This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug rtl-optimization/55489] [4.7/4.8 regression] insane PRE memory usage with PIE (translate.i)
- From: "bonzini at gnu dot org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Tue, 27 Nov 2012 15:50:44 +0000
- Subject: [Bug rtl-optimization/55489] [4.7/4.8 regression] insane PRE memory usage with PIE (translate.i)
- Auto-submitted: auto-generated
- References: <bug-55489-4@http.gcc.gnu.org/bugzilla/>
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55489
Paolo Bonzini <bonzini at gnu dot org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Known to work|4.8.0 |
Summary|[4.7 regression] insane PRE |[4.7/4.8 regression] insane
|memory usage with PIE |PRE memory usage with PIE
|(translate.i) |(translate.i)
Known to fail| |4.8.0
--- Comment #1 from Paolo Bonzini <bonzini at gnu dot org> 2012-11-27 15:50:44 UTC ---
Strictly not a regression, because the underlying bug is present since always.
But something between 4.4.6 and 4.7.2 caused memory usage to become much worse.
Also, marking it as not-working in 4.8 because the allocation of
reg_known_value is wrong:
vec_alloc (reg_known_value, maxreg - FIRST_PSEUDO_REGISTER);
and that's what causes the bug to be latent: set_reg_known_value becomes a
no-op, and get_reg_known_value always returns NULL.