Skip to contentSkip to conversation

block: return the correct bvec when checking for gaps

PATCH Superseded [email protected] 3 participants 5 messages Jun 3, 2021 – Jun 4, 2021

Subsystems (2)

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 block: return the correct bvec when checking for gaps

From: Long Li <[email protected]>

After commit 07173c3ec276 ("block: enable multipage bvecs"), a bvec can have multiple pages. But bio_will_gap() still assumes one page bvec while checking for merging. This causes data corruption on drivers relying on the correct merging on virt_boundary_mask.

Fix this by returning the multi-page bvec for testing gaps for merging.

Cc Jens Axboe <[email protected]>
Cc Johannes Thumshirn <[email protected]>
Cc Pavel Begunkov <[email protected]>
Cc Ming Lei <[email protected]>
Cc Tejun Heo <[email protected]>
Cc "Matthew Wilcox (Oracle)" <[email protected]>
Cc Jeffle Xu <[email protected]>
Fixes 07173c3ec276 ("block: enable multipage bvecs")
---
 include/linux/bio.h | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)
include/linux/bio.h +4 -7
index a0b4cfdf62a4..6b2f609ccfbf 100644
@@ -44,9 +44,6 @@ static inline unsigned int bio_max_segs(unsigned int nr_segs)
 #define bio_offset(bio)		bio_iter_offset((bio), (bio)->bi_iter)
 #define bio_iovec(bio)		bio_iter_iovec((bio), (bio)->bi_iter)
 
-#define bio_multiple_segments(bio)				\
-	((bio)->bi_iter.bi_size != bio_iovec(bio).bv_len)
-
 #define bvec_iter_sectors(iter)	((iter).bi_size >> 9)
 #define bvec_iter_end_sector(iter) ((iter).bi_sector + bvec_iter_sectors((iter)))
 
@@ -271,7 +268,7 @@ static inline void bio_clear_flag(struct bio *bio, unsigned int bit)
 
 static inline void bio_get_first_bvec(struct bio *bio, struct bio_vec *bv)
 {
-	*bv = bio_iovec(bio);
+	*bv = mp_bvec_iter_bvec(bio->bi_io_vec, bio->bi_iter);
 }
 
 static inline void bio_get_last_bvec(struct bio *bio, struct bio_vec *bv)
@@ -279,10 +276,10 @@ static inline void bio_get_last_bvec(struct bio *bio, struct bio_vec *bv)
 	struct bvec_iter iter = bio->bi_iter;
 	int idx;
 
-	if (unlikely(!bio_multiple_segments(bio))) {
-		*bv = bio_iovec(bio);
+	/* this bio has only one bvec */
+	*bv = mp_bvec_iter_bvec(bio->bi_io_vec, bio->bi_iter);
+	if (bv->bv_len == bio->bi_iter.bi_size)
 		return;
-	}
 
 	bio_advance_iter(bio, &iter, iter.bi_size);
 
Signature · 1 line
2.17.1
#1 lore raw

Hello Long,

Quoted 6 lines · On Thu, Jun 03, 2021 at 03:34:31PM -0700, [email protected] wrote:

From: Long Li <[email protected]>

After commit 07173c3ec276 ("block: enable multipage bvecs"), a bvec can have multiple pages. But bio_will_gap() still assumes one page bvec while checking for merging. This causes data corruption on drivers relying on the correct merging on virt_boundary_mask.

Can you explain the data corruption a bit?

IMO, either single page bvec or multipage bvec should be fine, because bio_will_gap() just checks if the last bvec of prev bio and the 1st bvec of next bio can be merged.

Quoted 44 lines

Fix this by returning the multi-page bvec for testing gaps for merging.

Cc Jens Axboe <[email protected]>
Cc Johannes Thumshirn <[email protected]>
Cc Pavel Begunkov <[email protected]>
Cc Ming Lei <[email protected]>
Cc Tejun Heo <[email protected]>
Cc "Matthew Wilcox (Oracle)" <[email protected]>
Cc Jeffle Xu <[email protected]>
Fixes 07173c3ec276 ("block: enable multipage bvecs")
---
 include/linux/bio.h | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)
include/linux/bio.h +1 -4
index a0b4cfdf62a4..6b2f609ccfbf 100644
@@ -44,9 +44,6 @@ static inline unsigned int bio_max_segs(unsigned int nr_segs)
 #define bio_offset(bio)		bio_iter_offset((bio), (bio)->bi_iter)
 #define bio_iovec(bio)		bio_iter_iovec((bio), (bio)->bi_iter)
 
-#define bio_multiple_segments(bio)				\
-	((bio)->bi_iter.bi_size != bio_iovec(bio).bv_len)
-
 #define bvec_iter_sectors(iter)	((iter).bi_size >> 9)
 #define bvec_iter_end_sector(iter) ((iter).bi_sector + bvec_iter_sectors((iter)))
 
@@ -271,7 +268,7 @@ static inline void bio_clear_flag(struct bio *bio, unsigned int bit)
 
 static inline void bio_get_first_bvec(struct bio *bio, struct bio_vec *bv)
 {
-	*bv = bio_iovec(bio);
+	*bv = mp_bvec_iter_bvec(bio->bi_io_vec, bio->bi_iter);
 }
 
 static inline void bio_get_last_bvec(struct bio *bio, struct bio_vec *bv)
@@ -279,10 +276,10 @@ static inline void bio_get_last_bvec(struct bio *bio, struct bio_vec *bv)
 	struct bvec_iter iter = bio->bi_iter;
 	int idx;

include/linux/bio.h static inline void bio_get_last_bvec(struct bio *bio, struct bio_vec *bv) · line 284

Quoted 10 lines
include/linux/bio.h +3 -3
 
-	if (unlikely(!bio_multiple_segments(bio))) {
-		*bv = bio_iovec(bio);
+	/* this bio has only one bvec */
+	*bv = mp_bvec_iter_bvec(bio->bi_io_vec, bio->bi_iter);
+	if (bv->bv_len == bio->bi_iter.bi_size)
 		return;
-	}
 
 	bio_advance_iter(bio, &iter, iter.bi_size);

The patch itself looks fine, given both bio_get_first_bvec() and bio_get_last_bvec() are used in bio_will_gap() only.

Thanks,
Ming

#2 lore raw
Quoted 12 lines

Subject: Re: [PATCH] block: return the correct bvec when checking for gaps

Hello Long,

Quoted 6 lines · On Thu, Jun 03, 2021 at 03:34:31PM -0700, [email protected] wrote:

From: Long Li <[email protected]>

After commit 07173c3ec276 ("block: enable multipage bvecs"), a bvec can have multiple pages. But bio_will_gap() still assumes one page bvec while checking for merging. This causes data corruption on drivers relying on the correct merging on virt_boundary_mask.

Quoted 5 lines

Can you explain the data corruption a bit?

IMO, either single page bvec or multipage bvec should be fine, because bio_will_gap() just checks if the last bvec of prev bio and the 1st bvec of next bio can be merged.

Hi Ming,

When bio_will_gap() calls into biovec_phys_mergeable (), seg_boundary_mask (queue_segment_boundary()) is used to test if the two bio_vecs can be merged. This test can succeed if only the 1st page in bvec is used, but at the same time it can fail if all the pages in bvec are used. In other words, if the pages in bvec go across the seg_boundary_mask, the test can potentially succeed if only the 1st page is tested, but can fail if all the pages are tested.

Later, when SCSI builds the SG list from BIOs (that calls into __blk_bios_map_sg), __blk_segment_map_sg_merge() calls biovec_phys_mergeable() doing the same test . This time it may fail if the pages in bvec go across the seg_boundary_mask (but tested okay in bio_will_gap() earlier, so those two BIOs were merged). If __blk_segment_map_sg_merge() fails, we end up with a broken SG list for drivers assuming the SG list not having offsets in intermediate pages.

In practice, usually a duplicate page (because merging fails) is put to the SG list. This page and all the pages afterwards in the SG list end up writing to the wrong sectors on disk.

Thanks,
Long

Quoted 66 lines

include/linux/bio.h static inline unsigned int bio_max_segs(unsigned int

Quoted 24 lines

Fix this by returning the multi-page bvec for testing gaps for merging.

Cc Jens Axboe <[email protected]>
Cc Johannes Thumshirn <[email protected]>
Cc Pavel Begunkov <[email protected]>
Cc Ming Lei <[email protected]>
Cc Tejun Heo <[email protected]>
Cc "Matthew Wilcox (Oracle)" <[email protected]>
Cc Jeffle Xu <[email protected]>
Fixes 07173c3ec276 ("block: enable multipage bvecs")
---
 include/linux/bio.h | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)
include/linux/bio.h index renamed from include/linux/bio.h +0 -0

a0b4cfdf62a4..6b2f609ccfbf 100644

include/linux/bio.h +0 -0
@@ -44,9 +44,6 @@ static inline unsigned int bio_max_segs(unsigned int

nr_segs)

Quoted 19 lines
include/linux/bio.h +1 -4
 #define bio_offset(bio)		bio_iter_offset((bio), (bio)->bi_iter)
 #define bio_iovec(bio)		bio_iter_iovec((bio), (bio)->bi_iter)
 
-#define bio_multiple_segments(bio)				\
-	((bio)->bi_iter.bi_size != bio_iovec(bio).bv_len)
-
 #define bvec_iter_sectors(iter)	((iter).bi_size >> 9)
 #define bvec_iter_end_sector(iter) ((iter).bi_sector +
 bvec_iter_sectors((iter)))
@@ -271,7 +268,7 @@ static inline void bio_clear_flag(struct bio *bio,
 unsigned int bit)
 
 static inline void bio_get_first_bvec(struct bio *bio, struct bio_vec
 *bv)  {
-	*bv = bio_iovec(bio);
+	*bv = mp_bvec_iter_bvec(bio->bi_io_vec, bio->bi_iter);
 }
 
Quoted 2 lines

static inline void bio_get_last_bvec(struct bio *bio, struct bio_vec
*bv) @@ -279,10 +276,10 @@ static inline void bio_get_last_bvec(struct

bio *bio, struct bio_vec *bv)

include/linux/bio.h static inline void bio_get_last_bvec(struct bio *bio, struct bio_vec *bv) · line 284

Quoted 12 lines
include/linux/bio.h +3 -3
 	struct bvec_iter iter = bio->bi_iter;
 	int idx;
 
-	if (unlikely(!bio_multiple_segments(bio))) {
-		*bv = bio_iovec(bio);
+	/* this bio has only one bvec */
+	*bv = mp_bvec_iter_bvec(bio->bi_io_vec, bio->bi_iter);
+	if (bv->bv_len == bio->bi_iter.bi_size)
 		return;
-	}
 
 	bio_advance_iter(bio, &iter, iter.bi_size);

The patch itself looks fine, given both bio_get_first_bvec() and bio_get_last_bvec() are used in bio_will_gap() only.

Thanks,
Ming

#3 lore raw
Quoted 18 lines · Long Li
Quoted 12 lines

Subject: Re: [PATCH] block: return the correct bvec when checking for gaps

Hello Long,

Quoted 6 lines · On Thu, Jun 03, 2021 at 03:34:31PM -0700, [email protected] wrote:

From: Long Li <[email protected]>

After commit 07173c3ec276 ("block: enable multipage bvecs"), a bvec can have multiple pages. But bio_will_gap() still assumes one page bvec while checking for merging. This causes data corruption on drivers relying on the correct merging on virt_boundary_mask.

Quoted 5 lines

Can you explain the data corruption a bit?

IMO, either single page bvec or multipage bvec should be fine, because bio_will_gap() just checks if the last bvec of prev bio and the 1st bvec of next bio can be merged.

Quoted 6 lines

Hi Ming,

When bio_will_gap() calls into biovec_phys_mergeable (), seg_boundary_mask (queue_segment_boundary()) is used to test if the two bio_vecs can be merged. This test can succeed if only the 1st page in bvec is used, but at the same time it can fail if all the pages in bvec are used. In other words, if the pages in bvec go across the seg_boundary_mask, the test can potentially succeed if only the 1st page is tested, but can fail if all the pages are tested.

Later, when SCSI builds the SG list from BIOs (that calls into __blk_bios_map_sg), __blk_segment_map_sg_merge() calls biovec_phys_mergeable() doing the same test . This time it may fail if the pages in bvec go across the seg_boundary_mask (but tested okay in bio_will_gap() earlier, so those two BIOs were merged). If __blk_segment_map_sg_merge() fails, we end up with a broken SG list for drivers assuming the SG list not having offsets in intermediate pages.

OK, the reason is that both bio_will_gap() and __blk_segment_map_sg_merge() have to use same approach to check if two bvecs from two bios can be mergeable.

Now __blk_segment_map_sg_merge() won't merge the 1st bvec of next bio into previous bio if the 1st bvec of next bio crosses segment boundary, so bio_will_gap() has to take same way to check if the two bvecs can be merged.

Please add the segment boundary and map SG list story in commit log, then the patch looks fine.

Thanks,
Ming

#4 lore raw
Quoted 47 lines

Subject: Re: [PATCH] block: return the correct bvec when checking for gaps

Quoted 6 lines · Long Li
Quoted 6 lines

Subject: Re: [PATCH] block: return the correct bvec when checking for gaps

Hello Long,

On Thu, Jun 03, 2021 at 03:34:31PM -0700, [email protected]

Quoted 9 lines · wrote:
Quoted 9 lines
Quoted 6 lines

From: Long Li <[email protected]>

After commit 07173c3ec276 ("block: enable multipage bvecs"), a bvec can have multiple pages. But bio_will_gap() still assumes one page bvec while checking for merging. This causes data corruption on drivers relying on the correct merging on virt_boundary_mask.

Can you explain the data corruption a bit?

Quoted 8 lines
Quoted 4 lines

IMO, either single page bvec or multipage bvec should be fine, because
bio_will_gap() just checks if the last bvec of prev bio and the 1st bvec of next bio can be merged.

Hi Ming,

When bio_will_gap() calls into biovec_phys_mergeable (),

seg_boundary_mask (queue_segment_boundary()) is used to test if the two bio_vecs can be merged. This test can succeed if only the 1st page in bvec is used, but at the same time it can fail if all the pages in bvec are used. In other words, if the pages in bvec go across the seg_boundary_mask, the test can potentially succeed if only the 1st page is tested, but can fail if all the pages are tested.

Quoted 2 lines

Later, when SCSI builds the SG list from BIOs (that calls into

__blk_bios_map_sg), __blk_segment_map_sg_merge() calls biovec_phys_mergeable() doing the same test . This time it may fail if the pages in bvec go across the seg_boundary_mask (but tested okay in bio_will_gap() earlier, so those two BIOs were merged). If __blk_segment_map_sg_merge() fails, we end up with a broken SG list for drivers assuming the SG list not having offsets in intermediate pages.

Quoted 1 line

OK, the reason is that both bio_will_gap() and __blk_segment_map_sg_merge() have to use same approach to check if two bvecs from two bios can be mergeable.

Quoted 6 lines

Now __blk_segment_map_sg_merge() won't merge the 1st bvec of next bio into previous bio if the 1st bvec of next bio crosses segment boundary, so bio_will_gap() has to take same way to check if the two bvecs can be merged.

Please add the segment boundary and map SG list story in commit log, then the patch looks fine.

Sure, I will send v2.

Long

Quoted 4 lines

Thanks,
Ming