FWD: ioctl h8300 specific problem fix

FWD: ioctl h8300 specific problem fix

[Andrew Lunn]

Hi Folks

Any suggests as to what we should do this with patch for ioctl.  I
don't like target specific #define in a generic file like this. We
could just make varargs the standard solution.

        Thanks
                Andrew

Index: io/fileio/current/ChangeLog
===================================================================
RCS file: /cvsroot/ecos-h8/ecos/packages/io/fileio/current/ChangeLog,v
retrieving revision 1.1.1.13
retrieving revision 1.2
diff -u -r1.1.1.13 -r1.2
--- io/fileio/current/ChangeLog	27 Jan 2005 02:38:13 -0000	1.1.1.13
+++ io/fileio/current/ChangeLog	27 Jan 2005 14:10:56 -0000	1.2
@@ -1,3 +1,10 @@
+2005-01-27  Yoshinori Sato  <ysato@users.sourceforge.jp>
+
+	* src/io.cxx (ioctl): Change a receipt of argument in variable length 
+	with h8300.
+	Because it is a register normally, cannot get argument rightly 
+	that don't agree with prototype.
+
 2005-01-22  Andrew Lunn  <andrew.lunn@ascom.ch>
 
 	* src/misc.cxx (cyg_fs_root_lookup): New function to find the mount
Index: io/fileio/current/src/io.cxx
===================================================================
RCS file: /cvsroot/ecos-h8/ecos/packages/io/fileio/current/src/io.cxx,v
retrieving revision 1.1.1.5
retrieving revision 1.2
diff -u -r1.1.1.5 -r1.2
--- io/fileio/current/src/io.cxx	4 Mar 2004 05:17:58 -0000	1.1.1.5
+++ io/fileio/current/src/io.cxx	27 Jan 2005 14:10:56 -0000	1.2
@@ -252,12 +252,29 @@
 //==========================================================================
 // ioctl
 
+#include <cyg/infra/diag.h>
+#if !defined(__H8300H__) && !defined(__H8300S__)
 __externC int ioctl( int fd, CYG_ADDRWORD com, CYG_ADDRWORD data )
 {
     FILEIO_ENTRY();
 
     int ret;
     cyg_file *fp;
+#else
+#include <stdarg.h>
+__externC int ioctl( int fd, CYG_ADDRWORD com, ... )
+{
+    FILEIO_ENTRY();
+
+    int ret;
+    cyg_file *fp;
+    CYG_ADDRWORD data;
+    va_list ap;
+
+    va_start(ap, com);
+    data = va_arg(ap, CYG_ADDRWORD);
+    va_end(ap);
+#endif
     
     fp = cyg_fp_get( fd );
 
-- 
Yoshinori Sato
<ysato@users.sourceforge.jp>

----- End forwarded message -----

Re: FWD: ioctl h8300 specific problem fix

[Gary Thomas]

On Sun, 2005-01-30 at 23:46 +0100, Andrew Lunn wrote:
> Hi Folks
> 
> Any suggests as to what we should do this with patch for ioctl.  I
> don't like target specific #define in a generic file like this. We
> could just make varargs the standard solution.

I think so - having platform/architecture code is not acceptable.

> 
>         Thanks
>                 Andrew
> 
> Index: io/fileio/current/ChangeLog
> ===================================================================
> RCS file: /cvsroot/ecos-h8/ecos/packages/io/fileio/current/ChangeLog,v
> retrieving revision 1.1.1.13
> retrieving revision 1.2
> diff -u -r1.1.1.13 -r1.2
> --- io/fileio/current/ChangeLog	27 Jan 2005 02:38:13 -0000	1.1.1.13
> +++ io/fileio/current/ChangeLog	27 Jan 2005 14:10:56 -0000	1.2
> @@ -1,3 +1,10 @@
> +2005-01-27  Yoshinori Sato  <ysato@users.sourceforge.jp>
> +
> +	* src/io.cxx (ioctl): Change a receipt of argument in variable length 
> +	with h8300.
> +	Because it is a register normally, cannot get argument rightly 
> +	that don't agree with prototype.
> +
>  2005-01-22  Andrew Lunn  <andrew.lunn@ascom.ch>
>  
>  	* src/misc.cxx (cyg_fs_root_lookup): New function to find the mount
> Index: io/fileio/current/src/io.cxx
> ===================================================================
> RCS file: /cvsroot/ecos-h8/ecos/packages/io/fileio/current/src/io.cxx,v
> retrieving revision 1.1.1.5
> retrieving revision 1.2
> diff -u -r1.1.1.5 -r1.2
> --- io/fileio/current/src/io.cxx	4 Mar 2004 05:17:58 -0000	1.1.1.5
> +++ io/fileio/current/src/io.cxx	27 Jan 2005 14:10:56 -0000	1.2
> @@ -252,12 +252,29 @@
>  //==========================================================================
>  // ioctl
>  
> +#include <cyg/infra/diag.h>
> +#if !defined(__H8300H__) && !defined(__H8300S__)
>  __externC int ioctl( int fd, CYG_ADDRWORD com, CYG_ADDRWORD data )
>  {
>      FILEIO_ENTRY();
>  
>      int ret;
>      cyg_file *fp;
> +#else
> +#include <stdarg.h>
> +__externC int ioctl( int fd, CYG_ADDRWORD com, ... )
> +{
> +    FILEIO_ENTRY();
> +
> +    int ret;
> +    cyg_file *fp;
> +    CYG_ADDRWORD data;
> +    va_list ap;
> +
> +    va_start(ap, com);
> +    data = va_arg(ap, CYG_ADDRWORD);
> +    va_end(ap);
> +#endif
>      
>      fp = cyg_fp_get( fd );
>  
-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------

Re: FWD: ioctl h8300 specific problem fix

[Nick Garnett]

Gary Thomas <gary@mlbassoc.com> writes:

> On Sun, 2005-01-30 at 23:46 +0100, Andrew Lunn wrote:
> > Hi Folks
> > 
> > Any suggests as to what we should do this with patch for ioctl.  I
> > don't like target specific #define in a generic file like this. We
> > could just make varargs the standard solution.
> 
> I think so - having platform/architecture code is not acceptable.

Agreed. I don't now recall why the H8300 has problems with this, but
if the change will work for all platforms, then it should be made. 

-- 
Nick Garnett                                          eCos Kernel Architect
http://www.ecoscentric.com                     The eCos and RedBoot experts
Besuchen Sie uns vom 22.-24.02.05 auf der Embedded World 2005, Stand 11-124
Visit us at Embedded World 2005, Nürnberg, Germany, 22-24 Feb, Stand 11-124

Re: FWD: ioctl h8300 specific problem fix

[Bart Veer]

>>>>> "Andrew" == Andrew Lunn <andrew@lunn.ch> writes:

    Andrew> Hi Folks
    Andrew> Any suggests as to what we should do this with patch for
    Andrew> ioctl. I don't like target specific #define in a generic
    Andrew> file like this. We could just make varargs the standard
    Andrew> solution.

We should make the varargs version the standard solution. ioctl() has
not been well-standardized in the past, but these days the varargs
version seems to be ubiquitous and is listed @
http://www.opengroup.org/onlinepubs/009695399/functions/ioctl.html

Bart