[RFA] arm_extract_return_value, big-endian

[RFA] arm_extract_return_value, big-endian

[Michael Snyder]

[-- Attachment #1: Type: text/plain, Size: 199 bytes --]

[This doesn't seem to have gone out.  Excuse me if it appears twice]

One of you arm-savvy guys check me out on this:
it seems logical, and it fixes two fails in callfuncs with -mbig-endian.

Michael

[-- Attachment #2: arm.patch --]
[-- Type: text/plain, Size: 848 bytes --]

2002-11-05  Michael Snyder  <msnyder@redhat.com>

	* arm-tdep.c (arm_extract_return_value): Handle offset of 
	small types on big-endian machines.

Index: arm-tdep.c
===================================================================
RCS file: /cvs/src/src/gdb/arm-tdep.c,v
retrieving revision 1.74
diff -p -r1.74 arm-tdep.c
*** arm-tdep.c	1 Nov 2002 21:21:49 -0000	1.74
--- arm-tdep.c	6 Nov 2002 01:54:36 -0000
*************** arm_extract_return_value (struct type *t
*** 2274,2279 ****
--- 2274,2284 ----
  	  break;
  	}
      }
+   else if (TARGET_BYTE_ORDER == BFD_ENDIAN_BIG)
+     memcpy (valbuf, 
+ 	    &regbuf[REGISTER_BYTE (ARM_A1_REGNUM)] 
+ 	     + (REGISTER_RAW_SIZE (ARM_A1_REGNUM) - TYPE_LENGTH (type)), 
+ 	    TYPE_LENGTH (type));
    else
      memcpy (valbuf, &regbuf[REGISTER_BYTE (ARM_A1_REGNUM)],
  	    TYPE_LENGTH (type));

Re: [RFA] arm_extract_return_value, big-endian

[Fernando Nasser]

Well Michael, nobody has said anything, so if it seem to work lets make 
the change.  It seems reasonable to me.

My only question would be if we don't have any abstraction of the 
"extract a small type from a register" operation somewhere (arch vector, 
or an old macro).  But you've probably have checked for that already.

Regards to all,
Fernando

Michael Snyder wrote:
> [This doesn't seem to have gone out.  Excuse me if it appears twice]
> 
> One of you arm-savvy guys check me out on this:
> it seems logical, and it fixes two fails in callfuncs with -mbig-endian.
> 
> Michael
> 
> 
> ------------------------------------------------------------------------
> 
> 2002-11-05  Michael Snyder  <msnyder@redhat.com>
> 
> 	* arm-tdep.c (arm_extract_return_value): Handle offset of 
> 	small types on big-endian machines.
> 
> Index: arm-tdep.c
> ===================================================================
> RCS file: /cvs/src/src/gdb/arm-tdep.c,v
> retrieving revision 1.74
> diff -p -r1.74 arm-tdep.c
> *** arm-tdep.c	1 Nov 2002 21:21:49 -0000	1.74
> --- arm-tdep.c	6 Nov 2002 01:54:36 -0000
> *************** arm_extract_return_value (struct type *t
> *** 2274,2279 ****
> --- 2274,2284 ----
>   	  break;
>   	}
>       }
> +   else if (TARGET_BYTE_ORDER == BFD_ENDIAN_BIG)
> +     memcpy (valbuf, 
> + 	    &regbuf[REGISTER_BYTE (ARM_A1_REGNUM)] 
> + 	     + (REGISTER_RAW_SIZE (ARM_A1_REGNUM) - TYPE_LENGTH (type)), 
> + 	    TYPE_LENGTH (type));
>     else
>       memcpy (valbuf, &regbuf[REGISTER_BYTE (ARM_A1_REGNUM)],
>   	    TYPE_LENGTH (type));


-- 
Fernando Nasser
Red Hat Canada Ltd.                     E-Mail:  fnasser@redhat.com
2323 Yonge Street, Suite #300
Toronto, Ontario   M4P 2C9

Re: [RFA] arm_extract_return_value, big-endian

[Fernando Nasser]

Humm..., I am having second thoughts about this.  Isn't the problem you 
are seeing the same problem of not having the values peoperly 
sign-extended?  I guess I will have to look into this closer (and my ARM 
manuals are not here right now either).

Please do not apply the patch yet Michael.  Sorry,. :-(

Fernando

Michael Snyder wrote:
> [This doesn't seem to have gone out.  Excuse me if it appears twice]
> 
> One of you arm-savvy guys check me out on this:
> it seems logical, and it fixes two fails in callfuncs with -mbig-endian.
> 
> Michael
> 
> 
> ------------------------------------------------------------------------
> 
> 2002-11-05  Michael Snyder  <msnyder@redhat.com>
> 
> 	* arm-tdep.c (arm_extract_return_value): Handle offset of 
> 	small types on big-endian machines.
> 
> Index: arm-tdep.c
> ===================================================================
> RCS file: /cvs/src/src/gdb/arm-tdep.c,v
> retrieving revision 1.74
> diff -p -r1.74 arm-tdep.c
> *** arm-tdep.c	1 Nov 2002 21:21:49 -0000	1.74
> --- arm-tdep.c	6 Nov 2002 01:54:36 -0000
> *************** arm_extract_return_value (struct type *t
> *** 2274,2279 ****
> --- 2274,2284 ----
>   	  break;
>   	}
>       }
> +   else if (TARGET_BYTE_ORDER == BFD_ENDIAN_BIG)
> +     memcpy (valbuf, 
> + 	    &regbuf[REGISTER_BYTE (ARM_A1_REGNUM)] 
> + 	     + (REGISTER_RAW_SIZE (ARM_A1_REGNUM) - TYPE_LENGTH (type)), 
> + 	    TYPE_LENGTH (type));
>     else
>       memcpy (valbuf, &regbuf[REGISTER_BYTE (ARM_A1_REGNUM)],
>   	    TYPE_LENGTH (type));


-- 
Fernando Nasser
Red Hat Canada Ltd.                     E-Mail:  fnasser@redhat.com
2323 Yonge Street, Suite #300
Toronto, Ontario   M4P 2C9

Re: [RFA] arm_extract_return_value, big-endian

[Richard Earnshaw]

fnasser@redhat.com said:
> Humm..., I am having second thoughts about this.  Isn't the problem
> you  are seeing the same problem of not having the values peoperly
> sign-extended?  


No.  In this case we really need to copy the least significant 1 (or 2) 
bytes into the 1 or 2 bytes in the valbuf target.  That means doing a copy 
from the higher addresses.   So in that respect, the patch is correct.

But it breaks the case where the return value is more than one word.

R.

Re: [RFA] arm_extract_return_value, big-endian

[Michael Snyder]

[-- Attachment #1: Type: text/plain, Size: 632 bytes --]

Richard Earnshaw wrote:
> 
> fnasser@redhat.com said:
> > Humm..., I am having second thoughts about this.  Isn't the problem
> > you  are seeing the same problem of not having the values peoperly
> > sign-extended?
> 
> No.  In this case we really need to copy the least significant 1 (or 2)
> bytes into the 1 or 2 bytes in the valbuf target.  That means doing a copy
> from the higher addresses.   So in that respect, the patch is correct.
> 
> But it breaks the case where the return value is more than one word.

Yes, I see that now.  Richard, how about a joint effort?
Would you be so kind as to fill in the empty else clause?

[-- Attachment #2: arm.diff --]
[-- Type: text/plain, Size: 793 bytes --]

Index: arm-tdep.c
===================================================================
RCS file: /cvs/src/src/gdb/arm-tdep.c,v
retrieving revision 1.74
diff -p -r1.74 arm-tdep.c
*** arm-tdep.c	1 Nov 2002 21:21:49 -0000	1.74
--- arm-tdep.c	7 Jan 2003 23:45:19 -0000
*************** arm_extract_return_value (struct type *t
*** 2274,2279 ****
--- 2274,2292 ----
  	  break;
  	}
      }
+   else if (TARGET_BYTE_ORDER == BFD_ENDIAN_BIG)
+     {
+       if (TYPE_LENGTH (type) <= REGISTER_SIZE)
+ 	{
+ 	  memcpy (valbuf, 
+ 		  &regbuf[REGISTER_BYTE (ARM_A1_REGNUM)] 
+ 		  + (REGISTER_RAW_SIZE (ARM_A1_REGNUM) - TYPE_LENGTH (type)), 
+ 		  TYPE_LENGTH (type));
+ 	}
+       else
+ 	{
+ 	}
+     }
    else
      memcpy (valbuf, &regbuf[REGISTER_BYTE (ARM_A1_REGNUM)],
  	    TYPE_LENGTH (type));

Re: [RFA] arm_extract_return_value, big-endian

[Richard Earnshaw]

> Richard Earnshaw wrote:
> > 
> > fnasser@redhat.com said:
> > > Humm..., I am having second thoughts about this.  Isn't the problem
> > > you  are seeing the same problem of not having the values peoperly
> > > sign-extended?
> > 
> > No.  In this case we really need to copy the least significant 1 (or 2)
> > bytes into the 1 or 2 bytes in the valbuf target.  That means doing a copy
> > from the higher addresses.   So in that respect, the patch is correct.
> > 
> > But it breaks the case where the return value is more than one word.
> 
> Yes, I see that now.  Richard, how about a joint effort?
> Would you be so kind as to fill in the empty else clause?

I think I've already fixed this.  Did you miss my request that you test it 
for me on a big-endian system?


http://sources.redhat.com/ml/gdb-patches/2002-12/msg00444.html

R.