[PATCH v6 0/3] limit bio max size

[PATCH v6 0/3] limit bio max size

[Changheun Lee]

I found a inefficient behavior from multipage bvec. Large chunk DIO
scenario is that. This patch series could be a solution to improve it.

Changheun Lee (3):
  bio: limit bio max size
  ufs: set QUEUE_FLAG_LIMIT_BIO_SIZE
  bio: add limit_bio_size sysfs

 Documentation/ABI/testing/sysfs-block | 10 ++++++++++
 Documentation/block/queue-sysfs.rst   |  7 +++++++
 block/bio.c                           | 13 ++++++++++++-
 block/blk-settings.c                  | 17 +++++++++++++++++
 block/blk-sysfs.c                     |  3 +++
 drivers/scsi/scsi_lib.c               |  2 ++
 drivers/scsi/ufs/ufshcd.c             |  1 +
 include/linux/bio.h                   |  2 +-
 include/linux/blkdev.h                |  4 ++++
 include/scsi/scsi_host.h              |  2 ++
 10 files changed, 59 insertions(+), 2 deletions(-)

-- 
2.29.0

[PATCH v6 1/3] bio: limit bio max size

[Changheun Lee]

bio size can grow up to 4GB when muli-page bvec is enabled.
but sometimes it would lead to inefficient behaviors.
in case of large chunk direct I/O, - 32MB chunk read in user space -
all pages for 32MB would be merged to a bio structure if the pages
physical addresses are contiguous. it makes some delay to submit
until merge complete. bio max size should be limited to a proper size.

When 32MB chunk read with direct I/O option is coming from userspace,
kernel behavior is below now in do_direct_IO() loop. it's timeline.

 | bio merge for 32MB. total 8,192 pages are merged.
 | total elapsed time is over 2ms.
 |------------------ ... ----------------------->|
                                                 | 8,192 pages merged a bio.
                                                 | at this time, first bio submit is done.
                                                 | 1 bio is split to 32 read request and issue.
                                                 |--------------->
                                                  |--------------->
                                                   |--------------->
                                                              ......
                                                                   |--------------->
                                                                    |--------------->|
                          total 19ms elapsed to complete 32MB read done from device. |

If bio max size is limited with 1MB, behavior is changed below.

 | bio merge for 1MB. 256 pages are merged for each bio.
 | total 32 bio will be made.
 | total elapsed time is over 2ms. it's same.
 | but, first bio submit timing is fast. about 100us.
 |--->|--->|--->|---> ... -->|--->|--->|--->|--->|
      | 256 pages merged a bio.
      | at this time, first bio submit is done.
      | and 1 read request is issued for 1 bio.
      |--------------->
           |--------------->
                |--------------->
                                      ......
                                                 |--------------->
                                                  |--------------->|
        total 17ms elapsed to complete 32MB read done from device. |

As a result, read request issue timing is faster if bio max size is limited.
Current kernel behavior with multipage bvec, super large bio can be created.
And it lead to delay first I/O request issue.

Signed-off-by: Changheun Lee <nanich.lee@samsung.com>
---
 block/bio.c            | 13 ++++++++++++-
 block/blk-settings.c   | 17 +++++++++++++++++
 include/linux/bio.h    |  2 +-
 include/linux/blkdev.h |  4 ++++
 4 files changed, 34 insertions(+), 2 deletions(-)

diff --git a/block/bio.c b/block/bio.c
index 50e579088aca..e4d6169106b6 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -255,6 +255,17 @@ void bio_init(struct bio *bio, struct bio_vec *table,
 }
 EXPORT_SYMBOL(bio_init);
 
+unsigned int bio_max_size(struct bio *bio)
+{
+	struct request_queue *q = bio->bi_disk->queue;
+
+	if (blk_queue_limit_bio_size(q))
+		return blk_queue_get_max_sectors(q, bio_op(bio))
+			<< SECTOR_SHIFT;
+
+	return UINT_MAX;
+}
+
 /**
  * bio_reset - reinitialize a bio
  * @bio:	bio to reset
@@ -866,7 +877,7 @@ bool __bio_try_merge_page(struct bio *bio, struct page *page,
 		struct bio_vec *bv = &bio->bi_io_vec[bio->bi_vcnt - 1];
 
 		if (page_is_mergeable(bv, page, len, off, same_page)) {
-			if (bio->bi_iter.bi_size > UINT_MAX - len) {
+			if (bio->bi_iter.bi_size > bio_max_size(bio) - len) {
 				*same_page = false;
 				return false;
 			}
diff --git a/block/blk-settings.c b/block/blk-settings.c
index b4aa2f37fab6..1d94b97cea4f 100644
--- a/block/blk-settings.c
+++ b/block/blk-settings.c
@@ -928,6 +928,23 @@ void blk_queue_set_zoned(struct gendisk *disk, enum blk_zoned_model model)
 }
 EXPORT_SYMBOL_GPL(blk_queue_set_zoned);
 
+/**
+ * blk_queue_set_limit_bio_size - set limit bio size flag
+ * @q:		the request queue for the device
+ * @limit:	limit bio size on(true), or off
+ *
+ * bio max size will be limited to queue max sectors size,
+ * if limit is true.
+ */
+void blk_queue_set_limit_bio_size(struct request_queue *q, bool limit)
+{
+	if (limit)
+		blk_queue_flag_set(QUEUE_FLAG_LIMIT_BIO_SIZE, q);
+	else
+		blk_queue_flag_clear(QUEUE_FLAG_LIMIT_BIO_SIZE, q);
+}
+EXPORT_SYMBOL_GPL(blk_queue_set_limit_bio_size);
+
 static int __init blk_settings_init(void)
 {
 	blk_max_low_pfn = max_low_pfn - 1;
diff --git a/include/linux/bio.h b/include/linux/bio.h
index d0246c92a6e8..830c784967c0 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -119,7 +119,7 @@ static inline bool bio_full(struct bio *bio, unsigned len)
 	if (bio->bi_vcnt >= bio->bi_max_vecs)
 		return true;
 
-	if (bio->bi_iter.bi_size > UINT_MAX - len)
+	if (bio->bi_iter.bi_size > bio_max_size(bio) - len)
 		return true;
 
 	return false;
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 158aefae1030..c69a6ed7a189 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -617,6 +617,7 @@ struct request_queue {
 #define QUEUE_FLAG_RQ_ALLOC_TIME 27	/* record rq->alloc_time_ns */
 #define QUEUE_FLAG_HCTX_ACTIVE	28	/* at least one blk-mq hctx is active */
 #define QUEUE_FLAG_NOWAIT       29	/* device supports NOWAIT */
+#define QUEUE_FLAG_LIMIT_BIO_SIZE 30	/* limit bio size */
 
 #define QUEUE_FLAG_MQ_DEFAULT	((1 << QUEUE_FLAG_IO_STAT) |		\
 				 (1 << QUEUE_FLAG_SAME_COMP) |		\
@@ -663,6 +664,8 @@ bool blk_queue_flag_test_and_set(unsigned int flag, struct request_queue *q);
 #define blk_queue_fua(q)	test_bit(QUEUE_FLAG_FUA, &(q)->queue_flags)
 #define blk_queue_registered(q)	test_bit(QUEUE_FLAG_REGISTERED, &(q)->queue_flags)
 #define blk_queue_nowait(q)	test_bit(QUEUE_FLAG_NOWAIT, &(q)->queue_flags)
+#define blk_queue_limit_bio_size(q)	\
+	test_bit(QUEUE_FLAG_LIMIT_BIO_SIZE, &(q)->queue_flags)
 
 extern void blk_set_pm_only(struct request_queue *q);
 extern void blk_clear_pm_only(struct request_queue *q);
@@ -1183,6 +1186,7 @@ extern void blk_queue_required_elevator_features(struct request_queue *q,
 						 unsigned int features);
 extern bool blk_queue_can_use_dma_map_merging(struct request_queue *q,
 					      struct device *dev);
+extern void blk_queue_set_limit_bio_size(struct request_queue *q, bool limit);
 
 /*
  * Number of physical segments as sent to the device.
-- 
2.29.0

[PATCH v6 2/3] ufs: set QUEUE_FLAG_LIMIT_BIO_SIZE

[Changheun Lee]

Set QUEUE_FLAG_LIMIT_BIO_SIZE queue flag to limit bio max size to
queue max sectors size for UFS device.

Signed-off-by: Changheun Lee <nanich.lee@samsung.com>
---
 drivers/scsi/scsi_lib.c   | 2 ++
 drivers/scsi/ufs/ufshcd.c | 1 +
 include/scsi/scsi_host.h  | 2 ++
 3 files changed, 5 insertions(+)

diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
index 7d52a11e1b61..73ce6ba7903a 100644
--- a/drivers/scsi/scsi_lib.c
+++ b/drivers/scsi/scsi_lib.c
@@ -1838,6 +1838,8 @@ void __scsi_init_queue(struct Scsi_Host *shost, struct request_queue *q)
 	 * Devices that require a bigger alignment can increase it later.
 	 */
 	blk_queue_dma_alignment(q, max(4, dma_get_cache_alignment()) - 1);
+
+	blk_queue_set_limit_bio_size(q, shost->limit_bio_size);
 }
 EXPORT_SYMBOL_GPL(__scsi_init_queue);
 
diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index d3d05e997c13..000eb5ab022e 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -9313,6 +9313,7 @@ int ufshcd_init(struct ufs_hba *hba, void __iomem *mmio_base, unsigned int irq)
 	host->max_channel = UFSHCD_MAX_CHANNEL;
 	host->unique_id = host->host_no;
 	host->max_cmd_len = UFS_CDB_SIZE;
+	host->limit_bio_size = true;
 
 	hba->max_pwr_info.is_valid = false;
 
diff --git a/include/scsi/scsi_host.h b/include/scsi/scsi_host.h
index e30fd963b97d..486f61588717 100644
--- a/include/scsi/scsi_host.h
+++ b/include/scsi/scsi_host.h
@@ -607,6 +607,8 @@ struct Scsi_Host {
 	unsigned int max_segment_size;
 	unsigned long dma_boundary;
 	unsigned long virt_boundary_mask;
+	unsigned int limit_bio_size;
+
 	/*
 	 * In scsi-mq mode, the number of hardware queues supported by the LLD.
 	 *
-- 
2.29.0

[PATCH v6 3/3] bio: add limit_bio_size sysfs

[Changheun Lee]

Add limit_bio_size block sysfs node to limit bio size.
Queue flag QUEUE_FLAG_LIMIT_BIO_SIZE will be set if limit_bio_size is set.
And bio max size will be limited by queue max sectors via
QUEUE_FLAG_LIMIT_BIO_SIZE set.

Signed-off-by: Changheun Lee <nanich.lee@samsung.com>
---
 Documentation/ABI/testing/sysfs-block | 10 ++++++++++
 Documentation/block/queue-sysfs.rst   |  7 +++++++
 block/blk-sysfs.c                     |  3 +++
 3 files changed, 20 insertions(+)

diff --git a/Documentation/ABI/testing/sysfs-block b/Documentation/ABI/testing/sysfs-block
index e34cdeeeb9d4..86a7b15410cf 100644
--- a/Documentation/ABI/testing/sysfs-block
+++ b/Documentation/ABI/testing/sysfs-block
@@ -316,3 +316,13 @@ Description:
 		does not complete in this time then the block driver timeout
 		handler is invoked. That timeout handler can decide to retry
 		the request, to fail it or to start a device recovery strategy.
+
+What:		/sys/block/<disk>/queue/limit_bio_size
+Date:		Feb, 2021
+Contact:	Changheun Lee <nanich.lee@samsung.com>
+Description:
+		(RW) Toggle for set/clear QUEUE_FLAG_LIMIT_BIO_SIZE queue flag.
+		Queue flag QUEUE_FLAG_LIMIT_BIO_SIZE will be set if limit_bio_size
+		is set. And bio max size will be limited by queue max sectors.
+		QUEUE_FLAG_LIMIT_BIO_SIZE will be cleared if limit_bio_size is
+		cleard. And limit of bio max size will be cleard.
diff --git a/Documentation/block/queue-sysfs.rst b/Documentation/block/queue-sysfs.rst
index 4dc7f0d499a8..220d183a4c11 100644
--- a/Documentation/block/queue-sysfs.rst
+++ b/Documentation/block/queue-sysfs.rst
@@ -286,4 +286,11 @@ sequential zones of zoned block devices (devices with a zoned attributed
 that reports "host-managed" or "host-aware"). This value is always 0 for
 regular block devices.
 
+limit_bio_size (RW)
+-------------------
+This indicates QUEUE_FLAG_LIMIT_BIO_SIZE queue flag value. And
+QUEUE_FLAG_LIMIT_BIO_SIZE can be changed via set(1)/clear(0) this node.
+bio max size will be limited by queue max sectors via set this node. And
+limit of bio max size will be cleard via clear this node.
+
 Jens Axboe <jens.axboe@oracle.com>, February 2009
diff --git a/block/blk-sysfs.c b/block/blk-sysfs.c
index 0f4f0c8a7825..4625d5319daf 100644
--- a/block/blk-sysfs.c
+++ b/block/blk-sysfs.c
@@ -294,6 +294,7 @@ QUEUE_SYSFS_BIT_FNS(nonrot, NONROT, 1);
 QUEUE_SYSFS_BIT_FNS(random, ADD_RANDOM, 0);
 QUEUE_SYSFS_BIT_FNS(iostats, IO_STAT, 0);
 QUEUE_SYSFS_BIT_FNS(stable_writes, STABLE_WRITES, 0);
+QUEUE_SYSFS_BIT_FNS(limit_bio_size, LIMIT_BIO_SIZE, 0);
 #undef QUEUE_SYSFS_BIT_FNS
 
 static ssize_t queue_zoned_show(struct request_queue *q, char *page)
@@ -625,6 +626,7 @@ QUEUE_RW_ENTRY(queue_nonrot, "rotational");
 QUEUE_RW_ENTRY(queue_iostats, "iostats");
 QUEUE_RW_ENTRY(queue_random, "add_random");
 QUEUE_RW_ENTRY(queue_stable_writes, "stable_writes");
+QUEUE_RW_ENTRY(queue_limit_bio_size, "limit_bio_size");
 
 static struct attribute *queue_attrs[] = {
 	&queue_requests_entry.attr,
@@ -659,6 +661,7 @@ static struct attribute *queue_attrs[] = {
 	&queue_rq_affinity_entry.attr,
 	&queue_iostats_entry.attr,
 	&queue_stable_writes_entry.attr,
+	&queue_limit_bio_size_entry.attr,
 	&queue_random_entry.attr,
 	&queue_poll_entry.attr,
 	&queue_wc_entry.attr,
-- 
2.29.0

Re: [PATCH v6 1/3] bio: limit bio max size

[kernel test robot]

[-- Attachment #1: Type: text/plain, Size: 3933 bytes --]

Hi Changheun,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on mkp-scsi/for-next]
[also build test ERROR on scsi/for-next linus/master v5.12-rc7]
[cannot apply to block/for-next next-20210409]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]

url:    https://github.com/0day-ci/linux/commits/Changheun-Lee/bio-limit-bio-max-size/20210412-115922
base:   https://git.kernel.org/pub/scm/linux/kernel/git/mkp/scsi.git for-next
config: i386-tinyconfig (attached as .config)
compiler: gcc-9 (Debian 9.3.0-22) 9.3.0
reproduce (this is a W=1 build):
        # https://github.com/0day-ci/linux/commit/5d00aa42f58575968c4a7a4b374addaf4f9a5624
        git remote add linux-review https://github.com/0day-ci/linux
        git fetch --no-tags linux-review Changheun-Lee/bio-limit-bio-max-size/20210412-115922
        git checkout 5d00aa42f58575968c4a7a4b374addaf4f9a5624
        # save the attached .config to linux build tree
        make W=1 ARCH=i386 

If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>

All errors (new ones prefixed by >>):

   In file included from include/linux/blkdev.h:19,
                    from include/linux/blk-cgroup.h:23,
                    from include/linux/writeback.h:14,
                    from include/linux/memcontrol.h:22,
                    from include/linux/swap.h:9,
                    from include/linux/suspend.h:5,
                    from arch/x86/kernel/asm-offsets.c:13:
   include/linux/bio.h: In function 'bio_full':
>> include/linux/bio.h:122:29: error: implicit declaration of function 'bio_max_size'; did you mean 'bio_max_segs'? [-Werror=implicit-function-declaration]
     122 |  if (bio->bi_iter.bi_size > bio_max_size(bio) - len)
         |                             ^~~~~~~~~~~~
         |                             bio_max_segs
   cc1: some warnings being treated as errors
--
   In file included from include/linux/blkdev.h:19,
                    from include/linux/blk-cgroup.h:23,
                    from include/linux/writeback.h:14,
                    from include/linux/memcontrol.h:22,
                    from include/linux/swap.h:9,
                    from include/linux/suspend.h:5,
                    from arch/x86/kernel/asm-offsets.c:13:
   include/linux/bio.h: In function 'bio_full':
>> include/linux/bio.h:122:29: error: implicit declaration of function 'bio_max_size'; did you mean 'bio_max_segs'? [-Werror=implicit-function-declaration]
     122 |  if (bio->bi_iter.bi_size > bio_max_size(bio) - len)
         |                             ^~~~~~~~~~~~
         |                             bio_max_segs
   cc1: some warnings being treated as errors
   make[2]: *** [scripts/Makefile.build:116: arch/x86/kernel/asm-offsets.s] Error 1
   make[2]: Target '__build' not remade because of errors.
   make[1]: *** [Makefile:1233: prepare0] Error 2
   make[1]: Target 'prepare' not remade because of errors.
   make: *** [Makefile:215: __sub-make] Error 2
   make: Target 'prepare' not remade because of errors.


vim +122 include/linux/bio.h

   108	
   109	/**
   110	 * bio_full - check if the bio is full
   111	 * @bio:	bio to check
   112	 * @len:	length of one segment to be added
   113	 *
   114	 * Return true if @bio is full and one segment with @len bytes can't be
   115	 * added to the bio, otherwise return false
   116	 */
   117	static inline bool bio_full(struct bio *bio, unsigned len)
   118	{
   119		if (bio->bi_vcnt >= bio->bi_max_vecs)
   120			return true;
   121	
 > 122		if (bio->bi_iter.bi_size > bio_max_size(bio) - len)
   123			return true;
   124	
   125		return false;
   126	}
   127	

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org

[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 7346 bytes --]

Re: [PATCH v6 1/3] bio: limit bio max size

[kernel test robot]

[-- Attachment #1: Type: text/plain, Size: 3667 bytes --]

Hi Changheun,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on mkp-scsi/for-next]
[also build test ERROR on scsi/for-next linus/master v5.12-rc7]
[cannot apply to block/for-next next-20210409]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]

url:    https://github.com/0day-ci/linux/commits/Changheun-Lee/bio-limit-bio-max-size/20210412-115922
base:   https://git.kernel.org/pub/scm/linux/kernel/git/mkp/scsi.git for-next
config: nios2-defconfig (attached as .config)
compiler: nios2-linux-gcc (GCC) 9.3.0
reproduce (this is a W=1 build):
        wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
        chmod +x ~/bin/make.cross
        # https://github.com/0day-ci/linux/commit/5d00aa42f58575968c4a7a4b374addaf4f9a5624
        git remote add linux-review https://github.com/0day-ci/linux
        git fetch --no-tags linux-review Changheun-Lee/bio-limit-bio-max-size/20210412-115922
        git checkout 5d00aa42f58575968c4a7a4b374addaf4f9a5624
        # save the attached .config to linux build tree
        COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross ARCH=nios2 

If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>

All error/warnings (new ones prefixed by >>):

   In file included from include/linux/blkdev.h:19,
                    from include/linux/blk-cgroup.h:23,
                    from include/linux/writeback.h:14,
                    from include/linux/memcontrol.h:22,
                    from include/linux/swap.h:9,
                    from block/bio.c:6:
   include/linux/bio.h: In function 'bio_full':
   include/linux/bio.h:122:29: error: implicit declaration of function 'bio_max_size'; did you mean 'bio_max_segs'? [-Werror=implicit-function-declaration]
     122 |  if (bio->bi_iter.bi_size > bio_max_size(bio) - len)
         |                             ^~~~~~~~~~~~
         |                             bio_max_segs
   block/bio.c: At top level:
>> block/bio.c:258:14: warning: no previous prototype for 'bio_max_size' [-Wmissing-prototypes]
     258 | unsigned int bio_max_size(struct bio *bio)
         |              ^~~~~~~~~~~~
>> block/bio.c:258:14: error: conflicting types for 'bio_max_size'
   In file included from include/linux/blkdev.h:19,
                    from include/linux/blk-cgroup.h:23,
                    from include/linux/writeback.h:14,
                    from include/linux/memcontrol.h:22,
                    from include/linux/swap.h:9,
                    from block/bio.c:6:
   include/linux/bio.h:122:29: note: previous implicit declaration of 'bio_max_size' was here
     122 |  if (bio->bi_iter.bi_size > bio_max_size(bio) - len)
         |                             ^~~~~~~~~~~~
   block/bio.c: In function 'bio_max_size':
>> block/bio.c:260:31: error: 'struct bio' has no member named 'bi_disk'
     260 |  struct request_queue *q = bio->bi_disk->queue;
         |                               ^~
   cc1: some warnings being treated as errors


vim +/bio_max_size +258 block/bio.c

   257	
 > 258	unsigned int bio_max_size(struct bio *bio)
   259	{
 > 260		struct request_queue *q = bio->bi_disk->queue;
   261	
   262		if (blk_queue_limit_bio_size(q))
   263			return blk_queue_get_max_sectors(q, bio_op(bio))
   264				<< SECTOR_SHIFT;
   265	
   266		return UINT_MAX;
   267	}
   268	

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org

[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 9951 bytes --]