Re: [tpmdd-devel] [PATCH v8 6/8] tpm: TPM 2.0 baseline support

[peterhuewe]

Hi 

On 8. Dezember 2014 14:56:15 MEZ, Jarkko Sakkinen <jarkko.sakkinen@xxxxxxxxxxxxxxx> wrote:
>On Fri, Dec 05, 2014 at 12:13:18AM +0100, Peter HÃwe wrote:
>> Am Mittwoch, 3. Dezember 2014, 03:28:35 schrieb Stefan Berger:
>> > On 12/02/2014 05:31 PM, Jarkko Sakkinen wrote:
>> > > +
>> > > +/**
>> > > + * tpm2_startup() - send startup command to the TPM chip
>> > > + * @chip:		TPM chip to use.
>> > > + * @startup_type	startup type. The value is either
>> > > + *			TPM_SU_CLEAR or TPM_SU_STATE.
>> > > + *
>> > > + * 0 is returned when the operation is successful. If a negative
>number
>> > > is + * returned it remarks a POSIX error code. If a positive
>number is
>> > > returned + * it remarks a TPM error.
>> > > + */
>> > > +int tpm2_startup(struct tpm_chip *chip, __be16 startup_type)
>> > > +{
>> > > +	struct tpm2_cmd cmd;
>> > > +
>> > > +	cmd.header.in = tpm2_startup_header;
>> > > +
>> > > +	cmd.params.startup_in.startup_type = startup_type;
>> > > +	return tpm_transmit_cmd(chip, &cmd, sizeof(cmd),
>> > > +				"attempting to start the TPM");
>> > > +}
>> > 
>> > I suppose you need to send this command because your firmware does
>not
>> > do it ?Following TPM1.2 I guess the BIOS / UEFI should send this
>instead
>> > and sending it later would actually be wrong. Hm, I don't find from
>> > where you are calling this... do you need it ? Can you remove it?
>> > 
>> >     Stefan
>> 
>> Hi,
>> 
>> I think it would be good to send a TPM2_Startup if the TPM sends a 
>> TPM_RC_INITIALIZE (0x100) - so it becomes atleast usable.
>> Of course the BIOS/UEFI/Firmware should send the TPM2_Startup, but if
>there is 
>> no such thing, I would prefer Linux to do it, rather than nobody.
>> (analog: This was done for embedded platforms with TPM1.2).
>> 
>> In the current situation (v9) it is not possible to use the TPM2 on a
>machine 
>> without bios integration. :( (so I cannot test here :( )
>
>Should the place be if sending self-test fails? I think the type should
>be TPM2_SU_CLEAR. Do you agree?
>

Yes. If the first command returns "invalid post init" or whatever it is called in tpm20 speech the driver should send the startup clear.


Not sure if we should send a startup state in the resume case (like on tpm1.2)

And also it might make sense to send a tpm_shutdown? (If we aren't already) I think even on a machine with bios integration we have to send this?


>All other issues are now fixed except this and STS3 bit issue that I 
>look for next. In my github there is tpm2-v10 branch now with fixes
>on top. I squash the fixes right after these two remaining issues are
>fixed.
>
>> Peter
>
>/Jarkko

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