Re: [tpmdd-devel] [PATCH RESEND 3/3] tpm-chip: Export TPM device to user space even when startup failed

On Wed, Aug 30, 2017 at 12:34:16PM +0200, Michal Suchánek wrote:
> On Wed, 30 Aug 2017 13:20:02 +0300
> Jarkko Sakkinen <[email protected]> wrote:
> 
> > On Wed, Aug 30, 2017 at 01:15:10PM +0300, Jarkko Sakkinen wrote:
> > > On Tue, Aug 29, 2017 at 03:17:39PM +0200, Michal Suchánek wrote:  
> > > > Hello,
> > > > 
> > > > On Tue, 29 Aug 2017 15:55:09 +0300
> > > > Jarkko Sakkinen <[email protected]> wrote:
> > > >   
> > > > > On Mon, Aug 28, 2017 at 05:15:58PM +0000,
> > > > > [email protected] wrote:  
> > > > > > But is that just because nobody bothered to implement the
> > > > > > necessary logic or for some other reason?    
> > > > > 
> > > > > We do not want user space to access broken hardware. It's a
> > > > > huge risk for system stability and potentially could be used
> > > > > for evil purposes.
> > > > > 
> > > > > This is not going to mainline as it is not suitable for general
> > > > > consumption. You must use a patched kernel if you want this.
> > > > > 
> > > > > /Jarkko
> > > > >   
> > > > 
> > > > It has been pointed out that userspace applications that use
> > > > direct IO access exist for the purpose. So using a kernel driver
> > > > is an improvement over that if the interface is otherwise sane.
> > > > 
> > > > What do you expect is the potential for instability or evil use?  
> > > 
> > > By definition the use of broken hardware can have unpredictable
> > > effects. Use a patched kernel if you want to do it.
> > > 
> > > /Jarkko  
> > 
> > I.e. too many unknown unknowns for mainline.
> > 
> > I could consider a solution for the TPM error case on self-test that
> > allows only restricted subset of commands.
> > 
> > The patch description did not go to *any* detail on how it is used so
> > practically it's unreviewable at this point. There's a big burder of
> > proof and now there's only hand waving.
> > 
> Hello,
> 
> there was a bug patched recently in which Linux was not sending the
> shutdown command on system shutdown. Presumably with this bug some TPMs
> consider being under attack and stop performing most functions.
> However, you should be able to read the log if this is implemented
> sanely. For that the TPM needs to be accessible.
> 
> There are probably other cases when the TPM might be useless for system
> use but it might be useful to access it. For example, does Linux handle
> uninitialized TPMs?
> 
> Thanks
> 
> Michal

Agreed. I still think it would make sense to start with a limited subset
of TPM commands, not with all-command-allowed.

I guess Alexander should be able to propose such subset.

/Jarkko