https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71644
rr7crecb3z at liamekaens dot com changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |rr7crecb3z at liamekaens dot
com
--- Comment #1 from rr7crecb3z at liamekaens dot com ---
7 years and this bug is STILL "UNCONFIRMED". I'm here because I ran into this
bug after building Python, using GCC 11.3.0.
The reproduction program provided by jus...@gmx.li back in 2016 STILL
reproduces the bug in GCC 11.3.0.
Can't there just be a way to disable the use of the MOVAPS instruction
entirely, without having to disable optimizations that don't have bugs? I'm
sick and tired of dealing with this bug. GCC obviously will never be able to be
trusted to use the MOVAPS instruction correctly.
