diff options
Diffstat (limited to 'drivers/nvdimm')
| -rw-r--r-- | drivers/nvdimm/btt_devs.c | 23 | ||||
| -rw-r--r-- | drivers/nvdimm/bus.c | 38 | ||||
| -rw-r--r-- | drivers/nvdimm/core.c | 19 | ||||
| -rw-r--r-- | drivers/nvdimm/dax_devs.c | 4 | ||||
| -rw-r--r-- | drivers/nvdimm/dimm_devs.c | 12 | ||||
| -rw-r--r-- | drivers/nvdimm/namespace_devs.c | 46 | ||||
| -rw-r--r-- | drivers/nvdimm/nd-core.h | 68 | ||||
| -rw-r--r-- | drivers/nvdimm/pfn_devs.c | 31 | ||||
| -rw-r--r-- | drivers/nvdimm/pmem.c | 205 | ||||
| -rw-r--r-- | drivers/nvdimm/pmem.h | 5 | ||||
| -rw-r--r-- | drivers/nvdimm/region.c | 2 | ||||
| -rw-r--r-- | drivers/nvdimm/region_devs.c | 20 | ||||
| -rw-r--r-- | drivers/nvdimm/security.c | 5 |
13 files changed, 252 insertions, 226 deletions
diff --git a/drivers/nvdimm/btt_devs.c b/drivers/nvdimm/btt_devs.c index e5a58520d398..fabbb31f2c35 100644 --- a/drivers/nvdimm/btt_devs.c +++ b/drivers/nvdimm/btt_devs.c @@ -50,14 +50,14 @@ static ssize_t sector_size_store(struct device *dev, struct nd_btt *nd_btt = to_nd_btt(dev); ssize_t rc; - nd_device_lock(dev); + device_lock(dev); nvdimm_bus_lock(dev); rc = nd_size_select_store(dev, buf, &nd_btt->lbasize, btt_lbasize_supported); dev_dbg(dev, "result: %zd wrote: %s%s", rc, buf, buf[len - 1] == '\n' ? "" : "\n"); nvdimm_bus_unlock(dev); - nd_device_unlock(dev); + device_unlock(dev); return rc ? rc : len; } @@ -79,11 +79,11 @@ static ssize_t uuid_store(struct device *dev, struct nd_btt *nd_btt = to_nd_btt(dev); ssize_t rc; - nd_device_lock(dev); + device_lock(dev); rc = nd_uuid_store(dev, &nd_btt->uuid, buf, len); dev_dbg(dev, "result: %zd wrote: %s%s", rc, buf, buf[len - 1] == '\n' ? "" : "\n"); - nd_device_unlock(dev); + device_unlock(dev); return rc ? rc : len; } @@ -108,13 +108,13 @@ static ssize_t namespace_store(struct device *dev, struct nd_btt *nd_btt = to_nd_btt(dev); ssize_t rc; - nd_device_lock(dev); + device_lock(dev); nvdimm_bus_lock(dev); rc = nd_namespace_store(dev, &nd_btt->ndns, buf, len); dev_dbg(dev, "result: %zd wrote: %s%s", rc, buf, buf[len - 1] == '\n' ? "" : "\n"); nvdimm_bus_unlock(dev); - nd_device_unlock(dev); + device_unlock(dev); return rc; } @@ -126,14 +126,14 @@ static ssize_t size_show(struct device *dev, struct nd_btt *nd_btt = to_nd_btt(dev); ssize_t rc; - nd_device_lock(dev); + device_lock(dev); if (dev->driver) rc = sprintf(buf, "%llu\n", nd_btt->size); else { /* no size to convey if the btt instance is disabled */ rc = -ENXIO; } - nd_device_unlock(dev); + device_unlock(dev); return rc; } @@ -178,6 +178,8 @@ bool is_nd_btt(struct device *dev) } EXPORT_SYMBOL(is_nd_btt); +static struct lock_class_key nvdimm_btt_key; + static struct device *__nd_btt_create(struct nd_region *nd_region, unsigned long lbasize, uuid_t *uuid, struct nd_namespace_common *ndns) @@ -205,6 +207,7 @@ static struct device *__nd_btt_create(struct nd_region *nd_region, dev->parent = &nd_region->dev; dev->type = &nd_btt_device_type; device_initialize(&nd_btt->dev); + lockdep_set_class(&nd_btt->dev.mutex, &nvdimm_btt_key); if (ndns && !__nd_attach_ndns(&nd_btt->dev, ndns, &nd_btt->ndns)) { dev_dbg(&ndns->dev, "failed, already claimed by %s\n", dev_name(ndns->claim)); @@ -225,7 +228,7 @@ struct device *nd_btt_create(struct nd_region *nd_region) { struct device *dev = __nd_btt_create(nd_region, 0, NULL, NULL); - __nd_device_register(dev); + nd_device_register(dev); return dev; } @@ -324,7 +327,7 @@ static int __nd_btt_probe(struct nd_btt *nd_btt, if (!nd_btt->uuid) return -ENOMEM; - __nd_device_register(&nd_btt->dev); + nd_device_register(&nd_btt->dev); return 0; } diff --git a/drivers/nvdimm/bus.c b/drivers/nvdimm/bus.c index 7b0d1443217a..a4fc17db707c 100644 --- a/drivers/nvdimm/bus.c +++ b/drivers/nvdimm/bus.c @@ -88,10 +88,7 @@ static int nvdimm_bus_probe(struct device *dev) dev->driver->name, dev_name(dev)); nvdimm_bus_probe_start(nvdimm_bus); - debug_nvdimm_lock(dev); rc = nd_drv->probe(dev); - debug_nvdimm_unlock(dev); - if ((rc == 0 || rc == -EOPNOTSUPP) && dev->parent && is_nd_region(dev->parent)) nd_region_advance_seeds(to_nd_region(dev->parent), dev); @@ -111,11 +108,8 @@ static void nvdimm_bus_remove(struct device *dev) struct module *provider = to_bus_provider(dev); struct nvdimm_bus *nvdimm_bus = walk_to_nvdimm_bus(dev); - if (nd_drv->remove) { - debug_nvdimm_lock(dev); + if (nd_drv->remove) nd_drv->remove(dev); - debug_nvdimm_unlock(dev); - } dev_dbg(&nvdimm_bus->dev, "%s.remove(%s)\n", dev->driver->name, dev_name(dev)); @@ -139,7 +133,7 @@ static void nvdimm_bus_shutdown(struct device *dev) void nd_device_notify(struct device *dev, enum nvdimm_event event) { - nd_device_lock(dev); + device_lock(dev); if (dev->driver) { struct nd_device_driver *nd_drv; @@ -147,7 +141,7 @@ void nd_device_notify(struct device *dev, enum nvdimm_event event) if (nd_drv->notify) nd_drv->notify(dev, event); } - nd_device_unlock(dev); + device_unlock(dev); } EXPORT_SYMBOL(nd_device_notify); @@ -334,6 +328,8 @@ struct nvdimm_bus *nvdimm_to_bus(struct nvdimm *nvdimm) } EXPORT_SYMBOL_GPL(nvdimm_to_bus); +static struct lock_class_key nvdimm_bus_key; + struct nvdimm_bus *nvdimm_bus_register(struct device *parent, struct nvdimm_bus_descriptor *nd_desc) { @@ -360,6 +356,7 @@ struct nvdimm_bus *nvdimm_bus_register(struct device *parent, nvdimm_bus->dev.bus = &nvdimm_bus_type; nvdimm_bus->dev.of_node = nd_desc->of_node; device_initialize(&nvdimm_bus->dev); + lockdep_set_class(&nvdimm_bus->dev.mutex, &nvdimm_bus_key); device_set_pm_not_required(&nvdimm_bus->dev); rc = dev_set_name(&nvdimm_bus->dev, "ndbus%d", nvdimm_bus->id); if (rc) @@ -511,7 +508,7 @@ static void nd_async_device_unregister(void *d, async_cookie_t cookie) put_device(dev); } -void __nd_device_register(struct device *dev) +void nd_device_register(struct device *dev) { if (!dev) return; @@ -537,12 +534,6 @@ void __nd_device_register(struct device *dev) async_schedule_dev_domain(nd_async_device_register, dev, &nd_async_domain); } - -void nd_device_register(struct device *dev) -{ - device_initialize(dev); - __nd_device_register(dev); -} EXPORT_SYMBOL(nd_device_register); void nd_device_unregister(struct device *dev, enum nd_async_mode mode) @@ -572,9 +563,9 @@ void nd_device_unregister(struct device *dev, enum nd_async_mode mode) * or otherwise let the async path handle it if the * unregistration was already queued. */ - nd_device_lock(dev); + device_lock(dev); killed = kill_device(dev); - nd_device_unlock(dev); + device_unlock(dev); if (!killed) return; @@ -724,6 +715,8 @@ static void ndctl_release(struct device *dev) kfree(dev); } +static struct lock_class_key nvdimm_ndctl_key; + int nvdimm_bus_create_ndctl(struct nvdimm_bus *nvdimm_bus) { dev_t devt = MKDEV(nvdimm_bus_major, nvdimm_bus->id); @@ -734,6 +727,7 @@ int nvdimm_bus_create_ndctl(struct nvdimm_bus *nvdimm_bus) if (!dev) return -ENOMEM; device_initialize(dev); + lockdep_set_class(&dev->mutex, &nvdimm_ndctl_key); device_set_pm_not_required(dev); dev->class = nd_class; dev->parent = &nvdimm_bus->dev; @@ -930,10 +924,10 @@ void wait_nvdimm_bus_probe_idle(struct device *dev) if (nvdimm_bus->probe_active == 0) break; nvdimm_bus_unlock(dev); - nd_device_unlock(dev); + device_unlock(dev); wait_event(nvdimm_bus->wait, nvdimm_bus->probe_active == 0); - nd_device_lock(dev); + device_lock(dev); nvdimm_bus_lock(dev); } while (true); } @@ -1167,7 +1161,7 @@ static int __nd_ioctl(struct nvdimm_bus *nvdimm_bus, struct nvdimm *nvdimm, goto out; } - nd_device_lock(dev); + device_lock(dev); nvdimm_bus_lock(dev); rc = nd_cmd_clear_to_send(nvdimm_bus, nvdimm, func, buf); if (rc) @@ -1189,7 +1183,7 @@ static int __nd_ioctl(struct nvdimm_bus *nvdimm_bus, struct nvdimm *nvdimm, out_unlock: nvdimm_bus_unlock(dev); - nd_device_unlock(dev); + device_unlock(dev); out: kfree(in_env); kfree(out_env); diff --git a/drivers/nvdimm/core.c b/drivers/nvdimm/core.c index 69a03358817f..d91799b71d23 100644 --- a/drivers/nvdimm/core.c +++ b/drivers/nvdimm/core.c @@ -215,7 +215,7 @@ EXPORT_SYMBOL_GPL(to_nvdimm_bus_dev); * * Enforce that uuids can only be changed while the device is disabled * (driver detached) - * LOCKING: expects nd_device_lock() is held on entry + * LOCKING: expects device_lock() is held on entry */ int nd_uuid_store(struct device *dev, uuid_t **uuid_out, const char *buf, size_t len) @@ -316,15 +316,15 @@ static DEVICE_ATTR_RO(provider); static int flush_namespaces(struct device *dev, void *data) { - nd_device_lock(dev); - nd_device_unlock(dev); + device_lock(dev); + device_unlock(dev); return 0; } static int flush_regions_dimms(struct device *dev, void *data) { - nd_device_lock(dev); - nd_device_unlock(dev); + device_lock(dev); + device_unlock(dev); device_for_each_child(dev, NULL, flush_namespaces); return 0; } @@ -368,9 +368,7 @@ static ssize_t capability_show(struct device *dev, if (!nd_desc->fw_ops) return -EOPNOTSUPP; - nvdimm_bus_lock(dev); cap = nd_desc->fw_ops->capability(nd_desc); - nvdimm_bus_unlock(dev); switch (cap) { case NVDIMM_FWA_CAP_QUIESCE: @@ -395,10 +393,8 @@ static ssize_t activate_show(struct device *dev, if (!nd_desc->fw_ops) return -EOPNOTSUPP; - nvdimm_bus_lock(dev); cap = nd_desc->fw_ops->capability(nd_desc); state = nd_desc->fw_ops->activate_state(nd_desc); - nvdimm_bus_unlock(dev); if (cap < NVDIMM_FWA_CAP_QUIESCE) return -EOPNOTSUPP; @@ -443,7 +439,6 @@ static ssize_t activate_store(struct device *dev, else return -EINVAL; - nvdimm_bus_lock(dev); state = nd_desc->fw_ops->activate_state(nd_desc); switch (state) { @@ -461,7 +456,6 @@ static ssize_t activate_store(struct device *dev, default: rc = -ENXIO; } - nvdimm_bus_unlock(dev); if (rc == 0) rc = len; @@ -484,10 +478,7 @@ static umode_t nvdimm_bus_firmware_visible(struct kobject *kobj, struct attribut if (!nd_desc->fw_ops) return 0; - nvdimm_bus_lock(dev); cap = nd_desc->fw_ops->capability(nd_desc); - nvdimm_bus_unlock(dev); - if (cap < NVDIMM_FWA_CAP_QUIESCE) return 0; diff --git a/drivers/nvdimm/dax_devs.c b/drivers/nvdimm/dax_devs.c index 99965077bac4..7f4a9d28b670 100644 --- a/drivers/nvdimm/dax_devs.c +++ b/drivers/nvdimm/dax_devs.c @@ -80,7 +80,7 @@ struct device *nd_dax_create(struct nd_region *nd_region) nd_dax = nd_dax_alloc(nd_region); if (nd_dax) dev = nd_pfn_devinit(&nd_dax->nd_pfn, NULL); - __nd_device_register(dev); + nd_device_register(dev); return dev; } @@ -119,7 +119,7 @@ int nd_dax_probe(struct device *dev, struct nd_namespace_common *ndns) nd_detach_ndns(dax_dev, &nd_pfn->ndns); put_device(dax_dev); } else - __nd_device_register(dax_dev); + nd_device_register(dax_dev); return rc; } diff --git a/drivers/nvdimm/dimm_devs.c b/drivers/nvdimm/dimm_devs.c index ee507eed42b5..c7c980577491 100644 --- a/drivers/nvdimm/dimm_devs.c +++ b/drivers/nvdimm/dimm_devs.c @@ -341,9 +341,9 @@ static ssize_t available_slots_show(struct device *dev, { ssize_t rc; - nd_device_lock(dev); + device_lock(dev); rc = __available_slots_show(dev_get_drvdata(dev), buf); - nd_device_unlock(dev); + device_unlock(dev); return rc; } @@ -386,12 +386,12 @@ static ssize_t security_store(struct device *dev, * done while probing is idle and the DIMM is not in active use * in any region. */ - nd_device_lock(dev); + device_lock(dev); nvdimm_bus_lock(dev); wait_nvdimm_bus_probe_idle(dev); rc = nvdimm_security_store(dev, buf, len); nvdimm_bus_unlock(dev); - nd_device_unlock(dev); + device_unlock(dev); return rc; } @@ -570,6 +570,8 @@ bool is_nvdimm(struct device *dev) return dev->type == &nvdimm_device_type; } +static struct lock_class_key nvdimm_key; + struct nvdimm *__nvdimm_create(struct nvdimm_bus *nvdimm_bus, void *provider_data, const struct attribute_group **groups, unsigned long flags, unsigned long cmd_mask, int num_flush, @@ -613,6 +615,8 @@ struct nvdimm *__nvdimm_create(struct nvdimm_bus *nvdimm_bus, /* get security state and extended (master) state */ nvdimm->sec.flags = nvdimm_security_flags(nvdimm, NVDIMM_USER); nvdimm->sec.ext_flags = nvdimm_security_flags(nvdimm, NVDIMM_MASTER); + device_initialize(dev); + lockdep_set_class(&dev->mutex, &nvdimm_key); nd_device_register(dev); return nvdimm; diff --git a/drivers/nvdimm/namespace_devs.c b/drivers/nvdimm/namespace_devs.c index 62b83b2e26e3..bf4f5c09d9b1 100644 --- a/drivers/nvdimm/namespace_devs.c +++ b/drivers/nvdimm/namespace_devs.c @@ -264,7 +264,7 @@ static ssize_t alt_name_store(struct device *dev, struct nd_region *nd_region = to_nd_region(dev->parent); ssize_t rc; - nd_device_lock(dev); + device_lock(dev); nvdimm_bus_lock(dev); wait_nvdimm_bus_probe_idle(dev); rc = __alt_name_store(dev, buf, len); @@ -272,7 +272,7 @@ static ssize_t alt_name_store(struct device *dev, rc = nd_namespace_label_update(nd_region, dev); dev_dbg(dev, "%s(%zd)\n", rc < 0 ? "fail " : "", rc); nvdimm_bus_unlock(dev); - nd_device_unlock(dev); + device_unlock(dev); return rc < 0 ? rc : len; } @@ -846,7 +846,7 @@ static ssize_t size_store(struct device *dev, if (rc) return rc; - nd_device_lock(dev); + device_lock(dev); nvdimm_bus_lock(dev); wait_nvdimm_bus_probe_idle(dev); rc = __size_store(dev, val); @@ -868,7 +868,7 @@ static ssize_t size_store(struct device *dev, dev_dbg(dev, "%llx %s (%d)\n", val, rc < 0 ? "fail" : "success", rc); nvdimm_bus_unlock(dev); - nd_device_unlock(dev); + device_unlock(dev); return rc < 0 ? rc : len; } @@ -1043,7 +1043,7 @@ static ssize_t uuid_store(struct device *dev, } else return -ENXIO; - nd_device_lock(dev); + device_lock(dev); nvdimm_bus_lock(dev); wait_nvdimm_bus_probe_idle(dev); if (to_ndns(dev)->claim) @@ -1059,7 +1059,7 @@ static ssize_t uuid_store(struct device *dev, dev_dbg(dev, "result: %zd wrote: %s%s", rc, buf, buf[len - 1] == '\n' ? "" : "\n"); nvdimm_bus_unlock(dev); - nd_device_unlock(dev); + device_unlock(dev); return rc < 0 ? rc : len; } @@ -1118,7 +1118,7 @@ static ssize_t sector_size_store(struct device *dev, } else return -ENXIO; - nd_device_lock(dev); + device_lock(dev); nvdimm_bus_lock(dev); if (to_ndns(dev)->claim) rc = -EBUSY; @@ -1129,7 +1129,7 @@ static ssize_t sector_size_store(struct device *dev, dev_dbg(dev, "result: %zd %s: %s%s", rc, rc < 0 ? "tried" : "wrote", buf, buf[len - 1] == '\n' ? "" : "\n"); nvdimm_bus_unlock(dev); - nd_device_unlock(dev); + device_unlock(dev); return rc ? rc : len; } @@ -1239,9 +1239,9 @@ static ssize_t holder_show(struct device *dev, struct nd_namespace_common *ndns = to_ndns(dev); ssize_t rc; - nd_device_lock(dev); + device_lock(dev); rc = sprintf(buf, "%s\n", ndns->claim ? dev_name(ndns->claim) : ""); - nd_device_unlock(dev); + device_unlock(dev); return rc; } @@ -1278,7 +1278,7 @@ static ssize_t holder_class_store(struct device *dev, struct nd_region *nd_region = to_nd_region(dev->parent); int rc; - nd_device_lock(dev); + device_lock(dev); nvdimm_bus_lock(dev); wait_nvdimm_bus_probe_idle(dev); rc = __holder_class_store(dev, buf); @@ -1286,7 +1286,7 @@ static ssize_t holder_class_store(struct device *dev, rc = nd_namespace_label_update(nd_region, dev); dev_dbg(dev, "%s(%d)\n", rc < 0 ? "fail " : "", rc); nvdimm_bus_unlock(dev); - nd_device_unlock(dev); + device_unlock(dev); return rc < 0 ? rc : len; } @@ -1297,7 +1297,7 @@ static ssize_t holder_class_show(struct device *dev, struct nd_namespace_common *ndns = to_ndns(dev); ssize_t rc; - nd_device_lock(dev); + device_lock(dev); if (ndns->claim_class == NVDIMM_CCLASS_NONE) rc = sprintf(buf, "\n"); else if ((ndns->claim_class == NVDIMM_CCLASS_BTT) || @@ -1309,7 +1309,7 @@ static ssize_t holder_class_show(struct device *dev, rc = sprintf(buf, "dax\n"); else rc = sprintf(buf, "<unknown>\n"); - nd_device_unlock(dev); + device_unlock(dev); return rc; } @@ -1323,7 +1323,7 @@ static ssize_t mode_show(struct device *dev, char *mode; ssize_t rc; - nd_device_lock(dev); + device_lock(dev); claim = ndns->claim; if (claim && is_nd_btt(claim)) mode = "safe"; @@ -1336,7 +1336,7 @@ static ssize_t mode_show(struct device *dev, else mode = "raw"; rc = sprintf(buf, "%s\n", mode); - nd_device_unlock(dev); + device_unlock(dev); return rc; } @@ -1456,8 +1456,8 @@ struct nd_namespace_common *nvdimm_namespace_common_probe(struct device *dev) * Flush any in-progess probes / removals in the driver * for the raw personality of this namespace. */ - nd_device_lock(&ndns->dev); - nd_device_unlock(&ndns->dev); + device_lock(&ndns->dev); + device_unlock(&ndns->dev); if (ndns->dev.driver) { dev_dbg(&ndns->dev, "is active, can't bind %s\n", dev_name(dev)); @@ -1830,6 +1830,8 @@ static struct device *nd_namespace_pmem_create(struct nd_region *nd_region) return dev; } +static struct lock_class_key nvdimm_namespace_key; + void nd_region_create_ns_seed(struct nd_region *nd_region) { WARN_ON(!is_nvdimm_bus_locked(&nd_region->dev)); @@ -1845,8 +1847,12 @@ void nd_region_create_ns_seed(struct nd_region *nd_region) */ if (!nd_region->ns_seed) dev_err(&nd_region->dev, "failed to create namespace\n"); - else + else { + device_initialize(nd_region->ns_seed); + lockdep_set_class(&nd_region->ns_seed->mutex, + &nvdimm_namespace_key); nd_device_register(nd_region->ns_seed); + } } void nd_region_create_dax_seed(struct nd_region *nd_region) @@ -2200,6 +2206,8 @@ int nd_region_register_namespaces(struct nd_region *nd_region, int *err) if (id < 0) break; dev_set_name(dev, "namespace%d.%d", nd_region->id, id); + device_initialize(dev); + lockdep_set_class(&dev->mutex, &nvdimm_namespace_key); nd_device_register(dev); } if (i) diff --git a/drivers/nvdimm/nd-core.h b/drivers/nvdimm/nd-core.h index 448f9dcb4bb7..cc86ee09d7c0 100644 --- a/drivers/nvdimm/nd-core.h +++ b/drivers/nvdimm/nd-core.h @@ -106,7 +106,7 @@ void nd_region_create_dax_seed(struct nd_region *nd_region); int nvdimm_bus_create_ndctl(struct nvdimm_bus *nvdimm_bus); void nvdimm_bus_destroy_ndctl(struct nvdimm_bus *nvdimm_bus); void nd_synchronize(void); -void __nd_device_register(struct device *dev); +void nd_device_register(struct device *dev); struct nd_label_id; char *nd_label_gen_id(struct nd_label_id *label_id, const uuid_t *uuid, u32 flags); @@ -161,70 +161,4 @@ static inline void devm_nsio_disable(struct device *dev, { } #endif - -#ifdef CONFIG_PROVE_NVDIMM_LOCKING -extern struct class *nd_class; - -enum { - LOCK_BUS, - LOCK_NDCTL, - LOCK_REGION, - LOCK_DIMM = LOCK_REGION, - LOCK_NAMESPACE, - LOCK_CLAIM, -}; - -static inline void debug_nvdimm_lock(struct device *dev) -{ - if (is_nd_region(dev)) - mutex_lock_nested(&dev->lockdep_mutex, LOCK_REGION); - else if (is_nvdimm(dev)) - mutex_lock_nested(&dev->lockdep_mutex, LOCK_DIMM); - else if (is_nd_btt(dev) || is_nd_pfn(dev) || is_nd_dax(dev)) - mutex_lock_nested(&dev->lockdep_mutex, LOCK_CLAIM); - else if (dev->parent && (is_nd_region(dev->parent))) - mutex_lock_nested(&dev->lockdep_mutex, LOCK_NAMESPACE); - else if (is_nvdimm_bus(dev)) - mutex_lock_nested(&dev->lockdep_mutex, LOCK_BUS); - else if (dev->class && dev->class == nd_class) - mutex_lock_nested(&dev->lockdep_mutex, LOCK_NDCTL); - else - dev_WARN(dev, "unknown lock level\n"); -} - -static inline void debug_nvdimm_unlock(struct device *dev) -{ - mutex_unlock(&dev->lockdep_mutex); -} - -static inline void nd_device_lock(struct device *dev) -{ - device_lock(dev); - debug_nvdimm_lock(dev); -} - -static inline void nd_device_unlock(struct device *dev) -{ - debug_nvdimm_unlock(dev); - device_unlock(dev); -} -#else -static inline void nd_device_lock(struct device *dev) -{ - device_lock(dev); -} - -static inline void nd_device_unlock(struct device *dev) -{ - device_unlock(dev); -} - -static inline void debug_nvdimm_lock(struct device *dev) -{ -} - -static inline void debug_nvdimm_unlock(struct device *dev) -{ -} -#endif #endif /* __ND_CORE_H__ */ diff --git a/drivers/nvdimm/pfn_devs.c b/drivers/nvdimm/pfn_devs.c index c31e184bfa45..0e92ab4b3283 100644 --- a/drivers/nvdimm/pfn_devs.c +++ b/drivers/nvdimm/pfn_devs.c @@ -55,7 +55,7 @@ static ssize_t mode_store(struct device *dev, struct nd_pfn *nd_pfn = to_nd_pfn_safe(dev); ssize_t rc = 0; - nd_device_lock(dev); + device_lock(dev); nvdimm_bus_lock(dev); if (dev->driver) rc = -EBUSY; @@ -77,7 +77,7 @@ static ssize_t mode_store(struct device *dev, dev_dbg(dev, "result: %zd wrote: %s%s", rc, buf, buf[len - 1] == '\n' ? "" : "\n"); nvdimm_bus_unlock(dev); - nd_device_unlock(dev); + device_unlock(dev); return rc ? rc : len; } @@ -123,14 +123,14 @@ static ssize_t align_store(struct device *dev, unsigned long aligns[MAX_NVDIMM_ALIGN] = { [0] = 0, }; ssize_t rc; - nd_device_lock(dev); + device_lock(dev); nvdimm_bus_lock(dev); rc = nd_size_select_store(dev, buf, &nd_pfn->align, nd_pfn_supported_alignments(aligns)); dev_dbg(dev, "result: %zd wrote: %s%s", rc, buf, buf[len - 1] == '\n' ? "" : "\n"); nvdimm_bus_unlock(dev); - nd_device_unlock(dev); + device_unlock(dev); return rc ? rc : len; } @@ -152,11 +152,11 @@ static ssize_t uuid_store(struct device *dev, struct nd_pfn *nd_pfn = to_nd_pfn_safe(dev); ssize_t rc; - nd_device_lock(dev); + device_lock(dev); rc = nd_uuid_store(dev, &nd_pfn->uuid, buf, len); dev_dbg(dev, "result: %zd wrote: %s%s", rc, buf, buf[len - 1] == '\n' ? "" : "\n"); - nd_device_unlock(dev); + device_unlock(dev); return rc ? rc : len; } @@ -181,13 +181,13 @@ static ssize_t namespace_store(struct device *dev, struct nd_pfn *nd_pfn = to_nd_pfn_safe(dev); ssize_t rc; - nd_device_lock(dev); + device_lock(dev); nvdimm_bus_lock(dev); rc = nd_namespace_store(dev, &nd_pfn->ndns, buf, len); dev_dbg(dev, "result: %zd wrote: %s%s", rc, buf, buf[len - 1] == '\n' ? "" : "\n"); nvdimm_bus_unlock(dev); - nd_device_unlock(dev); + device_unlock(dev); return rc; } @@ -199,7 +199,7 @@ static ssize_t resource_show(struct device *dev, struct nd_pfn *nd_pfn = to_nd_pfn_safe(dev); ssize_t rc; - nd_device_lock(dev); + device_lock(dev); if (dev->driver) { struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb; u64 offset = __le64_to_cpu(pfn_sb->dataoff); @@ -213,7 +213,7 @@ static ssize_t resource_show(struct device *dev, /* no address to convey if the pfn instance is disabled */ rc = -ENXIO; } - nd_device_unlock(dev); + device_unlock(dev); return rc; } @@ -225,7 +225,7 @@ static ssize_t size_show(struct device *dev, struct nd_pfn *nd_pfn = to_nd_pfn_safe(dev); ssize_t rc; - nd_device_lock(dev); + device_lock(dev); if (dev->driver) { struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb; u64 offset = __le64_to_cpu(pfn_sb->dataoff); @@ -241,7 +241,7 @@ static ssize_t size_show(struct device *dev, /* no size to convey if the pfn instance is disabled */ rc = -ENXIO; } - nd_device_unlock(dev); + device_unlock(dev); return rc; } @@ -291,6 +291,8 @@ bool is_nd_pfn(struct device *dev) } EXPORT_SYMBOL(is_nd_pfn); +static struct lock_class_key nvdimm_pfn_key; + struct device *nd_pfn_devinit(struct nd_pfn *nd_pfn, struct nd_namespace_common *ndns) { @@ -303,6 +305,7 @@ struct device *nd_pfn_devinit(struct nd_pfn *nd_pfn, nd_pfn->align = nd_pfn_default_alignment(); dev = &nd_pfn->dev; device_initialize(&nd_pfn->dev); + lockdep_set_class(&nd_pfn->dev.mutex, &nvdimm_pfn_key); if (ndns && !__nd_attach_ndns(&nd_pfn->dev, ndns, &nd_pfn->ndns)) { dev_dbg(&ndns->dev, "failed, already claimed by %s\n", dev_name(ndns->claim)); @@ -346,7 +349,7 @@ struct device *nd_pfn_create(struct nd_region *nd_region) nd_pfn = nd_pfn_alloc(nd_region); dev = nd_pfn_devinit(nd_pfn, NULL); - __nd_device_register(dev); + nd_device_register(dev); return dev; } @@ -643,7 +646,7 @@ int nd_pfn_probe(struct device *dev, struct nd_namespace_common *ndns) nd_detach_ndns(pfn_dev, &nd_pfn->ndns); put_device(pfn_dev); } else - __nd_device_register(pfn_dev); + nd_device_register(pfn_dev); return rc; } diff --git a/drivers/nvdimm/pmem.c b/drivers/nvdimm/pmem.c index 58d95242a836..629d10fcf53b 100644 --- a/drivers/nvdimm/pmem.c +++ b/drivers/nvdimm/pmem.c @@ -45,9 +45,25 @@ static struct nd_region *to_region(struct pmem_device *pmem) return to_nd_region(to_dev(pmem)->parent); } -static void hwpoison_clear(struct pmem_device *pmem, - phys_addr_t phys, unsigned int len) +static phys_addr_t to_phys(struct pmem_device *pmem, phys_addr_t offset) { + return pmem->phys_addr + offset; +} + +static sector_t to_sect(struct pmem_device *pmem, phys_addr_t offset) +{ + return (offset - pmem->data_offset) >> SECTOR_SHIFT; +} + +static phys_addr_t to_offset(struct pmem_device *pmem, sector_t sector) +{ + return (sector << SECTOR_SHIFT) + pmem->data_offset; +} + +static void pmem_mkpage_present(struct pmem_device *pmem, phys_addr_t offset, + unsigned int len) +{ + phys_addr_t phys = to_phys(pmem, offset); unsigned long pfn_start, pfn_end, pfn; /* only pmem in the linear map supports HWPoison */ @@ -69,33 +85,40 @@ static void hwpoison_clear(struct pmem_device *pmem, } } -static blk_status_t pmem_clear_poison(struct pmem_device *pmem, - phys_addr_t offset, unsigned int len) +static void pmem_clear_bb(struct pmem_device *pmem, sector_t sector, long blks) { - struct device *dev = to_dev(pmem); - sector_t sector; - long cleared; - blk_status_t rc = BLK_STS_OK; + if (blks == 0) + return; + badblocks_clear(&pmem->bb, sector, blks); + if (pmem->bb_state) + sysfs_notify_dirent(pmem->bb_state); +} - sector = (offset - pmem->data_offset) / 512; +static long __pmem_clear_poison(struct pmem_device *pmem, + phys_addr_t offset, unsigned int len) +{ + phys_addr_t phys = to_phys(pmem, offset); + long cleared = nvdimm_clear_poison(to_dev(pmem), phys, len); - cleared = nvdimm_clear_poison(dev, pmem->phys_addr + offset, len); - if (cleared < len) - rc = BLK_STS_IOERR; - if (cleared > 0 && cleared / 512) { - hwpoison_clear(pmem, pmem->phys_addr + offset, cleared); - cleared /= 512; - dev_dbg(dev, "%#llx clear %ld sector%s\n", - (unsigned long long) sector, cleared, - cleared > 1 ? "s" : ""); - badblocks_clear(&pmem->bb, sector, cleared); - if (pmem->bb_state) - sysfs_notify_dirent(pmem->bb_state); + if (cleared > 0) { + pmem_mkpage_present(pmem, offset, cleared); + arch_invalidate_pmem(pmem->virt_addr + offset, len); } + return cleared; +} + +static blk_status_t pmem_clear_poison(struct pmem_device *pmem, + phys_addr_t offset, unsigned int len) +{ + long cleared = __pmem_clear_poison(pmem, offset, len); - arch_invalidate_pmem(pmem->virt_addr + offset, len); + if (cleared < 0) + return BLK_STS_IOERR; - return rc; + pmem_clear_bb(pmem, to_sect(pmem, offset), cleared >> SECTOR_SHIFT); + if (cleared < len) + return BLK_STS_IOERR; + return BLK_STS_OK; } static void write_pmem(void *pmem_addr, struct page *page, @@ -143,7 +166,7 @@ static blk_status_t pmem_do_read(struct pmem_device *pmem, sector_t sector, unsigned int len) { blk_status_t rc; - phys_addr_t pmem_off = sector * 512 + pmem->data_offset; + phys_addr_t pmem_off = to_offset(pmem, sector); void *pmem_addr = pmem->virt_addr + pmem_off; if (unlikely(is_bad_pmem(&pmem->bb, sector, len))) @@ -158,36 +181,20 @@ static blk_status_t pmem_do_write(struct pmem_device *pmem, struct page *page, unsigned int page_off, sector_t sector, unsigned int len) { - blk_status_t rc = BLK_STS_OK; - bool bad_pmem = false; - phys_addr_t pmem_off = sector * 512 + pmem->data_offset; + phys_addr_t pmem_off = to_offset(pmem, sector); void *pmem_addr = pmem->virt_addr + pmem_off; - if (unlikely(is_bad_pmem(&pmem->bb, sector, len))) - bad_pmem = true; + if (unlikely(is_bad_pmem(&pmem->bb, sector, len))) { + blk_status_t rc = pmem_clear_poison(pmem, pmem_off, len); + + if (rc != BLK_STS_OK) + return rc; + } - /* - * Note that we write the data both before and after - * clearing poison. The write before clear poison - * handles situations where the latest written data is - * preserved and the clear poison operation simply marks - * the address range as valid without changing the data. - * In this case application software can assume that an - * interrupted write will either return the new good - * data or an error. - * - * However, if pmem_clear_poison() leaves the data in an - * indeterminate state we need to perform the write - * after clear poison. - */ flush_dcache_page(page); write_pmem(pmem_addr, page, page_off, len); - if (unlikely(bad_pmem)) { - rc = pmem_clear_poison(pmem, pmem_off, len); - write_pmem(pmem_addr, page, page_off, len); - } - return rc; + return BLK_STS_OK; } static void pmem_submit_bio(struct bio *bio) @@ -255,24 +262,47 @@ static int pmem_rw_page(struct block_device *bdev, sector_t sector, /* see "strong" declaration in tools/testing/nvdimm/pmem-dax.c */ __weak long __pmem_direct_access(struct pmem_device *pmem, pgoff_t pgoff, - long nr_pages, void **kaddr, pfn_t *pfn) + long nr_pages, enum dax_access_mode mode, void **kaddr, + pfn_t *pfn) { resource_size_t offset = PFN_PHYS(pgoff) + pmem->data_offset; - - if (unlikely(is_bad_pmem(&pmem->bb, PFN_PHYS(pgoff) / 512, - PFN_PHYS(nr_pages)))) - return -EIO; + sector_t sector = PFN_PHYS(pgoff) >> SECTOR_SHIFT; + unsigned int num = PFN_PHYS(nr_pages) >> SECTOR_SHIFT; + struct badblocks *bb = &pmem->bb; + sector_t first_bad; + int num_bad; if (kaddr) *kaddr = pmem->virt_addr + offset; if (pfn) *pfn = phys_to_pfn_t(pmem->phys_addr + offset, pmem->pfn_flags); + if (bb->count && + badblocks_check(bb, sector, num, &first_bad, &num_bad)) { + long actual_nr; + + if (mode != DAX_RECOVERY_WRITE) + return -EIO; + + /* + * Set the recovery stride is set to kernel page size because + * the underlying driver and firmware clear poison functions + * don't appear to handle large chunk(such as 2MiB) reliably. + */ + actual_nr = PHYS_PFN( + PAGE_ALIGN((first_bad - sector) << SECTOR_SHIFT)); + dev_dbg(pmem->bb.dev, "start sector(%llu), nr_pages(%ld), first_bad(%llu), actual_nr(%ld)\n", + sector, nr_pages, first_bad, actual_nr); + if (actual_nr) + return actual_nr; + return 1; + } + /* - * If badblocks are present, limit known good range to the - * requested range. + * If badblocks are present but not in the range, limit known good range + * to the requested range. */ - if (unlikely(pmem->bb.count)) + if (bb->count) return nr_pages; return PHYS_PFN(pmem->size - pmem->pfn_pad - offset); } @@ -294,16 +324,73 @@ static int pmem_dax_zero_page_range(struct dax_device *dax_dev, pgoff_t pgoff, } static long pmem_dax_direct_access(struct dax_device *dax_dev, - pgoff_t pgoff, long nr_pages, void **kaddr, pfn_t *pfn) + pgoff_t pgoff, long nr_pages, enum dax_access_mode mode, + void **kaddr, pfn_t *pfn) +{ + struct pmem_device *pmem = dax_get_private(dax_dev); + + return __pmem_direct_access(pmem, pgoff, nr_pages, mode, kaddr, pfn); +} + +/* + * The recovery write thread started out as a normal pwrite thread and + * when the filesystem was told about potential media error in the + * range, filesystem turns the normal pwrite to a dax_recovery_write. + * + * The recovery write consists of clearing media poison, clearing page + * HWPoison bit, reenable page-wide read-write permission, flush the + * caches and finally write. A competing pread thread will be held + * off during the recovery process since data read back might not be + * valid, and this is achieved by clearing the badblock records after + * the recovery write is complete. Competing recovery write threads + * are already serialized by writer lock held by dax_iomap_rw(). + */ +static size_t pmem_recovery_write(struct dax_device *dax_dev, pgoff_t pgoff, + void *addr, size_t bytes, struct iov_iter *i) { struct pmem_device *pmem = dax_get_private(dax_dev); + size_t olen, len, off; + phys_addr_t pmem_off; + struct device *dev = pmem->bb.dev; + long cleared; + + off = offset_in_page(addr); + len = PFN_PHYS(PFN_UP(off + bytes)); + if (!is_bad_pmem(&pmem->bb, PFN_PHYS(pgoff) >> SECTOR_SHIFT, len)) + return _copy_from_iter_flushcache(addr, bytes, i); + + /* + * Not page-aligned range cannot be recovered. This should not + * happen unless something else went wrong. + */ + if (off || !PAGE_ALIGNED(bytes)) { + dev_dbg(dev, "Found poison, but addr(%p) or bytes(%#zx) not page aligned\n", + addr, bytes); + return 0; + } + + pmem_off = PFN_PHYS(pgoff) + pmem->data_offset; + cleared = __pmem_clear_poison(pmem, pmem_off, len); + if (cleared > 0 && cleared < len) { + dev_dbg(dev, "poison cleared only %ld out of %zu bytes\n", + cleared, len); + return 0; + } + if (cleared < 0) { + dev_dbg(dev, "poison clear failed: %ld\n", cleared); + return 0; + } + + olen = _copy_from_iter_flushcache(addr, bytes, i); + pmem_clear_bb(pmem, to_sect(pmem, pmem_off), cleared >> SECTOR_SHIFT); - return __pmem_direct_access(pmem, pgoff, nr_pages, kaddr, pfn); + return olen; } static const struct dax_operations pmem_dax_ops = { .direct_access = pmem_dax_direct_access, .zero_page_range = pmem_dax_zero_page_range, + .recovery_write = pmem_recovery_write, }; static ssize_t write_cache_show(struct device *dev, @@ -573,7 +660,7 @@ static void nd_pmem_remove(struct device *dev) nvdimm_namespace_detach_btt(to_nd_btt(dev)); else { /* - * Note, this assumes nd_device_lock() context to not + * Note, this assumes device_lock() context to not * race nd_pmem_notify() */ sysfs_put(pmem->bb_state); diff --git a/drivers/nvdimm/pmem.h b/drivers/nvdimm/pmem.h index 1f51a2361429..392b0b38acb9 100644 --- a/drivers/nvdimm/pmem.h +++ b/drivers/nvdimm/pmem.h @@ -8,6 +8,8 @@ #include <linux/pfn_t.h> #include <linux/fs.h> +enum dax_access_mode; + /* this definition is in it's own header for tools/testing/nvdimm to consume */ struct pmem_device { /* One contiguous memory region per device */ @@ -28,7 +30,8 @@ struct pmem_device { }; long __pmem_direct_access(struct pmem_device *pmem, pgoff_t pgoff, - long nr_pages, void **kaddr, pfn_t *pfn); + long nr_pages, enum dax_access_mode mode, void **kaddr, + pfn_t *pfn); #ifdef CONFIG_MEMORY_FAILURE static inline bool test_and_clear_pmem_poison(struct page *page) diff --git a/drivers/nvdimm/region.c b/drivers/nvdimm/region.c index 188560b1c110..390123d293ea 100644 --- a/drivers/nvdimm/region.c +++ b/drivers/nvdimm/region.c @@ -95,7 +95,7 @@ static void nd_region_remove(struct device *dev) nvdimm_bus_unlock(dev); /* - * Note, this assumes nd_device_lock() context to not race + * Note, this assumes device_lock() context to not race * nd_region_notify() */ sysfs_put(nd_region->bb_state); diff --git a/drivers/nvdimm/region_devs.c b/drivers/nvdimm/region_devs.c index 0cb274c2b508..d976260eca7a 100644 --- a/drivers/nvdimm/region_devs.c +++ b/drivers/nvdimm/region_devs.c @@ -279,7 +279,7 @@ static ssize_t set_cookie_show(struct device *dev, * the v1.1 namespace label cookie definition. To read all this * data we need to wait for probing to settle. */ - nd_device_lock(dev); + device_lock(dev); nvdimm_bus_lock(dev); wait_nvdimm_bus_probe_idle(dev); if (nd_region->ndr_mappings) { @@ -296,7 +296,7 @@ static ssize_t set_cookie_show(struct device *dev, } } nvdimm_bus_unlock(dev); - nd_device_unlock(dev); + device_unlock(dev); if (rc) return rc; @@ -353,12 +353,12 @@ static ssize_t available_size_show(struct device *dev, * memory nvdimm_bus_lock() is dropped, but that's userspace's * problem to not race itself. */ - nd_device_lock(dev); + device_lock(dev); nvdimm_bus_lock(dev); wait_nvdimm_bus_probe_idle(dev); available = nd_region_available_dpa(nd_region); nvdimm_bus_unlock(dev); - nd_device_unlock(dev); + device_unlock(dev); return sprintf(buf, "%llu\n", available); } @@ -370,12 +370,12 @@ static ssize_t max_available_extent_show(struct device *dev, struct nd_region *nd_region = to_nd_region(dev); unsigned long long available = 0; - nd_device_lock(dev); + device_lock(dev); nvdimm_bus_lock(dev); wait_nvdimm_bus_probe_idle(dev); available = nd_region_allocatable_dpa(nd_region); nvdimm_bus_unlock(dev); - nd_device_unlock(dev); + device_unlock(dev); return sprintf(buf, "%llu\n", available); } @@ -549,12 +549,12 @@ static ssize_t region_badblocks_show(struct device *dev, struct nd_region *nd_region = to_nd_region(dev); ssize_t rc; - nd_device_lock(dev); + device_lock(dev); if (dev->driver) rc = badblocks_show(&nd_region->bb, buf, 0); else rc = -ENXIO; - nd_device_unlock(dev); + device_unlock(dev); return rc; } @@ -949,6 +949,8 @@ static unsigned long default_align(struct nd_region *nd_region) return align; } +static struct lock_class_key nvdimm_region_key; + static struct nd_region *nd_region_create(struct nvdimm_bus *nvdimm_bus, struct nd_region_desc *ndr_desc, const struct device_type *dev_type, const char *caller) @@ -1035,6 +1037,8 @@ static struct nd_region *nd_region_create(struct nvdimm_bus *nvdimm_bus, else nd_region->flush = NULL; + device_initialize(dev); + lockdep_set_class(&dev->mutex, &nvdimm_region_key); nd_device_register(dev); return nd_region; diff --git a/drivers/nvdimm/security.c b/drivers/nvdimm/security.c index 4b80150e4afa..b5aa55c61461 100644 --- a/drivers/nvdimm/security.c +++ b/drivers/nvdimm/security.c @@ -379,11 +379,6 @@ static int security_overwrite(struct nvdimm *nvdimm, unsigned int keyid) || !nvdimm->sec.flags) return -EOPNOTSUPP; - if (dev->driver == NULL) { - dev_dbg(dev, "Unable to overwrite while DIMM active.\n"); - return -EINVAL; - } - rc = check_security_state(nvdimm); if (rc) return rc; |