On Wed, Aug 09, 2023 at 11:54:48PM +0200, Jan Engelhardt wrote: > > On Wednesday 2023-08-09 23:40, Justin Stitt wrote: > >On Wed, Aug 9, 2023 at 1:19 PM Florian Westphal <fw@xxxxxxxxx> wrote: > >> > >> Justin Stitt <justinstitt@xxxxxxxxxx> wrote: > >> > Use `strscpy_pad` instead of `strncpy`. > >> > >> I don't think that any of these need zero-padding. > >It's a more consistent change with the rest of the series and I don't > >believe it has much different behavior to `strncpy` (other than > >NUL-termination) as that will continue to pad to `n` as well. > > > >Do you think the `_pad` for 1/7, 6/7 and 7/7 should be changed back to > >`strscpy` in a v3? I really am shooting in the dark as it is quite > >hard to tell whether or not a buffer is expected to be NUL-padded or > >not. > > I don't recall either NF userspace or kernelspace code doing memcmp > with name-like fields, so padding should not be strictly needed. My only concern with padding is just to make sure any buffers copied to userspace have been zeroed. I would need to take a close look at how buffers are passed around here to know for sure... -- Kees Cook
