Re: [PATCH v3 4/6] tpm: Support TPM2 sized buffers (TPM2B)

[Jarkko Sakkinen]

On Tue Nov 7, 2023 at 7:20 PM EET, Jerry Snitselaar wrote:
> On Tue, Oct 24, 2023 at 04:15:22AM +0300, Jarkko Sakkinen wrote:
> > Add boolean parameters @alloc and @sized to tpm_buf_init():
> > 
> > * If @alloc is set to false, buf->data is assumed to be pre-feeded and
> >   owned by the caller.
> > * If @sized is set to true, the buffer represents a sized buffer
> >   (TPM2B).
> > 
> > Signed-off-by: Jarkko Sakkinen <jarkko@xxxxxxxxxx>
> > ---
> >  drivers/char/tpm/tpm-buf.c                | 32 ++++++++++++++++-------
> >  drivers/char/tpm/tpm-sysfs.c              |  2 +-
> >  drivers/char/tpm/tpm1-cmd.c               | 14 +++++-----
> >  drivers/char/tpm/tpm2-cmd.c               | 22 ++++++++--------
> >  drivers/char/tpm/tpm2-space.c             |  4 +--
> >  drivers/char/tpm/tpm_vtpm_proxy.c         |  2 +-
> >  include/linux/tpm.h                       |  3 ++-
> >  security/keys/trusted-keys/trusted_tpm1.c |  4 +--
> >  security/keys/trusted-keys/trusted_tpm2.c |  6 ++---
> >  9 files changed, 51 insertions(+), 38 deletions(-)
> > 
> > diff --git a/drivers/char/tpm/tpm-buf.c b/drivers/char/tpm/tpm-buf.c
> > index fa9a4c51157a..f1d92d7e758d 100644
> > --- a/drivers/char/tpm/tpm-buf.c
> > +++ b/drivers/char/tpm/tpm-buf.c
> > @@ -7,22 +7,32 @@
> >  #include <linux/tpm.h>
> >  
> >  /**
> > - * tpm_buf_init() - Initialize from the heap
> > + * tpm_buf_init() - Initialize a TPM buffer
> >   * @buf:	A @tpm_buf
> > + * @sized:	Represent a sized buffer (TPM2B)
> > + * @alloc:	Allocate from the heap
> >   *
> >   * Initialize all structure fields to zero, allocate a page from the heap, and
>
> Depending on what the decision ends up being on the bools, flags,
> separate functions, or wrappers possibly an "if needed" should be
> tacked on to the end of "allocate a page from the heap" here.
>
>
> Flags would be better when coming across calls to the routine in the
> code than the bools, but I think switching to wrappers around
> a __tpm_buf_init for the different types would be good.

Yeah, I'll bake something based on this discussion.

BR, Jarkko