diff --git a/ChangeLog b/ChangeLog
index d98bfce94..7d6c1923f 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -27,6 +27,7 @@
 	* almost completely changed the storage interface (for custom storage)
 	* added support for hashing pieces in multiple threads
 
+	* fixed uTP upload performance issue
 	* fix missing support for DHT put salt
 
 1.0.1 release
diff --git a/docs/dht_store.rst b/docs/dht_store.rst
index f51f90625..7deb533df 100644
--- a/docs/dht_store.rst
+++ b/docs/dht_store.rst
@@ -284,12 +284,12 @@ some additional error codes.
 +------------+-----------------------------+
 | error-code | description                 |
 +============+=============================+
-| 205        | message (i.e. ``v`` field)  |
+| 205        | message (``v`` field)       |
 |            | too big.                    |
 +------------+-----------------------------+
 | 206        | invalid signature           |
 +------------+-----------------------------+
-| 207        | salt (i.e. ``salt`` field)  |
+| 207        | salt (``salt`` field)       |
 |            | too big.                    |
 +------------+-----------------------------+
 | 301        | the CAS hash mismatched,    |
@@ -362,7 +362,7 @@ and verification of the value and sequence number should be done as follows:
 
 On the storage node, the signature MUST be verified before accepting the store
 command. The data MUST be stored under the SHA-1 hash of the public key (as it
-appears in the bencoded dict).
+appears in the bencoded dict) and the salt (if present).
 
 On the requesting nodes, the key they get back from a ``get`` request MUST be
 verified to hash to the target ID the lookup was made for, as well as verifying
diff --git a/src/utp_stream.cpp b/src/utp_stream.cpp
index ae7a579af..16ab4853e 100644
--- a/src/utp_stream.cpp
+++ b/src/utp_stream.cpp
@@ -77,7 +77,7 @@ void utp_log(char const* fmt, ...)
 {
 	mutex::scoped_lock lock(log_file_holder.utp_log_mutex);
 	static ptime start = time_now_hires();
-	fprintf(log_file_holder.utp_log_file, "[%012"PRId64"] ", total_microseconds(time_now_hires() - start));
+	fprintf(log_file_holder.utp_log_file, "[%012" PRId64 "] ", total_microseconds(time_now_hires() - start));
 	va_list l;
 	va_start(l, fmt);
 	vfprintf(log_file_holder.utp_log_file, fmt, l);
@@ -3212,7 +3212,7 @@ void utp_socket_impl::do_ledbat(int acked_bytes, int delay, int in_flight)
 
 	// true if the upper layer is pushing enough data down the socket to be
 	// limited by the cwnd. If this is not the case, we should not adjust cwnd.
-	bool cwnd_saturated = (m_bytes_in_flight + acked_bytes + m_mtu > (m_cwnd << 16));
+	bool cwnd_saturated = (m_bytes_in_flight + acked_bytes + m_mtu > (m_cwnd >> 16));
 
 	// all of these are fixed points with 16 bits fraction portion
 	boost::int64_t window_factor = (boost::int64_t(acked_bytes) << 16) / in_flight;
@@ -3221,7 +3221,9 @@ void utp_socket_impl::do_ledbat(int acked_bytes, int delay, int in_flight)
   
 	if (delay >= target_delay)
 	{
-		UTP_LOGV("%8p: off_target: %d slow_start -> 0\n", this, target_delay - delay);
+		if (m_slow_start)
+			UTP_LOGV("%8p: off_target: %d slow_start -> 0\n", this, target_delay - delay);
+
 		m_sm->inc_stats_counter(counters::utp_samples_above_target);
 		m_slow_start = false;
 	}