On Fri 03 Dec 08:37 PST 2021, Arnaud POULIQUEN wrote:
> On 12/3/21 3:17 AM, Bjorn Andersson wrote:
> > On Mon 08 Nov 08:19 CST 2021, Arnaud Pouliquen wrote:
[..]
> >> +static inline int rpmsg_chrdev_eptdev_create(struct rpmsg_device *rpdev, struct device *parent,
> >> + struct rpmsg_channel_info chinfo)
> >> +{
> >> + /* This shouldn't be possible */
> >
> > But isn't it very much possible that userspace invokes this function
> > through the ioctl that you move to the separate rpmsg_ctrl driver?
> >
> >> + WARN_ON(1);
> >
> > Which would mean that this will spam the kernel with stack dumps.
>
> Good catch, I will suppress the WARM_ON. I propose also to return -ENXIO
> instead of -EINVAL to be aligned with other functions in rpmsg.h
>
ENXIO sounds better than EINVAL, let's go with that.
Thanks,
Bjorn
