ECC enhancements

[karel-m]

This PR is related to #34

[sjaeckel]

I think the build error can be fixed in several ways:

1. by making point compression optional and disabling it by default
2. creating a new release of libtommath with the required additional algorithm, waiting until it's published
3. building libtommath from the git repositories
4. combining 1&3 and creating a special recipe with point compression enabled and build ltm from git

I would prefer going for 2 or 4

[karel-m]

Or we can introduce in libtommath something like LTM_VERSION (e.g. 0x0042 for version 0.42) and then in libtomcrypt's ltm_desc.c:

/* sqrtmod_prime */
static int sqrtmod_prime(void *a, void *b, void *c)
{
   LTC_ARGCHK(a != NULL);
   LTC_ARGCHK(b != NULL);
   LTC_ARGCHK(c != NULL);
#if !defined(LTM_VERSION) || LTM_VERSION < 0x0042
   return CRYPT_LTM_TOO_OLD; 
#else
   return mpi_to_ltc_error(mp_sqrtmod_prime(a, b, c));
#endif
}

[karel-m]

@fperrad may I ask you for running your linting machine on miko-ecc-enhancements branch? Thanks.

[fperrad]

@karel-m linting issues:

missing prototypes in tomcrypt.h : ecc_sign_hash_ex & ecc_verify_hash_ex

Info 765: external 'ecc_sign_hash_ex(const unsigned char *, unsigned long, unsigned char *, unsigned long *, union Prng_state *, int, struct ecc_key *, int)' (line 24, file src/pk/ecc/ecc_sign_hash.c) could be made static
Info 765: external 'ecc_verify_hash_ex(const unsigned char *, unsigned long, const unsigned char *, unsigned long, int *, struct ecc_key *, int)' (line 34, file src/pk/ecc/ecc_verify_hash.c) could be made static

bad indentation line 30 in src/pk/ecc/ecc_export_raw.c

bad indentation line 62 in src/pk/ecc/ecc_sign_hash.c

bad indentation line 72, 74 & 75 in src/pk/ecc/ecc_import_raw.c

[sjaeckel]

can we somehow bring the coverage up again?

[karel-m]

can we somehow bring the coverage up again?

What has to be done to bring it up?

[sjaeckel]

1. These aren't tested at all

    ecc_sign_hash_rfc7518()
    ecc_verify_hash_rfc7518()
   

2. There are no tests for all the added functionality

    pk_get_oid()
   

3. The new tag field in DER/ltc_asn1_list

4. The case a==NULL in some of the code paths that are handled specially

[karel-m]

Oh yes, tests, I know.

[coveralls]

Coverage decreased (-3.8%) to 83.635% when pulling 1caf67f on miko-ecc-enhancements into 1b81848 on develop.

[sjaeckel]

I prefer not to use a default case when I switch on an enum as the compiler then warns me if I add elements to the enum and forget to modify the switch-case

[karel-m]

@sjaeckel this PR is too complex (rebasing is quite often a pain), I am considering the idea of splitting this PR into two parts: 1/ ASN.1 changes + 2/ EC-only stuf.

The part 1/ would be a new PR (which I want to merge to develop before continuing on the remaining EC-only part) comprising of approx. ede958b + subset of 01c319d

What do you think?

[coveralls]

Coverage decreased (-3.5%) to 84.284% when pulling c7c170d on miko-ecc-enhancements into fbe7d22 on develop.

[sjaeckel]

...splitting this PR into two parts: 1/ ASN.1 changes + 2/ EC-only stuf.

Very good idea, that also makes review easier. Can the EC part probably be split even more?

[sjaeckel]

We need a ltm-1.0 package backported to trusty to fix these build failures.
I had to remove the ltm from sid in #217 as I couldn't update clang otherwise...

[karel-m]

I have split this PR and my branch miko-ecc-enhancements to:

• ECC-step2: curves y^2 = x^3 + ax + b #236 ECC enhancements part 1 (curves y^2 = x^3 + ax + b) - branch pr/ecc-non-asn1-part
• ECC-step3: extra import/export (ASN.1 related) #187 ECC enhancements part 2 (ASN.1 changes) - branch pr/ecc-asn1-part (based on pr/ecc-non-asn1-part)

So basically current miko-ecc-enhancement is the same as pr/ecc-asn1-part

There fore I am closing this PR, please move futher discussion to #236 + #187