rgw: URL-decode S3 and Swift object-copy URLs

Fixes: http://tracker.ceph.com/issues/22121
Signed-off-by: Matt Benjamin <mbenjamin@redhat.com>

diff --git a/src/rgw/rgw_op.cc b/src/rgw/rgw_op.cc
index 119552821012238544f5256483a87dc25b49fb38..12d006cf3630db45ec3f9db3f224598e282be95f 100644 (file)
--- a/src/rgw/rgw_op.cc
+++ b/src/rgw/rgw_op.cc
@@ -2932,7 +2932,7 @@ void RGWDeleteBucket::execute()
 
 int RGWPutObj::verify_permission()
 {
-  if (copy_source) {
+  if (! copy_source.empty()) {
 
     RGWAccessControlPolicy cs_acl(s->cct);
     boost::optional<Policy> policy;
@@ -2945,8 +2945,8 @@ int RGWPutObj::verify_permission()
     store->set_prefetch_data(s->obj_ctx, obj);
 
     /* check source object permissions */
-    if (read_obj_policy(store, s, copy_source_bucket_info, cs_attrs, &cs_acl, policy,
-                       cs_bucket, cs_object) < 0) {
+    if (read_obj_policy(store, s, copy_source_bucket_info, cs_attrs, &cs_acl,
+                       policy, cs_bucket, cs_object) < 0) {
       return -EACCES;
     }
 
@@ -3405,7 +3405,7 @@ void RGWPutObj::execute()
     bufferlist data;
     if (fst > lst)
       break;
-    if (!copy_source) {
+    if (copy_source.empty()) {
       len = get_data(data);
     } else {
       uint64_t cur_lst = min(fst + s->cct->_conf->rgw_max_chunk_size - 1, lst);
@@ -4280,11 +4280,12 @@ void RGWDeleteObj::execute()
   }
 }
 
-
-bool RGWCopyObj::parse_copy_location(const string& url_src, string& bucket_name, rgw_obj_key& key)
+bool RGWCopyObj::parse_copy_location(const boost::string_view& url_src,
+                                    string& bucket_name,
+                                    rgw_obj_key& key)
 {
-  string name_str;
-  string params_str;
+  boost::string_view name_str;
+  boost::string_view params_str;
 
   size_t pos = url_src.find('?');
   if (pos == string::npos) {
@@ -4294,27 +4295,27 @@ bool RGWCopyObj::parse_copy_location(const string& url_src, string& bucket_name,
     params_str = url_src.substr(pos + 1);
   }
 
-  std::string dec_src = url_decode(name_str);
-  const char *src = dec_src.c_str();
-
-  if (*src == '/') ++src;
+  boost::string_view dec_src{name_str};
+  if (dec_src[0] == '/')
+    dec_src.remove_prefix(1);
 
-  string str(src);
-
-  pos = str.find('/');
+  pos = dec_src.find('/');
   if (pos ==string::npos)
     return false;
 
-  bucket_name = str.substr(0, pos);
-  key.name = str.substr(pos + 1);
+  boost::string_view bn_view{dec_src.substr(0, pos)};
+  bucket_name = std::string{bn_view.data(), bn_view.size()};
+
+  boost::string_view kn_view{dec_src.substr(pos + 1)};
+  key.name = std::string{kn_view.data(), kn_view.size()};
 
   if (key.name.empty()) {
     return false;
   }
 
-  if (!params_str.empty()) {
+  if (! params_str.empty()) {
     RGWHTTPArgs args;
-    args.set(params_str);
+    args.set(params_str.to_string());
     args.parse();
 
     key.instance = args.get("versionId", NULL);

diff --git a/src/rgw/rgw_op.h b/src/rgw/rgw_op.h
index 50fbb6865f48936ba04a36fba8151c68edf9f329..45958505cc37c59c512ba9bf5abac59549f77403 100644 (file)
--- a/src/rgw/rgw_op.h
+++ b/src/rgw/rgw_op.h
@@ -978,7 +978,7 @@ protected:
   const char *supplied_etag;
   const char *if_match;
   const char *if_nomatch;
-  const char *copy_source;
+  std::string copy_source;
   const char *copy_source_range;
   RGWBucketInfo copy_source_bucket_info;
   string copy_source_tenant_name;
@@ -1009,7 +1009,6 @@ public:
                 supplied_etag(NULL),
                 if_match(NULL),
                 if_nomatch(NULL),
-                copy_source(NULL),
                 copy_source_range(NULL),
                 copy_source_range_fst(0),
                 copy_source_range_lst(0),
@@ -1320,7 +1319,7 @@ public:
     copy_if_newer = false;
   }
 
-  static bool parse_copy_location(const string& src,
+  static bool parse_copy_location(const boost::string_view& src,
                                   string& bucket_name,
                                   rgw_obj_key& object);
 

diff --git a/src/rgw/rgw_rest_s3.cc b/src/rgw/rgw_rest_s3.cc
index b9b5d1b6ffea3d2d612198ed3bd3805ce821884f..ec774deb3b9add3ccbb4b4ee97216c424ce34b66 100644 (file)
--- a/src/rgw/rgw_rest_s3.cc
+++ b/src/rgw/rgw_rest_s3.cc
@@ -1252,29 +1252,33 @@ int RGWPutObj_ObjStore_S3::get_params()
 
   if_match = s->info.env->get("HTTP_IF_MATCH");
   if_nomatch = s->info.env->get("HTTP_IF_NONE_MATCH");
-  copy_source = s->info.env->get("HTTP_X_AMZ_COPY_SOURCE");
+  copy_source = url_decode(s->info.env->get("HTTP_X_AMZ_COPY_SOURCE", ""));
   copy_source_range = s->info.env->get("HTTP_X_AMZ_COPY_SOURCE_RANGE");
 
   /* handle x-amz-copy-source */
-
-  if (copy_source) {
-    if (*copy_source == '/') ++copy_source;
-    copy_source_bucket_name = copy_source;
+  boost::string_view cs_view(copy_source);
+  if (! cs_view.empty()) {
+    if (cs_view[0] == '/')
+      cs_view.remove_prefix(1);
+    copy_source_bucket_name = cs_view.to_string();
     pos = copy_source_bucket_name.find("/");
     if (pos == std::string::npos) {
       ret = -EINVAL;
       ldout(s->cct, 5) << "x-amz-copy-source bad format" << dendl;
       return ret;
     }
-    copy_source_object_name = copy_source_bucket_name.substr(pos + 1, copy_source_bucket_name.size());
+    copy_source_object_name =
+      copy_source_bucket_name.substr(pos + 1, copy_source_bucket_name.size());
     copy_source_bucket_name = copy_source_bucket_name.substr(0, pos);
 #define VERSION_ID_STR "?versionId="
     pos = copy_source_object_name.find(VERSION_ID_STR);
     if (pos == std::string::npos) {
       copy_source_object_name = url_decode(copy_source_object_name);
     } else {
-      copy_source_version_id = copy_source_object_name.substr(pos + sizeof(VERSION_ID_STR) - 1);
-      copy_source_object_name = url_decode(copy_source_object_name.substr(0, pos));
+      copy_source_version_id =
+       copy_source_object_name.substr(pos + sizeof(VERSION_ID_STR) - 1);
+      copy_source_object_name =
+       url_decode(copy_source_object_name.substr(0, pos));
     }
     pos = copy_source_bucket_name.find(":");
     if (pos == std::string::npos) {
@@ -1376,7 +1380,7 @@ void RGWPutObj_ObjStore_S3::send_response()
        s->cct->_conf->rgw_s3_success_create_obj_status);
       set_req_state_err(s, op_ret);
     }
-    if (!copy_source) {
+    if (copy_source.empty()) {
       dump_errno(s);
       dump_etag(s, etag);
       dump_content_length(s, 0);
@@ -3274,11 +3278,11 @@ int RGWHandler_REST_S3::init(RGWRados *store, struct req_state *s,
   s->has_acl_header = s->info.env->exists_prefix("HTTP_X_AMZ_GRANT");
 
   const char *copy_source = s->info.env->get("HTTP_X_AMZ_COPY_SOURCE");
-
-  if (copy_source && !s->info.env->get("HTTP_X_AMZ_COPY_SOURCE_RANGE")) {
-    ret = RGWCopyObj::parse_copy_location(copy_source,
-                                          s->init_state.src_bucket,
-                                          s->src_object);
+  if (copy_source &&
+      (! s->info.env->get("HTTP_X_AMZ_COPY_SOURCE_RANGE"))) {
+       ret = RGWCopyObj::parse_copy_location(url_decode(copy_source),
+                                           s->init_state.src_bucket,
+                                           s->src_object);
     if (!ret) {
       ldout(s->cct, 0) << "failed to parse copy location" << dendl;
       return -EINVAL; // XXX why not -ERR_INVALID_BUCKET_NAME or -ERR_BAD_URL?

diff --git a/src/rgw/rgw_rest_swift.cc b/src/rgw/rgw_rest_swift.cc
index a21f31bfea8ea6c9193166a255021e249e2af1b1..199c6a084d032fa442a76e3f666a078c85a97d82 100644 (file)
--- a/src/rgw/rgw_rest_swift.cc
+++ b/src/rgw/rgw_rest_swift.cc
@@ -2909,8 +2909,9 @@ int RGWHandler_REST_SWIFT::init(RGWRados* store, struct req_state* s,
 
   s->dialect = "swift";
 
-  const char *copy_source = s->info.env->get("HTTP_X_COPY_FROM");
-  if (copy_source) {
+  std::string copy_source =
+    url_decode(s->info.env->get("HTTP_X_COPY_FROM", ""));
+  if (! copy_source.empty()) {
     bool result = RGWCopyObj::parse_copy_location(copy_source, t->src_bucket,
                                                  s->src_object);
     if (!result)
@@ -2918,11 +2919,12 @@ int RGWHandler_REST_SWIFT::init(RGWRados* store, struct req_state* s,
   }
 
   if (s->op == OP_COPY) {
-    const char *req_dest = s->info.env->get("HTTP_DESTINATION");
-    if (!req_dest)
+    std::string req_dest =
+      url_decode(s->info.env->get("HTTP_DESTINATION", ""));
+    if (req_dest.empty())
       return -ERR_BAD_URL;
 
-    string dest_bucket_name;
+    std::string dest_bucket_name;
     rgw_obj_key dest_obj_key;
     bool result =
       RGWCopyObj::parse_copy_location(req_dest, dest_bucket_name,
@@ -2930,7 +2932,7 @@ int RGWHandler_REST_SWIFT::init(RGWRados* store, struct req_state* s,
     if (!result)
        return -ERR_BAD_URL;
 
-    string dest_object = dest_obj_key.name;
+    std::string dest_object = dest_obj_key.name;
 
     /* convert COPY operation into PUT */
     t->src_bucket = t->url_bucket;