ocfs2: always unmap in ocfs2_data_convert_worker()
Mmap-heavy clustered workloads were sometimes finding stale data on mmap
reads. The solution is to call unmap_mapping_range() on any down convert of
a data lock.
Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com>
diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c
index e622013..e335541 100644
--- a/fs/ocfs2/dlmglue.c
+++ b/fs/ocfs2/dlmglue.c
@@ -2718,6 +2718,15 @@
inode = ocfs2_lock_res_inode(lockres);
mapping = inode->i_mapping;
+ /*
+ * We need this before the filemap_fdatawrite() so that it can
+ * transfer the dirty bit from the PTE to the
+ * page. Unfortunately this means that even for EX->PR
+ * downconverts, we'll lose our mappings and have to build
+ * them up again.
+ */
+ unmap_mapping_range(mapping, 0, 0, 0);
+
if (filemap_fdatawrite(mapping)) {
mlog(ML_ERROR, "Could not sync inode %llu for downconvert!",
(unsigned long long)OCFS2_I(inode)->ip_blkno);
@@ -2725,7 +2734,6 @@
sync_mapping_buffers(mapping);
if (blocking == LKM_EXMODE) {
truncate_inode_pages(mapping, 0);
- unmap_mapping_range(mapping, 0, 0, 0);
} else {
/* We only need to wait on the I/O if we're not also
* truncating pages because truncate_inode_pages waits
