This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug rtl-optimization/15927] New: THUMB -O2: strength-reduced iteration variable ends up off by 1
- From: "danfuzz at milk dot com" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 11 Jun 2004 00:34:45 -0000
- Subject: [Bug rtl-optimization/15927] New: THUMB -O2: strength-reduced iteration variable ends up off by 1
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
When compiling a particular function with -O2 and targetting THUMB, I found that the compiler
generated incorrect code. In particular it looks like an iteration variable that was strength-reduced ends
up off by one: In each iteration, the value used for the variable in question appears to be the value that
ought to be used in the *subsequent* iteration.
The line in question that gets erroneously compiled is:
ic = scan->ics[n];
I'll attach the raw source (pared down but still verified to compile) and the annotated assembly file as
soon as I figure out how to do that with bugzilla.
--
Summary: THUMB -O2: strength-reduced iteration variable ends up
off by 1
Product: gcc
Version: 3.4.0
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: rtl-optimization
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: danfuzz at milk dot com
CC: gcc-bugs at gcc dot gnu dot org
GCC build triplet: powerpc-mac-osx
GCC host triplet: powerpc-mac-osx
GCC target triplet: arm-unknown-elf
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15927