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

[Jim Grosbach]

In http://reviews.llvm.org/D8217#144272, @colinl wrote:

> I'm definitely open to cleaner implementations.
>
> The issue with not doing anything in relaxInstruction() is relaxation will get stuck in an infinite loop.  The relaxation loop assumes that relaxInstruction will eventually change fixupNeedsRelaxation or mayNeedRelaxation to false.


Yeah, we'll likely need to do something there. My inclination is do something with the target-specific fixups. Perhaps use a different fixup kind which isn't relaxable?


REPOSITORY
  rL LLVM

http://reviews.llvm.org/D8217

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/