[v3] serial: liteuart: add IRQ support

[PATCH v3 12/14] serial: liteuart: add IRQ support for the RX path

Posted by Gabriel Somlo 3 years, 5 months ago

Add support for IRQ-driven RX. Support for the TX path will be added
in a separate commit.

Signed-off-by: Gabriel Somlo <gsomlo@gmail.com>
---
 drivers/tty/serial/liteuart.c | 61 +++++++++++++++++++++++++++++++----
 1 file changed, 54 insertions(+), 7 deletions(-)

diff --git a/drivers/tty/serial/liteuart.c b/drivers/tty/serial/liteuart.c
index cf1ce597b45e..e30adb30277f 100644
--- a/drivers/tty/serial/liteuart.c
+++ b/drivers/tty/serial/liteuart.c
@@ -6,6 +6,7 @@
  */
 
 #include <linux/console.h>
+#include <linux/interrupt.h>
 #include <linux/litex.h>
 #include <linux/module.h>
 #include <linux/of.h>
@@ -130,13 +131,29 @@ static void liteuart_rx_chars(struct uart_port *port)
 	tty_flip_buffer_push(&port->state->port);
 }
 
+static irqreturn_t liteuart_interrupt(int irq, void *data)
+{
+	struct liteuart_port *uart = data;
+	struct uart_port *port = &uart->port;
+	u8 isr = litex_read8(port->membase + OFF_EV_PENDING);
+
+	/* for now, only rx path triggers interrupts */
+	isr &= EV_RX;
+
+	spin_lock(&port->lock);
+	if (isr & EV_RX)
+		liteuart_rx_chars(port);
+	spin_unlock(&port->lock);
+
+	return IRQ_RETVAL(isr);
+}
+
 static void liteuart_timer(struct timer_list *t)
 {
 	struct liteuart_port *uart = from_timer(uart, t, timer);
 	struct uart_port *port = &uart->port;
 
-	liteuart_rx_chars(port);
-
+	liteuart_interrupt(0, port);
 	mod_timer(&uart->timer, jiffies + uart_poll_timeout(port));
 }
 
@@ -162,19 +179,42 @@ static unsigned int liteuart_get_mctrl(struct uart_port *port)
 static int liteuart_startup(struct uart_port *port)
 {
 	struct liteuart_port *uart = to_liteuart_port(port);
+	int ret;
+	u8 irq_mask = 0;
 
-	/* disable events */
-	litex_write8(port->membase + OFF_EV_ENABLE, 0);
+	if (port->irq) {
+		ret = request_irq(port->irq, liteuart_interrupt, 0,
+				  KBUILD_MODNAME, uart);
+		if (ret == 0) {
+			/* only enable rx interrupts at this time */
+			irq_mask = EV_RX;
+		} else {
+			pr_err(pr_fmt("line %d irq %d failed: using polling\n"),
+				port->line, port->irq);
+			port->irq = 0;
+		}
+	}
 
-	/* prepare timer for polling */
-	timer_setup(&uart->timer, liteuart_timer, 0);
-	mod_timer(&uart->timer, jiffies + uart_poll_timeout(port));
+	if (!port->irq) {
+		timer_setup(&uart->timer, liteuart_timer, 0);
+		mod_timer(&uart->timer, jiffies + uart_poll_timeout(port));
+	}
+
+	litex_write8(port->membase + OFF_EV_ENABLE, irq_mask);
 
 	return 0;
 }
 
 static void liteuart_shutdown(struct uart_port *port)
 {
+	struct liteuart_port *uart = to_liteuart_port(port);
+
+	litex_write8(port->membase + OFF_EV_ENABLE, 0);
+
+	if (port->irq)
+		free_irq(port->irq, port);
+	else
+		del_timer_sync(&uart->timer);
 }
 
 static void liteuart_set_termios(struct uart_port *port, struct ktermios *new,
@@ -263,6 +303,13 @@ static int liteuart_probe(struct platform_device *pdev)
 		goto err_erase_id;
 	}
 
+	/* get irq */
+	ret = platform_get_irq_optional(pdev, 0);
+	if (ret < 0 && ret != -ENXIO)
+		return ret;
+	if (ret > 0)
+		port->irq = ret;
+
 	/* values not from device tree */
 	port->dev = &pdev->dev;
 	port->iotype = UPIO_MEM;
-- 
2.37.3

Re: [PATCH v3 12/14] serial: liteuart: add IRQ support for the RX path

Posted by Ilpo Järvinen 3 years, 4 months ago

On Sat, 12 Nov 2022, Gabriel Somlo wrote:

> Add support for IRQ-driven RX. Support for the TX path will be added
> in a separate commit.
> 
> Signed-off-by: Gabriel Somlo <gsomlo@gmail.com>
> ---
>  drivers/tty/serial/liteuart.c | 61 +++++++++++++++++++++++++++++++----
>  1 file changed, 54 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/tty/serial/liteuart.c b/drivers/tty/serial/liteuart.c
> index cf1ce597b45e..e30adb30277f 100644
> --- a/drivers/tty/serial/liteuart.c
> +++ b/drivers/tty/serial/liteuart.c
> @@ -6,6 +6,7 @@
>   */
>  
>  #include <linux/console.h>
> +#include <linux/interrupt.h>
>  #include <linux/litex.h>
>  #include <linux/module.h>
>  #include <linux/of.h>
> @@ -130,13 +131,29 @@ static void liteuart_rx_chars(struct uart_port *port)
>  	tty_flip_buffer_push(&port->state->port);
>  }
>  
> +static irqreturn_t liteuart_interrupt(int irq, void *data)
> +{
> +	struct liteuart_port *uart = data;
> +	struct uart_port *port = &uart->port;
> +	u8 isr = litex_read8(port->membase + OFF_EV_PENDING);
> +
> +	/* for now, only rx path triggers interrupts */

Please don't add comment like this at all when your series removes it in a 
later patch.

> +	isr &= EV_RX;
> +
> +	spin_lock(&port->lock);
> +	if (isr & EV_RX)
> +		liteuart_rx_chars(port);
> +	spin_unlock(&port->lock);
> +
> +	return IRQ_RETVAL(isr);
> +}
> +
>  static void liteuart_timer(struct timer_list *t)
>  {
>  	struct liteuart_port *uart = from_timer(uart, t, timer);
>  	struct uart_port *port = &uart->port;
>  
> -	liteuart_rx_chars(port);
> -
> +	liteuart_interrupt(0, port);
>  	mod_timer(&uart->timer, jiffies + uart_poll_timeout(port));
>  }
>  
> @@ -162,19 +179,42 @@ static unsigned int liteuart_get_mctrl(struct uart_port *port)
>  static int liteuart_startup(struct uart_port *port)
>  {
>  	struct liteuart_port *uart = to_liteuart_port(port);
> +	int ret;
> +	u8 irq_mask = 0;
>  
> -	/* disable events */
> -	litex_write8(port->membase + OFF_EV_ENABLE, 0);
> +	if (port->irq) {
> +		ret = request_irq(port->irq, liteuart_interrupt, 0,
> +				  KBUILD_MODNAME, uart);
> +		if (ret == 0) {
> +			/* only enable rx interrupts at this time */

This comment seems pretty useless. Your code says very much the same.

-- 
 i.

> +			irq_mask = EV_RX;
> +		} else {
> +			pr_err(pr_fmt("line %d irq %d failed: using polling\n"),
> +				port->line, port->irq);
> +			port->irq = 0;
> +		}
> +	}
>  
> -	/* prepare timer for polling */
> -	timer_setup(&uart->timer, liteuart_timer, 0);
> -	mod_timer(&uart->timer, jiffies + uart_poll_timeout(port));
> +	if (!port->irq) {
> +		timer_setup(&uart->timer, liteuart_timer, 0);
> +		mod_timer(&uart->timer, jiffies + uart_poll_timeout(port));
> +	}
> +
> +	litex_write8(port->membase + OFF_EV_ENABLE, irq_mask);
>  
>  	return 0;
>  }
>  
>  static void liteuart_shutdown(struct uart_port *port)
>  {
> +	struct liteuart_port *uart = to_liteuart_port(port);
> +
> +	litex_write8(port->membase + OFF_EV_ENABLE, 0);
> +
> +	if (port->irq)
> +		free_irq(port->irq, port);
> +	else
> +		del_timer_sync(&uart->timer);
>  }
>  
>  static void liteuart_set_termios(struct uart_port *port, struct ktermios *new,
> @@ -263,6 +303,13 @@ static int liteuart_probe(struct platform_device *pdev)
>  		goto err_erase_id;
>  	}
>  
> +	/* get irq */
> +	ret = platform_get_irq_optional(pdev, 0);
> +	if (ret < 0 && ret != -ENXIO)
> +		return ret;
> +	if (ret > 0)
> +		port->irq = ret;
> +
>  	/* values not from device tree */
>  	port->dev = &pdev->dev;
>  	port->iotype = UPIO_MEM;
>

Re: [PATCH v3 12/14] serial: liteuart: add IRQ support for the RX path

Posted by Gabriel L. Somlo 3 years, 4 months ago

On Tue, Nov 15, 2022 at 06:00:11PM +0200, Ilpo Järvinen wrote:
> On Sat, 12 Nov 2022, Gabriel Somlo wrote:
> 
> > Add support for IRQ-driven RX. Support for the TX path will be added
> > in a separate commit.
> > 
> > Signed-off-by: Gabriel Somlo <gsomlo@gmail.com>
> > ---
> >  drivers/tty/serial/liteuart.c | 61 +++++++++++++++++++++++++++++++----
> >  1 file changed, 54 insertions(+), 7 deletions(-)
> > 
> > diff --git a/drivers/tty/serial/liteuart.c b/drivers/tty/serial/liteuart.c
> > index cf1ce597b45e..e30adb30277f 100644
> > --- a/drivers/tty/serial/liteuart.c
> > +++ b/drivers/tty/serial/liteuart.c
> > @@ -6,6 +6,7 @@
> >   */
> >  
> >  #include <linux/console.h>
> > +#include <linux/interrupt.h>
> >  #include <linux/litex.h>
> >  #include <linux/module.h>
> >  #include <linux/of.h>
> > @@ -130,13 +131,29 @@ static void liteuart_rx_chars(struct uart_port *port)
> >  	tty_flip_buffer_push(&port->state->port);
> >  }
> >  
> > +static irqreturn_t liteuart_interrupt(int irq, void *data)
> > +{
> > +	struct liteuart_port *uart = data;
> > +	struct uart_port *port = &uart->port;
> > +	u8 isr = litex_read8(port->membase + OFF_EV_PENDING);
> > +
> > +	/* for now, only rx path triggers interrupts */
> 
> Please don't add comment like this at all when your series removes it in a 
> later patch.

OK, I will remove it in v4.

> > +	isr &= EV_RX;
> > +
> > +	spin_lock(&port->lock);
> > +	if (isr & EV_RX)
> > +		liteuart_rx_chars(port);
> > +	spin_unlock(&port->lock);
> > +
> > +	return IRQ_RETVAL(isr);
> > +}
> > +
> >  static void liteuart_timer(struct timer_list *t)
> >  {
> >  	struct liteuart_port *uart = from_timer(uart, t, timer);
> >  	struct uart_port *port = &uart->port;
> >  
> > -	liteuart_rx_chars(port);
> > -
> > +	liteuart_interrupt(0, port);
> >  	mod_timer(&uart->timer, jiffies + uart_poll_timeout(port));
> >  }
> >  
> > @@ -162,19 +179,42 @@ static unsigned int liteuart_get_mctrl(struct uart_port *port)
> >  static int liteuart_startup(struct uart_port *port)
> >  {
> >  	struct liteuart_port *uart = to_liteuart_port(port);
> > +	int ret;
> > +	u8 irq_mask = 0;
> >  
> > -	/* disable events */
> > -	litex_write8(port->membase + OFF_EV_ENABLE, 0);
> > +	if (port->irq) {
> > +		ret = request_irq(port->irq, liteuart_interrupt, 0,
> > +				  KBUILD_MODNAME, uart);
> > +		if (ret == 0) {
> > +			/* only enable rx interrupts at this time */
> 
> This comment seems pretty useless. Your code says very much the same.

The comment was meant to let the reader know that the code is doing it
*intentionally* (rather than forgetting to enable tx irqs by mistake).
But I'm OK with removing this comment in v4 as well if you think
that's an overly paranoid and redundant thing to do... :)

Thanks,
--Gabriel
 
> -- 
>  i.
> 
> > +			irq_mask = EV_RX;
> > +		} else {
> > +			pr_err(pr_fmt("line %d irq %d failed: using polling\n"),
> > +				port->line, port->irq);
> > +			port->irq = 0;
> > +		}
> > +	}
> >  
> > -	/* prepare timer for polling */
> > -	timer_setup(&uart->timer, liteuart_timer, 0);
> > -	mod_timer(&uart->timer, jiffies + uart_poll_timeout(port));
> > +	if (!port->irq) {
> > +		timer_setup(&uart->timer, liteuart_timer, 0);
> > +		mod_timer(&uart->timer, jiffies + uart_poll_timeout(port));
> > +	}
> > +
> > +	litex_write8(port->membase + OFF_EV_ENABLE, irq_mask);
> >  
> >  	return 0;
> >  }
> >  
> >  static void liteuart_shutdown(struct uart_port *port)
> >  {
> > +	struct liteuart_port *uart = to_liteuart_port(port);
> > +
> > +	litex_write8(port->membase + OFF_EV_ENABLE, 0);
> > +
> > +	if (port->irq)
> > +		free_irq(port->irq, port);
> > +	else
> > +		del_timer_sync(&uart->timer);
> >  }
> >  
> >  static void liteuart_set_termios(struct uart_port *port, struct ktermios *new,
> > @@ -263,6 +303,13 @@ static int liteuart_probe(struct platform_device *pdev)
> >  		goto err_erase_id;
> >  	}
> >  
> > +	/* get irq */
> > +	ret = platform_get_irq_optional(pdev, 0);
> > +	if (ret < 0 && ret != -ENXIO)
> > +		return ret;
> > +	if (ret > 0)
> > +		port->irq = ret;
> > +
> >  	/* values not from device tree */
> >  	port->dev = &pdev->dev;
> >  	port->iotype = UPIO_MEM;
> > 
>

Re: [PATCH v3 12/14] serial: liteuart: add IRQ support for the RX path

Posted by Ilpo Järvinen 3 years, 4 months ago

On Tue, 15 Nov 2022, Gabriel L. Somlo wrote:

> On Tue, Nov 15, 2022 at 06:00:11PM +0200, Ilpo Järvinen wrote:
> > On Sat, 12 Nov 2022, Gabriel Somlo wrote:
> > 
> > > Add support for IRQ-driven RX. Support for the TX path will be added
> > > in a separate commit.
> > > 
> > > Signed-off-by: Gabriel Somlo <gsomlo@gmail.com>
> > > ---
> > >  drivers/tty/serial/liteuart.c | 61 +++++++++++++++++++++++++++++++----
> > >  1 file changed, 54 insertions(+), 7 deletions(-)
> > > 
> > > diff --git a/drivers/tty/serial/liteuart.c b/drivers/tty/serial/liteuart.c
> > > index cf1ce597b45e..e30adb30277f 100644
> > > --- a/drivers/tty/serial/liteuart.c
> > > +++ b/drivers/tty/serial/liteuart.c
> > > @@ -6,6 +6,7 @@
> > >   */
> > >  
> > >  #include <linux/console.h>
> > > +#include <linux/interrupt.h>
> > >  #include <linux/litex.h>
> > >  #include <linux/module.h>
> > >  #include <linux/of.h>
> > > @@ -130,13 +131,29 @@ static void liteuart_rx_chars(struct uart_port *port)
> > >  	tty_flip_buffer_push(&port->state->port);
> > >  }
> > >  
> > > +static irqreturn_t liteuart_interrupt(int irq, void *data)
> > > +{
> > > +	struct liteuart_port *uart = data;
> > > +	struct uart_port *port = &uart->port;
> > > +	u8 isr = litex_read8(port->membase + OFF_EV_PENDING);
> > > +
> > > +	/* for now, only rx path triggers interrupts */
> > 
> > Please don't add comment like this at all when your series removes it in a 
> > later patch.
> 
> OK, I will remove it in v4.
> 
> > > +	isr &= EV_RX;
> > > +
> > > +	spin_lock(&port->lock);
> > > +	if (isr & EV_RX)
> > > +		liteuart_rx_chars(port);
> > > +	spin_unlock(&port->lock);
> > > +
> > > +	return IRQ_RETVAL(isr);
> > > +}
> > > +
> > >  static void liteuart_timer(struct timer_list *t)
> > >  {
> > >  	struct liteuart_port *uart = from_timer(uart, t, timer);
> > >  	struct uart_port *port = &uart->port;
> > >  
> > > -	liteuart_rx_chars(port);
> > > -
> > > +	liteuart_interrupt(0, port);
> > >  	mod_timer(&uart->timer, jiffies + uart_poll_timeout(port));
> > >  }
> > >  
> > > @@ -162,19 +179,42 @@ static unsigned int liteuart_get_mctrl(struct uart_port *port)
> > >  static int liteuart_startup(struct uart_port *port)
> > >  {
> > >  	struct liteuart_port *uart = to_liteuart_port(port);
> > > +	int ret;
> > > +	u8 irq_mask = 0;
> > >  
> > > -	/* disable events */
> > > -	litex_write8(port->membase + OFF_EV_ENABLE, 0);
> > > +	if (port->irq) {
> > > +		ret = request_irq(port->irq, liteuart_interrupt, 0,
> > > +				  KBUILD_MODNAME, uart);
> > > +		if (ret == 0) {
> > > +			/* only enable rx interrupts at this time */
> > 
> > This comment seems pretty useless. Your code says very much the same.
> 
> The comment was meant to let the reader know that the code is doing it
> *intentionally* (rather than forgetting to enable tx irqs by mistake).
> But I'm OK with removing this comment in v4 as well if you think
> that's an overly paranoid and redundant thing to do... :)

I see. Reading the other comment first caused me to misinterpret this one 
to mean that only RX interrupts are implemented.

Maybe if you change "at this time" to "at startup" it would make it more 
obvious.

-- 
 i.

Re: [PATCH v3 12/14] serial: liteuart: add IRQ support for the RX path

Posted by Gabriel L. Somlo 3 years, 4 months ago

On Tue, Nov 15, 2022 at 06:21:00PM +0200, Ilpo Järvinen wrote:
> On Tue, 15 Nov 2022, Gabriel L. Somlo wrote:
> 
> > On Tue, Nov 15, 2022 at 06:00:11PM +0200, Ilpo Järvinen wrote:
> > > On Sat, 12 Nov 2022, Gabriel Somlo wrote:
> > > 
> > > > Add support for IRQ-driven RX. Support for the TX path will be added
> > > > in a separate commit.
> > > > 
> > > > Signed-off-by: Gabriel Somlo <gsomlo@gmail.com>
> > > > ---
> > > >  drivers/tty/serial/liteuart.c | 61 +++++++++++++++++++++++++++++++----
> > > >  1 file changed, 54 insertions(+), 7 deletions(-)
> > > > 
> > > > diff --git a/drivers/tty/serial/liteuart.c b/drivers/tty/serial/liteuart.c
> > > > index cf1ce597b45e..e30adb30277f 100644
> > > > --- a/drivers/tty/serial/liteuart.c
> > > > +++ b/drivers/tty/serial/liteuart.c
> > > > @@ -6,6 +6,7 @@
> > > >   */
> > > >  
> > > >  #include <linux/console.h>
> > > > +#include <linux/interrupt.h>
> > > >  #include <linux/litex.h>
> > > >  #include <linux/module.h>
> > > >  #include <linux/of.h>
> > > > @@ -130,13 +131,29 @@ static void liteuart_rx_chars(struct uart_port *port)
> > > >  	tty_flip_buffer_push(&port->state->port);
> > > >  }
> > > >  
> > > > +static irqreturn_t liteuart_interrupt(int irq, void *data)
> > > > +{
> > > > +	struct liteuart_port *uart = data;
> > > > +	struct uart_port *port = &uart->port;
> > > > +	u8 isr = litex_read8(port->membase + OFF_EV_PENDING);
> > > > +
> > > > +	/* for now, only rx path triggers interrupts */
> > > 
> > > Please don't add comment like this at all when your series removes it in a 
> > > later patch.
> > 
> > OK, I will remove it in v4.
> > 
> > > > +	isr &= EV_RX;
> > > > +
> > > > +	spin_lock(&port->lock);
> > > > +	if (isr & EV_RX)
> > > > +		liteuart_rx_chars(port);
> > > > +	spin_unlock(&port->lock);
> > > > +
> > > > +	return IRQ_RETVAL(isr);
> > > > +}
> > > > +
> > > >  static void liteuart_timer(struct timer_list *t)
> > > >  {
> > > >  	struct liteuart_port *uart = from_timer(uart, t, timer);
> > > >  	struct uart_port *port = &uart->port;
> > > >  
> > > > -	liteuart_rx_chars(port);
> > > > -
> > > > +	liteuart_interrupt(0, port);
> > > >  	mod_timer(&uart->timer, jiffies + uart_poll_timeout(port));
> > > >  }
> > > >  
> > > > @@ -162,19 +179,42 @@ static unsigned int liteuart_get_mctrl(struct uart_port *port)
> > > >  static int liteuart_startup(struct uart_port *port)
> > > >  {
> > > >  	struct liteuart_port *uart = to_liteuart_port(port);
> > > > +	int ret;
> > > > +	u8 irq_mask = 0;
> > > >  
> > > > -	/* disable events */
> > > > -	litex_write8(port->membase + OFF_EV_ENABLE, 0);
> > > > +	if (port->irq) {
> > > > +		ret = request_irq(port->irq, liteuart_interrupt, 0,
> > > > +				  KBUILD_MODNAME, uart);
> > > > +		if (ret == 0) {
> > > > +			/* only enable rx interrupts at this time */
> > > 
> > > This comment seems pretty useless. Your code says very much the same.
> > 
> > The comment was meant to let the reader know that the code is doing it
> > *intentionally* (rather than forgetting to enable tx irqs by mistake).
> > But I'm OK with removing this comment in v4 as well if you think
> > that's an overly paranoid and redundant thing to do... :)
> 
> I see. Reading the other comment first caused me to misinterpret this one 
> to mean that only RX interrupts are implemented.
> 
> Maybe if you change "at this time" to "at startup" it would make it more 
> obvious.
 
OK, I'll fix the comment per your suggestion rather than get rid of it.

Thanks again,
--G

> -- 
>  i.