Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1314382

[PATCH v2 2/5] dax: clear TOWRITE flag after flush is complete

Show all headers | View raw

Previously in dax_writeback_one() we cleared the PAGECACHE_TAG_TOWRITE flag
before we had actually flushed the tagged radix tree entry to media.  This
is incorrect because of the following race:

Thread 1				Thread 2
--------				--------
dax_writeback_mapping_range()
tag entry with PAGECACHE_TAG_TOWRITE
					dax_writeback_mapping_range()
					tag entry with PAGECACHE_TAG_TOWRITE
					dax_writeback_one()
					radix_tree_tag_clear(TOWRITE)
TOWRITE flag is no longer set,
  find_get_entries_tag() finds no
  entries, return
					flush entry to media

In this case thread 1 returns before the data for the dirty entry is
actually durable on media.

Fix this by only clearing the PAGECACHE_TAG_TOWRITE flag after all flushing
is complete.

Signed-off-by: Ross Zwisler <ross.zwisler@linux.intel.com>
Reported-by: Jan Kara <jack@suse.cz>
---
 fs/dax.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/fs/dax.c b/fs/dax.c
index cee9e1b..d589113 100644
--- a/fs/dax.c
+++ b/fs/dax.c
@@ -407,8 +407,6 @@ static int dax_writeback_one(struct block_device *bdev,
 	if (!radix_tree_tag_get(page_tree, index, PAGECACHE_TAG_TOWRITE))
 		goto unlock;
 
-	radix_tree_tag_clear(page_tree, index, PAGECACHE_TAG_TOWRITE);
-
 	if (WARN_ON_ONCE(type != RADIX_DAX_PTE && type != RADIX_DAX_PMD)) {
 		ret = -EIO;
 		goto unlock;
@@ -432,6 +430,10 @@ static int dax_writeback_one(struct block_device *bdev,
 	}
 
 	wb_cache_pmem(dax.addr, dax.size);
+
+	spin_lock_irq(&mapping->tree_lock);
+	radix_tree_tag_clear(page_tree, index, PAGECACHE_TAG_TOWRITE);
+	spin_unlock_irq(&mapping->tree_lock);
  unmap:
 	dax_unmap_atomic(bdev, &dax);
 	return ret;
-- 
2.5.0

Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

[PATCH v2 0/5] DAX fsync/msync fixes Ross Zwisler <ross.zwisler@linux.intel.com> - 2016-01-21 18:50 +0100
  [PATCH v2 4/5] dax: fix PMD handling for fsync/msync Ross Zwisler <ross.zwisler@linux.intel.com> - 2016-01-21 18:50 +0100
    Re: [PATCH v2 4/5] dax: fix PMD handling for fsync/msync Jan Kara <jack@suse.cz> - 2016-01-22 16:20 +0100
      Re: [PATCH v2 4/5] dax: fix PMD handling for fsync/msync Ross Zwisler <ross.zwisler@linux.intel.com> - 2016-01-22 17:10 +0100
  [PATCH v2 3/5] dax: improve documentation for fsync/msync Ross Zwisler <ross.zwisler@linux.intel.com> - 2016-01-21 18:50 +0100
    Re: [PATCH v2 3/5] dax: improve documentation for fsync/msync Jan Kara <jack@suse.cz> - 2016-01-22 16:10 +0100
      Re: [PATCH v2 3/5] dax: improve documentation for fsync/msync Ross Zwisler <ross.zwisler@linux.intel.com> - 2016-01-22 17:00 +0100
        RE: [PATCH v2 3/5] dax: improve documentation for fsync/msync "Elliott, Robert (Persistent Memory)" <elliott@hpe.com> - 2016-01-22 17:20 +0100
  [PATCH v2 2/5] dax: clear TOWRITE flag after flush is complete Ross Zwisler <ross.zwisler@linux.intel.com> - 2016-01-21 18:50 +0100
    Re: [PATCH v2 2/5] dax: clear TOWRITE flag after flush is complete Jan Kara <jack@suse.cz> - 2016-01-22 16:00 +0100
  [PATCH v2 1/5] dax: never rely on bh.b_dev being set by get_block() Ross Zwisler <ross.zwisler@linux.intel.com> - 2016-01-21 18:50 +0100
    Re: [PATCH v2 1/5] dax: never rely on bh.b_dev being set by  get_block() Jan Kara <jack@suse.cz> - 2016-01-22 16:00 +0100
  [PATCH v2 5/5] dax: fix clearing of holes in __dax_pmd_fault() Ross Zwisler <ross.zwisler@linux.intel.com> - 2016-01-21 18:50 +0100
    Re: [PATCH v2 5/5] dax: fix clearing of holes in __dax_pmd_fault() Jan Kara <jack@suse.cz> - 2016-01-22 16:40 +0100
      Re: [PATCH v2 5/5] dax: fix clearing of holes in __dax_pmd_fault() Ross Zwisler <ross.zwisler@linux.intel.com> - 2016-01-22 17:20 +0100
        Re: [PATCH v2 5/5] dax: fix clearing of holes in __dax_pmd_fault() Jan Kara <jack@suse.cz> - 2016-01-25 15:50 +0100

csiph-web