[riot-notifications] [RIOT-OS/RIOT] [WIP] Improving TinyDTLS package and dtls-echo example (#7615)
notifications at github.com
Wed Sep 20 15:04:41 CEST 2017
@rfuentess I'm on your side with the macro functions for dynamic allocations. As far as I can see the allocations we are talking about are the `hmac_context_storage`, `peer_storage`, `handshake_storage`, `security_storage` and `netq_storage`. In other words: every place Contiki uses [memb](http://contiki.sourceforge.net/docs/2.6/a01684.html) functions, which do exactly what you propose.
For the upstream repository I guess the least complex solution would be to provide one or more platform.h files containing not only constants but also platform specific function definitions, similar to micro-ecc's [platform-specific.inc](https://github.com/kmackay/micro-ecc/blob/master/platform-specific.inc), as our function signatures will not differ from contiki or *nix. However, I guess it all just depends how the tinyDTLS maintainers want to solve this.
About the PR itself: which gcc version have you been testing with? I'm getting a lot of implicit fallthrough warnings (which are enabled by default in GCC 7) building tinydtls:
/examples/dtls-echo/bin/pkg/native/tinydtls/uthash.h:449:21: error: this statement may fall through [-Werror=implicit-fallthrough=]
case 3: _hj_i += ( (unsigned)_hj_key << 16 ); \
I propose adding a `-Wno-implicit-fallthrough` to the CFLAGS or, which would be even better, fixing it upstream using [fallthrough comments](https://developers.redhat.com/blog/2017/03/10/wimplicit-fallthrough-in-gcc-7/).
While testing on native the dtls-echo server was working well, the client had to be restarted to get a second reply, but i've not yet looked into this further.
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the notifications