On Thu, Sep 16, 2021 at 11:31 AM zhiyong.tao <zhiyong.tao@xxxxxxxxxxxx> wrote: > On Mon, 2021-09-06 at 18:09 +0800, Chen-Yu Tsai wrote: > > > > > + if (!found) { > > > > > + dev_err(hw->dev, "Not support rsel value %d Ohm > > > > > for > > > > > pin = %d (%s)\n", > > > > > + arg, desc->number, desc->name); > > > > > + return -EOPNOTSUPP; > > > > > > > > I believe ENOTSUPP is the correct one. EOPNOTSUPP seems to be > > > > network > > > > related. > > > > > > > > > > if we change it as "ENOTSUPP", it will report warning"ENOTSUPP is > > > not > > > SUSV4 error code, prefer EOPNOTSUP" when checking patch. > > > > The context surrounding this warning seems to be that ENOTSUPP is > > hard > > for userspace to understand. AFAIK the return code here does not get > > passed to userspace? And the pinctrl core does check for EINVAL or > > ENOTSUPP, so I think this is a valid use case. > > > > Linus? > > > > Hi Linus, > > Do you have some suggestion for the warning issue? Ignore the warning. We use this code in the pinctrl subsystem. The pinctrl subsystem has no userspace API so that is certainly not a problem. Yours, Linus Walleij