re PR sanitizer/64344 ([UBSAN] ICE with -fsanitize=float-cast-overflow [ICE in -fsani...

re PR sanitizer/64344 ([UBSAN] ICE with -fsanitize=float-cast-overflow [ICE in -fsanitize=float-cast-overflow])

	PR sanitizer/64344
	* ubsan.h (ubsan_instrument_float_cast): Add ARG argument.
	* ubsan.c (ubsan_instrument_float_cast): Add ARG argument, pass
	it to libubsan handler instead of EXPR.  Fold comparisons earlier,
	if the result is integer_zerop, return NULL_TREE.
	* convert.c (convert_to_integer): Pass expr as ARG.
c/
	* c-typeck.c (convert_for_assignment, c_finish_return): For
	-fsanitize=float-cast-overflow casts from REAL_TYPE to integer/enum
	types also set in_late_binary_op around convert call.
	* c-convert.c (convert): For -fsanitize=float-cast-overflow REAL_TYPE
	to integral type casts, if not in_late_binary_op, pass c_fully_fold
	result on expr as last argument to ubsan_instrument_float_cast,
	if in_late_binary_op, don't use c_save_expr but save_expr.
testsuite/
	* c-c++-common/ubsan/pr64344-1.c: New test.
	* c-c++-common/ubsan/pr64344-2.c: New test.

From-SVN: r219201