From patchwork Sun Dec 20 09:18:41 2015 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Vishal Verma X-Patchwork-Id: 7891441 Return-Path: X-Original-To: patchwork-linux-nvdimm@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork1.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.136]) by patchwork1.web.kernel.org (Postfix) with ESMTP id 6AAC69F349 for ; Sun, 20 Dec 2015 09:19:23 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id 36D742041D for ; Sun, 20 Dec 2015 09:19:22 +0000 (UTC) Received: from ml01.01.org (ml01.01.org [198.145.21.10]) (using TLSv1.2 with cipher DHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 0B3F42041A for ; Sun, 20 Dec 2015 09:19:21 +0000 (UTC) Received: from ml01.vlan14.01.org (localhost [IPv6:::1]) by ml01.01.org (Postfix) with ESMTP id 012F41A1FAF; Sun, 20 Dec 2015 01:19:21 -0800 (PST) X-Original-To: linux-nvdimm@lists.01.org Delivered-To: linux-nvdimm@lists.01.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.136]) by ml01.01.org (Postfix) with ESMTP id B60601A1FAF for ; Sun, 20 Dec 2015 01:19:19 -0800 (PST) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id 9AF8120481; Sun, 20 Dec 2015 09:19:18 +0000 (UTC) Received: from gamestarV3L.home. (71-212-131-94.hlrn.qwest.net [71.212.131.94]) (using TLSv1.2 with cipher DHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id BC0BF2047D; Sun, 20 Dec 2015 09:19:16 +0000 (UTC) From: vishal@kernel.org To: linux-nvdimm@lists.01.org Subject: [PATCH 2/3] libnvdimm: Add a poison list Date: Sun, 20 Dec 2015 02:18:41 -0700 Message-Id: <1450603122-7205-3-git-send-email-vishal@kernel.org> X-Mailer: git-send-email 2.5.0 In-Reply-To: <1450603122-7205-1-git-send-email-vishal@kernel.org> References: <1450603122-7205-1-git-send-email-vishal@kernel.org> X-Spam-Status: No, score=-2.6 required=5.0 tests=BAYES_00, RCVD_IN_DNSWL_LOW, RP_MATCHES_RCVD, UNPARSEABLE_RELAY autolearn=unavailable version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP Cc: linux-acpi@vger.kernel.org, linux-fsdevel@vger.kernel.org X-BeenThere: linux-nvdimm@lists.01.org X-Mailman-Version: 2.1.17 Precedence: list List-Id: "Linux-nvdimm developer list." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , MIME-Version: 1.0 Errors-To: linux-nvdimm-bounces@lists.01.org Sender: "Linux-nvdimm" X-Virus-Scanned: ClamAV using ClamSMTP From: Vishal Verma During region creation, perform Address Range Scrubs (ARS) for the SPA (System Physical Address) ranges to retrieve known poison locations from firmware. Add a new data structure 'nd_poison' which is used as a list in nvdimm_bus to store these poison locations for subsequent consumption by the pmem driver. Signed-off-by: Vishal Verma --- drivers/acpi/nfit.c | 193 ++++++++++++++++++++++++++++++++++++++++++++++ drivers/nvdimm/core.c | 63 +++++++++++++++ drivers/nvdimm/nd-core.h | 1 + drivers/nvdimm/nd.h | 6 ++ include/linux/libnvdimm.h | 2 + 5 files changed, 265 insertions(+) diff --git a/drivers/acpi/nfit.c b/drivers/acpi/nfit.c index e7ed39b..bd8726d 100644 --- a/drivers/acpi/nfit.c +++ b/drivers/acpi/nfit.c @@ -15,6 +15,7 @@ #include #include #include +#include #include #include #include @@ -1473,6 +1474,195 @@ static void acpi_nfit_blk_region_disable(struct nvdimm_bus *nvdimm_bus, /* devm will free nfit_blk */ } +static int ars_get_cap(struct nvdimm_bus_descriptor *nd_desc, + struct nd_cmd_ars_cap *cmd, u64 addr, u64 length) +{ + cmd->address = addr; + cmd->length = length; + + return nd_desc->ndctl(nd_desc, NULL, ND_CMD_ARS_CAP, cmd, + sizeof(*cmd)); +} + +static int ars_do_start(struct nvdimm_bus_descriptor *nd_desc, + struct nd_cmd_ars_start *cmd, u64 addr, u64 length) +{ + int rc; + + cmd->address = addr; + cmd->length = length; + cmd->type = ND_ARS_PERSISTENT; + + while (1) { + rc = nd_desc->ndctl(nd_desc, NULL, ND_CMD_ARS_START, cmd, + sizeof(*cmd)); + if (rc) + return rc; + switch (cmd->status) { + case 0: + return 0; + case 1: + /* ARS unsupported, but we should never get here */ + return 0; + case 2: + return -EINVAL; + case 3: + /* ARS is in progress */ + msleep(1000); + break; + default: + return -ENXIO; + } + } +} + +static int ars_get_status(struct nvdimm_bus_descriptor *nd_desc, + struct nd_cmd_ars_status *cmd) +{ + int rc; + + while (1) { + rc = nd_desc->ndctl(nd_desc, NULL, ND_CMD_ARS_STATUS, cmd, + sizeof(*cmd)); + if (rc || cmd->status & 0xffff) + return -ENXIO; + + /* Check extended status (Upper two bytes) */ + switch (cmd->status & 0xffff0000) { + case 0: + return 0; + case 1: + /* ARS is in progress */ + msleep(1000); + break; + case 2: + /* No ARS performed for the current boot */ + default: + return -ENXIO; + } + } +} + +static int ars_status_process_records(struct nvdimm_bus *nvdimm_bus, + struct nd_cmd_ars_status *ars_status, u64 start) +{ + int rc; + u32 i; + + /* + * The address field returned by ars_status should be either + * less than or equal to the address we last started ARS for. + * The (start, length) returned by ars_status should also have + * non-zero overlap with the range we started ARS for. + * If this is not the case, bail. + */ + if (ars_status->address > start || + (ars_status->address + ars_status->length < start)) + return -ENXIO; + + for (i = 0; i < ars_status->num_records; i++) { + rc = nvdimm_bus_add_poison(nvdimm_bus, + ars_status->records[i].err_address, + ars_status->records[i].length); + if (rc) + return rc; + } + + return 0; +} + +static int acpi_nfit_find_poison(struct acpi_nfit_desc *acpi_desc, + struct nd_region_desc *ndr_desc) +{ + struct nvdimm_bus_descriptor *nd_desc = &acpi_desc->nd_desc; + struct nvdimm_bus *nvdimm_bus = acpi_desc->nvdimm_bus; + struct nd_cmd_ars_status *ars_status = NULL; + struct nd_cmd_ars_start *ars_start = NULL; + struct nd_cmd_ars_cap *ars_cap = NULL; + u64 start, len, cur, remaining; + int rc; + + ars_cap = kzalloc(sizeof(*ars_cap), GFP_KERNEL); + if (!ars_cap) + return -ENOMEM; + + start = ndr_desc->res->start; + len = ndr_desc->res->end - ndr_desc->res->start + 1; + + rc = ars_get_cap(nd_desc, ars_cap, start, len); + if (rc) + goto out; + + /* If ARS is unsupported, skip this but continue initialization */ + if (ars_cap->status & 0xffff) { + dev_warn(acpi_desc->dev, + "ARS unsupported, won't construct an error list\n"); + goto out; + } + + /* + * Check if a full-range ARS has been run. If so, use those results + * without having to start a new ARS. + */ + ars_status = kzalloc(ars_cap->max_ars_out + sizeof(*ars_status), + GFP_KERNEL); + if (!ars_status) { + rc = -ENOMEM; + goto out; + } + + rc = ars_get_status(nd_desc, ars_status); + if (rc) + goto out; + + if (ars_status->address <= start && + (ars_status->address + ars_status->length >= start + len)) { + rc = ars_status_process_records(nvdimm_bus, ars_status, start); + goto out; + } + + /* + * ARS_STATUS can overflow if the number of poison entries found is + * greater than the maximum buffer size (ars_cap->max_ars_out) + * To detect overflow, check if the length field of ars_status + * is less than the length we supplied. If so, process the + * error entries we got, adjust the start point, and start again + */ + ars_start = kzalloc(sizeof(*ars_start), GFP_KERNEL); + if (!ars_start) + return -ENOMEM; + + cur = start; + remaining = len; + do { + u64 done, end; + + rc = ars_do_start(nd_desc, ars_start, cur, remaining); + if (rc) + goto out; + + rc = ars_get_status(nd_desc, ars_status); + if (rc) + goto out; + + rc = ars_status_process_records(nvdimm_bus, ars_status, cur); + if (rc) + goto out; + + end = min(cur + remaining, + ars_status->address + ars_status->length); + done = end - cur; + cur += done; + remaining -= done; + } while (remaining); + + out: + kfree(ars_cap); + kfree(ars_start); + kfree(ars_status); + return rc; +} + static int acpi_nfit_init_mapping(struct acpi_nfit_desc *acpi_desc, struct nd_mapping *nd_mapping, struct nd_region_desc *ndr_desc, struct acpi_nfit_memory_map *memdev, @@ -1585,6 +1775,9 @@ static int acpi_nfit_register_region(struct acpi_nfit_desc *acpi_desc, nvdimm_bus = acpi_desc->nvdimm_bus; if (nfit_spa_type(spa) == NFIT_SPA_PM) { + rc = acpi_nfit_find_poison(acpi_desc, ndr_desc); + if (rc) + return rc; if (!nvdimm_pmem_region_create(nvdimm_bus, ndr_desc)) return -ENOMEM; } else if (nfit_spa_type(spa) == NFIT_SPA_VOLATILE) { diff --git a/drivers/nvdimm/core.c b/drivers/nvdimm/core.c index 82c49bb..8c882c8 100644 --- a/drivers/nvdimm/core.c +++ b/drivers/nvdimm/core.c @@ -325,6 +325,7 @@ struct nvdimm_bus *__nvdimm_bus_register(struct device *parent, if (!nvdimm_bus) return NULL; INIT_LIST_HEAD(&nvdimm_bus->list); + INIT_LIST_HEAD(&nvdimm_bus->poison_list); init_waitqueue_head(&nvdimm_bus->probe_wait); nvdimm_bus->id = ida_simple_get(&nd_ida, 0, 0, GFP_KERNEL); mutex_init(&nvdimm_bus->reconfig_mutex); @@ -359,6 +360,67 @@ struct nvdimm_bus *__nvdimm_bus_register(struct device *parent, } EXPORT_SYMBOL_GPL(__nvdimm_bus_register); +struct list_head *nvdimm_bus_get_poison_list(struct nvdimm_bus *nvdimm_bus) +{ + return &nvdimm_bus->poison_list; +} +EXPORT_SYMBOL_GPL(nvdimm_bus_get_poison_list); + +static int __add_poison(struct nvdimm_bus *nvdimm_bus, u64 addr, u64 length) +{ + struct nd_poison *pl; + + pl = kzalloc(sizeof(*pl), GFP_KERNEL); + if (!pl) + return -ENOMEM; + + pl->start = addr; + pl->length = length; + list_add_tail(&pl->list, &nvdimm_bus->poison_list); + + return 0; +} + +int nvdimm_bus_add_poison(struct nvdimm_bus *nvdimm_bus, u64 addr, u64 length) +{ + struct nd_poison *pl; + + if (list_empty(&nvdimm_bus->poison_list)) + return __add_poison(nvdimm_bus, addr, length); + + /* + * There is a chance this is a duplicate, check for those first. + * This will be the common case as ARS_STATUS returns all known + * errors in the SPA space, and we can't query it per region + */ + list_for_each_entry(pl, &nvdimm_bus->poison_list, list) + if (pl->start == addr) { + /* If length has changed, update this list entry */ + if (pl->length != length) + pl->length = length; + return 0; + } + + /* + * If not a duplicate or a simple length update, add the entry as is, + * as any overlapping ranges will get resolved when the list is consumed + * and converted to badblocks + */ + return __add_poison(nvdimm_bus, addr, length); +} +EXPORT_SYMBOL_GPL(nvdimm_bus_add_poison); + +static void free_poison_list(struct list_head *poison_list) +{ + struct nd_poison *pl, *next; + + list_for_each_entry_safe(pl, next, poison_list, list) { + list_del(&pl->list); + kfree(pl); + } + list_del_init(poison_list); +} + static int child_unregister(struct device *dev, void *data) { /* @@ -385,6 +447,7 @@ void nvdimm_bus_unregister(struct nvdimm_bus *nvdimm_bus) nd_synchronize(); device_for_each_child(&nvdimm_bus->dev, NULL, child_unregister); + free_poison_list(&nvdimm_bus->poison_list); nvdimm_bus_destroy_ndctl(nvdimm_bus); device_unregister(&nvdimm_bus->dev); diff --git a/drivers/nvdimm/nd-core.h b/drivers/nvdimm/nd-core.h index 159aed5..29acdaa 100644 --- a/drivers/nvdimm/nd-core.h +++ b/drivers/nvdimm/nd-core.h @@ -30,6 +30,7 @@ struct nvdimm_bus { struct list_head list; struct device dev; int id, probe_active; + struct list_head poison_list; struct mutex reconfig_mutex; }; diff --git a/drivers/nvdimm/nd.h b/drivers/nvdimm/nd.h index 417e521..ba91fcd 100644 --- a/drivers/nvdimm/nd.h +++ b/drivers/nvdimm/nd.h @@ -38,6 +38,12 @@ enum { #endif }; +struct nd_poison { + u64 start; + u64 length; + struct list_head list; +}; + struct nvdimm_drvdata { struct device *dev; int nsindex_size; diff --git a/include/linux/libnvdimm.h b/include/linux/libnvdimm.h index 3f021dc..43b82a0 100644 --- a/include/linux/libnvdimm.h +++ b/include/linux/libnvdimm.h @@ -116,6 +116,8 @@ static inline struct nd_blk_region_desc *to_blk_region_desc( } +struct list_head *nvdimm_bus_get_poison_list(struct nvdimm_bus *nvdimm_bus); +int nvdimm_bus_add_poison(struct nvdimm_bus *nvdimm_bus, u64 addr, u64 length); struct nvdimm_bus *__nvdimm_bus_register(struct device *parent, struct nvdimm_bus_descriptor *nfit_desc, struct module *module); #define nvdimm_bus_register(parent, desc) \