[PATCH 1/2] cpufreq: exynos: Fix unsigned variable being checked for negative value

[PATCH 1/2] cpufreq: exynos: Fix unsigned variable being checked for negative value

[Sachin Kamat]

exynos_cpufreq_scale function returns signed value which was
assigned to an unsigned variable and checked for negative value which
is always false. Hence make it signed.

Fixes the following smatch warnings:
drivers/cpufreq/exynos-cpufreq.c:83 exynos_cpufreq_scale() warn:
unsigned 'old_index' is never less than zero.
drivers/cpufreq/exynos-cpufreq.c:89 exynos_cpufreq_scale() warn:
unsigned 'index' is never less than zero.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
---
Rafael J. Wysocki suggested these 2 patches be taken through the
Samsung tree.
---
 drivers/cpufreq/exynos-cpufreq.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/cpufreq/exynos-cpufreq.c b/drivers/cpufreq/exynos-cpufreq.c
index 218b3ce..88401ba 100644
--- a/drivers/cpufreq/exynos-cpufreq.c
+++ b/drivers/cpufreq/exynos-cpufreq.c
@@ -65,7 +65,7 @@ static int exynos_cpufreq_scale(unsigned int target_freq)
 	struct cpufreq_policy *policy = cpufreq_cpu_get(0);
 	unsigned int arm_volt, safe_arm_volt = 0;
 	unsigned int mpll_freq_khz = exynos_info->mpll_freq_khz;
-	unsigned int index, old_index;
+	int index, old_index;
 	int ret = 0;
 
 	freqs.old = policy->cur;
-- 
1.7.4.1

[PATCH 2/2] cpufreq: exynos: Initialise return variable

[Sachin Kamat]

'ret' is undefined when the function returns from the first
'if' condition. Without this patch we get the following warning:

drivers/cpufreq/exynos-cpufreq.c: In function ‘exynos_target’:
drivers/cpufreq/exynos-cpufreq.c:182:2: warning:
‘ret’ may be used uninitialized in this function [-Wuninitialized]

Suggested-by: Jonghwan Choi <jhbird.choi@samsung.com>
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
---
 drivers/cpufreq/exynos-cpufreq.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/cpufreq/exynos-cpufreq.c b/drivers/cpufreq/exynos-cpufreq.c
index 88401ba..4268e46 100644
--- a/drivers/cpufreq/exynos-cpufreq.c
+++ b/drivers/cpufreq/exynos-cpufreq.c
@@ -159,7 +159,7 @@ static int exynos_target(struct cpufreq_policy *policy,
 {
 	struct cpufreq_frequency_table *freq_table = exynos_info->freq_table;
 	unsigned int index;
-	int ret;
+	int ret = 0;
 
 	mutex_lock(&cpufreq_lock);
 
-- 
1.7.4.1

RE: [PATCH 1/2] cpufreq: exynos: Fix unsigned variable being checked for negative value

[Kukjin Kim]

Sachin Kamat wrote:
> 
> exynos_cpufreq_scale function returns signed value which was
> assigned to an unsigned variable and checked for negative value which
> is always false. Hence make it signed.
> 
> Fixes the following smatch warnings:
> drivers/cpufreq/exynos-cpufreq.c:83 exynos_cpufreq_scale() warn:
> unsigned 'old_index' is never less than zero.
> drivers/cpufreq/exynos-cpufreq.c:89 exynos_cpufreq_scale() warn:
> unsigned 'index' is never less than zero.
> 
> Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
> ---
> Rafael J. Wysocki suggested these 2 patches be taken through the
> Samsung tree.
> ---
>  drivers/cpufreq/exynos-cpufreq.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
Looks OK to me, applied.

Thanks.

- Kukjin

RE: [PATCH 2/2] cpufreq: exynos: Initialise return variable

[Kukjin Kim]

Sachin Kamat wrote:
> 
> 'ret' is undefined when the function returns from the first
> 'if' condition. Without this patch we get the following warning:
> 
> drivers/cpufreq/exynos-cpufreq.c: In function ‘exynos_target’:
> drivers/cpufreq/exynos-cpufreq.c:182:2: warning:
> ‘ret’ may be used uninitialized in this function [-Wuninitialized]
> 
> Suggested-by: Jonghwan Choi <jhbird.choi@samsung.com>

I modified above to 'Reported-by' because checkpatch complains like following...

WARNING: Non-standard signature: Suggested-by:
#10:
Suggested-by: Jonghwan Choi <jhbird.choi@samsung.com>

> Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
> ---
>  drivers/cpufreq/exynos-cpufreq.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
Applied, thanks.

- Kukjin

Re: [PATCH 2/2] cpufreq: exynos: Initialise return variable

[Rafael J. Wysocki]

On Friday, January 25, 2013 10:26:41 AM Kukjin Kim wrote:
> Sachin Kamat wrote:
> > 
> > 'ret' is undefined when the function returns from the first
> > 'if' condition. Without this patch we get the following warning:
> > 
> > drivers/cpufreq/exynos-cpufreq.c: In function ‘exynos_target’:
> > drivers/cpufreq/exynos-cpufreq.c:182:2: warning:
> > ‘ret’ may be used uninitialized in this function [-Wuninitialized]
> > 
> > Suggested-by: Jonghwan Choi <jhbird.choi@samsung.com>
> 
> I modified above to 'Reported-by' because checkpatch complains like following...
> 
> WARNING: Non-standard signature: Suggested-by:
> #10:
> Suggested-by: Jonghwan Choi <jhbird.choi@samsung.com>

This tag has been used already, though, and I think it makes sense (more than
Reported-by in this case).  I think it may be regarded as standard regardless
of what checkpatch has to say about that. :-)

Thanks,
Rafael


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

RE: [PATCH 2/2] cpufreq: exynos: Initialise return variable

[Kukjin Kim]

Rafael J. Wysocki wrote:
> 
> On Friday, January 25, 2013 10:26:41 AM Kukjin Kim wrote:
> > Sachin Kamat wrote:
> > >
> > > 'ret' is undefined when the function returns from the first
> > > 'if' condition. Without this patch we get the following warning:
> > >
> > > drivers/cpufreq/exynos-cpufreq.c: In function ‘exynos_target’:
> > > drivers/cpufreq/exynos-cpufreq.c:182:2: warning:
> > > ‘ret’ may be used uninitialized in this function [-Wuninitialized]
> > >
> > > Suggested-by: Jonghwan Choi <jhbird.choi@samsung.com>
> >
> > I modified above to 'Reported-by' because checkpatch complains like
> following...
> >
> > WARNING: Non-standard signature: Suggested-by:
> > #10:
> > Suggested-by: Jonghwan Choi <jhbird.choi@samsung.com>
> 
> This tag has been used already, though, and I think it makes sense (more
> than  

Yes, agreed.

> Reported-by in this case).  I think it may be regarded as standard regardless
> of what checkpatch has to say about that. :-)
> 
I see. Let me keep the 'Reported-by' here.
Thanks for your opinion.

- Kukjin