RFA: put fpscr in right place in core file

RFA: put fpscr in right place in core file

[Jim Blandy]

This fixes a failure in gcore.exp.

2004-02-29  Jim Blandy  <jimb@redhat.com>

	* ppc-linux-nat.c (fill_fpregset): Place the 32-bit value of the
	fpscr in the right end of the 64-bit slot for it in the fpregset
	array.

Index: gdb/ppc-linux-nat.c
===================================================================
RCS file: /cvs/cvsfiles/devo/gdb/ppc-linux-nat.c,v
retrieving revision 2.19.6.4
diff -c -r2.19.6.4 ppc-linux-nat.c
*** gdb/ppc-linux-nat.c	29 Feb 2004 07:51:24 -0000	2.19.6.4
--- gdb/ppc-linux-nat.c	1 Mar 2004 04:51:15 -0000
***************
*** 1032,1038 ****
  	regcache_collect (FP0_REGNUM + regi, (char *) (*fpregsetp + regi));
      }
    if ((regno == -1) || regno == tdep->ppc_fpscr_regnum)
!     regcache_collect (tdep->ppc_fpscr_regnum, (char *) (*fpregsetp + regi));
  }
  
  
--- 1032,1044 ----
  	regcache_collect (FP0_REGNUM + regi, (char *) (*fpregsetp + regi));
      }
    if ((regno == -1) || regno == tdep->ppc_fpscr_regnum)
!     {
!       /* The FPSCR is stored in the low order word of the last
!          doubleword in the fpregset.  */
!       int offset = (TARGET_BYTE_ORDER == BFD_ENDIAN_BIG) ? 4 : 0;
!       regcache_collect (tdep->ppc_fpscr_regnum,
!                         (char *) (*fpregsetp + regi) + offset);
!     }
  }
  
  

Re: RFA: put fpscr in right place in core file

[Kevin Buettner]

On 24 Mar 2004 11:17:47 -0500
Jim Blandy <jimb@redhat.com> wrote:

> This fixes a failure in gcore.exp.
> 
> 2004-02-29  Jim Blandy  <jimb@redhat.com>
> 
> 	* ppc-linux-nat.c (fill_fpregset): Place the 32-bit value of the
> 	fpscr in the right end of the 64-bit slot for it in the fpregset
> 	array.

Okay.

Kevin

Re: RFA: put fpscr in right place in core file

[Jim Blandy]

Kevin Buettner <kevinb@redhat.com> writes:
> On 24 Mar 2004 11:17:47 -0500
> Jim Blandy <jimb@redhat.com> wrote:
> 
> > This fixes a failure in gcore.exp.
> > 
> > 2004-02-29  Jim Blandy  <jimb@redhat.com>
> > 
> > 	* ppc-linux-nat.c (fill_fpregset): Place the 32-bit value of the
> > 	fpscr in the right end of the 64-bit slot for it in the fpregset
> > 	array.
> 
> Okay.

Oops.  This got fixed last week.  Never mind.

2004-03-15  Andrew Cagney  <cagney@redhat.com>

	* ppc-tdep.h: Update copyright.
	(ppc_linux_supply_gregset, ppc_linux_supply_fpregset): Change
	function signatures to match "regsets.h".
	* ppc-linux-tdep.c: Include "regset.h".
	(ELF_GREGSET_SIZE): Delete.
	(right_supply_register): New function.
	(ppc_linux_supply_fpregset, ppc_linux_supply_gregset): Rewrite
	using right_supply_register.
	(ppc32_linux_supply_gregset, ppc64_linux_supply_gregset): New
	functions.
	(ppc64_linux_gregset, ppc32_linux_gregset): Define.
	(ppc_linux_init_abi): Register ppc_linux_regset_from_core_section.
	(_initialize_ppc_linux_tdep): Do not register
	ppc_linux_regset_core_fns.
	(ppc_linux_regset_from_core_section): Replace
	fetch_core_registers.
	(ppc_linux_regset_core_fns): Delete.
	* ppc-linux-nat.c: (right_fill_reg): New function.
	(supply_gregset): Update call to ppc_linux_supply_gregset.
	(fill_gregset): Clear the register set, use right_fill_reg.
	(supply_fpregset): Update call to ppc_linux_supply_fpregset.
	(fill_fpregset): Use right_fill_reg, correctly compute FP offsets.