On Tue, May 02, 2017 at 01:42:24PM +1000, NeilBrown wrote:
> Since commit 23688bf4f830 ("block: ensure to split after potentially
> bouncing a bio") blk_queue_bounce() is called *before*
> blk_queue_split().
> This means that:
> 1/ the comments blk_queue_split() about bounce buffers are
> irrelevant, and
> 2/ a very large bio (more than BIO_MAX_PAGES) will no longer be
> split before it arrives at blk_queue_bounce(), leading to the
> possibility that bio_clone_bioset() will fail and a NULL
> will be dereferenced.
>
> Separately, blk_queue_bounce() shouldn't use fs_bio_set as the bio
> being copied could be from the same set, and this could lead to a
> deadlock.
>
> So:
> - allocate 2 private biosets for blk_queue_bounce, one for
> splitting enormous bios and one for cloning bios.
> - add code to split a bio that exceeds BIO_MAX_PAGES.
> - Fix up the comments in blk_queue_split()
>
> Credit-to: Ming Lei <tom.leim...@gmail.com> (suggested using single
> bio_for_each_segment loop)
> Signed-off-by: NeilBrown <ne...@suse.com>
> ---
> block/blk-merge.c | 14 ++++----------
> block/bounce.c | 32 ++++++++++++++++++++++++++------
> 2 files changed, 30 insertions(+), 16 deletions(-)
>
> diff --git a/block/blk-merge.c b/block/blk-merge.c
> index d59074556703..51c84540d3bb 100644
> --- a/block/blk-merge.c
> +++ b/block/blk-merge.c
> @@ -117,17 +117,11 @@ static struct bio *blk_bio_segment_split(struct
> request_queue *q,
> * each holds at most BIO_MAX_PAGES bvecs because
> * bio_clone() can fail to allocate big bvecs.
> *
> - * It should have been better to apply the limit per
> - * request queue in which bio_clone() is involved,
> - * instead of globally. The biggest blocker is the
> - * bio_clone() in bio bounce.
> + * Those drivers which will need to use bio_clone()
> + * should tell us in some way. For now, impose the
> + * BIO_MAX_PAGES limit on all queues.
> *
> - * If bio is splitted by this reason, we should have
> - * allowed to continue bios merging, but don't do
> - * that now for making the change simple.
> - *
> - * TODO: deal with bio bounce's bio_clone() gracefully
> - * and convert the global limit into per-queue limit.
> + * TODO: handle users of bio_clone() differently.
> */
> if (bvecs++ >= BIO_MAX_PAGES)
> goto split;
> diff --git a/block/bounce.c b/block/bounce.c
> index 1cb5dd3a5da1..087ecc2dc66c 100644
> --- a/block/bounce.c
> +++ b/block/bounce.c
> @@ -26,6 +26,7 @@
> #define POOL_SIZE 64
> #define ISA_POOL_SIZE 16
>
> +struct bio_set *bounce_bio_set, *bounce_bio_split;
> static mempool_t *page_pool, *isa_page_pool;
>
> #if defined(CONFIG_HIGHMEM) || defined(CONFIG_NEED_BOUNCE_POOL)
> @@ -40,6 +41,14 @@ static __init int init_emergency_pool(void)
> BUG_ON(!page_pool);
> pr_info("pool size: %d pages\n", POOL_SIZE);
>
> + bounce_bio_set = bioset_create(BIO_POOL_SIZE, 0, BIOSET_NEED_BVECS);
> + BUG_ON(!bounce_bio_set);
> + if (bioset_integrity_create(bounce_bio_set, BIO_POOL_SIZE))
> + BUG_ON(1);
> +
> + bounce_bio_split = bioset_create(BIO_POOL_SIZE, 0, 0);
> + BUG_ON(!bounce_bio_split);
> +
> return 0;
> }
>
> @@ -186,15 +195,26 @@ static void __blk_queue_bounce(struct request_queue *q,
> struct bio **bio_orig,
> int rw = bio_data_dir(*bio_orig);
> struct bio_vec *to, from;
> struct bvec_iter iter;
> - unsigned i;
> + unsigned i = 0;
> + bool bounce = false;
> + int sectors = 0;
>
> - bio_for_each_segment(from, *bio_orig, iter)
> + bio_for_each_segment(from, *bio_orig, iter) {
> + if (i++ < BIO_MAX_PAGES)
> + sectors += from.bv_len >> 9;
> if (page_to_pfn(from.bv_page) > queue_bounce_pfn(q))
> - goto bounce;
> + bounce = true;
> + }
> + if (!bounce)
> + return;
>
> - return;
> -bounce:
> - bio = bio_clone_bioset(*bio_orig, GFP_NOIO, fs_bio_set);
> + if (sectors < bio_sectors(*bio_orig)) {
> + bio = bio_split(*bio_orig, sectors, GFP_NOIO, bounce_bio_split);
> + bio_chain(bio, *bio_orig);
> + generic_make_request(*bio_orig);
> + *bio_orig = bio;
> + }
> + bio = bio_clone_bioset(*bio_orig, GFP_NOIO, bounce_bio_set);
>
> bio_for_each_segment_all(to, bio, i) {
> struct page *page = to->bv_page;
Reviewed-by: Ming Lei <ming....@redhat.com>
Thanks,
Ming
