|GCC porting question firstname.lastname@example.org (C Jaiprakash, Noida) (2003-07-13)|
|Re: GCC porting question email@example.com (Tim Olson) (2003-07-15)|
|Re: GCC porting question firstname.lastname@example.org (Ken Rose) (2003-07-17)|
|Re: GCC porting question email@example.com (Michael Meissner) (2003-07-17)|
|From:||Ken Rose <firstname.lastname@example.org>|
|Date:||17 Jul 2003 00:28:50 -0400|
|Organization:||Posted via Supernews, http://www.supernews.com|
|Posted-Date:||17 Jul 2003 00:28:50 EDT|
"C Jaiprakash, Noida" wrote:
> Can gcc be ported to a machine which do not have diaplacement
> addressing mode?
Yes. I've done it.
> If yes then what sould macros related to base register be defined as?
> for ex ( REG_OK_FOR_BASE_P, BASE_REG_CLASS )
I was working with a machine where the 32 GP registers were OK for an
indirect load, hence the otherwise-magical 32s below. Here are the
relevant lines from tm.h.
#define BASE_REG_CLASS ALL_REGS
#define REGNO_OK_FOR_BASE_P(REGNO) \
((REGNO) < 32 || (unsigned) reg_renumber[REGNO] < 32)
# define REG_OK_FOR_BASE_P(X) 1
# define REG_OK_FOR_BASE_P(X) REGNO_OK_FOR_BASE_P (REGNO (X))
Tag me off-group if I can help further. (or on-group, if John would
Return to the
Search the comp.compilers archives again.