Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932797AbdDEJDd (ORCPT ); Wed, 5 Apr 2017 05:03:33 -0400 Received: from mga06.intel.com ([134.134.136.31]:42785 "EHLO mga06.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932783AbdDEJDb (ORCPT ); Wed, 5 Apr 2017 05:03:31 -0400 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.36,277,1486454400"; d="scan'208";a="73869380" Date: Wed, 5 Apr 2017 12:03:27 +0300 From: Jarkko Sakkinen To: Enric Balletbo i Serra Cc: Peter Huewe , Andrew Lunn , tpmdd-devel@lists.sourceforge.net, linux-kernel@vger.kernel.org, wsa@the-dreams.de, Bryan Freed Subject: Re: [PATCH v3] tpm: Apply a sane minimum adapterlimit value for retransmission. Message-ID: <20170405090327.pssnm3mjcpajpoy6@intel.com> References: <20170328152938.14539-1-enric.balletbo@collabora.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20170328152938.14539-1-enric.balletbo@collabora.com> Organization: Intel Finland Oy - BIC 0357606-4 - Westendinkatu 7, 02160 Espoo User-Agent: Mutt/1.6.2-neo (2016-08-21) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 4616 Lines: 138 On Tue, Mar 28, 2017 at 05:29:38PM +0200, Enric Balletbo i Serra wrote: > From: Bryan Freed > > When the I2C Infineon part is attached to an I2C adapter that imposes > a size limitation, large requests will fail with -EOPNOTSUPP. Retry > them with a sane minimum size without re-issuing the 0x05 command > as this appears to occasionally put the TPM in a bad state. > > Signed-off-by: Bryan Freed > [rework the patch to adapt to the feedback received] > Signed-off-by: Enric Balletbo i Serra > --- > This is a reworked version of the original patch based on the > suggestion made by Wolfram Sang to simply fall back to a sane minimum > when the maximum fails. > > Changes since v2: > - Do not remove faster transfers when chip is SLB9645 (Peter Huewe) > - Remember the adapterlimit length once it fails to not generate extra > i2c core messages (suggested by Andrew Lunn) > Changes since v1: > - Check the correct return value (-EOPNOTSUPP instead of -EINVAL) > - Fall back len to I2C_SMBUS_BLOCK_MAX if fails. > > drivers/char/tpm/tpm_i2c_infineon.c | 76 +++++++++++++++++++++++++++---------- > 1 file changed, 56 insertions(+), 20 deletions(-) > > diff --git a/drivers/char/tpm/tpm_i2c_infineon.c b/drivers/char/tpm/tpm_i2c_infineon.c > index 62ee44e..fdefcdb 100644 > --- a/drivers/char/tpm/tpm_i2c_infineon.c > +++ b/drivers/char/tpm/tpm_i2c_infineon.c > @@ -70,6 +70,7 @@ struct tpm_inf_dev { > u8 buf[TPM_BUFSIZE + sizeof(u8)]; /* max. buffer size + addr */ > struct tpm_chip *chip; > enum i2c_chip_type chip_type; > + unsigned int adapterlimit; > }; > > static struct tpm_inf_dev tpm_dev; > @@ -111,6 +112,7 @@ static int iic_tpm_read(u8 addr, u8 *buffer, size_t len) > > int rc = 0; > int count; > + unsigned int msglen = len; > > /* Lock the adapter for the duration of the whole sequence. */ > if (!tpm_dev.client->adapter->algo->master_xfer) > @@ -131,27 +133,61 @@ static int iic_tpm_read(u8 addr, u8 *buffer, size_t len) > usleep_range(SLEEP_DURATION_LOW, SLEEP_DURATION_HI); > } > } else { > - /* slb9635 protocol should work in all cases */ > - for (count = 0; count < MAX_COUNT; count++) { > - rc = __i2c_transfer(tpm_dev.client->adapter, &msg1, 1); > - if (rc > 0) > - break; /* break here to skip sleep */ > - > - usleep_range(SLEEP_DURATION_LOW, SLEEP_DURATION_HI); > - } > - > - if (rc <= 0) > - goto out; > - > - /* After the TPM has successfully received the register address > - * it needs some time, thus we're sleeping here again, before > - * retrieving the data > + /* Expect to send one command message and one data message, but > + * support looping over each or both if necessary. > */ > - for (count = 0; count < MAX_COUNT; count++) { > - usleep_range(SLEEP_DURATION_LOW, SLEEP_DURATION_HI); > - rc = __i2c_transfer(tpm_dev.client->adapter, &msg2, 1); > - if (rc > 0) > - break; > + while (len > 0) { > + /* slb9635 protocol should work in all cases */ > + for (count = 0; count < MAX_COUNT; count++) { > + rc = __i2c_transfer(tpm_dev.client->adapter, > + &msg1, 1); > + if (rc > 0) > + break; /* break here to skip sleep */ > + > + usleep_range(SLEEP_DURATION_LOW, > + SLEEP_DURATION_HI); > + } > + > + if (rc <= 0) > + goto out; > + > + /* After the TPM has successfully received the register > + * address it needs some time, thus we're sleeping here > + * again, before retrieving the data > + */ > + for (count = 0; count < MAX_COUNT; count++) { > + if (tpm_dev.adapterlimit) { > + msglen = min_t(unsigned int, > + tpm_dev.adapterlimit, > + len); > + msg2.len = msglen; > + } > + usleep_range(SLEEP_DURATION_LOW, > + SLEEP_DURATION_HI); > + rc = __i2c_transfer(tpm_dev.client->adapter, > + &msg2, 1); > + if (rc > 0) { > + /* Since len is unsigned, make doubly > + * sure we do not underflow it. > + */ > + if (msglen > len) > + len = 0; > + else > + len -= msglen; > + msg2.buf += msglen; > + break; > + } > + /* If the I2C adapter rejected the request (e.g > + * when the quirk read_max_len < len) fall back > + * to a sane minimum value and try again. > + */ > + if (rc == -EOPNOTSUPP) > + tpm_dev.adapterlimit = > + I2C_SMBUS_BLOCK_MAX; > + } > + > + if (rc <= 0) > + goto out; > } > } > > -- > 2.9.3 > Reviewed-by: Jarkko Sakkinen Peter, Andrew, anyone: Tested-by? /Jarkko