Make improvements

- Introduce portable sched_getcpu() api
- Support GCC's __target_clones__ feature
- Make fma() go faster on x86 in default mode
- Remove some asan checks from core libraries
- WinMain() now ensures $HOME and $USER are defined

libc/runtime/runtime.h

@@ -54,7 +54,7 @@ char *getlogin(void);
 int getlogin_r(char *, size_t);
 int login_tty(int);
 int getpagesize(void);
-int syncfs(int);
+int syncfs(int) dontthrow;
 int vhangup(void);
 int getdtablesize(void);
 int sethostname(const char *, size_t);