mbox series

[0/3] crypto: sun4i-ss: fix SHA1 on A33 SecuritySystem

Message ID 20191114144812.22747-1-clabbe.montjoie@gmail.com
Headers show
Series crypto: sun4i-ss: fix SHA1 on A33 SecuritySystem | expand

Message

Corentin Labbe Nov. 14, 2019, 2:48 p.m. UTC 
Thanks to Igor Pecovnik, I have now in my kernelCI lab, a sun8i-a33-olinuxino.
Strange behavour, crypto selftests was failling but only for SHA1 on
this A33 SoC.

This is due to the A33 SS having a difference with all other SS, it give SHA1 digest directly in BE.
This serie handle this difference.

Corentin Labbe (3):
  dt-bindings: crypto: add new compatible for A33 SS
  ARM: dts: sun8i: a33: add the new SS compatible
  crypto: sun4i-ss: add the A33 variant of SS

 .../crypto/allwinner,sun4i-a10-crypto.yaml    |  3 +++
 arch/arm/boot/dts/sun8i-a33.dtsi              |  3 ++-
 .../crypto/allwinner/sun4i-ss/sun4i-ss-core.c | 22 ++++++++++++++++++-
 .../crypto/allwinner/sun4i-ss/sun4i-ss-hash.c |  5 ++++-
 drivers/crypto/allwinner/sun4i-ss/sun4i-ss.h  |  9 ++++++++
 5 files changed, 39 insertions(+), 3 deletions(-)

-- 
2.23.0

Comments

Corentin Labbe Nov. 19, 2019, 7:39 a.m. UTC | #1 
On Mon, Nov 18, 2019 at 12:11:43PM +0100, Maxime Ripard wrote:
> Hi,

> 

> On Thu, Nov 14, 2019 at 03:48:11PM +0100, Corentin Labbe wrote:

> > Add the new A33 SecuritySystem compatible to the crypto node.

> >

> > Signed-off-by: Corentin Labbe <clabbe.montjoie@gmail.com>

> > ---

> >  arch/arm/boot/dts/sun8i-a33.dtsi | 3 ++-

> >  1 file changed, 2 insertions(+), 1 deletion(-)

> >

> > diff --git a/arch/arm/boot/dts/sun8i-a33.dtsi b/arch/arm/boot/dts/sun8i-a33.dtsi

> > index 1532a0e59af4..5680fa1de102 100644

> > --- a/arch/arm/boot/dts/sun8i-a33.dtsi

> > +++ b/arch/arm/boot/dts/sun8i-a33.dtsi

> > @@ -215,7 +215,8 @@

> >  		};

> >

> >  		crypto: crypto-engine@1c15000 {

> > -			compatible = "allwinner,sun4i-a10-crypto";

> > +			compatible = "allwinner,sun8i-a33-crypto",

> > +				     "allwinner,sun4i-a10-crypto";

> 

> If some algorithms aren't working properly, we can't really fall back

> to it, we should just use the a33 compatible.

> 


Since crypto selftest detect the problem, the fallback could be used and SS will just be in degraded mode (no sha1).
But since nobody reported this problem since 4 years (when SS was added in a33 dts), the absence of sha1 is clearly not an issue.

Regards