[U-Boot] tpm TIS TPMv2.0

[Miquel Raynal]

Hi Martin,

On Mon, 18 Jun 2018 08:20:20 +0000, "Hecht, Martin (Avnet Silica)"
<Martin.Hecht at avnet.eu> wrote:

> Hi Miquel,
> 
> > -----Original Message-----
> > From: Miquel Raynal [mailto:miquel.raynal at bootlin.com]
> > Sent: Montag, 18. Juni 2018 10:05
> > To: Hecht, Martin (Avnet Silica) <Martin.Hecht at avnet.eu>
> > Cc: sjg at chromium.org; u-boot at lists.denx.de
> > Subject: Re: [U-Boot] tpm TIS TPMv2.0
> > 
> > Hi Martin,
> > 
> > On Fri, 15 Jun 2018 13:34:07 +0000, "Hecht, Martin (Avnet Silica)"
> > <Martin.Hecht at avnet.eu> wrote:
> >   
> > > Hi Miquel, Simon,
> > >
> > > Is there any specific reason why the new tpm2_tis_spi_xfer doesn't  
> > support full duplex? It seems we did some work in parallel but you sent the
> > patches earlier. Is that codes tested against an existing TPM v2? I have a
> > working implementation what runs on SLB9670 including full duplex.
> > 
> > What do you mean exactly?
> > 
> > I don't think the TPM2 protocol makes real use of full-duplex unless for the
> > wait state between the host command and the actual xfer.  
> 
> You are right, TIS 1.3 FIFO doesn’t use full duplex in physical level. What I mean is that the driver you just wrote doesn't use the xfer function in that way that you can specify in and out parameters at same time. I did this in my implementation what gave me an easy chance to control the CS# of the TPM.

Do you need this CS# handling for more advanced features? Same question
for the in/out xfers?

> Can you tell me on what TPM did you test? For the SLB9670 the code
> doesn't work on my hardware.

I tested with a ST33TPHF20 SPI TPM.

I'm surprised it did not work with an SLB9670, I don't see anything in
the spec explaining this CS# specificity.

> For the code you wrote I'm considering to
> add a few lines to control the CS# in that way how my xfer is doing
> this for the SLB9670.

Yes please, share the patch and add me in cc so I could test it with
mine.

> On the other hand what about to use a xfer what
> can handle all three cases (in, out, in/out)?

As I did not implement any TPM command that needed it I did not care
about it. Of course if there is a need for it: it should be implemented
too. I contributed only basic support for essential commands (measured
boot, mainly) but please feel free to enhance the code to add more
features!

Regards,
Miquèl