linux-kernel - Re: [PATCH v1 12/12] tpm: TPM2 sysfs attributes

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: <20140924201938.GC10887@obsidianresearch.com>
Date:	Wed, 24 Sep 2014 14:19:38 -0600
From:	Jason Gunthorpe <jgunthorpe@...idianresearch.com>
To:	Jarkko Sakkinen <jarkko.sakkinen@...ux.intel.com>
Cc:	tpmdd-devel@...ts.sourceforge.net, Peter Huewe <peterhuewe@....de>,
	Marcel Selhorst <tpmdd@...horst.net>,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH v1 12/12] tpm: TPM2 sysfs attributes

On Wed, Sep 24, 2014 at 10:02:34PM +0300, Jarkko Sakkinen wrote:

> > The pcrs file never conformed to the sysfs rules, if TPM2 is getting a
> > whole new file set, I wouldn't mind seeing it not include the
> > non-conformant ones. What do you think?
> 
> I think that it's better to put extra focus on these sysfs attributes in
> first patch set because it's user space visible. What's wrong in the
> current pcrs file?

Each PCR should be a distinct sysfs file, probably with a
directory. One Value Per File is the rule.

> > > +static ssize_t caps_show(struct device *dev, struct device_attribute *attr,
> > > +			 char *buf)
> > > +{
> > 
> > Ditto.. The manfacturer number should probably be its own file
> 
> Maybe here would make sense to have three files:
> 
> - manufacturer
> - firmware_1
> - firmware_2
> 
> More or less following the name of the TPM properties in the
> specification.

Probably, maybe firmware_1/2 could be combined if they are the same
logical value? (I've always expressed it as firmware_1.firwmare_2?)

> I did not fully understand the comment about tpm2 flag. Why driver
> cannot set it when it initializes the device like with this based
> on value of the STS3?

I was talking about the /dev/ char device - a random application today
will open it and send TPM1 formed messages. Those should be refused
with EINVAL for a TPM2 chip unless the application declares via IOCTL
that it will be sending TPM2 messages.

Otherwise the API contract for the /dev/ device (write TPM1 formed
messages) is broken..

Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/