On Mon, Dec 14, 2009 at 02:35:25PM +0100, Jim Meyering wrote: > I didn't try to trigger the double free, but it looks feasible. > > >From 864cc161ffaf6c08cb980e66a673b53ac5ea52ab Mon Sep 17 00:00:00 2001 > From: Jim Meyering <meyering(a)redhat.com&gt; > Date: Mon, 14 Dec 2009 14:34:29 +0100 > Subject: [PATCH] virsh: avoid double-free > > * tools/virsh.c (vshCommandParse): Avoid double-free of "tkdata". > Set it to NULL immediately after free in the (cmd == NULL) case, > just as in the other case, in case the final free(tkdata) is > triggered by a syntax error. > --- > tools/virsh.c | 1 + > 1 files changed, 1 insertions(+), 0 deletions(-) > > diff --git a/tools/virsh.c b/tools/virsh.c > index 62c1270..0312945 100644 > --- a/tools/virsh.c > +++ b/tools/virsh.c > @@ -8044,6 +8044,7 @@ vshCommandParse(vshControl *ctl, char *cmdstr) > goto syntaxError; /* ... or ignore this command only? */ > } > free(tkdata); > + tkdata = NULL; > } else if (tk == VSH_TK_OPTION) { > if (!(opt = vshCmddefGetOption(cmd, tkdata))) { > vshError(ctl, ACK, that's the kind of things VIR_FREE() solved, but unfortunately it's not available for virsh.c ...