Re: [PATCH v4 1/2] tpm: Refactor tpm_transmit pulling out tpm_transfer function

[Stefan Berger]

On 05/15/2017 12:04 PM, Stefan Berger wrote:
> On 05/15/2017 08:40 AM, Jarkko Sakkinen wrote:
> > On Wed, May 10, 2017 at 07:54:21PM -0400, Stefan Berger wrote:
> > > Refactor tpm_transmit and pull out code sending the command
> > > and receiving the response and put this into tpm_transfer.
> > >
> > > Signed-off-by: Stefan Berger <stefanb@xxxxxxxxxxxxxxxxxx>
> > > ---
> > >   drivers/char/tpm/tpm-interface.c | 121 
> > > +++++++++++++++++++++++----------------
> > >   1 file changed, 73 insertions(+), 48 deletions(-)
> > >
> > > diff --git a/drivers/char/tpm/tpm-interface.c 
> > > b/drivers/char/tpm/tpm-interface.c
> > > index 158c1db..263b6d1 100644
> > > --- a/drivers/char/tpm/tpm-interface.c
> > > +++ b/drivers/char/tpm/tpm-interface.c
> > > @@ -370,67 +370,29 @@ static bool tpm_validate_command(struct 
> > > tpm_chip *chip,
> > >   }
> > >     /**
> > > - * tmp_transmit - Internal kernel interface to transmit TPM commands.
> > > + * tmp_transfer - Send a TPM command to the TPM and receive response
> > >    *
> > >    * @chip: TPM chip to use
> > >    * @buf: TPM command buffer
> > > + * @count: size of the TPM command
> > >    * @bufsiz: length of the TPM command buffer
> > > - * @flags: tpm transmit flags - bitmap
> > >    *
> > >    * Return:
> > > - *     0 when the operation is successful.
> > > + *     >0 when the operation is successful; returns response length
> > >    *     A negative number for system errors (errno).
> > >    */
> > > -ssize_t tpm_transmit(struct tpm_chip *chip, struct tpm_space *space,
> > > -             u8 *buf, size_t bufsiz, unsigned int flags)
> > > +ssize_t tpm_transfer(struct tpm_chip *chip, u8 *buf, u32 count, 
> > > size_t bufsiz)
> > Add instead a flag TPM_TRANSMIT_RAW (this name is just a suggestion)
> > that skips "prepare" and "commit" parts. That would save us from a
> > new export.
> >
> > Better way to make it less messy would be to add static functions
> > tpm_prepare_command and tpm_commit_command that would be always
> > called and would return immediately if flags contain TPM_TRANSMIT_RAW.
>
> I'll do that then, modifying the code to skip tpm_validate_command as 
> well.

But we don't want to recurse (in the 2nd patch) into 
chip->ops->request_locality(). Any suggestions ?

   Stefan