MDEV-24142 - Windows - do not use WaitOnAddress-based ssux_lock.

WaitOnAddress() turns out to be too CPU-heavy for the specific scenario,
which makes it prominent in profiler output on several benchmarks with
contended sux_lock.

The condition variable implementation does not show the same behavior.
Thus, defined SRWLOCK_DUMMY for Windows

srw_mutex should remain mapped to SRWLOCK on Windows (since SRWLOCK is
smaller).

storage/innobase/CMakeLists.txt

@@ -237,6 +237,7 @@ SET(INNOBASE_SOURCES
 	include/srv0mon.ic
 	include/srv0srv.h
 	include/srv0start.h
+	include/srw_lock.h
 	include/sux_lock.h
 	include/trx0i_s.h
 	include/trx0purge.h

storage/innobase/include/srw_lock.h

@@ -19,13 +19,13 @@ this program; if not, write to the Free Software Foundation, Inc.,
 #pragma once
 #include "univ.i"
 
-#if !(defined __linux__ || defined _WIN32 || defined __OpenBSD__)
+#if !(defined __linux__ || defined __OpenBSD__)
 # define SRW_LOCK_DUMMY
 #elif 0 // defined SAFE_MUTEX
 # define SRW_LOCK_DUMMY /* Use dummy implementation for debugging purposes */
 #endif
 
-#ifdef SRW_LOCK_DUMMY
+#if defined SRW_LOCK_DUMMY && !(defined _WIN32)
 /** An exclusive-only variant of srw_lock */
 class srw_mutex
 {