1 files changed, 100 insertions, 58 deletions

diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index ba6508455e18..1a8d32a4a5c3 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -22,7 +22,7 @@
 #include <linux/nvme_ioctl.h>
 #include <linux/pm_qos.h>
 #include <linux/ratelimit.h>
-#include <asm/unaligned.h>
+#include <linux/unaligned.h>
 
 #include "nvme.h"
 #include "fabrics.h"
@@ -42,6 +42,8 @@ struct nvme_ns_info {
 	bool is_readonly;
 	bool is_ready;
 	bool is_removed;
+	bool is_rotational;
+	bool no_vwc;
 };
 
 unsigned int admin_timeout = 60;
@@ -92,6 +94,17 @@ MODULE_PARM_DESC(apst_secondary_latency_tol_us,
 	"secondary APST latency tolerance in us");
 
 /*
+ * Older kernels didn't enable protection information if it was at an offset.
+ * Newer kernels do, so it breaks reads on the upgrade if such formats were
+ * used in prior kernels since the metadata written did not contain a valid
+ * checksum.
+ */
+static bool disable_pi_offsets = false;
+module_param(disable_pi_offsets, bool, 0444);
+MODULE_PARM_DESC(disable_pi_offsets,
+	"disable protection information if it has an offset");
+
+/*
  * nvme_wq - hosts nvme related works that are not reset or delete
  * nvme_reset_wq - hosts nvme reset works
  * nvme_delete_wq - hosts nvme delete works
@@ -1292,14 +1305,12 @@ static void nvme_queue_keep_alive_work(struct nvme_ctrl *ctrl)
 	queue_delayed_work(nvme_wq, &ctrl->ka_work, delay);
 }
 
-static enum rq_end_io_ret nvme_keep_alive_end_io(struct request *rq,
-						 blk_status_t status)
+static void nvme_keep_alive_finish(struct request *rq,
+		blk_status_t status, struct nvme_ctrl *ctrl)
 {
-	struct nvme_ctrl *ctrl = rq->end_io_data;
-	unsigned long flags;
-	bool startka = false;
 	unsigned long rtt = jiffies - (rq->deadline - rq->timeout);
 	unsigned long delay = nvme_keep_alive_work_period(ctrl);
+	enum nvme_ctrl_state state = nvme_ctrl_state(ctrl);
 
 	/*
 	 * Subtract off the keepalive RTT so nvme_keep_alive_work runs
@@ -1313,25 +1324,17 @@ static enum rq_end_io_ret nvme_keep_alive_end_io(struct request *rq,
 		delay = 0;
 	}
 
-	blk_mq_free_request(rq);
-
 	if (status) {
 		dev_err(ctrl->device,
 			"failed nvme_keep_alive_end_io error=%d\n",
 				status);
-		return RQ_END_IO_NONE;
+		return;
 	}
 
 	ctrl->ka_last_check_time = jiffies;
 	ctrl->comp_seen = false;
-	spin_lock_irqsave(&ctrl->lock, flags);
-	if (ctrl->state == NVME_CTRL_LIVE ||
-	    ctrl->state == NVME_CTRL_CONNECTING)
-		startka = true;
-	spin_unlock_irqrestore(&ctrl->lock, flags);
-	if (startka)
+	if (state == NVME_CTRL_LIVE || state == NVME_CTRL_CONNECTING)
 		queue_delayed_work(nvme_wq, &ctrl->ka_work, delay);
-	return RQ_END_IO_NONE;
 }
 
 static void nvme_keep_alive_work(struct work_struct *work)
@@ -1340,6 +1343,7 @@ static void nvme_keep_alive_work(struct work_struct *work)
 			struct nvme_ctrl, ka_work);
 	bool comp_seen = ctrl->comp_seen;
 	struct request *rq;
+	blk_status_t status;
 
 	ctrl->ka_last_check_time = jiffies;
 
@@ -1362,9 +1366,9 @@ static void nvme_keep_alive_work(struct work_struct *work)
 	nvme_init_request(rq, &ctrl->ka_cmd);
 
 	rq->timeout = ctrl->kato * HZ;
-	rq->end_io = nvme_keep_alive_end_io;
-	rq->end_io_data = ctrl;
-	blk_execute_rq_nowait(rq, false);
+	status = blk_execute_rq(rq, false);
+	nvme_keep_alive_finish(rq, status, ctrl);
+	blk_mq_free_request(rq);
 }
 
 static void nvme_start_keep_alive(struct nvme_ctrl *ctrl)
@@ -1399,17 +1403,30 @@ static void nvme_update_keep_alive(struct nvme_ctrl *ctrl,
 	nvme_start_keep_alive(ctrl);
 }
 
-/*
- * In NVMe 1.0 the CNS field was just a binary controller or namespace
- * flag, thus sending any new CNS opcodes has a big chance of not working.
- * Qemu unfortunately had that bug after reporting a 1.1 version compliance
- * (but not for any later version).
- */
-static bool nvme_ctrl_limited_cns(struct nvme_ctrl *ctrl)
+static bool nvme_id_cns_ok(struct nvme_ctrl *ctrl, u8 cns)
 {
-	if (ctrl->quirks & NVME_QUIRK_IDENTIFY_CNS)
-		return ctrl->vs < NVME_VS(1, 2, 0);
-	return ctrl->vs < NVME_VS(1, 1, 0);
+	/*
+	 * The CNS field occupies a full byte starting with NVMe 1.2
+	 */
+	if (ctrl->vs >= NVME_VS(1, 2, 0))
+		return true;
+
+	/*
+	 * NVMe 1.1 expanded the CNS value to two bits, which means values
+	 * larger than that could get truncated and treated as an incorrect
+	 * value.
+	 *
+	 * Qemu implemented 1.0 behavior for controllers claiming 1.1
+	 * compliance, so they need to be quirked here.
+	 */
+	if (ctrl->vs >= NVME_VS(1, 1, 0) &&
+	    !(ctrl->quirks & NVME_QUIRK_IDENTIFY_CNS))
+		return cns <= 3;
+
+	/*
+	 * NVMe 1.0 used a single bit for the CNS value.
+	 */
+	return cns <= 1;
 }
 
 static int nvme_identify_ctrl(struct nvme_ctrl *dev, struct nvme_id_ctrl **id)
@@ -1624,6 +1641,8 @@ static int nvme_ns_info_from_id_cs_indep(struct nvme_ctrl *ctrl,
 		info->is_shared = id->nmic & NVME_NS_NMIC_SHARED;
 		info->is_readonly = id->nsattr & NVME_NS_ATTR_RO;
 		info->is_ready = id->nstat & NVME_NSTAT_NRDY;
+		info->is_rotational = id->nsfeat & NVME_NS_ROTATIONAL;
+		info->no_vwc = id->nsfeat & NVME_NS_VWC_NOT_PRESENT;
 	}
 	kfree(id);
 	return ret;
@@ -1922,8 +1941,12 @@ static void nvme_configure_metadata(struct nvme_ctrl *ctrl,
 
 	if (head->pi_size && head->ms >= head->pi_size)
 		head->pi_type = id->dps & NVME_NS_DPS_PI_MASK;
-	if (!(id->dps & NVME_NS_DPS_PI_FIRST))
-		info->pi_offset = head->ms - head->pi_size;
+	if (!(id->dps & NVME_NS_DPS_PI_FIRST)) {
+		if (disable_pi_offsets)
+			head->pi_type = 0;
+		else
+			info->pi_offset = head->ms - head->pi_size;
+	}
 
 	if (ctrl->ops->flags & NVME_F_FABRICS) {
 		/*
@@ -2166,11 +2189,14 @@ static int nvme_update_ns_info_block(struct nvme_ns *ns,
 	    ns->head->ids.csi == NVME_CSI_ZNS)
 		nvme_update_zone_info(ns, &lim, &zi);
 
-	if (ns->ctrl->vwc & NVME_CTRL_VWC_PRESENT)
+	if ((ns->ctrl->vwc & NVME_CTRL_VWC_PRESENT) && !info->no_vwc)
 		lim.features |= BLK_FEAT_WRITE_CACHE | BLK_FEAT_FUA;
 	else
 		lim.features &= ~(BLK_FEAT_WRITE_CACHE | BLK_FEAT_FUA);
 
+	if (info->is_rotational)
+		lim.features |= BLK_FEAT_ROTATIONAL;
+
 	/*
 	 * Register a metadata profile for PI, or the plain non-integrity NVMe
 	 * metadata masquerading as Type 0 if supported, otherwise reject block
@@ -2458,8 +2484,13 @@ int nvme_enable_ctrl(struct nvme_ctrl *ctrl)
 	else
 		ctrl->ctrl_config = NVME_CC_CSS_NVM;
 
-	if (ctrl->cap & NVME_CAP_CRMS_CRWMS && ctrl->cap & NVME_CAP_CRMS_CRIMS)
-		ctrl->ctrl_config |= NVME_CC_CRIME;
+	/*
+	 * Setting CRIME results in CSTS.RDY before the media is ready. This
+	 * makes it possible for media related commands to return the error
+	 * NVME_SC_ADMIN_COMMAND_MEDIA_NOT_READY. Until the driver is
+	 * restructured to handle retries, disable CC.CRIME.
+	 */
+	ctrl->ctrl_config &= ~NVME_CC_CRIME;
 
 	ctrl->ctrl_config |= (NVME_CTRL_PAGE_SHIFT - 12) << NVME_CC_MPS_SHIFT;
 	ctrl->ctrl_config |= NVME_CC_AMS_RR | NVME_CC_SHN_NONE;
@@ -2489,10 +2520,7 @@ int nvme_enable_ctrl(struct nvme_ctrl *ctrl)
 		 * devices are known to get this wrong. Use the larger of the
 		 * two values.
 		 */
-		if (ctrl->ctrl_config & NVME_CC_CRIME)
-			ready_timeout = NVME_CRTO_CRIMT(crto);
-		else
-			ready_timeout = NVME_CRTO_CRWMT(crto);
+		ready_timeout = NVME_CRTO_CRWMT(crto);
 
 		if (ready_timeout < timeout)
 			dev_warn_once(ctrl->device, "bad crto:%x cap:%llx\n",
@@ -3111,7 +3139,7 @@ static int nvme_init_non_mdts_limits(struct nvme_ctrl *ctrl)
 		ctrl->max_zeroes_sectors = 0;
 
 	if (ctrl->subsys->subtype != NVME_NQN_NVME ||
-	    nvme_ctrl_limited_cns(ctrl) ||
+	    !nvme_id_cns_ok(ctrl, NVME_ID_CNS_CS_CTRL) ||
 	    test_bit(NVME_CTRL_SKIP_ID_CNS_CS, &ctrl->flags))
 		return 0;
 
@@ -3615,6 +3643,7 @@ static struct nvme_ns_head *nvme_alloc_ns_head(struct nvme_ctrl *ctrl,
 	head->ns_id = info->nsid;
 	head->ids = info->ids;
 	head->shared = info->is_shared;
+	head->rotational = info->is_rotational;
 	ratelimit_state_init(&head->rs_nuse, 5 * HZ, 1);
 	ratelimit_set_flags(&head->rs_nuse, RATELIMIT_MSG_ON_RELEASE);
 	kref_init(&head->ref);
@@ -3774,7 +3803,8 @@ struct nvme_ns *nvme_find_get_ns(struct nvme_ctrl *ctrl, unsigned nsid)
 	int srcu_idx;
 
 	srcu_idx = srcu_read_lock(&ctrl->srcu);
-	list_for_each_entry_rcu(ns, &ctrl->namespaces, list) {
+	list_for_each_entry_srcu(ns, &ctrl->namespaces, list,
+				 srcu_read_lock_held(&ctrl->srcu)) {
 		if (ns->head->ns_id == nsid) {
 			if (!nvme_get_ns(ns))
 				continue;
@@ -3995,7 +4025,7 @@ static void nvme_scan_ns(struct nvme_ctrl *ctrl, unsigned nsid)
 {
 	struct nvme_ns_info info = { .nsid = nsid };
 	struct nvme_ns *ns;
-	int ret;
+	int ret = 1;
 
 	if (nvme_identify_ns_descs(ctrl, &info))
 		return;
@@ -4012,9 +4042,10 @@ static void nvme_scan_ns(struct nvme_ctrl *ctrl, unsigned nsid)
 	 * set up a namespace.  If not fall back to the legacy version.
 	 */
 	if ((ctrl->cap & NVME_CAP_CRMS_CRIMS) ||
-	    (info.ids.csi != NVME_CSI_NVM && info.ids.csi != NVME_CSI_ZNS))
+	    (info.ids.csi != NVME_CSI_NVM && info.ids.csi != NVME_CSI_ZNS) ||
+	    ctrl->vs >= NVME_VS(2, 0, 0))
 		ret = nvme_ns_info_from_id_cs_indep(ctrl, &info);
-	else
+	if (ret > 0)
 		ret = nvme_ns_info_from_identify(ctrl, &info);
 
 	if (info.is_removed)
@@ -4207,7 +4238,7 @@ static void nvme_scan_work(struct work_struct *work)
 	}
 
 	mutex_lock(&ctrl->scan_lock);
-	if (nvme_ctrl_limited_cns(ctrl)) {
+	if (!nvme_id_cns_ok(ctrl, NVME_ID_CNS_NS_ACTIVE_LIST)) {
 		nvme_scan_ns_sequential(ctrl);
 	} else {
 		/*
@@ -4858,7 +4889,8 @@ void nvme_mark_namespaces_dead(struct nvme_ctrl *ctrl)
 	int srcu_idx;
 
 	srcu_idx = srcu_read_lock(&ctrl->srcu);
-	list_for_each_entry_rcu(ns, &ctrl->namespaces, list)
+	list_for_each_entry_srcu(ns, &ctrl->namespaces, list,
+				 srcu_read_lock_held(&ctrl->srcu))
 		blk_mark_disk_dead(ns->disk);
 	srcu_read_unlock(&ctrl->srcu, srcu_idx);
 }
@@ -4870,8 +4902,9 @@ void nvme_unfreeze(struct nvme_ctrl *ctrl)
 	int srcu_idx;
 
 	srcu_idx = srcu_read_lock(&ctrl->srcu);
-	list_for_each_entry_rcu(ns, &ctrl->namespaces, list)
-		blk_mq_unfreeze_queue(ns->queue);
+	list_for_each_entry_srcu(ns, &ctrl->namespaces, list,
+				 srcu_read_lock_held(&ctrl->srcu))
+		blk_mq_unfreeze_queue_non_owner(ns->queue);
 	srcu_read_unlock(&ctrl->srcu, srcu_idx);
 	clear_bit(NVME_CTRL_FROZEN, &ctrl->flags);
 }
@@ -4883,7 +4916,8 @@ int nvme_wait_freeze_timeout(struct nvme_ctrl *ctrl, long timeout)
 	int srcu_idx;
 
 	srcu_idx = srcu_read_lock(&ctrl->srcu);
-	list_for_each_entry_rcu(ns, &ctrl->namespaces, list) {
+	list_for_each_entry_srcu(ns, &ctrl->namespaces, list,
+				 srcu_read_lock_held(&ctrl->srcu)) {
 		timeout = blk_mq_freeze_queue_wait_timeout(ns->queue, timeout);
 		if (timeout <= 0)
 			break;
@@ -4899,7 +4933,8 @@ void nvme_wait_freeze(struct nvme_ctrl *ctrl)
 	int srcu_idx;
 
 	srcu_idx = srcu_read_lock(&ctrl->srcu);
-	list_for_each_entry_rcu(ns, &ctrl->namespaces, list)
+	list_for_each_entry_srcu(ns, &ctrl->namespaces, list,
+				 srcu_read_lock_held(&ctrl->srcu))
 		blk_mq_freeze_queue_wait(ns->queue);
 	srcu_read_unlock(&ctrl->srcu, srcu_idx);
 }
@@ -4912,8 +4947,14 @@ void nvme_start_freeze(struct nvme_ctrl *ctrl)
 
 	set_bit(NVME_CTRL_FROZEN, &ctrl->flags);
 	srcu_idx = srcu_read_lock(&ctrl->srcu);
-	list_for_each_entry_rcu(ns, &ctrl->namespaces, list)
-		blk_freeze_queue_start(ns->queue);
+	list_for_each_entry_srcu(ns, &ctrl->namespaces, list,
+				 srcu_read_lock_held(&ctrl->srcu))
+		/*
+		 * Typical non_owner use case is from pci driver, in which
+		 * start_freeze is called from timeout work function, but
+		 * unfreeze is done in reset work context
+		 */
+		blk_freeze_queue_start_non_owner(ns->queue);
 	srcu_read_unlock(&ctrl->srcu, srcu_idx);
 }
 EXPORT_SYMBOL_GPL(nvme_start_freeze);
@@ -4960,7 +5001,8 @@ void nvme_sync_io_queues(struct nvme_ctrl *ctrl)
 	int srcu_idx;
 
 	srcu_idx = srcu_read_lock(&ctrl->srcu);
-	list_for_each_entry_rcu(ns, &ctrl->namespaces, list)
+	list_for_each_entry_srcu(ns, &ctrl->namespaces, list,
+				 srcu_read_lock_held(&ctrl->srcu))
 		blk_sync_queue(ns->queue);
 	srcu_read_unlock(&ctrl->srcu, srcu_idx);
 }
@@ -5008,6 +5050,8 @@ static inline void _nvme_check_size(void)
 	BUILD_BUG_ON(sizeof(struct nvme_id_ctrl_nvm) != NVME_IDENTIFY_DATA_SIZE);
 	BUILD_BUG_ON(sizeof(struct nvme_lba_range_type) != 64);
 	BUILD_BUG_ON(sizeof(struct nvme_smart_log) != 512);
+	BUILD_BUG_ON(sizeof(struct nvme_endurance_group_log) != 512);
+	BUILD_BUG_ON(sizeof(struct nvme_rotational_media_log) != 512);
 	BUILD_BUG_ON(sizeof(struct nvme_dbbuf) != 64);
 	BUILD_BUG_ON(sizeof(struct nvme_directive_cmd) != 64);
 	BUILD_BUG_ON(sizeof(struct nvme_feat_host_behavior) != 512);
@@ -5016,22 +5060,20 @@ static inline void _nvme_check_size(void)
 
 static int __init nvme_core_init(void)
 {
+	unsigned int wq_flags = WQ_UNBOUND | WQ_MEM_RECLAIM | WQ_SYSFS;
 	int result = -ENOMEM;
 
 	_nvme_check_size();
 
-	nvme_wq = alloc_workqueue("nvme-wq",
-			WQ_UNBOUND | WQ_MEM_RECLAIM | WQ_SYSFS, 0);
+	nvme_wq = alloc_workqueue("nvme-wq", wq_flags, 0);
 	if (!nvme_wq)
 		goto out;
 
-	nvme_reset_wq = alloc_workqueue("nvme-reset-wq",
-			WQ_UNBOUND | WQ_MEM_RECLAIM | WQ_SYSFS, 0);
+	nvme_reset_wq = alloc_workqueue("nvme-reset-wq", wq_flags, 0);
 	if (!nvme_reset_wq)
 		goto destroy_wq;
 
-	nvme_delete_wq = alloc_workqueue("nvme-delete-wq",
-			WQ_UNBOUND | WQ_MEM_RECLAIM | WQ_SYSFS, 0);
+	nvme_delete_wq = alloc_workqueue("nvme-delete-wq", wq_flags, 0);
 	if (!nvme_delete_wq)
 		goto destroy_reset_wq;