This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: RFC/RFA: Fix bug with REE optimization corrupting extended registers

On 11/19/2015 10:18 AM, Bernd Schmidt wrote:

On 11/19/2015 06:06 PM, Jeff Law wrote:


Frankly, the overall structure of ree is a mess -- I've found it
incredibly difficult to follow every time I've had to look at it.


Yeah, no kidding. The check seems to be in the wrong place - it's done
very late when we're replacing things, and IMO we should just restrict
the candidates for optimization much earlier.
And I didn't make it any cleaner when I hacked it up to handle one more case :( I'm hoping that a year away will allow me to look and think about the code again without getting a horrid headache. 




Also, I want to apply the following. Ok if testing succeeds?


Bernd

clean-ree1.diff


commit ce68938b5150f5d41a54ed317ab97d98461be064
Author: Bernd Schmidt<bernds_cb1@t-online.de>
Date:   Thu Nov 19 17:38:15 2015 +0100

     Readability improvements in ree.c:combine_reaching_defs.

     	* ree.c (combine_reaching_defs): Simplify code by caching expressions
     	in variables.

Yes. That's good with me.

jeff
Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]