[PATCH 1/1] Char: ds1286, eliminate busy waiting

[PATCH 1/1] Char: ds1286, eliminate busy waiting

[Jiri Slaby]

ds1286_get_time(); is not called from atomic context, sleep for 20 ms is
better choice than a (home-made) busy waiting for such a situation.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
---
 drivers/char/ds1286.c |    4 +---
 1 files changed, 1 insertions(+), 3 deletions(-)

diff --git a/drivers/char/ds1286.c b/drivers/char/ds1286.c
index fb58493..5329d48 100644
--- a/drivers/char/ds1286.c
+++ b/drivers/char/ds1286.c
@@ -443,7 +443,6 @@ static void ds1286_get_time(struct rtc_time *rtc_tm)
 {
 	unsigned char save_control;
 	unsigned long flags;
-	unsigned long uip_watchdog = jiffies;
 
 	/*
 	 * read RTC once any update in progress is done. The update
@@ -456,8 +455,7 @@ static void ds1286_get_time(struct rtc_time *rtc_tm)
 	 */
 
 	if (ds1286_is_updating() != 0)
-		while (time_before(jiffies, uip_watchdog + 2*HZ/100))
-			barrier();
+		msleep(20);
 
 	/*
 	 * Only the values that we read from the RTC are set. We leave
-- 
1.5.6.2

Re: [PATCH 1/1] Char: ds1286, eliminate busy waiting

[Ralf Baechle]

On Fri, Jul 25, 2008 at 07:49:58PM +0200, Jiri Slaby wrote:

> ds1286_get_time(); is not called from atomic context, sleep for 20 ms is
> better choice than a (home-made) busy waiting for such a situation.
> 
> Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
> Cc: Ralf Baechle <ralf@linux-mips.org>

Looks ok to me I guess.  Though I don't really think it matters ...

Acked-by: Ralf Baechle <ralf@linux-mips.org>

The same condition also appears in drivers/char/rtc.c and maybe a few
others.  Rtc.c has been copies and modified several times.

  Ralf

Re: [PATCH 1/1] Char: ds1286, eliminate busy waiting

[Jiri Slaby]

Ralf Baechle napsal(a):
> On Fri, Jul 25, 2008 at 07:49:58PM +0200, Jiri Slaby wrote:
> 
>> ds1286_get_time(); is not called from atomic context, sleep for 20 ms is
>> better choice than a (home-made) busy waiting for such a situation.

> Looks ok to me I guess.  Though I don't really think it matters ...

I think RT people has different opinion ;).

> The same condition also appears in drivers/char/rtc.c and maybe a few
> others.  Rtc.c has been copies and modified several times.

In rtc.c there is something completely different:
         while (rtc_is_updating() != 0 &&
                time_before(jiffies, uip_watchdog + 2*HZ/100))
                 cpu_relax();

It's a conditional busy-waiting. It reads the rtc status after each busy 
cycle while in the ds1286 there was
         if (ds1286_is_updating() != 0)
                 while (time_before(jiffies, uip_watchdog + 2*HZ/100))
                         barrier();