[Cluster-devel] [GFS2 PATCH 3/3] gfs2: Use filemap_fdatawrite_range for gfs2_meta_sync

[Bob Peterson <rpeterso@redhat.com>]

Before this patch, function gfs2_meta_sync called filemap_fdatawrite to write
the address space for the metadata being synced. That's great for inodes, but
for resource groups, they all point to the same superblock-address space,
sdp->sd_aspace, with their own range. That meant every time an rgrp's metadata
was synced, it would write all of them instead of just the range.

This patch changes function gfs2_meta_sync so that instead it calls
filemap_fdatawrite_range and filemap_fdatawait_range with a range of addresses
appropriate to the object. For inodes, it specifies the whole range, as
per filemap_fdatawrite. For rgrps, it uses only the range required.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>

12345678901234567890123456789012345678901234567890123456789012345678901234567890
---
 fs/gfs2/lops.c | 14 +++++++++++---
 1 file changed, 11 insertions(+), 3 deletions(-)

diff --git a/fs/gfs2/lops.c b/fs/gfs2/lops.c
index ed69298dd824..94b0e89ee7b2 100644
--- a/fs/gfs2/lops.c
+++ b/fs/gfs2/lops.c
@@ -827,13 +827,21 @@ void gfs2_meta_sync(struct gfs2_glock *gl)
 {
 	struct address_space *mapping = gfs2_glock2aspace(gl);
 	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
+	loff_t start = 0;
+	loff_t end = LLONG_MAX;
 	int error;
 
-	if (mapping == NULL)
+	if (mapping == NULL) {
+		struct gfs2_rgrpd *rgd = gfs2_glock2rgrp(gl);
+		const unsigned bsize = sdp->sd_sb.sb_bsize;
+
 		mapping = &sdp->sd_aspace;
+		start = (rgd->rd_addr * bsize) & PAGE_MASK;
+		end = PAGE_ALIGN((rgd->rd_addr + rgd->rd_length) * bsize) - 1;
+	}
 
-	filemap_fdatawrite(mapping);
-	error = filemap_fdatawait(mapping);
+	filemap_fdatawrite_range(mapping, start, end);
+	error = filemap_fdatawait_range(mapping, start, end);
 
 	if (error)
 		gfs2_io_error(gl->gl_name.ln_sbd);
-- 
2.26.2