[PATCH] usb: musb: - kill the compile warning

[PATCH] usb: musb: - kill the compile warning

[Bryan Wu]

drivers/usb/musb/musb_core.c:1433: warning: assignment makes pointer
from integer without a cast
	hw_ep->target_regs = musb_read_target_reg_base(i, mbase);

static inline u16 musb_read_target_reg_base(u8 i, void __iomem *mbase);

this is a common bug, but a bug still ?  mbase is a 32/64 bit pointer,
but we return a u16 to assign to a pointer ?  seems odd :)

So return right pointer in the stub.

Cc: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
---
 drivers/usb/musb/musb_regs.h |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/usb/musb/musb_regs.h b/drivers/usb/musb/musb_regs.h
index de3b2f1..c2f6a16 100644
--- a/drivers/usb/musb/musb_regs.h
+++ b/drivers/usb/musb/musb_regs.h
@@ -333,7 +333,7 @@ static inline u16 musb_read_hwvers(void __iomem *mbase)
 
 static inline void __iomem *musb_read_target_reg_base(u8 i, void __iomem *mbase)
 {
-	return (MUSB_BUSCTL_OFFSET(i, 0) + mbase);
+	return (void __iomem *)(MUSB_BUSCTL_OFFSET(i, 0) + mbase);
 }
 
 static inline void musb_write_rxfunaddr(void __iomem *ep_target_regs,
@@ -473,9 +473,9 @@ static inline u16 musb_read_hwvers(void __iomem *mbase)
 	return 0;
 }
 
-static inline u16 musb_read_target_reg_base(u8 i, void __iomem *mbase)
+static inline void __iomem *musb_read_target_reg_base(u8 i, void __iomem *mbase)
 {
-	return 0;
+	return NULL;
 }
 
 static inline void musb_write_rxfunaddr(void __iomem *ep_target_regs,
-- 
1.5.6.3

Re: [PATCH] usb: musb: - kill the compile warning

[Felipe Balbi]

On Fri, Feb 06, 2009 at 06:23:30PM +0800, Bryan Wu wrote:
> drivers/usb/musb/musb_core.c:1433: warning: assignment makes pointer
> from integer without a cast
> 	hw_ep->target_regs = musb_read_target_reg_base(i, mbase);
> 
> static inline u16 musb_read_target_reg_base(u8 i, void __iomem *mbase);
> 
> this is a common bug, but a bug still ?  mbase is a 32/64 bit pointer,
> but we return a u16 to assign to a pointer ?  seems odd :)
> 
> So return right pointer in the stub.
> 
> Cc: Mike Frysinger <vapier.adi@gmail.com>
> Signed-off-by: Bryan Wu <cooloney@kernel.org>
> ---
>  drivers/usb/musb/musb_regs.h |    6 +++---
>  1 files changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/usb/musb/musb_regs.h b/drivers/usb/musb/musb_regs.h
> index de3b2f1..c2f6a16 100644
> --- a/drivers/usb/musb/musb_regs.h
> +++ b/drivers/usb/musb/musb_regs.h
> @@ -333,7 +333,7 @@ static inline u16 musb_read_hwvers(void __iomem *mbase)
>  
>  static inline void __iomem *musb_read_target_reg_base(u8 i, void __iomem *mbase)
>  {
> -	return (MUSB_BUSCTL_OFFSET(i, 0) + mbase);
> +	return (void __iomem *)(MUSB_BUSCTL_OFFSET(i, 0) + mbase);

please don't cast, we don't get this warning in non-blackfin builds
(well, omap at least)

> @@ -473,9 +473,9 @@ static inline u16 musb_read_hwvers(void __iomem *mbase)
>  	return 0;
>  }
>  
> -static inline u16 musb_read_target_reg_base(u8 i, void __iomem *mbase)
> +static inline void __iomem *musb_read_target_reg_base(u8 i, void __iomem *mbase)
>  {
> -	return 0;
> +	return NULL;

this looks fine

-- 
balbi

Re: [PATCH] usb: musb: - kill the compile warning

[David Brownell]

On Friday 06 February 2009, Bryan Wu wrote:
> -       return (MUSB_BUSCTL_OFFSET(i, 0) + mbase);
> +       return (void __iomem *)(MUSB_BUSCTL_OFFSET(i, 0) + mbase);

Why is MUSB_BUSCTL_OFFSET() returning something other than
an integer?  "mbase" is already "void __iomem *", so the only
way that should trigger a warning is if that OFFSET() macro
is doing something odd.

Re: [PATCH] usb: musb: - kill the compile warning

[Felipe Balbi]

On Thu, Feb 12, 2009 at 07:03:25AM +0100, David Brownell wrote:
> On Friday 06 February 2009, Bryan Wu wrote:
> > -       return (MUSB_BUSCTL_OFFSET(i, 0) + mbase);
> > +       return (void __iomem *)(MUSB_BUSCTL_OFFSET(i, 0) + mbase);
> 
> Why is MUSB_BUSCTL_OFFSET() returning something other than
> an integer?  "mbase" is already "void __iomem *", so the only
> way that should trigger a warning is if that OFFSET() macro
> is doing something odd.

The warning only comes on Blackfin's version of this, this hunk is
unnecessary

-- 
balbi

Re: [PATCH] usb: musb: - kill the compile warning

[David Brownell]

On Thursday 12 February 2009, Felipe Balbi wrote:
> On Thu, Feb 12, 2009 at 07:03:25AM +0100, David Brownell wrote:
> > On Friday 06 February 2009, Bryan Wu wrote:
> > > -       return (MUSB_BUSCTL_OFFSET(i, 0) + mbase);
> > > +       return (void __iomem *)(MUSB_BUSCTL_OFFSET(i, 0) + mbase);
> > 
> > Why is MUSB_BUSCTL_OFFSET() returning something other than
> > an integer?  "mbase" is already "void __iomem *", so the only
> > way that should trigger a warning is if that OFFSET() macro
> > is doing something odd.
> 
> The warning only comes on Blackfin's version of this, this hunk is
> unnecessary

We want to get rid of Blackfin-specific warnings too ... but
what I was hinting is that this can't be the right fix.

Re: [PATCH] usb: musb: - kill the compile warning

[Felipe Balbi]

On Fri, Feb 06, 2009 at 06:23:30PM +0800, Bryan Wu wrote:
> drivers/usb/musb/musb_core.c:1433: warning: assignment makes pointer
> from integer without a cast
> 	hw_ep->target_regs = musb_read_target_reg_base(i, mbase);
> 
> static inline u16 musb_read_target_reg_base(u8 i, void __iomem *mbase);
> 
> this is a common bug, but a bug still ?  mbase is a 32/64 bit pointer,
> but we return a u16 to assign to a pointer ?  seems odd :)
> 
> So return right pointer in the stub.
> 
> Cc: Mike Frysinger <vapier.adi@gmail.com>
> Signed-off-by: Bryan Wu <cooloney@kernel.org>

any new version coming ??

-- 
balbi

Re: [PATCH] usb: musb: - kill the compile warning

[Mike Frysinger]

On Wed, Feb 25, 2009 at 13:16, Felipe Balbi wrote:
> On Fri, Feb 06, 2009 at 06:23:30PM +0800, Bryan Wu wrote:
>> drivers/usb/musb/musb_core.c:1433: warning: assignment makes pointer
>> from integer without a cast
>>       hw_ep->target_regs = musb_read_target_reg_base(i, mbase);
>>
>> static inline u16 musb_read_target_reg_base(u8 i, void __iomem *mbase);
>>
>> this is a common bug, but a bug still ?  mbase is a 32/64 bit pointer,
>> but we return a u16 to assign to a pointer ?  seems odd :)
>>
>> So return right pointer in the stub.
>>
>> Cc: Mike Frysinger <vapier.adi@gmail.com>
>> Signed-off-by: Bryan Wu <cooloney@kernel.org>
>
> any new version coming ??

so we're all on the same page ...

we agree the first hunk is redundant as MUSB_BUSCTL_OFFSET() is an
integer constant and mbase is of the type that is returned.  which
means the first hunk can be dropped.

the 2nd hunk is 100% correct though ...
-mike