block: count 'ios' and 'sectors' when io is done for bio-based device

While using iostat for raid, I observed very strange 'await'
occasionally, and turns out it's due to that 'ios' and 'sectors' is
counted in bdev_start_io_acct(), while 'nsecs' is counted in
bdev_end_io_acct(). I'm not sure why they are ccounted like that
but I think this behaviour is obviously wrong because user will get
wrong disk stats.

Fix the problem by counting 'ios' and 'sectors' when io is done, like
what rq-based device does.

Fixes: 394ffa503b ("blk: introduce generic io stat accounting help function")
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20230223091226.1135678-1-yukuai1@huaweicloud.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This commit is contained in:
Yu Kuai
2023-02-23 17:12:26 +08:00
committed by Jens Axboe
parent 6030363199
commit 5f27571382
4 changed files with 15 additions and 20 deletions

View File

@@ -1446,11 +1446,10 @@ static inline void blk_wake_io_task(struct task_struct *waiter)
wake_up_process(waiter);
}
unsigned long bdev_start_io_acct(struct block_device *bdev,
unsigned int sectors, enum req_op op,
unsigned long bdev_start_io_acct(struct block_device *bdev, enum req_op op,
unsigned long start_time);
void bdev_end_io_acct(struct block_device *bdev, enum req_op op,
unsigned long start_time);
unsigned int sectors, unsigned long start_time);
unsigned long bio_start_io_acct(struct bio *bio);
void bio_end_io_acct_remapped(struct bio *bio, unsigned long start_time,