Skip to contentSkip to conversation

btrfs: scrub: fix failed to detect checksum error

PATCH In review Li Zhang 2 participants 3 messages Nov 6, 2022 – Nov 7, 2022 Btrfs file system

lore
Share & tools
lore

Li Zhang

author
#0 lore raw

PATCH 1/1 btrfs: scrub: fix failed to detect checksum error

[bug]
Scrub fails to detect checksum errors
To reproduce the problem:

$ truncate -s 250M loop_dev1
$ truncate -s 250M loop_dev2
$ losetup /dev/loop1 loop_dev1
$ losetup /dev/loop2 loop_dev2
$ mkfs.btrfs -mraid1 -draid1 /dev/loop1 /dev/loop2 -f
$ mount /dev/loop1 /mnt/
$ cp ~/btrfs/btrfs-progs/mkfs/main.c /mnt/
$ vim -b loop_dev1

....

         free(label);
         free(source_dir);
         exit(1);
success:
         exit(0);
}zhangli

....

$ sudo btrfs scrub start /mnt/
fsid:b66aa912-ae76-4b4b-881b-6a6840f06870 sock_path:/var/lib/btrfs/scrub.progress.b66aa912-ae76-4b4b-881b-6a6840f06870.
scrub started on /mnt/, fsid b66aa912-ae76-4b4b-881b-6a6840f06870 (pid=9894)
$ sudo btrfs scrub status /mnt/
UUID:             b66aa912-ae76-4b4b-881b-6a6840f06870
Scrub started:    Sun Nov  6 21:51:50 2022
Status:           finished
Duration:         0:00:00
Total to scrub:   392.00KiB
Rate:             0.00B/s
Error summary:    no errors found

[reason]
Because scrub only checks the first sector (scrub_sector) of the sblock (scrub_block), it does not check other sectors.

[implementation]
1. Move the set sector (scrub_sector) csum from scrub_extent to scrub_sectors, since each sector has an independent checksum.
2. In the scrub_checksum_data function, check all sectors in the sblock.
3. In the scrub_setup_recheck_block function,
use sector_index to set the sector csum.

[test]
The test method is the same as the problem reproduction. Can detect checksum errors and fix checksum errors Below is the scrub output.

$ sudo btrfs scrub start /mnt/
fsid:b66aa912-ae76-4b4b-881b-6a6840f06870 sock_path:/var/lib/btrfs/scrub.progress.b66aa912-ae76-4b4b-881b-6a6840f06870.
scrub started on /mnt/, fsid b66aa912-ae76-4b4b-881b-6a6840f06870 (pid=11089)
$ sudo btrfs scrub status /mnt/WARNING: errors detected during scrubbing, corrected
UUID:             b66aa912-ae76-4b4b-881b-6a6840f06870
Scrub started:    Sun Nov  6 22:15:02 2022
Status:           finished
Duration:         0:00:00
Total to scrub:   392.00KiB
Rate:             0.00B/s
Error summary:    csum=1
  Corrected:      1
  Uncorrectable:  0
  Unverified:     0

---
Issue: 537

 fs/btrfs/scrub.c | 58 ++++++++++++++++++++++++++++----------------------------
 1 file changed, 29 insertions(+), 29 deletions(-)
fs/btrfs/scrub.c +29 -29
index f260c53..56ee600 100644
@@ -404,7 +404,7 @@ static int scrub_write_sector_to_dev_replace(struct scrub_block *sblock,
 static void scrub_parity_put(struct scrub_parity *sparity);
 static int scrub_sectors(struct scrub_ctx *sctx, u64 logical, u32 len,
 			 u64 physical, struct btrfs_device *dev, u64 flags,
-			 u64 gen, int mirror_num, u8 *csum,
+			 u64 gen, int mirror_num,
 			 u64 physical_for_dev_replace);
 static void scrub_bio_end_io(struct bio *bio);
 static void scrub_bio_end_io_worker(struct work_struct *work);
@@ -420,6 +420,8 @@ static int scrub_add_sector_to_wr_bio(struct scrub_ctx *sctx,
 static void scrub_wr_bio_end_io(struct bio *bio);
 static void scrub_wr_bio_end_io_worker(struct work_struct *work);
 static void scrub_put_ctx(struct scrub_ctx *sctx);
+static int scrub_find_csum(struct scrub_ctx *sctx, u64 logical, u8 *csum);
+
 
 static inline int scrub_is_page_on_raid56(struct scrub_sector *sector)
 {
@@ -1516,7 +1518,7 @@ static int scrub_setup_recheck_block(struct scrub_block *original_sblock,
 			sector->have_csum = have_csum;
 			if (have_csum)
 				memcpy(sector->csum,
-				       original_sblock->sectors[0]->csum,
+				       original_sblock->sectors[sector_index]->csum,
 				       sctx->fs_info->csum_size);
 
 			scrub_stripe_index_and_offset(logical,
@@ -1984,21 +1986,22 @@ static int scrub_checksum_data(struct scrub_block *sblock)
 	u8 csum[BTRFS_CSUM_SIZE];
 	struct scrub_sector *sector;
 	char *kaddr;
+	int i;
 
 	BUG_ON(sblock->sector_count < 1);
-	sector = sblock->sectors[0];
-	if (!sector->have_csum)
-		return 0;
-
-	kaddr = scrub_sector_get_kaddr(sector);
 
 	shash->tfm = fs_info->csum_shash;
 	crypto_shash_init(shash);
+	for (i = 0; i < sblock->sector_count; i++) {
+		sector = sblock->sectors[i];
+		if (!sector->have_csum)
+			continue;
 
-	crypto_shash_digest(shash, kaddr, fs_info->sectorsize, csum);
-
-	if (memcmp(csum, sector->csum, fs_info->csum_size))
-		sblock->checksum_error = 1;
+		kaddr = scrub_sector_get_kaddr(sector);
+		crypto_shash_digest(shash, kaddr, fs_info->sectorsize, csum);
+		if (memcmp(csum, sector->csum, fs_info->csum_size))
+			sblock->checksum_error = 1;
+	}
 	return sblock->checksum_error;
 }
 
@@ -2400,12 +2403,14 @@ static void scrub_missing_raid56_pages(struct scrub_block *sblock)
 
 static int scrub_sectors(struct scrub_ctx *sctx, u64 logical, u32 len,
 		       u64 physical, struct btrfs_device *dev, u64 flags,
-		       u64 gen, int mirror_num, u8 *csum,
+		       u64 gen, int mirror_num,
 		       u64 physical_for_dev_replace)
 {
 	struct scrub_block *sblock;
 	const u32 sectorsize = sctx->fs_info->sectorsize;
 	int index;
+	u8 csum[BTRFS_CSUM_SIZE];
+	int have_csum;
 
 	sblock = alloc_scrub_block(sctx, dev, logical, physical,
 				   physical_for_dev_replace, mirror_num);
@@ -2424,7 +2429,6 @@ static int scrub_sectors(struct scrub_ctx *sctx, u64 logical, u32 len,
 		 * more memory for PAGE_SIZE > sectorsize case.
 		 */
 		u32 l = min(sectorsize, len);
-
 		sector = alloc_scrub_sector(sblock, logical, GFP_KERNEL);
 		if (!sector) {
 			spin_lock(&sctx->stat_lock);
@@ -2435,11 +2439,16 @@ static int scrub_sectors(struct scrub_ctx *sctx, u64 logical, u32 len,
 		}
 		sector->flags = flags;
 		sector->generation = gen;
-		if (csum) {
-			sector->have_csum = 1;
-			memcpy(sector->csum, csum, sctx->fs_info->csum_size);
-		} else {
-			sector->have_csum = 0;
+		if (flags & BTRFS_EXTENT_FLAG_DATA) {
+			/* push csums to sbio */
+			have_csum = scrub_find_csum(sctx, logical, csum);
+			if (have_csum == 0) {
+				++sctx->stat.no_csum;
+				sector->have_csum = 0;
+			} else {
+				sector->have_csum = 1;
+				memcpy(sector->csum, csum, sctx->fs_info->csum_size);
+			}
 		}
 		len -= l;
 		logical += l;
@@ -2669,7 +2678,6 @@ static int scrub_extent(struct scrub_ctx *sctx, struct map_lookup *map,
 	u64 src_physical = physical;
 	int src_mirror = mirror_num;
 	int ret;
-	u8 csum[BTRFS_CSUM_SIZE];
 	u32 blocksize;
 
 	/*
@@ -2715,17 +2723,9 @@ static int scrub_extent(struct scrub_ctx *sctx, struct map_lookup *map,
 				     &src_dev, &src_mirror);
 	while (len) {
 		u32 l = min(len, blocksize);
-		int have_csum = 0;
-
-		if (flags & BTRFS_EXTENT_FLAG_DATA) {
-			/* push csums to sbio */
-			have_csum = scrub_find_csum(sctx, logical, csum);
-			if (have_csum == 0)
-				++sctx->stat.no_csum;
-		}
 		ret = scrub_sectors(sctx, logical, l, src_physical, src_dev,
 				    flags, gen, src_mirror,
-				    have_csum ? csum : NULL, physical);
+				    physical);
 		if (ret)
 			return ret;
 		len -= l;
@@ -4155,7 +4155,7 @@ static noinline_for_stack int scrub_supers(struct scrub_ctx *sctx,
 
 		ret = scrub_sectors(sctx, bytenr, BTRFS_SUPER_INFO_SIZE, bytenr,
 				    scrub_dev, BTRFS_EXTENT_FLAG_SUPER, gen, i,
-				    NULL, bytenr);
+				    bytenr);
 		if (ret)
 			return ret;
 	}
Signature · 1 line
1.8.3.1
#1 lore raw
Quoted 37 lines · Li Zhang

[bug]
Scrub fails to detect checksum errors
To reproduce the problem:

$ truncate -s 250M loop_dev1
$ truncate -s 250M loop_dev2
$ losetup /dev/loop1 loop_dev1
$ losetup /dev/loop2 loop_dev2
$ mkfs.btrfs -mraid1 -draid1 /dev/loop1 /dev/loop2 -f
$ mount /dev/loop1 /mnt/
$ cp ~/btrfs/btrfs-progs/mkfs/main.c /mnt/
$ vim -b loop_dev1

....

          free(label);
          free(source_dir);
          exit(1);
success:
          exit(0);
}zhangli

....

$ sudo btrfs scrub start /mnt/
fsid:b66aa912-ae76-4b4b-881b-6a6840f06870 sock_path:/var/lib/btrfs/scrub.progress.b66aa912-ae76-4b4b-881b-6a6840f06870.
scrub started on /mnt/, fsid b66aa912-ae76-4b4b-881b-6a6840f06870 (pid=9894)
$ sudo btrfs scrub status /mnt/
UUID:             b66aa912-ae76-4b4b-881b-6a6840f06870
Scrub started:    Sun Nov  6 21:51:50 2022
Status:           finished
Duration:         0:00:00
Total to scrub:   392.00KiB
Rate:             0.00B/s
Error summary:    no errors found
Quoted 3 lines

[reason]
Because scrub only checks the first sector (scrub_sector) of the sblock (scrub_block), it does not check other sectors.

That's caused by commit 786672e9e1a3 ("btrfs: scrub: use larger block size for data extent scrub"), which enlarged data scrub extent size (previously always sectorsize, thus there will only be one sector per scrub_block, thus it always works before that commit).

I'd prefer a revert before we have better fix.

Quoted 51 lines

[implementation]
1. Move the set sector (scrub_sector) csum from scrub_extent to scrub_sectors, since each sector has an independent checksum.
2. In the scrub_checksum_data function, check all sectors in the sblock.
3. In the scrub_setup_recheck_block function,
use sector_index to set the sector csum.

[test]
The test method is the same as the problem reproduction. Can detect checksum errors and fix checksum errors Below is the scrub output.

$ sudo btrfs scrub start /mnt/
fsid:b66aa912-ae76-4b4b-881b-6a6840f06870 sock_path:/var/lib/btrfs/scrub.progress.b66aa912-ae76-4b4b-881b-6a6840f06870.
scrub started on /mnt/, fsid b66aa912-ae76-4b4b-881b-6a6840f06870 (pid=11089)
$ sudo btrfs scrub status /mnt/WARNING: errors detected during scrubbing, corrected
UUID:             b66aa912-ae76-4b4b-881b-6a6840f06870
Scrub started:    Sun Nov  6 22:15:02 2022
Status:           finished
Duration:         0:00:00
Total to scrub:   392.00KiB
Rate:             0.00B/s
Error summary:    csum=1
   Corrected:      1
   Uncorrectable:  0
   Unverified:     0

---
Issue: 537

  fs/btrfs/scrub.c | 58 ++++++++++++++++++++++++++++----------------------------
  1 file changed, 29 insertions(+), 29 deletions(-)
fs/btrfs/scrub.c +1 -1
index f260c53..56ee600 100644
@@ -404,7 +404,7 @@ static int scrub_write_sector_to_dev_replace(struct scrub_block *sblock,
  static void scrub_parity_put(struct scrub_parity *sparity);
  static int scrub_sectors(struct scrub_ctx *sctx, u64 logical, u32 len,
  			 u64 physical, struct btrfs_device *dev, u64 flags,
-			 u64 gen, int mirror_num, u8 *csum,
+			 u64 gen, int mirror_num,
  			 u64 physical_for_dev_replace);
  static void scrub_bio_end_io(struct bio *bio);
  static void scrub_bio_end_io_worker(struct work_struct *work);

fs/btrfs/scrub.c static int scrub_add_sector_to_wr_bio(struct scrub_ctx *sctx, · line 424

Quoted 5 lines
fs/btrfs/scrub.c +2 -0
@@ -420,6 +420,8 @@ static int scrub_add_sector_to_wr_bio(struct scrub_ctx *sctx,
  static void scrub_wr_bio_end_io(struct bio *bio);
  static void scrub_wr_bio_end_io_worker(struct work_struct *work);
  static void scrub_put_ctx(struct scrub_ctx *sctx);
+static int scrub_find_csum(struct scrub_ctx *sctx, u64 logical, u8 *csum);
+

I don't think this is the way to go, since we can have a extent up to STRIPE_LEN, going csum search again and again is not the proper way to go.

We need a function to grab a batch of csum in just one go.

Quoted 30 lines
fs/btrfs/scrub.c +5 -6
  static inline int scrub_is_page_on_raid56(struct scrub_sector *sector)
  {
@@ -1516,7 +1518,7 @@ static int scrub_setup_recheck_block(struct scrub_block *original_sblock,
  			sector->have_csum = have_csum;
  			if (have_csum)
  				memcpy(sector->csum,
-				       original_sblock->sectors[0]->csum,
+				       original_sblock->sectors[sector_index]->csum,
  				       sctx->fs_info->csum_size);
 
  			scrub_stripe_index_and_offset(logical,
@@ -1984,21 +1986,22 @@ static int scrub_checksum_data(struct scrub_block *sblock)
  	u8 csum[BTRFS_CSUM_SIZE];
  	struct scrub_sector *sector;
  	char *kaddr;
+	int i;
 
  	BUG_ON(sblock->sector_count < 1);
-	sector = sblock->sectors[0];
-	if (!sector->have_csum)
-		return 0;
-
-	kaddr = scrub_sector_get_kaddr(sector);
 
  	shash->tfm = fs_info->csum_shash;
  	crypto_shash_init(shash);
+	for (i = 0; i < sblock->sector_count; i++) {
+		sector = sblock->sectors[i];
+		if (!sector->have_csum)

fs/btrfs/scrub.c static int scrub_checksum_data(struct scrub_block *sblock) · line 2003

Quoted 10 lines
fs/btrfs/scrub.c +5 -4
+			continue;
 
-	crypto_shash_digest(shash, kaddr, fs_info->sectorsize, csum);
-
-	if (memcmp(csum, sector->csum, fs_info->csum_size))
-		sblock->checksum_error = 1;
+		kaddr = scrub_sector_get_kaddr(sector);
+		crypto_shash_digest(shash, kaddr, fs_info->sectorsize, csum);
+		if (memcmp(csum, sector->csum, fs_info->csum_size))
+			sblock->checksum_error = 1;

That would only increase checksum error by 1, but we may have multiple corruptions in that extent.

This hotfix is going to screw up scrub error reporting.

Thanks,
Qu

Quoted 86 lines
fs/btrfs/scrub.c +16 -18
+	}
  	return sblock->checksum_error;
  }
 
@@ -2400,12 +2403,14 @@ static void scrub_missing_raid56_pages(struct scrub_block *sblock)
 
  static int scrub_sectors(struct scrub_ctx *sctx, u64 logical, u32 len,
  		       u64 physical, struct btrfs_device *dev, u64 flags,
-		       u64 gen, int mirror_num, u8 *csum,
+		       u64 gen, int mirror_num,
  		       u64 physical_for_dev_replace)
  {
  	struct scrub_block *sblock;
  	const u32 sectorsize = sctx->fs_info->sectorsize;
  	int index;
+	u8 csum[BTRFS_CSUM_SIZE];
+	int have_csum;
 
  	sblock = alloc_scrub_block(sctx, dev, logical, physical,
  				   physical_for_dev_replace, mirror_num);
@@ -2424,7 +2429,6 @@ static int scrub_sectors(struct scrub_ctx *sctx, u64 logical, u32 len,
  		 * more memory for PAGE_SIZE > sectorsize case.
  		 */
  		u32 l = min(sectorsize, len);
-
  		sector = alloc_scrub_sector(sblock, logical, GFP_KERNEL);
  		if (!sector) {
  			spin_lock(&sctx->stat_lock);
@@ -2435,11 +2439,16 @@ static int scrub_sectors(struct scrub_ctx *sctx, u64 logical, u32 len,
  		}
  		sector->flags = flags;
  		sector->generation = gen;
-		if (csum) {
-			sector->have_csum = 1;
-			memcpy(sector->csum, csum, sctx->fs_info->csum_size);
-		} else {
-			sector->have_csum = 0;
+		if (flags & BTRFS_EXTENT_FLAG_DATA) {
+			/* push csums to sbio */
+			have_csum = scrub_find_csum(sctx, logical, csum);
+			if (have_csum == 0) {
+				++sctx->stat.no_csum;
+				sector->have_csum = 0;
+			} else {
+				sector->have_csum = 1;
+				memcpy(sector->csum, csum, sctx->fs_info->csum_size);
+			}
  		}
  		len -= l;
  		logical += l;
@@ -2669,7 +2678,6 @@ static int scrub_extent(struct scrub_ctx *sctx, struct map_lookup *map,
  	u64 src_physical = physical;
  	int src_mirror = mirror_num;
  	int ret;
-	u8 csum[BTRFS_CSUM_SIZE];
  	u32 blocksize;
 
  	/*
@@ -2715,17 +2723,9 @@ static int scrub_extent(struct scrub_ctx *sctx, struct map_lookup *map,
  				     &src_dev, &src_mirror);
  	while (len) {
  		u32 l = min(len, blocksize);
-		int have_csum = 0;
-
-		if (flags & BTRFS_EXTENT_FLAG_DATA) {
-			/* push csums to sbio */
-			have_csum = scrub_find_csum(sctx, logical, csum);
-			if (have_csum == 0)
-				++sctx->stat.no_csum;
-		}
  		ret = scrub_sectors(sctx, logical, l, src_physical, src_dev,
  				    flags, gen, src_mirror,
-				    have_csum ? csum : NULL, physical);
+				    physical);
  		if (ret)
  			return ret;
  		len -= l;
@@ -4155,7 +4155,7 @@ static noinline_for_stack int scrub_supers(struct scrub_ctx *sctx,
 
  		ret = scrub_sectors(sctx, bytenr, BTRFS_SUPER_INFO_SIZE, bytenr,
  				    scrub_dev, BTRFS_EXTENT_FLAG_SUPER, gen, i,
-				    NULL, bytenr);
+				    bytenr);
  		if (ret)
  			return ret;
  	}
author
#2 lore raw

Yes, I agree.
Would begin to work on it

Thanks,
Li

Quoted 248 lines · Qu Wenruo <[email protected]> 于2022年11月7日周一 06:57写道:
Quoted 37 lines · Li Zhang

[bug]
Scrub fails to detect checksum errors
To reproduce the problem:

$ truncate -s 250M loop_dev1
$ truncate -s 250M loop_dev2
$ losetup /dev/loop1 loop_dev1
$ losetup /dev/loop2 loop_dev2
$ mkfs.btrfs -mraid1 -draid1 /dev/loop1 /dev/loop2 -f
$ mount /dev/loop1 /mnt/
$ cp ~/btrfs/btrfs-progs/mkfs/main.c /mnt/
$ vim -b loop_dev1

....

          free(label);
          free(source_dir);
          exit(1);
success:
          exit(0);
}zhangli

....

$ sudo btrfs scrub start /mnt/
fsid:b66aa912-ae76-4b4b-881b-6a6840f06870 sock_path:/var/lib/btrfs/scrub.progress.b66aa912-ae76-4b4b-881b-6a6840f06870.
scrub started on /mnt/, fsid b66aa912-ae76-4b4b-881b-6a6840f06870 (pid=9894)
$ sudo btrfs scrub status /mnt/
UUID:             b66aa912-ae76-4b4b-881b-6a6840f06870
Scrub started:    Sun Nov  6 21:51:50 2022
Status:           finished
Duration:         0:00:00
Total to scrub:   392.00KiB
Rate:             0.00B/s
Error summary:    no errors found
Quoted 3 lines

[reason]
Because scrub only checks the first sector (scrub_sector) of the sblock (scrub_block), it does not check other sectors.

That's caused by commit 786672e9e1a3 ("btrfs: scrub: use larger block size for data extent scrub"), which enlarged data scrub extent size (previously always sectorsize, thus there will only be one sector per scrub_block, thus it always works before that commit).

I'd prefer a revert before we have better fix.

Quoted 51 lines

[implementation]
1. Move the set sector (scrub_sector) csum from scrub_extent to scrub_sectors, since each sector has an independent checksum.
2. In the scrub_checksum_data function, check all sectors in the sblock.
3. In the scrub_setup_recheck_block function,
use sector_index to set the sector csum.

[test]
The test method is the same as the problem reproduction. Can detect checksum errors and fix checksum errors Below is the scrub output.

$ sudo btrfs scrub start /mnt/
fsid:b66aa912-ae76-4b4b-881b-6a6840f06870 sock_path:/var/lib/btrfs/scrub.progress.b66aa912-ae76-4b4b-881b-6a6840f06870.
scrub started on /mnt/, fsid b66aa912-ae76-4b4b-881b-6a6840f06870 (pid=11089)
$ sudo btrfs scrub status /mnt/WARNING: errors detected during scrubbing, corrected
UUID:             b66aa912-ae76-4b4b-881b-6a6840f06870
Scrub started:    Sun Nov  6 22:15:02 2022
Status:           finished
Duration:         0:00:00
Total to scrub:   392.00KiB
Rate:             0.00B/s
Error summary:    csum=1
   Corrected:      1
   Uncorrectable:  0
   Unverified:     0

---
Issue: 537

  fs/btrfs/scrub.c | 58 ++++++++++++++++++++++++++++----------------------------
  1 file changed, 29 insertions(+), 29 deletions(-)
fs/btrfs/scrub.c +1 -1
index f260c53..56ee600 100644
@@ -404,7 +404,7 @@ static int scrub_write_sector_to_dev_replace(struct scrub_block *sblock,
  static void scrub_parity_put(struct scrub_parity *sparity);
  static int scrub_sectors(struct scrub_ctx *sctx, u64 logical, u32 len,
                       u64 physical, struct btrfs_device *dev, u64 flags,
-                      u64 gen, int mirror_num, u8 *csum,
+                      u64 gen, int mirror_num,
                       u64 physical_for_dev_replace);
  static void scrub_bio_end_io(struct bio *bio);
  static void scrub_bio_end_io_worker(struct work_struct *work);

fs/btrfs/scrub.c static int scrub_add_sector_to_wr_bio(struct scrub_ctx *sctx, · line 424

Quoted 5 lines
fs/btrfs/scrub.c +2 -0
@@ -420,6 +420,8 @@ static int scrub_add_sector_to_wr_bio(struct scrub_ctx *sctx,
  static void scrub_wr_bio_end_io(struct bio *bio);
  static void scrub_wr_bio_end_io_worker(struct work_struct *work);
  static void scrub_put_ctx(struct scrub_ctx *sctx);
+static int scrub_find_csum(struct scrub_ctx *sctx, u64 logical, u8 *csum);
+

I don't think this is the way to go, since we can have a extent up to STRIPE_LEN, going csum search again and again is not the proper way to go.

We need a function to grab a batch of csum in just one go.

Quoted 30 lines
fs/btrfs/scrub.c +5 -6
  static inline int scrub_is_page_on_raid56(struct scrub_sector *sector)
  {
@@ -1516,7 +1518,7 @@ static int scrub_setup_recheck_block(struct scrub_block *original_sblock,
                      sector->have_csum = have_csum;
                      if (have_csum)
                              memcpy(sector->csum,
-                                    original_sblock->sectors[0]->csum,
+                                    original_sblock->sectors[sector_index]->csum,
                                     sctx->fs_info->csum_size);
 
                      scrub_stripe_index_and_offset(logical,
@@ -1984,21 +1986,22 @@ static int scrub_checksum_data(struct scrub_block *sblock)
      u8 csum[BTRFS_CSUM_SIZE];
      struct scrub_sector *sector;
      char *kaddr;
+     int i;
 
      BUG_ON(sblock->sector_count < 1);
-     sector = sblock->sectors[0];
-     if (!sector->have_csum)
-             return 0;
-
-     kaddr = scrub_sector_get_kaddr(sector);
 
      shash->tfm = fs_info->csum_shash;
      crypto_shash_init(shash);
+     for (i = 0; i < sblock->sector_count; i++) {
+             sector = sblock->sectors[i];
+             if (!sector->have_csum)

fs/btrfs/scrub.c static int scrub_checksum_data(struct scrub_block *sblock) · line 2003

Quoted 10 lines
fs/btrfs/scrub.c +5 -4
+                     continue;
 
-     crypto_shash_digest(shash, kaddr, fs_info->sectorsize, csum);
-
-     if (memcmp(csum, sector->csum, fs_info->csum_size))
-             sblock->checksum_error = 1;
+             kaddr = scrub_sector_get_kaddr(sector);
+             crypto_shash_digest(shash, kaddr, fs_info->sectorsize, csum);
+             if (memcmp(csum, sector->csum, fs_info->csum_size))
+                     sblock->checksum_error = 1;

That would only increase checksum error by 1, but we may have multiple corruptions in that extent.

This hotfix is going to screw up scrub error reporting.

Thanks,
Qu

Quoted 86 lines
fs/btrfs/scrub.c +16 -18
+     }
      return sblock->checksum_error;
  }
 
@@ -2400,12 +2403,14 @@ static void scrub_missing_raid56_pages(struct scrub_block *sblock)
 
  static int scrub_sectors(struct scrub_ctx *sctx, u64 logical, u32 len,
                     u64 physical, struct btrfs_device *dev, u64 flags,
-                    u64 gen, int mirror_num, u8 *csum,
+                    u64 gen, int mirror_num,
                     u64 physical_for_dev_replace)
  {
      struct scrub_block *sblock;
      const u32 sectorsize = sctx->fs_info->sectorsize;
      int index;
+     u8 csum[BTRFS_CSUM_SIZE];
+     int have_csum;
 
      sblock = alloc_scrub_block(sctx, dev, logical, physical,
                                 physical_for_dev_replace, mirror_num);
@@ -2424,7 +2429,6 @@ static int scrub_sectors(struct scrub_ctx *sctx, u64 logical, u32 len,
               * more memory for PAGE_SIZE > sectorsize case.
               */
              u32 l = min(sectorsize, len);
-
              sector = alloc_scrub_sector(sblock, logical, GFP_KERNEL);
              if (!sector) {
                      spin_lock(&sctx->stat_lock);
@@ -2435,11 +2439,16 @@ static int scrub_sectors(struct scrub_ctx *sctx, u64 logical, u32 len,
              }
              sector->flags = flags;
              sector->generation = gen;
-             if (csum) {
-                     sector->have_csum = 1;
-                     memcpy(sector->csum, csum, sctx->fs_info->csum_size);
-             } else {
-                     sector->have_csum = 0;
+             if (flags & BTRFS_EXTENT_FLAG_DATA) {
+                     /* push csums to sbio */
+                     have_csum = scrub_find_csum(sctx, logical, csum);
+                     if (have_csum == 0) {
+                             ++sctx->stat.no_csum;
+                             sector->have_csum = 0;
+                     } else {
+                             sector->have_csum = 1;
+                             memcpy(sector->csum, csum, sctx->fs_info->csum_size);
+                     }
              }
              len -= l;
              logical += l;
@@ -2669,7 +2678,6 @@ static int scrub_extent(struct scrub_ctx *sctx, struct map_lookup *map,
      u64 src_physical = physical;
      int src_mirror = mirror_num;
      int ret;
-     u8 csum[BTRFS_CSUM_SIZE];
      u32 blocksize;
 
      /*
@@ -2715,17 +2723,9 @@ static int scrub_extent(struct scrub_ctx *sctx, struct map_lookup *map,
                                   &src_dev, &src_mirror);
      while (len) {
              u32 l = min(len, blocksize);
-             int have_csum = 0;
-
-             if (flags & BTRFS_EXTENT_FLAG_DATA) {
-                     /* push csums to sbio */
-                     have_csum = scrub_find_csum(sctx, logical, csum);
-                     if (have_csum == 0)
-                             ++sctx->stat.no_csum;
-             }
              ret = scrub_sectors(sctx, logical, l, src_physical, src_dev,
                                  flags, gen, src_mirror,
-                                 have_csum ? csum : NULL, physical);
+                                 physical);
              if (ret)
                      return ret;
              len -= l;
@@ -4155,7 +4155,7 @@ static noinline_for_stack int scrub_supers(struct scrub_ctx *sctx,
 
              ret = scrub_sectors(sctx, bytenr, BTRFS_SUPER_INFO_SIZE, bytenr,
                                  scrub_dev, BTRFS_EXTENT_FLAG_SUPER, gen, i,
-                                 NULL, bytenr);
+                                 bytenr);
              if (ret)
                      return ret;
      }