The following changes since commit 13d5f87cc3b94bfccc501142df4a7b12fee3a6e7: Merge remote-tracking branch 'remotes/rth-gitlab/tags/pull-axp-20210628' into staging (2021-06-29 10:02:42 +0100) are available in the Git repository at: https://gitlab.com/rth7680/qemu.git tags/pull-tcg-20210629 for you to fetch changes up to c86bd2dc4c1d37653c27293b2dacee6bb46bb995: tcg/riscv: Remove MO_BSWAP handling (2021-06-29 10:04:57 -0700) ---------------------------------------------------------------- TranslatorOps conversion for target/avr TranslatorOps conversion for target/cris TranslatorOps conversion for target/nios2 Simple vector operations on TCGv_i32 Host signal fixes for *BSD Improvements to tcg bswap operations ---------------------------------------------------------------- LIU Zhiwei (5): tcg: Add tcg_gen_vec_add{sub}16_i32 tcg: Add tcg_gen_vec_add{sub}8_i32 tcg: Add tcg_gen_vec_shl{shr}{sar}16i_i32 tcg: Add tcg_gen_vec_shl{shr}{sar}8i_i32 tcg: Implement tcg_gen_vec_add{sub}32_tl Richard Henderson (57): target/nios2: Replace DISAS_TB_JUMP with DISAS_NORETURN target/nios2: Use global cpu_env target/nios2: Use global cpu_R target/nios2: Add DisasContextBase to DisasContext target/nios2: Convert to TranslatorOps target/nios2: Remove assignment to env in handle_instruction target/nios2: Clean up goto in handle_instruction target/nios2: Inline handle_instruction target/nios2: Use pc_next for pc + 4 target/avr: Add DisasContextBase to DisasContext target/avr: Change ctx to DisasContext* in gen_intermediate_code target/avr: Convert to TranslatorOps target/cris: Add DisasContextBase to DisasContext target/cris: Remove DISAS_SWI target/cris: Replace DISAS_TB_JUMP with DISAS_NORETURN target/cris: Mark exceptions as DISAS_NORETURN target/cris: Fix use_goto_tb target/cris: Convert to TranslatorOps target/cris: Mark helper_raise_exception noreturn target/cris: Mark static arrays const target/cris: Fold unhandled X_FLAG changes into cpustate_changed target/cris: Set cpustate_changed for rfe/rfn target/cris: Add DISAS_UPDATE_NEXT target/cris: Add DISAS_DBRANCH target/cris: Use tcg_gen_lookup_and_goto_ptr target/cris: Improve JMP_INDIRECT target/cris: Remove dc->flagx_known target/cris: Do not exit tb for X_FLAG changes tcg: Add flags argument to bswap opcodes tcg/i386: Support bswap flags tcg/aarch64: Merge tcg_out_rev{16,32,64} tcg/aarch64: Support bswap flags tcg/arm: Support bswap flags tcg/ppc: Split out tcg_out_ext{8,16,32}s tcg/ppc: Split out tcg_out_sari{32,64} tcg/ppc: Split out tcg_out_bswap16 tcg/ppc: Split out tcg_out_bswap32 tcg/ppc: Split out tcg_out_bswap64 tcg/ppc: Support bswap flags tcg/ppc: Use power10 byte-reverse instructions tcg/s390: Support bswap flags tcg/mips: Support bswap flags in tcg_out_bswap16 tcg/mips: Support bswap flags in tcg_out_bswap32 tcg/tci: Support bswap flags tcg: Handle new bswap flags during optimize tcg: Add flags argument to tcg_gen_bswap16_*, tcg_gen_bswap32_i64 tcg: Make use of bswap flags in tcg_gen_qemu_ld_* tcg: Make use of bswap flags in tcg_gen_qemu_st_* target/arm: Improve REV32 target/arm: Improve vector REV target/arm: Improve REVSH target/i386: Improve bswap translation target/sh4: Improve swap.b translation target/mips: Fix gen_mxu_s32ldd_s32lddr tcg/arm: Unset TCG_TARGET_HAS_MEMORY_BSWAP tcg/aarch64: Unset TCG_TARGET_HAS_MEMORY_BSWAP tcg/riscv: Remove MO_BSWAP handling Warner Losh (1): tcg: Use correct trap number for page faults on *BSD systems include/tcg/tcg-op-gvec.h | 43 ++++ include/tcg/tcg-op.h | 8 +- include/tcg/tcg-opc.h | 10 +- include/tcg/tcg.h | 12 + target/cris/helper.h | 2 +- tcg/aarch64/tcg-target.h | 2 +- tcg/arm/tcg-target.h | 2 +- accel/tcg/user-exec.c | 20 +- target/arm/translate-a64.c | 21 +- target/arm/translate.c | 4 +- target/avr/translate.c | 284 ++++++++++++---------- target/cris/translate.c | 515 ++++++++++++++++++++-------------------- target/i386/tcg/translate.c | 14 +- target/mips/tcg/mxu_translate.c | 6 +- target/nios2/translate.c | 318 ++++++++++++------------- target/s390x/translate.c | 4 +- target/sh4/translate.c | 3 +- tcg/optimize.c | 56 ++++- tcg/tcg-op-gvec.c | 122 ++++++++++ tcg/tcg-op.c | 143 +++++++---- tcg/tcg.c | 28 +++ tcg/tci.c | 3 +- target/cris/translate_v10.c.inc | 17 +- tcg/aarch64/tcg-target.c.inc | 125 ++++------ tcg/arm/tcg-target.c.inc | 295 ++++++++++------------- tcg/i386/tcg-target.c.inc | 20 +- tcg/mips/tcg-target.c.inc | 102 ++++---- tcg/ppc/tcg-target.c.inc | 230 ++++++++++++------ tcg/riscv/tcg-target.c.inc | 64 ++--- tcg/s390/tcg-target.c.inc | 34 ++- tcg/tci/tcg-target.c.inc | 23 +- tcg/README | 22 +- 32 files changed, 1458 insertions(+), 1094 deletions(-) The only semantic of DISAS_TB_JUMP is that we've done goto_tb, which is the same as DISAS_NORETURN -- we've exited the tb. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/nios2/translate.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/target/nios2/translate.c b/target/nios2/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/nios2/translate.c +++ b/target/nios2/translate.c @@ -XXX,XX +XXX,XX @@ /* is_jmp field values */ #define DISAS_JUMP DISAS_TARGET_0 /* only pc was modified dynamically */ #define DISAS_UPDATE DISAS_TARGET_1 /* cpu state was modified dynamically */ -#define DISAS_TB_JUMP DISAS_TARGET_2 /* only pc was modified statically */ #define INSTRUCTION_FLG(func, flags) { (func), (flags) } #define INSTRUCTION(func) \ @@ -XXX,XX +XXX,XX @@ static void jmpi(DisasContext *dc, uint32_t code, uint32_t flags) { J_TYPE(instr, code); gen_goto_tb(dc, 0, (dc->pc & 0xF0000000) | (instr.imm26 << 2)); - dc->is_jmp = DISAS_TB_JUMP; + dc->is_jmp = DISAS_NORETURN; } static void call(DisasContext *dc, uint32_t code, uint32_t flags) @@ -XXX,XX +XXX,XX @@ static void br(DisasContext *dc, uint32_t code, uint32_t flags) I_TYPE(instr, code); gen_goto_tb(dc, 0, dc->pc + 4 + (instr.imm16.s & -4)); - dc->is_jmp = DISAS_TB_JUMP; + dc->is_jmp = DISAS_NORETURN; } static void gen_bxx(DisasContext *dc, uint32_t code, uint32_t flags) @@ -XXX,XX +XXX,XX @@ static void gen_bxx(DisasContext *dc, uint32_t code, uint32_t flags) gen_goto_tb(dc, 0, dc->pc + 4); gen_set_label(l1); gen_goto_tb(dc, 1, dc->pc + 4 + (instr.imm16.s & -4)); - dc->is_jmp = DISAS_TB_JUMP; + dc->is_jmp = DISAS_NORETURN; } /* Comparison instructions */ @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) break; case DISAS_NORETURN: - case DISAS_TB_JUMP: /* nothing more to generate */ break; } -- 2.25.1 We do not need to copy this into DisasContext. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/nios2/translate.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) diff --git a/target/nios2/translate.c b/target/nios2/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/nios2/translate.c +++ b/target/nios2/translate.c @@ -XXX,XX +XXX,XX @@ } typedef struct DisasContext { - TCGv_ptr cpu_env; TCGv *cpu_R; TCGv_i32 zero; int is_jmp; @@ -XXX,XX +XXX,XX @@ static void t_gen_helper_raise_exception(DisasContext *dc, TCGv_i32 tmp = tcg_const_i32(index); tcg_gen_movi_tl(dc->cpu_R[R_PC], dc->pc); - gen_helper_raise_exception(dc->cpu_env, tmp); + gen_helper_raise_exception(cpu_env, tmp); tcg_temp_free_i32(tmp); dc->is_jmp = DISAS_NORETURN; } @@ -XXX,XX +XXX,XX @@ static void rdctl(DisasContext *dc, uint32_t code, uint32_t flags) tcg_gen_mov_tl(dc->cpu_R[instr.c], dc->cpu_R[instr.imm5 + CR_BASE]); #ifdef DEBUG_MMU TCGv_i32 tmp = tcg_const_i32(instr.imm5 + CR_BASE); - gen_helper_mmu_read_debug(dc->cpu_R[instr.c], dc->cpu_env, tmp); + gen_helper_mmu_read_debug(dc->cpu_R[instr.c], cpu_env, tmp); tcg_temp_free_i32(tmp); #endif } @@ -XXX,XX +XXX,XX @@ static void wrctl(DisasContext *dc, uint32_t code, uint32_t flags) { #if !defined(CONFIG_USER_ONLY) TCGv_i32 tmp = tcg_const_i32(instr.imm5 + CR_BASE); - gen_helper_mmu_write(dc->cpu_env, tmp, load_gpr(dc, instr.a)); + gen_helper_mmu_write(cpu_env, tmp, load_gpr(dc, instr.a)); tcg_temp_free_i32(tmp); #endif break; @@ -XXX,XX +XXX,XX @@ static void wrctl(DisasContext *dc, uint32_t code, uint32_t flags) if (tb_cflags(dc->tb) & CF_USE_ICOUNT) { gen_io_start(); } - gen_helper_check_interrupts(dc->cpu_env); + gen_helper_check_interrupts(cpu_env); dc->is_jmp = DISAS_UPDATE; } #endif @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) int num_insns; /* Initialize DC */ - dc->cpu_env = cpu_env; dc->cpu_R = cpu_R; dc->is_jmp = DISAS_NEXT; dc->pc = tb->pc; -- 2.25.1 We do not need to copy this into DisasContext. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/nios2/translate.c | 73 +++++++++++++++++++--------------------- 1 file changed, 34 insertions(+), 39 deletions(-) diff --git a/target/nios2/translate.c b/target/nios2/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/nios2/translate.c +++ b/target/nios2/translate.c @@ -XXX,XX +XXX,XX @@ } typedef struct DisasContext { - TCGv *cpu_R; TCGv_i32 zero; int is_jmp; target_ulong pc; @@ -XXX,XX +XXX,XX @@ typedef struct DisasContext { bool singlestep_enabled; } DisasContext; +static TCGv cpu_R[NUM_CORE_REGS]; + typedef struct Nios2Instruction { void (*handler)(DisasContext *dc, uint32_t code, uint32_t flags); uint32_t flags; @@ -XXX,XX +XXX,XX @@ static TCGv load_zero(DisasContext *dc) static TCGv load_gpr(DisasContext *dc, uint8_t reg) { if (likely(reg != R_ZERO)) { - return dc->cpu_R[reg]; + return cpu_R[reg]; } else { return load_zero(dc); } @@ -XXX,XX +XXX,XX @@ static void t_gen_helper_raise_exception(DisasContext *dc, { TCGv_i32 tmp = tcg_const_i32(index); - tcg_gen_movi_tl(dc->cpu_R[R_PC], dc->pc); + tcg_gen_movi_tl(cpu_R[R_PC], dc->pc); gen_helper_raise_exception(cpu_env, tmp); tcg_temp_free_i32(tmp); dc->is_jmp = DISAS_NORETURN; @@ -XXX,XX +XXX,XX @@ static void gen_goto_tb(DisasContext *dc, int n, uint32_t dest) if (use_goto_tb(dc, dest)) { tcg_gen_goto_tb(n); - tcg_gen_movi_tl(dc->cpu_R[R_PC], dest); + tcg_gen_movi_tl(cpu_R[R_PC], dest); tcg_gen_exit_tb(tb, n); } else { - tcg_gen_movi_tl(dc->cpu_R[R_PC], dest); + tcg_gen_movi_tl(cpu_R[R_PC], dest); tcg_gen_exit_tb(NULL, 0); } } @@ -XXX,XX +XXX,XX @@ static void jmpi(DisasContext *dc, uint32_t code, uint32_t flags) static void call(DisasContext *dc, uint32_t code, uint32_t flags) { - tcg_gen_movi_tl(dc->cpu_R[R_RA], dc->pc + 4); + tcg_gen_movi_tl(cpu_R[R_RA], dc->pc + 4); jmpi(dc, code, flags); } @@ -XXX,XX +XXX,XX @@ static void gen_ldx(DisasContext *dc, uint32_t code, uint32_t flags) * the Nios2 CPU. */ if (likely(instr.b != R_ZERO)) { - data = dc->cpu_R[instr.b]; + data = cpu_R[instr.b]; } else { data = tcg_temp_new(); } @@ -XXX,XX +XXX,XX @@ static void gen_bxx(DisasContext *dc, uint32_t code, uint32_t flags) I_TYPE(instr, code); TCGLabel *l1 = gen_new_label(); - tcg_gen_brcond_tl(flags, dc->cpu_R[instr.a], dc->cpu_R[instr.b], l1); + tcg_gen_brcond_tl(flags, cpu_R[instr.a], cpu_R[instr.b], l1); gen_goto_tb(dc, 0, dc->pc + 4); gen_set_label(l1); gen_goto_tb(dc, 1, dc->pc + 4 + (instr.imm16.s & -4)); @@ -XXX,XX +XXX,XX @@ static void gen_bxx(DisasContext *dc, uint32_t code, uint32_t flags) static void (fname)(DisasContext *dc, uint32_t code, uint32_t flags) \ { \ I_TYPE(instr, (code)); \ - tcg_gen_setcondi_tl(flags, (dc)->cpu_R[instr.b], (dc)->cpu_R[instr.a], \ - (op3)); \ + tcg_gen_setcondi_tl(flags, cpu_R[instr.b], cpu_R[instr.a], (op3)); \ } gen_i_cmpxx(gen_cmpxxsi, instr.imm16.s) @@ -XXX,XX +XXX,XX @@ static void (fname)(DisasContext *dc, uint32_t code, uint32_t flags) \ if (unlikely(instr.b == R_ZERO)) { /* Store to R_ZERO is ignored */ \ return; \ } else if (instr.a == R_ZERO) { /* MOVxI optimizations */ \ - tcg_gen_movi_tl(dc->cpu_R[instr.b], (resimm) ? (op3) : 0); \ + tcg_gen_movi_tl(cpu_R[instr.b], (resimm) ? (op3) : 0); \ } else { \ - tcg_gen_##insn##_tl((dc)->cpu_R[instr.b], (dc)->cpu_R[instr.a], \ - (op3)); \ + tcg_gen_##insn##_tl(cpu_R[instr.b], cpu_R[instr.a], (op3)); \ } \ } @@ -XXX,XX +XXX,XX @@ static const Nios2Instruction i_type_instructions[] = { */ static void eret(DisasContext *dc, uint32_t code, uint32_t flags) { - tcg_gen_mov_tl(dc->cpu_R[CR_STATUS], dc->cpu_R[CR_ESTATUS]); - tcg_gen_mov_tl(dc->cpu_R[R_PC], dc->cpu_R[R_EA]); + tcg_gen_mov_tl(cpu_R[CR_STATUS], cpu_R[CR_ESTATUS]); + tcg_gen_mov_tl(cpu_R[R_PC], cpu_R[R_EA]); dc->is_jmp = DISAS_JUMP; } @@ -XXX,XX +XXX,XX @@ static void eret(DisasContext *dc, uint32_t code, uint32_t flags) /* PC <- ra */ static void ret(DisasContext *dc, uint32_t code, uint32_t flags) { - tcg_gen_mov_tl(dc->cpu_R[R_PC], dc->cpu_R[R_RA]); + tcg_gen_mov_tl(cpu_R[R_PC], cpu_R[R_RA]); dc->is_jmp = DISAS_JUMP; } @@ -XXX,XX +XXX,XX @@ static void ret(DisasContext *dc, uint32_t code, uint32_t flags) /* PC <- ba */ static void bret(DisasContext *dc, uint32_t code, uint32_t flags) { - tcg_gen_mov_tl(dc->cpu_R[R_PC], dc->cpu_R[R_BA]); + tcg_gen_mov_tl(cpu_R[R_PC], cpu_R[R_BA]); dc->is_jmp = DISAS_JUMP; } @@ -XXX,XX +XXX,XX @@ static void jmp(DisasContext *dc, uint32_t code, uint32_t flags) { R_TYPE(instr, code); - tcg_gen_mov_tl(dc->cpu_R[R_PC], load_gpr(dc, instr.a)); + tcg_gen_mov_tl(cpu_R[R_PC], load_gpr(dc, instr.a)); dc->is_jmp = DISAS_JUMP; } @@ -XXX,XX +XXX,XX @@ static void nextpc(DisasContext *dc, uint32_t code, uint32_t flags) R_TYPE(instr, code); if (likely(instr.c != R_ZERO)) { - tcg_gen_movi_tl(dc->cpu_R[instr.c], dc->pc + 4); + tcg_gen_movi_tl(cpu_R[instr.c], dc->pc + 4); } } @@ -XXX,XX +XXX,XX @@ static void callr(DisasContext *dc, uint32_t code, uint32_t flags) { R_TYPE(instr, code); - tcg_gen_mov_tl(dc->cpu_R[R_PC], load_gpr(dc, instr.a)); - tcg_gen_movi_tl(dc->cpu_R[R_RA], dc->pc + 4); + tcg_gen_mov_tl(cpu_R[R_PC], load_gpr(dc, instr.a)); + tcg_gen_movi_tl(cpu_R[R_RA], dc->pc + 4); dc->is_jmp = DISAS_JUMP; } @@ -XXX,XX +XXX,XX @@ static void rdctl(DisasContext *dc, uint32_t code, uint32_t flags) { #if !defined(CONFIG_USER_ONLY) if (likely(instr.c != R_ZERO)) { - tcg_gen_mov_tl(dc->cpu_R[instr.c], dc->cpu_R[instr.imm5 + CR_BASE]); + tcg_gen_mov_tl(cpu_R[instr.c], cpu_R[instr.imm5 + CR_BASE]); #ifdef DEBUG_MMU TCGv_i32 tmp = tcg_const_i32(instr.imm5 + CR_BASE); - gen_helper_mmu_read_debug(dc->cpu_R[instr.c], cpu_env, tmp); + gen_helper_mmu_read_debug(cpu_R[instr.c], cpu_env, tmp); tcg_temp_free_i32(tmp); #endif } @@ -XXX,XX +XXX,XX @@ static void rdctl(DisasContext *dc, uint32_t code, uint32_t flags) default: if (likely(instr.c != R_ZERO)) { - tcg_gen_mov_tl(dc->cpu_R[instr.c], dc->cpu_R[instr.imm5 + CR_BASE]); + tcg_gen_mov_tl(cpu_R[instr.c], cpu_R[instr.imm5 + CR_BASE]); } break; } @@ -XXX,XX +XXX,XX @@ static void wrctl(DisasContext *dc, uint32_t code, uint32_t flags) } default: - tcg_gen_mov_tl(dc->cpu_R[instr.imm5 + CR_BASE], load_gpr(dc, instr.a)); + tcg_gen_mov_tl(cpu_R[instr.imm5 + CR_BASE], load_gpr(dc, instr.a)); break; } @@ -XXX,XX +XXX,XX @@ static void gen_cmpxx(DisasContext *dc, uint32_t code, uint32_t flags) { R_TYPE(instr, code); if (likely(instr.c != R_ZERO)) { - tcg_gen_setcond_tl(flags, dc->cpu_R[instr.c], dc->cpu_R[instr.a], - dc->cpu_R[instr.b]); + tcg_gen_setcond_tl(flags, cpu_R[instr.c], cpu_R[instr.a], + cpu_R[instr.b]); } } @@ -XXX,XX +XXX,XX @@ static void (fname)(DisasContext *dc, uint32_t code, uint32_t flags) \ { \ R_TYPE(instr, (code)); \ if (likely(instr.c != R_ZERO)) { \ - tcg_gen_##insn((dc)->cpu_R[instr.c], load_gpr((dc), instr.a), \ - (op3)); \ + tcg_gen_##insn(cpu_R[instr.c], load_gpr((dc), instr.a), (op3)); \ } \ } @@ -XXX,XX +XXX,XX @@ static void (fname)(DisasContext *dc, uint32_t code, uint32_t flags) \ R_TYPE(instr, (code)); \ if (likely(instr.c != R_ZERO)) { \ TCGv t0 = tcg_temp_new(); \ - tcg_gen_##insn(t0, dc->cpu_R[instr.c], \ - load_gpr(dc, instr.a), load_gpr(dc, instr.b)); \ + tcg_gen_##insn(t0, cpu_R[instr.c], \ + load_gpr(dc, instr.a), load_gpr(dc, instr.b)); \ tcg_temp_free(t0); \ } \ } @@ -XXX,XX +XXX,XX @@ static void (fname)(DisasContext *dc, uint32_t code, uint32_t flags) \ if (likely(instr.c != R_ZERO)) { \ TCGv t0 = tcg_temp_new(); \ tcg_gen_andi_tl(t0, load_gpr((dc), instr.b), 31); \ - tcg_gen_##insn((dc)->cpu_R[instr.c], load_gpr((dc), instr.a), t0); \ + tcg_gen_##insn(cpu_R[instr.c], load_gpr((dc), instr.a), t0); \ tcg_temp_free(t0); \ } \ } @@ -XXX,XX +XXX,XX @@ static void divs(DisasContext *dc, uint32_t code, uint32_t flags) tcg_gen_or_tl(t2, t2, t3); tcg_gen_movi_tl(t3, 0); tcg_gen_movcond_tl(TCG_COND_NE, t1, t2, t3, t2, t1); - tcg_gen_div_tl(dc->cpu_R[instr.c], t0, t1); - tcg_gen_ext32s_tl(dc->cpu_R[instr.c], dc->cpu_R[instr.c]); + tcg_gen_div_tl(cpu_R[instr.c], t0, t1); + tcg_gen_ext32s_tl(cpu_R[instr.c], cpu_R[instr.c]); tcg_temp_free(t3); tcg_temp_free(t2); @@ -XXX,XX +XXX,XX @@ static void divu(DisasContext *dc, uint32_t code, uint32_t flags) tcg_gen_ext32u_tl(t0, load_gpr(dc, instr.a)); tcg_gen_ext32u_tl(t1, load_gpr(dc, instr.b)); tcg_gen_movcond_tl(TCG_COND_EQ, t1, t1, t2, t3, t1); - tcg_gen_divu_tl(dc->cpu_R[instr.c], t0, t1); - tcg_gen_ext32s_tl(dc->cpu_R[instr.c], dc->cpu_R[instr.c]); + tcg_gen_divu_tl(cpu_R[instr.c], t0, t1); + tcg_gen_ext32s_tl(cpu_R[instr.c], cpu_R[instr.c]); tcg_temp_free(t3); tcg_temp_free(t2); @@ -XXX,XX +XXX,XX @@ static const char * const regnames[] = { "rpc" }; -static TCGv cpu_R[NUM_CORE_REGS]; - #include "exec/gen-icount.h" static void gen_exception(DisasContext *dc, uint32_t excp) @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) int num_insns; /* Initialize DC */ - dc->cpu_R = cpu_R; dc->is_jmp = DISAS_NEXT; dc->pc = tb->pc; dc->tb = tb; -- 2.25.1 Migrate the is_jmp, tb and singlestep_enabled fields from DisasContext into the base. Use pc_first instead of tb->pc. Increment pc_next prior to decode, leaving the address of the current insn in dc->pc. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/nios2/translate.c | 70 +++++++++++++++++++++------------------- 1 file changed, 36 insertions(+), 34 deletions(-) diff --git a/target/nios2/translate.c b/target/nios2/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/nios2/translate.c +++ b/target/nios2/translate.c @@ -XXX,XX +XXX,XX @@ } typedef struct DisasContext { + DisasContextBase base; TCGv_i32 zero; - int is_jmp; target_ulong pc; - TranslationBlock *tb; int mem_idx; - bool singlestep_enabled; } DisasContext; static TCGv cpu_R[NUM_CORE_REGS]; @@ -XXX,XX +XXX,XX @@ static void t_gen_helper_raise_exception(DisasContext *dc, tcg_gen_movi_tl(cpu_R[R_PC], dc->pc); gen_helper_raise_exception(cpu_env, tmp); tcg_temp_free_i32(tmp); - dc->is_jmp = DISAS_NORETURN; + dc->base.is_jmp = DISAS_NORETURN; } static bool use_goto_tb(DisasContext *dc, uint32_t dest) { - if (unlikely(dc->singlestep_enabled)) { + if (unlikely(dc->base.singlestep_enabled)) { return false; } #ifndef CONFIG_USER_ONLY - return (dc->tb->pc & TARGET_PAGE_MASK) == (dest & TARGET_PAGE_MASK); + return (dc->base.pc_first & TARGET_PAGE_MASK) == (dest & TARGET_PAGE_MASK); #else return true; #endif @@ -XXX,XX +XXX,XX @@ static bool use_goto_tb(DisasContext *dc, uint32_t dest) static void gen_goto_tb(DisasContext *dc, int n, uint32_t dest) { - TranslationBlock *tb = dc->tb; + const TranslationBlock *tb = dc->base.tb; if (use_goto_tb(dc, dest)) { tcg_gen_goto_tb(n); @@ -XXX,XX +XXX,XX @@ static void gen_excp(DisasContext *dc, uint32_t code, uint32_t flags) static void gen_check_supervisor(DisasContext *dc) { - if (dc->tb->flags & CR_STATUS_U) { + if (dc->base.tb->flags & CR_STATUS_U) { /* CPU in user mode, privileged instruction called, stop. */ t_gen_helper_raise_exception(dc, EXCP_SUPERI); } @@ -XXX,XX +XXX,XX @@ static void jmpi(DisasContext *dc, uint32_t code, uint32_t flags) { J_TYPE(instr, code); gen_goto_tb(dc, 0, (dc->pc & 0xF0000000) | (instr.imm26 << 2)); - dc->is_jmp = DISAS_NORETURN; + dc->base.is_jmp = DISAS_NORETURN; } static void call(DisasContext *dc, uint32_t code, uint32_t flags) @@ -XXX,XX +XXX,XX @@ static void br(DisasContext *dc, uint32_t code, uint32_t flags) I_TYPE(instr, code); gen_goto_tb(dc, 0, dc->pc + 4 + (instr.imm16.s & -4)); - dc->is_jmp = DISAS_NORETURN; + dc->base.is_jmp = DISAS_NORETURN; } static void gen_bxx(DisasContext *dc, uint32_t code, uint32_t flags) @@ -XXX,XX +XXX,XX @@ static void gen_bxx(DisasContext *dc, uint32_t code, uint32_t flags) gen_goto_tb(dc, 0, dc->pc + 4); gen_set_label(l1); gen_goto_tb(dc, 1, dc->pc + 4 + (instr.imm16.s & -4)); - dc->is_jmp = DISAS_NORETURN; + dc->base.is_jmp = DISAS_NORETURN; } /* Comparison instructions */ @@ -XXX,XX +XXX,XX @@ static void eret(DisasContext *dc, uint32_t code, uint32_t flags) tcg_gen_mov_tl(cpu_R[CR_STATUS], cpu_R[CR_ESTATUS]); tcg_gen_mov_tl(cpu_R[R_PC], cpu_R[R_EA]); - dc->is_jmp = DISAS_JUMP; + dc->base.is_jmp = DISAS_JUMP; } /* PC <- ra */ @@ -XXX,XX +XXX,XX @@ static void ret(DisasContext *dc, uint32_t code, uint32_t flags) { tcg_gen_mov_tl(cpu_R[R_PC], cpu_R[R_RA]); - dc->is_jmp = DISAS_JUMP; + dc->base.is_jmp = DISAS_JUMP; } /* PC <- ba */ @@ -XXX,XX +XXX,XX @@ static void bret(DisasContext *dc, uint32_t code, uint32_t flags) { tcg_gen_mov_tl(cpu_R[R_PC], cpu_R[R_BA]); - dc->is_jmp = DISAS_JUMP; + dc->base.is_jmp = DISAS_JUMP; } /* PC <- rA */ @@ -XXX,XX +XXX,XX @@ static void jmp(DisasContext *dc, uint32_t code, uint32_t flags) tcg_gen_mov_tl(cpu_R[R_PC], load_gpr(dc, instr.a)); - dc->is_jmp = DISAS_JUMP; + dc->base.is_jmp = DISAS_JUMP; } /* rC <- PC + 4 */ @@ -XXX,XX +XXX,XX @@ static void callr(DisasContext *dc, uint32_t code, uint32_t flags) tcg_gen_mov_tl(cpu_R[R_PC], load_gpr(dc, instr.a)); tcg_gen_movi_tl(cpu_R[R_RA], dc->pc + 4); - dc->is_jmp = DISAS_JUMP; + dc->base.is_jmp = DISAS_JUMP; } /* rC <- ctlN */ @@ -XXX,XX +XXX,XX @@ static void wrctl(DisasContext *dc, uint32_t code, uint32_t flags) /* If interrupts were enabled using WRCTL, trigger them. */ #if !defined(CONFIG_USER_ONLY) if ((instr.imm5 + CR_BASE) == CR_STATUS) { - if (tb_cflags(dc->tb) & CF_USE_ICOUNT) { + if (tb_cflags(dc->base.tb) & CF_USE_ICOUNT) { gen_io_start(); } gen_helper_check_interrupts(cpu_env); - dc->is_jmp = DISAS_UPDATE; + dc->base.is_jmp = DISAS_UPDATE; } #endif } @@ -XXX,XX +XXX,XX @@ static void gen_exception(DisasContext *dc, uint32_t excp) tcg_gen_movi_tl(cpu_R[R_PC], dc->pc); gen_helper_raise_exception(cpu_env, tmp); tcg_temp_free_i32(tmp); - dc->is_jmp = DISAS_NORETURN; + dc->base.is_jmp = DISAS_NORETURN; } /* generate intermediate code for basic block 'tb'. */ @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) int num_insns; /* Initialize DC */ - dc->is_jmp = DISAS_NEXT; - dc->pc = tb->pc; - dc->tb = tb; + + dc->base.tb = tb; + dc->base.singlestep_enabled = cs->singlestep_enabled; + dc->base.is_jmp = DISAS_NEXT; + dc->base.pc_first = tb->pc; + dc->base.pc_next = tb->pc; + dc->mem_idx = cpu_mmu_index(env, false); - dc->singlestep_enabled = cs->singlestep_enabled; /* Set up instruction counts */ num_insns = 0; @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) gen_tb_start(tb); do { - tcg_gen_insn_start(dc->pc); + tcg_gen_insn_start(dc->base.pc_next); num_insns++; - if (unlikely(cpu_breakpoint_test(cs, dc->pc, BP_ANY))) { + if (unlikely(cpu_breakpoint_test(cs, dc->base.pc_next, BP_ANY))) { gen_exception(dc, EXCP_DEBUG); /* The address covered by the breakpoint must be included in [tb->pc, tb->pc + tb->size) in order to for it to be @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) gen_io_start(); } + dc->pc = dc->base.pc_next; + dc->base.pc_next += 4; + /* Decode an instruction */ handle_instruction(dc, env); - dc->pc += 4; - /* Translation stops when a conditional branch is encountered. * Otherwise the subsequent code could get translated several times. * Also stop translation when a page boundary is reached. This * ensures prefetch aborts occur at the right place. */ - } while (!dc->is_jmp && + } while (!dc->base.is_jmp && !tcg_op_buf_full() && num_insns < max_insns); /* Indicate where the next block should start */ - switch (dc->is_jmp) { + switch (dc->base.is_jmp) { case DISAS_NEXT: case DISAS_UPDATE: /* Save the current PC back into the CPU register */ - tcg_gen_movi_tl(cpu_R[R_PC], dc->pc); + tcg_gen_movi_tl(cpu_R[R_PC], dc->base.pc_next); tcg_gen_exit_tb(NULL, 0); break; @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) gen_tb_end(tb, num_insns); /* Mark instruction starts for the final generated instruction */ - tb->size = dc->pc - tb->pc; + tb->size = dc->base.pc_next - dc->base.pc_first; tb->icount = num_insns; #ifdef DEBUG_DISAS if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM) - && qemu_log_in_addr_range(tb->pc)) { + && qemu_log_in_addr_range(dc->base.pc_first)) { FILE *logfile = qemu_log_lock(); - qemu_log("IN: %s\n", lookup_symbol(tb->pc)); - log_target_disas(cs, tb->pc, dc->pc - tb->pc); + qemu_log("IN: %s\n", lookup_symbol(dc->base.pc_first)); + log_target_disas(cs, tb->pc, tb->size); qemu_log("\n"); qemu_log_unlock(logfile); } -- 2.25.1 Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/nios2/translate.c | 128 ++++++++++++++++++++------------------- 1 file changed, 65 insertions(+), 63 deletions(-) diff --git a/target/nios2/translate.c b/target/nios2/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/nios2/translate.c +++ b/target/nios2/translate.c @@ -XXX,XX +XXX,XX @@ static void gen_exception(DisasContext *dc, uint32_t excp) } /* generate intermediate code for basic block 'tb'. */ -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) +static void nios2_tr_init_disas_context(DisasContextBase *dcbase, CPUState *cs) { + DisasContext *dc = container_of(dcbase, DisasContext, base); CPUNios2State *env = cs->env_ptr; - DisasContext dc1, *dc = &dc1; - int num_insns; - - /* Initialize DC */ - - dc->base.tb = tb; - dc->base.singlestep_enabled = cs->singlestep_enabled; - dc->base.is_jmp = DISAS_NEXT; - dc->base.pc_first = tb->pc; - dc->base.pc_next = tb->pc; + int page_insns; dc->mem_idx = cpu_mmu_index(env, false); - /* Set up instruction counts */ - num_insns = 0; - if (max_insns > 1) { - int page_insns = (TARGET_PAGE_SIZE - (tb->pc & ~TARGET_PAGE_MASK)) / 4; - if (max_insns > page_insns) { - max_insns = page_insns; - } - } + /* Bound the number of insns to execute to those left on the page. */ + page_insns = -(dc->base.pc_first | TARGET_PAGE_MASK) / 4; + dc->base.max_insns = MIN(page_insns, dc->base.max_insns); +} - gen_tb_start(tb); - do { - tcg_gen_insn_start(dc->base.pc_next); - num_insns++; +static void nios2_tr_tb_start(DisasContextBase *db, CPUState *cs) +{ +} - if (unlikely(cpu_breakpoint_test(cs, dc->base.pc_next, BP_ANY))) { - gen_exception(dc, EXCP_DEBUG); - /* The address covered by the breakpoint must be included in - [tb->pc, tb->pc + tb->size) in order to for it to be - properly cleared -- thus we increment the PC here so that - the logic setting tb->size below does the right thing. */ - dc->pc += 4; - break; - } +static void nios2_tr_insn_start(DisasContextBase *dcbase, CPUState *cs) +{ + tcg_gen_insn_start(dcbase->pc_next); +} - if (num_insns == max_insns && (tb_cflags(tb) & CF_LAST_IO)) { - gen_io_start(); - } +static bool nios2_tr_breakpoint_check(DisasContextBase *dcbase, CPUState *cs, + const CPUBreakpoint *bp) +{ + DisasContext *dc = container_of(dcbase, DisasContext, base); - dc->pc = dc->base.pc_next; - dc->base.pc_next += 4; + gen_exception(dc, EXCP_DEBUG); + /* + * The address covered by the breakpoint must be included in + * [tb->pc, tb->pc + tb->size) in order to for it to be + * properly cleared -- thus we increment the PC here so that + * the logic setting tb->size below does the right thing. + */ + dc->base.pc_next += 4; + return true; +} - /* Decode an instruction */ - handle_instruction(dc, env); +static void nios2_tr_translate_insn(DisasContextBase *dcbase, CPUState *cs) +{ + DisasContext *dc = container_of(dcbase, DisasContext, base); + CPUNios2State *env = cs->env_ptr; - /* Translation stops when a conditional branch is encountered. - * Otherwise the subsequent code could get translated several times. - * Also stop translation when a page boundary is reached. This - * ensures prefetch aborts occur at the right place. */ - } while (!dc->base.is_jmp && - !tcg_op_buf_full() && - num_insns < max_insns); + dc->pc = dc->base.pc_next; + dc->base.pc_next += 4; + + /* Decode an instruction */ + handle_instruction(dc, env); +} + +static void nios2_tr_tb_stop(DisasContextBase *dcbase, CPUState *cs) +{ + DisasContext *dc = container_of(dcbase, DisasContext, base); /* Indicate where the next block should start */ switch (dc->base.is_jmp) { - case DISAS_NEXT: + case DISAS_TOO_MANY: case DISAS_UPDATE: /* Save the current PC back into the CPU register */ tcg_gen_movi_tl(cpu_R[R_PC], dc->base.pc_next); tcg_gen_exit_tb(NULL, 0); break; - default: case DISAS_JUMP: /* The jump will already have updated the PC register */ tcg_gen_exit_tb(NULL, 0); @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) case DISAS_NORETURN: /* nothing more to generate */ break; + + default: + g_assert_not_reached(); } +} - /* End off the block */ - gen_tb_end(tb, num_insns); +static void nios2_tr_disas_log(const DisasContextBase *dcbase, CPUState *cpu) +{ + qemu_log("IN: %s\n", lookup_symbol(dcbase->pc_first)); + log_target_disas(cpu, dcbase->pc_first, dcbase->tb->size); +} - /* Mark instruction starts for the final generated instruction */ - tb->size = dc->base.pc_next - dc->base.pc_first; - tb->icount = num_insns; +static const TranslatorOps nios2_tr_ops = { + .init_disas_context = nios2_tr_init_disas_context, + .tb_start = nios2_tr_tb_start, + .insn_start = nios2_tr_insn_start, + .breakpoint_check = nios2_tr_breakpoint_check, + .translate_insn = nios2_tr_translate_insn, + .tb_stop = nios2_tr_tb_stop, + .disas_log = nios2_tr_disas_log, +}; -#ifdef DEBUG_DISAS - if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM) - && qemu_log_in_addr_range(dc->base.pc_first)) { - FILE *logfile = qemu_log_lock(); - qemu_log("IN: %s\n", lookup_symbol(dc->base.pc_first)); - log_target_disas(cs, tb->pc, tb->size); - qemu_log("\n"); - qemu_log_unlock(logfile); - } -#endif +void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) +{ + DisasContext dc; + translator_loop(&nios2_tr_ops, &dc.base, cs, tb, max_insns); } void nios2_cpu_dump_state(CPUState *cs, FILE *f, int flags) -- 2.25.1 Direct assignments to env during translation do not work. As it happens, the only way we can get here is if env->pc is already set to dc->pc. We will trap on the first insn we execute anywhere on the page. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/nios2/translate.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/target/nios2/translate.c b/target/nios2/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/nios2/translate.c +++ b/target/nios2/translate.c @@ -XXX,XX +XXX,XX @@ static void handle_instruction(DisasContext *dc, CPUNios2State *env) uint32_t code; uint8_t op; const Nios2Instruction *instr; + #if defined(CONFIG_USER_ONLY) /* FIXME: Is this needed ? */ if (dc->pc >= 0x1000 && dc->pc < 0x2000) { - env->regs[R_PC] = dc->pc; t_gen_helper_raise_exception(dc, 0xaa); return; } #endif + code = cpu_ldl_code(env, dc->pc); op = get_opcode(code); -- 2.25.1 Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/nios2/translate.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) diff --git a/target/nios2/translate.c b/target/nios2/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/nios2/translate.c +++ b/target/nios2/translate.c @@ -XXX,XX +XXX,XX @@ static void handle_instruction(DisasContext *dc, CPUNios2State *env) op = get_opcode(code); if (unlikely(op >= ARRAY_SIZE(i_type_instructions))) { - goto illegal_op; + t_gen_helper_raise_exception(dc, EXCP_ILLEGAL); + return; } dc->zero = NULL; @@ -XXX,XX +XXX,XX @@ static void handle_instruction(DisasContext *dc, CPUNios2State *env) if (dc->zero) { tcg_temp_free(dc->zero); } - - return; - -illegal_op: - t_gen_helper_raise_exception(dc, EXCP_ILLEGAL); } static const char * const regnames[] = { -- 2.25.1 Move handle_instruction into nios2_tr_translate_insn as the only caller. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/nios2/translate.c | 66 +++++++++++++++++++--------------------- 1 file changed, 31 insertions(+), 35 deletions(-) diff --git a/target/nios2/translate.c b/target/nios2/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/nios2/translate.c +++ b/target/nios2/translate.c @@ -XXX,XX +XXX,XX @@ illegal_op: t_gen_helper_raise_exception(dc, EXCP_ILLEGAL); } -static void handle_instruction(DisasContext *dc, CPUNios2State *env) -{ - uint32_t code; - uint8_t op; - const Nios2Instruction *instr; - -#if defined(CONFIG_USER_ONLY) - /* FIXME: Is this needed ? */ - if (dc->pc >= 0x1000 && dc->pc < 0x2000) { - t_gen_helper_raise_exception(dc, 0xaa); - return; - } -#endif - - code = cpu_ldl_code(env, dc->pc); - op = get_opcode(code); - - if (unlikely(op >= ARRAY_SIZE(i_type_instructions))) { - t_gen_helper_raise_exception(dc, EXCP_ILLEGAL); - return; - } - - dc->zero = NULL; - - instr = &i_type_instructions[op]; - instr->handler(dc, code, instr->flags); - - if (dc->zero) { - tcg_temp_free(dc->zero); - } -} - static const char * const regnames[] = { "zero", "at", "r2", "r3", "r4", "r5", "r6", "r7", @@ -XXX,XX +XXX,XX @@ static void nios2_tr_translate_insn(DisasContextBase *dcbase, CPUState *cs) { DisasContext *dc = container_of(dcbase, DisasContext, base); CPUNios2State *env = cs->env_ptr; + const Nios2Instruction *instr; + uint32_t code, pc; + uint8_t op; - dc->pc = dc->base.pc_next; - dc->base.pc_next += 4; + pc = dc->base.pc_next; + dc->pc = pc; + dc->base.pc_next = pc + 4; /* Decode an instruction */ - handle_instruction(dc, env); + +#if defined(CONFIG_USER_ONLY) + /* FIXME: Is this needed ? */ + if (pc >= 0x1000 && pc < 0x2000) { + t_gen_helper_raise_exception(dc, 0xaa); + return; + } +#endif + + code = cpu_ldl_code(env, pc); + op = get_opcode(code); + + if (unlikely(op >= ARRAY_SIZE(i_type_instructions))) { + t_gen_helper_raise_exception(dc, EXCP_ILLEGAL); + return; + } + + dc->zero = NULL; + + instr = &i_type_instructions[op]; + instr->handler(dc, code, instr->flags); + + if (dc->zero) { + tcg_temp_free(dc->zero); + } } static void nios2_tr_tb_stop(DisasContextBase *dcbase, CPUState *cs) -- 2.25.1 We have pre-computed the next instruction address into dc->base.pc_next, so we might as well use it. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Suggested-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/nios2/translate.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/target/nios2/translate.c b/target/nios2/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/nios2/translate.c +++ b/target/nios2/translate.c @@ -XXX,XX +XXX,XX @@ static void jmpi(DisasContext *dc, uint32_t code, uint32_t flags) static void call(DisasContext *dc, uint32_t code, uint32_t flags) { - tcg_gen_movi_tl(cpu_R[R_RA], dc->pc + 4); + tcg_gen_movi_tl(cpu_R[R_RA], dc->base.pc_next); jmpi(dc, code, flags); } @@ -XXX,XX +XXX,XX @@ static void br(DisasContext *dc, uint32_t code, uint32_t flags) { I_TYPE(instr, code); - gen_goto_tb(dc, 0, dc->pc + 4 + (instr.imm16.s & -4)); + gen_goto_tb(dc, 0, dc->base.pc_next + (instr.imm16.s & -4)); dc->base.is_jmp = DISAS_NORETURN; } @@ -XXX,XX +XXX,XX @@ static void gen_bxx(DisasContext *dc, uint32_t code, uint32_t flags) TCGLabel *l1 = gen_new_label(); tcg_gen_brcond_tl(flags, cpu_R[instr.a], cpu_R[instr.b], l1); - gen_goto_tb(dc, 0, dc->pc + 4); + gen_goto_tb(dc, 0, dc->base.pc_next); gen_set_label(l1); - gen_goto_tb(dc, 1, dc->pc + 4 + (instr.imm16.s & -4)); + gen_goto_tb(dc, 1, dc->base.pc_next + (instr.imm16.s & -4)); dc->base.is_jmp = DISAS_NORETURN; } @@ -XXX,XX +XXX,XX @@ static void nextpc(DisasContext *dc, uint32_t code, uint32_t flags) R_TYPE(instr, code); if (likely(instr.c != R_ZERO)) { - tcg_gen_movi_tl(cpu_R[instr.c], dc->pc + 4); + tcg_gen_movi_tl(cpu_R[instr.c], dc->base.pc_next); } } @@ -XXX,XX +XXX,XX @@ static void callr(DisasContext *dc, uint32_t code, uint32_t flags) R_TYPE(instr, code); tcg_gen_mov_tl(cpu_R[R_PC], load_gpr(dc, instr.a)); - tcg_gen_movi_tl(cpu_R[R_RA], dc->pc + 4); + tcg_gen_movi_tl(cpu_R[R_RA], dc->base.pc_next); dc->base.is_jmp = DISAS_JUMP; } -- 2.25.1 Migrate the bstate, tb and singlestep_enabled fields from DisasContext into the base. Tested-by: Michael Rolnik <mrolnik@gmail.com> Reviewed-by: Michael Rolnik <mrolnik@gmail.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/avr/translate.c | 58 +++++++++++++++++++++--------------------- 1 file changed, 29 insertions(+), 29 deletions(-) diff --git a/target/avr/translate.c b/target/avr/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/avr/translate.c +++ b/target/avr/translate.c @@ -XXX,XX +XXX,XX @@ typedef struct DisasContext DisasContext; /* This is the state at translation time. */ struct DisasContext { - TranslationBlock *tb; + DisasContextBase base; CPUAVRState *env; CPUState *cs; @@ -XXX,XX +XXX,XX @@ struct DisasContext { /* Routine used to access memory */ int memidx; - int bstate; - int singlestep; /* * some AVR instructions can make the following instruction to be skipped @@ -XXX,XX +XXX,XX @@ static bool avr_have_feature(DisasContext *ctx, int feature) { if (!avr_feature(ctx->env, feature)) { gen_helper_unsupported(cpu_env); - ctx->bstate = DISAS_NORETURN; + ctx->base.is_jmp = DISAS_NORETURN; return false; } return true; @@ -XXX,XX +XXX,XX @@ static void gen_jmp_ez(DisasContext *ctx) { tcg_gen_deposit_tl(cpu_pc, cpu_r[30], cpu_r[31], 8, 8); tcg_gen_or_tl(cpu_pc, cpu_pc, cpu_eind); - ctx->bstate = DISAS_LOOKUP; + ctx->base.is_jmp = DISAS_LOOKUP; } static void gen_jmp_z(DisasContext *ctx) { tcg_gen_deposit_tl(cpu_pc, cpu_r[30], cpu_r[31], 8, 8); - ctx->bstate = DISAS_LOOKUP; + ctx->base.is_jmp = DISAS_LOOKUP; } static void gen_push_ret(DisasContext *ctx, int ret) @@ -XXX,XX +XXX,XX @@ static void gen_pop_ret(DisasContext *ctx, TCGv ret) static void gen_goto_tb(DisasContext *ctx, int n, target_ulong dest) { - TranslationBlock *tb = ctx->tb; + const TranslationBlock *tb = ctx->base.tb; - if (ctx->singlestep == 0) { + if (!ctx->base.singlestep_enabled) { tcg_gen_goto_tb(n); tcg_gen_movi_i32(cpu_pc, dest); tcg_gen_exit_tb(tb, n); @@ -XXX,XX +XXX,XX @@ static void gen_goto_tb(DisasContext *ctx, int n, target_ulong dest) gen_helper_debug(cpu_env); tcg_gen_exit_tb(NULL, 0); } - ctx->bstate = DISAS_NORETURN; + ctx->base.is_jmp = DISAS_NORETURN; } /* @@ -XXX,XX +XXX,XX @@ static bool trans_RET(DisasContext *ctx, arg_RET *a) { gen_pop_ret(ctx, cpu_pc); - ctx->bstate = DISAS_LOOKUP; + ctx->base.is_jmp = DISAS_LOOKUP; return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_RETI(DisasContext *ctx, arg_RETI *a) tcg_gen_movi_tl(cpu_If, 1); /* Need to return to main loop to re-evaluate interrupts. */ - ctx->bstate = DISAS_EXIT; + ctx->base.is_jmp = DISAS_EXIT; return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_BRBC(DisasContext *ctx, arg_BRBC *a) gen_goto_tb(ctx, 0, ctx->npc + a->imm); gen_set_label(not_taken); - ctx->bstate = DISAS_CHAIN; + ctx->base.is_jmp = DISAS_CHAIN; return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_BRBS(DisasContext *ctx, arg_BRBS *a) gen_goto_tb(ctx, 0, ctx->npc + a->imm); gen_set_label(not_taken); - ctx->bstate = DISAS_CHAIN; + ctx->base.is_jmp = DISAS_CHAIN; return true; } @@ -XXX,XX +XXX,XX @@ static TCGv gen_get_zaddr(void) */ static void gen_data_store(DisasContext *ctx, TCGv data, TCGv addr) { - if (ctx->tb->flags & TB_FLAGS_FULL_ACCESS) { + if (ctx->base.tb->flags & TB_FLAGS_FULL_ACCESS) { gen_helper_fullwr(cpu_env, data, addr); } else { tcg_gen_qemu_st8(data, addr, MMU_DATA_IDX); /* mem[addr] = data */ @@ -XXX,XX +XXX,XX @@ static void gen_data_store(DisasContext *ctx, TCGv data, TCGv addr) static void gen_data_load(DisasContext *ctx, TCGv data, TCGv addr) { - if (ctx->tb->flags & TB_FLAGS_FULL_ACCESS) { + if (ctx->base.tb->flags & TB_FLAGS_FULL_ACCESS) { gen_helper_fullrd(data, cpu_env, addr); } else { tcg_gen_qemu_ld8u(data, addr, MMU_DATA_IDX); /* data = mem[addr] */ @@ -XXX,XX +XXX,XX @@ static bool trans_BREAK(DisasContext *ctx, arg_BREAK *a) #ifdef BREAKPOINT_ON_BREAK tcg_gen_movi_tl(cpu_pc, ctx->npc - 1); gen_helper_debug(cpu_env); - ctx->bstate = DISAS_EXIT; + ctx->base.is_jmp = DISAS_EXIT; #else /* NOP */ #endif @@ -XXX,XX +XXX,XX @@ static bool trans_NOP(DisasContext *ctx, arg_NOP *a) static bool trans_SLEEP(DisasContext *ctx, arg_SLEEP *a) { gen_helper_sleep(cpu_env); - ctx->bstate = DISAS_NORETURN; + ctx->base.is_jmp = DISAS_NORETURN; return true; } @@ -XXX,XX +XXX,XX @@ static void translate(DisasContext *ctx) if (!decode_insn(ctx, opcode)) { gen_helper_unsupported(cpu_env); - ctx->bstate = DISAS_NORETURN; + ctx->base.is_jmp = DISAS_NORETURN; } } @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) { CPUAVRState *env = cs->env_ptr; DisasContext ctx = { - .tb = tb, + .base.tb = tb, + .base.is_jmp = DISAS_NEXT, + .base.pc_first = tb->pc, + .base.pc_next = tb->pc, + .base.singlestep_enabled = cs->singlestep_enabled, .cs = cs, .env = env, .memidx = 0, - .bstate = DISAS_NEXT, .skip_cond = TCG_COND_NEVER, - .singlestep = cs->singlestep_enabled, }; target_ulong pc_start = tb->pc / 2; int num_insns = 0; @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) */ max_insns = 1; } - if (ctx.singlestep) { + if (ctx.base.singlestep_enabled) { max_insns = 1; } @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) * b main - sets breakpoint at address 0x00000100 (code) * b *0x100 - sets breakpoint at address 0x00800100 (data) */ - if (unlikely(!ctx.singlestep && + if (unlikely(!ctx.base.singlestep_enabled && (cpu_breakpoint_test(cs, OFFSET_CODE + ctx.npc * 2, BP_ANY) || cpu_breakpoint_test(cs, OFFSET_DATA + ctx.npc * 2, BP_ANY)))) { canonicalize_skip(&ctx); @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) if (skip_label) { canonicalize_skip(&ctx); gen_set_label(skip_label); - if (ctx.bstate == DISAS_NORETURN) { - ctx.bstate = DISAS_CHAIN; + if (ctx.base.is_jmp == DISAS_NORETURN) { + ctx.base.is_jmp = DISAS_CHAIN; } } - } while (ctx.bstate == DISAS_NEXT + } while (ctx.base.is_jmp == DISAS_NEXT && num_insns < max_insns && (ctx.npc - pc_start) * 2 < TARGET_PAGE_SIZE - 4 && !tcg_op_buf_full()); @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) bool nonconst_skip = canonicalize_skip(&ctx); - switch (ctx.bstate) { + switch (ctx.base.is_jmp) { case DISAS_NORETURN: assert(!nonconst_skip); break; @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) tcg_gen_movi_tl(cpu_pc, ctx.npc); /* fall through */ case DISAS_LOOKUP: - if (!ctx.singlestep) { + if (!ctx.base.singlestep_enabled) { tcg_gen_lookup_and_goto_ptr(); break; } /* fall through */ case DISAS_EXIT: - if (ctx.singlestep) { + if (ctx.base.singlestep_enabled) { gen_helper_debug(cpu_env); } else { tcg_gen_exit_tb(NULL, 0); -- 2.25.1 Prepare for receiving it as a pointer input. Tested-by: Michael Rolnik <mrolnik@gmail.com> Reviewed-by: Michael Rolnik <mrolnik@gmail.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/avr/translate.c | 84 +++++++++++++++++++++--------------------- 1 file changed, 43 insertions(+), 41 deletions(-) diff --git a/target/avr/translate.c b/target/avr/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/avr/translate.c +++ b/target/avr/translate.c @@ -XXX,XX +XXX,XX @@ struct DisasContext { * used in the following manner (sketch) * * TCGLabel *skip_label = NULL; - * if (ctx.skip_cond != TCG_COND_NEVER) { + * if (ctx->skip_cond != TCG_COND_NEVER) { * skip_label = gen_new_label(); * tcg_gen_brcond_tl(skip_cond, skip_var0, skip_var1, skip_label); * } @@ -XXX,XX +XXX,XX @@ struct DisasContext { * free_skip_var0 = false; * } * - * translate(&ctx); + * translate(ctx); * * if (skip_label) { * gen_set_label(skip_label); @@ -XXX,XX +XXX,XX @@ static bool canonicalize_skip(DisasContext *ctx) void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) { CPUAVRState *env = cs->env_ptr; - DisasContext ctx = { + DisasContext ctx1 = { .base.tb = tb, .base.is_jmp = DISAS_NEXT, .base.pc_first = tb->pc, @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) .memidx = 0, .skip_cond = TCG_COND_NEVER, }; + DisasContext *ctx = &ctx1; target_ulong pc_start = tb->pc / 2; int num_insns = 0; @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) */ max_insns = 1; } - if (ctx.base.singlestep_enabled) { + if (ctx->base.singlestep_enabled) { max_insns = 1; } gen_tb_start(tb); - ctx.npc = pc_start; + ctx->npc = pc_start; if (tb->flags & TB_FLAGS_SKIP) { - ctx.skip_cond = TCG_COND_ALWAYS; - ctx.skip_var0 = cpu_skip; + ctx->skip_cond = TCG_COND_ALWAYS; + ctx->skip_var0 = cpu_skip; } do { TCGLabel *skip_label = NULL; /* translate current instruction */ - tcg_gen_insn_start(ctx.npc); + tcg_gen_insn_start(ctx->npc); num_insns++; /* @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) * b main - sets breakpoint at address 0x00000100 (code) * b *0x100 - sets breakpoint at address 0x00800100 (data) */ - if (unlikely(!ctx.base.singlestep_enabled && - (cpu_breakpoint_test(cs, OFFSET_CODE + ctx.npc * 2, BP_ANY) || - cpu_breakpoint_test(cs, OFFSET_DATA + ctx.npc * 2, BP_ANY)))) { - canonicalize_skip(&ctx); - tcg_gen_movi_tl(cpu_pc, ctx.npc); + if (unlikely(!ctx->base.singlestep_enabled && + (cpu_breakpoint_test(cs, OFFSET_CODE + ctx->npc * 2, BP_ANY) || + cpu_breakpoint_test(cs, OFFSET_DATA + ctx->npc * 2, BP_ANY)))) { + canonicalize_skip(ctx); + tcg_gen_movi_tl(cpu_pc, ctx->npc); gen_helper_debug(cpu_env); goto done_generating; } /* Conditionally skip the next instruction, if indicated. */ - if (ctx.skip_cond != TCG_COND_NEVER) { + if (ctx->skip_cond != TCG_COND_NEVER) { skip_label = gen_new_label(); - if (ctx.skip_var0 == cpu_skip) { + if (ctx->skip_var0 == cpu_skip) { /* * Copy cpu_skip so that we may zero it before the branch. * This ensures that cpu_skip is non-zero after the label * if and only if the skipped insn itself sets a skip. */ - ctx.free_skip_var0 = true; - ctx.skip_var0 = tcg_temp_new(); - tcg_gen_mov_tl(ctx.skip_var0, cpu_skip); + ctx->free_skip_var0 = true; + ctx->skip_var0 = tcg_temp_new(); + tcg_gen_mov_tl(ctx->skip_var0, cpu_skip); tcg_gen_movi_tl(cpu_skip, 0); } - if (ctx.skip_var1 == NULL) { - tcg_gen_brcondi_tl(ctx.skip_cond, ctx.skip_var0, 0, skip_label); + if (ctx->skip_var1 == NULL) { + tcg_gen_brcondi_tl(ctx->skip_cond, ctx->skip_var0, + 0, skip_label); } else { - tcg_gen_brcond_tl(ctx.skip_cond, ctx.skip_var0, - ctx.skip_var1, skip_label); - ctx.skip_var1 = NULL; + tcg_gen_brcond_tl(ctx->skip_cond, ctx->skip_var0, + ctx->skip_var1, skip_label); + ctx->skip_var1 = NULL; } - if (ctx.free_skip_var0) { - tcg_temp_free(ctx.skip_var0); - ctx.free_skip_var0 = false; + if (ctx->free_skip_var0) { + tcg_temp_free(ctx->skip_var0); + ctx->free_skip_var0 = false; } - ctx.skip_cond = TCG_COND_NEVER; - ctx.skip_var0 = NULL; + ctx->skip_cond = TCG_COND_NEVER; + ctx->skip_var0 = NULL; } - translate(&ctx); + translate(ctx); if (skip_label) { - canonicalize_skip(&ctx); + canonicalize_skip(ctx); gen_set_label(skip_label); - if (ctx.base.is_jmp == DISAS_NORETURN) { - ctx.base.is_jmp = DISAS_CHAIN; + if (ctx->base.is_jmp == DISAS_NORETURN) { + ctx->base.is_jmp = DISAS_CHAIN; } } - } while (ctx.base.is_jmp == DISAS_NEXT + } while (ctx->base.is_jmp == DISAS_NEXT && num_insns < max_insns - && (ctx.npc - pc_start) * 2 < TARGET_PAGE_SIZE - 4 + && (ctx->npc - pc_start) * 2 < TARGET_PAGE_SIZE - 4 && !tcg_op_buf_full()); if (tb->cflags & CF_LAST_IO) { gen_io_end(); } - bool nonconst_skip = canonicalize_skip(&ctx); + bool nonconst_skip = canonicalize_skip(ctx); - switch (ctx.base.is_jmp) { + switch (ctx->base.is_jmp) { case DISAS_NORETURN: assert(!nonconst_skip); break; @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) case DISAS_CHAIN: if (!nonconst_skip) { /* Note gen_goto_tb checks singlestep. */ - gen_goto_tb(&ctx, 1, ctx.npc); + gen_goto_tb(ctx, 1, ctx->npc); break; } - tcg_gen_movi_tl(cpu_pc, ctx.npc); + tcg_gen_movi_tl(cpu_pc, ctx->npc); /* fall through */ case DISAS_LOOKUP: - if (!ctx.base.singlestep_enabled) { + if (!ctx->base.singlestep_enabled) { tcg_gen_lookup_and_goto_ptr(); break; } /* fall through */ case DISAS_EXIT: - if (ctx.base.singlestep_enabled) { + if (ctx->base.singlestep_enabled) { gen_helper_debug(cpu_env); } else { tcg_gen_exit_tb(NULL, 0); @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) done_generating: gen_tb_end(tb, num_insns); - tb->size = (ctx.npc - pc_start) * 2; + tb->size = (ctx->npc - pc_start) * 2; tb->icount = num_insns; #ifdef DEBUG_DISAS -- 2.25.1 Tested-by: Michael Rolnik <mrolnik@gmail.com> Reviewed-by: Michael Rolnik <mrolnik@gmail.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/avr/translate.c | 234 ++++++++++++++++++++++------------------- 1 file changed, 128 insertions(+), 106 deletions(-) diff --git a/target/avr/translate.c b/target/avr/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/avr/translate.c +++ b/target/avr/translate.c @@ -XXX,XX +XXX,XX @@ static bool canonicalize_skip(DisasContext *ctx) return true; } -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) +static void gen_breakpoint(DisasContext *ctx) { + canonicalize_skip(ctx); + tcg_gen_movi_tl(cpu_pc, ctx->npc); + gen_helper_debug(cpu_env); + ctx->base.is_jmp = DISAS_NORETURN; +} + +static void avr_tr_init_disas_context(DisasContextBase *dcbase, CPUState *cs) +{ + DisasContext *ctx = container_of(dcbase, DisasContext, base); CPUAVRState *env = cs->env_ptr; - DisasContext ctx1 = { - .base.tb = tb, - .base.is_jmp = DISAS_NEXT, - .base.pc_first = tb->pc, - .base.pc_next = tb->pc, - .base.singlestep_enabled = cs->singlestep_enabled, - .cs = cs, - .env = env, - .memidx = 0, - .skip_cond = TCG_COND_NEVER, - }; - DisasContext *ctx = &ctx1; - target_ulong pc_start = tb->pc / 2; - int num_insns = 0; + uint32_t tb_flags = ctx->base.tb->flags; - if (tb->flags & TB_FLAGS_FULL_ACCESS) { - /* - * This flag is set by ST/LD instruction we will regenerate it ONLY - * with mem/cpu memory access instead of mem access - */ - max_insns = 1; - } - if (ctx->base.singlestep_enabled) { - max_insns = 1; - } + ctx->cs = cs; + ctx->env = env; + ctx->npc = ctx->base.pc_first / 2; - gen_tb_start(tb); - - ctx->npc = pc_start; - if (tb->flags & TB_FLAGS_SKIP) { + ctx->skip_cond = TCG_COND_NEVER; + if (tb_flags & TB_FLAGS_SKIP) { ctx->skip_cond = TCG_COND_ALWAYS; ctx->skip_var0 = cpu_skip; } - do { - TCGLabel *skip_label = NULL; - - /* translate current instruction */ - tcg_gen_insn_start(ctx->npc); - num_insns++; - + if (tb_flags & TB_FLAGS_FULL_ACCESS) { /* - * this is due to some strange GDB behavior - * let's assume main has address 0x100 - * b main - sets breakpoint at address 0x00000100 (code) - * b *0x100 - sets breakpoint at address 0x00800100 (data) + * This flag is set by ST/LD instruction we will regenerate it ONLY + * with mem/cpu memory access instead of mem access */ - if (unlikely(!ctx->base.singlestep_enabled && - (cpu_breakpoint_test(cs, OFFSET_CODE + ctx->npc * 2, BP_ANY) || - cpu_breakpoint_test(cs, OFFSET_DATA + ctx->npc * 2, BP_ANY)))) { - canonicalize_skip(ctx); - tcg_gen_movi_tl(cpu_pc, ctx->npc); - gen_helper_debug(cpu_env); - goto done_generating; - } + ctx->base.max_insns = 1; + } +} - /* Conditionally skip the next instruction, if indicated. */ - if (ctx->skip_cond != TCG_COND_NEVER) { - skip_label = gen_new_label(); - if (ctx->skip_var0 == cpu_skip) { - /* - * Copy cpu_skip so that we may zero it before the branch. - * This ensures that cpu_skip is non-zero after the label - * if and only if the skipped insn itself sets a skip. - */ - ctx->free_skip_var0 = true; - ctx->skip_var0 = tcg_temp_new(); - tcg_gen_mov_tl(ctx->skip_var0, cpu_skip); - tcg_gen_movi_tl(cpu_skip, 0); - } - if (ctx->skip_var1 == NULL) { - tcg_gen_brcondi_tl(ctx->skip_cond, ctx->skip_var0, - 0, skip_label); - } else { - tcg_gen_brcond_tl(ctx->skip_cond, ctx->skip_var0, - ctx->skip_var1, skip_label); - ctx->skip_var1 = NULL; - } - if (ctx->free_skip_var0) { - tcg_temp_free(ctx->skip_var0); - ctx->free_skip_var0 = false; - } - ctx->skip_cond = TCG_COND_NEVER; - ctx->skip_var0 = NULL; - } +static void avr_tr_tb_start(DisasContextBase *db, CPUState *cs) +{ +} - translate(ctx); +static void avr_tr_insn_start(DisasContextBase *dcbase, CPUState *cs) +{ + DisasContext *ctx = container_of(dcbase, DisasContext, base); - if (skip_label) { - canonicalize_skip(ctx); - gen_set_label(skip_label); - if (ctx->base.is_jmp == DISAS_NORETURN) { - ctx->base.is_jmp = DISAS_CHAIN; - } - } - } while (ctx->base.is_jmp == DISAS_NEXT - && num_insns < max_insns - && (ctx->npc - pc_start) * 2 < TARGET_PAGE_SIZE - 4 - && !tcg_op_buf_full()); + tcg_gen_insn_start(ctx->npc); +} - if (tb->cflags & CF_LAST_IO) { - gen_io_end(); +static bool avr_tr_breakpoint_check(DisasContextBase *dcbase, CPUState *cs, + const CPUBreakpoint *bp) +{ + DisasContext *ctx = container_of(dcbase, DisasContext, base); + + gen_breakpoint(ctx); + return true; +} + +static void avr_tr_translate_insn(DisasContextBase *dcbase, CPUState *cs) +{ + DisasContext *ctx = container_of(dcbase, DisasContext, base); + TCGLabel *skip_label = NULL; + + /* + * This is due to some strange GDB behavior + * Let's assume main has address 0x100: + * b main - sets breakpoint at address 0x00000100 (code) + * b *0x100 - sets breakpoint at address 0x00800100 (data) + * + * The translator driver has already taken care of the code pointer. + */ + if (!ctx->base.singlestep_enabled && + cpu_breakpoint_test(cs, OFFSET_DATA + ctx->base.pc_next, BP_ANY)) { + gen_breakpoint(ctx); + return; } + /* Conditionally skip the next instruction, if indicated. */ + if (ctx->skip_cond != TCG_COND_NEVER) { + skip_label = gen_new_label(); + if (ctx->skip_var0 == cpu_skip) { + /* + * Copy cpu_skip so that we may zero it before the branch. + * This ensures that cpu_skip is non-zero after the label + * if and only if the skipped insn itself sets a skip. + */ + ctx->free_skip_var0 = true; + ctx->skip_var0 = tcg_temp_new(); + tcg_gen_mov_tl(ctx->skip_var0, cpu_skip); + tcg_gen_movi_tl(cpu_skip, 0); + } + if (ctx->skip_var1 == NULL) { + tcg_gen_brcondi_tl(ctx->skip_cond, ctx->skip_var0, 0, skip_label); + } else { + tcg_gen_brcond_tl(ctx->skip_cond, ctx->skip_var0, + ctx->skip_var1, skip_label); + ctx->skip_var1 = NULL; + } + if (ctx->free_skip_var0) { + tcg_temp_free(ctx->skip_var0); + ctx->free_skip_var0 = false; + } + ctx->skip_cond = TCG_COND_NEVER; + ctx->skip_var0 = NULL; + } + + translate(ctx); + + ctx->base.pc_next = ctx->npc * 2; + + if (skip_label) { + canonicalize_skip(ctx); + gen_set_label(skip_label); + if (ctx->base.is_jmp == DISAS_NORETURN) { + ctx->base.is_jmp = DISAS_CHAIN; + } + } + + if (ctx->base.is_jmp == DISAS_NEXT) { + target_ulong page_first = ctx->base.pc_first & TARGET_PAGE_MASK; + + if ((ctx->base.pc_next - page_first) >= TARGET_PAGE_SIZE - 4) { + ctx->base.is_jmp = DISAS_TOO_MANY; + } + } +} + +static void avr_tr_tb_stop(DisasContextBase *dcbase, CPUState *cs) +{ + DisasContext *ctx = container_of(dcbase, DisasContext, base); bool nonconst_skip = canonicalize_skip(ctx); switch (ctx->base.is_jmp) { @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) default: g_assert_not_reached(); } +} -done_generating: - gen_tb_end(tb, num_insns); +static void avr_tr_disas_log(const DisasContextBase *dcbase, CPUState *cs) +{ + qemu_log("IN: %s\n", lookup_symbol(dcbase->pc_first)); + log_target_disas(cs, dcbase->pc_first, dcbase->tb->size); +} - tb->size = (ctx->npc - pc_start) * 2; - tb->icount = num_insns; +static const TranslatorOps avr_tr_ops = { + .init_disas_context = avr_tr_init_disas_context, + .tb_start = avr_tr_tb_start, + .insn_start = avr_tr_insn_start, + .breakpoint_check = avr_tr_breakpoint_check, + .translate_insn = avr_tr_translate_insn, + .tb_stop = avr_tr_tb_stop, + .disas_log = avr_tr_disas_log, +}; -#ifdef DEBUG_DISAS - if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM) - && qemu_log_in_addr_range(tb->pc)) { - FILE *fd; - fd = qemu_log_lock(); - qemu_log("IN: %s\n", lookup_symbol(tb->pc)); - log_target_disas(cs, tb->pc, tb->size); - qemu_log("\n"); - qemu_log_unlock(fd); - } -#endif +void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) +{ + DisasContext dc = { }; + translator_loop(&avr_tr_ops, &dc.base, cs, tb, max_insns); } void restore_state_to_opc(CPUAVRState *env, TranslationBlock *tb, -- 2.25.1 Migrate the is_jmp, tb and singlestep_enabled fields from DisasContext into the base. Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 49 +++++++++++++++++---------------- target/cris/translate_v10.c.inc | 4 +-- 2 files changed, 27 insertions(+), 26 deletions(-) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ static TCGv env_pc; /* This is the state at translation time. */ typedef struct DisasContext { + DisasContextBase base; + CRISCPU *cpu; target_ulong pc, ppc; @@ -XXX,XX +XXX,XX @@ typedef struct DisasContext { int clear_locked_irq; /* Clear the irq lockout. */ int cpustate_changed; unsigned int tb_flags; /* tb dependent flags. */ - int is_jmp; #define JMP_NOJMP 0 #define JMP_DIRECT 1 @@ -XXX,XX +XXX,XX @@ typedef struct DisasContext { uint32_t jmp_pc; int delayed_branch; - - TranslationBlock *tb; - int singlestep_enabled; } DisasContext; static void gen_BUG(DisasContext *dc, const char *file, int line) @@ -XXX,XX +XXX,XX @@ static void t_gen_cc_jmp(TCGv pc_true, TCGv pc_false) static inline bool use_goto_tb(DisasContext *dc, target_ulong dest) { #ifndef CONFIG_USER_ONLY - return (dc->tb->pc & TARGET_PAGE_MASK) == (dest & TARGET_PAGE_MASK) || + return (dc->base.pc_first & TARGET_PAGE_MASK) == (dest & TARGET_PAGE_MASK) || (dc->ppc & TARGET_PAGE_MASK) == (dest & TARGET_PAGE_MASK); #else return true; @@ -XXX,XX +XXX,XX @@ static void gen_goto_tb(DisasContext *dc, int n, target_ulong dest) if (use_goto_tb(dc, dest)) { tcg_gen_goto_tb(n); tcg_gen_movi_tl(env_pc, dest); - tcg_gen_exit_tb(dc->tb, n); + tcg_gen_exit_tb(dc->base.tb, n); } else { tcg_gen_movi_tl(env_pc, dest); tcg_gen_exit_tb(NULL, 0); @@ -XXX,XX +XXX,XX @@ static int dec_setclrf(CPUCRISState *env, DisasContext *dc) /* Break the TB if any of the SPI flag changes. */ if (flags & (P_FLAG | S_FLAG)) { tcg_gen_movi_tl(env_pc, dc->pc + 2); - dc->is_jmp = DISAS_UPDATE; + dc->base.is_jmp = DISAS_UPDATE; dc->cpustate_changed = 1; } /* For the I flag, only act on posedge. */ if ((flags & I_FLAG)) { tcg_gen_movi_tl(env_pc, dc->pc + 2); - dc->is_jmp = DISAS_UPDATE; + dc->base.is_jmp = DISAS_UPDATE; dc->cpustate_changed = 1; } @@ -XXX,XX +XXX,XX @@ static int dec_rfe_etc(CPUCRISState *env, DisasContext *dc) LOG_DIS("rfe\n"); cris_evaluate_flags(dc); gen_helper_rfe(cpu_env); - dc->is_jmp = DISAS_UPDATE; + dc->base.is_jmp = DISAS_UPDATE; break; case 5: /* rfn. */ LOG_DIS("rfn\n"); cris_evaluate_flags(dc); gen_helper_rfn(cpu_env); - dc->is_jmp = DISAS_UPDATE; + dc->base.is_jmp = DISAS_UPDATE; break; case 6: LOG_DIS("break %d\n", dc->op1); @@ -XXX,XX +XXX,XX @@ static int dec_rfe_etc(CPUCRISState *env, DisasContext *dc) /* Breaks start at 16 in the exception vector. */ t_gen_movi_env_TN(trap_vector, dc->op1 + 16); t_gen_raise_exception(EXCP_BREAK); - dc->is_jmp = DISAS_UPDATE; + dc->base.is_jmp = DISAS_UPDATE; break; default: printf("op2=%x\n", dc->op2); @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) * delayslot, like in real hw. */ pc_start = tb->pc & ~1; - dc->cpu = env_archcpu(env); - dc->tb = tb; - dc->is_jmp = DISAS_NEXT; + dc->base.tb = tb; + dc->base.pc_first = pc_start; + dc->base.pc_next = pc_start; + dc->base.is_jmp = DISAS_NEXT; + dc->base.singlestep_enabled = cs->singlestep_enabled; + + dc->cpu = env_archcpu(env); dc->ppc = pc_start; dc->pc = pc_start; - dc->singlestep_enabled = cs->singlestep_enabled; dc->flags_uptodate = 1; dc->flagx_known = 1; dc->flags_x = tb->flags & X_FLAG; @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) cris_evaluate_flags(dc); tcg_gen_movi_tl(env_pc, dc->pc); t_gen_raise_exception(EXCP_DEBUG); - dc->is_jmp = DISAS_UPDATE; + dc->base.is_jmp = DISAS_UPDATE; /* The address covered by the breakpoint must be included in [tb->pc, tb->pc + tb->size) in order to for it to be properly cleared -- thus we increment the PC here so that @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) gen_goto_tb(dc, 1, dc->jmp_pc); gen_set_label(l1); gen_goto_tb(dc, 0, dc->pc); - dc->is_jmp = DISAS_TB_JUMP; + dc->base.is_jmp = DISAS_TB_JUMP; dc->jmp = JMP_NOJMP; } else if (dc->jmp == JMP_DIRECT) { cris_evaluate_flags(dc); gen_goto_tb(dc, 0, dc->jmp_pc); - dc->is_jmp = DISAS_TB_JUMP; + dc->base.is_jmp = DISAS_TB_JUMP; dc->jmp = JMP_NOJMP; } else { TCGv c = tcg_const_tl(dc->pc); t_gen_cc_jmp(env_btarget, c); tcg_temp_free(c); - dc->is_jmp = DISAS_JUMP; + dc->base.is_jmp = DISAS_JUMP; } break; } @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) if (!(tb->pc & 1) && cs->singlestep_enabled) { break; } - } while (!dc->is_jmp && !dc->cpustate_changed + } while (!dc->base.is_jmp && !dc->cpustate_changed && !tcg_op_buf_full() && !singlestep && (dc->pc - page_start < TARGET_PAGE_SIZE) @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) npc = dc->pc; /* Force an update if the per-tb cpu state has changed. */ - if (dc->is_jmp == DISAS_NEXT + if (dc->base.is_jmp == DISAS_NEXT && (dc->cpustate_changed || !dc->flagx_known || (dc->flags_x != (tb->flags & X_FLAG)))) { - dc->is_jmp = DISAS_UPDATE; + dc->base.is_jmp = DISAS_UPDATE; tcg_gen_movi_tl(env_pc, npc); } /* Broken branch+delayslot sequence. */ @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) cris_evaluate_flags(dc); if (unlikely(cs->singlestep_enabled)) { - if (dc->is_jmp == DISAS_NEXT) { + if (dc->base.is_jmp == DISAS_NEXT) { tcg_gen_movi_tl(env_pc, npc); } t_gen_raise_exception(EXCP_DEBUG); } else { - switch (dc->is_jmp) { + switch (dc->base.is_jmp) { case DISAS_NEXT: gen_goto_tb(dc, 1, npc); break; diff --git a/target/cris/translate_v10.c.inc b/target/cris/translate_v10.c.inc index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate_v10.c.inc +++ b/target/cris/translate_v10.c.inc @@ -XXX,XX +XXX,XX @@ static unsigned int dec10_ind(CPUCRISState *env, DisasContext *dc) t_gen_mov_env_TN(trap_vector, c); tcg_temp_free(c); t_gen_raise_exception(EXCP_BREAK); - dc->is_jmp = DISAS_UPDATE; + dc->base.is_jmp = DISAS_UPDATE; return insn_len; } LOG_DIS("%d: jump.%d %d r%d r%d\n", __LINE__, size, @@ -XXX,XX +XXX,XX @@ static unsigned int crisv10_decoder(CPUCRISState *env, DisasContext *dc) if (dc->clear_prefix && dc->tb_flags & PFIX_FLAG) { dc->tb_flags &= ~PFIX_FLAG; tcg_gen_andi_tl(cpu_PR[PR_CCS], cpu_PR[PR_CCS], ~PFIX_FLAG); - if (dc->tb_flags != dc->tb->flags) { + if (dc->tb_flags != dc->base.tb->flags) { dc->cpustate_changed = 1; } } -- 2.25.1 This value is unused. Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ #define DISAS_JUMP DISAS_TARGET_0 /* only pc was modified dynamically */ #define DISAS_UPDATE DISAS_TARGET_1 /* cpu state was modified dynamically */ #define DISAS_TB_JUMP DISAS_TARGET_2 /* only pc was modified statically */ -#define DISAS_SWI DISAS_TARGET_3 /* Used by the decoder. */ #define EXTRACT_FIELD(src, start, end) \ @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) to find the next TB */ tcg_gen_exit_tb(NULL, 0); break; - case DISAS_SWI: case DISAS_TB_JUMP: /* nothing more to generate */ break; -- 2.25.1 The only semantic of DISAS_TB_JUMP is that we've done goto_tb, which is the same as DISAS_NORETURN -- we've exited the tb. Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ /* is_jmp field values */ #define DISAS_JUMP DISAS_TARGET_0 /* only pc was modified dynamically */ #define DISAS_UPDATE DISAS_TARGET_1 /* cpu state was modified dynamically */ -#define DISAS_TB_JUMP DISAS_TARGET_2 /* only pc was modified statically */ /* Used by the decoder. */ #define EXTRACT_FIELD(src, start, end) \ @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) gen_goto_tb(dc, 1, dc->jmp_pc); gen_set_label(l1); gen_goto_tb(dc, 0, dc->pc); - dc->base.is_jmp = DISAS_TB_JUMP; + dc->base.is_jmp = DISAS_NORETURN; dc->jmp = JMP_NOJMP; } else if (dc->jmp == JMP_DIRECT) { cris_evaluate_flags(dc); gen_goto_tb(dc, 0, dc->jmp_pc); - dc->base.is_jmp = DISAS_TB_JUMP; + dc->base.is_jmp = DISAS_NORETURN; dc->jmp = JMP_NOJMP; } else { TCGv c = tcg_const_tl(dc->pc); @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) to find the next TB */ tcg_gen_exit_tb(NULL, 0); break; - case DISAS_TB_JUMP: + case DISAS_NORETURN: /* nothing more to generate */ break; } -- 2.25.1 After we've raised the exception, we have left the TB. Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 5 +++-- target/cris/translate_v10.c.inc | 3 ++- 2 files changed, 5 insertions(+), 3 deletions(-) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ static int dec_rfe_etc(CPUCRISState *env, DisasContext *dc) -offsetof(CRISCPU, env) + offsetof(CPUState, halted)); tcg_gen_movi_tl(env_pc, dc->pc + 2); t_gen_raise_exception(EXCP_HLT); + dc->base.is_jmp = DISAS_NORETURN; return 2; } @@ -XXX,XX +XXX,XX @@ static int dec_rfe_etc(CPUCRISState *env, DisasContext *dc) /* Breaks start at 16 in the exception vector. */ t_gen_movi_env_TN(trap_vector, dc->op1 + 16); t_gen_raise_exception(EXCP_BREAK); - dc->base.is_jmp = DISAS_UPDATE; + dc->base.is_jmp = DISAS_NORETURN; break; default: printf("op2=%x\n", dc->op2); @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) cris_evaluate_flags(dc); tcg_gen_movi_tl(env_pc, dc->pc); t_gen_raise_exception(EXCP_DEBUG); - dc->base.is_jmp = DISAS_UPDATE; + dc->base.is_jmp = DISAS_NORETURN; /* The address covered by the breakpoint must be included in [tb->pc, tb->pc + tb->size) in order to for it to be properly cleared -- thus we increment the PC here so that diff --git a/target/cris/translate_v10.c.inc b/target/cris/translate_v10.c.inc index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate_v10.c.inc +++ b/target/cris/translate_v10.c.inc @@ -XXX,XX +XXX,XX @@ static inline void cris_illegal_insn(DisasContext *dc) { qemu_log_mask(LOG_GUEST_ERROR, "illegal insn at pc=%x\n", dc->pc); t_gen_raise_exception(EXCP_BREAK); + dc->base.is_jmp = DISAS_NORETURN; } static void gen_store_v10_conditional(DisasContext *dc, TCGv addr, TCGv val, @@ -XXX,XX +XXX,XX @@ static unsigned int dec10_ind(CPUCRISState *env, DisasContext *dc) t_gen_mov_env_TN(trap_vector, c); tcg_temp_free(c); t_gen_raise_exception(EXCP_BREAK); - dc->base.is_jmp = DISAS_UPDATE; + dc->base.is_jmp = DISAS_NORETURN; return insn_len; } LOG_DIS("%d: jump.%d %d r%d r%d\n", __LINE__, size, -- 2.25.1 Do not skip the page check for user-only -- mmap/mprotect can still change page mappings. Only check dc->base.pc_first, not dc->ppc -- the start page is the only one that's relevant. Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 9 ++------- 1 file changed, 2 insertions(+), 7 deletions(-) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ static void t_gen_cc_jmp(TCGv pc_true, TCGv pc_false) gen_set_label(l1); } -static inline bool use_goto_tb(DisasContext *dc, target_ulong dest) +static bool use_goto_tb(DisasContext *dc, target_ulong dest) { -#ifndef CONFIG_USER_ONLY - return (dc->base.pc_first & TARGET_PAGE_MASK) == (dest & TARGET_PAGE_MASK) || - (dc->ppc & TARGET_PAGE_MASK) == (dest & TARGET_PAGE_MASK); -#else - return true; -#endif + return ((dest ^ dc->base.pc_first) & TARGET_PAGE_MASK) == 0; } static void gen_goto_tb(DisasContext *dc, int n, target_ulong dest) -- 2.25.1 Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 317 ++++++++++++++++++++++------------------ 1 file changed, 174 insertions(+), 143 deletions(-) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ static unsigned int crisv32_decoder(CPUCRISState *env, DisasContext *dc) * */ -/* generate intermediate code for basic block 'tb'. */ -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) +static void cris_tr_init_disas_context(DisasContextBase *dcbase, CPUState *cs) { + DisasContext *dc = container_of(dcbase, DisasContext, base); CPUCRISState *env = cs->env_ptr; + uint32_t tb_flags = dc->base.tb->flags; uint32_t pc_start; - unsigned int insn_len; - struct DisasContext ctx; - struct DisasContext *dc = &ctx; - uint32_t page_start; - target_ulong npc; - int num_insns; if (env->pregs[PR_VR] == 32) { dc->decoder = crisv32_decoder; @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) dc->clear_locked_irq = 1; } - /* Odd PC indicates that branch is rexecuting due to exception in the + /* + * Odd PC indicates that branch is rexecuting due to exception in the * delayslot, like in real hw. */ - pc_start = tb->pc & ~1; - - dc->base.tb = tb; + pc_start = dc->base.pc_first & ~1; dc->base.pc_first = pc_start; dc->base.pc_next = pc_start; - dc->base.is_jmp = DISAS_NEXT; - dc->base.singlestep_enabled = cs->singlestep_enabled; dc->cpu = env_archcpu(env); dc->ppc = pc_start; dc->pc = pc_start; dc->flags_uptodate = 1; dc->flagx_known = 1; - dc->flags_x = tb->flags & X_FLAG; + dc->flags_x = tb_flags & X_FLAG; dc->cc_x_uptodate = 0; dc->cc_mask = 0; dc->update_cc = 0; dc->clear_prefix = 0; + dc->cpustate_changed = 0; cris_update_cc_op(dc, CC_OP_FLAGS, 4); dc->cc_size_uptodate = -1; /* Decode TB flags. */ - dc->tb_flags = tb->flags & (S_FLAG | P_FLAG | U_FLAG \ - | X_FLAG | PFIX_FLAG); - dc->delayed_branch = !!(tb->flags & 7); + dc->tb_flags = tb_flags & (S_FLAG | P_FLAG | U_FLAG | X_FLAG | PFIX_FLAG); + dc->delayed_branch = !!(tb_flags & 7); if (dc->delayed_branch) { dc->jmp = JMP_INDIRECT; } else { dc->jmp = JMP_NOJMP; } +} - dc->cpustate_changed = 0; +static void cris_tr_tb_start(DisasContextBase *db, CPUState *cpu) +{ +} - page_start = pc_start & TARGET_PAGE_MASK; - num_insns = 0; +static void cris_tr_insn_start(DisasContextBase *dcbase, CPUState *cpu) +{ + DisasContext *dc = container_of(dcbase, DisasContext, base); - gen_tb_start(tb); - do { - tcg_gen_insn_start(dc->delayed_branch == 1 - ? dc->ppc | 1 : dc->pc); - num_insns++; + tcg_gen_insn_start(dc->delayed_branch == 1 ? dc->ppc | 1 : dc->pc); +} - if (unlikely(cpu_breakpoint_test(cs, dc->pc, BP_ANY))) { +static bool cris_tr_breakpoint_check(DisasContextBase *dcbase, CPUState *cpu, + const CPUBreakpoint *bp) +{ + DisasContext *dc = container_of(dcbase, DisasContext, base); + + cris_evaluate_flags(dc); + tcg_gen_movi_tl(env_pc, dc->pc); + t_gen_raise_exception(EXCP_DEBUG); + dc->base.is_jmp = DISAS_NORETURN; + /* + * The address covered by the breakpoint must be included in + * [tb->pc, tb->pc + tb->size) in order to for it to be + * properly cleared -- thus we increment the PC here so that + * the logic setting tb->size below does the right thing. + */ + dc->pc += 2; + return true; +} + +static void cris_tr_translate_insn(DisasContextBase *dcbase, CPUState *cs) +{ + DisasContext *dc = container_of(dcbase, DisasContext, base); + CPUCRISState *env = cs->env_ptr; + unsigned int insn_len; + + /* Pretty disas. */ + LOG_DIS("%8.8x:\t", dc->pc); + + dc->clear_x = 1; + + insn_len = dc->decoder(env, dc); + dc->ppc = dc->pc; + dc->pc += insn_len; + dc->base.pc_next += insn_len; + + if (dc->base.is_jmp == DISAS_NORETURN) { + return; + } + + if (dc->clear_x) { + cris_clear_x_flag(dc); + } + + /* + * Check for delayed branches here. If we do it before + * actually generating any host code, the simulator will just + * loop doing nothing for on this program location. + */ + if (dc->delayed_branch && --dc->delayed_branch == 0) { + if (dc->base.tb->flags & 7) { + t_gen_movi_env_TN(dslot, 0); + } + + if (dc->cpustate_changed + || !dc->flagx_known + || (dc->flags_x != (dc->base.tb->flags & X_FLAG))) { + cris_store_direct_jmp(dc); + } + + if (dc->clear_locked_irq) { + dc->clear_locked_irq = 0; + t_gen_movi_env_TN(locked_irq, 0); + } + + if (dc->jmp == JMP_DIRECT_CC) { + TCGLabel *l1 = gen_new_label(); cris_evaluate_flags(dc); - tcg_gen_movi_tl(env_pc, dc->pc); - t_gen_raise_exception(EXCP_DEBUG); + + /* Conditional jmp. */ + tcg_gen_brcondi_tl(TCG_COND_EQ, env_btaken, 0, l1); + gen_goto_tb(dc, 1, dc->jmp_pc); + gen_set_label(l1); + gen_goto_tb(dc, 0, dc->pc); dc->base.is_jmp = DISAS_NORETURN; - /* The address covered by the breakpoint must be included in - [tb->pc, tb->pc + tb->size) in order to for it to be - properly cleared -- thus we increment the PC here so that - the logic setting tb->size below does the right thing. */ - dc->pc += 2; - break; + dc->jmp = JMP_NOJMP; + } else if (dc->jmp == JMP_DIRECT) { + cris_evaluate_flags(dc); + gen_goto_tb(dc, 0, dc->jmp_pc); + dc->base.is_jmp = DISAS_NORETURN; + dc->jmp = JMP_NOJMP; + } else { + TCGv c = tcg_const_tl(dc->pc); + t_gen_cc_jmp(env_btarget, c); + tcg_temp_free(c); + dc->base.is_jmp = DISAS_JUMP; } + } - /* Pretty disas. */ - LOG_DIS("%8.8x:\t", dc->pc); + /* Force an update if the per-tb cpu state has changed. */ + if (dc->base.is_jmp == DISAS_NEXT + && (dc->cpustate_changed + || !dc->flagx_known + || (dc->flags_x != (dc->base.tb->flags & X_FLAG)))) { + dc->base.is_jmp = DISAS_UPDATE; + tcg_gen_movi_tl(env_pc, dc->pc); + } - if (num_insns == max_insns && (tb_cflags(tb) & CF_LAST_IO)) { - gen_io_start(); - } - dc->clear_x = 1; + /* + * FIXME: Only the first insn in the TB should cross a page boundary. + * If we can detect the length of the next insn easily, we should. + * In the meantime, simply stop when we do cross. + */ + if (dc->base.is_jmp == DISAS_NEXT + && ((dc->pc ^ dc->base.pc_first) & TARGET_PAGE_MASK) != 0) { + dc->base.is_jmp = DISAS_TOO_MANY; + } +} - insn_len = dc->decoder(env, dc); - dc->ppc = dc->pc; - dc->pc += insn_len; - if (dc->clear_x) { - cris_clear_x_flag(dc); - } +static void cris_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) +{ + DisasContext *dc = container_of(dcbase, DisasContext, base); + DisasJumpType is_jmp = dc->base.is_jmp; + target_ulong npc = dc->pc; - /* Check for delayed branches here. If we do it before - actually generating any host code, the simulator will just - loop doing nothing for on this program location. */ - if (dc->delayed_branch) { - dc->delayed_branch--; - if (dc->delayed_branch == 0) { - if (tb->flags & 7) { - t_gen_movi_env_TN(dslot, 0); - } - if (dc->cpustate_changed || !dc->flagx_known - || (dc->flags_x != (tb->flags & X_FLAG))) { - cris_store_direct_jmp(dc); - } - - if (dc->clear_locked_irq) { - dc->clear_locked_irq = 0; - t_gen_movi_env_TN(locked_irq, 0); - } - - if (dc->jmp == JMP_DIRECT_CC) { - TCGLabel *l1 = gen_new_label(); - cris_evaluate_flags(dc); - - /* Conditional jmp. */ - tcg_gen_brcondi_tl(TCG_COND_EQ, - env_btaken, 0, l1); - gen_goto_tb(dc, 1, dc->jmp_pc); - gen_set_label(l1); - gen_goto_tb(dc, 0, dc->pc); - dc->base.is_jmp = DISAS_NORETURN; - dc->jmp = JMP_NOJMP; - } else if (dc->jmp == JMP_DIRECT) { - cris_evaluate_flags(dc); - gen_goto_tb(dc, 0, dc->jmp_pc); - dc->base.is_jmp = DISAS_NORETURN; - dc->jmp = JMP_NOJMP; - } else { - TCGv c = tcg_const_tl(dc->pc); - t_gen_cc_jmp(env_btarget, c); - tcg_temp_free(c); - dc->base.is_jmp = DISAS_JUMP; - } - break; - } - } - - /* If we are rexecuting a branch due to exceptions on - delay slots don't break. */ - if (!(tb->pc & 1) && cs->singlestep_enabled) { - break; - } - } while (!dc->base.is_jmp && !dc->cpustate_changed - && !tcg_op_buf_full() - && !singlestep - && (dc->pc - page_start < TARGET_PAGE_SIZE) - && num_insns < max_insns); + if (is_jmp == DISAS_NORETURN) { + /* If we have a broken branch+delayslot sequence, it's too late. */ + assert(dc->delayed_branch != 1); + return; + } if (dc->clear_locked_irq) { t_gen_movi_env_TN(locked_irq, 0); } - npc = dc->pc; - - /* Force an update if the per-tb cpu state has changed. */ - if (dc->base.is_jmp == DISAS_NEXT - && (dc->cpustate_changed || !dc->flagx_known - || (dc->flags_x != (tb->flags & X_FLAG)))) { - dc->base.is_jmp = DISAS_UPDATE; - tcg_gen_movi_tl(env_pc, npc); - } /* Broken branch+delayslot sequence. */ if (dc->delayed_branch == 1) { /* Set env->dslot to the size of the branch insn. */ @@ -XXX,XX +XXX,XX @@ void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) cris_evaluate_flags(dc); - if (unlikely(cs->singlestep_enabled)) { - if (dc->base.is_jmp == DISAS_NEXT) { + if (unlikely(dc->base.singlestep_enabled)) { + switch (is_jmp) { + case DISAS_TOO_MANY: tcg_gen_movi_tl(env_pc, npc); - } - t_gen_raise_exception(EXCP_DEBUG); - } else { - switch (dc->base.is_jmp) { - case DISAS_NEXT: - gen_goto_tb(dc, 1, npc); - break; - default: + /* fall through */ case DISAS_JUMP: case DISAS_UPDATE: - /* indicate that the hash table must be used - to find the next TB */ - tcg_gen_exit_tb(NULL, 0); - break; - case DISAS_NORETURN: - /* nothing more to generate */ + t_gen_raise_exception(EXCP_DEBUG); + return; + default: break; } + g_assert_not_reached(); } - gen_tb_end(tb, num_insns); - tb->size = dc->pc - pc_start; - tb->icount = num_insns; - -#ifdef DEBUG_DISAS -#if !DISAS_CRIS - if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM) - && qemu_log_in_addr_range(pc_start)) { - FILE *logfile = qemu_log_lock(); - qemu_log("--------------\n"); - qemu_log("IN: %s\n", lookup_symbol(pc_start)); - log_target_disas(cs, pc_start, dc->pc - pc_start); - qemu_log_unlock(logfile); + switch (is_jmp) { + case DISAS_TOO_MANY: + gen_goto_tb(dc, 0, npc); + break; + case DISAS_JUMP: + case DISAS_UPDATE: + /* Indicate that interupts must be re-evaluated before the next TB. */ + tcg_gen_exit_tb(NULL, 0); + break; + default: + g_assert_not_reached(); } -#endif -#endif +} + +static void cris_tr_disas_log(const DisasContextBase *dcbase, CPUState *cpu) +{ + if (!DISAS_CRIS) { + qemu_log("IN: %s\n", lookup_symbol(dcbase->pc_first)); + log_target_disas(cpu, dcbase->pc_first, dcbase->tb->size); + } +} + +static const TranslatorOps cris_tr_ops = { + .init_disas_context = cris_tr_init_disas_context, + .tb_start = cris_tr_tb_start, + .insn_start = cris_tr_insn_start, + .breakpoint_check = cris_tr_breakpoint_check, + .translate_insn = cris_tr_translate_insn, + .tb_stop = cris_tr_tb_stop, + .disas_log = cris_tr_disas_log, +}; + +void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns) +{ + DisasContext dc; + translator_loop(&cris_tr_ops, &dc.base, cs, tb, max_insns); } void cris_cpu_dump_state(CPUState *cs, FILE *f, int flags) -- 2.25.1 Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/helper.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/target/cris/helper.h b/target/cris/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/cris/helper.h +++ b/target/cris/helper.h @@ -XXX,XX +XXX,XX @@ -DEF_HELPER_2(raise_exception, void, env, i32) +DEF_HELPER_2(raise_exception, noreturn, env, i32) DEF_HELPER_2(tlb_flush_pid, void, env, i32) DEF_HELPER_2(spc_write, void, env, i32) DEF_HELPER_1(rfe, void, env) -- 2.25.1 Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 19 ++++++++++--------- target/cris/translate_v10.c.inc | 6 +++--- 2 files changed, 13 insertions(+), 12 deletions(-) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ static void gen_BUG(DisasContext *dc, const char *file, int line) cpu_abort(CPU(dc->cpu), "%s:%d pc=%x\n", file, line, dc->pc); } -static const char *regnames_v32[] = +static const char * const regnames_v32[] = { "$r0", "$r1", "$r2", "$r3", "$r4", "$r5", "$r6", "$r7", "$r8", "$r9", "$r10", "$r11", "$r12", "$r13", "$sp", "$acr", }; -static const char *pregnames_v32[] = + +static const char * const pregnames_v32[] = { "$bz", "$vr", "$pid", "$srs", "$wz", "$exs", "$eda", "$mof", @@ -XXX,XX +XXX,XX @@ static const char *pregnames_v32[] = }; /* We need this table to handle preg-moves with implicit width. */ -static int preg_sizes[] = { +static const int preg_sizes[] = { 1, /* bz. */ 1, /* vr. */ 4, /* pid. */ @@ -XXX,XX +XXX,XX @@ static inline void t_gen_swapw(TCGv d, TCGv s) ((T0 >> 5) & 0x02020202) | ((T0 >> 7) & 0x01010101)); */ -static inline void t_gen_swapr(TCGv d, TCGv s) +static void t_gen_swapr(TCGv d, TCGv s) { - struct { + static const struct { int shift; /* LSL when positive, LSR when negative. */ uint32_t mask; } bitrev[] = { @@ -XXX,XX +XXX,XX @@ static int dec_prep_alu_m(CPUCRISState *env, DisasContext *dc, #if DISAS_CRIS static const char *cc_name(int cc) { - static const char *cc_names[16] = { + static const char * const cc_names[16] = { "cc", "cs", "ne", "eq", "vc", "vs", "pl", "mi", "ls", "hi", "ge", "lt", "gt", "le", "a", "p" }; @@ -XXX,XX +XXX,XX @@ static int dec_null(CPUCRISState *env, DisasContext *dc) return 2; } -static struct decoder_info { +static const struct decoder_info { struct { uint32_t bits; uint32_t mask; @@ -XXX,XX +XXX,XX @@ void cris_cpu_dump_state(CPUState *cs, FILE *f, int flags) { CRISCPU *cpu = CRIS_CPU(cs); CPUCRISState *env = &cpu->env; - const char **regnames; - const char **pregnames; + const char * const *regnames; + const char * const *pregnames; int i; if (!env) { diff --git a/target/cris/translate_v10.c.inc b/target/cris/translate_v10.c.inc index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate_v10.c.inc +++ b/target/cris/translate_v10.c.inc @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "crisv10-decode.h" -static const char *regnames_v10[] = +static const char * const regnames_v10[] = { "$r0", "$r1", "$r2", "$r3", "$r4", "$r5", "$r6", "$r7", @@ -XXX,XX +XXX,XX @@ static const char *regnames_v10[] = "$r12", "$r13", "$sp", "$pc", }; -static const char *pregnames_v10[] = +static const char * const pregnames_v10[] = { "$bz", "$vr", "$p2", "$p3", "$wz", "$ccr", "$p6-prefix", "$mof", @@ -XXX,XX +XXX,XX @@ static const char *pregnames_v10[] = }; /* We need this table to handle preg-moves with implicit width. */ -static int preg_sizes_v10[] = { +static const int preg_sizes_v10[] = { 1, /* bz. */ 1, /* vr. */ 1, /* pid. */ -- 2.25.1 We really do this already, by including them into the same test. This just hoists the expression up a bit. Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ static void cris_tr_translate_insn(DisasContextBase *dcbase, CPUState *cs) cris_clear_x_flag(dc); } + /* Fold unhandled changes to X_FLAG into cpustate_changed. */ + dc->cpustate_changed |= !dc->flagx_known; + dc->cpustate_changed |= dc->flags_x != (dc->base.tb->flags & X_FLAG); + /* * Check for delayed branches here. If we do it before * actually generating any host code, the simulator will just @@ -XXX,XX +XXX,XX @@ static void cris_tr_translate_insn(DisasContextBase *dcbase, CPUState *cs) t_gen_movi_env_TN(dslot, 0); } - if (dc->cpustate_changed - || !dc->flagx_known - || (dc->flags_x != (dc->base.tb->flags & X_FLAG))) { + if (dc->cpustate_changed) { cris_store_direct_jmp(dc); } @@ -XXX,XX +XXX,XX @@ static void cris_tr_translate_insn(DisasContextBase *dcbase, CPUState *cs) } /* Force an update if the per-tb cpu state has changed. */ - if (dc->base.is_jmp == DISAS_NEXT - && (dc->cpustate_changed - || !dc->flagx_known - || (dc->flags_x != (dc->base.tb->flags & X_FLAG)))) { + if (dc->base.is_jmp == DISAS_NEXT && dc->cpustate_changed) { dc->base.is_jmp = DISAS_UPDATE; tcg_gen_movi_tl(env_pc, dc->pc); } -- 2.25.1 These insns set DISAS_UPDATE without cpustate_changed, which isn't quite right. Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ static int dec_rfe_etc(CPUCRISState *env, DisasContext *dc) cris_evaluate_flags(dc); gen_helper_rfe(cpu_env); dc->base.is_jmp = DISAS_UPDATE; + dc->cpustate_changed = true; break; case 5: /* rfn. */ @@ -XXX,XX +XXX,XX @@ static int dec_rfe_etc(CPUCRISState *env, DisasContext *dc) cris_evaluate_flags(dc); gen_helper_rfn(cpu_env); dc->base.is_jmp = DISAS_UPDATE; + dc->cpustate_changed = true; break; case 6: LOG_DIS("break %d\n", dc->op1); -- 2.25.1 Move this pc update into tb_stop. We will be able to re-use this code shortly. Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 20 +++++++++++++++----- 1 file changed, 15 insertions(+), 5 deletions(-) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ #define BUG() (gen_BUG(dc, __FILE__, __LINE__)) #define BUG_ON(x) ({if (x) BUG();}) -/* is_jmp field values */ -#define DISAS_JUMP DISAS_TARGET_0 /* only pc was modified dynamically */ -#define DISAS_UPDATE DISAS_TARGET_1 /* cpu state was modified dynamically */ +/* + * Target-specific is_jmp field values + */ +/* Only pc was modified dynamically */ +#define DISAS_JUMP DISAS_TARGET_0 +/* Cpu state was modified dynamically, including pc */ +#define DISAS_UPDATE DISAS_TARGET_1 +/* Cpu state was modified dynamically, excluding pc -- use npc */ +#define DISAS_UPDATE_NEXT DISAS_TARGET_2 /* Used by the decoder. */ #define EXTRACT_FIELD(src, start, end) \ @@ -XXX,XX +XXX,XX @@ static void cris_tr_translate_insn(DisasContextBase *dcbase, CPUState *cs) /* Force an update if the per-tb cpu state has changed. */ if (dc->base.is_jmp == DISAS_NEXT && dc->cpustate_changed) { - dc->base.is_jmp = DISAS_UPDATE; - tcg_gen_movi_tl(env_pc, dc->pc); + dc->base.is_jmp = DISAS_UPDATE_NEXT; + return; } /* @@ -XXX,XX +XXX,XX @@ static void cris_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) if (unlikely(dc->base.singlestep_enabled)) { switch (is_jmp) { case DISAS_TOO_MANY: + case DISAS_UPDATE_NEXT: tcg_gen_movi_tl(env_pc, npc); /* fall through */ case DISAS_JUMP: @@ -XXX,XX +XXX,XX @@ static void cris_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) case DISAS_TOO_MANY: gen_goto_tb(dc, 0, npc); break; + case DISAS_UPDATE_NEXT: + tcg_gen_movi_tl(env_pc, npc); + /* fall through */ case DISAS_JUMP: case DISAS_UPDATE: /* Indicate that interupts must be re-evaluated before the next TB. */ -- 2.25.1 Move delayed branch handling to tb_stop, where we can re-use other end-of-tb code, e.g. the evaluation of flags. Honor single stepping. Validate that we aren't losing state by overwriting is_jmp. Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 96 ++++++++++++++++++++++++----------------- 1 file changed, 56 insertions(+), 40 deletions(-) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ #define DISAS_UPDATE DISAS_TARGET_1 /* Cpu state was modified dynamically, excluding pc -- use npc */ #define DISAS_UPDATE_NEXT DISAS_TARGET_2 +/* PC update for delayed branch, see cpustate_changed otherwise */ +#define DISAS_DBRANCH DISAS_TARGET_3 /* Used by the decoder. */ #define EXTRACT_FIELD(src, start, end) \ @@ -XXX,XX +XXX,XX @@ static void cris_tr_translate_insn(DisasContextBase *dcbase, CPUState *cs) dc->cpustate_changed |= dc->flags_x != (dc->base.tb->flags & X_FLAG); /* - * Check for delayed branches here. If we do it before - * actually generating any host code, the simulator will just - * loop doing nothing for on this program location. + * All branches are delayed branches, handled immediately below. + * We don't expect to see odd combinations of exit conditions. */ + assert(dc->base.is_jmp == DISAS_NEXT || dc->cpustate_changed); + if (dc->delayed_branch && --dc->delayed_branch == 0) { - if (dc->base.tb->flags & 7) { - t_gen_movi_env_TN(dslot, 0); - } + dc->base.is_jmp = DISAS_DBRANCH; + return; + } - if (dc->cpustate_changed) { - cris_store_direct_jmp(dc); - } - - if (dc->clear_locked_irq) { - dc->clear_locked_irq = 0; - t_gen_movi_env_TN(locked_irq, 0); - } - - if (dc->jmp == JMP_DIRECT_CC) { - TCGLabel *l1 = gen_new_label(); - cris_evaluate_flags(dc); - - /* Conditional jmp. */ - tcg_gen_brcondi_tl(TCG_COND_EQ, env_btaken, 0, l1); - gen_goto_tb(dc, 1, dc->jmp_pc); - gen_set_label(l1); - gen_goto_tb(dc, 0, dc->pc); - dc->base.is_jmp = DISAS_NORETURN; - dc->jmp = JMP_NOJMP; - } else if (dc->jmp == JMP_DIRECT) { - cris_evaluate_flags(dc); - gen_goto_tb(dc, 0, dc->jmp_pc); - dc->base.is_jmp = DISAS_NORETURN; - dc->jmp = JMP_NOJMP; - } else { - TCGv c = tcg_const_tl(dc->pc); - t_gen_cc_jmp(env_btarget, c); - tcg_temp_free(c); - dc->base.is_jmp = DISAS_JUMP; - } + if (dc->base.is_jmp != DISAS_NEXT) { + return; } /* Force an update if the per-tb cpu state has changed. */ - if (dc->base.is_jmp == DISAS_NEXT && dc->cpustate_changed) { + if (dc->cpustate_changed) { dc->base.is_jmp = DISAS_UPDATE_NEXT; return; } @@ -XXX,XX +XXX,XX @@ static void cris_tr_translate_insn(DisasContextBase *dcbase, CPUState *cs) * If we can detect the length of the next insn easily, we should. * In the meantime, simply stop when we do cross. */ - if (dc->base.is_jmp == DISAS_NEXT - && ((dc->pc ^ dc->base.pc_first) & TARGET_PAGE_MASK) != 0) { + if ((dc->pc ^ dc->base.pc_first) & TARGET_PAGE_MASK) { dc->base.is_jmp = DISAS_TOO_MANY; } } @@ -XXX,XX +XXX,XX @@ static void cris_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) cris_evaluate_flags(dc); + /* Evaluate delayed branch destination and fold to another is_jmp case. */ + if (is_jmp == DISAS_DBRANCH) { + if (dc->base.tb->flags & 7) { + t_gen_movi_env_TN(dslot, 0); + } + + switch (dc->jmp) { + case JMP_DIRECT: + npc = dc->jmp_pc; + is_jmp = dc->cpustate_changed ? DISAS_UPDATE_NEXT : DISAS_TOO_MANY; + break; + + case JMP_DIRECT_CC: + /* + * Use a conditional branch if either taken or not-taken path + * can use goto_tb. If neither can, then treat it as indirect. + */ + if (likely(!dc->base.singlestep_enabled) + && likely(!dc->cpustate_changed) + && (use_goto_tb(dc, dc->jmp_pc) || use_goto_tb(dc, npc))) { + TCGLabel *not_taken = gen_new_label(); + + tcg_gen_brcondi_tl(TCG_COND_EQ, env_btaken, 0, not_taken); + gen_goto_tb(dc, 1, dc->jmp_pc); + gen_set_label(not_taken); + + /* not-taken case handled below. */ + is_jmp = DISAS_TOO_MANY; + break; + } + tcg_gen_movi_tl(env_btarget, dc->jmp_pc); + /* fall through */ + + case JMP_INDIRECT: + t_gen_cc_jmp(env_btarget, tcg_constant_tl(npc)); + is_jmp = dc->cpustate_changed ? DISAS_UPDATE : DISAS_JUMP; + break; + + default: + g_assert_not_reached(); + } + } + if (unlikely(dc->base.singlestep_enabled)) { switch (is_jmp) { case DISAS_TOO_MANY: -- 2.25.1 We can use this in gen_goto_tb and for DISAS_JUMP to indirectly chain to the next TB. Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ static void gen_goto_tb(DisasContext *dc, int n, target_ulong dest) tcg_gen_exit_tb(dc->base.tb, n); } else { tcg_gen_movi_tl(env_pc, dest); - tcg_gen_exit_tb(NULL, 0); + tcg_gen_lookup_and_goto_ptr(); } } @@ -XXX,XX +XXX,XX @@ static void cris_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) tcg_gen_movi_tl(env_pc, npc); /* fall through */ case DISAS_JUMP: + tcg_gen_lookup_and_goto_ptr(); + break; case DISAS_UPDATE: /* Indicate that interupts must be re-evaluated before the next TB. */ tcg_gen_exit_tb(NULL, 0); -- 2.25.1 Use movcond instead of brcond to set env_pc. Discard the btarget and btaken variables to improve register allocation and avoid unnecessary writeback. Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 22 ++++++++++------------ 1 file changed, 10 insertions(+), 12 deletions(-) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ static void t_gen_swapr(TCGv d, TCGv s) tcg_temp_free(org_s); } -static void t_gen_cc_jmp(TCGv pc_true, TCGv pc_false) -{ - TCGLabel *l1 = gen_new_label(); - - /* Conditional jmp. */ - tcg_gen_mov_tl(env_pc, pc_false); - tcg_gen_brcondi_tl(TCG_COND_EQ, env_btaken, 0, l1); - tcg_gen_mov_tl(env_pc, pc_true); - gen_set_label(l1); -} - static bool use_goto_tb(DisasContext *dc, target_ulong dest) { return ((dest ^ dc->base.pc_first) & TARGET_PAGE_MASK) == 0; @@ -XXX,XX +XXX,XX @@ static void cris_tr_tb_stop(DisasContextBase *dcbase, CPUState *cpu) /* fall through */ case JMP_INDIRECT: - t_gen_cc_jmp(env_btarget, tcg_constant_tl(npc)); + tcg_gen_movcond_tl(TCG_COND_NE, env_pc, + env_btaken, tcg_constant_tl(0), + env_btarget, tcg_constant_tl(npc)); is_jmp = dc->cpustate_changed ? DISAS_UPDATE : DISAS_JUMP; + + /* + * We have now consumed btaken and btarget. Hint to the + * tcg compiler that the writeback to env may be dropped. + */ + tcg_gen_discard_tl(env_btaken); + tcg_gen_discard_tl(env_btarget); break; default: -- 2.25.1 Ever since 2a44f7f17364, flagx_known is always true. Fold away all of the tests against the flag. Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 99 ++++++++------------------------- target/cris/translate_v10.c.inc | 6 +- 2 files changed, 24 insertions(+), 81 deletions(-) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ typedef struct DisasContext { int cc_x_uptodate; /* 1 - ccs, 2 - known | X_FLAG. 0 not up-to-date. */ int flags_uptodate; /* Whether or not $ccs is up-to-date. */ - int flagx_known; /* Whether or not flags_x has the x flag known at - translation time. */ int flags_x; int clear_x; /* Clear x after this insn? */ @@ -XXX,XX +XXX,XX @@ static inline void t_gen_add_flag(TCGv d, int flag) static inline void t_gen_addx_carry(DisasContext *dc, TCGv d) { - if (dc->flagx_known) { - if (dc->flags_x) { - TCGv c; - - c = tcg_temp_new(); - t_gen_mov_TN_preg(c, PR_CCS); - /* C flag is already at bit 0. */ - tcg_gen_andi_tl(c, c, C_FLAG); - tcg_gen_add_tl(d, d, c); - tcg_temp_free(c); - } - } else { - TCGv x, c; + if (dc->flags_x) { + TCGv c = tcg_temp_new(); - x = tcg_temp_new(); - c = tcg_temp_new(); - t_gen_mov_TN_preg(x, PR_CCS); - tcg_gen_mov_tl(c, x); - - /* Propagate carry into d if X is set. Branch free. */ + t_gen_mov_TN_preg(c, PR_CCS); + /* C flag is already at bit 0. */ tcg_gen_andi_tl(c, c, C_FLAG); - tcg_gen_andi_tl(x, x, X_FLAG); - tcg_gen_shri_tl(x, x, 4); - - tcg_gen_and_tl(x, x, c); - tcg_gen_add_tl(d, d, x); - tcg_temp_free(x); + tcg_gen_add_tl(d, d, c); tcg_temp_free(c); } } static inline void t_gen_subx_carry(DisasContext *dc, TCGv d) { - if (dc->flagx_known) { - if (dc->flags_x) { - TCGv c; - - c = tcg_temp_new(); - t_gen_mov_TN_preg(c, PR_CCS); - /* C flag is already at bit 0. */ - tcg_gen_andi_tl(c, c, C_FLAG); - tcg_gen_sub_tl(d, d, c); - tcg_temp_free(c); - } - } else { - TCGv x, c; + if (dc->flags_x) { + TCGv c = tcg_temp_new(); - x = tcg_temp_new(); - c = tcg_temp_new(); - t_gen_mov_TN_preg(x, PR_CCS); - tcg_gen_mov_tl(c, x); - - /* Propagate carry into d if X is set. Branch free. */ + t_gen_mov_TN_preg(c, PR_CCS); + /* C flag is already at bit 0. */ tcg_gen_andi_tl(c, c, C_FLAG); - tcg_gen_andi_tl(x, x, X_FLAG); - tcg_gen_shri_tl(x, x, 4); - - tcg_gen_and_tl(x, x, c); - tcg_gen_sub_tl(d, d, x); - tcg_temp_free(x); + tcg_gen_sub_tl(d, d, c); tcg_temp_free(c); } } @@ -XXX,XX +XXX,XX @@ static void gen_goto_tb(DisasContext *dc, int n, target_ulong dest) static inline void cris_clear_x_flag(DisasContext *dc) { - if (dc->flagx_known && dc->flags_x) { + if (dc->flags_x) { dc->flags_uptodate = 0; } - - dc->flagx_known = 1; dc->flags_x = 0; } @@ -XXX,XX +XXX,XX @@ static void cris_evaluate_flags(DisasContext *dc) break; } - if (dc->flagx_known) { - if (dc->flags_x) { - tcg_gen_ori_tl(cpu_PR[PR_CCS], cpu_PR[PR_CCS], X_FLAG); - } else if (dc->cc_op == CC_OP_FLAGS) { - tcg_gen_andi_tl(cpu_PR[PR_CCS], cpu_PR[PR_CCS], ~X_FLAG); - } + if (dc->flags_x) { + tcg_gen_ori_tl(cpu_PR[PR_CCS], cpu_PR[PR_CCS], X_FLAG); + } else if (dc->cc_op == CC_OP_FLAGS) { + tcg_gen_andi_tl(cpu_PR[PR_CCS], cpu_PR[PR_CCS], ~X_FLAG); } dc->flags_uptodate = 1; } @@ -XXX,XX +XXX,XX @@ static void cris_update_cc_op(DisasContext *dc, int op, int size) static inline void cris_update_cc_x(DisasContext *dc) { /* Save the x flag state at the time of the cc snapshot. */ - if (dc->flagx_known) { - if (dc->cc_x_uptodate == (2 | dc->flags_x)) { - return; - } - tcg_gen_movi_tl(cc_x, dc->flags_x); - dc->cc_x_uptodate = 2 | dc->flags_x; - } else { - tcg_gen_andi_tl(cc_x, cpu_PR[PR_CCS], X_FLAG); - dc->cc_x_uptodate = 1; + if (dc->cc_x_uptodate == (2 | dc->flags_x)) { + return; } + tcg_gen_movi_tl(cc_x, dc->flags_x); + dc->cc_x_uptodate = 2 | dc->flags_x; } /* Update cc prior to executing ALU op. Needs source operands untouched. */ @@ -XXX,XX +XXX,XX @@ static void gen_store (DisasContext *dc, TCGv addr, TCGv val, /* Conditional writes. We only support the kind were X and P are known at translation time. */ - if (dc->flagx_known && dc->flags_x && (dc->tb_flags & P_FLAG)) { + if (dc->flags_x && (dc->tb_flags & P_FLAG)) { dc->postinc = 0; cris_evaluate_flags(dc); tcg_gen_ori_tl(cpu_PR[PR_CCS], cpu_PR[PR_CCS], C_FLAG); @@ -XXX,XX +XXX,XX @@ static void gen_store (DisasContext *dc, TCGv addr, TCGv val, tcg_gen_qemu_st_tl(val, addr, mem_index, MO_TE + ctz32(size)); - if (dc->flagx_known && dc->flags_x) { + if (dc->flags_x) { cris_evaluate_flags(dc); tcg_gen_andi_tl(cpu_PR[PR_CCS], cpu_PR[PR_CCS], ~C_FLAG); } @@ -XXX,XX +XXX,XX @@ static int dec_addc_r(CPUCRISState *env, DisasContext *dc) LOG_DIS("addc $r%u, $r%u\n", dc->op1, dc->op2); cris_evaluate_flags(dc); + /* Set for this insn. */ - dc->flagx_known = 1; dc->flags_x = X_FLAG; cris_cc_mask(dc, CC_MASK_NZVC); @@ -XXX,XX +XXX,XX @@ static int dec_setclrf(CPUCRISState *env, DisasContext *dc) } if (flags & X_FLAG) { - dc->flagx_known = 1; if (set) { dc->flags_x = X_FLAG; } else { @@ -XXX,XX +XXX,XX @@ static int dec_addc_mr(CPUCRISState *env, DisasContext *dc) cris_evaluate_flags(dc); /* Set for this insn. */ - dc->flagx_known = 1; dc->flags_x = X_FLAG; cris_alu_m_alloc_temps(t); @@ -XXX,XX +XXX,XX @@ static void cris_tr_init_disas_context(DisasContextBase *dcbase, CPUState *cs) dc->ppc = pc_start; dc->pc = pc_start; dc->flags_uptodate = 1; - dc->flagx_known = 1; dc->flags_x = tb_flags & X_FLAG; dc->cc_x_uptodate = 0; dc->cc_mask = 0; @@ -XXX,XX +XXX,XX @@ static void cris_tr_translate_insn(DisasContextBase *dcbase, CPUState *cs) } /* Fold unhandled changes to X_FLAG into cpustate_changed. */ - dc->cpustate_changed |= !dc->flagx_known; dc->cpustate_changed |= dc->flags_x != (dc->base.tb->flags & X_FLAG); /* diff --git a/target/cris/translate_v10.c.inc b/target/cris/translate_v10.c.inc index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate_v10.c.inc +++ b/target/cris/translate_v10.c.inc @@ -XXX,XX +XXX,XX @@ static void gen_store_v10(DisasContext *dc, TCGv addr, TCGv val, cris_store_direct_jmp(dc); } - /* Conditional writes. We only support the kind were X is known - at translation time. */ - if (dc->flagx_known && dc->flags_x) { + /* Conditional writes. */ + if (dc->flags_x) { gen_store_v10_conditional(dc, addr, val, size, mem_index); return; } @@ -XXX,XX +XXX,XX @@ static unsigned int dec10_setclrf(DisasContext *dc) if (flags & X_FLAG) { - dc->flagx_known = 1; if (set) dc->flags_x = X_FLAG; else -- 2.25.1 We always know the exact value of X, that's all that matters. This avoids splitting the TB e.g. between "ax" and "addq". Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/cris/translate.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/target/cris/translate.c b/target/cris/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/cris/translate.c +++ b/target/cris/translate.c @@ -XXX,XX +XXX,XX @@ static void cris_tr_translate_insn(DisasContextBase *dcbase, CPUState *cs) cris_clear_x_flag(dc); } - /* Fold unhandled changes to X_FLAG into cpustate_changed. */ - dc->cpustate_changed |= dc->flags_x != (dc->base.tb->flags & X_FLAG); - /* * All branches are delayed branches, handled immediately below. * We don't expect to see odd combinations of exit conditions. -- 2.25.1 From: LIU Zhiwei <zhiwei_liu@c-sky.com> Implement tcg_gen_vec_add{sub}16_tl by adding corresponding i32 OP. Signed-off-by: LIU Zhiwei <zhiwei_liu@c-sky.com> Message-Id: <20210624105023.3852-2-zhiwei_liu@c-sky.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/tcg/tcg-op-gvec.h | 13 +++++++++++++ tcg/tcg-op-gvec.c | 28 ++++++++++++++++++++++++++++ 2 files changed, 41 insertions(+) diff --git a/include/tcg/tcg-op-gvec.h b/include/tcg/tcg-op-gvec.h index XXXXXXX..XXXXXXX 100644 --- a/include/tcg/tcg-op-gvec.h +++ b/include/tcg/tcg-op-gvec.h @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_sar16i_i64(TCGv_i64 d, TCGv_i64 a, int64_t); void tcg_gen_vec_rotl8i_i64(TCGv_i64 d, TCGv_i64 a, int64_t c); void tcg_gen_vec_rotl16i_i64(TCGv_i64 d, TCGv_i64 a, int64_t c); +/* 32-bit vector operations. */ +void tcg_gen_vec_add16_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b); + +void tcg_gen_vec_sub16_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b); + +#if TARGET_LONG_BITS == 64 +#define tcg_gen_vec_add16_tl tcg_gen_vec_add16_i64 +#define tcg_gen_vec_sub16_tl tcg_gen_vec_sub16_i64 +#else +#define tcg_gen_vec_add16_tl tcg_gen_vec_add16_i32 +#define tcg_gen_vec_sub16_tl tcg_gen_vec_sub16_i32 +#endif + #endif diff --git a/tcg/tcg-op-gvec.c b/tcg/tcg-op-gvec.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg-op-gvec.c +++ b/tcg/tcg-op-gvec.c @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_add16_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b) gen_addv_mask(d, a, b, m); } +void tcg_gen_vec_add16_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b) +{ + TCGv_i32 t1 = tcg_temp_new_i32(); + TCGv_i32 t2 = tcg_temp_new_i32(); + + tcg_gen_andi_i32(t1, a, ~0xffff); + tcg_gen_add_i32(t2, a, b); + tcg_gen_add_i32(t1, t1, b); + tcg_gen_deposit_i32(d, t1, t2, 0, 16); + + tcg_temp_free_i32(t1); + tcg_temp_free_i32(t2); +} + void tcg_gen_vec_add32_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b) { TCGv_i64 t1 = tcg_temp_new_i64(); @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_sub16_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b) gen_subv_mask(d, a, b, m); } +void tcg_gen_vec_sub16_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b) +{ + TCGv_i32 t1 = tcg_temp_new_i32(); + TCGv_i32 t2 = tcg_temp_new_i32(); + + tcg_gen_andi_i32(t1, b, ~0xffff); + tcg_gen_sub_i32(t2, a, b); + tcg_gen_sub_i32(t1, a, t1); + tcg_gen_deposit_i32(d, t1, t2, 0, 16); + + tcg_temp_free_i32(t1); + tcg_temp_free_i32(t2); +} + void tcg_gen_vec_sub32_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b) { TCGv_i64 t1 = tcg_temp_new_i64(); -- 2.25.1 From: LIU Zhiwei <zhiwei_liu@c-sky.com> Implement tcg_gen_vec_add{sub}8_tl by adding corresponging i32 OP. Signed-off-by: LIU Zhiwei <zhiwei_liu@c-sky.com> Message-Id: <20210624105023.3852-3-zhiwei_liu@c-sky.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/tcg/tcg-op-gvec.h | 6 ++++++ tcg/tcg-op-gvec.c | 38 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 44 insertions(+) diff --git a/include/tcg/tcg-op-gvec.h b/include/tcg/tcg-op-gvec.h index XXXXXXX..XXXXXXX 100644 --- a/include/tcg/tcg-op-gvec.h +++ b/include/tcg/tcg-op-gvec.h @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_rotl8i_i64(TCGv_i64 d, TCGv_i64 a, int64_t c); void tcg_gen_vec_rotl16i_i64(TCGv_i64 d, TCGv_i64 a, int64_t c); /* 32-bit vector operations. */ +void tcg_gen_vec_add8_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b); void tcg_gen_vec_add16_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b); +void tcg_gen_vec_sub8_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b); void tcg_gen_vec_sub16_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b); #if TARGET_LONG_BITS == 64 +#define tcg_gen_vec_add8_tl tcg_gen_vec_add8_i64 +#define tcg_gen_vec_sub8_tl tcg_gen_vec_sub8_i64 #define tcg_gen_vec_add16_tl tcg_gen_vec_add16_i64 #define tcg_gen_vec_sub16_tl tcg_gen_vec_sub16_i64 #else +#define tcg_gen_vec_add8_tl tcg_gen_vec_add8_i32 +#define tcg_gen_vec_sub8_tl tcg_gen_vec_sub8_i32 #define tcg_gen_vec_add16_tl tcg_gen_vec_add16_i32 #define tcg_gen_vec_sub16_tl tcg_gen_vec_sub16_i32 #endif diff --git a/tcg/tcg-op-gvec.c b/tcg/tcg-op-gvec.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg-op-gvec.c +++ b/tcg/tcg-op-gvec.c @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_add8_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b) gen_addv_mask(d, a, b, m); } +void tcg_gen_vec_add8_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b) +{ + TCGv_i32 m = tcg_constant_i32((int32_t)dup_const(MO_8, 0x80)); + TCGv_i32 t1 = tcg_temp_new_i32(); + TCGv_i32 t2 = tcg_temp_new_i32(); + TCGv_i32 t3 = tcg_temp_new_i32(); + + tcg_gen_andc_i32(t1, a, m); + tcg_gen_andc_i32(t2, b, m); + tcg_gen_xor_i32(t3, a, b); + tcg_gen_add_i32(d, t1, t2); + tcg_gen_and_i32(t3, t3, m); + tcg_gen_xor_i32(d, d, t3); + + tcg_temp_free_i32(t1); + tcg_temp_free_i32(t2); + tcg_temp_free_i32(t3); +} + void tcg_gen_vec_add16_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b) { TCGv_i64 m = tcg_constant_i64(dup_const(MO_16, 0x8000)); @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_sub8_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b) gen_subv_mask(d, a, b, m); } +void tcg_gen_vec_sub8_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b) +{ + TCGv_i32 m = tcg_constant_i32((int32_t)dup_const(MO_8, 0x80)); + TCGv_i32 t1 = tcg_temp_new_i32(); + TCGv_i32 t2 = tcg_temp_new_i32(); + TCGv_i32 t3 = tcg_temp_new_i32(); + + tcg_gen_or_i32(t1, a, m); + tcg_gen_andc_i32(t2, b, m); + tcg_gen_eqv_i32(t3, a, b); + tcg_gen_sub_i32(d, t1, t2); + tcg_gen_and_i32(t3, t3, m); + tcg_gen_xor_i32(d, d, t3); + + tcg_temp_free_i32(t1); + tcg_temp_free_i32(t2); + tcg_temp_free_i32(t3); +} + void tcg_gen_vec_sub16_i64(TCGv_i64 d, TCGv_i64 a, TCGv_i64 b) { TCGv_i64 m = tcg_constant_i64(dup_const(MO_16, 0x8000)); -- 2.25.1 From: LIU Zhiwei <zhiwei_liu@c-sky.com> Implement tcg_gen_vec_shl{shr}{sar}16i_tl by adding corresponging i32 OP. Signed-off-by: LIU Zhiwei <zhiwei_liu@c-sky.com> Message-Id: <20210624105023.3852-4-zhiwei_liu@c-sky.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/tcg/tcg-op-gvec.h | 10 ++++++++++ tcg/tcg-op-gvec.c | 28 ++++++++++++++++++++++++++++ 2 files changed, 38 insertions(+) diff --git a/include/tcg/tcg-op-gvec.h b/include/tcg/tcg-op-gvec.h index XXXXXXX..XXXXXXX 100644 --- a/include/tcg/tcg-op-gvec.h +++ b/include/tcg/tcg-op-gvec.h @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_add16_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b); void tcg_gen_vec_sub8_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b); void tcg_gen_vec_sub16_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b); +void tcg_gen_vec_shl16i_i32(TCGv_i32 d, TCGv_i32 a, int32_t); +void tcg_gen_vec_shr16i_i32(TCGv_i32 d, TCGv_i32 a, int32_t); +void tcg_gen_vec_sar16i_i32(TCGv_i32 d, TCGv_i32 a, int32_t); + #if TARGET_LONG_BITS == 64 #define tcg_gen_vec_add8_tl tcg_gen_vec_add8_i64 #define tcg_gen_vec_sub8_tl tcg_gen_vec_sub8_i64 #define tcg_gen_vec_add16_tl tcg_gen_vec_add16_i64 #define tcg_gen_vec_sub16_tl tcg_gen_vec_sub16_i64 +#define tcg_gen_vec_shl16i_tl tcg_gen_vec_shl16i_i64 +#define tcg_gen_vec_shr16i_tl tcg_gen_vec_shr16i_i64 +#define tcg_gen_vec_sar16i_tl tcg_gen_vec_sar16i_i64 #else #define tcg_gen_vec_add8_tl tcg_gen_vec_add8_i32 #define tcg_gen_vec_sub8_tl tcg_gen_vec_sub8_i32 #define tcg_gen_vec_add16_tl tcg_gen_vec_add16_i32 #define tcg_gen_vec_sub16_tl tcg_gen_vec_sub16_i32 +#define tcg_gen_vec_shl16i_tl tcg_gen_vec_shl16i_i32 +#define tcg_gen_vec_shr16i_tl tcg_gen_vec_shr16i_i32 +#define tcg_gen_vec_sar16i_tl tcg_gen_vec_sar16i_i32 #endif #endif diff --git a/tcg/tcg-op-gvec.c b/tcg/tcg-op-gvec.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg-op-gvec.c +++ b/tcg/tcg-op-gvec.c @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_shl16i_i64(TCGv_i64 d, TCGv_i64 a, int64_t c) tcg_gen_andi_i64(d, d, mask); } +void tcg_gen_vec_shl16i_i32(TCGv_i32 d, TCGv_i32 a, int32_t c) +{ + uint32_t mask = dup_const(MO_16, 0xffff << c); + tcg_gen_shli_i32(d, a, c); + tcg_gen_andi_i32(d, d, mask); +} + void tcg_gen_gvec_shli(unsigned vece, uint32_t dofs, uint32_t aofs, int64_t shift, uint32_t oprsz, uint32_t maxsz) { @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_shr16i_i64(TCGv_i64 d, TCGv_i64 a, int64_t c) tcg_gen_andi_i64(d, d, mask); } +void tcg_gen_vec_shr16i_i32(TCGv_i32 d, TCGv_i32 a, int32_t c) +{ + uint32_t mask = dup_const(MO_16, 0xffff >> c); + tcg_gen_shri_i32(d, a, c); + tcg_gen_andi_i32(d, d, mask); +} + void tcg_gen_gvec_shri(unsigned vece, uint32_t dofs, uint32_t aofs, int64_t shift, uint32_t oprsz, uint32_t maxsz) { @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_sar16i_i64(TCGv_i64 d, TCGv_i64 a, int64_t c) tcg_temp_free_i64(s); } +void tcg_gen_vec_sar16i_i32(TCGv_i32 d, TCGv_i32 a, int32_t c) +{ + uint32_t s_mask = dup_const(MO_16, 0x8000 >> c); + uint32_t c_mask = dup_const(MO_16, 0xffff >> c); + TCGv_i32 s = tcg_temp_new_i32(); + + tcg_gen_shri_i32(d, a, c); + tcg_gen_andi_i32(s, d, s_mask); /* isolate (shifted) sign bit */ + tcg_gen_andi_i32(d, d, c_mask); /* clear out bits above sign */ + tcg_gen_muli_i32(s, s, (2 << c) - 2); /* replicate isolated signs */ + tcg_gen_or_i32(d, d, s); /* include sign extension */ + tcg_temp_free_i32(s); +} + void tcg_gen_gvec_sari(unsigned vece, uint32_t dofs, uint32_t aofs, int64_t shift, uint32_t oprsz, uint32_t maxsz) { -- 2.25.1 From: LIU Zhiwei <zhiwei_liu@c-sky.com> Implement tcg_gen_vec_shl{shr}{sar}8i_tl by adding corresponging i32 OP. Signed-off-by: LIU Zhiwei <zhiwei_liu@c-sky.com> Message-Id: <20210624105023.3852-5-zhiwei_liu@c-sky.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/tcg/tcg-op-gvec.h | 10 ++++++++++ tcg/tcg-op-gvec.c | 28 ++++++++++++++++++++++++++++ 2 files changed, 38 insertions(+) diff --git a/include/tcg/tcg-op-gvec.h b/include/tcg/tcg-op-gvec.h index XXXXXXX..XXXXXXX 100644 --- a/include/tcg/tcg-op-gvec.h +++ b/include/tcg/tcg-op-gvec.h @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_add16_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b); void tcg_gen_vec_sub8_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b); void tcg_gen_vec_sub16_i32(TCGv_i32 d, TCGv_i32 a, TCGv_i32 b); +void tcg_gen_vec_shl8i_i32(TCGv_i32 d, TCGv_i32 a, int32_t); void tcg_gen_vec_shl16i_i32(TCGv_i32 d, TCGv_i32 a, int32_t); +void tcg_gen_vec_shr8i_i32(TCGv_i32 d, TCGv_i32 a, int32_t); void tcg_gen_vec_shr16i_i32(TCGv_i32 d, TCGv_i32 a, int32_t); +void tcg_gen_vec_sar8i_i32(TCGv_i32 d, TCGv_i32 a, int32_t); void tcg_gen_vec_sar16i_i32(TCGv_i32 d, TCGv_i32 a, int32_t); #if TARGET_LONG_BITS == 64 @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_sar16i_i32(TCGv_i32 d, TCGv_i32 a, int32_t); #define tcg_gen_vec_sub8_tl tcg_gen_vec_sub8_i64 #define tcg_gen_vec_add16_tl tcg_gen_vec_add16_i64 #define tcg_gen_vec_sub16_tl tcg_gen_vec_sub16_i64 +#define tcg_gen_vec_shl8i_tl tcg_gen_vec_shl8i_i64 +#define tcg_gen_vec_shr8i_tl tcg_gen_vec_shr8i_i64 +#define tcg_gen_vec_sar8i_tl tcg_gen_vec_sar8i_i64 #define tcg_gen_vec_shl16i_tl tcg_gen_vec_shl16i_i64 #define tcg_gen_vec_shr16i_tl tcg_gen_vec_shr16i_i64 #define tcg_gen_vec_sar16i_tl tcg_gen_vec_sar16i_i64 + #else #define tcg_gen_vec_add8_tl tcg_gen_vec_add8_i32 #define tcg_gen_vec_sub8_tl tcg_gen_vec_sub8_i32 #define tcg_gen_vec_add16_tl tcg_gen_vec_add16_i32 #define tcg_gen_vec_sub16_tl tcg_gen_vec_sub16_i32 +#define tcg_gen_vec_shl8i_tl tcg_gen_vec_shl8i_i32 +#define tcg_gen_vec_shr8i_tl tcg_gen_vec_shr8i_i32 +#define tcg_gen_vec_sar8i_tl tcg_gen_vec_sar8i_i32 #define tcg_gen_vec_shl16i_tl tcg_gen_vec_shl16i_i32 #define tcg_gen_vec_shr16i_tl tcg_gen_vec_shr16i_i32 #define tcg_gen_vec_sar16i_tl tcg_gen_vec_sar16i_i32 diff --git a/tcg/tcg-op-gvec.c b/tcg/tcg-op-gvec.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg-op-gvec.c +++ b/tcg/tcg-op-gvec.c @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_shl16i_i64(TCGv_i64 d, TCGv_i64 a, int64_t c) tcg_gen_andi_i64(d, d, mask); } +void tcg_gen_vec_shl8i_i32(TCGv_i32 d, TCGv_i32 a, int32_t c) +{ + uint32_t mask = dup_const(MO_8, 0xff << c); + tcg_gen_shli_i32(d, a, c); + tcg_gen_andi_i32(d, d, mask); +} + void tcg_gen_vec_shl16i_i32(TCGv_i32 d, TCGv_i32 a, int32_t c) { uint32_t mask = dup_const(MO_16, 0xffff << c); @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_shr16i_i64(TCGv_i64 d, TCGv_i64 a, int64_t c) tcg_gen_andi_i64(d, d, mask); } +void tcg_gen_vec_shr8i_i32(TCGv_i32 d, TCGv_i32 a, int32_t c) +{ + uint32_t mask = dup_const(MO_8, 0xff >> c); + tcg_gen_shri_i32(d, a, c); + tcg_gen_andi_i32(d, d, mask); +} + void tcg_gen_vec_shr16i_i32(TCGv_i32 d, TCGv_i32 a, int32_t c) { uint32_t mask = dup_const(MO_16, 0xffff >> c); @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_sar16i_i64(TCGv_i64 d, TCGv_i64 a, int64_t c) tcg_temp_free_i64(s); } +void tcg_gen_vec_sar8i_i32(TCGv_i32 d, TCGv_i32 a, int32_t c) +{ + uint32_t s_mask = dup_const(MO_8, 0x80 >> c); + uint32_t c_mask = dup_const(MO_8, 0xff >> c); + TCGv_i32 s = tcg_temp_new_i32(); + + tcg_gen_shri_i32(d, a, c); + tcg_gen_andi_i32(s, d, s_mask); /* isolate (shifted) sign bit */ + tcg_gen_muli_i32(s, s, (2 << c) - 2); /* replicate isolated signs */ + tcg_gen_andi_i32(d, d, c_mask); /* clear out bits above sign */ + tcg_gen_or_i32(d, d, s); /* include sign extension */ + tcg_temp_free_i32(s); +} + void tcg_gen_vec_sar16i_i32(TCGv_i32 d, TCGv_i32 a, int32_t c) { uint32_t s_mask = dup_const(MO_16, 0x8000 >> c); -- 2.25.1 From: LIU Zhiwei <zhiwei_liu@c-sky.com> Signed-off-by: LIU Zhiwei <zhiwei_liu@c-sky.com> Message-Id: <20210624105023.3852-6-zhiwei_liu@c-sky.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/tcg/tcg-op-gvec.h | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/include/tcg/tcg-op-gvec.h b/include/tcg/tcg-op-gvec.h index XXXXXXX..XXXXXXX 100644 --- a/include/tcg/tcg-op-gvec.h +++ b/include/tcg/tcg-op-gvec.h @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_sar16i_i32(TCGv_i32 d, TCGv_i32 a, int32_t); #define tcg_gen_vec_sub8_tl tcg_gen_vec_sub8_i64 #define tcg_gen_vec_add16_tl tcg_gen_vec_add16_i64 #define tcg_gen_vec_sub16_tl tcg_gen_vec_sub16_i64 +#define tcg_gen_vec_add32_tl tcg_gen_vec_add32_i64 +#define tcg_gen_vec_sub32_tl tcg_gen_vec_sub32_i64 #define tcg_gen_vec_shl8i_tl tcg_gen_vec_shl8i_i64 #define tcg_gen_vec_shr8i_tl tcg_gen_vec_shr8i_i64 #define tcg_gen_vec_sar8i_tl tcg_gen_vec_sar8i_i64 @@ -XXX,XX +XXX,XX @@ void tcg_gen_vec_sar16i_i32(TCGv_i32 d, TCGv_i32 a, int32_t); #define tcg_gen_vec_sub8_tl tcg_gen_vec_sub8_i32 #define tcg_gen_vec_add16_tl tcg_gen_vec_add16_i32 #define tcg_gen_vec_sub16_tl tcg_gen_vec_sub16_i32 +#define tcg_gen_vec_add32_tl tcg_gen_add_i32 +#define tcg_gen_vec_sub32_tl tcg_gen_sub_i32 #define tcg_gen_vec_shl8i_tl tcg_gen_vec_shl8i_i32 #define tcg_gen_vec_shr8i_tl tcg_gen_vec_shr8i_i32 #define tcg_gen_vec_sar8i_tl tcg_gen_vec_sar8i_i32 -- 2.25.1 From: Warner Losh <imp@bsdimp.com> The trap number for a page fault on BSD systems is T_PAGEFLT not 0xe -- 0xe is used by Linux and represents the intel hardware trap vector. The BSD kernels, however, translate this to T_PAGEFLT in their Xpage, Xtrap0e, Xtrap14, etc fault handlers. This is true for i386 and x86_64, though the name of the trap hanlder can very on the flavor of BSD. As far as I can tell, Linux doesn't provide a define for this value. Invent a new one (PAGE_FAULT_TRAP) and use it instead to avoid uglier ifdefs. Signed-off-by: Mark Johnston <markj@FreeBSD.org> Signed-off-by: Juergen Lock <nox@FreeBSD.org> [ Rework to avoid ifdefs and expand it to i386 ] Signed-off-by: Warner Losh <imp@bsdimp.com> Message-Id: <20210625045707.84534-3-imp@bsdimp.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- accel/tcg/user-exec.c | 20 ++++++++++++++++++-- 1 file changed, 18 insertions(+), 2 deletions(-) diff --git a/accel/tcg/user-exec.c b/accel/tcg/user-exec.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/user-exec.c +++ b/accel/tcg/user-exec.c @@ -XXX,XX +XXX,XX @@ void *probe_access(CPUArchState *env, target_ulong addr, int size, #if defined(__NetBSD__) #include <ucontext.h> +#include <machine/trap.h> #define EIP_sig(context) ((context)->uc_mcontext.__gregs[_REG_EIP]) #define TRAP_sig(context) ((context)->uc_mcontext.__gregs[_REG_TRAPNO]) #define ERROR_sig(context) ((context)->uc_mcontext.__gregs[_REG_ERR]) #define MASK_sig(context) ((context)->uc_sigmask) +#define PAGE_FAULT_TRAP T_PAGEFLT #elif defined(__FreeBSD__) || defined(__DragonFly__) #include <ucontext.h> +#include <machine/trap.h> #define EIP_sig(context) (*((unsigned long *)&(context)->uc_mcontext.mc_eip)) #define TRAP_sig(context) ((context)->uc_mcontext.mc_trapno) #define ERROR_sig(context) ((context)->uc_mcontext.mc_err) #define MASK_sig(context) ((context)->uc_sigmask) +#define PAGE_FAULT_TRAP T_PAGEFLT #elif defined(__OpenBSD__) +#include <machine/trap.h> #define EIP_sig(context) ((context)->sc_eip) #define TRAP_sig(context) ((context)->sc_trapno) #define ERROR_sig(context) ((context)->sc_err) #define MASK_sig(context) ((context)->sc_mask) +#define PAGE_FAULT_TRAP T_PAGEFLT #else #define EIP_sig(context) ((context)->uc_mcontext.gregs[REG_EIP]) #define TRAP_sig(context) ((context)->uc_mcontext.gregs[REG_TRAPNO]) #define ERROR_sig(context) ((context)->uc_mcontext.gregs[REG_ERR]) #define MASK_sig(context) ((context)->uc_sigmask) +#define PAGE_FAULT_TRAP 0xe #endif int cpu_signal_handler(int host_signum, void *pinfo, @@ -XXX,XX +XXX,XX @@ int cpu_signal_handler(int host_signum, void *pinfo, pc = EIP_sig(uc); trapno = TRAP_sig(uc); return handle_cpu_signal(pc, info, - trapno == 0xe ? (ERROR_sig(uc) >> 1) & 1 : 0, + trapno == PAGE_FAULT_TRAP ? + (ERROR_sig(uc) >> 1) & 1 : 0, &MASK_sig(uc)); } #elif defined(__x86_64__) #ifdef __NetBSD__ +#include <machine/trap.h> #define PC_sig(context) _UC_MACHINE_PC(context) #define TRAP_sig(context) ((context)->uc_mcontext.__gregs[_REG_TRAPNO]) #define ERROR_sig(context) ((context)->uc_mcontext.__gregs[_REG_ERR]) #define MASK_sig(context) ((context)->uc_sigmask) +#define PAGE_FAULT_TRAP T_PAGEFLT #elif defined(__OpenBSD__) +#include <machine/trap.h> #define PC_sig(context) ((context)->sc_rip) #define TRAP_sig(context) ((context)->sc_trapno) #define ERROR_sig(context) ((context)->sc_err) #define MASK_sig(context) ((context)->sc_mask) +#define PAGE_FAULT_TRAP T_PAGEFLT #elif defined(__FreeBSD__) || defined(__DragonFly__) #include <ucontext.h> +#include <machine/trap.h> #define PC_sig(context) (*((unsigned long *)&(context)->uc_mcontext.mc_rip)) #define TRAP_sig(context) ((context)->uc_mcontext.mc_trapno) #define ERROR_sig(context) ((context)->uc_mcontext.mc_err) #define MASK_sig(context) ((context)->uc_sigmask) +#define PAGE_FAULT_TRAP T_PAGEFLT #else #define PC_sig(context) ((context)->uc_mcontext.gregs[REG_RIP]) #define TRAP_sig(context) ((context)->uc_mcontext.gregs[REG_TRAPNO]) #define ERROR_sig(context) ((context)->uc_mcontext.gregs[REG_ERR]) #define MASK_sig(context) ((context)->uc_sigmask) +#define PAGE_FAULT_TRAP 0xe #endif int cpu_signal_handler(int host_signum, void *pinfo, @@ -XXX,XX +XXX,XX @@ int cpu_signal_handler(int host_signum, void *pinfo, pc = PC_sig(uc); return handle_cpu_signal(pc, info, - TRAP_sig(uc) == 0xe ? (ERROR_sig(uc) >> 1) & 1 : 0, + TRAP_sig(uc) == PAGE_FAULT_TRAP ? + (ERROR_sig(uc) >> 1) & 1 : 0, &MASK_sig(uc)); } -- 2.25.1 This will eventually simplify front-end usage, and will allow backends to unset TCG_TARGET_HAS_MEMORY_BSWAP without loss of optimization. The argument is added during expansion, not currently exposed to the front end translators. The backends currently only support a flags value of either TCG_BSWAP_IZ, or (TCG_BSWAP_IZ | TCG_BSWAP_OZ), since they all require zero top bytes and leave them that way. At the existing call sites we pass in (TCG_BSWAP_IZ | TCG_BSWAP_OZ), except for the flags-ignored cases of a 32-bit swap of a 32-bit value and or a 64-bit swap of a 64-bit value, where we pass 0. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/tcg/tcg-opc.h | 10 +++++----- include/tcg/tcg.h | 12 ++++++++++++ tcg/tcg-op.c | 13 ++++++++----- tcg/tcg.c | 28 ++++++++++++++++++++++++++++ tcg/README | 22 ++++++++++++++-------- 5 files changed, 67 insertions(+), 18 deletions(-) diff --git a/include/tcg/tcg-opc.h b/include/tcg/tcg-opc.h index XXXXXXX..XXXXXXX 100644 --- a/include/tcg/tcg-opc.h +++ b/include/tcg/tcg-opc.h @@ -XXX,XX +XXX,XX @@ DEF(ext8s_i32, 1, 1, 0, IMPL(TCG_TARGET_HAS_ext8s_i32)) DEF(ext16s_i32, 1, 1, 0, IMPL(TCG_TARGET_HAS_ext16s_i32)) DEF(ext8u_i32, 1, 1, 0, IMPL(TCG_TARGET_HAS_ext8u_i32)) DEF(ext16u_i32, 1, 1, 0, IMPL(TCG_TARGET_HAS_ext16u_i32)) -DEF(bswap16_i32, 1, 1, 0, IMPL(TCG_TARGET_HAS_bswap16_i32)) -DEF(bswap32_i32, 1, 1, 0, IMPL(TCG_TARGET_HAS_bswap32_i32)) +DEF(bswap16_i32, 1, 1, 1, IMPL(TCG_TARGET_HAS_bswap16_i32)) +DEF(bswap32_i32, 1, 1, 1, IMPL(TCG_TARGET_HAS_bswap32_i32)) DEF(not_i32, 1, 1, 0, IMPL(TCG_TARGET_HAS_not_i32)) DEF(neg_i32, 1, 1, 0, IMPL(TCG_TARGET_HAS_neg_i32)) DEF(andc_i32, 1, 2, 0, IMPL(TCG_TARGET_HAS_andc_i32)) @@ -XXX,XX +XXX,XX @@ DEF(ext32s_i64, 1, 1, 0, IMPL64 | IMPL(TCG_TARGET_HAS_ext32s_i64)) DEF(ext8u_i64, 1, 1, 0, IMPL64 | IMPL(TCG_TARGET_HAS_ext8u_i64)) DEF(ext16u_i64, 1, 1, 0, IMPL64 | IMPL(TCG_TARGET_HAS_ext16u_i64)) DEF(ext32u_i64, 1, 1, 0, IMPL64 | IMPL(TCG_TARGET_HAS_ext32u_i64)) -DEF(bswap16_i64, 1, 1, 0, IMPL64 | IMPL(TCG_TARGET_HAS_bswap16_i64)) -DEF(bswap32_i64, 1, 1, 0, IMPL64 | IMPL(TCG_TARGET_HAS_bswap32_i64)) -DEF(bswap64_i64, 1, 1, 0, IMPL64 | IMPL(TCG_TARGET_HAS_bswap64_i64)) +DEF(bswap16_i64, 1, 1, 1, IMPL64 | IMPL(TCG_TARGET_HAS_bswap16_i64)) +DEF(bswap32_i64, 1, 1, 1, IMPL64 | IMPL(TCG_TARGET_HAS_bswap32_i64)) +DEF(bswap64_i64, 1, 1, 1, IMPL64 | IMPL(TCG_TARGET_HAS_bswap64_i64)) DEF(not_i64, 1, 1, 0, IMPL64 | IMPL(TCG_TARGET_HAS_not_i64)) DEF(neg_i64, 1, 1, 0, IMPL64 | IMPL(TCG_TARGET_HAS_neg_i64)) DEF(andc_i64, 1, 2, 0, IMPL64 | IMPL(TCG_TARGET_HAS_andc_i64)) diff --git a/include/tcg/tcg.h b/include/tcg/tcg.h index XXXXXXX..XXXXXXX 100644 --- a/include/tcg/tcg.h +++ b/include/tcg/tcg.h @@ -XXX,XX +XXX,XX @@ typedef TCGv_ptr TCGv_env; /* Used to align parameters. See the comment before tcgv_i32_temp. */ #define TCG_CALL_DUMMY_ARG ((TCGArg)0) +/* + * Flags for the bswap opcodes. + * If IZ, the input is zero-extended, otherwise unknown. + * If OZ or OS, the output is zero- or sign-extended respectively, + * otherwise the high bits are undefined. + */ +enum { + TCG_BSWAP_IZ = 1, + TCG_BSWAP_OZ = 2, + TCG_BSWAP_OS = 4, +}; + typedef enum TCGTempVal { TEMP_VAL_DEAD, TEMP_VAL_REG, diff --git a/tcg/tcg-op.c b/tcg/tcg-op.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg-op.c +++ b/tcg/tcg-op.c @@ -XXX,XX +XXX,XX @@ void tcg_gen_ext16u_i32(TCGv_i32 ret, TCGv_i32 arg) void tcg_gen_bswap16_i32(TCGv_i32 ret, TCGv_i32 arg) { if (TCG_TARGET_HAS_bswap16_i32) { - tcg_gen_op2_i32(INDEX_op_bswap16_i32, ret, arg); + tcg_gen_op3i_i32(INDEX_op_bswap16_i32, ret, arg, + TCG_BSWAP_IZ | TCG_BSWAP_OZ); } else { TCGv_i32 t0 = tcg_temp_new_i32(); @@ -XXX,XX +XXX,XX @@ void tcg_gen_bswap16_i32(TCGv_i32 ret, TCGv_i32 arg) void tcg_gen_bswap32_i32(TCGv_i32 ret, TCGv_i32 arg) { if (TCG_TARGET_HAS_bswap32_i32) { - tcg_gen_op2_i32(INDEX_op_bswap32_i32, ret, arg); + tcg_gen_op3i_i32(INDEX_op_bswap32_i32, ret, arg, 0); } else { TCGv_i32 t0 = tcg_temp_new_i32(); TCGv_i32 t1 = tcg_temp_new_i32(); @@ -XXX,XX +XXX,XX @@ void tcg_gen_bswap16_i64(TCGv_i64 ret, TCGv_i64 arg) tcg_gen_bswap16_i32(TCGV_LOW(ret), TCGV_LOW(arg)); tcg_gen_movi_i32(TCGV_HIGH(ret), 0); } else if (TCG_TARGET_HAS_bswap16_i64) { - tcg_gen_op2_i64(INDEX_op_bswap16_i64, ret, arg); + tcg_gen_op3i_i64(INDEX_op_bswap16_i64, ret, arg, + TCG_BSWAP_IZ | TCG_BSWAP_OZ); } else { TCGv_i64 t0 = tcg_temp_new_i64(); @@ -XXX,XX +XXX,XX @@ void tcg_gen_bswap32_i64(TCGv_i64 ret, TCGv_i64 arg) tcg_gen_bswap32_i32(TCGV_LOW(ret), TCGV_LOW(arg)); tcg_gen_movi_i32(TCGV_HIGH(ret), 0); } else if (TCG_TARGET_HAS_bswap32_i64) { - tcg_gen_op2_i64(INDEX_op_bswap32_i64, ret, arg); + tcg_gen_op3i_i64(INDEX_op_bswap32_i64, ret, arg, + TCG_BSWAP_IZ | TCG_BSWAP_OZ); } else { TCGv_i64 t0 = tcg_temp_new_i64(); TCGv_i64 t1 = tcg_temp_new_i64(); @@ -XXX,XX +XXX,XX @@ void tcg_gen_bswap64_i64(TCGv_i64 ret, TCGv_i64 arg) tcg_temp_free_i32(t0); tcg_temp_free_i32(t1); } else if (TCG_TARGET_HAS_bswap64_i64) { - tcg_gen_op2_i64(INDEX_op_bswap64_i64, ret, arg); + tcg_gen_op3i_i64(INDEX_op_bswap64_i64, ret, arg, 0); } else { TCGv_i64 t0 = tcg_temp_new_i64(); TCGv_i64 t1 = tcg_temp_new_i64(); diff --git a/tcg/tcg.c b/tcg/tcg.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg.c +++ b/tcg/tcg.c @@ -XXX,XX +XXX,XX @@ static const char * const alignment_name[(MO_AMASK >> MO_ASHIFT) + 1] = { [MO_ALIGN_64 >> MO_ASHIFT] = "al64+", }; +static const char bswap_flag_name[][6] = { + [TCG_BSWAP_IZ] = "iz", + [TCG_BSWAP_OZ] = "oz", + [TCG_BSWAP_OS] = "os", + [TCG_BSWAP_IZ | TCG_BSWAP_OZ] = "iz,oz", + [TCG_BSWAP_IZ | TCG_BSWAP_OS] = "iz,os", +}; + static inline bool tcg_regset_single(TCGRegSet d) { return (d & (d - 1)) == 0; @@ -XXX,XX +XXX,XX @@ static void tcg_dump_ops(TCGContext *s, bool have_prefs) i = 1; } break; + case INDEX_op_bswap16_i32: + case INDEX_op_bswap16_i64: + case INDEX_op_bswap32_i32: + case INDEX_op_bswap32_i64: + case INDEX_op_bswap64_i64: + { + TCGArg flags = op->args[k]; + const char *name = NULL; + + if (flags < ARRAY_SIZE(bswap_flag_name)) { + name = bswap_flag_name[flags]; + } + if (name) { + col += qemu_log(",%s", name); + } else { + col += qemu_log(",$0x%" TCG_PRIlx, flags); + } + i = k = 1; + } + break; default: i = 0; break; diff --git a/tcg/README b/tcg/README index XXXXXXX..XXXXXXX 100644 --- a/tcg/README +++ b/tcg/README @@ -XXX,XX +XXX,XX @@ ext32u_i64 t0, t1 8, 16 or 32 bit sign/zero extension (both operands must have the same type) -* bswap16_i32/i64 t0, t1 +* bswap16_i32/i64 t0, t1, flags -16 bit byte swap on a 32/64 bit value. It assumes that the two/six high order -bytes are set to zero. +16 bit byte swap on the low bits of a 32/64 bit input. +If flags & TCG_BSWAP_IZ, then t1 is known to be zero-extended from bit 15. +If flags & TCG_BSWAP_OZ, then t0 will be zero-extended from bit 15. +If flags & TCG_BSWAP_OS, then t0 will be sign-extended from bit 15. +If neither TCG_BSWAP_OZ nor TCG_BSWAP_OS are set, then the bits of +t0 above bit 15 may contain any value. -* bswap32_i32/i64 t0, t1 +* bswap32_i64 t0, t1, flags -32 bit byte swap on a 32/64 bit value. With a 64 bit value, it assumes that -the four high order bytes are set to zero. +32 bit byte swap on a 64-bit value. The flags are the same as for bswap16, +except they apply from bit 31 instead of bit 15. -* bswap64_i64 t0, t1 +* bswap32_i32 t0, t1, flags +* bswap64_i64 t0, t1, flags -64 bit byte swap +32/64 bit byte swap. The flags are ignored, but still present +for consistency with the other bswap opcodes. * discard_i32/i64 t0 -- 2.25.1 Retain the current rorw bswap16 expansion for the zero-in/zero-out case. Otherwise, perform a wider bswap plus a right-shift or extend. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/i386/tcg-target.c.inc | 20 +++++++++++++++++++- 1 file changed, 19 insertions(+), 1 deletion(-) diff --git a/tcg/i386/tcg-target.c.inc b/tcg/i386/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/i386/tcg-target.c.inc +++ b/tcg/i386/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_op(TCGContext *s, TCGOpcode opc, break; OP_32_64(bswap16): - tcg_out_rolw_8(s, a0); + if (a2 & TCG_BSWAP_OS) { + /* Output must be sign-extended. */ + if (rexw) { + tcg_out_bswap64(s, a0); + tcg_out_shifti(s, SHIFT_SAR + rexw, a0, 48); + } else { + tcg_out_bswap32(s, a0); + tcg_out_shifti(s, SHIFT_SAR, a0, 16); + } + } else if ((a2 & (TCG_BSWAP_IZ | TCG_BSWAP_OZ)) == TCG_BSWAP_OZ) { + /* Output must be zero-extended, but input isn't. */ + tcg_out_bswap32(s, a0); + tcg_out_shifti(s, SHIFT_SHR, a0, 16); + } else { + tcg_out_rolw_8(s, a0); + } break; OP_32_64(bswap32): tcg_out_bswap32(s, a0); + if (rexw && (a2 & TCG_BSWAP_OS)) { + tcg_out_ext32s(s, a0, a0); + } break; OP_32_64(neg): -- 2.25.1 Pass in the input and output size. We currently use 3 of the 5 possible combinations; the others may be used by new tcg opcodes. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/aarch64/tcg-target.c.inc | 42 ++++++++++++++---------------------- 1 file changed, 16 insertions(+), 26 deletions(-) diff --git a/tcg/aarch64/tcg-target.c.inc b/tcg/aarch64/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/aarch64/tcg-target.c.inc +++ b/tcg/aarch64/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ typedef enum { /* Data-processing (1 source) instructions. */ I3507_CLZ = 0x5ac01000, I3507_RBIT = 0x5ac00000, - I3507_REV16 = 0x5ac00400, - I3507_REV32 = 0x5ac00800, - I3507_REV64 = 0x5ac00c00, + I3507_REV = 0x5ac00000, /* + size << 10 */ /* Data-processing (2 source) instructions. */ I3508_LSLV = 0x1ac02000, @@ -XXX,XX +XXX,XX @@ static void tcg_out_brcond(TCGContext *s, TCGType ext, TCGCond c, TCGArg a, } } -static inline void tcg_out_rev64(TCGContext *s, TCGReg rd, TCGReg rn) +static inline void tcg_out_rev(TCGContext *s, int ext, MemOp s_bits, + TCGReg rd, TCGReg rn) { - tcg_out_insn(s, 3507, REV64, TCG_TYPE_I64, rd, rn); -} - -static inline void tcg_out_rev32(TCGContext *s, TCGReg rd, TCGReg rn) -{ - tcg_out_insn(s, 3507, REV32, TCG_TYPE_I32, rd, rn); -} - -static inline void tcg_out_rev16(TCGContext *s, TCGReg rd, TCGReg rn) -{ - tcg_out_insn(s, 3507, REV16, TCG_TYPE_I32, rd, rn); + /* REV, REV16, REV32 */ + tcg_out_insn_3507(s, I3507_REV | (s_bits << 10), ext, rd, rn); } static inline void tcg_out_sxt(TCGContext *s, TCGType ext, MemOp s_bits, @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_ld_direct(TCGContext *s, MemOp memop, TCGType ext, case MO_UW: tcg_out_ldst_r(s, I3312_LDRH, data_r, addr_r, otype, off_r); if (bswap) { - tcg_out_rev16(s, data_r, data_r); + tcg_out_rev(s, TCG_TYPE_I32, MO_16, data_r, data_r); } break; case MO_SW: if (bswap) { tcg_out_ldst_r(s, I3312_LDRH, data_r, addr_r, otype, off_r); - tcg_out_rev16(s, data_r, data_r); + tcg_out_rev(s, TCG_TYPE_I32, MO_16, data_r, data_r); tcg_out_sxt(s, ext, MO_16, data_r, data_r); } else { tcg_out_ldst_r(s, (ext ? I3312_LDRSHX : I3312_LDRSHW), @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_ld_direct(TCGContext *s, MemOp memop, TCGType ext, case MO_UL: tcg_out_ldst_r(s, I3312_LDRW, data_r, addr_r, otype, off_r); if (bswap) { - tcg_out_rev32(s, data_r, data_r); + tcg_out_rev(s, TCG_TYPE_I32, MO_32, data_r, data_r); } break; case MO_SL: if (bswap) { tcg_out_ldst_r(s, I3312_LDRW, data_r, addr_r, otype, off_r); - tcg_out_rev32(s, data_r, data_r); + tcg_out_rev(s, TCG_TYPE_I32, MO_32, data_r, data_r); tcg_out_sxt(s, TCG_TYPE_I64, MO_32, data_r, data_r); } else { tcg_out_ldst_r(s, I3312_LDRSWX, data_r, addr_r, otype, off_r); @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_ld_direct(TCGContext *s, MemOp memop, TCGType ext, case MO_Q: tcg_out_ldst_r(s, I3312_LDRX, data_r, addr_r, otype, off_r); if (bswap) { - tcg_out_rev64(s, data_r, data_r); + tcg_out_rev(s, TCG_TYPE_I64, MO_64, data_r, data_r); } break; default: @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_st_direct(TCGContext *s, MemOp memop, break; case MO_16: if (bswap && data_r != TCG_REG_XZR) { - tcg_out_rev16(s, TCG_REG_TMP, data_r); + tcg_out_rev(s, TCG_TYPE_I32, MO_16, TCG_REG_TMP, data_r); data_r = TCG_REG_TMP; } tcg_out_ldst_r(s, I3312_STRH, data_r, addr_r, otype, off_r); break; case MO_32: if (bswap && data_r != TCG_REG_XZR) { - tcg_out_rev32(s, TCG_REG_TMP, data_r); + tcg_out_rev(s, TCG_TYPE_I32, MO_32, TCG_REG_TMP, data_r); data_r = TCG_REG_TMP; } tcg_out_ldst_r(s, I3312_STRW, data_r, addr_r, otype, off_r); break; case MO_64: if (bswap && data_r != TCG_REG_XZR) { - tcg_out_rev64(s, TCG_REG_TMP, data_r); + tcg_out_rev(s, TCG_TYPE_I64, MO_64, TCG_REG_TMP, data_r); data_r = TCG_REG_TMP; } tcg_out_ldst_r(s, I3312_STRX, data_r, addr_r, otype, off_r); @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, break; case INDEX_op_bswap64_i64: - tcg_out_rev64(s, a0, a1); + tcg_out_rev(s, TCG_TYPE_I64, MO_64, a0, a1); break; case INDEX_op_bswap32_i64: case INDEX_op_bswap32_i32: - tcg_out_rev32(s, a0, a1); + tcg_out_rev(s, TCG_TYPE_I32, MO_32, a0, a1); break; case INDEX_op_bswap16_i64: case INDEX_op_bswap16_i32: - tcg_out_rev16(s, a0, a1); + tcg_out_rev(s, TCG_TYPE_I32, MO_16, a0, a1); break; case INDEX_op_ext8s_i64: -- 2.25.1 Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/aarch64/tcg-target.c.inc | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/tcg/aarch64/tcg-target.c.inc b/tcg/aarch64/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/aarch64/tcg-target.c.inc +++ b/tcg/aarch64/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, tcg_out_rev(s, TCG_TYPE_I64, MO_64, a0, a1); break; case INDEX_op_bswap32_i64: + tcg_out_rev(s, TCG_TYPE_I32, MO_32, a0, a1); + if (a2 & TCG_BSWAP_OS) { + tcg_out_sxt(s, TCG_TYPE_I64, MO_32, a0, a0); + } + break; case INDEX_op_bswap32_i32: tcg_out_rev(s, TCG_TYPE_I32, MO_32, a0, a1); break; case INDEX_op_bswap16_i64: case INDEX_op_bswap16_i32: tcg_out_rev(s, TCG_TYPE_I32, MO_16, a0, a1); + if (a2 & TCG_BSWAP_OS) { + /* Output must be sign-extended. */ + tcg_out_sxt(s, ext, MO_16, a0, a0); + } else if ((a2 & (TCG_BSWAP_IZ | TCG_BSWAP_OZ)) == TCG_BSWAP_OZ) { + /* Output must be zero-extended, but input isn't. */ + tcg_out_uxt(s, MO_16, a0, a0); + } break; case INDEX_op_ext8s_i64: -- 2.25.1 Combine the three bswap16 routines, and differentiate via the flags. Use the correct flags combination from the load/store routines, and pass along the constant parameter from tcg_out_op. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/arm/tcg-target.c.inc | 101 ++++++++++++++++++++++++--------------- 1 file changed, 63 insertions(+), 38 deletions(-) diff --git a/tcg/arm/tcg-target.c.inc b/tcg/arm/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/arm/tcg-target.c.inc +++ b/tcg/arm/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_ext16u(TCGContext *s, int cond, } } -static inline void tcg_out_bswap16s(TCGContext *s, int cond, int rd, int rn) +static void tcg_out_bswap16(TCGContext *s, int cond, int rd, int rn, int flags) { if (use_armv6_instructions) { - /* revsh */ - tcg_out32(s, 0x06ff0fb0 | (cond << 28) | (rd << 12) | rn); - } else { - tcg_out_dat_reg(s, cond, ARITH_MOV, - TCG_REG_TMP, 0, rn, SHIFT_IMM_LSL(24)); - tcg_out_dat_reg(s, cond, ARITH_MOV, - TCG_REG_TMP, 0, TCG_REG_TMP, SHIFT_IMM_ASR(16)); - tcg_out_dat_reg(s, cond, ARITH_ORR, - rd, TCG_REG_TMP, rn, SHIFT_IMM_LSR(8)); - } -} + if (flags & TCG_BSWAP_OS) { + /* revsh */ + tcg_out32(s, 0x06ff0fb0 | (cond << 28) | (rd << 12) | rn); + return; + } -static inline void tcg_out_bswap16(TCGContext *s, int cond, int rd, int rn) -{ - if (use_armv6_instructions) { /* rev16 */ tcg_out32(s, 0x06bf0fb0 | (cond << 28) | (rd << 12) | rn); - } else { - tcg_out_dat_reg(s, cond, ARITH_MOV, - TCG_REG_TMP, 0, rn, SHIFT_IMM_LSL(24)); - tcg_out_dat_reg(s, cond, ARITH_MOV, - TCG_REG_TMP, 0, TCG_REG_TMP, SHIFT_IMM_LSR(16)); - tcg_out_dat_reg(s, cond, ARITH_ORR, - rd, TCG_REG_TMP, rn, SHIFT_IMM_LSR(8)); + if ((flags & (TCG_BSWAP_IZ | TCG_BSWAP_OZ)) == TCG_BSWAP_OZ) { + /* uxth */ + tcg_out32(s, 0x06ff0070 | (cond << 28) | (rd << 12) | rd); + } + return; } -} -/* swap the two low bytes assuming that the two high input bytes and the - two high output bit can hold any value. */ -static inline void tcg_out_bswap16st(TCGContext *s, int cond, int rd, int rn) -{ - if (use_armv6_instructions) { - /* rev16 */ - tcg_out32(s, 0x06bf0fb0 | (cond << 28) | (rd << 12) | rn); - } else { + if (flags == 0) { + /* + * For stores, no input or output extension: + * rn = xxAB + * lsr tmp, rn, #8 tmp = 0xxA + * and tmp, tmp, #0xff tmp = 000A + * orr rd, tmp, rn, lsl #8 rd = xABA + */ tcg_out_dat_reg(s, cond, ARITH_MOV, TCG_REG_TMP, 0, rn, SHIFT_IMM_LSR(8)); tcg_out_dat_imm(s, cond, ARITH_AND, TCG_REG_TMP, TCG_REG_TMP, 0xff); tcg_out_dat_reg(s, cond, ARITH_ORR, rd, TCG_REG_TMP, rn, SHIFT_IMM_LSL(8)); + return; } + + /* + * Byte swap, leaving the result at the top of the register. + * We will then shift down, zero or sign-extending. + */ + if (flags & TCG_BSWAP_IZ) { + /* + * rn = 00AB + * ror tmp, rn, #8 tmp = B00A + * orr tmp, tmp, tmp, lsl #16 tmp = BA00 + */ + tcg_out_dat_reg(s, cond, ARITH_MOV, + TCG_REG_TMP, 0, rn, SHIFT_IMM_ROR(8)); + tcg_out_dat_reg(s, cond, ARITH_ORR, + TCG_REG_TMP, TCG_REG_TMP, TCG_REG_TMP, + SHIFT_IMM_LSL(16)); + } else { + /* + * rn = xxAB + * and tmp, rn, #0xff00 tmp = 00A0 + * lsl tmp, tmp, #8 tmp = 0A00 + * orr tmp, tmp, rn, lsl #24 tmp = BA00 + */ + tcg_out_dat_rI(s, cond, ARITH_AND, TCG_REG_TMP, rn, 0xff00, 1); + tcg_out_dat_reg(s, cond, ARITH_MOV, + TCG_REG_TMP, 0, TCG_REG_TMP, SHIFT_IMM_LSL(8)); + tcg_out_dat_reg(s, cond, ARITH_ORR, + TCG_REG_TMP, TCG_REG_TMP, rn, SHIFT_IMM_LSL(24)); + } + tcg_out_dat_reg(s, cond, ARITH_MOV, rd, 0, TCG_REG_TMP, + (flags & TCG_BSWAP_OS + ? SHIFT_IMM_ASR(8) : SHIFT_IMM_LSR(8))); } static inline void tcg_out_bswap32(TCGContext *s, int cond, int rd, int rn) @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_qemu_ld_index(TCGContext *s, MemOp opc, case MO_UW: tcg_out_ld16u_r(s, COND_AL, datalo, addrlo, addend); if (bswap) { - tcg_out_bswap16(s, COND_AL, datalo, datalo); + tcg_out_bswap16(s, COND_AL, datalo, datalo, + TCG_BSWAP_IZ | TCG_BSWAP_OZ); } break; case MO_SW: if (bswap) { tcg_out_ld16u_r(s, COND_AL, datalo, addrlo, addend); - tcg_out_bswap16s(s, COND_AL, datalo, datalo); + tcg_out_bswap16(s, COND_AL, datalo, datalo, + TCG_BSWAP_IZ | TCG_BSWAP_OS); } else { tcg_out_ld16s_r(s, COND_AL, datalo, addrlo, addend); } @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_qemu_ld_direct(TCGContext *s, MemOp opc, case MO_UW: tcg_out_ld16u_8(s, COND_AL, datalo, addrlo, 0); if (bswap) { - tcg_out_bswap16(s, COND_AL, datalo, datalo); + tcg_out_bswap16(s, COND_AL, datalo, datalo, + TCG_BSWAP_IZ | TCG_BSWAP_OZ); } break; case MO_SW: if (bswap) { tcg_out_ld16u_8(s, COND_AL, datalo, addrlo, 0); - tcg_out_bswap16s(s, COND_AL, datalo, datalo); + tcg_out_bswap16(s, COND_AL, datalo, datalo, + TCG_BSWAP_IZ | TCG_BSWAP_OS); } else { tcg_out_ld16s_8(s, COND_AL, datalo, addrlo, 0); } @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_qemu_st_index(TCGContext *s, int cond, MemOp opc, break; case MO_16: if (bswap) { - tcg_out_bswap16st(s, cond, TCG_REG_R0, datalo); + tcg_out_bswap16(s, cond, TCG_REG_R0, datalo, 0); tcg_out_st16_r(s, cond, TCG_REG_R0, addrlo, addend); } else { tcg_out_st16_r(s, cond, datalo, addrlo, addend); @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_qemu_st_direct(TCGContext *s, MemOp opc, break; case MO_16: if (bswap) { - tcg_out_bswap16st(s, COND_AL, TCG_REG_R0, datalo); + tcg_out_bswap16(s, COND_AL, TCG_REG_R0, datalo, 0); tcg_out_st16_8(s, COND_AL, TCG_REG_R0, addrlo, 0); } else { tcg_out_st16_8(s, COND_AL, datalo, addrlo, 0); @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_op(TCGContext *s, TCGOpcode opc, break; case INDEX_op_bswap16_i32: - tcg_out_bswap16(s, COND_AL, args[0], args[1]); + tcg_out_bswap16(s, COND_AL, args[0], args[1], args[2]); break; case INDEX_op_bswap32_i32: tcg_out_bswap32(s, COND_AL, args[0], args[1]); -- 2.25.1 We will shortly require these in other context; make the expansion as clear as possible. Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/ppc/tcg-target.c.inc | 31 +++++++++++++++++++++---------- 1 file changed, 21 insertions(+), 10 deletions(-) diff --git a/tcg/ppc/tcg-target.c.inc b/tcg/ppc/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/ppc/tcg-target.c.inc +++ b/tcg/ppc/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_rlw(TCGContext *s, int op, TCGReg ra, TCGReg rs, tcg_out32(s, op | RA(ra) | RS(rs) | SH(sh) | MB(mb) | ME(me)); } +static inline void tcg_out_ext8s(TCGContext *s, TCGReg dst, TCGReg src) +{ + tcg_out32(s, EXTSB | RA(dst) | RS(src)); +} + +static inline void tcg_out_ext16s(TCGContext *s, TCGReg dst, TCGReg src) +{ + tcg_out32(s, EXTSH | RA(dst) | RS(src)); +} + +static inline void tcg_out_ext32s(TCGContext *s, TCGReg dst, TCGReg src) +{ + tcg_out32(s, EXTSW | RA(dst) | RS(src)); +} + static inline void tcg_out_ext32u(TCGContext *s, TCGReg dst, TCGReg src) { tcg_out_rld(s, RLDICL, dst, src, 0, 32); @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, const int const_args[TCG_MAX_OP_ARGS]) { TCGArg a0, a1, a2; - int c; switch (opc) { case INDEX_op_exit_tb: @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, case INDEX_op_ld8s_i32: case INDEX_op_ld8s_i64: tcg_out_mem_long(s, LBZ, LBZX, args[0], args[1], args[2]); - tcg_out32(s, EXTSB | RS(args[0]) | RA(args[0])); + tcg_out_ext8s(s, args[0], args[0]); break; case INDEX_op_ld16u_i32: case INDEX_op_ld16u_i64: @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, case INDEX_op_ext8s_i32: case INDEX_op_ext8s_i64: - c = EXTSB; - goto gen_ext; + tcg_out_ext8s(s, args[0], args[1]); + break; case INDEX_op_ext16s_i32: case INDEX_op_ext16s_i64: - c = EXTSH; - goto gen_ext; + tcg_out_ext16s(s, args[0], args[1]); + break; case INDEX_op_ext_i32_i64: case INDEX_op_ext32s_i64: - c = EXTSW; - goto gen_ext; - gen_ext: - tcg_out32(s, c | RS(args[1]) | RA(args[0])); + tcg_out_ext32s(s, args[0], args[1]); break; case INDEX_op_extu_i32_i64: tcg_out_ext32u(s, args[0], args[1]); -- 2.25.1 We will shortly require sari in other context; split out both for cleanliness sake. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/ppc/tcg-target.c.inc | 17 +++++++++++++---- 1 file changed, 13 insertions(+), 4 deletions(-) diff --git a/tcg/ppc/tcg-target.c.inc b/tcg/ppc/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/ppc/tcg-target.c.inc +++ b/tcg/ppc/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_shli64(TCGContext *s, TCGReg dst, TCGReg src, int c) tcg_out_rld(s, RLDICR, dst, src, c, 63 - c); } +static inline void tcg_out_sari32(TCGContext *s, TCGReg dst, TCGReg src, int c) +{ + /* Limit immediate shift count lest we create an illegal insn. */ + tcg_out32(s, SRAWI | RA(dst) | RS(src) | SH(c & 31)); +} + static inline void tcg_out_shri32(TCGContext *s, TCGReg dst, TCGReg src, int c) { tcg_out_rlw(s, RLWINM, dst, src, 32 - c, c, 31); @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_shri64(TCGContext *s, TCGReg dst, TCGReg src, int c) tcg_out_rld(s, RLDICL, dst, src, 64 - c, c); } +static inline void tcg_out_sari64(TCGContext *s, TCGReg dst, TCGReg src, int c) +{ + tcg_out32(s, SRADI | RA(dst) | RS(src) | SH(c & 0x1f) | ((c >> 4) & 2)); +} + /* Emit a move into ret of arg, if it can be done in one insn. */ static bool tcg_out_movi_one(TCGContext *s, TCGReg ret, tcg_target_long arg) { @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, break; case INDEX_op_sar_i32: if (const_args[2]) { - /* Limit immediate shift count lest we create an illegal insn. */ - tcg_out32(s, SRAWI | RS(args[1]) | RA(args[0]) | SH(args[2] & 31)); + tcg_out_sari32(s, args[0], args[1], args[2]); } else { tcg_out32(s, SRAW | SAB(args[1], args[0], args[2])); } @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, break; case INDEX_op_sar_i64: if (const_args[2]) { - int sh = SH(args[2] & 0x1f) | (((args[2] >> 5) & 1) << 1); - tcg_out32(s, SRADI | RA(args[0]) | RS(args[1]) | sh); + tcg_out_sari64(s, args[0], args[1], args[2]); } else { tcg_out32(s, SRAD | SAB(args[1], args[0], args[2])); } -- 2.25.1 With the use of a suitable temporary, we can use the same algorithm when src overlaps dst. The result is the same number of instructions either way. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/ppc/tcg-target.c.inc | 34 +++++++++++++++++++--------------- 1 file changed, 19 insertions(+), 15 deletions(-) diff --git a/tcg/ppc/tcg-target.c.inc b/tcg/ppc/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/ppc/tcg-target.c.inc +++ b/tcg/ppc/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_sari64(TCGContext *s, TCGReg dst, TCGReg src, int c) tcg_out32(s, SRADI | RA(dst) | RS(src) | SH(c & 0x1f) | ((c >> 4) & 2)); } +static void tcg_out_bswap16(TCGContext *s, TCGReg dst, TCGReg src) +{ + TCGReg tmp = dst == src ? TCG_REG_R0 : dst; + + /* + * In the following, + * dep(a, b, m) -> (a & ~m) | (b & m) + * + * Begin with: src = xxxxabcd + */ + /* tmp = rol32(src, 24) & 0x000000ff = 0000000c */ + tcg_out_rlw(s, RLWINM, tmp, src, 24, 24, 31); + /* tmp = dep(tmp, rol32(src, 8), 0x0000ff00) = 000000dc */ + tcg_out_rlw(s, RLWIMI, tmp, src, 8, 16, 23); + + tcg_out_mov(s, TCG_TYPE_REG, dst, tmp); +} + /* Emit a move into ret of arg, if it can be done in one insn. */ static bool tcg_out_movi_one(TCGContext *s, TCGReg ret, tcg_target_long arg) { @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, case INDEX_op_bswap16_i32: case INDEX_op_bswap16_i64: - a0 = args[0], a1 = args[1]; - /* a1 = abcd */ - if (a0 != a1) { - /* a0 = (a1 r<< 24) & 0xff # 000c */ - tcg_out_rlw(s, RLWINM, a0, a1, 24, 24, 31); - /* a0 = (a0 & ~0xff00) | (a1 r<< 8) & 0xff00 # 00dc */ - tcg_out_rlw(s, RLWIMI, a0, a1, 8, 16, 23); - } else { - /* r0 = (a1 r<< 8) & 0xff00 # 00d0 */ - tcg_out_rlw(s, RLWINM, TCG_REG_R0, a1, 8, 16, 23); - /* a0 = (a1 r<< 24) & 0xff # 000c */ - tcg_out_rlw(s, RLWINM, a0, a1, 24, 24, 31); - /* a0 = a0 | r0 # 00dc */ - tcg_out32(s, OR | SAB(TCG_REG_R0, a0, a0)); - } + tcg_out_bswap16(s, args[0], args[1]); break; case INDEX_op_bswap32_i32: -- 2.25.1 Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/ppc/tcg-target.c.inc | 38 ++++++++++++++++++++++---------------- 1 file changed, 22 insertions(+), 16 deletions(-) diff --git a/tcg/ppc/tcg-target.c.inc b/tcg/ppc/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/ppc/tcg-target.c.inc +++ b/tcg/ppc/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_bswap16(TCGContext *s, TCGReg dst, TCGReg src) tcg_out_mov(s, TCG_TYPE_REG, dst, tmp); } +static void tcg_out_bswap32(TCGContext *s, TCGReg dst, TCGReg src) +{ + TCGReg tmp = dst == src ? TCG_REG_R0 : dst; + + /* + * Stolen from gcc's builtin_bswap32. + * In the following, + * dep(a, b, m) -> (a & ~m) | (b & m) + * + * Begin with: src = xxxxabcd + */ + /* tmp = rol32(src, 8) & 0xffffffff = 0000bcda */ + tcg_out_rlw(s, RLWINM, tmp, src, 8, 0, 31); + /* tmp = dep(tmp, rol32(src, 24), 0xff000000) = 0000dcda */ + tcg_out_rlw(s, RLWIMI, tmp, src, 24, 0, 7); + /* tmp = dep(tmp, rol32(src, 24), 0x0000ff00) = 0000dcba */ + tcg_out_rlw(s, RLWIMI, tmp, src, 24, 16, 23); + + tcg_out_mov(s, TCG_TYPE_REG, dst, tmp); +} + /* Emit a move into ret of arg, if it can be done in one insn. */ static bool tcg_out_movi_one(TCGContext *s, TCGReg ret, tcg_target_long arg) { @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, case INDEX_op_bswap16_i64: tcg_out_bswap16(s, args[0], args[1]); break; - case INDEX_op_bswap32_i32: case INDEX_op_bswap32_i64: - /* Stolen from gcc's builtin_bswap32 */ - a1 = args[1]; - a0 = args[0] == a1 ? TCG_REG_R0 : args[0]; - - /* a1 = args[1] # abcd */ - /* a0 = rotate_left (a1, 8) # bcda */ - tcg_out_rlw(s, RLWINM, a0, a1, 8, 0, 31); - /* a0 = (a0 & ~0xff000000) | ((a1 r<< 24) & 0xff000000) # dcda */ - tcg_out_rlw(s, RLWIMI, a0, a1, 24, 0, 7); - /* a0 = (a0 & ~0x0000ff00) | ((a1 r<< 24) & 0x0000ff00) # dcba */ - tcg_out_rlw(s, RLWIMI, a0, a1, 24, 16, 23); - - if (a0 == TCG_REG_R0) { - tcg_out_mov(s, TCG_TYPE_REG, args[0], a0); - } + tcg_out_bswap32(s, args[0], args[1]); break; case INDEX_op_bswap64_i64: -- 2.25.1 Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/ppc/tcg-target.c.inc | 64 +++++++++++++++++++++------------------- 1 file changed, 34 insertions(+), 30 deletions(-) diff --git a/tcg/ppc/tcg-target.c.inc b/tcg/ppc/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/ppc/tcg-target.c.inc +++ b/tcg/ppc/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_bswap32(TCGContext *s, TCGReg dst, TCGReg src) tcg_out_mov(s, TCG_TYPE_REG, dst, tmp); } +static void tcg_out_bswap64(TCGContext *s, TCGReg dst, TCGReg src) +{ + TCGReg t0 = dst == src ? TCG_REG_R0 : dst; + TCGReg t1 = dst == src ? dst : TCG_REG_R0; + + /* + * In the following, + * dep(a, b, m) -> (a & ~m) | (b & m) + * + * Begin with: src = abcdefgh + */ + /* t0 = rol32(src, 8) & 0xffffffff = 0000fghe */ + tcg_out_rlw(s, RLWINM, t0, src, 8, 0, 31); + /* t0 = dep(t0, rol32(src, 24), 0xff000000) = 0000hghe */ + tcg_out_rlw(s, RLWIMI, t0, src, 24, 0, 7); + /* t0 = dep(t0, rol32(src, 24), 0x0000ff00) = 0000hgfe */ + tcg_out_rlw(s, RLWIMI, t0, src, 24, 16, 23); + + /* t0 = rol64(t0, 32) = hgfe0000 */ + tcg_out_rld(s, RLDICL, t0, t0, 32, 0); + /* t1 = rol64(src, 32) = efghabcd */ + tcg_out_rld(s, RLDICL, t1, src, 32, 0); + + /* t0 = dep(t0, rol32(t1, 24), 0xffffffff) = hgfebcda */ + tcg_out_rlw(s, RLWIMI, t0, t1, 8, 0, 31); + /* t0 = dep(t0, rol32(t1, 24), 0xff000000) = hgfedcda */ + tcg_out_rlw(s, RLWIMI, t0, t1, 24, 0, 7); + /* t0 = dep(t0, rol32(t1, 24), 0x0000ff00) = hgfedcba */ + tcg_out_rlw(s, RLWIMI, t0, t1, 24, 16, 23); + + tcg_out_mov(s, TCG_TYPE_REG, dst, t0); +} + /* Emit a move into ret of arg, if it can be done in one insn. */ static bool tcg_out_movi_one(TCGContext *s, TCGReg ret, tcg_target_long arg) { @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, case INDEX_op_bswap32_i64: tcg_out_bswap32(s, args[0], args[1]); break; - case INDEX_op_bswap64_i64: - a0 = args[0], a1 = args[1], a2 = TCG_REG_R0; - if (a0 == a1) { - a0 = TCG_REG_R0; - a2 = a1; - } - - /* a1 = # abcd efgh */ - /* a0 = rl32(a1, 8) # 0000 fghe */ - tcg_out_rlw(s, RLWINM, a0, a1, 8, 0, 31); - /* a0 = dep(a0, rl32(a1, 24), 0xff000000) # 0000 hghe */ - tcg_out_rlw(s, RLWIMI, a0, a1, 24, 0, 7); - /* a0 = dep(a0, rl32(a1, 24), 0x0000ff00) # 0000 hgfe */ - tcg_out_rlw(s, RLWIMI, a0, a1, 24, 16, 23); - - /* a0 = rl64(a0, 32) # hgfe 0000 */ - /* a2 = rl64(a1, 32) # efgh abcd */ - tcg_out_rld(s, RLDICL, a0, a0, 32, 0); - tcg_out_rld(s, RLDICL, a2, a1, 32, 0); - - /* a0 = dep(a0, rl32(a2, 8), 0xffffffff) # hgfe bcda */ - tcg_out_rlw(s, RLWIMI, a0, a2, 8, 0, 31); - /* a0 = dep(a0, rl32(a2, 24), 0xff000000) # hgfe dcda */ - tcg_out_rlw(s, RLWIMI, a0, a2, 24, 0, 7); - /* a0 = dep(a0, rl32(a2, 24), 0x0000ff00) # hgfe dcba */ - tcg_out_rlw(s, RLWIMI, a0, a2, 24, 16, 23); - - if (a0 == 0) { - tcg_out_mov(s, TCG_TYPE_REG, args[0], a0); - } + tcg_out_bswap64(s, args[0], args[1]); break; case INDEX_op_deposit_i32: -- 2.25.1 For INDEX_op_bswap32_i32, pass 0 for flags: input not zero-extended, output does not need extension within the host 64-bit register. Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/ppc/tcg-target.c.inc | 22 ++++++++++++++++------ 1 file changed, 16 insertions(+), 6 deletions(-) diff --git a/tcg/ppc/tcg-target.c.inc b/tcg/ppc/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/ppc/tcg-target.c.inc +++ b/tcg/ppc/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_sari64(TCGContext *s, TCGReg dst, TCGReg src, int c) tcg_out32(s, SRADI | RA(dst) | RS(src) | SH(c & 0x1f) | ((c >> 4) & 2)); } -static void tcg_out_bswap16(TCGContext *s, TCGReg dst, TCGReg src) +static void tcg_out_bswap16(TCGContext *s, TCGReg dst, TCGReg src, int flags) { TCGReg tmp = dst == src ? TCG_REG_R0 : dst; @@ -XXX,XX +XXX,XX @@ static void tcg_out_bswap16(TCGContext *s, TCGReg dst, TCGReg src) /* tmp = dep(tmp, rol32(src, 8), 0x0000ff00) = 000000dc */ tcg_out_rlw(s, RLWIMI, tmp, src, 8, 16, 23); - tcg_out_mov(s, TCG_TYPE_REG, dst, tmp); + if (flags & TCG_BSWAP_OS) { + tcg_out_ext16s(s, dst, tmp); + } else { + tcg_out_mov(s, TCG_TYPE_REG, dst, tmp); + } } -static void tcg_out_bswap32(TCGContext *s, TCGReg dst, TCGReg src) +static void tcg_out_bswap32(TCGContext *s, TCGReg dst, TCGReg src, int flags) { TCGReg tmp = dst == src ? TCG_REG_R0 : dst; @@ -XXX,XX +XXX,XX @@ static void tcg_out_bswap32(TCGContext *s, TCGReg dst, TCGReg src) /* tmp = dep(tmp, rol32(src, 24), 0x0000ff00) = 0000dcba */ tcg_out_rlw(s, RLWIMI, tmp, src, 24, 16, 23); - tcg_out_mov(s, TCG_TYPE_REG, dst, tmp); + if (flags & TCG_BSWAP_OS) { + tcg_out_ext32s(s, dst, tmp); + } else { + tcg_out_mov(s, TCG_TYPE_REG, dst, tmp); + } } static void tcg_out_bswap64(TCGContext *s, TCGReg dst, TCGReg src) @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, case INDEX_op_bswap16_i32: case INDEX_op_bswap16_i64: - tcg_out_bswap16(s, args[0], args[1]); + tcg_out_bswap16(s, args[0], args[1], args[2]); break; case INDEX_op_bswap32_i32: + tcg_out_bswap32(s, args[0], args[1], 0); + break; case INDEX_op_bswap32_i64: - tcg_out_bswap32(s, args[0], args[1]); + tcg_out_bswap32(s, args[0], args[1], args[2]); break; case INDEX_op_bswap64_i64: tcg_out_bswap64(s, args[0], args[1]); -- 2.25.1 Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/ppc/tcg-target.c.inc | 34 ++++++++++++++++++++++++++++++++++ 1 file changed, 34 insertions(+) diff --git a/tcg/ppc/tcg-target.c.inc b/tcg/ppc/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/ppc/tcg-target.c.inc +++ b/tcg/ppc/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static bool tcg_target_const_match(int64_t val, TCGType type, int ct) #define SRAD XO31(794) #define SRADI XO31(413<<1) +#define BRH XO31(219) +#define BRW XO31(155) +#define BRD XO31(187) + #define TW XO31( 4) #define TRAP (TW | TO(31)) @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_ext16s(TCGContext *s, TCGReg dst, TCGReg src) tcg_out32(s, EXTSH | RA(dst) | RS(src)); } +static inline void tcg_out_ext16u(TCGContext *s, TCGReg dst, TCGReg src) +{ + tcg_out32(s, ANDI | SAI(src, dst, 0xffff)); +} + static inline void tcg_out_ext32s(TCGContext *s, TCGReg dst, TCGReg src) { tcg_out32(s, EXTSW | RA(dst) | RS(src)); @@ -XXX,XX +XXX,XX @@ static void tcg_out_bswap16(TCGContext *s, TCGReg dst, TCGReg src, int flags) { TCGReg tmp = dst == src ? TCG_REG_R0 : dst; + if (have_isa_3_10) { + tcg_out32(s, BRH | RA(dst) | RS(src)); + if (flags & TCG_BSWAP_OS) { + tcg_out_ext16s(s, dst, dst); + } else if ((flags & (TCG_BSWAP_IZ | TCG_BSWAP_OZ)) == TCG_BSWAP_OZ) { + tcg_out_ext16u(s, dst, dst); + } + return; + } + /* * In the following, * dep(a, b, m) -> (a & ~m) | (b & m) @@ -XXX,XX +XXX,XX @@ static void tcg_out_bswap32(TCGContext *s, TCGReg dst, TCGReg src, int flags) { TCGReg tmp = dst == src ? TCG_REG_R0 : dst; + if (have_isa_3_10) { + tcg_out32(s, BRW | RA(dst) | RS(src)); + if (flags & TCG_BSWAP_OS) { + tcg_out_ext32s(s, dst, dst); + } else if ((flags & (TCG_BSWAP_IZ | TCG_BSWAP_OZ)) == TCG_BSWAP_OZ) { + tcg_out_ext32u(s, dst, dst); + } + return; + } + /* * Stolen from gcc's builtin_bswap32. * In the following, @@ -XXX,XX +XXX,XX @@ static void tcg_out_bswap64(TCGContext *s, TCGReg dst, TCGReg src) TCGReg t0 = dst == src ? TCG_REG_R0 : dst; TCGReg t1 = dst == src ? dst : TCG_REG_R0; + if (have_isa_3_10) { + tcg_out32(s, BRD | RA(dst) | RS(src)); + return; + } + /* * In the following, * dep(a, b, m) -> (a & ~m) | (b & m) -- 2.25.1 For INDEX_op_bswap16_i64, use 64-bit instructions so that we can easily provide the extension to 64-bits. Drop the special case, previously used, where the input is already zero-extended -- the minor code size savings is not worth the complication. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/s390/tcg-target.c.inc | 34 ++++++++++++++++++++++++++++------ 1 file changed, 28 insertions(+), 6 deletions(-) diff --git a/tcg/s390/tcg-target.c.inc b/tcg/s390/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/s390/tcg-target.c.inc +++ b/tcg/s390/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_op(TCGContext *s, TCGOpcode opc, tgen_ext16u(s, TCG_TYPE_I32, args[0], args[1]); break; - OP_32_64(bswap16): - /* The TCG bswap definition requires bits 0-47 already be zero. - Thus we don't need the G-type insns to implement bswap16_i64. */ - tcg_out_insn(s, RRE, LRVR, args[0], args[1]); - tcg_out_sh32(s, RS_SRL, args[0], TCG_REG_NONE, 16); + case INDEX_op_bswap16_i32: + a0 = args[0], a1 = args[1], a2 = args[2]; + tcg_out_insn(s, RRE, LRVR, a0, a1); + if (a2 & TCG_BSWAP_OS) { + tcg_out_sh32(s, RS_SRA, a0, TCG_REG_NONE, 16); + } else { + tcg_out_sh32(s, RS_SRL, a0, TCG_REG_NONE, 16); + } break; - OP_32_64(bswap32): + case INDEX_op_bswap16_i64: + a0 = args[0], a1 = args[1], a2 = args[2]; + tcg_out_insn(s, RRE, LRVGR, a0, a1); + if (a2 & TCG_BSWAP_OS) { + tcg_out_sh64(s, RSY_SRAG, a0, a0, TCG_REG_NONE, 48); + } else { + tcg_out_sh64(s, RSY_SRLG, a0, a0, TCG_REG_NONE, 48); + } + break; + + case INDEX_op_bswap32_i32: tcg_out_insn(s, RRE, LRVR, args[0], args[1]); break; + case INDEX_op_bswap32_i64: + a0 = args[0], a1 = args[1], a2 = args[2]; + tcg_out_insn(s, RRE, LRVR, a0, a1); + if (a2 & TCG_BSWAP_OS) { + tgen_ext32s(s, a0, a0); + } else if ((a2 & (TCG_BSWAP_IZ | TCG_BSWAP_OZ)) == TCG_BSWAP_OZ) { + tgen_ext32u(s, a0, a0); + } + break; case INDEX_op_add2_i32: if (const_args[4]) { -- 2.25.1 Merge tcg_out_bswap16 and tcg_out_bswap16s. Use the flags in the internal uses for loads and stores. Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/mips/tcg-target.c.inc | 63 +++++++++++++++++++-------------------- 1 file changed, 30 insertions(+), 33 deletions(-) diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/mips/tcg-target.c.inc +++ b/tcg/mips/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_movi(TCGContext *s, TCGType type, } } -static inline void tcg_out_bswap16(TCGContext *s, TCGReg ret, TCGReg arg) +static void tcg_out_bswap16(TCGContext *s, TCGReg ret, TCGReg arg, int flags) { + /* ret and arg can't be register tmp0 */ + tcg_debug_assert(ret != TCG_TMP0); + tcg_debug_assert(arg != TCG_TMP0); + + /* With arg = abcd: */ if (use_mips32r2_instructions) { - tcg_out_opc_reg(s, OPC_WSBH, ret, 0, arg); - } else { - /* ret and arg can't be register at */ - if (ret == TCG_TMP0 || arg == TCG_TMP0) { - tcg_abort(); + tcg_out_opc_reg(s, OPC_WSBH, ret, 0, arg); /* badc */ + if (flags & TCG_BSWAP_OS) { + tcg_out_opc_reg(s, OPC_SEH, ret, 0, ret); /* ssdc */ + } else if ((flags & (TCG_BSWAP_IZ | TCG_BSWAP_OZ)) == TCG_BSWAP_OZ) { + tcg_out_opc_imm(s, OPC_ANDI, ret, ret, 0xffff); /* 00dc */ } - - tcg_out_opc_sa(s, OPC_SRL, TCG_TMP0, arg, 8); - tcg_out_opc_sa(s, OPC_SLL, ret, arg, 8); - tcg_out_opc_imm(s, OPC_ANDI, ret, ret, 0xff00); - tcg_out_opc_reg(s, OPC_OR, ret, ret, TCG_TMP0); + return; } -} -static inline void tcg_out_bswap16s(TCGContext *s, TCGReg ret, TCGReg arg) -{ - if (use_mips32r2_instructions) { - tcg_out_opc_reg(s, OPC_WSBH, ret, 0, arg); - tcg_out_opc_reg(s, OPC_SEH, ret, 0, ret); + tcg_out_opc_sa(s, OPC_SRL, TCG_TMP0, arg, 8); /* 0abc */ + if (!(flags & TCG_BSWAP_IZ)) { + tcg_out_opc_imm(s, OPC_ANDI, TCG_TMP0, TCG_TMP0, 0x00ff); /* 000c */ + } + if (flags & TCG_BSWAP_OS) { + tcg_out_opc_sa(s, OPC_SLL, ret, arg, 24); /* d000 */ + tcg_out_opc_sa(s, OPC_SRA, ret, ret, 16); /* ssd0 */ } else { - /* ret and arg can't be register at */ - if (ret == TCG_TMP0 || arg == TCG_TMP0) { - tcg_abort(); + tcg_out_opc_sa(s, OPC_SLL, ret, arg, 8); /* bcd0 */ + if (flags & TCG_BSWAP_OZ) { + tcg_out_opc_imm(s, OPC_ANDI, ret, ret, 0xff00); /* 00d0 */ } - - tcg_out_opc_sa(s, OPC_SRL, TCG_TMP0, arg, 8); - tcg_out_opc_sa(s, OPC_SLL, ret, arg, 24); - tcg_out_opc_sa(s, OPC_SRA, ret, ret, 16); - tcg_out_opc_reg(s, OPC_OR, ret, ret, TCG_TMP0); } + tcg_out_opc_reg(s, OPC_OR, ret, ret, TCG_TMP0); /* ssdc */ } static void tcg_out_bswap_subr(TCGContext *s, const tcg_insn_unit *sub) @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_ld_direct(TCGContext *s, TCGReg lo, TCGReg hi, break; case MO_UW | MO_BSWAP: tcg_out_opc_imm(s, OPC_LHU, TCG_TMP1, base, 0); - tcg_out_bswap16(s, lo, TCG_TMP1); + tcg_out_bswap16(s, lo, TCG_TMP1, TCG_BSWAP_IZ | TCG_BSWAP_OZ); break; case MO_UW: tcg_out_opc_imm(s, OPC_LHU, lo, base, 0); break; case MO_SW | MO_BSWAP: tcg_out_opc_imm(s, OPC_LHU, TCG_TMP1, base, 0); - tcg_out_bswap16s(s, lo, TCG_TMP1); + tcg_out_bswap16(s, lo, TCG_TMP1, TCG_BSWAP_IZ | TCG_BSWAP_OS); break; case MO_SW: tcg_out_opc_imm(s, OPC_LH, lo, base, 0); @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_st_direct(TCGContext *s, TCGReg lo, TCGReg hi, break; case MO_16 | MO_BSWAP: - tcg_out_opc_imm(s, OPC_ANDI, TCG_TMP1, lo, 0xffff); - tcg_out_bswap16(s, TCG_TMP1, TCG_TMP1); + tcg_out_bswap16(s, TCG_TMP1, lo, 0); lo = TCG_TMP1; /* FALLTHRU */ case MO_16: @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_op(TCGContext *s, TCGOpcode opc, case INDEX_op_not_i64: i1 = OPC_NOR; goto do_unary; - case INDEX_op_bswap16_i32: - case INDEX_op_bswap16_i64: - i1 = OPC_WSBH; - goto do_unary; case INDEX_op_ext8s_i32: case INDEX_op_ext8s_i64: i1 = OPC_SEB; @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_op(TCGContext *s, TCGOpcode opc, tcg_out_opc_reg(s, i1, a0, TCG_REG_ZERO, a1); break; + case INDEX_op_bswap16_i32: + case INDEX_op_bswap16_i64: + tcg_out_bswap16(s, a0, a1, a2); + break; case INDEX_op_bswap32_i32: tcg_out_bswap32(s, a0, a1); break; -- 2.25.1 Merge tcg_out_bswap32 and tcg_out_bswap32s. Use the flags in the internal uses for loads and stores. For mips32r2 bswap32 with zero-extension, standardize on WSBH+ROTR+DEXT. This is the same number of insns as the previous DSBH+DSHD+DSRL but fits in better with the flags check. Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/mips/tcg-target.c.inc | 39 ++++++++++++++++----------------------- 1 file changed, 16 insertions(+), 23 deletions(-) diff --git a/tcg/mips/tcg-target.c.inc b/tcg/mips/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/mips/tcg-target.c.inc +++ b/tcg/mips/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_bswap_subr(TCGContext *s, const tcg_insn_unit *sub) tcg_debug_assert(ok); } -static void tcg_out_bswap32(TCGContext *s, TCGReg ret, TCGReg arg) +static void tcg_out_bswap32(TCGContext *s, TCGReg ret, TCGReg arg, int flags) { if (use_mips32r2_instructions) { tcg_out_opc_reg(s, OPC_WSBH, ret, 0, arg); tcg_out_opc_sa(s, OPC_ROTR, ret, ret, 16); + if (flags & TCG_BSWAP_OZ) { + tcg_out_opc_bf(s, OPC_DEXT, ret, ret, 31, 0); + } } else { - tcg_out_bswap_subr(s, bswap32_addr); - /* delay slot -- never omit the insn, like tcg_out_mov might. */ - tcg_out_opc_reg(s, OPC_OR, TCG_TMP0, arg, TCG_REG_ZERO); - tcg_out_mov(s, TCG_TYPE_I32, ret, TCG_TMP3); - } -} - -static void tcg_out_bswap32u(TCGContext *s, TCGReg ret, TCGReg arg) -{ - if (use_mips32r2_instructions) { - tcg_out_opc_reg(s, OPC_DSBH, ret, 0, arg); - tcg_out_opc_reg(s, OPC_DSHD, ret, 0, ret); - tcg_out_dsrl(s, ret, ret, 32); - } else { - tcg_out_bswap_subr(s, bswap32u_addr); + if (flags & TCG_BSWAP_OZ) { + tcg_out_bswap_subr(s, bswap32u_addr); + } else { + tcg_out_bswap_subr(s, bswap32_addr); + } /* delay slot -- never omit the insn, like tcg_out_mov might. */ tcg_out_opc_reg(s, OPC_OR, TCG_TMP0, arg, TCG_REG_ZERO); tcg_out_mov(s, TCG_TYPE_I32, ret, TCG_TMP3); @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_ld_direct(TCGContext *s, TCGReg lo, TCGReg hi, if (TCG_TARGET_REG_BITS == 64 && is_64) { if (use_mips32r2_instructions) { tcg_out_opc_imm(s, OPC_LWU, lo, base, 0); - tcg_out_bswap32u(s, lo, lo); + tcg_out_bswap32(s, lo, lo, TCG_BSWAP_IZ | TCG_BSWAP_OZ); } else { tcg_out_bswap_subr(s, bswap32u_addr); /* delay slot */ @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_ld_direct(TCGContext *s, TCGReg lo, TCGReg hi, case MO_SL | MO_BSWAP: if (use_mips32r2_instructions) { tcg_out_opc_imm(s, OPC_LW, lo, base, 0); - tcg_out_bswap32(s, lo, lo); + tcg_out_bswap32(s, lo, lo, 0); } else { tcg_out_bswap_subr(s, bswap32_addr); /* delay slot */ @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_st_direct(TCGContext *s, TCGReg lo, TCGReg hi, break; case MO_32 | MO_BSWAP: - tcg_out_bswap32(s, TCG_TMP3, lo); + tcg_out_bswap32(s, TCG_TMP3, lo, 0); lo = TCG_TMP3; /* FALLTHRU */ case MO_32: @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_st_direct(TCGContext *s, TCGReg lo, TCGReg hi, tcg_out_opc_imm(s, OPC_SW, TCG_TMP0, base, 0); tcg_out_opc_imm(s, OPC_SW, TCG_TMP1, base, 4); } else { - tcg_out_bswap32(s, TCG_TMP3, MIPS_BE ? lo : hi); + tcg_out_bswap32(s, TCG_TMP3, MIPS_BE ? lo : hi, 0); tcg_out_opc_imm(s, OPC_SW, TCG_TMP3, base, 0); - tcg_out_bswap32(s, TCG_TMP3, MIPS_BE ? hi : lo); + tcg_out_bswap32(s, TCG_TMP3, MIPS_BE ? hi : lo, 0); tcg_out_opc_imm(s, OPC_SW, TCG_TMP3, base, 4); } break; @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_op(TCGContext *s, TCGOpcode opc, tcg_out_bswap16(s, a0, a1, a2); break; case INDEX_op_bswap32_i32: - tcg_out_bswap32(s, a0, a1); + tcg_out_bswap32(s, a0, a1, 0); break; case INDEX_op_bswap32_i64: - tcg_out_bswap32u(s, a0, a1); + tcg_out_bswap32(s, a0, a1, a2); break; case INDEX_op_bswap64_i64: tcg_out_bswap64(s, a0, a1); -- 2.25.1 The existing interpreter zero-extends, ignoring high bits. Simply add a separate sign-extension opcode if required. Ensure that the interpreter supports ext16s when bswap16 is enabled. Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/tci.c | 3 ++- tcg/tci/tcg-target.c.inc | 23 ++++++++++++++++++++--- 2 files changed, 22 insertions(+), 4 deletions(-) diff --git a/tcg/tci.c b/tcg/tci.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/tci.c +++ b/tcg/tci.c @@ -XXX,XX +XXX,XX @@ uintptr_t QEMU_DISABLE_CFI tcg_qemu_tb_exec(CPUArchState *env, regs[r0] = (int8_t)regs[r1]; break; #endif -#if TCG_TARGET_HAS_ext16s_i32 || TCG_TARGET_HAS_ext16s_i64 +#if TCG_TARGET_HAS_ext16s_i32 || TCG_TARGET_HAS_ext16s_i64 || \ + TCG_TARGET_HAS_bswap16_i32 || TCG_TARGET_HAS_bswap16_i64 CASE_32_64(ext16s) tci_args_rr(insn, &r0, &r1); regs[r0] = (int16_t)regs[r1]; diff --git a/tcg/tci/tcg-target.c.inc b/tcg/tci/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/tci/tcg-target.c.inc +++ b/tcg/tci/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, const TCGArg args[TCG_MAX_OP_ARGS], const int const_args[TCG_MAX_OP_ARGS]) { + TCGOpcode exts; + switch (opc) { case INDEX_op_exit_tb: tcg_out_op_p(s, opc, (void *)args[0]); @@ -XXX,XX +XXX,XX @@ static void tcg_out_op(TCGContext *s, TCGOpcode opc, CASE_64(ext32u) /* Optional (TCG_TARGET_HAS_ext32u_i64). */ CASE_64(ext_i32) CASE_64(extu_i32) - CASE_32_64(bswap16) /* Optional (TCG_TARGET_HAS_bswap16_*). */ - CASE_32_64(bswap32) /* Optional (TCG_TARGET_HAS_bswap32_*). */ - CASE_64(bswap64) /* Optional (TCG_TARGET_HAS_bswap64_i64). */ CASE_32_64(ctpop) /* Optional (TCG_TARGET_HAS_ctpop_*). */ + case INDEX_op_bswap32_i32: /* Optional (TCG_TARGET_HAS_bswap32_i32). */ + case INDEX_op_bswap64_i64: /* Optional (TCG_TARGET_HAS_bswap64_i64). */ tcg_out_op_rr(s, opc, args[0], args[1]); break; + case INDEX_op_bswap16_i32: /* Optional (TCG_TARGET_HAS_bswap16_i32). */ + exts = INDEX_op_ext16s_i32; + goto do_bswap; + case INDEX_op_bswap16_i64: /* Optional (TCG_TARGET_HAS_bswap16_i64). */ + exts = INDEX_op_ext16s_i64; + goto do_bswap; + case INDEX_op_bswap32_i64: /* Optional (TCG_TARGET_HAS_bswap32_i64). */ + exts = INDEX_op_ext32s_i64; + do_bswap: + /* The base tci bswaps zero-extend, and ignore high bits. */ + tcg_out_op_rr(s, opc, args[0], args[1]); + if (args[2] & TCG_BSWAP_OS) { + tcg_out_op_rr(s, exts, args[0], args[0]); + } + break; + CASE_32_64(add2) CASE_32_64(sub2) tcg_out_op_rrrrrr(s, opc, args[0], args[1], args[2], -- 2.25.1 Notice when the input is known to be zero-extended and force the TCG_BSWAP_IZ flag on. Honor the TCG_BSWAP_OS bit during constant folding. Propagate the input to the output mask. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 56 +++++++++++++++++++++++++++++++++++++++++++++----- 1 file changed, 51 insertions(+), 5 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static uint64_t do_constant_folding_2(TCGOpcode op, uint64_t x, uint64_t y) return (uint16_t)x; CASE_OP_32_64(bswap16): - return bswap16(x); + x = bswap16(x); + return y & TCG_BSWAP_OS ? (int16_t)x : x; CASE_OP_32_64(bswap32): - return bswap32(x); + x = bswap32(x); + return y & TCG_BSWAP_OS ? (int32_t)x : x; case INDEX_op_bswap64_i64: return bswap64(x); @@ -XXX,XX +XXX,XX @@ void tcg_optimize(TCGContext *s) } break; + CASE_OP_32_64(bswap16): + mask = arg_info(op->args[1])->mask; + if (mask <= 0xffff) { + op->args[2] |= TCG_BSWAP_IZ; + } + mask = bswap16(mask); + switch (op->args[2] & (TCG_BSWAP_OZ | TCG_BSWAP_OS)) { + case TCG_BSWAP_OZ: + break; + case TCG_BSWAP_OS: + mask = (int16_t)mask; + break; + default: /* undefined high bits */ + mask |= MAKE_64BIT_MASK(16, 48); + break; + } + break; + + case INDEX_op_bswap32_i64: + mask = arg_info(op->args[1])->mask; + if (mask <= 0xffffffffu) { + op->args[2] |= TCG_BSWAP_IZ; + } + mask = bswap32(mask); + switch (op->args[2] & (TCG_BSWAP_OZ | TCG_BSWAP_OS)) { + case TCG_BSWAP_OZ: + break; + case TCG_BSWAP_OS: + mask = (int32_t)mask; + break; + default: /* undefined high bits */ + mask |= MAKE_64BIT_MASK(32, 32); + break; + } + break; + default: break; } @@ -XXX,XX +XXX,XX @@ void tcg_optimize(TCGContext *s) CASE_OP_32_64(ext16s): CASE_OP_32_64(ext16u): CASE_OP_32_64(ctpop): - CASE_OP_32_64(bswap16): - CASE_OP_32_64(bswap32): - case INDEX_op_bswap64_i64: case INDEX_op_ext32s_i64: case INDEX_op_ext32u_i64: case INDEX_op_ext_i32_i64: @@ -XXX,XX +XXX,XX @@ void tcg_optimize(TCGContext *s) } goto do_default; + CASE_OP_32_64(bswap16): + CASE_OP_32_64(bswap32): + case INDEX_op_bswap64_i64: + if (arg_is_const(op->args[1])) { + tmp = do_constant_folding(opc, arg_info(op->args[1])->val, + op->args[2]); + tcg_opt_gen_movi(s, &temps_used, op, op->args[0], tmp); + break; + } + goto do_default; + CASE_OP_32_64(add): CASE_OP_32_64(sub): CASE_OP_32_64(mul): -- 2.25.1 Implement the new semantics in the fallback expansion. Change all callers to supply the flags that keep the semantics unchanged locally. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/tcg/tcg-op.h | 8 +-- target/arm/translate-a64.c | 12 ++-- target/arm/translate.c | 2 +- target/i386/tcg/translate.c | 2 +- target/mips/tcg/mxu_translate.c | 2 +- target/s390x/translate.c | 4 +- target/sh4/translate.c | 2 +- tcg/tcg-op.c | 121 ++++++++++++++++++++++---------- 8 files changed, 99 insertions(+), 54 deletions(-) diff --git a/include/tcg/tcg-op.h b/include/tcg/tcg-op.h index XXXXXXX..XXXXXXX 100644 --- a/include/tcg/tcg-op.h +++ b/include/tcg/tcg-op.h @@ -XXX,XX +XXX,XX @@ void tcg_gen_ext8s_i32(TCGv_i32 ret, TCGv_i32 arg); void tcg_gen_ext16s_i32(TCGv_i32 ret, TCGv_i32 arg); void tcg_gen_ext8u_i32(TCGv_i32 ret, TCGv_i32 arg); void tcg_gen_ext16u_i32(TCGv_i32 ret, TCGv_i32 arg); -void tcg_gen_bswap16_i32(TCGv_i32 ret, TCGv_i32 arg); +void tcg_gen_bswap16_i32(TCGv_i32 ret, TCGv_i32 arg, int flags); void tcg_gen_bswap32_i32(TCGv_i32 ret, TCGv_i32 arg); void tcg_gen_smin_i32(TCGv_i32, TCGv_i32 arg1, TCGv_i32 arg2); void tcg_gen_smax_i32(TCGv_i32, TCGv_i32 arg1, TCGv_i32 arg2); @@ -XXX,XX +XXX,XX @@ void tcg_gen_ext32s_i64(TCGv_i64 ret, TCGv_i64 arg); void tcg_gen_ext8u_i64(TCGv_i64 ret, TCGv_i64 arg); void tcg_gen_ext16u_i64(TCGv_i64 ret, TCGv_i64 arg); void tcg_gen_ext32u_i64(TCGv_i64 ret, TCGv_i64 arg); -void tcg_gen_bswap16_i64(TCGv_i64 ret, TCGv_i64 arg); -void tcg_gen_bswap32_i64(TCGv_i64 ret, TCGv_i64 arg); +void tcg_gen_bswap16_i64(TCGv_i64 ret, TCGv_i64 arg, int flags); +void tcg_gen_bswap32_i64(TCGv_i64 ret, TCGv_i64 arg, int flags); void tcg_gen_bswap64_i64(TCGv_i64 ret, TCGv_i64 arg); void tcg_gen_smin_i64(TCGv_i64, TCGv_i64 arg1, TCGv_i64 arg2); void tcg_gen_smax_i64(TCGv_i64, TCGv_i64 arg1, TCGv_i64 arg2); @@ -XXX,XX +XXX,XX @@ void tcg_gen_stl_vec(TCGv_vec r, TCGv_ptr base, TCGArg offset, TCGType t); #define tcg_gen_ext32u_tl tcg_gen_mov_i32 #define tcg_gen_ext32s_tl tcg_gen_mov_i32 #define tcg_gen_bswap16_tl tcg_gen_bswap16_i32 -#define tcg_gen_bswap32_tl tcg_gen_bswap32_i32 +#define tcg_gen_bswap32_tl(D, S, F) tcg_gen_bswap32_i32(D, S) #define tcg_gen_bswap_tl tcg_gen_bswap32_i32 #define tcg_gen_concat_tl_i64 tcg_gen_concat_i32_i64 #define tcg_gen_extr_i64_tl tcg_gen_extr_i64_i32 diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_rev32(DisasContext *s, unsigned int sf, /* bswap32_i64 requires zero high word */ tcg_gen_ext32u_i64(tcg_tmp, tcg_rn); - tcg_gen_bswap32_i64(tcg_rd, tcg_tmp); + tcg_gen_bswap32_i64(tcg_rd, tcg_tmp, TCG_BSWAP_IZ | TCG_BSWAP_OZ); tcg_gen_shri_i64(tcg_tmp, tcg_rn, 32); - tcg_gen_bswap32_i64(tcg_tmp, tcg_tmp); + tcg_gen_bswap32_i64(tcg_tmp, tcg_tmp, TCG_BSWAP_IZ | TCG_BSWAP_OZ); tcg_gen_concat32_i64(tcg_rd, tcg_rd, tcg_tmp); tcg_temp_free_i64(tcg_tmp); } else { tcg_gen_ext32u_i64(tcg_rd, cpu_reg(s, rn)); - tcg_gen_bswap32_i64(tcg_rd, tcg_rd); + tcg_gen_bswap32_i64(tcg_rd, tcg_rd, TCG_BSWAP_IZ | TCG_BSWAP_OZ); } } @@ -XXX,XX +XXX,XX @@ static void handle_rev(DisasContext *s, int opcode, bool u, read_vec_element(s, tcg_tmp, rn, i, grp_size); switch (grp_size) { case MO_16: - tcg_gen_bswap16_i64(tcg_tmp, tcg_tmp); + tcg_gen_bswap16_i64(tcg_tmp, tcg_tmp, + TCG_BSWAP_IZ | TCG_BSWAP_OZ); break; case MO_32: - tcg_gen_bswap32_i64(tcg_tmp, tcg_tmp); + tcg_gen_bswap32_i64(tcg_tmp, tcg_tmp, + TCG_BSWAP_IZ | TCG_BSWAP_OZ); break; case MO_64: tcg_gen_bswap64_i64(tcg_tmp, tcg_tmp); diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ void gen_rev16(TCGv_i32 dest, TCGv_i32 var) static void gen_revsh(TCGv_i32 dest, TCGv_i32 var) { tcg_gen_ext16u_i32(var, var); - tcg_gen_bswap16_i32(var, var); + tcg_gen_bswap16_i32(var, var, TCG_BSWAP_IZ | TCG_BSWAP_OZ); tcg_gen_ext16s_i32(dest, var); } diff --git a/target/i386/tcg/translate.c b/target/i386/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/i386/tcg/translate.c +++ b/target/i386/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static target_ulong disas_insn(DisasContext *s, CPUState *cpu) { gen_op_mov_v_reg(s, MO_32, s->T0, reg); tcg_gen_ext32u_tl(s->T0, s->T0); - tcg_gen_bswap32_tl(s->T0, s->T0); + tcg_gen_bswap32_tl(s->T0, s->T0, TCG_BSWAP_IZ | TCG_BSWAP_OZ); gen_op_mov_reg_v(s, MO_32, reg, s->T0); } break; diff --git a/target/mips/tcg/mxu_translate.c b/target/mips/tcg/mxu_translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/mips/tcg/mxu_translate.c +++ b/target/mips/tcg/mxu_translate.c @@ -XXX,XX +XXX,XX @@ static void gen_mxu_s32ldd_s32lddr(DisasContext *ctx) if (sel == 1) { /* S32LDDR */ - tcg_gen_bswap32_tl(t1, t1); + tcg_gen_bswap32_tl(t1, t1, TCG_BSWAP_IZ | TCG_BSWAP_OZ); } gen_store_mxu_gpr(t1, XRa); diff --git a/target/s390x/translate.c b/target/s390x/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/translate.c +++ b/target/s390x/translate.c @@ -XXX,XX +XXX,XX @@ static DisasJumpType op_rosbg(DisasContext *s, DisasOps *o) static DisasJumpType op_rev16(DisasContext *s, DisasOps *o) { - tcg_gen_bswap16_i64(o->out, o->in2); + tcg_gen_bswap16_i64(o->out, o->in2, TCG_BSWAP_IZ | TCG_BSWAP_OZ); return DISAS_NEXT; } static DisasJumpType op_rev32(DisasContext *s, DisasOps *o) { - tcg_gen_bswap32_i64(o->out, o->in2); + tcg_gen_bswap32_i64(o->out, o->in2, TCG_BSWAP_IZ | TCG_BSWAP_OZ); return DISAS_NEXT; } diff --git a/target/sh4/translate.c b/target/sh4/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/sh4/translate.c +++ b/target/sh4/translate.c @@ -XXX,XX +XXX,XX @@ static void _decode_opc(DisasContext * ctx) { TCGv low = tcg_temp_new(); tcg_gen_ext16u_i32(low, REG(B7_4)); - tcg_gen_bswap16_i32(low, low); + tcg_gen_bswap16_i32(low, low, TCG_BSWAP_IZ | TCG_BSWAP_OZ); tcg_gen_deposit_i32(REG(B11_8), REG(B7_4), low, 0, 16); tcg_temp_free(low); } diff --git a/tcg/tcg-op.c b/tcg/tcg-op.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg-op.c +++ b/tcg/tcg-op.c @@ -XXX,XX +XXX,XX @@ void tcg_gen_ext16u_i32(TCGv_i32 ret, TCGv_i32 arg) } } -/* Note: we assume the two high bytes are set to zero */ -void tcg_gen_bswap16_i32(TCGv_i32 ret, TCGv_i32 arg) +void tcg_gen_bswap16_i32(TCGv_i32 ret, TCGv_i32 arg, int flags) { + /* Only one extension flag may be present. */ + tcg_debug_assert(!(flags & TCG_BSWAP_OS) || !(flags & TCG_BSWAP_OZ)); + if (TCG_TARGET_HAS_bswap16_i32) { - tcg_gen_op3i_i32(INDEX_op_bswap16_i32, ret, arg, - TCG_BSWAP_IZ | TCG_BSWAP_OZ); + tcg_gen_op3i_i32(INDEX_op_bswap16_i32, ret, arg, flags); } else { TCGv_i32 t0 = tcg_temp_new_i32(); + TCGv_i32 t1 = tcg_temp_new_i32(); - tcg_gen_ext8u_i32(t0, arg); - tcg_gen_shli_i32(t0, t0, 8); - tcg_gen_shri_i32(ret, arg, 8); - tcg_gen_or_i32(ret, ret, t0); + tcg_gen_shri_i32(t0, arg, 8); + if (!(flags & TCG_BSWAP_IZ)) { + tcg_gen_ext8u_i32(t0, t0); + } + + if (flags & TCG_BSWAP_OS) { + tcg_gen_shli_i32(t1, arg, 24); + tcg_gen_sari_i32(t1, t1, 16); + } else if (flags & TCG_BSWAP_OZ) { + tcg_gen_ext8u_i32(t1, arg); + tcg_gen_shli_i32(t1, t1, 8); + } else { + tcg_gen_shli_i32(t1, arg, 8); + } + + tcg_gen_or_i32(ret, t0, t1); tcg_temp_free_i32(t0); + tcg_temp_free_i32(t1); } } @@ -XXX,XX +XXX,XX @@ void tcg_gen_ext32u_i64(TCGv_i64 ret, TCGv_i64 arg) } } -/* Note: we assume the six high bytes are set to zero */ -void tcg_gen_bswap16_i64(TCGv_i64 ret, TCGv_i64 arg) +void tcg_gen_bswap16_i64(TCGv_i64 ret, TCGv_i64 arg, int flags) { + /* Only one extension flag may be present. */ + tcg_debug_assert(!(flags & TCG_BSWAP_OS) || !(flags & TCG_BSWAP_OZ)); + if (TCG_TARGET_REG_BITS == 32) { - tcg_gen_bswap16_i32(TCGV_LOW(ret), TCGV_LOW(arg)); - tcg_gen_movi_i32(TCGV_HIGH(ret), 0); + tcg_gen_bswap16_i32(TCGV_LOW(ret), TCGV_LOW(arg), flags); + if (flags & TCG_BSWAP_OS) { + tcg_gen_sari_i32(TCGV_HIGH(ret), TCGV_LOW(ret), 31); + } else { + tcg_gen_movi_i32(TCGV_HIGH(ret), 0); + } } else if (TCG_TARGET_HAS_bswap16_i64) { - tcg_gen_op3i_i64(INDEX_op_bswap16_i64, ret, arg, - TCG_BSWAP_IZ | TCG_BSWAP_OZ); + tcg_gen_op3i_i64(INDEX_op_bswap16_i64, ret, arg, flags); } else { TCGv_i64 t0 = tcg_temp_new_i64(); + TCGv_i64 t1 = tcg_temp_new_i64(); - tcg_gen_ext8u_i64(t0, arg); - tcg_gen_shli_i64(t0, t0, 8); - tcg_gen_shri_i64(ret, arg, 8); - tcg_gen_or_i64(ret, ret, t0); + tcg_gen_shri_i64(t0, arg, 8); + if (!(flags & TCG_BSWAP_IZ)) { + tcg_gen_ext8u_i64(t0, t0); + } + + if (flags & TCG_BSWAP_OS) { + tcg_gen_shli_i64(t1, arg, 56); + tcg_gen_sari_i64(t1, t1, 48); + } else if (flags & TCG_BSWAP_OZ) { + tcg_gen_ext8u_i64(t1, arg); + tcg_gen_shli_i64(t1, t1, 8); + } else { + tcg_gen_shli_i64(t1, arg, 8); + } + + tcg_gen_or_i64(ret, t0, t1); tcg_temp_free_i64(t0); + tcg_temp_free_i64(t1); } } -/* Note: we assume the four high bytes are set to zero */ -void tcg_gen_bswap32_i64(TCGv_i64 ret, TCGv_i64 arg) +void tcg_gen_bswap32_i64(TCGv_i64 ret, TCGv_i64 arg, int flags) { + /* Only one extension flag may be present. */ + tcg_debug_assert(!(flags & TCG_BSWAP_OS) || !(flags & TCG_BSWAP_OZ)); + if (TCG_TARGET_REG_BITS == 32) { tcg_gen_bswap32_i32(TCGV_LOW(ret), TCGV_LOW(arg)); - tcg_gen_movi_i32(TCGV_HIGH(ret), 0); + if (flags & TCG_BSWAP_OS) { + tcg_gen_sari_i32(TCGV_HIGH(ret), TCGV_LOW(ret), 31); + } else { + tcg_gen_movi_i32(TCGV_HIGH(ret), 0); + } } else if (TCG_TARGET_HAS_bswap32_i64) { - tcg_gen_op3i_i64(INDEX_op_bswap32_i64, ret, arg, - TCG_BSWAP_IZ | TCG_BSWAP_OZ); + tcg_gen_op3i_i64(INDEX_op_bswap32_i64, ret, arg, flags); } else { TCGv_i64 t0 = tcg_temp_new_i64(); TCGv_i64 t1 = tcg_temp_new_i64(); TCGv_i64 t2 = tcg_constant_i64(0x00ff00ff); - /* arg = ....abcd */ - tcg_gen_shri_i64(t0, arg, 8); /* t0 = .....abc */ - tcg_gen_and_i64(t1, arg, t2); /* t1 = .....b.d */ - tcg_gen_and_i64(t0, t0, t2); /* t0 = .....a.c */ - tcg_gen_shli_i64(t1, t1, 8); /* t1 = ....b.d. */ - tcg_gen_or_i64(ret, t0, t1); /* ret = ....badc */ + /* arg = xxxxabcd */ + tcg_gen_shri_i64(t0, arg, 8); /* t0 = .xxxxabc */ + tcg_gen_and_i64(t1, arg, t2); /* t1 = .....b.d */ + tcg_gen_and_i64(t0, t0, t2); /* t0 = .....a.c */ + tcg_gen_shli_i64(t1, t1, 8); /* t1 = ....b.d. */ + tcg_gen_or_i64(ret, t0, t1); /* ret = ....badc */ - tcg_gen_shli_i64(t1, ret, 48); /* t1 = dc...... */ - tcg_gen_shri_i64(t0, ret, 16); /* t0 = ......ba */ - tcg_gen_shri_i64(t1, t1, 32); /* t1 = ....dc.. */ - tcg_gen_or_i64(ret, t0, t1); /* ret = ....dcba */ + tcg_gen_shli_i64(t1, ret, 48); /* t1 = dc...... */ + tcg_gen_shri_i64(t0, ret, 16); /* t0 = ......ba */ + if (flags & TCG_BSWAP_OS) { + tcg_gen_sari_i64(t1, t1, 32); /* t1 = ssssdc.. */ + } else { + tcg_gen_shri_i64(t1, t1, 32); /* t1 = ....dc.. */ + } + tcg_gen_or_i64(ret, t0, t1); /* ret = ssssdcba */ tcg_temp_free_i64(t0); tcg_temp_free_i64(t1); @@ -XXX,XX +XXX,XX @@ void tcg_gen_qemu_ld_i32(TCGv_i32 val, TCGv addr, TCGArg idx, MemOp memop) if ((orig_memop ^ memop) & MO_BSWAP) { switch (orig_memop & MO_SIZE) { case MO_16: - tcg_gen_bswap16_i32(val, val); + tcg_gen_bswap16_i32(val, val, TCG_BSWAP_IZ | TCG_BSWAP_OZ); if (orig_memop & MO_SIGN) { tcg_gen_ext16s_i32(val, val); } @@ -XXX,XX +XXX,XX @@ void tcg_gen_qemu_st_i32(TCGv_i32 val, TCGv addr, TCGArg idx, MemOp memop) switch (memop & MO_SIZE) { case MO_16: tcg_gen_ext16u_i32(swap, val); - tcg_gen_bswap16_i32(swap, swap); + tcg_gen_bswap16_i32(swap, swap, TCG_BSWAP_IZ | TCG_BSWAP_OZ); break; case MO_32: tcg_gen_bswap32_i32(swap, val); @@ -XXX,XX +XXX,XX @@ void tcg_gen_qemu_ld_i64(TCGv_i64 val, TCGv addr, TCGArg idx, MemOp memop) if ((orig_memop ^ memop) & MO_BSWAP) { switch (orig_memop & MO_SIZE) { case MO_16: - tcg_gen_bswap16_i64(val, val); + tcg_gen_bswap16_i64(val, val, TCG_BSWAP_IZ | TCG_BSWAP_OZ); if (orig_memop & MO_SIGN) { tcg_gen_ext16s_i64(val, val); } break; case MO_32: - tcg_gen_bswap32_i64(val, val); + tcg_gen_bswap32_i64(val, val, TCG_BSWAP_IZ | TCG_BSWAP_OZ); if (orig_memop & MO_SIGN) { tcg_gen_ext32s_i64(val, val); } @@ -XXX,XX +XXX,XX @@ void tcg_gen_qemu_st_i64(TCGv_i64 val, TCGv addr, TCGArg idx, MemOp memop) switch (memop & MO_SIZE) { case MO_16: tcg_gen_ext16u_i64(swap, val); - tcg_gen_bswap16_i64(swap, swap); + tcg_gen_bswap16_i64(swap, swap, TCG_BSWAP_IZ | TCG_BSWAP_OZ); break; case MO_32: tcg_gen_ext32u_i64(swap, val); - tcg_gen_bswap32_i64(swap, swap); + tcg_gen_bswap32_i64(swap, swap, TCG_BSWAP_IZ | TCG_BSWAP_OZ); break; case MO_64: tcg_gen_bswap64_i64(swap, val); -- 2.25.1 We can perform any required sign-extension via TCG_BSWAP_OS. Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/tcg-op.c | 24 ++++++++++-------------- 1 file changed, 10 insertions(+), 14 deletions(-) diff --git a/tcg/tcg-op.c b/tcg/tcg-op.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg-op.c +++ b/tcg/tcg-op.c @@ -XXX,XX +XXX,XX @@ void tcg_gen_qemu_ld_i32(TCGv_i32 val, TCGv addr, TCGArg idx, MemOp memop) orig_memop = memop; if (!TCG_TARGET_HAS_MEMORY_BSWAP && (memop & MO_BSWAP)) { memop &= ~MO_BSWAP; - /* The bswap primitive requires zero-extended input. */ + /* The bswap primitive benefits from zero-extended input. */ if ((memop & MO_SSIZE) == MO_SW) { memop &= ~MO_SIGN; } @@ -XXX,XX +XXX,XX @@ void tcg_gen_qemu_ld_i32(TCGv_i32 val, TCGv addr, TCGArg idx, MemOp memop) if ((orig_memop ^ memop) & MO_BSWAP) { switch (orig_memop & MO_SIZE) { case MO_16: - tcg_gen_bswap16_i32(val, val, TCG_BSWAP_IZ | TCG_BSWAP_OZ); - if (orig_memop & MO_SIGN) { - tcg_gen_ext16s_i32(val, val); - } + tcg_gen_bswap16_i32(val, val, (orig_memop & MO_SIGN + ? TCG_BSWAP_IZ | TCG_BSWAP_OS + : TCG_BSWAP_IZ | TCG_BSWAP_OZ)); break; case MO_32: tcg_gen_bswap32_i32(val, val); @@ -XXX,XX +XXX,XX @@ void tcg_gen_qemu_ld_i64(TCGv_i64 val, TCGv addr, TCGArg idx, MemOp memop) orig_memop = memop; if (!TCG_TARGET_HAS_MEMORY_BSWAP && (memop & MO_BSWAP)) { memop &= ~MO_BSWAP; - /* The bswap primitive requires zero-extended input. */ + /* The bswap primitive benefits from zero-extended input. */ if ((memop & MO_SIGN) && (memop & MO_SIZE) < MO_64) { memop &= ~MO_SIGN; } @@ -XXX,XX +XXX,XX @@ void tcg_gen_qemu_ld_i64(TCGv_i64 val, TCGv addr, TCGArg idx, MemOp memop) plugin_gen_mem_callbacks(addr, info); if ((orig_memop ^ memop) & MO_BSWAP) { + int flags = (orig_memop & MO_SIGN + ? TCG_BSWAP_IZ | TCG_BSWAP_OS + : TCG_BSWAP_IZ | TCG_BSWAP_OZ); switch (orig_memop & MO_SIZE) { case MO_16: - tcg_gen_bswap16_i64(val, val, TCG_BSWAP_IZ | TCG_BSWAP_OZ); - if (orig_memop & MO_SIGN) { - tcg_gen_ext16s_i64(val, val); - } + tcg_gen_bswap16_i64(val, val, flags); break; case MO_32: - tcg_gen_bswap32_i64(val, val, TCG_BSWAP_IZ | TCG_BSWAP_OZ); - if (orig_memop & MO_SIGN) { - tcg_gen_ext32s_i64(val, val); - } + tcg_gen_bswap32_i64(val, val, flags); break; case MO_64: tcg_gen_bswap64_i64(val, val); -- 2.25.1 By removing TCG_BSWAP_IZ we indicate that the input is not zero-extended, and thus can remove an explicit extend. By removing TCG_BSWAP_OZ, we allow the implementation to leave high bits set, which will be ignored by the store. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/tcg-op.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) diff --git a/tcg/tcg-op.c b/tcg/tcg-op.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/tcg-op.c +++ b/tcg/tcg-op.c @@ -XXX,XX +XXX,XX @@ void tcg_gen_qemu_st_i32(TCGv_i32 val, TCGv addr, TCGArg idx, MemOp memop) swap = tcg_temp_new_i32(); switch (memop & MO_SIZE) { case MO_16: - tcg_gen_ext16u_i32(swap, val); - tcg_gen_bswap16_i32(swap, swap, TCG_BSWAP_IZ | TCG_BSWAP_OZ); + tcg_gen_bswap16_i32(swap, val, 0); break; case MO_32: tcg_gen_bswap32_i32(swap, val); @@ -XXX,XX +XXX,XX @@ void tcg_gen_qemu_st_i64(TCGv_i64 val, TCGv addr, TCGArg idx, MemOp memop) swap = tcg_temp_new_i64(); switch (memop & MO_SIZE) { case MO_16: - tcg_gen_ext16u_i64(swap, val); - tcg_gen_bswap16_i64(swap, swap, TCG_BSWAP_IZ | TCG_BSWAP_OZ); + tcg_gen_bswap16_i64(swap, val, 0); break; case MO_32: - tcg_gen_ext32u_i64(swap, val); - tcg_gen_bswap32_i64(swap, swap, TCG_BSWAP_IZ | TCG_BSWAP_OZ); + tcg_gen_bswap32_i64(swap, val, 0); break; case MO_64: tcg_gen_bswap64_i64(swap, val); -- 2.25.1 For the sf version, we are performing two 32-bit bswaps in either half of the register. This is equivalent to performing one 64-bit bswap followed by a rotate. For the non-sf version, we can remove TCG_BSWAP_IZ and the preceding zero-extension. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/arm/translate-a64.c | 17 ++++------------- 1 file changed, 4 insertions(+), 13 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_rev32(DisasContext *s, unsigned int sf, unsigned int rn, unsigned int rd) { TCGv_i64 tcg_rd = cpu_reg(s, rd); + TCGv_i64 tcg_rn = cpu_reg(s, rn); if (sf) { - TCGv_i64 tcg_tmp = tcg_temp_new_i64(); - TCGv_i64 tcg_rn = read_cpu_reg(s, rn, sf); - - /* bswap32_i64 requires zero high word */ - tcg_gen_ext32u_i64(tcg_tmp, tcg_rn); - tcg_gen_bswap32_i64(tcg_rd, tcg_tmp, TCG_BSWAP_IZ | TCG_BSWAP_OZ); - tcg_gen_shri_i64(tcg_tmp, tcg_rn, 32); - tcg_gen_bswap32_i64(tcg_tmp, tcg_tmp, TCG_BSWAP_IZ | TCG_BSWAP_OZ); - tcg_gen_concat32_i64(tcg_rd, tcg_rd, tcg_tmp); - - tcg_temp_free_i64(tcg_tmp); + tcg_gen_bswap64_i64(tcg_rd, tcg_rn); + tcg_gen_rotri_i64(tcg_rd, tcg_rd, 32); } else { - tcg_gen_ext32u_i64(tcg_rd, cpu_reg(s, rn)); - tcg_gen_bswap32_i64(tcg_rd, tcg_rd, TCG_BSWAP_IZ | TCG_BSWAP_OZ); + tcg_gen_bswap32_i64(tcg_rd, tcg_rn, TCG_BSWAP_OZ); } } -- 2.25.1 We can eliminate the requirement for a zero-extended output, because the following store will ignore any garbage high bits. Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/arm/translate-a64.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void handle_rev(DisasContext *s, int opcode, bool u, read_vec_element(s, tcg_tmp, rn, i, grp_size); switch (grp_size) { case MO_16: - tcg_gen_bswap16_i64(tcg_tmp, tcg_tmp, - TCG_BSWAP_IZ | TCG_BSWAP_OZ); + tcg_gen_bswap16_i64(tcg_tmp, tcg_tmp, TCG_BSWAP_IZ); break; case MO_32: - tcg_gen_bswap32_i64(tcg_tmp, tcg_tmp, - TCG_BSWAP_IZ | TCG_BSWAP_OZ); + tcg_gen_bswap32_i64(tcg_tmp, tcg_tmp, TCG_BSWAP_IZ); break; case MO_64: tcg_gen_bswap64_i64(tcg_tmp, tcg_tmp); -- 2.25.1 The new bswap flags can implement the semantics exactly. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/arm/translate.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ void gen_rev16(TCGv_i32 dest, TCGv_i32 var) /* Byteswap low halfword and sign extend. */ static void gen_revsh(TCGv_i32 dest, TCGv_i32 var) { - tcg_gen_ext16u_i32(var, var); - tcg_gen_bswap16_i32(var, var, TCG_BSWAP_IZ | TCG_BSWAP_OZ); - tcg_gen_ext16s_i32(dest, var); + tcg_gen_bswap16_i32(var, var, TCG_BSWAP_OS); } /* Dual 16-bit add. Result placed in t0 and t1 is marked as dead. -- 2.25.1 Use a break instead of an ifdefed else. There's no need to move the values through s->T0. Remove TCG_BSWAP_IZ and the preceding zero-extension. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/i386/tcg/translate.c | 14 ++++---------- 1 file changed, 4 insertions(+), 10 deletions(-) diff --git a/target/i386/tcg/translate.c b/target/i386/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/i386/tcg/translate.c +++ b/target/i386/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static target_ulong disas_insn(DisasContext *s, CPUState *cpu) reg = (b & 7) | REX_B(s); #ifdef TARGET_X86_64 if (dflag == MO_64) { - gen_op_mov_v_reg(s, MO_64, s->T0, reg); - tcg_gen_bswap64_i64(s->T0, s->T0); - gen_op_mov_reg_v(s, MO_64, reg, s->T0); - } else -#endif - { - gen_op_mov_v_reg(s, MO_32, s->T0, reg); - tcg_gen_ext32u_tl(s->T0, s->T0); - tcg_gen_bswap32_tl(s->T0, s->T0, TCG_BSWAP_IZ | TCG_BSWAP_OZ); - gen_op_mov_reg_v(s, MO_32, reg, s->T0); + tcg_gen_bswap64_i64(cpu_regs[reg], cpu_regs[reg]); + break; } +#endif + tcg_gen_bswap32_tl(cpu_regs[reg], cpu_regs[reg], TCG_BSWAP_OZ); break; case 0xd6: /* salc */ if (CODE64(s)) -- 2.25.1 Remove TCG_BSWAP_IZ and the preceding zero-extension. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/sh4/translate.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/target/sh4/translate.c b/target/sh4/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/sh4/translate.c +++ b/target/sh4/translate.c @@ -XXX,XX +XXX,XX @@ static void _decode_opc(DisasContext * ctx) case 0x6008: /* swap.b Rm,Rn */ { TCGv low = tcg_temp_new(); - tcg_gen_ext16u_i32(low, REG(B7_4)); - tcg_gen_bswap16_i32(low, low, TCG_BSWAP_IZ | TCG_BSWAP_OZ); + tcg_gen_bswap16_i32(low, REG(B7_4), 0); tcg_gen_deposit_i32(REG(B11_8), REG(B7_4), low, 0, 16); tcg_temp_free(low); } -- 2.25.1 There were two bugs here: (1) the required endianness was not present in the MemOp, and (2) we were not providing a zero-extended input to the bswap as semantics required. The best fix is to fold the bswap into the memory operation, producing the desired result directly. Acked-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/mips/tcg/mxu_translate.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/target/mips/tcg/mxu_translate.c b/target/mips/tcg/mxu_translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/mips/tcg/mxu_translate.c +++ b/target/mips/tcg/mxu_translate.c @@ -XXX,XX +XXX,XX @@ static void gen_mxu_s32ldd_s32lddr(DisasContext *ctx) tcg_gen_ori_tl(t1, t1, 0xFFFFF000); } tcg_gen_add_tl(t1, t0, t1); - tcg_gen_qemu_ld_tl(t1, t1, ctx->mem_idx, MO_SL); + tcg_gen_qemu_ld_tl(t1, t1, ctx->mem_idx, MO_TESL ^ (sel * MO_BSWAP)); - if (sel == 1) { - /* S32LDDR */ - tcg_gen_bswap32_tl(t1, t1, TCG_BSWAP_IZ | TCG_BSWAP_OZ); - } gen_store_mxu_gpr(t1, XRa); tcg_temp_free(t0); -- 2.25.1 Now that the middle-end can replicate the same tricks as tcg/arm used for optimizing bswap for signed loads and for stores, do not pretend to have these memory ops in the backend. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/arm/tcg-target.h | 2 +- tcg/arm/tcg-target.c.inc | 214 ++++++++++++++------------------------- 2 files changed, 77 insertions(+), 139 deletions(-) diff --git a/tcg/arm/tcg-target.h b/tcg/arm/tcg-target.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/arm/tcg-target.h +++ b/tcg/arm/tcg-target.h @@ -XXX,XX +XXX,XX @@ extern bool use_neon_instructions; #define TCG_TARGET_HAS_cmpsel_vec 0 #define TCG_TARGET_DEFAULT_MO (0) -#define TCG_TARGET_HAS_MEMORY_BSWAP 1 +#define TCG_TARGET_HAS_MEMORY_BSWAP 0 /* not defined -- call should be eliminated at compile time */ void tb_target_set_jmp_target(uintptr_t, uintptr_t, uintptr_t, uintptr_t); diff --git a/tcg/arm/tcg-target.c.inc b/tcg/arm/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/arm/tcg-target.c.inc +++ b/tcg/arm/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_vldst(TCGContext *s, ARMInsn insn, /* helper signature: helper_ret_ld_mmu(CPUState *env, target_ulong addr, * int mmu_idx, uintptr_t ra) */ -static void * const qemu_ld_helpers[16] = { +static void * const qemu_ld_helpers[8] = { [MO_UB] = helper_ret_ldub_mmu, [MO_SB] = helper_ret_ldsb_mmu, - - [MO_LEUW] = helper_le_lduw_mmu, - [MO_LEUL] = helper_le_ldul_mmu, - [MO_LEQ] = helper_le_ldq_mmu, - [MO_LESW] = helper_le_ldsw_mmu, - [MO_LESL] = helper_le_ldul_mmu, - - [MO_BEUW] = helper_be_lduw_mmu, - [MO_BEUL] = helper_be_ldul_mmu, - [MO_BEQ] = helper_be_ldq_mmu, - [MO_BESW] = helper_be_ldsw_mmu, - [MO_BESL] = helper_be_ldul_mmu, +#ifdef HOST_WORDS_BIGENDIAN + [MO_UW] = helper_be_lduw_mmu, + [MO_UL] = helper_be_ldul_mmu, + [MO_Q] = helper_be_ldq_mmu, + [MO_SW] = helper_be_ldsw_mmu, + [MO_SL] = helper_be_ldul_mmu, +#else + [MO_UW] = helper_le_lduw_mmu, + [MO_UL] = helper_le_ldul_mmu, + [MO_Q] = helper_le_ldq_mmu, + [MO_SW] = helper_le_ldsw_mmu, + [MO_SL] = helper_le_ldul_mmu, +#endif }; /* helper signature: helper_ret_st_mmu(CPUState *env, target_ulong addr, * uintxx_t val, int mmu_idx, uintptr_t ra) */ -static void * const qemu_st_helpers[16] = { - [MO_UB] = helper_ret_stb_mmu, - [MO_LEUW] = helper_le_stw_mmu, - [MO_LEUL] = helper_le_stl_mmu, - [MO_LEQ] = helper_le_stq_mmu, - [MO_BEUW] = helper_be_stw_mmu, - [MO_BEUL] = helper_be_stl_mmu, - [MO_BEQ] = helper_be_stq_mmu, +static void * const qemu_st_helpers[4] = { + [MO_8] = helper_ret_stb_mmu, +#ifdef HOST_WORDS_BIGENDIAN + [MO_16] = helper_be_stw_mmu, + [MO_32] = helper_be_stl_mmu, + [MO_64] = helper_be_stq_mmu, +#else + [MO_16] = helper_le_stw_mmu, + [MO_32] = helper_le_stl_mmu, + [MO_64] = helper_le_stq_mmu, +#endif }; /* Helper routines for marshalling helper function arguments into @@ -XXX,XX +XXX,XX @@ static bool tcg_out_qemu_ld_slow_path(TCGContext *s, TCGLabelQemuLdst *lb) icache usage. For pre-armv6, use the signed helpers since we do not have a single insn sign-extend. */ if (use_armv6_instructions) { - func = qemu_ld_helpers[opc & (MO_BSWAP | MO_SIZE)]; + func = qemu_ld_helpers[opc & MO_SIZE]; } else { - func = qemu_ld_helpers[opc & (MO_BSWAP | MO_SSIZE)]; + func = qemu_ld_helpers[opc & MO_SSIZE]; if (opc & MO_SIGN) { opc = MO_UL; } @@ -XXX,XX +XXX,XX @@ static bool tcg_out_qemu_st_slow_path(TCGContext *s, TCGLabelQemuLdst *lb) argreg = tcg_out_arg_reg32(s, argreg, TCG_REG_R14); /* Tail-call to the helper, which will return to the fast path. */ - tcg_out_goto(s, COND_AL, qemu_st_helpers[opc & (MO_BSWAP | MO_SIZE)]); + tcg_out_goto(s, COND_AL, qemu_st_helpers[opc & MO_SIZE]); return true; } #endif /* SOFTMMU */ @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_qemu_ld_index(TCGContext *s, MemOp opc, TCGReg datalo, TCGReg datahi, TCGReg addrlo, TCGReg addend) { - MemOp bswap = opc & MO_BSWAP; + /* Byte swapping is left to middle-end expansion. */ + tcg_debug_assert((opc & MO_BSWAP) == 0); switch (opc & MO_SSIZE) { case MO_UB: @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_qemu_ld_index(TCGContext *s, MemOp opc, break; case MO_UW: tcg_out_ld16u_r(s, COND_AL, datalo, addrlo, addend); - if (bswap) { - tcg_out_bswap16(s, COND_AL, datalo, datalo, - TCG_BSWAP_IZ | TCG_BSWAP_OZ); - } break; case MO_SW: - if (bswap) { - tcg_out_ld16u_r(s, COND_AL, datalo, addrlo, addend); - tcg_out_bswap16(s, COND_AL, datalo, datalo, - TCG_BSWAP_IZ | TCG_BSWAP_OS); - } else { - tcg_out_ld16s_r(s, COND_AL, datalo, addrlo, addend); - } + tcg_out_ld16s_r(s, COND_AL, datalo, addrlo, addend); break; case MO_UL: - default: tcg_out_ld32_r(s, COND_AL, datalo, addrlo, addend); - if (bswap) { - tcg_out_bswap32(s, COND_AL, datalo, datalo); - } break; case MO_Q: - { - TCGReg dl = (bswap ? datahi : datalo); - TCGReg dh = (bswap ? datalo : datahi); - - /* Avoid ldrd for user-only emulation, to handle unaligned. */ - if (USING_SOFTMMU && use_armv6_instructions - && (dl & 1) == 0 && dh == dl + 1) { - tcg_out_ldrd_r(s, COND_AL, dl, addrlo, addend); - } else if (dl != addend) { - tcg_out_ld32_rwb(s, COND_AL, dl, addend, addrlo); - tcg_out_ld32_12(s, COND_AL, dh, addend, 4); - } else { - tcg_out_dat_reg(s, COND_AL, ARITH_ADD, TCG_REG_TMP, - addend, addrlo, SHIFT_IMM_LSL(0)); - tcg_out_ld32_12(s, COND_AL, dl, TCG_REG_TMP, 0); - tcg_out_ld32_12(s, COND_AL, dh, TCG_REG_TMP, 4); - } - if (bswap) { - tcg_out_bswap32(s, COND_AL, dl, dl); - tcg_out_bswap32(s, COND_AL, dh, dh); - } + /* Avoid ldrd for user-only emulation, to handle unaligned. */ + if (USING_SOFTMMU && use_armv6_instructions + && (datalo & 1) == 0 && datahi == datalo + 1) { + tcg_out_ldrd_r(s, COND_AL, datalo, addrlo, addend); + } else if (datalo != addend) { + tcg_out_ld32_rwb(s, COND_AL, datalo, addend, addrlo); + tcg_out_ld32_12(s, COND_AL, datahi, addend, 4); + } else { + tcg_out_dat_reg(s, COND_AL, ARITH_ADD, TCG_REG_TMP, + addend, addrlo, SHIFT_IMM_LSL(0)); + tcg_out_ld32_12(s, COND_AL, datalo, TCG_REG_TMP, 0); + tcg_out_ld32_12(s, COND_AL, datahi, TCG_REG_TMP, 4); } break; + default: + g_assert_not_reached(); } } @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_qemu_ld_direct(TCGContext *s, MemOp opc, TCGReg datalo, TCGReg datahi, TCGReg addrlo) { - MemOp bswap = opc & MO_BSWAP; + /* Byte swapping is left to middle-end expansion. */ + tcg_debug_assert((opc & MO_BSWAP) == 0); switch (opc & MO_SSIZE) { case MO_UB: @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_qemu_ld_direct(TCGContext *s, MemOp opc, break; case MO_UW: tcg_out_ld16u_8(s, COND_AL, datalo, addrlo, 0); - if (bswap) { - tcg_out_bswap16(s, COND_AL, datalo, datalo, - TCG_BSWAP_IZ | TCG_BSWAP_OZ); - } break; case MO_SW: - if (bswap) { - tcg_out_ld16u_8(s, COND_AL, datalo, addrlo, 0); - tcg_out_bswap16(s, COND_AL, datalo, datalo, - TCG_BSWAP_IZ | TCG_BSWAP_OS); - } else { - tcg_out_ld16s_8(s, COND_AL, datalo, addrlo, 0); - } + tcg_out_ld16s_8(s, COND_AL, datalo, addrlo, 0); break; case MO_UL: - default: tcg_out_ld32_12(s, COND_AL, datalo, addrlo, 0); - if (bswap) { - tcg_out_bswap32(s, COND_AL, datalo, datalo); - } break; case MO_Q: - { - TCGReg dl = (bswap ? datahi : datalo); - TCGReg dh = (bswap ? datalo : datahi); - - /* Avoid ldrd for user-only emulation, to handle unaligned. */ - if (USING_SOFTMMU && use_armv6_instructions - && (dl & 1) == 0 && dh == dl + 1) { - tcg_out_ldrd_8(s, COND_AL, dl, addrlo, 0); - } else if (dl == addrlo) { - tcg_out_ld32_12(s, COND_AL, dh, addrlo, bswap ? 0 : 4); - tcg_out_ld32_12(s, COND_AL, dl, addrlo, bswap ? 4 : 0); - } else { - tcg_out_ld32_12(s, COND_AL, dl, addrlo, bswap ? 4 : 0); - tcg_out_ld32_12(s, COND_AL, dh, addrlo, bswap ? 0 : 4); - } - if (bswap) { - tcg_out_bswap32(s, COND_AL, dl, dl); - tcg_out_bswap32(s, COND_AL, dh, dh); - } + /* Avoid ldrd for user-only emulation, to handle unaligned. */ + if (USING_SOFTMMU && use_armv6_instructions + && (datalo & 1) == 0 && datahi == datalo + 1) { + tcg_out_ldrd_8(s, COND_AL, datalo, addrlo, 0); + } else if (datalo == addrlo) { + tcg_out_ld32_12(s, COND_AL, datahi, addrlo, 4); + tcg_out_ld32_12(s, COND_AL, datalo, addrlo, 0); + } else { + tcg_out_ld32_12(s, COND_AL, datalo, addrlo, 0); + tcg_out_ld32_12(s, COND_AL, datahi, addrlo, 4); } break; + default: + g_assert_not_reached(); } } @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_qemu_st_index(TCGContext *s, int cond, MemOp opc, TCGReg datalo, TCGReg datahi, TCGReg addrlo, TCGReg addend) { - MemOp bswap = opc & MO_BSWAP; + /* Byte swapping is left to middle-end expansion. */ + tcg_debug_assert((opc & MO_BSWAP) == 0); switch (opc & MO_SIZE) { case MO_8: tcg_out_st8_r(s, cond, datalo, addrlo, addend); break; case MO_16: - if (bswap) { - tcg_out_bswap16(s, cond, TCG_REG_R0, datalo, 0); - tcg_out_st16_r(s, cond, TCG_REG_R0, addrlo, addend); - } else { - tcg_out_st16_r(s, cond, datalo, addrlo, addend); - } + tcg_out_st16_r(s, cond, datalo, addrlo, addend); break; case MO_32: - default: - if (bswap) { - tcg_out_bswap32(s, cond, TCG_REG_R0, datalo); - tcg_out_st32_r(s, cond, TCG_REG_R0, addrlo, addend); - } else { - tcg_out_st32_r(s, cond, datalo, addrlo, addend); - } + tcg_out_st32_r(s, cond, datalo, addrlo, addend); break; case MO_64: /* Avoid strd for user-only emulation, to handle unaligned. */ - if (bswap) { - tcg_out_bswap32(s, cond, TCG_REG_R0, datahi); - tcg_out_st32_rwb(s, cond, TCG_REG_R0, addend, addrlo); - tcg_out_bswap32(s, cond, TCG_REG_R0, datalo); - tcg_out_st32_12(s, cond, TCG_REG_R0, addend, 4); - } else if (USING_SOFTMMU && use_armv6_instructions - && (datalo & 1) == 0 && datahi == datalo + 1) { + if (USING_SOFTMMU && use_armv6_instructions + && (datalo & 1) == 0 && datahi == datalo + 1) { tcg_out_strd_r(s, cond, datalo, addrlo, addend); } else { tcg_out_st32_rwb(s, cond, datalo, addend, addrlo); tcg_out_st32_12(s, cond, datahi, addend, 4); } break; + default: + g_assert_not_reached(); } } @@ -XXX,XX +XXX,XX @@ static inline void tcg_out_qemu_st_direct(TCGContext *s, MemOp opc, TCGReg datalo, TCGReg datahi, TCGReg addrlo) { - MemOp bswap = opc & MO_BSWAP; + /* Byte swapping is left to middle-end expansion. */ + tcg_debug_assert((opc & MO_BSWAP) == 0); switch (opc & MO_SIZE) { case MO_8: tcg_out_st8_12(s, COND_AL, datalo, addrlo, 0); break; case MO_16: - if (bswap) { - tcg_out_bswap16(s, COND_AL, TCG_REG_R0, datalo, 0); - tcg_out_st16_8(s, COND_AL, TCG_REG_R0, addrlo, 0); - } else { - tcg_out_st16_8(s, COND_AL, datalo, addrlo, 0); - } + tcg_out_st16_8(s, COND_AL, datalo, addrlo, 0); break; case MO_32: - default: - if (bswap) { - tcg_out_bswap32(s, COND_AL, TCG_REG_R0, datalo); - tcg_out_st32_12(s, COND_AL, TCG_REG_R0, addrlo, 0); - } else { - tcg_out_st32_12(s, COND_AL, datalo, addrlo, 0); - } + tcg_out_st32_12(s, COND_AL, datalo, addrlo, 0); break; case MO_64: /* Avoid strd for user-only emulation, to handle unaligned. */ - if (bswap) { - tcg_out_bswap32(s, COND_AL, TCG_REG_R0, datahi); - tcg_out_st32_12(s, COND_AL, TCG_REG_R0, addrlo, 0); - tcg_out_bswap32(s, COND_AL, TCG_REG_R0, datalo); - tcg_out_st32_12(s, COND_AL, TCG_REG_R0, addrlo, 4); - } else if (USING_SOFTMMU && use_armv6_instructions - && (datalo & 1) == 0 && datahi == datalo + 1) { + if (USING_SOFTMMU && use_armv6_instructions + && (datalo & 1) == 0 && datahi == datalo + 1) { tcg_out_strd_8(s, COND_AL, datalo, addrlo, 0); } else { tcg_out_st32_12(s, COND_AL, datalo, addrlo, 0); tcg_out_st32_12(s, COND_AL, datahi, addrlo, 4); } break; + default: + g_assert_not_reached(); } } -- 2.25.1 The memory bswap support in the aarch64 backend merely dates from a time when it was required. There is nothing special about the backend support that could not have been provided by the middle-end even prior to the introduction of the bswap flags. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/aarch64/tcg-target.h | 2 +- tcg/aarch64/tcg-target.c.inc | 87 +++++++++++++----------------------- 2 files changed, 32 insertions(+), 57 deletions(-) diff --git a/tcg/aarch64/tcg-target.h b/tcg/aarch64/tcg-target.h index XXXXXXX..XXXXXXX 100644 --- a/tcg/aarch64/tcg-target.h +++ b/tcg/aarch64/tcg-target.h @@ -XXX,XX +XXX,XX @@ typedef enum { #define TCG_TARGET_HAS_cmpsel_vec 0 #define TCG_TARGET_DEFAULT_MO (0) -#define TCG_TARGET_HAS_MEMORY_BSWAP 1 +#define TCG_TARGET_HAS_MEMORY_BSWAP 0 void tb_target_set_jmp_target(uintptr_t, uintptr_t, uintptr_t, uintptr_t); diff --git a/tcg/aarch64/tcg-target.c.inc b/tcg/aarch64/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/aarch64/tcg-target.c.inc +++ b/tcg/aarch64/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_cltz(TCGContext *s, TCGType ext, TCGReg d, /* helper signature: helper_ret_ld_mmu(CPUState *env, target_ulong addr, * TCGMemOpIdx oi, uintptr_t ra) */ -static void * const qemu_ld_helpers[16] = { - [MO_UB] = helper_ret_ldub_mmu, - [MO_LEUW] = helper_le_lduw_mmu, - [MO_LEUL] = helper_le_ldul_mmu, - [MO_LEQ] = helper_le_ldq_mmu, - [MO_BEUW] = helper_be_lduw_mmu, - [MO_BEUL] = helper_be_ldul_mmu, - [MO_BEQ] = helper_be_ldq_mmu, +static void * const qemu_ld_helpers[4] = { + [MO_8] = helper_ret_ldub_mmu, +#ifdef HOST_WORDS_BIGENDIAN + [MO_16] = helper_be_lduw_mmu, + [MO_32] = helper_be_ldul_mmu, + [MO_64] = helper_be_ldq_mmu, +#else + [MO_16] = helper_le_lduw_mmu, + [MO_32] = helper_le_ldul_mmu, + [MO_64] = helper_le_ldq_mmu, +#endif }; /* helper signature: helper_ret_st_mmu(CPUState *env, target_ulong addr, * uintxx_t val, TCGMemOpIdx oi, * uintptr_t ra) */ -static void * const qemu_st_helpers[16] = { - [MO_UB] = helper_ret_stb_mmu, - [MO_LEUW] = helper_le_stw_mmu, - [MO_LEUL] = helper_le_stl_mmu, - [MO_LEQ] = helper_le_stq_mmu, - [MO_BEUW] = helper_be_stw_mmu, - [MO_BEUL] = helper_be_stl_mmu, - [MO_BEQ] = helper_be_stq_mmu, +static void * const qemu_st_helpers[4] = { + [MO_8] = helper_ret_stb_mmu, +#ifdef HOST_WORDS_BIGENDIAN + [MO_16] = helper_be_stw_mmu, + [MO_32] = helper_be_stl_mmu, + [MO_64] = helper_be_stq_mmu, +#else + [MO_16] = helper_le_stw_mmu, + [MO_32] = helper_le_stl_mmu, + [MO_64] = helper_le_stq_mmu, +#endif }; static inline void tcg_out_adr(TCGContext *s, TCGReg rd, const void *target) @@ -XXX,XX +XXX,XX @@ static bool tcg_out_qemu_ld_slow_path(TCGContext *s, TCGLabelQemuLdst *lb) tcg_out_mov(s, TARGET_LONG_BITS == 64, TCG_REG_X1, lb->addrlo_reg); tcg_out_movi(s, TCG_TYPE_I32, TCG_REG_X2, oi); tcg_out_adr(s, TCG_REG_X3, lb->raddr); - tcg_out_call(s, qemu_ld_helpers[opc & (MO_BSWAP | MO_SIZE)]); + tcg_out_call(s, qemu_ld_helpers[opc & MO_SIZE]); if (opc & MO_SIGN) { tcg_out_sxt(s, lb->type, size, lb->datalo_reg, TCG_REG_X0); } else { @@ -XXX,XX +XXX,XX @@ static bool tcg_out_qemu_st_slow_path(TCGContext *s, TCGLabelQemuLdst *lb) tcg_out_mov(s, size == MO_64, TCG_REG_X2, lb->datalo_reg); tcg_out_movi(s, TCG_TYPE_I32, TCG_REG_X3, oi); tcg_out_adr(s, TCG_REG_X4, lb->raddr); - tcg_out_call(s, qemu_st_helpers[opc & (MO_BSWAP | MO_SIZE)]); + tcg_out_call(s, qemu_st_helpers[opc & MO_SIZE]); tcg_out_goto(s, lb->raddr); return true; } @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_ld_direct(TCGContext *s, MemOp memop, TCGType ext, TCGReg data_r, TCGReg addr_r, TCGType otype, TCGReg off_r) { - const MemOp bswap = memop & MO_BSWAP; + /* Byte swapping is left to middle-end expansion. */ + tcg_debug_assert((memop & MO_BSWAP) == 0); switch (memop & MO_SSIZE) { case MO_UB: @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_ld_direct(TCGContext *s, MemOp memop, TCGType ext, break; case MO_UW: tcg_out_ldst_r(s, I3312_LDRH, data_r, addr_r, otype, off_r); - if (bswap) { - tcg_out_rev(s, TCG_TYPE_I32, MO_16, data_r, data_r); - } break; case MO_SW: - if (bswap) { - tcg_out_ldst_r(s, I3312_LDRH, data_r, addr_r, otype, off_r); - tcg_out_rev(s, TCG_TYPE_I32, MO_16, data_r, data_r); - tcg_out_sxt(s, ext, MO_16, data_r, data_r); - } else { - tcg_out_ldst_r(s, (ext ? I3312_LDRSHX : I3312_LDRSHW), - data_r, addr_r, otype, off_r); - } + tcg_out_ldst_r(s, (ext ? I3312_LDRSHX : I3312_LDRSHW), + data_r, addr_r, otype, off_r); break; case MO_UL: tcg_out_ldst_r(s, I3312_LDRW, data_r, addr_r, otype, off_r); - if (bswap) { - tcg_out_rev(s, TCG_TYPE_I32, MO_32, data_r, data_r); - } break; case MO_SL: - if (bswap) { - tcg_out_ldst_r(s, I3312_LDRW, data_r, addr_r, otype, off_r); - tcg_out_rev(s, TCG_TYPE_I32, MO_32, data_r, data_r); - tcg_out_sxt(s, TCG_TYPE_I64, MO_32, data_r, data_r); - } else { - tcg_out_ldst_r(s, I3312_LDRSWX, data_r, addr_r, otype, off_r); - } + tcg_out_ldst_r(s, I3312_LDRSWX, data_r, addr_r, otype, off_r); break; case MO_Q: tcg_out_ldst_r(s, I3312_LDRX, data_r, addr_r, otype, off_r); - if (bswap) { - tcg_out_rev(s, TCG_TYPE_I64, MO_64, data_r, data_r); - } break; default: tcg_abort(); @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_st_direct(TCGContext *s, MemOp memop, TCGReg data_r, TCGReg addr_r, TCGType otype, TCGReg off_r) { - const MemOp bswap = memop & MO_BSWAP; + /* Byte swapping is left to middle-end expansion. */ + tcg_debug_assert((memop & MO_BSWAP) == 0); switch (memop & MO_SIZE) { case MO_8: tcg_out_ldst_r(s, I3312_STRB, data_r, addr_r, otype, off_r); break; case MO_16: - if (bswap && data_r != TCG_REG_XZR) { - tcg_out_rev(s, TCG_TYPE_I32, MO_16, TCG_REG_TMP, data_r); - data_r = TCG_REG_TMP; - } tcg_out_ldst_r(s, I3312_STRH, data_r, addr_r, otype, off_r); break; case MO_32: - if (bswap && data_r != TCG_REG_XZR) { - tcg_out_rev(s, TCG_TYPE_I32, MO_32, TCG_REG_TMP, data_r); - data_r = TCG_REG_TMP; - } tcg_out_ldst_r(s, I3312_STRW, data_r, addr_r, otype, off_r); break; case MO_64: - if (bswap && data_r != TCG_REG_XZR) { - tcg_out_rev(s, TCG_TYPE_I64, MO_64, TCG_REG_TMP, data_r); - data_r = TCG_REG_TMP; - } tcg_out_ldst_r(s, I3312_STRX, data_r, addr_r, otype, off_r); break; default: -- 2.25.1 TCG_TARGET_HAS_MEMORY_BSWAP is already unset for this backend, which means that MO_BSWAP be handled by the middle-end and will never be seen by the backend. Thus the indexes used with qemu_{ld,st}_helpers will always be zero. Tidy the comments and asserts in tcg_out_qemu_{ld,st}_direct. It is not that we do not handle bswap "yet", but never will. Acked-by: Alistair Francis <alistair.francis@wdc.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/riscv/tcg-target.c.inc | 64 ++++++++++++++++++++------------------ 1 file changed, 33 insertions(+), 31 deletions(-) diff --git a/tcg/riscv/tcg-target.c.inc b/tcg/riscv/tcg-target.c.inc index XXXXXXX..XXXXXXX 100644 --- a/tcg/riscv/tcg-target.c.inc +++ b/tcg/riscv/tcg-target.c.inc @@ -XXX,XX +XXX,XX @@ static void tcg_out_mb(TCGContext *s, TCGArg a0) /* helper signature: helper_ret_ld_mmu(CPUState *env, target_ulong addr, * TCGMemOpIdx oi, uintptr_t ra) */ -static void * const qemu_ld_helpers[16] = { - [MO_UB] = helper_ret_ldub_mmu, - [MO_SB] = helper_ret_ldsb_mmu, - [MO_LEUW] = helper_le_lduw_mmu, - [MO_LESW] = helper_le_ldsw_mmu, - [MO_LEUL] = helper_le_ldul_mmu, +static void * const qemu_ld_helpers[8] = { + [MO_UB] = helper_ret_ldub_mmu, + [MO_SB] = helper_ret_ldsb_mmu, +#ifdef HOST_WORDS_BIGENDIAN + [MO_UW] = helper_be_lduw_mmu, + [MO_SW] = helper_be_ldsw_mmu, + [MO_UL] = helper_be_ldul_mmu, #if TCG_TARGET_REG_BITS == 64 - [MO_LESL] = helper_le_ldsl_mmu, + [MO_SL] = helper_be_ldsl_mmu, #endif - [MO_LEQ] = helper_le_ldq_mmu, - [MO_BEUW] = helper_be_lduw_mmu, - [MO_BESW] = helper_be_ldsw_mmu, - [MO_BEUL] = helper_be_ldul_mmu, + [MO_Q] = helper_be_ldq_mmu, +#else + [MO_UW] = helper_le_lduw_mmu, + [MO_SW] = helper_le_ldsw_mmu, + [MO_UL] = helper_le_ldul_mmu, #if TCG_TARGET_REG_BITS == 64 - [MO_BESL] = helper_be_ldsl_mmu, + [MO_SL] = helper_le_ldsl_mmu, +#endif + [MO_Q] = helper_le_ldq_mmu, #endif - [MO_BEQ] = helper_be_ldq_mmu, }; /* helper signature: helper_ret_st_mmu(CPUState *env, target_ulong addr, * uintxx_t val, TCGMemOpIdx oi, * uintptr_t ra) */ -static void * const qemu_st_helpers[16] = { - [MO_UB] = helper_ret_stb_mmu, - [MO_LEUW] = helper_le_stw_mmu, - [MO_LEUL] = helper_le_stl_mmu, - [MO_LEQ] = helper_le_stq_mmu, - [MO_BEUW] = helper_be_stw_mmu, - [MO_BEUL] = helper_be_stl_mmu, - [MO_BEQ] = helper_be_stq_mmu, +static void * const qemu_st_helpers[4] = { + [MO_8] = helper_ret_stb_mmu, +#ifdef HOST_WORDS_BIGENDIAN + [MO_16] = helper_be_stw_mmu, + [MO_32] = helper_be_stl_mmu, + [MO_64] = helper_be_stq_mmu, +#else + [MO_16] = helper_le_stw_mmu, + [MO_32] = helper_le_stl_mmu, + [MO_64] = helper_le_stq_mmu, +#endif }; /* We don't support oversize guests */ @@ -XXX,XX +XXX,XX @@ static bool tcg_out_qemu_ld_slow_path(TCGContext *s, TCGLabelQemuLdst *l) tcg_out_movi(s, TCG_TYPE_PTR, a2, oi); tcg_out_movi(s, TCG_TYPE_PTR, a3, (tcg_target_long)l->raddr); - tcg_out_call(s, qemu_ld_helpers[opc & (MO_BSWAP | MO_SSIZE)]); + tcg_out_call(s, qemu_ld_helpers[opc & MO_SSIZE]); tcg_out_mov(s, (opc & MO_SIZE) == MO_64, l->datalo_reg, a0); tcg_out_goto(s, l->raddr); @@ -XXX,XX +XXX,XX @@ static bool tcg_out_qemu_st_slow_path(TCGContext *s, TCGLabelQemuLdst *l) tcg_out_movi(s, TCG_TYPE_PTR, a3, oi); tcg_out_movi(s, TCG_TYPE_PTR, a4, (tcg_target_long)l->raddr); - tcg_out_call(s, qemu_st_helpers[opc & (MO_BSWAP | MO_SSIZE)]); + tcg_out_call(s, qemu_st_helpers[opc & MO_SIZE]); tcg_out_goto(s, l->raddr); return true; @@ -XXX,XX +XXX,XX @@ static bool tcg_out_qemu_st_slow_path(TCGContext *s, TCGLabelQemuLdst *l) static void tcg_out_qemu_ld_direct(TCGContext *s, TCGReg lo, TCGReg hi, TCGReg base, MemOp opc, bool is_64) { - const MemOp bswap = opc & MO_BSWAP; - - /* We don't yet handle byteswapping, assert */ - g_assert(!bswap); + /* Byte swapping is left to middle-end expansion. */ + tcg_debug_assert((opc & MO_BSWAP) == 0); switch (opc & (MO_SSIZE)) { case MO_UB: @@ -XXX,XX +XXX,XX @@ static void tcg_out_qemu_ld(TCGContext *s, const TCGArg *args, bool is_64) static void tcg_out_qemu_st_direct(TCGContext *s, TCGReg lo, TCGReg hi, TCGReg base, MemOp opc) { - const MemOp bswap = opc & MO_BSWAP; - - /* We don't yet handle byteswapping, assert */ - g_assert(!bswap); + /* Byte swapping is left to middle-end expansion. */ + tcg_debug_assert((opc & MO_BSWAP) == 0); switch (opc & (MO_SSIZE)) { case MO_8: -- 2.25.1 The following changes since commit aa3a285b5bc56a4208b3b57d4a55291e9c260107: Merge tag 'mem-2024-12-21' of https://github.com/davidhildenbrand/qemu into staging (2024-12-22 14:33:27 -0500) are available in the Git repository at: https://gitlab.com/rth7680/qemu.git tags/pull-tcg-20241224 for you to fetch changes up to e4a8e093dc74be049f4829831dce76e5edab0003: accel/tcg: Move gen_intermediate_code to TCGCPUOps.translate_core (2024-12-24 08:32:15 -0800) ---------------------------------------------------------------- tcg/optimize: Remove in-flight mask data from OptContext fpu: Add float*_muladd_scalbn fpu: Remove float_muladd_halve_result fpu: Add float_round_nearest_even_max fpu: Add float_muladd_suppress_add_product_zero target/hexagon: Use float32_muladd accel/tcg: Move gen_intermediate_code to TCGCPUOps.translate_core ---------------------------------------------------------------- Ilya Leoshkevich (1): tests/tcg: Do not use inttypes.h in multiarch/system/memory.c Pierrick Bouvier (1): plugins: optimize cpu_index code generation Richard Henderson (70): tcg/optimize: Split out finish_bb, finish_ebb tcg/optimize: Split out fold_affected_mask tcg/optimize: Copy mask writeback to fold_masks tcg/optimize: Split out fold_masks_zs tcg/optimize: Augment s_mask from z_mask in fold_masks_zs tcg/optimize: Change representation of s_mask tcg/optimize: Use finish_folding in fold_add, fold_add_vec, fold_addsub2 tcg/optimize: Introduce const value accessors for TempOptInfo tcg/optimize: Use fold_masks_zs in fold_and tcg/optimize: Use fold_masks_zs in fold_andc tcg/optimize: Use fold_masks_zs in fold_bswap tcg/optimize: Use fold_masks_zs in fold_count_zeros tcg/optimize: Use fold_masks_z in fold_ctpop tcg/optimize: Use fold_and and fold_masks_z in fold_deposit tcg/optimize: Compute sign mask in fold_deposit tcg/optimize: Use finish_folding in fold_divide tcg/optimize: Use finish_folding in fold_dup, fold_dup2 tcg/optimize: Use fold_masks_s in fold_eqv tcg/optimize: Use fold_masks_z in fold_extract tcg/optimize: Use finish_folding in fold_extract2 tcg/optimize: Use fold_masks_zs in fold_exts tcg/optimize: Use fold_masks_z in fold_extu tcg/optimize: Use fold_masks_zs in fold_movcond tcg/optimize: Use finish_folding in fold_mul* tcg/optimize: Use fold_masks_s in fold_nand tcg/optimize: Use fold_masks_z in fold_neg_no_const tcg/optimize: Use fold_masks_s in fold_nor tcg/optimize: Use fold_masks_s in fold_not tcg/optimize: Use fold_masks_zs in fold_or tcg/optimize: Use fold_masks_zs in fold_orc tcg/optimize: Use fold_masks_zs in fold_qemu_ld tcg/optimize: Return true from fold_qemu_st, fold_tcg_st tcg/optimize: Use finish_folding in fold_remainder tcg/optimize: Distinguish simplification in fold_setcond_zmask tcg/optimize: Use fold_masks_z in fold_setcond tcg/optimize: Use fold_masks_s in fold_negsetcond tcg/optimize: Use fold_masks_z in fold_setcond2 tcg/optimize: Use finish_folding in fold_cmp_vec tcg/optimize: Use finish_folding in fold_cmpsel_vec tcg/optimize: Use fold_masks_zs in fold_sextract tcg/optimize: Use fold_masks_zs, fold_masks_s in fold_shift tcg/optimize: Simplify sign bit test in fold_shift tcg/optimize: Use finish_folding in fold_sub, fold_sub_vec tcg/optimize: Use fold_masks_zs in fold_tcg_ld tcg/optimize: Use finish_folding in fold_tcg_ld_memcopy tcg/optimize: Use fold_masks_zs in fold_xor tcg/optimize: Use finish_folding in fold_bitsel_vec tcg/optimize: Use finish_folding as default in tcg_optimize tcg/optimize: Remove z_mask, s_mask from OptContext tcg/optimize: Re-enable sign-mask optimizations tcg/optimize: Move fold_bitsel_vec into alphabetic sort tcg/optimize: Move fold_cmp_vec, fold_cmpsel_vec into alphabetic sort softfloat: Add float{16,32,64}_muladd_scalbn target/arm: Use float*_muladd_scalbn target/sparc: Use float*_muladd_scalbn softfloat: Remove float_muladd_halve_result softfloat: Add float_round_nearest_even_max softfloat: Add float_muladd_suppress_add_product_zero target/hexagon: Use float32_mul in helper_sfmpy target/hexagon: Use float32_muladd for helper_sffma target/hexagon: Use float32_muladd for helper_sffms target/hexagon: Use float32_muladd_scalbn for helper_sffma_sc target/hexagon: Use float32_muladd for helper_sffm[as]_lib target/hexagon: Remove internal_fmafx target/hexagon: Expand GEN_XF_ROUND target/hexagon: Remove Float target/hexagon: Remove Double target/hexagon: Use mulu64 for int128_mul_6464 target/hexagon: Simplify internal_mpyhh setup accel/tcg: Move gen_intermediate_code to TCGCPUOps.translate_core include/exec/translator.h | 14 - include/fpu/softfloat-types.h | 2 + include/fpu/softfloat.h | 14 +- include/hw/core/tcg-cpu-ops.h | 13 + target/alpha/cpu.h | 2 + target/arm/internals.h | 2 + target/avr/cpu.h | 2 + target/hexagon/cpu.h | 2 + target/hexagon/fma_emu.h | 3 - target/hppa/cpu.h | 2 + target/i386/tcg/helper-tcg.h | 2 + target/loongarch/internals.h | 2 + target/m68k/cpu.h | 2 + target/microblaze/cpu.h | 2 + target/mips/tcg/tcg-internal.h | 2 + target/openrisc/cpu.h | 2 + target/ppc/cpu.h | 2 + target/riscv/cpu.h | 3 + target/rx/cpu.h | 2 + target/s390x/s390x-internal.h | 2 + target/sh4/cpu.h | 2 + target/sparc/cpu.h | 2 + target/sparc/helper.h | 4 +- target/tricore/cpu.h | 2 + target/xtensa/cpu.h | 2 + accel/tcg/cpu-exec.c | 8 +- accel/tcg/plugin-gen.c | 9 + accel/tcg/translate-all.c | 8 +- fpu/softfloat.c | 63 +-- target/alpha/cpu.c | 1 + target/alpha/translate.c | 4 +- target/arm/cpu.c | 1 + target/arm/tcg/cpu-v7m.c | 1 + target/arm/tcg/helper-a64.c | 6 +- target/arm/tcg/translate.c | 5 +- target/avr/cpu.c | 1 + target/avr/translate.c | 6 +- target/hexagon/cpu.c | 1 + target/hexagon/fma_emu.c | 496 ++++++--------------- target/hexagon/op_helper.c | 125 ++---- target/hexagon/translate.c | 4 +- target/hppa/cpu.c | 1 + target/hppa/translate.c | 4 +- target/i386/tcg/tcg-cpu.c | 1 + target/i386/tcg/translate.c | 5 +- target/loongarch/cpu.c | 1 + target/loongarch/tcg/translate.c | 4 +- target/m68k/cpu.c | 1 + target/m68k/translate.c | 4 +- target/microblaze/cpu.c | 1 + target/microblaze/translate.c | 4 +- target/mips/cpu.c | 1 + target/mips/tcg/translate.c | 4 +- target/openrisc/cpu.c | 1 + target/openrisc/translate.c | 4 +- target/ppc/cpu_init.c | 1 + target/ppc/translate.c | 4 +- target/riscv/tcg/tcg-cpu.c | 1 + target/riscv/translate.c | 4 +- target/rx/cpu.c | 1 + target/rx/translate.c | 4 +- target/s390x/cpu.c | 1 + target/s390x/tcg/translate.c | 4 +- target/sh4/cpu.c | 1 + target/sh4/translate.c | 4 +- target/sparc/cpu.c | 1 + target/sparc/fop_helper.c | 8 +- target/sparc/translate.c | 84 ++-- target/tricore/cpu.c | 1 + target/tricore/translate.c | 5 +- target/xtensa/cpu.c | 1 + target/xtensa/translate.c | 4 +- tcg/optimize.c | 857 +++++++++++++++++++----------------- tests/tcg/multiarch/system/memory.c | 9 +- fpu/softfloat-parts.c.inc | 16 +- 75 files changed, 866 insertions(+), 1009 deletions(-) From: Ilya Leoshkevich <iii@linux.ibm.com> make check-tcg fails on Fedora with the following error message: alpha-linux-gnu-gcc [...] qemu/tests/tcg/multiarch/system/memory.c -o memory [...] qemu/tests/tcg/multiarch/system/memory.c:17:10: fatal error: inttypes.h: No such file or directory 17 | #include <inttypes.h> | ^~~~~~~~~~~~ compilation terminated. The reason is that Fedora has cross-compilers, but no cross-glibc headers. Fix by hardcoding the format specifiers and dropping the include. An alternative fix would be to introduce a configure check for inttypes.h. But this would make it impossible to use Fedora cross-compilers for softmmu tests, which used to work so far. Fixes: ecbcc9ead2f8 ("tests/tcg: add a system test to check memory instrumentation") Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-ID: <20241010085906.226249-1-iii@linux.ibm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tests/tcg/multiarch/system/memory.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) diff --git a/tests/tcg/multiarch/system/memory.c b/tests/tcg/multiarch/system/memory.c index XXXXXXX..XXXXXXX 100644 --- a/tests/tcg/multiarch/system/memory.c +++ b/tests/tcg/multiarch/system/memory.c @@ -XXX,XX +XXX,XX @@ #include <stdint.h> #include <stdbool.h> -#include <inttypes.h> #include <minilib.h> #ifndef CHECK_UNALIGNED @@ -XXX,XX +XXX,XX @@ int main(void) int i; bool ok = true; - ml_printf("Test data start: 0x%"PRIxPTR"\n", &test_data[0]); - ml_printf("Test data end: 0x%"PRIxPTR"\n", &test_data[TEST_SIZE]); + ml_printf("Test data start: 0x%lx\n", (unsigned long)&test_data[0]); + ml_printf("Test data end: 0x%lx\n", (unsigned long)&test_data[TEST_SIZE]); /* Run through the unsigned tests first */ for (i = 0; i < ARRAY_SIZE(init_ufns) && ok; i++) { @@ -XXX,XX +XXX,XX @@ int main(void) ok = do_signed_reads(true); } - ml_printf("Test data read: %"PRId32"\n", test_read_count); - ml_printf("Test data write: %"PRId32"\n", test_write_count); + ml_printf("Test data read: %lu\n", (unsigned long)test_read_count); + ml_printf("Test data write: %lu\n", (unsigned long)test_write_count); ml_printf("Test complete: %s\n", ok ? "PASSED" : "FAILED"); return ok ? 0 : -1; } -- 2.43.0 From: Pierrick Bouvier <pierrick.bouvier@linaro.org> When running with a single vcpu, we can return a constant instead of a load when accessing cpu_index. A side effect is that all tcg operations using it are optimized, most notably scoreboard access. When running a simple loop in user-mode, the speedup is around 20%. Signed-off-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-ID: <20241128213843.1023080-1-pierrick.bouvier@linaro.org> --- accel/tcg/plugin-gen.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/accel/tcg/plugin-gen.c b/accel/tcg/plugin-gen.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/plugin-gen.c +++ b/accel/tcg/plugin-gen.c @@ -XXX,XX +XXX,XX @@ static void gen_disable_mem_helper(void) static TCGv_i32 gen_cpu_index(void) { + /* + * Optimize when we run with a single vcpu. All values using cpu_index, + * including scoreboard index, will be optimized out. + * User-mode calls tb_flush when setting this flag. In system-mode, all + * vcpus are created before generating code. + */ + if (!tcg_cflags_has(current_cpu, CF_PARALLEL)) { + return tcg_constant_i32(current_cpu->cpu_index); + } TCGv_i32 cpu_index = tcg_temp_ebb_new_i32(); tcg_gen_ld_i32(cpu_index, tcg_env, -offsetof(ArchCPU, env) + offsetof(CPUState, cpu_index)); -- 2.43.0 Call them directly from the opcode switch statement in tcg_optimize, rather than in finish_folding based on opcode flags. Adjust folding of conditional branches to match. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 47 +++++++++++++++++++++++++++++++---------------- 1 file changed, 31 insertions(+), 16 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static void copy_propagate(OptContext *ctx, TCGOp *op, } } +static void finish_bb(OptContext *ctx) +{ + /* We only optimize memory barriers across basic blocks. */ + ctx->prev_mb = NULL; +} + +static void finish_ebb(OptContext *ctx) +{ + finish_bb(ctx); + /* We only optimize across extended basic blocks. */ + memset(&ctx->temps_used, 0, sizeof(ctx->temps_used)); + remove_mem_copy_all(ctx); +} + static void finish_folding(OptContext *ctx, TCGOp *op) { const TCGOpDef *def = &tcg_op_defs[op->opc]; int i, nb_oargs; - /* - * We only optimize extended basic blocks. If the opcode ends a BB - * and is not a conditional branch, reset all temp data. - */ - if (def->flags & TCG_OPF_BB_END) { - ctx->prev_mb = NULL; - if (!(def->flags & TCG_OPF_COND_BRANCH)) { - memset(&ctx->temps_used, 0, sizeof(ctx->temps_used)); - remove_mem_copy_all(ctx); - } - return; - } - nb_oargs = def->nb_oargs; for (i = 0; i < nb_oargs; i++) { TCGTemp *ts = arg_temp(op->args[i]); @@ -XXX,XX +XXX,XX @@ static bool fold_brcond(OptContext *ctx, TCGOp *op) if (i > 0) { op->opc = INDEX_op_br; op->args[0] = op->args[3]; + finish_ebb(ctx); + } else { + finish_bb(ctx); } - return false; + return true; } static bool fold_brcond2(OptContext *ctx, TCGOp *op) @@ -XXX,XX +XXX,XX @@ static bool fold_brcond2(OptContext *ctx, TCGOp *op) } op->opc = INDEX_op_br; op->args[0] = label; - break; + finish_ebb(ctx); + return true; } - return false; + + finish_bb(ctx); + return true; } static bool fold_bswap(OptContext *ctx, TCGOp *op) @@ -XXX,XX +XXX,XX @@ void tcg_optimize(TCGContext *s) CASE_OP_32_64_VEC(xor): done = fold_xor(&ctx, op); break; + case INDEX_op_set_label: + case INDEX_op_br: + case INDEX_op_exit_tb: + case INDEX_op_goto_tb: + case INDEX_op_goto_ptr: + finish_ebb(&ctx); + done = true; + break; default: break; } -- 2.43.0 There are only a few logical operations which can compute an "affected" mask. Split out handling of this optimization to a separate function, only to be called when applicable. Remove the a_mask field from OptContext, as the mask is no longer stored anywhere. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 42 +++++++++++++++++++++++++++--------------- 1 file changed, 27 insertions(+), 15 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ typedef struct OptContext { QSIMPLEQ_HEAD(, MemCopyInfo) mem_free; /* In flight values from optimization. */ - uint64_t a_mask; /* mask bit is 0 iff value identical to first input */ uint64_t z_mask; /* mask bit is 0 iff value bit is 0 */ uint64_t s_mask; /* mask of clrsb(value) bits */ TCGType type; @@ -XXX,XX +XXX,XX @@ static bool fold_const2_commutative(OptContext *ctx, TCGOp *op) static bool fold_masks(OptContext *ctx, TCGOp *op) { - uint64_t a_mask = ctx->a_mask; uint64_t z_mask = ctx->z_mask; uint64_t s_mask = ctx->s_mask; @@ -XXX,XX +XXX,XX @@ static bool fold_masks(OptContext *ctx, TCGOp *op) * type changing opcodes. */ if (ctx->type == TCG_TYPE_I32) { - a_mask = (int32_t)a_mask; z_mask = (int32_t)z_mask; s_mask |= MAKE_64BIT_MASK(32, 32); ctx->z_mask = z_mask; @@ -XXX,XX +XXX,XX @@ static bool fold_masks(OptContext *ctx, TCGOp *op) if (z_mask == 0) { return tcg_opt_gen_movi(ctx, op, op->args[0], 0); } + return false; +} + +/* + * An "affected" mask bit is 0 if and only if the result is identical + * to the first input. Thus if the entire mask is 0, the operation + * is equivalent to a copy. + */ +static bool fold_affected_mask(OptContext *ctx, TCGOp *op, uint64_t a_mask) +{ + if (ctx->type == TCG_TYPE_I32) { + a_mask = (uint32_t)a_mask; + } if (a_mask == 0) { return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[1]); } @@ -XXX,XX +XXX,XX @@ static bool fold_and(OptContext *ctx, TCGOp *op) * Known-zeros does not imply known-ones. Therefore unless * arg2 is constant, we can't infer affected bits from it. */ - if (arg_is_const(op->args[2])) { - ctx->a_mask = z1 & ~z2; + if (arg_is_const(op->args[2]) && + fold_affected_mask(ctx, op, z1 & ~z2)) { + return true; } return fold_masks(ctx, op); @@ -XXX,XX +XXX,XX @@ static bool fold_andc(OptContext *ctx, TCGOp *op) */ if (arg_is_const(op->args[2])) { uint64_t z2 = ~arg_info(op->args[2])->z_mask; - ctx->a_mask = z1 & ~z2; + if (fold_affected_mask(ctx, op, z1 & ~z2)) { + return true; + } z1 &= z2; } ctx->z_mask = z1; @@ -XXX,XX +XXX,XX @@ static bool fold_extract(OptContext *ctx, TCGOp *op) z_mask_old = arg_info(op->args[1])->z_mask; z_mask = extract64(z_mask_old, pos, len); - if (pos == 0) { - ctx->a_mask = z_mask_old ^ z_mask; + if (pos == 0 && fold_affected_mask(ctx, op, z_mask_old ^ z_mask)) { + return true; } ctx->z_mask = z_mask; ctx->s_mask = smask_from_zmask(z_mask); @@ -XXX,XX +XXX,XX @@ static bool fold_exts(OptContext *ctx, TCGOp *op) ctx->z_mask = z_mask; ctx->s_mask = s_mask; - if (!type_change) { - ctx->a_mask = s_mask & ~s_mask_old; + if (!type_change && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { + return true; } return fold_masks(ctx, op); @@ -XXX,XX +XXX,XX @@ static bool fold_extu(OptContext *ctx, TCGOp *op) ctx->z_mask = z_mask; ctx->s_mask = smask_from_zmask(z_mask); - if (!type_change) { - ctx->a_mask = z_mask_old ^ z_mask; + if (!type_change && fold_affected_mask(ctx, op, z_mask_old ^ z_mask)) { + return true; } return fold_masks(ctx, op); } @@ -XXX,XX +XXX,XX @@ static bool fold_sextract(OptContext *ctx, TCGOp *op) s_mask |= MAKE_64BIT_MASK(len, 64 - len); ctx->s_mask = s_mask; - if (pos == 0) { - ctx->a_mask = s_mask & ~s_mask_old; + if (pos == 0 && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { + return true; } return fold_masks(ctx, op); @@ -XXX,XX +XXX,XX @@ void tcg_optimize(TCGContext *s) } /* Assume all bits affected, no bits known zero, no sign reps. */ - ctx.a_mask = -1; ctx.z_mask = -1; ctx.s_mask = 0; -- 2.43.0 Use of fold_masks should be restricted to those opcodes that can reliably make use of it -- those with a single output, and from higher-level folders that set up the masks. Prepare for conversion of each folder in turn. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_masks(OptContext *ctx, TCGOp *op) { uint64_t z_mask = ctx->z_mask; uint64_t s_mask = ctx->s_mask; + const TCGOpDef *def = &tcg_op_defs[op->opc]; + TCGTemp *ts; + TempOptInfo *ti; + + /* Only single-output opcodes are supported here. */ + tcg_debug_assert(def->nb_oargs == 1); /* * 32-bit ops generate 32-bit results, which for the purpose of @@ -XXX,XX +XXX,XX @@ static bool fold_masks(OptContext *ctx, TCGOp *op) if (ctx->type == TCG_TYPE_I32) { z_mask = (int32_t)z_mask; s_mask |= MAKE_64BIT_MASK(32, 32); - ctx->z_mask = z_mask; - ctx->s_mask = s_mask; } if (z_mask == 0) { return tcg_opt_gen_movi(ctx, op, op->args[0], 0); } - return false; + + ts = arg_temp(op->args[0]); + reset_ts(ctx, ts); + + ti = ts_info(ts); + ti->z_mask = z_mask; + ti->s_mask = s_mask; + return true; } /* -- 2.43.0 Add a routine to which masks can be passed directly, rather than storing them into OptContext. To be used in upcoming patches. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_const2_commutative(OptContext *ctx, TCGOp *op) return fold_const2(ctx, op); } -static bool fold_masks(OptContext *ctx, TCGOp *op) +/* + * Record "zero" and "sign" masks for the single output of @op. + * See TempOptInfo definition of z_mask and s_mask. + * If z_mask allows, fold the output to constant zero. + */ +static bool fold_masks_zs(OptContext *ctx, TCGOp *op, + uint64_t z_mask, uint64_t s_mask) { - uint64_t z_mask = ctx->z_mask; - uint64_t s_mask = ctx->s_mask; const TCGOpDef *def = &tcg_op_defs[op->opc]; TCGTemp *ts; TempOptInfo *ti; @@ -XXX,XX +XXX,XX @@ static bool fold_masks(OptContext *ctx, TCGOp *op) return true; } +static bool fold_masks(OptContext *ctx, TCGOp *op) +{ + return fold_masks_zs(ctx, op, ctx->z_mask, ctx->s_mask); +} + /* * An "affected" mask bit is 0 if and only if the result is identical * to the first input. Thus if the entire mask is 0, the operation -- 2.43.0 Consider the passed s_mask to be a minimum deduced from either existing s_mask or from a sign-extension operation. We may be able to deduce more from the set of known zeros. Remove identical logic from several opcode folders. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 21 ++++++--------------- 1 file changed, 6 insertions(+), 15 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_const2_commutative(OptContext *ctx, TCGOp *op) * Record "zero" and "sign" masks for the single output of @op. * See TempOptInfo definition of z_mask and s_mask. * If z_mask allows, fold the output to constant zero. + * The passed s_mask may be augmented by z_mask. */ static bool fold_masks_zs(OptContext *ctx, TCGOp *op, uint64_t z_mask, uint64_t s_mask) @@ -XXX,XX +XXX,XX @@ static bool fold_masks_zs(OptContext *ctx, TCGOp *op, ti = ts_info(ts); ti->z_mask = z_mask; - ti->s_mask = s_mask; + ti->s_mask = s_mask | smask_from_zmask(z_mask); return true; } @@ -XXX,XX +XXX,XX @@ static bool fold_bswap(OptContext *ctx, TCGOp *op) default: g_assert_not_reached(); } - s_mask = smask_from_zmask(z_mask); + s_mask = 0; switch (op->args[2] & (TCG_BSWAP_OZ | TCG_BSWAP_OS)) { case TCG_BSWAP_OZ: break; @@ -XXX,XX +XXX,XX @@ static bool fold_bswap(OptContext *ctx, TCGOp *op) default: /* The high bits are undefined: force all bits above the sign to 1. */ z_mask |= sign << 1; - s_mask = 0; break; } ctx->z_mask = z_mask; @@ -XXX,XX +XXX,XX @@ static bool fold_count_zeros(OptContext *ctx, TCGOp *op) g_assert_not_reached(); } ctx->z_mask = arg_info(op->args[2])->z_mask | z_mask; - ctx->s_mask = smask_from_zmask(ctx->z_mask); return false; } @@ -XXX,XX +XXX,XX @@ static bool fold_ctpop(OptContext *ctx, TCGOp *op) default: g_assert_not_reached(); } - ctx->s_mask = smask_from_zmask(ctx->z_mask); return false; } @@ -XXX,XX +XXX,XX @@ static bool fold_extract(OptContext *ctx, TCGOp *op) return true; } ctx->z_mask = z_mask; - ctx->s_mask = smask_from_zmask(z_mask); return fold_masks(ctx, op); } @@ -XXX,XX +XXX,XX @@ static bool fold_extu(OptContext *ctx, TCGOp *op) } ctx->z_mask = z_mask; - ctx->s_mask = smask_from_zmask(z_mask); if (!type_change && fold_affected_mask(ctx, op, z_mask_old ^ z_mask)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool fold_qemu_ld(OptContext *ctx, TCGOp *op) int width = 8 * memop_size(mop); if (width < 64) { - ctx->s_mask = MAKE_64BIT_MASK(width, 64 - width); - if (!(mop & MO_SIGN)) { + if (mop & MO_SIGN) { + ctx->s_mask = MAKE_64BIT_MASK(width, 64 - width); + } else { ctx->z_mask = MAKE_64BIT_MASK(0, width); - ctx->s_mask <<= 1; } } @@ -XXX,XX +XXX,XX @@ static bool fold_setcond(OptContext *ctx, TCGOp *op) fold_setcond_tst_pow2(ctx, op, false); ctx->z_mask = 1; - ctx->s_mask = smask_from_zmask(1); return false; } @@ -XXX,XX +XXX,XX @@ static bool fold_setcond2(OptContext *ctx, TCGOp *op) } ctx->z_mask = 1; - ctx->s_mask = smask_from_zmask(1); return false; do_setcond_const: @@ -XXX,XX +XXX,XX @@ static bool fold_tcg_ld(OptContext *ctx, TCGOp *op) break; CASE_OP_32_64(ld8u): ctx->z_mask = MAKE_64BIT_MASK(0, 8); - ctx->s_mask = MAKE_64BIT_MASK(9, 55); break; CASE_OP_32_64(ld16s): ctx->s_mask = MAKE_64BIT_MASK(16, 48); break; CASE_OP_32_64(ld16u): ctx->z_mask = MAKE_64BIT_MASK(0, 16); - ctx->s_mask = MAKE_64BIT_MASK(17, 47); break; case INDEX_op_ld32s_i64: ctx->s_mask = MAKE_64BIT_MASK(32, 32); break; case INDEX_op_ld32u_i64: ctx->z_mask = MAKE_64BIT_MASK(0, 32); - ctx->s_mask = MAKE_64BIT_MASK(33, 31); break; default: g_assert_not_reached(); -- 2.43.0 Change the representation from sign bit repetitions to all bits equal to the sign bit, including the sign bit itself. The previous format has a problem in that it is difficult to recreate a valid sign mask after a shift operation: the "repetitions" part of the previous format meant that applying the same shift as for the value lead to an off-by-one value. The new format, including the sign bit itself, means that the sign mask can be manipulated in exactly the same way as the value, canonicalization is easier. Canonicalize the s_mask in fold_masks_zs, rather than requiring callers to do so. Treat 0 as a non-canonical but typeless input for no sign information, which will be reset as appropriate for the data type. We can easily fold in the data from z_mask while canonicalizing. Temporarily disable optimizations using s_mask while each operation is converted to use fold_masks_zs and to the new form. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 64 ++++++++++++-------------------------------------- 1 file changed, 15 insertions(+), 49 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ typedef struct TempOptInfo { QSIMPLEQ_HEAD(, MemCopyInfo) mem_copy; uint64_t val; uint64_t z_mask; /* mask bit is 0 if and only if value bit is 0 */ - uint64_t s_mask; /* a left-aligned mask of clrsb(value) bits. */ + uint64_t s_mask; /* mask bit is 1 if value bit matches msb */ } TempOptInfo; typedef struct OptContext { @@ -XXX,XX +XXX,XX @@ typedef struct OptContext { /* In flight values from optimization. */ uint64_t z_mask; /* mask bit is 0 iff value bit is 0 */ - uint64_t s_mask; /* mask of clrsb(value) bits */ + uint64_t s_mask; /* mask bit is 1 if value bit matches msb */ TCGType type; } OptContext; -/* Calculate the smask for a specific value. */ -static uint64_t smask_from_value(uint64_t value) -{ - int rep = clrsb64(value); - return ~(~0ull >> rep); -} - -/* - * Calculate the smask for a given set of known-zeros. - * If there are lots of zeros on the left, we can consider the remainder - * an unsigned field, and thus the corresponding signed field is one bit - * larger. - */ -static uint64_t smask_from_zmask(uint64_t zmask) -{ - /* - * Only the 0 bits are significant for zmask, thus the msb itself - * must be zero, else we have no sign information. - */ - int rep = clz64(zmask); - if (rep == 0) { - return 0; - } - rep -= 1; - return ~(~0ull >> rep); -} - -/* - * Recreate a properly left-aligned smask after manipulation. - * Some bit-shuffling, particularly shifts and rotates, may - * retain sign bits on the left, but may scatter disconnected - * sign bits on the right. Retain only what remains to the left. - */ -static uint64_t smask_from_smask(int64_t smask) -{ - /* Only the 1 bits are significant for smask */ - return smask_from_zmask(~smask); -} - static inline TempOptInfo *ts_info(TCGTemp *ts) { return ts->state_ptr; @@ -XXX,XX +XXX,XX @@ static void init_ts_info(OptContext *ctx, TCGTemp *ts) ti->is_const = true; ti->val = ts->val; ti->z_mask = ts->val; - ti->s_mask = smask_from_value(ts->val); + ti->s_mask = INT64_MIN >> clrsb64(ts->val); } else { ti->is_const = false; ti->z_mask = -1; @@ -XXX,XX +XXX,XX @@ static void finish_folding(OptContext *ctx, TCGOp *op) */ if (i == 0) { ts_info(ts)->z_mask = ctx->z_mask; - ts_info(ts)->s_mask = ctx->s_mask; } } } @@ -XXX,XX +XXX,XX @@ static bool fold_const2_commutative(OptContext *ctx, TCGOp *op) * The passed s_mask may be augmented by z_mask. */ static bool fold_masks_zs(OptContext *ctx, TCGOp *op, - uint64_t z_mask, uint64_t s_mask) + uint64_t z_mask, int64_t s_mask) { const TCGOpDef *def = &tcg_op_defs[op->opc]; TCGTemp *ts; TempOptInfo *ti; + int rep; /* Only single-output opcodes are supported here. */ tcg_debug_assert(def->nb_oargs == 1); @@ -XXX,XX +XXX,XX @@ static bool fold_masks_zs(OptContext *ctx, TCGOp *op, */ if (ctx->type == TCG_TYPE_I32) { z_mask = (int32_t)z_mask; - s_mask |= MAKE_64BIT_MASK(32, 32); + s_mask |= INT32_MIN; } if (z_mask == 0) { @@ -XXX,XX +XXX,XX @@ static bool fold_masks_zs(OptContext *ctx, TCGOp *op, ti = ts_info(ts); ti->z_mask = z_mask; - ti->s_mask = s_mask | smask_from_zmask(z_mask); + + /* Canonicalize s_mask and incorporate data from z_mask. */ + rep = clz64(~s_mask); + rep = MAX(rep, clz64(z_mask)); + rep = MAX(rep - 1, 0); + ti->s_mask = INT64_MIN >> rep; + return true; } @@ -XXX,XX +XXX,XX @@ static bool fold_exts(OptContext *ctx, TCGOp *op) ctx->z_mask = z_mask; ctx->s_mask = s_mask; - if (!type_change && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { + if (0 && !type_change && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool fold_sextract(OptContext *ctx, TCGOp *op) s_mask |= MAKE_64BIT_MASK(len, 64 - len); ctx->s_mask = s_mask; - if (pos == 0 && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { + if (0 && pos == 0 && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool fold_shift(OptContext *ctx, TCGOp *op) ctx->z_mask = do_constant_folding(op->opc, ctx->type, z_mask, sh); s_mask = do_constant_folding(op->opc, ctx->type, s_mask, sh); - ctx->s_mask = smask_from_smask(s_mask); return fold_masks(ctx, op); } -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static void finish_ebb(OptContext *ctx) remove_mem_copy_all(ctx); } -static void finish_folding(OptContext *ctx, TCGOp *op) +static bool finish_folding(OptContext *ctx, TCGOp *op) { const TCGOpDef *def = &tcg_op_defs[op->opc]; int i, nb_oargs; @@ -XXX,XX +XXX,XX @@ static void finish_folding(OptContext *ctx, TCGOp *op) ts_info(ts)->z_mask = ctx->z_mask; } } + return true; } /* @@ -XXX,XX +XXX,XX @@ static bool fold_add(OptContext *ctx, TCGOp *op) fold_xi_to_x(ctx, op, 0)) { return true; } - return false; + return finish_folding(ctx, op); } /* We cannot as yet do_constant_folding with vectors. */ @@ -XXX,XX +XXX,XX @@ static bool fold_add_vec(OptContext *ctx, TCGOp *op) fold_xi_to_x(ctx, op, 0)) { return true; } - return false; + return finish_folding(ctx, op); } static bool fold_addsub2(OptContext *ctx, TCGOp *op, bool add) @@ -XXX,XX +XXX,XX @@ static bool fold_addsub2(OptContext *ctx, TCGOp *op, bool add) op->args[4] = arg_new_constant(ctx, bl); op->args[5] = arg_new_constant(ctx, bh); } - return false; + return finish_folding(ctx, op); } static bool fold_add2(OptContext *ctx, TCGOp *op) -- 2.43.0 Introduce ti_is_const, ti_const_val, ti_is_const_val. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 20 +++++++++++++++++--- 1 file changed, 17 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static inline TempOptInfo *arg_info(TCGArg arg) return ts_info(arg_temp(arg)); } +static inline bool ti_is_const(TempOptInfo *ti) +{ + return ti->is_const; +} + +static inline uint64_t ti_const_val(TempOptInfo *ti) +{ + return ti->val; +} + +static inline bool ti_is_const_val(TempOptInfo *ti, uint64_t val) +{ + return ti_is_const(ti) && ti_const_val(ti) == val; +} + static inline bool ts_is_const(TCGTemp *ts) { - return ts_info(ts)->is_const; + return ti_is_const(ts_info(ts)); } static inline bool ts_is_const_val(TCGTemp *ts, uint64_t val) { - TempOptInfo *ti = ts_info(ts); - return ti->is_const && ti->val == val; + return ti_is_const_val(ts_info(ts), val); } static inline bool arg_is_const(TCGArg arg) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Sink mask computation below fold_affected_mask early exit. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 30 ++++++++++++++++-------------- 1 file changed, 16 insertions(+), 14 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_add2(OptContext *ctx, TCGOp *op) static bool fold_and(OptContext *ctx, TCGOp *op) { - uint64_t z1, z2; + uint64_t z1, z2, z_mask, s_mask; + TempOptInfo *t1, *t2; if (fold_const2_commutative(ctx, op) || fold_xi_to_i(ctx, op, 0) || @@ -XXX,XX +XXX,XX @@ static bool fold_and(OptContext *ctx, TCGOp *op) return true; } - z1 = arg_info(op->args[1])->z_mask; - z2 = arg_info(op->args[2])->z_mask; - ctx->z_mask = z1 & z2; - - /* - * Sign repetitions are perforce all identical, whether they are 1 or 0. - * Bitwise operations preserve the relative quantity of the repetitions. - */ - ctx->s_mask = arg_info(op->args[1])->s_mask - & arg_info(op->args[2])->s_mask; + t1 = arg_info(op->args[1]); + t2 = arg_info(op->args[2]); + z1 = t1->z_mask; + z2 = t2->z_mask; /* * Known-zeros does not imply known-ones. Therefore unless * arg2 is constant, we can't infer affected bits from it. */ - if (arg_is_const(op->args[2]) && - fold_affected_mask(ctx, op, z1 & ~z2)) { + if (ti_is_const(t2) && fold_affected_mask(ctx, op, z1 & ~z2)) { return true; } - return fold_masks(ctx, op); + z_mask = z1 & z2; + + /* + * Sign repetitions are perforce all identical, whether they are 1 or 0. + * Bitwise operations preserve the relative quantity of the repetitions. + */ + s_mask = t1->s_mask & t2->s_mask; + + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_andc(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Avoid double inversion of the value of second const operand. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 21 +++++++++++---------- 1 file changed, 11 insertions(+), 10 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_and(OptContext *ctx, TCGOp *op) static bool fold_andc(OptContext *ctx, TCGOp *op) { - uint64_t z1; + uint64_t z_mask, s_mask; + TempOptInfo *t1, *t2; if (fold_const2(ctx, op) || fold_xx_to_i(ctx, op, 0) || @@ -XXX,XX +XXX,XX @@ static bool fold_andc(OptContext *ctx, TCGOp *op) return true; } - z1 = arg_info(op->args[1])->z_mask; + t1 = arg_info(op->args[1]); + t2 = arg_info(op->args[2]); + z_mask = t1->z_mask; /* * Known-zeros does not imply known-ones. Therefore unless * arg2 is constant, we can't infer anything from it. */ - if (arg_is_const(op->args[2])) { - uint64_t z2 = ~arg_info(op->args[2])->z_mask; - if (fold_affected_mask(ctx, op, z1 & ~z2)) { + if (ti_is_const(t2)) { + uint64_t v2 = ti_const_val(t2); + if (fold_affected_mask(ctx, op, z_mask & v2)) { return true; } - z1 &= z2; + z_mask &= ~v2; } - ctx->z_mask = z1; - ctx->s_mask = arg_info(op->args[1])->s_mask - & arg_info(op->args[2])->s_mask; - return fold_masks(ctx, op); + s_mask = t1->s_mask & t2->s_mask; + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_brcond(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Always set s_mask along the BSWAP_OS path, since the result is being explicitly sign-extended. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 21 ++++++++++----------- 1 file changed, 10 insertions(+), 11 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_brcond2(OptContext *ctx, TCGOp *op) static bool fold_bswap(OptContext *ctx, TCGOp *op) { uint64_t z_mask, s_mask, sign; + TempOptInfo *t1 = arg_info(op->args[1]); - if (arg_is_const(op->args[1])) { - uint64_t t = arg_info(op->args[1])->val; - - t = do_constant_folding(op->opc, ctx->type, t, op->args[2]); - return tcg_opt_gen_movi(ctx, op, op->args[0], t); + if (ti_is_const(t1)) { + return tcg_opt_gen_movi(ctx, op, op->args[0], + do_constant_folding(op->opc, ctx->type, + ti_const_val(t1), + op->args[2])); } - z_mask = arg_info(op->args[1])->z_mask; - + z_mask = t1->z_mask; switch (op->opc) { case INDEX_op_bswap16_i32: case INDEX_op_bswap16_i64: @@ -XXX,XX +XXX,XX @@ static bool fold_bswap(OptContext *ctx, TCGOp *op) /* If the sign bit may be 1, force all the bits above to 1. */ if (z_mask & sign) { z_mask |= sign; - s_mask = sign << 1; } + /* The value and therefore s_mask is explicitly sign-extended. */ + s_mask = sign; break; default: /* The high bits are undefined: force all bits above the sign to 1. */ z_mask |= sign << 1; break; } - ctx->z_mask = z_mask; - ctx->s_mask = s_mask; - return fold_masks(ctx, op); + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_call(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Compute s_mask from the union of the maximum count and the op2 fallback for op1 being zero. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_call(OptContext *ctx, TCGOp *op) static bool fold_count_zeros(OptContext *ctx, TCGOp *op) { - uint64_t z_mask; + uint64_t z_mask, s_mask; + TempOptInfo *t1 = arg_info(op->args[1]); + TempOptInfo *t2 = arg_info(op->args[2]); - if (arg_is_const(op->args[1])) { - uint64_t t = arg_info(op->args[1])->val; + if (ti_is_const(t1)) { + uint64_t t = ti_const_val(t1); if (t != 0) { t = do_constant_folding(op->opc, ctx->type, t, 0); @@ -XXX,XX +XXX,XX @@ static bool fold_count_zeros(OptContext *ctx, TCGOp *op) default: g_assert_not_reached(); } - ctx->z_mask = arg_info(op->args[2])->z_mask | z_mask; - return false; + s_mask = ~z_mask; + z_mask |= t2->z_mask; + s_mask &= t2->s_mask; + + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_ctpop(OptContext *ctx, TCGOp *op) -- 2.43.0 Add fold_masks_z as a trivial wrapper around fold_masks_zs. Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_masks_zs(OptContext *ctx, TCGOp *op, return true; } +static bool fold_masks_z(OptContext *ctx, TCGOp *op, uint64_t z_mask) +{ + return fold_masks_zs(ctx, op, z_mask, 0); +} + static bool fold_masks(OptContext *ctx, TCGOp *op) { return fold_masks_zs(ctx, op, ctx->z_mask, ctx->s_mask); @@ -XXX,XX +XXX,XX @@ static bool fold_count_zeros(OptContext *ctx, TCGOp *op) static bool fold_ctpop(OptContext *ctx, TCGOp *op) { + uint64_t z_mask; + if (fold_const1(ctx, op)) { return true; } switch (ctx->type) { case TCG_TYPE_I32: - ctx->z_mask = 32 | 31; + z_mask = 32 | 31; break; case TCG_TYPE_I64: - ctx->z_mask = 64 | 63; + z_mask = 64 | 63; break; default: g_assert_not_reached(); } - return false; + return fold_masks_z(ctx, op, z_mask); } static bool fold_deposit(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. When we fold to and, use fold_and. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 35 +++++++++++++++++------------------ 1 file changed, 17 insertions(+), 18 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_ctpop(OptContext *ctx, TCGOp *op) static bool fold_deposit(OptContext *ctx, TCGOp *op) { + TempOptInfo *t1 = arg_info(op->args[1]); + TempOptInfo *t2 = arg_info(op->args[2]); + int ofs = op->args[3]; + int len = op->args[4]; TCGOpcode and_opc; + uint64_t z_mask; - if (arg_is_const(op->args[1]) && arg_is_const(op->args[2])) { - uint64_t t1 = arg_info(op->args[1])->val; - uint64_t t2 = arg_info(op->args[2])->val; - - t1 = deposit64(t1, op->args[3], op->args[4], t2); - return tcg_opt_gen_movi(ctx, op, op->args[0], t1); + if (ti_is_const(t1) && ti_is_const(t2)) { + return tcg_opt_gen_movi(ctx, op, op->args[0], + deposit64(ti_const_val(t1), ofs, len, + ti_const_val(t2))); } switch (ctx->type) { @@ -XXX,XX +XXX,XX @@ static bool fold_deposit(OptContext *ctx, TCGOp *op) } /* Inserting a value into zero at offset 0. */ - if (arg_is_const_val(op->args[1], 0) && op->args[3] == 0) { - uint64_t mask = MAKE_64BIT_MASK(0, op->args[4]); + if (ti_is_const_val(t1, 0) && ofs == 0) { + uint64_t mask = MAKE_64BIT_MASK(0, len); op->opc = and_opc; op->args[1] = op->args[2]; op->args[2] = arg_new_constant(ctx, mask); - ctx->z_mask = mask & arg_info(op->args[1])->z_mask; - return false; + return fold_and(ctx, op); } /* Inserting zero into a value. */ - if (arg_is_const_val(op->args[2], 0)) { - uint64_t mask = deposit64(-1, op->args[3], op->args[4], 0); + if (ti_is_const_val(t2, 0)) { + uint64_t mask = deposit64(-1, ofs, len, 0); op->opc = and_opc; op->args[2] = arg_new_constant(ctx, mask); - ctx->z_mask = mask & arg_info(op->args[1])->z_mask; - return false; + return fold_and(ctx, op); } - ctx->z_mask = deposit64(arg_info(op->args[1])->z_mask, - op->args[3], op->args[4], - arg_info(op->args[2])->z_mask); - return false; + z_mask = deposit64(t1->z_mask, ofs, len, t2->z_mask); + return fold_masks_z(ctx, op, z_mask); } static bool fold_divide(OptContext *ctx, TCGOp *op) -- 2.43.0 The input which overlaps the sign bit of the output can have its input s_mask propagated to the output s_mask. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_deposit(OptContext *ctx, TCGOp *op) TempOptInfo *t2 = arg_info(op->args[2]); int ofs = op->args[3]; int len = op->args[4]; + int width; TCGOpcode and_opc; - uint64_t z_mask; + uint64_t z_mask, s_mask; if (ti_is_const(t1) && ti_is_const(t2)) { return tcg_opt_gen_movi(ctx, op, op->args[0], @@ -XXX,XX +XXX,XX @@ static bool fold_deposit(OptContext *ctx, TCGOp *op) switch (ctx->type) { case TCG_TYPE_I32: and_opc = INDEX_op_and_i32; + width = 32; break; case TCG_TYPE_I64: and_opc = INDEX_op_and_i64; + width = 64; break; default: g_assert_not_reached(); @@ -XXX,XX +XXX,XX @@ static bool fold_deposit(OptContext *ctx, TCGOp *op) return fold_and(ctx, op); } + /* The s_mask from the top portion of the deposit is still valid. */ + if (ofs + len == width) { + s_mask = t2->s_mask << ofs; + } else { + s_mask = t1->s_mask & ~MAKE_64BIT_MASK(0, ofs + len); + } + z_mask = deposit64(t1->z_mask, ofs, len, t2->z_mask); - return fold_masks_z(ctx, op, z_mask); + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_divide(OptContext *ctx, TCGOp *op) -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_divide(OptContext *ctx, TCGOp *op) fold_xi_to_x(ctx, op, 1)) { return true; } - return false; + return finish_folding(ctx, op); } static bool fold_dup(OptContext *ctx, TCGOp *op) -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_dup(OptContext *ctx, TCGOp *op) t = dup_const(TCGOP_VECE(op), t); return tcg_opt_gen_movi(ctx, op, op->args[0], t); } - return false; + return finish_folding(ctx, op); } static bool fold_dup2(OptContext *ctx, TCGOp *op) @@ -XXX,XX +XXX,XX @@ static bool fold_dup2(OptContext *ctx, TCGOp *op) op->opc = INDEX_op_dup_vec; TCGOP_VECE(op) = MO_32; } - return false; + return finish_folding(ctx, op); } static bool fold_eqv(OptContext *ctx, TCGOp *op) -- 2.43.0 Add fold_masks_s as a trivial wrapper around fold_masks_zs. Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_masks_z(OptContext *ctx, TCGOp *op, uint64_t z_mask) return fold_masks_zs(ctx, op, z_mask, 0); } +static bool fold_masks_s(OptContext *ctx, TCGOp *op, uint64_t s_mask) +{ + return fold_masks_zs(ctx, op, -1, s_mask); +} + static bool fold_masks(OptContext *ctx, TCGOp *op) { return fold_masks_zs(ctx, op, ctx->z_mask, ctx->s_mask); @@ -XXX,XX +XXX,XX @@ static bool fold_dup2(OptContext *ctx, TCGOp *op) static bool fold_eqv(OptContext *ctx, TCGOp *op) { + uint64_t s_mask; + if (fold_const2_commutative(ctx, op) || fold_xi_to_x(ctx, op, -1) || fold_xi_to_not(ctx, op, 0)) { return true; } - ctx->s_mask = arg_info(op->args[1])->s_mask - & arg_info(op->args[2])->s_mask; - return false; + s_mask = arg_info(op->args[1])->s_mask + & arg_info(op->args[2])->s_mask; + return fold_masks_s(ctx, op, s_mask); } static bool fold_extract(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_eqv(OptContext *ctx, TCGOp *op) static bool fold_extract(OptContext *ctx, TCGOp *op) { uint64_t z_mask_old, z_mask; + TempOptInfo *t1 = arg_info(op->args[1]); int pos = op->args[2]; int len = op->args[3]; - if (arg_is_const(op->args[1])) { - uint64_t t; - - t = arg_info(op->args[1])->val; - t = extract64(t, pos, len); - return tcg_opt_gen_movi(ctx, op, op->args[0], t); + if (ti_is_const(t1)) { + return tcg_opt_gen_movi(ctx, op, op->args[0], + extract64(ti_const_val(t1), pos, len)); } - z_mask_old = arg_info(op->args[1])->z_mask; + z_mask_old = t1->z_mask; z_mask = extract64(z_mask_old, pos, len); if (pos == 0 && fold_affected_mask(ctx, op, z_mask_old ^ z_mask)) { return true; } - ctx->z_mask = z_mask; - return fold_masks(ctx, op); + return fold_masks_z(ctx, op, z_mask); } static bool fold_extract2(OptContext *ctx, TCGOp *op) -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_extract2(OptContext *ctx, TCGOp *op) } return tcg_opt_gen_movi(ctx, op, op->args[0], v1 | v2); } - return false; + return finish_folding(ctx, op); } static bool fold_exts(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Explicitly sign-extend z_mask instead of doing that manually. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 29 ++++++++++++----------------- 1 file changed, 12 insertions(+), 17 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_extract2(OptContext *ctx, TCGOp *op) static bool fold_exts(OptContext *ctx, TCGOp *op) { - uint64_t s_mask_old, s_mask, z_mask, sign; + uint64_t s_mask_old, s_mask, z_mask; bool type_change = false; + TempOptInfo *t1; if (fold_const1(ctx, op)) { return true; } - z_mask = arg_info(op->args[1])->z_mask; - s_mask = arg_info(op->args[1])->s_mask; + t1 = arg_info(op->args[1]); + z_mask = t1->z_mask; + s_mask = t1->s_mask; s_mask_old = s_mask; switch (op->opc) { CASE_OP_32_64(ext8s): - sign = INT8_MIN; - z_mask = (uint8_t)z_mask; + s_mask |= INT8_MIN; + z_mask = (int8_t)z_mask; break; CASE_OP_32_64(ext16s): - sign = INT16_MIN; - z_mask = (uint16_t)z_mask; + s_mask |= INT16_MIN; + z_mask = (int16_t)z_mask; break; case INDEX_op_ext_i32_i64: type_change = true; QEMU_FALLTHROUGH; case INDEX_op_ext32s_i64: - sign = INT32_MIN; - z_mask = (uint32_t)z_mask; + s_mask |= INT32_MIN; + z_mask = (int32_t)z_mask; break; default: g_assert_not_reached(); } - if (z_mask & sign) { - z_mask |= sign; - } - s_mask |= sign << 1; - - ctx->z_mask = z_mask; - ctx->s_mask = s_mask; if (0 && !type_change && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { return true; } - return fold_masks(ctx, op); + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_extu(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_extu(OptContext *ctx, TCGOp *op) g_assert_not_reached(); } - ctx->z_mask = z_mask; if (!type_change && fold_affected_mask(ctx, op, z_mask_old ^ z_mask)) { return true; } - return fold_masks(ctx, op); + + return fold_masks_z(ctx, op, z_mask); } static bool fold_mb(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 19 +++++++++++-------- 1 file changed, 11 insertions(+), 8 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_mov(OptContext *ctx, TCGOp *op) static bool fold_movcond(OptContext *ctx, TCGOp *op) { + uint64_t z_mask, s_mask; + TempOptInfo *tt, *ft; int i; /* If true and false values are the same, eliminate the cmp. */ @@ -XXX,XX +XXX,XX @@ static bool fold_movcond(OptContext *ctx, TCGOp *op) return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[4 - i]); } - ctx->z_mask = arg_info(op->args[3])->z_mask - | arg_info(op->args[4])->z_mask; - ctx->s_mask = arg_info(op->args[3])->s_mask - & arg_info(op->args[4])->s_mask; + tt = arg_info(op->args[3]); + ft = arg_info(op->args[4]); + z_mask = tt->z_mask | ft->z_mask; + s_mask = tt->s_mask & ft->s_mask; - if (arg_is_const(op->args[3]) && arg_is_const(op->args[4])) { - uint64_t tv = arg_info(op->args[3])->val; - uint64_t fv = arg_info(op->args[4])->val; + if (ti_is_const(tt) && ti_is_const(ft)) { + uint64_t tv = ti_const_val(tt); + uint64_t fv = ti_const_val(ft); TCGOpcode opc, negopc = 0; TCGCond cond = op->args[5]; @@ -XXX,XX +XXX,XX @@ static bool fold_movcond(OptContext *ctx, TCGOp *op) } } } - return false; + + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_mul(OptContext *ctx, TCGOp *op) -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_mul(OptContext *ctx, TCGOp *op) fold_xi_to_x(ctx, op, 1)) { return true; } - return false; + return finish_folding(ctx, op); } static bool fold_mul_highpart(OptContext *ctx, TCGOp *op) @@ -XXX,XX +XXX,XX @@ static bool fold_mul_highpart(OptContext *ctx, TCGOp *op) fold_xi_to_i(ctx, op, 0)) { return true; } - return false; + return finish_folding(ctx, op); } static bool fold_multiply2(OptContext *ctx, TCGOp *op) @@ -XXX,XX +XXX,XX @@ static bool fold_multiply2(OptContext *ctx, TCGOp *op) tcg_opt_gen_movi(ctx, op2, rh, h); return true; } - return false; + return finish_folding(ctx, op); } static bool fold_nand(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_multiply2(OptContext *ctx, TCGOp *op) static bool fold_nand(OptContext *ctx, TCGOp *op) { + uint64_t s_mask; + if (fold_const2_commutative(ctx, op) || fold_xi_to_not(ctx, op, -1)) { return true; } - ctx->s_mask = arg_info(op->args[1])->s_mask - & arg_info(op->args[2])->s_mask; - return false; + s_mask = arg_info(op->args[1])->s_mask + & arg_info(op->args[2])->s_mask; + return fold_masks_s(ctx, op, s_mask); } static bool fold_neg_no_const(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 9 ++------- 1 file changed, 2 insertions(+), 7 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_neg_no_const(OptContext *ctx, TCGOp *op) { /* Set to 1 all bits to the left of the rightmost. */ uint64_t z_mask = arg_info(op->args[1])->z_mask; - ctx->z_mask = -(z_mask & -z_mask); + z_mask = -(z_mask & -z_mask); - /* - * Because of fold_sub_to_neg, we want to always return true, - * via finish_folding. - */ - finish_folding(ctx, op); - return true; + return fold_masks_z(ctx, op, z_mask); } static bool fold_neg(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_neg(OptContext *ctx, TCGOp *op) static bool fold_nor(OptContext *ctx, TCGOp *op) { + uint64_t s_mask; + if (fold_const2_commutative(ctx, op) || fold_xi_to_not(ctx, op, 0)) { return true; } - ctx->s_mask = arg_info(op->args[1])->s_mask - & arg_info(op->args[2])->s_mask; - return false; + s_mask = arg_info(op->args[1])->s_mask + & arg_info(op->args[2])->s_mask; + return fold_masks_s(ctx, op, s_mask); } static bool fold_not(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_not(OptContext *ctx, TCGOp *op) if (fold_const1(ctx, op)) { return true; } - - ctx->s_mask = arg_info(op->args[1])->s_mask; - - /* Because of fold_to_not, we want to always return true, via finish. */ - finish_folding(ctx, op); - return true; + return fold_masks_s(ctx, op, arg_info(op->args[1])->s_mask); } static bool fold_or(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_not(OptContext *ctx, TCGOp *op) static bool fold_or(OptContext *ctx, TCGOp *op) { + uint64_t z_mask, s_mask; + TempOptInfo *t1, *t2; + if (fold_const2_commutative(ctx, op) || fold_xi_to_x(ctx, op, 0) || fold_xx_to_x(ctx, op)) { return true; } - ctx->z_mask = arg_info(op->args[1])->z_mask - | arg_info(op->args[2])->z_mask; - ctx->s_mask = arg_info(op->args[1])->s_mask - & arg_info(op->args[2])->s_mask; - return fold_masks(ctx, op); + t1 = arg_info(op->args[1]); + t2 = arg_info(op->args[2]); + z_mask = t1->z_mask | t2->z_mask; + s_mask = t1->s_mask & t2->s_mask; + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_orc(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_or(OptContext *ctx, TCGOp *op) static bool fold_orc(OptContext *ctx, TCGOp *op) { + uint64_t s_mask; + if (fold_const2(ctx, op) || fold_xx_to_i(ctx, op, -1) || fold_xi_to_x(ctx, op, -1) || @@ -XXX,XX +XXX,XX @@ static bool fold_orc(OptContext *ctx, TCGOp *op) return true; } - ctx->s_mask = arg_info(op->args[1])->s_mask - & arg_info(op->args[2])->s_mask; - return false; + s_mask = arg_info(op->args[1])->s_mask + & arg_info(op->args[2])->s_mask; + return fold_masks_s(ctx, op, s_mask); } static bool fold_qemu_ld(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Be careful not to call fold_masks_zs when the memory operation is wide enough to require multiple outputs, so split into two functions: fold_qemu_ld_1reg and fold_qemu_ld_2reg. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 26 +++++++++++++++++++++----- 1 file changed, 21 insertions(+), 5 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_orc(OptContext *ctx, TCGOp *op) return fold_masks_s(ctx, op, s_mask); } -static bool fold_qemu_ld(OptContext *ctx, TCGOp *op) +static bool fold_qemu_ld_1reg(OptContext *ctx, TCGOp *op) { const TCGOpDef *def = &tcg_op_defs[op->opc]; MemOpIdx oi = op->args[def->nb_oargs + def->nb_iargs]; MemOp mop = get_memop(oi); int width = 8 * memop_size(mop); + uint64_t z_mask = -1, s_mask = 0; if (width < 64) { if (mop & MO_SIGN) { - ctx->s_mask = MAKE_64BIT_MASK(width, 64 - width); + s_mask = MAKE_64BIT_MASK(width - 1, 64 - (width - 1)); } else { - ctx->z_mask = MAKE_64BIT_MASK(0, width); + z_mask = MAKE_64BIT_MASK(0, width); } } /* Opcodes that touch guest memory stop the mb optimization. */ ctx->prev_mb = NULL; - return false; + + return fold_masks_zs(ctx, op, z_mask, s_mask); +} + +static bool fold_qemu_ld_2reg(OptContext *ctx, TCGOp *op) +{ + /* Opcodes that touch guest memory stop the mb optimization. */ + ctx->prev_mb = NULL; + return finish_folding(ctx, op); } static bool fold_qemu_st(OptContext *ctx, TCGOp *op) @@ -XXX,XX +XXX,XX @@ void tcg_optimize(TCGContext *s) break; case INDEX_op_qemu_ld_a32_i32: case INDEX_op_qemu_ld_a64_i32: + done = fold_qemu_ld_1reg(&ctx, op); + break; case INDEX_op_qemu_ld_a32_i64: case INDEX_op_qemu_ld_a64_i64: + if (TCG_TARGET_REG_BITS == 64) { + done = fold_qemu_ld_1reg(&ctx, op); + break; + } + QEMU_FALLTHROUGH; case INDEX_op_qemu_ld_a32_i128: case INDEX_op_qemu_ld_a64_i128: - done = fold_qemu_ld(&ctx, op); + done = fold_qemu_ld_2reg(&ctx, op); break; case INDEX_op_qemu_st8_a32_i32: case INDEX_op_qemu_st8_a64_i32: -- 2.43.0 Stores have no output operands, and so need no further work. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_qemu_st(OptContext *ctx, TCGOp *op) { /* Opcodes that touch guest memory stop the mb optimization. */ ctx->prev_mb = NULL; - return false; + return true; } static bool fold_remainder(OptContext *ctx, TCGOp *op) @@ -XXX,XX +XXX,XX @@ static bool fold_tcg_st(OptContext *ctx, TCGOp *op) if (op->args[1] != tcgv_ptr_arg(tcg_env)) { remove_mem_copy_all(ctx); - return false; + return true; } switch (op->opc) { @@ -XXX,XX +XXX,XX @@ static bool fold_tcg_st(OptContext *ctx, TCGOp *op) g_assert_not_reached(); } remove_mem_copy_in(ctx, ofs, ofs + lm1); - return false; + return true; } static bool fold_tcg_st_memcopy(OptContext *ctx, TCGOp *op) @@ -XXX,XX +XXX,XX @@ static bool fold_tcg_st_memcopy(OptContext *ctx, TCGOp *op) TCGType type; if (op->args[1] != tcgv_ptr_arg(tcg_env)) { - fold_tcg_st(ctx, op); - return false; + return fold_tcg_st(ctx, op); } src = arg_temp(op->args[0]); @@ -XXX,XX +XXX,XX @@ static bool fold_tcg_st_memcopy(OptContext *ctx, TCGOp *op) last = ofs + tcg_type_size(type) - 1; remove_mem_copy_in(ctx, ofs, last); record_mem_copy(ctx, type, src, ofs, last); - return false; + return true; } static bool fold_xor(OptContext *ctx, TCGOp *op) -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_remainder(OptContext *ctx, TCGOp *op) fold_xx_to_i(ctx, op, 0)) { return true; } - return false; + return finish_folding(ctx, op); } static bool fold_setcond_zmask(OptContext *ctx, TCGOp *op, bool neg) -- 2.43.0 Change return from bool to int; distinguish between complete folding, simplification, and no change. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 22 ++++++++++++++-------- 1 file changed, 14 insertions(+), 8 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_remainder(OptContext *ctx, TCGOp *op) return finish_folding(ctx, op); } -static bool fold_setcond_zmask(OptContext *ctx, TCGOp *op, bool neg) +/* Return 1 if finished, -1 if simplified, 0 if unchanged. */ +static int fold_setcond_zmask(OptContext *ctx, TCGOp *op, bool neg) { uint64_t a_zmask, b_val; TCGCond cond; @@ -XXX,XX +XXX,XX @@ static bool fold_setcond_zmask(OptContext *ctx, TCGOp *op, bool neg) op->opc = xor_opc; op->args[2] = arg_new_constant(ctx, 1); } - return false; + return -1; } } - - return false; + return 0; } static void fold_setcond_tst_pow2(OptContext *ctx, TCGOp *op, bool neg) @@ -XXX,XX +XXX,XX @@ static bool fold_setcond(OptContext *ctx, TCGOp *op) return tcg_opt_gen_movi(ctx, op, op->args[0], i); } - if (fold_setcond_zmask(ctx, op, false)) { + i = fold_setcond_zmask(ctx, op, false); + if (i > 0) { return true; } - fold_setcond_tst_pow2(ctx, op, false); + if (i == 0) { + fold_setcond_tst_pow2(ctx, op, false); + } ctx->z_mask = 1; return false; @@ -XXX,XX +XXX,XX @@ static bool fold_negsetcond(OptContext *ctx, TCGOp *op) return tcg_opt_gen_movi(ctx, op, op->args[0], -i); } - if (fold_setcond_zmask(ctx, op, true)) { + i = fold_setcond_zmask(ctx, op, true); + if (i > 0) { return true; } - fold_setcond_tst_pow2(ctx, op, true); + if (i == 0) { + fold_setcond_tst_pow2(ctx, op, true); + } /* Value is {0,-1} so all bits are repetitions of the sign. */ ctx->s_mask = -1; -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_setcond(OptContext *ctx, TCGOp *op) fold_setcond_tst_pow2(ctx, op, false); } - ctx->z_mask = 1; - return false; + return fold_masks_z(ctx, op, 1); } static bool fold_negsetcond(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_negsetcond(OptContext *ctx, TCGOp *op) } /* Value is {0,-1} so all bits are repetitions of the sign. */ - ctx->s_mask = -1; - return false; + return fold_masks_s(ctx, op, -1); } static bool fold_setcond2(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_setcond2(OptContext *ctx, TCGOp *op) return fold_setcond(ctx, op); } - ctx->z_mask = 1; - return false; + return fold_masks_z(ctx, op, 1); do_setcond_const: return tcg_opt_gen_movi(ctx, op, op->args[0], i); -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_cmp_vec(OptContext *ctx, TCGOp *op) if (swap_commutative(NO_DEST, &op->args[1], &op->args[2])) { op->args[3] = tcg_swap_cond(op->args[3]); } - return false; + return finish_folding(ctx, op); } static bool fold_cmpsel_vec(OptContext *ctx, TCGOp *op) -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_cmpsel_vec(OptContext *ctx, TCGOp *op) if (swap_commutative(op->args[0], &op->args[4], &op->args[3])) { op->args[5] = tcg_invert_cond(op->args[5]); } - return false; + return finish_folding(ctx, op); } static bool fold_sextract(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 24 +++++++++--------------- 1 file changed, 9 insertions(+), 15 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_cmpsel_vec(OptContext *ctx, TCGOp *op) static bool fold_sextract(OptContext *ctx, TCGOp *op) { uint64_t z_mask, s_mask, s_mask_old; + TempOptInfo *t1 = arg_info(op->args[1]); int pos = op->args[2]; int len = op->args[3]; - if (arg_is_const(op->args[1])) { - uint64_t t; - - t = arg_info(op->args[1])->val; - t = sextract64(t, pos, len); - return tcg_opt_gen_movi(ctx, op, op->args[0], t); + if (ti_is_const(t1)) { + return tcg_opt_gen_movi(ctx, op, op->args[0], + sextract64(ti_const_val(t1), pos, len)); } - z_mask = arg_info(op->args[1])->z_mask; - z_mask = sextract64(z_mask, pos, len); - ctx->z_mask = z_mask; - - s_mask_old = arg_info(op->args[1])->s_mask; - s_mask = sextract64(s_mask_old, pos, len); - s_mask |= MAKE_64BIT_MASK(len, 64 - len); - ctx->s_mask = s_mask; + s_mask_old = t1->s_mask; + s_mask = s_mask_old >> pos; + s_mask |= -1ull << (len - 1); if (0 && pos == 0 && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { return true; } - return fold_masks(ctx, op); + z_mask = sextract64(t1->z_mask, pos, len); + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_shift(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 27 ++++++++++++++------------- 1 file changed, 14 insertions(+), 13 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_sextract(OptContext *ctx, TCGOp *op) static bool fold_shift(OptContext *ctx, TCGOp *op) { uint64_t s_mask, z_mask, sign; + TempOptInfo *t1, *t2; if (fold_const2(ctx, op) || fold_ix_to_i(ctx, op, 0) || @@ -XXX,XX +XXX,XX @@ static bool fold_shift(OptContext *ctx, TCGOp *op) return true; } - s_mask = arg_info(op->args[1])->s_mask; - z_mask = arg_info(op->args[1])->z_mask; + t1 = arg_info(op->args[1]); + t2 = arg_info(op->args[2]); + s_mask = t1->s_mask; + z_mask = t1->z_mask; - if (arg_is_const(op->args[2])) { - int sh = arg_info(op->args[2])->val; - - ctx->z_mask = do_constant_folding(op->opc, ctx->type, z_mask, sh); + if (ti_is_const(t2)) { + int sh = ti_const_val(t2); + z_mask = do_constant_folding(op->opc, ctx->type, z_mask, sh); s_mask = do_constant_folding(op->opc, ctx->type, s_mask, sh); - return fold_masks(ctx, op); + return fold_masks_zs(ctx, op, z_mask, s_mask); } switch (op->opc) { @@ -XXX,XX +XXX,XX @@ static bool fold_shift(OptContext *ctx, TCGOp *op) * Arithmetic right shift will not reduce the number of * input sign repetitions. */ - ctx->s_mask = s_mask; - break; + return fold_masks_s(ctx, op, s_mask); CASE_OP_32_64(shr): /* * If the sign bit is known zero, then logical right shift - * will not reduced the number of input sign repetitions. + * will not reduce the number of input sign repetitions. */ - sign = (s_mask & -s_mask) >> 1; + sign = -s_mask; if (sign && !(z_mask & sign)) { - ctx->s_mask = s_mask; + return fold_masks_s(ctx, op, s_mask); } break; default: break; } - return false; + return finish_folding(ctx, op); } static bool fold_sub_to_neg(OptContext *ctx, TCGOp *op) -- 2.43.0 Merge the two conditions, sign != 0 && !(z_mask & sign), by testing ~z_mask & sign. If sign == 0, the logical and will produce false. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_sextract(OptContext *ctx, TCGOp *op) static bool fold_shift(OptContext *ctx, TCGOp *op) { - uint64_t s_mask, z_mask, sign; + uint64_t s_mask, z_mask; TempOptInfo *t1, *t2; if (fold_const2(ctx, op) || @@ -XXX,XX +XXX,XX @@ static bool fold_shift(OptContext *ctx, TCGOp *op) * If the sign bit is known zero, then logical right shift * will not reduce the number of input sign repetitions. */ - sign = -s_mask; - if (sign && !(z_mask & sign)) { + if (~z_mask & -s_mask) { return fold_masks_s(ctx, op, s_mask); } break; -- 2.43.0 Duplicate fold_sub_vec into fold_sub instead of calling it, now that fold_sub_vec always returns true. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_sub_vec(OptContext *ctx, TCGOp *op) fold_sub_to_neg(ctx, op)) { return true; } - return false; + return finish_folding(ctx, op); } static bool fold_sub(OptContext *ctx, TCGOp *op) { - if (fold_const2(ctx, op) || fold_sub_vec(ctx, op)) { + if (fold_const2(ctx, op) || + fold_xx_to_i(ctx, op, 0) || + fold_xi_to_x(ctx, op, 0) || + fold_sub_to_neg(ctx, op)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool fold_sub(OptContext *ctx, TCGOp *op) ? INDEX_op_add_i32 : INDEX_op_add_i64); op->args[2] = arg_new_constant(ctx, -val); } - return false; + return finish_folding(ctx, op); } static bool fold_sub2(OptContext *ctx, TCGOp *op) -- 2.43.0 Avoid the use of the OptContext slots. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 16 +++++++++------- 1 file changed, 9 insertions(+), 7 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_sub2(OptContext *ctx, TCGOp *op) static bool fold_tcg_ld(OptContext *ctx, TCGOp *op) { + uint64_t z_mask = -1, s_mask = 0; + /* We can't do any folding with a load, but we can record bits. */ switch (op->opc) { CASE_OP_32_64(ld8s): - ctx->s_mask = MAKE_64BIT_MASK(8, 56); + s_mask = INT8_MIN; break; CASE_OP_32_64(ld8u): - ctx->z_mask = MAKE_64BIT_MASK(0, 8); + z_mask = MAKE_64BIT_MASK(0, 8); break; CASE_OP_32_64(ld16s): - ctx->s_mask = MAKE_64BIT_MASK(16, 48); + s_mask = INT16_MIN; break; CASE_OP_32_64(ld16u): - ctx->z_mask = MAKE_64BIT_MASK(0, 16); + z_mask = MAKE_64BIT_MASK(0, 16); break; case INDEX_op_ld32s_i64: - ctx->s_mask = MAKE_64BIT_MASK(32, 32); + s_mask = INT32_MIN; break; case INDEX_op_ld32u_i64: - ctx->z_mask = MAKE_64BIT_MASK(0, 32); + z_mask = MAKE_64BIT_MASK(0, 32); break; default: g_assert_not_reached(); } - return false; + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_tcg_ld_memcopy(OptContext *ctx, TCGOp *op) -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_tcg_ld_memcopy(OptContext *ctx, TCGOp *op) TCGType type; if (op->args[1] != tcgv_ptr_arg(tcg_env)) { - return false; + return finish_folding(ctx, op); } type = ctx->type; -- 2.43.0 Avoid the use of the OptContext slots. Find TempOptInfo once. Remove fold_masks as the function becomes unused. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 18 ++++++++---------- 1 file changed, 8 insertions(+), 10 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_masks_s(OptContext *ctx, TCGOp *op, uint64_t s_mask) return fold_masks_zs(ctx, op, -1, s_mask); } -static bool fold_masks(OptContext *ctx, TCGOp *op) -{ - return fold_masks_zs(ctx, op, ctx->z_mask, ctx->s_mask); -} - /* * An "affected" mask bit is 0 if and only if the result is identical * to the first input. Thus if the entire mask is 0, the operation @@ -XXX,XX +XXX,XX @@ static bool fold_tcg_st_memcopy(OptContext *ctx, TCGOp *op) static bool fold_xor(OptContext *ctx, TCGOp *op) { + uint64_t z_mask, s_mask; + TempOptInfo *t1, *t2; + if (fold_const2_commutative(ctx, op) || fold_xx_to_i(ctx, op, 0) || fold_xi_to_x(ctx, op, 0) || @@ -XXX,XX +XXX,XX @@ static bool fold_xor(OptContext *ctx, TCGOp *op) return true; } - ctx->z_mask = arg_info(op->args[1])->z_mask - | arg_info(op->args[2])->z_mask; - ctx->s_mask = arg_info(op->args[1])->s_mask - & arg_info(op->args[2])->s_mask; - return fold_masks(ctx, op); + t1 = arg_info(op->args[1]); + t2 = arg_info(op->args[2]); + z_mask = t1->z_mask | t2->z_mask; + s_mask = t1->s_mask & t2->s_mask; + return fold_masks_zs(ctx, op, z_mask, s_mask); } static bool fold_bitsel_vec(OptContext *ctx, TCGOp *op) -- 2.43.0 Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_bitsel_vec(OptContext *ctx, TCGOp *op) return fold_orc(ctx, op); } } - return false; + return finish_folding(ctx, op); } /* Propagate constants and copies, fold constant expressions. */ -- 2.43.0 All non-default cases now finish folding within each function. Do the same with the default case and assert it is done after. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ void tcg_optimize(TCGContext *s) done = true; break; default: + done = finish_folding(&ctx, op); break; } - - if (!done) { - finish_folding(&ctx, op); - } + tcg_debug_assert(done); } } -- 2.43.0 All mask setting is now done with parameters via fold_masks_*. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 13 ------------- 1 file changed, 13 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ typedef struct OptContext { QSIMPLEQ_HEAD(, MemCopyInfo) mem_free; /* In flight values from optimization. */ - uint64_t z_mask; /* mask bit is 0 iff value bit is 0 */ - uint64_t s_mask; /* mask bit is 1 if value bit matches msb */ TCGType type; } OptContext; @@ -XXX,XX +XXX,XX @@ static bool finish_folding(OptContext *ctx, TCGOp *op) for (i = 0; i < nb_oargs; i++) { TCGTemp *ts = arg_temp(op->args[i]); reset_ts(ctx, ts); - /* - * Save the corresponding known-zero/sign bits mask for the - * first output argument (only one supported so far). - */ - if (i == 0) { - ts_info(ts)->z_mask = ctx->z_mask; - } } return true; } @@ -XXX,XX +XXX,XX @@ void tcg_optimize(TCGContext *s) ctx.type = TCG_TYPE_I32; } - /* Assume all bits affected, no bits known zero, no sign reps. */ - ctx.z_mask = -1; - ctx.s_mask = 0; - /* * Process each opcode. * Sorted alphabetically by opcode as much as possible. -- 2.43.0 All instances of s_mask have been converted to the new representation. We can now re-enable usage. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_exts(OptContext *ctx, TCGOp *op) g_assert_not_reached(); } - if (0 && !type_change && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { + if (!type_change && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { return true; } @@ -XXX,XX +XXX,XX @@ static bool fold_sextract(OptContext *ctx, TCGOp *op) s_mask = s_mask_old >> pos; s_mask |= -1ull << (len - 1); - if (0 && pos == 0 && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { + if (pos == 0 && fold_affected_mask(ctx, op, s_mask & ~s_mask_old)) { return true; } -- 2.43.0 The big comment just above says functions should be sorted. Add forward declarations as needed. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 114 +++++++++++++++++++++++++------------------------ 1 file changed, 59 insertions(+), 55 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_xx_to_x(OptContext *ctx, TCGOp *op) * 3) those that produce information about the result value. */ +static bool fold_or(OptContext *ctx, TCGOp *op); +static bool fold_orc(OptContext *ctx, TCGOp *op); +static bool fold_xor(OptContext *ctx, TCGOp *op); + static bool fold_add(OptContext *ctx, TCGOp *op) { if (fold_const2_commutative(ctx, op) || @@ -XXX,XX +XXX,XX @@ static bool fold_andc(OptContext *ctx, TCGOp *op) return fold_masks_zs(ctx, op, z_mask, s_mask); } +static bool fold_bitsel_vec(OptContext *ctx, TCGOp *op) +{ + /* If true and false values are the same, eliminate the cmp. */ + if (args_are_copies(op->args[2], op->args[3])) { + return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[2]); + } + + if (arg_is_const(op->args[2]) && arg_is_const(op->args[3])) { + uint64_t tv = arg_info(op->args[2])->val; + uint64_t fv = arg_info(op->args[3])->val; + + if (tv == -1 && fv == 0) { + return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[1]); + } + if (tv == 0 && fv == -1) { + if (TCG_TARGET_HAS_not_vec) { + op->opc = INDEX_op_not_vec; + return fold_not(ctx, op); + } else { + op->opc = INDEX_op_xor_vec; + op->args[2] = arg_new_constant(ctx, -1); + return fold_xor(ctx, op); + } + } + } + if (arg_is_const(op->args[2])) { + uint64_t tv = arg_info(op->args[2])->val; + if (tv == -1) { + op->opc = INDEX_op_or_vec; + op->args[2] = op->args[3]; + return fold_or(ctx, op); + } + if (tv == 0 && TCG_TARGET_HAS_andc_vec) { + op->opc = INDEX_op_andc_vec; + op->args[2] = op->args[1]; + op->args[1] = op->args[3]; + return fold_andc(ctx, op); + } + } + if (arg_is_const(op->args[3])) { + uint64_t fv = arg_info(op->args[3])->val; + if (fv == 0) { + op->opc = INDEX_op_and_vec; + return fold_and(ctx, op); + } + if (fv == -1 && TCG_TARGET_HAS_orc_vec) { + op->opc = INDEX_op_orc_vec; + op->args[2] = op->args[1]; + op->args[1] = op->args[3]; + return fold_orc(ctx, op); + } + } + return finish_folding(ctx, op); +} + static bool fold_brcond(OptContext *ctx, TCGOp *op) { int i = do_constant_folding_cond1(ctx, op, NO_DEST, &op->args[0], @@ -XXX,XX +XXX,XX @@ static bool fold_xor(OptContext *ctx, TCGOp *op) return fold_masks_zs(ctx, op, z_mask, s_mask); } -static bool fold_bitsel_vec(OptContext *ctx, TCGOp *op) -{ - /* If true and false values are the same, eliminate the cmp. */ - if (args_are_copies(op->args[2], op->args[3])) { - return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[2]); - } - - if (arg_is_const(op->args[2]) && arg_is_const(op->args[3])) { - uint64_t tv = arg_info(op->args[2])->val; - uint64_t fv = arg_info(op->args[3])->val; - - if (tv == -1 && fv == 0) { - return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[1]); - } - if (tv == 0 && fv == -1) { - if (TCG_TARGET_HAS_not_vec) { - op->opc = INDEX_op_not_vec; - return fold_not(ctx, op); - } else { - op->opc = INDEX_op_xor_vec; - op->args[2] = arg_new_constant(ctx, -1); - return fold_xor(ctx, op); - } - } - } - if (arg_is_const(op->args[2])) { - uint64_t tv = arg_info(op->args[2])->val; - if (tv == -1) { - op->opc = INDEX_op_or_vec; - op->args[2] = op->args[3]; - return fold_or(ctx, op); - } - if (tv == 0 && TCG_TARGET_HAS_andc_vec) { - op->opc = INDEX_op_andc_vec; - op->args[2] = op->args[1]; - op->args[1] = op->args[3]; - return fold_andc(ctx, op); - } - } - if (arg_is_const(op->args[3])) { - uint64_t fv = arg_info(op->args[3])->val; - if (fv == 0) { - op->opc = INDEX_op_and_vec; - return fold_and(ctx, op); - } - if (fv == -1 && TCG_TARGET_HAS_orc_vec) { - op->opc = INDEX_op_orc_vec; - op->args[2] = op->args[1]; - op->args[1] = op->args[3]; - return fold_orc(ctx, op); - } - } - return finish_folding(ctx, op); -} - /* Propagate constants and copies, fold constant expressions. */ void tcg_optimize(TCGContext *s) { -- 2.43.0 The big comment just above says functions should be sorted. Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- tcg/optimize.c | 60 +++++++++++++++++++++++++------------------------- 1 file changed, 30 insertions(+), 30 deletions(-) diff --git a/tcg/optimize.c b/tcg/optimize.c index XXXXXXX..XXXXXXX 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -XXX,XX +XXX,XX @@ static bool fold_call(OptContext *ctx, TCGOp *op) return true; } +static bool fold_cmp_vec(OptContext *ctx, TCGOp *op) +{ + /* Canonicalize the comparison to put immediate second. */ + if (swap_commutative(NO_DEST, &op->args[1], &op->args[2])) { + op->args[3] = tcg_swap_cond(op->args[3]); + } + return finish_folding(ctx, op); +} + +static bool fold_cmpsel_vec(OptContext *ctx, TCGOp *op) +{ + /* If true and false values are the same, eliminate the cmp. */ + if (args_are_copies(op->args[3], op->args[4])) { + return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[3]); + } + + /* Canonicalize the comparison to put immediate second. */ + if (swap_commutative(NO_DEST, &op->args[1], &op->args[2])) { + op->args[5] = tcg_swap_cond(op->args[5]); + } + /* + * Canonicalize the "false" input reg to match the destination, + * so that the tcg backend can implement "move if true". + */ + if (swap_commutative(op->args[0], &op->args[4], &op->args[3])) { + op->args[5] = tcg_invert_cond(op->args[5]); + } + return finish_folding(ctx, op); +} + static bool fold_count_zeros(OptContext *ctx, TCGOp *op) { uint64_t z_mask, s_mask; @@ -XXX,XX +XXX,XX @@ static bool fold_setcond2(OptContext *ctx, TCGOp *op) return tcg_opt_gen_movi(ctx, op, op->args[0], i); } -static bool fold_cmp_vec(OptContext *ctx, TCGOp *op) -{ - /* Canonicalize the comparison to put immediate second. */ - if (swap_commutative(NO_DEST, &op->args[1], &op->args[2])) { - op->args[3] = tcg_swap_cond(op->args[3]); - } - return finish_folding(ctx, op); -} - -static bool fold_cmpsel_vec(OptContext *ctx, TCGOp *op) -{ - /* If true and false values are the same, eliminate the cmp. */ - if (args_are_copies(op->args[3], op->args[4])) { - return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[3]); - } - - /* Canonicalize the comparison to put immediate second. */ - if (swap_commutative(NO_DEST, &op->args[1], &op->args[2])) { - op->args[5] = tcg_swap_cond(op->args[5]); - } - /* - * Canonicalize the "false" input reg to match the destination, - * so that the tcg backend can implement "move if true". - */ - if (swap_commutative(op->args[0], &op->args[4], &op->args[3])) { - op->args[5] = tcg_invert_cond(op->args[5]); - } - return finish_folding(ctx, op); -} - static bool fold_sextract(OptContext *ctx, TCGOp *op) { uint64_t z_mask, s_mask, s_mask_old; -- 2.43.0 We currently have a flag, float_muladd_halve_result, to scale the result by 2**-1. Extend this to handle arbitrary scaling. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/fpu/softfloat.h | 6 ++++ fpu/softfloat.c | 58 ++++++++++++++++++++++----------------- fpu/softfloat-parts.c.inc | 7 +++-- 3 files changed, 44 insertions(+), 27 deletions(-) diff --git a/include/fpu/softfloat.h b/include/fpu/softfloat.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat.h +++ b/include/fpu/softfloat.h @@ -XXX,XX +XXX,XX @@ float16 float16_add(float16, float16, float_status *status); float16 float16_sub(float16, float16, float_status *status); float16 float16_mul(float16, float16, float_status *status); float16 float16_muladd(float16, float16, float16, int, float_status *status); +float16 float16_muladd_scalbn(float16, float16, float16, + int, int, float_status *status); float16 float16_div(float16, float16, float_status *status); float16 float16_scalbn(float16, int, float_status *status); float16 float16_min(float16, float16, float_status *status); @@ -XXX,XX +XXX,XX @@ float32 float32_mul(float32, float32, float_status *status); float32 float32_div(float32, float32, float_status *status); float32 float32_rem(float32, float32, float_status *status); float32 float32_muladd(float32, float32, float32, int, float_status *status); +float32 float32_muladd_scalbn(float32, float32, float32, + int, int, float_status *status); float32 float32_sqrt(float32, float_status *status); float32 float32_exp2(float32, float_status *status); float32 float32_log2(float32, float_status *status); @@ -XXX,XX +XXX,XX @@ float64 float64_mul(float64, float64, float_status *status); float64 float64_div(float64, float64, float_status *status); float64 float64_rem(float64, float64, float_status *status); float64 float64_muladd(float64, float64, float64, int, float_status *status); +float64 float64_muladd_scalbn(float64, float64, float64, + int, int, float_status *status); float64 float64_sqrt(float64, float_status *status); float64 float64_log2(float64, float_status *status); FloatRelation float64_compare(float64, float64, float_status *status); diff --git a/fpu/softfloat.c b/fpu/softfloat.c index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat.c +++ b/fpu/softfloat.c @@ -XXX,XX +XXX,XX @@ static FloatParts128 *parts128_mul(FloatParts128 *a, FloatParts128 *b, #define parts_mul(A, B, S) \ PARTS_GENERIC_64_128(mul, A)(A, B, S) -static FloatParts64 *parts64_muladd(FloatParts64 *a, FloatParts64 *b, - FloatParts64 *c, int flags, - float_status *s); -static FloatParts128 *parts128_muladd(FloatParts128 *a, FloatParts128 *b, - FloatParts128 *c, int flags, - float_status *s); +static FloatParts64 *parts64_muladd_scalbn(FloatParts64 *a, FloatParts64 *b, + FloatParts64 *c, int scale, + int flags, float_status *s); +static FloatParts128 *parts128_muladd_scalbn(FloatParts128 *a, FloatParts128 *b, + FloatParts128 *c, int scale, + int flags, float_status *s); -#define parts_muladd(A, B, C, Z, S) \ - PARTS_GENERIC_64_128(muladd, A)(A, B, C, Z, S) +#define parts_muladd_scalbn(A, B, C, Z, Y, S) \ + PARTS_GENERIC_64_128(muladd_scalbn, A)(A, B, C, Z, Y, S) static FloatParts64 *parts64_div(FloatParts64 *a, FloatParts64 *b, float_status *s); @@ -XXX,XX +XXX,XX @@ floatx80_mul(floatx80 a, floatx80 b, float_status *status) * Fused multiply-add */ -float16 QEMU_FLATTEN float16_muladd(float16 a, float16 b, float16 c, - int flags, float_status *status) +float16 QEMU_FLATTEN +float16_muladd_scalbn(float16 a, float16 b, float16 c, + int scale, int flags, float_status *status) { FloatParts64 pa, pb, pc, *pr; float16_unpack_canonical(&pa, a, status); float16_unpack_canonical(&pb, b, status); float16_unpack_canonical(&pc, c, status); - pr = parts_muladd(&pa, &pb, &pc, flags, status); + pr = parts_muladd_scalbn(&pa, &pb, &pc, scale, flags, status); return float16_round_pack_canonical(pr, status); } -static float32 QEMU_SOFTFLOAT_ATTR -soft_f32_muladd(float32 a, float32 b, float32 c, int flags, - float_status *status) +float16 float16_muladd(float16 a, float16 b, float16 c, + int flags, float_status *status) +{ + return float16_muladd_scalbn(a, b, c, 0, flags, status); +} + +float32 QEMU_SOFTFLOAT_ATTR +float32_muladd_scalbn(float32 a, float32 b, float32 c, + int scale, int flags, float_status *status) { FloatParts64 pa, pb, pc, *pr; float32_unpack_canonical(&pa, a, status); float32_unpack_canonical(&pb, b, status); float32_unpack_canonical(&pc, c, status); - pr = parts_muladd(&pa, &pb, &pc, flags, status); + pr = parts_muladd_scalbn(&pa, &pb, &pc, scale, flags, status); return float32_round_pack_canonical(pr, status); } -static float64 QEMU_SOFTFLOAT_ATTR -soft_f64_muladd(float64 a, float64 b, float64 c, int flags, - float_status *status) +float64 QEMU_SOFTFLOAT_ATTR +float64_muladd_scalbn(float64 a, float64 b, float64 c, + int scale, int flags, float_status *status) { FloatParts64 pa, pb, pc, *pr; float64_unpack_canonical(&pa, a, status); float64_unpack_canonical(&pb, b, status); float64_unpack_canonical(&pc, c, status); - pr = parts_muladd(&pa, &pb, &pc, flags, status); + pr = parts_muladd_scalbn(&pa, &pb, &pc, scale, flags, status); return float64_round_pack_canonical(pr, status); } @@ -XXX,XX +XXX,XX @@ float32_muladd(float32 xa, float32 xb, float32 xc, int flags, float_status *s) return ur.s; soft: - return soft_f32_muladd(ua.s, ub.s, uc.s, flags, s); + return float32_muladd_scalbn(ua.s, ub.s, uc.s, 0, flags, s); } float64 QEMU_FLATTEN @@ -XXX,XX +XXX,XX @@ float64_muladd(float64 xa, float64 xb, float64 xc, int flags, float_status *s) return ur.s; soft: - return soft_f64_muladd(ua.s, ub.s, uc.s, flags, s); + return float64_muladd_scalbn(ua.s, ub.s, uc.s, 0, flags, s); } float64 float64r32_muladd(float64 a, float64 b, float64 c, @@ -XXX,XX +XXX,XX @@ float64 float64r32_muladd(float64 a, float64 b, float64 c, float64_unpack_canonical(&pa, a, status); float64_unpack_canonical(&pb, b, status); float64_unpack_canonical(&pc, c, status); - pr = parts_muladd(&pa, &pb, &pc, flags, status); + pr = parts_muladd_scalbn(&pa, &pb, &pc, 0, flags, status); return float64r32_round_pack_canonical(pr, status); } @@ -XXX,XX +XXX,XX @@ bfloat16 QEMU_FLATTEN bfloat16_muladd(bfloat16 a, bfloat16 b, bfloat16 c, bfloat16_unpack_canonical(&pa, a, status); bfloat16_unpack_canonical(&pb, b, status); bfloat16_unpack_canonical(&pc, c, status); - pr = parts_muladd(&pa, &pb, &pc, flags, status); + pr = parts_muladd_scalbn(&pa, &pb, &pc, 0, flags, status); return bfloat16_round_pack_canonical(pr, status); } @@ -XXX,XX +XXX,XX @@ float128 QEMU_FLATTEN float128_muladd(float128 a, float128 b, float128 c, float128_unpack_canonical(&pa, a, status); float128_unpack_canonical(&pb, b, status); float128_unpack_canonical(&pc, c, status); - pr = parts_muladd(&pa, &pb, &pc, flags, status); + pr = parts_muladd_scalbn(&pa, &pb, &pc, 0, flags, status); return float128_round_pack_canonical(pr, status); } @@ -XXX,XX +XXX,XX @@ float32 float32_exp2(float32 a, float_status *status) float64_unpack_canonical(&rp, float64_one, status); for (i = 0 ; i < 15 ; i++) { + float64_unpack_canonical(&tp, float32_exp2_coefficients[i], status); - rp = *parts_muladd(&tp, &xnp, &rp, 0, status); + rp = *parts_muladd_scalbn(&tp, &xnp, &rp, 0, 0, status); xnp = *parts_mul(&xnp, &xp, status); } diff --git a/fpu/softfloat-parts.c.inc b/fpu/softfloat-parts.c.inc index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat-parts.c.inc +++ b/fpu/softfloat-parts.c.inc @@ -XXX,XX +XXX,XX @@ static FloatPartsN *partsN(mul)(FloatPartsN *a, FloatPartsN *b, * Requires A and C extracted into a double-sized structure to provide the * extra space for the widening multiply. */ -static FloatPartsN *partsN(muladd)(FloatPartsN *a, FloatPartsN *b, - FloatPartsN *c, int flags, float_status *s) +static FloatPartsN *partsN(muladd_scalbn)(FloatPartsN *a, FloatPartsN *b, + FloatPartsN *c, int scale, + int flags, float_status *s) { int ab_mask, abc_mask; FloatPartsW p_widen, c_widen; @@ -XXX,XX +XXX,XX @@ static FloatPartsN *partsN(muladd)(FloatPartsN *a, FloatPartsN *b, a->exp = p_widen.exp; return_normal: + /* TODO: Replace all use of float_muladd_halve_result with scale. */ if (flags & float_muladd_halve_result) { a->exp -= 1; } + a->exp += scale; finish_sign: if (flags & float_muladd_negate_result) { a->sign ^= 1; -- 2.43.0 Use the scalbn interface instead of float_muladd_halve_result. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/arm/tcg/helper-a64.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/target/arm/tcg/helper-a64.c b/target/arm/tcg/helper-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tcg/helper-a64.c +++ b/target/arm/tcg/helper-a64.c @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(rsqrtsf_f16)(uint32_t a, uint32_t b, float_status *fpst) (float16_is_infinity(b) && float16_is_zero(a))) { return float16_one_point_five; } - return float16_muladd(a, b, float16_three, float_muladd_halve_result, fpst); + return float16_muladd_scalbn(a, b, float16_three, -1, 0, fpst); } float32 HELPER(rsqrtsf_f32)(float32 a, float32 b, float_status *fpst) @@ -XXX,XX +XXX,XX @@ float32 HELPER(rsqrtsf_f32)(float32 a, float32 b, float_status *fpst) (float32_is_infinity(b) && float32_is_zero(a))) { return float32_one_point_five; } - return float32_muladd(a, b, float32_three, float_muladd_halve_result, fpst); + return float32_muladd_scalbn(a, b, float32_three, -1, 0, fpst); } float64 HELPER(rsqrtsf_f64)(float64 a, float64 b, float_status *fpst) @@ -XXX,XX +XXX,XX @@ float64 HELPER(rsqrtsf_f64)(float64 a, float64 b, float_status *fpst) (float64_is_infinity(b) && float64_is_zero(a))) { return float64_one_point_five; } - return float64_muladd(a, b, float64_three, float_muladd_halve_result, fpst); + return float64_muladd_scalbn(a, b, float64_three, -1, 0, fpst); } /* Floating-point reciprocal exponent - see FPRecpX in ARM ARM */ -- 2.43.0 Use the scalbn interface instead of float_muladd_halve_result. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/sparc/helper.h | 4 +- target/sparc/fop_helper.c | 8 ++-- target/sparc/translate.c | 80 +++++++++++++++++++++++---------------- 3 files changed, 54 insertions(+), 38 deletions(-) diff --git a/target/sparc/helper.h b/target/sparc/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/sparc/helper.h +++ b/target/sparc/helper.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_3(faddd, TCG_CALL_NO_WG, f64, env, f64, f64) DEF_HELPER_FLAGS_3(fsubd, TCG_CALL_NO_WG, f64, env, f64, f64) DEF_HELPER_FLAGS_3(fmuld, TCG_CALL_NO_WG, f64, env, f64, f64) DEF_HELPER_FLAGS_3(fdivd, TCG_CALL_NO_WG, f64, env, f64, f64) -DEF_HELPER_FLAGS_5(fmaddd, TCG_CALL_NO_WG, f64, env, f64, f64, f64, i32) +DEF_HELPER_FLAGS_6(fmaddd, TCG_CALL_NO_WG, f64, env, f64, f64, f64, s32, i32) DEF_HELPER_FLAGS_3(fnaddd, TCG_CALL_NO_WG, f64, env, f64, f64) DEF_HELPER_FLAGS_3(fnmuld, TCG_CALL_NO_WG, f64, env, f64, f64) @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_3(fadds, TCG_CALL_NO_WG, f32, env, f32, f32) DEF_HELPER_FLAGS_3(fsubs, TCG_CALL_NO_WG, f32, env, f32, f32) DEF_HELPER_FLAGS_3(fmuls, TCG_CALL_NO_WG, f32, env, f32, f32) DEF_HELPER_FLAGS_3(fdivs, TCG_CALL_NO_WG, f32, env, f32, f32) -DEF_HELPER_FLAGS_5(fmadds, TCG_CALL_NO_WG, f32, env, f32, f32, f32, i32) +DEF_HELPER_FLAGS_6(fmadds, TCG_CALL_NO_WG, f32, env, f32, f32, f32, s32, i32) DEF_HELPER_FLAGS_3(fnadds, TCG_CALL_NO_WG, f32, env, f32, f32) DEF_HELPER_FLAGS_3(fnmuls, TCG_CALL_NO_WG, f32, env, f32, f32) diff --git a/target/sparc/fop_helper.c b/target/sparc/fop_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/sparc/fop_helper.c +++ b/target/sparc/fop_helper.c @@ -XXX,XX +XXX,XX @@ Int128 helper_fsqrtq(CPUSPARCState *env, Int128 src) } float32 helper_fmadds(CPUSPARCState *env, float32 s1, - float32 s2, float32 s3, uint32_t op) + float32 s2, float32 s3, int32_t sc, uint32_t op) { - float32 ret = float32_muladd(s1, s2, s3, op, &env->fp_status); + float32 ret = float32_muladd_scalbn(s1, s2, s3, sc, op, &env->fp_status); check_ieee_exceptions(env, GETPC()); return ret; } float64 helper_fmaddd(CPUSPARCState *env, float64 s1, - float64 s2, float64 s3, uint32_t op) + float64 s2, float64 s3, int32_t sc, uint32_t op) { - float64 ret = float64_muladd(s1, s2, s3, op, &env->fp_status); + float64 ret = float64_muladd_scalbn(s1, s2, s3, sc, op, &env->fp_status); check_ieee_exceptions(env, GETPC()); return ret; } diff --git a/target/sparc/translate.c b/target/sparc/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/sparc/translate.c +++ b/target/sparc/translate.c @@ -XXX,XX +XXX,XX @@ static void gen_op_fabsq(TCGv_i128 dst, TCGv_i128 src) static void gen_op_fmadds(TCGv_i32 d, TCGv_i32 s1, TCGv_i32 s2, TCGv_i32 s3) { - gen_helper_fmadds(d, tcg_env, s1, s2, s3, tcg_constant_i32(0)); + TCGv_i32 z = tcg_constant_i32(0); + gen_helper_fmadds(d, tcg_env, s1, s2, s3, z, z); } static void gen_op_fmaddd(TCGv_i64 d, TCGv_i64 s1, TCGv_i64 s2, TCGv_i64 s3) { - gen_helper_fmaddd(d, tcg_env, s1, s2, s3, tcg_constant_i32(0)); + TCGv_i32 z = tcg_constant_i32(0); + gen_helper_fmaddd(d, tcg_env, s1, s2, s3, z, z); } static void gen_op_fmsubs(TCGv_i32 d, TCGv_i32 s1, TCGv_i32 s2, TCGv_i32 s3) { - int op = float_muladd_negate_c; - gen_helper_fmadds(d, tcg_env, s1, s2, s3, tcg_constant_i32(op)); + TCGv_i32 z = tcg_constant_i32(0); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_c); + gen_helper_fmadds(d, tcg_env, s1, s2, s3, z, op); } static void gen_op_fmsubd(TCGv_i64 d, TCGv_i64 s1, TCGv_i64 s2, TCGv_i64 s3) { - int op = float_muladd_negate_c; - gen_helper_fmaddd(d, tcg_env, s1, s2, s3, tcg_constant_i32(op)); + TCGv_i32 z = tcg_constant_i32(0); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_c); + gen_helper_fmaddd(d, tcg_env, s1, s2, s3, z, op); } static void gen_op_fnmsubs(TCGv_i32 d, TCGv_i32 s1, TCGv_i32 s2, TCGv_i32 s3) { - int op = float_muladd_negate_c | float_muladd_negate_result; - gen_helper_fmadds(d, tcg_env, s1, s2, s3, tcg_constant_i32(op)); + TCGv_i32 z = tcg_constant_i32(0); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_c | + float_muladd_negate_result); + gen_helper_fmadds(d, tcg_env, s1, s2, s3, z, op); } static void gen_op_fnmsubd(TCGv_i64 d, TCGv_i64 s1, TCGv_i64 s2, TCGv_i64 s3) { - int op = float_muladd_negate_c | float_muladd_negate_result; - gen_helper_fmaddd(d, tcg_env, s1, s2, s3, tcg_constant_i32(op)); + TCGv_i32 z = tcg_constant_i32(0); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_c | + float_muladd_negate_result); + gen_helper_fmaddd(d, tcg_env, s1, s2, s3, z, op); } static void gen_op_fnmadds(TCGv_i32 d, TCGv_i32 s1, TCGv_i32 s2, TCGv_i32 s3) { - int op = float_muladd_negate_result; - gen_helper_fmadds(d, tcg_env, s1, s2, s3, tcg_constant_i32(op)); + TCGv_i32 z = tcg_constant_i32(0); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_result); + gen_helper_fmadds(d, tcg_env, s1, s2, s3, z, op); } static void gen_op_fnmaddd(TCGv_i64 d, TCGv_i64 s1, TCGv_i64 s2, TCGv_i64 s3) { - int op = float_muladd_negate_result; - gen_helper_fmaddd(d, tcg_env, s1, s2, s3, tcg_constant_i32(op)); + TCGv_i32 z = tcg_constant_i32(0); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_result); + gen_helper_fmaddd(d, tcg_env, s1, s2, s3, z, op); } /* Use muladd to compute (1 * src1) + src2 / 2 with one rounding. */ static void gen_op_fhadds(TCGv_i32 d, TCGv_i32 s1, TCGv_i32 s2) { - TCGv_i32 one = tcg_constant_i32(float32_one); - int op = float_muladd_halve_result; - gen_helper_fmadds(d, tcg_env, one, s1, s2, tcg_constant_i32(op)); + TCGv_i32 fone = tcg_constant_i32(float32_one); + TCGv_i32 mone = tcg_constant_i32(-1); + TCGv_i32 op = tcg_constant_i32(0); + gen_helper_fmadds(d, tcg_env, fone, s1, s2, mone, op); } static void gen_op_fhaddd(TCGv_i64 d, TCGv_i64 s1, TCGv_i64 s2) { - TCGv_i64 one = tcg_constant_i64(float64_one); - int op = float_muladd_halve_result; - gen_helper_fmaddd(d, tcg_env, one, s1, s2, tcg_constant_i32(op)); + TCGv_i64 fone = tcg_constant_i64(float64_one); + TCGv_i32 mone = tcg_constant_i32(-1); + TCGv_i32 op = tcg_constant_i32(0); + gen_helper_fmaddd(d, tcg_env, fone, s1, s2, mone, op); } /* Use muladd to compute (1 * src1) - src2 / 2 with one rounding. */ static void gen_op_fhsubs(TCGv_i32 d, TCGv_i32 s1, TCGv_i32 s2) { - TCGv_i32 one = tcg_constant_i32(float32_one); - int op = float_muladd_negate_c | float_muladd_halve_result; - gen_helper_fmadds(d, tcg_env, one, s1, s2, tcg_constant_i32(op)); + TCGv_i32 fone = tcg_constant_i32(float32_one); + TCGv_i32 mone = tcg_constant_i32(-1); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_c); + gen_helper_fmadds(d, tcg_env, fone, s1, s2, mone, op); } static void gen_op_fhsubd(TCGv_i64 d, TCGv_i64 s1, TCGv_i64 s2) { - TCGv_i64 one = tcg_constant_i64(float64_one); - int op = float_muladd_negate_c | float_muladd_halve_result; - gen_helper_fmaddd(d, tcg_env, one, s1, s2, tcg_constant_i32(op)); + TCGv_i64 fone = tcg_constant_i64(float64_one); + TCGv_i32 mone = tcg_constant_i32(-1); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_c); + gen_helper_fmaddd(d, tcg_env, fone, s1, s2, mone, op); } /* Use muladd to compute -((1 * src1) + src2 / 2) with one rounding. */ static void gen_op_fnhadds(TCGv_i32 d, TCGv_i32 s1, TCGv_i32 s2) { - TCGv_i32 one = tcg_constant_i32(float32_one); - int op = float_muladd_negate_result | float_muladd_halve_result; - gen_helper_fmadds(d, tcg_env, one, s1, s2, tcg_constant_i32(op)); + TCGv_i32 fone = tcg_constant_i32(float32_one); + TCGv_i32 mone = tcg_constant_i32(-1); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_result); + gen_helper_fmadds(d, tcg_env, fone, s1, s2, mone, op); } static void gen_op_fnhaddd(TCGv_i64 d, TCGv_i64 s1, TCGv_i64 s2) { - TCGv_i64 one = tcg_constant_i64(float64_one); - int op = float_muladd_negate_result | float_muladd_halve_result; - gen_helper_fmaddd(d, tcg_env, one, s1, s2, tcg_constant_i32(op)); + TCGv_i64 fone = tcg_constant_i64(float64_one); + TCGv_i32 mone = tcg_constant_i32(-1); + TCGv_i32 op = tcg_constant_i32(float_muladd_negate_result); + gen_helper_fmaddd(d, tcg_env, fone, s1, s2, mone, op); } static void gen_op_fpexception_im(DisasContext *dc, int ftt) -- 2.43.0 All uses have been convered to float*_muladd_scalbn. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/fpu/softfloat.h | 3 --- fpu/softfloat.c | 6 ------ fpu/softfloat-parts.c.inc | 4 ---- 3 files changed, 13 deletions(-) diff --git a/include/fpu/softfloat.h b/include/fpu/softfloat.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat.h +++ b/include/fpu/softfloat.h @@ -XXX,XX +XXX,XX @@ bfloat16 bfloat16_squash_input_denormal(bfloat16 a, float_status *status); | Using these differs from negating an input or output before calling | the muladd function in that this means that a NaN doesn't have its | sign bit inverted before it is propagated. -| We also support halving the result before rounding, as a special -| case to support the ARM fused-sqrt-step instruction FRSQRTS. *----------------------------------------------------------------------------*/ enum { float_muladd_negate_c = 1, float_muladd_negate_product = 2, float_muladd_negate_result = 4, - float_muladd_halve_result = 8, }; /*---------------------------------------------------------------------------- diff --git a/fpu/softfloat.c b/fpu/softfloat.c index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat.c +++ b/fpu/softfloat.c @@ -XXX,XX +XXX,XX @@ float32_muladd(float32 xa, float32 xb, float32 xc, int flags, float_status *s) if (unlikely(!can_use_fpu(s))) { goto soft; } - if (unlikely(flags & float_muladd_halve_result)) { - goto soft; - } float32_input_flush3(&ua.s, &ub.s, &uc.s, s); if (unlikely(!f32_is_zon3(ua, ub, uc))) { @@ -XXX,XX +XXX,XX @@ float64_muladd(float64 xa, float64 xb, float64 xc, int flags, float_status *s) if (unlikely(!can_use_fpu(s))) { goto soft; } - if (unlikely(flags & float_muladd_halve_result)) { - goto soft; - } float64_input_flush3(&ua.s, &ub.s, &uc.s, s); if (unlikely(!f64_is_zon3(ua, ub, uc))) { diff --git a/fpu/softfloat-parts.c.inc b/fpu/softfloat-parts.c.inc index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat-parts.c.inc +++ b/fpu/softfloat-parts.c.inc @@ -XXX,XX +XXX,XX @@ static FloatPartsN *partsN(muladd_scalbn)(FloatPartsN *a, FloatPartsN *b, a->exp = p_widen.exp; return_normal: - /* TODO: Replace all use of float_muladd_halve_result with scale. */ - if (flags & float_muladd_halve_result) { - a->exp -= 1; - } a->exp += scale; finish_sign: if (flags & float_muladd_negate_result) { -- 2.43.0 This rounding mode is used by Hexagon. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/fpu/softfloat-types.h | 2 ++ fpu/softfloat-parts.c.inc | 3 +++ 2 files changed, 5 insertions(+) diff --git a/include/fpu/softfloat-types.h b/include/fpu/softfloat-types.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat-types.h +++ b/include/fpu/softfloat-types.h @@ -XXX,XX +XXX,XX @@ typedef enum __attribute__((__packed__)) { float_round_to_odd = 5, /* Not an IEEE rounding mode: round to closest odd, overflow to inf */ float_round_to_odd_inf = 6, + /* Not an IEEE rounding mode: round to nearest even, overflow to max */ + float_round_nearest_even_max = 7, } FloatRoundMode; /* diff --git a/fpu/softfloat-parts.c.inc b/fpu/softfloat-parts.c.inc index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat-parts.c.inc +++ b/fpu/softfloat-parts.c.inc @@ -XXX,XX +XXX,XX @@ static void partsN(uncanon_normal)(FloatPartsN *p, float_status *s, int exp, flags = 0; switch (s->float_rounding_mode) { + case float_round_nearest_even_max: + overflow_norm = true; + /* fall through */ case float_round_nearest_even: if (N > 64 && frac_lsb == 0) { inc = ((p->frac_hi & 1) || (p->frac_lo & round_mask) != frac_lsbm1 -- 2.43.0 Certain Hexagon instructions suppress changes to the result when the product of fma() is a true zero. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/fpu/softfloat.h | 5 +++++ fpu/softfloat.c | 3 +++ fpu/softfloat-parts.c.inc | 4 +++- 3 files changed, 11 insertions(+), 1 deletion(-) diff --git a/include/fpu/softfloat.h b/include/fpu/softfloat.h index XXXXXXX..XXXXXXX 100644 --- a/include/fpu/softfloat.h +++ b/include/fpu/softfloat.h @@ -XXX,XX +XXX,XX @@ bfloat16 bfloat16_squash_input_denormal(bfloat16 a, float_status *status); | Using these differs from negating an input or output before calling | the muladd function in that this means that a NaN doesn't have its | sign bit inverted before it is propagated. +| +| With float_muladd_suppress_add_product_zero, if A or B is zero +| such that the product is a true zero, then return C without addition. +| This preserves the sign of C when C is +/- 0. Used for Hexagon. *----------------------------------------------------------------------------*/ enum { float_muladd_negate_c = 1, float_muladd_negate_product = 2, float_muladd_negate_result = 4, + float_muladd_suppress_add_product_zero = 8, }; /*---------------------------------------------------------------------------- diff --git a/fpu/softfloat.c b/fpu/softfloat.c index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat.c +++ b/fpu/softfloat.c @@ -XXX,XX +XXX,XX @@ float32_muladd(float32 xa, float32 xb, float32 xc, int flags, float_status *s) if (unlikely(!can_use_fpu(s))) { goto soft; } + if (unlikely(flags & float_muladd_suppress_add_product_zero)) { + goto soft; + } float32_input_flush3(&ua.s, &ub.s, &uc.s, s); if (unlikely(!f32_is_zon3(ua, ub, uc))) { diff --git a/fpu/softfloat-parts.c.inc b/fpu/softfloat-parts.c.inc index XXXXXXX..XXXXXXX 100644 --- a/fpu/softfloat-parts.c.inc +++ b/fpu/softfloat-parts.c.inc @@ -XXX,XX +XXX,XX @@ static FloatPartsN *partsN(muladd_scalbn)(FloatPartsN *a, FloatPartsN *b, goto return_normal; } if (c->cls == float_class_zero) { - if (a->sign != c->sign) { + if (flags & float_muladd_suppress_add_product_zero) { + a->sign = c->sign; + } else if (a->sign != c->sign) { goto return_sub_zero; } goto return_zero; -- 2.43.0 There are no special cases for this instruction. Remove internal_mpyf as unused. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/fma_emu.h | 1 - target/hexagon/fma_emu.c | 8 -------- target/hexagon/op_helper.c | 2 +- 3 files changed, 1 insertion(+), 10 deletions(-) diff --git a/target/hexagon/fma_emu.h b/target/hexagon/fma_emu.h index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.h +++ b/target/hexagon/fma_emu.h @@ -XXX,XX +XXX,XX @@ int32_t float32_getexp(float32 f32); float32 infinite_float32(uint8_t sign); float32 internal_fmafx(float32 a, float32 b, float32 c, int scale, float_status *fp_status); -float32 internal_mpyf(float32 a, float32 b, float_status *fp_status); float64 internal_mpyhh(float64 a, float64 b, unsigned long long int accumulated, float_status *fp_status); diff --git a/target/hexagon/fma_emu.c b/target/hexagon/fma_emu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.c +++ b/target/hexagon/fma_emu.c @@ -XXX,XX +XXX,XX @@ float32 internal_fmafx(float32 a, float32 b, float32 c, int scale, return accum_round_float32(result, fp_status); } -float32 internal_mpyf(float32 a, float32 b, float_status *fp_status) -{ - if (float32_is_zero(a) || float32_is_zero(b)) { - return float32_mul(a, b, fp_status); - } - return internal_fmafx(a, b, float32_zero, 0, fp_status); -} - float64 internal_mpyhh(float64 a, float64 b, unsigned long long int accumulated, float_status *fp_status) diff --git a/target/hexagon/op_helper.c b/target/hexagon/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/op_helper.c +++ b/target/hexagon/op_helper.c @@ -XXX,XX +XXX,XX @@ float32 HELPER(sfmpy)(CPUHexagonState *env, float32 RsV, float32 RtV) { float32 RdV; arch_fpop_start(env); - RdV = internal_mpyf(RsV, RtV, &env->fp_status); + RdV = float32_mul(RsV, RtV, &env->fp_status); arch_fpop_end(env); return RdV; } -- 2.43.0 There are no special cases for this instruction. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/op_helper.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/target/hexagon/op_helper.c b/target/hexagon/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/op_helper.c +++ b/target/hexagon/op_helper.c @@ -XXX,XX +XXX,XX @@ float32 HELPER(sffma)(CPUHexagonState *env, float32 RxV, float32 RsV, float32 RtV) { arch_fpop_start(env); - RxV = internal_fmafx(RsV, RtV, RxV, 0, &env->fp_status); + RxV = float32_muladd(RsV, RtV, RxV, 0, &env->fp_status); arch_fpop_end(env); return RxV; } -- 2.43.0 There are no special cases for this instruction. Since hexagon always uses default-nan mode, explicitly negating the first input is unnecessary. Use float_muladd_negate_product instead. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/op_helper.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/target/hexagon/op_helper.c b/target/hexagon/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/op_helper.c +++ b/target/hexagon/op_helper.c @@ -XXX,XX +XXX,XX @@ float32 HELPER(sffma_sc)(CPUHexagonState *env, float32 RxV, float32 HELPER(sffms)(CPUHexagonState *env, float32 RxV, float32 RsV, float32 RtV) { - float32 neg_RsV; arch_fpop_start(env); - neg_RsV = float32_set_sign(RsV, float32_is_neg(RsV) ? 0 : 1); - RxV = internal_fmafx(neg_RsV, RtV, RxV, 0, &env->fp_status); + RxV = float32_muladd(RsV, RtV, RxV, float_muladd_negate_product, + &env->fp_status); arch_fpop_end(env); return RxV; } -- 2.43.0 This instruction has a special case that 0 * x + c returns c without the normal sign folding that comes with 0 + -0. Use the new float_muladd_suppress_add_product_zero to describe this. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/op_helper.c | 11 +++-------- 1 file changed, 3 insertions(+), 8 deletions(-) diff --git a/target/hexagon/op_helper.c b/target/hexagon/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/op_helper.c +++ b/target/hexagon/op_helper.c @@ -XXX,XX +XXX,XX @@ static float32 check_nan(float32 dst, float32 x, float_status *fp_status) float32 HELPER(sffma_sc)(CPUHexagonState *env, float32 RxV, float32 RsV, float32 RtV, float32 PuV) { - size4s_t tmp; arch_fpop_start(env); - RxV = check_nan(RxV, RxV, &env->fp_status); - RxV = check_nan(RxV, RsV, &env->fp_status); - RxV = check_nan(RxV, RtV, &env->fp_status); - tmp = internal_fmafx(RsV, RtV, RxV, fSXTN(8, 64, PuV), &env->fp_status); - if (!(float32_is_zero(RxV) && is_zero_prod(RsV, RtV))) { - RxV = tmp; - } + RxV = float32_muladd_scalbn(RsV, RtV, RxV, fSXTN(8, 64, PuV), + float_muladd_suppress_add_product_zero, + &env->fp_status); arch_fpop_end(env); return RxV; } -- 2.43.0 There are multiple special cases for this instruction. (1) The saturate to normal maximum instead of overflow to infinity is handled by the new float_round_nearest_even_max rounding mode. (2) The 0 * n + c special case is handled by the new float_muladd_suppress_add_product_zero flag. (3) The Inf - Inf -> 0 special case can be detected after the fact by examining float_flag_invalid_isi. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/op_helper.c | 105 +++++++++---------------------------- 1 file changed, 26 insertions(+), 79 deletions(-) diff --git a/target/hexagon/op_helper.c b/target/hexagon/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/op_helper.c +++ b/target/hexagon/op_helper.c @@ -XXX,XX +XXX,XX @@ float32 HELPER(sffma)(CPUHexagonState *env, float32 RxV, return RxV; } -static bool is_zero_prod(float32 a, float32 b) -{ - return ((float32_is_zero(a) && is_finite(b)) || - (float32_is_zero(b) && is_finite(a))); -} - -static float32 check_nan(float32 dst, float32 x, float_status *fp_status) -{ - float32 ret = dst; - if (float32_is_any_nan(x)) { - if (extract32(x, 22, 1) == 0) { - float_raise(float_flag_invalid, fp_status); - } - ret = make_float32(0xffffffff); /* nan */ - } - return ret; -} - float32 HELPER(sffma_sc)(CPUHexagonState *env, float32 RxV, float32 RsV, float32 RtV, float32 PuV) { @@ -XXX,XX +XXX,XX @@ float32 HELPER(sffms)(CPUHexagonState *env, float32 RxV, return RxV; } -static bool is_inf_prod(int32_t a, int32_t b) +static float32 do_sffma_lib(CPUHexagonState *env, float32 RxV, + float32 RsV, float32 RtV, int negate) { - return (float32_is_infinity(a) && float32_is_infinity(b)) || - (float32_is_infinity(a) && is_finite(b) && !float32_is_zero(b)) || - (float32_is_infinity(b) && is_finite(a) && !float32_is_zero(a)); + int flags; + + arch_fpop_start(env); + + set_float_rounding_mode(float_round_nearest_even_max, &env->fp_status); + RxV = float32_muladd(RsV, RtV, RxV, + negate | float_muladd_suppress_add_product_zero, + &env->fp_status); + + flags = get_float_exception_flags(&env->fp_status); + if (flags) { + /* Flags are suppressed by this instruction. */ + set_float_exception_flags(0, &env->fp_status); + + /* Return 0 for Inf - Inf. */ + if (flags & float_flag_invalid_isi) { + RxV = 0; + } + } + + arch_fpop_end(env); + return RxV; } float32 HELPER(sffma_lib)(CPUHexagonState *env, float32 RxV, float32 RsV, float32 RtV) { - bool infinp; - bool infminusinf; - float32 tmp; - - arch_fpop_start(env); - set_float_rounding_mode(float_round_nearest_even, &env->fp_status); - infminusinf = float32_is_infinity(RxV) && - is_inf_prod(RsV, RtV) && - (fGETBIT(31, RsV ^ RxV ^ RtV) != 0); - infinp = float32_is_infinity(RxV) || - float32_is_infinity(RtV) || - float32_is_infinity(RsV); - RxV = check_nan(RxV, RxV, &env->fp_status); - RxV = check_nan(RxV, RsV, &env->fp_status); - RxV = check_nan(RxV, RtV, &env->fp_status); - tmp = internal_fmafx(RsV, RtV, RxV, 0, &env->fp_status); - if (!(float32_is_zero(RxV) && is_zero_prod(RsV, RtV))) { - RxV = tmp; - } - set_float_exception_flags(0, &env->fp_status); - if (float32_is_infinity(RxV) && !infinp) { - RxV = RxV - 1; - } - if (infminusinf) { - RxV = 0; - } - arch_fpop_end(env); - return RxV; + return do_sffma_lib(env, RxV, RsV, RtV, 0); } float32 HELPER(sffms_lib)(CPUHexagonState *env, float32 RxV, float32 RsV, float32 RtV) { - bool infinp; - bool infminusinf; - float32 tmp; - - arch_fpop_start(env); - set_float_rounding_mode(float_round_nearest_even, &env->fp_status); - infminusinf = float32_is_infinity(RxV) && - is_inf_prod(RsV, RtV) && - (fGETBIT(31, RsV ^ RxV ^ RtV) == 0); - infinp = float32_is_infinity(RxV) || - float32_is_infinity(RtV) || - float32_is_infinity(RsV); - RxV = check_nan(RxV, RxV, &env->fp_status); - RxV = check_nan(RxV, RsV, &env->fp_status); - RxV = check_nan(RxV, RtV, &env->fp_status); - float32 minus_RsV = float32_sub(float32_zero, RsV, &env->fp_status); - tmp = internal_fmafx(minus_RsV, RtV, RxV, 0, &env->fp_status); - if (!(float32_is_zero(RxV) && is_zero_prod(RsV, RtV))) { - RxV = tmp; - } - set_float_exception_flags(0, &env->fp_status); - if (float32_is_infinity(RxV) && !infinp) { - RxV = RxV - 1; - } - if (infminusinf) { - RxV = 0; - } - arch_fpop_end(env); - return RxV; + return do_sffma_lib(env, RxV, RsV, RtV, float_muladd_negate_product); } float64 HELPER(dfmpyfix)(CPUHexagonState *env, float64 RssV, float64 RttV) -- 2.43.0 The function is now unused. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/fma_emu.h | 2 - target/hexagon/fma_emu.c | 171 --------------------------------------- 2 files changed, 173 deletions(-) diff --git a/target/hexagon/fma_emu.h b/target/hexagon/fma_emu.h index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.h +++ b/target/hexagon/fma_emu.h @@ -XXX,XX +XXX,XX @@ static inline uint32_t float32_getexp_raw(float32 f32) } int32_t float32_getexp(float32 f32); float32 infinite_float32(uint8_t sign); -float32 internal_fmafx(float32 a, float32 b, float32 c, - int scale, float_status *fp_status); float64 internal_mpyhh(float64 a, float64 b, unsigned long long int accumulated, float_status *fp_status); diff --git a/target/hexagon/fma_emu.c b/target/hexagon/fma_emu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.c +++ b/target/hexagon/fma_emu.c @@ -XXX,XX +XXX,XX @@ int32_t float64_getexp(float64 f64) return -1; } -static uint64_t float32_getmant(float32 f32) -{ - Float a = { .i = f32 }; - if (float32_is_normal(f32)) { - return a.mant | 1ULL << 23; - } - if (float32_is_zero(f32)) { - return 0; - } - if (float32_is_denormal(f32)) { - return a.mant; - } - return ~0ULL; -} - int32_t float32_getexp(float32 f32) { Float a = { .i = f32 }; @@ -XXX,XX +XXX,XX @@ float32 infinite_float32(uint8_t sign) } /* Return a maximum finite value with the requested sign */ -static float32 maxfinite_float32(uint8_t sign) -{ - if (sign) { - return make_float32(SF_MINUS_MAXF); - } else { - return make_float32(SF_MAXF); - } -} - -/* Return a zero value with requested sign */ -static float32 zero_float32(uint8_t sign) -{ - if (sign) { - return make_float32(0x80000000); - } else { - return float32_zero; - } -} - #define GEN_XF_ROUND(SUFFIX, MANTBITS, INF_EXP, INTERNAL_TYPE) \ static SUFFIX accum_round_##SUFFIX(Accum a, float_status * fp_status) \ { \ @@ -XXX,XX +XXX,XX @@ static SUFFIX accum_round_##SUFFIX(Accum a, float_status * fp_status) \ } GEN_XF_ROUND(float64, DF_MANTBITS, DF_INF_EXP, Double) -GEN_XF_ROUND(float32, SF_MANTBITS, SF_INF_EXP, Float) - -static bool is_inf_prod(float64 a, float64 b) -{ - return ((float64_is_infinity(a) && float64_is_infinity(b)) || - (float64_is_infinity(a) && is_finite(b) && (!float64_is_zero(b))) || - (float64_is_infinity(b) && is_finite(a) && (!float64_is_zero(a)))); -} - -static float64 special_fma(float64 a, float64 b, float64 c, - float_status *fp_status) -{ - float64 ret = make_float64(0); - - /* - * If A multiplied by B is an exact infinity and C is also an infinity - * but with the opposite sign, FMA returns NaN and raises invalid. - */ - uint8_t a_sign = float64_is_neg(a); - uint8_t b_sign = float64_is_neg(b); - uint8_t c_sign = float64_is_neg(c); - if (is_inf_prod(a, b) && float64_is_infinity(c)) { - if ((a_sign ^ b_sign) != c_sign) { - ret = make_float64(DF_NAN); - float_raise(float_flag_invalid, fp_status); - return ret; - } - } - if ((float64_is_infinity(a) && float64_is_zero(b)) || - (float64_is_zero(a) && float64_is_infinity(b))) { - ret = make_float64(DF_NAN); - float_raise(float_flag_invalid, fp_status); - return ret; - } - /* - * If none of the above checks are true and C is a NaN, - * a NaN shall be returned - * If A or B are NaN, a NAN shall be returned. - */ - if (float64_is_any_nan(a) || - float64_is_any_nan(b) || - float64_is_any_nan(c)) { - if (float64_is_any_nan(a) && (fGETBIT(51, a) == 0)) { - float_raise(float_flag_invalid, fp_status); - } - if (float64_is_any_nan(b) && (fGETBIT(51, b) == 0)) { - float_raise(float_flag_invalid, fp_status); - } - if (float64_is_any_nan(c) && (fGETBIT(51, c) == 0)) { - float_raise(float_flag_invalid, fp_status); - } - ret = make_float64(DF_NAN); - return ret; - } - /* - * We have checked for adding opposite-signed infinities. - * Other infinities return infinity with the correct sign - */ - if (float64_is_infinity(c)) { - ret = infinite_float64(c_sign); - return ret; - } - if (float64_is_infinity(a) || float64_is_infinity(b)) { - ret = infinite_float64(a_sign ^ b_sign); - return ret; - } - g_assert_not_reached(); -} - -static float32 special_fmaf(float32 a, float32 b, float32 c, - float_status *fp_status) -{ - float64 aa, bb, cc; - aa = float32_to_float64(a, fp_status); - bb = float32_to_float64(b, fp_status); - cc = float32_to_float64(c, fp_status); - return float64_to_float32(special_fma(aa, bb, cc, fp_status), fp_status); -} - -float32 internal_fmafx(float32 a, float32 b, float32 c, int scale, - float_status *fp_status) -{ - Accum prod; - Accum acc; - Accum result; - accum_init(&prod); - accum_init(&acc); - accum_init(&result); - - uint8_t a_sign = float32_is_neg(a); - uint8_t b_sign = float32_is_neg(b); - uint8_t c_sign = float32_is_neg(c); - if (float32_is_infinity(a) || - float32_is_infinity(b) || - float32_is_infinity(c)) { - return special_fmaf(a, b, c, fp_status); - } - if (float32_is_any_nan(a) || - float32_is_any_nan(b) || - float32_is_any_nan(c)) { - return special_fmaf(a, b, c, fp_status); - } - if ((scale == 0) && (float32_is_zero(a) || float32_is_zero(b))) { - float32 tmp = float32_mul(a, b, fp_status); - tmp = float32_add(tmp, c, fp_status); - return tmp; - } - - /* (a * 2**b) * (c * 2**d) == a*c * 2**(b+d) */ - prod.mant = int128_mul_6464(float32_getmant(a), float32_getmant(b)); - - /* - * Note: extracting the mantissa into an int is multiplying by - * 2**23, so adjust here - */ - prod.exp = float32_getexp(a) + float32_getexp(b) - SF_BIAS - 23; - prod.sign = a_sign ^ b_sign; - if (float32_is_zero(a) || float32_is_zero(b)) { - prod.exp = -2 * WAY_BIG_EXP; - } - if ((scale > 0) && float32_is_denormal(c)) { - acc.mant = int128_mul_6464(0, 0); - acc.exp = -WAY_BIG_EXP; - acc.sign = c_sign; - acc.sticky = 1; - result = accum_add(prod, acc); - } else if (!float32_is_zero(c)) { - acc.mant = int128_mul_6464(float32_getmant(c), 1); - acc.exp = float32_getexp(c); - acc.sign = c_sign; - result = accum_add(prod, acc); - } else { - result = prod; - } - result.exp += scale; - return accum_round_float32(result, fp_status); -} float64 internal_mpyhh(float64 a, float64 b, unsigned long long int accumulated, -- 2.43.0 This massive macro is now only used once. Expand it for use only by float64. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/fma_emu.c | 255 +++++++++++++++++++-------------------- 1 file changed, 127 insertions(+), 128 deletions(-) diff --git a/target/hexagon/fma_emu.c b/target/hexagon/fma_emu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.c +++ b/target/hexagon/fma_emu.c @@ -XXX,XX +XXX,XX @@ float32 infinite_float32(uint8_t sign) } /* Return a maximum finite value with the requested sign */ -#define GEN_XF_ROUND(SUFFIX, MANTBITS, INF_EXP, INTERNAL_TYPE) \ -static SUFFIX accum_round_##SUFFIX(Accum a, float_status * fp_status) \ -{ \ - if ((int128_gethi(a.mant) == 0) && (int128_getlo(a.mant) == 0) \ - && ((a.guard | a.round | a.sticky) == 0)) { \ - /* result zero */ \ - switch (fp_status->float_rounding_mode) { \ - case float_round_down: \ - return zero_##SUFFIX(1); \ - default: \ - return zero_##SUFFIX(0); \ - } \ - } \ - /* Normalize right */ \ - /* We want MANTBITS bits of mantissa plus the leading one. */ \ - /* That means that we want MANTBITS+1 bits, or 0x000000000000FF_FFFF */ \ - /* So we need to normalize right while the high word is non-zero and \ - * while the low word is nonzero when masked with 0xffe0_0000_0000_0000 */ \ - while ((int128_gethi(a.mant) != 0) || \ - ((int128_getlo(a.mant) >> (MANTBITS + 1)) != 0)) { \ - a = accum_norm_right(a, 1); \ - } \ - /* \ - * OK, now normalize left \ - * We want to normalize left until we have a leading one in bit 24 \ - * Theoretically, we only need to shift a maximum of one to the left if we \ - * shifted out lots of bits from B, or if we had no shift / 1 shift sticky \ - * should be 0 \ - */ \ - while ((int128_getlo(a.mant) & (1ULL << MANTBITS)) == 0) { \ - a = accum_norm_left(a); \ - } \ - /* \ - * OK, now we might need to denormalize because of potential underflow. \ - * We need to do this before rounding, and rounding might make us normal \ - * again \ - */ \ - while (a.exp <= 0) { \ - a = accum_norm_right(a, 1 - a.exp); \ - /* \ - * Do we have underflow? \ - * That's when we get an inexact answer because we ran out of bits \ - * in a denormal. \ - */ \ - if (a.guard || a.round || a.sticky) { \ - float_raise(float_flag_underflow, fp_status); \ - } \ - } \ - /* OK, we're relatively canonical... now we need to round */ \ - if (a.guard || a.round || a.sticky) { \ - float_raise(float_flag_inexact, fp_status); \ - switch (fp_status->float_rounding_mode) { \ - case float_round_to_zero: \ - /* Chop and we're done */ \ - break; \ - case float_round_up: \ - if (a.sign == 0) { \ - a.mant = int128_add(a.mant, int128_one()); \ - } \ - break; \ - case float_round_down: \ - if (a.sign != 0) { \ - a.mant = int128_add(a.mant, int128_one()); \ - } \ - break; \ - default: \ - if (a.round || a.sticky) { \ - /* round up if guard is 1, down if guard is zero */ \ - a.mant = int128_add(a.mant, int128_make64(a.guard)); \ - } else if (a.guard) { \ - /* exactly .5, round up if odd */ \ - a.mant = int128_add(a.mant, int128_and(a.mant, int128_one())); \ - } \ - break; \ - } \ - } \ - /* \ - * OK, now we might have carried all the way up. \ - * So we might need to shr once \ - * at least we know that the lsb should be zero if we rounded and \ - * got a carry out... \ - */ \ - if ((int128_getlo(a.mant) >> (MANTBITS + 1)) != 0) { \ - a = accum_norm_right(a, 1); \ - } \ - /* Overflow? */ \ - if (a.exp >= INF_EXP) { \ - /* Yep, inf result */ \ - float_raise(float_flag_overflow, fp_status); \ - float_raise(float_flag_inexact, fp_status); \ - switch (fp_status->float_rounding_mode) { \ - case float_round_to_zero: \ - return maxfinite_##SUFFIX(a.sign); \ - case float_round_up: \ - if (a.sign == 0) { \ - return infinite_##SUFFIX(a.sign); \ - } else { \ - return maxfinite_##SUFFIX(a.sign); \ - } \ - case float_round_down: \ - if (a.sign != 0) { \ - return infinite_##SUFFIX(a.sign); \ - } else { \ - return maxfinite_##SUFFIX(a.sign); \ - } \ - default: \ - return infinite_##SUFFIX(a.sign); \ - } \ - } \ - /* Underflow? */ \ - if (int128_getlo(a.mant) & (1ULL << MANTBITS)) { \ - /* Leading one means: No, we're normal. So, we should be done... */ \ - INTERNAL_TYPE ret; \ - ret.i = 0; \ - ret.sign = a.sign; \ - ret.exp = a.exp; \ - ret.mant = int128_getlo(a.mant); \ - return ret.i; \ - } \ - assert(a.exp == 1); \ - INTERNAL_TYPE ret; \ - ret.i = 0; \ - ret.sign = a.sign; \ - ret.exp = 0; \ - ret.mant = int128_getlo(a.mant); \ - return ret.i; \ +static float64 accum_round_float64(Accum a, float_status *fp_status) +{ + if ((int128_gethi(a.mant) == 0) && (int128_getlo(a.mant) == 0) + && ((a.guard | a.round | a.sticky) == 0)) { + /* result zero */ + switch (fp_status->float_rounding_mode) { + case float_round_down: + return zero_float64(1); + default: + return zero_float64(0); + } + } + /* + * Normalize right + * We want DF_MANTBITS bits of mantissa plus the leading one. + * That means that we want DF_MANTBITS+1 bits, or 0x000000000000FF_FFFF + * So we need to normalize right while the high word is non-zero and + * while the low word is nonzero when masked with 0xffe0_0000_0000_0000 + */ + while ((int128_gethi(a.mant) != 0) || + ((int128_getlo(a.mant) >> (DF_MANTBITS + 1)) != 0)) { + a = accum_norm_right(a, 1); + } + /* + * OK, now normalize left + * We want to normalize left until we have a leading one in bit 24 + * Theoretically, we only need to shift a maximum of one to the left if we + * shifted out lots of bits from B, or if we had no shift / 1 shift sticky + * should be 0 + */ + while ((int128_getlo(a.mant) & (1ULL << DF_MANTBITS)) == 0) { + a = accum_norm_left(a); + } + /* + * OK, now we might need to denormalize because of potential underflow. + * We need to do this before rounding, and rounding might make us normal + * again + */ + while (a.exp <= 0) { + a = accum_norm_right(a, 1 - a.exp); + /* + * Do we have underflow? + * That's when we get an inexact answer because we ran out of bits + * in a denormal. + */ + if (a.guard || a.round || a.sticky) { + float_raise(float_flag_underflow, fp_status); + } + } + /* OK, we're relatively canonical... now we need to round */ + if (a.guard || a.round || a.sticky) { + float_raise(float_flag_inexact, fp_status); + switch (fp_status->float_rounding_mode) { + case float_round_to_zero: + /* Chop and we're done */ + break; + case float_round_up: + if (a.sign == 0) { + a.mant = int128_add(a.mant, int128_one()); + } + break; + case float_round_down: + if (a.sign != 0) { + a.mant = int128_add(a.mant, int128_one()); + } + break; + default: + if (a.round || a.sticky) { + /* round up if guard is 1, down if guard is zero */ + a.mant = int128_add(a.mant, int128_make64(a.guard)); + } else if (a.guard) { + /* exactly .5, round up if odd */ + a.mant = int128_add(a.mant, int128_and(a.mant, int128_one())); + } + break; + } + } + /* + * OK, now we might have carried all the way up. + * So we might need to shr once + * at least we know that the lsb should be zero if we rounded and + * got a carry out... + */ + if ((int128_getlo(a.mant) >> (DF_MANTBITS + 1)) != 0) { + a = accum_norm_right(a, 1); + } + /* Overflow? */ + if (a.exp >= DF_INF_EXP) { + /* Yep, inf result */ + float_raise(float_flag_overflow, fp_status); + float_raise(float_flag_inexact, fp_status); + switch (fp_status->float_rounding_mode) { + case float_round_to_zero: + return maxfinite_float64(a.sign); + case float_round_up: + if (a.sign == 0) { + return infinite_float64(a.sign); + } else { + return maxfinite_float64(a.sign); + } + case float_round_down: + if (a.sign != 0) { + return infinite_float64(a.sign); + } else { + return maxfinite_float64(a.sign); + } + default: + return infinite_float64(a.sign); + } + } + /* Underflow? */ + if (int128_getlo(a.mant) & (1ULL << DF_MANTBITS)) { + /* Leading one means: No, we're normal. So, we should be done... */ + Double ret; + ret.i = 0; + ret.sign = a.sign; + ret.exp = a.exp; + ret.mant = int128_getlo(a.mant); + return ret.i; + } + assert(a.exp == 1); + Double ret; + ret.i = 0; + ret.sign = a.sign; + ret.exp = 0; + ret.mant = int128_getlo(a.mant); + return ret.i; } -GEN_XF_ROUND(float64, DF_MANTBITS, DF_INF_EXP, Double) - float64 internal_mpyhh(float64 a, float64 b, unsigned long long int accumulated, float_status *fp_status) -- 2.43.0 This structure, with bitfields, is incorrect for big-endian. Use the existing float32_getexp_raw which uses extract32. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/fma_emu.c | 16 +++------------- 1 file changed, 3 insertions(+), 13 deletions(-) diff --git a/target/hexagon/fma_emu.c b/target/hexagon/fma_emu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.c +++ b/target/hexagon/fma_emu.c @@ -XXX,XX +XXX,XX @@ typedef union { }; } Double; -typedef union { - float f; - uint32_t i; - struct { - uint32_t mant:23; - uint32_t exp:8; - uint32_t sign:1; - }; -} Float; - static uint64_t float64_getmant(float64 f64) { Double a = { .i = f64 }; @@ -XXX,XX +XXX,XX @@ int32_t float64_getexp(float64 f64) int32_t float32_getexp(float32 f32) { - Float a = { .i = f32 }; + int exp = float32_getexp_raw(f32); if (float32_is_normal(f32)) { - return a.exp; + return exp; } if (float32_is_denormal(f32)) { - return a.exp + 1; + return exp + 1; } return -1; } -- 2.43.0 This structure, with bitfields, is incorrect for big-endian. Use extract64 and deposit64 instead. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/fma_emu.c | 46 ++++++++++++++-------------------------- 1 file changed, 16 insertions(+), 30 deletions(-) diff --git a/target/hexagon/fma_emu.c b/target/hexagon/fma_emu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.c +++ b/target/hexagon/fma_emu.c @@ -XXX,XX +XXX,XX @@ #define WAY_BIG_EXP 4096 -typedef union { - double f; - uint64_t i; - struct { - uint64_t mant:52; - uint64_t exp:11; - uint64_t sign:1; - }; -} Double; - static uint64_t float64_getmant(float64 f64) { - Double a = { .i = f64 }; + uint64_t mant = extract64(f64, 0, 52); if (float64_is_normal(f64)) { - return a.mant | 1ULL << 52; + return mant | 1ULL << 52; } if (float64_is_zero(f64)) { return 0; } if (float64_is_denormal(f64)) { - return a.mant; + return mant; } return ~0ULL; } int32_t float64_getexp(float64 f64) { - Double a = { .i = f64 }; + int exp = extract64(f64, 52, 11); if (float64_is_normal(f64)) { - return a.exp; + return exp; } if (float64_is_denormal(f64)) { - return a.exp + 1; + return exp + 1; } return -1; } @@ -XXX,XX +XXX,XX @@ float32 infinite_float32(uint8_t sign) /* Return a maximum finite value with the requested sign */ static float64 accum_round_float64(Accum a, float_status *fp_status) { + uint64_t ret; + if ((int128_gethi(a.mant) == 0) && (int128_getlo(a.mant) == 0) && ((a.guard | a.round | a.sticky) == 0)) { /* result zero */ @@ -XXX,XX +XXX,XX @@ static float64 accum_round_float64(Accum a, float_status *fp_status) } } /* Underflow? */ - if (int128_getlo(a.mant) & (1ULL << DF_MANTBITS)) { + ret = int128_getlo(a.mant); + if (ret & (1ULL << DF_MANTBITS)) { /* Leading one means: No, we're normal. So, we should be done... */ - Double ret; - ret.i = 0; - ret.sign = a.sign; - ret.exp = a.exp; - ret.mant = int128_getlo(a.mant); - return ret.i; + ret = deposit64(ret, 52, 11, a.exp); + } else { + assert(a.exp == 1); + ret = deposit64(ret, 52, 11, 0); } - assert(a.exp == 1); - Double ret; - ret.i = 0; - ret.sign = a.sign; - ret.exp = 0; - ret.mant = int128_getlo(a.mant); - return ret.i; + ret = deposit64(ret, 63, 1, a.sign); + return ret; } float64 internal_mpyhh(float64 a, float64 b, -- 2.43.0 No need to open-code 64x64->128-bit multiplication. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/fma_emu.c | 32 +++----------------------------- 1 file changed, 3 insertions(+), 29 deletions(-) diff --git a/target/hexagon/fma_emu.c b/target/hexagon/fma_emu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.c +++ b/target/hexagon/fma_emu.c @@ -XXX,XX +XXX,XX @@ int32_t float32_getexp(float32 f32) return -1; } -static uint32_t int128_getw0(Int128 x) -{ - return int128_getlo(x); -} - -static uint32_t int128_getw1(Int128 x) -{ - return int128_getlo(x) >> 32; -} - static Int128 int128_mul_6464(uint64_t ai, uint64_t bi) { - Int128 a, b; - uint64_t pp0, pp1a, pp1b, pp1s, pp2; + uint64_t l, h; - a = int128_make64(ai); - b = int128_make64(bi); - pp0 = (uint64_t)int128_getw0(a) * (uint64_t)int128_getw0(b); - pp1a = (uint64_t)int128_getw1(a) * (uint64_t)int128_getw0(b); - pp1b = (uint64_t)int128_getw1(b) * (uint64_t)int128_getw0(a); - pp2 = (uint64_t)int128_getw1(a) * (uint64_t)int128_getw1(b); - - pp1s = pp1a + pp1b; - if ((pp1s < pp1a) || (pp1s < pp1b)) { - pp2 += (1ULL << 32); - } - uint64_t ret_low = pp0 + (pp1s << 32); - if ((ret_low < pp0) || (ret_low < (pp1s << 32))) { - pp2 += 1; - } - - return int128_make128(ret_low, pp2 + (pp1s >> 32)); + mulu64(&l, &h, ai, bi); + return int128_make128(l, h); } static Int128 int128_sub_borrow(Int128 a, Int128 b, int borrow) -- 2.43.0 Initialize x with accumulated via direct assignment, rather than multiplying by 1. Reviewed-by: Brian Cain <brian.cain@oss.qualcomm.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- target/hexagon/fma_emu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/target/hexagon/fma_emu.c b/target/hexagon/fma_emu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/fma_emu.c +++ b/target/hexagon/fma_emu.c @@ -XXX,XX +XXX,XX @@ float64 internal_mpyhh(float64 a, float64 b, float64_is_infinity(b)) { return float64_mul(a, b, fp_status); } - x.mant = int128_mul_6464(accumulated, 1); + x.mant = int128_make64(accumulated); x.sticky = sticky; prod = fGETUWORD(1, float64_getmant(a)) * fGETUWORD(1, float64_getmant(b)); x.mant = int128_add(x.mant, int128_mul_6464(prod, 0x100000000ULL)); -- 2.43.0 Convert all targets simultaneously, as the gen_intermediate_code function disappears from the target. While there are possible workarounds, they're larger than simply performing the conversion. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> --- include/exec/translator.h | 14 -------------- include/hw/core/tcg-cpu-ops.h | 13 +++++++++++++ target/alpha/cpu.h | 2 ++ target/arm/internals.h | 2 ++ target/avr/cpu.h | 2 ++ target/hexagon/cpu.h | 2 ++ target/hppa/cpu.h | 2 ++ target/i386/tcg/helper-tcg.h | 2 ++ target/loongarch/internals.h | 2 ++ target/m68k/cpu.h | 2 ++ target/microblaze/cpu.h | 2 ++ target/mips/tcg/tcg-internal.h | 2 ++ target/openrisc/cpu.h | 2 ++ target/ppc/cpu.h | 2 ++ target/riscv/cpu.h | 3 +++ target/rx/cpu.h | 2 ++ target/s390x/s390x-internal.h | 2 ++ target/sh4/cpu.h | 2 ++ target/sparc/cpu.h | 2 ++ target/tricore/cpu.h | 2 ++ target/xtensa/cpu.h | 2 ++ accel/tcg/cpu-exec.c | 8 +++++--- accel/tcg/translate-all.c | 8 +++++--- target/alpha/cpu.c | 1 + target/alpha/translate.c | 4 ++-- target/arm/cpu.c | 1 + target/arm/tcg/cpu-v7m.c | 1 + target/arm/tcg/translate.c | 5 ++--- target/avr/cpu.c | 1 + target/avr/translate.c | 6 +++--- target/hexagon/cpu.c | 1 + target/hexagon/translate.c | 4 ++-- target/hppa/cpu.c | 1 + target/hppa/translate.c | 4 ++-- target/i386/tcg/tcg-cpu.c | 1 + target/i386/tcg/translate.c | 5 ++--- target/loongarch/cpu.c | 1 + target/loongarch/tcg/translate.c | 4 ++-- target/m68k/cpu.c | 1 + target/m68k/translate.c | 4 ++-- target/microblaze/cpu.c | 1 + target/microblaze/translate.c | 4 ++-- target/mips/cpu.c | 1 + target/mips/tcg/translate.c | 4 ++-- target/openrisc/cpu.c | 1 + target/openrisc/translate.c | 4 ++-- target/ppc/cpu_init.c | 1 + target/ppc/translate.c | 4 ++-- target/riscv/tcg/tcg-cpu.c | 1 + target/riscv/translate.c | 4 ++-- target/rx/cpu.c | 1 + target/rx/translate.c | 4 ++-- target/s390x/cpu.c | 1 + target/s390x/tcg/translate.c | 4 ++-- target/sh4/cpu.c | 1 + target/sh4/translate.c | 4 ++-- target/sparc/cpu.c | 1 + target/sparc/translate.c | 4 ++-- target/tricore/cpu.c | 1 + target/tricore/translate.c | 5 ++--- target/xtensa/cpu.c | 1 + target/xtensa/translate.c | 4 ++-- 62 files changed, 121 insertions(+), 62 deletions(-) diff --git a/include/exec/translator.h b/include/exec/translator.h index XXXXXXX..XXXXXXX 100644 --- a/include/exec/translator.h +++ b/include/exec/translator.h @@ -XXX,XX +XXX,XX @@ #include "qemu/bswap.h" #include "exec/vaddr.h" -/** - * gen_intermediate_code - * @cpu: cpu context - * @tb: translation block - * @max_insns: max number of instructions to translate - * @pc: guest virtual program counter address - * @host_pc: host physical program counter address - * - * This function must be provided by the target, which should create - * the target-specific DisasContext, and then invoke translator_loop. - */ -void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc); - /** * DisasJumpType: * @DISAS_NEXT: Next instruction in program order. diff --git a/include/hw/core/tcg-cpu-ops.h b/include/hw/core/tcg-cpu-ops.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/core/tcg-cpu-ops.h +++ b/include/hw/core/tcg-cpu-ops.h @@ -XXX,XX +XXX,XX @@ struct TCGCPUOps { * Called when the first CPU is realized. */ void (*initialize)(void); + /** + * @translate_code: Translate guest instructions to TCGOps + * @cpu: cpu context + * @tb: translation block + * @max_insns: max number of instructions to translate + * @pc: guest virtual program counter address + * @host_pc: host physical program counter address + * + * This function must be provided by the target, which should create + * the target-specific DisasContext, and then invoke translator_loop. + */ + void (*translate_code)(CPUState *cpu, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); /** * @synchronize_from_tb: Synchronize state from a TCG #TranslationBlock * diff --git a/target/alpha/cpu.h b/target/alpha/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/alpha/cpu.h +++ b/target/alpha/cpu.h @@ -XXX,XX +XXX,XX @@ enum { }; void alpha_translate_init(void); +void alpha_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); #define CPU_RESOLVING_TYPE TYPE_ALPHA_CPU diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ void init_cpreg_list(ARMCPU *cpu); void arm_cpu_register_gdb_regs_for_features(ARMCPU *cpu); void arm_translate_init(void); +void arm_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); void arm_cpu_register_gdb_commands(ARMCPU *cpu); void aarch64_cpu_register_gdb_commands(ARMCPU *cpu, GString *, diff --git a/target/avr/cpu.h b/target/avr/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/avr/cpu.h +++ b/target/avr/cpu.h @@ -XXX,XX +XXX,XX @@ static inline void set_avr_feature(CPUAVRState *env, int feature) } void avr_cpu_tcg_init(void); +void avr_cpu_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); int cpu_avr_exec(CPUState *cpu); diff --git a/target/hexagon/cpu.h b/target/hexagon/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/cpu.h +++ b/target/hexagon/cpu.h @@ -XXX,XX +XXX,XX @@ static inline void cpu_get_tb_cpu_state(CPUHexagonState *env, vaddr *pc, typedef HexagonCPU ArchCPU; void hexagon_translate_init(void); +void hexagon_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); #include "exec/cpu-all.h" diff --git a/target/hppa/cpu.h b/target/hppa/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/hppa/cpu.h +++ b/target/hppa/cpu.h @@ -XXX,XX +XXX,XX @@ static inline int HPPA_BTLB_ENTRIES(CPUHPPAState *env) } void hppa_translate_init(void); +void hppa_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); #define CPU_RESOLVING_TYPE TYPE_HPPA_CPU diff --git a/target/i386/tcg/helper-tcg.h b/target/i386/tcg/helper-tcg.h index XXXXXXX..XXXXXXX 100644 --- a/target/i386/tcg/helper-tcg.h +++ b/target/i386/tcg/helper-tcg.h @@ -XXX,XX +XXX,XX @@ static inline target_long lshift(target_long x, int n) /* translate.c */ void tcg_x86_init(void); +void x86_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); /* excp_helper.c */ G_NORETURN void raise_exception(CPUX86State *env, int exception_index); diff --git a/target/loongarch/internals.h b/target/loongarch/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/loongarch/internals.h +++ b/target/loongarch/internals.h @@ -XXX,XX +XXX,XX @@ #define TARGET_VIRT_MASK MAKE_64BIT_MASK(0, TARGET_VIRT_ADDR_SPACE_BITS) void loongarch_translate_init(void); +void loongarch_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); void G_NORETURN do_raise_exception(CPULoongArchState *env, uint32_t exception, diff --git a/target/m68k/cpu.h b/target/m68k/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/m68k/cpu.h +++ b/target/m68k/cpu.h @@ -XXX,XX +XXX,XX @@ int m68k_cpu_gdb_read_register(CPUState *cpu, GByteArray *buf, int reg); int m68k_cpu_gdb_write_register(CPUState *cpu, uint8_t *buf, int reg); void m68k_tcg_init(void); +void m68k_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); void m68k_cpu_init_gdb(M68kCPU *cpu); uint32_t cpu_m68k_get_ccr(CPUM68KState *env); void cpu_m68k_set_ccr(CPUM68KState *env, uint32_t); diff --git a/target/microblaze/cpu.h b/target/microblaze/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/microblaze/cpu.h +++ b/target/microblaze/cpu.h @@ -XXX,XX +XXX,XX @@ static inline void mb_cpu_write_msr(CPUMBState *env, uint32_t val) } void mb_tcg_init(void); +void mb_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); #define CPU_RESOLVING_TYPE TYPE_MICROBLAZE_CPU diff --git a/target/mips/tcg/tcg-internal.h b/target/mips/tcg/tcg-internal.h index XXXXXXX..XXXXXXX 100644 --- a/target/mips/tcg/tcg-internal.h +++ b/target/mips/tcg/tcg-internal.h @@ -XXX,XX +XXX,XX @@ #include "cpu.h" void mips_tcg_init(void); +void mips_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); void mips_cpu_synchronize_from_tb(CPUState *cs, const TranslationBlock *tb); G_NORETURN void mips_cpu_do_unaligned_access(CPUState *cpu, vaddr addr, diff --git a/target/openrisc/cpu.h b/target/openrisc/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/openrisc/cpu.h +++ b/target/openrisc/cpu.h @@ -XXX,XX +XXX,XX @@ void openrisc_cpu_dump_state(CPUState *cpu, FILE *f, int flags); int openrisc_cpu_gdb_read_register(CPUState *cpu, GByteArray *buf, int reg); int openrisc_cpu_gdb_write_register(CPUState *cpu, uint8_t *buf, int reg); void openrisc_translate_init(void); +void openrisc_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); int print_insn_or1k(bfd_vma addr, disassemble_info *info); #ifndef CONFIG_USER_ONLY diff --git a/target/ppc/cpu.h b/target/ppc/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/ppc/cpu.h +++ b/target/ppc/cpu.h @@ -XXX,XX +XXX,XX @@ extern const VMStateDescription vmstate_ppc_cpu; /*****************************************************************************/ void ppc_translate_init(void); +void ppc_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); #if !defined(CONFIG_USER_ONLY) void ppc_store_sdr1(CPUPPCState *env, target_ulong value); diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/riscv/cpu.h +++ b/target/riscv/cpu.h @@ -XXX,XX +XXX,XX @@ RISCVException smstateen_acc_ok(CPURISCVState *env, int index, uint64_t bit); void riscv_cpu_set_mode(CPURISCVState *env, target_ulong newpriv, bool virt_en); void riscv_translate_init(void); +void riscv_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); + G_NORETURN void riscv_raise_exception(CPURISCVState *env, uint32_t exception, uintptr_t pc); diff --git a/target/rx/cpu.h b/target/rx/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/rx/cpu.h +++ b/target/rx/cpu.h @@ -XXX,XX +XXX,XX @@ int rx_cpu_gdb_read_register(CPUState *cpu, GByteArray *buf, int reg); int rx_cpu_gdb_write_register(CPUState *cpu, uint8_t *buf, int reg); void rx_translate_init(void); +void rx_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); void rx_cpu_unpack_psw(CPURXState *env, uint32_t psw, int rte); #include "exec/cpu-all.h" diff --git a/target/s390x/s390x-internal.h b/target/s390x/s390x-internal.h index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/s390x-internal.h +++ b/target/s390x/s390x-internal.h @@ -XXX,XX +XXX,XX @@ void handle_diag_308(CPUS390XState *env, uint64_t r1, uint64_t r3, /* translate.c */ void s390x_translate_init(void); +void s390x_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); void s390x_restore_state_to_opc(CPUState *cs, const TranslationBlock *tb, const uint64_t *data); diff --git a/target/sh4/cpu.h b/target/sh4/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/sh4/cpu.h +++ b/target/sh4/cpu.h @@ -XXX,XX +XXX,XX @@ G_NORETURN void superh_cpu_do_unaligned_access(CPUState *cpu, vaddr addr, uintptr_t retaddr); void sh4_translate_init(void); +void sh4_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); #if !defined(CONFIG_USER_ONLY) hwaddr superh_cpu_get_phys_page_debug(CPUState *cpu, vaddr addr); diff --git a/target/sparc/cpu.h b/target/sparc/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/sparc/cpu.h +++ b/target/sparc/cpu.h @@ -XXX,XX +XXX,XX @@ int sparc_cpu_memory_rw_debug(CPUState *cpu, vaddr addr, /* translate.c */ void sparc_tcg_init(void); +void sparc_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); /* fop_helper.c */ target_ulong cpu_get_fsr(CPUSPARCState *); diff --git a/target/tricore/cpu.h b/target/tricore/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/tricore/cpu.h +++ b/target/tricore/cpu.h @@ -XXX,XX +XXX,XX @@ FIELD(TB_FLAGS, PRIV, 0, 2) void cpu_state_reset(CPUTriCoreState *s); void tricore_tcg_init(void); +void tricore_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); static inline void cpu_get_tb_cpu_state(CPUTriCoreState *env, vaddr *pc, uint64_t *cs_base, uint32_t *flags) diff --git a/target/xtensa/cpu.h b/target/xtensa/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/xtensa/cpu.h +++ b/target/xtensa/cpu.h @@ -XXX,XX +XXX,XX @@ G_NORETURN void xtensa_cpu_do_unaligned_access(CPUState *cpu, vaddr addr, void xtensa_collect_sr_names(const XtensaConfig *config); void xtensa_translate_init(void); +void xtensa_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc); void **xtensa_get_regfile_by_name(const char *name, int entries, int bits); void xtensa_breakpoint_handler(CPUState *cs); void xtensa_register_core(XtensaConfigList *node); diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/cpu-exec.c +++ b/accel/tcg/cpu-exec.c @@ -XXX,XX +XXX,XX @@ bool tcg_exec_realizefn(CPUState *cpu, Error **errp) if (!tcg_target_initialized) { /* Check mandatory TCGCPUOps handlers */ + const TCGCPUOps *tcg_ops = cpu->cc->tcg_ops; #ifndef CONFIG_USER_ONLY - assert(cpu->cc->tcg_ops->cpu_exec_halt); - assert(cpu->cc->tcg_ops->cpu_exec_interrupt); + assert(tcg_ops->cpu_exec_halt); + assert(tcg_ops->cpu_exec_interrupt); #endif /* !CONFIG_USER_ONLY */ - cpu->cc->tcg_ops->initialize(); + assert(tcg_ops->translate_code); + tcg_ops->initialize(); tcg_target_initialized = true; } diff --git a/accel/tcg/translate-all.c b/accel/tcg/translate-all.c index XXXXXXX..XXXXXXX 100644 --- a/accel/tcg/translate-all.c +++ b/accel/tcg/translate-all.c @@ -XXX,XX +XXX,XX @@ static int setjmp_gen_code(CPUArchState *env, TranslationBlock *tb, tcg_func_start(tcg_ctx); - tcg_ctx->cpu = env_cpu(env); - gen_intermediate_code(env_cpu(env), tb, max_insns, pc, host_pc); + CPUState *cs = env_cpu(env); + tcg_ctx->cpu = cs; + cs->cc->tcg_ops->translate_code(cs, tb, max_insns, pc, host_pc); + assert(tb->size != 0); tcg_ctx->cpu = NULL; *max_insns = tb->icount; @@ -XXX,XX +XXX,XX @@ TranslationBlock *tb_gen_code(CPUState *cpu, /* * Overflow of code_gen_buffer, or the current slice of it. * - * TODO: We don't need to re-do gen_intermediate_code, nor + * TODO: We don't need to re-do tcg_ops->translate_code, nor * should we re-do the tcg optimization currently hidden * inside tcg_gen_code. All that should be required is to * flush the TBs, allocate a new TB, re-initialize it per diff --git a/target/alpha/cpu.c b/target/alpha/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/alpha/cpu.c +++ b/target/alpha/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps alpha_sysemu_ops = { static const TCGCPUOps alpha_tcg_ops = { .initialize = alpha_translate_init, + .translate_code = alpha_translate_code, .synchronize_from_tb = alpha_cpu_synchronize_from_tb, .restore_state_to_opc = alpha_restore_state_to_opc, diff --git a/target/alpha/translate.c b/target/alpha/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/alpha/translate.c +++ b/target/alpha/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps alpha_tr_ops = { .tb_stop = alpha_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void alpha_translate_code(CPUState *cpu, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc; translator_loop(cpu, tb, max_insns, pc, host_pc, &alpha_tr_ops, &dc.base); diff --git a/target/arm/cpu.c b/target/arm/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps arm_sysemu_ops = { #ifdef CONFIG_TCG static const TCGCPUOps arm_tcg_ops = { .initialize = arm_translate_init, + .translate_code = arm_translate_code, .synchronize_from_tb = arm_cpu_synchronize_from_tb, .debug_excp_handler = arm_debug_excp_handler, .restore_state_to_opc = arm_restore_state_to_opc, diff --git a/target/arm/tcg/cpu-v7m.c b/target/arm/tcg/cpu-v7m.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tcg/cpu-v7m.c +++ b/target/arm/tcg/cpu-v7m.c @@ -XXX,XX +XXX,XX @@ static void cortex_m55_initfn(Object *obj) static const TCGCPUOps arm_v7m_tcg_ops = { .initialize = arm_translate_init, + .translate_code = arm_translate_code, .synchronize_from_tb = arm_cpu_synchronize_from_tb, .debug_excp_handler = arm_debug_excp_handler, .restore_state_to_opc = arm_restore_state_to_opc, diff --git a/target/arm/tcg/translate.c b/target/arm/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/tcg/translate.c +++ b/target/arm/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps thumb_translator_ops = { .tb_stop = arm_tr_tb_stop, }; -/* generate intermediate code for basic block 'tb'. */ -void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void arm_translate_code(CPUState *cpu, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc = { }; const TranslatorOps *ops = &arm_translator_ops; diff --git a/target/avr/cpu.c b/target/avr/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/avr/cpu.c +++ b/target/avr/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps avr_sysemu_ops = { static const TCGCPUOps avr_tcg_ops = { .initialize = avr_cpu_tcg_init, + .translate_code = avr_cpu_translate_code, .synchronize_from_tb = avr_cpu_synchronize_from_tb, .restore_state_to_opc = avr_restore_state_to_opc, .cpu_exec_interrupt = avr_cpu_exec_interrupt, diff --git a/target/avr/translate.c b/target/avr/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/avr/translate.c +++ b/target/avr/translate.c @@ -XXX,XX +XXX,XX @@ static bool trans_WDR(DisasContext *ctx, arg_WDR *a) * * - translate() * - canonicalize_skip() - * - gen_intermediate_code() + * - translate_code() * - restore_state_to_opc() * */ @@ -XXX,XX +XXX,XX @@ static const TranslatorOps avr_tr_ops = { .tb_stop = avr_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void avr_cpu_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc = { }; translator_loop(cs, tb, max_insns, pc, host_pc, &avr_tr_ops, &dc.base); diff --git a/target/hexagon/cpu.c b/target/hexagon/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/cpu.c +++ b/target/hexagon/cpu.c @@ -XXX,XX +XXX,XX @@ static void hexagon_cpu_init(Object *obj) static const TCGCPUOps hexagon_tcg_ops = { .initialize = hexagon_translate_init, + .translate_code = hexagon_translate_code, .synchronize_from_tb = hexagon_cpu_synchronize_from_tb, .restore_state_to_opc = hexagon_restore_state_to_opc, }; diff --git a/target/hexagon/translate.c b/target/hexagon/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/hexagon/translate.c +++ b/target/hexagon/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps hexagon_tr_ops = { .tb_stop = hexagon_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void hexagon_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx; diff --git a/target/hppa/cpu.c b/target/hppa/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/hppa/cpu.c +++ b/target/hppa/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps hppa_sysemu_ops = { static const TCGCPUOps hppa_tcg_ops = { .initialize = hppa_translate_init, + .translate_code = hppa_translate_code, .synchronize_from_tb = hppa_cpu_synchronize_from_tb, .restore_state_to_opc = hppa_restore_state_to_opc, diff --git a/target/hppa/translate.c b/target/hppa/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/hppa/translate.c +++ b/target/hppa/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps hppa_tr_ops = { #endif }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void hppa_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx = { }; translator_loop(cs, tb, max_insns, pc, host_pc, &hppa_tr_ops, &ctx.base); diff --git a/target/i386/tcg/tcg-cpu.c b/target/i386/tcg/tcg-cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/i386/tcg/tcg-cpu.c +++ b/target/i386/tcg/tcg-cpu.c @@ -XXX,XX +XXX,XX @@ static bool x86_debug_check_breakpoint(CPUState *cs) static const TCGCPUOps x86_tcg_ops = { .initialize = tcg_x86_init, + .translate_code = x86_translate_code, .synchronize_from_tb = x86_cpu_synchronize_from_tb, .restore_state_to_opc = x86_restore_state_to_opc, .cpu_exec_enter = x86_cpu_exec_enter, diff --git a/target/i386/tcg/translate.c b/target/i386/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/i386/tcg/translate.c +++ b/target/i386/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps i386_tr_ops = { .tb_stop = i386_tr_tb_stop, }; -/* generate intermediate code for basic block 'tb'. */ -void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void x86_translate_code(CPUState *cpu, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc; diff --git a/target/loongarch/cpu.c b/target/loongarch/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/loongarch/cpu.c +++ b/target/loongarch/cpu.c @@ -XXX,XX +XXX,XX @@ static void loongarch_cpu_dump_state(CPUState *cs, FILE *f, int flags) static const TCGCPUOps loongarch_tcg_ops = { .initialize = loongarch_translate_init, + .translate_code = loongarch_translate_code, .synchronize_from_tb = loongarch_cpu_synchronize_from_tb, .restore_state_to_opc = loongarch_restore_state_to_opc, diff --git a/target/loongarch/tcg/translate.c b/target/loongarch/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/loongarch/tcg/translate.c +++ b/target/loongarch/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps loongarch_tr_ops = { .tb_stop = loongarch_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void loongarch_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx; diff --git a/target/m68k/cpu.c b/target/m68k/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/m68k/cpu.c +++ b/target/m68k/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps m68k_sysemu_ops = { static const TCGCPUOps m68k_tcg_ops = { .initialize = m68k_tcg_init, + .translate_code = m68k_translate_code, .restore_state_to_opc = m68k_restore_state_to_opc, #ifndef CONFIG_USER_ONLY diff --git a/target/m68k/translate.c b/target/m68k/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/m68k/translate.c +++ b/target/m68k/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps m68k_tr_ops = { .tb_stop = m68k_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void m68k_translate_code(CPUState *cpu, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc; translator_loop(cpu, tb, max_insns, pc, host_pc, &m68k_tr_ops, &dc.base); diff --git a/target/microblaze/cpu.c b/target/microblaze/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/microblaze/cpu.c +++ b/target/microblaze/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps mb_sysemu_ops = { static const TCGCPUOps mb_tcg_ops = { .initialize = mb_tcg_init, + .translate_code = mb_translate_code, .synchronize_from_tb = mb_cpu_synchronize_from_tb, .restore_state_to_opc = mb_restore_state_to_opc, diff --git a/target/microblaze/translate.c b/target/microblaze/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/microblaze/translate.c +++ b/target/microblaze/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps mb_tr_ops = { .tb_stop = mb_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void mb_translate_code(CPUState *cpu, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc; translator_loop(cpu, tb, max_insns, pc, host_pc, &mb_tr_ops, &dc.base); diff --git a/target/mips/cpu.c b/target/mips/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/mips/cpu.c +++ b/target/mips/cpu.c @@ -XXX,XX +XXX,XX @@ static const Property mips_cpu_properties[] = { #include "hw/core/tcg-cpu-ops.h" static const TCGCPUOps mips_tcg_ops = { .initialize = mips_tcg_init, + .translate_code = mips_translate_code, .synchronize_from_tb = mips_cpu_synchronize_from_tb, .restore_state_to_opc = mips_restore_state_to_opc, diff --git a/target/mips/tcg/translate.c b/target/mips/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/mips/tcg/translate.c +++ b/target/mips/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps mips_tr_ops = { .tb_stop = mips_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void mips_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx; diff --git a/target/openrisc/cpu.c b/target/openrisc/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/openrisc/cpu.c +++ b/target/openrisc/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps openrisc_sysemu_ops = { static const TCGCPUOps openrisc_tcg_ops = { .initialize = openrisc_translate_init, + .translate_code = openrisc_translate_code, .synchronize_from_tb = openrisc_cpu_synchronize_from_tb, .restore_state_to_opc = openrisc_restore_state_to_opc, diff --git a/target/openrisc/translate.c b/target/openrisc/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/openrisc/translate.c +++ b/target/openrisc/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps openrisc_tr_ops = { .tb_stop = openrisc_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void openrisc_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx; diff --git a/target/ppc/cpu_init.c b/target/ppc/cpu_init.c index XXXXXXX..XXXXXXX 100644 --- a/target/ppc/cpu_init.c +++ b/target/ppc/cpu_init.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps ppc_sysemu_ops = { static const TCGCPUOps ppc_tcg_ops = { .initialize = ppc_translate_init, + .translate_code = ppc_translate_code, .restore_state_to_opc = ppc_restore_state_to_opc, #ifdef CONFIG_USER_ONLY diff --git a/target/ppc/translate.c b/target/ppc/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/ppc/translate.c +++ b/target/ppc/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps ppc_tr_ops = { .tb_stop = ppc_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void ppc_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx; diff --git a/target/riscv/tcg/tcg-cpu.c b/target/riscv/tcg/tcg-cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/riscv/tcg/tcg-cpu.c +++ b/target/riscv/tcg/tcg-cpu.c @@ -XXX,XX +XXX,XX @@ static void riscv_restore_state_to_opc(CPUState *cs, static const TCGCPUOps riscv_tcg_ops = { .initialize = riscv_translate_init, + .translate_code = riscv_translate_code, .synchronize_from_tb = riscv_cpu_synchronize_from_tb, .restore_state_to_opc = riscv_restore_state_to_opc, diff --git a/target/riscv/translate.c b/target/riscv/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps riscv_tr_ops = { .tb_stop = riscv_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void riscv_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx; diff --git a/target/rx/cpu.c b/target/rx/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/rx/cpu.c +++ b/target/rx/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps rx_sysemu_ops = { static const TCGCPUOps rx_tcg_ops = { .initialize = rx_translate_init, + .translate_code = rx_translate_code, .synchronize_from_tb = rx_cpu_synchronize_from_tb, .restore_state_to_opc = rx_restore_state_to_opc, .tlb_fill = rx_cpu_tlb_fill, diff --git a/target/rx/translate.c b/target/rx/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/rx/translate.c +++ b/target/rx/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps rx_tr_ops = { .tb_stop = rx_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void rx_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc; diff --git a/target/s390x/cpu.c b/target/s390x/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/cpu.c +++ b/target/s390x/cpu.c @@ -XXX,XX +XXX,XX @@ void cpu_get_tb_cpu_state(CPUS390XState *env, vaddr *pc, static const TCGCPUOps s390_tcg_ops = { .initialize = s390x_translate_init, + .translate_code = s390x_translate_code, .restore_state_to_opc = s390x_restore_state_to_opc, #ifdef CONFIG_USER_ONLY diff --git a/target/s390x/tcg/translate.c b/target/s390x/tcg/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/s390x/tcg/translate.c +++ b/target/s390x/tcg/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps s390x_tr_ops = { .disas_log = s390x_tr_disas_log, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void s390x_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc; diff --git a/target/sh4/cpu.c b/target/sh4/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/sh4/cpu.c +++ b/target/sh4/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps sh4_sysemu_ops = { static const TCGCPUOps superh_tcg_ops = { .initialize = sh4_translate_init, + .translate_code = sh4_translate_code, .synchronize_from_tb = superh_cpu_synchronize_from_tb, .restore_state_to_opc = superh_restore_state_to_opc, diff --git a/target/sh4/translate.c b/target/sh4/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/sh4/translate.c +++ b/target/sh4/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps sh4_tr_ops = { .tb_stop = sh4_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void sh4_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx; diff --git a/target/sparc/cpu.c b/target/sparc/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/sparc/cpu.c +++ b/target/sparc/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps sparc_sysemu_ops = { static const TCGCPUOps sparc_tcg_ops = { .initialize = sparc_tcg_init, + .translate_code = sparc_translate_code, .synchronize_from_tb = sparc_cpu_synchronize_from_tb, .restore_state_to_opc = sparc_restore_state_to_opc, diff --git a/target/sparc/translate.c b/target/sparc/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/sparc/translate.c +++ b/target/sparc/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps sparc_tr_ops = { .tb_stop = sparc_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void sparc_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc = {}; diff --git a/target/tricore/cpu.c b/target/tricore/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/tricore/cpu.c +++ b/target/tricore/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps tricore_sysemu_ops = { static const TCGCPUOps tricore_tcg_ops = { .initialize = tricore_tcg_init, + .translate_code = tricore_translate_code, .synchronize_from_tb = tricore_cpu_synchronize_from_tb, .restore_state_to_opc = tricore_restore_state_to_opc, .tlb_fill = tricore_cpu_tlb_fill, diff --git a/target/tricore/translate.c b/target/tricore/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/tricore/translate.c +++ b/target/tricore/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps tricore_tr_ops = { .tb_stop = tricore_tr_tb_stop, }; - -void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void tricore_translate_code(CPUState *cs, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext ctx; translator_loop(cs, tb, max_insns, pc, host_pc, diff --git a/target/xtensa/cpu.c b/target/xtensa/cpu.c index XXXXXXX..XXXXXXX 100644 --- a/target/xtensa/cpu.c +++ b/target/xtensa/cpu.c @@ -XXX,XX +XXX,XX @@ static const struct SysemuCPUOps xtensa_sysemu_ops = { static const TCGCPUOps xtensa_tcg_ops = { .initialize = xtensa_translate_init, + .translate_code = xtensa_translate_code, .debug_excp_handler = xtensa_breakpoint_handler, .restore_state_to_opc = xtensa_restore_state_to_opc, diff --git a/target/xtensa/translate.c b/target/xtensa/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/xtensa/translate.c +++ b/target/xtensa/translate.c @@ -XXX,XX +XXX,XX @@ static const TranslatorOps xtensa_translator_ops = { .tb_stop = xtensa_tr_tb_stop, }; -void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int *max_insns, - vaddr pc, void *host_pc) +void xtensa_translate_code(CPUState *cpu, TranslationBlock *tb, + int *max_insns, vaddr pc, void *host_pc) { DisasContext dc = {}; translator_loop(cpu, tb, max_insns, pc, host_pc, -- 2.43.0