On Tue, Jan 17, 2012 at 11:38:22PM +0100, Johan Hovold wrote:
> On Mon, Jan 16, 2012 at 7:46 PM, Greg KH <gregkh@xxxxxxx> wrote:
> > On Mon, Jan 16, 2012 at 10:37:24AM -0800, Stephen Hemminger wrote:
> >> When using USB serial console, there is an infinite loop on shutdown.
> >> The console fills with the error message:
> >> ttyUSB0 - error submitting urb: -22
> >>
> >> It seems that writing the error message is causing the error, which
> >> cause the message, which causes the error, ...
> >> The following is needed to avoid the infinite loop.
> >
> > Ick.
> >
> >> Signed-off-by: Stephen Hemminger <shemminger@xxxxxxxxxx>
> >>
> >> --- a/drivers/usb/serial/generic.c 2012-01-10 10:56:58.667176726 -0800
> >> +++ b/drivers/usb/serial/generic.c 2012-01-16 09:37:26.754919670 -0800
> >> @@ -217,8 +217,6 @@ retry:
> >> clear_bit(i, &port->write_urbs_free);
> >> result = usb_submit_urb(urb, GFP_ATOMIC);
> >> if (result) {
> >> - dev_err(&port->dev, "%s - error submitting urb: %d\n",
> >> - __func__, result);
> >
> > How about we switch this to dev_dbg()? "Normal" users will never see
> > this then. Will that work for you?
>
> Or, may I suggest, something along the lines of
>
> if (result) {
> if (!port->port.console)
> dev_err(...
>
> This way the error will generally not get lost, and it will be (more) clear why
> we're using dev_dbg rather then dev_err (and thus less likely to be changed
> back in the future).
>
> The remaining custom write implementations should probably be audited for
> dev_err (and friends) in their write paths as well.
That sounds good, care to send a patch for this?
thanks,
greg k-h
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
