Conversation
…ind messages that include a Server Identifier option. Signed-off-by: ethan-thompson <ethan.thompson@networkradius.com>
…rmation-request messages IF the message includes a Server Identifier option AND the DUID in the option does not match the server's DUID, OR the message includes an IA option. So Server-ID is optional in this case and only if present should it be validated. Signed-off-by: ethan-thompson <ethan.thompson@networkradius.com>
… Instead, explicitly free vp before going to fail. Signed-off-by: ethan-thompson <ethan.thompson@networkradius.com>
…d to not reject them. Signed-off-by: ethan-thompson <ethan.thompson@networkradius.com>
Signed-off-by: ethan-thompson <ethan.thompson@networkradius.com>
Signed-off-by: ethan-thompson <ethan.thompson@networkradius.com>
…nteger. Signed-off-by: ethan-thompson <ethan.thompson@networkradius.com>
Signed-off-by: ethan-thompson <ethan.thompson@networkradius.com>
Signed-off-by: ethan-thompson <ethan.thompson@networkradius.com>
|
|
||
| static uint32_t instance_count = 0; | ||
| static bool instantiated = false; | ||
| static pthread_mutex_t init_mutex = PTHREAD_MUTEX_INITIALIZER; |
There was a problem hiding this comment.
What was the justification for from claude for needing to make this thread safe? I don't think this is needed.
There was a problem hiding this comment.
Alright, it didn't reference any call sites which actually needed this. Remove this commit and add a @note saying that init and deinit are only called by the main thred and do not need to be thread-safe. Ideally you'd add that to all init functions of the same type.
| append = true; | ||
| } | ||
|
|
||
| slen = decode_vsa(vp, &vp->vp_group, da, data + 4, len, decode_ctx); |
There was a problem hiding this comment.
I think the correct thing here, is to do the same as we do in the raw label, but output the raw attribute within the VSA vp we already allocated, NOT free the vp. @alandekok @ndptech do you agree?
There was a problem hiding this comment.
there's a lot of similar issues elsewhere. I'll take a look.
|
I've grabbed some of the commit manually, and dropped others. I've rewritten the "raw VSA" one, and hoisted the code out of the It still has to do a better job of decoding unknown / invalid VSAs, but that can wait for a future day. |
3 participants
Various DHCPv6 related fixes