Replace rust's BLAKE3 with llvm's implementation

rust/BLAKE3/CMakeLists.txt

@@ -1,3 +0,0 @@
-clickhouse_import_crate(MANIFEST_PATH Cargo.toml)
-target_include_directories(_ch_rust_blake3 INTERFACE include)
-add_library(ch_rust::blake3 ALIAS _ch_rust_blake3)

rust/BLAKE3/Cargo.toml

@@ -1,20 +0,0 @@
-[package]
-name = "_ch_rust_blake3"
-version = "0.1.0"
-
-# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
-
-[dependencies]
-blake3 = "1.2.0"
-libc = "0.2.132"
-
-[lib]
-crate-type = ["staticlib"]
-
-[profile.release]
-debug = true
-
-[profile.release-thinlto]
-inherits = "release"
-# BLAKE3 module requires "full" LTO (not "thin") to get additional 10% performance benefit
-lto = true

rust/BLAKE3/include/blake3.h

@@ -1,15 +0,0 @@
-#ifndef BLAKE3_H
-#define BLAKE3_H
-
-#include <cstdint>
-
-
-extern "C" {
-
-char *blake3_apply_shim(const char *begin, uint32_t _size, uint8_t *out_char_data);
-
-void blake3_free_char_pointer(char *ptr_to_free);
-
-} // extern "C"
-
-#endif /* BLAKE3_H */

rust/BLAKE3/src/lib.rs

@@ -1,31 +0,0 @@
-extern crate blake3;
-extern crate libc;
-
-use std::ffi::{CString};
-use std::slice;
-use std::os::raw::c_char;
-
-#[no_mangle]
-pub unsafe extern "C" fn blake3_apply_shim(
-    begin: *const c_char,
-    size: u32,
-    out_char_data: *mut u8,
-) -> *mut c_char {
-    if begin.is_null() {
-        let err_str = CString::new("input was a null pointer").unwrap();
-        return err_str.into_raw();
-    }
-    let input_res = slice::from_raw_parts(begin as *const u8, size as usize);
-    let mut hasher = blake3::Hasher::new();
-    hasher.update(input_res);
-    let mut reader = hasher.finalize_xof();
-
-    reader.fill(std::slice::from_raw_parts_mut(out_char_data, blake3::OUT_LEN));
-    std::ptr::null_mut()
-}
-
-// Freeing memory according to docs: https://doc.rust-lang.org/std/ffi/struct.CString.html#method.into_raw
-#[no_mangle]
-pub unsafe extern "C" fn blake3_free_char_pointer(ptr_to_free: *mut c_char) {
-    std::mem::drop(CString::from_raw(ptr_to_free));
-}

rust/CMakeLists.txt

@@ -99,6 +99,5 @@ function(add_rust_subdirectory src)
         VERBATIM)
 endfunction()
 
-add_rust_subdirectory (BLAKE3)
 add_rust_subdirectory (skim)
 add_rust_subdirectory (prql)

rust/Cargo.toml

@@ -1,7 +1,6 @@
 # workspace is required to vendor crates for all packages.
 [workspace]
 members = [
-    "BLAKE3",
     "skim",
     "prql",
 ]

src/Functions/FunctionsHashing.h

@@ -16,7 +16,7 @@
 #include <xxhash.h>
 
 #if USE_BLAKE3
-#    include <blake3.h>
+#    include <llvm/Support/BLAKE3.h>
 #endif
 
 #include <Common/SipHash.h>
@@ -833,13 +833,13 @@ struct ImplBLAKE3
 #else
     static void apply(const char * begin, const size_t size, unsigned char* out_char_data)
     {
-        auto err_msg = blake3_apply_shim(begin, safe_cast<uint32_t>(size), out_char_data);
-        if (err_msg != nullptr)
-        {
-            auto err_st = std::string(err_msg);
-            blake3_free_char_pointer(err_msg);
-            throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Function returned error message: {}", err_st);
-        }
+        static_assert(LLVM_BLAKE3_OUT_LEN == ImplBLAKE3::length);
+        auto & result = *reinterpret_cast<std::array<uint8_t, LLVM_BLAKE3_OUT_LEN> *>(out_char_data);
+
+        llvm::BLAKE3 hasher;
+        if (size > 0)
+            hasher.update(llvm::StringRef(begin, size));
+        hasher.final(result);
     }
 #endif
 };

src/configure_config.cmake

@@ -19,9 +19,6 @@ endif()
 if (TARGET ch_contrib::rdkafka)
     set(USE_RDKAFKA 1)
 endif()
-if (TARGET ch_rust::blake3)
-    set(USE_BLAKE3 1)
-endif()
 if (TARGET ch_rust::skim)
     set(USE_SKIM 1)
 endif()
@@ -103,6 +100,7 @@ endif()
 if (TARGET ch_contrib::llvm)
     set(USE_EMBEDDED_COMPILER ${ENABLE_EMBEDDED_COMPILER})
     set(USE_DWARF_PARSER ${ENABLE_DWARF_PARSER})
+    set(USE_BLAKE3 1)
 endif()
 if (TARGET ch_contrib::unixodbc)
     set(USE_ODBC 1)