This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: lcm improvements/fixes
- To: law at cygnus dot com, rth at cygnus dot com
- Subject: Re: lcm improvements/fixes
- From: Andrew Macleod <amacleod at cygnus dot com>
- Date: Thu, 11 Nov 1999 12:11:59 -0800
- Cc: gcc-patches at gcc dot gnu dot org
>>
>> On Wed, Nov 10, 1999 at 11:37:40PM -0700, Jeffrey A Law wrote:
>> > ! /* Do not clear the aux field for blocks which are
>> > ! predecessors of the EXIT block. That way we never
>> > ! add then to the worklist again. */
>>
>> We need to decide on protocol here. Either we have to always assume
>> AUX is garbage on entry to a routine, or we have to be dilligent about
>> resetting it to NULL on exit.
I was already assuming its always garbage. The next place using it might
not care if its been initialized, so zeroing it out would be a waste of time. If
someone wants it all zeros, let them set it to zeros. Has anyone been assuming
its zero's?
You're right, let make sure now that we dont have different protocols early.
I think Jeffs comment relates to a local clearing within the algorithm, and
isn't counting on anything coming in.
Andrew