> Kalyna is a Ukrainian block cipher with variable key and block sizes. Its
>> now available in the the library. The tracking issue and commits of
>> interest are:
>>
>> * https://github.com/weidai11/cryptopp/issues/411
>> * https://github.com/weidai11/cryptopp/commit/a5c67cfdd6ad
>> * https://github.com/weidai11/cryptopp/commit/1d7dfc69274d
>>
>> Its our first attempt at variable block sizes, so it may have some rough
>> edges.
>>
>> We also have a wiki page started at https://www.cryptopp.com/wiki/Kalyna.
>> Its a work in progress.
>>
>
> I got to run the cryptest.sh last night on Kalyna. I have good news and
> bad news...
>
> The good news is Kalyna is testing good on nearly every platform. The bad
> news is, it failed on my 6th gen Skylake. The Skylake runs Fedora with GCC
> 6.3.1, and its known to be a little more rigid with respect to dotting i's
> and crossing t's. For example, Red Hat will run memcpy's in reverse, which
> breaks regular memcpy if the buffers overlap. Overlapping buffers are
> usually undefined behavior, but other compilers and platforms are usually
> more accommodating.
>
> I'm guessing I introduced some undefined behavior. I'll track down the
> undefined behavior shortly.
>
Yeah, it was UB from the cache timing attack hardening (commit
9cf9f4235d52). A out-of-bounds read was performed. Bad code which tries to
access S[256] (from an array of S[4][256]):
word32 u ...;
for (unsigned int i=0; i<256; i+=cacheLineSize)
u &= *reinterpret_cast<const word32*>(KalynaTab::S+i);
Good code:
word64 u ...;
const byte* p = reinterpret_cast<const byte*>(KalynaTab::S);
for (unsigned int i=0; i<256; i+=cacheLineSize)
u &= *reinterpret_cast<const word64*>(p+i);
If anyone is interested, UBSan identified the problem. 'make ubsan' to the
rescue!
Jeff
--
--
You received this message because you are subscribed to the "Crypto++ Users"
Google Group.
To unsubscribe, send an email to cryptopp-users-unsubscr...@googlegroups.com.
More information about Crypto++ and this group is available at
http://www.cryptopp.com.
---
You received this message because you are subscribed to the Google Groups
"Crypto++ Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to cryptopp-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
