Add a sysctl vfs.read_min to complement the exiting vfs.read_max. It

defaults to 1, meaning that it's off.

When read-ahead is enabled on a file, the vfs cluster code deliberately
breaks a read into 2 I/O transactions; one to satisfy the actual read,
and one to perform read-ahead.  This makes sense in low-latency
circumstances, but often produces unbalanced i/o transactions that
penalize disks.  By setting vfs.read_min, we can tell the algorithm to
fetch a larger transaction that what we asked for, achieving the same
effect as the read-ahead but without the doubled, unbalanced transaction
and the slightly lower latency.  This significantly helps our workloads
with video streaming.

Submitted by:	emax
Reviewed by:	kib
Obtained from:	Netflix

1 files changed, 12 insertions, 0 deletions

diff --git a/sys/kern/vfs_cluster.c b/sys/kern/vfs_cluster.c
index 91a0443..d619960 100644
--- a/sys/kern/vfs_cluster.c
+++ b/sys/kern/vfs_cluster.c
@@ -76,6 +76,10 @@ static int read_max = 64;
 SYSCTL_INT(_vfs, OID_AUTO, read_max, CTLFLAG_RW, &read_max, 0,
     "Cluster read-ahead max block count");
 
+static int read_min = 1;
+SYSCTL_INT(_vfs, OID_AUTO, read_min, CTLFLAG_RW, &read_min, 0,
+    "Cluster read min block count");
+
 /* Page expended to mark partially backed buffers */
 extern vm_page_t	bogus_page;
 
@@ -166,6 +170,7 @@ cluster_read(struct vnode *vp, u_quad_t filesize, daddr_t lblkno, long size,
 	} else {
 		off_t firstread = bp->b_offset;
 		int nblks;
+		long minread;
 
 		KASSERT(bp->b_offset != NOOFFSET,
 		    ("cluster_read: no buffer offset"));
@@ -173,6 +178,13 @@ cluster_read(struct vnode *vp, u_quad_t filesize, daddr_t lblkno, long size,
 		ncontig = 0;
 
 		/*
+		 * Adjust totread if needed
+		 */
+		minread = read_min * size;
+		if (minread > totread)
+			totread = minread;
+
+		/*
 		 * Compute the total number of blocks that we should read
 		 * synchronously.
 		 */