[PATCH] Allowing MC backends to decide relaxation based on fixup resolution

[Joerg Sonnenberger]

On Tue, Mar 10, 2015 at 06:23:03PM +0000, colinl at codeaurora.org wrote:
> Currently a failure to resolve a fixup implies instruction relaxation
> which may not be correct for all targets.  In Hexagon, fixups might
> always or never be relaxed depending a combination of fixup type and
> instruction opcode.

I can't parse that paragraph. Could you please rewrite it?

Joerg