Change wording of warnings
This commit is contained in:
parent
f240ea0b50
commit
f9734b35b5
2 changed files with 25 additions and 27 deletions
|
@ -25,22 +25,20 @@
|
|||
* of non-safe primes both decreases the difficulty of the underlying
|
||||
* discrete logarithm problem and can lead to small subgroup attacks
|
||||
* leaking private exponent bits when invalid public keys are used
|
||||
* and not detected. This is especially relevant if the same DHM parameters
|
||||
* are reused for multiple key exchanges as in static DHM, while the
|
||||
* criticality of small-subgroup attacks is lower for ephemeral DHM.
|
||||
* and not detected. This is especially relevant if the same DHM
|
||||
* parameters are reused for multiple key exchanges as in static DHM,
|
||||
* while the criticality of small-subgroup attacks is lower for
|
||||
* ephemeral DHM.
|
||||
*
|
||||
* For performance reasons, the code does neither perform primality
|
||||
* nor safe primality tests, nor the expensive checks for invalid
|
||||
* subgroups.
|
||||
* subgroups. Moreover, even if these were performed, non-standardized
|
||||
* primes cannot be trusted because of the possibility of backdoors
|
||||
* that can't be effectively checked for.
|
||||
*
|
||||
* The possibility for the use of custom, non-safe primes in DHM
|
||||
* is a deficiency in the TLS protocol that has been adressed only
|
||||
* recently through the addition of the named group extension from
|
||||
* RFC 7919, which however is not yet implemented in Mbed TLS.
|
||||
*
|
||||
* If possible, we recommend to use elliptic curve based key
|
||||
* exchanges instead of DHM-based ones, because the former only
|
||||
* accepts standardized groups.
|
||||
* We therefore consider DHE a security risk. If possible, it is
|
||||
* recommended users should consider preferring other methods of
|
||||
* key exchange.
|
||||
*
|
||||
*/
|
||||
#ifndef MBEDTLS_DHM_H
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue