Reading the inode size brings in a new cacheline for IO submit, and
it's in the hot path being checked for every single IO. When doing
millions of IOs per core per second, this is noticeable overhead.
Cache the nr_sectors in the bdev itself.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
spin_lock(&bdev->bd_size_lock);
i_size_write(bdev->bd_inode, (loff_t)sectors << SECTOR_SHIFT);
+ bdev->bd_nr_sectors = sectors;
spin_unlock(&bdev->bd_size_lock);
}
EXPORT_SYMBOL(set_capacity);
{
spin_lock(&bdev->bd_size_lock);
i_size_write(bdev->bd_inode, (loff_t)sectors << SECTOR_SHIFT);
+ bdev->bd_nr_sectors = sectors;
spin_unlock(&bdev->bd_size_lock);
}
struct block_device {
sector_t bd_start_sect;
+ sector_t bd_nr_sectors;
struct disk_stats __percpu *bd_stats;
unsigned long bd_stamp;
bool bd_read_only; /* read-only policy */
return bdev->bd_start_sect;
}
-static inline loff_t bdev_nr_bytes(struct block_device *bdev)
+static inline sector_t bdev_nr_sectors(struct block_device *bdev)
{
- return i_size_read(bdev->bd_inode);
+ return bdev->bd_nr_sectors;
}
-static inline sector_t bdev_nr_sectors(struct block_device *bdev)
+static inline loff_t bdev_nr_bytes(struct block_device *bdev)
{
- return bdev_nr_bytes(bdev) >> SECTOR_SHIFT;
+ return bdev_nr_sectors(bdev) << SECTOR_SHIFT;
}
static inline sector_t get_capacity(struct gendisk *disk)