SM2/SM3/SM4 Support

[pedroalbanese]

Hi,

I read that LibreSSL supports SM2/SM3/SM4, but SM2 is not on the ecparam -list_curves. How to create a SM2 private key?

Thanks in advance.

[botovq]

On Mon, Oct 26, 2020 at 06:19:49AM -0700, pedroalbanese wrote: I read that LibreSSL supports SM2/SM3/SM4, but SM2 is not on the ecparam -list_curves. How to create an SM2 private key?

SM3 and SM4 are supported. There is an open pull request for SM2 which I worked on and tried to get in several times but has not yet received the necessary go-ahead. I can't give you an ETA.

[ronaldtse]

@botovq I just realised that this issue exist! Is there anything we can help with the original #513 ticket? (cc @ni4)

[ni4]

...as well as libressl/openbsd#105. At least those should be updated due to merge conflicts, I can handle that once finished with some other things.

[botovq]

@ronaldtse, @ni4: There is no need for you to do anything at this point. I have a rebased version of opensbsd/pull/105 that I should be able to start landing in steps soon. Thus, SM2 should end up being available in the next release (October).

[ronaldtse]

Thank you @botovq for the great news! Very happy to see it finally merged. It was a long standing PR which would have been 2.5 years by October 😄

[pedroalbanese]

Sorry fe bad English. I can't understood. Is it under development or is already implemented?

Version 3.3.3 do not list it.

Thanks.

[botovq]

On Sun, Aug 01, 2021 at 02:57:15PM -0700, Pedro Albanese wrote: Sorry fe bad English. I can't understood. Is it under development or is already implemented?

The code exists but is not yet part of LibreSSL.

Version 3.3.3 do not list it.

Yes, it is not yet there. It will almost surely be part of the next stable release 3.4.x.

[orbea]

@botovq Was SM2 support ever added? Is this issue outdated?

[botovq]

@orbea It was added to the tree but never enabled due to some code quality issues. I don't remember the details. While I would still like to have support, it will be some work to get there and it never landed high enough on our list of priorities. The situation for SM* is a bit similar to GOST: they are mandatory for a large number of people, but the existing implementations are not ideal.

In any case, the issue is still open and correctly so.

[ronaldtse]

Thanks @botovq for the update. In any case, we are still around to help with any remaining issues with our contributed code. We look forward having this work completed, which began all the way back from 2019!