[PATCH] scheduler register pressure excessive stack space

[law@redhat.com]

 In message <200205292102.RAA26694@makai.watson.ibm.com>, David Edelsohn 
writes:
 > 	This failure occurs in the stage2 compiler, so presumably the
 > stage1 compiler mis-compiles stage2.  Figuring out how the compiler is
 > miscompiled may not be easier than an execution c-torture test that
 > fails.  Maybe your flow buglet will fix the miscompilation, but I have not
 > seen the patch to try it.  I also cannot spend all my time debugging this
 > one problem.
Ah.  I wasn't aware it was triggering in stage2.  Ugh.  I'm probably not going
to be able to track this down anytime soon.


jeff