All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH 1/2] spi/pl022: timeout on polled transfer
@ 2011-05-19 12:13 ` Linus Walleij
  0 siblings, 0 replies; 8+ messages in thread
From: Linus Walleij @ 2011-05-19 12:13 UTC (permalink / raw)
  To: Grant Likely, spi-devel-general
  Cc: Linus Walleij, Lee Jones, linux-arm-kernel, Magnus Templing

From: Magnus Templing <magnus.templing@stericsson.com>

This adds the missing handling of polling timeouts and deletes
our last todo.

Signed-off-by: Magnus Templing <magnus.templing@stericsson.com>
Reviewed-by: Srinidhi Kasagar <srinidhi.kasagar@stericsson.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
 drivers/spi/amba-pl022.c |   22 ++++++++++++++--------
 1 files changed, 14 insertions(+), 8 deletions(-)

diff --git a/drivers/spi/amba-pl022.c b/drivers/spi/amba-pl022.c
index 08de58e..1f84a8a 100644
--- a/drivers/spi/amba-pl022.c
+++ b/drivers/spi/amba-pl022.c
@@ -24,11 +24,6 @@
  * GNU General Public License for more details.
  */
 
-/*
- * TODO:
- * - add timeout on polled transfers
- */
-
 #include <linux/init.h>
 #include <linux/module.h>
 #include <linux/device.h>
@@ -287,6 +282,8 @@
 
 #define CLEAR_ALL_INTERRUPTS  0x3
 
+#define SPI_POLLING_TIMEOUT 1000
+
 
 /*
  * The type of reading going on on this chip
@@ -1378,6 +1375,7 @@ static void do_polling_transfer(struct pl022 *pl022)
 	struct spi_transfer *transfer = NULL;
 	struct spi_transfer *previous = NULL;
 	struct chip_data *chip;
+	unsigned long timeout;
 
 	chip = pl022->cur_chip;
 	message = pl022->cur_msg;
@@ -1415,9 +1413,17 @@ static void do_polling_transfer(struct pl022 *pl022)
 		       SSP_CR1(pl022->virtbase));
 
 		dev_dbg(&pl022->adev->dev, "polling transfer ongoing ...\n");
-		/* FIXME: insert a timeout so we don't hang here indefinitely */
-		while (pl022->tx < pl022->tx_end || pl022->rx < pl022->rx_end)
+
+		timeout = jiffies + msecs_to_jiffies(SPI_POLLING_TIMEOUT);
+		while (pl022->tx < pl022->tx_end || pl022->rx < pl022->rx_end) {
+			if (time_after(jiffies, timeout)) {
+				dev_warn(&pl022->adev->dev,
+				"%s: timeout!\n", __func__);
+				message->state = STATE_ERROR;
+				goto out;
+			}
 			readwriter(pl022);
+		}
 
 		/* Update total byte transferred */
 		message->actual_length += pl022->cur_transfer->len;
@@ -1426,7 +1432,7 @@ static void do_polling_transfer(struct pl022 *pl022)
 		/* Move to next transfer */
 		message->state = next_transfer(pl022);
 	}
-
+out:
 	/* Handle end of message */
 	if (message->state == STATE_DONE)
 		message->status = 0;
-- 
1.7.3.2

^ permalink raw reply related	[flat|nested] 8+ messages in thread
* [PATCH 1/2] spi/pl022: timeout on polled transfer
@ 2011-05-19 12:13 ` Linus Walleij
  0 siblings, 0 replies; 8+ messages in thread
From: Linus Walleij @ 2011-05-19 12:13 UTC (permalink / raw)
  To: linux-arm-kernel

From: Magnus Templing <magnus.templing@stericsson.com>

This adds the missing handling of polling timeouts and deletes
our last todo.

Signed-off-by: Magnus Templing <magnus.templing@stericsson.com>
Reviewed-by: Srinidhi Kasagar <srinidhi.kasagar@stericsson.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
 drivers/spi/amba-pl022.c |   22 ++++++++++++++--------
 1 files changed, 14 insertions(+), 8 deletions(-)

diff --git a/drivers/spi/amba-pl022.c b/drivers/spi/amba-pl022.c
index 08de58e..1f84a8a 100644
--- a/drivers/spi/amba-pl022.c
+++ b/drivers/spi/amba-pl022.c
@@ -24,11 +24,6 @@
  * GNU General Public License for more details.
  */
 
-/*
- * TODO:
- * - add timeout on polled transfers
- */
-
 #include <linux/init.h>
 #include <linux/module.h>
 #include <linux/device.h>
@@ -287,6 +282,8 @@
 
 #define CLEAR_ALL_INTERRUPTS  0x3
 
+#define SPI_POLLING_TIMEOUT 1000
+
 
 /*
  * The type of reading going on on this chip
@@ -1378,6 +1375,7 @@ static void do_polling_transfer(struct pl022 *pl022)
 	struct spi_transfer *transfer = NULL;
 	struct spi_transfer *previous = NULL;
 	struct chip_data *chip;
+	unsigned long timeout;
 
 	chip = pl022->cur_chip;
 	message = pl022->cur_msg;
@@ -1415,9 +1413,17 @@ static void do_polling_transfer(struct pl022 *pl022)
 		       SSP_CR1(pl022->virtbase));
 
 		dev_dbg(&pl022->adev->dev, "polling transfer ongoing ...\n");
-		/* FIXME: insert a timeout so we don't hang here indefinitely */
-		while (pl022->tx < pl022->tx_end || pl022->rx < pl022->rx_end)
+
+		timeout = jiffies + msecs_to_jiffies(SPI_POLLING_TIMEOUT);
+		while (pl022->tx < pl022->tx_end || pl022->rx < pl022->rx_end) {
+			if (time_after(jiffies, timeout)) {
+				dev_warn(&pl022->adev->dev,
+				"%s: timeout!\n", __func__);
+				message->state = STATE_ERROR;
+				goto out;
+			}
 			readwriter(pl022);
+		}
 
 		/* Update total byte transferred */
 		message->actual_length += pl022->cur_transfer->len;
@@ -1426,7 +1432,7 @@ static void do_polling_transfer(struct pl022 *pl022)
 		/* Move to next transfer */
 		message->state = next_transfer(pl022);
 	}
-
+out:
 	/* Handle end of message */
 	if (message->state == STATE_DONE)
 		message->status = 0;
-- 
1.7.3.2

^ permalink raw reply related	[flat|nested] 8+ messages in thread
* Re: [PATCH 1/2] spi/pl022: timeout on polled transfer
  2011-05-19 12:13 ` Linus Walleij
@ 2011-05-19 12:44   ` Wolfram Sang
  -1 siblings, 0 replies; 8+ messages in thread
From: Wolfram Sang @ 2011-05-19 12:44 UTC (permalink / raw)
  To: Linus Walleij
  Cc: Linus Walleij, Magnus Templing, Grant Likely, spi-devel-general,
	Lee Jones, linux-arm-kernel


[-- Attachment #1.1: Type: text/plain, Size: 783 bytes --]

Hi,

> -		/* FIXME: insert a timeout so we don't hang here indefinitely */
> -		while (pl022->tx < pl022->tx_end || pl022->rx < pl022->rx_end)
> +
> +		timeout = jiffies + msecs_to_jiffies(SPI_POLLING_TIMEOUT);

Won't you miss the transfer if you get interrupted here longer than
SPI_POLLING_TIMEOUT?

> +		while (pl022->tx < pl022->tx_end || pl022->rx < pl022->rx_end) {
> +			if (time_after(jiffies, timeout)) {
> +				dev_warn(&pl022->adev->dev,
> +				"%s: timeout!\n", __func__);
> +				message->state = STATE_ERROR;
> +				goto out;
> +			}
>  			readwriter(pl022);
> +		}

Regards,

   Wolfram

-- 
Pengutronix e.K.                           | Wolfram Sang                |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

[-- Attachment #2: Type: text/plain, Size: 176 bytes --]

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 8+ messages in thread
* [PATCH 1/2] spi/pl022: timeout on polled transfer
@ 2011-05-19 12:44   ` Wolfram Sang
  0 siblings, 0 replies; 8+ messages in thread
From: Wolfram Sang @ 2011-05-19 12:44 UTC (permalink / raw)
  To: linux-arm-kernel

Hi,

> -		/* FIXME: insert a timeout so we don't hang here indefinitely */
> -		while (pl022->tx < pl022->tx_end || pl022->rx < pl022->rx_end)
> +
> +		timeout = jiffies + msecs_to_jiffies(SPI_POLLING_TIMEOUT);

Won't you miss the transfer if you get interrupted here longer than
SPI_POLLING_TIMEOUT?

> +		while (pl022->tx < pl022->tx_end || pl022->rx < pl022->rx_end) {
> +			if (time_after(jiffies, timeout)) {
> +				dev_warn(&pl022->adev->dev,
> +				"%s: timeout!\n", __func__);
> +				message->state = STATE_ERROR;
> +				goto out;
> +			}
>  			readwriter(pl022);
> +		}

Regards,

   Wolfram

-- 
Pengutronix e.K.                           | Wolfram Sang                |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20110519/d5c2ab71/attachment.sig>

^ permalink raw reply	[flat|nested] 8+ messages in thread
* Re: [PATCH 1/2] spi/pl022: timeout on polled transfer
  2011-05-19 12:44   ` Wolfram Sang
@ 2011-05-19 13:23     ` Linus Walleij
  -1 siblings, 0 replies; 8+ messages in thread
From: Linus Walleij @ 2011-05-19 13:23 UTC (permalink / raw)
  To: Wolfram Sang
  Cc: Linus Walleij, Magnus Templing, Grant Likely, spi-devel-general,
	Lee Jones, linux-arm-kernel

On Thu, May 19, 2011 at 2:44 PM, Wolfram Sang <w.sang@pengutronix.de> wrote:

> Hi,
>
>> -             /* FIXME: insert a timeout so we don't hang here indefinitely */
>> -             while (pl022->tx < pl022->tx_end || pl022->rx < pl022->rx_end)
>> +
>> +             timeout = jiffies + msecs_to_jiffies(SPI_POLLING_TIMEOUT);
>
> Won't you miss the transfer if you get interrupted here longer than
> SPI_POLLING_TIMEOUT?

Yeah... preempted for a second hm.

>> +             while (pl022->tx < pl022->tx_end || pl022->rx < pl022->rx_end) {
>> +                     if (time_after(jiffies, timeout)) {
>> +                             dev_warn(&pl022->adev->dev,
>> +                             "%s: timeout!\n", __func__);
>> +                             message->state = STATE_ERROR;
>> +                             goto out;
>> +                     }
>>                       readwriter(pl022);
>> +             }

What about we move readerwriter() above if (time_after...) then it atleast
gets one chance to run even if we're preempted for 10 seconds.

Any other design patterns that'd be better?

Yours,
Linus Walleij

^ permalink raw reply	[flat|nested] 8+ messages in thread
* [PATCH 1/2] spi/pl022: timeout on polled transfer
@ 2011-05-19 13:23     ` Linus Walleij
  0 siblings, 0 replies; 8+ messages in thread
From: Linus Walleij @ 2011-05-19 13:23 UTC (permalink / raw)
  To: linux-arm-kernel

On Thu, May 19, 2011 at 2:44 PM, Wolfram Sang <w.sang@pengutronix.de> wrote:

> Hi,
>
>> - ? ? ? ? ? ? /* FIXME: insert a timeout so we don't hang here indefinitely */
>> - ? ? ? ? ? ? while (pl022->tx < pl022->tx_end || pl022->rx < pl022->rx_end)
>> +
>> + ? ? ? ? ? ? timeout = jiffies + msecs_to_jiffies(SPI_POLLING_TIMEOUT);
>
> Won't you miss the transfer if you get interrupted here longer than
> SPI_POLLING_TIMEOUT?

Yeah... preempted for a second hm.

>> + ? ? ? ? ? ? while (pl022->tx < pl022->tx_end || pl022->rx < pl022->rx_end) {
>> + ? ? ? ? ? ? ? ? ? ? if (time_after(jiffies, timeout)) {
>> + ? ? ? ? ? ? ? ? ? ? ? ? ? ? dev_warn(&pl022->adev->dev,
>> + ? ? ? ? ? ? ? ? ? ? ? ? ? ? "%s: timeout!\n", __func__);
>> + ? ? ? ? ? ? ? ? ? ? ? ? ? ? message->state = STATE_ERROR;
>> + ? ? ? ? ? ? ? ? ? ? ? ? ? ? goto out;
>> + ? ? ? ? ? ? ? ? ? ? }
>> ? ? ? ? ? ? ? ? ? ? ? readwriter(pl022);
>> + ? ? ? ? ? ? }

What about we move readerwriter() above if (time_after...) then it atleast
gets one chance to run even if we're preempted for 10 seconds.

Any other design patterns that'd be better?

Yours,
Linus Walleij

^ permalink raw reply	[flat|nested] 8+ messages in thread
* Re: [PATCH 1/2] spi/pl022: timeout on polled transfer
  2011-05-19 13:23     ` Linus Walleij
@ 2011-05-19 14:23       ` Wolfram Sang
  -1 siblings, 0 replies; 8+ messages in thread
From: Wolfram Sang @ 2011-05-19 14:23 UTC (permalink / raw)
  To: Linus Walleij
  Cc: Linus Walleij, Magnus Templing, Grant Likely, spi-devel-general,
	Lee Jones, linux-arm-kernel


[-- Attachment #1.1: Type: text/plain, Size: 1396 bytes --]


> >> +             while (pl022->tx < pl022->tx_end || pl022->rx < pl022->rx_end) {
> >> +                     if (time_after(jiffies, timeout)) {
> >> +                             dev_warn(&pl022->adev->dev,
> >> +                             "%s: timeout!\n", __func__);
> >> +                             message->state = STATE_ERROR;
> >> +                             goto out;
> >> +                     }
> >>                       readwriter(pl022);
> >> +             }
> 
> What about we move readerwriter() above if (time_after...) then it atleast
> gets one chance to run even if we're preempted for 10 seconds.

This will fix the initial problem, but still exit early if e.g. the
preemption came on a subsequent iteration of the while loop, say during
readwriter().

> 
> Any other design patterns that'd be better?

A bit better might be:

	while (...) {
		time = jiffies;
		readwriter();
		if (time_after(time, timeout)...)
			...
	}

So you continue triggering transfers as long as the last one was in the
valid timespan. In case of an preemption, you made sure you tried the
whole timespan (+ the jitter) which is probably better than stop trying
too early.

Regards,

   Wolfram

-- 
Pengutronix e.K.                           | Wolfram Sang                |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

[-- Attachment #2: Type: text/plain, Size: 176 bytes --]

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 8+ messages in thread
* [PATCH 1/2] spi/pl022: timeout on polled transfer
@ 2011-05-19 14:23       ` Wolfram Sang
  0 siblings, 0 replies; 8+ messages in thread
From: Wolfram Sang @ 2011-05-19 14:23 UTC (permalink / raw)
  To: linux-arm-kernel


> >> + ? ? ? ? ? ? while (pl022->tx < pl022->tx_end || pl022->rx < pl022->rx_end) {
> >> + ? ? ? ? ? ? ? ? ? ? if (time_after(jiffies, timeout)) {
> >> + ? ? ? ? ? ? ? ? ? ? ? ? ? ? dev_warn(&pl022->adev->dev,
> >> + ? ? ? ? ? ? ? ? ? ? ? ? ? ? "%s: timeout!\n", __func__);
> >> + ? ? ? ? ? ? ? ? ? ? ? ? ? ? message->state = STATE_ERROR;
> >> + ? ? ? ? ? ? ? ? ? ? ? ? ? ? goto out;
> >> + ? ? ? ? ? ? ? ? ? ? }
> >> ? ? ? ? ? ? ? ? ? ? ? readwriter(pl022);
> >> + ? ? ? ? ? ? }
> 
> What about we move readerwriter() above if (time_after...) then it atleast
> gets one chance to run even if we're preempted for 10 seconds.

This will fix the initial problem, but still exit early if e.g. the
preemption came on a subsequent iteration of the while loop, say during
readwriter().

> 
> Any other design patterns that'd be better?

A bit better might be:

	while (...) {
		time = jiffies;
		readwriter();
		if (time_after(time, timeout)...)
			...
	}

So you continue triggering transfers as long as the last one was in the
valid timespan. In case of an preemption, you made sure you tried the
whole timespan (+ the jitter) which is probably better than stop trying
too early.

Regards,

   Wolfram

-- 
Pengutronix e.K.                           | Wolfram Sang                |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20110519/af841886/attachment.sig>

^ permalink raw reply	[flat|nested] 8+ messages in thread
end of thread, other threads:[~2011-05-19 14:23 UTC | newest]

Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-05-19 12:13 [PATCH 1/2] spi/pl022: timeout on polled transfer Linus Walleij
2011-05-19 12:13 ` Linus Walleij
2011-05-19 12:44 ` Wolfram Sang
2011-05-19 12:44   ` Wolfram Sang
2011-05-19 13:23   ` Linus Walleij
2011-05-19 13:23     ` Linus Walleij
2011-05-19 14:23     ` Wolfram Sang
2011-05-19 14:23       ` Wolfram Sang

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.