[PATCH] cpufreq: cpufreq-cpu0: clk_round_rate() can return a zero upon error

[PATCH] cpufreq: cpufreq-cpu0: clk_round_rate() can return a zero upon error

[Paul Walmsley]

Treat both negative and zero return values from clk_round_rate()
as errors.  This is needed since subsequent patches will convert
clk_round_rate()'s return value to be an unsigned type, rather
than a signed type, since some clock sources can generate rates
higher than (2^31)-1 Hz.

Eventually, when calling clk_round_rate(), only a return value of
zero will be considered a error.  All other values will be
considered valid rates.  The comparison against values less than
0 is kept to preserve the correct behavior in the meantime.

Signed-off-by: Paul Walmsley <paul@pwsan.com>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: Mike Turquette <mturquette@linaro.org>
---
Applies on v3.13-rc1.  See also:

http://marc.info/?l=linux-arm-kernel&m=138542591313620&w=2

  drivers/cpufreq/cpufreq-cpu0.c |    2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/cpufreq/cpufreq-cpu0.c b/drivers/cpufreq/cpufreq-cpu0.c
index d4585ce2346c..0faf756f6197 100644
--- a/drivers/cpufreq/cpufreq-cpu0.c
+++ b/drivers/cpufreq/cpufreq-cpu0.c
@@ -44,7 +44,7 @@ static int cpu0_set_target(struct cpufreq_policy *policy, unsigned int index)
  	int ret;

  	freq_Hz = clk_round_rate(cpu_clk, freq_table[index].frequency * 1000);
-	if (freq_Hz < 0)
+	if (freq_Hz <= 0)
  		freq_Hz = freq_table[index].frequency * 1000;

  	freq_exact = freq_Hz;

Re: [PATCH] cpufreq: cpufreq-cpu0: clk_round_rate() can return a zero upon error

[viresh kumar]

On Tuesday 26 November 2013 07:31 AM, Paul Walmsley wrote:
> 
> Treat both negative and zero return values from clk_round_rate()
> as errors.  This is needed since subsequent patches will convert
> clk_round_rate()'s return value to be an unsigned type, rather
> than a signed type, since some clock sources can generate rates
> higher than (2^31)-1 Hz.
> 
> Eventually, when calling clk_round_rate(), only a return value of
> zero will be considered a error.  All other values will be
> considered valid rates.  The comparison against values less than
> 0 is kept to preserve the correct behavior in the meantime.
> 
> Signed-off-by: Paul Walmsley <paul@pwsan.com>
> Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> Cc: Viresh Kumar <viresh.kumar@linaro.org>
> Cc: Mike Turquette <mturquette@linaro.org>
> ---
> Applies on v3.13-rc1.  See also:
> 
> http://marc.info/?l=linux-arm-kernel&m=138542591313620&w=2
> 
>  drivers/cpufreq/cpufreq-cpu0.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/cpufreq/cpufreq-cpu0.c b/drivers/cpufreq/cpufreq-cpu0.c
> index d4585ce2346c..0faf756f6197 100644
> --- a/drivers/cpufreq/cpufreq-cpu0.c
> +++ b/drivers/cpufreq/cpufreq-cpu0.c
> @@ -44,7 +44,7 @@ static int cpu0_set_target(struct cpufreq_policy *policy,
> unsigned int index)
>      int ret;
> 
>      freq_Hz = clk_round_rate(cpu_clk, freq_table[index].frequency * 1000);
> -    if (freq_Hz < 0)
> +    if (freq_Hz <= 0)
>          freq_Hz = freq_table[index].frequency * 1000;
> 
>      freq_exact = freq_Hz;

So, we will see another patch where you will do: s/<=/== ??
I am wondering if there is any other way we can get this solved, i.e. in a
single patchset.

Otherwise, for both SPEAr and cpu0 patches:

Acked-by: Viresh Kumar <viresh.kumar@linaro.org>

Re: [PATCH] cpufreq: cpufreq-cpu0: clk_round_rate() can return a zero upon error

[Paul Walmsley]

On 11/25/2013 09:03 PM, viresh kumar wrote:
> On Tuesday 26 November 2013 07:31 AM, Paul Walmsley wrote:
>> diff --git a/drivers/cpufreq/cpufreq-cpu0.c b/drivers/cpufreq/cpufreq-cpu0.c
>> index d4585ce2346c..0faf756f6197 100644
>> --- a/drivers/cpufreq/cpufreq-cpu0.c
>> +++ b/drivers/cpufreq/cpufreq-cpu0.c
>> @@ -44,7 +44,7 @@ static int cpu0_set_target(struct cpufreq_policy *policy,
>> unsigned int index)
>>       int ret;
>>
>>       freq_Hz = clk_round_rate(cpu_clk, freq_table[index].frequency * 1000);
>> -    if (freq_Hz < 0)
>> +    if (freq_Hz <= 0)
>>           freq_Hz = freq_table[index].frequency * 1000;
>>
>>       freq_exact = freq_Hz;
> So, we will see another patch where you will do: s/<=/== ??

Probably so for this driver - along with converting the type of freq_Hz 
to be u64 or unsigned long.  Not sure yet about all of the other 
drivers, since many of them are unlikely to see rates above (2^31)-1 Hz.

> I am wondering if there is any other way we can get this solved, i.e. in a
> single patchset.

I'm trying to avoid sending up a large series that touches drivers all 
over the tree :-(

> Otherwise, for both SPEAr and cpu0 patches:
>
> Acked-by: Viresh Kumar <viresh.kumar@linaro.org>

Thanks!  But I was instead hoping you might queue them up for merging 
for v3.14?  That should greatly reduce the risk of merge conflicts.

- Paul

Re: [PATCH] cpufreq: cpufreq-cpu0: clk_round_rate() can return a zero upon error

[Rafael J. Wysocki]

On Tuesday, November 26, 2013 03:50:05 PM Paul Walmsley wrote:
> On 11/25/2013 09:03 PM, viresh kumar wrote:
> > On Tuesday 26 November 2013 07:31 AM, Paul Walmsley wrote:
> >> diff --git a/drivers/cpufreq/cpufreq-cpu0.c b/drivers/cpufreq/cpufreq-cpu0.c
> >> index d4585ce2346c..0faf756f6197 100644
> >> --- a/drivers/cpufreq/cpufreq-cpu0.c
> >> +++ b/drivers/cpufreq/cpufreq-cpu0.c
> >> @@ -44,7 +44,7 @@ static int cpu0_set_target(struct cpufreq_policy *policy,
> >> unsigned int index)
> >>       int ret;
> >>
> >>       freq_Hz = clk_round_rate(cpu_clk, freq_table[index].frequency * 1000);
> >> -    if (freq_Hz < 0)
> >> +    if (freq_Hz <= 0)
> >>           freq_Hz = freq_table[index].frequency * 1000;
> >>
> >>       freq_exact = freq_Hz;
> > So, we will see another patch where you will do: s/<=/== ??
> 
> Probably so for this driver - along with converting the type of freq_Hz 
> to be u64 or unsigned long.  Not sure yet about all of the other 
> drivers, since many of them are unlikely to see rates above (2^31)-1 Hz.
> 
> > I am wondering if there is any other way we can get this solved, i.e. in a
> > single patchset.
> 
> I'm trying to avoid sending up a large series that touches drivers all 
> over the tree :-(
> 
> > Otherwise, for both SPEAr and cpu0 patches:
> >
> > Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
> 
> Thanks!  But I was instead hoping you might queue them up for merging 
> for v3.14?  That should greatly reduce the risk of merge conflicts.

I have a plan to queue them up for 3.14. :-)

Thanks!

-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.