On Tue, 2017-01-10 at 18:11 +0300, Alexey Brodkin wrote:
> With CONFIG_DYNAMIC_DEBUG if dyndbg enables debug output in
> 8250_port.c deadlock happens inevitably on UART IRQ handling.
>
> That's the problematic execution path:
> ---------------------------->8------------------------
> UART IRQ:
> serial8250_interrupt() ->
> serial8250_handle_irq(): lock "port->lock" ->
> pr_debug() ->
> serial8250_console_write(): bump in locked "port->lock".
>
> OR (if above pr_debug() gets removed):
> serial8250_tx_chars() ->
> pr_debug() ->
> serial8250_console_write(): bump in locked "port->lock".
> ---------------------------->8------------------------
>
> So let's get rid of those not that much useful debug entries.
>
> Discussed problem could be easily reproduced with QEMU for x86_64.
> As well as this fix could be mimicked with muting of dynamic debug for
> the problematic lines as simple as:
> ---------------------------->8------------------------
> dyndbg="+p; file 8250_port.c line 1756 -p; file 8250_port.c line 1822
> -p"
> ---------------------------->8------------------------
Before it was disabled globally for the driver, but someone converted
macro to pr_debug() and problem appears.
Since it's known issue I remember discussion where someone proposed to
schedule printk() in such cases. Don't know if the idea made upstream or
on its way.
FWIW:
Reviewed-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
>
> Signed-off-by: Alexey Brodkin <abrodkin@xxxxxxxxxxxx>
> Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> Cc: Jiri Slaby <jslaby@xxxxxxxx>
> Cc: Peter Hurley <peter@xxxxxxxxxxxxxxxxxx>
> Cc: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
> Cc: Phillip Raffeck <phillip.raffeck@xxxxxx>
> Cc: Anton Wuerfel <anton.wuerfel@xxxxxx>
> Cc: "Matwey V. Kornilov" <matwey@xxxxxxxxxx>
> Cc: Yegor Yefremov <yegorslists@xxxxxxxxxxxxxx>
> Cc: Thor Thayer <tthayer@xxxxxxxxxxxxxxxxxxxxx>
> ---
> drivers/tty/serial/8250/8250_port.c | 4 ----
> 1 file changed, 4 deletions(-)
>
> diff --git a/drivers/tty/serial/8250/8250_port.c
> b/drivers/tty/serial/8250/8250_port.c
> index fe4399b41df6..3cfdd745a97a 100644
> --- a/drivers/tty/serial/8250/8250_port.c
> +++ b/drivers/tty/serial/8250/8250_port.c
> @@ -1753,8 +1753,6 @@ void serial8250_tx_chars(struct uart_8250_port
> *up)
> if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
> uart_write_wakeup(port);
>
> - pr_debug("%s: THRE\n", __func__);
> -
> /*
> * With RPM enabled, we have to wait until the FIFO is empty
> before the
> * HW can go idle. So we get here once again with empty FIFO
> and disable
> @@ -1819,8 +1817,6 @@ int serial8250_handle_irq(struct uart_port
> *port, unsigned int iir)
>
> status = serial_port_in(port, UART_LSR);
>
> - pr_debug("%s: status = %x\n", __func__, status);
> -
> if (status & (UART_LSR_DR | UART_LSR_BI)) {
> if (!up->dma || handle_rx_dma(up, iir))
> status = serial8250_rx_chars(up, status);
--
Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
Intel Finland Oy
--
To unsubscribe from this list: send the line "unsubscribe linux-serial" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
