From add7f0df607453029d95663c8f54e5251f033f52 Mon Sep 17 00:00:00 2001 From: Patrick Date: Fri, 6 Mar 2026 21:58:16 +0000 Subject: [PATCH] inital commit for edwards25519 --- LICENSE | 232 ++++++++ README.MD | 13 + asm/fe_amd64_asm.go | 313 +++++++++++ asm/go.mod | 14 + doc.go | 14 + edwards25519.go | 504 +++++++++++++++++ edwards25519_test.go | 343 ++++++++++++ extra.go | 439 +++++++++++++++ extra_test.go | 346 ++++++++++++ field/fe.go | 474 ++++++++++++++++ field/fe_alias_test.go | 157 ++++++ field/fe_amd64.go | 15 + field/fe_amd64.s | 398 ++++++++++++++ field/fe_amd64_noasm.go | 20 + field/fe_bench_test.go | 82 +++ field/fe_extra.go | 63 +++ field/fe_extra_test.go | 50 ++ field/fe_generic.go | 303 +++++++++++ field/fe_test.go | 607 +++++++++++++++++++++ go.mod | 3 + go.sum | 0 pull.sh | 53 ++ scalar.go | 397 ++++++++++++++ scalar_alias_test.go | 124 +++++ scalar_fiat.go | 1128 +++++++++++++++++++++++++++++++++++++++ scalar_test.go | 303 +++++++++++ scalarmult.go | 237 ++++++++ scalarmult_test.go | 242 +++++++++ tables.go | 162 ++++++ tables_test.go | 145 +++++ 30 files changed, 7181 insertions(+) create mode 100644 LICENSE create mode 100644 README.MD create mode 100644 asm/fe_amd64_asm.go create mode 100644 asm/go.mod create mode 100644 doc.go create mode 100644 edwards25519.go create mode 100644 edwards25519_test.go create mode 100644 extra.go create mode 100644 extra_test.go create mode 100644 field/fe.go create mode 100644 field/fe_alias_test.go create mode 100644 field/fe_amd64.go create mode 100644 field/fe_amd64.s create mode 100644 field/fe_amd64_noasm.go create mode 100644 field/fe_bench_test.go create mode 100644 field/fe_extra.go create mode 100644 field/fe_extra_test.go create mode 100644 field/fe_generic.go create mode 100644 field/fe_test.go create mode 100644 go.mod create mode 100644 go.sum create mode 100644 pull.sh create mode 100644 scalar.go create mode 100644 scalar_alias_test.go create mode 100644 scalar_fiat.go create mode 100644 scalar_test.go create mode 100644 scalarmult.go create mode 100644 scalarmult_test.go create mode 100644 tables.go create mode 100644 tables_test.go diff --git a/LICENSE b/LICENSE new file mode 100644 index 0000000..a29981b --- /dev/null +++ b/LICENSE @@ -0,0 +1,232 @@ +GNU GENERAL PUBLIC LICENSE +Version 3, 29 June 2007 + +Copyright © 2007 Free Software Foundation, Inc. + +Everyone is permitted to copy and distribute verbatim copies of this license document, but changing it is not allowed. + +Preamble + +The GNU General Public License is a free, copyleft license for software and other kinds of works. + +The licenses for most software and other practical works are designed to take away your freedom to share and change the works. By contrast, the GNU General Public License is intended to guarantee your freedom to share and change all versions of a program--to make sure it remains free software for all its users. We, the Free Software Foundation, use the GNU General Public License for most of our software; it applies also to any other work released this way by its authors. You can apply it to your programs, too. + +When we speak of free software, we are referring to freedom, 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 them if you wish), that you receive source code or can get it if you want it, that you can change the software or use pieces of it in new free programs, and that you know you can do these things. + +To protect your rights, we need to prevent others from denying you these rights or asking you to surrender the rights. Therefore, you have certain responsibilities if you distribute copies of the software, or if you modify it: responsibilities to respect the freedom of others. + +For example, if you distribute copies of such a program, whether gratis or for a fee, you must pass on to the recipients the same freedoms that you received. You must make sure that they, too, receive or can get the source code. And you must show them these terms so they know their rights. + +Developers that use the GNU GPL protect your rights with two steps: (1) assert copyright on the software, and (2) offer you this License giving you legal permission to copy, distribute and/or modify it. + +For the developers' and authors' protection, the GPL clearly explains that there is no warranty for this free software. For both users' and authors' sake, the GPL requires that modified versions be marked as changed, so that their problems will not be attributed erroneously to authors of previous versions. + +Some devices are designed to deny users access to install or run modified versions of the software inside them, although the manufacturer can do so. This is fundamentally incompatible with the aim of protecting users' freedom to change the software. The systematic pattern of such abuse occurs in the area of products for individuals to use, which is precisely where it is most unacceptable. Therefore, we have designed this version of the GPL to prohibit the practice for those products. If such problems arise substantially in other domains, we stand ready to extend this provision to those domains in future versions of the GPL, as needed to protect the freedom of users. + +Finally, every program is threatened constantly by software patents. States should not allow patents to restrict development and use of software on general-purpose computers, but in those that do, we wish to avoid the special danger that patents applied to a free program could make it effectively proprietary. To prevent this, the GPL assures that patents cannot be used to render the program non-free. + +The precise terms and conditions for copying, distribution and modification follow. + +TERMS AND CONDITIONS + +0. Definitions. + +“This License” refers to version 3 of the GNU General Public License. + +“Copyright” also means copyright-like laws that apply to other kinds of works, such as semiconductor masks. + +“The Program” refers to any copyrightable work licensed under this License. Each licensee is addressed as “you”. “Licensees” and “recipients” may be individuals or organizations. + +To “modify” a work means to copy from or adapt all or part of the work in a fashion requiring copyright permission, other than the making of an exact copy. The resulting work is called a “modified version” of the earlier work or a work “based on” the earlier work. + +A “covered work” means either the unmodified Program or a work based on the Program. + +To “propagate” a work means to do anything with it that, without permission, would make you directly or secondarily liable for infringement under applicable copyright law, except executing it on a computer or modifying a private copy. Propagation includes copying, distribution (with or without modification), making available to the public, and in some countries other activities as well. + +To “convey” a work means any kind of propagation that enables other parties to make or receive copies. Mere interaction with a user through a computer network, with no transfer of a copy, is not conveying. + +An interactive user interface displays “Appropriate Legal Notices” to the extent that it includes a convenient and prominently visible feature that (1) displays an appropriate copyright notice, and (2) tells the user that there is no warranty for the work (except to the extent that warranties are provided), that licensees may convey the work under this License, and how to view a copy of this License. If the interface presents a list of user commands or options, such as a menu, a prominent item in the list meets this criterion. + +1. Source Code. +The “source code” for a work means the preferred form of the work for making modifications to it. “Object code” means any non-source form of a work. + +A “Standard Interface” means an interface that either is an official standard defined by a recognized standards body, or, in the case of interfaces specified for a particular programming language, one that is widely used among developers working in that language. + +The “System Libraries” of an executable work include anything, other than the work as a whole, that (a) is included in the normal form of packaging a Major Component, but which is not part of that Major Component, and (b) serves only to enable use of the work with that Major Component, or to implement a Standard Interface for which an implementation is available to the public in source code form. A “Major Component”, in this context, means a major essential component (kernel, window system, and so on) of the specific operating system (if any) on which the executable work runs, or a compiler used to produce the work, or an object code interpreter used to run it. + +The “Corresponding Source” for a work in object code form means all the source code needed to generate, install, and (for an executable work) run the object code and to modify the work, including scripts to control those activities. However, it does not include the work's System Libraries, or general-purpose tools or generally available free programs which are used unmodified in performing those activities but which are not part of the work. For example, Corresponding Source includes interface definition files associated with source files for the work, and the source code for shared libraries and dynamically linked subprograms that the work is specifically designed to require, such as by intimate data communication or control flow between those subprograms and other parts of the work. + +The Corresponding Source need not include anything that users can regenerate automatically from other parts of the Corresponding Source. + +The Corresponding Source for a work in source code form is that same work. + +2. Basic Permissions. +All rights granted under this License are granted for the term of copyright on the Program, and are irrevocable provided the stated conditions are met. This License explicitly affirms your unlimited permission to run the unmodified Program. The output from running a covered work is covered by this License only if the output, given its content, constitutes a covered work. This License acknowledges your rights of fair use or other equivalent, as provided by copyright law. + +You may make, run and propagate covered works that you do not convey, without conditions so long as your license otherwise remains in force. You may convey covered works to others for the sole purpose of having them make modifications exclusively for you, or provide you with facilities for running those works, provided that you comply with the terms of this License in conveying all material for which you do not control copyright. Those thus making or running the covered works for you must do so exclusively on your behalf, under your direction and control, on terms that prohibit them from making any copies of your copyrighted material outside their relationship with you. + +Conveying under any other circumstances is permitted solely under the conditions stated below. Sublicensing is not allowed; section 10 makes it unnecessary. + +3. Protecting Users' Legal Rights From Anti-Circumvention Law. +No covered work shall be deemed part of an effective technological measure under any applicable law fulfilling obligations under article 11 of the WIPO copyright treaty adopted on 20 December 1996, or similar laws prohibiting or restricting circumvention of such measures. + +When you convey a covered work, you waive any legal power to forbid circumvention of technological measures to the extent such circumvention is effected by exercising rights under this License with respect to the covered work, and you disclaim any intention to limit operation or modification of the work as a means of enforcing, against the work's users, your or third parties' legal rights to forbid circumvention of technological measures. + +4. Conveying Verbatim Copies. +You may convey verbatim copies of the Program's source code as you receive it, in any medium, provided that you conspicuously and appropriately publish on each copy an appropriate copyright notice; keep intact all notices stating that this License and any non-permissive terms added in accord with section 7 apply to the code; keep intact all notices of the absence of any warranty; and give all recipients a copy of this License along with the Program. + +You may charge any price or no price for each copy that you convey, and you may offer support or warranty protection for a fee. + +5. Conveying Modified Source Versions. +You may convey a work based on the Program, or the modifications to produce it from the Program, in the form of source code under the terms of section 4, provided that you also meet all of these conditions: + + a) The work must carry prominent notices stating that you modified it, and giving a relevant date. + + b) The work must carry prominent notices stating that it is released under this License and any conditions added under section 7. This requirement modifies the requirement in section 4 to “keep intact all notices”. + + c) You must license the entire work, as a whole, under this License to anyone who comes into possession of a copy. This License will therefore apply, along with any applicable section 7 additional terms, to the whole of the work, and all its parts, regardless of how they are packaged. This License gives no permission to license the work in any other way, but it does not invalidate such permission if you have separately received it. + + d) If the work has interactive user interfaces, each must display Appropriate Legal Notices; however, if the Program has interactive interfaces that do not display Appropriate Legal Notices, your work need not make them do so. + +A compilation of a covered work with other separate and independent works, which are not by their nature extensions of the covered work, and which are not combined with it such as to form a larger program, in or on a volume of a storage or distribution medium, is called an “aggregate” if the compilation and its resulting copyright are not used to limit the access or legal rights of the compilation's users beyond what the individual works permit. Inclusion of a covered work in an aggregate does not cause this License to apply to the other parts of the aggregate. + +6. Conveying Non-Source Forms. +You may convey a covered work in object code form under the terms of sections 4 and 5, provided that you also convey the machine-readable Corresponding Source under the terms of this License, in one of these ways: + + a) Convey the object code in, or embodied in, a physical product (including a physical distribution medium), accompanied by the Corresponding Source fixed on a durable physical medium customarily used for software interchange. + + b) Convey the object code in, or embodied in, a physical product (including a physical distribution medium), accompanied by a written offer, valid for at least three years and valid for as long as you offer spare parts or customer support for that product model, to give anyone who possesses the object code either (1) a copy of the Corresponding Source for all the software in the product that is covered by this License, on a durable physical medium customarily used for software interchange, for a price no more than your reasonable cost of physically performing this conveying of source, or (2) access to copy the Corresponding Source from a network server at no charge. + + c) Convey individual copies of the object code with a copy of the written offer to provide the Corresponding Source. This alternative is allowed only occasionally and noncommercially, and only if you received the object code with such an offer, in accord with subsection 6b. + + d) Convey the object code by offering access from a designated place (gratis or for a charge), and offer equivalent access to the Corresponding Source in the same way through the same place at no further charge. You need not require recipients to copy the Corresponding Source along with the object code. If the place to copy the object code is a network server, the Corresponding Source may be on a different server (operated by you or a third party) that supports equivalent copying facilities, provided you maintain clear directions next to the object code saying where to find the Corresponding Source. Regardless of what server hosts the Corresponding Source, you remain obligated to ensure that it is available for as long as needed to satisfy these requirements. + + e) Convey the object code using peer-to-peer transmission, provided you inform other peers where the object code and Corresponding Source of the work are being offered to the general public at no charge under subsection 6d. + +A separable portion of the object code, whose source code is excluded from the Corresponding Source as a System Library, need not be included in conveying the object code work. + +A “User Product” is either (1) a “consumer product”, which means any tangible personal property which is normally used for personal, family, or household purposes, or (2) anything designed or sold for incorporation into a dwelling. In determining whether a product is a consumer product, doubtful cases shall be resolved in favor of coverage. For a particular product received by a particular user, “normally used” refers to a typical or common use of that class of product, regardless of the status of the particular user or of the way in which the particular user actually uses, or expects or is expected to use, the product. A product is a consumer product regardless of whether the product has substantial commercial, industrial or non-consumer uses, unless such uses represent the only significant mode of use of the product. + +“Installation Information” for a User Product means any methods, procedures, authorization keys, or other information required to install and execute modified versions of a covered work in that User Product from a modified version of its Corresponding Source. The information must suffice to ensure that the continued functioning of the modified object code is in no case prevented or interfered with solely because modification has been made. + +If you convey an object code work under this section in, or with, or specifically for use in, a User Product, and the conveying occurs as part of a transaction in which the right of possession and use of the User Product is transferred to the recipient in perpetuity or for a fixed term (regardless of how the transaction is characterized), the Corresponding Source conveyed under this section must be accompanied by the Installation Information. But this requirement does not apply if neither you nor any third party retains the ability to install modified object code on the User Product (for example, the work has been installed in ROM). + +The requirement to provide Installation Information does not include a requirement to continue to provide support service, warranty, or updates for a work that has been modified or installed by the recipient, or for the User Product in which it has been modified or installed. Access to a network may be denied when the modification itself materially and adversely affects the operation of the network or violates the rules and protocols for communication across the network. + +Corresponding Source conveyed, and Installation Information provided, in accord with this section must be in a format that is publicly documented (and with an implementation available to the public in source code form), and must require no special password or key for unpacking, reading or copying. + +7. Additional Terms. +“Additional permissions” are terms that supplement the terms of this License by making exceptions from one or more of its conditions. Additional permissions that are applicable to the entire Program shall be treated as though they were included in this License, to the extent that they are valid under applicable law. If additional permissions apply only to part of the Program, that part may be used separately under those permissions, but the entire Program remains governed by this License without regard to the additional permissions. + +When you convey a copy of a covered work, you may at your option remove any additional permissions from that copy, or from any part of it. (Additional permissions may be written to require their own removal in certain cases when you modify the work.) You may place additional permissions on material, added by you to a covered work, for which you have or can give appropriate copyright permission. + +Notwithstanding any other provision of this License, for material you add to a covered work, you may (if authorized by the copyright holders of that material) supplement the terms of this License with terms: + + a) Disclaiming warranty or limiting liability differently from the terms of sections 15 and 16 of this License; or + + b) Requiring preservation of specified reasonable legal notices or author attributions in that material or in the Appropriate Legal Notices displayed by works containing it; or + + c) Prohibiting misrepresentation of the origin of that material, or requiring that modified versions of such material be marked in reasonable ways as different from the original version; or + + d) Limiting the use for publicity purposes of names of licensors or authors of the material; or + + e) Declining to grant rights under trademark law for use of some trade names, trademarks, or service marks; or + + f) Requiring indemnification of licensors and authors of that material by anyone who conveys the material (or modified versions of it) with contractual assumptions of liability to the recipient, for any liability that these contractual assumptions directly impose on those licensors and authors. + +All other non-permissive additional terms are considered “further restrictions” within the meaning of section 10. If the Program as you received it, or any part of it, contains a notice stating that it is governed by this License along with a term that is a further restriction, you may remove that term. If a license document contains a further restriction but permits relicensing or conveying under this License, you may add to a covered work material governed by the terms of that license document, provided that the further restriction does not survive such relicensing or conveying. + +If you add terms to a covered work in accord with this section, you must place, in the relevant source files, a statement of the additional terms that apply to those files, or a notice indicating where to find the applicable terms. + +Additional terms, permissive or non-permissive, may be stated in the form of a separately written license, or stated as exceptions; the above requirements apply either way. + +8. Termination. +You may not propagate or modify a covered work except as expressly provided under this License. Any attempt otherwise to propagate or modify it is void, and will automatically terminate your rights under this License (including any patent licenses granted under the third paragraph of section 11). + +However, if you cease all violation of this License, then your license from a particular copyright holder is reinstated (a) provisionally, unless and until the copyright holder explicitly and finally terminates your license, and (b) permanently, if the copyright holder fails to notify you of the violation by some reasonable means prior to 60 days after the cessation. + +Moreover, your license from a particular copyright holder is reinstated permanently if the copyright holder notifies you of the violation by some reasonable means, this is the first time you have received notice of violation of this License (for any work) from that copyright holder, and you cure the violation prior to 30 days after your receipt of the notice. + +Termination of your rights under this section does not terminate the licenses of parties who have received copies or rights from you under this License. If your rights have been terminated and not permanently reinstated, you do not qualify to receive new licenses for the same material under section 10. + +9. Acceptance Not Required for Having Copies. +You are not required to accept this License in order to receive or run a copy of the Program. Ancillary propagation of a covered work occurring solely as a consequence of using peer-to-peer transmission to receive a copy likewise does not require acceptance. However, nothing other than this License grants you permission to propagate or modify any covered work. These actions infringe copyright if you do not accept this License. Therefore, by modifying or propagating a covered work, you indicate your acceptance of this License to do so. + +10. Automatic Licensing of Downstream Recipients. +Each time you convey a covered work, the recipient automatically receives a license from the original licensors, to run, modify and propagate that work, subject to this License. You are not responsible for enforcing compliance by third parties with this License. + +An “entity transaction” is a transaction transferring control of an organization, or substantially all assets of one, or subdividing an organization, or merging organizations. If propagation of a covered work results from an entity transaction, each party to that transaction who receives a copy of the work also receives whatever licenses to the work the party's predecessor in interest had or could give under the previous paragraph, plus a right to possession of the Corresponding Source of the work from the predecessor in interest, if the predecessor has it or can get it with reasonable efforts. + +You may not impose any further restrictions on the exercise of the rights granted or affirmed under this License. For example, you may not impose a license fee, royalty, or other charge for exercise of rights granted under this License, and you may not initiate litigation (including a cross-claim or counterclaim in a lawsuit) alleging that any patent claim is infringed by making, using, selling, offering for sale, or importing the Program or any portion of it. + +11. Patents. +A “contributor” is a copyright holder who authorizes use under this License of the Program or a work on which the Program is based. The work thus licensed is called the contributor's “contributor version”. + +A contributor's “essential patent claims” are all patent claims owned or controlled by the contributor, whether already acquired or hereafter acquired, that would be infringed by some manner, permitted by this License, of making, using, or selling its contributor version, but do not include claims that would be infringed only as a consequence of further modification of the contributor version. For purposes of this definition, “control” includes the right to grant patent sublicenses in a manner consistent with the requirements of this License. + +Each contributor grants you a non-exclusive, worldwide, royalty-free patent license under the contributor's essential patent claims, to make, use, sell, offer for sale, import and otherwise run, modify and propagate the contents of its contributor version. + +In the following three paragraphs, a “patent license” is any express agreement or commitment, however denominated, not to enforce a patent (such as an express permission to practice a patent or covenant not to sue for patent infringement). To “grant” such a patent license to a party means to make such an agreement or commitment not to enforce a patent against the party. + +If you convey a covered work, knowingly relying on a patent license, and the Corresponding Source of the work is not available for anyone to copy, free of charge and under the terms of this License, through a publicly available network server or other readily accessible means, then you must either (1) cause the Corresponding Source to be so available, or (2) arrange to deprive yourself of the benefit of the patent license for this particular work, or (3) arrange, in a manner consistent with the requirements of this License, to extend the patent license to downstream recipients. “Knowingly relying” means you have actual knowledge that, but for the patent license, your conveying the covered work in a country, or your recipient's use of the covered work in a country, would infringe one or more identifiable patents in that country that you have reason to believe are valid. + +If, pursuant to or in connection with a single transaction or arrangement, you convey, or propagate by procuring conveyance of, a covered work, and grant a patent license to some of the parties receiving the covered work authorizing them to use, propagate, modify or convey a specific copy of the covered work, then the patent license you grant is automatically extended to all recipients of the covered work and works based on it. + +A patent license is “discriminatory” if it does not include within the scope of its coverage, prohibits the exercise of, or is conditioned on the non-exercise of one or more of the rights that are specifically granted under this License. You may not convey a covered work if you are a party to an arrangement with a third party that is in the business of distributing software, under which you make payment to the third party based on the extent of your activity of conveying the work, and under which the third party grants, to any of the parties who would receive the covered work from you, a discriminatory patent license (a) in connection with copies of the covered work conveyed by you (or copies made from those copies), or (b) primarily for and in connection with specific products or compilations that contain the covered work, unless you entered into that arrangement, or that patent license was granted, prior to 28 March 2007. + +Nothing in this License shall be construed as excluding or limiting any implied license or other defenses to infringement that may otherwise be available to you under applicable patent law. + +12. No Surrender of Others' Freedom. +If 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 convey a covered work so as to satisfy simultaneously your obligations under this License and any other pertinent obligations, then as a consequence you may not convey it at all. For example, if you agree to terms that obligate you to collect a royalty for further conveying from those to whom you convey the Program, the only way you could satisfy both those terms and this License would be to refrain entirely from conveying the Program. + +13. Use with the GNU Affero General Public License. +Notwithstanding any other provision of this License, you have permission to link or combine any covered work with a work licensed under version 3 of the GNU Affero General Public License into a single combined work, and to convey the resulting work. The terms of this License will continue to apply to the part which is the covered work, but the special requirements of the GNU Affero General Public License, section 13, concerning interaction through a network will apply to the combination as such. + +14. Revised Versions of this License. +The Free Software Foundation may publish revised and/or new versions of the GNU 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 Program specifies that a certain numbered version of the GNU General Public License “or any later version” applies to it, you have the option of following the terms and conditions either of that numbered version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of the GNU General Public License, you may choose any version ever published by the Free Software Foundation. + +If the Program specifies that a proxy can decide which future versions of the GNU General Public License can be used, that proxy's public statement of acceptance of a version permanently authorizes you to choose that version for the Program. + +Later license versions may give you additional or different permissions. However, no additional obligations are imposed on any author or copyright holder as a result of your choosing to follow a later version. + +15. Disclaimer of Warranty. +THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM “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 PROGRAM IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION. + +16. Limitation of Liability. +IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS THE PROGRAM 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 PROGRAM (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 PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. + +17. Interpretation of Sections 15 and 16. +If the disclaimer of warranty and limitation of liability provided above cannot be given local legal effect according to their terms, reviewing courts shall apply local law that most closely approximates an absolute waiver of all civil liability in connection with the Program, unless a warranty or assumption of liability accompanies a copy of the Program in return for a fee. + +END OF TERMS AND CONDITIONS + +How to Apply These Terms to Your New Programs + +If you develop a new program, and you want it to be of the greatest possible use to the public, the best way to achieve this is to make it free software which everyone can redistribute and change under these terms. + +To do so, attach the following notices to the program. It is safest to attach them to the start of each source file to most effectively state the exclusion of warranty; and each file should have at least the “copyright” line and a pointer to where the full notice is found. + + edwards25519 + Copyright (C) 2026 code + + This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version. + + This program 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 General Public License for more details. + + You should have received a copy of the GNU General Public License along with this program. If not, see . + +Also add information on how to contact you by electronic and paper mail. + +If the program does terminal interaction, make it output a short notice like this when it starts in an interactive mode: + + edwards25519 Copyright (C) 2026 code + This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'. + This is free software, and you are welcome to redistribute it under certain conditions; type `show c' for details. + +The hypothetical commands `show w' and `show c' should show the appropriate parts of the General Public License. Of course, your program's commands might be different; for a GUI interface, you would use an “about box”. + +You should also get your employer (if you work as a programmer) or school, if any, to sign a “copyright disclaimer” for the program, if necessary. For more information on this, and how to apply and follow the GNU GPL, see . + +The GNU General Public License does not permit incorporating your program into proprietary programs. If your program is a subroutine library, you may consider it more useful to permit linking proprietary applications with the library. If this is what you want to do, use the GNU Lesser General Public License instead of this License. But first, please read . diff --git a/README.MD b/README.MD new file mode 100644 index 0000000..93b6a94 --- /dev/null +++ b/README.MD @@ -0,0 +1,13 @@ +import "sources.truenas.cloud/code/edwards25519" + +# edwards25519 + +### This library implements the edwards25519 elliptic curve, exposing the necessary APIs to build a wide array of higher level primitives. + +The package tracks the upstream standard library package crypto/internal/fips140/edwards25519 and extends it with additional functionality. + +The code is originally derived from Adam Langley's internal implementation in the Go standard library, and includes George Tankersley's performance improvements. It was then further developed by Henry de Velence for use in ristretto255 and was finally merged back into the Go standard library as of Go 1.17. + +Most users don't need this package, and should instead use crypto/ed25519 for signatures, crypto/ecdh for Diffie-Hellman, or github.com/gtank/ristretto255 for prime order group logic. However, for anyone currently using a fork of the internal edwards25519 package or of github.com/ag1/edwards25519, this package should be a safer, faster, and more powerful alternative. + +Since this package is meant to curb proliferation of edwards25519 implementations in the Go ecosystem, it welcomes requests for new APIs or reviewable performance improvements. \ No newline at end of file diff --git a/asm/fe_amd64_asm.go b/asm/fe_amd64_asm.go new file mode 100644 index 0000000..5f319f8 --- /dev/null +++ b/asm/fe_amd64_asm.go @@ -0,0 +1,313 @@ +// Copyright (c) 2021 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package main + +import ( + "fmt" + + . "sources.truenas.cloud/code/avo/build" + . "sources.truenas.cloud/code/avo/gotypes" + . "sources.truenas.cloud/code/avo/operand" + . "sources.truenas.cloud/code/avo/reg" + + _ "sources.truenas.cloud/code/edwards25519/field" +) + +//go:generate go run . -out ../fe_amd64.s -stubs ../fe_amd64.go -pkg field + +func main() { + Package("sources.truenas.cloud/code/edwards25519/field") + ConstraintExpr("!purego") + feMul() + feSquare() + Generate() +} + +type namedComponent struct { + Component + name string +} + +func (c namedComponent) String() string { return c.name } + +type uint128 struct { + name string + hi, lo GPVirtual +} + +func (c uint128) String() string { return c.name } + +func feSquare() { + TEXT("feSquare", NOSPLIT, "func(out, a *Element)") + Doc("feSquare sets out = a * a. It works like feSquareGeneric.") + Pragma("noescape") + + a := Dereference(Param("a")) + l0 := namedComponent{a.Field("l0"), "l0"} + l1 := namedComponent{a.Field("l1"), "l1"} + l2 := namedComponent{a.Field("l2"), "l2"} + l3 := namedComponent{a.Field("l3"), "l3"} + l4 := namedComponent{a.Field("l4"), "l4"} + + // r0 = l0×l0 + 19×2×(l1×l4 + l2×l3) + r0 := uint128{"r0", GP64(), GP64()} + mul64(r0, 1, l0, l0) + addMul64(r0, 38, l1, l4) + addMul64(r0, 38, l2, l3) + + // r1 = 2×l0×l1 + 19×2×l2×l4 + 19×l3×l3 + r1 := uint128{"r1", GP64(), GP64()} + mul64(r1, 2, l0, l1) + addMul64(r1, 38, l2, l4) + addMul64(r1, 19, l3, l3) + + // r2 = = 2×l0×l2 + l1×l1 + 19×2×l3×l4 + r2 := uint128{"r2", GP64(), GP64()} + mul64(r2, 2, l0, l2) + addMul64(r2, 1, l1, l1) + addMul64(r2, 38, l3, l4) + + // r3 = = 2×l0×l3 + 2×l1×l2 + 19×l4×l4 + r3 := uint128{"r3", GP64(), GP64()} + mul64(r3, 2, l0, l3) + addMul64(r3, 2, l1, l2) + addMul64(r3, 19, l4, l4) + + // r4 = = 2×l0×l4 + 2×l1×l3 + l2×l2 + r4 := uint128{"r4", GP64(), GP64()} + mul64(r4, 2, l0, l4) + addMul64(r4, 2, l1, l3) + addMul64(r4, 1, l2, l2) + + Comment("First reduction chain") + maskLow51Bits := GP64() + MOVQ(Imm((1<<51)-1), maskLow51Bits) + c0, r0lo := shiftRightBy51(&r0) + c1, r1lo := shiftRightBy51(&r1) + c2, r2lo := shiftRightBy51(&r2) + c3, r3lo := shiftRightBy51(&r3) + c4, r4lo := shiftRightBy51(&r4) + maskAndAdd(r0lo, maskLow51Bits, c4, 19) + maskAndAdd(r1lo, maskLow51Bits, c0, 1) + maskAndAdd(r2lo, maskLow51Bits, c1, 1) + maskAndAdd(r3lo, maskLow51Bits, c2, 1) + maskAndAdd(r4lo, maskLow51Bits, c3, 1) + + Comment("Second reduction chain (carryPropagate)") + // c0 = r0 >> 51 + MOVQ(r0lo, c0) + SHRQ(Imm(51), c0) + // c1 = r1 >> 51 + MOVQ(r1lo, c1) + SHRQ(Imm(51), c1) + // c2 = r2 >> 51 + MOVQ(r2lo, c2) + SHRQ(Imm(51), c2) + // c3 = r3 >> 51 + MOVQ(r3lo, c3) + SHRQ(Imm(51), c3) + // c4 = r4 >> 51 + MOVQ(r4lo, c4) + SHRQ(Imm(51), c4) + maskAndAdd(r0lo, maskLow51Bits, c4, 19) + maskAndAdd(r1lo, maskLow51Bits, c0, 1) + maskAndAdd(r2lo, maskLow51Bits, c1, 1) + maskAndAdd(r3lo, maskLow51Bits, c2, 1) + maskAndAdd(r4lo, maskLow51Bits, c3, 1) + + Comment("Store output") + out := Dereference(Param("out")) + Store(r0lo, out.Field("l0")) + Store(r1lo, out.Field("l1")) + Store(r2lo, out.Field("l2")) + Store(r3lo, out.Field("l3")) + Store(r4lo, out.Field("l4")) + + RET() +} + +func feMul() { + TEXT("feMul", NOSPLIT, "func(out, a, b *Element)") + Doc("feMul sets out = a * b. It works like feMulGeneric.") + Pragma("noescape") + + a := Dereference(Param("a")) + a0 := namedComponent{a.Field("l0"), "a0"} + a1 := namedComponent{a.Field("l1"), "a1"} + a2 := namedComponent{a.Field("l2"), "a2"} + a3 := namedComponent{a.Field("l3"), "a3"} + a4 := namedComponent{a.Field("l4"), "a4"} + + b := Dereference(Param("b")) + b0 := namedComponent{b.Field("l0"), "b0"} + b1 := namedComponent{b.Field("l1"), "b1"} + b2 := namedComponent{b.Field("l2"), "b2"} + b3 := namedComponent{b.Field("l3"), "b3"} + b4 := namedComponent{b.Field("l4"), "b4"} + + // r0 = a0×b0 + 19×(a1×b4 + a2×b3 + a3×b2 + a4×b1) + r0 := uint128{"r0", GP64(), GP64()} + mul64(r0, 1, a0, b0) + addMul64(r0, 19, a1, b4) + addMul64(r0, 19, a2, b3) + addMul64(r0, 19, a3, b2) + addMul64(r0, 19, a4, b1) + + // r1 = a0×b1 + a1×b0 + 19×(a2×b4 + a3×b3 + a4×b2) + r1 := uint128{"r1", GP64(), GP64()} + mul64(r1, 1, a0, b1) + addMul64(r1, 1, a1, b0) + addMul64(r1, 19, a2, b4) + addMul64(r1, 19, a3, b3) + addMul64(r1, 19, a4, b2) + + // r2 = a0×b2 + a1×b1 + a2×b0 + 19×(a3×b4 + a4×b3) + r2 := uint128{"r2", GP64(), GP64()} + mul64(r2, 1, a0, b2) + addMul64(r2, 1, a1, b1) + addMul64(r2, 1, a2, b0) + addMul64(r2, 19, a3, b4) + addMul64(r2, 19, a4, b3) + + // r3 = a0×b3 + a1×b2 + a2×b1 + a3×b0 + 19×a4×b4 + r3 := uint128{"r3", GP64(), GP64()} + mul64(r3, 1, a0, b3) + addMul64(r3, 1, a1, b2) + addMul64(r3, 1, a2, b1) + addMul64(r3, 1, a3, b0) + addMul64(r3, 19, a4, b4) + + // r4 = a0×b4 + a1×b3 + a2×b2 + a3×b1 + a4×b0 + r4 := uint128{"r4", GP64(), GP64()} + mul64(r4, 1, a0, b4) + addMul64(r4, 1, a1, b3) + addMul64(r4, 1, a2, b2) + addMul64(r4, 1, a3, b1) + addMul64(r4, 1, a4, b0) + + Comment("First reduction chain") + maskLow51Bits := GP64() + MOVQ(Imm((1<<51)-1), maskLow51Bits) + c0, r0lo := shiftRightBy51(&r0) + c1, r1lo := shiftRightBy51(&r1) + c2, r2lo := shiftRightBy51(&r2) + c3, r3lo := shiftRightBy51(&r3) + c4, r4lo := shiftRightBy51(&r4) + maskAndAdd(r0lo, maskLow51Bits, c4, 19) + maskAndAdd(r1lo, maskLow51Bits, c0, 1) + maskAndAdd(r2lo, maskLow51Bits, c1, 1) + maskAndAdd(r3lo, maskLow51Bits, c2, 1) + maskAndAdd(r4lo, maskLow51Bits, c3, 1) + + Comment("Second reduction chain (carryPropagate)") + // c0 = r0 >> 51 + MOVQ(r0lo, c0) + SHRQ(Imm(51), c0) + // c1 = r1 >> 51 + MOVQ(r1lo, c1) + SHRQ(Imm(51), c1) + // c2 = r2 >> 51 + MOVQ(r2lo, c2) + SHRQ(Imm(51), c2) + // c3 = r3 >> 51 + MOVQ(r3lo, c3) + SHRQ(Imm(51), c3) + // c4 = r4 >> 51 + MOVQ(r4lo, c4) + SHRQ(Imm(51), c4) + maskAndAdd(r0lo, maskLow51Bits, c4, 19) + maskAndAdd(r1lo, maskLow51Bits, c0, 1) + maskAndAdd(r2lo, maskLow51Bits, c1, 1) + maskAndAdd(r3lo, maskLow51Bits, c2, 1) + maskAndAdd(r4lo, maskLow51Bits, c3, 1) + + Comment("Store output") + out := Dereference(Param("out")) + Store(r0lo, out.Field("l0")) + Store(r1lo, out.Field("l1")) + Store(r2lo, out.Field("l2")) + Store(r3lo, out.Field("l3")) + Store(r4lo, out.Field("l4")) + + RET() +} + +// mul64 sets r to i * aX * bX. +func mul64(r uint128, i int, aX, bX namedComponent) { + switch i { + case 1: + Comment(fmt.Sprintf("%s = %s×%s", r, aX, bX)) + Load(aX, RAX) + case 2: + Comment(fmt.Sprintf("%s = 2×%s×%s", r, aX, bX)) + Load(aX, RAX) + SHLQ(Imm(1), RAX) + default: + panic("unsupported i value") + } + MULQ(mustAddr(bX)) // RDX, RAX = RAX * bX + MOVQ(RAX, r.lo) + MOVQ(RDX, r.hi) +} + +// addMul64 sets r to r + i * aX * bX. +func addMul64(r uint128, i uint64, aX, bX namedComponent) { + switch i { + case 1: + Comment(fmt.Sprintf("%s += %s×%s", r, aX, bX)) + Load(aX, RAX) + case 2: + Comment(fmt.Sprintf("%s += %d×%s×%s", r, i, aX, bX)) + Load(aX, RAX) + SHLQ(U8(1), RAX) + case 19: + Comment(fmt.Sprintf("%s += %d×%s×%s", r, i, aX, bX)) + // 19 * v ==> v + (v+v*8)*2 + tmp := Load(aX, GP64()) + LEAQ(Mem{Base: tmp, Index: tmp, Scale: 8}, RAX) + LEAQ(Mem{Base: tmp, Index: RAX, Scale: 2}, RAX) + case 38: + Comment(fmt.Sprintf("%s += %d×%s×%s", r, i, aX, bX)) + // 38 * v ==> (v + (v+v*8)*2) * 2 + tmp := Load(aX, GP64()) + LEAQ(Mem{Base: tmp, Index: tmp, Scale: 8}, RAX) + LEAQ(Mem{Base: tmp, Index: RAX, Scale: 2}, RAX) + SHLQ(U8(1), RAX) + default: + Comment(fmt.Sprintf("%s += %d×%s×%s", r, i, aX, bX)) + IMUL3Q(Imm(i), Load(aX, GP64()), RAX) + } + MULQ(mustAddr(bX)) // RDX, RAX = RAX * bX + ADDQ(RAX, r.lo) + ADCQ(RDX, r.hi) +} + +// shiftRightBy51 returns r >> 51 and r.lo. +// +// After this function is called, the uint128 may not be used anymore. +func shiftRightBy51(r *uint128) (out, lo GPVirtual) { + out = r.hi + lo = r.lo + SHLQ(Imm(64-51), r.lo, r.hi) + r.lo, r.hi = nil, nil // make sure the uint128 is unusable + return +} + +// maskAndAdd sets r = r&mask + c*i. +func maskAndAdd(r, mask, c GPVirtual, i uint64) { + ANDQ(mask, r) + if i != 1 { + IMUL3Q(Imm(i), c, c) + } + ADDQ(c, r) +} + +func mustAddr(c Component) Op { + b, err := c.Resolve() + if err != nil { + panic(err) + } + return b.Addr +} diff --git a/asm/go.mod b/asm/go.mod new file mode 100644 index 0000000..624fdda --- /dev/null +++ b/asm/go.mod @@ -0,0 +1,14 @@ +module crypto/internal/fips140/edwards25519/field_asm + +go 1.26 + +require sources.truenas.cloud/code/avo v0.6.0 + +require golang.org/x/mod v0.14.0 + +require ( + golang.org/x/tools v0.16.1 + sources.truenas.cloud/code/edwards25519 v0.0.0 +) + +replace sources.truenas.cloud/edwards25519 v0.0.0 => ../.. diff --git a/doc.go b/doc.go new file mode 100644 index 0000000..f700164 --- /dev/null +++ b/doc.go @@ -0,0 +1,14 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519 +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package edwards25519 diff --git a/edwards25519.go b/edwards25519.go new file mode 100644 index 0000000..37e4f24 --- /dev/null +++ b/edwards25519.go @@ -0,0 +1,504 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519 +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package edwards25519 + +import ( + "errors" + + "sources.truenas.cloud/code/edwards25519/field" +) + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Point types. + +type projP1xP1 struct { + X, Y, Z, T field.Element +} + +type projP2 struct { + X, Y, Z field.Element +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Point represents a point on the edwards25519 curve. +// +// This type works similarly to math/big.Int, and all arguments and receivers +// are allowed to alias. +// +// The zero value is NOT valid, and it may be used only as a receiver. +type Point struct { + // Make the type not comparable (i.e. used with == or as a map key), as + // equivalent points can be represented by different Go values. + _ incomparable + + // The point is internally represented in extended coordinates (X, Y, Z, T) + // where x = X/Z, y = Y/Z, and xy = T/Z per https://eprint.iacr.org/2008/522. + x, y, z, t field.Element +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +type incomparable [0]func() + +func checkInitialized(points ...*Point) { + for _, p := range points { + if p.x == (field.Element{}) && p.y == (field.Element{}) { + panic("edwards25519: use of uninitialized Point") + } + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +type projCached struct { + YplusX, YminusX, Z, T2d field.Element +} + +type affineCached struct { + YplusX, YminusX, T2d field.Element +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Constructors. + +func (v *projP2) Zero() *projP2 { + v.X.Zero() + v.Y.One() + v.Z.One() + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// identity is the point at infinity. +var identity, _ = new(Point).SetBytes([]byte{ + 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}) + +// NewIdentityPoint returns a new Point set to the identity. +func NewIdentityPoint() *Point { + return new(Point).Set(identity) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// generator is the canonical curve basepoint. See TestGenerator for the +// correspondence of this encoding with the values in RFC 8032. +var generator, _ = new(Point).SetBytes([]byte{ + 0x58, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, + 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, + 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, + 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66}) + +// NewGeneratorPoint returns a new Point set to the canonical generator. +func NewGeneratorPoint() *Point { + return new(Point).Set(generator) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v *projCached) Zero() *projCached { + v.YplusX.One() + v.YminusX.One() + v.Z.One() + v.T2d.Zero() + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v *affineCached) Zero() *affineCached { + v.YplusX.One() + v.YminusX.One() + v.T2d.Zero() + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Assignments. + +// Set sets v = u, and returns v. +func (v *Point) Set(u *Point) *Point { + *v = *u + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Encoding. + +// Bytes returns the canonical 32-byte encoding of v, according to RFC 8032, +// Section 5.1.2. +func (v *Point) Bytes() []byte { + // This function is outlined to make the allocations inline in the caller + // rather than happen on the heap. + var buf [32]byte + return v.bytes(&buf) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v *Point) bytes(buf *[32]byte) []byte { + checkInitialized(v) + + var zInv, x, y field.Element + zInv.Invert(&v.z) // zInv = 1 / Z + x.Multiply(&v.x, &zInv) // x = X / Z + y.Multiply(&v.y, &zInv) // y = Y / Z + + out := copyFieldElement(buf, &y) + out[31] |= byte(x.IsNegative() << 7) + return out +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +var feOne = new(field.Element).One() + +// SetBytes sets v = x, where x is a 32-byte encoding of v. If x does not +// represent a valid point on the curve, SetBytes returns nil and an error and +// the receiver is unchanged. Otherwise, SetBytes returns v. +// +// Note that SetBytes accepts all non-canonical encodings of valid points. +// That is, it follows decoding rules that match most implementations in +// the ecosystem rather than RFC 8032. +func (v *Point) SetBytes(x []byte) (*Point, error) { + // Specifically, the non-canonical encodings that are accepted are + // 1) the ones where the field element is not reduced (see the + // (*field.Element).SetBytes docs) and + // 2) the ones where the x-coordinate is zero and the sign bit is set. + // + // Read more at https://hdevalence.ca/blog/2020-10-04-its-25519am, + // specifically the "Canonical A, R" section. + + y, err := new(field.Element).SetBytes(x) + if err != nil { + return nil, errors.New("edwards25519: invalid point encoding length") + } + + // -x² + y² = 1 + dx²y² + // x² + dx²y² = x²(dy² + 1) = y² - 1 + // x² = (y² - 1) / (dy² + 1) + + // u = y² - 1 + y2 := new(field.Element).Square(y) + u := new(field.Element).Subtract(y2, feOne) + + // v = dy² + 1 + vv := new(field.Element).Multiply(y2, d) + vv = vv.Add(vv, feOne) + + // x = +√(u/v) + xx, wasSquare := new(field.Element).SqrtRatio(u, vv) + if wasSquare == 0 { + return nil, errors.New("edwards25519: invalid point encoding") + } + + // Select the negative square root if the sign bit is set. + xxNeg := new(field.Element).Negate(xx) + xx = xx.Select(xxNeg, xx, int(x[31]>>7)) + + v.x.Set(xx) + v.y.Set(y) + v.z.One() + v.t.Multiply(xx, y) // xy = T / Z + + return v, nil +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func copyFieldElement(buf *[32]byte, v *field.Element) []byte { + copy(buf[:], v.Bytes()) + return buf[:] +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Conversions. + +func (v *projP2) FromP1xP1(p *projP1xP1) *projP2 { + v.X.Multiply(&p.X, &p.T) + v.Y.Multiply(&p.Y, &p.Z) + v.Z.Multiply(&p.Z, &p.T) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v *projP2) FromP3(p *Point) *projP2 { + v.X.Set(&p.x) + v.Y.Set(&p.y) + v.Z.Set(&p.z) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v *Point) fromP1xP1(p *projP1xP1) *Point { + v.x.Multiply(&p.X, &p.T) + v.y.Multiply(&p.Y, &p.Z) + v.z.Multiply(&p.Z, &p.T) + v.t.Multiply(&p.X, &p.Y) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v *Point) fromP2(p *projP2) *Point { + v.x.Multiply(&p.X, &p.Z) + v.y.Multiply(&p.Y, &p.Z) + v.z.Square(&p.Z) + v.t.Multiply(&p.X, &p.Y) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// d is a constant in the curve equation. +var d, _ = new(field.Element).SetBytes([]byte{ + 0xa3, 0x78, 0x59, 0x13, 0xca, 0x4d, 0xeb, 0x75, + 0xab, 0xd8, 0x41, 0x41, 0x4d, 0x0a, 0x70, 0x00, + 0x98, 0xe8, 0x79, 0x77, 0x79, 0x40, 0xc7, 0x8c, + 0x73, 0xfe, 0x6f, 0x2b, 0xee, 0x6c, 0x03, 0x52}) +var d2 = new(field.Element).Add(d, d) + +func (v *projCached) FromP3(p *Point) *projCached { + v.YplusX.Add(&p.y, &p.x) + v.YminusX.Subtract(&p.y, &p.x) + v.Z.Set(&p.z) + v.T2d.Multiply(&p.t, d2) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v *affineCached) FromP3(p *Point) *affineCached { + v.YplusX.Add(&p.y, &p.x) + v.YminusX.Subtract(&p.y, &p.x) + v.T2d.Multiply(&p.t, d2) + + var invZ field.Element + invZ.Invert(&p.z) + v.YplusX.Multiply(&v.YplusX, &invZ) + v.YminusX.Multiply(&v.YminusX, &invZ) + v.T2d.Multiply(&v.T2d, &invZ) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// (Re)addition and subtraction. + +// Add sets v = p + q, and returns v. +func (v *Point) Add(p, q *Point) *Point { + checkInitialized(p, q) + qCached := new(projCached).FromP3(q) + result := new(projP1xP1).Add(p, qCached) + return v.fromP1xP1(result) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Subtract sets v = p - q, and returns v. +func (v *Point) Subtract(p, q *Point) *Point { + checkInitialized(p, q) + qCached := new(projCached).FromP3(q) + result := new(projP1xP1).Sub(p, qCached) + return v.fromP1xP1(result) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v *projP1xP1) Add(p *Point, q *projCached) *projP1xP1 { + var YplusX, YminusX, PP, MM, TT2d, ZZ2 field.Element + + YplusX.Add(&p.y, &p.x) + YminusX.Subtract(&p.y, &p.x) + + PP.Multiply(&YplusX, &q.YplusX) + MM.Multiply(&YminusX, &q.YminusX) + TT2d.Multiply(&p.t, &q.T2d) + ZZ2.Multiply(&p.z, &q.Z) + + ZZ2.Add(&ZZ2, &ZZ2) + + v.X.Subtract(&PP, &MM) + v.Y.Add(&PP, &MM) + v.Z.Add(&ZZ2, &TT2d) + v.T.Subtract(&ZZ2, &TT2d) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v *projP1xP1) Sub(p *Point, q *projCached) *projP1xP1 { + var YplusX, YminusX, PP, MM, TT2d, ZZ2 field.Element + + YplusX.Add(&p.y, &p.x) + YminusX.Subtract(&p.y, &p.x) + + PP.Multiply(&YplusX, &q.YminusX) // flipped sign + MM.Multiply(&YminusX, &q.YplusX) // flipped sign + TT2d.Multiply(&p.t, &q.T2d) + ZZ2.Multiply(&p.z, &q.Z) + + ZZ2.Add(&ZZ2, &ZZ2) + + v.X.Subtract(&PP, &MM) + v.Y.Add(&PP, &MM) + v.Z.Subtract(&ZZ2, &TT2d) // flipped sign + v.T.Add(&ZZ2, &TT2d) // flipped sign + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v *projP1xP1) AddAffine(p *Point, q *affineCached) *projP1xP1 { + var YplusX, YminusX, PP, MM, TT2d, Z2 field.Element + + YplusX.Add(&p.y, &p.x) + YminusX.Subtract(&p.y, &p.x) + + PP.Multiply(&YplusX, &q.YplusX) + MM.Multiply(&YminusX, &q.YminusX) + TT2d.Multiply(&p.t, &q.T2d) + + Z2.Add(&p.z, &p.z) + + v.X.Subtract(&PP, &MM) + v.Y.Add(&PP, &MM) + v.Z.Add(&Z2, &TT2d) + v.T.Subtract(&Z2, &TT2d) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v *projP1xP1) SubAffine(p *Point, q *affineCached) *projP1xP1 { + var YplusX, YminusX, PP, MM, TT2d, Z2 field.Element + + YplusX.Add(&p.y, &p.x) + YminusX.Subtract(&p.y, &p.x) + + PP.Multiply(&YplusX, &q.YminusX) // flipped sign + MM.Multiply(&YminusX, &q.YplusX) // flipped sign + TT2d.Multiply(&p.t, &q.T2d) + + Z2.Add(&p.z, &p.z) + + v.X.Subtract(&PP, &MM) + v.Y.Add(&PP, &MM) + v.Z.Subtract(&Z2, &TT2d) // flipped sign + v.T.Add(&Z2, &TT2d) // flipped sign + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Doubling. + +func (v *projP1xP1) Double(p *projP2) *projP1xP1 { + var XX, YY, ZZ2, XplusYsq field.Element + + XX.Square(&p.X) + YY.Square(&p.Y) + ZZ2.Square(&p.Z) + ZZ2.Add(&ZZ2, &ZZ2) + XplusYsq.Add(&p.X, &p.Y) + XplusYsq.Square(&XplusYsq) + + v.Y.Add(&YY, &XX) + v.Z.Subtract(&YY, &XX) + + v.X.Subtract(&XplusYsq, &v.Y) + v.T.Subtract(&ZZ2, &v.Z) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Negation. + +// Negate sets v = -p, and returns v. +func (v *Point) Negate(p *Point) *Point { + checkInitialized(p) + v.x.Negate(&p.x) + v.y.Set(&p.y) + v.z.Set(&p.z) + v.t.Negate(&p.t) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Equal returns 1 if v is equivalent to u, and 0 otherwise. +func (v *Point) Equal(u *Point) int { + checkInitialized(v, u) + + var t1, t2, t3, t4 field.Element + t1.Multiply(&v.x, &u.z) + t2.Multiply(&u.x, &v.z) + t3.Multiply(&v.y, &u.z) + t4.Multiply(&u.y, &v.z) + + return t1.Equal(&t2) & t3.Equal(&t4) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Constant-time operations + +// Select sets v to a if cond == 1 and to b if cond == 0. +func (v *projCached) Select(a, b *projCached, cond int) *projCached { + v.YplusX.Select(&a.YplusX, &b.YplusX, cond) + v.YminusX.Select(&a.YminusX, &b.YminusX, cond) + v.Z.Select(&a.Z, &b.Z, cond) + v.T2d.Select(&a.T2d, &b.T2d, cond) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Select sets v to a if cond == 1 and to b if cond == 0. +func (v *affineCached) Select(a, b *affineCached, cond int) *affineCached { + v.YplusX.Select(&a.YplusX, &b.YplusX, cond) + v.YminusX.Select(&a.YminusX, &b.YminusX, cond) + v.T2d.Select(&a.T2d, &b.T2d, cond) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// CondNeg negates v if cond == 1 and leaves it unchanged if cond == 0. +func (v *projCached) CondNeg(cond int) *projCached { + v.YplusX.Swap(&v.YminusX, cond) + v.T2d.Select(new(field.Element).Negate(&v.T2d), &v.T2d, cond) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// CondNeg negates v if cond == 1 and leaves it unchanged if cond == 0. +func (v *affineCached) CondNeg(cond int) *affineCached { + v.YplusX.Swap(&v.YminusX, cond) + v.T2d.Select(new(field.Element).Negate(&v.T2d), &v.T2d, cond) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/edwards25519_test.go b/edwards25519_test.go new file mode 100644 index 0000000..5bcaf7e --- /dev/null +++ b/edwards25519_test.go @@ -0,0 +1,343 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519 +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package edwards25519 + +import ( + "encoding/hex" + "reflect" + "testing" + + "sources.truenas.cloud/code/edwards25519/field" +) + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +var B = NewGeneratorPoint() +var I = NewIdentityPoint() + +func checkOnCurve(t *testing.T, points ...*Point) { + t.Helper() + for i, p := range points { + if p.z.Equal(new(field.Element)) == 1 { + t.Errorf("point %d has Z == 0 (degenerate projective point)", i) + } + var XX, YY, ZZ, ZZZZ field.Element + XX.Square(&p.x) + YY.Square(&p.y) + ZZ.Square(&p.z) + ZZZZ.Square(&ZZ) + // -x² + y² = 1 + dx²y² + // -(X/Z)² + (Y/Z)² = 1 + d(X/Z)²(Y/Z)² + // (-X² + Y²)/Z² = 1 + (dX²Y²)/Z⁴ + // (-X² + Y²)*Z² = Z⁴ + dX²Y² + var lhs, rhs field.Element + lhs.Subtract(&YY, &XX).Multiply(&lhs, &ZZ) + rhs.Multiply(d, &XX).Multiply(&rhs, &YY).Add(&rhs, &ZZZZ) + if lhs.Equal(&rhs) != 1 { + t.Errorf("X, Y, and Z do not specify a point on the curve\nX = %v\nY = %v\nZ = %v", p.x, p.y, p.z) + } + // xy = T/Z + lhs.Multiply(&p.x, &p.y) + rhs.Multiply(&p.z, &p.t) + if lhs.Equal(&rhs) != 1 { + t.Errorf("point %d is not valid\nX = %v\nY = %v\nZ = %v", i, p.x, p.y, p.z) + } + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestGenerator(t *testing.T) { + // These are the coordinates of B from RFC 8032, Section 5.1, converted to + // little endian hex. + x := "1ad5258f602d56c9b2a7259560c72c695cdcd6fd31e2a4c0fe536ecdd3366921" + y := "5866666666666666666666666666666666666666666666666666666666666666" + if got := hex.EncodeToString(B.x.Bytes()); got != x { + t.Errorf("wrong B.x: got %s, expected %s", got, x) + } + if got := hex.EncodeToString(B.y.Bytes()); got != y { + t.Errorf("wrong B.y: got %s, expected %s", got, y) + } + if B.z.Equal(feOne) != 1 { + t.Errorf("wrong B.z: got %v, expected 1", B.z) + } + // Check that t is correct. + checkOnCurve(t, B) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestAddSubNegOnBasePoint(t *testing.T) { + checkLhs, checkRhs := &Point{}, &Point{} + + checkLhs.Add(B, B) + tmpP2 := new(projP2).FromP3(B) + tmpP1xP1 := new(projP1xP1).Double(tmpP2) + checkRhs.fromP1xP1(tmpP1xP1) + if checkLhs.Equal(checkRhs) != 1 { + t.Error("B + B != [2]B") + } + checkOnCurve(t, checkLhs, checkRhs) + + checkLhs.Subtract(B, B) + Bneg := new(Point).Negate(B) + checkRhs.Add(B, Bneg) + if checkLhs.Equal(checkRhs) != 1 { + t.Error("B - B != B + (-B)") + } + if I.Equal(checkLhs) != 1 { + t.Error("B - B != 0") + } + if I.Equal(checkRhs) != 1 { + t.Error("B + (-B) != 0") + } + checkOnCurve(t, checkLhs, checkRhs, Bneg) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestComparable(t *testing.T) { + if reflect.TypeOf(Point{}).Comparable() { + t.Error("Point is unexpectedly comparable") + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestInvalidEncodings(t *testing.T) { + // An invalid point, that also happens to have y > p. + invalid := "efffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff7f" + p := NewGeneratorPoint() + if out, err := p.SetBytes(decodeHex(invalid)); err == nil { + t.Error("expected error for invalid point") + } else if out != nil { + t.Error("SetBytes did not return nil on an invalid encoding") + } else if p.Equal(B) != 1 { + t.Error("the Point was modified while decoding an invalid encoding") + } + checkOnCurve(t, p) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestNonCanonicalPoints(t *testing.T) { + type test struct { + name string + encoding, canonical string + } + tests := []test{ + // Points with x = 0 and the sign bit set. With x = 0 the curve equation + // gives y² = 1, so y = ±1. 1 has two valid encodings. + { + "y=1,sign-", + "0100000000000000000000000000000000000000000000000000000000000080", + "0100000000000000000000000000000000000000000000000000000000000000", + }, + { + "y=p+1,sign-", + "eeffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff", + "0100000000000000000000000000000000000000000000000000000000000000", + }, + { + "y=p-1,sign-", + "ecffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff", + "ecffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff7f", + }, + + // Non-canonical y encodings with values 2²⁵⁵-19 (p) to 2²⁵⁵-1 (p+18). + { + "y=p,sign+", + "edffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff7f", + "0000000000000000000000000000000000000000000000000000000000000000", + }, + { + "y=p,sign-", + "edffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff", + "0000000000000000000000000000000000000000000000000000000000000080", + }, + { + "y=p+1,sign+", + "eeffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff7f", + "0100000000000000000000000000000000000000000000000000000000000000", + }, + // "y=p+1,sign-" is already tested above. + // p+2 is not a valid y-coordinate. + { + "y=p+3,sign+", + "f0ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff7f", + "0300000000000000000000000000000000000000000000000000000000000000", + }, + { + "y=p+3,sign-", + "f0ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff", + "0300000000000000000000000000000000000000000000000000000000000080", + }, + { + "y=p+4,sign+", + "f1ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff7f", + "0400000000000000000000000000000000000000000000000000000000000000", + }, + { + "y=p+4,sign-", + "f1ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff", + "0400000000000000000000000000000000000000000000000000000000000080", + }, + { + "y=p+5,sign+", + "f2ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff7f", + "0500000000000000000000000000000000000000000000000000000000000000", + }, + { + "y=p+5,sign-", + "f2ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff", + "0500000000000000000000000000000000000000000000000000000000000080", + }, + { + "y=p+6,sign+", + "f3ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff7f", + "0600000000000000000000000000000000000000000000000000000000000000", + }, + { + "y=p+6,sign-", + "f3ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff", + "0600000000000000000000000000000000000000000000000000000000000080", + }, + // p+7 is not a valid y-coordinate. + // p+8 is not a valid y-coordinate. + { + "y=p+9,sign+", + "f6ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff7f", + "0900000000000000000000000000000000000000000000000000000000000000", + }, + { + "y=p+9,sign-", + "f6ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff", + "0900000000000000000000000000000000000000000000000000000000000080", + }, + { + "y=p+10,sign+", + "f7ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff7f", + "0a00000000000000000000000000000000000000000000000000000000000000", + }, + { + "y=p+10,sign-", + "f7ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff", + "0a00000000000000000000000000000000000000000000000000000000000080", + }, + // p+11 is not a valid y-coordinate. + // p+12 is not a valid y-coordinate. + // p+13 is not a valid y-coordinate. + { + "y=p+14,sign+", + "fbffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff7f", + "0e00000000000000000000000000000000000000000000000000000000000000", + }, + { + "y=p+14,sign-", + "fbffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff", + "0e00000000000000000000000000000000000000000000000000000000000080", + }, + { + "y=p+15,sign+", + "fcffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff7f", + "0f00000000000000000000000000000000000000000000000000000000000000", + }, + { + "y=p+15,sign-", + "fcffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff", + "0f00000000000000000000000000000000000000000000000000000000000080", + }, + { + "y=p+16,sign+", + "fdffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff7f", + "1000000000000000000000000000000000000000000000000000000000000000", + }, + { + "y=p+16,sign-", + "fdffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff", + "1000000000000000000000000000000000000000000000000000000000000080", + }, + // p+17 is not a valid y-coordinate. + { + "y=p+18,sign+", + "ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff7f", + "1200000000000000000000000000000000000000000000000000000000000000", + }, + { + "y=p+18,sign-", + "ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff", + "1200000000000000000000000000000000000000000000000000000000000080", + }, + } + for _, tt := range tests { + t.Run(tt.name, func(t *testing.T) { + p1, err := new(Point).SetBytes(decodeHex(tt.encoding)) + if err != nil { + t.Fatalf("error decoding non-canonical point: %v", err) + } + p2, err := new(Point).SetBytes(decodeHex(tt.canonical)) + if err != nil { + t.Fatalf("error decoding canonical point: %v", err) + } + if p1.Equal(p2) != 1 { + t.Errorf("equivalent points are not equal: %v, %v", p1, p2) + } + if encoding := hex.EncodeToString(p1.Bytes()); encoding != tt.canonical { + t.Errorf("re-encoding does not match canonical; got %q, expected %q", encoding, tt.canonical) + } + checkOnCurve(t, p1, p2) + }) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +var testAllocationsSink byte + +func TestAllocations(t *testing.T) { + if allocs := testing.AllocsPerRun(100, func() { + p := NewIdentityPoint() + p.Add(p, NewGeneratorPoint()) + s := NewScalar() + testAllocationsSink ^= s.Bytes()[0] + testAllocationsSink ^= p.Bytes()[0] + }); allocs > 0 { + t.Errorf("expected zero allocations, got %0.1v", allocs) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func decodeHex(s string) []byte { + b, err := hex.DecodeString(s) + if err != nil { + panic(err) + } + return b +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func BenchmarkEncodingDecoding(b *testing.B) { + p := new(Point).Set(dalekScalarBasepoint) + for i := 0; i < b.N; i++ { + buf := p.Bytes() + _, err := p.SetBytes(buf) + if err != nil { + b.Fatal(err) + } + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/extra.go b/extra.go new file mode 100644 index 0000000..c31e080 --- /dev/null +++ b/extra.go @@ -0,0 +1,439 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519 +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package edwards25519 + +import ( + "errors" + "slices" + + "sources.truenas.cloud/code/edwards25519/field" +) + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// ExtendedCoordinates returns v in extended coordinates (X:Y:Z:T) where +// x = X/Z, y = Y/Z, and xy = T/Z as in https://eprint.iacr.org/2008/522. +func (v *Point) ExtendedCoordinates() (X, Y, Z, T *field.Element) { + // This function is outlined to make the allocations inline in the caller + // rather than happen on the heap. Don't change the style without making + // sure it doesn't increase the inliner cost. + var e [4]field.Element + X, Y, Z, T = v.extendedCoordinates(&e) + return +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v *Point) extendedCoordinates(e *[4]field.Element) (X, Y, Z, T *field.Element) { + checkInitialized(v) + X = e[0].Set(&v.x) + Y = e[1].Set(&v.y) + Z = e[2].Set(&v.z) + T = e[3].Set(&v.t) + return +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// SetExtendedCoordinates sets v = (X:Y:Z:T) in extended coordinates where +// x = X/Z, y = Y/Z, and xy = T/Z as in https://eprint.iacr.org/2008/522. +// +// If the coordinates are invalid or don't represent a valid point on the curve, +// SetExtendedCoordinates returns nil and an error and the receiver is +// unchanged. Otherwise, SetExtendedCoordinates returns v. +func (v *Point) SetExtendedCoordinates(X, Y, Z, T *field.Element) (*Point, error) { + if !isOnCurve(X, Y, Z, T) { + return nil, errors.New("edwards25519: invalid point coordinates") + } + v.x.Set(X) + v.y.Set(Y) + v.z.Set(Z) + v.t.Set(T) + return v, nil +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func isOnCurve(X, Y, Z, T *field.Element) bool { + var lhs, rhs field.Element + XX := new(field.Element).Square(X) + YY := new(field.Element).Square(Y) + ZZ := new(field.Element).Square(Z) + TT := new(field.Element).Square(T) + // -x² + y² = 1 + dx²y² + // -(X/Z)² + (Y/Z)² = 1 + d(T/Z)² + // -X² + Y² = Z² + dT² + lhs.Subtract(YY, XX) + rhs.Multiply(d, TT).Add(&rhs, ZZ) + if lhs.Equal(&rhs) != 1 { + return false + } + // xy = T/Z + // XY/Z² = T/Z + // XY = TZ + lhs.Multiply(X, Y) + rhs.Multiply(T, Z) + return lhs.Equal(&rhs) == 1 +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// BytesMontgomery converts v to a point on the birationally-equivalent +// Curve25519 Montgomery curve, and returns its canonical 32 bytes encoding +// according to RFC 7748. +// +// Note that BytesMontgomery only encodes the u-coordinate, so v and -v encode +// to the same value. If v is the identity point, BytesMontgomery returns 32 +// zero bytes, analogously to the X25519 function. +// +// The lack of an inverse operation (such as SetMontgomeryBytes) is deliberate: +// while every valid edwards25519 point has a unique u-coordinate Montgomery +// encoding, X25519 accepts inputs on the quadratic twist, which don't correspond +// to any edwards25519 point, and every other X25519 input corresponds to two +// edwards25519 points. +func (v *Point) BytesMontgomery() []byte { + // This function is outlined to make the allocations inline in the caller + // rather than happen on the heap. + var buf [32]byte + return v.bytesMontgomery(&buf) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v *Point) bytesMontgomery(buf *[32]byte) []byte { + checkInitialized(v) + + // RFC 7748, Section 4.1 provides the bilinear map to calculate the + // Montgomery u-coordinate + // + // u = (1 + y) / (1 - y) + // + // where y = Y / Z and therefore + // + // u = (Z + Y) / (Z - Y) + + var n, r, u field.Element + + n.Add(&v.z, &v.y) // n = Z + Y + r.Invert(r.Subtract(&v.z, &v.y)) // r = 1 / (Z - Y) + u.Multiply(&n, &r) // u = n * r + + return copyFieldElement(buf, &u) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// MultByCofactor sets v = 8 * p, and returns v. +func (v *Point) MultByCofactor(p *Point) *Point { + checkInitialized(p) + result := projP1xP1{} + pp := (&projP2{}).FromP3(p) + result.Double(pp) + pp.FromP1xP1(&result) + result.Double(pp) + pp.FromP1xP1(&result) + result.Double(pp) + return v.fromP1xP1(&result) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Given k > 0, set s = s**(2*k). +func (s *Scalar) pow2k(k int) { + for i := 0; i < k; i++ { + s.Multiply(s, s) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Invert sets s to the inverse of a nonzero scalar v, and returns s. +// +// If t is zero, Invert returns zero. +func (s *Scalar) Invert(t *Scalar) *Scalar { + // Uses a hardcoded sliding window of width 4. + var table [8]Scalar + var tt Scalar + tt.Multiply(t, t) + table[0] = *t + for i := 0; i < 7; i++ { + table[i+1].Multiply(&table[i], &tt) + } + // Now table = [t**1, t**3, t**5, t**7, t**9, t**11, t**13, t**15] + // so t**k = t[k/2] for odd k + + // To compute the sliding window digits, use the following Sage script: + + // sage: import itertools + // sage: def sliding_window(w,k): + // ....: digits = [] + // ....: while k > 0: + // ....: if k % 2 == 1: + // ....: kmod = k % (2**w) + // ....: digits.append(kmod) + // ....: k = k - kmod + // ....: else: + // ....: digits.append(0) + // ....: k = k // 2 + // ....: return digits + + // Now we can compute s roughly as follows: + + // sage: s = 1 + // sage: for coeff in reversed(sliding_window(4,l-2)): + // ....: s = s*s + // ....: if coeff > 0 : + // ....: s = s*t**coeff + + // This works on one bit at a time, with many runs of zeros. + // The digits can be collapsed into [(count, coeff)] as follows: + + // sage: [(len(list(group)),d) for d,group in itertools.groupby(sliding_window(4,l-2))] + + // Entries of the form (k, 0) turn into pow2k(k) + // Entries of the form (1, coeff) turn into a squaring and then a table lookup. + // We can fold the squaring into the previous pow2k(k) as pow2k(k+1). + + *s = table[1/2] + s.pow2k(127 + 1) + s.Multiply(s, &table[1/2]) + s.pow2k(4 + 1) + s.Multiply(s, &table[9/2]) + s.pow2k(3 + 1) + s.Multiply(s, &table[11/2]) + s.pow2k(3 + 1) + s.Multiply(s, &table[13/2]) + s.pow2k(3 + 1) + s.Multiply(s, &table[15/2]) + s.pow2k(4 + 1) + s.Multiply(s, &table[7/2]) + s.pow2k(4 + 1) + s.Multiply(s, &table[15/2]) + s.pow2k(3 + 1) + s.Multiply(s, &table[5/2]) + s.pow2k(3 + 1) + s.Multiply(s, &table[1/2]) + s.pow2k(4 + 1) + s.Multiply(s, &table[15/2]) + s.pow2k(4 + 1) + s.Multiply(s, &table[15/2]) + s.pow2k(4 + 1) + s.Multiply(s, &table[7/2]) + s.pow2k(3 + 1) + s.Multiply(s, &table[3/2]) + s.pow2k(4 + 1) + s.Multiply(s, &table[11/2]) + s.pow2k(5 + 1) + s.Multiply(s, &table[11/2]) + s.pow2k(9 + 1) + s.Multiply(s, &table[9/2]) + s.pow2k(3 + 1) + s.Multiply(s, &table[3/2]) + s.pow2k(4 + 1) + s.Multiply(s, &table[3/2]) + s.pow2k(4 + 1) + s.Multiply(s, &table[3/2]) + s.pow2k(4 + 1) + s.Multiply(s, &table[9/2]) + s.pow2k(3 + 1) + s.Multiply(s, &table[7/2]) + s.pow2k(3 + 1) + s.Multiply(s, &table[3/2]) + s.pow2k(3 + 1) + s.Multiply(s, &table[13/2]) + s.pow2k(3 + 1) + s.Multiply(s, &table[7/2]) + s.pow2k(4 + 1) + s.Multiply(s, &table[9/2]) + s.pow2k(3 + 1) + s.Multiply(s, &table[15/2]) + s.pow2k(4 + 1) + s.Multiply(s, &table[11/2]) + + return s +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// MultiScalarMult sets v = sum(scalars[i] * points[i]), and returns v. +// +// Execution time depends only on the lengths of the two slices, which must match. +func (v *Point) MultiScalarMult(scalars []*Scalar, points []*Point) *Point { + if len(scalars) != len(points) { + panic("edwards25519: called MultiScalarMult with different size inputs") + } + checkInitialized(points...) + + // Proceed as in the single-base case, but share doublings + // between each point in the multiscalar equation. + + // Build lookup tables for each point + tables := make([]projLookupTable, 0, 2) // avoid allocation for small sizes + tables = slices.Grow(tables, len(points))[:len(points)] + for i := range tables { + tables[i].FromP3(points[i]) + } + // Compute signed radix-16 digits for each scalar + digits := make([][64]int8, 0, 2) // avoid allocation for small sizes + digits = slices.Grow(digits, len(scalars))[:len(scalars)] + for i := range digits { + digits[i] = scalars[i].signedRadix16() + } + + // Unwrap first loop iteration to save computing 16*identity + multiple := &projCached{} + tmp1 := &projP1xP1{} + tmp2 := &projP2{} + // Lookup-and-add the appropriate multiple of each input point + v.Set(NewIdentityPoint()) + for j := range tables { + tables[j].SelectInto(multiple, digits[j][63]) + tmp1.Add(v, multiple) // tmp1 = v + x_(j,63)*Q in P1xP1 coords + v.fromP1xP1(tmp1) // update v + } + tmp2.FromP3(v) // set up tmp2 = v in P2 coords for next iteration + for i := 62; i >= 0; i-- { + tmp1.Double(tmp2) // tmp1 = 2*(prev) in P1xP1 coords + tmp2.FromP1xP1(tmp1) // tmp2 = 2*(prev) in P2 coords + tmp1.Double(tmp2) // tmp1 = 4*(prev) in P1xP1 coords + tmp2.FromP1xP1(tmp1) // tmp2 = 4*(prev) in P2 coords + tmp1.Double(tmp2) // tmp1 = 8*(prev) in P1xP1 coords + tmp2.FromP1xP1(tmp1) // tmp2 = 8*(prev) in P2 coords + tmp1.Double(tmp2) // tmp1 = 16*(prev) in P1xP1 coords + v.fromP1xP1(tmp1) // v = 16*(prev) in P3 coords + // Lookup-and-add the appropriate multiple of each input point + for j := range tables { + tables[j].SelectInto(multiple, digits[j][i]) + tmp1.Add(v, multiple) // tmp1 = v + x_(j,i)*Q in P1xP1 coords + v.fromP1xP1(tmp1) // update v + } + tmp2.FromP3(v) // set up tmp2 = v in P2 coords for next iteration + } + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// VarTimeMultiScalarMult sets v = sum(scalars[i] * points[i]), and returns v. +// +// Execution time depends on the inputs. +func (v *Point) VarTimeMultiScalarMult(scalars []*Scalar, points []*Point) *Point { + if len(scalars) != len(points) { + panic("edwards25519: called VarTimeMultiScalarMult with different size inputs") + } + checkInitialized(points...) + + // Generalize double-base NAF computation to arbitrary sizes. + // Here all the points are dynamic, so we only use the smaller + // tables. + + // Build lookup tables for each point + tables := make([]nafLookupTable5, len(points)) + for i := range tables { + tables[i].FromP3(points[i]) + } + // Compute a NAF for each scalar + nafs := make([][256]int8, len(scalars)) + for i := range nafs { + nafs[i] = scalars[i].nonAdjacentForm(5) + } + + multiple := &projCached{} + tmp1 := &projP1xP1{} + tmp2 := &projP2{} + tmp2.Zero() + + // Move from high to low bits, doubling the accumulator + // at each iteration and checking whether there is a nonzero + // coefficient to look up a multiple of. + // + // Skip trying to find the first nonzero coefficent, because + // searching might be more work than a few extra doublings. + for i := 255; i >= 0; i-- { + tmp1.Double(tmp2) + + for j := range nafs { + if nafs[j][i] > 0 { + v.fromP1xP1(tmp1) + tables[j].SelectInto(multiple, nafs[j][i]) + tmp1.Add(v, multiple) + } else if nafs[j][i] < 0 { + v.fromP1xP1(tmp1) + tables[j].SelectInto(multiple, -nafs[j][i]) + tmp1.Sub(v, multiple) + } + } + + tmp2.FromP1xP1(tmp1) + } + + v.fromP2(tmp2) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Select sets v to a if cond == 1 and to b if cond == 0. +func (v *Point) Select(a, b *Point, cond int) *Point { + checkInitialized(a, b) + v.x.Select(&a.x, &b.x, cond) + v.y.Select(&a.y, &b.y, cond) + v.z.Select(&a.z, &b.z, cond) + v.t.Select(&a.t, &b.t, cond) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Double sets v = p + p, and returns v. +func (v *Point) Double(p *Point) *Point { + checkInitialized(p) + + pp := new(projP2).FromP3(p) + p1 := new(projP1xP1).Double(pp) + return v.fromP1xP1(p1) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v *Point) addCached(p *Point, qCached *projCached) *Point { + result := new(projP1xP1).Add(p, qCached) + return v.fromP1xP1(result) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// ScalarMultSlow sets v = x * q, and returns v. It doesn't precompute a large +// table, so it is considerably slower, but requires less memory. +// +// The scalar multiplication is done in constant time. +func (v *Point) ScalarMultSlow(x *Scalar, q *Point) *Point { + checkInitialized(q) + + s := x.Bytes() + qCached := new(projCached).FromP3(q) + v.Set(NewIdentityPoint()) + t := new(Point) + + for i := 255; i >= 0; i-- { + v.Double(v) + t.addCached(v, qCached) + cond := (s[i/8] >> (i % 8)) & 1 + v.Select(t, v, int(cond)) + } + + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/extra_test.go b/extra_test.go new file mode 100644 index 0000000..882866d --- /dev/null +++ b/extra_test.go @@ -0,0 +1,346 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519 +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package edwards25519 + +import ( + "crypto/rand" + "encoding/hex" + "testing" + "testing/quick" +) + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// TestBytesMontgomery tests the SetBytesWithClamping+BytesMontgomery path +// equivalence to curve25519.X25519 for basepoint scalar multiplications. +// +// Note that you can't actually implement X25519 with this package because +// there is no SetBytesMontgomery, and it would not be possible to implement +// it properly: points on the twist would get rejected, and the Scalar returned +// by SetBytesWithClamping does not preserve its cofactor-clearing properties. +// +// Disabled to avoid the golang.org/x/crypto module dependency. +/* func TestBytesMontgomery(t *testing.T) { + f := func(scalar [32]byte) bool { + s := NewScalar().SetBytesWithClamping(scalar[:]) + p := (&Point{}).ScalarBaseMult(s) + got := p.BytesMontgomery() + want, _ := curve25519.X25519(scalar[:], curve25519.Basepoint) + return bytes.Equal(got, want) + } + if err := quick.Check(f, nil); err != nil { + t.Error(err) + } +} */ + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestBytesMontgomerySodium(t *testing.T) { + // Generated with libsodium.js 1.0.18 + // crypto_sign_keypair().publicKey + publicKey := "3bf918ffc2c955dc895bf145f566fb96623c1cadbe040091175764b5fde322c0" + p, err := (&Point{}).SetBytes(decodeHex(publicKey)) + if err != nil { + t.Fatal(err) + } + // crypto_sign_ed25519_pk_to_curve25519(publicKey) + want := "efc6c9d0738e9ea18d738ad4a2653631558931b0f1fde4dd58c436d19686dc28" + if got := hex.EncodeToString(p.BytesMontgomery()); got != want { + t.Errorf("got %q, want %q", got, want) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestBytesMontgomeryInfinity(t *testing.T) { + p := NewIdentityPoint() + want := "0000000000000000000000000000000000000000000000000000000000000000" + if got := hex.EncodeToString(p.BytesMontgomery()); got != want { + t.Errorf("got %q, want %q", got, want) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestMultByCofactor(t *testing.T) { + lowOrderBytes := "26e8958fc2b227b045c3f489f2ef98f0d5dfac05d3c63339b13802886d53fc85" + lowOrder, err := (&Point{}).SetBytes(decodeHex(lowOrderBytes)) + if err != nil { + t.Fatal(err) + } + + if p := (&Point{}).MultByCofactor(lowOrder); p.Equal(NewIdentityPoint()) != 1 { + t.Errorf("expected low order point * cofactor to be the identity") + } + + f := func(scalar [64]byte) bool { + s, _ := NewScalar().SetUniformBytes(scalar[:]) + p := (&Point{}).ScalarBaseMult(s) + p8 := (&Point{}).MultByCofactor(p) + checkOnCurve(t, p8) + + // 8 * p == (8 * s) * B + reprEight := [32]byte{8, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0} + scEight, _ := (&Scalar{}).SetCanonicalBytes(reprEight[:]) + s.Multiply(s, scEight) + pp := (&Point{}).ScalarBaseMult(s) + if p8.Equal(pp) != 1 { + return false + } + + // 8 * p == 8 * (lowOrder + p) + pp.Add(p, lowOrder) + pp.MultByCofactor(pp) + if p8.Equal(pp) != 1 { + return false + } + + // 8 * p == p + p + p + p + p + p + p + p + pp.Set(NewIdentityPoint()) + for i := 0; i < 8; i++ { + pp.Add(pp, p) + } + return p8.Equal(pp) == 1 + } + if err := quick.Check(f, nil); err != nil { + t.Error(err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestScalarInvert(t *testing.T) { + invertWorks := func(xInv Scalar, x notZeroScalar) bool { + xInv.Invert((*Scalar)(&x)) + var check Scalar + check.Multiply((*Scalar)(&x), &xInv) + + return check.Equal(scOne) == 1 && isReduced(xInv.Bytes()) + } + + if err := quick.Check(invertWorks, quickCheckConfig(32)); err != nil { + t.Error(err) + } + + randomScalar := *dalekScalar + randomInverse := NewScalar().Invert(&randomScalar) + var check Scalar + check.Multiply(&randomScalar, randomInverse) + + if check.Equal(scOne) == 0 || !isReduced(randomInverse.Bytes()) { + t.Error("inversion did not work") + } + + zero := NewScalar() + if xx := NewScalar().Invert(zero); xx.Equal(zero) != 1 { + t.Errorf("inverting zero did not return zero") + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestMultiScalarMultMatchesBaseMult(t *testing.T) { + multiScalarMultMatchesBaseMult := func(x, y, z Scalar) bool { + var p, q1, q2, q3, check Point + + p.MultiScalarMult([]*Scalar{&x, &y, &z}, []*Point{B, B, B}) + + q1.ScalarBaseMult(&x) + q2.ScalarBaseMult(&y) + q3.ScalarBaseMult(&z) + check.Add(&q1, &q2).Add(&check, &q3) + + checkOnCurve(t, &p, &check, &q1, &q2, &q3) + return p.Equal(&check) == 1 + } + + if err := quick.Check(multiScalarMultMatchesBaseMult, quickCheckConfig(32)); err != nil { + t.Error(err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestMultiScalarMultZeroReceiver(t *testing.T) { + // A zero-value (uninitialized) receiver should be handled correctly, + // producing a valid point on the curve. + var p Point + p.MultiScalarMult([]*Scalar{dalekScalar}, []*Point{B}) + + var check Point + check.ScalarBaseMult(dalekScalar) + + checkOnCurve(t, &p, &check) + if p.Equal(&check) != 1 { + t.Error("MultiScalarMult with zero-value receiver did not match ScalarBaseMult") + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestMultiScalarMultReceiverAliasing(t *testing.T) { + // The receiver v aliasing one of the input points should produce + // the correct result. + p := NewGeneratorPoint() + p.MultiScalarMult([]*Scalar{dalekScalar}, []*Point{p}) + + var check Point + check.ScalarBaseMult(dalekScalar) + + checkOnCurve(t, p, &check) + if p.Equal(&check) != 1 { + t.Error("MultiScalarMult with aliased receiver did not match ScalarBaseMult") + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestVarTimeMultiScalarMultMatchesBaseMult(t *testing.T) { + varTimeMultiScalarMultMatchesBaseMult := func(x, y, z Scalar) bool { + var p, q1, q2, q3, check Point + + p.VarTimeMultiScalarMult([]*Scalar{&x, &y, &z}, []*Point{B, B, B}) + + q1.ScalarBaseMult(&x) + q2.ScalarBaseMult(&y) + q3.ScalarBaseMult(&z) + check.Add(&q1, &q2).Add(&check, &q3) + + checkOnCurve(t, &p, &check, &q1, &q2, &q3) + return p.Equal(&check) == 1 + } + + if err := quick.Check(varTimeMultiScalarMultMatchesBaseMult, quickCheckConfig(32)); err != nil { + t.Error(err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestMultiScalarMult2NoAllocs(t *testing.T) { + p := NewIdentityPoint() + if allocs := testing.AllocsPerRun(100, func() { + p.MultiScalarMult([]*Scalar{dalekScalar, dalekScalar}, []*Point{B, B}) + }); allocs != 0 { + t.Errorf("MultiScalarMult allocated %v times, expected 0", allocs) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestScalarMultSlowMatchesMult(t *testing.T) { + scalarMultSlowMatchesMult := func(x, y Scalar) bool { + p := NewGeneratorPoint() + p.ScalarMultSlow(&x, p) + p.ScalarMultSlow(&y, p) + + q := NewGeneratorPoint() + q.ScalarMult(&x, B) + q.ScalarMult(&y, q) + + checkOnCurve(t, p, q) + return p.Equal(q) == 1 + } + + if err := quick.Check(scalarMultSlowMatchesMult, quickCheckConfig(32)); err != nil { + t.Error(err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func BenchmarkScalarMultSlow(b *testing.B) { + var p Point + x := dalekScalar + + for i := 0; i < b.N; i++ { + p.ScalarMultSlow(x, B) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func BenchmarkMultiScalarMultSize8(t *testing.B) { + var p Point + x := dalekScalar + + for i := 0; i < t.N; i++ { + p.MultiScalarMult([]*Scalar{x, x, x, x, x, x, x, x}, + []*Point{B, B, B, B, B, B, B, B}) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func BenchmarkScalarAddition(b *testing.B) { + var rnd [128]byte + rand.Read(rnd[:]) + s1, _ := (&Scalar{}).SetUniformBytes(rnd[0:64]) + s2, _ := (&Scalar{}).SetUniformBytes(rnd[64:128]) + t := &Scalar{} + + b.ResetTimer() + + for i := 0; i < b.N; i++ { + t.Add(s1, s2) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func BenchmarkScalarMultiplication(b *testing.B) { + var rnd [128]byte + rand.Read(rnd[:]) + s1, _ := (&Scalar{}).SetUniformBytes(rnd[0:64]) + s2, _ := (&Scalar{}).SetUniformBytes(rnd[64:128]) + t := &Scalar{} + + b.ResetTimer() + + for i := 0; i < b.N; i++ { + t.Multiply(s1, s2) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func BenchmarkScalarInversion(b *testing.B) { + var rnd [64]byte + rand.Read(rnd[:]) + s1, _ := (&Scalar{}).SetUniformBytes(rnd[0:64]) + + b.ResetTimer() + + for i := 0; i < b.N; i++ { + s1.Invert(s1) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func BenchmarkBytesMontgomery(b *testing.B) { + publicKey := "3bf918ffc2c955dc895bf145f566fb96623c1cadbe040091175764b5fde322c0" + p, err := (&Point{}).SetBytes(decodeHex(publicKey)) + if err != nil { + b.Fatal(err) + } + + b.ResetTimer() + + for i := 0; i < b.N; i++ { + _ = p.BytesMontgomery() + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/field/fe.go b/field/fe.go new file mode 100644 index 0000000..a780a39 --- /dev/null +++ b/field/fe.go @@ -0,0 +1,474 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519/field +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package field + +import ( + "crypto/subtle" + "encoding/binary" + "errors" + "math/bits" +) + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Element represents an element of the field GF(2^255-19). Note that this +// is not a cryptographically secure group, and should only be used to interact +// with edwards25519.Point coordinates. +// +// This type works similarly to math/big.Int, and all arguments and receivers +// are allowed to alias. +// +// The zero value is a valid zero element. +type Element struct { + // An element t represents the integer + // t.l0 + t.l1*2^51 + t.l2*2^102 + t.l3*2^153 + t.l4*2^204 + // + // Between operations, all limbs are expected to be lower than 2^52. + l0 uint64 + l1 uint64 + l2 uint64 + l3 uint64 + l4 uint64 +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +const maskLow51Bits uint64 = (1 << 51) - 1 + +var feZero = &Element{0, 0, 0, 0, 0} + +// Zero sets v = 0, and returns v. +func (v *Element) Zero() *Element { + *v = *feZero + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +var feOne = &Element{1, 0, 0, 0, 0} + +// One sets v = 1, and returns v. +func (v *Element) One() *Element { + *v = *feOne + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// reduce reduces v modulo 2^255 - 19 and returns it. +func (v *Element) reduce() *Element { + v.carryPropagate() + + // After the light reduction we now have a field element representation + // v < 2^255 + 2^13 * 19, but need v < 2^255 - 19. + + // If v >= 2^255 - 19, then v + 19 >= 2^255, which would overflow 2^255 - 1, + // generating a carry. That is, c will be 0 if v < 2^255 - 19, and 1 otherwise. + c := (v.l0 + 19) >> 51 + c = (v.l1 + c) >> 51 + c = (v.l2 + c) >> 51 + c = (v.l3 + c) >> 51 + c = (v.l4 + c) >> 51 + + // If v < 2^255 - 19 and c = 0, this will be a no-op. Otherwise, it's + // effectively applying the reduction identity to the carry. + v.l0 += 19 * c + + v.l1 += v.l0 >> 51 + v.l0 = v.l0 & maskLow51Bits + v.l2 += v.l1 >> 51 + v.l1 = v.l1 & maskLow51Bits + v.l3 += v.l2 >> 51 + v.l2 = v.l2 & maskLow51Bits + v.l4 += v.l3 >> 51 + v.l3 = v.l3 & maskLow51Bits + // no additional carry + v.l4 = v.l4 & maskLow51Bits + + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Add sets v = a + b, and returns v. +func (v *Element) Add(a, b *Element) *Element { + v.l0 = a.l0 + b.l0 + v.l1 = a.l1 + b.l1 + v.l2 = a.l2 + b.l2 + v.l3 = a.l3 + b.l3 + v.l4 = a.l4 + b.l4 + return v.carryPropagate() +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Subtract sets v = a - b, and returns v. +func (v *Element) Subtract(a, b *Element) *Element { + // We first add 2 * p, to guarantee the subtraction won't underflow, and + // then subtract b (which can be up to 2^255 + 2^13 * 19). + v.l0 = (a.l0 + 0xFFFFFFFFFFFDA) - b.l0 + v.l1 = (a.l1 + 0xFFFFFFFFFFFFE) - b.l1 + v.l2 = (a.l2 + 0xFFFFFFFFFFFFE) - b.l2 + v.l3 = (a.l3 + 0xFFFFFFFFFFFFE) - b.l3 + v.l4 = (a.l4 + 0xFFFFFFFFFFFFE) - b.l4 + return v.carryPropagate() +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Negate sets v = -a, and returns v. +func (v *Element) Negate(a *Element) *Element { + return v.Subtract(feZero, a) +} + +// Invert sets v = 1/z mod p, and returns v. +// +// If z == 0, Invert returns v = 0. +func (v *Element) Invert(z *Element) *Element { + // Inversion is implemented as exponentiation with exponent p − 2. It uses the + // same sequence of 255 squarings and 11 multiplications as [Curve25519]. + var z2, z9, z11, z2_5_0, z2_10_0, z2_20_0, z2_50_0, z2_100_0, t Element + + z2.Square(z) // 2 + t.Square(&z2) // 4 + t.Square(&t) // 8 + z9.Multiply(&t, z) // 9 + z11.Multiply(&z9, &z2) // 11 + t.Square(&z11) // 22 + z2_5_0.Multiply(&t, &z9) // 31 = 2^5 - 2^0 + + t.Square(&z2_5_0) // 2^6 - 2^1 + for i := 0; i < 4; i++ { + t.Square(&t) // 2^10 - 2^5 + } + z2_10_0.Multiply(&t, &z2_5_0) // 2^10 - 2^0 + + t.Square(&z2_10_0) // 2^11 - 2^1 + for i := 0; i < 9; i++ { + t.Square(&t) // 2^20 - 2^10 + } + z2_20_0.Multiply(&t, &z2_10_0) // 2^20 - 2^0 + + t.Square(&z2_20_0) // 2^21 - 2^1 + for i := 0; i < 19; i++ { + t.Square(&t) // 2^40 - 2^20 + } + t.Multiply(&t, &z2_20_0) // 2^40 - 2^0 + + t.Square(&t) // 2^41 - 2^1 + for i := 0; i < 9; i++ { + t.Square(&t) // 2^50 - 2^10 + } + z2_50_0.Multiply(&t, &z2_10_0) // 2^50 - 2^0 + + t.Square(&z2_50_0) // 2^51 - 2^1 + for i := 0; i < 49; i++ { + t.Square(&t) // 2^100 - 2^50 + } + z2_100_0.Multiply(&t, &z2_50_0) // 2^100 - 2^0 + + t.Square(&z2_100_0) // 2^101 - 2^1 + for i := 0; i < 99; i++ { + t.Square(&t) // 2^200 - 2^100 + } + t.Multiply(&t, &z2_100_0) // 2^200 - 2^0 + + t.Square(&t) // 2^201 - 2^1 + for i := 0; i < 49; i++ { + t.Square(&t) // 2^250 - 2^50 + } + t.Multiply(&t, &z2_50_0) // 2^250 - 2^0 + + t.Square(&t) // 2^251 - 2^1 + t.Square(&t) // 2^252 - 2^2 + t.Square(&t) // 2^253 - 2^3 + t.Square(&t) // 2^254 - 2^4 + t.Square(&t) // 2^255 - 2^5 + + return v.Multiply(&t, &z11) // 2^255 - 21 +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Set sets v = a, and returns v. +func (v *Element) Set(a *Element) *Element { + *v = *a + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// SetBytes sets v to x, where x is a 32-byte little-endian encoding. If x is +// not of the right length, SetBytes returns nil and an error, and the +// receiver is unchanged. +// +// Consistent with RFC 7748, the most significant bit (the high bit of the +// last byte) is ignored, and non-canonical values (2^255-19 through 2^255-1) +// are accepted. Note that this is laxer than specified by RFC 8032, but +// consistent with most Ed25519 implementations. +func (v *Element) SetBytes(x []byte) (*Element, error) { + if len(x) != 32 { + return nil, errors.New("edwards25519: invalid field element input size") + } + + // Bits 0:51 (bytes 0:8, bits 0:64, shift 0, mask 51). + v.l0 = binary.LittleEndian.Uint64(x[0:8]) + v.l0 &= maskLow51Bits + // Bits 51:102 (bytes 6:14, bits 48:112, shift 3, mask 51). + v.l1 = binary.LittleEndian.Uint64(x[6:14]) >> 3 + v.l1 &= maskLow51Bits + // Bits 102:153 (bytes 12:20, bits 96:160, shift 6, mask 51). + v.l2 = binary.LittleEndian.Uint64(x[12:20]) >> 6 + v.l2 &= maskLow51Bits + // Bits 153:204 (bytes 19:27, bits 152:216, shift 1, mask 51). + v.l3 = binary.LittleEndian.Uint64(x[19:27]) >> 1 + v.l3 &= maskLow51Bits + // Bits 204:255 (bytes 24:32, bits 192:256, shift 12, mask 51). + // Note: not bytes 25:33, shift 4, to avoid overread. + v.l4 = binary.LittleEndian.Uint64(x[24:32]) >> 12 + v.l4 &= maskLow51Bits + + return v, nil +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Bytes returns the canonical 32-byte little-endian encoding of v. +func (v *Element) Bytes() []byte { + // This function is outlined to make the allocations inline in the caller + // rather than happen on the heap. + var out [32]byte + return v.bytes(&out) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v *Element) bytes(out *[32]byte) []byte { + t := *v + t.reduce() + + // Pack five 51-bit limbs into four 64-bit words: + // + // 255 204 153 102 51 0 + // ├──l4──┼──l3──┼──l2──┼──l1──┼──l0──┤ + // ├───u3───┼───u2───┼───u1───┼───u0───┤ + // 256 192 128 64 0 + + u0 := t.l1<<51 | t.l0 + u1 := t.l2<<(102-64) | t.l1>>(64-51) + u2 := t.l3<<(153-128) | t.l2>>(128-102) + u3 := t.l4<<(204-192) | t.l3>>(192-153) + + binary.LittleEndian.PutUint64(out[0*8:], u0) + binary.LittleEndian.PutUint64(out[1*8:], u1) + binary.LittleEndian.PutUint64(out[2*8:], u2) + binary.LittleEndian.PutUint64(out[3*8:], u3) + + return out[:] +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Equal returns 1 if v and u are equal, and 0 otherwise. +func (v *Element) Equal(u *Element) int { + sa, sv := u.Bytes(), v.Bytes() + return subtle.ConstantTimeCompare(sa, sv) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// mask64Bits returns 0xffffffff if cond is 1, and 0 otherwise. +func mask64Bits(cond int) uint64 { return ^(uint64(cond) - 1) } + +// Select sets v to a if cond == 1, and to b if cond == 0. +func (v *Element) Select(a, b *Element, cond int) *Element { + m := mask64Bits(cond) + v.l0 = (m & a.l0) | (^m & b.l0) + v.l1 = (m & a.l1) | (^m & b.l1) + v.l2 = (m & a.l2) | (^m & b.l2) + v.l3 = (m & a.l3) | (^m & b.l3) + v.l4 = (m & a.l4) | (^m & b.l4) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Swap swaps v and u if cond == 1 or leaves them unchanged if cond == 0, and returns v. +func (v *Element) Swap(u *Element, cond int) { + m := mask64Bits(cond) + t := m & (v.l0 ^ u.l0) + v.l0 ^= t + u.l0 ^= t + t = m & (v.l1 ^ u.l1) + v.l1 ^= t + u.l1 ^= t + t = m & (v.l2 ^ u.l2) + v.l2 ^= t + u.l2 ^= t + t = m & (v.l3 ^ u.l3) + v.l3 ^= t + u.l3 ^= t + t = m & (v.l4 ^ u.l4) + v.l4 ^= t + u.l4 ^= t +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// IsNegative returns 1 if v is negative, and 0 otherwise. +func (v *Element) IsNegative() int { + return int(v.Bytes()[0] & 1) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Absolute sets v to |u|, and returns v. +func (v *Element) Absolute(u *Element) *Element { + return v.Select(new(Element).Negate(u), u, u.IsNegative()) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Multiply sets v = x * y, and returns v. +func (v *Element) Multiply(x, y *Element) *Element { + feMul(v, x, y) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Square sets v = x * x, and returns v. +func (v *Element) Square(x *Element) *Element { + feSquare(v, x) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Mult32 sets v = x * y, and returns v. +func (v *Element) Mult32(x *Element, y uint32) *Element { + x0lo, x0hi := mul51(x.l0, y) + x1lo, x1hi := mul51(x.l1, y) + x2lo, x2hi := mul51(x.l2, y) + x3lo, x3hi := mul51(x.l3, y) + x4lo, x4hi := mul51(x.l4, y) + v.l0 = x0lo + 19*x4hi // carried over per the reduction identity + v.l1 = x1lo + x0hi + v.l2 = x2lo + x1hi + v.l3 = x3lo + x2hi + v.l4 = x4lo + x3hi + // The hi portions are going to be only 32 bits, plus any previous excess, + // so we can skip the carry propagation. + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// mul51 returns lo + hi * 2⁵¹ = a * b. +func mul51(a uint64, b uint32) (lo uint64, hi uint64) { + mh, ml := bits.Mul64(a, uint64(b)) + lo = ml & maskLow51Bits + hi = (mh << 13) | (ml >> 51) + return +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Pow22523 set v = x^((p-5)/8), and returns v. (p-5)/8 is 2^252-3. +func (v *Element) Pow22523(x *Element) *Element { + var t0, t1, t2 Element + + t0.Square(x) // x^2 + t1.Square(&t0) // x^4 + t1.Square(&t1) // x^8 + t1.Multiply(x, &t1) // x^9 + t0.Multiply(&t0, &t1) // x^11 + t0.Square(&t0) // x^22 + t0.Multiply(&t1, &t0) // x^31 + t1.Square(&t0) // x^62 + for i := 1; i < 5; i++ { // x^992 + t1.Square(&t1) + } + t0.Multiply(&t1, &t0) // x^1023 -> 1023 = 2^10 - 1 + t1.Square(&t0) // 2^11 - 2 + for i := 1; i < 10; i++ { // 2^20 - 2^10 + t1.Square(&t1) + } + t1.Multiply(&t1, &t0) // 2^20 - 1 + t2.Square(&t1) // 2^21 - 2 + for i := 1; i < 20; i++ { // 2^40 - 2^20 + t2.Square(&t2) + } + t1.Multiply(&t2, &t1) // 2^40 - 1 + t1.Square(&t1) // 2^41 - 2 + for i := 1; i < 10; i++ { // 2^50 - 2^10 + t1.Square(&t1) + } + t0.Multiply(&t1, &t0) // 2^50 - 1 + t1.Square(&t0) // 2^51 - 2 + for i := 1; i < 50; i++ { // 2^100 - 2^50 + t1.Square(&t1) + } + t1.Multiply(&t1, &t0) // 2^100 - 1 + t2.Square(&t1) // 2^101 - 2 + for i := 1; i < 100; i++ { // 2^200 - 2^100 + t2.Square(&t2) + } + t1.Multiply(&t2, &t1) // 2^200 - 1 + t1.Square(&t1) // 2^201 - 2 + for i := 1; i < 50; i++ { // 2^250 - 2^50 + t1.Square(&t1) + } + t0.Multiply(&t1, &t0) // 2^250 - 1 + t0.Square(&t0) // 2^251 - 2 + t0.Square(&t0) // 2^252 - 4 + return v.Multiply(&t0, x) // 2^252 - 3 -> x^(2^252-3) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// sqrtM1 is 2^((p-1)/4), which squared is equal to -1 by Euler's Criterion. +var sqrtM1 = &Element{1718705420411056, 234908883556509, + 2233514472574048, 2117202627021982, 765476049583133} + +// SqrtRatio sets r to the non-negative square root of the ratio of u and v. +// +// If u/v is square, SqrtRatio returns r and 1. If u/v is not square, SqrtRatio +// sets r according to Section 4.3 of draft-irtf-cfrg-ristretto255-decaf448-00, +// and returns r and 0. +func (r *Element) SqrtRatio(u, v *Element) (R *Element, wasSquare int) { + t0 := new(Element) + + // r = (u * v3) * (u * v7)^((p-5)/8) + v2 := new(Element).Square(v) + uv3 := new(Element).Multiply(u, t0.Multiply(v2, v)) + uv7 := new(Element).Multiply(uv3, t0.Square(v2)) + rr := new(Element).Multiply(uv3, t0.Pow22523(uv7)) + + check := new(Element).Multiply(v, t0.Square(rr)) // check = v * r^2 + + uNeg := new(Element).Negate(u) + correctSignSqrt := check.Equal(u) + flippedSignSqrt := check.Equal(uNeg) + flippedSignSqrtI := check.Equal(t0.Multiply(uNeg, sqrtM1)) + + rPrime := new(Element).Multiply(rr, sqrtM1) // r_prime = SQRT_M1 * r + // r = CT_SELECT(r_prime IF flipped_sign_sqrt | flipped_sign_sqrt_i ELSE r) + rr.Select(rPrime, rr, flippedSignSqrt|flippedSignSqrtI) + + r.Absolute(rr) // Choose the nonnegative square root. + return r, correctSignSqrt | flippedSignSqrt +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/field/fe_alias_test.go b/field/fe_alias_test.go new file mode 100644 index 0000000..6e2bab5 --- /dev/null +++ b/field/fe_alias_test.go @@ -0,0 +1,157 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519/field +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package field + +import ( + "testing" + "testing/quick" +) + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func checkAliasingOneArg(f func(v, x *Element) *Element) func(v, x Element) bool { + return func(v, x Element) bool { + x1, v1 := x, x + + // Calculate a reference f(x) without aliasing. + if out := f(&v, &x); out != &v && isInBounds(out) { + return false + } + + // Test aliasing the argument and the receiver. + if out := f(&v1, &v1); out != &v1 || v1 != v { + return false + } + + // Ensure the arguments was not modified. + return x == x1 + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func checkAliasingTwoArgs(f func(v, x, y *Element) *Element) func(v, x, y Element) bool { + return func(v, x, y Element) bool { + x1, y1, v1 := x, y, Element{} + + // Calculate a reference f(x, y) without aliasing. + if out := f(&v, &x, &y); out != &v && isInBounds(out) { + return false + } + + // Test aliasing the first argument and the receiver. + v1 = x + if out := f(&v1, &v1, &y); out != &v1 || v1 != v { + return false + } + // Test aliasing the second argument and the receiver. + v1 = y + if out := f(&v1, &x, &v1); out != &v1 || v1 != v { + return false + } + + // Calculate a reference f(x, x) without aliasing. + if out := f(&v, &x, &x); out != &v { + return false + } + + // Test aliasing the first argument and the receiver. + v1 = x + if out := f(&v1, &v1, &x); out != &v1 || v1 != v { + return false + } + // Test aliasing the second argument and the receiver. + v1 = x + if out := f(&v1, &x, &v1); out != &v1 || v1 != v { + return false + } + // Test aliasing both arguments and the receiver. + v1 = x + if out := f(&v1, &v1, &v1); out != &v1 || v1 != v { + return false + } + + // Ensure the arguments were not modified. + return x == x1 && y == y1 + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// TestAliasing checks that receivers and arguments can alias each other without +// leading to incorrect results. That is, it ensures that it's safe to write +// +// v.Invert(v) +// +// or +// +// v.Add(v, v) +// +// without any of the inputs getting clobbered by the output being written. +func TestAliasing(t *testing.T) { + type target struct { + name string + oneArgF func(v, x *Element) *Element + twoArgsF func(v, x, y *Element) *Element + } + for _, tt := range []target{ + {name: "Absolute", oneArgF: (*Element).Absolute}, + {name: "Invert", oneArgF: (*Element).Invert}, + {name: "Negate", oneArgF: (*Element).Negate}, + {name: "Set", oneArgF: (*Element).Set}, + {name: "Square", oneArgF: (*Element).Square}, + {name: "Pow22523", oneArgF: (*Element).Pow22523}, + { + name: "Mult32", + oneArgF: func(v, x *Element) *Element { + return v.Mult32(x, 0xffffffff) + }, + }, + {name: "Multiply", twoArgsF: (*Element).Multiply}, + {name: "Add", twoArgsF: (*Element).Add}, + {name: "Subtract", twoArgsF: (*Element).Subtract}, + { + name: "SqrtRatio", + twoArgsF: func(v, x, y *Element) *Element { + r, _ := v.SqrtRatio(x, y) + return r + }, + }, + { + name: "Select0", + twoArgsF: func(v, x, y *Element) *Element { + return v.Select(x, y, 0) + }, + }, + { + name: "Select1", + twoArgsF: func(v, x, y *Element) *Element { + return v.Select(x, y, 1) + }, + }, + } { + var err error + switch { + case tt.oneArgF != nil: + err = quick.Check(checkAliasingOneArg(tt.oneArgF), quickCheckConfig(256)) + case tt.twoArgsF != nil: + err = quick.Check(checkAliasingTwoArgs(tt.twoArgsF), quickCheckConfig(256)) + } + if err != nil { + t.Errorf("%v: %v", tt.name, err) + } + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/field/fe_amd64.go b/field/fe_amd64.go new file mode 100644 index 0000000..00bf8f4 --- /dev/null +++ b/field/fe_amd64.go @@ -0,0 +1,15 @@ +// Code generated by command: go run fe_amd64_asm.go -out ../fe_amd64.s -stubs ../fe_amd64.go -pkg field. DO NOT EDIT. + +//go:build !purego + +package field + +// feMul sets out = a * b. It works like feMulGeneric. +// +//go:noescape +func feMul(out *Element, a *Element, b *Element) + +// feSquare sets out = a * a. It works like feSquareGeneric. +// +//go:noescape +func feSquare(out *Element, a *Element) diff --git a/field/fe_amd64.s b/field/fe_amd64.s new file mode 100644 index 0000000..5e06e24 --- /dev/null +++ b/field/fe_amd64.s @@ -0,0 +1,398 @@ +// Code generated by command: go run fe_amd64_asm.go -out ../fe_amd64.s -stubs ../fe_amd64.go -pkg field. DO NOT EDIT. + +//go:build !purego + +#include "textflag.h" + +// func feMul(out *Element, a *Element, b *Element) +TEXT ·feMul(SB), NOSPLIT, $0-24 + MOVQ a+8(FP), CX + MOVQ b+16(FP), BX + + // r0 = a0×b0 + MOVQ (CX), AX + MULQ (BX) + MOVQ AX, DI + MOVQ DX, SI + + // r0 += 19×a1×b4 + MOVQ 8(CX), DX + LEAQ (DX)(DX*8), AX + LEAQ (DX)(AX*2), AX + MULQ 32(BX) + ADDQ AX, DI + ADCQ DX, SI + + // r0 += 19×a2×b3 + MOVQ 16(CX), DX + LEAQ (DX)(DX*8), AX + LEAQ (DX)(AX*2), AX + MULQ 24(BX) + ADDQ AX, DI + ADCQ DX, SI + + // r0 += 19×a3×b2 + MOVQ 24(CX), DX + LEAQ (DX)(DX*8), AX + LEAQ (DX)(AX*2), AX + MULQ 16(BX) + ADDQ AX, DI + ADCQ DX, SI + + // r0 += 19×a4×b1 + MOVQ 32(CX), DX + LEAQ (DX)(DX*8), AX + LEAQ (DX)(AX*2), AX + MULQ 8(BX) + ADDQ AX, DI + ADCQ DX, SI + + // r1 = a0×b1 + MOVQ (CX), AX + MULQ 8(BX) + MOVQ AX, R9 + MOVQ DX, R8 + + // r1 += a1×b0 + MOVQ 8(CX), AX + MULQ (BX) + ADDQ AX, R9 + ADCQ DX, R8 + + // r1 += 19×a2×b4 + MOVQ 16(CX), DX + LEAQ (DX)(DX*8), AX + LEAQ (DX)(AX*2), AX + MULQ 32(BX) + ADDQ AX, R9 + ADCQ DX, R8 + + // r1 += 19×a3×b3 + MOVQ 24(CX), DX + LEAQ (DX)(DX*8), AX + LEAQ (DX)(AX*2), AX + MULQ 24(BX) + ADDQ AX, R9 + ADCQ DX, R8 + + // r1 += 19×a4×b2 + MOVQ 32(CX), DX + LEAQ (DX)(DX*8), AX + LEAQ (DX)(AX*2), AX + MULQ 16(BX) + ADDQ AX, R9 + ADCQ DX, R8 + + // r2 = a0×b2 + MOVQ (CX), AX + MULQ 16(BX) + MOVQ AX, R11 + MOVQ DX, R10 + + // r2 += a1×b1 + MOVQ 8(CX), AX + MULQ 8(BX) + ADDQ AX, R11 + ADCQ DX, R10 + + // r2 += a2×b0 + MOVQ 16(CX), AX + MULQ (BX) + ADDQ AX, R11 + ADCQ DX, R10 + + // r2 += 19×a3×b4 + MOVQ 24(CX), DX + LEAQ (DX)(DX*8), AX + LEAQ (DX)(AX*2), AX + MULQ 32(BX) + ADDQ AX, R11 + ADCQ DX, R10 + + // r2 += 19×a4×b3 + MOVQ 32(CX), DX + LEAQ (DX)(DX*8), AX + LEAQ (DX)(AX*2), AX + MULQ 24(BX) + ADDQ AX, R11 + ADCQ DX, R10 + + // r3 = a0×b3 + MOVQ (CX), AX + MULQ 24(BX) + MOVQ AX, R13 + MOVQ DX, R12 + + // r3 += a1×b2 + MOVQ 8(CX), AX + MULQ 16(BX) + ADDQ AX, R13 + ADCQ DX, R12 + + // r3 += a2×b1 + MOVQ 16(CX), AX + MULQ 8(BX) + ADDQ AX, R13 + ADCQ DX, R12 + + // r3 += a3×b0 + MOVQ 24(CX), AX + MULQ (BX) + ADDQ AX, R13 + ADCQ DX, R12 + + // r3 += 19×a4×b4 + MOVQ 32(CX), DX + LEAQ (DX)(DX*8), AX + LEAQ (DX)(AX*2), AX + MULQ 32(BX) + ADDQ AX, R13 + ADCQ DX, R12 + + // r4 = a0×b4 + MOVQ (CX), AX + MULQ 32(BX) + MOVQ AX, R15 + MOVQ DX, R14 + + // r4 += a1×b3 + MOVQ 8(CX), AX + MULQ 24(BX) + ADDQ AX, R15 + ADCQ DX, R14 + + // r4 += a2×b2 + MOVQ 16(CX), AX + MULQ 16(BX) + ADDQ AX, R15 + ADCQ DX, R14 + + // r4 += a3×b1 + MOVQ 24(CX), AX + MULQ 8(BX) + ADDQ AX, R15 + ADCQ DX, R14 + + // r4 += a4×b0 + MOVQ 32(CX), AX + MULQ (BX) + ADDQ AX, R15 + ADCQ DX, R14 + + // First reduction chain + MOVQ $0x0007ffffffffffff, AX + SHLQ $0x0d, DI, SI + SHLQ $0x0d, R9, R8 + SHLQ $0x0d, R11, R10 + SHLQ $0x0d, R13, R12 + SHLQ $0x0d, R15, R14 + ANDQ AX, DI + IMUL3Q $0x13, R14, R14 + ADDQ R14, DI + ANDQ AX, R9 + ADDQ SI, R9 + ANDQ AX, R11 + ADDQ R8, R11 + ANDQ AX, R13 + ADDQ R10, R13 + ANDQ AX, R15 + ADDQ R12, R15 + + // Second reduction chain (carryPropagate) + MOVQ DI, SI + SHRQ $0x33, SI + MOVQ R9, R8 + SHRQ $0x33, R8 + MOVQ R11, R10 + SHRQ $0x33, R10 + MOVQ R13, R12 + SHRQ $0x33, R12 + MOVQ R15, R14 + SHRQ $0x33, R14 + ANDQ AX, DI + IMUL3Q $0x13, R14, R14 + ADDQ R14, DI + ANDQ AX, R9 + ADDQ SI, R9 + ANDQ AX, R11 + ADDQ R8, R11 + ANDQ AX, R13 + ADDQ R10, R13 + ANDQ AX, R15 + ADDQ R12, R15 + + // Store output + MOVQ out+0(FP), AX + MOVQ DI, (AX) + MOVQ R9, 8(AX) + MOVQ R11, 16(AX) + MOVQ R13, 24(AX) + MOVQ R15, 32(AX) + RET + +// func feSquare(out *Element, a *Element) +TEXT ·feSquare(SB), NOSPLIT, $0-16 + MOVQ a+8(FP), CX + + // r0 = l0×l0 + MOVQ (CX), AX + MULQ (CX) + MOVQ AX, SI + MOVQ DX, BX + + // r0 += 38×l1×l4 + MOVQ 8(CX), DX + LEAQ (DX)(DX*8), AX + LEAQ (DX)(AX*2), AX + SHLQ $0x01, AX + MULQ 32(CX) + ADDQ AX, SI + ADCQ DX, BX + + // r0 += 38×l2×l3 + MOVQ 16(CX), DX + LEAQ (DX)(DX*8), AX + LEAQ (DX)(AX*2), AX + SHLQ $0x01, AX + MULQ 24(CX) + ADDQ AX, SI + ADCQ DX, BX + + // r1 = 2×l0×l1 + MOVQ (CX), AX + SHLQ $0x01, AX + MULQ 8(CX) + MOVQ AX, R8 + MOVQ DX, DI + + // r1 += 38×l2×l4 + MOVQ 16(CX), DX + LEAQ (DX)(DX*8), AX + LEAQ (DX)(AX*2), AX + SHLQ $0x01, AX + MULQ 32(CX) + ADDQ AX, R8 + ADCQ DX, DI + + // r1 += 19×l3×l3 + MOVQ 24(CX), DX + LEAQ (DX)(DX*8), AX + LEAQ (DX)(AX*2), AX + MULQ 24(CX) + ADDQ AX, R8 + ADCQ DX, DI + + // r2 = 2×l0×l2 + MOVQ (CX), AX + SHLQ $0x01, AX + MULQ 16(CX) + MOVQ AX, R10 + MOVQ DX, R9 + + // r2 += l1×l1 + MOVQ 8(CX), AX + MULQ 8(CX) + ADDQ AX, R10 + ADCQ DX, R9 + + // r2 += 38×l3×l4 + MOVQ 24(CX), DX + LEAQ (DX)(DX*8), AX + LEAQ (DX)(AX*2), AX + SHLQ $0x01, AX + MULQ 32(CX) + ADDQ AX, R10 + ADCQ DX, R9 + + // r3 = 2×l0×l3 + MOVQ (CX), AX + SHLQ $0x01, AX + MULQ 24(CX) + MOVQ AX, R12 + MOVQ DX, R11 + + // r3 += 2×l1×l2 + MOVQ 8(CX), AX + SHLQ $0x01, AX + MULQ 16(CX) + ADDQ AX, R12 + ADCQ DX, R11 + + // r3 += 19×l4×l4 + MOVQ 32(CX), DX + LEAQ (DX)(DX*8), AX + LEAQ (DX)(AX*2), AX + MULQ 32(CX) + ADDQ AX, R12 + ADCQ DX, R11 + + // r4 = 2×l0×l4 + MOVQ (CX), AX + SHLQ $0x01, AX + MULQ 32(CX) + MOVQ AX, R14 + MOVQ DX, R13 + + // r4 += 2×l1×l3 + MOVQ 8(CX), AX + SHLQ $0x01, AX + MULQ 24(CX) + ADDQ AX, R14 + ADCQ DX, R13 + + // r4 += l2×l2 + MOVQ 16(CX), AX + MULQ 16(CX) + ADDQ AX, R14 + ADCQ DX, R13 + + // First reduction chain + MOVQ $0x0007ffffffffffff, AX + SHLQ $0x0d, SI, BX + SHLQ $0x0d, R8, DI + SHLQ $0x0d, R10, R9 + SHLQ $0x0d, R12, R11 + SHLQ $0x0d, R14, R13 + ANDQ AX, SI + IMUL3Q $0x13, R13, R13 + ADDQ R13, SI + ANDQ AX, R8 + ADDQ BX, R8 + ANDQ AX, R10 + ADDQ DI, R10 + ANDQ AX, R12 + ADDQ R9, R12 + ANDQ AX, R14 + ADDQ R11, R14 + + // Second reduction chain (carryPropagate) + MOVQ SI, BX + SHRQ $0x33, BX + MOVQ R8, DI + SHRQ $0x33, DI + MOVQ R10, R9 + SHRQ $0x33, R9 + MOVQ R12, R11 + SHRQ $0x33, R11 + MOVQ R14, R13 + SHRQ $0x33, R13 + ANDQ AX, SI + IMUL3Q $0x13, R13, R13 + ADDQ R13, SI + ANDQ AX, R8 + ADDQ BX, R8 + ANDQ AX, R10 + ADDQ DI, R10 + ANDQ AX, R12 + ADDQ R9, R12 + ANDQ AX, R14 + ADDQ R11, R14 + + // Store output + MOVQ out+0(FP), AX + MOVQ SI, (AX) + MOVQ R8, 8(AX) + MOVQ R10, 16(AX) + MOVQ R12, 24(AX) + MOVQ R14, 32(AX) + RET diff --git a/field/fe_amd64_noasm.go b/field/fe_amd64_noasm.go new file mode 100644 index 0000000..eb187dc --- /dev/null +++ b/field/fe_amd64_noasm.go @@ -0,0 +1,20 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519/field +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +//go:build !amd64 || purego + +package field + +func feMul(v, x, y *Element) { feMulGeneric(v, x, y) } + +func feSquare(v, x *Element) { feSquareGeneric(v, x) } diff --git a/field/fe_bench_test.go b/field/fe_bench_test.go new file mode 100644 index 0000000..b945f84 --- /dev/null +++ b/field/fe_bench_test.go @@ -0,0 +1,82 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519/field +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package field + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +import "testing" + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func BenchmarkAdd(b *testing.B) { + x := new(Element).One() + y := new(Element).Add(x, x) + b.ResetTimer() + for i := 0; i < b.N; i++ { + x.Add(x, y) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func BenchmarkMultiply(b *testing.B) { + x := new(Element).One() + y := new(Element).Add(x, x) + b.ResetTimer() + for i := 0; i < b.N; i++ { + x.Multiply(x, y) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func BenchmarkSquare(b *testing.B) { + x := new(Element).Add(feOne, feOne) + b.ResetTimer() + for i := 0; i < b.N; i++ { + x.Square(x) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func BenchmarkInvert(b *testing.B) { + x := new(Element).Add(feOne, feOne) + b.ResetTimer() + for i := 0; i < b.N; i++ { + x.Invert(x) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func BenchmarkMult32(b *testing.B) { + x := new(Element).One() + b.ResetTimer() + for i := 0; i < b.N; i++ { + x.Mult32(x, 0xaa42aa42) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func BenchmarkBytes(b *testing.B) { + x := new(Element).One() + b.ResetTimer() + for i := 0; i < b.N; i++ { + x.Bytes() + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/field/fe_extra.go b/field/fe_extra.go new file mode 100644 index 0000000..2486c90 --- /dev/null +++ b/field/fe_extra.go @@ -0,0 +1,63 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519/field +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package field + +import "errors" + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// This file contains additional functionality that is not included in the +// upstream crypto/ed25519/edwards25519/field package. + +// SetWideBytes sets v to x, where x is a 64-byte little-endian encoding, which +// is reduced modulo the field order. If x is not of the right length, +// SetWideBytes returns nil and an error, and the receiver is unchanged. +// +// SetWideBytes is not necessary to select a uniformly distributed value, and is +// only provided for compatibility: SetBytes can be used instead as the chance +// of bias is less than 2⁻²⁵⁰. +func (v *Element) SetWideBytes(x []byte) (*Element, error) { + if len(x) != 64 { + return nil, errors.New("edwards25519: invalid SetWideBytes input size") + } + + // Split the 64 bytes into two elements, and extract the most significant + // bit of each, which is ignored by SetBytes. + lo, _ := new(Element).SetBytes(x[:32]) + loMSB := uint64(x[31] >> 7) + hi, _ := new(Element).SetBytes(x[32:]) + hiMSB := uint64(x[63] >> 7) + + // The output we want is + // + // v = lo + loMSB * 2²⁵⁵ + hi * 2²⁵⁶ + hiMSB * 2⁵¹¹ + // + // which applying the reduction identity comes out to + // + // v = lo + loMSB * 19 + hi * 2 * 19 + hiMSB * 2 * 19² + // + // l0 will be the sum of a 52 bits value (lo.l0), plus a 5 bits value + // (loMSB * 19), a 6 bits value (hi.l0 * 2 * 19), and a 10 bits value + // (hiMSB * 2 * 19²), so it fits in a uint64. + + v.l0 = lo.l0 + loMSB*19 + hi.l0*2*19 + hiMSB*2*19*19 + v.l1 = lo.l1 + hi.l1*2*19 + v.l2 = lo.l2 + hi.l2*2*19 + v.l3 = lo.l3 + hi.l3*2*19 + v.l4 = lo.l4 + hi.l4*2*19 + + return v.carryPropagate(), nil +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/field/fe_extra_test.go b/field/fe_extra_test.go new file mode 100644 index 0000000..ea5d6df --- /dev/null +++ b/field/fe_extra_test.go @@ -0,0 +1,50 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519/field +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package field + +import ( + "math/big" + "testing" + "testing/quick" +) + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +var bigP = new(big.Int).Sub(new(big.Int).Lsh(big.NewInt(1), 255), big.NewInt(19)) + +func TestSetWideBytes(t *testing.T) { + f1 := func(in [64]byte, fe Element) bool { + fe1 := new(Element).Set(&fe) + + if out, err := fe.SetWideBytes([]byte{42}); err == nil || out != nil || + fe.Equal(fe1) != 1 { + return false + } + + if out, err := fe.SetWideBytes(in[:]); err != nil || out != &fe { + return false + } + + b := new(big.Int).SetBytes(swapEndianness(in[:])) + fe1.fromBig(b.Mod(b, bigP)) + + return fe.Equal(fe1) == 1 && isInBounds(&fe) && isInBounds(fe1) + } + if err := quick.Check(f1, nil); err != nil { + t.Error(err) + } + +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/field/fe_generic.go b/field/fe_generic.go new file mode 100644 index 0000000..455c747 --- /dev/null +++ b/field/fe_generic.go @@ -0,0 +1,303 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519/field +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package field + +import "math/bits" + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// uint128 holds a 128-bit number as two 64-bit limbs, for use with the +// bits.Mul64 and bits.Add64 intrinsics. +type uint128 struct { + lo, hi uint64 +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// mul returns a * b. +func mul(a, b uint64) uint128 { + hi, lo := bits.Mul64(a, b) + return uint128{lo, hi} +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// addMul returns v + a * b. +func addMul(v uint128, a, b uint64) uint128 { + hi, lo := bits.Mul64(a, b) + lo, c := bits.Add64(lo, v.lo, 0) + hi, _ = bits.Add64(hi, v.hi, c) + return uint128{lo, hi} +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// mul19 returns v * 19. +func mul19(v uint64) uint64 { + // Using this approach seems to yield better optimizations than *19. + return v + (v+v<<3)<<1 +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// addMul19 returns v + 19 * a * b, where a and b are at most 52 bits. +func addMul19(v uint128, a, b uint64) uint128 { + hi, lo := bits.Mul64(mul19(a), b) + lo, c := bits.Add64(lo, v.lo, 0) + hi, _ = bits.Add64(hi, v.hi, c) + return uint128{lo, hi} +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// addMul38 returns v + 38 * a * b, where a and b are at most 52 bits. +func addMul38(v uint128, a, b uint64) uint128 { + hi, lo := bits.Mul64(mul19(a), b*2) + lo, c := bits.Add64(lo, v.lo, 0) + hi, _ = bits.Add64(hi, v.hi, c) + return uint128{lo, hi} +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// shiftRightBy51 returns a >> 51. a is assumed to be at most 115 bits. +func shiftRightBy51(a uint128) uint64 { + return (a.hi << (64 - 51)) | (a.lo >> 51) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func feMulGeneric(v, a, b *Element) { + a0 := a.l0 + a1 := a.l1 + a2 := a.l2 + a3 := a.l3 + a4 := a.l4 + + b0 := b.l0 + b1 := b.l1 + b2 := b.l2 + b3 := b.l3 + b4 := b.l4 + + // Limb multiplication works like pen-and-paper columnar multiplication, but + // with 51-bit limbs instead of digits. + // + // a4 a3 a2 a1 a0 x + // b4 b3 b2 b1 b0 = + // ------------------------ + // a4b0 a3b0 a2b0 a1b0 a0b0 + + // a4b1 a3b1 a2b1 a1b1 a0b1 + + // a4b2 a3b2 a2b2 a1b2 a0b2 + + // a4b3 a3b3 a2b3 a1b3 a0b3 + + // a4b4 a3b4 a2b4 a1b4 a0b4 = + // ---------------------------------------------- + // r8 r7 r6 r5 r4 r3 r2 r1 r0 + // + // We can then use the reduction identity (a * 2²⁵⁵ + b = a * 19 + b) to + // reduce the limbs that would overflow 255 bits. r5 * 2²⁵⁵ becomes 19 * r5, + // r6 * 2³⁰⁶ becomes 19 * r6 * 2⁵¹, etc. + // + // Reduction can be carried out simultaneously to multiplication. For + // example, we do not compute r5: whenever the result of a multiplication + // belongs to r5, like a1b4, we multiply it by 19 and add the result to r0. + // + // a4b0 a3b0 a2b0 a1b0 a0b0 + + // a3b1 a2b1 a1b1 a0b1 19×a4b1 + + // a2b2 a1b2 a0b2 19×a4b2 19×a3b2 + + // a1b3 a0b3 19×a4b3 19×a3b3 19×a2b3 + + // a0b4 19×a4b4 19×a3b4 19×a2b4 19×a1b4 = + // -------------------------------------- + // r4 r3 r2 r1 r0 + // + // Finally we add up the columns into wide, overlapping limbs. + + // r0 = a0×b0 + 19×(a1×b4 + a2×b3 + a3×b2 + a4×b1) + r0 := mul(a0, b0) + r0 = addMul19(r0, a1, b4) + r0 = addMul19(r0, a2, b3) + r0 = addMul19(r0, a3, b2) + r0 = addMul19(r0, a4, b1) + + // r1 = a0×b1 + a1×b0 + 19×(a2×b4 + a3×b3 + a4×b2) + r1 := mul(a0, b1) + r1 = addMul(r1, a1, b0) + r1 = addMul19(r1, a2, b4) + r1 = addMul19(r1, a3, b3) + r1 = addMul19(r1, a4, b2) + + // r2 = a0×b2 + a1×b1 + a2×b0 + 19×(a3×b4 + a4×b3) + r2 := mul(a0, b2) + r2 = addMul(r2, a1, b1) + r2 = addMul(r2, a2, b0) + r2 = addMul19(r2, a3, b4) + r2 = addMul19(r2, a4, b3) + + // r3 = a0×b3 + a1×b2 + a2×b1 + a3×b0 + 19×a4×b4 + r3 := mul(a0, b3) + r3 = addMul(r3, a1, b2) + r3 = addMul(r3, a2, b1) + r3 = addMul(r3, a3, b0) + r3 = addMul19(r3, a4, b4) + + // r4 = a0×b4 + a1×b3 + a2×b2 + a3×b1 + a4×b0 + r4 := mul(a0, b4) + r4 = addMul(r4, a1, b3) + r4 = addMul(r4, a2, b2) + r4 = addMul(r4, a3, b1) + r4 = addMul(r4, a4, b0) + + // After the multiplication, we need to reduce (carry) the five coefficients + // to obtain a result with limbs that are at most slightly larger than 2⁵¹, + // to respect the Element invariant. + // + // Overall, the reduction works the same as carryPropagate, except with + // wider inputs: we take the carry for each coefficient by shifting it right + // by 51, and add it to the limb above it. The top carry is multiplied by 19 + // according to the reduction identity and added to the lowest limb. + // + // The largest coefficient (r0) will be at most 111 bits, which guarantees + // that all carries are at most 111 - 51 = 60 bits, which fits in a uint64. + // + // r0 = a0×b0 + 19×(a1×b4 + a2×b3 + a3×b2 + a4×b1) + // r0 < 2⁵²×2⁵² + 19×(2⁵²×2⁵² + 2⁵²×2⁵² + 2⁵²×2⁵² + 2⁵²×2⁵²) + // r0 < (1 + 19 × 4) × 2⁵² × 2⁵² + // r0 < 2⁷ × 2⁵² × 2⁵² + // r0 < 2¹¹¹ + // + // Moreover, the top coefficient (r4) is at most 107 bits, so c4 is at most + // 56 bits, and c4 * 19 is at most 61 bits, which again fits in a uint64 and + // allows us to easily apply the reduction identity. + // + // r4 = a0×b4 + a1×b3 + a2×b2 + a3×b1 + a4×b0 + // r4 < 5 × 2⁵² × 2⁵² + // r4 < 2¹⁰⁷ + // + + c0 := shiftRightBy51(r0) + c1 := shiftRightBy51(r1) + c2 := shiftRightBy51(r2) + c3 := shiftRightBy51(r3) + c4 := shiftRightBy51(r4) + + rr0 := r0.lo&maskLow51Bits + mul19(c4) + rr1 := r1.lo&maskLow51Bits + c0 + rr2 := r2.lo&maskLow51Bits + c1 + rr3 := r3.lo&maskLow51Bits + c2 + rr4 := r4.lo&maskLow51Bits + c3 + + // Now all coefficients fit into 64-bit registers but are still too large to + // be passed around as an Element. We therefore do one last carry chain, + // where the carries will be small enough to fit in the wiggle room above 2⁵¹. + + v.l0 = rr0&maskLow51Bits + mul19(rr4>>51) + v.l1 = rr1&maskLow51Bits + rr0>>51 + v.l2 = rr2&maskLow51Bits + rr1>>51 + v.l3 = rr3&maskLow51Bits + rr2>>51 + v.l4 = rr4&maskLow51Bits + rr3>>51 +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func feSquareGeneric(v, a *Element) { + l0 := a.l0 + l1 := a.l1 + l2 := a.l2 + l3 := a.l3 + l4 := a.l4 + + // Squaring works precisely like multiplication above, but thanks to its + // symmetry we get to group a few terms together. + // + // l4 l3 l2 l1 l0 x + // l4 l3 l2 l1 l0 = + // ------------------------ + // l4l0 l3l0 l2l0 l1l0 l0l0 + + // l4l1 l3l1 l2l1 l1l1 l0l1 + + // l4l2 l3l2 l2l2 l1l2 l0l2 + + // l4l3 l3l3 l2l3 l1l3 l0l3 + + // l4l4 l3l4 l2l4 l1l4 l0l4 = + // ---------------------------------------------- + // r8 r7 r6 r5 r4 r3 r2 r1 r0 + // + // l4l0 l3l0 l2l0 l1l0 l0l0 + + // l3l1 l2l1 l1l1 l0l1 19×l4l1 + + // l2l2 l1l2 l0l2 19×l4l2 19×l3l2 + + // l1l3 l0l3 19×l4l3 19×l3l3 19×l2l3 + + // l0l4 19×l4l4 19×l3l4 19×l2l4 19×l1l4 = + // -------------------------------------- + // r4 r3 r2 r1 r0 + + // r0 = l0×l0 + 19×(l1×l4 + l2×l3 + l3×l2 + l4×l1) = l0×l0 + 19×2×(l1×l4 + l2×l3) + r0 := mul(l0, l0) + r0 = addMul38(r0, l1, l4) + r0 = addMul38(r0, l2, l3) + + // r1 = l0×l1 + l1×l0 + 19×(l2×l4 + l3×l3 + l4×l2) = 2×l0×l1 + 19×2×l2×l4 + 19×l3×l3 + r1 := mul(l0*2, l1) + r1 = addMul38(r1, l2, l4) + r1 = addMul19(r1, l3, l3) + + // r2 = l0×l2 + l1×l1 + l2×l0 + 19×(l3×l4 + l4×l3) = 2×l0×l2 + l1×l1 + 19×2×l3×l4 + r2 := mul(l0*2, l2) + r2 = addMul(r2, l1, l1) + r2 = addMul38(r2, l3, l4) + + // r3 = l0×l3 + l1×l2 + l2×l1 + l3×l0 + 19×l4×l4 = 2×l0×l3 + 2×l1×l2 + 19×l4×l4 + r3 := mul(l0*2, l3) + r3 = addMul(r3, l1*2, l2) + r3 = addMul19(r3, l4, l4) + + // r4 = l0×l4 + l1×l3 + l2×l2 + l3×l1 + l4×l0 = 2×l0×l4 + 2×l1×l3 + l2×l2 + r4 := mul(l0*2, l4) + r4 = addMul(r4, l1*2, l3) + r4 = addMul(r4, l2, l2) + + c0 := shiftRightBy51(r0) + c1 := shiftRightBy51(r1) + c2 := shiftRightBy51(r2) + c3 := shiftRightBy51(r3) + c4 := shiftRightBy51(r4) + + rr0 := r0.lo&maskLow51Bits + mul19(c4) + rr1 := r1.lo&maskLow51Bits + c0 + rr2 := r2.lo&maskLow51Bits + c1 + rr3 := r3.lo&maskLow51Bits + c2 + rr4 := r4.lo&maskLow51Bits + c3 + + v.l0 = rr0&maskLow51Bits + mul19(rr4>>51) + v.l1 = rr1&maskLow51Bits + rr0>>51 + v.l2 = rr2&maskLow51Bits + rr1>>51 + v.l3 = rr3&maskLow51Bits + rr2>>51 + v.l4 = rr4&maskLow51Bits + rr3>>51 +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// carryPropagate brings the limbs below 52 bits by applying the reduction +// identity (a * 2²⁵⁵ + b = a * 19 + b) to the l4 carry. +func (v *Element) carryPropagate() *Element { + // (l4>>51) is at most 64 - 51 = 13 bits, so (l4>>51)*19 is at most 18 bits, and + // the final l0 will be at most 52 bits. Similarly for the rest. + l0 := v.l0 + v.l0 = v.l0&maskLow51Bits + mul19(v.l4>>51) + v.l4 = v.l4&maskLow51Bits + v.l3>>51 + v.l3 = v.l3&maskLow51Bits + v.l2>>51 + v.l2 = v.l2&maskLow51Bits + v.l1>>51 + v.l1 = v.l1&maskLow51Bits + l0>>51 + + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/field/fe_test.go b/field/fe_test.go new file mode 100644 index 0000000..202b5ce --- /dev/null +++ b/field/fe_test.go @@ -0,0 +1,607 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519/field +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package field + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +import ( + "bytes" + "crypto/rand" + "encoding/hex" + "io" + "math/big" + "math/bits" + mathrand "math/rand" + "reflect" + "testing" + "testing/quick" +) + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v Element) String() string { + return hex.EncodeToString(v.Bytes()) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// quickCheckConfig returns a quick.Config that scales the max count by the +// given factor if the -short flag is not set. +func quickCheckConfig(slowScale int) *quick.Config { + cfg := new(quick.Config) + if !testing.Short() { + cfg.MaxCountScale = float64(slowScale) + } + return cfg +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func generateFieldElement(rand *mathrand.Rand) Element { + const maskLow52Bits = (1 << 52) - 1 + return Element{ + rand.Uint64() & maskLow52Bits, + rand.Uint64() & maskLow52Bits, + rand.Uint64() & maskLow52Bits, + rand.Uint64() & maskLow52Bits, + rand.Uint64() & maskLow52Bits, + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// weirdLimbs can be combined to generate a range of edge-case field elements. +// 0 and -1 are intentionally more weighted, as they combine well. +var ( + weirdLimbs51 = []uint64{ + 0, 0, 0, 0, + 1, + 19 - 1, + 19, + 0x2aaaaaaaaaaaa, + 0x5555555555555, + (1 << 51) - 20, + (1 << 51) - 19, + (1 << 51) - 1, (1 << 51) - 1, + (1 << 51) - 1, (1 << 51) - 1, + } + weirdLimbs52 = []uint64{ + 0, 0, 0, 0, 0, 0, + 1, + 19 - 1, + 19, + 0x2aaaaaaaaaaaa, + 0x5555555555555, + (1 << 51) - 20, + (1 << 51) - 19, + (1 << 51) - 1, (1 << 51) - 1, + (1 << 51) - 1, (1 << 51) - 1, + (1 << 51) - 1, (1 << 51) - 1, + 1 << 51, + (1 << 51) + 1, + (1 << 52) - 19, + (1 << 52) - 1, + } +) + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func generateWeirdFieldElement(rand *mathrand.Rand) Element { + return Element{ + weirdLimbs52[rand.Intn(len(weirdLimbs52))], + weirdLimbs51[rand.Intn(len(weirdLimbs51))], + weirdLimbs51[rand.Intn(len(weirdLimbs51))], + weirdLimbs51[rand.Intn(len(weirdLimbs51))], + weirdLimbs51[rand.Intn(len(weirdLimbs51))], + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (Element) Generate(rand *mathrand.Rand, size int) reflect.Value { + if rand.Intn(2) == 0 { + return reflect.ValueOf(generateWeirdFieldElement(rand)) + } + return reflect.ValueOf(generateFieldElement(rand)) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// isInBounds returns whether the element is within the expected bit size bounds +// after a light reduction. +func isInBounds(x *Element) bool { + return bits.Len64(x.l0) <= 52 && + bits.Len64(x.l1) <= 52 && + bits.Len64(x.l2) <= 52 && + bits.Len64(x.l3) <= 52 && + bits.Len64(x.l4) <= 52 +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestMultiplyDistributesOverAdd(t *testing.T) { + multiplyDistributesOverAdd := func(x, y, z Element) bool { + // Compute t1 = (x+y)*z + t1 := new(Element) + t1.Add(&x, &y) + t1.Multiply(t1, &z) + + // Compute t2 = x*z + y*z + t2 := new(Element) + t3 := new(Element) + t2.Multiply(&x, &z) + t3.Multiply(&y, &z) + t2.Add(t2, t3) + + return t1.Equal(t2) == 1 && isInBounds(t1) && isInBounds(t2) + } + + if err := quick.Check(multiplyDistributesOverAdd, quickCheckConfig(1024)); err != nil { + t.Error(err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestMul64to128(t *testing.T) { + a := uint64(5) + b := uint64(5) + r := mul(a, b) + if r.lo != 0x19 || r.hi != 0 { + t.Errorf("lo-range wide mult failed, got %d + %d*(2**64)", r.lo, r.hi) + } + + a = uint64(18014398509481983) // 2^54 - 1 + b = uint64(18014398509481983) // 2^54 - 1 + r = mul(a, b) + if r.lo != 0xff80000000000001 || r.hi != 0xfffffffffff { + t.Errorf("hi-range wide mult failed, got %d + %d*(2**64)", r.lo, r.hi) + } + + a = uint64(1125899906842661) + b = uint64(2097155) + r = mul(a, b) + r = addMul(r, a, b) + r = addMul(r, a, b) + r = addMul(r, a, b) + r = addMul(r, a, b) + if r.lo != 16888498990613035 || r.hi != 640 { + t.Errorf("wrong answer: %d + %d*(2**64)", r.lo, r.hi) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestSetBytesRoundTrip(t *testing.T) { + f1 := func(in [32]byte, fe Element) bool { + fe.SetBytes(in[:]) + + // Mask the most significant bit as it's ignored by SetBytes. (Now + // instead of earlier so we check the masking in SetBytes is working.) + in[len(in)-1] &= (1 << 7) - 1 + + return bytes.Equal(in[:], fe.Bytes()) && isInBounds(&fe) + } + if err := quick.Check(f1, nil); err != nil { + t.Errorf("failed bytes->FE->bytes round-trip: %v", err) + } + + f2 := func(fe, r Element) bool { + r.SetBytes(fe.Bytes()) + + // Intentionally not using Equal not to go through Bytes again. + // Calling reduce because both Generate and SetBytes can produce + // non-canonical representations. + fe.reduce() + r.reduce() + return fe == r + } + if err := quick.Check(f2, nil); err != nil { + t.Errorf("failed FE->bytes->FE round-trip: %v", err) + } + + // Check some fixed vectors from dalek + type feRTTest struct { + fe Element + b []byte + } + var tests = []feRTTest{ + { + fe: Element{358744748052810, 1691584618240980, 977650209285361, 1429865912637724, 560044844278676}, + b: []byte{74, 209, 69, 197, 70, 70, 161, 222, 56, 226, 229, 19, 112, 60, 25, 92, 187, 74, 222, 56, 50, 153, 51, 233, 40, 74, 57, 6, 160, 185, 213, 31}, + }, + { + fe: Element{84926274344903, 473620666599931, 365590438845504, 1028470286882429, 2146499180330972}, + b: []byte{199, 23, 106, 112, 61, 77, 216, 79, 186, 60, 11, 118, 13, 16, 103, 15, 42, 32, 83, 250, 44, 57, 204, 198, 78, 199, 253, 119, 146, 172, 3, 122}, + }, + } + + for _, tt := range tests { + b := tt.fe.Bytes() + fe, _ := new(Element).SetBytes(tt.b) + if !bytes.Equal(b, tt.b) || fe.Equal(&tt.fe) != 1 { + t.Errorf("Failed fixed roundtrip: %v", tt) + } + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func swapEndianness(buf []byte) []byte { + for i := 0; i < len(buf)/2; i++ { + buf[i], buf[len(buf)-i-1] = buf[len(buf)-i-1], buf[i] + } + return buf +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestBytesBigEquivalence(t *testing.T) { + f1 := func(in [32]byte, fe, fe1 Element) bool { + fe.SetBytes(in[:]) + + in[len(in)-1] &= (1 << 7) - 1 // mask the most significant bit + b := new(big.Int).SetBytes(swapEndianness(in[:])) + fe1.fromBig(b) + + if fe != fe1 { + return false + } + + buf := make([]byte, 32) + buf = swapEndianness(fe1.toBig().FillBytes(buf)) + + return bytes.Equal(fe.Bytes(), buf) && isInBounds(&fe) && isInBounds(&fe1) + } + if err := quick.Check(f1, nil); err != nil { + t.Error(err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// fromBig sets v = n, and returns v. The bit length of n must not exceed 256. +func (v *Element) fromBig(n *big.Int) *Element { + if n.BitLen() > 32*8 { + panic("edwards25519: invalid field element input size") + } + + buf := make([]byte, 0, 32) + for _, word := range n.Bits() { + for i := 0; i < bits.UintSize; i += 8 { + if len(buf) >= cap(buf) { + break + } + buf = append(buf, byte(word)) + word >>= 8 + } + } + + v.SetBytes(buf[:32]) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (v *Element) fromDecimal(s string) *Element { + n, ok := new(big.Int).SetString(s, 10) + if !ok { + panic("not a valid decimal: " + s) + } + return v.fromBig(n) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// toBig returns v as a big.Int. +func (v *Element) toBig() *big.Int { + buf := v.Bytes() + + words := make([]big.Word, 32*8/bits.UintSize) + for n := range words { + for i := 0; i < bits.UintSize; i += 8 { + if len(buf) == 0 { + break + } + words[n] |= big.Word(buf[0]) << big.Word(i) + buf = buf[1:] + } + } + + return new(big.Int).SetBits(words) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestDecimalConstants(t *testing.T) { + sqrtM1String := "19681161376707505956807079304988542015446066515923890162744021073123829784752" + if exp := new(Element).fromDecimal(sqrtM1String); sqrtM1.Equal(exp) != 1 { + t.Errorf("sqrtM1 is %v, expected %v", sqrtM1, exp) + } + // d is in the parent package, and we don't want to expose d or fromDecimal. + // dString := "37095705934669439343138083508754565189542113879843219016388785533085940283555" + // if exp := new(Element).fromDecimal(dString); d.Equal(exp) != 1 { + // t.Errorf("d is %v, expected %v", d, exp) + // } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestSetBytesRoundTripEdgeCases(t *testing.T) { + // TODO: values close to 0, close to 2^255-19, between 2^255-19 and 2^255-1, + // and between 2^255 and 2^256-1. Test both the documented SetBytes + // behavior, and that Bytes reduces them. +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Tests self-consistency between Multiply and Square. +func TestConsistency(t *testing.T) { + var x Element + var x2, x2sq Element + + x = Element{1, 1, 1, 1, 1} + x2.Multiply(&x, &x) + x2sq.Square(&x) + + if x2 != x2sq { + t.Fatalf("all ones failed\nmul: %x\nsqr: %x\n", x2, x2sq) + } + + var bytes [32]byte + + _, err := io.ReadFull(rand.Reader, bytes[:]) + if err != nil { + t.Fatal(err) + } + x.SetBytes(bytes[:]) + + x2.Multiply(&x, &x) + x2sq.Square(&x) + + if x2 != x2sq { + t.Fatalf("all ones failed\nmul: %x\nsqr: %x\n", x2, x2sq) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestEqual(t *testing.T) { + x := Element{1, 1, 1, 1, 1} + y := Element{5, 4, 3, 2, 1} + + eq := x.Equal(&x) + if eq != 1 { + t.Errorf("wrong about equality") + } + + eq = x.Equal(&y) + if eq != 0 { + t.Errorf("wrong about inequality") + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestInvert(t *testing.T) { + x := Element{1, 1, 1, 1, 1} + one := Element{1, 0, 0, 0, 0} + var xinv, r Element + + xinv.Invert(&x) + r.Multiply(&x, &xinv) + r.reduce() + + if one != r { + t.Errorf("inversion identity failed, got: %x", r) + } + + var bytes [32]byte + + _, err := io.ReadFull(rand.Reader, bytes[:]) + if err != nil { + t.Fatal(err) + } + x.SetBytes(bytes[:]) + + xinv.Invert(&x) + r.Multiply(&x, &xinv) + r.reduce() + + if one != r { + t.Errorf("random inversion identity failed, got: %x for field element %x", r, x) + } + + zero := Element{} + x.Set(&zero) + if xx := xinv.Invert(&x); xx != &xinv { + t.Errorf("inverting zero did not return the receiver") + } else if xinv.Equal(&zero) != 1 { + t.Errorf("inverting zero did not return zero") + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestSelectSwap(t *testing.T) { + a := Element{358744748052810, 1691584618240980, 977650209285361, 1429865912637724, 560044844278676} + b := Element{84926274344903, 473620666599931, 365590438845504, 1028470286882429, 2146499180330972} + + var c, d Element + + c.Select(&a, &b, 1) + d.Select(&a, &b, 0) + + if c.Equal(&a) != 1 || d.Equal(&b) != 1 { + t.Errorf("Select failed") + } + + c.Swap(&d, 0) + + if c.Equal(&a) != 1 || d.Equal(&b) != 1 { + t.Errorf("Swap failed") + } + + c.Swap(&d, 1) + + if c.Equal(&b) != 1 || d.Equal(&a) != 1 { + t.Errorf("Swap failed") + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestMult32(t *testing.T) { + mult32EquivalentToMul := func(x Element, y uint32) bool { + t1 := new(Element) + for i := 0; i < 100; i++ { + t1.Mult32(&x, y) + } + + ty := new(Element) + ty.l0 = uint64(y) + + t2 := new(Element) + for i := 0; i < 100; i++ { + t2.Multiply(&x, ty) + } + + return t1.Equal(t2) == 1 && isInBounds(t1) && isInBounds(t2) + } + + if err := quick.Check(mult32EquivalentToMul, quickCheckConfig(1024)); err != nil { + t.Error(err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestSqrtRatio(t *testing.T) { + // From draft-irtf-cfrg-ristretto255-decaf448-00, Appendix A.4. + type test struct { + u, v []byte + wasSquare int + r []byte + } + var tests = []test{ + // If u is 0, the function is defined to return (0, TRUE), even if v + // is zero. Note that where used in this package, the denominator v + // is never zero. + { + decodeHex("0000000000000000000000000000000000000000000000000000000000000000"), + decodeHex("0000000000000000000000000000000000000000000000000000000000000000"), + 1, decodeHex("0000000000000000000000000000000000000000000000000000000000000000"), + }, + // 0/1 == 0² + { + decodeHex("0000000000000000000000000000000000000000000000000000000000000000"), + decodeHex("0100000000000000000000000000000000000000000000000000000000000000"), + 1, decodeHex("0000000000000000000000000000000000000000000000000000000000000000"), + }, + // If u is non-zero and v is zero, defined to return (0, FALSE). + { + decodeHex("0100000000000000000000000000000000000000000000000000000000000000"), + decodeHex("0000000000000000000000000000000000000000000000000000000000000000"), + 0, decodeHex("0000000000000000000000000000000000000000000000000000000000000000"), + }, + // 2/1 is not square in this field. + { + decodeHex("0200000000000000000000000000000000000000000000000000000000000000"), + decodeHex("0100000000000000000000000000000000000000000000000000000000000000"), + 0, decodeHex("3c5ff1b5d8e4113b871bd052f9e7bcd0582804c266ffb2d4f4203eb07fdb7c54"), + }, + // 4/1 == 2² + { + decodeHex("0400000000000000000000000000000000000000000000000000000000000000"), + decodeHex("0100000000000000000000000000000000000000000000000000000000000000"), + 1, decodeHex("0200000000000000000000000000000000000000000000000000000000000000"), + }, + // 1/4 == (2⁻¹)² == (2^(p-2))² per Euler's theorem + { + decodeHex("0100000000000000000000000000000000000000000000000000000000000000"), + decodeHex("0400000000000000000000000000000000000000000000000000000000000000"), + 1, decodeHex("f6ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff3f"), + }, + } + + for i, tt := range tests { + u, _ := new(Element).SetBytes(tt.u) + v, _ := new(Element).SetBytes(tt.v) + want, _ := new(Element).SetBytes(tt.r) + got, wasSquare := new(Element).SqrtRatio(u, v) + if got.Equal(want) == 0 || wasSquare != tt.wasSquare { + t.Errorf("%d: got (%v, %v), want (%v, %v)", i, got, wasSquare, want, tt.wasSquare) + } + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestFeSquare(t *testing.T) { + asmLikeGeneric := func(a Element) bool { + t1 := a + t2 := a + + feSquareGeneric(&t1, &t1) + feSquare(&t2, &t2) + + if t1 != t2 { + t.Logf("got: %#v,\nexpected: %#v", t1, t2) + } + + return t1 == t2 && isInBounds(&t2) + } + + if err := quick.Check(asmLikeGeneric, quickCheckConfig(1024)); err != nil { + t.Error(err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestFeMul(t *testing.T) { + asmLikeGeneric := func(a, b Element) bool { + a1 := a + a2 := a + b1 := b + b2 := b + + feMulGeneric(&a1, &a1, &b1) + feMul(&a2, &a2, &b2) + + if a1 != a2 || b1 != b2 { + t.Logf("got: %#v,\nexpected: %#v", a1, a2) + t.Logf("got: %#v,\nexpected: %#v", b1, b2) + } + + return a1 == a2 && isInBounds(&a2) && + b1 == b2 && isInBounds(&b2) + } + + if err := quick.Check(asmLikeGeneric, quickCheckConfig(1024)); err != nil { + t.Error(err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func decodeHex(s string) []byte { + b, err := hex.DecodeString(s) + if err != nil { + panic(err) + } + return b +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/go.mod b/go.mod new file mode 100644 index 0000000..665921c --- /dev/null +++ b/go.mod @@ -0,0 +1,3 @@ +module sources.truenas.cloud/code/edwards25519 + +go 1.26 diff --git a/go.sum b/go.sum new file mode 100644 index 0000000..e69de29 diff --git a/pull.sh b/pull.sh new file mode 100644 index 0000000..a6e6998 --- /dev/null +++ b/pull.sh @@ -0,0 +1,53 @@ +#!/usr/bin/env bash +set -euo pipefail + +if [ "$#" -ne 1 ]; then + echo "Usage: $0 " + exit 1 +fi + +TAG="$1" +TMPDIR="$(mktemp -d)" + +cleanup() { + rm -rf "$TMPDIR" +} +trap cleanup EXIT + +command -v git >/dev/null +command -v git-filter-repo >/dev/null + +if [ -d "$HOME/go/.git" ]; then + REFERENCE=(--reference "$HOME/go" --dissociate) +else + REFERENCE=() +fi + +git -c advice.detachedHead=false clone --no-checkout "${REFERENCE[@]}" \ + -b "$TAG" https://go.googlesource.com/go.git "$TMPDIR" + +# Simplify the history graph by removing the dev.boringcrypto branches, whose +# merges end up empty after grafting anyway. This also fixes a weird quirk +# (maybe a git-filter-repo bug?) where only one file from an old path, +# src/crypto/ed25519/internal/edwards25519/const.go, would still exist in the +# filtered repo. +git -C "$TMPDIR" replace --graft f771edd7f9 99f1bf54eb +git -C "$TMPDIR" replace --graft 109c13b64f c2f96e686f +git -C "$TMPDIR" replace --graft aa4da4f189 912f075047 + +git -C "$TMPDIR" filter-repo --force \ + --paths-from-file /dev/stdin \ + --prune-empty always \ + --prune-degenerate always \ + --tag-callback 'tag.skip()' <<'EOF' +src/crypto/internal/fips140/edwards25519 +src/crypto/internal/edwards25519 +src/crypto/ed25519/internal/edwards25519 +EOF + +git fetch "$TMPDIR" +git update-ref "refs/heads/upstream/$TAG" FETCH_HEAD + +echo +echo "Fetched upstream history up to $TAG. Merge with:" +echo -e "\tgit merge --no-ff --no-commit --allow-unrelated-histories upstream/$TAG" \ No newline at end of file diff --git a/scalar.go b/scalar.go new file mode 100644 index 0000000..a629a1b --- /dev/null +++ b/scalar.go @@ -0,0 +1,397 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519 +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package edwards25519 + +import ( + "encoding/binary" + "errors" + "math/bits" +) + +// A Scalar is an integer modulo +// +// l = 2^252 + 27742317777372353535851937790883648493 +// +// which is the prime order of the edwards25519 group. +// +// This type works similarly to math/big.Int, and all arguments and +// receivers are allowed to alias. +// +// The zero value is a valid zero element. +type Scalar struct { + // s is the scalar in the Montgomery domain, in the format of the + // fiat-crypto implementation. + s fiatScalarMontgomeryDomainFieldElement +} + +// The field implementation in scalar_fiat.go is generated by the fiat-crypto +// project (https://github.com/mit-plv/fiat-crypto) at version v0.0.9 (23d2dbc) +// from a formally verified model. +// +// fiat-crypto code comes under the following license. +// +// Copyright (c) 2015-2020 The fiat-crypto Authors. All rights reserved. +// +// Redistribution and use in source and binary forms, with or without +// modification, are permitted provided that the following conditions are +// met: +// +// 1. Redistributions of source code must retain the above copyright +// notice, this list of conditions and the following disclaimer. +// +// THIS SOFTWARE IS PROVIDED BY the fiat-crypto authors "AS IS" +// AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, +// THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR +// PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL Berkeley Software Design, +// Inc. BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, +// EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, +// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR +// PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF +// LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING +// NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS +// SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. +// + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// NewScalar returns a new zero Scalar. +func NewScalar() *Scalar { + return &Scalar{} +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// MultiplyAdd sets s = x * y + z mod l, and returns s. It is equivalent to +// using Multiply and then Add. +func (s *Scalar) MultiplyAdd(x, y, z *Scalar) *Scalar { + // Make a copy of z in case it aliases s. + zCopy := new(Scalar).Set(z) + return s.Multiply(x, y).Add(s, zCopy) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Add sets s = x + y mod l, and returns s. +func (s *Scalar) Add(x, y *Scalar) *Scalar { + // s = 1 * x + y mod l + fiatScalarAdd(&s.s, &x.s, &y.s) + return s +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Subtract sets s = x - y mod l, and returns s. +func (s *Scalar) Subtract(x, y *Scalar) *Scalar { + // s = -1 * y + x mod l + fiatScalarSub(&s.s, &x.s, &y.s) + return s +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Negate sets s = -x mod l, and returns s. +func (s *Scalar) Negate(x *Scalar) *Scalar { + // s = -1 * x + 0 mod l + fiatScalarOpp(&s.s, &x.s) + return s +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Multiply sets s = x * y mod l, and returns s. +func (s *Scalar) Multiply(x, y *Scalar) *Scalar { + // s = x * y + 0 mod l + fiatScalarMul(&s.s, &x.s, &y.s) + return s +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Set sets s = x, and returns s. +func (s *Scalar) Set(x *Scalar) *Scalar { + *s = *x + return s +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// SetUniformBytes sets s = x mod l, where x is a 64-byte little-endian integer. +// If x is not of the right length, SetUniformBytes returns nil and an error, +// and the receiver is unchanged. +// +// SetUniformBytes can be used to set s to a uniformly distributed value given +// 64 uniformly distributed random bytes. +func (s *Scalar) SetUniformBytes(x []byte) (*Scalar, error) { + if len(x) != 64 { + return nil, errors.New("edwards25519: invalid SetUniformBytes input length") + } + + // We have a value x of 512 bits, but our fiatScalarFromBytes function + // expects an input lower than l, which is a little over 252 bits. + // + // Instead of writing a reduction function that operates on wider inputs, we + // can interpret x as the sum of three shorter values a, b, and c. + // + // x = a + b * 2^168 + c * 2^336 mod l + // + // We then precompute 2^168 and 2^336 modulo l, and perform the reduction + // with two multiplications and two additions. + + s.setShortBytes(x[:21]) + t := new(Scalar).setShortBytes(x[21:42]) + s.Add(s, t.Multiply(t, scalarTwo168)) + t.setShortBytes(x[42:]) + s.Add(s, t.Multiply(t, scalarTwo336)) + + return s, nil +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// scalarTwo168 and scalarTwo336 are 2^168 and 2^336 modulo l, encoded as a +// fiatScalarMontgomeryDomainFieldElement, which is a little-endian 4-limb value +// in the 2^256 Montgomery domain. +var scalarTwo168 = &Scalar{s: [4]uint64{0x5b8ab432eac74798, 0x38afddd6de59d5d7, + 0xa2c131b399411b7c, 0x6329a7ed9ce5a30}} +var scalarTwo336 = &Scalar{s: [4]uint64{0xbd3d108e2b35ecc5, 0x5c3a3718bdf9c90b, + 0x63aa97a331b4f2ee, 0x3d217f5be65cb5c}} + +// setShortBytes sets s = x mod l, where x is a little-endian integer shorter +// than 32 bytes. +func (s *Scalar) setShortBytes(x []byte) *Scalar { + if len(x) >= 32 { + panic("edwards25519: internal error: setShortBytes called with a long string") + } + var buf [32]byte + copy(buf[:], x) + fiatScalarFromBytes((*[4]uint64)(&s.s), &buf) + fiatScalarToMontgomery(&s.s, (*fiatScalarNonMontgomeryDomainFieldElement)(&s.s)) + return s +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// SetCanonicalBytes sets s = x, where x is a 32-byte little-endian encoding of +// s, and returns s. If x is not a canonical encoding of s, SetCanonicalBytes +// returns nil and an error, and the receiver is unchanged. +func (s *Scalar) SetCanonicalBytes(x []byte) (*Scalar, error) { + if len(x) != 32 { + return nil, errors.New("invalid scalar length") + } + if !isReduced(x) { + return nil, errors.New("invalid scalar encoding") + } + + fiatScalarFromBytes((*[4]uint64)(&s.s), (*[32]byte)(x)) + fiatScalarToMontgomery(&s.s, (*fiatScalarNonMontgomeryDomainFieldElement)(&s.s)) + + return s, nil +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// scalarMinusOneBytes is l - 1 in little endian. +var scalarMinusOneBytes = [32]byte{236, 211, 245, 92, 26, 99, 18, 88, 214, 156, 247, 162, 222, 249, 222, 20, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 16} + +// isReduced returns whether the given scalar in 32-byte little endian encoded +// form is reduced modulo l. +func isReduced(s []byte) bool { + if len(s) != 32 { + return false + } + + s0 := binary.LittleEndian.Uint64(s[:8]) + s1 := binary.LittleEndian.Uint64(s[8:16]) + s2 := binary.LittleEndian.Uint64(s[16:24]) + s3 := binary.LittleEndian.Uint64(s[24:]) + + l0 := binary.LittleEndian.Uint64(scalarMinusOneBytes[:8]) + l1 := binary.LittleEndian.Uint64(scalarMinusOneBytes[8:16]) + l2 := binary.LittleEndian.Uint64(scalarMinusOneBytes[16:24]) + l3 := binary.LittleEndian.Uint64(scalarMinusOneBytes[24:]) + + // Do a constant time subtraction chain scalarMinusOneBytes - s. If there is + // a borrow at the end, then s > scalarMinusOneBytes. + _, b := bits.Sub64(l0, s0, 0) + _, b = bits.Sub64(l1, s1, b) + _, b = bits.Sub64(l2, s2, b) + _, b = bits.Sub64(l3, s3, b) + return b == 0 +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// SetBytesWithClamping applies the buffer pruning described in RFC 8032, +// Section 5.1.5 (also known as clamping) and sets s to the result. The input +// must be 32 bytes, and it is not modified. If x is not of the right length, +// SetBytesWithClamping returns nil and an error, and the receiver is unchanged. +// +// Note that since Scalar values are always reduced modulo the prime order of +// the curve, the resulting value will not preserve any of the cofactor-clearing +// properties that clamping is meant to provide. It will however work as +// expected as long as it is applied to points on the prime order subgroup, like +// in Ed25519. In fact, it is lost to history why RFC 8032 adopted the +// irrelevant RFC 7748 clamping, but it is now required for compatibility. +func (s *Scalar) SetBytesWithClamping(x []byte) (*Scalar, error) { + // The description above omits the purpose of the high bits of the clamping + // for brevity, but those are also lost to reductions, and are also + // irrelevant to edwards25519 as they protect against a specific + // implementation bug that was once observed in a generic Montgomery ladder. + if len(x) != 32 { + return nil, errors.New("edwards25519: invalid SetBytesWithClamping input length") + } + + // We need to use the wide reduction from SetUniformBytes, since clamping + // sets the 2^254 bit, making the value higher than the order. + var wideBytes [64]byte + copy(wideBytes[:], x[:]) + wideBytes[0] &= 248 + wideBytes[31] &= 63 + wideBytes[31] |= 64 + return s.SetUniformBytes(wideBytes[:]) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Bytes returns the canonical 32-byte little-endian encoding of s. +func (s *Scalar) Bytes() []byte { + // This function is outlined to make the allocations inline in the caller + // rather than happen on the heap. + var encoded [32]byte + return s.bytes(&encoded) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (s *Scalar) bytes(out *[32]byte) []byte { + var ss fiatScalarNonMontgomeryDomainFieldElement + fiatScalarFromMontgomery(&ss, &s.s) + fiatScalarToBytes(out, (*[4]uint64)(&ss)) + return out[:] +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Equal returns 1 if s and t are equal, and 0 otherwise. +func (s *Scalar) Equal(t *Scalar) int { + var diff fiatScalarMontgomeryDomainFieldElement + fiatScalarSub(&diff, &s.s, &t.s) + var nonzero uint64 + fiatScalarNonzero(&nonzero, (*[4]uint64)(&diff)) + nonzero |= nonzero >> 32 + nonzero |= nonzero >> 16 + nonzero |= nonzero >> 8 + nonzero |= nonzero >> 4 + nonzero |= nonzero >> 2 + nonzero |= nonzero >> 1 + return int(^nonzero) & 1 +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// nonAdjacentForm computes a width-w non-adjacent form for this scalar. +// +// w must be between 2 and 8, or nonAdjacentForm will panic. +func (s *Scalar) nonAdjacentForm(w uint) [256]int8 { + // This implementation is adapted from the one + // in curve25519-dalek and is documented there: + // https://github.com/dalek-cryptography/curve25519-dalek/blob/f630041af28e9a405255f98a8a93adca18e4315b/src/scalar.rs#L800-L871 + b := s.Bytes() + if b[31] > 127 { + panic("scalar has high bit set illegally") + } + if w < 2 { + panic("w must be at least 2 by the definition of NAF") + } else if w > 8 { + panic("NAF digits must fit in int8") + } + + var naf [256]int8 + var digits [5]uint64 + + for i := 0; i < 4; i++ { + digits[i] = binary.LittleEndian.Uint64(b[i*8:]) + } + + width := uint64(1 << w) + windowMask := uint64(width - 1) + + pos := uint(0) + carry := uint64(0) + for pos < 256 { + indexU64 := pos / 64 + indexBit := pos % 64 + var bitBuf uint64 + if indexBit < 64-w { + // This window's bits are contained in a single u64 + bitBuf = digits[indexU64] >> indexBit + } else { + // Combine the current 64 bits with bits from the next 64 + bitBuf = (digits[indexU64] >> indexBit) | (digits[1+indexU64] << (64 - indexBit)) + } + + // Add carry into the current window + window := carry + (bitBuf & windowMask) + + if window&1 == 0 { + // If the window value is even, preserve the carry and continue. + // Why is the carry preserved? + // If carry == 0 and window & 1 == 0, + // then the next carry should be 0 + // If carry == 1 and window & 1 == 0, + // then bit_buf & 1 == 1 so the next carry should be 1 + pos += 1 + continue + } + + if window < width/2 { + carry = 0 + naf[pos] = int8(window) + } else { + carry = 1 + naf[pos] = int8(window) - int8(width) + } + + pos += w + } + return naf +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func (s *Scalar) signedRadix16() [64]int8 { + b := s.Bytes() + if b[31] > 127 { + panic("scalar has high bit set illegally") + } + + var digits [64]int8 + + // Compute unsigned radix-16 digits: + for i := 0; i < 32; i++ { + digits[2*i] = int8(b[i] & 15) + digits[2*i+1] = int8((b[i] >> 4) & 15) + } + + // Recenter coefficients: + for i := 0; i < 63; i++ { + carry := (digits[i] + 8) >> 4 + digits[i] -= carry << 4 + digits[i+1] += carry + } + + return digits +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/scalar_alias_test.go b/scalar_alias_test.go new file mode 100644 index 0000000..d432f04 --- /dev/null +++ b/scalar_alias_test.go @@ -0,0 +1,124 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519 +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package edwards25519 + +import ( + "testing" + "testing/quick" +) + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestScalarAliasing(t *testing.T) { + checkAliasingOneArg := func(f func(v, x *Scalar) *Scalar, v, x Scalar) bool { + x1, v1 := x, x + + // Calculate a reference f(x) without aliasing. + if out := f(&v, &x); out != &v || !isReduced(out.Bytes()) { + return false + } + + // Test aliasing the argument and the receiver. + if out := f(&v1, &v1); out != &v1 || v1 != v || !isReduced(out.Bytes()) { + return false + } + + // Ensure the arguments was not modified. + return x == x1 + } + + checkAliasingTwoArgs := func(f func(v, x, y *Scalar) *Scalar, v, x, y Scalar) bool { + x1, y1, v1 := x, y, Scalar{} + + // Calculate a reference f(x, y) without aliasing. + if out := f(&v, &x, &y); out != &v || !isReduced(out.Bytes()) { + return false + } + + // Test aliasing the first argument and the receiver. + v1 = x + if out := f(&v1, &v1, &y); out != &v1 || v1 != v || !isReduced(out.Bytes()) { + return false + } + // Test aliasing the second argument and the receiver. + v1 = y + if out := f(&v1, &x, &v1); out != &v1 || v1 != v || !isReduced(out.Bytes()) { + return false + } + + // Calculate a reference f(x, x) without aliasing. + if out := f(&v, &x, &x); out != &v || !isReduced(out.Bytes()) { + return false + } + + // Test aliasing the first argument and the receiver. + v1 = x + if out := f(&v1, &v1, &x); out != &v1 || v1 != v || !isReduced(out.Bytes()) { + return false + } + // Test aliasing the second argument and the receiver. + v1 = x + if out := f(&v1, &x, &v1); out != &v1 || v1 != v || !isReduced(out.Bytes()) { + return false + } + // Test aliasing both arguments and the receiver. + v1 = x + if out := f(&v1, &v1, &v1); out != &v1 || v1 != v || !isReduced(out.Bytes()) { + return false + } + + // Ensure the arguments were not modified. + return x == x1 && y == y1 + } + + for name, f := range map[string]any{ + "Negate": func(v, x Scalar) bool { + return checkAliasingOneArg((*Scalar).Negate, v, x) + }, + "Invert": func(v, x Scalar) bool { + return checkAliasingOneArg((*Scalar).Invert, v, x) + }, + "Multiply": func(v, x, y Scalar) bool { + return checkAliasingTwoArgs((*Scalar).Multiply, v, x, y) + }, + "Add": func(v, x, y Scalar) bool { + return checkAliasingTwoArgs((*Scalar).Add, v, x, y) + }, + "Subtract": func(v, x, y Scalar) bool { + return checkAliasingTwoArgs((*Scalar).Subtract, v, x, y) + }, + "MultiplyAdd1": func(v, x, y, fixed Scalar) bool { + return checkAliasingTwoArgs(func(v, x, y *Scalar) *Scalar { + return v.MultiplyAdd(&fixed, x, y) + }, v, x, y) + }, + "MultiplyAdd2": func(v, x, y, fixed Scalar) bool { + return checkAliasingTwoArgs(func(v, x, y *Scalar) *Scalar { + return v.MultiplyAdd(x, &fixed, y) + }, v, x, y) + }, + "MultiplyAdd3": func(v, x, y, fixed Scalar) bool { + return checkAliasingTwoArgs(func(v, x, y *Scalar) *Scalar { + return v.MultiplyAdd(x, y, &fixed) + }, v, x, y) + }, + } { + err := quick.Check(f, quickCheckConfig(32)) + if err != nil { + t.Errorf("%v: %v", name, err) + } + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/scalar_fiat.go b/scalar_fiat.go new file mode 100644 index 0000000..f9e14c5 --- /dev/null +++ b/scalar_fiat.go @@ -0,0 +1,1128 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519 +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// STOP!!! Code generated by Fiat Cryptography. DO NOT EDIT. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package edwards25519 + +import "math/bits" + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +type fiatScalarUint1 uint64 // We use uint64 instead of a more narrow type for performance reasons; see https://github.com/mit-plv/fiat-crypto/pull/1006#issuecomment-892625927 +type fiatScalarInt1 int64 // We use uint64 instead of a more narrow type for performance reasons; see https://github.com/mit-plv/fiat-crypto/pull/1006#issuecomment-892625927 + +// The type fiatScalarMontgomeryDomainFieldElement is a field element in the Montgomery domain. +// +// Bounds: [[0x0 ~> 0xffffffffffffffff], [0x0 ~> 0xffffffffffffffff], [0x0 ~> 0xffffffffffffffff], [0x0 ~> 0xffffffffffffffff]] +type fiatScalarMontgomeryDomainFieldElement [4]uint64 + +// The type fiatScalarNonMontgomeryDomainFieldElement is a field element NOT in the Montgomery domain. +// +// Bounds: [[0x0 ~> 0xffffffffffffffff], [0x0 ~> 0xffffffffffffffff], [0x0 ~> 0xffffffffffffffff], [0x0 ~> 0xffffffffffffffff]] +type fiatScalarNonMontgomeryDomainFieldElement [4]uint64 + +// fiatScalarCmovznzU64 is a single-word conditional move. +// +// Postconditions: +// +// out1 = (if arg1 = 0 then arg2 else arg3) +// +// Input Bounds: +// +// arg1: [0x0 ~> 0x1] +// arg2: [0x0 ~> 0xffffffffffffffff] +// arg3: [0x0 ~> 0xffffffffffffffff] +// +// Output Bounds: +// +// out1: [0x0 ~> 0xffffffffffffffff] +func fiatScalarCmovznzU64(out1 *uint64, arg1 fiatScalarUint1, arg2 uint64, arg3 uint64) { + x1 := (uint64(arg1) * 0xffffffffffffffff) + x2 := ((x1 & arg3) | ((^x1) & arg2)) + *out1 = x2 +} + +// fiatScalarMul multiplies two field elements in the Montgomery domain. +// +// Preconditions: +// +// 0 ≤ eval arg1 < m +// 0 ≤ eval arg2 < m +// +// Postconditions: +// +// eval (from_montgomery out1) mod m = (eval (from_montgomery arg1) * eval (from_montgomery arg2)) mod m +// 0 ≤ eval out1 < m +func fiatScalarMul(out1 *fiatScalarMontgomeryDomainFieldElement, arg1 *fiatScalarMontgomeryDomainFieldElement, arg2 *fiatScalarMontgomeryDomainFieldElement) { + x1 := arg1[1] + x2 := arg1[2] + x3 := arg1[3] + x4 := arg1[0] + var x5 uint64 + var x6 uint64 + x6, x5 = bits.Mul64(x4, arg2[3]) + var x7 uint64 + var x8 uint64 + x8, x7 = bits.Mul64(x4, arg2[2]) + var x9 uint64 + var x10 uint64 + x10, x9 = bits.Mul64(x4, arg2[1]) + var x11 uint64 + var x12 uint64 + x12, x11 = bits.Mul64(x4, arg2[0]) + var x13 uint64 + var x14 uint64 + x13, x14 = bits.Add64(x12, x9, uint64(0x0)) + var x15 uint64 + var x16 uint64 + x15, x16 = bits.Add64(x10, x7, uint64(fiatScalarUint1(x14))) + var x17 uint64 + var x18 uint64 + x17, x18 = bits.Add64(x8, x5, uint64(fiatScalarUint1(x16))) + x19 := (uint64(fiatScalarUint1(x18)) + x6) + var x20 uint64 + _, x20 = bits.Mul64(x11, 0xd2b51da312547e1b) + var x22 uint64 + var x23 uint64 + x23, x22 = bits.Mul64(x20, 0x1000000000000000) + var x24 uint64 + var x25 uint64 + x25, x24 = bits.Mul64(x20, 0x14def9dea2f79cd6) + var x26 uint64 + var x27 uint64 + x27, x26 = bits.Mul64(x20, 0x5812631a5cf5d3ed) + var x28 uint64 + var x29 uint64 + x28, x29 = bits.Add64(x27, x24, uint64(0x0)) + x30 := (uint64(fiatScalarUint1(x29)) + x25) + var x32 uint64 + _, x32 = bits.Add64(x11, x26, uint64(0x0)) + var x33 uint64 + var x34 uint64 + x33, x34 = bits.Add64(x13, x28, uint64(fiatScalarUint1(x32))) + var x35 uint64 + var x36 uint64 + x35, x36 = bits.Add64(x15, x30, uint64(fiatScalarUint1(x34))) + var x37 uint64 + var x38 uint64 + x37, x38 = bits.Add64(x17, x22, uint64(fiatScalarUint1(x36))) + var x39 uint64 + var x40 uint64 + x39, x40 = bits.Add64(x19, x23, uint64(fiatScalarUint1(x38))) + var x41 uint64 + var x42 uint64 + x42, x41 = bits.Mul64(x1, arg2[3]) + var x43 uint64 + var x44 uint64 + x44, x43 = bits.Mul64(x1, arg2[2]) + var x45 uint64 + var x46 uint64 + x46, x45 = bits.Mul64(x1, arg2[1]) + var x47 uint64 + var x48 uint64 + x48, x47 = bits.Mul64(x1, arg2[0]) + var x49 uint64 + var x50 uint64 + x49, x50 = bits.Add64(x48, x45, uint64(0x0)) + var x51 uint64 + var x52 uint64 + x51, x52 = bits.Add64(x46, x43, uint64(fiatScalarUint1(x50))) + var x53 uint64 + var x54 uint64 + x53, x54 = bits.Add64(x44, x41, uint64(fiatScalarUint1(x52))) + x55 := (uint64(fiatScalarUint1(x54)) + x42) + var x56 uint64 + var x57 uint64 + x56, x57 = bits.Add64(x33, x47, uint64(0x0)) + var x58 uint64 + var x59 uint64 + x58, x59 = bits.Add64(x35, x49, uint64(fiatScalarUint1(x57))) + var x60 uint64 + var x61 uint64 + x60, x61 = bits.Add64(x37, x51, uint64(fiatScalarUint1(x59))) + var x62 uint64 + var x63 uint64 + x62, x63 = bits.Add64(x39, x53, uint64(fiatScalarUint1(x61))) + var x64 uint64 + var x65 uint64 + x64, x65 = bits.Add64(uint64(fiatScalarUint1(x40)), x55, uint64(fiatScalarUint1(x63))) + var x66 uint64 + _, x66 = bits.Mul64(x56, 0xd2b51da312547e1b) + var x68 uint64 + var x69 uint64 + x69, x68 = bits.Mul64(x66, 0x1000000000000000) + var x70 uint64 + var x71 uint64 + x71, x70 = bits.Mul64(x66, 0x14def9dea2f79cd6) + var x72 uint64 + var x73 uint64 + x73, x72 = bits.Mul64(x66, 0x5812631a5cf5d3ed) + var x74 uint64 + var x75 uint64 + x74, x75 = bits.Add64(x73, x70, uint64(0x0)) + x76 := (uint64(fiatScalarUint1(x75)) + x71) + var x78 uint64 + _, x78 = bits.Add64(x56, x72, uint64(0x0)) + var x79 uint64 + var x80 uint64 + x79, x80 = bits.Add64(x58, x74, uint64(fiatScalarUint1(x78))) + var x81 uint64 + var x82 uint64 + x81, x82 = bits.Add64(x60, x76, uint64(fiatScalarUint1(x80))) + var x83 uint64 + var x84 uint64 + x83, x84 = bits.Add64(x62, x68, uint64(fiatScalarUint1(x82))) + var x85 uint64 + var x86 uint64 + x85, x86 = bits.Add64(x64, x69, uint64(fiatScalarUint1(x84))) + x87 := (uint64(fiatScalarUint1(x86)) + uint64(fiatScalarUint1(x65))) + var x88 uint64 + var x89 uint64 + x89, x88 = bits.Mul64(x2, arg2[3]) + var x90 uint64 + var x91 uint64 + x91, x90 = bits.Mul64(x2, arg2[2]) + var x92 uint64 + var x93 uint64 + x93, x92 = bits.Mul64(x2, arg2[1]) + var x94 uint64 + var x95 uint64 + x95, x94 = bits.Mul64(x2, arg2[0]) + var x96 uint64 + var x97 uint64 + x96, x97 = bits.Add64(x95, x92, uint64(0x0)) + var x98 uint64 + var x99 uint64 + x98, x99 = bits.Add64(x93, x90, uint64(fiatScalarUint1(x97))) + var x100 uint64 + var x101 uint64 + x100, x101 = bits.Add64(x91, x88, uint64(fiatScalarUint1(x99))) + x102 := (uint64(fiatScalarUint1(x101)) + x89) + var x103 uint64 + var x104 uint64 + x103, x104 = bits.Add64(x79, x94, uint64(0x0)) + var x105 uint64 + var x106 uint64 + x105, x106 = bits.Add64(x81, x96, uint64(fiatScalarUint1(x104))) + var x107 uint64 + var x108 uint64 + x107, x108 = bits.Add64(x83, x98, uint64(fiatScalarUint1(x106))) + var x109 uint64 + var x110 uint64 + x109, x110 = bits.Add64(x85, x100, uint64(fiatScalarUint1(x108))) + var x111 uint64 + var x112 uint64 + x111, x112 = bits.Add64(x87, x102, uint64(fiatScalarUint1(x110))) + var x113 uint64 + _, x113 = bits.Mul64(x103, 0xd2b51da312547e1b) + var x115 uint64 + var x116 uint64 + x116, x115 = bits.Mul64(x113, 0x1000000000000000) + var x117 uint64 + var x118 uint64 + x118, x117 = bits.Mul64(x113, 0x14def9dea2f79cd6) + var x119 uint64 + var x120 uint64 + x120, x119 = bits.Mul64(x113, 0x5812631a5cf5d3ed) + var x121 uint64 + var x122 uint64 + x121, x122 = bits.Add64(x120, x117, uint64(0x0)) + x123 := (uint64(fiatScalarUint1(x122)) + x118) + var x125 uint64 + _, x125 = bits.Add64(x103, x119, uint64(0x0)) + var x126 uint64 + var x127 uint64 + x126, x127 = bits.Add64(x105, x121, uint64(fiatScalarUint1(x125))) + var x128 uint64 + var x129 uint64 + x128, x129 = bits.Add64(x107, x123, uint64(fiatScalarUint1(x127))) + var x130 uint64 + var x131 uint64 + x130, x131 = bits.Add64(x109, x115, uint64(fiatScalarUint1(x129))) + var x132 uint64 + var x133 uint64 + x132, x133 = bits.Add64(x111, x116, uint64(fiatScalarUint1(x131))) + x134 := (uint64(fiatScalarUint1(x133)) + uint64(fiatScalarUint1(x112))) + var x135 uint64 + var x136 uint64 + x136, x135 = bits.Mul64(x3, arg2[3]) + var x137 uint64 + var x138 uint64 + x138, x137 = bits.Mul64(x3, arg2[2]) + var x139 uint64 + var x140 uint64 + x140, x139 = bits.Mul64(x3, arg2[1]) + var x141 uint64 + var x142 uint64 + x142, x141 = bits.Mul64(x3, arg2[0]) + var x143 uint64 + var x144 uint64 + x143, x144 = bits.Add64(x142, x139, uint64(0x0)) + var x145 uint64 + var x146 uint64 + x145, x146 = bits.Add64(x140, x137, uint64(fiatScalarUint1(x144))) + var x147 uint64 + var x148 uint64 + x147, x148 = bits.Add64(x138, x135, uint64(fiatScalarUint1(x146))) + x149 := (uint64(fiatScalarUint1(x148)) + x136) + var x150 uint64 + var x151 uint64 + x150, x151 = bits.Add64(x126, x141, uint64(0x0)) + var x152 uint64 + var x153 uint64 + x152, x153 = bits.Add64(x128, x143, uint64(fiatScalarUint1(x151))) + var x154 uint64 + var x155 uint64 + x154, x155 = bits.Add64(x130, x145, uint64(fiatScalarUint1(x153))) + var x156 uint64 + var x157 uint64 + x156, x157 = bits.Add64(x132, x147, uint64(fiatScalarUint1(x155))) + var x158 uint64 + var x159 uint64 + x158, x159 = bits.Add64(x134, x149, uint64(fiatScalarUint1(x157))) + var x160 uint64 + _, x160 = bits.Mul64(x150, 0xd2b51da312547e1b) + var x162 uint64 + var x163 uint64 + x163, x162 = bits.Mul64(x160, 0x1000000000000000) + var x164 uint64 + var x165 uint64 + x165, x164 = bits.Mul64(x160, 0x14def9dea2f79cd6) + var x166 uint64 + var x167 uint64 + x167, x166 = bits.Mul64(x160, 0x5812631a5cf5d3ed) + var x168 uint64 + var x169 uint64 + x168, x169 = bits.Add64(x167, x164, uint64(0x0)) + x170 := (uint64(fiatScalarUint1(x169)) + x165) + var x172 uint64 + _, x172 = bits.Add64(x150, x166, uint64(0x0)) + var x173 uint64 + var x174 uint64 + x173, x174 = bits.Add64(x152, x168, uint64(fiatScalarUint1(x172))) + var x175 uint64 + var x176 uint64 + x175, x176 = bits.Add64(x154, x170, uint64(fiatScalarUint1(x174))) + var x177 uint64 + var x178 uint64 + x177, x178 = bits.Add64(x156, x162, uint64(fiatScalarUint1(x176))) + var x179 uint64 + var x180 uint64 + x179, x180 = bits.Add64(x158, x163, uint64(fiatScalarUint1(x178))) + x181 := (uint64(fiatScalarUint1(x180)) + uint64(fiatScalarUint1(x159))) + var x182 uint64 + var x183 uint64 + x182, x183 = bits.Sub64(x173, 0x5812631a5cf5d3ed, uint64(0x0)) + var x184 uint64 + var x185 uint64 + x184, x185 = bits.Sub64(x175, 0x14def9dea2f79cd6, uint64(fiatScalarUint1(x183))) + var x186 uint64 + var x187 uint64 + x186, x187 = bits.Sub64(x177, uint64(0x0), uint64(fiatScalarUint1(x185))) + var x188 uint64 + var x189 uint64 + x188, x189 = bits.Sub64(x179, 0x1000000000000000, uint64(fiatScalarUint1(x187))) + var x191 uint64 + _, x191 = bits.Sub64(x181, uint64(0x0), uint64(fiatScalarUint1(x189))) + var x192 uint64 + fiatScalarCmovznzU64(&x192, fiatScalarUint1(x191), x182, x173) + var x193 uint64 + fiatScalarCmovznzU64(&x193, fiatScalarUint1(x191), x184, x175) + var x194 uint64 + fiatScalarCmovznzU64(&x194, fiatScalarUint1(x191), x186, x177) + var x195 uint64 + fiatScalarCmovznzU64(&x195, fiatScalarUint1(x191), x188, x179) + out1[0] = x192 + out1[1] = x193 + out1[2] = x194 + out1[3] = x195 +} + +// fiatScalarAdd adds two field elements in the Montgomery domain. +// +// Preconditions: +// +// 0 ≤ eval arg1 < m +// 0 ≤ eval arg2 < m +// +// Postconditions: +// +// eval (from_montgomery out1) mod m = (eval (from_montgomery arg1) + eval (from_montgomery arg2)) mod m +// 0 ≤ eval out1 < m +func fiatScalarAdd(out1 *fiatScalarMontgomeryDomainFieldElement, arg1 *fiatScalarMontgomeryDomainFieldElement, arg2 *fiatScalarMontgomeryDomainFieldElement) { + var x1 uint64 + var x2 uint64 + x1, x2 = bits.Add64(arg1[0], arg2[0], uint64(0x0)) + var x3 uint64 + var x4 uint64 + x3, x4 = bits.Add64(arg1[1], arg2[1], uint64(fiatScalarUint1(x2))) + var x5 uint64 + var x6 uint64 + x5, x6 = bits.Add64(arg1[2], arg2[2], uint64(fiatScalarUint1(x4))) + var x7 uint64 + var x8 uint64 + x7, x8 = bits.Add64(arg1[3], arg2[3], uint64(fiatScalarUint1(x6))) + var x9 uint64 + var x10 uint64 + x9, x10 = bits.Sub64(x1, 0x5812631a5cf5d3ed, uint64(0x0)) + var x11 uint64 + var x12 uint64 + x11, x12 = bits.Sub64(x3, 0x14def9dea2f79cd6, uint64(fiatScalarUint1(x10))) + var x13 uint64 + var x14 uint64 + x13, x14 = bits.Sub64(x5, uint64(0x0), uint64(fiatScalarUint1(x12))) + var x15 uint64 + var x16 uint64 + x15, x16 = bits.Sub64(x7, 0x1000000000000000, uint64(fiatScalarUint1(x14))) + var x18 uint64 + _, x18 = bits.Sub64(uint64(fiatScalarUint1(x8)), uint64(0x0), uint64(fiatScalarUint1(x16))) + var x19 uint64 + fiatScalarCmovznzU64(&x19, fiatScalarUint1(x18), x9, x1) + var x20 uint64 + fiatScalarCmovznzU64(&x20, fiatScalarUint1(x18), x11, x3) + var x21 uint64 + fiatScalarCmovznzU64(&x21, fiatScalarUint1(x18), x13, x5) + var x22 uint64 + fiatScalarCmovznzU64(&x22, fiatScalarUint1(x18), x15, x7) + out1[0] = x19 + out1[1] = x20 + out1[2] = x21 + out1[3] = x22 +} + +// fiatScalarSub subtracts two field elements in the Montgomery domain. +// +// Preconditions: +// +// 0 ≤ eval arg1 < m +// 0 ≤ eval arg2 < m +// +// Postconditions: +// +// eval (from_montgomery out1) mod m = (eval (from_montgomery arg1) - eval (from_montgomery arg2)) mod m +// 0 ≤ eval out1 < m +func fiatScalarSub(out1 *fiatScalarMontgomeryDomainFieldElement, arg1 *fiatScalarMontgomeryDomainFieldElement, arg2 *fiatScalarMontgomeryDomainFieldElement) { + var x1 uint64 + var x2 uint64 + x1, x2 = bits.Sub64(arg1[0], arg2[0], uint64(0x0)) + var x3 uint64 + var x4 uint64 + x3, x4 = bits.Sub64(arg1[1], arg2[1], uint64(fiatScalarUint1(x2))) + var x5 uint64 + var x6 uint64 + x5, x6 = bits.Sub64(arg1[2], arg2[2], uint64(fiatScalarUint1(x4))) + var x7 uint64 + var x8 uint64 + x7, x8 = bits.Sub64(arg1[3], arg2[3], uint64(fiatScalarUint1(x6))) + var x9 uint64 + fiatScalarCmovznzU64(&x9, fiatScalarUint1(x8), uint64(0x0), 0xffffffffffffffff) + var x10 uint64 + var x11 uint64 + x10, x11 = bits.Add64(x1, (x9 & 0x5812631a5cf5d3ed), uint64(0x0)) + var x12 uint64 + var x13 uint64 + x12, x13 = bits.Add64(x3, (x9 & 0x14def9dea2f79cd6), uint64(fiatScalarUint1(x11))) + var x14 uint64 + var x15 uint64 + x14, x15 = bits.Add64(x5, uint64(0x0), uint64(fiatScalarUint1(x13))) + var x16 uint64 + x16, _ = bits.Add64(x7, (x9 & 0x1000000000000000), uint64(fiatScalarUint1(x15))) + out1[0] = x10 + out1[1] = x12 + out1[2] = x14 + out1[3] = x16 +} + +// fiatScalarOpp negates a field element in the Montgomery domain. +// +// Preconditions: +// +// 0 ≤ eval arg1 < m +// +// Postconditions: +// +// eval (from_montgomery out1) mod m = -eval (from_montgomery arg1) mod m +// 0 ≤ eval out1 < m +func fiatScalarOpp(out1 *fiatScalarMontgomeryDomainFieldElement, arg1 *fiatScalarMontgomeryDomainFieldElement) { + var x1 uint64 + var x2 uint64 + x1, x2 = bits.Sub64(uint64(0x0), arg1[0], uint64(0x0)) + var x3 uint64 + var x4 uint64 + x3, x4 = bits.Sub64(uint64(0x0), arg1[1], uint64(fiatScalarUint1(x2))) + var x5 uint64 + var x6 uint64 + x5, x6 = bits.Sub64(uint64(0x0), arg1[2], uint64(fiatScalarUint1(x4))) + var x7 uint64 + var x8 uint64 + x7, x8 = bits.Sub64(uint64(0x0), arg1[3], uint64(fiatScalarUint1(x6))) + var x9 uint64 + fiatScalarCmovznzU64(&x9, fiatScalarUint1(x8), uint64(0x0), 0xffffffffffffffff) + var x10 uint64 + var x11 uint64 + x10, x11 = bits.Add64(x1, (x9 & 0x5812631a5cf5d3ed), uint64(0x0)) + var x12 uint64 + var x13 uint64 + x12, x13 = bits.Add64(x3, (x9 & 0x14def9dea2f79cd6), uint64(fiatScalarUint1(x11))) + var x14 uint64 + var x15 uint64 + x14, x15 = bits.Add64(x5, uint64(0x0), uint64(fiatScalarUint1(x13))) + var x16 uint64 + x16, _ = bits.Add64(x7, (x9 & 0x1000000000000000), uint64(fiatScalarUint1(x15))) + out1[0] = x10 + out1[1] = x12 + out1[2] = x14 + out1[3] = x16 +} + +// fiatScalarNonzero outputs a single non-zero word if the input is non-zero and zero otherwise. +// +// Preconditions: +// +// 0 ≤ eval arg1 < m +// +// Postconditions: +// +// out1 = 0 ↔ eval (from_montgomery arg1) mod m = 0 +// +// Input Bounds: +// +// arg1: [[0x0 ~> 0xffffffffffffffff], [0x0 ~> 0xffffffffffffffff], [0x0 ~> 0xffffffffffffffff], [0x0 ~> 0xffffffffffffffff]] +// +// Output Bounds: +// +// out1: [0x0 ~> 0xffffffffffffffff] +func fiatScalarNonzero(out1 *uint64, arg1 *[4]uint64) { + x1 := (arg1[0] | (arg1[1] | (arg1[2] | arg1[3]))) + *out1 = x1 +} + +// fiatScalarFromMontgomery translates a field element out of the Montgomery domain. +// +// Preconditions: +// +// 0 ≤ eval arg1 < m +// +// Postconditions: +// +// eval out1 mod m = (eval arg1 * ((2^64)⁻¹ mod m)^4) mod m +// 0 ≤ eval out1 < m +func fiatScalarFromMontgomery(out1 *fiatScalarNonMontgomeryDomainFieldElement, arg1 *fiatScalarMontgomeryDomainFieldElement) { + x1 := arg1[0] + var x2 uint64 + _, x2 = bits.Mul64(x1, 0xd2b51da312547e1b) + var x4 uint64 + var x5 uint64 + x5, x4 = bits.Mul64(x2, 0x1000000000000000) + var x6 uint64 + var x7 uint64 + x7, x6 = bits.Mul64(x2, 0x14def9dea2f79cd6) + var x8 uint64 + var x9 uint64 + x9, x8 = bits.Mul64(x2, 0x5812631a5cf5d3ed) + var x10 uint64 + var x11 uint64 + x10, x11 = bits.Add64(x9, x6, uint64(0x0)) + var x13 uint64 + _, x13 = bits.Add64(x1, x8, uint64(0x0)) + var x14 uint64 + var x15 uint64 + x14, x15 = bits.Add64(uint64(0x0), x10, uint64(fiatScalarUint1(x13))) + var x16 uint64 + var x17 uint64 + x16, x17 = bits.Add64(x14, arg1[1], uint64(0x0)) + var x18 uint64 + _, x18 = bits.Mul64(x16, 0xd2b51da312547e1b) + var x20 uint64 + var x21 uint64 + x21, x20 = bits.Mul64(x18, 0x1000000000000000) + var x22 uint64 + var x23 uint64 + x23, x22 = bits.Mul64(x18, 0x14def9dea2f79cd6) + var x24 uint64 + var x25 uint64 + x25, x24 = bits.Mul64(x18, 0x5812631a5cf5d3ed) + var x26 uint64 + var x27 uint64 + x26, x27 = bits.Add64(x25, x22, uint64(0x0)) + var x29 uint64 + _, x29 = bits.Add64(x16, x24, uint64(0x0)) + var x30 uint64 + var x31 uint64 + x30, x31 = bits.Add64((uint64(fiatScalarUint1(x17)) + (uint64(fiatScalarUint1(x15)) + (uint64(fiatScalarUint1(x11)) + x7))), x26, uint64(fiatScalarUint1(x29))) + var x32 uint64 + var x33 uint64 + x32, x33 = bits.Add64(x4, (uint64(fiatScalarUint1(x27)) + x23), uint64(fiatScalarUint1(x31))) + var x34 uint64 + var x35 uint64 + x34, x35 = bits.Add64(x5, x20, uint64(fiatScalarUint1(x33))) + var x36 uint64 + var x37 uint64 + x36, x37 = bits.Add64(x30, arg1[2], uint64(0x0)) + var x38 uint64 + var x39 uint64 + x38, x39 = bits.Add64(x32, uint64(0x0), uint64(fiatScalarUint1(x37))) + var x40 uint64 + var x41 uint64 + x40, x41 = bits.Add64(x34, uint64(0x0), uint64(fiatScalarUint1(x39))) + var x42 uint64 + _, x42 = bits.Mul64(x36, 0xd2b51da312547e1b) + var x44 uint64 + var x45 uint64 + x45, x44 = bits.Mul64(x42, 0x1000000000000000) + var x46 uint64 + var x47 uint64 + x47, x46 = bits.Mul64(x42, 0x14def9dea2f79cd6) + var x48 uint64 + var x49 uint64 + x49, x48 = bits.Mul64(x42, 0x5812631a5cf5d3ed) + var x50 uint64 + var x51 uint64 + x50, x51 = bits.Add64(x49, x46, uint64(0x0)) + var x53 uint64 + _, x53 = bits.Add64(x36, x48, uint64(0x0)) + var x54 uint64 + var x55 uint64 + x54, x55 = bits.Add64(x38, x50, uint64(fiatScalarUint1(x53))) + var x56 uint64 + var x57 uint64 + x56, x57 = bits.Add64(x40, (uint64(fiatScalarUint1(x51)) + x47), uint64(fiatScalarUint1(x55))) + var x58 uint64 + var x59 uint64 + x58, x59 = bits.Add64((uint64(fiatScalarUint1(x41)) + (uint64(fiatScalarUint1(x35)) + x21)), x44, uint64(fiatScalarUint1(x57))) + var x60 uint64 + var x61 uint64 + x60, x61 = bits.Add64(x54, arg1[3], uint64(0x0)) + var x62 uint64 + var x63 uint64 + x62, x63 = bits.Add64(x56, uint64(0x0), uint64(fiatScalarUint1(x61))) + var x64 uint64 + var x65 uint64 + x64, x65 = bits.Add64(x58, uint64(0x0), uint64(fiatScalarUint1(x63))) + var x66 uint64 + _, x66 = bits.Mul64(x60, 0xd2b51da312547e1b) + var x68 uint64 + var x69 uint64 + x69, x68 = bits.Mul64(x66, 0x1000000000000000) + var x70 uint64 + var x71 uint64 + x71, x70 = bits.Mul64(x66, 0x14def9dea2f79cd6) + var x72 uint64 + var x73 uint64 + x73, x72 = bits.Mul64(x66, 0x5812631a5cf5d3ed) + var x74 uint64 + var x75 uint64 + x74, x75 = bits.Add64(x73, x70, uint64(0x0)) + var x77 uint64 + _, x77 = bits.Add64(x60, x72, uint64(0x0)) + var x78 uint64 + var x79 uint64 + x78, x79 = bits.Add64(x62, x74, uint64(fiatScalarUint1(x77))) + var x80 uint64 + var x81 uint64 + x80, x81 = bits.Add64(x64, (uint64(fiatScalarUint1(x75)) + x71), uint64(fiatScalarUint1(x79))) + var x82 uint64 + var x83 uint64 + x82, x83 = bits.Add64((uint64(fiatScalarUint1(x65)) + (uint64(fiatScalarUint1(x59)) + x45)), x68, uint64(fiatScalarUint1(x81))) + x84 := (uint64(fiatScalarUint1(x83)) + x69) + var x85 uint64 + var x86 uint64 + x85, x86 = bits.Sub64(x78, 0x5812631a5cf5d3ed, uint64(0x0)) + var x87 uint64 + var x88 uint64 + x87, x88 = bits.Sub64(x80, 0x14def9dea2f79cd6, uint64(fiatScalarUint1(x86))) + var x89 uint64 + var x90 uint64 + x89, x90 = bits.Sub64(x82, uint64(0x0), uint64(fiatScalarUint1(x88))) + var x91 uint64 + var x92 uint64 + x91, x92 = bits.Sub64(x84, 0x1000000000000000, uint64(fiatScalarUint1(x90))) + var x94 uint64 + _, x94 = bits.Sub64(uint64(0x0), uint64(0x0), uint64(fiatScalarUint1(x92))) + var x95 uint64 + fiatScalarCmovznzU64(&x95, fiatScalarUint1(x94), x85, x78) + var x96 uint64 + fiatScalarCmovznzU64(&x96, fiatScalarUint1(x94), x87, x80) + var x97 uint64 + fiatScalarCmovznzU64(&x97, fiatScalarUint1(x94), x89, x82) + var x98 uint64 + fiatScalarCmovznzU64(&x98, fiatScalarUint1(x94), x91, x84) + out1[0] = x95 + out1[1] = x96 + out1[2] = x97 + out1[3] = x98 +} + +// fiatScalarToMontgomery translates a field element into the Montgomery domain. +// +// Preconditions: +// +// 0 ≤ eval arg1 < m +// +// Postconditions: +// +// eval (from_montgomery out1) mod m = eval arg1 mod m +// 0 ≤ eval out1 < m +func fiatScalarToMontgomery(out1 *fiatScalarMontgomeryDomainFieldElement, arg1 *fiatScalarNonMontgomeryDomainFieldElement) { + x1 := arg1[1] + x2 := arg1[2] + x3 := arg1[3] + x4 := arg1[0] + var x5 uint64 + var x6 uint64 + x6, x5 = bits.Mul64(x4, 0x399411b7c309a3d) + var x7 uint64 + var x8 uint64 + x8, x7 = bits.Mul64(x4, 0xceec73d217f5be65) + var x9 uint64 + var x10 uint64 + x10, x9 = bits.Mul64(x4, 0xd00e1ba768859347) + var x11 uint64 + var x12 uint64 + x12, x11 = bits.Mul64(x4, 0xa40611e3449c0f01) + var x13 uint64 + var x14 uint64 + x13, x14 = bits.Add64(x12, x9, uint64(0x0)) + var x15 uint64 + var x16 uint64 + x15, x16 = bits.Add64(x10, x7, uint64(fiatScalarUint1(x14))) + var x17 uint64 + var x18 uint64 + x17, x18 = bits.Add64(x8, x5, uint64(fiatScalarUint1(x16))) + var x19 uint64 + _, x19 = bits.Mul64(x11, 0xd2b51da312547e1b) + var x21 uint64 + var x22 uint64 + x22, x21 = bits.Mul64(x19, 0x1000000000000000) + var x23 uint64 + var x24 uint64 + x24, x23 = bits.Mul64(x19, 0x14def9dea2f79cd6) + var x25 uint64 + var x26 uint64 + x26, x25 = bits.Mul64(x19, 0x5812631a5cf5d3ed) + var x27 uint64 + var x28 uint64 + x27, x28 = bits.Add64(x26, x23, uint64(0x0)) + var x30 uint64 + _, x30 = bits.Add64(x11, x25, uint64(0x0)) + var x31 uint64 + var x32 uint64 + x31, x32 = bits.Add64(x13, x27, uint64(fiatScalarUint1(x30))) + var x33 uint64 + var x34 uint64 + x33, x34 = bits.Add64(x15, (uint64(fiatScalarUint1(x28)) + x24), uint64(fiatScalarUint1(x32))) + var x35 uint64 + var x36 uint64 + x35, x36 = bits.Add64(x17, x21, uint64(fiatScalarUint1(x34))) + var x37 uint64 + var x38 uint64 + x38, x37 = bits.Mul64(x1, 0x399411b7c309a3d) + var x39 uint64 + var x40 uint64 + x40, x39 = bits.Mul64(x1, 0xceec73d217f5be65) + var x41 uint64 + var x42 uint64 + x42, x41 = bits.Mul64(x1, 0xd00e1ba768859347) + var x43 uint64 + var x44 uint64 + x44, x43 = bits.Mul64(x1, 0xa40611e3449c0f01) + var x45 uint64 + var x46 uint64 + x45, x46 = bits.Add64(x44, x41, uint64(0x0)) + var x47 uint64 + var x48 uint64 + x47, x48 = bits.Add64(x42, x39, uint64(fiatScalarUint1(x46))) + var x49 uint64 + var x50 uint64 + x49, x50 = bits.Add64(x40, x37, uint64(fiatScalarUint1(x48))) + var x51 uint64 + var x52 uint64 + x51, x52 = bits.Add64(x31, x43, uint64(0x0)) + var x53 uint64 + var x54 uint64 + x53, x54 = bits.Add64(x33, x45, uint64(fiatScalarUint1(x52))) + var x55 uint64 + var x56 uint64 + x55, x56 = bits.Add64(x35, x47, uint64(fiatScalarUint1(x54))) + var x57 uint64 + var x58 uint64 + x57, x58 = bits.Add64(((uint64(fiatScalarUint1(x36)) + (uint64(fiatScalarUint1(x18)) + x6)) + x22), x49, uint64(fiatScalarUint1(x56))) + var x59 uint64 + _, x59 = bits.Mul64(x51, 0xd2b51da312547e1b) + var x61 uint64 + var x62 uint64 + x62, x61 = bits.Mul64(x59, 0x1000000000000000) + var x63 uint64 + var x64 uint64 + x64, x63 = bits.Mul64(x59, 0x14def9dea2f79cd6) + var x65 uint64 + var x66 uint64 + x66, x65 = bits.Mul64(x59, 0x5812631a5cf5d3ed) + var x67 uint64 + var x68 uint64 + x67, x68 = bits.Add64(x66, x63, uint64(0x0)) + var x70 uint64 + _, x70 = bits.Add64(x51, x65, uint64(0x0)) + var x71 uint64 + var x72 uint64 + x71, x72 = bits.Add64(x53, x67, uint64(fiatScalarUint1(x70))) + var x73 uint64 + var x74 uint64 + x73, x74 = bits.Add64(x55, (uint64(fiatScalarUint1(x68)) + x64), uint64(fiatScalarUint1(x72))) + var x75 uint64 + var x76 uint64 + x75, x76 = bits.Add64(x57, x61, uint64(fiatScalarUint1(x74))) + var x77 uint64 + var x78 uint64 + x78, x77 = bits.Mul64(x2, 0x399411b7c309a3d) + var x79 uint64 + var x80 uint64 + x80, x79 = bits.Mul64(x2, 0xceec73d217f5be65) + var x81 uint64 + var x82 uint64 + x82, x81 = bits.Mul64(x2, 0xd00e1ba768859347) + var x83 uint64 + var x84 uint64 + x84, x83 = bits.Mul64(x2, 0xa40611e3449c0f01) + var x85 uint64 + var x86 uint64 + x85, x86 = bits.Add64(x84, x81, uint64(0x0)) + var x87 uint64 + var x88 uint64 + x87, x88 = bits.Add64(x82, x79, uint64(fiatScalarUint1(x86))) + var x89 uint64 + var x90 uint64 + x89, x90 = bits.Add64(x80, x77, uint64(fiatScalarUint1(x88))) + var x91 uint64 + var x92 uint64 + x91, x92 = bits.Add64(x71, x83, uint64(0x0)) + var x93 uint64 + var x94 uint64 + x93, x94 = bits.Add64(x73, x85, uint64(fiatScalarUint1(x92))) + var x95 uint64 + var x96 uint64 + x95, x96 = bits.Add64(x75, x87, uint64(fiatScalarUint1(x94))) + var x97 uint64 + var x98 uint64 + x97, x98 = bits.Add64(((uint64(fiatScalarUint1(x76)) + (uint64(fiatScalarUint1(x58)) + (uint64(fiatScalarUint1(x50)) + x38))) + x62), x89, uint64(fiatScalarUint1(x96))) + var x99 uint64 + _, x99 = bits.Mul64(x91, 0xd2b51da312547e1b) + var x101 uint64 + var x102 uint64 + x102, x101 = bits.Mul64(x99, 0x1000000000000000) + var x103 uint64 + var x104 uint64 + x104, x103 = bits.Mul64(x99, 0x14def9dea2f79cd6) + var x105 uint64 + var x106 uint64 + x106, x105 = bits.Mul64(x99, 0x5812631a5cf5d3ed) + var x107 uint64 + var x108 uint64 + x107, x108 = bits.Add64(x106, x103, uint64(0x0)) + var x110 uint64 + _, x110 = bits.Add64(x91, x105, uint64(0x0)) + var x111 uint64 + var x112 uint64 + x111, x112 = bits.Add64(x93, x107, uint64(fiatScalarUint1(x110))) + var x113 uint64 + var x114 uint64 + x113, x114 = bits.Add64(x95, (uint64(fiatScalarUint1(x108)) + x104), uint64(fiatScalarUint1(x112))) + var x115 uint64 + var x116 uint64 + x115, x116 = bits.Add64(x97, x101, uint64(fiatScalarUint1(x114))) + var x117 uint64 + var x118 uint64 + x118, x117 = bits.Mul64(x3, 0x399411b7c309a3d) + var x119 uint64 + var x120 uint64 + x120, x119 = bits.Mul64(x3, 0xceec73d217f5be65) + var x121 uint64 + var x122 uint64 + x122, x121 = bits.Mul64(x3, 0xd00e1ba768859347) + var x123 uint64 + var x124 uint64 + x124, x123 = bits.Mul64(x3, 0xa40611e3449c0f01) + var x125 uint64 + var x126 uint64 + x125, x126 = bits.Add64(x124, x121, uint64(0x0)) + var x127 uint64 + var x128 uint64 + x127, x128 = bits.Add64(x122, x119, uint64(fiatScalarUint1(x126))) + var x129 uint64 + var x130 uint64 + x129, x130 = bits.Add64(x120, x117, uint64(fiatScalarUint1(x128))) + var x131 uint64 + var x132 uint64 + x131, x132 = bits.Add64(x111, x123, uint64(0x0)) + var x133 uint64 + var x134 uint64 + x133, x134 = bits.Add64(x113, x125, uint64(fiatScalarUint1(x132))) + var x135 uint64 + var x136 uint64 + x135, x136 = bits.Add64(x115, x127, uint64(fiatScalarUint1(x134))) + var x137 uint64 + var x138 uint64 + x137, x138 = bits.Add64(((uint64(fiatScalarUint1(x116)) + (uint64(fiatScalarUint1(x98)) + (uint64(fiatScalarUint1(x90)) + x78))) + x102), x129, uint64(fiatScalarUint1(x136))) + var x139 uint64 + _, x139 = bits.Mul64(x131, 0xd2b51da312547e1b) + var x141 uint64 + var x142 uint64 + x142, x141 = bits.Mul64(x139, 0x1000000000000000) + var x143 uint64 + var x144 uint64 + x144, x143 = bits.Mul64(x139, 0x14def9dea2f79cd6) + var x145 uint64 + var x146 uint64 + x146, x145 = bits.Mul64(x139, 0x5812631a5cf5d3ed) + var x147 uint64 + var x148 uint64 + x147, x148 = bits.Add64(x146, x143, uint64(0x0)) + var x150 uint64 + _, x150 = bits.Add64(x131, x145, uint64(0x0)) + var x151 uint64 + var x152 uint64 + x151, x152 = bits.Add64(x133, x147, uint64(fiatScalarUint1(x150))) + var x153 uint64 + var x154 uint64 + x153, x154 = bits.Add64(x135, (uint64(fiatScalarUint1(x148)) + x144), uint64(fiatScalarUint1(x152))) + var x155 uint64 + var x156 uint64 + x155, x156 = bits.Add64(x137, x141, uint64(fiatScalarUint1(x154))) + x157 := ((uint64(fiatScalarUint1(x156)) + (uint64(fiatScalarUint1(x138)) + (uint64(fiatScalarUint1(x130)) + x118))) + x142) + var x158 uint64 + var x159 uint64 + x158, x159 = bits.Sub64(x151, 0x5812631a5cf5d3ed, uint64(0x0)) + var x160 uint64 + var x161 uint64 + x160, x161 = bits.Sub64(x153, 0x14def9dea2f79cd6, uint64(fiatScalarUint1(x159))) + var x162 uint64 + var x163 uint64 + x162, x163 = bits.Sub64(x155, uint64(0x0), uint64(fiatScalarUint1(x161))) + var x164 uint64 + var x165 uint64 + x164, x165 = bits.Sub64(x157, 0x1000000000000000, uint64(fiatScalarUint1(x163))) + var x167 uint64 + _, x167 = bits.Sub64(uint64(0x0), uint64(0x0), uint64(fiatScalarUint1(x165))) + var x168 uint64 + fiatScalarCmovznzU64(&x168, fiatScalarUint1(x167), x158, x151) + var x169 uint64 + fiatScalarCmovznzU64(&x169, fiatScalarUint1(x167), x160, x153) + var x170 uint64 + fiatScalarCmovznzU64(&x170, fiatScalarUint1(x167), x162, x155) + var x171 uint64 + fiatScalarCmovznzU64(&x171, fiatScalarUint1(x167), x164, x157) + out1[0] = x168 + out1[1] = x169 + out1[2] = x170 + out1[3] = x171 +} + +// fiatScalarToBytes serializes a field element NOT in the Montgomery domain to bytes in little-endian order. +// +// Preconditions: +// +// 0 ≤ eval arg1 < m +// +// Postconditions: +// +// out1 = map (λ x, ⌊((eval arg1 mod m) mod 2^(8 * (x + 1))) / 2^(8 * x)⌋) [0..31] +// +// Input Bounds: +// +// arg1: [[0x0 ~> 0xffffffffffffffff], [0x0 ~> 0xffffffffffffffff], [0x0 ~> 0xffffffffffffffff], [0x0 ~> 0x1fffffffffffffff]] +// +// Output Bounds: +// +// out1: [[0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0x1f]] +func fiatScalarToBytes(out1 *[32]uint8, arg1 *[4]uint64) { + x1 := arg1[3] + x2 := arg1[2] + x3 := arg1[1] + x4 := arg1[0] + x5 := (uint8(x4) & 0xff) + x6 := (x4 >> 8) + x7 := (uint8(x6) & 0xff) + x8 := (x6 >> 8) + x9 := (uint8(x8) & 0xff) + x10 := (x8 >> 8) + x11 := (uint8(x10) & 0xff) + x12 := (x10 >> 8) + x13 := (uint8(x12) & 0xff) + x14 := (x12 >> 8) + x15 := (uint8(x14) & 0xff) + x16 := (x14 >> 8) + x17 := (uint8(x16) & 0xff) + x18 := uint8((x16 >> 8)) + x19 := (uint8(x3) & 0xff) + x20 := (x3 >> 8) + x21 := (uint8(x20) & 0xff) + x22 := (x20 >> 8) + x23 := (uint8(x22) & 0xff) + x24 := (x22 >> 8) + x25 := (uint8(x24) & 0xff) + x26 := (x24 >> 8) + x27 := (uint8(x26) & 0xff) + x28 := (x26 >> 8) + x29 := (uint8(x28) & 0xff) + x30 := (x28 >> 8) + x31 := (uint8(x30) & 0xff) + x32 := uint8((x30 >> 8)) + x33 := (uint8(x2) & 0xff) + x34 := (x2 >> 8) + x35 := (uint8(x34) & 0xff) + x36 := (x34 >> 8) + x37 := (uint8(x36) & 0xff) + x38 := (x36 >> 8) + x39 := (uint8(x38) & 0xff) + x40 := (x38 >> 8) + x41 := (uint8(x40) & 0xff) + x42 := (x40 >> 8) + x43 := (uint8(x42) & 0xff) + x44 := (x42 >> 8) + x45 := (uint8(x44) & 0xff) + x46 := uint8((x44 >> 8)) + x47 := (uint8(x1) & 0xff) + x48 := (x1 >> 8) + x49 := (uint8(x48) & 0xff) + x50 := (x48 >> 8) + x51 := (uint8(x50) & 0xff) + x52 := (x50 >> 8) + x53 := (uint8(x52) & 0xff) + x54 := (x52 >> 8) + x55 := (uint8(x54) & 0xff) + x56 := (x54 >> 8) + x57 := (uint8(x56) & 0xff) + x58 := (x56 >> 8) + x59 := (uint8(x58) & 0xff) + x60 := uint8((x58 >> 8)) + out1[0] = x5 + out1[1] = x7 + out1[2] = x9 + out1[3] = x11 + out1[4] = x13 + out1[5] = x15 + out1[6] = x17 + out1[7] = x18 + out1[8] = x19 + out1[9] = x21 + out1[10] = x23 + out1[11] = x25 + out1[12] = x27 + out1[13] = x29 + out1[14] = x31 + out1[15] = x32 + out1[16] = x33 + out1[17] = x35 + out1[18] = x37 + out1[19] = x39 + out1[20] = x41 + out1[21] = x43 + out1[22] = x45 + out1[23] = x46 + out1[24] = x47 + out1[25] = x49 + out1[26] = x51 + out1[27] = x53 + out1[28] = x55 + out1[29] = x57 + out1[30] = x59 + out1[31] = x60 +} + +// fiatScalarFromBytes deserializes a field element NOT in the Montgomery domain from bytes in little-endian order. +// +// Preconditions: +// +// 0 ≤ bytes_eval arg1 < m +// +// Postconditions: +// +// eval out1 mod m = bytes_eval arg1 mod m +// 0 ≤ eval out1 < m +// +// Input Bounds: +// +// arg1: [[0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0xff], [0x0 ~> 0x1f]] +// +// Output Bounds: +// +// out1: [[0x0 ~> 0xffffffffffffffff], [0x0 ~> 0xffffffffffffffff], [0x0 ~> 0xffffffffffffffff], [0x0 ~> 0x1fffffffffffffff]] +func fiatScalarFromBytes(out1 *[4]uint64, arg1 *[32]uint8) { + x1 := (uint64(arg1[31]) << 56) + x2 := (uint64(arg1[30]) << 48) + x3 := (uint64(arg1[29]) << 40) + x4 := (uint64(arg1[28]) << 32) + x5 := (uint64(arg1[27]) << 24) + x6 := (uint64(arg1[26]) << 16) + x7 := (uint64(arg1[25]) << 8) + x8 := arg1[24] + x9 := (uint64(arg1[23]) << 56) + x10 := (uint64(arg1[22]) << 48) + x11 := (uint64(arg1[21]) << 40) + x12 := (uint64(arg1[20]) << 32) + x13 := (uint64(arg1[19]) << 24) + x14 := (uint64(arg1[18]) << 16) + x15 := (uint64(arg1[17]) << 8) + x16 := arg1[16] + x17 := (uint64(arg1[15]) << 56) + x18 := (uint64(arg1[14]) << 48) + x19 := (uint64(arg1[13]) << 40) + x20 := (uint64(arg1[12]) << 32) + x21 := (uint64(arg1[11]) << 24) + x22 := (uint64(arg1[10]) << 16) + x23 := (uint64(arg1[9]) << 8) + x24 := arg1[8] + x25 := (uint64(arg1[7]) << 56) + x26 := (uint64(arg1[6]) << 48) + x27 := (uint64(arg1[5]) << 40) + x28 := (uint64(arg1[4]) << 32) + x29 := (uint64(arg1[3]) << 24) + x30 := (uint64(arg1[2]) << 16) + x31 := (uint64(arg1[1]) << 8) + x32 := arg1[0] + x33 := (x31 + uint64(x32)) + x34 := (x30 + x33) + x35 := (x29 + x34) + x36 := (x28 + x35) + x37 := (x27 + x36) + x38 := (x26 + x37) + x39 := (x25 + x38) + x40 := (x23 + uint64(x24)) + x41 := (x22 + x40) + x42 := (x21 + x41) + x43 := (x20 + x42) + x44 := (x19 + x43) + x45 := (x18 + x44) + x46 := (x17 + x45) + x47 := (x15 + uint64(x16)) + x48 := (x14 + x47) + x49 := (x13 + x48) + x50 := (x12 + x49) + x51 := (x11 + x50) + x52 := (x10 + x51) + x53 := (x9 + x52) + x54 := (x7 + uint64(x8)) + x55 := (x6 + x54) + x56 := (x5 + x55) + x57 := (x4 + x56) + x58 := (x3 + x57) + x59 := (x2 + x58) + x60 := (x1 + x59) + out1[0] = x39 + out1[1] = x46 + out1[2] = x53 + out1[3] = x60 +} diff --git a/scalar_test.go b/scalar_test.go new file mode 100644 index 0000000..d669411 --- /dev/null +++ b/scalar_test.go @@ -0,0 +1,303 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519 +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package edwards25519 + +import ( + "bytes" + "encoding/hex" + "math/big" + mathrand "math/rand" + "reflect" + "testing" + "testing/quick" +) + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// quickCheckConfig returns a quick.Config that scales the max count by the +// given factor if the -short flag is not set. +func quickCheckConfig(slowScale int) *quick.Config { + cfg := new(quick.Config) + if !testing.Short() { + cfg.MaxCountScale = float64(slowScale) + } + return cfg +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +var scOneBytes = [32]byte{1} +var scOne, _ = new(Scalar).SetCanonicalBytes(scOneBytes[:]) +var scMinusOne = new(Scalar).Subtract(new(Scalar), scOne) + +// Generate returns a valid (reduced modulo l) Scalar with a distribution +// weighted towards high, low, and edge values. +func (Scalar) Generate(rand *mathrand.Rand, size int) reflect.Value { + var s [32]byte + diceRoll := rand.Intn(100) + switch { + case diceRoll == 0: + case diceRoll == 1: + s = scOneBytes + case diceRoll == 2: + s = [32]byte(scMinusOne.Bytes()) + case diceRoll < 5: + // Generate a low scalar in [0, 2^125). + rand.Read(s[:16]) + s[15] &= (1 << 5) - 1 + case diceRoll < 10: + // Generate a high scalar in [2^252, 2^252 + 2^124). + s[31] = 1 << 4 + rand.Read(s[:16]) + s[15] &= (1 << 4) - 1 + default: + // Generate a valid scalar in [0, l) by returning [0, 2^252) which has a + // negligibly different distribution (the former has a 2^-127.6 chance + // of being out of the latter range). + rand.Read(s[:]) + s[31] &= (1 << 4) - 1 + } + + val := Scalar{} + fiatScalarFromBytes((*[4]uint64)(&val.s), &s) + fiatScalarToMontgomery(&val.s, (*fiatScalarNonMontgomeryDomainFieldElement)(&val.s)) + + return reflect.ValueOf(val) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestScalarGenerate(t *testing.T) { + f := func(sc Scalar) bool { + return isReduced(sc.Bytes()) + } + if err := quick.Check(f, quickCheckConfig(1024)); err != nil { + t.Errorf("generated unreduced scalar: %v", err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestScalarSetCanonicalBytes(t *testing.T) { + f1 := func(in [32]byte, sc Scalar) bool { + // Mask out top 4 bits to guarantee value falls in [0, l). + in[len(in)-1] &= (1 << 4) - 1 + if _, err := sc.SetCanonicalBytes(in[:]); err != nil { + return false + } + repr := sc.Bytes() + return bytes.Equal(in[:], repr) && isReduced(repr) + } + if err := quick.Check(f1, quickCheckConfig(1024)); err != nil { + t.Errorf("failed bytes->scalar->bytes round-trip: %v", err) + } + + f2 := func(sc1, sc2 Scalar) bool { + if _, err := sc2.SetCanonicalBytes(sc1.Bytes()); err != nil { + return false + } + return sc1 == sc2 + } + if err := quick.Check(f2, quickCheckConfig(1024)); err != nil { + t.Errorf("failed scalar->bytes->scalar round-trip: %v", err) + } + + expectReject := func(b []byte) { + t.Helper() + s := scOne + if out, err := s.SetCanonicalBytes(b[:]); err == nil { + t.Errorf("SetCanonicalBytes worked on a non-canonical value") + } else if s != scOne { + t.Errorf("SetCanonicalBytes modified its receiver") + } else if out != nil { + t.Errorf("SetCanonicalBytes did not return nil with an error") + } + } + + b := scMinusOne.Bytes() + b[0] += 1 + expectReject(b) + + b = scMinusOne.Bytes() + b[31] += 1 + expectReject(b) + + b = scMinusOne.Bytes() + b[31] |= 0b1000_0000 + expectReject(b) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestScalarSetUniformBytes(t *testing.T) { + mod, _ := new(big.Int).SetString("27742317777372353535851937790883648493", 10) + mod.Add(mod, new(big.Int).Lsh(big.NewInt(1), 252)) + f := func(in [64]byte, sc Scalar) bool { + sc.SetUniformBytes(in[:]) + repr := sc.Bytes() + if !isReduced(repr) { + return false + } + scBig := bigIntFromLittleEndianBytes(repr[:]) + inBig := bigIntFromLittleEndianBytes(in[:]) + return inBig.Mod(inBig, mod).Cmp(scBig) == 0 + } + if err := quick.Check(f, quickCheckConfig(1024)); err != nil { + t.Error(err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestScalarSetBytesWithClamping(t *testing.T) { + // Generated with libsodium.js 1.0.18 crypto_scalarmult_ed25519_base. + + random := "633d368491364dc9cd4c1bf891b1d59460face1644813240a313e61f2c88216e" + s, _ := new(Scalar).SetBytesWithClamping(decodeHex(random)) + p := new(Point).ScalarBaseMult(s) + want := "1d87a9026fd0126a5736fe1628c95dd419172b5b618457e041c9c861b2494a94" + if got := hex.EncodeToString(p.Bytes()); got != want { + t.Errorf("random: got %q, want %q", got, want) + } + + zero := "0000000000000000000000000000000000000000000000000000000000000000" + s, _ = new(Scalar).SetBytesWithClamping(decodeHex(zero)) + p = new(Point).ScalarBaseMult(s) + want = "693e47972caf527c7883ad1b39822f026f47db2ab0e1919955b8993aa04411d1" + if got := hex.EncodeToString(p.Bytes()); got != want { + t.Errorf("zero: got %q, want %q", got, want) + } + + one := "ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff" + s, _ = new(Scalar).SetBytesWithClamping(decodeHex(one)) + p = new(Point).ScalarBaseMult(s) + want = "12e9a68b73fd5aacdbcaf3e88c46fea6ebedb1aa84eed1842f07f8edab65e3a7" + if got := hex.EncodeToString(p.Bytes()); got != want { + t.Errorf("one: got %q, want %q", got, want) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func bigIntFromLittleEndianBytes(b []byte) *big.Int { + bb := make([]byte, len(b)) + for i := range b { + bb[i] = b[len(b)-i-1] + } + return new(big.Int).SetBytes(bb) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestScalarMultiplyDistributesOverAdd(t *testing.T) { + multiplyDistributesOverAdd := func(x, y, z Scalar) bool { + // Compute t1 = (x+y)*z + var t1 Scalar + t1.Add(&x, &y) + t1.Multiply(&t1, &z) + + // Compute t2 = x*z + y*z + var t2 Scalar + var t3 Scalar + t2.Multiply(&x, &z) + t3.Multiply(&y, &z) + t2.Add(&t2, &t3) + + reprT1, reprT2 := t1.Bytes(), t2.Bytes() + + return t1 == t2 && isReduced(reprT1) && isReduced(reprT2) + } + + if err := quick.Check(multiplyDistributesOverAdd, quickCheckConfig(1024)); err != nil { + t.Error(err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestScalarAddLikeSubNeg(t *testing.T) { + addLikeSubNeg := func(x, y Scalar) bool { + // Compute t1 = x - y + var t1 Scalar + t1.Subtract(&x, &y) + + // Compute t2 = -y + x + var t2 Scalar + t2.Negate(&y) + t2.Add(&t2, &x) + + return t1 == t2 && isReduced(t1.Bytes()) + } + + if err := quick.Check(addLikeSubNeg, quickCheckConfig(1024)); err != nil { + t.Error(err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestScalarNonAdjacentForm(t *testing.T) { + s, _ := (&Scalar{}).SetCanonicalBytes([]byte{ + 0x1a, 0x0e, 0x97, 0x8a, 0x90, 0xf6, 0x62, 0x2d, + 0x37, 0x47, 0x02, 0x3f, 0x8a, 0xd8, 0x26, 0x4d, + 0xa7, 0x58, 0xaa, 0x1b, 0x88, 0xe0, 0x40, 0xd1, + 0x58, 0x9e, 0x7b, 0x7f, 0x23, 0x76, 0xef, 0x09, + }) + + expectedNaf := [256]int8{ + 0, 13, 0, 0, 0, 0, 0, 0, 0, 7, 0, 0, 0, 0, 0, 0, -9, 0, 0, 0, 0, -11, 0, 0, 0, 0, 3, 0, 0, 0, 0, 1, + 0, 0, 0, 0, 9, 0, 0, 0, 0, -5, 0, 0, 0, 0, 0, 0, 3, 0, 0, 0, 0, 11, 0, 0, 0, 0, 11, 0, 0, 0, 0, 0, + -9, 0, 0, 0, 0, 0, -3, 0, 0, 0, 0, 9, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, -1, 0, 0, 0, 0, 0, 9, 0, + 0, 0, 0, -15, 0, 0, 0, 0, -7, 0, 0, 0, 0, -9, 0, 0, 0, 0, 0, 5, 0, 0, 0, 0, 13, 0, 0, 0, 0, 0, -3, 0, + 0, 0, 0, -11, 0, 0, 0, 0, -7, 0, 0, 0, 0, -13, 0, 0, 0, 0, 11, 0, 0, 0, 0, -9, 0, 0, 0, 0, 0, 1, 0, 0, + 0, 0, 0, -15, 0, 0, 0, 0, 1, 0, 0, 0, 0, 7, 0, 0, 0, 0, 0, 0, 0, 0, 5, 0, 0, 0, 0, 0, 13, 0, 0, 0, + 0, 0, 0, 11, 0, 0, 0, 0, 0, 15, 0, 0, 0, 0, 0, -9, 0, 0, 0, 0, 0, 0, 0, -1, 0, 0, 0, 0, 0, 0, 0, 7, + 0, 0, 0, 0, 0, -15, 0, 0, 0, 0, 0, 15, 0, 0, 0, 0, 15, 0, 0, 0, 0, 15, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, + } + + sNaf := s.nonAdjacentForm(5) + + for i := 0; i < 256; i++ { + if expectedNaf[i] != sNaf[i] { + t.Errorf("Wrong digit at position %d, got %d, expected %d", i, sNaf[i], expectedNaf[i]) + } + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +type notZeroScalar Scalar + +func (notZeroScalar) Generate(rand *mathrand.Rand, size int) reflect.Value { + var s Scalar + var isNonZero uint64 + for isNonZero == 0 { + s = Scalar{}.Generate(rand, size).Interface().(Scalar) + fiatScalarNonzero(&isNonZero, (*[4]uint64)(&s.s)) + } + return reflect.ValueOf(notZeroScalar(s)) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestScalarEqual(t *testing.T) { + if scOne.Equal(scMinusOne) == 1 { + t.Errorf("scOne.Equal(&scMinusOne) is true") + } + if scMinusOne.Equal(scMinusOne) == 0 { + t.Errorf("scMinusOne.Equal(&scMinusOne) is false") + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/scalarmult.go b/scalarmult.go new file mode 100644 index 0000000..b461959 --- /dev/null +++ b/scalarmult.go @@ -0,0 +1,237 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519 +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package edwards25519 + +import "sync" + +// basepointTable is a set of 32 affineLookupTables, where table i is generated +// from 256i * basepoint. It is precomputed the first time it's used. +func basepointTable() *[32]affineLookupTable { + basepointTablePrecomp.initOnce.Do(func() { + p := NewGeneratorPoint() + for i := 0; i < 32; i++ { + basepointTablePrecomp.table[i].FromP3(p) + for j := 0; j < 8; j++ { + p.Add(p, p) + } + } + }) + return &basepointTablePrecomp.table +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +var basepointTablePrecomp struct { + table [32]affineLookupTable + initOnce sync.Once +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// ScalarBaseMult sets v = x * B, where B is the canonical generator, and +// returns v. +// +// The scalar multiplication is done in constant time. +func (v *Point) ScalarBaseMult(x *Scalar) *Point { + basepointTable := basepointTable() + + // Write x = sum(x_i * 16^i) so x*B = sum( B*x_i*16^i ) + // as described in the Ed25519 paper + // + // Group even and odd coefficients + // x*B = x_0*16^0*B + x_2*16^2*B + ... + x_62*16^62*B + // + x_1*16^1*B + x_3*16^3*B + ... + x_63*16^63*B + // x*B = x_0*16^0*B + x_2*16^2*B + ... + x_62*16^62*B + // + 16*( x_1*16^0*B + x_3*16^2*B + ... + x_63*16^62*B) + // + // We use a lookup table for each i to get x_i*16^(2*i)*B + // and do four doublings to multiply by 16. + digits := x.signedRadix16() + + multiple := &affineCached{} + tmp1 := &projP1xP1{} + tmp2 := &projP2{} + + // Accumulate the odd components first + v.Set(NewIdentityPoint()) + for i := 1; i < 64; i += 2 { + basepointTable[i/2].SelectInto(multiple, digits[i]) + tmp1.AddAffine(v, multiple) + v.fromP1xP1(tmp1) + } + + // Multiply by 16 + tmp2.FromP3(v) // tmp2 = v in P2 coords + tmp1.Double(tmp2) // tmp1 = 2*v in P1xP1 coords + tmp2.FromP1xP1(tmp1) // tmp2 = 2*v in P2 coords + tmp1.Double(tmp2) // tmp1 = 4*v in P1xP1 coords + tmp2.FromP1xP1(tmp1) // tmp2 = 4*v in P2 coords + tmp1.Double(tmp2) // tmp1 = 8*v in P1xP1 coords + tmp2.FromP1xP1(tmp1) // tmp2 = 8*v in P2 coords + tmp1.Double(tmp2) // tmp1 = 16*v in P1xP1 coords + v.fromP1xP1(tmp1) // now v = 16*(odd components) + + // Accumulate the even components + for i := 0; i < 64; i += 2 { + basepointTable[i/2].SelectInto(multiple, digits[i]) + tmp1.AddAffine(v, multiple) + v.fromP1xP1(tmp1) + } + + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// ScalarMult sets v = x * q, and returns v. +// +// The scalar multiplication is done in constant time. +func (v *Point) ScalarMult(x *Scalar, q *Point) *Point { + checkInitialized(q) + + var table projLookupTable + table.FromP3(q) + + // Write x = sum(x_i * 16^i) + // so x*Q = sum( Q*x_i*16^i ) + // = Q*x_0 + 16*(Q*x_1 + 16*( ... + Q*x_63) ... ) + // <------compute inside out--------- + // + // We use the lookup table to get the x_i*Q values + // and do four doublings to compute 16*Q + digits := x.signedRadix16() + + // Unwrap first loop iteration to save computing 16*identity + multiple := &projCached{} + tmp1 := &projP1xP1{} + tmp2 := &projP2{} + table.SelectInto(multiple, digits[63]) + + v.Set(NewIdentityPoint()) + tmp1.Add(v, multiple) // tmp1 = x_63*Q in P1xP1 coords + for i := 62; i >= 0; i-- { + tmp2.FromP1xP1(tmp1) // tmp2 = (prev) in P2 coords + tmp1.Double(tmp2) // tmp1 = 2*(prev) in P1xP1 coords + tmp2.FromP1xP1(tmp1) // tmp2 = 2*(prev) in P2 coords + tmp1.Double(tmp2) // tmp1 = 4*(prev) in P1xP1 coords + tmp2.FromP1xP1(tmp1) // tmp2 = 4*(prev) in P2 coords + tmp1.Double(tmp2) // tmp1 = 8*(prev) in P1xP1 coords + tmp2.FromP1xP1(tmp1) // tmp2 = 8*(prev) in P2 coords + tmp1.Double(tmp2) // tmp1 = 16*(prev) in P1xP1 coords + v.fromP1xP1(tmp1) // v = 16*(prev) in P3 coords + table.SelectInto(multiple, digits[i]) + tmp1.Add(v, multiple) // tmp1 = x_i*Q + 16*(prev) in P1xP1 coords + } + v.fromP1xP1(tmp1) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// basepointNafTable is the nafLookupTable8 for the basepoint. +// It is precomputed the first time it's used. +func basepointNafTable() *nafLookupTable8 { + basepointNafTablePrecomp.initOnce.Do(func() { + basepointNafTablePrecomp.table.FromP3(NewGeneratorPoint()) + }) + return &basepointNafTablePrecomp.table +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +var basepointNafTablePrecomp struct { + table nafLookupTable8 + initOnce sync.Once +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// VarTimeDoubleScalarBaseMult sets v = a * A + b * B, where B is the canonical +// generator, and returns v. +// +// Execution time depends on the inputs. +func (v *Point) VarTimeDoubleScalarBaseMult(a *Scalar, A *Point, b *Scalar) *Point { + checkInitialized(A) + + // Similarly to the single variable-base approach, we compute + // digits and use them with a lookup table. However, because + // we are allowed to do variable-time operations, we don't + // need constant-time lookups or constant-time digit + // computations. + // + // So we use a non-adjacent form of some width w instead of + // radix 16. This is like a binary representation (one digit + // for each binary place) but we allow the digits to grow in + // magnitude up to 2^{w-1} so that the nonzero digits are as + // sparse as possible. Intuitively, this "condenses" the + // "mass" of the scalar onto sparse coefficients (meaning + // fewer additions). + + basepointNafTable := basepointNafTable() + var aTable nafLookupTable5 + aTable.FromP3(A) + // Because the basepoint is fixed, we can use a wider NAF + // corresponding to a bigger table. + aNaf := a.nonAdjacentForm(5) + bNaf := b.nonAdjacentForm(8) + + // Find the first nonzero coefficient. + i := 255 + for j := i; j >= 0; j-- { + if aNaf[j] != 0 || bNaf[j] != 0 { + break + } + } + + multA := &projCached{} + multB := &affineCached{} + tmp1 := &projP1xP1{} + tmp2 := &projP2{} + tmp2.Zero() + + // Move from high to low bits, doubling the accumulator + // at each iteration and checking whether there is a nonzero + // coefficient to look up a multiple of. + for ; i >= 0; i-- { + tmp1.Double(tmp2) + + // Only update v if we have a nonzero coeff to add in. + if aNaf[i] > 0 { + v.fromP1xP1(tmp1) + aTable.SelectInto(multA, aNaf[i]) + tmp1.Add(v, multA) + } else if aNaf[i] < 0 { + v.fromP1xP1(tmp1) + aTable.SelectInto(multA, -aNaf[i]) + tmp1.Sub(v, multA) + } + + if bNaf[i] > 0 { + v.fromP1xP1(tmp1) + basepointNafTable.SelectInto(multB, bNaf[i]) + tmp1.AddAffine(v, multB) + } else if bNaf[i] < 0 { + v.fromP1xP1(tmp1) + basepointNafTable.SelectInto(multB, -bNaf[i]) + tmp1.SubAffine(v, multB) + } + + tmp2.FromP1xP1(tmp1) + } + + v.fromP2(tmp2) + return v +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/scalarmult_test.go b/scalarmult_test.go new file mode 100644 index 0000000..880c74f --- /dev/null +++ b/scalarmult_test.go @@ -0,0 +1,242 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519 +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package edwards25519 + +import ( + "testing" + "testing/quick" +) + +var ( + // a random scalar generated using dalek. + dalekScalar, _ = (&Scalar{}).SetCanonicalBytes([]byte{219, 106, 114, 9, 174, 249, 155, 89, 69, 203, 201, 93, 92, 116, 234, 187, 78, 115, 103, 172, 182, 98, 62, 103, 187, 136, 13, 100, 248, 110, 12, 4}) + // the above, times the edwards25519 basepoint. + dalekScalarBasepoint, _ = new(Point).SetBytes([]byte{0xf4, 0xef, 0x7c, 0xa, 0x34, 0x55, 0x7b, 0x9f, 0x72, 0x3b, 0xb6, 0x1e, 0xf9, 0x46, 0x9, 0x91, 0x1c, 0xb9, 0xc0, 0x6c, 0x17, 0x28, 0x2d, 0x8b, 0x43, 0x2b, 0x5, 0x18, 0x6a, 0x54, 0x3e, 0x48}) +) + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestScalarMultSmallScalars(t *testing.T) { + var z Scalar + var p Point + p.ScalarMult(&z, B) + if I.Equal(&p) != 1 { + t.Error("0*B != 0") + } + checkOnCurve(t, &p) + + scEight, _ := (&Scalar{}).SetCanonicalBytes([]byte{1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}) + p.ScalarMult(scEight, B) + if B.Equal(&p) != 1 { + t.Error("1*B != 1") + } + checkOnCurve(t, &p) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestScalarMultVsDalek(t *testing.T) { + var p Point + p.ScalarMult(dalekScalar, B) + if dalekScalarBasepoint.Equal(&p) != 1 { + t.Error("Scalar mul does not match dalek") + } + checkOnCurve(t, &p) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestBaseMultVsDalek(t *testing.T) { + var p Point + p.ScalarBaseMult(dalekScalar) + if dalekScalarBasepoint.Equal(&p) != 1 { + t.Error("Scalar mul does not match dalek") + } + checkOnCurve(t, &p) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestVarTimeDoubleBaseMultVsDalek(t *testing.T) { + var p Point + var z Scalar + p.VarTimeDoubleScalarBaseMult(dalekScalar, B, &z) + if dalekScalarBasepoint.Equal(&p) != 1 { + t.Error("VarTimeDoubleScalarBaseMult fails with b=0") + } + checkOnCurve(t, &p) + p.VarTimeDoubleScalarBaseMult(&z, B, dalekScalar) + if dalekScalarBasepoint.Equal(&p) != 1 { + t.Error("VarTimeDoubleScalarBaseMult fails with a=0") + } + checkOnCurve(t, &p) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestScalarMultDistributesOverAdd(t *testing.T) { + scalarMultDistributesOverAdd := func(x, y Scalar) bool { + var z Scalar + z.Add(&x, &y) + var p, q, r, check Point + p.ScalarMult(&x, B) + q.ScalarMult(&y, B) + r.ScalarMult(&z, B) + check.Add(&p, &q) + checkOnCurve(t, &p, &q, &r, &check) + return check.Equal(&r) == 1 + } + + if err := quick.Check(scalarMultDistributesOverAdd, quickCheckConfig(32)); err != nil { + t.Error(err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestScalarMultNonIdentityPoint(t *testing.T) { + // Check whether p.ScalarMult and q.ScalaBaseMult give the same, + // when p and q are originally set to the base point. + + scalarMultNonIdentityPoint := func(x Scalar) bool { + var p, q Point + p.Set(B) + q.Set(B) + + p.ScalarMult(&x, B) + q.ScalarBaseMult(&x) + + checkOnCurve(t, &p, &q) + + return p.Equal(&q) == 1 + } + + if err := quick.Check(scalarMultNonIdentityPoint, quickCheckConfig(32)); err != nil { + t.Error(err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestBasepointTableGeneration(t *testing.T) { + // The basepoint table is 32 affineLookupTables, + // corresponding to (16^2i)*B for table i. + basepointTable := basepointTable() + + tmp1 := &projP1xP1{} + tmp2 := &projP2{} + tmp3 := &Point{} + tmp3.Set(B) + table := make([]affineLookupTable, 32) + for i := 0; i < 32; i++ { + // Build the table + table[i].FromP3(tmp3) + // Assert equality with the hardcoded one + if table[i] != basepointTable[i] { + t.Errorf("Basepoint table %d does not match", i) + } + + // Set p = (16^2)*p = 256*p = 2^8*p + tmp2.FromP3(tmp3) + for j := 0; j < 7; j++ { + tmp1.Double(tmp2) + tmp2.FromP1xP1(tmp1) + } + tmp1.Double(tmp2) + tmp3.fromP1xP1(tmp1) + checkOnCurve(t, tmp3) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestScalarMultMatchesBaseMult(t *testing.T) { + scalarMultMatchesBaseMult := func(x Scalar) bool { + var p, q Point + p.ScalarMult(&x, B) + q.ScalarBaseMult(&x) + checkOnCurve(t, &p, &q) + return p.Equal(&q) == 1 + } + + if err := quick.Check(scalarMultMatchesBaseMult, quickCheckConfig(32)); err != nil { + t.Error(err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestBasepointNafTableGeneration(t *testing.T) { + var table nafLookupTable8 + table.FromP3(B) + + if table != *basepointNafTable() { + t.Error("BasepointNafTable does not match") + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestVarTimeDoubleBaseMultMatchesBaseMult(t *testing.T) { + varTimeDoubleBaseMultMatchesBaseMult := func(x, y Scalar) bool { + var p, q1, q2, check Point + + p.VarTimeDoubleScalarBaseMult(&x, B, &y) + + q1.ScalarBaseMult(&x) + q2.ScalarBaseMult(&y) + check.Add(&q1, &q2) + + checkOnCurve(t, &p, &check, &q1, &q2) + return p.Equal(&check) == 1 + } + + if err := quick.Check(varTimeDoubleBaseMultMatchesBaseMult, quickCheckConfig(32)); err != nil { + t.Error(err) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Benchmarks. + +func BenchmarkScalarBaseMult(b *testing.B) { + var p Point + + for i := 0; i < b.N; i++ { + p.ScalarBaseMult(dalekScalar) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func BenchmarkScalarMult(b *testing.B) { + var p Point + + for i := 0; i < b.N; i++ { + p.ScalarMult(dalekScalar, B) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func BenchmarkVarTimeDoubleScalarBaseMult(b *testing.B) { + var p Point + + for i := 0; i < b.N; i++ { + p.VarTimeDoubleScalarBaseMult(dalekScalar, B, dalekScalar) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/tables.go b/tables.go new file mode 100644 index 0000000..45a0848 --- /dev/null +++ b/tables.go @@ -0,0 +1,162 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519 +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package edwards25519 + +import "crypto/subtle" + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// A dynamic lookup table for variable-base, constant-time scalar muls. +type projLookupTable struct { + points [8]projCached +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// A precomputed lookup table for fixed-base, constant-time scalar muls. +type affineLookupTable struct { + points [8]affineCached +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// A dynamic lookup table for variable-base, variable-time scalar muls. +type nafLookupTable5 struct { + points [8]projCached +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// A precomputed lookup table for fixed-base, variable-time scalar muls. +type nafLookupTable8 struct { + points [64]affineCached +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Constructors. + +// Builds a lookup table at runtime. Fast. +func (v *projLookupTable) FromP3(q *Point) { + // Goal: v.points[i] = (i+1)*Q, i.e., Q, 2Q, ..., 8Q + // This allows lookup of -8Q, ..., -Q, 0, Q, ..., 8Q + v.points[0].FromP3(q) + tmpP3 := Point{} + tmpP1xP1 := projP1xP1{} + for i := 0; i < 7; i++ { + // Compute (i+1)*Q as Q + i*Q and convert to a projCached + // This is needlessly complicated because the API has explicit + // receivers instead of creating stack objects and relying on RVO + v.points[i+1].FromP3(tmpP3.fromP1xP1(tmpP1xP1.Add(q, &v.points[i]))) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// This is not optimised for speed; fixed-base tables should be precomputed. +func (v *affineLookupTable) FromP3(q *Point) { + // Goal: v.points[i] = (i+1)*Q, i.e., Q, 2Q, ..., 8Q + // This allows lookup of -8Q, ..., -Q, 0, Q, ..., 8Q + v.points[0].FromP3(q) + tmpP3 := Point{} + tmpP1xP1 := projP1xP1{} + for i := 0; i < 7; i++ { + // Compute (i+1)*Q as Q + i*Q and convert to affineCached + v.points[i+1].FromP3(tmpP3.fromP1xP1(tmpP1xP1.AddAffine(q, &v.points[i]))) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Builds a lookup table at runtime. Fast. +func (v *nafLookupTable5) FromP3(q *Point) { + // Goal: v.points[i] = (2*i+1)*Q, i.e., Q, 3Q, 5Q, ..., 15Q + // This allows lookup of -15Q, ..., -3Q, -Q, 0, Q, 3Q, ..., 15Q + v.points[0].FromP3(q) + q2 := Point{} + q2.Add(q, q) + tmpP3 := Point{} + tmpP1xP1 := projP1xP1{} + for i := 0; i < 7; i++ { + v.points[i+1].FromP3(tmpP3.fromP1xP1(tmpP1xP1.Add(&q2, &v.points[i]))) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// This is not optimised for speed; fixed-base tables should be precomputed. +func (v *nafLookupTable8) FromP3(q *Point) { + v.points[0].FromP3(q) + q2 := Point{} + q2.Add(q, q) + tmpP3 := Point{} + tmpP1xP1 := projP1xP1{} + for i := 0; i < 63; i++ { + v.points[i+1].FromP3(tmpP3.fromP1xP1(tmpP1xP1.AddAffine(&q2, &v.points[i]))) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Selectors. + +// Set dest to x*Q, where -8 <= x <= 8, in constant time. +func (v *projLookupTable) SelectInto(dest *projCached, x int8) { + // Compute xabs = |x| + xmask := x >> 7 + xabs := uint8((x + xmask) ^ xmask) + + dest.Zero() + for j := 1; j <= 8; j++ { + // Set dest = j*Q if |x| = j + cond := subtle.ConstantTimeByteEq(xabs, uint8(j)) + dest.Select(&v.points[j-1], dest, cond) + } + // Now dest = |x|*Q, conditionally negate to get x*Q + dest.CondNeg(int(xmask & 1)) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Set dest to x*Q, where -8 <= x <= 8, in constant time. +func (v *affineLookupTable) SelectInto(dest *affineCached, x int8) { + // Compute xabs = |x| + xmask := x >> 7 + xabs := uint8((x + xmask) ^ xmask) + + dest.Zero() + for j := 1; j <= 8; j++ { + // Set dest = j*Q if |x| = j + cond := subtle.ConstantTimeByteEq(xabs, uint8(j)) + dest.Select(&v.points[j-1], dest, cond) + } + // Now dest = |x|*Q, conditionally negate to get x*Q + dest.CondNeg(int(xmask & 1)) +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Given odd x with 0 < x < 2^4, return x*Q (in variable time). +func (v *nafLookupTable5) SelectInto(dest *projCached, x int8) { + *dest = v.points[x/2] +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +// Given odd x with 0 < x < 2^7, return x*Q (in variable time). +func (v *nafLookupTable8) SelectInto(dest *affineCached, x int8) { + *dest = v.points[x/2] +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// diff --git a/tables_test.go b/tables_test.go new file mode 100644 index 0000000..19217a6 --- /dev/null +++ b/tables_test.go @@ -0,0 +1,145 @@ +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Source: https://sources.truenas.cloud/code +// Import: sources.truenas.cloud/code/edwards25519 +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +// +// Copyright (c) 2019 The Go Authors. All rights Reserved. +// Use of this source code is goverened by a BSD-style +// license that can be found in the LICENSE file. +// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +package edwards25519 + +import "testing" + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestProjLookupTable(t *testing.T) { + var table projLookupTable + table.FromP3(B) + + var tmp1, tmp2, tmp3 projCached + table.SelectInto(&tmp1, 6) + table.SelectInto(&tmp2, -2) + table.SelectInto(&tmp3, -4) + // Expect T1 + T2 + T3 = identity + + var accP1xP1 projP1xP1 + accP3 := NewIdentityPoint() + + accP1xP1.Add(accP3, &tmp1) + accP3.fromP1xP1(&accP1xP1) + accP1xP1.Add(accP3, &tmp2) + accP3.fromP1xP1(&accP1xP1) + accP1xP1.Add(accP3, &tmp3) + accP3.fromP1xP1(&accP1xP1) + + if accP3.Equal(I) != 1 { + t.Errorf("Consistency check on ProjLookupTable.SelectInto failed! %x %x %x", tmp1, tmp2, tmp3) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestAffineLookupTable(t *testing.T) { + var table affineLookupTable + table.FromP3(B) + + var tmp1, tmp2, tmp3 affineCached + table.SelectInto(&tmp1, 3) + table.SelectInto(&tmp2, -7) + table.SelectInto(&tmp3, 4) + // Expect T1 + T2 + T3 = identity + + var accP1xP1 projP1xP1 + accP3 := NewIdentityPoint() + + accP1xP1.AddAffine(accP3, &tmp1) + accP3.fromP1xP1(&accP1xP1) + accP1xP1.AddAffine(accP3, &tmp2) + accP3.fromP1xP1(&accP1xP1) + accP1xP1.AddAffine(accP3, &tmp3) + accP3.fromP1xP1(&accP1xP1) + + if accP3.Equal(I) != 1 { + t.Errorf("Consistency check on ProjLookupTable.SelectInto failed! %x %x %x", tmp1, tmp2, tmp3) + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestNafLookupTable5(t *testing.T) { + var table nafLookupTable5 + table.FromP3(B) + + var tmp1, tmp2, tmp3, tmp4 projCached + table.SelectInto(&tmp1, 9) + table.SelectInto(&tmp2, 11) + table.SelectInto(&tmp3, 7) + table.SelectInto(&tmp4, 13) + // Expect T1 + T2 = T3 + T4 + + var accP1xP1 projP1xP1 + lhs := NewIdentityPoint() + rhs := NewIdentityPoint() + + accP1xP1.Add(lhs, &tmp1) + lhs.fromP1xP1(&accP1xP1) + accP1xP1.Add(lhs, &tmp2) + lhs.fromP1xP1(&accP1xP1) + + accP1xP1.Add(rhs, &tmp3) + rhs.fromP1xP1(&accP1xP1) + accP1xP1.Add(rhs, &tmp4) + rhs.fromP1xP1(&accP1xP1) + + if lhs.Equal(rhs) != 1 { + t.Errorf("Consistency check on nafLookupTable5 failed") + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// + +func TestNafLookupTable8(t *testing.T) { + var table nafLookupTable8 + table.FromP3(B) + + var tmp1, tmp2, tmp3, tmp4 affineCached + table.SelectInto(&tmp1, 49) + table.SelectInto(&tmp2, 11) + table.SelectInto(&tmp3, 35) + table.SelectInto(&tmp4, 25) + // Expect T1 + T2 = T3 + T4 + + var accP1xP1 projP1xP1 + lhs := NewIdentityPoint() + rhs := NewIdentityPoint() + + accP1xP1.AddAffine(lhs, &tmp1) + lhs.fromP1xP1(&accP1xP1) + accP1xP1.AddAffine(lhs, &tmp2) + lhs.fromP1xP1(&accP1xP1) + + accP1xP1.AddAffine(rhs, &tmp3) + rhs.fromP1xP1(&accP1xP1) + accP1xP1.AddAffine(rhs, &tmp4) + rhs.fromP1xP1(&accP1xP1) + + if lhs.Equal(rhs) != 1 { + t.Errorf("Consistency check on nafLookupTable8 failed") + } +} + +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// +////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////