saving modes.
/// onSleep(), like onDelay(), must return a bool value, if true, runCoopTasks performs the
/// default housekeeping actions, otherwise it skips those.
void runCoopTasks(const Delegate<void(const CoopTaskBase* const task)>& reaper = nullptr,
const Delegate<bool(uint32_t ms)>& onDelay = nullptr, const Delegate<bool()>& onSleep = nullptr);
#endif // __CoopTaskBase_h

2130
arduino-cli/libraries/CoopTask/src/circular_queue/Delegate.h Normal file
View File

File diff suppressed because it is too large Load Diff

567
arduino-cli/libraries/CoopTask/src/circular_queue/MultiDelegate.h Normal file
View File

@@ -0,0 +1,567 @@
/*
MultiDelegate.h - A queue or event multiplexer based on the efficient Delegate
class
Copyright (c) 2019-2020 Dirk O. Kaar. All rights reserved.
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#ifndef __MULTIDELEGATE_H
#define __MULTIDELEGATE_H
#include <iterator>
#if defined(ESP8266) || defined(ESP32) || !defined(ARDUINO)
#include <atomic>
#else
#include "circular_queue/ghostl.h"
#endif
#if defined(ESP8266)
#include <interrupts.h>
using esp8266::InterruptLock;
#elif defined(ARDUINO)
class InterruptLock {
public:
InterruptLock() {
noInterrupts();
}
~InterruptLock() {
interrupts();
}
};
#else
#include <mutex>
#endif
namespace
{
template< typename Delegate, typename R, bool ISQUEUE = false, typename... P>
struct CallP
{
static R execute(Delegate& del, P... args)
{
return del(std::forward<P...>(args...));
}
};
template< typename Delegate, bool ISQUEUE, typename... P>
struct CallP<Delegate, void, ISQUEUE, P...>
{
static bool execute(Delegate& del, P... args)
{
del(std::forward<P...>(args...));
return true;
}
};
template< typename Delegate, typename R, bool ISQUEUE = false>
struct Call
{
static R execute(Delegate& del)
{
return del();
}
};
template< typename Delegate, bool ISQUEUE>
struct Call<Delegate, void, ISQUEUE>
{
static bool execute(Delegate& del)
{
del();
return true;
}
};
}
namespace delegate
{
namespace detail
{
template< typename Delegate, typename R, bool ISQUEUE = false, size_t QUEUE_CAPACITY = 32, typename... P>
class MultiDelegatePImpl
{
public:
MultiDelegatePImpl() = default;
~MultiDelegatePImpl()
{
*this = nullptr;
}
MultiDelegatePImpl(const MultiDelegatePImpl&) = delete;
MultiDelegatePImpl& operator=(const MultiDelegatePImpl&) = delete;
MultiDelegatePImpl(MultiDelegatePImpl&& md)
{
first = md.first;
last = md.last;
unused = md.unused;
nodeCount = md.nodeCount;
md.first = nullptr;
md.last = nullptr;
md.unused = nullptr;
md.nodeCount = 0;
}
MultiDelegatePImpl(const Delegate& del)
{
add(del);
}
MultiDelegatePImpl(Delegate&& del)
{
add(std::move(del));
}
MultiDelegatePImpl& operator=(MultiDelegatePImpl&& md)
{
first = md.first;
last = md.last;
unused = md.unused;
nodeCount = md.nodeCount;
md.first = nullptr;
md.last = nullptr;
md.unused = nullptr;
md.nodeCount = 0;
return *this;
}
MultiDelegatePImpl& operator=(std::nullptr_t)
{
if (last)
last->mNext = unused;
if (first)
unused = first;
while (unused)
{
auto to_delete = unused;
unused = unused->mNext;
delete(to_delete);
}
return *this;
}
MultiDelegatePImpl& operator+=(const Delegate& del)
{
add(del);
return *this;
}
MultiDelegatePImpl& operator+=(Delegate&& del)
{
add(std::move(del));
return *this;
}
protected:
struct Node_t
{
~Node_t()
{
mDelegate = nullptr; // special overload in Delegate
}
Node_t* mNext = nullptr;
Delegate mDelegate;
};
Node_t* first = nullptr;
Node_t* last = nullptr;
Node_t* unused = nullptr;
size_t nodeCount = 0;
// Returns a pointer to an unused Node_t,
// or if none are available allocates a new one,
// or nullptr if limit is reached
Node_t* IRAM_ATTR get_node_unsafe()
{
Node_t* result = nullptr;
// try to get an item from unused items list
if (unused)
{
result = unused;
unused = unused->mNext;
}
// if no unused items, and count not too high, allocate a new one
else if (nodeCount < QUEUE_CAPACITY)
{
#if defined(ESP8266) || defined(ESP32)
result = new (std::nothrow) Node_t;
#else
result = new Node_t;
#endif
if (result)
++nodeCount;
}
return result;
}
void recycle_node_unsafe(Node_t* node)
{
node->mDelegate = nullptr; // special overload in Delegate
node->mNext = unused;
unused = node;
}
#ifndef ARDUINO
std::mutex mutex_unused;
#endif
public:
class iterator : public std::iterator<std::forward_iterator_tag, Delegate>
{
public:
Node_t* current = nullptr;
Node_t* prev = nullptr;
const Node_t* stop = nullptr;
iterator(MultiDelegatePImpl& md) : current(md.first), stop(md.last) {}
iterator() = default;
iterator(const iterator&) = default;
iterator& operator=(const iterator&) = default;
iterator& operator=(iterator&&) = default;
operator bool() const
{
return current && stop;
}
bool operator==(const iterator& rhs) const
{
return current == rhs.current;
}
bool operator!=(const iterator& rhs) const
{
return !operator==(rhs);
}
Delegate& operator*() const
{
return current->mDelegate;
}
Delegate* operator->() const
{
return &current->mDelegate;
}
iterator& operator++() // prefix
{
if (current && stop != current)
{
prev = current;
current = current->mNext;
}
else
current = nullptr; // end
return *this;
}
iterator& operator++(int) // postfix
{
iterator tmp(*this);
operator++();
return tmp;
}
};
iterator begin()
{
return iterator(*this);
}
iterator end() const
{
return iterator();
}
const Delegate* IRAM_ATTR add(const Delegate& del)
{
return add(Delegate(del));
}
const Delegate* IRAM_ATTR add(Delegate&& del)
{
if (!del)
return nullptr;
#ifdef ARDUINO
InterruptLock lockAllInterruptsInThisScope;
#else
std::lock_guard<std::mutex> lock(mutex_unused);
#endif
Node_t* item = ISQUEUE ? get_node_unsafe() :
#if defined(ESP8266) || defined(ESP32)
new (std::nothrow) Node_t;
#else
new Node_t;
#endif
if (!item)
return nullptr;
item->mDelegate = std::move(del);
item->mNext = nullptr;
if (last)
last->mNext = item;
else
first = item;
last = item;
return &item->mDelegate;
}
iterator erase(iterator it)
{
if (!it)
return end();
#ifdef ARDUINO
InterruptLock lockAllInterruptsInThisScope;
#else
std::lock_guard<std::mutex> lock(mutex_unused);
#endif
auto to_recycle = it.current;
if (last == it.current)
last = it.prev;
it.current = it.current->mNext;
if (it.prev)
{
it.prev->mNext = it.current;
}
else
{
first = it.current;
}
if (ISQUEUE)
recycle_node_unsafe(to_recycle);
else
delete to_recycle;
return it;
}
bool erase(const Delegate* const del)
{
auto it = begin();
while (it)
{
if (del == &(*it))
{
erase(it);
return true;
}
++it;
}
return false;
}
operator bool() const
{
return first;
}
R operator()(P... args)
{
auto it = begin();
if (!it)
return {};
static std::atomic<bool> fence(false);
// prevent recursive calls
#if defined(ARDUINO) && !defined(ESP32)
if (fence.load()) return {};
fence.store(true);
#else
if (fence.exchange(true)) return {};
#endif
R result;
do
{
result = CallP<Delegate, R, ISQUEUE, P...>::execute(*it, args...);
if (result && ISQUEUE)
it = erase(it);
else
++it;
#if defined(ESP8266) || defined(ESP32)
// running callbacks might last too long for watchdog etc.
optimistic_yield(10000);
#endif
} while (it);
fence.store(false);
return result;
}
};
template< typename Delegate, typename R = void, bool ISQUEUE = false, size_t QUEUE_CAPACITY = 32>
class MultiDelegateImpl : public MultiDelegatePImpl<Delegate, R, ISQUEUE, QUEUE_CAPACITY>
{
public:
using MultiDelegatePImpl<Delegate, R, ISQUEUE, QUEUE_CAPACITY>::MultiDelegatePImpl;
R operator()()
{
auto it = this->begin();
if (!it)
return {};
static std::atomic<bool> fence(false);
// prevent recursive calls
#if defined(ARDUINO) && !defined(ESP32)
if (fence.load()) return {};
fence.store(true);
#else
if (fence.exchange(true)) return {};
#endif
R result;
do
{
result = Call<Delegate, R, ISQUEUE>::execute(*it);
if (result && ISQUEUE)
it = this->erase(it);
else
++it;
#if defined(ESP8266) || defined(ESP32)
// running callbacks might last too long for watchdog etc.
optimistic_yield(10000);
#endif
} while (it);
fence.store(false);
return result;
}
};
template< typename Delegate, typename R, bool ISQUEUE, size_t QUEUE_CAPACITY, typename... P> class MultiDelegate;
template< typename Delegate, typename R, bool ISQUEUE, size_t QUEUE_CAPACITY, typename... P>
class MultiDelegate<Delegate, R(P...), ISQUEUE, QUEUE_CAPACITY> : public MultiDelegatePImpl<Delegate, R, ISQUEUE, QUEUE_CAPACITY, P...>
{
public:
using MultiDelegatePImpl<Delegate, R, ISQUEUE, QUEUE_CAPACITY, P...>::MultiDelegatePImpl;
};
template< typename Delegate, typename R, bool ISQUEUE, size_t QUEUE_CAPACITY>
class MultiDelegate<Delegate, R(), ISQUEUE, QUEUE_CAPACITY> : public MultiDelegateImpl<Delegate, R, ISQUEUE, QUEUE_CAPACITY>
{
public:
using MultiDelegateImpl<Delegate, R, ISQUEUE, QUEUE_CAPACITY>::MultiDelegateImpl;
};
template< typename Delegate, bool ISQUEUE, size_t QUEUE_CAPACITY, typename... P>
class MultiDelegate<Delegate, void(P...), ISQUEUE, QUEUE_CAPACITY> : public MultiDelegatePImpl<Delegate, void, ISQUEUE, QUEUE_CAPACITY, P...>
{
public:
using MultiDelegatePImpl<Delegate, void, ISQUEUE, QUEUE_CAPACITY, P...>::MultiDelegatePImpl;
void operator()(P... args)
{
auto it = this->begin();
if (!it)
return;
static std::atomic<bool> fence(false);
// prevent recursive calls
#if defined(ARDUINO) && !defined(ESP32)
if (fence.load()) return;
fence.store(true);
#else
if (fence.exchange(true)) return;
#endif
do
{
CallP<Delegate, void, ISQUEUE, P...>::execute(*it, args...);
if (ISQUEUE)
it = this->erase(it);
else
++it;
#if defined(ESP8266) || defined(ESP32)
// running callbacks might last too long for watchdog etc.
optimistic_yield(10000);
#endif
} while (it);
fence.store(false);
}
};
template< typename Delegate, bool ISQUEUE, size_t QUEUE_CAPACITY>
class MultiDelegate<Delegate, void(), ISQUEUE, QUEUE_CAPACITY> : public MultiDelegateImpl<Delegate, void, ISQUEUE, QUEUE_CAPACITY>
{
public:
using MultiDelegateImpl<Delegate, void, ISQUEUE, QUEUE_CAPACITY>::MultiDelegateImpl;
void operator()()
{
auto it = this->begin();
if (!it)
return;
static std::atomic<bool> fence(false);
// prevent recursive calls
#if defined(ARDUINO) && !defined(ESP32)
if (fence.load()) return;
fence.store(true);
#else
if (fence.exchange(true)) return;
#endif
do
{
Call<Delegate, void, ISQUEUE>::execute(*it);
if (ISQUEUE)
it = this->erase(it);
else
++it;
#if defined(ESP8266) || defined(ESP32)
// running callbacks might last too long for watchdog etc.
optimistic_yield(10000);
#endif
} while (it);
fence.store(false);
}
};
}
}
/**
The MultiDelegate class template can be specialized to either a queue or an event multiplexer.
It is designed to be used with Delegate, the efficient runtime wrapper for C function ptr and C++ std::function.
@tparam Delegate specifies the concrete type that MultiDelegate bases the queue or event multiplexer on.
@tparam ISQUEUE modifies the generated MultiDelegate class in subtle ways. In queue mode (ISQUEUE == true),
the value of QUEUE_CAPACITY enforces the maximum number of simultaneous items the queue can contain.
This is exploited to minimize the use of new and delete by reusing already allocated items, thus
reducing heap fragmentation. In event multiplexer mode (ISQUEUE = false), new and delete are
used for allocation of the event handler items.
If the result type of the function call operator of Delegate is void, calling a MultiDelegate queue
removes each item after calling it; a Multidelegate event multiplexer keeps event handlers until
explicitly removed.
If the result type of the function call operator of Delegate is non-void, in a MultiDelegate queue
the type-conversion to bool of that result determines if the item is immediately removed or kept
after each call: if true is returned, the item is removed. A Multidelegate event multiplexer keeps event
handlers until they are explicitly removed.
@tparam QUEUE_CAPACITY is only used if ISQUEUE == true. Then, it sets the maximum capacity that the queue dynamically
allocates from the heap. Unused items are not returned to the heap, but are managed by the MultiDelegate
instance during its own lifetime for efficiency.
*/
template< typename Delegate, bool ISQUEUE = false, size_t QUEUE_CAPACITY = 32>
class MultiDelegate : public delegate::detail::MultiDelegate<Delegate, typename Delegate::target_type, ISQUEUE, QUEUE_CAPACITY>
{
public:
using delegate::detail::MultiDelegate<Delegate, typename Delegate::target_type, ISQUEUE, QUEUE_CAPACITY>::MultiDelegate;
};
#endif // __MULTIDELEGATE_H

393
arduino-cli/libraries/CoopTask/src/circular_queue/circular_queue.h Normal file
View File

@@ -0,0 +1,393 @@
/*
circular_queue.h - Implementation of a lock-free circular queue for EspSoftwareSerial.
Copyright (c) 2019 Dirk O. Kaar. All rights reserved.
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#ifndef __circular_queue_h
#define __circular_queue_h
#ifdef ARDUINO
#include <Arduino.h>
#endif
#if defined(ESP8266) || defined(ESP32) || !defined(ARDUINO)
#include <atomic>
#include <memory>
#include <algorithm>
#include "Delegate.h"
using std::min;
#else
#include "ghostl.h"
#endif
#if !defined(ESP32) && !defined(ESP8266)
#define IRAM_ATTR
#endif
/*!
@brief Instance class for a single-producer, single-consumer circular queue / ring buffer (FIFO).
This implementation is lock-free between producer and consumer for the available(), peek(),
pop(), and push() type functions.
*/
template< typename T, typename ForEachArg = void >
class circular_queue
{
public:
/*!
@brief Constructs a valid, but zero-capacity dummy queue.
*/
circular_queue() : m_bufSize(1)
{
m_inPos.store(0);
m_outPos.store(0);
}
/*!
@brief Constructs a queue of the given maximum capacity.
*/
circular_queue(const size_t capacity) : m_bufSize(capacity + 1), m_buffer(new T[m_bufSize])
{
m_inPos.store(0);
m_outPos.store(0);
}
circular_queue(circular_queue&& cq) :
m_bufSize(cq.m_bufSize), m_buffer(cq.m_buffer), m_inPos(cq.m_inPos.load()), m_outPos(cq.m_outPos.load())
{}
~circular_queue()
{
m_buffer.reset();
}
circular_queue(const circular_queue&) = delete;
circular_queue& operator=(circular_queue&& cq)
{
m_bufSize = cq.m_bufSize;
m_buffer = cq.m_buffer;
m_inPos.store(cq.m_inPos.load());
m_outPos.store(cq.m_outPos.load());
}
circular_queue& operator=(const circular_queue&) = delete;
/*!
@brief Get the numer of elements the queue can hold at most.
*/
size_t capacity() const
{
return m_bufSize - 1;
}
/*!
@brief Resize the queue. The available elements in the queue are preserved.
This is not lock-free and concurrent producer or consumer access
will lead to corruption.
@return True if the new capacity could accommodate the present elements in
the queue, otherwise nothing is done and false is returned.
*/
bool capacity(const size_t cap);
/*!
@brief Discard all data in the queue.
*/
void flush()
{
m_outPos.store(m_inPos.load());
}
/*!
@brief Get a snapshot number of elements that can be retrieved by pop.
*/
size_t available() const
{
int avail = static_cast<int>(m_inPos.load() - m_outPos.load());
if (avail < 0) avail += m_bufSize;
return avail;
}
/*!
@brief Get the remaining free elementes for pushing.
*/
size_t available_for_push() const
{
int avail = static_cast<int>(m_outPos.load() - m_inPos.load()) - 1;
if (avail < 0) avail += m_bufSize;
return avail;
}
/*!
@brief Peek at the next element pop will return without removing it from the queue.
@return An rvalue copy of the next element that can be popped. If the queue is empty,
return an rvalue copy of the element that is pending the next push.
*/
T peek() const
{
const auto outPos = m_outPos.load(std::memory_order_relaxed);
std::atomic_thread_fence(std::memory_order_acquire);
return m_buffer[outPos];
}
/*!
@brief Peek at the next pending input value.
@return A reference to the next element that can be pushed.
*/
inline T& IRAM_ATTR pushpeek() __attribute__((always_inline))
{
const auto inPos = m_inPos.load(std::memory_order_relaxed);
std::atomic_thread_fence(std::memory_order_acquire);
return m_buffer[inPos];
}
/*!
@brief Release the next pending input value, accessible by pushpeek(), into the queue.
@return true if the queue accepted the value, false if the queue
was full.
*/
inline bool IRAM_ATTR push() __attribute__((always_inline))
{
const auto inPos = m_inPos.load(std::memory_order_acquire);
const size_t next = (inPos + 1) % m_bufSize;
if (next == m_outPos.load(std::memory_order_relaxed)) {
return false;
}
std::atomic_thread_fence(std::memory_order_acquire);
m_inPos.store(next, std::memory_order_release);
return true;
}
/*!
@brief Move the rvalue parameter into the queue.
@return true if the queue accepted the value, false if the queue
was full.
*/
inline bool IRAM_ATTR push(T&& val) __attribute__((always_inline))
{
const auto inPos = m_inPos.load(std::memory_order_acquire);
const size_t next = (inPos + 1) % m_bufSize;
if (next == m_outPos.load(std::memory_order_relaxed)) {
return false;
}
std::atomic_thread_fence(std::memory_order_acquire);
m_buffer[inPos] = std::move(val);
std::atomic_thread_fence(std::memory_order_release);
m_inPos.store(next, std::memory_order_release);
return true;
}
/*!
@brief Push a copy of the parameter into the queue.
@return true if the queue accepted the value, false if the queue
was full.
*/
inline bool IRAM_ATTR push(const T& val) __attribute__((always_inline))
{
T v(val);
return push(std::move(v));
}
#if defined(ESP8266) || defined(ESP32) || !defined(ARDUINO)
/*!
@brief Push copies of multiple elements from a buffer into the queue,
in order, beginning at buffer's head.
@return The number of elements actually copied into the queue, counted
from the buffer head.
*/
size_t push_n(const T* buffer, size_t size);
#endif
/*!
@brief Pop the next available element from the queue.
@return An rvalue copy of the popped element, or a default
value of type T if the queue is empty.
*/
T pop();
#if defined(ESP8266) || defined(ESP32) || !defined(ARDUINO)
/*!
@brief Pop multiple elements in ordered sequence from the queue to a buffer.
If buffer is nullptr, simply discards up to size elements from the queue.
@return The number of elements actually popped from the queue to
buffer.
*/
size_t pop_n(T* buffer, size_t size);
#endif
/*!
@brief Iterate over and remove each available element from queue,
calling back fun with an rvalue reference of every single element.
*/
#if defined(ESP8266) || defined(ESP32) || !defined(ARDUINO)
void for_each(const Delegate<void(T&&), ForEachArg>& fun);
#else
void for_each(Delegate<void(T&&), ForEachArg> fun);
#endif
/*!
@brief In reverse order, iterate over, pop and optionally requeue each available element from the queue,
calling back fun with a reference of every single element.
Requeuing is dependent on the return boolean of the callback function. If it
returns true, the requeue occurs.
*/
#if defined(ESP8266) || defined(ESP32) || !defined(ARDUINO)
bool for_each_rev_requeue(const Delegate<bool(T&), ForEachArg>& fun);
#else
bool for_each_rev_requeue(Delegate<bool(T&), ForEachArg> fun);
#endif
protected:
const T defaultValue = {};
size_t m_bufSize;
#if defined(ESP8266) || defined(ESP32) || !defined(ARDUINO)
std::unique_ptr<T[]> m_buffer;
#else
std::unique_ptr<T> m_buffer;
#endif
std::atomic<size_t> m_inPos;
std::atomic<size_t> m_outPos;
};
template< typename T, typename ForEachArg >
bool circular_queue<T, ForEachArg>::capacity(const size_t cap)
{
if (cap + 1 == m_bufSize) return true;
else if (available() > cap) return false;
std::unique_ptr<T[] > buffer(new T[cap + 1]);
const auto available = pop_n(buffer, cap);
m_buffer.reset(buffer);
m_bufSize = cap + 1;
std::atomic_thread_fence(std::memory_order_release);
m_inPos.store(available, std::memory_order_relaxed);
m_outPos.store(0, std::memory_order_release);
return true;
}
#if defined(ESP8266) || defined(ESP32) || !defined(ARDUINO)
template< typename T, typename ForEachArg >
size_t circular_queue<T, ForEachArg>::push_n(const T* buffer, size_t size)
{
const auto inPos = m_inPos.load(std::memory_order_acquire);
const auto outPos = m_outPos.load(std::memory_order_relaxed);
size_t blockSize = (outPos > inPos) ? outPos - 1 - inPos : (outPos == 0) ? m_bufSize - 1 - inPos : m_bufSize - inPos;
blockSize = min(size, blockSize);
if (!blockSize) return 0;
int next = (inPos + blockSize) % m_bufSize;
std::atomic_thread_fence(std::memory_order_acquire);
auto dest = m_buffer.get() + inPos;
std::copy_n(std::make_move_iterator(buffer), blockSize, dest);
size = min(size - blockSize, outPos > 1 ? static_cast<size_t>(outPos - next - 1) : 0);
next += size;
dest = m_buffer.get();
std::copy_n(std::make_move_iterator(buffer + blockSize), size, dest);
std::atomic_thread_fence(std::memory_order_release);
m_inPos.store(next, std::memory_order_release);
return blockSize + size;
}
#endif
template< typename T, typename ForEachArg >
T circular_queue<T, ForEachArg>::pop()
{
const auto outPos = m_outPos.load(std::memory_order_acquire);
if (m_inPos.load(std::memory_order_relaxed) == outPos) return defaultValue;
std::atomic_thread_fence(std::memory_order_acquire);
auto val = std::move(m_buffer[outPos]);
std::atomic_thread_fence(std::memory_order_release);
m_outPos.store((outPos + 1) % m_bufSize, std::memory_order_release);
return val;
}
#if defined(ESP8266) || defined(ESP32) || !defined(ARDUINO)
template< typename T, typename ForEachArg >
size_t circular_queue<T, ForEachArg>::pop_n(T* buffer, size_t size) {
size_t avail = size = min(size, available());
if (!avail) return 0;
const auto outPos = m_outPos.load(std::memory_order_acquire);
size_t n = min(avail, static_cast<size_t>(m_bufSize - outPos));
std::atomic_thread_fence(std::memory_order_acquire);
if (buffer) {
buffer = std::copy_n(std::make_move_iterator(m_buffer.get() + outPos), n, buffer);
avail -= n;
std::copy_n(std::make_move_iterator(m_buffer.get()), avail, buffer);
}
std::atomic_thread_fence(std::memory_order_release);
m_outPos.store((outPos + size) % m_bufSize, std::memory_order_release);
return size;
}
#endif
template< typename T, typename ForEachArg >
#if defined(ESP8266) || defined(ESP32) || !defined(ARDUINO)
void circular_queue<T, ForEachArg>::for_each(const Delegate<void(T&&), ForEachArg>& fun)
#else
void circular_queue<T, ForEachArg>::for_each(Delegate<void(T&&), ForEachArg> fun)
#endif
{
auto outPos = m_outPos.load(std::memory_order_acquire);
const auto inPos = m_inPos.load(std::memory_order_relaxed);
std::atomic_thread_fence(std::memory_order_acquire);
while (outPos != inPos)
{
fun(std::move(m_buffer[outPos]));
std::atomic_thread_fence(std::memory_order_release);
outPos = (outPos + 1) % m_bufSize;
m_outPos.store(outPos, std::memory_order_release);
}
}
template< typename T, typename ForEachArg >
#if defined(ESP8266) || defined(ESP32) || !defined(ARDUINO)
bool circular_queue<T, ForEachArg>::for_each_rev_requeue(const Delegate<bool(T&), ForEachArg>& fun)
#else
bool circular_queue<T, ForEachArg>::for_each_rev_requeue(Delegate<bool(T&), ForEachArg> fun)
#endif
{
auto inPos0 = circular_queue<T, ForEachArg>::m_inPos.load(std::memory_order_acquire);
auto outPos = circular_queue<T, ForEachArg>::m_outPos.load(std::memory_order_relaxed);
std::atomic_thread_fence(std::memory_order_acquire);
if (outPos == inPos0) return false;
auto pos = inPos0;
auto outPos1 = inPos0;
const auto posDecr = circular_queue<T, ForEachArg>::m_bufSize - 1;
do {
pos = (pos + posDecr) % circular_queue<T, ForEachArg>::m_bufSize;
T&& val = std::move(circular_queue<T, ForEachArg>::m_buffer[pos]);
if (fun(val))
{
outPos1 = (outPos1 + posDecr) % circular_queue<T, ForEachArg>::m_bufSize;
if (outPos1 != pos) circular_queue<T, ForEachArg>::m_buffer[outPos1] = std::move(val);
}
} while (pos != outPos);
circular_queue<T, ForEachArg>::m_outPos.store(outPos1, std::memory_order_release);
return true;
}
#endif // __circular_queue_h

200
arduino-cli/libraries/CoopTask/src/circular_queue/circular_queue_mp.h Normal file
View File

@@ -0,0 +1,200 @@
/*
circular_queue_mp.h - Implementation of a lock-free circular queue for EspSoftwareSerial.
Copyright (c) 2019 Dirk O. Kaar. All rights reserved.
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#ifndef __circular_queue_mp_h
#define __circular_queue_mp_h
#include "circular_queue.h"
#ifdef ESP8266
#include "interrupts.h"
#else
#include <mutex>
#endif
/*!
@brief Instance class for a multi-producer, single-consumer circular queue / ring buffer (FIFO).
This implementation is lock-free between producers and consumer for the available(), peek(),
pop(), and push() type functions, but is guarded to safely allow only a single producer
at any instant.
*/
template< typename T, typename ForEachArg = void >
class circular_queue_mp : protected circular_queue<T, ForEachArg>
{
public:
circular_queue_mp() = default;
circular_queue_mp(const size_t capacity) : circular_queue<T, ForEachArg>(capacity)
{}
circular_queue_mp(circular_queue<T, ForEachArg>&& cq) : circular_queue<T, ForEachArg>(std::move(cq))
{}
using circular_queue<T, ForEachArg>::operator=;
using circular_queue<T, ForEachArg>::capacity;
using circular_queue<T, ForEachArg>::flush;
using circular_queue<T, ForEachArg>::available;
using circular_queue<T, ForEachArg>::available_for_push;
using circular_queue<T, ForEachArg>::peek;
using circular_queue<T, ForEachArg>::pop;
using circular_queue<T, ForEachArg>::pop_n;
using circular_queue<T, ForEachArg>::for_each;
using circular_queue<T, ForEachArg>::for_each_rev_requeue;
/*!
@brief Resize the queue. The available elements in the queue are preserved.
This is not lock-free, but safe, concurrent producer or consumer access
is guarded.
@return True if the new capacity could accommodate the present elements in
the queue, otherwise nothing is done and false is returned.
*/
bool capacity(const size_t cap)
{
#ifdef ESP8266
esp8266::InterruptLock lock;
#else
std::lock_guard<std::mutex> lock(m_pushMtx);
#endif
return circular_queue<T, ForEachArg>::capacity(cap);
}
bool IRAM_ATTR push() = delete;
/*!
@brief Move the rvalue parameter into the queue, guarded
for multiple concurrent producers.
@return true if the queue accepted the value, false if the queue
was full.
*/
bool IRAM_ATTR push(T&& val)
{
#ifdef ESP8266
esp8266::InterruptLock lock;
#else
std::lock_guard<std::mutex> lock(m_pushMtx);
#endif
return circular_queue<T, ForEachArg>::push(std::move(val));
}
/*!
@brief Push a copy of the parameter into the queue, guarded
for multiple concurrent producers.
@return true if the queue accepted the value, false if the queue
was full.
*/
bool IRAM_ATTR push(const T& val)
{
#ifdef ESP8266
esp8266::InterruptLock lock;
#else
std::lock_guard<std::mutex> lock(m_pushMtx);
#endif
return circular_queue<T, ForEachArg>::push(val);
}
/*!
@brief Push copies of multiple elements from a buffer into the queue,
in order, beginning at buffer's head. This is guarded for
multiple producers, push_n() is atomic.
@return The number of elements actually copied into the queue, counted
from the buffer head.
*/
size_t push_n(const T* buffer, size_t size)
{
#ifdef ESP8266
esp8266::InterruptLock lock;
#else
std::lock_guard<std::mutex> lock(m_pushMtx);
#endif
return circular_queue<T, ForEachArg>::push_n(buffer, size);
}
/*!
@brief Pops the next available element from the queue, requeues
it immediately.
@return A reference to the just requeued element, or the default
value of type T if the queue is empty.
*/
T& pop_requeue();
/*!
@brief Iterate over, pop and optionally requeue each available element from the queue,
calling back fun with a reference of every single element.
Requeuing is dependent on the return boolean of the callback function. If it
returns true, the requeue occurs.
*/
bool for_each_requeue(const Delegate<bool(T&), ForEachArg>& fun);
#ifndef ESP8266
protected:
std::mutex m_pushMtx;
#endif
};
template< typename T, typename ForEachArg >
T& circular_queue_mp<T, ForEachArg>::pop_requeue()
{
#ifdef ESP8266
esp8266::InterruptLock lock;
#else
std::lock_guard<std::mutex> lock(m_pushMtx);
#endif
const auto outPos = circular_queue<T, ForEachArg>::m_outPos.load(std::memory_order_acquire);
const auto inPos = circular_queue<T, ForEachArg>::m_inPos.load(std::memory_order_relaxed);
std::atomic_thread_fence(std::memory_order_acquire);
if (inPos == outPos) return circular_queue<T, ForEachArg>::defaultValue;
T& val = circular_queue<T, ForEachArg>::m_buffer[inPos] = std::move(circular_queue<T, ForEachArg>::m_buffer[outPos]);
const auto bufSize = circular_queue<T, ForEachArg>::m_bufSize;
std::atomic_thread_fence(std::memory_order_release);
circular_queue<T, ForEachArg>::m_outPos.store((outPos + 1) % bufSize, std::memory_order_relaxed);
circular_queue<T, ForEachArg>::m_inPos.store((inPos + 1) % bufSize, std::memory_order_release);
return val;
}
template< typename T, typename ForEachArg >
bool circular_queue_mp<T, ForEachArg>::for_each_requeue(const Delegate<bool(T&), ForEachArg>& fun)
{
auto inPos0 = circular_queue<T, ForEachArg>::m_inPos.load(std::memory_order_acquire);
auto outPos = circular_queue<T, ForEachArg>::m_outPos.load(std::memory_order_relaxed);
std::atomic_thread_fence(std::memory_order_acquire);
if (outPos == inPos0) return false;
do {
T&& val = std::move(circular_queue<T, ForEachArg>::m_buffer[outPos]);
if (fun(val))
{
#ifdef ESP8266
esp8266::InterruptLock lock;
#else
std::lock_guard<std::mutex> lock(m_pushMtx);
#endif
std::atomic_thread_fence(std::memory_order_release);
auto inPos = circular_queue<T, ForEachArg>::m_inPos.load(std::memory_order_relaxed);
std::atomic_thread_fence(std::memory_order_acquire);
circular_queue<T, ForEachArg>::m_buffer[inPos] = std::move(val);
std::atomic_thread_fence(std::memory_order_release);
circular_queue<T, ForEachArg>::m_inPos.store((inPos + 1) % circular_queue<T, ForEachArg>::m_bufSize, std::memory_order_release);
}
else
{
std::atomic_thread_fence(std::memory_order_release);
}
outPos = (outPos + 1) % circular_queue<T, ForEachArg>::m_bufSize;
circular_queue<T, ForEachArg>::m_outPos.store(outPos, std::memory_order_release);
} while (outPos != inPos0);
return true;
}
#endif // __circular_queue_mp_h

94
arduino-cli/libraries/CoopTask/src/circular_queue/ghostl.h Normal file
View File

@@ -0,0 +1,94 @@
/*
ghostl.h - Implementation of a bare-bones, mostly no-op, C++ STL shell
that allows building some Arduino ESP8266/ESP32
libraries on Aruduino AVR.
Copyright (c) 2019 Dirk O. Kaar. All rights reserved.
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#ifndef __ghostl_h
#define __ghostl_h
#if defined(ARDUINO_ARCH_SAMD)
#include <atomic>
#endif
using size_t = decltype(sizeof(char));
namespace std
{
#if !defined(ARDUINO_ARCH_SAMD)
typedef enum memory_order {
memory_order_relaxed,
memory_order_acquire,
memory_order_release,
memory_order_seq_cst
} memory_order;
template< typename T > class atomic {
private:
T value;
public:
atomic() {}
atomic(T desired) { value = desired; }
void store(T desired, std::memory_order = std::memory_order_seq_cst) volatile noexcept { value = desired; }
T load(std::memory_order = std::memory_order_seq_cst) const volatile noexcept { return value; }
};
inline void atomic_thread_fence(std::memory_order order) noexcept {}
template< typename T > T&& move(T& t) noexcept { return static_cast<T&&>(t); }
#endif
template< typename T, size_t long N > struct array
{
T _M_elems[N];
decltype(sizeof(0)) size() const { return N; }
T& operator[](decltype(sizeof(0)) i) { return _M_elems[i]; }
const T& operator[](decltype(sizeof(0)) i) const { return _M_elems[i]; }
};
template< typename T > class unique_ptr
{
public:
using pointer = T*;
unique_ptr() noexcept : ptr(nullptr) {}
unique_ptr(pointer p) : ptr(p) {}
pointer operator->() const noexcept { return ptr; }
T& operator[](decltype(sizeof(0)) i) const { return ptr[i]; }
void reset(pointer p = pointer()) noexcept
{
delete ptr;
ptr = p;
}
T& operator*() const { return *ptr; }
private:
pointer ptr;
};
template< typename T > using function = T*;
using nullptr_t = decltype(nullptr);
template<typename T>
struct identity {
typedef T type;
};
template <typename T>
inline T&& forward(typename identity<T>::type& t) noexcept
{
return static_cast<typename identity<T>::type&&>(t);
}
}
#endif // __ghostl_h