On Tue, May 26, 2020 at 03:17:46PM +0800, zhangyi (F) wrote:
> Pick out ll_rw_one_block() helper function from ll_rw_block() for
> submitting one locked buffer for reading/writing.
That should probably read factor out instead of pick out.
>
> Signed-off-by: zhangyi (F) <yi.zhang@xxxxxxxxxx>
> ---
> fs/buffer.c | 41 ++++++++++++++++++++++---------------
> include/linux/buffer_head.h | 1 +
> 2 files changed, 26 insertions(+), 16 deletions(-)
>
> diff --git a/fs/buffer.c b/fs/buffer.c
> index a60f60396cfa..3a2226f88b2d 100644
> --- a/fs/buffer.c
> +++ b/fs/buffer.c
> @@ -3081,6 +3081,29 @@ int submit_bh(int op, int op_flags, struct buffer_head *bh)
> }
> EXPORT_SYMBOL(submit_bh);
>
> +void ll_rw_one_block(int op, int op_flags, struct buffer_head *bh)
> +{
> + BUG_ON(!buffer_locked(bh));
> +
> + if (op == WRITE) {
> + if (test_clear_buffer_dirty(bh)) {
> + bh->b_end_io = end_buffer_write_sync;
> + get_bh(bh);
> + submit_bh(op, op_flags, bh);
> + return;
> + }
> + } else {
> + if (!buffer_uptodate(bh)) {
> + bh->b_end_io = end_buffer_read_sync;
> + get_bh(bh);
> + submit_bh(op, op_flags, bh);
> + return;
> + }
> + }
> + unlock_buffer(bh);
> +}
> +EXPORT_SYMBOL(ll_rw_one_block);
I don't think you want separate read and write sides. In fact I'm not
sure you want the helper at all. At this point just open coding it
rather than adding more overhead to core code might be a better idea.
