Received-SPF: pass (google.com: best guess record for domain of linux-bluetooth-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   "=?utf-8?Q?Micha=C5=82?= Lowas-Rzechonek" 
        <michal.lowas-rzechonek@silvair.com>
Date:   Mon, 8 Jul 2019 20:22:11 +0200
To:     "Stotland, Inga" <inga.stotland@intel.com>
Cc:     "jakub.witowski@silvair.com" <jakub.witowski@silvair.com>,
        "linux-bluetooth@vger.kernel.org" <linux-bluetooth@vger.kernel.org>,
        "Gix, Brian" <brian.gix@intel.com>
Subject: Re: [RFC BlueZ 0/1] Validate element indexation
Message-ID: <20190708182211.yxkucondnw5x6ay7@kynes>
Mail-Followup-To: "Stotland, Inga" <inga.stotland@intel.com>,
        "jakub.witowski@silvair.com" <jakub.witowski@silvair.com>,
        "linux-bluetooth@vger.kernel.org" <linux-bluetooth@vger.kernel.org>,
        "Gix, Brian" <brian.gix@intel.com>
References: <20190708141314.13950-1-jakub.witowski@silvair.com>
 <65651744c67ae13bd95ffac88fae42f0f6c16c61.camel@intel.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <65651744c67ae13bd95ffac88fae42f0f6c16c61.camel@intel.com>
User-Agent: NeoMutt/20180716
Sender: linux-bluetooth-owner@vger.kernel.org
Precedence: bulk

Inga, Jakub,

On 07/08, Stotland, Inga wrote:
> I agree that the validation for the gaps is needed. Interesting point
> about max number of elements...
> 
> I wonder if a better check woul be to we to add to construct
> composition data as a validation point to make sure it fits in mesh
> message. Plus, an additional strict check can be done when Attach
> method is called: stored composition can be byte compared to the one
> dynamically generated from collected properties...

If I read that correctly, this means we would need a way to build
Composition Data on the fly, during get_manager_object_cb processing.

I think it would be possible to get rid of validate_model_property
function - instead, we could build a temporary mesh_node instance
using information provided by the application as-is, and then:

- in case of existing nodes, generate Composition Data from both
  existing and temporary instances, and byte-compare the two

- in case of new nodes, simply save the temporary instace to 'nodes'
  list

All of that assumes that Composition Data generationchecks that:
 - everything fits into a buffer (this is already done)
 - mandatory models are present
 - indexation is OK

I think this would make things slightly more consistent, and we would
get rid of most "is_new" checks during attach/join/create/import.

regard
-- 
Michał Lowas-Rzechonek <michal.lowas-rzechonek@silvair.com>
Silvair http://silvair.com
Jasnogórska 44, 31-358 Krakow, POLAND