From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:48609) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1fGpdI-000593-7M for qemu-devel@nongnu.org; Thu, 10 May 2018 13:45:37 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1fGpdG-0000vN-Ro for qemu-devel@nongnu.org; Thu, 10 May 2018 13:45:36 -0400 Received: from orth.archaic.org.uk ([2001:8b0:1d0::2]:41606) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1fGpdG-0000uH-Il for qemu-devel@nongnu.org; Thu, 10 May 2018 13:45:34 -0400 Received: from pm215 by orth.archaic.org.uk with local (Exim 4.89) (envelope-from ) id 1fGpdF-0003a1-Hl for qemu-devel@nongnu.org; Thu, 10 May 2018 18:45:33 +0100 From: Peter Maydell Date: Thu, 10 May 2018 18:45:14 +0100 Message-Id: <20180510174519.11264-17-peter.maydell@linaro.org> In-Reply-To: <20180510174519.11264-1-peter.maydell@linaro.org> References: <20180510174519.11264-1-peter.maydell@linaro.org> Subject: [Qemu-devel] [PULL 16/21] target/arm: Implement CAS and CASP List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org From: Richard Henderson Reviewed-by: Peter Maydell Signed-off-by: Richard Henderson Message-id: 20180508151437.4232-10-richard.henderson@linaro.org Signed-off-by: Peter Maydell --- target/arm/helper-a64.h | 2 + target/arm/helper-a64.c | 43 ++++++++++++++ target/arm/translate-a64.c | 119 ++++++++++++++++++++++++++++++++++++- 3 files changed, 161 insertions(+), 3 deletions(-) diff --git a/target/arm/helper-a64.h b/target/arm/helper-a64.h index ef4ddfe9d8..b8028ac98c 100644 --- a/target/arm/helper-a64.h +++ b/target/arm/helper-a64.h @@ -51,6 +51,8 @@ DEF_HELPER_FLAGS_4(paired_cmpxchg64_le_parallel, TCG_CALL_NO_WG, DEF_HELPER_FLAGS_4(paired_cmpxchg64_be, TCG_CALL_NO_WG, i64, env, i64, i64, i64) DEF_HELPER_FLAGS_4(paired_cmpxchg64_be_parallel, TCG_CALL_NO_WG, i64, env, i64, i64, i64) +DEF_HELPER_5(casp_le_parallel, void, env, i32, i64, i64, i64) +DEF_HELPER_5(casp_be_parallel, void, env, i32, i64, i64, i64) DEF_HELPER_FLAGS_3(advsimd_maxh, TCG_CALL_NO_RWG, f16, f16, f16, ptr) DEF_HELPER_FLAGS_3(advsimd_minh, TCG_CALL_NO_RWG, f16, f16, f16, ptr) DEF_HELPER_FLAGS_3(advsimd_maxnumh, TCG_CALL_NO_RWG, f16, f16, f16, ptr) diff --git a/target/arm/helper-a64.c b/target/arm/helper-a64.c index afb25ad20c..549ed3513e 100644 --- a/target/arm/helper-a64.c +++ b/target/arm/helper-a64.c @@ -636,6 +636,49 @@ uint64_t HELPER(paired_cmpxchg64_be_parallel)(CPUARMState *env, uint64_t addr, return do_paired_cmpxchg64_be(env, addr, new_lo, new_hi, true, GETPC()); } +/* Writes back the old data into Rs. */ +void HELPER(casp_le_parallel)(CPUARMState *env, uint32_t rs, uint64_t addr, + uint64_t new_lo, uint64_t new_hi) +{ + uintptr_t ra = GETPC(); +#ifndef CONFIG_ATOMIC128 + cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); +#else + Int128 oldv, cmpv, newv; + + cmpv = int128_make128(env->xregs[rs], env->xregs[rs + 1]); + newv = int128_make128(new_lo, new_hi); + + int mem_idx = cpu_mmu_index(env, false); + TCGMemOpIdx oi = make_memop_idx(MO_LEQ | MO_ALIGN_16, mem_idx); + oldv = helper_atomic_cmpxchgo_le_mmu(env, addr, cmpv, newv, oi, ra); + + env->xregs[rs] = int128_getlo(oldv); + env->xregs[rs + 1] = int128_gethi(oldv); +#endif +} + +void HELPER(casp_be_parallel)(CPUARMState *env, uint32_t rs, uint64_t addr, + uint64_t new_hi, uint64_t new_lo) +{ + uintptr_t ra = GETPC(); +#ifndef CONFIG_ATOMIC128 + cpu_loop_exit_atomic(ENV_GET_CPU(env), ra); +#else + Int128 oldv, cmpv, newv; + + cmpv = int128_make128(env->xregs[rs + 1], env->xregs[rs]); + newv = int128_make128(new_lo, new_hi); + + int mem_idx = cpu_mmu_index(env, false); + TCGMemOpIdx oi = make_memop_idx(MO_LEQ | MO_ALIGN_16, mem_idx); + oldv = helper_atomic_cmpxchgo_be_mmu(env, addr, cmpv, newv, oi, ra); + + env->xregs[rs + 1] = int128_getlo(oldv); + env->xregs[rs] = int128_gethi(oldv); +#endif +} + /* * AdvSIMD half-precision */ diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index 86989fda6c..fa60cf908f 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -2114,6 +2114,103 @@ static void gen_store_exclusive(DisasContext *s, int rd, int rt, int rt2, tcg_gen_movi_i64(cpu_exclusive_addr, -1); } +static void gen_compare_and_swap(DisasContext *s, int rs, int rt, + int rn, int size) +{ + TCGv_i64 tcg_rs = cpu_reg(s, rs); + TCGv_i64 tcg_rt = cpu_reg(s, rt); + int memidx = get_mem_index(s); + TCGv_i64 addr = cpu_reg_sp(s, rn); + + if (rn == 31) { + gen_check_sp_alignment(s); + } + tcg_gen_atomic_cmpxchg_i64(tcg_rs, addr, tcg_rs, tcg_rt, memidx, + size | MO_ALIGN | s->be_data); +} + +static void gen_compare_and_swap_pair(DisasContext *s, int rs, int rt, + int rn, int size) +{ + TCGv_i64 s1 = cpu_reg(s, rs); + TCGv_i64 s2 = cpu_reg(s, rs + 1); + TCGv_i64 t1 = cpu_reg(s, rt); + TCGv_i64 t2 = cpu_reg(s, rt + 1); + TCGv_i64 addr = cpu_reg_sp(s, rn); + int memidx = get_mem_index(s); + + if (rn == 31) { + gen_check_sp_alignment(s); + } + + if (size == 2) { + TCGv_i64 cmp = tcg_temp_new_i64(); + TCGv_i64 val = tcg_temp_new_i64(); + + if (s->be_data == MO_LE) { + tcg_gen_concat32_i64(val, t1, t2); + tcg_gen_concat32_i64(cmp, s1, s2); + } else { + tcg_gen_concat32_i64(val, t2, t1); + tcg_gen_concat32_i64(cmp, s2, s1); + } + + tcg_gen_atomic_cmpxchg_i64(cmp, addr, cmp, val, memidx, + MO_64 | MO_ALIGN | s->be_data); + tcg_temp_free_i64(val); + + if (s->be_data == MO_LE) { + tcg_gen_extr32_i64(s1, s2, cmp); + } else { + tcg_gen_extr32_i64(s2, s1, cmp); + } + tcg_temp_free_i64(cmp); + } else if (tb_cflags(s->base.tb) & CF_PARALLEL) { + TCGv_i32 tcg_rs = tcg_const_i32(rs); + + if (s->be_data == MO_LE) { + gen_helper_casp_le_parallel(cpu_env, tcg_rs, addr, t1, t2); + } else { + gen_helper_casp_be_parallel(cpu_env, tcg_rs, addr, t1, t2); + } + tcg_temp_free_i32(tcg_rs); + } else { + TCGv_i64 d1 = tcg_temp_new_i64(); + TCGv_i64 d2 = tcg_temp_new_i64(); + TCGv_i64 a2 = tcg_temp_new_i64(); + TCGv_i64 c1 = tcg_temp_new_i64(); + TCGv_i64 c2 = tcg_temp_new_i64(); + TCGv_i64 zero = tcg_const_i64(0); + + /* Load the two words, in memory order. */ + tcg_gen_qemu_ld_i64(d1, addr, memidx, + MO_64 | MO_ALIGN_16 | s->be_data); + tcg_gen_addi_i64(a2, addr, 8); + tcg_gen_qemu_ld_i64(d2, addr, memidx, MO_64 | s->be_data); + + /* Compare the two words, also in memory order. */ + tcg_gen_setcond_i64(TCG_COND_EQ, c1, d1, s1); + tcg_gen_setcond_i64(TCG_COND_EQ, c2, d2, s2); + tcg_gen_and_i64(c2, c2, c1); + + /* If compare equal, write back new data, else write back old data. */ + tcg_gen_movcond_i64(TCG_COND_NE, c1, c2, zero, t1, d1); + tcg_gen_movcond_i64(TCG_COND_NE, c2, c2, zero, t2, d2); + tcg_gen_qemu_st_i64(c1, addr, memidx, MO_64 | s->be_data); + tcg_gen_qemu_st_i64(c2, a2, memidx, MO_64 | s->be_data); + tcg_temp_free_i64(a2); + tcg_temp_free_i64(c1); + tcg_temp_free_i64(c2); + tcg_temp_free_i64(zero); + + /* Write back the data from memory to Rs. */ + tcg_gen_mov_i64(s1, d1); + tcg_gen_mov_i64(s2, d2); + tcg_temp_free_i64(d1); + tcg_temp_free_i64(d2); + } +} + /* Update the Sixty-Four bit (SF) registersize. This logic is derived * from the ARMv8 specs for LDR (Shared decode for all encodings). */ @@ -2214,10 +2311,16 @@ static void disas_ldst_excl(DisasContext *s, uint32_t insn) gen_store_exclusive(s, rs, rt, rt2, tcg_addr, size, true); return; } - /* CASP / CASPL */ + if (rt2 == 31 + && ((rt | rs) & 1) == 0 + && arm_dc_feature(s, ARM_FEATURE_V8_ATOMICS)) { + /* CASP / CASPL */ + gen_compare_and_swap_pair(s, rs, rt, rn, size | 2); + return; + } break; - case 0x6: case 0x7: /* CASP / LDXP */ + case 0x6: case 0x7: /* CASPA / LDXP */ if (size & 2) { /* LDXP / LDAXP */ if (rn == 31) { gen_check_sp_alignment(s); @@ -2230,13 +2333,23 @@ static void disas_ldst_excl(DisasContext *s, uint32_t insn) } return; } - /* CASPA / CASPAL */ + if (rt2 == 31 + && ((rt | rs) & 1) == 0 + && arm_dc_feature(s, ARM_FEATURE_V8_ATOMICS)) { + /* CASPA / CASPAL */ + gen_compare_and_swap_pair(s, rs, rt, rn, size | 2); + return; + } break; case 0xa: /* CAS */ case 0xb: /* CASL */ case 0xe: /* CASA */ case 0xf: /* CASAL */ + if (rt2 == 31 && arm_dc_feature(s, ARM_FEATURE_V8_ATOMICS)) { + gen_compare_and_swap(s, rs, rt, rn, size); + return; + } break; } unallocated_encoding(s); -- 2.17.0