nvme: wire-up uring-cmd support for io-passthru on char-device.

Introduce handler for fops->uring_cmd(), implementing async passthru
on char device (/dev/ngX). The handler supports newly introduced
operation NVME_URING_CMD_IO. This operates on a new structure
nvme_uring_cmd, which is similar to struct nvme_passthru_cmd64 but
without the embedded 8b result field. This field is not needed since
uring-cmd allows to return additional result via big-CQE.

Signed-off-by: Kanchan Joshi <joshi.k@samsung.com>
Signed-off-by: Anuj Gupta <anuj20.g@samsung.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20220511054750.20432-5-joshi.k@samsung.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>

5 files changed, 220 insertions, 3 deletions

diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index e1846d04817f..682df98db341 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -3699,6 +3699,7 @@ static const struct file_operations nvme_ns_chr_fops = {
 	.release	= nvme_ns_chr_release,
 	.unlocked_ioctl	= nvme_ns_chr_ioctl,
 	.compat_ioctl	= compat_ptr_ioctl,
+	.uring_cmd	= nvme_ns_chr_uring_cmd,
 };
 
 static int nvme_add_ns_cdev(struct nvme_ns *ns)
diff --git a/drivers/nvme/host/ioctl.c b/drivers/nvme/host/ioctl.c
index 8d2569b656cc..92d695262d8f 100644
--- a/drivers/nvme/host/ioctl.c
+++ b/drivers/nvme/host/ioctl.c
@@ -5,6 +5,7 @@
  */
 #include <linux/ptrace.h>	/* for force_successful_syscall_return */
 #include <linux/nvme_ioctl.h>
+#include <linux/io_uring.h>
 #include "nvme.h"
 
 /*
@@ -66,7 +67,8 @@ static int nvme_finish_user_metadata(struct request *req, void __user *ubuf,
 static struct request *nvme_alloc_user_request(struct request_queue *q,
 		struct nvme_command *cmd, void __user *ubuffer,
 		unsigned bufflen, void __user *meta_buffer, unsigned meta_len,
-		u32 meta_seed, void **metap, unsigned timeout, bool vec)
+		u32 meta_seed, void **metap, unsigned timeout, bool vec,
+		unsigned int rq_flags, blk_mq_req_flags_t blk_flags)
 {
 	bool write = nvme_is_write(cmd);
 	struct nvme_ns *ns = q->queuedata;
@@ -76,7 +78,7 @@ static struct request *nvme_alloc_user_request(struct request_queue *q,
 	void *meta = NULL;
 	int ret;
 
-	req = blk_mq_alloc_request(q, nvme_req_op(cmd), 0);
+	req = blk_mq_alloc_request(q, nvme_req_op(cmd) | rq_flags, blk_flags);
 	if (IS_ERR(req))
 		return req;
 	nvme_init_request(req, cmd);
@@ -140,7 +142,7 @@ static int nvme_submit_user_cmd(struct request_queue *q,
 	int ret;
 
 	req = nvme_alloc_user_request(q, cmd, ubuffer, bufflen, meta_buffer,
-			meta_len, meta_seed, &meta, timeout, vec);
+			meta_len, meta_seed, &meta, timeout, vec, 0, 0);
 	if (IS_ERR(req))
 		return PTR_ERR(req);
 
@@ -330,6 +332,139 @@ static int nvme_user_cmd64(struct nvme_ctrl *ctrl, struct nvme_ns *ns,
 	return status;
 }
 
+struct nvme_uring_data {
+	__u64	metadata;
+	__u64	addr;
+	__u32	data_len;
+	__u32	metadata_len;
+	__u32	timeout_ms;
+};
+
+/*
+ * This overlays struct io_uring_cmd pdu.
+ * Expect build errors if this grows larger than that.
+ */
+struct nvme_uring_cmd_pdu {
+	union {
+		struct bio *bio;
+		struct request *req;
+	};
+	void *meta; /* kernel-resident buffer */
+	void __user *meta_buffer;
+	u32 meta_len;
+};
+
+static inline struct nvme_uring_cmd_pdu *nvme_uring_cmd_pdu(
+		struct io_uring_cmd *ioucmd)
+{
+	return (struct nvme_uring_cmd_pdu *)&ioucmd->pdu;
+}
+
+static void nvme_uring_task_cb(struct io_uring_cmd *ioucmd)
+{
+	struct nvme_uring_cmd_pdu *pdu = nvme_uring_cmd_pdu(ioucmd);
+	struct request *req = pdu->req;
+	struct bio *bio = req->bio;
+	int status;
+	u64 result;
+
+	if (nvme_req(req)->flags & NVME_REQ_CANCELLED)
+		status = -EINTR;
+	else
+		status = nvme_req(req)->status;
+
+	result = le64_to_cpu(nvme_req(req)->result.u64);
+
+	if (pdu->meta)
+		status = nvme_finish_user_metadata(req, pdu->meta_buffer,
+					pdu->meta, pdu->meta_len, status);
+	if (bio)
+		blk_rq_unmap_user(bio);
+	blk_mq_free_request(req);
+
+	io_uring_cmd_done(ioucmd, status, result);
+}
+
+static void nvme_uring_cmd_end_io(struct request *req, blk_status_t err)
+{
+	struct io_uring_cmd *ioucmd = req->end_io_data;
+	struct nvme_uring_cmd_pdu *pdu = nvme_uring_cmd_pdu(ioucmd);
+	/* extract bio before reusing the same field for request */
+	struct bio *bio = pdu->bio;
+
+	pdu->req = req;
+	req->bio = bio;
+	/* this takes care of moving rest of completion-work to task context */
+	io_uring_cmd_complete_in_task(ioucmd, nvme_uring_task_cb);
+}
+
+static int nvme_uring_cmd_io(struct nvme_ctrl *ctrl, struct nvme_ns *ns,
+		struct io_uring_cmd *ioucmd, unsigned int issue_flags)
+{
+	struct nvme_uring_cmd_pdu *pdu = nvme_uring_cmd_pdu(ioucmd);
+	const struct nvme_uring_cmd *cmd = ioucmd->cmd;
+	struct request_queue *q = ns ? ns->queue : ctrl->admin_q;
+	struct nvme_uring_data d;
+	struct nvme_command c;
+	struct request *req;
+	unsigned int rq_flags = 0;
+	blk_mq_req_flags_t blk_flags = 0;
+	void *meta = NULL;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EACCES;
+
+	c.common.opcode = READ_ONCE(cmd->opcode);
+	c.common.flags = READ_ONCE(cmd->flags);
+	if (c.common.flags)
+		return -EINVAL;
+
+	c.common.command_id = 0;
+	c.common.nsid = cpu_to_le32(cmd->nsid);
+	if (!nvme_validate_passthru_nsid(ctrl, ns, le32_to_cpu(c.common.nsid)))
+		return -EINVAL;
+
+	c.common.cdw2[0] = cpu_to_le32(READ_ONCE(cmd->cdw2));
+	c.common.cdw2[1] = cpu_to_le32(READ_ONCE(cmd->cdw3));
+	c.common.metadata = 0;
+	c.common.dptr.prp1 = c.common.dptr.prp2 = 0;
+	c.common.cdw10 = cpu_to_le32(READ_ONCE(cmd->cdw10));
+	c.common.cdw11 = cpu_to_le32(READ_ONCE(cmd->cdw11));
+	c.common.cdw12 = cpu_to_le32(READ_ONCE(cmd->cdw12));
+	c.common.cdw13 = cpu_to_le32(READ_ONCE(cmd->cdw13));
+	c.common.cdw14 = cpu_to_le32(READ_ONCE(cmd->cdw14));
+	c.common.cdw15 = cpu_to_le32(READ_ONCE(cmd->cdw15));
+
+	d.metadata = READ_ONCE(cmd->metadata);
+	d.addr = READ_ONCE(cmd->addr);
+	d.data_len = READ_ONCE(cmd->data_len);
+	d.metadata_len = READ_ONCE(cmd->metadata_len);
+	d.timeout_ms = READ_ONCE(cmd->timeout_ms);
+
+	if (issue_flags & IO_URING_F_NONBLOCK) {
+		rq_flags = REQ_NOWAIT;
+		blk_flags = BLK_MQ_REQ_NOWAIT;
+	}
+
+	req = nvme_alloc_user_request(q, &c, nvme_to_user_ptr(d.addr),
+			d.data_len, nvme_to_user_ptr(d.metadata),
+			d.metadata_len, 0, &meta, d.timeout_ms ?
+			msecs_to_jiffies(d.timeout_ms) : 0, 0, rq_flags,
+			blk_flags);
+	if (IS_ERR(req))
+		return PTR_ERR(req);
+	req->end_io_data = ioucmd;
+
+	/* to free bio on completion, as req->bio will be null at that time */
+	pdu->bio = req->bio;
+	pdu->meta = meta;
+	pdu->meta_buffer = nvme_to_user_ptr(d.metadata);
+	pdu->meta_len = d.metadata_len;
+
+	blk_execute_rq_nowait(req, 0, nvme_uring_cmd_end_io);
+	return -EIOCBQUEUED;
+}
+
 static bool is_ctrl_ioctl(unsigned int cmd)
 {
 	if (cmd == NVME_IOCTL_ADMIN_CMD || cmd == NVME_IOCTL_ADMIN64_CMD)
@@ -421,6 +556,42 @@ long nvme_ns_chr_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
 	return __nvme_ioctl(ns, cmd, (void __user *)arg);
 }
 
+static int nvme_ns_uring_cmd(struct nvme_ns *ns, struct io_uring_cmd *ioucmd,
+			     unsigned int issue_flags)
+{
+	struct nvme_ctrl *ctrl = ns->ctrl;
+	int ret;
+
+	BUILD_BUG_ON(sizeof(struct nvme_uring_cmd_pdu) > sizeof(ioucmd->pdu));
+
+	/* IOPOLL not supported yet */
+	if (issue_flags & IO_URING_F_IOPOLL)
+		return -EOPNOTSUPP;
+
+	/* NVMe passthrough requires bit SQE/CQE support */
+	if ((issue_flags & (IO_URING_F_SQE128|IO_URING_F_CQE32)) !=
+	    (IO_URING_F_SQE128|IO_URING_F_CQE32))
+		return -EOPNOTSUPP;
+
+	switch (ioucmd->cmd_op) {
+	case NVME_URING_CMD_IO:
+		ret = nvme_uring_cmd_io(ctrl, ns, ioucmd, issue_flags);
+		break;
+	default:
+		ret = -ENOTTY;
+	}
+
+	return ret;
+}
+
+int nvme_ns_chr_uring_cmd(struct io_uring_cmd *ioucmd, unsigned int issue_flags)
+{
+	struct nvme_ns *ns = container_of(file_inode(ioucmd->file)->i_cdev,
+			struct nvme_ns, cdev);
+
+	return nvme_ns_uring_cmd(ns, ioucmd, issue_flags);
+}
+
 #ifdef CONFIG_NVME_MULTIPATH
 static int nvme_ns_head_ctrl_ioctl(struct nvme_ns *ns, unsigned int cmd,
 		void __user *argp, struct nvme_ns_head *head, int srcu_idx)
@@ -487,6 +658,21 @@ out_unlock:
 	srcu_read_unlock(&head->srcu, srcu_idx);
 	return ret;
 }
+
+int nvme_ns_head_chr_uring_cmd(struct io_uring_cmd *ioucmd,
+		unsigned int issue_flags)
+{
+	struct cdev *cdev = file_inode(ioucmd->file)->i_cdev;
+	struct nvme_ns_head *head = container_of(cdev, struct nvme_ns_head, cdev);
+	int srcu_idx = srcu_read_lock(&head->srcu);
+	struct nvme_ns *ns = nvme_find_path(head);
+	int ret = -EINVAL;
+
+	if (ns)
+		ret = nvme_ns_uring_cmd(ns, ioucmd, issue_flags);
+	srcu_read_unlock(&head->srcu, srcu_idx);
+	return ret;
+}
 #endif /* CONFIG_NVME_MULTIPATH */
 
 static int nvme_dev_user_cmd(struct nvme_ctrl *ctrl, void __user *argp)
diff --git a/drivers/nvme/host/multipath.c b/drivers/nvme/host/multipath.c
index d464fdf978fb..d3e2440d8abb 100644
--- a/drivers/nvme/host/multipath.c
+++ b/drivers/nvme/host/multipath.c
@@ -437,6 +437,7 @@ static const struct file_operations nvme_ns_head_chr_fops = {
 	.release	= nvme_ns_head_chr_release,
 	.unlocked_ioctl	= nvme_ns_head_chr_ioctl,
 	.compat_ioctl	= compat_ptr_ioctl,
+	.uring_cmd	= nvme_ns_head_chr_uring_cmd,
 };
 
 static int nvme_add_ns_head_cdev(struct nvme_ns_head *head)
diff --git a/drivers/nvme/host/nvme.h b/drivers/nvme/host/nvme.h
index a2b53ca63335..086ccbdd7003 100644
--- a/drivers/nvme/host/nvme.h
+++ b/drivers/nvme/host/nvme.h
@@ -782,6 +782,10 @@ long nvme_ns_head_chr_ioctl(struct file *file, unsigned int cmd,
 		unsigned long arg);
 long nvme_dev_ioctl(struct file *file, unsigned int cmd,
 		unsigned long arg);
+int nvme_ns_chr_uring_cmd(struct io_uring_cmd *ioucmd,
+		unsigned int issue_flags);
+int nvme_ns_head_chr_uring_cmd(struct io_uring_cmd *ioucmd,
+		unsigned int issue_flags);
 int nvme_getgeo(struct block_device *bdev, struct hd_geometry *geo);
 
 extern const struct attribute_group *nvme_ns_id_attr_groups[];
diff --git a/include/uapi/linux/nvme_ioctl.h b/include/uapi/linux/nvme_ioctl.h
index b2e43185e3b5..04e458c649ab 100644
--- a/include/uapi/linux/nvme_ioctl.h
+++ b/include/uapi/linux/nvme_ioctl.h
@@ -70,6 +70,28 @@ struct nvme_passthru_cmd64 {
 	__u64	result;
 };
 
+/* same as struct nvme_passthru_cmd64, minus the 8b result field */
+struct nvme_uring_cmd {
+	__u8	opcode;
+	__u8	flags;
+	__u16	rsvd1;
+	__u32	nsid;
+	__u32	cdw2;
+	__u32	cdw3;
+	__u64	metadata;
+	__u64	addr;
+	__u32	metadata_len;
+	__u32	data_len;
+	__u32	cdw10;
+	__u32	cdw11;
+	__u32	cdw12;
+	__u32	cdw13;
+	__u32	cdw14;
+	__u32	cdw15;
+	__u32	timeout_ms;
+	__u32   rsvd2;
+};
+
 #define nvme_admin_cmd nvme_passthru_cmd
 
 #define NVME_IOCTL_ID		_IO('N', 0x40)
@@ -83,4 +105,7 @@ struct nvme_passthru_cmd64 {
 #define NVME_IOCTL_IO64_CMD	_IOWR('N', 0x48, struct nvme_passthru_cmd64)
 #define NVME_IOCTL_IO64_CMD_VEC	_IOWR('N', 0x49, struct nvme_passthru_cmd64)
 
+/* io_uring async commands: */
+#define NVME_URING_CMD_IO	_IOWR('N', 0x80, struct nvme_uring_cmd)
+
 #endif /* _UAPI_LINUX_NVME_IOCTL_H */