Merge pull request ceph#638 from ceph/wip-bloom

bloom filter cleanups, encodability, and unit tests

COPYING

@@ -23,6 +23,10 @@ Files: src/include/ceph_hash.cc
 Copyright: None
 License: Public domain
 
+Files: src/common/bloom_filter.hpp
+Copyright: Copyright (C) 2000 Arash Partow <[email protected]>
+License: Boost Software License, Version 1.0
+
 Files: m4/acx_pthread.m4
 Copyright: Steven G. Johnson <[email protected]>
 License: GPLWithACException

debian/copyright

@@ -23,6 +23,10 @@ Files: src/include/ceph_hash.cc
 Copyright: None
 License: Public domain
 
+Files: src/common/bloom_filter.hpp
+Copyright: Copyright (C) 2000 Arash Partow
+License: Boost Software License, Version 1.0
+
 Files: m4/acx_pthread.m4
 Copyright: Steven G. Johnson <[email protected]>
 License: GPLWithACException

src/common/Makefile.am

@@ -65,7 +65,8 @@ libcommon_la_SOURCES = \
 	common/ceph_strings.cc \
 	common/ceph_frag.cc \
 	common/addr_parsing.c \
-	common/hobject.cc
+	common/hobject.cc \
+	common/bloom_filter.cc
 
 if LINUX
 libcommon_la_SOURCES += common/secret.c
@@ -97,6 +98,7 @@ LIBCOMMON_DEPS += libcommon_crc.la
 noinst_LTLIBRARIES += libcommon_crc.la
 
 noinst_HEADERS += \
+	common/bloom_filter.hpp \
 	common/sctp_crc32.h \
 	common/crc32c_intel_baseline.h \
 	common/crc32c_intel_fast.h

src/common/bloom_filter.cc

@@ -0,0 +1,76 @@
+// -*- mode:C++; tab-width:8; c-basic-offset:2; indent-tabs-mode:t -*-
+// vim: ts=8 sw=2 smarttab
+
+#include "include/types.h"
+#include "common/bloom_filter.hpp"
+
+void bloom_filter::encode(bufferlist& bl) const
+{
+  ENCODE_START(1, 1, bl);
+  ::encode((uint64_t)salt_count_, bl);
+  ::encode((uint64_t)table_size_, bl);
+  ::encode((uint64_t)inserted_element_count_, bl);
+  ::encode((uint64_t)random_seed_, bl);
+  bufferptr bp((const char*)bit_table_, raw_table_size_);
+  ::encode(bp, bl);
+  ENCODE_FINISH(bl);
+}
+
+void bloom_filter::decode(bufferlist::iterator& p)
+{
+  DECODE_START(1, p);
+  uint64_t v;
+  ::decode(v, p);
+  salt_count_ = v;
+  ::decode(v, p);
+  table_size_ = v;
+  ::decode(v, p);
+  inserted_element_count_ = v;
+  ::decode(v, p);
+  random_seed_ = v;
+  bufferlist t;
+  ::decode(t, p);
+
+  salt_.clear();
+  generate_unique_salt();
+  raw_table_size_ = t.length();
+  assert(raw_table_size_ == table_size_ / bits_per_char);
+  delete bit_table_;
+  bit_table_ = new cell_type[raw_table_size_];
+  t.copy(0, raw_table_size_, (char *)bit_table_);
+
+  DECODE_FINISH(p);
+}
+
+void bloom_filter::dump(Formatter *f) const
+{
+  f->dump_unsigned("salt_count", salt_count_);
+  f->dump_unsigned("table_size", table_size_);
+  f->dump_unsigned("raw_table_size", raw_table_size_);
+  f->dump_unsigned("insert_count", inserted_element_count_);
+  f->dump_unsigned("random_seed", random_seed_);
+
+  f->open_array_section("salt_table");
+  for (std::vector<bloom_type>::const_iterator i = salt_.begin(); i != salt_.end(); ++i)
+    f->dump_unsigned("salt", *i);
+  f->close_section();
+
+  f->open_array_section("bit_table");
+  for (unsigned i = 0; i < raw_table_size_; ++i)
+    f->dump_unsigned("byte", (unsigned)bit_table_[i]);
+  f->close_section();
+}
+
+void bloom_filter::generate_test_instances(list<bloom_filter*>& ls)
+{
+  ls.push_back(new bloom_filter(10, .5, 1));
+  ls.push_back(new bloom_filter(10, .5, 1));
+  ls.back()->insert("foo");
+  ls.back()->insert("bar");
+  ls.push_back(new bloom_filter(50, .5, 1));
+  ls.back()->insert("foo");
+  ls.back()->insert("bar");
+  ls.back()->insert("baz");
+  ls.back()->insert("boof");
+  ls.back()->insert("boogggg");
+}