This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug bootstrap/41241] [4.5 regression] bootstrap comparison failure
- From: "vmakarov at redhat dot com" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 4 Sep 2009 15:25:17 -0000
- Subject: [Bug bootstrap/41241] [4.5 regression] bootstrap comparison failure
- References: <bug-41241-226@http.gcc.gnu.org/bugzilla/>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Comment #38 from vmakarov at redhat dot com 2009-09-04 15:25 -------
Finally, I think I found what is the reason for all these bootstrap failures.
I've changed code dealing with register preferences. Now regnfo::reg_pref is
not null from the start and some passes (now it is regmove) using
reg_(preferred,alternative)_class returns uninitialized data instead of
GENERAL_REGS as before.
I don't see sense for these passes to use this info (it was in the old RA too:
regclass initializing reg_pref was called only in local-alloc first time)
because code in the passes assumes that reg_preferred_class returns right class
(not GENERAL_REGS).
I'll work for a work around. I hope to send it today. The right solution will
take more time (may be week or two).
I apologize for all that trouble I cause with my patch. But it permitted to
find some pitfalls in RA (and the old one too).
Thanks for all your help.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41241