On 7/1/2024 10:39 AM, Christoph Hellwig wrote:
> +/*
> + * Integrity payloads can either be owned by the submitter, in which case
> + * bio_uninit will free them, or owned and generated by the block layer,
> + * in which case we'll verify them here (for reads) and free them before
> + * the bio is handed back to the submitted.
> + */
> +bool __bio_integrity_endio(struct bio *bio);
> static inline bool bio_integrity_endio(struct bio *bio)
> {
> - if (bio_integrity(bio))
> + struct bio_integrity_payload *bip = bio_integrity(bio);
> +
> + if (bip && (bip->bip_flags & BIP_BLOCK_INTEGRITY))
> return __bio_integrity_endio(bio);
The patch will cause regression for nvme-passthrough. For that
completion order is:
(a) bio_endio()
(b) req->end_io
(c) blk_rq_unmap_user.
And current code ensures that integrity is freed explicitly only after
(a) and (b).
With the patch, integrity will get freed during (a) itself.
There are two places in bio_endio() that can free the integrity.
It first calls bio_integrity_endio() - which is handled fine above.
But it also calls bio_uninit() - which will free the integrity. We don't
want that to happen before passthrough gets the chance to unpin/copy-back.
