Add support for EUI48 and EUI64 RRs

CMakeLists.txt

@@ -138,7 +138,6 @@ target_sources(zone PRIVATE
   src/log.c
   src/parser.c
   src/fallback/parser.c
-  src/generic/base16.c
   src/generic/base64.c)
 
 add_executable(zone-bench src/bench.c src/fallback/bench.c)

src/fallback/base16.h

@@ -0,0 +1,158 @@
+/*
+ * base16.h -- some useful comment
+ *
+ * Copyright (c) 2022, NLnet Labs. All rights reserved.
+ *
+ * SPDX-License-Identifier: BSD-3-Clause
+ *
+ */
+#ifndef BASE16_H
+#define BASE16_H
+
+// hexadecimal characters decode to 0x00 - 0x0f
+// delimiting characters decode to 0x80
+// non-delimiters, non-hexadecimal characters decode to 0x90
+
+// eui48 and eui64 depend on this table too
+static const uint8_t b16rmap[256] = {
+  // end-of-file (0x00)
+  0x80, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0x00 - 0x07
+  // tab (0x09), line feed (0x0a), carriage return (0x0d)
+  0x90, 0x80, 0x80, 0x90, 0x90, 0x80, 0x90, 0x90, // 0x08 - 0x0f
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0x10 - 0x17
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0x18 - 0x1f
+  // space (0x20), quote (0x22)
+  0x80, 0x90, 0x80, 0x90, 0x90, 0x90, 0x90, 0x90, // 0x20 - 0x27
+  // left paren (0x28), right paren (0x29)
+  0x80, 0x80, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0x28 - 0x2f
+  // digits
+  0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, // 0x30 - 0x37
+  // semicolon (0x3b)
+  0x08, 0x09, 0x90, 0x80, 0x90, 0x90, 0x90, 0x90, // 0x38 - 0x3f
+  // upper case
+  0x90, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, 0x90, // 0x40 - 0x47
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0x48 - 0x4f
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0x50 - 0x57
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0x58 - 0x5f
+  // lower case
+  0x90, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, 0x90, // 0x60 - 0x67
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0x68 - 0x6f
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0x70 - 0x77
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0x78 - 0x7f
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0x80 - 0x87
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0x88 - 0x8f
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0x90 - 0x97
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0x98 - 0x9f
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0xa0 - 0xa7
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0xa8 - 0xaf
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0xb0 - 0xb7
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0xb8 - 0xbf
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0xc0 - 0xc7
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0xc8 - 0xcf
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0xd0 - 0xd7
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0xd8 - 0xdf
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0xe0 - 0xe7
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0xe8 - 0xef
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0xf0 - 0xf7
+  0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, 0x90, // 0xf8 - 0xff
+};
+
+// FIXME: RFC3597 section 5 states each word of data must contain an even
+//        number of hexadecimal digits. The same is not true for DS records.
+//        RFC4043 section 5.3 merely states whitespace is allowed within the
+//        hexadecimal text. Words containing an uneven number of hexadecimal
+//        digits are impractical, but supported (BIND supports uneven
+//        sequences)
+zone_nonnull_all
+static zone_really_inline int32_t parse_base16(
+  zone_parser_t *parser,
+  const zone_type_info_t *type,
+  const zone_field_info_t *field,
+  token_t *token)
+{
+  int32_t r;
+
+  if ((r = have_contiguous(parser, type, field, token)) < 0)
+    return r;
+
+  uint8_t x0 = 0x80, x1 = 0x80;
+  uint8_t *w = &parser->rdata->octets[parser->rdata->length];
+  const uint8_t *ws = w, *we = &parser->rdata->octets[ZONE_RDATA_LIMIT];
+  const char *p;
+
+  do {
+    p = token->data;
+    if (zone_unlikely(!(x0 & 0x80))) {
+      x1 = b16rmap[(uint8_t)p[0]];
+      if (x1 & 0x80)
+        SYNTAX_ERROR(parser, "Invalid %s in %s record", NAME(field), TNAME(type));
+      w[0] = (uint8_t)((x0 << 4) | x1);
+      w += 1; p += 1;
+    }
+
+    while (w < we) {
+      x0 = b16rmap[ (uint8_t)(p[0]) ];
+      x1 = b16rmap[ (uint8_t)(p[1]) ];
+      if ((x0 | x1) & 0x80)
+        break;
+      w[0] = (uint8_t)((x0 << 4) | x1);
+      w += 1; p += 2;
+    }
+
+    if (x0 == 0x90 || (x0 != 0x80 && x1 != 0x80))
+      SYNTAX_ERROR(parser, "Invalid %s in %s record", NAME(field), TNAME(type));
+
+    lex(parser, token);
+  } while (token->code == CONTIGUOUS && w < we);
+
+  if (w - ws == 1 || w >= we || x0 != 0x80)
+    SYNTAX_ERROR(parser, "Invalid %s in %s record", NAME(field), TNAME(type));
+
+  if ((r = have_delimiter(parser, type, token)) < 0)
+    return r;
+
+  parser->rdata->length += (size_t)(w - ws);
+  return ZONE_BLOB;
+}
+
+zone_nonnull_all
+static zone_really_inline int32_t parse_salt(
+  zone_parser_t *parser,
+  const zone_type_info_t *type,
+  const zone_field_info_t *field,
+  const token_t *token)
+{
+  int32_t r;
+
+  if ((r = have_contiguous(parser, type, field, token)) < 0)
+    return r;
+
+  const char *p = token->data;
+
+  if (p[0] == '-' && !is_contiguous((uint8_t)p[1])) {
+    parser->rdata->octets[parser->rdata->length++] = 0;
+    return ZONE_STRING;
+  }
+
+  uint8_t x0 = 0x80, x1 = 0x80;
+  uint8_t *ws = &parser->rdata->octets[parser->rdata->length], *w = ws + 1;
+  const uint8_t *we = w + 255;
+
+  while (w < we) {
+    x0 = b16rmap[ (uint8_t)(p[0]) ];
+    x1 = b16rmap[ (uint8_t)(p[1]) ];
+    if ((x0 | x1) & 0x80)
+      break;
+    w[0] = (uint8_t)((x0 << 4) | x1);
+    w += 1; p += 2;
+  }
+
+  if (w - ws == 1 || w >= we || x0 != 0x80)
+    SYNTAX_ERROR(parser, "Invalid %s in %s record", NAME(field), TNAME(type));
+
+  *ws = (uint8_t)((w - ws) - 1);
+  parser->rdata->length += (size_t)(w - ws);
+  return ZONE_STRING;
+}
+
+#endif // BASE16_H

src/fallback/eui.h

@@ -0,0 +1,111 @@
+/*
+ * eui48.h -- some useful comment
+ *
+ * Copyright (c) 2023, NLnet Labs. All rights reserved.
+ *
+ * SPDX-License-Identifier: BSD-3-Clause
+ *
+ */
+#ifndef EUI_H
+#define EUI_H
+
+// RFC7043 section 3.2: xx-xx-xx-xx-xx-xx
+zone_nonnull_all
+static zone_really_inline int32_t parse_eui48(
+  zone_parser_t *parser,
+  const zone_type_info_t *type,
+  const zone_field_info_t *field,
+  const token_t *token)
+{
+  uint8_t c, x[12];
+  uint8_t *w = parser->rdata->octets + parser->rdata->length;
+  const uint8_t *p = (uint8_t *)token->data;
+
+  c = (p[ 2] == '-') & (p[ 5] == '-') &
+      (p[ 8] == '-') & (p[11] == '-') &
+      (p[14] == '-');
+  if (!c)
+    SYNTAX_ERROR(parser, "Invalid %s in %s", NAME(field), TNAME(type));
+
+  x[ 0] = b16rmap[p[ 0]];
+  x[ 1] = b16rmap[p[ 1]];
+  x[ 2] = b16rmap[p[ 3]];
+  x[ 3] = b16rmap[p[ 4]];
+  x[ 4] = b16rmap[p[ 6]];
+  x[ 5] = b16rmap[p[ 7]];
+  x[ 6] = b16rmap[p[ 9]];
+  x[ 7] = b16rmap[p[10]];
+  x[ 8] = b16rmap[p[12]];
+  x[ 9] = b16rmap[p[13]];
+  x[10] = b16rmap[p[15]];
+  x[11] = b16rmap[p[16]];
+
+  c = x[ 0] | x[ 1] | x[ 2] | x[ 3] | x[ 4] | x[ 5] |
+      x[ 6] | x[ 7] | x[ 8] | x[ 9] | x[10] | x[11];
+  if (c & 0x80)
+    SYNTAX_ERROR(parser, "Invalid %s in %s", NAME(field), TNAME(type));
+
+  w[0] = (uint8_t)((x[ 0] << 4) | x[ 1]);
+  w[1] = (uint8_t)((x[ 2] << 4) | x[ 3]);
+  w[2] = (uint8_t)((x[ 4] << 4) | x[ 5]);
+  w[3] = (uint8_t)((x[ 6] << 4) | x[ 7]);
+  w[4] = (uint8_t)((x[ 8] << 4) | x[ 9]);
+  w[5] = (uint8_t)((x[10] << 4) | x[11]);
+
+  parser->rdata->length += 6;
+  return ZONE_EUI48;
+}
+
+// RFC7043 section 4.2, require xx-xx-xx-xx-xx-xx-xx-xx
+zone_nonnull_all
+static zone_really_inline int32_t parse_eui64(
+  zone_parser_t *parser,
+  const zone_type_info_t *type,
+  const zone_field_info_t *field,
+  const token_t *token)
+{
+  uint8_t c, x[16];
+  uint8_t *w = parser->rdata->octets + parser->rdata->length;
+  const uint8_t *p = (uint8_t *)token->data;
+
+  c = (p[ 2] == '-') & (p[ 5] == '-') & (p[ 8] == '-') & (p[11] == '-') &
+      (p[14] == '-') & (p[17] == '-') & (p[20] == '-');
+  if (!c)
+    SYNTAX_ERROR(parser, "Invalid %s in %s", NAME(field), TNAME(type));
+
+  x[ 0] = b16rmap[p[ 0]];
+  x[ 1] = b16rmap[p[ 1]];
+  x[ 2] = b16rmap[p[ 3]];
+  x[ 3] = b16rmap[p[ 4]];
+  x[ 4] = b16rmap[p[ 6]];
+  x[ 5] = b16rmap[p[ 7]];
+  x[ 6] = b16rmap[p[ 9]];
+  x[ 7] = b16rmap[p[10]];
+  x[ 8] = b16rmap[p[12]];
+  x[ 9] = b16rmap[p[13]];
+  x[10] = b16rmap[p[15]];
+  x[11] = b16rmap[p[16]];
+  x[12] = b16rmap[p[18]];
+  x[13] = b16rmap[p[19]];
+  x[14] = b16rmap[p[21]];
+  x[15] = b16rmap[p[22]];
+
+  c = x[ 0] | x[ 1] | x[ 2] | x[ 3] | x[ 4] | x[ 5] | x[ 6] | x[ 7] |
+      x[ 8] | x[ 9] | x[10] | x[11] | x[12] | x[13] | x[14] | x[15];
+  if (c & 0x80)
+    SYNTAX_ERROR(parser, "Invalid %s in %s", NAME(field), TNAME(type));
+
+  w[0] = (uint8_t)((x[ 0] << 4) | x[ 1]);
+  w[1] = (uint8_t)((x[ 2] << 4) | x[ 3]);
+  w[2] = (uint8_t)((x[ 4] << 4) | x[ 5]);
+  w[3] = (uint8_t)((x[ 6] << 4) | x[ 7]);
+  w[4] = (uint8_t)((x[ 8] << 4) | x[ 9]);
+  w[5] = (uint8_t)((x[10] << 4) | x[11]);
+  w[6] = (uint8_t)((x[12] << 4) | x[13]);
+  w[7] = (uint8_t)((x[14] << 4) | x[15]);
+
+  parser->rdata->length += 8;
+  return ZONE_EUI64;
+}
+
+#endif // EUI_H

src/fallback/parser.c

@@ -19,12 +19,13 @@
 #include "fallback/ip4.h"
 #include "generic/ip6.h"
 #include "fallback/text.h"
-#include "generic/base16.h"
+#include "fallback/base16.h"
 #include "fallback/base32.h"
 #include "generic/base64.h"
 #include "generic/nsec.h"
 #include "generic/caa.h"
 #include "generic/ilnp64.h"
+#include "fallback/eui.h"
 #include "visit.h"
 #include "types.h"
 #include "fallback/type.h"