[patch]: 2 more needed changes for sim-fpu.c & psim

[patch]: 2 more needed changes for sim-fpu.c & psim

[matthew green]

these provide the missing sim_io_error function that's required for
sim-fpu.c to actually link & be used.


OK?


.mrg.


2001-12-14  matthew green  <mrg@redhat.com>

	* main.c: Include "defs.h", "bfd.h", "callback.h" and "remote-sim.h".
	(sim_io_error): New function.
	* sim_calls.c: (sim_io_error): New function.


Index: main.c
===================================================================
RCS file: /cvs/src/src/sim/ppc/main.c,v
retrieving revision 1.1.1.1
diff -p -r1.1.1.1 main.c
*** main.c	1999/04/16 01:35:11	1.1.1.1
--- main.c	2001/12/14 01:11:24
***************
*** 30,35 ****
--- 30,40 ----
  #include "device.h" /* FIXME: psim should provide the interface */
  #include "events.h" /* FIXME: psim should provide the interface */
  
+ #include "defs.h"
+ #include "bfd.h"
+ #include "callback.h"
+ #include "remote-sim.h"
+ 
  #ifdef HAVE_STDLIB_H
  #include <stdlib.h>
  #endif
*************** sim_io_flush_stdoutput(void)
*** 220,225 ****
--- 225,248 ----
      error("sim_io_flush_stdoutput: invalid switch\n");
      break;
    }
+ }
+ 
+ void
+ sim_io_error (SIM_DESC sd, const char *msg, ...)
+ {
+   va_list ap;
+   va_start(ap, msg);
+   vprintf(msg, ap);
+   printf("\n");
+   va_end(ap);
+ 
+ #if 0 /* FIXME: `simulation' is static to ppc/main.c */
+   /* any final clean up */
+   if (ppc_trace[trace_print_info] && simulation != NULL)
+     psim_print_info (simulation, ppc_trace[trace_print_info]);
+ #endif
+ 
+   exit (1);
  }
  
  
Index: sim_calls.c
===================================================================
RCS file: /cvs/src/src/sim/ppc/sim_calls.c,v
retrieving revision 1.2
diff -p -r1.2 sim_calls.c
*** sim_calls.c	2000/12/08 01:52:41	1.2
--- sim_calls.c	2001/12/14 01:11:28
*************** sim_io_flush_stdoutput(void)
*** 441,446 ****
--- 441,464 ----
    }
  }
  
+ void
+ sim_io_error (SIM_DESC sd, const char *msg, ...)
+ {
+   va_list ap;
+   va_start(ap, msg);
+   vprintf(msg, ap);
+   printf("\n");
+   va_end(ap);
+ 
+ #if 0 /* FIXME: `simulation' is static to ppc/main.c */
+   /* any final clean up */
+   if (ppc_trace[trace_print_info] && simulation != NULL)
+     psim_print_info (simulation, ppc_trace[trace_print_info]);
+ #endif
+ 
+   exit (1);
+ }
+ 
  /****/
  
  void *

Re: [patch]: 2 more needed changes for sim-fpu.c & psim

[Andrew Cagney]

 
> + #include "defs.h"
> + #include "bfd.h"
> + #include "callback.h"
> + #include "remote-sim.h"


I'm puzzled by some of the includes.  Well "defs.h" to be honest.   I 
know the PPC has a nasty hack so that it can access the GDB register 
names but I'm not sure that main.c - a standalone program - should 
include "defs.h".


>   #ifdef HAVE_STDLIB_H
>   #include <stdlib.h>
> #endif
> *************** sim_io_flush_stdoutput(void)
> *** 220,225 ****
> --- 225,248 ----
>       error("sim_io_flush_stdoutput: invalid switch\n");
>       break;
>     }
> + }
> + 
> + void
> + sim_io_error (SIM_DESC sd, const char *msg, ...)
> + {
> +   va_list ap;
> +   va_start(ap, msg);
> +   vprintf(msg, ap);
> +   printf("\n");
> +   va_end(ap);


I think the error stuff should go to stderr.

 From memory, for the sim_call.c, the error stuff should go to the callback.

Andrew

re: [patch]: 2 more needed changes for sim-fpu.c & psim

[matthew green]

   
   I'm puzzled by some of the includes.  Well "defs.h" to be honest.   I 
   know the PPC has a nasty hack so that it can access the GDB register 
   names but I'm not sure that main.c - a standalone program - should 
   include "defs.h".

main.c doesn't need defs.h, right.
   
   I think the error stuff should go to stderr.

hmm, i choose stdout because the "error" function a couple of pages
above does the same...
   
   From memory, for the sim_call.c, the error stuff should go to the callback.

OK, i've done this..  how's this?



	* main.c: Include "bfd.h", "callback.h" and "remote-sim.h".
	(sim_io_error): New function.
	* sim_calls.c: (sim_io_error): New function.


Index: sim_calls.c
===================================================================
RCS file: /cvs/src/src/sim/ppc/sim_calls.c,v
retrieving revision 1.2
diff -p -r1.2 sim_calls.c
*** sim_calls.c	2000/12/08 01:52:41	1.2
--- sim_calls.c	2001/12/15 04:14:35
*************** sim_io_flush_stdoutput(void)
*** 441,446 ****
--- 441,456 ----
    }
  }
  
+ void
+ sim_io_error (SIM_DESC sd, const char *fmt, ...)
+ {
+   va_list ap;
+   va_start(ap, fmt);
+   callbacks->evprintf_filtered (callbacks, fmt, ap);
+   va_end(ap);
+   callbacks->error (callbacks, "");
+ }
+ 
  /****/
  
  void *
Index: main.c
===================================================================
RCS file: /cvs/src/src/sim/ppc/main.c,v
retrieving revision 1.1.1.1
diff -p -r1.1.1.1 main.c
*** main.c	1999/04/16 01:35:11	1.1.1.1
--- main.c	2001/12/15 04:14:35
***************
*** 30,35 ****
--- 30,39 ----
  #include "device.h" /* FIXME: psim should provide the interface */
  #include "events.h" /* FIXME: psim should provide the interface */
  
+ #include "bfd.h"
+ #include "callback.h"
+ #include "remote-sim.h"
+ 
  #ifdef HAVE_STDLIB_H
  #include <stdlib.h>
  #endif
*************** sim_io_flush_stdoutput(void)
*** 220,225 ****
--- 224,245 ----
      error("sim_io_flush_stdoutput: invalid switch\n");
      break;
    }
+ }
+ 
+ void
+ sim_io_error (SIM_DESC sd, const char *msg, ...)
+ {
+   va_list ap;
+   va_start(ap, msg);
+   vprintf(msg, ap);
+   printf("\n");
+   va_end(ap);
+ 
+   /* any final clean up */
+   if (ppc_trace[trace_print_info] && simulation != NULL)
+     psim_print_info (simulation, ppc_trace[trace_print_info]);
+ 
+   exit (1);
  }
  
  

Re: [patch]: 2 more needed changes for sim-fpu.c & psim

[Andrew Cagney]

>  I'm puzzled by some of the includes.  Well "defs.h" to be honest.   I 
>    know the PPC has a nasty hack so that it can access the GDB register 
>    names but I'm not sure that main.c - a standalone program - should 
>    include "defs.h".
> 
> main.c doesn't need defs.h, right.
>    
>    I think the error stuff should go to stderr.
> 
> hmm, i choose stdout because the "error" function a couple of pages
> above does the same...


Ah, ok then.


>    From memory, for the sim_call.c, the error stuff should go to the callback.
> 
> OK, i've done this..  how's this?
> 
> 
> 
> 	* main.c: Include "bfd.h", "callback.h" and "remote-sim.h".
> 	(sim_io_error): New function.
> 	* sim_calls.c: (sim_io_error): New function.
> 

Yes fine.

Thanks!
Andrew