This is the mail archive of the
mailing list for the GCC project.
Re: [PATCH RS6000] adding an unsigned SI and a DI together and subtracting an unsigned SI from an DI
- From: David Edelsohn <dje at watson dot ibm dot com>
- To: Richard Henderson <rth at redhat dot com>, Andrew Pinski <pinskia at physics dot uc dot edu>, gcc-patches at gcc dot gnu dot org
- Date: Wed, 14 May 2003 12:55:22 -0400
- Subject: Re: [PATCH RS6000] adding an unsigned SI and a DI together and subtracting an unsigned SI from an DI
>>>>> Richard Henderson writes:
Richard> It'd be nice if the fixed-point exception register was
Richard> represented in gcc, so you could split these, but...
The XER register is defined in the machine description, but is not
modeled. It is unclear how much of a win it is to schedule those
instructions. XLC does not, for instance.
What is the current recommendation with regard to use of
gen_lowpart/gen_highpart vs operand_subword vs simplify_gen_subreg? The
rs6000 port historically has used operand_subword in 32-bit mode, which is
somewhat deprecated but allows the endian option to be described by an
argument. All variants lead to simplify_gen_subreg. This patch is using
gen_lowpart/gen_highpart for that purpose.