PythonExtra/py/asmarm.c
Damien George a70a4e6688 py/emitglue: Always flush caches when assigning native ARM code.
Prior to this commit, cache flushing for ARM native code was done only in
the assembler code asm_thumb_end_pass()/asm_arm_end_pass(), at the last
pass of the assembler.  But this misses flushing the cache when loading
native code from an .mpy file, ie in persistentcode.c.

The change here makes sure the cache is always flushed/cleaned/invalidated
when assigning native code on ARM architectures.

This problem was found running tests/micropython/import_mpy_native_gc.py on
the mimxrt port.

Signed-off-by: Damien George <damien@micropython.org>
2021-06-05 11:03:04 +10:00

373 lines
11 KiB
C

/*
* This file is part of the MicroPython project, http://micropython.org/
*
* The MIT License (MIT)
*
* Copyright (c) 2014 Fabian Vogt
* Copyright (c) 2013, 2014 Damien P. George
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
#include <stdio.h>
#include <assert.h>
#include <string.h>
#include "py/mpconfig.h"
// wrapper around everything in this file
#if MICROPY_EMIT_ARM
#include "py/asmarm.h"
#define SIGNED_FIT24(x) (((x) & 0xff800000) == 0) || (((x) & 0xff000000) == 0xff000000)
// Insert word into instruction flow
STATIC void emit(asm_arm_t *as, uint op) {
uint8_t *c = mp_asm_base_get_cur_to_write_bytes(&as->base, 4);
if (c != NULL) {
*(uint32_t *)c = op;
}
}
// Insert word into instruction flow, add "ALWAYS" condition code
STATIC void emit_al(asm_arm_t *as, uint op) {
emit(as, op | ASM_ARM_CC_AL);
}
// Basic instructions without condition code
STATIC uint asm_arm_op_push(uint reglist) {
// stmfd sp!, {reglist}
return 0x92d0000 | (reglist & 0xFFFF);
}
STATIC uint asm_arm_op_pop(uint reglist) {
// ldmfd sp!, {reglist}
return 0x8bd0000 | (reglist & 0xFFFF);
}
STATIC uint asm_arm_op_mov_reg(uint rd, uint rn) {
// mov rd, rn
return 0x1a00000 | (rd << 12) | rn;
}
STATIC uint asm_arm_op_mov_imm(uint rd, uint imm) {
// mov rd, #imm
return 0x3a00000 | (rd << 12) | imm;
}
STATIC uint asm_arm_op_mvn_imm(uint rd, uint imm) {
// mvn rd, #imm
return 0x3e00000 | (rd << 12) | imm;
}
STATIC uint asm_arm_op_add_imm(uint rd, uint rn, uint imm) {
// add rd, rn, #imm
return 0x2800000 | (rn << 16) | (rd << 12) | (imm & 0xFF);
}
STATIC uint asm_arm_op_add_reg(uint rd, uint rn, uint rm) {
// add rd, rn, rm
return 0x0800000 | (rn << 16) | (rd << 12) | rm;
}
STATIC uint asm_arm_op_sub_imm(uint rd, uint rn, uint imm) {
// sub rd, rn, #imm
return 0x2400000 | (rn << 16) | (rd << 12) | (imm & 0xFF);
}
STATIC uint asm_arm_op_sub_reg(uint rd, uint rn, uint rm) {
// sub rd, rn, rm
return 0x0400000 | (rn << 16) | (rd << 12) | rm;
}
STATIC uint asm_arm_op_mul_reg(uint rd, uint rm, uint rs) {
// mul rd, rm, rs
assert(rd != rm);
return 0x0000090 | (rd << 16) | (rs << 8) | rm;
}
STATIC uint asm_arm_op_and_reg(uint rd, uint rn, uint rm) {
// and rd, rn, rm
return 0x0000000 | (rn << 16) | (rd << 12) | rm;
}
STATIC uint asm_arm_op_eor_reg(uint rd, uint rn, uint rm) {
// eor rd, rn, rm
return 0x0200000 | (rn << 16) | (rd << 12) | rm;
}
STATIC uint asm_arm_op_orr_reg(uint rd, uint rn, uint rm) {
// orr rd, rn, rm
return 0x1800000 | (rn << 16) | (rd << 12) | rm;
}
void asm_arm_bkpt(asm_arm_t *as) {
// bkpt #0
emit_al(as, 0x1200070);
}
// locals:
// - stored on the stack in ascending order
// - numbered 0 through num_locals-1
// - SP points to first local
//
// | SP
// v
// l0 l1 l2 ... l(n-1)
// ^ ^
// | low address | high address in RAM
void asm_arm_entry(asm_arm_t *as, int num_locals) {
assert(num_locals >= 0);
as->stack_adjust = 0;
as->push_reglist = 1 << ASM_ARM_REG_R1
| 1 << ASM_ARM_REG_R2
| 1 << ASM_ARM_REG_R3
| 1 << ASM_ARM_REG_R4
| 1 << ASM_ARM_REG_R5
| 1 << ASM_ARM_REG_R6
| 1 << ASM_ARM_REG_R7
| 1 << ASM_ARM_REG_R8;
// Only adjust the stack if there are more locals than usable registers
if (num_locals > 3) {
as->stack_adjust = num_locals * 4;
// Align stack to 8 bytes
if (num_locals & 1) {
as->stack_adjust += 4;
}
}
emit_al(as, asm_arm_op_push(as->push_reglist | 1 << ASM_ARM_REG_LR));
if (as->stack_adjust > 0) {
emit_al(as, asm_arm_op_sub_imm(ASM_ARM_REG_SP, ASM_ARM_REG_SP, as->stack_adjust));
}
}
void asm_arm_exit(asm_arm_t *as) {
if (as->stack_adjust > 0) {
emit_al(as, asm_arm_op_add_imm(ASM_ARM_REG_SP, ASM_ARM_REG_SP, as->stack_adjust));
}
emit_al(as, asm_arm_op_pop(as->push_reglist | (1 << ASM_ARM_REG_PC)));
}
void asm_arm_push(asm_arm_t *as, uint reglist) {
emit_al(as, asm_arm_op_push(reglist));
}
void asm_arm_pop(asm_arm_t *as, uint reglist) {
emit_al(as, asm_arm_op_pop(reglist));
}
void asm_arm_mov_reg_reg(asm_arm_t *as, uint reg_dest, uint reg_src) {
emit_al(as, asm_arm_op_mov_reg(reg_dest, reg_src));
}
size_t asm_arm_mov_reg_i32(asm_arm_t *as, uint rd, int imm) {
// Insert immediate into code and jump over it
emit_al(as, 0x59f0000 | (rd << 12)); // ldr rd, [pc]
emit_al(as, 0xa000000); // b pc
size_t loc = mp_asm_base_get_code_pos(&as->base);
emit(as, imm);
return loc;
}
void asm_arm_mov_reg_i32_optimised(asm_arm_t *as, uint rd, int imm) {
// TODO: There are more variants of immediate values
if ((imm & 0xFF) == imm) {
emit_al(as, asm_arm_op_mov_imm(rd, imm));
} else if (imm < 0 && imm >= -256) {
// mvn is "move not", not "move negative"
emit_al(as, asm_arm_op_mvn_imm(rd, ~imm));
} else {
asm_arm_mov_reg_i32(as, rd, imm);
}
}
void asm_arm_mov_local_reg(asm_arm_t *as, int local_num, uint rd) {
// str rd, [sp, #local_num*4]
emit_al(as, 0x58d0000 | (rd << 12) | (local_num << 2));
}
void asm_arm_mov_reg_local(asm_arm_t *as, uint rd, int local_num) {
// ldr rd, [sp, #local_num*4]
emit_al(as, 0x59d0000 | (rd << 12) | (local_num << 2));
}
void asm_arm_cmp_reg_i8(asm_arm_t *as, uint rd, int imm) {
// cmp rd, #imm
emit_al(as, 0x3500000 | (rd << 16) | (imm & 0xFF));
}
void asm_arm_cmp_reg_reg(asm_arm_t *as, uint rd, uint rn) {
// cmp rd, rn
emit_al(as, 0x1500000 | (rd << 16) | rn);
}
void asm_arm_setcc_reg(asm_arm_t *as, uint rd, uint cond) {
emit(as, asm_arm_op_mov_imm(rd, 1) | cond); // movCOND rd, #1
emit(as, asm_arm_op_mov_imm(rd, 0) | (cond ^ (1 << 28))); // mov!COND rd, #0
}
void asm_arm_add_reg_reg_reg(asm_arm_t *as, uint rd, uint rn, uint rm) {
// add rd, rn, rm
emit_al(as, asm_arm_op_add_reg(rd, rn, rm));
}
void asm_arm_sub_reg_reg_reg(asm_arm_t *as, uint rd, uint rn, uint rm) {
// sub rd, rn, rm
emit_al(as, asm_arm_op_sub_reg(rd, rn, rm));
}
void asm_arm_mul_reg_reg_reg(asm_arm_t *as, uint rd, uint rs, uint rm) {
// rs and rm are swapped because of restriction rd!=rm
// mul rd, rm, rs
emit_al(as, asm_arm_op_mul_reg(rd, rm, rs));
}
void asm_arm_and_reg_reg_reg(asm_arm_t *as, uint rd, uint rn, uint rm) {
// and rd, rn, rm
emit_al(as, asm_arm_op_and_reg(rd, rn, rm));
}
void asm_arm_eor_reg_reg_reg(asm_arm_t *as, uint rd, uint rn, uint rm) {
// eor rd, rn, rm
emit_al(as, asm_arm_op_eor_reg(rd, rn, rm));
}
void asm_arm_orr_reg_reg_reg(asm_arm_t *as, uint rd, uint rn, uint rm) {
// orr rd, rn, rm
emit_al(as, asm_arm_op_orr_reg(rd, rn, rm));
}
void asm_arm_mov_reg_local_addr(asm_arm_t *as, uint rd, int local_num) {
// add rd, sp, #local_num*4
emit_al(as, asm_arm_op_add_imm(rd, ASM_ARM_REG_SP, local_num << 2));
}
void asm_arm_mov_reg_pcrel(asm_arm_t *as, uint reg_dest, uint label) {
assert(label < as->base.max_num_labels);
mp_uint_t dest = as->base.label_offsets[label];
mp_int_t rel = dest - as->base.code_offset;
rel -= 12 + 8; // adjust for load of rel, and then PC+8 prefetch of add_reg_reg_reg
// To load rel int reg_dest, insert immediate into code and jump over it
emit_al(as, 0x59f0000 | (reg_dest << 12)); // ldr rd, [pc]
emit_al(as, 0xa000000); // b pc
emit(as, rel);
// Do reg_dest += PC
asm_arm_add_reg_reg_reg(as, reg_dest, reg_dest, ASM_ARM_REG_PC);
}
void asm_arm_lsl_reg_reg(asm_arm_t *as, uint rd, uint rs) {
// mov rd, rd, lsl rs
emit_al(as, 0x1a00010 | (rd << 12) | (rs << 8) | rd);
}
void asm_arm_lsr_reg_reg(asm_arm_t *as, uint rd, uint rs) {
// mov rd, rd, lsr rs
emit_al(as, 0x1a00030 | (rd << 12) | (rs << 8) | rd);
}
void asm_arm_asr_reg_reg(asm_arm_t *as, uint rd, uint rs) {
// mov rd, rd, asr rs
emit_al(as, 0x1a00050 | (rd << 12) | (rs << 8) | rd);
}
void asm_arm_ldr_reg_reg(asm_arm_t *as, uint rd, uint rn, uint byte_offset) {
// ldr rd, [rn, #off]
emit_al(as, 0x5900000 | (rn << 16) | (rd << 12) | byte_offset);
}
void asm_arm_ldrh_reg_reg(asm_arm_t *as, uint rd, uint rn) {
// ldrh rd, [rn]
emit_al(as, 0x1d000b0 | (rn << 16) | (rd << 12));
}
void asm_arm_ldrb_reg_reg(asm_arm_t *as, uint rd, uint rn) {
// ldrb rd, [rn]
emit_al(as, 0x5d00000 | (rn << 16) | (rd << 12));
}
void asm_arm_str_reg_reg(asm_arm_t *as, uint rd, uint rm, uint byte_offset) {
// str rd, [rm, #off]
emit_al(as, 0x5800000 | (rm << 16) | (rd << 12) | byte_offset);
}
void asm_arm_strh_reg_reg(asm_arm_t *as, uint rd, uint rm) {
// strh rd, [rm]
emit_al(as, 0x1c000b0 | (rm << 16) | (rd << 12));
}
void asm_arm_strb_reg_reg(asm_arm_t *as, uint rd, uint rm) {
// strb rd, [rm]
emit_al(as, 0x5c00000 | (rm << 16) | (rd << 12));
}
void asm_arm_str_reg_reg_reg(asm_arm_t *as, uint rd, uint rm, uint rn) {
// str rd, [rm, rn, lsl #2]
emit_al(as, 0x7800100 | (rm << 16) | (rd << 12) | rn);
}
void asm_arm_strh_reg_reg_reg(asm_arm_t *as, uint rd, uint rm, uint rn) {
// strh doesn't support scaled register index
emit_al(as, 0x1a00080 | (ASM_ARM_REG_R8 << 12) | rn); // mov r8, rn, lsl #1
emit_al(as, 0x18000b0 | (rm << 16) | (rd << 12) | ASM_ARM_REG_R8); // strh rd, [rm, r8]
}
void asm_arm_strb_reg_reg_reg(asm_arm_t *as, uint rd, uint rm, uint rn) {
// strb rd, [rm, rn]
emit_al(as, 0x7c00000 | (rm << 16) | (rd << 12) | rn);
}
void asm_arm_bcc_label(asm_arm_t *as, int cond, uint label) {
assert(label < as->base.max_num_labels);
mp_uint_t dest = as->base.label_offsets[label];
mp_int_t rel = dest - as->base.code_offset;
rel -= 8; // account for instruction prefetch, PC is 8 bytes ahead of this instruction
rel >>= 2; // in ARM mode the branch target is 32-bit aligned, so the 2 LSB are omitted
if (SIGNED_FIT24(rel)) {
emit(as, cond | 0xa000000 | (rel & 0xffffff));
} else {
printf("asm_arm_bcc: branch does not fit in 24 bits\n");
}
}
void asm_arm_b_label(asm_arm_t *as, uint label) {
asm_arm_bcc_label(as, ASM_ARM_CC_AL, label);
}
void asm_arm_bl_ind(asm_arm_t *as, uint fun_id, uint reg_temp) {
// The table offset should fit into the ldr instruction
assert(fun_id < (0x1000 / 4));
emit_al(as, asm_arm_op_mov_reg(ASM_ARM_REG_LR, ASM_ARM_REG_PC)); // mov lr, pc
emit_al(as, 0x597f000 | (fun_id << 2)); // ldr pc, [r7, #fun_id*4]
}
void asm_arm_bx_reg(asm_arm_t *as, uint reg_src) {
emit_al(as, 0x012fff10 | reg_src);
}
#endif // MICROPY_EMIT_ARM