3 files changed, 488 insertions, 0 deletions

diff --git a/ChangeLog b/ChangeLog
index 0d8152f00..34b6d17f9 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,8 @@
+20131009
+ - (djm) [openbsd-compat/arc4random.c openbsd-compat/chacha_private.h] Pull
+   in OpenBSD implementation of arc4random, shortly to replace the existing
+   bsd-arc4random.c
+
 20130922
  - (dtucker) [platform.c platform.h sshd.c] bz#2156: restore Linux oom_adj
    setting when handling SIGHUP to maintain behaviour over retart.  Patch
diff --git a/openbsd-compat/arc4random.c b/openbsd-compat/arc4random.c
new file mode 100644
index 000000000..356e23181
--- /dev/null
+++ b/openbsd-compat/arc4random.c
@@ -0,0 +1,261 @@
+/*      $OpenBSD: arc4random.c,v 1.25 2013/10/01 18:34:57 markus Exp $  */
+
+/*
+ * Copyright (c) 1996, David Mazieres <dm@uun.org>
+ * Copyright (c) 2008, Damien Miller <djm@openbsd.org>
+ * Copyright (c) 2013, Markus Friedl <markus@openbsd.org>
+ *
+ * Permission to use, copy, modify, and distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+/*
+ * ChaCha based random number generator for OpenBSD.
+ */
+
+#include <fcntl.h>
+#include <limits.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+#include <sys/types.h>
+#include <sys/param.h>
+#include <sys/time.h>
+#include <sys/sysctl.h>
+#include "thread_private.h"
+
+#define KEYSTREAM_ONLY
+#include "chacha_private.h"
+
+#ifdef __GNUC__
+#define inline __inline
+#else                           /* !__GNUC__ */
+#define inline
+#endif                          /* !__GNUC__ */
+
+#define KEYSZ   32
+#define IVSZ    8
+#define BLOCKSZ 64
+#define RSBUFSZ (16*BLOCKSZ)
+static int rs_initialized;
+static pid_t rs_stir_pid;
+static chacha_ctx rs;           /* chacha context for random keystream */
+static u_char rs_buf[RSBUFSZ];  /* keystream blocks */
+static size_t rs_have;          /* valid bytes at end of rs_buf */
+static size_t rs_count;         /* bytes till reseed */
+
+static inline void _rs_rekey(u_char *dat, size_t datlen);
+
+static inline void
+_rs_init(u_char *buf, size_t n)
+{
+        if (n < KEYSZ + IVSZ)
+                return;
+        chacha_keysetup(&rs, buf, KEYSZ * 8, 0);
+        chacha_ivsetup(&rs, buf + KEYSZ);
+}
+
+static void
+_rs_stir(void)
+{
+        int     mib[2];
+        size_t  len;
+        u_char rnd[KEYSZ + IVSZ];
+
+        mib[0] = CTL_KERN;
+        mib[1] = KERN_ARND;
+
+        len = sizeof(rnd);
+        sysctl(mib, 2, rnd, &len, NULL, 0);
+
+        if (!rs_initialized) {
+                rs_initialized = 1;
+                _rs_init(rnd, sizeof(rnd));
+        } else
+                _rs_rekey(rnd, sizeof(rnd));
+        memset(rnd, 0, sizeof(rnd));
+
+        /* invalidate rs_buf */
+        rs_have = 0;
+        memset(rs_buf, 0, RSBUFSZ);
+
+        rs_count = 1600000;
+}
+
+static inline void
+_rs_stir_if_needed(size_t len)
+{
+        pid_t pid = getpid();
+
+        if (rs_count <= len || !rs_initialized || rs_stir_pid != pid) {
+                rs_stir_pid = pid;
+                _rs_stir();
+        } else
+                rs_count -= len;
+}
+
+static inline void
+_rs_rekey(u_char *dat, size_t datlen)
+{
+#ifndef KEYSTREAM_ONLY
+        memset(rs_buf, 0,RSBUFSZ);
+#endif
+        /* fill rs_buf with the keystream */
+        chacha_encrypt_bytes(&rs, rs_buf, rs_buf, RSBUFSZ);
+        /* mix in optional user provided data */
+        if (dat) {
+                size_t i, m;
+
+                m = MIN(datlen, KEYSZ + IVSZ);
+                for (i = 0; i < m; i++)
+                        rs_buf[i] ^= dat[i];
+        }
+        /* immediately reinit for backtracking resistance */
+        _rs_init(rs_buf, KEYSZ + IVSZ);
+        memset(rs_buf, 0, KEYSZ + IVSZ);
+        rs_have = RSBUFSZ - KEYSZ - IVSZ;
+}
+
+static inline void
+_rs_random_buf(void *_buf, size_t n)
+{
+        u_char *buf = (u_char *)_buf;
+        size_t m;
+
+        _rs_stir_if_needed(n);
+        while (n > 0) {
+                if (rs_have > 0) {
+                        m = MIN(n, rs_have);
+                        memcpy(buf, rs_buf + RSBUFSZ - rs_have, m);
+                        memset(rs_buf + RSBUFSZ - rs_have, 0, m);
+                        buf += m;
+                        n -= m;
+                        rs_have -= m;
+                }
+                if (rs_have == 0)
+                        _rs_rekey(NULL, 0);
+        }
+}
+
+static inline void
+_rs_random_u32(u_int32_t *val)
+{
+        _rs_stir_if_needed(sizeof(*val));
+        if (rs_have < sizeof(*val))
+                _rs_rekey(NULL, 0);
+        memcpy(val, rs_buf + RSBUFSZ - rs_have, sizeof(*val));
+        memset(rs_buf + RSBUFSZ - rs_have, 0, sizeof(*val));
+        rs_have -= sizeof(*val);
+        return;
+}
+
+void
+arc4random_stir(void)
+{
+        _ARC4_LOCK();
+        _rs_stir();
+        _ARC4_UNLOCK();
+}
+
+void
+arc4random_addrandom(u_char *dat, int datlen)
+{
+        int m;
+
+        _ARC4_LOCK();
+        if (!rs_initialized)
+                _rs_stir();
+        while (datlen > 0) {
+                m = MIN(datlen, KEYSZ + IVSZ);
+                _rs_rekey(dat, m);
+                dat += m;
+                datlen -= m;
+        }
+        _ARC4_UNLOCK();
+}
+
+u_int32_t
+arc4random(void)
+{
+        u_int32_t val;
+
+        _ARC4_LOCK();
+        _rs_random_u32(&val);
+        _ARC4_UNLOCK();
+        return val;
+}
+
+void
+arc4random_buf(void *buf, size_t n)
+{
+        _ARC4_LOCK();
+        _rs_random_buf(buf, n);
+        _ARC4_UNLOCK();
+}
+
+/*
+ * Calculate a uniformly distributed random number less than upper_bound
+ * avoiding "modulo bias".
+ *
+ * Uniformity is achieved by generating new random numbers until the one
+ * returned is outside the range [0, 2**32 % upper_bound).  This
+ * guarantees the selected random number will be inside
+ * [2**32 % upper_bound, 2**32) which maps back to [0, upper_bound)
+ * after reduction modulo upper_bound.
+ */
+u_int32_t
+arc4random_uniform(u_int32_t upper_bound)
+{
+        u_int32_t r, min;
+
+        if (upper_bound < 2)
+                return 0;
+
+        /* 2**32 % x == (2**32 - x) % x */
+        min = -upper_bound % upper_bound;
+
+        /*
+         * This could theoretically loop forever but each retry has
+         * p > 0.5 (worst case, usually far better) of selecting a
+         * number inside the range we need, so it should rarely need
+         * to re-roll.
+         */
+        for (;;) {
+                r = arc4random();
+                if (r >= min)
+                        break;
+        }
+
+        return r % upper_bound;
+}
+
+#if 0
+/*-------- Test code for i386 --------*/
+#include <stdio.h>
+#include <machine/pctr.h>
+int
+main(int argc, char **argv)
+{
+        const int iter = 1000000;
+        int     i;
+        pctrval v;
+
+        v = rdtsc();
+        for (i = 0; i < iter; i++)
+                arc4random();
+        v = rdtsc() - v;
+        v /= iter;
+
+        printf("%qd cycles\n", v);
+        exit(0);
+}
+#endif
diff --git a/openbsd-compat/chacha_private.h b/openbsd-compat/chacha_private.h
new file mode 100644
index 000000000..7c3680fa6
--- /dev/null
+++ b/openbsd-compat/chacha_private.h
@@ -0,0 +1,222 @@
+/*
+chacha-merged.c version 20080118
+D. J. Bernstein
+Public domain.
+*/
+
+/* $OpenBSD: chacha_private.h,v 1.2 2013/10/04 07:02:27 djm Exp $ */
+
+typedef unsigned char u8;
+typedef unsigned int u32;
+
+typedef struct
+{
+  u32 input[16]; /* could be compressed */
+} chacha_ctx;
+
+#define U8C(v) (v##U)
+#define U32C(v) (v##U)
+
+#define U8V(v) ((u8)(v) & U8C(0xFF))
+#define U32V(v) ((u32)(v) & U32C(0xFFFFFFFF))
+
+#define ROTL32(v, n) \
+  (U32V((v) << (n)) | ((v) >> (32 - (n))))
+
+#define U8TO32_LITTLE(p) \
+  (((u32)((p)[0])      ) | \
+   ((u32)((p)[1]) <<  8) | \
+   ((u32)((p)[2]) << 16) | \
+   ((u32)((p)[3]) << 24))
+
+#define U32TO8_LITTLE(p, v) \
+  do { \
+    (p)[0] = U8V((v)      ); \
+    (p)[1] = U8V((v) >>  8); \
+    (p)[2] = U8V((v) >> 16); \
+    (p)[3] = U8V((v) >> 24); \
+  } while (0)
+
+#define ROTATE(v,c) (ROTL32(v,c))
+#define XOR(v,w) ((v) ^ (w))
+#define PLUS(v,w) (U32V((v) + (w)))
+#define PLUSONE(v) (PLUS((v),1))
+
+#define QUARTERROUND(a,b,c,d) \
+  a = PLUS(a,b); d = ROTATE(XOR(d,a),16); \
+  c = PLUS(c,d); b = ROTATE(XOR(b,c),12); \
+  a = PLUS(a,b); d = ROTATE(XOR(d,a), 8); \
+  c = PLUS(c,d); b = ROTATE(XOR(b,c), 7);
+
+static const char sigma[16] = "expand 32-byte k";
+static const char tau[16] = "expand 16-byte k";
+
+static void
+chacha_keysetup(chacha_ctx *x,const u8 *k,u32 kbits,u32 ivbits)
+{
+  const char *constants;
+
+  x->input[4] = U8TO32_LITTLE(k + 0);
+  x->input[5] = U8TO32_LITTLE(k + 4);
+  x->input[6] = U8TO32_LITTLE(k + 8);
+  x->input[7] = U8TO32_LITTLE(k + 12);
+  if (kbits == 256) { /* recommended */
+    k += 16;
+    constants = sigma;
+  } else { /* kbits == 128 */
+    constants = tau;
+  }
+  x->input[8] = U8TO32_LITTLE(k + 0);
+  x->input[9] = U8TO32_LITTLE(k + 4);
+  x->input[10] = U8TO32_LITTLE(k + 8);
+  x->input[11] = U8TO32_LITTLE(k + 12);
+  x->input[0] = U8TO32_LITTLE(constants + 0);
+  x->input[1] = U8TO32_LITTLE(constants + 4);
+  x->input[2] = U8TO32_LITTLE(constants + 8);
+  x->input[3] = U8TO32_LITTLE(constants + 12);
+}
+
+static void
+chacha_ivsetup(chacha_ctx *x,const u8 *iv)
+{
+  x->input[12] = 0;
+  x->input[13] = 0;
+  x->input[14] = U8TO32_LITTLE(iv + 0);
+  x->input[15] = U8TO32_LITTLE(iv + 4);
+}
+
+static void
+chacha_encrypt_bytes(chacha_ctx *x,const u8 *m,u8 *c,u32 bytes)
+{
+  u32 x0, x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, x11, x12, x13, x14, x15;
+  u32 j0, j1, j2, j3, j4, j5, j6, j7, j8, j9, j10, j11, j12, j13, j14, j15;
+  u8 *ctarget = NULL;
+  u8 tmp[64];
+  u_int i;
+
+  if (!bytes) return;
+
+  j0 = x->input[0];
+  j1 = x->input[1];
+  j2 = x->input[2];
+  j3 = x->input[3];
+  j4 = x->input[4];
+  j5 = x->input[5];
+  j6 = x->input[6];
+  j7 = x->input[7];
+  j8 = x->input[8];
+  j9 = x->input[9];
+  j10 = x->input[10];
+  j11 = x->input[11];
+  j12 = x->input[12];
+  j13 = x->input[13];
+  j14 = x->input[14];
+  j15 = x->input[15];
+
+  for (;;) {
+    if (bytes < 64) {
+      for (i = 0;i < bytes;++i) tmp[i] = m[i];
+      m = tmp;
+      ctarget = c;
+      c = tmp;
+    }
+    x0 = j0;
+    x1 = j1;
+    x2 = j2;
+    x3 = j3;
+    x4 = j4;
+    x5 = j5;
+    x6 = j6;
+    x7 = j7;
+    x8 = j8;
+    x9 = j9;
+    x10 = j10;
+    x11 = j11;
+    x12 = j12;
+    x13 = j13;
+    x14 = j14;
+    x15 = j15;
+    for (i = 20;i > 0;i -= 2) {
+      QUARTERROUND( x0, x4, x8,x12)
+      QUARTERROUND( x1, x5, x9,x13)
+      QUARTERROUND( x2, x6,x10,x14)
+      QUARTERROUND( x3, x7,x11,x15)
+      QUARTERROUND( x0, x5,x10,x15)
+      QUARTERROUND( x1, x6,x11,x12)
+      QUARTERROUND( x2, x7, x8,x13)
+      QUARTERROUND( x3, x4, x9,x14)
+    }
+    x0 = PLUS(x0,j0);
+    x1 = PLUS(x1,j1);
+    x2 = PLUS(x2,j2);
+    x3 = PLUS(x3,j3);
+    x4 = PLUS(x4,j4);
+    x5 = PLUS(x5,j5);
+    x6 = PLUS(x6,j6);
+    x7 = PLUS(x7,j7);
+    x8 = PLUS(x8,j8);
+    x9 = PLUS(x9,j9);
+    x10 = PLUS(x10,j10);
+    x11 = PLUS(x11,j11);
+    x12 = PLUS(x12,j12);
+    x13 = PLUS(x13,j13);
+    x14 = PLUS(x14,j14);
+    x15 = PLUS(x15,j15);
+
+#ifndef KEYSTREAM_ONLY
+    x0 = XOR(x0,U8TO32_LITTLE(m + 0));
+    x1 = XOR(x1,U8TO32_LITTLE(m + 4));
+    x2 = XOR(x2,U8TO32_LITTLE(m + 8));
+    x3 = XOR(x3,U8TO32_LITTLE(m + 12));
+    x4 = XOR(x4,U8TO32_LITTLE(m + 16));
+    x5 = XOR(x5,U8TO32_LITTLE(m + 20));
+    x6 = XOR(x6,U8TO32_LITTLE(m + 24));
+    x7 = XOR(x7,U8TO32_LITTLE(m + 28));
+    x8 = XOR(x8,U8TO32_LITTLE(m + 32));
+    x9 = XOR(x9,U8TO32_LITTLE(m + 36));
+    x10 = XOR(x10,U8TO32_LITTLE(m + 40));
+    x11 = XOR(x11,U8TO32_LITTLE(m + 44));
+    x12 = XOR(x12,U8TO32_LITTLE(m + 48));
+    x13 = XOR(x13,U8TO32_LITTLE(m + 52));
+    x14 = XOR(x14,U8TO32_LITTLE(m + 56));
+    x15 = XOR(x15,U8TO32_LITTLE(m + 60));
+#endif
+
+    j12 = PLUSONE(j12);
+    if (!j12) {
+      j13 = PLUSONE(j13);
+      /* stopping at 2^70 bytes per nonce is user's responsibility */
+    }
+
+    U32TO8_LITTLE(c + 0,x0);
+    U32TO8_LITTLE(c + 4,x1);
+    U32TO8_LITTLE(c + 8,x2);
+    U32TO8_LITTLE(c + 12,x3);
+    U32TO8_LITTLE(c + 16,x4);
+    U32TO8_LITTLE(c + 20,x5);
+    U32TO8_LITTLE(c + 24,x6);
+    U32TO8_LITTLE(c + 28,x7);
+    U32TO8_LITTLE(c + 32,x8);
+    U32TO8_LITTLE(c + 36,x9);
+    U32TO8_LITTLE(c + 40,x10);
+    U32TO8_LITTLE(c + 44,x11);
+    U32TO8_LITTLE(c + 48,x12);
+    U32TO8_LITTLE(c + 52,x13);
+    U32TO8_LITTLE(c + 56,x14);
+    U32TO8_LITTLE(c + 60,x15);
+
+    if (bytes <= 64) {
+      if (bytes < 64) {
+        for (i = 0;i < bytes;++i) ctarget[i] = c[i];
+      }
+      x->input[12] = j12;
+      x->input[13] = j13;
+      return;
+    }
+    bytes -= 64;
+    c += 64;
+#ifndef KEYSTREAM_ONLY
+    m += 64;
+#endif
+  }
+}