feat: support mmap for marisa trie (#29613)

this supports mmap for marisa trie index
related https://github.com/milvus-io/milvus/issues/21866

Signed-off-by: yah01 <yang.cen@zilliz.com>

internal/core/src/common/File.h

@@ -13,6 +13,7 @@
 
 #include <string>
 #include "common/EasyAssert.h"
+#include "common/Types.h"
 #include "fmt/core.h"
 #include <fcntl.h>
 #include <unistd.h>
@@ -51,6 +52,11 @@ class File {
         return write(fd_, buf, size);
     }
 
+    offset_t
+    Seek(offset_t offset, int whence) {
+        return lseek(fd_, offset, whence);
+    }
+
     void
     Close() {
         close(fd_);

internal/core/src/index/StringIndexMarisa.cpp

@@ -17,11 +17,16 @@
 #include <boost/uuid/uuid.hpp>
 #include <boost/uuid/uuid_io.hpp>
 #include <boost/uuid/uuid_generators.hpp>
+#include <cstring>
 #include <memory>
 #include <stdlib.h>
 #include <stdio.h>
 #include <fcntl.h>
+#include <sys/errno.h>
+#include <sys/mman.h>
+#include <unistd.h>
 
+#include "common/File.h"
 #include "common/Types.h"
 #include "common/EasyAssert.h"
 #include "common/Exception.h"
@@ -249,28 +254,29 @@ StringIndexMarisa::LoadWithoutAssemble(const BinarySet& set,
                                        const Config& config) {
     auto uuid = boost::uuids::random_generator()();
     auto uuid_string = boost::uuids::to_string(uuid);
-    auto file = std::string("/tmp/") + uuid_string;
+    auto file_name = std::string("/tmp/") + uuid_string;
 
     auto index = set.GetByName(MARISA_TRIE_INDEX);
     auto len = index->size;
 
-    auto fd = open(
-        file.c_str(), O_RDWR | O_CREAT | O_EXCL, S_IRUSR | S_IWUSR | S_IXUSR);
-    lseek(fd, 0, SEEK_SET);
-
-    auto status = write(fd, index->data.get(), len);
-    if (status != len) {
-        close(fd);
-        remove(file.c_str());
+    auto file = File::Open(file_name, O_RDWR | O_CREAT | O_EXCL);
+    auto written = file.Write(index->data.get(), len);
+    if (written != len) {
+        file.Close();
+        remove(file_name.c_str());
         throw SegcoreError(
             ErrorCode::UnistdError,
-            "write index to fd error, errorCode is " + std::to_string(status));
+            fmt::format("write index to fd error: {}", strerror(errno)));
     }
 
-    lseek(fd, 0, SEEK_SET);
-    trie_.read(fd);
-    close(fd);
-    remove(file.c_str());
+    file.Seek(0, SEEK_SET);
+    if (config.contains(kEnableMmap)) {
+        trie_.mmap(file_name.c_str());
+    } else {
+        trie_.read(file.Descriptor());
+    }
+    // make sure the file would be removed after we unmap & close it
+    unlink(file_name.c_str());
 
     auto str_ids = set.GetByName(MARISA_STR_IDS);
     auto str_ids_len = str_ids->size;

scripts/run_intergration_test.sh

@@ -31,7 +31,7 @@ beginTime=`date +%s`
 
 for d in $(go list ./tests/integration/...); do
     echo "$d"
-    go test -race -tags dynamic -v -coverpkg=./... -coverprofile=profile.out -covermode=atomic "$d" -timeout=20m
+    go test -race -tags dynamic -v -coverpkg=./... -coverprofile=profile.out -covermode=atomic "$d" -timeout=30m
     if [ -f profile.out ]; then
         grep -v kafka profile.out | grep -v planparserv2/generated | grep -v mocks | sed '1d' >> ${FILE_COVERAGE_INFO}
         rm profile.out