[PATCH -for-next 1/2] block: introduce BIO_IN_FLIGHT flag

[PATCH -for-next 1/2] block: introduce BIO_IN_FLIGHT flag

[Namhyung Kim]

BIO_IN_FLIGHT flag is used for tracing block I/O completion.
This patch fixes tracing bio-based devices - except DM which
inserts completion tracepoint explicitly - that could not be
traced such event using blktrace.

It won't affect tracing normal (request-based) disk devices
and nested bio handling paths.

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
---
 block/blk-core.c          |    5 +++++
 fs/bio.c                  |    9 +++++++++
 include/linux/blk_types.h |    1 +
 3 files changed, 15 insertions(+), 0 deletions(-)

diff --git a/block/blk-core.c b/block/blk-core.c
index a99514191ce2..d8005ffc480d 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -162,6 +162,9 @@ static void req_bio_endio(struct request *rq, struct bio *bio,
 	if (unlikely(rq->cmd_flags & REQ_QUIET))
 		set_bit(BIO_QUIET, &bio->bi_flags);
 
+	/* completion event was already reported in blk_update_request */
+	clear_bit(BIO_IN_FLIGHT, &bio->bi_flags);
+
 	bio->bi_size -= nbytes;
 	bio->bi_sector += (nbytes >> 9);
 
@@ -1571,6 +1574,8 @@ void generic_make_request(struct bio *bio)
 	do {
 		struct request_queue *q = bdev_get_queue(bio->bi_bdev);
 
+		set_bit(BIO_IN_FLIGHT, &bio->bi_flags);
+
 		q->make_request_fn(q, bio);
 
 		bio = bio_list_pop(current->bio_list);
diff --git a/fs/bio.c b/fs/bio.c
index 9bfade8a609b..c510e93b6312 100644
--- a/fs/bio.c
+++ b/fs/bio.c
@@ -1447,6 +1447,15 @@ void bio_endio(struct bio *bio, int error)
 	else if (!test_bit(BIO_UPTODATE, &bio->bi_flags))
 		error = -EIO;
 
+	if (test_bit(BIO_IN_FLIGHT, &bio->bi_flags)) {
+		struct request_queue *q = bdev_get_queue(bio->bi_bdev);
+
+		trace_block_bio_complete(q, bio, error);
+
+		/* prevent duplicated completion event report */
+		clear_bit(BIO_IN_FLIGHT, &bio->bi_flags);
+	}
+
 	if (bio->bi_end_io)
 		bio->bi_end_io(bio, error);
 }
diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
index 71fc53bb8f1c..146295babc98 100644
--- a/include/linux/blk_types.h
+++ b/include/linux/blk_types.h
@@ -98,6 +98,7 @@ struct bio {
 #define BIO_FS_INTEGRITY 10	/* fs owns integrity data, not block layer */
 #define BIO_QUIET	11	/* Make BIO Quiet */
 #define BIO_MAPPED_INTEGRITY 12/* integrity metadata has been remapped */
+#define BIO_IN_FLIGHT	13	/* report I/O completion event */
 #define bio_flagged(bio, flag)	((bio)->bi_flags & (1 << (flag)))
 
 /*
-- 
1.7.6

[PATCH 2/2] block: don't export block_bio_complete tracepoint

[Namhyung Kim]

Now bio_endio() contains the tracepoint in it, so we don't
need to have it twice in DM. Plus, as the only external user
of the tracepoint was the DM, we can unexport the symbol.

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Cc: dm-devel@redhat.com
---
 block/blk-core.c |    1 -
 drivers/md/dm.c  |    1 -
 2 files changed, 0 insertions(+), 2 deletions(-)

diff --git a/block/blk-core.c b/block/blk-core.c
index d8005ffc480d..5c81ef057e6a 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -36,7 +36,6 @@
 
 EXPORT_TRACEPOINT_SYMBOL_GPL(block_bio_remap);
 EXPORT_TRACEPOINT_SYMBOL_GPL(block_rq_remap);
-EXPORT_TRACEPOINT_SYMBOL_GPL(block_bio_complete);
 
 /*
  * For the allocated request tables
diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index 7b986e77b75e..a8223d92a52b 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -636,7 +636,6 @@ static void dec_pending(struct dm_io *io, int error)
 			queue_io(md, bio);
 		} else {
 			/* done with normal IO or empty flush */
-			trace_block_bio_complete(md->queue, bio, io_error);
 			bio_endio(bio, io_error);
 		}
 	}
-- 
1.7.6

[PATCH -block/for-next 0/2] blktrace: bio-based device tracing improvement v3

[Namhyung Kim]

Hello,

The blktrace is used to report block device activities to user space
using kernel tracepoint but it was focused to block I/O request struct.
Thus bio-based devices (i.e. loop, ram, md, ...) which don't make use
of the request structure could not be supported well - the tool only
can detect a limited number of events such as queuing, cloning and
remapping but it cannot know when the I/O activity is completed.

bio_endio(), the I/O completion callback, can be used to fix this
problem by adding appropriate tracepoint in it. However it was called
from other paths too (normal request-based block devices and some of
nested block I/O handling routines) so that we should recognize such
cases to prevent duplicated reports. In this series, BIO_IN_FLIGHT flag
is introduced and used for that purpose.

Note that (bio-based) dm already supported completion report by adding
the tracepoint into the path manually. With this patches, it will be
converted to use generic mechanism.

Changes from v2:
 * rebased on current block/for-next

Changes from v1:
 * kill __bio_endio()
 * use BIO_IN_FLIGHT flag

Any feedbacks are welcome.

Thanks.


Namhyung Kim (2):
  block: introduce BIO_IN_FLIGHT flag
  block: don't export block_bio_complete tracepoint

 block/blk-core.c          |    6 +++++-
 drivers/md/dm.c           |    1 -
 fs/bio.c                  |    7 +++++++
 include/linux/blk_types.h |    1 +
 4 files changed, 13 insertions(+), 2 deletions(-)

--
1.7.6

[PATCH 2/2] block: don't export block_bio_complete tracepoint

[Namhyung Kim]

Now bio_endio() contains the tracepoint in it, so we don't
need to have it twice in DM. Plus, as the only external user
of the tracepoint was the DM, we can unexport the symbol.

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Cc: dm-devel@redhat.com
---
 block/blk-core.c |    1 -
 drivers/md/dm.c  |    1 -
 2 files changed, 0 insertions(+), 2 deletions(-)

diff --git a/block/blk-core.c b/block/blk-core.c
index f61310323954..f7eab1932543 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -37,7 +37,6 @@
 
 EXPORT_TRACEPOINT_SYMBOL_GPL(block_bio_remap);
 EXPORT_TRACEPOINT_SYMBOL_GPL(block_rq_remap);
-EXPORT_TRACEPOINT_SYMBOL_GPL(block_bio_complete);
 
 DEFINE_IDA(blk_queue_ida);
 
diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index 4720f68f817e..01185fa0eb74 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -648,7 +648,6 @@ static void dec_pending(struct dm_io *io, int error)
 			queue_io(md, bio);
 		} else {
 			/* done with normal IO or empty flush */
-			trace_block_bio_complete(md->queue, bio, io_error);
 			bio_endio(bio, io_error);
 		}
 	}
-- 
1.7.6