On 09.03.2022 22:49:49, Vincent Mailhol wrote:
> > But rx_offload needs the mailbox_read function, even if it's a dummy,
> > because can_rx_offload_add_fifo() checks:
> >
> > if (!offload->mailbox_read)
> > return -EINVAL;
>
> I think that there should not be a dummy functions like this one.
>
> Either we agree that using can_rx_offload without implementing
> the mailbox_read() is OK and in that case, the can_rx_offload
> framework should be modified to allow mailbox_read() to be a NULL
> pointer.
>
> Either it is not the case and you use the more classic
> netif_rx().
>
> And I do not have the answer. I haven't studied can_rx_offload
> enough to be a judge here. Sorry.
>
> @Marc, any thoughts?
Use can_rx_offload_add_manual() instead.
> > > > +/* Send a can_frame to a TTY. */
> > > > +static netdev_tx_t elmcan_netdev_start_xmit(struct sk_buff *skb,
> > > > + struct net_device *dev)
> > > > +{
> > > > + struct elmcan *elm = netdev_priv(dev);
> > > > + struct can_frame *frame = (struct can_frame *)skb->data;
> > > > +
> > > > + if (skb->len != sizeof(struct can_frame))
> > > > + goto out;
> > >
> > > Isn’t this aleardy guaranteed by the upper layers?
> >
> > Copy-pasta from slcan.c - will look into it.
>
> Also give a look at can_dropped_invalid_skb().
ACK
Marc
--
Pengutronix e.K. | Marc Kleine-Budde |
Embedded Linux | https://www.pengutronix.de |
Vertretung West/Dortmund | Phone: +49-231-2826-924 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
Attachment:
signature.asc
Description: PGP signature
