runtime/cgo: fix compilation on Windows with CFLAGS -Wall -Werror

also fix one out-of-date comment in cmd/ld/pe.c as well.

R=golang-dev, alex.brainman
CC=golang-dev
https://golang.org/cl/6843061

src/cmd/ld/pe.c

@@ -659,7 +659,7 @@ asmbpe(void)
 	// for other threads we specify stack size in runtime explicitly
 	// (runtime knows whether cgo is enabled or not).
 	// If you change stack reserve sizes here,
-	// change them in runtime/cgo/windows_386/amd64.c as well.
+	// change STACKSIZE in runtime/cgo/gcc_windows_{386,amd64}.c as well.
 	if(!iscgo) {
 		set(SizeOfStackReserve, 0x00010000);
 		set(SizeOfStackCommit, 0x0000ffff);

src/pkg/runtime/cgo/gcc_windows_386.c

@@ -4,9 +4,10 @@
 
 #define WIN32_LEAN_AND_MEAN
 #include <windows.h>
+#include <process.h>
 #include "libcgo.h"
 
-static void *threadentry(void*);
+static void threadentry(void*);
 
 /* 1MB is default stack size for 32-bit Windows.
    Allocation granularity on Windows is typically 64 KB.
@@ -28,7 +29,7 @@ libcgo_sys_thread_start(ThreadStart *ts)
 	_beginthread(threadentry, 0, ts);
 }
 
-static void*
+static void
 threadentry(void *v)
 {
 	ThreadStart ts;
@@ -55,5 +56,4 @@ threadentry(void *v)
 	crosscall_386(ts.fn);
 	
 	LocalFree(tls0);
-	return nil;
 }

src/pkg/runtime/cgo/gcc_windows_amd64.c

@@ -4,9 +4,10 @@
 
 #define WIN64_LEAN_AND_MEAN
 #include <windows.h>
+#include <process.h>
 #include "libcgo.h"
 
-static void *threadentry(void*);
+static void threadentry(void*);
 
 /* 2MB is default stack size for 64-bit Windows.
    Allocation granularity on Windows is typically 64 KB.
@@ -28,7 +29,7 @@ libcgo_sys_thread_start(ThreadStart *ts)
 	_beginthread(threadentry, 0, ts);
 }
 
-static void*
+static void
 threadentry(void *v)
 {
 	ThreadStart ts;
@@ -53,5 +54,4 @@ threadentry(void *v)
 	);
 
 	crosscall_amd64(ts.fn);
-	return nil;
 }