e3d0dae7cf
The existing implementation of DTLSv1_listen() is fundamentally flawed. This function is used in DTLS solutions to listen for new incoming connections from DTLS clients. A client will send an initial ClientHello. The server will respond with a HelloVerifyRequest containing a unique cookie. The client the responds with a second ClientHello - which this time contains the cookie. Once the cookie has been verified then DTLSv1_listen() returns to user code, which is typically expected to continue the handshake with a call to (for example) SSL_accept(). Whilst listening for incoming ClientHellos, the underlying BIO is usually in an unconnected state. Therefore ClientHellos can come in from *any* peer. The arrival of the first ClientHello without the cookie, and the second one with it, could be interspersed with other intervening messages from different clients. The whole purpose of this mechanism is as a defence against DoS attacks. The idea is to avoid allocating state on the server until the client has verified that it is capable of receiving messages at the address it claims to come from. However the existing DTLSv1_listen() implementation completely fails to do this. It attempts to super-impose itself on the standard state machine and reuses all of this code. However the standard state machine expects to operate in a stateful manner with a single client, and this can cause various problems. A second more minor issue is that the return codes from this function are quite confused, with no distinction made between fatal and non-fatal errors. Most user code treats all errors as non-fatal, and simply retries the call to DTLSv1_listen(). This commit completely rewrites the implementation of DTLSv1_listen() and provides a stand alone implementation that does not rely on the existing state machine. It also provides more consistent return codes. Reviewed-by: Andy Polyakov <appro@openssl.org> |
||
|---|---|---|
| .. | ||
| dtls1_bitmap.c | ||
| README | ||
| rec_layer_d1.c | ||
| rec_layer_s3.c | ||
| record.h | ||
| record_locl.h | ||
| ssl3_buffer.c | ||
| ssl3_record.c | ||
Record Layer Design
===================
This file provides some guidance on the thinking behind the design of the
record layer code to aid future maintenance.
The record layer is divided into a number of components. At the time of writing
there are four: SSL3_RECORD, SSL3_BUFFER, DLTS1_BITMAP and RECORD_LAYER. Each
of these components is defined by:
1) A struct definition of the same name as the component
2) A set of source files that define the functions for that component
3) A set of accessor macros
All struct definitions are in record.h. The functions and macros are either
defined in record.h or record_locl.h dependent on whether they are intended to
be private to the record layer, or whether they form part of the API to the rest
of libssl.
The source files map to components as follows:
dtls1_bitmap.c -> DTLS1_BITMAP component
ssl3_buffer.c -> SSL3_BUFFER component
ssl3_record.c -> SSL3_RECORD component
rec_layer_s3.c, rec_layer_d1.c -> RECORD_LAYER component
The RECORD_LAYER component is a facade pattern, i.e. it provides a simplified
interface to the record layer for the rest of libssl. The other 3 components are
entirely private to the record layer and therefore should never be accessed
directly by libssl.
Any component can directly access its own members - they are private to that
component, e.g. ssl3_buffer.c can access members of the SSL3_BUFFER struct
without using a macro. No component can directly access the members of another
component, e.g. ssl3_buffer cannot reach inside the RECORD_LAYER component to
directly access its members. Instead components use accessor macros, so if code
in ssl3_buffer.c wants to access the members of the RECORD_LAYER it uses the
RECORD_LAYER_* macros.
Conceptually it looks like this:
libssl
|
---------------------------|-----record.h--------------------------------------
|
_______V______________
| |
| RECORD_LAYER |
| |
| rec_layer_s3.c |
| ^ |
| _________|__________ |
|| ||
|| DTLS1_RECORD_LAYER ||
|| ||
|| rec_layer_d1.c ||
||____________________||
|______________________|
record_locl.h ^ ^ ^
_________________| | |_________________
| | |
_____V_________ ______V________ _______V________
| | | | | |
| SSL3_BUFFER | | SSL3_RECORD | | DTLS1_BITMAP |
| |--->| | | |
| ssl3_buffer.c | | ssl3_record.c | | dtls1_bitmap.c |
|_______________| |_______________| |________________|
The two RECORD_LAYER source files build on each other, i.e.
the main one is rec_layer_s3.c which provides the core SSL/TLS layer. The second
one is rec_layer_d1.c which builds off of the SSL/TLS code to provide DTLS
specific capabilities. It uses some DTLS specific RECORD_LAYER component members
which should only be accessed from rec_layer_d1.c. These are held in the
DTLS1_RECORD_LAYER struct.