This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug middle-end/79454] [7 Regression] c-c++-common/ubsan/overflow-vec-*.c FAILs on some 64-bit BE targets
- From: "jakub at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Fri, 10 Feb 2017 15:17:33 +0000
- Subject: [Bug middle-end/79454] [7 Regression] c-c++-common/ubsan/overflow-vec-*.c FAILs on some 64-bit BE targets
- Auto-submitted: auto-generated
- References: <bug-79454-4@http.gcc.gnu.org/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79454
--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 40712
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=40712&action=edit
gcc7-pr79454.patch
Untested fix. The problem was if the vector type had non-BLKmode, but not
vector mode, like e.g. for the 16xchar vector TImode, we actually performed the
addition to get the result in TImode rather than in V16QImode (not really
supported), or saving the result inside of the loop from each separate
comparison.