[PATCH 15/16] readahead: disable look-ahead for loopback file

[Wu Fengguang]

Loopback files normally contain filesystem, in which case there are already
proper look-aheads in the upper layer, more look-aheads on the loopback file
only ruin the read-ahead hit rate.

Signed-off-by: Wu Fengguang <wfg@xxxxxxxxxxxxxxxx>
---

I'd like to thank Tero Grundstr?m for uncovering the loopback problem.

 drivers/block/loop.c |    6 ++++++
 include/linux/fs.h   |    1 +
 mm/readahead.c       |    8 ++++++++
 3 files changed, 15 insertions(+)

--- linux-2.6.14-mm1.orig/include/linux/fs.h
+++ linux-2.6.14-mm1/include/linux/fs.h
@@ -585,6 +585,7 @@ struct file_ra_state {
 };
 #define RA_FLAG_MISS 0x01	/* a cache miss occured against this file */
 #define RA_FLAG_INCACHE 0x02	/* file is already in cache */
+#define RA_FLAG_NO_LOOKAHEAD	(1<<31)	/* disable look-ahead */
 
 struct file {
 	/*
--- linux-2.6.14-mm1.orig/drivers/block/loop.c
+++ linux-2.6.14-mm1/drivers/block/loop.c
@@ -769,6 +769,12 @@ static int loop_set_fd(struct loop_devic
 	mapping = file->f_mapping;
 	inode = mapping->host;
 
+	/*
+	 * The upper layer should already do proper look-ahead,
+	 * one more look-ahead here only ruins the cache hit rate.
+	 */
+	file->f_ra.flags |= RA_FLAG_NO_LOOKAHEAD;
+
 	if (!(file->f_mode & FMODE_WRITE))
 		lo_flags |= LO_FLAGS_READ_ONLY;
 
--- linux-2.6.14-mm1.orig/mm/readahead.c
+++ linux-2.6.14-mm1/mm/readahead.c
@@ -1272,6 +1272,11 @@ static inline void ra_state_update(struc
 	if (ra_size < old_ra && ra_cache_hit(ra, 0))
 		ra_account(ra, RA_EVENT_READAHEAD_SHRINK, old_ra - ra_size);
 #endif
+
+	/* Disable look-ahead for loopback file. */
+	if (unlikely(ra->flags & RA_FLAG_NO_LOOKAHEAD))
+		la_size = 0;
+
 	ra_addup_cache_hit(ra);
 	ra->ra_index = ra->readahead_index;
 	ra->la_index = ra->lookahead_index;
@@ -1628,6 +1633,9 @@ static int query_page_cache(struct addre
 	if (count < ra_max)
 		goto out;
 
+	if (unlikely(ra->flags & RA_FLAG_NO_LOOKAHEAD))
+		goto out;
+
 	/*
 	 * Check the far pages coarsely.
 	 * The big count here helps increase la_size.

--
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/