Re: GCM: cra_name == cra_driver_name
From: Stephan Mueller <hidden>
Date: 2016-01-26 13:14:42
Am Dienstag, 26. Januar 2016, 19:13:52 schrieb Herbert Xu: Hi Herbert,
On Mon, Jan 25, 2016 at 10:48:38PM +0100, Stephan Mueller wrote:quoted
Hi Herbert, with the current cryptodev tree, when I call a compound AEAD cipher with cra_driver_name for the first time, this cra_driver_name gets registered also as cra_name. Note, using other compound ciphers (like HMAC or symmetric ciphers), I do not see that error. With that registering, I am not able to resolve the cipher with usual common cra_name. Note, when I initialized an AEAD cipher with a cra_name, then all works as it used to be.Is this a new problem? I thought it always worked like this.
At least I have seen that now for the first time. But maybe I have not tested this order of invocation before.
I was going to fix it to generate a proper cra_name after I'm done with the skcipher conversion.
Ok, thanks.
Cheers,
Ciao Stephan