On Thu, Feb 13, 2020 at 07:22:46PM +0100, Pavel Machek wrote:
> Hi!
>
> > commit 277375b864e8147975b064b513f491e2a910e66a upstream
> >
> > In case the cable is not connected then the target gets into
> > an infinite wait for tx empty.
> > Add a timeout to the tx empty wait.
>
> Was this tested? Because it does not work...
>
> > Reported-by: Jean-Francois Dagenais <jeff.dagenais@xxxxxxxxx>
> > Signed-off-by: Shubhrajyoti Datta <shubhrajyoti.datta@xxxxxxxxxx>
> > Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> > Cc: stable <stable@xxxxxxxxxxxxxxx> # 4.19
> > Signed-off-by: Mathieu Poirier <mathieu.poirier@xxxxxxxxxx>
> > Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
>
> > @@ -681,16 +683,20 @@ static void cdns_uart_set_termios(struct uart_port *port,
> ...
> > + int err;
> >
> > spin_lock_irqsave(&port->lock, flags);
> >
> > /* Wait for the transmit FIFO to empty before making changes */
> > if (!(readl(port->membase + CDNS_UART_CR) &
> > CDNS_UART_CR_TX_DIS)) {
> > - while (!(readl(port->membase + CDNS_UART_SR) &
> > - CDNS_UART_SR_TXEMPTY)) {
> > - cpu_relax();
> > + err = readl_poll_timeout(port->membase + CDNS_UART_SR,
> > + val, (val & CDNS_UART_SR_TXEMPTY),
> > + 1000, TX_TIMEOUT);
> > + if (err) {
> > + dev_err(port->dev, "timed out waiting for tx empty");
> > + return;
> > }
> > }
> >
>
> It will return with lock held and interrupts disabled. Mainline takes
> spinlock later, so it does not have a problem.
>
> Merging 107475685abfdee504bb0ef4824f15797f6d2d4d before this one
> should fix the problem.
Good catch, that's a mess. I'll go queue this up now.
greg k-h
