Re: [bug] crypto/vmx/p8_ghash memory corruption in 4.8-rc7

Hi,

On Tue, Sep 27, 2016 at 05:01:03AM -0400, Jan Stancek wrote:

So, if we extended p8_ghash_desc_ctx to accommodate fallback_desc's ctx
and then provided statesize/import/export, would that be acceptable?

struct p8_ghash_desc_ctx {
        ...
        struct shash_desc fallback_desc;
+       char fallback_ctx[sizeof(struct ghash_desc_ctx)];

I think so. That's the solution mentioned by Herbert. The only drawback
is that we will need to fix "ghash-generic" as the fallback
implementation in order to know beforehand its descsize.

However I would keep the p8_ghash_desc_ctx the way it is and I would
sum sizeof(struct ghash_desc_ctx) to the algorithm descsize instead.

Let me put a quick patch together to test this.

Also, does that mean that padlock_sha has similar problem?
It does not seem to reserve any space for fallback __ctx and it calls
init()/update()/export() with padlock_sha_desc's fallback:

struct padlock_sha_desc {
        struct shash_desc fallback;
};

static struct shash_alg sha1_alg = {
        .descsize       =       sizeof(struct padlock_sha_desc),

Yeah. It still seems to me that padlock-sha has the same problem. Maybe
we are missing something...

--
Regards,
Marcelo