Return to BigIntConfig.in CVS log

Up to [ELWIX - Embedded LightWeight unIX -] / embedaddon / axTLS / ssl

axTLS

    1: #
    2: # For a description of the syntax of this configuration file,
    3: # see scripts/config/Kconfig-language.txt
    4: #
    5: 
    6: menu "BigInt Options"
    7:     depends on !CONFIG_SSL_SKELETON_MODE
    8: 
    9: choice
   10:     prompt "Reduction Algorithm"
   11:     default CONFIG_BIGINT_BARRETT
   12: 
   13: config CONFIG_BIGINT_CLASSICAL
   14:     bool "Classical"
   15:     help
   16:         Classical uses standard division. It has no limitations and is
   17:         theoretically the slowest due to the divisions used. For this particular
   18:         implementation it is surprisingly quite fast.
   19: 
   20: config CONFIG_BIGINT_MONTGOMERY
   21:     bool "Montgomery"
   22:     help
   23:         Montgomery uses simple addition and multiplication to achieve its
   24:         performance.  It has the limitation that 0 <= x, y < m, and so is not 
   25:         used when CRT is active.
   26: 
   27:         This option will not be normally selected.
   28: 
   29: config CONFIG_BIGINT_BARRETT
   30:     bool "Barrett"
   31:     help
   32:         Barrett performs expensive precomputation before reduction and partial
   33:         multiplies for computational speed.
   34: 
   35:         It is about 40% faster than Classical/Montgomery with the expense of
   36:         about 2kB, and so this option is normally selected.
   37: 
   38: endchoice
   39: 
   40: config CONFIG_BIGINT_CRT
   41:     bool "Chinese Remainder Theorem (CRT)"
   42:     default y
   43:     help 
   44:         Allow the Chinese Remainder Theorem (CRT) to be used.
   45: 
   46:         Uses a number of extra coefficients from the private key to improve the
   47:         performance of a decryption. This feature is one of the most 
   48:         significant performance improvements (it reduces a decryption time by 
   49:         over 3 times).
   50: 
   51:         This option should be selected.
   52: 
   53: config CONFIG_BIGINT_KARATSUBA
   54:     bool "Karatsuba Multiplication"
   55:     default n
   56:     help
   57:         Allow Karasuba multiplication to be used.
   58:  
   59:         Uses 3 multiplications (plus a number of additions/subtractions) 
   60:         instead of 4. Multiplications are O(N^2) but addition/subtraction 
   61:         is O(N) hence for large numbers is beneficial. For this project, the 
   62:         effect was only useful for 4096 bit keys (for 32 bit processors). For
   63:         8 bit processors this option might be a possibility.
   64:         
   65:         It costs about 2kB to enable it.
   66: 
   67: config MUL_KARATSUBA_THRESH
   68:     int "Karatsuba Multiplication Theshold"
   69:     default 20
   70:     depends on CONFIG_BIGINT_KARATSUBA
   71:     help
   72:         The minimum number of components needed before Karasuba muliplication
   73:         is used.
   74:  
   75:         This is very dependent on the speed/implementation of bi_add()/
   76:         bi_subtract(). There is a bit of trial and error here and will be
   77:         at a different point for different architectures.
   78: 
   79: config SQU_KARATSUBA_THRESH
   80:     int "Karatsuba Square Threshold"
   81:     default 40
   82:     depends on CONFIG_BIGINT_KARATSUBA && CONFIG_BIGINT_SQUARE
   83:     help    
   84:         The minimum number of components needed before Karatsuba squaring
   85:         is used.
   86:  
   87:         This is very dependent on the speed/implementation of bi_add()/
   88:         bi_subtract(). There is a bit of trial and error here and will be
   89:         at a different point for different architectures.
   90: 
   91: config CONFIG_BIGINT_SLIDING_WINDOW
   92:     bool "Sliding Window Exponentiation"
   93:     default y
   94:     help
   95:         Allow Sliding-Window Exponentiation to be used.
   96:  
   97:         Potentially processes more than 1 bit at a time when doing 
   98:         exponentiation. The sliding-window technique reduces the number of 
   99:         precomputations compared to other precomputed techniques.
  100: 
  101:         It results in a considerable performance improvement with it enabled
  102:         (it halves the decryption time) and so should be selected.
  103: 
  104: config CONFIG_BIGINT_SQUARE
  105:     bool "Square Algorithm"
  106:     default y
  107:     help
  108:         Allow squaring to be used instead of a multiplication. It uses
  109:         1/2 of the standard multiplies to obtain its performance.  
  110:         It gives a 20% speed improvement overall and so should be selected.
  111: 
  112: config CONFIG_BIGINT_CHECK_ON
  113:     bool "BigInt Integrity Checking"
  114:     default n if !CONFIG_DEBUG
  115:     default y if CONFIG_DEBUG
  116:     help
  117:         This is used when developing bigint algorithms. It performs a sanity
  118:         check on all operations at the expense of speed.
  119:         
  120:         This option is only selected when developing and should normally be
  121:         turned off.
  122: 
  123: choice
  124:     prompt "Integer Size"
  125:     default CONFIG_INTEGER_32BIT
  126: 
  127: config CONFIG_INTEGER_32BIT
  128:     bool "32"
  129:     help
  130:         The native integer size is 32 bits or higher.
  131:         
  132: 
  133: config CONFIG_INTEGER_16BIT
  134:     bool "16"
  135:     help
  136:         The native integer size is 16 bits.
  137: 
  138: config CONFIG_INTEGER_8BIT
  139:     bool "8"
  140:     help
  141:         The native integer size is 8 bits.
  142: 
  143: endchoice
  144: endmenu
  145: