linux-kernel - Re: [PATCH v2 5/6] tpm: add driver for cr50 on SPI

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]

Date:   Thu, 18 Jul 2019 11:07:34 -0700
From:   Stephen Boyd <swboyd@...omium.org>
To:     Alexander Steffen <Alexander.Steffen@...ineon.com>,
        Jarkko Sakkinen <jarkko.sakkinen@...ux.intel.com>,
        Peter Huewe <peterhuewe@....de>
Cc:     Andrey Pronin <apronin@...omium.org>, linux-kernel@...r.kernel.org,
        Jason Gunthorpe <jgg@...pe.ca>, Arnd Bergmann <arnd@...db.de>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        linux-integrity@...r.kernel.org,
        Duncan Laurie <dlaurie@...omium.org>,
        Guenter Roeck <groeck@...omium.org>
Subject: Re: [PATCH v2 5/6] tpm: add driver for cr50 on SPI

Quoting Alexander Steffen (2019-07-18 09:47:14)
> On 17.07.2019 21:57, Stephen Boyd wrote:
> > 
> > I think the idea is to let users override the quality if they decide
> > that they don't want to use the default value specified in the driver.
> 
> But isn't this something that applies to all TPMs, not only cr50? So 
> shouldn't this parameter be added to one of the global modules (tpm? 
> tpm_tis_core?) instead? Or do all low-level drivers (tpm_tis, 
> tpm_tis_spi, ...) need this parameter to provide a consistent interface 
> for the user?

Looking at commit 7a64c5597aa4 ("tpm: Allow tpm_tis drivers to set hwrng
quality.") I think all low-level drivers need to set the hwrng quality
somehow. I'm not sure how tpm_tis_spi will do that in general, but at
least for cr50 we have derived this quality number.

I can move this module parameter to tpm_tis_core.c, but then it will be
a global hwrng quality override for whatever tpm is registered through
tpm_tis_core instead of per-tpm driver. This is sort of a problem right
now too if we have two tpm_tis_spi devices. I can drop this parameter if
you want.

> 
> > 
> > Do you want me to describe something further?
> > 
> >> For example, struct
> >> cr50_spi_phy contains both tx_buf and rx_buf, whereas tpm_tis_spi uses a
> >> single iobuf, that is allocated via devm_kmalloc instead of being part
> >> of the struct. Maybe the difference matters, maybe not, who knows?
> > 
> > Ok. Are you asking if this is a full-duplex SPI device?
> 
> No, this was meant as an example for the previous question. As far as I 
> understood it, cr50 is basically compliant to the spec implemented by 
> tpm_tis_spi, but needs special handling in some cases. Therefore, I'd 
> expect a driver for cr50 to look exactly like tpm_tis_spi except for the 
> special bits here and there. The way buffers are allocated within the 
> driver is probably not something that should differ because of the TPM chip.
> 

Ok.