[patches] Re: Patch: x86_64 disassembler support

[Alan Modra]

On Thu, 4 Jan 2001, Jan Hubicka wrote:

> I still believe it makes sense to keep tables as similar as possible,
> but if you think otherwise, just let me know and I can remove the relevant
> bits.

It's OK.  I know how easy it is to accidentally leave old patches in code,
which is why I was asking about the change.

> +static bfd_signed_vma
> +get32s ()
> +{
> +  bfd_signed_vma x = 0;
> +
> +  FETCH_DATA (the_info, codep + 4);
> +  x = *codep++ & (bfd_signed_vma) 0xff;
> +  x |= (*codep++ & (bfd_signed_vma) 0xff) << 8;
> +  x |= (*codep++ & (bfd_signed_vma) 0xff) << 16;
> +  x |= (*codep++ & (bfd_signed_vma) 0xff) << 24;
> +
> +  if ((x & ~(((bfd_signed_vma) 2 << 31) - 1)) == 0)

The above line is redundant, isn't it?

> +    x = (x ^ ((bfd_signed_vma) 1 << 31)) - ((bfd_signed_vma) 1 << 31);

If you've run the binutils testsuite, then this is OK to commit.

Alan
-- 
Linuxcare.  Support for the Revolution.