diff mbox

[03/16] tty: serial: 8250_core: read only RX if there is something in the FIFO

Message ID 20150212192333.GA18882@linutronix.de (mailing list archive)
State New, archived
Headers show

Commit Message

Sebastian Andrzej Siewior Feb. 12, 2015, 7:23 p.m. UTC
* Peter Hurley | 2015-02-12 11:32:04 [-0500]:

>That said, I don't think serial8250_do_startup() is really doing that much
>for OMAP h/w startup; open-coding what omap_8250 really needs is probably
>< 10 loc.

10 loc? I have a few more. serial8250_clear_fifos(),
serial_link_irq_chain() aren't exported. serial8250_set_mctrl() can
maybe accessed via uart_ops->set_mctrl(). Maybe I'm not removing the
obvious not required code but here it looks better to just a BUG flag for
the Omap.


Sebastian

Comments

Peter Hurley Feb. 12, 2015, 7:55 p.m. UTC | #1
On 02/12/2015 02:23 PM, Sebastian Andrzej Siewior wrote:
> * Peter Hurley | 2015-02-12 11:32:04 [-0500]:
> 
>> That said, I don't think serial8250_do_startup() is really doing that much
>> for OMAP h/w startup; open-coding what omap_8250 really needs is probably
>> < 10 loc.
> 
> 10 loc? I have a few more.

:)

> serial8250_clear_fifos(),
> serial_link_irq_chain() aren't exported. serial8250_set_mctrl() can
> maybe accessed via uart_ops->set_mctrl(). Maybe I'm not removing the
> obvious not required code but here it looks better to just a BUG flag for
> the Omap.

Ok.

FWIW,

> --- a/drivers/tty/serial/8250/8250_omap.c
> +++ b/drivers/tty/serial/8250/8250_omap.c
> @@ -557,9 +557,74 @@ static int omap_8250_startup(struct uart_port *port)
>  
>  	pm_runtime_get_sync(port->dev);
>  
> -	ret = serial8250_do_startup(port);
> -	if (ret)
> -		goto err;
> +	up->mcr = 0;
> +
> +	/*
> +	 * Clear the FIFO buffers and disable them.
> +	 * (they will be reenabled in set_termios())
> +	 */
> +	serial8250_clear_fifos(up);

For omap this would just be:

	serial_out(up, UART_FCR, UART_FCR_CLEAR_RCVR | UART_FCR_CLEAR_XMIT);

In any event, the fifo enable/disable is probably not happening since
FIFO_EN is ignored unless the divisor == 0.

> +
> +	/*
> +	 * Clear the interrupt registers.
> +	 */
> +	if (serial_port_in(port, UART_LSR) & UART_LSR_DR)
> +		serial_port_in(port, UART_RX);
> +	serial_port_in(port, UART_IIR);
> +	serial_port_in(port, UART_MSR);
> +
> +	retval = serial_link_irq_chain(up);
> +	if (retval)
> +		goto out;

omap doesn't really need the legacy irq chain handling; this could just be
request_irq().

In the 8250 split I'll be posting soon, all the irq chaining and
polling-via-timeout workarounds stays in the universal/legacy driver so
other 8250 drivers can opt-out.

> +
> +	/*
> +	 * Now, initialize the UART
> +	 */
> +	serial_port_out(port, UART_LCR, UART_LCR_WLEN8);
> +

--- >% ---
> +	spin_lock_irqsave(&port->lock, flags);
> +	/*
> +	 * Most PC uarts need OUT2 raised to enable interrupts.
> +	 */
> +	if (port->irq)
> +		up->port.mctrl |= TIOCM_OUT2;
> +
> +	serial8250_set_mctrl(port, port->mctrl);
> +
> +	spin_unlock_irqrestore(&port->lock, flags);
> +

None of this is required because there is no OUT2 on omap.
---------

--- >% ----
> +	/*
> +	 * Clear the interrupt registers again for luck, and clear the
> +	 * saved flags to avoid getting false values from polling
> +	 * routines or the previous session.
> +	 */
> +	if (serial_port_in(port, UART_LSR) & UART_LSR_DR)
> +		serial_port_in(port, UART_RX);
> +	serial_port_in(port, UART_IIR);
> +	serial_port_in(port, UART_MSR);

None of this is required because none of the probing is taking place.
------------


> +	up->lsr_saved_flags = 0;
> +	up->msr_saved_flags = 0;
> +
> +	/*
> +	 * Request DMA channels for both RX and TX.
> +	 */
> +	if (up->dma) {
> +		retval = serial8250_request_dma(up);
> +		if (retval) {
> +			pr_warn_ratelimited("ttyS%d - failed to request DMA\n",
> +					    serial_index(port));
> +			up->dma = NULL;
> +		}
> +	}
> +
> +	/*
> +	 * Finally, enable interrupts.  Note: Modem status interrupts
> +	 * are set via set_termios(), which will be occurring imminently
> +	 * anyway, so we don't enable them here.
> +	 */
> +	up->ier = UART_IER_RLSI | UART_IER_RDI;
> +	serial_port_out(port, UART_IER, up->ier);
> +
>  
>  #ifdef CONFIG_PM
>  	up->capabilities |= UART_CAP_RPM;
> 
> Sebastian
>
Sebastian Andrzej Siewior Feb. 12, 2015, 8:34 p.m. UTC | #2
On 02/12/2015 08:55 PM, Peter Hurley wrote:
> On 02/12/2015 02:23 PM, Sebastian Andrzej Siewior wrote:
>> * Peter Hurley | 2015-02-12 11:32:04 [-0500]:
>>
>>> That said, I don't think serial8250_do_startup() is really doing that much
>>> for OMAP h/w startup; open-coding what omap_8250 really needs is probably
>>> < 10 loc.
>>
>> 10 loc? I have a few more.
> 
> :)
> 
>> serial8250_clear_fifos(),
>> serial_link_irq_chain() aren't exported. serial8250_set_mctrl() can
>> maybe accessed via uart_ops->set_mctrl(). Maybe I'm not removing the
>> obvious not required code but here it looks better to just a BUG flag for
>> the Omap.
> 
> Ok.

Okay. I will try to post something tomorrow.

>> --- a/drivers/tty/serial/8250/8250_omap.c
>> +++ b/drivers/tty/serial/8250/8250_omap.c
>> +
>> +	/*
>> +	 * Clear the interrupt registers.
>> +	 */
>> +	if (serial_port_in(port, UART_LSR) & UART_LSR_DR)
>> +		serial_port_in(port, UART_RX);
>> +	serial_port_in(port, UART_IIR);
>> +	serial_port_in(port, UART_MSR);
>> +
>> +	retval = serial_link_irq_chain(up);
>> +	if (retval)
>> +		goto out;
> 
> omap doesn't really need the legacy irq chain handling; this could just be
> request_irq().
> 
> In the 8250 split I'll be posting soon, all the irq chaining and
> polling-via-timeout workarounds stays in the universal/legacy driver so
> other 8250 drivers can opt-out.

Ah. This sounds interesting.

Sebastian
diff mbox

Patch

--- a/drivers/tty/serial/8250/8250_omap.c
+++ b/drivers/tty/serial/8250/8250_omap.c
@@ -557,9 +557,74 @@  static int omap_8250_startup(struct uart_port *port)
 
 	pm_runtime_get_sync(port->dev);
 
-	ret = serial8250_do_startup(port);
-	if (ret)
-		goto err;
+	up->mcr = 0;
+
+	/*
+	 * Clear the FIFO buffers and disable them.
+	 * (they will be reenabled in set_termios())
+	 */
+	serial8250_clear_fifos(up);
+
+	/*
+	 * Clear the interrupt registers.
+	 */
+	if (serial_port_in(port, UART_LSR) & UART_LSR_DR)
+		serial_port_in(port, UART_RX);
+	serial_port_in(port, UART_IIR);
+	serial_port_in(port, UART_MSR);
+
+	retval = serial_link_irq_chain(up);
+	if (retval)
+		goto out;
+
+	/*
+	 * Now, initialize the UART
+	 */
+	serial_port_out(port, UART_LCR, UART_LCR_WLEN8);
+
+	spin_lock_irqsave(&port->lock, flags);
+	/*
+	 * Most PC uarts need OUT2 raised to enable interrupts.
+	 */
+	if (port->irq)
+		up->port.mctrl |= TIOCM_OUT2;
+
+	serial8250_set_mctrl(port, port->mctrl);
+
+	spin_unlock_irqrestore(&port->lock, flags);
+
+	/*
+	 * Clear the interrupt registers again for luck, and clear the
+	 * saved flags to avoid getting false values from polling
+	 * routines or the previous session.
+	 */
+	if (serial_port_in(port, UART_LSR) & UART_LSR_DR)
+		serial_port_in(port, UART_RX);
+	serial_port_in(port, UART_IIR);
+	serial_port_in(port, UART_MSR);
+	up->lsr_saved_flags = 0;
+	up->msr_saved_flags = 0;
+
+	/*
+	 * Request DMA channels for both RX and TX.
+	 */
+	if (up->dma) {
+		retval = serial8250_request_dma(up);
+		if (retval) {
+			pr_warn_ratelimited("ttyS%d - failed to request DMA\n",
+					    serial_index(port));
+			up->dma = NULL;
+		}
+	}
+
+	/*
+	 * Finally, enable interrupts.  Note: Modem status interrupts
+	 * are set via set_termios(), which will be occurring imminently
+	 * anyway, so we don't enable them here.
+	 */
+	up->ier = UART_IER_RLSI | UART_IER_RDI;
+	serial_port_out(port, UART_IER, up->ier);
+
 
 #ifdef CONFIG_PM
 	up->capabilities |= UART_CAP_RPM;