This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug middle-end/61409] [5/6/7 regression] -Wmaybe-uninitialized false-positive with -O2
- From: "law at redhat dot com" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Wed, 16 Nov 2016 19:52:49 +0000
- Subject: [Bug middle-end/61409] [5/6/7 regression] -Wmaybe-uninitialized false-positive with -O2
- Auto-submitted: auto-generated
- References: <bug-61409-4@http.gcc.gnu.org/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61409
--- Comment #24 from Jeffrey A. Law <law at redhat dot com> ---
OK. It's coming back to me now. And yes, Aldy, it was the edge 9->6 :-)
So we have a PHI argument that references an uninitialized variable. There is
a control predicate for that PHI argument, call it p.
The use of the result of the PHI is also guarded. In this particular case the
guard is !p.
Thus there is no path through the CFG which uses the uninitialized variable.
We ought to be able to look at the guard of the PHI argument as well as the
guard for the use, at least that's the theory. Now onward to look at your
patch...