On Tue, Sep 7, 2021 at 12:12 AM Michael Meissner via Gcc-patches <
gcc-patches@gcc.gnu.org> wrote:
> This patch fixes the breakage in the PowerPC due to a recent change in
> SUBREG
> behavior. While it is arguable that the patch that caused the breakage
> should
> be reverted, this patch should be a bandage to prevent these changes from
> happening again.
>
FYI I'm dealing with the same problem in the RISC-V backend via
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102211
My current proposed solution is a new predicate f_register_operand to be
used in patterns that use the f constraint which disallows the subregs that
can't work in FP registers due to NaN-boxing. This just borrows a few
lines of code that were deleted from validate_subreg and puts it in a
RISC-V specific predicate function. It worked for newlib and glibc
builds. I'm trying bootstrap testing now, but my target is a little slower
than yours so this will take some time.
Jim
