rgw: consolidate code that implements hashing algorithms

Copies of the same hashing algorithm are scattered throughout the rgw
codebase. Consolidate them.

Signed-off-by: J. Eric Ivancich <ivancich@redhat.com>

diff --git a/src/rgw/rgw_gc.cc b/src/rgw/rgw_gc.cc
index e29af16ed1762379f302745a141d31aef819b6e8..bf64925abda588a90918314202fce9bbd72dcf91 100644 (file)
--- a/src/rgw/rgw_gc.cc
+++ b/src/rgw/rgw_gc.cc
@@ -43,7 +43,7 @@ void RGWGC::finalize()
 
 int RGWGC::tag_index(const string& tag)
 {
-  return rgw_shards_hash(tag, max_objs);
+  return rgw_shard_id(tag, max_objs);
 }
 
 void RGWGC::add_chain(ObjectWriteOperation& op, cls_rgw_obj_chain& chain, const string& tag)

diff --git a/src/rgw/rgw_rados.cc b/src/rgw/rgw_rados.cc
index b9181d47359af4af0989340b721c086953557e44..eb4eafa576f7abc15f4060356ee22929a0621ec6 100644 (file)
--- a/src/rgw/rgw_rados.cc
+++ b/src/rgw/rgw_rados.cc
@@ -5121,7 +5121,7 @@ next:
 
 int RGWRados::key_to_shard_id(const string& key, int max_shards)
 {
-  return rgw_shards_hash(key, max_shards);
+  return rgw_shard_id(key, max_shards);
 }
 
 void RGWRados::shard_name(const string& prefix, unsigned max_shards, const string& key, string& name, int *shard_id)
@@ -5302,10 +5302,7 @@ int RGWRados::objexp_key_shard(const rgw_obj_index_key& key)
 {
   string obj_key = key.name + key.instance;
   int num_shards = cct->_conf->rgw_objexp_hints_num_shards;
-  uint32_t sid = ceph_str_hash_linux(obj_key.c_str(), obj_key.size());
-  uint32_t sid2 = sid ^ ((sid & 0xFF) << 24);
-  sid = rgw_shards_mod(sid2, num_shards);
-  return sid;
+  return rgw_bucket_shard_index(obj_key, num_shards);
 }
 
 static string objexp_hint_get_keyext(const string& tenant_name,
@@ -13384,9 +13381,7 @@ int RGWRados::get_target_shard_id(const RGWBucketInfo& bucket_info, const string
           *shard_id = -1;
         }
       } else {
-        uint32_t sid = ceph_str_hash_linux(obj_key.c_str(), obj_key.size());
-        uint32_t sid2 = sid ^ ((sid & 0xFF) << 24);
-        sid = rgw_shards_mod(sid2, bucket_info.num_shards);
+        uint32_t sid = rgw_bucket_shard_index(obj_key, bucket_info.num_shards);
         if (shard_id) {
           *shard_id = (int)sid;
         }
@@ -13424,9 +13419,7 @@ int RGWRados::get_bucket_index_object(const string& bucket_oid_base, const strin
           *shard_id = -1;
         }
       } else {
-        uint32_t sid = ceph_str_hash_linux(obj_key.c_str(), obj_key.size());
-        uint32_t sid2 = sid ^ ((sid & 0xFF) << 24);
-        sid = rgw_shards_mod(sid2, num_shards);
+        uint32_t sid = rgw_bucket_shard_index(obj_key, num_shards);
         char buf[bucket_oid_base.size() + 32];
         snprintf(buf, sizeof(buf), "%s.%d", bucket_oid_base.c_str(), sid);
         (*bucket_obj) = buf;

diff --git a/src/rgw/rgw_rados.h b/src/rgw/rgw_rados.h
index 220aa5e90a2f3b14953102661112161e9b31b6b4..09ae86c64dd6b189e7b0f0da201eedd5102a77ea 100644 (file)
--- a/src/rgw/rgw_rados.h
+++ b/src/rgw/rgw_rados.h
@@ -57,6 +57,7 @@ class RGWReshardWait;
 #define RGW_SHARDS_PRIME_0 7877
 #define RGW_SHARDS_PRIME_1 65521
 
+// only called by rgw_shard_id and rgw_bucket_shard_index
 static inline int rgw_shards_mod(unsigned hval, int max_shards)
 {
   if (max_shards <= RGW_SHARDS_PRIME_0) {
@@ -65,9 +66,19 @@ static inline int rgw_shards_mod(unsigned hval, int max_shards)
   return hval % RGW_SHARDS_PRIME_1 % max_shards;
 }
 
-static inline int rgw_shards_hash(const string& key, int max_shards)
+// used for logging and tagging
+static inline int rgw_shard_id(const string& key, int max_shards)
 {
-  return rgw_shards_mod(ceph_str_hash_linux(key.c_str(), key.size()), max_shards);
+  return rgw_shards_mod(ceph_str_hash_linux(key.c_str(), key.size()),
+                       max_shards);
+}
+
+// used for bucket indices
+static inline uint32_t rgw_bucket_shard_index(const std::string& key,
+                                             int num_shards) {
+  uint32_t sid = ceph_str_hash_linux(key.c_str(), key.size());
+  uint32_t sid2 = sid ^ ((sid & 0xFF) << 24);
+  return rgw_shards_mod(sid2, num_shards);
 }
 
 static inline int rgw_shards_max()

diff --git a/src/rgw/rgw_reshard.cc b/src/rgw/rgw_reshard.cc
index 0389ca33dce293c4fba356db6b3568d5416b1e76..c4d04ae84456c278a15261ebd179c02e9a6e6ac7 100644 (file)
--- a/src/rgw/rgw_reshard.cc
+++ b/src/rgw/rgw_reshard.cc
@@ -588,9 +588,8 @@ void RGWReshard::get_bucket_logshard_oid(const string& tenant, const string& buc
   uint32_t sid = ceph_str_hash_linux(key.c_str(), key.size());
   uint32_t sid2 = sid ^ ((sid & 0xFF) << 24);
   sid = sid2 % MAX_RESHARD_LOGSHARDS_PRIME % num_logshards;
-  int logshard = sid % num_logshards;
 
-  get_logshard_oid(logshard, oid);
+  get_logshard_oid(int(sid), oid);
 }
 
 int RGWReshard::add(cls_rgw_reshard_entry& entry)