[Cluster-devel] [PATCH 02/20] gfs2: Don't ignore inode write errors during inode_go_sync

[Andreas Gruenbacher]

From: Bob Peterson <rpeterso at redhat.com>

Before for this patch, function inode_go_sync ignored io errors
during inode_go_sync, overwriting them with metadata write errors:

		error = filemap_fdatawait(mapping);
		mapping_set_error(mapping, error);
	}
	error = filemap_fdatawait(metamapping);
	...
	return error;

So any errors returned by the inode write would be forgotten if the
metadata write succeeded. This patch still does both writes, but
only sets error if it's still zero. That way, any errors will be
reported by to the caller, do_xmote, which will take appropriate
action and report the error.

Signed-off-by: Bob Peterson <rpeterso at redhat.com>
Signed-off-by: Andreas Gruenbacher <agruenba at redhat.com>
---
 fs/gfs2/glops.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c
index 9e9c7a4b8c66..4862dae868a2 100644
--- a/fs/gfs2/glops.c
+++ b/fs/gfs2/glops.c
@@ -268,7 +268,7 @@ static int inode_go_sync(struct gfs2_glock *gl)
 	struct gfs2_inode *ip = gfs2_glock2inode(gl);
 	int isreg = ip && S_ISREG(ip->i_inode.i_mode);
 	struct address_space *metamapping = gfs2_glock2aspace(gl);
-	int error = 0;
+	int error = 0, ret;
 
 	if (isreg) {
 		if (test_and_clear_bit(GIF_SW_PAGED, &ip->i_flags))
@@ -289,8 +289,10 @@ static int inode_go_sync(struct gfs2_glock *gl)
 		error = filemap_fdatawait(mapping);
 		mapping_set_error(mapping, error);
 	}
-	error = filemap_fdatawait(metamapping);
-	mapping_set_error(metamapping, error);
+	ret = filemap_fdatawait(metamapping);
+	mapping_set_error(metamapping, ret);
+	if (!error)
+		error = ret;
 	gfs2_ail_empty_gl(gl);
 	/*
 	 * Writeback of the data mapping may cause the dirty flag to be set
-- 
2.26.2