[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180305111107.GH25377@linux.intel.com>
Date: Mon, 5 Mar 2018 13:11:07 +0200
From: Jarkko Sakkinen <jarkko.sakkinen@...ux.intel.com>
To: J Freyensee <why2jjj.linux@...il.com>
Cc: linux-integrity@...r.kernel.org,
linux-security-module@...r.kernel.org, keyrings@...r.kernel.org,
Peter Huewe <peterhuewe@....de>,
Jason Gunthorpe <jgg@...pe.ca>, Arnd Bergmann <arnd@...db.de>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 3/5] tpm: migrate tpm2_probe() to use struct tpm_buf
On Thu, Mar 01, 2018 at 02:10:17PM -0800, J Freyensee wrote:
> .
> .
> .
> I'm new to this area of the kernel, but I'm not getting these lines:
>
> > + rc = tpm_transmit_cmd(chip, NULL, buf.data, PAGE_SIZE, 0, 0, NULL);
> > + tpm_buf_destroy(&buf);
> > if (rc < 0)
> Why is this if() check not directly after the tpm_transmit_cmd() call that
> sets rc? Is it correct you want to destroy buf regardless of the
> tpm_transmit_cmd() outcome?
> > return rc;
> > -
> > - if (be16_to_cpu(cmd.header.out.tag) == TPM2_ST_NO_SESSIONS)
> > + out = (struct tpm_output_header *)buf.data;
>
> So buf has been destroyed, buf.data sill has something valid to assign to
> out?
> > + if (be16_to_cpu(out->tag) == TPM2_ST_NO_SESSIONS)
> > chip->flags |= TPM_CHIP_FLAG_TPM2;
> > return 0;
> Thanks,
> Jay
Nope it is a regression in the patch. Thank you :-) tpm_buf_destroy()
can be called if the response data is not needed other than everything
went OK (tpm_transmit_cmd() already digs this info).
/Jarkko
Powered by blists - more mailing lists