Linux-Block Archive on lore.kernel.org
 help / Atom feed
* [PATCHSET v1] io_uring IO interface
@ 2019-01-08 16:56 Jens Axboe
  2019-01-08 16:56 ` [PATCH 01/16] fs: add an iopoll method to struct file_operations Jens Axboe
                   ` (16 more replies)
  0 siblings, 17 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch; +Cc: hch, jmoyer, avi

After some arm twisting from Christoph, I finally caved and divorced the
aio-poll patches from aio/libaio itself. The io_uring interface itself
is useful and efficient, and after rebasing all the new goodies on top
of that, there was little reason to retail the aio connection.

Hence io_uring was born. This is what I previously called scqring for
aio, but now as a standalone entity. Patch #5 adds the core of this
interface, but in short, it has two main data structures:

struct io_uring_iocb {
	__u8	opcode;
	__u8	flags;
	__u16	ioprio;
	__s32	fd;
	__u64	off;
	union {
		void	*addr;
		__u64	__pad;
	};
	__u32	len;
	union {
		__kernel_rwf_t	rw_flags;
		__u32		__resv;
	};
};

struct io_uring_event {
	__u64	index;		/* what iocb this event came from */
	__s32	res;		/* result code for this event */
	__u32	flags;
};

The SQ ring is an array of indexes into an array of io_uring_iocbs,
which describe the IO to be done. The SQ ring is an array of
io_uring_events, which describe a completion event. Both of these rings
are mapped into the application through mmap(2), at special magic
offsets. The application manipulates the ring directly, and then
communicates with the kernel through these two system calls:

io_uring_setup(entries, iovecs, params)
	Sets up a context for doing async IO. On success, returns a file
	descriptor that the application can mmap to gain access to the
	SQ ring, CQ ring, and io_uring_iocbs.

io_uring_enter(fd, to_submit, min_complete, flags)
	Initiates IO against the rings mapped to this fd, or waits for
	them to complete, or both The behavior is controlled by the
	parameters passed in. If 'min_complete' is non-zero, then we'll
	try and submit new IO. If IORING_ENTER_GETEVENTS is set, the
	kernel will wait for 'min_complete' events, if they aren't
	already available.

I've started a liburing git repo for this, which contains some helpers
for doing IO without having to muck with the ring directly, setting up
an io_uring context, etc. Clone that here:

git://git.kernel.dk/liburing

In terms of usage, there's also a small test app here:

http://git.kernel.dk/cgit/fio/plain/t/io_uring.c

and if you're into fio, there's a io_uring engine included with that as
well for test purposes.

In terms of features, this has everything that the prior aio-poll
postings did. Later patches add support for polled IO, fixed buffers,
kernel side submission and polling, buffered aio, etc. Also a number of
bug fixes in here from previous postings.

Series is against 5.0-rc1, and can also be found in my io_uring branch.
For now just x86-64 has the system calls wired up, and liburing also
only supports x86-64. The latter just needs system call numbers and
reasonable read/write barrier defines to work, however.

 Documentation/filesystems/vfs.txt      |    3 +
 arch/x86/entry/syscalls/syscall_64.tbl |    2 +
 block/bio.c                            |   59 +-
 fs/Makefile                            |    2 +-
 fs/block_dev.c                         |   19 +-
 fs/file.c                              |   15 +-
 fs/file_table.c                        |    9 +-
 fs/gfs2/file.c                         |    2 +
 fs/io_uring.c                          | 1907 ++++++++++++++++++++++++
 fs/iomap.c                             |   48 +-
 fs/xfs/xfs_file.c                      |    1 +
 include/linux/bio.h                    |   14 +
 include/linux/blk_types.h              |    1 +
 include/linux/file.h                   |    2 +
 include/linux/fs.h                     |    6 +-
 include/linux/iomap.h                  |    1 +
 include/linux/syscalls.h               |    5 +
 include/uapi/linux/io_uring.h          |  115 ++
 kernel/sys_ni.c                        |    2 +
 19 files changed, 2173 insertions(+), 40 deletions(-)

-- 
Jens Axboe



^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 01/16] fs: add an iopoll method to struct file_operations
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
@ 2019-01-08 16:56 ` Jens Axboe
  2019-01-08 16:56 ` [PATCH 02/16] block: wire up block device iopoll method Jens Axboe
                   ` (15 subsequent siblings)
  16 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

From: Christoph Hellwig <hch@lst.de>

This new methods is used to explicitly poll for I/O completion for an
iocb.  It must be called for any iocb submitted asynchronously (that
is with a non-null ki_complete) which has the IOCB_HIPRI flag set.

The method is assisted by a new ki_cookie field in struct iocb to store
the polling cookie.

TODO: we can probably union ki_cookie with the existing hint and I/O
priority fields to avoid struct kiocb growth.

Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 Documentation/filesystems/vfs.txt | 3 +++
 include/linux/fs.h                | 2 ++
 2 files changed, 5 insertions(+)

diff --git a/Documentation/filesystems/vfs.txt b/Documentation/filesystems/vfs.txt
index 8dc8e9c2913f..761c6fd24a53 100644
--- a/Documentation/filesystems/vfs.txt
+++ b/Documentation/filesystems/vfs.txt
@@ -857,6 +857,7 @@ struct file_operations {
 	ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
 	ssize_t (*read_iter) (struct kiocb *, struct iov_iter *);
 	ssize_t (*write_iter) (struct kiocb *, struct iov_iter *);
+	int (*iopoll)(struct kiocb *kiocb, bool spin);
 	int (*iterate) (struct file *, struct dir_context *);
 	int (*iterate_shared) (struct file *, struct dir_context *);
 	__poll_t (*poll) (struct file *, struct poll_table_struct *);
@@ -902,6 +903,8 @@ otherwise noted.
 
   write_iter: possibly asynchronous write with iov_iter as source
 
+  iopoll: called when aio wants to poll for completions on HIPRI iocbs
+
   iterate: called when the VFS needs to read the directory contents
 
   iterate_shared: called when the VFS needs to read the directory contents
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 811c77743dad..ccb0b7a63aa5 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -310,6 +310,7 @@ struct kiocb {
 	int			ki_flags;
 	u16			ki_hint;
 	u16			ki_ioprio; /* See linux/ioprio.h */
+	unsigned int		ki_cookie; /* for ->iopoll */
 } __randomize_layout;
 
 static inline bool is_sync_kiocb(struct kiocb *kiocb)
@@ -1786,6 +1787,7 @@ struct file_operations {
 	ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
 	ssize_t (*read_iter) (struct kiocb *, struct iov_iter *);
 	ssize_t (*write_iter) (struct kiocb *, struct iov_iter *);
+	int (*iopoll)(struct kiocb *kiocb, bool spin);
 	int (*iterate) (struct file *, struct dir_context *);
 	int (*iterate_shared) (struct file *, struct dir_context *);
 	__poll_t (*poll) (struct file *, struct poll_table_struct *);
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 02/16] block: wire up block device iopoll method
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
  2019-01-08 16:56 ` [PATCH 01/16] fs: add an iopoll method to struct file_operations Jens Axboe
@ 2019-01-08 16:56 ` Jens Axboe
  2019-01-08 16:56 ` [PATCH 03/16] block: add bio_set_polled() helper Jens Axboe
                   ` (14 subsequent siblings)
  16 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

From: Christoph Hellwig <hch@lst.de>

Just call blk_poll on the iocb cookie, we can derive the block device
from the inode trivially.

Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/block_dev.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/fs/block_dev.c b/fs/block_dev.c
index c546cdce77e6..5415579f3e14 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -279,6 +279,14 @@ struct blkdev_dio {
 
 static struct bio_set blkdev_dio_pool;
 
+static int blkdev_iopoll(struct kiocb *kiocb, bool wait)
+{
+	struct block_device *bdev = I_BDEV(kiocb->ki_filp->f_mapping->host);
+	struct request_queue *q = bdev_get_queue(bdev);
+
+	return blk_poll(q, READ_ONCE(kiocb->ki_cookie), wait);
+}
+
 static void blkdev_bio_end_io(struct bio *bio)
 {
 	struct blkdev_dio *dio = bio->bi_private;
@@ -396,6 +404,7 @@ __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages)
 				bio->bi_opf |= REQ_HIPRI;
 
 			qc = submit_bio(bio);
+			WRITE_ONCE(iocb->ki_cookie, qc);
 			break;
 		}
 
@@ -2068,6 +2077,7 @@ const struct file_operations def_blk_fops = {
 	.llseek		= block_llseek,
 	.read_iter	= blkdev_read_iter,
 	.write_iter	= blkdev_write_iter,
+	.iopoll		= blkdev_iopoll,
 	.mmap		= generic_file_mmap,
 	.fsync		= blkdev_fsync,
 	.unlocked_ioctl	= block_ioctl,
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 03/16] block: add bio_set_polled() helper
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
  2019-01-08 16:56 ` [PATCH 01/16] fs: add an iopoll method to struct file_operations Jens Axboe
  2019-01-08 16:56 ` [PATCH 02/16] block: wire up block device iopoll method Jens Axboe
@ 2019-01-08 16:56 ` Jens Axboe
  2019-01-10  9:43   ` Ming Lei
  2019-01-08 16:56 ` [PATCH 04/16] iomap: wire up the iopoll method Jens Axboe
                   ` (13 subsequent siblings)
  16 siblings, 1 reply; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

For the upcoming async polled IO, we can't sleep allocating requests.
If we do, then we introduce a deadlock where the submitter already
has async polled IO in-flight, but can't wait for them to complete
since polled requests must be active found and reaped.

Utilize the helper in the blockdev DIRECT_IO code.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/block_dev.c      |  4 ++--
 include/linux/bio.h | 14 ++++++++++++++
 2 files changed, 16 insertions(+), 2 deletions(-)

diff --git a/fs/block_dev.c b/fs/block_dev.c
index 5415579f3e14..2ebd2a0d7789 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -233,7 +233,7 @@ __blkdev_direct_IO_simple(struct kiocb *iocb, struct iov_iter *iter,
 		task_io_account_write(ret);
 	}
 	if (iocb->ki_flags & IOCB_HIPRI)
-		bio.bi_opf |= REQ_HIPRI;
+		bio_set_polled(&bio, iocb);
 
 	qc = submit_bio(&bio);
 	for (;;) {
@@ -401,7 +401,7 @@ __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages)
 		nr_pages = iov_iter_npages(iter, BIO_MAX_PAGES);
 		if (!nr_pages) {
 			if (iocb->ki_flags & IOCB_HIPRI)
-				bio->bi_opf |= REQ_HIPRI;
+				bio_set_polled(bio, iocb);
 
 			qc = submit_bio(bio);
 			WRITE_ONCE(iocb->ki_cookie, qc);
diff --git a/include/linux/bio.h b/include/linux/bio.h
index 7380b094dcca..f6f0a2b3cbc8 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -823,5 +823,19 @@ static inline int bio_integrity_add_page(struct bio *bio, struct page *page,
 
 #endif /* CONFIG_BLK_DEV_INTEGRITY */
 
+/*
+ * Mark a bio as polled. Note that for async polled IO, the caller must
+ * expect -EWOULDBLOCK if we cannot allocate a request (or other resources).
+ * We cannot block waiting for requests on polled IO, as those completions
+ * must be found by the caller. This is different than IRQ driven IO, where
+ * it's safe to wait for IO to complete.
+ */
+static inline void bio_set_polled(struct bio *bio, struct kiocb *kiocb)
+{
+	bio->bi_opf |= REQ_HIPRI;
+	if (!is_sync_kiocb(kiocb))
+		bio->bi_opf |= REQ_NOWAIT;
+}
+
 #endif /* CONFIG_BLOCK */
 #endif /* __LINUX_BIO_H */
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 04/16] iomap: wire up the iopoll method
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
                   ` (2 preceding siblings ...)
  2019-01-08 16:56 ` [PATCH 03/16] block: add bio_set_polled() helper Jens Axboe
@ 2019-01-08 16:56 ` Jens Axboe
  2019-01-08 16:56 ` [PATCH 05/16] Add io_uring IO interface Jens Axboe
                   ` (12 subsequent siblings)
  16 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

From: Christoph Hellwig <hch@lst.de>

Store the request queue the last bio was submitted to in the iocb
private data in addition to the cookie so that we find the right block
device.  Also refactor the common direct I/O bio submission code into a
nice little helper.

Signed-off-by: Christoph Hellwig <hch@lst.de>

Modified to use bio_set_polled().

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/gfs2/file.c        |  2 ++
 fs/iomap.c            | 43 ++++++++++++++++++++++++++++---------------
 fs/xfs/xfs_file.c     |  1 +
 include/linux/iomap.h |  1 +
 4 files changed, 32 insertions(+), 15 deletions(-)

diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c
index a2dea5bc0427..58a768e59712 100644
--- a/fs/gfs2/file.c
+++ b/fs/gfs2/file.c
@@ -1280,6 +1280,7 @@ const struct file_operations gfs2_file_fops = {
 	.llseek		= gfs2_llseek,
 	.read_iter	= gfs2_file_read_iter,
 	.write_iter	= gfs2_file_write_iter,
+	.iopoll		= iomap_dio_iopoll,
 	.unlocked_ioctl	= gfs2_ioctl,
 	.mmap		= gfs2_mmap,
 	.open		= gfs2_open,
@@ -1310,6 +1311,7 @@ const struct file_operations gfs2_file_fops_nolock = {
 	.llseek		= gfs2_llseek,
 	.read_iter	= gfs2_file_read_iter,
 	.write_iter	= gfs2_file_write_iter,
+	.iopoll		= iomap_dio_iopoll,
 	.unlocked_ioctl	= gfs2_ioctl,
 	.mmap		= gfs2_mmap,
 	.open		= gfs2_open,
diff --git a/fs/iomap.c b/fs/iomap.c
index a3088fae567b..4ee50b76b4a1 100644
--- a/fs/iomap.c
+++ b/fs/iomap.c
@@ -1454,6 +1454,28 @@ struct iomap_dio {
 	};
 };
 
+int iomap_dio_iopoll(struct kiocb *kiocb, bool spin)
+{
+	struct request_queue *q = READ_ONCE(kiocb->private);
+
+	if (!q)
+		return 0;
+	return blk_poll(q, READ_ONCE(kiocb->ki_cookie), spin);
+}
+EXPORT_SYMBOL_GPL(iomap_dio_iopoll);
+
+static void iomap_dio_submit_bio(struct iomap_dio *dio, struct iomap *iomap,
+		struct bio *bio)
+{
+	atomic_inc(&dio->ref);
+
+	if (dio->iocb->ki_flags & IOCB_HIPRI)
+		bio_set_polled(bio, dio->iocb);
+
+	dio->submit.last_queue = bdev_get_queue(iomap->bdev);
+	dio->submit.cookie = submit_bio(bio);
+}
+
 static ssize_t iomap_dio_complete(struct iomap_dio *dio)
 {
 	struct kiocb *iocb = dio->iocb;
@@ -1566,7 +1588,7 @@ static void iomap_dio_bio_end_io(struct bio *bio)
 	}
 }
 
-static blk_qc_t
+static void
 iomap_dio_zero(struct iomap_dio *dio, struct iomap *iomap, loff_t pos,
 		unsigned len)
 {
@@ -1580,15 +1602,10 @@ iomap_dio_zero(struct iomap_dio *dio, struct iomap *iomap, loff_t pos,
 	bio->bi_private = dio;
 	bio->bi_end_io = iomap_dio_bio_end_io;
 
-	if (dio->iocb->ki_flags & IOCB_HIPRI)
-		flags |= REQ_HIPRI;
-
 	get_page(page);
 	__bio_add_page(bio, page, len, 0);
 	bio_set_op_attrs(bio, REQ_OP_WRITE, flags);
-
-	atomic_inc(&dio->ref);
-	return submit_bio(bio);
+	iomap_dio_submit_bio(dio, iomap, bio);
 }
 
 static loff_t
@@ -1691,9 +1708,6 @@ iomap_dio_bio_actor(struct inode *inode, loff_t pos, loff_t length,
 				bio_set_pages_dirty(bio);
 		}
 
-		if (dio->iocb->ki_flags & IOCB_HIPRI)
-			bio->bi_opf |= REQ_HIPRI;
-
 		iov_iter_advance(dio->submit.iter, n);
 
 		dio->size += n;
@@ -1701,11 +1715,7 @@ iomap_dio_bio_actor(struct inode *inode, loff_t pos, loff_t length,
 		copied += n;
 
 		nr_pages = iov_iter_npages(&iter, BIO_MAX_PAGES);
-
-		atomic_inc(&dio->ref);
-
-		dio->submit.last_queue = bdev_get_queue(iomap->bdev);
-		dio->submit.cookie = submit_bio(bio);
+		iomap_dio_submit_bio(dio, iomap, bio);
 	} while (nr_pages);
 
 	/*
@@ -1916,6 +1926,9 @@ iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter,
 	if (dio->flags & IOMAP_DIO_WRITE_FUA)
 		dio->flags &= ~IOMAP_DIO_NEED_SYNC;
 
+	WRITE_ONCE(iocb->ki_cookie, dio->submit.cookie);
+	WRITE_ONCE(iocb->private, dio->submit.last_queue);
+
 	if (!atomic_dec_and_test(&dio->ref)) {
 		if (!dio->wait_for_completion)
 			return -EIOCBQUEUED;
diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index e47425071e65..60c2da41f0fc 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -1203,6 +1203,7 @@ const struct file_operations xfs_file_operations = {
 	.write_iter	= xfs_file_write_iter,
 	.splice_read	= generic_file_splice_read,
 	.splice_write	= iter_file_splice_write,
+	.iopoll		= iomap_dio_iopoll,
 	.unlocked_ioctl	= xfs_file_ioctl,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl	= xfs_file_compat_ioctl,
diff --git a/include/linux/iomap.h b/include/linux/iomap.h
index 9a4258154b25..0fefb5455bda 100644
--- a/include/linux/iomap.h
+++ b/include/linux/iomap.h
@@ -162,6 +162,7 @@ typedef int (iomap_dio_end_io_t)(struct kiocb *iocb, ssize_t ret,
 		unsigned flags);
 ssize_t iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter,
 		const struct iomap_ops *ops, iomap_dio_end_io_t end_io);
+int iomap_dio_iopoll(struct kiocb *kiocb, bool spin);
 
 #ifdef CONFIG_SWAP
 struct file;
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 05/16] Add io_uring IO interface
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
                   ` (3 preceding siblings ...)
  2019-01-08 16:56 ` [PATCH 04/16] iomap: wire up the iopoll method Jens Axboe
@ 2019-01-08 16:56 ` Jens Axboe
  2019-01-09 12:10   ` Christoph Hellwig
  2019-01-08 16:56 ` [PATCH 06/16] io_uring: support for IO polling Jens Axboe
                   ` (11 subsequent siblings)
  16 siblings, 1 reply; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

The submission queue (SQ) and completion queue (CQ) rings are shared
between the application and the kernel. This eliminates the need to
copy data back and forth to submit and complete IO.

IO submissions use the io_uring_iocb data structure, and completions
are generated in the form of io_uring_event data structures. The SQ
ring is an index into the iocb_io_uring array, which makes it possible
to submit a batch of IOs without them being contiguous in the ring.
The CQ ring is always contiguous, as completion events are inherently
unordered and can point to any io_uring_iocb.

Two new system calls are added for this:

io_uring_setup(entries, iovecs, params)
	Sets up a context for doing async IO. On success, returns a file
	descriptor that the application can mmap to gain access to the
	SQ ring, CQ ring, and io_uring_iocbs.

io_uring_enter(fd, to_submit, min_complete, flags)
	Initiates IO against the rings mapped to this fd, or waits for
	them to complete, or both The behavior is controlled by the
	parameters passed in. If 'min_complete' is non-zero, then we'll
	try and submit new IO. If IORING_ENTER_GETEVENTS is set, the
	kernel will wait for 'min_complete' events, if they aren't
	already available.

With this setup, it's possible to do async IO with a single system
call. Future developments will enable polled IO with this interface,
and polled submission as well. The latter will enable an application
to do IO without doing ANY system calls at all.

For IRQ driven IO, an application only needs to enter the kernel for
completions if it wants to wait for them to occur.

Sample application: http://git.kernel.dk/cgit/fio/plain/t/io_uring.c

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 arch/x86/entry/syscalls/syscall_64.tbl |   2 +
 fs/Makefile                            |   2 +-
 fs/io_uring.c                          | 849 +++++++++++++++++++++++++
 include/linux/syscalls.h               |   5 +
 include/uapi/linux/io_uring.h          | 101 +++
 kernel/sys_ni.c                        |   2 +
 6 files changed, 960 insertions(+), 1 deletion(-)
 create mode 100644 fs/io_uring.c
 create mode 100644 include/uapi/linux/io_uring.h

diff --git a/arch/x86/entry/syscalls/syscall_64.tbl b/arch/x86/entry/syscalls/syscall_64.tbl
index f0b1709a5ffb..453ff7a79002 100644
--- a/arch/x86/entry/syscalls/syscall_64.tbl
+++ b/arch/x86/entry/syscalls/syscall_64.tbl
@@ -343,6 +343,8 @@
 332	common	statx			__x64_sys_statx
 333	common	io_pgetevents		__x64_sys_io_pgetevents
 334	common	rseq			__x64_sys_rseq
+335	common	io_uring_setup		__x64_sys_io_uring_setup
+336	common	io_uring_enter		__x64_sys_io_uring_enter
 
 #
 # x32-specific system call numbers start at 512 to avoid cache impact
diff --git a/fs/Makefile b/fs/Makefile
index 293733f61594..9ef9987b4192 100644
--- a/fs/Makefile
+++ b/fs/Makefile
@@ -29,7 +29,7 @@ obj-$(CONFIG_SIGNALFD)		+= signalfd.o
 obj-$(CONFIG_TIMERFD)		+= timerfd.o
 obj-$(CONFIG_EVENTFD)		+= eventfd.o
 obj-$(CONFIG_USERFAULTFD)	+= userfaultfd.o
-obj-$(CONFIG_AIO)               += aio.o
+obj-$(CONFIG_AIO)               += aio.o io_uring.o
 obj-$(CONFIG_FS_DAX)		+= dax.o
 obj-$(CONFIG_FS_ENCRYPTION)	+= crypto/
 obj-$(CONFIG_FILE_LOCKING)      += locks.o
diff --git a/fs/io_uring.c b/fs/io_uring.c
new file mode 100644
index 000000000000..ae2b886282bb
--- /dev/null
+++ b/fs/io_uring.c
@@ -0,0 +1,849 @@
+/*
+ * Shared application/kernel submission and completion ring pairs, for
+ * supporting fast/efficient IO.
+ *
+ * Copyright (C) 2019 Jens Axboe
+ */
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/syscalls.h>
+#include <linux/refcount.h>
+#include <linux/uio.h>
+
+#include <linux/sched/signal.h>
+#include <linux/fs.h>
+#include <linux/file.h>
+#include <linux/mm.h>
+#include <linux/mman.h>
+#include <linux/mmu_context.h>
+#include <linux/percpu.h>
+#include <linux/slab.h>
+#include <linux/workqueue.h>
+#include <linux/blkdev.h>
+#include <linux/anon_inodes.h>
+
+#include <linux/uaccess.h>
+#include <linux/nospec.h>
+
+#include <uapi/linux/io_uring.h>
+
+#include "internal.h"
+
+struct io_uring {
+	u32 head ____cacheline_aligned_in_smp;
+	u32 tail ____cacheline_aligned_in_smp;
+};
+
+struct io_sq_ring {
+	struct io_uring		r;
+	u32			ring_mask;
+	u32			ring_entries;
+	u32			dropped;
+	u32			flags;
+	u32			array[0];
+};
+
+struct io_cq_ring {
+	struct io_uring		r;
+	u32			ring_mask;
+	u32			ring_entries;
+	u32			overflow;
+	struct io_uring_event	events[0];
+};
+
+struct io_iocb_ring {
+	struct			io_sq_ring *ring;
+	unsigned		entries;
+	unsigned		ring_mask;
+	struct io_uring_iocb	*iocbs;
+};
+
+struct io_event_ring {
+	struct io_cq_ring	*ring;
+	unsigned		entries;
+	unsigned		ring_mask;
+};
+
+struct io_ring_ctx {
+	struct percpu_ref	refs;
+
+	unsigned int		flags;
+	unsigned int		max_reqs;
+
+	struct io_iocb_ring	sq_ring;
+	struct io_event_ring	cq_ring;
+
+	struct work_struct	work;
+
+	struct {
+		struct mutex uring_lock;
+	} ____cacheline_aligned_in_smp;
+
+	struct {
+		struct mutex    ring_lock;
+		wait_queue_head_t wait;
+	} ____cacheline_aligned_in_smp;
+
+	struct {
+		spinlock_t      completion_lock;
+	} ____cacheline_aligned_in_smp;
+};
+
+struct fsync_iocb {
+	struct work_struct	work;
+	struct file		*file;
+	bool			datasync;
+};
+
+struct io_kiocb {
+	union {
+		struct kiocb		rw;
+		struct fsync_iocb	fsync;
+	};
+
+	struct io_ring_ctx	*ki_ctx;
+	unsigned long		ki_index;
+	struct list_head	ki_list;
+	unsigned long		ki_flags;
+};
+
+#define IO_PLUG_THRESHOLD		2
+
+static struct kmem_cache *kiocb_cachep, *ioctx_cachep;
+
+static const struct file_operations io_scqring_fops;
+
+static void io_ring_ctx_free(struct work_struct *work);
+static void io_ring_ctx_ref_free(struct percpu_ref *ref);
+
+static struct io_ring_ctx *io_ring_ctx_alloc(struct io_uring_params *p)
+{
+	struct io_ring_ctx *ctx;
+
+	ctx = kmem_cache_zalloc(ioctx_cachep, GFP_KERNEL);
+	if (!ctx)
+		return NULL;
+
+	if (percpu_ref_init(&ctx->refs, io_ring_ctx_ref_free, 0, GFP_KERNEL)) {
+		kmem_cache_free(ioctx_cachep, ctx);
+		return NULL;
+	}
+
+	ctx->flags = p->flags;
+	ctx->max_reqs = p->sq_entries;
+
+	INIT_WORK(&ctx->work, io_ring_ctx_free);
+
+	spin_lock_init(&ctx->completion_lock);
+	mutex_init(&ctx->ring_lock);
+	init_waitqueue_head(&ctx->wait);
+	mutex_init(&ctx->uring_lock);
+
+	return ctx;
+}
+
+static void io_inc_cqring(struct io_ring_ctx *ctx)
+{
+	struct io_cq_ring *ring = ctx->cq_ring.ring;
+
+	ring->r.tail++;
+	smp_wmb();
+}
+
+static struct io_uring_event *io_peek_cqring(struct io_ring_ctx *ctx)
+{
+	struct io_cq_ring *ring = ctx->cq_ring.ring;
+	unsigned tail;
+
+	smp_rmb();
+	tail = READ_ONCE(ring->r.tail);
+	if (tail + 1 == READ_ONCE(ring->r.head))
+		return NULL;
+
+	return &ring->events[tail & ctx->cq_ring.ring_mask];
+}
+
+static struct io_kiocb *io_get_req(struct io_ring_ctx *ctx)
+{
+	struct io_kiocb *req;
+
+	req = kmem_cache_alloc(kiocb_cachep, GFP_KERNEL);
+	if (!req)
+		return NULL;
+
+	percpu_ref_get(&ctx->refs);
+	req->ki_ctx = ctx;
+	INIT_LIST_HEAD(&req->ki_list);
+	req->ki_flags = 0;
+	return req;
+}
+
+static inline void iocb_put(struct io_kiocb *iocb)
+{
+	percpu_ref_put(&iocb->ki_ctx->refs);
+	kmem_cache_free(kiocb_cachep, iocb);
+}
+
+static void io_complete_iocb(struct io_ring_ctx *ctx, struct io_kiocb *iocb)
+{
+	if (waitqueue_active(&ctx->wait))
+		wake_up(&ctx->wait);
+	iocb_put(iocb);
+}
+
+static void kiocb_end_write(struct kiocb *kiocb)
+{
+	if (kiocb->ki_flags & IOCB_WRITE) {
+		struct inode *inode = file_inode(kiocb->ki_filp);
+
+		/*
+		 * Tell lockdep we inherited freeze protection from submission
+		 * thread.
+		 */
+		if (S_ISREG(inode->i_mode))
+			__sb_writers_acquired(inode->i_sb, SB_FREEZE_WRITE);
+		file_end_write(kiocb->ki_filp);
+	}
+}
+
+static void io_fill_event(struct io_uring_event *ev, struct io_kiocb *kiocb,
+			  long res, unsigned flags)
+{
+	ev->index = kiocb->ki_index;
+	ev->res = res;
+	ev->flags = flags;
+}
+
+static void io_cqring_fill_event(struct io_kiocb *iocb, long res,
+				 unsigned ev_flags)
+{
+	struct io_ring_ctx *ctx = iocb->ki_ctx;
+	struct io_uring_event *ev;
+	unsigned long flags;
+
+	/*
+	 * If we can't get a cq entry, userspace overflowed the
+	 * submission (by quite a lot). Increment the overflow count in
+	 * the ring.
+	 */
+	spin_lock_irqsave(&ctx->completion_lock, flags);
+	ev = io_peek_cqring(ctx);
+	if (ev) {
+		io_fill_event(ev, iocb, res, ev_flags);
+		io_inc_cqring(ctx);
+	} else
+		ctx->cq_ring.ring->overflow++;
+	spin_unlock_irqrestore(&ctx->completion_lock, flags);
+}
+
+static void io_complete_scqring(struct io_kiocb *iocb, long res, unsigned flags)
+{
+	io_cqring_fill_event(iocb, res, flags);
+	io_complete_iocb(iocb->ki_ctx, iocb);
+}
+
+static void io_complete_scqring_rw(struct kiocb *kiocb, long res, long res2)
+{
+	struct io_kiocb *iocb = container_of(kiocb, struct io_kiocb, rw);
+
+	kiocb_end_write(kiocb);
+
+	fput(kiocb->ki_filp);
+	io_complete_scqring(iocb, res, 0);
+}
+
+static int io_prep_rw(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb)
+{
+	struct kiocb *req = &kiocb->rw;
+	int ret;
+
+	req->ki_filp = fget(iocb->fd);
+	if (unlikely(!req->ki_filp))
+		return -EBADF;
+	req->ki_pos = iocb->off;
+	req->ki_flags = iocb_flags(req->ki_filp);
+	req->ki_hint = ki_hint_validate(file_write_hint(req->ki_filp));
+	if (iocb->ioprio) {
+		ret = ioprio_check_cap(iocb->ioprio);
+		if (ret)
+			goto out_fput;
+
+		req->ki_ioprio = iocb->ioprio;
+	} else
+		req->ki_ioprio = get_current_ioprio();
+
+	ret = kiocb_set_rw_flags(req, iocb->rw_flags);
+	if (unlikely(ret))
+		goto out_fput;
+
+	/* no one is going to poll for this I/O */
+	req->ki_flags &= ~IOCB_HIPRI;
+	req->ki_complete = io_complete_scqring_rw;
+	return 0;
+out_fput:
+	fput(req->ki_filp);
+	return ret;
+}
+
+static int io_setup_rw(int rw, const struct io_uring_iocb *iocb,
+		       struct iovec **iovec, struct iov_iter *iter)
+{
+	void __user *buf = (void __user *)(uintptr_t)iocb->addr;
+	size_t ret;
+
+	ret = import_single_range(rw, buf, iocb->len, *iovec, iter);
+	*iovec = NULL;
+	return ret;
+}
+
+static inline void io_rw_done(struct kiocb *req, ssize_t ret)
+{
+	switch (ret) {
+	case -EIOCBQUEUED:
+		break;
+	case -ERESTARTSYS:
+	case -ERESTARTNOINTR:
+	case -ERESTARTNOHAND:
+	case -ERESTART_RESTARTBLOCK:
+		/*
+		 * There's no easy way to restart the syscall since other AIO's
+		 * may be already running. Just fail this IO with EINTR.
+		 */
+		ret = -EINTR;
+		/*FALLTHRU*/
+	default:
+		req->ki_complete(req, ret, 0);
+	}
+}
+
+static ssize_t io_read(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb)
+{
+	struct iovec inline_vecs[UIO_FASTIOV], *iovec = inline_vecs;
+	struct kiocb *req = &kiocb->rw;
+	struct iov_iter iter;
+	struct file *file;
+	ssize_t ret;
+
+	ret = io_prep_rw(kiocb, iocb);
+	if (ret)
+		return ret;
+	file = req->ki_filp;
+
+	ret = -EBADF;
+	if (unlikely(!(file->f_mode & FMODE_READ)))
+		goto out_fput;
+	ret = -EINVAL;
+	if (unlikely(!file->f_op->read_iter))
+		goto out_fput;
+
+	ret = io_setup_rw(READ, iocb, &iovec, &iter);
+	if (ret)
+		goto out_fput;
+
+	ret = rw_verify_area(READ, file, &req->ki_pos, iov_iter_count(&iter));
+	if (!ret)
+		io_rw_done(req, call_read_iter(file, req, &iter));
+	kfree(iovec);
+out_fput:
+	if (unlikely(ret))
+		fput(file);
+	return ret;
+}
+
+static ssize_t io_write(struct io_kiocb *kiocb,
+			const struct io_uring_iocb *iocb)
+{
+	struct iovec inline_vecs[UIO_FASTIOV], *iovec = inline_vecs;
+	struct kiocb *req = &kiocb->rw;
+	struct iov_iter iter;
+	struct file *file;
+	ssize_t ret;
+
+	ret = io_prep_rw(kiocb, iocb);
+	if (ret)
+		return ret;
+	file = req->ki_filp;
+
+	ret = -EBADF;
+	if (unlikely(!(file->f_mode & FMODE_WRITE)))
+		goto out_fput;
+	ret = -EINVAL;
+	if (unlikely(!file->f_op->write_iter))
+		goto out_fput;
+
+	ret = io_setup_rw(WRITE, iocb, &iovec, &iter);
+	if (ret)
+		goto out_fput;
+	ret = rw_verify_area(WRITE, file, &req->ki_pos, iov_iter_count(&iter));
+	if (!ret) {
+		/*
+		 * Open-code file_start_write here to grab freeze protection,
+		 * which will be released by another thread in
+		 * io_complete_rw().  Fool lockdep by telling it the lock got
+		 * released so that it doesn't complain about the held lock when
+		 * we return to userspace.
+		 */
+		if (S_ISREG(file_inode(file)->i_mode)) {
+			__sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, true);
+			__sb_writers_release(file_inode(file)->i_sb, SB_FREEZE_WRITE);
+		}
+		req->ki_flags |= IOCB_WRITE;
+		io_rw_done(req, call_write_iter(file, req, &iter));
+	}
+	kfree(iovec);
+out_fput:
+	if (unlikely(ret))
+		fput(file);
+	return ret;
+}
+
+static void io_fsync_work(struct work_struct *work)
+{
+	struct fsync_iocb *req = container_of(work, struct fsync_iocb, work);
+	struct io_kiocb *iocb = container_of(req, struct io_kiocb, fsync);
+	int ret;
+
+	ret = vfs_fsync(req->file, req->datasync);
+	fput(req->file);
+
+	io_complete_scqring(iocb, ret, 0);
+}
+
+static int io_fsync(struct fsync_iocb *req, const struct io_uring_iocb *iocb,
+		    bool datasync)
+{
+	if (unlikely(iocb->addr || iocb->off || iocb->len || iocb->__resv))
+		return -EINVAL;
+
+	req->file = fget(iocb->fd);
+	if (unlikely(!req->file))
+		return -EBADF;
+	if (unlikely(!req->file->f_op->fsync)) {
+		fput(req->file);
+		return -EINVAL;
+	}
+
+	req->datasync = datasync;
+	INIT_WORK(&req->work, io_fsync_work);
+	schedule_work(&req->work);
+	return 0;
+}
+
+static int __io_submit_one(struct io_ring_ctx *ctx,
+			   const struct io_uring_iocb *iocb,
+			   unsigned long ki_index)
+{
+	struct io_kiocb *req;
+	ssize_t ret;
+
+	/* enforce forwards compatibility on users */
+	if (unlikely(iocb->flags))
+		return -EINVAL;
+
+	req = io_get_req(ctx);
+	if (unlikely(!req))
+		return -EAGAIN;
+
+	ret = -EINVAL;
+	if (ki_index >= ctx->max_reqs)
+		goto out_put_req;
+	req->ki_index = ki_index;
+
+	ret = -EINVAL;
+	switch (iocb->opcode) {
+	case IORING_OP_READ:
+		ret = io_read(req, iocb);
+		break;
+	case IORING_OP_WRITE:
+		ret = io_write(req, iocb);
+		break;
+	case IORING_OP_FSYNC:
+		ret = io_fsync(&req->fsync, iocb, false);
+		break;
+	case IORING_OP_FDSYNC:
+		ret = io_fsync(&req->fsync, iocb, true);
+		break;
+	default:
+		ret = -EINVAL;
+		break;
+	}
+
+	/*
+	 * If ret is 0, ->ki_complete() has either been called, or will get
+	 * called later on. Anything else, we need to free the req.
+	 */
+	if (ret)
+		goto out_put_req;
+	return 0;
+out_put_req:
+	iocb_put(req);
+	return ret;
+}
+
+static void io_inc_sqring(struct io_ring_ctx *ctx)
+{
+	struct io_sq_ring *ring = ctx->sq_ring.ring;
+
+	ring->r.head++;
+	smp_wmb();
+}
+
+static const struct io_uring_iocb *io_peek_sqring(struct io_ring_ctx *ctx,
+						  unsigned *iocb_index)
+{
+	struct io_sq_ring *ring = ctx->sq_ring.ring;
+	unsigned head;
+
+	smp_rmb();
+	head = READ_ONCE(ring->r.head);
+	if (head == READ_ONCE(ring->r.tail))
+		return NULL;
+
+	head = ring->array[head & ctx->sq_ring.ring_mask];
+	if (head < ctx->sq_ring.entries) {
+		*iocb_index = head;
+		return &ctx->sq_ring.iocbs[head];
+	}
+
+	/* drop invalid entries */
+	ring->r.head++;
+	ring->dropped++;
+	smp_wmb();
+	return NULL;
+}
+
+static int io_ring_submit(struct io_ring_ctx *ctx, unsigned int to_submit)
+{
+	int i, ret = 0, submit = 0;
+	struct blk_plug plug;
+
+	if (to_submit > IO_PLUG_THRESHOLD)
+		blk_start_plug(&plug);
+
+	for (i = 0; i < to_submit; i++) {
+		const struct io_uring_iocb *iocb;
+		unsigned iocb_index;
+
+		iocb = io_peek_sqring(ctx, &iocb_index);
+		if (!iocb)
+			break;
+
+		ret = __io_submit_one(ctx, iocb, iocb_index);
+		if (ret)
+			break;
+
+		submit++;
+		io_inc_sqring(ctx);
+	}
+
+	if (to_submit > IO_PLUG_THRESHOLD)
+		blk_finish_plug(&plug);
+
+	return submit ? submit : ret;
+}
+
+/*
+ * Wait until events become available, if we don't already have some. The
+ * application must reap them itself, as they reside on the shared cq ring.
+ */
+static int io_cqring_wait(struct io_ring_ctx *ctx, int min_events)
+{
+	struct io_cq_ring *ring = ctx->cq_ring.ring;
+	DEFINE_WAIT(wait);
+	int ret;
+
+	smp_rmb();
+	if (ring->r.head != ring->r.tail)
+		return 0;
+	if (!min_events)
+		return 0;
+
+	do {
+		prepare_to_wait(&ctx->wait, &wait, TASK_INTERRUPTIBLE);
+
+		ret = 0;
+		smp_rmb();
+		if (ring->r.head != ring->r.tail)
+			break;
+
+		schedule();
+
+		ret = -EINTR;
+		if (signal_pending(current))
+			break;
+	} while (1);
+
+	finish_wait(&ctx->wait, &wait);
+	return ring->r.head == ring->r.tail ? ret : 0;
+}
+
+static int __io_uring_enter(struct io_ring_ctx *ctx, unsigned to_submit,
+			    unsigned min_complete, unsigned flags)
+{
+	int ret = 0;
+
+	if (to_submit) {
+		ret = io_ring_submit(ctx, to_submit);
+		if (ret < 0)
+			return ret;
+	}
+	if (flags & IORING_ENTER_GETEVENTS) {
+		int get_ret;
+
+		if (!ret && to_submit)
+			min_complete = 0;
+
+		get_ret = io_cqring_wait(ctx, min_complete);
+		if (get_ret < 0 && !ret)
+			ret = get_ret;
+	}
+
+	return ret;
+}
+
+static void io_free_scq_urings(struct io_ring_ctx *ctx)
+{
+	if (ctx->sq_ring.ring) {
+		page_frag_free(ctx->sq_ring.ring);
+		ctx->sq_ring.ring = NULL;
+	}
+	if (ctx->sq_ring.iocbs) {
+		page_frag_free(ctx->sq_ring.iocbs);
+		ctx->sq_ring.iocbs = NULL;
+	}
+	if (ctx->cq_ring.ring) {
+		page_frag_free(ctx->cq_ring.ring);
+		ctx->cq_ring.ring = NULL;
+	}
+}
+
+static void io_ring_ctx_free(struct work_struct *work)
+{
+	struct io_ring_ctx *ctx = container_of(work, struct io_ring_ctx, work);
+
+	io_free_scq_urings(ctx);
+	percpu_ref_exit(&ctx->refs);
+	kmem_cache_free(ioctx_cachep, ctx);
+}
+
+static void io_ring_ctx_ref_free(struct percpu_ref *ref)
+{
+	struct io_ring_ctx *ctx = container_of(ref, struct io_ring_ctx, refs);
+
+	schedule_work(&ctx->work);
+}
+
+static int io_scqring_release(struct inode *inode, struct file *file)
+{
+	struct io_ring_ctx *ctx = file->private_data;
+
+	file->private_data = NULL;
+	percpu_ref_kill(&ctx->refs);
+	return 0;
+}
+
+static int io_scqring_mmap(struct file *file, struct vm_area_struct *vma)
+{
+	loff_t offset = (loff_t) vma->vm_pgoff << PAGE_SHIFT;
+	unsigned long sz = vma->vm_end - vma->vm_start;
+	struct io_ring_ctx *ctx = file->private_data;
+	unsigned long pfn;
+	struct page *page;
+	void *ptr;
+
+	switch (offset) {
+	case IORING_OFF_SQ_RING:
+		ptr = ctx->sq_ring.ring;
+		break;
+	case IORING_OFF_IOCB:
+		ptr = ctx->sq_ring.iocbs;
+		break;
+	case IORING_OFF_CQ_RING:
+		ptr = ctx->cq_ring.ring;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	page = virt_to_head_page(ptr);
+	if (sz > (PAGE_SIZE << compound_order(page)))
+		return -EINVAL;
+
+	pfn = virt_to_phys(ptr) >> PAGE_SHIFT;
+	return remap_pfn_range(vma, vma->vm_start, pfn, sz, vma->vm_page_prot);
+}
+
+SYSCALL_DEFINE4(io_uring_enter, unsigned int, fd, u32, to_submit,
+		u32, min_complete, u32, flags)
+{
+	long ret = -EBADF;
+	struct fd f;
+
+	f = fdget(fd);
+	if (f.file) {
+		struct io_ring_ctx *ctx;
+
+		ret = -EOPNOTSUPP;
+		if (f.file->f_op != &io_scqring_fops)
+			goto err;
+
+		ctx = f.file->private_data;
+		ret = -EBUSY;
+		if (!mutex_trylock(&ctx->uring_lock))
+			goto err;
+
+		ret = __io_uring_enter(ctx, to_submit, min_complete, flags);
+		mutex_unlock(&ctx->uring_lock);
+err:
+		fdput(f);
+	}
+
+	return ret;
+}
+
+static const struct file_operations io_scqring_fops = {
+	.release	= io_scqring_release,
+	.mmap		= io_scqring_mmap,
+};
+
+static void *io_mem_alloc(size_t size)
+{
+	gfp_t gfp_flags = GFP_KERNEL | __GFP_ZERO | __GFP_NOWARN | __GFP_COMP |
+				__GFP_NORETRY;
+
+	return (void *) __get_free_pages(gfp_flags, get_order(size));
+}
+
+static int io_allocate_scq_urings(struct io_ring_ctx *ctx,
+				  struct io_uring_params *p)
+{
+	struct io_sq_ring *sq_ring;
+	struct io_cq_ring *cq_ring;
+
+	sq_ring = io_mem_alloc(struct_size(sq_ring, array, p->sq_entries));
+	if (!sq_ring)
+		return -ENOMEM;
+
+	ctx->sq_ring.ring = sq_ring;
+	sq_ring->ring_mask = p->sq_entries - 1;
+	sq_ring->ring_entries = p->sq_entries;
+	ctx->sq_ring.ring_mask = sq_ring->ring_mask;
+	ctx->sq_ring.entries = sq_ring->ring_entries;
+
+	ctx->sq_ring.iocbs = io_mem_alloc(sizeof(struct io_uring_iocb) *
+						p->sq_entries);
+	if (!ctx->sq_ring.iocbs)
+		goto err;
+
+	cq_ring = io_mem_alloc(struct_size(cq_ring, events, p->cq_entries));
+	if (!cq_ring)
+		goto err;
+
+	ctx->cq_ring.ring = cq_ring;
+	cq_ring->ring_mask = p->cq_entries - 1;
+	cq_ring->ring_entries = p->cq_entries;
+	ctx->cq_ring.ring_mask = cq_ring->ring_mask;
+	ctx->cq_ring.entries = cq_ring->ring_entries;
+	return 0;
+err:
+	io_free_scq_urings(ctx);
+	return -ENOMEM;
+}
+
+static void io_fill_offsets(struct io_uring_params *p)
+{
+	memset(&p->sq_off, 0, sizeof(p->sq_off));
+	p->sq_off.head = offsetof(struct io_sq_ring, r.head);
+	p->sq_off.tail = offsetof(struct io_sq_ring, r.tail);
+	p->sq_off.ring_mask = offsetof(struct io_sq_ring, ring_mask);
+	p->sq_off.ring_entries = offsetof(struct io_sq_ring, ring_entries);
+	p->sq_off.flags = offsetof(struct io_sq_ring, flags);
+	p->sq_off.dropped = offsetof(struct io_sq_ring, dropped);
+	p->sq_off.array = offsetof(struct io_sq_ring, array);
+
+	memset(&p->cq_off, 0, sizeof(p->cq_off));
+	p->cq_off.head = offsetof(struct io_cq_ring, r.head);
+	p->cq_off.tail = offsetof(struct io_cq_ring, r.tail);
+	p->cq_off.ring_mask = offsetof(struct io_cq_ring, ring_mask);
+	p->cq_off.ring_entries = offsetof(struct io_cq_ring, ring_entries);
+	p->cq_off.overflow = offsetof(struct io_cq_ring, overflow);
+	p->cq_off.events = offsetof(struct io_cq_ring, events);
+}
+
+static int io_uring_create(unsigned entries, struct io_uring_params *p)
+{
+	struct io_ring_ctx *ctx;
+	int ret;
+
+	/*
+	 * Use twice as many entries for the CQ ring. It's possible for the
+	 * application to drive a higher depth than the size of the SQ ring,
+	 * since the iocbs are only used at submission time. This allows for
+	 * some flexibility in overcommitting a bit.
+	 */
+	p->sq_entries = roundup_pow_of_two(entries);
+	p->cq_entries = 2 * p->sq_entries;
+
+	ctx = io_ring_ctx_alloc(p);
+	if (!ctx)
+		return -ENOMEM;
+
+	ret = io_allocate_scq_urings(ctx, p);
+	if (ret)
+		goto err;
+
+	ret = anon_inode_getfd("[io_uring]", &io_scqring_fops, ctx,
+				O_RDWR | O_CLOEXEC);
+	if (ret < 0)
+		goto err;
+
+	io_fill_offsets(p);
+	return ret;
+err:
+	percpu_ref_kill(&ctx->refs);
+	return ret;
+}
+
+/*
+ * sys_io_uring_setup:
+ *	Sets up an aio uring context, and returns the fd. Applications asks
+ *	for a ring size, we return the actual sq/cq ring sizes (among other
+ *	things) in the params structure passed in.
+ */
+SYSCALL_DEFINE3(io_uring_setup, u32, entries, struct iovec __user *, iovecs,
+		struct io_uring_params __user *, params)
+{
+	struct io_uring_params p;
+	long ret;
+	int i;
+
+	if (copy_from_user(&p, params, sizeof(p)))
+		return -EFAULT;
+	for (i = 0; i < ARRAY_SIZE(p.resv); i++) {
+		if (p.resv[i])
+			return -EINVAL;
+	}
+
+	if (p.flags)
+		return -EINVAL;
+	if (iovecs)
+		return -EINVAL;
+
+	ret = io_uring_create(entries, &p);
+	if (ret < 0)
+		return ret;
+
+	if (copy_to_user(params, &p, sizeof(p)))
+		return -EFAULT;
+
+	return ret;
+}
+
+static int __init io_uring_setup(void)
+{
+	kiocb_cachep = KMEM_CACHE(io_kiocb, SLAB_HWCACHE_ALIGN | SLAB_PANIC);
+	ioctx_cachep = KMEM_CACHE(io_ring_ctx, SLAB_HWCACHE_ALIGN | SLAB_PANIC);
+	return 0;
+};
+__initcall(io_uring_setup);
diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h
index 257cccba3062..6d40939f65cd 100644
--- a/include/linux/syscalls.h
+++ b/include/linux/syscalls.h
@@ -69,6 +69,7 @@ struct file_handle;
 struct sigaltstack;
 struct rseq;
 union bpf_attr;
+struct io_uring_params;
 
 #include <linux/types.h>
 #include <linux/aio_abi.h>
@@ -309,6 +310,10 @@ asmlinkage long sys_io_pgetevents_time32(aio_context_t ctx_id,
 				struct io_event __user *events,
 				struct old_timespec32 __user *timeout,
 				const struct __aio_sigset *sig);
+asmlinkage long sys_io_uring_setup(u32 entries, struct iovec __user *iov,
+				struct io_uring_params __user *p);
+asmlinkage long sys_io_uring_enter(unsigned int fd, u32 to_submit,
+				u32 min_complete, u32 flags);
 
 /* fs/xattr.c */
 asmlinkage long sys_setxattr(const char __user *path, const char __user *name,
diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
new file mode 100644
index 000000000000..c31ac84d9f53
--- /dev/null
+++ b/include/uapi/linux/io_uring.h
@@ -0,0 +1,101 @@
+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
+/*
+ * Header file for the io_uring interface.
+ *
+ * Copyright (C) 2019 Jens Axboe
+ * Copyright (C) 2019 Christoph Hellwig
+ */
+#ifndef LINUX_IO_URING_H
+#define LINUX_IO_URING_H
+
+#include <linux/fs.h>
+#include <linux/types.h>
+
+/*
+ * IO submission data structure
+ */
+struct io_uring_iocb {
+	__u8	opcode;
+	__u8	flags;
+	__u16	ioprio;
+	__s32	fd;
+	__u64	off;
+	union {
+		void	*addr;
+		__u64	__pad;
+	};
+	__u32	len;
+	union {
+		__kernel_rwf_t	rw_flags;
+		__u32		__resv;
+	};
+};
+
+#define IORING_OP_READ		1
+#define IORING_OP_WRITE		2
+#define IORING_OP_FSYNC		3
+#define IORING_OP_FDSYNC	4
+
+/*
+ * IO completion data structure
+ */
+struct io_uring_event {
+	__u64	index;		/* what iocb this event came from */
+	__s32	res;		/* result code for this event */
+	__u32	flags;
+};
+
+/*
+ * io_uring_event->flags
+ */
+#define IOEV_FLAG_CACHEHIT	(1 << 0)	/* IO did not hit media */
+
+/*
+ * Magic offsets for the application to mmap the data it needs
+ */
+#define IORING_OFF_SQ_RING		0ULL
+#define IORING_OFF_CQ_RING		0x8000000ULL
+#define IORING_OFF_IOCB			0x10000000ULL
+
+/*
+ * Filled with the offset for mmap(2)
+ */
+struct io_sqring_offsets {
+	__u32 head;
+	__u32 tail;
+	__u32 ring_mask;
+	__u32 ring_entries;
+	__u32 flags;
+	__u32 dropped;
+	__u32 array;
+	__u32 resv[3];
+};
+
+struct io_cqring_offsets {
+	__u32 head;
+	__u32 tail;
+	__u32 ring_mask;
+	__u32 ring_entries;
+	__u32 overflow;
+	__u32 events;
+	__u32 resv[4];
+};
+
+/*
+ * io_uring_enter(2) flags
+ */
+#define IORING_ENTER_GETEVENTS	(1 << 0)
+
+/*
+ * Passed in for io_uring_setup(2). Copied back with updated info on success
+ */
+struct io_uring_params {
+	__u32 sq_entries;
+	__u32 cq_entries;
+	__u32 flags;
+	__u16 resv[10];
+	struct io_sqring_offsets sq_off;
+	struct io_cqring_offsets cq_off;
+};
+
+#endif
diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c
index ab9d0e3c6d50..ee5e523564bb 100644
--- a/kernel/sys_ni.c
+++ b/kernel/sys_ni.c
@@ -46,6 +46,8 @@ COND_SYSCALL(io_getevents);
 COND_SYSCALL(io_pgetevents);
 COND_SYSCALL_COMPAT(io_getevents);
 COND_SYSCALL_COMPAT(io_pgetevents);
+COND_SYSCALL(io_uring_setup);
+COND_SYSCALL(io_uring_enter);
 
 /* fs/xattr.c */
 
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 06/16] io_uring: support for IO polling
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
                   ` (4 preceding siblings ...)
  2019-01-08 16:56 ` [PATCH 05/16] Add io_uring IO interface Jens Axboe
@ 2019-01-08 16:56 ` Jens Axboe
  2019-01-09 12:11   ` Christoph Hellwig
  2019-01-08 16:56 ` [PATCH 07/16] io_uring: add submission side request cache Jens Axboe
                   ` (10 subsequent siblings)
  16 siblings, 1 reply; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

Add polled variants of the read and write commands. These act like their
non-polled counterparts, except we expect to poll for completion of
them.

To use polling, io_uring_setup() must be used with the
IORING_SETUP_IOPOLL flag being set. It is illegal to mix and match
polled and non-polled IO on an io_uring.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/io_uring.c                 | 227 +++++++++++++++++++++++++++++++++-
 include/uapi/linux/io_uring.h |  10 +-
 2 files changed, 227 insertions(+), 10 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index ae2b886282bb..02eab2f42c63 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -76,7 +76,14 @@ struct io_ring_ctx {
 
 	struct work_struct	work;
 
+	/* iopoll submission state */
 	struct {
+		spinlock_t poll_lock;
+		struct list_head poll_submitted;
+	} ____cacheline_aligned_in_smp;
+
+	struct {
+		struct list_head poll_completing;
 		struct mutex uring_lock;
 	} ____cacheline_aligned_in_smp;
 
@@ -106,10 +113,14 @@ struct io_kiocb {
 	unsigned long		ki_index;
 	struct list_head	ki_list;
 	unsigned long		ki_flags;
+#define KIOCB_F_IOPOLL_COMPLETED	0	/* polled IO has completed */
+#define KIOCB_F_IOPOLL_EAGAIN		1	/* submission got EAGAIN */
 };
 
 #define IO_PLUG_THRESHOLD		2
 
+#define IO_IOPOLL_BATCH	8
+
 static struct kmem_cache *kiocb_cachep, *ioctx_cachep;
 
 static const struct file_operations io_scqring_fops;
@@ -138,6 +149,9 @@ static struct io_ring_ctx *io_ring_ctx_alloc(struct io_uring_params *p)
 	spin_lock_init(&ctx->completion_lock);
 	mutex_init(&ctx->ring_lock);
 	init_waitqueue_head(&ctx->wait);
+	spin_lock_init(&ctx->poll_lock);
+	INIT_LIST_HEAD(&ctx->poll_submitted);
+	INIT_LIST_HEAD(&ctx->poll_completing);
 	mutex_init(&ctx->uring_lock);
 
 	return ctx;
@@ -185,6 +199,15 @@ static inline void iocb_put(struct io_kiocb *iocb)
 	kmem_cache_free(kiocb_cachep, iocb);
 }
 
+static void iocb_put_many(struct io_ring_ctx *ctx, void **iocbs, int *nr)
+{
+	if (*nr) {
+		percpu_ref_put_many(&ctx->refs, *nr);
+		kmem_cache_free_bulk(kiocb_cachep, *nr, iocbs);
+		*nr = 0;
+	}
+}
+
 static void io_complete_iocb(struct io_ring_ctx *ctx, struct io_kiocb *iocb)
 {
 	if (waitqueue_active(&ctx->wait))
@@ -192,6 +215,134 @@ static void io_complete_iocb(struct io_ring_ctx *ctx, struct io_kiocb *iocb)
 	iocb_put(iocb);
 }
 
+/*
+ * Find and free completed poll iocbs
+ */
+static void io_iopoll_reap(struct io_ring_ctx *ctx, unsigned int *nr_events)
+{
+	void *iocbs[IO_IOPOLL_BATCH];
+	struct io_kiocb *iocb, *n;
+	int to_free = 0;
+
+	list_for_each_entry_safe(iocb, n, &ctx->poll_completing, ki_list) {
+		if (!test_bit(KIOCB_F_IOPOLL_COMPLETED, &iocb->ki_flags))
+			continue;
+		if (to_free == ARRAY_SIZE(iocbs))
+			iocb_put_many(ctx, iocbs, &to_free);
+
+		list_del(&iocb->ki_list);
+		iocbs[to_free++] = iocb;
+
+		fput(iocb->rw.ki_filp);
+		(*nr_events)++;
+	}
+
+	if (to_free)
+		iocb_put_many(ctx, iocbs, &to_free);
+}
+
+/*
+ * Poll for a mininum of 'min' events, and a maximum of 'max'. Note that if
+ * min == 0 we consider that a non-spinning poll check - we'll still enter
+ * the driver poll loop, but only as a non-spinning completion check.
+ */
+static int io_iopoll_getevents(struct io_ring_ctx *ctx, unsigned int *nr_events,
+				long min)
+{
+	struct io_kiocb *iocb;
+	int found, polled, ret;
+
+	/*
+	 * Check if we already have done events that satisfy what we need
+	 */
+	if (!list_empty(&ctx->poll_completing)) {
+		io_iopoll_reap(ctx, nr_events);
+		if (min && *nr_events >= min)
+			return 0;
+	}
+
+	/*
+	 * Take in a new working set from the submitted list, if possible.
+	 */
+	if (!list_empty_careful(&ctx->poll_submitted)) {
+		spin_lock(&ctx->poll_lock);
+		list_splice_init(&ctx->poll_submitted, &ctx->poll_completing);
+		spin_unlock(&ctx->poll_lock);
+	}
+
+	if (list_empty(&ctx->poll_completing))
+		return 0;
+
+	/*
+	 * Check again now that we have a new batch.
+	 */
+	io_iopoll_reap(ctx, nr_events);
+	if (min && *nr_events >= min)
+		return 0;
+
+	polled = found = 0;
+	list_for_each_entry(iocb, &ctx->poll_completing, ki_list) {
+		/*
+		 * Poll for needed events with spin == true, anything after
+		 * that we just check if we have more, up to max.
+		 */
+		bool spin = !polled || *nr_events < min;
+		struct kiocb *kiocb = &iocb->rw;
+
+		if (test_bit(KIOCB_F_IOPOLL_COMPLETED, &iocb->ki_flags))
+			break;
+
+		found++;
+		ret = kiocb->ki_filp->f_op->iopoll(kiocb, spin);
+		if (ret < 0)
+			return ret;
+
+		polled += ret;
+	}
+
+	io_iopoll_reap(ctx, nr_events);
+	if (*nr_events >= min)
+		return 0;
+	return found;
+}
+
+/*
+ * We can't just wait for polled events to come to us, we have to actively
+ * find and complete them.
+ */
+static void io_iopoll_reap_events(struct io_ring_ctx *ctx)
+{
+	if (!(ctx->flags & IORING_SETUP_IOPOLL))
+		return;
+
+	while (!list_empty_careful(&ctx->poll_submitted) ||
+	       !list_empty(&ctx->poll_completing)) {
+		unsigned int nr_events = 0;
+
+		io_iopoll_getevents(ctx, &nr_events, 1);
+	}
+}
+
+static int io_iopoll_check(struct io_ring_ctx *ctx, unsigned *nr_events,
+			   long min)
+{
+	int ret = 0;
+
+	while (!*nr_events || !need_resched()) {
+		int tmin = 0;
+
+		if (*nr_events < min)
+			tmin = min - *nr_events;
+
+		ret = io_iopoll_getevents(ctx, nr_events, tmin);
+		if (ret <= 0)
+			break;
+		ret = 0;
+	}
+
+	return ret;
+}
+
 static void kiocb_end_write(struct kiocb *kiocb)
 {
 	if (kiocb->ki_flags & IOCB_WRITE) {
@@ -253,8 +404,23 @@ static void io_complete_scqring_rw(struct kiocb *kiocb, long res, long res2)
 	io_complete_scqring(iocb, res, 0);
 }
 
+static void io_complete_scqring_iopoll(struct kiocb *kiocb, long res, long res2)
+{
+	struct io_kiocb *iocb = container_of(kiocb, struct io_kiocb, rw);
+
+	kiocb_end_write(kiocb);
+
+	if (unlikely(res == -EAGAIN)) {
+		set_bit(KIOCB_F_IOPOLL_EAGAIN, &iocb->ki_flags);
+	} else {
+		io_cqring_fill_event(iocb, res, 0);
+		set_bit(KIOCB_F_IOPOLL_COMPLETED, &iocb->ki_flags);
+	}
+}
+
 static int io_prep_rw(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb)
 {
+	struct io_ring_ctx *ctx = kiocb->ki_ctx;
 	struct kiocb *req = &kiocb->rw;
 	int ret;
 
@@ -277,9 +443,19 @@ static int io_prep_rw(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb)
 	if (unlikely(ret))
 		goto out_fput;
 
-	/* no one is going to poll for this I/O */
-	req->ki_flags &= ~IOCB_HIPRI;
-	req->ki_complete = io_complete_scqring_rw;
+	if (ctx->flags & IORING_SETUP_IOPOLL) {
+		ret = -EOPNOTSUPP;
+		if (!(req->ki_flags & IOCB_DIRECT) ||
+		    !req->ki_filp->f_op->iopoll)
+			goto out_fput;
+
+		req->ki_flags |= IOCB_HIPRI;
+		req->ki_complete = io_complete_scqring_iopoll;
+	} else {
+		/* no one is going to poll for this I/O */
+		req->ki_flags &= ~IOCB_HIPRI;
+		req->ki_complete = io_complete_scqring_rw;
+	}
 	return 0;
 out_fput:
 	fput(req->ki_filp);
@@ -317,6 +493,30 @@ static inline void io_rw_done(struct kiocb *req, ssize_t ret)
 	}
 }
 
+/*
+ * After the iocb has been issued, it's safe to be found on the poll list.
+ * Adding the kiocb to the list AFTER submission ensures that we don't
+ * find it from a io_getevents() thread before the issuer is done accessing
+ * the kiocb cookie.
+ */
+static void io_iopoll_iocb_issued(struct io_kiocb *kiocb)
+{
+	/*
+	 * For fast devices, IO may have already completed. If it has, add
+	 * it to the front so we find it first. We can't add to the poll_done
+	 * list as that's unlocked from the completion side.
+	 */
+	const int front = test_bit(KIOCB_F_IOPOLL_COMPLETED, &kiocb->ki_flags);
+	struct io_ring_ctx *ctx = kiocb->ki_ctx;
+
+	spin_lock(&ctx->poll_lock);
+	if (front)
+		list_add(&kiocb->ki_list, &ctx->poll_submitted);
+	else
+		list_add_tail(&kiocb->ki_list, &ctx->poll_submitted);
+	spin_unlock(&ctx->poll_lock);
+}
+
 static ssize_t io_read(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb)
 {
 	struct iovec inline_vecs[UIO_FASTIOV], *iovec = inline_vecs;
@@ -459,9 +659,13 @@ static int __io_submit_one(struct io_ring_ctx *ctx,
 		ret = io_write(req, iocb);
 		break;
 	case IORING_OP_FSYNC:
+		if (ctx->flags & IORING_SETUP_IOPOLL)
+			break;
 		ret = io_fsync(&req->fsync, iocb, false);
 		break;
 	case IORING_OP_FDSYNC:
+		if (ctx->flags & IORING_SETUP_IOPOLL)
+			break;
 		ret = io_fsync(&req->fsync, iocb, true);
 		break;
 	default:
@@ -475,6 +679,13 @@ static int __io_submit_one(struct io_ring_ctx *ctx,
 	 */
 	if (ret)
 		goto out_put_req;
+	if (ctx->flags & IORING_SETUP_IOPOLL) {
+		if (test_bit(KIOCB_F_IOPOLL_EAGAIN, &req->ki_flags)) {
+			ret = -EAGAIN;
+			goto out_put_req;
+		}
+		io_iopoll_iocb_issued(req);
+	}
 	return 0;
 out_put_req:
 	iocb_put(req);
@@ -589,12 +800,17 @@ static int __io_uring_enter(struct io_ring_ctx *ctx, unsigned to_submit,
 			return ret;
 	}
 	if (flags & IORING_ENTER_GETEVENTS) {
+		unsigned nr_events = 0;
 		int get_ret;
 
 		if (!ret && to_submit)
 			min_complete = 0;
 
-		get_ret = io_cqring_wait(ctx, min_complete);
+		if (ctx->flags & IORING_SETUP_IOPOLL)
+			get_ret = io_iopoll_check(ctx, &nr_events,
+							min_complete);
+		else
+			get_ret = io_cqring_wait(ctx, min_complete);
 		if (get_ret < 0 && !ret)
 			ret = get_ret;
 	}
@@ -622,6 +838,7 @@ static void io_ring_ctx_free(struct work_struct *work)
 {
 	struct io_ring_ctx *ctx = container_of(work, struct io_ring_ctx, work);
 
+	io_iopoll_reap_events(ctx);
 	io_free_scq_urings(ctx);
 	percpu_ref_exit(&ctx->refs);
 	kmem_cache_free(ioctx_cachep, ctx);
@@ -825,7 +1042,7 @@ SYSCALL_DEFINE3(io_uring_setup, u32, entries, struct iovec __user *, iovecs,
 			return -EINVAL;
 	}
 
-	if (p.flags)
+	if (p.flags & ~IORING_SETUP_IOPOLL)
 		return -EINVAL;
 	if (iovecs)
 		return -EINVAL;
diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
index c31ac84d9f53..f7ba30747816 100644
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -31,6 +31,11 @@ struct io_uring_iocb {
 	};
 };
 
+/*
+ * io_uring_setup() flags
+ */
+#define IORING_SETUP_IOPOLL	(1 << 0)	/* io_context is polled */
+
 #define IORING_OP_READ		1
 #define IORING_OP_WRITE		2
 #define IORING_OP_FSYNC		3
@@ -45,11 +50,6 @@ struct io_uring_event {
 	__u32	flags;
 };
 
-/*
- * io_uring_event->flags
- */
-#define IOEV_FLAG_CACHEHIT	(1 << 0)	/* IO did not hit media */
-
 /*
  * Magic offsets for the application to mmap the data it needs
  */
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 07/16] io_uring: add submission side request cache
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
                   ` (5 preceding siblings ...)
  2019-01-08 16:56 ` [PATCH 06/16] io_uring: support for IO polling Jens Axboe
@ 2019-01-08 16:56 ` Jens Axboe
  2019-01-08 16:56 ` [PATCH 08/16] fs: add fget_many() and fput_many() Jens Axboe
                   ` (9 subsequent siblings)
  16 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

We have to add each submitted polled request to the io_context
poll_submitted list, which means we have to grab the poll_lock. We
already use the block plug to batch submissions if we're doing a batch
of IO submissions, extend that to cover the poll requests internally as
well.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/io_uring.c | 122 +++++++++++++++++++++++++++++++++++++++++++-------
 1 file changed, 106 insertions(+), 16 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 02eab2f42c63..9f36eb728208 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -121,6 +121,21 @@ struct io_kiocb {
 
 #define IO_IOPOLL_BATCH	8
 
+struct io_submit_state {
+	struct io_ring_ctx *ctx;
+
+	struct blk_plug plug;
+#ifdef CONFIG_BLOCK
+	struct blk_plug_cb plug_cb;
+#endif
+
+	/*
+	 * Polled iocbs that have been submitted, but not added to the ctx yet
+	 */
+	struct list_head req_list;
+	unsigned int req_count;
+};
+
 static struct kmem_cache *kiocb_cachep, *ioctx_cachep;
 
 static const struct file_operations io_scqring_fops;
@@ -494,21 +509,29 @@ static inline void io_rw_done(struct kiocb *req, ssize_t ret)
 }
 
 /*
- * After the iocb has been issued, it's safe to be found on the poll list.
- * Adding the kiocb to the list AFTER submission ensures that we don't
- * find it from a io_getevents() thread before the issuer is done accessing
- * the kiocb cookie.
+ * Called either at the end of IO submission, or through a plug callback
+ * because we're going to schedule. Moves out local batch of requests to
+ * the ctx poll list, so they can be found for polling + reaping.
  */
-static void io_iopoll_iocb_issued(struct io_kiocb *kiocb)
+static void io_flush_state_reqs(struct io_ring_ctx *ctx,
+				 struct io_submit_state *state)
 {
+	spin_lock(&ctx->poll_lock);
+	list_splice_tail_init(&state->req_list, &ctx->poll_submitted);
+	spin_unlock(&ctx->poll_lock);
+	state->req_count = 0;
+}
+
+static void io_iopoll_iocb_add_list(struct io_kiocb *kiocb)
+{
+	const int front = test_bit(KIOCB_F_IOPOLL_COMPLETED, &kiocb->ki_flags);
+	struct io_ring_ctx *ctx = kiocb->ki_ctx;
+
 	/*
 	 * For fast devices, IO may have already completed. If it has, add
 	 * it to the front so we find it first. We can't add to the poll_done
 	 * list as that's unlocked from the completion side.
 	 */
-	const int front = test_bit(KIOCB_F_IOPOLL_COMPLETED, &kiocb->ki_flags);
-	struct io_ring_ctx *ctx = kiocb->ki_ctx;
-
 	spin_lock(&ctx->poll_lock);
 	if (front)
 		list_add(&kiocb->ki_list, &ctx->poll_submitted);
@@ -517,6 +540,33 @@ static void io_iopoll_iocb_issued(struct io_kiocb *kiocb)
 	spin_unlock(&ctx->poll_lock);
 }
 
+static void io_iopoll_iocb_add_state(struct io_submit_state *state,
+				     struct io_kiocb *kiocb)
+{
+	if (test_bit(KIOCB_F_IOPOLL_COMPLETED, &kiocb->ki_flags))
+		list_add(&kiocb->ki_list, &state->req_list);
+	else
+		list_add_tail(&kiocb->ki_list, &state->req_list);
+
+	if (++state->req_count >= IO_IOPOLL_BATCH)
+		io_flush_state_reqs(state->ctx, state);
+}
+
+/*
+ * After the iocb has been issued, it's safe to be found on the poll list.
+ * Adding the kiocb to the list AFTER submission ensures that we don't
+ * find it from a io_getevents() thread before the issuer is done accessing
+ * the kiocb cookie.
+ */
+static void io_iopoll_iocb_issued(struct io_submit_state *state,
+				  struct io_kiocb *kiocb)
+{
+	if (!state || !IS_ENABLED(CONFIG_BLOCK))
+		io_iopoll_iocb_add_list(kiocb);
+	else
+		io_iopoll_iocb_add_state(state, kiocb);
+}
+
 static ssize_t io_read(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb)
 {
 	struct iovec inline_vecs[UIO_FASTIOV], *iovec = inline_vecs;
@@ -632,7 +682,8 @@ static int io_fsync(struct fsync_iocb *req, const struct io_uring_iocb *iocb,
 
 static int __io_submit_one(struct io_ring_ctx *ctx,
 			   const struct io_uring_iocb *iocb,
-			   unsigned long ki_index)
+			   unsigned long ki_index,
+			   struct io_submit_state *state)
 {
 	struct io_kiocb *req;
 	ssize_t ret;
@@ -684,7 +735,7 @@ static int __io_submit_one(struct io_ring_ctx *ctx,
 			ret = -EAGAIN;
 			goto out_put_req;
 		}
-		io_iopoll_iocb_issued(req);
+		io_iopoll_iocb_issued(state, req);
 	}
 	return 0;
 out_put_req:
@@ -692,6 +743,43 @@ static int __io_submit_one(struct io_ring_ctx *ctx,
 	return ret;
 }
 
+#ifdef CONFIG_BLOCK
+static void io_state_unplug(struct blk_plug_cb *cb, bool from_schedule)
+{
+	struct io_submit_state *state;
+
+	state = container_of(cb, struct io_submit_state, plug_cb);
+	if (!list_empty(&state->req_list))
+		io_flush_state_reqs(state->ctx, state);
+}
+#endif
+
+/*
+ * Batched submission is done, ensure local IO is flushed out.
+ */
+static void io_submit_state_end(struct io_submit_state *state)
+{
+	blk_finish_plug(&state->plug);
+	if (!list_empty(&state->req_list))
+		io_flush_state_reqs(state->ctx, state);
+}
+
+/*
+ * Start submission side cache.
+ */
+static void io_submit_state_start(struct io_submit_state *state,
+				  struct io_ring_ctx *ctx)
+{
+	state->ctx = ctx;
+	INIT_LIST_HEAD(&state->req_list);
+	state->req_count = 0;
+#ifdef CONFIG_BLOCK
+	state->plug_cb.callback = io_state_unplug;
+	blk_start_plug(&state->plug);
+	list_add(&state->plug_cb.list, &state->plug.cb_list);
+#endif
+}
+
 static void io_inc_sqring(struct io_ring_ctx *ctx)
 {
 	struct io_sq_ring *ring = ctx->sq_ring.ring;
@@ -726,11 +814,13 @@ static const struct io_uring_iocb *io_peek_sqring(struct io_ring_ctx *ctx,
 
 static int io_ring_submit(struct io_ring_ctx *ctx, unsigned int to_submit)
 {
+	struct io_submit_state state, *statep = NULL;
 	int i, ret = 0, submit = 0;
-	struct blk_plug plug;
 
-	if (to_submit > IO_PLUG_THRESHOLD)
-		blk_start_plug(&plug);
+	if (to_submit > IO_PLUG_THRESHOLD) {
+		io_submit_state_start(&state, ctx);
+		statep = &state;
+	}
 
 	for (i = 0; i < to_submit; i++) {
 		const struct io_uring_iocb *iocb;
@@ -740,7 +830,7 @@ static int io_ring_submit(struct io_ring_ctx *ctx, unsigned int to_submit)
 		if (!iocb)
 			break;
 
-		ret = __io_submit_one(ctx, iocb, iocb_index);
+		ret = __io_submit_one(ctx, iocb, iocb_index, statep);
 		if (ret)
 			break;
 
@@ -748,8 +838,8 @@ static int io_ring_submit(struct io_ring_ctx *ctx, unsigned int to_submit)
 		io_inc_sqring(ctx);
 	}
 
-	if (to_submit > IO_PLUG_THRESHOLD)
-		blk_finish_plug(&plug);
+	if (statep)
+		io_submit_state_end(statep);
 
 	return submit ? submit : ret;
 }
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 08/16] fs: add fget_many() and fput_many()
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
                   ` (6 preceding siblings ...)
  2019-01-08 16:56 ` [PATCH 07/16] io_uring: add submission side request cache Jens Axboe
@ 2019-01-08 16:56 ` Jens Axboe
  2019-01-08 16:56 ` [PATCH 09/16] io_uring: use fget/fput_many() for file references Jens Axboe
                   ` (8 subsequent siblings)
  16 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

Some uses cases repeatedly get and put references to the same file, but
the only exposed interface is doing these one at the time. As each of
these entail an atomic inc or dec on a shared structure, that cost can
add up.

Add fget_many(), which works just like fget(), except it takes an
argument for how many references to get on the file. Ditto fput_many(),
which can drop an arbitrary number of references to a file.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/file.c            | 15 ++++++++++-----
 fs/file_table.c      |  9 +++++++--
 include/linux/file.h |  2 ++
 include/linux/fs.h   |  4 +++-
 4 files changed, 22 insertions(+), 8 deletions(-)

diff --git a/fs/file.c b/fs/file.c
index 3209ee271c41..e0d7ce70e860 100644
--- a/fs/file.c
+++ b/fs/file.c
@@ -705,7 +705,7 @@ void do_close_on_exec(struct files_struct *files)
 	spin_unlock(&files->file_lock);
 }
 
-static struct file *__fget(unsigned int fd, fmode_t mask)
+static struct file *__fget(unsigned int fd, fmode_t mask, unsigned int refs)
 {
 	struct files_struct *files = current->files;
 	struct file *file;
@@ -720,7 +720,7 @@ static struct file *__fget(unsigned int fd, fmode_t mask)
 		 */
 		if (file->f_mode & mask)
 			file = NULL;
-		else if (!get_file_rcu(file))
+		else if (!get_file_rcu_many(file, refs))
 			goto loop;
 	}
 	rcu_read_unlock();
@@ -728,15 +728,20 @@ static struct file *__fget(unsigned int fd, fmode_t mask)
 	return file;
 }
 
+struct file *fget_many(unsigned int fd, unsigned int refs)
+{
+	return __fget(fd, FMODE_PATH, refs);
+}
+
 struct file *fget(unsigned int fd)
 {
-	return __fget(fd, FMODE_PATH);
+	return fget_many(fd, 1);
 }
 EXPORT_SYMBOL(fget);
 
 struct file *fget_raw(unsigned int fd)
 {
-	return __fget(fd, 0);
+	return __fget(fd, 0, 1);
 }
 EXPORT_SYMBOL(fget_raw);
 
@@ -767,7 +772,7 @@ static unsigned long __fget_light(unsigned int fd, fmode_t mask)
 			return 0;
 		return (unsigned long)file;
 	} else {
-		file = __fget(fd, mask);
+		file = __fget(fd, mask, 1);
 		if (!file)
 			return 0;
 		return FDPUT_FPUT | (unsigned long)file;
diff --git a/fs/file_table.c b/fs/file_table.c
index 5679e7fcb6b0..155d7514a094 100644
--- a/fs/file_table.c
+++ b/fs/file_table.c
@@ -326,9 +326,9 @@ void flush_delayed_fput(void)
 
 static DECLARE_DELAYED_WORK(delayed_fput_work, delayed_fput);
 
-void fput(struct file *file)
+void fput_many(struct file *file, unsigned int refs)
 {
-	if (atomic_long_dec_and_test(&file->f_count)) {
+	if (atomic_long_sub_and_test(refs, &file->f_count)) {
 		struct task_struct *task = current;
 
 		if (likely(!in_interrupt() && !(task->flags & PF_KTHREAD))) {
@@ -347,6 +347,11 @@ void fput(struct file *file)
 	}
 }
 
+void fput(struct file *file)
+{
+	fput_many(file, 1);
+}
+
 /*
  * synchronous analog of fput(); for kernel threads that might be needed
  * in some umount() (and thus can't use flush_delayed_fput() without
diff --git a/include/linux/file.h b/include/linux/file.h
index 6b2fb032416c..3fcddff56bc4 100644
--- a/include/linux/file.h
+++ b/include/linux/file.h
@@ -13,6 +13,7 @@
 struct file;
 
 extern void fput(struct file *);
+extern void fput_many(struct file *, unsigned int);
 
 struct file_operations;
 struct vfsmount;
@@ -44,6 +45,7 @@ static inline void fdput(struct fd fd)
 }
 
 extern struct file *fget(unsigned int fd);
+extern struct file *fget_many(unsigned int fd, unsigned int refs);
 extern struct file *fget_raw(unsigned int fd);
 extern unsigned long __fdget(unsigned int fd);
 extern unsigned long __fdget_raw(unsigned int fd);
diff --git a/include/linux/fs.h b/include/linux/fs.h
index ccb0b7a63aa5..acaad78b6781 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -952,7 +952,9 @@ static inline struct file *get_file(struct file *f)
 	atomic_long_inc(&f->f_count);
 	return f;
 }
-#define get_file_rcu(x) atomic_long_inc_not_zero(&(x)->f_count)
+#define get_file_rcu_many(x, cnt)	\
+	atomic_long_add_unless(&(x)->f_count, (cnt), 0)
+#define get_file_rcu(x) get_file_rcu_many((x), 1)
 #define fput_atomic(x)	atomic_long_add_unless(&(x)->f_count, -1, 1)
 #define file_count(x)	atomic_long_read(&(x)->f_count)
 
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 09/16] io_uring: use fget/fput_many() for file references
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
                   ` (7 preceding siblings ...)
  2019-01-08 16:56 ` [PATCH 08/16] fs: add fget_many() and fput_many() Jens Axboe
@ 2019-01-08 16:56 ` Jens Axboe
  2019-01-08 16:56 ` [PATCH 10/16] io_uring: split kiocb init from allocation Jens Axboe
                   ` (7 subsequent siblings)
  16 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

On the submission side, add file reference batching to the
io_submit_state. We get as many references as the number of iocbs we
are submitting, and drop unused ones if we end up switching files. The
assumption here is that we're usually only dealing with one fd, and if
there are multiple, hopefuly they are at least somewhat ordered. Could
trivially be extended to cover multiple fds, if needed.

On the completion side we do the same thing, except this is trivially
done just locally in io_iopoll_reap().

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/io_uring.c | 105 +++++++++++++++++++++++++++++++++++++++++++-------
 1 file changed, 92 insertions(+), 13 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 9f36eb728208..afbaebb63012 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -134,6 +134,15 @@ struct io_submit_state {
 	 */
 	struct list_head req_list;
 	unsigned int req_count;
+
+	/*
+	 * File reference cache
+	 */
+	struct file *file;
+	unsigned int fd;
+	unsigned int has_refs;
+	unsigned int used_refs;
+	unsigned int ios_left;
 };
 
 static struct kmem_cache *kiocb_cachep, *ioctx_cachep;
@@ -237,7 +246,8 @@ static void io_iopoll_reap(struct io_ring_ctx *ctx, unsigned int *nr_events)
 {
 	void *iocbs[IO_IOPOLL_BATCH];
 	struct io_kiocb *iocb, *n;
-	int to_free = 0;
+	int file_count, to_free = 0;
+	struct file *file = NULL;
 
 	list_for_each_entry_safe(iocb, n, &ctx->poll_completing, ki_list) {
 		if (!test_bit(KIOCB_F_IOPOLL_COMPLETED, &iocb->ki_flags))
@@ -248,10 +258,27 @@ static void io_iopoll_reap(struct io_ring_ctx *ctx, unsigned int *nr_events)
 		list_del(&iocb->ki_list);
 		iocbs[to_free++] = iocb;
 
-		fput(iocb->rw.ki_filp);
+		/*
+		 * Batched puts of the same file, to avoid dirtying the
+		 * file usage count multiple times, if avoidable.
+		 */
+		if (!file) {
+			file = iocb->rw.ki_filp;
+			file_count = 1;
+		} else if (file == iocb->rw.ki_filp) {
+			file_count++;
+		} else {
+			fput_many(file, file_count);
+			file = iocb->rw.ki_filp;
+			file_count = 1;
+		}
+
 		(*nr_events)++;
 	}
 
+	if (file)
+		fput_many(file, file_count);
+
 	if (to_free)
 		iocb_put_many(ctx, iocbs, &to_free);
 }
@@ -433,13 +460,60 @@ static void io_complete_scqring_iopoll(struct kiocb *kiocb, long res, long res2)
 	}
 }
 
-static int io_prep_rw(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb)
+static void io_file_put(struct io_submit_state *state, struct file *file)
+{
+	if (!state) {
+		fput(file);
+	} else if (state->file) {
+		int diff = state->has_refs - state->used_refs;
+
+		if (diff)
+			fput_many(state->file, diff);
+		state->file = NULL;
+	}
+}
+
+/*
+ * Get as many references to a file as we have IOs left in this submission,
+ * assuming most submissions are for one file, or at least that each file
+ * has more than one submission.
+ */
+static struct file *io_file_get(struct io_submit_state *state, int fd)
+{
+	if (!state)
+		return fget(fd);
+
+	if (!state->file) {
+get_file:
+		state->file = fget_many(fd, state->ios_left);
+		if (!state->file)
+			return NULL;
+
+		state->fd = fd;
+		state->has_refs = state->ios_left;
+		state->used_refs = 1;
+		state->ios_left--;
+		return state->file;
+	}
+
+	if (state->fd == fd) {
+		state->used_refs++;
+		state->ios_left--;
+		return state->file;
+	}
+
+	io_file_put(state, NULL);
+	goto get_file;
+}
+
+static int io_prep_rw(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb,
+		      struct io_submit_state *state)
 {
 	struct io_ring_ctx *ctx = kiocb->ki_ctx;
 	struct kiocb *req = &kiocb->rw;
 	int ret;
 
-	req->ki_filp = fget(iocb->fd);
+	req->ki_filp = io_file_get(state, iocb->fd);
 	if (unlikely(!req->ki_filp))
 		return -EBADF;
 	req->ki_pos = iocb->off;
@@ -473,7 +547,7 @@ static int io_prep_rw(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb)
 	}
 	return 0;
 out_fput:
-	fput(req->ki_filp);
+	io_file_put(state, req->ki_filp);
 	return ret;
 }
 
@@ -567,7 +641,8 @@ static void io_iopoll_iocb_issued(struct io_submit_state *state,
 		io_iopoll_iocb_add_state(state, kiocb);
 }
 
-static ssize_t io_read(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb)
+static ssize_t io_read(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb,
+		       struct io_submit_state *state)
 {
 	struct iovec inline_vecs[UIO_FASTIOV], *iovec = inline_vecs;
 	struct kiocb *req = &kiocb->rw;
@@ -575,7 +650,7 @@ static ssize_t io_read(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb)
 	struct file *file;
 	ssize_t ret;
 
-	ret = io_prep_rw(kiocb, iocb);
+	ret = io_prep_rw(kiocb, iocb, state);
 	if (ret)
 		return ret;
 	file = req->ki_filp;
@@ -602,7 +677,8 @@ static ssize_t io_read(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb)
 }
 
 static ssize_t io_write(struct io_kiocb *kiocb,
-			const struct io_uring_iocb *iocb)
+			const struct io_uring_iocb *iocb,
+			struct io_submit_state *state)
 {
 	struct iovec inline_vecs[UIO_FASTIOV], *iovec = inline_vecs;
 	struct kiocb *req = &kiocb->rw;
@@ -610,7 +686,7 @@ static ssize_t io_write(struct io_kiocb *kiocb,
 	struct file *file;
 	ssize_t ret;
 
-	ret = io_prep_rw(kiocb, iocb);
+	ret = io_prep_rw(kiocb, iocb, state);
 	if (ret)
 		return ret;
 	file = req->ki_filp;
@@ -704,10 +780,10 @@ static int __io_submit_one(struct io_ring_ctx *ctx,
 	ret = -EINVAL;
 	switch (iocb->opcode) {
 	case IORING_OP_READ:
-		ret = io_read(req, iocb);
+		ret = io_read(req, iocb, state);
 		break;
 	case IORING_OP_WRITE:
-		ret = io_write(req, iocb);
+		ret = io_write(req, iocb, state);
 		break;
 	case IORING_OP_FSYNC:
 		if (ctx->flags & IORING_SETUP_IOPOLL)
@@ -762,17 +838,20 @@ static void io_submit_state_end(struct io_submit_state *state)
 	blk_finish_plug(&state->plug);
 	if (!list_empty(&state->req_list))
 		io_flush_state_reqs(state->ctx, state);
+	io_file_put(state, NULL);
 }
 
 /*
  * Start submission side cache.
  */
 static void io_submit_state_start(struct io_submit_state *state,
-				  struct io_ring_ctx *ctx)
+				  struct io_ring_ctx *ctx, unsigned max_ios)
 {
 	state->ctx = ctx;
 	INIT_LIST_HEAD(&state->req_list);
 	state->req_count = 0;
+	state->file = NULL;
+	state->ios_left = max_ios;
 #ifdef CONFIG_BLOCK
 	state->plug_cb.callback = io_state_unplug;
 	blk_start_plug(&state->plug);
@@ -818,7 +897,7 @@ static int io_ring_submit(struct io_ring_ctx *ctx, unsigned int to_submit)
 	int i, ret = 0, submit = 0;
 
 	if (to_submit > IO_PLUG_THRESHOLD) {
-		io_submit_state_start(&state, ctx);
+		io_submit_state_start(&state, ctx, to_submit);
 		statep = &state;
 	}
 
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 10/16] io_uring: split kiocb init from allocation
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
                   ` (8 preceding siblings ...)
  2019-01-08 16:56 ` [PATCH 09/16] io_uring: use fget/fput_many() for file references Jens Axboe
@ 2019-01-08 16:56 ` Jens Axboe
  2019-01-09 12:12   ` Christoph Hellwig
  2019-01-08 16:56 ` [PATCH 11/16] io_uring: batch io_kiocb allocation Jens Axboe
                   ` (6 subsequent siblings)
  16 siblings, 1 reply; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

In preparation from having pre-allocated requests, that we then just
need to initialize before use.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/io_uring.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index afbaebb63012..11d045f0f799 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -202,6 +202,14 @@ static struct io_uring_event *io_peek_cqring(struct io_ring_ctx *ctx)
 	return &ring->events[tail & ctx->cq_ring.ring_mask];
 }
 
+static void io_req_init(struct io_ring_ctx *ctx, struct io_kiocb *req)
+{
+	percpu_ref_get(&ctx->refs);
+	req->ki_ctx = ctx;
+	INIT_LIST_HEAD(&req->ki_list);
+	req->ki_flags = 0;
+}
+
 static struct io_kiocb *io_get_req(struct io_ring_ctx *ctx)
 {
 	struct io_kiocb *req;
@@ -210,10 +218,7 @@ static struct io_kiocb *io_get_req(struct io_ring_ctx *ctx)
 	if (!req)
 		return NULL;
 
-	percpu_ref_get(&ctx->refs);
-	req->ki_ctx = ctx;
-	INIT_LIST_HEAD(&req->ki_list);
-	req->ki_flags = 0;
+	io_req_init(ctx, req);
 	return req;
 }
 
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 11/16] io_uring: batch io_kiocb allocation
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
                   ` (9 preceding siblings ...)
  2019-01-08 16:56 ` [PATCH 10/16] io_uring: split kiocb init from allocation Jens Axboe
@ 2019-01-08 16:56 ` Jens Axboe
  2019-01-09 12:13   ` Christoph Hellwig
  2019-01-08 16:56 ` [PATCH 12/16] block: implement bio helper to add iter bvec pages to bio Jens Axboe
                   ` (5 subsequent siblings)
  16 siblings, 1 reply; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

Similarly to how we use the state->ios_left to know how many references
to get to a file, we can use it to allocate the io_kiocb's we need in
bulk.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/io_uring.c | 41 +++++++++++++++++++++++++++++++++++------
 1 file changed, 35 insertions(+), 6 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 11d045f0f799..62778d7ffb8d 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -135,6 +135,13 @@ struct io_submit_state {
 	struct list_head req_list;
 	unsigned int req_count;
 
+	/*
+	 * io_kiocb alloc cache
+	 */
+	void *iocbs[IO_IOPOLL_BATCH];
+	unsigned int free_iocbs;
+	unsigned int cur_iocb;
+
 	/*
 	 * File reference cache
 	 */
@@ -210,15 +217,33 @@ static void io_req_init(struct io_ring_ctx *ctx, struct io_kiocb *req)
 	req->ki_flags = 0;
 }
 
-static struct io_kiocb *io_get_req(struct io_ring_ctx *ctx)
+static struct io_kiocb *io_get_req(struct io_ring_ctx *ctx,
+				   struct io_submit_state *state)
 {
 	struct io_kiocb *req;
 
-	req = kmem_cache_alloc(kiocb_cachep, GFP_KERNEL);
-	if (!req)
-		return NULL;
+	if (!state)
+		req = kmem_cache_alloc(kiocb_cachep, GFP_KERNEL);
+	else if (!state->free_iocbs) {
+		size_t size;
+		int ret;
+
+		size = min_t(size_t, state->ios_left, ARRAY_SIZE(state->iocbs));
+		ret = kmem_cache_alloc_bulk(kiocb_cachep, GFP_KERNEL, size,
+						state->iocbs);
+		if (ret <= 0)
+			return ERR_PTR(-ENOMEM);
+		state->free_iocbs = ret - 1;
+		state->cur_iocb = 1;
+		req = state->iocbs[0];
+	} else {
+		req = state->iocbs[state->cur_iocb];
+		state->free_iocbs--;
+		state->cur_iocb++;
+	}
 
-	io_req_init(ctx, req);
+	if (req)
+		io_req_init(ctx, req);
 	return req;
 }
 
@@ -773,7 +798,7 @@ static int __io_submit_one(struct io_ring_ctx *ctx,
 	if (unlikely(iocb->flags))
 		return -EINVAL;
 
-	req = io_get_req(ctx);
+	req = io_get_req(ctx, state);
 	if (unlikely(!req))
 		return -EAGAIN;
 
@@ -844,6 +869,9 @@ static void io_submit_state_end(struct io_submit_state *state)
 	if (!list_empty(&state->req_list))
 		io_flush_state_reqs(state->ctx, state);
 	io_file_put(state, NULL);
+	if (state->free_iocbs)
+		kmem_cache_free_bulk(kiocb_cachep, state->free_iocbs,
+					&state->iocbs[state->cur_iocb]);
 }
 
 /*
@@ -855,6 +883,7 @@ static void io_submit_state_start(struct io_submit_state *state,
 	state->ctx = ctx;
 	INIT_LIST_HEAD(&state->req_list);
 	state->req_count = 0;
+	state->free_iocbs = 0;
 	state->file = NULL;
 	state->ios_left = max_ios;
 #ifdef CONFIG_BLOCK
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 12/16] block: implement bio helper to add iter bvec pages to bio
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
                   ` (10 preceding siblings ...)
  2019-01-08 16:56 ` [PATCH 11/16] io_uring: batch io_kiocb allocation Jens Axboe
@ 2019-01-08 16:56 ` Jens Axboe
  2019-01-08 16:56 ` [PATCH 13/16] io_uring: add support for pre-mapped user IO buffers Jens Axboe
                   ` (4 subsequent siblings)
  16 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

For an ITER_BVEC, we can just iterate the iov and add the pages
to the bio directly. This requires that the caller doesn't releases
the pages on IO completion, we add a BIO_HOLD_PAGES flag for that.

The current two callers of bio_iov_iter_get_pages() are updated to
check if they need to release pages on completion. This makes them
work with bvecs that contain kernel mapped pages already.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 block/bio.c               | 59 ++++++++++++++++++++++++++++++++-------
 fs/block_dev.c            |  5 ++--
 fs/iomap.c                |  5 ++--
 include/linux/blk_types.h |  1 +
 4 files changed, 56 insertions(+), 14 deletions(-)

diff --git a/block/bio.c b/block/bio.c
index 4db1008309ed..7af4f45d2ed6 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -828,6 +828,23 @@ int bio_add_page(struct bio *bio, struct page *page,
 }
 EXPORT_SYMBOL(bio_add_page);
 
+static int __bio_iov_bvec_add_pages(struct bio *bio, struct iov_iter *iter)
+{
+	const struct bio_vec *bv = iter->bvec;
+	unsigned int len;
+	size_t size;
+
+	len = min_t(size_t, bv->bv_len, iter->count);
+	size = bio_add_page(bio, bv->bv_page, len,
+				bv->bv_offset + iter->iov_offset);
+	if (size == len) {
+		iov_iter_advance(iter, size);
+		return 0;
+	}
+
+	return -EINVAL;
+}
+
 #define PAGE_PTRS_PER_BVEC     (sizeof(struct bio_vec) / sizeof(struct page *))
 
 /**
@@ -876,23 +893,43 @@ static int __bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter)
 }
 
 /**
- * bio_iov_iter_get_pages - pin user or kernel pages and add them to a bio
+ * bio_iov_iter_get_pages - add user or kernel pages to a bio
  * @bio: bio to add pages to
- * @iter: iov iterator describing the region to be mapped
+ * @iter: iov iterator describing the region to be added
+ *
+ * This takes either an iterator pointing to user memory, or one pointing to
+ * kernel pages (BVEC iterator). If we're adding user pages, we pin them and
+ * map them into the kernel. On IO completion, the caller should put those
+ * pages. If we're adding kernel pages, we just have to add the pages to the
+ * bio directly. We don't grab an extra reference to those pages (the user
+ * should already have that), and we don't put the page on IO completion.
+ * The caller needs to check if the bio is flagged BIO_HOLD_PAGES on IO
+ * completion. If it isn't, then pages should be released.
  *
- * Pins pages from *iter and appends them to @bio's bvec array. The
- * pages will have to be released using put_page() when done.
  * The function tries, but does not guarantee, to pin as many pages as
- * fit into the bio, or are requested in *iter, whatever is smaller.
- * If MM encounters an error pinning the requested pages, it stops.
- * Error is returned only if 0 pages could be pinned.
+ * fit into the bio, or are requested in *iter, whatever is smaller. If
+ * MM encounters an error pinning the requested pages, it stops. Error
+ * is returned only if 0 pages could be pinned.
  */
 int bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter)
 {
+	const bool is_bvec = iov_iter_is_bvec(iter);
 	unsigned short orig_vcnt = bio->bi_vcnt;
 
+	/*
+	 * If this is a BVEC iter, then the pages are kernel pages. Don't
+	 * release them on IO completion.
+	 */
+	if (is_bvec)
+		bio_set_flag(bio, BIO_HOLD_PAGES);
+
 	do {
-		int ret = __bio_iov_iter_get_pages(bio, iter);
+		int ret;
+
+		if (is_bvec)
+			ret = __bio_iov_bvec_add_pages(bio, iter);
+		else
+			ret = __bio_iov_iter_get_pages(bio, iter);
 
 		if (unlikely(ret))
 			return bio->bi_vcnt > orig_vcnt ? 0 : ret;
@@ -1634,7 +1671,8 @@ static void bio_dirty_fn(struct work_struct *work)
 		next = bio->bi_private;
 
 		bio_set_pages_dirty(bio);
-		bio_release_pages(bio);
+		if (!bio_flagged(bio, BIO_HOLD_PAGES))
+			bio_release_pages(bio);
 		bio_put(bio);
 	}
 }
@@ -1650,7 +1688,8 @@ void bio_check_pages_dirty(struct bio *bio)
 			goto defer;
 	}
 
-	bio_release_pages(bio);
+	if (!bio_flagged(bio, BIO_HOLD_PAGES))
+		bio_release_pages(bio);
 	bio_put(bio);
 	return;
 defer:
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 2ebd2a0d7789..b7742014c9de 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -324,8 +324,9 @@ static void blkdev_bio_end_io(struct bio *bio)
 		struct bio_vec *bvec;
 		int i;
 
-		bio_for_each_segment_all(bvec, bio, i)
-			put_page(bvec->bv_page);
+		if (!bio_flagged(bio, BIO_HOLD_PAGES))
+			bio_for_each_segment_all(bvec, bio, i)
+				put_page(bvec->bv_page);
 		bio_put(bio);
 	}
 }
diff --git a/fs/iomap.c b/fs/iomap.c
index 4ee50b76b4a1..0a64c9c51203 100644
--- a/fs/iomap.c
+++ b/fs/iomap.c
@@ -1582,8 +1582,9 @@ static void iomap_dio_bio_end_io(struct bio *bio)
 		struct bio_vec *bvec;
 		int i;
 
-		bio_for_each_segment_all(bvec, bio, i)
-			put_page(bvec->bv_page);
+		if (!bio_flagged(bio, BIO_HOLD_PAGES))
+			bio_for_each_segment_all(bvec, bio, i)
+				put_page(bvec->bv_page);
 		bio_put(bio);
 	}
 }
diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
index 5c7e7f859a24..97e206855cd3 100644
--- a/include/linux/blk_types.h
+++ b/include/linux/blk_types.h
@@ -215,6 +215,7 @@ struct bio {
 /*
  * bio flags
  */
+#define BIO_HOLD_PAGES	0	/* don't put O_DIRECT pages */
 #define BIO_SEG_VALID	1	/* bi_phys_segments valid */
 #define BIO_CLONED	2	/* doesn't own data */
 #define BIO_BOUNCED	3	/* bio is a bounce bio */
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 13/16] io_uring: add support for pre-mapped user IO buffers
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
                   ` (11 preceding siblings ...)
  2019-01-08 16:56 ` [PATCH 12/16] block: implement bio helper to add iter bvec pages to bio Jens Axboe
@ 2019-01-08 16:56 ` Jens Axboe
  2019-01-09 12:16   ` Christoph Hellwig
  2019-01-08 16:56 ` [PATCH 14/16] io_uring: support kernel side submission Jens Axboe
                   ` (3 subsequent siblings)
  16 siblings, 1 reply; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

If we have fixed user buffers, we can map them into the kernel when we
setup the io_context. That avoids the need to do get_user_pages() for
each and every IO.

To utilize this feature, the application must set
IORING_SETUP_FIXEDBUFS and pass in an array of iovecs that contain the
desired buffer addresses and lengths. These buffers can then be mapped
into the kernel for the life time of the io_uring, as opposed to just
the duration of the each single IO. The application can then use the
IORING_OP_{READ,WRITE}_FIXED to perform IO to these fixed locations.

It's perfectly valid to setup a larger buffer, and then sometimes only
use parts of it for an IO. As long as the range is within the originally
mapped region, it will work just fine.

A limit of 4M is imposed as the largest buffer we currently support.
There's nothing preventing us from going larger, but we need some cap,
and 4M seemed like it would definitely be big enough. RLIMIT_MEMLOCK
is used to cap the total amount of memory pinned.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/io_uring.c                 | 212 +++++++++++++++++++++++++++++++---
 include/uapi/linux/io_uring.h |   3 +
 2 files changed, 201 insertions(+), 14 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 62778d7ffb8d..92129f867824 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -22,6 +22,8 @@
 #include <linux/workqueue.h>
 #include <linux/blkdev.h>
 #include <linux/anon_inodes.h>
+#include <linux/sizes.h>
+#include <linux/nospec.h>
 
 #include <linux/uaccess.h>
 #include <linux/nospec.h>
@@ -65,6 +67,13 @@ struct io_event_ring {
 	unsigned		ring_mask;
 };
 
+struct io_mapped_ubuf {
+	u64		ubuf;
+	size_t		len;
+	struct		bio_vec *bvec;
+	unsigned int	nr_bvecs;
+};
+
 struct io_ring_ctx {
 	struct percpu_ref	refs;
 
@@ -74,6 +83,9 @@ struct io_ring_ctx {
 	struct io_iocb_ring	sq_ring;
 	struct io_event_ring	cq_ring;
 
+	/* if used, fixed mapped user buffers */
+	struct io_mapped_ubuf	*user_bufs;
+
 	struct work_struct	work;
 
 	/* iopoll submission state */
@@ -581,13 +593,45 @@ static int io_prep_rw(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb,
 	return ret;
 }
 
-static int io_setup_rw(int rw, const struct io_uring_iocb *iocb,
-		       struct iovec **iovec, struct iov_iter *iter)
+static int io_setup_rw(int rw, struct io_kiocb *kiocb,
+		       const struct io_uring_iocb *iocb, struct iovec **iovec,
+		       struct iov_iter *iter, bool kaddr)
 {
 	void __user *buf = (void __user *)(uintptr_t)iocb->addr;
 	size_t ret;
 
-	ret = import_single_range(rw, buf, iocb->len, *iovec, iter);
+	if (!kaddr) {
+		ret = import_single_range(rw, buf, iocb->len, *iovec, iter);
+	} else {
+		struct io_ring_ctx *ctx = kiocb->ki_ctx;
+		struct io_mapped_ubuf *imu;
+		size_t len = iocb->len;
+		size_t offset;
+		int index;
+
+		/* __io_submit_one() already validated the index */
+		index = array_index_nospec(kiocb->ki_index,
+						ctx->max_reqs);
+		imu = &ctx->user_bufs[index];
+		if ((unsigned long) iocb->addr < imu->ubuf ||
+		    (unsigned long) iocb->addr + len > imu->ubuf + imu->len) {
+			ret = -EFAULT;
+			goto err;
+		}
+
+		/*
+		 * May not be a start of buffer, set size appropriately
+		 * and advance us to the beginning.
+		 */
+		offset = (unsigned long) iocb->addr - imu->ubuf;
+		iov_iter_bvec(iter, rw, imu->bvec, imu->nr_bvecs,
+				offset + len);
+		if (offset)
+			iov_iter_advance(iter, offset);
+		ret = 0;
+
+	}
+err:
 	*iovec = NULL;
 	return ret;
 }
@@ -672,7 +716,7 @@ static void io_iopoll_iocb_issued(struct io_submit_state *state,
 }
 
 static ssize_t io_read(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb,
-		       struct io_submit_state *state)
+		       struct io_submit_state *state, bool kaddr)
 {
 	struct iovec inline_vecs[UIO_FASTIOV], *iovec = inline_vecs;
 	struct kiocb *req = &kiocb->rw;
@@ -692,7 +736,7 @@ static ssize_t io_read(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb,
 	if (unlikely(!file->f_op->read_iter))
 		goto out_fput;
 
-	ret = io_setup_rw(READ, iocb, &iovec, &iter);
+	ret = io_setup_rw(READ, kiocb, iocb, &iovec, &iter, kaddr);
 	if (ret)
 		goto out_fput;
 
@@ -708,7 +752,7 @@ static ssize_t io_read(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb,
 
 static ssize_t io_write(struct io_kiocb *kiocb,
 			const struct io_uring_iocb *iocb,
-			struct io_submit_state *state)
+			struct io_submit_state *state, bool kaddr)
 {
 	struct iovec inline_vecs[UIO_FASTIOV], *iovec = inline_vecs;
 	struct kiocb *req = &kiocb->rw;
@@ -728,7 +772,7 @@ static ssize_t io_write(struct io_kiocb *kiocb,
 	if (unlikely(!file->f_op->write_iter))
 		goto out_fput;
 
-	ret = io_setup_rw(WRITE, iocb, &iovec, &iter);
+	ret = io_setup_rw(WRITE, kiocb, iocb, &iovec, &iter, kaddr);
 	if (ret)
 		goto out_fput;
 	ret = rw_verify_area(WRITE, file, &req->ki_pos, iov_iter_count(&iter));
@@ -810,10 +854,16 @@ static int __io_submit_one(struct io_ring_ctx *ctx,
 	ret = -EINVAL;
 	switch (iocb->opcode) {
 	case IORING_OP_READ:
-		ret = io_read(req, iocb, state);
+		ret = io_read(req, iocb, state, false);
+		break;
+	case IORING_OP_READ_FIXED:
+		ret = io_read(req, iocb, state, true);
 		break;
 	case IORING_OP_WRITE:
-		ret = io_write(req, iocb, state);
+		ret = io_write(req, iocb, state, false);
+		break;
+	case IORING_OP_WRITE_FIXED:
+		ret = io_write(req, iocb, state, true);
 		break;
 	case IORING_OP_FSYNC:
 		if (ctx->flags & IORING_SETUP_IOPOLL)
@@ -1021,6 +1071,127 @@ static int __io_uring_enter(struct io_ring_ctx *ctx, unsigned to_submit,
 	return ret;
 }
 
+static void io_iocb_buffer_unmap(struct io_ring_ctx *ctx)
+{
+	int i, j;
+
+	if (!ctx->user_bufs)
+		return;
+
+	for (i = 0; i < ctx->max_reqs; i++) {
+		struct io_mapped_ubuf *imu = &ctx->user_bufs[i];
+
+		for (j = 0; j < imu->nr_bvecs; j++)
+			put_page(imu->bvec[j].bv_page);
+
+		kfree(imu->bvec);
+		imu->nr_bvecs = 0;
+	}
+
+	kfree(ctx->user_bufs);
+	ctx->user_bufs = NULL;
+}
+
+static int io_iocb_buffer_map(struct io_ring_ctx *ctx,
+			      struct iovec __user *iovecs)
+{
+	unsigned long total_pages, page_limit;
+	struct page **pages = NULL;
+	int i, j, got_pages = 0;
+	int ret = -EINVAL;
+
+	ctx->user_bufs = kcalloc(ctx->max_reqs, sizeof(struct io_mapped_ubuf),
+					GFP_KERNEL);
+	if (!ctx->user_bufs)
+		return -ENOMEM;
+
+	/* Don't allow more pages than we can safely lock */
+	total_pages = 0;
+	page_limit = rlimit(RLIMIT_MEMLOCK) >> PAGE_SHIFT;
+
+	for (i = 0; i < ctx->max_reqs; i++) {
+		struct io_mapped_ubuf *imu = &ctx->user_bufs[i];
+		unsigned long off, start, end, ubuf;
+		int pret, nr_pages;
+		struct iovec iov;
+		size_t size;
+
+		ret = -EFAULT;
+		if (copy_from_user(&iov, &iovecs[i], sizeof(iov)))
+			goto err;
+
+		/*
+		 * Don't impose further limits on the size and buffer
+		 * constraints here, we'll -EINVAL later when IO is
+		 * submitted if they are wrong.
+		 */
+		ret = -EFAULT;
+		if (!iov.iov_base)
+			goto err;
+
+		/* arbitrary limit, but we need something */
+		if (iov.iov_len > SZ_4M)
+			goto err;
+
+		ubuf = (unsigned long) iov.iov_base;
+		end = (ubuf + iov.iov_len + PAGE_SIZE - 1) >> PAGE_SHIFT;
+		start = ubuf >> PAGE_SHIFT;
+		nr_pages = end - start;
+
+		ret = -ENOMEM;
+		if (total_pages + nr_pages > page_limit)
+			goto err;
+
+		if (!pages || nr_pages > got_pages) {
+			kfree(pages);
+			pages = kmalloc(nr_pages * sizeof(struct page *),
+					GFP_KERNEL);
+			if (!pages)
+				goto err;
+			got_pages = nr_pages;
+		}
+
+		imu->bvec = kmalloc(nr_pages * sizeof(struct bio_vec),
+					GFP_KERNEL);
+		if (!imu->bvec)
+			goto err;
+
+		down_write(&current->mm->mmap_sem);
+		pret = get_user_pages(ubuf, nr_pages, 1, pages, NULL);
+		up_write(&current->mm->mmap_sem);
+
+		if (pret < nr_pages) {
+			if (pret < 0)
+				ret = pret;
+			goto err;
+		}
+
+		off = ubuf & ~PAGE_MASK;
+		size = iov.iov_len;
+		for (j = 0; j < nr_pages; j++) {
+			size_t vec_len;
+
+			vec_len = min_t(size_t, size, PAGE_SIZE - off);
+			imu->bvec[j].bv_page = pages[j];
+			imu->bvec[j].bv_len = vec_len;
+			imu->bvec[j].bv_offset = off;
+			off = 0;
+			size -= vec_len;
+		}
+		/* store original address for later verification */
+		imu->ubuf = ubuf;
+		imu->len = iov.iov_len;
+		imu->nr_bvecs = nr_pages;
+		total_pages += nr_pages;
+	}
+	kfree(pages);
+	return 0;
+err:
+	kfree(pages);
+	io_iocb_buffer_unmap(ctx);
+	return ret;
+}
+
 static void io_free_scq_urings(struct io_ring_ctx *ctx)
 {
 	if (ctx->sq_ring.ring) {
@@ -1043,6 +1214,7 @@ static void io_ring_ctx_free(struct work_struct *work)
 
 	io_iopoll_reap_events(ctx);
 	io_free_scq_urings(ctx);
+	io_iocb_buffer_unmap(ctx);
 	percpu_ref_exit(&ctx->refs);
 	kmem_cache_free(ioctx_cachep, ctx);
 }
@@ -1191,11 +1363,19 @@ static void io_fill_offsets(struct io_uring_params *p)
 	p->cq_off.events = offsetof(struct io_cq_ring, events);
 }
 
-static int io_uring_create(unsigned entries, struct io_uring_params *p)
+static int io_uring_create(unsigned entries, struct io_uring_params *p,
+			   struct iovec __user *iovecs)
 {
 	struct io_ring_ctx *ctx;
 	int ret;
 
+	/*
+	 * We don't use the iovecs without fixed buffers being asked for.
+	 * Error out if they don't match.
+	 */
+	if (!(p->flags & IORING_SETUP_FIXEDBUFS) && iovecs)
+		return -EINVAL;
+
 	/*
 	 * Use twice as many entries for the CQ ring. It's possible for the
 	 * application to drive a higher depth than the size of the SQ ring,
@@ -1213,6 +1393,12 @@ static int io_uring_create(unsigned entries, struct io_uring_params *p)
 	if (ret)
 		goto err;
 
+	if (p->flags & IORING_SETUP_FIXEDBUFS) {
+		ret = io_iocb_buffer_map(ctx, iovecs);
+		if (ret)
+			goto err;
+	}
+
 	ret = anon_inode_getfd("[io_uring]", &io_scqring_fops, ctx,
 				O_RDWR | O_CLOEXEC);
 	if (ret < 0)
@@ -1245,12 +1431,10 @@ SYSCALL_DEFINE3(io_uring_setup, u32, entries, struct iovec __user *, iovecs,
 			return -EINVAL;
 	}
 
-	if (p.flags & ~IORING_SETUP_IOPOLL)
-		return -EINVAL;
-	if (iovecs)
+	if (p.flags & ~(IORING_SETUP_IOPOLL | IORING_SETUP_FIXEDBUFS))
 		return -EINVAL;
 
-	ret = io_uring_create(entries, &p);
+	ret = io_uring_create(entries, &p, iovecs);
 	if (ret < 0)
 		return ret;
 
diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
index f7ba30747816..925fd6ca3f38 100644
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -35,11 +35,14 @@ struct io_uring_iocb {
  * io_uring_setup() flags
  */
 #define IORING_SETUP_IOPOLL	(1 << 0)	/* io_context is polled */
+#define IORING_SETUP_FIXEDBUFS	(1 << 1)	/* IO buffers are fixed */
 
 #define IORING_OP_READ		1
 #define IORING_OP_WRITE		2
 #define IORING_OP_FSYNC		3
 #define IORING_OP_FDSYNC	4
+#define IORING_OP_READ_FIXED	5
+#define IORING_OP_WRITE_FIXED	6
 
 /*
  * IO completion data structure
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 14/16] io_uring: support kernel side submission
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
                   ` (12 preceding siblings ...)
  2019-01-08 16:56 ` [PATCH 13/16] io_uring: add support for pre-mapped user IO buffers Jens Axboe
@ 2019-01-08 16:56 ` Jens Axboe
  2019-01-09 19:06   ` Christoph Hellwig
  2019-01-08 16:56 ` [PATCH 15/16] io_uring: add submission polling Jens Axboe
                   ` (2 subsequent siblings)
  16 siblings, 1 reply; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

Add support for backing the io_uring fd with either a thread, or a
workqueue and letting those handle the submission for us. This can
be used to reduce overhead for submission, or to always make submission
async. The latter is particularly useful for buffered aio, which is
now fully async with this feature.

For polled IO, we could have the kernel side thread hammer on the SQ
ring and submit when it finds IO. This would mean that an application
would NEVER have to enter the kernel to do IO! Didn't add this yet,
but it would be trivial to add.

If an application sets IORING_SETUP_SCQTHREAD, the io_uring gets a
single thread backing. If used with buffered IO, this will limit the
device queue depth to 1, but it will be async, IOs will simply be
serialized.

Or an application can set IORING_SETUP_SQWQ, in which case the urings
get a work queue backing. The concurrency level is the mininum of twice
the available CPUs, or the queue depth specific for the context. For
this mode, we attempt to do buffered reads inline, in case they are
cached. So we should only punt to a workqueue, if we would have to block
to get our data.

Tested with polling, no polling, fixedbufs, no fixedbufs, buffered,
O_DIRECT.

See this sample application for how to use it:

http://git.kernel.dk/cgit/fio/plain/t/io_uring.c

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/io_uring.c                 | 405 ++++++++++++++++++++++++++++++++--
 include/uapi/linux/io_uring.h |   5 +-
 2 files changed, 387 insertions(+), 23 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 92129f867824..e6a808a89b78 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -14,6 +14,7 @@
 #include <linux/sched/signal.h>
 #include <linux/fs.h>
 #include <linux/file.h>
+#include <linux/fdtable.h>
 #include <linux/mm.h>
 #include <linux/mman.h>
 #include <linux/mmu_context.h>
@@ -24,6 +25,8 @@
 #include <linux/anon_inodes.h>
 #include <linux/sizes.h>
 #include <linux/nospec.h>
+#include <linux/kthread.h>
+#include <linux/sched/mm.h>
 
 #include <linux/uaccess.h>
 #include <linux/nospec.h>
@@ -58,6 +61,7 @@ struct io_iocb_ring {
 	struct			io_sq_ring *ring;
 	unsigned		entries;
 	unsigned		ring_mask;
+	unsigned		sq_thread_cpu;
 	struct io_uring_iocb	*iocbs;
 };
 
@@ -74,6 +78,14 @@ struct io_mapped_ubuf {
 	unsigned int	nr_bvecs;
 };
 
+struct io_sq_offload {
+	struct task_struct	*thread;	/* if using a thread */
+	struct workqueue_struct	*wq;	/* wq offload */
+	struct mm_struct	*mm;
+	struct files_struct	*files;
+	wait_queue_head_t	wait;
+};
+
 struct io_ring_ctx {
 	struct percpu_ref	refs;
 
@@ -88,6 +100,9 @@ struct io_ring_ctx {
 
 	struct work_struct	work;
 
+	/* sq ring submitter thread, if used */
+	struct io_sq_offload	sq_offload;
+
 	/* iopoll submission state */
 	struct {
 		spinlock_t poll_lock;
@@ -127,6 +142,7 @@ struct io_kiocb {
 	unsigned long		ki_flags;
 #define KIOCB_F_IOPOLL_COMPLETED	0	/* polled IO has completed */
 #define KIOCB_F_IOPOLL_EAGAIN		1	/* submission got EAGAIN */
+#define KIOCB_F_FORCE_NONBLOCK		2	/* inline submission attempt */
 };
 
 #define IO_PLUG_THRESHOLD		2
@@ -164,6 +180,18 @@ struct io_submit_state {
 	unsigned int ios_left;
 };
 
+struct iocb_submit {
+	const struct io_uring_iocb *iocb;
+	unsigned int index;
+};
+
+struct io_work {
+	struct work_struct work;
+	struct io_ring_ctx *ctx;
+	struct io_uring_iocb iocb;
+	unsigned iocb_index;
+};
+
 static struct kmem_cache *kiocb_cachep, *ioctx_cachep;
 
 static const struct file_operations io_scqring_fops;
@@ -442,18 +470,17 @@ static void kiocb_end_write(struct kiocb *kiocb)
 	}
 }
 
-static void io_fill_event(struct io_uring_event *ev, struct io_kiocb *kiocb,
+static void io_fill_event(struct io_uring_event *ev, unsigned ki_index,
 			  long res, unsigned flags)
 {
-	ev->index = kiocb->ki_index;
+	ev->index = ki_index;
 	ev->res = res;
 	ev->flags = flags;
 }
 
-static void io_cqring_fill_event(struct io_kiocb *iocb, long res,
-				 unsigned ev_flags)
+static void io_cqring_fill_event(struct io_ring_ctx *ctx, unsigned ki_index,
+				 long res, unsigned ev_flags)
 {
-	struct io_ring_ctx *ctx = iocb->ki_ctx;
 	struct io_uring_event *ev;
 	unsigned long flags;
 
@@ -465,7 +492,7 @@ static void io_cqring_fill_event(struct io_kiocb *iocb, long res,
 	spin_lock_irqsave(&ctx->completion_lock, flags);
 	ev = io_peek_cqring(ctx);
 	if (ev) {
-		io_fill_event(ev, iocb, res, ev_flags);
+		io_fill_event(ev, ki_index, res, ev_flags);
 		io_inc_cqring(ctx);
 	} else
 		ctx->cq_ring.ring->overflow++;
@@ -474,10 +501,24 @@ static void io_cqring_fill_event(struct io_kiocb *iocb, long res,
 
 static void io_complete_scqring(struct io_kiocb *iocb, long res, unsigned flags)
 {
-	io_cqring_fill_event(iocb, res, flags);
+	io_cqring_fill_event(iocb->ki_ctx, iocb->ki_index, res, flags);
 	io_complete_iocb(iocb->ki_ctx, iocb);
 }
 
+static void io_fill_cq_error(struct io_ring_ctx *ctx, unsigned ki_index,
+			     long error)
+{
+	io_cqring_fill_event(ctx, ki_index, error, 0);
+
+	/*
+	 * for thread offload, app could already be sleeping in io_ring_enter()
+	 * before we get to flag the error. wake them up, if needed.
+	 */
+	if (ctx->flags & (IORING_SETUP_SQTHREAD | IORING_SETUP_SQWQ))
+		if (waitqueue_active(&ctx->wait))
+			wake_up(&ctx->wait);
+}
+
 static void io_complete_scqring_rw(struct kiocb *kiocb, long res, long res2)
 {
 	struct io_kiocb *iocb = container_of(kiocb, struct io_kiocb, rw);
@@ -485,6 +526,7 @@ static void io_complete_scqring_rw(struct kiocb *kiocb, long res, long res2)
 	kiocb_end_write(kiocb);
 
 	fput(kiocb->ki_filp);
+
 	io_complete_scqring(iocb, res, 0);
 }
 
@@ -497,7 +539,7 @@ static void io_complete_scqring_iopoll(struct kiocb *kiocb, long res, long res2)
 	if (unlikely(res == -EAGAIN)) {
 		set_bit(KIOCB_F_IOPOLL_EAGAIN, &iocb->ki_flags);
 	} else {
-		io_cqring_fill_event(iocb, res, 0);
+		io_cqring_fill_event(iocb->ki_ctx, iocb->ki_index, res, 0);
 		set_bit(KIOCB_F_IOPOLL_COMPLETED, &iocb->ki_flags);
 	}
 }
@@ -549,7 +591,7 @@ static struct file *io_file_get(struct io_submit_state *state, int fd)
 }
 
 static int io_prep_rw(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb,
-		      struct io_submit_state *state)
+		      struct io_submit_state *state, bool force_nonblock)
 {
 	struct io_ring_ctx *ctx = kiocb->ki_ctx;
 	struct kiocb *req = &kiocb->rw;
@@ -573,6 +615,10 @@ static int io_prep_rw(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb,
 	ret = kiocb_set_rw_flags(req, iocb->rw_flags);
 	if (unlikely(ret))
 		goto out_fput;
+	if (force_nonblock) {
+		req->ki_flags |= IOCB_NOWAIT;
+		set_bit(KIOCB_F_FORCE_NONBLOCK, &kiocb->ki_flags);
+	}
 
 	if (ctx->flags & IORING_SETUP_IOPOLL) {
 		ret = -EOPNOTSUPP;
@@ -716,7 +762,7 @@ static void io_iopoll_iocb_issued(struct io_submit_state *state,
 }
 
 static ssize_t io_read(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb,
-		       struct io_submit_state *state, bool kaddr)
+		       struct io_submit_state *state, bool kaddr, bool nonblock)
 {
 	struct iovec inline_vecs[UIO_FASTIOV], *iovec = inline_vecs;
 	struct kiocb *req = &kiocb->rw;
@@ -724,7 +770,7 @@ static ssize_t io_read(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb,
 	struct file *file;
 	ssize_t ret;
 
-	ret = io_prep_rw(kiocb, iocb, state);
+	ret = io_prep_rw(kiocb, iocb, state, nonblock);
 	if (ret)
 		return ret;
 	file = req->ki_filp;
@@ -741,8 +787,18 @@ static ssize_t io_read(struct io_kiocb *kiocb, const struct io_uring_iocb *iocb,
 		goto out_fput;
 
 	ret = rw_verify_area(READ, file, &req->ki_pos, iov_iter_count(&iter));
-	if (!ret)
-		io_rw_done(req, call_read_iter(file, req, &iter));
+	if (!ret) {
+		ssize_t ret2;
+
+		/*
+		 * Catch -EAGAIN return for forced non-blocking submission
+		 */
+		ret2 = call_read_iter(file, req, &iter);
+		if (!nonblock || ret2 != -EAGAIN)
+			io_rw_done(req, ret2);
+		else
+			ret = -EAGAIN;
+	}
 	kfree(iovec);
 out_fput:
 	if (unlikely(ret))
@@ -760,7 +816,7 @@ static ssize_t io_write(struct io_kiocb *kiocb,
 	struct file *file;
 	ssize_t ret;
 
-	ret = io_prep_rw(kiocb, iocb, state);
+	ret = io_prep_rw(kiocb, iocb, state, false);
 	if (ret)
 		return ret;
 	file = req->ki_filp;
@@ -833,7 +889,7 @@ static int io_fsync(struct fsync_iocb *req, const struct io_uring_iocb *iocb,
 static int __io_submit_one(struct io_ring_ctx *ctx,
 			   const struct io_uring_iocb *iocb,
 			   unsigned long ki_index,
-			   struct io_submit_state *state)
+			   struct io_submit_state *state, bool force_nonblock)
 {
 	struct io_kiocb *req;
 	ssize_t ret;
@@ -854,10 +910,10 @@ static int __io_submit_one(struct io_ring_ctx *ctx,
 	ret = -EINVAL;
 	switch (iocb->opcode) {
 	case IORING_OP_READ:
-		ret = io_read(req, iocb, state, false);
+		ret = io_read(req, iocb, state, false, force_nonblock);
 		break;
 	case IORING_OP_READ_FIXED:
-		ret = io_read(req, iocb, state, true);
+		ret = io_read(req, iocb, state, true, force_nonblock);
 		break;
 	case IORING_OP_WRITE:
 		ret = io_write(req, iocb, state, false);
@@ -993,7 +1049,7 @@ static int io_ring_submit(struct io_ring_ctx *ctx, unsigned int to_submit)
 		if (!iocb)
 			break;
 
-		ret = __io_submit_one(ctx, iocb, iocb_index, statep);
+		ret = __io_submit_one(ctx, iocb, iocb_index, statep, false);
 		if (ret)
 			break;
 
@@ -1042,15 +1098,239 @@ static int io_cqring_wait(struct io_ring_ctx *ctx, int min_events)
 	return ring->r.head == ring->r.tail ? ret : 0;
 }
 
+static int io_submit_iocbs(struct io_ring_ctx *ctx, struct iocb_submit *iocbs,
+			   unsigned int nr, struct mm_struct *cur_mm,
+			   bool mm_fault)
+{
+	struct io_submit_state state, *statep = NULL;
+	int ret, i, submitted = 0;
+
+	if (nr > IO_PLUG_THRESHOLD) {
+		io_submit_state_start(&state, ctx, nr);
+		statep = &state;
+	}
+
+	for (i = 0; i < nr; i++) {
+		if (unlikely(mm_fault))
+			ret = -EFAULT;
+		else
+			ret = __io_submit_one(ctx, iocbs[i].iocb,
+						iocbs[i].index, statep, false);
+		if (!ret) {
+			submitted++;
+			continue;
+		}
+
+		io_fill_cq_error(ctx, iocbs[i].index, ret);
+	}
+
+	if (statep)
+		io_submit_state_end(&state);
+
+	return submitted;
+}
+
+/*
+ * sq thread only supports O_DIRECT or FIXEDBUFS IO
+ */
+static int io_sq_thread(void *data)
+{
+	struct iocb_submit iocbs[IO_IOPOLL_BATCH];
+	struct io_ring_ctx *ctx = data;
+	struct io_sq_offload *sqo = &ctx->sq_offload;
+	struct mm_struct *cur_mm = NULL;
+	struct files_struct *old_files;
+	mm_segment_t old_fs;
+	DEFINE_WAIT(wait);
+
+	old_files = current->files;
+	current->files = sqo->files;
+
+	old_fs = get_fs();
+	set_fs(USER_DS);
+
+	while (!kthread_should_stop()) {
+		const struct io_uring_iocb *iocb;
+		bool mm_fault = false;
+		unsigned iocb_index;
+		int i;
+
+		iocb = io_peek_sqring(ctx, &iocb_index);
+		if (!iocb) {
+			/*
+			 * Drop cur_mm before scheduling, we can't hold it for
+			 * long periods (or over schedule()). Do this before
+			 * adding ourselves to the waitqueue, as the unuse/drop
+			 * may sleep.
+			 */
+			if (cur_mm) {
+				unuse_mm(cur_mm);
+				mmput(cur_mm);
+				cur_mm = NULL;
+			}
+
+			prepare_to_wait(&sqo->wait, &wait, TASK_INTERRUPTIBLE);
+			iocb = io_peek_sqring(ctx, &iocb_index);
+			if (!iocb) {
+				if (kthread_should_park())
+					kthread_parkme();
+				if (kthread_should_stop()) {
+					finish_wait(&sqo->wait, &wait);
+					break;
+				}
+				if (signal_pending(current))
+					flush_signals(current);
+				schedule();
+			}
+			finish_wait(&sqo->wait, &wait);
+			if (!iocb)
+				continue;
+		}
+
+		/* If ->mm is set, we're not doing FIXEDBUFS */
+		if (sqo->mm && !cur_mm) {
+			mm_fault = !mmget_not_zero(sqo->mm);
+			if (!mm_fault) {
+				use_mm(sqo->mm);
+				cur_mm = sqo->mm;
+			}
+		}
+
+		i = 0;
+		do {
+			if (i == ARRAY_SIZE(iocbs))
+				break;
+			iocbs[i].iocb = iocb;
+			iocbs[i].index = iocb_index;
+			++i;
+			io_inc_sqring(ctx);
+		} while ((iocb = io_peek_sqring(ctx, &iocb_index)) != NULL);
+
+		io_submit_iocbs(ctx, iocbs, i, cur_mm, mm_fault);
+	}
+	current->files = old_files;
+	set_fs(old_fs);
+	if (cur_mm) {
+		unuse_mm(cur_mm);
+		mmput(cur_mm);
+	}
+	return 0;
+}
+
+static void io_sq_wq_submit_work(struct work_struct *work)
+{
+	struct io_work *iw = container_of(work, struct io_work, work);
+	struct io_ring_ctx *ctx = iw->ctx;
+	struct io_sq_offload *sqo = &ctx->sq_offload;
+	mm_segment_t old_fs = get_fs();
+	struct files_struct *old_files;
+	int ret;
+
+	old_files = current->files;
+	current->files = sqo->files;
+
+	if (sqo->mm) {
+		if (!mmget_not_zero(sqo->mm)) {
+			ret = -EFAULT;
+			goto err;
+		}
+		use_mm(sqo->mm);
+	}
+
+	set_fs(USER_DS);
+
+	ret = __io_submit_one(ctx, &iw->iocb, iw->iocb_index, NULL, false);
+
+	set_fs(old_fs);
+	if (sqo->mm) {
+		unuse_mm(sqo->mm);
+		mmput(sqo->mm);
+	}
+
+err:
+	if (ret)
+		io_fill_cq_error(ctx, iw->iocb_index, ret);
+	current->files = old_files;
+	kfree(iw);
+}
+
+/*
+ * If this is a read, try a cached inline read first. If the IO is in the
+ * page cache, we can satisfy it without blocking and without having to
+ * punt to a threaded execution. This is much faster, particularly for
+ * lower queue depth IO, and it's always a lot more efficient.
+ */
+static bool io_sq_try_inline(struct io_ring_ctx *ctx,
+			     const struct io_uring_iocb *iocb, unsigned index)
+{
+	int ret;
+
+	if (iocb->opcode != IORING_OP_READ &&
+	    iocb->opcode != IORING_OP_READ_FIXED)
+		return false;
+
+	ret = __io_submit_one(ctx, iocb, index, NULL, true);
+
+	/*
+	 * If we get -EAGAIN, return false to submit out-of-line. Any other
+	 * result and we're done, call will fill in CQ ring event.
+	 */
+	return ret != -EAGAIN;
+}
+
+static int io_sq_wq_submit(struct io_ring_ctx *ctx, unsigned int to_submit)
+{
+	const struct io_uring_iocb *iocb;
+	struct io_work *work;
+	unsigned iocb_index;
+	int ret, queued;
+
+	ret = queued = 0;
+	while ((iocb = io_peek_sqring(ctx, &iocb_index)) != NULL) {
+		ret = io_sq_try_inline(ctx, iocb, iocb_index);
+		if (!ret) {
+			work = kmalloc(sizeof(*work), GFP_KERNEL);
+			if (!work) {
+				ret = -ENOMEM;
+				break;
+			}
+			memcpy(&work->iocb, iocb, sizeof(*iocb));
+			io_inc_sqring(ctx);
+			work->iocb_index = iocb_index;
+			INIT_WORK(&work->work, io_sq_wq_submit_work);
+			work->ctx = ctx;
+			queue_work(ctx->sq_offload.wq, &work->work);
+		}
+		queued++;
+		if (queued == to_submit)
+			break;
+	}
+
+	return queued ? queued : ret;
+}
+
 static int __io_uring_enter(struct io_ring_ctx *ctx, unsigned to_submit,
 			    unsigned min_complete, unsigned flags)
 {
 	int ret = 0;
 
 	if (to_submit) {
-		ret = io_ring_submit(ctx, to_submit);
-		if (ret < 0)
-			return ret;
+		/*
+		 * Three options here:
+		 * 1) We have an sq thread, just wake it up to do submissions
+		 * 2) We have an sq wq, queue a work item for each iocb
+		 * 3) Submit directly
+		 */
+		if (ctx->flags & IORING_SETUP_SQTHREAD) {
+			wake_up(&ctx->sq_offload.wait);
+			ret = to_submit;
+		} else if (ctx->flags & IORING_SETUP_SQWQ) {
+			ret = io_sq_wq_submit(ctx, to_submit);
+		} else {
+			ret = io_ring_submit(ctx, to_submit);
+			if (ret < 0)
+				return ret;
+		}
 	}
 	if (flags & IORING_ENTER_GETEVENTS) {
 		unsigned nr_events = 0;
@@ -1192,6 +1472,78 @@ static int io_iocb_buffer_map(struct io_ring_ctx *ctx,
 	return ret;
 }
 
+static int io_sq_thread(void *);
+
+static int io_sq_thread_start(struct io_ring_ctx *ctx)
+{
+	struct io_sq_offload *sqo = &ctx->sq_offload;
+	struct io_iocb_ring *ring = &ctx->sq_ring;
+	int ret;
+
+	memset(sqo, 0, sizeof(*sqo));
+	init_waitqueue_head(&sqo->wait);
+
+	if (!(ctx->flags & IORING_SETUP_FIXEDBUFS))
+		sqo->mm = current->mm;
+
+	ret = -EBADF;
+	sqo->files = get_files_struct(current);
+	if (!sqo->files)
+		goto err;
+
+	if (ctx->flags & IORING_SETUP_SQTHREAD) {
+		sqo->thread = kthread_create_on_cpu(io_sq_thread, ctx,
+							ring->sq_thread_cpu,
+							"io_uring-sq");
+		if (IS_ERR(sqo->thread)) {
+			ret = PTR_ERR(sqo->thread);
+			sqo->thread = NULL;
+			goto err;
+		}
+		wake_up_process(sqo->thread);
+	} else if (ctx->flags & IORING_SETUP_SQWQ) {
+		int concurrency;
+
+		/* Do QD, or 2 * CPUS, whatever is smallest */
+		concurrency = min(ring->entries - 1, 2 * num_online_cpus());
+		sqo->wq = alloc_workqueue("io_ring-wq",
+						WQ_UNBOUND | WQ_FREEZABLE,
+						concurrency);
+		if (!sqo->wq) {
+			ret = -ENOMEM;
+			goto err;
+		}
+	}
+
+	return 0;
+err:
+	if (sqo->files) {
+		put_files_struct(sqo->files);
+		sqo->files = NULL;
+	}
+	if (sqo->mm)
+		sqo->mm = NULL;
+	return ret;
+}
+
+static void io_sq_thread_stop(struct io_ring_ctx *ctx)
+{
+	struct io_sq_offload *sqo = &ctx->sq_offload;
+
+	if (sqo->thread) {
+		kthread_park(sqo->thread);
+		kthread_stop(sqo->thread);
+		sqo->thread = NULL;
+	} else if (sqo->wq) {
+		destroy_workqueue(sqo->wq);
+		sqo->wq = NULL;
+	}
+	if (sqo->files) {
+		put_files_struct(sqo->files);
+		sqo->files = NULL;
+	}
+}
+
 static void io_free_scq_urings(struct io_ring_ctx *ctx)
 {
 	if (ctx->sq_ring.ring) {
@@ -1212,6 +1564,7 @@ static void io_ring_ctx_free(struct work_struct *work)
 {
 	struct io_ring_ctx *ctx = container_of(work, struct io_ring_ctx, work);
 
+	io_sq_thread_stop(ctx);
 	io_iopoll_reap_events(ctx);
 	io_free_scq_urings(ctx);
 	io_iocb_buffer_unmap(ctx);
@@ -1398,6 +1751,13 @@ static int io_uring_create(unsigned entries, struct io_uring_params *p,
 		if (ret)
 			goto err;
 	}
+	if (p->flags & (IORING_SETUP_SQTHREAD | IORING_SETUP_SQWQ)) {
+		ctx->sq_ring.sq_thread_cpu = p->sq_thread_cpu;
+
+		ret = io_sq_thread_start(ctx);
+		if (ret)
+			goto err;
+	}
 
 	ret = anon_inode_getfd("[io_uring]", &io_scqring_fops, ctx,
 				O_RDWR | O_CLOEXEC);
@@ -1431,7 +1791,8 @@ SYSCALL_DEFINE3(io_uring_setup, u32, entries, struct iovec __user *, iovecs,
 			return -EINVAL;
 	}
 
-	if (p.flags & ~(IORING_SETUP_IOPOLL | IORING_SETUP_FIXEDBUFS))
+	if (p.flags & ~(IORING_SETUP_IOPOLL | IORING_SETUP_FIXEDBUFS |
+			IORING_SETUP_SQTHREAD | IORING_SETUP_SQWQ))
 		return -EINVAL;
 
 	ret = io_uring_create(entries, &p, iovecs);
diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
index 925fd6ca3f38..4f0a8ce49f9a 100644
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -36,6 +36,8 @@ struct io_uring_iocb {
  */
 #define IORING_SETUP_IOPOLL	(1 << 0)	/* io_context is polled */
 #define IORING_SETUP_FIXEDBUFS	(1 << 1)	/* IO buffers are fixed */
+#define IORING_SETUP_SQTHREAD	(1 << 2)	/* Use SQ thread */
+#define IORING_SETUP_SQWQ	(1 << 3)	/* Use SQ workqueue */
 
 #define IORING_OP_READ		1
 #define IORING_OP_WRITE		2
@@ -96,7 +98,8 @@ struct io_uring_params {
 	__u32 sq_entries;
 	__u32 cq_entries;
 	__u32 flags;
-	__u16 resv[10];
+	__u16 sq_thread_cpu;
+	__u16 resv[9];
 	struct io_sqring_offsets sq_off;
 	struct io_cqring_offsets cq_off;
 };
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 15/16] io_uring: add submission polling
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
                   ` (13 preceding siblings ...)
  2019-01-08 16:56 ` [PATCH 14/16] io_uring: support kernel side submission Jens Axboe
@ 2019-01-08 16:56 ` Jens Axboe
  2019-01-08 16:56 ` [PATCH 16/16] io_uring: add io_uring_event cache hit information Jens Axboe
  2019-01-09 16:00 ` [PATCHSET v1] io_uring IO interface Matthew Wilcox
  16 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

This enables an application to do IO, without ever entering the kernel.
By using the SQ ring to fill in new events and watching for completions
on the CQ ring, we can submit and reap IOs without doing a single system
call. The kernel side thread will poll for new submissions, and in case
of HIPRI/polled IO, it'll also poll for completions.

For O_DIRECT, we can do this with just SQTHREAD being enabled. For
buffered aio, we need the workqueue as well. If we can satisfy the
buffered inline from the SQTHREAD, we do that. If not, we punt to the
workqueue. This is just like buffered aio off the io_uring_enter(2)
system call.

Proof of concept. If the thread has been idle for 1 second, it will set
sq_ring->flags |= IORING_SQ_NEED_WAKEUP. The application will have to
call io_uring_enter() to start things back up again. If IO is kept busy,
that will never be needed. Basically an application that has this
feature enabled will guard it's io_uring_enter(2) call with:

barrier();
if (*sq_ring->flags & IORING_SQ_NEED_WAKEUP)
	io_uring_enter(fd, to_submit, 0, 0);

instead of calling it unconditionally.

Improvements:

1) Maybe have smarter backoff. Busy loop for X time, then go to
   monitor/mwait, finally the schedule we have now after an idle
   second. Might not be worth the complexity.

2) Probably want the application to pass in the appropriate grace
   period, not hard code it at 1 second.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/io_uring.c                 | 135 ++++++++++++++++++++++++++++------
 include/uapi/linux/io_uring.h |   3 +
 2 files changed, 115 insertions(+), 23 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index e6a808a89b78..6c10841e4342 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -80,7 +80,8 @@ struct io_mapped_ubuf {
 
 struct io_sq_offload {
 	struct task_struct	*thread;	/* if using a thread */
-	struct workqueue_struct	*wq;	/* wq offload */
+	bool			thread_poll;
+	struct workqueue_struct	*wq;		/* wq offload */
 	struct mm_struct	*mm;
 	struct files_struct	*files;
 	wait_queue_head_t	wait;
@@ -198,6 +199,7 @@ static const struct file_operations io_scqring_fops;
 
 static void io_ring_ctx_free(struct work_struct *work);
 static void io_ring_ctx_ref_free(struct percpu_ref *ref);
+static void io_sq_wq_submit_work(struct work_struct *work);
 
 static struct io_ring_ctx *io_ring_ctx_alloc(struct io_uring_params *p)
 {
@@ -1098,27 +1100,59 @@ static int io_cqring_wait(struct io_ring_ctx *ctx, int min_events)
 	return ring->r.head == ring->r.tail ? ret : 0;
 }
 
+static int io_queue_async_work(struct io_ring_ctx *ctx, struct iocb_submit *is)
+{
+	struct io_work *work;
+
+	work = kmalloc(sizeof(*work), GFP_KERNEL);
+	if (work) {
+		memcpy(&work->iocb, is->iocb, sizeof(*is->iocb));
+		work->iocb_index = is->index;
+		INIT_WORK(&work->work, io_sq_wq_submit_work);
+		work->ctx = ctx;
+		queue_work(ctx->sq_offload.wq, &work->work);
+		return 0;
+	}
+
+	return -ENOMEM;
+}
+
 static int io_submit_iocbs(struct io_ring_ctx *ctx, struct iocb_submit *iocbs,
 			   unsigned int nr, struct mm_struct *cur_mm,
 			   bool mm_fault)
 {
 	struct io_submit_state state, *statep = NULL;
 	int ret, i, submitted = 0;
+	bool force_nonblock;
 
 	if (nr > IO_PLUG_THRESHOLD) {
 		io_submit_state_start(&state, ctx, nr);
 		statep = &state;
 	}
 
+	/*
+	 * Having both a thread and a workqueue only makes sense for buffered
+	 * IO, where we can't submit in an async fashion. Use the NOWAIT
+	 * trick from the SQ thread, and punt to the workqueue if we can't
+	 * satisfy this iocb without blocking. This is only necessary
+	 * for buffered IO with sqthread polled submission.
+	 */
+	force_nonblock = (ctx->flags & IORING_SETUP_SQWQ) != 0;
+
 	for (i = 0; i < nr; i++) {
-		if (unlikely(mm_fault))
+		if (unlikely(mm_fault)) {
 			ret = -EFAULT;
-		else
+		} else {
 			ret = __io_submit_one(ctx, iocbs[i].iocb,
-						iocbs[i].index, statep, false);
-		if (!ret) {
-			submitted++;
-			continue;
+						iocbs[i].index, statep,
+						force_nonblock);
+			/* nogo, submit to workqueue */
+			if (force_nonblock && ret == -EAGAIN)
+				ret = io_queue_async_work(ctx, &iocbs[i]);
+			if (!ret) {
+				submitted++;
+				continue;
+			}
 		}
 
 		io_fill_cq_error(ctx, iocbs[i].index, ret);
@@ -1131,7 +1165,10 @@ static int io_submit_iocbs(struct io_ring_ctx *ctx, struct iocb_submit *iocbs,
 }
 
 /*
- * sq thread only supports O_DIRECT or FIXEDBUFS IO
+ * SQ thread is woken if the app asked for offloaded submission. This can
+ * be either O_DIRECT, in which case we do submissions directly, or it can
+ * be buffered IO, in which case we do them inline if we can do so without
+ * blocking. If we can't, then we punt to a workqueue.
  */
 static int io_sq_thread(void *data)
 {
@@ -1142,6 +1179,8 @@ static int io_sq_thread(void *data)
 	struct files_struct *old_files;
 	mm_segment_t old_fs;
 	DEFINE_WAIT(wait);
+	unsigned inflight;
+	unsigned long timeout;
 
 	old_files = current->files;
 	current->files = sqo->files;
@@ -1149,14 +1188,43 @@ static int io_sq_thread(void *data)
 	old_fs = get_fs();
 	set_fs(USER_DS);
 
+	timeout = inflight = 0;
 	while (!kthread_should_stop()) {
 		const struct io_uring_iocb *iocb;
 		bool mm_fault = false;
 		unsigned iocb_index;
 		int i;
 
+		if (sqo->thread_poll && inflight) {
+			unsigned int nr_events = 0;
+
+			/*
+			 * Normal IO, just pretend everything completed.
+			 * We don't have to poll completions for that.
+			 */
+			if (ctx->flags & IORING_SETUP_IOPOLL)
+				io_iopoll_check(ctx, &nr_events, 0);
+			else
+				nr_events = inflight;
+
+			inflight -= nr_events;
+			if (!inflight)
+				timeout = jiffies + HZ;
+		}
+
 		iocb = io_peek_sqring(ctx, &iocb_index);
 		if (!iocb) {
+			/*
+			 * If we're polling, let us spin for a second without
+			 * work before going to sleep.
+			 */
+			if (sqo->thread_poll) {
+				if (inflight || !time_after(jiffies, timeout)) {
+					cpu_relax();
+					continue;
+				}
+			}
+
 			/*
 			 * Drop cur_mm before scheduling, we can't hold it for
 			 * long periods (or over schedule()). Do this before
@@ -1170,6 +1238,16 @@ static int io_sq_thread(void *data)
 			}
 
 			prepare_to_wait(&sqo->wait, &wait, TASK_INTERRUPTIBLE);
+
+			/* Tell userspace we may need a wakeup call */
+			if (sqo->thread_poll) {
+				struct io_sq_ring *ring;
+
+				ring = ctx->sq_ring.ring;
+				ring->flags |= IORING_SQ_NEED_WAKEUP;
+				smp_wmb();
+			}
+
 			iocb = io_peek_sqring(ctx, &iocb_index);
 			if (!iocb) {
 				if (kthread_should_park())
@@ -1181,6 +1259,13 @@ static int io_sq_thread(void *data)
 				if (signal_pending(current))
 					flush_signals(current);
 				schedule();
+
+				if (sqo->thread_poll) {
+					struct io_sq_ring *ring;
+
+					ring = ctx->sq_ring.ring;
+					ring->flags &= ~IORING_SQ_NEED_WAKEUP;
+				}
 			}
 			finish_wait(&sqo->wait, &wait);
 			if (!iocb)
@@ -1206,7 +1291,7 @@ static int io_sq_thread(void *data)
 			io_inc_sqring(ctx);
 		} while ((iocb = io_peek_sqring(ctx, &iocb_index)) != NULL);
 
-		io_submit_iocbs(ctx, iocbs, i, cur_mm, mm_fault);
+		inflight += io_submit_iocbs(ctx, iocbs, i, cur_mm, mm_fault);
 	}
 	current->files = old_files;
 	set_fs(old_fs);
@@ -1281,7 +1366,6 @@ static bool io_sq_try_inline(struct io_ring_ctx *ctx,
 static int io_sq_wq_submit(struct io_ring_ctx *ctx, unsigned int to_submit)
 {
 	const struct io_uring_iocb *iocb;
-	struct io_work *work;
 	unsigned iocb_index;
 	int ret, queued;
 
@@ -1289,18 +1373,17 @@ static int io_sq_wq_submit(struct io_ring_ctx *ctx, unsigned int to_submit)
 	while ((iocb = io_peek_sqring(ctx, &iocb_index)) != NULL) {
 		ret = io_sq_try_inline(ctx, iocb, iocb_index);
 		if (!ret) {
-			work = kmalloc(sizeof(*work), GFP_KERNEL);
-			if (!work) {
-				ret = -ENOMEM;
+			struct iocb_submit is = {
+				.iocb = iocb,
+				.index = iocb_index
+			};
+
+			ret = io_queue_async_work(ctx, &is);
+			if (ret)
 				break;
-			}
-			memcpy(&work->iocb, iocb, sizeof(*iocb));
-			io_inc_sqring(ctx);
-			work->iocb_index = iocb_index;
-			INIT_WORK(&work->work, io_sq_wq_submit_work);
-			work->ctx = ctx;
-			queue_work(ctx->sq_offload.wq, &work->work);
 		}
+
+		io_inc_sqring(ctx);
 		queued++;
 		if (queued == to_submit)
 			break;
@@ -1491,6 +1574,9 @@ static int io_sq_thread_start(struct io_ring_ctx *ctx)
 	if (!sqo->files)
 		goto err;
 
+	if (ctx->flags & IORING_SETUP_SQPOLL)
+		sqo->thread_poll = true;
+
 	if (ctx->flags & IORING_SETUP_SQTHREAD) {
 		sqo->thread = kthread_create_on_cpu(io_sq_thread, ctx,
 							ring->sq_thread_cpu,
@@ -1501,7 +1587,8 @@ static int io_sq_thread_start(struct io_ring_ctx *ctx)
 			goto err;
 		}
 		wake_up_process(sqo->thread);
-	} else if (ctx->flags & IORING_SETUP_SQWQ) {
+	}
+	if (ctx->flags & IORING_SETUP_SQWQ) {
 		int concurrency;
 
 		/* Do QD, or 2 * CPUS, whatever is smallest */
@@ -1534,7 +1621,8 @@ static void io_sq_thread_stop(struct io_ring_ctx *ctx)
 		kthread_park(sqo->thread);
 		kthread_stop(sqo->thread);
 		sqo->thread = NULL;
-	} else if (sqo->wq) {
+	}
+	if (sqo->wq) {
 		destroy_workqueue(sqo->wq);
 		sqo->wq = NULL;
 	}
@@ -1792,7 +1880,8 @@ SYSCALL_DEFINE3(io_uring_setup, u32, entries, struct iovec __user *, iovecs,
 	}
 
 	if (p.flags & ~(IORING_SETUP_IOPOLL | IORING_SETUP_FIXEDBUFS |
-			IORING_SETUP_SQTHREAD | IORING_SETUP_SQWQ))
+			IORING_SETUP_SQTHREAD | IORING_SETUP_SQWQ |
+			IORING_SETUP_SQPOLL))
 		return -EINVAL;
 
 	ret = io_uring_create(entries, &p, iovecs);
diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
index 4f0a8ce49f9a..bd665d38dd97 100644
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -38,6 +38,7 @@ struct io_uring_iocb {
 #define IORING_SETUP_FIXEDBUFS	(1 << 1)	/* IO buffers are fixed */
 #define IORING_SETUP_SQTHREAD	(1 << 2)	/* Use SQ thread */
 #define IORING_SETUP_SQWQ	(1 << 3)	/* Use SQ workqueue */
+#define IORING_SETUP_SQPOLL	(1 << 4)	/* SQ thread polls */
 
 #define IORING_OP_READ		1
 #define IORING_OP_WRITE		2
@@ -76,6 +77,8 @@ struct io_sqring_offsets {
 	__u32 resv[3];
 };
 
+#define IORING_SQ_NEED_WAKEUP	(1 << 0) /* needs io_uring_enter wakeup */
+
 struct io_cqring_offsets {
 	__u32 head;
 	__u32 tail;
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 16/16] io_uring: add io_uring_event cache hit information
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
                   ` (14 preceding siblings ...)
  2019-01-08 16:56 ` [PATCH 15/16] io_uring: add submission polling Jens Axboe
@ 2019-01-08 16:56 ` Jens Axboe
  2019-01-09 16:00 ` [PATCHSET v1] io_uring IO interface Matthew Wilcox
  16 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-08 16:56 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

Add hint on whether a read was served out of the page cache, or if it
hit media. This is useful for buffered async IO, O_DIRECT reads would
never have this set (for obvious reasons).

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/io_uring.c                 | 6 +++++-
 include/uapi/linux/io_uring.h | 5 +++++
 2 files changed, 10 insertions(+), 1 deletion(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 6c10841e4342..50b9cfa8c861 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -524,12 +524,16 @@ static void io_fill_cq_error(struct io_ring_ctx *ctx, unsigned ki_index,
 static void io_complete_scqring_rw(struct kiocb *kiocb, long res, long res2)
 {
 	struct io_kiocb *iocb = container_of(kiocb, struct io_kiocb, rw);
+	unsigned ev_flags = 0;
 
 	kiocb_end_write(kiocb);
 
 	fput(kiocb->ki_filp);
 
-	io_complete_scqring(iocb, res, 0);
+	if (res > 0 && test_bit(KIOCB_F_FORCE_NONBLOCK, &iocb->ki_flags))
+		ev_flags = IOEV_FLAG_CACHEHIT;
+
+	io_complete_scqring(iocb, res, ev_flags);
 }
 
 static void io_complete_scqring_iopoll(struct kiocb *kiocb, long res, long res2)
diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
index bd665d38dd97..7dd21126f142 100644
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -56,6 +56,11 @@ struct io_uring_event {
 	__u32	flags;
 };
 
+/*
+ * io_uring_event->flags
+ */
+#define IOEV_FLAG_CACHEHIT	(1 << 0)	/* IO did not hit media */
+
 /*
  * Magic offsets for the application to mmap the data it needs
  */
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 05/16] Add io_uring IO interface
  2019-01-08 16:56 ` [PATCH 05/16] Add io_uring IO interface Jens Axboe
@ 2019-01-09 12:10   ` Christoph Hellwig
  2019-01-09 15:53     ` Jens Axboe
  0 siblings, 1 reply; 39+ messages in thread
From: Christoph Hellwig @ 2019-01-09 12:10 UTC (permalink / raw)
  To: Jens Axboe
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, hch, jmoyer, avi

> index 293733f61594..9ef9987b4192 100644
> --- a/fs/Makefile
> +++ b/fs/Makefile
> @@ -29,7 +29,7 @@ obj-$(CONFIG_SIGNALFD)		+= signalfd.o
>  obj-$(CONFIG_TIMERFD)		+= timerfd.o
>  obj-$(CONFIG_EVENTFD)		+= eventfd.o
>  obj-$(CONFIG_USERFAULTFD)	+= userfaultfd.o
> -obj-$(CONFIG_AIO)               += aio.o
> +obj-$(CONFIG_AIO)               += aio.o io_uring.o

It is probablt worth adding a new config symbol for the uring as no
code is shared with aio.

> diff --git a/fs/io_uring.c b/fs/io_uring.c
> new file mode 100644
> index 000000000000..ae2b886282bb
> --- /dev/null
> +++ b/fs/io_uring.c
> @@ -0,0 +1,849 @@
> +/*
> + * Shared application/kernel submission and completion ring pairs, for
> + * supporting fast/efficient IO.
> + *
> + * Copyright (C) 2019 Jens Axboe
> + */

Add an SPDX header to all new files, please.

> +struct io_sq_ring {
> +	struct io_uring		r;
> +	u32			ring_mask;
> +	u32			ring_entries;
> +	u32			dropped;
> +	u32			flags;
> +	u32			array[0];
> +};

field[0] is a legacy gcc extension, the proper C99+ way is field[].

> +
> +struct io_iocb_ring {
> +	struct			io_sq_ring *ring;
> +	unsigned		entries;
> +	unsigned		ring_mask;
> +	struct io_uring_iocb	*iocbs;
> +};
> +
> +struct io_event_ring {
> +	struct io_cq_ring	*ring;
> +	unsigned		entries;
> +	unsigned		ring_mask;
> +};

Btw, do we really need there structures?  It would seem simpler
to just embedd them into the containing structure as:

	struct io_sq_ring	*sq_ring;
	unsigned		sq_ring_entries;
	unsigned		sq_ring_mask;
	struct io_uring_iocb	*sq_ring_iocbs;

	struct io_cq_ring	*cq_ring;
	unsigned		cq_ring_entries;
	unsigned		cq_ring_mask;
	

> +struct io_ring_ctx {
> +	struct percpu_ref	refs;
> +
> +	unsigned int		flags;
> +	unsigned int		max_reqs;

max_reqs can probably go away in favour of the sq ring nr_entries
field.

> +	struct io_iocb_ring	sq_ring;
> +	struct io_event_ring	cq_ring;
> +
> +	struct work_struct	work;
> +
> +	struct {
> +		struct mutex uring_lock;
> +	} ____cacheline_aligned_in_smp;
> +
> +	struct {
> +		struct mutex    ring_lock;
> +		wait_queue_head_t wait;
> +	} ____cacheline_aligned_in_smp;
> +
> +	struct {
> +		spinlock_t      completion_lock;
> +	} ____cacheline_aligned_in_smp;
> +};

Can you take a deep look if we need to keep all of ring_lock,
completion_lock and the later added poll locking?  From a quick look
is isn't entirely clear what the locking strategy on the completion
side is.  It needs to be documented and can hopefully be simplified.

> +struct fsync_iocb {
> +	struct work_struct	work;
> +	struct file		*file;
> +	bool			datasync;
> +};

Do we actually need this?  Can't we just reuse the later thread
offload for fsync?  Maybe just add fsync support once everything else
is done to make that simpler.

> +static const struct file_operations io_scqring_fops;
> +
> +static void io_ring_ctx_free(struct work_struct *work);
> +static void io_ring_ctx_ref_free(struct percpu_ref *ref);

Can you try to avoid to need the forward delcaration?  (except for the
fops, where we probably need it).

>
> +
> +static struct io_ring_ctx *io_ring_ctx_alloc(struct io_uring_params *p)
> +{
> +	struct io_ring_ctx *ctx;
> +
> +	ctx = kmem_cache_zalloc(ioctx_cachep, GFP_KERNEL);
> +	if (!ctx)
> +		return NULL;

Do we really need an explicit slab for the contexts?

> +static struct io_kiocb *io_get_req(struct io_ring_ctx *ctx)

Maybe replace the req name with something matching the structure
name?  (and more on the structure name later).

> +{
> +	struct io_kiocb *req;
> +
> +	req = kmem_cache_alloc(kiocb_cachep, GFP_KERNEL);
> +	if (!req)
> +		return NULL;
> +
> +	percpu_ref_get(&ctx->refs);
> +	req->ki_ctx = ctx;
> +	INIT_LIST_HEAD(&req->ki_list);

We never do a list_empty ceck on ki_list, so there should be no need
to initialize it.

> +static void io_fill_event(struct io_uring_event *ev, struct io_kiocb *kiocb,
> +			  long res, unsigned flags)
> +{
> +	ev->index = kiocb->ki_index;
> +	ev->res = res;
> +	ev->flags = flags;
> +}

Probably no need for this helper.

> +static void io_complete_scqring(struct io_kiocb *iocb, long res, unsigned flags)
> +{
> +	io_cqring_fill_event(iocb, res, flags);
> +	io_complete_iocb(iocb->ki_ctx, iocb);
> +}

Probably no need for this helper either.

> +	ret = kiocb_set_rw_flags(req, iocb->rw_flags);
> +	if (unlikely(ret))
> +		goto out_fput;
> +
> +	/* no one is going to poll for this I/O */
> +	req->ki_flags &= ~IOCB_HIPRI;

Now that we don't have the aio legacy to deal with should we just
reject IOCB_HIPRI on a non-polled context?

> +static int io_setup_rw(int rw, const struct io_uring_iocb *iocb,
> +		       struct iovec **iovec, struct iov_iter *iter)
> +{
> +	void __user *buf = (void __user *)(uintptr_t)iocb->addr;
> +	size_t ret;
> +
> +	ret = import_single_range(rw, buf, iocb->len, *iovec, iter);
> +	*iovec = NULL;
> +	return ret;
> +}

Is there any point in supporting non-vectored operations here?

> +		if (S_ISREG(file_inode(file)->i_mode)) {
> +			__sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, true);
> +			__sb_writers_release(file_inode(file)->i_sb, SB_FREEZE_WRITE);
> +		}

Overly long lines.

> +static int __io_submit_one(struct io_ring_ctx *ctx,
> +			   const struct io_uring_iocb *iocb,
> +			   unsigned long ki_index)

Maybe calls this io_ring_submit_one?  Or generally find a nice prefix
for all the functions in this file?

> +	f = fdget(fd);
> +	if (f.file) {
> +		struct io_ring_ctx *ctx;

Please just return early on fialure instead of forcing another level
of indentation.

> +
> +	ctx->sq_ring.iocbs = io_mem_alloc(sizeof(struct io_uring_iocb) *
> +						p->sq_entries);

Use array_size().

> +/*
> + * sys_io_uring_setup:
> + *	Sets up an aio uring context, and returns the fd. Applications asks
> + *	for a ring size, we return the actual sq/cq ring sizes (among other
> + *	things) in the params structure passed in.
> + */

Can we drop this odd aio-style comment format?  In fact the syscall
documentation probably just belongs into the man page only anyway.

Same for the uring_enter syscall.

> +struct io_uring_iocb {

Should we just call this io_uring_sqe?

> +/*
> + * IO completion data structure
> + */
> +struct io_uring_event {
> +	__u64	index;		/* what iocb this event came from */
> +	__s32	res;		/* result code for this event */
> +	__u32	flags;
> +};

io_uring_cqe?

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 06/16] io_uring: support for IO polling
  2019-01-08 16:56 ` [PATCH 06/16] io_uring: support for IO polling Jens Axboe
@ 2019-01-09 12:11   ` Christoph Hellwig
  2019-01-09 15:53     ` Jens Axboe
  0 siblings, 1 reply; 39+ messages in thread
From: Christoph Hellwig @ 2019-01-09 12:11 UTC (permalink / raw)
  To: Jens Axboe
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, hch, jmoyer, avi

On Tue, Jan 08, 2019 at 09:56:35AM -0700, Jens Axboe wrote:
> Add polled variants of the read and write commands. These act like their
> non-polled counterparts, except we expect to poll for completion of
> them.

These aren't really need command variants, but a different type of context.

>  	case IORING_OP_FSYNC:
> +		if (ctx->flags & IORING_SETUP_IOPOLL)
> +			break;
>  		ret = io_fsync(&req->fsync, iocb, false);
>  		break;
>  	case IORING_OP_FDSYNC:
> +		if (ctx->flags & IORING_SETUP_IOPOLL)
> +			break;
>  		ret = io_fsync(&req->fsync, iocb, true);

I'd move this check into io_fsync to avoid a little duplication.

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 10/16] io_uring: split kiocb init from allocation
  2019-01-08 16:56 ` [PATCH 10/16] io_uring: split kiocb init from allocation Jens Axboe
@ 2019-01-09 12:12   ` Christoph Hellwig
  2019-01-09 16:56     ` Jens Axboe
  0 siblings, 1 reply; 39+ messages in thread
From: Christoph Hellwig @ 2019-01-09 12:12 UTC (permalink / raw)
  To: Jens Axboe
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, hch, jmoyer, avi

On Tue, Jan 08, 2019 at 09:56:39AM -0700, Jens Axboe wrote:
> In preparation from having pre-allocated requests, that we then just
> need to initialize before use.
> 
> Signed-off-by: Jens Axboe <axboe@kernel.dk>
> ---
>  fs/io_uring.c | 13 +++++++++----
>  1 file changed, 9 insertions(+), 4 deletions(-)
> 
> diff --git a/fs/io_uring.c b/fs/io_uring.c
> index afbaebb63012..11d045f0f799 100644
> --- a/fs/io_uring.c
> +++ b/fs/io_uring.c
> @@ -202,6 +202,14 @@ static struct io_uring_event *io_peek_cqring(struct io_ring_ctx *ctx)
>  	return &ring->events[tail & ctx->cq_ring.ring_mask];
>  }
>  
> +static void io_req_init(struct io_ring_ctx *ctx, struct io_kiocb *req)
> +{
> +	percpu_ref_get(&ctx->refs);
> +	req->ki_ctx = ctx;
> +	INIT_LIST_HEAD(&req->ki_list);
> +	req->ki_flags = 0;

We still only have a single caller of this in the final tree, and
I don't think this function helps.  So I'd just drop the patch.

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 11/16] io_uring: batch io_kiocb allocation
  2019-01-08 16:56 ` [PATCH 11/16] io_uring: batch io_kiocb allocation Jens Axboe
@ 2019-01-09 12:13   ` Christoph Hellwig
  2019-01-09 16:57     ` Jens Axboe
  0 siblings, 1 reply; 39+ messages in thread
From: Christoph Hellwig @ 2019-01-09 12:13 UTC (permalink / raw)
  To: Jens Axboe
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, hch, jmoyer, avi

> +	if (!state)
> +		req = kmem_cache_alloc(kiocb_cachep, GFP_KERNEL);

Just return an error here if kmem_cache_alloc fails.

> +	if (req)
> +		io_req_init(ctx, req);

Because all the other ones can't reached this with a NULL req.

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 13/16] io_uring: add support for pre-mapped user IO buffers
  2019-01-08 16:56 ` [PATCH 13/16] io_uring: add support for pre-mapped user IO buffers Jens Axboe
@ 2019-01-09 12:16   ` Christoph Hellwig
  2019-01-09 17:06     ` Jens Axboe
  0 siblings, 1 reply; 39+ messages in thread
From: Christoph Hellwig @ 2019-01-09 12:16 UTC (permalink / raw)
  To: Jens Axboe
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, hch, jmoyer, avi

> +static int io_setup_rw(int rw, struct io_kiocb *kiocb,
> +		       const struct io_uring_iocb *iocb, struct iovec **iovec,
> +		       struct iov_iter *iter, bool kaddr)
>  {
>  	void __user *buf = (void __user *)(uintptr_t)iocb->addr;
>  	size_t ret;
>  
> -	ret = import_single_range(rw, buf, iocb->len, *iovec, iter);
> +	if (!kaddr) {
> +		ret = import_single_range(rw, buf, iocb->len, *iovec, iter);
> +	} else {
> +		struct io_ring_ctx *ctx = kiocb->ki_ctx;
> +		struct io_mapped_ubuf *imu;
> +		size_t len = iocb->len;
> +		size_t offset;
> +		int index;
> +
> +		/* __io_submit_one() already validated the index */
> +		index = array_index_nospec(kiocb->ki_index,
> +						ctx->max_reqs);
> +		imu = &ctx->user_bufs[index];
> +		if ((unsigned long) iocb->addr < imu->ubuf ||
> +		    (unsigned long) iocb->addr + len > imu->ubuf + imu->len) {
> +			ret = -EFAULT;
> +			goto err;
> +		}
> +
> +		/*
> +		 * May not be a start of buffer, set size appropriately
> +		 * and advance us to the beginning.
> +		 */
> +		offset = (unsigned long) iocb->addr - imu->ubuf;
> +		iov_iter_bvec(iter, rw, imu->bvec, imu->nr_bvecs,
> +				offset + len);
> +		if (offset)
> +			iov_iter_advance(iter, offset);
> +		ret = 0;
> +

Please split this code in a separate helper.

>  	if (unlikely(!file->f_op->read_iter))
>  		goto out_fput;
>  
> -	ret = io_setup_rw(READ, iocb, &iovec, &iter);
> +	ret = io_setup_rw(READ, kiocb, iocb, &iovec, &iter, kaddr);

And I'd personally just call that helper here based on the opcode and
avoid magic bool arguments.

> +		down_write(&current->mm->mmap_sem);
> +		pret = get_user_pages(ubuf, nr_pages, 1, pages, NULL);
> +		up_write(&current->mm->mmap_sem);

This needs to be get_user_pages_longterm.

> +	 * We don't use the iovecs without fixed buffers being asked for.
> +	 * Error out if they don't match.
> +	 */
> +	if (!(p->flags & IORING_SETUP_FIXEDBUFS) && iovecs)
> +		return -EINVAL;

I don't think we need the IORING_SETUP_FIXEDBUFS flag at all, as a
non-zero iovecs pointer is enough of an indication.

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 05/16] Add io_uring IO interface
  2019-01-09 12:10   ` Christoph Hellwig
@ 2019-01-09 15:53     ` Jens Axboe
  2019-01-09 18:30       ` Christoph Hellwig
  0 siblings, 1 reply; 39+ messages in thread
From: Jens Axboe @ 2019-01-09 15:53 UTC (permalink / raw)
  To: Christoph Hellwig
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, jmoyer, avi

On 1/9/19 5:10 AM, Christoph Hellwig wrote:
>> index 293733f61594..9ef9987b4192 100644
>> --- a/fs/Makefile
>> +++ b/fs/Makefile
>> @@ -29,7 +29,7 @@ obj-$(CONFIG_SIGNALFD)		+= signalfd.o
>>  obj-$(CONFIG_TIMERFD)		+= timerfd.o
>>  obj-$(CONFIG_EVENTFD)		+= eventfd.o
>>  obj-$(CONFIG_USERFAULTFD)	+= userfaultfd.o
>> -obj-$(CONFIG_AIO)               += aio.o
>> +obj-$(CONFIG_AIO)               += aio.o io_uring.o
> 
> It is probablt worth adding a new config symbol for the uring as no
> code is shared with aio.

Agreed, done.

>> diff --git a/fs/io_uring.c b/fs/io_uring.c
>> new file mode 100644
>> index 000000000000..ae2b886282bb
>> --- /dev/null
>> +++ b/fs/io_uring.c
>> @@ -0,0 +1,849 @@
>> +/*
>> + * Shared application/kernel submission and completion ring pairs, for
>> + * supporting fast/efficient IO.
>> + *
>> + * Copyright (C) 2019 Jens Axboe
>> + */
> 
> Add an SPDX header to all new files, please.

Done

>> +struct io_sq_ring {
>> +	struct io_uring		r;
>> +	u32			ring_mask;
>> +	u32			ring_entries;
>> +	u32			dropped;
>> +	u32			flags;
>> +	u32			array[0];
>> +};
> 
> field[0] is a legacy gcc extension, the proper C99+ way is field[].

Fixed

>> +struct io_iocb_ring {
>> +	struct			io_sq_ring *ring;
>> +	unsigned		entries;
>> +	unsigned		ring_mask;
>> +	struct io_uring_iocb	*iocbs;
>> +};
>> +
>> +struct io_event_ring {
>> +	struct io_cq_ring	*ring;
>> +	unsigned		entries;
>> +	unsigned		ring_mask;
>> +};
> 
> Btw, do we really need there structures?  It would seem simpler
> to just embedd them into the containing structure as:
> 
> 	struct io_sq_ring	*sq_ring;
> 	unsigned		sq_ring_entries;
> 	unsigned		sq_ring_mask;
> 	struct io_uring_iocb	*sq_ring_iocbs;
> 
> 	struct io_cq_ring	*cq_ring;
> 	unsigned		cq_ring_entries;
> 	unsigned		cq_ring_mask;

Yeah, I guess we use it directly in so few places that we may as well
just get rid of the structs for these.

> 	
> 
>> +struct io_ring_ctx {
>> +	struct percpu_ref	refs;
>> +
>> +	unsigned int		flags;
>> +	unsigned int		max_reqs;
> 
> max_reqs can probably go away in favour of the sq ring nr_entries
> field.

Indeed, killed.

>> +	struct io_iocb_ring	sq_ring;
>> +	struct io_event_ring	cq_ring;
>> +
>> +	struct work_struct	work;
>> +
>> +	struct {
>> +		struct mutex uring_lock;
>> +	} ____cacheline_aligned_in_smp;
>> +
>> +	struct {
>> +		struct mutex    ring_lock;
>> +		wait_queue_head_t wait;
>> +	} ____cacheline_aligned_in_smp;
>> +
>> +	struct {
>> +		spinlock_t      completion_lock;
>> +	} ____cacheline_aligned_in_smp;
>> +};
> 
> Can you take a deep look if we need to keep all of ring_lock,
> completion_lock and the later added poll locking?  From a quick look
> is isn't entirely clear what the locking strategy on the completion
> side is.  It needs to be documented and can hopefully be simplified.

I think we just need to kill ring_lock, it's actually not even used.
I'll take a closer look at the locking as well.

> 
>> +struct fsync_iocb {
>> +	struct work_struct	work;
>> +	struct file		*file;
>> +	bool			datasync;
>> +};
> 
> Do we actually need this?  Can't we just reuse the later thread
> offload for fsync?  Maybe just add fsync support once everything else
> is done to make that simpler.

We can just use the sq thread, but we don't always have that backing. I
guess we could create it lazily if an fsync comes in. I'll take a look
at adding that as a separate thing.

>> +static const struct file_operations io_scqring_fops;
>> +
>> +static void io_ring_ctx_free(struct work_struct *work);
>> +static void io_ring_ctx_ref_free(struct percpu_ref *ref);
> 
> Can you try to avoid to need the forward delcaration?  (except for the
> fops, where we probably need it).

I got rid of one of them in my current tree already, I'll see if I can
dump the other one.

>> +static struct io_ring_ctx *io_ring_ctx_alloc(struct io_uring_params *p)
>> +{
>> +	struct io_ring_ctx *ctx;
>> +
>> +	ctx = kmem_cache_zalloc(ioctx_cachep, GFP_KERNEL);
>> +	if (!ctx)
>> +		return NULL;
> 
> Do we really need an explicit slab for the contexts?

Not sure, guess it depends on the frequency of them. But I suspect that
it won't matter one bit, I'll kill this slab.

> 
>> +static struct io_kiocb *io_get_req(struct io_ring_ctx *ctx)
> 
> Maybe replace the req name with something matching the structure
> name?  (and more on the structure name later).

Make sense.

>> +{
>> +	struct io_kiocb *req;
>> +
>> +	req = kmem_cache_alloc(kiocb_cachep, GFP_KERNEL);
>> +	if (!req)
>> +		return NULL;
>> +
>> +	percpu_ref_get(&ctx->refs);
>> +	req->ki_ctx = ctx;
>> +	INIT_LIST_HEAD(&req->ki_list);
> 
> We never do a list_empty ceck on ki_list, so there should be no need
> to initialize it.

Killed

>> +static void io_fill_event(struct io_uring_event *ev, struct io_kiocb *kiocb,
>> +			  long res, unsigned flags)
>> +{
>> +	ev->index = kiocb->ki_index;
>> +	ev->res = res;
>> +	ev->flags = flags;
>> +}
> 
> Probably no need for this helper.

Killed. Also realized that we're missing a store ordering barrier after
filling in 'ev', but before incrementing the ring.

>> +static void io_complete_scqring(struct io_kiocb *iocb, long res, unsigned flags)
>> +{
>> +	io_cqring_fill_event(iocb, res, flags);
>> +	io_complete_iocb(iocb->ki_ctx, iocb);
>> +}
> 
> Probably no need for this helper either.

Killed

> 
>> +	ret = kiocb_set_rw_flags(req, iocb->rw_flags);
>> +	if (unlikely(ret))
>> +		goto out_fput;
>> +
>> +	/* no one is going to poll for this I/O */
>> +	req->ki_flags &= ~IOCB_HIPRI;
> 
> Now that we don't have the aio legacy to deal with should we just
> reject IOCB_HIPRI on a non-polled context?

Yes I think so, we don't have any legacy behavior to adhere to.

> 
>> +static int io_setup_rw(int rw, const struct io_uring_iocb *iocb,
>> +		       struct iovec **iovec, struct iov_iter *iter)
>> +{
>> +	void __user *buf = (void __user *)(uintptr_t)iocb->addr;
>> +	size_t ret;
>> +
>> +	ret = import_single_range(rw, buf, iocb->len, *iovec, iter);
>> +	*iovec = NULL;
>> +	return ret;
>> +}
> 
> Is there any point in supporting non-vectored operations here?

Not sure I follow?

>> +		if (S_ISREG(file_inode(file)->i_mode)) {
>> +			__sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, true);
>> +			__sb_writers_release(file_inode(file)->i_sb, SB_FREEZE_WRITE);
>> +		}
> 
> Overly long lines.

Fixed

>> +static int __io_submit_one(struct io_ring_ctx *ctx,
>> +			   const struct io_uring_iocb *iocb,
>> +			   unsigned long ki_index)
> 
> Maybe calls this io_ring_submit_one?  Or generally find a nice prefix
> for all the functions in this file?

Agree, some of this is leftover cruft from the aio side. I'll clean it
up.

>> +	f = fdget(fd);
>> +	if (f.file) {
>> +		struct io_ring_ctx *ctx;
> 
> Please just return early on fialure instead of forcing another level
> of indentation.

Sure, done.

> 
>> +
>> +	ctx->sq_ring.iocbs = io_mem_alloc(sizeof(struct io_uring_iocb) *
>> +						p->sq_entries);
> 
> Use array_size().

Done

>> +/*
>> + * sys_io_uring_setup:
>> + *	Sets up an aio uring context, and returns the fd. Applications asks
>> + *	for a ring size, we return the actual sq/cq ring sizes (among other
>> + *	things) in the params structure passed in.
>> + */
> 
> Can we drop this odd aio-style comment format?  In fact the syscall
> documentation probably just belongs into the man page only anyway.
> 
> Same for the uring_enter syscall.

Sure, not a big deal to me, dropped.

>> +struct io_uring_iocb {
> 
> Should we just call this io_uring_sqe?
> 
>> +/*
>> + * IO completion data structure
>> + */
>> +struct io_uring_event {
>> +	__u64	index;		/* what iocb this event came from */
>> +	__s32	res;		/* result code for this event */
>> +	__u32	flags;
>> +};
> 
> io_uring_cqe?

I'm fine with that, I like the symmetry of the names.

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 06/16] io_uring: support for IO polling
  2019-01-09 12:11   ` Christoph Hellwig
@ 2019-01-09 15:53     ` Jens Axboe
  0 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-09 15:53 UTC (permalink / raw)
  To: Christoph Hellwig
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, jmoyer, avi

On 1/9/19 5:11 AM, Christoph Hellwig wrote:
> On Tue, Jan 08, 2019 at 09:56:35AM -0700, Jens Axboe wrote:
>> Add polled variants of the read and write commands. These act like their
>> non-polled counterparts, except we expect to poll for completion of
>> them.
> 
> These aren't really need command variants, but a different type of context.

That is phrased poorly, I'll fix that.

>>  	case IORING_OP_FSYNC:
>> +		if (ctx->flags & IORING_SETUP_IOPOLL)
>> +			break;
>>  		ret = io_fsync(&req->fsync, iocb, false);
>>  		break;
>>  	case IORING_OP_FDSYNC:
>> +		if (ctx->flags & IORING_SETUP_IOPOLL)
>> +			break;
>>  		ret = io_fsync(&req->fsync, iocb, true);
> 
> I'd move this check into io_fsync to avoid a little duplication.

Done

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCHSET v1] io_uring IO interface
  2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
                   ` (15 preceding siblings ...)
  2019-01-08 16:56 ` [PATCH 16/16] io_uring: add io_uring_event cache hit information Jens Axboe
@ 2019-01-09 16:00 ` Matthew Wilcox
  2019-01-09 16:27   ` Chris Mason
  16 siblings, 1 reply; 39+ messages in thread
From: Matthew Wilcox @ 2019-01-09 16:00 UTC (permalink / raw)
  To: Jens Axboe
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, hch, jmoyer, avi

On Tue, Jan 08, 2019 at 09:56:29AM -0700, Jens Axboe wrote:
> After some arm twisting from Christoph, I finally caved and divorced the
> aio-poll patches from aio/libaio itself. The io_uring interface itself
> is useful and efficient, and after rebasing all the new goodies on top
> of that, there was little reason to retail the aio connection.
> 
> Hence io_uring was born. This is what I previously called scqring for
> aio, but now as a standalone entity. Patch #5 adds the core of this
> interface, but in short, it has two main data structures:

Please can we call it something that looks a little less like io_urine?
aio_ring?  io_ring?  ring_io?

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCHSET v1] io_uring IO interface
  2019-01-09 16:00 ` [PATCHSET v1] io_uring IO interface Matthew Wilcox
@ 2019-01-09 16:27   ` Chris Mason
  0 siblings, 0 replies; 39+ messages in thread
From: Chris Mason @ 2019-01-09 16:27 UTC (permalink / raw)
  To: Matthew Wilcox
  Cc: Jens Axboe, linux-fsdevel, linux-aio, linux-block, linux-arch,
	hch, jmoyer, avi

On 9 Jan 2019, at 11:00, Matthew Wilcox wrote:

> On Tue, Jan 08, 2019 at 09:56:29AM -0700, Jens Axboe wrote:
>> After some arm twisting from Christoph, I finally caved and divorced 
>> the
>> aio-poll patches from aio/libaio itself. The io_uring interface 
>> itself
>> is useful and efficient, and after rebasing all the new goodies on 
>> top
>> of that, there was little reason to retail the aio connection.
>>
>> Hence io_uring was born. This is what I previously called scqring for
>> aio, but now as a standalone entity. Patch #5 adds the core of this
>> interface, but in short, it has two main data structures:
>
> Please can we call it something that looks a little less like 
> io_urine?
> aio_ring?  io_ring?  ring_io?

We must be using the same fonts, I've been harassing Jens with jokes 
about this for days.  But, I totally missed the link between io_uring 
and https://twitter.com/axboe/status/927571366085246976 until just now.  
It's clear that Jens has been planning this for over a year, and he was 
just waiting for us to catch on.


Really though, io_uring is growing on me.  The kernel is full of rings 
and it does help set it apart.

-chris

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 10/16] io_uring: split kiocb init from allocation
  2019-01-09 12:12   ` Christoph Hellwig
@ 2019-01-09 16:56     ` Jens Axboe
  0 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-09 16:56 UTC (permalink / raw)
  To: Christoph Hellwig
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, jmoyer, avi

On 1/9/19 5:12 AM, Christoph Hellwig wrote:
> On Tue, Jan 08, 2019 at 09:56:39AM -0700, Jens Axboe wrote:
>> In preparation from having pre-allocated requests, that we then just
>> need to initialize before use.
>>
>> Signed-off-by: Jens Axboe <axboe@kernel.dk>
>> ---
>>  fs/io_uring.c | 13 +++++++++----
>>  1 file changed, 9 insertions(+), 4 deletions(-)
>>
>> diff --git a/fs/io_uring.c b/fs/io_uring.c
>> index afbaebb63012..11d045f0f799 100644
>> --- a/fs/io_uring.c
>> +++ b/fs/io_uring.c
>> @@ -202,6 +202,14 @@ static struct io_uring_event *io_peek_cqring(struct io_ring_ctx *ctx)
>>  	return &ring->events[tail & ctx->cq_ring.ring_mask];
>>  }
>>  
>> +static void io_req_init(struct io_ring_ctx *ctx, struct io_kiocb *req)
>> +{
>> +	percpu_ref_get(&ctx->refs);
>> +	req->ki_ctx = ctx;
>> +	INIT_LIST_HEAD(&req->ki_list);
>> +	req->ki_flags = 0;
> 
> We still only have a single caller of this in the final tree, and
> I don't think this function helps.  So I'd just drop the patch.

Dropped

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 11/16] io_uring: batch io_kiocb allocation
  2019-01-09 12:13   ` Christoph Hellwig
@ 2019-01-09 16:57     ` Jens Axboe
  2019-01-09 19:03       ` Christoph Hellwig
  0 siblings, 1 reply; 39+ messages in thread
From: Jens Axboe @ 2019-01-09 16:57 UTC (permalink / raw)
  To: Christoph Hellwig
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, jmoyer, avi

On 1/9/19 5:13 AM, Christoph Hellwig wrote:
>> +	if (!state)
>> +		req = kmem_cache_alloc(kiocb_cachep, GFP_KERNEL);
> 
> Just return an error here if kmem_cache_alloc fails.
> 
>> +	if (req)
>> +		io_req_init(ctx, req);
> 
> Because all the other ones can't reached this with a NULL req.

This is different in the current tree, since I properly fixed the
ctx ref issue.

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 13/16] io_uring: add support for pre-mapped user IO buffers
  2019-01-09 12:16   ` Christoph Hellwig
@ 2019-01-09 17:06     ` Jens Axboe
  0 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-09 17:06 UTC (permalink / raw)
  To: Christoph Hellwig
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, jmoyer, avi

On 1/9/19 5:16 AM, Christoph Hellwig wrote:
>> +static int io_setup_rw(int rw, struct io_kiocb *kiocb,
>> +		       const struct io_uring_iocb *iocb, struct iovec **iovec,
>> +		       struct iov_iter *iter, bool kaddr)
>>  {
>>  	void __user *buf = (void __user *)(uintptr_t)iocb->addr;
>>  	size_t ret;
>>  
>> -	ret = import_single_range(rw, buf, iocb->len, *iovec, iter);
>> +	if (!kaddr) {
>> +		ret = import_single_range(rw, buf, iocb->len, *iovec, iter);
>> +	} else {
>> +		struct io_ring_ctx *ctx = kiocb->ki_ctx;
>> +		struct io_mapped_ubuf *imu;
>> +		size_t len = iocb->len;
>> +		size_t offset;
>> +		int index;
>> +
>> +		/* __io_submit_one() already validated the index */
>> +		index = array_index_nospec(kiocb->ki_index,
>> +						ctx->max_reqs);
>> +		imu = &ctx->user_bufs[index];
>> +		if ((unsigned long) iocb->addr < imu->ubuf ||
>> +		    (unsigned long) iocb->addr + len > imu->ubuf + imu->len) {
>> +			ret = -EFAULT;
>> +			goto err;
>> +		}
>> +
>> +		/*
>> +		 * May not be a start of buffer, set size appropriately
>> +		 * and advance us to the beginning.
>> +		 */
>> +		offset = (unsigned long) iocb->addr - imu->ubuf;
>> +		iov_iter_bvec(iter, rw, imu->bvec, imu->nr_bvecs,
>> +				offset + len);
>> +		if (offset)
>> +			iov_iter_advance(iter, offset);
>> +		ret = 0;
>> +
> 
> Please split this code in a separate helper.

Done

>>  	if (unlikely(!file->f_op->read_iter))
>>  		goto out_fput;
>>  
>> -	ret = io_setup_rw(READ, iocb, &iovec, &iter);
>> +	ret = io_setup_rw(READ, kiocb, iocb, &iovec, &iter, kaddr);
> 
> And I'd personally just call that helper here based on the opcode and
> avoid magic bool arguments.

Then we can also fold the switch cases, cleans it up.

>> +		down_write(&current->mm->mmap_sem);
>> +		pret = get_user_pages(ubuf, nr_pages, 1, pages, NULL);
>> +		up_write(&current->mm->mmap_sem);
> 
> This needs to be get_user_pages_longterm.

Done

>> +	 * We don't use the iovecs without fixed buffers being asked for.
>> +	 * Error out if they don't match.
>> +	 */
>> +	if (!(p->flags & IORING_SETUP_FIXEDBUFS) && iovecs)
>> +		return -EINVAL;
> 
> I don't think we need the IORING_SETUP_FIXEDBUFS flag at all, as a
> non-zero iovecs pointer is enough of an indication.

Good point, no point in that redundancy. Fixed.

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 05/16] Add io_uring IO interface
  2019-01-09 15:53     ` Jens Axboe
@ 2019-01-09 18:30       ` Christoph Hellwig
  2019-01-09 20:07         ` Jens Axboe
  0 siblings, 1 reply; 39+ messages in thread
From: Christoph Hellwig @ 2019-01-09 18:30 UTC (permalink / raw)
  To: Jens Axboe
  Cc: Christoph Hellwig, linux-fsdevel, linux-aio, linux-block,
	linux-arch, jmoyer, avi

On Wed, Jan 09, 2019 at 08:53:31AM -0700, Jens Axboe wrote:
> >> +static int io_setup_rw(int rw, const struct io_uring_iocb *iocb,
> >> +		       struct iovec **iovec, struct iov_iter *iter)
> >> +{
> >> +	void __user *buf = (void __user *)(uintptr_t)iocb->addr;
> >> +	size_t ret;
> >> +
> >> +	ret = import_single_range(rw, buf, iocb->len, *iovec, iter);
> >> +	*iovec = NULL;
> >> +	return ret;
> >> +}
> > 
> > Is there any point in supporting non-vectored operations here?
> 
> Not sure I follow?

This version only supports non-vectored read and write, that is
the equivalent of pread/pwrite.  Many AIO users really need vectored
operations, that is preadv/pwritev semantics indirecting through
a struct iovec array.  The non-vectored version can be trivially
emulated using a vector of 1, which is what we do in the kernel
I/O stack everywhere.  So I think we should just support the vectored
version here, and not the non-vectored one.  See my io_uring branch
for the sketeched implementation.

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 11/16] io_uring: batch io_kiocb allocation
  2019-01-09 16:57     ` Jens Axboe
@ 2019-01-09 19:03       ` Christoph Hellwig
  2019-01-09 20:08         ` Jens Axboe
  0 siblings, 1 reply; 39+ messages in thread
From: Christoph Hellwig @ 2019-01-09 19:03 UTC (permalink / raw)
  To: Jens Axboe
  Cc: Christoph Hellwig, linux-fsdevel, linux-aio, linux-block,
	linux-arch, jmoyer, avi

On Wed, Jan 09, 2019 at 09:57:59AM -0700, Jens Axboe wrote:
> On 1/9/19 5:13 AM, Christoph Hellwig wrote:
> >> +	if (!state)
> >> +		req = kmem_cache_alloc(kiocb_cachep, GFP_KERNEL);
> > 
> > Just return an error here if kmem_cache_alloc fails.
> > 
> >> +	if (req)
> >> +		io_req_init(ctx, req);
> > 
> > Because all the other ones can't reached this with a NULL req.
> 
> This is different in the current tree, since I properly fixed the
> ctx ref issue.

Your tree does a percpu_ref_tryget very first, and then leaks that if
kmem_cache_alloc_bulk fails, and also is inconsistent for NULL vs
ERR_PTR returns.  I think you want something like this on top:

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 35d055dcbc22..6c95749e9601 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -250,14 +250,6 @@ static struct io_uring_event *io_peek_cqring(struct io_ring_ctx *ctx)
 	return &ring->events[tail & ctx->cq_ring.ring_mask];
 }
 
-static bool io_req_init(struct io_ring_ctx *ctx, struct io_kiocb *req)
-{
-	req->ki_ctx = ctx;
-	INIT_LIST_HEAD(&req->ki_list);
-	req->ki_flags = 0;
-	return true;
-}
-
 static void io_ring_drop_ctx_ref(struct io_ring_ctx *ctx, unsigned refs)
 {
 	percpu_ref_put_many(&ctx->refs, refs);
@@ -274,9 +266,11 @@ static struct io_kiocb *io_get_req(struct io_ring_ctx *ctx,
 	if (!percpu_ref_tryget(&ctx->refs))
 		return NULL;
 
-	if (!state)
+	if (!state) {
 		req = kmem_cache_alloc(kiocb_cachep, GFP_KERNEL);
-	else if (!state->free_iocbs) {
+		if (!req)
+			goto out_drop_ref;
+	} else if (!state->free_iocbs) {
 		size_t size;
 		int ret;
 
@@ -284,7 +278,7 @@ static struct io_kiocb *io_get_req(struct io_ring_ctx *ctx,
 		ret = kmem_cache_alloc_bulk(kiocb_cachep, GFP_KERNEL, size,
 						state->iocbs);
 		if (ret <= 0)
-			return ERR_PTR(-ENOMEM);
+			goto out_drop_ref;
 		state->free_iocbs = ret - 1;
 		state->cur_iocb = 1;
 		req = state->iocbs[0];
@@ -294,11 +288,11 @@ static struct io_kiocb *io_get_req(struct io_ring_ctx *ctx,
 		state->cur_iocb++;
 	}
 
-	if (req) {
-		io_req_init(ctx, req);
-		return req;
-	}
+	req->ki_ctx = ctx;
+	req->ki_flags = 0;
+	return req;
 
+out_drop_ref:
 	io_ring_drop_ctx_ref(ctx, 1);
 	return NULL;
 }

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 14/16] io_uring: support kernel side submission
  2019-01-08 16:56 ` [PATCH 14/16] io_uring: support kernel side submission Jens Axboe
@ 2019-01-09 19:06   ` Christoph Hellwig
  2019-01-09 20:49     ` Jens Axboe
  0 siblings, 1 reply; 39+ messages in thread
From: Christoph Hellwig @ 2019-01-09 19:06 UTC (permalink / raw)
  To: Jens Axboe
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, hch, jmoyer, avi

> +struct iocb_submit {
> +	const struct io_uring_iocb *iocb;
> +	unsigned int index;
> +};
> +
> +struct io_work {
> +	struct work_struct work;
> +	struct io_ring_ctx *ctx;
> +	struct io_uring_iocb iocb;
> +	unsigned iocb_index;
> +};

I think we should use struct iocb_submit everywhere where we pass
an iocb + index, including in the io_work structure.  Here is how
I did that to my old tree:

http://git.infradead.org/users/hch/misc.git/commitdiff/65929ed6f143a1c996305a10a15fd7f64d32595f

Btw, I think we can also remove the separate io_work allocation,
just do the io_get_req in the submission context, and overlay
the io_work onto the iocb in a union of some sort.  This avoids
a whole memory allocation.

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 05/16] Add io_uring IO interface
  2019-01-09 18:30       ` Christoph Hellwig
@ 2019-01-09 20:07         ` Jens Axboe
  0 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-09 20:07 UTC (permalink / raw)
  To: Christoph Hellwig
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, jmoyer, avi

On 1/9/19 11:30 AM, Christoph Hellwig wrote:
> On Wed, Jan 09, 2019 at 08:53:31AM -0700, Jens Axboe wrote:
>>>> +static int io_setup_rw(int rw, const struct io_uring_iocb *iocb,
>>>> +		       struct iovec **iovec, struct iov_iter *iter)
>>>> +{
>>>> +	void __user *buf = (void __user *)(uintptr_t)iocb->addr;
>>>> +	size_t ret;
>>>> +
>>>> +	ret = import_single_range(rw, buf, iocb->len, *iovec, iter);
>>>> +	*iovec = NULL;
>>>> +	return ret;
>>>> +}
>>>
>>> Is there any point in supporting non-vectored operations here?
>>
>> Not sure I follow?
> 
> This version only supports non-vectored read and write, that is
> the equivalent of pread/pwrite.  Many AIO users really need vectored
> operations, that is preadv/pwritev semantics indirecting through
> a struct iovec array.  The non-vectored version can be trivially
> emulated using a vector of 1, which is what we do in the kernel
> I/O stack everywhere.  So I think we should just support the vectored
> version here, and not the non-vectored one.  See my io_uring branch
> for the sketeched implementation.

OK, I see what you mean, so only supported the vectored version.
Probably makes more sense, I'll make the change.

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 11/16] io_uring: batch io_kiocb allocation
  2019-01-09 19:03       ` Christoph Hellwig
@ 2019-01-09 20:08         ` Jens Axboe
  0 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-09 20:08 UTC (permalink / raw)
  To: Christoph Hellwig
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, jmoyer, avi

On 1/9/19 12:03 PM, Christoph Hellwig wrote:
> On Wed, Jan 09, 2019 at 09:57:59AM -0700, Jens Axboe wrote:
>> On 1/9/19 5:13 AM, Christoph Hellwig wrote:
>>>> +	if (!state)
>>>> +		req = kmem_cache_alloc(kiocb_cachep, GFP_KERNEL);
>>>
>>> Just return an error here if kmem_cache_alloc fails.
>>>
>>>> +	if (req)
>>>> +		io_req_init(ctx, req);
>>>
>>> Because all the other ones can't reached this with a NULL req.
>>
>> This is different in the current tree, since I properly fixed the
>> ctx ref issue.
> 
> Your tree does a percpu_ref_tryget very first, and then leaks that if
> kmem_cache_alloc_bulk fails, and also is inconsistent for NULL vs
> ERR_PTR returns.  I think you want something like this on top:

I fixed it up while doing the rebase already. Just haven't pushed anything
out until I've had a chance to run it through testing.

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 14/16] io_uring: support kernel side submission
  2019-01-09 19:06   ` Christoph Hellwig
@ 2019-01-09 20:49     ` Jens Axboe
  0 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-09 20:49 UTC (permalink / raw)
  To: Christoph Hellwig
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, jmoyer, avi

On 1/9/19 12:06 PM, Christoph Hellwig wrote:
>> +struct iocb_submit {
>> +	const struct io_uring_iocb *iocb;
>> +	unsigned int index;
>> +};
>> +
>> +struct io_work {
>> +	struct work_struct work;
>> +	struct io_ring_ctx *ctx;
>> +	struct io_uring_iocb iocb;
>> +	unsigned iocb_index;
>> +};
> 
> I think we should use struct iocb_submit everywhere where we pass
> an iocb + index, including in the io_work structure.  Here is how
> I did that to my old tree:
> 
> http://git.infradead.org/users/hch/misc.git/commitdiff/65929ed6f143a1c996305a10a15fd7f64d32595f

I like that unification.

> Btw, I think we can also remove the separate io_work allocation,
> just do the io_get_req in the submission context, and overlay
> the io_work onto the iocb in a union of some sort.  This avoids
> a whole memory allocation.

I'll take a look at that.

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 03/16] block: add bio_set_polled() helper
  2019-01-08 16:56 ` [PATCH 03/16] block: add bio_set_polled() helper Jens Axboe
@ 2019-01-10  9:43   ` Ming Lei
  2019-01-10 16:05     ` Jens Axboe
  0 siblings, 1 reply; 39+ messages in thread
From: Ming Lei @ 2019-01-10  9:43 UTC (permalink / raw)
  To: Jens Axboe
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, hch, jmoyer, avi

On Tue, Jan 08, 2019 at 09:56:32AM -0700, Jens Axboe wrote:
> For the upcoming async polled IO, we can't sleep allocating requests.
> If we do, then we introduce a deadlock where the submitter already
> has async polled IO in-flight, but can't wait for them to complete
> since polled requests must be active found and reaped.
> 
> Utilize the helper in the blockdev DIRECT_IO code.
> 
> Signed-off-by: Jens Axboe <axboe@kernel.dk>
> ---
>  fs/block_dev.c      |  4 ++--
>  include/linux/bio.h | 14 ++++++++++++++
>  2 files changed, 16 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/block_dev.c b/fs/block_dev.c
> index 5415579f3e14..2ebd2a0d7789 100644
> --- a/fs/block_dev.c
> +++ b/fs/block_dev.c
> @@ -233,7 +233,7 @@ __blkdev_direct_IO_simple(struct kiocb *iocb, struct iov_iter *iter,
>  		task_io_account_write(ret);
>  	}
>  	if (iocb->ki_flags & IOCB_HIPRI)
> -		bio.bi_opf |= REQ_HIPRI;
> +		bio_set_polled(&bio, iocb);
>  
>  	qc = submit_bio(&bio);
>  	for (;;) {
> @@ -401,7 +401,7 @@ __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages)
>  		nr_pages = iov_iter_npages(iter, BIO_MAX_PAGES);
>  		if (!nr_pages) {
>  			if (iocb->ki_flags & IOCB_HIPRI)
> -				bio->bi_opf |= REQ_HIPRI;
> +				bio_set_polled(bio, iocb);
>  
>  			qc = submit_bio(bio);
>  			WRITE_ONCE(iocb->ki_cookie, qc);
> diff --git a/include/linux/bio.h b/include/linux/bio.h
> index 7380b094dcca..f6f0a2b3cbc8 100644
> --- a/include/linux/bio.h
> +++ b/include/linux/bio.h
> @@ -823,5 +823,19 @@ static inline int bio_integrity_add_page(struct bio *bio, struct page *page,
>  
>  #endif /* CONFIG_BLK_DEV_INTEGRITY */
>  
> +/*
> + * Mark a bio as polled. Note that for async polled IO, the caller must
> + * expect -EWOULDBLOCK if we cannot allocate a request (or other resources).
> + * We cannot block waiting for requests on polled IO, as those completions
> + * must be found by the caller. This is different than IRQ driven IO, where
> + * it's safe to wait for IO to complete.
> + */
> +static inline void bio_set_polled(struct bio *bio, struct kiocb *kiocb)
> +{
> +	bio->bi_opf |= REQ_HIPRI;
> +	if (!is_sync_kiocb(kiocb))
> +		bio->bi_opf |= REQ_NOWAIT;
> +}
> +

REQ_NOWAIT doesn't cover allocating split bio, is that a issue?

BTW, could you explain a bit about the deadlock in case of sleep from
request allocation?

Thanks,
Ming

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: [PATCH 03/16] block: add bio_set_polled() helper
  2019-01-10  9:43   ` Ming Lei
@ 2019-01-10 16:05     ` Jens Axboe
  0 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-10 16:05 UTC (permalink / raw)
  To: Ming Lei
  Cc: linux-fsdevel, linux-aio, linux-block, linux-arch, hch, jmoyer, avi

On 1/10/19 2:43 AM, Ming Lei wrote:
> On Tue, Jan 08, 2019 at 09:56:32AM -0700, Jens Axboe wrote:
>> For the upcoming async polled IO, we can't sleep allocating requests.
>> If we do, then we introduce a deadlock where the submitter already
>> has async polled IO in-flight, but can't wait for them to complete
>> since polled requests must be active found and reaped.
>>
>> Utilize the helper in the blockdev DIRECT_IO code.
>>
>> Signed-off-by: Jens Axboe <axboe@kernel.dk>
>> ---
>>  fs/block_dev.c      |  4 ++--
>>  include/linux/bio.h | 14 ++++++++++++++
>>  2 files changed, 16 insertions(+), 2 deletions(-)
>>
>> diff --git a/fs/block_dev.c b/fs/block_dev.c
>> index 5415579f3e14..2ebd2a0d7789 100644
>> --- a/fs/block_dev.c
>> +++ b/fs/block_dev.c
>> @@ -233,7 +233,7 @@ __blkdev_direct_IO_simple(struct kiocb *iocb, struct iov_iter *iter,
>>  		task_io_account_write(ret);
>>  	}
>>  	if (iocb->ki_flags & IOCB_HIPRI)
>> -		bio.bi_opf |= REQ_HIPRI;
>> +		bio_set_polled(&bio, iocb);
>>  
>>  	qc = submit_bio(&bio);
>>  	for (;;) {
>> @@ -401,7 +401,7 @@ __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages)
>>  		nr_pages = iov_iter_npages(iter, BIO_MAX_PAGES);
>>  		if (!nr_pages) {
>>  			if (iocb->ki_flags & IOCB_HIPRI)
>> -				bio->bi_opf |= REQ_HIPRI;
>> +				bio_set_polled(bio, iocb);
>>  
>>  			qc = submit_bio(bio);
>>  			WRITE_ONCE(iocb->ki_cookie, qc);
>> diff --git a/include/linux/bio.h b/include/linux/bio.h
>> index 7380b094dcca..f6f0a2b3cbc8 100644
>> --- a/include/linux/bio.h
>> +++ b/include/linux/bio.h
>> @@ -823,5 +823,19 @@ static inline int bio_integrity_add_page(struct bio *bio, struct page *page,
>>  
>>  #endif /* CONFIG_BLK_DEV_INTEGRITY */
>>  
>> +/*
>> + * Mark a bio as polled. Note that for async polled IO, the caller must
>> + * expect -EWOULDBLOCK if we cannot allocate a request (or other resources).
>> + * We cannot block waiting for requests on polled IO, as those completions
>> + * must be found by the caller. This is different than IRQ driven IO, where
>> + * it's safe to wait for IO to complete.
>> + */
>> +static inline void bio_set_polled(struct bio *bio, struct kiocb *kiocb)
>> +{
>> +	bio->bi_opf |= REQ_HIPRI;
>> +	if (!is_sync_kiocb(kiocb))
>> +		bio->bi_opf |= REQ_NOWAIT;
>> +}
>> +
> 
> REQ_NOWAIT doesn't cover allocating split bio, is that a issue?

Yes, that might be an issue. I'll look into what we should do about that,
for now it's not a huge problem.

> BTW, could you explain a bit about the deadlock in case of sleep from
> request allocation?

It's more a live lock I guess, but the issue is that with polled IO, we don't
get an IRQ. For normal IO, if you run out, you can just sleep and wait for
an IRQ to come in, trigger a completion (or multiple), which will then wake
you up. For polled IO, you have to find those completions. Hence if you just
go to sleep, nobody is going to find those completions for you. You'll then
be waiting forever for an event, that will never trigger.

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 04/16] iomap: wire up the iopoll method
       [not found] <20190115025531.13985-1-axboe@kernel.dk>
@ 2019-01-15  2:55 ` Jens Axboe
  0 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-15  2:55 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

From: Christoph Hellwig <hch@lst.de>

Store the request queue the last bio was submitted to in the iocb
private data in addition to the cookie so that we find the right block
device.  Also refactor the common direct I/O bio submission code into a
nice little helper.

Signed-off-by: Christoph Hellwig <hch@lst.de>

Modified to use bio_set_polled().

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/gfs2/file.c        |  2 ++
 fs/iomap.c            | 43 ++++++++++++++++++++++++++++---------------
 fs/xfs/xfs_file.c     |  1 +
 include/linux/iomap.h |  1 +
 4 files changed, 32 insertions(+), 15 deletions(-)

diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c
index a2dea5bc0427..58a768e59712 100644
--- a/fs/gfs2/file.c
+++ b/fs/gfs2/file.c
@@ -1280,6 +1280,7 @@ const struct file_operations gfs2_file_fops = {
 	.llseek		= gfs2_llseek,
 	.read_iter	= gfs2_file_read_iter,
 	.write_iter	= gfs2_file_write_iter,
+	.iopoll		= iomap_dio_iopoll,
 	.unlocked_ioctl	= gfs2_ioctl,
 	.mmap		= gfs2_mmap,
 	.open		= gfs2_open,
@@ -1310,6 +1311,7 @@ const struct file_operations gfs2_file_fops_nolock = {
 	.llseek		= gfs2_llseek,
 	.read_iter	= gfs2_file_read_iter,
 	.write_iter	= gfs2_file_write_iter,
+	.iopoll		= iomap_dio_iopoll,
 	.unlocked_ioctl	= gfs2_ioctl,
 	.mmap		= gfs2_mmap,
 	.open		= gfs2_open,
diff --git a/fs/iomap.c b/fs/iomap.c
index a3088fae567b..4ee50b76b4a1 100644
--- a/fs/iomap.c
+++ b/fs/iomap.c
@@ -1454,6 +1454,28 @@ struct iomap_dio {
 	};
 };
 
+int iomap_dio_iopoll(struct kiocb *kiocb, bool spin)
+{
+	struct request_queue *q = READ_ONCE(kiocb->private);
+
+	if (!q)
+		return 0;
+	return blk_poll(q, READ_ONCE(kiocb->ki_cookie), spin);
+}
+EXPORT_SYMBOL_GPL(iomap_dio_iopoll);
+
+static void iomap_dio_submit_bio(struct iomap_dio *dio, struct iomap *iomap,
+		struct bio *bio)
+{
+	atomic_inc(&dio->ref);
+
+	if (dio->iocb->ki_flags & IOCB_HIPRI)
+		bio_set_polled(bio, dio->iocb);
+
+	dio->submit.last_queue = bdev_get_queue(iomap->bdev);
+	dio->submit.cookie = submit_bio(bio);
+}
+
 static ssize_t iomap_dio_complete(struct iomap_dio *dio)
 {
 	struct kiocb *iocb = dio->iocb;
@@ -1566,7 +1588,7 @@ static void iomap_dio_bio_end_io(struct bio *bio)
 	}
 }
 
-static blk_qc_t
+static void
 iomap_dio_zero(struct iomap_dio *dio, struct iomap *iomap, loff_t pos,
 		unsigned len)
 {
@@ -1580,15 +1602,10 @@ iomap_dio_zero(struct iomap_dio *dio, struct iomap *iomap, loff_t pos,
 	bio->bi_private = dio;
 	bio->bi_end_io = iomap_dio_bio_end_io;
 
-	if (dio->iocb->ki_flags & IOCB_HIPRI)
-		flags |= REQ_HIPRI;
-
 	get_page(page);
 	__bio_add_page(bio, page, len, 0);
 	bio_set_op_attrs(bio, REQ_OP_WRITE, flags);
-
-	atomic_inc(&dio->ref);
-	return submit_bio(bio);
+	iomap_dio_submit_bio(dio, iomap, bio);
 }
 
 static loff_t
@@ -1691,9 +1708,6 @@ iomap_dio_bio_actor(struct inode *inode, loff_t pos, loff_t length,
 				bio_set_pages_dirty(bio);
 		}
 
-		if (dio->iocb->ki_flags & IOCB_HIPRI)
-			bio->bi_opf |= REQ_HIPRI;
-
 		iov_iter_advance(dio->submit.iter, n);
 
 		dio->size += n;
@@ -1701,11 +1715,7 @@ iomap_dio_bio_actor(struct inode *inode, loff_t pos, loff_t length,
 		copied += n;
 
 		nr_pages = iov_iter_npages(&iter, BIO_MAX_PAGES);
-
-		atomic_inc(&dio->ref);
-
-		dio->submit.last_queue = bdev_get_queue(iomap->bdev);
-		dio->submit.cookie = submit_bio(bio);
+		iomap_dio_submit_bio(dio, iomap, bio);
 	} while (nr_pages);
 
 	/*
@@ -1916,6 +1926,9 @@ iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter,
 	if (dio->flags & IOMAP_DIO_WRITE_FUA)
 		dio->flags &= ~IOMAP_DIO_NEED_SYNC;
 
+	WRITE_ONCE(iocb->ki_cookie, dio->submit.cookie);
+	WRITE_ONCE(iocb->private, dio->submit.last_queue);
+
 	if (!atomic_dec_and_test(&dio->ref)) {
 		if (!dio->wait_for_completion)
 			return -EIOCBQUEUED;
diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index e47425071e65..60c2da41f0fc 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -1203,6 +1203,7 @@ const struct file_operations xfs_file_operations = {
 	.write_iter	= xfs_file_write_iter,
 	.splice_read	= generic_file_splice_read,
 	.splice_write	= iter_file_splice_write,
+	.iopoll		= iomap_dio_iopoll,
 	.unlocked_ioctl	= xfs_file_ioctl,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl	= xfs_file_compat_ioctl,
diff --git a/include/linux/iomap.h b/include/linux/iomap.h
index 9a4258154b25..0fefb5455bda 100644
--- a/include/linux/iomap.h
+++ b/include/linux/iomap.h
@@ -162,6 +162,7 @@ typedef int (iomap_dio_end_io_t)(struct kiocb *iocb, ssize_t ret,
 		unsigned flags);
 ssize_t iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter,
 		const struct iomap_ops *ops, iomap_dio_end_io_t end_io);
+int iomap_dio_iopoll(struct kiocb *kiocb, bool spin);
 
 #ifdef CONFIG_SWAP
 struct file;
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

* [PATCH 04/16] iomap: wire up the iopoll method
  2019-01-12 21:29 [PATCHSET v3] " Jens Axboe
@ 2019-01-12 21:29 ` Jens Axboe
  0 siblings, 0 replies; 39+ messages in thread
From: Jens Axboe @ 2019-01-12 21:29 UTC (permalink / raw)
  To: linux-fsdevel, linux-aio, linux-block, linux-arch
  Cc: hch, jmoyer, avi, Jens Axboe

From: Christoph Hellwig <hch@lst.de>

Store the request queue the last bio was submitted to in the iocb
private data in addition to the cookie so that we find the right block
device.  Also refactor the common direct I/O bio submission code into a
nice little helper.

Signed-off-by: Christoph Hellwig <hch@lst.de>

Modified to use bio_set_polled().

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/gfs2/file.c        |  2 ++
 fs/iomap.c            | 43 ++++++++++++++++++++++++++++---------------
 fs/xfs/xfs_file.c     |  1 +
 include/linux/iomap.h |  1 +
 4 files changed, 32 insertions(+), 15 deletions(-)

diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c
index a2dea5bc0427..58a768e59712 100644
--- a/fs/gfs2/file.c
+++ b/fs/gfs2/file.c
@@ -1280,6 +1280,7 @@ const struct file_operations gfs2_file_fops = {
 	.llseek		= gfs2_llseek,
 	.read_iter	= gfs2_file_read_iter,
 	.write_iter	= gfs2_file_write_iter,
+	.iopoll		= iomap_dio_iopoll,
 	.unlocked_ioctl	= gfs2_ioctl,
 	.mmap		= gfs2_mmap,
 	.open		= gfs2_open,
@@ -1310,6 +1311,7 @@ const struct file_operations gfs2_file_fops_nolock = {
 	.llseek		= gfs2_llseek,
 	.read_iter	= gfs2_file_read_iter,
 	.write_iter	= gfs2_file_write_iter,
+	.iopoll		= iomap_dio_iopoll,
 	.unlocked_ioctl	= gfs2_ioctl,
 	.mmap		= gfs2_mmap,
 	.open		= gfs2_open,
diff --git a/fs/iomap.c b/fs/iomap.c
index a3088fae567b..4ee50b76b4a1 100644
--- a/fs/iomap.c
+++ b/fs/iomap.c
@@ -1454,6 +1454,28 @@ struct iomap_dio {
 	};
 };
 
+int iomap_dio_iopoll(struct kiocb *kiocb, bool spin)
+{
+	struct request_queue *q = READ_ONCE(kiocb->private);
+
+	if (!q)
+		return 0;
+	return blk_poll(q, READ_ONCE(kiocb->ki_cookie), spin);
+}
+EXPORT_SYMBOL_GPL(iomap_dio_iopoll);
+
+static void iomap_dio_submit_bio(struct iomap_dio *dio, struct iomap *iomap,
+		struct bio *bio)
+{
+	atomic_inc(&dio->ref);
+
+	if (dio->iocb->ki_flags & IOCB_HIPRI)
+		bio_set_polled(bio, dio->iocb);
+
+	dio->submit.last_queue = bdev_get_queue(iomap->bdev);
+	dio->submit.cookie = submit_bio(bio);
+}
+
 static ssize_t iomap_dio_complete(struct iomap_dio *dio)
 {
 	struct kiocb *iocb = dio->iocb;
@@ -1566,7 +1588,7 @@ static void iomap_dio_bio_end_io(struct bio *bio)
 	}
 }
 
-static blk_qc_t
+static void
 iomap_dio_zero(struct iomap_dio *dio, struct iomap *iomap, loff_t pos,
 		unsigned len)
 {
@@ -1580,15 +1602,10 @@ iomap_dio_zero(struct iomap_dio *dio, struct iomap *iomap, loff_t pos,
 	bio->bi_private = dio;
 	bio->bi_end_io = iomap_dio_bio_end_io;
 
-	if (dio->iocb->ki_flags & IOCB_HIPRI)
-		flags |= REQ_HIPRI;
-
 	get_page(page);
 	__bio_add_page(bio, page, len, 0);
 	bio_set_op_attrs(bio, REQ_OP_WRITE, flags);
-
-	atomic_inc(&dio->ref);
-	return submit_bio(bio);
+	iomap_dio_submit_bio(dio, iomap, bio);
 }
 
 static loff_t
@@ -1691,9 +1708,6 @@ iomap_dio_bio_actor(struct inode *inode, loff_t pos, loff_t length,
 				bio_set_pages_dirty(bio);
 		}
 
-		if (dio->iocb->ki_flags & IOCB_HIPRI)
-			bio->bi_opf |= REQ_HIPRI;
-
 		iov_iter_advance(dio->submit.iter, n);
 
 		dio->size += n;
@@ -1701,11 +1715,7 @@ iomap_dio_bio_actor(struct inode *inode, loff_t pos, loff_t length,
 		copied += n;
 
 		nr_pages = iov_iter_npages(&iter, BIO_MAX_PAGES);
-
-		atomic_inc(&dio->ref);
-
-		dio->submit.last_queue = bdev_get_queue(iomap->bdev);
-		dio->submit.cookie = submit_bio(bio);
+		iomap_dio_submit_bio(dio, iomap, bio);
 	} while (nr_pages);
 
 	/*
@@ -1916,6 +1926,9 @@ iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter,
 	if (dio->flags & IOMAP_DIO_WRITE_FUA)
 		dio->flags &= ~IOMAP_DIO_NEED_SYNC;
 
+	WRITE_ONCE(iocb->ki_cookie, dio->submit.cookie);
+	WRITE_ONCE(iocb->private, dio->submit.last_queue);
+
 	if (!atomic_dec_and_test(&dio->ref)) {
 		if (!dio->wait_for_completion)
 			return -EIOCBQUEUED;
diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index e47425071e65..60c2da41f0fc 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -1203,6 +1203,7 @@ const struct file_operations xfs_file_operations = {
 	.write_iter	= xfs_file_write_iter,
 	.splice_read	= generic_file_splice_read,
 	.splice_write	= iter_file_splice_write,
+	.iopoll		= iomap_dio_iopoll,
 	.unlocked_ioctl	= xfs_file_ioctl,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl	= xfs_file_compat_ioctl,
diff --git a/include/linux/iomap.h b/include/linux/iomap.h
index 9a4258154b25..0fefb5455bda 100644
--- a/include/linux/iomap.h
+++ b/include/linux/iomap.h
@@ -162,6 +162,7 @@ typedef int (iomap_dio_end_io_t)(struct kiocb *iocb, ssize_t ret,
 		unsigned flags);
 ssize_t iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter,
 		const struct iomap_ops *ops, iomap_dio_end_io_t end_io);
+int iomap_dio_iopoll(struct kiocb *kiocb, bool spin);
 
 #ifdef CONFIG_SWAP
 struct file;
-- 
2.17.1


^ permalink raw reply	[flat|nested] 39+ messages in thread

end of thread, back to index

Thread overview: 39+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-01-08 16:56 [PATCHSET v1] io_uring IO interface Jens Axboe
2019-01-08 16:56 ` [PATCH 01/16] fs: add an iopoll method to struct file_operations Jens Axboe
2019-01-08 16:56 ` [PATCH 02/16] block: wire up block device iopoll method Jens Axboe
2019-01-08 16:56 ` [PATCH 03/16] block: add bio_set_polled() helper Jens Axboe
2019-01-10  9:43   ` Ming Lei
2019-01-10 16:05     ` Jens Axboe
2019-01-08 16:56 ` [PATCH 04/16] iomap: wire up the iopoll method Jens Axboe
2019-01-08 16:56 ` [PATCH 05/16] Add io_uring IO interface Jens Axboe
2019-01-09 12:10   ` Christoph Hellwig
2019-01-09 15:53     ` Jens Axboe
2019-01-09 18:30       ` Christoph Hellwig
2019-01-09 20:07         ` Jens Axboe
2019-01-08 16:56 ` [PATCH 06/16] io_uring: support for IO polling Jens Axboe
2019-01-09 12:11   ` Christoph Hellwig
2019-01-09 15:53     ` Jens Axboe
2019-01-08 16:56 ` [PATCH 07/16] io_uring: add submission side request cache Jens Axboe
2019-01-08 16:56 ` [PATCH 08/16] fs: add fget_many() and fput_many() Jens Axboe
2019-01-08 16:56 ` [PATCH 09/16] io_uring: use fget/fput_many() for file references Jens Axboe
2019-01-08 16:56 ` [PATCH 10/16] io_uring: split kiocb init from allocation Jens Axboe
2019-01-09 12:12   ` Christoph Hellwig
2019-01-09 16:56     ` Jens Axboe
2019-01-08 16:56 ` [PATCH 11/16] io_uring: batch io_kiocb allocation Jens Axboe
2019-01-09 12:13   ` Christoph Hellwig
2019-01-09 16:57     ` Jens Axboe
2019-01-09 19:03       ` Christoph Hellwig
2019-01-09 20:08         ` Jens Axboe
2019-01-08 16:56 ` [PATCH 12/16] block: implement bio helper to add iter bvec pages to bio Jens Axboe
2019-01-08 16:56 ` [PATCH 13/16] io_uring: add support for pre-mapped user IO buffers Jens Axboe
2019-01-09 12:16   ` Christoph Hellwig
2019-01-09 17:06     ` Jens Axboe
2019-01-08 16:56 ` [PATCH 14/16] io_uring: support kernel side submission Jens Axboe
2019-01-09 19:06   ` Christoph Hellwig
2019-01-09 20:49     ` Jens Axboe
2019-01-08 16:56 ` [PATCH 15/16] io_uring: add submission polling Jens Axboe
2019-01-08 16:56 ` [PATCH 16/16] io_uring: add io_uring_event cache hit information Jens Axboe
2019-01-09 16:00 ` [PATCHSET v1] io_uring IO interface Matthew Wilcox
2019-01-09 16:27   ` Chris Mason
2019-01-12 21:29 [PATCHSET v3] " Jens Axboe
2019-01-12 21:29 ` [PATCH 04/16] iomap: wire up the iopoll method Jens Axboe
     [not found] <20190115025531.13985-1-axboe@kernel.dk>
2019-01-15  2:55 ` Jens Axboe

Linux-Block Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-block/0 linux-block/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-block linux-block/ https://lore.kernel.org/linux-block \
		linux-block@vger.kernel.org linux-block@archiver.kernel.org
	public-inbox-index linux-block


Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-block


AGPL code for this site: git clone https://public-inbox.org/ public-inbox