From 08e6b3c6e3a054f566367740c94b8c1d18e52056 Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Wed, 13 Jun 2018 09:08:36 -0700 Subject: libnvdimm: Introduce locked DIMM capacity support When a DIMM is locked its namespace label area may not be. Introduce the distinction of locked namespaces to allow namespace enumeration while the capacity is locked. Signed-off-by: Dan Williams --- drivers/nvdimm/dimm.c | 24 ++++++++++++++++++++++-- drivers/nvdimm/namespace_devs.c | 23 +++++++++++++++++++++++ drivers/nvdimm/nd.h | 1 + 3 files changed, 46 insertions(+), 2 deletions(-) (limited to 'drivers') diff --git a/drivers/nvdimm/dimm.c b/drivers/nvdimm/dimm.c index 233907889f96..6c8fb7590838 100644 --- a/drivers/nvdimm/dimm.c +++ b/drivers/nvdimm/dimm.c @@ -34,6 +34,9 @@ static int nvdimm_probe(struct device *dev) return rc; } + /* reset locked, to be validated below... */ + nvdimm_clear_locked(dev); + ndd = kzalloc(sizeof(*ndd), GFP_KERNEL); if (!ndd) return -ENOMEM; @@ -48,12 +51,30 @@ static int nvdimm_probe(struct device *dev) get_device(dev); kref_init(&ndd->kref); + /* + * EACCES failures reading the namespace label-area-properties + * are interpreted as the DIMM capacity being locked but the + * namespace labels themselves being accessible. + */ rc = nvdimm_init_nsarea(ndd); - if (rc == -EACCES) + if (rc == -EACCES) { + /* + * See nvdimm_namespace_common_probe() where we fail to + * allow namespaces to probe while the DIMM is locked, + * but we do allow for namespace enumeration. + */ nvdimm_set_locked(dev); + rc = 0; + } if (rc) goto err; + /* + * EACCES failures reading the namespace label-data are + * interpreted as the label area being locked in addition to the + * DIMM capacity. We fail the dimm probe to prevent regions from + * attempting to parse the label area. + */ rc = nvdimm_init_config_data(ndd); if (rc == -EACCES) nvdimm_set_locked(dev); @@ -72,7 +93,6 @@ static int nvdimm_probe(struct device *dev) if (rc == 0) nvdimm_set_aliasing(dev); } - nvdimm_clear_locked(dev); nvdimm_bus_unlock(dev); if (rc) diff --git a/drivers/nvdimm/namespace_devs.c b/drivers/nvdimm/namespace_devs.c index 28afdd668905..cb322f2bc605 100644 --- a/drivers/nvdimm/namespace_devs.c +++ b/drivers/nvdimm/namespace_devs.c @@ -1144,6 +1144,26 @@ resource_size_t nvdimm_namespace_capacity(struct nd_namespace_common *ndns) } EXPORT_SYMBOL(nvdimm_namespace_capacity); +bool nvdimm_namespace_locked(struct nd_namespace_common *ndns) +{ + int i; + bool locked = false; + struct device *dev = &ndns->dev; + struct nd_region *nd_region = to_nd_region(dev->parent); + + for (i = 0; i < nd_region->ndr_mappings; i++) { + struct nd_mapping *nd_mapping = &nd_region->mapping[i]; + struct nvdimm *nvdimm = nd_mapping->nvdimm; + + if (test_bit(NDD_LOCKED, &nvdimm->flags)) { + dev_dbg(dev, "%s locked\n", nvdimm_name(nvdimm)); + locked = true; + } + } + return locked; +} +EXPORT_SYMBOL(nvdimm_namespace_locked); + static ssize_t size_show(struct device *dev, struct device_attribute *attr, char *buf) { @@ -1695,6 +1715,9 @@ struct nd_namespace_common *nvdimm_namespace_common_probe(struct device *dev) } } + if (nvdimm_namespace_locked(ndns)) + return ERR_PTR(-EACCES); + size = nvdimm_namespace_capacity(ndns); if (size < ND_MIN_NAMESPACE_SIZE) { dev_dbg(&ndns->dev, "%pa, too small must be at least %#x\n", diff --git a/drivers/nvdimm/nd.h b/drivers/nvdimm/nd.h index 32e0364b48b9..9d17abd9f8d0 100644 --- a/drivers/nvdimm/nd.h +++ b/drivers/nvdimm/nd.h @@ -357,6 +357,7 @@ struct resource *nvdimm_allocate_dpa(struct nvdimm_drvdata *ndd, struct nd_label_id *label_id, resource_size_t start, resource_size_t n); resource_size_t nvdimm_namespace_capacity(struct nd_namespace_common *ndns); +bool nvdimm_namespace_locked(struct nd_namespace_common *ndns); struct nd_namespace_common *nvdimm_namespace_common_probe(struct device *dev); int nvdimm_namespace_attach_btt(struct nd_namespace_common *ndns); int nvdimm_namespace_detach_btt(struct nd_btt *nd_btt); -- cgit v1.2.3 From 099b07a25fff6ee1b37f3ca7f4d2b53186326db0 Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Wed, 13 Jun 2018 09:06:52 -0700 Subject: acpi, nfit: Prefer _DSM over _LSR for namespace label reads The _LSR method indicates locked status via error-code-3 returned in the _LSR payload. When any error is returned the payload of _LSR is truncated to a zero-length buffer. The _DSM path in comparison allows system software to retrieve the locked status *and* namespace label area contents. Signed-off-by: Dan Williams --- drivers/acpi/nfit/core.c | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/acpi/nfit/core.c b/drivers/acpi/nfit/core.c index 471402cee1f1..2be8373153ed 100644 --- a/drivers/acpi/nfit/core.c +++ b/drivers/acpi/nfit/core.c @@ -1697,7 +1697,7 @@ static int acpi_nfit_add_dimm(struct acpi_nfit_desc *acpi_desc, { struct acpi_device *adev, *adev_dimm; struct device *dev = acpi_desc->dev; - unsigned long dsm_mask; + unsigned long dsm_mask, label_mask; const guid_t *guid; int i; int family = -1; @@ -1769,6 +1769,16 @@ static int acpi_nfit_add_dimm(struct acpi_nfit_desc *acpi_desc, 1ULL << i)) set_bit(i, &nfit_mem->dsm_mask); + /* + * Prefer the NVDIMM_FAMILY_INTEL label read commands if present + * due to their better semantics handling locked capacity. + */ + label_mask = 1 << ND_CMD_GET_CONFIG_SIZE | 1 << ND_CMD_GET_CONFIG_DATA + | 1 << ND_CMD_SET_CONFIG_DATA; + if (family == NVDIMM_FAMILY_INTEL + && (dsm_mask & label_mask) == label_mask) + return 0; + if (acpi_nvdimm_has_method(adev_dimm, "_LSI") && acpi_nvdimm_has_method(adev_dimm, "_LSR")) { dev_dbg(dev, "%s: has _LSR\n", dev_name(&adev_dimm->dev)); -- cgit v1.2.3 From 12e3129e29b406c41bc89231092a20d79dbf802c Mon Sep 17 00:00:00 2001 From: Keith Busch Date: Tue, 24 Jul 2018 15:07:57 -0600 Subject: libnvdimm: Use max contiguous area for namespace size This patch will find the max contiguous area to determine the largest pmem namespace size that can be created. If the requested size exceeds the largest available, ENOSPC error will be returned. This fixes the allocation underrun error and wrong error return code that have otherwise been observed as the following kernel warning: WARNING: CPU: PID: at drivers/nvdimm/namespace_devs.c:913 size_store Fixes: a1f3e4d6a0c3 ("libnvdimm, region: update nd_region_available_dpa() for multi-pmem support") Cc: Signed-off-by: Keith Busch Reviewed-by: Vishal Verma Signed-off-by: Dave Jiang --- drivers/nvdimm/dimm_devs.c | 31 +++++++++++++++++++++++++++++++ drivers/nvdimm/namespace_devs.c | 6 +++--- drivers/nvdimm/nd-core.h | 8 ++++++++ drivers/nvdimm/region_devs.c | 24 ++++++++++++++++++++++++ 4 files changed, 66 insertions(+), 3 deletions(-) (limited to 'drivers') diff --git a/drivers/nvdimm/dimm_devs.c b/drivers/nvdimm/dimm_devs.c index 8d348b22ba45..863cabc35215 100644 --- a/drivers/nvdimm/dimm_devs.c +++ b/drivers/nvdimm/dimm_devs.c @@ -536,6 +536,37 @@ resource_size_t nd_blk_available_dpa(struct nd_region *nd_region) return info.available; } +/** + * nd_pmem_max_contiguous_dpa - For the given dimm+region, return the max + * contiguous unallocated dpa range. + * @nd_region: constrain available space check to this reference region + * @nd_mapping: container of dpa-resource-root + labels + */ +resource_size_t nd_pmem_max_contiguous_dpa(struct nd_region *nd_region, + struct nd_mapping *nd_mapping) +{ + struct nvdimm_drvdata *ndd = to_ndd(nd_mapping); + struct nvdimm_bus *nvdimm_bus; + resource_size_t max = 0; + struct resource *res; + + /* if a dimm is disabled the available capacity is zero */ + if (!ndd) + return 0; + + nvdimm_bus = walk_to_nvdimm_bus(ndd->dev); + if (__reserve_free_pmem(&nd_region->dev, nd_mapping->nvdimm)) + return 0; + for_each_dpa_resource(ndd, res) { + if (strcmp(res->name, "pmem-reserve") != 0) + continue; + if (resource_size(res) > max) + max = resource_size(res); + } + release_free_pmem(nvdimm_bus, nd_mapping); + return max; +} + /** * nd_pmem_available_dpa - for the given dimm+region account unallocated dpa * @nd_mapping: container of dpa-resource-root + labels diff --git a/drivers/nvdimm/namespace_devs.c b/drivers/nvdimm/namespace_devs.c index cb322f2bc605..4a4266250c28 100644 --- a/drivers/nvdimm/namespace_devs.c +++ b/drivers/nvdimm/namespace_devs.c @@ -799,7 +799,7 @@ static int merge_dpa(struct nd_region *nd_region, return 0; } -static int __reserve_free_pmem(struct device *dev, void *data) +int __reserve_free_pmem(struct device *dev, void *data) { struct nvdimm *nvdimm = data; struct nd_region *nd_region; @@ -836,7 +836,7 @@ static int __reserve_free_pmem(struct device *dev, void *data) return 0; } -static void release_free_pmem(struct nvdimm_bus *nvdimm_bus, +void release_free_pmem(struct nvdimm_bus *nvdimm_bus, struct nd_mapping *nd_mapping) { struct nvdimm_drvdata *ndd = to_ndd(nd_mapping); @@ -1032,7 +1032,7 @@ static ssize_t __size_store(struct device *dev, unsigned long long val) allocated += nvdimm_allocated_dpa(ndd, &label_id); } - available = nd_region_available_dpa(nd_region); + available = nd_region_allocatable_dpa(nd_region); if (val > available + allocated) return -ENOSPC; diff --git a/drivers/nvdimm/nd-core.h b/drivers/nvdimm/nd-core.h index 79274ead54fb..ac68072fb8cd 100644 --- a/drivers/nvdimm/nd-core.h +++ b/drivers/nvdimm/nd-core.h @@ -100,6 +100,14 @@ struct nd_region; struct nvdimm_drvdata; struct nd_mapping; void nd_mapping_free_labels(struct nd_mapping *nd_mapping); + +int __reserve_free_pmem(struct device *dev, void *data); +void release_free_pmem(struct nvdimm_bus *nvdimm_bus, + struct nd_mapping *nd_mapping); + +resource_size_t nd_pmem_max_contiguous_dpa(struct nd_region *nd_region, + struct nd_mapping *nd_mapping); +resource_size_t nd_region_allocatable_dpa(struct nd_region *nd_region); resource_size_t nd_pmem_available_dpa(struct nd_region *nd_region, struct nd_mapping *nd_mapping, resource_size_t *overlap); resource_size_t nd_blk_available_dpa(struct nd_region *nd_region); diff --git a/drivers/nvdimm/region_devs.c b/drivers/nvdimm/region_devs.c index ec3543b83330..c30d5af02cc2 100644 --- a/drivers/nvdimm/region_devs.c +++ b/drivers/nvdimm/region_devs.c @@ -389,6 +389,30 @@ resource_size_t nd_region_available_dpa(struct nd_region *nd_region) return available; } +resource_size_t nd_region_allocatable_dpa(struct nd_region *nd_region) +{ + resource_size_t available = 0; + int i; + + if (is_memory(&nd_region->dev)) + available = PHYS_ADDR_MAX; + + WARN_ON(!is_nvdimm_bus_locked(&nd_region->dev)); + for (i = 0; i < nd_region->ndr_mappings; i++) { + struct nd_mapping *nd_mapping = &nd_region->mapping[i]; + + if (is_memory(&nd_region->dev)) + available = min(available, + nd_pmem_max_contiguous_dpa(nd_region, + nd_mapping)); + else if (is_nd_blk(&nd_region->dev)) + available += nd_blk_available_dpa(nd_region); + } + if (is_memory(&nd_region->dev)) + return available * nd_region->ndr_mappings; + return available; +} + static ssize_t available_size_show(struct device *dev, struct device_attribute *attr, char *buf) { -- cgit v1.2.3 From 1e687220ef2d95a582e13d8de79932bfe32fdfa8 Mon Sep 17 00:00:00 2001 From: Keith Busch Date: Tue, 24 Jul 2018 15:07:58 -0600 Subject: libnvdimm: Export max available extent The 'available_size' attribute showing the combined total of all unallocated space isn't always useful to know how large of a namespace a user may be able to allocate if the region is fragmented. This patch will export the largest extent of unallocated space that may be allocated to create a new namespace. Signed-off-by: Keith Busch Reviewed-by: Vishal Verma Signed-off-by: Dave Jiang --- drivers/nvdimm/region_devs.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) (limited to 'drivers') diff --git a/drivers/nvdimm/region_devs.c b/drivers/nvdimm/region_devs.c index c30d5af02cc2..fa37afcd43ff 100644 --- a/drivers/nvdimm/region_devs.c +++ b/drivers/nvdimm/region_devs.c @@ -434,6 +434,21 @@ static ssize_t available_size_show(struct device *dev, } static DEVICE_ATTR_RO(available_size); +static ssize_t max_available_extent_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct nd_region *nd_region = to_nd_region(dev); + unsigned long long available = 0; + + nvdimm_bus_lock(dev); + wait_nvdimm_bus_probe_idle(dev); + available = nd_region_allocatable_dpa(nd_region); + nvdimm_bus_unlock(dev); + + return sprintf(buf, "%llu\n", available); +} +static DEVICE_ATTR_RO(max_available_extent); + static ssize_t init_namespaces_show(struct device *dev, struct device_attribute *attr, char *buf) { @@ -585,6 +600,7 @@ static struct attribute *nd_region_attributes[] = { &dev_attr_read_only.attr, &dev_attr_set_cookie.attr, &dev_attr_available_size.attr, + &dev_attr_max_available_extent.attr, &dev_attr_namespace_seed.attr, &dev_attr_init_namespaces.attr, &dev_attr_badblocks.attr, -- cgit v1.2.3 From cc3d3458d46f702b1d118fca8647dc9157887d9b Mon Sep 17 00:00:00 2001 From: Dave Jiang Date: Fri, 27 Jul 2018 09:04:17 -0700 Subject: acpi/nfit: queue issuing of ars when an uc error notification comes in When the ACPI UC error notifier gets called and ARS_REQ bit is set with the passed in flag, we can receive -EBUSY if ARS_REQ bit is already set for the nfit_spa->ars_state. When that happens, the ARS request is dropped. That can potentially cause us to miss the unreported errors that the on going ARS request does not receive. Add an ARS_REQ_REDO state that will request short ARS upon ARS completion to grab any errors we missed. Signed-off-by: Dave Jiang Reviewed-by: Vishal Verma --- drivers/acpi/nfit/core.c | 12 +++++++++--- drivers/acpi/nfit/nfit.h | 1 + 2 files changed, 10 insertions(+), 3 deletions(-) (limited to 'drivers') diff --git a/drivers/acpi/nfit/core.c b/drivers/acpi/nfit/core.c index 2be8373153ed..31f6325ab3b5 100644 --- a/drivers/acpi/nfit/core.c +++ b/drivers/acpi/nfit/core.c @@ -2567,7 +2567,12 @@ static void ars_complete(struct acpi_nfit_desc *acpi_desc, test_bit(ARS_SHORT, &nfit_spa->ars_state) ? "short" : "long"); clear_bit(ARS_SHORT, &nfit_spa->ars_state); - set_bit(ARS_DONE, &nfit_spa->ars_state); + if (test_and_clear_bit(ARS_REQ_REDO, &nfit_spa->ars_state)) { + set_bit(ARS_SHORT, &nfit_spa->ars_state); + set_bit(ARS_REQ, &nfit_spa->ars_state); + dev_dbg(dev, "ARS: processing scrub request received while in progress\n"); + } else + set_bit(ARS_DONE, &nfit_spa->ars_state); } static int ars_status_process_records(struct acpi_nfit_desc *acpi_desc) @@ -3242,9 +3247,10 @@ int acpi_nfit_ars_rescan(struct acpi_nfit_desc *acpi_desc, unsigned long flags) if (test_bit(ARS_FAILED, &nfit_spa->ars_state)) continue; - if (test_and_set_bit(ARS_REQ, &nfit_spa->ars_state)) + if (test_and_set_bit(ARS_REQ, &nfit_spa->ars_state)) { busy++; - else { + set_bit(ARS_REQ_REDO, &nfit_spa->ars_state); + } else { if (test_bit(ARS_SHORT, &flags)) set_bit(ARS_SHORT, &nfit_spa->ars_state); scheduled++; diff --git a/drivers/acpi/nfit/nfit.h b/drivers/acpi/nfit/nfit.h index 7d15856a739f..29466961426e 100644 --- a/drivers/acpi/nfit/nfit.h +++ b/drivers/acpi/nfit/nfit.h @@ -119,6 +119,7 @@ enum nfit_dimm_notifiers { enum nfit_ars_state { ARS_REQ, + ARS_REQ_REDO, ARS_DONE, ARS_SHORT, ARS_FAILED, -- cgit v1.2.3 From 46a590cde0bd17cc983e40e74c195311cc8ba555 Mon Sep 17 00:00:00 2001 From: Huaisheng Ye Date: Mon, 30 Jul 2018 15:15:43 +0800 Subject: libnvdimm, pmem: kaddr and pfn can be NULL to ->direct_access() pmem_direct_access() needs to check the validity of pointers kaddr and pfn for NULL assignment. If anyone equals to NULL, it doesn't need to calculate the value. If pointer equals to NULL, that is to say callers may have no need for kaddr or pfn, so this patch is prepared for allowing them to pass in NULL instead of having to pass in a pointer or local variable that they then just throw away. Signed-off-by: Huaisheng Ye Reviewed-by: Ross Zwisler Signed-off-by: Dave Jiang --- drivers/nvdimm/pmem.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'drivers') diff --git a/drivers/nvdimm/pmem.c b/drivers/nvdimm/pmem.c index 68940356cad3..f3096564acb2 100644 --- a/drivers/nvdimm/pmem.c +++ b/drivers/nvdimm/pmem.c @@ -227,8 +227,11 @@ __weak long __pmem_direct_access(struct pmem_device *pmem, pgoff_t pgoff, if (unlikely(is_bad_pmem(&pmem->bb, PFN_PHYS(pgoff) / 512, PFN_PHYS(nr_pages)))) return -EIO; - *kaddr = pmem->virt_addr + offset; - *pfn = phys_to_pfn_t(pmem->phys_addr + offset, pmem->pfn_flags); + + if (kaddr) + *kaddr = pmem->virt_addr + offset; + if (pfn) + *pfn = phys_to_pfn_t(pmem->phys_addr + offset, pmem->pfn_flags); /* * If badblocks are present, limit known good range to the -- cgit v1.2.3 From 036a82ca90328960d66b7a5c0e1908475d534d62 Mon Sep 17 00:00:00 2001 From: Huaisheng Ye Date: Mon, 30 Jul 2018 15:15:44 +0800 Subject: s390, dcssblk: kaddr and pfn can be NULL to ->direct_access() dcssblk_direct_access() needs to check the validity of pointers kaddr and pfn for NULL assignment. If anyone equals to NULL, it doesn't need to calculate the value. If either of them is equal to NULL, that is to say callers may have no need for kaddr or pfn, so this patch is prepared for allowing them to pass in NULL instead of having to pass in a pointer or local variable that they then just throw away. Signed-off-by: Huaisheng Ye Reviewed-by: Ross Zwisler Acked-by: Martin Schwidefsky Signed-off-by: Dave Jiang --- drivers/s390/block/dcssblk.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'drivers') diff --git a/drivers/s390/block/dcssblk.c b/drivers/s390/block/dcssblk.c index ed607288e696..23e526cda5c1 100644 --- a/drivers/s390/block/dcssblk.c +++ b/drivers/s390/block/dcssblk.c @@ -922,9 +922,11 @@ __dcssblk_direct_access(struct dcssblk_dev_info *dev_info, pgoff_t pgoff, unsigned long dev_sz; dev_sz = dev_info->end - dev_info->start + 1; - *kaddr = (void *) dev_info->start + offset; - *pfn = __pfn_to_pfn_t(PFN_DOWN(dev_info->start + offset), - PFN_DEV|PFN_SPECIAL); + if (kaddr) + *kaddr = (void *) dev_info->start + offset; + if (pfn) + *pfn = __pfn_to_pfn_t(PFN_DOWN(dev_info->start + offset), + PFN_DEV|PFN_SPECIAL); return (dev_sz - offset) / PAGE_SIZE; } -- cgit v1.2.3 From e0b401e3fef0f81fd87434793db1adb0d7568058 Mon Sep 17 00:00:00 2001 From: Huaisheng Ye Date: Mon, 30 Jul 2018 15:15:46 +0800 Subject: dax/super: Do not request a pointer kaddr when not required Function __bdev_dax_supported doesn't need to get local pointer kaddr from direct_access. Using NULL instead of having to pass in a useless local pointer that caller then just throw away. Signed-off-by: Huaisheng Ye Reviewed-by: Ross Zwisler Signed-off-by: Dave Jiang --- drivers/dax/super.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'drivers') diff --git a/drivers/dax/super.c b/drivers/dax/super.c index 903d9c473749..c2f42b1f3363 100644 --- a/drivers/dax/super.c +++ b/drivers/dax/super.c @@ -88,7 +88,6 @@ bool __bdev_dax_supported(struct block_device *bdev, int blocksize) bool dax_enabled = false; pgoff_t pgoff; int err, id; - void *kaddr; pfn_t pfn; long len; char buf[BDEVNAME_SIZE]; @@ -114,7 +113,7 @@ bool __bdev_dax_supported(struct block_device *bdev, int blocksize) } id = dax_read_lock(); - len = dax_direct_access(dax_dev, pgoff, 1, &kaddr, &pfn); + len = dax_direct_access(dax_dev, pgoff, 1, NULL, &pfn); dax_read_unlock(id); put_dax(dax_dev); -- cgit v1.2.3 From f742267ae91faf241d2892878208dffc586ffe06 Mon Sep 17 00:00:00 2001 From: Huaisheng Ye Date: Mon, 30 Jul 2018 15:15:47 +0800 Subject: md/dm-writecache: Don't request pointer dummy_addr when not required Function persistent_memory_claim doesn't need to get local pointer dummy_addr from direct_access. Using NULL instead of having to pass in a useless local pointer that caller then just throw away. Suggested-by: Ross Zwisler Signed-off-by: Huaisheng Ye Reviewed-by: Ross Zwisler Acked-by: Mike Snitzer Signed-off-by: Dave Jiang --- drivers/md/dm-writecache.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'drivers') diff --git a/drivers/md/dm-writecache.c b/drivers/md/dm-writecache.c index 5961c7794ef3..1839236a0f81 100644 --- a/drivers/md/dm-writecache.c +++ b/drivers/md/dm-writecache.c @@ -267,9 +267,8 @@ static int persistent_memory_claim(struct dm_writecache *wc) i = 0; do { long daa; - void *dummy_addr; daa = dax_direct_access(wc->ssd_dev->dax_dev, i, p - i, - &dummy_addr, &pfn); + NULL, &pfn); if (daa <= 0) { r = daa ? daa : -EINVAL; goto err3; -- cgit v1.2.3 From b7751410c180a05fdc21268f8661b1480169b0df Mon Sep 17 00:00:00 2001 From: Stefan Hajnoczi Date: Tue, 31 Jul 2018 15:32:46 +0100 Subject: device-dax: avoid hang on error before devm_memremap_pages() dax_pmem_percpu_exit() waits for dax_pmem_percpu_release() to invoke the dax_pmem->cmp completion. Unfortunately this approach to cleaning up the percpu_ref only works after devm_memremap_pages() was successful. If devm_add_action_or_reset() or devm_memremap_pages() fails, dax_pmem_percpu_release() is not invoked. Therefore dax_pmem_percpu_exit() hangs waiting for the completion: rc = devm_add_action_or_reset(dev, dax_pmem_percpu_exit, &dax_pmem->ref); if (rc) return rc; dax_pmem->pgmap.ref = &dax_pmem->ref; addr = devm_memremap_pages(dev, &dax_pmem->pgmap); Avoid the hang by calling percpu_ref_exit() in the error paths instead of going through dax_pmem_percpu_exit(). Signed-off-by: Stefan Hajnoczi Signed-off-by: Dave Jiang --- drivers/dax/pmem.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) (limited to 'drivers') diff --git a/drivers/dax/pmem.c b/drivers/dax/pmem.c index fd49b24fd6af..99e2aace8078 100644 --- a/drivers/dax/pmem.c +++ b/drivers/dax/pmem.c @@ -105,15 +105,19 @@ static int dax_pmem_probe(struct device *dev) if (rc) return rc; - rc = devm_add_action_or_reset(dev, dax_pmem_percpu_exit, - &dax_pmem->ref); - if (rc) + rc = devm_add_action(dev, dax_pmem_percpu_exit, &dax_pmem->ref); + if (rc) { + percpu_ref_exit(&dax_pmem->ref); return rc; + } dax_pmem->pgmap.ref = &dax_pmem->ref; addr = devm_memremap_pages(dev, &dax_pmem->pgmap); - if (IS_ERR(addr)) + if (IS_ERR(addr)) { + devm_remove_action(dev, dax_pmem_percpu_exit, &dax_pmem->ref); + percpu_ref_exit(&dax_pmem->ref); return PTR_ERR(addr); + } rc = devm_add_action_or_reset(dev, dax_pmem_percpu_kill, &dax_pmem->ref); -- cgit v1.2.3 From 286e87718103acdf85f4ed323a37e4839a8a7c05 Mon Sep 17 00:00:00 2001 From: Vishal Verma Date: Fri, 10 Aug 2018 13:23:15 -0600 Subject: libnvdimm: fix ars_status output length calculation Commit efda1b5d87cb ("acpi, nfit, libnvdimm: fix / harden ars_status output length handling") Introduced additional hardening for ambiguity in the ACPI spec for ars_status output sizing. However, it had a couple of cases mixed up. Where it should have been checking for (and returning) "out_field[1] - 4" it was using "out_field[1] - 8" and vice versa. This caused a four byte discrepancy in the buffer size passed on to the command handler, and in some cases, this caused memory corruption like: ./daxdev-errors.sh: line 76: 24104 Aborted (core dumped) ./daxdev-errors $busdev $region malloc(): memory corruption Program received signal SIGABRT, Aborted. [...] #5 0x00007ffff7865a2e in calloc () from /lib64/libc.so.6 #6 0x00007ffff7bc2970 in ndctl_bus_cmd_new_ars_status (ars_cap=ars_cap@entry=0x6153b0) at ars.c:136 #7 0x0000000000401644 in check_ars_status (check=0x7fffffffdeb0, bus=0x604c20) at daxdev-errors.c:144 #8 test_daxdev_clear_error (region_name=, bus_name=) at daxdev-errors.c:332 Cc: Cc: Dave Jiang Cc: Keith Busch Cc: Lukasz Dorau Cc: Dan Williams Fixes: efda1b5d87cb ("acpi, nfit, libnvdimm: fix / harden ars_status output length handling") Signed-off-by: Vishal Verma Reviewed-by: Keith Busch Signed-of-by: Dave Jiang --- drivers/nvdimm/bus.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers') diff --git a/drivers/nvdimm/bus.c b/drivers/nvdimm/bus.c index 27902a8799b1..8aae6dcc839f 100644 --- a/drivers/nvdimm/bus.c +++ b/drivers/nvdimm/bus.c @@ -812,9 +812,9 @@ u32 nd_cmd_out_size(struct nvdimm *nvdimm, int cmd, * overshoots the remainder by 4 bytes, assume it was * including 'status'. */ - if (out_field[1] - 8 == remainder) + if (out_field[1] - 4 == remainder) return remainder; - return out_field[1] - 4; + return out_field[1] - 8; } else if (cmd == ND_CMD_CALL) { struct nd_cmd_pkg *pkg = (struct nd_cmd_pkg *) in_field; -- cgit v1.2.3