Improve locks and signals

- Introduce fast spinlock API
- Double rand64() perf w/ spinlock
- Improve raise() on New Technology
- Support gettid() across platforms
- Implement SA_NODEFER on New Technology
- Move the lock intrinsics into LIBC_INTRIN
- Make SIGTRAP recoverable on New Technology
- Block SIGCHLD in wait4() on New Technology
- Add threading prototypes for XNU and FreeBSD
- Rewrite abort() fixing its minor bugs on XNU/NT
- Shave down a lot of the content in libc/bits/bits.h
- Let signal handlers modify CPU registers on New Technology

libc/intrin/lockxchg.h

@@ -0,0 +1,29 @@
+#ifndef COSMOPOLITAN_LIBC_INTRIN_LOCKXCHG_H_
+#define COSMOPOLITAN_LIBC_INTRIN_LOCKXCHG_H_
+#if !(__ASSEMBLER__ + __LINKER__ + 0)
+COSMOPOLITAN_C_START_
+
+intptr_t lockxchg(void *, void *, size_t);
+
+#if defined(__GNUC__) && !defined(__STRICT_ANSI__)
+/**
+ * Exchanges *MEMORY into *LOCALVAR w/ one operation.
+ *
+ * @param MEMORY is uint𝑘_t[hasatleast 1] where 𝑘 ∈ {8,16,32,64}
+ * @param LOCALVAR is uint𝑘_t[hasatleast 1]
+ * @return LOCALVAR[0]
+ * @see xchg()
+ */
+#define lockxchg(MEMORY, LOCALVAR)                            \
+  ({                                                          \
+    asm("xchg\t%0,%1" : "+%m"(*(MEMORY)), "+r"(*(LOCALVAR))); \
+    *(LOCALVAR);                                              \
+  })
+#else
+#define lockxchg(MEM, VAR) \
+  lockxchg(MEM, VAR, sizeof(*(MEM)) / (sizeof(*(MEM)) == sizeof(*(VAR))))
+#endif /* __GNUC__ && !__STRICT_ANSI__ */
+
+COSMOPOLITAN_C_END_
+#endif /* !(__ASSEMBLER__ + __LINKER__ + 0) */
+#endif /* COSMOPOLITAN_LIBC_INTRIN_LOCKXCHG_H_ */