diff --git a/EspSoftwareSerial/LICENSE b/EspSoftwareSerial/LICENSE new file mode 100644 index 00000000..f166cc57 --- /dev/null +++ b/EspSoftwareSerial/LICENSE @@ -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. + + + Copyright (C) + + 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. + + , 1 April 1990 + Ty Coon, President of Vice + +That's all there is to it! \ No newline at end of file diff --git a/EspSoftwareSerial/README.md b/EspSoftwareSerial/README.md new file mode 100644 index 00000000..2d14c9e6 --- /dev/null +++ b/EspSoftwareSerial/README.md @@ -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 +``` diff --git a/EspSoftwareSerial/examples/loopback/loopback.ino b/EspSoftwareSerial/examples/loopback/loopback.ino new file mode 100644 index 00000000..b612bdec --- /dev/null +++ b/EspSoftwareSerial/examples/loopback/loopback.ino @@ -0,0 +1,263 @@ +#include + +// 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(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(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(); + } +} diff --git a/EspSoftwareSerial/examples/onewiretest/onewiretest.ino b/EspSoftwareSerial/examples/onewiretest/onewiretest.ino new file mode 100644 index 00000000..3e96401b --- /dev/null +++ b/EspSoftwareSerial/examples/onewiretest/onewiretest.ino @@ -0,0 +1,48 @@ +#include +#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(); + } +} diff --git a/EspSoftwareSerial/examples/repeater/repeater.ino b/EspSoftwareSerial/examples/repeater/repeater.ino new file mode 100644 index 00000000..fa5566de --- /dev/null +++ b/EspSoftwareSerial/examples/repeater/repeater.ino @@ -0,0 +1,183 @@ +#include + +// 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(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; + } +} diff --git a/EspSoftwareSerial/examples/servoTester/servoTester.ino b/EspSoftwareSerial/examples/servoTester/servoTester.ino new file mode 100644 index 00000000..cbc784d8 --- /dev/null +++ b/EspSoftwareSerial/examples/servoTester/servoTester.ino @@ -0,0 +1,115 @@ +#include +#include + +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(); + } +} diff --git a/EspSoftwareSerial/examples/swsertest/swsertest.ino b/EspSoftwareSerial/examples/swsertest/swsertest.ino new file mode 100644 index 00000000..a047c1be --- /dev/null +++ b/EspSoftwareSerial/examples/swsertest/swsertest.ino @@ -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 + +#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(); + } + +} diff --git a/EspSoftwareSerial/keywords.txt b/EspSoftwareSerial/keywords.txt new file mode 100644 index 00000000..52d48ab3 --- /dev/null +++ b/EspSoftwareSerial/keywords.txt @@ -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 diff --git a/EspSoftwareSerial/library.json b/EspSoftwareSerial/library.json new file mode 100644 index 00000000..3415b056 --- /dev/null +++ b/EspSoftwareSerial/library.json @@ -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": "*" +} diff --git a/EspSoftwareSerial/library.properties b/EspSoftwareSerial/library.properties new file mode 100644 index 00000000..bd052153 --- /dev/null +++ b/EspSoftwareSerial/library.properties @@ -0,0 +1,9 @@ +name=EspSoftwareSerial +version=6.6.1 +author=Peter Lerup, Dirk Kaar +maintainer=Peter Lerup +sentence=Implementation of the Arduino software serial for ESP8266/ESP32. +paragraph= +category=Signal Input/Output +url=https://github.com/plerup/espsoftwareserial/ +architectures=esp8266,esp32 diff --git a/EspSoftwareSerial/src/SoftwareSerial.cpp b/EspSoftwareSerial/src/SoftwareSerial.cpp new file mode 100644 index 00000000..0e7b8e08 --- /dev/null +++ b/EspSoftwareSerial/src/SoftwareSerial.cpp @@ -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 + +#ifdef ESP32 +#define xt_rsil(a) (a) +#define xt_wsr_ps(a) +#endif + +constexpr uint8_t BYTE_ALL_BITS_SET = ~static_cast(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(config & 070); + m_stopBits = 1 + ((config & 0300) ? 1 : 0); + m_pduBits = m_dataBits + static_cast(m_parityMode) + m_stopBits; + m_bitCycles = (ESP.getCpuFreqMHz() * 1000000UL + baud / 2) / baud; + m_intTxEnabled = true; + if (isValidGPIOpin(m_rxPin)) { + std::unique_ptr > buffer(new circular_queue((bufCapacity > 0) ? bufCapacity : 64)); + m_buffer = move(buffer); + if (m_parityMode) + { + std::unique_ptr > parityBuffer(new circular_queue((bufCapacity > 0) ? (bufCapacity + 7) / 8 : 8)); + m_parityBuffer = move(parityBuffer); + m_parityInPos = m_parityOutPos = 1; + } + std::unique_ptr > isrBuffer(new circular_queue((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(rxBitISR), this, CHANGE); + else + attachInterruptArg(digitalPinToInterrupt(m_rxPin), reinterpret_cast(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(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 handler) { + receiveHandler = handler; +} + +void SoftwareSerial::perform_work() { + if (!m_rxValid) { return; } + rxBits(); + if (receiveHandler) { + int avail = m_buffer->available(); + if (avail) { receiveHandler(avail); } + } +} diff --git a/EspSoftwareSerial/src/SoftwareSerial.h b/EspSoftwareSerial/src/SoftwareSerial.h new file mode 100644 index 00000000..371e3731 --- /dev/null +++ b/EspSoftwareSerial/src/SoftwareSerial.h @@ -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 + +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(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(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(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(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 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 > m_buffer; + std::unique_ptr > 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 > m_isrBuffer; + std::atomic m_isrOverflow; + uint32_t m_isrLastCycle; + bool m_rxCurParity = false; + Delegate receiveHandler; +}; + +#endif // __SoftwareSerial_h diff --git a/EspSoftwareSerial/src/circular_queue/Delegate.h b/EspSoftwareSerial/src/circular_queue/Delegate.h new file mode 100644 index 00000000..bd19c66e --- /dev/null +++ b/EspSoftwareSerial/src/circular_queue/Delegate.h @@ -0,0 +1,1786 @@ +/* +Delegate.h - An efficient interchangeable C function ptr and C++ std::function delegate +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 __Delegate_h +#define __Delegate_h + +#if defined(ESP8266) +#include +#elif defined(ESP32) +#include +#else +#define ICACHE_RAM_ATTR +#define IRAM_ATTR +#endif + +#if !defined(ARDUINO) || defined(ESP8266) || defined(ESP32) +#include +#include +#else +#include "circular_queue/ghostl.h" +#endif + +namespace detail +{ + template + static R IRAM_ATTR vPtrToFunPtrExec(void* fn, P... args) + { + using target_type = R(P...); + return reinterpret_cast(fn)(std::forward(args...)); + } + +#if !defined(ARDUINO) || defined(ESP8266) || defined(ESP32) + template + class DelegatePImpl { + public: + using target_type = R(P...); + protected: + using FunPtr = target_type*; + using FunAPtr = R(*)(A, P...); + using FunVPPtr = R(*)(void*, P...); + using FunctionType = std::function; + public: + DelegatePImpl() + { + kind = FP; + fn = nullptr; + } + + DelegatePImpl(std::nullptr_t) + { + kind = FP; + fn = nullptr; + } + + ~DelegatePImpl() + { + if (FUNC == kind) + functional.~FunctionType(); + else if (FPA == kind) + obj.~A(); + } + + DelegatePImpl(const DelegatePImpl& del) + { + kind = del.kind; + if (FUNC == del.kind) + { + new (&functional) FunctionType(del.functional); + } + else if (FPA == del.kind) + { + fnA = del.fnA; + new (&obj) A(del.obj); + } + else + { + fn = del.fn; + } + } + + DelegatePImpl(DelegatePImpl&& del) + { + kind = del.kind; + if (FUNC == del.kind) + { + new (&functional) FunctionType(std::move(del.functional)); + } + else if (FPA == del.kind) + { + fnA = del.fnA; + new (&obj) A(std::move(del.obj)); + } + else + { + fn = del.fn; + } + } + + DelegatePImpl(FunAPtr fnA, const A& obj) + { + kind = FPA; + DelegatePImpl::fnA = fnA; + new (&this->obj) A(obj); + } + + DelegatePImpl(FunAPtr fnA, A&& obj) + { + kind = FPA; + DelegatePImpl::fnA = fnA; + new (&this->obj) A(std::move(obj)); + } + + DelegatePImpl(FunPtr fn) + { + kind = FP; + DelegatePImpl::fn = fn; + } + + template DelegatePImpl(F functional) + { + kind = FUNC; + new (&this->functional) FunctionType(std::forward(functional)); + } + + DelegatePImpl& operator=(const DelegatePImpl& del) + { + if (this == &del) return *this; + if (kind != del.kind) + { + if (FUNC == kind) + { + functional.~FunctionType(); + } + else if (FPA == kind) + { + obj.~A(); + } + if (FUNC == del.kind) + { + new (&this->functional) FunctionType(); + } + else if (FPA == del.kind) + { + new (&obj) A; + } + kind = del.kind; + } + if (FUNC == del.kind) + { + functional = del.functional; + } + else if (FPA == del.kind) + { + fnA = del.fnA; + obj = del.obj; + } + else + { + fn = del.fn; + } + return *this; + } + + DelegatePImpl& operator=(DelegatePImpl&& del) + { + if (this == &del) return *this; + if (kind != del.kind) + { + if (FUNC == kind) + { + functional.~FunctionType(); + } + else if (FPA == kind) + { + obj.~A(); + } + if (FUNC == del.kind) + { + new (&this->functional) FunctionType(); + } + else if (FPA == del.kind) + { + new (&obj) A; + } + kind = del.kind; + } + if (FUNC == del.kind) + { + functional = std::move(del.functional); + } + else if (FPA == del.kind) + { + fnA = del.fnA; + obj = std::move(del.obj); + } + else + { + fn = del.fn; + } + return *this; + } + + DelegatePImpl& operator=(FunPtr fn) + { + if (FUNC == kind) + { + functional.~FunctionType(); + } + else if (FPA == kind) + { + obj.~A(); + } + kind = FP; + this->fn = fn; + return *this; + } + + DelegatePImpl& IRAM_ATTR operator=(std::nullptr_t) + { + if (FUNC == kind) + { + functional.~FunctionType(); + } + else if (FPA == kind) + { + obj.~A(); + } + kind = FP; + fn = nullptr; + return *this; + } + + operator bool() const + { + if (FP == kind) + { + return fn; + } + else if (FPA == kind) + { + return fnA; + } + else + { + return functional ? true : false; + } + } + + static R IRAM_ATTR vPtrToFunAPtrExec(void* self, P... args) + { + return static_cast(self)->fnA( + static_cast(self)->obj, + std::forward(args...)); + }; + + operator FunVPPtr() const + { + if (FP == kind) + { + return vPtrToFunPtrExec; + } + else if (FPA == kind) + { + return vPtrToFunAPtrExec; + } + else + { + return [](void* self, P... args) -> R + { + return static_cast(self)->functional(std::forward(args...)); + }; + } + } + + void* arg() const + { + if (FP == kind) + { + return reinterpret_cast(fn); + } + else + { + return const_cast(this); + } + } + + operator FunctionType() const + { + if (FP == kind) + { + return fn; + } + else if (FPA == kind) + { + return [this](P... args) { return fnA(obj, std::forward(args...)); }; + } + else + { + return functional; + } + } + + R IRAM_ATTR operator()(P... args) const + { + if (FP == kind) + { + return fn(std::forward(args...)); + } + else if (FPA == kind) + { + return fnA(obj, std::forward(args...)); + } + else + { + return functional(std::forward(args...)); + } + } + + protected: + enum { FUNC, FP, FPA } kind; + union { + FunctionType functional; + FunPtr fn; + struct { + FunAPtr fnA; + A obj; + }; + }; + }; +#else + template + class DelegatePImpl { + public: + using target_type = R(P...); + protected: + using FunPtr = target_type*; + using FunAPtr = R(*)(A, P...); + using FunVPPtr = R(*)(void*, P...); + public: + DelegatePImpl() + { + kind = FP; + fn = nullptr; + } + + DelegatePImpl(std::nullptr_t) + { + kind = FP; + fn = nullptr; + } + + DelegatePImpl(const DelegatePImpl& del) + { + kind = del.kind; + if (FPA == del.kind) + { + fnA = del.fnA; + obj = del.obj; + } + else + { + fn = del.fn; + } + } + + DelegatePImpl(DelegatePImpl&& del) + { + kind = del.kind; + if (FPA == del.kind) + { + fnA = del.fnA; + obj = std::move(del.obj); + } + else + { + fn = del.fn; + } + } + + DelegatePImpl(FunAPtr fnA, const A& obj) + { + kind = FPA; + DelegatePImpl::fnA = fnA; + this->obj = obj; + } + + DelegatePImpl(FunAPtr fnA, A&& obj) + { + kind = FPA; + DelegatePImpl::fnA = fnA; + this->obj = std::move(obj); + } + + DelegatePImpl(FunPtr fn) + { + kind = FP; + DelegatePImpl::fn = fn; + } + + template DelegatePImpl(F fn) + { + kind = FP; + DelegatePImpl::fn = std::forward(fn); + } + + DelegatePImpl& operator=(const DelegatePImpl& del) + { + if (this == &del) return *this; + if (kind != del.kind) + { + if (FPA == kind) + { + obj = {}; + } + kind = del.kind; + } + if (FPA == del.kind) + { + fnA = del.fnA; + obj = del.obj; + } + else + { + fn = del.fn; + } + return *this; + } + + DelegatePImpl& operator=(DelegatePImpl&& del) + { + if (this == &del) return *this; + if (kind != del.kind) + { + if (FPA == kind) + { + obj = {}; + } + kind = del.kind; + } + if (FPA == del.kind) + { + fnA = del.fnA; + obj = std::move(del.obj); + } + else + { + fn = del.fn; + } + return *this; + } + + DelegatePImpl& operator=(FunPtr fn) + { + if (FPA == kind) + { + obj = {}; + } + kind = FP; + this->fn = fn; + return *this; + } + + DelegatePImpl& IRAM_ATTR operator=(std::nullptr_t) + { + if (FPA == kind) + { + obj = {}; + } + kind = FP; + fn = nullptr; + return *this; + } + + operator bool() const + { + if (FP == kind) + { + return fn; + } + else + { + return fnA; + } + } + + static R IRAM_ATTR vPtrToFunAPtrExec(void* self, P... args) + { + return static_cast(self)->fnA( + static_cast(self)->obj, + std::forward(args...)); + }; + + operator FunVPPtr() const + { + if (FP == kind) + { + return vPtrToFunPtrExec; + } + else + { + return vPtrToFunAPtrExec; + } + } + + void* arg() const + { + if (FP == kind) + { + return reinterpret_cast(fn); + } + else + { + return const_cast(this); + } + } + + R IRAM_ATTR operator()(P... args) const + { + if (FP == kind) + { + return fn(std::forward(args...)); + } + else + { + return fnA(obj, std::forward(args...)); + } + } + + protected: + enum { FP, FPA } kind; + union { + FunPtr fn; + FunAPtr fnA; + }; + A obj; + }; +#endif + +#if !defined(ARDUINO) || defined(ESP8266) || defined(ESP32) + template + class DelegatePImpl { + public: + using target_type = R(P...); + protected: + using FunPtr = target_type*; + using FunctionType = std::function; + using FunVPPtr = R(*)(void*, P...); + public: + DelegatePImpl() + { + kind = FP; + fn = nullptr; + } + + DelegatePImpl(std::nullptr_t) + { + kind = FP; + fn = nullptr; + } + + ~DelegatePImpl() + { + if (FUNC == kind) + functional.~FunctionType(); + } + + DelegatePImpl(const DelegatePImpl& del) + { + kind = del.kind; + if (FUNC == del.kind) + { + new (&functional) FunctionType(del.functional); + } + else + { + fn = del.fn; + } + } + + DelegatePImpl(DelegatePImpl&& del) + { + kind = del.kind; + if (FUNC == del.kind) + { + new (&functional) FunctionType(std::move(del.functional)); + } + else + { + fn = del.fn; + } + } + + DelegatePImpl(FunPtr fn) + { + kind = FP; + DelegatePImpl::fn = fn; + } + + template DelegatePImpl(F functional) + { + kind = FUNC; + new (&this->functional) FunctionType(std::forward(functional)); + } + + DelegatePImpl& operator=(const DelegatePImpl& del) + { + if (this == &del) return *this; + if (FUNC == kind && FUNC != del.kind) + { + functional.~FunctionType(); + } + else if (FUNC != kind && FUNC == del.kind) + { + new (&this->functional) FunctionType(); + } + kind = del.kind; + if (FUNC == del.kind) + { + functional = del.functional; + } + else + { + fn = del.fn; + } + return *this; + } + + DelegatePImpl& operator=(DelegatePImpl&& del) + { + if (this == &del) return *this; + if (FUNC == kind && FUNC != del.kind) + { + functional.~FunctionType(); + } + else if (FUNC != kind && FUNC == del.kind) + { + new (&this->functional) FunctionType(); + } + kind = del.kind; + if (FUNC == del.kind) + { + functional = std::move(del.functional); + } + else + { + fn = del.fn; + } + return *this; + } + + DelegatePImpl& operator=(FunPtr fn) + { + if (FUNC == kind) + { + functional.~FunctionType(); + kind = FP; + } + DelegatePImpl::fn = fn; + return *this; + } + + DelegatePImpl& IRAM_ATTR operator=(std::nullptr_t) + { + if (FUNC == kind) + { + functional.~FunctionType(); + } + kind = FP; + fn = nullptr; + return *this; + } + + operator bool() const + { + if (FP == kind) + { + return fn; + } + else + { + return functional ? true : false; + } + } + + operator FunVPPtr() const + { + if (FP == kind) + { + return vPtrToFunPtrExec; + } + else + { + return [](void* self, P... args) -> R + { + return static_cast(self)->functional(std::forward(args...)); + }; + } + } + + void* arg() const + { + if (FP == kind) + { + return reinterpret_cast(fn); + } + else + { + return const_cast(this); + } + } + + operator FunctionType() const + { + if (FP == kind) + { + return fn; + } + else + { + return functional; + } + } + + R IRAM_ATTR operator()(P... args) const + { + if (FP == kind) + { + return fn(std::forward(args...)); + } + else + { + return functional(std::forward(args...)); + } + } + + protected: + enum { FUNC, FP } kind; + union { + FunctionType functional; + FunPtr fn; + }; + }; +#else + template + class DelegatePImpl { + public: + using target_type = R(P...); + protected: + using FunPtr = target_type*; + using FunVPPtr = R(*)(void*, P...); + public: + DelegatePImpl() + { + fn = nullptr; + } + + DelegatePImpl(std::nullptr_t) + { + fn = nullptr; + } + + DelegatePImpl(const DelegatePImpl& del) + { + fn = del.fn; + } + + DelegatePImpl(DelegatePImpl&& del) + { + fn = std::move(del.fn); + } + + DelegatePImpl(FunPtr fn) + { + DelegatePImpl::fn = fn; + } + + template DelegatePImpl(F fn) + { + DelegatePImpl::fn = std::forward(fn); + } + + DelegatePImpl& operator=(const DelegatePImpl& del) + { + if (this == &del) return *this; + fn = del.fn; + return *this; + } + + DelegatePImpl& operator=(DelegatePImpl&& del) + { + if (this == &del) return *this; + fn = std::move(del.fn); + return *this; + } + + DelegatePImpl& operator=(FunPtr fn) + { + DelegatePImpl::fn = fn; + return *this; + } + + DelegatePImpl& IRAM_ATTR operator=(std::nullptr_t) + { + fn = nullptr; + return *this; + } + + operator bool() const + { + return fn; + } + + operator FunVPPtr() const + { + return vPtrToFunPtrExec; + } + + void* arg() const + { + return reinterpret_cast(fn); + } + + R IRAM_ATTR operator()(P... args) const + { + return fn(std::forward(args...)); + } + + protected: + FunPtr fn; + }; +#endif + +#if !defined(ARDUINO) || defined(ESP8266) || defined(ESP32) + template + class DelegateImpl { + public: + using target_type = R(); + protected: + using FunPtr = target_type*; + using FunAPtr = R(*)(A); + using FunctionType = std::function; + using FunVPPtr = R(*)(void*); + public: + DelegateImpl() + { + kind = FP; + fn = nullptr; + } + + DelegateImpl(std::nullptr_t) + { + kind = FP; + fn = nullptr; + } + + ~DelegateImpl() + { + if (FUNC == kind) + functional.~FunctionType(); + else if (FPA == kind) + obj.~A(); + } + + DelegateImpl(const DelegateImpl& del) + { + kind = del.kind; + if (FUNC == del.kind) + { + new (&functional) FunctionType(del.functional); + } + else if (FPA == del.kind) + { + fnA = del.fnA; + new (&obj) A(del.obj); + } + else + { + fn = del.fn; + } + } + + DelegateImpl(DelegateImpl&& del) + { + kind = del.kind; + if (FUNC == del.kind) + { + new (&functional) FunctionType(std::move(del.functional)); + } + else if (FPA == del.kind) + { + fnA = del.fnA; + new (&obj) A(std::move(del.obj)); + } + else + { + fn = del.fn; + } + } + + DelegateImpl(FunAPtr fnA, const A& obj) + { + kind = FPA; + DelegateImpl::fnA = fnA; + new (&this->obj) A(obj); + } + + DelegateImpl(FunAPtr fnA, A&& obj) + { + kind = FPA; + DelegateImpl::fnA = fnA; + new (&this->obj) A(std::move(obj)); + } + + DelegateImpl(FunPtr fn) + { + kind = FP; + DelegateImpl::fn = fn; + } + + template DelegateImpl(F functional) + { + kind = FUNC; + new (&this->functional) FunctionType(std::forward(functional)); + } + + DelegateImpl& operator=(const DelegateImpl& del) + { + if (this == &del) return *this; + if (kind != del.kind) + { + if (FUNC == kind) + { + functional.~FunctionType(); + } + else if (FPA == kind) + { + obj.~A(); + } + if (FUNC == del.kind) + { + new (&this->functional) FunctionType(); + } + else if (FPA == del.kind) + { + new (&obj) A; + } + kind = del.kind; + } + if (FUNC == del.kind) + { + functional = del.functional; + } + else if (FPA == del.kind) + { + fnA = del.fnA; + obj = del.obj; + } + else + { + fn = del.fn; + } + return *this; + } + + DelegateImpl& operator=(DelegateImpl&& del) + { + if (this == &del) return *this; + if (kind != del.kind) + { + if (FUNC == kind) + { + functional.~FunctionType(); + } + else if (FPA == kind) + { + obj.~A(); + } + if (FUNC == del.kind) + { + new (&this->functional) FunctionType(); + } + else if (FPA == del.kind) + { + new (&obj) A; + } + kind = del.kind; + } + if (FUNC == del.kind) + { + functional = std::move(del.functional); + } + else if (FPA == del.kind) + { + fnA = del.fnA; + obj = std::move(del.obj); + } + else + { + fn = del.fn; + } + return *this; + } + + DelegateImpl& operator=(FunPtr fn) + { + if (FUNC == kind) + { + functional.~FunctionType(); + } + else if (FPA == kind) + { + obj.~A(); + } + kind = FP; + this->fn = fn; + return *this; + } + + DelegateImpl& IRAM_ATTR operator=(std::nullptr_t) + { + if (FUNC == kind) + { + functional.~FunctionType(); + } + else if (FPA == kind) + { + obj.~A(); + } + kind = FP; + fn = nullptr; + return *this; + } + + operator bool() const + { + if (FP == kind) + { + return fn; + } + else if (FPA == kind) + { + return fnA; + } + else + { + return functional ? true : false; + } + } + + static R IRAM_ATTR vPtrToFunAPtrExec(void* self) + { + return static_cast(self)->fnA( + static_cast(self)->obj); + }; + + operator FunVPPtr() const + { + if (FP == kind) + { + return reinterpret_cast(fn); + } + else if (FPA == kind) + { + return vPtrToFunAPtrExec; + } + else + { + return [](void* self) -> R + { + return static_cast(self)->functional(); + }; + } + } + + void* arg() const + { + if (FP == kind) + { + return nullptr; + } + else + { + return const_cast(this); + } + } + + operator FunctionType() const + { + if (FP == kind) + { + return fn; + } + else if (FPA == kind) + { + return [this]() { return fnA(obj); }; + } + else + { + return functional; + } + } + + R IRAM_ATTR operator()() const + { + if (FP == kind) + { + return fn(); + } + else if (FPA == kind) + { + return fnA(obj); + } + else + { + return functional(); + } + } + + protected: + enum { FUNC, FP, FPA } kind; + union { + FunctionType functional; + FunPtr fn; + struct { + FunAPtr fnA; + A obj; + }; + }; + }; +#else + template + class DelegateImpl { + public: + using target_type = R(); + protected: + using FunPtr = target_type*; + using FunAPtr = R(*)(A); + using FunVPPtr = R(*)(void*); + public: + DelegateImpl() + { + kind = FP; + fn = nullptr; + } + + DelegateImpl(std::nullptr_t) + { + kind = FP; + fn = nullptr; + } + + DelegateImpl(const DelegateImpl& del) + { + kind = del.kind; + if (FPA == del.kind) + { + fnA = del.fnA; + obj = del.obj; + } + else + { + fn = del.fn; + } + } + + DelegateImpl(DelegateImpl&& del) + { + kind = del.kind; + if (FPA == del.kind) + { + fnA = del.fnA; + obj = std::move(del.obj); + } + else + { + fn = del.fn; + } + } + + DelegateImpl(FunAPtr fnA, const A& obj) + { + kind = FPA; + DelegateImpl::fnA = fnA; + this->obj = obj; + } + + DelegateImpl(FunAPtr fnA, A&& obj) + { + kind = FPA; + DelegateImpl::fnA = fnA; + this->obj = std::move(obj); + } + + DelegateImpl(FunPtr fn) + { + kind = FP; + DelegateImpl::fn = fn; + } + + template DelegateImpl(F fn) + { + kind = FP; + DelegateImpl::fn = std::forward(fn); + } + + DelegateImpl& operator=(const DelegateImpl& del) + { + if (this == &del) return *this; + if (kind != del.kind) + { + if (FPA == kind) + { + obj = {}; + } + kind = del.kind; + } + if (FPA == del.kind) + { + fnA = del.fnA; + obj = del.obj; + } + else + { + fn = del.fn; + } + return *this; + } + + DelegateImpl& operator=(DelegateImpl&& del) + { + if (this == &del) return *this; + if (kind != del.kind) + { + if (FPA == kind) + { + obj = {}; + } + kind = del.kind; + } + if (FPA == del.kind) + { + fnA = del.fnA; + obj = std::move(del.obj); + } + else + { + fn = del.fn; + } + return *this; + } + + DelegateImpl& operator=(FunPtr fn) + { + if (FPA == kind) + { + obj = {}; + } + kind = FP; + this->fn = fn; + return *this; + } + + DelegateImpl& IRAM_ATTR operator=(std::nullptr_t) + { + if (FPA == kind) + { + obj = {}; + } + kind = FP; + fn = nullptr; + return *this; + } + + operator bool() const + { + if (FP == kind) + { + return fn; + } + else + { + return fnA; + } + } + + static R IRAM_ATTR vPtrToFunAPtrExec(void* self) + { + return static_cast(self)->fnA( + static_cast(self)->obj); + }; + + operator FunVPPtr() const + { + if (FP == kind) + { + return reinterpret_cast(fn); + } + else + { + return vPtrToFunAPtrExec; + } + } + + void* arg() const + { + if (FP == kind) + { + return nullptr; + } + else + { + return const_cast(this); + } + } + + R IRAM_ATTR operator()() const + { + if (FP == kind) + { + return fn(); + } + else + { + return fnA(obj); + } + } + + protected: + enum { FP, FPA } kind; + union { + FunPtr fn; + FunAPtr fnA; + }; + A obj; + }; +#endif + +#if !defined(ARDUINO) || defined(ESP8266) || defined(ESP32) + template + class DelegateImpl { + public: + using target_type = R(); + protected: + using FunPtr = target_type*; + using FunctionType = std::function; + using FunVPPtr = R(*)(void*); + public: + DelegateImpl() + { + kind = FP; + fn = nullptr; + } + + DelegateImpl(std::nullptr_t) + { + kind = FP; + fn = nullptr; + } + + ~DelegateImpl() + { + if (FUNC == kind) + functional.~FunctionType(); + } + + DelegateImpl(const DelegateImpl& del) + { + kind = del.kind; + if (FUNC == del.kind) + { + new (&functional) FunctionType(del.functional); + } + else + { + fn = del.fn; + } + } + + DelegateImpl(DelegateImpl&& del) + { + kind = del.kind; + if (FUNC == del.kind) + { + new (&functional) FunctionType(std::move(del.functional)); + } + else + { + fn = del.fn; + } + } + + DelegateImpl(FunPtr fn) + { + kind = FP; + DelegateImpl::fn = fn; + } + + template DelegateImpl(F functional) + { + kind = FUNC; + new (&this->functional) FunctionType(std::forward(functional)); + } + + DelegateImpl& operator=(const DelegateImpl& del) + { + if (this == &del) return *this; + if (FUNC == kind && FUNC != del.kind) + { + functional.~FunctionType(); + } + else if (FUNC != kind && FUNC == del.kind) + { + new (&this->functional) FunctionType(); + } + kind = del.kind; + if (FUNC == del.kind) + { + functional = del.functional; + } + else + { + fn = del.fn; + } + return *this; + } + + DelegateImpl& operator=(DelegateImpl&& del) + { + if (this == &del) return *this; + if (FUNC == kind && FUNC != del.kind) + { + functional.~FunctionType(); + } + else if (FUNC != kind && FUNC == del.kind) + { + new (&this->functional) FunctionType(); + } + kind = del.kind; + if (FUNC == del.kind) + { + functional = std::move(del.functional); + } + else + { + fn = del.fn; + } + return *this; + } + + DelegateImpl& operator=(FunPtr fn) + { + if (FUNC == kind) + { + functional.~FunctionType(); + kind = FP; + } + DelegateImpl::fn = fn; + return *this; + } + + DelegateImpl& IRAM_ATTR operator=(std::nullptr_t) + { + if (FUNC == kind) + { + functional.~FunctionType(); + } + kind = FP; + fn = nullptr; + return *this; + } + + operator bool() const + { + if (FP == kind) + { + return fn; + } + else + { + return functional ? true : false; + } + } + + operator FunVPPtr() const + { + if (FP == kind) + { + return reinterpret_cast(fn); + } + else + { + return [](void* self) -> R + { + return static_cast(self)->functional(); + }; + } + } + + void* arg() const + { + if (FP == kind) + { + return nullptr; + } + else + { + return const_cast(this); + } + } + + operator FunctionType() const + { + if (FP == kind) + { + return fn; + } + else + { + return functional; + } + } + + R IRAM_ATTR operator()() const + { + if (FP == kind) + { + return fn(); + } + else + { + return functional(); + } + } + + protected: + enum { FUNC, FP } kind; + union { + FunctionType functional; + FunPtr fn; + }; + }; +#else + template + class DelegateImpl { + public: + using target_type = R(); + protected: + using FunPtr = target_type*; + using FunVPPtr = R(*)(void*); + public: + DelegateImpl() + { + fn = nullptr; + } + + DelegateImpl(std::nullptr_t) + { + fn = nullptr; + } + + DelegateImpl(const DelegateImpl& del) + { + fn = del.fn; + } + + DelegateImpl(DelegateImpl&& del) + { + fn = std::move(del.fn); + } + + DelegateImpl(FunPtr fn) + { + DelegateImpl::fn = fn; + } + + template DelegateImpl(F fn) + { + DelegateImpl::fn = std::forward(fn); + } + + DelegateImpl& operator=(const DelegateImpl& del) + { + if (this == &del) return *this; + fn = del.fn; + return *this; + } + + DelegateImpl& operator=(DelegateImpl&& del) + { + if (this == &del) return *this; + fn = std::move(del.fn); + return *this; + } + + DelegateImpl& operator=(FunPtr fn) + { + DelegateImpl::fn = fn; + return *this; + } + + DelegateImpl& IRAM_ATTR operator=(std::nullptr_t) + { + fn = nullptr; + return *this; + } + + operator bool() const + { + return fn; + } + + operator FunVPPtr() const + { + return reinterpret_cast(fn); + } + + void* arg() const + { + return nullptr; + } + + R IRAM_ATTR operator()() const + { + return fn(); + } + + protected: + FunPtr fn; + }; +#endif + + template + class Delegate : private detail::DelegatePImpl + { + private: + using typename detail::DelegatePImpl::FunVPPtr; +#if !defined(ARDUINO) || defined(ESP8266) || defined(ESP32) + using typename detail::DelegatePImpl::FunctionType; +#endif + public: + using detail::DelegatePImpl::target_type; + using detail::DelegatePImpl::DelegatePImpl; + using detail::DelegatePImpl::operator=; + using detail::DelegatePImpl::operator bool; + using detail::DelegatePImpl::operator FunVPPtr; + using detail::DelegatePImpl::arg; +#if !defined(ARDUINO) || defined(ESP8266) || defined(ESP32) + using detail::DelegatePImpl::operator FunctionType; +#endif + using detail::DelegatePImpl::operator(); + }; + + template + class Delegate : private detail::DelegatePImpl + { + private: + using typename detail::DelegatePImpl::FunVPPtr; +#if !defined(ARDUINO) || defined(ESP8266) || defined(ESP32) + using typename detail::DelegatePImpl::FunctionType; +#endif + public: + using detail::DelegatePImpl::target_type; + using detail::DelegatePImpl::DelegatePImpl; + using detail::DelegatePImpl::operator=; + using detail::DelegatePImpl::operator bool; +#if !defined(ARDUINO) || defined(ESP8266) || defined(ESP32) + using detail::DelegatePImpl::operator FunctionType; +#endif + using detail::DelegatePImpl::operator(); + operator FunVPPtr() const + { + if (detail::DelegatePImpl::FPA == detail::DelegatePImpl::kind) + { + return reinterpret_cast(detail::DelegatePImpl::fnA); + } + else + { + return detail::DelegatePImpl::operator FunVPPtr(); + } + } + void* arg() const + { + if (detail::DelegatePImpl::FPA == detail::DelegatePImpl::kind) + { + return detail::DelegatePImpl::obj; + } + else + { + return detail::DelegatePImpl::arg(); + } + } + }; + + template + class Delegate : private detail::DelegateImpl + { + private: + using typename detail::DelegateImpl::FunVPPtr; +#if !defined(ARDUINO) || defined(ESP8266) || defined(ESP32) + using typename detail::DelegateImpl::FunctionType; +#endif + public: + using detail::DelegateImpl::target_type; + using detail::DelegateImpl::DelegateImpl; + using detail::DelegateImpl::operator=; + using detail::DelegateImpl::operator bool; + using detail::DelegateImpl::operator FunVPPtr; + using detail::DelegateImpl::arg; +#if !defined(ARDUINO) || defined(ESP8266) || defined(ESP32) + using detail::DelegateImpl::operator FunctionType; +#endif + using detail::DelegateImpl::operator(); + }; + + template + class Delegate : private detail::DelegateImpl + { + private: + using typename detail::DelegateImpl::FunVPPtr; +#if !defined(ARDUINO) || defined(ESP8266) || defined(ESP32) + using typename detail::DelegateImpl::FunctionType; +#endif + public: + using detail::DelegateImpl::target_type; + using detail::DelegateImpl::DelegateImpl; + using detail::DelegateImpl::operator=; + using detail::DelegateImpl::operator bool; +#if !defined(ARDUINO) || defined(ESP8266) || defined(ESP32) + using detail::DelegateImpl::operator FunctionType; +#endif + using detail::DelegateImpl::operator(); + operator FunVPPtr() const + { + if (detail::DelegateImpl::FPA == detail::DelegateImpl::kind) + { + return reinterpret_cast(detail::DelegateImpl::fnA); + } + else + { + return detail::DelegateImpl::operator FunVPPtr(); + } + } + void* arg() const + { + if (detail::DelegateImpl::FPA == detail::DelegateImpl::kind) + { + return detail::DelegateImpl::obj; + } + else + { + return detail::DelegateImpl::arg(); + } + } + }; + +} + +template class Delegate; +template class Delegate : public detail::Delegate +{ +public: + using detail::Delegate::Delegate; +}; +template class Delegate : public detail::Delegate +{ +public: + using detail::Delegate::Delegate; +}; + +#endif // __Delegate_h diff --git a/EspSoftwareSerial/src/circular_queue/MultiDelegate.h b/EspSoftwareSerial/src/circular_queue/MultiDelegate.h new file mode 100644 index 00000000..1fd4188d --- /dev/null +++ b/EspSoftwareSerial/src/circular_queue/MultiDelegate.h @@ -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 +#else +#include "circular_queue/ghostl.h" +#endif + +#if defined(ESP8266) +#include +using esp8266::InterruptLock; +#elif defined(ARDUINO) +class InterruptLock { +public: + InterruptLock() { + noInterrupts(); + } + ~InterruptLock() { + interrupts(); + } +}; +#else +#include +#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(args...)) ? !ISQUEUE : ISQUEUE; + } + }; + + template< typename Delegate, bool ISQUEUE, typename... P> + struct CallP + { + static bool execute(Delegate& del, P... args) + { + del(std::forward(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 + { + 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 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 == ¤t->mDelegate) + { + // remove callback from stack +#ifdef ARDUINO + InterruptLock lockAllInterruptsInThisScope; +#else + std::lock_guard 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 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::execute(current->mDelegate, args...)) + { + // remove callback from stack +#ifdef ARDUINO + InterruptLock lockAllInterruptsInThisScope; +#else + std::lock_guard 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 + { + protected: + using typename MultiDelegatePImpl::Node_t; + using MultiDelegatePImpl::first; + using MultiDelegatePImpl::last; + using MultiDelegatePImpl::unused; + using MultiDelegatePImpl::nodeCount; + using MultiDelegatePImpl::recycle_node_unsafe; +#ifndef ARDUINO + using MultiDelegatePImpl::mutex_unused; +#endif + + public: + using MultiDelegatePImpl::MultiDelegatePImpl; + + void operator()() + { + auto current = first; + if (!current) + return; + + static std::atomic 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::execute(current->mDelegate)) + { + // remove callback from stack +#ifdef ARDUINO + InterruptLock lockAllInterruptsInThisScope; +#else + std::lock_guard 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 : public MultiDelegatePImpl + { + public: + using MultiDelegatePImpl::MultiDelegatePImpl; + }; + + template< typename Delegate, typename R, bool ISQUEUE, uint32_t QUEUE_CAPACITY> + class MultiDelegate : public MultiDelegateImpl + { + public: + using MultiDelegateImpl::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 +{ +public: + using detail::MultiDelegate::MultiDelegate; +}; + +#endif // __MULTIDELEGATE_H diff --git a/EspSoftwareSerial/src/circular_queue/circular_queue.h b/EspSoftwareSerial/src/circular_queue/circular_queue.h new file mode 100644 index 00000000..46e3f66e --- /dev/null +++ b/EspSoftwareSerial/src/circular_queue/circular_queue.h @@ -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 +#endif + +#if defined(ESP8266) || defined(ESP32) || !defined(ARDUINO) +#include +#include +#include +#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(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(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& fun); +#else + void for_each(Delegate 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& fun); +#else + bool for_each_rev_requeue(Delegate fun); +#endif + +protected: + const T defaultValue = {}; + unsigned m_bufSize; +#if defined(ESP8266) || defined(ESP32) || !defined(ARDUINO) + std::unique_ptr m_buffer; +#else + std::unique_ptr m_buffer; +#endif + std::atomic m_inPos; + std::atomic m_outPos; +}; + +template< typename T, typename ForEachArg > +bool circular_queue::capacity(const size_t cap) +{ + if (cap + 1 == m_bufSize) return true; + else if (available() > cap) return false; + std::unique_ptr 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::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::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::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(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::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::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(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::for_each(const Delegate& fun) +#else +void circular_queue::for_each(Delegate 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::for_each_rev_requeue(const Delegate& fun) +#else +bool circular_queue::for_each_rev_requeue(Delegate fun) +#endif +{ + auto inPos0 = circular_queue::m_inPos.load(std::memory_order_acquire); + auto outPos = circular_queue::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::m_bufSize - 1; + do { + pos = (pos + posDecr) % circular_queue::m_bufSize; + T&& val = std::move(circular_queue::m_buffer[pos]); + if (fun(val)) + { + outPos1 = (outPos1 + posDecr) % circular_queue::m_bufSize; + if (outPos1 != pos) circular_queue::m_buffer[outPos1] = std::move(val); + } + } while (pos != outPos); + circular_queue::m_outPos.store(outPos1, std::memory_order_release); + return true; +} + +#endif // __circular_queue_h diff --git a/EspSoftwareSerial/src/circular_queue/circular_queue_mp.h b/EspSoftwareSerial/src/circular_queue/circular_queue_mp.h new file mode 100644 index 00000000..7024247a --- /dev/null +++ b/EspSoftwareSerial/src/circular_queue/circular_queue_mp.h @@ -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 +#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 +{ +public: + circular_queue_mp() = default; + circular_queue_mp(const size_t capacity) : circular_queue(capacity) + {} + circular_queue_mp(circular_queue&& cq) : circular_queue(std::move(cq)) + {} + using circular_queue::operator=; + using circular_queue::capacity; + using circular_queue::flush; + using circular_queue::available; + using circular_queue::available_for_push; + using circular_queue::peek; + using circular_queue::pop; + using circular_queue::pop_n; + using circular_queue::for_each; + using circular_queue::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 lock(m_pushMtx); +#endif + return circular_queue::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 lock(m_pushMtx); +#endif + return circular_queue::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 lock(m_pushMtx); +#endif + return circular_queue::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 lock(m_pushMtx); +#endif + return circular_queue::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& fun); + +#ifndef ESP8266 +protected: + std::mutex m_pushMtx; +#endif +}; + +template< typename T, typename ForEachArg > +T& circular_queue_mp::pop_requeue() +{ +#ifdef ESP8266 + esp8266::InterruptLock lock; +#else + std::lock_guard lock(m_pushMtx); +#endif + const auto outPos = circular_queue::m_outPos.load(std::memory_order_acquire); + const auto inPos = circular_queue::m_inPos.load(std::memory_order_relaxed); + std::atomic_thread_fence(std::memory_order_acquire); + if (inPos == outPos) return circular_queue::defaultValue; + T& val = circular_queue::m_buffer[inPos] = std::move(circular_queue::m_buffer[outPos]); + const auto bufSize = circular_queue::m_bufSize; + std::atomic_thread_fence(std::memory_order_release); + circular_queue::m_outPos.store((outPos + 1) % bufSize, std::memory_order_relaxed); + circular_queue::m_inPos.store((inPos + 1) % bufSize, std::memory_order_release); + return val; +} + +template< typename T, typename ForEachArg > +bool circular_queue_mp::for_each_requeue(const Delegate& fun) +{ + auto inPos0 = circular_queue::m_inPos.load(std::memory_order_acquire); + auto outPos = circular_queue::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::m_buffer[outPos]); + if (fun(val)) + { +#ifdef ESP8266 + esp8266::InterruptLock lock; +#else + std::lock_guard lock(m_pushMtx); +#endif + std::atomic_thread_fence(std::memory_order_release); + auto inPos = circular_queue::m_inPos.load(std::memory_order_relaxed); + std::atomic_thread_fence(std::memory_order_acquire); + circular_queue::m_buffer[inPos] = std::move(val); + std::atomic_thread_fence(std::memory_order_release); + circular_queue::m_inPos.store((inPos + 1) % circular_queue::m_bufSize, std::memory_order_release); + } + else + { + std::atomic_thread_fence(std::memory_order_release); + } + outPos = (outPos + 1) % circular_queue::m_bufSize; + circular_queue::m_outPos.store(outPos, std::memory_order_release); + } while (outPos != inPos0); + return true; +} + +#endif // __circular_queue_mp_h diff --git a/EspSoftwareSerial/src/circular_queue/ghostl.h b/EspSoftwareSerial/src/circular_queue/ghostl.h new file mode 100644 index 00000000..11683805 --- /dev/null +++ b/EspSoftwareSerial/src/circular_queue/ghostl.h @@ -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 +#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); } +#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 + struct identity { + typedef T type; + }; + + template + inline T&& forward(typename identity::type& t) noexcept + { + return static_cast::type&&>(t); + } +} + +#endif // __ghostl_h