From ee692a21e9bf8354bd3ec816f1cf4bff8619ed77 Mon Sep 17 00:00:00 2001
From: Jens Axboe <axboe@kernel.dk>
Date: Wed, 11 May 2022 11:17:45 +0530
Subject: [PATCH] fs,io_uring: add infrastructure for uring-cmd
Git-commit: ee692a21e9bf8354bd3ec816f1cf4bff8619ed77
Patch-mainline: v5.19-rc1
References: bsc#1205205

file_operations->uring_cmd is a file private handler.
This is somewhat similar to ioctl but hopefully a lot more sane and
useful as it can be used to enable many io_uring capabilities for the
underlying operation.

IORING_OP_URING_CMD is a file private kind of request. io_uring doesn't
know what is in this command type, it's for the provider of ->uring_cmd()
to deal with.

Co-developed-by: Kanchan Joshi <joshi.k@samsung.com>
Signed-off-by: Kanchan Joshi <joshi.k@samsung.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20220511054750.20432-2-joshi.k@samsung.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Gabriel Krisman Bertazi <krisman@suse.de>
---
 fs/io_uring.c                 |  123 +++++++++++++++++++++++++++++++++++++++---
 include/linux/fs.h            |    2 
 include/linux/io_uring.h      |   33 +++++++++++
 include/uapi/linux/io_uring.h |   19 ++++--
 4 files changed, 164 insertions(+), 13 deletions(-)

--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -200,13 +200,6 @@ struct io_rings {
 	struct io_uring_cqe	cqes[] ____cacheline_aligned_in_smp;
 };
 
-enum io_uring_cmd_flags {
-	IO_URING_F_COMPLETE_DEFER	= 1,
-	IO_URING_F_UNLOCKED		= 2,
-	/* int's last bit, sign checks are usually faster than a bit test */
-	IO_URING_F_NONBLOCK		= INT_MIN,
-};
-
 struct io_mapped_ubuf {
 	u64		ubuf;
 	u64		ubuf_end;
@@ -953,6 +946,7 @@ struct io_kiocb {
 		struct io_hardlink	hardlink;
 		struct io_msg		msg;
 		struct io_nop           nop;
+		struct io_uring_cmd     uring_cmd;
 	};
 
 	u8				opcode;
@@ -1031,6 +1025,14 @@ struct io_cancel_data {
 	int seq;
 };
 
+/*
+ * The URING_CMD payload starts at 'cmd' in the first sqe, and continues into
+ * the following sqe if SQE128 is used.
+ */
+#define uring_cmd_pdu_size(is_sqe128)				\
+	((1 + !!(is_sqe128)) * sizeof(struct io_uring_sqe) -	\
+		offsetof(struct io_uring_sqe, cmd))
+
 struct io_op_def {
 	/* needs req->file assigned */
 	unsigned		needs_file : 1;
@@ -1229,6 +1231,12 @@ static const struct io_op_def io_op_defs
 		.needs_file		= 1,
 		.iopoll			= 1,
 	},
+	[IORING_OP_URING_CMD] = {
+		.needs_file		= 1,
+		.plug			= 1,
+		.needs_async_setup	= 1,
+		.async_size		= uring_cmd_pdu_size(1),
+	},
 };
 
 /* requests with any of those set should undergo io_disarm_next() */
@@ -1358,6 +1366,8 @@ const char *io_uring_get_opcode(u8 opcod
 		return "LINKAT";
 	case IORING_OP_MSG_RING:
 		return "MSG_RING";
+	case IORING_OP_URING_CMD:
+		return "URING_CMD";
 	case IORING_OP_LAST:
 		return "INVALID";
 	}
@@ -4588,6 +4598,96 @@ static int io_linkat(struct io_kiocb *re
 	return 0;
 }
 
+static void io_uring_cmd_work(struct io_kiocb *req, bool *locked)
+{
+	req->uring_cmd.task_work_cb(&req->uring_cmd);
+}
+
+void io_uring_cmd_complete_in_task(struct io_uring_cmd *ioucmd,
+			void (*task_work_cb)(struct io_uring_cmd *))
+{
+	struct io_kiocb *req = container_of(ioucmd, struct io_kiocb, uring_cmd);
+
+	req->uring_cmd.task_work_cb = task_work_cb;
+	req->io_task_work.func = io_uring_cmd_work;
+	io_req_task_work_add(req, !!(req->ctx->flags & IORING_SETUP_SQPOLL));
+}
+EXPORT_SYMBOL_GPL(io_uring_cmd_complete_in_task);
+
+/*
+ * Called by consumers of io_uring_cmd, if they originally returned
+ * -EIOCBQUEUED upon receiving the command.
+ */
+void io_uring_cmd_done(struct io_uring_cmd *ioucmd, ssize_t ret, ssize_t res2)
+{
+	struct io_kiocb *req = container_of(ioucmd, struct io_kiocb, uring_cmd);
+
+	if (ret < 0)
+		req_set_fail(req);
+	if (req->ctx->flags & IORING_SETUP_CQE32)
+		__io_req_complete32(req, 0, ret, 0, res2, 0);
+	else
+		io_req_complete(req, ret);
+}
+EXPORT_SYMBOL_GPL(io_uring_cmd_done);
+
+static int io_uring_cmd_prep_async(struct io_kiocb *req)
+{
+	size_t cmd_size;
+
+	cmd_size = uring_cmd_pdu_size(req->ctx->flags & IORING_SETUP_SQE128);
+
+	memcpy(req->async_data, req->uring_cmd.cmd, cmd_size);
+	return 0;
+}
+
+static int io_uring_cmd_prep(struct io_kiocb *req,
+			     const struct io_uring_sqe *sqe)
+{
+	struct io_uring_cmd *ioucmd = &req->uring_cmd;
+
+	if (sqe->rw_flags)
+		return -EINVAL;
+	ioucmd->cmd = sqe->cmd;
+	ioucmd->cmd_op = READ_ONCE(sqe->cmd_op);
+	return 0;
+}
+
+static int io_uring_cmd(struct io_kiocb *req, unsigned int issue_flags)
+{
+	struct io_uring_cmd *ioucmd = &req->uring_cmd;
+	struct io_ring_ctx *ctx = req->ctx;
+	struct file *file = req->file;
+	int ret;
+
+	if (!req->file->f_op->uring_cmd)
+		return -EOPNOTSUPP;
+
+	if (ctx->flags & IORING_SETUP_SQE128)
+		issue_flags |= IO_URING_F_SQE128;
+	if (ctx->flags & IORING_SETUP_CQE32)
+		issue_flags |= IO_URING_F_CQE32;
+	if (ctx->flags & IORING_SETUP_IOPOLL)
+		issue_flags |= IO_URING_F_IOPOLL;
+
+	if (req_has_async_data(req))
+		ioucmd->cmd = req->async_data;
+
+	ret = file->f_op->uring_cmd(ioucmd, issue_flags);
+	if (ret == -EAGAIN) {
+		if (!req_has_async_data(req)) {
+			if (io_alloc_async_data(req))
+				return -ENOMEM;
+			io_uring_cmd_prep_async(req);
+		}
+		return -EAGAIN;
+	}
+
+	if (ret != -EIOCBQUEUED)
+		io_uring_cmd_done(ioucmd, ret, 0);
+	return 0;
+}
+
 static int io_shutdown_prep(struct io_kiocb *req,
 			    const struct io_uring_sqe *sqe)
 {
@@ -7412,6 +7512,8 @@ static int io_req_prep(struct io_kiocb *
 		return io_linkat_prep(req, sqe);
 	case IORING_OP_MSG_RING:
 		return io_msg_ring_prep(req, sqe);
+	case IORING_OP_URING_CMD:
+		return io_uring_cmd_prep(req, sqe);
 	}
 
 	printk_once(KERN_WARNING "io_uring: unhandled opcode %d\n",
@@ -7444,6 +7546,8 @@ static int io_req_prep_async(struct io_k
 		return io_recvmsg_prep_async(req);
 	case IORING_OP_CONNECT:
 		return io_connect_prep_async(req);
+	case IORING_OP_URING_CMD:
+		return io_uring_cmd_prep_async(req);
 	}
 	printk_once(KERN_WARNING "io_uring: prep_async() bad opcode %d\n",
 		    req->opcode);
@@ -7714,6 +7818,9 @@ static int io_issue_sqe(struct io_kiocb
 	case IORING_OP_MSG_RING:
 		ret = io_msg_ring(req, issue_flags);
 		break;
+	case IORING_OP_URING_CMD:
+		ret = io_uring_cmd(req, issue_flags);
+		break;
 	default:
 		ret = -EINVAL;
 		break;
@@ -12326,6 +12433,8 @@ static int __init io_uring_init(void)
 
 	BUILD_BUG_ON(sizeof(atomic_t) != sizeof(u32));
 
+	BUILD_BUG_ON(sizeof(struct io_uring_cmd) > 64);
+
 	req_cachep = KMEM_CACHE(io_kiocb, SLAB_HWCACHE_ALIGN | SLAB_PANIC |
 				SLAB_ACCOUNT);
 	return 0;
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -2058,6 +2058,7 @@ struct dir_context {
 #define REMAP_FILE_ADVISORY		(REMAP_FILE_CAN_SHORTEN)
 
 struct iov_iter;
+struct io_uring_cmd;
 
 struct file_operations {
 	struct module *owner;
@@ -2100,6 +2101,7 @@ struct file_operations {
 				   struct file *file_out, loff_t pos_out,
 				   loff_t len, unsigned int remap_flags);
 	int (*fadvise)(struct file *, loff_t, loff_t, int);
+	int (*uring_cmd)(struct io_uring_cmd *ioucmd, unsigned int issue_flags);
 } __randomize_layout;
 
 struct inode_operations {
--- a/include/linux/io_uring.h
+++ b/include/linux/io_uring.h
@@ -5,7 +5,32 @@
 #include <linux/sched.h>
 #include <linux/xarray.h>
 
+enum io_uring_cmd_flags {
+	IO_URING_F_COMPLETE_DEFER	= 1,
+	IO_URING_F_UNLOCKED		= 2,
+	/* int's last bit, sign checks are usually faster than a bit test */
+	IO_URING_F_NONBLOCK		= INT_MIN,
+
+	/* ctx state flags, for URING_CMD */
+	IO_URING_F_SQE128		= 4,
+	IO_URING_F_CQE32		= 8,
+	IO_URING_F_IOPOLL		= 16,
+};
+
+struct io_uring_cmd {
+	struct file	*file;
+	const void	*cmd;
+	/* callback to defer completions to task context */
+	void (*task_work_cb)(struct io_uring_cmd *cmd);
+	u32		cmd_op;
+	u32		pad;
+	u8		pdu[32]; /* available inline for free use */
+};
+
 #if defined(CONFIG_IO_URING)
+void io_uring_cmd_done(struct io_uring_cmd *cmd, ssize_t ret, ssize_t res2);
+void io_uring_cmd_complete_in_task(struct io_uring_cmd *ioucmd,
+			void (*task_work_cb)(struct io_uring_cmd *));
 struct sock *io_uring_get_socket(struct file *file);
 void __io_uring_cancel(bool cancel_all);
 void __io_uring_free(struct task_struct *tsk);
@@ -30,6 +55,14 @@ static inline void io_uring_free(struct
 		__io_uring_free(tsk);
 }
 #else
+static inline void io_uring_cmd_done(struct io_uring_cmd *cmd, ssize_t ret,
+		ssize_t ret2)
+{
+}
+static inline void io_uring_cmd_complete_in_task(struct io_uring_cmd *ioucmd,
+			void (*task_work_cb)(struct io_uring_cmd *))
+{
+}
 static inline struct sock *io_uring_get_socket(struct file *file)
 {
 	return NULL;
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -22,6 +22,7 @@ struct io_uring_sqe {
 	union {
 		__u64	off;	/* offset into file */
 		__u64	addr2;
+		__u32	cmd_op;
 	};
 	union {
 		__u64	addr;	/* pointer to buffer or iovecs */
@@ -60,13 +61,18 @@ struct io_uring_sqe {
 		__s32	splice_fd_in;
 		__u32	file_index;
 	};
-	__u64	__pad2[2];
 
-	/*
-	 * If the ring is initialized with IORING_SETUP_SQE128, then this field
-	 * contains 64-bytes of padding, doubling the size of the SQE.
-	 */
-	__u64   __big_sqe_pad[0];
+	 union {
+		 struct {
+			 __u64   addr3;
+			 __u64   __pad2[1];
+		 };
+		 /*
+		  * If the ring is initialized with IORING_SETUP_SQE128, then
+		  * this field is used for 80 bytes of arbitrary command data
+		  */
+		 __u8    cmd[0];
+	 };
 };
 
 /*
@@ -163,6 +169,7 @@ enum io_uring_op {
 	IORING_OP_SYMLINKAT,
 	IORING_OP_LINKAT,
 	IORING_OP_MSG_RING,
+	IORING_OP_URING_CMD,
 
 	/* this goes last, obviously */
 	IORING_OP_LAST,