This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
machine description - 'high' meaning
- To: <gcc at gcc dot gnu dot org>
- Subject: machine description - 'high' meaning
- From: "Dragos Badea" <badrian at ss dot pub dot ro>
- Date: Wed, 12 Apr 2000 20:46:41 +0300
have tried to describe a double precision multiplication operation by a
define_expand. My problem is -I guess- that I don't know if (high)
expression really considers the high portion of a register, as I want to
specify.
So I come into such dump fragment, where the REG_EQUAL note is totally
wrong with my intention.
(insn 13 12 14 (set (reg:SI 38)
(mult:SI (zero_extend:SI (subreg:HI (reg:SI 35) 0))
(zero_extend:SI (subreg:HI (reg:SI 36) 0)))) -1 (nil)
(nil))
(insn 14 13 16 (set (reg:SI 37)
(mult:SI (high:SI (reg:SI 35))
(zero_extend:SI (subreg:HI (reg:SI 36) 0)))) -1 (nil)
(expr_list:REG_EQUAL (mult:SI (reg:SI 35)
(reg:SI 36))
(nil)))
In case my assumption about (high) is wrong, could you guide me please
through the right expression?
Thank you so much,
Adrian