If queue_dma_alignment(rq->q), then blk_rq_map_user_iov() will call bio_copy_user_iov() and not bio_map_user_iov(). So, bio_map_user_iov() does not need to handle the queue_dma_alignment(rq->q) case in any special way. Remove this dead branch from bio_map_user_iov(). Suggested-by: Christoph Hellwig <hch@xxxxxx> Signed-off-by: Lukas Bulwahn <lukas.bulwahn@xxxxxxxxx> Reviewed-by: Christoph Hellwig <hch@xxxxxx> --- still applies on next-20220307 added Christoph's Reviewed-by tag from: https://lore.kernel.org/all/20220104124724.GA21670@xxxxxx/ Jens, this patch must have fell through the cracks... please pick this minor non-urgent clean-up patch. block/blk-map.c | 41 ++++++++++++++++++----------------------- 1 file changed, 18 insertions(+), 23 deletions(-) diff --git a/block/blk-map.c b/block/blk-map.c index c7f71d83eff1..2a1d7c5f791e 100644 --- a/block/blk-map.c +++ b/block/blk-map.c @@ -260,31 +260,26 @@ static int bio_map_user_iov(struct request *rq, struct iov_iter *iter, npages = DIV_ROUND_UP(offs + bytes, PAGE_SIZE); - if (unlikely(offs & queue_dma_alignment(rq->q))) { - ret = -EINVAL; - j = 0; - } else { - for (j = 0; j < npages; j++) { - struct page *page = pages[j]; - unsigned int n = PAGE_SIZE - offs; - bool same_page = false; - - if (n > bytes) - n = bytes; - - if (!bio_add_hw_page(rq->q, bio, page, n, offs, - max_sectors, &same_page)) { - if (same_page) - put_page(page); - break; - } - - added += n; - bytes -= n; - offs = 0; + for (j = 0; j < npages; j++) { + struct page *page = pages[j]; + unsigned int n = PAGE_SIZE - offs; + bool same_page = false; + + if (n > bytes) + n = bytes; + + if (!bio_add_hw_page(rq->q, bio, page, n, offs, + max_sectors, &same_page)) { + if (same_page) + put_page(page); + break; } - iov_iter_advance(iter, added); + + added += n; + bytes -= n; + offs = 0; } + iov_iter_advance(iter, added); /* * release the pages we didn't map into the bio, if any */ -- 2.17.1
