Hi,

On Wed, 15 Sep 1999, Ignatios Souvatzis wrote:

> On Thu, Sep 09, 1999 at 01:01:48PM +0100, Richard Earnshaw wrote:
> > 
> > I posted a PR with a patch for this some time ago -- see bin/7565
> > 
> > Nothing's been done with the PR :-(
> This has been fixed in -current.

Hmmm... some part of the patch changed :

  /* Instruction condition field values.  */
! #define INST_EQ               0x00000000
! #define INST_NE               0x10000000
! #define INST_CS               0x20000000
! #define INST_CC               0x30000000
! #define INST_MI               0x40000000
! #define INST_PL               0x50000000
! #define INST_VS               0x60000000
! #define INST_VC               0x70000000
! #define INST_HI               0x80000000
! #define INST_LS               0x90000000
! #define INST_GE               0xa0000000
! #define INST_LT               0xb0000000
! #define INST_GT               0xc0000000
! #define INST_LE               0xd0000000
! #define INST_AL               0xe0000000
! #define INST_NV               0xf0000000

to :  
  /* Instruction condition field values.  */
! #define INST_EQ               0x0
! #define INST_NE               0x1
! #define INST_CS               0x2
! #define INST_CC               0x3
! #define INST_MI               0x4
! #define INST_PL               0x5
! #define INST_VS               0x6
! #define INST_VC               0x7
! #define INST_HI               0x8
! #define INST_LS               0x9
! #define INST_GE               0xa
! #define INST_LT               0xb
! #define INST_GT               0xc
! #define INST_LE               0xd
! #define INST_AL               0xe
! #define INST_NV               0xf


I find this change kinda wierd; Are these flags now shifted? or is gcc big
endian oriented ? 

Puzzeled, and without a gcc source :-)

Reinoud