On Mon, 10 Oct 2016 19:54:01 +0800 Li Qiang <[email protected]> wrote:
> I think the 'xattr_fidp->fs.xattr.name' will not leak as the > 'v9fs_string_copy' will first free the fs.xattr.name. > Indeed! > Thanks. > Cheers. -- Greg > 2016-10-10 19:28 GMT+08:00 Greg Kurz <[email protected]>: > > > On Mon, 10 Oct 2016 13:20:51 +0200 > > Greg Kurz <[email protected]> wrote: > > > > > On Mon, 10 Oct 2016 03:41:38 -0700 > > > Li Qiang <[email protected]> wrote: > > > > > > > From: Li Qiang <[email protected]> > > > > > > > > The 'fs.xattr.value' field in V9fsFidState object doesn't consider the > > > > situation that this field has been allocated previously. Every time, it > > > > will be allocated directly. This leads a host memory leak issue. This > > > > patch fix this. > > > > > > > > Signed-off-by: Li Qiang <[email protected]> > > > > --- > > > > > > I'll add to the changelog that this may happen if the client sends a > > > Txattrcreate message with the same fid number before the fid was > > > clunked. > > > > > > Reviewed-by: Greg Kurz <[email protected]> > > > > > > > Oops I may have answered to fast... what about xattr_fidp->fs.xattr.name ? > > It looks like it is leaked the same way... > > > > > > hw/9pfs/9p.c | 1 + > > > > 1 file changed, 1 insertion(+) > > > > > > > > diff --git a/hw/9pfs/9p.c b/hw/9pfs/9p.c > > > > index 8751c19..e4040dc 100644 > > > > --- a/hw/9pfs/9p.c > > > > +++ b/hw/9pfs/9p.c > > > > @@ -3282,6 +3282,7 @@ static void v9fs_xattrcreate(void *opaque) > > > > xattr_fidp->fs.xattr.flags = flags; > > > > v9fs_string_init(&xattr_fidp->fs.xattr.name); > > > > v9fs_string_copy(&xattr_fidp->fs.xattr.name, &name); > > > > + g_free(xattr_fidp->fs.xattr.value); > > > > xattr_fidp->fs.xattr.value = g_malloc0(size); > > > > err = offset; > > > > put_fid(pdu, file_fidp); > > > > > > >
