From: Matt Benjamin <mbenjamin@redhat.com>
Date: Fri, 12 Feb 2016 22:02:53 +0000 (-0500)
Subject: rgw: add rgw_log_http_headers option
X-Git-Tag: v10.2.6~35^2~2
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=8838bcbabcd2c9b3d7aec2403cad0bad4ad9a057;p=ceph.git

rgw: add rgw_log_http_headers option

Tracks headers that should be handled conditionally (currently,
can only log, so using minimal structure to represent the
mapping).

Adds map of custom headers to rgw_log_entry, and populate it with
headers pre-selected for custom logging in RGWREST.  Added to encoder
and Formatter output.

Some additional optimization possible.

(cherry picked from commit b82919a152217b3cd49afdc28bb890f329c2742a)

Signed-off-by: Matt Benjamin <mbenjamin@redhat.com>
---

diff --git a/src/common/config_opts.h b/src/common/config_opts.h
index 78e67e2c8fc3..2c0bc536218d 100644
--- a/src/common/config_opts.h
+++ b/src/common/config_opts.h
@@ -1423,6 +1423,7 @@ OPTION(rgw_objexp_hints_num_shards, OPT_U32, 127) // maximum number of parts in
 OPTION(rgw_objexp_chunk_size, OPT_U32, 100) // maximum number of entries in a single operation when processing objexp data
 
 OPTION(rgw_enable_static_website, OPT_BOOL, false) // enable static website feature
+OPTION(rgw_log_http_headers, OPT_STR, "" ) // list of HTTP headers to log when seen, ignores case (e.g., http_x_forwarded_for
 
 OPTION(rgw_num_async_rados_threads, OPT_INT, 32) // num of threads to use for async rados operations
 OPTION(rgw_md_notify_interval_msec, OPT_INT, 200) // metadata changes notification interval to followers
diff --git a/src/include/str_list.h b/src/include/str_list.h
index 4ba0cadd960c..12cf1280854e 100644
--- a/src/include/str_list.h
+++ b/src/include/str_list.h
@@ -91,4 +91,12 @@ inline std::string str_join(const std::vector<std::string>& v, std::string sep)
   return r;
 }
 
+static inline std::vector<std::string> get_str_vec(const std::string& str)
+{
+  std::vector<std::string> str_vec;
+  const char *delims = ";,= \t";
+  get_str_vec(str, delims, str_vec);
+  return std::move(str_vec);
+}
+
 #endif
diff --git a/src/rgw/librgw.cc b/src/rgw/librgw.cc
index 1b4decc8aba0..288365e6d687 100644
--- a/src/rgw/librgw.cc
+++ b/src/rgw/librgw.cc
@@ -286,7 +286,8 @@ namespace rgw {
       dout(0) << "ERROR: io->complete_request() returned " << r << dendl;
     }
     if (should_log) {
-      rgw_log_op(store, s, (op ? op->name() : "unknown"), olog);
+      rgw_log_op(store, nullptr /* !rest */, s,
+		 (op ? op->name() : "unknown"), olog);
     }
 
     int http_ret = s->err.http_ret;
diff --git a/src/rgw/rgw_env.cc b/src/rgw/rgw_env.cc
index ffa479be73f5..6436ee8577fa 100644
--- a/src/rgw/rgw_env.cc
+++ b/src/rgw/rgw_env.cc
@@ -7,6 +7,8 @@
 #include <string>
 #include <map>
 
+#include "include/assert.h"
+
 #define dout_subsys ceph_subsys_rgw
 
 RGWEnv::RGWEnv()
diff --git a/src/rgw/rgw_log.cc b/src/rgw/rgw_log.cc
index 3bef606b1041..c444c2787e1c 100644
--- a/src/rgw/rgw_log.cc
+++ b/src/rgw/rgw_log.cc
@@ -12,6 +12,7 @@
 #include "rgw_acl.h"
 #include "rgw_rados.h"
 #include "rgw_client_io.h"
+#include "rgw_rest.h"
 
 #define dout_subsys ceph_subsys_rgw
 
@@ -250,6 +251,15 @@ void rgw_format_ops_log_entry(struct rgw_log_entry& entry, Formatter *formatter)
   formatter->dump_int("total_time", total_time);
   formatter->dump_string("user_agent",  entry.user_agent);
   formatter->dump_string("referrer",  entry.referrer);
+  if (entry.x_headers.size() > 0) {
+    formatter->open_array_section("http_x_headers");
+    for (const auto& iter: entry.x_headers) {
+      formatter->open_object_section(iter.first.c_str());
+      formatter->dump_string(iter.first.c_str(), iter.second);
+      formatter->close_section();
+    }
+    formatter->close_section();
+  }
   formatter->close_section();
 }
 
@@ -290,7 +300,8 @@ void OpsLogSocket::log(struct rgw_log_entry& entry)
   append_output(bl);
 }
 
-int rgw_log_op(RGWRados *store, struct req_state *s, const string& op_name, OpsLogSocket *olog)
+int rgw_log_op(RGWRados *store, RGWREST* const rest, struct req_state *s,
+	       const string& op_name, OpsLogSocket *olog)
 {
   struct rgw_log_entry entry;
   string bucket_id;
@@ -339,6 +350,18 @@ int rgw_log_op(RGWRados *store, struct req_state *s, const string& op_name, OpsL
   set_param_str(s, "REQUEST_URI", entry.uri);
   set_param_str(s, "REQUEST_METHOD", entry.op);
 
+  /* custom header logging */
+  if (rest) {
+    if (rest->log_x_headers()) {
+      for (const auto& iter : s->info.env->get_map()) {
+	if (rest->log_x_header(iter.first)) {
+	  entry.x_headers.insert(
+	    rgw_log_entry::headers_map::value_type(iter.first, iter.second));
+	}
+      }
+    }
+  }
+
   entry.user = s->user->user_id.to_str();
   if (s->object_acl)
     entry.object_owner = s->object_acl->get_owner().get_id();
diff --git a/src/rgw/rgw_log.h b/src/rgw/rgw_log.h
index 51acfdf8f9d4..ab800c7fa2ab 100644
--- a/src/rgw/rgw_log.h
+++ b/src/rgw/rgw_log.h
@@ -4,6 +4,7 @@
 #ifndef CEPH_RGW_LOG_H
 #define CEPH_RGW_LOG_H
 
+#include <boost/container/flat_set.hpp>
 #include "rgw_common.h"
 #include "include/utime.h"
 #include "common/Formatter.h"
@@ -12,6 +13,9 @@
 class RGWRados;
 
 struct rgw_log_entry {
+
+  using headers_map = std::map<std::string, std::string>;
+
   rgw_user object_owner;
   rgw_user bucket_owner;
   string bucket;
@@ -30,9 +34,10 @@ struct rgw_log_entry {
   string user_agent;
   string referrer;
   string bucket_id;
+  headers_map x_headers;
 
   void encode(bufferlist &bl) const {
-    ENCODE_START(8, 5, bl);
+    ENCODE_START(9, 5, bl);
     ::encode(object_owner.id, bl);
     ::encode(bucket_owner.id, bl);
     ::encode(bucket, bl);
@@ -54,6 +59,7 @@ struct rgw_log_entry {
     ::encode(obj, bl);
     ::encode(object_owner, bl);
     ::encode(bucket_owner, bl);
+    ::encode(x_headers, bl);
     ENCODE_FINISH(bl);
   }
   void decode(bufferlist::iterator &p) {
@@ -100,6 +106,9 @@ struct rgw_log_entry {
       ::decode(object_owner, p);
       ::decode(bucket_owner, p);
     }
+    if (struct_v >= 9) {
+      ::decode(x_headers, p);
+    }
     DECODE_FINISH(p);
   }
   void dump(Formatter *f) const;
@@ -123,10 +132,14 @@ public:
   void log(struct rgw_log_entry& entry);
 };
 
-int rgw_log_op(RGWRados *store, struct req_state *s, const string& op_name, OpsLogSocket *olog);
+class RGWREST;
+
+int rgw_log_op(RGWRados *store, RGWREST* const rest, struct req_state *s,
+	       const string& op_name, OpsLogSocket *olog);
 void rgw_log_usage_init(CephContext *cct, RGWRados *store);
 void rgw_log_usage_finalize();
-void rgw_format_ops_log_entry(struct rgw_log_entry& entry, Formatter *formatter);
+void rgw_format_ops_log_entry(struct rgw_log_entry& entry,
+			      Formatter *formatter);
 
-#endif
+#endif /* CEPH_RGW_LOG_H */
 
diff --git a/src/rgw/rgw_main.cc b/src/rgw/rgw_main.cc
index b5adf5ebd25d..d9de14563334 100644
--- a/src/rgw/rgw_main.cc
+++ b/src/rgw/rgw_main.cc
@@ -356,6 +356,9 @@ int main(int argc, const char **argv)
     rest.register_resource(g_conf->rgw_admin_entry, admin_resource);
   }
 
+  /* Header custom behavior */
+  rest.register_x_headers(g_conf->rgw_log_http_headers);
+
   OpsLogSocket *olog = NULL;
 
   if (!g_conf->rgw_ops_log_socket_path.empty()) {
diff --git a/src/rgw/rgw_metadata.cc b/src/rgw/rgw_metadata.cc
index ea6e7932b8d0..cad08313d7fc 100644
--- a/src/rgw/rgw_metadata.cc
+++ b/src/rgw/rgw_metadata.cc
@@ -13,6 +13,8 @@
 #include "rgw_cr_rados.h"
 #include "rgw_boost_asio_yield.h"
 
+#include "include/assert.h"
+
 #define dout_subsys ceph_subsys_rgw
 
 void LogStatusDump::dump(Formatter *f) const {
diff --git a/src/rgw/rgw_period_history.cc b/src/rgw/rgw_period_history.cc
index 29926236b46a..eff0e78ad9de 100644
--- a/src/rgw/rgw_period_history.cc
+++ b/src/rgw/rgw_period_history.cc
@@ -4,6 +4,8 @@
 #include "rgw_period_history.h"
 #include "rgw_rados.h"
 
+#include "include/assert.h"
+
 #define dout_subsys ceph_subsys_rgw
 
 #undef dout_prefix
diff --git a/src/rgw/rgw_process.cc b/src/rgw/rgw_process.cc
index 6a92c49ce673..97b4d13814a7 100644
--- a/src/rgw/rgw_process.cc
+++ b/src/rgw/rgw_process.cc
@@ -81,6 +81,7 @@ int process_request(RGWRados* store, RGWREST* rest, RGWRequest* req,
     abort_early(s, NULL, -ERR_METHOD_NOT_ALLOWED, handler);
     goto done;
   }
+
   req->op = op;
   dout(10) << "op=" << typeid(*op).name() << dendl;
 
@@ -180,7 +181,7 @@ done:
     dout(0) << "ERROR: client_io->complete_request() returned " << r << dendl;
   }
   if (should_log) {
-    rgw_log_op(store, s, (op ? op->name() : "unknown"), olog);
+    rgw_log_op(store, rest, s, (op ? op->name() : "unknown"), olog);
   }
 
   int http_ret = s->err.http_ret;
diff --git a/src/rgw/rgw_rest.cc b/src/rgw/rgw_rest.cc
index 1feb8fe556b4..8d2fb4345695 100644
--- a/src/rgw/rgw_rest.cc
+++ b/src/rgw/rgw_rest.cc
@@ -4,6 +4,7 @@
 #include <errno.h>
 #include <limits.h>
 
+#include <boost/algorithm/string.hpp>
 #include "common/Formatter.h"
 #include "common/HTMLFormatter.h"
 #include "common/utf8.h"
@@ -1631,6 +1632,15 @@ RGWRESTMgr *RGWRESTMgr::get_resource_mgr(struct req_state *s, const string& uri,
   return this;
 }
 
+void RGWREST::register_x_headers(const string& s_headers)
+{
+  std::vector<std::string> hdrs = get_str_vec(s_headers);
+  for (auto& hdr : hdrs) {
+    boost::algorithm::to_upper(hdr); // XXX
+    (void) x_headers.insert(hdr);
+  }
+}
+
 RGWRESTMgr::~RGWRESTMgr()
 {
   map<string, RGWRESTMgr *>::iterator iter;
diff --git a/src/rgw/rgw_rest.h b/src/rgw/rgw_rest.h
index 96947b3c6608..ba22c1d2de3a 100644
--- a/src/rgw/rgw_rest.h
+++ b/src/rgw/rgw_rest.h
@@ -6,6 +6,9 @@
 
 #define TIME_BUF_SIZE 128
 
+#include <boost/utility/string_ref.hpp>
+#include <boost/container/flat_set.hpp>
+#include "common/sstring.hh"
 #include "common/ceph_json.h"
 #include "include/assert.h" /* needed because of common/ceph_json.h */
 #include "rgw_op.h"
@@ -370,6 +373,7 @@ public:
 
 class RGWHandler_REST : public RGWHandler {
 protected:
+
   virtual bool is_obj_update_op() { return false; }
   virtual RGWOp *op_get() { return NULL; }
   virtual RGWOp *op_put() { return NULL; }
@@ -411,6 +415,7 @@ class RGWHandler_REST_S3;
 
 class RGWRESTMgr {
   bool should_log;
+
 protected:
   map<string, RGWRESTMgr *> resource_mgrs;
   multimap<size_t, string> resources_by_size;
@@ -435,6 +440,8 @@ public:
 class RGWLibIO;
 
 class RGWREST {
+  using x_header = basic_sstring<char, uint16_t, 32>;
+  boost::container::flat_set<x_header> x_headers;
   RGWRESTMgr mgr;
 
   static int preprocess(struct req_state *s, RGWClientIO *sio);
@@ -448,6 +455,7 @@ public:
 			  RGWLibIO *io, RGWRESTMgr **pmgr,
 			  int *init_error);
 #endif
+
   void put_handler(RGWHandler_REST *handler) {
     mgr.put_handler(handler);
   }
@@ -459,9 +467,20 @@ public:
 
     mgr.register_resource(resource, m);
   }
+
   void register_default_mgr(RGWRESTMgr *m) {
     mgr.register_default_mgr(m);
   }
+
+  void register_x_headers(const std::string& headers);
+
+  bool log_x_headers(void) {
+    return (x_headers.size() > 0);
+  }
+
+  bool log_x_header(const std::string& header) {
+    return (x_headers.find(header) != x_headers.end());
+  }
 };
 
 static constexpr int64_t NO_CONTENT_LENGTH = -1;