[PATCH] headers_check fix: linux/pps.h

[PATCH] headers_check fix: linux/pps.h

[Jaswinder Singh Rajput]

fix the following 'make headers_check' warnings:

  usr/include/linux/pps.h:52: found __[us]{8,16,32,64} type without #include <linux/types.h>

Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
---
 include/linux/pps.h |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)

diff --git a/include/linux/pps.h b/include/linux/pps.h
index cfe5c72..0194ab0 100644
--- a/include/linux/pps.h
+++ b/include/linux/pps.h
@@ -22,6 +22,8 @@
 #ifndef _PPS_H_
 #define _PPS_H_
 
+#include <linux/types.h>
+
 #define PPS_VERSION		"5.3.6"
 #define PPS_MAX_SOURCES		16		/* should be enough... */
 
-- 
1.6.0.6

Re: [PATCH] headers_check fix: linux/pps.h

[Sam Ravnborg]

On Sat, Jun 27, 2009 at 10:03:34PM +0530, Jaswinder Singh Rajput wrote:
> 
> fix the following 'make headers_check' warnings:
> 
>   usr/include/linux/pps.h:52: found __[us]{8,16,32,64} type without #include <linux/types.h>
> 
> Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
> ---
>  include/linux/pps.h |    2 ++
>  1 files changed, 2 insertions(+), 0 deletions(-)
> 
> diff --git a/include/linux/pps.h b/include/linux/pps.h
> index cfe5c72..0194ab0 100644
> --- a/include/linux/pps.h
> +++ b/include/linux/pps.h
> @@ -22,6 +22,8 @@
>  #ifndef _PPS_H_
>  #define _PPS_H_
>  
> +#include <linux/types.h>
> +
>  #define PPS_VERSION		"5.3.6"
>  #define PPS_MAX_SOURCES		16		/* should be enough... */

That file has other issues that should be addresses too.
1) It uses int rather than wide specific types
2) It uses structs with questionable alignmner as per David's comment

Keeping the warning until all issues are sorted out is preferred.
If we 'fix' the warning then we loose the reminder that this file
needs to be eyeballed.

This is not a quest to eliminate warnings - this is a quest to
raise the quality and correctness of the exported headers.
The warnings serves as reminders where to focus attention.

	Sam

Re: [PATCH] headers_check fix: linux/pps.h

[Jaswinder Singh Rajput]

On Sat, 2009-06-27 at 23:53 +0200, Sam Ravnborg wrote:
> On Sat, Jun 27, 2009 at 10:03:34PM +0530, Jaswinder Singh Rajput wrote:
> > 
> > fix the following 'make headers_check' warnings:
> > 
> >   usr/include/linux/pps.h:52: found __[us]{8,16,32,64} type without #include <linux/types.h>
> > 
> > Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
> > ---
> >  include/linux/pps.h |    2 ++
> >  1 files changed, 2 insertions(+), 0 deletions(-)
> > 
> > diff --git a/include/linux/pps.h b/include/linux/pps.h
> > index cfe5c72..0194ab0 100644
> > --- a/include/linux/pps.h
> > +++ b/include/linux/pps.h
> > @@ -22,6 +22,8 @@
> >  #ifndef _PPS_H_
> >  #define _PPS_H_
> >  
> > +#include <linux/types.h>
> > +
> >  #define PPS_VERSION		"5.3.6"
> >  #define PPS_MAX_SOURCES		16		/* should be enough... */
> 
> That file has other issues that should be addresses too.
> 1) It uses int rather than wide specific types
> 2) It uses structs with questionable alignmner as per David's comment
> 

This should be send in different patch.

> Keeping the warning until all issues are sorted out is preferred.
> If we 'fix' the warning then we loose the reminder that this file
> needs to be eyeballed.
> 

You mean files which do not get headers_check warning are absolutely OK.
This is totally insane.

This are different issues and need to send by different series of
patches. Please do not mix up things.

> This is not a quest to eliminate warnings - this is a quest to
> raise the quality and correctness of the exported headers.
> The warnings serves as reminders where to focus attention.
> 

ditto.

--
JSR

Re: [PATCH] headers_check fix: linux/pps.h

[Jaswinder Singh Rajput]

On Sun, 2009-06-28 at 10:47 +0530, Jaswinder Singh Rajput wrote:
> On Sat, 2009-06-27 at 23:53 +0200, Sam Ravnborg wrote:
> > On Sat, Jun 27, 2009 at 10:03:34PM +0530, Jaswinder Singh Rajput wrote:
> > > 
> > > fix the following 'make headers_check' warnings:
> > > 
> > >   usr/include/linux/pps.h:52: found __[us]{8,16,32,64} type without #include <linux/types.h>
> > > 
> > > Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
> > > ---
> > >  include/linux/pps.h |    2 ++
> > >  1 files changed, 2 insertions(+), 0 deletions(-)
> > > 
> > > diff --git a/include/linux/pps.h b/include/linux/pps.h
> > > index cfe5c72..0194ab0 100644
> > > --- a/include/linux/pps.h
> > > +++ b/include/linux/pps.h
> > > @@ -22,6 +22,8 @@
> > >  #ifndef _PPS_H_
> > >  #define _PPS_H_
> > >  
> > > +#include <linux/types.h>
> > > +
> > >  #define PPS_VERSION		"5.3.6"
> > >  #define PPS_MAX_SOURCES		16		/* should be enough... */
> > 
> > That file has other issues that should be addresses too.
> > 1) It uses int rather than wide specific types
> > 2) It uses structs with questionable alignmner as per David's comment
> > 
> 
> This should be send in different patch.
> 
> > Keeping the warning until all issues are sorted out is preferred.
> > If we 'fix' the warning then we loose the reminder that this file
> > needs to be eyeballed.
> > 
> 
> You mean files which do not get headers_check warning are absolutely OK.
> This is totally insane.
> 
> This are different issues and need to send by different series of
> patches. Please do not mix up things.

If others maintainers will be like you they will never get any patch. If
they keep on saying go and fix complete source file first.

Do you get some idea, what are you requesting for.

If you find the problem why do not you submit patches ?

Thanks,
--
JSR