target-arm queue: the big things in here are SVE in system emulation mode, and v8M stack limit checking; there are also a handful of smaller fixes. thanks -- PMM The following changes since commit 079911cb6e26898e16f5bb56ef4f9d33cf92d32d: Merge remote-tracking branch 'remotes/rth/tags/pull-fpu-20181005' into staging (2018-10-08 12:44:35 +0100) are available in the Git repository at: https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20181008 for you to fetch changes up to 74e2e59b8d0a68be0956310fc349179c89fd7be0: hw/display/bcm2835_fb: Silence Coverity warning about multiply overflow (2018-10-08 14:55:05 +0100) ---------------------------------------------------------------- target-arm queue: * target/arm: fix error in a code comment * virt: Suppress external aborts on virt-2.10 and earlier * target/arm: Correct condition for v8M callee stack push * target/arm: Don't read r4 from v8M exception stackframe twice * target/arm: Support SVE in system emulation mode * target/arm: Implement v8M hardware stack limit checking * hw/display/bcm2835_fb: Silence Coverity warning about multiply overflow ---------------------------------------------------------------- Dongjiu Geng (1): target/arm: fix code comments error Peter Maydell (17): virt: Suppress external aborts on virt-2.10 and earlier target/arm: Correct condition for v8M callee stack push target/arm: Don't read r4 from v8M exception stackframe twice target/arm: Define new TBFLAG for v8M stack checking target/arm: Define new EXCP type for v8M stack overflows target/arm: Move v7m_using_psp() to internals.h target/arm: Add v8M stack checks on ADD/SUB/MOV of SP target/arm: Add some comments in Thumb decode target/arm: Add v8M stack checks on exception entry target/arm: Add v8M stack limit checks on NS function calls target/arm: Add v8M stack checks for LDRD/STRD (imm) target/arm: Add v8M stack checks for Thumb2 LDM/STM target/arm: Add v8M stack checks for T32 load/store single target/arm: Add v8M stack checks for Thumb push/pop target/arm: Add v8M stack checks for VLDM/VSTM target/arm: Add v8M stack checks for MSR to SP_NS hw/display/bcm2835_fb: Silence Coverity warning about multiply overflow Richard Henderson (15): target/arm: Define ID_AA64ZFR0_EL1 target/arm: Adjust sve_exception_el target/arm: Pass in current_el to fp and sve_exception_el target/arm: Handle SVE vector length changes in system mode target/arm: Adjust aarch64_cpu_dump_state for system mode SVE target/arm: Clear unused predicate bits for LD1RQ target/arm: Rewrite helper_sve_ld1*_r using pages target/arm: Rewrite helper_sve_ld[234]*_r target/arm: Rewrite helper_sve_st[1234]*_r target/arm: Split contiguous loads for endianness target/arm: Split contiguous stores for endianness target/arm: Rewrite vector gather loads target/arm: Rewrite vector gather stores target/arm: Rewrite vector gather first-fault loads target/arm: Pass TCGMemOpIdx to sve memory helpers target/arm/cpu.h | 17 + target/arm/helper-sve.h | 385 ++++++--- target/arm/helper.h | 2 + target/arm/internals.h | 44 + target/arm/kvm_arm.h | 4 +- target/arm/translate.h | 1 + hw/arm/virt.c | 2 + hw/display/bcm2835_fb.c | 2 +- target/arm/cpu64.c | 42 - target/arm/helper.c | 345 +++++--- target/arm/kvm.c | 2 +- target/arm/op_helper.c | 24 +- target/arm/sve_helper.c | 1961 ++++++++++++++++++++++++++++++-------------- target/arm/translate-a64.c | 8 +- target/arm/translate-sve.c | 670 ++++++++++----- target/arm/translate.c | 198 ++++- 16 files changed, 2611 insertions(+), 1096 deletions(-) From: Dongjiu Geng <gengdongjiu@huawei.com> The parameter of kvm_arm_init_cpreg_list() is ARMCPU instead of CPUState, so correct the note to make it match the code. Signed-off-by: Dongjiu Geng <gengdongjiu@huawei.com> Message-id: 1538069046-5757-1-git-send-email-gengdongjiu@huawei.com Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/kvm_arm.h | 4 ++-- target/arm/kvm.c | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/target/arm/kvm_arm.h b/target/arm/kvm_arm.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/kvm_arm.h +++ b/target/arm/kvm_arm.h @@ -XXX,XX +XXX,XX @@ void kvm_arm_register_device(MemoryRegion *mr, uint64_t devid, uint64_t group, /** * kvm_arm_init_cpreg_list: - * @cs: CPUState + * @cpu: ARMCPU * - * Initialize the CPUState's cpreg list according to the kernel's + * Initialize the ARMCPU cpreg list according to the kernel's * definition of what CPU registers it knows about (and throw away * the previous TCG-created cpreg list). * diff --git a/target/arm/kvm.c b/target/arm/kvm.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/kvm.c +++ b/target/arm/kvm.c @@ -XXX,XX +XXX,XX @@ static int compare_u64(const void *a, const void *b) return 0; } -/* Initialize the CPUState's cpreg list according to the kernel's +/* Initialize the ARMCPU cpreg list according to the kernel's * definition of what CPU registers it knows about (and throw away * the previous TCG-created cpreg list). */ -- 2.19.0 In commit c79c0a314c43b78 we enabled emulation of external aborts when the guest attempts to access a physical address with no mapped device. In commit 4672cbd7bed88dc6 we suppress this for most legacy boards to prevent breakage of previously working guests, but we didn't suppress it in the 'virt' board, with the rationale "we know that guests won't try to prod devices that we don't describe in the device tree or ACPI tables". This is mostly true, but we've had a report of a Linux guest image that this did break. The problem seems to be that the guest is (incorrectly) configured with a DEBUG_UART_PHYS value that tells it there is a uart at 0x10009000 (which is true for vexpress but not for virt), so in early bootup the kernel probes this bogus address. This is a misconfigured guest, so we don't need to worry about it too much, but we can arrange that guests that ran on QEMU v2.10 (before c79c0a314c43b78) will still run on the "virt-2.10" board model, by suppressing external aborts only for that version and earlier. This seems a reasonable compromise: "virt-2.10" is supposed to behave the same way that "virt" did in the 2.10 release, and making it do that provides a usable workaround for guests with bugs like this. Cc: qemu-stable@nongnu.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20180925144127.31965-1-peter.maydell@linaro.org Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> --- hw/arm/virt.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/hw/arm/virt.c b/hw/arm/virt.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/virt.c +++ b/hw/arm/virt.c @@ -XXX,XX +XXX,XX @@ static void virt_machine_2_10_options(MachineClass *mc) { virt_machine_2_11_options(mc); SET_MACHINE_COMPAT(mc, VIRT_COMPAT_2_10); + /* before 2.11 we never faulted accesses to bad addresses */ + mc->ignore_memory_transaction_failures = true; } DEFINE_VIRT_MACHINE(2, 10) -- 2.19.0 In v7m_exception_taken() we were incorrectly using a "LR bit EXCRET.ES is 1" check when it should be 0 (compare the pseudocode ExceptionTaken() function). This meant we didn't stack the callee-saved registers when tailchaining from a NonSecure to a Secure exception. Cc: qemu-stable@nongnu.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181002145940.30931-1-peter.maydell@linaro.org --- 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 XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void v7m_exception_taken(ARMCPU *cpu, uint32_t lr, bool dotailchain, * not already saved. */ if (lr & R_V7M_EXCRET_DCRS_MASK && - !(dotailchain && (lr & R_V7M_EXCRET_ES_MASK))) { + !(dotailchain && !(lr & R_V7M_EXCRET_ES_MASK))) { push_failed = v7m_push_callee_stack(cpu, lr, dotailchain, ignore_stackfaults); } -- 2.19.0 A cut-and-paste error meant we were reading r4 from the v8M callee-saves exception stack frame twice. This is harmless since it just meant we did two memory accesses to the same location, but it's unnecessary. Delete it. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181002150304.2287-1-peter.maydell@linaro.org --- target/arm/helper.c | 1 - 1 file changed, 1 deletion(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void do_v7m_exception_exit(ARMCPU *cpu) } pop_ok = pop_ok && - v7m_stack_read(cpu, &env->regs[4], frameptr + 0x8, mmu_idx) && v7m_stack_read(cpu, &env->regs[4], frameptr + 0x8, mmu_idx) && v7m_stack_read(cpu, &env->regs[5], frameptr + 0xc, mmu_idx) && v7m_stack_read(cpu, &env->regs[6], frameptr + 0x10, mmu_idx) && -- 2.19.0 From: Richard Henderson <richard.henderson@linaro.org> Given that the only field defined for this new register may only be 0, we don't actually need to change anything except the name. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181005175350.30752-2-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ void register_cp_regs_for_features(ARMCPU *cpu) .opc0 = 3, .opc1 = 0, .crn = 0, .crm = 4, .opc2 = 3, .access = PL1_R, .type = ARM_CP_CONST, .resetvalue = 0 }, - { .name = "ID_AA64PFR4_EL1_RESERVED", .state = ARM_CP_STATE_AA64, + { .name = "ID_AA64ZFR0_EL1", .state = ARM_CP_STATE_AA64, .opc0 = 3, .opc1 = 0, .crn = 0, .crm = 4, .opc2 = 4, .access = PL1_R, .type = ARM_CP_CONST, + /* At present, only SVEver == 0 is defined anyway. */ .resetvalue = 0 }, { .name = "ID_AA64PFR5_EL1_RESERVED", .state = ARM_CP_STATE_AA64, .opc0 = 3, .opc1 = 0, .crn = 0, .crm = 4, .opc2 = 5, -- 2.19.0 From: Richard Henderson <richard.henderson@linaro.org> Check for EL3 before testing CPTR_EL3.EZ. Return 0 when the exception should be routed via AdvSIMDFPAccessTrap. Mirror the structure of CheckSVEEnabled more closely. Fixes: 5be5e8eda78 Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181005175350.30752-3-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 96 ++++++++++++++++++++++----------------------- 1 file changed, 46 insertions(+), 50 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo debug_lpae_cp_reginfo[] = { REGINFO_SENTINEL }; -/* Return the exception level to which SVE-disabled exceptions should - * be taken, or 0 if SVE is enabled. +/* Return the exception level to which exceptions should be taken + * via SVEAccessTrap. If an exception should be routed through + * AArch64.AdvSIMDFPAccessTrap, return 0; fp_exception_el should + * take care of raising that exception. + * C.f. the ARM pseudocode function CheckSVEEnabled. */ static int sve_exception_el(CPUARMState *env) { #ifndef CONFIG_USER_ONLY unsigned current_el = arm_current_el(env); - /* The CPACR.ZEN controls traps to EL1: - * 0, 2 : trap EL0 and EL1 accesses - * 1 : trap only EL0 accesses - * 3 : trap no accesses + if (current_el <= 1) { + bool disabled = false; + + /* The CPACR.ZEN controls traps to EL1: + * 0, 2 : trap EL0 and EL1 accesses + * 1 : trap only EL0 accesses + * 3 : trap no accesses + */ + if (!extract32(env->cp15.cpacr_el1, 16, 1)) { + disabled = true; + } else if (!extract32(env->cp15.cpacr_el1, 17, 1)) { + disabled = current_el == 0; + } + if (disabled) { + /* route_to_el2 */ + return (arm_feature(env, ARM_FEATURE_EL2) + && !arm_is_secure(env) + && (env->cp15.hcr_el2 & HCR_TGE) ? 2 : 1); + } + + /* Check CPACR.FPEN. */ + if (!extract32(env->cp15.cpacr_el1, 20, 1)) { + disabled = true; + } else if (!extract32(env->cp15.cpacr_el1, 21, 1)) { + disabled = current_el == 0; + } + if (disabled) { + return 0; + } + } + + /* CPTR_EL2. Since TZ and TFP are positive, + * they will be zero when EL2 is not present. */ - switch (extract32(env->cp15.cpacr_el1, 16, 2)) { - default: - if (current_el <= 1) { - /* Trap to PL1, which might be EL1 or EL3 */ - if (arm_is_secure(env) && !arm_el_is_aa64(env, 3)) { - return 3; - } - return 1; + if (current_el <= 2 && !arm_is_secure_below_el3(env)) { + if (env->cp15.cptr_el[2] & CPTR_TZ) { + return 2; } - break; - case 1: - if (current_el == 0) { - return 1; + if (env->cp15.cptr_el[2] & CPTR_TFP) { + return 0; } - break; - case 3: - break; } - /* Similarly for CPACR.FPEN, after having checked ZEN. */ - switch (extract32(env->cp15.cpacr_el1, 20, 2)) { - default: - if (current_el <= 1) { - if (arm_is_secure(env) && !arm_el_is_aa64(env, 3)) { - return 3; - } - return 1; - } - break; - case 1: - if (current_el == 0) { - return 1; - } - break; - case 3: - break; - } - - /* CPTR_EL2. Check both TZ and TFP. */ - if (current_el <= 2 - && (env->cp15.cptr_el[2] & (CPTR_TFP | CPTR_TZ)) - && !arm_is_secure_below_el3(env)) { - return 2; - } - - /* CPTR_EL3. Check both EZ and TFP. */ - if (!(env->cp15.cptr_el[3] & CPTR_EZ) - || (env->cp15.cptr_el[3] & CPTR_TFP)) { + /* CPTR_EL3. Since EZ is negative we must check for EL3. */ + if (arm_feature(env, ARM_FEATURE_EL3) + && !(env->cp15.cptr_el[3] & CPTR_EZ)) { return 3; } #endif -- 2.19.0 From: Richard Henderson <richard.henderson@linaro.org> We are going to want to determine whether sve is enabled for EL other than current. Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181005175350.30752-4-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 21 +++++++++------------ 1 file changed, 9 insertions(+), 12 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo debug_lpae_cp_reginfo[] = { * take care of raising that exception. * C.f. the ARM pseudocode function CheckSVEEnabled. */ -static int sve_exception_el(CPUARMState *env) +static int sve_exception_el(CPUARMState *env, int el) { #ifndef CONFIG_USER_ONLY - unsigned current_el = arm_current_el(env); - - if (current_el <= 1) { + if (el <= 1) { bool disabled = false; /* The CPACR.ZEN controls traps to EL1: @@ -XXX,XX +XXX,XX @@ static int sve_exception_el(CPUARMState *env) if (!extract32(env->cp15.cpacr_el1, 16, 1)) { disabled = true; } else if (!extract32(env->cp15.cpacr_el1, 17, 1)) { - disabled = current_el == 0; + disabled = el == 0; } if (disabled) { /* route_to_el2 */ @@ -XXX,XX +XXX,XX @@ static int sve_exception_el(CPUARMState *env) if (!extract32(env->cp15.cpacr_el1, 20, 1)) { disabled = true; } else if (!extract32(env->cp15.cpacr_el1, 21, 1)) { - disabled = current_el == 0; + disabled = el == 0; } if (disabled) { return 0; @@ -XXX,XX +XXX,XX @@ static int sve_exception_el(CPUARMState *env) /* CPTR_EL2. Since TZ and TFP are positive, * they will be zero when EL2 is not present. */ - if (current_el <= 2 && !arm_is_secure_below_el3(env)) { + if (el <= 2 && !arm_is_secure_below_el3(env)) { if (env->cp15.cptr_el[2] & CPTR_TZ) { return 2; } @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(crc32c)(uint32_t acc, uint32_t val, uint32_t bytes) /* Return the exception level to which FP-disabled exceptions should * be taken, or 0 if FP is enabled. */ -static inline int fp_exception_el(CPUARMState *env) +static int fp_exception_el(CPUARMState *env, int cur_el) { #ifndef CONFIG_USER_ONLY int fpen; - int cur_el = arm_current_el(env); /* CPACR and the CPTR registers don't exist before v6, so FP is * always accessible @@ -XXX,XX +XXX,XX @@ void cpu_get_tb_cpu_state(CPUARMState *env, target_ulong *pc, target_ulong *cs_base, uint32_t *pflags) { ARMMMUIdx mmu_idx = core_to_arm_mmu_idx(env, cpu_mmu_index(env, false)); - int fp_el = fp_exception_el(env); + int current_el = arm_current_el(env); + int fp_el = fp_exception_el(env, current_el); uint32_t flags; if (is_a64(env)) { @@ -XXX,XX +XXX,XX @@ void cpu_get_tb_cpu_state(CPUARMState *env, target_ulong *pc, flags |= (arm_regime_tbi1(env, mmu_idx) << ARM_TBFLAG_TBI1_SHIFT); if (arm_feature(env, ARM_FEATURE_SVE)) { - int sve_el = sve_exception_el(env); + int sve_el = sve_exception_el(env, current_el); uint32_t zcr_len; /* If SVE is disabled, but FP is enabled, @@ -XXX,XX +XXX,XX @@ void cpu_get_tb_cpu_state(CPUARMState *env, target_ulong *pc, if (sve_el != 0 && fp_el == 0) { zcr_len = 0; } else { - int current_el = arm_current_el(env); ARMCPU *cpu = arm_env_get_cpu(env); zcr_len = cpu->sve_max_vq - 1; -- 2.19.0 From: Richard Henderson <richard.henderson@linaro.org> SVE vector length can change when changing EL, or when writing to one of the ZCR_ELn registers. For correctness, our implementation requires that predicate bits that are inaccessible are never set. Which means noticing length changes and zeroing the appropriate register bits. Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181005175350.30752-5-richard.henderson@linaro.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 4 ++ target/arm/cpu64.c | 42 ------------- target/arm/helper.c | 133 +++++++++++++++++++++++++++++++++++++---- target/arm/op_helper.c | 1 + 4 files changed, 125 insertions(+), 55 deletions(-) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ int arm_cpu_write_elf32_note(WriteCoreDumpFunction f, CPUState *cs, int aarch64_cpu_gdb_read_register(CPUState *cpu, uint8_t *buf, int reg); int aarch64_cpu_gdb_write_register(CPUState *cpu, uint8_t *buf, int reg); void aarch64_sve_narrow_vq(CPUARMState *env, unsigned vq); +void aarch64_sve_change_el(CPUARMState *env, int old_el, int new_el); +#else +static inline void aarch64_sve_narrow_vq(CPUARMState *env, unsigned vq) { } +static inline void aarch64_sve_change_el(CPUARMState *env, int o, int n) { } #endif target_ulong do_arm_semihosting(CPUARMState *env); diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu64.c +++ b/target/arm/cpu64.c @@ -XXX,XX +XXX,XX @@ static void aarch64_cpu_register_types(void) } type_init(aarch64_cpu_register_types) - -/* The manual says that when SVE is enabled and VQ is widened the - * implementation is allowed to zero the previously inaccessible - * portion of the registers. The corollary to that is that when - * SVE is enabled and VQ is narrowed we are also allowed to zero - * the now inaccessible portion of the registers. - * - * The intent of this is that no predicate bit beyond VQ is ever set. - * Which means that some operations on predicate registers themselves - * may operate on full uint64_t or even unrolled across the maximum - * uint64_t[4]. Performing 4 bits of host arithmetic unconditionally - * may well be cheaper than conditionals to restrict the operation - * to the relevant portion of a uint16_t[16]. - * - * TODO: Need to call this for changes to the real system registers - * and EL state changes. - */ -void aarch64_sve_narrow_vq(CPUARMState *env, unsigned vq) -{ - int i, j; - uint64_t pmask; - - assert(vq >= 1 && vq <= ARM_MAX_VQ); - assert(vq <= arm_env_get_cpu(env)->sve_max_vq); - - /* Zap the high bits of the zregs. */ - for (i = 0; i < 32; i++) { - memset(&env->vfp.zregs[i].d[2 * vq], 0, 16 * (ARM_MAX_VQ - vq)); - } - - /* Zap the high bits of the pregs and ffr. */ - pmask = 0; - if (vq & 3) { - pmask = ~(-1ULL << (16 * (vq & 3))); - } - for (j = vq / 4; j < ARM_MAX_VQ / 4; j++) { - for (i = 0; i < 17; ++i) { - env->vfp.pregs[i].p[j] &= pmask; - } - pmask = 0; - } -} diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static int sve_exception_el(CPUARMState *env, int el) return 0; } +/* + * Given that SVE is enabled, return the vector length for EL. + */ +static uint32_t sve_zcr_len_for_el(CPUARMState *env, int el) +{ + ARMCPU *cpu = arm_env_get_cpu(env); + uint32_t zcr_len = cpu->sve_max_vq - 1; + + if (el <= 1) { + zcr_len = MIN(zcr_len, 0xf & (uint32_t)env->vfp.zcr_el[1]); + } + if (el < 2 && arm_feature(env, ARM_FEATURE_EL2)) { + zcr_len = MIN(zcr_len, 0xf & (uint32_t)env->vfp.zcr_el[2]); + } + if (el < 3 && arm_feature(env, ARM_FEATURE_EL3)) { + zcr_len = MIN(zcr_len, 0xf & (uint32_t)env->vfp.zcr_el[3]); + } + return zcr_len; +} + static void zcr_write(CPUARMState *env, const ARMCPRegInfo *ri, uint64_t value) { + int cur_el = arm_current_el(env); + int old_len = sve_zcr_len_for_el(env, cur_el); + int new_len; + /* Bits other than [3:0] are RAZ/WI. */ raw_write(env, ri, value & 0xf); + + /* + * Because we arrived here, we know both FP and SVE are enabled; + * otherwise we would have trapped access to the ZCR_ELn register. + */ + new_len = sve_zcr_len_for_el(env, cur_el); + if (new_len < old_len) { + aarch64_sve_narrow_vq(env, new_len + 1); + } } static const ARMCPRegInfo zcr_el1_reginfo = { @@ -XXX,XX +XXX,XX @@ static void arm_cpu_do_interrupt_aarch64(CPUState *cs) unsigned int new_el = env->exception.target_el; target_ulong addr = env->cp15.vbar_el[new_el]; unsigned int new_mode = aarch64_pstate_mode(new_el, true); + unsigned int cur_el = arm_current_el(env); - if (arm_current_el(env) < new_el) { + aarch64_sve_change_el(env, cur_el, new_el); + + if (cur_el < new_el) { /* Entry vector offset depends on whether the implemented EL * immediately lower than the target level is using AArch32 or AArch64 */ @@ -XXX,XX +XXX,XX @@ void cpu_get_tb_cpu_state(CPUARMState *env, target_ulong *pc, if (sve_el != 0 && fp_el == 0) { zcr_len = 0; } else { - ARMCPU *cpu = arm_env_get_cpu(env); - - zcr_len = cpu->sve_max_vq - 1; - if (current_el <= 1) { - zcr_len = MIN(zcr_len, 0xf & (uint32_t)env->vfp.zcr_el[1]); - } - if (current_el < 2 && arm_feature(env, ARM_FEATURE_EL2)) { - zcr_len = MIN(zcr_len, 0xf & (uint32_t)env->vfp.zcr_el[2]); - } - if (current_el < 3 && arm_feature(env, ARM_FEATURE_EL3)) { - zcr_len = MIN(zcr_len, 0xf & (uint32_t)env->vfp.zcr_el[3]); - } + zcr_len = sve_zcr_len_for_el(env, current_el); } flags |= sve_el << ARM_TBFLAG_SVEEXC_EL_SHIFT; flags |= zcr_len << ARM_TBFLAG_ZCR_LEN_SHIFT; @@ -XXX,XX +XXX,XX @@ void cpu_get_tb_cpu_state(CPUARMState *env, target_ulong *pc, *pflags = flags; *cs_base = 0; } + +#ifdef TARGET_AARCH64 +/* + * The manual says that when SVE is enabled and VQ is widened the + * implementation is allowed to zero the previously inaccessible + * portion of the registers. The corollary to that is that when + * SVE is enabled and VQ is narrowed we are also allowed to zero + * the now inaccessible portion of the registers. + * + * The intent of this is that no predicate bit beyond VQ is ever set. + * Which means that some operations on predicate registers themselves + * may operate on full uint64_t or even unrolled across the maximum + * uint64_t[4]. Performing 4 bits of host arithmetic unconditionally + * may well be cheaper than conditionals to restrict the operation + * to the relevant portion of a uint16_t[16]. + */ +void aarch64_sve_narrow_vq(CPUARMState *env, unsigned vq) +{ + int i, j; + uint64_t pmask; + + assert(vq >= 1 && vq <= ARM_MAX_VQ); + assert(vq <= arm_env_get_cpu(env)->sve_max_vq); + + /* Zap the high bits of the zregs. */ + for (i = 0; i < 32; i++) { + memset(&env->vfp.zregs[i].d[2 * vq], 0, 16 * (ARM_MAX_VQ - vq)); + } + + /* Zap the high bits of the pregs and ffr. */ + pmask = 0; + if (vq & 3) { + pmask = ~(-1ULL << (16 * (vq & 3))); + } + for (j = vq / 4; j < ARM_MAX_VQ / 4; j++) { + for (i = 0; i < 17; ++i) { + env->vfp.pregs[i].p[j] &= pmask; + } + pmask = 0; + } +} + +/* + * Notice a change in SVE vector size when changing EL. + */ +void aarch64_sve_change_el(CPUARMState *env, int old_el, int new_el) +{ + int old_len, new_len; + + /* Nothing to do if no SVE. */ + if (!arm_feature(env, ARM_FEATURE_SVE)) { + return; + } + + /* Nothing to do if FP is disabled in either EL. */ + if (fp_exception_el(env, old_el) || fp_exception_el(env, new_el)) { + return; + } + + /* + * DDI0584A.d sec 3.2: "If SVE instructions are disabled or trapped + * at ELx, or not available because the EL is in AArch32 state, then + * for all purposes other than a direct read, the ZCR_ELx.LEN field + * has an effective value of 0". + * + * Consider EL2 (aa64, vq=4) -> EL0 (aa32) -> EL1 (aa64, vq=0). + * If we ignore aa32 state, we would fail to see the vq4->vq0 transition + * from EL2->EL1. Thus we go ahead and narrow when entering aa32 so that + * we already have the correct register contents when encountering the + * vq0->vq0 transition between EL0->EL1. + */ + old_len = (arm_el_is_aa64(env, old_el) && !sve_exception_el(env, old_el) + ? sve_zcr_len_for_el(env, old_el) : 0); + new_len = (arm_el_is_aa64(env, new_el) && !sve_exception_el(env, new_el) + ? sve_zcr_len_for_el(env, new_el) : 0); + + /* When changing vector length, clear inaccessible state. */ + if (new_len < old_len) { + aarch64_sve_narrow_vq(env, new_len + 1); + } +} +#endif diff --git a/target/arm/op_helper.c b/target/arm/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/op_helper.c +++ b/target/arm/op_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(exception_return)(CPUARMState *env) "AArch64 EL%d PC 0x%" PRIx64 "\n", cur_el, new_el, env->pc); } + aarch64_sve_change_el(env, cur_el, new_el); qemu_mutex_lock_iothread(); arm_call_el_change_hook(arm_env_get_cpu(env)); -- 2.19.0 From: Richard Henderson <richard.henderson@linaro.org> Use the existing helpers to determine if (1) the fpu is enabled, (2) sve state is enabled, and (3) the current sve vector length. Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181005175350.30752-6-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/cpu.h | 4 ++++ target/arm/helper.c | 6 +++--- target/arm/translate-a64.c | 8 ++++++-- 3 files changed, 13 insertions(+), 5 deletions(-) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ target_ulong do_arm_semihosting(CPUARMState *env); void aarch64_sync_32_to_64(CPUARMState *env); void aarch64_sync_64_to_32(CPUARMState *env); +int fp_exception_el(CPUARMState *env, int cur_el); +int sve_exception_el(CPUARMState *env, int cur_el); +uint32_t sve_zcr_len_for_el(CPUARMState *env, int el); + static inline bool is_a64(CPUARMState *env) { return env->aarch64; diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo debug_lpae_cp_reginfo[] = { * take care of raising that exception. * C.f. the ARM pseudocode function CheckSVEEnabled. */ -static int sve_exception_el(CPUARMState *env, int el) +int sve_exception_el(CPUARMState *env, int el) { #ifndef CONFIG_USER_ONLY if (el <= 1) { @@ -XXX,XX +XXX,XX @@ static int sve_exception_el(CPUARMState *env, int el) /* * Given that SVE is enabled, return the vector length for EL. */ -static uint32_t sve_zcr_len_for_el(CPUARMState *env, int el) +uint32_t sve_zcr_len_for_el(CPUARMState *env, int el) { ARMCPU *cpu = arm_env_get_cpu(env); uint32_t zcr_len = cpu->sve_max_vq - 1; @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(crc32c)(uint32_t acc, uint32_t val, uint32_t bytes) /* Return the exception level to which FP-disabled exceptions should * be taken, or 0 if FP is enabled. */ -static int fp_exception_el(CPUARMState *env, int cur_el) +int fp_exception_el(CPUARMState *env, int cur_el) { #ifndef CONFIG_USER_ONLY int fpen; diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ void aarch64_cpu_dump_state(CPUState *cs, FILE *f, cpu_fprintf(f, "\n"); return; } + if (fp_exception_el(env, el) != 0) { + cpu_fprintf(f, " FPU disabled\n"); + return; + } cpu_fprintf(f, " FPCR=%08x FPSR=%08x\n", vfp_get_fpcr(env), vfp_get_fpsr(env)); - if (arm_feature(env, ARM_FEATURE_SVE)) { - int j, zcr_len = env->vfp.zcr_el[1] & 0xf; /* fix for system mode */ + if (arm_feature(env, ARM_FEATURE_SVE) && sve_exception_el(env, el) == 0) { + int j, zcr_len = sve_zcr_len_for_el(env, el); for (i = 0; i <= FFR_PRED_NUM; i++) { bool eol; -- 2.19.0 From: Richard Henderson <richard.henderson@linaro.org> The 16-byte load only uses 16 predicate bits. But while reusing the other load infrastructure, we find other bits that are set and trigger an assert. To avoid this and retain the assert, zero-extend the predicate that we pass to the LD1 helper. Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com> Reported-by: Laurent Desnogues <laurent.desnogues@gmail.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181005175350.30752-7-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 25 +++++++++++++++++++++++-- 1 file changed, 23 insertions(+), 2 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static void do_ldrq(DisasContext *s, int zt, int pg, TCGv_i64 addr, int msz) unsigned vsz = vec_full_reg_size(s); TCGv_ptr t_pg; TCGv_i32 desc; + int poff; /* Load the first quadword using the normal predicated load helpers. */ desc = tcg_const_i32(simd_desc(16, 16, zt)); - t_pg = tcg_temp_new_ptr(); - tcg_gen_addi_ptr(t_pg, cpu_env, pred_full_reg_offset(s, pg)); + poff = pred_full_reg_offset(s, pg); + if (vsz > 16) { + /* + * Zero-extend the first 16 bits of the predicate into a temporary. + * This avoids triggering an assert making sure we don't have bits + * set within a predicate beyond VQ, but we have lowered VQ to 1 + * for this load operation. + */ + TCGv_i64 tmp = tcg_temp_new_i64(); +#ifdef HOST_WORDS_BIGENDIAN + poff += 6; +#endif + tcg_gen_ld16u_i64(tmp, cpu_env, poff); + + poff = offsetof(CPUARMState, vfp.preg_tmp); + tcg_gen_st_i64(tmp, cpu_env, poff); + tcg_temp_free_i64(tmp); + } + + t_pg = tcg_temp_new_ptr(); + tcg_gen_addi_ptr(t_pg, cpu_env, poff); + fns[msz](cpu_env, t_pg, addr, desc); tcg_temp_free_ptr(t_pg); -- 2.19.0 From: Richard Henderson <richard.henderson@linaro.org> Uses tlb_vaddr_to_host for correct operation with softmmu. Optimize for accesses within a single page or pair of pages. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181005175350.30752-8-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sve_helper.c | 731 +++++++++++++++++++++++++++++++--------- 1 file changed, 569 insertions(+), 162 deletions(-) diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ static void swap_memmove(void *vd, void *vs, size_t n) } } +/* Similarly for memset of 0. */ +static void swap_memzero(void *vd, size_t n) +{ + uintptr_t d = (uintptr_t)vd; + uintptr_t o = (d | n) & 7; + size_t i; + + /* Usually, the first bit of a predicate is set, so N is 0. */ + if (likely(n == 0)) { + return; + } + +#ifndef HOST_WORDS_BIGENDIAN + o = 0; +#endif + switch (o) { + case 0: + memset(vd, 0, n); + break; + + case 4: + for (i = 0; i < n; i += 4) { + *(uint32_t *)H1_4(d + i) = 0; + } + break; + + case 2: + case 6: + for (i = 0; i < n; i += 2) { + *(uint16_t *)H1_2(d + i) = 0; + } + break; + + default: + for (i = 0; i < n; i++) { + *(uint8_t *)H1(d + i) = 0; + } + break; + } +} + void HELPER(sve_ext)(void *vd, void *vn, void *vm, uint32_t desc) { intptr_t opr_sz = simd_oprsz(desc); @@ -XXX,XX +XXX,XX @@ void HELPER(sve_fcmla_zpzzz_d)(CPUARMState *env, void *vg, uint32_t desc) /* * Load contiguous data, protected by a governing predicate. */ -#define DO_LD1(NAME, FN, TYPEE, TYPEM, H) \ -static void do_##NAME(CPUARMState *env, void *vd, void *vg, \ - target_ulong addr, intptr_t oprsz, \ - uintptr_t ra) \ -{ \ - intptr_t i = 0; \ - do { \ - uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); \ - do { \ - TYPEM m = 0; \ - if (pg & 1) { \ - m = FN(env, addr, ra); \ - } \ - *(TYPEE *)(vd + H(i)) = m; \ - i += sizeof(TYPEE), pg >>= sizeof(TYPEE); \ - addr += sizeof(TYPEM); \ - } while (i & 15); \ - } while (i < oprsz); \ -} \ -void HELPER(NAME)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - do_##NAME(env, &env->vfp.zregs[simd_data(desc)], vg, \ - addr, simd_oprsz(desc), GETPC()); \ + +/* + * Load elements into @vd, controlled by @vg, from @host + @mem_ofs. + * Memory is valid through @host + @mem_max. The register element + * indicies are inferred from @mem_ofs, as modified by the types for + * which the helper is built. Return the @mem_ofs of the first element + * not loaded (which is @mem_max if they are all loaded). + * + * For softmmu, we have fully validated the guest page. For user-only, + * we cannot fully validate without taking the mmap lock, but since we + * know the access is within one host page, if any access is valid they + * all must be valid. However, when @vg is all false, it may be that + * no access is valid. + */ +typedef intptr_t sve_ld1_host_fn(void *vd, void *vg, void *host, + intptr_t mem_ofs, intptr_t mem_max); + +/* + * Load one element into @vd + @reg_off from (@env, @vaddr, @ra). + * The controlling predicate is known to be true. + */ +typedef void sve_ld1_tlb_fn(CPUARMState *env, void *vd, intptr_t reg_off, + target_ulong vaddr, int mmu_idx, uintptr_t ra); + +/* + * Generate the above primitives. + */ + +#define DO_LD_HOST(NAME, H, TYPEE, TYPEM, HOST) \ +static intptr_t sve_##NAME##_host(void *vd, void *vg, void *host, \ + intptr_t mem_off, const intptr_t mem_max) \ +{ \ + intptr_t reg_off = mem_off * (sizeof(TYPEE) / sizeof(TYPEM)); \ + uint64_t *pg = vg; \ + while (mem_off + sizeof(TYPEM) <= mem_max) { \ + TYPEM val = 0; \ + if (likely((pg[reg_off >> 6] >> (reg_off & 63)) & 1)) { \ + val = HOST(host + mem_off); \ + } \ + *(TYPEE *)(vd + H(reg_off)) = val; \ + mem_off += sizeof(TYPEM), reg_off += sizeof(TYPEE); \ + } \ + return mem_off; \ } +#ifdef CONFIG_SOFTMMU +#define DO_LD_TLB(NAME, H, TYPEE, TYPEM, HOST, MOEND, TLB) \ +static void sve_##NAME##_tlb(CPUARMState *env, void *vd, intptr_t reg_off, \ + target_ulong addr, int mmu_idx, uintptr_t ra) \ +{ \ + TCGMemOpIdx oi = make_memop_idx(ctz32(sizeof(TYPEM)) | MOEND, mmu_idx); \ + TYPEM val = TLB(env, addr, oi, ra); \ + *(TYPEE *)(vd + H(reg_off)) = val; \ +} +#else +#define DO_LD_TLB(NAME, H, TYPEE, TYPEM, HOST, MOEND, TLB) \ +static void sve_##NAME##_tlb(CPUARMState *env, void *vd, intptr_t reg_off, \ + target_ulong addr, int mmu_idx, uintptr_t ra) \ +{ \ + TYPEM val = HOST(g2h(addr)); \ + *(TYPEE *)(vd + H(reg_off)) = val; \ +} +#endif + +#define DO_LD_PRIM_1(NAME, H, TE, TM) \ + DO_LD_HOST(NAME, H, TE, TM, ldub_p) \ + DO_LD_TLB(NAME, H, TE, TM, ldub_p, 0, helper_ret_ldub_mmu) + +DO_LD_PRIM_1(ld1bb, H1, uint8_t, uint8_t) +DO_LD_PRIM_1(ld1bhu, H1_2, uint16_t, uint8_t) +DO_LD_PRIM_1(ld1bhs, H1_2, uint16_t, int8_t) +DO_LD_PRIM_1(ld1bsu, H1_4, uint32_t, uint8_t) +DO_LD_PRIM_1(ld1bss, H1_4, uint32_t, int8_t) +DO_LD_PRIM_1(ld1bdu, , uint64_t, uint8_t) +DO_LD_PRIM_1(ld1bds, , uint64_t, int8_t) + +#define DO_LD_PRIM_2(NAME, end, MOEND, H, TE, TM, PH, PT) \ + DO_LD_HOST(NAME##_##end, H, TE, TM, PH##_##end##_p) \ + DO_LD_TLB(NAME##_##end, H, TE, TM, PH##_##end##_p, \ + MOEND, helper_##end##_##PT##_mmu) + +DO_LD_PRIM_2(ld1hh, le, MO_LE, H1_2, uint16_t, uint16_t, lduw, lduw) +DO_LD_PRIM_2(ld1hsu, le, MO_LE, H1_4, uint32_t, uint16_t, lduw, lduw) +DO_LD_PRIM_2(ld1hss, le, MO_LE, H1_4, uint32_t, int16_t, lduw, lduw) +DO_LD_PRIM_2(ld1hdu, le, MO_LE, , uint64_t, uint16_t, lduw, lduw) +DO_LD_PRIM_2(ld1hds, le, MO_LE, , uint64_t, int16_t, lduw, lduw) + +DO_LD_PRIM_2(ld1ss, le, MO_LE, H1_4, uint32_t, uint32_t, ldl, ldul) +DO_LD_PRIM_2(ld1sdu, le, MO_LE, , uint64_t, uint32_t, ldl, ldul) +DO_LD_PRIM_2(ld1sds, le, MO_LE, , uint64_t, int32_t, ldl, ldul) + +DO_LD_PRIM_2(ld1dd, le, MO_LE, , uint64_t, uint64_t, ldq, ldq) + +DO_LD_PRIM_2(ld1hh, be, MO_BE, H1_2, uint16_t, uint16_t, lduw, lduw) +DO_LD_PRIM_2(ld1hsu, be, MO_BE, H1_4, uint32_t, uint16_t, lduw, lduw) +DO_LD_PRIM_2(ld1hss, be, MO_BE, H1_4, uint32_t, int16_t, lduw, lduw) +DO_LD_PRIM_2(ld1hdu, be, MO_BE, , uint64_t, uint16_t, lduw, lduw) +DO_LD_PRIM_2(ld1hds, be, MO_BE, , uint64_t, int16_t, lduw, lduw) + +DO_LD_PRIM_2(ld1ss, be, MO_BE, H1_4, uint32_t, uint32_t, ldl, ldul) +DO_LD_PRIM_2(ld1sdu, be, MO_BE, , uint64_t, uint32_t, ldl, ldul) +DO_LD_PRIM_2(ld1sds, be, MO_BE, , uint64_t, int32_t, ldl, ldul) + +DO_LD_PRIM_2(ld1dd, be, MO_BE, , uint64_t, uint64_t, ldq, ldq) + +#undef DO_LD_TLB +#undef DO_LD_HOST +#undef DO_LD_PRIM_1 +#undef DO_LD_PRIM_2 + +/* + * Skip through a sequence of inactive elements in the guarding predicate @vg, + * beginning at @reg_off bounded by @reg_max. Return the offset of the active + * element >= @reg_off, or @reg_max if there were no active elements at all. + */ +static intptr_t find_next_active(uint64_t *vg, intptr_t reg_off, + intptr_t reg_max, int esz) +{ + uint64_t pg_mask = pred_esz_masks[esz]; + uint64_t pg = (vg[reg_off >> 6] & pg_mask) >> (reg_off & 63); + + /* In normal usage, the first element is active. */ + if (likely(pg & 1)) { + return reg_off; + } + + if (pg == 0) { + reg_off &= -64; + do { + reg_off += 64; + if (unlikely(reg_off >= reg_max)) { + /* The entire predicate was false. */ + return reg_max; + } + pg = vg[reg_off >> 6] & pg_mask; + } while (pg == 0); + } + reg_off += ctz64(pg); + + /* We should never see an out of range predicate bit set. */ + tcg_debug_assert(reg_off < reg_max); + return reg_off; +} + +/* + * Return the maximum offset <= @mem_max which is still within the page + * referenced by @base + @mem_off. + */ +static intptr_t max_for_page(target_ulong base, intptr_t mem_off, + intptr_t mem_max) +{ + target_ulong addr = base + mem_off; + intptr_t split = -(intptr_t)(addr | TARGET_PAGE_MASK); + return MIN(split, mem_max - mem_off) + mem_off; +} + +static inline void set_helper_retaddr(uintptr_t ra) +{ +#ifdef CONFIG_USER_ONLY + helper_retaddr = ra; +#endif +} + +/* + * The result of tlb_vaddr_to_host for user-only is just g2h(x), + * which is always non-null. Elide the useless test. + */ +static inline bool test_host_page(void *host) +{ +#ifdef CONFIG_USER_ONLY + return true; +#else + return likely(host != NULL); +#endif +} + +/* + * Common helper for all contiguous one-register predicated loads. + */ +static void sve_ld1_r(CPUARMState *env, void *vg, const target_ulong addr, + uint32_t desc, const uintptr_t retaddr, + const int esz, const int msz, + sve_ld1_host_fn *host_fn, + sve_ld1_tlb_fn *tlb_fn) +{ + void *vd = &env->vfp.zregs[simd_data(desc)]; + const int diffsz = esz - msz; + const intptr_t reg_max = simd_oprsz(desc); + const intptr_t mem_max = reg_max >> diffsz; + const int mmu_idx = cpu_mmu_index(env, false); + ARMVectorReg scratch; + void *host; + intptr_t split, reg_off, mem_off; + + /* Find the first active element. */ + reg_off = find_next_active(vg, 0, reg_max, esz); + if (unlikely(reg_off == reg_max)) { + /* The entire predicate was false; no load occurs. */ + memset(vd, 0, reg_max); + return; + } + mem_off = reg_off >> diffsz; + set_helper_retaddr(retaddr); + + /* + * If the (remaining) load is entirely within a single page, then: + * For softmmu, and the tlb hits, then no faults will occur; + * For user-only, either the first load will fault or none will. + * We can thus perform the load directly to the destination and + * Vd will be unmodified on any exception path. + */ + split = max_for_page(addr, mem_off, mem_max); + if (likely(split == mem_max)) { + host = tlb_vaddr_to_host(env, addr + mem_off, MMU_DATA_LOAD, mmu_idx); + if (test_host_page(host)) { + mem_off = host_fn(vd, vg, host - mem_off, mem_off, mem_max); + tcg_debug_assert(mem_off == mem_max); + set_helper_retaddr(0); + /* After having taken any fault, zero leading inactive elements. */ + swap_memzero(vd, reg_off); + return; + } + } + + /* + * Perform the predicated read into a temporary, thus ensuring + * if the load of the last element faults, Vd is not modified. + */ +#ifdef CONFIG_USER_ONLY + swap_memzero(&scratch, reg_off); + host_fn(&scratch, vg, g2h(addr), mem_off, mem_max); +#else + memset(&scratch, 0, reg_max); + goto start; + while (1) { + reg_off = find_next_active(vg, reg_off, reg_max, esz); + if (reg_off >= reg_max) { + break; + } + mem_off = reg_off >> diffsz; + split = max_for_page(addr, mem_off, mem_max); + + start: + if (split - mem_off >= (1 << msz)) { + /* At least one whole element on this page. */ + host = tlb_vaddr_to_host(env, addr + mem_off, + MMU_DATA_LOAD, mmu_idx); + if (host) { + mem_off = host_fn(&scratch, vg, host - mem_off, + mem_off, split); + reg_off = mem_off << diffsz; + continue; + } + } + + /* + * Perform one normal read. This may fault, longjmping out to the + * main loop in order to raise an exception. It may succeed, and + * as a side-effect load the TLB entry for the next round. Finally, + * in the extremely unlikely case we're performing this operation + * on I/O memory, it may succeed but not bring in the TLB entry. + * But even then we have still made forward progress. + */ + tlb_fn(env, &scratch, reg_off, addr + mem_off, mmu_idx, retaddr); + reg_off += 1 << esz; + } +#endif + + set_helper_retaddr(0); + memcpy(vd, &scratch, reg_max); +} + +#define DO_LD1_1(NAME, ESZ) \ +void HELPER(sve_##NAME##_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ld1_r(env, vg, addr, desc, GETPC(), ESZ, 0, \ + sve_##NAME##_host, sve_##NAME##_tlb); \ +} + +/* TODO: Propagate the endian check back to the translator. */ +#define DO_LD1_2(NAME, ESZ, MSZ) \ +void HELPER(sve_##NAME##_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + if (arm_cpu_data_is_big_endian(env)) { \ + sve_ld1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, \ + sve_##NAME##_be_host, sve_##NAME##_be_tlb); \ + } else { \ + sve_ld1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, \ + sve_##NAME##_le_host, sve_##NAME##_le_tlb); \ + } \ +} + +DO_LD1_1(ld1bb, 0) +DO_LD1_1(ld1bhu, 1) +DO_LD1_1(ld1bhs, 1) +DO_LD1_1(ld1bsu, 2) +DO_LD1_1(ld1bss, 2) +DO_LD1_1(ld1bdu, 3) +DO_LD1_1(ld1bds, 3) + +DO_LD1_2(ld1hh, 1, 1) +DO_LD1_2(ld1hsu, 2, 1) +DO_LD1_2(ld1hss, 2, 1) +DO_LD1_2(ld1hdu, 3, 1) +DO_LD1_2(ld1hds, 3, 1) + +DO_LD1_2(ld1ss, 2, 2) +DO_LD1_2(ld1sdu, 3, 2) +DO_LD1_2(ld1sds, 3, 2) + +DO_LD1_2(ld1dd, 3, 3) + +#undef DO_LD1_1 +#undef DO_LD1_2 + #define DO_LD2(NAME, FN, TYPEE, TYPEM, H) \ void HELPER(NAME)(CPUARMState *env, void *vg, \ target_ulong addr, uint32_t desc) \ @@ -XXX,XX +XXX,XX @@ void HELPER(NAME)(CPUARMState *env, void *vg, \ } \ } -DO_LD1(sve_ld1bhu_r, cpu_ldub_data_ra, uint16_t, uint8_t, H1_2) -DO_LD1(sve_ld1bhs_r, cpu_ldsb_data_ra, uint16_t, int8_t, H1_2) -DO_LD1(sve_ld1bsu_r, cpu_ldub_data_ra, uint32_t, uint8_t, H1_4) -DO_LD1(sve_ld1bss_r, cpu_ldsb_data_ra, uint32_t, int8_t, H1_4) -DO_LD1(sve_ld1bdu_r, cpu_ldub_data_ra, uint64_t, uint8_t, ) -DO_LD1(sve_ld1bds_r, cpu_ldsb_data_ra, uint64_t, int8_t, ) - -DO_LD1(sve_ld1hsu_r, cpu_lduw_data_ra, uint32_t, uint16_t, H1_4) -DO_LD1(sve_ld1hss_r, cpu_ldsw_data_ra, uint32_t, int16_t, H1_4) -DO_LD1(sve_ld1hdu_r, cpu_lduw_data_ra, uint64_t, uint16_t, ) -DO_LD1(sve_ld1hds_r, cpu_ldsw_data_ra, uint64_t, int16_t, ) - -DO_LD1(sve_ld1sdu_r, cpu_ldl_data_ra, uint64_t, uint32_t, ) -DO_LD1(sve_ld1sds_r, cpu_ldl_data_ra, uint64_t, int32_t, ) - -DO_LD1(sve_ld1bb_r, cpu_ldub_data_ra, uint8_t, uint8_t, H1) DO_LD2(sve_ld2bb_r, cpu_ldub_data_ra, uint8_t, uint8_t, H1) DO_LD3(sve_ld3bb_r, cpu_ldub_data_ra, uint8_t, uint8_t, H1) DO_LD4(sve_ld4bb_r, cpu_ldub_data_ra, uint8_t, uint8_t, H1) -DO_LD1(sve_ld1hh_r, cpu_lduw_data_ra, uint16_t, uint16_t, H1_2) DO_LD2(sve_ld2hh_r, cpu_lduw_data_ra, uint16_t, uint16_t, H1_2) DO_LD3(sve_ld3hh_r, cpu_lduw_data_ra, uint16_t, uint16_t, H1_2) DO_LD4(sve_ld4hh_r, cpu_lduw_data_ra, uint16_t, uint16_t, H1_2) -DO_LD1(sve_ld1ss_r, cpu_ldl_data_ra, uint32_t, uint32_t, H1_4) DO_LD2(sve_ld2ss_r, cpu_ldl_data_ra, uint32_t, uint32_t, H1_4) DO_LD3(sve_ld3ss_r, cpu_ldl_data_ra, uint32_t, uint32_t, H1_4) DO_LD4(sve_ld4ss_r, cpu_ldl_data_ra, uint32_t, uint32_t, H1_4) -DO_LD1(sve_ld1dd_r, cpu_ldq_data_ra, uint64_t, uint64_t, ) DO_LD2(sve_ld2dd_r, cpu_ldq_data_ra, uint64_t, uint64_t, ) DO_LD3(sve_ld3dd_r, cpu_ldq_data_ra, uint64_t, uint64_t, ) DO_LD4(sve_ld4dd_r, cpu_ldq_data_ra, uint64_t, uint64_t, ) -#undef DO_LD1 #undef DO_LD2 #undef DO_LD3 #undef DO_LD4 /* * Load contiguous data, first-fault and no-fault. + * + * For user-only, one could argue that we should hold the mmap_lock during + * the operation so that there is no race between page_check_range and the + * load operation. However, unmapping pages out from under a running thread + * is extraordinarily unlikely. This theoretical race condition also affects + * linux-user/ in its get_user/put_user macros. + * + * TODO: Construct some helpers, written in assembly, that interact with + * handle_cpu_signal to produce memory ops which can properly report errors + * without racing. */ -#ifdef CONFIG_USER_ONLY - /* Fault on byte I. All bits in FFR from I are cleared. The vector * result from I is CONSTRAINED UNPREDICTABLE; we choose the MERGE * option, which leaves subsequent data unchanged. @@ -XXX,XX +XXX,XX @@ static void record_fault(CPUARMState *env, uintptr_t i, uintptr_t oprsz) } } -/* Hold the mmap lock during the operation so that there is no race - * between page_check_range and the load operation. We expect the - * usual case to have no faults at all, so we check the whole range - * first and if successful defer to the normal load operation. - * - * TODO: Change mmap_lock to a rwlock so that multiple readers - * can run simultaneously. This will probably help other uses - * within QEMU as well. +/* + * Common helper for all contiguous first-fault loads. */ -#define DO_LDFF1(PART, FN, TYPEE, TYPEM, H) \ -static void do_sve_ldff1##PART(CPUARMState *env, void *vd, void *vg, \ - target_ulong addr, intptr_t oprsz, \ - bool first, uintptr_t ra) \ -{ \ - intptr_t i = 0; \ - do { \ - uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); \ - do { \ - TYPEM m = 0; \ - if (pg & 1) { \ - if (!first && \ - unlikely(page_check_range(addr, sizeof(TYPEM), \ - PAGE_READ))) { \ - record_fault(env, i, oprsz); \ - return; \ - } \ - m = FN(env, addr, ra); \ - first = false; \ - } \ - *(TYPEE *)(vd + H(i)) = m; \ - i += sizeof(TYPEE), pg >>= sizeof(TYPEE); \ - addr += sizeof(TYPEM); \ - } while (i & 15); \ - } while (i < oprsz); \ -} \ -void HELPER(sve_ldff1##PART)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - intptr_t oprsz = simd_oprsz(desc); \ - unsigned rd = simd_data(desc); \ - void *vd = &env->vfp.zregs[rd]; \ - mmap_lock(); \ - if (likely(page_check_range(addr, oprsz, PAGE_READ) == 0)) { \ - do_sve_ld1##PART(env, vd, vg, addr, oprsz, GETPC()); \ - } else { \ - do_sve_ldff1##PART(env, vd, vg, addr, oprsz, true, GETPC()); \ - } \ - mmap_unlock(); \ -} +static void sve_ldff1_r(CPUARMState *env, void *vg, const target_ulong addr, + uint32_t desc, const uintptr_t retaddr, + const int esz, const int msz, + sve_ld1_host_fn *host_fn, + sve_ld1_tlb_fn *tlb_fn) +{ + void *vd = &env->vfp.zregs[simd_data(desc)]; + const int diffsz = esz - msz; + const intptr_t reg_max = simd_oprsz(desc); + const intptr_t mem_max = reg_max >> diffsz; + const int mmu_idx = cpu_mmu_index(env, false); + intptr_t split, reg_off, mem_off; + void *host; -/* No-fault loads are like first-fault loads without the - * first faulting special case. - */ -#define DO_LDNF1(PART) \ -void HELPER(sve_ldnf1##PART)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - intptr_t oprsz = simd_oprsz(desc); \ - unsigned rd = simd_data(desc); \ - void *vd = &env->vfp.zregs[rd]; \ - mmap_lock(); \ - if (likely(page_check_range(addr, oprsz, PAGE_READ) == 0)) { \ - do_sve_ld1##PART(env, vd, vg, addr, oprsz, GETPC()); \ - } else { \ - do_sve_ldff1##PART(env, vd, vg, addr, oprsz, false, GETPC()); \ - } \ - mmap_unlock(); \ -} + /* Skip to the first active element. */ + reg_off = find_next_active(vg, 0, reg_max, esz); + if (unlikely(reg_off == reg_max)) { + /* The entire predicate was false; no load occurs. */ + memset(vd, 0, reg_max); + return; + } + mem_off = reg_off >> diffsz; + set_helper_retaddr(retaddr); + /* + * If the (remaining) load is entirely within a single page, then: + * For softmmu, and the tlb hits, then no faults will occur; + * For user-only, either the first load will fault or none will. + * We can thus perform the load directly to the destination and + * Vd will be unmodified on any exception path. + */ + split = max_for_page(addr, mem_off, mem_max); + if (likely(split == mem_max)) { + host = tlb_vaddr_to_host(env, addr + mem_off, MMU_DATA_LOAD, mmu_idx); + if (test_host_page(host)) { + mem_off = host_fn(vd, vg, host - mem_off, mem_off, mem_max); + tcg_debug_assert(mem_off == mem_max); + set_helper_retaddr(0); + /* After any fault, zero any leading inactive elements. */ + swap_memzero(vd, reg_off); + return; + } + } + +#ifdef CONFIG_USER_ONLY + /* + * The page(s) containing this first element at ADDR+MEM_OFF must + * be valid. Considering that this first element may be misaligned + * and cross a page boundary itself, take the rest of the page from + * the last byte of the element. + */ + split = max_for_page(addr, mem_off + (1 << msz) - 1, mem_max); + mem_off = host_fn(vd, vg, g2h(addr), mem_off, split); + + /* After any fault, zero any leading inactive elements. */ + swap_memzero(vd, reg_off); + reg_off = mem_off << diffsz; #else + /* + * Perform one normal read, which will fault or not. + * But it is likely to bring the page into the tlb. + */ + tlb_fn(env, vd, reg_off, addr + mem_off, mmu_idx, retaddr); -/* TODO: System mode is not yet supported. - * This would probably use tlb_vaddr_to_host. - */ -#define DO_LDFF1(PART, FN, TYPEE, TYPEM, H) \ -void HELPER(sve_ldff1##PART)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - g_assert_not_reached(); \ -} - -#define DO_LDNF1(PART) \ -void HELPER(sve_ldnf1##PART)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - g_assert_not_reached(); \ -} + /* After any fault, zero any leading predicated false elts. */ + swap_memzero(vd, reg_off); + mem_off += 1 << msz; + reg_off += 1 << esz; + /* Try again to read the balance of the page. */ + split = max_for_page(addr, mem_off - 1, mem_max); + if (split >= (1 << msz)) { + host = tlb_vaddr_to_host(env, addr + mem_off, MMU_DATA_LOAD, mmu_idx); + if (host) { + mem_off = host_fn(vd, vg, host - mem_off, mem_off, split); + reg_off = mem_off << diffsz; + } + } #endif -DO_LDFF1(bb_r, cpu_ldub_data_ra, uint8_t, uint8_t, H1) -DO_LDFF1(bhu_r, cpu_ldub_data_ra, uint16_t, uint8_t, H1_2) -DO_LDFF1(bhs_r, cpu_ldsb_data_ra, uint16_t, int8_t, H1_2) -DO_LDFF1(bsu_r, cpu_ldub_data_ra, uint32_t, uint8_t, H1_4) -DO_LDFF1(bss_r, cpu_ldsb_data_ra, uint32_t, int8_t, H1_4) -DO_LDFF1(bdu_r, cpu_ldub_data_ra, uint64_t, uint8_t, ) -DO_LDFF1(bds_r, cpu_ldsb_data_ra, uint64_t, int8_t, ) + set_helper_retaddr(0); + record_fault(env, reg_off, reg_max); +} -DO_LDFF1(hh_r, cpu_lduw_data_ra, uint16_t, uint16_t, H1_2) -DO_LDFF1(hsu_r, cpu_lduw_data_ra, uint32_t, uint16_t, H1_4) -DO_LDFF1(hss_r, cpu_ldsw_data_ra, uint32_t, int8_t, H1_4) -DO_LDFF1(hdu_r, cpu_lduw_data_ra, uint64_t, uint16_t, ) -DO_LDFF1(hds_r, cpu_ldsw_data_ra, uint64_t, int16_t, ) +/* + * Common helper for all contiguous no-fault loads. + */ +static void sve_ldnf1_r(CPUARMState *env, void *vg, const target_ulong addr, + uint32_t desc, const int esz, const int msz, + sve_ld1_host_fn *host_fn) +{ + void *vd = &env->vfp.zregs[simd_data(desc)]; + const int diffsz = esz - msz; + const intptr_t reg_max = simd_oprsz(desc); + const intptr_t mem_max = reg_max >> diffsz; + const int mmu_idx = cpu_mmu_index(env, false); + intptr_t split, reg_off, mem_off; + void *host; -DO_LDFF1(ss_r, cpu_ldl_data_ra, uint32_t, uint32_t, H1_4) -DO_LDFF1(sdu_r, cpu_ldl_data_ra, uint64_t, uint32_t, ) -DO_LDFF1(sds_r, cpu_ldl_data_ra, uint64_t, int32_t, ) +#ifdef CONFIG_USER_ONLY + host = tlb_vaddr_to_host(env, addr, MMU_DATA_LOAD, mmu_idx); + if (likely(page_check_range(addr, mem_max, PAGE_READ) == 0)) { + /* The entire operation is valid and will not fault. */ + host_fn(vd, vg, host, 0, mem_max); + return; + } +#endif -DO_LDFF1(dd_r, cpu_ldq_data_ra, uint64_t, uint64_t, ) + /* There will be no fault, so we may modify in advance. */ + memset(vd, 0, reg_max); -#undef DO_LDFF1 + /* Skip to the first active element. */ + reg_off = find_next_active(vg, 0, reg_max, esz); + if (unlikely(reg_off == reg_max)) { + /* The entire predicate was false; no load occurs. */ + return; + } + mem_off = reg_off >> diffsz; -DO_LDNF1(bb_r) -DO_LDNF1(bhu_r) -DO_LDNF1(bhs_r) -DO_LDNF1(bsu_r) -DO_LDNF1(bss_r) -DO_LDNF1(bdu_r) -DO_LDNF1(bds_r) +#ifdef CONFIG_USER_ONLY + if (page_check_range(addr + mem_off, 1 << msz, PAGE_READ) == 0) { + /* At least one load is valid; take the rest of the page. */ + split = max_for_page(addr, mem_off + (1 << msz) - 1, mem_max); + mem_off = host_fn(vd, vg, host, mem_off, split); + reg_off = mem_off << diffsz; + } +#else + /* + * If the address is not in the TLB, we have no way to bring the + * entry into the TLB without also risking a fault. Note that + * the corollary is that we never load from an address not in RAM. + * + * This last is out of spec, in a weird corner case. + * Per the MemNF/MemSingleNF pseudocode, a NF load from Device memory + * must not actually hit the bus -- it returns UNKNOWN data instead. + * But if you map non-RAM with Normal memory attributes and do a NF + * load then it should access the bus. (Nobody ought actually do this + * in the real world, obviously.) + * + * Then there are the annoying special cases with watchpoints... + * + * TODO: Add a form of tlb_fill that does not raise an exception, + * with a form of tlb_vaddr_to_host and a set of loads to match. + * The non_fault_vaddr_to_host would handle everything, usually, + * and the loads would handle the iomem path for watchpoints. + */ + host = tlb_vaddr_to_host(env, addr + mem_off, MMU_DATA_LOAD, mmu_idx); + split = max_for_page(addr, mem_off, mem_max); + if (host && split >= (1 << msz)) { + mem_off = host_fn(vd, vg, host - mem_off, mem_off, split); + reg_off = mem_off << diffsz; + } +#endif -DO_LDNF1(hh_r) -DO_LDNF1(hsu_r) -DO_LDNF1(hss_r) -DO_LDNF1(hdu_r) -DO_LDNF1(hds_r) + record_fault(env, reg_off, reg_max); +} -DO_LDNF1(ss_r) -DO_LDNF1(sdu_r) -DO_LDNF1(sds_r) +#define DO_LDFF1_LDNF1_1(PART, ESZ) \ +void HELPER(sve_ldff1##PART##_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldff1_r(env, vg, addr, desc, GETPC(), ESZ, 0, \ + sve_ld1##PART##_host, sve_ld1##PART##_tlb); \ +} \ +void HELPER(sve_ldnf1##PART##_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldnf1_r(env, vg, addr, desc, ESZ, 0, sve_ld1##PART##_host); \ +} -DO_LDNF1(dd_r) +/* TODO: Propagate the endian check back to the translator. */ +#define DO_LDFF1_LDNF1_2(PART, ESZ, MSZ) \ +void HELPER(sve_ldff1##PART##_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + if (arm_cpu_data_is_big_endian(env)) { \ + sve_ldff1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, \ + sve_ld1##PART##_be_host, sve_ld1##PART##_be_tlb); \ + } else { \ + sve_ldff1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, \ + sve_ld1##PART##_le_host, sve_ld1##PART##_le_tlb); \ + } \ +} \ +void HELPER(sve_ldnf1##PART##_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + if (arm_cpu_data_is_big_endian(env)) { \ + sve_ldnf1_r(env, vg, addr, desc, ESZ, MSZ, \ + sve_ld1##PART##_be_host); \ + } else { \ + sve_ldnf1_r(env, vg, addr, desc, ESZ, MSZ, \ + sve_ld1##PART##_le_host); \ + } \ +} -#undef DO_LDNF1 +DO_LDFF1_LDNF1_1(bb, 0) +DO_LDFF1_LDNF1_1(bhu, 1) +DO_LDFF1_LDNF1_1(bhs, 1) +DO_LDFF1_LDNF1_1(bsu, 2) +DO_LDFF1_LDNF1_1(bss, 2) +DO_LDFF1_LDNF1_1(bdu, 3) +DO_LDFF1_LDNF1_1(bds, 3) + +DO_LDFF1_LDNF1_2(hh, 1, 1) +DO_LDFF1_LDNF1_2(hsu, 2, 1) +DO_LDFF1_LDNF1_2(hss, 2, 1) +DO_LDFF1_LDNF1_2(hdu, 3, 1) +DO_LDFF1_LDNF1_2(hds, 3, 1) + +DO_LDFF1_LDNF1_2(ss, 2, 2) +DO_LDFF1_LDNF1_2(sdu, 3, 2) +DO_LDFF1_LDNF1_2(sds, 3, 2) + +DO_LDFF1_LDNF1_2(dd, 3, 3) + +#undef DO_LDFF1_LDNF1_1 +#undef DO_LDFF1_LDNF1_2 /* * Store contiguous data, protected by a governing predicate. -- 2.19.0 From: Richard Henderson <richard.henderson@linaro.org> Use the same *_tlb primitives as we use for ld1. For linux-user, this hoists the set of helper_retaddr. For softmmu, hoists the computation of the current mmu_idx outside the loop, fixes the endianness problem, and moves the main loop out of a macro and into an inlined function. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181005175350.30752-9-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sve_helper.c | 210 ++++++++++++++++++++++------------------ 1 file changed, 117 insertions(+), 93 deletions(-) diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ DO_LD1_2(ld1dd, 3, 3) #undef DO_LD1_1 #undef DO_LD1_2 -#define DO_LD2(NAME, FN, TYPEE, TYPEM, H) \ -void HELPER(NAME)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - intptr_t i, oprsz = simd_oprsz(desc); \ - intptr_t ra = GETPC(); \ - unsigned rd = simd_data(desc); \ - void *d1 = &env->vfp.zregs[rd]; \ - void *d2 = &env->vfp.zregs[(rd + 1) & 31]; \ - for (i = 0; i < oprsz; ) { \ - uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); \ - do { \ - TYPEM m1 = 0, m2 = 0; \ - if (pg & 1) { \ - m1 = FN(env, addr, ra); \ - m2 = FN(env, addr + sizeof(TYPEM), ra); \ - } \ - *(TYPEE *)(d1 + H(i)) = m1; \ - *(TYPEE *)(d2 + H(i)) = m2; \ - i += sizeof(TYPEE), pg >>= sizeof(TYPEE); \ - addr += 2 * sizeof(TYPEM); \ - } while (i & 15); \ - } \ +/* + * Common helpers for all contiguous 2,3,4-register predicated loads. + */ +static void sve_ld2_r(CPUARMState *env, void *vg, target_ulong addr, + uint32_t desc, int size, uintptr_t ra, + sve_ld1_tlb_fn *tlb_fn) +{ + const int mmu_idx = cpu_mmu_index(env, false); + intptr_t i, oprsz = simd_oprsz(desc); + unsigned rd = simd_data(desc); + ARMVectorReg scratch[2] = { }; + + set_helper_retaddr(ra); + for (i = 0; i < oprsz; ) { + uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); + do { + if (pg & 1) { + tlb_fn(env, &scratch[0], i, addr, mmu_idx, ra); + tlb_fn(env, &scratch[1], i, addr + size, mmu_idx, ra); + } + i += size, pg >>= size; + addr += 2 * size; + } while (i & 15); + } + set_helper_retaddr(0); + + /* Wait until all exceptions have been raised to write back. */ + memcpy(&env->vfp.zregs[rd], &scratch[0], oprsz); + memcpy(&env->vfp.zregs[(rd + 1) & 31], &scratch[1], oprsz); } -#define DO_LD3(NAME, FN, TYPEE, TYPEM, H) \ -void HELPER(NAME)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - intptr_t i, oprsz = simd_oprsz(desc); \ - intptr_t ra = GETPC(); \ - unsigned rd = simd_data(desc); \ - void *d1 = &env->vfp.zregs[rd]; \ - void *d2 = &env->vfp.zregs[(rd + 1) & 31]; \ - void *d3 = &env->vfp.zregs[(rd + 2) & 31]; \ - for (i = 0; i < oprsz; ) { \ - uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); \ - do { \ - TYPEM m1 = 0, m2 = 0, m3 = 0; \ - if (pg & 1) { \ - m1 = FN(env, addr, ra); \ - m2 = FN(env, addr + sizeof(TYPEM), ra); \ - m3 = FN(env, addr + 2 * sizeof(TYPEM), ra); \ - } \ - *(TYPEE *)(d1 + H(i)) = m1; \ - *(TYPEE *)(d2 + H(i)) = m2; \ - *(TYPEE *)(d3 + H(i)) = m3; \ - i += sizeof(TYPEE), pg >>= sizeof(TYPEE); \ - addr += 3 * sizeof(TYPEM); \ - } while (i & 15); \ - } \ +static void sve_ld3_r(CPUARMState *env, void *vg, target_ulong addr, + uint32_t desc, int size, uintptr_t ra, + sve_ld1_tlb_fn *tlb_fn) +{ + const int mmu_idx = cpu_mmu_index(env, false); + intptr_t i, oprsz = simd_oprsz(desc); + unsigned rd = simd_data(desc); + ARMVectorReg scratch[3] = { }; + + set_helper_retaddr(ra); + for (i = 0; i < oprsz; ) { + uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); + do { + if (pg & 1) { + tlb_fn(env, &scratch[0], i, addr, mmu_idx, ra); + tlb_fn(env, &scratch[1], i, addr + size, mmu_idx, ra); + tlb_fn(env, &scratch[2], i, addr + 2 * size, mmu_idx, ra); + } + i += size, pg >>= size; + addr += 3 * size; + } while (i & 15); + } + set_helper_retaddr(0); + + /* Wait until all exceptions have been raised to write back. */ + memcpy(&env->vfp.zregs[rd], &scratch[0], oprsz); + memcpy(&env->vfp.zregs[(rd + 1) & 31], &scratch[1], oprsz); + memcpy(&env->vfp.zregs[(rd + 2) & 31], &scratch[2], oprsz); } -#define DO_LD4(NAME, FN, TYPEE, TYPEM, H) \ -void HELPER(NAME)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - intptr_t i, oprsz = simd_oprsz(desc); \ - intptr_t ra = GETPC(); \ - unsigned rd = simd_data(desc); \ - void *d1 = &env->vfp.zregs[rd]; \ - void *d2 = &env->vfp.zregs[(rd + 1) & 31]; \ - void *d3 = &env->vfp.zregs[(rd + 2) & 31]; \ - void *d4 = &env->vfp.zregs[(rd + 3) & 31]; \ - for (i = 0; i < oprsz; ) { \ - uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); \ - do { \ - TYPEM m1 = 0, m2 = 0, m3 = 0, m4 = 0; \ - if (pg & 1) { \ - m1 = FN(env, addr, ra); \ - m2 = FN(env, addr + sizeof(TYPEM), ra); \ - m3 = FN(env, addr + 2 * sizeof(TYPEM), ra); \ - m4 = FN(env, addr + 3 * sizeof(TYPEM), ra); \ - } \ - *(TYPEE *)(d1 + H(i)) = m1; \ - *(TYPEE *)(d2 + H(i)) = m2; \ - *(TYPEE *)(d3 + H(i)) = m3; \ - *(TYPEE *)(d4 + H(i)) = m4; \ - i += sizeof(TYPEE), pg >>= sizeof(TYPEE); \ - addr += 4 * sizeof(TYPEM); \ - } while (i & 15); \ - } \ +static void sve_ld4_r(CPUARMState *env, void *vg, target_ulong addr, + uint32_t desc, int size, uintptr_t ra, + sve_ld1_tlb_fn *tlb_fn) +{ + const int mmu_idx = cpu_mmu_index(env, false); + intptr_t i, oprsz = simd_oprsz(desc); + unsigned rd = simd_data(desc); + ARMVectorReg scratch[4] = { }; + + set_helper_retaddr(ra); + for (i = 0; i < oprsz; ) { + uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); + do { + if (pg & 1) { + tlb_fn(env, &scratch[0], i, addr, mmu_idx, ra); + tlb_fn(env, &scratch[1], i, addr + size, mmu_idx, ra); + tlb_fn(env, &scratch[2], i, addr + 2 * size, mmu_idx, ra); + tlb_fn(env, &scratch[3], i, addr + 3 * size, mmu_idx, ra); + } + i += size, pg >>= size; + addr += 4 * size; + } while (i & 15); + } + set_helper_retaddr(0); + + /* Wait until all exceptions have been raised to write back. */ + memcpy(&env->vfp.zregs[rd], &scratch[0], oprsz); + memcpy(&env->vfp.zregs[(rd + 1) & 31], &scratch[1], oprsz); + memcpy(&env->vfp.zregs[(rd + 2) & 31], &scratch[2], oprsz); + memcpy(&env->vfp.zregs[(rd + 3) & 31], &scratch[3], oprsz); } -DO_LD2(sve_ld2bb_r, cpu_ldub_data_ra, uint8_t, uint8_t, H1) -DO_LD3(sve_ld3bb_r, cpu_ldub_data_ra, uint8_t, uint8_t, H1) -DO_LD4(sve_ld4bb_r, cpu_ldub_data_ra, uint8_t, uint8_t, H1) +#define DO_LDN_1(N) \ +void __attribute__((flatten)) HELPER(sve_ld##N##bb_r) \ + (CPUARMState *env, void *vg, target_ulong addr, uint32_t desc) \ +{ \ + sve_ld##N##_r(env, vg, addr, desc, 1, GETPC(), sve_ld1bb_tlb); \ +} -DO_LD2(sve_ld2hh_r, cpu_lduw_data_ra, uint16_t, uint16_t, H1_2) -DO_LD3(sve_ld3hh_r, cpu_lduw_data_ra, uint16_t, uint16_t, H1_2) -DO_LD4(sve_ld4hh_r, cpu_lduw_data_ra, uint16_t, uint16_t, H1_2) +#define DO_LDN_2(N, SUFF, SIZE) \ +void __attribute__((flatten)) HELPER(sve_ld##N##SUFF##_r) \ + (CPUARMState *env, void *vg, target_ulong addr, uint32_t desc) \ +{ \ + sve_ld##N##_r(env, vg, addr, desc, SIZE, GETPC(), \ + arm_cpu_data_is_big_endian(env) \ + ? sve_ld1##SUFF##_be_tlb : sve_ld1##SUFF##_le_tlb); \ +} -DO_LD2(sve_ld2ss_r, cpu_ldl_data_ra, uint32_t, uint32_t, H1_4) -DO_LD3(sve_ld3ss_r, cpu_ldl_data_ra, uint32_t, uint32_t, H1_4) -DO_LD4(sve_ld4ss_r, cpu_ldl_data_ra, uint32_t, uint32_t, H1_4) +DO_LDN_1(2) +DO_LDN_1(3) +DO_LDN_1(4) -DO_LD2(sve_ld2dd_r, cpu_ldq_data_ra, uint64_t, uint64_t, ) -DO_LD3(sve_ld3dd_r, cpu_ldq_data_ra, uint64_t, uint64_t, ) -DO_LD4(sve_ld4dd_r, cpu_ldq_data_ra, uint64_t, uint64_t, ) +DO_LDN_2(2, hh, 2) +DO_LDN_2(3, hh, 2) +DO_LDN_2(4, hh, 2) -#undef DO_LD2 -#undef DO_LD3 -#undef DO_LD4 +DO_LDN_2(2, ss, 4) +DO_LDN_2(3, ss, 4) +DO_LDN_2(4, ss, 4) + +DO_LDN_2(2, dd, 8) +DO_LDN_2(3, dd, 8) +DO_LDN_2(4, dd, 8) + +#undef DO_LDN_1 +#undef DO_LDN_2 /* * Load contiguous data, first-fault and no-fault. -- 2.19.0 From: Richard Henderson <richard.henderson@linaro.org> This fixes the endianness problem for softmmu, and moves the main loop out of a macro and into an inlined function. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181005175350.30752-10-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sve_helper.c | 351 ++++++++++++++++++++-------------------- 1 file changed, 172 insertions(+), 179 deletions(-) diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ typedef intptr_t sve_ld1_host_fn(void *vd, void *vg, void *host, */ typedef void sve_ld1_tlb_fn(CPUARMState *env, void *vd, intptr_t reg_off, target_ulong vaddr, int mmu_idx, uintptr_t ra); +typedef sve_ld1_tlb_fn sve_st1_tlb_fn; /* * Generate the above primitives. @@ -XXX,XX +XXX,XX @@ DO_LDFF1_LDNF1_2(dd, 3, 3) /* * Store contiguous data, protected by a governing predicate. */ -#define DO_ST1(NAME, FN, TYPEE, TYPEM, H) \ -void HELPER(NAME)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - intptr_t i, oprsz = simd_oprsz(desc); \ - intptr_t ra = GETPC(); \ - unsigned rd = simd_data(desc); \ - void *vd = &env->vfp.zregs[rd]; \ - for (i = 0; i < oprsz; ) { \ - uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); \ - do { \ - if (pg & 1) { \ - TYPEM m = *(TYPEE *)(vd + H(i)); \ - FN(env, addr, m, ra); \ - } \ - i += sizeof(TYPEE), pg >>= sizeof(TYPEE); \ - addr += sizeof(TYPEM); \ - } while (i & 15); \ - } \ + +#ifdef CONFIG_SOFTMMU +#define DO_ST_TLB(NAME, H, TYPEM, HOST, MOEND, TLB) \ +static void sve_##NAME##_tlb(CPUARMState *env, void *vd, intptr_t reg_off, \ + target_ulong addr, int mmu_idx, uintptr_t ra) \ +{ \ + TCGMemOpIdx oi = make_memop_idx(ctz32(sizeof(TYPEM)) | MOEND, mmu_idx); \ + TLB(env, addr, *(TYPEM *)(vd + H(reg_off)), oi, ra); \ } - -#define DO_ST1_D(NAME, FN, TYPEM) \ -void HELPER(NAME)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - intptr_t i, oprsz = simd_oprsz(desc) / 8; \ - intptr_t ra = GETPC(); \ - unsigned rd = simd_data(desc); \ - uint64_t *d = &env->vfp.zregs[rd].d[0]; \ - uint8_t *pg = vg; \ - for (i = 0; i < oprsz; i += 1) { \ - if (pg[H1(i)] & 1) { \ - FN(env, addr, d[i], ra); \ - } \ - addr += sizeof(TYPEM); \ - } \ +#else +#define DO_ST_TLB(NAME, H, TYPEM, HOST, MOEND, TLB) \ +static void sve_##NAME##_tlb(CPUARMState *env, void *vd, intptr_t reg_off, \ + target_ulong addr, int mmu_idx, uintptr_t ra) \ +{ \ + HOST(g2h(addr), *(TYPEM *)(vd + H(reg_off))); \ } +#endif -#define DO_ST2(NAME, FN, TYPEE, TYPEM, H) \ -void HELPER(NAME)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - intptr_t i, oprsz = simd_oprsz(desc); \ - intptr_t ra = GETPC(); \ - unsigned rd = simd_data(desc); \ - void *d1 = &env->vfp.zregs[rd]; \ - void *d2 = &env->vfp.zregs[(rd + 1) & 31]; \ - for (i = 0; i < oprsz; ) { \ - uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); \ - do { \ - if (pg & 1) { \ - TYPEM m1 = *(TYPEE *)(d1 + H(i)); \ - TYPEM m2 = *(TYPEE *)(d2 + H(i)); \ - FN(env, addr, m1, ra); \ - FN(env, addr + sizeof(TYPEM), m2, ra); \ - } \ - i += sizeof(TYPEE), pg >>= sizeof(TYPEE); \ - addr += 2 * sizeof(TYPEM); \ - } while (i & 15); \ - } \ -} +DO_ST_TLB(st1bb, H1, uint8_t, stb_p, 0, helper_ret_stb_mmu) +DO_ST_TLB(st1bh, H1_2, uint16_t, stb_p, 0, helper_ret_stb_mmu) +DO_ST_TLB(st1bs, H1_4, uint32_t, stb_p, 0, helper_ret_stb_mmu) +DO_ST_TLB(st1bd, , uint64_t, stb_p, 0, helper_ret_stb_mmu) -#define DO_ST3(NAME, FN, TYPEE, TYPEM, H) \ -void HELPER(NAME)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - intptr_t i, oprsz = simd_oprsz(desc); \ - intptr_t ra = GETPC(); \ - unsigned rd = simd_data(desc); \ - void *d1 = &env->vfp.zregs[rd]; \ - void *d2 = &env->vfp.zregs[(rd + 1) & 31]; \ - void *d3 = &env->vfp.zregs[(rd + 2) & 31]; \ - for (i = 0; i < oprsz; ) { \ - uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); \ - do { \ - if (pg & 1) { \ - TYPEM m1 = *(TYPEE *)(d1 + H(i)); \ - TYPEM m2 = *(TYPEE *)(d2 + H(i)); \ - TYPEM m3 = *(TYPEE *)(d3 + H(i)); \ - FN(env, addr, m1, ra); \ - FN(env, addr + sizeof(TYPEM), m2, ra); \ - FN(env, addr + 2 * sizeof(TYPEM), m3, ra); \ - } \ - i += sizeof(TYPEE), pg >>= sizeof(TYPEE); \ - addr += 3 * sizeof(TYPEM); \ - } while (i & 15); \ - } \ -} +DO_ST_TLB(st1hh_le, H1_2, uint16_t, stw_le_p, MO_LE, helper_le_stw_mmu) +DO_ST_TLB(st1hs_le, H1_4, uint32_t, stw_le_p, MO_LE, helper_le_stw_mmu) +DO_ST_TLB(st1hd_le, , uint64_t, stw_le_p, MO_LE, helper_le_stw_mmu) -#define DO_ST4(NAME, FN, TYPEE, TYPEM, H) \ -void HELPER(NAME)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - intptr_t i, oprsz = simd_oprsz(desc); \ - intptr_t ra = GETPC(); \ - unsigned rd = simd_data(desc); \ - void *d1 = &env->vfp.zregs[rd]; \ - void *d2 = &env->vfp.zregs[(rd + 1) & 31]; \ - void *d3 = &env->vfp.zregs[(rd + 2) & 31]; \ - void *d4 = &env->vfp.zregs[(rd + 3) & 31]; \ - for (i = 0; i < oprsz; ) { \ - uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); \ - do { \ - if (pg & 1) { \ - TYPEM m1 = *(TYPEE *)(d1 + H(i)); \ - TYPEM m2 = *(TYPEE *)(d2 + H(i)); \ - TYPEM m3 = *(TYPEE *)(d3 + H(i)); \ - TYPEM m4 = *(TYPEE *)(d4 + H(i)); \ - FN(env, addr, m1, ra); \ - FN(env, addr + sizeof(TYPEM), m2, ra); \ - FN(env, addr + 2 * sizeof(TYPEM), m3, ra); \ - FN(env, addr + 3 * sizeof(TYPEM), m4, ra); \ - } \ - i += sizeof(TYPEE), pg >>= sizeof(TYPEE); \ - addr += 4 * sizeof(TYPEM); \ - } while (i & 15); \ - } \ -} +DO_ST_TLB(st1ss_le, H1_4, uint32_t, stl_le_p, MO_LE, helper_le_stl_mmu) +DO_ST_TLB(st1sd_le, , uint64_t, stl_le_p, MO_LE, helper_le_stl_mmu) -DO_ST1(sve_st1bh_r, cpu_stb_data_ra, uint16_t, uint8_t, H1_2) -DO_ST1(sve_st1bs_r, cpu_stb_data_ra, uint32_t, uint8_t, H1_4) -DO_ST1_D(sve_st1bd_r, cpu_stb_data_ra, uint8_t) +DO_ST_TLB(st1dd_le, , uint64_t, stq_le_p, MO_LE, helper_le_stq_mmu) -DO_ST1(sve_st1hs_r, cpu_stw_data_ra, uint32_t, uint16_t, H1_4) -DO_ST1_D(sve_st1hd_r, cpu_stw_data_ra, uint16_t) +DO_ST_TLB(st1hh_be, H1_2, uint16_t, stw_be_p, MO_BE, helper_be_stw_mmu) +DO_ST_TLB(st1hs_be, H1_4, uint32_t, stw_be_p, MO_BE, helper_be_stw_mmu) +DO_ST_TLB(st1hd_be, , uint64_t, stw_be_p, MO_BE, helper_be_stw_mmu) -DO_ST1_D(sve_st1sd_r, cpu_stl_data_ra, uint32_t) +DO_ST_TLB(st1ss_be, H1_4, uint32_t, stl_be_p, MO_BE, helper_be_stl_mmu) +DO_ST_TLB(st1sd_be, , uint64_t, stl_be_p, MO_BE, helper_be_stl_mmu) -DO_ST1(sve_st1bb_r, cpu_stb_data_ra, uint8_t, uint8_t, H1) -DO_ST2(sve_st2bb_r, cpu_stb_data_ra, uint8_t, uint8_t, H1) -DO_ST3(sve_st3bb_r, cpu_stb_data_ra, uint8_t, uint8_t, H1) -DO_ST4(sve_st4bb_r, cpu_stb_data_ra, uint8_t, uint8_t, H1) +DO_ST_TLB(st1dd_be, , uint64_t, stq_be_p, MO_BE, helper_be_stq_mmu) -DO_ST1(sve_st1hh_r, cpu_stw_data_ra, uint16_t, uint16_t, H1_2) -DO_ST2(sve_st2hh_r, cpu_stw_data_ra, uint16_t, uint16_t, H1_2) -DO_ST3(sve_st3hh_r, cpu_stw_data_ra, uint16_t, uint16_t, H1_2) -DO_ST4(sve_st4hh_r, cpu_stw_data_ra, uint16_t, uint16_t, H1_2) +#undef DO_ST_TLB -DO_ST1(sve_st1ss_r, cpu_stl_data_ra, uint32_t, uint32_t, H1_4) -DO_ST2(sve_st2ss_r, cpu_stl_data_ra, uint32_t, uint32_t, H1_4) -DO_ST3(sve_st3ss_r, cpu_stl_data_ra, uint32_t, uint32_t, H1_4) -DO_ST4(sve_st4ss_r, cpu_stl_data_ra, uint32_t, uint32_t, H1_4) - -DO_ST1_D(sve_st1dd_r, cpu_stq_data_ra, uint64_t) - -void HELPER(sve_st2dd_r)(CPUARMState *env, void *vg, - target_ulong addr, uint32_t desc) +/* + * Common helpers for all contiguous 1,2,3,4-register predicated stores. + */ +static void sve_st1_r(CPUARMState *env, void *vg, target_ulong addr, + uint32_t desc, const uintptr_t ra, + const int esize, const int msize, + sve_st1_tlb_fn *tlb_fn) { - intptr_t i, oprsz = simd_oprsz(desc) / 8; - intptr_t ra = GETPC(); + const int mmu_idx = cpu_mmu_index(env, false); + intptr_t i, oprsz = simd_oprsz(desc); unsigned rd = simd_data(desc); - uint64_t *d1 = &env->vfp.zregs[rd].d[0]; - uint64_t *d2 = &env->vfp.zregs[(rd + 1) & 31].d[0]; - uint8_t *pg = vg; + void *vd = &env->vfp.zregs[rd]; - for (i = 0; i < oprsz; i += 1) { - if (pg[H1(i)] & 1) { - cpu_stq_data_ra(env, addr, d1[i], ra); - cpu_stq_data_ra(env, addr + 8, d2[i], ra); - } - addr += 2 * 8; + set_helper_retaddr(ra); + for (i = 0; i < oprsz; ) { + uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); + do { + if (pg & 1) { + tlb_fn(env, vd, i, addr, mmu_idx, ra); + } + i += esize, pg >>= esize; + addr += msize; + } while (i & 15); } + set_helper_retaddr(0); } -void HELPER(sve_st3dd_r)(CPUARMState *env, void *vg, - target_ulong addr, uint32_t desc) +static void sve_st2_r(CPUARMState *env, void *vg, target_ulong addr, + uint32_t desc, const uintptr_t ra, + const int esize, const int msize, + sve_st1_tlb_fn *tlb_fn) { - intptr_t i, oprsz = simd_oprsz(desc) / 8; - intptr_t ra = GETPC(); + const int mmu_idx = cpu_mmu_index(env, false); + intptr_t i, oprsz = simd_oprsz(desc); unsigned rd = simd_data(desc); - uint64_t *d1 = &env->vfp.zregs[rd].d[0]; - uint64_t *d2 = &env->vfp.zregs[(rd + 1) & 31].d[0]; - uint64_t *d3 = &env->vfp.zregs[(rd + 2) & 31].d[0]; - uint8_t *pg = vg; + void *d1 = &env->vfp.zregs[rd]; + void *d2 = &env->vfp.zregs[(rd + 1) & 31]; - for (i = 0; i < oprsz; i += 1) { - if (pg[H1(i)] & 1) { - cpu_stq_data_ra(env, addr, d1[i], ra); - cpu_stq_data_ra(env, addr + 8, d2[i], ra); - cpu_stq_data_ra(env, addr + 16, d3[i], ra); - } - addr += 3 * 8; + set_helper_retaddr(ra); + for (i = 0; i < oprsz; ) { + uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); + do { + if (pg & 1) { + tlb_fn(env, d1, i, addr, mmu_idx, ra); + tlb_fn(env, d2, i, addr + msize, mmu_idx, ra); + } + i += esize, pg >>= esize; + addr += 2 * msize; + } while (i & 15); } + set_helper_retaddr(0); } -void HELPER(sve_st4dd_r)(CPUARMState *env, void *vg, - target_ulong addr, uint32_t desc) +static void sve_st3_r(CPUARMState *env, void *vg, target_ulong addr, + uint32_t desc, const uintptr_t ra, + const int esize, const int msize, + sve_st1_tlb_fn *tlb_fn) { - intptr_t i, oprsz = simd_oprsz(desc) / 8; - intptr_t ra = GETPC(); + const int mmu_idx = cpu_mmu_index(env, false); + intptr_t i, oprsz = simd_oprsz(desc); unsigned rd = simd_data(desc); - uint64_t *d1 = &env->vfp.zregs[rd].d[0]; - uint64_t *d2 = &env->vfp.zregs[(rd + 1) & 31].d[0]; - uint64_t *d3 = &env->vfp.zregs[(rd + 2) & 31].d[0]; - uint64_t *d4 = &env->vfp.zregs[(rd + 3) & 31].d[0]; - uint8_t *pg = vg; + void *d1 = &env->vfp.zregs[rd]; + void *d2 = &env->vfp.zregs[(rd + 1) & 31]; + void *d3 = &env->vfp.zregs[(rd + 2) & 31]; - for (i = 0; i < oprsz; i += 1) { - if (pg[H1(i)] & 1) { - cpu_stq_data_ra(env, addr, d1[i], ra); - cpu_stq_data_ra(env, addr + 8, d2[i], ra); - cpu_stq_data_ra(env, addr + 16, d3[i], ra); - cpu_stq_data_ra(env, addr + 24, d4[i], ra); - } - addr += 4 * 8; + set_helper_retaddr(ra); + for (i = 0; i < oprsz; ) { + uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); + do { + if (pg & 1) { + tlb_fn(env, d1, i, addr, mmu_idx, ra); + tlb_fn(env, d2, i, addr + msize, mmu_idx, ra); + tlb_fn(env, d3, i, addr + 2 * msize, mmu_idx, ra); + } + i += esize, pg >>= esize; + addr += 3 * msize; + } while (i & 15); } + set_helper_retaddr(0); } +static void sve_st4_r(CPUARMState *env, void *vg, target_ulong addr, + uint32_t desc, const uintptr_t ra, + const int esize, const int msize, + sve_st1_tlb_fn *tlb_fn) +{ + const int mmu_idx = cpu_mmu_index(env, false); + intptr_t i, oprsz = simd_oprsz(desc); + unsigned rd = simd_data(desc); + void *d1 = &env->vfp.zregs[rd]; + void *d2 = &env->vfp.zregs[(rd + 1) & 31]; + void *d3 = &env->vfp.zregs[(rd + 2) & 31]; + void *d4 = &env->vfp.zregs[(rd + 3) & 31]; + + set_helper_retaddr(ra); + for (i = 0; i < oprsz; ) { + uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); + do { + if (pg & 1) { + tlb_fn(env, d1, i, addr, mmu_idx, ra); + tlb_fn(env, d2, i, addr + msize, mmu_idx, ra); + tlb_fn(env, d3, i, addr + 2 * msize, mmu_idx, ra); + tlb_fn(env, d4, i, addr + 3 * msize, mmu_idx, ra); + } + i += esize, pg >>= esize; + addr += 4 * msize; + } while (i & 15); + } + set_helper_retaddr(0); +} + +#define DO_STN_1(N, NAME, ESIZE) \ +void __attribute__((flatten)) HELPER(sve_st##N##NAME##_r) \ + (CPUARMState *env, void *vg, target_ulong addr, uint32_t desc) \ +{ \ + sve_st##N##_r(env, vg, addr, desc, GETPC(), ESIZE, 1, \ + sve_st1##NAME##_tlb); \ +} + +#define DO_STN_2(N, NAME, ESIZE, MSIZE) \ +void __attribute__((flatten)) HELPER(sve_st##N##NAME##_r) \ + (CPUARMState *env, void *vg, target_ulong addr, uint32_t desc) \ +{ \ + sve_st##N##_r(env, vg, addr, desc, GETPC(), ESIZE, MSIZE, \ + arm_cpu_data_is_big_endian(env) \ + ? sve_st1##NAME##_be_tlb : sve_st1##NAME##_le_tlb); \ +} + +DO_STN_1(1, bb, 1) +DO_STN_1(1, bh, 2) +DO_STN_1(1, bs, 4) +DO_STN_1(1, bd, 8) +DO_STN_1(2, bb, 1) +DO_STN_1(3, bb, 1) +DO_STN_1(4, bb, 1) + +DO_STN_2(1, hh, 2, 2) +DO_STN_2(1, hs, 4, 2) +DO_STN_2(1, hd, 8, 2) +DO_STN_2(2, hh, 2, 2) +DO_STN_2(3, hh, 2, 2) +DO_STN_2(4, hh, 2, 2) + +DO_STN_2(1, ss, 4, 4) +DO_STN_2(1, sd, 8, 4) +DO_STN_2(2, ss, 4, 4) +DO_STN_2(3, ss, 4, 4) +DO_STN_2(4, ss, 4, 4) + +DO_STN_2(1, dd, 8, 8) +DO_STN_2(2, dd, 8, 8) +DO_STN_2(3, dd, 8, 8) +DO_STN_2(4, dd, 8, 8) + +#undef DO_STN_1 +#undef DO_STN_2 + /* Loads with a vector index. */ #define DO_LD1_ZPZ_S(NAME, TYPEI, TYPEM, FN) \ -- 2.19.0 From: Richard Henderson <richard.henderson@linaro.org> We can choose the endianness at translation time, rather than re-computing it at execution time. Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181005175350.30752-11-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sve.h | 117 +++++++++++++++------- target/arm/sve_helper.c | 70 ++++++------- target/arm/translate-sve.c | 196 +++++++++++++++++++++++++------------ 3 files changed, 252 insertions(+), 131 deletions(-) diff --git a/target/arm/helper-sve.h b/target/arm/helper-sve.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sve.h +++ b/target/arm/helper-sve.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(sve_ld2bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ld3bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ld4bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld1hh_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld2hh_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld3hh_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld4hh_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hh_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2hh_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3hh_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4hh_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld1ss_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld2ss_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld3ss_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld4ss_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hh_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2hh_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3hh_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4hh_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld1dd_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld2dd_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld3dd_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld4dd_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1ss_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2ss_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3ss_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4ss_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1ss_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2ss_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3ss_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4ss_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1dd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2dd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3dd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4dd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1dd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld2dd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld3dd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld4dd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ld1bhu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ld1bsu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(sve_ld1bhs_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ld1bss_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ld1bds_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld1hsu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld1hdu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld1hss_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld1hds_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hsu_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hdu_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hss_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hds_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld1sdu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ld1sds_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hsu_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hdu_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hss_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1hds_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1sdu_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1sds_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ld1sdu_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ld1sds_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldff1bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldff1bhu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(sve_ldff1bhs_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldff1bss_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldff1bds_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldff1hh_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldff1hsu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldff1hdu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldff1hss_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldff1hds_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hh_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hsu_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hdu_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hss_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hds_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldff1ss_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldff1sdu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldff1sds_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hh_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hsu_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hdu_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hss_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1hds_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldff1dd_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1ss_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1sdu_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1sds_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldff1ss_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1sdu_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1sds_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldff1dd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldff1dd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldnf1bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldnf1bhu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(sve_ldnf1bhs_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldnf1bss_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_ldnf1bds_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldnf1hh_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldnf1hsu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldnf1hdu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldnf1hss_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldnf1hds_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hh_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hsu_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hdu_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hss_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hds_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldnf1ss_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldnf1sdu_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldnf1sds_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hh_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hsu_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hdu_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hss_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1hds_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_ldnf1dd_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1ss_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1sdu_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1sds_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldnf1ss_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1sdu_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1sds_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_ldnf1dd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_ldnf1dd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_st1bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_st2bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(sve_##NAME##_r)(CPUARMState *env, void *vg, \ sve_##NAME##_host, sve_##NAME##_tlb); \ } -/* TODO: Propagate the endian check back to the translator. */ #define DO_LD1_2(NAME, ESZ, MSZ) \ -void HELPER(sve_##NAME##_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ -{ \ - if (arm_cpu_data_is_big_endian(env)) { \ - sve_ld1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, \ - sve_##NAME##_be_host, sve_##NAME##_be_tlb); \ - } else { \ - sve_ld1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, \ - sve_##NAME##_le_host, sve_##NAME##_le_tlb); \ - } \ +void HELPER(sve_##NAME##_le_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ld1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, \ + sve_##NAME##_le_host, sve_##NAME##_le_tlb); \ +} \ +void HELPER(sve_##NAME##_be_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ld1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, \ + sve_##NAME##_be_host, sve_##NAME##_be_tlb); \ } DO_LD1_1(ld1bb, 0) @@ -XXX,XX +XXX,XX @@ void __attribute__((flatten)) HELPER(sve_ld##N##bb_r) \ } #define DO_LDN_2(N, SUFF, SIZE) \ -void __attribute__((flatten)) HELPER(sve_ld##N##SUFF##_r) \ +void __attribute__((flatten)) HELPER(sve_ld##N##SUFF##_le_r) \ (CPUARMState *env, void *vg, target_ulong addr, uint32_t desc) \ { \ sve_ld##N##_r(env, vg, addr, desc, SIZE, GETPC(), \ - arm_cpu_data_is_big_endian(env) \ - ? sve_ld1##SUFF##_be_tlb : sve_ld1##SUFF##_le_tlb); \ + sve_ld1##SUFF##_le_tlb); \ +} \ +void __attribute__((flatten)) HELPER(sve_ld##N##SUFF##_be_r) \ + (CPUARMState *env, void *vg, target_ulong addr, uint32_t desc) \ +{ \ + sve_ld##N##_r(env, vg, addr, desc, SIZE, GETPC(), \ + sve_ld1##SUFF##_be_tlb); \ } DO_LDN_1(2) @@ -XXX,XX +XXX,XX @@ void HELPER(sve_ldnf1##PART##_r)(CPUARMState *env, void *vg, \ sve_ldnf1_r(env, vg, addr, desc, ESZ, 0, sve_ld1##PART##_host); \ } -/* TODO: Propagate the endian check back to the translator. */ #define DO_LDFF1_LDNF1_2(PART, ESZ, MSZ) \ -void HELPER(sve_ldff1##PART##_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ +void HELPER(sve_ldff1##PART##_le_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ { \ - if (arm_cpu_data_is_big_endian(env)) { \ - sve_ldff1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, \ - sve_ld1##PART##_be_host, sve_ld1##PART##_be_tlb); \ - } else { \ - sve_ldff1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, \ - sve_ld1##PART##_le_host, sve_ld1##PART##_le_tlb); \ - } \ + sve_ldff1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, \ + sve_ld1##PART##_le_host, sve_ld1##PART##_le_tlb); \ } \ -void HELPER(sve_ldnf1##PART##_r)(CPUARMState *env, void *vg, \ - target_ulong addr, uint32_t desc) \ +void HELPER(sve_ldnf1##PART##_le_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ { \ - if (arm_cpu_data_is_big_endian(env)) { \ - sve_ldnf1_r(env, vg, addr, desc, ESZ, MSZ, \ - sve_ld1##PART##_be_host); \ - } else { \ - sve_ldnf1_r(env, vg, addr, desc, ESZ, MSZ, \ - sve_ld1##PART##_le_host); \ - } \ + sve_ldnf1_r(env, vg, addr, desc, ESZ, MSZ, sve_ld1##PART##_le_host); \ +} \ +void HELPER(sve_ldff1##PART##_be_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldff1_r(env, vg, addr, desc, GETPC(), ESZ, MSZ, \ + sve_ld1##PART##_be_host, sve_ld1##PART##_be_tlb); \ +} \ +void HELPER(sve_ldnf1##PART##_be_r)(CPUARMState *env, void *vg, \ + target_ulong addr, uint32_t desc) \ +{ \ + sve_ldnf1_r(env, vg, addr, desc, ESZ, MSZ, sve_ld1##PART##_be_host); \ } DO_LDFF1_LDNF1_1(bb, 0) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static void do_mem_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, static void do_ld_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, int dtype, int nreg) { - static gen_helper_gvec_mem * const fns[16][4] = { - { gen_helper_sve_ld1bb_r, gen_helper_sve_ld2bb_r, - gen_helper_sve_ld3bb_r, gen_helper_sve_ld4bb_r }, - { gen_helper_sve_ld1bhu_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bsu_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bdu_r, NULL, NULL, NULL }, + static gen_helper_gvec_mem * const fns[2][16][4] = { + /* Little-endian */ + { { gen_helper_sve_ld1bb_r, gen_helper_sve_ld2bb_r, + gen_helper_sve_ld3bb_r, gen_helper_sve_ld4bb_r }, + { gen_helper_sve_ld1bhu_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bsu_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bdu_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1sds_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hh_r, gen_helper_sve_ld2hh_r, - gen_helper_sve_ld3hh_r, gen_helper_sve_ld4hh_r }, - { gen_helper_sve_ld1hsu_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hdu_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1sds_le_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hh_le_r, gen_helper_sve_ld2hh_le_r, + gen_helper_sve_ld3hh_le_r, gen_helper_sve_ld4hh_le_r }, + { gen_helper_sve_ld1hsu_le_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hdu_le_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hds_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1hss_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1ss_r, gen_helper_sve_ld2ss_r, - gen_helper_sve_ld3ss_r, gen_helper_sve_ld4ss_r }, - { gen_helper_sve_ld1sdu_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hds_le_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hss_le_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1ss_le_r, gen_helper_sve_ld2ss_le_r, + gen_helper_sve_ld3ss_le_r, gen_helper_sve_ld4ss_le_r }, + { gen_helper_sve_ld1sdu_le_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bds_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bss_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1bhs_r, NULL, NULL, NULL }, - { gen_helper_sve_ld1dd_r, gen_helper_sve_ld2dd_r, - gen_helper_sve_ld3dd_r, gen_helper_sve_ld4dd_r }, + { gen_helper_sve_ld1bds_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bss_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bhs_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1dd_le_r, gen_helper_sve_ld2dd_le_r, + gen_helper_sve_ld3dd_le_r, gen_helper_sve_ld4dd_le_r } }, + + /* Big-endian */ + { { gen_helper_sve_ld1bb_r, gen_helper_sve_ld2bb_r, + gen_helper_sve_ld3bb_r, gen_helper_sve_ld4bb_r }, + { gen_helper_sve_ld1bhu_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bsu_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bdu_r, NULL, NULL, NULL }, + + { gen_helper_sve_ld1sds_be_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hh_be_r, gen_helper_sve_ld2hh_be_r, + gen_helper_sve_ld3hh_be_r, gen_helper_sve_ld4hh_be_r }, + { gen_helper_sve_ld1hsu_be_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hdu_be_r, NULL, NULL, NULL }, + + { gen_helper_sve_ld1hds_be_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1hss_be_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1ss_be_r, gen_helper_sve_ld2ss_be_r, + gen_helper_sve_ld3ss_be_r, gen_helper_sve_ld4ss_be_r }, + { gen_helper_sve_ld1sdu_be_r, NULL, NULL, NULL }, + + { gen_helper_sve_ld1bds_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bss_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1bhs_r, NULL, NULL, NULL }, + { gen_helper_sve_ld1dd_be_r, gen_helper_sve_ld2dd_be_r, + gen_helper_sve_ld3dd_be_r, gen_helper_sve_ld4dd_be_r } } }; - gen_helper_gvec_mem *fn = fns[dtype][nreg]; + gen_helper_gvec_mem *fn = fns[s->be_data == MO_BE][dtype][nreg]; /* While there are holes in the table, they are not * accessible via the instruction encoding. @@ -XXX,XX +XXX,XX @@ static bool trans_LD_zpri(DisasContext *s, arg_rpri_load *a, uint32_t insn) static bool trans_LDFF1_zprr(DisasContext *s, arg_rprr_load *a, uint32_t insn) { - static gen_helper_gvec_mem * const fns[16] = { - gen_helper_sve_ldff1bb_r, - gen_helper_sve_ldff1bhu_r, - gen_helper_sve_ldff1bsu_r, - gen_helper_sve_ldff1bdu_r, + static gen_helper_gvec_mem * const fns[2][16] = { + /* Little-endian */ + { gen_helper_sve_ldff1bb_r, + gen_helper_sve_ldff1bhu_r, + gen_helper_sve_ldff1bsu_r, + gen_helper_sve_ldff1bdu_r, - gen_helper_sve_ldff1sds_r, - gen_helper_sve_ldff1hh_r, - gen_helper_sve_ldff1hsu_r, - gen_helper_sve_ldff1hdu_r, + gen_helper_sve_ldff1sds_le_r, + gen_helper_sve_ldff1hh_le_r, + gen_helper_sve_ldff1hsu_le_r, + gen_helper_sve_ldff1hdu_le_r, - gen_helper_sve_ldff1hds_r, - gen_helper_sve_ldff1hss_r, - gen_helper_sve_ldff1ss_r, - gen_helper_sve_ldff1sdu_r, + gen_helper_sve_ldff1hds_le_r, + gen_helper_sve_ldff1hss_le_r, + gen_helper_sve_ldff1ss_le_r, + gen_helper_sve_ldff1sdu_le_r, - gen_helper_sve_ldff1bds_r, - gen_helper_sve_ldff1bss_r, - gen_helper_sve_ldff1bhs_r, - gen_helper_sve_ldff1dd_r, + gen_helper_sve_ldff1bds_r, + gen_helper_sve_ldff1bss_r, + gen_helper_sve_ldff1bhs_r, + gen_helper_sve_ldff1dd_le_r }, + + /* Big-endian */ + { gen_helper_sve_ldff1bb_r, + gen_helper_sve_ldff1bhu_r, + gen_helper_sve_ldff1bsu_r, + gen_helper_sve_ldff1bdu_r, + + gen_helper_sve_ldff1sds_be_r, + gen_helper_sve_ldff1hh_be_r, + gen_helper_sve_ldff1hsu_be_r, + gen_helper_sve_ldff1hdu_be_r, + + gen_helper_sve_ldff1hds_be_r, + gen_helper_sve_ldff1hss_be_r, + gen_helper_sve_ldff1ss_be_r, + gen_helper_sve_ldff1sdu_be_r, + + gen_helper_sve_ldff1bds_r, + gen_helper_sve_ldff1bss_r, + gen_helper_sve_ldff1bhs_r, + gen_helper_sve_ldff1dd_be_r }, }; if (sve_access_check(s)) { TCGv_i64 addr = new_tmp_a64(s); tcg_gen_shli_i64(addr, cpu_reg(s, a->rm), dtype_msz(a->dtype)); tcg_gen_add_i64(addr, addr, cpu_reg_sp(s, a->rn)); - do_mem_zpa(s, a->rd, a->pg, addr, fns[a->dtype]); + do_mem_zpa(s, a->rd, a->pg, addr, fns[s->be_data == MO_BE][a->dtype]); } return true; } static bool trans_LDNF1_zpri(DisasContext *s, arg_rpri_load *a, uint32_t insn) { - static gen_helper_gvec_mem * const fns[16] = { - gen_helper_sve_ldnf1bb_r, - gen_helper_sve_ldnf1bhu_r, - gen_helper_sve_ldnf1bsu_r, - gen_helper_sve_ldnf1bdu_r, + static gen_helper_gvec_mem * const fns[2][16] = { + /* Little-endian */ + { gen_helper_sve_ldnf1bb_r, + gen_helper_sve_ldnf1bhu_r, + gen_helper_sve_ldnf1bsu_r, + gen_helper_sve_ldnf1bdu_r, - gen_helper_sve_ldnf1sds_r, - gen_helper_sve_ldnf1hh_r, - gen_helper_sve_ldnf1hsu_r, - gen_helper_sve_ldnf1hdu_r, + gen_helper_sve_ldnf1sds_le_r, + gen_helper_sve_ldnf1hh_le_r, + gen_helper_sve_ldnf1hsu_le_r, + gen_helper_sve_ldnf1hdu_le_r, - gen_helper_sve_ldnf1hds_r, - gen_helper_sve_ldnf1hss_r, - gen_helper_sve_ldnf1ss_r, - gen_helper_sve_ldnf1sdu_r, + gen_helper_sve_ldnf1hds_le_r, + gen_helper_sve_ldnf1hss_le_r, + gen_helper_sve_ldnf1ss_le_r, + gen_helper_sve_ldnf1sdu_le_r, - gen_helper_sve_ldnf1bds_r, - gen_helper_sve_ldnf1bss_r, - gen_helper_sve_ldnf1bhs_r, - gen_helper_sve_ldnf1dd_r, + gen_helper_sve_ldnf1bds_r, + gen_helper_sve_ldnf1bss_r, + gen_helper_sve_ldnf1bhs_r, + gen_helper_sve_ldnf1dd_le_r }, + + /* Big-endian */ + { gen_helper_sve_ldnf1bb_r, + gen_helper_sve_ldnf1bhu_r, + gen_helper_sve_ldnf1bsu_r, + gen_helper_sve_ldnf1bdu_r, + + gen_helper_sve_ldnf1sds_be_r, + gen_helper_sve_ldnf1hh_be_r, + gen_helper_sve_ldnf1hsu_be_r, + gen_helper_sve_ldnf1hdu_be_r, + + gen_helper_sve_ldnf1hds_be_r, + gen_helper_sve_ldnf1hss_be_r, + gen_helper_sve_ldnf1ss_be_r, + gen_helper_sve_ldnf1sdu_be_r, + + gen_helper_sve_ldnf1bds_r, + gen_helper_sve_ldnf1bss_r, + gen_helper_sve_ldnf1bhs_r, + gen_helper_sve_ldnf1dd_be_r }, }; if (sve_access_check(s)) { @@ -XXX,XX +XXX,XX @@ static bool trans_LDNF1_zpri(DisasContext *s, arg_rpri_load *a, uint32_t insn) TCGv_i64 addr = new_tmp_a64(s); tcg_gen_addi_i64(addr, cpu_reg_sp(s, a->rn), off); - do_mem_zpa(s, a->rd, a->pg, addr, fns[a->dtype]); + do_mem_zpa(s, a->rd, a->pg, addr, fns[s->be_data == MO_BE][a->dtype]); } return true; } static void do_ldrq(DisasContext *s, int zt, int pg, TCGv_i64 addr, int msz) { - static gen_helper_gvec_mem * const fns[4] = { - gen_helper_sve_ld1bb_r, gen_helper_sve_ld1hh_r, - gen_helper_sve_ld1ss_r, gen_helper_sve_ld1dd_r, + static gen_helper_gvec_mem * const fns[2][4] = { + { gen_helper_sve_ld1bb_r, gen_helper_sve_ld1hh_le_r, + gen_helper_sve_ld1ss_le_r, gen_helper_sve_ld1dd_le_r }, + { gen_helper_sve_ld1bb_r, gen_helper_sve_ld1hh_be_r, + gen_helper_sve_ld1ss_be_r, gen_helper_sve_ld1dd_be_r }, }; unsigned vsz = vec_full_reg_size(s); TCGv_ptr t_pg; @@ -XXX,XX +XXX,XX @@ static void do_ldrq(DisasContext *s, int zt, int pg, TCGv_i64 addr, int msz) t_pg = tcg_temp_new_ptr(); tcg_gen_addi_ptr(t_pg, cpu_env, poff); - fns[msz](cpu_env, t_pg, addr, desc); + fns[s->be_data == MO_BE][msz](cpu_env, t_pg, addr, desc); tcg_temp_free_ptr(t_pg); tcg_temp_free_i32(desc); -- 2.19.0 From: Richard Henderson <richard.henderson@linaro.org> We can choose the endianness at translation time, rather than re-computing it at execution time. Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181005175350.30752-12-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sve.h | 48 +++++++++++++++++-------- target/arm/sve_helper.c | 11 ++++-- target/arm/translate-sve.c | 72 +++++++++++++++++++++++++++++--------- 3 files changed, 96 insertions(+), 35 deletions(-) diff --git a/target/arm/helper-sve.h b/target/arm/helper-sve.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sve.h +++ b/target/arm/helper-sve.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(sve_st2bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_st3bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_st4bb_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_st1hh_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_st2hh_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_st3hh_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_st4hh_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1hh_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2hh_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3hh_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4hh_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_st1ss_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_st2ss_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_st3ss_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_st4ss_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1hh_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2hh_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3hh_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4hh_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_st1dd_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_st2dd_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_st3dd_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_st4dd_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1ss_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2ss_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3ss_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4ss_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1ss_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2ss_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3ss_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4ss_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1dd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2dd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3dd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4dd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) + +DEF_HELPER_FLAGS_4(sve_st1dd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st2dd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st3dd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st4dd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_st1bh_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_st1bs_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_4(sve_st1bd_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_st1hs_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_st1hd_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1hs_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1hd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1hs_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1hd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) -DEF_HELPER_FLAGS_4(sve_st1sd_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1sd_le_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) +DEF_HELPER_FLAGS_4(sve_st1sd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldbsu_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ void __attribute__((flatten)) HELPER(sve_st##N##NAME##_r) \ } #define DO_STN_2(N, NAME, ESIZE, MSIZE) \ -void __attribute__((flatten)) HELPER(sve_st##N##NAME##_r) \ +void __attribute__((flatten)) HELPER(sve_st##N##NAME##_le_r) \ (CPUARMState *env, void *vg, target_ulong addr, uint32_t desc) \ { \ sve_st##N##_r(env, vg, addr, desc, GETPC(), ESIZE, MSIZE, \ - arm_cpu_data_is_big_endian(env) \ - ? sve_st1##NAME##_be_tlb : sve_st1##NAME##_le_tlb); \ + sve_st1##NAME##_le_tlb); \ +} \ +void __attribute__((flatten)) HELPER(sve_st##N##NAME##_be_r) \ + (CPUARMState *env, void *vg, target_ulong addr, uint32_t desc) \ +{ \ + sve_st##N##_r(env, vg, addr, desc, GETPC(), ESIZE, MSIZE, \ + sve_st1##NAME##_be_tlb); \ } DO_STN_1(1, bb, 1) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_LD1R_zpri(DisasContext *s, arg_rpri_load *a, uint32_t insn) static void do_st_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, int msz, int esz, int nreg) { - static gen_helper_gvec_mem * const fn_single[4][4] = { - { gen_helper_sve_st1bb_r, gen_helper_sve_st1bh_r, - gen_helper_sve_st1bs_r, gen_helper_sve_st1bd_r }, - { NULL, gen_helper_sve_st1hh_r, - gen_helper_sve_st1hs_r, gen_helper_sve_st1hd_r }, - { NULL, NULL, - gen_helper_sve_st1ss_r, gen_helper_sve_st1sd_r }, - { NULL, NULL, NULL, gen_helper_sve_st1dd_r }, + static gen_helper_gvec_mem * const fn_single[2][4][4] = { + { { gen_helper_sve_st1bb_r, + gen_helper_sve_st1bh_r, + gen_helper_sve_st1bs_r, + gen_helper_sve_st1bd_r }, + { NULL, + gen_helper_sve_st1hh_le_r, + gen_helper_sve_st1hs_le_r, + gen_helper_sve_st1hd_le_r }, + { NULL, NULL, + gen_helper_sve_st1ss_le_r, + gen_helper_sve_st1sd_le_r }, + { NULL, NULL, NULL, + gen_helper_sve_st1dd_le_r } }, + { { gen_helper_sve_st1bb_r, + gen_helper_sve_st1bh_r, + gen_helper_sve_st1bs_r, + gen_helper_sve_st1bd_r }, + { NULL, + gen_helper_sve_st1hh_be_r, + gen_helper_sve_st1hs_be_r, + gen_helper_sve_st1hd_be_r }, + { NULL, NULL, + gen_helper_sve_st1ss_be_r, + gen_helper_sve_st1sd_be_r }, + { NULL, NULL, NULL, + gen_helper_sve_st1dd_be_r } }, }; - static gen_helper_gvec_mem * const fn_multiple[3][4] = { - { gen_helper_sve_st2bb_r, gen_helper_sve_st2hh_r, - gen_helper_sve_st2ss_r, gen_helper_sve_st2dd_r }, - { gen_helper_sve_st3bb_r, gen_helper_sve_st3hh_r, - gen_helper_sve_st3ss_r, gen_helper_sve_st3dd_r }, - { gen_helper_sve_st4bb_r, gen_helper_sve_st4hh_r, - gen_helper_sve_st4ss_r, gen_helper_sve_st4dd_r }, + static gen_helper_gvec_mem * const fn_multiple[2][3][4] = { + { { gen_helper_sve_st2bb_r, + gen_helper_sve_st2hh_le_r, + gen_helper_sve_st2ss_le_r, + gen_helper_sve_st2dd_le_r }, + { gen_helper_sve_st3bb_r, + gen_helper_sve_st3hh_le_r, + gen_helper_sve_st3ss_le_r, + gen_helper_sve_st3dd_le_r }, + { gen_helper_sve_st4bb_r, + gen_helper_sve_st4hh_le_r, + gen_helper_sve_st4ss_le_r, + gen_helper_sve_st4dd_le_r } }, + { { gen_helper_sve_st2bb_r, + gen_helper_sve_st2hh_be_r, + gen_helper_sve_st2ss_be_r, + gen_helper_sve_st2dd_be_r }, + { gen_helper_sve_st3bb_r, + gen_helper_sve_st3hh_be_r, + gen_helper_sve_st3ss_be_r, + gen_helper_sve_st3dd_be_r }, + { gen_helper_sve_st4bb_r, + gen_helper_sve_st4hh_be_r, + gen_helper_sve_st4ss_be_r, + gen_helper_sve_st4dd_be_r } }, }; gen_helper_gvec_mem *fn; + int be = s->be_data == MO_BE; if (nreg == 0) { /* ST1 */ - fn = fn_single[msz][esz]; + fn = fn_single[be][msz][esz]; } else { /* ST2, ST3, ST4 -- msz == esz, enforced by encoding */ assert(msz == esz); - fn = fn_multiple[nreg - 1][msz]; + fn = fn_multiple[be][nreg - 1][msz]; } assert(fn != NULL); do_mem_zpa(s, zt, pg, addr, fn); -- 2.19.0 From: Richard Henderson <richard.henderson@linaro.org> This fixes the endianness problem for softmmu, and moves the main loop out of a macro and into an inlined function. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181005175350.30752-13-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sve.h | 84 +++++++++---- target/arm/sve_helper.c | 225 ++++++++++++++++++++++++---------- target/arm/translate-sve.c | 244 +++++++++++++++++++++++++------------ 3 files changed, 386 insertions(+), 167 deletions(-) diff --git a/target/arm/helper-sve.h b/target/arm/helper-sve.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sve.h +++ b/target/arm/helper-sve.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(sve_st1sd_be_r, TCG_CALL_NO_WG, void, env, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldbsu_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldhsu_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhsu_le_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldssu_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhsu_be_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldss_le_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldss_be_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldbss_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldhss_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhss_le_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhss_be_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldbsu_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldhsu_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhsu_le_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldssu_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhsu_be_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldss_le_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldss_be_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldbss_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldhss_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhss_le_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldhss_be_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldbdu_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldhdu_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhdu_le_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldsdu_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhdu_be_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldddu_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldsdu_le_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsdu_be_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_le_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_be_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldbds_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldhds_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhds_le_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldsds_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhds_be_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_le_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_be_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldbdu_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldhdu_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhdu_le_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldsdu_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhdu_be_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldddu_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldsdu_le_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsdu_be_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_le_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_be_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldbds_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldhds_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhds_le_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldsds_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhds_be_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_le_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_be_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldbdu_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldhdu_zd, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhdu_le_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldsdu_zd, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhdu_be_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldddu_zd, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldsdu_le_zd, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsdu_be_zd, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_le_zd, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_lddd_be_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldbds_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldhds_zd, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhds_le_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldsds_zd, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldhds_be_zd, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_le_zd, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldsds_be_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldffbsu_zsu, TCG_CALL_NO_WG, diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ DO_STN_2(4, dd, 8, 8) #undef DO_STN_1 #undef DO_STN_2 -/* Loads with a vector index. */ +/* + * Loads with a vector index. + */ -#define DO_LD1_ZPZ_S(NAME, TYPEI, TYPEM, FN) \ -void HELPER(NAME)(CPUARMState *env, void *vd, void *vg, void *vm, \ - target_ulong base, uint32_t desc) \ -{ \ - intptr_t i, oprsz = simd_oprsz(desc); \ - unsigned scale = simd_data(desc); \ - uintptr_t ra = GETPC(); \ - for (i = 0; i < oprsz; ) { \ - uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); \ - do { \ - TYPEM m = 0; \ - if (pg & 1) { \ - target_ulong off = *(TYPEI *)(vm + H1_4(i)); \ - m = FN(env, base + (off << scale), ra); \ - } \ - *(uint32_t *)(vd + H1_4(i)) = m; \ - i += 4, pg >>= 4; \ - } while (i & 15); \ - } \ +/* + * Load the element at @reg + @reg_ofs, sign or zero-extend as needed. + */ +typedef target_ulong zreg_off_fn(void *reg, intptr_t reg_ofs); + +static target_ulong off_zsu_s(void *reg, intptr_t reg_ofs) +{ + return *(uint32_t *)(reg + H1_4(reg_ofs)); } -#define DO_LD1_ZPZ_D(NAME, TYPEI, TYPEM, FN) \ -void HELPER(NAME)(CPUARMState *env, void *vd, void *vg, void *vm, \ - target_ulong base, uint32_t desc) \ -{ \ - intptr_t i, oprsz = simd_oprsz(desc) / 8; \ - unsigned scale = simd_data(desc); \ - uintptr_t ra = GETPC(); \ - uint64_t *d = vd, *m = vm; uint8_t *pg = vg; \ - for (i = 0; i < oprsz; i++) { \ - TYPEM mm = 0; \ - if (pg[H1(i)] & 1) { \ - target_ulong off = (TYPEI)m[i]; \ - mm = FN(env, base + (off << scale), ra); \ - } \ - d[i] = mm; \ - } \ +static target_ulong off_zss_s(void *reg, intptr_t reg_ofs) +{ + return *(int32_t *)(reg + H1_4(reg_ofs)); } -DO_LD1_ZPZ_S(sve_ldbsu_zsu, uint32_t, uint8_t, cpu_ldub_data_ra) -DO_LD1_ZPZ_S(sve_ldhsu_zsu, uint32_t, uint16_t, cpu_lduw_data_ra) -DO_LD1_ZPZ_S(sve_ldssu_zsu, uint32_t, uint32_t, cpu_ldl_data_ra) -DO_LD1_ZPZ_S(sve_ldbss_zsu, uint32_t, int8_t, cpu_ldub_data_ra) -DO_LD1_ZPZ_S(sve_ldhss_zsu, uint32_t, int16_t, cpu_lduw_data_ra) +static target_ulong off_zsu_d(void *reg, intptr_t reg_ofs) +{ + return (uint32_t)*(uint64_t *)(reg + reg_ofs); +} -DO_LD1_ZPZ_S(sve_ldbsu_zss, int32_t, uint8_t, cpu_ldub_data_ra) -DO_LD1_ZPZ_S(sve_ldhsu_zss, int32_t, uint16_t, cpu_lduw_data_ra) -DO_LD1_ZPZ_S(sve_ldssu_zss, int32_t, uint32_t, cpu_ldl_data_ra) -DO_LD1_ZPZ_S(sve_ldbss_zss, int32_t, int8_t, cpu_ldub_data_ra) -DO_LD1_ZPZ_S(sve_ldhss_zss, int32_t, int16_t, cpu_lduw_data_ra) +static target_ulong off_zss_d(void *reg, intptr_t reg_ofs) +{ + return (int32_t)*(uint64_t *)(reg + reg_ofs); +} -DO_LD1_ZPZ_D(sve_ldbdu_zsu, uint32_t, uint8_t, cpu_ldub_data_ra) -DO_LD1_ZPZ_D(sve_ldhdu_zsu, uint32_t, uint16_t, cpu_lduw_data_ra) -DO_LD1_ZPZ_D(sve_ldsdu_zsu, uint32_t, uint32_t, cpu_ldl_data_ra) -DO_LD1_ZPZ_D(sve_ldddu_zsu, uint32_t, uint64_t, cpu_ldq_data_ra) -DO_LD1_ZPZ_D(sve_ldbds_zsu, uint32_t, int8_t, cpu_ldub_data_ra) -DO_LD1_ZPZ_D(sve_ldhds_zsu, uint32_t, int16_t, cpu_lduw_data_ra) -DO_LD1_ZPZ_D(sve_ldsds_zsu, uint32_t, int32_t, cpu_ldl_data_ra) +static target_ulong off_zd_d(void *reg, intptr_t reg_ofs) +{ + return *(uint64_t *)(reg + reg_ofs); +} -DO_LD1_ZPZ_D(sve_ldbdu_zss, int32_t, uint8_t, cpu_ldub_data_ra) -DO_LD1_ZPZ_D(sve_ldhdu_zss, int32_t, uint16_t, cpu_lduw_data_ra) -DO_LD1_ZPZ_D(sve_ldsdu_zss, int32_t, uint32_t, cpu_ldl_data_ra) -DO_LD1_ZPZ_D(sve_ldddu_zss, int32_t, uint64_t, cpu_ldq_data_ra) -DO_LD1_ZPZ_D(sve_ldbds_zss, int32_t, int8_t, cpu_ldub_data_ra) -DO_LD1_ZPZ_D(sve_ldhds_zss, int32_t, int16_t, cpu_lduw_data_ra) -DO_LD1_ZPZ_D(sve_ldsds_zss, int32_t, int32_t, cpu_ldl_data_ra) +static void sve_ld1_zs(CPUARMState *env, void *vd, void *vg, void *vm, + target_ulong base, uint32_t desc, uintptr_t ra, + zreg_off_fn *off_fn, sve_ld1_tlb_fn *tlb_fn) +{ + const int mmu_idx = cpu_mmu_index(env, false); + intptr_t i, oprsz = simd_oprsz(desc); + unsigned scale = simd_data(desc); + ARMVectorReg scratch = { }; -DO_LD1_ZPZ_D(sve_ldbdu_zd, uint64_t, uint8_t, cpu_ldub_data_ra) -DO_LD1_ZPZ_D(sve_ldhdu_zd, uint64_t, uint16_t, cpu_lduw_data_ra) -DO_LD1_ZPZ_D(sve_ldsdu_zd, uint64_t, uint32_t, cpu_ldl_data_ra) -DO_LD1_ZPZ_D(sve_ldddu_zd, uint64_t, uint64_t, cpu_ldq_data_ra) -DO_LD1_ZPZ_D(sve_ldbds_zd, uint64_t, int8_t, cpu_ldub_data_ra) -DO_LD1_ZPZ_D(sve_ldhds_zd, uint64_t, int16_t, cpu_lduw_data_ra) -DO_LD1_ZPZ_D(sve_ldsds_zd, uint64_t, int32_t, cpu_ldl_data_ra) + set_helper_retaddr(ra); + for (i = 0; i < oprsz; ) { + uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); + do { + if (likely(pg & 1)) { + target_ulong off = off_fn(vm, i); + tlb_fn(env, &scratch, i, base + (off << scale), mmu_idx, ra); + } + i += 4, pg >>= 4; + } while (i & 15); + } + set_helper_retaddr(0); + + /* Wait until all exceptions have been raised to write back. */ + memcpy(vd, &scratch, oprsz); +} + +static void sve_ld1_zd(CPUARMState *env, void *vd, void *vg, void *vm, + target_ulong base, uint32_t desc, uintptr_t ra, + zreg_off_fn *off_fn, sve_ld1_tlb_fn *tlb_fn) +{ + const int mmu_idx = cpu_mmu_index(env, false); + intptr_t i, oprsz = simd_oprsz(desc) / 8; + unsigned scale = simd_data(desc); + ARMVectorReg scratch = { }; + + set_helper_retaddr(ra); + for (i = 0; i < oprsz; i++) { + uint8_t pg = *(uint8_t *)(vg + H1(i)); + if (likely(pg & 1)) { + target_ulong off = off_fn(vm, i * 8); + tlb_fn(env, &scratch, i * 8, base + (off << scale), mmu_idx, ra); + } + } + set_helper_retaddr(0); + + /* Wait until all exceptions have been raised to write back. */ + memcpy(vd, &scratch, oprsz * 8); +} + +#define DO_LD1_ZPZ_S(MEM, OFS) \ +void __attribute__((flatten)) HELPER(sve_ld##MEM##_##OFS) \ + (CPUARMState *env, void *vd, void *vg, void *vm, \ + target_ulong base, uint32_t desc) \ +{ \ + sve_ld1_zs(env, vd, vg, vm, base, desc, GETPC(), \ + off_##OFS##_s, sve_ld1##MEM##_tlb); \ +} + +#define DO_LD1_ZPZ_D(MEM, OFS) \ +void __attribute__((flatten)) HELPER(sve_ld##MEM##_##OFS) \ + (CPUARMState *env, void *vd, void *vg, void *vm, \ + target_ulong base, uint32_t desc) \ +{ \ + sve_ld1_zd(env, vd, vg, vm, base, desc, GETPC(), \ + off_##OFS##_d, sve_ld1##MEM##_tlb); \ +} + +DO_LD1_ZPZ_S(bsu, zsu) +DO_LD1_ZPZ_S(bsu, zss) +DO_LD1_ZPZ_D(bdu, zsu) +DO_LD1_ZPZ_D(bdu, zss) +DO_LD1_ZPZ_D(bdu, zd) + +DO_LD1_ZPZ_S(bss, zsu) +DO_LD1_ZPZ_S(bss, zss) +DO_LD1_ZPZ_D(bds, zsu) +DO_LD1_ZPZ_D(bds, zss) +DO_LD1_ZPZ_D(bds, zd) + +DO_LD1_ZPZ_S(hsu_le, zsu) +DO_LD1_ZPZ_S(hsu_le, zss) +DO_LD1_ZPZ_D(hdu_le, zsu) +DO_LD1_ZPZ_D(hdu_le, zss) +DO_LD1_ZPZ_D(hdu_le, zd) + +DO_LD1_ZPZ_S(hsu_be, zsu) +DO_LD1_ZPZ_S(hsu_be, zss) +DO_LD1_ZPZ_D(hdu_be, zsu) +DO_LD1_ZPZ_D(hdu_be, zss) +DO_LD1_ZPZ_D(hdu_be, zd) + +DO_LD1_ZPZ_S(hss_le, zsu) +DO_LD1_ZPZ_S(hss_le, zss) +DO_LD1_ZPZ_D(hds_le, zsu) +DO_LD1_ZPZ_D(hds_le, zss) +DO_LD1_ZPZ_D(hds_le, zd) + +DO_LD1_ZPZ_S(hss_be, zsu) +DO_LD1_ZPZ_S(hss_be, zss) +DO_LD1_ZPZ_D(hds_be, zsu) +DO_LD1_ZPZ_D(hds_be, zss) +DO_LD1_ZPZ_D(hds_be, zd) + +DO_LD1_ZPZ_S(ss_le, zsu) +DO_LD1_ZPZ_S(ss_le, zss) +DO_LD1_ZPZ_D(sdu_le, zsu) +DO_LD1_ZPZ_D(sdu_le, zss) +DO_LD1_ZPZ_D(sdu_le, zd) + +DO_LD1_ZPZ_S(ss_be, zsu) +DO_LD1_ZPZ_S(ss_be, zss) +DO_LD1_ZPZ_D(sdu_be, zsu) +DO_LD1_ZPZ_D(sdu_be, zss) +DO_LD1_ZPZ_D(sdu_be, zd) + +DO_LD1_ZPZ_D(sds_le, zsu) +DO_LD1_ZPZ_D(sds_le, zss) +DO_LD1_ZPZ_D(sds_le, zd) + +DO_LD1_ZPZ_D(sds_be, zsu) +DO_LD1_ZPZ_D(sds_be, zss) +DO_LD1_ZPZ_D(sds_be, zd) + +DO_LD1_ZPZ_D(dd_le, zsu) +DO_LD1_ZPZ_D(dd_le, zss) +DO_LD1_ZPZ_D(dd_le, zd) + +DO_LD1_ZPZ_D(dd_be, zsu) +DO_LD1_ZPZ_D(dd_be, zss) +DO_LD1_ZPZ_D(dd_be, zd) + +#undef DO_LD1_ZPZ_S +#undef DO_LD1_ZPZ_D /* First fault loads with a vector index. */ diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static void do_mem_zpz(DisasContext *s, int zt, int pg, int zm, int scale, tcg_temp_free_i32(desc); } -/* Indexed by [ff][xs][u][msz]. */ -static gen_helper_gvec_mem_scatter * const gather_load_fn32[2][2][2][3] = { - { { { gen_helper_sve_ldbss_zsu, - gen_helper_sve_ldhss_zsu, - NULL, }, - { gen_helper_sve_ldbsu_zsu, - gen_helper_sve_ldhsu_zsu, - gen_helper_sve_ldssu_zsu, } }, - { { gen_helper_sve_ldbss_zss, - gen_helper_sve_ldhss_zss, - NULL, }, - { gen_helper_sve_ldbsu_zss, - gen_helper_sve_ldhsu_zss, - gen_helper_sve_ldssu_zss, } } }, +/* Indexed by [be][ff][xs][u][msz]. */ +static gen_helper_gvec_mem_scatter * const gather_load_fn32[2][2][2][2][3] = { + /* Little-endian */ + { { { { gen_helper_sve_ldbss_zsu, + gen_helper_sve_ldhss_le_zsu, + NULL, }, + { gen_helper_sve_ldbsu_zsu, + gen_helper_sve_ldhsu_le_zsu, + gen_helper_sve_ldss_le_zsu, } }, + { { gen_helper_sve_ldbss_zss, + gen_helper_sve_ldhss_le_zss, + NULL, }, + { gen_helper_sve_ldbsu_zss, + gen_helper_sve_ldhsu_le_zss, + gen_helper_sve_ldss_le_zss, } } }, - { { { gen_helper_sve_ldffbss_zsu, - gen_helper_sve_ldffhss_zsu, - NULL, }, - { gen_helper_sve_ldffbsu_zsu, - gen_helper_sve_ldffhsu_zsu, - gen_helper_sve_ldffssu_zsu, } }, - { { gen_helper_sve_ldffbss_zss, - gen_helper_sve_ldffhss_zss, - NULL, }, - { gen_helper_sve_ldffbsu_zss, - gen_helper_sve_ldffhsu_zss, - gen_helper_sve_ldffssu_zss, } } } + /* First-fault */ + { { { gen_helper_sve_ldffbss_zsu, + gen_helper_sve_ldffhss_zsu, + NULL, }, + { gen_helper_sve_ldffbsu_zsu, + gen_helper_sve_ldffhsu_zsu, + gen_helper_sve_ldffssu_zsu, } }, + { { gen_helper_sve_ldffbss_zss, + gen_helper_sve_ldffhss_zss, + NULL, }, + { gen_helper_sve_ldffbsu_zss, + gen_helper_sve_ldffhsu_zss, + gen_helper_sve_ldffssu_zss, } } } }, + + /* Big-endian */ + { { { { gen_helper_sve_ldbss_zsu, + gen_helper_sve_ldhss_be_zsu, + NULL, }, + { gen_helper_sve_ldbsu_zsu, + gen_helper_sve_ldhsu_be_zsu, + gen_helper_sve_ldss_be_zsu, } }, + { { gen_helper_sve_ldbss_zss, + gen_helper_sve_ldhss_be_zss, + NULL, }, + { gen_helper_sve_ldbsu_zss, + gen_helper_sve_ldhsu_be_zss, + gen_helper_sve_ldss_be_zss, } } }, + + /* First-fault */ + { { { gen_helper_sve_ldffbss_zsu, + gen_helper_sve_ldffhss_zsu, + NULL, }, + { gen_helper_sve_ldffbsu_zsu, + gen_helper_sve_ldffhsu_zsu, + gen_helper_sve_ldffssu_zsu, } }, + { { gen_helper_sve_ldffbss_zss, + gen_helper_sve_ldffhss_zss, + NULL, }, + { gen_helper_sve_ldffbsu_zss, + gen_helper_sve_ldffhsu_zss, + gen_helper_sve_ldffssu_zss, } } } }, }; /* Note that we overload xs=2 to indicate 64-bit offset. */ -static gen_helper_gvec_mem_scatter * const gather_load_fn64[2][3][2][4] = { - { { { gen_helper_sve_ldbds_zsu, - gen_helper_sve_ldhds_zsu, - gen_helper_sve_ldsds_zsu, - NULL, }, - { gen_helper_sve_ldbdu_zsu, - gen_helper_sve_ldhdu_zsu, - gen_helper_sve_ldsdu_zsu, - gen_helper_sve_ldddu_zsu, } }, - { { gen_helper_sve_ldbds_zss, - gen_helper_sve_ldhds_zss, - gen_helper_sve_ldsds_zss, - NULL, }, - { gen_helper_sve_ldbdu_zss, - gen_helper_sve_ldhdu_zss, - gen_helper_sve_ldsdu_zss, - gen_helper_sve_ldddu_zss, } }, - { { gen_helper_sve_ldbds_zd, - gen_helper_sve_ldhds_zd, - gen_helper_sve_ldsds_zd, - NULL, }, - { gen_helper_sve_ldbdu_zd, - gen_helper_sve_ldhdu_zd, - gen_helper_sve_ldsdu_zd, - gen_helper_sve_ldddu_zd, } } }, +static gen_helper_gvec_mem_scatter * const gather_load_fn64[2][2][3][2][4] = { + /* Little-endian */ + { { { { gen_helper_sve_ldbds_zsu, + gen_helper_sve_ldhds_le_zsu, + gen_helper_sve_ldsds_le_zsu, + NULL, }, + { gen_helper_sve_ldbdu_zsu, + gen_helper_sve_ldhdu_le_zsu, + gen_helper_sve_ldsdu_le_zsu, + gen_helper_sve_lddd_le_zsu, } }, + { { gen_helper_sve_ldbds_zss, + gen_helper_sve_ldhds_le_zss, + gen_helper_sve_ldsds_le_zss, + NULL, }, + { gen_helper_sve_ldbdu_zss, + gen_helper_sve_ldhdu_le_zss, + gen_helper_sve_ldsdu_le_zss, + gen_helper_sve_lddd_le_zss, } }, + { { gen_helper_sve_ldbds_zd, + gen_helper_sve_ldhds_le_zd, + gen_helper_sve_ldsds_le_zd, + NULL, }, + { gen_helper_sve_ldbdu_zd, + gen_helper_sve_ldhdu_le_zd, + gen_helper_sve_ldsdu_le_zd, + gen_helper_sve_lddd_le_zd, } } }, - { { { gen_helper_sve_ldffbds_zsu, - gen_helper_sve_ldffhds_zsu, - gen_helper_sve_ldffsds_zsu, - NULL, }, - { gen_helper_sve_ldffbdu_zsu, - gen_helper_sve_ldffhdu_zsu, - gen_helper_sve_ldffsdu_zsu, - gen_helper_sve_ldffddu_zsu, } }, - { { gen_helper_sve_ldffbds_zss, - gen_helper_sve_ldffhds_zss, - gen_helper_sve_ldffsds_zss, - NULL, }, - { gen_helper_sve_ldffbdu_zss, - gen_helper_sve_ldffhdu_zss, - gen_helper_sve_ldffsdu_zss, - gen_helper_sve_ldffddu_zss, } }, - { { gen_helper_sve_ldffbds_zd, - gen_helper_sve_ldffhds_zd, - gen_helper_sve_ldffsds_zd, - NULL, }, - { gen_helper_sve_ldffbdu_zd, - gen_helper_sve_ldffhdu_zd, - gen_helper_sve_ldffsdu_zd, - gen_helper_sve_ldffddu_zd, } } } + /* First-fault */ + { { { gen_helper_sve_ldffbds_zsu, + gen_helper_sve_ldffhds_zsu, + gen_helper_sve_ldffsds_zsu, + NULL, }, + { gen_helper_sve_ldffbdu_zsu, + gen_helper_sve_ldffhdu_zsu, + gen_helper_sve_ldffsdu_zsu, + gen_helper_sve_ldffddu_zsu, } }, + { { gen_helper_sve_ldffbds_zss, + gen_helper_sve_ldffhds_zss, + gen_helper_sve_ldffsds_zss, + NULL, }, + { gen_helper_sve_ldffbdu_zss, + gen_helper_sve_ldffhdu_zss, + gen_helper_sve_ldffsdu_zss, + gen_helper_sve_ldffddu_zss, } }, + { { gen_helper_sve_ldffbds_zd, + gen_helper_sve_ldffhds_zd, + gen_helper_sve_ldffsds_zd, + NULL, }, + { gen_helper_sve_ldffbdu_zd, + gen_helper_sve_ldffhdu_zd, + gen_helper_sve_ldffsdu_zd, + gen_helper_sve_ldffddu_zd, } } } }, + + /* Big-endian */ + { { { { gen_helper_sve_ldbds_zsu, + gen_helper_sve_ldhds_be_zsu, + gen_helper_sve_ldsds_be_zsu, + NULL, }, + { gen_helper_sve_ldbdu_zsu, + gen_helper_sve_ldhdu_be_zsu, + gen_helper_sve_ldsdu_be_zsu, + gen_helper_sve_lddd_be_zsu, } }, + { { gen_helper_sve_ldbds_zss, + gen_helper_sve_ldhds_be_zss, + gen_helper_sve_ldsds_be_zss, + NULL, }, + { gen_helper_sve_ldbdu_zss, + gen_helper_sve_ldhdu_be_zss, + gen_helper_sve_ldsdu_be_zss, + gen_helper_sve_lddd_be_zss, } }, + { { gen_helper_sve_ldbds_zd, + gen_helper_sve_ldhds_be_zd, + gen_helper_sve_ldsds_be_zd, + NULL, }, + { gen_helper_sve_ldbdu_zd, + gen_helper_sve_ldhdu_be_zd, + gen_helper_sve_ldsdu_be_zd, + gen_helper_sve_lddd_be_zd, } } }, + + /* First-fault */ + { { { gen_helper_sve_ldffbds_zsu, + gen_helper_sve_ldffhds_zsu, + gen_helper_sve_ldffsds_zsu, + NULL, }, + { gen_helper_sve_ldffbdu_zsu, + gen_helper_sve_ldffhdu_zsu, + gen_helper_sve_ldffsdu_zsu, + gen_helper_sve_ldffddu_zsu, } }, + { { gen_helper_sve_ldffbds_zss, + gen_helper_sve_ldffhds_zss, + gen_helper_sve_ldffsds_zss, + NULL, }, + { gen_helper_sve_ldffbdu_zss, + gen_helper_sve_ldffhdu_zss, + gen_helper_sve_ldffsdu_zss, + gen_helper_sve_ldffddu_zss, } }, + { { gen_helper_sve_ldffbds_zd, + gen_helper_sve_ldffhds_zd, + gen_helper_sve_ldffsds_zd, + NULL, }, + { gen_helper_sve_ldffbdu_zd, + gen_helper_sve_ldffhdu_zd, + gen_helper_sve_ldffsdu_zd, + gen_helper_sve_ldffddu_zd, } } } }, }; static bool trans_LD1_zprz(DisasContext *s, arg_LD1_zprz *a, uint32_t insn) { gen_helper_gvec_mem_scatter *fn = NULL; + int be = s->be_data == MO_BE; if (!sve_access_check(s)) { return true; @@ -XXX,XX +XXX,XX @@ static bool trans_LD1_zprz(DisasContext *s, arg_LD1_zprz *a, uint32_t insn) switch (a->esz) { case MO_32: - fn = gather_load_fn32[a->ff][a->xs][a->u][a->msz]; + fn = gather_load_fn32[be][a->ff][a->xs][a->u][a->msz]; break; case MO_64: - fn = gather_load_fn64[a->ff][a->xs][a->u][a->msz]; + fn = gather_load_fn64[be][a->ff][a->xs][a->u][a->msz]; break; } assert(fn != NULL); @@ -XXX,XX +XXX,XX @@ static bool trans_LD1_zprz(DisasContext *s, arg_LD1_zprz *a, uint32_t insn) static bool trans_LD1_zpiz(DisasContext *s, arg_LD1_zpiz *a, uint32_t insn) { gen_helper_gvec_mem_scatter *fn = NULL; + int be = s->be_data == MO_BE; TCGv_i64 imm; if (a->esz < a->msz || (a->esz == a->msz && !a->u)) { @@ -XXX,XX +XXX,XX @@ static bool trans_LD1_zpiz(DisasContext *s, arg_LD1_zpiz *a, uint32_t insn) switch (a->esz) { case MO_32: - fn = gather_load_fn32[a->ff][0][a->u][a->msz]; + fn = gather_load_fn32[be][a->ff][0][a->u][a->msz]; break; case MO_64: - fn = gather_load_fn64[a->ff][2][a->u][a->msz]; + fn = gather_load_fn64[be][a->ff][2][a->u][a->msz]; break; } assert(fn != NULL); -- 2.19.0 From: Richard Henderson <richard.henderson@linaro.org> This fixes the endianness problem for softmmu, and moves the main loop out of a macro and into an inlined function. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181005175350.30752-14-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sve.h | 52 ++++++++++---- target/arm/sve_helper.c | 139 ++++++++++++++++++++++++------------- target/arm/translate-sve.c | 74 +++++++++++++------- 3 files changed, 177 insertions(+), 88 deletions(-) diff --git a/target/arm/helper-sve.h b/target/arm/helper-sve.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sve.h +++ b/target/arm/helper-sve.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_6(sve_ldffsds_zd, TCG_CALL_NO_WG, DEF_HELPER_FLAGS_6(sve_stbs_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_sths_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_sths_le_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_stss_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_sths_be_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stss_le_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stss_be_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_stbs_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_sths_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_sths_le_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_stss_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_sths_be_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stss_le_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stss_be_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_stbd_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_sthd_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_sthd_le_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_stsd_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_sthd_be_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_stdd_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_stsd_le_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stsd_be_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_le_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_be_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_stbd_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_sthd_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_sthd_le_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_stsd_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_sthd_be_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_stdd_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_stsd_le_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stsd_be_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_le_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_be_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_stbd_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_sthd_zd, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_sthd_le_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_stsd_zd, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_sthd_be_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_stdd_zd, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_stsd_le_zd, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stsd_be_zd, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_le_zd, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_stdd_be_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ DO_LDFF1_ZPZ_D(sve_ldffsds_zd, uint64_t, int32_t, cpu_ldl_data_ra) /* Stores with a vector index. */ -#define DO_ST1_ZPZ_S(NAME, TYPEI, FN) \ -void HELPER(NAME)(CPUARMState *env, void *vd, void *vg, void *vm, \ - target_ulong base, uint32_t desc) \ -{ \ - intptr_t i, oprsz = simd_oprsz(desc); \ - unsigned scale = simd_data(desc); \ - uintptr_t ra = GETPC(); \ - for (i = 0; i < oprsz; ) { \ - uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); \ - do { \ - if (likely(pg & 1)) { \ - target_ulong off = *(TYPEI *)(vm + H1_4(i)); \ - uint32_t d = *(uint32_t *)(vd + H1_4(i)); \ - FN(env, base + (off << scale), d, ra); \ - } \ - i += sizeof(uint32_t), pg >>= sizeof(uint32_t); \ - } while (i & 15); \ - } \ +static void sve_st1_zs(CPUARMState *env, void *vd, void *vg, void *vm, + target_ulong base, uint32_t desc, uintptr_t ra, + zreg_off_fn *off_fn, sve_ld1_tlb_fn *tlb_fn) +{ + const int mmu_idx = cpu_mmu_index(env, false); + intptr_t i, oprsz = simd_oprsz(desc); + unsigned scale = simd_data(desc); + + set_helper_retaddr(ra); + for (i = 0; i < oprsz; ) { + uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); + do { + if (likely(pg & 1)) { + target_ulong off = off_fn(vm, i); + tlb_fn(env, vd, i, base + (off << scale), mmu_idx, ra); + } + i += 4, pg >>= 4; + } while (i & 15); + } + set_helper_retaddr(0); } -#define DO_ST1_ZPZ_D(NAME, TYPEI, FN) \ -void HELPER(NAME)(CPUARMState *env, void *vd, void *vg, void *vm, \ - target_ulong base, uint32_t desc) \ -{ \ - intptr_t i, oprsz = simd_oprsz(desc) / 8; \ - unsigned scale = simd_data(desc); \ - uintptr_t ra = GETPC(); \ - uint64_t *d = vd, *m = vm; uint8_t *pg = vg; \ - for (i = 0; i < oprsz; i++) { \ - if (likely(pg[H1(i)] & 1)) { \ - target_ulong off = (target_ulong)(TYPEI)m[i] << scale; \ - FN(env, base + off, d[i], ra); \ - } \ - } \ +static void sve_st1_zd(CPUARMState *env, void *vd, void *vg, void *vm, + target_ulong base, uint32_t desc, uintptr_t ra, + zreg_off_fn *off_fn, sve_ld1_tlb_fn *tlb_fn) +{ + const int mmu_idx = cpu_mmu_index(env, false); + intptr_t i, oprsz = simd_oprsz(desc) / 8; + unsigned scale = simd_data(desc); + + set_helper_retaddr(ra); + for (i = 0; i < oprsz; i++) { + uint8_t pg = *(uint8_t *)(vg + H1(i)); + if (likely(pg & 1)) { + target_ulong off = off_fn(vm, i * 8); + tlb_fn(env, vd, i * 8, base + (off << scale), mmu_idx, ra); + } + } + set_helper_retaddr(0); } -DO_ST1_ZPZ_S(sve_stbs_zsu, uint32_t, cpu_stb_data_ra) -DO_ST1_ZPZ_S(sve_sths_zsu, uint32_t, cpu_stw_data_ra) -DO_ST1_ZPZ_S(sve_stss_zsu, uint32_t, cpu_stl_data_ra) +#define DO_ST1_ZPZ_S(MEM, OFS) \ +void __attribute__((flatten)) HELPER(sve_st##MEM##_##OFS) \ + (CPUARMState *env, void *vd, void *vg, void *vm, \ + target_ulong base, uint32_t desc) \ +{ \ + sve_st1_zs(env, vd, vg, vm, base, desc, GETPC(), \ + off_##OFS##_s, sve_st1##MEM##_tlb); \ +} -DO_ST1_ZPZ_S(sve_stbs_zss, int32_t, cpu_stb_data_ra) -DO_ST1_ZPZ_S(sve_sths_zss, int32_t, cpu_stw_data_ra) -DO_ST1_ZPZ_S(sve_stss_zss, int32_t, cpu_stl_data_ra) +#define DO_ST1_ZPZ_D(MEM, OFS) \ +void __attribute__((flatten)) HELPER(sve_st##MEM##_##OFS) \ + (CPUARMState *env, void *vd, void *vg, void *vm, \ + target_ulong base, uint32_t desc) \ +{ \ + sve_st1_zd(env, vd, vg, vm, base, desc, GETPC(), \ + off_##OFS##_d, sve_st1##MEM##_tlb); \ +} -DO_ST1_ZPZ_D(sve_stbd_zsu, uint32_t, cpu_stb_data_ra) -DO_ST1_ZPZ_D(sve_sthd_zsu, uint32_t, cpu_stw_data_ra) -DO_ST1_ZPZ_D(sve_stsd_zsu, uint32_t, cpu_stl_data_ra) -DO_ST1_ZPZ_D(sve_stdd_zsu, uint32_t, cpu_stq_data_ra) +DO_ST1_ZPZ_S(bs, zsu) +DO_ST1_ZPZ_S(hs_le, zsu) +DO_ST1_ZPZ_S(hs_be, zsu) +DO_ST1_ZPZ_S(ss_le, zsu) +DO_ST1_ZPZ_S(ss_be, zsu) -DO_ST1_ZPZ_D(sve_stbd_zss, int32_t, cpu_stb_data_ra) -DO_ST1_ZPZ_D(sve_sthd_zss, int32_t, cpu_stw_data_ra) -DO_ST1_ZPZ_D(sve_stsd_zss, int32_t, cpu_stl_data_ra) -DO_ST1_ZPZ_D(sve_stdd_zss, int32_t, cpu_stq_data_ra) +DO_ST1_ZPZ_S(bs, zss) +DO_ST1_ZPZ_S(hs_le, zss) +DO_ST1_ZPZ_S(hs_be, zss) +DO_ST1_ZPZ_S(ss_le, zss) +DO_ST1_ZPZ_S(ss_be, zss) -DO_ST1_ZPZ_D(sve_stbd_zd, uint64_t, cpu_stb_data_ra) -DO_ST1_ZPZ_D(sve_sthd_zd, uint64_t, cpu_stw_data_ra) -DO_ST1_ZPZ_D(sve_stsd_zd, uint64_t, cpu_stl_data_ra) -DO_ST1_ZPZ_D(sve_stdd_zd, uint64_t, cpu_stq_data_ra) +DO_ST1_ZPZ_D(bd, zsu) +DO_ST1_ZPZ_D(hd_le, zsu) +DO_ST1_ZPZ_D(hd_be, zsu) +DO_ST1_ZPZ_D(sd_le, zsu) +DO_ST1_ZPZ_D(sd_be, zsu) +DO_ST1_ZPZ_D(dd_le, zsu) +DO_ST1_ZPZ_D(dd_be, zsu) + +DO_ST1_ZPZ_D(bd, zss) +DO_ST1_ZPZ_D(hd_le, zss) +DO_ST1_ZPZ_D(hd_be, zss) +DO_ST1_ZPZ_D(sd_le, zss) +DO_ST1_ZPZ_D(sd_be, zss) +DO_ST1_ZPZ_D(dd_le, zss) +DO_ST1_ZPZ_D(dd_be, zss) + +DO_ST1_ZPZ_D(bd, zd) +DO_ST1_ZPZ_D(hd_le, zd) +DO_ST1_ZPZ_D(hd_be, zd) +DO_ST1_ZPZ_D(sd_le, zd) +DO_ST1_ZPZ_D(sd_be, zd) +DO_ST1_ZPZ_D(dd_le, zd) +DO_ST1_ZPZ_D(dd_be, zd) + +#undef DO_ST1_ZPZ_S +#undef DO_ST1_ZPZ_D diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_LD1_zpiz(DisasContext *s, arg_LD1_zpiz *a, uint32_t insn) return true; } -/* Indexed by [xs][msz]. */ -static gen_helper_gvec_mem_scatter * const scatter_store_fn32[2][3] = { - { gen_helper_sve_stbs_zsu, - gen_helper_sve_sths_zsu, - gen_helper_sve_stss_zsu, }, - { gen_helper_sve_stbs_zss, - gen_helper_sve_sths_zss, - gen_helper_sve_stss_zss, }, +/* Indexed by [be][xs][msz]. */ +static gen_helper_gvec_mem_scatter * const scatter_store_fn32[2][2][3] = { + /* Little-endian */ + { { gen_helper_sve_stbs_zsu, + gen_helper_sve_sths_le_zsu, + gen_helper_sve_stss_le_zsu, }, + { gen_helper_sve_stbs_zss, + gen_helper_sve_sths_le_zss, + gen_helper_sve_stss_le_zss, } }, + /* Big-endian */ + { { gen_helper_sve_stbs_zsu, + gen_helper_sve_sths_be_zsu, + gen_helper_sve_stss_be_zsu, }, + { gen_helper_sve_stbs_zss, + gen_helper_sve_sths_be_zss, + gen_helper_sve_stss_be_zss, } }, }; /* Note that we overload xs=2 to indicate 64-bit offset. */ -static gen_helper_gvec_mem_scatter * const scatter_store_fn64[3][4] = { - { gen_helper_sve_stbd_zsu, - gen_helper_sve_sthd_zsu, - gen_helper_sve_stsd_zsu, - gen_helper_sve_stdd_zsu, }, - { gen_helper_sve_stbd_zss, - gen_helper_sve_sthd_zss, - gen_helper_sve_stsd_zss, - gen_helper_sve_stdd_zss, }, - { gen_helper_sve_stbd_zd, - gen_helper_sve_sthd_zd, - gen_helper_sve_stsd_zd, - gen_helper_sve_stdd_zd, }, +static gen_helper_gvec_mem_scatter * const scatter_store_fn64[2][3][4] = { + /* Little-endian */ + { { gen_helper_sve_stbd_zsu, + gen_helper_sve_sthd_le_zsu, + gen_helper_sve_stsd_le_zsu, + gen_helper_sve_stdd_le_zsu, }, + { gen_helper_sve_stbd_zss, + gen_helper_sve_sthd_le_zss, + gen_helper_sve_stsd_le_zss, + gen_helper_sve_stdd_le_zss, }, + { gen_helper_sve_stbd_zd, + gen_helper_sve_sthd_le_zd, + gen_helper_sve_stsd_le_zd, + gen_helper_sve_stdd_le_zd, } }, + /* Big-endian */ + { { gen_helper_sve_stbd_zsu, + gen_helper_sve_sthd_be_zsu, + gen_helper_sve_stsd_be_zsu, + gen_helper_sve_stdd_be_zsu, }, + { gen_helper_sve_stbd_zss, + gen_helper_sve_sthd_be_zss, + gen_helper_sve_stsd_be_zss, + gen_helper_sve_stdd_be_zss, }, + { gen_helper_sve_stbd_zd, + gen_helper_sve_sthd_be_zd, + gen_helper_sve_stsd_be_zd, + gen_helper_sve_stdd_be_zd, } }, }; static bool trans_ST1_zprz(DisasContext *s, arg_ST1_zprz *a, uint32_t insn) { gen_helper_gvec_mem_scatter *fn; + int be = s->be_data == MO_BE; if (a->esz < a->msz || (a->msz == 0 && a->scale)) { return false; @@ -XXX,XX +XXX,XX @@ static bool trans_ST1_zprz(DisasContext *s, arg_ST1_zprz *a, uint32_t insn) } switch (a->esz) { case MO_32: - fn = scatter_store_fn32[a->xs][a->msz]; + fn = scatter_store_fn32[be][a->xs][a->msz]; break; case MO_64: - fn = scatter_store_fn64[a->xs][a->msz]; + fn = scatter_store_fn64[be][a->xs][a->msz]; break; default: g_assert_not_reached(); @@ -XXX,XX +XXX,XX @@ static bool trans_ST1_zprz(DisasContext *s, arg_ST1_zprz *a, uint32_t insn) static bool trans_ST1_zpiz(DisasContext *s, arg_ST1_zpiz *a, uint32_t insn) { gen_helper_gvec_mem_scatter *fn = NULL; + int be = s->be_data == MO_BE; TCGv_i64 imm; if (a->esz < a->msz) { @@ -XXX,XX +XXX,XX @@ static bool trans_ST1_zpiz(DisasContext *s, arg_ST1_zpiz *a, uint32_t insn) switch (a->esz) { case MO_32: - fn = scatter_store_fn32[0][a->msz]; + fn = scatter_store_fn32[be][0][a->msz]; break; case MO_64: - fn = scatter_store_fn64[2][a->msz]; + fn = scatter_store_fn64[be][2][a->msz]; break; } assert(fn != NULL); -- 2.19.0 From: Richard Henderson <richard.henderson@linaro.org> This implements the feature for softmmu, and moves the main loop out of a macro and into a function. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181005175350.30752-15-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sve.h | 84 ++++++++--- target/arm/sve_helper.c | 290 +++++++++++++++++++++++++++---------- target/arm/translate-sve.c | 84 +++++------ 3 files changed, 321 insertions(+), 137 deletions(-) diff --git a/target/arm/helper-sve.h b/target/arm/helper-sve.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sve.h +++ b/target/arm/helper-sve.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_6(sve_ldsds_be_zd, TCG_CALL_NO_WG, DEF_HELPER_FLAGS_6(sve_ldffbsu_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffhsu_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhsu_le_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffssu_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhsu_be_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffss_le_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffss_be_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldffbss_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffhss_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhss_le_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhss_be_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldffbsu_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffhsu_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhsu_le_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffssu_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhsu_be_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffss_le_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffss_be_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldffbss_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffhss_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhss_le_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffhss_be_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldffbdu_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffhdu_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhdu_le_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffsdu_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhdu_be_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffddu_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffsdu_le_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsdu_be_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_le_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_be_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldffbds_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffhds_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhds_le_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffsds_zsu, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhds_be_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_le_zsu, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_be_zsu, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldffbdu_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffhdu_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhdu_le_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffsdu_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhdu_be_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffddu_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffsdu_le_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsdu_be_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_le_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_be_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldffbds_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffhds_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhds_le_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffsds_zss, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhds_be_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_le_zss, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_be_zss, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldffbdu_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffhdu_zd, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhdu_le_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffsdu_zd, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhdu_be_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffddu_zd, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffsdu_le_zd, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsdu_be_zd, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_le_zd, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffdd_be_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_ldffbds_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffhds_zd, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhds_le_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) -DEF_HELPER_FLAGS_6(sve_ldffsds_zd, TCG_CALL_NO_WG, +DEF_HELPER_FLAGS_6(sve_ldffhds_be_zd, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_le_zd, TCG_CALL_NO_WG, + void, env, ptr, ptr, ptr, tl, i32) +DEF_HELPER_FLAGS_6(sve_ldffsds_be_zd, TCG_CALL_NO_WG, void, env, ptr, ptr, ptr, tl, i32) DEF_HELPER_FLAGS_6(sve_stbs_zsu, TCG_CALL_NO_WG, diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ DO_LD1_ZPZ_D(dd_be, zd) /* First fault loads with a vector index. */ -#ifdef CONFIG_USER_ONLY +/* Load one element into VD+REG_OFF from (ENV,VADDR) without faulting. + * The controlling predicate is known to be true. Return true if the + * load was successful. + */ +typedef bool sve_ld1_nf_fn(CPUARMState *env, void *vd, intptr_t reg_off, + target_ulong vaddr, int mmu_idx); -#define DO_LDFF1_ZPZ(NAME, TYPEE, TYPEI, TYPEM, FN, H) \ -void HELPER(NAME)(CPUARMState *env, void *vd, void *vg, void *vm, \ - target_ulong base, uint32_t desc) \ -{ \ - intptr_t i, oprsz = simd_oprsz(desc); \ - unsigned scale = simd_data(desc); \ - uintptr_t ra = GETPC(); \ - bool first = true; \ - mmap_lock(); \ - for (i = 0; i < oprsz; ) { \ - uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); \ - do { \ - TYPEM m = 0; \ - if (pg & 1) { \ - target_ulong off = *(TYPEI *)(vm + H(i)); \ - target_ulong addr = base + (off << scale); \ - if (!first && \ - page_check_range(addr, sizeof(TYPEM), PAGE_READ)) { \ - record_fault(env, i, oprsz); \ - goto exit; \ - } \ - m = FN(env, addr, ra); \ - first = false; \ - } \ - *(TYPEE *)(vd + H(i)) = m; \ - i += sizeof(TYPEE), pg >>= sizeof(TYPEE); \ - } while (i & 15); \ - } \ - exit: \ - mmap_unlock(); \ +#ifdef CONFIG_SOFTMMU +#define DO_LD_NF(NAME, H, TYPEE, TYPEM, HOST) \ +static bool sve_ld##NAME##_nf(CPUARMState *env, void *vd, intptr_t reg_off, \ + target_ulong addr, int mmu_idx) \ +{ \ + target_ulong next_page = -(addr | TARGET_PAGE_MASK); \ + if (likely(next_page - addr >= sizeof(TYPEM))) { \ + void *host = tlb_vaddr_to_host(env, addr, MMU_DATA_LOAD, mmu_idx); \ + if (likely(host)) { \ + TYPEM val = HOST(host); \ + *(TYPEE *)(vd + H(reg_off)) = val; \ + return true; \ + } \ + } \ + return false; \ } - #else - -#define DO_LDFF1_ZPZ(NAME, TYPEE, TYPEI, TYPEM, FN, H) \ -void HELPER(NAME)(CPUARMState *env, void *vd, void *vg, void *vm, \ - target_ulong base, uint32_t desc) \ -{ \ - g_assert_not_reached(); \ +#define DO_LD_NF(NAME, H, TYPEE, TYPEM, HOST) \ +static bool sve_ld##NAME##_nf(CPUARMState *env, void *vd, intptr_t reg_off, \ + target_ulong addr, int mmu_idx) \ +{ \ + if (likely(page_check_range(addr, sizeof(TYPEM), PAGE_READ))) { \ + TYPEM val = HOST(g2h(addr)); \ + *(TYPEE *)(vd + H(reg_off)) = val; \ + return true; \ + } \ + return false; \ } - #endif -#define DO_LDFF1_ZPZ_S(NAME, TYPEI, TYPEM, FN) \ - DO_LDFF1_ZPZ(NAME, uint32_t, TYPEI, TYPEM, FN, H1_4) -#define DO_LDFF1_ZPZ_D(NAME, TYPEI, TYPEM, FN) \ - DO_LDFF1_ZPZ(NAME, uint64_t, TYPEI, TYPEM, FN, ) +DO_LD_NF(bsu, H1_4, uint32_t, uint8_t, ldub_p) +DO_LD_NF(bss, H1_4, uint32_t, int8_t, ldsb_p) +DO_LD_NF(bdu, , uint64_t, uint8_t, ldub_p) +DO_LD_NF(bds, , uint64_t, int8_t, ldsb_p) -DO_LDFF1_ZPZ_S(sve_ldffbsu_zsu, uint32_t, uint8_t, cpu_ldub_data_ra) -DO_LDFF1_ZPZ_S(sve_ldffhsu_zsu, uint32_t, uint16_t, cpu_lduw_data_ra) -DO_LDFF1_ZPZ_S(sve_ldffssu_zsu, uint32_t, uint32_t, cpu_ldl_data_ra) -DO_LDFF1_ZPZ_S(sve_ldffbss_zsu, uint32_t, int8_t, cpu_ldub_data_ra) -DO_LDFF1_ZPZ_S(sve_ldffhss_zsu, uint32_t, int16_t, cpu_lduw_data_ra) +DO_LD_NF(hsu_le, H1_4, uint32_t, uint16_t, lduw_le_p) +DO_LD_NF(hss_le, H1_4, uint32_t, int16_t, ldsw_le_p) +DO_LD_NF(hsu_be, H1_4, uint32_t, uint16_t, lduw_be_p) +DO_LD_NF(hss_be, H1_4, uint32_t, int16_t, ldsw_be_p) +DO_LD_NF(hdu_le, , uint64_t, uint16_t, lduw_le_p) +DO_LD_NF(hds_le, , uint64_t, int16_t, ldsw_le_p) +DO_LD_NF(hdu_be, , uint64_t, uint16_t, lduw_be_p) +DO_LD_NF(hds_be, , uint64_t, int16_t, ldsw_be_p) -DO_LDFF1_ZPZ_S(sve_ldffbsu_zss, int32_t, uint8_t, cpu_ldub_data_ra) -DO_LDFF1_ZPZ_S(sve_ldffhsu_zss, int32_t, uint16_t, cpu_lduw_data_ra) -DO_LDFF1_ZPZ_S(sve_ldffssu_zss, int32_t, uint32_t, cpu_ldl_data_ra) -DO_LDFF1_ZPZ_S(sve_ldffbss_zss, int32_t, int8_t, cpu_ldub_data_ra) -DO_LDFF1_ZPZ_S(sve_ldffhss_zss, int32_t, int16_t, cpu_lduw_data_ra) +DO_LD_NF(ss_le, H1_4, uint32_t, uint32_t, ldl_le_p) +DO_LD_NF(ss_be, H1_4, uint32_t, uint32_t, ldl_be_p) +DO_LD_NF(sdu_le, , uint64_t, uint32_t, ldl_le_p) +DO_LD_NF(sds_le, , uint64_t, int32_t, ldl_le_p) +DO_LD_NF(sdu_be, , uint64_t, uint32_t, ldl_be_p) +DO_LD_NF(sds_be, , uint64_t, int32_t, ldl_be_p) -DO_LDFF1_ZPZ_D(sve_ldffbdu_zsu, uint32_t, uint8_t, cpu_ldub_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffhdu_zsu, uint32_t, uint16_t, cpu_lduw_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffsdu_zsu, uint32_t, uint32_t, cpu_ldl_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffddu_zsu, uint32_t, uint64_t, cpu_ldq_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffbds_zsu, uint32_t, int8_t, cpu_ldub_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffhds_zsu, uint32_t, int16_t, cpu_lduw_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffsds_zsu, uint32_t, int32_t, cpu_ldl_data_ra) +DO_LD_NF(dd_le, , uint64_t, uint64_t, ldq_le_p) +DO_LD_NF(dd_be, , uint64_t, uint64_t, ldq_be_p) -DO_LDFF1_ZPZ_D(sve_ldffbdu_zss, int32_t, uint8_t, cpu_ldub_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffhdu_zss, int32_t, uint16_t, cpu_lduw_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffsdu_zss, int32_t, uint32_t, cpu_ldl_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffddu_zss, int32_t, uint64_t, cpu_ldq_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffbds_zss, int32_t, int8_t, cpu_ldub_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffhds_zss, int32_t, int16_t, cpu_lduw_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffsds_zss, int32_t, int32_t, cpu_ldl_data_ra) +/* + * Common helper for all gather first-faulting loads. + */ +static inline void sve_ldff1_zs(CPUARMState *env, void *vd, void *vg, void *vm, + target_ulong base, uint32_t desc, uintptr_t ra, + zreg_off_fn *off_fn, sve_ld1_tlb_fn *tlb_fn, + sve_ld1_nf_fn *nonfault_fn) +{ + const int mmu_idx = cpu_mmu_index(env, false); + intptr_t reg_off, reg_max = simd_oprsz(desc); + unsigned scale = simd_data(desc); + target_ulong addr; -DO_LDFF1_ZPZ_D(sve_ldffbdu_zd, uint64_t, uint8_t, cpu_ldub_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffhdu_zd, uint64_t, uint16_t, cpu_lduw_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffsdu_zd, uint64_t, uint32_t, cpu_ldl_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffddu_zd, uint64_t, uint64_t, cpu_ldq_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffbds_zd, uint64_t, int8_t, cpu_ldub_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffhds_zd, uint64_t, int16_t, cpu_lduw_data_ra) -DO_LDFF1_ZPZ_D(sve_ldffsds_zd, uint64_t, int32_t, cpu_ldl_data_ra) + /* Skip to the first true predicate. */ + reg_off = find_next_active(vg, 0, reg_max, MO_32); + if (likely(reg_off < reg_max)) { + /* Perform one normal read, which will fault or not. */ + set_helper_retaddr(ra); + addr = off_fn(vm, reg_off); + addr = base + (addr << scale); + tlb_fn(env, vd, reg_off, addr, mmu_idx, ra); + + /* The rest of the reads will be non-faulting. */ + set_helper_retaddr(0); + } + + /* After any fault, zero the leading predicated false elements. */ + swap_memzero(vd, reg_off); + + while (likely((reg_off += 4) < reg_max)) { + uint64_t pg = *(uint64_t *)(vg + (reg_off >> 6) * 8); + if (likely((pg >> (reg_off & 63)) & 1)) { + addr = off_fn(vm, reg_off); + addr = base + (addr << scale); + if (!nonfault_fn(env, vd, reg_off, addr, mmu_idx)) { + record_fault(env, reg_off, reg_max); + break; + } + } else { + *(uint32_t *)(vd + H1_4(reg_off)) = 0; + } + } +} + +static inline void sve_ldff1_zd(CPUARMState *env, void *vd, void *vg, void *vm, + target_ulong base, uint32_t desc, uintptr_t ra, + zreg_off_fn *off_fn, sve_ld1_tlb_fn *tlb_fn, + sve_ld1_nf_fn *nonfault_fn) +{ + const int mmu_idx = cpu_mmu_index(env, false); + intptr_t reg_off, reg_max = simd_oprsz(desc); + unsigned scale = simd_data(desc); + target_ulong addr; + + /* Skip to the first true predicate. */ + reg_off = find_next_active(vg, 0, reg_max, MO_64); + if (likely(reg_off < reg_max)) { + /* Perform one normal read, which will fault or not. */ + set_helper_retaddr(ra); + addr = off_fn(vm, reg_off); + addr = base + (addr << scale); + tlb_fn(env, vd, reg_off, addr, mmu_idx, ra); + + /* The rest of the reads will be non-faulting. */ + set_helper_retaddr(0); + } + + /* After any fault, zero the leading predicated false elements. */ + swap_memzero(vd, reg_off); + + while (likely((reg_off += 8) < reg_max)) { + uint8_t pg = *(uint8_t *)(vg + H1(reg_off >> 3)); + if (likely(pg & 1)) { + addr = off_fn(vm, reg_off); + addr = base + (addr << scale); + if (!nonfault_fn(env, vd, reg_off, addr, mmu_idx)) { + record_fault(env, reg_off, reg_max); + break; + } + } else { + *(uint64_t *)(vd + reg_off) = 0; + } + } +} + +#define DO_LDFF1_ZPZ_S(MEM, OFS) \ +void HELPER(sve_ldff##MEM##_##OFS) \ + (CPUARMState *env, void *vd, void *vg, void *vm, \ + target_ulong base, uint32_t desc) \ +{ \ + sve_ldff1_zs(env, vd, vg, vm, base, desc, GETPC(), \ + off_##OFS##_s, sve_ld1##MEM##_tlb, sve_ld##MEM##_nf); \ +} + +#define DO_LDFF1_ZPZ_D(MEM, OFS) \ +void HELPER(sve_ldff##MEM##_##OFS) \ + (CPUARMState *env, void *vd, void *vg, void *vm, \ + target_ulong base, uint32_t desc) \ +{ \ + sve_ldff1_zd(env, vd, vg, vm, base, desc, GETPC(), \ + off_##OFS##_d, sve_ld1##MEM##_tlb, sve_ld##MEM##_nf); \ +} + +DO_LDFF1_ZPZ_S(bsu, zsu) +DO_LDFF1_ZPZ_S(bsu, zss) +DO_LDFF1_ZPZ_D(bdu, zsu) +DO_LDFF1_ZPZ_D(bdu, zss) +DO_LDFF1_ZPZ_D(bdu, zd) + +DO_LDFF1_ZPZ_S(bss, zsu) +DO_LDFF1_ZPZ_S(bss, zss) +DO_LDFF1_ZPZ_D(bds, zsu) +DO_LDFF1_ZPZ_D(bds, zss) +DO_LDFF1_ZPZ_D(bds, zd) + +DO_LDFF1_ZPZ_S(hsu_le, zsu) +DO_LDFF1_ZPZ_S(hsu_le, zss) +DO_LDFF1_ZPZ_D(hdu_le, zsu) +DO_LDFF1_ZPZ_D(hdu_le, zss) +DO_LDFF1_ZPZ_D(hdu_le, zd) + +DO_LDFF1_ZPZ_S(hsu_be, zsu) +DO_LDFF1_ZPZ_S(hsu_be, zss) +DO_LDFF1_ZPZ_D(hdu_be, zsu) +DO_LDFF1_ZPZ_D(hdu_be, zss) +DO_LDFF1_ZPZ_D(hdu_be, zd) + +DO_LDFF1_ZPZ_S(hss_le, zsu) +DO_LDFF1_ZPZ_S(hss_le, zss) +DO_LDFF1_ZPZ_D(hds_le, zsu) +DO_LDFF1_ZPZ_D(hds_le, zss) +DO_LDFF1_ZPZ_D(hds_le, zd) + +DO_LDFF1_ZPZ_S(hss_be, zsu) +DO_LDFF1_ZPZ_S(hss_be, zss) +DO_LDFF1_ZPZ_D(hds_be, zsu) +DO_LDFF1_ZPZ_D(hds_be, zss) +DO_LDFF1_ZPZ_D(hds_be, zd) + +DO_LDFF1_ZPZ_S(ss_le, zsu) +DO_LDFF1_ZPZ_S(ss_le, zss) +DO_LDFF1_ZPZ_D(sdu_le, zsu) +DO_LDFF1_ZPZ_D(sdu_le, zss) +DO_LDFF1_ZPZ_D(sdu_le, zd) + +DO_LDFF1_ZPZ_S(ss_be, zsu) +DO_LDFF1_ZPZ_S(ss_be, zss) +DO_LDFF1_ZPZ_D(sdu_be, zsu) +DO_LDFF1_ZPZ_D(sdu_be, zss) +DO_LDFF1_ZPZ_D(sdu_be, zd) + +DO_LDFF1_ZPZ_D(sds_le, zsu) +DO_LDFF1_ZPZ_D(sds_le, zss) +DO_LDFF1_ZPZ_D(sds_le, zd) + +DO_LDFF1_ZPZ_D(sds_be, zsu) +DO_LDFF1_ZPZ_D(sds_be, zss) +DO_LDFF1_ZPZ_D(sds_be, zd) + +DO_LDFF1_ZPZ_D(dd_le, zsu) +DO_LDFF1_ZPZ_D(dd_le, zss) +DO_LDFF1_ZPZ_D(dd_le, zd) + +DO_LDFF1_ZPZ_D(dd_be, zsu) +DO_LDFF1_ZPZ_D(dd_be, zss) +DO_LDFF1_ZPZ_D(dd_be, zd) /* Stores with a vector index. */ diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static gen_helper_gvec_mem_scatter * const gather_load_fn32[2][2][2][2][3] = { /* First-fault */ { { { gen_helper_sve_ldffbss_zsu, - gen_helper_sve_ldffhss_zsu, + gen_helper_sve_ldffhss_le_zsu, NULL, }, { gen_helper_sve_ldffbsu_zsu, - gen_helper_sve_ldffhsu_zsu, - gen_helper_sve_ldffssu_zsu, } }, + gen_helper_sve_ldffhsu_le_zsu, + gen_helper_sve_ldffss_le_zsu, } }, { { gen_helper_sve_ldffbss_zss, - gen_helper_sve_ldffhss_zss, + gen_helper_sve_ldffhss_le_zss, NULL, }, { gen_helper_sve_ldffbsu_zss, - gen_helper_sve_ldffhsu_zss, - gen_helper_sve_ldffssu_zss, } } } }, + gen_helper_sve_ldffhsu_le_zss, + gen_helper_sve_ldffss_le_zss, } } } }, /* Big-endian */ { { { { gen_helper_sve_ldbss_zsu, @@ -XXX,XX +XXX,XX @@ static gen_helper_gvec_mem_scatter * const gather_load_fn32[2][2][2][2][3] = { /* First-fault */ { { { gen_helper_sve_ldffbss_zsu, - gen_helper_sve_ldffhss_zsu, + gen_helper_sve_ldffhss_be_zsu, NULL, }, { gen_helper_sve_ldffbsu_zsu, - gen_helper_sve_ldffhsu_zsu, - gen_helper_sve_ldffssu_zsu, } }, + gen_helper_sve_ldffhsu_be_zsu, + gen_helper_sve_ldffss_be_zsu, } }, { { gen_helper_sve_ldffbss_zss, - gen_helper_sve_ldffhss_zss, + gen_helper_sve_ldffhss_be_zss, NULL, }, { gen_helper_sve_ldffbsu_zss, - gen_helper_sve_ldffhsu_zss, - gen_helper_sve_ldffssu_zss, } } } }, + gen_helper_sve_ldffhsu_be_zss, + gen_helper_sve_ldffss_be_zss, } } } }, }; /* Note that we overload xs=2 to indicate 64-bit offset. */ @@ -XXX,XX +XXX,XX @@ static gen_helper_gvec_mem_scatter * const gather_load_fn64[2][2][3][2][4] = { /* First-fault */ { { { gen_helper_sve_ldffbds_zsu, - gen_helper_sve_ldffhds_zsu, - gen_helper_sve_ldffsds_zsu, + gen_helper_sve_ldffhds_le_zsu, + gen_helper_sve_ldffsds_le_zsu, NULL, }, { gen_helper_sve_ldffbdu_zsu, - gen_helper_sve_ldffhdu_zsu, - gen_helper_sve_ldffsdu_zsu, - gen_helper_sve_ldffddu_zsu, } }, + gen_helper_sve_ldffhdu_le_zsu, + gen_helper_sve_ldffsdu_le_zsu, + gen_helper_sve_ldffdd_le_zsu, } }, { { gen_helper_sve_ldffbds_zss, - gen_helper_sve_ldffhds_zss, - gen_helper_sve_ldffsds_zss, + gen_helper_sve_ldffhds_le_zss, + gen_helper_sve_ldffsds_le_zss, NULL, }, { gen_helper_sve_ldffbdu_zss, - gen_helper_sve_ldffhdu_zss, - gen_helper_sve_ldffsdu_zss, - gen_helper_sve_ldffddu_zss, } }, + gen_helper_sve_ldffhdu_le_zss, + gen_helper_sve_ldffsdu_le_zss, + gen_helper_sve_ldffdd_le_zss, } }, { { gen_helper_sve_ldffbds_zd, - gen_helper_sve_ldffhds_zd, - gen_helper_sve_ldffsds_zd, + gen_helper_sve_ldffhds_le_zd, + gen_helper_sve_ldffsds_le_zd, NULL, }, { gen_helper_sve_ldffbdu_zd, - gen_helper_sve_ldffhdu_zd, - gen_helper_sve_ldffsdu_zd, - gen_helper_sve_ldffddu_zd, } } } }, + gen_helper_sve_ldffhdu_le_zd, + gen_helper_sve_ldffsdu_le_zd, + gen_helper_sve_ldffdd_le_zd, } } } }, /* Big-endian */ { { { { gen_helper_sve_ldbds_zsu, @@ -XXX,XX +XXX,XX @@ static gen_helper_gvec_mem_scatter * const gather_load_fn64[2][2][3][2][4] = { /* First-fault */ { { { gen_helper_sve_ldffbds_zsu, - gen_helper_sve_ldffhds_zsu, - gen_helper_sve_ldffsds_zsu, + gen_helper_sve_ldffhds_be_zsu, + gen_helper_sve_ldffsds_be_zsu, NULL, }, { gen_helper_sve_ldffbdu_zsu, - gen_helper_sve_ldffhdu_zsu, - gen_helper_sve_ldffsdu_zsu, - gen_helper_sve_ldffddu_zsu, } }, + gen_helper_sve_ldffhdu_be_zsu, + gen_helper_sve_ldffsdu_be_zsu, + gen_helper_sve_ldffdd_be_zsu, } }, { { gen_helper_sve_ldffbds_zss, - gen_helper_sve_ldffhds_zss, - gen_helper_sve_ldffsds_zss, + gen_helper_sve_ldffhds_be_zss, + gen_helper_sve_ldffsds_be_zss, NULL, }, { gen_helper_sve_ldffbdu_zss, - gen_helper_sve_ldffhdu_zss, - gen_helper_sve_ldffsdu_zss, - gen_helper_sve_ldffddu_zss, } }, + gen_helper_sve_ldffhdu_be_zss, + gen_helper_sve_ldffsdu_be_zss, + gen_helper_sve_ldffdd_be_zss, } }, { { gen_helper_sve_ldffbds_zd, - gen_helper_sve_ldffhds_zd, - gen_helper_sve_ldffsds_zd, + gen_helper_sve_ldffhds_be_zd, + gen_helper_sve_ldffsds_be_zd, NULL, }, { gen_helper_sve_ldffbdu_zd, - gen_helper_sve_ldffhdu_zd, - gen_helper_sve_ldffsdu_zd, - gen_helper_sve_ldffddu_zd, } } } }, + gen_helper_sve_ldffhdu_be_zd, + gen_helper_sve_ldffsdu_be_zd, + gen_helper_sve_ldffdd_be_zd, } } } }, }; static bool trans_LD1_zprz(DisasContext *s, arg_LD1_zprz *a, uint32_t insn) -- 2.19.0 From: Richard Henderson <richard.henderson@linaro.org> There is quite a lot of code required to compute cpu_mem_index, or even put together the full TCGMemOpIdx. This can easily be done at translation time. Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181005175350.30752-16-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/internals.h | 5 ++ target/arm/sve_helper.c | 138 +++++++++++++++++++------------------ target/arm/translate-sve.c | 67 +++++++++++------- 3 files changed, 121 insertions(+), 89 deletions(-) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ static inline uint32_t arm_debug_exception_fsr(CPUARMState *env) } } +/* Note make_memop_idx reserves 4 bits for mmu_idx, and MO_BSWAP is bit 3. + * Thus a TCGMemOpIdx, without any MO_ALIGN bits, fits in 8 bits. + */ +#define MEMOPIDX_SHIFT 8 + #endif diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "cpu.h" +#include "internals.h" #include "exec/exec-all.h" #include "exec/cpu_ldst.h" #include "exec/helper-proto.h" @@ -XXX,XX +XXX,XX @@ typedef intptr_t sve_ld1_host_fn(void *vd, void *vg, void *host, * The controlling predicate is known to be true. */ typedef void sve_ld1_tlb_fn(CPUARMState *env, void *vd, intptr_t reg_off, - target_ulong vaddr, int mmu_idx, uintptr_t ra); + target_ulong vaddr, TCGMemOpIdx oi, uintptr_t ra); typedef sve_ld1_tlb_fn sve_st1_tlb_fn; /* @@ -XXX,XX +XXX,XX @@ static intptr_t sve_##NAME##_host(void *vd, void *vg, void *host, \ #ifdef CONFIG_SOFTMMU #define DO_LD_TLB(NAME, H, TYPEE, TYPEM, HOST, MOEND, TLB) \ static void sve_##NAME##_tlb(CPUARMState *env, void *vd, intptr_t reg_off, \ - target_ulong addr, int mmu_idx, uintptr_t ra) \ + target_ulong addr, TCGMemOpIdx oi, uintptr_t ra) \ { \ - TCGMemOpIdx oi = make_memop_idx(ctz32(sizeof(TYPEM)) | MOEND, mmu_idx); \ TYPEM val = TLB(env, addr, oi, ra); \ *(TYPEE *)(vd + H(reg_off)) = val; \ } #else #define DO_LD_TLB(NAME, H, TYPEE, TYPEM, HOST, MOEND, TLB) \ static void sve_##NAME##_tlb(CPUARMState *env, void *vd, intptr_t reg_off, \ - target_ulong addr, int mmu_idx, uintptr_t ra) \ + target_ulong addr, TCGMemOpIdx oi, uintptr_t ra) \ { \ TYPEM val = HOST(g2h(addr)); \ *(TYPEE *)(vd + H(reg_off)) = val; \ @@ -XXX,XX +XXX,XX @@ static void sve_ld1_r(CPUARMState *env, void *vg, const target_ulong addr, sve_ld1_host_fn *host_fn, sve_ld1_tlb_fn *tlb_fn) { - void *vd = &env->vfp.zregs[simd_data(desc)]; + const TCGMemOpIdx oi = extract32(desc, SIMD_DATA_SHIFT, MEMOPIDX_SHIFT); + const int mmu_idx = get_mmuidx(oi); + const unsigned rd = extract32(desc, SIMD_DATA_SHIFT + MEMOPIDX_SHIFT, 5); + void *vd = &env->vfp.zregs[rd]; const int diffsz = esz - msz; const intptr_t reg_max = simd_oprsz(desc); const intptr_t mem_max = reg_max >> diffsz; - const int mmu_idx = cpu_mmu_index(env, false); ARMVectorReg scratch; void *host; intptr_t split, reg_off, mem_off; @@ -XXX,XX +XXX,XX @@ static void sve_ld1_r(CPUARMState *env, void *vg, const target_ulong addr, * on I/O memory, it may succeed but not bring in the TLB entry. * But even then we have still made forward progress. */ - tlb_fn(env, &scratch, reg_off, addr + mem_off, mmu_idx, retaddr); + tlb_fn(env, &scratch, reg_off, addr + mem_off, oi, retaddr); reg_off += 1 << esz; } #endif @@ -XXX,XX +XXX,XX @@ static void sve_ld2_r(CPUARMState *env, void *vg, target_ulong addr, uint32_t desc, int size, uintptr_t ra, sve_ld1_tlb_fn *tlb_fn) { - const int mmu_idx = cpu_mmu_index(env, false); + const TCGMemOpIdx oi = extract32(desc, SIMD_DATA_SHIFT, MEMOPIDX_SHIFT); + const unsigned rd = extract32(desc, SIMD_DATA_SHIFT + MEMOPIDX_SHIFT, 5); intptr_t i, oprsz = simd_oprsz(desc); - unsigned rd = simd_data(desc); ARMVectorReg scratch[2] = { }; set_helper_retaddr(ra); @@ -XXX,XX +XXX,XX @@ static void sve_ld2_r(CPUARMState *env, void *vg, target_ulong addr, uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); do { if (pg & 1) { - tlb_fn(env, &scratch[0], i, addr, mmu_idx, ra); - tlb_fn(env, &scratch[1], i, addr + size, mmu_idx, ra); + tlb_fn(env, &scratch[0], i, addr, oi, ra); + tlb_fn(env, &scratch[1], i, addr + size, oi, ra); } i += size, pg >>= size; addr += 2 * size; @@ -XXX,XX +XXX,XX @@ static void sve_ld3_r(CPUARMState *env, void *vg, target_ulong addr, uint32_t desc, int size, uintptr_t ra, sve_ld1_tlb_fn *tlb_fn) { - const int mmu_idx = cpu_mmu_index(env, false); + const TCGMemOpIdx oi = extract32(desc, SIMD_DATA_SHIFT, MEMOPIDX_SHIFT); + const unsigned rd = extract32(desc, SIMD_DATA_SHIFT + MEMOPIDX_SHIFT, 5); intptr_t i, oprsz = simd_oprsz(desc); - unsigned rd = simd_data(desc); ARMVectorReg scratch[3] = { }; set_helper_retaddr(ra); @@ -XXX,XX +XXX,XX @@ static void sve_ld3_r(CPUARMState *env, void *vg, target_ulong addr, uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); do { if (pg & 1) { - tlb_fn(env, &scratch[0], i, addr, mmu_idx, ra); - tlb_fn(env, &scratch[1], i, addr + size, mmu_idx, ra); - tlb_fn(env, &scratch[2], i, addr + 2 * size, mmu_idx, ra); + tlb_fn(env, &scratch[0], i, addr, oi, ra); + tlb_fn(env, &scratch[1], i, addr + size, oi, ra); + tlb_fn(env, &scratch[2], i, addr + 2 * size, oi, ra); } i += size, pg >>= size; addr += 3 * size; @@ -XXX,XX +XXX,XX @@ static void sve_ld4_r(CPUARMState *env, void *vg, target_ulong addr, uint32_t desc, int size, uintptr_t ra, sve_ld1_tlb_fn *tlb_fn) { - const int mmu_idx = cpu_mmu_index(env, false); + const TCGMemOpIdx oi = extract32(desc, SIMD_DATA_SHIFT, MEMOPIDX_SHIFT); + const unsigned rd = extract32(desc, SIMD_DATA_SHIFT + MEMOPIDX_SHIFT, 5); intptr_t i, oprsz = simd_oprsz(desc); - unsigned rd = simd_data(desc); ARMVectorReg scratch[4] = { }; set_helper_retaddr(ra); @@ -XXX,XX +XXX,XX @@ static void sve_ld4_r(CPUARMState *env, void *vg, target_ulong addr, uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); do { if (pg & 1) { - tlb_fn(env, &scratch[0], i, addr, mmu_idx, ra); - tlb_fn(env, &scratch[1], i, addr + size, mmu_idx, ra); - tlb_fn(env, &scratch[2], i, addr + 2 * size, mmu_idx, ra); - tlb_fn(env, &scratch[3], i, addr + 3 * size, mmu_idx, ra); + tlb_fn(env, &scratch[0], i, addr, oi, ra); + tlb_fn(env, &scratch[1], i, addr + size, oi, ra); + tlb_fn(env, &scratch[2], i, addr + 2 * size, oi, ra); + tlb_fn(env, &scratch[3], i, addr + 3 * size, oi, ra); } i += size, pg >>= size; addr += 4 * size; @@ -XXX,XX +XXX,XX @@ static void sve_ldff1_r(CPUARMState *env, void *vg, const target_ulong addr, sve_ld1_host_fn *host_fn, sve_ld1_tlb_fn *tlb_fn) { - void *vd = &env->vfp.zregs[simd_data(desc)]; + const TCGMemOpIdx oi = extract32(desc, SIMD_DATA_SHIFT, MEMOPIDX_SHIFT); + const int mmu_idx = get_mmuidx(oi); + const unsigned rd = extract32(desc, SIMD_DATA_SHIFT + MEMOPIDX_SHIFT, 5); + void *vd = &env->vfp.zregs[rd]; const int diffsz = esz - msz; const intptr_t reg_max = simd_oprsz(desc); const intptr_t mem_max = reg_max >> diffsz; - const int mmu_idx = cpu_mmu_index(env, false); intptr_t split, reg_off, mem_off; void *host; @@ -XXX,XX +XXX,XX @@ static void sve_ldff1_r(CPUARMState *env, void *vg, const target_ulong addr, * Perform one normal read, which will fault or not. * But it is likely to bring the page into the tlb. */ - tlb_fn(env, vd, reg_off, addr + mem_off, mmu_idx, retaddr); + tlb_fn(env, vd, reg_off, addr + mem_off, oi, retaddr); /* After any fault, zero any leading predicated false elts. */ swap_memzero(vd, reg_off); @@ -XXX,XX +XXX,XX @@ static void sve_ldnf1_r(CPUARMState *env, void *vg, const target_ulong addr, uint32_t desc, const int esz, const int msz, sve_ld1_host_fn *host_fn) { - void *vd = &env->vfp.zregs[simd_data(desc)]; + const unsigned rd = extract32(desc, SIMD_DATA_SHIFT + MEMOPIDX_SHIFT, 5); + void *vd = &env->vfp.zregs[rd]; const int diffsz = esz - msz; const intptr_t reg_max = simd_oprsz(desc); const intptr_t mem_max = reg_max >> diffsz; @@ -XXX,XX +XXX,XX @@ DO_LDFF1_LDNF1_2(dd, 3, 3) #ifdef CONFIG_SOFTMMU #define DO_ST_TLB(NAME, H, TYPEM, HOST, MOEND, TLB) \ static void sve_##NAME##_tlb(CPUARMState *env, void *vd, intptr_t reg_off, \ - target_ulong addr, int mmu_idx, uintptr_t ra) \ + target_ulong addr, TCGMemOpIdx oi, uintptr_t ra) \ { \ - TCGMemOpIdx oi = make_memop_idx(ctz32(sizeof(TYPEM)) | MOEND, mmu_idx); \ TLB(env, addr, *(TYPEM *)(vd + H(reg_off)), oi, ra); \ } #else #define DO_ST_TLB(NAME, H, TYPEM, HOST, MOEND, TLB) \ static void sve_##NAME##_tlb(CPUARMState *env, void *vd, intptr_t reg_off, \ - target_ulong addr, int mmu_idx, uintptr_t ra) \ + target_ulong addr, TCGMemOpIdx oi, uintptr_t ra) \ { \ HOST(g2h(addr), *(TYPEM *)(vd + H(reg_off))); \ } @@ -XXX,XX +XXX,XX @@ static void sve_st1_r(CPUARMState *env, void *vg, target_ulong addr, const int esize, const int msize, sve_st1_tlb_fn *tlb_fn) { - const int mmu_idx = cpu_mmu_index(env, false); + const TCGMemOpIdx oi = extract32(desc, SIMD_DATA_SHIFT, MEMOPIDX_SHIFT); + const unsigned rd = extract32(desc, SIMD_DATA_SHIFT + MEMOPIDX_SHIFT, 5); intptr_t i, oprsz = simd_oprsz(desc); - unsigned rd = simd_data(desc); void *vd = &env->vfp.zregs[rd]; set_helper_retaddr(ra); @@ -XXX,XX +XXX,XX @@ static void sve_st1_r(CPUARMState *env, void *vg, target_ulong addr, uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); do { if (pg & 1) { - tlb_fn(env, vd, i, addr, mmu_idx, ra); + tlb_fn(env, vd, i, addr, oi, ra); } i += esize, pg >>= esize; addr += msize; @@ -XXX,XX +XXX,XX @@ static void sve_st2_r(CPUARMState *env, void *vg, target_ulong addr, const int esize, const int msize, sve_st1_tlb_fn *tlb_fn) { - const int mmu_idx = cpu_mmu_index(env, false); + const TCGMemOpIdx oi = extract32(desc, SIMD_DATA_SHIFT, MEMOPIDX_SHIFT); + const unsigned rd = extract32(desc, SIMD_DATA_SHIFT + MEMOPIDX_SHIFT, 5); intptr_t i, oprsz = simd_oprsz(desc); - unsigned rd = simd_data(desc); void *d1 = &env->vfp.zregs[rd]; void *d2 = &env->vfp.zregs[(rd + 1) & 31]; @@ -XXX,XX +XXX,XX @@ static void sve_st2_r(CPUARMState *env, void *vg, target_ulong addr, uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); do { if (pg & 1) { - tlb_fn(env, d1, i, addr, mmu_idx, ra); - tlb_fn(env, d2, i, addr + msize, mmu_idx, ra); + tlb_fn(env, d1, i, addr, oi, ra); + tlb_fn(env, d2, i, addr + msize, oi, ra); } i += esize, pg >>= esize; addr += 2 * msize; @@ -XXX,XX +XXX,XX @@ static void sve_st3_r(CPUARMState *env, void *vg, target_ulong addr, const int esize, const int msize, sve_st1_tlb_fn *tlb_fn) { - const int mmu_idx = cpu_mmu_index(env, false); + const TCGMemOpIdx oi = extract32(desc, SIMD_DATA_SHIFT, MEMOPIDX_SHIFT); + const unsigned rd = extract32(desc, SIMD_DATA_SHIFT + MEMOPIDX_SHIFT, 5); intptr_t i, oprsz = simd_oprsz(desc); - unsigned rd = simd_data(desc); void *d1 = &env->vfp.zregs[rd]; void *d2 = &env->vfp.zregs[(rd + 1) & 31]; void *d3 = &env->vfp.zregs[(rd + 2) & 31]; @@ -XXX,XX +XXX,XX @@ static void sve_st3_r(CPUARMState *env, void *vg, target_ulong addr, uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); do { if (pg & 1) { - tlb_fn(env, d1, i, addr, mmu_idx, ra); - tlb_fn(env, d2, i, addr + msize, mmu_idx, ra); - tlb_fn(env, d3, i, addr + 2 * msize, mmu_idx, ra); + tlb_fn(env, d1, i, addr, oi, ra); + tlb_fn(env, d2, i, addr + msize, oi, ra); + tlb_fn(env, d3, i, addr + 2 * msize, oi, ra); } i += esize, pg >>= esize; addr += 3 * msize; @@ -XXX,XX +XXX,XX @@ static void sve_st4_r(CPUARMState *env, void *vg, target_ulong addr, const int esize, const int msize, sve_st1_tlb_fn *tlb_fn) { - const int mmu_idx = cpu_mmu_index(env, false); + const TCGMemOpIdx oi = extract32(desc, SIMD_DATA_SHIFT, MEMOPIDX_SHIFT); + const unsigned rd = extract32(desc, SIMD_DATA_SHIFT + MEMOPIDX_SHIFT, 5); intptr_t i, oprsz = simd_oprsz(desc); - unsigned rd = simd_data(desc); void *d1 = &env->vfp.zregs[rd]; void *d2 = &env->vfp.zregs[(rd + 1) & 31]; void *d3 = &env->vfp.zregs[(rd + 2) & 31]; @@ -XXX,XX +XXX,XX @@ static void sve_st4_r(CPUARMState *env, void *vg, target_ulong addr, uint16_t pg = *(uint16_t *)(vg + H1_2(i >> 3)); do { if (pg & 1) { - tlb_fn(env, d1, i, addr, mmu_idx, ra); - tlb_fn(env, d2, i, addr + msize, mmu_idx, ra); - tlb_fn(env, d3, i, addr + 2 * msize, mmu_idx, ra); - tlb_fn(env, d4, i, addr + 3 * msize, mmu_idx, ra); + tlb_fn(env, d1, i, addr, oi, ra); + tlb_fn(env, d2, i, addr + msize, oi, ra); + tlb_fn(env, d3, i, addr + 2 * msize, oi, ra); + tlb_fn(env, d4, i, addr + 3 * msize, oi, ra); } i += esize, pg >>= esize; addr += 4 * msize; @@ -XXX,XX +XXX,XX @@ static void sve_ld1_zs(CPUARMState *env, void *vd, void *vg, void *vm, target_ulong base, uint32_t desc, uintptr_t ra, zreg_off_fn *off_fn, sve_ld1_tlb_fn *tlb_fn) { - const int mmu_idx = cpu_mmu_index(env, false); + const TCGMemOpIdx oi = extract32(desc, SIMD_DATA_SHIFT, MEMOPIDX_SHIFT); + const int scale = extract32(desc, SIMD_DATA_SHIFT + MEMOPIDX_SHIFT, 2); intptr_t i, oprsz = simd_oprsz(desc); - unsigned scale = simd_data(desc); ARMVectorReg scratch = { }; set_helper_retaddr(ra); @@ -XXX,XX +XXX,XX @@ static void sve_ld1_zs(CPUARMState *env, void *vd, void *vg, void *vm, do { if (likely(pg & 1)) { target_ulong off = off_fn(vm, i); - tlb_fn(env, &scratch, i, base + (off << scale), mmu_idx, ra); + tlb_fn(env, &scratch, i, base + (off << scale), oi, ra); } i += 4, pg >>= 4; } while (i & 15); @@ -XXX,XX +XXX,XX @@ static void sve_ld1_zd(CPUARMState *env, void *vd, void *vg, void *vm, target_ulong base, uint32_t desc, uintptr_t ra, zreg_off_fn *off_fn, sve_ld1_tlb_fn *tlb_fn) { - const int mmu_idx = cpu_mmu_index(env, false); + const TCGMemOpIdx oi = extract32(desc, SIMD_DATA_SHIFT, MEMOPIDX_SHIFT); + const int scale = extract32(desc, SIMD_DATA_SHIFT + MEMOPIDX_SHIFT, 2); intptr_t i, oprsz = simd_oprsz(desc) / 8; - unsigned scale = simd_data(desc); ARMVectorReg scratch = { }; set_helper_retaddr(ra); @@ -XXX,XX +XXX,XX @@ static void sve_ld1_zd(CPUARMState *env, void *vd, void *vg, void *vm, uint8_t pg = *(uint8_t *)(vg + H1(i)); if (likely(pg & 1)) { target_ulong off = off_fn(vm, i * 8); - tlb_fn(env, &scratch, i * 8, base + (off << scale), mmu_idx, ra); + tlb_fn(env, &scratch, i * 8, base + (off << scale), oi, ra); } } set_helper_retaddr(0); @@ -XXX,XX +XXX,XX @@ typedef bool sve_ld1_nf_fn(CPUARMState *env, void *vd, intptr_t reg_off, #ifdef CONFIG_SOFTMMU #define DO_LD_NF(NAME, H, TYPEE, TYPEM, HOST) \ static bool sve_ld##NAME##_nf(CPUARMState *env, void *vd, intptr_t reg_off, \ - target_ulong addr, int mmu_idx) \ + target_ulong addr, int mmu_idx) \ { \ target_ulong next_page = -(addr | TARGET_PAGE_MASK); \ if (likely(next_page - addr >= sizeof(TYPEM))) { \ @@ -XXX,XX +XXX,XX @@ static inline void sve_ldff1_zs(CPUARMState *env, void *vd, void *vg, void *vm, zreg_off_fn *off_fn, sve_ld1_tlb_fn *tlb_fn, sve_ld1_nf_fn *nonfault_fn) { - const int mmu_idx = cpu_mmu_index(env, false); + const TCGMemOpIdx oi = extract32(desc, SIMD_DATA_SHIFT, MEMOPIDX_SHIFT); + const int mmu_idx = get_mmuidx(oi); + const int scale = extract32(desc, SIMD_DATA_SHIFT + MEMOPIDX_SHIFT, 2); intptr_t reg_off, reg_max = simd_oprsz(desc); - unsigned scale = simd_data(desc); target_ulong addr; /* Skip to the first true predicate. */ @@ -XXX,XX +XXX,XX @@ static inline void sve_ldff1_zs(CPUARMState *env, void *vd, void *vg, void *vm, set_helper_retaddr(ra); addr = off_fn(vm, reg_off); addr = base + (addr << scale); - tlb_fn(env, vd, reg_off, addr, mmu_idx, ra); + tlb_fn(env, vd, reg_off, addr, oi, ra); /* The rest of the reads will be non-faulting. */ set_helper_retaddr(0); @@ -XXX,XX +XXX,XX @@ static inline void sve_ldff1_zd(CPUARMState *env, void *vd, void *vg, void *vm, zreg_off_fn *off_fn, sve_ld1_tlb_fn *tlb_fn, sve_ld1_nf_fn *nonfault_fn) { - const int mmu_idx = cpu_mmu_index(env, false); + const TCGMemOpIdx oi = extract32(desc, SIMD_DATA_SHIFT, MEMOPIDX_SHIFT); + const int mmu_idx = get_mmuidx(oi); + const int scale = extract32(desc, SIMD_DATA_SHIFT + MEMOPIDX_SHIFT, 2); intptr_t reg_off, reg_max = simd_oprsz(desc); - unsigned scale = simd_data(desc); target_ulong addr; /* Skip to the first true predicate. */ @@ -XXX,XX +XXX,XX @@ static inline void sve_ldff1_zd(CPUARMState *env, void *vd, void *vg, void *vm, set_helper_retaddr(ra); addr = off_fn(vm, reg_off); addr = base + (addr << scale); - tlb_fn(env, vd, reg_off, addr, mmu_idx, ra); + tlb_fn(env, vd, reg_off, addr, oi, ra); /* The rest of the reads will be non-faulting. */ set_helper_retaddr(0); @@ -XXX,XX +XXX,XX @@ static void sve_st1_zs(CPUARMState *env, void *vd, void *vg, void *vm, target_ulong base, uint32_t desc, uintptr_t ra, zreg_off_fn *off_fn, sve_ld1_tlb_fn *tlb_fn) { - const int mmu_idx = cpu_mmu_index(env, false); + const TCGMemOpIdx oi = extract32(desc, SIMD_DATA_SHIFT, MEMOPIDX_SHIFT); + const int scale = extract32(desc, SIMD_DATA_SHIFT + MEMOPIDX_SHIFT, 2); intptr_t i, oprsz = simd_oprsz(desc); - unsigned scale = simd_data(desc); set_helper_retaddr(ra); for (i = 0; i < oprsz; ) { @@ -XXX,XX +XXX,XX @@ static void sve_st1_zs(CPUARMState *env, void *vd, void *vg, void *vm, do { if (likely(pg & 1)) { target_ulong off = off_fn(vm, i); - tlb_fn(env, vd, i, base + (off << scale), mmu_idx, ra); + tlb_fn(env, vd, i, base + (off << scale), oi, ra); } i += 4, pg >>= 4; } while (i & 15); @@ -XXX,XX +XXX,XX @@ static void sve_st1_zd(CPUARMState *env, void *vd, void *vg, void *vm, target_ulong base, uint32_t desc, uintptr_t ra, zreg_off_fn *off_fn, sve_ld1_tlb_fn *tlb_fn) { - const int mmu_idx = cpu_mmu_index(env, false); + const TCGMemOpIdx oi = extract32(desc, SIMD_DATA_SHIFT, MEMOPIDX_SHIFT); + const int scale = extract32(desc, SIMD_DATA_SHIFT + MEMOPIDX_SHIFT, 2); intptr_t i, oprsz = simd_oprsz(desc) / 8; - unsigned scale = simd_data(desc); set_helper_retaddr(ra); for (i = 0; i < oprsz; i++) { uint8_t pg = *(uint8_t *)(vg + H1(i)); if (likely(pg & 1)) { target_ulong off = off_fn(vm, i * 8); - tlb_fn(env, vd, i * 8, base + (off << scale), mmu_idx, ra); + tlb_fn(env, vd, i * 8, base + (off << scale), oi, ra); } } set_helper_retaddr(0); diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static const uint8_t dtype_esz[16] = { 3, 2, 1, 3 }; +static TCGMemOpIdx sve_memopidx(DisasContext *s, int dtype) +{ + return make_memop_idx(s->be_data | dtype_mop[dtype], get_mem_index(s)); +} + static void do_mem_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, - gen_helper_gvec_mem *fn) + int dtype, gen_helper_gvec_mem *fn) { unsigned vsz = vec_full_reg_size(s); TCGv_ptr t_pg; - TCGv_i32 desc; + TCGv_i32 t_desc; + int desc; /* For e.g. LD4, there are not enough arguments to pass all 4 * registers as pointers, so encode the regno into the data field. * For consistency, do this even for LD1. */ - desc = tcg_const_i32(simd_desc(vsz, vsz, zt)); + desc = sve_memopidx(s, dtype); + desc |= zt << MEMOPIDX_SHIFT; + desc = simd_desc(vsz, vsz, desc); + t_desc = tcg_const_i32(desc); t_pg = tcg_temp_new_ptr(); tcg_gen_addi_ptr(t_pg, cpu_env, pred_full_reg_offset(s, pg)); - fn(cpu_env, t_pg, addr, desc); + fn(cpu_env, t_pg, addr, t_desc); tcg_temp_free_ptr(t_pg); - tcg_temp_free_i32(desc); + tcg_temp_free_i32(t_desc); } static void do_ld_zpa(DisasContext *s, int zt, int pg, @@ -XXX,XX +XXX,XX @@ static void do_ld_zpa(DisasContext *s, int zt, int pg, * accessible via the instruction encoding. */ assert(fn != NULL); - do_mem_zpa(s, zt, pg, addr, fn); + do_mem_zpa(s, zt, pg, addr, dtype, fn); } static bool trans_LD_zprr(DisasContext *s, arg_rprr_load *a, uint32_t insn) @@ -XXX,XX +XXX,XX @@ static bool trans_LDFF1_zprr(DisasContext *s, arg_rprr_load *a, uint32_t insn) TCGv_i64 addr = new_tmp_a64(s); tcg_gen_shli_i64(addr, cpu_reg(s, a->rm), dtype_msz(a->dtype)); tcg_gen_add_i64(addr, addr, cpu_reg_sp(s, a->rn)); - do_mem_zpa(s, a->rd, a->pg, addr, fns[s->be_data == MO_BE][a->dtype]); + do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, + fns[s->be_data == MO_BE][a->dtype]); } return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_LDNF1_zpri(DisasContext *s, arg_rpri_load *a, uint32_t insn) TCGv_i64 addr = new_tmp_a64(s); tcg_gen_addi_i64(addr, cpu_reg_sp(s, a->rn), off); - do_mem_zpa(s, a->rd, a->pg, addr, fns[s->be_data == MO_BE][a->dtype]); + do_mem_zpa(s, a->rd, a->pg, addr, a->dtype, + fns[s->be_data == MO_BE][a->dtype]); } return true; } @@ -XXX,XX +XXX,XX @@ static void do_ldrq(DisasContext *s, int zt, int pg, TCGv_i64 addr, int msz) }; unsigned vsz = vec_full_reg_size(s); TCGv_ptr t_pg; - TCGv_i32 desc; - int poff; + TCGv_i32 t_desc; + int desc, poff; /* Load the first quadword using the normal predicated load helpers. */ - desc = tcg_const_i32(simd_desc(16, 16, zt)); + desc = sve_memopidx(s, msz_dtype(msz)); + desc |= zt << MEMOPIDX_SHIFT; + desc = simd_desc(16, 16, desc); + t_desc = tcg_const_i32(desc); poff = pred_full_reg_offset(s, pg); if (vsz > 16) { @@ -XXX,XX +XXX,XX @@ static void do_ldrq(DisasContext *s, int zt, int pg, TCGv_i64 addr, int msz) t_pg = tcg_temp_new_ptr(); tcg_gen_addi_ptr(t_pg, cpu_env, poff); - fns[s->be_data == MO_BE][msz](cpu_env, t_pg, addr, desc); + fns[s->be_data == MO_BE][msz](cpu_env, t_pg, addr, t_desc); tcg_temp_free_ptr(t_pg); - tcg_temp_free_i32(desc); + tcg_temp_free_i32(t_desc); /* Replicate that first quadword. */ if (vsz > 16) { @@ -XXX,XX +XXX,XX @@ static void do_st_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, fn = fn_multiple[be][nreg - 1][msz]; } assert(fn != NULL); - do_mem_zpa(s, zt, pg, addr, fn); + do_mem_zpa(s, zt, pg, addr, msz_dtype(msz), fn); } static bool trans_ST_zprr(DisasContext *s, arg_rprr_store *a, uint32_t insn) @@ -XXX,XX +XXX,XX @@ static bool trans_ST_zpri(DisasContext *s, arg_rpri_store *a, uint32_t insn) *** SVE gather loads / scatter stores */ -static void do_mem_zpz(DisasContext *s, int zt, int pg, int zm, int scale, - TCGv_i64 scalar, gen_helper_gvec_mem_scatter *fn) +static void do_mem_zpz(DisasContext *s, int zt, int pg, int zm, + int scale, TCGv_i64 scalar, int msz, + gen_helper_gvec_mem_scatter *fn) { unsigned vsz = vec_full_reg_size(s); - TCGv_i32 desc = tcg_const_i32(simd_desc(vsz, vsz, scale)); TCGv_ptr t_zm = tcg_temp_new_ptr(); TCGv_ptr t_pg = tcg_temp_new_ptr(); TCGv_ptr t_zt = tcg_temp_new_ptr(); + TCGv_i32 t_desc; + int desc; + + desc = sve_memopidx(s, msz_dtype(msz)); + desc |= scale << MEMOPIDX_SHIFT; + desc = simd_desc(vsz, vsz, desc); + t_desc = tcg_const_i32(desc); tcg_gen_addi_ptr(t_pg, cpu_env, pred_full_reg_offset(s, pg)); tcg_gen_addi_ptr(t_zm, cpu_env, vec_full_reg_offset(s, zm)); tcg_gen_addi_ptr(t_zt, cpu_env, vec_full_reg_offset(s, zt)); - fn(cpu_env, t_zt, t_pg, t_zm, scalar, desc); + fn(cpu_env, t_zt, t_pg, t_zm, scalar, t_desc); tcg_temp_free_ptr(t_zt); tcg_temp_free_ptr(t_zm); tcg_temp_free_ptr(t_pg); - tcg_temp_free_i32(desc); + tcg_temp_free_i32(t_desc); } /* Indexed by [be][ff][xs][u][msz]. */ @@ -XXX,XX +XXX,XX @@ static bool trans_LD1_zprz(DisasContext *s, arg_LD1_zprz *a, uint32_t insn) assert(fn != NULL); do_mem_zpz(s, a->rd, a->pg, a->rm, a->scale * a->msz, - cpu_reg_sp(s, a->rn), fn); + cpu_reg_sp(s, a->rn), a->msz, fn); return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_LD1_zpiz(DisasContext *s, arg_LD1_zpiz *a, uint32_t insn) * by loading the immediate into the scalar parameter. */ imm = tcg_const_i64(a->imm << a->msz); - do_mem_zpz(s, a->rd, a->pg, a->rn, 0, imm, fn); + do_mem_zpz(s, a->rd, a->pg, a->rn, 0, imm, a->msz, fn); tcg_temp_free_i64(imm); return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_ST1_zprz(DisasContext *s, arg_ST1_zprz *a, uint32_t insn) g_assert_not_reached(); } do_mem_zpz(s, a->rd, a->pg, a->rm, a->scale * a->msz, - cpu_reg_sp(s, a->rn), fn); + cpu_reg_sp(s, a->rn), a->msz, fn); return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_ST1_zpiz(DisasContext *s, arg_ST1_zpiz *a, uint32_t insn) * by loading the immediate into the scalar parameter. */ imm = tcg_const_i64(a->imm << a->msz); - do_mem_zpz(s, a->rd, a->pg, a->rn, 0, imm, fn); + do_mem_zpz(s, a->rd, a->pg, a->rn, 0, imm, a->msz, fn); tcg_temp_free_i64(imm); return true; } -- 2.19.0 The Arm v8M architecture includes hardware stack limit checking. When certain instructions update the stack pointer, if the new value of SP is below the limit set in the associated limit register then an exception is taken. Add a TB flag that tracks whether the limit-checking code needs to be emitted. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Message-id: 20181002163556.10279-2-peter.maydell@linaro.org --- target/arm/cpu.h | 7 +++++++ target/arm/translate.h | 1 + target/arm/helper.c | 10 ++++++++++ target/arm/translate.c | 1 + 4 files changed, 19 insertions(+) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ FIELD(V7M_CCR, UNALIGN_TRP, 3, 1) FIELD(V7M_CCR, DIV_0_TRP, 4, 1) FIELD(V7M_CCR, BFHFNMIGN, 8, 1) FIELD(V7M_CCR, STKALIGN, 9, 1) +FIELD(V7M_CCR, STKOFHFNMIGN, 10, 1) FIELD(V7M_CCR, DC, 16, 1) FIELD(V7M_CCR, IC, 17, 1) +FIELD(V7M_CCR, BP, 18, 1) /* V7M SCR bits */ FIELD(V7M_SCR, SLEEPONEXIT, 1, 1) @@ -XXX,XX +XXX,XX @@ static inline bool arm_cpu_data_is_big_endian(CPUARMState *env) /* For M profile only, Handler (ie not Thread) mode */ #define ARM_TBFLAG_HANDLER_SHIFT 21 #define ARM_TBFLAG_HANDLER_MASK (1 << ARM_TBFLAG_HANDLER_SHIFT) +/* For M profile only, whether we should generate stack-limit checks */ +#define ARM_TBFLAG_STACKCHECK_SHIFT 22 +#define ARM_TBFLAG_STACKCHECK_MASK (1 << ARM_TBFLAG_STACKCHECK_SHIFT) /* Bit usage when in AArch64 state */ #define ARM_TBFLAG_TBI0_SHIFT 0 /* TBI0 for EL0/1 or TBI for EL2/3 */ @@ -XXX,XX +XXX,XX @@ static inline bool arm_cpu_data_is_big_endian(CPUARMState *env) (((F) & ARM_TBFLAG_BE_DATA_MASK) >> ARM_TBFLAG_BE_DATA_SHIFT) #define ARM_TBFLAG_HANDLER(F) \ (((F) & ARM_TBFLAG_HANDLER_MASK) >> ARM_TBFLAG_HANDLER_SHIFT) +#define ARM_TBFLAG_STACKCHECK(F) \ + (((F) & ARM_TBFLAG_STACKCHECK_MASK) >> ARM_TBFLAG_STACKCHECK_SHIFT) #define ARM_TBFLAG_TBI0(F) \ (((F) & ARM_TBFLAG_TBI0_MASK) >> ARM_TBFLAG_TBI0_SHIFT) #define ARM_TBFLAG_TBI1(F) \ diff --git a/target/arm/translate.h b/target/arm/translate.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.h +++ b/target/arm/translate.h @@ -XXX,XX +XXX,XX @@ typedef struct DisasContext { int vec_stride; bool v7m_handler_mode; bool v8m_secure; /* true if v8M and we're in Secure mode */ + bool v8m_stackcheck; /* true if we need to perform v8M stack limit checks */ /* Immediate value in AArch32 SVC insn; must be set if is_jmp == DISAS_SWI * so that top level loop can generate correct syndrome information. */ diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ void cpu_get_tb_cpu_state(CPUARMState *env, target_ulong *pc, flags |= ARM_TBFLAG_HANDLER_MASK; } + /* v8M always applies stack limit checks unless CCR.STKOFHFNMIGN is + * suppressing them because the requested execution priority is less than 0. + */ + if (arm_feature(env, ARM_FEATURE_V8) && + arm_feature(env, ARM_FEATURE_M) && + !((mmu_idx & ARM_MMU_IDX_M_NEGPRI) && + (env->v7m.ccr[env->v7m.secure] & R_V7M_CCR_STKOFHFNMIGN_MASK))) { + flags |= ARM_TBFLAG_STACKCHECK_MASK; + } + *pflags = flags; *cs_base = 0; } diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static void arm_tr_init_disas_context(DisasContextBase *dcbase, CPUState *cs) dc->v7m_handler_mode = ARM_TBFLAG_HANDLER(dc->base.tb->flags); dc->v8m_secure = arm_feature(env, ARM_FEATURE_M_SECURITY) && regime_is_secure(env, dc->mmu_idx); + dc->v8m_stackcheck = ARM_TBFLAG_STACKCHECK(dc->base.tb->flags); dc->cp_regs = cpu->cp_regs; dc->features = env->features; -- 2.19.0 Define EXCP_STKOF, and arrange for it to cause us to take a UsageFault with CFSR.STKOF set. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181002163556.10279-3-peter.maydell@linaro.org --- target/arm/cpu.h | 2 ++ target/arm/helper.c | 5 +++++ 2 files changed, 7 insertions(+) diff --git a/target/arm/cpu.h b/target/arm/cpu.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/cpu.h +++ b/target/arm/cpu.h @@ -XXX,XX +XXX,XX @@ #define EXCP_SEMIHOST 16 /* semihosting call */ #define EXCP_NOCP 17 /* v7M NOCP UsageFault */ #define EXCP_INVSTATE 18 /* v7M INVSTATE UsageFault */ +#define EXCP_STKOF 19 /* v8M STKOF UsageFault */ /* NB: add new EXCP_ defines to the array in arm_log_exception() too */ #define ARMV7M_EXCP_RESET 1 @@ -XXX,XX +XXX,XX @@ FIELD(V7M_CFSR, UNDEFINSTR, 16 + 0, 1) FIELD(V7M_CFSR, INVSTATE, 16 + 1, 1) FIELD(V7M_CFSR, INVPC, 16 + 2, 1) FIELD(V7M_CFSR, NOCP, 16 + 3, 1) +FIELD(V7M_CFSR, STKOF, 16 + 4, 1) FIELD(V7M_CFSR, UNALIGNED, 16 + 8, 1) FIELD(V7M_CFSR, DIVBYZERO, 16 + 9, 1) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static void arm_log_exception(int idx) [EXCP_SEMIHOST] = "Semihosting call", [EXCP_NOCP] = "v7M NOCP UsageFault", [EXCP_INVSTATE] = "v7M INVSTATE UsageFault", + [EXCP_STKOF] = "v8M STKOF UsageFault", }; if (idx >= 0 && idx < ARRAY_SIZE(excnames)) { @@ -XXX,XX +XXX,XX @@ void arm_v7m_cpu_do_interrupt(CPUState *cs) armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_USAGE, env->v7m.secure); env->v7m.cfsr[env->v7m.secure] |= R_V7M_CFSR_INVSTATE_MASK; break; + case EXCP_STKOF: + armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_USAGE, env->v7m.secure); + env->v7m.cfsr[env->v7m.secure] |= R_V7M_CFSR_STKOF_MASK; + break; case EXCP_SWI: /* The PC already points to the next instruction. */ armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_SVC, env->v7m.secure); -- 2.19.0 We're going to want v7m_using_psp() in op_helper.c in the next patch, so move it from helper.c to internals.h. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181002163556.10279-4-peter.maydell@linaro.org --- target/arm/internals.h | 16 ++++++++++++++++ target/arm/helper.c | 12 ------------ 2 files changed, 16 insertions(+), 12 deletions(-) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ static inline uint32_t arm_debug_exception_fsr(CPUARMState *env) */ #define MEMOPIDX_SHIFT 8 +/** + * v7m_using_psp: Return true if using process stack pointer + * Return true if the CPU is currently using the process stack + * pointer, or false if it is using the main stack pointer. + */ +static inline bool v7m_using_psp(CPUARMState *env) +{ + /* Handler mode always uses the main stack; for thread mode + * the CONTROL.SPSEL bit determines the answer. + * Note that in v7M it is not possible to be in Handler mode with + * CONTROL.SPSEL non-zero, but in v8M it is, so we must check both. + */ + return !arm_v7m_is_handler_mode(env) && + env->v7m.control[env->v7m.secure] & R_V7M_CONTROL_SPSEL_MASK; +} + #endif diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ pend_fault: return false; } -/* Return true if we're using the process stack pointer (not the MSP) */ -static bool v7m_using_psp(CPUARMState *env) -{ - /* Handler mode always uses the main stack; for thread mode - * the CONTROL.SPSEL bit determines the answer. - * Note that in v7M it is not possible to be in Handler mode with - * CONTROL.SPSEL non-zero, but in v8M it is, so we must check both. - */ - return !arm_v7m_is_handler_mode(env) && - env->v7m.control[env->v7m.secure] & R_V7M_CONTROL_SPSEL_MASK; -} - /* Write to v7M CONTROL.SPSEL bit for the specified security bank. * This may change the current stack pointer between Main and Process * stack pointers if it is done for the CONTROL register for the current -- 2.19.0 Add code to insert calls to a helper function to do the stack limit checking when we handle these forms of instruction that write to SP: * ADD (SP plus immediate) * ADD (SP plus register) * SUB (SP minus immediate) * SUB (SP minus register) * MOV (register) Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181002163556.10279-5-peter.maydell@linaro.org --- target/arm/helper.h | 2 ++ target/arm/internals.h | 14 ++++++++ target/arm/op_helper.c | 19 ++++++++++ target/arm/translate.c | 80 +++++++++++++++++++++++++++++++++++++----- 4 files changed, 106 insertions(+), 9 deletions(-) diff --git a/target/arm/helper.h b/target/arm/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.h +++ b/target/arm/helper.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_2(v7m_blxns, void, env, i32) DEF_HELPER_3(v7m_tt, i32, env, i32, i32) +DEF_HELPER_2(v8m_stackcheck, void, env, i32) + DEF_HELPER_4(access_check_cp_reg, void, env, ptr, i32, i32) DEF_HELPER_3(set_cp_reg, void, env, ptr, i32) DEF_HELPER_2(get_cp_reg, i32, env, ptr) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ static inline bool v7m_using_psp(CPUARMState *env) env->v7m.control[env->v7m.secure] & R_V7M_CONTROL_SPSEL_MASK; } +/** + * v7m_sp_limit: Return SP limit for current CPU state + * Return the SP limit value for the current CPU security state + * and stack pointer. + */ +static inline uint32_t v7m_sp_limit(CPUARMState *env) +{ + if (v7m_using_psp(env)) { + return env->v7m.psplim[env->v7m.secure]; + } else { + return env->v7m.msplim[env->v7m.secure]; + } +} + #endif diff --git a/target/arm/op_helper.c b/target/arm/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/op_helper.c +++ b/target/arm/op_helper.c @@ -XXX,XX +XXX,XX @@ void arm_cpu_do_transaction_failed(CPUState *cs, hwaddr physaddr, #endif /* !defined(CONFIG_USER_ONLY) */ +void HELPER(v8m_stackcheck)(CPUARMState *env, uint32_t newvalue) +{ + /* + * Perform the v8M stack limit check for SP updates from translated code, + * raising an exception if the limit is breached. + */ + if (newvalue < v7m_sp_limit(env)) { + CPUState *cs = CPU(arm_env_get_cpu(env)); + + /* + * Stack limit exceptions are a rare case, so rather than syncing + * PC/condbits before the call, we use cpu_restore_state() to + * get them right before raising the exception. + */ + cpu_restore_state(cs, GETPC(), true); + raise_exception(env, EXCP_STKOF, 0, 1); + } +} + uint32_t HELPER(add_setq)(CPUARMState *env, uint32_t a, uint32_t b) { uint32_t res = a + b; diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static void store_reg(DisasContext *s, int reg, TCGv_i32 var) tcg_temp_free_i32(var); } +/* + * Variant of store_reg which applies v8M stack-limit checks before updating + * SP. If the check fails this will result in an exception being taken. + * We disable the stack checks for CONFIG_USER_ONLY because we have + * no idea what the stack limits should be in that case. + * If stack checking is not being done this just acts like store_reg(). + */ +static void store_sp_checked(DisasContext *s, TCGv_i32 var) +{ +#ifndef CONFIG_USER_ONLY + if (s->v8m_stackcheck) { + gen_helper_v8m_stackcheck(cpu_env, var); + } +#endif + store_reg(s, 13, var); +} + /* Value extensions. */ #define gen_uxtb(var) tcg_gen_ext8u_i32(var, var) #define gen_uxth(var) tcg_gen_ext16u_i32(var, var) @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) if (gen_thumb2_data_op(s, op, conds, 0, tmp, tmp2)) goto illegal_op; tcg_temp_free_i32(tmp2); - if (rd != 15) { + if (rd == 13 && + ((op == 2 && rn == 15) || + (op == 8 && rn == 13) || + (op == 13 && rn == 13))) { + /* MOV SP, ... or ADD SP, SP, ... or SUB SP, SP, ... */ + store_sp_checked(s, tmp); + } else if (rd != 15) { store_reg(s, rd, tmp); } else { tcg_temp_free_i32(tmp); @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) gen_jmp(s, s->pc + offset); } } else { - /* Data processing immediate. */ + /* + * 0b1111_0xxx_xxxx_0xxx_xxxx_xxxx + * - Data-processing (modified immediate, plain binary immediate) + */ if (insn & (1 << 25)) { + /* + * 0b1111_0x1x_xxxx_0xxx_xxxx_xxxx + * - Data-processing (plain binary immediate) + */ if (insn & (1 << 24)) { if (insn & (1 << 20)) goto illegal_op; @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) tmp = tcg_temp_new_i32(); tcg_gen_movi_i32(tmp, imm); } + store_reg(s, rd, tmp); } else { /* Add/sub 12-bit immediate. */ if (rn == 15) { @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) offset += imm; tmp = tcg_temp_new_i32(); tcg_gen_movi_i32(tmp, offset); + store_reg(s, rd, tmp); } else { tmp = load_reg(s, rn); if (insn & (1 << 23)) tcg_gen_subi_i32(tmp, tmp, imm); else tcg_gen_addi_i32(tmp, tmp, imm); + if (rn == 13 && rd == 13) { + /* ADD SP, SP, imm or SUB SP, SP, imm */ + store_sp_checked(s, tmp); + } else { + store_reg(s, rd, tmp); + } } } - store_reg(s, rd, tmp); } } else { + /* + * 0b1111_0x0x_xxxx_0xxx_xxxx_xxxx + * - Data-processing (modified immediate) + */ int shifter_out = 0; /* modified 12-bit immediate. */ shift = ((insn & 0x04000000) >> 23) | ((insn & 0x7000) >> 12); @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) goto illegal_op; tcg_temp_free_i32(tmp2); rd = (insn >> 8) & 0xf; - if (rd != 15) { + if (rd == 13 && rn == 13 + && (op == 8 || op == 13)) { + /* ADD(S) SP, SP, imm or SUB(S) SP, SP, imm */ + store_sp_checked(s, tmp); + } else if (rd != 15) { store_reg(s, rd, tmp); } else { tcg_temp_free_i32(tmp); @@ -XXX,XX +XXX,XX @@ static void disas_thumb_insn(DisasContext *s, uint32_t insn) tmp2 = load_reg(s, rm); tcg_gen_add_i32(tmp, tmp, tmp2); tcg_temp_free_i32(tmp2); - store_reg(s, rd, tmp); + if (rd == 13) { + /* ADD SP, SP, reg */ + store_sp_checked(s, tmp); + } else { + store_reg(s, rd, tmp); + } break; case 1: /* cmp */ tmp = load_reg(s, rd); @@ -XXX,XX +XXX,XX @@ static void disas_thumb_insn(DisasContext *s, uint32_t insn) break; case 2: /* mov/cpy */ tmp = load_reg(s, rm); - store_reg(s, rd, tmp); + if (rd == 13) { + /* MOV SP, reg */ + store_sp_checked(s, tmp); + } else { + store_reg(s, rd, tmp); + } break; case 3: { @@ -XXX,XX +XXX,XX @@ static void disas_thumb_insn(DisasContext *s, uint32_t insn) break; case 10: - /* add to high reg */ + /* + * 0b1010_xxxx_xxxx_xxxx + * - Add PC/SP (immediate) + */ rd = (insn >> 8) & 7; if (insn & (1 << 11)) { /* SP */ @@ -XXX,XX +XXX,XX @@ static void disas_thumb_insn(DisasContext *s, uint32_t insn) op = (insn >> 8) & 0xf; switch (op) { case 0: - /* adjust stack pointer */ + /* + * 0b1011_0000_xxxx_xxxx + * - ADD (SP plus immediate) + * - SUB (SP minus immediate) + */ tmp = load_reg(s, 13); val = (insn & 0x7f) * 4; if (insn & (1 << 7)) val = -(int32_t)val; tcg_gen_addi_i32(tmp, tmp, val); - store_reg(s, 13, tmp); + store_sp_checked(s, tmp); break; case 2: /* sign/zero extend. */ -- 2.19.0 Add some comments to the Thumb decoder indicating what bits of the instruction have been decoded at various points in the code. This is not an exhaustive set of comments; we're gradually adding comments as we work with particular bits of the code. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181002163556.10279-6-peter.maydell@linaro.org --- target/arm/translate.c | 20 +++++++++++++++++--- 1 file changed, 17 insertions(+), 3 deletions(-) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) tmp2 = load_reg(s, rm); if ((insn & 0x70) != 0) goto illegal_op; + /* + * 0b1111_1010_0xxx_xxxx_1111_xxxx_0000_xxxx: + * - MOV, MOVS (register-shifted register), flagsetting + */ op = (insn >> 21) & 3; logic_cc = (insn & (1 << 20)) != 0; gen_arm_shift_reg(tmp, op, tmp2, logic_cc); @@ -XXX,XX +XXX,XX @@ static void disas_thumb_insn(DisasContext *s, uint32_t insn) rd = insn & 7; op = (insn >> 11) & 3; if (op == 3) { - /* add/subtract */ + /* + * 0b0001_1xxx_xxxx_xxxx + * - Add, subtract (three low registers) + * - Add, subtract (two low registers and immediate) + */ rn = (insn >> 3) & 7; tmp = load_reg(s, rn); if (insn & (1 << 10)) { @@ -XXX,XX +XXX,XX @@ static void disas_thumb_insn(DisasContext *s, uint32_t insn) } break; case 2: case 3: - /* arithmetic large immediate */ + /* + * 0b001x_xxxx_xxxx_xxxx + * - Add, subtract, compare, move (one low register and immediate) + */ op = (insn >> 11) & 3; rd = (insn >> 8) & 0x7; if (op == 0) { /* mov */ @@ -XXX,XX +XXX,XX @@ static void disas_thumb_insn(DisasContext *s, uint32_t insn) break; } - /* data processing register */ + /* + * 0b0100_00xx_xxxx_xxxx + * - Data-processing (two low registers) + */ rd = insn & 7; rm = (insn >> 3) & 7; op = (insn >> 6) & 0xf; -- 2.19.0 Add checks for breaches of the v8M stack limit when the stack pointer is decremented to push the exception frame for exception entry. Note that the exception-entry case is unique in that the stack pointer is updated to be the limit value if the limit is hit (per rule R_ZLZG). Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181002163556.10279-7-peter.maydell@linaro.org --- target/arm/helper.c | 54 ++++++++++++++++++++++++++++++++++++++------- 1 file changed, 46 insertions(+), 8 deletions(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static bool v7m_push_callee_stack(ARMCPU *cpu, uint32_t lr, bool dotailchain, uint32_t frameptr; ARMMMUIdx mmu_idx; bool stacked_ok; + uint32_t limit; + bool want_psp; if (dotailchain) { bool mode = lr & R_V7M_EXCRET_MODE_MASK; @@ -XXX,XX +XXX,XX @@ static bool v7m_push_callee_stack(ARMCPU *cpu, uint32_t lr, bool dotailchain, mmu_idx = arm_v7m_mmu_idx_for_secstate_and_priv(env, M_REG_S, priv); frame_sp_p = get_v7m_sp_ptr(env, M_REG_S, mode, lr & R_V7M_EXCRET_SPSEL_MASK); + want_psp = mode && (lr & R_V7M_EXCRET_SPSEL_MASK); + if (want_psp) { + limit = env->v7m.psplim[M_REG_S]; + } else { + limit = env->v7m.msplim[M_REG_S]; + } } else { mmu_idx = core_to_arm_mmu_idx(env, cpu_mmu_index(env, false)); frame_sp_p = &env->regs[13]; + limit = v7m_sp_limit(env); } frameptr = *frame_sp_p - 0x28; + if (frameptr < limit) { + /* + * Stack limit failure: set SP to the limit value, and generate + * STKOF UsageFault. Stack pushes below the limit must not be + * performed. It is IMPDEF whether pushes above the limit are + * performed; we choose not to. + */ + qemu_log_mask(CPU_LOG_INT, + "...STKOF during callee-saves register stacking\n"); + env->v7m.cfsr[env->v7m.secure] |= R_V7M_CFSR_STKOF_MASK; + armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_USAGE, + env->v7m.secure); + *frame_sp_p = limit; + return true; + } /* Write as much of the stack frame as we can. A write failure may * cause us to pend a derived exception. @@ -XXX,XX +XXX,XX @@ static bool v7m_push_callee_stack(ARMCPU *cpu, uint32_t lr, bool dotailchain, v7m_stack_write(cpu, frameptr + 0x24, env->regs[11], mmu_idx, ignore_faults); - /* Update SP regardless of whether any of the stack accesses failed. - * When we implement v8M stack limit checking then this attempt to - * update SP might also fail and result in a derived exception. - */ + /* Update SP regardless of whether any of the stack accesses failed. */ *frame_sp_p = frameptr; return !stacked_ok; @@ -XXX,XX +XXX,XX @@ static bool v7m_push_stack(ARMCPU *cpu) frameptr -= 0x20; + if (arm_feature(env, ARM_FEATURE_V8)) { + uint32_t limit = v7m_sp_limit(env); + + if (frameptr < limit) { + /* + * Stack limit failure: set SP to the limit value, and generate + * STKOF UsageFault. Stack pushes below the limit must not be + * performed. It is IMPDEF whether pushes above the limit are + * performed; we choose not to. + */ + qemu_log_mask(CPU_LOG_INT, + "...STKOF during stacking\n"); + env->v7m.cfsr[env->v7m.secure] |= R_V7M_CFSR_STKOF_MASK; + armv7m_nvic_set_pending(env->nvic, ARMV7M_EXCP_USAGE, + env->v7m.secure); + env->regs[13] = limit; + return true; + } + } + /* Write as much of the stack frame as we can. If we fail a stack * write this will result in a derived exception being pended * (which may be taken in preference to the one we started with @@ -XXX,XX +XXX,XX @@ static bool v7m_push_stack(ARMCPU *cpu) v7m_stack_write(cpu, frameptr + 24, env->regs[15], mmu_idx, false) && v7m_stack_write(cpu, frameptr + 28, xpsr, mmu_idx, false); - /* Update SP regardless of whether any of the stack accesses failed. - * When we implement v8M stack limit checking then this attempt to - * update SP might also fail and result in a derived exception. - */ + /* Update SP regardless of whether any of the stack accesses failed. */ env->regs[13] = frameptr; return !stacked_ok; -- 2.19.0 Check the v8M stack limits when pushing the frame for a non-secure function call via BLXNS. In order to be able to generate the exception we need to promote raise_exception() from being local to op_helper.c so we can call it from helper.c. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181002163556.10279-8-peter.maydell@linaro.org --- target/arm/internals.h | 9 +++++++++ target/arm/helper.c | 4 ++++ target/arm/op_helper.c | 4 ++-- 3 files changed, 15 insertions(+), 2 deletions(-) diff --git a/target/arm/internals.h b/target/arm/internals.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/internals.h +++ b/target/arm/internals.h @@ -XXX,XX +XXX,XX @@ FIELD(V7M_EXCRET, RES1, 7, 25) /* including the must-be-1 prefix */ #define M_FAKE_FSR_NSC_EXEC 0xf /* NS executing in S&NSC memory */ #define M_FAKE_FSR_SFAULT 0xe /* SecureFault INVTRAN, INVEP or AUVIOL */ +/** + * raise_exception: Raise the specified exception. + * Raise a guest exception with the specified value, syndrome register + * and target exception level. This should be called from helper functions, + * and never returns because we will longjump back up to the CPU main loop. + */ +void QEMU_NORETURN raise_exception(CPUARMState *env, uint32_t excp, + uint32_t syndrome, uint32_t target_el); + /* * For AArch64, map a given EL to an index in the banked_spsr array. * Note that this mapping and the AArch32 mapping defined in bank_number() diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(v7m_blxns)(CPUARMState *env, uint32_t dest) "BLXNS with misaligned SP is UNPREDICTABLE\n"); } + if (sp < v7m_sp_limit(env)) { + raise_exception(env, EXCP_STKOF, 0, 1); + } + saved_psr = env->v7m.exception; if (env->v7m.control[M_REG_S] & R_V7M_CONTROL_SFPA_MASK) { saved_psr |= XPSR_SFPA; diff --git a/target/arm/op_helper.c b/target/arm/op_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/op_helper.c +++ b/target/arm/op_helper.c @@ -XXX,XX +XXX,XX @@ #define SIGNBIT (uint32_t)0x80000000 #define SIGNBIT64 ((uint64_t)1 << 63) -static void raise_exception(CPUARMState *env, uint32_t excp, - uint32_t syndrome, uint32_t target_el) +void raise_exception(CPUARMState *env, uint32_t excp, + uint32_t syndrome, uint32_t target_el) { CPUState *cs = CPU(arm_env_get_cpu(env)); -- 2.19.0 Add the v8M stack checks for: * LDRD (immediate) * STRD (immediate) Loads and stores are more complicated than ADD/SUB/MOV, because we must ensure that memory accesses below the stack limit are not performed, so we can't simply do the check when we actually update SP. For these instructions, if the stack limit check triggers we must not: * perform any memory access below the SP limit * update PC, SP or the load/store base register but it is IMPDEF whether we: * perform any accesses above or equal to the SP limit * update destination registers for loads For QEMU we choose to always check the limit before doing any other part of the load or store, so we won't update any registers or perform any memory accesses. It is UNKNOWN whether the limit check triggers for a load or store where the initial SP value is below the limit and one of the stores would be below the limit, but the writeback moves SP to above the limit. For QEMU we choose to trigger the check in this situation. Note that limit checks happen only for loads and stores which update SP via writeback; they do not happen for loads and stores which simply use SP as a base register. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181002163556.10279-9-peter.maydell@linaro.org --- target/arm/translate.c | 27 +++++++++++++++++++++++++-- 1 file changed, 25 insertions(+), 2 deletions(-) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) * 0b1111_1001_x11x_xxxx_xxxx_xxxx_xxxx_xxxx * - load/store dual (pre-indexed) */ + bool wback = extract32(insn, 21, 1); + if (rn == 15) { if (insn & (1 << 21)) { /* UNPREDICTABLE */ @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) addr = load_reg(s, rn); } offset = (insn & 0xff) * 4; - if ((insn & (1 << 23)) == 0) + if ((insn & (1 << 23)) == 0) { offset = -offset; + } + + if (s->v8m_stackcheck && rn == 13 && wback) { + /* + * Here 'addr' is the current SP; if offset is +ve we're + * moving SP up, else down. It is UNKNOWN whether the limit + * check triggers when SP starts below the limit and ends + * up above it; check whichever of the current and final + * SP is lower, so QEMU will trigger in that situation. + */ + if ((int32_t)offset < 0) { + TCGv_i32 newsp = tcg_temp_new_i32(); + + tcg_gen_addi_i32(newsp, addr, offset); + gen_helper_v8m_stackcheck(cpu_env, newsp); + tcg_temp_free_i32(newsp); + } else { + gen_helper_v8m_stackcheck(cpu_env, addr); + } + } + if (insn & (1 << 24)) { tcg_gen_addi_i32(addr, addr, offset); offset = 0; @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) gen_aa32_st32(s, tmp, addr, get_mem_index(s)); tcg_temp_free_i32(tmp); } - if (insn & (1 << 21)) { + if (wback) { /* Base writeback. */ tcg_gen_addi_i32(addr, addr, offset - 4); store_reg(s, rn, addr); -- 2.19.0 Add the v8M stack checks for: * LDM (T2 encoding) * STM (T2 encoding) This includes the 32-bit encodings of the instructions listed in v8M ARM ARM rule R_YVWT as * LDM, LDMIA, LDMFD * LDMDB, LDMEA * POP (multiple registers) * PUSH (muliple registers) * STM, STMIA, STMEA * STMDB, STMFD We perform the stack limit before doing any other part of the load or store. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181002163556.10279-10-peter.maydell@linaro.org --- target/arm/translate.c | 19 ++++++++++++++++++- 1 file changed, 18 insertions(+), 1 deletion(-) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) } else { int i, loaded_base = 0; TCGv_i32 loaded_var; + bool wback = extract32(insn, 21, 1); /* Load/store multiple. */ addr = load_reg(s, rn); offset = 0; @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) if (insn & (1 << i)) offset += 4; } + if (insn & (1 << 24)) { tcg_gen_addi_i32(addr, addr, -offset); } + if (s->v8m_stackcheck && rn == 13 && wback) { + /* + * If the writeback is incrementing SP rather than + * decrementing it, and the initial SP is below the + * stack limit but the final written-back SP would + * be above, then then we must not perform any memory + * accesses, but it is IMPDEF whether we generate + * an exception. We choose to do so in this case. + * At this point 'addr' is the lowest address, so + * either the original SP (if incrementing) or our + * final SP (if decrementing), so that's what we check. + */ + gen_helper_v8m_stackcheck(cpu_env, addr); + } + loaded_var = NULL; for (i = 0; i < 16; i++) { if ((insn & (1 << i)) == 0) @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) if (loaded_base) { store_reg(s, rn, loaded_var); } - if (insn & (1 << 21)) { + if (wback) { /* Base register writeback. */ if (insn & (1 << 24)) { tcg_gen_addi_i32(addr, addr, -offset); -- 2.19.0 Add v8M stack checks for the instructions in the T32 "load/store single" encoding class: these are the "immediate pre-indexed" and "immediate, post-indexed" LDR and STR instructions. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181002163556.10279-11-peter.maydell@linaro.org --- target/arm/translate.c | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) imm = -imm; /* Fall through. */ case 0xf: /* Pre-increment. */ - tcg_gen_addi_i32(addr, addr, imm); writeback = 1; break; default: @@ -XXX,XX +XXX,XX @@ static void disas_thumb2_insn(DisasContext *s, uint32_t insn) issinfo = writeback ? ISSInvalid : rs; + if (s->v8m_stackcheck && rn == 13 && writeback) { + /* + * Stackcheck. Here we know 'addr' is the current SP; + * if imm is +ve we're moving SP up, else down. It is + * UNKNOWN whether the limit check triggers when SP starts + * below the limit and ends up above it; we chose to do so. + */ + if ((int32_t)imm < 0) { + TCGv_i32 newsp = tcg_temp_new_i32(); + + tcg_gen_addi_i32(newsp, addr, imm); + gen_helper_v8m_stackcheck(cpu_env, newsp); + tcg_temp_free_i32(newsp); + } else { + gen_helper_v8m_stackcheck(cpu_env, addr); + } + } + + if (writeback && !postinc) { + tcg_gen_addi_i32(addr, addr, imm); + } + if (insn & (1 << 20)) { /* Load. */ tmp = tcg_temp_new_i32(); -- 2.19.0 Add v8M stack checks for the 16-bit Thumb push/pop encodings: STMDB, STMFD, LDM, LDMIA, LDMFD. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181002163556.10279-12-peter.maydell@linaro.org --- target/arm/translate.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static void disas_thumb_insn(DisasContext *s, uint32_t insn) store_reg(s, rd, tmp); break; case 4: case 5: case 0xc: case 0xd: - /* push/pop */ + /* + * 0b1011_x10x_xxxx_xxxx + * - push/pop + */ addr = load_reg(s, 13); if (insn & (1 << 8)) offset = 4; @@ -XXX,XX +XXX,XX @@ static void disas_thumb_insn(DisasContext *s, uint32_t insn) if ((insn & (1 << 11)) == 0) { tcg_gen_addi_i32(addr, addr, -offset); } + + if (s->v8m_stackcheck) { + /* + * Here 'addr' is the lower of "old SP" and "new SP"; + * if this is a pop that starts below the limit and ends + * above it, it is UNKNOWN whether the limit check triggers; + * we choose to trigger. + */ + gen_helper_v8m_stackcheck(cpu_env, addr); + } + for (i = 0; i < 8; i++) { if (insn & (1 << i)) { if (insn & (1 << 11)) { -- 2.19.0 Add the v8M stack checks for the VLDM/VSTM (aka VPUSH/VPOP) instructions. This code is currently unreachable because we haven't yet implemented M profile floating point support, but since the change is simple, we add it now because otherwise we're likely to forget to do it later. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181002163556.10279-13-peter.maydell@linaro.org --- target/arm/translate.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/target/arm/translate.c b/target/arm/translate.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.c +++ b/target/arm/translate.c @@ -XXX,XX +XXX,XX @@ static int disas_vfp_insn(DisasContext *s, uint32_t insn) if (insn & (1 << 24)) /* pre-decrement */ tcg_gen_addi_i32(addr, addr, -((insn & 0xff) << 2)); + if (s->v8m_stackcheck && rn == 13 && w) { + /* + * Here 'addr' is the lowest address we will store to, + * and is either the old SP (if post-increment) or + * the new SP (if pre-decrement). For post-increment + * where the old value is below the limit and the new + * value is above, it is UNKNOWN whether the limit check + * triggers; we choose to trigger. + */ + gen_helper_v8m_stackcheck(cpu_env, addr); + } + if (dp) offset = 8; else -- 2.19.0 Updating the NS stack pointer via MSR to SP_NS should include a check whether the new SP value is below the stack limit. No other kinds of update to the various stack pointer and limit registers via MSR should perform a check. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20181002163556.10279-14-peter.maydell@linaro.org --- target/arm/helper.c | 14 +++++++++++++- 1 file changed, 13 insertions(+), 1 deletion(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(v7m_msr)(CPUARMState *env, uint32_t maskreg, uint32_t val) * currently in handler mode or not, using the NS CONTROL.SPSEL. */ bool spsel = env->v7m.control[M_REG_NS] & R_V7M_CONTROL_SPSEL_MASK; + bool is_psp = !arm_v7m_is_handler_mode(env) && spsel; + uint32_t limit; if (!env->v7m.secure) { return; } - if (!arm_v7m_is_handler_mode(env) && spsel) { + + limit = is_psp ? env->v7m.psplim[false] : env->v7m.msplim[false]; + + if (val < limit) { + CPUState *cs = CPU(arm_env_get_cpu(env)); + + cpu_restore_state(cs, GETPC(), true); + raise_exception(env, EXCP_STKOF, 0, 1); + } + + if (is_psp) { env->v7m.other_ss_psp = val; } else { env->v7m.other_ss_msp = val; -- 2.19.0 Coverity complains (CID 1395628) that the multiply in the calculation of the framebuffer base is performed as 32x32 but then used in a context that takes a 64-bit hwaddr. This can't actually ever overflow the 32-bit result, because of the constraints placed on the s->config values in bcm2835_fb_validate_config(). But we can placate Coverity anyway, by explicitly casting one of the inputs to a hwaddr, so the whole expression is calculated with 64-bit arithmetic. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20181005133012.26490-1-peter.maydell@linaro.org --- hw/display/bcm2835_fb.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/display/bcm2835_fb.c b/hw/display/bcm2835_fb.c index XXXXXXX..XXXXXXX 100644 --- a/hw/display/bcm2835_fb.c +++ b/hw/display/bcm2835_fb.c @@ -XXX,XX +XXX,XX @@ static void fb_update_display(void *opaque) } if (s->invalidate) { - hwaddr base = s->config.base + xoff + yoff * src_width; + hwaddr base = s->config.base + xoff + (hwaddr)yoff * src_width; framebuffer_update_memory_section(&s->fbsection, s->dma_mr, base, s->config.yres, src_width); -- 2.19.0 Nothing earth-shaking in here, just a lot of refactoring and cleanup and a few bugfixes. I suspect I'll have another pullreq to come in the early part of next week... The following changes since commit 19591e9e0938ea5066984553c256a043bd5d822f: Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging (2020-08-27 16:59:02 +0100) are available in the Git repository at: https://git.linaro.org/people/pmaydell/qemu-arm.git tags/pull-target-arm-20200828 for you to fetch changes up to ed78849d9711805bda37ee026018d6ee7a606d0e: target/arm: Convert sq{, r}dmulh to gvec for aa64 advsimd (2020-08-28 10:02:50 +0100) ---------------------------------------------------------------- target-arm queue: * target/arm: Cleanup and refactoring preparatory to SVE2 * armsse: Define ARMSSEClass correctly * hw/misc/unimp: Improve information provided in log messages * hw/qdev-clock: Avoid calling qdev_connect_clock_in after DeviceRealize * hw/arm/xilinx_zynq: Call qdev_connect_clock_in() before DeviceRealize * hw/net/allwinner-sun8i-emac: Use AddressSpace for DMA transfers * hw/sd/allwinner-sdhost: Use AddressSpace for DMA transfers * target/arm: Fill in the WnR syndrome bit in mte_check_fail * target/arm: Clarify HCR_EL2 ARMCPRegInfo type * hw/arm/musicpal: Use AddressSpace for DMA transfers * hw/clock: Minor cleanups * hw/arm/sbsa-ref: fix typo breaking PCIe IRQs ---------------------------------------------------------------- Eduardo Habkost (1): armsse: Define ARMSSEClass correctly Graeme Gregory (1): hw/arm/sbsa-ref: fix typo breaking PCIe IRQs Philippe Mathieu-Daudé (14): hw/clock: Remove unused clock_init*() functions hw/clock: Let clock_set() return boolean value hw/clock: Only propagate clock changes if the clock is changed hw/arm/musicpal: Use AddressSpace for DMA transfers target/arm: Clarify HCR_EL2 ARMCPRegInfo type hw/sd/allwinner-sdhost: Use AddressSpace for DMA transfers hw/net/allwinner-sun8i-emac: Use AddressSpace for DMA transfers hw/arm/xilinx_zynq: Uninline cadence_uart_create() hw/arm/xilinx_zynq: Call qdev_connect_clock_in() before DeviceRealize hw/qdev-clock: Uninline qdev_connect_clock_in() hw/qdev-clock: Avoid calling qdev_connect_clock_in after DeviceRealize hw/misc/unimp: Display value after offset hw/misc/unimp: Display the value with width of the access size hw/misc/unimp: Display the offset with width of the region size Richard Henderson (19): target/arm: Pass the entire mte descriptor to mte_check_fail target/arm: Fill in the WnR syndrome bit in mte_check_fail qemu/int128: Add int128_lshift target/arm: Split out gen_gvec_fn_zz target/arm: Split out gen_gvec_fn_zzz, do_zzz_fn target/arm: Rearrange {sve,fp}_check_access assert target/arm: Merge do_vector2_p into do_mov_p target/arm: Clean up 4-operand predicate expansion target/arm: Use tcg_gen_gvec_bitsel for trans_SEL_pppp target/arm: Split out gen_gvec_ool_zzzp target/arm: Merge helper_sve_clr_* and helper_sve_movz_* target/arm: Split out gen_gvec_ool_zzp target/arm: Split out gen_gvec_ool_zzz target/arm: Split out gen_gvec_ool_zz target/arm: Tidy SVE tszimm shift formats target/arm: Generalize inl_qrdmlah_* helper functions target/arm: Convert integer multiply (indexed) to gvec for aa64 advsimd target/arm: Convert integer multiply-add (indexed) to gvec for aa64 advsimd target/arm: Convert sq{, r}dmulh to gvec for aa64 advsimd include/hw/arm/armsse.h | 2 +- include/hw/char/cadence_uart.h | 17 -- include/hw/clock.h | 30 +-- include/hw/misc/unimp.h | 1 + include/hw/net/allwinner-sun8i-emac.h | 6 + include/hw/qdev-clock.h | 8 +- include/hw/sd/allwinner-sdhost.h | 6 + include/qemu/int128.h | 16 ++ target/arm/helper-sve.h | 5 - target/arm/helper.h | 28 +++ target/arm/translate.h | 1 + target/arm/sve.decode | 35 ++- hw/arm/allwinner-a10.c | 2 + hw/arm/allwinner-h3.c | 4 + hw/arm/armsse.c | 1 + hw/arm/musicpal.c | 45 ++-- hw/arm/sbsa-ref.c | 2 +- hw/arm/xilinx_zynq.c | 24 +- hw/core/clock.c | 7 +- hw/core/qdev-clock.c | 6 + hw/misc/unimp.c | 14 +- hw/net/allwinner-sun8i-emac.c | 46 ++-- hw/sd/allwinner-sdhost.c | 37 +++- target/arm/helper.c | 1 - target/arm/mte_helper.c | 19 +- target/arm/sve_helper.c | 70 ++---- target/arm/translate-a64.c | 110 ++++++++-- target/arm/translate-sve.c | 399 ++++++++++++++-------------------- target/arm/vec_helper.c | 182 +++++++++++----- 29 files changed, 629 insertions(+), 495 deletions(-) From: Graeme Gregory <graeme@nuviainc.com> Fixing a typo in a previous patch that translated an "i" to a 1 and therefore breaking the allocation of PCIe interrupts. This was discovered when virtio-net-pci devices ceased to function correctly. Cc: qemu-stable@nongnu.org Fixes: 48ba18e6d3f3 ("hw/arm/sbsa-ref: Simplify by moving the gic in the machine state") Signed-off-by: Graeme Gregory <graeme@nuviainc.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200821083853.356490-1-graeme@nuviainc.com Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/sbsa-ref.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hw/arm/sbsa-ref.c b/hw/arm/sbsa-ref.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/sbsa-ref.c +++ b/hw/arm/sbsa-ref.c @@ -XXX,XX +XXX,XX @@ static void create_pcie(SBSAMachineState *sms) for (i = 0; i < GPEX_NUM_IRQS; i++) { sysbus_connect_irq(SYS_BUS_DEVICE(dev), i, - qdev_get_gpio_in(sms->gic, irq + 1)); + qdev_get_gpio_in(sms->gic, irq + i)); gpex_set_irq_num(GPEX_HOST(dev), i, irq + i); } -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> clock_init*() inlined funtions are simple wrappers around clock_set*() and are not used. Remove them in favor of clock_set*(). Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200806123858.30058-2-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/clock.h | 13 ------------- 1 file changed, 13 deletions(-) diff --git a/include/hw/clock.h b/include/hw/clock.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/clock.h +++ b/include/hw/clock.h @@ -XXX,XX +XXX,XX @@ static inline bool clock_is_enabled(const Clock *clk) return clock_get(clk) != 0; } -static inline void clock_init(Clock *clk, uint64_t value) -{ - clock_set(clk, value); -} -static inline void clock_init_hz(Clock *clk, uint64_t value) -{ - clock_set_hz(clk, value); -} -static inline void clock_init_ns(Clock *clk, uint64_t value) -{ - clock_set_ns(clk, value); -} - #endif /* QEMU_HW_CLOCK_H */ -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Let clock_set() return a boolean value whether the clock has been updated or not. Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200806123858.30058-3-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/clock.h | 12 +++++++----- hw/core/clock.c | 7 ++++++- 2 files changed, 13 insertions(+), 6 deletions(-) diff --git a/include/hw/clock.h b/include/hw/clock.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/clock.h +++ b/include/hw/clock.h @@ -XXX,XX +XXX,XX @@ void clock_set_source(Clock *clk, Clock *src); * @value: the clock's value, 0 means unclocked * * Set the local cached period value of @clk to @value. + * + * @return: true if the clock is changed. */ -void clock_set(Clock *clk, uint64_t value); +bool clock_set(Clock *clk, uint64_t value); -static inline void clock_set_hz(Clock *clk, unsigned hz) +static inline bool clock_set_hz(Clock *clk, unsigned hz) { - clock_set(clk, CLOCK_PERIOD_FROM_HZ(hz)); + return clock_set(clk, CLOCK_PERIOD_FROM_HZ(hz)); } -static inline void clock_set_ns(Clock *clk, unsigned ns) +static inline bool clock_set_ns(Clock *clk, unsigned ns) { - clock_set(clk, CLOCK_PERIOD_FROM_NS(ns)); + return clock_set(clk, CLOCK_PERIOD_FROM_NS(ns)); } /** diff --git a/hw/core/clock.c b/hw/core/clock.c index XXXXXXX..XXXXXXX 100644 --- a/hw/core/clock.c +++ b/hw/core/clock.c @@ -XXX,XX +XXX,XX @@ void clock_clear_callback(Clock *clk) clock_set_callback(clk, NULL, NULL); } -void clock_set(Clock *clk, uint64_t period) +bool clock_set(Clock *clk, uint64_t period) { + if (clk->period == period) { + return false; + } trace_clock_set(CLOCK_PATH(clk), CLOCK_PERIOD_TO_NS(clk->period), CLOCK_PERIOD_TO_NS(period)); clk->period = period; + + return true; } static void clock_propagate_period(Clock *clk, bool call_callbacks) -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Avoid propagating the clock change when the clock does not change. Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200806123858.30058-4-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/clock.h | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/include/hw/clock.h b/include/hw/clock.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/clock.h +++ b/include/hw/clock.h @@ -XXX,XX +XXX,XX @@ void clock_propagate(Clock *clk); */ static inline void clock_update(Clock *clk, uint64_t value) { - clock_set(clk, value); - clock_propagate(clk); + if (clock_set(clk, value)) { + clock_propagate(clk); + } } static inline void clock_update_hz(Clock *clk, unsigned hz) -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Allow the device to execute the DMA transfers in a different AddressSpace. We keep using the system_memory address space, but via the proper dma_memory_access() API. Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200814125533.4047-1-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/musicpal.c | 45 +++++++++++++++++++++++++++++++-------------- 1 file changed, 31 insertions(+), 14 deletions(-) diff --git a/hw/arm/musicpal.c b/hw/arm/musicpal.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/musicpal.c +++ b/hw/arm/musicpal.c @@ -XXX,XX +XXX,XX @@ #include "hw/audio/wm8750.h" #include "sysemu/block-backend.h" #include "sysemu/runstate.h" +#include "sysemu/dma.h" #include "exec/address-spaces.h" #include "ui/pixel_ops.h" #include "qemu/cutils.h" @@ -XXX,XX +XXX,XX @@ typedef struct mv88w8618_eth_state { MemoryRegion iomem; qemu_irq irq; + MemoryRegion *dma_mr; + AddressSpace dma_as; uint32_t smir; uint32_t icr; uint32_t imr; @@ -XXX,XX +XXX,XX @@ typedef struct mv88w8618_eth_state { NICConf conf; } mv88w8618_eth_state; -static void eth_rx_desc_put(uint32_t addr, mv88w8618_rx_desc *desc) +static void eth_rx_desc_put(AddressSpace *dma_as, uint32_t addr, + mv88w8618_rx_desc *desc) { cpu_to_le32s(&desc->cmdstat); cpu_to_le16s(&desc->bytes); cpu_to_le16s(&desc->buffer_size); cpu_to_le32s(&desc->buffer); cpu_to_le32s(&desc->next); - cpu_physical_memory_write(addr, desc, sizeof(*desc)); + dma_memory_write(dma_as, addr, desc, sizeof(*desc)); } -static void eth_rx_desc_get(uint32_t addr, mv88w8618_rx_desc *desc) +static void eth_rx_desc_get(AddressSpace *dma_as, uint32_t addr, + mv88w8618_rx_desc *desc) { - cpu_physical_memory_read(addr, desc, sizeof(*desc)); + dma_memory_read(dma_as, addr, desc, sizeof(*desc)); le32_to_cpus(&desc->cmdstat); le16_to_cpus(&desc->bytes); le16_to_cpus(&desc->buffer_size); @@ -XXX,XX +XXX,XX @@ static ssize_t eth_receive(NetClientState *nc, const uint8_t *buf, size_t size) continue; } do { - eth_rx_desc_get(desc_addr, &desc); + eth_rx_desc_get(&s->dma_as, desc_addr, &desc); if ((desc.cmdstat & MP_ETH_RX_OWN) && desc.buffer_size >= size) { - cpu_physical_memory_write(desc.buffer + s->vlan_header, + dma_memory_write(&s->dma_as, desc.buffer + s->vlan_header, buf, size); desc.bytes = size + s->vlan_header; desc.cmdstat &= ~MP_ETH_RX_OWN; @@ -XXX,XX +XXX,XX @@ static ssize_t eth_receive(NetClientState *nc, const uint8_t *buf, size_t size) if (s->icr & s->imr) { qemu_irq_raise(s->irq); } - eth_rx_desc_put(desc_addr, &desc); + eth_rx_desc_put(&s->dma_as, desc_addr, &desc); return size; } desc_addr = desc.next; @@ -XXX,XX +XXX,XX @@ static ssize_t eth_receive(NetClientState *nc, const uint8_t *buf, size_t size) return size; } -static void eth_tx_desc_put(uint32_t addr, mv88w8618_tx_desc *desc) +static void eth_tx_desc_put(AddressSpace *dma_as, uint32_t addr, + mv88w8618_tx_desc *desc) { cpu_to_le32s(&desc->cmdstat); cpu_to_le16s(&desc->res); cpu_to_le16s(&desc->bytes); cpu_to_le32s(&desc->buffer); cpu_to_le32s(&desc->next); - cpu_physical_memory_write(addr, desc, sizeof(*desc)); + dma_memory_write(dma_as, addr, desc, sizeof(*desc)); } -static void eth_tx_desc_get(uint32_t addr, mv88w8618_tx_desc *desc) +static void eth_tx_desc_get(AddressSpace *dma_as, uint32_t addr, + mv88w8618_tx_desc *desc) { - cpu_physical_memory_read(addr, desc, sizeof(*desc)); + dma_memory_read(dma_as, addr, desc, sizeof(*desc)); le32_to_cpus(&desc->cmdstat); le16_to_cpus(&desc->res); le16_to_cpus(&desc->bytes); @@ -XXX,XX +XXX,XX @@ static void eth_send(mv88w8618_eth_state *s, int queue_index) int len; do { - eth_tx_desc_get(desc_addr, &desc); + eth_tx_desc_get(&s->dma_as, desc_addr, &desc); next_desc = desc.next; if (desc.cmdstat & MP_ETH_TX_OWN) { len = desc.bytes; if (len < 2048) { - cpu_physical_memory_read(desc.buffer, buf, len); + dma_memory_read(&s->dma_as, desc.buffer, buf, len); qemu_send_packet(qemu_get_queue(s->nic), buf, len); } desc.cmdstat &= ~MP_ETH_TX_OWN; s->icr |= 1 << (MP_ETH_IRQ_TXLO_BIT - queue_index); - eth_tx_desc_put(desc_addr, &desc); + eth_tx_desc_put(&s->dma_as, desc_addr, &desc); } desc_addr = next_desc; } while (desc_addr != s->tx_queue[queue_index]); @@ -XXX,XX +XXX,XX @@ static void mv88w8618_eth_realize(DeviceState *dev, Error **errp) { mv88w8618_eth_state *s = MV88W8618_ETH(dev); + if (!s->dma_mr) { + error_setg(errp, TYPE_MV88W8618_ETH " 'dma-memory' link not set"); + return; + } + + address_space_init(&s->dma_as, s->dma_mr, "emac-dma"); s->nic = qemu_new_nic(&net_mv88w8618_info, &s->conf, object_get_typename(OBJECT(dev)), dev->id, s); } @@ -XXX,XX +XXX,XX @@ static const VMStateDescription mv88w8618_eth_vmsd = { static Property mv88w8618_eth_properties[] = { DEFINE_NIC_PROPERTIES(mv88w8618_eth_state, conf), + DEFINE_PROP_LINK("dma-memory", mv88w8618_eth_state, dma_mr, + TYPE_MEMORY_REGION, MemoryRegion *), DEFINE_PROP_END_OF_LIST(), }; @@ -XXX,XX +XXX,XX @@ static void musicpal_init(MachineState *machine) qemu_check_nic_model(&nd_table[0], "mv88w8618"); dev = qdev_new(TYPE_MV88W8618_ETH); qdev_set_nic_properties(dev, &nd_table[0]); + object_property_set_link(OBJECT(dev), "dma-memory", + OBJECT(get_system_memory()), &error_fatal); sysbus_realize_and_unref(SYS_BUS_DEVICE(dev), &error_fatal); sysbus_mmio_map(SYS_BUS_DEVICE(dev), 0, MP_ETH_BASE); sysbus_connect_irq(SYS_BUS_DEVICE(dev), 0, pic[MP_ETH_IRQ]); -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> In commit ce4afed839 ("target/arm: Implement AArch32 HCR and HCR2") the HCR_EL2 register has been changed from type NO_RAW (no underlying state and does not support raw access for state saving/loading) to type CONST (TCG can assume the value to be constant), removing the read/write accessors. We forgot to remove the previous type ARM_CP_NO_RAW. This is not really a problem since the field is overwritten. However it makes code review confuse, so remove it. Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200812111223.7787-1-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.c | 1 - 1 file changed, 1 deletion(-) diff --git a/target/arm/helper.c b/target/arm/helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.c +++ b/target/arm/helper.c @@ -XXX,XX +XXX,XX @@ static const ARMCPRegInfo el3_no_el2_cp_reginfo[] = { .access = PL2_RW, .readfn = arm_cp_read_zero, .writefn = arm_cp_write_ignore }, { .name = "HCR_EL2", .state = ARM_CP_STATE_BOTH, - .type = ARM_CP_NO_RAW, .opc0 = 3, .opc1 = 4, .crn = 1, .crm = 1, .opc2 = 0, .access = PL2_RW, .type = ARM_CP_CONST, .resetvalue = 0 }, -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> We need more information than just the mmu_idx in order to create the proper exception syndrome. Only change the function signature so far. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200813200816.3037186-2-richard.henderson@linaro.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/mte_helper.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(stzgm_tags)(CPUARMState *env, uint64_t ptr, uint64_t val) } /* Record a tag check failure. */ -static void mte_check_fail(CPUARMState *env, int mmu_idx, +static void mte_check_fail(CPUARMState *env, uint32_t desc, uint64_t dirty_ptr, uintptr_t ra) { + int mmu_idx = FIELD_EX32(desc, MTEDESC, MIDX); ARMMMUIdx arm_mmu_idx = core_to_aa64_mmu_idx(mmu_idx); int el, reg_el, tcf, select; uint64_t sctlr; @@ -XXX,XX +XXX,XX @@ uint64_t mte_check1(CPUARMState *env, uint32_t desc, } if (unlikely(!mte_probe1_int(env, desc, ptr, ra, bit55))) { - int mmu_idx = FIELD_EX32(desc, MTEDESC, MIDX); - mte_check_fail(env, mmu_idx, ptr, ra); + mte_check_fail(env, desc, ptr, ra); } return useronly_clean_ptr(ptr); @@ -XXX,XX +XXX,XX @@ uint64_t mte_checkN(CPUARMState *env, uint32_t desc, fail_ofs = tag_first + n * TAG_GRANULE - ptr; fail_ofs = ROUND_UP(fail_ofs, esize); - mte_check_fail(env, mmu_idx, ptr + fail_ofs, ra); + mte_check_fail(env, desc, ptr + fail_ofs, ra); } done: @@ -XXX,XX +XXX,XX @@ uint64_t HELPER(mte_check_zva)(CPUARMState *env, uint32_t desc, uint64_t ptr) fail: /* Locate the first nibble that differs. */ i = ctz64(mem_tag ^ ptr_tag) >> 4; - mte_check_fail(env, mmu_idx, align_ptr + i * TAG_GRANULE, ra); + mte_check_fail(env, desc, align_ptr + i * TAG_GRANULE, ra); done: return useronly_clean_ptr(ptr); -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> According to AArch64.TagCheckFault, none of the other ISS values are provided, so we do not need to go so far as merge_syn_data_abort. But we were missing the WnR bit. Tested-by: Andrey Konovalov <andreyknvl@google.com> Reported-by: Andrey Konovalov <andreyknvl@google.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Message-id: 20200813200816.3037186-3-richard.henderson@linaro.org Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/mte_helper.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/target/arm/mte_helper.c b/target/arm/mte_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/mte_helper.c +++ b/target/arm/mte_helper.c @@ -XXX,XX +XXX,XX @@ static void mte_check_fail(CPUARMState *env, uint32_t desc, { int mmu_idx = FIELD_EX32(desc, MTEDESC, MIDX); ARMMMUIdx arm_mmu_idx = core_to_aa64_mmu_idx(mmu_idx); - int el, reg_el, tcf, select; + int el, reg_el, tcf, select, is_write, syn; uint64_t sctlr; reg_el = regime_el(env, arm_mmu_idx); @@ -XXX,XX +XXX,XX @@ static void mte_check_fail(CPUARMState *env, uint32_t desc, */ cpu_restore_state(env_cpu(env), ra, true); env->exception.vaddress = dirty_ptr; - raise_exception(env, EXCP_DATA_ABORT, - syn_data_abort_no_iss(el != 0, 0, 0, 0, 0, 0, 0x11), - exception_target_el(env)); + + is_write = FIELD_EX32(desc, MTEDESC, WRITE); + syn = syn_data_abort_no_iss(el != 0, 0, 0, 0, 0, is_write, 0x11); + raise_exception(env, EXCP_DATA_ABORT, syn, exception_target_el(env)); /* noreturn, but fall through to the assert anyway */ case 0: -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Allow the device to execute the DMA transfers in a different AddressSpace. The A10 and H3 SoC keep using the system_memory address space, but via the proper dma_memory_access() API. Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Niek Linnenbank <nieklinnenbank@gmail.com> Reviewed-by: Niek Linnenbank <nieklinnenbank@gmail.com> Message-id: 20200814110057.307-1-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/sd/allwinner-sdhost.h | 6 ++++++ hw/arm/allwinner-a10.c | 2 ++ hw/arm/allwinner-h3.c | 2 ++ hw/sd/allwinner-sdhost.c | 37 ++++++++++++++++++++++++++------ 4 files changed, 41 insertions(+), 6 deletions(-) diff --git a/include/hw/sd/allwinner-sdhost.h b/include/hw/sd/allwinner-sdhost.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/sd/allwinner-sdhost.h +++ b/include/hw/sd/allwinner-sdhost.h @@ -XXX,XX +XXX,XX @@ typedef struct AwSdHostState { /** Interrupt output signal to notify CPU */ qemu_irq irq; + /** Memory region where DMA transfers are done */ + MemoryRegion *dma_mr; + + /** Address space used internally for DMA transfers */ + AddressSpace dma_as; + /** Number of bytes left in current DMA transfer */ uint32_t transfer_cnt; diff --git a/hw/arm/allwinner-a10.c b/hw/arm/allwinner-a10.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/allwinner-a10.c +++ b/hw/arm/allwinner-a10.c @@ -XXX,XX +XXX,XX @@ static void aw_a10_realize(DeviceState *dev, Error **errp) } /* SD/MMC */ + object_property_set_link(OBJECT(&s->mmc0), "dma-memory", + OBJECT(get_system_memory()), &error_fatal); sysbus_realize(SYS_BUS_DEVICE(&s->mmc0), &error_fatal); sysbus_mmio_map(SYS_BUS_DEVICE(&s->mmc0), 0, AW_A10_MMC0_BASE); sysbus_connect_irq(SYS_BUS_DEVICE(&s->mmc0), 0, qdev_get_gpio_in(dev, 32)); diff --git a/hw/arm/allwinner-h3.c b/hw/arm/allwinner-h3.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/allwinner-h3.c +++ b/hw/arm/allwinner-h3.c @@ -XXX,XX +XXX,XX @@ static void allwinner_h3_realize(DeviceState *dev, Error **errp) sysbus_mmio_map(SYS_BUS_DEVICE(&s->sid), 0, s->memmap[AW_H3_SID]); /* SD/MMC */ + object_property_set_link(OBJECT(&s->mmc0), "dma-memory", + OBJECT(get_system_memory()), &error_fatal); sysbus_realize(SYS_BUS_DEVICE(&s->mmc0), &error_fatal); sysbus_mmio_map(SYS_BUS_DEVICE(&s->mmc0), 0, s->memmap[AW_H3_MMC0]); sysbus_connect_irq(SYS_BUS_DEVICE(&s->mmc0), 0, diff --git a/hw/sd/allwinner-sdhost.c b/hw/sd/allwinner-sdhost.c index XXXXXXX..XXXXXXX 100644 --- a/hw/sd/allwinner-sdhost.c +++ b/hw/sd/allwinner-sdhost.c @@ -XXX,XX +XXX,XX @@ #include "qemu/log.h" #include "qemu/module.h" #include "qemu/units.h" +#include "qapi/error.h" #include "sysemu/blockdev.h" +#include "sysemu/dma.h" +#include "hw/qdev-properties.h" #include "hw/irq.h" #include "hw/sd/allwinner-sdhost.h" #include "migration/vmstate.h" @@ -XXX,XX +XXX,XX @@ static uint32_t allwinner_sdhost_process_desc(AwSdHostState *s, uint8_t buf[1024]; /* Read descriptor */ - cpu_physical_memory_read(desc_addr, desc, sizeof(*desc)); + dma_memory_read(&s->dma_as, desc_addr, desc, sizeof(*desc)); if (desc->size == 0) { desc->size = klass->max_desc_size; } else if (desc->size > klass->max_desc_size) { @@ -XXX,XX +XXX,XX @@ static uint32_t allwinner_sdhost_process_desc(AwSdHostState *s, /* Write to SD bus */ if (is_write) { - cpu_physical_memory_read((desc->addr & DESC_SIZE_MASK) + num_done, - buf, buf_bytes); + dma_memory_read(&s->dma_as, + (desc->addr & DESC_SIZE_MASK) + num_done, + buf, buf_bytes); sdbus_write_data(&s->sdbus, buf, buf_bytes); /* Read from SD bus */ } else { sdbus_read_data(&s->sdbus, buf, buf_bytes); - cpu_physical_memory_write((desc->addr & DESC_SIZE_MASK) + num_done, - buf, buf_bytes); + dma_memory_write(&s->dma_as, + (desc->addr & DESC_SIZE_MASK) + num_done, + buf, buf_bytes); } num_done += buf_bytes; } /* Clear hold flag and flush descriptor */ desc->status &= ~DESC_STATUS_HOLD; - cpu_physical_memory_write(desc_addr, desc, sizeof(*desc)); + dma_memory_write(&s->dma_as, desc_addr, desc, sizeof(*desc)); return num_done; } @@ -XXX,XX +XXX,XX @@ static const VMStateDescription vmstate_allwinner_sdhost = { } }; +static Property allwinner_sdhost_properties[] = { + DEFINE_PROP_LINK("dma-memory", AwSdHostState, dma_mr, + TYPE_MEMORY_REGION, MemoryRegion *), + DEFINE_PROP_END_OF_LIST(), +}; + static void allwinner_sdhost_init(Object *obj) { AwSdHostState *s = AW_SDHOST(obj); @@ -XXX,XX +XXX,XX @@ static void allwinner_sdhost_init(Object *obj) sysbus_init_irq(SYS_BUS_DEVICE(s), &s->irq); } +static void allwinner_sdhost_realize(DeviceState *dev, Error **errp) +{ + AwSdHostState *s = AW_SDHOST(dev); + + if (!s->dma_mr) { + error_setg(errp, TYPE_AW_SDHOST " 'dma-memory' link not set"); + return; + } + + address_space_init(&s->dma_as, s->dma_mr, "sdhost-dma"); +} + static void allwinner_sdhost_reset(DeviceState *dev) { AwSdHostState *s = AW_SDHOST(dev); @@ -XXX,XX +XXX,XX @@ static void allwinner_sdhost_class_init(ObjectClass *klass, void *data) dc->reset = allwinner_sdhost_reset; dc->vmsd = &vmstate_allwinner_sdhost; + dc->realize = allwinner_sdhost_realize; + device_class_set_props(dc, allwinner_sdhost_properties); } static void allwinner_sdhost_sun4i_class_init(ObjectClass *klass, void *data) -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Allow the device to execute the DMA transfers in a different AddressSpace. The H3 SoC keeps using the system_memory address space, but via the proper dma_memory_access() API. Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Niek Linnenbank <nieklinnenbank@gmail.com> Tested-by: Niek Linnenbank <nieklinnenbank@gmail.com> Message-id: 20200814122907.27732-1-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/net/allwinner-sun8i-emac.h | 6 ++++ hw/arm/allwinner-h3.c | 2 ++ hw/net/allwinner-sun8i-emac.c | 46 +++++++++++++++++---------- 3 files changed, 38 insertions(+), 16 deletions(-) diff --git a/include/hw/net/allwinner-sun8i-emac.h b/include/hw/net/allwinner-sun8i-emac.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/net/allwinner-sun8i-emac.h +++ b/include/hw/net/allwinner-sun8i-emac.h @@ -XXX,XX +XXX,XX @@ typedef struct AwSun8iEmacState { /** Interrupt output signal to notify CPU */ qemu_irq irq; + /** Memory region where DMA transfers are done */ + MemoryRegion *dma_mr; + + /** Address space used internally for DMA transfers */ + AddressSpace dma_as; + /** Generic Network Interface Controller (NIC) for networking API */ NICState *nic; diff --git a/hw/arm/allwinner-h3.c b/hw/arm/allwinner-h3.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/allwinner-h3.c +++ b/hw/arm/allwinner-h3.c @@ -XXX,XX +XXX,XX @@ static void allwinner_h3_realize(DeviceState *dev, Error **errp) qemu_check_nic_model(&nd_table[0], TYPE_AW_SUN8I_EMAC); qdev_set_nic_properties(DEVICE(&s->emac), &nd_table[0]); } + object_property_set_link(OBJECT(&s->emac), "dma-memory", + OBJECT(get_system_memory()), &error_fatal); sysbus_realize(SYS_BUS_DEVICE(&s->emac), &error_fatal); sysbus_mmio_map(SYS_BUS_DEVICE(&s->emac), 0, s->memmap[AW_H3_EMAC]); sysbus_connect_irq(SYS_BUS_DEVICE(&s->emac), 0, diff --git a/hw/net/allwinner-sun8i-emac.c b/hw/net/allwinner-sun8i-emac.c index XXXXXXX..XXXXXXX 100644 --- a/hw/net/allwinner-sun8i-emac.c +++ b/hw/net/allwinner-sun8i-emac.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "qemu/units.h" +#include "qapi/error.h" #include "hw/sysbus.h" #include "migration/vmstate.h" #include "net/net.h" @@ -XXX,XX +XXX,XX @@ #include "net/checksum.h" #include "qemu/module.h" #include "exec/cpu-common.h" +#include "sysemu/dma.h" #include "hw/net/allwinner-sun8i-emac.h" /* EMAC register offsets */ @@ -XXX,XX +XXX,XX @@ static void allwinner_sun8i_emac_update_irq(AwSun8iEmacState *s) qemu_set_irq(s->irq, (s->int_sta & s->int_en) != 0); } -static uint32_t allwinner_sun8i_emac_next_desc(FrameDescriptor *desc, +static uint32_t allwinner_sun8i_emac_next_desc(AwSun8iEmacState *s, + FrameDescriptor *desc, size_t min_size) { uint32_t paddr = desc->next; - cpu_physical_memory_read(paddr, desc, sizeof(*desc)); + dma_memory_read(&s->dma_as, paddr, desc, sizeof(*desc)); if ((desc->status & DESC_STATUS_CTL) && (desc->status2 & DESC_STATUS2_BUF_SIZE_MASK) >= min_size) { @@ -XXX,XX +XXX,XX @@ static uint32_t allwinner_sun8i_emac_next_desc(FrameDescriptor *desc, } } -static uint32_t allwinner_sun8i_emac_get_desc(FrameDescriptor *desc, +static uint32_t allwinner_sun8i_emac_get_desc(AwSun8iEmacState *s, + FrameDescriptor *desc, uint32_t start_addr, size_t min_size) { @@ -XXX,XX +XXX,XX @@ static uint32_t allwinner_sun8i_emac_get_desc(FrameDescriptor *desc, /* Note that the list is a cycle. Last entry points back to the head. */ while (desc_addr != 0) { - cpu_physical_memory_read(desc_addr, desc, sizeof(*desc)); + dma_memory_read(&s->dma_as, desc_addr, desc, sizeof(*desc)); if ((desc->status & DESC_STATUS_CTL) && (desc->status2 & DESC_STATUS2_BUF_SIZE_MASK) >= min_size) { @@ -XXX,XX +XXX,XX @@ static uint32_t allwinner_sun8i_emac_rx_desc(AwSun8iEmacState *s, FrameDescriptor *desc, size_t min_size) { - return allwinner_sun8i_emac_get_desc(desc, s->rx_desc_curr, min_size); + return allwinner_sun8i_emac_get_desc(s, desc, s->rx_desc_curr, min_size); } static uint32_t allwinner_sun8i_emac_tx_desc(AwSun8iEmacState *s, FrameDescriptor *desc, size_t min_size) { - return allwinner_sun8i_emac_get_desc(desc, s->tx_desc_head, min_size); + return allwinner_sun8i_emac_get_desc(s, desc, s->tx_desc_head, min_size); } -static void allwinner_sun8i_emac_flush_desc(FrameDescriptor *desc, +static void allwinner_sun8i_emac_flush_desc(AwSun8iEmacState *s, + FrameDescriptor *desc, uint32_t phys_addr) { - cpu_physical_memory_write(phys_addr, desc, sizeof(*desc)); + dma_memory_write(&s->dma_as, phys_addr, desc, sizeof(*desc)); } static bool allwinner_sun8i_emac_can_receive(NetClientState *nc) @@ -XXX,XX +XXX,XX @@ static ssize_t allwinner_sun8i_emac_receive(NetClientState *nc, << RX_DESC_STATUS_FRM_LEN_SHIFT; } - cpu_physical_memory_write(desc.addr, buf, desc_bytes); - allwinner_sun8i_emac_flush_desc(&desc, s->rx_desc_curr); + dma_memory_write(&s->dma_as, desc.addr, buf, desc_bytes); + allwinner_sun8i_emac_flush_desc(s, &desc, s->rx_desc_curr); trace_allwinner_sun8i_emac_receive(s->rx_desc_curr, desc.addr, desc_bytes); @@ -XXX,XX +XXX,XX @@ static ssize_t allwinner_sun8i_emac_receive(NetClientState *nc, bytes_left -= desc_bytes; /* Move to the next descriptor */ - s->rx_desc_curr = allwinner_sun8i_emac_next_desc(&desc, 64); + s->rx_desc_curr = allwinner_sun8i_emac_next_desc(s, &desc, 64); if (!s->rx_desc_curr) { /* Not enough buffer space available */ s->int_sta |= INT_STA_RX_BUF_UA; @@ -XXX,XX +XXX,XX @@ static void allwinner_sun8i_emac_transmit(AwSun8iEmacState *s) desc.status |= TX_DESC_STATUS_LENGTH_ERR; break; } - cpu_physical_memory_read(desc.addr, packet_buf + packet_bytes, bytes); + dma_memory_read(&s->dma_as, desc.addr, packet_buf + packet_bytes, bytes); packet_bytes += bytes; desc.status &= ~DESC_STATUS_CTL; - allwinner_sun8i_emac_flush_desc(&desc, s->tx_desc_curr); + allwinner_sun8i_emac_flush_desc(s, &desc, s->tx_desc_curr); /* After the last descriptor, send the packet */ if (desc.status2 & TX_DESC_STATUS2_LAST_DESC) { @@ -XXX,XX +XXX,XX @@ static void allwinner_sun8i_emac_transmit(AwSun8iEmacState *s) packet_bytes = 0; transmitted++; } - s->tx_desc_curr = allwinner_sun8i_emac_next_desc(&desc, 0); + s->tx_desc_curr = allwinner_sun8i_emac_next_desc(s, &desc, 0); } /* Raise transmit completed interrupt */ @@ -XXX,XX +XXX,XX @@ static uint64_t allwinner_sun8i_emac_read(void *opaque, hwaddr offset, break; case REG_TX_CUR_BUF: /* Transmit Current Buffer */ if (s->tx_desc_curr != 0) { - cpu_physical_memory_read(s->tx_desc_curr, &desc, sizeof(desc)); + dma_memory_read(&s->dma_as, s->tx_desc_curr, &desc, sizeof(desc)); value = desc.addr; } else { value = 0; @@ -XXX,XX +XXX,XX @@ static uint64_t allwinner_sun8i_emac_read(void *opaque, hwaddr offset, break; case REG_RX_CUR_BUF: /* Receive Current Buffer */ if (s->rx_desc_curr != 0) { - cpu_physical_memory_read(s->rx_desc_curr, &desc, sizeof(desc)); + dma_memory_read(&s->dma_as, s->rx_desc_curr, &desc, sizeof(desc)); value = desc.addr; } else { value = 0; @@ -XXX,XX +XXX,XX @@ static void allwinner_sun8i_emac_realize(DeviceState *dev, Error **errp) { AwSun8iEmacState *s = AW_SUN8I_EMAC(dev); + if (!s->dma_mr) { + error_setg(errp, TYPE_AW_SUN8I_EMAC " 'dma-memory' link not set"); + return; + } + + address_space_init(&s->dma_as, s->dma_mr, "emac-dma"); + qemu_macaddr_default_if_unset(&s->conf.macaddr); s->nic = qemu_new_nic(&net_allwinner_sun8i_emac_info, &s->conf, object_get_typename(OBJECT(dev)), dev->id, s); @@ -XXX,XX +XXX,XX @@ static void allwinner_sun8i_emac_realize(DeviceState *dev, Error **errp) static Property allwinner_sun8i_emac_properties[] = { DEFINE_NIC_PROPERTIES(AwSun8iEmacState, conf), DEFINE_PROP_UINT8("phy-addr", AwSun8iEmacState, mii_phy_addr, 0), + DEFINE_PROP_LINK("dma-memory", AwSun8iEmacState, dma_mr, + TYPE_MEMORY_REGION, MemoryRegion *), DEFINE_PROP_END_OF_LIST(), }; -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> As we want to call qdev_connect_clock_in() before the device is realized, we need to uninline cadence_uart_create() first. Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Message-id: 20200803105647.22223-2-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/char/cadence_uart.h | 17 ----------------- hw/arm/xilinx_zynq.c | 14 ++++++++++++-- 2 files changed, 12 insertions(+), 19 deletions(-) diff --git a/include/hw/char/cadence_uart.h b/include/hw/char/cadence_uart.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/char/cadence_uart.h +++ b/include/hw/char/cadence_uart.h @@ -XXX,XX +XXX,XX @@ typedef struct { Clock *refclk; } CadenceUARTState; -static inline DeviceState *cadence_uart_create(hwaddr addr, - qemu_irq irq, - Chardev *chr) -{ - DeviceState *dev; - SysBusDevice *s; - - dev = qdev_new(TYPE_CADENCE_UART); - s = SYS_BUS_DEVICE(dev); - qdev_prop_set_chr(dev, "chardev", chr); - sysbus_realize_and_unref(s, &error_fatal); - sysbus_mmio_map(s, 0, addr); - sysbus_connect_irq(s, 0, irq); - - return dev; -} - #endif diff --git a/hw/arm/xilinx_zynq.c b/hw/arm/xilinx_zynq.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xilinx_zynq.c +++ b/hw/arm/xilinx_zynq.c @@ -XXX,XX +XXX,XX @@ static void zynq_init(MachineState *machine) sysbus_create_simple(TYPE_CHIPIDEA, 0xE0002000, pic[53 - IRQ_OFFSET]); sysbus_create_simple(TYPE_CHIPIDEA, 0xE0003000, pic[76 - IRQ_OFFSET]); - dev = cadence_uart_create(0xE0000000, pic[59 - IRQ_OFFSET], serial_hd(0)); + dev = qdev_new(TYPE_CADENCE_UART); + busdev = SYS_BUS_DEVICE(dev); + qdev_prop_set_chr(dev, "chardev", serial_hd(0)); + sysbus_realize_and_unref(busdev, &error_fatal); + sysbus_mmio_map(busdev, 0, 0xE0000000); + sysbus_connect_irq(busdev, 0, pic[59 - IRQ_OFFSET]); qdev_connect_clock_in(dev, "refclk", qdev_get_clock_out(slcr, "uart0_ref_clk")); - dev = cadence_uart_create(0xE0001000, pic[82 - IRQ_OFFSET], serial_hd(1)); + dev = qdev_new(TYPE_CADENCE_UART); + busdev = SYS_BUS_DEVICE(dev); + qdev_prop_set_chr(dev, "chardev", serial_hd(1)); + sysbus_realize_and_unref(busdev, &error_fatal); + sysbus_mmio_map(busdev, 0, 0xE0001000); + sysbus_connect_irq(busdev, 0, pic[82 - IRQ_OFFSET]); qdev_connect_clock_in(dev, "refclk", qdev_get_clock_out(slcr, "uart1_ref_clk")); -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Clock canonical name is set in device_set_realized (see the block added to hw/core/qdev.c in commit 0e6934f264). If we connect a clock after the device is realized, this code is not executed. This is currently not a problem as this name is only used for trace events, however this disrupt tracing. Fix by calling qdev_connect_clock_in() before realizing. Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Message-id: 20200803105647.22223-3-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/arm/xilinx_zynq.c | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) diff --git a/hw/arm/xilinx_zynq.c b/hw/arm/xilinx_zynq.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/xilinx_zynq.c +++ b/hw/arm/xilinx_zynq.c @@ -XXX,XX +XXX,XX @@ static void zynq_init(MachineState *machine) 1, 0x0066, 0x0022, 0x0000, 0x0000, 0x0555, 0x2aa, 0); - /* Create slcr, keep a pointer to connect clocks */ - slcr = qdev_new("xilinx,zynq_slcr"); - sysbus_realize_and_unref(SYS_BUS_DEVICE(slcr), &error_fatal); - sysbus_mmio_map(SYS_BUS_DEVICE(slcr), 0, 0xF8000000); - /* Create the main clock source, and feed slcr with it */ zynq_machine->ps_clk = CLOCK(object_new(TYPE_CLOCK)); object_property_add_child(OBJECT(zynq_machine), "ps_clk", OBJECT(zynq_machine->ps_clk)); object_unref(OBJECT(zynq_machine->ps_clk)); clock_set_hz(zynq_machine->ps_clk, PS_CLK_FREQUENCY); + + /* Create slcr, keep a pointer to connect clocks */ + slcr = qdev_new("xilinx,zynq_slcr"); qdev_connect_clock_in(slcr, "ps_clk", zynq_machine->ps_clk); + sysbus_realize_and_unref(SYS_BUS_DEVICE(slcr), &error_fatal); + sysbus_mmio_map(SYS_BUS_DEVICE(slcr), 0, 0xF8000000); dev = qdev_new(TYPE_A9MPCORE_PRIV); qdev_prop_set_uint32(dev, "num-cpu", 1); @@ -XXX,XX +XXX,XX @@ static void zynq_init(MachineState *machine) dev = qdev_new(TYPE_CADENCE_UART); busdev = SYS_BUS_DEVICE(dev); qdev_prop_set_chr(dev, "chardev", serial_hd(0)); + qdev_connect_clock_in(dev, "refclk", + qdev_get_clock_out(slcr, "uart0_ref_clk")); sysbus_realize_and_unref(busdev, &error_fatal); sysbus_mmio_map(busdev, 0, 0xE0000000); sysbus_connect_irq(busdev, 0, pic[59 - IRQ_OFFSET]); - qdev_connect_clock_in(dev, "refclk", - qdev_get_clock_out(slcr, "uart0_ref_clk")); dev = qdev_new(TYPE_CADENCE_UART); busdev = SYS_BUS_DEVICE(dev); qdev_prop_set_chr(dev, "chardev", serial_hd(1)); + qdev_connect_clock_in(dev, "refclk", + qdev_get_clock_out(slcr, "uart1_ref_clk")); sysbus_realize_and_unref(busdev, &error_fatal); sysbus_mmio_map(busdev, 0, 0xE0001000); sysbus_connect_irq(busdev, 0, pic[82 - IRQ_OFFSET]); - qdev_connect_clock_in(dev, "refclk", - qdev_get_clock_out(slcr, "uart1_ref_clk")); sysbus_create_varargs("cadence_ttc", 0xF8001000, pic[42-IRQ_OFFSET], pic[43-IRQ_OFFSET], pic[44-IRQ_OFFSET], NULL); -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> We want to assert the device is not realized. To avoid overloading this header including "hw/qdev-core.h", uninline the function first. Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Message-id: 20200803105647.22223-4-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/qdev-clock.h | 6 +----- hw/core/qdev-clock.c | 5 +++++ 2 files changed, 6 insertions(+), 5 deletions(-) diff --git a/include/hw/qdev-clock.h b/include/hw/qdev-clock.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/qdev-clock.h +++ b/include/hw/qdev-clock.h @@ -XXX,XX +XXX,XX @@ Clock *qdev_get_clock_out(DeviceState *dev, const char *name); * Set the source clock of input clock @name of device @dev to @source. * @source period update will be propagated to @name clock. */ -static inline void qdev_connect_clock_in(DeviceState *dev, const char *name, - Clock *source) -{ - clock_set_source(qdev_get_clock_in(dev, name), source); -} +void qdev_connect_clock_in(DeviceState *dev, const char *name, Clock *source); /** * qdev_alias_clock: diff --git a/hw/core/qdev-clock.c b/hw/core/qdev-clock.c index XXXXXXX..XXXXXXX 100644 --- a/hw/core/qdev-clock.c +++ b/hw/core/qdev-clock.c @@ -XXX,XX +XXX,XX @@ Clock *qdev_alias_clock(DeviceState *dev, const char *name, return ncl->clock; } + +void qdev_connect_clock_in(DeviceState *dev, const char *name, Clock *source) +{ + clock_set_source(qdev_get_clock_in(dev, name), source); +} -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> Clock canonical name is set in device_set_realized (see the block added to hw/core/qdev.c in commit 0e6934f264). If we connect a clock after the device is realized, this code is not executed. This is currently not a problem as this name is only used for trace events, however this disrupt tracing. Add a comment to document qdev_connect_clock_in() must be called before the device is realized, and assert this condition. Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Message-id: 20200803105647.22223-5-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/qdev-clock.h | 2 ++ hw/core/qdev-clock.c | 1 + 2 files changed, 3 insertions(+) diff --git a/include/hw/qdev-clock.h b/include/hw/qdev-clock.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/qdev-clock.h +++ b/include/hw/qdev-clock.h @@ -XXX,XX +XXX,XX @@ Clock *qdev_get_clock_out(DeviceState *dev, const char *name); * * Set the source clock of input clock @name of device @dev to @source. * @source period update will be propagated to @name clock. + * + * Must be called before @dev is realized. */ void qdev_connect_clock_in(DeviceState *dev, const char *name, Clock *source); diff --git a/hw/core/qdev-clock.c b/hw/core/qdev-clock.c index XXXXXXX..XXXXXXX 100644 --- a/hw/core/qdev-clock.c +++ b/hw/core/qdev-clock.c @@ -XXX,XX +XXX,XX @@ Clock *qdev_alias_clock(DeviceState *dev, const char *name, void qdev_connect_clock_in(DeviceState *dev, const char *name, Clock *source) { + assert(!dev->realized); clock_set_source(qdev_get_clock_in(dev, name), source); } -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> To better align the read/write accesses, display the value after the offset (read accesses only display the offset). Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200812190206.31595-2-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/misc/unimp.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/hw/misc/unimp.c b/hw/misc/unimp.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/unimp.c +++ b/hw/misc/unimp.c @@ -XXX,XX +XXX,XX @@ static uint64_t unimp_read(void *opaque, hwaddr offset, unsigned size) { UnimplementedDeviceState *s = UNIMPLEMENTED_DEVICE(opaque); - qemu_log_mask(LOG_UNIMP, "%s: unimplemented device read " + qemu_log_mask(LOG_UNIMP, "%s: unimplemented device read " "(size %d, offset 0x%" HWADDR_PRIx ")\n", s->name, size, offset); return 0; @@ -XXX,XX +XXX,XX @@ static void unimp_write(void *opaque, hwaddr offset, UnimplementedDeviceState *s = UNIMPLEMENTED_DEVICE(opaque); qemu_log_mask(LOG_UNIMP, "%s: unimplemented device write " - "(size %d, value 0x%" PRIx64 - ", offset 0x%" HWADDR_PRIx ")\n", - s->name, size, value, offset); + "(size %d, offset 0x%" HWADDR_PRIx + ", value 0x%" PRIx64 ")\n", + s->name, size, offset, value); } static const MemoryRegionOps unimp_ops = { -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> To quickly notice the access size, display the value with the width of the access (i.e. 16-bit access is displayed 0x0000, while 8-bit access 0x00). Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200812190206.31595-3-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- hw/misc/unimp.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/hw/misc/unimp.c b/hw/misc/unimp.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/unimp.c +++ b/hw/misc/unimp.c @@ -XXX,XX +XXX,XX @@ static void unimp_write(void *opaque, hwaddr offset, qemu_log_mask(LOG_UNIMP, "%s: unimplemented device write " "(size %d, offset 0x%" HWADDR_PRIx - ", value 0x%" PRIx64 ")\n", - s->name, size, offset, value); + ", value 0x%0*" PRIx64 ")\n", + s->name, size, offset, size << 1, value); } static const MemoryRegionOps unimp_ops = { -- 2.20.1 From: Philippe Mathieu-Daudé <f4bug@amsat.org> To have a better idea of how big is the region where the offset belongs, display the value with the width of the region size (i.e. a region of 0x1000 bytes uses 0x000 format). Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-id: 20200812190206.31595-4-f4bug@amsat.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/misc/unimp.h | 1 + hw/misc/unimp.c | 10 ++++++---- 2 files changed, 7 insertions(+), 4 deletions(-) diff --git a/include/hw/misc/unimp.h b/include/hw/misc/unimp.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/misc/unimp.h +++ b/include/hw/misc/unimp.h @@ -XXX,XX +XXX,XX @@ typedef struct { SysBusDevice parent_obj; MemoryRegion iomem; + unsigned offset_fmt_width; char *name; uint64_t size; } UnimplementedDeviceState; diff --git a/hw/misc/unimp.c b/hw/misc/unimp.c index XXXXXXX..XXXXXXX 100644 --- a/hw/misc/unimp.c +++ b/hw/misc/unimp.c @@ -XXX,XX +XXX,XX @@ static uint64_t unimp_read(void *opaque, hwaddr offset, unsigned size) UnimplementedDeviceState *s = UNIMPLEMENTED_DEVICE(opaque); qemu_log_mask(LOG_UNIMP, "%s: unimplemented device read " - "(size %d, offset 0x%" HWADDR_PRIx ")\n", - s->name, size, offset); + "(size %d, offset 0x%0*" HWADDR_PRIx ")\n", + s->name, size, s->offset_fmt_width, offset); return 0; } @@ -XXX,XX +XXX,XX @@ static void unimp_write(void *opaque, hwaddr offset, UnimplementedDeviceState *s = UNIMPLEMENTED_DEVICE(opaque); qemu_log_mask(LOG_UNIMP, "%s: unimplemented device write " - "(size %d, offset 0x%" HWADDR_PRIx + "(size %d, offset 0x%0*" HWADDR_PRIx ", value 0x%0*" PRIx64 ")\n", - s->name, size, offset, size << 1, value); + s->name, size, s->offset_fmt_width, offset, size << 1, value); } static const MemoryRegionOps unimp_ops = { @@ -XXX,XX +XXX,XX @@ static void unimp_realize(DeviceState *dev, Error **errp) return; } + s->offset_fmt_width = DIV_ROUND_UP(64 - clz64(s->size - 1), 4); + memory_region_init_io(&s->iomem, OBJECT(s), &unimp_ops, s, s->name, s->size); sysbus_init_mmio(SYS_BUS_DEVICE(s), &s->iomem); -- 2.20.1 From: Eduardo Habkost <ehabkost@redhat.com> TYPE_ARM_SSE is a TYPE_SYS_BUS_DEVICE subclass, but ARMSSEClass::parent_class is declared as DeviceClass. It never caused any problems by pure luck: We were not setting class_size for TYPE_ARM_SSE, so class_size of TYPE_SYS_BUS_DEVICE was being used (sizeof(SysBusDeviceClass)). This made the system allocate enough memory for TYPE_ARM_SSE devices even though ARMSSEClass was too small for a sysbus device. Additionally, the ARMSSEClass::info field ended up at the same offset as SysBusDeviceClass::explicit_ofw_unit_address. This would make sysbus_get_fw_dev_path() crash for the device. Luckily, sysbus_get_fw_dev_path() never gets called for TYPE_ARM_SSE devices, because qdev_get_fw_dev_path() is only used by the boot device code, and TYPE_ARM_SSE devices don't appear at the fw_boot_order list. Signed-off-by: Eduardo Habkost <ehabkost@redhat.com> Message-id: 20200826181006.4097163-1-ehabkost@redhat.com Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/hw/arm/armsse.h | 2 +- hw/arm/armsse.c | 1 + 2 files changed, 2 insertions(+), 1 deletion(-) diff --git a/include/hw/arm/armsse.h b/include/hw/arm/armsse.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/arm/armsse.h +++ b/include/hw/arm/armsse.h @@ -XXX,XX +XXX,XX @@ typedef struct ARMSSE { typedef struct ARMSSEInfo ARMSSEInfo; typedef struct ARMSSEClass { - DeviceClass parent_class; + SysBusDeviceClass parent_class; const ARMSSEInfo *info; } ARMSSEClass; diff --git a/hw/arm/armsse.c b/hw/arm/armsse.c index XXXXXXX..XXXXXXX 100644 --- a/hw/arm/armsse.c +++ b/hw/arm/armsse.c @@ -XXX,XX +XXX,XX @@ static const TypeInfo armsse_info = { .name = TYPE_ARMSSE, .parent = TYPE_SYS_BUS_DEVICE, .instance_size = sizeof(ARMSSE), + .class_size = sizeof(ARMSSEClass), .instance_init = armsse_init, .abstract = true, .interfaces = (InterfaceInfo[]) { -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Add left-shift to match the existing right-shift. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-2-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- include/qemu/int128.h | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/include/qemu/int128.h b/include/qemu/int128.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/int128.h +++ b/include/qemu/int128.h @@ -XXX,XX +XXX,XX @@ static inline Int128 int128_rshift(Int128 a, int n) return a >> n; } +static inline Int128 int128_lshift(Int128 a, int n) +{ + return a << n; +} + static inline Int128 int128_add(Int128 a, Int128 b) { return a + b; @@ -XXX,XX +XXX,XX @@ static inline Int128 int128_rshift(Int128 a, int n) } } +static inline Int128 int128_lshift(Int128 a, int n) +{ + uint64_t l = a.lo << (n & 63); + if (n >= 64) { + return int128_make128(0, l); + } else if (n > 0) { + return int128_make128(l, (a.hi << n) | (a.lo >> (64 - n))); + } + return a; +} + static inline Int128 int128_add(Int128 a, Int128 b) { uint64_t lo = a.lo + b.lo; -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Model the new function on gen_gvec_fn2 in translate-a64.c, but indicating which kind of register and in which order. Since there is only one user of do_vector2_z, fold it into do_mov_z. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-3-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static int pred_gvec_reg_size(DisasContext *s) } /* Invoke a vector expander on two Zregs. */ -static bool do_vector2_z(DisasContext *s, GVecGen2Fn *gvec_fn, - int esz, int rd, int rn) + +static void gen_gvec_fn_zz(DisasContext *s, GVecGen2Fn *gvec_fn, + int esz, int rd, int rn) { - if (sve_access_check(s)) { - unsigned vsz = vec_full_reg_size(s); - gvec_fn(esz, vec_full_reg_offset(s, rd), - vec_full_reg_offset(s, rn), vsz, vsz); - } - return true; + unsigned vsz = vec_full_reg_size(s); + gvec_fn(esz, vec_full_reg_offset(s, rd), + vec_full_reg_offset(s, rn), vsz, vsz); } /* Invoke a vector expander on three Zregs. */ @@ -XXX,XX +XXX,XX @@ static bool do_vector3_z(DisasContext *s, GVecGen3Fn *gvec_fn, /* Invoke a vector move on two Zregs. */ static bool do_mov_z(DisasContext *s, int rd, int rn) { - return do_vector2_z(s, tcg_gen_gvec_mov, 0, rd, rn); + if (sve_access_check(s)) { + gen_gvec_fn_zz(s, tcg_gen_gvec_mov, MO_8, rd, rn); + } + return true; } /* Initialize a Zreg with replications of a 64-bit immediate. */ -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Model gen_gvec_fn_zzz on gen_gvec_fn3 in translate-a64.c, but indicating which kind of register and in which order. Model do_zzz_fn on the other do_foo functions that take an argument set and verify sve enabled. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-4-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 43 +++++++++++++++++++++----------------- 1 file changed, 24 insertions(+), 19 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static void gen_gvec_fn_zz(DisasContext *s, GVecGen2Fn *gvec_fn, } /* Invoke a vector expander on three Zregs. */ -static bool do_vector3_z(DisasContext *s, GVecGen3Fn *gvec_fn, - int esz, int rd, int rn, int rm) +static void gen_gvec_fn_zzz(DisasContext *s, GVecGen3Fn *gvec_fn, + int esz, int rd, int rn, int rm) { - if (sve_access_check(s)) { - unsigned vsz = vec_full_reg_size(s); - gvec_fn(esz, vec_full_reg_offset(s, rd), - vec_full_reg_offset(s, rn), - vec_full_reg_offset(s, rm), vsz, vsz); - } - return true; + unsigned vsz = vec_full_reg_size(s); + gvec_fn(esz, vec_full_reg_offset(s, rd), + vec_full_reg_offset(s, rn), + vec_full_reg_offset(s, rm), vsz, vsz); } /* Invoke a vector move on two Zregs. */ @@ -XXX,XX +XXX,XX @@ const uint64_t pred_esz_masks[4] = { *** SVE Logical - Unpredicated Group */ +static bool do_zzz_fn(DisasContext *s, arg_rrr_esz *a, GVecGen3Fn *gvec_fn) +{ + if (sve_access_check(s)) { + gen_gvec_fn_zzz(s, gvec_fn, a->esz, a->rd, a->rn, a->rm); + } + return true; +} + static bool trans_AND_zzz(DisasContext *s, arg_rrr_esz *a) { - return do_vector3_z(s, tcg_gen_gvec_and, 0, a->rd, a->rn, a->rm); + return do_zzz_fn(s, a, tcg_gen_gvec_and); } static bool trans_ORR_zzz(DisasContext *s, arg_rrr_esz *a) { - return do_vector3_z(s, tcg_gen_gvec_or, 0, a->rd, a->rn, a->rm); + return do_zzz_fn(s, a, tcg_gen_gvec_or); } static bool trans_EOR_zzz(DisasContext *s, arg_rrr_esz *a) { - return do_vector3_z(s, tcg_gen_gvec_xor, 0, a->rd, a->rn, a->rm); + return do_zzz_fn(s, a, tcg_gen_gvec_xor); } static bool trans_BIC_zzz(DisasContext *s, arg_rrr_esz *a) { - return do_vector3_z(s, tcg_gen_gvec_andc, 0, a->rd, a->rn, a->rm); + return do_zzz_fn(s, a, tcg_gen_gvec_andc); } /* @@ -XXX,XX +XXX,XX @@ static bool trans_BIC_zzz(DisasContext *s, arg_rrr_esz *a) static bool trans_ADD_zzz(DisasContext *s, arg_rrr_esz *a) { - return do_vector3_z(s, tcg_gen_gvec_add, a->esz, a->rd, a->rn, a->rm); + return do_zzz_fn(s, a, tcg_gen_gvec_add); } static bool trans_SUB_zzz(DisasContext *s, arg_rrr_esz *a) { - return do_vector3_z(s, tcg_gen_gvec_sub, a->esz, a->rd, a->rn, a->rm); + return do_zzz_fn(s, a, tcg_gen_gvec_sub); } static bool trans_SQADD_zzz(DisasContext *s, arg_rrr_esz *a) { - return do_vector3_z(s, tcg_gen_gvec_ssadd, a->esz, a->rd, a->rn, a->rm); + return do_zzz_fn(s, a, tcg_gen_gvec_ssadd); } static bool trans_SQSUB_zzz(DisasContext *s, arg_rrr_esz *a) { - return do_vector3_z(s, tcg_gen_gvec_sssub, a->esz, a->rd, a->rn, a->rm); + return do_zzz_fn(s, a, tcg_gen_gvec_sssub); } static bool trans_UQADD_zzz(DisasContext *s, arg_rrr_esz *a) { - return do_vector3_z(s, tcg_gen_gvec_usadd, a->esz, a->rd, a->rn, a->rm); + return do_zzz_fn(s, a, tcg_gen_gvec_usadd); } static bool trans_UQSUB_zzz(DisasContext *s, arg_rrr_esz *a) { - return do_vector3_z(s, tcg_gen_gvec_ussub, a->esz, a->rd, a->rn, a->rm); + return do_zzz_fn(s, a, tcg_gen_gvec_ussub); } /* -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> We want to ensure that access is checked by the time we ask for a specific fp/vector register. We want to ensure that we do not emit two lots of code to raise an exception. But sometimes it's difficult to cleanly organize the code such that we never pass through sve_check_access exactly once. Allow multiple calls so long as the result is true, that is, no exception to be raised. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-5-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate.h | 1 + target/arm/translate-a64.c | 27 ++++++++++++++++----------- 2 files changed, 17 insertions(+), 11 deletions(-) diff --git a/target/arm/translate.h b/target/arm/translate.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate.h +++ b/target/arm/translate.h @@ -XXX,XX +XXX,XX @@ typedef struct DisasContext { * that it is set at the point where we actually touch the FP regs. */ bool fp_access_checked; + bool sve_access_checked; /* ARMv8 single-step state (this is distinct from the QEMU gdbstub * single-step support). */ diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void do_vec_ld(DisasContext *s, int destidx, int element, * unallocated-encoding checks (otherwise the syndrome information * for the resulting exception will be incorrect). */ -static inline bool fp_access_check(DisasContext *s) +static bool fp_access_check(DisasContext *s) { - assert(!s->fp_access_checked); - s->fp_access_checked = true; + if (s->fp_excp_el) { + assert(!s->fp_access_checked); + s->fp_access_checked = true; - if (!s->fp_excp_el) { - return true; + gen_exception_insn(s, s->pc_curr, EXCP_UDEF, + syn_fp_access_trap(1, 0xe, false), s->fp_excp_el); + return false; } - - gen_exception_insn(s, s->pc_curr, EXCP_UDEF, - syn_fp_access_trap(1, 0xe, false), s->fp_excp_el); - return false; + s->fp_access_checked = true; + return true; } /* Check that SVE access is enabled. If it is, return true. @@ -XXX,XX +XXX,XX @@ static inline bool fp_access_check(DisasContext *s) bool sve_access_check(DisasContext *s) { if (s->sve_excp_el) { - gen_exception_insn(s, s->pc_curr, EXCP_UDEF, syn_sve_access_trap(), - s->sve_excp_el); + assert(!s->sve_access_checked); + s->sve_access_checked = true; + + gen_exception_insn(s, s->pc_curr, EXCP_UDEF, + syn_sve_access_trap(), s->sve_excp_el); return false; } + s->sve_access_checked = true; return fp_access_check(s); } @@ -XXX,XX +XXX,XX @@ static void disas_a64_insn(CPUARMState *env, DisasContext *s) s->base.pc_next += 4; s->fp_access_checked = false; + s->sve_access_checked = false; if (dc_isar_feature(aa64_bti, s)) { if (s->base.num_insns == 1) { -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> This is the only user of the function. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-6-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 19 ++++++------------- 1 file changed, 6 insertions(+), 13 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static void do_dupi_z(DisasContext *s, int rd, uint64_t word) tcg_gen_gvec_dup_imm(MO_64, vec_full_reg_offset(s, rd), vsz, vsz, word); } -/* Invoke a vector expander on two Pregs. */ -static bool do_vector2_p(DisasContext *s, GVecGen2Fn *gvec_fn, - int esz, int rd, int rn) -{ - if (sve_access_check(s)) { - unsigned psz = pred_gvec_reg_size(s); - gvec_fn(esz, pred_full_reg_offset(s, rd), - pred_full_reg_offset(s, rn), psz, psz); - } - return true; -} - /* Invoke a vector expander on three Pregs. */ static bool do_vector3_p(DisasContext *s, GVecGen3Fn *gvec_fn, int esz, int rd, int rn, int rm) @@ -XXX,XX +XXX,XX @@ static bool do_vecop4_p(DisasContext *s, const GVecGen4 *gvec_op, /* Invoke a vector move on two Pregs. */ static bool do_mov_p(DisasContext *s, int rd, int rn) { - return do_vector2_p(s, tcg_gen_gvec_mov, 0, rd, rn); + if (sve_access_check(s)) { + unsigned psz = pred_gvec_reg_size(s); + tcg_gen_gvec_mov(MO_8, pred_full_reg_offset(s, rd), + pred_full_reg_offset(s, rn), psz, psz); + } + return true; } /* Set the cpu flags as per a return from an SVE helper. */ -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Move the check for !S into do_pppp_flags, which allows to merge in do_vecop4_p. Split out gen_gvec_fn_ppp without sve_access_check, to mirror gen_gvec_fn_zzz. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-7-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 111 ++++++++++++++----------------------- 1 file changed, 43 insertions(+), 68 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static void do_dupi_z(DisasContext *s, int rd, uint64_t word) } /* Invoke a vector expander on three Pregs. */ -static bool do_vector3_p(DisasContext *s, GVecGen3Fn *gvec_fn, - int esz, int rd, int rn, int rm) +static void gen_gvec_fn_ppp(DisasContext *s, GVecGen3Fn *gvec_fn, + int rd, int rn, int rm) { - if (sve_access_check(s)) { - unsigned psz = pred_gvec_reg_size(s); - gvec_fn(esz, pred_full_reg_offset(s, rd), - pred_full_reg_offset(s, rn), - pred_full_reg_offset(s, rm), psz, psz); - } - return true; -} - -/* Invoke a vector operation on four Pregs. */ -static bool do_vecop4_p(DisasContext *s, const GVecGen4 *gvec_op, - int rd, int rn, int rm, int rg) -{ - if (sve_access_check(s)) { - unsigned psz = pred_gvec_reg_size(s); - tcg_gen_gvec_4(pred_full_reg_offset(s, rd), - pred_full_reg_offset(s, rn), - pred_full_reg_offset(s, rm), - pred_full_reg_offset(s, rg), - psz, psz, gvec_op); - } - return true; + unsigned psz = pred_gvec_reg_size(s); + gvec_fn(MO_64, pred_full_reg_offset(s, rd), + pred_full_reg_offset(s, rn), + pred_full_reg_offset(s, rm), psz, psz); } /* Invoke a vector move on two Pregs. */ @@ -XXX,XX +XXX,XX @@ static bool do_pppp_flags(DisasContext *s, arg_rprr_s *a, int mofs = pred_full_reg_offset(s, a->rm); int gofs = pred_full_reg_offset(s, a->pg); + if (!a->s) { + tcg_gen_gvec_4(dofs, nofs, mofs, gofs, psz, psz, gvec_op); + return true; + } + if (psz == 8) { /* Do the operation and the flags generation in temps. */ TCGv_i64 pd = tcg_temp_new_i64(); @@ -XXX,XX +XXX,XX @@ static bool trans_AND_pppp(DisasContext *s, arg_rprr_s *a) .fno = gen_helper_sve_and_pppp, .prefer_i64 = TCG_TARGET_REG_BITS == 64, }; - if (a->s) { - return do_pppp_flags(s, a, &op); - } else if (a->rn == a->rm) { - if (a->pg == a->rn) { - return do_mov_p(s, a->rd, a->rn); - } else { - return do_vector3_p(s, tcg_gen_gvec_and, 0, a->rd, a->rn, a->pg); + + if (!a->s) { + if (!sve_access_check(s)) { + return true; + } + if (a->rn == a->rm) { + if (a->pg == a->rn) { + do_mov_p(s, a->rd, a->rn); + } else { + gen_gvec_fn_ppp(s, tcg_gen_gvec_and, a->rd, a->rn, a->pg); + } + return true; + } else if (a->pg == a->rn || a->pg == a->rm) { + gen_gvec_fn_ppp(s, tcg_gen_gvec_and, a->rd, a->rn, a->rm); + return true; } - } else if (a->pg == a->rn || a->pg == a->rm) { - return do_vector3_p(s, tcg_gen_gvec_and, 0, a->rd, a->rn, a->rm); - } else { - return do_vecop4_p(s, &op, a->rd, a->rn, a->rm, a->pg); } + return do_pppp_flags(s, a, &op); } static void gen_bic_pg_i64(TCGv_i64 pd, TCGv_i64 pn, TCGv_i64 pm, TCGv_i64 pg) @@ -XXX,XX +XXX,XX @@ static bool trans_BIC_pppp(DisasContext *s, arg_rprr_s *a) .fno = gen_helper_sve_bic_pppp, .prefer_i64 = TCG_TARGET_REG_BITS == 64, }; - if (a->s) { - return do_pppp_flags(s, a, &op); - } else if (a->pg == a->rn) { - return do_vector3_p(s, tcg_gen_gvec_andc, 0, a->rd, a->rn, a->rm); - } else { - return do_vecop4_p(s, &op, a->rd, a->rn, a->rm, a->pg); + + if (!a->s && a->pg == a->rn) { + if (sve_access_check(s)) { + gen_gvec_fn_ppp(s, tcg_gen_gvec_andc, a->rd, a->rn, a->rm); + } + return true; } + return do_pppp_flags(s, a, &op); } static void gen_eor_pg_i64(TCGv_i64 pd, TCGv_i64 pn, TCGv_i64 pm, TCGv_i64 pg) @@ -XXX,XX +XXX,XX @@ static bool trans_EOR_pppp(DisasContext *s, arg_rprr_s *a) .fno = gen_helper_sve_eor_pppp, .prefer_i64 = TCG_TARGET_REG_BITS == 64, }; - if (a->s) { - return do_pppp_flags(s, a, &op); - } else { - return do_vecop4_p(s, &op, a->rd, a->rn, a->rm, a->pg); - } + return do_pppp_flags(s, a, &op); } static void gen_sel_pg_i64(TCGv_i64 pd, TCGv_i64 pn, TCGv_i64 pm, TCGv_i64 pg) @@ -XXX,XX +XXX,XX @@ static bool trans_SEL_pppp(DisasContext *s, arg_rprr_s *a) .fno = gen_helper_sve_sel_pppp, .prefer_i64 = TCG_TARGET_REG_BITS == 64, }; + if (a->s) { return false; - } else { - return do_vecop4_p(s, &op, a->rd, a->rn, a->rm, a->pg); } + return do_pppp_flags(s, a, &op); } static void gen_orr_pg_i64(TCGv_i64 pd, TCGv_i64 pn, TCGv_i64 pm, TCGv_i64 pg) @@ -XXX,XX +XXX,XX @@ static bool trans_ORR_pppp(DisasContext *s, arg_rprr_s *a) .fno = gen_helper_sve_orr_pppp, .prefer_i64 = TCG_TARGET_REG_BITS == 64, }; - if (a->s) { - return do_pppp_flags(s, a, &op); - } else if (a->pg == a->rn && a->rn == a->rm) { + + if (!a->s && a->pg == a->rn && a->rn == a->rm) { return do_mov_p(s, a->rd, a->rn); - } else { - return do_vecop4_p(s, &op, a->rd, a->rn, a->rm, a->pg); } + return do_pppp_flags(s, a, &op); } static void gen_orn_pg_i64(TCGv_i64 pd, TCGv_i64 pn, TCGv_i64 pm, TCGv_i64 pg) @@ -XXX,XX +XXX,XX @@ static bool trans_ORN_pppp(DisasContext *s, arg_rprr_s *a) .fno = gen_helper_sve_orn_pppp, .prefer_i64 = TCG_TARGET_REG_BITS == 64, }; - if (a->s) { - return do_pppp_flags(s, a, &op); - } else { - return do_vecop4_p(s, &op, a->rd, a->rn, a->rm, a->pg); - } + return do_pppp_flags(s, a, &op); } static void gen_nor_pg_i64(TCGv_i64 pd, TCGv_i64 pn, TCGv_i64 pm, TCGv_i64 pg) @@ -XXX,XX +XXX,XX @@ static bool trans_NOR_pppp(DisasContext *s, arg_rprr_s *a) .fno = gen_helper_sve_nor_pppp, .prefer_i64 = TCG_TARGET_REG_BITS == 64, }; - if (a->s) { - return do_pppp_flags(s, a, &op); - } else { - return do_vecop4_p(s, &op, a->rd, a->rn, a->rm, a->pg); - } + return do_pppp_flags(s, a, &op); } static void gen_nand_pg_i64(TCGv_i64 pd, TCGv_i64 pn, TCGv_i64 pm, TCGv_i64 pg) @@ -XXX,XX +XXX,XX @@ static bool trans_NAND_pppp(DisasContext *s, arg_rprr_s *a) .fno = gen_helper_sve_nand_pppp, .prefer_i64 = TCG_TARGET_REG_BITS == 64, }; - if (a->s) { - return do_pppp_flags(s, a, &op); - } else { - return do_vecop4_p(s, &op, a->rd, a->rn, a->rm, a->pg); - } + return do_pppp_flags(s, a, &op); } /* -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> The gvec operation was added after the initial implementation of the SEL instruction and was missed in the conversion. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-8-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 31 ++++++++----------------------- 1 file changed, 8 insertions(+), 23 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_EOR_pppp(DisasContext *s, arg_rprr_s *a) return do_pppp_flags(s, a, &op); } -static void gen_sel_pg_i64(TCGv_i64 pd, TCGv_i64 pn, TCGv_i64 pm, TCGv_i64 pg) -{ - tcg_gen_and_i64(pn, pn, pg); - tcg_gen_andc_i64(pm, pm, pg); - tcg_gen_or_i64(pd, pn, pm); -} - -static void gen_sel_pg_vec(unsigned vece, TCGv_vec pd, TCGv_vec pn, - TCGv_vec pm, TCGv_vec pg) -{ - tcg_gen_and_vec(vece, pn, pn, pg); - tcg_gen_andc_vec(vece, pm, pm, pg); - tcg_gen_or_vec(vece, pd, pn, pm); -} - static bool trans_SEL_pppp(DisasContext *s, arg_rprr_s *a) { - static const GVecGen4 op = { - .fni8 = gen_sel_pg_i64, - .fniv = gen_sel_pg_vec, - .fno = gen_helper_sve_sel_pppp, - .prefer_i64 = TCG_TARGET_REG_BITS == 64, - }; - if (a->s) { return false; } - return do_pppp_flags(s, a, &op); + if (sve_access_check(s)) { + unsigned psz = pred_gvec_reg_size(s); + tcg_gen_gvec_bitsel(MO_8, pred_full_reg_offset(s, a->rd), + pred_full_reg_offset(s, a->pg), + pred_full_reg_offset(s, a->rn), + pred_full_reg_offset(s, a->rm), psz, psz); + } + return true; } static void gen_orr_pg_i64(TCGv_i64 pd, TCGv_i64 pn, TCGv_i64 pm, TCGv_i64 pg) -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Model after gen_gvec_fn_zzz et al. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-9-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 35 ++++++++++++++++------------------- 1 file changed, 16 insertions(+), 19 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static int pred_gvec_reg_size(DisasContext *s) return size_for_gvec(pred_full_reg_size(s)); } -/* Invoke a vector expander on two Zregs. */ +/* Invoke an out-of-line helper on 3 Zregs and a predicate. */ +static void gen_gvec_ool_zzzp(DisasContext *s, gen_helper_gvec_4 *fn, + int rd, int rn, int rm, int pg, int data) +{ + unsigned vsz = vec_full_reg_size(s); + tcg_gen_gvec_4_ool(vec_full_reg_offset(s, rd), + vec_full_reg_offset(s, rn), + vec_full_reg_offset(s, rm), + pred_full_reg_offset(s, pg), + vsz, vsz, data, fn); +} +/* Invoke a vector expander on two Zregs. */ static void gen_gvec_fn_zz(DisasContext *s, GVecGen2Fn *gvec_fn, int esz, int rd, int rn) { @@ -XXX,XX +XXX,XX @@ static bool trans_UQSUB_zzz(DisasContext *s, arg_rrr_esz *a) static bool do_zpzz_ool(DisasContext *s, arg_rprr_esz *a, gen_helper_gvec_4 *fn) { - unsigned vsz = vec_full_reg_size(s); if (fn == NULL) { return false; } if (sve_access_check(s)) { - tcg_gen_gvec_4_ool(vec_full_reg_offset(s, a->rd), - vec_full_reg_offset(s, a->rn), - vec_full_reg_offset(s, a->rm), - pred_full_reg_offset(s, a->pg), - vsz, vsz, 0, fn); + gen_gvec_ool_zzzp(s, fn, a->rd, a->rn, a->rm, a->pg, 0); } return true; } @@ -XXX,XX +XXX,XX @@ static void do_sel_z(DisasContext *s, int rd, int rn, int rm, int pg, int esz) gen_helper_sve_sel_zpzz_b, gen_helper_sve_sel_zpzz_h, gen_helper_sve_sel_zpzz_s, gen_helper_sve_sel_zpzz_d }; - unsigned vsz = vec_full_reg_size(s); - tcg_gen_gvec_4_ool(vec_full_reg_offset(s, rd), - vec_full_reg_offset(s, rn), - vec_full_reg_offset(s, rm), - pred_full_reg_offset(s, pg), - vsz, vsz, 0, fns[esz]); + gen_gvec_ool_zzzp(s, fns[esz], rd, rn, rm, pg, 0); } #define DO_ZPZZ(NAME, name) \ @@ -XXX,XX +XXX,XX @@ static bool trans_RBIT(DisasContext *s, arg_rpr_esz *a) static bool trans_SPLICE(DisasContext *s, arg_rprr_esz *a) { if (sve_access_check(s)) { - unsigned vsz = vec_full_reg_size(s); - tcg_gen_gvec_4_ool(vec_full_reg_offset(s, a->rd), - vec_full_reg_offset(s, a->rn), - vec_full_reg_offset(s, a->rm), - pred_full_reg_offset(s, a->pg), - vsz, vsz, a->esz, gen_helper_sve_splice); + gen_gvec_ool_zzzp(s, gen_helper_sve_splice, + a->rd, a->rn, a->rm, a->pg, 0); } return true; } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> The existing clr functions have only one vector argument, and so can only clear in place. The existing movz functions have two vector arguments, and so can clear while moving. Merge them, with a flag that controls the sense of active vs inactive elements being cleared. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-10-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper-sve.h | 5 --- target/arm/sve_helper.c | 70 ++++++++------------------------------ target/arm/translate-sve.c | 53 +++++++++++------------------ 3 files changed, 34 insertions(+), 94 deletions(-) diff --git a/target/arm/helper-sve.h b/target/arm/helper-sve.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper-sve.h +++ b/target/arm/helper-sve.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_3(sve_uminv_h, TCG_CALL_NO_RWG, i64, ptr, ptr, i32) DEF_HELPER_FLAGS_3(sve_uminv_s, TCG_CALL_NO_RWG, i64, ptr, ptr, i32) DEF_HELPER_FLAGS_3(sve_uminv_d, TCG_CALL_NO_RWG, i64, ptr, ptr, i32) -DEF_HELPER_FLAGS_3(sve_clr_b, TCG_CALL_NO_RWG, void, ptr, ptr, i32) -DEF_HELPER_FLAGS_3(sve_clr_h, TCG_CALL_NO_RWG, void, ptr, ptr, i32) -DEF_HELPER_FLAGS_3(sve_clr_s, TCG_CALL_NO_RWG, void, ptr, ptr, i32) -DEF_HELPER_FLAGS_3(sve_clr_d, TCG_CALL_NO_RWG, void, ptr, ptr, i32) - DEF_HELPER_FLAGS_4(sve_movz_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sve_movz_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sve_movz_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(sve_pnext)(void *vd, void *vg, uint32_t pred_desc) return flags; } -/* Store zero into every active element of Zd. We will use this for two - * and three-operand predicated instructions for which logic dictates a - * zero result. In particular, logical shift by element size, which is - * otherwise undefined on the host. - * - * For element sizes smaller than uint64_t, we use tables to expand - * the N bits of the controlling predicate to a byte mask, and clear - * those bytes. +/* + * Copy Zn into Zd, and store zero into inactive elements. + * If inv, store zeros into the active elements. */ -void HELPER(sve_clr_b)(void *vd, void *vg, uint32_t desc) -{ - intptr_t i, opr_sz = simd_oprsz(desc) / 8; - uint64_t *d = vd; - uint8_t *pg = vg; - for (i = 0; i < opr_sz; i += 1) { - d[i] &= ~expand_pred_b(pg[H1(i)]); - } -} - -void HELPER(sve_clr_h)(void *vd, void *vg, uint32_t desc) -{ - intptr_t i, opr_sz = simd_oprsz(desc) / 8; - uint64_t *d = vd; - uint8_t *pg = vg; - for (i = 0; i < opr_sz; i += 1) { - d[i] &= ~expand_pred_h(pg[H1(i)]); - } -} - -void HELPER(sve_clr_s)(void *vd, void *vg, uint32_t desc) -{ - intptr_t i, opr_sz = simd_oprsz(desc) / 8; - uint64_t *d = vd; - uint8_t *pg = vg; - for (i = 0; i < opr_sz; i += 1) { - d[i] &= ~expand_pred_s(pg[H1(i)]); - } -} - -void HELPER(sve_clr_d)(void *vd, void *vg, uint32_t desc) -{ - intptr_t i, opr_sz = simd_oprsz(desc) / 8; - uint64_t *d = vd; - uint8_t *pg = vg; - for (i = 0; i < opr_sz; i += 1) { - if (pg[H1(i)] & 1) { - d[i] = 0; - } - } -} - -/* Copy Zn into Zd, and store zero into inactive elements. */ void HELPER(sve_movz_b)(void *vd, void *vn, void *vg, uint32_t desc) { intptr_t i, opr_sz = simd_oprsz(desc) / 8; + uint64_t inv = -(uint64_t)(simd_data(desc) & 1); uint64_t *d = vd, *n = vn; uint8_t *pg = vg; + for (i = 0; i < opr_sz; i += 1) { - d[i] = n[i] & expand_pred_b(pg[H1(i)]); + d[i] = n[i] & (expand_pred_b(pg[H1(i)]) ^ inv); } } void HELPER(sve_movz_h)(void *vd, void *vn, void *vg, uint32_t desc) { intptr_t i, opr_sz = simd_oprsz(desc) / 8; + uint64_t inv = -(uint64_t)(simd_data(desc) & 1); uint64_t *d = vd, *n = vn; uint8_t *pg = vg; + for (i = 0; i < opr_sz; i += 1) { - d[i] = n[i] & expand_pred_h(pg[H1(i)]); + d[i] = n[i] & (expand_pred_h(pg[H1(i)]) ^ inv); } } void HELPER(sve_movz_s)(void *vd, void *vn, void *vg, uint32_t desc) { intptr_t i, opr_sz = simd_oprsz(desc) / 8; + uint64_t inv = -(uint64_t)(simd_data(desc) & 1); uint64_t *d = vd, *n = vn; uint8_t *pg = vg; + for (i = 0; i < opr_sz; i += 1) { - d[i] = n[i] & expand_pred_s(pg[H1(i)]); + d[i] = n[i] & (expand_pred_s(pg[H1(i)]) ^ inv); } } @@ -XXX,XX +XXX,XX @@ void HELPER(sve_movz_d)(void *vd, void *vn, void *vg, uint32_t desc) intptr_t i, opr_sz = simd_oprsz(desc) / 8; uint64_t *d = vd, *n = vn; uint8_t *pg = vg; + uint8_t inv = simd_data(desc); + for (i = 0; i < opr_sz; i += 1) { - d[i] = n[i] & -(uint64_t)(pg[H1(i)] & 1); + d[i] = n[i] & -(uint64_t)((pg[H1(i)] ^ inv) & 1); } } diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static bool trans_SADDV(DisasContext *s, arg_rpr_esz *a) *** SVE Shift by Immediate - Predicated Group */ -/* Store zero into every active element of Zd. We will use this for two - * and three-operand predicated instructions for which logic dictates a - * zero result. +/* + * Copy Zn into Zd, storing zeros into inactive elements. + * If invert, store zeros into the active elements. */ -static bool do_clr_zp(DisasContext *s, int rd, int pg, int esz) -{ - static gen_helper_gvec_2 * const fns[4] = { - gen_helper_sve_clr_b, gen_helper_sve_clr_h, - gen_helper_sve_clr_s, gen_helper_sve_clr_d, - }; - if (sve_access_check(s)) { - unsigned vsz = vec_full_reg_size(s); - tcg_gen_gvec_2_ool(vec_full_reg_offset(s, rd), - pred_full_reg_offset(s, pg), - vsz, vsz, 0, fns[esz]); - } - return true; -} - -/* Copy Zn into Zd, storing zeros into inactive elements. */ -static void do_movz_zpz(DisasContext *s, int rd, int rn, int pg, int esz) +static bool do_movz_zpz(DisasContext *s, int rd, int rn, int pg, + int esz, bool invert) { static gen_helper_gvec_3 * const fns[4] = { gen_helper_sve_movz_b, gen_helper_sve_movz_h, gen_helper_sve_movz_s, gen_helper_sve_movz_d, }; - unsigned vsz = vec_full_reg_size(s); - tcg_gen_gvec_3_ool(vec_full_reg_offset(s, rd), - vec_full_reg_offset(s, rn), - pred_full_reg_offset(s, pg), - vsz, vsz, 0, fns[esz]); + + if (sve_access_check(s)) { + unsigned vsz = vec_full_reg_size(s); + tcg_gen_gvec_3_ool(vec_full_reg_offset(s, rd), + vec_full_reg_offset(s, rn), + pred_full_reg_offset(s, pg), + vsz, vsz, invert, fns[esz]); + } + return true; } static bool do_zpzi_ool(DisasContext *s, arg_rpri_esz *a, @@ -XXX,XX +XXX,XX @@ static bool trans_LSR_zpzi(DisasContext *s, arg_rpri_esz *a) /* Shift by element size is architecturally valid. For logical shifts, it is a zeroing operation. */ if (a->imm >= (8 << a->esz)) { - return do_clr_zp(s, a->rd, a->pg, a->esz); + return do_movz_zpz(s, a->rd, a->rd, a->pg, a->esz, true); } else { return do_zpzi_ool(s, a, fns[a->esz]); } @@ -XXX,XX +XXX,XX @@ static bool trans_LSL_zpzi(DisasContext *s, arg_rpri_esz *a) /* Shift by element size is architecturally valid. For logical shifts, it is a zeroing operation. */ if (a->imm >= (8 << a->esz)) { - return do_clr_zp(s, a->rd, a->pg, a->esz); + return do_movz_zpz(s, a->rd, a->rd, a->pg, a->esz, true); } else { return do_zpzi_ool(s, a, fns[a->esz]); } @@ -XXX,XX +XXX,XX @@ static bool trans_ASRD(DisasContext *s, arg_rpri_esz *a) /* Shift by element size is architecturally valid. For arithmetic right shift for division, it is a zeroing operation. */ if (a->imm >= (8 << a->esz)) { - return do_clr_zp(s, a->rd, a->pg, a->esz); + return do_movz_zpz(s, a->rd, a->rd, a->pg, a->esz, true); } else { return do_zpzi_ool(s, a, fns[a->esz]); } @@ -XXX,XX +XXX,XX @@ static bool trans_LD1R_zpri(DisasContext *s, arg_rpri_load *a) /* Zero the inactive elements. */ gen_set_label(over); - do_movz_zpz(s, a->rd, a->rd, a->pg, esz); - return true; + return do_movz_zpz(s, a->rd, a->rd, a->pg, esz, false); } static void do_st_zpa(DisasContext *s, int zt, int pg, TCGv_i64 addr, @@ -XXX,XX +XXX,XX @@ static bool trans_MOVPRFX_m(DisasContext *s, arg_rpr_esz *a) static bool trans_MOVPRFX_z(DisasContext *s, arg_rpr_esz *a) { - if (sve_access_check(s)) { - do_movz_zpz(s, a->rd, a->rn, a->pg, a->esz); - } - return true; + return do_movz_zpz(s, a->rd, a->rn, a->pg, a->esz, false); } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Model after gen_gvec_fn_zzz et al. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-11-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 29 ++++++++++++++--------------- 1 file changed, 14 insertions(+), 15 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static int pred_gvec_reg_size(DisasContext *s) return size_for_gvec(pred_full_reg_size(s)); } +/* Invoke an out-of-line helper on 2 Zregs and a predicate. */ +static void gen_gvec_ool_zzp(DisasContext *s, gen_helper_gvec_3 *fn, + int rd, int rn, int pg, int data) +{ + unsigned vsz = vec_full_reg_size(s); + tcg_gen_gvec_3_ool(vec_full_reg_offset(s, rd), + vec_full_reg_offset(s, rn), + pred_full_reg_offset(s, pg), + vsz, vsz, data, fn); +} + /* Invoke an out-of-line helper on 3 Zregs and a predicate. */ static void gen_gvec_ool_zzzp(DisasContext *s, gen_helper_gvec_4 *fn, int rd, int rn, int rm, int pg, int data) @@ -XXX,XX +XXX,XX @@ static bool do_zpz_ool(DisasContext *s, arg_rpr_esz *a, gen_helper_gvec_3 *fn) return false; } if (sve_access_check(s)) { - unsigned vsz = vec_full_reg_size(s); - tcg_gen_gvec_3_ool(vec_full_reg_offset(s, a->rd), - vec_full_reg_offset(s, a->rn), - pred_full_reg_offset(s, a->pg), - vsz, vsz, 0, fn); + gen_gvec_ool_zzp(s, fn, a->rd, a->rn, a->pg, 0); } return true; } @@ -XXX,XX +XXX,XX @@ static bool do_movz_zpz(DisasContext *s, int rd, int rn, int pg, }; if (sve_access_check(s)) { - unsigned vsz = vec_full_reg_size(s); - tcg_gen_gvec_3_ool(vec_full_reg_offset(s, rd), - vec_full_reg_offset(s, rn), - pred_full_reg_offset(s, pg), - vsz, vsz, invert, fns[esz]); + gen_gvec_ool_zzp(s, fns[esz], rd, rn, pg, invert); } return true; } @@ -XXX,XX +XXX,XX @@ static bool do_zpzi_ool(DisasContext *s, arg_rpri_esz *a, gen_helper_gvec_3 *fn) { if (sve_access_check(s)) { - unsigned vsz = vec_full_reg_size(s); - tcg_gen_gvec_3_ool(vec_full_reg_offset(s, a->rd), - vec_full_reg_offset(s, a->rn), - pred_full_reg_offset(s, a->pg), - vsz, vsz, a->imm, fn); + gen_gvec_ool_zzp(s, fn, a->rd, a->rn, a->pg, a->imm); } return true; } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-12-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 53 +++++++++++++------------------------- 1 file changed, 18 insertions(+), 35 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static int pred_gvec_reg_size(DisasContext *s) return size_for_gvec(pred_full_reg_size(s)); } +/* Invoke an out-of-line helper on 3 Zregs. */ +static void gen_gvec_ool_zzz(DisasContext *s, gen_helper_gvec_3 *fn, + int rd, int rn, int rm, int data) +{ + unsigned vsz = vec_full_reg_size(s); + tcg_gen_gvec_3_ool(vec_full_reg_offset(s, rd), + vec_full_reg_offset(s, rn), + vec_full_reg_offset(s, rm), + vsz, vsz, data, fn); +} + /* Invoke an out-of-line helper on 2 Zregs and a predicate. */ static void gen_gvec_ool_zzp(DisasContext *s, gen_helper_gvec_3 *fn, int rd, int rn, int pg, int data) @@ -XXX,XX +XXX,XX @@ static bool do_zzw_ool(DisasContext *s, arg_rrr_esz *a, gen_helper_gvec_3 *fn) return false; } if (sve_access_check(s)) { - unsigned vsz = vec_full_reg_size(s); - tcg_gen_gvec_3_ool(vec_full_reg_offset(s, a->rd), - vec_full_reg_offset(s, a->rn), - vec_full_reg_offset(s, a->rm), - vsz, vsz, 0, fn); + gen_gvec_ool_zzz(s, fn, a->rd, a->rn, a->rm, 0); } return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_RDVL(DisasContext *s, arg_RDVL *a) static bool do_adr(DisasContext *s, arg_rrri *a, gen_helper_gvec_3 *fn) { if (sve_access_check(s)) { - unsigned vsz = vec_full_reg_size(s); - tcg_gen_gvec_3_ool(vec_full_reg_offset(s, a->rd), - vec_full_reg_offset(s, a->rn), - vec_full_reg_offset(s, a->rm), - vsz, vsz, a->imm, fn); + gen_gvec_ool_zzz(s, fn, a->rd, a->rn, a->rm, a->imm); } return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_FTSSEL(DisasContext *s, arg_rrr_esz *a) return false; } if (sve_access_check(s)) { - unsigned vsz = vec_full_reg_size(s); - tcg_gen_gvec_3_ool(vec_full_reg_offset(s, a->rd), - vec_full_reg_offset(s, a->rn), - vec_full_reg_offset(s, a->rm), - vsz, vsz, 0, fns[a->esz]); + gen_gvec_ool_zzz(s, fns[a->esz], a->rd, a->rn, a->rm, 0); } return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_TBL(DisasContext *s, arg_rrr_esz *a) }; if (sve_access_check(s)) { - unsigned vsz = vec_full_reg_size(s); - tcg_gen_gvec_3_ool(vec_full_reg_offset(s, a->rd), - vec_full_reg_offset(s, a->rn), - vec_full_reg_offset(s, a->rm), - vsz, vsz, 0, fns[a->esz]); + gen_gvec_ool_zzz(s, fns[a->esz], a->rd, a->rn, a->rm, 0); } return true; } @@ -XXX,XX +XXX,XX @@ static bool do_zzz_data_ool(DisasContext *s, arg_rrr_esz *a, int data, gen_helper_gvec_3 *fn) { if (sve_access_check(s)) { - unsigned vsz = vec_full_reg_size(s); - tcg_gen_gvec_3_ool(vec_full_reg_offset(s, a->rd), - vec_full_reg_offset(s, a->rn), - vec_full_reg_offset(s, a->rm), - vsz, vsz, data, fn); + gen_gvec_ool_zzz(s, fn, a->rd, a->rn, a->rm, data); } return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_DOT_zzz(DisasContext *s, arg_DOT_zzz *a) }; if (sve_access_check(s)) { - unsigned vsz = vec_full_reg_size(s); - tcg_gen_gvec_3_ool(vec_full_reg_offset(s, a->rd), - vec_full_reg_offset(s, a->rn), - vec_full_reg_offset(s, a->rm), - vsz, vsz, 0, fns[a->u][a->sz]); + gen_gvec_ool_zzz(s, fns[a->u][a->sz], a->rd, a->rn, a->rm, 0); } return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_DOT_zzx(DisasContext *s, arg_DOT_zzx *a) }; if (sve_access_check(s)) { - unsigned vsz = vec_full_reg_size(s); - tcg_gen_gvec_3_ool(vec_full_reg_offset(s, a->rd), - vec_full_reg_offset(s, a->rn), - vec_full_reg_offset(s, a->rm), - vsz, vsz, a->index, fns[a->u][a->sz]); + gen_gvec_ool_zzz(s, fns[a->u][a->sz], a->rd, a->rn, a->rm, a->index); } return true; } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-13-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/translate-sve.c | 20 ++++++++++++-------- 1 file changed, 12 insertions(+), 8 deletions(-) diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -XXX,XX +XXX,XX @@ static int pred_gvec_reg_size(DisasContext *s) return size_for_gvec(pred_full_reg_size(s)); } +/* Invoke an out-of-line helper on 2 Zregs. */ +static void gen_gvec_ool_zz(DisasContext *s, gen_helper_gvec_2 *fn, + int rd, int rn, int data) +{ + unsigned vsz = vec_full_reg_size(s); + tcg_gen_gvec_2_ool(vec_full_reg_offset(s, rd), + vec_full_reg_offset(s, rn), + vsz, vsz, data, fn); +} + /* Invoke an out-of-line helper on 3 Zregs. */ static void gen_gvec_ool_zzz(DisasContext *s, gen_helper_gvec_3 *fn, int rd, int rn, int rm, int data) @@ -XXX,XX +XXX,XX @@ static bool trans_FEXPA(DisasContext *s, arg_rr_esz *a) return false; } if (sve_access_check(s)) { - unsigned vsz = vec_full_reg_size(s); - tcg_gen_gvec_2_ool(vec_full_reg_offset(s, a->rd), - vec_full_reg_offset(s, a->rn), - vsz, vsz, 0, fns[a->esz]); + gen_gvec_ool_zz(s, fns[a->esz], a->rd, a->rn, 0); } return true; } @@ -XXX,XX +XXX,XX @@ static bool trans_REV_v(DisasContext *s, arg_rr_esz *a) }; if (sve_access_check(s)) { - unsigned vsz = vec_full_reg_size(s); - tcg_gen_gvec_2_ool(vec_full_reg_offset(s, a->rd), - vec_full_reg_offset(s, a->rn), - vsz, vsz, 0, fns[a->esz]); + gen_gvec_ool_zz(s, fns[a->esz], a->rd, a->rn, 0); } return true; } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Rather than require the user to fill in the immediate (shl or shr), create full formats that include the immediate. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-14-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/sve.decode | 35 ++++++++++++++++------------------- 1 file changed, 16 insertions(+), 19 deletions(-) diff --git a/target/arm/sve.decode b/target/arm/sve.decode index XXXXXXX..XXXXXXX 100644 --- a/target/arm/sve.decode +++ b/target/arm/sve.decode @@ -XXX,XX +XXX,XX @@ @rd_rn_i6 ........ ... rn:5 ..... imm:s6 rd:5 &rri # Two register operand, one immediate operand, with predicate, -# element size encoded as TSZHL. User must fill in imm. -@rdn_pg_tszimm ........ .. ... ... ... pg:3 ..... rd:5 \ - &rpri_esz rn=%reg_movprfx esz=%tszimm_esz +# element size encoded as TSZHL. +@rdn_pg_tszimm_shl ........ .. ... ... ... pg:3 ..... rd:5 \ + &rpri_esz rn=%reg_movprfx esz=%tszimm_esz imm=%tszimm_shl +@rdn_pg_tszimm_shr ........ .. ... ... ... pg:3 ..... rd:5 \ + &rpri_esz rn=%reg_movprfx esz=%tszimm_esz imm=%tszimm_shr # Similarly without predicate. -@rd_rn_tszimm ........ .. ... ... ...... rn:5 rd:5 \ - &rri_esz esz=%tszimm16_esz +@rd_rn_tszimm_shl ........ .. ... ... ...... rn:5 rd:5 \ + &rri_esz esz=%tszimm16_esz imm=%tszimm16_shl +@rd_rn_tszimm_shr ........ .. ... ... ...... rn:5 rd:5 \ + &rri_esz esz=%tszimm16_esz imm=%tszimm16_shr # Two register operand, one immediate operand, with 4-bit predicate. # User must fill in imm. @@ -XXX,XX +XXX,XX @@ UMINV 00000100 .. 001 011 001 ... ..... ..... @rd_pg_rn ### SVE Shift by Immediate - Predicated Group # SVE bitwise shift by immediate (predicated) -ASR_zpzi 00000100 .. 000 000 100 ... .. ... ..... \ - @rdn_pg_tszimm imm=%tszimm_shr -LSR_zpzi 00000100 .. 000 001 100 ... .. ... ..... \ - @rdn_pg_tszimm imm=%tszimm_shr -LSL_zpzi 00000100 .. 000 011 100 ... .. ... ..... \ - @rdn_pg_tszimm imm=%tszimm_shl -ASRD 00000100 .. 000 100 100 ... .. ... ..... \ - @rdn_pg_tszimm imm=%tszimm_shr +ASR_zpzi 00000100 .. 000 000 100 ... .. ... ..... @rdn_pg_tszimm_shr +LSR_zpzi 00000100 .. 000 001 100 ... .. ... ..... @rdn_pg_tszimm_shr +LSL_zpzi 00000100 .. 000 011 100 ... .. ... ..... @rdn_pg_tszimm_shl +ASRD 00000100 .. 000 100 100 ... .. ... ..... @rdn_pg_tszimm_shr # SVE bitwise shift by vector (predicated) ASR_zpzz 00000100 .. 010 000 100 ... ..... ..... @rdn_pg_rm @@ -XXX,XX +XXX,XX @@ RDVL 00000100 101 11111 01010 imm:s6 rd:5 ### SVE Bitwise Shift - Unpredicated Group # SVE bitwise shift by immediate (unpredicated) -ASR_zzi 00000100 .. 1 ..... 1001 00 ..... ..... \ - @rd_rn_tszimm imm=%tszimm16_shr -LSR_zzi 00000100 .. 1 ..... 1001 01 ..... ..... \ - @rd_rn_tszimm imm=%tszimm16_shr -LSL_zzi 00000100 .. 1 ..... 1001 11 ..... ..... \ - @rd_rn_tszimm imm=%tszimm16_shl +ASR_zzi 00000100 .. 1 ..... 1001 00 ..... ..... @rd_rn_tszimm_shr +LSR_zzi 00000100 .. 1 ..... 1001 01 ..... ..... @rd_rn_tszimm_shr +LSL_zzi 00000100 .. 1 ..... 1001 11 ..... ..... @rd_rn_tszimm_shl # SVE bitwise shift by wide elements (unpredicated) # Note esz != 3 -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Unify add/sub helpers and add a parameter for rounding. This will allow saturating non-rounding to reuse this code. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> [PMM: fixed accidental use of '=' rather than '+=' in do_sqrdmlah_s] Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-15-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/vec_helper.c | 80 +++++++++++++++-------------------------- 1 file changed, 29 insertions(+), 51 deletions(-) diff --git a/target/arm/vec_helper.c b/target/arm/vec_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vec_helper.c +++ b/target/arm/vec_helper.c @@ -XXX,XX +XXX,XX @@ #endif /* Signed saturating rounding doubling multiply-accumulate high half, 16-bit */ -static int16_t inl_qrdmlah_s16(int16_t src1, int16_t src2, - int16_t src3, uint32_t *sat) +static int16_t do_sqrdmlah_h(int16_t src1, int16_t src2, int16_t src3, + bool neg, bool round, uint32_t *sat) { - /* Simplify: + /* + * Simplify: * = ((a3 << 16) + ((e1 * e2) << 1) + (1 << 15)) >> 16 * = ((a3 << 15) + (e1 * e2) + (1 << 14)) >> 15 */ int32_t ret = (int32_t)src1 * src2; - ret = ((int32_t)src3 << 15) + ret + (1 << 14); + if (neg) { + ret = -ret; + } + ret += ((int32_t)src3 << 15) + (round << 14); ret >>= 15; + if (ret != (int16_t)ret) { *sat = 1; - ret = (ret < 0 ? -0x8000 : 0x7fff); + ret = (ret < 0 ? INT16_MIN : INT16_MAX); } return ret; } @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(neon_qrdmlah_s16)(CPUARMState *env, uint32_t src1, uint32_t src2, uint32_t src3) { uint32_t *sat = &env->vfp.qc[0]; - uint16_t e1 = inl_qrdmlah_s16(src1, src2, src3, sat); - uint16_t e2 = inl_qrdmlah_s16(src1 >> 16, src2 >> 16, src3 >> 16, sat); + uint16_t e1 = do_sqrdmlah_h(src1, src2, src3, false, true, sat); + uint16_t e2 = do_sqrdmlah_h(src1 >> 16, src2 >> 16, src3 >> 16, + false, true, sat); return deposit32(e1, 16, 16, e2); } @@ -XXX,XX +XXX,XX @@ void HELPER(gvec_qrdmlah_s16)(void *vd, void *vn, void *vm, uintptr_t i; for (i = 0; i < opr_sz / 2; ++i) { - d[i] = inl_qrdmlah_s16(n[i], m[i], d[i], vq); + d[i] = do_sqrdmlah_h(n[i], m[i], d[i], false, true, vq); } clear_tail(d, opr_sz, simd_maxsz(desc)); } -/* Signed saturating rounding doubling multiply-subtract high half, 16-bit */ -static int16_t inl_qrdmlsh_s16(int16_t src1, int16_t src2, - int16_t src3, uint32_t *sat) -{ - /* Similarly, using subtraction: - * = ((a3 << 16) - ((e1 * e2) << 1) + (1 << 15)) >> 16 - * = ((a3 << 15) - (e1 * e2) + (1 << 14)) >> 15 - */ - int32_t ret = (int32_t)src1 * src2; - ret = ((int32_t)src3 << 15) - ret + (1 << 14); - ret >>= 15; - if (ret != (int16_t)ret) { - *sat = 1; - ret = (ret < 0 ? -0x8000 : 0x7fff); - } - return ret; -} - uint32_t HELPER(neon_qrdmlsh_s16)(CPUARMState *env, uint32_t src1, uint32_t src2, uint32_t src3) { uint32_t *sat = &env->vfp.qc[0]; - uint16_t e1 = inl_qrdmlsh_s16(src1, src2, src3, sat); - uint16_t e2 = inl_qrdmlsh_s16(src1 >> 16, src2 >> 16, src3 >> 16, sat); + uint16_t e1 = do_sqrdmlah_h(src1, src2, src3, true, true, sat); + uint16_t e2 = do_sqrdmlah_h(src1 >> 16, src2 >> 16, src3 >> 16, + true, true, sat); return deposit32(e1, 16, 16, e2); } @@ -XXX,XX +XXX,XX @@ void HELPER(gvec_qrdmlsh_s16)(void *vd, void *vn, void *vm, uintptr_t i; for (i = 0; i < opr_sz / 2; ++i) { - d[i] = inl_qrdmlsh_s16(n[i], m[i], d[i], vq); + d[i] = do_sqrdmlah_h(n[i], m[i], d[i], true, true, vq); } clear_tail(d, opr_sz, simd_maxsz(desc)); } /* Signed saturating rounding doubling multiply-accumulate high half, 32-bit */ -static int32_t inl_qrdmlah_s32(int32_t src1, int32_t src2, - int32_t src3, uint32_t *sat) +static int32_t do_sqrdmlah_s(int32_t src1, int32_t src2, int32_t src3, + bool neg, bool round, uint32_t *sat) { /* Simplify similarly to int_qrdmlah_s16 above. */ int64_t ret = (int64_t)src1 * src2; - ret = ((int64_t)src3 << 31) + ret + (1 << 30); + if (neg) { + ret = -ret; + } + ret += ((int64_t)src3 << 31) + (round << 30); ret >>= 31; + if (ret != (int32_t)ret) { *sat = 1; ret = (ret < 0 ? INT32_MIN : INT32_MAX); @@ -XXX,XX +XXX,XX @@ uint32_t HELPER(neon_qrdmlah_s32)(CPUARMState *env, int32_t src1, int32_t src2, int32_t src3) { uint32_t *sat = &env->vfp.qc[0]; - return inl_qrdmlah_s32(src1, src2, src3, sat); + return do_sqrdmlah_s(src1, src2, src3, false, true, sat); } void HELPER(gvec_qrdmlah_s32)(void *vd, void *vn, void *vm, @@ -XXX,XX +XXX,XX @@ void HELPER(gvec_qrdmlah_s32)(void *vd, void *vn, void *vm, uintptr_t i; for (i = 0; i < opr_sz / 4; ++i) { - d[i] = inl_qrdmlah_s32(n[i], m[i], d[i], vq); + d[i] = do_sqrdmlah_s(n[i], m[i], d[i], false, true, vq); } clear_tail(d, opr_sz, simd_maxsz(desc)); } -/* Signed saturating rounding doubling multiply-subtract high half, 32-bit */ -static int32_t inl_qrdmlsh_s32(int32_t src1, int32_t src2, - int32_t src3, uint32_t *sat) -{ - /* Simplify similarly to int_qrdmlsh_s16 above. */ - int64_t ret = (int64_t)src1 * src2; - ret = ((int64_t)src3 << 31) - ret + (1 << 30); - ret >>= 31; - if (ret != (int32_t)ret) { - *sat = 1; - ret = (ret < 0 ? INT32_MIN : INT32_MAX); - } - return ret; -} - uint32_t HELPER(neon_qrdmlsh_s32)(CPUARMState *env, int32_t src1, int32_t src2, int32_t src3) { uint32_t *sat = &env->vfp.qc[0]; - return inl_qrdmlsh_s32(src1, src2, src3, sat); + return do_sqrdmlah_s(src1, src2, src3, true, true, sat); } void HELPER(gvec_qrdmlsh_s32)(void *vd, void *vn, void *vm, @@ -XXX,XX +XXX,XX @@ void HELPER(gvec_qrdmlsh_s32)(void *vd, void *vn, void *vm, uintptr_t i; for (i = 0; i < opr_sz / 4; ++i) { - d[i] = inl_qrdmlsh_s32(n[i], m[i], d[i], vq); + d[i] = do_sqrdmlah_s(n[i], m[i], d[i], true, true, vq); } clear_tail(d, opr_sz, simd_maxsz(desc)); } -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-19-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.h | 4 ++++ target/arm/translate-a64.c | 16 ++++++++++++++++ target/arm/vec_helper.c | 29 +++++++++++++++++++++++++---- 3 files changed, 45 insertions(+), 4 deletions(-) diff --git a/target/arm/helper.h b/target/arm/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.h +++ b/target/arm/helper.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(gvec_uaba_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(gvec_uaba_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(gvec_uaba_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_mul_idx_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_mul_idx_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(gvec_mul_idx_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + #ifdef TARGET_AARCH64 #include "helper-a64.h" #include "helper-sve.h" diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_indexed(DisasContext *s, uint32_t insn) data, gen_helper_gvec_fmlal_idx_a64); } return; + + case 0x08: /* MUL */ + if (!is_long && !is_scalar) { + static gen_helper_gvec_3 * const fns[3] = { + gen_helper_gvec_mul_idx_h, + gen_helper_gvec_mul_idx_s, + gen_helper_gvec_mul_idx_d, + }; + tcg_gen_gvec_3_ool(vec_full_reg_offset(s, rd), + vec_full_reg_offset(s, rn), + vec_full_reg_offset(s, rm), + is_q ? 16 : 8, vec_full_reg_size(s), + index, fns[size - 1]); + return; + } + break; } if (size == 3) { diff --git a/target/arm/vec_helper.c b/target/arm/vec_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vec_helper.c +++ b/target/arm/vec_helper.c @@ -XXX,XX +XXX,XX @@ DO_3OP(gvec_rsqrts_d, helper_rsqrtsf_f64, float64) */ #define DO_MUL_IDX(NAME, TYPE, H) \ +void HELPER(NAME)(void *vd, void *vn, void *vm, uint32_t desc) \ +{ \ + intptr_t i, j, oprsz = simd_oprsz(desc), segment = 16 / sizeof(TYPE); \ + intptr_t idx = simd_data(desc); \ + TYPE *d = vd, *n = vn, *m = vm; \ + for (i = 0; i < oprsz / sizeof(TYPE); i += segment) { \ + TYPE mm = m[H(i + idx)]; \ + for (j = 0; j < segment; j++) { \ + d[i + j] = n[i + j] * mm; \ + } \ + } \ + clear_tail(d, oprsz, simd_maxsz(desc)); \ +} + +DO_MUL_IDX(gvec_mul_idx_h, uint16_t, H2) +DO_MUL_IDX(gvec_mul_idx_s, uint32_t, H4) +DO_MUL_IDX(gvec_mul_idx_d, uint64_t, ) + +#undef DO_MUL_IDX + +#define DO_FMUL_IDX(NAME, TYPE, H) \ void HELPER(NAME)(void *vd, void *vn, void *vm, void *stat, uint32_t desc) \ { \ intptr_t i, j, oprsz = simd_oprsz(desc), segment = 16 / sizeof(TYPE); \ @@ -XXX,XX +XXX,XX @@ void HELPER(NAME)(void *vd, void *vn, void *vm, void *stat, uint32_t desc) \ clear_tail(d, oprsz, simd_maxsz(desc)); \ } -DO_MUL_IDX(gvec_fmul_idx_h, float16, H2) -DO_MUL_IDX(gvec_fmul_idx_s, float32, H4) -DO_MUL_IDX(gvec_fmul_idx_d, float64, ) +DO_FMUL_IDX(gvec_fmul_idx_h, float16, H2) +DO_FMUL_IDX(gvec_fmul_idx_s, float32, H4) +DO_FMUL_IDX(gvec_fmul_idx_d, float64, ) -#undef DO_MUL_IDX +#undef DO_FMUL_IDX #define DO_FMLA_IDX(NAME, TYPE, H) \ void HELPER(NAME)(void *vd, void *vn, void *vm, void *va, \ -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-20-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.h | 14 ++++++++++++++ target/arm/translate-a64.c | 34 ++++++++++++++++++++++++++++++++++ target/arm/vec_helper.c | 25 +++++++++++++++++++++++++ 3 files changed, 73 insertions(+) diff --git a/target/arm/helper.h b/target/arm/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.h +++ b/target/arm/helper.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_4(gvec_mul_idx_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(gvec_mul_idx_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(gvec_mul_idx_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_mla_idx_h, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_mla_idx_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_mla_idx_d, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_5(gvec_mls_idx_h, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_mls_idx_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(gvec_mls_idx_d, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) + #ifdef TARGET_AARCH64 #include "helper-a64.h" #include "helper-sve.h" diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void disas_simd_indexed(DisasContext *s, uint32_t insn) return; } break; + + case 0x10: /* MLA */ + if (!is_long && !is_scalar) { + static gen_helper_gvec_4 * const fns[3] = { + gen_helper_gvec_mla_idx_h, + gen_helper_gvec_mla_idx_s, + gen_helper_gvec_mla_idx_d, + }; + tcg_gen_gvec_4_ool(vec_full_reg_offset(s, rd), + vec_full_reg_offset(s, rn), + vec_full_reg_offset(s, rm), + vec_full_reg_offset(s, rd), + is_q ? 16 : 8, vec_full_reg_size(s), + index, fns[size - 1]); + return; + } + break; + + case 0x14: /* MLS */ + if (!is_long && !is_scalar) { + static gen_helper_gvec_4 * const fns[3] = { + gen_helper_gvec_mls_idx_h, + gen_helper_gvec_mls_idx_s, + gen_helper_gvec_mls_idx_d, + }; + tcg_gen_gvec_4_ool(vec_full_reg_offset(s, rd), + vec_full_reg_offset(s, rn), + vec_full_reg_offset(s, rm), + vec_full_reg_offset(s, rd), + is_q ? 16 : 8, vec_full_reg_size(s), + index, fns[size - 1]); + return; + } + break; } if (size == 3) { diff --git a/target/arm/vec_helper.c b/target/arm/vec_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vec_helper.c +++ b/target/arm/vec_helper.c @@ -XXX,XX +XXX,XX @@ DO_MUL_IDX(gvec_mul_idx_d, uint64_t, ) #undef DO_MUL_IDX +#define DO_MLA_IDX(NAME, TYPE, OP, H) \ +void HELPER(NAME)(void *vd, void *vn, void *vm, void *va, uint32_t desc) \ +{ \ + intptr_t i, j, oprsz = simd_oprsz(desc), segment = 16 / sizeof(TYPE); \ + intptr_t idx = simd_data(desc); \ + TYPE *d = vd, *n = vn, *m = vm, *a = va; \ + for (i = 0; i < oprsz / sizeof(TYPE); i += segment) { \ + TYPE mm = m[H(i + idx)]; \ + for (j = 0; j < segment; j++) { \ + d[i + j] = a[i + j] OP n[i + j] * mm; \ + } \ + } \ + clear_tail(d, oprsz, simd_maxsz(desc)); \ +} + +DO_MLA_IDX(gvec_mla_idx_h, uint16_t, +, H2) +DO_MLA_IDX(gvec_mla_idx_s, uint32_t, +, H4) +DO_MLA_IDX(gvec_mla_idx_d, uint64_t, +, ) + +DO_MLA_IDX(gvec_mls_idx_h, uint16_t, -, H2) +DO_MLA_IDX(gvec_mls_idx_s, uint32_t, -, H4) +DO_MLA_IDX(gvec_mls_idx_d, uint64_t, -, ) + +#undef DO_MLA_IDX + #define DO_FMUL_IDX(NAME, TYPE, H) \ void HELPER(NAME)(void *vd, void *vn, void *vm, void *stat, uint32_t desc) \ { \ -- 2.20.1 From: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Message-id: 20200815013145.539409-21-richard.henderson@linaro.org Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- target/arm/helper.h | 10 ++++++++ target/arm/translate-a64.c | 33 ++++++++++++++++++-------- target/arm/vec_helper.c | 48 ++++++++++++++++++++++++++++++++++++++ 3 files changed, 81 insertions(+), 10 deletions(-) diff --git a/target/arm/helper.h b/target/arm/helper.h index XXXXXXX..XXXXXXX 100644 --- a/target/arm/helper.h +++ b/target/arm/helper.h @@ -XXX,XX +XXX,XX @@ DEF_HELPER_FLAGS_5(gvec_mls_idx_s, TCG_CALL_NO_RWG, DEF_HELPER_FLAGS_5(gvec_mls_idx_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(neon_sqdmulh_h, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(neon_sqdmulh_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_5(neon_sqrdmulh_h, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(neon_sqrdmulh_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) + #ifdef TARGET_AARCH64 #include "helper-a64.h" #include "helper-sve.h" diff --git a/target/arm/translate-a64.c b/target/arm/translate-a64.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/translate-a64.c +++ b/target/arm/translate-a64.c @@ -XXX,XX +XXX,XX @@ static void gen_gvec_op3_fpst(DisasContext *s, bool is_q, int rd, int rn, tcg_temp_free_ptr(fpst); } +/* Expand a 3-operand + qc + operation using an out-of-line helper. */ +static void gen_gvec_op3_qc(DisasContext *s, bool is_q, int rd, int rn, + int rm, gen_helper_gvec_3_ptr *fn) +{ + TCGv_ptr qc_ptr = tcg_temp_new_ptr(); + + tcg_gen_addi_ptr(qc_ptr, cpu_env, offsetof(CPUARMState, vfp.qc)); + tcg_gen_gvec_3_ptr(vec_full_reg_offset(s, rd), + vec_full_reg_offset(s, rn), + vec_full_reg_offset(s, rm), qc_ptr, + is_q ? 16 : 8, vec_full_reg_size(s), 0, fn); + tcg_temp_free_ptr(qc_ptr); +} + /* Set ZF and NF based on a 64 bit result. This is alas fiddlier * than the 32 bit equivalent. */ @@ -XXX,XX +XXX,XX @@ static void disas_simd_3same_int(DisasContext *s, uint32_t insn) gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_mla, size); } return; + case 0x16: /* SQDMULH, SQRDMULH */ + { + static gen_helper_gvec_3_ptr * const fns[2][2] = { + { gen_helper_neon_sqdmulh_h, gen_helper_neon_sqrdmulh_h }, + { gen_helper_neon_sqdmulh_s, gen_helper_neon_sqrdmulh_s }, + }; + gen_gvec_op3_qc(s, is_q, rd, rn, rm, fns[size - 1][u]); + } + return; case 0x11: if (!u) { /* CMTST */ gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_cmtst, size); @@ -XXX,XX +XXX,XX @@ static void disas_simd_3same_int(DisasContext *s, uint32_t insn) genenvfn = fns[size][u]; break; } - case 0x16: /* SQDMULH, SQRDMULH */ - { - static NeonGenTwoOpEnvFn * const fns[2][2] = { - { gen_helper_neon_qdmulh_s16, gen_helper_neon_qrdmulh_s16 }, - { gen_helper_neon_qdmulh_s32, gen_helper_neon_qrdmulh_s32 }, - }; - assert(size == 1 || size == 2); - genenvfn = fns[size - 1][u]; - break; - } default: g_assert_not_reached(); } diff --git a/target/arm/vec_helper.c b/target/arm/vec_helper.c index XXXXXXX..XXXXXXX 100644 --- a/target/arm/vec_helper.c +++ b/target/arm/vec_helper.c @@ -XXX,XX +XXX,XX @@ void HELPER(gvec_qrdmlsh_s16)(void *vd, void *vn, void *vm, clear_tail(d, opr_sz, simd_maxsz(desc)); } +void HELPER(neon_sqdmulh_h)(void *vd, void *vn, void *vm, + void *vq, uint32_t desc) +{ + intptr_t i, opr_sz = simd_oprsz(desc); + int16_t *d = vd, *n = vn, *m = vm; + + for (i = 0; i < opr_sz / 2; ++i) { + d[i] = do_sqrdmlah_h(n[i], m[i], 0, false, false, vq); + } + clear_tail(d, opr_sz, simd_maxsz(desc)); +} + +void HELPER(neon_sqrdmulh_h)(void *vd, void *vn, void *vm, + void *vq, uint32_t desc) +{ + intptr_t i, opr_sz = simd_oprsz(desc); + int16_t *d = vd, *n = vn, *m = vm; + + for (i = 0; i < opr_sz / 2; ++i) { + d[i] = do_sqrdmlah_h(n[i], m[i], 0, false, true, vq); + } + clear_tail(d, opr_sz, simd_maxsz(desc)); +} + /* Signed saturating rounding doubling multiply-accumulate high half, 32-bit */ static int32_t do_sqrdmlah_s(int32_t src1, int32_t src2, int32_t src3, bool neg, bool round, uint32_t *sat) @@ -XXX,XX +XXX,XX @@ void HELPER(gvec_qrdmlsh_s32)(void *vd, void *vn, void *vm, clear_tail(d, opr_sz, simd_maxsz(desc)); } +void HELPER(neon_sqdmulh_s)(void *vd, void *vn, void *vm, + void *vq, uint32_t desc) +{ + intptr_t i, opr_sz = simd_oprsz(desc); + int32_t *d = vd, *n = vn, *m = vm; + + for (i = 0; i < opr_sz / 4; ++i) { + d[i] = do_sqrdmlah_s(n[i], m[i], 0, false, false, vq); + } + clear_tail(d, opr_sz, simd_maxsz(desc)); +} + +void HELPER(neon_sqrdmulh_s)(void *vd, void *vn, void *vm, + void *vq, uint32_t desc) +{ + intptr_t i, opr_sz = simd_oprsz(desc); + int32_t *d = vd, *n = vn, *m = vm; + + for (i = 0; i < opr_sz / 4; ++i) { + d[i] = do_sqrdmlah_s(n[i], m[i], 0, false, true, vq); + } + clear_tail(d, opr_sz, simd_maxsz(desc)); +} + /* Integer 8 and 16-bit dot-product. * * Note that for the loops herein, host endianness does not matter -- 2.20.1