From 763e10f7bb6413c745edf36e9e806c3d626c8262 Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Thu, 21 Feb 2019 18:17:45 +0000 Subject: [PATCH 01/21] hw/arm/armsse: Fix memory leak in error-exit path MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Coverity points out (CID 1398632, CID 1398650) that we leak a couple of allocated strings in the error-exit code path for setting up the MHUs in the ARMSSE. Fix this bug by moving the allocate-and-free of each string to be closer to the use, so we do the free before doing the error-exit check. Fixes: f8574705f62b38a ("hw/arm/armsse: Add unimplemented-device stubs for MHUs") Signed-off-by: Peter Maydell Reviewed-by: Philippe Mathieu-Daudé Message-id: 20190215113707.24553-1-peter.maydell@linaro.org --- hw/arm/armsse.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) diff --git a/hw/arm/armsse.c b/hw/arm/armsse.c index 9a8c49547d..d0207dbabc 100644 --- a/hw/arm/armsse.c +++ b/hw/arm/armsse.c @@ -762,26 +762,28 @@ static void armsse_realize(DeviceState *dev, Error **errp) if (info->has_mhus) { for (i = 0; i < ARRAY_SIZE(s->mhu); i++) { - char *name = g_strdup_printf("MHU%d", i); - char *port = g_strdup_printf("port[%d]", i + 3); + char *name; + char *port; + name = g_strdup_printf("MHU%d", i); qdev_prop_set_string(DEVICE(&s->mhu[i]), "name", name); qdev_prop_set_uint64(DEVICE(&s->mhu[i]), "size", 0x1000); object_property_set_bool(OBJECT(&s->mhu[i]), true, "realized", &err); + g_free(name); if (err) { error_propagate(errp, err); return; } + port = g_strdup_printf("port[%d]", i + 3); mr = sysbus_mmio_get_region(SYS_BUS_DEVICE(&s->mhu[i]), 0); object_property_set_link(OBJECT(&s->apb_ppc0), OBJECT(mr), port, &err); + g_free(port); if (err) { error_propagate(errp, err); return; } - g_free(name); - g_free(port); } } From cff21316c666c8053b1f425577e324038d0ca30d Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Thu, 21 Feb 2019 18:17:45 +0000 Subject: [PATCH 02/21] target/arm: v8M MPU should use background region as default, not always The "background region" for a v8M MPU is a default which will be used (if enabled, and if the access is privileged) if the access does not match any specific MPU region. We were incorrectly using it always (by putting the condition at the wrong nesting level). This meant that we would always return the default background permissions rather than the correct permissions for a specific region, and also that we would not return the right information in response to a TT instruction. Move the check for the background region to the same place in the logic as the equivalent v8M MPUCheck() pseudocode puts it. This in turn means we must adjust the condition we use to detect matches in multiple regions to avoid false-positives. Signed-off-by: Peter Maydell Reviewed-by: Richard Henderson Message-id: 20190214113408.10214-1-peter.maydell@linaro.org --- target/arm/helper.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index a018eb23fe..fe054897c7 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -11387,9 +11387,11 @@ static bool pmsav8_mpu_lookup(CPUARMState *env, uint32_t address, hit = true; } else if (m_is_ppb_region(env, address)) { hit = true; - } else if (pmsav7_use_background_region(cpu, mmu_idx, is_user)) { - hit = true; } else { + if (pmsav7_use_background_region(cpu, mmu_idx, is_user)) { + hit = true; + } + for (n = (int)cpu->pmsav7_dregion - 1; n >= 0; n--) { /* region search */ /* Note that the base address is bits [31:5] from the register @@ -11427,7 +11429,7 @@ static bool pmsav8_mpu_lookup(CPUARMState *env, uint32_t address, *is_subpage = true; } - if (hit) { + if (matchregion != -1) { /* Multiple regions match -- always a failure (unlike * PMSAv7 where highest-numbered-region wins) */ From 67da43d668320e1bcb0a0195aaf2de4ff2a001a0 Mon Sep 17 00:00:00 2001 From: Aaron Lindsay OS Date: Thu, 21 Feb 2019 18:17:45 +0000 Subject: [PATCH 03/21] target/arm: Stop unintentional sign extension in pmu_init This was introduced by commit bf8d09694ccc07487cd73d7562081fdaec3370c8 target/arm: Don't clear supported PMU events when initializing PMCEID1 and identified by Coverity (CID 1398645). Signed-off-by: Aaron Lindsay Reported-by: Peter Maydell Message-id: 20190219144621.450-1-aaron@os.amperecomputing.com Reviewed-by: Peter Maydell Signed-off-by: Peter Maydell --- target/arm/helper.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index fe054897c7..c2630fa1ea 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -1167,7 +1167,7 @@ void pmu_init(ARMCPU *cpu) if (cnt->supported(&cpu->env)) { supported_event_map[cnt->number] = i; - uint64_t event_mask = 1 << (cnt->number & 0x1f); + uint64_t event_mask = 1ULL << (cnt->number & 0x1f); if (cnt->number & 0x20) { cpu->pmceid1 |= event_mask; } else { From 3c3ff68492c2d00bd8cb39ed2d02bdaf5caf5cb8 Mon Sep 17 00:00:00 2001 From: Richard Henderson Date: Thu, 21 Feb 2019 18:17:45 +0000 Subject: [PATCH 04/21] target/arm: Restructure disas_fp_int_conv For opcodes 0-5, move some if conditions into the structure of a switch statement. For opcodes 6 & 7, decode everything at once with a second switch. Signed-off-by: Richard Henderson Message-id: 20190215192302.27855-2-richard.henderson@linaro.org Reviewed-by: Peter Maydell Signed-off-by: Peter Maydell --- target/arm/translate-a64.c | 94 ++++++++++++++++++++------------------ 1 file changed, 49 insertions(+), 45 deletions(-) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index af8e4fd4be..dbce24fe32 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -6541,68 +6541,72 @@ static void disas_fp_int_conv(DisasContext *s, uint32_t insn) int type = extract32(insn, 22, 2); bool sbit = extract32(insn, 29, 1); bool sf = extract32(insn, 31, 1); + bool itof = false; if (sbit) { - unallocated_encoding(s); - return; + goto do_unallocated; } - if (opcode > 5) { - /* FMOV */ - bool itof = opcode & 1; - - if (rmode >= 2) { - unallocated_encoding(s); - return; - } - - switch (sf << 3 | type << 1 | rmode) { - case 0x0: /* 32 bit */ - case 0xa: /* 64 bit */ - case 0xd: /* 64 bit to top half of quad */ - break; - case 0x6: /* 16-bit float, 32-bit int */ - case 0xe: /* 16-bit float, 64-bit int */ - if (dc_isar_feature(aa64_fp16, s)) { - break; - } - /* fallthru */ - default: - /* all other sf/type/rmode combinations are invalid */ - unallocated_encoding(s); - return; - } - - if (!fp_access_check(s)) { - return; - } - handle_fmov(s, rd, rn, type, itof); - } else { - /* actual FP conversions */ - bool itof = extract32(opcode, 1, 1); - - if (rmode != 0 && opcode > 1) { - unallocated_encoding(s); - return; + switch (opcode) { + case 2: /* SCVTF */ + case 3: /* UCVTF */ + itof = true; + /* fallthru */ + case 4: /* FCVTAS */ + case 5: /* FCVTAU */ + if (rmode != 0) { + goto do_unallocated; } + /* fallthru */ + case 0: /* FCVT[NPMZ]S */ + case 1: /* FCVT[NPMZ]U */ switch (type) { case 0: /* float32 */ case 1: /* float64 */ break; case 3: /* float16 */ - if (dc_isar_feature(aa64_fp16, s)) { - break; + if (!dc_isar_feature(aa64_fp16, s)) { + goto do_unallocated; } - /* fallthru */ + break; default: - unallocated_encoding(s); - return; + goto do_unallocated; } - if (!fp_access_check(s)) { return; } handle_fpfpcvt(s, rd, rn, opcode, itof, rmode, 64, sf, type); + break; + + default: + switch (sf << 7 | type << 5 | rmode << 3 | opcode) { + case 0b01100110: /* FMOV half <-> 32-bit int */ + case 0b01100111: + case 0b11100110: /* FMOV half <-> 64-bit int */ + case 0b11100111: + if (!dc_isar_feature(aa64_fp16, s)) { + goto do_unallocated; + } + /* fallthru */ + case 0b00000110: /* FMOV 32-bit */ + case 0b00000111: + case 0b10100110: /* FMOV 64-bit */ + case 0b10100111: + case 0b11001110: /* FMOV top half of 128-bit */ + case 0b11001111: + if (!fp_access_check(s)) { + return; + } + itof = opcode & 1; + handle_fmov(s, rd, rn, type, itof); + break; + + default: + do_unallocated: + unallocated_encoding(s); + return; + } + break; } } From 37356079fcdb34e13abbed8ea0c00ca880c31247 Mon Sep 17 00:00:00 2001 From: Richard Henderson Date: Thu, 21 Feb 2019 18:17:45 +0000 Subject: [PATCH 05/21] target/arm: Split out vfp_helper.c Move all of the fp helpers out of helper.c into a new file. This is code movement only. Since helper.c has no copyright header, take the one from cpu.h for the new file. Signed-off-by: Richard Henderson Message-id: 20190215192302.27855-3-richard.henderson@linaro.org Reviewed-by: Peter Maydell Signed-off-by: Peter Maydell --- target/arm/Makefile.objs | 2 +- target/arm/helper.c | 1062 ------------------------------------- target/arm/vfp_helper.c | 1088 ++++++++++++++++++++++++++++++++++++++ 3 files changed, 1089 insertions(+), 1063 deletions(-) create mode 100644 target/arm/vfp_helper.c diff --git a/target/arm/Makefile.objs b/target/arm/Makefile.objs index 1a4fc06448..6bdcc65c2c 100644 --- a/target/arm/Makefile.objs +++ b/target/arm/Makefile.objs @@ -5,7 +5,7 @@ obj-$(call land,$(CONFIG_KVM),$(call lnot,$(TARGET_AARCH64))) += kvm32.o obj-$(call land,$(CONFIG_KVM),$(TARGET_AARCH64)) += kvm64.o obj-$(call lnot,$(CONFIG_KVM)) += kvm-stub.o obj-y += translate.o op_helper.o helper.o cpu.o -obj-y += neon_helper.o iwmmxt_helper.o vec_helper.o +obj-y += neon_helper.o iwmmxt_helper.o vec_helper.o vfp_helper.o obj-y += gdbstub.o obj-$(TARGET_AARCH64) += cpu64.o translate-a64.o helper-a64.o gdbstub64.o obj-$(TARGET_AARCH64) += pauth_helper.o diff --git a/target/arm/helper.c b/target/arm/helper.c index c2630fa1ea..fbaa801cea 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -12681,1068 +12681,6 @@ uint32_t HELPER(sel_flags)(uint32_t flags, uint32_t a, uint32_t b) return (a & mask) | (b & ~mask); } -/* VFP support. We follow the convention used for VFP instructions: - Single precision routines have a "s" suffix, double precision a - "d" suffix. */ - -/* Convert host exception flags to vfp form. */ -static inline int vfp_exceptbits_from_host(int host_bits) -{ - int target_bits = 0; - - if (host_bits & float_flag_invalid) - target_bits |= 1; - if (host_bits & float_flag_divbyzero) - target_bits |= 2; - if (host_bits & float_flag_overflow) - target_bits |= 4; - if (host_bits & (float_flag_underflow | float_flag_output_denormal)) - target_bits |= 8; - if (host_bits & float_flag_inexact) - target_bits |= 0x10; - if (host_bits & float_flag_input_denormal) - target_bits |= 0x80; - return target_bits; -} - -uint32_t HELPER(vfp_get_fpscr)(CPUARMState *env) -{ - uint32_t i, fpscr; - - fpscr = env->vfp.xregs[ARM_VFP_FPSCR] - | (env->vfp.vec_len << 16) - | (env->vfp.vec_stride << 20); - - i = get_float_exception_flags(&env->vfp.fp_status); - i |= get_float_exception_flags(&env->vfp.standard_fp_status); - /* FZ16 does not generate an input denormal exception. */ - i |= (get_float_exception_flags(&env->vfp.fp_status_f16) - & ~float_flag_input_denormal); - fpscr |= vfp_exceptbits_from_host(i); - - i = env->vfp.qc[0] | env->vfp.qc[1] | env->vfp.qc[2] | env->vfp.qc[3]; - fpscr |= i ? FPCR_QC : 0; - - return fpscr; -} - -uint32_t vfp_get_fpscr(CPUARMState *env) -{ - return HELPER(vfp_get_fpscr)(env); -} - -/* Convert vfp exception flags to target form. */ -static inline int vfp_exceptbits_to_host(int target_bits) -{ - int host_bits = 0; - - if (target_bits & 1) - host_bits |= float_flag_invalid; - if (target_bits & 2) - host_bits |= float_flag_divbyzero; - if (target_bits & 4) - host_bits |= float_flag_overflow; - if (target_bits & 8) - host_bits |= float_flag_underflow; - if (target_bits & 0x10) - host_bits |= float_flag_inexact; - if (target_bits & 0x80) - host_bits |= float_flag_input_denormal; - return host_bits; -} - -void HELPER(vfp_set_fpscr)(CPUARMState *env, uint32_t val) -{ - int i; - uint32_t changed = env->vfp.xregs[ARM_VFP_FPSCR]; - - /* When ARMv8.2-FP16 is not supported, FZ16 is RES0. */ - if (!cpu_isar_feature(aa64_fp16, arm_env_get_cpu(env))) { - val &= ~FPCR_FZ16; - } - - /* - * We don't implement trapped exception handling, so the - * trap enable bits, IDE|IXE|UFE|OFE|DZE|IOE are all RAZ/WI (not RES0!) - * - * If we exclude the exception flags, IOC|DZC|OFC|UFC|IXC|IDC - * (which are stored in fp_status), and the other RES0 bits - * in between, then we clear all of the low 16 bits. - */ - env->vfp.xregs[ARM_VFP_FPSCR] = val & 0xf7c80000; - env->vfp.vec_len = (val >> 16) & 7; - env->vfp.vec_stride = (val >> 20) & 3; - - /* - * The bit we set within fpscr_q is arbitrary; the register as a - * whole being zero/non-zero is what counts. - */ - env->vfp.qc[0] = val & FPCR_QC; - env->vfp.qc[1] = 0; - env->vfp.qc[2] = 0; - env->vfp.qc[3] = 0; - - changed ^= val; - if (changed & (3 << 22)) { - i = (val >> 22) & 3; - switch (i) { - case FPROUNDING_TIEEVEN: - i = float_round_nearest_even; - break; - case FPROUNDING_POSINF: - i = float_round_up; - break; - case FPROUNDING_NEGINF: - i = float_round_down; - break; - case FPROUNDING_ZERO: - i = float_round_to_zero; - break; - } - set_float_rounding_mode(i, &env->vfp.fp_status); - set_float_rounding_mode(i, &env->vfp.fp_status_f16); - } - if (changed & FPCR_FZ16) { - bool ftz_enabled = val & FPCR_FZ16; - set_flush_to_zero(ftz_enabled, &env->vfp.fp_status_f16); - set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status_f16); - } - if (changed & FPCR_FZ) { - bool ftz_enabled = val & FPCR_FZ; - set_flush_to_zero(ftz_enabled, &env->vfp.fp_status); - set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status); - } - if (changed & FPCR_DN) { - bool dnan_enabled = val & FPCR_DN; - set_default_nan_mode(dnan_enabled, &env->vfp.fp_status); - set_default_nan_mode(dnan_enabled, &env->vfp.fp_status_f16); - } - - /* The exception flags are ORed together when we read fpscr so we - * only need to preserve the current state in one of our - * float_status values. - */ - i = vfp_exceptbits_to_host(val); - set_float_exception_flags(i, &env->vfp.fp_status); - set_float_exception_flags(0, &env->vfp.fp_status_f16); - set_float_exception_flags(0, &env->vfp.standard_fp_status); -} - -void vfp_set_fpscr(CPUARMState *env, uint32_t val) -{ - HELPER(vfp_set_fpscr)(env, val); -} - -#define VFP_HELPER(name, p) HELPER(glue(glue(vfp_,name),p)) - -#define VFP_BINOP(name) \ -float32 VFP_HELPER(name, s)(float32 a, float32 b, void *fpstp) \ -{ \ - float_status *fpst = fpstp; \ - return float32_ ## name(a, b, fpst); \ -} \ -float64 VFP_HELPER(name, d)(float64 a, float64 b, void *fpstp) \ -{ \ - float_status *fpst = fpstp; \ - return float64_ ## name(a, b, fpst); \ -} -VFP_BINOP(add) -VFP_BINOP(sub) -VFP_BINOP(mul) -VFP_BINOP(div) -VFP_BINOP(min) -VFP_BINOP(max) -VFP_BINOP(minnum) -VFP_BINOP(maxnum) -#undef VFP_BINOP - -float32 VFP_HELPER(neg, s)(float32 a) -{ - return float32_chs(a); -} - -float64 VFP_HELPER(neg, d)(float64 a) -{ - return float64_chs(a); -} - -float32 VFP_HELPER(abs, s)(float32 a) -{ - return float32_abs(a); -} - -float64 VFP_HELPER(abs, d)(float64 a) -{ - return float64_abs(a); -} - -float32 VFP_HELPER(sqrt, s)(float32 a, CPUARMState *env) -{ - return float32_sqrt(a, &env->vfp.fp_status); -} - -float64 VFP_HELPER(sqrt, d)(float64 a, CPUARMState *env) -{ - return float64_sqrt(a, &env->vfp.fp_status); -} - -static void softfloat_to_vfp_compare(CPUARMState *env, int cmp) -{ - uint32_t flags; - switch (cmp) { - case float_relation_equal: - flags = 0x6; - break; - case float_relation_less: - flags = 0x8; - break; - case float_relation_greater: - flags = 0x2; - break; - case float_relation_unordered: - flags = 0x3; - break; - default: - g_assert_not_reached(); - } - env->vfp.xregs[ARM_VFP_FPSCR] = - deposit32(env->vfp.xregs[ARM_VFP_FPSCR], 28, 4, flags); -} - -/* XXX: check quiet/signaling case */ -#define DO_VFP_cmp(p, type) \ -void VFP_HELPER(cmp, p)(type a, type b, CPUARMState *env) \ -{ \ - softfloat_to_vfp_compare(env, \ - type ## _compare_quiet(a, b, &env->vfp.fp_status)); \ -} \ -void VFP_HELPER(cmpe, p)(type a, type b, CPUARMState *env) \ -{ \ - softfloat_to_vfp_compare(env, \ - type ## _compare(a, b, &env->vfp.fp_status)); \ -} -DO_VFP_cmp(s, float32) -DO_VFP_cmp(d, float64) -#undef DO_VFP_cmp - -/* Integer to float and float to integer conversions */ - -#define CONV_ITOF(name, ftype, fsz, sign) \ -ftype HELPER(name)(uint32_t x, void *fpstp) \ -{ \ - float_status *fpst = fpstp; \ - return sign##int32_to_##float##fsz((sign##int32_t)x, fpst); \ -} - -#define CONV_FTOI(name, ftype, fsz, sign, round) \ -sign##int32_t HELPER(name)(ftype x, void *fpstp) \ -{ \ - float_status *fpst = fpstp; \ - if (float##fsz##_is_any_nan(x)) { \ - float_raise(float_flag_invalid, fpst); \ - return 0; \ - } \ - return float##fsz##_to_##sign##int32##round(x, fpst); \ -} - -#define FLOAT_CONVS(name, p, ftype, fsz, sign) \ - CONV_ITOF(vfp_##name##to##p, ftype, fsz, sign) \ - CONV_FTOI(vfp_to##name##p, ftype, fsz, sign, ) \ - CONV_FTOI(vfp_to##name##z##p, ftype, fsz, sign, _round_to_zero) - -FLOAT_CONVS(si, h, uint32_t, 16, ) -FLOAT_CONVS(si, s, float32, 32, ) -FLOAT_CONVS(si, d, float64, 64, ) -FLOAT_CONVS(ui, h, uint32_t, 16, u) -FLOAT_CONVS(ui, s, float32, 32, u) -FLOAT_CONVS(ui, d, float64, 64, u) - -#undef CONV_ITOF -#undef CONV_FTOI -#undef FLOAT_CONVS - -/* floating point conversion */ -float64 VFP_HELPER(fcvtd, s)(float32 x, CPUARMState *env) -{ - return float32_to_float64(x, &env->vfp.fp_status); -} - -float32 VFP_HELPER(fcvts, d)(float64 x, CPUARMState *env) -{ - return float64_to_float32(x, &env->vfp.fp_status); -} - -/* VFP3 fixed point conversion. */ -#define VFP_CONV_FIX_FLOAT(name, p, fsz, isz, itype) \ -float##fsz HELPER(vfp_##name##to##p)(uint##isz##_t x, uint32_t shift, \ - void *fpstp) \ -{ return itype##_to_##float##fsz##_scalbn(x, -shift, fpstp); } - -#define VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, isz, itype, ROUND, suff) \ -uint##isz##_t HELPER(vfp_to##name##p##suff)(float##fsz x, uint32_t shift, \ - void *fpst) \ -{ \ - if (unlikely(float##fsz##_is_any_nan(x))) { \ - float_raise(float_flag_invalid, fpst); \ - return 0; \ - } \ - return float##fsz##_to_##itype##_scalbn(x, ROUND, shift, fpst); \ -} - -#define VFP_CONV_FIX(name, p, fsz, isz, itype) \ -VFP_CONV_FIX_FLOAT(name, p, fsz, isz, itype) \ -VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, isz, itype, \ - float_round_to_zero, _round_to_zero) \ -VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, isz, itype, \ - get_float_rounding_mode(fpst), ) - -#define VFP_CONV_FIX_A64(name, p, fsz, isz, itype) \ -VFP_CONV_FIX_FLOAT(name, p, fsz, isz, itype) \ -VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, isz, itype, \ - get_float_rounding_mode(fpst), ) - -VFP_CONV_FIX(sh, d, 64, 64, int16) -VFP_CONV_FIX(sl, d, 64, 64, int32) -VFP_CONV_FIX_A64(sq, d, 64, 64, int64) -VFP_CONV_FIX(uh, d, 64, 64, uint16) -VFP_CONV_FIX(ul, d, 64, 64, uint32) -VFP_CONV_FIX_A64(uq, d, 64, 64, uint64) -VFP_CONV_FIX(sh, s, 32, 32, int16) -VFP_CONV_FIX(sl, s, 32, 32, int32) -VFP_CONV_FIX_A64(sq, s, 32, 64, int64) -VFP_CONV_FIX(uh, s, 32, 32, uint16) -VFP_CONV_FIX(ul, s, 32, 32, uint32) -VFP_CONV_FIX_A64(uq, s, 32, 64, uint64) - -#undef VFP_CONV_FIX -#undef VFP_CONV_FIX_FLOAT -#undef VFP_CONV_FLOAT_FIX_ROUND -#undef VFP_CONV_FIX_A64 - -uint32_t HELPER(vfp_sltoh)(uint32_t x, uint32_t shift, void *fpst) -{ - return int32_to_float16_scalbn(x, -shift, fpst); -} - -uint32_t HELPER(vfp_ultoh)(uint32_t x, uint32_t shift, void *fpst) -{ - return uint32_to_float16_scalbn(x, -shift, fpst); -} - -uint32_t HELPER(vfp_sqtoh)(uint64_t x, uint32_t shift, void *fpst) -{ - return int64_to_float16_scalbn(x, -shift, fpst); -} - -uint32_t HELPER(vfp_uqtoh)(uint64_t x, uint32_t shift, void *fpst) -{ - return uint64_to_float16_scalbn(x, -shift, fpst); -} - -uint32_t HELPER(vfp_toshh)(uint32_t x, uint32_t shift, void *fpst) -{ - if (unlikely(float16_is_any_nan(x))) { - float_raise(float_flag_invalid, fpst); - return 0; - } - return float16_to_int16_scalbn(x, get_float_rounding_mode(fpst), - shift, fpst); -} - -uint32_t HELPER(vfp_touhh)(uint32_t x, uint32_t shift, void *fpst) -{ - if (unlikely(float16_is_any_nan(x))) { - float_raise(float_flag_invalid, fpst); - return 0; - } - return float16_to_uint16_scalbn(x, get_float_rounding_mode(fpst), - shift, fpst); -} - -uint32_t HELPER(vfp_toslh)(uint32_t x, uint32_t shift, void *fpst) -{ - if (unlikely(float16_is_any_nan(x))) { - float_raise(float_flag_invalid, fpst); - return 0; - } - return float16_to_int32_scalbn(x, get_float_rounding_mode(fpst), - shift, fpst); -} - -uint32_t HELPER(vfp_toulh)(uint32_t x, uint32_t shift, void *fpst) -{ - if (unlikely(float16_is_any_nan(x))) { - float_raise(float_flag_invalid, fpst); - return 0; - } - return float16_to_uint32_scalbn(x, get_float_rounding_mode(fpst), - shift, fpst); -} - -uint64_t HELPER(vfp_tosqh)(uint32_t x, uint32_t shift, void *fpst) -{ - if (unlikely(float16_is_any_nan(x))) { - float_raise(float_flag_invalid, fpst); - return 0; - } - return float16_to_int64_scalbn(x, get_float_rounding_mode(fpst), - shift, fpst); -} - -uint64_t HELPER(vfp_touqh)(uint32_t x, uint32_t shift, void *fpst) -{ - if (unlikely(float16_is_any_nan(x))) { - float_raise(float_flag_invalid, fpst); - return 0; - } - return float16_to_uint64_scalbn(x, get_float_rounding_mode(fpst), - shift, fpst); -} - -/* Set the current fp rounding mode and return the old one. - * The argument is a softfloat float_round_ value. - */ -uint32_t HELPER(set_rmode)(uint32_t rmode, void *fpstp) -{ - float_status *fp_status = fpstp; - - uint32_t prev_rmode = get_float_rounding_mode(fp_status); - set_float_rounding_mode(rmode, fp_status); - - return prev_rmode; -} - -/* Set the current fp rounding mode in the standard fp status and return - * the old one. This is for NEON instructions that need to change the - * rounding mode but wish to use the standard FPSCR values for everything - * else. Always set the rounding mode back to the correct value after - * modifying it. - * The argument is a softfloat float_round_ value. - */ -uint32_t HELPER(set_neon_rmode)(uint32_t rmode, CPUARMState *env) -{ - float_status *fp_status = &env->vfp.standard_fp_status; - - uint32_t prev_rmode = get_float_rounding_mode(fp_status); - set_float_rounding_mode(rmode, fp_status); - - return prev_rmode; -} - -/* Half precision conversions. */ -float32 HELPER(vfp_fcvt_f16_to_f32)(uint32_t a, void *fpstp, uint32_t ahp_mode) -{ - /* Squash FZ16 to 0 for the duration of conversion. In this case, - * it would affect flushing input denormals. - */ - float_status *fpst = fpstp; - flag save = get_flush_inputs_to_zero(fpst); - set_flush_inputs_to_zero(false, fpst); - float32 r = float16_to_float32(a, !ahp_mode, fpst); - set_flush_inputs_to_zero(save, fpst); - return r; -} - -uint32_t HELPER(vfp_fcvt_f32_to_f16)(float32 a, void *fpstp, uint32_t ahp_mode) -{ - /* Squash FZ16 to 0 for the duration of conversion. In this case, - * it would affect flushing output denormals. - */ - float_status *fpst = fpstp; - flag save = get_flush_to_zero(fpst); - set_flush_to_zero(false, fpst); - float16 r = float32_to_float16(a, !ahp_mode, fpst); - set_flush_to_zero(save, fpst); - return r; -} - -float64 HELPER(vfp_fcvt_f16_to_f64)(uint32_t a, void *fpstp, uint32_t ahp_mode) -{ - /* Squash FZ16 to 0 for the duration of conversion. In this case, - * it would affect flushing input denormals. - */ - float_status *fpst = fpstp; - flag save = get_flush_inputs_to_zero(fpst); - set_flush_inputs_to_zero(false, fpst); - float64 r = float16_to_float64(a, !ahp_mode, fpst); - set_flush_inputs_to_zero(save, fpst); - return r; -} - -uint32_t HELPER(vfp_fcvt_f64_to_f16)(float64 a, void *fpstp, uint32_t ahp_mode) -{ - /* Squash FZ16 to 0 for the duration of conversion. In this case, - * it would affect flushing output denormals. - */ - float_status *fpst = fpstp; - flag save = get_flush_to_zero(fpst); - set_flush_to_zero(false, fpst); - float16 r = float64_to_float16(a, !ahp_mode, fpst); - set_flush_to_zero(save, fpst); - return r; -} - -#define float32_two make_float32(0x40000000) -#define float32_three make_float32(0x40400000) -#define float32_one_point_five make_float32(0x3fc00000) - -float32 HELPER(recps_f32)(float32 a, float32 b, CPUARMState *env) -{ - float_status *s = &env->vfp.standard_fp_status; - if ((float32_is_infinity(a) && float32_is_zero_or_denormal(b)) || - (float32_is_infinity(b) && float32_is_zero_or_denormal(a))) { - if (!(float32_is_zero(a) || float32_is_zero(b))) { - float_raise(float_flag_input_denormal, s); - } - return float32_two; - } - return float32_sub(float32_two, float32_mul(a, b, s), s); -} - -float32 HELPER(rsqrts_f32)(float32 a, float32 b, CPUARMState *env) -{ - float_status *s = &env->vfp.standard_fp_status; - float32 product; - if ((float32_is_infinity(a) && float32_is_zero_or_denormal(b)) || - (float32_is_infinity(b) && float32_is_zero_or_denormal(a))) { - if (!(float32_is_zero(a) || float32_is_zero(b))) { - float_raise(float_flag_input_denormal, s); - } - return float32_one_point_five; - } - product = float32_mul(a, b, s); - return float32_div(float32_sub(float32_three, product, s), float32_two, s); -} - -/* NEON helpers. */ - -/* Constants 256 and 512 are used in some helpers; we avoid relying on - * int->float conversions at run-time. */ -#define float64_256 make_float64(0x4070000000000000LL) -#define float64_512 make_float64(0x4080000000000000LL) -#define float16_maxnorm make_float16(0x7bff) -#define float32_maxnorm make_float32(0x7f7fffff) -#define float64_maxnorm make_float64(0x7fefffffffffffffLL) - -/* Reciprocal functions - * - * The algorithm that must be used to calculate the estimate - * is specified by the ARM ARM, see FPRecipEstimate()/RecipEstimate - */ - -/* See RecipEstimate() - * - * input is a 9 bit fixed point number - * input range 256 .. 511 for a number from 0.5 <= x < 1.0. - * result range 256 .. 511 for a number from 1.0 to 511/256. - */ - -static int recip_estimate(int input) -{ - int a, b, r; - assert(256 <= input && input < 512); - a = (input * 2) + 1; - b = (1 << 19) / a; - r = (b + 1) >> 1; - assert(256 <= r && r < 512); - return r; -} - -/* - * Common wrapper to call recip_estimate - * - * The parameters are exponent and 64 bit fraction (without implicit - * bit) where the binary point is nominally at bit 52. Returns a - * float64 which can then be rounded to the appropriate size by the - * callee. - */ - -static uint64_t call_recip_estimate(int *exp, int exp_off, uint64_t frac) -{ - uint32_t scaled, estimate; - uint64_t result_frac; - int result_exp; - - /* Handle sub-normals */ - if (*exp == 0) { - if (extract64(frac, 51, 1) == 0) { - *exp = -1; - frac <<= 2; - } else { - frac <<= 1; - } - } - - /* scaled = UInt('1':fraction<51:44>) */ - scaled = deposit32(1 << 8, 0, 8, extract64(frac, 44, 8)); - estimate = recip_estimate(scaled); - - result_exp = exp_off - *exp; - result_frac = deposit64(0, 44, 8, estimate); - if (result_exp == 0) { - result_frac = deposit64(result_frac >> 1, 51, 1, 1); - } else if (result_exp == -1) { - result_frac = deposit64(result_frac >> 2, 50, 2, 1); - result_exp = 0; - } - - *exp = result_exp; - - return result_frac; -} - -static bool round_to_inf(float_status *fpst, bool sign_bit) -{ - switch (fpst->float_rounding_mode) { - case float_round_nearest_even: /* Round to Nearest */ - return true; - case float_round_up: /* Round to +Inf */ - return !sign_bit; - case float_round_down: /* Round to -Inf */ - return sign_bit; - case float_round_to_zero: /* Round to Zero */ - return false; - } - - g_assert_not_reached(); -} - -uint32_t HELPER(recpe_f16)(uint32_t input, void *fpstp) -{ - float_status *fpst = fpstp; - float16 f16 = float16_squash_input_denormal(input, fpst); - uint32_t f16_val = float16_val(f16); - uint32_t f16_sign = float16_is_neg(f16); - int f16_exp = extract32(f16_val, 10, 5); - uint32_t f16_frac = extract32(f16_val, 0, 10); - uint64_t f64_frac; - - if (float16_is_any_nan(f16)) { - float16 nan = f16; - if (float16_is_signaling_nan(f16, fpst)) { - float_raise(float_flag_invalid, fpst); - nan = float16_silence_nan(f16, fpst); - } - if (fpst->default_nan_mode) { - nan = float16_default_nan(fpst); - } - return nan; - } else if (float16_is_infinity(f16)) { - return float16_set_sign(float16_zero, float16_is_neg(f16)); - } else if (float16_is_zero(f16)) { - float_raise(float_flag_divbyzero, fpst); - return float16_set_sign(float16_infinity, float16_is_neg(f16)); - } else if (float16_abs(f16) < (1 << 8)) { - /* Abs(value) < 2.0^-16 */ - float_raise(float_flag_overflow | float_flag_inexact, fpst); - if (round_to_inf(fpst, f16_sign)) { - return float16_set_sign(float16_infinity, f16_sign); - } else { - return float16_set_sign(float16_maxnorm, f16_sign); - } - } else if (f16_exp >= 29 && fpst->flush_to_zero) { - float_raise(float_flag_underflow, fpst); - return float16_set_sign(float16_zero, float16_is_neg(f16)); - } - - f64_frac = call_recip_estimate(&f16_exp, 29, - ((uint64_t) f16_frac) << (52 - 10)); - - /* result = sign : result_exp<4:0> : fraction<51:42> */ - f16_val = deposit32(0, 15, 1, f16_sign); - f16_val = deposit32(f16_val, 10, 5, f16_exp); - f16_val = deposit32(f16_val, 0, 10, extract64(f64_frac, 52 - 10, 10)); - return make_float16(f16_val); -} - -float32 HELPER(recpe_f32)(float32 input, void *fpstp) -{ - float_status *fpst = fpstp; - float32 f32 = float32_squash_input_denormal(input, fpst); - uint32_t f32_val = float32_val(f32); - bool f32_sign = float32_is_neg(f32); - int f32_exp = extract32(f32_val, 23, 8); - uint32_t f32_frac = extract32(f32_val, 0, 23); - uint64_t f64_frac; - - if (float32_is_any_nan(f32)) { - float32 nan = f32; - if (float32_is_signaling_nan(f32, fpst)) { - float_raise(float_flag_invalid, fpst); - nan = float32_silence_nan(f32, fpst); - } - if (fpst->default_nan_mode) { - nan = float32_default_nan(fpst); - } - return nan; - } else if (float32_is_infinity(f32)) { - return float32_set_sign(float32_zero, float32_is_neg(f32)); - } else if (float32_is_zero(f32)) { - float_raise(float_flag_divbyzero, fpst); - return float32_set_sign(float32_infinity, float32_is_neg(f32)); - } else if (float32_abs(f32) < (1ULL << 21)) { - /* Abs(value) < 2.0^-128 */ - float_raise(float_flag_overflow | float_flag_inexact, fpst); - if (round_to_inf(fpst, f32_sign)) { - return float32_set_sign(float32_infinity, f32_sign); - } else { - return float32_set_sign(float32_maxnorm, f32_sign); - } - } else if (f32_exp >= 253 && fpst->flush_to_zero) { - float_raise(float_flag_underflow, fpst); - return float32_set_sign(float32_zero, float32_is_neg(f32)); - } - - f64_frac = call_recip_estimate(&f32_exp, 253, - ((uint64_t) f32_frac) << (52 - 23)); - - /* result = sign : result_exp<7:0> : fraction<51:29> */ - f32_val = deposit32(0, 31, 1, f32_sign); - f32_val = deposit32(f32_val, 23, 8, f32_exp); - f32_val = deposit32(f32_val, 0, 23, extract64(f64_frac, 52 - 23, 23)); - return make_float32(f32_val); -} - -float64 HELPER(recpe_f64)(float64 input, void *fpstp) -{ - float_status *fpst = fpstp; - float64 f64 = float64_squash_input_denormal(input, fpst); - uint64_t f64_val = float64_val(f64); - bool f64_sign = float64_is_neg(f64); - int f64_exp = extract64(f64_val, 52, 11); - uint64_t f64_frac = extract64(f64_val, 0, 52); - - /* Deal with any special cases */ - if (float64_is_any_nan(f64)) { - float64 nan = f64; - if (float64_is_signaling_nan(f64, fpst)) { - float_raise(float_flag_invalid, fpst); - nan = float64_silence_nan(f64, fpst); - } - if (fpst->default_nan_mode) { - nan = float64_default_nan(fpst); - } - return nan; - } else if (float64_is_infinity(f64)) { - return float64_set_sign(float64_zero, float64_is_neg(f64)); - } else if (float64_is_zero(f64)) { - float_raise(float_flag_divbyzero, fpst); - return float64_set_sign(float64_infinity, float64_is_neg(f64)); - } else if ((f64_val & ~(1ULL << 63)) < (1ULL << 50)) { - /* Abs(value) < 2.0^-1024 */ - float_raise(float_flag_overflow | float_flag_inexact, fpst); - if (round_to_inf(fpst, f64_sign)) { - return float64_set_sign(float64_infinity, f64_sign); - } else { - return float64_set_sign(float64_maxnorm, f64_sign); - } - } else if (f64_exp >= 2045 && fpst->flush_to_zero) { - float_raise(float_flag_underflow, fpst); - return float64_set_sign(float64_zero, float64_is_neg(f64)); - } - - f64_frac = call_recip_estimate(&f64_exp, 2045, f64_frac); - - /* result = sign : result_exp<10:0> : fraction<51:0>; */ - f64_val = deposit64(0, 63, 1, f64_sign); - f64_val = deposit64(f64_val, 52, 11, f64_exp); - f64_val = deposit64(f64_val, 0, 52, f64_frac); - return make_float64(f64_val); -} - -/* The algorithm that must be used to calculate the estimate - * is specified by the ARM ARM. - */ - -static int do_recip_sqrt_estimate(int a) -{ - int b, estimate; - - assert(128 <= a && a < 512); - if (a < 256) { - a = a * 2 + 1; - } else { - a = (a >> 1) << 1; - a = (a + 1) * 2; - } - b = 512; - while (a * (b + 1) * (b + 1) < (1 << 28)) { - b += 1; - } - estimate = (b + 1) / 2; - assert(256 <= estimate && estimate < 512); - - return estimate; -} - - -static uint64_t recip_sqrt_estimate(int *exp , int exp_off, uint64_t frac) -{ - int estimate; - uint32_t scaled; - - if (*exp == 0) { - while (extract64(frac, 51, 1) == 0) { - frac = frac << 1; - *exp -= 1; - } - frac = extract64(frac, 0, 51) << 1; - } - - if (*exp & 1) { - /* scaled = UInt('01':fraction<51:45>) */ - scaled = deposit32(1 << 7, 0, 7, extract64(frac, 45, 7)); - } else { - /* scaled = UInt('1':fraction<51:44>) */ - scaled = deposit32(1 << 8, 0, 8, extract64(frac, 44, 8)); - } - estimate = do_recip_sqrt_estimate(scaled); - - *exp = (exp_off - *exp) / 2; - return extract64(estimate, 0, 8) << 44; -} - -uint32_t HELPER(rsqrte_f16)(uint32_t input, void *fpstp) -{ - float_status *s = fpstp; - float16 f16 = float16_squash_input_denormal(input, s); - uint16_t val = float16_val(f16); - bool f16_sign = float16_is_neg(f16); - int f16_exp = extract32(val, 10, 5); - uint16_t f16_frac = extract32(val, 0, 10); - uint64_t f64_frac; - - if (float16_is_any_nan(f16)) { - float16 nan = f16; - if (float16_is_signaling_nan(f16, s)) { - float_raise(float_flag_invalid, s); - nan = float16_silence_nan(f16, s); - } - if (s->default_nan_mode) { - nan = float16_default_nan(s); - } - return nan; - } else if (float16_is_zero(f16)) { - float_raise(float_flag_divbyzero, s); - return float16_set_sign(float16_infinity, f16_sign); - } else if (f16_sign) { - float_raise(float_flag_invalid, s); - return float16_default_nan(s); - } else if (float16_is_infinity(f16)) { - return float16_zero; - } - - /* Scale and normalize to a double-precision value between 0.25 and 1.0, - * preserving the parity of the exponent. */ - - f64_frac = ((uint64_t) f16_frac) << (52 - 10); - - f64_frac = recip_sqrt_estimate(&f16_exp, 44, f64_frac); - - /* result = sign : result_exp<4:0> : estimate<7:0> : Zeros(2) */ - val = deposit32(0, 15, 1, f16_sign); - val = deposit32(val, 10, 5, f16_exp); - val = deposit32(val, 2, 8, extract64(f64_frac, 52 - 8, 8)); - return make_float16(val); -} - -float32 HELPER(rsqrte_f32)(float32 input, void *fpstp) -{ - float_status *s = fpstp; - float32 f32 = float32_squash_input_denormal(input, s); - uint32_t val = float32_val(f32); - uint32_t f32_sign = float32_is_neg(f32); - int f32_exp = extract32(val, 23, 8); - uint32_t f32_frac = extract32(val, 0, 23); - uint64_t f64_frac; - - if (float32_is_any_nan(f32)) { - float32 nan = f32; - if (float32_is_signaling_nan(f32, s)) { - float_raise(float_flag_invalid, s); - nan = float32_silence_nan(f32, s); - } - if (s->default_nan_mode) { - nan = float32_default_nan(s); - } - return nan; - } else if (float32_is_zero(f32)) { - float_raise(float_flag_divbyzero, s); - return float32_set_sign(float32_infinity, float32_is_neg(f32)); - } else if (float32_is_neg(f32)) { - float_raise(float_flag_invalid, s); - return float32_default_nan(s); - } else if (float32_is_infinity(f32)) { - return float32_zero; - } - - /* Scale and normalize to a double-precision value between 0.25 and 1.0, - * preserving the parity of the exponent. */ - - f64_frac = ((uint64_t) f32_frac) << 29; - - f64_frac = recip_sqrt_estimate(&f32_exp, 380, f64_frac); - - /* result = sign : result_exp<4:0> : estimate<7:0> : Zeros(15) */ - val = deposit32(0, 31, 1, f32_sign); - val = deposit32(val, 23, 8, f32_exp); - val = deposit32(val, 15, 8, extract64(f64_frac, 52 - 8, 8)); - return make_float32(val); -} - -float64 HELPER(rsqrte_f64)(float64 input, void *fpstp) -{ - float_status *s = fpstp; - float64 f64 = float64_squash_input_denormal(input, s); - uint64_t val = float64_val(f64); - bool f64_sign = float64_is_neg(f64); - int f64_exp = extract64(val, 52, 11); - uint64_t f64_frac = extract64(val, 0, 52); - - if (float64_is_any_nan(f64)) { - float64 nan = f64; - if (float64_is_signaling_nan(f64, s)) { - float_raise(float_flag_invalid, s); - nan = float64_silence_nan(f64, s); - } - if (s->default_nan_mode) { - nan = float64_default_nan(s); - } - return nan; - } else if (float64_is_zero(f64)) { - float_raise(float_flag_divbyzero, s); - return float64_set_sign(float64_infinity, float64_is_neg(f64)); - } else if (float64_is_neg(f64)) { - float_raise(float_flag_invalid, s); - return float64_default_nan(s); - } else if (float64_is_infinity(f64)) { - return float64_zero; - } - - f64_frac = recip_sqrt_estimate(&f64_exp, 3068, f64_frac); - - /* result = sign : result_exp<4:0> : estimate<7:0> : Zeros(44) */ - val = deposit64(0, 61, 1, f64_sign); - val = deposit64(val, 52, 11, f64_exp); - val = deposit64(val, 44, 8, extract64(f64_frac, 52 - 8, 8)); - return make_float64(val); -} - -uint32_t HELPER(recpe_u32)(uint32_t a, void *fpstp) -{ - /* float_status *s = fpstp; */ - int input, estimate; - - if ((a & 0x80000000) == 0) { - return 0xffffffff; - } - - input = extract32(a, 23, 9); - estimate = recip_estimate(input); - - return deposit32(0, (32 - 9), 9, estimate); -} - -uint32_t HELPER(rsqrte_u32)(uint32_t a, void *fpstp) -{ - int estimate; - - if ((a & 0xc0000000) == 0) { - return 0xffffffff; - } - - estimate = do_recip_sqrt_estimate(extract32(a, 23, 9)); - - return deposit32(0, 23, 9, estimate); -} - -/* VFPv4 fused multiply-accumulate */ -float32 VFP_HELPER(muladd, s)(float32 a, float32 b, float32 c, void *fpstp) -{ - float_status *fpst = fpstp; - return float32_muladd(a, b, c, 0, fpst); -} - -float64 VFP_HELPER(muladd, d)(float64 a, float64 b, float64 c, void *fpstp) -{ - float_status *fpst = fpstp; - return float64_muladd(a, b, c, 0, fpst); -} - -/* ARMv8 round to integral */ -float32 HELPER(rints_exact)(float32 x, void *fp_status) -{ - return float32_round_to_int(x, fp_status); -} - -float64 HELPER(rintd_exact)(float64 x, void *fp_status) -{ - return float64_round_to_int(x, fp_status); -} - -float32 HELPER(rints)(float32 x, void *fp_status) -{ - int old_flags = get_float_exception_flags(fp_status), new_flags; - float32 ret; - - ret = float32_round_to_int(x, fp_status); - - /* Suppress any inexact exceptions the conversion produced */ - if (!(old_flags & float_flag_inexact)) { - new_flags = get_float_exception_flags(fp_status); - set_float_exception_flags(new_flags & ~float_flag_inexact, fp_status); - } - - return ret; -} - -float64 HELPER(rintd)(float64 x, void *fp_status) -{ - int old_flags = get_float_exception_flags(fp_status), new_flags; - float64 ret; - - ret = float64_round_to_int(x, fp_status); - - new_flags = get_float_exception_flags(fp_status); - - /* Suppress any inexact exceptions the conversion produced */ - if (!(old_flags & float_flag_inexact)) { - new_flags = get_float_exception_flags(fp_status); - set_float_exception_flags(new_flags & ~float_flag_inexact, fp_status); - } - - return ret; -} - -/* Convert ARM rounding mode to softfloat */ -int arm_rmode_to_sf(int rmode) -{ - switch (rmode) { - case FPROUNDING_TIEAWAY: - rmode = float_round_ties_away; - break; - case FPROUNDING_ODD: - /* FIXME: add support for TIEAWAY and ODD */ - qemu_log_mask(LOG_UNIMP, "arm: unimplemented rounding mode: %d\n", - rmode); - /* fall through for now */ - case FPROUNDING_TIEEVEN: - default: - rmode = float_round_nearest_even; - break; - case FPROUNDING_POSINF: - rmode = float_round_up; - break; - case FPROUNDING_NEGINF: - rmode = float_round_down; - break; - case FPROUNDING_ZERO: - rmode = float_round_to_zero; - break; - } - return rmode; -} - /* CRC helpers. * The upper bytes of val (above the number specified by 'bytes') must have * been zeroed out by the caller. diff --git a/target/arm/vfp_helper.c b/target/arm/vfp_helper.c new file mode 100644 index 0000000000..74d3030c47 --- /dev/null +++ b/target/arm/vfp_helper.c @@ -0,0 +1,1088 @@ +/* + * ARM VFP floating-point operations + * + * Copyright (c) 2003 Fabrice Bellard + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; if not, see . + */ + +#include "qemu/osdep.h" +#include "qemu/log.h" +#include "cpu.h" +#include "exec/helper-proto.h" +#include "fpu/softfloat.h" +#include "internals.h" + + +/* VFP support. We follow the convention used for VFP instructions: + Single precision routines have a "s" suffix, double precision a + "d" suffix. */ + +/* Convert host exception flags to vfp form. */ +static inline int vfp_exceptbits_from_host(int host_bits) +{ + int target_bits = 0; + + if (host_bits & float_flag_invalid) + target_bits |= 1; + if (host_bits & float_flag_divbyzero) + target_bits |= 2; + if (host_bits & float_flag_overflow) + target_bits |= 4; + if (host_bits & (float_flag_underflow | float_flag_output_denormal)) + target_bits |= 8; + if (host_bits & float_flag_inexact) + target_bits |= 0x10; + if (host_bits & float_flag_input_denormal) + target_bits |= 0x80; + return target_bits; +} + +uint32_t HELPER(vfp_get_fpscr)(CPUARMState *env) +{ + uint32_t i, fpscr; + + fpscr = env->vfp.xregs[ARM_VFP_FPSCR] + | (env->vfp.vec_len << 16) + | (env->vfp.vec_stride << 20); + + i = get_float_exception_flags(&env->vfp.fp_status); + i |= get_float_exception_flags(&env->vfp.standard_fp_status); + /* FZ16 does not generate an input denormal exception. */ + i |= (get_float_exception_flags(&env->vfp.fp_status_f16) + & ~float_flag_input_denormal); + fpscr |= vfp_exceptbits_from_host(i); + + i = env->vfp.qc[0] | env->vfp.qc[1] | env->vfp.qc[2] | env->vfp.qc[3]; + fpscr |= i ? FPCR_QC : 0; + + return fpscr; +} + +uint32_t vfp_get_fpscr(CPUARMState *env) +{ + return HELPER(vfp_get_fpscr)(env); +} + +/* Convert vfp exception flags to target form. */ +static inline int vfp_exceptbits_to_host(int target_bits) +{ + int host_bits = 0; + + if (target_bits & 1) + host_bits |= float_flag_invalid; + if (target_bits & 2) + host_bits |= float_flag_divbyzero; + if (target_bits & 4) + host_bits |= float_flag_overflow; + if (target_bits & 8) + host_bits |= float_flag_underflow; + if (target_bits & 0x10) + host_bits |= float_flag_inexact; + if (target_bits & 0x80) + host_bits |= float_flag_input_denormal; + return host_bits; +} + +void HELPER(vfp_set_fpscr)(CPUARMState *env, uint32_t val) +{ + int i; + uint32_t changed = env->vfp.xregs[ARM_VFP_FPSCR]; + + /* When ARMv8.2-FP16 is not supported, FZ16 is RES0. */ + if (!cpu_isar_feature(aa64_fp16, arm_env_get_cpu(env))) { + val &= ~FPCR_FZ16; + } + + /* + * We don't implement trapped exception handling, so the + * trap enable bits, IDE|IXE|UFE|OFE|DZE|IOE are all RAZ/WI (not RES0!) + * + * If we exclude the exception flags, IOC|DZC|OFC|UFC|IXC|IDC + * (which are stored in fp_status), and the other RES0 bits + * in between, then we clear all of the low 16 bits. + */ + env->vfp.xregs[ARM_VFP_FPSCR] = val & 0xf7c80000; + env->vfp.vec_len = (val >> 16) & 7; + env->vfp.vec_stride = (val >> 20) & 3; + + /* + * The bit we set within fpscr_q is arbitrary; the register as a + * whole being zero/non-zero is what counts. + */ + env->vfp.qc[0] = val & FPCR_QC; + env->vfp.qc[1] = 0; + env->vfp.qc[2] = 0; + env->vfp.qc[3] = 0; + + changed ^= val; + if (changed & (3 << 22)) { + i = (val >> 22) & 3; + switch (i) { + case FPROUNDING_TIEEVEN: + i = float_round_nearest_even; + break; + case FPROUNDING_POSINF: + i = float_round_up; + break; + case FPROUNDING_NEGINF: + i = float_round_down; + break; + case FPROUNDING_ZERO: + i = float_round_to_zero; + break; + } + set_float_rounding_mode(i, &env->vfp.fp_status); + set_float_rounding_mode(i, &env->vfp.fp_status_f16); + } + if (changed & FPCR_FZ16) { + bool ftz_enabled = val & FPCR_FZ16; + set_flush_to_zero(ftz_enabled, &env->vfp.fp_status_f16); + set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status_f16); + } + if (changed & FPCR_FZ) { + bool ftz_enabled = val & FPCR_FZ; + set_flush_to_zero(ftz_enabled, &env->vfp.fp_status); + set_flush_inputs_to_zero(ftz_enabled, &env->vfp.fp_status); + } + if (changed & FPCR_DN) { + bool dnan_enabled = val & FPCR_DN; + set_default_nan_mode(dnan_enabled, &env->vfp.fp_status); + set_default_nan_mode(dnan_enabled, &env->vfp.fp_status_f16); + } + + /* The exception flags are ORed together when we read fpscr so we + * only need to preserve the current state in one of our + * float_status values. + */ + i = vfp_exceptbits_to_host(val); + set_float_exception_flags(i, &env->vfp.fp_status); + set_float_exception_flags(0, &env->vfp.fp_status_f16); + set_float_exception_flags(0, &env->vfp.standard_fp_status); +} + +void vfp_set_fpscr(CPUARMState *env, uint32_t val) +{ + HELPER(vfp_set_fpscr)(env, val); +} + +#define VFP_HELPER(name, p) HELPER(glue(glue(vfp_,name),p)) + +#define VFP_BINOP(name) \ +float32 VFP_HELPER(name, s)(float32 a, float32 b, void *fpstp) \ +{ \ + float_status *fpst = fpstp; \ + return float32_ ## name(a, b, fpst); \ +} \ +float64 VFP_HELPER(name, d)(float64 a, float64 b, void *fpstp) \ +{ \ + float_status *fpst = fpstp; \ + return float64_ ## name(a, b, fpst); \ +} +VFP_BINOP(add) +VFP_BINOP(sub) +VFP_BINOP(mul) +VFP_BINOP(div) +VFP_BINOP(min) +VFP_BINOP(max) +VFP_BINOP(minnum) +VFP_BINOP(maxnum) +#undef VFP_BINOP + +float32 VFP_HELPER(neg, s)(float32 a) +{ + return float32_chs(a); +} + +float64 VFP_HELPER(neg, d)(float64 a) +{ + return float64_chs(a); +} + +float32 VFP_HELPER(abs, s)(float32 a) +{ + return float32_abs(a); +} + +float64 VFP_HELPER(abs, d)(float64 a) +{ + return float64_abs(a); +} + +float32 VFP_HELPER(sqrt, s)(float32 a, CPUARMState *env) +{ + return float32_sqrt(a, &env->vfp.fp_status); +} + +float64 VFP_HELPER(sqrt, d)(float64 a, CPUARMState *env) +{ + return float64_sqrt(a, &env->vfp.fp_status); +} + +static void softfloat_to_vfp_compare(CPUARMState *env, int cmp) +{ + uint32_t flags; + switch (cmp) { + case float_relation_equal: + flags = 0x6; + break; + case float_relation_less: + flags = 0x8; + break; + case float_relation_greater: + flags = 0x2; + break; + case float_relation_unordered: + flags = 0x3; + break; + default: + g_assert_not_reached(); + } + env->vfp.xregs[ARM_VFP_FPSCR] = + deposit32(env->vfp.xregs[ARM_VFP_FPSCR], 28, 4, flags); +} + +/* XXX: check quiet/signaling case */ +#define DO_VFP_cmp(p, type) \ +void VFP_HELPER(cmp, p)(type a, type b, CPUARMState *env) \ +{ \ + softfloat_to_vfp_compare(env, \ + type ## _compare_quiet(a, b, &env->vfp.fp_status)); \ +} \ +void VFP_HELPER(cmpe, p)(type a, type b, CPUARMState *env) \ +{ \ + softfloat_to_vfp_compare(env, \ + type ## _compare(a, b, &env->vfp.fp_status)); \ +} +DO_VFP_cmp(s, float32) +DO_VFP_cmp(d, float64) +#undef DO_VFP_cmp + +/* Integer to float and float to integer conversions */ + +#define CONV_ITOF(name, ftype, fsz, sign) \ +ftype HELPER(name)(uint32_t x, void *fpstp) \ +{ \ + float_status *fpst = fpstp; \ + return sign##int32_to_##float##fsz((sign##int32_t)x, fpst); \ +} + +#define CONV_FTOI(name, ftype, fsz, sign, round) \ +sign##int32_t HELPER(name)(ftype x, void *fpstp) \ +{ \ + float_status *fpst = fpstp; \ + if (float##fsz##_is_any_nan(x)) { \ + float_raise(float_flag_invalid, fpst); \ + return 0; \ + } \ + return float##fsz##_to_##sign##int32##round(x, fpst); \ +} + +#define FLOAT_CONVS(name, p, ftype, fsz, sign) \ + CONV_ITOF(vfp_##name##to##p, ftype, fsz, sign) \ + CONV_FTOI(vfp_to##name##p, ftype, fsz, sign, ) \ + CONV_FTOI(vfp_to##name##z##p, ftype, fsz, sign, _round_to_zero) + +FLOAT_CONVS(si, h, uint32_t, 16, ) +FLOAT_CONVS(si, s, float32, 32, ) +FLOAT_CONVS(si, d, float64, 64, ) +FLOAT_CONVS(ui, h, uint32_t, 16, u) +FLOAT_CONVS(ui, s, float32, 32, u) +FLOAT_CONVS(ui, d, float64, 64, u) + +#undef CONV_ITOF +#undef CONV_FTOI +#undef FLOAT_CONVS + +/* floating point conversion */ +float64 VFP_HELPER(fcvtd, s)(float32 x, CPUARMState *env) +{ + return float32_to_float64(x, &env->vfp.fp_status); +} + +float32 VFP_HELPER(fcvts, d)(float64 x, CPUARMState *env) +{ + return float64_to_float32(x, &env->vfp.fp_status); +} + +/* VFP3 fixed point conversion. */ +#define VFP_CONV_FIX_FLOAT(name, p, fsz, isz, itype) \ +float##fsz HELPER(vfp_##name##to##p)(uint##isz##_t x, uint32_t shift, \ + void *fpstp) \ +{ return itype##_to_##float##fsz##_scalbn(x, -shift, fpstp); } + +#define VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, isz, itype, ROUND, suff) \ +uint##isz##_t HELPER(vfp_to##name##p##suff)(float##fsz x, uint32_t shift, \ + void *fpst) \ +{ \ + if (unlikely(float##fsz##_is_any_nan(x))) { \ + float_raise(float_flag_invalid, fpst); \ + return 0; \ + } \ + return float##fsz##_to_##itype##_scalbn(x, ROUND, shift, fpst); \ +} + +#define VFP_CONV_FIX(name, p, fsz, isz, itype) \ +VFP_CONV_FIX_FLOAT(name, p, fsz, isz, itype) \ +VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, isz, itype, \ + float_round_to_zero, _round_to_zero) \ +VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, isz, itype, \ + get_float_rounding_mode(fpst), ) + +#define VFP_CONV_FIX_A64(name, p, fsz, isz, itype) \ +VFP_CONV_FIX_FLOAT(name, p, fsz, isz, itype) \ +VFP_CONV_FLOAT_FIX_ROUND(name, p, fsz, isz, itype, \ + get_float_rounding_mode(fpst), ) + +VFP_CONV_FIX(sh, d, 64, 64, int16) +VFP_CONV_FIX(sl, d, 64, 64, int32) +VFP_CONV_FIX_A64(sq, d, 64, 64, int64) +VFP_CONV_FIX(uh, d, 64, 64, uint16) +VFP_CONV_FIX(ul, d, 64, 64, uint32) +VFP_CONV_FIX_A64(uq, d, 64, 64, uint64) +VFP_CONV_FIX(sh, s, 32, 32, int16) +VFP_CONV_FIX(sl, s, 32, 32, int32) +VFP_CONV_FIX_A64(sq, s, 32, 64, int64) +VFP_CONV_FIX(uh, s, 32, 32, uint16) +VFP_CONV_FIX(ul, s, 32, 32, uint32) +VFP_CONV_FIX_A64(uq, s, 32, 64, uint64) + +#undef VFP_CONV_FIX +#undef VFP_CONV_FIX_FLOAT +#undef VFP_CONV_FLOAT_FIX_ROUND +#undef VFP_CONV_FIX_A64 + +uint32_t HELPER(vfp_sltoh)(uint32_t x, uint32_t shift, void *fpst) +{ + return int32_to_float16_scalbn(x, -shift, fpst); +} + +uint32_t HELPER(vfp_ultoh)(uint32_t x, uint32_t shift, void *fpst) +{ + return uint32_to_float16_scalbn(x, -shift, fpst); +} + +uint32_t HELPER(vfp_sqtoh)(uint64_t x, uint32_t shift, void *fpst) +{ + return int64_to_float16_scalbn(x, -shift, fpst); +} + +uint32_t HELPER(vfp_uqtoh)(uint64_t x, uint32_t shift, void *fpst) +{ + return uint64_to_float16_scalbn(x, -shift, fpst); +} + +uint32_t HELPER(vfp_toshh)(uint32_t x, uint32_t shift, void *fpst) +{ + if (unlikely(float16_is_any_nan(x))) { + float_raise(float_flag_invalid, fpst); + return 0; + } + return float16_to_int16_scalbn(x, get_float_rounding_mode(fpst), + shift, fpst); +} + +uint32_t HELPER(vfp_touhh)(uint32_t x, uint32_t shift, void *fpst) +{ + if (unlikely(float16_is_any_nan(x))) { + float_raise(float_flag_invalid, fpst); + return 0; + } + return float16_to_uint16_scalbn(x, get_float_rounding_mode(fpst), + shift, fpst); +} + +uint32_t HELPER(vfp_toslh)(uint32_t x, uint32_t shift, void *fpst) +{ + if (unlikely(float16_is_any_nan(x))) { + float_raise(float_flag_invalid, fpst); + return 0; + } + return float16_to_int32_scalbn(x, get_float_rounding_mode(fpst), + shift, fpst); +} + +uint32_t HELPER(vfp_toulh)(uint32_t x, uint32_t shift, void *fpst) +{ + if (unlikely(float16_is_any_nan(x))) { + float_raise(float_flag_invalid, fpst); + return 0; + } + return float16_to_uint32_scalbn(x, get_float_rounding_mode(fpst), + shift, fpst); +} + +uint64_t HELPER(vfp_tosqh)(uint32_t x, uint32_t shift, void *fpst) +{ + if (unlikely(float16_is_any_nan(x))) { + float_raise(float_flag_invalid, fpst); + return 0; + } + return float16_to_int64_scalbn(x, get_float_rounding_mode(fpst), + shift, fpst); +} + +uint64_t HELPER(vfp_touqh)(uint32_t x, uint32_t shift, void *fpst) +{ + if (unlikely(float16_is_any_nan(x))) { + float_raise(float_flag_invalid, fpst); + return 0; + } + return float16_to_uint64_scalbn(x, get_float_rounding_mode(fpst), + shift, fpst); +} + +/* Set the current fp rounding mode and return the old one. + * The argument is a softfloat float_round_ value. + */ +uint32_t HELPER(set_rmode)(uint32_t rmode, void *fpstp) +{ + float_status *fp_status = fpstp; + + uint32_t prev_rmode = get_float_rounding_mode(fp_status); + set_float_rounding_mode(rmode, fp_status); + + return prev_rmode; +} + +/* Set the current fp rounding mode in the standard fp status and return + * the old one. This is for NEON instructions that need to change the + * rounding mode but wish to use the standard FPSCR values for everything + * else. Always set the rounding mode back to the correct value after + * modifying it. + * The argument is a softfloat float_round_ value. + */ +uint32_t HELPER(set_neon_rmode)(uint32_t rmode, CPUARMState *env) +{ + float_status *fp_status = &env->vfp.standard_fp_status; + + uint32_t prev_rmode = get_float_rounding_mode(fp_status); + set_float_rounding_mode(rmode, fp_status); + + return prev_rmode; +} + +/* Half precision conversions. */ +float32 HELPER(vfp_fcvt_f16_to_f32)(uint32_t a, void *fpstp, uint32_t ahp_mode) +{ + /* Squash FZ16 to 0 for the duration of conversion. In this case, + * it would affect flushing input denormals. + */ + float_status *fpst = fpstp; + flag save = get_flush_inputs_to_zero(fpst); + set_flush_inputs_to_zero(false, fpst); + float32 r = float16_to_float32(a, !ahp_mode, fpst); + set_flush_inputs_to_zero(save, fpst); + return r; +} + +uint32_t HELPER(vfp_fcvt_f32_to_f16)(float32 a, void *fpstp, uint32_t ahp_mode) +{ + /* Squash FZ16 to 0 for the duration of conversion. In this case, + * it would affect flushing output denormals. + */ + float_status *fpst = fpstp; + flag save = get_flush_to_zero(fpst); + set_flush_to_zero(false, fpst); + float16 r = float32_to_float16(a, !ahp_mode, fpst); + set_flush_to_zero(save, fpst); + return r; +} + +float64 HELPER(vfp_fcvt_f16_to_f64)(uint32_t a, void *fpstp, uint32_t ahp_mode) +{ + /* Squash FZ16 to 0 for the duration of conversion. In this case, + * it would affect flushing input denormals. + */ + float_status *fpst = fpstp; + flag save = get_flush_inputs_to_zero(fpst); + set_flush_inputs_to_zero(false, fpst); + float64 r = float16_to_float64(a, !ahp_mode, fpst); + set_flush_inputs_to_zero(save, fpst); + return r; +} + +uint32_t HELPER(vfp_fcvt_f64_to_f16)(float64 a, void *fpstp, uint32_t ahp_mode) +{ + /* Squash FZ16 to 0 for the duration of conversion. In this case, + * it would affect flushing output denormals. + */ + float_status *fpst = fpstp; + flag save = get_flush_to_zero(fpst); + set_flush_to_zero(false, fpst); + float16 r = float64_to_float16(a, !ahp_mode, fpst); + set_flush_to_zero(save, fpst); + return r; +} + +#define float32_two make_float32(0x40000000) +#define float32_three make_float32(0x40400000) +#define float32_one_point_five make_float32(0x3fc00000) + +float32 HELPER(recps_f32)(float32 a, float32 b, CPUARMState *env) +{ + float_status *s = &env->vfp.standard_fp_status; + if ((float32_is_infinity(a) && float32_is_zero_or_denormal(b)) || + (float32_is_infinity(b) && float32_is_zero_or_denormal(a))) { + if (!(float32_is_zero(a) || float32_is_zero(b))) { + float_raise(float_flag_input_denormal, s); + } + return float32_two; + } + return float32_sub(float32_two, float32_mul(a, b, s), s); +} + +float32 HELPER(rsqrts_f32)(float32 a, float32 b, CPUARMState *env) +{ + float_status *s = &env->vfp.standard_fp_status; + float32 product; + if ((float32_is_infinity(a) && float32_is_zero_or_denormal(b)) || + (float32_is_infinity(b) && float32_is_zero_or_denormal(a))) { + if (!(float32_is_zero(a) || float32_is_zero(b))) { + float_raise(float_flag_input_denormal, s); + } + return float32_one_point_five; + } + product = float32_mul(a, b, s); + return float32_div(float32_sub(float32_three, product, s), float32_two, s); +} + +/* NEON helpers. */ + +/* Constants 256 and 512 are used in some helpers; we avoid relying on + * int->float conversions at run-time. */ +#define float64_256 make_float64(0x4070000000000000LL) +#define float64_512 make_float64(0x4080000000000000LL) +#define float16_maxnorm make_float16(0x7bff) +#define float32_maxnorm make_float32(0x7f7fffff) +#define float64_maxnorm make_float64(0x7fefffffffffffffLL) + +/* Reciprocal functions + * + * The algorithm that must be used to calculate the estimate + * is specified by the ARM ARM, see FPRecipEstimate()/RecipEstimate + */ + +/* See RecipEstimate() + * + * input is a 9 bit fixed point number + * input range 256 .. 511 for a number from 0.5 <= x < 1.0. + * result range 256 .. 511 for a number from 1.0 to 511/256. + */ + +static int recip_estimate(int input) +{ + int a, b, r; + assert(256 <= input && input < 512); + a = (input * 2) + 1; + b = (1 << 19) / a; + r = (b + 1) >> 1; + assert(256 <= r && r < 512); + return r; +} + +/* + * Common wrapper to call recip_estimate + * + * The parameters are exponent and 64 bit fraction (without implicit + * bit) where the binary point is nominally at bit 52. Returns a + * float64 which can then be rounded to the appropriate size by the + * callee. + */ + +static uint64_t call_recip_estimate(int *exp, int exp_off, uint64_t frac) +{ + uint32_t scaled, estimate; + uint64_t result_frac; + int result_exp; + + /* Handle sub-normals */ + if (*exp == 0) { + if (extract64(frac, 51, 1) == 0) { + *exp = -1; + frac <<= 2; + } else { + frac <<= 1; + } + } + + /* scaled = UInt('1':fraction<51:44>) */ + scaled = deposit32(1 << 8, 0, 8, extract64(frac, 44, 8)); + estimate = recip_estimate(scaled); + + result_exp = exp_off - *exp; + result_frac = deposit64(0, 44, 8, estimate); + if (result_exp == 0) { + result_frac = deposit64(result_frac >> 1, 51, 1, 1); + } else if (result_exp == -1) { + result_frac = deposit64(result_frac >> 2, 50, 2, 1); + result_exp = 0; + } + + *exp = result_exp; + + return result_frac; +} + +static bool round_to_inf(float_status *fpst, bool sign_bit) +{ + switch (fpst->float_rounding_mode) { + case float_round_nearest_even: /* Round to Nearest */ + return true; + case float_round_up: /* Round to +Inf */ + return !sign_bit; + case float_round_down: /* Round to -Inf */ + return sign_bit; + case float_round_to_zero: /* Round to Zero */ + return false; + } + + g_assert_not_reached(); +} + +uint32_t HELPER(recpe_f16)(uint32_t input, void *fpstp) +{ + float_status *fpst = fpstp; + float16 f16 = float16_squash_input_denormal(input, fpst); + uint32_t f16_val = float16_val(f16); + uint32_t f16_sign = float16_is_neg(f16); + int f16_exp = extract32(f16_val, 10, 5); + uint32_t f16_frac = extract32(f16_val, 0, 10); + uint64_t f64_frac; + + if (float16_is_any_nan(f16)) { + float16 nan = f16; + if (float16_is_signaling_nan(f16, fpst)) { + float_raise(float_flag_invalid, fpst); + nan = float16_silence_nan(f16, fpst); + } + if (fpst->default_nan_mode) { + nan = float16_default_nan(fpst); + } + return nan; + } else if (float16_is_infinity(f16)) { + return float16_set_sign(float16_zero, float16_is_neg(f16)); + } else if (float16_is_zero(f16)) { + float_raise(float_flag_divbyzero, fpst); + return float16_set_sign(float16_infinity, float16_is_neg(f16)); + } else if (float16_abs(f16) < (1 << 8)) { + /* Abs(value) < 2.0^-16 */ + float_raise(float_flag_overflow | float_flag_inexact, fpst); + if (round_to_inf(fpst, f16_sign)) { + return float16_set_sign(float16_infinity, f16_sign); + } else { + return float16_set_sign(float16_maxnorm, f16_sign); + } + } else if (f16_exp >= 29 && fpst->flush_to_zero) { + float_raise(float_flag_underflow, fpst); + return float16_set_sign(float16_zero, float16_is_neg(f16)); + } + + f64_frac = call_recip_estimate(&f16_exp, 29, + ((uint64_t) f16_frac) << (52 - 10)); + + /* result = sign : result_exp<4:0> : fraction<51:42> */ + f16_val = deposit32(0, 15, 1, f16_sign); + f16_val = deposit32(f16_val, 10, 5, f16_exp); + f16_val = deposit32(f16_val, 0, 10, extract64(f64_frac, 52 - 10, 10)); + return make_float16(f16_val); +} + +float32 HELPER(recpe_f32)(float32 input, void *fpstp) +{ + float_status *fpst = fpstp; + float32 f32 = float32_squash_input_denormal(input, fpst); + uint32_t f32_val = float32_val(f32); + bool f32_sign = float32_is_neg(f32); + int f32_exp = extract32(f32_val, 23, 8); + uint32_t f32_frac = extract32(f32_val, 0, 23); + uint64_t f64_frac; + + if (float32_is_any_nan(f32)) { + float32 nan = f32; + if (float32_is_signaling_nan(f32, fpst)) { + float_raise(float_flag_invalid, fpst); + nan = float32_silence_nan(f32, fpst); + } + if (fpst->default_nan_mode) { + nan = float32_default_nan(fpst); + } + return nan; + } else if (float32_is_infinity(f32)) { + return float32_set_sign(float32_zero, float32_is_neg(f32)); + } else if (float32_is_zero(f32)) { + float_raise(float_flag_divbyzero, fpst); + return float32_set_sign(float32_infinity, float32_is_neg(f32)); + } else if (float32_abs(f32) < (1ULL << 21)) { + /* Abs(value) < 2.0^-128 */ + float_raise(float_flag_overflow | float_flag_inexact, fpst); + if (round_to_inf(fpst, f32_sign)) { + return float32_set_sign(float32_infinity, f32_sign); + } else { + return float32_set_sign(float32_maxnorm, f32_sign); + } + } else if (f32_exp >= 253 && fpst->flush_to_zero) { + float_raise(float_flag_underflow, fpst); + return float32_set_sign(float32_zero, float32_is_neg(f32)); + } + + f64_frac = call_recip_estimate(&f32_exp, 253, + ((uint64_t) f32_frac) << (52 - 23)); + + /* result = sign : result_exp<7:0> : fraction<51:29> */ + f32_val = deposit32(0, 31, 1, f32_sign); + f32_val = deposit32(f32_val, 23, 8, f32_exp); + f32_val = deposit32(f32_val, 0, 23, extract64(f64_frac, 52 - 23, 23)); + return make_float32(f32_val); +} + +float64 HELPER(recpe_f64)(float64 input, void *fpstp) +{ + float_status *fpst = fpstp; + float64 f64 = float64_squash_input_denormal(input, fpst); + uint64_t f64_val = float64_val(f64); + bool f64_sign = float64_is_neg(f64); + int f64_exp = extract64(f64_val, 52, 11); + uint64_t f64_frac = extract64(f64_val, 0, 52); + + /* Deal with any special cases */ + if (float64_is_any_nan(f64)) { + float64 nan = f64; + if (float64_is_signaling_nan(f64, fpst)) { + float_raise(float_flag_invalid, fpst); + nan = float64_silence_nan(f64, fpst); + } + if (fpst->default_nan_mode) { + nan = float64_default_nan(fpst); + } + return nan; + } else if (float64_is_infinity(f64)) { + return float64_set_sign(float64_zero, float64_is_neg(f64)); + } else if (float64_is_zero(f64)) { + float_raise(float_flag_divbyzero, fpst); + return float64_set_sign(float64_infinity, float64_is_neg(f64)); + } else if ((f64_val & ~(1ULL << 63)) < (1ULL << 50)) { + /* Abs(value) < 2.0^-1024 */ + float_raise(float_flag_overflow | float_flag_inexact, fpst); + if (round_to_inf(fpst, f64_sign)) { + return float64_set_sign(float64_infinity, f64_sign); + } else { + return float64_set_sign(float64_maxnorm, f64_sign); + } + } else if (f64_exp >= 2045 && fpst->flush_to_zero) { + float_raise(float_flag_underflow, fpst); + return float64_set_sign(float64_zero, float64_is_neg(f64)); + } + + f64_frac = call_recip_estimate(&f64_exp, 2045, f64_frac); + + /* result = sign : result_exp<10:0> : fraction<51:0>; */ + f64_val = deposit64(0, 63, 1, f64_sign); + f64_val = deposit64(f64_val, 52, 11, f64_exp); + f64_val = deposit64(f64_val, 0, 52, f64_frac); + return make_float64(f64_val); +} + +/* The algorithm that must be used to calculate the estimate + * is specified by the ARM ARM. + */ + +static int do_recip_sqrt_estimate(int a) +{ + int b, estimate; + + assert(128 <= a && a < 512); + if (a < 256) { + a = a * 2 + 1; + } else { + a = (a >> 1) << 1; + a = (a + 1) * 2; + } + b = 512; + while (a * (b + 1) * (b + 1) < (1 << 28)) { + b += 1; + } + estimate = (b + 1) / 2; + assert(256 <= estimate && estimate < 512); + + return estimate; +} + + +static uint64_t recip_sqrt_estimate(int *exp , int exp_off, uint64_t frac) +{ + int estimate; + uint32_t scaled; + + if (*exp == 0) { + while (extract64(frac, 51, 1) == 0) { + frac = frac << 1; + *exp -= 1; + } + frac = extract64(frac, 0, 51) << 1; + } + + if (*exp & 1) { + /* scaled = UInt('01':fraction<51:45>) */ + scaled = deposit32(1 << 7, 0, 7, extract64(frac, 45, 7)); + } else { + /* scaled = UInt('1':fraction<51:44>) */ + scaled = deposit32(1 << 8, 0, 8, extract64(frac, 44, 8)); + } + estimate = do_recip_sqrt_estimate(scaled); + + *exp = (exp_off - *exp) / 2; + return extract64(estimate, 0, 8) << 44; +} + +uint32_t HELPER(rsqrte_f16)(uint32_t input, void *fpstp) +{ + float_status *s = fpstp; + float16 f16 = float16_squash_input_denormal(input, s); + uint16_t val = float16_val(f16); + bool f16_sign = float16_is_neg(f16); + int f16_exp = extract32(val, 10, 5); + uint16_t f16_frac = extract32(val, 0, 10); + uint64_t f64_frac; + + if (float16_is_any_nan(f16)) { + float16 nan = f16; + if (float16_is_signaling_nan(f16, s)) { + float_raise(float_flag_invalid, s); + nan = float16_silence_nan(f16, s); + } + if (s->default_nan_mode) { + nan = float16_default_nan(s); + } + return nan; + } else if (float16_is_zero(f16)) { + float_raise(float_flag_divbyzero, s); + return float16_set_sign(float16_infinity, f16_sign); + } else if (f16_sign) { + float_raise(float_flag_invalid, s); + return float16_default_nan(s); + } else if (float16_is_infinity(f16)) { + return float16_zero; + } + + /* Scale and normalize to a double-precision value between 0.25 and 1.0, + * preserving the parity of the exponent. */ + + f64_frac = ((uint64_t) f16_frac) << (52 - 10); + + f64_frac = recip_sqrt_estimate(&f16_exp, 44, f64_frac); + + /* result = sign : result_exp<4:0> : estimate<7:0> : Zeros(2) */ + val = deposit32(0, 15, 1, f16_sign); + val = deposit32(val, 10, 5, f16_exp); + val = deposit32(val, 2, 8, extract64(f64_frac, 52 - 8, 8)); + return make_float16(val); +} + +float32 HELPER(rsqrte_f32)(float32 input, void *fpstp) +{ + float_status *s = fpstp; + float32 f32 = float32_squash_input_denormal(input, s); + uint32_t val = float32_val(f32); + uint32_t f32_sign = float32_is_neg(f32); + int f32_exp = extract32(val, 23, 8); + uint32_t f32_frac = extract32(val, 0, 23); + uint64_t f64_frac; + + if (float32_is_any_nan(f32)) { + float32 nan = f32; + if (float32_is_signaling_nan(f32, s)) { + float_raise(float_flag_invalid, s); + nan = float32_silence_nan(f32, s); + } + if (s->default_nan_mode) { + nan = float32_default_nan(s); + } + return nan; + } else if (float32_is_zero(f32)) { + float_raise(float_flag_divbyzero, s); + return float32_set_sign(float32_infinity, float32_is_neg(f32)); + } else if (float32_is_neg(f32)) { + float_raise(float_flag_invalid, s); + return float32_default_nan(s); + } else if (float32_is_infinity(f32)) { + return float32_zero; + } + + /* Scale and normalize to a double-precision value between 0.25 and 1.0, + * preserving the parity of the exponent. */ + + f64_frac = ((uint64_t) f32_frac) << 29; + + f64_frac = recip_sqrt_estimate(&f32_exp, 380, f64_frac); + + /* result = sign : result_exp<4:0> : estimate<7:0> : Zeros(15) */ + val = deposit32(0, 31, 1, f32_sign); + val = deposit32(val, 23, 8, f32_exp); + val = deposit32(val, 15, 8, extract64(f64_frac, 52 - 8, 8)); + return make_float32(val); +} + +float64 HELPER(rsqrte_f64)(float64 input, void *fpstp) +{ + float_status *s = fpstp; + float64 f64 = float64_squash_input_denormal(input, s); + uint64_t val = float64_val(f64); + bool f64_sign = float64_is_neg(f64); + int f64_exp = extract64(val, 52, 11); + uint64_t f64_frac = extract64(val, 0, 52); + + if (float64_is_any_nan(f64)) { + float64 nan = f64; + if (float64_is_signaling_nan(f64, s)) { + float_raise(float_flag_invalid, s); + nan = float64_silence_nan(f64, s); + } + if (s->default_nan_mode) { + nan = float64_default_nan(s); + } + return nan; + } else if (float64_is_zero(f64)) { + float_raise(float_flag_divbyzero, s); + return float64_set_sign(float64_infinity, float64_is_neg(f64)); + } else if (float64_is_neg(f64)) { + float_raise(float_flag_invalid, s); + return float64_default_nan(s); + } else if (float64_is_infinity(f64)) { + return float64_zero; + } + + f64_frac = recip_sqrt_estimate(&f64_exp, 3068, f64_frac); + + /* result = sign : result_exp<4:0> : estimate<7:0> : Zeros(44) */ + val = deposit64(0, 61, 1, f64_sign); + val = deposit64(val, 52, 11, f64_exp); + val = deposit64(val, 44, 8, extract64(f64_frac, 52 - 8, 8)); + return make_float64(val); +} + +uint32_t HELPER(recpe_u32)(uint32_t a, void *fpstp) +{ + /* float_status *s = fpstp; */ + int input, estimate; + + if ((a & 0x80000000) == 0) { + return 0xffffffff; + } + + input = extract32(a, 23, 9); + estimate = recip_estimate(input); + + return deposit32(0, (32 - 9), 9, estimate); +} + +uint32_t HELPER(rsqrte_u32)(uint32_t a, void *fpstp) +{ + int estimate; + + if ((a & 0xc0000000) == 0) { + return 0xffffffff; + } + + estimate = do_recip_sqrt_estimate(extract32(a, 23, 9)); + + return deposit32(0, 23, 9, estimate); +} + +/* VFPv4 fused multiply-accumulate */ +float32 VFP_HELPER(muladd, s)(float32 a, float32 b, float32 c, void *fpstp) +{ + float_status *fpst = fpstp; + return float32_muladd(a, b, c, 0, fpst); +} + +float64 VFP_HELPER(muladd, d)(float64 a, float64 b, float64 c, void *fpstp) +{ + float_status *fpst = fpstp; + return float64_muladd(a, b, c, 0, fpst); +} + +/* ARMv8 round to integral */ +float32 HELPER(rints_exact)(float32 x, void *fp_status) +{ + return float32_round_to_int(x, fp_status); +} + +float64 HELPER(rintd_exact)(float64 x, void *fp_status) +{ + return float64_round_to_int(x, fp_status); +} + +float32 HELPER(rints)(float32 x, void *fp_status) +{ + int old_flags = get_float_exception_flags(fp_status), new_flags; + float32 ret; + + ret = float32_round_to_int(x, fp_status); + + /* Suppress any inexact exceptions the conversion produced */ + if (!(old_flags & float_flag_inexact)) { + new_flags = get_float_exception_flags(fp_status); + set_float_exception_flags(new_flags & ~float_flag_inexact, fp_status); + } + + return ret; +} + +float64 HELPER(rintd)(float64 x, void *fp_status) +{ + int old_flags = get_float_exception_flags(fp_status), new_flags; + float64 ret; + + ret = float64_round_to_int(x, fp_status); + + new_flags = get_float_exception_flags(fp_status); + + /* Suppress any inexact exceptions the conversion produced */ + if (!(old_flags & float_flag_inexact)) { + new_flags = get_float_exception_flags(fp_status); + set_float_exception_flags(new_flags & ~float_flag_inexact, fp_status); + } + + return ret; +} + +/* Convert ARM rounding mode to softfloat */ +int arm_rmode_to_sf(int rmode) +{ + switch (rmode) { + case FPROUNDING_TIEAWAY: + rmode = float_round_ties_away; + break; + case FPROUNDING_ODD: + /* FIXME: add support for TIEAWAY and ODD */ + qemu_log_mask(LOG_UNIMP, "arm: unimplemented rounding mode: %d\n", + rmode); + /* fall through for now */ + case FPROUNDING_TIEEVEN: + default: + rmode = float_round_nearest_even; + break; + case FPROUNDING_POSINF: + rmode = float_round_up; + break; + case FPROUNDING_NEGINF: + rmode = float_round_down; + break; + case FPROUNDING_ZERO: + rmode = float_round_to_zero; + break; + } + return rmode; +} From e80941bd64cc388554770fd72334e9e7d459a1ef Mon Sep 17 00:00:00 2001 From: Richard Henderson Date: Thu, 21 Feb 2019 18:17:45 +0000 Subject: [PATCH 06/21] target/arm: Rearrange Floating-point data-processing (2 regs) There are lots of special cases within these insns. Split the major argument decode/loading/saving into no_output (compares), rd_is_dp, and rm_is_dp. We still need to special case argument load for compare (rd as input, rm as zero) and vcvt fixed (rd as input+output), but lots of special cases do disappear. Now that we have a full switch at the beginning, hoist the ISA checks from the code generation. Signed-off-by: Richard Henderson Message-id: 20190215192302.27855-4-richard.henderson@linaro.org Reviewed-by: Peter Maydell Signed-off-by: Peter Maydell --- target/arm/translate.c | 227 ++++++++++++++++++++--------------------- 1 file changed, 111 insertions(+), 116 deletions(-) diff --git a/target/arm/translate.c b/target/arm/translate.c index dac737f6ca..64c5fe0df3 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -3639,52 +3639,108 @@ static int disas_vfp_insn(DisasContext *s, uint32_t insn) } } else { /* data processing */ + bool rd_is_dp = dp; + bool rm_is_dp = dp; + bool no_output = false; + /* The opcode is in bits 23, 21, 20 and 6. */ op = ((insn >> 20) & 8) | ((insn >> 19) & 6) | ((insn >> 6) & 1); - if (dp) { - if (op == 15) { - /* rn is opcode */ - rn = ((insn >> 15) & 0x1e) | ((insn >> 7) & 1); - } else { - /* rn is register number */ - VFP_DREG_N(rn, insn); - } + rn = VFP_SREG_N(insn); - if (op == 15 && (rn == 15 || ((rn & 0x1c) == 0x18) || - ((rn & 0x1e) == 0x6))) { - /* Integer or single/half precision destination. */ - rd = VFP_SREG_D(insn); - } else { - VFP_DREG_D(rd, insn); - } - if (op == 15 && - (((rn & 0x1c) == 0x10) || ((rn & 0x14) == 0x14) || - ((rn & 0x1e) == 0x4))) { - /* VCVT from int or half precision is always from S reg - * regardless of dp bit. VCVT with immediate frac_bits - * has same format as SREG_M. + if (op == 15) { + /* rn is opcode, encoded as per VFP_SREG_N. */ + switch (rn) { + case 0x00: /* vmov */ + case 0x01: /* vabs */ + case 0x02: /* vneg */ + case 0x03: /* vsqrt */ + break; + + case 0x04: /* vcvtb.f64.f16, vcvtb.f32.f16 */ + case 0x05: /* vcvtt.f64.f16, vcvtt.f32.f16 */ + /* + * VCVTB, VCVTT: only present with the halfprec extension + * UNPREDICTABLE if bit 8 is set prior to ARMv8 + * (we choose to UNDEF) */ - rm = VFP_SREG_M(insn); - } else { - VFP_DREG_M(rm, insn); + if ((dp && !arm_dc_feature(s, ARM_FEATURE_V8)) || + !arm_dc_feature(s, ARM_FEATURE_VFP_FP16)) { + return 1; + } + rm_is_dp = false; + break; + case 0x06: /* vcvtb.f16.f32, vcvtb.f16.f64 */ + case 0x07: /* vcvtt.f16.f32, vcvtt.f16.f64 */ + if ((dp && !arm_dc_feature(s, ARM_FEATURE_V8)) || + !arm_dc_feature(s, ARM_FEATURE_VFP_FP16)) { + return 1; + } + rd_is_dp = false; + break; + + case 0x08: case 0x0a: /* vcmp, vcmpz */ + case 0x09: case 0x0b: /* vcmpe, vcmpez */ + no_output = true; + break; + + case 0x0c: /* vrintr */ + case 0x0d: /* vrintz */ + case 0x0e: /* vrintx */ + break; + + case 0x0f: /* vcvt double<->single */ + rd_is_dp = !dp; + break; + + case 0x10: /* vcvt.fxx.u32 */ + case 0x11: /* vcvt.fxx.s32 */ + rm_is_dp = false; + break; + case 0x18: /* vcvtr.u32.fxx */ + case 0x19: /* vcvtz.u32.fxx */ + case 0x1a: /* vcvtr.s32.fxx */ + case 0x1b: /* vcvtz.s32.fxx */ + rd_is_dp = false; + break; + + case 0x14: /* vcvt fp <-> fixed */ + case 0x15: + case 0x16: + case 0x17: + case 0x1c: + case 0x1d: + case 0x1e: + case 0x1f: + if (!arm_dc_feature(s, ARM_FEATURE_VFP3)) { + return 1; + } + /* Immediate frac_bits has same format as SREG_M. */ + rm_is_dp = false; + break; + + default: + return 1; } + } else if (dp) { + /* rn is register number */ + VFP_DREG_N(rn, insn); + } + + if (rd_is_dp) { + VFP_DREG_D(rd, insn); + } else { + rd = VFP_SREG_D(insn); + } + if (rm_is_dp) { + VFP_DREG_M(rm, insn); } else { - rn = VFP_SREG_N(insn); - if (op == 15 && rn == 15) { - /* Double precision destination. */ - VFP_DREG_D(rd, insn); - } else { - rd = VFP_SREG_D(insn); - } - /* NB that we implicitly rely on the encoding for the frac_bits - * in VCVT of fixed to float being the same as that of an SREG_M - */ rm = VFP_SREG_M(insn); } veclen = s->vec_len; - if (op == 15 && rn > 3) + if (op == 15 && rn > 3) { veclen = 0; + } /* Shut up compiler warnings. */ delta_m = 0; @@ -3720,55 +3776,28 @@ static int disas_vfp_insn(DisasContext *s, uint32_t insn) /* Load the initial operands. */ if (op == 15) { switch (rn) { - case 16: - case 17: - /* Integer source */ - gen_mov_F0_vreg(0, rm); - break; - case 8: - case 9: - /* Compare */ + case 0x08: case 0x09: /* Compare */ gen_mov_F0_vreg(dp, rd); gen_mov_F1_vreg(dp, rm); break; - case 10: - case 11: - /* Compare with zero */ + case 0x0a: case 0x0b: /* Compare with zero */ gen_mov_F0_vreg(dp, rd); gen_vfp_F1_ld0(dp); break; - case 20: - case 21: - case 22: - case 23: - case 28: - case 29: - case 30: - case 31: + case 0x14: /* vcvt fp <-> fixed */ + case 0x15: + case 0x16: + case 0x17: + case 0x1c: + case 0x1d: + case 0x1e: + case 0x1f: /* Source and destination the same. */ gen_mov_F0_vreg(dp, rd); break; - case 4: - case 5: - case 6: - case 7: - /* VCVTB, VCVTT: only present with the halfprec extension - * UNPREDICTABLE if bit 8 is set prior to ARMv8 - * (we choose to UNDEF) - */ - if ((dp && !arm_dc_feature(s, ARM_FEATURE_V8)) || - !arm_dc_feature(s, ARM_FEATURE_VFP_FP16)) { - return 1; - } - if (!extract32(rn, 1, 1)) { - /* Half precision source. */ - gen_mov_F0_vreg(0, rm); - break; - } - /* Otherwise fall through */ default: /* One source operand. */ - gen_mov_F0_vreg(dp, rm); + gen_mov_F0_vreg(rm_is_dp, rm); break; } } else { @@ -4047,10 +4076,11 @@ static int disas_vfp_insn(DisasContext *s, uint32_t insn) break; } case 15: /* single<->double conversion */ - if (dp) + if (dp) { gen_helper_vfp_fcvtsd(cpu_F0s, cpu_F0d, cpu_env); - else + } else { gen_helper_vfp_fcvtds(cpu_F0d, cpu_F0s, cpu_env); + } break; case 16: /* fuito */ gen_vfp_uito(dp, 0); @@ -4059,27 +4089,15 @@ static int disas_vfp_insn(DisasContext *s, uint32_t insn) gen_vfp_sito(dp, 0); break; case 20: /* fshto */ - if (!arm_dc_feature(s, ARM_FEATURE_VFP3)) { - return 1; - } gen_vfp_shto(dp, 16 - rm, 0); break; case 21: /* fslto */ - if (!arm_dc_feature(s, ARM_FEATURE_VFP3)) { - return 1; - } gen_vfp_slto(dp, 32 - rm, 0); break; case 22: /* fuhto */ - if (!arm_dc_feature(s, ARM_FEATURE_VFP3)) { - return 1; - } gen_vfp_uhto(dp, 16 - rm, 0); break; case 23: /* fulto */ - if (!arm_dc_feature(s, ARM_FEATURE_VFP3)) { - return 1; - } gen_vfp_ulto(dp, 32 - rm, 0); break; case 24: /* ftoui */ @@ -4095,57 +4113,34 @@ static int disas_vfp_insn(DisasContext *s, uint32_t insn) gen_vfp_tosiz(dp, 0); break; case 28: /* ftosh */ - if (!arm_dc_feature(s, ARM_FEATURE_VFP3)) { - return 1; - } gen_vfp_tosh(dp, 16 - rm, 0); break; case 29: /* ftosl */ - if (!arm_dc_feature(s, ARM_FEATURE_VFP3)) { - return 1; - } gen_vfp_tosl(dp, 32 - rm, 0); break; case 30: /* ftouh */ - if (!arm_dc_feature(s, ARM_FEATURE_VFP3)) { - return 1; - } gen_vfp_touh(dp, 16 - rm, 0); break; case 31: /* ftoul */ - if (!arm_dc_feature(s, ARM_FEATURE_VFP3)) { - return 1; - } gen_vfp_toul(dp, 32 - rm, 0); break; default: /* undefined */ - return 1; + g_assert_not_reached(); } break; default: /* undefined */ return 1; } - /* Write back the result. */ - if (op == 15 && (rn >= 8 && rn <= 11)) { - /* Comparison, do nothing. */ - } else if (op == 15 && dp && ((rn & 0x1c) == 0x18 || - (rn & 0x1e) == 0x6)) { - /* VCVT double to int: always integer result. - * VCVT double to half precision is always a single - * precision result. - */ - gen_mov_vreg_F0(0, rd); - } else if (op == 15 && rn == 15) { - /* conversion */ - gen_mov_vreg_F0(!dp, rd); - } else { - gen_mov_vreg_F0(dp, rd); + /* Write back the result, if any. */ + if (!no_output) { + gen_mov_vreg_F0(rd_is_dp, rd); } /* break out of the loop if we have finished */ - if (veclen == 0) + if (veclen == 0) { break; + } if (op == 15 && delta_m == 0) { /* single source one-many */ From 6c1f6f2733a7692793135ea5ce72b829add99a50 Mon Sep 17 00:00:00 2001 From: Richard Henderson Date: Thu, 21 Feb 2019 18:17:46 +0000 Subject: [PATCH 07/21] target/arm: Implement ARMv8.3-JSConv Signed-off-by: Richard Henderson Message-id: 20190215192302.27855-5-richard.henderson@linaro.org Reviewed-by: Peter Maydell [PMM: fixed a couple of comment typos] Signed-off-by: Peter Maydell --- target/arm/cpu.c | 1 + target/arm/cpu.h | 10 +++++ target/arm/cpu64.c | 2 + target/arm/helper.h | 3 ++ target/arm/translate-a64.c | 26 +++++++++++ target/arm/translate.c | 10 +++++ target/arm/vfp_helper.c | 88 ++++++++++++++++++++++++++++++++++++++ 7 files changed, 140 insertions(+) diff --git a/target/arm/cpu.c b/target/arm/cpu.c index edf6e0e1f1..8ea6569088 100644 --- a/target/arm/cpu.c +++ b/target/arm/cpu.c @@ -2001,6 +2001,7 @@ static void arm_max_initfn(Object *obj) cpu->isar.id_isar5 = t; t = cpu->isar.id_isar6; + t = FIELD_DP32(t, ID_ISAR6, JSCVT, 1); t = FIELD_DP32(t, ID_ISAR6, DP, 1); cpu->isar.id_isar6 = t; diff --git a/target/arm/cpu.h b/target/arm/cpu.h index 84ae6849c2..1eea1a408b 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -3273,6 +3273,11 @@ static inline bool isar_feature_aa32_vcma(const ARMISARegisters *id) return FIELD_EX32(id->id_isar5, ID_ISAR5, VCMA) != 0; } +static inline bool isar_feature_aa32_jscvt(const ARMISARegisters *id) +{ + return FIELD_EX32(id->id_isar6, ID_ISAR6, JSCVT) != 0; +} + static inline bool isar_feature_aa32_dp(const ARMISARegisters *id) { return FIELD_EX32(id->id_isar6, ID_ISAR6, DP) != 0; @@ -3351,6 +3356,11 @@ static inline bool isar_feature_aa64_dp(const ARMISARegisters *id) return FIELD_EX64(id->id_aa64isar0, ID_AA64ISAR0, DP) != 0; } +static inline bool isar_feature_aa64_jscvt(const ARMISARegisters *id) +{ + return FIELD_EX64(id->id_aa64isar1, ID_AA64ISAR1, JSCVT) != 0; +} + static inline bool isar_feature_aa64_fcma(const ARMISARegisters *id) { return FIELD_EX64(id->id_aa64isar1, ID_AA64ISAR1, FCMA) != 0; diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c index eff0f164dd..69e4134f79 100644 --- a/target/arm/cpu64.c +++ b/target/arm/cpu64.c @@ -311,6 +311,7 @@ static void aarch64_max_initfn(Object *obj) cpu->isar.id_aa64isar0 = t; t = cpu->isar.id_aa64isar1; + t = FIELD_DP64(t, ID_AA64ISAR1, JSCVT, 1); t = FIELD_DP64(t, ID_AA64ISAR1, FCMA, 1); t = FIELD_DP64(t, ID_AA64ISAR1, APA, 1); /* PAuth, architected only */ t = FIELD_DP64(t, ID_AA64ISAR1, API, 0); @@ -344,6 +345,7 @@ static void aarch64_max_initfn(Object *obj) cpu->isar.id_isar5 = u; u = cpu->isar.id_isar6; + u = FIELD_DP32(u, ID_ISAR6, JSCVT, 1); u = FIELD_DP32(u, ID_ISAR6, DP, 1); cpu->isar.id_isar6 = u; diff --git a/target/arm/helper.h b/target/arm/helper.h index 923e8e1525..747cb64d29 100644 --- a/target/arm/helper.h +++ b/target/arm/helper.h @@ -218,6 +218,9 @@ DEF_HELPER_FLAGS_2(rintd_exact, TCG_CALL_NO_RWG, f64, f64, ptr) DEF_HELPER_FLAGS_2(rints, TCG_CALL_NO_RWG, f32, f32, ptr) DEF_HELPER_FLAGS_2(rintd, TCG_CALL_NO_RWG, f64, f64, ptr) +DEF_HELPER_FLAGS_2(vjcvt, TCG_CALL_NO_RWG, i32, f64, env) +DEF_HELPER_FLAGS_2(fjcvtzs, TCG_CALL_NO_RWG, i64, f64, ptr) + /* neon_helper.c */ DEF_HELPER_FLAGS_3(neon_qadd_u8, TCG_CALL_NO_RWG, i32, env, i32, i32) DEF_HELPER_FLAGS_3(neon_qadd_s8, TCG_CALL_NO_RWG, i32, env, i32, i32) diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index dbce24fe32..c56e878787 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -6526,6 +6526,24 @@ static void handle_fmov(DisasContext *s, int rd, int rn, int type, bool itof) } } +static void handle_fjcvtzs(DisasContext *s, int rd, int rn) +{ + TCGv_i64 t = read_fp_dreg(s, rn); + TCGv_ptr fpstatus = get_fpstatus_ptr(false); + + gen_helper_fjcvtzs(t, t, fpstatus); + + tcg_temp_free_ptr(fpstatus); + + tcg_gen_ext32u_i64(cpu_reg(s, rd), t); + tcg_gen_extrh_i64_i32(cpu_ZF, t); + tcg_gen_movi_i32(cpu_CF, 0); + tcg_gen_movi_i32(cpu_NF, 0); + tcg_gen_movi_i32(cpu_VF, 0); + + tcg_temp_free_i64(t); +} + /* Floating point <-> integer conversions * 31 30 29 28 24 23 22 21 20 19 18 16 15 10 9 5 4 0 * +----+---+---+-----------+------+---+-------+-----+-------------+----+----+ @@ -6601,6 +6619,14 @@ static void disas_fp_int_conv(DisasContext *s, uint32_t insn) handle_fmov(s, rd, rn, type, itof); break; + case 0b00111110: /* FJCVTZS */ + if (!dc_isar_feature(aa64_jscvt, s)) { + goto do_unallocated; + } else if (fp_access_check(s)) { + handle_fjcvtzs(s, rd, rn); + } + break; + default: do_unallocated: unallocated_encoding(s); diff --git a/target/arm/translate.c b/target/arm/translate.c index 64c5fe0df3..c1175798ac 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -3718,6 +3718,13 @@ static int disas_vfp_insn(DisasContext *s, uint32_t insn) rm_is_dp = false; break; + case 0x13: /* vjcvt */ + if (!dp || !dc_isar_feature(aa32_jscvt, s)) { + return 1; + } + rd_is_dp = false; + break; + default: return 1; } @@ -4088,6 +4095,9 @@ static int disas_vfp_insn(DisasContext *s, uint32_t insn) case 17: /* fsito */ gen_vfp_sito(dp, 0); break; + case 19: /* vjcvt */ + gen_helper_vjcvt(cpu_F0s, cpu_F0d, cpu_env); + break; case 20: /* fshto */ gen_vfp_shto(dp, 16 - rm, 0); break; diff --git a/target/arm/vfp_helper.c b/target/arm/vfp_helper.c index 74d3030c47..cc7f9f5cb1 100644 --- a/target/arm/vfp_helper.c +++ b/target/arm/vfp_helper.c @@ -1086,3 +1086,91 @@ int arm_rmode_to_sf(int rmode) } return rmode; } + +/* + * Implement float64 to int32_t conversion without saturation; + * the result is supplied modulo 2^32. + */ +uint64_t HELPER(fjcvtzs)(float64 value, void *vstatus) +{ + float_status *status = vstatus; + uint32_t exp, sign; + uint64_t frac; + uint32_t inexact = 1; /* !Z */ + + sign = extract64(value, 63, 1); + exp = extract64(value, 52, 11); + frac = extract64(value, 0, 52); + + if (exp == 0) { + /* While not inexact for IEEE FP, -0.0 is inexact for JavaScript. */ + inexact = sign; + if (frac != 0) { + if (status->flush_inputs_to_zero) { + float_raise(float_flag_input_denormal, status); + } else { + float_raise(float_flag_inexact, status); + inexact = 1; + } + } + frac = 0; + } else if (exp == 0x7ff) { + /* This operation raises Invalid for both NaN and overflow (Inf). */ + float_raise(float_flag_invalid, status); + frac = 0; + } else { + int true_exp = exp - 1023; + int shift = true_exp - 52; + + /* Restore implicit bit. */ + frac |= 1ull << 52; + + /* Shift the fraction into place. */ + if (shift >= 0) { + /* The number is so large we must shift the fraction left. */ + if (shift >= 64) { + /* The fraction is shifted out entirely. */ + frac = 0; + } else { + frac <<= shift; + } + } else if (shift > -64) { + /* Normal case -- shift right and notice if bits shift out. */ + inexact = (frac << (64 + shift)) != 0; + frac >>= -shift; + } else { + /* The fraction is shifted out entirely. */ + frac = 0; + } + + /* Notice overflow or inexact exceptions. */ + if (true_exp > 31 || frac > (sign ? 0x80000000ull : 0x7fffffff)) { + /* Overflow, for which this operation raises invalid. */ + float_raise(float_flag_invalid, status); + inexact = 1; + } else if (inexact) { + float_raise(float_flag_inexact, status); + } + + /* Honor the sign. */ + if (sign) { + frac = -frac; + } + } + + /* Pack the result and the env->ZF representation of Z together. */ + return deposit64(frac, 32, 32, inexact); +} + +uint32_t HELPER(vjcvt)(float64 value, CPUARMState *env) +{ + uint64_t pair = HELPER(fjcvtzs)(value, &env->vfp.fp_status); + uint32_t result = pair; + uint32_t z = (pair >> 32) == 0; + + /* Store Z, clear NCV, in FPSCR.NZCV. */ + env->vfp.xregs[ARM_VFP_FPSCR] + = (env->vfp.xregs[ARM_VFP_FPSCR] & ~CPSR_NZCV) | (z * CPSR_Z); + + return result; +} From 37e571f1e0101f9aab29235afa651028723eb4c5 Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Thu, 21 Feb 2019 18:17:46 +0000 Subject: [PATCH 08/21] hw/misc/tz-ppc: Support having unused ports in the middle of the range The Peripheral Protection Controller's handling of unused ports is that if there is nothing connected to the port's downstream then it does not create the sysbus MMIO region for the upstream end of the port. This results in odd behaviour when there is an unused port in the middle of the range: since sysbus MMIO regions are implicitly consecutively allocated, any used ports above the unused ones end up with sysbus MMIO region numbers that don't match the port number. Avoid this numbering mismatch by creating dummy MMIO regions for the unused ports. This doesn't change anything for our existing boards, which don't have any gaps in the middle of the port ranges they use; but it will be needed for the Musca board. Signed-off-by: Peter Maydell Reviewed-by: Richard Henderson --- hw/misc/tz-ppc.c | 32 ++++++++++++++++++++++++++++++++ include/hw/misc/tz-ppc.h | 8 +++++++- 2 files changed, 39 insertions(+), 1 deletion(-) diff --git a/hw/misc/tz-ppc.c b/hw/misc/tz-ppc.c index 3dd045c15f..2e04837bea 100644 --- a/hw/misc/tz-ppc.c +++ b/hw/misc/tz-ppc.c @@ -181,6 +181,21 @@ static const MemoryRegionOps tz_ppc_ops = { .endianness = DEVICE_LITTLE_ENDIAN, }; +static bool tz_ppc_dummy_accepts(void *opaque, hwaddr addr, + unsigned size, bool is_write, + MemTxAttrs attrs) +{ + /* + * Board code should never map the upstream end of an unused port, + * so we should never try to make a memory access to it. + */ + g_assert_not_reached(); +} + +static const MemoryRegionOps tz_ppc_dummy_ops = { + .valid.accepts = tz_ppc_dummy_accepts, +}; + static void tz_ppc_reset(DeviceState *dev) { TZPPC *s = TZ_PPC(dev); @@ -210,16 +225,33 @@ static void tz_ppc_realize(DeviceState *dev, Error **errp) SysBusDevice *sbd = SYS_BUS_DEVICE(dev); TZPPC *s = TZ_PPC(dev); int i; + int max_port = 0; /* We can't create the upstream end of the port until realize, * as we don't know the size of the MR used as the downstream until then. */ for (i = 0; i < TZ_NUM_PORTS; i++) { + if (s->port[i].downstream) { + max_port = i; + } + } + + for (i = 0; i <= max_port; i++) { TZPPCPort *port = &s->port[i]; char *name; uint64_t size; if (!port->downstream) { + /* + * Create dummy sysbus MMIO region so the sysbus region + * numbering doesn't get out of sync with the port numbers. + * The size is entirely arbitrary. + */ + name = g_strdup_printf("tz-ppc-dummy-port[%d]", i); + memory_region_init_io(&port->upstream, obj, &tz_ppc_dummy_ops, + port, name, 0x10000); + sysbus_init_mmio(sbd, &port->upstream); + g_free(name); continue; } diff --git a/include/hw/misc/tz-ppc.h b/include/hw/misc/tz-ppc.h index fc8b806e4d..080d6e2ec1 100644 --- a/include/hw/misc/tz-ppc.h +++ b/include/hw/misc/tz-ppc.h @@ -38,7 +38,13 @@ * * QEMU interface: * + sysbus MMIO regions 0..15: MemoryRegions defining the upstream end - * of each of the 16 ports of the PPC + * of each of the 16 ports of the PPC. When a port is unused (i.e. no + * downstream MemoryRegion is connected to it) at the end of the 0..15 + * range then no sysbus MMIO region is created for its upstream. When an + * unused port lies in the middle of the range with other used ports at + * higher port numbers, a dummy MMIO region is created to ensure that + * port N's upstream is always sysbus MMIO region N. Dummy regions should + * not be mapped, and will assert if any access is made to them. * + Property "port[0..15]": MemoryRegion defining the downstream device(s) * for each of the 16 ports of the PPC * + Named GPIO inputs "cfg_nonsec[0..15]": set to 1 if the port should be From b0de99f3e9bc7a0cacbff0ff8517379d915dd5f6 Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Thu, 21 Feb 2019 18:17:46 +0000 Subject: [PATCH 09/21] hw/timer/pl031: Allow use as an embedded-struct device MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Create a new include file for the pl031's device struct, type macros, etc, so that it can be instantiated using the "embedded struct" coding style. Signed-off-by: Peter Maydell Reviewed-by: Philippe Mathieu-Daudé Reviewed-by: Richard Henderson --- MAINTAINERS | 1 + hw/timer/pl031.c | 25 +---------------------- include/hw/timer/pl031.h | 44 ++++++++++++++++++++++++++++++++++++++++ 3 files changed, 46 insertions(+), 24 deletions(-) create mode 100644 include/hw/timer/pl031.h diff --git a/MAINTAINERS b/MAINTAINERS index b1d786cfd8..db31bc4486 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -491,6 +491,7 @@ F: hw/sd/pl181.c F: hw/ssi/pl022.c F: include/hw/ssi/pl022.h F: hw/timer/pl031.c +F: include/hw/timer/pl031.h F: include/hw/arm/primecell.h F: hw/timer/cmsdk-apb-timer.c F: include/hw/timer/cmsdk-apb-timer.h diff --git a/hw/timer/pl031.c b/hw/timer/pl031.c index d3aacce80d..f774dcd522 100644 --- a/hw/timer/pl031.c +++ b/hw/timer/pl031.c @@ -12,6 +12,7 @@ */ #include "qemu/osdep.h" +#include "hw/timer/pl031.h" #include "hw/sysbus.h" #include "qemu/timer.h" #include "sysemu/sysemu.h" @@ -36,30 +37,6 @@ do { printf("pl031: " fmt , ## __VA_ARGS__); } while (0) #define RTC_MIS 0x18 /* Masked interrupt status register */ #define RTC_ICR 0x1c /* Interrupt clear register */ -#define TYPE_PL031 "pl031" -#define PL031(obj) OBJECT_CHECK(PL031State, (obj), TYPE_PL031) - -typedef struct PL031State { - SysBusDevice parent_obj; - - MemoryRegion iomem; - QEMUTimer *timer; - qemu_irq irq; - - /* Needed to preserve the tick_count across migration, even if the - * absolute value of the rtc_clock is different on the source and - * destination. - */ - uint32_t tick_offset_vmstate; - uint32_t tick_offset; - - uint32_t mr; - uint32_t lr; - uint32_t cr; - uint32_t im; - uint32_t is; -} PL031State; - static const unsigned char pl031_id[] = { 0x31, 0x10, 0x14, 0x00, /* Device ID */ 0x0d, 0xf0, 0x05, 0xb1 /* Cell ID */ diff --git a/include/hw/timer/pl031.h b/include/hw/timer/pl031.h new file mode 100644 index 0000000000..99416d8ba5 --- /dev/null +++ b/include/hw/timer/pl031.h @@ -0,0 +1,44 @@ +/* + * ARM AMBA PrimeCell PL031 RTC + * + * Copyright (c) 2007 CodeSourcery + * + * This file is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + * Contributions after 2012-01-13 are licensed under the terms of the + * GNU GPL, version 2 or (at your option) any later version. + */ + +#ifndef HW_TIMER_PL031 +#define HW_TIMER_PL031 + +#include "hw/sysbus.h" + +#define TYPE_PL031 "pl031" +#define PL031(obj) OBJECT_CHECK(PL031State, (obj), TYPE_PL031) + +typedef struct PL031State { + SysBusDevice parent_obj; + + MemoryRegion iomem; + QEMUTimer *timer; + qemu_irq irq; + + /* + * Needed to preserve the tick_count across migration, even if the + * absolute value of the rtc_clock is different on the source and + * destination. + */ + uint32_t tick_offset_vmstate; + uint32_t tick_offset; + + uint32_t mr; + uint32_t lr; + uint32_t cr; + uint32_t im; + uint32_t is; +} PL031State; + +#endif From dd849ef2c9d57a329c6001c58dbdf49de712349c Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Thu, 21 Feb 2019 18:17:46 +0000 Subject: [PATCH 10/21] hw/timer/pl031: Convert to using trace events MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Convert the debug printing in the PL031 device to use trace events, and augment it to cover the interesting parts of device operation. Signed-off-by: Peter Maydell Reviewed-by: Philippe Mathieu-Daudé Reviewed-by: Richard Henderson --- hw/timer/pl031.c | 55 +++++++++++++++++++++++-------------------- hw/timer/trace-events | 6 +++++ 2 files changed, 36 insertions(+), 25 deletions(-) diff --git a/hw/timer/pl031.c b/hw/timer/pl031.c index f774dcd522..274ad47a33 100644 --- a/hw/timer/pl031.c +++ b/hw/timer/pl031.c @@ -18,15 +18,7 @@ #include "sysemu/sysemu.h" #include "qemu/cutils.h" #include "qemu/log.h" - -//#define DEBUG_PL031 - -#ifdef DEBUG_PL031 -#define DPRINTF(fmt, ...) \ -do { printf("pl031: " fmt , ## __VA_ARGS__); } while (0) -#else -#define DPRINTF(fmt, ...) do {} while(0) -#endif +#include "trace.h" #define RTC_DR 0x00 /* Data read register */ #define RTC_MR 0x04 /* Match register */ @@ -44,7 +36,10 @@ static const unsigned char pl031_id[] = { static void pl031_update(PL031State *s) { - qemu_set_irq(s->irq, s->is & s->im); + uint32_t flags = s->is & s->im; + + trace_pl031_irq_state(flags); + qemu_set_irq(s->irq, flags); } static void pl031_interrupt(void * opaque) @@ -52,7 +47,7 @@ static void pl031_interrupt(void * opaque) PL031State *s = (PL031State *)opaque; s->is = 1; - DPRINTF("Alarm raised\n"); + trace_pl031_alarm_raised(); pl031_update(s); } @@ -69,7 +64,7 @@ static void pl031_set_alarm(PL031State *s) /* The timer wraps around. This subtraction also wraps in the same way, and gives correct results when alarm < now_ticks. */ ticks = s->mr - pl031_get_count(s); - DPRINTF("Alarm set in %ud ticks\n", ticks); + trace_pl031_set_alarm(ticks); if (ticks == 0) { timer_del(s->timer); pl031_interrupt(s); @@ -83,38 +78,49 @@ static uint64_t pl031_read(void *opaque, hwaddr offset, unsigned size) { PL031State *s = (PL031State *)opaque; - - if (offset >= 0xfe0 && offset < 0x1000) - return pl031_id[(offset - 0xfe0) >> 2]; + uint64_t r; switch (offset) { case RTC_DR: - return pl031_get_count(s); + r = pl031_get_count(s); + break; case RTC_MR: - return s->mr; + r = s->mr; + break; case RTC_IMSC: - return s->im; + r = s->im; + break; case RTC_RIS: - return s->is; + r = s->is; + break; case RTC_LR: - return s->lr; + r = s->lr; + break; case RTC_CR: /* RTC is permanently enabled. */ - return 1; + r = 1; + break; case RTC_MIS: - return s->is & s->im; + r = s->is & s->im; + break; + case 0xfe0 ... 0xfff: + r = pl031_id[(offset - 0xfe0) >> 2]; + break; case RTC_ICR: qemu_log_mask(LOG_GUEST_ERROR, "pl031: read of write-only register at offset 0x%x\n", (int)offset); + r = 0; break; default: qemu_log_mask(LOG_GUEST_ERROR, "pl031_read: Bad offset 0x%x\n", (int)offset); + r = 0; break; } - return 0; + trace_pl031_read(offset, r); + return r; } static void pl031_write(void * opaque, hwaddr offset, @@ -122,6 +128,7 @@ static void pl031_write(void * opaque, hwaddr offset, { PL031State *s = (PL031State *)opaque; + trace_pl031_write(offset, value); switch (offset) { case RTC_LR: @@ -134,7 +141,6 @@ static void pl031_write(void * opaque, hwaddr offset, break; case RTC_IMSC: s->im = value & 1; - DPRINTF("Interrupt mask %d\n", s->im); pl031_update(s); break; case RTC_ICR: @@ -142,7 +148,6 @@ static void pl031_write(void * opaque, hwaddr offset, cleared when bit 0 of the written value is set. However the arm926e documentation (DDI0287B) states that the interrupt is cleared when any value is written. */ - DPRINTF("Interrupt cleared"); s->is = 0; pl031_update(s); break; diff --git a/hw/timer/trace-events b/hw/timer/trace-events index 0144a68951..12eb505fee 100644 --- a/hw/timer/trace-events +++ b/hw/timer/trace-events @@ -77,3 +77,9 @@ xlnx_zynqmp_rtc_gettime(int year, int month, int day, int hour, int min, int sec nrf51_timer_read(uint64_t addr, uint32_t value, unsigned size) "read addr 0x%" PRIx64 " data 0x%" PRIx32 " size %u" nrf51_timer_write(uint64_t addr, uint32_t value, unsigned size) "write addr 0x%" PRIx64 " data 0x%" PRIx32 " size %u" +# hw/timer/pl031.c +pl031_irq_state(int level) "irq state %d" +pl031_read(uint32_t addr, uint32_t value) "addr 0x%08x value 0x%08x" +pl031_write(uint32_t addr, uint32_t value) "addr 0x%08x value 0x%08x" +pl031_alarm_raised(void) "alarm raised" +pl031_set_alarm(uint32_t ticks) "alarm set for %u ticks" From 694cf209990f7b7a6f6f30b21010bab92c985175 Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Thu, 21 Feb 2019 18:17:46 +0000 Subject: [PATCH 11/21] hw/char/pl011: Allow use as an embedded-struct device MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Create a new include file for the pl011's device struct, type macros, etc, so that it can be instantiated using the "embedded struct" coding style. Signed-off-by: Peter Maydell Reviewed-by: Philippe Mathieu-Daudé Reviewed-by: Richard Henderson --- hw/char/pl011.c | 31 ++----------------------------- include/hw/char/pl011.h | 34 ++++++++++++++++++++++++++++++++++ 2 files changed, 36 insertions(+), 29 deletions(-) diff --git a/hw/char/pl011.c b/hw/char/pl011.c index 2aa277fc4f..0c4711e402 100644 --- a/hw/char/pl011.c +++ b/hw/char/pl011.c @@ -8,39 +8,12 @@ */ #include "qemu/osdep.h" +#include "hw/char/pl011.h" #include "hw/sysbus.h" #include "chardev/char-fe.h" #include "qemu/log.h" #include "trace.h" -#define TYPE_PL011 "pl011" -#define PL011(obj) OBJECT_CHECK(PL011State, (obj), TYPE_PL011) - -typedef struct PL011State { - SysBusDevice parent_obj; - - MemoryRegion iomem; - uint32_t readbuff; - uint32_t flags; - uint32_t lcr; - uint32_t rsr; - uint32_t cr; - uint32_t dmacr; - uint32_t int_enabled; - uint32_t int_level; - uint32_t read_fifo[16]; - uint32_t ilpr; - uint32_t ibrd; - uint32_t fbrd; - uint32_t ifl; - int read_pos; - int read_count; - int read_trigger; - CharBackend chr; - qemu_irq irq; - const unsigned char *id; -} PL011State; - #define PL011_INT_TX 0x20 #define PL011_INT_RX 0x10 @@ -357,7 +330,7 @@ static void pl011_luminary_init(Object *obj) } static const TypeInfo pl011_luminary_info = { - .name = "pl011_luminary", + .name = TYPE_PL011_LUMINARY, .parent = TYPE_PL011, .instance_init = pl011_luminary_init, }; diff --git a/include/hw/char/pl011.h b/include/hw/char/pl011.h index 83649324b6..1b52bfd5c9 100644 --- a/include/hw/char/pl011.h +++ b/include/hw/char/pl011.h @@ -15,6 +15,40 @@ #ifndef HW_PL011_H #define HW_PL011_H +#include "hw/sysbus.h" +#include "chardev/char-fe.h" + +#define TYPE_PL011 "pl011" +#define PL011(obj) OBJECT_CHECK(PL011State, (obj), TYPE_PL011) + +/* This shares the same struct (and cast macro) as the base pl011 device */ +#define TYPE_PL011_LUMINARY "pl011_luminary" + +typedef struct PL011State { + SysBusDevice parent_obj; + + MemoryRegion iomem; + uint32_t readbuff; + uint32_t flags; + uint32_t lcr; + uint32_t rsr; + uint32_t cr; + uint32_t dmacr; + uint32_t int_enabled; + uint32_t int_level; + uint32_t read_fifo[16]; + uint32_t ilpr; + uint32_t ibrd; + uint32_t fbrd; + uint32_t ifl; + int read_pos; + int read_count; + int read_trigger; + CharBackend chr; + qemu_irq irq; + const unsigned char *id; +} PL011State; + static inline DeviceState *pl011_create(hwaddr addr, qemu_irq irq, Chardev *chr) From a3c1ca56c0a6ec368a7876f2331b037d066b0b27 Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Thu, 21 Feb 2019 18:17:46 +0000 Subject: [PATCH 12/21] hw/char/pl011: Support all interrupt lines The PL011 UART has six interrupt lines: * RX (receive data) * TX (transmit data) * RT (receive timeout) * MS (modem status) * E (errors) * combined (logical OR of all the above) So far we have only emulated the combined interrupt line; add support for the others, so that boards that wire them up to different interrupt controller inputs can do so. Signed-off-by: Peter Maydell Reviewed-by: Richard Henderson --- hw/char/pl011.c | 46 +++++++++++++++++++++++++++++++++++++++-- include/hw/char/pl011.h | 2 +- 2 files changed, 45 insertions(+), 3 deletions(-) diff --git a/hw/char/pl011.c b/hw/char/pl011.c index 0c4711e402..29f4e5eb22 100644 --- a/hw/char/pl011.c +++ b/hw/char/pl011.c @@ -7,6 +7,17 @@ * This code is licensed under the GPL. */ +/* + * QEMU interface: + * + sysbus MMIO region 0: device registers + * + sysbus IRQ 0: UARTINTR (combined interrupt line) + * + sysbus IRQ 1: UARTRXINTR (receive FIFO interrupt line) + * + sysbus IRQ 2: UARTTXINTR (transmit FIFO interrupt line) + * + sysbus IRQ 3: UARTRTINTR (receive timeout interrupt line) + * + sysbus IRQ 4: UARTMSINTR (momem status interrupt line) + * + sysbus IRQ 5: UARTEINTR (error interrupt line) + */ + #include "qemu/osdep.h" #include "hw/char/pl011.h" #include "hw/sysbus.h" @@ -22,18 +33,46 @@ #define PL011_FLAG_TXFF 0x20 #define PL011_FLAG_RXFE 0x10 +/* Interrupt status bits in UARTRIS, UARTMIS, UARTIMSC */ +#define INT_OE (1 << 10) +#define INT_BE (1 << 9) +#define INT_PE (1 << 8) +#define INT_FE (1 << 7) +#define INT_RT (1 << 6) +#define INT_TX (1 << 5) +#define INT_RX (1 << 4) +#define INT_DSR (1 << 3) +#define INT_DCD (1 << 2) +#define INT_CTS (1 << 1) +#define INT_RI (1 << 0) +#define INT_E (INT_OE | INT_BE | INT_PE | INT_FE) +#define INT_MS (INT_RI | INT_DSR | INT_DCD | INT_CTS) + static const unsigned char pl011_id_arm[8] = { 0x11, 0x10, 0x14, 0x00, 0x0d, 0xf0, 0x05, 0xb1 }; static const unsigned char pl011_id_luminary[8] = { 0x11, 0x00, 0x18, 0x01, 0x0d, 0xf0, 0x05, 0xb1 }; +/* Which bits in the interrupt status matter for each outbound IRQ line ? */ +static const uint32_t irqmask[] = { + INT_E | INT_MS | INT_RT | INT_TX | INT_RX, /* combined IRQ */ + INT_RX, + INT_TX, + INT_RT, + INT_MS, + INT_E, +}; + static void pl011_update(PL011State *s) { uint32_t flags; + int i; flags = s->int_level & s->int_enabled; trace_pl011_irq_state(flags != 0); - qemu_set_irq(s->irq, flags != 0); + for (i = 0; i < ARRAY_SIZE(s->irq); i++) { + qemu_set_irq(s->irq[i], (flags & irqmask[i]) != 0); + } } static uint64_t pl011_read(void *opaque, hwaddr offset, @@ -284,10 +323,13 @@ static void pl011_init(Object *obj) { SysBusDevice *sbd = SYS_BUS_DEVICE(obj); PL011State *s = PL011(obj); + int i; memory_region_init_io(&s->iomem, OBJECT(s), &pl011_ops, s, "pl011", 0x1000); sysbus_init_mmio(sbd, &s->iomem); - sysbus_init_irq(sbd, &s->irq); + for (i = 0; i < ARRAY_SIZE(s->irq); i++) { + sysbus_init_irq(sbd, &s->irq[i]); + } s->read_trigger = 1; s->ifl = 0x12; diff --git a/include/hw/char/pl011.h b/include/hw/char/pl011.h index 1b52bfd5c9..dad3cf2912 100644 --- a/include/hw/char/pl011.h +++ b/include/hw/char/pl011.h @@ -45,7 +45,7 @@ typedef struct PL011State { int read_count; int read_trigger; CharBackend chr; - qemu_irq irq; + qemu_irq irq[6]; const unsigned char *id; } PL011State; From 76b09fafaf54051ccc0620169ae5b72c87f4f547 Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Thu, 21 Feb 2019 18:17:46 +0000 Subject: [PATCH 13/21] hw/char/pl011: Use '0x' prefix when logging hex numbers MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The pl011 logs when the guest makes a bad access. It prints the address offset in hex but confusingly omits the '0x' prefix; add it. Signed-off-by: Peter Maydell Reviewed-by: Philippe Mathieu-Daudé Reviewed-by: Richard Henderson --- hw/char/pl011.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/hw/char/pl011.c b/hw/char/pl011.c index 29f4e5eb22..e5dd448f85 100644 --- a/hw/char/pl011.c +++ b/hw/char/pl011.c @@ -143,7 +143,7 @@ static uint64_t pl011_read(void *opaque, hwaddr offset, break; default: qemu_log_mask(LOG_GUEST_ERROR, - "pl011_read: Bad offset %x\n", (int)offset); + "pl011_read: Bad offset 0x%x\n", (int)offset); r = 0; break; } @@ -232,7 +232,7 @@ static void pl011_write(void *opaque, hwaddr offset, break; default: qemu_log_mask(LOG_GUEST_ERROR, - "pl011_write: Bad offset %x\n", (int)offset); + "pl011_write: Bad offset 0x%x\n", (int)offset); } } From 74ecf7677b72084b25ace9de3191abe3afdaeff6 Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Thu, 21 Feb 2019 18:17:47 +0000 Subject: [PATCH 14/21] hw/arm/armsse: Document SRAM_ADDR_WIDTH property in header comment In commit 4b635cf7a95e501211 we added a QOM property to the ARMSSE object, but forgot to add it to the documentation comment in the header. Correct the omission. Fixes: 4b635cf7a95e501211 ("hw/arm/armsse: Make SRAM bank size configurable") Signed-off-by: Peter Maydell Reviewed-by: Richard Henderson --- include/hw/arm/armsse.h | 2 ++ 1 file changed, 2 insertions(+) diff --git a/include/hw/arm/armsse.h b/include/hw/arm/armsse.h index f800bafb14..444605b44d 100644 --- a/include/hw/arm/armsse.h +++ b/include/hw/arm/armsse.h @@ -46,6 +46,8 @@ * being the same for both, to avoid having to have separate Property * lists for different variants. This restriction can be relaxed later * if necessary.) + * + QOM property "SRAM_ADDR_WIDTH" sets the number of bits used for the + * address of each SRAM bank (and thus the total amount of internal SRAM) * + Named GPIO inputs "EXP_IRQ" 0..n are the expansion interrupts for CPU 0, * which are wired to its NVIC lines 32 .. n+32 * + Named GPIO inputs "EXP_CPU1_IRQ" 0..n are the expansion interrupts for From 321874196d0fff06eff1df43b20cbadcfe699aad Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Thu, 21 Feb 2019 18:17:47 +0000 Subject: [PATCH 15/21] hw/arm/armsse: Allow boards to specify init-svtor MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The Musca boards have DAPLink firmware that sets the initial secure VTOR value (the location of the vector table) differently depending on the boot mode (from flash, from RAM, etc). Export the init-svtor as a QOM property of the ARMSSE object so that the board can change it. Signed-off-by: Peter Maydell Reviewed-by: Philippe Mathieu-Daudé Reviewed-by: Richard Henderson --- hw/arm/armsse.c | 8 ++++---- include/hw/arm/armsse.h | 3 +++ 2 files changed, 7 insertions(+), 4 deletions(-) diff --git a/hw/arm/armsse.c b/hw/arm/armsse.c index d0207dbabc..50da41f64c 100644 --- a/hw/arm/armsse.c +++ b/hw/arm/armsse.c @@ -505,11 +505,10 @@ static void armsse_realize(DeviceState *dev, Error **errp) * the INITSVTOR* registers before powering up the CPUs in any case, * so the hardware's default value doesn't matter. QEMU doesn't emulate * the control processor, so instead we behave in the way that the - * firmware does. All boards currently known about have firmware that - * sets the INITSVTOR0 and INITSVTOR1 registers to 0x10000000, like the - * IoTKit default. We can make this more configurable if necessary. + * firmware does. The initial value is configurable by the board code + * to match whatever its firmware does. */ - qdev_prop_set_uint32(cpudev, "init-svtor", 0x10000000); + qdev_prop_set_uint32(cpudev, "init-svtor", s->init_svtor); /* * Start all CPUs except CPU0 powered down. In real hardware it is * a configurable property of the SSE-200 which CPUs start powered up @@ -1187,6 +1186,7 @@ static Property armsse_properties[] = { DEFINE_PROP_UINT32("EXP_NUMIRQ", ARMSSE, exp_numirq, 64), DEFINE_PROP_UINT32("MAINCLK", ARMSSE, mainclk_frq, 0), DEFINE_PROP_UINT32("SRAM_ADDR_WIDTH", ARMSSE, sram_addr_width, 15), + DEFINE_PROP_UINT32("init-svtor", ARMSSE, init_svtor, 0x10000000), DEFINE_PROP_END_OF_LIST() }; diff --git a/include/hw/arm/armsse.h b/include/hw/arm/armsse.h index 444605b44d..84879f40dd 100644 --- a/include/hw/arm/armsse.h +++ b/include/hw/arm/armsse.h @@ -48,6 +48,8 @@ * if necessary.) * + QOM property "SRAM_ADDR_WIDTH" sets the number of bits used for the * address of each SRAM bank (and thus the total amount of internal SRAM) + * + QOM property "init-svtor" sets the initial value of the CPU SVTOR register + * (where it expects to load the PC and SP from the vector table on reset) * + Named GPIO inputs "EXP_IRQ" 0..n are the expansion interrupts for CPU 0, * which are wired to its NVIC lines 32 .. n+32 * + Named GPIO inputs "EXP_CPU1_IRQ" 0..n are the expansion interrupts for @@ -204,6 +206,7 @@ typedef struct ARMSSE { uint32_t exp_numirq; uint32_t mainclk_frq; uint32_t sram_addr_width; + uint32_t init_svtor; } ARMSSE; typedef struct ARMSSEInfo ARMSSEInfo; From 8f69a4c15d977d7d66c46143f74fbba68efcc0f8 Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Thu, 21 Feb 2019 18:17:47 +0000 Subject: [PATCH 16/21] hw/arm/musca.c: Implement models of the Musca-A and -B1 boards The Musca-A and Musca-B1 development boards are based on the SSE-200 subsystem for embedded. Implement an initial skeleton model of these boards, which are similar but not identical. This commit creates the board model with the SSE and the IRQ splitters to wire IRQs up to its two CPUs. As yet there are no devices and no memory: these will be added later. Signed-off-by: Peter Maydell Reviewed-by: Richard Henderson --- MAINTAINERS | 6 + default-configs/arm-softmmu.mak | 1 + hw/arm/Makefile.objs | 1 + hw/arm/musca.c | 197 ++++++++++++++++++++++++++++++++ 4 files changed, 205 insertions(+) create mode 100644 hw/arm/musca.c diff --git a/MAINTAINERS b/MAINTAINERS index db31bc4486..6ae55ff732 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -634,6 +634,12 @@ F: include/hw/misc/iotkit-sysinfo.h F: hw/misc/armsse-cpuid.c F: include/hw/misc/armsse-cpuid.h +Musca +M: Peter Maydell +L: qemu-arm@nongnu.org +S: Maintained +F: hw/arm/musca.c + Musicpal M: Jan Kiszka M: Peter Maydell diff --git a/default-configs/arm-softmmu.mak b/default-configs/arm-softmmu.mak index 734ca721e9..87ad267494 100644 --- a/default-configs/arm-softmmu.mak +++ b/default-configs/arm-softmmu.mak @@ -89,6 +89,7 @@ CONFIG_TUSB6010=y CONFIG_IMX=y CONFIG_MAINSTONE=y CONFIG_MPS2=y +CONFIG_MUSCA=y CONFIG_NSERIES=y CONFIG_RASPI=y CONFIG_REALVIEW=y diff --git a/hw/arm/Makefile.objs b/hw/arm/Makefile.objs index fa40e8d641..fa57c7c770 100644 --- a/hw/arm/Makefile.objs +++ b/hw/arm/Makefile.objs @@ -35,6 +35,7 @@ obj-$(CONFIG_ASPEED_SOC) += aspeed_soc.o aspeed.o obj-$(CONFIG_MPS2) += mps2.o obj-$(CONFIG_MPS2) += mps2-tz.o obj-$(CONFIG_MSF2) += msf2-soc.o msf2-som.o +obj-$(CONFIG_MUSCA) += musca.o obj-$(CONFIG_ARMSSE) += armsse.o obj-$(CONFIG_FSL_IMX7) += fsl-imx7.o mcimx7d-sabre.o obj-$(CONFIG_ARM_SMMUV3) += smmu-common.o smmuv3.o diff --git a/hw/arm/musca.c b/hw/arm/musca.c new file mode 100644 index 0000000000..cc624c7d16 --- /dev/null +++ b/hw/arm/musca.c @@ -0,0 +1,197 @@ +/* + * Arm Musca-B1 test chip board emulation + * + * Copyright (c) 2019 Linaro Limited + * Written by Peter Maydell + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 or + * (at your option) any later version. + */ + +/* + * The Musca boards are a reference implementation of a system using + * the SSE-200 subsystem for embedded: + * https://developer.arm.com/products/system-design/development-boards/iot-test-chips-and-boards/musca-a-test-chip-board + * https://developer.arm.com/products/system-design/development-boards/iot-test-chips-and-boards/musca-b-test-chip-board + * We model the A and B1 variants of this board, as described in the TRMs: + * http://infocenter.arm.com/help/topic/com.arm.doc.101107_0000_00_en/index.html + * http://infocenter.arm.com/help/topic/com.arm.doc.101312_0000_00_en/index.html + */ + +#include "qemu/osdep.h" +#include "qemu/error-report.h" +#include "qapi/error.h" +#include "exec/address-spaces.h" +#include "hw/arm/arm.h" +#include "hw/arm/armsse.h" +#include "hw/boards.h" +#include "hw/core/split-irq.h" + +#define MUSCA_NUMIRQ_MAX 96 + +typedef enum MuscaType { + MUSCA_A, + MUSCA_B1, +} MuscaType; + +typedef struct { + MachineClass parent; + MuscaType type; + uint32_t init_svtor; + int sram_addr_width; + int num_irqs; +} MuscaMachineClass; + +typedef struct { + MachineState parent; + + ARMSSE sse; + SplitIRQ cpu_irq_splitter[MUSCA_NUMIRQ_MAX]; +} MuscaMachineState; + +#define TYPE_MUSCA_MACHINE "musca" +#define TYPE_MUSCA_A_MACHINE MACHINE_TYPE_NAME("musca-a") +#define TYPE_MUSCA_B1_MACHINE MACHINE_TYPE_NAME("musca-b1") + +#define MUSCA_MACHINE(obj) \ + OBJECT_CHECK(MuscaMachineState, obj, TYPE_MUSCA_MACHINE) +#define MUSCA_MACHINE_GET_CLASS(obj) \ + OBJECT_GET_CLASS(MuscaMachineClass, obj, TYPE_MUSCA_MACHINE) +#define MUSCA_MACHINE_CLASS(klass) \ + OBJECT_CLASS_CHECK(MuscaMachineClass, klass, TYPE_MUSCA_MACHINE) + +/* + * Main SYSCLK frequency in Hz + * TODO this should really be different for the two cores, but we + * don't model that in our SSE-200 model yet. + */ +#define SYSCLK_FRQ 40000000 + +static void musca_init(MachineState *machine) +{ + MuscaMachineState *mms = MUSCA_MACHINE(machine); + MuscaMachineClass *mmc = MUSCA_MACHINE_GET_CLASS(mms); + MachineClass *mc = MACHINE_GET_CLASS(machine); + MemoryRegion *system_memory = get_system_memory(); + DeviceState *ssedev; + int i; + + assert(mmc->num_irqs <= MUSCA_NUMIRQ_MAX); + + if (strcmp(machine->cpu_type, mc->default_cpu_type) != 0) { + error_report("This board can only be used with CPU %s", + mc->default_cpu_type); + exit(1); + } + + sysbus_init_child_obj(OBJECT(machine), "sse-200", &mms->sse, + sizeof(mms->sse), TYPE_SSE200); + ssedev = DEVICE(&mms->sse); + object_property_set_link(OBJECT(&mms->sse), OBJECT(system_memory), + "memory", &error_fatal); + qdev_prop_set_uint32(ssedev, "EXP_NUMIRQ", mmc->num_irqs); + qdev_prop_set_uint32(ssedev, "init-svtor", mmc->init_svtor); + qdev_prop_set_uint32(ssedev, "SRAM_ADDR_WIDTH", mmc->sram_addr_width); + qdev_prop_set_uint32(ssedev, "MAINCLK", SYSCLK_FRQ); + object_property_set_bool(OBJECT(&mms->sse), true, "realized", + &error_fatal); + + /* + * We need to create splitters to feed the IRQ inputs + * for each CPU in the SSE-200 from each device in the board. + */ + for (i = 0; i < mmc->num_irqs; i++) { + char *name = g_strdup_printf("musca-irq-splitter%d", i); + SplitIRQ *splitter = &mms->cpu_irq_splitter[i]; + + object_initialize_child(OBJECT(machine), name, + splitter, sizeof(*splitter), + TYPE_SPLIT_IRQ, &error_fatal, NULL); + g_free(name); + + object_property_set_int(OBJECT(splitter), 2, "num-lines", + &error_fatal); + object_property_set_bool(OBJECT(splitter), true, "realized", + &error_fatal); + qdev_connect_gpio_out(DEVICE(splitter), 0, + qdev_get_gpio_in_named(ssedev, "EXP_IRQ", i)); + qdev_connect_gpio_out(DEVICE(splitter), 1, + qdev_get_gpio_in_named(ssedev, + "EXP_CPU1_IRQ", i)); + } + + armv7m_load_kernel(ARM_CPU(first_cpu), machine->kernel_filename, 0x2000000); +} + +static void musca_class_init(ObjectClass *oc, void *data) +{ + MachineClass *mc = MACHINE_CLASS(oc); + + mc->default_cpus = 2; + mc->min_cpus = mc->default_cpus; + mc->max_cpus = mc->default_cpus; + mc->default_cpu_type = ARM_CPU_TYPE_NAME("cortex-m33"); + mc->init = musca_init; +} + +static void musca_a_class_init(ObjectClass *oc, void *data) +{ + MachineClass *mc = MACHINE_CLASS(oc); + MuscaMachineClass *mmc = MUSCA_MACHINE_CLASS(oc); + + mc->desc = "ARM Musca-A board (dual Cortex-M33)"; + mmc->type = MUSCA_A; + mmc->init_svtor = 0x10200000; + mmc->sram_addr_width = 15; + mmc->num_irqs = 64; +} + +static void musca_b1_class_init(ObjectClass *oc, void *data) +{ + MachineClass *mc = MACHINE_CLASS(oc); + MuscaMachineClass *mmc = MUSCA_MACHINE_CLASS(oc); + + mc->desc = "ARM Musca-B1 board (dual Cortex-M33)"; + mmc->type = MUSCA_B1; + /* + * This matches the DAPlink firmware which boots from QSPI. There + * is also a firmware blob which boots from the eFlash, which + * uses init_svtor = 0x1A000000. QEMU doesn't currently support that, + * though we could in theory expose a machine property on the command + * line to allow the user to request eFlash boot. + */ + mmc->init_svtor = 0x10000000; + mmc->sram_addr_width = 17; + mmc->num_irqs = 96; +} + +static const TypeInfo musca_info = { + .name = TYPE_MUSCA_MACHINE, + .parent = TYPE_MACHINE, + .abstract = true, + .instance_size = sizeof(MuscaMachineState), + .class_size = sizeof(MuscaMachineClass), + .class_init = musca_class_init, +}; + +static const TypeInfo musca_a_info = { + .name = TYPE_MUSCA_A_MACHINE, + .parent = TYPE_MUSCA_MACHINE, + .class_init = musca_a_class_init, +}; + +static const TypeInfo musca_b1_info = { + .name = TYPE_MUSCA_B1_MACHINE, + .parent = TYPE_MUSCA_MACHINE, + .class_init = musca_b1_class_init, +}; + +static void musca_machine_init(void) +{ + type_register_static(&musca_info); + type_register_static(&musca_a_info); + type_register_static(&musca_b1_info); +} + +type_init(musca_machine_init); From ae3bc714014deecc9e6d21750a1ed72cf551e36f Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Thu, 21 Feb 2019 18:17:47 +0000 Subject: [PATCH 17/21] hw/arm/musca: Add PPCs Many of the devices on the Musca board live behind TrustZone Peripheral Protection Controllers (PPCs); add models of the PPCs, using a similar scheme to the MPS2 board models. This commit wires up the PPCs with "unimplemented device" stubs behind them in the correct places in the address map. Signed-off-by: Peter Maydell Reviewed-by: Richard Henderson --- hw/arm/musca.c | 289 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 289 insertions(+) diff --git a/hw/arm/musca.c b/hw/arm/musca.c index cc624c7d16..8774e0b87b 100644 --- a/hw/arm/musca.c +++ b/hw/arm/musca.c @@ -27,8 +27,11 @@ #include "hw/arm/armsse.h" #include "hw/boards.h" #include "hw/core/split-irq.h" +#include "hw/misc/tz-ppc.h" +#include "hw/misc/unimp.h" #define MUSCA_NUMIRQ_MAX 96 +#define MUSCA_PPC_MAX 3 typedef enum MuscaType { MUSCA_A, @@ -48,6 +51,24 @@ typedef struct { ARMSSE sse; SplitIRQ cpu_irq_splitter[MUSCA_NUMIRQ_MAX]; + SplitIRQ sec_resp_splitter; + TZPPC ppc[MUSCA_PPC_MAX]; + MemoryRegion container; + UnimplementedDeviceState eflash[2]; + UnimplementedDeviceState qspi; + UnimplementedDeviceState mpc[5]; + UnimplementedDeviceState mhu[2]; + UnimplementedDeviceState pwm[3]; + UnimplementedDeviceState i2s; + UnimplementedDeviceState uart[2]; + UnimplementedDeviceState i2c[2]; + UnimplementedDeviceState spi; + UnimplementedDeviceState scc; + UnimplementedDeviceState timer; + UnimplementedDeviceState rtc; + UnimplementedDeviceState pvt; + UnimplementedDeviceState sdio; + UnimplementedDeviceState gpio; } MuscaMachineState; #define TYPE_MUSCA_MACHINE "musca" @@ -68,6 +89,94 @@ typedef struct { */ #define SYSCLK_FRQ 40000000 +/* + * Most of the devices in the Musca board sit behind Peripheral Protection + * Controllers. These data structures define the layout of which devices + * sit behind which PPCs. + * The devfn for each port is a function which creates, configures + * and initializes the device, returning the MemoryRegion which + * needs to be plugged into the downstream end of the PPC port. + */ +typedef MemoryRegion *MakeDevFn(MuscaMachineState *mms, void *opaque, + const char *name, hwaddr size); + +typedef struct PPCPortInfo { + const char *name; + MakeDevFn *devfn; + void *opaque; + hwaddr addr; + hwaddr size; +} PPCPortInfo; + +typedef struct PPCInfo { + const char *name; + PPCPortInfo ports[TZ_NUM_PORTS]; +} PPCInfo; + +static MemoryRegion *make_unimp_dev(MuscaMachineState *mms, + void *opaque, const char *name, hwaddr size) +{ + /* + * Initialize, configure and realize a TYPE_UNIMPLEMENTED_DEVICE, + * and return a pointer to its MemoryRegion. + */ + UnimplementedDeviceState *uds = opaque; + + sysbus_init_child_obj(OBJECT(mms), name, uds, + sizeof(UnimplementedDeviceState), + TYPE_UNIMPLEMENTED_DEVICE); + qdev_prop_set_string(DEVICE(uds), "name", name); + qdev_prop_set_uint64(DEVICE(uds), "size", size); + object_property_set_bool(OBJECT(uds), true, "realized", &error_fatal); + return sysbus_mmio_get_region(SYS_BUS_DEVICE(uds), 0); +} + +static MemoryRegion *make_musca_a_devs(MuscaMachineState *mms, void *opaque, + const char *name, hwaddr size) +{ + /* + * Create the container MemoryRegion for all the devices that live + * behind the Musca-A PPC's single port. These devices don't have a PPC + * port each, but we use the PPCPortInfo struct as a convenient way + * to describe them. Note that addresses here are relative to the base + * address of the PPC port region: 0x40100000, and devices appear both + * at the 0x4... NS region and the 0x5... S region. + */ + int i; + MemoryRegion *container = &mms->container; + + const PPCPortInfo devices[] = { + { "uart0", make_unimp_dev, &mms->uart[0], 0x1000, 0x1000 }, + { "uart1", make_unimp_dev, &mms->uart[1], 0x2000, 0x1000 }, + { "spi", make_unimp_dev, &mms->spi, 0x3000, 0x1000 }, + { "i2c0", make_unimp_dev, &mms->i2c[0], 0x4000, 0x1000 }, + { "i2c1", make_unimp_dev, &mms->i2c[1], 0x5000, 0x1000 }, + { "i2s", make_unimp_dev, &mms->i2s, 0x6000, 0x1000 }, + { "pwm0", make_unimp_dev, &mms->pwm[0], 0x7000, 0x1000 }, + { "rtc", make_unimp_dev, &mms->rtc, 0x8000, 0x1000 }, + { "qspi", make_unimp_dev, &mms->qspi, 0xa000, 0x1000 }, + { "timer", make_unimp_dev, &mms->timer, 0xb000, 0x1000 }, + { "scc", make_unimp_dev, &mms->scc, 0xc000, 0x1000 }, + { "pwm1", make_unimp_dev, &mms->pwm[1], 0xe000, 0x1000 }, + { "pwm2", make_unimp_dev, &mms->pwm[2], 0xf000, 0x1000 }, + { "gpio", make_unimp_dev, &mms->gpio, 0x10000, 0x1000 }, + { "mpc0", make_unimp_dev, &mms->mpc[0], 0x12000, 0x1000 }, + { "mpc1", make_unimp_dev, &mms->mpc[1], 0x13000, 0x1000 }, + }; + + memory_region_init(container, OBJECT(mms), "musca-device-container", size); + + for (i = 0; i < ARRAY_SIZE(devices); i++) { + const PPCPortInfo *pinfo = &devices[i]; + MemoryRegion *mr; + + mr = pinfo->devfn(mms, pinfo->opaque, pinfo->name, pinfo->size); + memory_region_add_subregion(container, pinfo->addr, mr); + } + + return &mms->container; +} + static void musca_init(MachineState *machine) { MuscaMachineState *mms = MUSCA_MACHINE(machine); @@ -75,6 +184,9 @@ static void musca_init(MachineState *machine) MachineClass *mc = MACHINE_GET_CLASS(machine); MemoryRegion *system_memory = get_system_memory(); DeviceState *ssedev; + DeviceState *dev_splitter; + const PPCInfo *ppcs; + int num_ppcs; int i; assert(mmc->num_irqs <= MUSCA_NUMIRQ_MAX); @@ -121,6 +233,183 @@ static void musca_init(MachineState *machine) "EXP_CPU1_IRQ", i)); } + /* + * The sec_resp_cfg output from the SSE-200 must be split into multiple + * lines, one for each of the PPCs we create here. + */ + object_initialize(&mms->sec_resp_splitter, sizeof(mms->sec_resp_splitter), + TYPE_SPLIT_IRQ); + object_property_add_child(OBJECT(machine), "sec-resp-splitter", + OBJECT(&mms->sec_resp_splitter), &error_fatal); + object_property_set_int(OBJECT(&mms->sec_resp_splitter), + ARRAY_SIZE(mms->ppc), "num-lines", &error_fatal); + object_property_set_bool(OBJECT(&mms->sec_resp_splitter), true, + "realized", &error_fatal); + dev_splitter = DEVICE(&mms->sec_resp_splitter); + qdev_connect_gpio_out_named(ssedev, "sec_resp_cfg", 0, + qdev_get_gpio_in(dev_splitter, 0)); + + /* + * Most of the devices in the board are behind Peripheral Protection + * Controllers. The required order for initializing things is: + * + initialize the PPC + * + initialize, configure and realize downstream devices + * + connect downstream device MemoryRegions to the PPC + * + realize the PPC + * + map the PPC's MemoryRegions to the places in the address map + * where the downstream devices should appear + * + wire up the PPC's control lines to the SSE object + * + * The PPC mapping differs for the -A and -B1 variants; the -A version + * is much simpler, using only a single port of a single PPC and putting + * all the devices behind that. + */ + const PPCInfo a_ppcs[] = { { + .name = "ahb_ppcexp0", + .ports = { + { "musca-devices", make_musca_a_devs, 0, 0x40100000, 0x100000 }, + }, + }, + }; + + /* + * Devices listed with an 0x4.. address appear in both the NS 0x4.. region + * and the 0x5.. S region. Devices listed with an 0x5.. address appear + * only in the S region. + */ + const PPCInfo b1_ppcs[] = { { + .name = "apb_ppcexp0", + .ports = { + { "eflash0", make_unimp_dev, &mms->eflash[0], + 0x52400000, 0x1000 }, + { "eflash1", make_unimp_dev, &mms->eflash[1], + 0x52500000, 0x1000 }, + { "qspi", make_unimp_dev, &mms->qspi, 0x42800000, 0x100000 }, + { "mpc0", make_unimp_dev, &mms->mpc[0], 0x52000000, 0x1000 }, + { "mpc1", make_unimp_dev, &mms->mpc[1], 0x52100000, 0x1000 }, + { "mpc2", make_unimp_dev, &mms->mpc[2], 0x52200000, 0x1000 }, + { "mpc3", make_unimp_dev, &mms->mpc[3], 0x52300000, 0x1000 }, + { "mhu0", make_unimp_dev, &mms->mhu[0], 0x42600000, 0x100000 }, + { "mhu1", make_unimp_dev, &mms->mhu[1], 0x42700000, 0x100000 }, + { }, /* port 9: unused */ + { }, /* port 10: unused */ + { }, /* port 11: unused */ + { }, /* port 12: unused */ + { }, /* port 13: unused */ + { "mpc4", make_unimp_dev, &mms->mpc[4], 0x52e00000, 0x1000 }, + }, + }, { + .name = "apb_ppcexp1", + .ports = { + { "pwm0", make_unimp_dev, &mms->pwm[0], 0x40101000, 0x1000 }, + { "pwm1", make_unimp_dev, &mms->pwm[1], 0x40102000, 0x1000 }, + { "pwm2", make_unimp_dev, &mms->pwm[2], 0x40103000, 0x1000 }, + { "i2s", make_unimp_dev, &mms->i2s, 0x40104000, 0x1000 }, + { "uart0", make_unimp_dev, &mms->uart[0], 0x40105000, 0x1000 }, + { "uart1", make_unimp_dev, &mms->uart[1], 0x40106000, 0x1000 }, + { "i2c0", make_unimp_dev, &mms->i2c[0], 0x40108000, 0x1000 }, + { "i2c1", make_unimp_dev, &mms->i2c[1], 0x40109000, 0x1000 }, + { "spi", make_unimp_dev, &mms->spi, 0x4010a000, 0x1000 }, + { "scc", make_unimp_dev, &mms->scc, 0x5010b000, 0x1000 }, + { "timer", make_unimp_dev, &mms->timer, 0x4010c000, 0x1000 }, + { "rtc", make_unimp_dev, &mms->rtc, 0x4010d000, 0x1000 }, + { "pvt", make_unimp_dev, &mms->pvt, 0x4010e000, 0x1000 }, + { "sdio", make_unimp_dev, &mms->sdio, 0x4010f000, 0x1000 }, + }, + }, { + .name = "ahb_ppcexp0", + .ports = { + { }, /* port 0: unused */ + { "gpio", make_unimp_dev, &mms->gpio, 0x41000000, 0x1000 }, + }, + }, + }; + + switch (mmc->type) { + case MUSCA_A: + ppcs = a_ppcs; + num_ppcs = ARRAY_SIZE(a_ppcs); + break; + case MUSCA_B1: + ppcs = b1_ppcs; + num_ppcs = ARRAY_SIZE(b1_ppcs); + break; + default: + g_assert_not_reached(); + } + assert(num_ppcs <= MUSCA_PPC_MAX); + + for (i = 0; i < num_ppcs; i++) { + const PPCInfo *ppcinfo = &ppcs[i]; + TZPPC *ppc = &mms->ppc[i]; + DeviceState *ppcdev; + int port; + char *gpioname; + + sysbus_init_child_obj(OBJECT(machine), ppcinfo->name, ppc, + sizeof(TZPPC), TYPE_TZ_PPC); + ppcdev = DEVICE(ppc); + + for (port = 0; port < TZ_NUM_PORTS; port++) { + const PPCPortInfo *pinfo = &ppcinfo->ports[port]; + MemoryRegion *mr; + char *portname; + + if (!pinfo->devfn) { + continue; + } + + mr = pinfo->devfn(mms, pinfo->opaque, pinfo->name, pinfo->size); + portname = g_strdup_printf("port[%d]", port); + object_property_set_link(OBJECT(ppc), OBJECT(mr), + portname, &error_fatal); + g_free(portname); + } + + object_property_set_bool(OBJECT(ppc), true, "realized", &error_fatal); + + for (port = 0; port < TZ_NUM_PORTS; port++) { + const PPCPortInfo *pinfo = &ppcinfo->ports[port]; + + if (!pinfo->devfn) { + continue; + } + sysbus_mmio_map(SYS_BUS_DEVICE(ppc), port, pinfo->addr); + + gpioname = g_strdup_printf("%s_nonsec", ppcinfo->name); + qdev_connect_gpio_out_named(ssedev, gpioname, port, + qdev_get_gpio_in_named(ppcdev, + "cfg_nonsec", + port)); + g_free(gpioname); + gpioname = g_strdup_printf("%s_ap", ppcinfo->name); + qdev_connect_gpio_out_named(ssedev, gpioname, port, + qdev_get_gpio_in_named(ppcdev, + "cfg_ap", port)); + g_free(gpioname); + } + + gpioname = g_strdup_printf("%s_irq_enable", ppcinfo->name); + qdev_connect_gpio_out_named(ssedev, gpioname, 0, + qdev_get_gpio_in_named(ppcdev, + "irq_enable", 0)); + g_free(gpioname); + gpioname = g_strdup_printf("%s_irq_clear", ppcinfo->name); + qdev_connect_gpio_out_named(ssedev, gpioname, 0, + qdev_get_gpio_in_named(ppcdev, + "irq_clear", 0)); + g_free(gpioname); + gpioname = g_strdup_printf("%s_irq_status", ppcinfo->name); + qdev_connect_gpio_out_named(ppcdev, "irq", 0, + qdev_get_gpio_in_named(ssedev, + gpioname, 0)); + g_free(gpioname); + + qdev_connect_gpio_out(dev_splitter, i, + qdev_get_gpio_in_named(ppcdev, + "cfg_sec_resp", 0)); + } + armv7m_load_kernel(ARM_CPU(first_cpu), machine->kernel_filename, 0x2000000); } From 33293e504926b73c8deb304619ef84fbd7c980bd Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Thu, 21 Feb 2019 18:17:47 +0000 Subject: [PATCH 18/21] hw/arm/musca: Add MPCs The Musca board puts its SRAM and flash behind TrustZone Memory Protection Controllers (MPCs). Each MPC sits between the CPU and the RAM/flash, and also has a set of memory mapped control registers. Wire up the MPCs, and the memory behind them. For the moment we implement the flash as simple ROM, which cannot be reprogrammed by the guest. Signed-off-by: Peter Maydell Reviewed-by: Richard Henderson --- hw/arm/musca.c | 155 ++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 147 insertions(+), 8 deletions(-) diff --git a/hw/arm/musca.c b/hw/arm/musca.c index 8774e0b87b..c4095c4579 100644 --- a/hw/arm/musca.c +++ b/hw/arm/musca.c @@ -27,11 +27,15 @@ #include "hw/arm/armsse.h" #include "hw/boards.h" #include "hw/core/split-irq.h" +#include "hw/misc/tz-mpc.h" #include "hw/misc/tz-ppc.h" #include "hw/misc/unimp.h" #define MUSCA_NUMIRQ_MAX 96 #define MUSCA_PPC_MAX 3 +#define MUSCA_MPC_MAX 5 + +typedef struct MPCInfo MPCInfo; typedef enum MuscaType { MUSCA_A, @@ -44,19 +48,23 @@ typedef struct { uint32_t init_svtor; int sram_addr_width; int num_irqs; + const MPCInfo *mpc_info; + int num_mpcs; } MuscaMachineClass; typedef struct { MachineState parent; ARMSSE sse; + /* RAM and flash */ + MemoryRegion ram[MUSCA_MPC_MAX]; SplitIRQ cpu_irq_splitter[MUSCA_NUMIRQ_MAX]; SplitIRQ sec_resp_splitter; TZPPC ppc[MUSCA_PPC_MAX]; MemoryRegion container; UnimplementedDeviceState eflash[2]; UnimplementedDeviceState qspi; - UnimplementedDeviceState mpc[5]; + TZMPC mpc[MUSCA_MPC_MAX]; UnimplementedDeviceState mhu[2]; UnimplementedDeviceState pwm[3]; UnimplementedDeviceState i2s; @@ -69,6 +77,7 @@ typedef struct { UnimplementedDeviceState pvt; UnimplementedDeviceState sdio; UnimplementedDeviceState gpio; + UnimplementedDeviceState cryptoisland; } MuscaMachineState; #define TYPE_MUSCA_MACHINE "musca" @@ -131,6 +140,131 @@ static MemoryRegion *make_unimp_dev(MuscaMachineState *mms, return sysbus_mmio_get_region(SYS_BUS_DEVICE(uds), 0); } +typedef enum MPCInfoType { + MPC_RAM, + MPC_ROM, + MPC_CRYPTOISLAND, +} MPCInfoType; + +struct MPCInfo { + const char *name; + hwaddr addr; + hwaddr size; + MPCInfoType type; +}; + +/* Order of the MPCs here must match the order of the bits in SECMPCINTSTATUS */ +static const MPCInfo a_mpc_info[] = { { + .name = "qspi", + .type = MPC_ROM, + .addr = 0x00200000, + .size = 0x00800000, + }, { + .name = "sram", + .type = MPC_RAM, + .addr = 0x00000000, + .size = 0x00200000, + } +}; + +static const MPCInfo b1_mpc_info[] = { { + .name = "qspi", + .type = MPC_ROM, + .addr = 0x00000000, + .size = 0x02000000, + }, { + .name = "sram", + .type = MPC_RAM, + .addr = 0x0a400000, + .size = 0x00080000, + }, { + .name = "eflash0", + .type = MPC_ROM, + .addr = 0x0a000000, + .size = 0x00200000, + }, { + .name = "eflash1", + .type = MPC_ROM, + .addr = 0x0a200000, + .size = 0x00200000, + }, { + .name = "cryptoisland", + .type = MPC_CRYPTOISLAND, + .addr = 0x0a000000, + .size = 0x00200000, + } +}; + +static MemoryRegion *make_mpc(MuscaMachineState *mms, void *opaque, + const char *name, hwaddr size) +{ + /* + * Create an MPC and the RAM or flash behind it. + * MPC 0: eFlash 0 + * MPC 1: eFlash 1 + * MPC 2: SRAM + * MPC 3: QSPI flash + * MPC 4: CryptoIsland + * For now we implement the flash regions as ROM (ie not programmable) + * (with their control interface memory regions being unimplemented + * stubs behind the PPCs). + * The whole CryptoIsland region behind its MPC is an unimplemented stub. + */ + MuscaMachineClass *mmc = MUSCA_MACHINE_GET_CLASS(mms); + TZMPC *mpc = opaque; + int i = mpc - &mms->mpc[0]; + MemoryRegion *downstream; + MemoryRegion *upstream; + UnimplementedDeviceState *uds; + char *mpcname; + const MPCInfo *mpcinfo = mmc->mpc_info; + + mpcname = g_strdup_printf("%s-mpc", mpcinfo[i].name); + + switch (mpcinfo[i].type) { + case MPC_ROM: + downstream = &mms->ram[i]; + memory_region_init_rom(downstream, NULL, mpcinfo[i].name, + mpcinfo[i].size, &error_fatal); + break; + case MPC_RAM: + downstream = &mms->ram[i]; + memory_region_init_ram(downstream, NULL, mpcinfo[i].name, + mpcinfo[i].size, &error_fatal); + break; + case MPC_CRYPTOISLAND: + /* We don't implement the CryptoIsland yet */ + uds = &mms->cryptoisland; + sysbus_init_child_obj(OBJECT(mms), name, uds, + sizeof(UnimplementedDeviceState), + TYPE_UNIMPLEMENTED_DEVICE); + qdev_prop_set_string(DEVICE(uds), "name", mpcinfo[i].name); + qdev_prop_set_uint64(DEVICE(uds), "size", mpcinfo[i].size); + object_property_set_bool(OBJECT(uds), true, "realized", &error_fatal); + downstream = sysbus_mmio_get_region(SYS_BUS_DEVICE(uds), 0); + break; + default: + g_assert_not_reached(); + } + + sysbus_init_child_obj(OBJECT(mms), mpcname, mpc, sizeof(mms->mpc[0]), + TYPE_TZ_MPC); + object_property_set_link(OBJECT(mpc), OBJECT(downstream), + "downstream", &error_fatal); + object_property_set_bool(OBJECT(mpc), true, "realized", &error_fatal); + /* Map the upstream end of the MPC into system memory */ + upstream = sysbus_mmio_get_region(SYS_BUS_DEVICE(mpc), 1); + memory_region_add_subregion(get_system_memory(), mpcinfo[i].addr, upstream); + /* and connect its interrupt to the SSE-200 */ + qdev_connect_gpio_out_named(DEVICE(mpc), "irq", 0, + qdev_get_gpio_in_named(DEVICE(&mms->sse), + "mpcexp_status", i)); + + g_free(mpcname); + /* Return the register interface MR for our caller to map behind the PPC */ + return sysbus_mmio_get_region(SYS_BUS_DEVICE(mpc), 0); +} + static MemoryRegion *make_musca_a_devs(MuscaMachineState *mms, void *opaque, const char *name, hwaddr size) { @@ -160,8 +294,8 @@ static MemoryRegion *make_musca_a_devs(MuscaMachineState *mms, void *opaque, { "pwm1", make_unimp_dev, &mms->pwm[1], 0xe000, 0x1000 }, { "pwm2", make_unimp_dev, &mms->pwm[2], 0xf000, 0x1000 }, { "gpio", make_unimp_dev, &mms->gpio, 0x10000, 0x1000 }, - { "mpc0", make_unimp_dev, &mms->mpc[0], 0x12000, 0x1000 }, - { "mpc1", make_unimp_dev, &mms->mpc[1], 0x13000, 0x1000 }, + { "mpc0", make_mpc, &mms->mpc[0], 0x12000, 0x1000 }, + { "mpc1", make_mpc, &mms->mpc[1], 0x13000, 0x1000 }, }; memory_region_init(container, OBJECT(mms), "musca-device-container", size); @@ -190,6 +324,7 @@ static void musca_init(MachineState *machine) int i; assert(mmc->num_irqs <= MUSCA_NUMIRQ_MAX); + assert(mmc->num_mpcs <= MUSCA_MPC_MAX); if (strcmp(machine->cpu_type, mc->default_cpu_type) != 0) { error_report("This board can only be used with CPU %s", @@ -285,10 +420,10 @@ static void musca_init(MachineState *machine) { "eflash1", make_unimp_dev, &mms->eflash[1], 0x52500000, 0x1000 }, { "qspi", make_unimp_dev, &mms->qspi, 0x42800000, 0x100000 }, - { "mpc0", make_unimp_dev, &mms->mpc[0], 0x52000000, 0x1000 }, - { "mpc1", make_unimp_dev, &mms->mpc[1], 0x52100000, 0x1000 }, - { "mpc2", make_unimp_dev, &mms->mpc[2], 0x52200000, 0x1000 }, - { "mpc3", make_unimp_dev, &mms->mpc[3], 0x52300000, 0x1000 }, + { "mpc0", make_mpc, &mms->mpc[0], 0x52000000, 0x1000 }, + { "mpc1", make_mpc, &mms->mpc[1], 0x52100000, 0x1000 }, + { "mpc2", make_mpc, &mms->mpc[2], 0x52200000, 0x1000 }, + { "mpc3", make_mpc, &mms->mpc[3], 0x52300000, 0x1000 }, { "mhu0", make_unimp_dev, &mms->mhu[0], 0x42600000, 0x100000 }, { "mhu1", make_unimp_dev, &mms->mhu[1], 0x42700000, 0x100000 }, { }, /* port 9: unused */ @@ -296,7 +431,7 @@ static void musca_init(MachineState *machine) { }, /* port 11: unused */ { }, /* port 12: unused */ { }, /* port 13: unused */ - { "mpc4", make_unimp_dev, &mms->mpc[4], 0x52e00000, 0x1000 }, + { "mpc4", make_mpc, &mms->mpc[4], 0x52e00000, 0x1000 }, }, }, { .name = "apb_ppcexp1", @@ -434,6 +569,8 @@ static void musca_a_class_init(ObjectClass *oc, void *data) mmc->init_svtor = 0x10200000; mmc->sram_addr_width = 15; mmc->num_irqs = 64; + mmc->mpc_info = a_mpc_info; + mmc->num_mpcs = ARRAY_SIZE(a_mpc_info); } static void musca_b1_class_init(ObjectClass *oc, void *data) @@ -453,6 +590,8 @@ static void musca_b1_class_init(ObjectClass *oc, void *data) mmc->init_svtor = 0x10000000; mmc->sram_addr_width = 17; mmc->num_irqs = 96; + mmc->mpc_info = b1_mpc_info; + mmc->num_mpcs = ARRAY_SIZE(b1_mpc_info); } static const TypeInfo musca_info = { From 4db6a761a570941570e287e21e8126995218c654 Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Thu, 21 Feb 2019 18:17:47 +0000 Subject: [PATCH 19/21] hw/arm/musca: Wire up PL031 RTC Wire up the PL031 RTC for the Musca board. Signed-off-by: Peter Maydell Reviewed-by: Richard Henderson --- hw/arm/musca.c | 26 +++++++++++++++++++++++--- 1 file changed, 23 insertions(+), 3 deletions(-) diff --git a/hw/arm/musca.c b/hw/arm/musca.c index c4095c4579..378912b738 100644 --- a/hw/arm/musca.c +++ b/hw/arm/musca.c @@ -30,6 +30,7 @@ #include "hw/misc/tz-mpc.h" #include "hw/misc/tz-ppc.h" #include "hw/misc/unimp.h" +#include "hw/timer/pl031.h" #define MUSCA_NUMIRQ_MAX 96 #define MUSCA_PPC_MAX 3 @@ -73,7 +74,7 @@ typedef struct { UnimplementedDeviceState spi; UnimplementedDeviceState scc; UnimplementedDeviceState timer; - UnimplementedDeviceState rtc; + PL031State rtc; UnimplementedDeviceState pvt; UnimplementedDeviceState sdio; UnimplementedDeviceState gpio; @@ -98,6 +99,14 @@ typedef struct { */ #define SYSCLK_FRQ 40000000 +static qemu_irq get_sse_irq_in(MuscaMachineState *mms, int irqno) +{ + /* Return a qemu_irq which will signal IRQ n to all CPUs in the SSE. */ + assert(irqno < MUSCA_NUMIRQ_MAX); + + return qdev_get_gpio_in(DEVICE(&mms->cpu_irq_splitter[irqno]), 0); +} + /* * Most of the devices in the Musca board sit behind Peripheral Protection * Controllers. These data structures define the layout of which devices @@ -265,6 +274,17 @@ static MemoryRegion *make_mpc(MuscaMachineState *mms, void *opaque, return sysbus_mmio_get_region(SYS_BUS_DEVICE(mpc), 0); } +static MemoryRegion *make_rtc(MuscaMachineState *mms, void *opaque, + const char *name, hwaddr size) +{ + PL031State *rtc = opaque; + + sysbus_init_child_obj(OBJECT(mms), name, rtc, sizeof(mms->rtc), TYPE_PL031); + object_property_set_bool(OBJECT(rtc), true, "realized", &error_fatal); + sysbus_connect_irq(SYS_BUS_DEVICE(rtc), 0, get_sse_irq_in(mms, 39)); + return sysbus_mmio_get_region(SYS_BUS_DEVICE(rtc), 0); +} + static MemoryRegion *make_musca_a_devs(MuscaMachineState *mms, void *opaque, const char *name, hwaddr size) { @@ -287,7 +307,7 @@ static MemoryRegion *make_musca_a_devs(MuscaMachineState *mms, void *opaque, { "i2c1", make_unimp_dev, &mms->i2c[1], 0x5000, 0x1000 }, { "i2s", make_unimp_dev, &mms->i2s, 0x6000, 0x1000 }, { "pwm0", make_unimp_dev, &mms->pwm[0], 0x7000, 0x1000 }, - { "rtc", make_unimp_dev, &mms->rtc, 0x8000, 0x1000 }, + { "rtc", make_rtc, &mms->rtc, 0x8000, 0x1000 }, { "qspi", make_unimp_dev, &mms->qspi, 0xa000, 0x1000 }, { "timer", make_unimp_dev, &mms->timer, 0xb000, 0x1000 }, { "scc", make_unimp_dev, &mms->scc, 0xc000, 0x1000 }, @@ -447,7 +467,7 @@ static void musca_init(MachineState *machine) { "spi", make_unimp_dev, &mms->spi, 0x4010a000, 0x1000 }, { "scc", make_unimp_dev, &mms->scc, 0x5010b000, 0x1000 }, { "timer", make_unimp_dev, &mms->timer, 0x4010c000, 0x1000 }, - { "rtc", make_unimp_dev, &mms->rtc, 0x4010d000, 0x1000 }, + { "rtc", make_rtc, &mms->rtc, 0x4010d000, 0x1000 }, { "pvt", make_unimp_dev, &mms->pvt, 0x4010e000, 0x1000 }, { "sdio", make_unimp_dev, &mms->sdio, 0x4010f000, 0x1000 }, }, From 1486f1bac3eb81b43138aa5a7e1434798fe4bd97 Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Thu, 21 Feb 2019 18:17:47 +0000 Subject: [PATCH 20/21] hw/arm/musca: Wire up PL011 UARTs Wire up the two PL011 UARTs in the Musca board. Signed-off-by: Peter Maydell Reviewed-by: Richard Henderson --- hw/arm/musca.c | 34 +++++++++++++++++++++++++++++----- 1 file changed, 29 insertions(+), 5 deletions(-) diff --git a/hw/arm/musca.c b/hw/arm/musca.c index 378912b738..23aff43f4b 100644 --- a/hw/arm/musca.c +++ b/hw/arm/musca.c @@ -23,9 +23,11 @@ #include "qemu/error-report.h" #include "qapi/error.h" #include "exec/address-spaces.h" +#include "sysemu/sysemu.h" #include "hw/arm/arm.h" #include "hw/arm/armsse.h" #include "hw/boards.h" +#include "hw/char/pl011.h" #include "hw/core/split-irq.h" #include "hw/misc/tz-mpc.h" #include "hw/misc/tz-ppc.h" @@ -69,7 +71,7 @@ typedef struct { UnimplementedDeviceState mhu[2]; UnimplementedDeviceState pwm[3]; UnimplementedDeviceState i2s; - UnimplementedDeviceState uart[2]; + PL011State uart[2]; UnimplementedDeviceState i2c[2]; UnimplementedDeviceState spi; UnimplementedDeviceState scc; @@ -285,6 +287,28 @@ static MemoryRegion *make_rtc(MuscaMachineState *mms, void *opaque, return sysbus_mmio_get_region(SYS_BUS_DEVICE(rtc), 0); } +static MemoryRegion *make_uart(MuscaMachineState *mms, void *opaque, + const char *name, hwaddr size) +{ + PL011State *uart = opaque; + int i = uart - &mms->uart[0]; + int irqbase = 7 + i * 6; + SysBusDevice *s; + + sysbus_init_child_obj(OBJECT(mms), name, uart, sizeof(mms->uart[0]), + TYPE_PL011); + qdev_prop_set_chr(DEVICE(uart), "chardev", serial_hd(i)); + object_property_set_bool(OBJECT(uart), true, "realized", &error_fatal); + s = SYS_BUS_DEVICE(uart); + sysbus_connect_irq(s, 0, get_sse_irq_in(mms, irqbase + 5)); /* combined */ + sysbus_connect_irq(s, 1, get_sse_irq_in(mms, irqbase + 0)); /* RX */ + sysbus_connect_irq(s, 2, get_sse_irq_in(mms, irqbase + 1)); /* TX */ + sysbus_connect_irq(s, 3, get_sse_irq_in(mms, irqbase + 2)); /* RT */ + sysbus_connect_irq(s, 4, get_sse_irq_in(mms, irqbase + 3)); /* MS */ + sysbus_connect_irq(s, 5, get_sse_irq_in(mms, irqbase + 4)); /* E */ + return sysbus_mmio_get_region(SYS_BUS_DEVICE(uart), 0); +} + static MemoryRegion *make_musca_a_devs(MuscaMachineState *mms, void *opaque, const char *name, hwaddr size) { @@ -300,8 +324,8 @@ static MemoryRegion *make_musca_a_devs(MuscaMachineState *mms, void *opaque, MemoryRegion *container = &mms->container; const PPCPortInfo devices[] = { - { "uart0", make_unimp_dev, &mms->uart[0], 0x1000, 0x1000 }, - { "uart1", make_unimp_dev, &mms->uart[1], 0x2000, 0x1000 }, + { "uart0", make_uart, &mms->uart[0], 0x1000, 0x1000 }, + { "uart1", make_uart, &mms->uart[1], 0x2000, 0x1000 }, { "spi", make_unimp_dev, &mms->spi, 0x3000, 0x1000 }, { "i2c0", make_unimp_dev, &mms->i2c[0], 0x4000, 0x1000 }, { "i2c1", make_unimp_dev, &mms->i2c[1], 0x5000, 0x1000 }, @@ -460,8 +484,8 @@ static void musca_init(MachineState *machine) { "pwm1", make_unimp_dev, &mms->pwm[1], 0x40102000, 0x1000 }, { "pwm2", make_unimp_dev, &mms->pwm[2], 0x40103000, 0x1000 }, { "i2s", make_unimp_dev, &mms->i2s, 0x40104000, 0x1000 }, - { "uart0", make_unimp_dev, &mms->uart[0], 0x40105000, 0x1000 }, - { "uart1", make_unimp_dev, &mms->uart[1], 0x40106000, 0x1000 }, + { "uart0", make_uart, &mms->uart[0], 0x40105000, 0x1000 }, + { "uart1", make_uart, &mms->uart[1], 0x40106000, 0x1000 }, { "i2c0", make_unimp_dev, &mms->i2c[0], 0x40108000, 0x1000 }, { "i2c1", make_unimp_dev, &mms->i2c[1], 0x40109000, 0x1000 }, { "spi", make_unimp_dev, &mms->spi, 0x4010a000, 0x1000 }, From 3733f80308d2a7f23f5e39b039e0547aba6c07f1 Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Thu, 21 Feb 2019 18:17:48 +0000 Subject: [PATCH 21/21] hw/arm/armsse: Make 0x5... alias region work for per-CPU devices MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The region 0x40010000 .. 0x4001ffff and its secure-only alias at 0x50010000... are for per-CPU devices. We implement this by giving each CPU its own container memory region, where the per-CPU devices live. Unfortunately, the alias region which makes devices mapped at 0x4... addresses also appear at 0x5... is only implemented in the overall "all CPUs" container. The effect of this bug is that the CPU_IDENTITY register block appears only at 0x4001f000, but not at the 0x5001f000 alias where it should also appear. Guests (like very recent Arm Trusted Firmware-M) which try to access it at 0x5001f000 will crash. Fix this by moving the handling for this alias from the "all CPUs" container to the per-CPU container. (We leave the aliases for 0x1... and 0x3... in the overall container, because there are no per-CPU devices there.) Signed-off-by: Peter Maydell Message-id: 20190215180500.6906-1-peter.maydell@linaro.org Reviewed-by: Alex Bennée --- hw/arm/armsse.c | 26 ++++++++++++++++---------- include/hw/arm/armsse.h | 2 +- 2 files changed, 17 insertions(+), 11 deletions(-) diff --git a/hw/arm/armsse.c b/hw/arm/armsse.c index 50da41f64c..129e7ea7fe 100644 --- a/hw/arm/armsse.c +++ b/hw/arm/armsse.c @@ -110,15 +110,16 @@ static bool irq_is_common[32] = { /* 30, 31: reserved */ }; -/* Create an alias region of @size bytes starting at @base +/* + * Create an alias region in @container of @size bytes starting at @base * which mirrors the memory starting at @orig. */ -static void make_alias(ARMSSE *s, MemoryRegion *mr, const char *name, - hwaddr base, hwaddr size, hwaddr orig) +static void make_alias(ARMSSE *s, MemoryRegion *mr, MemoryRegion *container, + const char *name, hwaddr base, hwaddr size, hwaddr orig) { - memory_region_init_alias(mr, NULL, name, &s->container, orig, size); + memory_region_init_alias(mr, NULL, name, container, orig, size); /* The alias is even lower priority than unimplemented_device regions */ - memory_region_add_subregion_overlap(&s->container, base, mr, -1500); + memory_region_add_subregion_overlap(container, base, mr, -1500); } static void irq_status_forwarder(void *opaque, int n, int level) @@ -607,16 +608,21 @@ static void armsse_realize(DeviceState *dev, Error **errp) } /* Set up the big aliases first */ - make_alias(s, &s->alias1, "alias 1", 0x10000000, 0x10000000, 0x00000000); - make_alias(s, &s->alias2, "alias 2", 0x30000000, 0x10000000, 0x20000000); + make_alias(s, &s->alias1, &s->container, "alias 1", + 0x10000000, 0x10000000, 0x00000000); + make_alias(s, &s->alias2, &s->container, + "alias 2", 0x30000000, 0x10000000, 0x20000000); /* The 0x50000000..0x5fffffff region is not a pure alias: it has * a few extra devices that only appear there (generally the * control interfaces for the protection controllers). * We implement this by mapping those devices over the top of this - * alias MR at a higher priority. + * alias MR at a higher priority. Some of the devices in this range + * are per-CPU, so we must put this alias in the per-cpu containers. */ - make_alias(s, &s->alias3, "alias 3", 0x50000000, 0x10000000, 0x40000000); - + for (i = 0; i < info->num_cpus; i++) { + make_alias(s, &s->alias3[i], &s->cpu_container[i], + "alias 3", 0x50000000, 0x10000000, 0x40000000); + } /* Security controller */ object_property_set_bool(OBJECT(&s->secctl), true, "realized", &err); diff --git a/include/hw/arm/armsse.h b/include/hw/arm/armsse.h index 84879f40dd..7ef871c7df 100644 --- a/include/hw/arm/armsse.h +++ b/include/hw/arm/armsse.h @@ -186,7 +186,7 @@ typedef struct ARMSSE { MemoryRegion cpu_container[SSE_MAX_CPUS]; MemoryRegion alias1; MemoryRegion alias2; - MemoryRegion alias3; + MemoryRegion alias3[SSE_MAX_CPUS]; MemoryRegion sram[MAX_SRAM_BANKS]; qemu_irq *exp_irqs[SSE_MAX_CPUS];