linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] i2c: octeon: Make the timeout 1*HZ instead of 1 jiffy
@ 2011-01-24 12:05 Bernhard Walle
  2011-01-24 12:14 ` Jean Delvare
  0 siblings, 1 reply; 6+ messages in thread
From: Bernhard Walle @ 2011-01-24 12:05 UTC (permalink / raw)
  To: ben-linux; +Cc: ddaney, ralf, linux-i2c, linux-kernel

Since the timeout in the struct i2c_adapter use jiffies as unit and a
timeout should be independent of CONFIG_HZ and most drivers that specify
a timeout here use just HZ, change the i2c-octeon driver to use 1*HZ
(which is a second), too.

Signed-off-by: Bernhard Walle <walle@corscience.de>
---
 drivers/i2c/busses/i2c-octeon.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/i2c/busses/i2c-octeon.c b/drivers/i2c/busses/i2c-octeon.c
index 0fe1102..537a4ee 100644
--- a/drivers/i2c/busses/i2c-octeon.c
+++ b/drivers/i2c/busses/i2c-octeon.c
@@ -460,7 +460,7 @@ static struct i2c_adapter octeon_i2c_ops = {
 	.owner = THIS_MODULE,
 	.name = "OCTEON adapter",
 	.algo = &octeon_i2c_algo,
-	.timeout = 2,
+	.timeout = HZ,
 };
 
 /**
-- 
1.7.1


^ permalink raw reply related	[flat|nested] 6+ messages in thread

* Re: [PATCH] i2c: octeon: Make the timeout 1*HZ instead of 1 jiffy
  2011-01-24 12:05 [PATCH] i2c: octeon: Make the timeout 1*HZ instead of 1 jiffy Bernhard Walle
@ 2011-01-24 12:14 ` Jean Delvare
  2011-01-24 12:37   ` Bernhard Walle
  0 siblings, 1 reply; 6+ messages in thread
From: Jean Delvare @ 2011-01-24 12:14 UTC (permalink / raw)
  To: Bernhard Walle; +Cc: ben-linux, ddaney, ralf, linux-i2c, linux-kernel

Hi Bernhard,

On Mon, 24 Jan 2011 13:05:41 +0100, Bernhard Walle wrote:
> Since the timeout in the struct i2c_adapter use jiffies as unit and a
> timeout should be independent of CONFIG_HZ and most drivers that specify
> a timeout here use just HZ, change the i2c-octeon driver to use 1*HZ
> (which is a second), too.
> 
> Signed-off-by: Bernhard Walle <walle@corscience.de>
> ---
>  drivers/i2c/busses/i2c-octeon.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/i2c/busses/i2c-octeon.c b/drivers/i2c/busses/i2c-octeon.c
> index 0fe1102..537a4ee 100644
> --- a/drivers/i2c/busses/i2c-octeon.c
> +++ b/drivers/i2c/busses/i2c-octeon.c
> @@ -460,7 +460,7 @@ static struct i2c_adapter octeon_i2c_ops = {
>  	.owner = THIS_MODULE,
>  	.name = "OCTEON adapter",
>  	.algo = &octeon_i2c_algo,
> -	.timeout = 2,
> +	.timeout = HZ,
>  };
>  
>  /**

This is a large timeout value, much larger than any value the
i2c-octeon driver was using today for any possible value of HZ.
Wouldn't it make more sense to set it to HZ / 50, which is 2 for HZ =
100, leaving the setting unchanged for this common case at least?

-- 
Jean Delvare

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] i2c: octeon: Make the timeout 1*HZ instead of 1 jiffy
  2011-01-24 12:14 ` Jean Delvare
@ 2011-01-24 12:37   ` Bernhard Walle
  2011-01-24 12:40     ` [PATCH] i2c: octeon: Make the timeout independent of CONFIG_HZ Bernhard Walle
  0 siblings, 1 reply; 6+ messages in thread
From: Bernhard Walle @ 2011-01-24 12:37 UTC (permalink / raw)
  To: Jean Delvare; +Cc: ben-linux, ddaney, ralf, linux-i2c, linux-kernel

[-- Attachment #1: Type: text/plain, Size: 965 bytes --]

Hi Jean,

Am 24.01.2011 13:14, schrieb Jean Delvare:
>>  	.owner = THIS_MODULE,
>>  	.name = "OCTEON adapter",
>>  	.algo = &octeon_i2c_algo,
>> -	.timeout = 2,
>> +	.timeout = HZ,
>>  };
> 
> This is a large timeout value, much larger than any value the
> i2c-octeon driver was using today for any possible value of HZ.
> Wouldn't it make more sense to set it to HZ / 50, which is 2 for HZ =
> 100, leaving the setting unchanged for this common case at least?

To be honest, I had problems with timeout=2 that go away with
timeout=HZ. But I cannot be sure that this is a problem with my hardware
that is not present with other devices.

The drivers that use HZ are i2c-mpc.c and i2c-pca-platform.c, but I
don't know the hardware that much so that I really can compare the values.

So yes, you're right. A patch modifying the timeout value should be an
extra patch, so I will resend the current patch to use HZ/50.


Regards,
Bernhard


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 900 bytes --]

^ permalink raw reply	[flat|nested] 6+ messages in thread

* [PATCH] i2c: octeon: Make the timeout independent of CONFIG_HZ
  2011-01-24 12:37   ` Bernhard Walle
@ 2011-01-24 12:40     ` Bernhard Walle
  2011-01-24 14:00       ` Jean Delvare
  0 siblings, 1 reply; 6+ messages in thread
From: Bernhard Walle @ 2011-01-24 12:40 UTC (permalink / raw)
  To: ben-linux; +Cc: ddaney, ralf, linux-i2c, linux-kernel

Since the timeout in the struct i2c_adapter use jiffies as unit and a
timeout should be independent of CONFIG_HZ, use a value of HZ/50 which
is equal to the current value 2 for CONFIG_HZ=100.

Signed-off-by: Bernhard Walle <walle@corscience.de>
---
 drivers/i2c/busses/i2c-octeon.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/i2c/busses/i2c-octeon.c b/drivers/i2c/busses/i2c-octeon.c
index 4cffda5..d07e5c7 100644
--- a/drivers/i2c/busses/i2c-octeon.c
+++ b/drivers/i2c/busses/i2c-octeon.c
@@ -469,7 +469,7 @@ static struct i2c_adapter octeon_i2c_ops = {
 	.owner = THIS_MODULE,
 	.name = "OCTEON adapter",
 	.algo = &octeon_i2c_algo,
-	.timeout = 2,
+	.timeout = HZ/50,
 };
 
 /**
-- 
1.7.1


^ permalink raw reply related	[flat|nested] 6+ messages in thread

* Re: [PATCH] i2c: octeon: Make the timeout independent of CONFIG_HZ
  2011-01-24 12:40     ` [PATCH] i2c: octeon: Make the timeout independent of CONFIG_HZ Bernhard Walle
@ 2011-01-24 14:00       ` Jean Delvare
  2011-03-16 13:32         ` Jean Delvare
  0 siblings, 1 reply; 6+ messages in thread
From: Jean Delvare @ 2011-01-24 14:00 UTC (permalink / raw)
  To: Bernhard Walle; +Cc: ben-linux, ddaney, ralf, linux-i2c, linux-kernel

On Mon, 24 Jan 2011 13:40:26 +0100, Bernhard Walle wrote:
> Since the timeout in the struct i2c_adapter use jiffies as unit and a
> timeout should be independent of CONFIG_HZ, use a value of HZ/50 which
> is equal to the current value 2 for CONFIG_HZ=100.
> 
> Signed-off-by: Bernhard Walle <walle@corscience.de>

Acked-by: Jean Delvare <khali@linux-fr.org>

Ben, please apply.

> ---
>  drivers/i2c/busses/i2c-octeon.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/i2c/busses/i2c-octeon.c b/drivers/i2c/busses/i2c-octeon.c
> index 4cffda5..d07e5c7 100644
> --- a/drivers/i2c/busses/i2c-octeon.c
> +++ b/drivers/i2c/busses/i2c-octeon.c
> @@ -469,7 +469,7 @@ static struct i2c_adapter octeon_i2c_ops = {
>  	.owner = THIS_MODULE,
>  	.name = "OCTEON adapter",
>  	.algo = &octeon_i2c_algo,
> -	.timeout = 2,
> +	.timeout = HZ/50,
>  };
>  
>  /**


-- 
Jean Delvare

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] i2c: octeon: Make the timeout independent of CONFIG_HZ
  2011-01-24 14:00       ` Jean Delvare
@ 2011-03-16 13:32         ` Jean Delvare
  0 siblings, 0 replies; 6+ messages in thread
From: Jean Delvare @ 2011-03-16 13:32 UTC (permalink / raw)
  To: ben-linux; +Cc: Bernhard Walle, ddaney, ralf, linux-i2c, linux-kernel

On Mon, 24 Jan 2011 15:00:05 +0100, Jean Delvare wrote:
> On Mon, 24 Jan 2011 13:40:26 +0100, Bernhard Walle wrote:
> > Since the timeout in the struct i2c_adapter use jiffies as unit and a
> > timeout should be independent of CONFIG_HZ, use a value of HZ/50 which
> > is equal to the current value 2 for CONFIG_HZ=100.
> > 
> > Signed-off-by: Bernhard Walle <walle@corscience.de>
> 
> Acked-by: Jean Delvare <khali@linux-fr.org>
> 
> Ben, please apply.

Ben, I can't see this fix in your tree, it seems it got lost?

> > ---
> >  drivers/i2c/busses/i2c-octeon.c |    2 +-
> >  1 files changed, 1 insertions(+), 1 deletions(-)
> > 
> > diff --git a/drivers/i2c/busses/i2c-octeon.c b/drivers/i2c/busses/i2c-octeon.c
> > index 4cffda5..d07e5c7 100644
> > --- a/drivers/i2c/busses/i2c-octeon.c
> > +++ b/drivers/i2c/busses/i2c-octeon.c
> > @@ -469,7 +469,7 @@ static struct i2c_adapter octeon_i2c_ops = {
> >  	.owner = THIS_MODULE,
> >  	.name = "OCTEON adapter",
> >  	.algo = &octeon_i2c_algo,
> > -	.timeout = 2,
> > +	.timeout = HZ/50,
> >  };
> >  
> >  /**
> 
> 


-- 
Jean Delvare

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2011-03-16 13:33 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-01-24 12:05 [PATCH] i2c: octeon: Make the timeout 1*HZ instead of 1 jiffy Bernhard Walle
2011-01-24 12:14 ` Jean Delvare
2011-01-24 12:37   ` Bernhard Walle
2011-01-24 12:40     ` [PATCH] i2c: octeon: Make the timeout independent of CONFIG_HZ Bernhard Walle
2011-01-24 14:00       ` Jean Delvare
2011-03-16 13:32         ` Jean Delvare

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).