Make fixes and improvements

- Document more compiler flags
- Expose new __print_maps() api
- Better overflow checking in mmap()
- Improve the shell example somewhat
- Fix minor runtime bugs regarding stacks
- Make kill() on fork()+execve()'d children work
- Support CLONE_CHILD_CLEARTID for proper joining
- Fix recent possible deadlock regression with --ftrace

libc/intrin/lockcmpxchgp.h

@@ -0,0 +1,24 @@
+#ifndef COSMOPOLITAN_LIBC_INTRIN_LOCKCMPXCHGP_H_
+#define COSMOPOLITAN_LIBC_INTRIN_LOCKCMPXCHGP_H_
+#include "libc/bits/asmflag.h"
+#if !(__ASSEMBLER__ + __LINKER__ + 0)
+COSMOPOLITAN_C_START_
+
+#if defined(__GNUC__) && !defined(__STRICT_ANSI__) && defined(__x86__)
+#define _lockcmpxchgp(IN_OUT_IFTHING, IN_OUT_ISEQUALTOME, IN_REPLACEITWITHME)  \
+  ({                                                                           \
+    bool DidIt;                                                                \
+    autotype(IN_OUT_IFTHING) IfThing = (IN_OUT_IFTHING);                       \
+    typeof(IfThing) IsEqualToMe = (IN_OUT_ISEQUALTOME);                        \
+    typeof(*IfThing) ReplaceItWithMe = (IN_REPLACEITWITHME);                   \
+    asm volatile(ZFLAG_ASM("lock cmpxchg\t%3,%1")                              \
+                 : ZFLAG_CONSTRAINT(DidIt), "+m"(*IfThing), "+a"(*IsEqualToMe) \
+                 : "r"(ReplaceItWithMe)                                        \
+                 : "cc");                                                      \
+    DidIt;                                                                     \
+  })
+#endif /* GNUC && !ANSI && x86 */
+
+COSMOPOLITAN_C_END_
+#endif /* !(__ASSEMBLER__ + __LINKER__ + 0) */
+#endif /* COSMOPOLITAN_LIBC_INTRIN_LOCKCMPXCHGP_H_ */