Increased ratio of allowed vnodes on freelist to 1/4th of the total. This

is more representative of worst case situations of 4 files/directory. (If
that last sentence doesn't make any sense, I'm not surprised. It's rather
compilcated how this all fits together....).
This should fix a problem that Ed Hudson has been complaining about where
directories with lots of symlinks could cause excessive disk I/O.
This commit is contained in:
David Greenman 1995-05-12 04:24:53 +00:00
parent 8af2296bcb
commit 15f1b09648
2 changed files with 6 additions and 6 deletions

View File

@ -36,7 +36,7 @@
* SUCH DAMAGE.
*
* @(#)vfs_subr.c 8.13 (Berkeley) 4/18/94
* $Id: vfs_subr.c,v 1.27 1995/04/09 06:02:46 davidg Exp $
* $Id: vfs_subr.c,v 1.28 1995/04/16 11:33:33 davidg Exp $
*/
/*
@ -333,10 +333,10 @@ getnewvnode(tag, mp, vops, vpp)
* is a silly thing to do.
* 2. we havn't filled our pool yet
* We don't want to trash the incore (VM-)vnodecache.
* 3. if less that 1/16th of our vnodes are free.
* 3. if less that 1/4th of our vnodes are free.
* We don't want to trash the namei cache either.
*/
if (freevnodes < (numvnodes >> 4) ||
if (freevnodes < (numvnodes >> 2) ||
numvnodes < desiredvnodes ||
vp == NULL) {
vp = (struct vnode *) malloc((u_long) sizeof *vp,

View File

@ -36,7 +36,7 @@
* SUCH DAMAGE.
*
* @(#)vfs_subr.c 8.13 (Berkeley) 4/18/94
* $Id: vfs_subr.c,v 1.27 1995/04/09 06:02:46 davidg Exp $
* $Id: vfs_subr.c,v 1.28 1995/04/16 11:33:33 davidg Exp $
*/
/*
@ -333,10 +333,10 @@ getnewvnode(tag, mp, vops, vpp)
* is a silly thing to do.
* 2. we havn't filled our pool yet
* We don't want to trash the incore (VM-)vnodecache.
* 3. if less that 1/16th of our vnodes are free.
* 3. if less that 1/4th of our vnodes are free.
* We don't want to trash the namei cache either.
*/
if (freevnodes < (numvnodes >> 4) ||
if (freevnodes < (numvnodes >> 2) ||
numvnodes < desiredvnodes ||
vp == NULL) {
vp = (struct vnode *) malloc((u_long) sizeof *vp,