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