From: Catalin Vasile Subject: Re: [PATCH] crypto: caam: add backlogging support Date: Fri, 13 May 2016 12:55:09 +0000 Message-ID: References: <1462540733-2170-1-git-send-email-cata.vasile@nxp.com> <20160510094622.GA15860@gondor.apana.org.au> ,<20160511105411.GB25016@gondor.apana.org.au> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT Cc: "linux-crypto@vger.kernel.org" , "linux-crypto-owner@vger.kernel.org" , Horia Ioan Geanta Neag , Alexandru Porosanu , Scott Wood To: Herbert Xu Return-path: Received: from mail-db3on0066.outbound.protection.outlook.com ([157.55.234.66]:1281 "EHLO emea01-db3-obe.outbound.protection.outlook.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S932127AbcEMNK7 convert rfc822-to-8bit (ORCPT ); Fri, 13 May 2016 09:10:59 -0400 In-Reply-To: <20160511105411.GB25016@gondor.apana.org.au> Content-Language: en-US Sender: linux-crypto-owner@vger.kernel.org List-ID: > ________________________________________ > From: Herbert Xu > Sent: Wednesday, May 11, 2016 1:54 PM > To: Catalin Vasile > Cc: linux-crypto@vger.kernel.org; linux-crypto-owner@vger.kernel.org; Horia Ioan Geanta Neag; Alexandru Porosanu; Scott Wood > Subject: Re: [PATCH] crypto: caam: add backlogging support> > On Wed, May 11, 2016 at 07:53:19AM +0000, Catalin Vasile wrote: > > > > Every request will be queued and eventually done. > > The hardware equipment has a constraint on the number of tfms it can have. > > Is there a requirement to support an infinite number of tfms on a device? > > > > > You should use a software queue instead.>> > As I said you drivers are always supposed to have a software queue > in addition to any hardware queues. There is no reason why your > driver should be special in this regard.> Hi Herbert, We are inclining towards a hardware backlogging solution because on low end devices the software queue does not scale very well. Regarding your concerns, we are now evaluating adding a software queue based fallback solution: if the hardware backlogging fails because no hardware slots are available, backlog the job inside a crypto software queue. Is this hybrid configurable solution acceptable? Namely, if the hardware queue is full, to fallback to the crypto software queue. > Cheers, > -- > Email: Herbert Xu > Home Page: http://gondor.apana.org.au/~herbert/ > PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt