From a586cb4937fa328a857c8d7fbefb986d347bd4a7 Mon Sep 17 00:00:00 2001 From: Yasunori Goto Date: Fri, 22 Sep 2017 16:46:26 +0900 Subject: nfit_test Make private definitions to command emulation Move private definitions to command emulation. These definitions were originally defined at include/uapi/linux/ndctl.h, but they are used at only nfit_test emulation now. Signed-off-by: Yasunori Goto Reviewed-by: Vishal Verma Signed-off-by: Dan Williams --- tools/testing/nvdimm/test/nfit_test.h | 47 +++++++++++++++++++++++++++++++++++ 1 file changed, 47 insertions(+) (limited to 'tools') diff --git a/tools/testing/nvdimm/test/nfit_test.h b/tools/testing/nvdimm/test/nfit_test.h index d3d63dd5ed38..52c83be9dcfa 100644 --- a/tools/testing/nvdimm/test/nfit_test.h +++ b/tools/testing/nvdimm/test/nfit_test.h @@ -32,6 +32,53 @@ struct nfit_test_resource { void *buf; }; +#define ND_TRANSLATE_SPA_STATUS_INVALID_SPA 2 + +/* nfit commands */ +enum nfit_cmd_num { + NFIT_CMD_TRANSLATE_SPA = 5, + NFIT_CMD_ARS_INJECT_SET = 7, + NFIT_CMD_ARS_INJECT_CLEAR = 8, + NFIT_CMD_ARS_INJECT_GET = 9, +}; + +struct nd_cmd_translate_spa { + __u64 spa; + __u32 status; + __u8 flags; + __u8 _reserved[3]; + __u64 translate_length; + __u32 num_nvdimms; + struct nd_nvdimm_device { + __u32 nfit_device_handle; + __u32 _reserved; + __u64 dpa; + } __packed devices[0]; + +} __packed; + +struct nd_cmd_ars_err_inj { + __u64 err_inj_spa_range_base; + __u64 err_inj_spa_range_length; + __u8 err_inj_options; + __u32 status; +} __packed; + +struct nd_cmd_ars_err_inj_clr { + __u64 err_inj_clr_spa_range_base; + __u64 err_inj_clr_spa_range_length; + __u32 status; +} __packed; + +struct nd_cmd_ars_err_inj_stat { + __u32 status; + __u32 inj_err_rec_count; + struct nd_error_stat_query_record { + __u64 err_inj_stat_spa_range_base; + __u64 err_inj_stat_spa_range_length; + } __packed record[0]; +} __packed; + union acpi_object; typedef void *acpi_handle; -- cgit v1.2.3-58-ga151 From 10246dc84dfcabb05f2d29d442c9de967fd61270 Mon Sep 17 00:00:00 2001 From: Yasunori Goto Date: Fri, 22 Sep 2017 16:48:38 +0900 Subject: acpi nfit: nfit_test supports translate SPA To test ndctl list which use interface of Translate SPA, nfit_test needs to emulates it. This test module searches region which includes SPA and returns 1 dimm handle which is last one. Signed-off-by: Yasunori Goto Reviewed-by: Vishal Verma Signed-off-by: Dan Williams --- tools/testing/nvdimm/test/nfit.c | 104 ++++++++++++++++++++++++++++++++++++++- 1 file changed, 103 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/testing/nvdimm/test/nfit.c b/tools/testing/nvdimm/test/nfit.c index bef419d4266d..3a3bad348d56 100644 --- a/tools/testing/nvdimm/test/nfit.c +++ b/tools/testing/nvdimm/test/nfit.c @@ -342,6 +342,84 @@ static int nfit_test_cmd_clear_error(struct nd_cmd_clear_error *clear_err, return 0; } +struct region_search_spa { + u64 addr; + struct nd_region *region; +}; + +static int is_region_device(struct device *dev) +{ + return !strncmp(dev->kobj.name, "region", 6); +} + +static int nfit_test_search_region_spa(struct device *dev, void *data) +{ + struct region_search_spa *ctx = data; + struct nd_region *nd_region; + resource_size_t ndr_end; + + if (!is_region_device(dev)) + return 0; + + nd_region = to_nd_region(dev); + ndr_end = nd_region->ndr_start + nd_region->ndr_size; + + if (ctx->addr >= nd_region->ndr_start && ctx->addr < ndr_end) { + ctx->region = nd_region; + return 1; + } + + return 0; +} + +static int nfit_test_search_spa(struct nvdimm_bus *bus, + struct nd_cmd_translate_spa *spa) +{ + int ret; + struct nd_region *nd_region = NULL; + struct nvdimm *nvdimm = NULL; + struct nd_mapping *nd_mapping = NULL; + struct region_search_spa ctx = { + .addr = spa->spa, + .region = NULL, + }; + u64 dpa; + + ret = device_for_each_child(&bus->dev, &ctx, + nfit_test_search_region_spa); + + if (!ret) + return -ENODEV; + + nd_region = ctx.region; + + dpa = ctx.addr - nd_region->ndr_start; + + /* + * last dimm is selected for test + */ + nd_mapping = &nd_region->mapping[nd_region->ndr_mappings - 1]; + nvdimm = nd_mapping->nvdimm; + + spa->devices[0].nfit_device_handle = handle[nvdimm->id]; + spa->num_nvdimms = 1; + spa->devices[0].dpa = dpa; + + return 0; +} + +static int nfit_test_cmd_translate_spa(struct nvdimm_bus *bus, + struct nd_cmd_translate_spa *spa, unsigned int buf_len) +{ + if (buf_len < spa->translate_length) + return -EINVAL; + + if (nfit_test_search_spa(bus, spa) < 0 || !spa->num_nvdimms) + spa->status = 2; + + return 0; +} + static int nfit_test_cmd_smart(struct nd_cmd_smart *smart, unsigned int buf_len) { static const struct nd_smart_payload smart_data = { @@ -449,6 +527,26 @@ static int nfit_test_ctl(struct nvdimm_bus_descriptor *nd_desc, } } else { struct ars_state *ars_state = &t->ars_state; + struct nd_cmd_pkg *call_pkg = buf; + + if (!nd_desc) + return -ENOTTY; + + if (cmd == ND_CMD_CALL) { + func = call_pkg->nd_command; + + buf_len = call_pkg->nd_size_in + call_pkg->nd_size_out; + buf = (void *) call_pkg->nd_payload; + + switch (func) { + case NFIT_CMD_TRANSLATE_SPA: + rc = nfit_test_cmd_translate_spa( + acpi_desc->nvdimm_bus, buf, buf_len); + return rc; + default: + return -ENOTTY; + } + } if (!nd_desc || !test_bit(cmd, &nd_desc->cmd_mask)) return -ENOTTY; @@ -1430,7 +1528,9 @@ static void nfit_test0_setup(struct nfit_test *t) set_bit(ND_CMD_ARS_START, &acpi_desc->bus_cmd_force_en); set_bit(ND_CMD_ARS_STATUS, &acpi_desc->bus_cmd_force_en); set_bit(ND_CMD_CLEAR_ERROR, &acpi_desc->bus_cmd_force_en); + set_bit(ND_CMD_CALL, &acpi_desc->bus_cmd_force_en); set_bit(ND_CMD_SMART_THRESHOLD, &acpi_desc->dimm_cmd_force_en); + set_bit(NFIT_CMD_TRANSLATE_SPA, &acpi_desc->bus_nfit_cmd_force_en); } static void nfit_test1_setup(struct nfit_test *t) @@ -1613,10 +1713,12 @@ static int nfit_ctl_test(struct device *dev) .cmd_mask = 1UL << ND_CMD_ARS_CAP | 1UL << ND_CMD_ARS_START | 1UL << ND_CMD_ARS_STATUS - | 1UL << ND_CMD_CLEAR_ERROR, + | 1UL << ND_CMD_CLEAR_ERROR + | 1UL << ND_CMD_CALL, .module = THIS_MODULE, .provider_name = "ACPI.NFIT", .ndctl = acpi_nfit_ctl, + .bus_dsm_mask = 1UL << NFIT_CMD_TRANSLATE_SPA, }, .dev = &adev->dev, }; -- cgit v1.2.3-58-ga151 From aa9ad44a42b4cf4387f8ecddaf8e51707fdcda5a Mon Sep 17 00:00:00 2001 From: Dave Jiang Date: Wed, 23 Aug 2017 12:48:26 -0700 Subject: libnvdimm: move poison list functions to a new 'badrange' file nfit_test needs to use the poison list manipulation code as well. Make it more generic and in the process rename poison to badrange, and move all the related helpers to a new file. Signed-off-by: Dave Jiang [vishal: Add badrange.o to nfit_test's Kbuild] [vishal: add a missed include in bus.c for the new badrange functions] [vishal: rename all instances of 'be' to 'bre'] Signed-off-by: Vishal Verma Signed-off-by: Dan Williams --- drivers/acpi/nfit/core.c | 2 +- drivers/acpi/nfit/mce.c | 2 +- drivers/nvdimm/Makefile | 1 + drivers/nvdimm/badrange.c | 294 ++++++++++++++++++++++++++++++++++++++++++++ drivers/nvdimm/bus.c | 24 ++-- drivers/nvdimm/core.c | 260 +-------------------------------------- drivers/nvdimm/nd-core.h | 3 +- drivers/nvdimm/nd.h | 6 - include/linux/libnvdimm.h | 21 +++- tools/testing/nvdimm/Kbuild | 1 + 10 files changed, 332 insertions(+), 282 deletions(-) create mode 100644 drivers/nvdimm/badrange.c (limited to 'tools') diff --git a/drivers/acpi/nfit/core.c b/drivers/acpi/nfit/core.c index aa75d67ceee9..8043bfde7c63 100644 --- a/drivers/acpi/nfit/core.c +++ b/drivers/acpi/nfit/core.c @@ -2486,7 +2486,7 @@ static int ars_status_process_records(struct acpi_nfit_desc *acpi_desc, if (ars_status->out_length < 44 + sizeof(struct nd_ars_record) * (i + 1)) break; - rc = nvdimm_bus_add_poison(nvdimm_bus, + rc = nvdimm_bus_add_badrange(nvdimm_bus, ars_status->records[i].err_address, ars_status->records[i].length); if (rc) diff --git a/drivers/acpi/nfit/mce.c b/drivers/acpi/nfit/mce.c index feeb95d574fa..b92921439657 100644 --- a/drivers/acpi/nfit/mce.c +++ b/drivers/acpi/nfit/mce.c @@ -67,7 +67,7 @@ static int nfit_handle_mce(struct notifier_block *nb, unsigned long val, continue; /* If this fails due to an -ENOMEM, there is little we can do */ - nvdimm_bus_add_poison(acpi_desc->nvdimm_bus, + nvdimm_bus_add_badrange(acpi_desc->nvdimm_bus, ALIGN(mce->addr, L1_CACHE_BYTES), L1_CACHE_BYTES); nvdimm_region_notify(nfit_spa->nd_region, diff --git a/drivers/nvdimm/Makefile b/drivers/nvdimm/Makefile index 909554c3f955..ca6d325438a3 100644 --- a/drivers/nvdimm/Makefile +++ b/drivers/nvdimm/Makefile @@ -20,6 +20,7 @@ libnvdimm-y += region_devs.o libnvdimm-y += region.o libnvdimm-y += namespace_devs.o libnvdimm-y += label.o +libnvdimm-y += badrange.o libnvdimm-$(CONFIG_ND_CLAIM) += claim.o libnvdimm-$(CONFIG_BTT) += btt_devs.o libnvdimm-$(CONFIG_NVDIMM_PFN) += pfn_devs.o diff --git a/drivers/nvdimm/badrange.c b/drivers/nvdimm/badrange.c new file mode 100644 index 000000000000..0b67dcf43234 --- /dev/null +++ b/drivers/nvdimm/badrange.c @@ -0,0 +1,294 @@ +/* + * Copyright(c) 2017 Intel Corporation. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of version 2 of the GNU General Public License as + * published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + */ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include "nd-core.h" +#include "nd.h" + +void badrange_init(struct badrange *badrange) +{ + INIT_LIST_HEAD(&badrange->list); + spin_lock_init(&badrange->lock); +} +EXPORT_SYMBOL_GPL(badrange_init); + +static void append_badrange_entry(struct badrange *badrange, + struct badrange_entry *bre, u64 addr, u64 length) +{ + lockdep_assert_held(&badrange->lock); + bre->start = addr; + bre->length = length; + list_add_tail(&bre->list, &badrange->list); +} + +static int alloc_and_append_badrange_entry(struct badrange *badrange, + u64 addr, u64 length, gfp_t flags) +{ + struct badrange_entry *bre; + + bre = kzalloc(sizeof(*bre), flags); + if (!bre) + return -ENOMEM; + + append_badrange_entry(badrange, bre, addr, length); + return 0; +} + +static int add_badrange(struct badrange *badrange, u64 addr, u64 length) +{ + struct badrange_entry *bre, *bre_new; + + spin_unlock(&badrange->lock); + bre_new = kzalloc(sizeof(*bre_new), GFP_KERNEL); + spin_lock(&badrange->lock); + + if (list_empty(&badrange->list)) { + if (!bre_new) + return -ENOMEM; + append_badrange_entry(badrange, bre_new, addr, length); + return 0; + } + + /* + * 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(bre, &badrange->list, list) + if (bre->start == addr) { + /* If length has changed, update this list entry */ + if (bre->length != length) + bre->length = length; + kfree(bre_new); + 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 + */ + if (!bre_new) + return -ENOMEM; + append_badrange_entry(badrange, bre_new, addr, length); + + return 0; +} + +int badrange_add(struct badrange *badrange, u64 addr, u64 length) +{ + int rc; + + spin_lock(&badrange->lock); + rc = add_badrange(badrange, addr, length); + spin_unlock(&badrange->lock); + + return rc; +} +EXPORT_SYMBOL_GPL(badrange_add); + +void badrange_forget(struct badrange *badrange, phys_addr_t start, + unsigned int len) +{ + struct list_head *badrange_list = &badrange->list; + u64 clr_end = start + len - 1; + struct badrange_entry *bre, *next; + + spin_lock(&badrange->lock); + WARN_ON_ONCE(list_empty(badrange_list)); + + /* + * [start, clr_end] is the badrange interval being cleared. + * [bre->start, bre_end] is the badrange_list entry we're comparing + * the above interval against. The badrange list entry may need + * to be modified (update either start or length), deleted, or + * split into two based on the overlap characteristics + */ + + list_for_each_entry_safe(bre, next, badrange_list, list) { + u64 bre_end = bre->start + bre->length - 1; + + /* Skip intervals with no intersection */ + if (bre_end < start) + continue; + if (bre->start > clr_end) + continue; + /* Delete completely overlapped badrange entries */ + if ((bre->start >= start) && (bre_end <= clr_end)) { + list_del(&bre->list); + kfree(bre); + continue; + } + /* Adjust start point of partially cleared entries */ + if ((start <= bre->start) && (clr_end > bre->start)) { + bre->length -= clr_end - bre->start + 1; + bre->start = clr_end + 1; + continue; + } + /* Adjust bre->length for partial clearing at the tail end */ + if ((bre->start < start) && (bre_end <= clr_end)) { + /* bre->start remains the same */ + bre->length = start - bre->start; + continue; + } + /* + * If clearing in the middle of an entry, we split it into + * two by modifying the current entry to represent one half of + * the split, and adding a new entry for the second half. + */ + if ((bre->start < start) && (bre_end > clr_end)) { + u64 new_start = clr_end + 1; + u64 new_len = bre_end - new_start + 1; + + /* Add new entry covering the right half */ + alloc_and_append_badrange_entry(badrange, new_start, + new_len, GFP_NOWAIT); + /* Adjust this entry to cover the left half */ + bre->length = start - bre->start; + continue; + } + } + spin_unlock(&badrange->lock); +} +EXPORT_SYMBOL_GPL(badrange_forget); + +static void set_badblock(struct badblocks *bb, sector_t s, int num) +{ + dev_dbg(bb->dev, "Found a bad range (0x%llx, 0x%llx)\n", + (u64) s * 512, (u64) num * 512); + /* this isn't an error as the hardware will still throw an exception */ + if (badblocks_set(bb, s, num, 1)) + dev_info_once(bb->dev, "%s: failed for sector %llx\n", + __func__, (u64) s); +} + +/** + * __add_badblock_range() - Convert a physical address range to bad sectors + * @bb: badblocks instance to populate + * @ns_offset: namespace offset where the error range begins (in bytes) + * @len: number of bytes of badrange to be added + * + * This assumes that the range provided with (ns_offset, len) is within + * the bounds of physical addresses for this namespace, i.e. lies in the + * interval [ns_start, ns_start + ns_size) + */ +static void __add_badblock_range(struct badblocks *bb, u64 ns_offset, u64 len) +{ + const unsigned int sector_size = 512; + sector_t start_sector, end_sector; + u64 num_sectors; + u32 rem; + + start_sector = div_u64(ns_offset, sector_size); + end_sector = div_u64_rem(ns_offset + len, sector_size, &rem); + if (rem) + end_sector++; + num_sectors = end_sector - start_sector; + + if (unlikely(num_sectors > (u64)INT_MAX)) { + u64 remaining = num_sectors; + sector_t s = start_sector; + + while (remaining) { + int done = min_t(u64, remaining, INT_MAX); + + set_badblock(bb, s, done); + remaining -= done; + s += done; + } + } else + set_badblock(bb, start_sector, num_sectors); +} + +static void badblocks_populate(struct badrange *badrange, + struct badblocks *bb, const struct resource *res) +{ + struct badrange_entry *bre; + + if (list_empty(&badrange->list)) + return; + + list_for_each_entry(bre, &badrange->list, list) { + u64 bre_end = bre->start + bre->length - 1; + + /* Discard intervals with no intersection */ + if (bre_end < res->start) + continue; + if (bre->start > res->end) + continue; + /* Deal with any overlap after start of the namespace */ + if (bre->start >= res->start) { + u64 start = bre->start; + u64 len; + + if (bre_end <= res->end) + len = bre->length; + else + len = res->start + resource_size(res) + - bre->start; + __add_badblock_range(bb, start - res->start, len); + continue; + } + /* + * Deal with overlap for badrange starting before + * the namespace. + */ + if (bre->start < res->start) { + u64 len; + + if (bre_end < res->end) + len = bre->start + bre->length - res->start; + else + len = resource_size(res); + __add_badblock_range(bb, 0, len); + } + } +} + +/** + * nvdimm_badblocks_populate() - Convert a list of badranges to badblocks + * @region: parent region of the range to interrogate + * @bb: badblocks instance to populate + * @res: resource range to consider + * + * The badrange list generated during bus initialization may contain + * multiple, possibly overlapping physical address ranges. Compare each + * of these ranges to the resource range currently being initialized, + * and add badblocks entries for all matching sub-ranges + */ +void nvdimm_badblocks_populate(struct nd_region *nd_region, + struct badblocks *bb, const struct resource *res) +{ + struct nvdimm_bus *nvdimm_bus; + + if (!is_memory(&nd_region->dev)) { + dev_WARN_ONCE(&nd_region->dev, 1, + "%s only valid for pmem regions\n", __func__); + return; + } + nvdimm_bus = walk_to_nvdimm_bus(&nd_region->dev); + + nvdimm_bus_lock(&nvdimm_bus->dev); + badblocks_populate(&nvdimm_bus->badrange, bb, res); + nvdimm_bus_unlock(&nvdimm_bus->dev); +} +EXPORT_SYMBOL_GPL(nvdimm_badblocks_populate); diff --git a/drivers/nvdimm/bus.c b/drivers/nvdimm/bus.c index baf283986a7e..0a5e6cd758fe 100644 --- a/drivers/nvdimm/bus.c +++ b/drivers/nvdimm/bus.c @@ -11,6 +11,7 @@ * General Public License for more details. */ #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt +#include #include #include #include @@ -221,7 +222,7 @@ static void nvdimm_account_cleared_poison(struct nvdimm_bus *nvdimm_bus, phys_addr_t phys, u64 cleared) { if (cleared > 0) - nvdimm_forget_poison(nvdimm_bus, phys, cleared); + badrange_forget(&nvdimm_bus->badrange, phys, cleared); if (cleared > 0 && cleared / 512) nvdimm_clear_badblocks_regions(nvdimm_bus, phys, cleared); @@ -344,11 +345,10 @@ struct nvdimm_bus *nvdimm_bus_register(struct device *parent, return NULL; INIT_LIST_HEAD(&nvdimm_bus->list); INIT_LIST_HEAD(&nvdimm_bus->mapping_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); - spin_lock_init(&nvdimm_bus->poison_lock); + badrange_init(&nvdimm_bus->badrange); if (nvdimm_bus->id < 0) { kfree(nvdimm_bus); return NULL; @@ -395,15 +395,15 @@ static int child_unregister(struct device *dev, void *data) return 0; } -static void free_poison_list(struct list_head *poison_list) +static void free_badrange_list(struct list_head *badrange_list) { - struct nd_poison *pl, *next; + struct badrange_entry *bre, *next; - list_for_each_entry_safe(pl, next, poison_list, list) { - list_del(&pl->list); - kfree(pl); + list_for_each_entry_safe(bre, next, badrange_list, list) { + list_del(&bre->list); + kfree(bre); } - list_del_init(poison_list); + list_del_init(badrange_list); } static int nd_bus_remove(struct device *dev) @@ -417,9 +417,9 @@ static int nd_bus_remove(struct device *dev) nd_synchronize(); device_for_each_child(&nvdimm_bus->dev, NULL, child_unregister); - spin_lock(&nvdimm_bus->poison_lock); - free_poison_list(&nvdimm_bus->poison_list); - spin_unlock(&nvdimm_bus->poison_lock); + spin_lock(&nvdimm_bus->badrange.lock); + free_badrange_list(&nvdimm_bus->badrange.list); + spin_unlock(&nvdimm_bus->badrange.lock); nvdimm_bus_destroy_ndctl(nvdimm_bus); diff --git a/drivers/nvdimm/core.c b/drivers/nvdimm/core.c index bb71f0cf8f5d..1dc527660637 100644 --- a/drivers/nvdimm/core.c +++ b/drivers/nvdimm/core.c @@ -398,265 +398,11 @@ struct attribute_group nvdimm_bus_attribute_group = { }; EXPORT_SYMBOL_GPL(nvdimm_bus_attribute_group); -static void set_badblock(struct badblocks *bb, sector_t s, int num) +int nvdimm_bus_add_badrange(struct nvdimm_bus *nvdimm_bus, u64 addr, u64 length) { - dev_dbg(bb->dev, "Found a poison range (0x%llx, 0x%llx)\n", - (u64) s * 512, (u64) num * 512); - /* this isn't an error as the hardware will still throw an exception */ - if (badblocks_set(bb, s, num, 1)) - dev_info_once(bb->dev, "%s: failed for sector %llx\n", - __func__, (u64) s); + return badrange_add(&nvdimm_bus->badrange, addr, length); } - -/** - * __add_badblock_range() - Convert a physical address range to bad sectors - * @bb: badblocks instance to populate - * @ns_offset: namespace offset where the error range begins (in bytes) - * @len: number of bytes of poison to be added - * - * This assumes that the range provided with (ns_offset, len) is within - * the bounds of physical addresses for this namespace, i.e. lies in the - * interval [ns_start, ns_start + ns_size) - */ -static void __add_badblock_range(struct badblocks *bb, u64 ns_offset, u64 len) -{ - const unsigned int sector_size = 512; - sector_t start_sector, end_sector; - u64 num_sectors; - u32 rem; - - start_sector = div_u64(ns_offset, sector_size); - end_sector = div_u64_rem(ns_offset + len, sector_size, &rem); - if (rem) - end_sector++; - num_sectors = end_sector - start_sector; - - if (unlikely(num_sectors > (u64)INT_MAX)) { - u64 remaining = num_sectors; - sector_t s = start_sector; - - while (remaining) { - int done = min_t(u64, remaining, INT_MAX); - - set_badblock(bb, s, done); - remaining -= done; - s += done; - } - } else - set_badblock(bb, start_sector, num_sectors); -} - -static void badblocks_populate(struct list_head *poison_list, - struct badblocks *bb, const struct resource *res) -{ - struct nd_poison *pl; - - if (list_empty(poison_list)) - return; - - list_for_each_entry(pl, poison_list, list) { - u64 pl_end = pl->start + pl->length - 1; - - /* Discard intervals with no intersection */ - if (pl_end < res->start) - continue; - if (pl->start > res->end) - continue; - /* Deal with any overlap after start of the namespace */ - if (pl->start >= res->start) { - u64 start = pl->start; - u64 len; - - if (pl_end <= res->end) - len = pl->length; - else - len = res->start + resource_size(res) - - pl->start; - __add_badblock_range(bb, start - res->start, len); - continue; - } - /* Deal with overlap for poison starting before the namespace */ - if (pl->start < res->start) { - u64 len; - - if (pl_end < res->end) - len = pl->start + pl->length - res->start; - else - len = resource_size(res); - __add_badblock_range(bb, 0, len); - } - } -} - -/** - * nvdimm_badblocks_populate() - Convert a list of poison ranges to badblocks - * @region: parent region of the range to interrogate - * @bb: badblocks instance to populate - * @res: resource range to consider - * - * The poison list generated during bus initialization may contain - * multiple, possibly overlapping physical address ranges. Compare each - * of these ranges to the resource range currently being initialized, - * and add badblocks entries for all matching sub-ranges - */ -void nvdimm_badblocks_populate(struct nd_region *nd_region, - struct badblocks *bb, const struct resource *res) -{ - struct nvdimm_bus *nvdimm_bus; - struct list_head *poison_list; - - if (!is_memory(&nd_region->dev)) { - dev_WARN_ONCE(&nd_region->dev, 1, - "%s only valid for pmem regions\n", __func__); - return; - } - nvdimm_bus = walk_to_nvdimm_bus(&nd_region->dev); - poison_list = &nvdimm_bus->poison_list; - - nvdimm_bus_lock(&nvdimm_bus->dev); - badblocks_populate(poison_list, bb, res); - nvdimm_bus_unlock(&nvdimm_bus->dev); -} -EXPORT_SYMBOL_GPL(nvdimm_badblocks_populate); - -static void append_poison_entry(struct nvdimm_bus *nvdimm_bus, - struct nd_poison *pl, u64 addr, u64 length) -{ - lockdep_assert_held(&nvdimm_bus->poison_lock); - pl->start = addr; - pl->length = length; - list_add_tail(&pl->list, &nvdimm_bus->poison_list); -} - -static int add_poison(struct nvdimm_bus *nvdimm_bus, u64 addr, u64 length, - gfp_t flags) -{ - struct nd_poison *pl; - - pl = kzalloc(sizeof(*pl), flags); - if (!pl) - return -ENOMEM; - - append_poison_entry(nvdimm_bus, pl, addr, length); - return 0; -} - -static int bus_add_poison(struct nvdimm_bus *nvdimm_bus, u64 addr, u64 length) -{ - struct nd_poison *pl, *pl_new; - - spin_unlock(&nvdimm_bus->poison_lock); - pl_new = kzalloc(sizeof(*pl_new), GFP_KERNEL); - spin_lock(&nvdimm_bus->poison_lock); - - if (list_empty(&nvdimm_bus->poison_list)) { - if (!pl_new) - return -ENOMEM; - append_poison_entry(nvdimm_bus, pl_new, addr, length); - return 0; - } - - /* - * 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; - kfree(pl_new); - 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 - */ - if (!pl_new) - return -ENOMEM; - append_poison_entry(nvdimm_bus, pl_new, addr, length); - - return 0; -} - -int nvdimm_bus_add_poison(struct nvdimm_bus *nvdimm_bus, u64 addr, u64 length) -{ - int rc; - - spin_lock(&nvdimm_bus->poison_lock); - rc = bus_add_poison(nvdimm_bus, addr, length); - spin_unlock(&nvdimm_bus->poison_lock); - - return rc; -} -EXPORT_SYMBOL_GPL(nvdimm_bus_add_poison); - -void nvdimm_forget_poison(struct nvdimm_bus *nvdimm_bus, phys_addr_t start, - unsigned int len) -{ - struct list_head *poison_list = &nvdimm_bus->poison_list; - u64 clr_end = start + len - 1; - struct nd_poison *pl, *next; - - spin_lock(&nvdimm_bus->poison_lock); - WARN_ON_ONCE(list_empty(poison_list)); - - /* - * [start, clr_end] is the poison interval being cleared. - * [pl->start, pl_end] is the poison_list entry we're comparing - * the above interval against. The poison list entry may need - * to be modified (update either start or length), deleted, or - * split into two based on the overlap characteristics - */ - - list_for_each_entry_safe(pl, next, poison_list, list) { - u64 pl_end = pl->start + pl->length - 1; - - /* Skip intervals with no intersection */ - if (pl_end < start) - continue; - if (pl->start > clr_end) - continue; - /* Delete completely overlapped poison entries */ - if ((pl->start >= start) && (pl_end <= clr_end)) { - list_del(&pl->list); - kfree(pl); - continue; - } - /* Adjust start point of partially cleared entries */ - if ((start <= pl->start) && (clr_end > pl->start)) { - pl->length -= clr_end - pl->start + 1; - pl->start = clr_end + 1; - continue; - } - /* Adjust pl->length for partial clearing at the tail end */ - if ((pl->start < start) && (pl_end <= clr_end)) { - /* pl->start remains the same */ - pl->length = start - pl->start; - continue; - } - /* - * If clearing in the middle of an entry, we split it into - * two by modifying the current entry to represent one half of - * the split, and adding a new entry for the second half. - */ - if ((pl->start < start) && (pl_end > clr_end)) { - u64 new_start = clr_end + 1; - u64 new_len = pl_end - new_start + 1; - - /* Add new entry covering the right half */ - add_poison(nvdimm_bus, new_start, new_len, GFP_NOWAIT); - /* Adjust this entry to cover the left half */ - pl->length = start - pl->start; - continue; - } - } - spin_unlock(&nvdimm_bus->poison_lock); -} -EXPORT_SYMBOL_GPL(nvdimm_forget_poison); +EXPORT_SYMBOL_GPL(nvdimm_bus_add_badrange); #ifdef CONFIG_BLK_DEV_INTEGRITY int nd_integrity_init(struct gendisk *disk, unsigned long meta_size) diff --git a/drivers/nvdimm/nd-core.h b/drivers/nvdimm/nd-core.h index 86bc19ae30da..79274ead54fb 100644 --- a/drivers/nvdimm/nd-core.h +++ b/drivers/nvdimm/nd-core.h @@ -29,10 +29,9 @@ struct nvdimm_bus { struct list_head list; struct device dev; int id, probe_active; - struct list_head poison_list; struct list_head mapping_list; struct mutex reconfig_mutex; - spinlock_t poison_lock; + struct badrange badrange; }; struct nvdimm { diff --git a/drivers/nvdimm/nd.h b/drivers/nvdimm/nd.h index 156be00e1f76..e958f3724c41 100644 --- a/drivers/nvdimm/nd.h +++ b/drivers/nvdimm/nd.h @@ -34,12 +34,6 @@ enum { NVDIMM_IO_ATOMIC = 1, }; -struct nd_poison { - u64 start; - u64 length; - struct list_head list; -}; - struct nvdimm_drvdata { struct device *dev; int nslabel_size; diff --git a/include/linux/libnvdimm.h b/include/linux/libnvdimm.h index 3eaad2fbf284..f8109ddb5ef1 100644 --- a/include/linux/libnvdimm.h +++ b/include/linux/libnvdimm.h @@ -18,6 +18,18 @@ #include #include #include +#include + +struct badrange_entry { + u64 start; + u64 length; + struct list_head list; +}; + +struct badrange { + struct list_head list; + spinlock_t lock; +}; enum { /* when a dimm supports both PMEM and BLK access a label is required */ @@ -129,9 +141,12 @@ static inline struct nd_blk_region_desc *to_blk_region_desc( } -int nvdimm_bus_add_poison(struct nvdimm_bus *nvdimm_bus, u64 addr, u64 length); -void nvdimm_forget_poison(struct nvdimm_bus *nvdimm_bus, - phys_addr_t start, unsigned int len); +void badrange_init(struct badrange *badrange); +int badrange_add(struct badrange *badrange, u64 addr, u64 length); +void badrange_forget(struct badrange *badrange, phys_addr_t start, + unsigned int len); +int nvdimm_bus_add_badrange(struct nvdimm_bus *nvdimm_bus, u64 addr, + u64 length); struct nvdimm_bus *nvdimm_bus_register(struct device *parent, struct nvdimm_bus_descriptor *nfit_desc); void nvdimm_bus_unregister(struct nvdimm_bus *nvdimm_bus); diff --git a/tools/testing/nvdimm/Kbuild b/tools/testing/nvdimm/Kbuild index d870520da68b..3272ab5a5cb9 100644 --- a/tools/testing/nvdimm/Kbuild +++ b/tools/testing/nvdimm/Kbuild @@ -69,6 +69,7 @@ libnvdimm-y += $(NVDIMM_SRC)/region_devs.o libnvdimm-y += $(NVDIMM_SRC)/region.o libnvdimm-y += $(NVDIMM_SRC)/namespace_devs.o libnvdimm-y += $(NVDIMM_SRC)/label.o +libnvdimm-y += $(NVDIMM_SRC)/badrange.o libnvdimm-$(CONFIG_ND_CLAIM) += $(NVDIMM_SRC)/claim.o libnvdimm-$(CONFIG_BTT) += $(NVDIMM_SRC)/btt_devs.o libnvdimm-$(CONFIG_NVDIMM_PFN) += $(NVDIMM_SRC)/pfn_devs.o -- cgit v1.2.3-58-ga151 From 9fb1a1903345fea598f48277576a3589a972b72e Mon Sep 17 00:00:00 2001 From: Dave Jiang Date: Mon, 30 Oct 2017 13:22:20 -0700 Subject: nfit_test: add error injection DSMs Add nfit_test emulation for the new ACPI 6.2 error injectino DSMs. This will allow unit tests to selectively inject the errors they wish to test for. Signed-off-by: Dave Jiang [vishal: Move injection functions to ND_CMD_CALL] [vishal: Add support for the notification option] [vishal: move an nfit_test private definition into a local header] Signed-off-by: Vishal Verma Signed-off-by: Dan Williams --- tools/testing/nvdimm/test/nfit.c | 187 +++++++++++++++++++++++++++++----- tools/testing/nvdimm/test/nfit_test.h | 5 + 2 files changed, 168 insertions(+), 24 deletions(-) (limited to 'tools') diff --git a/tools/testing/nvdimm/test/nfit.c b/tools/testing/nvdimm/test/nfit.c index 3a3bad348d56..788b1a5d6c5a 100644 --- a/tools/testing/nvdimm/test/nfit.c +++ b/tools/testing/nvdimm/test/nfit.c @@ -168,8 +168,12 @@ struct nfit_test { spinlock_t lock; } ars_state; struct device *dimm_dev[NUM_DCR]; + struct badrange badrange; + struct work_struct work; }; +static struct workqueue_struct *nfit_wq; + static struct nfit_test *to_nfit_test(struct device *dev) { struct platform_device *pdev = to_platform_device(dev); @@ -234,48 +238,68 @@ static int nfit_test_cmd_set_config_data(struct nd_cmd_set_config_hdr *nd_cmd, return rc; } -#define NFIT_TEST_ARS_RECORDS 4 #define NFIT_TEST_CLEAR_ERR_UNIT 256 static int nfit_test_cmd_ars_cap(struct nd_cmd_ars_cap *nd_cmd, unsigned int buf_len) { + int ars_recs; + if (buf_len < sizeof(*nd_cmd)) return -EINVAL; + /* for testing, only store up to n records that fit within 4k */ + ars_recs = SZ_4K / sizeof(struct nd_ars_record); + nd_cmd->max_ars_out = sizeof(struct nd_cmd_ars_status) - + NFIT_TEST_ARS_RECORDS * sizeof(struct nd_ars_record); + + ars_recs * sizeof(struct nd_ars_record); nd_cmd->status = (ND_ARS_PERSISTENT | ND_ARS_VOLATILE) << 16; nd_cmd->clear_err_unit = NFIT_TEST_CLEAR_ERR_UNIT; return 0; } -/* - * Initialize the ars_state to return an ars_result 1 second in the future with - * a 4K error range in the middle of the requested address range. - */ -static void post_ars_status(struct ars_state *ars_state, u64 addr, u64 len) +static void post_ars_status(struct ars_state *ars_state, + struct badrange *badrange, u64 addr, u64 len) { struct nd_cmd_ars_status *ars_status; struct nd_ars_record *ars_record; + struct badrange_entry *be; + u64 end = addr + len - 1; + int i = 0; ars_state->deadline = jiffies + 1*HZ; ars_status = ars_state->ars_status; ars_status->status = 0; - ars_status->out_length = sizeof(struct nd_cmd_ars_status) - + sizeof(struct nd_ars_record); ars_status->address = addr; ars_status->length = len; ars_status->type = ND_ARS_PERSISTENT; - ars_status->num_records = 1; - ars_record = &ars_status->records[0]; - ars_record->handle = 0; - ars_record->err_address = addr + len / 2; - ars_record->length = SZ_4K; + + spin_lock(&badrange->lock); + list_for_each_entry(be, &badrange->list, list) { + u64 be_end = be->start + be->length - 1; + u64 rstart, rend; + + /* skip entries outside the range */ + if (be_end < addr || be->start > end) + continue; + + rstart = (be->start < addr) ? addr : be->start; + rend = (be_end < end) ? be_end : end; + ars_record = &ars_status->records[i]; + ars_record->handle = 0; + ars_record->err_address = rstart; + ars_record->length = rend - rstart + 1; + i++; + } + spin_unlock(&badrange->lock); + ars_status->num_records = i; + ars_status->out_length = sizeof(struct nd_cmd_ars_status) + + i * sizeof(struct nd_ars_record); } -static int nfit_test_cmd_ars_start(struct ars_state *ars_state, +static int nfit_test_cmd_ars_start(struct nfit_test *t, + struct ars_state *ars_state, struct nd_cmd_ars_start *ars_start, unsigned int buf_len, int *cmd_rc) { @@ -289,7 +313,7 @@ static int nfit_test_cmd_ars_start(struct ars_state *ars_state, } else { ars_start->status = 0; ars_start->scrub_time = 1; - post_ars_status(ars_state, ars_start->address, + post_ars_status(ars_state, &t->badrange, ars_start->address, ars_start->length); *cmd_rc = 0; } @@ -456,6 +480,93 @@ static int nfit_test_cmd_smart_threshold(struct nd_cmd_smart_threshold *smart_t, return 0; } +static void uc_error_notify(struct work_struct *work) +{ + struct nfit_test *t = container_of(work, typeof(*t), work); + + __acpi_nfit_notify(&t->pdev.dev, t, NFIT_NOTIFY_UC_MEMORY_ERROR); +} + +static int nfit_test_cmd_ars_error_inject(struct nfit_test *t, + struct nd_cmd_ars_err_inj *err_inj, unsigned int buf_len) +{ + int rc; + + if (buf_len < sizeof(*err_inj)) { + rc = -EINVAL; + goto err; + } + + if (err_inj->err_inj_spa_range_length <= 0) { + rc = -EINVAL; + goto err; + } + + rc = badrange_add(&t->badrange, err_inj->err_inj_spa_range_base, + err_inj->err_inj_spa_range_length); + if (rc < 0) + goto err; + + if (err_inj->err_inj_options & (1 << ND_ARS_ERR_INJ_OPT_NOTIFY)) + queue_work(nfit_wq, &t->work); + + err_inj->status = 0; + return 0; + +err: + err_inj->status = NFIT_ARS_INJECT_INVALID; + return rc; +} + +static int nfit_test_cmd_ars_inject_clear(struct nfit_test *t, + struct nd_cmd_ars_err_inj_clr *err_clr, unsigned int buf_len) +{ + int rc; + + if (buf_len < sizeof(*err_clr)) { + rc = -EINVAL; + goto err; + } + + if (err_clr->err_inj_clr_spa_range_length <= 0) { + rc = -EINVAL; + goto err; + } + + badrange_forget(&t->badrange, err_clr->err_inj_clr_spa_range_base, + err_clr->err_inj_clr_spa_range_length); + + err_clr->status = 0; + return 0; + +err: + err_clr->status = NFIT_ARS_INJECT_INVALID; + return rc; +} + +static int nfit_test_cmd_ars_inject_status(struct nfit_test *t, + struct nd_cmd_ars_err_inj_stat *err_stat, + unsigned int buf_len) +{ + struct badrange_entry *be; + int max = SZ_4K / sizeof(struct nd_error_stat_query_record); + int i = 0; + + err_stat->status = 0; + spin_lock(&t->badrange.lock); + list_for_each_entry(be, &t->badrange.list, list) { + err_stat->record[i].err_inj_stat_spa_range_base = be->start; + err_stat->record[i].err_inj_stat_spa_range_length = be->length; + i++; + if (i > max) + break; + } + spin_unlock(&t->badrange.lock); + err_stat->inj_err_rec_count = i; + + return 0; +} + static int nfit_test_ctl(struct nvdimm_bus_descriptor *nd_desc, struct nvdimm *nvdimm, unsigned int cmd, void *buf, unsigned int buf_len, int *cmd_rc) @@ -543,6 +654,18 @@ static int nfit_test_ctl(struct nvdimm_bus_descriptor *nd_desc, rc = nfit_test_cmd_translate_spa( acpi_desc->nvdimm_bus, buf, buf_len); return rc; + case NFIT_CMD_ARS_INJECT_SET: + rc = nfit_test_cmd_ars_error_inject(t, buf, + buf_len); + return rc; + case NFIT_CMD_ARS_INJECT_CLEAR: + rc = nfit_test_cmd_ars_inject_clear(t, buf, + buf_len); + return rc; + case NFIT_CMD_ARS_INJECT_GET: + rc = nfit_test_cmd_ars_inject_status(t, buf, + buf_len); + return rc; default: return -ENOTTY; } @@ -556,8 +679,8 @@ static int nfit_test_ctl(struct nvdimm_bus_descriptor *nd_desc, rc = nfit_test_cmd_ars_cap(buf, buf_len); break; case ND_CMD_ARS_START: - rc = nfit_test_cmd_ars_start(ars_state, buf, buf_len, - cmd_rc); + rc = nfit_test_cmd_ars_start(t, ars_state, buf, + buf_len, cmd_rc); break; case ND_CMD_ARS_STATUS: rc = nfit_test_cmd_ars_status(ars_state, buf, buf_len, @@ -664,10 +787,9 @@ static struct nfit_test_resource *nfit_test_lookup(resource_size_t addr) static int ars_state_init(struct device *dev, struct ars_state *ars_state) { + /* for testing, only store up to n records that fit within 4k */ ars_state->ars_status = devm_kzalloc(dev, - sizeof(struct nd_cmd_ars_status) - + sizeof(struct nd_ars_record) * NFIT_TEST_ARS_RECORDS, - GFP_KERNEL); + sizeof(struct nd_cmd_ars_status) + SZ_4K, GFP_KERNEL); if (!ars_state->ars_status) return -ENOMEM; spin_lock_init(&ars_state->lock); @@ -1517,7 +1639,8 @@ static void nfit_test0_setup(struct nfit_test *t) + i * sizeof(u64); } - post_ars_status(&t->ars_state, t->spa_set_dma[0], SPA0_SIZE); + post_ars_status(&t->ars_state, &t->badrange, t->spa_set_dma[0], + SPA0_SIZE); acpi_desc = &t->acpi_desc; set_bit(ND_CMD_GET_CONFIG_SIZE, &acpi_desc->dimm_cmd_force_en); @@ -1531,6 +1654,9 @@ static void nfit_test0_setup(struct nfit_test *t) set_bit(ND_CMD_CALL, &acpi_desc->bus_cmd_force_en); set_bit(ND_CMD_SMART_THRESHOLD, &acpi_desc->dimm_cmd_force_en); set_bit(NFIT_CMD_TRANSLATE_SPA, &acpi_desc->bus_nfit_cmd_force_en); + set_bit(NFIT_CMD_ARS_INJECT_SET, &acpi_desc->bus_nfit_cmd_force_en); + set_bit(NFIT_CMD_ARS_INJECT_CLEAR, &acpi_desc->bus_nfit_cmd_force_en); + set_bit(NFIT_CMD_ARS_INJECT_GET, &acpi_desc->bus_nfit_cmd_force_en); } static void nfit_test1_setup(struct nfit_test *t) @@ -1620,7 +1746,8 @@ static void nfit_test1_setup(struct nfit_test *t) dcr->code = NFIT_FIC_BYTE; dcr->windows = 0; - post_ars_status(&t->ars_state, t->spa_set_dma[0], SPA2_SIZE); + post_ars_status(&t->ars_state, &t->badrange, t->spa_set_dma[0], + SPA2_SIZE); acpi_desc = &t->acpi_desc; set_bit(ND_CMD_ARS_CAP, &acpi_desc->bus_cmd_force_en); @@ -1718,7 +1845,10 @@ static int nfit_ctl_test(struct device *dev) .module = THIS_MODULE, .provider_name = "ACPI.NFIT", .ndctl = acpi_nfit_ctl, - .bus_dsm_mask = 1UL << NFIT_CMD_TRANSLATE_SPA, + .bus_dsm_mask = 1UL << NFIT_CMD_TRANSLATE_SPA + | 1UL << NFIT_CMD_ARS_INJECT_SET + | 1UL << NFIT_CMD_ARS_INJECT_CLEAR + | 1UL << NFIT_CMD_ARS_INJECT_GET, }, .dev = &adev->dev, }; @@ -2017,6 +2147,10 @@ static __init int nfit_test_init(void) nfit_test_setup(nfit_test_lookup, nfit_test_evaluate_dsm); + nfit_wq = create_singlethread_workqueue("nfit"); + if (!nfit_wq) + return -ENOMEM; + nfit_test_dimm = class_create(THIS_MODULE, "nfit_test_dimm"); if (IS_ERR(nfit_test_dimm)) { rc = PTR_ERR(nfit_test_dimm); @@ -2033,6 +2167,7 @@ static __init int nfit_test_init(void) goto err_register; } INIT_LIST_HEAD(&nfit_test->resources); + badrange_init(&nfit_test->badrange); switch (i) { case 0: nfit_test->num_pm = NUM_PM; @@ -2068,6 +2203,7 @@ static __init int nfit_test_init(void) goto err_register; instances[i] = nfit_test; + INIT_WORK(&nfit_test->work, uc_error_notify); } rc = platform_driver_register(&nfit_test_driver); @@ -2076,6 +2212,7 @@ static __init int nfit_test_init(void) return 0; err_register: + destroy_workqueue(nfit_wq); for (i = 0; i < NUM_NFITS; i++) if (instances[i]) platform_device_unregister(&instances[i]->pdev); @@ -2091,6 +2228,8 @@ static __exit void nfit_test_exit(void) { int i; + flush_workqueue(nfit_wq); + destroy_workqueue(nfit_wq); for (i = 0; i < NUM_NFITS; i++) platform_device_unregister(&instances[i]->pdev); platform_driver_unregister(&nfit_test_driver); diff --git a/tools/testing/nvdimm/test/nfit_test.h b/tools/testing/nvdimm/test/nfit_test.h index 52c83be9dcfa..113b44675a71 100644 --- a/tools/testing/nvdimm/test/nfit_test.h +++ b/tools/testing/nvdimm/test/nfit_test.h @@ -33,6 +33,11 @@ struct nfit_test_resource { }; #define ND_TRANSLATE_SPA_STATUS_INVALID_SPA 2 +#define NFIT_ARS_INJECT_INVALID 2 + +enum err_inj_options { + ND_ARS_ERR_INJ_OPT_NOTIFY = 0, +}; /* nfit commands */ enum nfit_cmd_num { -- cgit v1.2.3-58-ga151 From 5e096ef3b29e6113a7aa6a7dc4871987efa99bcb Mon Sep 17 00:00:00 2001 From: Vishal Verma Date: Mon, 30 Oct 2017 13:22:23 -0700 Subject: nfit_test: when clearing poison, also remove badrange entries The injected badrange entries can only be cleared from the kernel's accounting by writing to the affected blocks, so when such a write sends the clear errror DSM to nfit_test, also clear the ranges from nfit_test's badrange list. This lets an 'ARS Inject error status' DSM to return the correct status, omitting the cleared ranges. Cc: Dave Jiang Cc: Dan Williams Signed-off-by: Vishal Verma Reviewed-by: Dave Jiang Signed-off-by: Dan Williams --- tools/testing/nvdimm/test/nfit.c | 12 ++++-------- 1 file changed, 4 insertions(+), 8 deletions(-) (limited to 'tools') diff --git a/tools/testing/nvdimm/test/nfit.c b/tools/testing/nvdimm/test/nfit.c index 788b1a5d6c5a..79c5cb23693f 100644 --- a/tools/testing/nvdimm/test/nfit.c +++ b/tools/testing/nvdimm/test/nfit.c @@ -344,7 +344,8 @@ static int nfit_test_cmd_ars_status(struct ars_state *ars_state, return 0; } -static int nfit_test_cmd_clear_error(struct nd_cmd_clear_error *clear_err, +static int nfit_test_cmd_clear_error(struct nfit_test *t, + struct nd_cmd_clear_error *clear_err, unsigned int buf_len, int *cmd_rc) { const u64 mask = NFIT_TEST_CLEAR_ERR_UNIT - 1; @@ -354,12 +355,7 @@ static int nfit_test_cmd_clear_error(struct nd_cmd_clear_error *clear_err, if ((clear_err->address & mask) || (clear_err->length & mask)) return -EINVAL; - /* - * Report 'all clear' success for all commands even though a new - * scrub will find errors again. This is enough to have the - * error removed from the 'badblocks' tracking in the pmem - * driver. - */ + badrange_forget(&t->badrange, clear_err->address, clear_err->length); clear_err->status = 0; clear_err->cleared = clear_err->length; *cmd_rc = 0; @@ -687,7 +683,7 @@ static int nfit_test_ctl(struct nvdimm_bus_descriptor *nd_desc, cmd_rc); break; case ND_CMD_CLEAR_ERROR: - rc = nfit_test_cmd_clear_error(buf, buf_len, cmd_rc); + rc = nfit_test_cmd_clear_error(t, buf, buf_len, cmd_rc); break; default: return -ENOTTY; -- cgit v1.2.3-58-ga151 From 1c9725974074a047f6080eecc62c50a8e840d050 Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Wed, 1 Nov 2017 16:36:30 +0100 Subject: mm: introduce MAP_SHARED_VALIDATE, a mechanism to safely define new mmap flags The mmap(2) syscall suffers from the ABI anti-pattern of not validating unknown flags. However, proposals like MAP_SYNC need a mechanism to define new behavior that is known to fail on older kernels without the support. Define a new MAP_SHARED_VALIDATE flag pattern that is guaranteed to fail on all legacy mmap implementations. It is worth noting that the original proposal was for a standalone MAP_VALIDATE flag. However, when that could not be supported by all archs Linus observed: I see why you *think* you want a bitmap. You think you want a bitmap because you want to make MAP_VALIDATE be part of MAP_SYNC etc, so that people can do ret = mmap(NULL, size, PROT_READ | PROT_WRITE, MAP_SHARED | MAP_SYNC, fd, 0); and "know" that MAP_SYNC actually takes. And I'm saying that whole wish is bogus. You're fundamentally depending on special semantics, just make it explicit. It's already not portable, so don't try to make it so. Rename that MAP_VALIDATE as MAP_SHARED_VALIDATE, make it have a value of 0x3, and make people do ret = mmap(NULL, size, PROT_READ | PROT_WRITE, MAP_SHARED_VALIDATE | MAP_SYNC, fd, 0); and then the kernel side is easier too (none of that random garbage playing games with looking at the "MAP_VALIDATE bit", but just another case statement in that map type thing. Boom. Done. Similar to ->fallocate() we also want the ability to validate the support for new flags on a per ->mmap() 'struct file_operations' instance basis. Towards that end arrange for flags to be generically validated against a mmap_supported_flags exported by 'struct file_operations'. By default all existing flags are implicitly supported, but new flags require MAP_SHARED_VALIDATE and per-instance-opt-in. Cc: Jan Kara Cc: Arnd Bergmann Cc: Andy Lutomirski Cc: Andrew Morton Suggested-by: Christoph Hellwig Suggested-by: Linus Torvalds Reviewed-by: Ross Zwisler Signed-off-by: Dan Williams Signed-off-by: Jan Kara Signed-off-by: Dan Williams --- arch/alpha/include/uapi/asm/mman.h | 1 + arch/mips/include/uapi/asm/mman.h | 1 + arch/parisc/include/uapi/asm/mman.h | 1 + arch/xtensa/include/uapi/asm/mman.h | 1 + include/linux/fs.h | 1 + include/linux/mman.h | 39 ++++++++++++++++++++++++++++ include/uapi/asm-generic/mman-common.h | 1 + mm/mmap.c | 15 +++++++++++ tools/include/uapi/asm-generic/mman-common.h | 1 + 9 files changed, 61 insertions(+) (limited to 'tools') diff --git a/arch/alpha/include/uapi/asm/mman.h b/arch/alpha/include/uapi/asm/mman.h index 3b26cc62dadb..f6d118aaedb9 100644 --- a/arch/alpha/include/uapi/asm/mman.h +++ b/arch/alpha/include/uapi/asm/mman.h @@ -11,6 +11,7 @@ #define MAP_SHARED 0x01 /* Share changes */ #define MAP_PRIVATE 0x02 /* Changes are private */ +#define MAP_SHARED_VALIDATE 0x03 /* share + validate extension flags */ #define MAP_TYPE 0x0f /* Mask for type of mapping (OSF/1 is _wrong_) */ #define MAP_FIXED 0x100 /* Interpret addr exactly */ #define MAP_ANONYMOUS 0x10 /* don't use a file */ diff --git a/arch/mips/include/uapi/asm/mman.h b/arch/mips/include/uapi/asm/mman.h index da3216007fe0..93268e4cd3c7 100644 --- a/arch/mips/include/uapi/asm/mman.h +++ b/arch/mips/include/uapi/asm/mman.h @@ -28,6 +28,7 @@ */ #define MAP_SHARED 0x001 /* Share changes */ #define MAP_PRIVATE 0x002 /* Changes are private */ +#define MAP_SHARED_VALIDATE 0x003 /* share + validate extension flags */ #define MAP_TYPE 0x00f /* Mask for type of mapping */ #define MAP_FIXED 0x010 /* Interpret addr exactly */ diff --git a/arch/parisc/include/uapi/asm/mman.h b/arch/parisc/include/uapi/asm/mman.h index 775b5d5e41a1..bca652aa1677 100644 --- a/arch/parisc/include/uapi/asm/mman.h +++ b/arch/parisc/include/uapi/asm/mman.h @@ -11,6 +11,7 @@ #define MAP_SHARED 0x01 /* Share changes */ #define MAP_PRIVATE 0x02 /* Changes are private */ +#define MAP_SHARED_VALIDATE 0x03 /* share + validate extension flags */ #define MAP_TYPE 0x03 /* Mask for type of mapping */ #define MAP_FIXED 0x04 /* Interpret addr exactly */ #define MAP_ANONYMOUS 0x10 /* don't use a file */ diff --git a/arch/xtensa/include/uapi/asm/mman.h b/arch/xtensa/include/uapi/asm/mman.h index b15b278aa314..9ab426374714 100644 --- a/arch/xtensa/include/uapi/asm/mman.h +++ b/arch/xtensa/include/uapi/asm/mman.h @@ -35,6 +35,7 @@ */ #define MAP_SHARED 0x001 /* Share changes */ #define MAP_PRIVATE 0x002 /* Changes are private */ +#define MAP_SHARED_VALIDATE 0x003 /* share + validate extension flags */ #define MAP_TYPE 0x00f /* Mask for type of mapping */ #define MAP_FIXED 0x010 /* Interpret addr exactly */ diff --git a/include/linux/fs.h b/include/linux/fs.h index 13dab191a23e..57added3201d 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1701,6 +1701,7 @@ struct file_operations { long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long); long (*compat_ioctl) (struct file *, unsigned int, unsigned long); int (*mmap) (struct file *, struct vm_area_struct *); + unsigned long mmap_supported_flags; int (*open) (struct inode *, struct file *); int (*flush) (struct file *, fl_owner_t id); int (*release) (struct inode *, struct file *); diff --git a/include/linux/mman.h b/include/linux/mman.h index edb6cf6a81ed..74452e3f2536 100644 --- a/include/linux/mman.h +++ b/include/linux/mman.h @@ -7,6 +7,45 @@ #include #include +/* + * Arrange for legacy / undefined architecture specific flags to be + * ignored by default in LEGACY_MAP_MASK. + */ +#ifndef MAP_32BIT +#define MAP_32BIT 0 +#endif +#ifndef MAP_HUGE_2MB +#define MAP_HUGE_2MB 0 +#endif +#ifndef MAP_HUGE_1GB +#define MAP_HUGE_1GB 0 +#endif +#ifndef MAP_UNINITIALIZED +#define MAP_UNINITIALIZED 0 +#endif + +/* + * The historical set of flags that all mmap implementations implicitly + * support when a ->mmap_validate() op is not provided in file_operations. + */ +#define LEGACY_MAP_MASK (MAP_SHARED \ + | MAP_PRIVATE \ + | MAP_FIXED \ + | MAP_ANONYMOUS \ + | MAP_DENYWRITE \ + | MAP_EXECUTABLE \ + | MAP_UNINITIALIZED \ + | MAP_GROWSDOWN \ + | MAP_LOCKED \ + | MAP_NORESERVE \ + | MAP_POPULATE \ + | MAP_NONBLOCK \ + | MAP_STACK \ + | MAP_HUGETLB \ + | MAP_32BIT \ + | MAP_HUGE_2MB \ + | MAP_HUGE_1GB) + extern int sysctl_overcommit_memory; extern int sysctl_overcommit_ratio; extern unsigned long sysctl_overcommit_kbytes; diff --git a/include/uapi/asm-generic/mman-common.h b/include/uapi/asm-generic/mman-common.h index 203268f9231e..8ce7f5a0800f 100644 --- a/include/uapi/asm-generic/mman-common.h +++ b/include/uapi/asm-generic/mman-common.h @@ -16,6 +16,7 @@ #define MAP_SHARED 0x01 /* Share changes */ #define MAP_PRIVATE 0x02 /* Changes are private */ +#define MAP_SHARED_VALIDATE 0x03 /* share + validate extension flags */ #define MAP_TYPE 0x0f /* Mask for type of mapping */ #define MAP_FIXED 0x10 /* Interpret addr exactly */ #define MAP_ANONYMOUS 0x20 /* don't use a file */ diff --git a/mm/mmap.c b/mm/mmap.c index 680506faceae..924839fac0e6 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -1387,9 +1387,24 @@ unsigned long do_mmap(struct file *file, unsigned long addr, if (file) { struct inode *inode = file_inode(file); + unsigned long flags_mask; + + flags_mask = LEGACY_MAP_MASK | file->f_op->mmap_supported_flags; switch (flags & MAP_TYPE) { case MAP_SHARED: + /* + * Force use of MAP_SHARED_VALIDATE with non-legacy + * flags. E.g. MAP_SYNC is dangerous to use with + * MAP_SHARED as you don't know which consistency model + * you will get. We silently ignore unsupported flags + * with MAP_SHARED to preserve backward compatibility. + */ + flags &= LEGACY_MAP_MASK; + /* fall through */ + case MAP_SHARED_VALIDATE: + if (flags & ~flags_mask) + return -EOPNOTSUPP; if ((prot&PROT_WRITE) && !(file->f_mode&FMODE_WRITE)) return -EACCES; diff --git a/tools/include/uapi/asm-generic/mman-common.h b/tools/include/uapi/asm-generic/mman-common.h index 203268f9231e..8ce7f5a0800f 100644 --- a/tools/include/uapi/asm-generic/mman-common.h +++ b/tools/include/uapi/asm-generic/mman-common.h @@ -16,6 +16,7 @@ #define MAP_SHARED 0x01 /* Share changes */ #define MAP_PRIVATE 0x02 /* Changes are private */ +#define MAP_SHARED_VALIDATE 0x03 /* share + validate extension flags */ #define MAP_TYPE 0x0f /* Mask for type of mapping */ #define MAP_FIXED 0x10 /* Interpret addr exactly */ #define MAP_ANONYMOUS 0x20 /* don't use a file */ -- cgit v1.2.3-58-ga151 From 41cb3301c04d7f0b046ed296a34ef261c456dde8 Mon Sep 17 00:00:00 2001 From: Vishal Verma Date: Tue, 7 Nov 2017 15:48:21 -0700 Subject: tools/testing/nvdimm: stricter bounds checking for error injection commands Ensure that the in/out sizes passed in the nd_cmd_package are sane for the fixed output size commands (i.e. inject error and clear injected error). Reported-by: Dariusz Dokupil Signed-off-by: Vishal Verma Signed-off-by: Dan Williams --- tools/testing/nvdimm/test/nfit.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/testing/nvdimm/test/nfit.c b/tools/testing/nvdimm/test/nfit.c index 79c5cb23693f..edab68c2e221 100644 --- a/tools/testing/nvdimm/test/nfit.c +++ b/tools/testing/nvdimm/test/nfit.c @@ -488,7 +488,7 @@ static int nfit_test_cmd_ars_error_inject(struct nfit_test *t, { int rc; - if (buf_len < sizeof(*err_inj)) { + if (buf_len != sizeof(*err_inj)) { rc = -EINVAL; goto err; } @@ -519,7 +519,7 @@ static int nfit_test_cmd_ars_inject_clear(struct nfit_test *t, { int rc; - if (buf_len < sizeof(*err_clr)) { + if (buf_len != sizeof(*err_clr)) { rc = -EINVAL; goto err; } -- cgit v1.2.3-58-ga151 From fb2a1748355161e050e9f49f1ea9a0ae707a148b Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Sun, 12 Nov 2017 14:54:23 -0800 Subject: tools/testing/nvdimm: unit test clear-error commands Validate command parsing in acpi_nfit_ctl for the clear error command. This tests for a crash condition introduced by commit 4b27db7e26cd "acpi, nfit: add support for the _LSI, _LSR, and _LSW label methods". Cc: Vishal Verma Signed-off-by: Dan Williams --- tools/testing/nvdimm/test/nfit.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) (limited to 'tools') diff --git a/tools/testing/nvdimm/test/nfit.c b/tools/testing/nvdimm/test/nfit.c index edab68c2e221..7217b2b953b5 100644 --- a/tools/testing/nvdimm/test/nfit.c +++ b/tools/testing/nvdimm/test/nfit.c @@ -1812,6 +1812,7 @@ static int nfit_ctl_test(struct device *dev) unsigned long mask, cmd_size, offset; union { struct nd_cmd_get_config_size cfg_size; + struct nd_cmd_clear_error clear_err; struct nd_cmd_ars_status ars_stat; struct nd_cmd_ars_cap ars_cap; char buf[sizeof(struct nd_cmd_ars_status) @@ -1995,6 +1996,23 @@ static int nfit_ctl_test(struct device *dev) return -EIO; } + /* test clear error */ + cmd_size = sizeof(cmds.clear_err); + cmds.clear_err = (struct nd_cmd_clear_error) { + .length = 512, + .cleared = 512, + }; + rc = setup_result(cmds.buf, cmd_size); + if (rc) + return rc; + rc = acpi_nfit_ctl(&acpi_desc->nd_desc, NULL, ND_CMD_CLEAR_ERROR, + cmds.buf, cmd_size, &cmd_rc); + if (rc < 0 || cmd_rc) { + dev_dbg(dev, "%s: failed at: %d rc: %d cmd_rc: %d\n", + __func__, __LINE__, rc, cmd_rc); + return -EIO; + } + return 0; } -- cgit v1.2.3-58-ga151