Add code points for PADDED variant of Falcon [skip ci] #362
Conversation
So done: LGTM. Thanks for preparing this PR, @SWilson4 ! |
SWilson4
force-pushed
the
sw-falcon-padded-tracker
branch
from
48f6f08 to
75beb11
Compare
oqs-template/oqs-sig-info.md
Outdated
| | falcon512 | 20211101 | 3 | 1 | 0xfed7 | 1.3.9999.3.11 | | ||
| | falcon512 **hybrid with** p256 | 20211101 | 3 | 1 | 0xfed8 | 1.3.9999.3.12 | | ||
| | falcon512 **hybrid with** rsa3072 | 20211101 | 3 | 1 | 0xfed9 | 1.3.9999.3.13 | | ||
| | falcon512 | PQClean Round 3 version | 3 | 1 | 0xfeae | 1.3.9999.3.6 | |
There was a problem hiding this comment.
This seems fishy: Why do OIDs 1.3.999.3.6-8 change algorithm version identifier (from "20211101" to "PQClean Round 3 version"), but neither OIDs, nor code points? Or asked the other way around: Why does the same algorithm version "20211101" get different OIDs?
There was a problem hiding this comment.
I did it this way because due to an old PQClean bug, the "20211101" version of Falcon in PQClean / liboqs / oqs-provider doesn't actually implement the upstream 20211101-version spec; the long KATs don't match. The new updates are to correctly implement (hopefully) the 20211101 version of the spec.
I suppose (?) that we shouldn't retroactively re-version things, though... what do you think would be the best way to resolve this?
There was a problem hiding this comment.
Ouch. No, retroactively re-versioning doesn't sound like a good approach. I do now understand the reference to PQClean, though. But this shows that we have an extraordinarily weak and brittle approach to versioning. Created open-quantum-safe/liboqs#1719 to track/improve. For this issue, what about adding to the old OIDs the (admittedly longish) term "PQClean Round 3 version labelled 20211101" ? That way, it becomes clear to the reader of this file that the same version ID had been used but had a different functionality.
SWilson4
force-pushed
the
sw-falcon-padded-tracker
branch
from
75beb11 to
fadaf5a
Compare
SWilson4
force-pushed
the
sw-falcon-padded-tracker
branch
from
fadaf5a to
73f03fc
Compare
|
open-quantum-safe/liboqs#1710 has now landed, so this PR should be ready to go. |
…safe#362) * Update OIDs to reflect Falcon KAT changes * Update "old" version string to reflect KAT mismatch
…safe#362) * Update OIDs to reflect Falcon KAT changes * Update "old" version string to reflect KAT mismatch Signed-off-by: Felipe Ventura <[email protected]>
…safe#362) * Update OIDs to reflect Falcon KAT changes * Update "old" version string to reflect KAT mismatch
…safe#362) * Update OIDs to reflect Falcon KAT changes * Update "old" version string to reflect KAT mismatch
…safe#362) * Update OIDs to reflect Falcon KAT changes * Update "old" version string to reflect KAT mismatch Signed-off-by: Felipe Ventura <[email protected]>
Pending merge of open-quantum-safe/liboqs#1710. I assigned the next available OIDs from what appeared to be the appropriate space (1.3.9999.3). Please double-check the changes to
generate.ymlto make sure that I did everything correctly.