On 16.12.2024 18:34:28, Kees Cook wrote: > The proto_ops::getname callback was long ago backed by sockaddr_storage, > but the replacement of it for sockaddr was never done. Plumb it through > all the getname() callbacks, adjust prototypes, and fix casts. > > There are a few cases where the backing object is _not_ a sockaddr_storage > and converting it looks painful. In those cases, they use a cast to > struct sockaddr_storage. They appear well bounds-checked, so the risk > is no worse that we have currently. > > Other casts to sockaddr are removed, though to avoid spilling this > change into BPF (which becomes a much larger set of changes), cast the > sockaddr_storage instances there to sockaddr for the time being. > > In theory this could be split up into per-caller patches that add more > casts that all later get removed, but it seemed like there are few > enough callers that it seems feasible to do this in a single patch. Most > conversions are mechanical, so review should be fairly easy. (Famous > last words.) > > Signed-off-by: Kees Cook <kees@xxxxxxxxxx> > --- > net/can/isotp.c | 3 +- > net/can/j1939/socket.c | 2 +- > net/can/raw.c | 2 +- Acked-by: Marc Kleine-Budde <mkl@xxxxxxxxxxxxxx> # for net/can regards, Marc -- Pengutronix e.K. | Marc Kleine-Budde | Embedded Linux | https://www.pengutronix.de | Vertretung Nürnberg | Phone: +49-5121-206917-129 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-9 |
Attachment:
signature.asc
Description: PGP signature
