[pycrypto] Crypto.Random crashes due to unaligned access

Greg Price gnprice at gmail.com
Sun Oct 27 15:35:21 PDT 2013

On Sun, Oct 27, 2013 at 3:07 PM, Sebastian Ramacher
<sebastian+lists at ramacher.at> wrote:
> I debugged this for a while and the problem is not _mm_loadu_si128.
> That's fine. It generates the correct movdqu instruction for that. The
> problem is the rk[0] = ... part. On amd64, ek and dk from block_state
> get aligned at 16 byte boundaries and everything works out properly.
> However, on i386 this does not appear to be true.

Thanks for looking into this!  That is exactly right.  I debugged this
further later on the evening of the 23rd and came to the same
conclusion.  Unfortunately the mailing list has been eating my
messages. :(  The only one you've seen is the first one, in Dwayne's
reply to it on-list.

If this message comes through to the list I'll forward my subsequent
messages to it shortly; in any case I'll forward them to you.

> I have some ideas how to fix it and will hopefully come up with a patch
> next week.



More information about the pycrypto mailing list