summaryrefslogtreecommitdiffstats
path: root/mysys/my_new.cc
diff options
context:
space:
mode:
Diffstat (limited to 'mysys/my_new.cc')
-rw-r--r--mysys/my_new.cc117
1 files changed, 117 insertions, 0 deletions
diff --git a/mysys/my_new.cc b/mysys/my_new.cc
new file mode 100644
index 00000000..ae92d6a2
--- /dev/null
+++ b/mysys/my_new.cc
@@ -0,0 +1,117 @@
+/* Copyright (c) 2000, 2001, 2003-2006 MySQL AB, 2009 Sun Microsystems, Inc.
+ Use is subject to license terms.
+
+ 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; version 2 of the License.
+
+ 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, write to the Free Software
+ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1335 USA */
+
+/*
+ This is a replacement of new/delete operators to be used when compiling
+ with gcc 3.0.x to avoid including libstdc++
+
+ It is also used to make all memory allocations to go through
+ my_malloc/my_free wrappers (for debugging/safemalloc and accounting)
+*/
+
+#include "mysys_priv.h"
+#include <new>
+
+/*
+ We don't yet enable the my new operators by default.
+ The reasons (for MariaDB) are:
+
+ - There are several global objects in plugins (wsrep_info, InnoDB,
+ tpool) that allocates data with 'new'. These objects are not
+ freed properly before exit() is called and safemalloc will report
+ these as lost memory. The proper fix is to ensure that all plugins
+ either ensure that all objects frees there data or the global object are
+ changed to a pointer that as allocated and freed on demand.
+ Doing this will make it easier to find leaks and also speed up plugin
+ loads when we don't have to initialize a lot of objects until they
+ are really needed.
+ - Rocksdb calls malloc_usable_size, that will crash if used with new based
+ on my_malloc. One suggested fix would be to not define
+ ROCKSDB_MALLOC_USABLE_SIZE if MYSYS_USE_NEW is defined.
+
+ When the above is fixed, we can add enable ADD_DEFINITIONS(-DUSE_MYSYS_NEW)
+ in CMakeLists.txt
+*/
+
+#if defined(USE_MYSYS_NEW)
+
+void *operator new (size_t sz)
+{
+ return (void *) my_malloc(key_memory_new, sz ? sz : 1, MYF(0));
+}
+
+void *operator new[] (size_t sz)
+{
+ return (void *) my_malloc(key_memory_new, sz ? sz : 1, MYF(0));
+}
+
+void* operator new(std::size_t sz, const std::nothrow_t&) throw()
+{
+ return (void *) my_malloc(key_memory_new, sz ? sz : 1, MYF(0));
+}
+
+void* operator new[](std::size_t sz, const std::nothrow_t&) throw()
+{
+ return (void *) my_malloc(key_memory_new, sz ? sz : 1, MYF(0));
+}
+
+void operator delete (void *ptr, std::size_t) throw ()
+{
+ my_free(ptr);
+}
+
+void operator delete (void *ptr) throw ()
+{
+ my_free(ptr);
+}
+
+void operator delete[] (void *ptr) throw ()
+{
+ my_free(ptr);
+}
+
+void operator delete[] (void *ptr, std::size_t) throw ()
+{
+ my_free(ptr);
+}
+
+void operator delete(void* ptr, const std::nothrow_t&) throw()
+{
+ my_free(ptr);
+}
+
+void operator delete[](void* ptr, const std::nothrow_t&) throw()
+{
+ my_free(ptr);
+}
+
+C_MODE_START
+
+int __cxa_pure_virtual()
+{
+ assert(! "Aborted: pure virtual method called.");
+ return 0;
+}
+
+C_MODE_END
+#else
+/*
+ Define a dummy symbol, just to avoid compiler/linker warnings
+ about compiling an essentially empty file.
+*/
+int my_new_cc_symbol;
+#endif /* USE_MYSYS_NEW */
+