aboutsummaryrefslogtreecommitdiff
path: root/GNUmakefile
diff options
context:
space:
mode:
authorRob Austein <sra@hactrn.net>2018-04-02 09:46:47 -0400
committerRob Austein <sra@hactrn.net>2018-04-02 09:46:47 -0400
commit9d701b8a4da62de27b20f928f4385c46ece33d7f (patch)
treeba3eba1e4d20f7c0e9c5ce9ca8b34c92c33901ce /GNUmakefile
parentd8f5f37c2dffab9083fa4c3a89032fabb100f627 (diff)
First cut at reusing RSA blinding factors.rsa-blind-mutation
General technique here suggested by Peter Gutman. If I got the math wrong, blame me, not Peter. This has not yet been tested to confirm that it returns correct results when using the blinding factors cache, and preliminary timing results suggest that we may be chasing the wrong performance problem. Unclear whether we'll ever really want to integrate this change, but pushing it on a branch to get it into repository history. If we do end up using this, the blinding factors cache will need minor redesign, principally to use the external SDRAM because main memory has gotten kind of full. Some way to clear the cache when restarting the HSM would be nice, probably requires a `hal_rsa_init()` function.
Diffstat (limited to 'GNUmakefile')
0 files changed, 0 insertions, 0 deletions