[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <56C3886F.3050009@intel.com>
Date: Tue, 16 Feb 2016 12:37:03 -0800
From: Tadeusz Struk <tadeusz.struk@...el.com>
To: David Howells <dhowells@...hat.com>
Cc: herbert@...dor.apana.org.au, linux-kernel@...r.kernel.org,
linux-security-module@...r.kernel.org,
linux-crypto@...r.kernel.org, zohar@...ux.vnet.ibm.com
Subject: Re: [PATCH v5 1/3] crypto: KEYS: convert public key and digsig asym
to the akcipher api
Hi David,
On 02/11/2016 02:08 AM, David Howells wrote:
> David Howells <dhowells@...hat.com> wrote:
>
>> > Tadeusz Struk <tadeusz.struk@...el.com> wrote:
>> >
>>>> > > > Why didn't you put the RSA signature parsing - ie. where the OID and the other
>>>> > > > bits are checked - into crypto/rsa.c?
>>>> > > >
>>> > >
>>> > > Do you want to get rid of the crypto/asymmetric_keys/rsa.c completely?
>>> > > I wanted to make the conversion churn as small as possible.
>>> > > I can move it in a subsequent patch if you want.
>> >
>> > I was thinking of getting rid of it completely, yes.
>> >
>> > But I was wondering if you had some other motivation, such as keeping the
>> > crypto layer purely the mathematical operation.
>> >
>> > It's an interesting question where to draw the line, actually. The answer may
>> > hinge on what things like the TPM do. I should ask the TPM folks.
> Looking in the TPM emulator, the TPM_Sign operation indeed puts the wrappings
> on, so this needs to go into the crypto layer.
David,
If you could tell me what repo should I work with then I can prepare subsequent patch to
move this to a new file under crypto/
Thanks,
--
TS
Powered by blists - more mailing lists