Add files via upload

This commit is contained in:
August Quint 2020-02-05 16:34:37 +01:00 committed by GitHub
parent 090a1c6e35
commit 8e8d2bf8ee
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
17 changed files with 5126 additions and 0 deletions

502
EspSoftwareSerial/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!

124
EspSoftwareSerial/README.md Normal file
View File

@ -0,0 +1,124 @@
# EspSoftwareSerial
## Implementation of the Arduino software serial library for the ESP8266 / ESP32
This fork implements interrupt service routine best practice.
In the receive interrupt, instead of blocking for whole bytes
at a time - voiding any near-realtime behavior of the CPU - only level
change and timestamp are recorded. The more time consuming phase
detection and byte assembly are done in the main code.
Except at high bitrates, depending on other ongoing activity,
interrupts in particular, this software serial adapter
supports full duplex receive and send. At high bitrates (115200bps)
send bit timing can be improved at the expense of blocking concurrent
full duplex receives, with the ``SoftwareSerial::enableIntTx(false)`` function call.
The same functionality is given as the corresponding AVR library but
several instances can be active at the same time. Speed up to 115200 baud
is supported. Besides a constructor compatible to the AVR SoftwareSerial class,
and updated constructor that takes no arguments exists, instead the ``begin()``
function can handle the pin assignments and logic inversion.
It also has optional input buffer capacity arguments for byte buffer and ISR bit buffer.
This way, it is a better drop-in replacement for the hardware serial APIs on the ESP MCUs.
Please note that due to the fact that the ESPs always have other activities
ongoing, there will be some inexactness in interrupt timings. This may
lead to inevitable, but few, bit errors when having heavy data traffic
at high baud rates.
## Resource optimization
The memory footprint can be optimized to just fit the amount of expected
incoming asynchronous data.
For this, the ``SoftwareSerial`` constructor provides two arguments. First, the
octet buffer capacity for assembled received octets can be set. Read calls are
satisfied from this buffer, freeing it in return.
Second, the signal edge detection buffer of 32bit fields can be resized.
One octet may require up to to 10 fields, but fewer may be needed,
depending on the bit pattern. Any read or write calls check this buffer
to assemble received octets, thus promoting completed octets to the octet
buffer, freeing fields in the edge detection buffer.
Look at the swsertest.ino example. There, on reset, ASCII characters ' ' to 'z'
are sent. This happens not as a block write, but in a single write call per
character. As the example uses a local loopback wire, every outgoing bit is
immediately received back. Therefore, any single write call causes up to
10 fields - depending on the exact bit pattern - to be occupied in the signal
edge detection buffer. In turn, as explained before, each single write call
also causes received bit assembly to be performed, promoting these bits from
the signal edge detection buffer to the octet buffer as soon as possible.
Explaining by way of contrast, if during a a single write call, perhaps because
of using block writing, more than a single octet is received, there will be a
need for more than 10 fields in the signal edge detection buffer.
The necessary capacity of the octet buffer only depends on the amount of incoming
data until the next read call.
For the swsertest.ino example, this results in the following optimized
constructor arguments to spend only the minimum RAM on buffers required:
The octet buffer capacity (``bufCapacity``) is 93 (91 characters net plus two tolerance).
The signal edge detection buffer capacity (``isrBufCapacity``) is 10, as each octet has
10 bits on the wire, which are immediately received during the write, and each
write call causes the signal edge detection to promote the previously sent and
received bits to the octet buffer.
In a more generalized scenario, calculate the bits (use message size in octets
times 10) that may be asynchronously received to determine the value for
``isrBufCapacity`` in the constructor. Also use the number of received octets
that must be buffered for reading as the value of ``bufCapacity``.
The more frequently your code calls write or read functions, the greater the
chances are that you can reduce the ``isrBufCapacity`` footprint without losing data,
and each time you call read to fetch from the octet buffer, you reduce the
need for space there.
## SoftwareSerialConfig and parity
The configuration of the data stream is done via a ``SoftwareSerialConfig``
argument to ``begin()``. Word lengths can be set to between 5 and 8 bits, parity
can be N(one), O(dd) or E(ven) and 1 or 2 stop bits can be used. The default is
``SWSERIAL_8N1`` using 8 bits, no parity and 1 stop bit but any combination can
be used, e.g. ``SWSERIAL_7E2``. If using EVEN or ODD parity, any parity errors
can be detected with the ``peekParityError()`` function. Note that parity
checking must be done before ``read()``, as the parity information is removed
from the buffer when reading the corresponding byte.
To allow flexible 9-bit and data/addressing protocols, the additional parity
modes MARK and SPACE are also available. Furthermore, the parity mode can be
individually set in each call to ``write()``.
This allows a simple implementation of protocols where the parity bit is used to
distinguish between data and addresses/commands ("9-bit" protocols). First set
up SoftwareSerial with parity mode SPACE, e.g. ``SWSERIAL_8S1``. This will add a
parity bit to every byte sent, setting it to logical zero (SPACE parity).
To detect incoming bytes with the parity bit set (MARK parity), use the
``peekParityError()`` function. To send a byte with the parity bit set, just add
``MARK`` as the second argument when writing, e.g. ``write(ch, MARK)``.
## Using and updating EspSoftwareSerial in the esp8266com/esp8266 Arduino build environment
EspSoftwareSerial is both part of the BSP download for ESP8266 in Arduino,
and it is set up as a Git submodule in the esp8266 source tree,
specifically in ``.../esp8266/libraries/SoftwareSerial`` when using a Github
repository clone in your Arduino sketchbook hardware directory.
This supersedes any version of EspSoftwareSerial installed for instance via
the Arduino library manager, it is not required to install EspSoftwareSerial
for the ESP8266 separately at all, but doing so has ill effect.
The responsible maintainer of the esp8266 repository has kindly shared the
following command line instructions to use, if one wishes to manually
update EspSoftwareSerial to a newer release than pulled in via the ESP8266 Arduino BSP:
To update esp8266/arduino SoftwareSerial submodule to lastest master:
Clean it (optional):
```shell
$ rm -rf libraries/SoftwareSerial
$ git submodule update --init
```
Now update it:
```shell
$ cd libraries/SoftwareSerial
$ git checkout master
$ git pull
```

View File

@ -0,0 +1,263 @@
#include <SoftwareSerial.h>
// On ESP8266:
// Local SoftwareSerial loopback, connect D5 (rx) and D6 (tx).
// For local hardware loopback, connect D5 to D8 (tx), D6 to D7 (rx).
// For hardware send/sink, connect D7 (rx) and D8 (tx).
// Hint: The logger is run at 9600bps such that enableIntTx(true) can remain unchanged. Blocking
// interrupts severely impacts the ability of the SoftwareSerial devices to operate concurrently
// and/or in duplex mode.
// Operating in software serial full duplex mode, runs at 19200bps and few errors (~2.5%).
// Operating in software serial half duplex mode (both loopback and repeater),
// runs at 57600bps with nearly no errors.
// Operating loopback in full duplex, and repeater in half duplex, runs at 38400bps with nearly no errors.
// On ESP32:
// For SoftwareSerial or hardware send/sink, connect D5 (rx) and D6 (tx).
// Hardware Serial2 defaults to D4 (rx), D3 (tx).
// For local hardware loopback, connect D5 (rx) to D3 (tx), D6 (tx) to D4 (rx).
#if defined(ESP8266) && !defined(D5)
#define D5 (14)
#define D6 (12)
#define D7 (13)
#define D8 (15)
#define TX (1)
#endif
// Pick only one of HWLOOPBACK, HWSOURCESWSINK, or HWSOURCESINK
//#define HWLOOPBACK 1
//#define HWSOURCESWSINK 1
//#define HWSOURCESINK 1
#define HALFDUPLEX 1
#ifdef ESP32
constexpr int IUTBITRATE = 19200;
#else
constexpr int IUTBITRATE = 19200;
#endif
#if defined(ESP8266)
constexpr SoftwareSerialConfig swSerialConfig = SWSERIAL_8E1;
constexpr SerialConfig hwSerialConfig = SERIAL_8E1;
#elif defined(ESP32)
constexpr SoftwareSerialConfig swSerialConfig = SWSERIAL_8E1;
constexpr uint32_t hwSerialConfig = SERIAL_8E1;
#else
constexpr unsigned swSerialConfig = 3;
#endif
constexpr bool invert = false;
constexpr int BLOCKSIZE = 16; // use fractions of 256
unsigned long start;
String effTxTxt("eff. tx: ");
String effRxTxt("eff. rx: ");
int txCount;
int rxCount;
int expected;
int rxErrors;
int rxParityErrors;
constexpr int ReportInterval = IUTBITRATE / 8;
#if defined(ESP8266)
#if defined(HWLOOPBACK) || defined(HWSOURCESWSINK)
HardwareSerial& hwSerial(Serial);
SoftwareSerial serialIUT;
SoftwareSerial logger;
#elif defined(HWSOURCESINK)
HardwareSerial& serialIUT(Serial);
SoftwareSerial logger;
#else
SoftwareSerial serialIUT;
HardwareSerial& logger(Serial);
#endif
#elif defined(ESP32)
#if defined(HWLOOPBACK) || defined (HWSOURCESWSINK)
HardwareSerial& hwSerial(Serial2);
SoftwareSerial serialIUT;
#elif defined(HWSOURCESINK)
HardwareSerial& serialIUT(Serial2);
#else
SoftwareSerial serialIUT;
#endif
HardwareSerial& logger(Serial);
#else
SoftwareSerial serialIUT(14, 12);
HardwareSerial& logger(Serial);
#endif
void setup() {
#if defined(ESP8266)
#if defined(HWLOOPBACK) || defined(HWSOURCESINK) || defined(HWSOURCESWSINK)
Serial.begin(IUTBITRATE, hwSerialConfig, SERIAL_FULL, 1, invert);
Serial.swap();
Serial.setRxBufferSize(2 * BLOCKSIZE);
logger.begin(9600, SWSERIAL_8N1, -1, TX);
#else
logger.begin(9600);
#endif
#if !defined(HWSOURCESINK)
serialIUT.begin(IUTBITRATE, swSerialConfig, D5, D6, invert, 2 * BLOCKSIZE);
#ifdef HALFDUPLEX
serialIUT.enableIntTx(false);
#endif
#endif
#elif defined(ESP32)
#if defined(HWLOOPBACK) || defined(HWSOURCESWSINK)
Serial2.begin(IUTBITRATE, hwSerialConfig, D4, D3, invert);
Serial2.setRxBufferSize(2 * BLOCKSIZE);
#elif defined(HWSOURCESINK)
serialIUT.begin(IUTBITRATE, hwSerialConfig, D5, D6, invert);
serialIUT.setRxBufferSize(2 * BLOCKSIZE);
#endif
#if !defined(HWSOURCESINK)
serialIUT.begin(IUTBITRATE, swSerialConfig, D5, D6, invert, 2 * BLOCKSIZE);
#ifdef HALFDUPLEX
serialIUT.enableIntTx(false);
#endif
#endif
logger.begin(9600);
#else
#if !defined(HWSOURCESINK)
serialIUT.begin(IUTBITRATE);
#endif
logger.begin(9600);
#endif
logger.println("Loopback example for EspSoftwareSerial");
start = micros();
txCount = 0;
rxCount = 0;
rxErrors = 0;
rxParityErrors = 0;
expected = -1;
}
unsigned char c = 0;
void loop() {
#ifdef HALFDUPLEX
char block[BLOCKSIZE];
#endif
char inBuf[BLOCKSIZE];
for (int i = 0; i < BLOCKSIZE; ++i) {
#ifndef HALFDUPLEX
#ifdef HWSOURCESWSINK
hwSerial.write(c);
#else
serialIUT.write(c);
#endif
#ifdef HWLOOPBACK
int avail = hwSerial.available();
while ((0 == (i % 8)) && avail > 0) {
int inCnt = hwSerial.read(inBuf, min(avail, min(BLOCKSIZE, hwSerial.availableForWrite())));
hwSerial.write(inBuf, inCnt);
avail -= inCnt;
}
#endif
#else
block[i] = c;
#endif
c = (c + 1) % 256;
++txCount;
}
#ifdef HALFDUPLEX
#ifdef HWSOURCESWSINK
hwSerial.write(block, BLOCKSIZE);
#else
serialIUT.write(block, BLOCKSIZE);
#endif
#endif
#ifdef HWSOURCESINK
#if defined(ESP8266)
if (serialIUT.hasOverrun()) { logger.println("serialIUT.overrun"); }
#endif
#else
if (serialIUT.overflow()) { logger.println("serialIUT.overflow"); }
#endif
int inCnt;
uint32_t deadlineStart;
#ifdef HWLOOPBACK
// starting deadline for the first bytes to become readable
deadlineStart = ESP.getCycleCount();
inCnt = 0;
while ((ESP.getCycleCount() - deadlineStart) < (1000000UL * 12 * BLOCKSIZE) / IUTBITRATE * 24 * ESP.getCpuFreqMHz()) {
int avail = hwSerial.available();
inCnt += hwSerial.read(&inBuf[inCnt], min(avail, min(BLOCKSIZE - inCnt, hwSerial.availableForWrite())));
if (inCnt >= BLOCKSIZE) { break; }
// wait for more outstanding bytes to trickle in
if (avail) deadlineStart = ESP.getCycleCount();
}
hwSerial.write(inBuf, inCnt);
#endif
// starting deadline for the first bytes to come in
deadlineStart = ESP.getCycleCount();
inCnt = 0;
while ((ESP.getCycleCount() - deadlineStart) < (1000000UL * 12 * BLOCKSIZE) / IUTBITRATE * 8 * ESP.getCpuFreqMHz()) {
int avail;
if (0 != (swSerialConfig & 070))
avail = serialIUT.available();
else
avail = serialIUT.read(inBuf, BLOCKSIZE);
for (int i = 0; i < avail; ++i)
{
unsigned char r;
if (0 != (swSerialConfig & 070))
r = serialIUT.read();
else
r = inBuf[i];
if (expected == -1) { expected = r; }
else {
expected = (expected + 1) % (1UL << (5 + swSerialConfig % 4));
}
if (r != expected) {
++rxErrors;
expected = -1;
}
#ifndef HWSOURCESINK
if (serialIUT.readParity() != (static_cast<bool>(swSerialConfig & 010) ? serialIUT.parityOdd(r) : serialIUT.parityEven(r)))
{
++rxParityErrors;
}
#endif
++rxCount;
++inCnt;
}
if (inCnt >= BLOCKSIZE) { break; }
// wait for more outstanding bytes to trickle in
if (avail) deadlineStart = ESP.getCycleCount();
}
const uint32_t interval = micros() - start;
if (txCount >= ReportInterval && interval) {
uint8_t wordBits = (5 + swSerialConfig % 4) + static_cast<bool>(swSerialConfig & 070) + 1 + ((swSerialConfig & 0300) ? 1 : 0);
logger.println(String("tx/rx: ") + txCount + "/" + rxCount);
const long txCps = txCount * (1000000.0 / interval);
const long rxCps = rxCount * (1000000.0 / interval);
logger.print(effTxTxt + wordBits * txCps + "bps, "
+ effRxTxt + wordBits * rxCps + "bps, "
+ rxErrors + " errors (" + 100.0 * rxErrors / (!rxErrors ? 1 : rxCount) + "%)");
if (0 != (swSerialConfig & 070))
{
logger.print(" ("); logger.print(rxParityErrors); logger.println(" parity errors)");
}
else
{
logger.println();
}
txCount = 0;
rxCount = 0;
rxErrors = 0;
rxParityErrors = 0;
expected = -1;
// resync
delay(1000UL * 12 * BLOCKSIZE / IUTBITRATE * 16);
serialIUT.flush();
start = micros();
}
}

View File

@ -0,0 +1,48 @@
#include <ESP8266WiFi.h>
#include "SoftwareSerial.h"
SoftwareSerial swSer1;
SoftwareSerial swSer2;
void setup() {
delay(2000);
Serial.begin(115200);
Serial.println("\nOne Wire Half Duplex Serial Tester");
swSer1.begin(115200, SWSERIAL_8N1, 12, 12, false, 256);
swSer1.enableIntTx(true);
swSer2.begin(115200, SWSERIAL_8N1, 14, 14, false, 256);
swSer2.enableIntTx(true);
}
void loop() {
Serial.println("\n\nTesting on swSer1");
Serial.print("Enter something to send using swSer1.");
checkSwSerial(&swSer1);
Serial.println("\n\nTesting on swSer2");
Serial.print("Enter something to send using swSer2.");
checkSwSerial(&swSer2);
}
void checkSwSerial(SoftwareSerial* ss) {
byte ch;
while (!Serial.available());
ss->enableTx(true);
while (Serial.available()) {
ch = Serial.read();
ss->write(ch);
}
ss->enableTx(false);
// wait 1 second for the reply from SOftwareSerial if any
delay(1000);
if (ss->available()) {
Serial.print("\nResult:");
while (ss->available()) {
ch = (byte)ss->read();
Serial.print(ch < 0x01 ? " 0" : " ");
Serial.print(ch, HEX);
}
Serial.println();
}
}

View File

@ -0,0 +1,183 @@
#include <SoftwareSerial.h>
// On ESP8266:
// SoftwareSerial loopback for remote source (loopback.ino), or hardware loopback.
// Connect source D5 (rx) to local D8 (tx), source D6 (tx) to local D7 (rx).
// Hint: The logger is run at 9600bps such that enableIntTx(true) can remain unchanged. Blocking
// interrupts severely impacts the ability of the SoftwareSerial devices to operate concurrently
// and/or in duplex mode.
// On ESP32:
// For software or hardware loopback, connect source rx to local D8 (tx), source tx to local D7 (rx).
#if defined(ESP8266) && !defined(D5)
#define D5 (14)
#define D6 (12)
#define D7 (13)
#define D8 (15)
#define TX (1)
#endif
#define HWLOOPBACK 1
#define HALFDUPLEX 1
#ifdef ESP32
constexpr int IUTBITRATE = 19200;
#else
constexpr int IUTBITRATE = 19200;
#endif
#if defined(ESP8266)
constexpr SoftwareSerialConfig swSerialConfig = SWSERIAL_8E1;
constexpr SerialConfig hwSerialConfig = SERIAL_8E1;
#elif defined(ESP32)
constexpr SoftwareSerialConfig swSerialConfig = SWSERIAL_8E1;
constexpr uint32_t hwSerialConfig = SERIAL_8E1;
#else
constexpr unsigned swSerialConfig = 3;
#endif
constexpr bool invert = false;
constexpr int BLOCKSIZE = 16; // use fractions of 256
unsigned long start;
String bitRateTxt("Effective data rate: ");
int rxCount;
int seqErrors;
int parityErrors;
int expected;
constexpr int ReportInterval = IUTBITRATE / 8;
#if defined(ESP8266)
#if defined(HWLOOPBACK)
HardwareSerial& repeater(Serial);
SoftwareSerial logger;
#else
SoftwareSerial repeater;
HardwareSerial& logger(Serial);
#endif
#elif defined(ESP32)
#if defined(HWLOOPBACK)
HardwareSerial& repeater(Serial2);
#else
SoftwareSerial repeater;
#endif
HardwareSerial& logger(Serial);
#else
SoftwareSerial repeater(14, 12);
HardwareSerial& logger(Serial);
#endif
void setup() {
#if defined(ESP8266)
#if defined(HWLOOPBACK)
repeater.begin(IUTBITRATE, hwSerialConfig, SERIAL_FULL, 1, invert);
repeater.swap();
repeater.setRxBufferSize(2 * BLOCKSIZE);
logger.begin(9600, SWSERIAL_8N1, -1, TX);
#else
repeater.begin(IUTBITRATE, swSerialConfig, D7, D8, invert, 4 * BLOCKSIZE);
#ifdef HALFDUPLEX
repeater.enableIntTx(false);
#endif
logger.begin(9600);
#endif
#elif defined(ESP32)
#if defined(HWLOOPBACK)
repeater.begin(IUTBITRATE, hwSerialConfig, D7, D8, invert);
repeater.setRxBufferSize(2 * BLOCKSIZE);
#else
repeater.begin(IUTBITRATE, swSerialConfig, D7, D8, invert, 4 * BLOCKSIZE);
#ifdef HALFDUPLEX
repeater.enableIntTx(false);
#endif
#endif
logger.begin(9600);
#else
repeater.begin(IUTBITRATE);
logger.begin(9600);
#endif
logger.println("Repeater example for EspSoftwareSerial");
start = micros();
rxCount = 0;
seqErrors = 0;
parityErrors = 0;
expected = -1;
}
void loop() {
#ifdef HWLOOPBACK
#if defined(ESP8266)
if (repeater.hasOverrun()) { logger.println("repeater.overrun"); }
#endif
#else
if (repeater.overflow()) { logger.println("repeater.overflow"); }
#endif
#ifdef HALFDUPLEX
char block[BLOCKSIZE];
#endif
// starting deadline for the first bytes to come in
uint32_t deadlineStart = ESP.getCycleCount();
int inCnt = 0;
while ((ESP.getCycleCount() - deadlineStart) < (1000000UL * 12 * BLOCKSIZE) / IUTBITRATE * 24 * ESP.getCpuFreqMHz()) {
int avail = repeater.available();
for (int i = 0; i < avail; ++i)
{
int r = repeater.read();
if (r == -1) { logger.println("read() == -1"); }
if (expected == -1) { expected = r; }
else {
expected = (expected + 1) % (1UL << (5 + swSerialConfig % 4));
}
if (r != expected) {
++seqErrors;
expected = -1;
}
#ifndef HWLOOPBACK
if (repeater.readParity() != (static_cast<bool>(swSerialConfig & 010) ? repeater.parityOdd(r) : repeater.parityEven(r)))
{
++parityErrors;
}
#endif
++rxCount;
#ifdef HALFDUPLEX
block[inCnt] = r;
#else
repeater.write(r);
#endif
if (++inCnt >= BLOCKSIZE) { break; }
}
if (inCnt >= BLOCKSIZE) { break; }
// wait for more outstanding bytes to trickle in
if (avail) deadlineStart = ESP.getCycleCount();
}
#ifdef HALFDUPLEX
repeater.write(block, inCnt);
#endif
if (rxCount >= ReportInterval) {
auto end = micros();
unsigned long interval = end - start;
long cps = rxCount * (1000000.0 / interval);
long seqErrorsps = seqErrors * (1000000.0 / interval);
logger.print(bitRateTxt + 10 * cps + "bps, "
+ seqErrorsps + "cps seq. errors (" + 100.0 * seqErrors / rxCount + "%)");
#ifndef HWLOOPBACK
if (0 != (swSerialConfig & 070))
{
logger.print(" ("); logger.print(parityErrors); logger.print(" parity errors)");
}
else
#endif
{
logger.println();
}
start = end;
rxCount = 0;
seqErrors = 0;
parityErrors = 0;
expected = -1;
}
}

View File

@ -0,0 +1,115 @@
#include <ESP8266WiFi.h>
#include <SoftwareSerial.h>
SoftwareSerial swSer;
byte buf[10] = { 0xFA, 0xAF,0x00,0x00,0x00, 0x00, 0x00, 0x00, 0x00, 0xED };
byte cmd[10] = { 0xFA, 0xAF,0x00,0x00,0x00, 0x00, 0x00, 0x00, 0x00, 0xED };
byte ver[10] = { 0xFC, 0xCF,0x00,0xAA,0x41, 0x16, 0x51, 0x01, 0x00, 0xED };
void setup() {
delay(2000);
Serial.begin(115200);
Serial.println("\nAlpha 1S Servo Tester");
swSer.begin(115200, SWSERIAL_8N1, 12, 12, false, 256);
}
void loop() {
for (int i = 1; i <= 32; i++) {
GetVersion(i);
delay(100);
}
SetLED(1, 0);
GoPos(1, 0, 50);
delay(1000);
GoPos(1, 90, 50);
delay(1000);
GoPos(1, 100, 50);
delay(1000);
SetLED(1, 1);
delay(2000);
}
void GetVersion(byte id) {
memcpy(buf, cmd, 10);
buf[0] = 0xFC;
buf[1] = 0xCF;
buf[2] = id;
buf[3] = 0x01;
SendCommand();
}
void GoPos(byte id, byte Pos, byte Time) {
memcpy(buf, cmd, 10);
buf[2] = id;
buf[3] = 0x01;
buf[4] = Pos;
buf[5] = Time;
buf[6] = 0x00;
buf[7] = Time;
SendCommand();
}
void GetPos(byte id) {
memcpy(buf, cmd, 10);
buf[2] = id;
buf[3] = 0x02;
SendCommand();
}
void SetLED(byte id, byte mode) {
memcpy(buf, cmd, 10);
buf[2] = id;
buf[3] = 0x04;
buf[4] = mode;
SendCommand();
}
void SendCommand() {
SendCommand(true);
}
void SendCommand(bool checkResult) {
byte sum = 0;
for (int i = 2; i < 8; i++) {
sum += buf[i];
}
buf[8] = sum;
ShowCommand();
swSer.flush();
swSer.enableTx(true);
swSer.write(buf, 10);
swSer.enableTx(false);
if (checkResult) checkReturn();
}
void ShowCommand() {
Serial.print(millis());
Serial.print(" OUT>>");
for (int i = 0; i < 10; i++) {
Serial.print((buf[i] < 0x10 ? " 0" : " "));
Serial.print(buf[i], HEX);
}
Serial.println();
}
void checkReturn() {
unsigned long startMs = millis();
while (((millis() - startMs) < 500) && (!swSer.available()));
if (swSer.available()) {
Serial.print(millis());
Serial.print(" IN>>>");
while (swSer.available()) {
byte ch = (byte)swSer.read();
Serial.print((ch < 0x10 ? " 0" : " "));
Serial.print(ch, HEX);
}
Serial.println();
}
}

View File

@ -0,0 +1,47 @@
// On ESP8266:
// At 80MHz runs up 57600ps, and at 160MHz CPU frequency up to 115200bps with only negligible errors.
// Connect pin 12 to 14.
#include <SoftwareSerial.h>
#if defined(ESP8266) && !defined(D5)
#define D5 (14)
#define D6 (12)
#define D7 (13)
#define D8 (15)
#endif
#ifdef ESP32
#define BAUD_RATE 57600
#else
#define BAUD_RATE 57600
#endif
// Reminder: the buffer size optimizations here, in particular the isrBufSize that only accommodates
// a single 8N1 word, are on the basis that any char written to the loopback SoftwareSerial adapter gets read
// before another write is performed. Block writes with a size greater than 1 would usually fail.
SoftwareSerial swSer;
void setup() {
Serial.begin(115200);
swSer.begin(BAUD_RATE, SWSERIAL_8N1, D5, D6, false, 95, 11);
Serial.println("\nSoftware serial test started");
for (char ch = ' '; ch <= 'z'; ch++) {
swSer.write(ch);
}
swSer.println("");
}
void loop() {
while (swSer.available() > 0) {
Serial.write(swSer.read());
yield();
}
while (Serial.available() > 0) {
swSer.write(Serial.read());
yield();
}
}

View File

@ -0,0 +1,43 @@
#######################################
# Syntax Coloring Map for SoftwareSerial
# (esp8266)
#######################################
#######################################
# Datatypes (KEYWORD1)
#######################################
SoftwareSerial KEYWORD1
#######################################
# Methods and Functions (KEYWORD2)
#######################################
begin KEYWORD2
baudRate KEYWORD2
setTransmitEnablePin KEYWORD2
enableIntTx KEYWORD2
overflow KEYWORD2
available KEYWORD2
peek KEYWORD2
read KEYWORD2
flush KEYWORD2
write KEYWORD2
enableRx KEYWORD2
enableTx KEYWORD2
listen KEYWORD2
end KEYWORD2
isListening KEYWORD2
stopListening KEYWORD2
onReceive KEYWORD2
perform_work KEYWORD2
#######################################
# Constants (LITERAL1)
#######################################
SW_SERIAL_UNUSED_PIN LITERAL1
SWSERIAL_5N1 LITERAL1
SWSERIAL_6N1 LITERAL1
SWSERIAL_7N1 LITERAL1
SWSERIAL_8N1 LITERAL1

View File

@ -0,0 +1,15 @@
{
"name": "EspSoftwareSerial",
"version": "6.6.1",
"keywords": [
"serial", "io", "softwareserial"
],
"description": "Implementation of the Arduino software serial for ESP8266/ESP32.",
"repository":
{
"type": "git",
"url": "https://github.com/plerup/espsoftwareserial"
},
"frameworks": "arduino",
"platforms": "*"
}

View File

@ -0,0 +1,9 @@
name=EspSoftwareSerial
version=6.6.1
author=Peter Lerup, Dirk Kaar
maintainer=Peter Lerup <peter@lerup.com>
sentence=Implementation of the Arduino software serial for ESP8266/ESP32.
paragraph=
category=Signal Input/Output
url=https://github.com/plerup/espsoftwareserial/
architectures=esp8266,esp32

View File

@ -0,0 +1,542 @@
/*
SoftwareSerial.cpp - Implementation of the Arduino software serial for ESP8266/ESP32.
Copyright (c) 2015-2016 Peter Lerup. All rights reserved.
Copyright (c) 2018-2019 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 "SoftwareSerial.h"
#include <Arduino.h>
#ifdef ESP32
#define xt_rsil(a) (a)
#define xt_wsr_ps(a)
#endif
constexpr uint8_t BYTE_ALL_BITS_SET = ~static_cast<uint8_t>(0);
SoftwareSerial::SoftwareSerial() {
m_isrOverflow = false;
}
SoftwareSerial::SoftwareSerial(int8_t rxPin, int8_t txPin, bool invert)
{
m_isrOverflow = false;
m_rxPin = rxPin;
m_txPin = txPin;
m_invert = invert;
}
SoftwareSerial::~SoftwareSerial() {
end();
}
bool SoftwareSerial::isValidGPIOpin(int8_t pin) {
#if defined(ESP8266)
return (pin >= 0 && pin <= 5) || (pin >= 12 && pin <= 15);
#elif defined(ESP32)
return pin == 0 || pin == 2 || (pin >= 4 && pin <= 5) || (pin >= 12 && pin <= 19) ||
(pin >= 21 && pin <= 23) || (pin >= 25 && pin <= 27) || (pin >= 32 && pin <= 35);
#else
return true;
#endif
}
void SoftwareSerial::begin(uint32_t baud, SoftwareSerialConfig config,
int8_t rxPin, int8_t txPin,
bool invert, int bufCapacity, int isrBufCapacity) {
if (-1 != rxPin) m_rxPin = rxPin;
if (-1 != txPin) m_txPin = txPin;
m_oneWire = (m_rxPin == m_txPin);
m_invert = invert;
m_dataBits = 5 + (config & 07);
m_parityMode = static_cast<SoftwareSerialParity>(config & 070);
m_stopBits = 1 + ((config & 0300) ? 1 : 0);
m_pduBits = m_dataBits + static_cast<bool>(m_parityMode) + m_stopBits;
m_bitCycles = (ESP.getCpuFreqMHz() * 1000000UL + baud / 2) / baud;
m_intTxEnabled = true;
if (isValidGPIOpin(m_rxPin)) {
std::unique_ptr<circular_queue<uint8_t> > buffer(new circular_queue<uint8_t>((bufCapacity > 0) ? bufCapacity : 64));
m_buffer = move(buffer);
if (m_parityMode)
{
std::unique_ptr<circular_queue<uint8_t> > parityBuffer(new circular_queue<uint8_t>((bufCapacity > 0) ? (bufCapacity + 7) / 8 : 8));
m_parityBuffer = move(parityBuffer);
m_parityInPos = m_parityOutPos = 1;
}
std::unique_ptr<circular_queue<uint32_t> > isrBuffer(new circular_queue<uint32_t>((isrBufCapacity > 0) ? isrBufCapacity : (sizeof(uint8_t) * 8 + 2) * bufCapacity));
m_isrBuffer = move(isrBuffer);
if (m_buffer && (!m_parityMode || m_parityBuffer) && m_isrBuffer) {
m_rxValid = true;
pinMode(m_rxPin, INPUT_PULLUP);
}
}
if (isValidGPIOpin(m_txPin)
#ifdef ESP8266
|| ((m_txPin == 16) && !m_oneWire)) {
#else
) {
#endif
m_txValid = true;
if (!m_oneWire) {
pinMode(m_txPin, OUTPUT);
digitalWrite(m_txPin, !m_invert);
}
}
if (!m_rxEnabled) { enableRx(true); }
}
void SoftwareSerial::end()
{
enableRx(false);
m_txValid = false;
if (m_buffer) {
m_buffer.reset();
}
m_parityBuffer.reset();
if (m_isrBuffer) {
m_isrBuffer.reset();
}
}
uint32_t SoftwareSerial::baudRate() {
return ESP.getCpuFreqMHz() * 1000000UL / m_bitCycles;
}
void SoftwareSerial::setTransmitEnablePin(int8_t txEnablePin) {
if (isValidGPIOpin(txEnablePin)) {
m_txEnableValid = true;
m_txEnablePin = txEnablePin;
pinMode(m_txEnablePin, OUTPUT);
digitalWrite(m_txEnablePin, LOW);
}
else {
m_txEnableValid = false;
}
}
void SoftwareSerial::enableIntTx(bool on) {
m_intTxEnabled = on;
}
void SoftwareSerial::enableTx(bool on) {
if (m_txValid && m_oneWire) {
if (on) {
enableRx(false);
pinMode(m_txPin, OUTPUT);
digitalWrite(m_txPin, !m_invert);
}
else {
pinMode(m_rxPin, INPUT_PULLUP);
enableRx(true);
}
}
}
void SoftwareSerial::enableRx(bool on) {
if (m_rxValid) {
if (on) {
m_rxCurBit = m_pduBits - 1;
// Init to stop bit level and current cycle
m_isrLastCycle = (ESP.getCycleCount() | 1) ^ m_invert;
if (m_bitCycles >= (ESP.getCpuFreqMHz() * 1000000UL) / 74880UL)
attachInterruptArg(digitalPinToInterrupt(m_rxPin), reinterpret_cast<void (*)(void*)>(rxBitISR), this, CHANGE);
else
attachInterruptArg(digitalPinToInterrupt(m_rxPin), reinterpret_cast<void (*)(void*)>(rxBitSyncISR), this, m_invert ? RISING : FALLING);
}
else {
detachInterrupt(digitalPinToInterrupt(m_rxPin));
}
m_rxEnabled = on;
}
}
int SoftwareSerial::read() {
if (!m_rxValid) { return -1; }
if (!m_buffer->available()) {
rxBits();
if (!m_buffer->available()) { return -1; }
}
auto val = m_buffer->pop();
if (m_parityBuffer)
{
m_lastReadParity = m_parityBuffer->peek() & m_parityOutPos;
m_parityOutPos <<= 1;
if (!m_parityOutPos)
{
m_parityOutPos = 1;
m_parityBuffer->pop();
}
}
return val;
}
size_t SoftwareSerial::read(uint8_t * buffer, size_t size) {
if (!m_rxValid) { return 0; }
size_t avail;
if (0 == (avail = m_buffer->pop_n(buffer, size))) {
rxBits();
avail = m_buffer->pop_n(buffer, size);
}
if (!avail) return 0;
if (m_parityBuffer) {
uint32_t parityBits = avail;
while (m_parityOutPos >>= 1) ++parityBits;
m_parityOutPos = (1 << (parityBits % 8));
m_parityBuffer->pop_n(nullptr, parityBits / 8);
}
return avail;
}
size_t SoftwareSerial::readBytes(uint8_t * buffer, size_t size) {
if (!m_rxValid || !size) { return 0; }
size_t count = 0;
const auto start = millis();
do {
count += read(&buffer[count], size - count);
if (count >= size) break;
yield();
} while (millis() - start < _timeout);
return count;
}
int SoftwareSerial::available() {
if (!m_rxValid) { return 0; }
rxBits();
int avail = m_buffer->available();
if (!avail) {
optimistic_yield(10000UL);
}
return avail;
}
void ICACHE_RAM_ATTR SoftwareSerial::preciseDelay(bool sync) {
if (!sync)
{
// Reenable interrupts while delaying to avoid other tasks piling up
if (!m_intTxEnabled) { xt_wsr_ps(m_savedPS); }
auto expired = ESP.getCycleCount() - m_periodStart;
if (expired < m_periodDuration)
{
auto ms = (m_periodDuration - expired) / ESP.getCpuFreqMHz() / 1000UL;
if (ms) delay(ms);
}
while ((ESP.getCycleCount() - m_periodStart) < m_periodDuration) { optimistic_yield(10000); }
// Disable interrupts again
if (!m_intTxEnabled) { m_savedPS = xt_rsil(15); }
}
else
{
while ((ESP.getCycleCount() - m_periodStart) < m_periodDuration) {}
}
m_periodDuration = 0;
m_periodStart = ESP.getCycleCount();
}
void ICACHE_RAM_ATTR SoftwareSerial::writePeriod(
uint32_t dutyCycle, uint32_t offCycle, bool withStopBit) {
preciseDelay(true);
if (dutyCycle)
{
digitalWrite(m_txPin, HIGH);
m_periodDuration += dutyCycle;
if (offCycle || (withStopBit && !m_invert)) preciseDelay(!withStopBit || m_invert);
}
if (offCycle)
{
digitalWrite(m_txPin, LOW);
m_periodDuration += offCycle;
if (withStopBit && m_invert) preciseDelay(false);
}
}
size_t SoftwareSerial::write(uint8_t byte) {
return write(&byte, 1);
}
size_t SoftwareSerial::write(uint8_t byte, SoftwareSerialParity parity) {
return write(&byte, 1, parity);
}
size_t SoftwareSerial::write(const uint8_t * buffer, size_t size) {
return write(buffer, size, m_parityMode);
}
size_t ICACHE_RAM_ATTR SoftwareSerial::write(const uint8_t * buffer, size_t size, SoftwareSerialParity parity) {
if (m_rxValid) { rxBits(); }
if (!m_txValid) { return -1; }
if (m_txEnableValid) {
digitalWrite(m_txEnablePin, HIGH);
}
// Stop bit: if inverted, LOW, otherwise HIGH
bool b = !m_invert;
uint32_t dutyCycle = 0;
uint32_t offCycle = 0;
if (!m_intTxEnabled) {
// Disable interrupts in order to get a clean transmit timing
m_savedPS = xt_rsil(15);
}
const uint32_t dataMask = ((1UL << m_dataBits) - 1);
bool withStopBit = true;
m_periodDuration = 0;
m_periodStart = ESP.getCycleCount();
for (size_t cnt = 0; cnt < size; ++cnt) {
uint8_t byte = ~buffer[cnt] & dataMask;
// push LSB start-data-parity-stop bit pattern into uint32_t
// Stop bits: HIGH
uint32_t word = ~0UL;
// parity bit, if any
if (parity && m_parityMode)
{
uint32_t parityBit;
switch (parity)
{
case SWSERIAL_PARITY_EVEN:
// from inverted, so use odd parity
parityBit = byte;
parityBit ^= parityBit >> 4;
parityBit &= 0xf;
parityBit = (0x9669 >> parityBit) & 1;
break;
case SWSERIAL_PARITY_ODD:
// from inverted, so use even parity
parityBit = byte;
parityBit ^= parityBit >> 4;
parityBit &= 0xf;
parityBit = (0x6996 >> parityBit) & 1;
break;
case SWSERIAL_PARITY_MARK:
parityBit = false;
break;
case SWSERIAL_PARITY_SPACE:
// suppresses warning parityBit uninitialized
default:
parityBit = true;
break;
}
word ^= parityBit << m_dataBits;
}
word ^= byte;
// Stop bit: LOW
word <<= 1;
if (m_invert) word = ~word;
for (int i = 0; i <= m_pduBits; ++i) {
bool pb = b;
b = word & (1UL << i);
if (!pb && b) {
writePeriod(dutyCycle, offCycle, withStopBit);
withStopBit = false;
dutyCycle = offCycle = 0;
}
if (b) {
dutyCycle += m_bitCycles;
}
else {
offCycle += m_bitCycles;
}
}
withStopBit = true;
}
writePeriod(dutyCycle, offCycle, true);
if (!m_intTxEnabled) {
// restore the interrupt state
xt_wsr_ps(m_savedPS);
}
if (m_txEnableValid) {
digitalWrite(m_txEnablePin, LOW);
}
return size;
}
void SoftwareSerial::flush() {
if (!m_rxValid) { return; }
m_buffer->flush();
if (m_parityBuffer)
{
m_parityInPos = m_parityOutPos = 1;
m_parityBuffer->flush();
}
}
bool SoftwareSerial::overflow() {
bool res = m_overflow;
m_overflow = false;
return res;
}
int SoftwareSerial::peek() {
if (!m_rxValid) { return -1; }
if (!m_buffer->available()) {
rxBits();
if (!m_buffer->available()) return -1;
}
auto val = m_buffer->peek();
if (m_parityBuffer) m_lastReadParity = m_parityBuffer->peek() & m_parityOutPos;
return val;
}
void SoftwareSerial::rxBits() {
int isrAvail = m_isrBuffer->available();
#ifdef ESP8266
if (m_isrOverflow.load()) {
m_overflow = true;
m_isrOverflow.store(false);
}
#else
if (m_isrOverflow.exchange(false)) {
m_overflow = true;
}
#endif
// stop bit can go undetected if leading data bits are at same level
// and there was also no next start bit yet, so one byte may be pending.
// low-cost check first
if (!isrAvail && m_rxCurBit >= -1 && m_rxCurBit < m_pduBits - m_stopBits) {
uint32_t detectionCycles = (m_pduBits - m_stopBits - m_rxCurBit) * m_bitCycles;
if (ESP.getCycleCount() - m_isrLastCycle > detectionCycles) {
// Produce faux stop bit level, prevents start bit maldetection
// cycle's LSB is repurposed for the level bit
rxBits(((m_isrLastCycle + detectionCycles) | 1) ^ m_invert);
}
}
m_isrBuffer->for_each([this](const uint32_t& isrCycle) { rxBits(isrCycle); });
}
void SoftwareSerial::rxBits(const uint32_t & isrCycle) {
bool level = (m_isrLastCycle & 1) ^ m_invert;
// error introduced by edge value in LSB of isrCycle is negligible
int32_t cycles = isrCycle - m_isrLastCycle;
m_isrLastCycle = isrCycle;
uint8_t bits = cycles / m_bitCycles;
if (cycles % m_bitCycles > (m_bitCycles >> 1)) ++bits;
while (bits > 0) {
// start bit detection
if (m_rxCurBit >= (m_pduBits - 1)) {
// leading edge of start bit
if (level) break;
m_rxCurBit = -1;
--bits;
continue;
}
// data bits
if (m_rxCurBit >= -1 && m_rxCurBit < (m_dataBits - 1)) {
int8_t dataBits = min(bits, static_cast<uint8_t>(m_dataBits - 1 - m_rxCurBit));
m_rxCurBit += dataBits;
bits -= dataBits;
m_rxCurByte >>= dataBits;
if (level) { m_rxCurByte |= (BYTE_ALL_BITS_SET << (8 - dataBits)); }
continue;
}
// parity bit
if (m_parityMode && m_rxCurBit == (m_dataBits - 1)) {
++m_rxCurBit;
--bits;
m_rxCurParity = level;
continue;
}
// stop bits
if (m_rxCurBit < (m_pduBits - m_stopBits - 1)) {
++m_rxCurBit;
--bits;
continue;
}
if (m_rxCurBit == (m_pduBits - m_stopBits - 1)) {
// Store the received value in the buffer unless we have an overflow
// if not high stop bit level, discard word
if (level)
{
m_rxCurByte >>= (sizeof(uint8_t) * 8 - m_dataBits);
if (!m_buffer->push(m_rxCurByte)) {
m_overflow = true;
}
else {
if (m_parityBuffer)
{
if (m_rxCurParity) {
m_parityBuffer->pushpeek() |= m_parityInPos;
}
else {
m_parityBuffer->pushpeek() &= ~m_parityInPos;
}
m_parityInPos <<= 1;
if (!m_parityInPos)
{
m_parityBuffer->push();
m_parityInPos = 1;
}
}
}
}
m_rxCurBit = m_pduBits;
// reset to 0 is important for masked bit logic
m_rxCurByte = 0;
m_rxCurParity = false;
break;
}
break;
}
}
void ICACHE_RAM_ATTR SoftwareSerial::rxBitISR(SoftwareSerial * self) {
uint32_t curCycle = ESP.getCycleCount();
bool level = digitalRead(self->m_rxPin);
// Store level and cycle in the buffer unless we have an overflow
// cycle's LSB is repurposed for the level bit
if (!self->m_isrBuffer->push((curCycle | 1U) ^ !level)) self->m_isrOverflow.store(true);
}
void ICACHE_RAM_ATTR SoftwareSerial::rxBitSyncISR(SoftwareSerial * self) {
uint32_t start = ESP.getCycleCount();
uint32_t wait = self->m_bitCycles - 172U;
bool level = self->m_invert;
// Store level and cycle in the buffer unless we have an overflow
// cycle's LSB is repurposed for the level bit
if (!self->m_isrBuffer->push(((start + wait) | 1U) ^ !level)) self->m_isrOverflow.store(true);
for (uint32_t i = 0; i < self->m_pduBits; ++i) {
while (ESP.getCycleCount() - start < wait) {};
wait += self->m_bitCycles;
// Store level and cycle in the buffer unless we have an overflow
// cycle's LSB is repurposed for the level bit
if (digitalRead(self->m_rxPin) != level)
{
if (!self->m_isrBuffer->push(((start + wait) | 1U) ^ level)) self->m_isrOverflow.store(true);
level = !level;
}
}
}
void SoftwareSerial::onReceive(Delegate<void(int available), void*> handler) {
receiveHandler = handler;
}
void SoftwareSerial::perform_work() {
if (!m_rxValid) { return; }
rxBits();
if (receiveHandler) {
int avail = m_buffer->available();
if (avail) { receiveHandler(avail); }
}
}

View File

@ -0,0 +1,255 @@
/*
SoftwareSerial.h
SoftwareSerial.cpp - Implementation of the Arduino software serial for ESP8266/ESP32.
Copyright (c) 2015-2016 Peter Lerup. All rights reserved.
Copyright (c) 2018-2019 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 __SoftwareSerial_h
#define __SoftwareSerial_h
#include "circular_queue/circular_queue.h"
#include <Stream.h>
enum SoftwareSerialParity : uint8_t {
SWSERIAL_PARITY_NONE = 000,
SWSERIAL_PARITY_EVEN = 020,
SWSERIAL_PARITY_ODD = 030,
SWSERIAL_PARITY_MARK = 040,
SWSERIAL_PARITY_SPACE = 070,
};
enum SoftwareSerialConfig {
SWSERIAL_5N1 = SWSERIAL_PARITY_NONE,
SWSERIAL_6N1,
SWSERIAL_7N1,
SWSERIAL_8N1,
SWSERIAL_5E1 = SWSERIAL_PARITY_EVEN,
SWSERIAL_6E1,
SWSERIAL_7E1,
SWSERIAL_8E1,
SWSERIAL_5O1 = SWSERIAL_PARITY_ODD,
SWSERIAL_6O1,
SWSERIAL_7O1,
SWSERIAL_8O1,
SWSERIAL_5M1 = SWSERIAL_PARITY_MARK,
SWSERIAL_6M1,
SWSERIAL_7M1,
SWSERIAL_8M1,
SWSERIAL_5S1 = SWSERIAL_PARITY_SPACE,
SWSERIAL_6S1,
SWSERIAL_7S1,
SWSERIAL_8S1,
SWSERIAL_5N2 = 0200 | SWSERIAL_PARITY_NONE,
SWSERIAL_6N2,
SWSERIAL_7N2,
SWSERIAL_8N2,
SWSERIAL_5E2 = 0200 | SWSERIAL_PARITY_EVEN,
SWSERIAL_6E2,
SWSERIAL_7E2,
SWSERIAL_8E2,
SWSERIAL_5O2 = 0200 | SWSERIAL_PARITY_ODD,
SWSERIAL_6O2,
SWSERIAL_7O2,
SWSERIAL_8O2,
SWSERIAL_5M2 = 0200 | SWSERIAL_PARITY_MARK,
SWSERIAL_6M2,
SWSERIAL_7M2,
SWSERIAL_8M2,
SWSERIAL_5S2 = 0200 | SWSERIAL_PARITY_SPACE,
SWSERIAL_6S2,
SWSERIAL_7S2,
SWSERIAL_8S2,
};
/// This class is compatible with the corresponding AVR one, however,
/// the constructor takes no arguments, for compatibility with the
/// HardwareSerial class.
/// Instead, the begin() function handles pin assignments and logic inversion.
/// It also has optional input buffer capacity arguments for byte buffer and ISR bit buffer.
/// Bitrates up to at least 115200 can be used.
class SoftwareSerial : public Stream {
public:
SoftwareSerial();
/// Ctor to set defaults for pins.
/// @param rxPin the GPIO pin used for RX
/// @param txPin -1 for onewire protocol, GPIO pin used for twowire TX
SoftwareSerial(int8_t rxPin, int8_t txPin = -1, bool invert = false);
SoftwareSerial(const SoftwareSerial&) = delete;
SoftwareSerial& operator= (const SoftwareSerial&) = delete;
virtual ~SoftwareSerial();
/// Configure the SoftwareSerial object for use.
/// @param baud the TX/RX bitrate
/// @param config sets databits, parity, and stop bit count
/// @param rxPin -1 or default: either no RX pin, or keeps the rxPin set in the ctor
/// @param txPin -1 or default: either no TX pin (onewire), or keeps the txPin set in the ctor
/// @param invert true: uses invert line level logic
/// @param bufCapacity the capacity for the received bytes buffer
/// @param isrBufCapacity 0: derived from bufCapacity. The capacity of the internal asynchronous
/// bit receive buffer, a suggested size is bufCapacity times the sum of
/// start, data, parity and stop bit count.
void begin(uint32_t baud, SoftwareSerialConfig config,
int8_t rxPin, int8_t txPin, bool invert,
int bufCapacity = 64, int isrBufCapacity = 0);
void begin(uint32_t baud, SoftwareSerialConfig config,
int8_t rxPin, int8_t txPin) {
begin(baud, config, rxPin, txPin, m_invert);
}
void begin(uint32_t baud, SoftwareSerialConfig config,
int8_t rxPin) {
begin(baud, config, rxPin, m_txPin, m_invert);
}
void begin(uint32_t baud, SoftwareSerialConfig config = SWSERIAL_8N1) {
begin(baud, config, m_rxPin, m_txPin, m_invert);
}
uint32_t baudRate();
/// Transmit control pin.
void setTransmitEnablePin(int8_t txEnablePin);
/// Enable or disable interrupts during tx.
void enableIntTx(bool on);
bool overflow();
int available() override;
int availableForWrite() {
if (!m_txValid) return 0;
return 1;
}
int peek() override;
int read() override;
/// @returns The verbatim parity bit associated with the last read() or peek() call
bool readParity()
{
return m_lastReadParity;
}
/// @returns The calculated bit for even parity of the parameter byte
static bool parityEven(uint8_t byte) {
byte ^= byte >> 4;
byte &= 0xf;
return (0x6996 >> byte) & 1;
}
/// @returns The calculated bit for odd parity of the parameter byte
static bool parityOdd(uint8_t byte) {
byte ^= byte >> 4;
byte &= 0xf;
return (0x9669 >> byte) & 1;
}
/// The read(buffer, size) functions are non-blocking, the same as readBytes but without timeout
size_t read(uint8_t* buffer, size_t size);
/// The read(buffer, size) functions are non-blocking, the same as readBytes but without timeout
size_t read(char* buffer, size_t size) {
return read(reinterpret_cast<uint8_t*>(buffer), size);
}
/// @returns The number of bytes read into buffer, up to size. Times out if the limit set through
/// Stream::setTimeout() is reached.
size_t readBytes(uint8_t* buffer, size_t size) override;
/// @returns The number of bytes read into buffer, up to size. Times out if the limit set through
/// Stream::setTimeout() is reached.
size_t readBytes(char* buffer, size_t size) override {
return readBytes(reinterpret_cast<uint8_t*>(buffer), size);
}
void flush() override;
size_t write(uint8_t byte) override;
size_t write(uint8_t byte, SoftwareSerialParity parity);
size_t write(const uint8_t* buffer, size_t size) override;
size_t write(const char* buffer, size_t size) {
return write(reinterpret_cast<const uint8_t*>(buffer), size);
}
size_t write(const uint8_t* buffer, size_t size, SoftwareSerialParity parity);
size_t write(const char* buffer, size_t size, SoftwareSerialParity parity) {
return write(reinterpret_cast<const uint8_t*>(buffer), size, parity);
}
operator bool() const { return m_rxValid || m_txValid; }
/// Disable or enable interrupts on the rx pin.
void enableRx(bool on);
/// One wire control.
void enableTx(bool on);
// AVR compatibility methods.
bool listen() { enableRx(true); return true; }
void end();
bool isListening() { return m_rxEnabled; }
bool stopListening() { enableRx(false); return true; }
/// Set an event handler for received data.
void onReceive(Delegate<void(int available), void*> handler);
/// Run the internal processing and event engine. Can be iteratively called
/// from loop, or otherwise scheduled.
void perform_work();
using Print::write;
private:
// If sync is false, it's legal to exceed the deadline, for instance,
// by enabling interrupts.
void preciseDelay(bool sync);
// If withStopBit is set, either cycle contains a stop bit.
// If dutyCycle == 0, the level is not forced to HIGH.
// If offCycle == 0, the level remains unchanged from dutyCycle.
void writePeriod(
uint32_t dutyCycle, uint32_t offCycle, bool withStopBit);
bool isValidGPIOpin(int8_t pin);
/* check m_rxValid that calling is safe */
void rxBits();
void rxBits(const uint32_t& isrCycle);
static void rxBitISR(SoftwareSerial* self);
static void rxBitSyncISR(SoftwareSerial* self);
// Member variables
int8_t m_rxPin = -1;
int8_t m_txPin = -1;
int8_t m_txEnablePin = -1;
uint8_t m_dataBits;
bool m_oneWire;
bool m_rxValid = false;
bool m_rxEnabled = false;
bool m_txValid = false;
bool m_txEnableValid = false;
bool m_invert;
/// PDU bits include data, parity and stop bits; the start bit is not counted.
uint8_t m_pduBits;
bool m_intTxEnabled;
SoftwareSerialParity m_parityMode;
uint8_t m_stopBits;
bool m_lastReadParity;
bool m_overflow = false;
uint32_t m_bitCycles;
uint8_t m_parityInPos;
uint8_t m_parityOutPos;
int8_t m_rxCurBit; // 0 thru (m_pduBits - m_stopBits - 1): data/parity bits. -1: start bit. (m_pduBits - 1): stop bit.
uint8_t m_rxCurByte = 0;
std::unique_ptr<circular_queue<uint8_t> > m_buffer;
std::unique_ptr<circular_queue<uint8_t> > m_parityBuffer;
uint32_t m_periodStart;
uint32_t m_periodDuration;
uint32_t m_savedPS = 0;
// the ISR stores the relative bit times in the buffer. The inversion corrected level is used as sign bit (2's complement):
// 1 = positive including 0, 0 = negative.
std::unique_ptr<circular_queue<uint32_t> > m_isrBuffer;
std::atomic<bool> m_isrOverflow;
uint32_t m_isrLastCycle;
bool m_rxCurParity = false;
Delegate<void(int available), void*> receiveHandler;
};
#endif // __SoftwareSerial_h

File diff suppressed because it is too large Load Diff

View File

@ -0,0 +1,503 @@
/*
MultiDelegate.h - A queue or event multiplexer based on the efficient Delegate
class
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 __MULTIDELEGATE_H
#define __MULTIDELEGATE_H
#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 detail
{
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...)) ? !ISQUEUE : ISQUEUE;
}
};
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 !ISQUEUE;
}
};
template< typename Delegate, typename R, bool ISQUEUE = false>
struct Call
{
static R execute(Delegate& del)
{
return del() ? !ISQUEUE : ISQUEUE;
}
};
template< typename Delegate, bool ISQUEUE>
struct Call<Delegate, void, ISQUEUE>
{
static bool execute(Delegate& del)
{
del();
return !ISQUEUE;
}
};
};
template< typename Delegate, typename R = void, bool ISQUEUE = false, uint32_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;
uint32_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:
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;
}
bool remove(const Delegate* del)
{
auto current = first;
if (!current)
return false;
Node_t* prev = nullptr;
do
{
if (del == &current->mDelegate)
{
// remove callback from stack
#ifdef ARDUINO
InterruptLock lockAllInterruptsInThisScope;
#else
std::lock_guard<std::mutex> lock(mutex_unused);
#endif
auto to_recycle = current;
// removing rLast
if (last == current)
last = prev;
current = current->mNext;
if (prev)
{
prev->mNext = current;
}
else
{
first = current;
}
if (ISQUEUE)
recycle_node_unsafe(to_recycle);
else
delete to_recycle;
return true;
}
else
{
prev = current;
current = current->mNext;
}
} while (current);
return false;
}
void operator()(P... args)
{
auto current = first;
if (!current)
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
Node_t* prev = nullptr;
// prevent execution of new callbacks during this run
auto stop = last;
bool done;
do
{
done = current == stop;
if (!CallP<Delegate, R, ISQUEUE, P...>::execute(current->mDelegate, args...))
{
// remove callback from stack
#ifdef ARDUINO
InterruptLock lockAllInterruptsInThisScope;
#else
std::lock_guard<std::mutex> lock(mutex_unused);
#endif
auto to_recycle = current;
// removing rLast
if (last == current)
last = prev;
current = current->mNext;
if (prev)
{
prev->mNext = current;
}
else
{
first = current;
}
if (ISQUEUE)
recycle_node_unsafe(to_recycle);
else
delete to_recycle;
}
else
{
prev = current;
current = current->mNext;
}
#if defined(ESP8266) || defined(ESP32)
// running callbacks might last too long for watchdog etc.
optimistic_yield(10000);
#endif
} while (current && !done);
fence.store(false);
}
};
template< typename Delegate, typename R = void, bool ISQUEUE = false, uint32_t QUEUE_CAPACITY = 32>
class MultiDelegateImpl : public MultiDelegatePImpl<Delegate, R, ISQUEUE, QUEUE_CAPACITY>
{
protected:
using typename MultiDelegatePImpl<Delegate, R, ISQUEUE, QUEUE_CAPACITY>::Node_t;
using MultiDelegatePImpl<Delegate, R, ISQUEUE, QUEUE_CAPACITY>::first;
using MultiDelegatePImpl<Delegate, R, ISQUEUE, QUEUE_CAPACITY>::last;
using MultiDelegatePImpl<Delegate, R, ISQUEUE, QUEUE_CAPACITY>::unused;
using MultiDelegatePImpl<Delegate, R, ISQUEUE, QUEUE_CAPACITY>::nodeCount;
using MultiDelegatePImpl<Delegate, R, ISQUEUE, QUEUE_CAPACITY>::recycle_node_unsafe;
#ifndef ARDUINO
using MultiDelegatePImpl<Delegate, R, ISQUEUE, QUEUE_CAPACITY>::mutex_unused;
#endif
public:
using MultiDelegatePImpl<Delegate, R, ISQUEUE, QUEUE_CAPACITY>::MultiDelegatePImpl;
void operator()()
{
auto current = first;
if (!current)
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
Node_t* prev = nullptr;
// prevent execution of new callbacks during this run
auto stop = last;
bool done;
do
{
done = current == stop;
if (!Call<Delegate, R, ISQUEUE>::execute(current->mDelegate))
{
// remove callback from stack
#ifdef ARDUINO
InterruptLock lockAllInterruptsInThisScope;
#else
std::lock_guard<std::mutex> lock(mutex_unused);
#endif
auto to_recycle = current;
// removing rLast
if (last == current)
last = prev;
current = current->mNext;
if (prev)
{
prev->mNext = current;
}
else
{
first = current;
}
if (ISQUEUE)
recycle_node_unsafe(to_recycle);
else
delete to_recycle;
}
else
{
prev = current;
current = current->mNext;
}
#if defined(ESP8266) || defined(ESP32)
// running callbacks might last too long for watchdog etc.
optimistic_yield(10000);
#endif
} while (current && !done);
fence.store(false);
}
};
template< typename Delegate, typename R, bool ISQUEUE, uint32_t QUEUE_CAPACITY, typename... P> class MultiDelegate;
template< typename Delegate, typename R, bool ISQUEUE, uint32_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, uint32_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;
};
};
/**
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, the type-conversion to bool
of that result determines if the item is immediately removed or kept after each call: a Multidelegate
queue removes an item only if true is returned, but a Multidelegate event multiplexer removes event
handlers that return false.
@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, uint32_t QUEUE_CAPACITY = 32>
class MultiDelegate : public detail::MultiDelegate<Delegate, typename Delegate::target_type, ISQUEUE, QUEUE_CAPACITY>
{
public:
using detail::MultiDelegate<Delegate, typename Delegate::target_type, ISQUEUE, QUEUE_CAPACITY>::MultiDelegate;
};
#endif // __MULTIDELEGATE_H

View File

@ -0,0 +1,399 @@
/*
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 ICACHE_RAM_ATTR
#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.
*/
T& IRAM_ATTR pushpeek()
{
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.
*/
bool IRAM_ATTR push();
/*!
@brief Move the rvalue parameter into the queue.
@return true if the queue accepted the value, false if the queue
was full.
*/
bool IRAM_ATTR push(T&& val);
/*!
@brief Push a copy of the parameter into the queue.
@return true if the queue accepted the value, false if the queue
was full.
*/
bool IRAM_ATTR push(const T& val)
{
return push(T(val));
}
#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 = {};
unsigned 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<unsigned> m_inPos;
std::atomic<unsigned> 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;
}
template< typename T, typename ForEachArg >
bool IRAM_ATTR circular_queue<T, ForEachArg>::push()
{
const auto inPos = m_inPos.load(std::memory_order_acquire);
const unsigned 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;
}
template< typename T, typename ForEachArg >
bool IRAM_ATTR circular_queue<T, ForEachArg>::push(T&& val)
{
const auto inPos = m_inPos.load(std::memory_order_acquire);
const unsigned 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;
}
#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

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>::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>::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

View File

@ -0,0 +1,92 @@
/*
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
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, unsigned 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