tcg/loongarch64: Add direct jump support

Similar to the ARM64, LoongArch has PC-relative instructions such as
PCADDU18I. These instructions can be used to support direct jump for
LoongArch. Additionally, if instruction "B offset" can cover the target
address(target is within ±128MB range), a single "B offset" plus a nop
will be used by "tb_target_set_jump_target".

Signed-off-by: Qi Hu <huqi@loongson.cn>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: WANG Xuerui <git@xen0n.name>
Message-Id: <20221015092754.91971-1-huqi@loongson.cn>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This commit is contained in:
Qi Hu 2022-10-15 17:27:54 +08:00 committed by Richard Henderson
parent 0585105c80
commit f072a1ae7f
2 changed files with 50 additions and 7 deletions

View File

@ -1031,6 +1031,36 @@ static void tcg_out_qemu_st(TCGContext *s, const TCGArg *args)
#endif #endif
} }
/* LoongArch uses `andi zero, zero, 0` as NOP. */
#define NOP OPC_ANDI
static void tcg_out_nop(TCGContext *s)
{
tcg_out32(s, NOP);
}
void tb_target_set_jmp_target(uintptr_t tc_ptr, uintptr_t jmp_rx,
uintptr_t jmp_rw, uintptr_t addr)
{
tcg_insn_unit i1, i2;
ptrdiff_t upper, lower;
ptrdiff_t offset = (ptrdiff_t)(addr - jmp_rx) >> 2;
if (offset == sextreg(offset, 0, 26)) {
i1 = encode_sd10k16_insn(OPC_B, offset);
i2 = NOP;
} else {
tcg_debug_assert(offset == sextreg(offset, 0, 36));
lower = (int16_t)offset;
upper = (offset - lower) >> 16;
i1 = encode_dsj20_insn(OPC_PCADDU18I, TCG_REG_TMP0, upper);
i2 = encode_djsk16_insn(OPC_JIRL, TCG_REG_ZERO, TCG_REG_TMP0, lower);
}
uint64_t pair = ((uint64_t)i2 << 32) | i1;
qatomic_set((uint64_t *)jmp_rw, pair);
flush_idcache_range(jmp_rx, jmp_rw, 8);
}
/* /*
* Entry-points * Entry-points
*/ */
@ -1058,10 +1088,20 @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc,
break; break;
case INDEX_op_goto_tb: case INDEX_op_goto_tb:
assert(s->tb_jmp_insn_offset == 0); tcg_debug_assert(s->tb_jmp_insn_offset != NULL);
/* indirect jump method */ /*
tcg_out_ld(s, TCG_TYPE_PTR, TCG_REG_TMP0, TCG_REG_ZERO, * Ensure that patch area is 8-byte aligned so that an
(uintptr_t)(s->tb_jmp_target_addr + a0)); * atomic write can be used to patch the target address.
*/
if ((uintptr_t)s->code_ptr & 7) {
tcg_out_nop(s);
}
s->tb_jmp_insn_offset[a0] = tcg_current_code_size(s);
/*
* actual branch destination will be patched by
* tb_target_set_jmp_target later
*/
tcg_out_opc_pcaddu18i(s, TCG_REG_TMP0, 0);
tcg_out_opc_jirl(s, TCG_REG_ZERO, TCG_REG_TMP0, 0); tcg_out_opc_jirl(s, TCG_REG_ZERO, TCG_REG_TMP0, 0);
set_jmp_reset_offset(s, a0); set_jmp_reset_offset(s, a0);
break; break;

View File

@ -42,7 +42,11 @@
#define TCG_TARGET_INSN_UNIT_SIZE 4 #define TCG_TARGET_INSN_UNIT_SIZE 4
#define TCG_TARGET_NB_REGS 32 #define TCG_TARGET_NB_REGS 32
#define MAX_CODE_GEN_BUFFER_SIZE SIZE_MAX /*
* PCADDU18I + JIRL sequence can give 20 + 16 + 2 = 38 bits
* signed offset, which is +/- 128 GiB.
*/
#define MAX_CODE_GEN_BUFFER_SIZE (128 * GiB)
typedef enum { typedef enum {
TCG_REG_ZERO, TCG_REG_ZERO,
@ -123,7 +127,7 @@ typedef enum {
#define TCG_TARGET_HAS_clz_i32 1 #define TCG_TARGET_HAS_clz_i32 1
#define TCG_TARGET_HAS_ctz_i32 1 #define TCG_TARGET_HAS_ctz_i32 1
#define TCG_TARGET_HAS_ctpop_i32 0 #define TCG_TARGET_HAS_ctpop_i32 0
#define TCG_TARGET_HAS_direct_jump 0 #define TCG_TARGET_HAS_direct_jump 1
#define TCG_TARGET_HAS_brcond2 0 #define TCG_TARGET_HAS_brcond2 0
#define TCG_TARGET_HAS_setcond2 0 #define TCG_TARGET_HAS_setcond2 0
#define TCG_TARGET_HAS_qemu_st8_i32 0 #define TCG_TARGET_HAS_qemu_st8_i32 0
@ -166,7 +170,6 @@ typedef enum {
#define TCG_TARGET_HAS_muluh_i64 1 #define TCG_TARGET_HAS_muluh_i64 1
#define TCG_TARGET_HAS_mulsh_i64 1 #define TCG_TARGET_HAS_mulsh_i64 1
/* not defined -- call should be eliminated at compile time */
void tb_target_set_jmp_target(uintptr_t, uintptr_t, uintptr_t, uintptr_t); void tb_target_set_jmp_target(uintptr_t, uintptr_t, uintptr_t, uintptr_t);
#define TCG_TARGET_DEFAULT_MO (0) #define TCG_TARGET_DEFAULT_MO (0)