Re: Tidy up arm-elf header files

[Richard Earnshaw <rearnsha at arm dot com>]

> >  So here is a patch that does this.  I have not checked it in yet as
> >  I do not have the means to test the arm-linux or arm-conix
> >  configurations.  Richard, Phil can you have a look at this patch and
> >  tell me if you think that it is OK ?
> 
> I can't bootstrap the trunk on arm-linux at all right now.  I get this 
> badness while building stage 2:
> 
> ../../gcc/gcc/dwarf2out.c: In function `value_format':
> ../../gcc/gcc/dwarf2out.c:5955: In basic block 17:
> ../../gcc/gcc/dwarf2out.c:5955: Flow control insn inside a basic block
> (jump_insn 350 349 351 (parallel[ 
>             (set (pc)
>                 (if_then_else (leu (reg:SI 89)
>                         (const_int 7 [0x7]))
>                     (mem:SI (plus:SI (mult:SI (reg:SI 89)
>                                 (const_int 4 [0x4]))
>                             (label_ref 351)) 0)
>                     (label_ref 245)))
>             (clobber (reg:CC 24 cc))
>             (use (label_ref 351))
>         ] ) -1 (nil)
>     (nil))
> ../../gcc/gcc/dwarf2out.c:5955: confused by earlier errors, bailing out

Yep, I've been seeing this as well.  It looks like a basic block has been 
relocated and that the end marker has been changed to point to the 
addr*vec rather than the dispatch insn.  I haven't been able to track down 
exactly where this happens yet, except that it occurs after the gcse dump 
is made.

R.