Date: Fri, 17 Mar 2017 22:37:18 +0200
From: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
To: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Cc: Jerry Snitselaar <jsnitsel@redhat.com>,
        Jarkko Sakkinen <jarkko.sakkinen@iki.fi>,
        tpmdd-devel@lists.sourceforge.net,
        linux-security-module@vger.kernel.org, gang.wei@intel.com,
        Peter Huewe <peterhuewe@gmx.de>, Marcel Selhorst <tpmdd@selhorst.net>,
        open list <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v3] tpm_crb: request and relinquish locality 0
Message-ID: <20170317203718.3z45mfvnrgjzwn7l@intel.com>
References: <20170315055738.11088-1-jarkko.sakkinen@iki.fi>
 <87h92uvtce.fsf@redhat.com>
 <87fuidx5mt.fsf@redhat.com>
 <20170316115543.jw23imcvkcriaam5@intel.com>
 <878to3x2h2.fsf@redhat.com>
 <20170317171116.GA29996@obsidianresearch.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20170317171116.GA29996@obsidianresearch.com>
Organization: Intel Finland Oy - BIC 0357606-4 - Westendinkatu 7, 02160 Espoo
User-Agent: Mutt/1.6.2-neo (2016-08-21)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1034
Lines: 28

On Fri, Mar 17, 2017 at 11:11:16AM -0600, Jason Gunthorpe wrote:
> On Fri, Mar 17, 2017 at 10:00:41AM -0700, Jerry Snitselaar wrote:
> 
> > > Changing the return value to -EBUSY was a stupid mistake from my side.
> > >
> > > I'll try revise this a bit in a way that the API will allow positive
> > > value for stating that the given locality has been already taking.
> > 
> > Is there a big performance hit with requesting and releasing locality?
> > If instead it just released it when release_locality is called I think
> > the changes are pretty minor.
> 
> If you can measure please let us know :)
> 
> This is all very old it may not actually make any sense..
> 
> .. and as I said earlier if we want to 'cache' the locality for
> performance then the core code should do it.
> 
> I kinda thought the point of releasing the locality was to allow other
> platform things to access the TPM, so I'm confused why TIS wouldn't
> always release it as well..
> 
> Jason

I would propose to make tpm_tis_core to work like that.

/Jarkko