[PATCH v2 0/2] fixes and improvement for sprd hwspinlock

[PATCH v2 0/2] fixes and improvement for sprd hwspinlock

[Chunyan Zhang]

From: Chunyan Zhang <chunyan.zhang@unisoc.com>

Changes since v1:
* Change to add __maybe_unsed rather than not use of_match_ptr().

Chunyan Zhang (2):
  hwspinlock: sprd: fixed warning of unused variable
    'sprd_hwspinlock_of_match'
  hwspinlock: sprd: use module_platform_driver() instead postcore
    initcall

 drivers/hwspinlock/sprd_hwspinlock.c | 15 ++-------------
 1 file changed, 2 insertions(+), 13 deletions(-)

-- 
2.20.1

[PATCH v2 1/2] hwspinlock: sprd: fixed warning of unused variable 'sprd_hwspinlock_of_match'

[Chunyan Zhang]

From: Chunyan Zhang <chunyan.zhang@unisoc.com>

The macro function of_match_ptr() is NULL if CONFIG_OF is not set, then
Clang compiler would complain the of_device_id variable is unused.

But using of_match_ptr() is space saving, for this case, the unused structure
'sprd_hwspinlock_of_match' would be not built into symbol table if CONFIG_OF
is not set, probably depends on the compiler though.

So adding __maybe_unsed seems a good approach to fix this warning.

Reported-by: kernel test robot <lkp@intel.com>
Fixes: d8c8bbbb1aba ("hwspinlock: sprd: Add hardware spinlock driver")
Signed-off-by: Chunyan Zhang <chunyan.zhang@unisoc.com>
---
 drivers/hwspinlock/sprd_hwspinlock.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/hwspinlock/sprd_hwspinlock.c b/drivers/hwspinlock/sprd_hwspinlock.c
index 36dc8038bbb4..4c63e2546064 100644
--- a/drivers/hwspinlock/sprd_hwspinlock.c
+++ b/drivers/hwspinlock/sprd_hwspinlock.c
@@ -138,7 +138,7 @@ static int sprd_hwspinlock_probe(struct platform_device *pdev)
 					 SPRD_HWLOCKS_NUM);
 }
 
-static const struct of_device_id sprd_hwspinlock_of_match[] = {
+static const __maybe_unused struct of_device_id sprd_hwspinlock_of_match[] = {
 	{ .compatible = "sprd,hwspinlock-r3p0", },
 	{ /* sentinel */ }
 };
-- 
2.20.1

[PATCH v2 2/2] hwspinlock: sprd: use module_platform_driver() instead postcore initcall

[Chunyan Zhang]

From: Chunyan Zhang <chunyan.zhang@unisoc.com>

The hardware spinlock devices are defined in the DT, there's no need for
init calls order, remove boilerplate code by using module_platform_driver.

Signed-off-by: Chunyan Zhang <chunyan.zhang@unisoc.com>
---
 drivers/hwspinlock/sprd_hwspinlock.c | 13 +------------
 1 file changed, 1 insertion(+), 12 deletions(-)

diff --git a/drivers/hwspinlock/sprd_hwspinlock.c b/drivers/hwspinlock/sprd_hwspinlock.c
index 4c63e2546064..19d1924044e5 100644
--- a/drivers/hwspinlock/sprd_hwspinlock.c
+++ b/drivers/hwspinlock/sprd_hwspinlock.c
@@ -151,18 +151,7 @@ static struct platform_driver sprd_hwspinlock_driver = {
 		.of_match_table = of_match_ptr(sprd_hwspinlock_of_match),
 	},
 };
-
-static int __init sprd_hwspinlock_init(void)
-{
-	return platform_driver_register(&sprd_hwspinlock_driver);
-}
-postcore_initcall(sprd_hwspinlock_init);
-
-static void __exit sprd_hwspinlock_exit(void)
-{
-	platform_driver_unregister(&sprd_hwspinlock_driver);
-}
-module_exit(sprd_hwspinlock_exit);
+module_platform_driver(sprd_hwspinlock_driver);
 
 MODULE_LICENSE("GPL v2");
 MODULE_DESCRIPTION("Hardware spinlock driver for Spreadtrum");
-- 
2.20.1

Re: [PATCH v2 1/2] hwspinlock: sprd: fixed warning of unused variable 'sprd_hwspinlock_of_match'

[Bjorn Andersson]

On Mon 02 Nov 17:34 CST 2020, Baolin Wang wrote:

> On Friday, October 30, 2020, Chunyan Zhang <zhang.lyra@gmail.com> wrote:
> > From: Chunyan Zhang <chunyan.zhang@unisoc.com>
> >
> > The macro function of_match_ptr() is NULL if CONFIG_OF is not set, then
> > Clang compiler would complain the of_device_id variable is unused.
> >
> > But using of_match_ptr() is space saving, for this case, the unused
> structure
> > 'sprd_hwspinlock_of_match' would be not built into symbol table if
> CONFIG_OF
> > is not set, probably depends on the compiler though.
> >
> > So adding __maybe_unsed seems a good approach to fix this warning.
> >
> > Reported-by: kernel test robot <lkp@intel.com>
> > Fixes: d8c8bbbb1aba ("hwspinlock: sprd: Add hardware spinlock driver")
> > Signed-off-by: Chunyan Zhang <chunyan.zhang@unisoc.com>
> > ---
> 
> I'd like to remove this wrapper, and just depend on the CONFIG_OF. But I
> have no objection for this patch.  So
> Reviewed-by: Baolin Wang <baolin.wang7@gmail.com>
> 

As noted on the first line of the probe function, this driver isn't
going to do anything when CONFIG_OF is disabled - and I don't think we
should worry too much about space savings during COMPILE_TEST.

So I would prefer that we simply drop the of_match_ptr()

But I believe that's what you're saying as well Baolin?

Regards,
Bjorn

> 
> >  drivers/hwspinlock/sprd_hwspinlock.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/hwspinlock/sprd_hwspinlock.c
> b/drivers/hwspinlock/sprd_hwspinlock.c
> > index 36dc8038bbb4..4c63e2546064 100644
> > --- a/drivers/hwspinlock/sprd_hwspinlock.c
> > +++ b/drivers/hwspinlock/sprd_hwspinlock.c
> > @@ -138,7 +138,7 @@ static int sprd_hwspinlock_probe(struct
> platform_device *pdev)
> >                                          SPRD_HWLOCKS_NUM);
> >  }
> >
> > -static const struct of_device_id sprd_hwspinlock_of_match[] = {
> > +static const __maybe_unused struct of_device_id
> sprd_hwspinlock_of_match[] = {
> >         { .compatible = "sprd,hwspinlock-r3p0", },
> >         { /* sentinel */ }
> >  };
> > --
> > 2.20.1
> >
> >
> 
> -- 
> Baolin Wang

Re: [PATCH v2 2/2] hwspinlock: sprd: use module_platform_driver() instead postcore initcall

[Bjorn Andersson]

On Thu 29 Oct 22:46 CDT 2020, Chunyan Zhang wrote:

> From: Chunyan Zhang <chunyan.zhang@unisoc.com>
> 
> The hardware spinlock devices are defined in the DT, there's no need for
> init calls order, remove boilerplate code by using module_platform_driver.
> 
> Signed-off-by: Chunyan Zhang <chunyan.zhang@unisoc.com>

Applied this patch with Baolin's r-b.

Thank you,
Bjorn

> ---
>  drivers/hwspinlock/sprd_hwspinlock.c | 13 +------------
>  1 file changed, 1 insertion(+), 12 deletions(-)
> 
> diff --git a/drivers/hwspinlock/sprd_hwspinlock.c b/drivers/hwspinlock/sprd_hwspinlock.c
> index 4c63e2546064..19d1924044e5 100644
> --- a/drivers/hwspinlock/sprd_hwspinlock.c
> +++ b/drivers/hwspinlock/sprd_hwspinlock.c
> @@ -151,18 +151,7 @@ static struct platform_driver sprd_hwspinlock_driver = {
>  		.of_match_table = of_match_ptr(sprd_hwspinlock_of_match),
>  	},
>  };
> -
> -static int __init sprd_hwspinlock_init(void)
> -{
> -	return platform_driver_register(&sprd_hwspinlock_driver);
> -}
> -postcore_initcall(sprd_hwspinlock_init);
> -
> -static void __exit sprd_hwspinlock_exit(void)
> -{
> -	platform_driver_unregister(&sprd_hwspinlock_driver);
> -}
> -module_exit(sprd_hwspinlock_exit);
> +module_platform_driver(sprd_hwspinlock_driver);
>  
>  MODULE_LICENSE("GPL v2");
>  MODULE_DESCRIPTION("Hardware spinlock driver for Spreadtrum");
> -- 
> 2.20.1
> 

Re: [PATCH v2 1/2] hwspinlock: sprd: fixed warning of unused variable 'sprd_hwspinlock_of_match'

[Chunyan Zhang]

On Tue, 3 Nov 2020 at 07:58, Bjorn Andersson <bjorn.andersson@linaro.org> wrote:
>
> On Mon 02 Nov 17:34 CST 2020, Baolin Wang wrote:
>
> > On Friday, October 30, 2020, Chunyan Zhang <zhang.lyra@gmail.com> wrote:
> > > From: Chunyan Zhang <chunyan.zhang@unisoc.com>
> > >
> > > The macro function of_match_ptr() is NULL if CONFIG_OF is not set, then
> > > Clang compiler would complain the of_device_id variable is unused.
> > >
> > > But using of_match_ptr() is space saving, for this case, the unused
> > structure
> > > 'sprd_hwspinlock_of_match' would be not built into symbol table if
> > CONFIG_OF
> > > is not set, probably depends on the compiler though.
> > >
> > > So adding __maybe_unsed seems a good approach to fix this warning.
> > >
> > > Reported-by: kernel test robot <lkp@intel.com>
> > > Fixes: d8c8bbbb1aba ("hwspinlock: sprd: Add hardware spinlock driver")
> > > Signed-off-by: Chunyan Zhang <chunyan.zhang@unisoc.com>
> > > ---
> >
> > I'd like to remove this wrapper, and just depend on the CONFIG_OF. But I
> > have no objection for this patch.  So
> > Reviewed-by: Baolin Wang <baolin.wang7@gmail.com>
> >
>
> As noted on the first line of the probe function, this driver isn't
> going to do anything when CONFIG_OF is disabled - and I don't think we
> should worry too much about space savings during COMPILE_TEST.
>
> So I would prefer that we simply drop the of_match_ptr()

Ok, that's what the v1 does.
Bjorn, could you please pick up the 1st patch in v1 [1]?

Thanks,
Chunyan

[1] https://lkml.org/lkml/2020/10/26/87


>
> But I believe that's what you're saying as well Baolin?
>
> Regards,
> Bjorn
>
> >
> > >  drivers/hwspinlock/sprd_hwspinlock.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/hwspinlock/sprd_hwspinlock.c
> > b/drivers/hwspinlock/sprd_hwspinlock.c
> > > index 36dc8038bbb4..4c63e2546064 100644
> > > --- a/drivers/hwspinlock/sprd_hwspinlock.c
> > > +++ b/drivers/hwspinlock/sprd_hwspinlock.c
> > > @@ -138,7 +138,7 @@ static int sprd_hwspinlock_probe(struct
> > platform_device *pdev)
> > >                                          SPRD_HWLOCKS_NUM);
> > >  }
> > >
> > > -static const struct of_device_id sprd_hwspinlock_of_match[] = {
> > > +static const __maybe_unused struct of_device_id
> > sprd_hwspinlock_of_match[] = {
> > >         { .compatible = "sprd,hwspinlock-r3p0", },
> > >         { /* sentinel */ }
> > >  };
> > > --
> > > 2.20.1
> > >
> > >
> >
> > --
> > Baolin Wang