This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug tree-optimization/34618] [4.3 regression] ICE with -fmudflap and vectorization
- From: "irar at il dot ibm dot com" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Subject: [Bug tree-optimization/34618] [4.3 regression] ICE with -fmudflap and vectorization
- References: <bug-34618-1771@http.gcc.gnu.org/bugzilla/>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Comment #1 from irar at il dot ibm dot com 2008-01-01 10:02 -------
The failure occurs in is_gimple_val() for this variable. The problem is that
DECL_GIMPLE_REG_P for it is false. The vectorizer sets this value to be true
for every variable of the vector type that the vectorizer creates.
No ICE occurs when the complete unrolling pass is disabled.
I wonder if the unroller should duplicate this marking (set DECL_GIMPLE_REG_P
to true) or the vectorizer should somehow create vector vars differently? And
how this is related to mudflap?..
Ira
--
irar at il dot ibm dot com changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |NEW
Ever Confirmed|0 |1
Last reconfirmed|0000-00-00 00:00:00 |2008-01-01 10:02:49
date| |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34618