Any thoughts on this?
On Mon, Jun 06, 2016 at 01:36:05PM -0400, David Hill wrote:
> Hello -
>
> Clang reports a possible user of an uninitalized variable in
> nfs_vnops.c line 2605.
>
> attrflag is uninitialized when calling nfsm_postop_attr(), which is a
> macro that only sets attrflag if (info.nmi_mrep != NULL). I am not sure
> if that is possible, but maybe a KASSERT(info.nmi_mrep != NULL) before
> nfsm_postop_attr() would catch a bug...
>
> Or just silence it with:
>
> Index: nfs_vnops.c
> ===================================================================
> RCS file: /cvs/src/sys/nfs/nfs_vnops.c,v
> retrieving revision 1.169
> diff -u -p -r1.169 nfs_vnops.c
> --- nfs_vnops.c 29 Apr 2016 14:40:36 -0000 1.169
> +++ nfs_vnops.c 6 Jun 2016 17:34:24 -0000
> @@ -2562,7 +2562,7 @@ nfs_lookitup(struct vnode *dvp, char *na
> struct vnode *newvp = NULL;
> struct nfsnode *np, *dnp = VTONFS(dvp);
> caddr_t cp2;
> - int error = 0, fhlen, attrflag;
> + int error = 0, fhlen, attrflag = 0;
> nfsfh_t *nfhp;
>
> info.nmi_v3 = NFS_ISV3(dvp);
>
