On Wed, Feb 12, 2020 at 12:07:28PM -0500, Vivek Goyal wrote:
> Create a new helper dax_pgoff() which will replace bdev_dax_pgoff(). Difference
> between two is that dax_pgoff() takes in "sector_t dax_offset" as an argument
> instead of "struct block_device".
>
> dax_offset specifies any offset into dax device which should be added to
> sector while calculating pgoff.
>
> Signed-off-by: Vivek Goyal <vgoyal@xxxxxxxxxx>
> ---
> drivers/dax/super.c | 12 ++++++++++++
> include/linux/dax.h | 1 +
> 2 files changed, 13 insertions(+)
>
> diff --git a/drivers/dax/super.c b/drivers/dax/super.c
> index 0aa4b6bc5101..e9daa30e4250 100644
> --- a/drivers/dax/super.c
> +++ b/drivers/dax/super.c
> @@ -56,6 +56,18 @@ int bdev_dax_pgoff(struct block_device *bdev, sector_t sector, size_t size,
> }
> EXPORT_SYMBOL(bdev_dax_pgoff);
>
> +int dax_pgoff(sector_t dax_offset, sector_t sector, size_t size, pgoff_t *pgoff)
Please add a kerneldoc document. I can't really make sense of what
dax_offset and sector mean here and why they are passed separately.
> +{
> + phys_addr_t phys_off = (dax_offset + sector) * 512;
<< SECTOR_SHIFT;
> +
> + if (pgoff)
> + *pgoff = PHYS_PFN(phys_off);
What is the use case of not passing a pgoff argument?
> + if (phys_off % PAGE_SIZE || size % PAGE_SIZE)
> + return -EINVAL;
> + return 0;
> +}
> +EXPORT_SYMBOL(dax_pgoff);
EXPORT_SYMBOL_GPL, please.
