Make numerous improvements

- Python static hello world now 1.8mb
- Python static fully loaded now 10mb
- Python HTTPS client now uses MbedTLS
- Python REPL now completes import stmts
- Increase stack size for Python for now
- Begin synthesizing posixpath and ntpath
- Restore Python \N{UNICODE NAME} support
- Restore Python NFKD symbol normalization
- Add optimized code path for Intel SHA-NI
- Get more Python unit tests passing faster
- Get Python help() pagination working on NT
- Python hashlib now supports MbedTLS PBKDF2
- Make memcpy/memmove/memcmp/bcmp/etc. faster
- Add Mersenne Twister and Vigna to LIBC_RAND
- Provide privileged __printf() for error code
- Fix zipos opendir() so that it reports ENOTDIR
- Add basic chmod() implementation for Windows NT
- Add Cosmo's best functions to Python cosmo module
- Pin function trace indent depth to that of caller
- Show memory diagram on invalid access in MODE=dbg
- Differentiate stack overflow on crash in MODE=dbg
- Add stb_truetype and tools for analyzing font files
- Upgrade to UNICODE 13 and reduce its binary footprint
- COMPILE.COM now logs resource usage of build commands
- Start implementing basic poll() support on bare metal
- Set getauxval(AT_EXECFN) to GetModuleFileName() on NT
- Add descriptions to strerror() in non-TINY build modes
- Add COUNTBRANCH() macro to help with micro-optimizations
- Make error / backtrace / asan / memory code more unbreakable
- Add fast perfect C implementation of μ-Law and a-Law audio codecs
- Make strtol() functions consistent with other libc implementations
- Improve Linenoise implementation (see also github.com/jart/bestline)
- COMPILE.COM now suppresses stdout/stderr of successful build commands

libc/rand/rdrand.c

@@ -16,8 +16,62 @@
 │ TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR             │
 │ PERFORMANCE OF THIS SOFTWARE.                                                │
 ╚─────────────────────────────────────────────────────────────────────────────*/
+#include "libc/bits/bits.h"
+#include "libc/errno.h"
+#include "libc/nexgen32e/x86feature.h"
 #include "libc/rand/rand.h"
+#include "libc/sysv/consts/grnd.h"
 
-uint64_t rdrand(void) {
-  return rand64();
+STATIC_YOINK("rdrand_init");
+
+static noinline uint64_t rdrand_failover(void) {
+  int f;
+  size_t i;
+  ssize_t r;
+  volatile uint64_t b;
+  register uint64_t x;
+  for (f = GRND_RANDOM | GRND_NONBLOCK, i = 0; i < 8; i += r) {
+    if ((r = getrandom((char *)&b + i, 8 - i, f)) <= 0) {
+      if (r == -1 && errno == EINTR) {
+        r = 0;
+      } else if (r == -1 && errno == EAGAIN) {
+        f = 0;
+      } else {
+        return rand64();
+      }
+    }
+  }
+  x = b;
+  b = 0;
+  return x;
+}
+
+/**
+ * Retrieves 64-bits of hardware random data from RDRAND instruction.
+ *
+ * If RDRAND isn't available (we check CPUID and we also disable it
+ * automatically for microarchitectures where it's slow or buggy) then
+ * we try getrandom(), RtlGenRandom(), or sysctl(KERN_ARND). If those
+ * aren't available then we try /dev/urandom and if that fails, we try
+ * getauxval(AT_RANDOM), and if not we finally use RDTSC and getpid().
+ *
+ * This function takes between 10 nanoseconds to several microseconds.
+ *
+ * @see rngset(), rdseed(), rand64()
+ */
+uint64_t rdrand(void) {
+  int i;
+  char cf;
+  uint64_t x;
+  if (X86_HAVE(RDRND)) {
+    for (i = 0; i < 10; ++i) {
+      asm volatile(CFLAG_ASM("rdrand\t%1")
+                   : CFLAG_CONSTRAINT(cf), "=r"(x)
+                   : /* no inputs */
+                   : "cc");
+      if (cf) return x;
+      asm volatile("pause");
+    }
+  }
+  return rdrand_failover();
 }