diff --git a/src/runtime/asm_386.s b/src/runtime/asm_386.s
index 1495246a25e..c401741ef90 100644
--- a/src/runtime/asm_386.s
+++ b/src/runtime/asm_386.s
@@ -479,6 +479,9 @@ TEXT runtime·atomicloaduintptr(SB), NOSPLIT, $0-8
 TEXT runtime·atomicloaduint(SB), NOSPLIT, $0-8
 	JMP	runtime·atomicload(SB)
 
+TEXT runtime·atomicstoreuintptr(SB), NOSPLIT, $0-8
+	JMP	runtime·atomicstore(SB)
+
 // bool runtime·cas64(uint64 *val, uint64 old, uint64 new)
 // Atomically:
 //	if(*val == *old){
diff --git a/src/runtime/asm_amd64.s b/src/runtime/asm_amd64.s
index 3f7f6084101..e21270d8cc4 100644
--- a/src/runtime/asm_amd64.s
+++ b/src/runtime/asm_amd64.s
@@ -491,6 +491,9 @@ TEXT runtime·atomicloaduintptr(SB), NOSPLIT, $0-16
 TEXT runtime·atomicloaduint(SB), NOSPLIT, $0-16
 	JMP	runtime·atomicload64(SB)
 
+TEXT runtime·atomicstoreuintptr(SB), NOSPLIT, $0-16
+	JMP	runtime·atomicstore64(SB)
+
 // bool casp(void **val, void *old, void *new)
 // Atomically:
 //	if(*val == old){
diff --git a/src/runtime/asm_amd64p32.s b/src/runtime/asm_amd64p32.s
index 13a16425680..c2bc91a3f5a 100644
--- a/src/runtime/asm_amd64p32.s
+++ b/src/runtime/asm_amd64p32.s
@@ -440,6 +440,9 @@ TEXT runtime·atomicloaduintptr(SB), NOSPLIT, $0-12
 TEXT runtime·atomicloaduint(SB), NOSPLIT, $0-12
 	JMP	runtime·atomicload(SB)
 
+TEXT runtime·atomicstoreuintptr(SB), NOSPLIT, $0-12
+	JMP	runtime·atomicstore(SB)
+
 // bool	runtime·cas64(uint64 *val, uint64 old, uint64 new)
 // Atomically:
 //	if(*val == *old){
diff --git a/src/runtime/asm_arm.s b/src/runtime/asm_arm.s
index 36fb022f951..a1535aeec34 100644
--- a/src/runtime/asm_arm.s
+++ b/src/runtime/asm_arm.s
@@ -724,6 +724,9 @@ TEXT runtime·atomicloaduintptr(SB),NOSPLIT,$0-8
 TEXT runtime·atomicloaduint(SB),NOSPLIT,$0-8
 	B	runtime·atomicload(SB)
 
+TEXT runtime·atomicstoreuintptr(SB),NOSPLIT,$0-8
+	B	runtime·atomicstore(SB)
+
 // AES hashing not implemented for ARM
 TEXT runtime·aeshash(SB),NOSPLIT,$-4-0
 	MOVW	$0, R0
diff --git a/src/runtime/os_windows.c b/src/runtime/os_windows.c
index 77f99062cf3..6337dde2af8 100644
--- a/src/runtime/os_windows.c
+++ b/src/runtime/os_windows.c
@@ -268,19 +268,19 @@ runtime·mpreinit(M *mp)
 void
 runtime·minit(void)
 {
-	void *thandle;
+	uintptr thandle;
 
 	// -1 = current process, -2 = current thread
 	runtime·stdcall7(runtime·DuplicateHandle, -1, -2, -1, (uintptr)&thandle, 0, 0, DUPLICATE_SAME_ACCESS);
-	runtime·atomicstorep(&g->m->thread, thandle);
+	runtime·atomicstoreuintptr(&g->m->thread, thandle);
 }
 
 // Called from dropm to undo the effect of an minit.
 void
 runtime·unminit(void)
 {
-	runtime·stdcall1(runtime·CloseHandle, (uintptr)g->m->thread);
-	g->m->thread = nil;
+	runtime·stdcall1(runtime·CloseHandle, g->m->thread);
+	g->m->thread = 0;
 }
 
 // Described in http://www.dcl.hpi.uni-potsdam.de/research/WRK/2007/08/getting-os-information-the-kuser_shared_data-structure/
@@ -532,7 +532,7 @@ void
 runtime·profileloop1(void)
 {
 	M *mp, *allm;
-	void *thread;
+	uintptr thread;
 
 	runtime·stdcall2(runtime·SetThreadPriority, -2, THREAD_PRIORITY_HIGHEST);
 
@@ -540,11 +540,11 @@ runtime·profileloop1(void)
 		runtime·stdcall2(runtime·WaitForSingleObject, (uintptr)profiletimer, -1);
 		allm = runtime·atomicloadp(&runtime·allm);
 		for(mp = allm; mp != nil; mp = mp->alllink) {
-			thread = runtime·atomicloadp(&mp->thread);
+			thread = runtime·atomicloaduintptr(&mp->thread);
 			// Do not profile threads blocked on Notes,
 			// this includes idle worker threads,
 			// idle timer thread, idle heap scavenger, etc.
-			if(thread == nil || mp->profilehz == 0 || mp->blocked)
+			if(thread == 0 || mp->profilehz == 0 || mp->blocked)
 				continue;
 			runtime·stdcall1(runtime·SuspendThread, (uintptr)thread);
 			if(mp->profilehz != 0 && !mp->blocked)
diff --git a/src/runtime/runtime.h b/src/runtime/runtime.h
index c4d87860894..27a809a07ed 100644
--- a/src/runtime/runtime.h
+++ b/src/runtime/runtime.h
@@ -371,7 +371,7 @@ struct	M
 	uintptr scalararg[4];	// scalar argument/return for mcall
 	void*   ptrarg[4];	// pointer argument/return for mcall
 #ifdef GOOS_windows
-	void*	thread;		// thread handle
+	uintptr	thread;		// thread handle
 	// these are here because they are too large to be on the stack
 	// of low-level NOSPLIT functions.
 	LibCall	libcall;
@@ -885,7 +885,9 @@ void	runtime·atomicstore(uint32 volatile*, uint32);
 void	runtime·atomicstore64(uint64 volatile*, uint64);
 uint64	runtime·atomicload64(uint64 volatile*);
 void*	runtime·atomicloadp(void* volatile*);
+uintptr	runtime·atomicloaduintptr(uintptr volatile*);
 void	runtime·atomicstorep(void* volatile*, void*);
+void	runtime·atomicstoreuintptr(uintptr volatile*, uintptr);
 void	runtime·atomicor8(byte volatile*, byte);
 
 void	runtime·setg(G*);