linux-ide.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH 2/2] drivers: libata-core: Use usleep_range() instead of msleep() for short sleeps (<20 ms)
@ 2016-01-08  5:18 Daniel Walker
  2016-01-08 15:36 ` Tejun Heo
  0 siblings, 1 reply; 2+ messages in thread
From: Daniel Walker @ 2016-01-08  5:18 UTC (permalink / raw)
  To: Tejun Heo
  Cc: Anil Veliyankara Madam, xe-kernel, Shikha Jain, linux-ide,
	linux-kernel

From: Anil Veliyankara Madam <aveliyan@cisco.com>

Since msleep() may sleep longer than intended time for values less
than 20ms, this patch allows the use of usleep_range for waits less
that 20ms. usleep_range is a finer precision implementation of
msleep and is designed to be a drop-in replacement for udelay
where a precise sleep/busy-wait is unnecessary.

More details can be found at http://lkml.org/lkml/2007/8/3/250
and in Documentation/timers/timers-howto.txt.

This change has been done to improve the performace in PIO6 mode
which is used by viking flash.

Cc: xe-kernel@external.cisco.com
Signed-off-by: Anil Veliyankara Madam <aveliyan@cisco.com>
Signed-off-by: Shikha Jain <shikjain@cisco.com>
---
 drivers/ata/libata-core.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/drivers/ata/libata-core.c b/drivers/ata/libata-core.c
index b79cb10..3abf0ea 100644
--- a/drivers/ata/libata-core.c
+++ b/drivers/ata/libata-core.c
@@ -50,6 +50,7 @@
 #include <linux/blkdev.h>
 #include <linux/delay.h>
 #include <linux/timer.h>
+#include <linux/time.h>
 #include <linux/interrupt.h>
 #include <linux/completion.h>
 #include <linux/suspend.h>
@@ -6697,7 +6698,12 @@ void ata_msleep(struct ata_port *ap, unsigned int msecs)
 	if (owns_eh)
 		ata_eh_release(ap);
 
-	msleep(msecs);
+	if (msecs < 20) {
+		unsigned long usecs = msecs * USEC_PER_MSEC;
+		usleep_range(usecs, usecs + 50);
+	} else {
+		msleep(msecs);
+	}
 
 	if (owns_eh)
 		ata_eh_acquire(ap);
-- 
2.1.4

^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: [PATCH 2/2] drivers: libata-core: Use usleep_range() instead of msleep() for short sleeps (<20 ms)
  2016-01-08  5:18 [PATCH 2/2] drivers: libata-core: Use usleep_range() instead of msleep() for short sleeps (<20 ms) Daniel Walker
@ 2016-01-08 15:36 ` Tejun Heo
  0 siblings, 0 replies; 2+ messages in thread
From: Tejun Heo @ 2016-01-08 15:36 UTC (permalink / raw)
  To: Daniel Walker
  Cc: Anil Veliyankara Madam, xe-kernel, Shikha Jain, linux-ide,
	linux-kernel

On Thu, Jan 07, 2016 at 09:18:52PM -0800, Daniel Walker wrote:
> From: Anil Veliyankara Madam <aveliyan@cisco.com>
> 
> Since msleep() may sleep longer than intended time for values less
> than 20ms, this patch allows the use of usleep_range for waits less
> that 20ms. usleep_range is a finer precision implementation of
> msleep and is designed to be a drop-in replacement for udelay
> where a precise sleep/busy-wait is unnecessary.
> 
> More details can be found at http://lkml.org/lkml/2007/8/3/250
> and in Documentation/timers/timers-howto.txt.
> 
> This change has been done to improve the performace in PIO6 mode
> which is used by viking flash.
> 
> Cc: xe-kernel@external.cisco.com
> Signed-off-by: Anil Veliyankara Madam <aveliyan@cisco.com>
> Signed-off-by: Shikha Jain <shikjain@cisco.com>

Applied to libata/for-4.5.

Thanks.

-- 
tejun

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2016-01-08 15:36 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-01-08  5:18 [PATCH 2/2] drivers: libata-core: Use usleep_range() instead of msleep() for short sleeps (<20 ms) Daniel Walker
2016-01-08 15:36 ` Tejun Heo

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).