| Message ID | 20190408110238.30478-1-ming.lei@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [V3] block: don't use for-inside-for in bio_for_each_segment_all | expand |
Looks good,
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
On 4/8/19 5:02 AM, Ming Lei wrote: > Commit 6dc4f100c175 ("block: allow bio_for_each_segment_all() to > iterate over multi-page bvec") changes bio_for_each_segment_all() > to use for-inside-for. > > This way breaks all bio_for_each_segment_all() call with error out > branch via 'break', since now 'break' can only break from the inner > loop. > > Fixes this issue by implementing bio_for_each_segment_all() via > single 'for' loop, and now the logic is very similar with normal > bvec iterator. Applied, thanks Ming.
On Mon, Apr 08, 2019 at 08:19:59AM -0600, Jens Axboe wrote: > On 4/8/19 5:02 AM, Ming Lei wrote: > > Commit 6dc4f100c175 ("block: allow bio_for_each_segment_all() to > > iterate over multi-page bvec") changes bio_for_each_segment_all() > > to use for-inside-for. > > > > This way breaks all bio_for_each_segment_all() call with error out > > branch via 'break', since now 'break' can only break from the inner > > loop. > > > > Fixes this issue by implementing bio_for_each_segment_all() via > > single 'for' loop, and now the logic is very similar with normal > > bvec iterator. > > Applied, thanks Ming. What's the merge target please? This affects btrfs testing so I wonder if we should start adding workarounds or just wait for the next rc. Thanks.
On 4/11/19 7:26 AM, David Sterba wrote: > On Mon, Apr 08, 2019 at 08:19:59AM -0600, Jens Axboe wrote: >> On 4/8/19 5:02 AM, Ming Lei wrote: >>> Commit 6dc4f100c175 ("block: allow bio_for_each_segment_all() to >>> iterate over multi-page bvec") changes bio_for_each_segment_all() >>> to use for-inside-for. >>> >>> This way breaks all bio_for_each_segment_all() call with error out >>> branch via 'break', since now 'break' can only break from the inner >>> loop. >>> >>> Fixes this issue by implementing bio_for_each_segment_all() via >>> single 'for' loop, and now the logic is very similar with normal >>> bvec iterator. >> >> Applied, thanks Ming. > > What's the merge target please? This affects btrfs testing so I wonder > if we should start adding workarounds or just wait for the next rc. > Thanks. It's queued up in my for-linus, so it's for this series. It'll go upstream tomorrow.
diff --git a/include/linux/bio.h b/include/linux/bio.h index bb6090aa165d..e584673c1881 100644 --- a/include/linux/bio.h +++ b/include/linux/bio.h @@ -120,19 +120,23 @@ static inline bool bio_full(struct bio *bio) return bio->bi_vcnt >= bio->bi_max_vecs; } -#define mp_bvec_for_each_segment(bv, bvl, i, iter_all) \ - for (bv = bvec_init_iter_all(&iter_all); \ - (iter_all.done < (bvl)->bv_len) && \ - (mp_bvec_next_segment((bvl), &iter_all), 1); \ - iter_all.done += bv->bv_len, i += 1) +static inline bool bio_next_segment(const struct bio *bio, + struct bvec_iter_all *iter) +{ + if (iter->idx >= bio->bi_vcnt) + return false; + + bvec_advance(&bio->bi_io_vec[iter->idx], iter); + return true; +} /* * drivers should _never_ use the all version - the bio may have been split * before it got to the driver and the driver won't own all of it */ -#define bio_for_each_segment_all(bvl, bio, i, iter_all) \ - for (i = 0, iter_all.idx = 0; iter_all.idx < (bio)->bi_vcnt; iter_all.idx++) \ - mp_bvec_for_each_segment(bvl, &((bio)->bi_io_vec[iter_all.idx]), i, iter_all) +#define bio_for_each_segment_all(bvl, bio, i, iter) \ + for (i = 0, bvl = bvec_init_iter_all(&iter); \ + bio_next_segment((bio), &iter); i++) static inline void bio_advance_iter(struct bio *bio, struct bvec_iter *iter, unsigned bytes) diff --git a/include/linux/bvec.h b/include/linux/bvec.h index f6275c4da13a..3bc91879e1e2 100644 --- a/include/linux/bvec.h +++ b/include/linux/bvec.h @@ -145,18 +145,18 @@ static inline bool bvec_iter_advance(const struct bio_vec *bv, static inline struct bio_vec *bvec_init_iter_all(struct bvec_iter_all *iter_all) { - iter_all->bv.bv_page = NULL; iter_all->done = 0; + iter_all->idx = 0; return &iter_all->bv; } -static inline void mp_bvec_next_segment(const struct bio_vec *bvec, - struct bvec_iter_all *iter_all) +static inline void bvec_advance(const struct bio_vec *bvec, + struct bvec_iter_all *iter_all) { struct bio_vec *bv = &iter_all->bv; - if (bv->bv_page) { + if (iter_all->done) { bv->bv_page = nth_page(bv->bv_page, 1); bv->bv_offset = 0; } else { @@ -165,6 +165,12 @@ static inline void mp_bvec_next_segment(const struct bio_vec *bvec, } bv->bv_len = min_t(unsigned int, PAGE_SIZE - bv->bv_offset, bvec->bv_len - iter_all->done); + iter_all->done += bv->bv_len; + + if (iter_all->done == bvec->bv_len) { + iter_all->idx++; + iter_all->done = 0; + } } /*
Commit 6dc4f100c175 ("block: allow bio_for_each_segment_all() to iterate over multi-page bvec") changes bio_for_each_segment_all() to use for-inside-for. This way breaks all bio_for_each_segment_all() call with error out branch via 'break', since now 'break' can only break from the inner loop. Fixes this issue by implementing bio_for_each_segment_all() via single 'for' loop, and now the logic is very similar with normal bvec iterator. Cc: Qu Wenruo <quwenruo.btrfs@gmx.com> Cc: linux-btrfs@vger.kernel.org Cc: linux-fsdevel@vger.kernel.org Cc: Omar Sandoval <osandov@fb.com> Cc: Christoph Hellwig <hch@lst.de> Reported-and-Tested-by: Qu Wenruo <quwenruo.btrfs@gmx.com> Fixes: 6dc4f100c175 ("block: allow bio_for_each_segment_all() to iterate over multi-page bvec") Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Ming Lei <ming.lei@redhat.com> --- V3: - don't rename for avoiding conflict with hch's patches V2: - address comments: avoid explicated expressiones in macro, naming include/linux/bio.h | 20 ++++++++++++-------- include/linux/bvec.h | 14 ++++++++++---- 2 files changed, 22 insertions(+), 12 deletions(-)