Re: [PATCH] crypto: print self-test pass notices in fips mode

[Jarod Wilson <jarod@redhat.com>]

On Wednesday 29 April 2009 09:21:53 Jarod Wilson wrote:
> On Wednesday 29 April 2009 09:18:17 Herbert Xu wrote:
> > On Tue, Apr 28, 2009 at 09:21:35PM -0400, Jarod Wilson wrote:
> > >
> > > diff --git a/crypto/testmgr.c b/crypto/testmgr.c
> > > index 39ffa69..d0cc85c 100644
> > > --- a/crypto/testmgr.c
> > > +++ b/crypto/testmgr.c
> > > @@ -2149,6 +2149,10 @@ notest:
> > >  test_done:
> > >  	if (fips_enabled && rc)
> > >  		panic("%s: %s alg self test failed in fips mode!\n", driver, alg);
> > > +	/* fips mode requires we print out self-test success notices */
> > > +	if (fips_enabled && !rc && strncmp(alg, "ctr(aes", 7))
> > > +		printk(KERN_INFO "alg: self-tests for %s (%s) passed\n",
> > > +		       driver, alg);
> > 
> > What is this strncmp crap for?
> 
> To avoid claiming we successfully self-tested ctr(aes) when its
> not actually directly testable. Was intended to go sort of hand
> in hand with the other patch to suppress 'no self test' messages
> for ctr(aes) when in fips mode. Of course, since at this point,
> we've run ecb(aes), and that's what's suggested as the way to
> test ctr(aes)[*], perhaps we don't need to
> suppress it.
> 
> [*] well, along with the sign-off from the lab that the counter
> code is acceptable

So this might actually be another argument in favor of adding a "this
algo isn't really testable" flag as Neil suggested...

-- 
Jarod Wilson
jarod@redhat.com