Kernel/Net: Don't use TimePrecision::Precise when delaying TCP ACKs

If the peer is able to wait for ~500ms before receiving the ACK, it will
survive waiting the few more milliseconds that come with the imprecision
of the Coarse setting.

On x86_64, during a SFTP file transfer, the time spent on fetching the
precise clock can be as high as 10% of the total runtime.

Kernel/Net/TCPSocket.cpp

@@ -396,7 +396,7 @@ bool TCPSocket::should_delay_next_ack() const
         return false;
 
     // RFC 1122 says we should not delay ACKs for more than 500 milliseconds.
-    if (TimeManagement::the().monotonic_time(TimePrecision::Precise) >= m_last_ack_sent_time + Duration::from_milliseconds(500))
+    if (TimeManagement::the().monotonic_time() >= m_last_ack_sent_time + Duration::from_milliseconds(500))
         return false;
 
     return true;