[PATCH STABLE 5.18 0/2] btrfs: raid56 backports to reduce destructive RMW

[PATCH STABLE 5.18 0/2] btrfs: raid56 backports to reduce destructive RMW

[Qu Wenruo]

Hi Greg and Sasha,

This two patches are backports for v5.18 branch.

These two patches are reducing the chance of destructive RMW cycle,
where btrfs can use corrupted data to generate new P/Q, thus making some
repairable data unrepairable.

Those patches are more important than what I initially thought, thus
unfortunately they are not CCed to stable by themselves.

Furthermore due to recent refactors/renames, there are quite some member
change related to those patches, thus have to be manually backported.
(The v5.18 backport is more like the v5.15 backport, with small tweaks
due to member naming change).

One of the fastest way to verify the behavior is the existing btrfs/125
test case from fstests. (not in auto group AFAIK).

Qu Wenruo (2):
  btrfs: only write the sectors in the vertical stripe which has data
    stripes
  btrfs: raid56: don't trust any cached sector in
    __raid56_parity_recover()

 fs/btrfs/raid56.c | 74 ++++++++++++++++++++++++++++++++++++-----------
 1 file changed, 57 insertions(+), 17 deletions(-)

-- 
2.37.0

[PATCH STABLE 5.18 1/2] btrfs: only write the sectors in the vertical stripe which has data stripes

[Qu Wenruo]

commit bd8f7e627703ca5707833d623efcd43f104c7b3f upstream.

If we have only 8K partial write at the beginning of a full RAID56
stripe, we will write the following contents:

                    0  8K           32K             64K
Disk 1	(data):     |XX|            |               |
Disk 2  (data):     |               |               |
Disk 3  (parity):   |XXXXXXXXXXXXXXX|XXXXXXXXXXXXXXX|

|X| means the sector will be written back to disk.

Note that, although we won't write any sectors from disk 2, but we will
write the full 64KiB of parity to disk.

This behavior is fine for now, but not for the future (especially for
RAID56J, as we waste quite some space to journal the unused parity
stripes).

So here we will also utilize the btrfs_raid_bio::dbitmap, anytime we
queue a higher level bio into an rbio, we will update rbio::dbitmap to
indicate which vertical stripes we need to writeback.

And at finish_rmw(), we also check dbitmap to see if we need to write
any sector in the vertical stripe.

So after the patch, above example will only lead to the following
writeback pattern:

                    0  8K           32K             64K
Disk 1	(data):     |XX|            |               |
Disk 2  (data):     |               |               |
Disk 3  (parity):   |XX|            |               |

Cc: stable@vger.kernel.org # 5.18
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
---
 fs/btrfs/raid56.c | 55 +++++++++++++++++++++++++++++++++++++++++++----
 1 file changed, 51 insertions(+), 4 deletions(-)

diff --git a/fs/btrfs/raid56.c b/fs/btrfs/raid56.c
index 0e239a4c3b26..ca92a20d5c27 100644
--- a/fs/btrfs/raid56.c
+++ b/fs/btrfs/raid56.c
@@ -323,6 +323,9 @@ static void merge_rbio(struct btrfs_raid_bio *dest,
 {
 	bio_list_merge(&dest->bio_list, &victim->bio_list);
 	dest->bio_list_bytes += victim->bio_list_bytes;
+	/* Also inherit the bitmaps from @victim. */
+	bitmap_or(dest->dbitmap, victim->dbitmap, dest->dbitmap,
+		  dest->stripe_npages);
 	dest->generic_bio_cnt += victim->generic_bio_cnt;
 	bio_list_init(&victim->bio_list);
 }
@@ -864,6 +867,12 @@ static void rbio_orig_end_io(struct btrfs_raid_bio *rbio, blk_status_t err)
 
 	if (rbio->generic_bio_cnt)
 		btrfs_bio_counter_sub(rbio->bioc->fs_info, rbio->generic_bio_cnt);
+	/*
+	 * Clear the data bitmap, as the rbio may be cached for later usage.
+	 * do this before before unlock_stripe() so there will be no new bio
+	 * for this bio.
+	 */
+	bitmap_clear(rbio->dbitmap, 0, rbio->stripe_npages);
 
 	/*
 	 * At this moment, rbio->bio_list is empty, however since rbio does not
@@ -1195,6 +1204,9 @@ static noinline void finish_rmw(struct btrfs_raid_bio *rbio)
 	else
 		BUG();
 
+	/* We should have at least one data sector. */
+	ASSERT(bitmap_weight(rbio->dbitmap, rbio->stripe_npages));
+
 	/* at this point we either have a full stripe,
 	 * or we've read the full stripe from the drive.
 	 * recalculate the parity and write the new results.
@@ -1266,6 +1278,11 @@ static noinline void finish_rmw(struct btrfs_raid_bio *rbio)
 	for (stripe = 0; stripe < rbio->real_stripes; stripe++) {
 		for (pagenr = 0; pagenr < rbio->stripe_npages; pagenr++) {
 			struct page *page;
+
+			/* This vertical stripe has no data, skip it. */
+			if (!test_bit(pagenr, rbio->dbitmap))
+				continue;
+
 			if (stripe < rbio->nr_data) {
 				page = page_in_rbio(rbio, stripe, pagenr, 1);
 				if (!page)
@@ -1290,6 +1307,11 @@ static noinline void finish_rmw(struct btrfs_raid_bio *rbio)
 
 		for (pagenr = 0; pagenr < rbio->stripe_npages; pagenr++) {
 			struct page *page;
+
+			/* This vertical stripe has no data, skip it. */
+			if (!test_bit(pagenr, rbio->dbitmap))
+				continue;
+
 			if (stripe < rbio->nr_data) {
 				page = page_in_rbio(rbio, stripe, pagenr, 1);
 				if (!page)
@@ -1713,6 +1735,33 @@ static void btrfs_raid_unplug(struct blk_plug_cb *cb, bool from_schedule)
 	run_plug(plug);
 }
 
+/* Add the original bio into rbio->bio_list, and update rbio::dbitmap. */
+static void rbio_add_bio(struct btrfs_raid_bio *rbio, struct bio *orig_bio)
+{
+	const struct btrfs_fs_info *fs_info = rbio->bioc->fs_info;
+	const u64 orig_logical = orig_bio->bi_iter.bi_sector << SECTOR_SHIFT;
+	const u64 full_stripe_start = rbio->bioc->raid_map[0];
+	const u32 orig_len = orig_bio->bi_iter.bi_size;
+	const u32 sectorsize = fs_info->sectorsize;
+	u64 cur_logical;
+
+	ASSERT(orig_logical >= full_stripe_start &&
+	       orig_logical + orig_len <= full_stripe_start +
+	       rbio->nr_data * rbio->stripe_len);
+
+	bio_list_add(&rbio->bio_list, orig_bio);
+	rbio->bio_list_bytes += orig_bio->bi_iter.bi_size;
+
+	/* Update the dbitmap. */
+	for (cur_logical = orig_logical; cur_logical < orig_logical + orig_len;
+	     cur_logical += sectorsize) {
+		int bit = ((u32)(cur_logical - full_stripe_start) >>
+			   fs_info->sectorsize_bits) % rbio->stripe_npages;
+
+		set_bit(bit, rbio->dbitmap);
+	}
+}
+
 /*
  * our main entry point for writes from the rest of the FS.
  */
@@ -1730,9 +1779,8 @@ int raid56_parity_write(struct bio *bio, struct btrfs_io_context *bioc,
 		btrfs_put_bioc(bioc);
 		return PTR_ERR(rbio);
 	}
-	bio_list_add(&rbio->bio_list, bio);
-	rbio->bio_list_bytes = bio->bi_iter.bi_size;
 	rbio->operation = BTRFS_RBIO_WRITE;
+	rbio_add_bio(rbio, bio);
 
 	btrfs_bio_counter_inc_noblocked(fs_info);
 	rbio->generic_bio_cnt = 1;
@@ -2134,8 +2182,7 @@ int raid56_parity_recover(struct bio *bio, struct btrfs_io_context *bioc,
 	}
 
 	rbio->operation = BTRFS_RBIO_READ_REBUILD;
-	bio_list_add(&rbio->bio_list, bio);
-	rbio->bio_list_bytes = bio->bi_iter.bi_size;
+	rbio_add_bio(rbio, bio);
 
 	rbio->faila = find_logical_bio_stripe(rbio, bio);
 	if (rbio->faila == -1) {
-- 
2.37.0

[PATCH STABLE 5.18 2/2] btrfs: raid56: don't trust any cached sector in __raid56_parity_recover()

[Qu Wenruo]

commit f6065f8edeb25f4a9dfe0b446030ad995a84a088 upstream.

[BUG]
There is a small workload which will always fail with recent kernel:
(A simplified version from btrfs/125 test case)

  mkfs.btrfs -f -m raid5 -d raid5 -b 1G $dev1 $dev2 $dev3
  mount $dev1 $mnt
  xfs_io -f -c "pwrite -S 0xee 0 1M" $mnt/file1
  sync
  umount $mnt
  btrfs dev scan -u $dev3
  mount -o degraded $dev1 $mnt
  xfs_io -f -c "pwrite -S 0xff 0 128M" $mnt/file2
  umount $mnt
  btrfs dev scan
  mount $dev1 $mnt
  btrfs balance start --full-balance $mnt
  umount $mnt

The failure is always failed to read some tree blocks:

  BTRFS info (device dm-4): relocating block group 217710592 flags data|raid5
  BTRFS error (device dm-4): parent transid verify failed on 38993920 wanted 9 found 7
  BTRFS error (device dm-4): parent transid verify failed on 38993920 wanted 9 found 7
  ...

[CAUSE]
With the recently added debug output, we can see all RAID56 operations
related to full stripe 38928384:

  56.1183: raid56_read_partial: full_stripe=38928384 devid=2 type=DATA1 offset=0 opf=0x0 physical=9502720 len=65536
  56.1185: raid56_read_partial: full_stripe=38928384 devid=3 type=DATA2 offset=16384 opf=0x0 physical=9519104 len=16384
  56.1185: raid56_read_partial: full_stripe=38928384 devid=3 type=DATA2 offset=49152 opf=0x0 physical=9551872 len=16384
  56.1187: raid56_write_stripe: full_stripe=38928384 devid=3 type=DATA2 offset=0 opf=0x1 physical=9502720 len=16384
  56.1188: raid56_write_stripe: full_stripe=38928384 devid=3 type=DATA2 offset=32768 opf=0x1 physical=9535488 len=16384
  56.1188: raid56_write_stripe: full_stripe=38928384 devid=1 type=PQ1 offset=0 opf=0x1 physical=30474240 len=16384
  56.1189: raid56_write_stripe: full_stripe=38928384 devid=1 type=PQ1 offset=32768 opf=0x1 physical=30507008 len=16384
  56.1218: raid56_write_stripe: full_stripe=38928384 devid=3 type=DATA2 offset=49152 opf=0x1 physical=9551872 len=16384
  56.1219: raid56_write_stripe: full_stripe=38928384 devid=1 type=PQ1 offset=49152 opf=0x1 physical=30523392 len=16384
  56.2721: raid56_parity_recover: full stripe=38928384 eb=39010304 mirror=2
  56.2723: raid56_parity_recover: full stripe=38928384 eb=39010304 mirror=2
  56.2724: raid56_parity_recover: full stripe=38928384 eb=39010304 mirror=2

Before we enter raid56_parity_recover(), we have triggered some metadata
write for the full stripe 38928384, this leads to us to read all the
sectors from disk.

Furthermore, btrfs raid56 write will cache its calculated P/Q sectors to
avoid unnecessary read.

This means, for that full stripe, after any partial write, we will have
stale data, along with P/Q calculated using that stale data.

Thankfully due to patch "btrfs: only write the sectors in the vertical stripe
which has data stripes" we haven't submitted all the corrupted P/Q to disk.

When we really need to recover certain range, aka in
raid56_parity_recover(), we will use the cached rbio, along with its
cached sectors (the full stripe is all cached).

This explains why we have no event raid56_scrub_read_recover()
triggered.

Since we have the cached P/Q which is calculated using the stale data,
the recovered one will just be stale.

In our particular test case, it will always return the same incorrect
metadata, thus causing the same error message "parent transid verify
failed on 39010304 wanted 9 found 7" again and again.

[BTRFS DESTRUCTIVE RMW PROBLEM]

Test case btrfs/125 (and above workload) always has its trouble with
the destructive read-modify-write (RMW) cycle:

        0       32K     64K
Data1:  | Good  | Good  |
Data2:  | Bad   | Bad   |
Parity: | Good  | Good  |

In above case, if we trigger any write into Data1, we will use the bad
data in Data2 to re-generate parity, killing the only chance to recovery
Data2, thus Data2 is lost forever.

This destructive RMW cycle is not specific to btrfs RAID56, but there
are some btrfs specific behaviors making the case even worse:

- Btrfs will cache sectors for unrelated vertical stripes.

  In above example, if we're only writing into 0~32K range, btrfs will
  still read data range (32K ~ 64K) of Data1, and (64K~128K) of Data2.
  This behavior is to cache sectors for later update.

  Incidentally commit d4e28d9b5f04 ("btrfs: raid56: make steal_rbio()
  subpage compatible") has a bug which makes RAID56 to never trust the
  cached sectors, thus slightly improve the situation for recovery.

  Unfortunately, follow up fix "btrfs: update stripe_sectors::uptodate in
  steal_rbio" will revert the behavior back to the old one.

- Btrfs raid56 partial write will update all P/Q sectors and cache them

  This means, even if data at (64K ~ 96K) of Data2 is free space, and
  only (96K ~ 128K) of Data2 is really stale data.
  And we write into that (96K ~ 128K), we will update all the parity
  sectors for the full stripe.

  This unnecessary behavior will completely kill the chance of recovery.

  Thankfully, an unrelated optimization "btrfs: only write the sectors
  in the vertical stripe which has data stripes" will prevent
  submitting the write bio for untouched vertical sectors.

  That optimization will keep the on-disk P/Q untouched for a chance for
  later recovery.

[FIX]
Although we have no good way to completely fix the destructive RMW
(unless we go full scrub for each partial write), we can still limit the
damage.

With patch "btrfs: only write the sectors in the vertical stripe which
has data stripes" now we won't really submit the P/Q of unrelated
vertical stripes, so the on-disk P/Q should still be fine.

Now we really need to do is just drop all the cached sectors when doing
recovery.

By this, we have a chance to read the original P/Q from disk, and have a
chance to recover the stale data, while still keep the cache to speed up
regular write path.

In fact, just dropping all the cache for recovery path is good enough to
allow the test case btrfs/125 along with the small script to pass
reliably.

The lack of metadata write after the degraded mount, and forced metadata
COW is saving us this time.

So this patch will fix the behavior by not trust any cache in
__raid56_parity_recover(), to solve the problem while still keep the
cache useful.

But please note that this test pass DOES NOT mean we have solved the
destructive RMW problem, we just do better damage control a little
better.

Related patches:

- btrfs: only write the sectors in the vertical stripe
- d4e28d9b5f04 ("btrfs: raid56: make steal_rbio() subpage compatible")
- btrfs: update stripe_sectors::uptodate in steal_rbio

Cc: stable@vger.kernel.org # 5.18
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
---
 fs/btrfs/raid56.c | 19 ++++++-------------
 1 file changed, 6 insertions(+), 13 deletions(-)

diff --git a/fs/btrfs/raid56.c b/fs/btrfs/raid56.c
index ca92a20d5c27..39c4c513bf97 100644
--- a/fs/btrfs/raid56.c
+++ b/fs/btrfs/raid56.c
@@ -2084,9 +2084,12 @@ static int __raid56_parity_recover(struct btrfs_raid_bio *rbio)
 	atomic_set(&rbio->error, 0);
 
 	/*
-	 * read everything that hasn't failed.  Thanks to the
-	 * stripe cache, it is possible that some or all of these
-	 * pages are going to be uptodate.
+	 * Read everything that hasn't failed. However this time we will
+	 * not trust any cached sector.
+	 * As we may read out some stale data but higher layer is not reading
+	 * that stale part.
+	 *
+	 * So here we always re-read everything in recovery path.
 	 */
 	for (stripe = 0; stripe < rbio->real_stripes; stripe++) {
 		if (rbio->faila == stripe || rbio->failb == stripe) {
@@ -2095,16 +2098,6 @@ static int __raid56_parity_recover(struct btrfs_raid_bio *rbio)
 		}
 
 		for (pagenr = 0; pagenr < rbio->stripe_npages; pagenr++) {
-			struct page *p;
-
-			/*
-			 * the rmw code may have already read this
-			 * page in
-			 */
-			p = rbio_stripe_page(rbio, stripe, pagenr);
-			if (PageUptodate(p))
-				continue;
-
 			ret = rbio_add_io_page(rbio, &bio_list,
 				       rbio_stripe_page(rbio, stripe, pagenr),
 				       stripe, pagenr, rbio->stripe_len);
-- 
2.37.0

Re: [PATCH STABLE 5.18 0/2] btrfs: raid56 backports to reduce destructive RMW

[Qu Wenruo]

On 2022/8/4 16:10, Qu Wenruo wrote:
> Hi Greg and Sasha,
> 
> This two patches are backports for v5.18 branch.
> 
> These two patches are reducing the chance of destructive RMW cycle,
> where btrfs can use corrupted data to generate new P/Q, thus making some
> repairable data unrepairable.
> 
> Those patches are more important than what I initially thought, thus
> unfortunately they are not CCed to stable by themselves.
> 
> Furthermore due to recent refactors/renames, there are quite some member
> change related to those patches, thus have to be manually backported.
> (The v5.18 backport is more like the v5.15 backport, with small tweaks
> due to member naming change).

Just to mention, thankfully the crash in v5.15 backports are not 
affecting v5.18 backports.

The uninitialized btrfs_raid_bio::fs_info is solved by the commit 
731ccf15c952 ("btrfs: make sure btrfs_io_context::fs_info is always 
initialized"), which is already in the tree.

Thanks,
Qu
> 
> One of the fastest way to verify the behavior is the existing btrfs/125
> test case from fstests. (not in auto group AFAIK).
> 
> Qu Wenruo (2):
>    btrfs: only write the sectors in the vertical stripe which has data
>      stripes
>    btrfs: raid56: don't trust any cached sector in
>      __raid56_parity_recover()
> 
>   fs/btrfs/raid56.c | 74 ++++++++++++++++++++++++++++++++++++-----------
>   1 file changed, 57 insertions(+), 17 deletions(-)
> 

Re: [PATCH STABLE 5.18 0/2] btrfs: raid56 backports to reduce destructive RMW

[Greg KH]

On Thu, Aug 04, 2022 at 04:10:57PM +0800, Qu Wenruo wrote:
> Hi Greg and Sasha,
> 
> This two patches are backports for v5.18 branch.

I also need these for the 5.19.y branch if we were to take them into
5.18.y as you do not want anyone to suffer a regression when moving to
the newer kernel release.

So I'll wait for those to be sent before taking any of these.

thanks,

greg k-h

Re: [PATCH STABLE 5.18 0/2] btrfs: raid56 backports to reduce destructive RMW

[Greg KH]

On Mon, Aug 08, 2022 at 03:39:18PM +0200, Greg KH wrote:
> On Thu, Aug 04, 2022 at 04:10:57PM +0800, Qu Wenruo wrote:
> > Hi Greg and Sasha,
> > 
> > This two patches are backports for v5.18 branch.
> 
> I also need these for the 5.19.y branch if we were to take them into
> 5.18.y as you do not want anyone to suffer a regression when moving to
> the newer kernel release.
> 
> So I'll wait for those to be sent before taking any of these.

I've dropped all of these btrfs backports from my "to review" queue now.
Please fix them all up, get the needed acks, and then resend them and I
will be glad to reconsider them at that point in time.

thanks,

greg k-h

Re: [PATCH STABLE 5.18 0/2] btrfs: raid56 backports to reduce destructive RMW

[Qu Wenruo]

On 2022/8/13 21:13, Greg KH wrote:
> On Mon, Aug 08, 2022 at 03:39:18PM +0200, Greg KH wrote:
>> On Thu, Aug 04, 2022 at 04:10:57PM +0800, Qu Wenruo wrote:
>>> Hi Greg and Sasha,
>>>
>>> This two patches are backports for v5.18 branch.
>>
>> I also need these for the 5.19.y branch if we were to take them into
>> 5.18.y as you do not want anyone to suffer a regression when moving to
>> the newer kernel release.
>>
>> So I'll wait for those to be sent before taking any of these.
>
> I've dropped all of these btrfs backports from my "to review" queue now.
> Please fix them all up, get the needed acks, and then resend them and I
> will be glad to reconsider them at that point in time.

To David,

Mind to give an ACK for these backports?

As mentioned in the pull digest, these two backports are to reduce the
possibility of data corruption of RAID56.

Thanks,
Qu

>
> thanks,
>
> greg k-h

Re: [PATCH STABLE 5.18 0/2] btrfs: raid56 backports to reduce destructive RMW

[David Sterba]

On Sun, Aug 14, 2022 at 06:17:45AM +0800, Qu Wenruo wrote:
> 
> 
> On 2022/8/13 21:13, Greg KH wrote:
> > On Mon, Aug 08, 2022 at 03:39:18PM +0200, Greg KH wrote:
> >> On Thu, Aug 04, 2022 at 04:10:57PM +0800, Qu Wenruo wrote:
> >>> Hi Greg and Sasha,
> >>>
> >>> This two patches are backports for v5.18 branch.
> >>
> >> I also need these for the 5.19.y branch if we were to take them into
> >> 5.18.y as you do not want anyone to suffer a regression when moving to
> >> the newer kernel release.
> >>
> >> So I'll wait for those to be sent before taking any of these.
> >
> > I've dropped all of these btrfs backports from my "to review" queue now.
> > Please fix them all up, get the needed acks, and then resend them and I
> > will be glad to reconsider them at that point in time.
> 
> To David,
> 
> Mind to give an ACK for these backports?

Acked-by: David Sterba <dsterba@suse.com>

[PATCH STABLE 5.18 0/2] btrfs: raid56: backports to reduce corruption

[Qu Wenruo]

This is the backport for v5.18.x stable branch.

The full explananation can be found here:
https://lore.kernel.org/linux-btrfs/Yv85PTBsDhrQITZp@kroah.com/T/#t

Difference between v5.18.x and v5.19.x backports:

- Subpage sector::uptodate related changes
  Since v5.18 and older branches doesn't support subpage RAID56, these
  uptodate checks are now in their older PageUptodate form.

- Various member naming changes
  Like btrfs_raid_bio::stripe_npages -> stripe_nsector.

Qu Wenruo (2):
  btrfs: only write the sectors in the vertical stripe which has data
    stripes
  btrfs: raid56: don't trust any cached sector in
    __raid56_parity_recover()

 fs/btrfs/raid56.c | 74 ++++++++++++++++++++++++++++++++++++-----------
 1 file changed, 57 insertions(+), 17 deletions(-)

-- 
2.37.1

[PATCH STABLE 5.18 2/2] btrfs: raid56: don't trust any cached sector in __raid56_parity_recover()

[Qu Wenruo]

commit f6065f8edeb25f4a9dfe0b446030ad995a84a088 upstream.

[BUG]
There is a small workload which will always fail with recent kernel:
(A simplified version from btrfs/125 test case)

  mkfs.btrfs -f -m raid5 -d raid5 -b 1G $dev1 $dev2 $dev3
  mount $dev1 $mnt
  xfs_io -f -c "pwrite -S 0xee 0 1M" $mnt/file1
  sync
  umount $mnt
  btrfs dev scan -u $dev3
  mount -o degraded $dev1 $mnt
  xfs_io -f -c "pwrite -S 0xff 0 128M" $mnt/file2
  umount $mnt
  btrfs dev scan
  mount $dev1 $mnt
  btrfs balance start --full-balance $mnt
  umount $mnt

The failure is always failed to read some tree blocks:

  BTRFS info (device dm-4): relocating block group 217710592 flags data|raid5
  BTRFS error (device dm-4): parent transid verify failed on 38993920 wanted 9 found 7
  BTRFS error (device dm-4): parent transid verify failed on 38993920 wanted 9 found 7
  ...

[CAUSE]
With the recently added debug output, we can see all RAID56 operations
related to full stripe 38928384:

  56.1183: raid56_read_partial: full_stripe=38928384 devid=2 type=DATA1 offset=0 opf=0x0 physical=9502720 len=65536
  56.1185: raid56_read_partial: full_stripe=38928384 devid=3 type=DATA2 offset=16384 opf=0x0 physical=9519104 len=16384
  56.1185: raid56_read_partial: full_stripe=38928384 devid=3 type=DATA2 offset=49152 opf=0x0 physical=9551872 len=16384
  56.1187: raid56_write_stripe: full_stripe=38928384 devid=3 type=DATA2 offset=0 opf=0x1 physical=9502720 len=16384
  56.1188: raid56_write_stripe: full_stripe=38928384 devid=3 type=DATA2 offset=32768 opf=0x1 physical=9535488 len=16384
  56.1188: raid56_write_stripe: full_stripe=38928384 devid=1 type=PQ1 offset=0 opf=0x1 physical=30474240 len=16384
  56.1189: raid56_write_stripe: full_stripe=38928384 devid=1 type=PQ1 offset=32768 opf=0x1 physical=30507008 len=16384
  56.1218: raid56_write_stripe: full_stripe=38928384 devid=3 type=DATA2 offset=49152 opf=0x1 physical=9551872 len=16384
  56.1219: raid56_write_stripe: full_stripe=38928384 devid=1 type=PQ1 offset=49152 opf=0x1 physical=30523392 len=16384
  56.2721: raid56_parity_recover: full stripe=38928384 eb=39010304 mirror=2
  56.2723: raid56_parity_recover: full stripe=38928384 eb=39010304 mirror=2
  56.2724: raid56_parity_recover: full stripe=38928384 eb=39010304 mirror=2

Before we enter raid56_parity_recover(), we have triggered some metadata
write for the full stripe 38928384, this leads to us to read all the
sectors from disk.

Furthermore, btrfs raid56 write will cache its calculated P/Q sectors to
avoid unnecessary read.

This means, for that full stripe, after any partial write, we will have
stale data, along with P/Q calculated using that stale data.

Thankfully due to patch "btrfs: only write the sectors in the vertical stripe
which has data stripes" we haven't submitted all the corrupted P/Q to disk.

When we really need to recover certain range, aka in
raid56_parity_recover(), we will use the cached rbio, along with its
cached sectors (the full stripe is all cached).

This explains why we have no event raid56_scrub_read_recover()
triggered.

Since we have the cached P/Q which is calculated using the stale data,
the recovered one will just be stale.

In our particular test case, it will always return the same incorrect
metadata, thus causing the same error message "parent transid verify
failed on 39010304 wanted 9 found 7" again and again.

[BTRFS DESTRUCTIVE RMW PROBLEM]

Test case btrfs/125 (and above workload) always has its trouble with
the destructive read-modify-write (RMW) cycle:

        0       32K     64K
Data1:  | Good  | Good  |
Data2:  | Bad   | Bad   |
Parity: | Good  | Good  |

In above case, if we trigger any write into Data1, we will use the bad
data in Data2 to re-generate parity, killing the only chance to recovery
Data2, thus Data2 is lost forever.

This destructive RMW cycle is not specific to btrfs RAID56, but there
are some btrfs specific behaviors making the case even worse:

- Btrfs will cache sectors for unrelated vertical stripes.

  In above example, if we're only writing into 0~32K range, btrfs will
  still read data range (32K ~ 64K) of Data1, and (64K~128K) of Data2.
  This behavior is to cache sectors for later update.

  Incidentally commit d4e28d9b5f04 ("btrfs: raid56: make steal_rbio()
  subpage compatible") has a bug which makes RAID56 to never trust the
  cached sectors, thus slightly improve the situation for recovery.

  Unfortunately, follow up fix "btrfs: update stripe_sectors::uptodate in
  steal_rbio" will revert the behavior back to the old one.

- Btrfs raid56 partial write will update all P/Q sectors and cache them

  This means, even if data at (64K ~ 96K) of Data2 is free space, and
  only (96K ~ 128K) of Data2 is really stale data.
  And we write into that (96K ~ 128K), we will update all the parity
  sectors for the full stripe.

  This unnecessary behavior will completely kill the chance of recovery.

  Thankfully, an unrelated optimization "btrfs: only write the sectors
  in the vertical stripe which has data stripes" will prevent
  submitting the write bio for untouched vertical sectors.

  That optimization will keep the on-disk P/Q untouched for a chance for
  later recovery.

[FIX]
Although we have no good way to completely fix the destructive RMW
(unless we go full scrub for each partial write), we can still limit the
damage.

With patch "btrfs: only write the sectors in the vertical stripe which
has data stripes" now we won't really submit the P/Q of unrelated
vertical stripes, so the on-disk P/Q should still be fine.

Now we really need to do is just drop all the cached sectors when doing
recovery.

By this, we have a chance to read the original P/Q from disk, and have a
chance to recover the stale data, while still keep the cache to speed up
regular write path.

In fact, just dropping all the cache for recovery path is good enough to
allow the test case btrfs/125 along with the small script to pass
reliably.

The lack of metadata write after the degraded mount, and forced metadata
COW is saving us this time.

So this patch will fix the behavior by not trust any cache in
__raid56_parity_recover(), to solve the problem while still keep the
cache useful.

But please note that this test pass DOES NOT mean we have solved the
destructive RMW problem, we just do better damage control a little
better.

Related patches:

- btrfs: only write the sectors in the vertical stripe
- d4e28d9b5f04 ("btrfs: raid56: make steal_rbio() subpage compatible")
- btrfs: update stripe_sectors::uptodate in steal_rbio

Acked-by: David Sterba <dsterba@suse.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
---
 fs/btrfs/raid56.c | 19 ++++++-------------
 1 file changed, 6 insertions(+), 13 deletions(-)

diff --git a/fs/btrfs/raid56.c b/fs/btrfs/raid56.c
index ca92a20d5c27..39c4c513bf97 100644
--- a/fs/btrfs/raid56.c
+++ b/fs/btrfs/raid56.c
@@ -2084,9 +2084,12 @@ static int __raid56_parity_recover(struct btrfs_raid_bio *rbio)
 	atomic_set(&rbio->error, 0);
 
 	/*
-	 * read everything that hasn't failed.  Thanks to the
-	 * stripe cache, it is possible that some or all of these
-	 * pages are going to be uptodate.
+	 * Read everything that hasn't failed. However this time we will
+	 * not trust any cached sector.
+	 * As we may read out some stale data but higher layer is not reading
+	 * that stale part.
+	 *
+	 * So here we always re-read everything in recovery path.
 	 */
 	for (stripe = 0; stripe < rbio->real_stripes; stripe++) {
 		if (rbio->faila == stripe || rbio->failb == stripe) {
@@ -2095,16 +2098,6 @@ static int __raid56_parity_recover(struct btrfs_raid_bio *rbio)
 		}
 
 		for (pagenr = 0; pagenr < rbio->stripe_npages; pagenr++) {
-			struct page *p;
-
-			/*
-			 * the rmw code may have already read this
-			 * page in
-			 */
-			p = rbio_stripe_page(rbio, stripe, pagenr);
-			if (PageUptodate(p))
-				continue;
-
 			ret = rbio_add_io_page(rbio, &bio_list,
 				       rbio_stripe_page(rbio, stripe, pagenr),
 				       stripe, pagenr, rbio->stripe_len);
-- 
2.37.1