x86 asm: move rdrand from x86-assembly-cheat

README.adoc

@@ -12416,6 +12416,22 @@ link:userland/arch/x86_64/loop.S[LOOP]
 
 Vs <<x86-jcc-instructions,Jcc>>: https://stackoverflow.com/questions/6805692/x86-assembly-programming-loops-with-ecx-and-loop-instruction-versus-jmp-jcond Holy CISC!
 
+=== x86 random number generator instructions
+
+<<intel-manual-1>> 5.1.15 Random Number Generator Instructions
+
+Example: link:userland/arch/x86_64/rdrand.S[RDRAND]
+
+If you run that executable multiple times, it prints a random number every time to stdout.
+
+RDRAND is a true random number generator!
+
+This Intel engineer says its based on quantum effects: https://stackoverflow.com/questions/17616960/true-random-numbers-with-c11-and-rdrand/18004959#18004959
+
+Generated some polemic when kernel devs wanted to use it as part of `/dev/random`, because it could be used as a cryptographic backdoor by Intel since it is a black box.
+
+RDRAND sets the carry flag when data is ready so we must loop if the carry flag isn't set.
+
 === x86 SIMD
 
 History:

lkmc.c

@@ -56,6 +56,14 @@ void lkmc_assert_memcmp(
     }
 }
 
+void lkmc_print_hex_64(uint64_t x) {
+    printf("0x%016" PRIx64, x);
+}
+
+void lkmc_print_newline() {
+    printf("\n");
+}
+
 #if defined(__aarch64__)
 #define LKMC_SYSREG_READ_WRITE(type, name) \
     type LKMC_CONCAT(LKMC_CONCAT(LKMC_SYSREG_SYMBOL_PREFIX, name), _read(void)) { \

userland/arch/x86_64/rdrand.S

@@ -0,0 +1,11 @@
+/* https://github.com/cirosantilli/linux-kernel-module-cheat#x86-random-number-generator-instructions */
+
+#include <lkmc.h>
+
+LKMC_PROLOGUE
+1:
+    rdrand %rdi
+    jnc 1b
+    call lkmc_print_hex_64
+    call lkmc_print_newline
+LKMC_EPILOGUE