From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751328AbbGKOqV (ORCPT ); Sat, 11 Jul 2015 10:46:21 -0400 Received: from mga09.intel.com ([134.134.136.24]:59668 "EHLO mga09.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751216AbbGKOqT (ORCPT ); Sat, 11 Jul 2015 10:46:19 -0400 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.15,453,1432623600"; d="scan'208";a="745136167" Subject: [PATCH v2 0/6] Correctness fixes for NFIT BLK I/O path From: Dan Williams To: linux-nvdimm@ml01.01.org Cc: linux-acpi@vger.kernel.org, ross.zwisler@linux.intel.com, "Rafael J. Wysocki" , linux-kernel@vger.kernel.org Date: Sat, 11 Jul 2015 10:40:37 -0400 Message-ID: <20150711142806.25013.16924.stgit@dwillia2-desk3.amr.corp.intel.com> User-Agent: StGit/0.17.1-8-g92dd MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Changes since v1 [1]: 1/ Fixup the unit test in response to the 4.2-rc1 merge (handle the new usage of ioremap_wt). 2/ Add unit test infrastructure for NFIT "flush" sub-table entries 3/ Fix the mock DSM implementation's usage of ENOTTY 4/ Replace usage of devm_ioremap_resource() with devm_ioremap_nocache() for mapping flush table addresses. Multiple DIMMs are allowed to share the same flush address. Also, the side effect of stray writes to a flush address are simply extra memory controller buffer flushes, so we don't strictly need exclusive resource ownership. 5/ Fixed the warning "unable to guarantee persistence of writes" to consider the case of a platform without PMEM API support, but *with* flush address entries. 6/ Fixed the "latch" flag retrieval code to proceed with a sane default in the case the "DIMM flags" DSM is not implemented. [1]: https://lists.01.org/pipermail/linux-nvdimm/2015-July/001499.html Summary: This set of fixes to the BLK I/O path is a bit larger than I would have liked for an -rc update, but they are necessary for correct operation of a BLK-mode namespace. Originally posted as part of the initial PMEM API submission at the end of May [2], they got deferred while we iterated on the details of memremap_pmem(), memcpy_to_pmem(), and wmb_pmem(). Now that those APIs are upstream these patches finish off enabling persistence guarantees for BLK-mode persistent memory. [2]: https://lists.01.org/pipermail/linux-nvdimm/2015-May/000929.html --- Dan Williams (3): tools/testing/nvdimm: mock ioremap_wt tools/testing/nvdimm: fix return code for unimplemented commands tools/testing/nvdimm: add mock acpi_nfit_flush_address entries to nfit_test Ross Zwisler (3): pmem: add maintainer for include/linux/pmem.h nfit: update block I/O path to use PMEM API nfit: add support for NVDIMM "latch" flag MAINTAINERS | 1 drivers/acpi/nfit.c | 130 ++++++++++++++++++++++++++++++++++--- drivers/acpi/nfit.h | 20 +++++- tools/testing/nvdimm/Kbuild | 3 + tools/testing/nvdimm/test/iomap.c | 27 ++++++++ tools/testing/nvdimm/test/nfit.c | 52 ++++++++++++++- 6 files changed, 217 insertions(+), 16 deletions(-)