ocaml/asmcomp/arm/proc.ml

363 lines
12 KiB
OCaml

(**************************************************************************)
(* *)
(* OCaml *)
(* *)
(* Benedikt Meurer, University of Siegen *)
(* *)
(* Copyright 1998 Institut National de Recherche en Informatique et *)
(* en Automatique. *)
(* Copyright 2012 Benedikt Meurer. *)
(* *)
(* All rights reserved. This file is distributed under the terms of *)
(* the GNU Lesser General Public License version 2.1, with the *)
(* special exception on linking described in the file LICENSE. *)
(* *)
(**************************************************************************)
(* Description of the ARM processor *)
open Misc
open Cmm
open Reg
open Arch
open Mach
(* Instruction selection *)
let word_addressed = false
(* Registers available for register allocation *)
(* Integer register map:
r0 - r3 general purpose (not preserved)
r4 - r7 general purpose (preserved)
r8 trap pointer (preserved)
r9 platform register, usually reserved
r10 allocation pointer (preserved)
r11 domain state pointer (preserved)
r12 intra-procedural scratch register (not preserved)
r13 stack pointer
r14 return address
r15 program counter
Floating-point register map (VFPv{2,3}):
d0 - d7 general purpose (not preserved)
d8 - d15 general purpose (preserved)
d16 - d31 general purpose (not preserved), VFPv3 only
*)
let int_reg_name =
[| "r0"; "r1"; "r2"; "r3"; "r4"; "r5"; "r6"; "r7"; "r12" |]
let float_reg_name =
[| "d0"; "d1"; "d2"; "d3"; "d4"; "d5"; "d6"; "d7";
"d8"; "d9"; "d10"; "d11"; "d12"; "d13"; "d14"; "d15";
"d16"; "d17"; "d18"; "d19"; "d20"; "d21"; "d22"; "d23";
"d24"; "d25"; "d26"; "d27"; "d28"; "d29"; "d30"; "d31" |]
(* We have three register classes:
0 for integer registers
1 for VFPv2 and VFPv3-D16
2 for VFPv3
This way we can choose between VFPv2/VFPv3-D16 and VFPv3
at (ocamlopt) runtime using command line switches.
*)
let num_register_classes = 3
let register_class r =
match (r.typ, !fpu) with
| (Val | Int | Addr), _ -> 0
| Float, VFPv2 -> 1
| Float, VFPv3_D16 -> 1
| Float, _ -> 2
let num_available_registers =
[| 9; 16; 32 |]
let first_available_register =
[| 0; 100; 100 |]
let register_name r =
if r < 100 then int_reg_name.(r) else float_reg_name.(r - 100)
let rotate_registers = true
(* Representation of hard registers by pseudo-registers *)
let hard_int_reg =
let v = Array.make 9 Reg.dummy in
for i = 0 to 8 do
v.(i) <- Reg.at_location Int (Reg i)
done;
v
let hard_float_reg =
let v = Array.make 32 Reg.dummy in
for i = 0 to 31 do
v.(i) <- Reg.at_location Float (Reg(100 + i))
done;
v
let all_phys_regs =
Array.append hard_int_reg hard_float_reg
let phys_reg n =
if n < 100 then hard_int_reg.(n) else hard_float_reg.(n - 100)
let stack_slot slot ty =
Reg.at_location ty (Stack slot)
(* Calling conventions *)
let loc_int last_int make_stack int ofs =
if !int <= last_int then begin
let l = phys_reg !int in
incr int; l
end else begin
let l = stack_slot (make_stack !ofs) Int in
ofs := !ofs + size_int; l
end
let loc_float last_float make_stack float ofs =
assert (abi = EABI_HF);
assert (!fpu >= VFPv2);
if !float <= last_float then begin
let l = phys_reg !float in
incr float; l
end else begin
ofs := Misc.align !ofs size_float;
let l = stack_slot (make_stack !ofs) Float in
ofs := !ofs + size_float; l
end
let loc_int_pair last_int make_stack int ofs =
(* 64-bit quantities split across two registers must either be in a
consecutive pair of registers where the lowest numbered is an
even-numbered register; or in a stack slot that is 8-byte aligned. *)
int := Misc.align !int 2;
if !int <= last_int - 1 then begin
let reg_lower = phys_reg !int in
let reg_upper = phys_reg (1 + !int) in
int := !int + 2;
[| reg_lower; reg_upper |]
end else begin
let size_int64 = size_int * 2 in
ofs := Misc.align !ofs size_int64;
let stack_lower = stack_slot (make_stack !ofs) Int in
let stack_upper = stack_slot (make_stack (size_int + !ofs)) Int in
ofs := !ofs + size_int64;
[| stack_lower; stack_upper |]
end
let calling_conventions first_int last_int first_float last_float make_stack
arg =
let loc = Array.make (Array.length arg) Reg.dummy in
let int = ref first_int in
let float = ref first_float in
let ofs = ref 0 in
for i = 0 to Array.length arg - 1 do
match arg.(i) with
| Val | Int | Addr ->
loc.(i) <- loc_int last_int make_stack int ofs
| Float ->
loc.(i) <- loc_float last_float make_stack float ofs
done;
(loc, Misc.align !ofs 8) (* keep stack 8-aligned *)
let incoming ofs = Incoming ofs
let outgoing ofs = Outgoing ofs
let not_supported _ofs = fatal_error "Proc.loc_results: cannot call"
(* OCaml calling convention:
first integer args in r0...r7
first float args in d0...d15 (EABI+VFP)
remaining args on stack.
Return values in r0...r7 or d0...d15. *)
let max_arguments_for_tailcalls = 8
let loc_arguments arg =
calling_conventions 0 7 100 115 outgoing arg
let loc_parameters arg =
let (loc, _) = calling_conventions 0 7 100 115 incoming arg in loc
let loc_results res =
let (loc, _) = calling_conventions 0 7 100 115 not_supported res in loc
(* C calling convention:
first integer args in r0...r3
first 64-bit integer args in r0-r1, r2-r3
first float args in d0...d7 (EABI+VFP)
first float args in r0-r1, r2-r3 (soft FP)
remaining args on stack.
Return values in r0, r0-r1, or d0. *)
let external_calling_conventions first_int last_int first_float last_float
make_stack ty_args =
let loc = Array.make (List.length ty_args) [| Reg.dummy |] in
let int = ref first_int in
let float = ref first_float in
let ofs = ref 0 in
List.iteri
(fun i ty_arg ->
match ty_arg with
| XInt | XInt32 ->
loc.(i) <- [| loc_int last_int make_stack int ofs |]
| XInt64 ->
loc.(i) <- loc_int_pair last_int make_stack int ofs
| XFloat ->
loc.(i) <-
(if abi = EABI_HF
then [| loc_float last_float make_stack float ofs |]
else loc_int_pair last_int make_stack int ofs))
ty_args;
(loc, Misc.align !ofs 8) (* keep stack 8-aligned *)
let loc_external_arguments ty_args =
external_calling_conventions 0 3 100 107 outgoing ty_args
let loc_external_results res =
let (loc, _) = calling_conventions 0 1 100 100 not_supported res
in loc
let loc_exn_bucket = phys_reg 0
(* See "DWARF for the ARM architecture" available from developer.arm.com. *)
let int_dwarf_reg_numbers =
[| 0; 1; 2; 3; 4; 5; 6; 7; 12 |]
let float_dwarf_reg_numbers_legacy =
[| 64; 65; 66; 67; 68; 69; 70; 71;
72; 73; 74; 75; 76; 77; 78; 79;
80; 81; 82; 83; 84; 85; 86; 87;
88; 89; 90; 91; 92; 93; 94; 95;
|]
let float_dwarf_reg_numbers =
[| 256; 257; 258; 259; 260; 261; 262; 263;
264; 265; 266; 267; 268; 269; 270; 271;
272; 273; 274; 275; 276; 277; 278; 279;
280; 281; 282; 283; 284; 285; 286; 287;
|]
let dwarf_register_numbers ~reg_class =
match reg_class with
| 0 -> int_dwarf_reg_numbers
| 1 ->
(* Section 3.1 note 4 says that the "new" VFPv3 register numberings
(as per [float_dwarf_reg_numbers]) should be used for VFPv2 as well.
However we believe that for <= ARMv6 we should use the legacy VFPv2
numberings. *)
if !arch <= ARMv6 then float_dwarf_reg_numbers_legacy
else float_dwarf_reg_numbers
| 2 -> float_dwarf_reg_numbers
| _ -> Misc.fatal_errorf "Bad register class %d" reg_class
let stack_ptr_dwarf_register_number = 13
(* Volatile registers: none *)
let regs_are_volatile _rs = false
(* Registers destroyed by operations *)
let destroyed_at_alloc = (* r0-r6, d0-d15 preserved *)
Array.of_list (List.map
phys_reg
[7;8;
116;117;118;119;120;121;122;123;
124;125;126;127;128;129;130;131])
let destroyed_at_c_call =
Array.of_list (List.map
phys_reg
(match abi with
EABI -> (* r4-r7 preserved *)
[0;1;2;3;8;
100;101;102;103;104;105;106;107;
108;109;110;111;112;113;114;115;
116;117;118;119;120;121;122;123;
124;125;126;127;128;129;130;131]
| EABI_HF -> (* r4-r7, d8-d15 preserved *)
[0;1;2;3;8;
100;101;102;103;104;105;106;107;
116;117;118;119;120;121;122;123;
124;125;126;127;128;129;130;131]))
let destroyed_at_oper = function
Iop(Icall_ind _ | Icall_imm _)
| Iop(Iextcall { alloc = true; _ }) ->
all_phys_regs
| Iop(Iextcall { alloc = false; _}) ->
destroyed_at_c_call
| Iop(Ialloc _) ->
destroyed_at_alloc
| Iop(Iconst_symbol _) when !Clflags.pic_code ->
[| phys_reg 3; phys_reg 8 |] (* r3 and r12 destroyed *)
| Iop(Iintop Imulh) when !arch < ARMv6 ->
[| phys_reg 8 |] (* r12 destroyed *)
| Iop(Iintop (Icomp _) | Iintop_imm(Icomp _, _))
when !arch >= ARMv8 && !thumb ->
[| phys_reg 3 |] (* r3 destroyed *)
| Iop(Iintoffloat | Ifloatofint | Iload(Single, _) | Istore(Single, _, _)) ->
[| phys_reg 107 |] (* d7 (s14-s15) destroyed *)
| _ -> [||]
let destroyed_at_raise = all_phys_regs
(* lr is destroyed at [Lreloadretaddr], but lr is not used for register
allocation, and thus does not need to (and indeed cannot) occur here. *)
let destroyed_at_reloadretaddr = [| |]
(* Maximal register pressure *)
let safe_register_pressure = function
Iextcall _ -> if abi = EABI then 0 else 4
| Ialloc _ -> if abi = EABI then 0 else 7
| Iconst_symbol _ when !Clflags.pic_code -> 7
| Iintop Imulh when !arch < ARMv6 -> 8
| _ -> 9
let max_register_pressure = function
Iextcall _ -> if abi = EABI then [| 4; 0; 0 |] else [| 4; 8; 8 |]
| Ialloc _ -> if abi = EABI then [| 7; 0; 0 |] else [| 7; 8; 8 |]
| Iconst_symbol _ when !Clflags.pic_code -> [| 7; 16; 32 |]
| Iintoffloat | Ifloatofint
| Iload(Single, _) | Istore(Single, _, _) -> [| 9; 15; 31 |]
| Iintop Imulh when !arch < ARMv6 -> [| 8; 16; 32 |]
| _ -> [| 9; 16; 32 |]
(* Pure operations (without any side effect besides updating their result
registers). *)
let op_is_pure = function
| Icall_ind _ | Icall_imm _ | Itailcall_ind _ | Itailcall_imm _
| Iextcall _ | Istackoffset _ | Istore _ | Ialloc _
| Iintop(Icheckbound _) | Iintop_imm(Icheckbound _, _)
| Ispecific(Ishiftcheckbound _) -> false
| _ -> true
(* Layout of the stack *)
let frame_required fd =
let num_stack_slots = fd.fun_num_stack_slots in
fd.fun_contains_calls
|| num_stack_slots.(0) > 0
|| num_stack_slots.(1) > 0
|| num_stack_slots.(2) > 0
let prologue_required fd =
frame_required fd
(* Calling the assembler *)
let assemble_file infile outfile =
Ccomp.command (Config.asm ^ " " ^
(String.concat " " (Misc.debug_prefix_map_flags ())) ^
" -o " ^ Filename.quote outfile ^ " " ^ Filename.quote infile)
let init () = ()