Line data    Source code

       1             : #pragma once
       2             : 
       3             : #include <array>
       4             : #include <cstdint>
       5             : 
       6             : // A set of tables for validating that a character is in a specific
       7             : // character set. Used to validate RFC compliance for various HTTP protocol elements.
       8             : 
       9             : namespace Envoy {
      10             : namespace Http {
      11             : 
      12       49075 : inline constexpr bool testCharInTable(const std::array<uint32_t, 8>& table, char c) {
      13             :   // CPU cache friendly version of a lookup in a bit table of size 256.
      14             :   // The table is organized as 8 32 bit words.
      15             :   // This function looks up a bit from the `table` at the index `c`.
      16             :   // This function is used to test whether a character `c` is allowed
      17             :   // or not based on the value of a bit at index `c`.
      18       49075 :   uint8_t tmp = static_cast<uint8_t>(c);
      19             :   // The `tmp >> 5` determines which of the 8 uint32_t words has the bit at index `uc`.
      20             :   // The `0x80000000 >> (tmp & 0x1f)` determines the index of the bit within the 32 bit word.
      21       49075 :   return (table[tmp >> 5] & (0x80000000 >> (tmp & 0x1f))) != 0;
      22       49075 : }
      23             : 
      24             : // Header name character table.
      25             : // From RFC 9110, https://www.rfc-editor.org/rfc/rfc9110.html#section-5.1:
      26             : //
      27             : // SPELLCHECKER(off)
      28             : // header-field   = field-name ":" OWS field-value OWS
      29             : // field-name     = token
      30             : // token          = 1*tchar
      31             : //
      32             : // tchar          = "!" / "#" / "$" / "%" / "&" / "'" / "*"
      33             : //                / "+" / "-" / "." / "^" / "_" / "`" / "|" / "~"
      34             : //                / DIGIT / ALPHA
      35             : // SPELLCHECKER(on)
      36             : inline constexpr std::array<uint32_t, 8> kGenericHeaderNameCharTable = {
      37             :     // control characters
      38             :     0b00000000000000000000000000000000,
      39             :     // !"#$%&'()*+,-./0123456789:;<=>?
      40             :     0b01011111001101101111111111000000,
      41             :     //@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_
      42             :     0b01111111111111111111111111100011,
      43             :     //`abcdefghijklmnopqrstuvwxyz{|}~
      44             :     0b11111111111111111111111111101010,
      45             :     // extended ascii
      46             :     0b00000000000000000000000000000000,
      47             :     0b00000000000000000000000000000000,
      48             :     0b00000000000000000000000000000000,
      49             :     0b00000000000000000000000000000000,
      50             : };
      51             : 
      52             : // A URI query and fragment character table. From RFC 3986:
      53             : // https://datatracker.ietf.org/doc/html/rfc3986#section-3.4
      54             : //
      55             : // SPELLCHECKER(off)
      56             : // query       = *( pchar / "/" / "?" )
      57             : // fragment    = *( pchar / "/" / "?" )
      58             : // SPELLCHECKER(on)
      59             : inline constexpr std::array<uint32_t, 8> kUriQueryAndFragmentCharTable = {
      60             :     // control characters
      61             :     0b00000000000000000000000000000000,
      62             :     // !"#$%&'()*+,-./0123456789:;<=>?
      63             :     0b01001111111111111111111111110101,
      64             :     //@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_
      65             :     0b11111111111111111111111111100001,
      66             :     //`abcdefghijklmnopqrstuvwxyz{|}~
      67             :     0b01111111111111111111111111100010,
      68             :     // extended ascii
      69             :     0b00000000000000000000000000000000,
      70             :     0b00000000000000000000000000000000,
      71             :     0b00000000000000000000000000000000,
      72             :     0b00000000000000000000000000000000,
      73             : };
      74             : 
      75             : } // namespace Http
      76             : } // namespace Envoy