Răsfoiți Sursa

[JFFS2] Relax threshold for triggering GC due to dirty blocks.

Instead of matching resv_blocks_gcmerge, which is only about 3, instead
match resv_blocks_gctrigger, which includes a proportion of the total
device size.

These ought to become tunable from userspace, at some point.

Signed-off-by: David Woodhouse <dwmw2@infradead.org>
David Woodhouse 17 ani în urmă
părinte
comite
85becc535b
1 a modificat fișierele cu 1 adăugiri și 1 ștergeri
  1. 1 1
      fs/jffs2/build.c

+ 1 - 1
fs/jffs2/build.c

@@ -289,7 +289,7 @@ static void jffs2_calc_trigger_levels(struct jffs2_sb_info *c)
 	   trigger the GC thread even if we don't _need_ the space. When we
 	   trigger the GC thread even if we don't _need_ the space. When we
 	   can't mark nodes obsolete on the medium, the old dirty nodes cause
 	   can't mark nodes obsolete on the medium, the old dirty nodes cause
 	   performance problems because we have to inspect and discard them. */
 	   performance problems because we have to inspect and discard them. */
-	c->vdirty_blocks_gctrigger = c->resv_blocks_gcmerge;
+	c->vdirty_blocks_gctrigger = c->resv_blocks_gctrigger;
 	if (jffs2_can_mark_obsolete(c))
 	if (jffs2_can_mark_obsolete(c))
 		c->vdirty_blocks_gctrigger *= 10;
 		c->vdirty_blocks_gctrigger *= 10;