aboutsummaryrefslogtreecommitdiffstats
path: root/fs/ubifs/gc.c
diff options
context:
space:
mode:
authorArtem Bityutskiy <Artem.Bityutskiy@nokia.com>2009-01-19 12:57:27 -0500
committerArtem Bityutskiy <Artem.Bityutskiy@nokia.com>2009-01-20 03:10:47 -0500
commit7078202e55b565582fcbd831a8dd3069bdc72610 (patch)
tree0f8cabaa23a05210cc1b95a97d7cab490f38ebca /fs/ubifs/gc.c
parenta50412e3f8ce95d7ed558370d7dde5171fd04283 (diff)
UBIFS: document dark_wm and dead_wm better
Just add more commentaries. Also some commentary fixes for lprops flags. Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Diffstat (limited to 'fs/ubifs/gc.c')
-rw-r--r--fs/ubifs/gc.c20
1 files changed, 20 insertions, 0 deletions
diff --git a/fs/ubifs/gc.c b/fs/ubifs/gc.c
index 9832f9abe28e..b2e5f1133377 100644
--- a/fs/ubifs/gc.c
+++ b/fs/ubifs/gc.c
@@ -31,6 +31,26 @@
31 * to be reused. Garbage collection will cause the number of dirty index nodes 31 * to be reused. Garbage collection will cause the number of dirty index nodes
32 * to grow, however sufficient space is reserved for the index to ensure the 32 * to grow, however sufficient space is reserved for the index to ensure the
33 * commit will never run out of space. 33 * commit will never run out of space.
34 *
35 * Notes about dead watermark. At current UBIFS implementation we assume that
36 * LEBs which have less than @c->dead_wm bytes of free + dirty space are full
37 * and not worth garbage-collecting. The dead watermark is one min. I/O unit
38 * size, or min. UBIFS node size, depending on what is greater. Indeed, UBIFS
39 * Garbage Collector has to synchronize the GC head's write buffer before
40 * returning, so this is about wasting one min. I/O unit. However, UBIFS GC can
41 * actually reclaim even very small pieces of dirty space by garbage collecting
42 * enough dirty LEBs, but we do not bother doing this at this implementation.
43 *
44 * Notes about dark watermark. The results of GC work depends on how big are
45 * the UBIFS nodes GC deals with. Large nodes make GC waste more space. Indeed,
46 * if GC move data from LEB A to LEB B and nodes in LEB A are large, GC would
47 * have to waste large pieces of free space at the end of LEB B, because nodes
48 * from LEB A would not fit. And the worst situation is when all nodes are of
49 * maximum size. So dark watermark is the amount of free + dirty space in LEB
50 * which are guaranteed to be reclaimable. If LEB has less space, the GC migh
51 * be unable to reclaim it. So, LEBs with free + dirty greater than dark
52 * watermark are "good" LEBs from GC's point of few. The other LEBs are not so
53 * good, and GC takes extra care when moving them.
34 */ 54 */
35 55
36#include <linux/pagemap.h> 56#include <linux/pagemap.h>