0001-Hook-UMASH-in-demerphq-s-SMHasher.patch

From 421553b35908d45560040822c0d39a98cc15758b Mon Sep 17 00:00:00 2001
From: Paul Khuong <pkhuong@backtrace.io>
Date: Wed, 26 Aug 2020 23:33:06 +0000
Subject: [PATCH] Hook UMASH in demerphq's SMHasher

---
 CMakeLists.txt |  5 +++--
 build.sh       |  2 +-
 main.cpp       | 16 +++++++++++++--
 umash.hpp      | 55 ++++++++++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 73 insertions(+), 5 deletions(-)
 create mode 100644 umash.hpp

diff --git a/CMakeLists.txt b/CMakeLists.txt
index 57b3116..4c8326d 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -94,8 +94,8 @@ if (CMAKE_COMPILER_IS_GNUCC
     set(CMAKE_C_FLAGS   "${CMAKE_C_FLAGS} -maes")
     set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -maes")
   ENDIF (AES_TRUE)
-  set(CMAKE_C_FLAGS   "${CMAKE_C_FLAGS} -march=native")
-  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -march=native")
+  set(CMAKE_C_FLAGS   "${CMAKE_C_FLAGS} -march=native -mtune=native")
+  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -march=native -mtune=native")
 elseif ("${CMAKE_CXX_COMPILER_ID}" STREQUAL "MSVC")
   # using Visual Studio C++
   set(CMAKE_C_FLAGS   "${CMAKE_C_FLAGS} /arch:SSE2")
@@ -234,6 +234,7 @@ add_library(SMHasherSupport
   mum.cc
   jody_hash32.c
   jody_hash64.c
+  umash.c
 )
 
 add_executable(
diff --git a/build.sh b/build.sh
index b4b9e75..08cdf33 100755
--- a/build.sh
+++ b/build.sh
@@ -9,7 +9,7 @@ case `uname -s` in
 #Darwin)  CXX=g++-mp-6  CC=gcc-mp-6 cmake .. ;;
 Darwin)  cmake .. ;;
 FreeBSD) CXX=clang++ CC=clang cmake .. ;;
-*)       CXX=clang++-4.0  CC=clang-4.0 cmake ..   ;;
+*)       cmake ..   ;;
 esac
 make -j4 $@
 cd ..
diff --git a/main.cpp b/main.cpp
index 7d887a1..aa6ba35 100644
--- a/main.cpp
+++ b/main.cpp
@@ -15,6 +15,7 @@
 #endif
 #include "md5.h"
 #include "siphash.h"
+#include "umash.hpp"
 #include <stdio.h>
 #include <time.h>
 #include "VERSION.h"
@@ -468,6 +469,19 @@ HashInfo g_hashes[] =
   { "MUM", "github.com/vnmakarov/mum-hash",
     64, 64, 64, 0x73F6BFD7,
     NULL, mum_hash_with_state_test },
+
+  { "umash128", "umash 128",
+    64, 8 * sizeof(struct umash_params), 128,
+    0xFB18192F, umash_state_from_seed, umash128_with_state },
+  { "umash64", "umash 64",
+    64, 8 * sizeof(struct umash_params), 64,
+    0x17608BD7, umash_state_from_seed, umash_with_state },
+  { "umash32", "umash 32",
+    64, 8 * sizeof(struct umash_params), 32,
+    0x228589FD, umash_state_from_seed, umash32_with_state },
+  { "umash32_hi", "umash 32 hi",
+    64, 8 * sizeof(struct umash_params), 32,
+    0x799576D1, umash_state_from_seed, umash32_hi_with_state },
 };
 int g_hashes_sizeof= sizeof(g_hashes);
 
@@ -759,8 +773,6 @@ int main ( int argc, char ** argv )
     }
   }
 
-  // Code runs on the 3rd CPU by default
-  SetAffinity((1 << 2));
 
 
   //----------
diff --git a/umash.hpp b/umash.hpp
new file mode 100644
index 0000000..23d7f98
--- /dev/null
+++ b/umash.hpp
@@ -0,0 +1,55 @@
+#pragma once
+
+#include <assert.h>
+#include <string.h>
+
+#include "umash.h"
+
+void
+umash_state_from_seed(int seedbits, const void *seed, void *state)
+{
+	uint64_t bits;
+
+	memcpy(&bits, seed, sizeof(bits));
+	umash_params_derive(static_cast<umash_params *>(state), bits, nullptr);
+	return;
+}
+
+void
+umash128_with_state(const void *blob, int len, const void *state, void *out)
+{
+	struct umash_fp umash;
+
+	umash = umash_fprint(static_cast<const umash_params *>(state), 0, blob, len);
+	memcpy(out, &umash, sizeof(umash));
+	return;
+}
+void
+umash_with_state(const void *blob, int len, const void *state, void *out)
+{
+	uint64_t umash;
+
+	umash = umash_full(static_cast<const umash_params *>(state), 0, 0, blob, len);
+	memcpy(out, &umash, sizeof(umash));
+	return;
+}
+
+void
+umash32_with_state(const void *blob, int len, const void *state, void *out)
+{
+	uint32_t umash;
+
+	umash = umash_full(static_cast<const umash_params *>(state), 0, 0, blob, len);
+	memcpy(out, &umash, sizeof(umash));
+	return;
+}
+
+void
+umash32_hi_with_state(const void *blob, int len, const void *state, void *out)
+{
+	uint32_t umash;
+
+	umash = umash_full(static_cast<const umash_params *>(state), 0, 0, blob, len) >> 32;
+	memcpy(out, &umash, sizeof(umash));
+	return;
+}
-- 
2.17.1