Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1185975 > unrolled thread

[PATCH v7 0/6] block: loop: improve loop with AIO

Back to article view | Back to linux.kernel

Contents

  [PATCH v7 0/6] block: loop: improve loop with AIO Ming Lei <ming.lei@canonical.com> - 2015-07-16 17:40 +0200
    [PATCH v7 5/6] block: loop: introduce ioctl command of LOOP_SET_DIRECT_IO Ming Lei <ming.lei@canonical.com> - 2015-07-16 17:40 +0200
    [PATCH v7 4/6] block: loop: prepare for supporing direct IO Ming Lei <ming.lei@canonical.com> - 2015-07-16 17:40 +0200
    [PATCH v7 2/6] block: loop: set QUEUE_FLAG_NOMERGES for request queue of loop Ming Lei <ming.lei@canonical.com> - 2015-07-16 17:40 +0200
    [PATCH v7 1/6] fs: direct-io: don't dirtying pages for ITER_BVEC/ITER_KVEC direct read Ming Lei <ming.lei@canonical.com> - 2015-07-16 17:40 +0200
      Re: [PATCH v7 1/6] fs: direct-io: don't dirtying pages for  ITER_BVEC/ITER_KVEC direct read Dave Kleikamp <dave.kleikamp@oracle.com> - 2015-07-16 18:50 +0200
    [PATCH v7 6/6] block: loop: support DIO & AIO Ming Lei <ming.lei@canonical.com> - 2015-07-16 17:40 +0200

#1185975 — [PATCH v7 0/6] block: loop: improve loop with AIO

Hi Guys,

There are about 3 advantages to use direct I/O and AIO on
read/write loop's backing file:

1) double cache can be avoided, then memory usage gets
decreased a lot

2) not like user space direct I/O, there isn't cost of
pinning pages

3) avoid context switch for obtaining good throughput
- in buffered file read, random I/O throughput is often obtained
only if they are submitted concurrently from lots of tasks; but for
sequential I/O, most of times they can be hit from page cache, so
concurrent submissions often introduce unnecessary context switch
and can't improve throughput much. There was such discussion[1]
to use non-blocking I/O to improve the problem for application.
- with direct I/O and AIO, concurrent submissions can be
avoided and random read throughput can't be affected meantime

So this patchset trys to improve loop via AIO, and about 45% memory
usage can be decreased, see detailed data in commit log of patch4,
also IO throughput isn't affected too.

V7:
        - only 4/6 and 5/6 updated
        - update direct io after lo->offset is changed(4/6)
        - fix updating LO_FLAGS_DIRECT_IO in __loop_update_dio()(4/6)
        - introduce ioctl cmd of LOOP_SET_DIRECT_IO(5/6) for case of
          'mount -o loop'
        - losetup git tree:
            http://kernel.ubuntu.com/git/ming/util-linux.git/log/?h=loop-dio-v7
        - how to set direct io
            losetup --direct-io 1 /dev/loopN 
        - how to clear direct io
            losetup --direct-io 0 /dev/loopN 
        - how to show if direct io is used for accessing backing file
            losetup -l
V6:
        - only patch 4 and patch 5 get updated
        - check lo->lo_offset to decide if direct IO can be supported(4/5)
        - introduce one flag for userspace(losetup) to keep updated
        if using direct I/O to access backing file(4/5)
        - implement patches for util-linux(losetup) so that losetup can
        enable direct I/O feature:(4/5)

           http://kernel.ubuntu.com/git/ming/util-linux.git/log/?h=losetup-dio

        - remove the direct IO control interface from sysfs(4/5)
        - handle partial read in case of direct read (5/5)
        - add more comments for direct IO (5/5)
V5:
        - don't introduce IOCB_DONT_DIRTY_PAGE and bypass dirtying
        for ITER_KVEC and ITER_BVEC direct IO(read), as required by
        Christoph

V4:
        - add detailed commit log for 'use kthread_work'
        - allow userspace(sysfs, losetup) to decide if dio/aio is
        used as suggested by Christoph and Dave Chinner
        - only use dio if the backing block device's min io size
        is 512 as pointed by Dave Chinner & Christoph
V3:
        - based on Al's iov_iter work and Christoph's kiocb changes
        - use kthread_work
        - introduce IOCB_DONT_DIRTY_PAGE flag
        - set QUEUE_FLAG_NOMERGES for loop's request queue
V2:
        - remove 'extra' parameter to aio_kernel_alloc()
        - try to avoid memory allcation inside queue req callback
        - introduce 'use_mq' sysfs file for enabling kernel aio or disabling it
V1:
        - link:
                http://marc.info/?t=140803157700004&r=1&w=2
        - improve failure path in aio_kernel_submit()

 drivers/block/loop.c      | 263 ++++++++++++++++++++++++++++++++++++----------
 drivers/block/loop.h      |  13 +--
 fs/direct-io.c            |   9 +-
 include/uapi/linux/loop.h |   2 +
 4 files changed, 221 insertions(+), 66 deletions(-)


Thanks,
Ming


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [next] | [standalone]

#1185978 — [PATCH v7 5/6] block: loop: introduce ioctl command of LOOP_SET_DIRECT_IO

If loop block is mounted via 'mount -o loop', it isn't easy
to pass file descriptor opened as O_DIRECT, so this patch
introduces a new command to support direct IO for this case.

Cc: linux-api@vger.kernel.org
Signed-off-by: Ming Lei <ming.lei@canonical.com>
---
 drivers/block/loop.c      | 19 +++++++++++++++++++
 include/uapi/linux/loop.h |  1 +
 2 files changed, 20 insertions(+)

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index 35aa3dd..3f26a0a 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -1214,6 +1214,20 @@ static int loop_set_capacity(struct loop_device *lo, struct block_device *bdev)
 	return figure_loop_size(lo, lo->lo_offset, lo->lo_sizelimit);
 }
 
+static int loop_set_dio(struct loop_device *lo, unsigned long arg)
+{
+	int error = -ENXIO;
+	if (lo->lo_state != Lo_bound)
+		goto out;
+
+	__loop_update_dio(lo, !!arg);
+	if (lo->use_dio == !!arg)
+		return 0;
+	error = -EINVAL;
+ out:
+	return error;
+}
+
 static int lo_ioctl(struct block_device *bdev, fmode_t mode,
 	unsigned int cmd, unsigned long arg)
 {
@@ -1257,6 +1271,11 @@ static int lo_ioctl(struct block_device *bdev, fmode_t mode,
 		if ((mode & FMODE_WRITE) || capable(CAP_SYS_ADMIN))
 			err = loop_set_capacity(lo, bdev);
 		break;
+	case LOOP_SET_DIRECT_IO:
+		err = -EPERM;
+		if ((mode & FMODE_WRITE) || capable(CAP_SYS_ADMIN))
+			err = loop_set_dio(lo, arg);
+		break;
 	default:
 		err = lo->ioctl ? lo->ioctl(lo, cmd, arg) : -EINVAL;
 	}
diff --git a/include/uapi/linux/loop.h b/include/uapi/linux/loop.h
index 949851c..c8125ec 100644
--- a/include/uapi/linux/loop.h
+++ b/include/uapi/linux/loop.h
@@ -87,6 +87,7 @@ struct loop_info64 {
 #define LOOP_GET_STATUS64	0x4C05
 #define LOOP_CHANGE_FD		0x4C06
 #define LOOP_SET_CAPACITY	0x4C07
+#define LOOP_SET_DIRECT_IO	0x4C08
 
 /* /dev/loop-control interface */
 #define LOOP_CTL_ADD		0x4C80
-- 
1.9.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1185983 — [PATCH v7 4/6] block: loop: prepare for supporing direct IO

This patches provides one interface for enabling direct IO
from user space:

	- userspace(such as losetup) can pass 'file' which is
	opened/fcntl as O_DIRECT

Also __loop_update_dio() is introduced to check if direct I/O
can be used on current loop setting.

The last big change is to introduce LO_FLAGS_DIRECT_IO flag
for userspace to know if direct IO is used to access backing
file.

Cc: linux-api@vger.kernel.org
Signed-off-by: Ming Lei <ming.lei@canonical.com>
---
 drivers/block/loop.c      | 65 ++++++++++++++++++++++++++++++++++++++++++++++-
 drivers/block/loop.h      |  1 +
 include/uapi/linux/loop.h |  1 +
 3 files changed, 66 insertions(+), 1 deletion(-)

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index ce94b92..35aa3dd 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -164,6 +164,49 @@ static loff_t get_loop_size(struct loop_device *lo, struct file *file)
 	return get_size(lo->lo_offset, lo->lo_sizelimit, file);
 }
 
+static void __loop_update_dio(struct loop_device *lo, bool dio)
+{
+	struct file *file = lo->lo_backing_file;
+	struct inode *inode = file->f_mapping->host;
+	bool use_dio;
+
+	/*
+	 * loop block's logical block size is 512, now
+	 * we support direct I/O only if the backing
+	 * block devices' minimize I/O size is 512 and
+	 * the offset is aligned with 512.
+	 */
+	if (dio) {
+		if (inode->i_sb->s_bdev &&
+			bdev_io_min(inode->i_sb->s_bdev) == 512 &&
+			!(lo->lo_offset & 511))
+			use_dio = true;
+		else
+			use_dio = false;
+	} else {
+		use_dio = false;
+	}
+
+	if (lo->use_dio == use_dio)
+		return;
+
+	/* flush dirty pages before changing direct IO */
+	vfs_fsync(file, 0);
+
+	/*
+	 * The flag of LO_FLAGS_DIRECT_IO is handled similarly with
+	 * LO_FLAGS_READ_ONLY, both are set from kernel, and losetup
+	 * will get updated by ioctl(LOOP_GET_STATUS)
+	 */
+	blk_mq_freeze_queue(lo->lo_queue);
+	lo->use_dio = use_dio;
+	if (use_dio)
+		lo->lo_flags |= LO_FLAGS_DIRECT_IO;
+	else
+		lo->lo_flags &= ~LO_FLAGS_DIRECT_IO;
+	blk_mq_unfreeze_queue(lo->lo_queue);
+}
+
 static int
 figure_loop_size(struct loop_device *lo, loff_t offset, loff_t sizelimit)
 {
@@ -173,8 +216,12 @@ figure_loop_size(struct loop_device *lo, loff_t offset, loff_t sizelimit)
 
 	if (unlikely((loff_t)x != size))
 		return -EFBIG;
-	if (lo->lo_offset != offset)
+	if (lo->lo_offset != offset) {
 		lo->lo_offset = offset;
+
+		/* update dio if lo_offset is changed*/
+		__loop_update_dio(lo, lo->use_dio);
+	}
 	if (lo->lo_sizelimit != sizelimit)
 		lo->lo_sizelimit = sizelimit;
 	set_capacity(lo->lo_disk, x);
@@ -421,6 +468,11 @@ struct switch_request {
 	struct completion wait;
 };
 
+static inline void loop_update_dio(struct loop_device *lo)
+{
+	__loop_update_dio(lo, io_is_direct(lo->lo_backing_file));
+}
+
 /*
  * Do the actual switch; called from the BIO completion routine
  */
@@ -441,6 +493,7 @@ static void do_loop_switch(struct loop_device *lo, struct switch_request *p)
 		mapping->host->i_bdev->bd_block_size : PAGE_SIZE;
 	lo->old_gfp_mask = mapping_gfp_mask(mapping);
 	mapping_set_gfp_mask(mapping, lo->old_gfp_mask & ~(__GFP_IO|__GFP_FS));
+	loop_update_dio(lo);
 }
 
 /*
@@ -627,11 +680,19 @@ static ssize_t loop_attr_partscan_show(struct loop_device *lo, char *buf)
 	return sprintf(buf, "%s\n", partscan ? "1" : "0");
 }
 
+static ssize_t loop_attr_dio_show(struct loop_device *lo, char *buf)
+{
+	int dio = (lo->lo_flags & LO_FLAGS_DIRECT_IO);
+
+	return sprintf(buf, "%s\n", dio ? "1" : "0");
+}
+
 LOOP_ATTR_RO(backing_file);
 LOOP_ATTR_RO(offset);
 LOOP_ATTR_RO(sizelimit);
 LOOP_ATTR_RO(autoclear);
 LOOP_ATTR_RO(partscan);
+LOOP_ATTR_RO(dio);
 
 static struct attribute *loop_attrs[] = {
 	&loop_attr_backing_file.attr,
@@ -639,6 +700,7 @@ static struct attribute *loop_attrs[] = {
 	&loop_attr_sizelimit.attr,
 	&loop_attr_autoclear.attr,
 	&loop_attr_partscan.attr,
+	&loop_attr_dio.attr,
 	NULL,
 };
 
@@ -783,6 +845,7 @@ static int loop_set_fd(struct loop_device *lo, fmode_t mode,
 	if (!(lo_flags & LO_FLAGS_READ_ONLY) && file->f_op->fsync)
 		blk_queue_flush(lo->lo_queue, REQ_FLUSH);
 
+	loop_update_dio(lo);
 	set_capacity(lo->lo_disk, size);
 	bd_set_size(bdev, size << 9);
 	loop_sysfs_init(lo);
diff --git a/drivers/block/loop.h b/drivers/block/loop.h
index b6c7d21..d1de221 100644
--- a/drivers/block/loop.h
+++ b/drivers/block/loop.h
@@ -58,6 +58,7 @@ struct loop_device {
 	struct mutex		lo_ctl_mutex;
 	struct kthread_worker	worker;
 	struct task_struct	*worker_task;
+	bool			use_dio;
 
 	struct request_queue	*lo_queue;
 	struct blk_mq_tag_set	tag_set;
diff --git a/include/uapi/linux/loop.h b/include/uapi/linux/loop.h
index e0cecd2..949851c 100644
--- a/include/uapi/linux/loop.h
+++ b/include/uapi/linux/loop.h
@@ -21,6 +21,7 @@ enum {
 	LO_FLAGS_READ_ONLY	= 1,
 	LO_FLAGS_AUTOCLEAR	= 4,
 	LO_FLAGS_PARTSCAN	= 8,
+	LO_FLAGS_DIRECT_IO	= 16,
 };
 
 #include <asm/posix_types.h>	/* for __kernel_old_dev_t */
-- 
1.9.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1185991 — [PATCH v7 2/6] block: loop: set QUEUE_FLAG_NOMERGES for request queue of loop

It doesn't make sense to enable merge because the I/O
submitted to backing file is handled page by page.

Signed-off-by: Ming Lei <ming.lei@canonical.com>
---
 drivers/block/loop.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index f7a4c9d..e4381f8 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -1595,6 +1595,12 @@ static int loop_add(struct loop_device **l, int i)
 	}
 	lo->lo_queue->queuedata = lo;
 
+	/*
+	 * It doesn't make sense to enable merge because the I/O
+	 * submitted to backing file is handled page by page.
+	 */
+	queue_flag_set_unlocked(QUEUE_FLAG_NOMERGES, lo->lo_queue);
+
 	INIT_LIST_HEAD(&lo->write_cmd_head);
 	INIT_WORK(&lo->write_work, loop_queue_write_work);
 
-- 
1.9.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1185997 — [PATCH v7 1/6] fs: direct-io: don't dirtying pages for ITER_BVEC/ITER_KVEC direct read

When direct read IO is submitted from kernel, it is often
unnecessary to dirty pages, for example of loop, dirtying pages
have been considered in the upper filesystem(over loop) side
already, and they don't need to be dirtied again.

So this patch doesn't dirtying pages for ITER_BVEC/ITER_KVEC
direct read, and loop should be the 1st case to use ITER_BVEC/ITER_KVEC
for direct read I/O.

The patch is based on previous Dave's patch.

Cc: Dave Kleikamp <dave.kleikamp@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ming Lei <ming.lei@canonical.com>
---
 fs/direct-io.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/fs/direct-io.c b/fs/direct-io.c
index 745d234..07e4b28 100644
--- a/fs/direct-io.c
+++ b/fs/direct-io.c
@@ -120,6 +120,7 @@ struct dio {
 	int page_errors;		/* errno from get_user_pages() */
 	int is_async;			/* is IO async ? */
 	bool defer_completion;		/* defer AIO completion to workqueue? */
+	bool should_dirty;		/* if pages should be dirtied */
 	int io_error;			/* IO error in completion path */
 	unsigned long refcount;		/* direct_io_worker() and bios */
 	struct bio *bio_list;		/* singly linked via bi_private */
@@ -393,7 +394,7 @@ static inline void dio_bio_submit(struct dio *dio, struct dio_submit *sdio)
 	dio->refcount++;
 	spin_unlock_irqrestore(&dio->bio_lock, flags);
 
-	if (dio->is_async && dio->rw == READ)
+	if (dio->is_async && dio->rw == READ && dio->should_dirty)
 		bio_set_pages_dirty(bio);
 
 	if (sdio->submit_io)
@@ -464,13 +465,14 @@ static int dio_bio_complete(struct dio *dio, struct bio *bio)
 	if (!uptodate)
 		dio->io_error = -EIO;
 
-	if (dio->is_async && dio->rw == READ) {
+	if (dio->is_async && dio->rw == READ && dio->should_dirty) {
 		bio_check_pages_dirty(bio);	/* transfers ownership */
 	} else {
 		bio_for_each_segment_all(bvec, bio, i) {
 			struct page *page = bvec->bv_page;
 
-			if (dio->rw == READ && !PageCompound(page))
+			if (dio->rw == READ && !PageCompound(page) &&
+					dio->should_dirty)
 				set_page_dirty_lock(page);
 			page_cache_release(page);
 		}
@@ -1217,6 +1219,7 @@ do_blockdev_direct_IO(struct kiocb *iocb, struct inode *inode,
 	spin_lock_init(&dio->bio_lock);
 	dio->refcount = 1;
 
+	dio->should_dirty = (iter->type == ITER_IOVEC);
 	sdio.iter = iter;
 	sdio.final_block_in_request =
 		(offset + iov_iter_count(iter)) >> blkbits;
-- 
1.9.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1186068 — Re: [PATCH v7 1/6] fs: direct-io: don't dirtying pages for ITER_BVEC/ITER_KVEC direct read

On 07/16/2015 10:37 AM, Ming Lei wrote:
> When direct read IO is submitted from kernel, it is often
> unnecessary to dirty pages, for example of loop, dirtying pages
> have been considered in the upper filesystem(over loop) side
> already, and they don't need to be dirtied again.
> 
> So this patch doesn't dirtying pages for ITER_BVEC/ITER_KVEC
> direct read, and loop should be the 1st case to use ITER_BVEC/ITER_KVEC
> for direct read I/O.
> 
> The patch is based on previous Dave's patch.
> 
> Cc: Dave Kleikamp <dave.kleikamp@oracle.com>
> Reviewed-by: Christoph Hellwig <hch@lst.de>
> Signed-off-by: Ming Lei <ming.lei@canonical.com>

Reviewed-by: Dave Kleikamp <dave.kleikamp@oracle.com>

> ---
>  fs/direct-io.c | 9 ++++++---
>  1 file changed, 6 insertions(+), 3 deletions(-)
> 
> diff --git a/fs/direct-io.c b/fs/direct-io.c
> index 745d234..07e4b28 100644
> --- a/fs/direct-io.c
> +++ b/fs/direct-io.c
> @@ -120,6 +120,7 @@ struct dio {
>  	int page_errors;		/* errno from get_user_pages() */
>  	int is_async;			/* is IO async ? */
>  	bool defer_completion;		/* defer AIO completion to workqueue? */
> +	bool should_dirty;		/* if pages should be dirtied */
>  	int io_error;			/* IO error in completion path */
>  	unsigned long refcount;		/* direct_io_worker() and bios */
>  	struct bio *bio_list;		/* singly linked via bi_private */
> @@ -393,7 +394,7 @@ static inline void dio_bio_submit(struct dio *dio, struct dio_submit *sdio)
>  	dio->refcount++;
>  	spin_unlock_irqrestore(&dio->bio_lock, flags);
>  
> -	if (dio->is_async && dio->rw == READ)
> +	if (dio->is_async && dio->rw == READ && dio->should_dirty)
>  		bio_set_pages_dirty(bio);
>  
>  	if (sdio->submit_io)
> @@ -464,13 +465,14 @@ static int dio_bio_complete(struct dio *dio, struct bio *bio)
>  	if (!uptodate)
>  		dio->io_error = -EIO;
>  
> -	if (dio->is_async && dio->rw == READ) {
> +	if (dio->is_async && dio->rw == READ && dio->should_dirty) {
>  		bio_check_pages_dirty(bio);	/* transfers ownership */
>  	} else {
>  		bio_for_each_segment_all(bvec, bio, i) {
>  			struct page *page = bvec->bv_page;
>  
> -			if (dio->rw == READ && !PageCompound(page))
> +			if (dio->rw == READ && !PageCompound(page) &&
> +					dio->should_dirty)
>  				set_page_dirty_lock(page);
>  			page_cache_release(page);
>  		}
> @@ -1217,6 +1219,7 @@ do_blockdev_direct_IO(struct kiocb *iocb, struct inode *inode,
>  	spin_lock_init(&dio->bio_lock);
>  	dio->refcount = 1;
>  
> +	dio->should_dirty = (iter->type == ITER_IOVEC);
>  	sdio.iter = iter;
>  	sdio.final_block_in_request =
>  		(offset + iov_iter_count(iter)) >> blkbits;
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [next] | [standalone]

#1185999 — [PATCH v7 6/6] block: loop: support DIO & AIO

There are about 3 advantages to use direct I/O and AIO on
read/write loop's backing file:

1) double cache can be avoided, then memory usage gets
decreased a lot

2) not like user space direct I/O, there isn't cost of
pinning pages

3) avoid context switch for obtaining good throughput
- in buffered file read, random I/O top throughput is often obtained
only if they are submitted concurrently from lots of tasks; but for
sequential I/O, most of times they can be hit from page cache, so
concurrent submissions often introduce unnecessary context switch
and can't improve throughput much. There was such discussion[1]
to use non-blocking I/O to improve the problem for application.
- with direct I/O and AIO, concurrent submissions can be
avoided and random read throughput can't be affected meantime

Follows my fio test result:

1. 16 jobs fio test inside ext4 file system over loop block
1) How to run
	- linux kernel: 4.1.0-rc2-next-20150506 with the patchset
	- the loop block is over one image on HDD.
	- linux psync, 16 jobs, size 400M, ext4 over loop block
	- test result: IOPS from fio output

2) Throughput result:
        -------------------------------------------------------------
        test cases          |randread   |read   |randwrite  |write  |
        -------------------------------------------------------------
        base                |240        |8705   |3763       |20914
        -------------------------------------------------------------
        base+loop aio       |242        |9258   |4577       |21451
        -------------------------------------------------------------

3) context switch
        - context switch decreased by ~16% with loop aio for randread,
	and decreased by ~33% for read

4) memory usage
	- After these four tests with loop aio: ~10% memory becomes used
	- After these four tests without loop aio: more than 55% memory
	becomes used

2. single job fio test inside ext4 file system over loop block(for Maxim Patlasov)
1) How to run
	- linux kernel: 4.1.0-rc2-next-20150506 with the patchset
	- the loop block is over one image on HDD.
	- linux psync, 1 job, size 4000M, ext4 over loop block
	- test result: IOPS from fio output

2) Throughput result:
        -------------------------------------------------------------
        test cases          |randread   |read   |randwrite  |write  |
        -------------------------------------------------------------
        base                |109        |21180  |4192       |22782
        -------------------------------------------------------------
        base+loop aio       |114        |21018  |5404       |22670
        -------------------------------------------------------------

3) context switch
        - context switch decreased by ~10% with loop aio for randread,
	and decreased by ~50% for read

4) memory usage
	- After these four tests with loop aio: ~10% memory becomes used
	- After these four tests without loop aio: more than 55% memory
	becomes used

Both 'context switch' and 'memory usage' data are got from sar.

[1] https://lwn.net/Articles/612483/
[2] sar graph when running fio over loop without the patchset
http://kernel.ubuntu.com/~ming/block/loop-aio/v3/lo-nonaio.pdf

[3] sar graph when running fio over loop with the patchset
http://kernel.ubuntu.com/~ming/block/loop-aio/v3/lo-aio.pdf

[4] sar graph when running fio over loop without the patchset
http://kernel.ubuntu.com/~ming/block/loop-aio/v3/lo-nonaio-1job.pdf

[5] sar graph when running fio over loop with the patchset
http://kernel.ubuntu.com/~ming/block/loop-aio/v3/lo-aio-1job.pdf

Signed-off-by: Ming Lei <ming.lei@canonical.com>
---
 drivers/block/loop.c | 98 ++++++++++++++++++++++++++++++++++++++++++++++++++--
 drivers/block/loop.h |  2 ++
 2 files changed, 97 insertions(+), 3 deletions(-)

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index 3f26a0a..8099049 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -436,6 +436,90 @@ static int lo_req_flush(struct loop_device *lo, struct request *rq)
 	return ret;
 }
 
+static inline void handle_partial_read(struct loop_cmd *cmd, long bytes)
+{
+	if (bytes < 0 || (cmd->rq->cmd_flags & REQ_WRITE))
+		return;
+
+	if (unlikely(bytes < blk_rq_bytes(cmd->rq))) {
+		struct bio *bio = cmd->rq->bio;
+
+		bio_advance(bio, bytes);
+		zero_fill_bio(bio);
+	}
+}
+
+static void lo_rw_aio_complete(struct kiocb *iocb, long ret, long ret2)
+{
+	struct loop_cmd *cmd = container_of(iocb, struct loop_cmd, iocb);
+	struct request *rq = cmd->rq;
+
+	handle_partial_read(cmd, ret);
+
+	if (ret > 0)
+		ret = 0;
+	else if (ret < 0)
+		ret = -EIO;
+
+	rq->errors = ret;
+	blk_mq_complete_request(rq);
+}
+
+static int lo_rw_aio(struct loop_device *lo, struct loop_cmd *cmd,
+		     loff_t pos, bool rw)
+{
+	struct iov_iter iter;
+	struct bio_vec *bvec;
+	struct bio *bio = cmd->rq->bio;
+	struct file *file = lo->lo_backing_file;
+	int ret;
+
+	/* nomerge for loop request queue */
+	WARN_ON(cmd->rq->bio != cmd->rq->biotail);
+
+	bvec = __bvec_iter_bvec(bio->bi_io_vec, bio->bi_iter);
+	iov_iter_bvec(&iter, ITER_BVEC | rw, bvec,
+		      bio_segments(bio), blk_rq_bytes(cmd->rq));
+
+	cmd->iocb.ki_pos = pos;
+	cmd->iocb.ki_filp = file;
+	cmd->iocb.ki_complete = lo_rw_aio_complete;
+	cmd->iocb.ki_flags = IOCB_DIRECT;
+
+	if (rw == WRITE)
+		ret = file->f_op->write_iter(&cmd->iocb, &iter);
+	else
+		ret = file->f_op->read_iter(&cmd->iocb, &iter);
+
+	if (ret != -EIOCBQUEUED)
+		cmd->iocb.ki_complete(&cmd->iocb, ret, 0);
+	return 0;
+}
+
+
+static inline int lo_rw_simple(struct loop_device *lo,
+		struct request *rq, loff_t pos, bool rw)
+{
+	struct loop_cmd *cmd = blk_mq_rq_to_pdu(rq);
+
+	if (cmd->use_aio)
+		return lo_rw_aio(lo, cmd, pos, rw);
+
+	/*
+	 * lo_write_simple and lo_read_simple should have been covered
+	 * by io submit style function like lo_rw_aio(), one blocker
+	 * is that lo_read_simple() need to call flush_dcache_page after
+	 * the page is written from kernel, and it isn't easy to handle
+	 * this in io submit style function which submits all segments
+	 * of the req at one time. And direct read IO doesn't need to
+	 * run flush_dcache_page().
+	 */
+	if (rw == WRITE)
+		return lo_write_simple(lo, rq, pos);
+	else
+		return lo_read_simple(lo, rq, pos);
+}
+
 static int do_req_filebacked(struct loop_device *lo, struct request *rq)
 {
 	loff_t pos;
@@ -451,13 +535,13 @@ static int do_req_filebacked(struct loop_device *lo, struct request *rq)
 		else if (lo->transfer)
 			ret = lo_write_transfer(lo, rq, pos);
 		else
-			ret = lo_write_simple(lo, rq, pos);
+			ret = lo_rw_simple(lo, rq, pos, WRITE);
 
 	} else {
 		if (lo->transfer)
 			ret = lo_read_transfer(lo, rq, pos);
 		else
-			ret = lo_read_simple(lo, rq, pos);
+			ret = lo_rw_simple(lo, rq, pos, READ);
 	}
 
 	return ret;
@@ -1556,6 +1640,12 @@ static int loop_queue_rq(struct blk_mq_hw_ctx *hctx,
 	if (lo->lo_state != Lo_bound)
 		return -EIO;
 
+	if (lo->use_dio && !lo->transfer &&
+			!(cmd->rq->cmd_flags & (REQ_FLUSH | REQ_DISCARD)))
+		cmd->use_aio = true;
+	else
+		cmd->use_aio = false;
+
 	queue_kthread_work(&lo->worker, &cmd->work);
 
 	return BLK_MQ_RQ_QUEUE_OK;
@@ -1575,7 +1665,9 @@ static void loop_handle_cmd(struct loop_cmd *cmd)
  failed:
 	if (ret)
 		cmd->rq->errors = -EIO;
-	blk_mq_complete_request(cmd->rq);
+	/* complete non-aio request */
+	if (!cmd->use_aio || ret)
+		blk_mq_complete_request(cmd->rq);
 }
 
 static void loop_queue_work(struct kthread_work *work)
diff --git a/drivers/block/loop.h b/drivers/block/loop.h
index d1de221..fb2237c 100644
--- a/drivers/block/loop.h
+++ b/drivers/block/loop.h
@@ -69,6 +69,8 @@ struct loop_cmd {
 	struct kthread_work work;
 	struct request *rq;
 	struct list_head list;
+	bool use_aio;           /* use AIO interface to handle I/O */
+	struct kiocb iocb;
 };
 
 /* Support for loadable transfer modules */
-- 
1.9.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web