linux-kernel - Re: [PATCH] tpm: drop chip->is_open and chip->duration

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <20161115051106.u2xoduwf2kpcznv3@intel.com>
Date:   Mon, 14 Nov 2016 21:11:54 -0800
From:   Jarkko Sakkinen <jarkko.sakkinen@...ux.intel.com>
To:     Jason Gunthorpe <jgunthorpe@...idianresearch.com>
Cc:     tpmdd-devel@...ts.sourceforge.net,
        linux-security-module@...r.kernel.org,
        Peter Huewe <peterhuewe@....de>,
        Marcel Selhorst <tpmdd@...horst.net>,
        Christophe Ricard <christophe.ricard@...il.com>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] tpm: drop chip->is_open and chip->duration_adjusted

On Mon, Nov 14, 2016 at 09:30:01PM -0700, Jason Gunthorpe wrote:
> On Mon, Nov 14, 2016 at 03:44:58PM -0800, Jarkko Sakkinen wrote:
> > Use atomic bitops for chip->flags so that we do not need chip->is_open
> > and chip->duration_adjusted anymore.
> 
> I don't know if it s a really great idea to use atomic bit ops for
> things that do not need to be atomic.. It makes the locking scheme
> less clear. is open is genuinely different since it relies on the
> atomic for correctness.

The way I see it is one of the status flags bound to chip among the
others. I do not see this cause too much harm for clarity. It eases
debugging the driver a bit because you get more state out of 'flags'.

It also makes code little a bit more robust as flags is independent of
locks.

How strong is your opposition here? I do not see any exceptional damage
done but see some subtle but still significant benefits.

> Merging is_duration makes lots of sense though

Also timeout_adjusted should be merged (for some reason missed it).

> Jason

/Jarkko