Select cipher suites (2.)

[boaks]

Add callback for peer specific selection of cipher suites.

[boaks]

As preparation for the implementation of RFC5746.

[boaks]

:-) the tests makefile doesn't support a common .o for the server and client.
The "fast and dirty" would be to include the .c

Any other ideas?

[boaks]

Updated to include the dtls_user_parameters_t of PR #182.

[boaks]

Uups, sorry the source formatting of the previous add_ccm got lost.
I will fix it tomorrow.

[boaks]

Polished :-).

[obgm]

A few minor comments.

I suggest the following change to tests/Makefile.in instead of including the dtls_cipher_util.c sources.

diff --git a/tests/Makefile.in b/tests/Makefile.in
index a3097d9..6b6d8bb 100644
--- a/tests/Makefile.in
+++ b/tests/Makefile.in
@@ -28,10 +28,10 @@ top_srcdir:= @top_srcdir@
 
 # files and flags
 SOURCES:= dtls-server.c ccm-test.c \
-  dtls-client.c
+  dtls-client.c dtls_ciphers_util.c
   #cbc_aes128-test.c #dsrv-test.c
 OBJECTS:= $(patsubst %.c, %.o, $(SOURCES))
-PROGRAMS:= $(patsubst %.c, %, $(SOURCES))
+PROGRAMS:=dtls-server dtls-client ccm-test
 HEADERS:=
 CFLAGS:=-Wall -std=c99 @CFLAGS@ @WARNING_CFLAGS@ $(EXTRA_CFLAGS) -D_GNU_SOURCE
 CPPFLAGS:=-I$(top_srcdir) @CPPFLAGS@
@@ -47,6 +47,8 @@ FILES:=Makefile.in $(SOURCES) ccm-testdata.c #cbc_aes128-testdata.c
 
 all:   $(PROGRAMS)
 
+dtls-client dtls-server:: dtls_ciphers_util.o
+
 check:
        echo DISTDIR: $(DISTDIR)
        echo top_builddir: $(top_builddir)

[boaks]

Currently the console lines of the client and server exceeds the 80 characters. I would prefer to fix that with the PR for RFC5746.

[obgm]

Thanks for doing the updates. Do you have an opinion on the discussion topics that you have not addressed so far?

[boaks]

Do you have an opinion on the discussion topics that you have not addressed so far?

Depends on the topics. To be frank, this PR spans now a long time. So I'm not sure, If I forgot something, if something is done in an other PR or what ever. If you like, just list or "renew" this other topics and we see, how we can deal with them.

[obgm]

Here is the list. "Don't care" would be a valid reaction....:

• Select cipher suites (2.) #148 (comment)
• Select cipher suites (2.) #148 (comment)
• Select cipher suites (2.) #148 (comment)

Typos (did not check if they are fixed and just did not show up with this comment; feel free to ignore):

• Select cipher suites (2.) #148 (comment)
• Select cipher suites (2.) #148 (comment)

[boaks]

Except the topic with using utlist/uthash instead of a own array based implementation the other topics of the list are addressed. Please have a final check, then I would go to squash and merge this PR together with PR #147 into "main".