Hi,
My target architecture has 32 bit registers and uses register pairs
for 64 bit values. What makes it special is that whenever a register
pair is written by a 64 bit operation, the individual registers should
not be read by 32 bit operations. The reverse is also true: whenever a
register pair i
Hi,
I am looking for a way to specify that data is not aligned and that therefore
word accesses should be expanded into byte accesses. For example:
void f(__unaligned int *a)
{
*a = 0;
}
Expands into 4 byte stores. Any suggestions on how to implement this?
Regards,
Jan
Hi,
I am trying to define vec_init. I can only do splats on register
values. So
operands[1] of vec_init should be registers and those registers should
be the same.
Here is some code:
(define_expand "vec_init"
[(match_operand:VEC_MODE32 0 "register_operand" "")
(match_operand 1 "
Hi,
I want to introduce a __int40 data type in my port for 40 bits integers. I
do:
type = make_signed_type(40);
lang_hooks.types.register_builtin_type(type, "__int40");
The result is that __int40 variables are mapped on DImode with all kinds of
shifts and ands to obtain a 40 bits data type.
Hi,
I need to handle certain pragmas such that something special is generated in
the assembly output roughly at the place that corresponds to the place of
the pragma in the source code. Doing this in gcc in a clean way is not
possible I guess, so I have to apply a hack. The hack that I want to
