On 2021/06/25 15:49, Christoph Hellwig wrote:
> Keep the op-specific flag last and make a little more sense of the
> comments.
>
> Signed-off-by: Christoph Hellwig <hch@xxxxxx>
> ---
> include/linux/blk_types.h | 11 +++++------
> 1 file changed, 5 insertions(+), 6 deletions(-)
>
> diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
> index db61f7df1823..c80756e9f16e 100644
> --- a/include/linux/blk_types.h
> +++ b/include/linux/blk_types.h
> @@ -381,15 +381,14 @@ enum req_flag_bits {
> * work item to avoid such priority inversions.
> */
> __REQ_CGROUP_PUNT,
> + __REQ_HIPRI, /* High Priority I/O (polled) */
REQ_HIPRI requests may not have an actual high priority assigned (rq->ioprio).
So can we change the comment to only say "/* Polled I/O */".
Polled I/O is indeed semantically somewhat equivalent to a high priority
request, but I prefer we clearly distinguish it from it the ioprio stuff.
> + __REQ_SWAP, /* swap I/O */
>
> - /* command specific flags for REQ_OP_WRITE_ZEROES: */
> - __REQ_NOUNMAP, /* do not free blocks when zeroing */
> + __REQ_DRV, /* for driver use */
>
> - __REQ_HIPRI,
> + /* command specific flag for REQ_OP_WRITE_ZEROES: */
> + __REQ_NOUNMAP, /* do not free blocks when zeroing */
>
> - /* for driver use */
> - __REQ_DRV,
> - __REQ_SWAP, /* swapping request. */
> __REQ_NR_BITS, /* stops here */
> };
>
>
--
Damien Le Moal
Western Digital Research
