GFS2: Add Orlov allocator
Just like ext3, this works on the root directory and any directory
with the +T flag set. Also, just like ext3, any subdirectory created
in one of the just mentioned cases will be allocated to a random
resource group (GFS2 equivalent of a block group).
If you are creating a set of directories, each of which will contain a
job running on a different node, then by setting +T on the parent
directory before creating the subdirectories, each will land up in a
different resource group, and thus resource group contention between
nodes will be kept to a minimum.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
index bdf3e64..99a6197 100644
--- a/fs/gfs2/rgrp.c
+++ b/fs/gfs2/rgrp.c
@@ -16,6 +16,7 @@
#include <linux/prefetch.h>
#include <linux/blkdev.h>
#include <linux/rbtree.h>
+#include <linux/random.h>
#include "gfs2.h"
#include "incore.h"
@@ -1763,6 +1764,15 @@
return tdiff > (msecs * 1000 * 1000);
}
+static u32 gfs2_orlov_skip(const struct gfs2_inode *ip)
+{
+ const struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
+ u32 skip;
+
+ get_random_bytes(&skip, sizeof(skip));
+ return skip % sdp->sd_rgrps;
+}
+
static bool gfs2_select_rgrp(struct gfs2_rgrpd **pos, const struct gfs2_rgrpd *begin)
{
struct gfs2_rgrpd *rgd = *pos;
@@ -1784,7 +1794,7 @@
* Returns: errno
*/
-int gfs2_inplace_reserve(struct gfs2_inode *ip, u32 requested)
+int gfs2_inplace_reserve(struct gfs2_inode *ip, u32 requested, u32 aflags)
{
struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
struct gfs2_rgrpd *begin = NULL;
@@ -1792,6 +1802,7 @@
int error = 0, rg_locked, flags = 0;
u64 last_unlinked = NO_BLOCK;
int loops = 0;
+ u32 skip = 0;
if (sdp->sd_args.ar_rgrplvb)
flags |= GL_SKIP;
@@ -1805,6 +1816,8 @@
} else {
rs->rs_rbm.rgd = begin = gfs2_blk2rgrpd(sdp, ip->i_goal, 1);
}
+ if (S_ISDIR(ip->i_inode.i_mode) && (aflags & GFS2_AF_ORLOV))
+ skip = gfs2_orlov_skip(ip);
if (rs->rs_rbm.rgd == NULL)
return -EBADSLT;
@@ -1813,6 +1826,8 @@
if (!gfs2_glock_is_locked_by_me(rs->rs_rbm.rgd->rd_gl)) {
rg_locked = 0;
+ if (skip && skip--)
+ goto next_rgrp;
if (!gfs2_rs_active(rs) && (loops < 2) &&
gfs2_rgrp_used_recently(rs, 1000) &&
gfs2_rgrp_congested(rs->rs_rbm.rgd, loops))
@@ -1871,6 +1886,8 @@
/* Find the next rgrp, and continue looking */
if (gfs2_select_rgrp(&rs->rs_rbm.rgd, begin))
continue;
+ if (skip)
+ continue;
/* If we've scanned all the rgrps, but found no free blocks
* then this checks for some less likely conditions before