[PATCH v2 2/2] Add support for GOST R 34.10-2001/-2012 signatures

Dmitry Eremin-Solenikov dbaryshkov at gmail.com
Wed Oct 16 18:13:12 CEST 2013


On Wed, Oct 16, 2013 at 6:35 PM, Werner Koch <wk at gnupg.org> wrote:
> Hi,
> Thanks for the patches.  I just pushed them.
>> +      "GOST2001-test", 256, 0,
> What is the reason that you used the "-test" suffix? Is tehre a standard
> name for this curve?  I am fine with GOSTxxxx, though.

Because they are "test" curves defined by the GOST standards. RFC 4357
e.g. names the first curve as 'id-GostR3410-2001-TestParamSet' and tells that
And adds that 'Use of the test parameter sets [...] is NOT RECOMMENDED.'

Unfortunately these parameter sets are the only parameters defined
by standard and thus they are used to verify implementation.

> p.s.
> "make check" currently often fails in keygen.  This is due to a
> problem with Ed25519 with ECDSA - I will look into that.

And strangely enough it aborts in 50% of runs. Sometimes it does, sometimes
it just outputs a note regarding testkey and and exits normally.
I failed to capture a problem either via gdb or via valgrind.

With best wishes

More information about the Gcrypt-devel mailing list