This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Fix several endless loops in cse_insn for certain target types combinations
On May 17, 2005, at 12:34 PM, Igor Shevlyakov wrote:
But that means to introduce fake type, I didn't plan to use. This will
require to add several new patterns into .md file (at least movoi),
handle calling convention for OI type and god knows what else. It will
affect the performance in other places since it will take one more
iteration from MIN_TYPE_INT to MAX_TYPE_INT. Plus, building a cross
compiler to support 256-bit type will be nightmare. Since part of
constant optimizations are still performed in HOST_WIDE_INT type, even
128-bit support on 32-bit host is broken. I've avoided this problem by
moving into AMD64 host, but 256-bit will bring that saga back.
I think the problem with 128bit support on 32bit host is your bug, you
need to set
need_64bit_hwint to yes in config.gcc for your target.
Thanks,
Andrew Pinski