v3.1.3

RandomX: optimized loading from scratchpad

CHANGELOG.md

@@ -1,343 +1,50 @@
-# v2.99.5-beta
-- [#1066](https://github.com/xmrig/xmrig/issues/1066#issuecomment-518080529) Fixed crash and added error message if pool not ready for RandomX.
-- [#1092](https://github.com/xmrig/xmrig/issues/1092) Fixed crash if wrong CPU affinity used.
-- [#1103](https://github.com/xmrig/xmrig/issues/1103) Improved auto configuration for RandomX for CPUs where L2 cache is limiting factor.
+# v3.1.3
+- [#1180](https://github.com/xmrig/xmrig/issues/1180) Fixed possible duplicated shares after algorithm switching.
+- Fixed wrong config file permissions after write (only gcc builds on recent Windows 10 affected).
+
+# v3.1.2
+- Many RandomX optimizations and fixes.
+  - [#1132](https://github.com/xmrig/xmrig/issues/1132) Fixed build on CentOS 7.
+  - [#1163](https://github.com/xmrig/xmrig/pull/1163) Optimized soft AES code, up to +30% hashrate on CPU without AES support and other optimizations.
+  - [#1166](https://github.com/xmrig/xmrig/pull/1166) Fixed crash when initialize dataset with big threads count (eg 272).
+  - [#1168](https://github.com/xmrig/xmrig/pull/1168) Optimized loading from scratchpad.
+- [#1128](https://github.com/xmrig/xmrig/issues/1128) Fixed CMake 2.8 compatibility.
+
+# v3.1.1
+- [#1133](https://github.com/xmrig/xmrig/issues/1133) Fixed syslog regression.
+- [#1138](https://github.com/xmrig/xmrig/issues/1138) Fixed multiple network bugs.
+- [#1141](https://github.com/xmrig/xmrig/issues/1141) Fixed log in background mode.
+- [#1142](https://github.com/xmrig/xmrig/pull/1142) RandomX hashrate improved by 0.5-1.5% depending on variant and CPU.
+- [#1146](https://github.com/xmrig/xmrig/pull/1146) Fixed race condition in RandomX thread init.
+- [#1148](https://github.com/xmrig/xmrig/pull/1148) Fixed, on Linux linker marking entire executable as having an executable stack.
+- Fixed, for Argon2 algorithms command line options like `--threads` was ignored.
+- Fixed command line options for single pool, free order allowed again.
+
+# v3.1.0
+- [#1107](https://github.com/xmrig/xmrig/issues/1107#issuecomment-522235892) Added Argon2 algorithm family: `argon2/chukwa` and `argon2/wrkz`.
+
+# v3.0.0
+- **[#1111](https://github.com/xmrig/xmrig/pull/1111) Added RandomX (`rx/test`) algorithm for testing and benchmarking.**
+- **[#1036](https://github.com/xmrig/xmrig/pull/1036) Added RandomWOW (`rx/wow`) algorithm for [Wownero](http://wownero.org/).**
+- **[#1050](https://github.com/xmrig/xmrig/pull/1050) Added RandomXL (`rx/loki`) algorithm for [Loki](https://loki.network/).**
+- **[#1077](https://github.com/xmrig/xmrig/issues/1077) Added NUMA support via hwloc**.
+- **Added flexible [multi algorithm](doc/CPU.md) configuration.**
+- **Added unlimited switching between incompatible algorithms, all mining options can be changed in runtime.**
+- [#257](https://github.com/xmrig/xmrig-nvidia/pull/257) New logging subsystem, file and syslog now always without colors.
+- [#314](https://github.com/xmrig/xmrig-proxy/issues/314) Added donate over proxy feature.
+- [#1007](https://github.com/xmrig/xmrig/issues/1007) Old HTTP API backend based on libmicrohttpd, replaced to custom HTTP server (libuv + http_parser).
+- [#1010](https://github.com/xmrig/xmrig/pull/1010#issuecomment-482632107) Added daemon support (solo mining).
+- [#1066](https://github.com/xmrig/xmrig/issues/1066#issuecomment-518080529) Added error message if pool not ready for RandomX.
 - [#1105](https://github.com/xmrig/xmrig/issues/1105) Improved auto configuration for `cn-pico` algorithm.
-- [#1106](https://github.com/xmrig/xmrig/issues/1106) Fixed `hugepages` field in summary API. 
-- Added alternative short format for CPU threads.
-- Changed format for CPU threads with intensity above 1.
-- Name for reference RandomX configuration changed to `rx/text` to avoid potential conflicts in future.
-
-# v2.99.4-beta
-- [#1062](https://github.com/xmrig/xmrig/issues/1062) Fixed 32 bit support. **32 bit is slow and deprecated**.
-- [#1088](https://github.com/xmrig/xmrig/pull/1088) Fixed macOS compilation.
-- [#1095](https://github.com/xmrig/xmrig/pull/1095) Fixed compatibility with hwloc 1.10.x.
-- Optimized RandomX initialization and switching, fixed rare crash when re-initialize dataset.
-- Fixed ARM build with hwloc.
-
-# v2.99.3-beta
-- [#1082](https://github.com/xmrig/xmrig/issues/1082) Fixed hwloc auto configuration on AMD FX CPUs.
+- Added commands `pause` and `resume` via JSON RPC 2.0 API (`POST /json_rpc`).
 - Added command line option `--export-topology` for export hwloc topology to a XML file.
-
-# v2.99.2-beta
-- [#1077](https://github.com/xmrig/xmrig/issues/1077) Added NUMA support via **hwloc**.
-- Fixed miner freeze when switch between RandomX variants.
-- Fixed dataset initialization speed on Linux if thread affinity was used.
-
-# v2.99.1-beta
-- [#1072](https://github.com/xmrig/xmrig/issues/1072) Fixed RandomX `seed_hash` re-initialization.
-
-# v2.99.0-beta
-- [#1050](https://github.com/xmrig/xmrig/pull/1050) Added RandomXL algorithm for [Loki](https://loki.network/), algorithm name used by miner is `randomx/loki` or `rx/loki`.
-- Added [flexible](https://github.com/xmrig/xmrig/blob/evo/doc/CPU.md) multi algorithm configuration.
-- Added unlimited switching between incompatible algorithms, all mining options can be changed in runtime.
 - Breaked backward compatibility with previous configs and command line, `variant` option replaced to `algo`, global option `algo` removed, all CPU related settings moved to `cpu` object.
 - Options `av`, `safe` and `max-cpu-usage` removed.
 - Algorithm `cn/msr` renamed to `cn/fast`.
 - Algorithm `cn/xtl` removed.
 - API endpoint `GET /1/threads` replaced to `GET /2/backends`.
-
-# v2.16.0-beta
-- [#1036](https://github.com/xmrig/xmrig/pull/1036) Added RandomWOW (RandomX with different preferences) algorithm support for [Wownero](http://wownero.org/).
-  - Algorithm name used by miner is `randomx/wow` or `rx/wow`.
-  - Currently runtime algorithm switching NOT supported with other algorithms.
-
-# v2.15.4-beta
 - Added global uptime and extended connection information in API.
-- API now return current algorithm instead of global algorithm specified in config.
-- This version also include all changes from stable version v2.14.4.
+- API now return current algorithm.
 
-# v2.15.3-beta
-- [#1014](https://github.com/xmrig/xmrig/issues/1014) Fixed regression, default value for `algo` option was not applied.
-
-# v2.15.2-beta
-- [#1010](https://github.com/xmrig/xmrig/pull/1010#issuecomment-482632107) Added daemon support (solo mining).
-- [#1012](https://github.com/xmrig/xmrig/pull/1012) Fixed compatibility with clang 9.
-- Config subsystem was rewritten, internally JSON is primary format now.
-- Fixed regression, big HTTP responses was truncated.
-
-# v2.15.1-beta
-- [#1007](https://github.com/xmrig/xmrig/issues/1007) Old HTTP API backend based on libmicrohttpd, replaced to custom HTTP server (libuv + http_parser).
-- [#257](https://github.com/xmrig/xmrig-nvidia/pull/257) New logging subsystem, file and syslog now always without colors.
-
-# v2.15.0-beta
-- [#314](https://github.com/xmrig/xmrig-proxy/issues/314) Added donate over proxy feature.
-  - Added new option `donate-over-proxy`.
-  - Added real graceful exit.
-  
-# v2.14.4
-- [#992](https://github.com/xmrig/xmrig/pull/992)  Fixed compilation with Clang 3.5.
-- [#1012](https://github.com/xmrig/xmrig/pull/1012) Fixed compilation with Clang 9.0.
-- In HTTP API for unknown hashrate now used `null` instead of `0.0`.
-- Fixed MSVC 2019 version detection.
-- Removed obsolete automatic variants.
-
-# v2.14.1
-* [#975](https://github.com/xmrig/xmrig/issues/975) Fixed crash on Linux if double thread mode used.
-
-# v2.14.0
-- **[#969](https://github.com/xmrig/xmrig/pull/969) Added new algorithm `cryptonight/rwz`, short alias `cn/rwz` (also known as CryptoNight ReverseWaltz), for upcoming [Graft](https://www.graft.network/) fork.**
-- **[#931](https://github.com/xmrig/xmrig/issues/931) Added new algorithm `cryptonight/zls`, short alias `cn/zls` for [Zelerius Network](https://zelerius.org) fork.**
-- **[#940](https://github.com/xmrig/xmrig/issues/940) Added new algorithm `cryptonight/double`, short alias `cn/double` (also known as CryptoNight HeavyX), for [X-CASH](https://x-cash.org/).**
-- [#951](https://github.com/xmrig/xmrig/issues/951#issuecomment-469581529) Fixed crash if AVX was disabled on OS level.
-- [#952](https://github.com/xmrig/xmrig/issues/952) Fixed compile error on some Linux.
-- [#957](https://github.com/xmrig/xmrig/issues/957#issuecomment-468890667) Added support for embedded config.
-- [#958](https://github.com/xmrig/xmrig/pull/958) Fixed incorrect user agent on ARM platforms.
-- [#968](https://github.com/xmrig/xmrig/pull/968) Optimized `cn/r` algorithm performance.
-
-# v2.13.1
-- [#946](https://github.com/xmrig/xmrig/pull/946) Optimized software AES implementations for CPUs without hardware AES support. `cn/r`, `cn/wow` up to 2.6 times faster, 4-9% improvements for other algorithms.
-
-# v2.13.0
-- **[#938](https://github.com/xmrig/xmrig/issues/938) Added support for new algorithm `cryptonight/r`, short alias `cn/r` (also known as CryptoNightR or CryptoNight variant 4), for upcoming [Monero](https://www.getmonero.org/) fork on March 9, thanks [@SChernykh](https://github.com/SChernykh).**
-- [#939](https://github.com/xmrig/xmrig/issues/939) Added support for dynamic (runtime) pools reload.
-- [#932](https://github.com/xmrig/xmrig/issues/932) Fixed `cn-pico` hashrate drop, regression since v2.11.0.
-
-# v2.12.0
-- [#929](https://github.com/xmrig/xmrig/pull/929) Added support for new algorithm `cryptonight/wow`, short alias `cn/wow` (also known as CryptonightR), for upcoming [Wownero](http://wownero.org) fork on February 14.
-
-# v2.11.0
-- [#928](https://github.com/xmrig/xmrig/issues/928) Added support for new algorithm `cryptonight/gpu`, short alias `cn/gpu` (original name `cryptonight-gpu`), for upcoming [Ryo currency](https://ryo-currency.com) fork on February 14.
-- [#749](https://github.com/xmrig/xmrig/issues/749) Added support for detect hardware AES in runtime on ARMv8 platforms.
-- [#292](https://github.com/xmrig/xmrig/issues/292) Fixed build on ARMv8 platforms if compiler not support hardware AES.
-
-# v2.10.0
-- [#904](https://github.com/xmrig/xmrig/issues/904) Added new algorithm `cn-pico/trtl` (aliases `cryptonight-turtle`, `cn-trtl`) for upcoming TurtleCoin (TRTL) fork.
-- Default value for option `max-cpu-usage` changed to `100` also this option now deprecated.
-
-# v2.9.4
-- [#913](https://github.com/xmrig/xmrig/issues/913) Fixed Masari (MSR) support (this update required for upcoming fork).
-- [#915](https://github.com/xmrig/xmrig/pull/915) Improved security, JIT memory now read-only after patching.
-
-# v2.9.3
-- [#909](https://github.com/xmrig/xmrig/issues/909) Fixed compile errors on FreeBSD.
-- [#912](https://github.com/xmrig/xmrig/pull/912) Fixed, C++ implementation of `cn/half` was produce up to 13% of invalid hashes.
-
-# v2.9.2
-- [#907](https://github.com/xmrig/xmrig/pull/907) Fixed crash on Linux.
-
-# v2.9.1
-- Restored compatibility with https://stellite.hashvault.pro.
-
-# v2.9.0
-- [#899](https://github.com/xmrig/xmrig/issues/899) Added support for new algorithm `cn/half` for Masari and Stellite forks.
-- [#834](https://github.com/xmrig/xmrig/pull/834) Added ASM optimized code for AMD Bulldozer.
-- [#839](https://github.com/xmrig/xmrig/issues/839) Fixed FreeBSD compile.
-- [#857](https://github.com/xmrig/xmrig/pull/857) Fixed impossible to build for macOS without clang.
-
-# v2.8.3
-- [#813](https://github.com/xmrig/xmrig/issues/813) Fixed critical bug with Minergate pool and variant 2.
-
-# v2.8.1
-- [#768](https://github.com/xmrig/xmrig/issues/768) Fixed build with Visual Studio 2015.
-- [#769](https://github.com/xmrig/xmrig/issues/769) Fixed regression, some ANSI escape sequences was in log with disabled colors.
-- [#777](https://github.com/xmrig/xmrig/issues/777) Better report about pool connection issues. 
-- Simplified checks for ASM auto detection, only AES support necessary.
-- Added missing options to `--help` output.
-
-# v2.8.0
-- **[#753](https://github.com/xmrig/xmrig/issues/753) Added new algorithm [CryptoNight variant 2](https://github.com/xmrig/xmrig/issues/753) for Monero fork, thanks [@SChernykh](https://github.com/SChernykh).**
-  - Added global and per thread option `"asm"` and and command line equivalent.
-- **[#758](https://github.com/xmrig/xmrig/issues/758) Added SSL/TLS support for secure connections to pools.**
-  - Added per pool options `"tls"` and `"tls-fingerprint"` and command line equivalents.
-- [#767](https://github.com/xmrig/xmrig/issues/767) Added config autosave feature, same with GPU miners.  
-- [#245](https://github.com/xmrig/xmrig-proxy/issues/245) Fixed API ID collision when run multiple miners on same machine.
-- [#757](https://github.com/xmrig/xmrig/issues/757) Fixed send buffer overflow.
-
-# v2.6.4
-- [#700](https://github.com/xmrig/xmrig/issues/700) `cryptonight-lite/ipbc` replaced to `cryptonight-heavy/tube` for **Bittube (TUBE)**.
-- Added `cryptonight/rto` (cryptonight variant 1 with IPBC/TUBE mod) variant for **Arto (RTO)** coin.
-- Added `cryptonight/xao` (original cryptonight with bigger iteration count) variant for **Alloy (XAO)** coin.
-- Better variant detection for **nicehash.com** and **minergate.com**.
-- [#692](https://github.com/xmrig/xmrig/issues/692) Added support for specify both algorithm and variant via single `algo` option.
-
-# v2.6.3
-- **Added support for new cryptonight-heavy variant xhv** (`cn-heavy/xhv`) for upcoming Haven Protocol fork.
-- **Added support for new cryptonight variant msr** (`cn/msr`) also known as `cryptonight-fast` for upcoming Masari fork.
-- Added new detailed hashrate report.
-- [#446](https://github.com/xmrig/xmrig/issues/446) Likely fixed SIGBUS error on 32 bit ARM CPUs.
-- [#551](https://github.com/xmrig/xmrig/issues/551) Fixed `cn-heavy` algorithm on ARMv8.
-- [#614](https://github.com/xmrig/xmrig/issues/614) Fixed display issue with huge pages percentage when colors disabled.
-- [#615](https://github.com/xmrig/xmrig/issues/615) Fixed build without libcpuid.
-- [#629](https://github.com/xmrig/xmrig/pull/629) Fixed file logging with non-seekable files.
-- [#672](https://github.com/xmrig/xmrig/pull/672) Reverted back `cryptonight-light` and exit if no valid algorithm specified.
-
-# v2.6.2
- - [#607](https://github.com/xmrig/xmrig/issues/607) Fixed donation bug.
- - [#610](https://github.com/xmrig/xmrig/issues/610) Fixed ARM build.
-
-# v2.6.1
- - [#168](https://github.com/xmrig/xmrig-proxy/issues/168) Added support for [mining algorithm negotiation](https://github.com/xmrig/xmrig-proxy/blob/dev/doc/STRATUM_EXT.md#1-mining-algorithm-negotiation).
- - Added IPBC coin support, base algorithm `cn-lite` variant `ipbc`.
- - [#581](https://github.com/xmrig/xmrig/issues/581) Added support for upcoming Stellite (XTL) fork, base algorithm `cn` variant `xtl`, variant can set now, no need do it after fork.
- - Added support for **rig-id** stratum protocol extensions, compatible with xmr-stak.
- - Changed behavior for option `variant=-1` for `cryptonight`, now variant is `1` by default, if you mine old coins need change `variant` to `0`.
- - A lot of small fixes and better unification with proxy code.
-
-# v2.6.0-beta3
-- [#563](https://github.com/xmrig/xmrig/issues/563) **Added [advanced threads mode](https://github.com/xmrig/xmrig/issues/563), now possible configure each thread individually.**
-- [#255](https://github.com/xmrig/xmrig/issues/563) Low power mode extended to **triple**, **quard** and **penta** modes.
-- [#519](https://github.com/xmrig/xmrig/issues/519) Fixed high donation levels, improved donation start time randomization.
-- [#554](https://github.com/xmrig/xmrig/issues/554) Fixed regression with `print-time` option.
-
-# v2.6.0-beta2
-- Improved performance for `cryptonight v7` especially in double hash mode.
-- [#499](https://github.com/xmrig/xmrig/issues/499) IPv6 disabled for internal HTTP API by default, was causing issues on some systems.
-- Added short aliases for algorithm names: `cn`, `cn-lite` and `cn-heavy`.
-- Fixed regressions (v2.6.0-beta1 affected)
-  - [#494](https://github.com/xmrig/xmrig/issues/494) Command line option `--donate-level` was broken.
-  - [#502](https://github.com/xmrig/xmrig/issues/502) Build without libmicrohttpd was broken.
-  - Fixed nonce calculation for `--av 4` (software AES, double hash) was causing reduction of effective hashrate and rejected shares on nicehash.
-
-# v2.6.0-beta1
- - [#476](https://github.com/xmrig/xmrig/issues/476) **Added Cryptonight-Heavy support for Sumokoin ASIC resistance fork.**
- - HTTP server now runs in main loop, it make possible easy extend API without worry about thread synchronization.
- - Added initial graceful reload support, miner will reload configuration if config file changed, disabled by default until it will be fully implemented and tested.
- - Added API endpoint `PUT /1/config` to update current config.
- - Added API endpoint `GET /1/config` to get current active config.
- - Added API endpoint `GET /1/threads` to get current active threads configuration.
- - API endpoint `GET /` now deprecated, use `GET /1/summary` instead.
- - Added `--api-no-ipv6` and similar config option to disable IPv6 support for HTTP API.
- - Added `--api-no-restricted` to enable full access to api, this option has no effect if `--api-access-token` not specified.
-
-# v2.5.3
-- Fixed critical bug, in some cases miner was can't recovery connection and switch to failover pool, version 2.5.2 affected. If you use v2.6.0-beta3 this issue doesn't concern you.
-- [#499](https://github.com/xmrig/xmrig/issues/499) IPv6 support disabled for internal HTTP API.
-- Added workaround for nicehash.com if you use `cryptonightv7.<region>.nicehash.com` option `variant=1` will be set automatically.
-
-# v2.5.2
-- [#448](https://github.com/xmrig/xmrig/issues/478) Fixed broken reconnect.
-
-# v2.5.1
-- [#454](https://github.com/xmrig/xmrig/issues/454) Fixed build with libmicrohttpd version below v0.9.35.
-- [#456](https://github.com/xmrig/xmrig/issues/459) Verbose errors related to donation pool was not fully silenced.
-- [#459](https://github.com/xmrig/xmrig/issues/459) Fixed regression (version 2.5.0 affected) with connection to **xmr.f2pool.com**.
-
-# v2.5.0
-- [#434](https://github.com/xmrig/xmrig/issues/434) **Added support for Monero v7 PoW, scheduled on April 6.**
-- Added full IPv6 support.
-- Added protocol extension, when use the miner with xmrig-proxy 2.5+ no more need manually specify `nicehash` option.
-- [#123](https://github.com/xmrig/xmrig-proxy/issues/123) Fixed regression (all versions since 2.4 affected) fragmented responses from pool/proxy was parsed incorrectly.
-- [#428](https://github.com/xmrig/xmrig/issues/428) Fixed regression (version 2.4.5 affected) with CPU cache size detection.
-
-# v2.4.5
-- [#324](https://github.com/xmrig/xmrig/pull/324) Fixed build without libmicrohttpd (CMake cache issue).
-- [#341](https://github.com/xmrig/xmrig/issues/341) Fixed wrong exit code and added command line option `--dry-run`.
-- [#385](https://github.com/xmrig/xmrig/pull/385) Up to 20% performance increase for non-AES CPU and fixed Intel Core 2 cache detection.
-
-# v2.4.4
- - Added libmicrohttpd version to --version output.
- - Fixed bug in singal handler, in some cases miner wasn't shutdown properly.
- - Fixed recent MSVC 2017 version detection.
- - [#279](https://github.com/xmrig/xmrig/pull/279) Fixed build on some macOS versions.
-
-# v2.4.3
- - [#94](https://github.com/xmrig/xmrig/issues/94#issuecomment-342019257) [#216](https://github.com/xmrig/xmrig/issues/216) Added **ARMv8** and **ARMv7** support. Hardware AES supported, thanks [Imran Yusuff](https://github.com/imranyusuff).
- - [#157](https://github.com/xmrig/xmrig/issues/157) [#196](https://github.com/xmrig/xmrig/issues/196) Fixed Linux compile issues.
- - [#184](https://github.com/xmrig/xmrig/issues/184) Fixed cache size detection for CPUs with disabled Hyper-Threading.
- - [#200](https://github.com/xmrig/xmrig/issues/200) In some cases miner was doesn't write log to stdout.
-
-# v2.4.2
- - [#60](https://github.com/xmrig/xmrig/issues/60) Added FreeBSD support, thanks [vcambur](https://github.com/vcambur).
- - [#153](https://github.com/xmrig/xmrig/issues/153) Fixed issues with dwarfpool.com.
- 
-# v2.4.1
-  - [#147](https://github.com/xmrig/xmrig/issues/147) Fixed comparability with monero-stratum.
-
-# v2.4.0
- - Added [HTTP API](https://github.com/xmrig/xmrig/wiki/API).
- - Added comments support in config file.
- - libjansson replaced to rapidjson.
- - [#98](https://github.com/xmrig/xmrig/issues/98) Ignore `keepalive` option with minergate.com and nicehash.com.
- - [#101](https://github.com/xmrig/xmrig/issues/101) Fixed MSVC 2017 (15.3) compile time version detection.
- - [#108](https://github.com/xmrig/xmrig/issues/108) Silently ignore invalid values for `donate-level` option.
- - [#111](https://github.com/xmrig/xmrig/issues/111) Fixed build without AEON support.
- 
-# v2.3.1
-- [#68](https://github.com/xmrig/xmrig/issues/68) Fixed compatibility with Docker containers, was nothing print on console.
-
-# v2.3.0
-- Added `--cpu-priority` option (0 idle, 2 normal to 5 highest).
-- Added `--user-agent` option, to set custom user-agent string for pool. For example `cpuminer-multi/0.1`.
-- Added `--no-huge-pages` option, to disable huge pages support.
-- [#62](https://github.com/xmrig/xmrig/issues/62) Don't send the login to the dev pool.
-- Force reconnect if pool block miner IP address. helps switch to backup pool.
-- Fixed: failed open default config file if path contains non English characters.
-- Fixed: error occurred if try use unavailable stdin or stdout, regression since version 2.2.0.
-- Fixed: message about huge pages support successfully enabled on Windows was not shown in release builds.
-
-# v2.2.1
-- Fixed [terminal issues](https://github.com/xmrig/xmrig-proxy/issues/2#issuecomment-319914085) after exit on Linux and OS X.
-
-# v2.2.0
-- [#46](https://github.com/xmrig/xmrig/issues/46) Restored config file support. Now possible use multiple config files and combine with command line options also added support for default config.
-- Improved colors support on Windows, now used uv_tty, legacy code removed.
-- QuickEdit Mode now disabled on Windows.
-- Added interactive commands in console window:: **h**ashrate, **p**ause, **r**esume.
-- Fixed autoconf mode for AMD FX CPUs.
-
-# v2.1.0
-- [#40](https://github.com/xmrig/xmrig/issues/40)
-Improved miner shutdown, fixed crash on exit for Linux and OS X.
-- Fixed, login request was contain malformed JSON if username or password has some special characters for example `\`. 
-- [#220](https://github.com/fireice-uk/xmr-stak-cpu/pull/220) Better support for Round Robin DNS, IP address now always chosen randomly instead of stuck on first one.
-- Changed donation address, new [xmrig-proxy](https://github.com/xmrig/xmrig-proxy) is coming soon.
-
-# v2.0.2
-- Better deal with possible duplicate jobs from pool, show warning and ignore duplicates.
-- For Windows builds libuv updated to version 1.13.1 and gcc to 7.1.0.
-
-# v2.0.1
- - [#27](https://github.com/xmrig/xmrig/issues/27) Fixed possibility crash on 32bit systems.
-
-# v2.0.0
- - Option `--backup-url` removed, instead now possibility specify multiple pools for example: `-o example1.com:3333 -u user1 -p password1 -k -o example2.com:5555 -u user2 -o example3.com:4444 -u user3`
- - [#15](https://github.com/xmrig/xmrig/issues/15) Added option `-l, --log-file=FILE` to write log to file.
- - [#15](https://github.com/xmrig/xmrig/issues/15) Added option `-S, --syslog` to use syslog for logging, Linux only.
- - [#18](https://github.com/xmrig/xmrig/issues/18) Added nice messages for accepted/rejected shares with diff and network latency.
- - [#20](https://github.com/xmrig/xmrig/issues/20) Fixed `--cpu-affinity` for more than 32 threads.
- - Fixed Windows XP support.
- - Fixed regression, option `--no-color` was not fully disable colored output.
- - Show resolved pool IP address in miner output.
- 
-# v1.0.1
-- Fix broken software AES implementation, app has crashed if CPU not support AES-NI, only version 1.0.0 affected.
-
-# v1.0.0
-- Miner complete rewritten in C++ with libuv.
-- This version should be fully compatible (except config file) with previos versions, many new nice features will come in next versions.
-- This is still beta. If you found regression, stability or perfomance issues or have an idea for new feature please fell free to open new [issue](https://github.com/xmrig/xmrig/issues/new).
-- Added new option `--print-time=N`, print hashrate report every N seconds.
-- New hashrate reports, by default every 60 secons.
-- Added Microsoft Visual C++ 2015 and 2017 support.
-- Removed dependency on libcurl.
-- To compile this version from source please switch to [dev](https://github.com/xmrig/xmrig/tree/dev) branch.
-
-# v0.8.2
-- Fixed L2 cache size detection for AMD CPUs (Bulldozer/Piledriver/Steamroller/Excavator architecture).
-
-# v0.8.2
-- Fixed L2 cache size detection for AMD CPUs (Bulldozer/Piledriver/Steamroller/Excavator architecture).
-- Fixed gcc 7.1 support.
-
-# v0.8.1
-- Added nicehash support, detects automaticaly by pool URL, for example `cryptonight.eu.nicehash.com:3355` or manually via option `--nicehash`.
-
-# v0.8.0
-- Added double hash mode, also known as lower power mode. `--av=2` and `--av=4`.
-- Added smart automatic CPU configuration. Default threads count now depends on size of the L3 cache of CPU.
-- Added CryptoNight-Lite support for AEON `-a cryptonight-lite`.
-- Added `--max-cpu-usage` option for auto CPU configuration mode.
-- Added `--safe` option for adjust threads and algorithm variations to current CPU.
-- No more manual steps to enable huge pages on Windows. XMRig will do it automatically.
-- Removed BMI2 algorithm variation.
-- Removed default pool URL.
-
-# v0.6.0
-- Added automatic cryptonight self test.
-- New software AES algorithm variation. Will be automatically selected if cpu not support AES-NI.
-- Added 32 bit builds.
-- Documented [algorithm variations](https://github.com/xmrig/xmrig#algorithm-variations).
-
-# v0.5.0
-- Initial public release.
+# Previous versions
+[doc/CHANGELOG_OLD.md](doc/CHANGELOG_OLD.md)

CMakeLists.txt

@@ -1,14 +1,15 @@
 cmake_minimum_required(VERSION 2.8)
 project(xmrig)
 
-option(WITH_LIBCPUID        "Use Libcpuid" ON)
-option(WITH_HWLOC           "Use hwloc" ON)
-option(WITH_CN_LITE         "CryptoNight-Lite support" ON)
-option(WITH_CN_HEAVY        "CryptoNight-Heavy support" ON)
-option(WITH_CN_PICO         "CryptoNight-Pico support" ON)
-option(WITH_CN_GPU          "CryptoNight-GPU support" ON)
-option(WITH_RANDOMX         "RandomX support" ON)
-option(WITH_HTTP            "HTTP protocol support (client/server)" ON)
+option(WITH_LIBCPUID        "Enable libcpuid support" ON)
+option(WITH_HWLOC           "Enable hwloc support" ON)
+option(WITH_CN_LITE         "Enable CryptoNight-Lite algorithms family" ON)
+option(WITH_CN_HEAVY        "Enable CryptoNight-Heavy algorithms family" ON)
+option(WITH_CN_PICO         "Enable CryptoNight-Pico algorithm" ON)
+option(WITH_CN_GPU          "Enable CryptoNight-GPU algorithm" ON)
+option(WITH_RANDOMX         "Enable RandomX algorithms family" ON)
+option(WITH_ARGON2          "Enable Argon2 algorithms family" ON)
+option(WITH_HTTP            "Enable HTTP protocol support (client/server)" ON)
 option(WITH_DEBUG_LOG       "Enable debug log output" OFF)
 option(WITH_TLS             "Enable OpenSSL support" ON)
 option(WITH_ASM             "Enable ASM PoW implementations" ON)
@@ -32,7 +33,6 @@ set(HEADERS
     "${HEADERS_BASE}"
     "${HEADERS_BASE_HTTP}"
     "${HEADERS_BACKEND}"
-    src/api/interfaces/IApiListener.h
     src/App.h
     src/core/config/Config_default.h
     src/core/config/Config_platform.h
@@ -152,66 +152,9 @@ add_definitions(/DUNICODE)
 
 find_package(UV REQUIRED)
 
-if (WITH_RANDOMX)
-    include_directories(src/crypto/randomx)
-    add_definitions(/DXMRIG_ALGO_RANDOMX)
-    set(SOURCES_CRYPTO
-        "${SOURCES_CRYPTO}"
-        src/crypto/randomx/aes_hash.cpp
-        src/crypto/randomx/allocator.cpp
-        src/crypto/randomx/argon2_core.c
-        src/crypto/randomx/argon2_ref.c
-        src/crypto/randomx/blake2_generator.cpp
-        src/crypto/randomx/blake2/blake2b.c
-        src/crypto/randomx/bytecode_machine.cpp
-        src/crypto/randomx/dataset.cpp
-        src/crypto/randomx/instructions_portable.cpp
-        src/crypto/randomx/randomx.cpp
-        src/crypto/randomx/reciprocal.c
-        src/crypto/randomx/soft_aes.cpp
-        src/crypto/randomx/superscalar.cpp
-        src/crypto/randomx/virtual_machine.cpp
-        src/crypto/randomx/virtual_memory.cpp
-        src/crypto/randomx/vm_compiled_light.cpp
-        src/crypto/randomx/vm_compiled.cpp
-        src/crypto/randomx/vm_interpreted_light.cpp
-        src/crypto/randomx/vm_interpreted.cpp
-        src/crypto/rx/Rx.cpp
-        src/crypto/rx/Rx.h
-        src/crypto/rx/RxAlgo.cpp
-        src/crypto/rx/RxAlgo.h
-        src/crypto/rx/RxCache.cpp
-        src/crypto/rx/RxCache.h
-        src/crypto/rx/RxConfig.cpp
-        src/crypto/rx/RxConfig.h
-        src/crypto/rx/RxDataset.cpp
-        src/crypto/rx/RxDataset.h
-        src/crypto/rx/RxVm.cpp
-        src/crypto/rx/RxVm.h
-    )
-    if (NOT ARCH_ID)
-        set(ARCH_ID ${CMAKE_HOST_SYSTEM_PROCESSOR})
-    endif()
-    if (CMAKE_C_COMPILER_ID MATCHES MSVC)
-        enable_language(ASM_MASM)
-        list(APPEND SOURCES_CRYPTO
-             src/crypto/randomx/jit_compiler_x86_static.asm
-             src/crypto/randomx/jit_compiler_x86.cpp
-            )
-    elseif (NOT XMRIG_ARM AND CMAKE_SIZEOF_VOID_P EQUAL 8)
-        list(APPEND SOURCES_CRYPTO
-             src/crypto/randomx/jit_compiler_x86_static.S
-             src/crypto/randomx/jit_compiler_x86.cpp
-            )
-        # cheat because cmake and ccache hate each other
-        set_property(SOURCE src/crypto/randomx/jit_compiler_x86_static.S PROPERTY LANGUAGE C)
-    endif()
-else()
-    remove_definitions(/DXMRIG_ALGO_RANDOMX)
-endif()
-
 include(cmake/flags.cmake)
-
+include(cmake/randomx.cmake)
+include(cmake/argon2.cmake)
 include(cmake/OpenSSL.cmake)
 include(cmake/asm.cmake)
 include(cmake/cn-gpu.cmake)
@@ -232,22 +175,6 @@ if (WITH_EMBEDDED_CONFIG)
     add_definitions(/DXMRIG_FEATURE_EMBEDDED_CONFIG)
 endif()
 
-if (WITH_HTTP)
-    set(HTTP_SOURCES
-        src/api/Api.cpp
-        src/api/Api.h
-        src/api/Httpd.cpp
-        src/api/Httpd.h
-        src/api/interfaces/IApiRequest.h
-        src/api/requests/ApiRequest.cpp
-        src/api/requests/ApiRequest.h
-        src/api/requests/HttpApiRequest.cpp
-        src/api/requests/HttpApiRequest.h
-        )
-else()
-    set(HTTP_SOURCES "")
-endif()
-
 include_directories(src)
 include_directories(src/3rdparty)
 include_directories(${UV_INCLUDE_DIR})
@@ -260,5 +187,5 @@ if (WITH_DEBUG_LOG)
     add_definitions(/DAPP_DEBUG)
 endif()
 
-add_executable(${CMAKE_PROJECT_NAME} ${HEADERS} ${SOURCES} ${SOURCES_OS} ${SOURCES_CPUID} ${HEADERS_CRYPTO} ${SOURCES_CRYPTO} ${SOURCES_SYSLOG} ${HTTP_SOURCES} ${TLS_SOURCES} ${XMRIG_ASM_SOURCES} ${CN_GPU_SOURCES})
-target_link_libraries(${CMAKE_PROJECT_NAME} ${XMRIG_ASM_LIBRARY} ${OPENSSL_LIBRARIES} ${UV_LIBRARIES} ${EXTRA_LIBS} ${CPUID_LIB})
+add_executable(${CMAKE_PROJECT_NAME} ${HEADERS} ${SOURCES} ${SOURCES_OS} ${SOURCES_CPUID} ${HEADERS_CRYPTO} ${SOURCES_CRYPTO} ${SOURCES_SYSLOG} ${TLS_SOURCES} ${XMRIG_ASM_SOURCES} ${CN_GPU_SOURCES})
+target_link_libraries(${CMAKE_PROJECT_NAME} ${XMRIG_ASM_LIBRARY} ${OPENSSL_LIBRARIES} ${UV_LIBRARIES} ${EXTRA_LIBS} ${CPUID_LIB} ${ARGON2_LIBRARY})

README.md

@@ -7,136 +7,79 @@
 [![GitHub stars](https://img.shields.io/github/stars/xmrig/xmrig.svg)](https://github.com/xmrig/xmrig/stargazers)
 [![GitHub forks](https://img.shields.io/github/forks/xmrig/xmrig.svg)](https://github.com/xmrig/xmrig/network)
 
-XMRig is a high performance Monero (XMR) CPU miner, with official support for Windows.
-Originally based on cpuminer-multi with heavy optimizations/rewrites and removing a lot of legacy code, since version 1.0.0 completely rewritten from scratch on C++.
+XMRig is a high performance RandomX and CryptoNight CPU miner, with official support for Windows.
 
 * This is the **CPU-mining** version, there is also a [NVIDIA GPU version](https://github.com/xmrig/xmrig-nvidia) and [AMD GPU version]( https://github.com/xmrig/xmrig-amd).
-* [Roadmap](https://github.com/xmrig/xmrig/issues/106) for next releases.
 
-<img src="http://i.imgur.com/Ymumes5.png" width="670" >
+<img src="doc/screenshot.png" width="808" >
 
 #### Table of contents
-* [Features](#features)
 * [Download](#download)
 * [Usage](#usage)
-* [Algorithm variations](#algorithm-variations)
 * [Build](https://github.com/xmrig/xmrig/wiki/Build)
-* [Common Issues](#common-issues)
-* [Other information](#other-information)
 * [Donations](#donations)
-* [Release checksums](#release-checksums)
 * [Contacts](#contacts)
 
-## Features
-* High performance.
-* Official Windows support.
-* Small Windows executable, without dependencies.
-* x86/x64 support.
-* Support for backup (failover) mining server.
-* keepalived support.
-* Command line options compatible with cpuminer.
-* CryptoNight-Lite support for AEON.
-* Smart automatic [CPU configuration](https://github.com/xmrig/xmrig/wiki/Threads).
-* Nicehash support
-* It's open source software.
-
 ## Download
 * Binary releases: https://github.com/xmrig/xmrig/releases
 * Git tree: https://github.com/xmrig/xmrig.git
   * Clone with `git clone https://github.com/xmrig/xmrig.git` :hammer: [Build instructions](https://github.com/xmrig/xmrig/wiki/Build).
 
 ## Usage
-Use [config.xmrig.com](https://config.xmrig.com/xmrig) to generate, edit or share configurations.
+Preferend way to configure miner is [JSON config file](src/config.json) as more flexible and human frendly, command line interface not cover all features, for example mining profiles for different algorithms. Most impotant options can be changed in runtime without miner restart by editing config or via API.
 
 ### Options
 ```
-  -a, --algo=ALGO          specify the algorithm to use
-                             cryptonight
-                             cryptonight-lite
-                             cryptonight-heavy
-  -o, --url=URL            URL of mining server
-  -O, --userpass=U:P       username:password pair for mining server
-  -u, --user=USERNAME      username for mining server
-  -p, --pass=PASSWORD      password for mining server
-      --rig-id=ID          rig identifier for pool-side statistics (needs pool support)
-  -t, --threads=N          number of miner threads
-  -v, --av=N               algorithm variation, 0 auto select
-  -k, --keepalive          send keepalived packet for prevent timeout (needs pool support)
-      --nicehash           enable nicehash.com support
-      --tls                enable SSL/TLS support (needs pool support)
-      --tls-fingerprint=F  pool TLS certificate fingerprint, if set enable strict certificate pinning
-  -r, --retries=N          number of times to retry before switch to backup server (default: 5)
-  -R, --retry-pause=N      time to pause between retries (default: 5)
-      --cpu-affinity       set process affinity to CPU core(s), mask 0x3 for cores 0 and 1
-      --cpu-priority       set process priority (0 idle, 2 normal to 5 highest)
-      --no-huge-pages      disable huge pages support
-      --no-color           disable colored output
-      --variant            algorithm PoW variant
-      --donate-level=N     donate level, default 5% (5 minutes in 100 minutes)
-      --user-agent         set custom user-agent string for pool
-  -B, --background         run the miner in the background
-  -c, --config=FILE        load a JSON-format configuration file
-  -l, --log-file=FILE      log all output to a file
-  -S, --syslog             use system log for output messages
-      --max-cpu-usage=N    maximum CPU usage for automatic threads mode (default 75)
-      --safe               safe adjust threads and av settings for current CPU
-      --asm=ASM            ASM code for cn/2, possible values: auto, none, intel, ryzen.
-      --print-time=N       print hashrate report every N seconds
-      --api-port=N         port for the miner API
-      --api-access-token=T access token for API
-      --api-worker-id=ID   custom worker-id for API
-      --api-id=ID          custom instance ID for API
-      --api-ipv6           enable IPv6 support for API
-      --api-no-restricted  enable full remote access (only if API token set)
-      --dry-run            test configuration and exit
-  -h, --help               display this help and exit
-  -V, --version            output version information and exit
+  -a, --algo=ALGO               specify the algorithm to use
+                                  cn/r, cn/2, cn/1, cn/0, cn/double, cn/half, cn/fast,
+                                  cn/rwz, cn/zls, cn/xao, cn/rto, cn/gpu,
+                                  cn-lite/1,
+                                  cn-heavy/xhv, cn-heavy/tube, cn-heavy/0,
+                                  cn-pico,
+                                  rx/wow, rx/loki
+  -o, --url=URL                 URL of mining server
+  -O, --userpass=U:P            username:password pair for mining server
+  -u, --user=USERNAME           username for mining server
+  -p, --pass=PASSWORD           password for mining server
+      --rig-id=ID               rig identifier for pool-side statistics (needs pool support)
+  -t, --threads=N               number of miner threads
+  -v, --av=N                    algorithm variation, 0 auto select
+  -k, --keepalive               send keepalived packet for prevent timeout (needs pool support)
+      --nicehash                enable nicehash.com support
+      --tls                     enable SSL/TLS support (needs pool support)
+      --tls-fingerprint=F       pool TLS certificate fingerprint, if set enable strict certificate pinning
+      --daemon                  use daemon RPC instead of pool for solo mining
+      --daemon-poll-interval=N  daemon poll interval in milliseconds (default: 1000)
+  -r, --retries=N               number of times to retry before switch to backup server (default: 5)
+  -R, --retry-pause=N           time to pause between retries (default: 5)
+      --cpu-affinity            set process affinity to CPU core(s), mask 0x3 for cores 0 and 1
+      --cpu-priority            set process priority (0 idle, 2 normal to 5 highest)
+      --no-huge-pages           disable huge pages support
+      --no-color                disable colored output
+      --donate-level=N          donate level, default 5% (5 minutes in 100 minutes)
+      --user-agent              set custom user-agent string for pool
+  -B, --background              run the miner in the background
+  -c, --config=FILE             load a JSON-format configuration file
+  -l, --log-file=FILE           log all output to a file
+      --asm=ASM                 ASM optimizations, possible values: auto, none, intel, ryzen, bulldozer.
+      --print-time=N            print hashrate report every N seconds
+      --api-worker-id=ID        custom worker-id for API
+      --api-id=ID               custom instance ID for API
+      --http-enabled            enable HTTP API
+      --http-host=HOST          bind host for HTTP API (default: 127.0.0.1)
+      --http-port=N             bind port for HTTP API
+      --http-access-token=T     access token for HTTP API
+      --http-no-restricted      enable full remote access to HTTP API (only if access token set)
+      --randomx-init=N          threads count to initialize RandomX dataset
+      --randomx-no-numa         disable NUMA support for RandomX
+      --export-topology         export hwloc topology to a XML file and exit
+      --dry-run                 test configuration and exit
+  -h, --help                    display this help and exit
+  -V, --version                 output version information and exit
 ```
 
-Also you can use configuration via config file, default name **config.json**. Some options available only via config file: [`autosave`](https://github.com/xmrig/xmrig/issues/767), [`hw-aes`](https://github.com/xmrig/xmrig/issues/563). `watch` option currently not implemented in miners only in proxy.
-
-## Algorithm variations
-
-- `av` option used for automatic and simple threads mode (when you specify only threads count).
-- For [advanced threads mode](https://github.com/xmrig/xmrig/issues/563) each thread configured individually and `av` option not used.
-
-| av | Hashes per round | Hardware AES |
-|----|------------------|--------------|
-| 1  | 1 (Single)       | yes          |
-| 2  | 2 (Double)       | yes          |
-| 3  | 1 (Single)       | no           |
-| 4  | 2 (Double)       | no           |
-| 5  | 3 (Triple)       | yes          |
-| 6  | 4 (Quard)        | yes          |
-| 7  | 5 (Penta)        | yes          |
-| 8  | 3 (Triple)       | no           |
-| 9  | 4 (Quard)        | no           |
-| 10 | 5 (Penta)        | no           |
-
-## Common Issues
-### HUGE PAGES unavailable
-* Run XMRig as Administrator.
-* Since version 0.8.0 XMRig automatically enables SeLockMemoryPrivilege for current user, but reboot or sign out still required. [Manual instruction](https://msdn.microsoft.com/en-gb/library/ms190730.aspx).
-
-## Other information
-* No HTTP support, only stratum protocol support.
-* Default donation 5% (5 minutes in 100 minutes) can be reduced to 1% via option `donate-level`.
-
-
-### CPU mining performance
-* **Intel i7-7700** - 307 H/s (4 threads)
-* **AMD Ryzen 7 1700X** - 560 H/s (8 threads)
-
-Please note performance is highly dependent on system load. The numbers above are obtained on an idle system. Tasks heavily using a processor cache, such as video playback, can greatly degrade hashrate. Optimal number of threads depends on the size of the L3 cache of a processor, 1 thread requires 2 MB of cache.
-
-### Maximum performance checklist
-* Idle operating system.
-* Do not exceed optimal thread count.
-* Use modern CPUs with AES-NI instruction set.
-* Try setup optimal cpu affinity.
-* Enable fast memory (Large/Huge pages).
-
 ## Donations
+* Default donation 5% (5 minutes in 100 minutes) can be reduced to 1% via option `donate-level` or disabled in source code.
 * XMR: `48edfHu7V9Z84YzzMa6fUueoELZ9ZRXq9VetWzYGzKt52XU5xvqgzYnDK9URnRoJMk1j8nLwEVsaSWJ4fhdUyZijBGUicoD`
 * BTC: `1P7ujsXeX7GxQwHNnJsRMgAdNkFZmNVqJT`
 

cmake/argon2.cmake

@@ -0,0 +1,18 @@
+if (WITH_ARGON2)
+    add_definitions(/DXMRIG_ALGO_ARGON2)
+
+    list(APPEND HEADERS_CRYPTO
+        src/crypto/argon2/Hash.h
+        src/crypto/argon2/Impl.h
+    )
+
+    list(APPEND SOURCES_CRYPTO
+        src/crypto/argon2/Impl.cpp
+    )
+
+    add_subdirectory(src/3rdparty/argon2)
+    set(ARGON2_LIBRARY argon2)
+else()
+    remove_definitions(/DXMRIG_ALGO_ARGON2)
+    set(ARGON2_LIBRARY "")
+endif()

cmake/flags.cmake

@@ -2,6 +2,9 @@ set(CMAKE_CXX_STANDARD_REQUIRED ON)
 set(CMAKE_CXX_EXTENSIONS OFF)
 set(CMAKE_CXX_STANDARD 11)
 
+set(CMAKE_C_STANDARD 99)
+set(CMAKE_C_STANDARD_REQUIRED ON)
+
 if ("${CMAKE_BUILD_TYPE}" STREQUAL "")
     set(CMAKE_BUILD_TYPE Release)
 endif()
@@ -13,11 +16,10 @@ endif()
 include(CheckSymbolExists)
 
 if (CMAKE_CXX_COMPILER_ID MATCHES GNU)
-
     set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -Wall -Wno-strict-aliasing")
     set(CMAKE_C_FLAGS_RELEASE "${CMAKE_C_FLAGS_RELEASE} -Ofast")
 
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wall -fexceptions -fno-rtti -Wno-class-memaccess")
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wall -fexceptions -fno-rtti -Wno-strict-aliasing -Wno-class-memaccess")
     set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -Ofast -s")
 
     if (XMRIG_ARMv8)
@@ -46,6 +48,7 @@ if (CMAKE_CXX_COMPILER_ID MATCHES GNU)
     add_definitions(/D_GNU_SOURCE)
 
     if (${CMAKE_VERSION} VERSION_LESS "3.1.0")
+        set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -std=c99")
         set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11")
     endif()
 

cmake/randomx.cmake

@@ -0,0 +1,57 @@
+if (WITH_RANDOMX)
+    add_definitions(/DXMRIG_ALGO_RANDOMX)
+
+    list(APPEND HEADERS_CRYPTO
+        src/crypto/rx/Rx.h
+        src/crypto/rx/RxAlgo.h
+        src/crypto/rx/RxCache.h
+        src/crypto/rx/RxConfig.h
+        src/crypto/rx/RxDataset.h
+        src/crypto/rx/RxVm.h
+    )
+
+    list(APPEND SOURCES_CRYPTO
+        src/crypto/randomx/aes_hash.cpp
+        src/crypto/randomx/allocator.cpp
+        src/crypto/randomx/argon2_core.c
+        src/crypto/randomx/argon2_ref.c
+        src/crypto/randomx/blake2_generator.cpp
+        src/crypto/randomx/blake2/blake2b.c
+        src/crypto/randomx/bytecode_machine.cpp
+        src/crypto/randomx/dataset.cpp
+        src/crypto/randomx/instructions_portable.cpp
+        src/crypto/randomx/randomx.cpp
+        src/crypto/randomx/reciprocal.c
+        src/crypto/randomx/soft_aes.cpp
+        src/crypto/randomx/superscalar.cpp
+        src/crypto/randomx/virtual_machine.cpp
+        src/crypto/randomx/virtual_memory.cpp
+        src/crypto/randomx/vm_compiled_light.cpp
+        src/crypto/randomx/vm_compiled.cpp
+        src/crypto/randomx/vm_interpreted_light.cpp
+        src/crypto/randomx/vm_interpreted.cpp
+        src/crypto/rx/Rx.cpp
+        src/crypto/rx/RxAlgo.cpp
+        src/crypto/rx/RxCache.cpp
+        src/crypto/rx/RxConfig.cpp
+        src/crypto/rx/RxDataset.cpp
+        src/crypto/rx/RxVm.cpp
+    )
+
+    if (CMAKE_C_COMPILER_ID MATCHES MSVC)
+        enable_language(ASM_MASM)
+        list(APPEND SOURCES_CRYPTO
+             src/crypto/randomx/jit_compiler_x86_static.asm
+             src/crypto/randomx/jit_compiler_x86.cpp
+            )
+    elseif (NOT XMRIG_ARM AND CMAKE_SIZEOF_VOID_P EQUAL 8)
+        list(APPEND SOURCES_CRYPTO
+             src/crypto/randomx/jit_compiler_x86_static.S
+             src/crypto/randomx/jit_compiler_x86.cpp
+            )
+        # cheat because cmake and ccache hate each other
+        set_property(SOURCE src/crypto/randomx/jit_compiler_x86_static.S PROPERTY LANGUAGE C)
+    endif()
+else()
+    remove_definitions(/DXMRIG_ALGO_RANDOMX)
+endif()

doc/ALGORITHMS.md

@@ -23,28 +23,30 @@ Since version 3 mining [algorithm](#algorithm-names) should specified for each p
  
  ## Algorithm names
 
-| Name            | Memory | Notes                                                                                |
-|-----------------|--------|--------------------------------------------------------------------------------------|
-| `cn/0`          |   2 MB | CryptoNight (original)                                                               |
-| `cn/1`          |   2 MB | CryptoNight variant 1 also known as `Monero7` and `CryptoNightV7`.                   |
-| `cn/2`          |   2 MB | CryptoNight variant 2.                                                               |
-| `cn/r`          |   2 MB | CryptoNightR (Monero's variant 4).                                                   |
-| `cn/wow`        |   2 MB | CryptoNightR (Wownero).                                                              |
-| `cn/fast`       |   2 MB | CryptoNight variant 1 with half iterations.                                          |
-| `cn/half`       |   2 MB | CryptoNight variant 2 with half iterations (Masari/Torque)                           |
-| `cn/xao`        |   2 MB | CryptoNight variant 0 (modified, Alloy only)                                         |
-| `cn/rto`        |   2 MB | CryptoNight variant 1 (modified, Arto only)                                          |
-| `cn/rwz`        |   2 MB | CryptoNight variant 2 with 3/4 iterations and reversed shuffle operation (Graft).    |
-| `cn/zls`        |   2 MB | CryptoNight variant 2 with 3/4 iterations (Zelerius).                                |
-| `cn/double`     |   2 MB | CryptoNight variant 2 with double iterations (X-CASH).                               |
-| `cn/gpu`        |   2 MB | CryptoNight-GPU (RYO).                                                               |
-| `cn-lite/0`     |   1 MB | CryptoNight-Lite variant 0.                                                          |
-| `cn-lite/1`     |   1 MB | CryptoNight-Lite variant 1.                                                          |
-| `cn-heavy/0`    |   4 MB | CryptoNight-Heavy       .                                                            |
-| `cn-heavy/xhv`  |   4 MB | CryptoNight-Heavy (modified, TUBE only).                                             |
-| `cn-heavy/tube` |   4 MB | CryptoNight-Heavy (modified, Haven Protocol only).                                   |
-| `cn-pico`       | 256 KB | TurtleCoin (TRTL)                                                                    |
-| `rx/test`       |   2 MB | RandomX (reference configuration).                                                   |
-| `rx/0`          |   2 MB | RandomX (reference configuration), reserved for future use.                          |
-| `rx/wow`        |   1 MB | RandomWOW (Wownero).                                                                 |
-| `rx/loki`       |   2 MB | RandomXL (Loki).                                                                     |
+| Name | Memory | Version | Notes |
+|------|--------|---------|-------|
+| `argon2/chukwa` | 512 KB | 3.1.0+ | Argon2id (Chukwa). |
+| `argon2/wrkz` | 256 KB | 3.1.0+ | Argon2id (WRKZ) |
+| `rx/test` | 2 MB | 3.0.0+ | RandomX (reference configuration). |
+| `rx/0` | 2 MB | 3.0.0+ | RandomX (reference configuration), reserved for future use. |
+| `rx/wow` | 1 MB | 3.0.0+ | RandomWOW. |
+| `rx/loki` | 2 MB | 3.0.0+ | RandomXL. |
+| `cn/fast` | 2 MB | 3.0.0+ | CryptoNight variant 1 with half iterations. |
+| `cn/rwz` | 2 MB | 2.14.0+ | CryptoNight variant 2 with 3/4 iterations and reversed shuffle operation. |
+| `cn/zls` | 2 MB | 2.14.0+ | CryptoNight variant 2 with 3/4 iterations. |
+| `cn/double` | 2 MB | 2.14.0+ | CryptoNight variant 2 with double iterations. |
+| `cn/r` | 2 MB | 2.13.0+ | CryptoNightR (Monero's variant 4). |
+| `cn/wow` | 2 MB | 2.12.0+ | CryptoNightR (Wownero). |
+| `cn/gpu` | 2 MB | 2.11.0+ | CryptoNight-GPU. |
+| `cn-pico` | 256 KB | 2.10.0+ | CryptoNight-Pico. |
+| `cn/half` | 2 MB | 2.9.0+ | CryptoNight variant 2 with half iterations. |
+| `cn/2` | 2 MB | 2.8.0+ | CryptoNight variant 2. |
+| `cn/xao` | 2 MB | 2.6.4+ | CryptoNight variant 0 (modified). |
+| `cn/rto` | 2 MB | 2.6.4+ | CryptoNight variant 1 (modified). |
+| `cn-heavy/tube` | 4 MB | 2.6.4+ | CryptoNight-Heavy (modified). |
+| `cn-heavy/xhv` | 4 MB | 2.6.3+ | CryptoNight-Heavy (modified). |
+| `cn-heavy/0` | 4 MB | 2.6.0+ | CryptoNight-Heavy. |
+| `cn/1` | 2 MB | 2.5.0+ | CryptoNight variant 1. |
+| `cn-lite/1` | 1 MB | 2.5.0+ | CryptoNight-Lite variant 1. |
+| `cn-lite/0` | 1 MB | 0.8.0+ | CryptoNight-Lite variant 0. |
+| `cn/0` | 2 MB | 0.5.0+ | CryptoNight (original). |

doc/CHANGELOG_OLD.md

@@ -0,0 +1,349 @@
+# v2.99.6-beta
+- Added commands `pause` and `resume` via JSON RPC 2.0 API (`POST /json_rpc`).
+- Fixed autoconfig regression (since 2.99.5), mostly `rx/wow` was affected by this bug.
+- Fixed user job recovery after donation round.
+- Information about AVX2 CPU feature how hidden in miner summary.
+
+# v2.99.5-beta
+- [#1066](https://github.com/xmrig/xmrig/issues/1066#issuecomment-518080529) Fixed crash and added error message if pool not ready for RandomX.
+- [#1092](https://github.com/xmrig/xmrig/issues/1092) Fixed crash if wrong CPU affinity used.
+- [#1103](https://github.com/xmrig/xmrig/issues/1103) Improved auto configuration for RandomX for CPUs where L2 cache is limiting factor.
+- [#1105](https://github.com/xmrig/xmrig/issues/1105) Improved auto configuration for `cn-pico` algorithm.
+- [#1106](https://github.com/xmrig/xmrig/issues/1106) Fixed `hugepages` field in summary API. 
+- Added alternative short format for CPU threads.
+- Changed format for CPU threads with intensity above 1.
+- Name for reference RandomX configuration changed to `rx/test` to avoid potential conflicts in future.
+
+# v2.99.4-beta
+- [#1062](https://github.com/xmrig/xmrig/issues/1062) Fixed 32 bit support. **32 bit is slow and deprecated**.
+- [#1088](https://github.com/xmrig/xmrig/pull/1088) Fixed macOS compilation.
+- [#1095](https://github.com/xmrig/xmrig/pull/1095) Fixed compatibility with hwloc 1.10.x.
+- Optimized RandomX initialization and switching, fixed rare crash when re-initialize dataset.
+- Fixed ARM build with hwloc.
+
+# v2.99.3-beta
+- [#1082](https://github.com/xmrig/xmrig/issues/1082) Fixed hwloc auto configuration on AMD FX CPUs.
+- Added command line option `--export-topology` for export hwloc topology to a XML file.
+
+# v2.99.2-beta
+- [#1077](https://github.com/xmrig/xmrig/issues/1077) Added NUMA support via **hwloc**.
+- Fixed miner freeze when switch between RandomX variants.
+- Fixed dataset initialization speed on Linux if thread affinity was used.
+
+# v2.99.1-beta
+- [#1072](https://github.com/xmrig/xmrig/issues/1072) Fixed RandomX `seed_hash` re-initialization.
+
+# v2.99.0-beta
+- [#1050](https://github.com/xmrig/xmrig/pull/1050) Added RandomXL algorithm for [Loki](https://loki.network/), algorithm name used by miner is `randomx/loki` or `rx/loki`.
+- Added [flexible](https://github.com/xmrig/xmrig/blob/evo/doc/CPU.md) multi algorithm configuration.
+- Added unlimited switching between incompatible algorithms, all mining options can be changed in runtime.
+- Breaked backward compatibility with previous configs and command line, `variant` option replaced to `algo`, global option `algo` removed, all CPU related settings moved to `cpu` object.
+- Options `av`, `safe` and `max-cpu-usage` removed.
+- Algorithm `cn/msr` renamed to `cn/fast`.
+- Algorithm `cn/xtl` removed.
+- API endpoint `GET /1/threads` replaced to `GET /2/backends`.
+
+# v2.16.0-beta
+- [#1036](https://github.com/xmrig/xmrig/pull/1036) Added RandomWOW (RandomX with different preferences) algorithm support for [Wownero](http://wownero.org/).
+  - Algorithm name used by miner is `randomx/wow` or `rx/wow`.
+  - Currently runtime algorithm switching NOT supported with other algorithms.
+
+# v2.15.4-beta
+- Added global uptime and extended connection information in API.
+- API now return current algorithm instead of global algorithm specified in config.
+- This version also include all changes from stable version v2.14.4.
+
+# v2.15.3-beta
+- [#1014](https://github.com/xmrig/xmrig/issues/1014) Fixed regression, default value for `algo` option was not applied.
+
+# v2.15.2-beta
+- [#1010](https://github.com/xmrig/xmrig/pull/1010#issuecomment-482632107) Added daemon support (solo mining).
+- [#1012](https://github.com/xmrig/xmrig/pull/1012) Fixed compatibility with clang 9.
+- Config subsystem was rewritten, internally JSON is primary format now.
+- Fixed regression, big HTTP responses was truncated.
+
+# v2.15.1-beta
+- [#1007](https://github.com/xmrig/xmrig/issues/1007) Old HTTP API backend based on libmicrohttpd, replaced to custom HTTP server (libuv + http_parser).
+- [#257](https://github.com/xmrig/xmrig-nvidia/pull/257) New logging subsystem, file and syslog now always without colors.
+
+# v2.15.0-beta
+- [#314](https://github.com/xmrig/xmrig-proxy/issues/314) Added donate over proxy feature.
+  - Added new option `donate-over-proxy`.
+  - Added real graceful exit.
+  
+# v2.14.4
+- [#992](https://github.com/xmrig/xmrig/pull/992)  Fixed compilation with Clang 3.5.
+- [#1012](https://github.com/xmrig/xmrig/pull/1012) Fixed compilation with Clang 9.0.
+- In HTTP API for unknown hashrate now used `null` instead of `0.0`.
+- Fixed MSVC 2019 version detection.
+- Removed obsolete automatic variants.
+
+# v2.14.1
+* [#975](https://github.com/xmrig/xmrig/issues/975) Fixed crash on Linux if double thread mode used.
+
+# v2.14.0
+- **[#969](https://github.com/xmrig/xmrig/pull/969) Added new algorithm `cryptonight/rwz`, short alias `cn/rwz` (also known as CryptoNight ReverseWaltz), for upcoming [Graft](https://www.graft.network/) fork.**
+- **[#931](https://github.com/xmrig/xmrig/issues/931) Added new algorithm `cryptonight/zls`, short alias `cn/zls` for [Zelerius Network](https://zelerius.org) fork.**
+- **[#940](https://github.com/xmrig/xmrig/issues/940) Added new algorithm `cryptonight/double`, short alias `cn/double` (also known as CryptoNight HeavyX), for [X-CASH](https://x-cash.org/).**
+- [#951](https://github.com/xmrig/xmrig/issues/951#issuecomment-469581529) Fixed crash if AVX was disabled on OS level.
+- [#952](https://github.com/xmrig/xmrig/issues/952) Fixed compile error on some Linux.
+- [#957](https://github.com/xmrig/xmrig/issues/957#issuecomment-468890667) Added support for embedded config.
+- [#958](https://github.com/xmrig/xmrig/pull/958) Fixed incorrect user agent on ARM platforms.
+- [#968](https://github.com/xmrig/xmrig/pull/968) Optimized `cn/r` algorithm performance.
+
+# v2.13.1
+- [#946](https://github.com/xmrig/xmrig/pull/946) Optimized software AES implementations for CPUs without hardware AES support. `cn/r`, `cn/wow` up to 2.6 times faster, 4-9% improvements for other algorithms.
+
+# v2.13.0
+- **[#938](https://github.com/xmrig/xmrig/issues/938) Added support for new algorithm `cryptonight/r`, short alias `cn/r` (also known as CryptoNightR or CryptoNight variant 4), for upcoming [Monero](https://www.getmonero.org/) fork on March 9, thanks [@SChernykh](https://github.com/SChernykh).**
+- [#939](https://github.com/xmrig/xmrig/issues/939) Added support for dynamic (runtime) pools reload.
+- [#932](https://github.com/xmrig/xmrig/issues/932) Fixed `cn-pico` hashrate drop, regression since v2.11.0.
+
+# v2.12.0
+- [#929](https://github.com/xmrig/xmrig/pull/929) Added support for new algorithm `cryptonight/wow`, short alias `cn/wow` (also known as CryptonightR), for upcoming [Wownero](http://wownero.org) fork on February 14.
+
+# v2.11.0
+- [#928](https://github.com/xmrig/xmrig/issues/928) Added support for new algorithm `cryptonight/gpu`, short alias `cn/gpu` (original name `cryptonight-gpu`), for upcoming [Ryo currency](https://ryo-currency.com) fork on February 14.
+- [#749](https://github.com/xmrig/xmrig/issues/749) Added support for detect hardware AES in runtime on ARMv8 platforms.
+- [#292](https://github.com/xmrig/xmrig/issues/292) Fixed build on ARMv8 platforms if compiler not support hardware AES.
+
+# v2.10.0
+- [#904](https://github.com/xmrig/xmrig/issues/904) Added new algorithm `cn-pico/trtl` (aliases `cryptonight-turtle`, `cn-trtl`) for upcoming TurtleCoin (TRTL) fork.
+- Default value for option `max-cpu-usage` changed to `100` also this option now deprecated.
+
+# v2.9.4
+- [#913](https://github.com/xmrig/xmrig/issues/913) Fixed Masari (MSR) support (this update required for upcoming fork).
+- [#915](https://github.com/xmrig/xmrig/pull/915) Improved security, JIT memory now read-only after patching.
+
+# v2.9.3
+- [#909](https://github.com/xmrig/xmrig/issues/909) Fixed compile errors on FreeBSD.
+- [#912](https://github.com/xmrig/xmrig/pull/912) Fixed, C++ implementation of `cn/half` was produce up to 13% of invalid hashes.
+
+# v2.9.2
+- [#907](https://github.com/xmrig/xmrig/pull/907) Fixed crash on Linux.
+
+# v2.9.1
+- Restored compatibility with https://stellite.hashvault.pro.
+
+# v2.9.0
+- [#899](https://github.com/xmrig/xmrig/issues/899) Added support for new algorithm `cn/half` for Masari and Stellite forks.
+- [#834](https://github.com/xmrig/xmrig/pull/834) Added ASM optimized code for AMD Bulldozer.
+- [#839](https://github.com/xmrig/xmrig/issues/839) Fixed FreeBSD compile.
+- [#857](https://github.com/xmrig/xmrig/pull/857) Fixed impossible to build for macOS without clang.
+
+# v2.8.3
+- [#813](https://github.com/xmrig/xmrig/issues/813) Fixed critical bug with Minergate pool and variant 2.
+
+# v2.8.1
+- [#768](https://github.com/xmrig/xmrig/issues/768) Fixed build with Visual Studio 2015.
+- [#769](https://github.com/xmrig/xmrig/issues/769) Fixed regression, some ANSI escape sequences was in log with disabled colors.
+- [#777](https://github.com/xmrig/xmrig/issues/777) Better report about pool connection issues. 
+- Simplified checks for ASM auto detection, only AES support necessary.
+- Added missing options to `--help` output.
+
+# v2.8.0
+- **[#753](https://github.com/xmrig/xmrig/issues/753) Added new algorithm [CryptoNight variant 2](https://github.com/xmrig/xmrig/issues/753) for Monero fork, thanks [@SChernykh](https://github.com/SChernykh).**
+  - Added global and per thread option `"asm"` and and command line equivalent.
+- **[#758](https://github.com/xmrig/xmrig/issues/758) Added SSL/TLS support for secure connections to pools.**
+  - Added per pool options `"tls"` and `"tls-fingerprint"` and command line equivalents.
+- [#767](https://github.com/xmrig/xmrig/issues/767) Added config autosave feature, same with GPU miners.  
+- [#245](https://github.com/xmrig/xmrig-proxy/issues/245) Fixed API ID collision when run multiple miners on same machine.
+- [#757](https://github.com/xmrig/xmrig/issues/757) Fixed send buffer overflow.
+
+# v2.6.4
+- [#700](https://github.com/xmrig/xmrig/issues/700) `cryptonight-lite/ipbc` replaced to `cryptonight-heavy/tube` for **Bittube (TUBE)**.
+- Added `cryptonight/rto` (cryptonight variant 1 with IPBC/TUBE mod) variant for **Arto (RTO)** coin.
+- Added `cryptonight/xao` (original cryptonight with bigger iteration count) variant for **Alloy (XAO)** coin.
+- Better variant detection for **nicehash.com** and **minergate.com**.
+- [#692](https://github.com/xmrig/xmrig/issues/692) Added support for specify both algorithm and variant via single `algo` option.
+
+# v2.6.3
+- **Added support for new cryptonight-heavy variant xhv** (`cn-heavy/xhv`) for upcoming Haven Protocol fork.
+- **Added support for new cryptonight variant msr** (`cn/msr`) also known as `cryptonight-fast` for upcoming Masari fork.
+- Added new detailed hashrate report.
+- [#446](https://github.com/xmrig/xmrig/issues/446) Likely fixed SIGBUS error on 32 bit ARM CPUs.
+- [#551](https://github.com/xmrig/xmrig/issues/551) Fixed `cn-heavy` algorithm on ARMv8.
+- [#614](https://github.com/xmrig/xmrig/issues/614) Fixed display issue with huge pages percentage when colors disabled.
+- [#615](https://github.com/xmrig/xmrig/issues/615) Fixed build without libcpuid.
+- [#629](https://github.com/xmrig/xmrig/pull/629) Fixed file logging with non-seekable files.
+- [#672](https://github.com/xmrig/xmrig/pull/672) Reverted back `cryptonight-light` and exit if no valid algorithm specified.
+
+# v2.6.2
+ - [#607](https://github.com/xmrig/xmrig/issues/607) Fixed donation bug.
+ - [#610](https://github.com/xmrig/xmrig/issues/610) Fixed ARM build.
+
+# v2.6.1
+ - [#168](https://github.com/xmrig/xmrig-proxy/issues/168) Added support for [mining algorithm negotiation](https://github.com/xmrig/xmrig-proxy/blob/dev/doc/STRATUM_EXT.md#1-mining-algorithm-negotiation).
+ - Added IPBC coin support, base algorithm `cn-lite` variant `ipbc`.
+ - [#581](https://github.com/xmrig/xmrig/issues/581) Added support for upcoming Stellite (XTL) fork, base algorithm `cn` variant `xtl`, variant can set now, no need do it after fork.
+ - Added support for **rig-id** stratum protocol extensions, compatible with xmr-stak.
+ - Changed behavior for option `variant=-1` for `cryptonight`, now variant is `1` by default, if you mine old coins need change `variant` to `0`.
+ - A lot of small fixes and better unification with proxy code.
+
+# v2.6.0-beta3
+- [#563](https://github.com/xmrig/xmrig/issues/563) **Added [advanced threads mode](https://github.com/xmrig/xmrig/issues/563), now possible configure each thread individually.**
+- [#255](https://github.com/xmrig/xmrig/issues/563) Low power mode extended to **triple**, **quard** and **penta** modes.
+- [#519](https://github.com/xmrig/xmrig/issues/519) Fixed high donation levels, improved donation start time randomization.
+- [#554](https://github.com/xmrig/xmrig/issues/554) Fixed regression with `print-time` option.
+
+# v2.6.0-beta2
+- Improved performance for `cryptonight v7` especially in double hash mode.
+- [#499](https://github.com/xmrig/xmrig/issues/499) IPv6 disabled for internal HTTP API by default, was causing issues on some systems.
+- Added short aliases for algorithm names: `cn`, `cn-lite` and `cn-heavy`.
+- Fixed regressions (v2.6.0-beta1 affected)
+  - [#494](https://github.com/xmrig/xmrig/issues/494) Command line option `--donate-level` was broken.
+  - [#502](https://github.com/xmrig/xmrig/issues/502) Build without libmicrohttpd was broken.
+  - Fixed nonce calculation for `--av 4` (software AES, double hash) was causing reduction of effective hashrate and rejected shares on nicehash.
+
+# v2.6.0-beta1
+ - [#476](https://github.com/xmrig/xmrig/issues/476) **Added Cryptonight-Heavy support for Sumokoin ASIC resistance fork.**
+ - HTTP server now runs in main loop, it make possible easy extend API without worry about thread synchronization.
+ - Added initial graceful reload support, miner will reload configuration if config file changed, disabled by default until it will be fully implemented and tested.
+ - Added API endpoint `PUT /1/config` to update current config.
+ - Added API endpoint `GET /1/config` to get current active config.
+ - Added API endpoint `GET /1/threads` to get current active threads configuration.
+ - API endpoint `GET /` now deprecated, use `GET /1/summary` instead.
+ - Added `--api-no-ipv6` and similar config option to disable IPv6 support for HTTP API.
+ - Added `--api-no-restricted` to enable full access to api, this option has no effect if `--api-access-token` not specified.
+
+# v2.5.3
+- Fixed critical bug, in some cases miner was can't recovery connection and switch to failover pool, version 2.5.2 affected. If you use v2.6.0-beta3 this issue doesn't concern you.
+- [#499](https://github.com/xmrig/xmrig/issues/499) IPv6 support disabled for internal HTTP API.
+- Added workaround for nicehash.com if you use `cryptonightv7.<region>.nicehash.com` option `variant=1` will be set automatically.
+
+# v2.5.2
+- [#448](https://github.com/xmrig/xmrig/issues/478) Fixed broken reconnect.
+
+# v2.5.1
+- [#454](https://github.com/xmrig/xmrig/issues/454) Fixed build with libmicrohttpd version below v0.9.35.
+- [#456](https://github.com/xmrig/xmrig/issues/459) Verbose errors related to donation pool was not fully silenced.
+- [#459](https://github.com/xmrig/xmrig/issues/459) Fixed regression (version 2.5.0 affected) with connection to **xmr.f2pool.com**.
+
+# v2.5.0
+- [#434](https://github.com/xmrig/xmrig/issues/434) **Added support for Monero v7 PoW, scheduled on April 6.**
+- Added full IPv6 support.
+- Added protocol extension, when use the miner with xmrig-proxy 2.5+ no more need manually specify `nicehash` option.
+- [#123](https://github.com/xmrig/xmrig-proxy/issues/123) Fixed regression (all versions since 2.4 affected) fragmented responses from pool/proxy was parsed incorrectly.
+- [#428](https://github.com/xmrig/xmrig/issues/428) Fixed regression (version 2.4.5 affected) with CPU cache size detection.
+
+# v2.4.5
+- [#324](https://github.com/xmrig/xmrig/pull/324) Fixed build without libmicrohttpd (CMake cache issue).
+- [#341](https://github.com/xmrig/xmrig/issues/341) Fixed wrong exit code and added command line option `--dry-run`.
+- [#385](https://github.com/xmrig/xmrig/pull/385) Up to 20% performance increase for non-AES CPU and fixed Intel Core 2 cache detection.
+
+# v2.4.4
+ - Added libmicrohttpd version to --version output.
+ - Fixed bug in singal handler, in some cases miner wasn't shutdown properly.
+ - Fixed recent MSVC 2017 version detection.
+ - [#279](https://github.com/xmrig/xmrig/pull/279) Fixed build on some macOS versions.
+
+# v2.4.3
+ - [#94](https://github.com/xmrig/xmrig/issues/94#issuecomment-342019257) [#216](https://github.com/xmrig/xmrig/issues/216) Added **ARMv8** and **ARMv7** support. Hardware AES supported, thanks [Imran Yusuff](https://github.com/imranyusuff).
+ - [#157](https://github.com/xmrig/xmrig/issues/157) [#196](https://github.com/xmrig/xmrig/issues/196) Fixed Linux compile issues.
+ - [#184](https://github.com/xmrig/xmrig/issues/184) Fixed cache size detection for CPUs with disabled Hyper-Threading.
+ - [#200](https://github.com/xmrig/xmrig/issues/200) In some cases miner was doesn't write log to stdout.
+
+# v2.4.2
+ - [#60](https://github.com/xmrig/xmrig/issues/60) Added FreeBSD support, thanks [vcambur](https://github.com/vcambur).
+ - [#153](https://github.com/xmrig/xmrig/issues/153) Fixed issues with dwarfpool.com.
+ 
+# v2.4.1
+  - [#147](https://github.com/xmrig/xmrig/issues/147) Fixed comparability with monero-stratum.
+
+# v2.4.0
+ - Added [HTTP API](https://github.com/xmrig/xmrig/wiki/API).
+ - Added comments support in config file.
+ - libjansson replaced to rapidjson.
+ - [#98](https://github.com/xmrig/xmrig/issues/98) Ignore `keepalive` option with minergate.com and nicehash.com.
+ - [#101](https://github.com/xmrig/xmrig/issues/101) Fixed MSVC 2017 (15.3) compile time version detection.
+ - [#108](https://github.com/xmrig/xmrig/issues/108) Silently ignore invalid values for `donate-level` option.
+ - [#111](https://github.com/xmrig/xmrig/issues/111) Fixed build without AEON support.
+ 
+# v2.3.1
+- [#68](https://github.com/xmrig/xmrig/issues/68) Fixed compatibility with Docker containers, was nothing print on console.
+
+# v2.3.0
+- Added `--cpu-priority` option (0 idle, 2 normal to 5 highest).
+- Added `--user-agent` option, to set custom user-agent string for pool. For example `cpuminer-multi/0.1`.
+- Added `--no-huge-pages` option, to disable huge pages support.
+- [#62](https://github.com/xmrig/xmrig/issues/62) Don't send the login to the dev pool.
+- Force reconnect if pool block miner IP address. helps switch to backup pool.
+- Fixed: failed open default config file if path contains non English characters.
+- Fixed: error occurred if try use unavailable stdin or stdout, regression since version 2.2.0.
+- Fixed: message about huge pages support successfully enabled on Windows was not shown in release builds.
+
+# v2.2.1
+- Fixed [terminal issues](https://github.com/xmrig/xmrig-proxy/issues/2#issuecomment-319914085) after exit on Linux and OS X.
+
+# v2.2.0
+- [#46](https://github.com/xmrig/xmrig/issues/46) Restored config file support. Now possible use multiple config files and combine with command line options also added support for default config.
+- Improved colors support on Windows, now used uv_tty, legacy code removed.
+- QuickEdit Mode now disabled on Windows.
+- Added interactive commands in console window:: **h**ashrate, **p**ause, **r**esume.
+- Fixed autoconf mode for AMD FX CPUs.
+
+# v2.1.0
+- [#40](https://github.com/xmrig/xmrig/issues/40)
+Improved miner shutdown, fixed crash on exit for Linux and OS X.
+- Fixed, login request was contain malformed JSON if username or password has some special characters for example `\`. 
+- [#220](https://github.com/fireice-uk/xmr-stak-cpu/pull/220) Better support for Round Robin DNS, IP address now always chosen randomly instead of stuck on first one.
+- Changed donation address, new [xmrig-proxy](https://github.com/xmrig/xmrig-proxy) is coming soon.
+
+# v2.0.2
+- Better deal with possible duplicate jobs from pool, show warning and ignore duplicates.
+- For Windows builds libuv updated to version 1.13.1 and gcc to 7.1.0.
+
+# v2.0.1
+ - [#27](https://github.com/xmrig/xmrig/issues/27) Fixed possibility crash on 32bit systems.
+
+# v2.0.0
+ - Option `--backup-url` removed, instead now possibility specify multiple pools for example: `-o example1.com:3333 -u user1 -p password1 -k -o example2.com:5555 -u user2 -o example3.com:4444 -u user3`
+ - [#15](https://github.com/xmrig/xmrig/issues/15) Added option `-l, --log-file=FILE` to write log to file.
+ - [#15](https://github.com/xmrig/xmrig/issues/15) Added option `-S, --syslog` to use syslog for logging, Linux only.
+ - [#18](https://github.com/xmrig/xmrig/issues/18) Added nice messages for accepted/rejected shares with diff and network latency.
+ - [#20](https://github.com/xmrig/xmrig/issues/20) Fixed `--cpu-affinity` for more than 32 threads.
+ - Fixed Windows XP support.
+ - Fixed regression, option `--no-color` was not fully disable colored output.
+ - Show resolved pool IP address in miner output.
+ 
+# v1.0.1
+- Fix broken software AES implementation, app has crashed if CPU not support AES-NI, only version 1.0.0 affected.
+
+# v1.0.0
+- Miner complete rewritten in C++ with libuv.
+- This version should be fully compatible (except config file) with previos versions, many new nice features will come in next versions.
+- This is still beta. If you found regression, stability or perfomance issues or have an idea for new feature please fell free to open new [issue](https://github.com/xmrig/xmrig/issues/new).
+- Added new option `--print-time=N`, print hashrate report every N seconds.
+- New hashrate reports, by default every 60 secons.
+- Added Microsoft Visual C++ 2015 and 2017 support.
+- Removed dependency on libcurl.
+- To compile this version from source please switch to [dev](https://github.com/xmrig/xmrig/tree/dev) branch.
+
+# v0.8.2
+- Fixed L2 cache size detection for AMD CPUs (Bulldozer/Piledriver/Steamroller/Excavator architecture).
+
+# v0.8.2
+- Fixed L2 cache size detection for AMD CPUs (Bulldozer/Piledriver/Steamroller/Excavator architecture).
+- Fixed gcc 7.1 support.
+
+# v0.8.1
+- Added nicehash support, detects automaticaly by pool URL, for example `cryptonight.eu.nicehash.com:3355` or manually via option `--nicehash`.
+
+# v0.8.0
+- Added double hash mode, also known as lower power mode. `--av=2` and `--av=4`.
+- Added smart automatic CPU configuration. Default threads count now depends on size of the L3 cache of CPU.
+- Added CryptoNight-Lite support for AEON `-a cryptonight-lite`.
+- Added `--max-cpu-usage` option for auto CPU configuration mode.
+- Added `--safe` option for adjust threads and algorithm variations to current CPU.
+- No more manual steps to enable huge pages on Windows. XMRig will do it automatically.
+- Removed BMI2 algorithm variation.
+- Removed default pool URL.
+
+# v0.6.0
+- Added automatic cryptonight self test.
+- New software AES algorithm variation. Will be automatically selected if cpu not support AES-NI.
+- Added 32 bit builds.
+- Documented [algorithm variations](https://github.com/xmrig/xmrig#algorithm-variations).
+
+# v0.5.0
+- Initial public release.

doc/CPU.md

@@ -1,7 +1,5 @@
 # CPU backend
 
-**Information in this document actual to version 2.99.5+**
-
 All CPU related settings contains in one `cpu` object in config file, CPU backend allow specify multiple profiles and allow switch between them without restrictions by pool request or config change. Default auto-configuration create reasonable minimum of profiles which cover all supported algorithms.
 
 ### Example
@@ -93,3 +91,6 @@ Mining threads priority, value from `1` (lowest priority) to `5` (highest possib
 
 #### `asm`
 Enable/configure or disable ASM optimizations. Possible values: `true`, `false`, `"intel"`, `"ryzen"`, `"bulldozer"`.
+
+#### `argon2-impl` (since v3.1.0)
+Allow override automatically detected Argon2 implementation, this option added mostly for debug purposes, default value `null` means autodetect. Other possible values: `"x86_64"`, `"SSE2"`, `"SSSE3"`, `"XOP"`, `"AVX2"`, `"AVX-512F"`. Manual selection has no safe guards, if you CPU not support required instuctions, miner will crash.

doc/build/CMAKE_OPTIONS.md

@@ -0,0 +1,39 @@
+# CMake options
+This document contains list of useful cmake options.
+
+## Algorithms
+
+* **`-DWITH_CN_LITE=OFF`** disable all CryptoNight-Lite algorithms (`cn-lite/0`, `cn-lite/1`).
+* **`-DWITH_CN_HEAVY=OFF`** disable all CryptoNight-Heavy algorithms (`cn-heavy/0`, `cn-heavy/xhv`, `cn-heavy/tube`).
+* **`-DWITH_CN_PICO=OFF`** disable CryptoNight-Pico algorithm (`cn-pico`).
+* **`-DWITH_CN_GPU=OFF`** disable CryptoNight-GPU algorithm (`cn/gpu`).
+* **`-DWITH_RANDOMX=OFF`** disable RandomX algorithms (`rx/loki`, `rx/wow`).
+* **`-DWITH_ARGON2=OFF`** disable Argon2 algorithms (`argon2/chukwa`, `argon2/wrkz`).
+
+## Features
+
+* **`-DWITH_HWLOC=OFF`**
+disable [hwloc](https://github.com/xmrig/xmrig/issues/1077) support.
+Disabling this feature is not recommended in most cases.
+This feature add external dependency to libhwloc (1.10.0+) (except MSVC builds).
+* **`-DWITH_LIBCPUID=OFF`** disable built in libcpuid support, this feature always disabled if hwloc enabled, if both hwloc and libcpuid disabled auto configuration for CPU will very limited.
+* **`-DWITH_HTTP=OFF`** disable built in HTTP support, this feature used for HTTP API and daemon (solo mining) support.
+* **`-DWITH_TLS=OFF`** disable SSL/TLS support (secure connections to pool). This feature add external dependency to OpenSSL.
+* **`-DWITH_ASM=OFF`** disable assembly optimizations for modern CryptoNight algorithms.
+* **`-DWITH_EMBEDDED_CONFIG=ON`** Enable [embedded](https://github.com/xmrig/xmrig/issues/957) config support.
+
+## Debug options
+
+* **`-DWITH_DEBUG_LOG=ON`** enable debug log (mostly network requests).
+* **`-DHWLOC_DEBUG=ON`** enable some debug log for hwloc.
+* **`-DCMAKE_BUILD_TYPE=Debug`** enable debug build, only useful for investigate crashes, this option slow down miner.
+
+## Special build options
+
+* **`-DXMRIG_DEPS=<path>`** path to precompiled dependensices https://github.com/xmrig/xmrig-deps
+* **`-DARM_TARGET=<number>`** override ARM target, possible values `7` (ARMv7) and `8` (ARMv8).
+* **`-DUV_INCLUDE_DIR=<path>`** custom path to libuv headers.
+* **`-DUV_LIBRARY=<path>`** custom path to libuv library.
+* **`-DHWLOC_INCLUDE_DIR=<path>`** custom path to hwloc headers.
+* **`-DHWLOC_LIBRARY=<path>`** custom path to hwloc library.
+* **`-DOPENSSL_ROOT_DIR=<path>`** custom path to OpenSSL.

src/3rdparty/argon2.h

@@ -0,0 +1,33 @@
+/* XMRig
+ * Copyright 2010      Jeff Garzik <jgarzik@pobox.com>
+ * Copyright 2012-2014 pooler      <pooler@litecoinpool.org>
+ * Copyright 2014      Lucas Jones <https://github.com/lucasjones>
+ * Copyright 2014-2016 Wolf9466    <https://github.com/OhGodAPet>
+ * Copyright 2016      Jay D Dee   <jayddee246@gmail.com>
+ * Copyright 2017-2018 XMR-Stak    <https://github.com/fireice-uk>, <https://github.com/psychocrypt>
+ * Copyright 2018-2019 SChernykh   <https://github.com/SChernykh>
+ * Copyright 2016-2019 XMRig       <https://github.com/xmrig>, <support@xmrig.com>
+ *
+ *   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 <http://www.gnu.org/licenses/>.
+ */
+
+
+#ifndef XMRIG_3RDPARTY_ARGON2_H
+#define XMRIG_3RDPARTY_ARGON2_H
+
+
+#include "3rdparty/argon2/include/argon2.h"
+
+
+#endif /* XMRIG_3RDPARTY_ARGON2_H */

src/3rdparty/argon2/CMakeLists.txt

@@ -0,0 +1,88 @@
+cmake_minimum_required(VERSION 2.8)
+
+project(argon2 C)
+set(CMAKE_C_STANDARD 99)
+set(CMAKE_C_STANDARD_REQUIRED ON)
+
+include(CheckCSourceCompiles)
+
+set(ARGON2_SOURCES
+    lib/argon2.c
+    lib/core.c
+    lib/encoding.c
+    lib/genkat.c
+    lib/impl-select.c
+    lib/blake2/blake2.c
+    )
+
+set(ARGON2_X86_64_ENABLED ON)
+set(ARGON2_X86_64_LIBS    argon2-sse2 argon2-ssse3 argon2-xop argon2-avx2 argon2-avx512f)
+set(ARGON2_X86_64_SOURCES arch/x86_64/lib/argon2-arch.c arch/x86_64/lib/cpu-flags.c)
+
+if (CMAKE_C_COMPILER_ID MATCHES MSVC)
+    function(add_feature_impl FEATURE MSVC_FLAG DEF)
+        add_library(argon2-${FEATURE} STATIC arch/x86_64/lib/argon2-${FEATURE}.c)
+        target_include_directories(argon2-${FEATURE} PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/include)
+        target_include_directories(argon2-${FEATURE} PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/lib)
+        set_target_properties(argon2-${FEATURE} PROPERTIES POSITION_INDEPENDENT_CODE True)
+
+        target_compile_options(argon2-${FEATURE} PRIVATE ${MSVC_FLAG})
+        target_compile_definitions(argon2-${FEATURE} PRIVATE ${DEF})
+    endfunction()
+
+    add_feature_impl(sse2    ""              HAVE_SSE2)
+    add_feature_impl(ssse3   "/arch:SSSE3"   HAVE_SSSE3)
+    add_feature_impl(xop     ""              HAVE_XOP)
+    add_feature_impl(avx2    "/arch:AVX2"    HAVE_AVX2)
+    add_feature_impl(avx512f "/arch:AVX512F" HAVE_AVX512F)
+elseif (NOT XMRIG_ARM AND CMAKE_SIZEOF_VOID_P EQUAL 8)
+    function(add_feature_impl FEATURE GCC_FLAG DEF)
+        add_library(argon2-${FEATURE} STATIC arch/x86_64/lib/argon2-${FEATURE}.c)
+        target_include_directories(argon2-${FEATURE} PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/include)
+        target_include_directories(argon2-${FEATURE} PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/lib)
+        set_target_properties(argon2-${FEATURE} PROPERTIES POSITION_INDEPENDENT_CODE True)
+
+        message("-- argon2: detecting feature '${FEATURE}'...")
+        file(READ arch/x86_64/src/test-feature-${FEATURE}.c SOURCE_CODE)
+
+        # try without flag:
+        check_c_source_compiles("${SOURCE_CODE}" FEATURE_${FEATURE}_NOFLAG)
+        set(HAS_FEATURE ${FEATURE_${FEATURE}_NOFLAG})
+        if (NOT "${HAS_FEATURE}")
+            # try with -m<feature> flag:
+            set(CMAKE_REQUIRED_FLAGS ${GCC_FLAG})
+            check_c_source_compiles("${SOURCE_CODE}" FEATURE_${FEATURE}_FLAG)
+            set(CMAKE_REQUIRED_FLAGS "")
+
+            set(HAS_FEATURE ${FEATURE_${FEATURE}_FLAG})
+            if (${HAS_FEATURE})
+                target_compile_options(argon2-${FEATURE} PRIVATE ${GCC_FLAG})
+            endif()
+        endif()
+
+        if (${HAS_FEATURE})
+            message("-- argon2: feature '${FEATURE}' detected!")
+            target_compile_definitions(argon2-${FEATURE} PRIVATE ${DEF})
+        endif()
+    endfunction()
+
+    add_feature_impl(sse2    -msse2    HAVE_SSE2)
+    add_feature_impl(ssse3   -mssse3   HAVE_SSSE3)
+    add_feature_impl(xop     -mxop     HAVE_XOP)
+    add_feature_impl(avx2    -mavx2    HAVE_AVX2)
+    add_feature_impl(avx512f -mavx512f HAVE_AVX512F)
+else()
+    set(ARGON2_X86_64_ENABLED OFF)
+    list(APPEND ARGON2_SOURCES arch/generic/lib/argon2-arch.c)
+endif()
+
+if (ARGON2_X86_64_ENABLED)
+    set(ARGON2_LIBS ${ARGON2_X86_64_LIBS})
+    list(APPEND ARGON2_SOURCES ${ARGON2_X86_64_SOURCES})
+endif()
+
+add_library(argon2 STATIC ${ARGON2_SOURCES})
+target_link_libraries(argon2 ${ARGON2_LIBS})
+
+target_include_directories(argon2 PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/include)
+target_include_directories(argon2 PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/lib)

src/3rdparty/argon2/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2016 Ondrej Mosnáček
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

src/3rdparty/argon2/README.md

@@ -0,0 +1,58 @@
+# Argon2 [![Build Status](https://travis-ci.org/WOnder93/argon2.svg?branch=master)](https://travis-ci.org/WOnder93/argon2)
+A multi-arch library implementing the Argon2 password hashing algorithm.
+
+This project is based on the [original source code](https://github.com/P-H-C/phc-winner-argon2) by the Argon2 authors. The goal of this project is to provide efficient Argon2 implementations for various HW architectures (x86, SSE, ARM, PowerPC, ...).
+
+For the x86_64 architecture, the library implements a simple CPU dispatch which automatically selects the best implementation based on CPU flags and quick benchmarks.
+
+# Building
+## Using GNU autotools
+
+To prepare the build environment, run:
+```bash
+autoreconf -i
+./configure
+```
+
+After that, just run `make` to build the library.
+
+### Running tests
+After configuring the build environment, run `make check` to run the tests.
+
+### Architecture options
+You can specify the target architecture by passing the `--host=...` flag to `./configure`.
+
+Supported architectures:
+ * `x86_64` – 64-bit x86 architecture
+ * `generic` – use generic C impementation
+
+## Using CMake
+
+To prepare the build environment, run:
+```bash
+cmake -DCMAKE_BUILD_TYPE=Release .
+```
+
+Then you can run `make` to build the library.
+
+## Using QMake/Qt Creator
+A [QMake](http://doc.qt.io/qt-4.8/qmake-manual.html) project is also available in the `qmake` directory. You can open it in the [Qt Creator IDE](http://wiki.qt.io/Category:Tools::QtCreator) or build it from terminal:
+```bash
+cd qmake
+# see table below for the list of possible ARCH and CONFIG values
+qmake ARCH=... CONFIG+=...
+make
+```
+
+### Architecture options
+For QMake builds you can configure support for different architectures. Use the `ARCH` variable to choose the architecture and the `CONFIG` variable to set additional options.
+
+Supported architectures:
+ * `x86_64` – 64-bit x86 architecture
+   * QMake config flags:
+     * `USE_SSE2` – use SSE2 instructions
+     * `USE_SSSE3` – use SSSE3 instructions
+     * `USE_XOP` – use XOP instructions
+     * `USE_AVX2` – use AVX2 instructions
+     * `USE_AVX512F` – use AVX-512F instructions
+ * `generic` – use generic C impementation

src/3rdparty/argon2/arch/generic/lib/argon2-arch.c

@@ -0,0 +1,20 @@
+#include <stdint.h>
+#include <string.h>
+#include <stdlib.h>
+
+#include "impl-select.h"
+
+#define rotr64(x, n) (((x) >> (n)) | ((x) << (64 - (n))))
+
+#include "argon2-template-64.h"
+
+void fill_segment_default(const argon2_instance_t *instance,
+                          argon2_position_t position)
+{
+    fill_segment_64(instance, position);
+}
+
+void argon2_get_impl_list(argon2_impl_list *list)
+{
+    list->count = 0;
+}

src/3rdparty/argon2/arch/x86_64/lib/argon2-arch.c

@@ -0,0 +1,41 @@
+#include <stdint.h>
+#include <string.h>
+#include <stdlib.h>
+
+#include "impl-select.h"
+
+#include "cpu-flags.h"
+#include "argon2-sse2.h"
+#include "argon2-ssse3.h"
+#include "argon2-xop.h"
+#include "argon2-avx2.h"
+#include "argon2-avx512f.h"
+
+/* NOTE: there is no portable intrinsic for 64-bit rotate, but any
+ * sane compiler should be able to compile this into a ROR instruction: */
+#define rotr64(x, n) ((x) >> (n)) | ((x) << (64 - (n)))
+
+#include "argon2-template-64.h"
+
+void fill_segment_default(const argon2_instance_t *instance,
+                          argon2_position_t position)
+{
+    fill_segment_64(instance, position);
+}
+
+void argon2_get_impl_list(argon2_impl_list *list)
+{
+    static const argon2_impl IMPLS[] = {
+        { "x86_64",     NULL,           fill_segment_default },
+        { "SSE2",       check_sse2,     fill_segment_sse2 },
+        { "SSSE3",      check_ssse3,    fill_segment_ssse3 },
+        { "XOP",        check_xop,      fill_segment_xop },
+        { "AVX2",       check_avx2,     fill_segment_avx2 },
+        { "AVX-512F",   check_avx512f,  fill_segment_avx512f },
+    };
+
+    cpu_flags_get();
+
+    list->count = sizeof(IMPLS) / sizeof(IMPLS[0]);
+    list->entries = IMPLS;
+}

src/3rdparty/argon2/arch/x86_64/lib/argon2-avx2.c

@@ -0,0 +1,347 @@
+#include "argon2-avx2.h"
+
+#ifdef HAVE_AVX2
+#include <string.h>
+
+#ifdef __GNUC__
+#   include <x86intrin.h>
+#else
+#   include <intrin.h>
+#endif
+
+#include "cpu-flags.h"
+
+#define r16 (_mm256_setr_epi8( \
+     2,  3,  4,  5,  6,  7,  0,  1, \
+    10, 11, 12, 13, 14, 15,  8,  9, \
+    18, 19, 20, 21, 22, 23, 16, 17, \
+    26, 27, 28, 29, 30, 31, 24, 25))
+
+#define r24 (_mm256_setr_epi8( \
+     3,  4,  5,  6,  7,  0,  1,  2, \
+    11, 12, 13, 14, 15,  8,  9, 10, \
+    19, 20, 21, 22, 23, 16, 17, 18, \
+    27, 28, 29, 30, 31, 24, 25, 26))
+
+#define ror64_16(x) _mm256_shuffle_epi8((x), r16)
+#define ror64_24(x) _mm256_shuffle_epi8((x), r24)
+#define ror64_32(x) _mm256_shuffle_epi32((x), _MM_SHUFFLE(2, 3, 0, 1))
+#define ror64_63(x) \
+    _mm256_xor_si256(_mm256_srli_epi64((x), 63), _mm256_add_epi64((x), (x)))
+
+static __m256i f(__m256i x, __m256i y)
+{
+    __m256i z = _mm256_mul_epu32(x, y);
+    return _mm256_add_epi64(_mm256_add_epi64(x, y), _mm256_add_epi64(z, z));
+}
+
+#define G1(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        A0 = f(A0, B0); \
+        A1 = f(A1, B1); \
+\
+        D0 = _mm256_xor_si256(D0, A0); \
+        D1 = _mm256_xor_si256(D1, A1); \
+\
+        D0 = ror64_32(D0); \
+        D1 = ror64_32(D1); \
+\
+        C0 = f(C0, D0); \
+        C1 = f(C1, D1); \
+\
+        B0 = _mm256_xor_si256(B0, C0); \
+        B1 = _mm256_xor_si256(B1, C1); \
+\
+        B0 = ror64_24(B0); \
+        B1 = ror64_24(B1); \
+    } while ((void)0, 0)
+
+#define G2(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        A0 = f(A0, B0); \
+        A1 = f(A1, B1); \
+\
+        D0 = _mm256_xor_si256(D0, A0); \
+        D1 = _mm256_xor_si256(D1, A1); \
+\
+        D0 = ror64_16(D0); \
+        D1 = ror64_16(D1); \
+\
+        C0 = f(C0, D0); \
+        C1 = f(C1, D1); \
+\
+        B0 = _mm256_xor_si256(B0, C0); \
+        B1 = _mm256_xor_si256(B1, C1); \
+\
+        B0 = ror64_63(B0); \
+        B1 = ror64_63(B1); \
+    } while ((void)0, 0)
+
+#define DIAGONALIZE1(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        B0 = _mm256_permute4x64_epi64(B0, _MM_SHUFFLE(0, 3, 2, 1)); \
+        B1 = _mm256_permute4x64_epi64(B1, _MM_SHUFFLE(0, 3, 2, 1)); \
+\
+        C0 = _mm256_permute4x64_epi64(C0, _MM_SHUFFLE(1, 0, 3, 2)); \
+        C1 = _mm256_permute4x64_epi64(C1, _MM_SHUFFLE(1, 0, 3, 2)); \
+\
+        D0 = _mm256_permute4x64_epi64(D0, _MM_SHUFFLE(2, 1, 0, 3)); \
+        D1 = _mm256_permute4x64_epi64(D1, _MM_SHUFFLE(2, 1, 0, 3)); \
+    } while ((void)0, 0)
+
+#define UNDIAGONALIZE1(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        B0 = _mm256_permute4x64_epi64(B0, _MM_SHUFFLE(2, 1, 0, 3)); \
+        B1 = _mm256_permute4x64_epi64(B1, _MM_SHUFFLE(2, 1, 0, 3)); \
+\
+        C0 = _mm256_permute4x64_epi64(C0, _MM_SHUFFLE(1, 0, 3, 2)); \
+        C1 = _mm256_permute4x64_epi64(C1, _MM_SHUFFLE(1, 0, 3, 2)); \
+\
+        D0 = _mm256_permute4x64_epi64(D0, _MM_SHUFFLE(0, 3, 2, 1)); \
+        D1 = _mm256_permute4x64_epi64(D1, _MM_SHUFFLE(0, 3, 2, 1)); \
+    } while ((void)0, 0)
+
+#define DIAGONALIZE2(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        __m256i tmp1, tmp2; \
+        tmp1 = _mm256_blend_epi32(B0, B1, 0xCC); \
+        tmp2 = _mm256_blend_epi32(B0, B1, 0x33); \
+        B1 = _mm256_permute4x64_epi64(tmp1, _MM_SHUFFLE(2,3,0,1)); \
+        B0 = _mm256_permute4x64_epi64(tmp2, _MM_SHUFFLE(2,3,0,1)); \
+\
+        tmp1 = C0; \
+        C0 = C1; \
+        C1 = tmp1; \
+\
+        tmp1 = _mm256_blend_epi32(D0, D1, 0xCC); \
+        tmp2 = _mm256_blend_epi32(D0, D1, 0x33); \
+        D0 = _mm256_permute4x64_epi64(tmp1, _MM_SHUFFLE(2,3,0,1)); \
+        D1 = _mm256_permute4x64_epi64(tmp2, _MM_SHUFFLE(2,3,0,1)); \
+    } while ((void)0, 0)
+
+#define UNDIAGONALIZE2(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        __m256i tmp1, tmp2; \
+        tmp1 = _mm256_blend_epi32(B0, B1, 0xCC); \
+        tmp2 = _mm256_blend_epi32(B0, B1, 0x33); \
+        B0 = _mm256_permute4x64_epi64(tmp1, _MM_SHUFFLE(2,3,0,1)); \
+        B1 = _mm256_permute4x64_epi64(tmp2, _MM_SHUFFLE(2,3,0,1)); \
+\
+        tmp1 = C0; \
+        C0 = C1; \
+        C1 = tmp1; \
+\
+        tmp1 = _mm256_blend_epi32(D0, D1, 0xCC); \
+        tmp2 = _mm256_blend_epi32(D0, D1, 0x33); \
+        D1 = _mm256_permute4x64_epi64(tmp1, _MM_SHUFFLE(2,3,0,1)); \
+        D0 = _mm256_permute4x64_epi64(tmp2, _MM_SHUFFLE(2,3,0,1)); \
+    } while ((void)0, 0)
+
+#define BLAKE2_ROUND1(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        G1(A0, B0, C0, D0, A1, B1, C1, D1); \
+        G2(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+        DIAGONALIZE1(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+        G1(A0, B0, C0, D0, A1, B1, C1, D1); \
+        G2(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+        UNDIAGONALIZE1(A0, B0, C0, D0, A1, B1, C1, D1); \
+    } while ((void)0, 0)
+
+#define BLAKE2_ROUND2(A0, A1, B0, B1, C0, C1, D0, D1) \
+    do { \
+        G1(A0, B0, C0, D0, A1, B1, C1, D1); \
+        G2(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+        DIAGONALIZE2(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+        G1(A0, B0, C0, D0, A1, B1, C1, D1); \
+        G2(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+        UNDIAGONALIZE2(A0, B0, C0, D0, A1, B1, C1, D1); \
+    } while ((void)0, 0)
+
+enum {
+    ARGON2_HWORDS_IN_BLOCK = ARGON2_OWORDS_IN_BLOCK / 2,
+};
+
+static void fill_block(__m256i *s, const block *ref_block, block *next_block,
+                       int with_xor)
+{
+    __m256i block_XY[ARGON2_HWORDS_IN_BLOCK];
+    unsigned int i;
+
+    if (with_xor) {
+        for (i = 0; i < ARGON2_HWORDS_IN_BLOCK; i++) {
+            s[i] =_mm256_xor_si256(
+                s[i], _mm256_loadu_si256((const __m256i *)ref_block->v + i));
+            block_XY[i] = _mm256_xor_si256(
+                s[i], _mm256_loadu_si256((const __m256i *)next_block->v + i));
+        }
+
+    } else {
+        for (i = 0; i < ARGON2_HWORDS_IN_BLOCK; i++) {
+            block_XY[i] = s[i] =_mm256_xor_si256(
+                s[i], _mm256_loadu_si256((const __m256i *)ref_block->v + i));
+        }
+    }
+
+    for (i = 0; i < 4; ++i) {
+        BLAKE2_ROUND1(
+            s[8 * i + 0], s[8 * i + 1], s[8 * i + 2], s[8 * i + 3],
+            s[8 * i + 4], s[8 * i + 5], s[8 * i + 6], s[8 * i + 7]);
+    }
+
+    for (i = 0; i < 4; ++i) {
+        BLAKE2_ROUND2(
+            s[4 * 0 + i], s[4 * 1 + i], s[4 * 2 + i], s[4 * 3 + i],
+            s[4 * 4 + i], s[4 * 5 + i], s[4 * 6 + i], s[4 * 7 + i]);
+    }
+
+    for (i = 0; i < ARGON2_HWORDS_IN_BLOCK; i++) {
+        s[i] = _mm256_xor_si256(s[i], block_XY[i]);
+        _mm256_storeu_si256((__m256i *)next_block->v + i, s[i]);
+    }
+}
+
+static void next_addresses(block *address_block, block *input_block)
+{
+    /*Temporary zero-initialized blocks*/
+    __m256i zero_block[ARGON2_HWORDS_IN_BLOCK];
+    __m256i zero2_block[ARGON2_HWORDS_IN_BLOCK];
+
+    memset(zero_block, 0, sizeof(zero_block));
+    memset(zero2_block, 0, sizeof(zero2_block));
+
+    /*Increasing index counter*/
+    input_block->v[6]++;
+
+    /*First iteration of G*/
+    fill_block(zero_block, input_block, address_block, 0);
+
+    /*Second iteration of G*/
+    fill_block(zero2_block, address_block, address_block, 0);
+}
+
+void fill_segment_avx2(const argon2_instance_t *instance,
+                       argon2_position_t position)
+{
+    block *ref_block = NULL, *curr_block = NULL;
+    block address_block, input_block;
+    uint64_t pseudo_rand, ref_index, ref_lane;
+    uint32_t prev_offset, curr_offset;
+    uint32_t starting_index, i;
+    __m256i state[ARGON2_HWORDS_IN_BLOCK];
+    int data_independent_addressing;
+
+    if (instance == NULL) {
+        return;
+    }
+
+    data_independent_addressing = (instance->type == Argon2_i) ||
+            (instance->type == Argon2_id && (position.pass == 0) &&
+             (position.slice < ARGON2_SYNC_POINTS / 2));
+
+    if (data_independent_addressing) {
+        init_block_value(&input_block, 0);
+
+        input_block.v[0] = position.pass;
+        input_block.v[1] = position.lane;
+        input_block.v[2] = position.slice;
+        input_block.v[3] = instance->memory_blocks;
+        input_block.v[4] = instance->passes;
+        input_block.v[5] = instance->type;
+    }
+
+    starting_index = 0;
+
+    if ((0 == position.pass) && (0 == position.slice)) {
+        starting_index = 2; /* we have already generated the first two blocks */
+
+        /* Don't forget to generate the first block of addresses: */
+        if (data_independent_addressing) {
+            next_addresses(&address_block, &input_block);
+        }
+    }
+
+    /* Offset of the current block */
+    curr_offset = position.lane * instance->lane_length +
+                  position.slice * instance->segment_length + starting_index;
+
+    if (0 == curr_offset % instance->lane_length) {
+        /* Last block in this lane */
+        prev_offset = curr_offset + instance->lane_length - 1;
+    } else {
+        /* Previous block */
+        prev_offset = curr_offset - 1;
+    }
+
+    memcpy(state, ((instance->memory + prev_offset)->v), ARGON2_BLOCK_SIZE);
+
+    for (i = starting_index; i < instance->segment_length;
+         ++i, ++curr_offset, ++prev_offset) {
+        /*1.1 Rotating prev_offset if needed */
+        if (curr_offset % instance->lane_length == 1) {
+            prev_offset = curr_offset - 1;
+        }
+
+        /* 1.2 Computing the index of the reference block */
+        /* 1.2.1 Taking pseudo-random value from the previous block */
+        if (data_independent_addressing) {
+            if (i % ARGON2_ADDRESSES_IN_BLOCK == 0) {
+                next_addresses(&address_block, &input_block);
+            }
+            pseudo_rand = address_block.v[i % ARGON2_ADDRESSES_IN_BLOCK];
+        } else {
+            pseudo_rand = instance->memory[prev_offset].v[0];
+        }
+
+        /* 1.2.2 Computing the lane of the reference block */
+        ref_lane = ((pseudo_rand >> 32)) % instance->lanes;
+
+        if ((position.pass == 0) && (position.slice == 0)) {
+            /* Can not reference other lanes yet */
+            ref_lane = position.lane;
+        }
+
+        /* 1.2.3 Computing the number of possible reference block within the
+         * lane.
+         */
+        position.index = i;
+        ref_index = index_alpha(instance, &position, pseudo_rand & 0xFFFFFFFF,
+                                ref_lane == position.lane);
+
+        /* 2 Creating a new block */
+        ref_block =
+            instance->memory + instance->lane_length * ref_lane + ref_index;
+        curr_block = instance->memory + curr_offset;
+
+        /* version 1.2.1 and earlier: overwrite, not XOR */
+        if (0 == position.pass || ARGON2_VERSION_10 == instance->version) {
+            fill_block(state, ref_block, curr_block, 0);
+        } else {
+            fill_block(state, ref_block, curr_block, 1);
+        }
+    }
+}
+
+int check_avx2(void)
+{
+    return cpu_flags_have_avx2();
+}
+
+#else
+
+void fill_segment_avx2(const argon2_instance_t *instance,
+                       argon2_position_t position)
+{
+}
+
+int check_avx2(void)
+{
+    return 0;
+}
+
+#endif

src/3rdparty/argon2/arch/x86_64/lib/argon2-avx2.h

@@ -0,0 +1,11 @@
+#ifndef ARGON2_AVX2_H
+#define ARGON2_AVX2_H
+
+#include "core.h"
+
+void fill_segment_avx2(const argon2_instance_t *instance,
+                       argon2_position_t position);
+
+int check_avx2(void);
+
+#endif // ARGON2_AVX2_H

src/3rdparty/argon2/arch/x86_64/lib/argon2-avx512f.c

@@ -0,0 +1,332 @@
+#include "argon2-avx512f.h"
+
+#ifdef HAVE_AVX512F
+#include <stdint.h>
+#include <string.h>
+
+#ifdef __GNUC__
+#   include <x86intrin.h>
+#else
+#   include <intrin.h>
+#endif
+
+#include "cpu-flags.h"
+
+#define ror64(x, n) _mm512_ror_epi64((x), (n))
+
+static __m512i f(__m512i x, __m512i y)
+{
+    __m512i z = _mm512_mul_epu32(x, y);
+    return _mm512_add_epi64(_mm512_add_epi64(x, y), _mm512_add_epi64(z, z));
+}
+
+#define G1(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        A0 = f(A0, B0); \
+        A1 = f(A1, B1); \
+\
+        D0 = _mm512_xor_si512(D0, A0); \
+        D1 = _mm512_xor_si512(D1, A1); \
+\
+        D0 = ror64(D0, 32); \
+        D1 = ror64(D1, 32); \
+\
+        C0 = f(C0, D0); \
+        C1 = f(C1, D1); \
+\
+        B0 = _mm512_xor_si512(B0, C0); \
+        B1 = _mm512_xor_si512(B1, C1); \
+\
+        B0 = ror64(B0, 24); \
+        B1 = ror64(B1, 24); \
+    } while ((void)0, 0)
+
+#define G2(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        A0 = f(A0, B0); \
+        A1 = f(A1, B1); \
+\
+        D0 = _mm512_xor_si512(D0, A0); \
+        D1 = _mm512_xor_si512(D1, A1); \
+\
+        D0 = ror64(D0, 16); \
+        D1 = ror64(D1, 16); \
+\
+        C0 = f(C0, D0); \
+        C1 = f(C1, D1); \
+\
+        B0 = _mm512_xor_si512(B0, C0); \
+        B1 = _mm512_xor_si512(B1, C1); \
+\
+        B0 = ror64(B0, 63); \
+        B1 = ror64(B1, 63); \
+    } while ((void)0, 0)
+
+#define DIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        B0 = _mm512_permutex_epi64(B0, _MM_SHUFFLE(0, 3, 2, 1)); \
+        B1 = _mm512_permutex_epi64(B1, _MM_SHUFFLE(0, 3, 2, 1)); \
+\
+        C0 = _mm512_permutex_epi64(C0, _MM_SHUFFLE(1, 0, 3, 2)); \
+        C1 = _mm512_permutex_epi64(C1, _MM_SHUFFLE(1, 0, 3, 2)); \
+\
+        D0 = _mm512_permutex_epi64(D0, _MM_SHUFFLE(2, 1, 0, 3)); \
+        D1 = _mm512_permutex_epi64(D1, _MM_SHUFFLE(2, 1, 0, 3)); \
+    } while ((void)0, 0)
+
+#define UNDIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        B0 = _mm512_permutex_epi64(B0, _MM_SHUFFLE(2, 1, 0, 3)); \
+        B1 = _mm512_permutex_epi64(B1, _MM_SHUFFLE(2, 1, 0, 3)); \
+\
+        C0 = _mm512_permutex_epi64(C0, _MM_SHUFFLE(1, 0, 3, 2)); \
+        C1 = _mm512_permutex_epi64(C1, _MM_SHUFFLE(1, 0, 3, 2)); \
+\
+        D0 = _mm512_permutex_epi64(D0, _MM_SHUFFLE(0, 3, 2, 1)); \
+        D1 = _mm512_permutex_epi64(D1, _MM_SHUFFLE(0, 3, 2, 1)); \
+    } while ((void)0, 0)
+
+#define BLAKE2_ROUND(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        G1(A0, B0, C0, D0, A1, B1, C1, D1); \
+        G2(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+        DIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+        G1(A0, B0, C0, D0, A1, B1, C1, D1); \
+        G2(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+        UNDIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1); \
+    } while ((void)0, 0)
+
+#define SWAP_HALVES(A0, A1) \
+    do { \
+        __m512i t0, t1; \
+        t0 = _mm512_shuffle_i64x2(A0, A1, _MM_SHUFFLE(1, 0, 1, 0)); \
+        t1 = _mm512_shuffle_i64x2(A0, A1, _MM_SHUFFLE(3, 2, 3, 2)); \
+        A0 = t0; \
+        A1 = t1; \
+    } while((void)0, 0)
+
+#define SWAP_QUARTERS(A0, A1) \
+    do { \
+        SWAP_HALVES(A0, A1); \
+        A0 = _mm512_permutexvar_epi64(_mm512_setr_epi64(0, 1, 4, 5, 2, 3, 6, 7), A0); \
+        A1 = _mm512_permutexvar_epi64(_mm512_setr_epi64(0, 1, 4, 5, 2, 3, 6, 7), A1); \
+    } while((void)0, 0)
+
+#define UNSWAP_QUARTERS(A0, A1) \
+    do { \
+        A0 = _mm512_permutexvar_epi64(_mm512_setr_epi64(0, 1, 4, 5, 2, 3, 6, 7), A0); \
+        A1 = _mm512_permutexvar_epi64(_mm512_setr_epi64(0, 1, 4, 5, 2, 3, 6, 7), A1); \
+        SWAP_HALVES(A0, A1); \
+    } while((void)0, 0)
+
+#define BLAKE2_ROUND1(A0, C0, B0, D0, A1, C1, B1, D1) \
+    do { \
+        SWAP_HALVES(A0, B0); \
+        SWAP_HALVES(C0, D0); \
+        SWAP_HALVES(A1, B1); \
+        SWAP_HALVES(C1, D1); \
+        BLAKE2_ROUND(A0, B0, C0, D0, A1, B1, C1, D1); \
+        SWAP_HALVES(A0, B0); \
+        SWAP_HALVES(C0, D0); \
+        SWAP_HALVES(A1, B1); \
+        SWAP_HALVES(C1, D1); \
+    } while ((void)0, 0)
+
+#define BLAKE2_ROUND2(A0, A1, B0, B1, C0, C1, D0, D1) \
+    do { \
+        SWAP_QUARTERS(A0, A1); \
+        SWAP_QUARTERS(B0, B1); \
+        SWAP_QUARTERS(C0, C1); \
+        SWAP_QUARTERS(D0, D1); \
+        BLAKE2_ROUND(A0, B0, C0, D0, A1, B1, C1, D1); \
+        UNSWAP_QUARTERS(A0, A1); \
+        UNSWAP_QUARTERS(B0, B1); \
+        UNSWAP_QUARTERS(C0, C1); \
+        UNSWAP_QUARTERS(D0, D1); \
+    } while ((void)0, 0)
+
+enum {
+    ARGON2_VECS_IN_BLOCK = ARGON2_OWORDS_IN_BLOCK / 4,
+};
+
+static void fill_block(__m512i *s, const block *ref_block, block *next_block,
+                       int with_xor)
+{
+    __m512i block_XY[ARGON2_VECS_IN_BLOCK];
+    unsigned int i;
+
+    if (with_xor) {
+        for (i = 0; i < ARGON2_VECS_IN_BLOCK; i++) {
+            s[i] =_mm512_xor_si512(
+                s[i], _mm512_loadu_si512((const __m512i *)ref_block->v + i));
+            block_XY[i] = _mm512_xor_si512(
+                s[i], _mm512_loadu_si512((const __m512i *)next_block->v + i));
+        }
+
+    } else {
+        for (i = 0; i < ARGON2_VECS_IN_BLOCK; i++) {
+            block_XY[i] = s[i] =_mm512_xor_si512(
+                s[i], _mm512_loadu_si512((const __m512i *)ref_block->v + i));
+        }
+    }
+
+    for (i = 0; i < 2; ++i) {
+        BLAKE2_ROUND1(
+            s[8 * i + 0], s[8 * i + 1], s[8 * i + 2], s[8 * i + 3],
+            s[8 * i + 4], s[8 * i + 5], s[8 * i + 6], s[8 * i + 7]);
+    }
+
+    for (i = 0; i < 2; ++i) {
+        BLAKE2_ROUND2(
+            s[2 * 0 + i], s[2 * 1 + i], s[2 * 2 + i], s[2 * 3 + i],
+            s[2 * 4 + i], s[2 * 5 + i], s[2 * 6 + i], s[2 * 7 + i]);
+    }
+
+    for (i = 0; i < ARGON2_VECS_IN_BLOCK; i++) {
+        s[i] = _mm512_xor_si512(s[i], block_XY[i]);
+        _mm512_storeu_si512((__m512i *)next_block->v + i, s[i]);
+    }
+}
+
+static void next_addresses(block *address_block, block *input_block)
+{
+    /*Temporary zero-initialized blocks*/
+    __m512i zero_block[ARGON2_VECS_IN_BLOCK];
+    __m512i zero2_block[ARGON2_VECS_IN_BLOCK];
+
+    memset(zero_block, 0, sizeof(zero_block));
+    memset(zero2_block, 0, sizeof(zero2_block));
+
+    /*Increasing index counter*/
+    input_block->v[6]++;
+
+    /*First iteration of G*/
+    fill_block(zero_block, input_block, address_block, 0);
+
+    /*Second iteration of G*/
+    fill_block(zero2_block, address_block, address_block, 0);
+}
+
+void fill_segment_avx512f(const argon2_instance_t *instance,
+                          argon2_position_t position)
+{
+    block *ref_block = NULL, *curr_block = NULL;
+    block address_block, input_block;
+    uint64_t pseudo_rand, ref_index, ref_lane;
+    uint32_t prev_offset, curr_offset;
+    uint32_t starting_index, i;
+    __m512i state[ARGON2_VECS_IN_BLOCK];
+    int data_independent_addressing;
+
+    if (instance == NULL) {
+        return;
+    }
+
+    data_independent_addressing = (instance->type == Argon2_i) ||
+            (instance->type == Argon2_id && (position.pass == 0) &&
+             (position.slice < ARGON2_SYNC_POINTS / 2));
+
+    if (data_independent_addressing) {
+        init_block_value(&input_block, 0);
+
+        input_block.v[0] = position.pass;
+        input_block.v[1] = position.lane;
+        input_block.v[2] = position.slice;
+        input_block.v[3] = instance->memory_blocks;
+        input_block.v[4] = instance->passes;
+        input_block.v[5] = instance->type;
+    }
+
+    starting_index = 0;
+
+    if ((0 == position.pass) && (0 == position.slice)) {
+        starting_index = 2; /* we have already generated the first two blocks */
+
+        /* Don't forget to generate the first block of addresses: */
+        if (data_independent_addressing) {
+            next_addresses(&address_block, &input_block);
+        }
+    }
+
+    /* Offset of the current block */
+    curr_offset = position.lane * instance->lane_length +
+                  position.slice * instance->segment_length + starting_index;
+
+    if (0 == curr_offset % instance->lane_length) {
+        /* Last block in this lane */
+        prev_offset = curr_offset + instance->lane_length - 1;
+    } else {
+        /* Previous block */
+        prev_offset = curr_offset - 1;
+    }
+
+    memcpy(state, ((instance->memory + prev_offset)->v), ARGON2_BLOCK_SIZE);
+
+    for (i = starting_index; i < instance->segment_length;
+         ++i, ++curr_offset, ++prev_offset) {
+        /*1.1 Rotating prev_offset if needed */
+        if (curr_offset % instance->lane_length == 1) {
+            prev_offset = curr_offset - 1;
+        }
+
+        /* 1.2 Computing the index of the reference block */
+        /* 1.2.1 Taking pseudo-random value from the previous block */
+        if (data_independent_addressing) {
+            if (i % ARGON2_ADDRESSES_IN_BLOCK == 0) {
+                next_addresses(&address_block, &input_block);
+            }
+            pseudo_rand = address_block.v[i % ARGON2_ADDRESSES_IN_BLOCK];
+        } else {
+            pseudo_rand = instance->memory[prev_offset].v[0];
+        }
+
+        /* 1.2.2 Computing the lane of the reference block */
+        ref_lane = ((pseudo_rand >> 32)) % instance->lanes;
+
+        if ((position.pass == 0) && (position.slice == 0)) {
+            /* Can not reference other lanes yet */
+            ref_lane = position.lane;
+        }
+
+        /* 1.2.3 Computing the number of possible reference block within the
+         * lane.
+         */
+        position.index = i;
+        ref_index = index_alpha(instance, &position, pseudo_rand & 0xFFFFFFFF,
+                                ref_lane == position.lane);
+
+        /* 2 Creating a new block */
+        ref_block =
+            instance->memory + instance->lane_length * ref_lane + ref_index;
+        curr_block = instance->memory + curr_offset;
+
+        /* version 1.2.1 and earlier: overwrite, not XOR */
+        if (0 == position.pass || ARGON2_VERSION_10 == instance->version) {
+            fill_block(state, ref_block, curr_block, 0);
+        } else {
+            fill_block(state, ref_block, curr_block, 1);
+        }
+    }
+}
+
+int check_avx512f(void)
+{
+    return cpu_flags_have_avx512f();
+}
+
+#else
+
+void fill_segment_avx512f(const argon2_instance_t *instance,
+                          argon2_position_t position)
+{
+}
+
+int check_avx512f(void)
+{
+    return 0;
+}
+
+#endif

src/3rdparty/argon2/arch/x86_64/lib/argon2-avx512f.h

@@ -0,0 +1,11 @@
+#ifndef ARGON2_AVX512F_H
+#define ARGON2_AVX512F_H
+
+#include "core.h"
+
+void fill_segment_avx512f(const argon2_instance_t *instance,
+                          argon2_position_t position);
+
+int check_avx512f(void);
+
+#endif // ARGON2_AVX512F_H

src/3rdparty/argon2/arch/x86_64/lib/argon2-sse2.c

@@ -0,0 +1,128 @@
+#include "argon2-sse2.h"
+
+#ifdef HAVE_SSE2
+#ifdef __GNUC__
+#   include <x86intrin.h>
+#else
+#   include <intrin.h>
+#endif
+
+#include "cpu-flags.h"
+
+#define ror64_16(x) \
+    _mm_shufflehi_epi16( \
+        _mm_shufflelo_epi16((x), _MM_SHUFFLE(0, 3, 2, 1)), \
+        _MM_SHUFFLE(0, 3, 2, 1))
+#define ror64_24(x) \
+    _mm_xor_si128(_mm_srli_epi64((x), 24), _mm_slli_epi64((x), 40))
+#define ror64_32(x) _mm_shuffle_epi32((x), _MM_SHUFFLE(2, 3, 0, 1))
+#define ror64_63(x) \
+    _mm_xor_si128(_mm_srli_epi64((x), 63), _mm_add_epi64((x), (x)))
+
+static __m128i f(__m128i x, __m128i y)
+{
+    __m128i z = _mm_mul_epu32(x, y);
+    return _mm_add_epi64(_mm_add_epi64(x, y), _mm_add_epi64(z, z));
+}
+
+#define G1(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        A0 = f(A0, B0); \
+        A1 = f(A1, B1); \
+\
+        D0 = _mm_xor_si128(D0, A0); \
+        D1 = _mm_xor_si128(D1, A1); \
+\
+        D0 = ror64_32(D0); \
+        D1 = ror64_32(D1); \
+\
+        C0 = f(C0, D0); \
+        C1 = f(C1, D1); \
+\
+        B0 = _mm_xor_si128(B0, C0); \
+        B1 = _mm_xor_si128(B1, C1); \
+\
+        B0 = ror64_24(B0); \
+        B1 = ror64_24(B1); \
+    } while ((void)0, 0)
+
+#define G2(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        A0 = f(A0, B0); \
+        A1 = f(A1, B1); \
+\
+        D0 = _mm_xor_si128(D0, A0); \
+        D1 = _mm_xor_si128(D1, A1); \
+\
+        D0 = ror64_16(D0); \
+        D1 = ror64_16(D1); \
+\
+        C0 = f(C0, D0); \
+        C1 = f(C1, D1); \
+\
+        B0 = _mm_xor_si128(B0, C0); \
+        B1 = _mm_xor_si128(B1, C1); \
+\
+        B0 = ror64_63(B0); \
+        B1 = ror64_63(B1); \
+    } while ((void)0, 0)
+
+#define DIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        __m128i t0 = D0; \
+        __m128i t1 = B0; \
+        D0 = _mm_unpackhi_epi64(D1, _mm_unpacklo_epi64(t0, t0)); \
+        D1 = _mm_unpackhi_epi64(t0, _mm_unpacklo_epi64(D1, D1)); \
+        B0 = _mm_unpackhi_epi64(B0, _mm_unpacklo_epi64(B1, B1)); \
+        B1 = _mm_unpackhi_epi64(B1, _mm_unpacklo_epi64(t1, t1)); \
+    } while ((void)0, 0)
+
+#define UNDIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        __m128i t0 = B0; \
+        __m128i t1 = D0; \
+        B0 = _mm_unpackhi_epi64(B1, _mm_unpacklo_epi64(B0, B0)); \
+        B1 = _mm_unpackhi_epi64(t0, _mm_unpacklo_epi64(B1, B1)); \
+        D0 = _mm_unpackhi_epi64(D0, _mm_unpacklo_epi64(D1, D1)); \
+        D1 = _mm_unpackhi_epi64(D1, _mm_unpacklo_epi64(t1, t1)); \
+    } while ((void)0, 0)
+
+#define BLAKE2_ROUND(A0, A1, B0, B1, C0, C1, D0, D1) \
+    do { \
+        G1(A0, B0, C0, D0, A1, B1, C1, D1); \
+        G2(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+        DIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+        G1(A0, B0, C1, D0, A1, B1, C0, D1); \
+        G2(A0, B0, C1, D0, A1, B1, C0, D1); \
+\
+        UNDIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1); \
+    } while ((void)0, 0)
+
+#include "argon2-template-128.h"
+
+void fill_segment_sse2(const argon2_instance_t *instance,
+                       argon2_position_t position)
+{
+    fill_segment_128(instance, position);
+}
+
+int check_sse2(void)
+{
+    return cpu_flags_have_sse2();
+}
+
+#else
+
+void fill_segment_sse2(const argon2_instance_t *instance,
+                       argon2_position_t position)
+{
+}
+
+int check_sse2(void)
+{
+    return 0;
+}
+
+#endif

src/3rdparty/argon2/arch/x86_64/lib/argon2-sse2.h

@@ -0,0 +1,11 @@
+#ifndef ARGON2_SSE2_H
+#define ARGON2_SSE2_H
+
+#include "core.h"
+
+void fill_segment_sse2(const argon2_instance_t *instance,
+                       argon2_position_t position);
+
+int check_sse2(void);
+
+#endif // ARGON2_SSE2_H

src/3rdparty/argon2/arch/x86_64/lib/argon2-ssse3.c

@@ -0,0 +1,140 @@
+#include "argon2-ssse3.h"
+
+#ifdef HAVE_SSSE3
+#include <string.h>
+
+#ifdef __GNUC__
+#   include <x86intrin.h>
+#else
+#   include <intrin.h>
+#endif
+
+#include "cpu-flags.h"
+
+#define r16 (_mm_setr_epi8( \
+     2,  3,  4,  5,  6,  7,  0,  1, \
+    10, 11, 12, 13, 14, 15,  8,  9))
+
+#define r24 (_mm_setr_epi8( \
+     3,  4,  5,  6,  7,  0,  1,  2, \
+    11, 12, 13, 14, 15,  8,  9, 10))
+
+#define ror64_16(x) _mm_shuffle_epi8((x), r16)
+#define ror64_24(x) _mm_shuffle_epi8((x), r24)
+#define ror64_32(x) _mm_shuffle_epi32((x), _MM_SHUFFLE(2, 3, 0, 1))
+#define ror64_63(x) \
+    _mm_xor_si128(_mm_srli_epi64((x), 63), _mm_add_epi64((x), (x)))
+
+static __m128i f(__m128i x, __m128i y)
+{
+    __m128i z = _mm_mul_epu32(x, y);
+    return _mm_add_epi64(_mm_add_epi64(x, y), _mm_add_epi64(z, z));
+}
+
+#define G1(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        A0 = f(A0, B0); \
+        A1 = f(A1, B1); \
+\
+        D0 = _mm_xor_si128(D0, A0); \
+        D1 = _mm_xor_si128(D1, A1); \
+\
+        D0 = ror64_32(D0); \
+        D1 = ror64_32(D1); \
+\
+        C0 = f(C0, D0); \
+        C1 = f(C1, D1); \
+\
+        B0 = _mm_xor_si128(B0, C0); \
+        B1 = _mm_xor_si128(B1, C1); \
+\
+        B0 = ror64_24(B0); \
+        B1 = ror64_24(B1); \
+    } while ((void)0, 0)
+
+#define G2(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        A0 = f(A0, B0); \
+        A1 = f(A1, B1); \
+\
+        D0 = _mm_xor_si128(D0, A0); \
+        D1 = _mm_xor_si128(D1, A1); \
+\
+        D0 = ror64_16(D0); \
+        D1 = ror64_16(D1); \
+\
+        C0 = f(C0, D0); \
+        C1 = f(C1, D1); \
+\
+        B0 = _mm_xor_si128(B0, C0); \
+        B1 = _mm_xor_si128(B1, C1); \
+\
+        B0 = ror64_63(B0); \
+        B1 = ror64_63(B1); \
+    } while ((void)0, 0)
+
+#define DIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        __m128i t0 = _mm_alignr_epi8(B1, B0, 8); \
+        __m128i t1 = _mm_alignr_epi8(B0, B1, 8); \
+        B0 = t0; \
+        B1 = t1; \
+\
+        t0 = _mm_alignr_epi8(D1, D0, 8); \
+        t1 = _mm_alignr_epi8(D0, D1, 8); \
+        D0 = t1; \
+        D1 = t0; \
+    } while ((void)0, 0)
+
+#define UNDIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        __m128i t0 = _mm_alignr_epi8(B0, B1, 8); \
+        __m128i t1 = _mm_alignr_epi8(B1, B0, 8); \
+        B0 = t0; \
+        B1 = t1; \
+\
+        t0 = _mm_alignr_epi8(D0, D1, 8); \
+        t1 = _mm_alignr_epi8(D1, D0, 8); \
+        D0 = t1; \
+        D1 = t0; \
+    } while ((void)0, 0)
+
+#define BLAKE2_ROUND(A0, A1, B0, B1, C0, C1, D0, D1) \
+    do { \
+        G1(A0, B0, C0, D0, A1, B1, C1, D1); \
+        G2(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+        DIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+        G1(A0, B0, C1, D0, A1, B1, C0, D1); \
+        G2(A0, B0, C1, D0, A1, B1, C0, D1); \
+\
+        UNDIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1); \
+    } while ((void)0, 0)
+
+#include "argon2-template-128.h"
+
+void fill_segment_ssse3(const argon2_instance_t *instance,
+                        argon2_position_t position)
+{
+    fill_segment_128(instance, position);
+}
+
+int check_ssse3(void)
+{
+    return cpu_flags_have_ssse3();
+}
+
+#else
+
+void fill_segment_ssse3(const argon2_instance_t *instance,
+                        argon2_position_t position)
+{
+}
+
+int check_ssse3(void)
+{
+    return 0;
+}
+
+#endif

src/3rdparty/argon2/arch/x86_64/lib/argon2-ssse3.h

@@ -0,0 +1,11 @@
+#ifndef ARGON2_SSSE3_H
+#define ARGON2_SSSE3_H
+
+#include "core.h"
+
+void fill_segment_ssse3(const argon2_instance_t *instance,
+                        argon2_position_t position);
+
+int check_ssse3(void);
+
+#endif // ARGON2_SSSE3_H

src/3rdparty/argon2/arch/x86_64/lib/argon2-template-128.h

@@ -0,0 +1,168 @@
+#include <string.h>
+
+#ifdef __GNUC__
+#   include <x86intrin.h>
+#else
+#   include <intrin.h>
+#endif
+
+#include "core.h"
+
+static void fill_block(__m128i *s, const block *ref_block, block *next_block,
+                       int with_xor)
+{
+    __m128i block_XY[ARGON2_OWORDS_IN_BLOCK];
+    unsigned int i;
+
+    if (with_xor) {
+        for (i = 0; i < ARGON2_OWORDS_IN_BLOCK; i++) {
+            s[i] = _mm_xor_si128(
+                        s[i], _mm_loadu_si128((const __m128i *)ref_block->v + i));
+            block_XY[i] = _mm_xor_si128(
+                        s[i], _mm_loadu_si128((const __m128i *)next_block->v + i));
+        }
+    } else {
+        for (i = 0; i < ARGON2_OWORDS_IN_BLOCK; i++) {
+            block_XY[i] = s[i] = _mm_xor_si128(
+                        s[i], _mm_loadu_si128((const __m128i *)ref_block->v + i));
+        }
+    }
+
+    for (i = 0; i < 8; ++i) {
+        BLAKE2_ROUND(
+            s[8 * i + 0], s[8 * i + 1], s[8 * i + 2], s[8 * i + 3],
+            s[8 * i + 4], s[8 * i + 5], s[8 * i + 6], s[8 * i + 7]);
+    }
+
+    for (i = 0; i < 8; ++i) {
+        BLAKE2_ROUND(
+            s[8 * 0 + i], s[8 * 1 + i], s[8 * 2 + i], s[8 * 3 + i],
+            s[8 * 4 + i], s[8 * 5 + i], s[8 * 6 + i], s[8 * 7 + i]);
+    }
+
+    for (i = 0; i < ARGON2_OWORDS_IN_BLOCK; i++) {
+        s[i] = _mm_xor_si128(s[i], block_XY[i]);
+        _mm_storeu_si128((__m128i *)next_block->v + i, s[i]);
+    }
+}
+
+static void next_addresses(block *address_block, block *input_block)
+{
+    /*Temporary zero-initialized blocks*/
+    __m128i zero_block[ARGON2_OWORDS_IN_BLOCK];
+    __m128i zero2_block[ARGON2_OWORDS_IN_BLOCK];
+
+    memset(zero_block, 0, sizeof(zero_block));
+    memset(zero2_block, 0, sizeof(zero2_block));
+
+    /*Increasing index counter*/
+    input_block->v[6]++;
+
+    /*First iteration of G*/
+    fill_block(zero_block, input_block, address_block, 0);
+
+    /*Second iteration of G*/
+    fill_block(zero2_block, address_block, address_block, 0);
+}
+
+static void fill_segment_128(const argon2_instance_t *instance,
+                             argon2_position_t position)
+{
+    block *ref_block = NULL, *curr_block = NULL;
+    block address_block, input_block;
+    uint64_t pseudo_rand, ref_index, ref_lane;
+    uint32_t prev_offset, curr_offset;
+    uint32_t starting_index, i;
+    __m128i state[ARGON2_OWORDS_IN_BLOCK];
+    int data_independent_addressing;
+
+    if (instance == NULL) {
+        return;
+    }
+
+    data_independent_addressing = (instance->type == Argon2_i) ||
+            (instance->type == Argon2_id && (position.pass == 0) &&
+             (position.slice < ARGON2_SYNC_POINTS / 2));
+
+    if (data_independent_addressing) {
+        init_block_value(&input_block, 0);
+
+        input_block.v[0] = position.pass;
+        input_block.v[1] = position.lane;
+        input_block.v[2] = position.slice;
+        input_block.v[3] = instance->memory_blocks;
+        input_block.v[4] = instance->passes;
+        input_block.v[5] = instance->type;
+    }
+
+    starting_index = 0;
+
+    if ((0 == position.pass) && (0 == position.slice)) {
+        starting_index = 2; /* we have already generated the first two blocks */
+
+        /* Don't forget to generate the first block of addresses: */
+        if (data_independent_addressing) {
+            next_addresses(&address_block, &input_block);
+        }
+    }
+
+    /* Offset of the current block */
+    curr_offset = position.lane * instance->lane_length +
+                  position.slice * instance->segment_length + starting_index;
+
+    if (0 == curr_offset % instance->lane_length) {
+        /* Last block in this lane */
+        prev_offset = curr_offset + instance->lane_length - 1;
+    } else {
+        /* Previous block */
+        prev_offset = curr_offset - 1;
+    }
+
+    memcpy(state, ((instance->memory + prev_offset)->v), ARGON2_BLOCK_SIZE);
+
+    for (i = starting_index; i < instance->segment_length;
+         ++i, ++curr_offset, ++prev_offset) {
+        /*1.1 Rotating prev_offset if needed */
+        if (curr_offset % instance->lane_length == 1) {
+            prev_offset = curr_offset - 1;
+        }
+
+        /* 1.2 Computing the index of the reference block */
+        /* 1.2.1 Taking pseudo-random value from the previous block */
+        if (data_independent_addressing) {
+            if (i % ARGON2_ADDRESSES_IN_BLOCK == 0) {
+                next_addresses(&address_block, &input_block);
+            }
+            pseudo_rand = address_block.v[i % ARGON2_ADDRESSES_IN_BLOCK];
+        } else {
+            pseudo_rand = instance->memory[prev_offset].v[0];
+        }
+
+        /* 1.2.2 Computing the lane of the reference block */
+        ref_lane = ((pseudo_rand >> 32)) % instance->lanes;
+
+        if ((position.pass == 0) && (position.slice == 0)) {
+            /* Can not reference other lanes yet */
+            ref_lane = position.lane;
+        }
+
+        /* 1.2.3 Computing the number of possible reference block within the
+         * lane.
+         */
+        position.index = i;
+        ref_index = index_alpha(instance, &position, pseudo_rand & 0xFFFFFFFF,
+                                ref_lane == position.lane);
+
+        /* 2 Creating a new block */
+        ref_block =
+            instance->memory + instance->lane_length * ref_lane + ref_index;
+        curr_block = instance->memory + curr_offset;
+
+        /* version 1.2.1 and earlier: overwrite, not XOR */
+        if (0 == position.pass || ARGON2_VERSION_10 == instance->version) {
+            fill_block(state, ref_block, curr_block, 0);
+        } else {
+            fill_block(state, ref_block, curr_block, 1);
+        }
+    }
+}

src/3rdparty/argon2/arch/x86_64/lib/argon2-xop.c

@@ -0,0 +1,128 @@
+#include "argon2-xop.h"
+
+#ifdef HAVE_XOP
+#include <string.h>
+
+#ifdef __GNUC__
+#   include <x86intrin.h>
+#else
+#   include <intrin.h>
+#endif
+
+#include "cpu-flags.h"
+
+#define ror64(x, c) _mm_roti_epi64((x), -(c))
+
+static __m128i f(__m128i x, __m128i y)
+{
+    __m128i z = _mm_mul_epu32(x, y);
+    return _mm_add_epi64(_mm_add_epi64(x, y), _mm_add_epi64(z, z));
+}
+
+#define G1(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        A0 = f(A0, B0); \
+        A1 = f(A1, B1); \
+\
+        D0 = _mm_xor_si128(D0, A0); \
+        D1 = _mm_xor_si128(D1, A1); \
+\
+        D0 = ror64(D0, 32); \
+        D1 = ror64(D1, 32); \
+\
+        C0 = f(C0, D0); \
+        C1 = f(C1, D1); \
+\
+        B0 = _mm_xor_si128(B0, C0); \
+        B1 = _mm_xor_si128(B1, C1); \
+\
+        B0 = ror64(B0, 24); \
+        B1 = ror64(B1, 24); \
+    } while ((void)0, 0)
+
+#define G2(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        A0 = f(A0, B0); \
+        A1 = f(A1, B1); \
+\
+        D0 = _mm_xor_si128(D0, A0); \
+        D1 = _mm_xor_si128(D1, A1); \
+\
+        D0 = ror64(D0, 16); \
+        D1 = ror64(D1, 16); \
+\
+        C0 = f(C0, D0); \
+        C1 = f(C1, D1); \
+\
+        B0 = _mm_xor_si128(B0, C0); \
+        B1 = _mm_xor_si128(B1, C1); \
+\
+        B0 = ror64(B0, 63); \
+        B1 = ror64(B1, 63); \
+    } while ((void)0, 0)
+
+#define DIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        __m128i t0 = _mm_alignr_epi8(B1, B0, 8); \
+        __m128i t1 = _mm_alignr_epi8(B0, B1, 8); \
+        B0 = t0; \
+        B1 = t1; \
+\
+        t0 = _mm_alignr_epi8(D1, D0, 8); \
+        t1 = _mm_alignr_epi8(D0, D1, 8); \
+        D0 = t1; \
+        D1 = t0; \
+    } while ((void)0, 0)
+
+#define UNDIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1) \
+    do { \
+        __m128i t0 = _mm_alignr_epi8(B0, B1, 8); \
+        __m128i t1 = _mm_alignr_epi8(B1, B0, 8); \
+        B0 = t0; \
+        B1 = t1; \
+\
+        t0 = _mm_alignr_epi8(D0, D1, 8); \
+        t1 = _mm_alignr_epi8(D1, D0, 8); \
+        D0 = t1; \
+        D1 = t0; \
+    } while ((void)0, 0)
+
+#define BLAKE2_ROUND(A0, A1, B0, B1, C0, C1, D0, D1) \
+    do { \
+        G1(A0, B0, C0, D0, A1, B1, C1, D1); \
+        G2(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+        DIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+        G1(A0, B0, C1, D0, A1, B1, C0, D1); \
+        G2(A0, B0, C1, D0, A1, B1, C0, D1); \
+\
+        UNDIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1); \
+    } while ((void)0, 0)
+
+#include "argon2-template-128.h"
+
+void fill_segment_xop(const argon2_instance_t *instance,
+                      argon2_position_t position)
+{
+    fill_segment_128(instance, position);
+}
+
+int check_xop(void)
+{
+    return cpu_flags_have_xop();
+}
+
+#else
+
+void fill_segment_xop(const argon2_instance_t *instance,
+                      argon2_position_t position)
+{
+}
+
+int check_xop(void)
+{
+    return 0;
+}
+
+#endif

src/3rdparty/argon2/arch/x86_64/lib/argon2-xop.h

@@ -0,0 +1,11 @@
+#ifndef ARGON2_XOP_H
+#define ARGON2_XOP_H
+
+#include "core.h"
+
+void fill_segment_xop(const argon2_instance_t *instance,
+                      argon2_position_t position);
+
+int check_xop(void);
+
+#endif // ARGON2_XOP_H

src/3rdparty/argon2/arch/x86_64/lib/cpu-flags.c

@@ -0,0 +1,129 @@
+#include <stdbool.h>
+#include <stdint.h>
+
+
+#include "cpu-flags.h"
+
+#include <stdio.h>
+
+#ifdef _MSC_VER
+#   include <intrin.h>
+#else
+#   include <cpuid.h>
+#endif
+
+#ifndef bit_OSXSAVE
+#   define bit_OSXSAVE (1 << 27)
+#endif
+
+#ifndef bit_SSE2
+#   define bit_SSE2 (1 << 26)
+#endif
+
+#ifndef bit_SSSE3
+#   define bit_SSSE3 (1 << 9)
+#endif
+
+#ifndef bit_AVX2
+#   define bit_AVX2 (1 << 5)
+#endif
+
+#ifndef bit_AVX512F
+#   define bit_AVX512F (1 << 16)
+#endif
+
+#ifndef bit_XOP
+#   define bit_XOP (1 << 11)
+#endif
+
+#define PROCESSOR_INFO    (1)
+#define EXTENDED_FEATURES (7)
+
+#define EAX_Reg  (0)
+#define EBX_Reg  (1)
+#define ECX_Reg  (2)
+#define EDX_Reg  (3)
+
+
+enum {
+    X86_64_FEATURE_SSE2     = (1 << 0),
+    X86_64_FEATURE_SSSE3    = (1 << 1),
+    X86_64_FEATURE_XOP      = (1 << 2),
+    X86_64_FEATURE_AVX2     = (1 << 3),
+    X86_64_FEATURE_AVX512F  = (1 << 4),
+};
+
+static unsigned int cpu_flags;
+
+
+static inline void cpuid(uint32_t level, int32_t output[4])
+{
+#   ifdef _MSC_VER
+    __cpuid(output, (int) level);
+#   else
+    __cpuid_count(level, 0, output[0], output[1], output[2], output[3]);
+#   endif
+}
+
+
+static bool has_feature(uint32_t level, uint32_t reg, int32_t bit)
+{
+    int32_t cpu_info[4] = { 0 };
+    cpuid(level, cpu_info);
+
+    return (cpu_info[reg] & bit) != 0;
+}
+
+
+void cpu_flags_get(void)
+{
+    if (has_feature(PROCESSOR_INFO, EDX_Reg, bit_SSE2)) {
+        cpu_flags |= X86_64_FEATURE_SSE2;
+    }
+
+    if (has_feature(PROCESSOR_INFO, ECX_Reg, bit_SSSE3)) {
+        cpu_flags |= X86_64_FEATURE_SSSE3;
+    }
+
+    if (!has_feature(PROCESSOR_INFO, ECX_Reg, bit_OSXSAVE)) {
+        return;
+    }
+
+    if (has_feature(EXTENDED_FEATURES, EBX_Reg, bit_AVX2)) {
+        cpu_flags |= X86_64_FEATURE_AVX2;
+    }
+
+    if (has_feature(EXTENDED_FEATURES, EBX_Reg, bit_AVX512F)) {
+        cpu_flags |= X86_64_FEATURE_AVX512F;
+    }
+
+    if (has_feature(0x80000001, ECX_Reg, bit_XOP)) {
+        cpu_flags |= X86_64_FEATURE_XOP;
+    }
+}
+
+int cpu_flags_have_sse2(void)
+{
+    return cpu_flags & X86_64_FEATURE_SSE2;
+}
+
+int cpu_flags_have_ssse3(void)
+{
+    return cpu_flags & X86_64_FEATURE_SSSE3;
+}
+
+int cpu_flags_have_xop(void)
+{
+    return cpu_flags & X86_64_FEATURE_XOP;
+}
+
+int cpu_flags_have_avx2(void)
+{
+    return cpu_flags & X86_64_FEATURE_AVX2;
+}
+
+int cpu_flags_have_avx512f(void)
+{
+    return cpu_flags & X86_64_FEATURE_AVX512F;
+}
+

src/3rdparty/argon2/arch/x86_64/lib/cpu-flags.h

@@ -0,0 +1,12 @@
+#ifndef ARGON2_CPU_FLAGS_H
+#define ARGON2_CPU_FLAGS_H
+
+void cpu_flags_get(void);
+
+int cpu_flags_have_sse2(void);
+int cpu_flags_have_ssse3(void);
+int cpu_flags_have_xop(void);
+int cpu_flags_have_avx2(void);
+int cpu_flags_have_avx512f(void);
+
+#endif // ARGON2_CPU_FLAGS_H

src/3rdparty/argon2/arch/x86_64/src/test-feature-avx2.c

@@ -0,0 +1,8 @@
+#include <x86intrin.h>
+
+void function_avx2(__m256i *dst, const __m256i *a, const __m256i *b)
+{
+    *dst = _mm256_xor_si256(*a, *b);
+}
+
+int main(void) { return 0; }

src/3rdparty/argon2/arch/x86_64/src/test-feature-avx512f.c

@@ -0,0 +1,8 @@
+#include <x86intrin.h>
+
+void function_avx512f(__m512i *dst, const __m512i *a)
+{
+    *dst = _mm512_ror_epi64(*a, 57);
+}
+
+int main(void) { return 0; }

src/3rdparty/argon2/arch/x86_64/src/test-feature-sse2.c

@@ -0,0 +1,8 @@
+#include <x86intrin.h>
+
+void function_sse2(__m128i *dst, const __m128i *a, const __m128i *b)
+{
+    *dst = _mm_xor_si128(*a, *b);
+}
+
+int main(void) { return 0; }

src/3rdparty/argon2/arch/x86_64/src/test-feature-ssse3.c

@@ -0,0 +1,8 @@
+#include <x86intrin.h>
+
+void function_ssse3(__m128i *dst, const __m128i *a, const __m128i *b)
+{
+    *dst = _mm_shuffle_epi8(*a, *b);
+}
+
+int main(void) { return 0; }

src/3rdparty/argon2/arch/x86_64/src/test-feature-xop.c

@@ -0,0 +1,8 @@
+#include <x86intrin.h>
+
+void function_xop(__m128i *dst, const __m128i *a, int b)
+{
+    *dst = _mm_roti_epi64(*a, b);
+}
+
+int main(void) { return 0; }

src/3rdparty/argon2/include/argon2.h

@@ -0,0 +1,465 @@
+/*
+ * Argon2 source code package
+ *
+ * Written by Daniel Dinu and Dmitry Khovratovich, 2015
+ *
+ * This work is licensed under a Creative Commons CC0 1.0 License/Waiver.
+ *
+ * You should have received a copy of the CC0 Public Domain Dedication
+ * along with this software. If not, see
+ * <http://creativecommons.org/publicdomain/zero/1.0/>.
+ */
+
+#ifndef ARGON2_H
+#define ARGON2_H
+
+#include <stdint.h>
+#include <stddef.h>
+#include <stdio.h>
+#include <limits.h>
+
+/* Symbols visibility control */
+#define ARGON2_PUBLIC
+
+#if defined(__cplusplus)
+extern "C" {
+#endif
+
+/*
+ * Argon2 input parameter restrictions
+ */
+
+/* Minimum and maximum number of lanes (degree of parallelism) */
+#define ARGON2_MIN_LANES UINT32_C(1)
+#define ARGON2_MAX_LANES UINT32_C(0xFFFFFF)
+
+/* Minimum and maximum number of threads */
+#define ARGON2_MIN_THREADS UINT32_C(1)
+#define ARGON2_MAX_THREADS UINT32_C(0xFFFFFF)
+
+/* Number of synchronization points between lanes per pass */
+#define ARGON2_SYNC_POINTS UINT32_C(4)
+
+/* Minimum and maximum digest size in bytes */
+#define ARGON2_MIN_OUTLEN UINT32_C(4)
+#define ARGON2_MAX_OUTLEN UINT32_C(0xFFFFFFFF)
+
+/* Minimum and maximum number of memory blocks (each of BLOCK_SIZE bytes) */
+#define ARGON2_MIN_MEMORY (2 * ARGON2_SYNC_POINTS) /* 2 blocks per slice */
+
+#define ARGON2_MIN(a, b) ((a) < (b) ? (a) : (b))
+/* Max memory size is addressing-space/2, topping at 2^32 blocks (4 TB) */
+#define ARGON2_MAX_MEMORY_BITS                                                 \
+    ARGON2_MIN(UINT32_C(32), (sizeof(void *) * CHAR_BIT - 10 - 1))
+#define ARGON2_MAX_MEMORY                                                      \
+    ARGON2_MIN(UINT32_C(0xFFFFFFFF), UINT64_C(1) << ARGON2_MAX_MEMORY_BITS)
+
+/* Minimum and maximum number of passes */
+#define ARGON2_MIN_TIME UINT32_C(1)
+#define ARGON2_MAX_TIME UINT32_C(0xFFFFFFFF)
+
+/* Minimum and maximum password length in bytes */
+#define ARGON2_MIN_PWD_LENGTH UINT32_C(0)
+#define ARGON2_MAX_PWD_LENGTH UINT32_C(0xFFFFFFFF)
+
+/* Minimum and maximum associated data length in bytes */
+#define ARGON2_MIN_AD_LENGTH UINT32_C(0)
+#define ARGON2_MAX_AD_LENGTH UINT32_C(0xFFFFFFFF)
+
+/* Minimum and maximum salt length in bytes */
+#define ARGON2_MIN_SALT_LENGTH UINT32_C(8)
+#define ARGON2_MAX_SALT_LENGTH UINT32_C(0xFFFFFFFF)
+
+/* Minimum and maximum key length in bytes */
+#define ARGON2_MIN_SECRET UINT32_C(0)
+#define ARGON2_MAX_SECRET UINT32_C(0xFFFFFFFF)
+
+/* Flags to determine which fields are securely wiped (default = no wipe). */
+#define ARGON2_DEFAULT_FLAGS UINT32_C(0)
+#define ARGON2_FLAG_CLEAR_PASSWORD (UINT32_C(1) << 0)
+#define ARGON2_FLAG_CLEAR_SECRET (UINT32_C(1) << 1)
+#define ARGON2_FLAG_GENKAT (UINT32_C(1) << 3)
+
+/* Global flag to determine if we are wiping internal memory buffers. This flag
+ * is defined in core.c and deafults to 1 (wipe internal memory). */
+extern int FLAG_clear_internal_memory;
+
+/* Error codes */
+typedef enum Argon2_ErrorCodes {
+    ARGON2_OK = 0,
+
+    ARGON2_OUTPUT_PTR_NULL = -1,
+
+    ARGON2_OUTPUT_TOO_SHORT = -2,
+    ARGON2_OUTPUT_TOO_LONG = -3,
+
+    ARGON2_PWD_TOO_SHORT = -4,
+    ARGON2_PWD_TOO_LONG = -5,
+
+    ARGON2_SALT_TOO_SHORT = -6,
+    ARGON2_SALT_TOO_LONG = -7,
+
+    ARGON2_AD_TOO_SHORT = -8,
+    ARGON2_AD_TOO_LONG = -9,
+
+    ARGON2_SECRET_TOO_SHORT = -10,
+    ARGON2_SECRET_TOO_LONG = -11,
+
+    ARGON2_TIME_TOO_SMALL = -12,
+    ARGON2_TIME_TOO_LARGE = -13,
+
+    ARGON2_MEMORY_TOO_LITTLE = -14,
+    ARGON2_MEMORY_TOO_MUCH = -15,
+
+    ARGON2_LANES_TOO_FEW = -16,
+    ARGON2_LANES_TOO_MANY = -17,
+
+    ARGON2_PWD_PTR_MISMATCH = -18,    /* NULL ptr with non-zero length */
+    ARGON2_SALT_PTR_MISMATCH = -19,   /* NULL ptr with non-zero length */
+    ARGON2_SECRET_PTR_MISMATCH = -20, /* NULL ptr with non-zero length */
+    ARGON2_AD_PTR_MISMATCH = -21,     /* NULL ptr with non-zero length */
+
+    ARGON2_MEMORY_ALLOCATION_ERROR = -22,
+
+    ARGON2_FREE_MEMORY_CBK_NULL = -23,
+    ARGON2_ALLOCATE_MEMORY_CBK_NULL = -24,
+
+    ARGON2_INCORRECT_PARAMETER = -25,
+    ARGON2_INCORRECT_TYPE = -26,
+
+    ARGON2_OUT_PTR_MISMATCH = -27,
+
+    ARGON2_THREADS_TOO_FEW = -28,
+    ARGON2_THREADS_TOO_MANY = -29,
+
+    ARGON2_MISSING_ARGS = -30,
+
+    ARGON2_ENCODING_FAIL = -31,
+
+    ARGON2_DECODING_FAIL = -32,
+
+    ARGON2_THREAD_FAIL = -33,
+
+    ARGON2_DECODING_LENGTH_FAIL = -34,
+
+    ARGON2_VERIFY_MISMATCH = -35
+} argon2_error_codes;
+
+/* Memory allocator types --- for external allocation */
+typedef int (*allocate_fptr)(uint8_t **memory, size_t bytes_to_allocate);
+typedef void (*deallocate_fptr)(uint8_t *memory, size_t bytes_to_allocate);
+
+/* Argon2 external data structures */
+
+/*
+ *****
+ * Context: structure to hold Argon2 inputs:
+ *  output array and its length,
+ *  password and its length,
+ *  salt and its length,
+ *  secret and its length,
+ *  associated data and its length,
+ *  number of passes, amount of used memory (in KBytes, can be rounded up a bit)
+ *  number of parallel threads that will be run.
+ * All the parameters above affect the output hash value.
+ * Additionally, two function pointers can be provided to allocate and
+ * deallocate the memory (if NULL, memory will be allocated internally).
+ * Also, three flags indicate whether to erase password, secret as soon as they
+ * are pre-hashed (and thus not needed anymore), and the entire memory
+ *****
+ * Simplest situation: you have output array out[8], password is stored in
+ * pwd[32], salt is stored in salt[16], you do not have keys nor associated
+ * data. You need to spend 1 GB of RAM and you run 5 passes of Argon2d with
+ * 4 parallel lanes.
+ * You want to erase the password, but you're OK with last pass not being
+ * erased. You want to use the default memory allocator.
+ * Then you initialize:
+ Argon2_Context(out,8,pwd,32,salt,16,NULL,0,NULL,0,5,1<<20,4,4,NULL,NULL,true,false,false,false)
+ */
+typedef struct Argon2_Context {
+    uint8_t *out;    /* output array */
+    uint32_t outlen; /* digest length */
+
+    uint8_t *pwd;    /* password array */
+    uint32_t pwdlen; /* password length */
+
+    uint8_t *salt;    /* salt array */
+    uint32_t saltlen; /* salt length */
+
+    uint8_t *secret;    /* key array */
+    uint32_t secretlen; /* key length */
+
+    uint8_t *ad;    /* associated data array */
+    uint32_t adlen; /* associated data length */
+
+    uint32_t t_cost;  /* number of passes */
+    uint32_t m_cost;  /* amount of memory requested (KB) */
+    uint32_t lanes;   /* number of lanes */
+    uint32_t threads; /* maximum number of threads */
+
+    uint32_t version; /* version number */
+
+    allocate_fptr allocate_cbk; /* pointer to memory allocator */
+    deallocate_fptr free_cbk;   /* pointer to memory deallocator */
+
+    uint32_t flags; /* array of bool options */
+} argon2_context;
+
+/* Argon2 primitive type */
+typedef enum Argon2_type {
+    Argon2_d = 0,
+    Argon2_i = 1,
+    Argon2_id = 2
+} argon2_type;
+
+/* Version of the algorithm */
+typedef enum Argon2_version {
+    ARGON2_VERSION_10 = 0x10,
+    ARGON2_VERSION_13 = 0x13,
+    ARGON2_VERSION_NUMBER = ARGON2_VERSION_13
+} argon2_version;
+
+/*
+ * Function that gives the string representation of an argon2_type.
+ * @param type The argon2_type that we want the string for
+ * @param uppercase Whether the string should have the first letter uppercase
+ * @return NULL if invalid type, otherwise the string representation.
+ */
+ARGON2_PUBLIC const char *argon2_type2string(argon2_type type, int uppercase);
+
+/*
+ * Function that performs memory-hard hashing with certain degree of parallelism
+ * @param  context  Pointer to the Argon2 internal structure
+ * @return Error code if smth is wrong, ARGON2_OK otherwise
+ */
+ARGON2_PUBLIC int argon2_ctx(argon2_context *context, argon2_type type);
+
+/**
+ * Hashes a password with Argon2i, producing an encoded hash
+ * @param t_cost Number of iterations
+ * @param m_cost Sets memory usage to m_cost kibibytes
+ * @param parallelism Number of threads and compute lanes
+ * @param pwd Pointer to password
+ * @param pwdlen Password size in bytes
+ * @param salt Pointer to salt
+ * @param saltlen Salt size in bytes
+ * @param hashlen Desired length of the hash in bytes
+ * @param encoded Buffer where to write the encoded hash
+ * @param encodedlen Size of the buffer (thus max size of the encoded hash)
+ * @pre   Different parallelism levels will give different results
+ * @pre   Returns ARGON2_OK if successful
+ */
+ARGON2_PUBLIC int argon2i_hash_encoded(const uint32_t t_cost,
+                                       const uint32_t m_cost,
+                                       const uint32_t parallelism,
+                                       const void *pwd, const size_t pwdlen,
+                                       const void *salt, const size_t saltlen,
+                                       const size_t hashlen, char *encoded,
+                                       const size_t encodedlen);
+
+/**
+ * Hashes a password with Argon2i, producing a raw hash by allocating memory at
+ * @hash
+ * @param t_cost Number of iterations
+ * @param m_cost Sets memory usage to m_cost kibibytes
+ * @param parallelism Number of threads and compute lanes
+ * @param pwd Pointer to password
+ * @param pwdlen Password size in bytes
+ * @param salt Pointer to salt
+ * @param saltlen Salt size in bytes
+ * @param hash Buffer where to write the raw hash - updated by the function
+ * @param hashlen Desired length of the hash in bytes
+ * @pre   Different parallelism levels will give different results
+ * @pre   Returns ARGON2_OK if successful
+ */
+ARGON2_PUBLIC int argon2i_hash_raw(const uint32_t t_cost, const uint32_t m_cost,
+                                   const uint32_t parallelism, const void *pwd,
+                                   const size_t pwdlen, const void *salt,
+                                   const size_t saltlen, void *hash,
+                                   const size_t hashlen);
+
+ARGON2_PUBLIC int argon2d_hash_encoded(const uint32_t t_cost,
+                                       const uint32_t m_cost,
+                                       const uint32_t parallelism,
+                                       const void *pwd, const size_t pwdlen,
+                                       const void *salt, const size_t saltlen,
+                                       const size_t hashlen, char *encoded,
+                                       const size_t encodedlen);
+
+ARGON2_PUBLIC int argon2d_hash_raw(const uint32_t t_cost,
+                                   const uint32_t m_cost,
+                                   const uint32_t parallelism, const void *pwd,
+                                   const size_t pwdlen, const void *salt,
+                                   const size_t saltlen, void *hash,
+                                   const size_t hashlen);
+
+ARGON2_PUBLIC int argon2id_hash_encoded(const uint32_t t_cost,
+                                        const uint32_t m_cost,
+                                        const uint32_t parallelism,
+                                        const void *pwd, const size_t pwdlen,
+                                        const void *salt, const size_t saltlen,
+                                        const size_t hashlen, char *encoded,
+                                        const size_t encodedlen);
+
+ARGON2_PUBLIC int argon2id_hash_raw(const uint32_t t_cost,
+                                    const uint32_t m_cost,
+                                    const uint32_t parallelism, const void *pwd,
+                                    const size_t pwdlen, const void *salt,
+                                    const size_t saltlen, void *hash,
+                                    const size_t hashlen);
+
+ARGON2_PUBLIC int argon2id_hash_raw_ex(const uint32_t t_cost,
+                                       const uint32_t m_cost,
+                                       const uint32_t parallelism, const void *pwd,
+                                       const size_t pwdlen, const void *salt,
+                                       const size_t saltlen, void *hash,
+                                       const size_t hashlen,
+                                       void *memory);
+
+/* generic function underlying the above ones */
+ARGON2_PUBLIC int argon2_hash(const uint32_t t_cost, const uint32_t m_cost,
+                              const uint32_t parallelism, const void *pwd,
+                              const size_t pwdlen, const void *salt,
+                              const size_t saltlen, void *hash,
+                              const size_t hashlen, char *encoded,
+                              const size_t encodedlen, argon2_type type,
+                              const uint32_t version);
+
+/**
+ * Verifies a password against an encoded string
+ * Encoded string is restricted as in validate_inputs()
+ * @param encoded String encoding parameters, salt, hash
+ * @param pwd Pointer to password
+ * @pre   Returns ARGON2_OK if successful
+ */
+ARGON2_PUBLIC int argon2i_verify(const char *encoded, const void *pwd,
+                                 const size_t pwdlen);
+
+ARGON2_PUBLIC int argon2d_verify(const char *encoded, const void *pwd,
+                                 const size_t pwdlen);
+
+ARGON2_PUBLIC int argon2id_verify(const char *encoded, const void *pwd,
+                                  const size_t pwdlen);
+
+/* generic function underlying the above ones */
+ARGON2_PUBLIC int argon2_verify(const char *encoded, const void *pwd,
+                                const size_t pwdlen, argon2_type type);
+
+/**
+ * Argon2d: Version of Argon2 that picks memory blocks depending
+ * on the password and salt. Only for side-channel-free
+ * environment!!
+ *****
+ * @param  context  Pointer to current Argon2 context
+ * @return  Zero if successful, a non zero error code otherwise
+ */
+ARGON2_PUBLIC int argon2d_ctx(argon2_context *context);
+
+/**
+ * Argon2i: Version of Argon2 that picks memory blocks
+ * independent on the password and salt. Good for side-channels,
+ * but worse w.r.t. tradeoff attacks if only one pass is used.
+ *****
+ * @param  context  Pointer to current Argon2 context
+ * @return  Zero if successful, a non zero error code otherwise
+ */
+ARGON2_PUBLIC int argon2i_ctx(argon2_context *context);
+
+/**
+ * Argon2id: Version of Argon2 where the first half-pass over memory is
+ * password-independent, the rest are password-dependent (on the password and
+ * salt). OK against side channels (they reduce to 1/2-pass Argon2i), and
+ * better with w.r.t. tradeoff attacks (similar to Argon2d).
+ *****
+ * @param  context  Pointer to current Argon2 context
+ * @return  Zero if successful, a non zero error code otherwise
+ */
+ARGON2_PUBLIC int argon2id_ctx(argon2_context *context);
+
+/**
+ * Verify if a given password is correct for Argon2d hashing
+ * @param  context  Pointer to current Argon2 context
+ * @param  hash  The password hash to verify. The length of the hash is
+ * specified by the context outlen member
+ * @return  Zero if successful, a non zero error code otherwise
+ */
+ARGON2_PUBLIC int argon2d_verify_ctx(argon2_context *context, const char *hash);
+
+/**
+ * Verify if a given password is correct for Argon2i hashing
+ * @param  context  Pointer to current Argon2 context
+ * @param  hash  The password hash to verify. The length of the hash is
+ * specified by the context outlen member
+ * @return  Zero if successful, a non zero error code otherwise
+ */
+ARGON2_PUBLIC int argon2i_verify_ctx(argon2_context *context, const char *hash);
+
+/**
+ * Verify if a given password is correct for Argon2id hashing
+ * @param  context  Pointer to current Argon2 context
+ * @param  hash  The password hash to verify. The length of the hash is
+ * specified by the context outlen member
+ * @return  Zero if successful, a non zero error code otherwise
+ */
+ARGON2_PUBLIC int argon2id_verify_ctx(argon2_context *context,
+                                      const char *hash);
+
+/* generic function underlying the above ones */
+ARGON2_PUBLIC int argon2_verify_ctx(argon2_context *context, const char *hash,
+                                    argon2_type type);
+
+/**
+ * Get the associated error message for given error code
+ * @return  The error message associated with the given error code
+ */
+ARGON2_PUBLIC const char *argon2_error_message(int error_code);
+
+/**
+ * Returns the encoded hash length for the given input parameters
+ * @param t_cost  Number of iterations
+ * @param m_cost  Memory usage in kibibytes
+ * @param parallelism  Number of threads; used to compute lanes
+ * @param saltlen  Salt size in bytes
+ * @param hashlen  Hash size in bytes
+ * @param type The argon2_type that we want the encoded length for
+ * @return  The encoded hash length in bytes
+ */
+ARGON2_PUBLIC size_t argon2_encodedlen(uint32_t t_cost, uint32_t m_cost,
+                                       uint32_t parallelism, uint32_t saltlen,
+                                       uint32_t hashlen, argon2_type type);
+
+/* signals availability of argon2_select_impl: */
+#define ARGON2_SELECTABLE_IMPL
+
+/**
+ * Selects the fastest available optimized implementation.
+ * @param out The file for debug output (e. g. stderr; pass NULL for no
+ * debug output)
+ * @param prefix What to print before each line; NULL is equivalent to empty
+ * string
+ */
+ARGON2_PUBLIC void argon2_select_impl();
+ARGON2_PUBLIC const char *argon2_get_impl_name();
+ARGON2_PUBLIC int argon2_select_impl_by_name(const char *name);
+
+/* signals support for passing preallocated memory: */
+#define ARGON2_PREALLOCATED_MEMORY
+
+ARGON2_PUBLIC size_t argon2_memory_size(uint32_t m_cost, uint32_t parallelism);
+
+/**
+ * Function that performs memory-hard hashing with certain degree of parallelism
+ * @param context       Pointer to the Argon2 internal structure
+ * @param type          The Argon2 type
+ * @param memory        Preallocated memory for blocks (or NULL)
+ * @param memory_size   The size of preallocated memory
+ * @return Error code if smth is wrong, ARGON2_OK otherwise
+ */
+ARGON2_PUBLIC int argon2_ctx_mem(argon2_context *context, argon2_type type,
+                                 void *memory, size_t memory_size);
+
+#if defined(__cplusplus)
+}
+#endif
+
+#endif

src/3rdparty/argon2/lib/argon2-template-64.h

@@ -0,0 +1,193 @@
+#include <string.h>
+
+#include "core.h"
+
+#define MASK_32 UINT64_C(0xFFFFFFFF)
+
+#define F(x, y) ((x) + (y) + 2 * ((x) & MASK_32) * ((y) & MASK_32))
+
+#define G(a, b, c, d) \
+    do { \
+        a = F(a, b); \
+        d = rotr64(d ^ a, 32); \
+        c = F(c, d); \
+        b = rotr64(b ^ c, 24); \
+        a = F(a, b); \
+        d = rotr64(d ^ a, 16); \
+        c = F(c, d); \
+        b = rotr64(b ^ c, 63); \
+    } while ((void)0, 0)
+
+#define BLAKE2_ROUND_NOMSG(v0, v1, v2, v3, v4, v5, v6, v7, \
+                           v8, v9, v10, v11, v12, v13, v14, v15) \
+    do { \
+        G(v0, v4, v8,  v12); \
+        G(v1, v5, v9,  v13); \
+        G(v2, v6, v10, v14); \
+        G(v3, v7, v11, v15); \
+        G(v0, v5, v10, v15); \
+        G(v1, v6, v11, v12); \
+        G(v2, v7, v8,  v13); \
+        G(v3, v4, v9,  v14); \
+    } while ((void)0, 0)
+
+#define BLAKE2_ROUND_NOMSG1(v) \
+    BLAKE2_ROUND_NOMSG( \
+        (v)[ 0], (v)[ 1], (v)[ 2], (v)[ 3], \
+        (v)[ 4], (v)[ 5], (v)[ 6], (v)[ 7], \
+        (v)[ 8], (v)[ 9], (v)[10], (v)[11], \
+        (v)[12], (v)[13], (v)[14], (v)[15])
+
+#define BLAKE2_ROUND_NOMSG2(v) \
+    BLAKE2_ROUND_NOMSG( \
+        (v)[  0], (v)[  1], (v)[ 16], (v)[ 17], \
+        (v)[ 32], (v)[ 33], (v)[ 48], (v)[ 49], \
+        (v)[ 64], (v)[ 65], (v)[ 80], (v)[ 81], \
+        (v)[ 96], (v)[ 97], (v)[112], (v)[113])
+
+static void fill_block(const block *prev_block, const block *ref_block,
+                       block *next_block, int with_xor)
+{
+    block blockR, block_tmp;
+
+    copy_block(&blockR, ref_block);
+    xor_block(&blockR, prev_block);
+    copy_block(&block_tmp, &blockR);
+    if (with_xor) {
+        xor_block(&block_tmp, next_block);
+    }
+
+    /* Apply Blake2 on columns of 64-bit words: (0,1,...,15) , then
+    (16,17,..31)... finally (112,113,...127) */
+    BLAKE2_ROUND_NOMSG1(blockR.v + 0 * 16);
+    BLAKE2_ROUND_NOMSG1(blockR.v + 1 * 16);
+    BLAKE2_ROUND_NOMSG1(blockR.v + 2 * 16);
+    BLAKE2_ROUND_NOMSG1(blockR.v + 3 * 16);
+    BLAKE2_ROUND_NOMSG1(blockR.v + 4 * 16);
+    BLAKE2_ROUND_NOMSG1(blockR.v + 5 * 16);
+    BLAKE2_ROUND_NOMSG1(blockR.v + 6 * 16);
+    BLAKE2_ROUND_NOMSG1(blockR.v + 7 * 16);
+
+    /* Apply Blake2 on rows of 64-bit words: (0,1,16,17,...112,113), then
+    (2,3,18,19,...,114,115).. finally (14,15,30,31,...,126,127) */
+    BLAKE2_ROUND_NOMSG2(blockR.v + 0 * 2);
+    BLAKE2_ROUND_NOMSG2(blockR.v + 1 * 2);
+    BLAKE2_ROUND_NOMSG2(blockR.v + 2 * 2);
+    BLAKE2_ROUND_NOMSG2(blockR.v + 3 * 2);
+    BLAKE2_ROUND_NOMSG2(blockR.v + 4 * 2);
+    BLAKE2_ROUND_NOMSG2(blockR.v + 5 * 2);
+    BLAKE2_ROUND_NOMSG2(blockR.v + 6 * 2);
+    BLAKE2_ROUND_NOMSG2(blockR.v + 7 * 2);
+
+    copy_block(next_block, &block_tmp);
+    xor_block(next_block, &blockR);
+}
+
+static void next_addresses(block *address_block, block *input_block,
+                           const block *zero_block)
+{
+    input_block->v[6]++;
+    fill_block(zero_block, input_block, address_block, 0);
+    fill_block(zero_block, address_block, address_block, 0);
+}
+
+static void fill_segment_64(const argon2_instance_t *instance,
+                            argon2_position_t position)
+{
+    block *ref_block, *curr_block, *prev_block;
+    block address_block, input_block, zero_block;
+    uint64_t pseudo_rand, ref_index, ref_lane;
+    uint32_t prev_offset, curr_offset;
+    uint32_t starting_index, i;
+    int data_independent_addressing;
+
+    if (instance == NULL) {
+        return;
+    }
+
+    data_independent_addressing = (instance->type == Argon2_i) ||
+            (instance->type == Argon2_id && (position.pass == 0) &&
+             (position.slice < ARGON2_SYNC_POINTS / 2));
+
+    if (data_independent_addressing) {
+        init_block_value(&zero_block, 0);
+        init_block_value(&input_block, 0);
+
+        input_block.v[0] = position.pass;
+        input_block.v[1] = position.lane;
+        input_block.v[2] = position.slice;
+        input_block.v[3] = instance->memory_blocks;
+        input_block.v[4] = instance->passes;
+        input_block.v[5] = instance->type;
+    }
+
+    starting_index = 0;
+
+    if ((0 == position.pass) && (0 == position.slice)) {
+        starting_index = 2; /* we have already generated the first two blocks */
+
+        /* Don't forget to generate the first block of addresses: */
+        if (data_independent_addressing) {
+            next_addresses(&address_block, &input_block, &zero_block);
+        }
+    }
+
+    /* Offset of the current block */
+    curr_offset = position.lane * instance->lane_length +
+                  position.slice * instance->segment_length + starting_index;
+
+    if (0 == curr_offset % instance->lane_length) {
+        /* Last block in this lane */
+        prev_offset = curr_offset + instance->lane_length - 1;
+    } else {
+        /* Previous block */
+        prev_offset = curr_offset - 1;
+    }
+
+    for (i = starting_index; i < instance->segment_length;
+         ++i, ++curr_offset, ++prev_offset) {
+        /*1.1 Rotating prev_offset if needed */
+        if (curr_offset % instance->lane_length == 1) {
+            prev_offset = curr_offset - 1;
+        }
+
+        /* 1.2 Computing the index of the reference block */
+        /* 1.2.1 Taking pseudo-random value from the previous block */
+        if (data_independent_addressing) {
+            if (i % ARGON2_ADDRESSES_IN_BLOCK == 0) {
+                next_addresses(&address_block, &input_block, &zero_block);
+            }
+            pseudo_rand = address_block.v[i % ARGON2_ADDRESSES_IN_BLOCK];
+        } else {
+            pseudo_rand = instance->memory[prev_offset].v[0];
+        }
+
+        /* 1.2.2 Computing the lane of the reference block */
+        ref_lane = ((pseudo_rand >> 32)) % instance->lanes;
+
+        if ((position.pass == 0) && (position.slice == 0)) {
+            /* Can not reference other lanes yet */
+            ref_lane = position.lane;
+        }
+
+        /* 1.2.3 Computing the number of possible reference block within the
+         * lane.
+         */
+        position.index = i;
+        ref_index = index_alpha(instance, &position, pseudo_rand & 0xFFFFFFFF,
+                                ref_lane == position.lane);
+
+        /* 2 Creating a new block */
+        ref_block =
+            instance->memory + instance->lane_length * ref_lane + ref_index;
+        curr_block = instance->memory + curr_offset;
+        prev_block = instance->memory + prev_offset;
+
+        /* version 1.2.1 and earlier: overwrite, not XOR */
+        if (0 == position.pass || ARGON2_VERSION_10 == instance->version) {
+            fill_block(prev_block, ref_block, curr_block, 0);
+        } else {
+            fill_block(prev_block, ref_block, curr_block, 1);
+        }
+    }
+}

src/3rdparty/argon2/lib/argon2.c

@@ -0,0 +1,504 @@
+/*
+ * Argon2 source code package
+ *
+ * Written by Daniel Dinu and Dmitry Khovratovich, 2015
+ *
+ * This work is licensed under a Creative Commons CC0 1.0 License/Waiver.
+ *
+ * You should have received a copy of the CC0 Public Domain Dedication along
+ * with
+ * this software. If not, see
+ * <http://creativecommons.org/publicdomain/zero/1.0/>.
+ */
+
+#include <string.h>
+#include <stdlib.h>
+#include <stdio.h>
+
+#include "argon2.h"
+#include "encoding.h"
+#include "core.h"
+
+const char *argon2_type2string(argon2_type type, int uppercase) {
+    switch (type) {
+        case Argon2_d:
+            return uppercase ? "Argon2d" : "argon2d";
+        case Argon2_i:
+            return uppercase ? "Argon2i" : "argon2i";
+        case Argon2_id:
+            return uppercase ? "Argon2id" : "argon2id";
+    }
+
+    return NULL;
+}
+
+static void argon2_compute_memory_blocks(uint32_t *memory_blocks,
+                                         uint32_t *segment_length,
+                                         uint32_t m_cost, uint32_t lanes)
+{
+    /* Minimum memory_blocks = 8L blocks, where L is the number of lanes */
+    *memory_blocks = m_cost;
+    if (*memory_blocks < 2 * ARGON2_SYNC_POINTS * lanes) {
+        *memory_blocks = 2 * ARGON2_SYNC_POINTS * lanes;
+    }
+
+    *segment_length = *memory_blocks / (lanes * ARGON2_SYNC_POINTS);
+    /* Ensure that all segments have equal length */
+    *memory_blocks = *segment_length * (lanes * ARGON2_SYNC_POINTS);
+}
+
+size_t argon2_memory_size(uint32_t m_cost, uint32_t parallelism) {
+    uint32_t memory_blocks, segment_length;
+    argon2_compute_memory_blocks(&memory_blocks, &segment_length, m_cost,
+                                 parallelism);
+    return memory_blocks * ARGON2_BLOCK_SIZE;
+}
+
+int argon2_ctx_mem(argon2_context *context, argon2_type type, void *memory,
+                   size_t memory_size) {
+    /* 1. Validate all inputs */
+    int result = validate_inputs(context);
+    uint32_t memory_blocks, segment_length;
+    argon2_instance_t instance;
+
+    if (ARGON2_OK != result) {
+        return result;
+    }
+
+    if (Argon2_d != type && Argon2_i != type && Argon2_id != type) {
+        return ARGON2_INCORRECT_TYPE;
+    }
+
+    /* 2. Align memory size */
+    argon2_compute_memory_blocks(&memory_blocks, &segment_length,
+                                 context->m_cost, context->lanes);
+
+    /* check for sufficient memory size: */
+    if (memory != NULL && (memory_size % ARGON2_BLOCK_SIZE != 0 ||
+                           memory_size / ARGON2_BLOCK_SIZE < memory_blocks)) {
+        return ARGON2_MEMORY_ALLOCATION_ERROR;
+    }
+
+    instance.version = context->version;
+    instance.memory = (block *)memory;
+    instance.passes = context->t_cost;
+    instance.memory_blocks = memory_blocks;
+    instance.segment_length = segment_length;
+    instance.lane_length = segment_length * ARGON2_SYNC_POINTS;
+    instance.lanes = context->lanes;
+    instance.threads = context->threads;
+    instance.type = type;
+    instance.print_internals = !!(context->flags & ARGON2_FLAG_GENKAT);
+    instance.keep_memory = memory != NULL;
+
+    if (instance.threads > instance.lanes) {
+        instance.threads = instance.lanes;
+    }
+
+    /* 3. Initialization: Hashing inputs, allocating memory, filling first
+     * blocks
+     */
+    result = initialize(&instance, context);
+
+    if (ARGON2_OK != result) {
+        return result;
+    }
+
+    /* 4. Filling memory */
+    result = fill_memory_blocks(&instance);
+
+    if (ARGON2_OK != result) {
+        return result;
+    }
+    /* 5. Finalization */
+    finalize(context, &instance);
+
+    return ARGON2_OK;
+}
+
+int argon2_ctx(argon2_context *context, argon2_type type) {
+    return argon2_ctx_mem(context, type, NULL, 0);
+}
+
+int argon2_hash(const uint32_t t_cost, const uint32_t m_cost,
+                const uint32_t parallelism, const void *pwd,
+                const size_t pwdlen, const void *salt, const size_t saltlen,
+                void *hash, const size_t hashlen, char *encoded,
+                const size_t encodedlen, argon2_type type,
+                const uint32_t version){
+
+    argon2_context context;
+    int result;
+    uint8_t *out;
+
+    if (pwdlen > ARGON2_MAX_PWD_LENGTH) {
+        return ARGON2_PWD_TOO_LONG;
+    }
+
+    if (saltlen > ARGON2_MAX_SALT_LENGTH) {
+        return ARGON2_SALT_TOO_LONG;
+    }
+
+    if (hashlen > ARGON2_MAX_OUTLEN) {
+        return ARGON2_OUTPUT_TOO_LONG;
+    }
+
+    if (hashlen < ARGON2_MIN_OUTLEN) {
+        return ARGON2_OUTPUT_TOO_SHORT;
+    }
+
+    out = malloc(hashlen);
+    if (!out) {
+        return ARGON2_MEMORY_ALLOCATION_ERROR;
+    }
+
+    context.out = (uint8_t *)out;
+    context.outlen = (uint32_t)hashlen;
+    context.pwd = CONST_CAST(uint8_t *)pwd;
+    context.pwdlen = (uint32_t)pwdlen;
+    context.salt = CONST_CAST(uint8_t *)salt;
+    context.saltlen = (uint32_t)saltlen;
+    context.secret = NULL;
+    context.secretlen = 0;
+    context.ad = NULL;
+    context.adlen = 0;
+    context.t_cost = t_cost;
+    context.m_cost = m_cost;
+    context.lanes = parallelism;
+    context.threads = parallelism;
+    context.allocate_cbk = NULL;
+    context.free_cbk = NULL;
+    context.flags = ARGON2_DEFAULT_FLAGS;
+    context.version = version;
+
+    result = argon2_ctx(&context, type);
+
+    if (result != ARGON2_OK) {
+        clear_internal_memory(out, hashlen);
+        free(out);
+        return result;
+    }
+
+    /* if raw hash requested, write it */
+    if (hash) {
+        memcpy(hash, out, hashlen);
+    }
+
+    /* if encoding requested, write it */
+    if (encoded && encodedlen) {
+        if (encode_string(encoded, encodedlen, &context, type) != ARGON2_OK) {
+            clear_internal_memory(out, hashlen); /* wipe buffers if error */
+            clear_internal_memory(encoded, encodedlen);
+            free(out);
+            return ARGON2_ENCODING_FAIL;
+        }
+    }
+    clear_internal_memory(out, hashlen);
+    free(out);
+
+    return ARGON2_OK;
+}
+
+int argon2i_hash_encoded(const uint32_t t_cost, const uint32_t m_cost,
+                         const uint32_t parallelism, const void *pwd,
+                         const size_t pwdlen, const void *salt,
+                         const size_t saltlen, const size_t hashlen,
+                         char *encoded, const size_t encodedlen) {
+
+    return argon2_hash(t_cost, m_cost, parallelism, pwd, pwdlen, salt, saltlen,
+                       NULL, hashlen, encoded, encodedlen, Argon2_i,
+                       ARGON2_VERSION_NUMBER);
+}
+
+int argon2i_hash_raw(const uint32_t t_cost, const uint32_t m_cost,
+                     const uint32_t parallelism, const void *pwd,
+                     const size_t pwdlen, const void *salt,
+                     const size_t saltlen, void *hash, const size_t hashlen) {
+
+    return argon2_hash(t_cost, m_cost, parallelism, pwd, pwdlen, salt, saltlen,
+                       hash, hashlen, NULL, 0, Argon2_i, ARGON2_VERSION_NUMBER);
+}
+
+int argon2d_hash_encoded(const uint32_t t_cost, const uint32_t m_cost,
+                         const uint32_t parallelism, const void *pwd,
+                         const size_t pwdlen, const void *salt,
+                         const size_t saltlen, const size_t hashlen,
+                         char *encoded, const size_t encodedlen) {
+
+    return argon2_hash(t_cost, m_cost, parallelism, pwd, pwdlen, salt, saltlen,
+                       NULL, hashlen, encoded, encodedlen, Argon2_d,
+                       ARGON2_VERSION_NUMBER);
+}
+
+int argon2d_hash_raw(const uint32_t t_cost, const uint32_t m_cost,
+                     const uint32_t parallelism, const void *pwd,
+                     const size_t pwdlen, const void *salt,
+                     const size_t saltlen, void *hash, const size_t hashlen) {
+
+    return argon2_hash(t_cost, m_cost, parallelism, pwd, pwdlen, salt, saltlen,
+                       hash, hashlen, NULL, 0, Argon2_d, ARGON2_VERSION_NUMBER);
+}
+
+int argon2id_hash_encoded(const uint32_t t_cost, const uint32_t m_cost,
+                          const uint32_t parallelism, const void *pwd,
+                          const size_t pwdlen, const void *salt,
+                          const size_t saltlen, const size_t hashlen,
+                          char *encoded, const size_t encodedlen) {
+
+    return argon2_hash(t_cost, m_cost, parallelism, pwd, pwdlen, salt, saltlen,
+                       NULL, hashlen, encoded, encodedlen, Argon2_id,
+                       ARGON2_VERSION_NUMBER);
+}
+
+int argon2id_hash_raw(const uint32_t t_cost, const uint32_t m_cost,
+                      const uint32_t parallelism, const void *pwd,
+                      const size_t pwdlen, const void *salt,
+                      const size_t saltlen, void *hash, const size_t hashlen) {
+    return argon2_hash(t_cost, m_cost, parallelism, pwd, pwdlen, salt, saltlen,
+                       hash, hashlen, NULL, 0, Argon2_id,
+                       ARGON2_VERSION_NUMBER);
+}
+
+int argon2id_hash_raw_ex(const uint32_t t_cost, const uint32_t m_cost,
+                         const uint32_t parallelism, const void *pwd,
+                         const size_t pwdlen, const void *salt,
+                         const size_t saltlen, void *hash, const size_t hashlen, void *memory) {
+    argon2_context context;
+
+    context.out = (uint8_t *)hash;
+    context.outlen = (uint32_t)hashlen;
+    context.pwd = CONST_CAST(uint8_t *)pwd;
+    context.pwdlen = (uint32_t)pwdlen;
+    context.salt = CONST_CAST(uint8_t *)salt;
+    context.saltlen = (uint32_t)saltlen;
+    context.secret = NULL;
+    context.secretlen = 0;
+    context.ad = NULL;
+    context.adlen = 0;
+    context.t_cost = t_cost;
+    context.m_cost = m_cost;
+    context.lanes = parallelism;
+    context.threads = parallelism;
+    context.allocate_cbk = NULL;
+    context.free_cbk = NULL;
+    context.flags = ARGON2_DEFAULT_FLAGS;
+    context.version = ARGON2_VERSION_NUMBER;
+
+    return argon2_ctx_mem(&context, Argon2_id, memory, m_cost * 1024);
+}
+
+static int argon2_compare(const uint8_t *b1, const uint8_t *b2, size_t len) {
+    size_t i;
+    uint8_t d = 0U;
+
+    for (i = 0U; i < len; i++) {
+        d |= b1[i] ^ b2[i];
+    }
+    return (int)((1 & ((d - 1) >> 8)) - 1);
+}
+
+int argon2_verify(const char *encoded, const void *pwd, const size_t pwdlen,
+                  argon2_type type) {
+
+    argon2_context ctx;
+    uint8_t *desired_result = NULL;
+
+    int ret = ARGON2_OK;
+
+    size_t encoded_len;
+    uint32_t max_field_len;
+
+    if (pwdlen > ARGON2_MAX_PWD_LENGTH) {
+        return ARGON2_PWD_TOO_LONG;
+    }
+
+    if (encoded == NULL) {
+        return ARGON2_DECODING_FAIL;
+    }
+
+    encoded_len = strlen(encoded);
+    if (encoded_len > UINT32_MAX) {
+        return ARGON2_DECODING_FAIL;
+    }
+
+    /* No field can be longer than the encoded length */
+    max_field_len = (uint32_t)encoded_len;
+
+    ctx.saltlen = max_field_len;
+    ctx.outlen = max_field_len;
+
+    ctx.salt = malloc(ctx.saltlen);
+    ctx.out = malloc(ctx.outlen);
+    if (!ctx.salt || !ctx.out) {
+        ret = ARGON2_MEMORY_ALLOCATION_ERROR;
+        goto fail;
+    }
+
+    ctx.pwd = (uint8_t *)pwd;
+    ctx.pwdlen = (uint32_t)pwdlen;
+
+    ret = decode_string(&ctx, encoded, type);
+    if (ret != ARGON2_OK) {
+        goto fail;
+    }
+
+    /* Set aside the desired result, and get a new buffer. */
+    desired_result = ctx.out;
+    ctx.out = malloc(ctx.outlen);
+    if (!ctx.out) {
+        ret = ARGON2_MEMORY_ALLOCATION_ERROR;
+        goto fail;
+    }
+
+    ret = argon2_verify_ctx(&ctx, (char *)desired_result, type);
+    if (ret != ARGON2_OK) {
+        goto fail;
+    }
+
+fail:
+    free(ctx.salt);
+    free(ctx.out);
+    free(desired_result);
+
+    return ret;
+}
+
+int argon2i_verify(const char *encoded, const void *pwd, const size_t pwdlen) {
+
+    return argon2_verify(encoded, pwd, pwdlen, Argon2_i);
+}
+
+int argon2d_verify(const char *encoded, const void *pwd, const size_t pwdlen) {
+
+    return argon2_verify(encoded, pwd, pwdlen, Argon2_d);
+}
+
+int argon2id_verify(const char *encoded, const void *pwd, const size_t pwdlen) {
+
+    return argon2_verify(encoded, pwd, pwdlen, Argon2_id);
+}
+
+int argon2d_ctx(argon2_context *context) {
+    return argon2_ctx(context, Argon2_d);
+}
+
+int argon2i_ctx(argon2_context *context) {
+    return argon2_ctx(context, Argon2_i);
+}
+
+int argon2id_ctx(argon2_context *context) {
+    return argon2_ctx(context, Argon2_id);
+}
+
+int argon2_verify_ctx(argon2_context *context, const char *hash,
+                      argon2_type type) {
+    int ret = argon2_ctx(context, type);
+    if (ret != ARGON2_OK) {
+        return ret;
+    }
+
+    if (argon2_compare((uint8_t *)hash, context->out, context->outlen)) {
+        return ARGON2_VERIFY_MISMATCH;
+    }
+
+    return ARGON2_OK;
+}
+
+int argon2d_verify_ctx(argon2_context *context, const char *hash) {
+    return argon2_verify_ctx(context, hash, Argon2_d);
+}
+
+int argon2i_verify_ctx(argon2_context *context, const char *hash) {
+    return argon2_verify_ctx(context, hash, Argon2_i);
+}
+
+int argon2id_verify_ctx(argon2_context *context, const char *hash) {
+    return argon2_verify_ctx(context, hash, Argon2_id);
+}
+
+const char *argon2_error_message(int error_code) {
+    switch (error_code) {
+    case ARGON2_OK:
+        return "OK";
+    case ARGON2_OUTPUT_PTR_NULL:
+        return "Output pointer is NULL";
+    case ARGON2_OUTPUT_TOO_SHORT:
+        return "Output is too short";
+    case ARGON2_OUTPUT_TOO_LONG:
+        return "Output is too long";
+    case ARGON2_PWD_TOO_SHORT:
+        return "Password is too short";
+    case ARGON2_PWD_TOO_LONG:
+        return "Password is too long";
+    case ARGON2_SALT_TOO_SHORT:
+        return "Salt is too short";
+    case ARGON2_SALT_TOO_LONG:
+        return "Salt is too long";
+    case ARGON2_AD_TOO_SHORT:
+        return "Associated data is too short";
+    case ARGON2_AD_TOO_LONG:
+        return "Associated data is too long";
+    case ARGON2_SECRET_TOO_SHORT:
+        return "Secret is too short";
+    case ARGON2_SECRET_TOO_LONG:
+        return "Secret is too long";
+    case ARGON2_TIME_TOO_SMALL:
+        return "Time cost is too small";
+    case ARGON2_TIME_TOO_LARGE:
+        return "Time cost is too large";
+    case ARGON2_MEMORY_TOO_LITTLE:
+        return "Memory cost is too small";
+    case ARGON2_MEMORY_TOO_MUCH:
+        return "Memory cost is too large";
+    case ARGON2_LANES_TOO_FEW:
+        return "Too few lanes";
+    case ARGON2_LANES_TOO_MANY:
+        return "Too many lanes";
+    case ARGON2_PWD_PTR_MISMATCH:
+        return "Password pointer is NULL, but password length is not 0";
+    case ARGON2_SALT_PTR_MISMATCH:
+        return "Salt pointer is NULL, but salt length is not 0";
+    case ARGON2_SECRET_PTR_MISMATCH:
+        return "Secret pointer is NULL, but secret length is not 0";
+    case ARGON2_AD_PTR_MISMATCH:
+        return "Associated data pointer is NULL, but ad length is not 0";
+    case ARGON2_MEMORY_ALLOCATION_ERROR:
+        return "Memory allocation error";
+    case ARGON2_FREE_MEMORY_CBK_NULL:
+        return "The free memory callback is NULL";
+    case ARGON2_ALLOCATE_MEMORY_CBK_NULL:
+        return "The allocate memory callback is NULL";
+    case ARGON2_INCORRECT_PARAMETER:
+        return "Argon2_Context context is NULL";
+    case ARGON2_INCORRECT_TYPE:
+        return "There is no such version of Argon2";
+    case ARGON2_OUT_PTR_MISMATCH:
+        return "Output pointer mismatch";
+    case ARGON2_THREADS_TOO_FEW:
+        return "Not enough threads";
+    case ARGON2_THREADS_TOO_MANY:
+        return "Too many threads";
+    case ARGON2_MISSING_ARGS:
+        return "Missing arguments";
+    case ARGON2_ENCODING_FAIL:
+        return "Encoding failed";
+    case ARGON2_DECODING_FAIL:
+        return "Decoding failed";
+    case ARGON2_THREAD_FAIL:
+        return "Threading failure";
+    case ARGON2_DECODING_LENGTH_FAIL:
+        return "Some of encoded parameters are too long or too short";
+    case ARGON2_VERIFY_MISMATCH:
+        return "The password does not match the supplied hash";
+    default:
+        return "Unknown error code";
+    }
+}
+
+size_t argon2_encodedlen(uint32_t t_cost, uint32_t m_cost, uint32_t parallelism,
+                         uint32_t saltlen, uint32_t hashlen, argon2_type type) {
+    return strlen("$$v=$m=,t=,p=$$") + strlen(argon2_type2string(type, 0)) +
+            numlen(t_cost) + numlen(m_cost) + numlen(parallelism) +
+            b64len(saltlen) + b64len(hashlen) + numlen(ARGON2_VERSION_NUMBER) +
+            1;
+}

src/3rdparty/argon2/lib/blake2/blake2-impl.h

@@ -0,0 +1,90 @@
+#ifndef ARGON2_BLAKE2_IMPL_H
+#define ARGON2_BLAKE2_IMPL_H
+
+#include <stdint.h>
+
+/* Argon2 Team - Begin Code */
+/*
+   Not an exhaustive list, but should cover the majority of modern platforms
+   Additionally, the code will always be correct---this is only a performance
+   tweak.
+*/
+#if (defined(__BYTE_ORDER__) &&                                                \
+     (__BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__)) ||                           \
+    defined(__LITTLE_ENDIAN__) || defined(__ARMEL__) || defined(__MIPSEL__) || \
+    defined(__AARCH64EL__) || defined(__amd64__) || defined(__i386__) ||       \
+    defined(_M_IX86) || defined(_M_X64) || defined(_M_AMD64) ||                \
+    defined(_M_ARM)
+#define NATIVE_LITTLE_ENDIAN
+#endif
+/* Argon2 Team - End Code */
+
+static inline uint32_t load32(const void *src) {
+#if defined(NATIVE_LITTLE_ENDIAN)
+    return *(const uint32_t *)src;
+#else
+    const uint8_t *p = (const uint8_t *)src;
+    uint32_t w = *p++;
+    w |= (uint32_t)(*p++) << 8;
+    w |= (uint32_t)(*p++) << 16;
+    w |= (uint32_t)(*p++) << 24;
+    return w;
+#endif
+}
+
+static inline uint64_t load64(const void *src) {
+#if defined(NATIVE_LITTLE_ENDIAN)
+    return *(const uint64_t *)src;
+#else
+    const uint8_t *p = (const uint8_t *)src;
+    uint64_t w = *p++;
+    w |= (uint64_t)(*p++) << 8;
+    w |= (uint64_t)(*p++) << 16;
+    w |= (uint64_t)(*p++) << 24;
+    w |= (uint64_t)(*p++) << 32;
+    w |= (uint64_t)(*p++) << 40;
+    w |= (uint64_t)(*p++) << 48;
+    w |= (uint64_t)(*p++) << 56;
+    return w;
+#endif
+}
+
+static inline void store32(void *dst, uint32_t w) {
+#if defined(NATIVE_LITTLE_ENDIAN)
+    *(uint32_t *)dst = w;
+#else
+    uint8_t *p = (uint8_t *)dst;
+    *p++ = (uint8_t)w;
+    w >>= 8;
+    *p++ = (uint8_t)w;
+    w >>= 8;
+    *p++ = (uint8_t)w;
+    w >>= 8;
+    *p++ = (uint8_t)w;
+#endif
+}
+
+static inline void store64(void *dst, uint64_t w) {
+#if defined(NATIVE_LITTLE_ENDIAN)
+    *(uint64_t *)dst = w;
+#else
+    uint8_t *p = (uint8_t *)dst;
+    *p++ = (uint8_t)w;
+    w >>= 8;
+    *p++ = (uint8_t)w;
+    w >>= 8;
+    *p++ = (uint8_t)w;
+    w >>= 8;
+    *p++ = (uint8_t)w;
+    w >>= 8;
+    *p++ = (uint8_t)w;
+    w >>= 8;
+    *p++ = (uint8_t)w;
+    w >>= 8;
+    *p++ = (uint8_t)w;
+    w >>= 8;
+    *p++ = (uint8_t)w;
+#endif
+}
+
+#endif // ARGON2_BLAKE2_IMPL_H

src/3rdparty/argon2/lib/blake2/blake2.c

@@ -0,0 +1,225 @@
+#include <string.h>
+
+#include "blake2/blake2.h"
+#include "blake2/blake2-impl.h"
+
+#include "core.h"
+
+static const uint64_t blake2b_IV[8] = {
+    UINT64_C(0x6a09e667f3bcc908), UINT64_C(0xbb67ae8584caa73b),
+    UINT64_C(0x3c6ef372fe94f82b), UINT64_C(0xa54ff53a5f1d36f1),
+    UINT64_C(0x510e527fade682d1), UINT64_C(0x9b05688c2b3e6c1f),
+    UINT64_C(0x1f83d9abfb41bd6b), UINT64_C(0x5be0cd19137e2179)
+};
+
+#define rotr64(x, n) (((x) >> (n)) | ((x) << (64 - (n))))
+
+static const unsigned int blake2b_sigma[12][16] = {
+    {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15},
+    {14, 10, 4, 8, 9, 15, 13, 6, 1, 12, 0, 2, 11, 7, 5, 3},
+    {11, 8, 12, 0, 5, 2, 15, 13, 10, 14, 3, 6, 7, 1, 9, 4},
+    {7, 9, 3, 1, 13, 12, 11, 14, 2, 6, 5, 10, 4, 0, 15, 8},
+    {9, 0, 5, 7, 2, 4, 10, 15, 14, 1, 11, 12, 6, 8, 3, 13},
+    {2, 12, 6, 10, 0, 11, 8, 3, 4, 13, 7, 5, 15, 14, 1, 9},
+    {12, 5, 1, 15, 14, 13, 4, 10, 0, 7, 6, 3, 9, 2, 8, 11},
+    {13, 11, 7, 14, 12, 1, 3, 9, 5, 0, 15, 4, 8, 6, 2, 10},
+    {6, 15, 14, 9, 11, 3, 0, 8, 12, 2, 13, 7, 1, 4, 10, 5},
+    {10, 2, 8, 4, 7, 6, 1, 5, 15, 11, 9, 14, 3, 12, 13, 0},
+    {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15},
+    {14, 10, 4, 8, 9, 15, 13, 6, 1, 12, 0, 2, 11, 7, 5, 3},
+};
+
+#define G(m, r, i, a, b, c, d) \
+    do { \
+        a = a + b + m[blake2b_sigma[r][2 * i + 0]]; \
+        d = rotr64(d ^ a, 32); \
+        c = c + d; \
+        b = rotr64(b ^ c, 24); \
+        a = a + b + m[blake2b_sigma[r][2 * i + 1]]; \
+        d = rotr64(d ^ a, 16); \
+        c = c + d; \
+        b = rotr64(b ^ c, 63); \
+    } while ((void)0, 0)
+
+#define ROUND(m, v, r) \
+    do { \
+        G(m, r, 0, v[0], v[4], v[ 8], v[12]); \
+        G(m, r, 1, v[1], v[5], v[ 9], v[13]); \
+        G(m, r, 2, v[2], v[6], v[10], v[14]); \
+        G(m, r, 3, v[3], v[7], v[11], v[15]); \
+        G(m, r, 4, v[0], v[5], v[10], v[15]); \
+        G(m, r, 5, v[1], v[6], v[11], v[12]); \
+        G(m, r, 6, v[2], v[7], v[ 8], v[13]); \
+        G(m, r, 7, v[3], v[4], v[ 9], v[14]); \
+    } while ((void)0, 0)
+
+void blake2b_compress(blake2b_state *S, const void *block, uint64_t f0)
+{
+    uint64_t m[16];
+    uint64_t v[16];
+
+    m[ 0] = load64((const uint64_t *)block +  0);
+    m[ 1] = load64((const uint64_t *)block +  1);
+    m[ 2] = load64((const uint64_t *)block +  2);
+    m[ 3] = load64((const uint64_t *)block +  3);
+    m[ 4] = load64((const uint64_t *)block +  4);
+    m[ 5] = load64((const uint64_t *)block +  5);
+    m[ 6] = load64((const uint64_t *)block +  6);
+    m[ 7] = load64((const uint64_t *)block +  7);
+    m[ 8] = load64((const uint64_t *)block +  8);
+    m[ 9] = load64((const uint64_t *)block +  9);
+    m[10] = load64((const uint64_t *)block + 10);
+    m[11] = load64((const uint64_t *)block + 11);
+    m[12] = load64((const uint64_t *)block + 12);
+    m[13] = load64((const uint64_t *)block + 13);
+    m[14] = load64((const uint64_t *)block + 14);
+    m[15] = load64((const uint64_t *)block + 15);
+
+    v[ 0] = S->h[0];
+    v[ 1] = S->h[1];
+    v[ 2] = S->h[2];
+    v[ 3] = S->h[3];
+    v[ 4] = S->h[4];
+    v[ 5] = S->h[5];
+    v[ 6] = S->h[6];
+    v[ 7] = S->h[7];
+    v[ 8] = blake2b_IV[0];
+    v[ 9] = blake2b_IV[1];
+    v[10] = blake2b_IV[2];
+    v[11] = blake2b_IV[3];
+    v[12] = blake2b_IV[4] ^ S->t[0];
+    v[13] = blake2b_IV[5] ^ S->t[1];
+    v[14] = blake2b_IV[6] ^ f0;
+    v[15] = blake2b_IV[7];
+
+    ROUND(m, v, 0);
+    ROUND(m, v, 1);
+    ROUND(m, v, 2);
+    ROUND(m, v, 3);
+    ROUND(m, v, 4);
+    ROUND(m, v, 5);
+    ROUND(m, v, 6);
+    ROUND(m, v, 7);
+    ROUND(m, v, 8);
+    ROUND(m, v, 9);
+    ROUND(m, v, 10);
+    ROUND(m, v, 11);
+
+    S->h[0] ^= v[0] ^ v[ 8];
+    S->h[1] ^= v[1] ^ v[ 9];
+    S->h[2] ^= v[2] ^ v[10];
+    S->h[3] ^= v[3] ^ v[11];
+    S->h[4] ^= v[4] ^ v[12];
+    S->h[5] ^= v[5] ^ v[13];
+    S->h[6] ^= v[6] ^ v[14];
+    S->h[7] ^= v[7] ^ v[15];
+}
+
+static void blake2b_increment_counter(blake2b_state *S, uint64_t inc)
+{
+    S->t[0] += inc;
+    S->t[1] += (S->t[0] < inc);
+}
+
+static void blake2b_init_state(blake2b_state *S)
+{
+    memcpy(S->h, blake2b_IV, sizeof(S->h));
+    S->t[1] = S->t[0] = 0;
+    S->buflen = 0;
+}
+
+void blake2b_init(blake2b_state *S, size_t outlen)
+{
+    blake2b_init_state(S);
+    /* XOR initial state with param block: */
+    S->h[0] ^= (uint64_t)outlen | (UINT64_C(1) << 16) | (UINT64_C(1) << 24);
+}
+
+void blake2b_update(blake2b_state *S, const void *in, size_t inlen)
+{
+    const uint8_t *pin = (const uint8_t *)in;
+
+    if (S->buflen + inlen > BLAKE2B_BLOCKBYTES) {
+        size_t left = S->buflen;
+        size_t fill = BLAKE2B_BLOCKBYTES - left;
+        memcpy(&S->buf[left], pin, fill);
+        blake2b_increment_counter(S, BLAKE2B_BLOCKBYTES);
+        blake2b_compress(S, S->buf, 0);
+        S->buflen = 0;
+        inlen -= fill;
+        pin += fill;
+        /* Avoid buffer copies when possible */
+        while (inlen > BLAKE2B_BLOCKBYTES) {
+            blake2b_increment_counter(S, BLAKE2B_BLOCKBYTES);
+            blake2b_compress(S, pin, 0);
+            inlen -= BLAKE2B_BLOCKBYTES;
+            pin += BLAKE2B_BLOCKBYTES;
+        }
+    }
+    memcpy(&S->buf[S->buflen], pin, inlen);
+    S->buflen += inlen;
+}
+
+void blake2b_final(blake2b_state *S, void *out, size_t outlen)
+{
+    uint8_t buffer[BLAKE2B_OUTBYTES] = {0};
+    unsigned int i;
+
+    blake2b_increment_counter(S, S->buflen);
+    memset(&S->buf[S->buflen], 0, BLAKE2B_BLOCKBYTES - S->buflen); /* Padding */
+    blake2b_compress(S, S->buf, UINT64_C(0xFFFFFFFFFFFFFFFF));
+
+    for (i = 0; i < 8; ++i) { /* Output full hash to temp buffer */
+        store64(buffer + i * sizeof(uint64_t), S->h[i]);
+    }
+
+    memcpy(out, buffer, outlen);
+    clear_internal_memory(buffer, sizeof(buffer));
+    clear_internal_memory(S->buf, sizeof(S->buf));
+    clear_internal_memory(S->h, sizeof(S->h));
+}
+
+void blake2b_long(void *out, size_t outlen, const void *in, size_t inlen)
+{
+    uint8_t *pout = (uint8_t *)out;
+    blake2b_state blake_state;
+    uint8_t outlen_bytes[sizeof(uint32_t)] = {0};
+
+    store32(outlen_bytes, (uint32_t)outlen);
+    if (outlen <= BLAKE2B_OUTBYTES) {
+        blake2b_init(&blake_state, outlen);
+        blake2b_update(&blake_state, outlen_bytes, sizeof(outlen_bytes));
+        blake2b_update(&blake_state, in, inlen);
+        blake2b_final(&blake_state, pout, outlen);
+    } else {
+        uint32_t toproduce;
+        uint8_t out_buffer[BLAKE2B_OUTBYTES];
+
+        blake2b_init(&blake_state, BLAKE2B_OUTBYTES);
+        blake2b_update(&blake_state, outlen_bytes, sizeof(outlen_bytes));
+        blake2b_update(&blake_state, in, inlen);
+        blake2b_final(&blake_state, out_buffer, BLAKE2B_OUTBYTES);
+
+        memcpy(pout, out_buffer, BLAKE2B_OUTBYTES / 2);
+        pout += BLAKE2B_OUTBYTES / 2;
+        toproduce = (uint32_t)outlen - BLAKE2B_OUTBYTES / 2;
+
+        while (toproduce > BLAKE2B_OUTBYTES) {
+            blake2b_init(&blake_state, BLAKE2B_OUTBYTES);
+            blake2b_update(&blake_state, out_buffer, BLAKE2B_OUTBYTES);
+            blake2b_final(&blake_state, out_buffer, BLAKE2B_OUTBYTES);
+
+            memcpy(pout, out_buffer, BLAKE2B_OUTBYTES / 2);
+            pout += BLAKE2B_OUTBYTES / 2;
+            toproduce -= BLAKE2B_OUTBYTES / 2;
+        }
+
+        blake2b_init(&blake_state, toproduce);
+        blake2b_update(&blake_state, out_buffer, BLAKE2B_OUTBYTES);
+        blake2b_final(&blake_state, out_buffer, toproduce);
+
+        memcpy(pout, out_buffer, toproduce);
+
+        clear_internal_memory(out_buffer, sizeof(out_buffer));
+    }
+}

src/3rdparty/argon2/lib/blake2/blake2.h

@@ -0,0 +1,30 @@
+#ifndef ARGON2_BLAKE2_H
+#define ARGON2_BLAKE2_H
+
+#include <stddef.h>
+#include <stdint.h>
+
+enum blake2b_constant {
+    BLAKE2B_BLOCKBYTES = 128,
+    BLAKE2B_OUTBYTES = 64,
+    BLAKE2B_KEYBYTES = 64,
+    BLAKE2B_SALTBYTES = 16,
+    BLAKE2B_PERSONALBYTES = 16
+};
+
+typedef struct __blake2b_state {
+    uint64_t h[8];
+    uint64_t t[2];
+    uint8_t buf[BLAKE2B_BLOCKBYTES];
+    size_t buflen;
+} blake2b_state;
+
+/* Streaming API */
+void blake2b_init(blake2b_state *S, size_t outlen);
+void blake2b_update(blake2b_state *S, const void *in, size_t inlen);
+void blake2b_final(blake2b_state *S, void *out, size_t outlen);
+
+void blake2b_long(void *out, size_t outlen, const void *in, size_t inlen);
+
+#endif // ARGON2_BLAKE2_H
+

src/3rdparty/argon2/lib/core.c

@@ -0,0 +1,539 @@
+/*
+ * Argon2 source code package
+ *
+ * Written by Daniel Dinu and Dmitry Khovratovich, 2015
+ *
+ * This work is licensed under a Creative Commons CC0 1.0 License/Waiver.
+ *
+ * You should have received a copy of the CC0 Public Domain Dedication along
+ * with
+ * this software. If not, see
+ * <http://creativecommons.org/publicdomain/zero/1.0/>.
+ */
+
+/*For memory wiping*/
+#ifdef _MSC_VER
+#include <windows.h>
+#include <winbase.h> /* For SecureZeroMemory */
+#endif
+#if defined __STDC_LIB_EXT1__
+#define __STDC_WANT_LIB_EXT1__ 1
+#endif
+#define VC_GE_2005(version) (version >= 1400)
+
+#include <inttypes.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "core.h"
+#include "blake2/blake2.h"
+#include "blake2/blake2-impl.h"
+
+#include "genkat.h"
+
+#if defined(__clang__)
+#if __has_attribute(optnone)
+#define NOT_OPTIMIZED __attribute__((optnone))
+#endif
+#elif defined(__GNUC__)
+#define GCC_VERSION                                                            \
+    (__GNUC__ * 10000 + __GNUC_MINOR__ * 100 + __GNUC_PATCHLEVEL__)
+#if GCC_VERSION >= 40400
+#define NOT_OPTIMIZED __attribute__((optimize("O0")))
+#endif
+#endif
+#ifndef NOT_OPTIMIZED
+#define NOT_OPTIMIZED
+#endif
+
+/***************Instance and Position constructors**********/
+void init_block_value(block *b, uint8_t in) { memset(b->v, in, sizeof(b->v)); }
+
+void copy_block(block *dst, const block *src) {
+    memcpy(dst->v, src->v, sizeof(uint64_t) * ARGON2_QWORDS_IN_BLOCK);
+}
+
+void xor_block(block *dst, const block *src) {
+    int i;
+    for (i = 0; i < ARGON2_QWORDS_IN_BLOCK; ++i) {
+        dst->v[i] ^= src->v[i];
+    }
+}
+
+static void load_block(block *dst, const void *input) {
+    unsigned i;
+    for (i = 0; i < ARGON2_QWORDS_IN_BLOCK; ++i) {
+        dst->v[i] = load64((const uint8_t *)input + i * sizeof(dst->v[i]));
+    }
+}
+
+static void store_block(void *output, const block *src) {
+    unsigned i;
+    for (i = 0; i < ARGON2_QWORDS_IN_BLOCK; ++i) {
+        store64((uint8_t *)output + i * sizeof(src->v[i]), src->v[i]);
+    }
+}
+
+/***************Memory functions*****************/
+
+int allocate_memory(const argon2_context *context,
+                    argon2_instance_t *instance) {
+    size_t blocks = instance->memory_blocks;
+    size_t memory_size = blocks * ARGON2_BLOCK_SIZE;
+
+    /* 0. Check for memory supplied by user: */
+    /* NOTE: Sufficient memory size is already checked in argon2_ctx_mem() */
+    if (instance->memory != NULL) {
+        return ARGON2_OK;
+    }
+
+    /* 1. Check for multiplication overflow */
+    if (blocks != 0 && memory_size / ARGON2_BLOCK_SIZE != blocks) {
+        return ARGON2_MEMORY_ALLOCATION_ERROR;
+    }
+
+    /* 2. Try to allocate with appropriate allocator */
+    if (context->allocate_cbk) {
+        (context->allocate_cbk)((uint8_t **)&instance->memory, memory_size);
+    } else {
+        instance->memory = malloc(memory_size);
+    }
+
+    if (instance->memory == NULL) {
+        return ARGON2_MEMORY_ALLOCATION_ERROR;
+    }
+
+    return ARGON2_OK;
+}
+
+void free_memory(const argon2_context *context,
+                 const argon2_instance_t *instance) {
+    size_t memory_size = instance->memory_blocks * ARGON2_BLOCK_SIZE;
+
+    clear_internal_memory(instance->memory, memory_size);
+
+    if (instance->keep_memory) {
+        /* user-supplied memory -- do not free */
+        return;
+    }
+
+    if (context->free_cbk) {
+        (context->free_cbk)((uint8_t *)instance->memory, memory_size);
+    } else {
+        free(instance->memory);
+    }
+}
+
+void NOT_OPTIMIZED secure_wipe_memory(void *v, size_t n) {
+#if defined(_MSC_VER) && VC_GE_2005(_MSC_VER)
+    SecureZeroMemory(v, n);
+#elif defined memset_s
+    memset_s(v, n, 0, n);
+#elif defined(__OpenBSD__)
+    explicit_bzero(v, n);
+#else
+    static void *(*const volatile memset_sec)(void *, int, size_t) = &memset;
+    memset_sec(v, 0, n);
+#endif
+}
+
+/* Memory clear flag defaults to true. */
+int FLAG_clear_internal_memory = 0;
+void clear_internal_memory(void *v, size_t n) {
+    if (FLAG_clear_internal_memory && v) {
+        secure_wipe_memory(v, n);
+    }
+}
+
+void finalize(const argon2_context *context, argon2_instance_t *instance) {
+    if (context != NULL && instance != NULL) {
+        block blockhash;
+        uint32_t l;
+
+        copy_block(&blockhash, instance->memory + instance->lane_length - 1);
+
+        /* XOR the last blocks */
+        for (l = 1; l < instance->lanes; ++l) {
+            uint32_t last_block_in_lane =
+                l * instance->lane_length + (instance->lane_length - 1);
+            xor_block(&blockhash, instance->memory + last_block_in_lane);
+        }
+
+        /* Hash the result */
+        {
+            uint8_t blockhash_bytes[ARGON2_BLOCK_SIZE];
+            store_block(blockhash_bytes, &blockhash);
+            blake2b_long(context->out, context->outlen, blockhash_bytes,
+                         ARGON2_BLOCK_SIZE);
+            /* clear blockhash and blockhash_bytes */
+            clear_internal_memory(blockhash.v, ARGON2_BLOCK_SIZE);
+            clear_internal_memory(blockhash_bytes, ARGON2_BLOCK_SIZE);
+        }
+
+        if (instance->print_internals) {
+            print_tag(context->out, context->outlen);
+        }
+
+        free_memory(context, instance);
+    }
+}
+
+uint32_t index_alpha(const argon2_instance_t *instance,
+                     const argon2_position_t *position, uint32_t pseudo_rand,
+                     int same_lane) {
+    /*
+     * Pass 0:
+     *      This lane : all already finished segments plus already constructed
+     * blocks in this segment
+     *      Other lanes : all already finished segments
+     * Pass 1+:
+     *      This lane : (SYNC_POINTS - 1) last segments plus already constructed
+     * blocks in this segment
+     *      Other lanes : (SYNC_POINTS - 1) last segments
+     */
+    uint32_t reference_area_size;
+    uint64_t relative_position;
+    uint32_t start_position, absolute_position;
+
+    if (0 == position->pass) {
+        /* First pass */
+        if (0 == position->slice) {
+            /* First slice */
+            reference_area_size =
+                position->index - 1; /* all but the previous */
+        } else {
+            if (same_lane) {
+                /* The same lane => add current segment */
+                reference_area_size =
+                    position->slice * instance->segment_length +
+                    position->index - 1;
+            } else {
+                reference_area_size =
+                    position->slice * instance->segment_length +
+                    ((position->index == 0) ? (-1) : 0);
+            }
+        }
+    } else {
+        /* Second pass */
+        if (same_lane) {
+            reference_area_size = instance->lane_length -
+                                  instance->segment_length + position->index -
+                                  1;
+        } else {
+            reference_area_size = instance->lane_length -
+                                  instance->segment_length +
+                                  ((position->index == 0) ? (-1) : 0);
+        }
+    }
+
+    /* 1.2.4. Mapping pseudo_rand to 0..<reference_area_size-1> and produce
+     * relative position */
+    relative_position = pseudo_rand;
+    relative_position = relative_position * relative_position >> 32;
+    relative_position = reference_area_size - 1 -
+                        (reference_area_size * relative_position >> 32);
+
+    /* 1.2.5 Computing starting position */
+    start_position = 0;
+
+    if (0 != position->pass) {
+        start_position = (position->slice == ARGON2_SYNC_POINTS - 1)
+                             ? 0
+                             : (position->slice + 1) * instance->segment_length;
+    }
+
+    /* 1.2.6. Computing absolute position */
+    absolute_position = (start_position + relative_position) %
+                        instance->lane_length; /* absolute position */
+    return absolute_position;
+}
+
+/* Single-threaded version for p=1 case */
+static int fill_memory_blocks_st(argon2_instance_t *instance) {
+    uint32_t r, s, l;
+
+    for (r = 0; r < instance->passes; ++r) {
+        for (s = 0; s < ARGON2_SYNC_POINTS; ++s) {
+            for (l = 0; l < instance->lanes; ++l) {
+                argon2_position_t position = { r, l, (uint8_t)s, 0 };
+                fill_segment(instance, position);
+            }
+        }
+
+        if (instance->print_internals) {
+            internal_kat(instance, r); /* Print all memory blocks */
+        }
+    }
+    return ARGON2_OK;
+}
+
+int fill_memory_blocks(argon2_instance_t *instance) {
+    if (instance == NULL || instance->lanes == 0) {
+        return ARGON2_INCORRECT_PARAMETER;
+    }
+
+    return fill_memory_blocks_st(instance);
+}
+
+int validate_inputs(const argon2_context *context) {
+    if (NULL == context) {
+        return ARGON2_INCORRECT_PARAMETER;
+    }
+
+    if (NULL == context->out) {
+        return ARGON2_OUTPUT_PTR_NULL;
+    }
+
+    /* Validate output length */
+    if (ARGON2_MIN_OUTLEN > context->outlen) {
+        return ARGON2_OUTPUT_TOO_SHORT;
+    }
+
+    if (ARGON2_MAX_OUTLEN < context->outlen) {
+        return ARGON2_OUTPUT_TOO_LONG;
+    }
+
+    /* Validate password (required param) */
+    if (NULL == context->pwd) {
+        if (0 != context->pwdlen) {
+            return ARGON2_PWD_PTR_MISMATCH;
+        }
+    }
+
+    if (ARGON2_MIN_PWD_LENGTH > context->pwdlen) {
+        return ARGON2_PWD_TOO_SHORT;
+    }
+
+    if (ARGON2_MAX_PWD_LENGTH < context->pwdlen) {
+        return ARGON2_PWD_TOO_LONG;
+    }
+
+    /* Validate salt (required param) */
+    if (NULL == context->salt) {
+        if (0 != context->saltlen) {
+            return ARGON2_SALT_PTR_MISMATCH;
+        }
+    }
+
+    if (ARGON2_MIN_SALT_LENGTH > context->saltlen) {
+        return ARGON2_SALT_TOO_SHORT;
+    }
+
+    if (ARGON2_MAX_SALT_LENGTH < context->saltlen) {
+        return ARGON2_SALT_TOO_LONG;
+    }
+
+    /* Validate secret (optional param) */
+    if (NULL == context->secret) {
+        if (0 != context->secretlen) {
+            return ARGON2_SECRET_PTR_MISMATCH;
+        }
+    } else {
+        if (ARGON2_MIN_SECRET > context->secretlen) {
+            return ARGON2_SECRET_TOO_SHORT;
+        }
+        if (ARGON2_MAX_SECRET < context->secretlen) {
+            return ARGON2_SECRET_TOO_LONG;
+        }
+    }
+
+    /* Validate associated data (optional param) */
+    if (NULL == context->ad) {
+        if (0 != context->adlen) {
+            return ARGON2_AD_PTR_MISMATCH;
+        }
+    } else {
+        if (ARGON2_MIN_AD_LENGTH > context->adlen) {
+            return ARGON2_AD_TOO_SHORT;
+        }
+        if (ARGON2_MAX_AD_LENGTH < context->adlen) {
+            return ARGON2_AD_TOO_LONG;
+        }
+    }
+
+    /* Validate memory cost */
+    if (ARGON2_MIN_MEMORY > context->m_cost) {
+        return ARGON2_MEMORY_TOO_LITTLE;
+    }
+
+    if (ARGON2_MAX_MEMORY < context->m_cost) {
+        return ARGON2_MEMORY_TOO_MUCH;
+    }
+
+    if (context->m_cost < 8 * context->lanes) {
+        return ARGON2_MEMORY_TOO_LITTLE;
+    }
+
+    /* Validate time cost */
+    if (ARGON2_MIN_TIME > context->t_cost) {
+        return ARGON2_TIME_TOO_SMALL;
+    }
+
+    if (ARGON2_MAX_TIME < context->t_cost) {
+        return ARGON2_TIME_TOO_LARGE;
+    }
+
+    /* Validate lanes */
+    if (ARGON2_MIN_LANES > context->lanes) {
+        return ARGON2_LANES_TOO_FEW;
+    }
+
+    if (ARGON2_MAX_LANES < context->lanes) {
+        return ARGON2_LANES_TOO_MANY;
+    }
+
+    /* Validate threads */
+    if (ARGON2_MIN_THREADS > context->threads) {
+        return ARGON2_THREADS_TOO_FEW;
+    }
+
+    if (ARGON2_MAX_THREADS < context->threads) {
+        return ARGON2_THREADS_TOO_MANY;
+    }
+
+    if (NULL != context->allocate_cbk && NULL == context->free_cbk) {
+        return ARGON2_FREE_MEMORY_CBK_NULL;
+    }
+
+    if (NULL == context->allocate_cbk && NULL != context->free_cbk) {
+        return ARGON2_ALLOCATE_MEMORY_CBK_NULL;
+    }
+
+    return ARGON2_OK;
+}
+
+void fill_first_blocks(uint8_t *blockhash, const argon2_instance_t *instance) {
+    uint32_t l;
+    /* Make the first and second block in each lane as G(H0||0||i) or
+       G(H0||1||i) */
+    uint8_t blockhash_bytes[ARGON2_BLOCK_SIZE];
+    for (l = 0; l < instance->lanes; ++l) {
+
+        store32(blockhash + ARGON2_PREHASH_DIGEST_LENGTH, 0);
+        store32(blockhash + ARGON2_PREHASH_DIGEST_LENGTH + 4, l);
+        blake2b_long(blockhash_bytes, ARGON2_BLOCK_SIZE, blockhash,
+                     ARGON2_PREHASH_SEED_LENGTH);
+        load_block(&instance->memory[l * instance->lane_length + 0],
+                   blockhash_bytes);
+
+        store32(blockhash + ARGON2_PREHASH_DIGEST_LENGTH, 1);
+        blake2b_long(blockhash_bytes, ARGON2_BLOCK_SIZE, blockhash,
+                     ARGON2_PREHASH_SEED_LENGTH);
+        load_block(&instance->memory[l * instance->lane_length + 1],
+                   blockhash_bytes);
+    }
+    clear_internal_memory(blockhash_bytes, ARGON2_BLOCK_SIZE);
+}
+
+void initial_hash(uint8_t *blockhash, argon2_context *context,
+                  argon2_type type) {
+    blake2b_state BlakeHash;
+    uint8_t value[sizeof(uint32_t)];
+
+    if (NULL == context || NULL == blockhash) {
+        return;
+    }
+
+    blake2b_init(&BlakeHash, ARGON2_PREHASH_DIGEST_LENGTH);
+
+    store32(&value, context->lanes);
+    blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+    store32(&value, context->outlen);
+    blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+    store32(&value, context->m_cost);
+    blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+    store32(&value, context->t_cost);
+    blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+    store32(&value, context->version);
+    blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+    store32(&value, (uint32_t)type);
+    blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+    store32(&value, context->pwdlen);
+    blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+    if (context->pwd != NULL) {
+        blake2b_update(&BlakeHash, (const uint8_t *)context->pwd,
+                       context->pwdlen);
+
+        if (context->flags & ARGON2_FLAG_CLEAR_PASSWORD) {
+            secure_wipe_memory(context->pwd, context->pwdlen);
+            context->pwdlen = 0;
+        }
+    }
+
+    store32(&value, context->saltlen);
+    blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+    if (context->salt != NULL) {
+        blake2b_update(&BlakeHash, (const uint8_t *)context->salt,
+                       context->saltlen);
+    }
+
+    store32(&value, context->secretlen);
+    blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+    if (context->secret != NULL) {
+        blake2b_update(&BlakeHash, (const uint8_t *)context->secret,
+                       context->secretlen);
+
+        if (context->flags & ARGON2_FLAG_CLEAR_SECRET) {
+            secure_wipe_memory(context->secret, context->secretlen);
+            context->secretlen = 0;
+        }
+    }
+
+    store32(&value, context->adlen);
+    blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+    if (context->ad != NULL) {
+        blake2b_update(&BlakeHash, (const uint8_t *)context->ad,
+                       context->adlen);
+    }
+
+    blake2b_final(&BlakeHash, blockhash, ARGON2_PREHASH_DIGEST_LENGTH);
+}
+
+int initialize(argon2_instance_t *instance, argon2_context *context) {
+    uint8_t blockhash[ARGON2_PREHASH_SEED_LENGTH];
+    int result = ARGON2_OK;
+
+    if (instance == NULL || context == NULL)
+        return ARGON2_INCORRECT_PARAMETER;
+    instance->context_ptr = context;
+
+    /* 1. Memory allocation */
+
+    result = allocate_memory(context, instance);
+    if (result != ARGON2_OK) {
+        return result;
+    }
+
+    /* 2. Initial hashing */
+    /* H_0 + 8 extra bytes to produce the first blocks */
+    /* uint8_t blockhash[ARGON2_PREHASH_SEED_LENGTH]; */
+    /* Hashing all inputs */
+    initial_hash(blockhash, context, instance->type);
+    /* Zeroing 8 extra bytes */
+    clear_internal_memory(blockhash + ARGON2_PREHASH_DIGEST_LENGTH,
+                          ARGON2_PREHASH_SEED_LENGTH -
+                              ARGON2_PREHASH_DIGEST_LENGTH);
+
+    if (instance->print_internals) {
+        initial_kat(blockhash, context, instance->type);
+    }
+
+    /* 3. Creating first blocks, we always have at least two blocks in a slice
+     */
+    fill_first_blocks(blockhash, instance);
+    /* Clearing the hash */
+    clear_internal_memory(blockhash, ARGON2_PREHASH_SEED_LENGTH);
+
+    return ARGON2_OK;
+}

src/3rdparty/argon2/lib/core.h

@@ -0,0 +1,226 @@
+/*
+ * Argon2 source code package
+ *
+ * Written by Daniel Dinu and Dmitry Khovratovich, 2015
+ *
+ * This work is licensed under a Creative Commons CC0 1.0 License/Waiver.
+ *
+ * You should have received a copy of the CC0 Public Domain Dedication along
+ * with
+ * this software. If not, see
+ * <http://creativecommons.org/publicdomain/zero/1.0/>.
+ */
+
+#ifndef ARGON2_CORE_H
+#define ARGON2_CORE_H
+
+#include "argon2.h"
+
+#if defined(_MSC_VER)
+#define ALIGN(n) __declspec(align(16))
+#elif defined(__GNUC__) || defined(__clang)
+#define ALIGN(x) __attribute__((__aligned__(x)))
+#else
+#define ALIGN(x)
+#endif
+
+#define CONST_CAST(x) (x)(uintptr_t)
+
+/**********************Argon2 internal constants*******************************/
+
+enum argon2_core_constants {
+    /* Memory block size in bytes */
+    ARGON2_BLOCK_SIZE = 1024,
+    ARGON2_QWORDS_IN_BLOCK = ARGON2_BLOCK_SIZE / 8,
+    ARGON2_OWORDS_IN_BLOCK = ARGON2_BLOCK_SIZE / 16,
+
+    /* Number of pseudo-random values generated by one call to Blake in Argon2i
+       to
+       generate reference block positions */
+    ARGON2_ADDRESSES_IN_BLOCK = 128,
+
+    /* Pre-hashing digest length and its extension*/
+    ARGON2_PREHASH_DIGEST_LENGTH = 64,
+    ARGON2_PREHASH_SEED_LENGTH = 72
+};
+
+/*************************Argon2 internal data types***********************/
+
+/*
+ * Structure for the (1KB) memory block implemented as 128 64-bit words.
+ * Memory blocks can be copied, XORed. Internal words can be accessed by [] (no
+ * bounds checking).
+ */
+typedef struct block_ { uint64_t v[ARGON2_QWORDS_IN_BLOCK]; } block;
+
+/*****************Functions that work with the block******************/
+
+/* Initialize each byte of the block with @in */
+void init_block_value(block *b, uint8_t in);
+
+/* Copy block @src to block @dst */
+void copy_block(block *dst, const block *src);
+
+/* XOR @src onto @dst bytewise */
+void xor_block(block *dst, const block *src);
+
+/*
+ * Argon2 instance: memory pointer, number of passes, amount of memory, type,
+ * and derived values.
+ * Used to evaluate the number and location of blocks to construct in each
+ * thread
+ */
+typedef struct Argon2_instance_t {
+    block *memory;          /* Memory pointer */
+    uint32_t version;
+    uint32_t passes;        /* Number of passes */
+    uint32_t memory_blocks; /* Number of blocks in memory */
+    uint32_t segment_length;
+    uint32_t lane_length;
+    uint32_t lanes;
+    uint32_t threads;
+    argon2_type type;
+    int print_internals; /* whether to print the memory blocks */
+    int keep_memory;
+    argon2_context *context_ptr; /* points back to original context */
+} argon2_instance_t;
+
+/*
+ * Argon2 position: where we construct the block right now. Used to distribute
+ * work between threads.
+ */
+typedef struct Argon2_position_t {
+    uint32_t pass;
+    uint32_t lane;
+    uint8_t slice;
+    uint32_t index;
+} argon2_position_t;
+
+/*Struct that holds the inputs for thread handling FillSegment*/
+typedef struct Argon2_thread_data {
+    argon2_instance_t *instance_ptr;
+    argon2_position_t pos;
+} argon2_thread_data;
+
+/*************************Argon2 core functions********************************/
+
+/* Allocates memory to the given pointer, uses the appropriate allocator as
+ * specified in the context. Total allocated memory is num*size.
+ * @param context argon2_context which specifies the allocator
+ * @param instance the Argon2 instance
+ * @return ARGON2_OK if memory is allocated successfully
+ */
+int allocate_memory(const argon2_context *context,
+                    argon2_instance_t *instance);
+
+/*
+ * Frees memory at the given pointer, uses the appropriate deallocator as
+ * specified in the context. Also cleans the memory using clear_internal_memory.
+ * @param context argon2_context which specifies the deallocator
+ * @param instance the Argon2 instance
+ */
+void free_memory(const argon2_context *context,
+                 const argon2_instance_t *instance);
+
+/* Function that securely cleans the memory. This ignores any flags set
+ * regarding clearing memory. Usually one just calls clear_internal_memory.
+ * @param mem Pointer to the memory
+ * @param s Memory size in bytes
+ */
+void secure_wipe_memory(void *v, size_t n);
+
+/* Function that securely clears the memory if FLAG_clear_internal_memory is
+ * set. If the flag isn't set, this function does nothing.
+ * @param mem Pointer to the memory
+ * @param s Memory size in bytes
+ */
+ARGON2_PUBLIC void clear_internal_memory(void *v, size_t n);
+
+/*
+ * Computes absolute position of reference block in the lane following a skewed
+ * distribution and using a pseudo-random value as input
+ * @param instance Pointer to the current instance
+ * @param position Pointer to the current position
+ * @param pseudo_rand 32-bit pseudo-random value used to determine the position
+ * @param same_lane Indicates if the block will be taken from the current lane.
+ * If so we can reference the current segment
+ * @pre All pointers must be valid
+ */
+uint32_t index_alpha(const argon2_instance_t *instance,
+                     const argon2_position_t *position, uint32_t pseudo_rand,
+                     int same_lane);
+
+/*
+ * Function that validates all inputs against predefined restrictions and return
+ * an error code
+ * @param context Pointer to current Argon2 context
+ * @return ARGON2_OK if everything is all right, otherwise one of error codes
+ * (all defined in <argon2.h>
+ */
+int validate_inputs(const argon2_context *context);
+
+/*
+ * Hashes all the inputs into @a blockhash[PREHASH_DIGEST_LENGTH], clears
+ * password and secret if needed
+ * @param  context  Pointer to the Argon2 internal structure containing memory
+ * pointer, and parameters for time and space requirements.
+ * @param  blockhash Buffer for pre-hashing digest
+ * @param  type Argon2 type
+ * @pre    @a blockhash must have at least @a PREHASH_DIGEST_LENGTH bytes
+ * allocated
+ */
+void initial_hash(uint8_t *blockhash, argon2_context *context,
+                  argon2_type type);
+
+/*
+ * Function creates first 2 blocks per lane
+ * @param instance Pointer to the current instance
+ * @param blockhash Pointer to the pre-hashing digest
+ * @pre blockhash must point to @a PREHASH_SEED_LENGTH allocated values
+ */
+void fill_first_blocks(uint8_t *blockhash, const argon2_instance_t *instance);
+
+/*
+ * Function allocates memory, hashes the inputs with Blake,  and creates first
+ * two blocks. Returns the pointer to the main memory with 2 blocks per lane
+ * initialized
+ * @param  context  Pointer to the Argon2 internal structure containing memory
+ * pointer, and parameters for time and space requirements.
+ * @param  instance Current Argon2 instance
+ * @return Zero if successful, -1 if memory failed to allocate. @context->state
+ * will be modified if successful.
+ */
+int initialize(argon2_instance_t *instance, argon2_context *context);
+
+/*
+ * XORing the last block of each lane, hashing it, making the tag. Deallocates
+ * the memory.
+ * @param context Pointer to current Argon2 context (use only the out parameters
+ * from it)
+ * @param instance Pointer to current instance of Argon2
+ * @pre instance->state must point to necessary amount of memory
+ * @pre context->out must point to outlen bytes of memory
+ * @pre if context->free_cbk is not NULL, it should point to a function that
+ * deallocates memory
+ */
+void finalize(const argon2_context *context, argon2_instance_t *instance);
+
+/*
+ * Function that fills the segment using previous segments also from other
+ * threads
+ * @param instance Pointer to the current instance
+ * @param position Current position
+ * @pre all block pointers must be valid
+ */
+void fill_segment(const argon2_instance_t *instance,
+                  argon2_position_t position);
+
+/*
+ * Function that fills the entire memory t_cost times based on the first two
+ * blocks in each lane
+ * @param instance Pointer to the current instance
+ * @return ARGON2_OK if successful, @context->state
+ */
+int fill_memory_blocks(argon2_instance_t *instance);
+
+#endif

src/3rdparty/argon2/lib/encoding.c

@@ -0,0 +1,432 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <limits.h>
+#include "encoding.h"
+#include "core.h"
+
+/*
+ * Example code for a decoder and encoder of "hash strings", with Argon2
+ * parameters.
+ *
+ * This code comprises three sections:
+ *
+ *   -- The first section contains generic Base64 encoding and decoding
+ *   functions. It is conceptually applicable to any hash function
+ *   implementation that uses Base64 to encode and decode parameters,
+ *   salts and outputs. It could be made into a library, provided that
+ *   the relevant functions are made public (non-static) and be given
+ *   reasonable names to avoid collisions with other functions.
+ *
+ *   -- The second section is specific to Argon2. It encodes and decodes
+ *   the parameters, salts and outputs. It does not compute the hash
+ *   itself.
+ *
+ * The code was originally written by Thomas Pornin <pornin@bolet.org>,
+ * to whom comments and remarks may be sent. It is released under what
+ * should amount to Public Domain or its closest equivalent; the
+ * following mantra is supposed to incarnate that fact with all the
+ * proper legal rituals:
+ *
+ * ---------------------------------------------------------------------
+ * This file is provided under the terms of Creative Commons CC0 1.0
+ * Public Domain Dedication. To the extent possible under law, the
+ * author (Thomas Pornin) has waived all copyright and related or
+ * neighboring rights to this file. This work is published from: Canada.
+ * ---------------------------------------------------------------------
+ *
+ * Copyright (c) 2015 Thomas Pornin
+ */
+
+/* ==================================================================== */
+/*
+ * Common code; could be shared between different hash functions.
+ *
+ * Note: the Base64 functions below assume that uppercase letters (resp.
+ * lowercase letters) have consecutive numerical codes, that fit on 8
+ * bits. All modern systems use ASCII-compatible charsets, where these
+ * properties are true. If you are stuck with a dinosaur of a system
+ * that still defaults to EBCDIC then you already have much bigger
+ * interoperability issues to deal with.
+ */
+
+/*
+ * Some macros for constant-time comparisons. These work over values in
+ * the 0..255 range. Returned value is 0x00 on "false", 0xFF on "true".
+ */
+#define EQ(x, y) ((((0U - ((unsigned)(x) ^ (unsigned)(y))) >> 8) & 0xFF) ^ 0xFF)
+#define GT(x, y) ((((unsigned)(y) - (unsigned)(x)) >> 8) & 0xFF)
+#define GE(x, y) (GT(y, x) ^ 0xFF)
+#define LT(x, y) GT(y, x)
+#define LE(x, y) GE(y, x)
+
+/*
+ * Convert value x (0..63) to corresponding Base64 character.
+ */
+static int b64_byte_to_char(unsigned x) {
+    return (LT(x, 26) & (x + 'A')) |
+           (GE(x, 26) & LT(x, 52) & (x + ('a' - 26))) |
+           (GE(x, 52) & LT(x, 62) & (x + ('0' - 52))) | (EQ(x, 62) & '+') |
+           (EQ(x, 63) & '/');
+}
+
+/*
+ * Convert character c to the corresponding 6-bit value. If character c
+ * is not a Base64 character, then 0xFF (255) is returned.
+ */
+static unsigned b64_char_to_byte(int c) {
+    unsigned x;
+
+    x = (GE(c, 'A') & LE(c, 'Z') & (c - 'A')) |
+        (GE(c, 'a') & LE(c, 'z') & (c - ('a' - 26))) |
+        (GE(c, '0') & LE(c, '9') & (c - ('0' - 52))) | (EQ(c, '+') & 62) |
+        (EQ(c, '/') & 63);
+    return x | (EQ(x, 0) & (EQ(c, 'A') ^ 0xFF));
+}
+
+/*
+ * Convert some bytes to Base64. 'dst_len' is the length (in characters)
+ * of the output buffer 'dst'; if that buffer is not large enough to
+ * receive the result (including the terminating 0), then (size_t)-1
+ * is returned. Otherwise, the zero-terminated Base64 string is written
+ * in the buffer, and the output length (counted WITHOUT the terminating
+ * zero) is returned.
+ */
+static size_t to_base64(char *dst, size_t dst_len, const void *src,
+                        size_t src_len) {
+    size_t olen;
+    const unsigned char *buf;
+    unsigned acc, acc_len;
+
+    olen = (src_len / 3) << 2;
+    switch (src_len % 3) {
+    case 2:
+        olen++;
+    /* fall through */
+    case 1:
+        olen += 2;
+        break;
+    }
+    if (dst_len <= olen) {
+        return (size_t)-1;
+    }
+    acc = 0;
+    acc_len = 0;
+    buf = (const unsigned char *)src;
+    while (src_len-- > 0) {
+        acc = (acc << 8) + (*buf++);
+        acc_len += 8;
+        while (acc_len >= 6) {
+            acc_len -= 6;
+            *dst++ = (char)b64_byte_to_char((acc >> acc_len) & 0x3F);
+        }
+    }
+    if (acc_len > 0) {
+        *dst++ = (char)b64_byte_to_char((acc << (6 - acc_len)) & 0x3F);
+    }
+    *dst++ = 0;
+    return olen;
+}
+
+/*
+ * Decode Base64 chars into bytes. The '*dst_len' value must initially
+ * contain the length of the output buffer '*dst'; when the decoding
+ * ends, the actual number of decoded bytes is written back in
+ * '*dst_len'.
+ *
+ * Decoding stops when a non-Base64 character is encountered, or when
+ * the output buffer capacity is exceeded. If an error occurred (output
+ * buffer is too small, invalid last characters leading to unprocessed
+ * buffered bits), then NULL is returned; otherwise, the returned value
+ * points to the first non-Base64 character in the source stream, which
+ * may be the terminating zero.
+ */
+static const char *from_base64(void *dst, size_t *dst_len, const char *src) {
+    size_t len;
+    unsigned char *buf;
+    unsigned acc, acc_len;
+
+    buf = (unsigned char *)dst;
+    len = 0;
+    acc = 0;
+    acc_len = 0;
+    for (;;) {
+        unsigned d;
+
+        d = b64_char_to_byte(*src);
+        if (d == 0xFF) {
+            break;
+        }
+        src++;
+        acc = (acc << 6) + d;
+        acc_len += 6;
+        if (acc_len >= 8) {
+            acc_len -= 8;
+            if ((len++) >= *dst_len) {
+                return NULL;
+            }
+            *buf++ = (acc >> acc_len) & 0xFF;
+        }
+    }
+
+    /*
+     * If the input length is equal to 1 modulo 4 (which is
+     * invalid), then there will remain 6 unprocessed bits;
+     * otherwise, only 0, 2 or 4 bits are buffered. The buffered
+     * bits must also all be zero.
+     */
+    if (acc_len > 4 || (acc & (((unsigned)1 << acc_len) - 1)) != 0) {
+        return NULL;
+    }
+    *dst_len = len;
+    return src;
+}
+
+/*
+ * Decode decimal integer from 'str'; the value is written in '*v'.
+ * Returned value is a pointer to the next non-decimal character in the
+ * string. If there is no digit at all, or the value encoding is not
+ * minimal (extra leading zeros), or the value does not fit in an
+ * 'unsigned long', then NULL is returned.
+ */
+static const char *decode_decimal(const char *str, unsigned long *v) {
+    const char *orig;
+    unsigned long acc;
+
+    acc = 0;
+    for (orig = str;; str++) {
+        int c;
+
+        c = *str;
+        if (c < '0' || c > '9') {
+            break;
+        }
+        c -= '0';
+        if (acc > (ULONG_MAX / 10)) {
+            return NULL;
+        }
+        acc *= 10;
+        if ((unsigned long)c > (ULONG_MAX - acc)) {
+            return NULL;
+        }
+        acc += (unsigned long)c;
+    }
+    if (str == orig || (*orig == '0' && str != (orig + 1))) {
+        return NULL;
+    }
+    *v = acc;
+    return str;
+}
+
+/* ==================================================================== */
+/*
+ * Code specific to Argon2.
+ *
+ * The code below applies the following format:
+ *
+ *  $argon2<T>[$v=<num>]$m=<num>,t=<num>,p=<num>$<bin>$<bin>
+ *
+ * where <T> is either 'd', 'id', or 'i', <num> is a decimal integer (positive,
+ * fits in an 'unsigned long'), and <bin> is Base64-encoded data (no '=' padding
+ * characters, no newline or whitespace).
+ *
+ * The last two binary chunks (encoded in Base64) are, in that order,
+ * the salt and the output. Both are required. The binary salt length and the
+ * output length must be in the allowed ranges defined in argon2.h.
+ *
+ * The ctx struct must contain buffers large enough to hold the salt and pwd
+ * when it is fed into decode_string.
+ */
+
+int decode_string(argon2_context *ctx, const char *str, argon2_type type) {
+
+/* check for prefix */
+#define CC(prefix)                                                             \
+    do {                                                                       \
+        size_t cc_len = strlen(prefix);                                        \
+        if (strncmp(str, prefix, cc_len) != 0) {                               \
+            return ARGON2_DECODING_FAIL;                                       \
+        }                                                                      \
+        str += cc_len;                                                         \
+    } while ((void)0, 0)
+
+/* optional prefix checking with supplied code */
+#define CC_opt(prefix, code)                                                   \
+    do {                                                                       \
+        size_t cc_len = strlen(prefix);                                        \
+        if (strncmp(str, prefix, cc_len) == 0) {                               \
+            str += cc_len;                                                     \
+            { code; }                                                          \
+        }                                                                      \
+    } while ((void)0, 0)
+
+/* Decoding prefix into uint32_t decimal */
+#define DECIMAL_U32(x)                                                         \
+    do {                                                                       \
+        unsigned long dec_x;                                                   \
+        str = decode_decimal(str, &dec_x);                                     \
+        if (str == NULL || dec_x > UINT32_MAX) {                               \
+            return ARGON2_DECODING_FAIL;                                       \
+        }                                                                      \
+        (x) = (uint32_t)dec_x;                                                           \
+    } while ((void)0, 0)
+
+/* Decoding base64 into a binary buffer */
+#define BIN(buf, max_len, len)                                                 \
+    do {                                                                       \
+        size_t bin_len = (max_len);                                            \
+        str = from_base64(buf, &bin_len, str);                                 \
+        if (str == NULL || bin_len > UINT32_MAX) {                             \
+            return ARGON2_DECODING_FAIL;                                       \
+        }                                                                      \
+        (len) = (uint32_t)bin_len;                                             \
+    } while ((void)0, 0)
+
+    size_t maxsaltlen = ctx->saltlen;
+    size_t maxoutlen = ctx->outlen;
+    int validation_result;
+    const char* type_string;
+
+    /* We should start with the argon2_type we are using */
+    type_string = argon2_type2string(type, 0);
+    if (!type_string) {
+        return ARGON2_INCORRECT_TYPE;
+    }
+
+    CC("$");
+    CC(type_string);
+
+    /* Reading the version number if the default is suppressed */
+    ctx->version = ARGON2_VERSION_10;
+    CC_opt("$v=", DECIMAL_U32(ctx->version));
+
+    CC("$m=");
+    DECIMAL_U32(ctx->m_cost);
+    CC(",t=");
+    DECIMAL_U32(ctx->t_cost);
+    CC(",p=");
+    DECIMAL_U32(ctx->lanes);
+    ctx->threads = ctx->lanes;
+
+    CC("$");
+    BIN(ctx->salt, maxsaltlen, ctx->saltlen);
+    CC("$");
+    BIN(ctx->out, maxoutlen, ctx->outlen);
+
+    /* The rest of the fields get the default values */
+    ctx->secret = NULL;
+    ctx->secretlen = 0;
+    ctx->ad = NULL;
+    ctx->adlen = 0;
+    ctx->allocate_cbk = NULL;
+    ctx->free_cbk = NULL;
+    ctx->flags = ARGON2_DEFAULT_FLAGS;
+
+    /* On return, must have valid context */
+    validation_result = validate_inputs(ctx);
+    if (validation_result != ARGON2_OK) {
+        return validation_result;
+    }
+
+    /* Can't have any additional characters */
+    if (*str == 0) {
+        return ARGON2_OK;
+    } else {
+        return ARGON2_DECODING_FAIL;
+    }
+#undef CC
+#undef CC_opt
+#undef DECIMAL_U32
+#undef BIN
+}
+
+int encode_string(char *dst, size_t dst_len, argon2_context *ctx,
+                  argon2_type type) {
+#define SS(str)                                                                \
+    do {                                                                       \
+        size_t pp_len = strlen(str);                                           \
+        if (pp_len >= dst_len) {                                               \
+            return ARGON2_ENCODING_FAIL;                                       \
+        }                                                                      \
+        memcpy(dst, str, pp_len + 1);                                          \
+        dst += pp_len;                                                         \
+        dst_len -= pp_len;                                                     \
+    } while ((void)0, 0)
+
+#define SX(x)                                                                  \
+    do {                                                                       \
+        char tmp[30];                                                          \
+        sprintf(tmp, "%lu", (unsigned long)(x));                               \
+        SS(tmp);                                                               \
+    } while ((void)0, 0)
+
+#define SB(buf, len)                                                           \
+    do {                                                                       \
+        size_t sb_len = to_base64(dst, dst_len, buf, len);                     \
+        if (sb_len == (size_t)-1) {                                            \
+            return ARGON2_ENCODING_FAIL;                                       \
+        }                                                                      \
+        dst += sb_len;                                                         \
+        dst_len -= sb_len;                                                     \
+    } while ((void)0, 0)
+
+    const char* type_string = argon2_type2string(type, 0);
+    int validation_result = validate_inputs(ctx);
+
+    if (!type_string) {
+        return ARGON2_ENCODING_FAIL;
+    }
+
+    if (validation_result != ARGON2_OK) {
+        return validation_result;
+    }
+
+    SS("$");
+    SS(type_string);
+
+    SS("$v=");
+    SX(ctx->version);
+
+    SS("$m=");
+    SX(ctx->m_cost);
+    SS(",t=");
+    SX(ctx->t_cost);
+    SS(",p=");
+    SX(ctx->lanes);
+
+    SS("$");
+    SB(ctx->salt, ctx->saltlen);
+
+    SS("$");
+    SB(ctx->out, ctx->outlen);
+    return ARGON2_OK;
+
+#undef SS
+#undef SX
+#undef SB
+}
+
+size_t b64len(uint32_t len) {
+    size_t olen = ((size_t)len / 3) << 2;
+
+    switch (len % 3) {
+    case 2:
+        olen++;
+    /* fall through */
+    case 1:
+        olen += 2;
+        break;
+    }
+
+    return olen;
+}
+
+size_t numlen(uint32_t num) {
+    size_t len = 1;
+    while (num >= 10) {
+        ++len;
+        num = num / 10;
+    }
+    return len;
+}
+

src/3rdparty/argon2/lib/encoding.h

@@ -0,0 +1,40 @@
+#ifndef ENCODING_H
+#define ENCODING_H
+#include "argon2.h"
+
+#define ARGON2_MAX_DECODED_LANES UINT32_C(255)
+#define ARGON2_MIN_DECODED_SALT_LEN UINT32_C(8)
+#define ARGON2_MIN_DECODED_OUT_LEN UINT32_C(12)
+
+/*
+* encode an Argon2 hash string into the provided buffer. 'dst_len'
+* contains the size, in characters, of the 'dst' buffer; if 'dst_len'
+* is less than the number of required characters (including the
+* terminating 0), then this function returns ARGON2_ENCODING_ERROR.
+*
+* on success, ARGON2_OK is returned.
+*/
+int encode_string(char *dst, size_t dst_len, argon2_context *ctx,
+                  argon2_type type);
+
+/*
+* Decodes an Argon2 hash string into the provided structure 'ctx'.
+* The only fields that must be set prior to this call are ctx.saltlen and
+* ctx.outlen (which must be the maximal salt and out length values that are
+* allowed), ctx.salt and ctx.out (which must be buffers of the specified
+* length), and ctx.pwd and ctx.pwdlen which must hold a valid password.
+*
+* Invalid input string causes an error. On success, the ctx is valid and all
+* fields have been initialized.
+*
+* Returned value is ARGON2_OK on success, other ARGON2_ codes on error.
+*/
+int decode_string(argon2_context *ctx, const char *str, argon2_type type);
+
+/* Returns the length of the encoded byte stream with length len */
+size_t b64len(uint32_t len);
+
+/* Returns the length of the encoded number num */
+size_t numlen(uint32_t num);
+
+#endif

src/3rdparty/argon2/lib/genkat.c

@@ -0,0 +1,117 @@
+/*
+ * Argon2 source code package
+ *
+ * Written by Daniel Dinu and Dmitry Khovratovich, 2015
+ *
+ * This work is licensed under a Creative Commons CC0 1.0 License/Waiver.
+ *
+ * You should have received a copy of the CC0 Public Domain Dedication along
+ * with
+ * this software. If not, see
+ * <http://creativecommons.org/publicdomain/zero/1.0/>.
+ */
+
+#include <inttypes.h>
+#include <stdio.h>
+
+#include "genkat.h"
+
+void initial_kat(const uint8_t *blockhash, const argon2_context *context,
+                 argon2_type type) {
+    unsigned i;
+
+    if (blockhash != NULL && context != NULL) {
+        printf("=======================================\n");
+
+        printf("%s version number %d\n", argon2_type2string(type, 1),
+               context->version);
+
+        printf("=======================================\n");
+
+
+        printf("Memory: %u KiB, Iterations: %u, Parallelism: %u lanes, Tag "
+               "length: %u bytes\n",
+               context->m_cost, context->t_cost, context->lanes,
+               context->outlen);
+
+        printf("Password[%u]: ", context->pwdlen);
+
+        if (context->flags & ARGON2_FLAG_CLEAR_PASSWORD) {
+            printf("CLEARED\n");
+        } else {
+            for (i = 0; i < context->pwdlen; ++i) {
+                printf("%2.2x ", ((unsigned char *)context->pwd)[i]);
+            }
+
+            printf("\n");
+        }
+
+        printf("Salt[%u]: ", context->saltlen);
+
+        for (i = 0; i < context->saltlen; ++i) {
+            printf("%2.2x ", ((unsigned char *)context->salt)[i]);
+        }
+
+        printf("\n");
+
+        printf("Secret[%u]: ", context->secretlen);
+
+        if (context->flags & ARGON2_FLAG_CLEAR_SECRET) {
+            printf("CLEARED\n");
+        } else {
+            for (i = 0; i < context->secretlen; ++i) {
+                printf("%2.2x ", ((unsigned char *)context->secret)[i]);
+            }
+
+            printf("\n");
+        }
+
+        printf("Associated data[%u]: ", context->adlen);
+
+        for (i = 0; i < context->adlen; ++i) {
+            printf("%2.2x ", ((unsigned char *)context->ad)[i]);
+        }
+
+        printf("\n");
+
+        printf("Pre-hashing digest: ");
+
+        for (i = 0; i < ARGON2_PREHASH_DIGEST_LENGTH; ++i) {
+            printf("%2.2x ", ((unsigned char *)blockhash)[i]);
+        }
+
+        printf("\n");
+    }
+}
+
+void print_tag(const void *out, uint32_t outlen) {
+    unsigned i;
+    if (out != NULL) {
+        printf("Tag: ");
+
+        for (i = 0; i < outlen; ++i) {
+            printf("%2.2x ", ((uint8_t *)out)[i]);
+        }
+
+        printf("\n");
+    }
+}
+
+void internal_kat(const argon2_instance_t *instance, uint32_t pass) {
+
+    if (instance != NULL) {
+        uint32_t i, j;
+        printf("\n After pass %u:\n", pass);
+
+        for (i = 0; i < instance->memory_blocks; ++i) {
+            uint32_t how_many_words =
+                (instance->memory_blocks > ARGON2_QWORDS_IN_BLOCK)
+                    ? 1
+                    : ARGON2_QWORDS_IN_BLOCK;
+
+            for (j = 0; j < how_many_words; ++j)
+                printf("Block %.4u [%3u]: %016" PRIx64 "\n", i, j,
+                       instance->memory[i].v[j]);
+        }
+    }
+}

src/3rdparty/argon2/lib/genkat.h

@@ -0,0 +1,47 @@
+/*
+ * Argon2 source code package
+ *
+ * Written by Daniel Dinu and Dmitry Khovratovich, 2015
+ *
+ * This work is licensed under a Creative Commons CC0 1.0 License/Waiver.
+ *
+ * You should have received a copy of the CC0 Public Domain Dedication along
+ * with
+ * this software. If not, see
+ * <http://creativecommons.org/publicdomain/zero/1.0/>.
+ */
+
+#ifndef ARGON2_KAT_H
+#define ARGON2_KAT_H
+
+#include "core.h"
+
+/*
+ * Initial KAT function that prints the inputs to the file
+ * @param  blockhash  Array that contains pre-hashing digest
+ * @param  context Holds inputs
+ * @param  type Argon2 type
+ * @pre blockhash must point to INPUT_INITIAL_HASH_LENGTH bytes
+ * @pre context member pointers must point to allocated memory of size according
+ * to the length values
+ */
+void initial_kat(const uint8_t *blockhash, const argon2_context *context,
+                 argon2_type type);
+
+/*
+ * Function that prints the output tag
+ * @param  out  output array pointer
+ * @param  outlen digest length
+ * @pre out must point to @a outlen bytes
+ **/
+void print_tag(const void *out, uint32_t outlen);
+
+/*
+ * Function that prints the internal state at given moment
+ * @param  instance pointer to the current instance
+ * @param  pass current pass number
+ * @pre instance must have necessary memory allocated
+ **/
+void internal_kat(const argon2_instance_t *instance, uint32_t pass);
+
+#endif

src/3rdparty/argon2/lib/impl-select.c

@@ -0,0 +1,115 @@
+#include <time.h>
+#include <string.h>
+
+#include "impl-select.h"
+
+#include "argon2.h"
+
+#define BENCH_SAMPLES 1024
+#define BENCH_MEM_BLOCKS 512
+
+static argon2_impl selected_argon_impl = {
+    "default", NULL, fill_segment_default
+};
+
+/* the benchmark routine is not thread-safe, so we can use a global var here: */
+static block memory[BENCH_MEM_BLOCKS];
+
+static uint64_t benchmark_impl(const argon2_impl *impl) {
+    clock_t time;
+    unsigned int i;
+    uint64_t bench;
+    argon2_instance_t instance;
+    argon2_position_t pos;
+
+    memset(memory, 0, sizeof(memory));
+
+    instance.version = ARGON2_VERSION_NUMBER;
+    instance.memory = memory;
+    instance.passes = 1;
+    instance.memory_blocks = BENCH_MEM_BLOCKS;
+    instance.segment_length = BENCH_MEM_BLOCKS / ARGON2_SYNC_POINTS;
+    instance.lane_length = instance.segment_length * ARGON2_SYNC_POINTS;
+    instance.lanes = 1;
+    instance.threads = 1;
+    instance.type = Argon2_i;
+
+    pos.lane = 0;
+    pos.pass = 0;
+    pos.slice = 0;
+    pos.index = 0;
+
+    /* warm-up cache: */
+    impl->fill_segment(&instance, pos);
+
+    /* OK, now measure: */
+    bench = 0;
+    time = clock();
+    for (i = 0; i < BENCH_SAMPLES; i++) {
+        impl->fill_segment(&instance, pos);
+    }
+    time = clock() - time;
+    bench = (uint64_t)time;
+    return bench;
+}
+
+void argon2_select_impl()
+{
+    argon2_impl_list impls;
+    unsigned int i;
+    const argon2_impl *best_impl = NULL;
+    uint64_t best_bench = UINT_MAX;
+
+    argon2_get_impl_list(&impls);
+
+    for (i = 0; i < impls.count; i++) {
+        const argon2_impl *impl = &impls.entries[i];
+        uint64_t bench;
+
+        if (impl->check != NULL && !impl->check()) {
+            continue;
+        }
+
+        bench = benchmark_impl(impl);
+
+        if (bench < best_bench) {
+            best_bench = bench;
+            best_impl = impl;
+        }
+    }
+
+    if (best_impl != NULL) {
+        selected_argon_impl = *best_impl;
+    }
+}
+
+void fill_segment(const argon2_instance_t *instance, argon2_position_t position)
+{
+    selected_argon_impl.fill_segment(instance, position);
+}
+
+const char *argon2_get_impl_name()
+{
+    return selected_argon_impl.name;
+}
+
+
+int argon2_select_impl_by_name(const char *name)
+{
+    argon2_impl_list impls;
+    unsigned int i;
+
+    argon2_get_impl_list(&impls);
+
+    for (i = 0; i < impls.count; i++) {
+        const argon2_impl *impl = &impls.entries[i];
+
+        if (strcmp(impl->name, name) == 0) {
+            selected_argon_impl = *impl;
+
+            return 1;
+        }
+    }
+
+    return 0;
+}

src/3rdparty/argon2/lib/impl-select.h

@@ -0,0 +1,23 @@
+#ifndef ARGON2_IMPL_SELECT_H
+#define ARGON2_IMPL_SELECT_H
+
+#include "core.h"
+
+typedef struct Argon2_impl {
+    const char *name;
+    int (*check)(void);
+    void (*fill_segment)(const argon2_instance_t *instance,
+                         argon2_position_t position);
+} argon2_impl;
+
+typedef struct Argon2_impl_list {
+    const argon2_impl *entries;
+    size_t count;
+} argon2_impl_list;
+
+void argon2_get_impl_list(argon2_impl_list *list);
+void fill_segment_default(const argon2_instance_t *instance,
+                          argon2_position_t position);
+
+#endif // ARGON2_IMPL_SELECT_H
+

src/App.cpp

@@ -28,7 +28,6 @@
 #include <uv.h>
 
 
-#include "api/Api.h"
 #include "App.h"
 #include "backend/cpu/Cpu.h"
 #include "base/io/Console.h"

src/Summary.cpp

@@ -71,13 +71,12 @@ static void print_cpu(Config *)
 {
     const ICpuInfo *info = Cpu::info();
 
-    Log::print(GREEN_BOLD(" * ") WHITE_BOLD("%-13s%s (%zu)") " %sx64 %sAES %sAVX2",
+    Log::print(GREEN_BOLD(" * ") WHITE_BOLD("%-13s%s (%zu)") " %sx64 %sAES",
                "CPU",
                info->brand(),
                info->packages(),
                info->isX64()   ? GREEN_BOLD_S : RED_BOLD_S "-",
-               info->hasAES()  ? GREEN_BOLD_S : RED_BOLD_S "-",
-               info->hasAVX2() ? GREEN_BOLD_S : RED_BOLD_S "-"
+               info->hasAES()  ? GREEN_BOLD_S : RED_BOLD_S "-"
                );
 #   if defined(XMRIG_FEATURE_LIBCPUID) || defined (XMRIG_FEATURE_HWLOC)
     Log::print(WHITE_BOLD("   %-13s") BLACK_BOLD("L2:") WHITE_BOLD("%.1f MB") BLACK_BOLD(" L3:") WHITE_BOLD("%.1f MB")

src/api/requests/HttpApiRequest.cpp

@@ -1,87 +0,0 @@
-/* XMRig
- * Copyright 2010      Jeff Garzik <jgarzik@pobox.com>
- * Copyright 2012-2014 pooler      <pooler@litecoinpool.org>
- * Copyright 2014      Lucas Jones <https://github.com/lucasjones>
- * Copyright 2014-2016 Wolf9466    <https://github.com/OhGodAPet>
- * Copyright 2016      Jay D Dee   <jayddee246@gmail.com>
- * Copyright 2017-2018 XMR-Stak    <https://github.com/fireice-uk>, <https://github.com/psychocrypt>
- * Copyright 2018-2019 SChernykh   <https://github.com/SChernykh>
- * Copyright 2016-2019 XMRig       <https://github.com/xmrig>, <support@xmrig.com>
- *
- *   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 <http://www.gnu.org/licenses/>.
- */
-
-
-#include "api/requests/HttpApiRequest.h"
-#include "base/net/http/HttpData.h"
-#include "rapidjson/error/en.h"
-
-
-xmrig::HttpApiRequest::HttpApiRequest(const HttpData &req, bool restricted) :
-    ApiRequest(SOURCE_HTTP, restricted),
-    m_parsed(false),
-    m_req(req),
-    m_res(req.id()),
-    m_url(req.url.c_str())
-{
-    if (method() == METHOD_GET) {
-        if (url() == "/1/summary" || url() == "/2/summary" || url() == "/api.json") {
-            m_type = REQ_SUMMARY;
-        }
-    }
-
-    if (url().size() > 4) {
-        if (memcmp(url().data(), "/2/", 3) == 0) {
-            m_version = 2;
-        }
-    }
-}
-
-
-const rapidjson::Value &xmrig::HttpApiRequest::json() const
-{
-    return m_body;
-}
-
-
-xmrig::IApiRequest::Method xmrig::HttpApiRequest::method() const
-{
-    return static_cast<IApiRequest::Method>(m_req.method);
-}
-
-
-void xmrig::HttpApiRequest::accept()
-{
-    using namespace rapidjson;
-
-    ApiRequest::accept();
-
-    if (!m_parsed && !m_req.body.empty()) {
-        m_parsed = true;
-        m_body.Parse<kParseCommentsFlag | kParseTrailingCommasFlag>(m_req.body.c_str());
-
-        if (m_body.HasParseError()) {
-            reply().AddMember("error", StringRef(GetParseError_En(m_body.GetParseError())), doc().GetAllocator());;
-        }
-    }
-}
-
-
-void xmrig::HttpApiRequest::done(int status)
-{
-    ApiRequest::done(status);
-
-    m_res.setStatus(status);
-    m_res.end();
-}

src/backend/common/Threads.cpp

@@ -61,7 +61,11 @@ size_t xmrig::Threads<T>::read(const rapidjson::Value &value)
             if (!threads.isEmpty()) {
                 move(member.name.GetString(), std::move(threads));
             }
+        }
+    }
 
+    for (auto &member : value.GetObject()) {
+        if (member.value.IsArray() || member.value.IsObject()) {
             continue;
         }
 

src/backend/common/WorkerJob.h

@@ -56,6 +56,7 @@ public:
         }
 
         if (index() == 1 && job.index() == 0 && job == m_jobs[0]) {
+            m_index = 0;
             return;
         }
 

src/backend/common/Workers.cpp

@@ -166,18 +166,7 @@ namespace xmrig {
 template<>
 xmrig::IWorker *xmrig::Workers<CpuLaunchData>::create(Thread<CpuLaunchData> *handle)
 {
-    const int intensity = handle->config().intensity;
-
-#   if defined(XMRIG_ALGO_RANDOMX) || defined(XMRIG_ALGO_CN_GPU)
-    if (intensity > handle->config().algorithm.maxIntensity()) {
-        LOG_WARN("CPU thread %zu warning: \"intensity %d not supported for %s algorithm\".", handle->index(), handle->config().intensity, handle->config().algorithm.shortName());
-
-        return new CpuWorker<1>(handle->index(), handle->config());
-    }
-#   endif
-
-
-    switch (intensity) {
+    switch (handle->config().intensity) {
     case 1:
         return new CpuWorker<1>(handle->index(), handle->config());
 

src/backend/cpu/CpuBackend.cpp

@@ -43,6 +43,11 @@
 #include "rapidjson/document.h"
 
 
+#ifdef XMRIG_ALGO_ARGON2
+#   include "crypto/argon2/Impl.h"
+#endif
+
+
 namespace xmrig {
 
 
@@ -143,9 +148,9 @@ xmrig::CpuBackend::~CpuBackend()
 }
 
 
-std::pair<size_t, size_t> xmrig::CpuBackend::hugePages() const
+std::pair<unsigned, unsigned> xmrig::CpuBackend::hugePages() const
 {
-    std::pair<size_t, size_t> pages(0, 0);
+    std::pair<unsigned, unsigned> pages(0, 0);
 
 #   ifdef XMRIG_ALGO_RANDOMX
     if (d_ptr->algo.family() == Algorithm::RANDOM_X) {
@@ -192,8 +197,17 @@ const xmrig::String &xmrig::CpuBackend::type() const
 }
 
 
-void xmrig::CpuBackend::prepare(const Job &)
+void xmrig::CpuBackend::prepare(const Job &nextJob)
 {
+#   ifdef XMRIG_ALGO_ARGON2
+    if (nextJob.algorithm().family() == Algorithm::ARGON2 && argon2::Impl::select(d_ptr->controller->config()->cpu().argon2Impl())) {
+        LOG_INFO("%s use " WHITE_BOLD("argon2") " implementation " CSI "1;%dm" "%s",
+                 tag,
+                 argon2::Impl::name() == "default" ? 33 : 32,
+                 argon2::Impl::name().data()
+                 );
+    }
+#   endif
 }
 
 
@@ -319,6 +333,10 @@ rapidjson::Value xmrig::CpuBackend::toJSON(rapidjson::Document &doc) const
     out.AddMember("asm", false, allocator);
 #   endif
 
+#   ifdef XMRIG_ALGO_ARGON2
+    out.AddMember("argon2-impl", argon2::Impl::name().toJSON(), allocator);
+#   endif
+
     const auto pages = hugePages();
 
     rapidjson::Value hugepages(rapidjson::kArrayType);

src/backend/cpu/CpuBackend.h

@@ -46,7 +46,7 @@ public:
     CpuBackend(Controller *controller);
     ~CpuBackend() override;
 
-    std::pair<size_t, size_t> hugePages() const;
+    std::pair<unsigned, unsigned> hugePages() const;
 
 protected:
     bool isEnabled() const override;

src/backend/cpu/CpuConfig.cpp

@@ -62,6 +62,11 @@ static const char *kRx    = "rx";
 static const char *kRxWOW = "rx/wow";
 #endif
 
+#ifdef XMRIG_ALGO_ARGON2
+static const char *kArgon2     = "argon2";
+static const char *kArgon2Impl = "argon2-impl";
+#endif
+
 extern template class Threads<CpuThreads>;
 
 }
@@ -94,6 +99,10 @@ rapidjson::Value xmrig::CpuConfig::toJSON(rapidjson::Document &doc) const
     obj.AddMember(StringRef(kAsm), m_assembly.toJSON(), allocator);
 #   endif
 
+#   ifdef XMRIG_ALGO_ARGON2
+    obj.AddMember(StringRef(kArgon2Impl), m_argon2Impl.toJSON(), allocator);
+#   endif
+
     m_threads.toJSON(obj, doc);
 
     return obj;
@@ -119,7 +128,7 @@ std::vector<xmrig::CpuLaunchData> xmrig::CpuConfig::get(const Miner *miner, cons
 }
 
 
-void xmrig::CpuConfig::read(const rapidjson::Value &value)
+void xmrig::CpuConfig::read(const rapidjson::Value &value, uint32_t version)
 {
     if (value.IsObject()) {
         m_enabled       = Json::getBool(value, kEnabled, m_enabled);
@@ -132,9 +141,17 @@ void xmrig::CpuConfig::read(const rapidjson::Value &value)
         m_assembly = Json::getValue(value, kAsm);
 #       endif
 
+#       ifdef XMRIG_ALGO_ARGON2
+        m_argon2Impl = Json::getString(value, kArgon2Impl);
+#       endif
+
         if (!m_threads.read(value)) {
             generate();
         }
+
+        if (version == 0) {
+            generateArgon2();
+        }
     }
     else if (value.IsBool() && value.IsFalse()) {
         m_enabled = false;
@@ -174,6 +191,16 @@ void xmrig::CpuConfig::generate()
     m_threads.move(kRx, cpu->threads(Algorithm::RX_0));
     m_threads.move(kRxWOW, cpu->threads(Algorithm::RX_WOW));
 #   endif
+
+    generateArgon2();
+}
+
+
+void xmrig::CpuConfig::generateArgon2()
+{
+#   ifdef XMRIG_ALGO_ARGON2
+    m_threads.move(kArgon2, Cpu::info()->threads(Algorithm::AR2_CHUKWA));
+#   endif
 }
 
 

src/backend/cpu/CpuConfig.h

@@ -49,17 +49,19 @@ public:
     bool isHwAES() const;
     rapidjson::Value toJSON(rapidjson::Document &doc) const;
     std::vector<CpuLaunchData> get(const Miner *miner, const Algorithm &algorithm) const;
-    void read(const rapidjson::Value &value);
+    void read(const rapidjson::Value &value, uint32_t version);
 
     inline bool isEnabled() const                       { return m_enabled; }
     inline bool isHugePages() const                     { return m_hugePages; }
     inline bool isShouldSave() const                    { return m_shouldSave; }
     inline const Assembly &assembly() const             { return m_assembly; }
+    inline const String &argon2Impl() const             { return m_argon2Impl; }
     inline const Threads<CpuThreads> &threads() const   { return m_threads; }
     inline int priority() const                         { return m_priority; }
 
 private:
     void generate();
+    void generateArgon2();
     void setAesMode(const rapidjson::Value &aesMode);
 
     inline void setPriority(int priority)   { m_priority = (priority >= -1 && priority <= 5) ? priority : -1; }
@@ -70,6 +72,7 @@ private:
     bool m_hugePages     = true;
     bool m_shouldSave    = false;
     int m_priority       = -1;
+    String m_argon2Impl;
     Threads<CpuThreads> m_threads;
 };
 

src/backend/cpu/CpuLaunchData.cpp

@@ -24,6 +24,9 @@
  */
 
 
+#include <algorithm>
+
+
 #include "backend/cpu/CpuLaunchData.h"
 #include "backend/cpu/CpuConfig.h"
 
@@ -33,10 +36,10 @@ xmrig::CpuLaunchData::CpuLaunchData(const Miner *miner, const Algorithm &algorit
     assembly(config.assembly()),
     hugePages(config.isHugePages()),
     hwAES(config.isHwAES()),
-    intensity(thread.intensity()),
     priority(config.priority()),
     affinity(thread.affinity()),
-    miner(miner)
+    miner(miner),
+    intensity(std::min<uint32_t>(thread.intensity(), algorithm.maxIntensity()))
 {
 }
 

src/backend/cpu/CpuLaunchData.h

@@ -58,10 +58,10 @@ public:
     const Assembly assembly;
     const bool hugePages;
     const bool hwAES;
-    const int intensity;
     const int priority;
     const int64_t affinity;
     const Miner *miner;
+    const uint32_t intensity;
 };
 
 

src/backend/cpu/CpuThread.cpp

@@ -31,11 +31,11 @@
 xmrig::CpuThread::CpuThread(const rapidjson::Value &value)
 {
     if (value.IsArray() && value.Size() >= 2) {
-        m_intensity = value[0].GetInt();
+        m_intensity = value[0].GetUint();
         m_affinity  = value[1].GetInt();
     }
     else if (value.IsInt()) {
-        m_intensity = -1;
+        m_intensity = 0;
         m_affinity  = value.GetInt();
     }
 }
@@ -44,7 +44,7 @@ xmrig::CpuThread::CpuThread(const rapidjson::Value &value)
 rapidjson::Value xmrig::CpuThread::toJSON(rapidjson::Document &doc) const
 {
     using namespace rapidjson;
-    if (m_intensity == -1) {
+    if (m_intensity == 0) {
         return Value(m_affinity);
     }
 

src/backend/cpu/CpuThread.h

@@ -36,14 +36,14 @@ class CpuThread
 {
 public:
     inline constexpr CpuThread() {}
-    inline constexpr CpuThread(int64_t affinity, int intensity) : m_intensity(intensity), m_affinity(affinity) {}
+    inline constexpr CpuThread(int64_t affinity, uint32_t intensity) : m_affinity(affinity), m_intensity(intensity) {}
 
     CpuThread(const rapidjson::Value &value);
 
     inline bool isEqual(const CpuThread &other) const       { return other.m_affinity == m_affinity && other.m_intensity == m_intensity; }
-    inline bool isValid() const                             { return m_intensity == -1 || (m_intensity >= 1 && m_intensity <= 5); }
-    inline int intensity() const                            { return m_intensity == -1 ? 1 : m_intensity; }
+    inline bool isValid() const                             { return m_intensity <= 5; }
     inline int64_t affinity() const                         { return m_affinity; }
+    inline uint32_t intensity() const                       { return m_intensity == 0 ? 1 : m_intensity; }
 
     inline bool operator!=(const CpuThread &other) const    { return !isEqual(other); }
     inline bool operator==(const CpuThread &other) const    { return isEqual(other); }
@@ -51,8 +51,8 @@ public:
     rapidjson::Value toJSON(rapidjson::Document &doc) const;
 
 private:
-    int m_intensity     = -1;
-    int64_t m_affinity  = -1;
+    int64_t m_affinity   = -1;
+    uint32_t m_intensity = 0;
 };
 
 

src/backend/cpu/CpuThreads.cpp

@@ -94,7 +94,7 @@ xmrig::CpuThreads::CpuThreads(const rapidjson::Value &value)
         }
     }
     else if (value.IsObject()) {
-        int intensity        = Json::getInt(value, kIntensity, 1);
+        uint32_t intensity   = Json::getUint(value, kIntensity, 1);
         const size_t threads = std::min<unsigned>(Json::getUint(value, kThreads), 1024);
         m_affinity           = getAffinityMask(Json::getValue(value, kAffinity));
         m_format             = ObjectFormat;
@@ -110,7 +110,7 @@ xmrig::CpuThreads::CpuThreads(const rapidjson::Value &value)
 }
 
 
-xmrig::CpuThreads::CpuThreads(size_t count, int intensity)
+xmrig::CpuThreads::CpuThreads(size_t count, uint32_t intensity)
 {
     m_data.reserve(count);
 

src/backend/cpu/CpuThreads.h

@@ -42,13 +42,13 @@ public:
     inline CpuThreads(size_t count) : m_data(count) {}
 
     CpuThreads(const rapidjson::Value &value);
-    CpuThreads(size_t count, int intensity);
+    CpuThreads(size_t count, uint32_t intensity);
 
     inline bool isEmpty() const                             { return m_data.empty(); }
     inline const std::vector<CpuThread> &data() const       { return m_data; }
     inline size_t count() const                             { return m_data.size(); }
     inline void add(CpuThread &&thread)                     { m_data.push_back(thread); }
-    inline void add(int64_t affinity, int intensity)        { add(CpuThread(affinity, intensity)); }
+    inline void add(int64_t affinity, uint32_t intensity)   { add(CpuThread(affinity, intensity)); }
     inline void reserve(size_t capacity)                    { m_data.reserve(capacity); }
 
     rapidjson::Value toJSON(rapidjson::Document &doc) const;

src/backend/cpu/CpuWorker.cpp

@@ -90,6 +90,8 @@ void xmrig::CpuWorker<N>::allocateRandomX_VM()
         if (Nonce::sequence(Nonce::CPU) == 0) {
             return;
         }
+
+        dataset = Rx::dataset(m_job.currentJob(), m_node);
     }
 
     if (!m_vm) {
@@ -156,6 +158,13 @@ bool xmrig::CpuWorker<N>::selfTest()
     }
 #   endif
 
+#   ifdef XMRIG_ALGO_ARGON2
+    if (m_algorithm.family() == Algorithm::ARGON2) {
+        return verify(Algorithm::AR2_CHUKWA, argon2_chukwa_test_out) &&
+               verify(Algorithm::AR2_WRKZ, argon2_wrkz_test_out);
+    }
+#   endif
+
     return false;
 }
 

src/backend/cpu/platform/AdvancedCpuInfo.cpp

@@ -145,7 +145,7 @@ xmrig::CpuThreads xmrig::AdvancedCpuInfo::threads(const Algorithm &algorithm) co
         count = threads() / 2;
     }
 
-    int intensity = algorithm.maxIntensity() == 1 ? -1 : 1;
+    uint32_t intensity = algorithm.maxIntensity() == 1 ? 0 : 1;
 
 #   ifdef XMRIG_ALGO_CN_PICO
     if (algorithm == Algorithm::CN_PICO_0 && (count / cores()) >= 2) {

src/backend/cpu/platform/BasicCpuInfo.cpp

@@ -52,7 +52,6 @@
 
 #define VENDOR_ID                  (0)
 #define PROCESSOR_INFO             (1)
-#define CACHE_TLB_DESCRIPTOR       (2)
 #define EXTENDED_FEATURES          (7)
 #define PROCESSOR_BRAND_STRING_1   (0x80000002)
 #define PROCESSOR_BRAND_STRING_2   (0x80000003)
@@ -222,5 +221,11 @@ xmrig::CpuThreads xmrig::BasicCpuInfo::threads(const Algorithm &algorithm) const
     }
 #   endif
 
+#   ifdef XMRIG_ALGO_ARGON2
+    if (algorithm.family() == Algorithm::ARGON2) {
+        return count;
+    }
+#   endif
+
     return CpuThreads(std::max<size_t>(count / 2, 1), 1);
 }

src/backend/cpu/platform/HwlocCpuInfo.cpp

@@ -246,13 +246,14 @@ void xmrig::HwlocCpuInfo::processTopLevelCache(hwloc_obj_t cache, const Algorith
     findByType(cache, HWLOC_OBJ_CORE, [&cores](hwloc_obj_t found) { cores.emplace_back(found); });
 
     size_t L3               = cache->attr->cache.size;
+    const bool L3_exclusive = isCacheExclusive(cache);
     size_t L2               = 0;
     int L2_associativity    = 0;
     size_t extra            = 0;
     const size_t scratchpad = algorithm.l3();
-    int intensity           = algorithm.maxIntensity() == 1 ? -1 : 1;
+    uint32_t intensity      = algorithm.maxIntensity() == 1 ? 0 : 1;
 
-    if (cache->attr->cache.depth == 3 && isCacheExclusive(cache)) {
+    if (cache->attr->cache.depth == 3) {
         for (size_t i = 0; i < cache->arity; ++i) {
             hwloc_obj_t l2 = cache->children[i];
             if (!isCacheObject(l2) || l2->attr == nullptr) {
@@ -262,7 +263,7 @@ void xmrig::HwlocCpuInfo::processTopLevelCache(hwloc_obj_t cache, const Algorith
             L2 += l2->attr->cache.size;
             L2_associativity = l2->attr->cache.associativity;
 
-            if (l2->attr->cache.size >= scratchpad) {
+            if (L3_exclusive && l2->attr->cache.size >= scratchpad) {
                 extra += scratchpad;
             }
         }

src/base/api/Api.cpp

@@ -32,9 +32,9 @@
 
 
 #include "3rdparty/http-parser/http_parser.h"
-#include "api/Api.h"
-#include "api/interfaces/IApiListener.h"
-#include "api/requests/HttpApiRequest.h"
+#include "base/api/Api.h"
+#include "base/api/interfaces/IApiListener.h"
+#include "base/api/requests/HttpApiRequest.h"
 #include "base/kernel/Base.h"
 #include "base/tools/Buffer.h"
 #include "base/tools/Chrono.h"
@@ -45,7 +45,7 @@
 
 
 #ifdef XMRIG_FEATURE_HTTP
-#   include "api/Httpd.h"
+#   include "base/api/Httpd.h"
 #endif
 
 

src/base/api/Httpd.cpp

@@ -24,8 +24,8 @@
 
 
 #include "3rdparty/http-parser/http_parser.h"
-#include "api/Api.h"
-#include "api/Httpd.h"
+#include "base/api/Api.h"
+#include "base/api/Httpd.h"
 #include "base/io/log/Log.h"
 #include "base/net/http/HttpApiResponse.h"
 #include "base/net/http/HttpData.h"

src/base/api/interfaces/IApiRequest.h

@@ -54,16 +54,28 @@ public:
 
     enum RequestType {
         REQ_UNKNOWN,
-        REQ_SUMMARY
+        REQ_SUMMARY,
+        REQ_JSON_RPC
+    };
+
+
+    enum ErrorCode : int {
+        RPC_PARSE_ERROR      = -32700,
+        RPC_INVALID_REQUEST  = -32600,
+        RPC_METHOD_NOT_FOUND = -32601,
+        RPC_INVALID_PARAMS   = -32602
     };
 
 
     virtual ~IApiRequest() = default;
 
+    virtual bool accept()                                        = 0;
+    virtual bool hasParseError() const                           = 0;
     virtual bool isDone() const                                  = 0;
     virtual bool isNew() const                                   = 0;
     virtual bool isRestricted() const                            = 0;
     virtual const rapidjson::Value &json() const                 = 0;
+    virtual const String &rpcMethod() const                      = 0;
     virtual const String &url() const                            = 0;
     virtual int version() const                                  = 0;
     virtual Method method() const                                = 0;
@@ -71,7 +83,6 @@ public:
     virtual rapidjson::Value &reply()                            = 0;
     virtual RequestType type() const                             = 0;
     virtual Source source() const                                = 0;
-    virtual void accept()                                        = 0;
     virtual void done(int status)                                = 0;
 };
 

src/base/api/requests/ApiRequest.cpp

@@ -23,7 +23,7 @@
  */
 
 
-#include "api/requests/ApiRequest.h"
+#include "base/api/requests/ApiRequest.h"
 
 
 xmrig::ApiRequest::ApiRequest(Source source, bool restricted) :

src/base/api/requests/ApiRequest.h

@@ -27,7 +27,8 @@
 #define XMRIG_APIREQUEST_H
 
 
-#include "api/interfaces/IApiRequest.h"
+#include "base/api/interfaces/IApiRequest.h"
+#include "base/tools/String.h"
 
 
 namespace xmrig {
@@ -40,28 +41,30 @@ public:
     ~ApiRequest() override;
 
 protected:
-    inline bool isDone() const override          { return m_state == STATE_DONE; }
-    inline bool isNew() const override           { return m_state == STATE_NEW; }
-    inline bool isRestricted() const override    { return m_restricted; }
-    inline int version() const override          { return m_version; }
-    inline RequestType type() const override     { return m_type; }
-    inline Source source() const override        { return m_source; }
-    inline void accept() override                { m_state = STATE_ACCEPTED; }
-    inline void done(int) override               { m_state = STATE_DONE; }
-
-    int m_version      = 1;
-    RequestType m_type = REQ_UNKNOWN;
-
-private:
     enum State {
         STATE_NEW,
         STATE_ACCEPTED,
         STATE_DONE
     };
 
+    inline bool accept() override                   { m_state = STATE_ACCEPTED; return true; }
+    inline bool isDone() const override             { return m_state == STATE_DONE; }
+    inline bool isNew() const override              { return m_state == STATE_NEW; }
+    inline bool isRestricted() const override       { return m_restricted; }
+    inline const String &rpcMethod() const override { return m_rpcMethod; }
+    inline int version() const override             { return m_version; }
+    inline RequestType type() const override        { return m_type; }
+    inline Source source() const override           { return m_source; }
+    inline void done(int) override                  { m_state = STATE_DONE; }
+
+    int m_version       = 1;
+    RequestType m_type  = REQ_UNKNOWN;
+    State m_state       = STATE_NEW;
+    String m_rpcMethod;
+
+private:
     bool m_restricted;
     Source m_source;
-    State m_state = STATE_NEW;
 };
 
 

src/base/api/requests/HttpApiRequest.cpp

@@ -0,0 +1,178 @@
+/* XMRig
+ * Copyright 2010      Jeff Garzik <jgarzik@pobox.com>
+ * Copyright 2012-2014 pooler      <pooler@litecoinpool.org>
+ * Copyright 2014      Lucas Jones <https://github.com/lucasjones>
+ * Copyright 2014-2016 Wolf9466    <https://github.com/OhGodAPet>
+ * Copyright 2016      Jay D Dee   <jayddee246@gmail.com>
+ * Copyright 2017-2018 XMR-Stak    <https://github.com/fireice-uk>, <https://github.com/psychocrypt>
+ * Copyright 2018-2019 SChernykh   <https://github.com/SChernykh>
+ * Copyright 2016-2019 XMRig       <https://github.com/xmrig>, <support@xmrig.com>
+ *
+ *   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 <http://www.gnu.org/licenses/>.
+ */
+
+
+#include "3rdparty/http-parser/http_parser.h"
+#include "base/api/requests/HttpApiRequest.h"
+#include "base/io/json/Json.h"
+#include "base/net/http/HttpData.h"
+#include "rapidjson/error/en.h"
+
+
+namespace xmrig {
+
+
+static const char *kError  = "error";
+static const char *kId     = "id";
+static const char *kResult = "result";
+
+
+static inline const char *rpcError(int code) {
+    switch (code) {
+    case IApiRequest::RPC_PARSE_ERROR:
+        return "Parse error";
+
+    case IApiRequest::RPC_INVALID_REQUEST:
+        return "Invalid Request";
+
+    case IApiRequest::RPC_METHOD_NOT_FOUND:
+        return "Method not found";
+
+    case IApiRequest::RPC_INVALID_PARAMS:
+        return "Invalid params";
+    }
+
+    return "Internal error";
+}
+
+
+} // namespace xmrig
+
+
+xmrig::HttpApiRequest::HttpApiRequest(const HttpData &req, bool restricted) :
+    ApiRequest(SOURCE_HTTP, restricted),
+    m_req(req),
+    m_res(req.id()),
+    m_url(req.url.c_str())
+{
+    if (method() == METHOD_GET) {
+        if (url() == "/1/summary" || url() == "/2/summary" || url() == "/api.json") {
+            m_type = REQ_SUMMARY;
+        }
+    }
+
+    if (method() == METHOD_POST && url() == "/json_rpc") {
+        m_type = REQ_JSON_RPC;
+        accept();
+
+        if (hasParseError()) {
+            done(RPC_PARSE_ERROR);
+
+            return;
+        }
+
+        m_rpcMethod = Json::getString(json(), "method");
+        if (m_rpcMethod.isEmpty()) {
+            done(RPC_INVALID_REQUEST);
+
+            return;
+        }
+
+        m_state = STATE_NEW;
+
+        return;
+    }
+
+    if (url().size() > 4) {
+        if (memcmp(url().data(), "/2/", 3) == 0) {
+            m_version = 2;
+        }
+    }
+}
+
+
+bool xmrig::HttpApiRequest::accept()
+{
+    using namespace rapidjson;
+
+    ApiRequest::accept();
+
+    if (m_parsed == 0 && !m_req.body.empty()) {
+        m_body.Parse<kParseCommentsFlag | kParseTrailingCommasFlag>(m_req.body.c_str());
+        m_parsed = m_body.HasParseError() ? 2 : 1;
+
+        if (!hasParseError()) {
+            return true;
+        }
+
+        if (type() != REQ_JSON_RPC) {
+            reply().AddMember(StringRef(kError), StringRef(GetParseError_En(m_body.GetParseError())), doc().GetAllocator());
+        }
+
+        return false;
+    }
+
+    return hasParseError();
+}
+
+
+const rapidjson::Value &xmrig::HttpApiRequest::json() const
+{
+    return m_body;
+}
+
+
+xmrig::IApiRequest::Method xmrig::HttpApiRequest::method() const
+{
+    return static_cast<IApiRequest::Method>(m_req.method);
+}
+
+
+void xmrig::HttpApiRequest::done(int status)
+{
+    ApiRequest::done(status);
+
+    if (type() == REQ_JSON_RPC) {
+        using namespace rapidjson;
+        auto &allocator = doc().GetAllocator();
+
+        m_res.setStatus(HTTP_STATUS_OK);
+
+        if (status != HTTP_STATUS_OK) {
+            if (status == HTTP_STATUS_NOT_FOUND) {
+                status = RPC_METHOD_NOT_FOUND;
+            }
+
+            Value error(kObjectType);
+            error.AddMember("code",    status, allocator);
+            error.AddMember("message", StringRef(rpcError(status)), allocator);
+
+            reply().AddMember(StringRef(kError), error, allocator);
+        }
+        else if (!reply().HasMember(kResult)) {
+            Value result(kObjectType);
+            result.AddMember("status", "OK", allocator);
+
+            reply().AddMember(StringRef(kResult), result, allocator);
+        }
+
+        reply().AddMember("jsonrpc", "2.0", allocator);
+        reply().AddMember(StringRef(kId), Value().CopyFrom(Json::getValue(json(), kId), allocator), allocator);
+    }
+    else {
+        m_res.setStatus(status);
+    }
+
+    m_res.end();
+}

src/base/api/requests/HttpApiRequest.h

@@ -27,7 +27,7 @@
 #define XMRIG_HTTPAPIREQUEST_H
 
 
-#include "api/requests/ApiRequest.h"
+#include "base/api/requests/ApiRequest.h"
 #include "base/net/http/HttpApiResponse.h"
 #include "base/tools/String.h"
 
@@ -44,19 +44,20 @@ public:
     HttpApiRequest(const HttpData &req, bool restricted);
 
 protected:
+    inline bool hasParseError() const override           { return m_parsed == 2; }
+    inline const String &url() const override            { return m_url; }
     inline rapidjson::Document &doc() override           { return m_res.doc(); }
     inline rapidjson::Value &reply() override            { return m_res.doc(); }
-    inline const String &url() const override            { return m_url; }
 
+    bool accept() override;
     const rapidjson::Value &json() const override;
     Method method() const override;
-    void accept() override;
     void done(int status) override;
 
 private:
-    bool m_parsed;
     const HttpData &m_req;
     HttpApiResponse m_res;
+    int m_parsed = 0;
     rapidjson::Document m_body;
     String m_url;
 };

src/base/base.cmake

@@ -1,4 +1,5 @@
 set(HEADERS_BASE
+    src/base/api/interfaces/IApiListener.h
     src/base/io/Console.h
     src/base/io/json/Json.h
     src/base/io/json/JsonChain.h
@@ -114,6 +115,11 @@ endif()
 if (WITH_HTTP)
     set(HEADERS_BASE_HTTP
         src/3rdparty/http-parser/http_parser.h
+        src/base/api/Api.h
+        src/base/api/Httpd.h
+        src/base/api/interfaces/IApiRequest.h
+        src/base/api/requests/ApiRequest.h
+        src/base/api/requests/HttpApiRequest.h
         src/base/kernel/interfaces/IHttpListener.h
         src/base/kernel/interfaces/IJsonReader.h
         src/base/kernel/interfaces/ITcpServerListener.h
@@ -129,6 +135,10 @@ if (WITH_HTTP)
 
     set(SOURCES_BASE_HTTP
         src/3rdparty/http-parser/http_parser.c
+        src/base/api/Api.cpp
+        src/base/api/Httpd.cpp
+        src/base/api/requests/ApiRequest.cpp
+        src/base/api/requests/HttpApiRequest.cpp
         src/base/net/http/HttpApiResponse.cpp
         src/base/net/http/HttpClient.cpp
         src/base/net/http/HttpContext.cpp

src/base/io/json/Json_win.cpp

@@ -28,6 +28,7 @@
 
 #ifdef __GNUC__
 #   include <fcntl.h>
+#   include <sys/stat.h>
 #   include <ext/stdio_filebuf.h>
 #endif
 
@@ -102,7 +103,7 @@ bool xmrig::Json::save(const char *fileName, const rapidjson::Document &doc)
         return false;
     }
 #   elif defined(__GNUC__)
-    const int fd = _wopen(toUtf16(fileName).c_str(), _O_WRONLY | _O_BINARY | _O_CREAT | _O_TRUNC);
+    const int fd = _wopen(toUtf16(fileName).c_str(), _O_WRONLY | _O_BINARY | _O_CREAT | _O_TRUNC, _S_IWRITE);
     if (fd == -1) {
         return false;
     }

src/base/io/log/Log.cpp

@@ -91,6 +91,10 @@ public:
 
         std::lock_guard<std::mutex> lock(m_mutex);
 
+        if (Log::background && m_backends.empty()) {
+            return;
+        }
+
         timestamp(level, size, offset);
         color(level, size);
 
@@ -111,7 +115,7 @@ public:
         if (!m_backends.empty()) {
             for (ILogBackend *backend : m_backends) {
                 backend->print(level, m_buf, offset, size, true);
-                backend->print(level, txt.c_str(), offset, txt.size(), false);
+                backend->print(level, txt.c_str(), offset ? (offset - 11) : 0, txt.size(), false);
             }
         }
         else {
@@ -190,8 +194,9 @@ private:
 };
 
 
-bool Log::colors   = true;
-LogPrivate *Log::d = new LogPrivate();
+bool Log::background = false;
+bool Log::colors     = true;
+LogPrivate *Log::d   = new LogPrivate();
 
 
 } /* namespace xmrig */

src/base/io/log/Log.h

@@ -54,6 +54,7 @@ public:
     static void print(const char *fmt, ...);
     static void print(Level level, const char *fmt, ...);
 
+    static bool background;
     static bool colors;
 
 private:

src/base/io/log/backends/SysLog.h

@@ -37,7 +37,7 @@ class SysLog : public ILogBackend
 {
 public:
     SysLog();
-    ~SysLog();
+    ~SysLog() override;
 
 protected:
     void print(int level, const char *line, size_t offset, size_t size, bool colors) override;

src/base/kernel/Base.cpp

@@ -47,15 +47,8 @@
 
 
 #ifdef XMRIG_FEATURE_API
-#   include "api/Api.h"
-#   include "api/interfaces/IApiRequest.h"
-#endif
-
-
-#ifdef XMRIG_FEATURE_EMBEDDED_CONFIG
-#   include "core/config/Config_default.h"
-#endif
-
+#   include "base/api/Api.h"
+#   include "base/api/interfaces/IApiRequest.h"
 
 namespace xmrig {
 
@@ -63,6 +56,12 @@ static const char *kConfigPathV1 = "/1/config";
 static const char *kConfigPathV2 = "/2/config";
 
 } // namespace xmrig
+#endif
+
+
+#ifdef XMRIG_FEATURE_EMBEDDED_CONFIG
+#   include "core/config/Config_default.h"
+#endif
 
 
 class xmrig::BasePrivate
@@ -185,7 +184,10 @@ int xmrig::Base::init()
     Platform::setProcessPriority(config()->cpu().priority());
 #   endif
 
-    if (!config()->isBackground()) {
+    if (config()->isBackground()) {
+        Log::background = true;
+    }
+    else {
         Log::add(new ConsoleLog());
     }
 

src/base/kernel/Base.h

@@ -26,7 +26,7 @@
 #define XMRIG_BASE_H
 
 
-#include "api/interfaces/IApiListener.h"
+#include "base/api/interfaces/IApiListener.h"
 #include "base/kernel/interfaces/IConfigListener.h"
 #include "base/kernel/interfaces/IWatcherListener.h"
 #include "rapidjson/fwd.h"

src/base/kernel/config/BaseConfig.cpp

@@ -134,6 +134,7 @@ bool xmrig::BaseConfig::read(const IJsonReader &reader, const char *fileName)
     Log::colors    = reader.getBool("colors", Log::colors);
     m_logFile      = reader.getString("log-file");
     m_userAgent    = reader.getString("user-agent");
+    m_version      = reader.getUint("version");
 
     setPrintTime(reader.getUint("print-time", 60));
 

src/base/kernel/config/BaseConfig.h

@@ -56,6 +56,7 @@ public:
     inline const String &apiId() const             { return m_apiId; }
     inline const String &apiWorkerId() const       { return m_apiWorkerId; }
     inline uint32_t printTime() const              { return m_printTime; }
+    inline uint32_t version() const                { return m_version; }
 
     inline bool isWatch() const override                   { return m_watch && !m_fileName.isNull(); }
     inline const String &fileName() const override         { return m_fileName; }
@@ -80,7 +81,8 @@ protected:
     String m_fileName;
     String m_logFile;
     String m_userAgent;
-    uint32_t m_printTime;
+    uint32_t m_printTime = 60;
+    uint32_t m_version   = 0;
 
 private:
     inline void setPrintTime(uint32_t printTime) { if (printTime <= 3600) { m_printTime = printTime; } }

src/base/kernel/config/BaseTransform.cpp

@@ -33,11 +33,12 @@
 #endif
 
 
-#include "base/kernel/config/BaseTransform.h"
-#include "base/kernel/Process.h"
-#include "base/io/log/Log.h"
-#include "base/kernel/interfaces/IConfig.h"
 #include "base/io/json/JsonChain.h"
+#include "base/io/log/Log.h"
+#include "base/kernel/config/BaseTransform.h"
+#include "base/kernel/interfaces/IConfig.h"
+#include "base/kernel/Process.h"
+#include "base/net/stratum/Pool.h"
 #include "core/config/Config_platform.h"
 
 
@@ -138,7 +139,19 @@ void xmrig::BaseTransform::transform(rapidjson::Document &doc, int key, const ch
         break;
 
     case IConfig::UrlKey: /* --url */
-        return add(doc, kPools, "url", arg, true);
+    {
+        if (!doc.HasMember(kPools)) {
+            doc.AddMember(rapidjson::StringRef(kPools), rapidjson::kArrayType, doc.GetAllocator());
+        }
+
+        rapidjson::Value &array = doc[kPools];
+        if (array.Size() == 0 || Pool(array[array.Size() - 1]).isValid()) {
+            array.PushBack(rapidjson::kObjectType, doc.GetAllocator());
+        }
+
+        set(doc, array[array.Size() - 1], "url", arg);
+        break;
+    }
 
     case IConfig::UserKey: /* --user */
         return add(doc, kPools, "user", arg);

src/base/net/stratum/BaseClient.cpp

@@ -42,7 +42,8 @@ xmrig::BaseClient::BaseClient(int id, IClientListener *listener) :
     m_retries(5),
     m_failures(0),
     m_state(UnconnectedState),
-    m_retryPause(5000)
+    m_retryPause(5000),
+    m_enabled(true)
 {
 }
 

src/base/net/stratum/BaseClient.h

@@ -64,7 +64,8 @@ protected:
         HostLookupState,
         ConnectingState,
         ConnectedState,
-        ClosingState
+        ClosingState,
+        ReconnectingState
     };
 
     inline bool isQuiet() const { return m_quiet || m_failures >= m_retries; }

src/base/net/stratum/Client.cpp

@@ -70,21 +70,15 @@ static const char *states[] = {
     "host-lookup",
     "connecting",
     "connected",
-    "closing"
+    "closing",
+    "reconnecting"
 };
 #endif
 
 
 xmrig::Client::Client(int id, const char *agent, IClientListener *listener) :
     BaseClient(id, listener),
-    m_agent(agent),
-    m_tls(nullptr),
-    m_expire(0),
-    m_jobs(0),
-    m_keepAlive(0),
-    m_key(0),
-    m_stream(nullptr),
-    m_socket(nullptr)
+    m_agent(agent)
 {
     m_key = m_storage.add(this);
     m_dns = new Dns(this);
@@ -234,10 +228,16 @@ void xmrig::Client::tick(uint64_t now)
         else if (m_keepAlive && now > m_keepAlive) {
             ping();
         }
+
+        return;
     }
 
-    if (m_expire && now > m_expire && m_state == ConnectingState) {
-        connect();
+    if (m_state == ReconnectingState && m_expire && now > m_expire) {
+        return connect();
+    }
+
+    if (m_state == ConnectingState && m_expire && now > m_expire) {
+        return reconnect();
     }
 }
 
@@ -447,7 +447,6 @@ int xmrig::Client::resolve(const String &host)
 {
     setState(HostLookupState);
 
-    m_expire = 0;
     m_recvBuf.reset();
 
     if (m_failures == -1) {
@@ -754,6 +753,8 @@ void xmrig::Client::parseResponse(int64_t id, const rapidjson::Value &result, co
 void xmrig::Client::ping()
 {
     send(snprintf(m_sendBuf, sizeof(m_sendBuf), "{\"id\":%" PRId64 ",\"jsonrpc\":\"2.0\",\"method\":\"keepalived\",\"params\":{\"id\":\"%s\"}}\n", m_sequence, m_rpcId.data()));
+
+    m_keepAlive = 0;
 }
 
 
@@ -810,12 +811,10 @@ void xmrig::Client::reconnect()
         return m_listener->onClose(this, -1);
     }
 
-    setState(ConnectingState);
+    setState(ReconnectingState);
 
     m_failures++;
     m_listener->onClose(this, static_cast<int>(m_failures));
-
-    m_expire = Chrono::steadyMSecs() + m_retryPause;
 }
 
 
@@ -827,6 +826,23 @@ void xmrig::Client::setState(SocketState state)
         return;
     }
 
+    switch (state) {
+    case HostLookupState:
+        m_expire = 0;
+        break;
+
+    case ConnectingState:
+        m_expire = Chrono::steadyMSecs() + kConnectTimeout;
+        break;
+
+    case ReconnectingState:
+        m_expire = Chrono::steadyMSecs() + m_retryPause;
+        break;
+
+    default:
+        break;
+    }
+
     m_state = state;
 }
 
@@ -884,6 +900,12 @@ void xmrig::Client::onConnect(uv_connect_t *req, int status)
             LOG_ERR("[%s] connect error: \"%s\"", client->url(), uv_strerror(status));
         }
 
+        if (client->state() != ConnectingState) {
+            LOG_ERR("[%s] connect error: \"invalid state: %d\"", client->url(), client->state());
+
+            return;
+        }
+
         delete req;
         client->close();
         return;

src/base/net/stratum/Client.h

@@ -56,12 +56,13 @@ class JobResult;
 class Client : public BaseClient, public IDnsListener, public ILineListener
 {
 public:
-    constexpr static int kResponseTimeout = 20 * 1000;
+    constexpr static uint64_t kConnectTimeout  = 20 * 1000;
+    constexpr static uint64_t kResponseTimeout = 20 * 1000;
 
 #   ifdef XMRIG_FEATURE_TLS
-    constexpr static int kInputBufferSize = 1024 * 16;
+    constexpr static size_t kInputBufferSize = 1024 * 16;
 #   else
-    constexpr static int kInputBufferSize = 1024 * 2;
+    constexpr static size_t kInputBufferSize = 1024 * 2;
 #   endif
 
     Client(int id, const char *agent, IClientListener *listener);
@@ -122,19 +123,19 @@ private:
 
     static inline Client *getClient(void *data) { return m_storage.get(data); }
 
-    char m_sendBuf[2048];
+    char m_sendBuf[2048] = { 0 };
     const char *m_agent;
     Dns *m_dns;
     RecvBuf<kInputBufferSize> m_recvBuf;
     std::bitset<EXT_MAX> m_extensions;
     String m_rpcId;
-    Tls *m_tls;
-    uint64_t m_expire;
-    uint64_t m_jobs;
-    uint64_t m_keepAlive;
-    uintptr_t m_key;
-    uv_stream_t *m_stream;
-    uv_tcp_t *m_socket;
+    Tls *m_tls                  = nullptr;
+    uint64_t m_expire           = 0;
+    uint64_t m_jobs             = 0;
+    uint64_t m_keepAlive        = 0;
+    uintptr_t m_key             = 0;
+    uv_stream_t *m_stream       = nullptr;
+    uv_tcp_t *m_socket          = nullptr;
 
     static Storage<Client> m_storage;
 };

src/base/net/stratum/Pool.cpp

@@ -200,6 +200,9 @@ bool xmrig::Pool::isEqual(const Pool &other) const
 bool xmrig::Pool::parse(const char *url)
 {
     assert(url != nullptr);
+    if (url == nullptr) {
+        return false;
+    }
 
     const char *p    = strstr(url, "://");
     const char *base = url;

src/config.json

@@ -11,6 +11,7 @@
         "restricted": true
     },
     "autosave": true,
+    "version": 1,
     "background": false,
     "colors": true,
     "randomx": {
@@ -23,6 +24,7 @@
         "hw-aes": null,
         "priority": null,
         "asm": true,
+        "argon2-impl": null,
         "cn/0": false,
         "cn-lite/0": false
     },

src/core/Miner.cpp

@@ -45,8 +45,8 @@
 
 
 #ifdef XMRIG_FEATURE_API
-#   include "api/Api.h"
-#   include "api/interfaces/IApiRequest.h"
+#   include "base/api/Api.h"
+#   include "base/api/interfaces/IApiRequest.h"
 #endif
 
 
@@ -112,16 +112,15 @@ public:
     {
         active = true;
 
+        if (reset) {
+            Nonce::reset(job.index());
+        }
+
         for (IBackend *backend : backends) {
             backend->setJob(job);
         }
 
-        if (reset) {
-            Nonce::reset(job.index());
-        }
-        else {
-            Nonce::touch();
-        }
+        Nonce::touch();
 
         if (enabled) {
             Nonce::pause(false);;
@@ -477,5 +476,17 @@ void xmrig::Miner::onRequest(IApiRequest &request)
             d_ptr->getBackends(request.reply(), request.doc());
         }
     }
+    else if (request.type() == IApiRequest::REQ_JSON_RPC) {
+        if (request.rpcMethod() == "pause") {
+            request.accept();
+
+            setEnabled(false);
+        }
+        else if (request.rpcMethod() == "resume") {
+            request.accept();
+
+            setEnabled(true);
+        }
+    }
 }
 #endif

src/core/Miner.h

@@ -29,7 +29,7 @@
 #include <vector>
 
 
-#include "api/interfaces/IApiListener.h"
+#include "base/api/interfaces/IApiListener.h"
 #include "base/kernel/interfaces/IBaseListener.h"
 #include "base/kernel/interfaces/ITimerListener.h"
 #include "crypto/common/Algorithm.h"

src/core/config/Config.cpp

@@ -40,7 +40,8 @@
 
 namespace xmrig {
 
-static const char *kCPU     = "cpu";
+static const char *kCPU                  = "cpu";
+static constexpr const uint32_t kVersion = 1;
 
 #ifdef XMRIG_ALGO_RANDOMX
 static const char *kRandomX = "randomx";
@@ -54,13 +55,27 @@ xmrig::Config::Config() : BaseConfig()
 }
 
 
+bool xmrig::Config::isShouldSave() const
+{
+    if (!isAutoSave()) {
+        return false;
+    }
+
+    if (version() < kVersion) {
+        return true;
+    }
+
+    return (m_shouldSave || m_upgrade || m_cpu.isShouldSave());
+}
+
+
 bool xmrig::Config::read(const IJsonReader &reader, const char *fileName)
 {
     if (!BaseConfig::read(reader, fileName)) {
         return false;
     }
 
-    m_cpu.read(reader.getValue(kCPU));
+    m_cpu.read(reader.getValue(kCPU), version());
 
 #   ifdef XMRIG_ALGO_RANDOMX
     if (!m_rx.read(reader.getValue(kRandomX))) {
@@ -81,12 +96,13 @@ void xmrig::Config::getJSON(rapidjson::Document &doc) const
     auto &allocator = doc.GetAllocator();
 
     Value api(kObjectType);
-    api.AddMember("id",           m_apiId.toJSON(), allocator);
-    api.AddMember("worker-id",    m_apiWorkerId.toJSON(), allocator);
+    api.AddMember("id",                m_apiId.toJSON(), allocator);
+    api.AddMember("worker-id",         m_apiWorkerId.toJSON(), allocator);
 
     doc.AddMember("api",               api, allocator);
     doc.AddMember("http",              m_http.toJSON(doc), allocator);
     doc.AddMember("autosave",          isAutoSave(), allocator);
+    doc.AddMember("version",           kVersion, allocator);
     doc.AddMember("background",        isBackground(), allocator);
     doc.AddMember("colors",            Log::colors, allocator);
 

src/core/config/Config.h

@@ -50,10 +50,10 @@ class Config : public BaseConfig
 public:
     Config();
 
+    bool isShouldSave() const;
     bool read(const IJsonReader &reader, const char *fileName) override;
     void getJSON(rapidjson::Document &doc) const override;
 
-    inline bool isShouldSave() const        { return (m_shouldSave || m_upgrade || m_cpu.isShouldSave()) && isAutoSave(); }
     inline const CpuConfig &cpu() const     { return m_cpu; }
 
 #   ifdef XMRIG_ALGO_RANDOMX

src/core/config/ConfigTransform.cpp

@@ -96,6 +96,8 @@ void xmrig::ConfigTransform::finalize(rapidjson::Document &doc)
     BaseTransform::finalize(doc);
 
     if (m_threads) {
+        doc.AddMember("version", 1, allocator);
+
         if (!doc.HasMember(kCpu)) {
             doc.AddMember(StringRef(kCpu), Value(kObjectType), allocator);
         }