Re: [PATCH 2/9] serial: sh-sci: Check if TX data was written to device in .tx_empty()

[Claudiu Beznea <claudiu.beznea@tuxon.dev>]

Hi, Greg,

On 07.11.2024 10:47, Greg KH wrote:
> On Wed, Nov 06, 2024 at 02:01:11PM +0200, Claudiu wrote:
>> From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
>>
>> On the Renesas RZ/G3S, when doing suspend to RAM, the uart_suspend_port()
>> is called. The uart_suspend_port() calls 3 times the
>> struct uart_port::ops::tx_empty() before shutting down the port.
>>
>> According to the documentation, the struct uart_port::ops::tx_empty()
>> API tests whether the transmitter FIFO and shifter for the port is
>> empty.
>>
>> The Renesas RZ/G3S SCIFA IP reports the number of data units stored in the
>> transmit FIFO through the FDR (FIFO Data Count Register). The data units
>> in the FIFOs are written in the shift register and transmitted from there.
>> The TEND bit in the Serial Status Register reports if the data was
>> transmitted from the shift register.
>>
>> In the previous code, in the tx_empty() API implemented by the sh-sci
>> driver, it is considered that the TX is empty if the hardware reports the
>> TEND bit set and the number of data units in the FIFO is zero.
>>
>> According to the HW manual, the TEND bit has the following meaning:
>>
>> 0: Transmission is in the waiting state or in progress.
>> 1: Transmission is completed.
>>
>> It has been noticed that when opening the serial device w/o using it and
>> then switch to a power saving mode, the tx_empty() call in the
>> uart_port_suspend() function fails, leading to the "Unable to drain
>> transmitter" message being printed on the console. This is because the
>> TEND=0 if nothing has been transmitted and the FIFOs are empty. As the
>> TEND=0 has double meaning (waiting state, in progress) we can't
>> determined the scenario described above.
>>
>> Add a software workaround for this. This sets a variable if any data has
>> been sent on the serial console (when using PIO) or if the DMA callback has
>> been called (meaning something has been transmitted).
>>
>> Fixes: 73a19e4c0301 ("serial: sh-sci: Add DMA support.")
>> Cc: stable@vger.kernel.org
>> Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
>> ---
>>  drivers/tty/serial/sh-sci.c | 11 +++++++++++
>>  1 file changed, 11 insertions(+)
>>
>> diff --git a/drivers/tty/serial/sh-sci.c b/drivers/tty/serial/sh-sci.c
>> index df523c744423..8e2d534401fa 100644
>> --- a/drivers/tty/serial/sh-sci.c
>> +++ b/drivers/tty/serial/sh-sci.c
>> @@ -153,6 +153,7 @@ struct sci_port {
>>  	int				rx_trigger;
>>  	struct timer_list		rx_fifo_timer;
>>  	int				rx_fifo_timeout;
>> +	atomic_t			first_time_tx;
> 
> Don't use an atomic variable for an informational thing like this, it is
> racy and doesn't work properly.  Either use a real lock (because you
> care about the locking stuff here), or just use a boolean and live with
> any potential races.

OK, I'll drop it and use a boolean.

> 
> 
> 
>>  	u16				hscif_tot;
>>  
>>  	bool has_rtscts;
>> @@ -850,6 +851,7 @@ static void sci_transmit_chars(struct uart_port *port)
>>  {
>>  	struct tty_port *tport = &port->state->port;
>>  	unsigned int stopped = uart_tx_stopped(port);
>> +	struct sci_port *s = to_sci_port(port);
>>  	unsigned short status;
>>  	unsigned short ctrl;
>>  	int count;
>> @@ -885,6 +887,7 @@ static void sci_transmit_chars(struct uart_port *port)
>>  		}
>>  
>>  		sci_serial_out(port, SCxTDR, c);
>> +		atomic_set(&s->first_time_tx, 1);
>>  
>>  		port->icount.tx++;
>>  	} while (--count > 0);
>> @@ -1241,6 +1244,8 @@ static void sci_dma_tx_complete(void *arg)
>>  	if (kfifo_len(&tport->xmit_fifo) < WAKEUP_CHARS)
>>  		uart_write_wakeup(port);
>>  
>> +	atomic_set(&s->first_time_tx, 1);
>> +
>>  	if (!kfifo_is_empty(&tport->xmit_fifo)) {
>>  		s->cookie_tx = 0;
>>  		schedule_work(&s->work_tx);
>> @@ -2076,6 +2081,10 @@ static unsigned int sci_tx_empty(struct uart_port *port)
>>  {
>>  	unsigned short status = sci_serial_in(port, SCxSR);
>>  	unsigned short in_tx_fifo = sci_txfill(port);
>> +	struct sci_port *s = to_sci_port(port);
>> +
>> +	if (!atomic_read(&s->first_time_tx))
>> +		return TIOCSER_TEMT;
> 
> See, what happens here if the value changes right after you check it?

I agree. I am aware if it.

I chose this approach (w/o locking) as I noticed (as of my code checking)
that this function is called in kernel through uart_ioctl(),
uart_wait_until_sent(), uart_suspend_port().

The uart_wait_until_sent(), uart_suspend_port() are implementing a multiple
try approach when checking the ops::tx_timeout() return value.

I haven't checked any user space application but considered that it might
work in a similar way.

I will switch to a boolean in the next version.

Thank you,
Claudiu Beznea


> Being an atomic doesn't mean anything :(
> 
> thanks,
> 
> greg k-h