When stubs to interpreter are generated, the address of the interpreter is known. So we can either generate a direct jump or an adrp-based jump. An option also is to emit the address of the interpreter once, at the start of the stub section. So we can use `ldr rscratch1, adr; br rscratch1` instead of:

0x0000ffff79bd2570: mov x8, #0xe58c // #58764
0x0000ffff79bd2574: movk x8, #0x793b, lsl #16
0x0000ffff79bd2578: movk x8, #0xffff, lsl #32
0x0000ffff79bd257c: br x8