From: Steffen Klassert Subject: Re: [dm-devel] [PATCH] DM-CRYPT: Scale to multiple CPUs Date: Wed, 2 Jun 2010 08:30:33 +0200 Message-ID: <20100602063033.GS2184@secunet.com> References: <20100531160425.GA20344@basil.fritz.box> <20100601043901.GA25693@gondor.apana.org.au> <20100602051403.GA5784@gondor.apana.org.au> <20100602060702.GA6319@gondor.apana.org.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20100602060702.GA6319@gondor.apana.org.au> Sender: linux-kernel-owner@vger.kernel.org To: Herbert Xu Cc: Mikulas Patocka , device-mapper development , linux-kernel@vger.kernel.org, agk@redhat.com, ak@linux.intel.com List-Id: dm-devel.ids On Wed, Jun 02, 2010 at 04:07:02PM +1000, Herbert Xu wrote: > On Wed, Jun 02, 2010 at 01:25:11AM -0400, Mikulas Patocka wrote: > > > > Are you sure that you know what you're talking about? pcrypt_alloc > > contains this: > > switch (algt->type & algt->mask & CRYPTO_ALG_TYPE_MASK) { > > case CRYPTO_ALG_TYPE_AEAD: > > return pcrypt_alloc_aead(tb); > > } > > > > return ERR_PTR(-EINVAL); > > --- so for anything other byt AEAD it returns -EINVAL. > > So someone needs to write the ablkcipher plugins for pcrypt. > I did that already for one of the older (remote softirq based) version of pcrypt. I just never pushed it out because I was focused on IPsec and the softirq based version I was not that usefull for dm-crypt. Now, that we use workqueues internally, dm-crypt could benefit from pcrypt too. Let me see whether I can respin the old ablkcipher plugin. Steffen