Change the code in the experimental nfs client to avoid flushing

writes upon close when a write delegation is held by the client.
This should be safe to do, now that nfsv4 Close operations are
delayed until ncl_inactive() is called for the vnode.

Approved by:	kib (mentor)
This commit is contained in:
rmacklem 2009-05-22 15:01:47 +00:00
parent f360591bdc
commit 7c92f3c6dc

View File

@ -687,12 +687,8 @@ nfs_close(struct vop_close_args *ap)
int cm = newnfs_commit_on_close ? 1 : 0;
error = ncl_flush(vp, MNT_WAIT, cred, ap->a_td, cm);
/* np->n_flag &= ~NMODIFIED; */
} else if (NFS_ISV4(vp)) {
int cm;
if (newnfs_commit_on_close != 0)
cm = 1;
else
cm = nfscl_mustflush(vp);
} else if (NFS_ISV4(vp) && nfscl_mustflush(vp)) {
int cm = newnfs_commit_on_close ? 1 : 0;
error = ncl_flush(vp, MNT_WAIT, cred, ap->a_td, cm);
/* as above w.r.t. races when clearing NMODIFIED */
/* np->n_flag &= ~NMODIFIED; */