Currently sctx->stat.last_physical only got updated in the following
cases:
- When the last stripe of a non-RAID56 chunk is scrubbed
  This implies a pitfall, if the last stripe is at the chunk boundary,
  and we finished the scrub of the whole chunk, we won't update
  last_physical at all until the next chunk.
- When a P/Q stripe of a RAID56 chunk is scrubbed
This leads the following two problems:
- sctx->stat.last_physical is not updated for a almost full chunk
  This is especially bad, affecting scrub resume, as the resume would
  start from last_physical, causing unnecessary re-scrub.
- "btrfs scrub status" will not report any progress for a long time
Fix the problem by properly updating @last_physical after each stripe is
scrubbed.
And since we're here, for the sake of consistency, use spin lock to
protect the update of @last_physical, just like all the remaining
call sites touching sctx->stat.
Reported-by: Michel Palleau <michel.palleau@gmail.com>
Link: https://lore.kernel.org/linux-btrfs/CAMFk-+igFTv2E8svg=cQ6o3e6CrR5QwgQ3Ok9EyRaEvvthpqCQ@mail.gmail.com/
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
 
                stripe = &sctx->stripes[i];
 
                wait_scrub_stripe_io(stripe);
+               spin_lock(&sctx->stat_lock);
+               sctx->stat.last_physical = stripe->physical + stripe_length(stripe);
+               spin_unlock(&sctx->stat_lock);
                scrub_reset_stripe(stripe);
        }
 out:
                                         cur_physical, &found_logical);
                if (ret > 0) {
                        /* No more extent, just update the accounting */
+                       spin_lock(&sctx->stat_lock);
                        sctx->stat.last_physical = physical + logical_length;
+                       spin_unlock(&sctx->stat_lock);
                        ret = 0;
                        break;
                }
                        stripe_logical += chunk_logical;
                        ret = scrub_raid56_parity_stripe(sctx, scrub_dev, bg,
                                                         map, stripe_logical);
+                       spin_lock(&sctx->stat_lock);
+                       sctx->stat.last_physical = min(physical + BTRFS_STRIPE_LEN,
+                                                      physical_end);
+                       spin_unlock(&sctx->stat_lock);
                        if (ret)
                                goto out;
                        goto next;