d/freebsd-skq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

- Extend the vnode_free_list_mtx to cover numvnodes and freevnodes. This

Browse Source 
was done only some of the time before, and now it is uniformly applied.
This commit is contained in:
jeff 2002-08-13 05:29:48 +00:00
parent efe8049646
commit d18378e088
1 changed files with 15 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
sys/kern/vfs_subr.c
View File

@ -144,7 +144,12 @@ int nfs_mount_type = -1;
/* To keep more than one thread at a time from running vfs_getnewfsid */
static struct mtx mntid_mtx;
/* For any iteration/modification of vnode_free_list */
/*
* Lock for any access to the following:
* vnode_free_list
* numvnodes
* freevnodes
*/
static struct mtx vnode_free_list_mtx;
/*
@ -616,11 +621,14 @@ vnlru_proc(void)
s = splbio();
for (;;) {
kthread_suspend_check(p);
mtx_lock(&vnode_free_list_mtx);
if (numvnodes - freevnodes <= desiredvnodes * 9 / 10) {
mtx_unlock(&vnode_free_list_mtx);
vnlruproc_sig = 0;
tsleep(vnlruproc, PVFS, "vlruwt", 0);
continue;
}
mtx_unlock(&vnode_free_list_mtx);
done = 0;
mtx_lock(&mountlist_mtx);
for (mp = TAILQ_FIRST(&mountlist); mp != NULL; mp = nmp) {
@ -741,6 +749,8 @@ getnewvnode(tag, mp, vops, vpp)
struct mount *vnmp;
s = splbio();
mtx_lock(&vnode_free_list_mtx);
/*
* Try to reuse vnodes if we hit the max. This situation only
* occurs in certain large-memory (2G+) situations. We cannot
@ -758,8 +768,6 @@ getnewvnode(tag, mp, vops, vpp)
* good minimum for good LRU performance.
*/
mtx_lock(&vnode_free_list_mtx);
if (freevnodes >= wantfreevnodes && numvnodes >= minvnodes) {
int error;
int count;
@ -870,6 +878,7 @@ getnewvnode(tag, mp, vops, vpp)
splx(s);
#if 0
mp_fixme("This code does not lock access to numvnodes && freevnodes.");
vnodeallocs++;
if (vnodeallocs % vnoderecycleperiod == 0 &&
freevnodes < vnoderecycleminfreevn &&
@ -2761,6 +2770,9 @@ sysctl_vnode(SYSCTL_HANDLER_ARGS)
struct vnode *vp;
int error, len, n;
/*
* Stale numvnodes access is not fatal here.
*/
req->lock = 0;
len = (numvnodes + KINFO_VNODESLOP) * sizeof *xvn;
if (!req->oldptr)