On 2022/05/13 18:13, Keith Busch wrote:
> From: Keith Busch <kbusch@xxxxxxxxxx>
>
> If the iterator has an offset, filling a bio to the max bvecs may result
> in a size that isn't aligned to the block size. Mask off bytes for the
> bio being constructed.
>
> Signed-off-by: Keith Busch <kbusch@xxxxxxxxxx>
> ---
> block/bio.c | 4 ++++
> 1 file changed, 4 insertions(+)
>
> diff --git a/block/bio.c b/block/bio.c
> index 4259125e16ab..b42a9e3ff068 100644
> --- a/block/bio.c
> +++ b/block/bio.c
> @@ -1144,6 +1144,7 @@ static int __bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter)
> {
> unsigned short nr_pages = bio->bi_max_vecs - bio->bi_vcnt;
> unsigned short entries_left = bio->bi_max_vecs - bio->bi_vcnt;
> + struct request_queue *q = bdev_get_queue(bio->bi_bdev);
> struct bio_vec *bv = bio->bi_io_vec + bio->bi_vcnt;
> struct page **pages = (struct page **)bv;
> bool same_page = false;
> @@ -1160,6 +1161,9 @@ static int __bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter)
> pages += entries_left * (PAGE_PTRS_PER_BVEC - 1);
>
> size = iov_iter_get_pages(iter, pages, LONG_MAX, nr_pages, &offset);
> + if (size > 0)
> + size = size & ~(queue_logical_block_size(q) - 1);
I think that __bio_iov_append_get_pages() needs the same change. And given that
both __bio_iov_append_get_pages() and __bio_iov_iter_get_pages() start with
iov_iter_get_pages(), should we do that and check the size in the single caller:
bio_iov_iter_get_pages() ?
> +
> if (unlikely(size <= 0))
> return size ? size : -EFAULT;
>
--
Damien Le Moal
Western Digital Research
