Merge tag 'nvme-5.19-2022-05-19' of git://git.infradead.org/nvme into for-5.19/drivers

Pull NVMe updates from Christoph:

"nvme updates for Linux 5.19

 - set non-mdts limits in nvme_scan_work (Chaitanya Kulkarni)
 - add support for TP4084 - Time-to-Ready Enhancements (me)"

* tag 'nvme-5.19-2022-05-19' of git://git.infradead.org/nvme:
  nvme: set non-mdts limits in nvme_scan_work
  nvme: add support for TP4084 - Time-to-Ready Enhancements

3 files changed, 117 insertions, 10 deletions

diff --git a/drivers/nvme/host/constants.c b/drivers/nvme/host/constants.c
index 1dd1d78de295..4910543f00ff 100644
--- a/drivers/nvme/host/constants.c
+++ b/drivers/nvme/host/constants.c
@@ -91,6 +91,7 @@ static const char * const nvme_statuses[] = {
 	[NVME_SC_NS_WRITE_PROTECTED] = "Namespace is Write Protected",
 	[NVME_SC_CMD_INTERRUPTED] = "Command Interrupted",
 	[NVME_SC_TRANSIENT_TR_ERR] = "Transient Transport Error",
+	[NVME_SC_ADMIN_COMMAND_MEDIA_NOT_READY] = "Admin Command Media Not Ready",
 	[NVME_SC_INVALID_IO_CMD_SET] = "Invalid IO Command Set",
 	[NVME_SC_LBA_RANGE] = "LBA Out of Range",
 	[NVME_SC_CAP_EXCEEDED] = "Capacity Exceeded",
diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index 42f9772abc4d..a7660e4be55b 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -1427,6 +1427,32 @@ out_free_id:
 	return error;
 }
 
+static int nvme_identify_ns_cs_indep(struct nvme_ctrl *ctrl, unsigned nsid,
+			struct nvme_id_ns_cs_indep **id)
+{
+	struct nvme_command c = {
+		.identify.opcode	= nvme_admin_identify,
+		.identify.nsid		= cpu_to_le32(nsid),
+		.identify.cns		= NVME_ID_CNS_NS_CS_INDEP,
+	};
+	int ret;
+
+	*id = kmalloc(sizeof(**id), GFP_KERNEL);
+	if (!*id)
+		return -ENOMEM;
+
+	ret = nvme_submit_sync_cmd(ctrl->admin_q, &c, *id, sizeof(**id));
+	if (ret) {
+		dev_warn(ctrl->device,
+			 "Identify namespace (CS independent) failed (%d)\n",
+			 ret);
+		kfree(*id);
+		return ret;
+	}
+
+	return 0;
+}
+
 static int nvme_features(struct nvme_ctrl *dev, u8 op, unsigned int fid,
 		unsigned int dword11, void *buffer, size_t buflen, u32 *result)
 {
@@ -2103,10 +2129,9 @@ static const struct block_device_operations nvme_bdev_ops = {
 	.pr_ops		= &nvme_pr_ops,
 };
 
-static int nvme_wait_ready(struct nvme_ctrl *ctrl, u64 cap, bool enabled)
+static int nvme_wait_ready(struct nvme_ctrl *ctrl, u32 timeout, bool enabled)
 {
-	unsigned long timeout =
-		((NVME_CAP_TIMEOUT(cap) + 1) * HZ / 2) + jiffies;
+	unsigned long timeout_jiffies = ((timeout + 1) * HZ / 2) + jiffies;
 	u32 csts, bit = enabled ? NVME_CSTS_RDY : 0;
 	int ret;
 
@@ -2119,7 +2144,7 @@ static int nvme_wait_ready(struct nvme_ctrl *ctrl, u64 cap, bool enabled)
 		usleep_range(1000, 2000);
 		if (fatal_signal_pending(current))
 			return -EINTR;
-		if (time_after(jiffies, timeout)) {
+		if (time_after(jiffies, timeout_jiffies)) {
 			dev_err(ctrl->device,
 				"Device not ready; aborting %s, CSTS=0x%x\n",
 				enabled ? "initialisation" : "reset", csts);
@@ -2150,13 +2175,14 @@ int nvme_disable_ctrl(struct nvme_ctrl *ctrl)
 	if (ctrl->quirks & NVME_QUIRK_DELAY_BEFORE_CHK_RDY)
 		msleep(NVME_QUIRK_DELAY_AMOUNT);
 
-	return nvme_wait_ready(ctrl, ctrl->cap, false);
+	return nvme_wait_ready(ctrl, NVME_CAP_TIMEOUT(ctrl->cap), false);
 }
 EXPORT_SYMBOL_GPL(nvme_disable_ctrl);
 
 int nvme_enable_ctrl(struct nvme_ctrl *ctrl)
 {
 	unsigned dev_page_min;
+	u32 timeout;
 	int ret;
 
 	ret = ctrl->ops->reg_read64(ctrl, NVME_REG_CAP, &ctrl->cap);
@@ -2177,6 +2203,27 @@ int nvme_enable_ctrl(struct nvme_ctrl *ctrl)
 		ctrl->ctrl_config = NVME_CC_CSS_CSI;
 	else
 		ctrl->ctrl_config = NVME_CC_CSS_NVM;
+
+	if (ctrl->cap & NVME_CAP_CRMS_CRWMS) {
+		u32 crto;
+
+		ret = ctrl->ops->reg_read32(ctrl, NVME_REG_CRTO, &crto);
+		if (ret) {
+			dev_err(ctrl->device, "Reading CRTO failed (%d)\n",
+				ret);
+			return ret;
+		}
+
+		if (ctrl->cap & NVME_CAP_CRMS_CRIMS) {
+			ctrl->ctrl_config |= NVME_CC_CRIME;
+			timeout = NVME_CRTO_CRIMT(crto);
+		} else {
+			timeout = NVME_CRTO_CRWMT(crto);
+		}
+	} else {
+		timeout = NVME_CAP_TIMEOUT(ctrl->cap);
+	}
+
 	ctrl->ctrl_config |= (NVME_CTRL_PAGE_SHIFT - 12) << NVME_CC_MPS_SHIFT;
 	ctrl->ctrl_config |= NVME_CC_AMS_RR | NVME_CC_SHN_NONE;
 	ctrl->ctrl_config |= NVME_CC_IOSQES | NVME_CC_IOCQES;
@@ -2185,7 +2232,7 @@ int nvme_enable_ctrl(struct nvme_ctrl *ctrl)
 	ret = ctrl->ops->reg_write32(ctrl, NVME_REG_CC, ctrl->ctrl_config);
 	if (ret)
 		return ret;
-	return nvme_wait_ready(ctrl, ctrl->cap, true);
+	return nvme_wait_ready(ctrl, timeout, true);
 }
 EXPORT_SYMBOL_GPL(nvme_enable_ctrl);
 
@@ -3082,10 +3129,6 @@ int nvme_init_ctrl_finish(struct nvme_ctrl *ctrl)
 	if (ret)
 		return ret;
 
-	ret = nvme_init_non_mdts_limits(ctrl);
-	if (ret < 0)
-		return ret;
-
 	ret = nvme_configure_apst(ctrl);
 	if (ret < 0)
 		return ret;
@@ -4092,11 +4135,26 @@ out:
 static void nvme_validate_or_alloc_ns(struct nvme_ctrl *ctrl, unsigned nsid)
 {
 	struct nvme_ns_ids ids = { };
+	struct nvme_id_ns_cs_indep *id;
 	struct nvme_ns *ns;
+	bool ready = true;
 
 	if (nvme_identify_ns_descs(ctrl, nsid, &ids))
 		return;
 
+	/*
+	 * Check if the namespace is ready.  If not ignore it, we will get an
+	 * AEN once it becomes ready and restart the scan.
+	 */
+	if ((ctrl->cap & NVME_CAP_CRMS_CRIMS) &&
+	    !nvme_identify_ns_cs_indep(ctrl, nsid, &id)) {
+		ready = id->nstat & NVME_NSTAT_NRDY;
+		kfree(id);
+	}
+
+	if (!ready)
+		return;
+
 	ns = nvme_find_get_ns(ctrl, nsid);
 	if (ns) {
 		nvme_validate_ns(ns, &ids);
@@ -4239,11 +4297,26 @@ static void nvme_scan_work(struct work_struct *work)
 {
 	struct nvme_ctrl *ctrl =
 		container_of(work, struct nvme_ctrl, scan_work);
+	int ret;
 
 	/* No tagset on a live ctrl means IO queues could not created */
 	if (ctrl->state != NVME_CTRL_LIVE || !ctrl->tagset)
 		return;
 
+	/*
+	 * Identify controller limits can change at controller reset due to
+	 * new firmware download, even though it is not common we cannot ignore
+	 * such scenario. Controller's non-mdts limits are reported in the unit
+	 * of logical blocks that is dependent on the format of attached
+	 * namespace. Hence re-read the limits at the time of ns allocation.
+	 */
+	ret = nvme_init_non_mdts_limits(ctrl);
+	if (ret < 0) {
+		dev_warn(ctrl->device,
+			"reading non-mdts-limits failed: %d\n", ret);
+		return;
+	}
+
 	if (test_and_clear_bit(NVME_AER_NOTICE_NS_CHANGED, &ctrl->events)) {
 		dev_info(ctrl->device, "rescanning namespaces.\n");
 		nvme_clear_changed_ns_log(ctrl);
@@ -4841,6 +4914,8 @@ static inline void _nvme_check_size(void)
 	BUILD_BUG_ON(sizeof(struct nvme_command) != 64);
 	BUILD_BUG_ON(sizeof(struct nvme_id_ctrl) != NVME_IDENTIFY_DATA_SIZE);
 	BUILD_BUG_ON(sizeof(struct nvme_id_ns) != NVME_IDENTIFY_DATA_SIZE);
+	BUILD_BUG_ON(sizeof(struct nvme_id_ns_cs_indep) !=
+			NVME_IDENTIFY_DATA_SIZE);
 	BUILD_BUG_ON(sizeof(struct nvme_id_ns_zns) != NVME_IDENTIFY_DATA_SIZE);
 	BUILD_BUG_ON(sizeof(struct nvme_id_ns_nvm) != NVME_IDENTIFY_DATA_SIZE);
 	BUILD_BUG_ON(sizeof(struct nvme_id_ctrl_zns) != NVME_IDENTIFY_DATA_SIZE);
diff --git a/include/linux/nvme.h b/include/linux/nvme.h
index 5f6d432fa06a..29ec3e3481ff 100644
--- a/include/linux/nvme.h
+++ b/include/linux/nvme.h
@@ -137,6 +137,7 @@ enum {
 	NVME_REG_CMBMSC = 0x0050,	/* Controller Memory Buffer Memory
 					 * Space Control
 					 */
+	NVME_REG_CRTO	= 0x0068,	/* Controller Ready Timeouts */
 	NVME_REG_PMRCAP	= 0x0e00,	/* Persistent Memory Capabilities */
 	NVME_REG_PMRCTL	= 0x0e04,	/* Persistent Memory Region Control */
 	NVME_REG_PMRSTS	= 0x0e08,	/* Persistent Memory Region Status */
@@ -161,6 +162,9 @@ enum {
 #define NVME_CMB_BIR(cmbloc)	((cmbloc) & 0x7)
 #define NVME_CMB_OFST(cmbloc)	(((cmbloc) >> 12) & 0xfffff)
 
+#define NVME_CRTO_CRIMT(crto)	((crto) >> 16)
+#define NVME_CRTO_CRWMT(crto)	((crto) & 0xffff)
+
 enum {
 	NVME_CMBSZ_SQS		= 1 << 0,
 	NVME_CMBSZ_CQS		= 1 << 1,
@@ -204,6 +208,7 @@ enum {
 	NVME_CC_SHN_MASK	= 3 << NVME_CC_SHN_SHIFT,
 	NVME_CC_IOSQES		= NVME_NVM_IOSQES << NVME_CC_IOSQES_SHIFT,
 	NVME_CC_IOCQES		= NVME_NVM_IOCQES << NVME_CC_IOCQES_SHIFT,
+	NVME_CC_CRIME		= 1 << 24,
 };
 
 enum {
@@ -227,6 +232,11 @@ enum {
 	NVME_CAP_CSS_CSI	= 1 << 6,
 };
 
+enum {
+	NVME_CAP_CRMS_CRIMS	= 1ULL << 59,
+	NVME_CAP_CRMS_CRWMS	= 1ULL << 60,
+};
+
 struct nvme_id_power_state {
 	__le16			max_power;	/* centiwatts */
 	__u8			rsvd2;
@@ -414,6 +424,21 @@ struct nvme_id_ns {
 	__u8			vs[3712];
 };
 
+/* I/O Command Set Independent Identify Namespace Data Structure */
+struct nvme_id_ns_cs_indep {
+	__u8			nsfeat;
+	__u8			nmic;
+	__u8			rescap;
+	__u8			fpi;
+	__le32			anagrpid;
+	__u8			nsattr;
+	__u8			rsvd9;
+	__le16			nvmsetid;
+	__le16			endgid;
+	__u8			nstat;
+	__u8			rsvd15[4081];
+};
+
 struct nvme_zns_lbafe {
 	__le64			zsze;
 	__u8			zdes;
@@ -478,6 +503,7 @@ enum {
 	NVME_ID_CNS_NS_DESC_LIST	= 0x03,
 	NVME_ID_CNS_CS_NS		= 0x05,
 	NVME_ID_CNS_CS_CTRL		= 0x06,
+	NVME_ID_CNS_NS_CS_INDEP		= 0x08,
 	NVME_ID_CNS_NS_PRESENT_LIST	= 0x10,
 	NVME_ID_CNS_NS_PRESENT		= 0x11,
 	NVME_ID_CNS_CTRL_NS_LIST	= 0x12,
@@ -532,6 +558,10 @@ enum {
 };
 
 enum {
+	NVME_NSTAT_NRDY		= 1 << 0,
+};
+
+enum {
 	NVME_NVM_NS_16B_GUARD	= 0,
 	NVME_NVM_NS_32B_GUARD	= 1,
 	NVME_NVM_NS_64B_GUARD	= 2,
@@ -1592,6 +1622,7 @@ enum {
 	NVME_SC_NS_WRITE_PROTECTED	= 0x20,
 	NVME_SC_CMD_INTERRUPTED		= 0x21,
 	NVME_SC_TRANSIENT_TR_ERR	= 0x22,
+	NVME_SC_ADMIN_COMMAND_MEDIA_NOT_READY = 0x24,
 	NVME_SC_INVALID_IO_CMD_SET	= 0x2C,
 
 	NVME_SC_LBA_RANGE		= 0x80,