Re: [PATCH v2 04/10] block: Handle meta bounce buffer correctly in case of split
From: Christoph Hellwig <hch@lst.de>
Date: 2024-06-27 06:16:17
Also in:
io-uring, linux-nvme
On Wed, Jun 26, 2024 at 03:36:54PM +0530, Anuj Gupta wrote:
quoted hunk ↗ jump to hunk
@@ -105,9 +105,12 @@ static void bio_integrity_unpin_bvec(struct bio_vec *bv, int nr_vecs, static void bio_integrity_uncopy_user(struct bio_integrity_payload *bip) { + struct bio *bio = bip->bip_bio; + struct blk_integrity *bi = blk_get_integrity(bio->bi_bdev->bd_disk); unsigned short nr_vecs = bip->bip_max_vcnt - 1; struct bio_vec *copy = &bip->bip_vec[1]; - size_t bytes = bip->bip_iter.bi_size; + size_t bytes = bio_integrity_bytes(bi, + bvec_iter_sectors(bip->bio_iter));
Maybe add a well documented helper that calculates the metadata bytes based on the iter given that this is probably going to become more common now that we're doing proper cloning?
- bip->bip_flags = bip_src->bip_flags & ~BIP_BLOCK_INTEGRITY; - + bip->bip_flags = bip_src->bip_flags & ~(BIP_BLOCK_INTEGRITY | + BIP_COPY_USER);
We're probably better off say what flags should be cloned and not which ones should not. Preferably with a new #define in bio.h.