Re: [PATCH] block: fail unaligned bio from submit_bio_noacct()

[Ming Lei <ming.lei@redhat.com>]

On Thu, Mar 21, 2024 at 08:14:24AM -0700, Bart Van Assche wrote:
> On 3/21/24 06:16, Ming Lei wrote:
> > +static bool bio_check_alignment(struct bio *bio, struct request_queue *q)
> > +{
> > +	unsigned int bs = q->limits.logical_block_size;
> > +	unsigned int size = bio->bi_iter.bi_size;
> > +
> > +	if (size & (bs - 1))
> > +		return false;
> > +
> > +	if (size && ((bio->bi_iter.bi_sector << SECTOR_SHIFT) & (bs - 1)))
> > +		return false;
> Why "size &&"? It doesn't harm to reject unaligned bios if size == 0 and
> it will reduce the number of if-tests in the hot path.

It doesn't make sense to check the alignment for bio without data.

> 
> Why to shift bio->bi_iter.bi_sector left instead of shifting (bs - 1)
> right?

unit of bs is bytes, so .bi_sector needs to be converted to byte first.


Thanks,
Ming