Memory offsets in inline assembly

There are various ways of specifying addresses in A64 assembler.

/*
[base{,#0}]                 Simple register (exclusive) - Immediate Offset
[base{,#imm}]               Offset - Immediate Offset
[base,Xm{,LSL #imm}]        Offset - Register Offset
[base,Wm,(S|U)XTW {#imm}]   Offset - Extended Register Offset
[base,#imm]!                Pre-indexed - Immediate Offset
[base],#imm                 Post-indexed - Immediate Offset
label                       PC-relative (literal) load - Immediate Offset
*/

      


I would like to use "Offset - Immediate Offset" in inline assembly.

__asm__("ldp x8, x9, %0, 16 \n\t"
        :
        : "m" (*somePointer)
        : "x8", "x9");

      

This translates to

ldp x8, x9, [x0], 16

      

My goal is to achieve

ldp x8, x9, [x0, 16]

      

How to write such instructions with inline assembler?

+2


source to share


1 answer


I don't have a 64-bit ARM toolchain to test this, but you should do something like this:

asm("ldp x8, x9, %0\n\t"
    :
    : "Ump" (*((__int128 *) somePointer + 1))
    : "x8", "x9");

      

The constraint Ump

limits the memory operand to those permitted by the integer LDP instruction, otherwise it acts as a constraint m

. If it is somePointer

already a pointer to a 128-bit type, you can simply use somePointer[1]

as an operand.



If the above doesn't work, David Wolferd's suggestion should:

asm("ldp x8, x9, [%0, %1]" 
    : 
    : "r" (somePointer), "i"(16), "m" (*somePointer) 
    : "x8", "x9"); 

      

0


source







All Articles