Add workaround for data race in libstdc++'s locale(const locale&, Facet*) constructor

7 files changed, 66 insertions, 8 deletions

diff --git a/build2/b.cxx b/build2/b.cxx
index 5ea4f02..4a446ac 100644
--- a/build2/b.cxx
+++ b/build2/b.cxx
@@ -445,7 +445,7 @@ main (int argc, char* argv[])
       config_preprocess_create = &config::preprocess_create;
 
       bm["dist"]    = mf {&dist::boot, &dist::init};
-      bm["test"]    = mf {&test::boot, &test::init};
+      bm["test"]    = test::build2_test_load ();
       bm["install"] = mf {&install::boot, &install::init};
       bm["version"] = mf {&version::boot, &version::init};
 
diff --git a/build2/test/init.cxx b/build2/test/init.cxx
index 725d557..1f5a3ae 100644
--- a/build2/test/init.cxx
+++ b/build2/test/init.cxx
@@ -15,6 +15,8 @@
 #include <build2/test/target.hxx>
 #include <build2/test/operation.hxx>
 
+#include <build2/test/script/regex.hxx> // script::regex::init()
+
 using namespace std;
 using namespace butl;
 
@@ -217,5 +219,13 @@ namespace build2
 
       return true;
     }
+
+    module_functions
+    build2_test_load ()
+    {
+      script::regex::init ();
+
+      return module_functions {&boot, &init};
+    }
   }
 }
diff --git a/build2/test/init.hxx b/build2/test/init.hxx
index f429645..5272a4d 100644
--- a/build2/test/init.hxx
+++ b/build2/test/init.hxx
@@ -25,6 +25,9 @@ namespace build2
           bool,
           bool,
           const variable_map&);
+
+    extern "C" module_functions
+    build2_test_load ();
   }
 }
 
diff --git a/build2/test/script/regex.cxx b/build2/test/script/regex.cxx
index 6b15266..bbd1738 100644
--- a/build2/test/script/regex.cxx
+++ b/build2/test/script/regex.cxx
@@ -2,6 +2,8 @@
 // copyright : Copyright (c) 2014-2019 Code Synthesis Ltd
 // license   : MIT; see accompanying LICENSE file
 
+#include <locale>
+
 #include <build2/test/script/regex.hxx>
 
 using namespace std;
@@ -163,11 +165,40 @@ namespace build2
 
         // line_char_locale
         //
+
+        // An exemplar locale with the std::ctype<line_char> facet.  It is
+        // used for the subsequent line char locale objects creation (see
+        // below) which normally ends up with a shallow copy of a reference-
+        // counted object.
+        //
+        // Note that creating the line char locales from the exemplar is not
+        // merely an optimization: there is a data race in the libstdc++ (at
+        // least as of GCC 9.1) implementation of the locale(const locale&,
+        // Facet*) constructor (bug #91057).
+        //
+        // Also note that we install the facet in init() rather than during
+        // the object creation to avoid a race with the std::locale-related
+        // global variables initialization.
+        //
+        static locale line_char_locale_exemplar;
+
+        void
+        init ()
+        {
+          line_char_locale_exemplar =
+            locale (locale (),
+                    new std::ctype<line_char> ()); // Hidden by ctype bitmask.
+        }
+
         line_char_locale::
         line_char_locale ()
-            : locale (locale (),
-                      new std::ctype<line_char> ()) // Hidden by ctype bitmask.
+            : locale (line_char_locale_exemplar)
         {
+          // Make sure init() has been called.
+          //
+          // Note: has_facet() is hidden by a private function in libc++.
+          //
+          assert (std::has_facet<std::ctype<line_char>> (*this));
         }
 
         // char_regex
diff --git a/build2/test/script/regex.hxx b/build2/test/script/regex.hxx
index 500c21b..33a4cba 100644
--- a/build2/test/script/regex.hxx
+++ b/build2/test/script/regex.hxx
@@ -344,6 +344,13 @@ namespace build2
           //
           line_char_locale ();
         };
+
+        // Initialize the testscript regex global state. Should be called once
+        // prior to creating objects of types from this namespace. Note: not
+        // thread-safe.
+        //
+        void
+        init ();
       }
     }
   }
diff --git a/build2/test/script/regex.test.cxx b/build2/test/script/regex.test.cxx
index 7b89e4d..1e48f97 100644
--- a/build2/test/script/regex.test.cxx
+++ b/build2/test/script/regex.test.cxx
@@ -19,6 +19,8 @@ main ()
   using cf = char_flags;
   using cr = char_regex;
 
+  init (); // Initializes the testscript regex global state.
+
   // Test line_char.
   //
   {
@@ -183,7 +185,6 @@ main ()
     using ct = ctype<lc>;
 
     line_char_locale l;
-    assert (has_facet<ct> (l));
 
     // It is better not to create q facet on stack as it is
     // reference-countable.
diff --git a/libbuild2/module.hxx b/libbuild2/module.hxx
index 5fbed9c..7d94837 100644
--- a/libbuild2/module.hxx
+++ b/libbuild2/module.hxx
@@ -56,12 +56,18 @@ namespace build2
     module_init_function* init;
   };
 
-  // The register() function will be written in C++ and will be called from
-  // C++ but we need to suppress name mangling to be able to use dlsym() and
-  // equivalent.
+  // The build2_<modname>_load() function will be written in C++ and will be
+  // called from C++ but we need to suppress name mangling to be able to use
+  // dlsym() or equivalent.
+  //
+  // Note that the load() function is guaranteed to be called during serial
+  // execution (either from main() or during the load phase).
+  //
+  // @@ I wonder if returning a "C++ struct" (it contains pointer to functions
+  //    with signatures containing C++ type) is kosher.
   //
   extern "C"
-  using module_register_function = module_functions ();
+  using module_load_function = module_functions ();
 
   // Loaded modules state.
   //