On Mon, Oct 23, 2023 at 12:40 PM Mathieu Poirier
<mathieu.poirier@xxxxxxxxxx> wrote:
>
> On Mon, Oct 23, 2023 at 06:12:28PM +0000, Justin Stitt wrote:
> > strncpy() is deprecated for use on NUL-terminated destination strings
> > [1] and as such we should prefer more robust and less ambiguous string
> > interfaces.
> >
> > This patch replaces 3 callsites of strncpy().
> >
> > The first two populate the destination buffer `nsm.name` -- which we
> > expect to be NUL-terminated based on their use with format strings.
> >
> > Firstly, as I understand it, virtio_rpmsg_announce_create() creates an
> > rpmsg_ns_msg and sends via:
> >
> > virtio_rpmsg_bus.c:
> > 336: err = rpmsg_sendto(rpdev->ept, &nsm, sizeof(nsm), RPMSG_NS_ADDR);
> >
> > ... which uses:
> > virtio_rpmsg_sendto() -> rpmsg_send_offchannel_raw()
> >
> > ... which copies its data into an rpmsg_hdr `msg` in virtio_rpmsg_bus.c
> > 618: memcpy(msg->data, data, len);
> >
> > This callback is invoked when a message is received from the remote
> > processor:
> >
> > rpmsg_ns.c:
> > 30: /* invoked when a name service announcement arrives */
> > 31: static int rpmsg_ns_cb(struct rpmsg_device *rpdev, void *data, int len,
> > 32: void *priv, u32 src)
> > 33: {
> > 34: struct rpmsg_ns_msg *msg = data;
> > ...
> > 50: /* don't trust the remote processor for null terminating the name */
> > 51: msg->name[RPMSG_NAME_SIZE - 1] = '\0';
> >
> > ... which leads into the use of `name` within a format string:
> > rpmsg_ns.c:
> > 57: dev_info(dev, "%sing channel %s addr 0x%x\n",
> > 58: rpmsg32_to_cpu(rpdev, msg->flags) & RPMSG_NS_DESTROY ?
> > 59: "destroy" : "creat", msg->name, chinfo.dst);
> >
> > We can also observe that `nsm` is not zero-initialized and as such we
> > should maintain the NUL-padding behavior that strncpy() provides:
> >
> > virtio_rpmsg_bus.c:
> > 330: struct rpmsg_ns_msg nsm;
> >
> > Considering the above, a suitable replacement is `strscpy_pad` due to
> > the fact that it guarantees both NUL-termination and NUL-padding on the
> > destination buffer.
> >
> > Now, for the third and final destination buffer rpdev->id.name we can
> > just go for strscpy() (not _pad()) as rpdev points to &vch->rpdev:
> > | rpdev = &vch->rpdev;
> >
> > ... and vch is zero-allocated:
> > | vch = kzalloc(sizeof(*vch), GFP_KERNEL);
> >
> > ... this renders any additional NUL-byte assignments (like the ones
> > strncpy() or strscpy_pad() does) redundant.
> >
> > Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#strncpy-on-nul-terminated-strings [1]
> > Link: https://github.com/KSPP/linux/issues/90
> > Cc: linux-hardening@xxxxxxxxxxxxxxx
> > Signed-off-by: Justin Stitt <justinstitt@xxxxxxxxxx>
> > ---
> > Changes in v2:
> > - update commit msg (thanks Mathieu)
> > - Link to v1: https://lore.kernel.org/r/20231021-strncpy-drivers-rpmsg-virtio_rpmsg_bus-c-v1-1-8abb919cbe24@xxxxxxxxxx
> > ---
> > Note: build-tested only.
> >
> > Found with: $ rg "strncpy\("
> > ---
> > drivers/rpmsg/virtio_rpmsg_bus.c | 6 +++---
> > 1 file changed, 3 insertions(+), 3 deletions(-)
> >
>
> I have applied all 3 of your patches.
Cheers!
>
> Thanks,
> Mathieu
>
> > diff --git a/drivers/rpmsg/virtio_rpmsg_bus.c b/drivers/rpmsg/virtio_rpmsg_bus.c
> > index 905ac7910c98..dc87965f8164 100644
> > --- a/drivers/rpmsg/virtio_rpmsg_bus.c
> > +++ b/drivers/rpmsg/virtio_rpmsg_bus.c
> > @@ -329,7 +329,7 @@ static int virtio_rpmsg_announce_create(struct rpmsg_device *rpdev)
> > virtio_has_feature(vrp->vdev, VIRTIO_RPMSG_F_NS)) {
> > struct rpmsg_ns_msg nsm;
> >
> > - strncpy(nsm.name, rpdev->id.name, RPMSG_NAME_SIZE);
> > + strscpy_pad(nsm.name, rpdev->id.name, sizeof(nsm.name));
> > nsm.addr = cpu_to_rpmsg32(rpdev, rpdev->ept->addr);
> > nsm.flags = cpu_to_rpmsg32(rpdev, RPMSG_NS_CREATE);
> >
> > @@ -353,7 +353,7 @@ static int virtio_rpmsg_announce_destroy(struct rpmsg_device *rpdev)
> > virtio_has_feature(vrp->vdev, VIRTIO_RPMSG_F_NS)) {
> > struct rpmsg_ns_msg nsm;
> >
> > - strncpy(nsm.name, rpdev->id.name, RPMSG_NAME_SIZE);
> > + strscpy_pad(nsm.name, rpdev->id.name, sizeof(nsm.name));
> > nsm.addr = cpu_to_rpmsg32(rpdev, rpdev->ept->addr);
> > nsm.flags = cpu_to_rpmsg32(rpdev, RPMSG_NS_DESTROY);
> >
> > @@ -424,7 +424,7 @@ static struct rpmsg_device *__rpmsg_create_channel(struct virtproc_info *vrp,
> > */
> > rpdev->announce = rpdev->src != RPMSG_ADDR_ANY;
> >
> > - strncpy(rpdev->id.name, chinfo->name, RPMSG_NAME_SIZE);
> > + strscpy(rpdev->id.name, chinfo->name, sizeof(rpdev->id.name));
> >
> > rpdev->dev.parent = &vrp->vdev->dev;
> > rpdev->dev.release = virtio_rpmsg_release_device;
> >
> > ---
> > base-commit: 9c5d00cb7b6bbc5a7965d9ab7d223b5402d1f02c
> > change-id: 20231020-strncpy-drivers-rpmsg-virtio_rpmsg_bus-c-dba15db4e890
> >
> > Best regards,
> > --
> > Justin Stitt <justinstitt@xxxxxxxxxx>
> >
