kernel - NFS - fix additional problems with readdirplus

* Ok, give up trying to hack a fix for readdirplus.   Instead, do the
  fix the right by properly reordering namecache lookups and vnodes.

* Do not create a namecache entry for '.' or '..'.  These entries are
  superfluous (ignored by the lookup code).

diff --git a/sys/vfs/nfs/nfs_vnops.c b/sys/vfs/nfs/nfs_vnops.c
index ddbc2fc..0cf7dea 100644 (file)
--- a/sys/vfs/nfs/nfs_vnops.c
+++ b/sys/vfs/nfs/nfs_vnops.c
@@ -245,6 +245,22 @@ SYSCTL_INT(_vfs_nfs, OID_AUTO, access_cache_misses, CTLFLAG_RD,
 #define        NFSV3ACCESS_ALL (NFSV3ACCESS_READ | NFSV3ACCESS_MODIFY          \
                         | NFSV3ACCESS_EXTEND | NFSV3ACCESS_EXECUTE     \
                         | NFSV3ACCESS_DELETE | NFSV3ACCESS_LOOKUP)
+
+/*
+ * Returns whether a name component is a degenerate '.' or '..'.
+ */
+static __inline
+int
+nlcdegenerate(struct nlcomponent *nlc)
+{
+       if (nlc->nlc_namelen == 1 && nlc->nlc_nameptr[0] == '.')
+               return(1);
+       if (nlc->nlc_namelen == 2 &&
+           nlc->nlc_nameptr[0] == '.' && nlc->nlc_nameptr[1] == '.')
+               return(1);
+       return(0);
+}
+
 static int
 nfs3_access_otw(struct vnode *vp, int wmode,
                struct thread *td, struct ucred *cred)
@@ -2475,7 +2491,6 @@ int
 nfs_readdirplusrpc_uio(struct vnode *vp, struct uio *uiop)
 {
        int len, left;
-       int lkstatus;
        struct nfs_dirent *dp;
        u_int32_t *tl;
        struct vnode *newvp;
@@ -2522,10 +2537,14 @@ nfs_readdirplusrpc_uio(struct vnode *vp, struct uio *uiop)
         * If there is no cookie, assume directory was stale.
         */
        cookiep = nfs_getcookie(dnp, uiop->uio_offset, 0);
-       if (cookiep)
+       if (cookiep) {
                cookie = *cookiep;
-       else
+       } else {
+               if (dnch.ncp)
+                       cache_drop(&dnch);
                return (NFSERR_BAD_COOKIE);
+       }
+
        /*
         * Loop around doing readdir rpc's of size nm_readdirsize
         * truncated to a multiple of DIRBLKSIZ.
@@ -2608,8 +2627,9 @@ nfs_readdirplusrpc_uio(struct vnode *vp, struct uio *uiop)
                        if (bigenough) {
                                cookie.nfsuquad[0] = *tl++;
                                cookie.nfsuquad[1] = *tl++;
-                       } else
+                       } else {
                                tl += 2;
+                       }
 
                        /*
                         * Since the attributes are before the file handle
@@ -2625,58 +2645,38 @@ nfs_readdirplusrpc_uio(struct vnode *vp, struct uio *uiop)
                            doit = fxdr_unsigned(int, *tl);
                            if (doit) {
                                NEGATIVEOUT(fhsize = nfsm_getfh(&info, &fhp));
-                               if (NFS_CMPFH(dnp, fhp, fhsize)) {
-                                   vref(vp);
-                                   newvp = vp;
-                                   np = dnp;
-                               } else {
-                                   error = nfs_nget(vp->v_mount, fhp,
-                                                    fhsize, &np);
-                                   if (error)
-                                       doit = 0;
-                                   else
-                                       newvp = NFSTOV(np);
-                               }
                            }
-                           if (doit && bigenough) {
-                               dpossav2 = info.dpos;
-                               info.dpos = dpossav1;
-                               mdsav2 = info.md;
-                               info.md = mdsav1;
-                               ERROROUT(nfsm_loadattr(&info, newvp, NULL));
-                               info.dpos = dpossav2;
-                               info.md = mdsav2;
-                               dp->nfs_type =
-                                   IFTODT(VTTOIF(np->n_vattr.va_type));
+                           if (doit && bigenough && !nlcdegenerate(&nlc) &&
+                               !NFS_CMPFH(dnp, fhp, fhsize)
+                           ) {
                                if (dnch.ncp) {
 #if 0
                                    kprintf("NFS/READDIRPLUS, ENTER %*.*s\n",
                                        nlc.nlc_namelen, nlc.nlc_namelen,
                                        nlc.nlc_nameptr);
 #endif
-                                   /*
-                                    * Work around a vp/namecache deadlock.
-                                    * Namecache lookups should be run
-                                    * without any vp locks held.
-                                    */
-                                   lkstatus = vn_islocked(vp);
-                                   if (lkstatus == LK_EXCLUSIVE ||
-                                       lkstatus == LK_SHARED) {
-                                           vn_unlock(vp);
-                                   }
-                                   if (newvp != vp)
-                                           vn_unlock(newvp);
                                    nch = cache_nlookup(&dnch, &nlc);
                                    cache_setunresolved(&nch);
-                                   nfs_cache_setvp(&nch, newvp,
-                                                   nfspos_cache_timeout);
-                                   cache_put(&nch);
-                                   if (lkstatus == LK_EXCLUSIVE ||
-                                       lkstatus == LK_SHARED) {
-                                           vn_lock(vp, lkstatus | LK_RETRY);
+                                   error = nfs_nget(vp->v_mount, fhp,
+                                                    fhsize, &np);
+                                   if (error == 0) {
+                                       newvp = NFSTOV(np);
+                                       dpossav2 = info.dpos;
+                                       info.dpos = dpossav1;
+                                       mdsav2 = info.md;
+                                       info.md = mdsav1;
+                                       ERROROUT(nfsm_loadattr(&info, newvp,
+                                                              NULL));
+                                       info.dpos = dpossav2;
+                                       info.md = mdsav2;
+                                       dp->nfs_type =
+                                           IFTODT(VTTOIF(np->n_vattr.va_type));
+                                       nfs_cache_setvp(&nch, newvp,
+                                                       nfspos_cache_timeout);
+                                       vput(newvp);
+                                       newvp = NULLVP;
                                    }
-                                   if (newvp != vp)
-                                           vn_lock(newvp, LK_EXCLUSIVE | LK_RETRY);
+                                   cache_put(&nch);
                                } else {
                                    kprintf("Warning: NFS/rddirplus, "
                                            "UNABLE TO ENTER %*.*s\n",
@@ -2690,13 +2690,6 @@ nfs_readdirplusrpc_uio(struct vnode *vp, struct uio *uiop)
                            i = fxdr_unsigned(int, *tl);
                            ERROROUT(nfsm_adv(&info, nfsm_rndup(i)));
                        }
-                       if (newvp != NULLVP) {
-                           if (newvp == vp)
-                               vrele(newvp);
-                           else
-                               vput(newvp);
-                           newvp = NULLVP;
-                       }
                        NULLOUT(tl = nfsm_dissect(&info, NFSX_UNSIGNED));
                        more_dirs = fxdr_unsigned(int, *tl);
                }
@@ -2727,9 +2720,9 @@ nfs_readdirplusrpc_uio(struct vnode *vp, struct uio *uiop)
         * We are now either at the end of the directory or have filled the
         * block.
         */
-       if (bigenough)
+       if (bigenough) {
                dnp->n_direofoffset = uiop->uio_offset;
-       else {
+       } else {
                if (uiop->uio_resid > 0)
                        kprintf("EEK! readdirplusrpc resid > 0\n");
                cookiep = nfs_getcookie(dnp, uiop->uio_offset, 1);