linux-kernel - Re: [PATCH v2 00/26] Improve DVB documentation and reduce its gap

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20170904084309.7382864a@recife.lan>
Date:   Mon, 4 Sep 2017 08:46:16 -0300
From:   Mauro Carvalho Chehab <mchehab@...pensource.com>
To:     Honza Petrouš <jpetrous@...il.com>
Cc:     Linux Doc Mailing List <linux-doc@...r.kernel.org>,
        Linux Media Mailing List <linux-media@...r.kernel.org>,
        Mauro Carvalho Chehab <mchehab@...radead.org>,
        linux-kernel@...r.kernel.org, Jonathan Corbet <corbet@....net>
Subject: Re: [PATCH v2 00/26] Improve DVB documentation and reduce its gap

Em Mon, 4 Sep 2017 11:40:59 +0200
Honza Petrouš <jpetrous@...il.com> escreveu:

> > So, IMHO, the interface is broken by design. Perhaps that's
> > the reason why no upstream driver uses it.  
> 
> I have the same feeling regarding brokenness.
> 
> >
> > What seems to be a much better design would be to use the demux
> > set filter ioctls and route the PIDs to the right CA.
> >  
> 
> I don't have access to any programmer reference documentation
> for any modern DVB-enabled SoC, but I see two possible scenario
> of connecting descramblers to the demuxes (most of modern SoCs
> have more then one demux) - static one, when every demux has
> predefined descramblers already connected to it and dynamic ones,
> when any descrambler can be connected to the any demux.

I don't have access to the documentation either, but I know
some designs that have multiple demods that are dynamically set.
Some hardware even allow to dynamically change the maximum amount
of filters per demod at runtime.

> From that reason I vote to have some descrambler specific ioctl,
> which allow more flexibility then if we add it to the filter set ioctl.

I suspect that doing it at the demod does a lot more sense.

Anyway, someone should come with a driver requiring it upstream
for us to discuss and find the better alternatives to support.

Thanks,
Mauro